Читать книгу Electron Transfer - Shunichi Fukuzumi - Страница 12

Multistep electron‐transfer processes have been utilized to attain a long distance charge separation (CS), mimicking the natural photosynthetic reaction center (vide supra). However, a significant amount of energy is lost during the multistep electron‐transfer processes to reach the final CS state [11–15]. In photosynthesis a two–step photoexcitation, the so‐called “Z‐scheme,” is thereby required to recover the energy loss via the multistep electron‐transfer processes and to gain strong oxidizing power to oxidize water as well as high reducing power to reduce NAD⁺ coenzyme [4]. The design and synthesis of molecular machinery mimicking such an elaborated “Z‐scheme” in nature seems far beyond our capability, and even if it could be done, the synthetic cost would certainly preclude any type of practical application. Thus, it is highly desired to design simple molecular electron donor–acceptor dyads that are capable of fast CS but can retain slow charge recombination (CR). Theoretically, it is possible to obtain such an electron donor–acceptor dyad, because the CS lifetime increases with increasing driving force of electron transfer in the Marcus inverted region (vide supra). However, the driving force of electron transfer should be lower than the triplet excited state of one of the components of donor–acceptor dyads. Otherwise, the CS state would decay rapidly to the triplet excited state in the Marcus normal region rather than to the ground state in the Marcus inverted region [35].

A number of simple donor–acceptor dyads have been designed and synthesized to attain long‐lived CS state, where the donor and acceptor molecules are linked with a short spacer to minimize the solvent reorganization energy [45–50]. Efficient photoinduced electron transfer occurs in a zinc imidazoporphyrin–C₆₀ dyad (ZnImP–C₆₀) with a short linkage to form the CS state (ZnImP^·+–C₆₀^·−) with the rate constant of 1.4 × 10¹⁰ s⁻¹ (Scheme 4.1) [45]. The CS state (1.34 eV) is lower in energy than both the triplet excited states of C₆₀ (1.50 eV) and ZnImP (1.36 eV) [45]. The CS state, produced upon photoexcitation of ZnImP–C₆₀, is detected by the transient absorption spectrum, which has absorption bands at 700 and 1040 nm due to ZnImP^·+ and C₆₀^·−, respectively [45]. The CS state decays by back electron transfer to the ground state, obeying first‐order kinetics with a rate constant of 3.9 × 10³ s⁻¹ (the lifetime is 260 μs) at 298 K [45]. At 278 K the lifetime of the CS state was determined as 310 μs, which is much longer than those of conventional donor–acceptor dyads with longer spacers [7–9].

Scheme 4.1 Formation of a long‐lived CS state of a zinc imidazoporphyrin–C₆₀ dyad (ZnImP–C₆₀) with a short linkage (Ar = 3,5‐Bu^t₂C₆H₃).

Source: Kashiwagi et al. 2003 [45]. Reproduced with permission of American Chemical Society.

An electron donor–acceptor dyad linked with a short spacer containing Au(III) and Zn(II) porphyrins (ZnPQ–AuPQ⁺ in Scheme 4.2) also affords a long‐lived electron‐transfer state with a lifetime of 10 μs in nonpolar solvents such as cyclohexane [46]. The introduction of quinoxaline to the gold porphyrin results in a lowering of the electron‐transfer state energy. In contrast to the case of neutral donor–acceptor dyads, the energy of the electron‐transfer state (ZnPQ^·+–AuPQ) becomes smaller in a less polar solvent, which is lower than the energies of the triplet excited states of ZnPQ (1.32 eV) and AuPQ⁺ (1.64 eV) [46]. Photoinduced electron transfer occurs from the singlet excited state of the ZnPQ (¹ZnPQ^*) to the metal center of the AuPQ⁺ moiety to produce ZnPQ^·+–Au^IIPQ. The observed long lifetime of ZnPQ^·+–Au^IIPQ results from a small reorganization energy for the metal‐centered electron transfer of AuPQ⁺ in nonpolar solvents due to the small change in solvation upon electron transfer as compared with that in polar solvents [46]. In a polar solvent such as benzonitrile (PhCN), no CS state was observed, but instead only the triplet–triplet absorption due to ³ZnPQ^*–AuPQ⁺ was observed [46]. The absence of an observable CS state in PhCN is ascribed to the much slower photoinduced electron transfer due to the large reorganization energy as compared with that in nonpolar solvents allowing an efficient intersystem crossing process in the ZnPQ–AuPQ⁺ dyad to produce the triplet excited state ³ZnPQ^*–AuPQ⁺ [46].

