Bifunctional Catalysis Drives Efficient and Ethanol‐Producing Aqueous Zn‐CO2 Batteries Authors: Divyani Gupta, Jinshuo Zou, Anoja Kawsihan, Jodie A Yuwono, Zhaoliang Wu, Jun Yu, Yameng Fan, Sailin Liu, Jianfeng Mao, Zaiping Guo Published: 2026 ABSTRACT ABSTRACTBifunctional electrocatalysts that simultaneously drive CO2 reduction (CO2RR) and oxygen evolution (OER) in neutral electrolytes are essential for aqueous Zn‐CO2 batteries (AZCBs), but conventional designs prioritize CO2RR and are often incompatible with OER, resulting in low energy efficiency, poor reversibility, and rapid degradation of batteries. Here, we introduce a unified catalyst architecture integrating Cu and Ni single atoms with boron‐, nitrogen‐co‐doped carbon (CuNi@BNC‐T). Heteroatom co‐doping stabilizes high single‐atom loadings, enhances metal‐support interactions, and suppresses carbon corrosion, while N‐coordinated Cu+ and Ni3+ sites cooperatively boost CO2RR‐OER, enabling ethanol formation and low‐overpotential OER. This strategy achieves a 0.7 V voltage gap at 5 mA cm−2, 522 h cycling at 20 mA cm−2, and first ethanol production via AZCBs with energy efficiency of 88% (flow‐cell). By elucidating previously unresolved degradation pathways, this work also correlates catalyst dynamics with battery failure, establishing design principles for durable bifunctional catalysts in aqueous Zn‐CO2 systems.Keywords: aqueous Zn‐batteries, bifunctional catalyst, CO2 reduction reaction, O2 evolution reaction, single atom atalysts, Zn‐CO2 batteries Introduction Aqueous zinc‐CO2 batteries (AZCBs) offer a promising approach to simultaneously meet energy demands and mitigate CO2 emissions by combining safe energy storage with CO2 conversion into value‐added products [1]. However, their practical deployment is hindered by short cycle lifetimes (<100 h) and low round‐trip efficiencies (<50%) [2]. These limitations arise primarily from fundamental cathodic barriers, including poor CO2 solubility and diffusion, sluggish kinetics, and insufficient bifunctional activity toward CO2 reduction (CO2RR, discharge) and O2 evolution (OER, charge) reactions in neutral electrolytes [3]. Although considerable efforts have focused on cathode catalyst development [4], the distinct mechanistic requirements of CO2RR and OER typically demand separate catalysts, leading to increased complexity, reduced active‐site density and high costs [2, 5, 6]. Current cathode designs prioritize CO2RR while largely neglecting OER, leading to large voltage gaps (>2 V) and poor reversibility even at low currents (∼1–5 mA cm−2) [3]. Although recent studies have extended cycling to several tens of hours, the mechanistic origins of performance decay and the principles for designing durable bifunctional catalysts remain unclear [4]. Benchmark AZCB catalysts—including copper‐based materials, doped carbons, and metal‐carbon composites, exhibit high CO2RR performance but lack the OER performance and structural robustness required for long‐term cycling [3]. Designing robust bifunctional catalysts with optimal CO2RR‐OER activity and enhanced stability is therefore essential to improving cathode reversibility and extending cycle life [3, 7, 8]. Single‐atom alloys (SAAs) have attracted attention for bifunctional electrocatalysis because their distinct active sites [9] can modulate intermediate adsorption energies by breaking linear scaling relationships through hybrid‐orbital overlap between adjacent metals [10]. However, their performance is limited by low loadings (<1 at.%) and structural instability [11]. Heteroatom co‐doped carbon supports can overcome these limitations by enhancing single‐atom loading, distribution, and stability [12]. Here, we present a unified catalyst design integrating copper‐nickel SAAs with boron‐ and nitrogen‐co‐doped carbon (BNC) to simultaneously address activity and stability challenges. In this architecture, Cu‐Ni SAAs provide bifunctional activity, with Cu centers facilitating CO2RR to C2+ products and Ni sites enabling low‐overpotential OER in neutral pH (Figure 1a) [13, 14]. Embedding metal single atoms within the BNC matrix further mitigates stability and loading limitations. Heteroatom co‐doping enhances single‐atom density and dispersion, strengthens metal‐support interactions, and suppresses carbon corrosion [15], a major barrier to long‐term OER. Beyond stabilization, B and N dopants modulate the local electronic structure of carbon, promoting C2+ selectivity during CO2RR [16, 17]. Together, the rationally designed CuNi@BNC catalyst provides a synergistic solution to the longstanding activity‐stability trade‐offs in AZCB cathodes. These functionalities enable a substantially reduced voltage gap, extended cycling, and ethanol production during discharge. (a) Schematic illustration of the working mechanism of Zn‐CO2 battery and bifunctional catalyst design to address shortcomings at cathode, where Cu sites promote CO2RR toward C2+ products, while Ni facilitates OER and heteroatom co‐doping stabilizes single atoms and tunes carbon electronic structure for enhanced activity and stability under dynamic redox conditions. (b) Field‐emission scanning electron microscopy (FE‐SEM) image (inset: high‐magnification) representing uniform hexagonal morphology and (c) dark‐field scanning TEM (STEM) image confirming the particle morphology in CuNi@BNC‐750. (d) High‐angle annular dark‐field STEM (HAADF‐STEM) image showing atomically dispersed bright spots assigned to isolated Cu/Ni single atoms (yellow circles). (e) Electron energy loss spectroscopy (EELS) mapping confirming homogeneous elemental distribution in CuNi@BNC‐750. Implementing a flow‐type AZCB further enhances CO2 diffusion, improving discharge kinetics and energy efficiency compared with an H‐cell configuration. The catalyst undergoes dynamic yet reversible evolution during repeated CO2RR‐OER cycles, a critical requirement for long‐term cycling. Density functional theory (DFT) calculations reveal the synergistic roles of N‐coordinated Cu+ and Ni3+ sites in promoting both CO2RR and OER kinetics. Mechanistic analysis indicates that early‐stage battery failure originates from anode degradation, whereas long‐term failure results from irreversible catalyst oxidation. Together, these findings establish fundamental design principles for durable bifunctional catalysts and illuminate the degradation pathways that govern cathode performance in rechargeable AZCBs. Results and Discussion Catalyst Synthesis and Characterization CuNi@BNC‐T catalysts were synthesized by pyrolyzing a Cu, Ni‐doped zeolitic imidazole framework (ZIF‐8) precursor (Experimental section, Figures S1 and S2) with urea and phenylboronic acid at 700, 750, and 800°C under Ar for 2 h, respectively, and followed by acid treatment (Figure S3A). Field‐ emission scanning electron microscopy (FE‐SEM) revealed hexagonal morphologies for CuNi@BNC‐700 and 750 (Figures 1b and S3A), however, upon further increasing the pyrolysis temperature, the morphology was distorted (CuNi@BNC‐800, Figure S3A). High‐ angle annular dark‐field scanning transmission electron microscopy (HAADF‐STEM) confirmed hexagonal morphology and uniformly dispersed bright spots corresponding to Cu and Ni atoms (Figure 1c,d) in CuNi@BNC‐750, with the highest metal loading (Ni: 3.54 wt.%, Cu: 3.50 wt.