Multi-Scale Copper–Cobalt-Supported Carbon Catalysts for Efficient CO₂ and O₂ Reduction

Abstract

A sequenced ultrasonic atomization coupled with a pyrolysis process is developed to synthesize a series of cross-scale (Co/Cu)-NC catalysts. The catalysts demonstrate high metal utilization efficiency with a metal loading of 22.45 ± 0.07 wt%. Electrochemical evaluations for the oxygen reduction reaction (ORR) suggest that the best (Co/Cu)-NC catalysts are prepared with a Co/Cu ratio of 1/1 and a calcination temperature of 800 °C, which achieve a half-wave potential of 0.87 V and an electrochemical impedance spectroscopy semicircle radius as low as 30 ohms. Linear sweep voltammetry measurements indicate that (Co/Cu)-NC catalysts exhibit the highest current density. Under a potential of −0.73 V versus the reversible hydrogen electrode, (Co/Cu)-NC catalysts demonstrate long-term stability with the CO Faradaic efficiency of about 70% for catalyzing carbon dioxide reduction reaction (CO₂RR). Overall, the above metrics identify CoCu-800 as the optimal bifunctional catalyst among the tested samples for ORR and CO₂RR under the investigated conditions.

1. Introduction

In recent years, the increasing demand for energy and the rapid depletion of traditional fossil fuels have had significant adverse effects on environment. To address these challenges, transitioning to alternative, renewable, and clean energy sources has become a critical strategy. As part of this transition, efficient and environmentally friendly fuel cells have attracted considerable attention [1,2,3,4,5]. Notably, clean fuel cells, or hydrogen fuel cells, significantly rely on the oxygen reduction reaction (ORR) as a crucial process [6]. In parallel, the excessive use of fossil fuels and alcohol fuel cells has dramatically increased carbon dioxide emissions, exacerbating global warming and air pollution [7]. Techniques to recover and convert carbon dioxide are therefore essential to mitigate these pressing environmental issues. Among various approaches, electrocatalytic carbon dioxide reduction reaction (CO₂RR) stands out for its environmental compatibility and energy efficiency when compared with other methods (e.g., photocatalysis and thermos-catalysis) [8]. As a result, developing high-performance multifunctional catalysts capable of simultaneously facilitating both ORR and CO₂RR has become a pressing priority.

Platinum-group noble metals are well known as benchmark catalysts for ORR [9,10,11,12]. However, their high cost and limited scalability significantly restrict their widespread commercial use. As an alternative, transition metals (TMs), particularly iron-group elements, have drawn increasing attention as non-noble or non-precious metal catalysts due to their abundance and low cost. Among them, cobalt (Co)-based materials are among the most extensively studied TM-based catalysts for electrocatalytic ORR because of their ability to effectively adsorb key reactants, such as oxygen (O₂), during the reaction process [13,14,15]. More importantly, Co can be stabilized as Co-N_x motifs in N-doped carbon after pyrolysis. Such Co-N-C catalysts have been widely recognized to deliver meaningful ORR activity and durability, providing a realistic non-noble pathway toward practical devices such as alkaline fuel cells and Zn-air batteries [15,16]. In addition, recent studies have demonstrated that copper (Cu)-containing catalysts and Cu-involved composites also possess notable ORR catalytic activity, highlighting their potential as another kind of viable TM catalyst [17,18,19,20,21]. For example, Cu(I)-N coordinated sites and atomically dispersed Cu-N-C configurations have been reported to exhibit competitive ORR kinetics in alkaline media, supporting Cu as an effective and earth-abundant alternative component [22,23]. Notably, combining Co and Cu in a single N-doped carbon host is further motivated by the possibility of Co-Cu electronic and coordination coupling: Cu can modulate the local nitrogen configuration and metal electronic structure, while Co contributes strong ORR active Co-N_x motifs, together offering a tunable platform to balance activity and durability beyond single-metal systems [24]. This shift from noble metals to non-noble or non-precious TMs represents a critical step toward cost-effective and scalable ORR catalyst development.

