Tri-coordinated PdNP architecture for simultaneous capture, activation, and catalytic conversion of dilute CO₂via multisite synergy

Abstract

Palladium nanoparticles (PdNPs) are valuable for their unique properties, but their catalytic applications are frequently limited by the balance between the reactivity and coordination modes. Herein, we present a universal strategy for preparing ultrafine PdNPs (1c–4c, 1.6–2.9 nm) with multiple active sites through three coordination modes involving acetylglucose (AcGlu), N-heterocyclic olefin (NHO), and π bonds from AcGlu–MeIm–Pd (1b–4b). PdNP 3c demonstrates superior performance in carboxylative–cyclization-cross-coupling tri-component reactions involving dilute CO₂ (15%), propargylic amines, and aryl iodides, achieving the highest turnover frequency of 600 h⁻¹ while maintaining activity over 7 cycles. Mechanistic insights via in situ-NMR reveal that the multiple active sites enable simultaneous activation of all substrates, forming AcGlu–CO₂ philes and NHO–CO₂ adducts. Notably, these are the first PdNPs that enable the simultaneous capture, activation, and catalytic conversion of in situ-activated dilute CO₂, propargylic amine, and aryl iodide, potentially paving the way for new highly reactive transition metal NPs and novel nanomaterials.

---

Zhonggao Zhou

Zhonggao Zhou is a professor at the College of Chemistry and Materials Science at Gannan Normal University (GNNU) in China and has concurrently worked as Director of the Analytical and Testing Center since 2024. Before his PhD, he worked at Shicheng Middle School and GNNU. He obtained his PhD degree in Organic Chemistry from Fujian Normal University (FJNU) in 2017, under the supervision of Prof. Qidan Ling and Hongyu Zhen. He is very passionate about teaching and learning. His current research focuses on atomic-level construction of novel active sites from single atoms, dual-site metals, clusters and/or nanoparticles toward CO2 capture, activation, and thermo-catalytic conversion.

---

Introduction

The significant increase in carbon dioxide (CO₂) emissions from fossil fuels over the past 35 years has disrupted the global carbon balance and contributed to increasing temperatures, prompting intensified efforts to explore CO₂ as a safe, affordable, and abundant renewable C1 resource.^1–3 Current catalytic strategies for the efficient conversion of pure and dilute (5–30%, including industrial flue gas, coal-fired flue gas, etc.) CO₂ into valuable chemicals, including salicylic acid, cyclic carbonate, dimethyl carbonate, and 2-oxazolidinone (OZD), remain a significant challenge.^4–8 To achieve high conversion of CO₂ and other substrates to OZDs, all reactants must be maximally activated (including temperature, pressure, and additives, Fig. 1a, eqn (1)).^9,10 As proposed in our previous reports^11–13 and those of other scientists,^14–18 the simultaneous in situ capture and activation of CO₂ increase the efficiency of the catalytic reaction involved with dilute CO₂. A variety of noble metal catalysts, such as Pd,^19,20 Au,^21,22 and Ag,^23–25 have been investigated for their ability to transform CO₂ into OZDs. Among them, the Pd-catalyzed multicomponent reaction of CO₂ is widely recognized as a significant valuable method for preparing OZDs, enabling the generation of sophisticated molecules from commercially available or readily accessible starting materials.^26–28 Recently, the synthesis of a variety of metal–Se-containing OZDs through electrochemistry and Cu catalysts has reached a high level of maturity (Fig. 1a, eqn (2)).^29,30 Therefore, developing high performance metal catalysts to enhance the multicomponent reaction of dilute CO₂ into OZDs is highly desirable. However, these reactions often suffer from low efficiency, product diversity, and poor selectivity. In recent decades, substantial research has focused on enhancing catalytic performance of metal nanoparticles (MNPs)/clusters by adjusting their size, alloying, controlling the morphology, and controlling the microstructure of the ancillary ligand.^31–33 To date, the enhanced catalysis of MNPs using ancillary ligands has been influenced by several interrelated factors, including (1) the size effects of the MNPs,³⁴ (2) the effective surface control of the MNPs,^35,36 (3) supported substrate activation,³⁷ (4) the structural sensitivity of strong metal–ancillary ligand interactions,^38–40 and so on. As these factors are intertwined in nanocatalysts, achieving a precise understanding of how ancillary ligands influence the catalytic performance of MNPs is challenging. To fulfill these requirements for auxiliary ligands, N-heterocyclic carbenes (NHCs),^41–43 N-heterocyclic imines (NHIs),^44,45 and N-heterocyclic olefins (NHOs),^46–48 which possess suitable electron–donor capabilities and structural tunability, are promising candidates (Fig. 1b). Among them, well-established NHC ligands are commonly employed for high-performance metal surface modification.^31,49,50 Additionally, the N atom in the imidazolium heterocycle facilitates π–π interactions with metal species.⁵¹ As analogs of classic NHC ligands, NHO type ligands can function as either Lewis acids (LAs) or Lewis bases (LBs) depending on their substituents and are widely used in catalytic reactions.^48,52 The Lu group reported that the catalytic activity of NHO–CO₂ adducts is significantly greater than that of NHC–CO₂ adducts in the reactions of CO₂ and propargylic alcohols.⁴⁶ Additionally, Batista and colleagues demonstrated through density functional theory calculations that transition metal complexes incorporating NHO-type ligands hold substantial promise for CO₂ conversion.⁵³ Moreover, while NHOs in metal materials have been extensively reported for catalysis,^40,54,55 their use as supports for MNPs remains an emerging and unreported research direction, offering a promising opportunity to explore their synthesis and catalytic performance. In addition, previous reports have shown that acetylated sugars and alkane carboxylic acids can stabilize many MNPs, such as Fe, Co, Ni, Cu, and Zn.⁵⁶ Interestingly, the acetyl (Ac–) unit is a LA, unlike the majority of other capping agents, which are generally LBs and activate CO₂ through LAB interactions.⁵⁷ The resulting MNPs would be captured in the network of acetylated sugars. Furthermore, previous studies have shown that AcGlu, with its CO₂ affinity, can also activate CO₂ through strong hydrogen bonding.⁵⁸ Our findings further suggest that incorporating it as a modifying group in metal catalysts could enhance both CO₂ activation and the catalytic performance.^11–13

Fig. 1 (a) Synthesis of OZDs from propargylic amines. (b) Representative protection methods of NHC derivatives for MNPs. (c) A universal method for the preparation of PdNPs protected by AcGlu–MeIm. Right: schematic diagram of the structure of AcGlu–MeIm–PdNPs. Left: schematic diagram of the coordination mode of AcGlu–MeIm with PdNPs and CO₂.