%; Table S1A). Homogeneous elemental distribution is further evidenced by electron energy loss spectroscopy (EELS) mapping in Figure 1e. X‐ray diffraction peaks of CuNi@BNC‐T catalysts display broad peaks at around 25.4° and 43.3° corresponding to (002) and (101) planes of graphitic carbon (Figure S3A), with increased crystallinity at high pyrolysis temperature [18]. No peaks were observed corresponding to the metal and/or metal oxides (expected during high‐temperature pyrolysis) after acid treatment. X‐ray photoelectron spectroscopy (XPS) was employed to gain insight into the elemental chemical states of CuNi@BNC‐750, which revealed the presence of expected elements, including Cu, Ni, C, B, and N, signifying the successful doping of Cu and Ni in carbon B, N co‐doped carbon (Figure S3A). More interestingly, FESEM‐EDS (energy‐ dispersive X‐ ray spectroscopy) elemental analysis and dot mapping were acquired for CuNi@NC‐750 control catalyst to study the effect of boron doping. The boron‐free catalyst shows a comparatively less uniform distribution of Ni and Cu (Figure S3C), suggesting weaker anchoring and thus metal aggregation, which is supported by the reduced atomic concentration of Cu and Ni (Table S1A). The deconvoluted Cu 2p spectra in CuNi@BNC‐750 showed the existence of Cu+ (932.9 and 952.9 eV), while Ni 2p spectra showed characteristic peaks at 854.75 & 863.75 eV and 873.12 & 880.95 eV, which can be ascribed to Ni 2p3/2 and Ni 2p1/2 states of Ni2+ & corresponding satellite peaks [19], respectively (Figure S3A) [20, 21]. These multiple oxidation states in metals are crucial for bifunctional electrocatalytic activity. XPS also revealed C─O (287.3 eV), C─N (285.9 eV), C═C (284.8 eV) and C─B (284 eV) bonding (C 1s), while N 1s spectra confirmed pyridinic, Cu‐N/Ni‐N/N‐B, pyrrolic, graphitic nitrogen at 398.6, 399.6, 400.6 and 401.4 eV. B 1s spectra further confirmed B─O (192.3 eV), B─N (191 eV), and B─C (190.2 eV) coordination [22], verifying successful heteroatom incorporation and metal coordination within carbon. X‐ray absorption spectroscopy (XAS) probed the atomic structure of CuNi@BNC‐750, where Cu K‐edge x‐ray absorption near‐edge structure (XANES) spectrum was located between Cu (0) foil and copper (II) phthalocyanine (CuPc), consistent with the Cu+ oxidation state identified by Cu 2p XPS spectrum (Figure 2a). Likewise, Ni K‐edge XANES spectrum was located between Ni foil and nickel (II) phthalocyanine (NiPc), with an absorption edge closer to that of the NiPc reference, indicating an oxidized Ni species (Figure 2b). The relatively weak pre‐edge feature is consistent with a symmetric Ni‐N coordination environment. Together with the Ni 2p XPS results, these findings confirm that the Ni centers predominantly exist in the +2 oxidation state. Extended x‐ray absorption fine structure (EXAFS) showed Cu─N and Ni─N coordination without Cu─Cu and Ni─Ni signals, confirming atomic dispersion (Figure 2c,d) [23, 24]. Isolated metallic features are supported by wavelet transform (WT) contour plots (Figure 2e,f), exhibiting signals at 5.1 and 3.9 Å−1 for Cu and Ni, respectively (Figure S4A). Near‐edge x‐ray absorption fine structure (NEXAFS) identified local coordination of carbon, boron, and nitrogen in CuNi@BNC‐750 (Figure 2g‐i). C K‐edge spectra revealed peaks due to heteroatom‐induced defects (C─N/C─B), dipole transition of C 1s core electron to π*C = C, π*C‐N/C‐B, π*Cu─N─C/Ni─N─C, and σ*C─C orbitals [24]. N K‐edge depicted characteristic π* transitions of pyridinic, pyrrolic, and graphitic nitrogen, and σ* C─N bonding [24]. B K‐edge showed π* transitions of B─N/B─C and BOx moieties and σ* transitions due to tetrahedral B─O coordination [25, 26]. These results validate B, N co‐doping and their electronic interactions with atomically dispersed Cu and Ni. More importantly, the role of B, N co‐doping in regulating the local coordination of Cu and Ni species was studied by comparing the Cu and Ni K‐edge XANES and EXAFS spectra of the control CuNi@NC‐750 catalyst (Figure S4B). In contrast to CuNi@BNC‐750, CuNi@NC‐750 revealed the co‐existence of Cu─N/Cu─Cu and Ni─N/Ni─Ni coordination, indicating partial aggregation of both Cu and Ni species. These results demonstrate that B, N co‐doping regulates the local coordination of the Cu and Ni sites while suppressing metal aggregation. (a) Cu K‐edge x‐ray absorption near‐edge structure (XANES) spectra of CuNi@BNC‐750 compared with Cu foil and copper phthalocyanine (CuPc) references, indicating the presence of Cu+ species. (b) Ni K‐edge XANES spectra of CuNi@BNC‐750 alongside Ni foil and nickel phthalocyanine (NiPc), suggesting an oxidized Ni state. Extended x‐ray absorption fine structure (FT‐EXAFS) spectra at the (c) Cu K‐edge and (d) Ni K‐edge showing dominant Cu‐N and Ni‐N coordination without Cu─Cu and Ni─Ni bonding, respectively. (e, f) Wavelet transform (WT)‐EXAFS contour plots of Cu and Ni, respectively, supporting Cu‐N and Ni‐N coordination in CuNi@BNC‐750. (g) C K‐edge near‐edge x‐ray absorption fine structure (NEXAFS) spectra displaying characteristic transitions corresponding to defects: ‘a’ C‐N/C‐B at 283 eV, and dipole transition of C 1s core electron to: ‘b’ π*C = C at 284.7 eV, ‘c1’ π*C‐N/C‐B at 287.4 eV, ‘c2’ π*Cu‐N‐C/Ni‐N‐C at 289 eV and ‘d’ σ*C‐C orbitals at 293.8 eV. (h) N K‐edge NEXAFS spectra showing π* transitions of: ‘e1’ pyridinic‐N at 399 eV, ‘e2’ pyrrolic‐N at 399.8 eV, and ‘f’ graphitic‐N at 401.6 eV, along with σ*‐transition of: ‘g’ C─N bonds, respectively. (i) B K‐edge NEXAFS spectra with π* transitions of: ‘h’ B‐N/B‐C at 191.8 eV and ‘i’ BOx at 193.9 eV, followed by σ* features: ‘j’ 199.3 eV, ‘k’ 200.9 eV and ‘l’ 202.7 eV, attributed to tetrahedral B‐O coordination. These electronic structure analyses reveal the distinct local electronic environments in CuNi@BNC‐750. Bifunctional Activity Electrochemical Characterizations CO2RR activity of CuNi@BNC‐T was evaluated via linear sweep voltammetry (LSV) in CO2 and Ar‐saturated 0.1 M KHCO3 using a three‐electrode H‐cell setup (Figures 3a and S5). CuNi@BNC‐T catalysts exhibit high current density in CO2‐saturated 0.1 M KHCO3 electrolyte than in Ar, substantiating the catalytic activity toward CO2RR (Figure S5). Since the catalyst contains Ni and the substrate used in Ni foam (NF), LSV was acquired for bare NF to see its contribution toward CO2RR (Figure S5). Bare NF shows negligible activity, affirming the exclusive role of CuNi@BNC‐750 catalyst toward CO2RR rather than the substrate. To further substantiate the negligible contribution of NF to the electrocatalytic activity, LSV was recorded for the catalyst coated on an inert carbon paper (CP) substrate, which didn't show appreciable changes in both the current density and onset potential (Figure S5). Furthermore, control samples (monometallic Cu@BNC‐750 and Ni@BNC‐750, Figure S5) were also studied, where they revealed an inferior performance than the bimetallic counterparts, highlighting the synergistic role of Cu and Ni in facilitating CO2 reduction kinetics [23, 27]. This inferior performance is allied with considerably lower double layer capacitance (Cdl) and electrochemical active surface area (ECSA) of Cu@BNC‐750 (0.23 mF & 5.75 cm2) and Ni@BNC‐750 (0.261 mF & 6.525 cm2), respectively as compared to CuNi@BNC‐750 catalyst (C dl ≈ 1.43 mF, ECSA ≈ 35.75 cm2), as shown in Figure S8B and Table S4A. The larger ECSA of CuNi@BNC‐750 accounts for the superior CO2RR activity than the control catalysts owing to the enhanced exposure of the electrochemically accessible active sites [28]. (a) Linear sweep voltammetry (LSV) curves in CO2‐saturated 0.1 M KHCO3 representing superior CO2 reduction reaction (CO2RR) performance of CuNi@BNC‐750 compared to other variants (scan rate: 5 mV s−1, 99% iR‐corrected). (b) Bar graph representing CO2RR Faradaic efficiency toward ethanol