CO₂RR, simultaneously, faces tremendous challenges, including high activation energy requirements and the generation of diverse, complex products. Among these complex products, carbon monoxide (CO) has gained considerable attention due to its high economic value [25,26]. As a result, the development of catalysts with both high activity and excellent CO selectivity has become a critical focus of research [27,28,29,30]. Among the various transition metals, Cu-based catalysts are of considerable interest due to their unique characteristics, such as a high electrochemically active surface area (ECSA) and an optimized local electronic structure [31]. However, due to the unique electronic properties of Cu, its catalytic products are diverse, ranging from C₁ products like formic acid and CO to C₂ products such as ethylene and ethanol [32]. Bimetallic catalysts are widely used to enhance the catalytic performance of the CO₂RR, typically through structural optimization and tuning of surface acidic sites [33,34,35]. The reduction of CO₂ to CO generally involves the formation of a *COOH intermediate, which then transforms into *CO and ultimately produces CO [36]. Studies show that incorporating Co into Cu-based catalysts effectively modulates the adsorption strength of intermediates, leading to stronger binding of *COOH and weaker binding of *CO [37,38,39,40]. This modulation promotes the formation of *COOH and the desorption of CO, thereby accelerating CO production and improving the catalyst’s selectivity for CO.

Up to now, significant attention has been given to the development of advanced catalysts [41], particularly those involving single-atom sites [42], sub-nanometer clusters (especially multi-element superatoms) [43], and nano-heterostructures [44]. These materials are believed to have great potential for enhancing catalytic efficiency and metal utilization. Single-atom catalysts offer high metal utilization and activity per metal atom [45], but they suffer from low density and significant metal loss during preparation [46], especially during acid washing processes. This results in reduced metal utilization in large-scale commercial applications [47]. In contrast, nanoparticles larger than 3 nm contain a smaller proportion of surface atoms with catalytic potential, which limits their effectiveness [48]. Sub-nanometer superatoms offer a promising alternative due to their high surface atom ratio and enhanced defect sites, which increase the density of active sites [49]. However, controlling the atomic structure, microstructure, and electronic properties of these materials, especially in multi-element systems, remains a challenge [50]. Current research in this area is still in its early stages. Additionally, achieving pure superatoms often leads to significant metal loss during post-synthesis processing, further reducing metal utilization in large-scale applications [51].

Therefore, we designed multi-scale copper–cobalt-supported carbon catalysts with the aim of achieving near-zero reactant waste during synthesis while maintaining high catalytic performance. Carbon materials are widely used as catalyst supports because of their low cost and high electrical conductivity. During processing, surface defects and unsaturated bonds can be introduced and serve as anchoring sites for dispersed and stable metal species. During thermal treatment, some metal elements react with carbon to form metal carbides, strengthening the interaction between the metal and the carbon support. This enhanced interaction not only stabilizes the catalysts but also improves their overall performance [52]. By integrating the advantages of CoCu bimetallic catalysts and carbon-based supports, this design offers a promising approach for achieving efficient and scalable ORR and CO₂-to-CO conversion.

This study synthesized (Co/Cu)-NC catalysts with varying calcination temperatures and metal ratios using a modified sequential ultrasonic atomization and high-temperature pyrolysis process [53,54,55]. By integrating ultrasonic atomization with controlled hydrolysis conditions, this method not only achieves uniform particle distribution and dispersion but also minimizes metal precursor loss, offering a more sustainable synthesis route compared to conventional methods. Electrocatalytic evaluations revealed that the best catalysts can be obtained using a Co/Cu ratio of 1/1 and an annealing temperature of 800 °C after optimization of microstructures and synthesis kinetics, exhibiting superior ORR activity compared to other samples and commercial Pt/C. Furthermore, this catalyst demonstrated an impressive CO Faradaic efficiency (FE) of approximately 70% during CO₂RR, and outstanding stability. These findings highlight the potential of (Co/Cu)-NC as a scalable and multifunctional catalyst platform for ORR and CO₂RR.

2. Materials and Methods

The (Co/Cu)-NC catalysts were synthesized via an ultrasonic atomization coupled with a temperature- and atmosphere-controlled pyrolysis process (Figure 1). Details of chemicals, synthesis, and electrochemical measurements of (Co/Cu)-NC catalysts are also included in the Supporting Information.

Figure 1. Scheme of the ultrasonic atomization coupled with a pyrolysis process to synthesize (Co/Cu)-NC catalysts is a figure. Schemes follow the same formatting.

First, the alkaline solution is atomized into micron-sized droplets using ultrasonic atomization and sprayed into a salt solution to improve reaction uniformity. After centrifugation to obtain the gel, 1,10-phenanthroline and carbon black are added and stirred to ensure uniform mixing. The mixture then undergoes centrifugation, washing, and drying. Finally, the sample is stabilized by high-temperature calcination, yielding the desired material. In this experiment, to determine the optimal calcination temperature and element ratio for preparing (Co/Cu)-NC materials, samples with various Co/Cu ratios were synthesized at different calcination temperatures.