Scheme 4.2 Formation of a long‐lived CS state of ZnPQ–AuPQ⁺ in nonpolar solvents (Ar = 3,5‐Bu^t₂C₆H₃).

Source: Fukuzumi et al. 2003 [46]. Reproduced with permission of American Chemical Society.

Figure 4.1 Structure of a closely linked ZnCh–C₆₀ dyad.

Source: Ohkubo et al. 2004 [47]. Reproduced with permission of John Wiley & Sons.

A closely linked zinc chlorin–fullerene dyad (ZnCh–C₆₀ in Figure 4.1) affords a longer CS lifetime as compared with other zinc chlorin–fullerene dyads with longer spacers [47–51]. A deoxygenated PhCN solution containing ZnCh–C₆₀ gives rise upon a 388 nm laser pulse to a transient absorption maximum at 460 nm due to the singlet excited state of ZnCh [47]. The decay rate constant was determined as 1.0 × 10¹¹ s^–1, which agrees with the value determined from the fluorescence lifetime measurements [47]. The decay of absorbance at 460 nm due to ¹ZnCh^* is accompanied by an increase in absorbance at 590 nm due to ZnCh^·+ [47]. This indicates that electron transfer from ¹ZnCh^* to C₆₀ occurs rapidly to form the CS state, ZnCh^·+–C₆₀^·–. The CS state decays via back electron transfer to the ground state rather than to the triplet excited state, because the CS state is lower in energy (1.26 eV) than the triplet excited states of both C₆₀ (1.50 eV) and ZnCh (1.36–1.45 eV) [47]. The lifetime of the CS state is determined as 230 μs at 298 K. The large temperature dependence of the CS lifetime is observed and the lifetime of the CS state at 123 K becomes as long as 120 seconds [47].

Covalently and non‐covalently linked porphyrin–quinone dyads constitute one of the most extensively investigated photosynthetic reaction center models, in which the fast photoinduced electron transfer from the porphyrin singlet excited state to the quinone occurs to produce the CS state, mimicking well the photosynthetic electron transfer [52–54]. Unfortunately, the CR rates of the CS state of porphyrin–quinone dyads are also fast and the CS lifetimes are mostly on the order of picoseconds or subnanoseconds in solution [52–54]. In general, a three‐dimensional C₆₀ is superior to a two‐dimensional quinone in terms of the smaller reorganization of electron transfer of C₆₀ as compared with quinone (vide supra) to attain the long‐lived CS state [31–33,55]. When the geometry between a porphyrin ring and quinone is optimized by using hydrogen bonds, which can also control the redox potentials of quinones, however, a surprisingly long lifetime up to one microsecond has been attained [56]. In a series of ZnP–n–Q (n = 3, 6, 10) in Scheme 4.3, the hydrogen bond between two amide groups provides a structural scaffold to assemble the donor (ZnP) and the acceptor (Q) moiety, leading to attaining the long‐lived CS state [56].

Scheme 4.3 Zinc porphyrin–quinone linked dyads (ZnP–n–Q; n = 3, 6, 10) with hydrogen bonds.

Source: Okamoto and Fukuzumi 2005 [56]. Reproduced with permission of American Chemical Society.

