Fullerene–porphyrin hybrid nanoparticles that generate activated oxygen by photoirradiation

Abstract

The preparation of water-dispersible hybrid nanoparticles comprising fullerene and porphyrin from cyclodextrin complexes is described. In the presence of polyethylene glycol, C₆₀ fullerene and porphyrin were expelled from the cyclodextrin cavity to form fullerene–porphyrin hybrid nanoparticles in water. The fullerene–porphyrin hybrid nanoparticles exhibit improved singlet oxygen generation ability under photoirradiation compared with that of C₆₀ nanoparticles.

---

Water-dispersible colloidal fullerene assemblies, referred to as fullerene nanoparticles (NPs), have recently received increasing attention.^1–3 Fullerene NPs are negatively charged and can be dispersed in water in the absence of any solubilizer. Fullerene NPs demonstrate promise within biological and medical applications, as radical scavengers and photosensitizers for photodynamic therapy. To further extend the applications of fullerene NPs, additional hybridization with desired functional molecules is required. Porphyrin and associated derivatives are highly promising candidates for hybridization with fullerenes to increase photoactivity.⁴ Numerous studies on the complexation of fullerenes with porphyrin molecules using synthetic organic chemistry^5–7 or supramolecular chemistry^8–10 have been reported. Although fullerene NPs have been intensively studied over the last decade, no reliable method to achieve the hybridization of porphyrins with fullerene NPs has been proposed.

Poly(ethylene glycol) monomethyl ether (PEG) was recently observed to accelerate the decomposition of fullerene C₆₀–γ-CD complexes in water, which leads to the rapid aggregation of C₆₀ to form water-dispersible C₆₀ NPs.¹¹ In this method, C₆₀–γ-CD complexes can exist as stable isolated molecules in water, enabling the precise size control and step-wise growth of C₆₀ NPs.^12,13 Herein, the preparation of hybrid NPs comprising C₆₀ and hydrophobic porphyrin molecules are reported. C₆₀–γ-CD and porphyrin-trimethyl-β-cyclodextrin (por–TMe-β-CD) complexes are mixed in water in the presence of PEG. Both complexes decompose through the interaction of PEG with the CDs, leading to the formation of C₆₀–porphyrin hybrid NPs (denoted as C₆₀–por NPs). The C₆₀–por NPs are negatively charged and easily disperse in water. Additionally, the ability of C₆₀–por NPs to generate activated oxygen is also evaluated.

The C₆₀–γ-CD complex^14,15 and 1–TMe-β-CD complex (Fig. 1)^16–18 were prepared according to a previously described procedure (see the ESI† for details). The ¹H NMR spectrum of the mixed solution comprising the C₆₀–γ-CD complex and PEG after 1 h incubation at 80 °C shows that the peaks attributed to the C₆₀–γ-CD complexes completely disappeared (Fig. S1†). Hence, the ¹H NMR data confirm the decomposition of the C₆₀–γ-CD complexes and the formation of water-dispersible C₆₀ NPs, as previously reported.^11–13 The effect of PEG on 1–TMe-β-CD complexes was also investigated by ¹H-NMR as shown in Fig. S2.† After incubating the mixed solution of 1–TMe-β-CD complex and PEG ([1] = 0.1 mM, [PEG] = 5.0 g L⁻¹) for 1 h at room temperature, peaks attributed to the 1–TMe-β-CD complex were still evident at 4.97 ppm and above 7.7 ppm (Fig. S2(i)†). Hence, PEG has no influence upon the 1–TMe-β-CD structure at room temperature. Conversely, after incubating for 1 h at 80 °C, a dark purple precipitate formed and the aforementioned ¹H NMR peaks completely disappeared (Fig. S2(ii)†). In the absence of PEG, the 1–TMe-β-CD complex was stable in water both at room temperature (Fig. S2(iii)†) and 80 °C (Fig. S2(iv)†). These results suggest a decomposition route of 1–TMe-β-CD by interaction with PEG at 80 °C, with a concomitant formation of non-dispersible large aggregates.

Fig. 1 Chemical structures of porphyrin derivatives used in this study.

