Self-supplying coreactant radical and structural distortion induced by carbonate ligand in metal–organic framework for anomalous deep-red Self-electrochemiluminescence

Abstract

Self-electrochemiluminescence (self-ECL) can deal with the problems of limited electron transport efficiency, reduced signal stability, and redundant ECL processes. A feasible solution is to assemble the luminescent components and coreactant radicals together. Here a self-ECL material was designed relying on a europium-based metal–organic framework (Eu-MOF, ZL-2) with 1,10-phenanthroline and CO₃²⁻ as ligands under strong alkaline conditions. Upon oxidation of the CO₃²⁻ ligand to produce C₂O₆²⁻, OH˙ could be generated to act as a coreactant radical, which reacted with the luminophore radical of oxidized ZL-2 to realize self-ECL without an extra coreactant. The coordination of the CO₃²⁻ ligand induced Eu³⁺ to occupy some sites in the tetrakaidecahedron structure of ZL-2, which caused the structure to distort, and thus triggered the unusual electric-dipole transfer transition ⁵D₀–⁷F₄. Therefore, ZL-2 was endowed with deep-red luminescence in the first near-infrared (NIR-I) region, which filled a gap in self-NIR-ECL applications. This discovery suggested a novel promising dual-function mechanism of the CO₃²⁻ ligand, undoubtedly broadening the application and development of MOFs in the ECL field.

---

Introduction

In light of its low background, wide dynamic range and simple operation, electrochemiluminescence (ECL) has emerged as a pivotal technology for clinical diagnosis, drug screening, food and environmental detection, and has been universally acknowledged.^1–3 In particular, ECL emission in the deep-red area displays superior optical penetration, less photodamage, and lower optical scattering,⁴ leading to important applications in detection, sensing, optical communication, night vision and other fields.^4–6 The luminophore, which generates high-energy species through redox reactions to realize ECL emission, is regarded as a decisive factor in ECL efficiency. However, in most cases, its high ECL performance is only achieved with the assistance of coreactants. The reaction between two phases reduces electron transport efficiency, which lowers the ECL signal and causes the stability to deteriorate. Moreover, the excessive use of coreactants not only causes redundancy in the design of the ECL process but also generates toxic by-products.^7,8 Therefore, the development of novel ECL materials integrating luminophore and coreactant is essential to broaden the application of ECL technology.^9,10

MOFs have the characteristics of large active specific surface area, self-assembly properties, etc.,^11–13 which can satisfy the demand for the independent design of a high-efficiency luminophore. To integrate the luminophore and coreactant in MOFs for self-ECL, Lei et al. used 9,10-di(p-carboxyphenyl)-anthracene (DPA) and 1,4-diazabicyclo[2.2.2]octane (D-H₂) as ligands and Zn²⁺ as a bridge to prepare an m-MOF.⁷ The presence of DPA as luminophore and D-H₂ as coreactant resulted in a self-ECL function for the m-MOF. They further used N,N-diethylenediamine (DEDA) as coreactant to synthesize a covalent organic framework (COF) luminophore for self-ECL, which exhibited an extraordinary 1008-fold signal compared with the ECL system of COF and external coreactant DEDA.¹⁴ These results indicate that the self-assembly of coreactants into the framework can greatly shorten the distance between the luminophore and the generated coreactant radical to accelerate the reaction rate, effectively improving the ECL efficiency of framework materials.

Coreactant radicals like OH˙ and SO₄˙⁻ can be generated by a functional ligand with carboxyl or sulfhydryl. Moreover, the oxidation of carbonate (CO₃²⁻) can also generate OH˙ through the mediation of peroxydicarbonate (C₂O₆²⁻).¹⁵ Under strong alkaline conditions, CO₃²⁻ can be produced through the decarboxylation of amino acids to provide multiple binding sites for the coordination of the metal ion as a bridge in the structure of MOFs.¹⁶ In view of the eminent ECL performance of lanthanide metal–organic frameworks (Ln-MOFs), the innate advantages of rich ladderlike electronic levels and 4f–4f transitions in Ln-MOFs, which endow them with high luminescent efficiency and long lifetime.^13,17,18 This work has designed a Eu-MOF (ZL-2) as a self-ECL emitter using glycine (Gly) to provide CO₃²⁻ ligand for the generation of coreactant radical OH˙ and 1,10-phenanthroline (Phen) as the first ligand to increase the structural stability. Surprisingly, ZL-2 showed a rare deep-red self-ECL due to the characteristic luminescence of ⁵D₀–⁷F₄ transition emission for Eu³⁺ and the self-production of OH˙. Fluorescence measurements and theoretical calculation demonstrated that the CO₃²⁻ ligand twisted the tetrakaidecahedron structure to form a distorted antiprism structure, which resulted in the electric-dipole ⁵D₀–⁷F₄ transition with the strongest luminescence intensity. Therefore, ZL-2 demonstrates the prospects of the CO₃²⁻ ligand in the ECL field, and could broaden the ECL application of MOF materials in clinical and environmental detection.