Samples are denoted as Co_mCu_n to indicate the precursor Co/Cu ratio of m/n (with CoCu representing 1/1 for simplicity). For the calcination-temperature series at Co/Cu = 1/1, samples are further labeled as CoCu-T, where T = 0, 400, 800, and 1000 denote the calcination temperature (°C); T = 0 refers to the uncalcined (dried) precursor.

3. Results

In this section, three samples with Co/Cu = 1/1, 3/7, and 7/3 (all annealed at 800 °C) were selected for detailed characterization. They are denoted as CoCu, Co₃Cu₇, and Co₇Cu₃.

3.1. Fundamental Characterization of the Catalysts

Scanning electron microscopy (SEM) images (Figure S1a,b) and transmission electron microscopy (TEM) images (Figure 2a,b) reveal the presence of relatively large metal particles with diameters around 300 nm in the (Co/Cu)-NC materials. Elemental mapping of two representative particles via energy-dispersive X-ray spectroscopy (EDS) (Figure S1c) indicates that these particles are primarily composed of copper, with a minor presence of cobalt. This phenomenon is attributed to the aggregation of metal species during high-temperature annealing, where the mobility of metal atoms leads to the formation of large particles. Previous studies have reported that, compared to cobalt, copper exhibits a higher tendency to agglomerate during pyrolysis [37], which is consistent with the observations in this work. Moreover, considering the efficient utilization of metal resources and the feasibility for large-scale production, no additional treatment was applied to remove the large particles. This approach minimizes metal waste and may allow the larger Cu-rich particles to provide auxiliary electrochemical contributions, primarily by offering additional conductive pathways and facilitating interfacial charge transfer, thereby improving the utilization of the highly dispersed sub-nanometer/nanometer interfaces.

Figure 2. (a,b) TEM images of large particles in the CoCu sample.

TEM images (Figure 3) suggest that, in addition to the presence of large metal particles, the (Co/Cu)-NC materials contain a population of superatoms with diameters around 1 nm (many less than 1.0 nm: red circles; or some at most 2.0 nm) (Figure 3b,f) and a few nanoparticles with diameters of approximately 3–5 nm (yellow circles) (Figure 3c,d). High-resolution transmission electron microscopy (HR-TEM) images further reveal that some nanoparticles are free, while others are supported on the assemblies (Figure 3f). Precise lattice fringe measurements, along with X-ray diffraction (XRD) results (Figure 4g) and comparison with PDF standard reference cards, indicate that the lattice plane with d = 2.04 Å corresponds to the (111) plane of Co (JCPDS:89-7093), while the lattice plane with d = 2.09 Å corresponds to the (111) plane of Cu (JCPDS:70-3038). EDS line scans were collected across representative nanoparticles (Figure 4a,b) as well as cluster-rich, sub-nanometer-dense regions (Figure 4c,d). Both Co and Cu signals are observed in these regions, with Cu generally exhibiting higher intensity. Here, the line scans are intended to provide qualitative or semi-quantitative evidence of co-localization and relative trends rather than rigorous, particle-level Cu/Co quantification. Consistently, EDS elemental maps (Figure 4e) reveal Co-Cu co-localization on the carbon support with an overall Cu-enriched distribution, which may arise from synthesis conditions favoring Cu deposition and partial Co loss or incomplete Co reduction during processing. Statistical analysis of >100 features yields an average diameter of 1.4 ± 0.9 nm (Figure 4f). Overall, the catalyst exhibits a hierarchical multi-scale architecture comprising cluster-rich sub-nanometer domains, nanometer-scale nanoparticles, and occasional larger Cu-rich particles dispersed on an N-doped carbon framework. These motifs are better viewed as a cooperative system rather than simply additive components: the sub-nanometer/nanometer features provide highly accessible interfacial sites, whereas the carbon framework and larger-scale motifs mainly help maintain electronic/ionic transport continuity and facilitate interfacial charge transfer, thereby improving the utilization of the small-scale interfaces.

Figure 3. (a) Wide-viewed TEM images of region 1 of the CoCu sample, (b,c) the corresponding HR-TEM images of region 1 of the CoCu sample, (d,e) wide-viewed TEM images of region 2 of the CoCu sample, (f) the corresponding HR-TEM images of region 2 of the CoCu sample. (Dashed boxes: enlarged areas; yellow circles: nanoparticles; red circles: superatoms).

Figure 4. (a) TEM image of an individual nanoparticle in the CoCu sample, (b) its corresponding EDS element line scanning, (c) TEM image of super-atom clusters in the CoCu sample, (d) its corresponding EDS element line scanning, and (e) STEM image of the CoCu sample with corresponding EDS elemental distribution, (f) statistical distribution of particle size, (g) XRD patterns of CoCu, Co₃Cu₇, and Co₇Cu₃ samples.