To obtain hybrid NPs comprising C₆₀ and 1 (C₆₀–1 NPs), PEG (M_w = 2000) was added to an aqueous solution containing C₆₀–γ-CD and 1–TMe-β-CD complexes ([C₆₀] = [1] = 0.1 mM, [PEG] = 5.0 g L⁻¹), which were thereafter incubated at room temperature or 80 °C. The ¹H NMR spectrum of the mixed solution at room temperature shows peaks attributed to γ-CD in the C₆₀–γ-CD complex at 5.03 ppm, TMe-β-CD in the 1–TMe-β-CD complex at 4.97 ppm, and 1 in the 1–TMe-β-CD complex in the region of 7.6–8.5 ppm (Fig. 2a(i)). The data indicate that PEG fails to induce the decomposition of the C₆₀–γ-CD and 1–TMe-β-CD complexes at room temperature. Conversely, after the mixture was heated at 80 °C for 1 h, the peaks attributed to the C₆₀–γ-CD and 1–TMe-β-CD complexes completely disappeared (Fig. 2a(ii)). Hence, C₆₀–γ-CD and 1–TMe-β-CD were decomposed at 80 °C, in the presence of PEG. The solution after being subjected to heat treatment at 80 °C for 1 h, was dark purple in the absence of any precipitate. The hydrodynamic diameter and ζ-potential of the reacted solution were 125 nm (polydispersity index = 0.21) and −20.2 mV, respectively. Water dispersible fullerene NPs typically exhibit negative ζ-potentials, the origin of which still requires elucidating.^19,20 Hence, the formation of water-dispersible nano-composites, C₆₀–1 NPs, is suggested.

Fig. 2 (a) ¹H NMR spectra of mixed solutions comprising fullerene C₆₀–γ-cyclodextrin (CD) and 1–trimethyl (TMe)-β-CD complexes ([C₆₀] = [1] = 0.1 mM) (i) before and (ii) after heating at 80 °C for 1 h, in the presence of polyethylene glycol (PEG) (5 g L⁻¹). Open circles: free γ-CD, filled circles: C₆₀–γ-CD, open diamonds: free TMe-β-CD, and filled diamonds: porphyrin–TMe-β-CD (por–TMe-β-CD) complex. The spectra at 7.6–8.5 ppm, are amplified five-fold. (b) Ultraviolet-visible (UV/Vis) absorption spectra of the mixed solution comprising C₆₀–γ-CD and 1–TMe-β-CD complexes ([C₆₀] = [1] = 0.1 mM) with PEG (5 g L⁻¹), before (dashed line) and after (solid line) heating at 80 °C for 1 h. (c) UV/Vis absorption spectra of the mixed solution comprising the C₆₀–γ-CD complex as a function of the 1–TMe-β-CD complex concentration ([C₆₀] = 0.1 mM, [1] = 0–0.2 mM) with PEG (5 g L⁻¹) after heating at 80 °C for 1 h.

The por–TMe-β-CD complexes using 2–6 (Fig. 1), were also prepared adopting the same procedure as that for the 1–TMe-β-CD complex. Each por–TMe-β-CD complex solution was mixed with C₆₀–γ-CD and PEG ([C₆₀] = [por] = 0.1 mM, [PEG] = 5.0 g L⁻¹). The ¹H NMR spectrum of each individual mixed solution after being incubated for 1 h at 80 °C, is shown in Fig. S3.† In the ¹H NMR spectrum of the mixture comprising C₆₀–γ-CD and 2–TMe-β-CD, the peaks attributed to these complexes at 5.03, 4.98, and 7.60–8.50 ppm, completely disappeared after being subjected to incubation for 1 h at 80 °C, without precipitation (Fig. S3a†). Similar changes in the ¹H NMR spectrum of the mixture comprising C₆₀–γ-CD and 3–TMe-β-CD complexes are observed, as shown in Fig. S3b.† The data suggest that the 2–TMe-β-CD and 3–TMe-β-CD complexes can be decomposed in a similar manner as the C₆₀–γ-CD complexes, and imply the formation of C₆₀–2 and C₆₀–3 NPs.