Herein, we designed four ultrafine PdNPs (1.6–2.9 nm) with exceptional reactivity stabilized through a unique tri-coordination system combining acetyl group interactions, NHO coordination, and imidazolium π-bonding (AcGlu–MeIm–PdNPs 1c–4c, Fig. 1c) for the carboxylative–cyclization-cross-coupling tri-component reactions (4C-TCRs) involving dilute CO₂ (simulated flue gas, CO₂/N₂, vol/vol ratio 15 : 85), propargylic amines, and aryl iodides (Fig. 1a, eqn (3)).

AcGlu–MeIm–PdNPs can interact with CO₂ to form AcGlu–CO₂ philes and NHO–CO₂ adducts, facilitating the in situ capture, activation, and transformation of CO₂. PdNP 3c promotes the 4C-TCRs of CO₂, propargylic amine, and aryl iodide under mild conditions, resulting in OZDs with yields reaching up to 99% and a turnover frequency (TOF) of 600 h⁻¹. Moreover, this PdNP catalyst showed excellent stability during the catalytic process, maintaining its activity after 7 cycles. The innovative approach may be extended to synthesize ultrafine, highly reactive, and multi-coordinated MNPs, potentially leading to the development of novel nanomaterials with significant applications.

Results and discussion

Design, synthesis, and characterization of AcGlu–MeIm salts and PdNPs

At first, AcGlu–MeIm bromide salts (AcGlu–MeIm–Br 1a–4a, R = Me, ⁿBu, Bn, and PFB, Fig. 2a and S1†), each with a functional group at the N-position, were synthesized from D-glucose in three modified steps.¹² The last step was the direct reaction of N-substituted imidazole derivatives with 2-(2,3,4,6-tetra-O-acetyl-β-D-glucopyranosyloxy)-ethyl bromide (R*–Br). Subsequently, AcGlu–MeIm tetrachloropalladate salts (AcGlu–MeIm–Pd 1b–4b) were synthesized via a biphasic anion exchange reaction of the corresponding 1a–4a with sodium tetrachloropalladate (Na₂PdCl₄) in nearly quantitative yields (94–98%). These AcGlu–MeIm salts (1a–4a and 1b–4b) are soluble in water and polar organic solvents such as methanol (MeOH), ethanol (EtOH), acetone, N,N-dimethyl sulfoxide (DMSO), trichloromethane (TCM), dichloromethane (DCM), etc., but are insoluble in low-polarity to nonpolar solvents such as diethyl ether and petroleum ether. Finally, black solid PdNPs (AcGlu–MeIm–PdNPs 1c–4c) were synthesized by reducing 1b–4b using sodium borohydride (NaBH₄) in yields ranging from 96–99%. These PdNPs 1c–4c can be easily purified and redispersed in fresh solvent for further use. All the AcGlu–MeIm salts and PdNPs (1a–4a, 1b–4b, and 1c–4c) were characterized via nuclear magnetic resonance (NMR), Fourier transform infrared spectroscopy (FT-IR), high-performance liquid chromatography coupled with positive ion electrospray ionization high-resolution mass spectrometry (HPLC/ESI-HR-MS), ultraviolet-visible spectroscopy (UV), inductively coupled plasma-optical emission spectrometry (ICP-OES), and thermogravimetric analysis (TGA). PdNPs 1c–4c were further analyzed via high-resolution transmission electron microscopy (HR-TEM), selected-area electron diffraction (SAED), X-ray diffraction (XRD), and X-ray photoelectron spectroscopy (XPS).

Fig. 2 (a) Synthesis procedure and structural schematic of AcGlu–MeIm salts (1a–4a and 1b–4b) and PdNPs (1c–4c). (b) ¹H NMR spectra of 3a–3c. (c) ¹³C NMR spectra of 3a–3c. (d) FT-IR spectra of 3a–3c. (e) Expanded positive-mode LC-HR-MS of 3a–3c (the red line indicates the experimental data; the black line indicates the simulated isotope mode). (f) UV-vis absorption spectra of 3a–3c in EtOH. (g) TGA (black) and DTG (green) curves of 3c.

Taking AcGlu–MeIm salts (3a and 3b) and the corresponding PdNP 3c as examples, we can observe a slight chemical shift (upfield or downfield) of the sugar heterocycles in the range of 4.5–5.6 ppm, accompanied by signal broadening in the ¹H NMR spectra (Fig. 2b). Compared with those of 3a and 3b, the signal intensity of PdNP 3c decreased. Notably, the chemical shift of C₂–CH₃ exhibited continuous movement, moving from 2.76 ppm in AcGlu–MeIm–Br 3a to 3.01 ppm in the downfield region of AcGlu–MeIm–Pd 3b and ultimately returning to 2.76 ppm in the upfield region of PdNP 3c. Additionally, the integral area reduced from 3.00 to 2.28, suggesting that 24% of C₂–CH₃ has coordinated with PdNPs (Fig. 2b and S2†). Similar results were also found in the ¹³C NMR spectrum of C₂–CH₃ (Fig. 2c). However, the signals of C=C in the imidazolium ring exhibited unique changes, moving from 122.4 and 121.3 ppm in AcGlu–MeIm–Br 3a to 126.9 and 119.4 ppm in the downfield/upfield region of AcGlu–MeIm–Pd 3b and ultimately returning to 122.6 and 121.3 ppm of PdNP 3c. Compared with those of 3a and 3b, all the signals of PdNP 3c were broadened and weakened, likely influenced by the close contact between AcGlu, NHO, π bonds, and PdNPs. FT-IR analysis revealed a moderate absorption peak at 1369 cm⁻¹ for the C–N bond in the imidazolium ring of AcGlu–MeIm–Br 3a (Fig. 2d). After ion exchange to form AcGlu–MeIm–Pd 3b, the absorption intensity decreased. Reduction of AcGlu–MeIm–Pd 3b to PdNP 3c led to a further peak intensity decrease, yielding a weak absorption signal. PdNP 3c formation eliminated absorption peaks at 906, 1525, and 1582 cm⁻¹ in AcGlu–MeIm–Br 3a, while inducing a new peak at 800 cm⁻¹. Notably, in the C–H stretching region (2850–2900 cm⁻¹), C₂–CH₃ stretching vibration intensity markedly increased and red-shifted to 2926 cm⁻¹. Additionally, a Pd–CH₂–Im signal peak appeared at 2852 cm⁻¹. Overall, these experimental results indicate significant changes in the electronic environment of AcGlu–MeIm salts (3a and 3b) and PdNP 3c, suggesting unique coordination between PdNPs and C=O units in AcGlu groups, coordination of C2–CH₃ in the form of NHO, and π bonds in the imidazolium heterocycle. The Br⁻, PdCl₄²⁻, and PdNPs were undetectable in the HPLC/ESI-HR-MS analysis, with only AcGlu–MeIm cations observed. These results indicate consistency among AcGlu–MeIm salts (1a–4a and 1b–4b) and PdNPs (1c–4c) (Fig. 2e and Table S1†), further confirming that the AcGlu–MeIm cations could be stably detected via mass spectrometry. As shown in the UV-vis spectra (Fig. 2f), PdNP 3c exhibited a prominent absorption band at 273 nm, markedly different from those observed for 3a (258 nm) and 3b (263 nm). The CO₂ adsorption isotherms (Fig. S8a†) of PdNP 3c demonstrate a microporous-dominant structure (BET surface area: 62 m² g⁻¹) with a saturation capacity of 3.34 mmol g⁻¹ at 1.0 bar and 25 °C. Remarkably, the material maintains sustained CO₂ uptake under supercritical conditions (P/P₀ > 1, 1–100 kPa, Fig. S8b†) while exhibiting complete desorption reversibility (>95% cycling stability). Notably, it shows exceptional selectivity for trace CO₂ capture (0.1 wt%, Fig. S8c†), attributed to the generated AcGlu-CO₂ philes.

