Aggregation-induced photocatalytic activity and efficient photocatalytic hydrogen evolution of amphiphilic rhodamines in water

The development of photocatalysts is an essential task for clean energy generation and establishing a sustainable society. This paper describes the aggregation-induced photocatalytic activity (AI-PCA) of amphiphilic rhodamines and photocatalytic functions of the supramolecular assemblies. The supramolecular assemblies consisting of amphiphilic rhodamines with octadecyl alkyl chains exhibited significant photocatalytic activity under visible light irradiation in water, while the corresponding monomeric rhodamines did not exhibit photocatalytic activity. The studies on the photocatalytic mechanism by spectroscopic and microscopic analyses clearly demonstrated the AI-PCA of the rhodamines. Moreover, the supramolecular assemblies of the rhodamines exhibited excellent photocatalytic hydrogen evolution rates (up to 5.9 mmol g−1 h−1).


Introduction
Photocatalysts are promising materials for the conversion of solar energy into storable chemical energy and are expected to contribute signicantly to clean and renewable energy generation. 1 In 1974, Fujishima and Honda reported photocatalytic water-splitting using a titanium dioxide electrode, demonstrating the possibility of articial photosynthesis. 2 Since then, a wide range of photocatalysts, based on inorganic, 3 molecular, 4 and polymeric 5 compounds, have been actively developed. Besides their application in articial photosynthesis, the redox reactivity of photocatalysts has been utilized for environmental remediation, 6 organic synthesis, 7 and photodynamic therapy. 8 The emergence and development of new photocatalysts have contributed to the progress in articial photosynthesis and generated new opportunities in the related elds. 9 Based on these backgrounds, we explored a new class of photocatalysts and focused on supramolecular assemblies. The photophysical properties of supramolecular assemblies are different from those of the constituting monomers because of the interaction between the adjacent molecules. 10 Various characteristic aggregation-induced photophysical phenomena, such as aggregation-caused quenching (ACQ), 11 aggregationinduced enhanced emission, 12 light-harvesting, 13 and nonlinear optical phenomena (e.g., photon upconversion 14 and singlet ssion 15 ) have been intensively studied and applied to solar energy collection, 16 molecular sensing, 17 and biological applications (e.g., bioimaging, 18 optogenetics, 19 and phototherapy 20 ). However, aggregation-induced photocatalytic activity (AI-PCA) has never been reported despite the high potential for a novel photocatalytic material. Taking into account previous reports on aggregation-induced triplet excited state generation 21 and charge carrier migration 22 in self-assembled nanostructures of organic dyes, we considered that various organic dyes may cause AI-PCA. These phenomena cause elongation of the excited state lifetime 23 and increasing collision frequency with substrates, 24 which are important for the progression of photocatalytic reactions. AI-PCA would lead to expansion of the molecular design of photocatalysts that enables adjustment of absorption wavelength and redox potential. In addition, selfassembled supramolecular photocatalysts (SA-SPCs) possessing AI-PCA are expected to produce unprecedented photocatalytic so-materials 25 (gel, liquid crystal, membrane etc.) taking advantages of the unique properties of supramolecular assemblies 26 (e.g. reversibility and stimuli-responsiveness).
Herein, we demonstrate the AI-PCA of amphiphilic rhodamines (Fig. 1a). Rhodamines are very common hydrophilic organic dyes with excellent photophysical properties, such as high light absorption and quantum yield, which can be tuned through chemical modication. 27 The two SA-SPCs composed of amphiphilic rhodamines (rhodamine B (RhB) and rhodamine 19 (Rh19) (Fig. 1b)) exhibited photocatalytic activity under visible light irradiation in water, while the monomeric rhodamines did not exhibit photocatalytic activity. In particular, the SA-SPCs exhibited excellent hydrogen evolution rates.

Molecular design of the rhodamine derivatives
Four rhodamine derivatives with short and long alkyl chains were used in this study (Fig. 1b). Two common rhodamines with different absorption bands, RhB and Rh19, were selected as hydrophilic organic dyes to examine the concept of AI-PCA. The amphiphilic rhodamines with octadecyl alkyl chains (RhB-C18 and Rh19-C18) were expected to form supramolecular assemblies in water through hydrophobic interaction between the alkyl chains. More hydrophilic rhodamine derivatives with an ethyl ester group (RhB-C2 and Rh19-C2) compared to those with octadecyl alkyl chains were prepared as control compounds to evaluate the effect of self-assembly on the photocatalytic activity.
The photophysical properties of monomeric RhB-C2 and RhB-C18 were evaluated from their UV-vis absorption (UV-vis) and photoluminescence (PL) spectra measured in dimethyl sulfoxide (DMSO), which is a good solvent for these compounds ( Fig. 2a and b). The absorption spectra of RhB-C2 and RhB-C18 corresponded well with each other (Fig. 2a) and both the compounds exhibited absorption maxima at 566 nm. RhB-C2 and RhB-C18 also exhibited similar PL spectra with an emission peak at l em ¼ 592 nm (Fig. 2b). These results indicate that the electronic states of RhB-C2 and RhB-C18 are quite similar despite the different alkyl chain lengths. We thus conclude that this pair is suitable for evaluating the effect of self-assembly on their photophysical and photocatalytic properties. Further, Rh19-C2 and Rh19-C18 also exhibited similar UV-vis and PL spectra ( Fig. S1a and b †) with maxima at l abs ¼ 539 nm and l em ¼ 565 nm, respectively, which indicate that the Rh19-C2/Rh19-C18 pair has similar electronic states regardless of the alkyl chain length.