The ¹H NMR spectra of the mixed solutions comprising C₆₀–γ-CD and 4, 5, or 6–TMe-β-CD complexes failed to show peaks attributed to the C₆₀–γ-CD complex, and peaks associated with the respective por–TMe-β-CD complexes were observed after incubation for 1 h at 80 °C (Fig. S3c–e,† respectively). Hence, the data suggest that the C₆₀–γ-CD complex decomposed in the presence of PEG, and the 4, 5, and 6–TMe-β-CD complexes were observed to be stable without decomposition at 80 °C. There have been reports suggesting the strong interaction of water-soluble tetraphenyl porphyrins with TMe-β-CDs.^21,22 Polar substituents prompt the penetration of the polarized porphyrin rims into the TMe-β-CD cavity. Porphyrins 4–6 possess polar substituents, which are suggested to enable the formation of stable TMe-β-CD complexes. Furthermore, the size of the β-CD cavity is sufficiently narrow to prevent any strong interaction with PEG.²³ Thus, PEG-induced decomposition of the 4, 5, and 6–TMe-β-CD complexes is not possible.

The absorption behavior of C₆₀–1 NPs was investigated using ultraviolet-visible (UV/Vis) spectroscopy. In the UV/Vis spectra, the characteristic peak of solvated C₆₀–γ-CD, at 333 nm shifted to 344 nm after heating at 80 °C for 1 h (Fig. 2b). An additional broad absorption at 400–550 nm is also apparent, which is characteristic of solid-state crystalline C₆₀ and arises from the electronic interactions between adjacent C₆₀ molecules.^24,25 The characteristic peak of the solvated 1–TMe-β-CD complex at 415 nm, shifted to 432 nm, with induced broadening after being subjected to heat treatment at 80 °C for 1 h (Fig. 2b). In the absence of C₆₀–γ-CD complexes, the characteristic absorption peak attributed to the solvated 1–TMe-β-CD complex completely disappeared after heating for 1 h at 80 °C, in the presence of PEG (Fig. S4†). The data show that 1, when expelled from the TMe-β-CD cavities, forms non-dispersible precipitates in the absence of C₆₀. For 1 dispelled from the TMe-β-CD cavities to be stably dispersed in water, formation of co-aggregates with C₆₀ may be a prerequisite. C₆₀–2 and C₆₀–3 NPs also show similar UV-Vis absorption spectra after being subjected to heating at 80 °C for 1 h, as shown in Fig. S5a and b,† respectively.

To further elucidate the composite formation of C₆₀ and 1, the influence of 1 concentration on C₆₀–1 NP formation was investigated by UV/Vis spectroscopy (Fig. 2c). The intensity of the absorption peak at 432 nm increased as a function of 1 concentration from 0.05 to 0.1 mM. Conversely, the absorption peak at 345 nm, derived from the formation of fullerene NPs, shifted to 338 nm, with increasing concentration of 1. This absorption peak derives from the fullerene nanoparticle size, and as the size decreased (i.e., the NPs became smaller), the peak blue-shifted.¹¹ The results suggest that in the presence of 1, the fullerene interaction might be disturbed, or smaller C₆₀ NPs might form. For C₆₀–1 NPs formulated with 0.2 mM of 1 ([C₆₀] = 0.1 mM, [1] = 0.2 mM), the absorption peak derived from the Soret band of 1 split into two peaks (Fig. 2c). The absorption peak at the shorter wavelength of 415 nm is consistent with that of the 1–TMe-β-CD complex. The absorption peak at the longer wavelength of 431 nm is almost consistent with the absorption peaks in the UV/Vis spectra of C₆₀–1 NPs fabricated with 0.05 and 0.1 mM of 1. The findings indicate that in the sample comprising 0.2 mM of 1, a portion of the 1–TMe-β-CD complexes remained in the complex state after heating for 1 h at 80 °C, in the presence of PEG. The absorption peak at 338 nm, which reflects the state of fullerene NPs, is similar to that of C₆₀–1 NPs fabricated with 0.1 mM of 1. When C₆₀ and 1 form co-aggregated NPs, the ratio of 1 to C₆₀ is thought to be limited to ∼1 : 1.