As described above, the closely linked donor–acceptor dyads afford long‐lived CS states. As long as porphyrins and C₆₀ are used as components of donor–acceptor dyads, however, the low lying triplet energies of porphyrins and C₆₀ have precluded to attain the long‐lived CS states with a higher energy than the triplet energies [35]. In such a case, it is highly desired to find a chromophore that has a high triplet energy and a small λ value of electron transfer. Among many choromophores, acridinium ion is the best candidate for such a purpose, since the λ value for the electron self‐exchange between the acridinium ion and the corresponding one‐electron reduced radical (acridinyl radical) is the smallest (0.3 eV) among the redox‐active organic compounds [57]. Another important property of acridinium ion is a high triplet excited energy [58,59]. Thus, an electron donor moiety (mesityl group) is directly connected at the 9‐position of the acridinium ion to yield 9‐mesityl‐10‐methylacridinium ion (Acr⁺–Mes) [60], in which the solvent reorganization of electron transfer is minimized because of the short linkage between the donor and acceptor moieties. The X‐ray crystal structure of Acr⁺–Mes is shown in Figure 4.2a [60]. The dihedral angle made by aromatic ring planes is perpendicular and therefore there is no π conjugation between the donor and acceptor moieties. Indeed, the absorption and fluorescence spectra of Acr⁺–Mes are superpositions of the spectra of each component, i.e. mesitylene and 10‐methylacridinium ion. The highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) orbitals of Acr⁺–Mes calculated by a density functional theory (DFT) method with Gaussian 98 (B3LYP/6‐31G* basis set) are localized on mesitylene and acridinium moieties (Figure 4.2b,c), respectively [60]. The energy of the electron‐transfer state (Acr^·–Mes^·+) in PhCN is determined by the redox potentials of each component of Acr⁺–Mes as 2.37 eV [60].

Figure 4.2 (a) X‐ray crystal structure of Acr⁺–Mes. (b) HOMO and (c) LUMO orbitals calculated by DFT method with Gaussian 98 (B3LYP/6‐31G* basis set). (d) Plot of k_BET/T vs. T⁻¹.

Source: Fukuzumi et al. 2004 [60]. Reproduced with permission of American Chemical Society.

Photoirradiation of a deaerated PhCN solution of Acr⁺–Mes by a nanosecond laser light at 430 nm results in the formation of Acr^·–Mes^·+ with a quantum yield close to unity (98%) via photoinduced electron transfer from the mesitylene moiety to the singlet excited state of the acridinium ion moiety (¹Acr^+*–Mes) [60]. The intramolecular back electron transfer from the Acr^· moiety to the Mes^·+ moiety in Acr^·–Mes^·+ was too slow to compete with the intermolecular transfer (k_BET) in Figure 4.2d, agreeing with the Marcus equation in the deeply inverted region (Eq. (2.1)). The lifetime of the electron‐transfer state in frozen medium becomes longer with decreasing temperature to approach a virtually infinite value at 77 K [60]. However, the decay time profile of Acr^·–Mes^·+ in solution obeyed second‐order kinetics (NOT first‐order kinetics) [60]. This is the same as the case of Fc⁺–ZnP–H₂P–C₆₀^·−, in which bimolecular back electron transfer predominates due to the slow intramolecular back electron transfer (vide supra) [39]. In contrast, the decay of Acr^·–Mes^·+ obeys first‐order kinetics in PhCN at high temperatures [60]. This indicates that the rate of the intramolecular back electron transfer of Acr^·–Mes^·+ becomes much faster than the rate of the intermolecular back electron transfer at higher temperatures because of the larger activation energy of the former than that of the latter. Such a remarkable result has sparked a flurry of work by others in the field of artificial photosynthesis [61].