Results and discussion

Single crystal structural analysis of ZL-2

Single-crystal X-ray diffraction analysis showed crystallization of ZL-2 in the space group C12/c1 (Table S1†) via approximately two-fold of the asymmetric unit. The winglike asymmetric unit contained one Eu(III) ion, two coordinated Phen ligands, two coordinated CO₃²⁻ ligands and one coordinated water molecule (Fig. 1A). The nine-coordinated Eu(III) ions were occupied by five O atoms and four N atoms to form a [Eu₂(CO₃)₂(H₂O)₂] cluster with distorted tetrakaidecahedron coordination geometry (Fig. 1B and C), which contained a CO₃²⁻ bidentate ligand in the coordination. Five O atoms coordinated to one Eu(III) ion from one chelated η²-CO₃²⁻ ligand [Eu1–O1 = 2.395(3) Å and Eu1–O2 = 2.361(2) Å], one bridged μ₂-η2-CO₃²⁻ ligand [Eu1–O4 = 2.4375(6) Å and Eu1–O5 = 2.448(2) Å] and one water molecule [Eu1–O13 = 2.439(2) Å]. Four N atoms were from two chelating η²-Phen ligands and coordinated to one Eu(III) ion [Eu1–N1 = 2.602(3) Å, Eu1–N2 = 2.576(3) Å, Eu1–N3 = 2.576(3) Å and Eu1–N4 = 2.591(3) Å]. The infinite 2D parallelogram network structure of [Eu₂(CO₃)₂(H₂O)₂], SBUs were cross-connected to each other to form a 3D framework by μ₂-η²-CO₃²⁻ ligands and abundant H₂O molecules (Fig. 1C and D). The π–π stacking interactions resulting from the intersection of the coordinated Phen group and the neighbouring [Eu₂(CO₃)₂(H₂O)₂] SBUs stabilized the adjacent intersecting frameworks (Fig. 1D). In the whole ZL-2, the O–Eu–O, O–Eu–N and N-Eu–N bond angles were in the ranges of 53.75(8)–150.12(6)°, 70.17(9)^°–152.08(9)° and 62.41(8)–152.90(9)°, respectively. The coordinated O–C–O bond angles were 115.3(3)° and 118.3(2)°. To sum up, the bond lengths and angles of ZL-2 were in the normal ranges (Table S2†). The XRD patterns of ZL-2 and ZL-2S are exhibited in Fig. S1.†

Fig. 1 (A) Coordination environment of Eu(III) ions in ZL-2 (symmetry codes: #1 − x, y, 3/2 − z). (B) Coordination polyhedral geometry of Eu(III) ions. (C) The layer structure with multiple coordination units. (D) Stick model representation of a single 3D framework. Atom color scheme: C, dark grey; N, light blue; O, rosy; H, white; Eu, green.

Self-ECL mechanism of ZL-2

The self-ECL property of ZL-2 in PBS was first investigated with ECL–potential curves. Neither Eu³⁺-modified nor CO₃²⁻-modified GCEs showed any ECL response in the potential range of +0.4 to +1.8 V, while Phen-modified GCEs showed weak ECL emission peaking at about +0.9 V (Fig. S2†), which should be attributed to the oxidation of Phen to produce the excited species. In contrast, ZL-2-modified GCE showed a high ECL intensity of 6784 a.u. at +1.72 V in 0.1 M pH 6.38 PBS. Moreover, the ECL intensity depended on solution pH, and the highest ECL intensity occurred at pH 6.38 (Fig. 2A). This variation was unexpectedly consistent with the distribution score curve of carbonic acid. Therefore, the self-ECL behavior of ZL-2 was to some extent related to the CO₃²⁻ ligand in its structure. To confirm this speculation, a single-ligand Eu-MOF without CO₃²⁻ (ZL-2S) was synthesized for comparison of the ECL performance. Unlike ZL-2-modified GCE, ZL-2S-modified GCE showed only a very weak ECL peak at +1.28 V (Fig. 2B), similar to the ECL peak of Phen-modified GCE (Fig. S2†). The more positive ECL peak potential than free Phen-modified GCE could be attributed to coordination between Phen and metal ion Eu(III) to form a new luminous component in single-ligand Eu-MOF. Moreover, the self-ECL signals of ZL-2S under different pH conditions were not significantly different and were all of very low intensity, approximately 480 a.u. (Fig. S3†). Therefore, the much stronger ECL emission of ZL-2-modified GCE at +1.72 V further demonstrated the importance of the CO₃²⁻ ligand in the self-ECL process.