and respective current density at different applied potentials over CuNi@BNC‐750, showing the increasing current density at more negative potentials and maximum ethanol selectivity at ‐0.4 V vs. RHE. (c) LSV curves demonstrating oxygen evolution reaction (OER) activity with low overpotential for CuNi@BNC‐750 (scan rate: 5 mV s−1, 99% iR‐corrected) in CO2‐saturated 0.1 M KHCO3. (d) Nyquist plots obtained from electrochemical impedance spectroscopy (EIS) depicting lowest charge transfer resistance (R ct) for CuNi@BNC‐750 (inset: equivalent circuit diagram) and thus better kinetics at electrode‐electrolyte interface. (e) Raman spectrum of CuNi@BNC‐T catalysts with corresponding D, G, and 2D bands affirming the graphitic and defective carbon features. (f) Bar graph comparison of intensity ratios (ID/IG and I2D/IG) extracted from Raman spectra, signifying degree of disorder and graphitization in as‐synthesized catalysts. Amongst catalytic variants, CuNi@BNC‐750 outperformed (Table S2A) in terms of current density (‐72 mA cm−2 at ‐1.2 V vs. RHE). Chronoamperometry (CA) measurements were carried out at various applied potentials (‐0.2 to ‐0.6 V vs. RHE) over CuNi@BNC‐750 to determine CO2RR product selectivity (Figure S6). As expected, the current density increased with more negative potentials, signifying the enhanced electrochemical kinetics over CuNi@BNC‐750. Initially, the gaseous products were quantified using gas chromatography, which revealed H2 as the only product (Figure S6). The corresponding H2 Faradaic efficiencies reached 90% at ‐0.6 V, while the HER suppression was witnessed at ‐0.4 V vs. RHE. CuNi@BNC‐750 revealed maximum ethanol production with 54% selectivity at ‐0.4 V and a current density of ‐6.5 mA cm−2, consistent with HER suppression (Figure 3b; Figures S6 and S7A). CA was also conducted on CuNi@BNC‐750 on CP at ‐0.4 V vs. RHE, which revealed a FE of 53.86% toward ethanol production, affirming the catalyst selectivity toward C2+ products (Figure S7A). Besides, post‐CO2RR FE‐SEM, EDS dot mapping and Raman spectroscopy were performed on CuNi@BNC‐750 on CP and compared with the catalyst on NF, respectively. No evident differences in morphology, elemental distribution, and ID/IG ratios were witnessed, indicating that activity is due to the catalyst and is not influenced by substrate‐induced effects during electrolysis (Figure S7B). To gain deeper insights into the high ethanol selectivity, control measurements were conducted using monometallic counterparts (Cu@BNC‐750 and Ni@BNC‐750) as well as CuNi@NC‐750 catalyst (without B co‐doping). Under identical CO2RR conditions (‐0.4 V vs. RHE), a mixture of gaseous and liquid products was observed including H2, C1 (CO, CH4, formic acid) and C2 (acetate) over CuNi@NC‐750, with no detectable ethanol or ethylene. At the same time, Ni@BNC and Cu@BNC both revealed the formation of H2 and CO, whilst Cu@BNC‐750 also produced HCOOH additionally (Figure S7C and Table S2B). These findings suggest that alongside synergistic Cu‐Ni interactions, B, N co‐doping is equally essential to promote selective ethanol production. A plausible explanation could be the cooperativity effect by B, N co‐doped carbon which stabilizes CO adsorption and its dimerization [16, 29], promoting sequential hydrogenation of oxygenated intermediates to ethanol, while suppressing CO bond cleavage and the competing ethylene pathway [30, 31]. This is further supported by the DFT calculations, which demonstrate that boron moieties facilitate the co‐adsorption of CO near Cu‐Ni active sites, thereby enhancing C‐C coupling and selective ethanol formation (Figure S13C,D). CuNi@BNC‐750 also exhibited superior OER activity, requiring only 331 mV overpotential to achieve 10 mA cm−2 with a low Tafel slope (167.39 mV dec−1), surpassing CuNi@BNC‐700 (295.49 mV dec−1) and CuNi@BNC‐800 (393.52 mV dec−1), thus indicating the superior reaction kinetics (Figures 3c and S8A). This OER activity could also be exclusively related to the catalyst, irrespective of the substrate used (Figure S8A). Electrochemical impedance spectroscopy (EIS) revealed lowest charge‐transfer resistance (R ct = 13.6 Ω, Figure 3d) and largest electrochemical surface area (35.75 cm2), confirming boosted kinetics and enhanced number of electrocatalytic active sites (Figure S8A and Tables S3 and 4A). Raman spectrum of CuNi@BNC‐T catalysts displayed characteristic D, G, and 2D bands in carbon, with CuNi@BNC‐750 exhibiting highest ID/IG ratio (1.13), suggesting higher defect density (Figure 3e,f and Table S5) [32]. The decreasing I2D/IG ratio with increasing pyrolysis temperatures reflects changes in electronic structure, defect environment, and graphitic order of the heteroatom co‐doped carbon in CuNi@BNC‐T catalysts [33]. Interestingly, the changes in heteroatom content (B, N, O) of the CuNi@BNC‐T were determined by their respective XPS survey spectrum (Figure S3B). As summarized in Table S1B, the B content increased from 1.35 to 4.45 at.% with increasing pyrolysis temperature, whereas the N and O contents exhibited non‐monotonic variations with decreased concentration at 750°C and subsequent increase at 800°C. These inferences point toward the temperature‐dependent redistribution and stabilization of heteroatoms in carbon framework rather than a uniform increase in the total doping concentration. These structural features, in addition to stable OER (24 h) and CO2RR (10 h) operation at 10 mA cm−2 (Figures S9 and S10), highlight the bifunctional activity and catalyst robustness, signifying the application of CuNi@BNC‐750 as a cathode material in Zn‐CO2 batteries. Notably, the catalyst's stability after OER could be correlated with the enhanced corrosion resistance of carbon support during applied oxidation conditions, owing to the B, N co‐doping in the carbon framework. Chronoamperometric tests were performed for CuNi@BNC‐750 and control CuNi@NC‐750 catalysts at 1.8 V vs. RHE for 20 min. followed by Tafel analysis to calculate the corrosion current density that is, Icorr (Figure S8C and Table S4B). The extrapolated Icorr of CuNi@BNC‐750 (2.25 × 10−5 mA cm−2) was much lower than that of CuNi@NC‐750 (1.29 × 10−4 mA cm−2), confirming that heteroatom co‐doping enhances the carbon stability while suppressing its corrosion during OER. Mechanistic Investigations Post‐stability XPS was performed to investigate the evolution of active sites in CuNi@BNC‐750 (Figures S9 and S10). After CO2RR, catalyst showed carbon coordination rearrangements with retained structural integrity. Deconvoluted Cu 2p spectra revealed prominent Cu+ peaks, while Ni 2p spectra showed a slight positive binding energy shift, suggesting that Cu+ facilitates CO2 adsorption/activation and adjacent Ni2+ accelerates charge transfer [34]. Enhanced C = C and C‐O contributions were witnessed in deconvoluted C 1s XPS spectrum, while pyrrolic‐N features dominated pyridinic‐N contributions (N 1s XPS spectrum). Notably, a clear and high intensity peak for metal‐N coordination is witnessed, suggesting the exclusive role of metal‐N moieties toward CO2RR. In addition, B‐N coordination exhibited maximum intensity as compared to B─O and B─C bonds, signifying the role of BN bonding for CO2 reduction to ethanol over CuNi@BNC‐750. These results confirmed structural rearrangements and dominant roles of C = C, Cu‐ N/Ni‐N/N‐B and B‐N toward CO2RR [22, 35]. After OER, Ni 2p spectra revealed Ni3+ formation consistent with OER‐active NiOOH [36], accompanied by slight shifts in Cu 2p spectra, indicating promoted electron transfer [34]. Carbon rearrangement was