Morphological observations of the hybrid NPs were also undertaken. In the absence of the por–TMe-β-CD complex, C₆₀ NPs possessing fairly monodisperse size distributions were observed (Fig. S6a†). The average diameter of the individual NPs, determined from the transmission electron microscopy (TEM) images, is 82 nm. C₆₀ NPs have been previously reported to exhibit lattice fringes and diffraction patterns, which suggests that C₆₀ NPs maintain the face-centered cubic (fcc) crystalline structure.¹¹ C₆₀–1 NPs prepared with 0.05 mM C₆₀ and 0.1 mM 1, possessed irregular shapes (Fig. 3a and b, respectively). The average diameter of C₆₀–1 NPs, determined by TEM, is 119 nm (Fig. 3a and S6b†), demonstrating the larger C₆₀–1 NP size than that of the C₆₀ NPs (82 nm). Increasing the concentration of 1 to 0.1 mM results in the average diameter of C₆₀–1 NPs to increase to 131 nm (Fig. 3b and S6c†). Similar morphology is observed from the TEM micrographs of C₆₀–2 and C₆₀–3 NPs having average diameters of 109 nm and 144 nm, respectively (Fig. 3c and d, respectively). A lower PEG molecular weight or a lower reaction temperature during C₆₀ NP formation via C₆₀–γ-CD complexes have been reported to induce an increase in the diameter of the C₆₀ NPs.^11,12 Thus, the findings suggest that slower reaction conditions result in less nucleation and a larger NP formation. Porphyrin 3 possesses a methoxy substituent at the para position of the phenyl group and is more polar than 1 or 2. Previous reports have demonstrated that the higher the polarity of the phenyl group, the more stable the complex with TMe-β-CD,^21,22 which indicates that 3–TMe-β-CD is more stable than 1– or 2–TMe-β-CD in water. Thus, the aforementioned decomposition, which results from the interaction with PEG, is slower in 3 with a concomitant increase in the NP size.

Fig. 3 Transmission electron microscopy (TEM) images of C₆₀–1 nanoparticles (NPs) prepared with (a) 0.05 and (b) 0.1 mM of the 1–TMe-β-CD complex. TEM images of (c) C₆₀–2 and (d) C₆₀–3 NPs. Scale bars in images (a–d) are 100 nm. (e) High resolution TEM micrograph and selected-area electron diffraction pattern of C₆₀–1 NP. Scale bar is 10 nm. (f) ¹³C NMR spectra of (i) C₆₀ NPs and (ii) C₆₀–1 NPs. (g) Illustration of a C₆₀–por NP. A portion of C₆₀ form crystalline structures, while a portion of the porphyrin molecules interact with C₆₀ at the molecular level.

In the high-resolution TEM micrograph (Fig. 3e), the C₆₀–1 NPs only exhibited partial lattice fringes, and hence did not show clear diffraction patterns compared with the C₆₀ NPs (inset in Fig. 3e).¹¹ The findings demonstrate the highly amorphous nature of the C₆₀–1 NPs, and that the C₆₀ crystal structure was only retained in part. ¹³C NMR spectra also provide important information about the structure of the C₆₀–1 NPs. A characteristic C₆₀ cluster signal at 142.4 ppm was detected in both C₆₀ NPs and C₆₀–1 NPs (Fig. 3f).²⁶ The C₆₀–1 NP dispersions also exhibited several small new signals at 141.5, 143.3, and 143.7 ppm, as shown in Fig. 3f(ii). When a fullerene and a porphyrin molecule form a stable complex in solution, the C₆₀ signal shifts depending on the interaction type between the fullerene and the porphyrin molecule.²⁷ Thus, the porphyrin molecule interacted with the aggregate of C₆₀ in C₆₀–1, as illustrated in Fig. 3g.

Some porphyrin molecules can form a co-crystal with fullerene C₆₀.²⁸ To form a crystal structure, not only the interaction between molecules but also the relationship with the solvent, such as gradually changing the polarity of the solvent or removing the solvent, are important. In our system, porphyrin molecules that are pseudo-dissolved by TMe-β-CDs are added to water, which is a poor solvent for porphyrins, through the interaction of PEG with TMe-β-CDs. The porphyrin molecule should immediately aggregate and have difficulty forming a crystal structure. Furthermore, because water is also a poor solvent for fullerene C₆₀, C₆₀ also immediately aggregates in water. Thus, it should be extremely difficult for C₆₀ and porphyrin molecules to regularly associate to form a co-crystal structure.