Benniston et al. claimed that the triplet excited state of the acridinium ion moiety (³Acr^+*–Mes) might be formed rather than the electron‐transfer state (Acr^·–Mes^·+) and that the energy of ³Acr^+*–Mes is lower than that of Acr^·–Mes^·+ [62]. They reported that the triplet excitation energy of Acr⁺–Mes was 1.96 eV based on the phosphorescence spectrum [62]. If this value were correct, the one‐electron oxidation potential (E_ox) of ³Acr^+*–Mes would be −0.08 V vs. saturated calomel electrode (SCE), which is determined from the one‐electron oxidation potential of the Mes moiety (1.88 V) [60] and the triplet excitation energy (1.96 V). In such a case, electron transfer from the triplet excited state of Acr⁺–Mes to N,N‐dihexylnaphthalenediimide (NIm: E_red = –0.46 V vs. SCE) would be energetically impossible judging from the positive free energy change of electron transfer (0.38 eV). However, the addition of NIm (1.0 × 10⁻³ M) to a PhCN solution of Acr⁺–Mes and laser photoexcitation results in the formation of NIm^·− as detected clearly by the well‐known absorption bands at 480 and 720 nm [63,64], accompanied by the decay of transient absorption at 510 nm due to the Acr^· moiety of the electron‐transfer (ET) state as shown in Figure 4.3a [65]. Similarly, the addition of aniline (3.0 × 10⁻⁵ M) to a PhCN solution of Acr⁺–Mes results in the formation of aniline radical cation (λ_max = 430 nm) [66], accompanied by decay of the Mes^·+ moiety at 500 nm as shown in Figure 4.3b [65]. The rate constant of formation of aniline radical cation was determined to be 5.6 × 10⁹ M⁻¹ s⁻¹, which is close to the diffusion rate constant in PhCN [60]. Thus, the photogenerated state of Acr⁺–Mes has both the reducing and oxidizing abilities to reduce NIm and to oxidize aniline, respectively. Only the electron‐transfer state (Acr^·–Mes^·+) has such a dual ability, which has now been well confirmed for electron‐transfer oxidation of many electron donors with Acr^·–Mes^·+ and electron‐transfer reduction of many electron acceptors such as hexyl viologen, p‐benzoquinone, and Selectfluor (fluorinating reagent) with Acr^·–Mes^·+ [67–70]. However, this conclusion is contradictory to the reported triplet energy (1.96 eV), which is lower in energy than the ET state [62]. This contradiction comes from an acridine impurity, which may be left in the preparation of by Benniston et al. who synthesized the compound via methylation of the corresponding acridine [62]. The yield of acridinium ion is about 50–70% after reflux at high temperature for a few days [62]. In such a case, acridine may remain as an impurity even after purification of the acridinium ion by recrystallization. When Acr⁺–Mes was prepared by the Grignard reaction of 10‐methyl‐9(10H)acridone with 2‐mesitylmagnesium bromide, there was no acridine [60]. Thus, Acr⁺–Mes without acridine afforded no phosphorescence spectrum in both deaerated glassy 2‐MeTHF and ethanol at 77 K. It is well known that acridine derivatives exhibit phosphorescence at 15650–15850 cm⁻¹ [71]. It was confirmed that the phosphorescence maximum of 9‐phenylacridine in glassy 2‐MeTHF at 77 K afforded the same spectrum reported by Benniston et al. [62] Thus, the reported low triplet energy of Acr⁺–Mes, which contradicts our results on the long‐lived electron‐transfer state, results from the acridine impurity contained in Acr⁺–Mes used by Benniston et al. who also reported that photoirradiation of a PhCN solution of Acr⁺–Mes results in the formation of the acridinyl radical (Acr^·–Mes) [62]. They implied that this stable radical species could be mistaken as a long‐lived electron‐transfer state [62]. When PhCN is purified, however, no change in the absorption spectrum is observed [60,65]. The formation of Acr^·–Mes results from electron transfer from a donor impurity contained in unpurified PhCN (e.g. aniline) to the Mes^·+ moiety of Acr^·–Mes^·+ as indicated in Figure 4.3b. Even an extremely small amount (5.0 × 10⁻⁵ M) of aniline is enough to react with Acr^·–Mes^·+ to produce Acr^·–Mes, which is stable due to the bulky Mes substituent, because the lifetime of Acr^·–Mes^·+ is long enough to react with such a small concentration of an electron donor. It should be noted that no net photochemical reaction occurs without a donor impurity because the long‐lived Acr^·–Mes^·+ decays via bimolecular back electron transfer to the ground state [60,65]. Thus, misleading effects of impurities indeed result from the long‐lived electron‐transfer state, which has both oxidizing and reducing abilities.

Figure 4.3 Transient absorption spectra of Acr⁺–Mes (5.0 × 10⁻⁵ M) in deaerated MeCN at 298 K taken at 2 and 20 μs after laser excitation at 430 nm in the presence of (a) N,N‐dihexylnaphthalenediimide (1.0 × 10⁻³ M) or (b) aniline (3.0 × 10⁻⁵ M). Inset: Time profiles of the absorbance decay at 510 nm and the rise at 720 nm and (b) the decay at 500 nm and the rise at 430 nm.

Source: Fukuzumi and coworkers 2005 [65]. Reproduced with permission of Royal Society of Chemistry.