Self-assembling properties of the rhodamines in water
To examine the self-assembly properties, the UV-vis and PL spectra of RhB-C2 and RhB-C18 were recorded in water (concentration: 5.0 mM) ( Fig. 2c and d). The UV-vis spectrum of RhB-C2 in water is similar to that in DMSO with a slightly red-shied absorption maximum (l abs : 559 nm (DMSO), 562 nm (water)) ( Fig. 2c). In contrast, RhB-C18 exhibited a broad spectrum with split peaks at 530 and 559 nm that can be assigned to the aggregation states of RhB-C18, 28 which suggests the formation of a supramolecular assembly of RhB-C18 in water. The PL spectra of RhB-C2 and RhB-C18 were signicantly different (Fig. 2d). RhB-C2 exhibited an intense emission in water, whereas RhB-C18 exhibited a very weak emission. This suggests ACQ in RhB-C18. 11 Furthermore, the addition of a nonionic surfactant (Triton X-100, 0.3 vol%) to the RhB-C18 aqueous solution drastically increased the PL intensity ( Fig. S2b †), which suggests the dissociation of the RhB-C18 supramolecular assembly by Triton X-100. The photophysical properties of the Rh19-C2/Rh19-C18 pair exhibited the same trend as those of the RhB-C2/RhB-C18 pair (Fig. S1c, d and S2c, d †). Rh19-C18 formed supramolecular assemblies, whereas Rh19-C2 did not form supramolecular assemblies in water. With the increasing concentration, the absorbances of the rhodamine derivatives measured in water increased linearly ( Fig. S3 and S4 †); this indicates that RhB-C2 and Rh19-C2 did  not form supramolecular assemblies until 50 mM (Fig. S3 †), and the excessive aggregation of RhB-C18 and Rh19-C18 did not occur at least until 100 mM (Fig. S4 †).
Dynamic light scattering (DLS) measurements and transmission electron microscopy (TEM) observations were performed to conrm the formation of supramolecular assemblies. DLS measurements indicated the presence of supramolecular assemblies of RhB-C18 and Rh19-C18 having average sizes of 200 and 82 nm, respectively ( Fig. S5a and b †). TEM revealed the formation of spherical supramolecular assemblies of RhB-C18 and Rh19-C18 (Fig. S5c and d †). The selected area electron diffraction (SAED) patterns obtained by TEM exhibited diffused rings and no diffraction spots, which indicate the amorphous nature of the rhodamine supramolecular assemblies (Fig. S5e  and f †).