The concentration of singlet oxygen molecules (¹O₂, Type-II energy transfer pathway) generated by photoirradiation was measured according to a chemical method using 9,10-anthracenediyl-bis(methylene) dimalonic acid (ABDA)^15,29 as a marker to determine the biological activities of C₆₀ NPs, C₆₀–1 NPs, C₆₀–2 NPs, and C₆₀–3 NPs. The absorption of ABDA at the absorption maximum (380 nm) was monitored as a function of irradiation time ([C₆₀] = 0.1 mM, [por] = 0 or 0.1 mM). Under visible-light irradiation at wavelengths > 620 nm, C₆₀–1 NPs, C₆₀–2 NPs, and C₆₀–3 NPs generated higher levels of ¹O₂ than C₆₀ (Fig. 4a). These results show that the ¹O₂ photoproduction abilities of the C₆₀–por NPs were higher than that of the C₆₀ NPs. There are no significant differences in the ¹O₂ photoproduction abilities of C₆₀–1 NPs, C₆₀–2 NPs, and C₆₀–3 NPs, which suggests that the structure of the porphyrin has an insignificant influence on the ability of the hybrid NPs. The generation of formazan, via the reduction of nitroblue tetrazolium (NBT) by oxygen radicals (O₂˙⁻), is observed as an increase of absorption intensity at 560 nm.³⁰ The reduction of NBT by O₂˙⁻ was scarcely detected in solutions containing C₆₀–1 NPs, C₆₀–2 NPs, and C₆₀–3 NPs under photoirradiation, even though formazan was readily detected in the positive control sample in the presence of reduced nicotinamide adenine dinucleotide (NADH) (Fig. 4b). The results suggest that the reactive oxygen species produced by C₆₀–1 NPs, C₆₀–2 NPs, and C₆₀–3 NPs are predominantly ¹O₂ generated by a Type-II reaction.¹⁸

Fig. 4 (a) ¹O₂ generation by NPs. Bleaching of 9,10-anthracenediyl-bis(methylene) dimalonic acid (ABDA) was monitored as a function of the decrease in the absorbance at 380 nm, for C₆₀ NPs (black circles and solid line), C₆₀–1 NPs (red circles and solid line), C₆₀–2 NPs (blue circles and solid line), and C₆₀–3 NPs (green circles and solid line) ([C₆₀] = 15 μM, [por] = 0 or 15 μM, [ABDA] = 25 μM). (b) O₂˙⁻ generation by NPs. The amount of formazan generated by the reduction of nitroblue tetrazolium (NBT) in the presence of O₂˙⁻ was analyzed by the absorbance at 560 nm, of C₆₀–1 NPs (red circles) and C₆₀–2 NPs (blue circles) in the absence (solid lines) and presence (dashed lines) of NADH ([C₆₀] = 15 μM, [por] = 15 μM, [NBT] = 200 μM, [NADH] = 0 or 625 μM). All samples were photoirradiated at >620 nm, in O₂-saturated aqueous solutions.

In summary, the preparation of hybrid C₆₀–porphyrin NPs was achieved via a guest exchange reaction comprising porphyrin CD complexes and C₆₀. Seven C₆₀–por NP derivatives with various moieties were prepared. CD porphyrin complexes possessing phenyl and methoxyphenyl moieties were decomposed in the presence of PEG at the same time as C₆₀–γ-CD complexes and formed NPs with C₆₀. Porphyrins containing a hydrophilic moiety form stable complexes with TMe-β-CD and fail to co-aggregate with C₆₀. The C₆₀–por NPs are negatively charged and are easily dispersed and stable in water. The ¹O₂ generation ability of C₆₀–por NPs under photoirradiation (>620 nm) is greater than that of C₆₀ NPs. The findings herein demonstrate a new method to fabricate fullerene–porphyrin composite materials, which provides a route to highly functional fullerene-based materials.

Conflicts of interest

The authors declare no conflicts of interests.

Acknowledgements

This work was supported by the Japan Society for the Promotion of Science, KAKENHI (Grant No. JP16H04133 and JP19K15523).