Owing to the hygroscopic nature of the AcGlu–MeIm–PdNPs, water evaporation occurred during pyrolysis according to TGA (Fig. 2g and S9†). TGA and derivative thermogravimetric analysis (DTG) indicated that the Ac– groups in PdNPs 1c–4c began to degrade between 155 and 200 °C. As the temperature increases between 270 and 288 °C, all the Ac– units decomposed. The imidazolium heterocycle was decomposed completely at 350–470 °C. The Pd contents of PdNPs 1c–4c determined by ICP-OES were 22.95, 18.60, 18.14, and 18.82% (Table S2†), respectively. Based on the combined TGA and ICP-OES results, the metal-to-ligand ratio ranged from 1.11 : 1 to 1.39 : 1, indicating a consistent composition across these samples. HR-TEM (Fig. 3a–d and S10–12†) revealed that the PdNPs 1c–4c had a narrow size distribution (Fig. S13†), with mean sizes of 2.1 (±0.4), 2.5 (±0.4), 2.4 (±0.4), and 2.0 (±0.4) nm, respectively. The main diffraction rings of PdNP 3c can be assigned to the Pd FCC phase corresponding to the (111), (311), and (222) crystal planes according to SAED (Fig. 3b).⁵⁹ The lattice fringe spacing of 0.227 nm can be ascribed to the (111) plane of Pd (Fig. 3c, left), whereas the lattice fringe spacing of 0.245 nm corresponds to the (311) plane of Pd (Fig. 3c, right). The XRD patterns of PdNPs 1c–4c (Fig. S14 and 15†) indicate that the peaks at 2θ values of approximately 39.79° and 80.27° correspond to the PDF#87-0641 (111) and (311) crystal planes, respectively. Notably, PdNP 3c demonstrated the highest level of crystallinity among these PdNPs. In summary, we can conclude that the metal is polycrystalline in PdNPs.

Fig. 3 (a) TEM image of 3c. Scale bar: 20 nm. (b) SAED pattern of 3c. (c) HR-TEM images of 3c. Enlarged image (top) and lattice fringes of metallic PdNPs (bottom). Scale bar: 5.0 nm. The red dashed box indicates the metallic Pd that formed in the aggregates. (d) Size distribution histogram of 3c. (e) Partial XPS spectrum of Pd 3d for 3c.

The XPS spectra of Pd 3d for PdNPs 1c–4c show two sets of peaks corresponding to the Pd(0) and Pd(II) states (Fig. S16†). Fitting analysis revealed that the binding energies of the Pd(0) 3d_5/2 and 3d_3/2 orbitals are approximately 340.4 and 335.2 eV, respectively, whereas the binding energies of the Pd(II) 3d_5/2 and 3d_3/2 orbitals are approximately 342.5 and 337.3 eV, respectively, for PdNP 3c (Fig. 3e).⁶⁰ The reduction rates of Pd(0) and Pd(II) for PdNPs 1c and 2c surpassed 50%, with PdNP 3c exhibiting the highest reduction efficiency (∼76%). The C 1s XPS spectrum for PdNP 3c was resolved into components corresponding to C–C/C–H (∼284.8 eV), C–N (∼401.1 eV), and C=O (∼532.1 eV). The Pd 3d, C 1s, N 1s, and O 1s XPS spectra of PdNPs 1c, 2c, and 4c exhibited similar results to those of PdNP 3c.

Catalytic activity evaluation

Initially, the catalytic performance of these designed Pd catalysts for the 4C-TCR of CO₂, N-(4-fluorobenzyl)but-2-yn-1-amine (5a), and 3-bromoiodobenzene (6a) was evaluated. To simulate real industrial waste gas generated from fossil fuels, we utilized a gas proportioner that delivers a continuous and stable source of CO₂ (Fig. S17;† the flow rate is set at 40 sccm). First, we evaluated the bases that generally have a significant influence on the performance of the catalytic system (Fig. 4b and Table S3†). Moderate to good yields (57–87%) of the product (Z)-5-(1-(3-bromophenyl)ethylidene)-3-(4-fluorobenzyl)-oxazoliidin-2-one (7a) were obtained. As expected, no product was detected without a base, suggesting that the base acts as a reaction accelerator.^61,62 Among various organic and inorganic bases, NaO^tBu was identified as the best choice, with an 87% yield, while the weak organic base trimethylamine (Et₃N) did not promote the reaction. The 4C-TCR was further optimized in various solvents. In addition to DMSO, other organic solvents (Diox, THF, MeCN, toluene, MeOH, and EtOH) hardly promoted the reaction (Fig. 4c and Table S3†). DMAc, NMP, DMF, and DEF, which are weakly basic organic amine solvents containing carbonyl groups, afforded catalytic yields of 15, 7, 0, and 0%, respectively. Notably, the formed DMSO–CO₂ complex may increase the activity and solubility of CO₂ in DMSO.⁶³