also observed while the heteroatom co‐doped carbon coordination remained intact (Figure S9). Deconvoluted C 1s XPS spectrum revealed an increased intensity for C═C bonds, whereas the N 1s spectra demonstrated predominant role of pyrrolic‐N species over pyridinic‐N (before OER). B 1s spectrum revealed surface oxidation with increased intensity of B‐O coordination than B─N and B─C bonds. Peaks related with metal‐N coordination are well‐retained, suggesting the retention of metal active centers during OER. These inferences highlight the dynamic evolution in CuNi@BNC‐750 and synergistic role of Cu and Ni toward bifunctional activity, where the surface reconstruction may generate additional active sites, while retaining structural robustness critical for long‐term electrocatalysis. Catalyst evolution was further understood by in situ Raman spectroscopy under switching CO2RR‐ OER conditions (Figure 4a). A band at 351 cm−1 (M‐CO stretching [37]) emerged during CO2RR but disappeared upon OER, where new peaks at 468 (Eg bending) and 580 cm−1 (A1g stretching) confirmed NiOOH formation [38] after OER (∼1800 s). Presence of D and G bands throughout indicated carbon stability, while ID/IG ratio decreased from 1.13 (open‐circuit) to 1.0 after CO2RR (900 s) and 0.84 after OER (1200 s), demonstrating carbon rearrangement. The ratio decreased during CO2RR (reduced disorder) but increased during OER, reflecting increased degree of disorder [32]. Raman shifts around 2030 cm−1 (atop *COL) and 1870 cm−1 (bridge *COB) confirmed CO adsorption and CO‐CO coupling for C2+ products, with gradual decrease in intensity reflecting CO consumption and subsequent hydrogenation [39]. Absence of *CO bands during OER illustrates the catalyst's ability to dynamically switch active sites, a prerequisite for robust AZCB cathodes where CO2RR and OER occur at same electrode. (a) In situ Raman spectrum acquired under switching CO2RR‐OER conditions (‐0.4 V vs. RHE during CO2RR, 1.6 V vs. RHE during OER, electrolyte: 1 M KHCO3 + CO2 flow at 20 rpm) revealing: M‐CO formation during CO2RR and NiOOH formation during OER; dynamic carbon rearrangement showing progressive decrease of ID/IG ratio from OCP to CO2RR (reduced disorder) followed by recovery during OER (increased disorder); atop and bridged *CO adsorption and CO‐CO coupling during CO2RR with diminishing intensity consistent with hydrogenation. Free energy profiles for (b) OER and (c‐d) CO2RR over N‐coordinated Cu‐ (teal) and Ni‐ (green) active sites in CuNi@BNC‐750 catalyst along with the adsorbed intermediates configuration (top‐view). These inferences signify the synergistic role of Cu and Ni single atoms to facilitate bifunctional CO2RR‐OER activity over designed catalyst. DFT calculations further reinforced CO2RR‐OER mechanism over N‐coordinated Ni and Cu sites in CuNi@BNC‐750 (Figure S11). For OER (Figures 4b and S12; Tables S6 and S7, the potential‐limiting step was *OH → *O on the Ni (∆G = 2.43 eV) and *O → *OOH on Cu (∆G = 2.03 eV). Ni sites facilitate *OH adsorption, while Cu sites facilitate subsequent *OH → *O conversion. Notably, the free‐energy change for *O → *OOH is lower on Ni (0.57 eV) than on Cu (2.03 eV), suggesting that Ni stabilizes *OOH intermediates more effectively. The synergy between Cu─Ni effectively balances the adsorption energies for later steps and releases O2 with a reduced energy barrier in CuNi@BNC‐750. For CO2RR, the C─C coupling step was investigated, which is crucial for ethanol formation. Based on the experimentally verified adsorbed *CO intermediates and CO─CO coupling process revealed by in situ Raman spectroscopy, DFT calculations were performed to elucidate the energetically favorable reaction pathways toward ethanol. The second *CO involved in the CO─CO coupling step (*C2O2) is assumed to originate from the continuous electrochemical CO2 reduction, which is supported by the calculated free‐energy profile showing that CO desorption is a thermodynamically favorable downhill reaction (Table S9B). Cu‐ sites preferentially facilitate *C2OOH and *CCO intermediates with reduced ∆G (0.04 (*C2O2 → *C2OOH) and 0.32 eV (*C2OOH → *CCO)) compared to Ni (Figure 4c; Tables S8 and S9), indicating better energetics toward C─C coupling of Cu sites. It is noteworthy to mention that *CCO intermediate preferentially adsorbs at the Cu‐Ni bridge position, with thermodynamically less favorable atop adsorption on either the Cu or Ni site. Therefore, the stable CuNi*CCO bridge configuration is considered while analyzing the reaction pathway (Figure S13B). Besides, the effect of boron toward *CO‐CO co‐adsorption was studied by considering experimentally verified B─N, B─C, and B─O coordination environments. The B‐O species, linked to the carbon framework through B─N and B─C bonds, was optimized for lowest‐energy configurations near Cu, Ni, and Cu‐Ni active sites and subsequently utilized to calculate the co‐adsorption energies of two *CO intermediates (Figure S13C). As shown in Figure S13D, the boron incorporation thermodynamically favors the *CO co‐adsorption at metal‐active sites, with the adsorption energies of 1.71 eV in the absence of B‐O and 0.63, 0.66, and 0.65 in the presence of B‐O near Cu, Ni, and both active sites, respectively. Thus, boron in CuNi@BNC‐750 is crucial for C‐C coupling over metal active sites during CO2RR. Thereafter, the ethanol pathway was studied via formation of *CHCO, where the*CH‐CHO route was thermodynamically favored (Figure 4d) over *COH‐CH (Figure S13A and Table S10). Subsequent hydrogenation proceeds through the *CH2‐CHO intermediate, followed by formation of *CH2CH2O, which is energetically more favorable than another possible *CH3CHO intermediate. The *CH2CH2O pathway exhibits a lower potential‐limiting step (0.98 eV on Ni) than the *CH3CHO pathway (1.43 eV on Cu), indicating that CH2CH2O is the preferred intermediate for the subsequent formation of CH3CH2O and ultimately ethanol. Downhill energetics is witnessed from *CH‐CHO onwards to ethanol at both Cu‐ and Ni sites. Cu sites require negligible energies for *CH2‐ CHO and *CH3–CH2O conversion reactions and also stabilizes *CH3–CH2O more effectively than Ni sites, thereby accelerating ethanol formation. In contrast, Ni sites exhibit higher energy requirements at key hydrogenation steps (*C2OOH → *CCO and *CH2–CHO → *CH3–CH2O). Ethanol desorption requires almost same energy at both Ni and Cu; however, the Cu sites demonstrate better hydrogenation thermodynamics with smaller energy penalties during hydrogenation. Overall, these results depict the hybrid role of Ni and Cu sites toward CO2 activation to *CO, where synergy between Cu and Ni provides energetically favorable routes for critical C─C bond formation through stabilizing Cu*CO‐Ni*CO and CuNi*CCO as compared to isolated Cu or Ni centers, whilst Cu sites preferentially stabilize C‐C coupling and subsequent hydrogenation for selective CO2RR to ethanol. Zn‐CO2 battery Performance Zn‐CO2 battery was assembled in an H‐cell using CuNi@BNC‐750 cathode, Zn foil anode, 1 M KHCO3 catholyte (CO2‐saturated) and 1 M KOH + 0.2 M Zn(ac)2 anolyte. An open circuit potential of 1.36 V was obtained (Figure 5a) with discharge and charge voltages of 1.36–0.4 V and 1.5–2.4 V across 0.1–20 mA cm−2 (Figure 5b), maintaining a voltage gap of 0.7 V at 5 mA cm−2, which is reduced to the order of ∼1–1.5 V when compared with most of the reported literature (Figure 5c and Table S11). In addition, a peak power density of 12.4 mW cm−2 was achieved (Figure S14), outperforming reported catalysts (Table S11). The battery revealed