Notes and references

1. G. V. Andrievsky, V. K. Klochkov, A. B. Bordyuh and G. I. Dovbeshko, Chem. Phys. Lett., 2002, 364, 8–17.
2. K. J. Moor, S. D. Snow and J. H. Kim, Environ. Sci. Technol., 2015, 49, 5990–5998.
3. N. O. McHedlov-Petrossyan, Chem. Rev., 2013, 113, 5149–5193.
4. D. M. Guldi, Chem. Soc. Rev., 2002, 31, 22–36.
5. H. Imahori, K. Tamaki, D. M. Guldi, C. P. Luo, M. Fujitsuka, O. Ito, Y. Sakata and S. Fukuzumi, J. Am. Chem. Soc., 2001, 123, 2607–2617.
6. C. O. Obondi, G. N. Lim and F. D'Souza, J. Phys. Chem. C, 2015, 119, 176–185.
7. S. M. Rezayat, S. V. S. Boushehri, B. Salmanian, A. H. Omidvari, S. Tarighat, S. Esmaeili, S. Sarkar, N. Amirshahi, R. N. Alyautdin, M. A. Orlova, I. V. Trushkov, A. L. Buchachenko, K. C. Liu and D. A. Kuznetsov, Eur. J. Med. Chem., 2009, 44, 1554–1569.
8. M. Ayabe, A. Ikeda, Y. Kubo, M. Takeuchi and S. Shinkai, Angew. Chem., Int. Ed., 2002, 41, 2790–2792.
9. P. D. W. Boyd and C. A. Reed, Acc. Chem. Res., 2005, 38, 235–242.
10. F. D'Souza and O. Ito, Coord. Chem. Rev., 2005, 249, 1410–1422.
11. K. Sugikawa, K. Kozawa, M. Ueda and A. Ikeda, RSC Adv., 2016, 6, 74696–74699.
12. K. Sugikawa, K. Kozawa, M. Ueda and A. Ikeda, Chem.–Eur. J., 2017, 23, 13704–13710.
13. K. Sugikawa, Y. Inoue, K. Kozawa and A. Ikeda, ChemNanoMat, 2018, 4, 682–687.
14. K. Sugikawa, A. Kubo and A. Ikeda, Chem. Lett., 2016, 45, 60–62.
15. Y. Doi, A. Ikeda, M. Akiyama, M. Nagano, T. Shigematsu, T. Ogawa, T. Takeya and T. Nagasaki, Chem.–Eur. J., 2008, 14, 8892–8897.
16. A. Ikeda, S. Hino, T. Mae, Y. Tsuchiya, K. Sugikawa, M. Tsukamoto, K. Yasuhara, H. Shigeto, H. Funabashi, A. Kuroda and M. Akiyama, RSC Adv., 2015, 5, 105279–105287.
17. Y. Tsuchiya, T. Shiraki, T. Matsumoto, K. Sugikawa, K. Sada, A. Yamano and S. Shinkai, Chem.–Eur. J., 2012, 18, 456–465.
18. T. Nakaya, B. Horiguchi, S. Hino, K. Sugikawa, H. Funabashi, A. Kuroda and A. Ikeda, Photochem. Photobiol. Sci., 2019, 18, 459–466.
19. R. G. Alargova, S. Deguchi and K. Tsujii, J. Am. Chem. Soc., 2001, 123, 10460–10467.
20. J. Brant, H. Lecoanet, M. Hotze and M. Wiesner, Environ. Sci. Technol., 2005, 39, 6343–6351.
21. K. Kano, R. Nishiyabu, T. Asada and Y. Kuroda, J. Am. Chem. Soc., 2002, 124, 9937–9944.
22. T. CaroFiglio, R. Fornasier, V. Lucchini, C. Rosso and U. Tonellato, Tetrahedron Lett., 1996, 37, 8019–8022.
23. A. Harada, Coord. Chem. Rev., 1996, 148, 115–133.
24. Y. Ishibashi, M. Arinishi, T. Katayama, H. Miyasaka and T. Asahi, Chem. Lett., 2012, 41, 1104–1106.
25. X. Chang and P. J. Vikesland, Environ. Sci. Technol., 2011, 45, 9967–9974.
26. C. S. Yannoni, R. D. Johnson, G. Meijer, D. S. Bethune and J. R. Salem, J. Phys. Chem., 1991, 95, 9–10.
27. D. Y. Sun, F. S. Tham, C. A. Reed, L. Chaker and P. D. W. Boyd, J. Am. Chem. Soc., 2002, 124, 6604–6612.
28. P. D. W. Boyd, M. C. Hodgson, C. E. F. Rickard, A. G. Oliver, L. Chaker, P. J. Brothers, R. D. Bolskar, F. S. Tham and C. A. Reed, J. Am. Chem. Soc., 1999, 121, 10487–10495.
29. B. A. Lindig, M. A. J. Rodgers and A. P. Schaap, J. Am. Chem. Soc., 1980, 102, 5590–5593.
30. I. Nakanishi, S. Fukuzumi, T. Konishi, K. Ohkubo, M. Fujitsuka, O. Ito and N. Miyata, J. Phys. Chem. B, 2002, 106, 2372–2380.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details and Fig. S1–S6. See DOI: 10.1039/d0ra09387d

---