Subsequently, inductively coupled plasma optical emission spectrometry (ICP-OES) was conducted to quantitatively analyze the metal content in the (Co/Cu)-NC samples (Table S1). To ensure the accuracy and reproducibility of the results, each element was measured three times. The results show that the mass fractions of Cu and Co are 16.75 ± 0.05 wt% and 5.70 ± 0.02 wt%, respectively. These values are in good agreement with the morphological and structural analyses obtained from SEM, TEM, and XRD, further confirming the presence and distribution states of the metal species in the material. Notably, the total metal loading reaches 22.45 ± 0.07 wt%, indicating a highly efficient utilization of metal precursors and a substantial reduction in metal loss during the synthesis process. In contrast, conventional single-atom catalysts (SACs) generally suffer from low metal loading and inefficient precursor utilization, which leads to significant metal waste and limits their feasibility for practical industrial-scale applications. Therefore, the synthetic strategy developed in this work not only improves metal utilization but also offers excellent scalability and potential for real-world deployment.

The crystal phases of CoCu, Co₃Cu₇, and Co₇Cu₃ samples were characterized using XRD, with the results shown in Figure 4g. For all samples, a broad diffraction peak associated with the graphite (002) plane is observed at around 25.6°. Characteristic peaks of metallic Cu are observed at 43.32°, 50.44°, and 74.12°, corresponding to the (111), (200), and (220) crystal planes, respectively. Similarly, metallic Co peaks are identified at 44.35°, 51.67°, and 75.98°, associated with the same crystal planes. Additionally, the peaks corresponding to metallic Cu and metallic Co appear separately. Analysis of the binary phase diagram of the Co-Cu alloy reveals that Co-rich and Cu-rich phases exhibit phase separation over a wide range of temperatures and compositions [56]. This suggests that Co and Cu have limited miscibility and do not readily form a homogeneous alloy. Consequently, a Co-Cu alloy is not formed in the investigated material.

Raman spectroscopy was conducted on the above samples, with the results presented in Figure S2. Two peaks at ~1341 cm⁻¹ and 1600 cm⁻¹ correspond to the D-band and G-band of graphite, respectively. The I_D/I_G ratio reflects the structural disorder in graphite, which is increased with the presence of more and more defects or impurities [57,58]. The I_D/I_G ratios are 1.133, 1.220, and 1.323 for the CoCu, Co₃Cu₇, and Co₇Cu₃ catalysts, respectively. A lowest I_D/I_G ratio in the CoCu sample indicates this sample has the least concentration of defects or impurities [59], which favors increasing its electrical conductivity. Moderate defect levels can increase the number of active sites, enhancing catalytic performance, whereas excessive defects may reduce conductivity and structural stability, ultimately lowering efficiency [60,61]. Importantly, the variation of I_D/I_G among different Cu/Co ratios suggests that adjusting the Cu-Co composition is accompanied by a concurrent modulation of the carbon microstructure/defect environment rather than merely changing metal contents. Such a coupled change in the conductive/defect framework can influence interfacial charge transfer and the accessibility/utilization of metal-associated sites, providing a structural basis for the Cu-Co synergy observed electrochemically. Based on our findings in this article, the CoCu sample may appear to achieve the optimized balance between defect density and electrical conductivity, resulting in the most favorable overall catalytic performance among the three samples. This Raman-based trend alone, however, should be viewed as a supportive descriptor rather than a standalone proof of activity, and the catalytic ranking is ultimately established by the electrochemical metrics.