a stable discharge performance at different applied current densities, where increased current led to reduced voltage, consequently (Figure S14). Further, upon quantifying the discharge product via 1H‐NMR spectroscopy, ethanol was identified as a sole liquid product (Figure S14). The production rate of ethanol demonstrated a near‐linear increase with increasing discharge currents, reaching the maximum at 20 mA cm−2 with a yield rate of around 8 mmol h−1 cm−2 after 0.5 h of battery discharge. Long‐term cycling at 20 mA cm−2 delivered stable voltage output for 230 h and extended to 522 h after replacing the Zn anode and anolyte for two times (mechanical recharging [40], Figures 5d and S15). The initial galvanostatic charge‐discharge cycles (Figure S15) reveal well‐defined plateaus with a stable voltage response, suggesting facile redox processes and minimal polarization. However, upon prolonged cycling for 230 h, the discharge voltage dropped to 0 V while charge voltage remained stable, necessitating mechanical recharging to restart the battery operation. Battery failure at this stage corresponds to the Zn dendrite formation and anode passivation, where mechanical recharging exposes fresh Zn metallic surface and restores battery reversibility. Upon first mechanical recharge, the voltage profile is largely restored, but with slight broadening of charge plateau and increased slope of discharge plateau due to the increased resistance and partial electrochemical activity loss. Nevertheless, the cycling behavior remains stable, indicating that the mechanical recharging restored most of the electrode activity and that the initial loss of battery performance was due to anode degradation rather than cathode decay. After further discharge capacity decay, battery was mechanically recharged for the second time (Figure S15). Although the charge‐discharge plateaus were still visible and the voltage response were maintained, the charge plateau shows further broadening while discharge plateau displays increased slopes. This could be attributed to cathode degradation (surface passivation, increased resistance, hindered mass transport), since anode degradation was not significant at this stage. This reveals that mechanical recharging cannot completely restore the battery performance once irreversible cathode degradation occurs. Overall, the stable voltage response in due course of battery cycling with mechanical recharging suggests the robustness of CuNi@BNC‐750 catalyst. With each mechanical recharge, the charge plateaus broaden while the slope of discharge plateaus increase. This indicates that cathode degradation is the limiting factor in long‐term Zn‐CO2 battery cycling, with major effects on discharge process whilst charge kinetics remain relatively stable. 1H‐NMR analysis of catholyte after 230 h showed minimal ethanol production (0.2 mg h−1 mgcat −1), likely due to its oxidation into acetate (15 mg h−1 mgcat −1, Figure S16A). Ethanol oxidation didn't alter charging voltages, consistent with its sluggish kinetics compared to OER on CuNi@BNC‐750 (Figure S16A). Thus, charging is governed by OER as a dominant reaction, whilst ethanol oxidation observed during prolonged cycling could be attributed to the solution‐phase processes in the electrolyte (non‐faradaic/chemical reactions). To substantiate the remarkable performance of CuNi@BNC‐750 cathode, control Zn‐CO2 batteries were assembled using the monometallic catalyst counterparts (Ni@BNC‐750 and Cu@BNC‐750). The charge‐discharge polarization curves (Figure S16B) demonstrate much larger voltage gaps than CuNi@BNC‐750 cathode, that too at much lower current densities. More precisely, the Ni@BNC‐750 cathode exhibits a voltage gap of ∼2.1 V at 0.5 mA cm−2, while the Cu@BNC‐750 cathode shows a voltage gap of ∼2.2 V at 2 mA cm−2. The control cathodes could not sustain a high discharge voltage or a low voltage gap at high current density of 20 mA cm−2, which is in sharp contrast to CuNi@BNC‐750 cathode. Consequently, long‐term cycling tests were performed at lower current densities of 0.5 and 2 mA cm−2 for Ni@BNC‐750 and Cu@BNC‐750‐based batteries, respectively (Figure S16B). A much lower cycling stability for only ∼74 (Ni@BNC‐750) and ∼80 h (Cu@BNC‐750) was achieved (1 min charge/1 min discharge per cycle), as compared to CuNi@BNC‐750‐based battery (230 h, without mechanical recharging) at higher current density of 20 mA cm−2 and extended charge‐discharge time (5 min) per cycle. These outcomes are conclusive of the synergistic effect between Cu and Ni active sites to enhance the bifunctional CO2RR‐OER activity at cathode, thereby minimizing voltage gap and extending the cycle lifetime at high current densities.Figure 5e compares the cycling stability for reported cathodes in the literature with CuNi@BNC‐750, revealing the extended cycle life with less mechanical recharging needs. (a) Open‐circuit voltage profile for as‐assembled Zn‐CO2 battery (inset: H‐cell configuration) showing stable response for over 1000 s. (b) Galvanostatic charge‐discharge polarization curves displaying a voltage gap <2 V across 0.1‐20 mA cm−2 with voltage gaps low as 0.7 V at 5 mA cm−2 and 1.1 V at 10 mA cm−2. (c) Comparison of charge‐discharge voltage gap for reported Zn‐CO2 battery cathodes extracted from Table S11, showing the reduced voltage gaps (even at high current density) in this work. (d) Long‐term cycling in H‐cell configuration (1 cycle: 5 min. discharge + 5 min. charge), maintaining stable operation for 230 h and extended to 522 h upon mechanical recharging (Zn anode and anolyte refreshing), until cathode deactivation. (e) Comparison of cycling performance of CuNi@BNC‐750 cathode‐based Zn‐CO2 battery with reported literature, showing the extended cycle life and high achievable current density. (f) Battery cycling in flow‐cell configuration (1 cycle: 5 min. discharge + 5 min. charge) with voltage cutoffs (Discharge: 0.5 V, Charge: 3.0 V), revealing improved discharge voltage initially but compromised charging process under mass‐transport limitations as well as shortened cycle durations (1 min. discharge, 1 min. charge) due to fast capacity decay (rapidly reaching cutoff limits). In view of practical deployment, a flow‐cell configuration (Figure 5f) with continuous electrolyte circulation was implemented, operated with voltage cutoffs (0.5‐3.0 V) to avoid overcharging and deep discharge under enhanced mass‐transport conditions [41]. Enhanced CO2 diffusion improved discharge voltage (∼0.8 V), though charging voltage increased to 2.7 V possibly due to reduced OH− accumulation near electrode and the battery reached cutoffs after 70 h [42]. Nonetheless, stable cycling for 139 h was achieved with reduced charge‐discharge times without mechanical recharging, producing ethanol (18.92 mg h−1 mgcat. −1) during discharge (Figure S17). The cycling trends observed in flow‐type Zn‐CO2 battery were different from those witnessed in H‐type battery. Initial cycling shows distinct charge‐discharge plateaus with stable redox behavior and minimal polarization (Figure S17). The voltage plateaus remain well‐defined and stable even after 50 h, indicating electrode robustness and facile kinetics at electrode‐electrolyte interface. However, discharge voltage progressively reduced while charge voltage increased, as a consequence, the battery reached cutoffs quickly (0.5 V and 3 V) with reduced discharge and charge durations (1 min after 70 h), suggesting increased resistance and diffusion limitations (Figure S17). The charge plateau altered toward a sharp and step‐like