Fig. 4 (a) Pd-catalyzed 4C-TCR. PdNPs 1c–4c (1.0 mol%), 5a (0.2 mmol), 6a (0.1 mmol), and base (0.15 mmol) were dissolved in solvent (1.0 mL) at 60 °C and then placed in a Schlenk tube connected to a bottle of CO₂ through a gas proportioner with a 40 sccm flow rate for 3.0 h. All yields were determined via¹H NMR. (b) Yields were influenced by various bases, 4c (1.0 mol%, 0.69 mg), and DMSO. (c) Yields were influenced by different solvents, 4c (1.0 mol%, 0.69 mg), and NaO^tBu. (d) Yields were influenced by PdNPs 1c–4c (1.0 mol%). (e) Yields and TOFs were influenced by various volumes of DMSO, 3c (2.0 mol%, 1.18 mg), and 5 min. (f) Yields and TOFs were influenced by different Pd catalysts (2.0 mol%, 1.18 mg). (g) Assessment of the catalytic cycling performance of 3c (3.0 mol%, 1.77 mg). (h) The 4C-TCR scope of simulated flue gas, propargylic amines, and aryl iodides.

In addition, the DMSO–Pd complex reduced the activation barrier and improved the activity of the Pd catalyst.¹⁹ The effects of different temperatures on the PdNP 4c-catalyzed 4C-TCRs were further investigated (Table S4,† entries 1–4). At room temperature (RT, 25 °C), only a 25% yield of 7a was obtained, with yields decreasing from 90% to 81% as the temperature increased from 40 °C to 80 °C. Similar results were also obtained for the reaction catalyzed by PdNP 3c (entries 5–8). These results indicate that high temperatures are not conducive to this reaction, primarily because of the decomposition of the in situ-formed AcGlu–CO₂ philes, NHO–CO₂ adducts, and the DMSO–CO₂ complex at elevated temperatures.^12,64–66 The combination of NaO^tBu and DMSO was identified as the optimal reaction condition, resulting in a yield of 88% at 60 °C (Table S4,† entry 6). PdNP 3c with the Bn unit was observed to provide an excellent result with 88% yield (Fig. 4d and Table S5†), but PdNPs 1c, 2c, and 4c with Me, ⁿBu, and PFB substituents presented inferior results (75, 78, and 87%), indicating that the substituents have significant influences on the catalytic activity. Decreasing the catalyst dosage led to a notable decline in yield. Specifically, at dosages of 1.0 mol% (0.59 mg), 0.5 mol% (0.29 mg), 0.3 mol% (0.18 mg), and 0 mol% (0 mg), the yields were 87, 17, 11, and 0%, respectively (Table S5†). When AcGlu–MeIm–Pd 3b and 4b were used as catalysts, the yields were relatively low, at 55 and 50%, respectively. When simulated flue gas (CO₂/N₂, vol/vol ratio 15 : 85) was utilized as a CO₂ source for catalytic conversion, similar yields can also be obtained (88 and 87%) under the same conditions. Through the optimization of PdNP 3c and the reaction time, a 100% yield can be obtained with 2.0 mol% (1.18 mg) for 20 min and 3.0 mol% (1.77 mg) for 10 min (Table S6,† entries 4 and 6). The optimal catalytic performance was achieved with a catalyst-to-solvent ratio of 2.0 mol% (1.18 mg) catalyst to 2.0 mL solvent, yielding 100% and a TOF of 600 h⁻¹ (Fig. 4e and Table S7†). The results showed that as the concentration of PdNP 3c decreased, the yield of the product increased from 67% to 100%. When commercial catalysts such as Pd(OAc)₂, Na₂PdCl₄, Pd(PPh₃)₄, PdCl₂, PdCl₂(dppf), and Pd/C (10 wt%) were used, the reaction time was 3.0 h, and the catalysts demonstrated yields ranging from 33% to 86% with TOFs of 3.0–14 h⁻¹ (Fig. 4f and Table S8†). Compared with these homogeneous and nanocatalysts, PdNP 3c achieved a significantly higher yield (100%) and TOF (600 h⁻¹). PdNP 3c (3.0 mol%, 1.77 mg) was selected for the recycling reaction to evaluate its stability in the synthesis of 7a. Following each cycle, fresh substrates 5a and 6a were co-added into the catalytic system to assess the performance in the PdNP 3c-catalyzed 4C-TCR. Catalytic cycling without any post treatment was used to evaluate the stability of PdNP 3c. Notably, no significant loss of catalytic activity was observed after 7 consecutive cycles, and the yield and conversion exceeded 95% (Fig. 4g and Table S9†). The catalytic system maintained consistent coloration over seven cycles (Fig. S18†), demonstrating the robust stabilization of PdNP 3c against aggregation or leaching. This stability was further corroborated by ¹H NMR (Fig. S19†), which revealed complete retention of the characteristic signals for imidazolium (δ 7.6–7.7 ppm), glucopyranosyl ring (δ 3.8–5.4 ppm), and acetyl groups (δ 1.9–2.1 ppm). HR-TEM of PdNP 3c after seven cycles (Fig. S20a†) confirmed that the NPs remained well-dispersed without partial aggregation, exhibiting an average diameter of 3.0 ± 0.5 nm (Fig. S20b†), representing a 50% size increase from the initial size. The increase in crystallite size compared to the initial state is likely due to mild sintering or partial ligand detachment during cycling. These results demonstrate the exceptional stability and catalytic performance of PdNP 3c.

During the optimization process, we observed that the conversion rate of 5a varied significantly with the amounts of 5a and 6a and the reaction time. Consequently, a series of detailed kinetic experiments were conducted. The differential rate equation for the reaction was derived based on previous literature reports:⁶⁷

Y = −ln(1 − x) + α·[PA]₀·x (2)

The reaction rate (r) is governed by the intrinsic rate constant (k), initial propargylic amine concentration ([PA]₀), initial aryl iodide concentration ([ArI]₀), substrate inhibition constant (K_s), and conversion of propargylic amine (x). The reaction kinetics were investigated using eqn (1). By incorporating a logarithmic term for unconverted substrate and a concentration-dependent correction factor (α), eqn (2) was derived to linearize reaction progress data under substrate inhibition conditions. Assuming pseudo-first-order kinetics, the apparent reaction rate (k_app) was calculated from time-dependent conversion data (eqn (3)), where t represents reaction time. A linear relationship between ln(k_app) and ln([PA]₀) was analyzed.