Fig. 2 (A) ECL–time curves of ZL-2-modified GCE in 0.1 M PBS at different pH. (B) ECL–potential curves of ZL-2-modified and ZL-2S-modified GCEs in 0.1 M PBS (pH 6.38). (C) Current–potential curves of rotating-ring-ZL-2-modified disk electrode at 1600 rpm in 0.1 M PBS (pH 6.38). Inset: the selectivity for H₂O₂ production at ZL-2-modified disk electrode. (D) EPR spectrum of ZL-2 in DMPO dispersion for CO₃˙⁻ and OH˙. (E) ECL transients of ZL-2-modified GCE with steps from 0 V to 1.2, 1.56 and 1.8 V in 0.1 M pH 6.38 PBS.

In the anodic process, ZL-2-modified GCE showed an oxidation peak at +1.56 V (Fig. S4†), consistent with the reported potential of the maximum degree of transformation from CO₃²⁻ to HCO₃⁻ linked to the production of CO₃˙⁻ and C₂O₆²⁻.^15,19 Thus, the ECL process included the oxidation of the CO₃²⁻ ligand to produce peroxydicarbonate C₂O₆²⁻, which could then react with H₂O to produce hydrogen peroxide (H₂O₂).^15,20 Subsequently, the redox process of ZL-2-modified GCE for the generation of H₂O₂ was studied using rotating ring-disk electrode measurement.²⁰ The produced H₂O₂ could be detected by the Pt ring electrode, which exhibited a modest onset potential of +0.38 V with a ring collection efficiency of 37% (Fig. 2C). According to calculation,²¹ the H₂O₂ selectivity was 50.13% under the 2e⁻ ORR process with the electron transfer number (n) approaching 2.7 (Fig. 2C and S5†). These results reflected the generation of H₂O₂ in the process of converting CO₃²⁻ into C₂O₆²⁻. Upon the production of H₂O₂, the OH˙ radical was activated.^19,22–24 An electron paramagnetic resonance experiment with DMPO as a radical scavenger demonstrated the production of CO₃˙⁻, accompanied by the obvious formation of OH˙ (Fig. 2D), confirming the occurrence of oxidization of CO₃²⁻ into C₂O₆²⁻ and the production of H₂O₂ in turn. Furthermore, the ECL transients test with different stepping pulses confirmed the function of OH˙ (Fig. 2E). At the initial stepping potential of 0 V, the ECL transient of ZL-2 was not observed until the potential was stepped to +1.2 V, at which the oxidation of CO₃²⁻ took place (Fig. S4†), and OH˙ was also generated. Afterward, the strongest degree of oxidation was reached at +1.56 V, while the ECL transient intensity of ZL-2 was 3420 a.u. Powered by the whole oxidation up to the potential of +1.8 V, the accumulation of OH˙ reached its maximum, leading to the highest ECL transient with a stable intensity of 6784 a.u. In addition, the ECL transient intensity of ZL-2S-modified GCE in PBS containing free CO₃²⁻ showed a similar growth trend with a lower signal value (Fig. S6†). Conversely, the ECL transient intensity of ZL-2S-modified GCE in PBS was maintained at the low value of 480 a.u. with increasing stepping potential (Fig. S7†). Therefore, these results indicated that the CO₃²⁻ ligand had a coreactant effect. The generated OH˙ prompted the self-ECL of ZL-2, thanks to the source of the CO₃²⁻ ligand.