curve, while discharge plateau showed a similar rectangular curve but with an increased slope. This point toward increased resistance, electrode passivation, and hindered electron and mass transport during long‐term cycling. Under flow‐cell conditions, ethanol was produced as a major product (production rate: 18.92 mg h−1 mgcat. −1), as confirmed from 1H‐NMR spectra in Figure S17d. Further, discharge was conducted at a fixed current density of 20 mA cm−2 under H‐cell and flow‐cell configurations (Figure S17e), which revealed a better specific capacity under flow‐cell conditions (552 mA h g−1) than in H‐cell conditions (331 mAh g−1). Though, it is important to highlight that OER activity is suppressed under flow conditions, and blocked anolyte circulation is witnessed during cycling due to ZnO precipitation, highlighting the importance of anolyte management, gas diffusion electrode design, and cell engineering to ensure stable cycling under practical conditions. Flow configuration also yielded higher specific capacity (552 mA h g−1) and energy efficiency (88.9%) compared with H‐cell (Table S12). It is important to mention that ZnO precipitation in flow‐channels after ∼80 h affected anolyte circulation [43], contributing to anolyte drying and performance decay. This highlights the significance of electrode engineering, anolyte management, cell architecture, and operating conditions to ensure long cycling in Zn‐CO2 batteries under realistic conditions. Battery Failure and Cathode Degradation Mechanism Battery failure mechanism was investigated after long‐term AZCB cycling. Zn anodes cycled at 20 mA cm−2 revealed dendritic growth and surface passivation (Figure S18). Photographs of pristine and cycled anodes showed evident deposition on metallic anode surface. Further, FE‐SEM analysis revealed that after 100 cycles, the initially smooth Zn surface became rough, which upon further cycling (1000 cycles) produced large zinc oxide/hydroxide chunks alongside dendrites, increasing until battery failure. Ex situ Infrared (IR) spectroscopy (Figure S19) displayed characteristic Zn‐O stretching (406, 479, 508, 546 cm−1) and Zn‐OH bending (772 cm−1) vibrations, confirming zinc oxide and zinc hydroxide formation [44, 45, 46]. Additional bands at 714 and 843 cm−1 (CO3 2− bending) and at 1030, 1086 and 1440 cm−1 (carbonate or bicarbonate complexes, C‐O stretching), indicated the presence of carbonate and/or zinc carbonate [47, 48]. Post‐cycling XRD corroborated the presence of metallic Zn, ZnO, Zn(OH)2, and ZnCO3 (Figure S19). Thus, battery failure after 230 h arises from dendrite growth and anode passivation. Thereafter, the battery could be mechanically recharged twice and operate for 522 h, after which dendrites weren't observed, pointing toward cathode degradation as the reason behind failure. FE‐SEM, STEM‐HAADF, and synchrotron soft x‐ray spectroscopy measurements of cycled cathodes revealed catalyst aggregation (Figure S20) and structural evolution (Figure S21). The post‐cycling HAADF‐STEM imaging confirmed the retention and uniform atomic dispersion of Cu and Ni single atoms, demonstrating the cathode stability upon cycling (Figure S20), which allowed mechanical recharging to restore the battery performance after 230 h. FE‐SEM images displayed some agglomeration after 230 h with higher retention of hexagonal morphology, which gradually transformed to the dense cauliflower‐like agglomerates upon battery failure after 522 h. NEXAFS Ni L‐edge spectra exhibited split L3 envelope (852.7, 854 eV) with corresponding L2 (869.9, 871 eV) due to octahedral Ni2+ (3d8: t2g6eg2) [49]. Post‐failure, the intensity of low‐energy peak (852.7 eV) increased significantly, accompanied by high energy peak shift to 854.7 eV (multiplet broadening) [50]. A similar increase was observed at low‐energy L2 peak (869.9 eV), indicating generation of t2g holes from partial Ni2+ oxidation to high‐spin Ni3+ (d7: t2g5eg2) and enhanced Ni‐O covalency [50]. Cu L‐edge spectra also revealed L2,3 edges, where L3 peaks (934, 936 eV) indicated the presence of Cu+ (3d10) [51]. At battery failure, the spectra shifted toward lower energies (930.8, 934.2 eV), consistent with Cu+ to Cu2+ oxidation and surface reconstruction [51]. The C K‐edge spectra (Figure S21) retained graphitic and heteroatom‐doping features with emerging π* transitions (288‐291 eV) due to stronger Cu/Ni‐N‐C coordination and surface oxidation (C = O). This indicates carbon rearrangement and surface oxidation under prolonged cycling conditions. N K‐edge spectra (Figure S21) displayed a new π* N─O transition (405.2 eV) due to oxidized nitrogen species, along with suppressed pyridinic/pyrrolic‐N signals, representing oxidation induced deactivation of N‐sites. Besides, the characteristic π* and σ* transitions in B K‐edge spectra evolved (Figure S21), showing a dominant π* feature (∼194 eV) and low‐energy π* contribution (191.8 eV). This point toward the severe loss of B‐N/B‐C moieties and BOx formation, supported by strong σ* transitions from oxidized boron species. These transformations indicate cathode deactivation via oxidation of metal‐centers and heteroatom co‐doped carbon involving carbon rearrangement with surface oxidation, loss of N‐based coordination sites via oxidative deactivation and severe boron oxidation. Figure 6a demonstrates battery failure at 20 mA cm−2, initiated by anode degradation and followed by long‐term cathode deactivation. For deeper insights into cathode degradation, cycling was executed at low current (5 mA cm−2) with cutoffs (2.5 V charge, 1.0 V discharge; Figure 6b and S22). Stable cycling persisted for 68 h, after which both charge (∼2 V to 1.8 V) and discharge (∼1.2 to 1 V) voltages dropped and discharge time abruptly reduced from 5 to 1 min, indicating catalyst degradation and increased interfacial resistance. In the early stages (Figures S22), the charge‐discharge behavior remains stable (triangular) within the cutoff window and reflects stable plateaus, suggesting robust electrode kinetics at the interface. With extended duration (∼70 h), cycling becomes unstable, with rapid discharge duration decay from 5 to 1 min per cycle, though charging remains stable. This indicates the electrode surface passivation and sluggish discharge kinetics. The initial triangular curves also shifted toward a rectangular shape, suggesting electrode degradation and/or rearrangement. The charge response sustained within the cutoff limits while the discharge duration shortened to 1 min by reaching cutoffs rapidly beyond 1100 h, until the cutoffs are not removed. When voltage cutoffs are removed (1154 h), the battery performance rapidly decayed due to deep discharging and complete cell failure occurred within 1 h, as evidenced by the collapse of charge‐discharge symmetry and discharge plateaus. Even the mechanical recharging couldn't restore the performance, confirming the irreversible cathode degradation. Meanwhile, ethanol production dominated at 5 mA cm−2 till 68 h, but its relative intensity decreased after extended cycling and acetate production (3 mg h−1 mgcat. −1) was witnessed in addition to ethanol (1.5 mg h−1 mgcat. −1). This describes the role of current density toward increased ethanol oxidation to acetate during cycling. Thereafter, the battery reached a new quasi‐steady state with reduced discharge capacity, suggesting rate‐dependent degradation [52, 53]. Battery operated for >1100 h with stable charge (5 min) and rapid discharge (1 min) but removing cutoffs caused complete capacity loss within 24 h. Post‐failure analysis identified