To investigate the elemental composition and electronic status of key elements, X-ray photoelectron spectroscopy (XPS) was performed for the CoCu (Figure 5), Co₃Cu₇ (Figure S3), and Co₇Cu₃ (Figure S4) samples. The XPS spectra identify all the constituent elements (C, N, O, Co, and Cu) in the three samples. The high-resolution XPS spectra of the five elements were analyzed sequentially via the theoretical area fitting process for different element valences and/or possible species. The C 1s XPS spectrum reveals four bonding states—C-C, C-N, C-Co/Cu, and C-O. These bonds are attributed to the addition of carbon black and 1,10-phenanthroline in the preparation and to the formation of bonds between carbon atoms and metal elements during pyrolysis. The high-resolution O 1s spectrum is fitted by three peaks: C-O, O-H and M-O (M = Co, Cu). For Co 2p, the main features at ~780.5 eV (Co 2p^3/2) and ~795.6 eV (Co 2p^1/2) are accompanied by additional multiplet/satellite contributions. Notably, the Co 2p^1/2 region often appears broader than the 2p^3/2 region because it represents an envelope of overlapping Co²⁺/Co³⁺ contributions together with multiplet splitting and shake-up satellites. For Cu 2p, the spectra exhibit the expected spin–orbit doublets, with the low-binding-energy components (e.g., ~933.1 eV for 2p^3/2 and ~952.7 eV for 2p^1/2) attributable to Cu⁰/Cu⁺, while the higher-binding-energy components (e.g., ~934.7 eV and ~954.1 eV) are consistent with Cu²⁺ related contributions. Where applicable, Cu²⁺ shake-up satellite features were also included and discussed to support the presence of Cu²⁺. The N 1s spectra were deconvoluted into pyridinic, pyrrolic, and graphitic N, and their proportions are summarized in Table 1 [62]. The pyridinic-N fraction decreases from 48.2% (Co₇Cu₃) to 33.3% (CoCu) and 30.4% (Co₃Cu₇). Therefore, the higher pyridinic-N fraction in the Co-rich sample implies more available M-N (M = Co/Cu) coordination sites and a modified local electronic environment because the lone-pair electrons of pyridinic N are more accessible for metal coordination [63]. Meanwhile, the Cu 2p line-shape comparison suggests that Co incorporation is accompanied by a redistribution of Cu surface chemical states, consistent with a more Cu⁰/Cu⁺-dominated character and an attenuated Cu²⁺-related signature in the Co-rich samples. Overall, N 1s together with Co 2p/Cu 2p indicate that varying the Cu/Co ratio simultaneously tunes the N-coordination environment and Cu electronic structure, thereby affecting interfacial charge transfer/kinetics in ORR and CO₂RR.

Figure 5. (a) XPS survey spectrum of the CoCu sample, (b) high-resolution C 1s spectrum, (c) high-resolution N 1s spectrum, (d) high-resolution O 1s spectrum, (e) high-resolution Co 2p spectrum, (f) high-resolution Cu 2p spectrum.

Table 1. Proportions of Pyridinic Nitrogen, Pyrrolic Nitrogen, and Graphitic Nitrogen in CoCu, Co₃Cu₇, and Co₇Cu₃ samples.

Sample	Pyridinic Nitrogen (%)	Pyrrolic Nitrogen (%)	Graphitic Nitrogen (%)
CoCu	33.3	40.4	26.3
Co3Cu7	30.4	51.8	17.9
Co7Cu3	48.2	32.8	19.0

3.2. Electrocatalytic Activity for the ORR

ORR electrocatalytic performance of the prepared samples was evaluated using a three-electrode system. The calcination temperatures for the enhanced catalytic performance were first optimized using samples with a constant Co/Cu ratio of 1/1. These samples were subjected to calcination at temperatures of 400 °C, 800 °C, and 1000 °C. Their ORR catalytic performance was evaluated.

As shown in Figure 6, the linear sweep voltammetry (LSV) curves (Figure 6a) demonstrate that the unannealed sample and the samples annealed at 400 °C, 800 °C, and 1000 °C achieve half-wave potentials of 0.58 V, 0.53 V, 0.87 V and 0.71 V, respectively. Notably, only the 800 °C-annealed sample outperforms the commercial Pt/C catalyst, which has a half-wave potential of 0.85 V (Figure 6b). Furthermore, a lower Tafel slope indicates a less pronounced change in overpotential with increasing current density, suggesting enhanced electrocatalytic performance. The catalysts annealed at 800 °C (127.17 mV/dec) and 1000 °C (105.33 mV/dec) exhibit smaller Tafel slopes than the commercial Pt/C catalyst, indicating improved catalytic performance (Figure 6c). The electrochemical impedance spectroscopy (EIS) results indicate that, after calcination, the impedance arc radii of these samples are smaller than that of the commercial Pt/C catalyst. At calcination temperatures of 400 °C, 800 °C, and 1000 °C, the impedance radii are approximately 40 ohms, 30 ohms, and 35 ohms, respectively, with the smallest impedance radius observed at the 800 °C calcination temperature (Figure 6d). Thus, compared with Pt/C, the calcined samples exhibit more favorable interfacial charge-transfer/transport behavior under the investigated conditions, and CoCu-800 further shows the most pronounced transport advantage within the temperature series.

Figure 6. Electrocatalytic performance for ORR of CoCu samples calcination at different temperatures: (a) LSV curves; (b) half-wave potential comparison; (c) Tafel slopes; (d) EIS results.