To further explore the self-ECL mechanism, the luminous component was confirmed. As mentioned earlier, the ECL curve of Phen-modified GCEs showed a weak ECL peak at +0.9 V, while the ECL peaks of ZL-2-modified and ZL-2S-modified electrodes occurred at +1.72 and +1.28 V (Fig. S2† and 2B), indicating that the luminous component of as-synthesized framework ZL-2 was not the Phen ligand. Although the UV-vis spectra of both ZL-2 and ZL-2S exhibited the characteristic absorption peaks belonging to the Phen ligand, the existence of the second ligand CO₃²⁻ resulted in a red shift and extended the absorption band (Fig. S8†), revealing improved coordination stability.²⁵ Thus the CO₃²⁻ ligand changed the structure of the MOF.

The fluorescence spectra of ZL-2 showed excitation bands at 350 and 387 nm (Fig. 3A), which were matched with the ⁷F₀–⁵D₄ and ⁷F₀–⁵L₇ transitions of Eu³⁺, and the emission bands centered at 594, 618, 653 and 705 nm were ascribed to ⁵D₀–⁷F₁, ⁵D₀–⁷F₂, ⁵D₀–⁷F₃ and ⁵D₀–⁷F₄ transitions of Eu³⁺.^5,26 These phenomena reflected the antenna effect in ZL-2 that was the energy transfer process from the Phen ligand to Eu³⁺. Electron DFT calculation and natural transition orbital (NTO) analysis were used to theoretically verify the antenna effect in ZL-2. As shown in Fig. 3B, the transition from the ground state (S₀) to the excited singlet state (S₁) was concentrated in [Eu₂(CO₃)₂(H₂O)₂] SBUs, and the transition of NTO 276 to 277 was the dominant proportion (67.44%). In terms of the changes in electron cloud and fluorescence spectra, it could be concluded that the luminescence center in ZL-2 was the Eu unit. The energy gap (E_S1 − E_T1) of 2038.97 cm⁻¹ between S₁ and the triplet first excited state (T₁) of the Phen ligand as well as the energy difference of 5496.65 cm⁻¹ between T₁ of Phen and the excited state of Eu³⁺ (E_T1 − E_Eu³⁺) met the occurrence criteria of the antenna effect.²⁷ Thus the antenna effect could be described in a schematic diagram (Fig. 3C). These results confirmed that the self-ECL mechanism of ZL-2 was accomplished through the Eu unit as the luminophore and the CO₃²⁻ ligand as the coreactant. The oxidation of CO₃²⁻ produced CO₃˙⁻ to form C₂O₆²⁻ and then reacted with water to produce H₂O₂ followed by OH˙.^15,24 The OH˙ radical reacted with the electro-oxidized product of ZL-2 (ZL-2˙⁺ radical) to generate excited ZL-2*, accompanied by energy transfer from the Phen ligand to the central Eu unit, thus achieving self-ECL. Moreover, the process of CO₃²⁻ oxidation to induce the formation of OH˙ created a cycle to supply CO₃²⁻ continuously, maintaining the steady ECL signal. The self-ECL mechanism diagram of ZL-2 is shown in Fig. 3D.

Fig. 3 (A) Excitation and emission spectra of ZL-2. (B) NTO isodensity surfaces of S₁ for ZL-2 clusters (isovalue = 0.02 a.u.; the numbers show the contributions of S₁ to the orbital). Green and blue regions represent the positive and negative orbital phases, respectively. (C) Simple model for the energy transfer processes of the Phen ligand to Eu³⁺ and the f–f transition emission mechanism of Eu³⁺ in ZL-2. (D) Schematic diagram of the deep-red self-electrochemiluminescence mechanism in ZL-2.