dendrite growth and ZnO, Zn(OH)2 formation with reduced intensity for carbonate‐related vibrations (Figure S23). Mechanical recharging failed to recover performance and failure recurred within 24 h. FE‐SEM of cycled cathode revealed morphology collapse over time, forming irregular micro‐clusters and flake‐like structures (Figure S24). Soft x‐ray spectroscopy (Figure S25) revealed partial oxidation from Ni2+ to Ni3+ after 68 h, which intensified at battery failure, while complete loss of active Cu+ due to oxidation to Cu2+ is witnessed after 68 h [51] followed by slight Cu+ redeposition after failure. On the other hand, heteroatom co‐doped carbon showed analogous, but less pronounced oxidation than at higher current density. EIS was acquired to understand the progressive degradation of CuNi@BNC‐750 catalysts during long‐term battery cycling. Before cycling, the Nyquist plot revealed a well‐defined semicircle without a Warburg tail, confirming the dominant charge‐transfer process at electrode‐electrolyte interface. A very small solution and charge‐transfer resistance (R s ∼3 Ω; R ct ∼25 Ω) validates the rapid interfacial kinetics. The increased R ct (63 Ω) after 68 h could be correlated with the catalyst surface passivation [54], resulting in slow reaction kinetics at the interface and thus the onset of catalyst deactivation. An entirely different behavior is observed at battery failure stage, where the plot behavior changed to a steep Warburg‐like arc (nearly 45° slope, extending to ∼200 Ω on real axis) and increased solution resistance (R s ∼40 Ω), depicting a diffusion‐controlled process due to blocked active sites and restricted ion transport [55]. This complete change in behavior from kinetic‐controlled to diffusion‐limited transport is indicative of severe degradation of active sites and restricted mass transport and that ion‐diffusion, not charge transfer, governs the electrochemical kinetics. Thus, EIS substantiated gradual catalyst degradation where the cathode transitions from a kinetically controlled to a diffusion‐limited process during cycling (Figures 6c and S26). Overall, these findings establish that irreversible oxidation of metal‐sites and heteroatom co‐doped carbon drives cathode deactivation due to loss of active sites, surface passivation and increased interfacial resistance, resulting in long‐term performance loss. Especially, the transformation of Cu+ to Cu2+ leads to discharge performance decay and stabilizing Cu+ active‐sites under charging potentials will be critical to further enhance the catalyst stability and thus battery reversibility. (a) Schematic representation of battery failure at 20 mA cm−2 revealing the initial failure due to anode degradation followed by long‐term failure due to cathode deactivation. (b) Long‐term cycling performance of CuNi@BNC‐750 equipped Zn‐CO2 battery at 5 mA cm−2 with voltage cutoffs (Discharge: 1.0 V, Charge: 2.0 V), representing initial stable regime followed by structural rearrangement, voltage decay, and transition to quasi‐steady state with reduced capacity. (c) Schematic demonstration of rate‐dependent cathode degradation mechanism at 5 mA cm−2, demonstrating the irreversible surface oxidation of active sites in CuNi@BNC‐750 and a shift in reaction kinetics from kinetic‐controlled to diffusion‐limited, resulting in cathode degradation. Conclusion This work demonstrates that a Cu‐Ni single atom alloy embedded in B, N co‐doped carbon enables efficient and durable bifunctional catalysis for CO2 reduction and O2 evolution in AZCBs. Synergistic N‐coordinated Cu and Ni sites stabilize C‐C coupling for ethanol formation while facilitating low‐overpotential OER in neutral media. In situ Raman analysis reveals dynamic yet reversible catalyst evolution during repeated CO2RR‐OER cycling, underpinning long‐term stability. Consequently, CuNi@BNC‐750 achieve a reduced voltage gap of 0.7 V at 5 mA cm−2, sustained cycling for 522 h, and C2+ product formation during discharge. Mechanistic investigations identify zinc dendrite growth and anode passivation as the primary causes of early‐stage failure, while long‐term degradation arises from cathode deactivation through progressive oxidation of metal‐moieties and heteroatom co‐doped carbon and interfacial passivation. At low current densities, limited mass transport exacerbates passivation and accelerates structural rearrangement, shifting the cathode‐electrolyte interface from kinetically controlled to diffusion‐limited behavior. Beyond catalyst design, the flow‐type configuration underscores the critical role of cell engineering in achieving practical performance. Enhanced electrolyte circulation and CO2 transport improve discharge kinetics and energy efficiency but introduce new challenges: OH− depletion elevates charging voltages, ZnO precipitation impedes anolyte circulation, and electrolyte drying ultimately limits operation. Therefore, extending battery lifetime requires not only stable bifunctional catalysts but also optimized electrolyte flow, gas management and interface engineering. Moreover, discharge‐product evolution warrants careful consideration, as ethanol formed during discharge gradually oxidizes to acetate during extended cycling. Overall, this work realizes a long‐life aqueous Zn‐CO2 battery through the combined optimization of bifunctional catalyst design and system engineering, offering a comprehensive roadmap for integrating CO2 conversion with sustainable energy storage. Experimental Section Catalyst synthesis: CuNi@BNC‐T catalysts were synthesized by pyrolyzing a Cu, Ni doped zeolitic imidazole framework (ZIF‐8) precursor, obtained via a modified ZIF‐8 synthesis using nitrate salts of copper, nickel, zinc, and 2‐methylimidazole in methanol. Briefly, 7.4 mmol of zinc nitrate (Zn(NO3)2, 1.4 g), 7.5 mmol of cupric nitrate (Cu(NO3)2, 1.4 g) and 7.7 mmol of nickel nitrate (Ni(NO3)2, 1.4 g) were dissolved in 150 mL methanol under continuous stirring (A). Subsequently, a 150 mL solution of 64 mmol 2‐methylimidazole (C4H6N2, 5.25 g) was prepared in deionized water and added dropwise to A under continuous stirring for 1 h. The above solution mixture was kept at rest for 24 h at room temperature to yield CuNi@ZIF‐8 precipitates. Precipitates were collected via centrifugation and washed with fresh methanol 5–6 times and dried under vacuum overnight at 60°C. Similarly, the control samples including Cu@ZIF‐8, Ni@ZIF‐8 and ZIF‐8 were also prepared, but without the addition of Ni, Cu and Cu, Ni precursors respectively. It is noteworthy to mention that the synthesis was carried out under inert Ar atmosphere to avoid metal oxidation and facilitate metal doping in ZIF‐8. Thereafter, CuNi@ZIF‐8 precursor was mixed with urea (N‐dopant) and phenylboronic acid (B‐dopant) and pyrolyzed at different temperatures (T = 700, 750, 800°C) under inert Ar atmosphere for 2 h at a heating rate of 5°C min−1. The black precipitates were collected and treated with acid to remove metal and/or metal oxide nanoparticles (if any) and washed with deionized water 2–3 times to remove impurities. CuNi@BNC‐T catalysts were then dried overnight under vacuum at 60°C. Likewise, control samples were pyrolyzed at an optimal temperature of 750°C to synthesize Cu@BNC‐750, Ni@BNC‐750 respectively. In addition, control CuNi@NC‐750 catalyst was synthesized using the pyrolysis of CuNi@ZIF‐8 precursor under same conditions, but without the addition of B‐dopant. Electrochemical half‐cell measurements: Half‐cell electrochemical CO2RR‐OER performance were conducted under room temperature conditions