Mechanistic study on the photocatalytic activity of the rhodamine SA-SPCs
The photoreaction of SA-SPCs is considered to occur via two types of mechanisms involving electron and/or energy transfer processes. 30 Under aerobic conditions, in the energy transfer process, singlet oxygen ( 1 O 2 ) is commonly involved in the photocatalytic reaction, whereas oxygen radicals such as superoxide anion radicals (O 2 c À ) and hydroxyl radicals (OHc À ) are involved in the electron transfer mechanism (Fig. S9 †). 31 To explain the mechanism of the photocatalytic reaction, electron spin resonance (ESR) spectroscopy was performed using 4-hydroxy-2,2,6,6-tetramethylpiperidine (4-OH-TEMP) 32 and 5,5-dimethyl-1-pyrroline N-oxide (DMPO) 33 as spin trap reagents to detect 1 O 2 and oxygen radicals (e.g. O 2 c À , OHc À ), respectively ( Fig. S10a and b †). The ESR experimental conditions were rst determined using Rose Bengal (Fig. S10c †) as a standard. 34 The ESR spectra of RhB-C2 and RhB-C18 exhibited a characteristic 1 : 1 : 1 triplet corresponding to the TEMPOL radical (Fig. 4a). RhB generates a triplet state despite the low quantum yield of intersystem crossing (quantum yield (F T ): 0.006). 35 Therefore, it is reasonable that RhB-C2 exhibited an ESR signal for the TEMPOL radical. No signicant differences were observed in the signal intensities of the TEMPOL radicals of RhB-C2 and RhB-C18, which implies that the rate of energy transfer to oxygen did not drastically change aer self-assembly. These results suggest that the photocatalytic reaction of RhB-C18 does not occur through the energy transfer mechanism. Subsequently, we evaluated the generation of oxygen radical species via the electron transfer mechanism. The ESR spectrum for RhB-C2 in DMPO did not exhibit a clear signal for a DMPO adduct (Fig. 4b), whereas the ESR spectrum of RhB-C18 exhibited a signal for a DMPO hydroxyl radical adduct (DMPO-OH). Since the superoxide anion (O 2 c À ) is unstable in aqueous media, it reacts with protons immediately upon addition to an aqueous medium (Fig. S9a †) and does not react with DMPO to form DMPO-OOH. Hence, no DMPO-OOH peak appeared in the ESR spectrum. However, we conrmed the generation of a hydroxyl radical (OHc À ) that was produced by the chain reaction starting from O 2 c À ions through electron transfer in water (Fig. 4b, c and This journal is © The Royal Society of Chemistry 2020 Chem. Sci., 2020, 11, 11843-11848 | 11845 S9a †). The experimental results showed that RhB-C18 exhibits photocatalytic activity mainly through an electron transfer mechanism. Further, the ESR spectra of the Rh19-C2/Rh19-C18 pair were similar to those of the RhB-C2/RhB-C18 pair (Fig. S11 †).
The time-courses of photocatalytic hydrogen evolution by RhB-C2 and RhB-C18 are shown in Fig. 5b. Aer light irradiation (l > 360 nm, 300 W (Xe lamp)), Pt nanoparticles were formed ( Fig. S14c-f †) and RhB-C18 exhibited hydrogen generation, while RhB-C2 did not display hydrogen evolution. The RhB-C18 SA-SPC functioned for 80 min without any decrease in the photocatalytic activity, and the average hydrogen evolution rate (HER) was determined to be 3.7 mmol g À1 h À1 (Fig. 5b), which is comparable to that of other excellent organic systems such as g-C 3 N 4 (0.67 mmol g À1 h À1 ) 37 and a covalent organic framework (10.1 mmol g À1 h À1 ). 38 80 min aer light irradiation, the photocatalytic activities of the SA-SPCs decreased due to the decomposition of rhodamines. Rh19-C18 exhibited photocatalytic hydrogen evolution (HER: 2.9 mmol g À1 h À1 ), while Rh19-C2 did not (Fig. S15a †). One of the reasons for the high HER would be intermolecular electron migration among the rhodamines. 39 The photoirradiation of rhodamines generated intermolecular charge separation states, and the migration between the rhodamines may have facilitated efficient electron transfer to the Pt nanoparticle. The apparent quantum efficiencies of RhB-C18 and Rh19-C18 were 0.059 and 0.039% under these conditions, respectively.
To examine the effects of the SA-SPC concentration on the hydrogen evolution reaction, the SA-SPC concentrations were increased from 50 to 100 mM. The amorphous self-assembled spherical structures were almost unchanged (Fig. S16, † average particle diameter: RhB-C18: 223 nm, Rh19-C18: 122 nm) aer the increase. The hydrogen evolution rates decreased from 3.7 to 2.4 mmol g À1 h À1 for RhB-C18 (Fig. S17a †). On the other hand, in the case of Rh19-C18, the hydrogen evolution rate signicantly increased from 2.9 to 5.9 mmol g À1 h À1 (Fig. S17b †). These results indicate that the hydrogen evolution rates are signicantly affected by the concentration of the SA-SPC. The versatile factors including the size, morphology, surface area, uidity, and electric state of the SA-SPC, and interactions between the SA-SPC and Pt nanoparticles or ascorbic acid would have a sensitive effect on the hydrogen evolution. A detailed understanding of the changes of RhB-C18 and Rh19-C18 for photocatalytic hydrogen evolution rates is currently difficult. We will study how each factor has effects on photocatalytic activity in future studies.

Conclusions
In summary, we demonstrated the AI-PCA of two rhodamine derivatives. The rhodamine SA-SPCs showed excellent photocatalytic hydrogen evolution rates (up to 5.9 mmol g À1 h À1 ). ESR spectroscopic analysis revealed that the photocatalytic reaction proceeded via an electron transfer mechanism. We think that the concept of AI-PCA might be applicable to a wide range of photoactive molecules. Further investigations on the effects of organic dyes, morphologies, and the molecular arrangements of supramolecular assemblies on the photocatalytic activity of SA-SPCs, and the detailed mechanism of AI-PCA are currently underway.

Conflicts of interest
There are no conicts to declare.