Considering these results, the 800 °C-annealed sample exhibits superior electrocatalytic activity and conductivity. It shows the highest half-wave potential, the smallest impedance radius, and a lower Tafel slope. These results surpass those of samples calcined at other temperatures and the commercial Pt/C nanocatalyst, highlighting its enhanced electrocatalytic performance. Thus, 800 °C is the most suitable calcination temperature. Furthermore, our study demonstrates that when synthesized at an optimal calcination temperature, multi-scale structures achieve superior metal utilization and significantly improve electrocatalytic performance through a simple and cost-effective method, achieving better performance than Pt/C catalyst.

Then, to determine the optimal compositional ratio for ORR, Co_xCu_y-NC samples were synthesized with x/y ratios of 0/1, 1/9, 3/7, 7/3, 9/1, and 1/0 at a calcination temperature of 800 °C. The electrochemical performance of these samples is summarized in Figure 7. Analysis of the LSV curves (Figure 7a) shows that the half-wave potentials are 0.73 V for the Cu-NC sample, 0.78 V for the Co₁Cu₉ sample, 0.8 V for the Co₃Cu₇ sample, 0.82 V for the Co₇Cu₃ sample, 0.77 V for the Co₉Cu₁ sample, and 0.63 V for the Co-NC sample (Figure 7b), suggesting that samples with the CoCu ratio of 1/1 have the best performance. Notably, the monometallic controls (Cu-NC and Co-NC) show much lower E_1/2 values (0.73 V and 0.63 V), whereas introducing both metals leads to a clear increase in E_1/2 across the intermediate Co/Cu ratios (0.78–0.87 V). This comparison indicates that the ORR enhancement cannot be explained by a single-metal contribution alone and supports a Cu-Co coupled effect within the N-doped carbon framework. EIS indicates that the impedance of these samples is in the range of 70~80 ohms, and the Co₃Cu₇ sample shows the best performance (Figure 7c) at approximately 70 ohms, which suggests that CoCu has the lowest electron/ion transfer resistance and improved conductivity. Their Tafel slopes (Figure 7d,e) are smaller than those of the commercial Pt/C nanocatalyst under identical test conditions, indicating faster ORR kinetics and a reduced kinetic overpotential demand compared with the Pt/C catalyst. Among the tested samples, the Co₉Cu₁ sample exhibits the smallest Tafel slope of only 61.27 mV/dec, implying that its kinetic current increases more rapidly with potential and that the rate-determining electrochemical steps are more favorably promoted than on Pt/C catalyst. These results further demonstrate that under optimal Co/Cu ratios (Co/Cu = 1/1, 7/3 and 3/7), the catalysts exhibit enhanced ORR catalytic performance than Pt/C catalyst, owing to their distinct microstructural characteristics that can provide more accessible interfacial sites and more efficient charge-transfer/transport pathways than Pt/C, thereby strengthening the overall kinetic response reflected by the lower Tafel slopes.

Figure 7. Electrocatalytic performance for ORR of CoCu samples with varying ratios: (a) LSV curves; (b) half-wave potential comparison; (c) EIS results; (d,e) Tafel slope analysis; (f) stability during ORR.

Stability, a critical factor in evaluating catalysts, is assessed for the Co₃Cu₇, CoCu, and Co₇Cu₃ samples, which exhibit superior activity (Figure 7f). After 10,000 s of operation at 0.85 V, the current retention of the Co₃Cu₇ and Co₇Cu₃ samples is approximately 50% of their initial values, while the CoCu sample retains about 40%. All these values surpass the performance of the commercial Pt/C nanocatalyst, which retains only 32%. The stability of CoCu is slightly lower than that of Co₃Cu₇ and Co₇Cu₃, likely due to its higher half-wave potential, which increases the catalyst surface’s reactivity. This heightened reactivity makes the catalyst more prone to degradation processes such as oxidation or corrosion [64], leading to a faster decline in catalytic activity even though it still outperforms Pt/C in current retention under the identical protocol.

The ORR electrocatalytic performance tests indicate that the multi-scale catalysts, synthesized via a modified sequential ultrasonic atomization and high-temperature pyrolysis process, exhibit favorable Tafel slopes and a lower EIS semicircle radius compared to conventional Pt/C nanocatalyst under varying temperature and elemental ratio conditions. Additionally, the catalysts demonstrate an excellent half-wave potential. Notably, the catalyst achieves the best ORR performance when the Co/Cu ratio is 1/1 and the calcination temperature is 800 °C. In line with this, the best ORR performer is also characterized by a smaller EIS semicircle and favorable kinetic metrics, suggesting that the overall ORR trend is governed by the combined effects of intrinsic site activity and interfacial charge-transfer/transport, rather than by a single descriptor. These results suggest that the engineered microstructure significantly enhances ORR electrocatalytic activity. Based on the combined N 1s quantification and the ORR trends of the monometallic controls versus the CoCu series, the ORR activity is most consistently attributed to Co-N_x-related sites coupled with the N-doped carbon environment, with Cu mainly acting as a synergistic regulator of the local electronic and interfacial charge-transfer properties rather than serving as the dominant ORR active phase.

