2. Materials and Methods
All reagents were commercially sourced. Mesitylene, iron powder, bromine,
N-bromosuccinimide, benzoyl peroxide, sodium acetate, thionyl chloride, potassium carbonate, 1-bromo-3-chloropropane,
m-aminophenol, 1,5-dibromopentane,
N-ethyldiisopropylamine, copper powder, and organic solvents were purchased from FUJIFILM Wako pure Chemical Co., Ltd. (Osaka, Japan). 2-Bromophenol and 3-methoxyphenol were purchased from Tokyo Chemical Industries Co., Ltd. (Tokyo, Japan). Melting points were determined using a Yanaco MP-500P micro-melting point apparatus.
1H (600 or 400 MHz) and
13C (150 or 100 MHz) nuclear magnetic resonance (NMR) spectra were recorded using a Bruker AVANCE instrument. The following abbreviations were used to describe the multiplicities: singlet (s), doublet (d), triplet (t), doublet of doublets (dd), and multiplet (m). Absorption spectra were recorded using a JASCO V-560 instrument. High-resolution mass spectrometry (HRMS) spectra were recorded on a Thermo Fisher Scientific (Tokyo, Japan), Exactive Plus Orbitrap mass spectrometer for ionization. Only relatively intense peaks and structurally diagnostic mass spectral fragment ion peaks are reported. All quantum chemical calculations were performed using the Gaussian 16 program (Revision C.01) [
30]. The optimized structures of symmetrical and asymmetrical triskelion-shaped molecules were determined using density functional theory (DFT) calculations. All DFT calculations were performed using the ωB97-XD long-range corrected hybrid functional and 6-31G(d) basis set. Geometry optimizations were performed for the symmetrical derivatives while maintaining the
C3 symmetry. Normal mode analysis calculations were conducted at the same level of theory to ensure that local minima optimized structures were obtained. Electronic excitation energies and natural transition orbitals [
31] were estimated for all triskelion-shaped derivatives using time-dependent (TD)-DFT calculations. Fluorescence spectra were collected using a HITACHI F-4500 fluorescence spectrometer and a JASCO FP-8550 spectrofluorometer. Relative fluorescence quantum yields (
FF) were determined using rhodamine B in EtOH as the standard (
FF = 0.9). Dynamic light scattering (DLS) measurements were recorded using a Sysmex Zetasizer NanoZS instrument.
The syntheses of
1A–
E and
8A–
E have been described in our previous reports [
28,
29].
Synthesis of 9. To a 50 mL recovery flask equipped with a reflux condenser were added K2CO3 (28 mg, 0.20 mmol), 7d (50 mg, 0.20 mmol), 6 (102 mg, 0.20 mmol), and acetone (15 mL). The reaction mixture was refluxed overnight. After cooling to ambient temperature, residual solids were removed by filtration and washed with CHCl3. The filtrate was then evaporated under reduced pressure. The residue was purified by column chromatography on silica gel (CH2Cl2: hexane = 4:1 (v/v)) to obtain 9 (81 mg, 0.11 mmol) as yellow block crystals in 17% yield. Mp = 155 °C (decomp.). 1H NMR (400 MHz, CDCl3) δ 7.44 (d, J = 8.8 Hz, 1H), 6.88 (d, J = 2.8 Hz, 1H), 6.76 (dd, J = 2.8 and 8.8 Hz, 1H), 3.17 (t, J = 5.6 Hz, 4H), 1.70 (m, 4H), 1.62–1.58 (m, 2H). 13C NMR (100 MHz, CDCl3) δ 165.2, 161.5, 152.5, 147.9, 141.8, 138.2, 133.7, 117.6, 116.0, 114.7, 110.2, 102.6, 50.0, 25.6, 24.2. UV-vis (CH2Cl2, c = 1.0 × 10−5 M) λmax (ε) 261 (17,300) nm. HRMS (ESI, positive mode): m/z calcd for C20H14Br4Cl2NO4: [M + H]+ 721.6987; found: 721.6986.
Synthesis of 10. To a 50 mL recovery flask equipped with a reflux condenser were added K2CO3 (83 mg, 0.60 mmol), 7d (152 mg, 0.60 mmol), 6 (150 mg, 0.30 mmol), and acetone (15 mL). The reaction mixture was refluxed overnight. After cooling to ambient temperature, the residual solids were removed by filtration and washed with CHCl3. The filtrate was then evaporated under reduced pressure and the residue was purified by column chromatography on silica gel (CH2Cl2: hexane = 4:1 (v/v)) to obtain 10 (140 mg, 0.15 mmol) as yellow block crystals in 50% yield. Mp = 185 °C (decomp.). 1H NMR (400 MHz, CDCl3) δ 7.45–7.43 (m, 2H), 6.92–6.91 (m, 2H), 6.77–6.74 (m, 2H), 3.19–3.16 (m, 8H), 1.73–1.68 (m, 8H), 1.62–1.58 (m, 4H). 13C NMR (100 MHz, CDCl3) δ 165.5, 161.9, 152.5, 147.9, 141.7, 138.0, 133.7, 120.0, 117.1, 116.0, 110.3, 102.8, 50.1, 25.6, 24.2. UV-vis (CH2Cl2, c = 1.0 × 10–5 M) λmax (ε) 316 (4400), 263 (34,100) nm. HRMS (ESI, positive mode): m/z calcd for C31H26Br5ClN2O5: [M + H]+ 942.7459; found: 942.7455.