Ligand-induced deep-red emission of self-ECL

In addition to acting as a coreactant and providing radicals for the self-ECL of ZL-2, the introduction of the CO₃²⁻ ligand affected ZL-2 at the structural level, triggering ZL-2 to produce a rare deep-red luminescence. The emission peak of the ⁵D₀–⁷F₄ transition at 705 nm for ZL-2 (Fig. 3A) was different from those of common europium-based materials, in which the strongest emission peak was attributable to the ⁵D₀–⁷F₂ transition at 618 nm.⁵ Similarly, the strongest emission peak of ZL-2S was also located at 618 nm, and even the peak at 705 nm was not observed (Fig. 4A), indicating that the presence of CO₃²⁻ induced in ZL-2 the ability for NIR luminescence. The luminescence intensity was enhanced after introducing CO₃²⁻, besides the change in the dominant emission transition, which was discovered from a comparison of emission peak intensities (Fig. S9†) and the enhanced ECL signals in the presence of CO₃²⁻ (Fig. S6 and 7†). To investigate the effect of CO₃²⁻, the fluorescence lifetimes of two major emission wavelengths at 618 and 705 nm for ZL-2 and ZL-2S were measured respectively (Fig. 4B–D). Compared to ZL-2S, the fluorescence lifetime of ZL-2 at 618 nm showed a slight increase from 579.38 μs to 616.05 μs. In particular, the lifetime at 705 nm was significantly improved from 397.62 μs to 534.76 μs. These phenomena brought about geometric distortion due to the addition of CO₃²⁻.^5,28,29 The CO₃²⁻ ligand led to the substitution of Eu³⁺ to certain sites in the original tetrakaidecahedron structure (Fig. 4E). Therefore, the coordination polyhedron in ZL-2 presented a distorted square antiprism structure to trigger intense ⁵D₀–⁷F₄ transition emission, inducing brighter deep-red luminescence, which could be observed under illumination by a UV lamp (365 nm) (Fig. 4E). The designed synthesis of the deep-red ZL-2 was a significant discovery for self-NIR-ECL applications.

Fig. 4 (A) Excitation and emission spectra of ZL-2S. (B) CIE chromaticity diagram of ⁵D₀–⁷F₁, ⁵D₀–⁷F₂ and ⁵D₀–⁷F₄ transitions in ZL-2. FL lifetimes for ZL-2 and ZL-2S under emission wavelengths of (C) 705 nm and (D) 618 nm. (E) Comparison of distorted tetrakaidecahedron structure of ZL-2 with the normal tetrakaidecahedron model, and corresponding luminescence photographs of ZL-2S and ZL-2 under a UV lamp (365 nm).

Conclusions

An Eu-MOF featuring self-NIR-ECL emission was designed with Phen as the first ligand and CO₃²⁻ provided by Gly under strong alkaline conditions as the second ligand. Along with the antenna effect from the Phen ligand to the luminescence center Eu unit, the CO₃²⁻ ligand can provide an OH˙ radical for the generation of excited Eu-MOF (ZL-2*) upon oxidation of CO₃²⁻ in an anodic ECL process. Meanwhile, the introduction of the CO₃²⁻ ligand twisted the tetrakaidecahedron structure of ZL-2 by inducing Eu³⁺ to occupy some sites, triggering the unusual electric-dipole transfer transition of ⁵D₀–⁷F₄. Moreover, the cycle of the CO₃²⁻ ligand further enhanced the transition, leading to strong ECL emission in the deep-red region. This work extended the reported self-ECL emitters to the NIR field and broadened the function of the CO₃²⁻ ligand, which are beneficial to the development of self-ECL in related applications.

Data availability

All experimental procedures and ESI tables and figures are available in the ESI.† Crystallographic data for the structures reported in this paper are also deposited in the Cambridge Crystallographic Data Center with CCDC reference numbers 2427309 for the ZL-2.

Author contributions

L. Z. and X. S. designed the research, analyzed the data and wrote the manuscript. Z. X. carried out the experiments. C. D. provided the funds. H. J. supervised the research and writing. All authors have participated in modification of the manuscript and given approval to the final manuscript.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This study was supported by the National Natural Science Foundation of China (22404090 and 22374086), the Taishan Scholar Foundation of Shandong Province (tstp20231227), the Natural Science Foundation of Shandong Province (ZR2024QB064). All of authors express their sincere thanks.