3.3. Electrocatalytic Activity for the CO₂RR

A study on the relationship between microstructure and ORR electrocatalytic performance of (Co/Cu)-NC catalysts at different Co/Cu ratios indicates that samples with Co/Cu ratio of 1/1 have the best overall EC performance. Therefore, the samples with Co/Cu ratio of 1/1 were selected, and their CO₂RR electrochemical catalysis performance was evaluated after further optimization of the calcination effects at 400 °C, 800 °C, and 1000 °C, resulting in four types of samples.

LSV tests were performed for all samples in CO₂- and N₂-saturated solutions, and the results are shown in Figure 8a. The current density for all samples is higher in CO₂-saturated solutions than in N₂-saturated ones, indicating additional CO₂RR-related cathodic current relative to N₂ [65]. Among these, the CoCu-800 sample exhibits the highest current density among the tested samples [65,66]. Tafel slopes, derived from the LSV curves (Figure 8b), are 614.54, 281.54, 277.35, and 289.51 mV/dec for the samples CoCu-0, CoCu-400, CoCu-800 and CoCu-1000, respectively. Among the samples, CoCu-800 has the lowest Tafel slope (277.35 mV/dec). Furthermore, the cyclic voltammetry (CV) testing (Figure S5) was used to evaluate the double-layer capacitance (Cdl) as an indicator of the electrochemical surface area. A larger Cdl therefore indicates a larger accessible interface, but it does not necessarily quantify the absolute density of catalytically effective active sites nor guarantee higher activity [67]. The test results show that the CoCu-1000 has the largest Cdl (ECSA proxy), with a Cdl value of 10.22 mF/cm², followed by the CoCu-800 sample at 9.53 mF/cm² (Figure 8c). Notably, although CoCu-1000 exhibits a slightly larger Cdl than CoCu-800, its catalytic activity is lower. This apparent discrepancy indicates that the accessible interface (ECSA proxy) is not the sole determinant of performance. These results suggest that CoCu-800 may possess a more favorable population of effective interfacial sites and higher intrinsic activity per accessible interface, consistent with its lower Tafel slope and higher CO₂RR-related current response under the investigated conditions. Such behavior is reasonable for thermally treated carbon-supported metal catalysts, where increasing the pyrolysis temperature may change the local coordination/electronic environment of active motifs and thus the intrinsic activity, even if the electrochemically accessible interface increases. Overall, CoCu-800 delivers the highest CO₂RR-related current response and the lowest Tafel slope among the tested catalysts, supporting its favorable CO₂RR performance under the investigated conditions.

Figure 8. Electrochemical performance of the CoCu sample for CO₂RR: (a) LSV curves in CO₂- and N₂-saturated solutions; (b) Tafel slopes derived from the LSV data; (c) double-layer capacitance calculated from CV curves; (d) FE of gas-phase products for CoCu-800 measured using a GCE; (e) FE of gas-phase products for CoCu-800 measured using carbon paper as the working electrode; (f) stability test results for CoCu-800 at −0.73 V over 12 h.

To further evaluate its CO₂RR performance, the product analysis during the catalytic reaction was conducted, whose results are presented in Figure 8d. For the CoCu-800 sample, the maximum FE for CO reaches 69.6% at a potential of −0.93 V. However, accumulation of CO₂ bubbles on the electrode surface was clearly observed, possibly due to the limited working area of the glassy carbon electrode (GCE), which impeded gas release and caused unstable currents. To address this limitation, a carbon paper electrode with an effective area of 1 cm² was used for gas-phase CO detection. With this electrode, the maximum FE for CO can reach 68.7% at a potential of −0.73 V (Figure 8e). The FE for CO obtained with the two electrodes shows no significant difference. Overall, the CO selectivity under the H-cell conditions is reasonable but not intended to represent an exceptional record value; instead, the key merits are the good reproducibility and electrode adaptability, as evidenced by the consistent FE for CO obtained on both GCE and carbon paper. However, when using the GCE, small amounts of methane are detected, with its production increasing significantly at −1.13 V. This is primarily because CO and H₂ are major products. On the GCE, gas bubbles adhere to the surface, prolonging the coexistence of H₂ and CO. This promotes secondary reactions between them, ultimately forming CH₄ and H₂O [68,69]. In contrast, the porous carbon paper electrode with a larger effective area facilitates faster bubble detachment and a more stable gas–liquid–solid interface, thereby improving CO₂ mass transport and mitigating the local accumulation or residence of CO and H₂ near the catalyst surface. As a result, the product distribution on carbon paper is dominated by the primary CO and H₂ pathways, while the likelihood of secondary hydrogenation reactions leading to CH₄ is suppressed. Therefore, the observed electrode-dependent shift in product distribution is mainly attributed to differences in local reaction environment and mass transport/bubble management, rather than a fundamental change in the intrinsic catalytic selectivity of the catalyst.