The reaction displays first-order dependence on both propargylic amine and aryl iodide concentrations under intrinsic conditions (Fig. S21a–d†), consistent with a second-order (1 + 1) kinetic regime. However, substrate inhibition (Fig. S21e–f†) perturbs the apparent kinetics, leading to deviations from classical second-order behavior.

After the optimal 4C-TCR conditions were obtained, the scope of propargylic amines, aryl iodides, and simulated flue gas was explored. Sixteen examples of OZDs were synthesized with yields ranging from 81% to 100% (Fig. 4h and Table S10†). In all cases, the only byproducts were oxazolidinones (0% to 19%) formed via the carboxylative–cyclization of propargylic amines and CO₂. Propargylic amines featuring various R₁ groups, including both electron-withdrawing and electron-donating substituents such as F, CF₃, OMe, and ⁿBu, demonstrate excellent tolerance and afford target OZDs 7b–7e in yields ranging from 83% to 92%. Additionally, the substrate with 3,5-difluororophenyl can effectively generate 7f (88%) at a catalyst loading of 3.0 mol% (1.77 mg). Modifying the substituents on aryl iodobenzene with both electron-withdrawing and electron-donating groups, such as Cl, CF₃, CN, and Ac, yields favorable results, producing the desired OZDs 7h–7k (85–100%). Notably, when the F unit was used as a substituent, extending the reaction time resulted in a substantial yield of 81% (7k and 7l). Finally, flexible modification of the substituents of propargylic amines and aryl iodobenzenes successfully yields the desired OZDs 7m–7p (83–100%). The catalytic performance of PdNP 3c demonstrates remarkable superiority in 4C-TCRs (Table S11†), achieving quantitative yield (99%) with an unprecedented TOF (600 h⁻¹) in just 5 minutes at 60 °C. This catalyst exhibits unmatched efficiency (highest TOF), rapid kinetics (fastest reaction time), mild operating conditions, and superior atom economy (full conversion with only 1.18 mg loading), outperforming all reported systems. In summary, AcGlu–MeIm–PdNPs represent a novel and highly efficient Pd nanocatalyst, demonstrating remarkable catalytic efficiency across a wide range of substrates.

Mechanistic considerations

To elucidate the catalytic mechanism and actual active sites in the PdNP 3c-catalyzed 4C-TCR, we first performed numerous ¹³C isotope labeling experiments. Compared with the ¹³C NMR spectrum of the unlabeled product 7a, the relative intensity of the signal assigned to C5 by the chemical shift of 155.3 ppm of C=O in ¹³C-labeled 7a was significantly greater after the catalytic reaction was conducted in an atmosphere of ¹³CO₂ (Fig. S22†). Moreover, ¹³C isotope labeling FT-IR experiments further revealed that the signal of C=O in ¹³C-labeled 7a blueshifted from 1765 to 1768 cm⁻¹ (Fig. S23†). Therefore, by combining these ¹³C isotope labeling experiments, it can be concluded that CO₂ has been successfully converted into OZD in the PdNP 3c-catalyzed 4C-TCR. Second, the characteristic peaks of substrate 5a (1.80, 3.22, and 3.70 ppm) progressively diminished over time (Fig. 5a). In contrast, the signals corresponding to the newly generated species, Ar–CH₂ (I′, 4.40 ppm), N–CH₂–C (II′, 4.21 ppm), and C=C–CH₃ (III′, 1.98 ppm) in product 7a, markedly increased. These alterations suggest the involvement of the NH group of 5a in the reaction with CO₂, resulting in the formation of carbamate, which subsequently contributes to the generation of OZD. Moreover, understanding the interactions between the catalyst and substrates is essential. Through hydrogen–deuterium exchange ¹H NMR experiments in DMSO-d₆, we identified that the broad peak at 2.39 ppm was attributed to the NH of 5a (Fig. S24†). The addition of PdNP 3c to 5a caused a 19.35 Hz downfield shift in the NH peak, along with a significant decrease in signal intensity, especially after heating in ¹H NMR (Fig. S25†). Concurrently, from the ¹³C NMR data, C1, C2, and C3 exhibited downfield shifts of 6.63 and 6.09 Hz and upfield shifts of 6.46 Hz prior to heating, totaling 16.47, 14.14, and 15.55 Hz after heating, respectively (Fig. S26†). Notably, the C1 signal showed a downfield shift of 6.63 Hz, followed by an additional shift of 9.84 Hz after heating. These chemical shifts indicate that PdNP 3c facilitates the activation of the NH and C≡C units of 5a. The interactions between PdNP 3c and 6a were similarly determined via¹H NMR. The results revealed that PdNP 3c activated C1–H and C1–C on the aromatic ring of 6a while exerting minimal influence on other positions (Fig. S27 and S28†). These findings imply that C1, which is adjacent to the iodine in 6a, is the most active site for the reaction. The interactions of substrates 5a and 6a were investigated by separately and simultaneously adding PdNP 3c and NaO^tBu to their DMSO-d₆ solutions (Fig. 5b). The signals of Ar–CH₂ (I, 3.70 ppm), C≡C–CH₂ (II, 3.22 ppm), and NH (IV, 2.36 ppm) became weak and broad, indicating the activation of these protons in 5a. Notably, when NaO^tBu, PdNP 3c, and both were introduced together to the DMSO-d₆ solution of 5a, the NH (IV, 2.36 ppm) signal shifted upfield by 3.79, 7.09, and 13.57 ppm, respectively. CO₂ was bubbled into the DMSO-d₆ solution of PdNP 3c to identify the interaction sites between them, which was first evaluated at RT (Fig. 5c). The chemical shift of protons in imidazolium heterocycle, C2–CH₃, and Ac– units exhibited upfield shifts of 3.2, 2.0, and 1.5–1.9 Hz, respectively, in PdNP 3c. Upon heating to 60 °C for 30 min, the chemical shift values further increased to 5.6, 3.7, and 2.4–3.3 Hz. Additionally, the integral area of C2–CH₃ decreased from 2.28 to 2.06, after heating to 60 °C for 30 min and further decreased to 1.79 (Fig. 5c). To further elucidate the interactions between CO₂ and the catalytic system, CO₂ was bubbled through a DMSO-d₆ solution of PdNP 3c/NaO^tBu (Fig. 5d). The protons in imidazolium shifted downfield by 4.4 and 13.9 Hz, and the protons in Ac– shifted by 2.0 and 4.3 Hz at RT. The integral area of C2–CH₃ decreased from 2.28 to 2.10 and further decreased to 1.95 after the reaction with CO₂, suggesting that 6, 12, and 17% of NHO–CO₂ have been formed, respectively. This suggests that the combination of PdNP 3c/NaO^tBu/CO₂/heating for 30 min is optimal for NHO–CO₂. After the introduction of CO₂, the signal peak originally located at 10.1 ppm belonging to C3–C becomes weak, and a signal peak located at 173.8 ppm corresponds to the C=O bond of the NHO–CO₂ adducts newly generated through ¹³C NMR (Fig. S29†). After 240 h at RT and 4 h at 90 °C, the NHO–CO₂ adducts reverted back to C2–CH₃ partially. The results indicate that NHO–CO₂ adducts are formed and exhibit moderate stability. These findings highlight the pivotal roles of the 2-methylimidazolium and Ac– moieties in dilute CO₂in situ-capture and activation. On the basis of our experimental results and literature, we propose a plausible mechanism for the AcGlu–MeIm–PdNP-catalyzed 4C-TCR (Fig. 5e). Based on mechanistic investigations and prior literature,²⁶ we propose a catalytic cycle as follows: base-mediated deprotonation of the 2-methylimidazolium zwitterion (resonance hybrid A ↔ A′) generates the NHO (A′) intermediate, which subsequently activates CO₂ through nucleophilic attack (cycle A) to form NHO–CO₂ adducts (B). Simultaneously, the Ac– groups enhance local CO₂ concentration through the formation of AcGlu–CO₂–philes (C). Meanwhile, in situ-generated Pd(0) activates the C≡C bond of propargylic amine, while deprotonation of the NH group by NaO^tBu in propargylic amine (cycle B) affords intermediate (D). Subsequent electrophilic attack by CO₂ yields the carbamate intermediate (D′). Oxidative addition of Ar–I to Pd(0) forms the active intermediate Ar–Pd–I (E). Nucleophilic attack by the carbamate oxygen on the Pd-activated alkyne in F drives intramolecular cyclization to form G. Reductive elimination from the metal Pd(II) intermediate (G) regenerates Pd(0) and releases the product (H) (cycle C). The initial capture and activation of CO₂ by AcGlu–MeIm–PdNPs, along with the interaction between the C≡C bond in propargylic amines and Pd, facilitated the catalytic process. This catalyst exhibited a notable integrated synergistic effect, combining CO₂ capture, activation, and conversion within a spatially organized multifunctional catalytic system.