In contrast to the photoirradiation of a purified PhCN solution of Acr⁺–Mes at 298 K, which results in no change in the absorption spectrum (Figure 4.4a), when the photoirradiation of the same solution was performed at low temperatures (213–243 K) with a 1000 W high‐pressure mercury lamp through the UV light cutting filter (>390 nm) and the sample was cooled to 77 K, the color of the frozen sample at 77 K was clearly changed as shown in the inset of Figure 4.4b. When a glassy 2‐methyltetrahydrofuran (2‐MeTHF) is employed for the photoirradiation of Acr⁺–Mes at low temperature, the resulting glassy solution measured at 77 K affords the absorption spectrum due to the electron‐transfer state, which consists of the absorption bands of the Acr^· moiety (500 nm) and the Mes^·+ moiety (470 nm) as shown in Figure 4.4b. No decay of the absorption due to the electron‐transfer state in Figure 4.2b was observed until liquid nitrogen ran out [65].

The long lifetime of the ET state of Acr⁺–Mes has allowed observing the structural change in the Acr⁺–Mes(ClO₄⁻) crystal upon photoinduced ET directly by using laser pump and X‐ray probe crystallographic analysis (Figure 4.5) [72]. Upon photoexcitation of the crystal of Acr⁺–Mes(ClO₄⁻), the N‐methyl group of the Acr⁺ moiety was bent and its bending angle was 10.3(16)° when the N‐methyl carbon moved 0.27(4) Å away from the mean plane of the ring as shown in Figure 4.5 [72]. This bending is caused by the photoinduced electron transfer from the Mes moiety to the Acr⁺ moiety to produce Acr^·–Mes^·+, because the sp² carbon of the N‐methyl group of Acr⁺ is changed to the sp³ carbon in the one‐electron reduced state (Acr^·) [72]. The bending of the N‐methyl group by photoexcitation was accompanied by the rotation and movement of the ClO₄⁻ by the electrostatic interaction with the Mes^·+ moiety (Figure 4.5) [72]. Thus, the observed bending of the N‐methyl group and the movement of ClO₄⁻ provide strong evidence for the generation of the ET state of Acr⁺–Mes upon photoexcitation. In contrast to the case of Acr⁺–Mes, no geometrical difference was observed upon photoexcitation of Acr⁺–Ph, which does not afford the ET state [72].

Figure 4.4 (a) UV–vis spectral change in the steady‐state photolysis of a deaerated PhCN solution of Acr⁺–Mes (3.3 × 10^–5 M). Spectra were recorded at 90‐second interval. (b) UV–vis absorption spectra obtained by photoirradiation with high‐pressure mercury lamp of deaerated 2‐MeTHF glasses of Acr⁺–Mes at 77 K. Inset: picture images of frozen PhCN solutions of Acr⁺–Mes before and after photoirradiation at low temperatures and taken at 77 K.

Figure 4.5 (a) Diagram of the reaction cavity: (left) diagram around the N‐methyl group, with numbers indicating the volumes of the divided cavity formed by the dotted line; (right) drawing around ClO₄⁻. (b) Cooperative photoinduced geometrical changes. The dashed line indicates the suggested Mes^·+⋯ClO₄⁻ electrostatic interaction.

Source: Hoshino et al. 2012 [72]. Reproduced with permission of American Chemical Society.

Immobilization of Acr⁺–Mes has also been achieved by incorporating Acr⁺–Mes cation into nanosized mesoporous silica–alumina (AlMCM‐41), which has cation exchange sites to obtain a nanocomposite (Acr⁺–Mes@AlMCM‐41) [73]. The shape and size of nanosized AlMCM‐41 were controlled by changing the preparation conditions as shown in Figure 4.6, where TEM images reveal a tubular or rod‐like (tAlMCM‐41) morphology in the diameter of 50–100 nm with the length of 0.2–2 μm array (part a) and also a sphere morphology (sAlMCM‐41, part b) [73]. The X‐ray powder pattern of tAlMCM‐41 exhibited a well‐resolved pattern with a prominent peak (100) observed at c. 2θ = 2.56°, indicating a highly ordered material with a hexagonal array [73]. Uniform channels c. 4 nm in diameter exist in a tube. Because the Acr⁺–Mes molecular size is small enough as compared with the pore size of mesoporous silica with its diameter of more than 3 nm, cation exchange with Acr⁺–Mes occurs spontaneously upon mixing Na⁺–exchanged AlMCM‐41 with Acr⁺–Mes in acetonitrile [73]. The cation exchange percentages of tAlMCM‐41 and sAlMCM‐41 by Acr⁺–Mes were determined to be 16% and 18%, respectively [73]. The Acr⁺–Mes incorporated into AlMCM‐41 is stable without leaching out in acetonitrile at room temperature [73].