Catalyst stability, an essential performance metric, was further evaluated using the carbon paper electrode. A 12 h stability test for the CoCu-800 sample was conducted at a potential of −0.73 V (Figure 8f). The red dots represent CO evolution, while the blue triangles represent H₂ evolution. The current density remains stable at approximately −7.3 mA/cm² with minimal fluctuation. Meanwhile, the FE for CO is consistently maintained at around 70%, while the FE for H₂ remains steady at approximately 30%. These results align well with the early findings, confirming the excellent stability and catalytic performance of the CoCu-800 sample under prolonged operation. This durability is rationalized by the multi-scale metal distribution in CoCu-800: ultrafinely dispersed Co/Cu species anchored on the N-doped carbon provide plentiful, stable reaction interfaces, while the coexisting larger metal-containing domains together with the hierarchical carbon framework maintain continuous charge-transfer/transport pathways and mechanical integrity, thereby suppressing metal redistribution or aggregation and improving resistance to electrochemical restructuring during extended electrolysis, which ultimately enables both steady current output and sustained CO selectivity.

The (Co/Cu)-NC material synthesized in this study demonstrates a high metal utilization efficiency while maintaining good yield and structural stability. According to previous reports, CoCu nanocatalysts and CoCu cluster catalysts generally exhibit lower overall catalytic performance compared to the materials developed in this work [37,39,70]. Although the yield is somewhat lower than that reported for certain Co-Cu dual-atom catalysts [38], the present system offers advantages in terms of synthetic simplicity, efficient precursor utilization, and scalability for mass production. Therefore, rather than emphasizing only activity metrics, this strategy highlights practical merits, including stable CO-selective operation during 12 h electrolysis with nearly constant current density and an FE for CO maintained at about 70%, a low-step and potentially cost-effective sequential ultrasonic atomization–pyrolysis synthesis, and improved resource and metal utilization evidenced by the relatively high total metal loading of 22.45 ± 0.07 wt% from ICP analysis; together, these attributes make the present system a resource-efficient and scalable CoCu platform that is promising for scale-up and practical implementation. Considering the TEM-EDS evidence for abundant Cu-rich sub-nanometer clusters and nanoparticles and the fact that Cdl does not directly track CO₂RR activity across temperatures, the CO-selective CO₂RR performance is most plausibly governed by Cu-based sub-nanometer and nanoscale surface sites with favorable local coordination and electronic states, whereas the large Cu-rich particles are more likely secondary contributors, for example, through conductivity or interfacial modulation, rather than being the primary active centers.

4. Conclusions

In summary, multi-scale (Co/Cu)-NC catalysts were synthesized with controlled Co/Cu ratios via an ultrasonic spray atomization method coupled with pyrolysis and calcination, achieving high metal utilization. In ORR, the CoCu sample (Co/Cu = 1/1) calcined at 800 °C shows the highest half-wave potential among the tested samples (0.87 V). Under the same testing conditions, its E_1/2 exceeds that of commercial Pt/C. Further electrocatalytic CO₂RR tests for the CoCu series annealed at various temperatures show that the sample annealed at 800 °C delivers the highest overall CO₂RR response among the tested catalysts, maintaining a CO yield of approximately 70%. Overall, the above quantitative ORR and CO₂RR results support the bifunctional performance of the CoCu-800 catalyst under the investigated conditions. By tuning the Co/Cu ratio and calcination temperature, the electrocatalytic metrics can be systematically optimized within this catalyst family. Importantly, beyond activity metrics, the CoCu-800 catalyst also demonstrates practical merits in stability and reproducibility, as evidenced by sustained operation with stable CO selectivity in CO₂RR, and it maintains competitive durability trends in ORR compared with Pt/C under identical protocols. Moreover, the sequential ultrasonic spray atomization–pyrolysis route offers a streamlined and scalable synthesis, highlighting its potential for scale-up and cost-effective large-scale production. Generally, this method can be applied to synthesize multi-element and multi-scale catalysts for large-scale production economically.