using a three‐electrode configuration in a two compartment H‐cell equipped with (working electrode (WE): catalyst coated NF, reference electrode (RE): Ag/AgCl/3 M KCl and counter electrode (CE): Pt mesh for OER, graphite rod for CO2RR, electrolyte: CO2‐saturated 0.1 M KHCO3, membrane: Nafion 117). Membrane was cleansed through boiling in deionized water for 0.5 h and later in H2O2 (5%) and 0.05 M H2SO4 for 0.5 and 1 h at 80°C, followed by rinsing with fresh deionized water. Catalyst ink was prepared by initially grinding the material in mortar pestle for 1 h and later dissolving 2.5 mg of catalyst powder in isopropyl alcohol (40 µL) and deionized water (960 µL), followed by ultrasonication in bath sonicator till homogeneously dispersed. NF substrate was cleansed using HNO3 for 1 min. under ultrasonication and dried in oven at 60°C. Catalyst ink was coated on the NF substrate and dried at room temperature to prepare WE. All electrochemical measurements including LSV, CA, EIS and cyclic voltammetry (CV) were performed using CHI‐760E Potentiostat. The electrolyte was initially purged with CO2 for 20 min., and a continuous flow was maintained (20 SCCM) throughout the electrochemical tests. Control experiments were conducted under Ar‐environments similarly. All the potentials were converted to reversible hydrogen electrode (RHE) using the conversion formula: In situ Raman analysis: In situ Raman measurements were conducted to evaluate the catalyst structural evolution during repeated CO2RR and OER, mimicking Zn‐CO2 battery cathode. An in situ electrochemical Raman cell was used comprising graphite rod CE, Ag/AgCl reference electrode, and catalyst‐loaded working electrode. CO2‐saturated 1 M KHCO3 electrolyte was used with the catholyte flow rate fixed at 20 rpm. A steady‐state experimental mode was used following measurements at open circuit potential followed by consecutive CO2RR (900 s), OER (1200 s), CO2RR (1500 s), and OER (1800 s). The CO2RR‐OER experiments were performed using CHI potentiostat at constant potentials of ‐0.4 V and 1.6 V (vs. RHE), respectively. All the Raman spectrum were recorded at the same scanning area repeatedly thrice. Theoretical analysis: Theoretical insights into the CO2RR and OER mechanisms over Cu and Ni single atoms coordinated with N atoms in as‐designed CuNi@BNC‐750 catalyst were investigated using DFT. DFT calculations were performed using the Projector Augmented Wave (PAW) method [56, 57] as implemented in the Vienna Ab initio Simulation Package (VASP) [58, 59]. The calculations were completed with a plane‐wave cut‐off energy of 600 eV and Gamma k‐points mesh of 3×3×1. The electronic self‐consistent calculation was converged to 1×10−6 eV, and ionic relaxation steps were performed using the conjugate‐gradient method (IBRION = 2) and continued until the total force on each atom dropped below a tolerance of 1×10−2 eV/Å. The generalized gradient approximation (GGA) was used for the exchange‐correlation functionals as parameterized by Perdew‐Burke‐Ernzerhof (PBE) [60]. The dispersion correction was also included in this study by using DFT D‐3 method [61]. Free energy reaction pathways for CO2RR and OER were calculated based on the mechanisms in Tables S6,S8 and S10. The corrections for entropy and zero‐point‐energy for H2, O2, and H2O have also been considered in the free energy calculations. Zn‐CO2 battery measurements: Zn‐CO2 battery was assembled in both H‐cell and flow‐cell configurations. For H‐cell configuration, an airtight two compartment H‐cell was used comprising Zn foil anode and NF‐based cathode with CuNi@BNC‐750 catalyst ink (8:1:1 ratio of catalyst (80 mg), polyvinylidene difluoride (PVDF, 10 mg) and Vulcan carbon (VC, 10 mg) in N‐methyl pyrrolidone solvent) coated on one side (loading: 5 mg cm−2) and gas diffusion layer (1:1 mixture of PVDF and VC in NMP) on the other side, separated by Nafion‐117 membrane separator. A mixture of 1 M KOH and 0.02 M Zn(CH3COO)2 was used as anolyte while the cathode was immersed in 1 M KHCO3 catholyte with continuous CO2 purging at 5 SCCM controlled by a mass‐flow controller. The battery cycling was conducted using NEWARE BTS4000 series battery cycler run by NEWARE BTS 8.1 software. Likewise, the flow‐type battery was assembled using the same components but using a flow‐cell, where both anolyte and catholyte were circulated using mechanical pumps at 20 rpm and CO2 gas was fed directly to the GDL side instead of purging in the catholyte. Author Contributions Divyani Gupta: conceptualization, methodology, data curation, investigation, validation, formal analysis, visualization, writing – original draft, writing – review and editing. Jinshuo Zou: formal analysis, writing – review and editing, data curation, investigation, visualization, writing – original draft. Anoja Kawsihan: data curation, investigation, formal analysis, visualization, writing – review and editing, writing – original draft. Jodie A. Yuwono: data curation, investigation, formal analysis, visualization, writing – review and editing, writing – original draft. Zhaoliang Wu: data ration, investigation, writing – review and editing, formal analysis. Yameng Fan: data curation, investigation, formal analysis, writing – review and editing. Jun Yu: data curation, formal analysis, writing – review and editing, investigation. Sailin Liu: data curation, investigation, writing – review and editing, formal analysis. Jianfeng Mao: conceptualization, methodology, validation, supervision, funding acquisition, writing – review and editing, visualization, project administration. Zaiping Guo: conceptualization, methodology, validation, visualization, funding acquisition, project administration, writing – review and editing, supervision. Funding Australian Research Council (Nos. CE230100017, DP240102353 and FL210100050) and Australian synchrotron (Nos. M23925, M23853, M23396, M22417, M21904, M23408 and M21848) Conflicts of Interest The authors declare no conflicts of interest. Supporting information Supporting File: adma74846‐sup‐0001‐SuppMat.docx. Acknowledgements We acknowledge Australian Research Council (CE230100017, DP240102353 and FL210100050) for funding support. Part of this work was carried out at the Soft X‐ray (SXR) beamline (M23925, M23853, M23396, M22417, M21904) and X‐ray absorption spectroscopy beamline (M23408, M21848) of the Australian Synchrotron. We acknowledge the Adelaide Microscopy center for field‐emission scanning electron microscope (FE‐SEM), transmission electron microscope (TEM) and scanning TEM measurements. This research is supported by Sandland Bequest Zou and AINSE Ltd. Early Career Researcher Grant (ECRG‐J. Zou). J.A.Y acknowledge the high‐performance computing supports from National Computing Infrastructure (NCI) Australia and support from ARC via Discovery Project Scheme (DE250101071). Open access publishing facilitated by Adelaide University, as part of the Wiley ‐ Adelaide University agreement via the Council of Australasian University Librarians Contributor Information Divyani Gupta, Email: divyani.gupta@adelaide.edu.au, Email: guptadiv15@gmail.com. Jianfeng Mao, Email: jianfeng.mao@adelaide.edu.au. Zaiping Guo, Email: zaipiguo@cityu.edu.hk. Data Availability Statement The data that supports the findings of this study are available in the Supporting Information of this article. References Associated Data Supplementary Materials Supporting File: adma74846‐sup‐0001‐SuppMat.docx. Data Availability Statement The data that supports the findings of this study are available in the Supporting Information of this article.