Notes and references

1. Z. Ding, B. M. Quinn, S. K. Haram, L. E. Pell, B. A. Korgel and A. J. Bard, Science, 2002, 296, 1293–1296.
2. X. Song, X. Ren, W. Zhao, L. Zhao, S. Wang, C. Luo, Y. Li and Q. Wei, Anal. Chem., 2022, 94, 12531–12537.
3. X. Song, L. Zhao, C. Luo, X. Ren, L. Yang and Q. Wei, Anal. Chem., 2021, 93, 9704–9710.
4. Y. Hong, W. Geng, T. Zhang, G. Gong, C. Li, C. Zheng, F. Liu, J. Qian, M. Chen and B. Tang, Angew. Chem., Int. Ed., 2022, 61, e202209590.
5. H. Zhang, Z. Yang, H. Li, J. Li, C. Liu, H. Jiang, Y. Liu, R. Wang, Z. Xie and J. Zhu, Laser Photonics Rev., 2024, 18, 2400463.
6. M. Hesari, M. S. Workentin and Z. Ding, Chem. Sci., 2014, 5, 3814–3822.
7. D. Zhu, Y. Zhang, S. Bao, N. Wang, S. Yu, R. Luo, J. Ma, H. Ju and J. Lei, J. Am. Chem. Soc., 2021, 143, 3049–3053.
8. X. Wang, S. Xiao, C. Yang, C. Hu, X. Wang, S. Zhen, C. Huang and Y. Li, Anal. Chem., 2021, 93, 14178–14186.
9. N. Wang, H. Gao, Y. Li, G. Li, W. Chen, Z. Jin, J. Lei, Q. Wei and H. Ju, Angew. Chem., Int. Ed., 2021, 60, 197–201.
10. L. Li, X. Hu, Z. Tang, C. Wang and H. Ju, Anal. Chem., 2024, 96, 4308–4313.
11. D. E. Barry, D. F. Caffrey and T. Gunnlaugsson, Chem. Soc. Rev., 2016, 45, 3244–3274.
12. D. Jiang, C. Huang, J. Zhu, P. Wang, Z. Liu and D. Fang, Coord. Chem. Rev., 2021, 444, 214064.
13. L. Zhao, X. Ren, Y. Du, Z. Gao, H. Ma, H. Wang, Y. Li, Q. Wei, H. Ju and D. Wu, Adv. Funct. Mater., 2024, 34, 2410886.
14. X. Meng, L. Zheng, R. Luo, W. Kong, Z. Xu, P. Dong, J. Ma and J. Lei, Angew. Chem., Int. Ed., 2024, 63, e202402373.
15. I. A. Fiorani, G. Valenti, N. Kamoshida, F. Paolucci and Y. Einaga, J. Am. Chem. Soc., 2020, 142, 1518–1525.
16. S. J. Wang, M. Wahiduzzaman, L. Davis, A. Tissot, W. Shepard, J. Marrot, C. Martineau-Corcos, D. Hamdane, G. Maurin, S. Devautour-Vinot and C. Serre, Nat. Commun., 2018, 9, 4937.
17. X. Wang, K. Gopalsamy, G. Clavier, G. Maurin, B. Ding, A. Tissot and C. Serre, Chem. Sci., 2024, 15, 6488–6499.
18. L. Zhao, X. Song, X. Ren, H. Wang, D. Fan, D. Wu and Q. Wei, Biosens. Bioelectron., 2021, 191, 113409.
19. L. Xia, F. Chen, J. Li, S. Chen, J. Bai, T. Zhou, L. Li, Q. Xu and B. Zhou, J. Hazard. Mater., 2020, 389, 122140.
20. Y. Wen, Y. Feng, J. Wei, T. Zhang, C. Cai, J. Sun, X. Qian and Y. Zhao, Chem. Sci., 2024, 15, 18969–18976.
21. W. Gu, L. Hu, W. Hong, X. Jia, J. Li and E. Wang, Chem. Sci., 2016, 7, 4167–4173.
22. X. Dou, Q. Zhang, S. N. A. Shah, M. Khan, K. Uchiyama and J. Lin, Chem. Sci., 2019, 10, 497–500.
23. S. T. Putnam and J. Rodríguez-López, Chem. Sci., 2024, 15, 10036–10045.
24. H. Zhu, X. Lv, Y. Wu, W. Wang, Y. Wu, S. Yan and Y. Chen, Nat. Commun., 2024, 15, 8846.
25. H. Wang, P. He, H. Yan and M. Gong, Sens. Actuators, B, 2011, 156, 6–11.
26. L. Zhao, X. Song, X. Ren, D. Fan, Q. Wei and D. Wu, Anal. Chem., 2021, 93, 8613–8621.
27. W. Zhou, W. Lin, Y. Chen, X. Dai, Z. Liu and Y. Liu, Chem. Sci., 2022, 13, 573–579.
28. Z. Sun, J. Li, H. Qian, Y. Sakka, T. S. Suzuki and B. Lu, J. Adv. Ceram., 2024, 13, 113–123.
29. P. Hu, X. Guo, J. Hu, C. Deng and R. Cui, Chem. Rev., 2024, 12, 2301760.

---

Footnote

† Electronic supplementary information (ESI) available: X-ray crystallographic files (CIF), chemicals and apparatus, Experimental procedures, Tables S1, S2 and Fig. S1–S9. CCDC 2427309. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d5sc02359a

---