Fig. 5 (a) The reactions of 5a, 6a, and CO₂ were monitored by ¹H NMR. Reaction conditions: 5a (0.1 mmol), 6a (0.2 mmol), PdNP 3c (1.0 mol%, 0.59 mg), and CO₂ (15%) were reacted in DMSO-d₆ at 60 °C. (b) Activation of 5a and 6a in different systems was monitored by ¹H NMR in DMSO-d₆. (c) ¹H NMR spectra of PdNP 3c, PdNP 3c/CO₂, and PdNP 3c/CO₂ heated at 60 °C for 30 min. (d) ¹H NMR spectra of PdNP 3c/NaO^tBu, PdNP 3c/NaO^tBu/CO₂, and PdNP 3c/NaO^tBu/CO₂ heated at 60 °C for 30 min. (e) Plausible mechanism for PdNP 3c-catalyzed 4C-TCRs.

Conclusions

In summary, we designed four ultrafine and highly reactive PdNP catalysts (AcGlu–MeIm–PdNPs 1c–4c, 1.6–2.9 nm) protected by multiple active site ligands for 4C-TCRs involving simulated flue gas, propargylic amines, and aryl iodides. The NMR, XPS, and FT-IR spectroscopy results suggest that the PdNPs were stabilized in three coordination modes. HR-TEM, SAED, and XRD confirmed the existence of Pd in polycrystalline form. AcGlu–MeIm–PdNPs can react with CO₂ to form AcGlu–CO₂–philes and NHO–CO₂ adducts, facilitating the in situ–capture, activation, and transformation of dilute CO₂. The PdNP site functions as an electrophilic LA, initially activating the aryl iodide to create an intermediate, followed by the activation of the propargylic amine. This promotes the 4C-TCR of CO₂, propargylic amine, and aryl iodide under mild conditions, resulting in OZDs with yields reaching up to 99% and a record TOF of 600 h⁻¹. Moreover, PdNP 3c demonstrated excellent stability during the catalytic process, with its catalytic activity remaining unchanged even after 7 cycles. The robust electronic and moderate steric effects of AcGlu–MeIm cations are considered essential for achieving excellent catalytic performance. Owing to its wide applicability to various AcGlu–MeIm cations, solvent systems, MNPs, and broad reaction scope, this approach is highly promising for overcoming the compromise between the reactivity and stability of MNP-based materials in diverse fields, including catalysis, coordination chemistry, and materials science.

Data availability

The data that support the findings of this study are available in the main text and the ESI.†

Author contributions

Y. C. and Z. Z. conceived and designed the study, supervised the project, and acquired funding; Q. Z., J. L., and J. C. conducted the material synthesis, catalytic performance tests, and structural characterization; all authors participated in data analysis and interpretation; Q. Z., J. L., and J. C. drafted the manuscript with input from Y. C. and Z. Z.; all authors critically reviewed and approved the final manuscript. Q. Z., J. L., and J. C. contributed equally as co-first authors.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (22261002) and the Key Project of Jiangxi Provincial Natural Science Foundation (20232ACB203016).