Upon photoexcitation of Acr⁺–Mes@tAlMCM‐41 suspended in MeCN, photoinduced ET from the Mes moiety to the singlet excited state of the Acr⁺ moiety occurred within 10 ps to produce the ET state as detected by laser flash photolysis and electron paramagnetic resonance (EPR) measurements [60,67]. In contrast to the case in solution (vide supra), no bimolecular decay of the ET state occurs because each Acr⁺–Mes molecule is isolated inside AlMCM‐41 [73]. The lifetime of the ET state of Acr⁺–Mes@tAlMCM‐41 suspended in acetonitrile was determined to be 2.3 seconds at 198 K, which is much longer than that in solution because of the inhibition of bimolecular BET in AlMCM‐41 as illustrated in Figure 4.6 [73]. Thus, incorporation of a simple electron donor–acceptor dyad into AlMCM‐41 has made it possible to elongate the lifetime of the charge‐separated state, which is longer than that of the bacterial photosynthetic reaction center (one second) [74].

Figure 4.6 Transmission electron microscope (TEM) images of (a) tAlMCM‐41 and (b) sAlMCM‐41 (the high‐resolution image of tAlMCM‐41 is inserted in (a)). (c) Reaction scheme of photocatalytic oxygenation of p‐xylene with Acr⁺–Mes and [(tmpa)Cu^II]²⁺ incorporated into sAlMCM‐41.

Source: Fukuzumi et al. 2012 [73]. Reproduced with permission of PNAS.

The triplet ET state of Acr^·–Mes^·+@tAlMCM‐41 was detected by an EPR spectrum measured at 4 K, which exhibited a fine structure together with a strong sharp signal at g = 4.0 [73]. The distance between two electron spins was determined from the zero‐field splitting parameters to be 7.7 Å, which agrees with the expected distance of 7.2 Å between an sp² carbon atom at the 4 position of the mesityl moiety and sp² carbon atoms at the 3 and 6 positions of the acridinyl moiety [73]. Polycondensation of Acr⁺–Mes‐bridged organosilane in the presence of a nonionic surfactant is also reported to yield a mesostructured organosilica solid with a functional framework that exhibited long‐lived photoinduced CS [75].

Nano‐sized charge‐separated molecules can also be obtained by using single‐walled carbon nanotubes (SWNTs) [76], which exhibit excellent chemical and physical properties as revealed by various potential applications [77–81]. Extensive efforts have so far been devoted to assemble electron donor and acceptor molecules on SWNTs [82–88]. However, the fine control of size (i.e. length) of SWNTs remains a formidable challenge, because SWNTs have seamless cylindrical structures made up of a hexagonal carbon network, which leads to the difficulty of solubilization/functionalization without treatment with strong acid or vigorous sonication [89–92]. On the other hand, the cup‐stacked carbon nanotubes (CSCNTs) that consist of cup‐shaped nanocarbon (CNC) units, which stack via van der Waals attractions, have merited special attention from the viewpoint of the conventional carbon nanotube alternatives [93–96]. The tube–tube van der Waals energy between CNCs has been counterbalanced by the thermal or photoinduced electron transfer multi‐electron reduction due to electrostatic repulsion, resulting in the highly dispersible CNCs with size homogeneity [97,98].

The CNCs with controlled size have been functionalized with a large number of porphyrin molecules [99]. The general procedure for the synthesis of the porphyrin‐functionalized cup‐shaped nanocarbons [CNC–(H₂P)_n] is shown in Figure 4.7a [99]. The CNCs are first functionalized with aniline as the precursor for further functionalization with porphyrins. The aniline‐functionalized nanocarbons react with the porphyrin derivatives to construct the nanohybrids.

The structure of the CNCs of the CNC–(H₂P)_n nanohybrids is shown by the TEM in Figure 4.7b, which reveals a CNC with a hollow core along the length of the nanocup with well‐controlled diameter (c. 50 nm) and size (c. 100 nm) [99]. The weight percentage of porphyrins attached to the CNCs was determined by thermogravimetric analysis (TGA) and elemental analysis to be ca. 20% [99]. This corresponds to one functional group per 640 carbon atoms of the nanocup framework for CNC–(H₂P)_n nanohybrid. Thus, the π‐framework of the CNC is not destroyed despite attachment of a large number of porphyrin molecules on the CNC.

Spectroscopic evidence for the covalent functionalization of CNC–(H₂P)_n nanohybrid was obtained by an intensity increase of the Raman signal at 1353 cm⁻¹ (D band) in the functionalized CNC as compared with the pristine CSCNTs [99], because the D band has been used for monitoring the process of functionalization that transforms sp² to sp³ sites [99]. The UV–vis absorption spectrum of CNC–(H₂P)_n nanohybrid agreed with that of the superposition of ref‐H₂P [tetrakis(N‐octadecyl‐4‐aminocarboxyphenyl)porphyrin] and CNCs, indicating that there is no significant interaction between attached porphyrins and CSCNTs in the ground states [99].

The fluorescence lifetime of CNC–(H₂P)_n was determined to be 3.0 ± 0.1 ns, which is much shorter than that of ref‐H₂P (14.1 ± 0.1 ns) [99]. The fluorescence emission at 650 nm was also quenched in CNC–(H₂P)_n [99]. The short fluorescence lifetime of CNC–(H₂P)_n and an efficient fluorescence quenching of porphyrins in CNC–(H₂P)_n as compared to the ref‐H₂P may result from the photoinduced electron transfer from the singlet excited state of H₂P (¹H₂P^*) to CNC in CNC–(H₂P)_n. The occurrence of photoinduced electron transfer to afford the charge‐separated (CS) state of CNC–(H₂P)_n was confirmed by nanosecond laser flash photolysis measurements in Figure 4.8, where the absorption bands in the visible and near infrared (NIR) regions are attributed to H₂P^·+, which are clearly different from the triplet–triplet absorption of ref‐H₂P [99]. The formation of the CS state was also confirmed by EPR measurements under photoirradiation of CNC–(H₂P)_n in frozen N,N‐diemthylformamide (DMF) at 153 K. The observed isotropic EPR signal at g = 2.0044 agrees with that of ref‐H₂P^·+ produced by one‐electron oxidation with [Ru(bpy)₃]³⁺ (bpy = 2,2′‐bipyridine) in deaerated CHCl₃ [99]. The EPR signal corresponding to the reduced carbon‐based nanomaterials was too broad to be detected, probably due to delocalization of electrons in CNC [99].

Figure 4.7 (a) Synthetic procedure of CNC–(H₂P)_n. (b) TEM image of CNC–(H₂P)_n.

Source: Ohtani et al. 2009 [99]. Reproduced with permission of John Wiley & Sons.

The CS state of CNC–(H₂P)_n detected in Figure 4.8a decays obeying clean first‐order kinetics: the first‐order plots at different initial CS concentrations afford linear correlations with the same slope (Figure 4.8b) [99]. Thus, the decay of the CS state results from back electron transfer in the nanohybrid rather than intermolecular back electron transfer from CNC^·− to H₂P^·+. The CS lifetime was determined from the first‐order plots in Figure 4.8b to be 0.64 ± 0.01 ms, which is the longest lifetime ever reported for electron donor‐attached nanocarbon materials [99]. Such a long CS lifetime may be ascribed to the efficient electron migration in the CNCs following CS.

Figure 4.8 (a) Transient absorption spectra of (a) CNC–(H₂P)_n taken at 20 and 1.8 ms after laser excitation at 426 nm and (b) ref‐H₂P in deaerated DMF at 298 K taken at 100 ms and 1.6 ms after laser excitation at 426 nm. (c) Decay time profiles and (d) first‐order plots at 470 nm with different laser powers (5, 3, 2, and 1 mJ/pulse).

Source: Ohtani et al. 2009 [99]. Reproduced with permission of John Wiley & Sons.