Notes and references

1. C. Hepburn, E. Adlen, J. Beddington, E. A. Carter, S. Fuss, N. Mac Dowell, J. C. Minx, P. Smith and C. K. Williams, Nature, 2019, 575, 87–97.
2. R. M. Rapagnani, R. J. Dunscomb, A. A. Fresh and I. A. Tonks, Nat. Chem., 2022, 14, 877–883.
3. X. Liu, C. Wang, C. Chen, Z. Pan, C. Gao, W. Lai, J. Zhao, T. Tian and W. Xiao, Chem. Soc. Rev., 2025, 544, 216927.
4. S. Wang and C. Xi, Chem. Soc. Rev., 2019, 48, 382–404.
5. S.-L. Hou, J. Dong, X.-Y. Zhao, X.-S. Li, F.-Y. Ren, J. Zhao and B. Zhao, Angew. Chem., Int. Ed., 2023, 62, e202305213.
6. P. Rani, R. Das and C. M. Nagaraja, Inorg. Chem. Front., 2025, 12, 430–478.
7. M. Shen, W. Guo, L. Tong, L. Wang, P. K. Chu, S. Kawi and Y. Ding, Chem. Soc. Rev., 2025, 54, 2762–2831.
8. J. S. Stanley, H. N. Pauker, E. Kuker, V. Dong, R. J. Nielsen and J. Y. Yang, J. Am. Chem. Soc., 2025, 147, 16099–16106.
9. L. Li, Y. Lv, H. Sheng, Y. Du, H. Li, Y. Yun, Z. Zhang, H. Yu and M. Zhu, Nat. Commun., 2023, 14, 6989.
10. W.-Z. Qiao, Y.-J. Wang, S. Li, R. Wang, J. Wu and S.-Q. Zang, CCS Chem., 2024, 6, 2131–2141.
11. L. Kong, Z. Tao, Y. Li, H. Gong, Y. Bai, L. Li, X. Zhang, Z. Zhou and Y. Chen, Chem. Sci., 2024, 15, 15321–15331.
12. Z.-G. Zhou, P. He, J. Li, J. Zhang, G. Xu, S.-Y. Zhang, X.-X. Deng, Z.-Y. Du, G.-T. Luo, H.-Y. Zhen, Y. Chen and C.-T. He, Org. Chem. Front., 2023, 10, 2045–2053.
13. L. Kong, Y. Liu, Z. Zhou, Z. He, Y. Bai, G. Xu, Y. Pan, H. Zhen and Y. Chen, J. Catal., 2025, 443, 115992.
14. Y. Liang, Q. Xia, J.-Y. Yang, X.-J. Meng, X.-M. Hu and Y.-M. Pan, Chem. Eng. J., 2024, 498, 155305.
15. D. Sengupta, S. Bose, X. Wang, N. M. Schweitzer, C. D. Malliakas, H. Xie, J. Duncan, K. O. Kirlikovali, T. Yildirim and O. K. Farha, J. Am. Chem. Soc., 2024, 146, 27006–27013.
16. X.-S. Li, J. Zhao, S.-L. Hou, H. Xu, Z.-L. Liang, Y. Shi and B. Zhao, CCS Chem., 2024, 6, 1–14.
17. Y.-N. Gong, S.-Y. Lv, H.-Y. Yang, W.-J. Shi, J.-J. Wang, L. Jiang, D.-C. Zhong and T.-B. Lu, CCS Chem., 2024, 6, 3030–3040.
18. F.-Y. Ren, C. Hu, W.-B. Huang, L.-H. Duan, Y.-Z. Meng, X.-L. Li, Z. Fang, X.-Y. Zhao, W. Wang, X.-S. Li, J. Zhao, X.-Y. Zhang, S.-L. Hou, H. Xu, Y. Shi, L.-N. He and B. Zhao, Angew. Chem., Int. Ed., 2025, e202503898.
19. P. Brunel, J. Monot, C. E. Kefalidis, L. Maron, B. Martin-Vaca and D. Bourissou, ACS Catal., 2017, 7, 2652–2660.
20. Z.-G. Zhou, J. Zhang, Q. Xie, J. Li, P.-P. Xu, X.-F. Zhang, Z.-Y. Du, C.-T. He and X.-M. Chen, Renewables, 2024, 2, 405–413.
21. T. A. C. A. Bayrakdar, F. Nahra, J. V. Davis, M. M. Gamage, B. Captain, M. Temprado, M. Marazzi, M. Saab, K. V. Hecke, D. Ormerod, C. D. Hoff and S. P. Nolan, Organometallics, 2020, 39, 2907–2916.
22. D. Wen, Q. Zheng, S. Yang, H. Zhu and T. Tu, J. Catal., 2023, 418, 64–69.
23. R. Das, V. Parihar and C. M. Nagaraja, Inorg. Chem. Front., 2022, 9, 2583–2593.
24. Z. Zhang, Z.-H. Zhang, F. Zhou and J. Zhou, Org. Lett., 2021, 23, 2726–2730.
25. X. Chen, J. Y. Song, J. Zheng, Y. M. Wang, J. Luo, P. Weng, B. C. Cai, X. C. Lin, G. H. Ning and D. Li, J. Am. Chem. Soc., 2024, 146, 19271–19278.
26. P. García-Domínguez, L. Fehr, G. Rusconi and C. Nevado, Chem. Sci., 2016, 7, 3914–3918.
27. C. Zhou, Y. Dong, J. T. Yu, S. Sun and J. Cheng, Chem. Commun., 2019, 55, 13685–13688.
28. Z. Ma, P. Chen, C. Wu, Y. Liang and Y. Pan, Asian J. Org. Chem., 2025, 14, e202400492.
29. X.-Q. Zhou, H.-T. Tang, F.-H. Cui, Y. Liang, S.-H. Li and Y.-M. Pan, Green Chem., 2023, 25, 5024–5029.
30. T. K. Xiong, Q. Xia, X. Q. Zhou, S. H. Li, F. H. Cui, H. T. Tang, Y. M. Pan and Y. Liang, Adv. Synth. Catal., 2023, 365, 2183–2187.
31. M. Koy, P. Bellotti, M. Das and F. Glorius, Nat. Catal., 2021, 4, 352–363.
32. H. Guan, C. Harris and S. Sun, Acc. Chem. Res., 2023, 56, 1591–1601.
33. F. Zheng, W. Zhang, Q. Guo, B. Yu, D. Wang and W. Chen, Coord. Chem. Rev., 2024, 502, 215603.
34. R. J. Lewis, M. Koy, M. Macino, M. Das, J. H. Carter, D. J. Morgan, T. E. Davies, J. B. Ernst, S. J. Freakley, F. Glorius and G. J. Hutchings, J. Am. Chem. Soc., 2022, 144, 15431–15436.
35. Y. Zhang, J. Zhang, B. Zhang, R. Si, B. Han, F. Hong, Y. Niu, L. Sun, L. Li, B. Qiao, K. Sun, J. Huang and M. Haruta, Nat. Commun., 2020, 11, 558.
36. M. Cargnello, V. V. T. Doan-Nguyen, T. R. Gordon, R. E. Diaz, E. A. Stach, R. J. Gorte, P. Fornasiero and C. B. Murray, Science, 2013, 341, 771–773.
37. M. Esmaeilirad, Z. Jiang, A. M. Harzandi, A. Kondori, M. Tamadoni Saray, C. U. Segre, R. Shahbazian-Yassar, A. M. Rappe and M. Asadi, Nat. Energy, 2023, 8, 891–900.
38. Y. Wang, J. P. Chang, R. Xu, S. Bai, D. Wang, G. P. Yang, L. Y. Sun, P. Li and Y. F. Han, Chem. Soc. Rev., 2021, 50, 13559–13586.
39. J. Ren, M. Koy, H. Osthues, B. S. Lammers, C. Gutheil, M. Nyenhuis, Q. Zheng, Y. Xiao, L. Huang, A. Nalop, Q. Dai, H.-J. Gao, H. Mönig, N. L. Doltsinis, H. Fuchs and F. Glorius, Nat. Chem., 2023, 15, 1737–1744.
40. I. Berg, L. Schio, J. Reitz, E. Molteni, L. Lahav, C. G. Bolaños, A. Goldoni, C. Grazioli, G. Fratesi, M. M. Hansmann, L. Floreano and E. Gross, Angew. Chem., Int. Ed., 2023, 62, e202311832.
41. A. J. Arduengo III, R. L. Harlow and M. Kline, J. Am. Chem. Soc., 1991, 113, 361–363.
42. B. Qie, Z. Wang, J. Jiang, Z. Zhang, P. H. Jacobse, J. Lu, X. Li, F. Liu, A. N. Alexandrova, S. G. Louie, M. F. Crommie and F. R. Fischer, Science, 2024, 384, 895–901.
43. K. Miharu, N. Bunta, F. Hayato, Y. Kosuke, K. Takuya and T. Mamoru, Science, 2023, 379, 484–488.
44. T. Ochiai, D. Franz and S. Inoue, Chem. Soc. Rev., 2016, 45, 6327–6344.
45. J. Ren, M. Das, H. Osthues, M. Nyenhuis, B. Schulze Lammers, E. Kolodzeiski, H. Mönig, S. Amirjalayer, H. Fuchs, N. L. Doltsinis and F. Glorius, J. Am. Chem. Soc., 2024, 146, 7288–7294.
46. Y. B. Wang, Y. M. Wang, W. Z. Zhang and X. B. Lu, J. Am. Chem. Soc., 2013, 135, 11996–12003.
47. M. M. Hansmann, P. W. Antoni and H. Pesch, Angew. Chem., Int. Ed., 2020, 59, 5782–5787.
48. Q. Sun, A. Eitzinger, R. Esken, P. W. Antoni, R. J. Mayer, A. R. Ofial and M. M. Hansmann, Angew. Chem., Int. Ed., 2024, 63, e202318283.
49. H. Shen, G. Tian, Z. Xu, L. Wang, Q. Wu, Y. Zhang, B. K. Teo and N. Zheng, Coord. Chem. Rev., 2022, 458, 214425.
50. P. Bellotti, M. Koy, M. N. Hopkinson and F. Glorius, Nat. Rev. Chem., 2021, 5, 711–725.
51. M. R. Narouz, S. Takano, P. A. Lummis, T. I. Levchenko, A. Nazemi, S. Kaappa, S. Malola, G. Yousefalizadeh, L. A. Calhoun, K. G. Stamplecoskie, H. Häkkinen, T. Tsukuda and C. M. Crudden, J. Am. Chem. Soc., 2019, 141, 14997–15002.
52. S. Naumann, A. W. Thomas and A. P. Dove, Angew. Chem., Int. Ed., 2015, 54, 9550–9554.
53. A. P. de Lima Batista, A. G. S. de Oliveira-Filho and S. E. Galembeck, ChemistrySelect, 2017, 2, 4648–4654.
54. F. Landwehr, M. Das, S. Tosoni, J. J. Navarro, A. Das, M. Koy, M. Heyde, G. Pacchioni, F. Glorius and B. R. Cuenya, Adv. Mater. Interfaces, 2024, 11, 2400378.
55. T. Zevart, B. Pinter, M. Lozinsek, D. Urankar, R. D. Jansen-van Vuuren and J. Kosmrlj, Dalton Trans., 2024, 53, 8915–8925.
56. K. Bhattacharjee and B. L. V. Prasad, Chem. Soc. Rev., 2023, 52, 2573–2595.
57. P. Raveendran and S. L. Wallen, J. Am. Chem. Soc., 2002, 124, 7274–7275.
58. P. Raveendran and S. L. Wallen, J. Am. Chem. Soc., 2002, 124, 12590–12599.
59. S. Yue, C. S. Praveen, A. Klyushin, A. Fedorov, M. Hashimoto, Q. Li, T. Jones, P. Liu, W. Yu, M. G. Willinger and X. Huang, Nat. Commun., 2024, 15, 4678.
60. D. O. Prima, N. S. Kulikovskaya, R. A. Novikov, A. Y. Kostyukovich, J. V. Burykina, V. M. Chernyshev and V. P. Ananikov, Angew. Chem., Int. Ed., 2024, 63, e202317468.
61. H. Zhang, W. Zhang, F. Liu, Z. H. Luo, K. Gao and M. Liu, AlChE J., 2024, 70, e18488.
62. T. Qu, X. Duan, Z. Ma, F. Liu, J. Ma, K. Gao and M. Liu, Chem. Eng. Sci., 2024, 297, 120298.
63. F. Nahra and C. S. J. Cazin, Chem. Soc. Rev., 2021, 50, 3094–3142.
64. D. Yu and Y. Zhang, Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 20184–20189.
65. C. He, D.-H. Si, Y.-B. Huang and R. Cao, Angew. Chem., Int. Ed., 2022, 61, e202207478.
66. X. Guo, F. Zhang, Z. Yang, L. Gao, R. Wei and G. Xiao, Chem. Eng. J., 2023, 475, 146453.
67. S. R. Hackett, V. R. T. Zanotelli, W. Xu, J. Goya, J. O. Park, D. H. Perlman, P. A. Gibney, D. Botstein, J. D. Storey and J. D. Rabinowitz, Science, 2016, 354, aaf2786.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ta02845k

‡ Q. Z., J. L. and J. C. contributed equally to this work.

---