Synthesis of 8aad. To a 50 mL recovery flask equipped with a reflux condenser were added K2CO3 (39 mg, 0.28 mmol), 7a (30 μL, 0.28 mmol), 9 (100 mg, 0.14 mmol), and acetone (15 mL). The reaction mixture was refluxed overnight. After cooling to ambient temperature, the residual solids were removed by filtration and washed with CHCl3. The filtrate was then evaporated under reduced pressure. The residue was purified by column chromatography on silica gel (CH2Cl2: hexane = 4:1 (v/v)) to obtain 8aad (103 mg, 0.10 mmol) as yellow block crystals in 74% yield. Mp = 98 °C (decomp.). 1H NMR (400 MHz, CDCl3) δ 7.69 (dd, J = 1.6 and 8.0 Hz, 2H), 7.50 (dd, J = 1.6 and 8.0 Hz, 2H), 7.45 (d, J = 8.8 Hz, 2H), 7.41 (dd, J = 1.6 and 8.0 Hz, 1H), 3.18 (t, J = 5.6 Hz, 4H), 1.73–1.68 (m, 4H), 1.62–1.58 (m,2H). 13C NMR (100 MHz, CDCl3) δ 162.2, 162.1, 152.5, 148.0, 147.6, 137.9, 137.7, 134.1, 133.7, 128.7, 128.3, 123.5, 119.6, 119.5, 116.0, 115.6, 110.4, 102.8, 50.1, 25.6, 24.2. UV-vis (CH2Cl2, c = 1.0 × 10–5 M) λmax (ε) 266 (15,500) nm. HRMS (ESI, positive mode): m/z calcd for C32H21Br6NO6: [M + H]+ 995.6481; found: 995.6480.
Synthesis of 8add. To a 50 mL recovery flask equipped with a reflux condenser were added K2CO3 (15 mg, 0.11 mmol), 7a (11 μL, 0.11 mmol), 10 (101 mg, 0.11 mmol), and acetone (15 mL). The reaction mixture was refluxed overnight. After cooling to ambient temperature, the residual solids were removed by filtration and washed with CHCl3. The filtrate was then evaporated under reduced pressure. The residue was purified by column chromatography on silica gel (CH2Cl2: hexane = 4:1 (v/v)) to obtain 8add (66 mg, 0.06 mmol) as colorless block crystals in 58% yield. Mp = 268 °C (decomp.). 1H NMR (400 MHz, CDCl3) δ 7.70–7.68 (m, 1H), 7.52–7.49 (m, 1H), 7.46–7.45 (m, 2H), 7.44–7.40 (m, 1H), 7.24–7.20 (m, 1H), 6.96–6.95 (m, 1H), 6.77–6.74 (m, 2H), 3.20–3.17 (m, 8H), 1.74–1.69 (m, 8H), 1.63–1.59 (m, 4H). 13C NMR (100 MHz, CDCl3) δ 162.3, 162.2, 152.5, 148.1, 147.7, 138.0, 134.2, 133.7, 128.7, 128.3, 123.5, 119.7, 119.5, 116.0, 115.6, 110.5, 102.9, 50.2, 25.7, 24.2. UV-vis (CH2Cl2, c = 1.0 × 10–5 M) λmax (ε) 264 (31,700) nm. HRMS (ESI, positive mode): m/z calcd for C37H30Br6N2O6: [M + H]+ 1078.7216; found: 1078.7217.
Synthesis of 8ccd. To a 50 mL recovery flask equipped with a reflux condenser were added K2CO3 (40 mg, 0.28 mmol), 7c (70 mg, 0.28 mmol), 9 (101 mg, 0.14 mmol), and acetone (15 mL). The reaction mixture was refluxed overnight. After cooling to ambient temperature, the residual solids were removed by filtration and washed with CHCl3. The filtrate was then evaporated under reduced pressure. The residue was purified by column chromatography on silica gel (CH2Cl2) to obtain 8ccd (78 mg, 0.07 mmol) as yellow block crystals in 52% yield. Mp = 197 °C (decomp.). 1H NMR (400 MHz, CDCl3) δ 7.45–7.42 (m, 3H), 6.96–6.95 (m, 1H), 6.76–6.73 (m, 3H), 6.55–6.52 (m, 2H), 3.19–3.16 (m, 4H), 2.97 (s, 12H), 1.73–1.68 (m, 4H), 1.62–1.58 (m, 2H). 13C NMR (100 MHz, CDCl3) δ 162.3, 162.3, 152.5, 150.8, 148.1, 137.9, 137.9, 133.6, 119.5, 119.5, 115.9, 112.4, 110.5, 106.7, 102.9, 100.4, 50.1, 40.6, 25.7, 24.2. UV-vis (CH2Cl2, c = 1.2 × 10–5 M) λmax (ε) 311 (7400), 266 (38,300), 261 (38,100), 255 (31,400) nm. HRMS (ESI, positive mode): m/z calcd for C36H32Br6N3O6: [M + H]+ 1081.7325; found: 1078.7327.
Synthesis of 8cdd. To a 50 mL recovery flask equipped with a reflux condenser were added K2CO3 (20 mg, 0.14 mmol), 7c (30 mg, 0.14 mmol), 10 (131 mg, 0.14 mmol), and acetone (20 mL). The reaction mixture was refluxed overnight. After cooling to ambient temperature, the residual solids were removed by filtration and washed with CHCl3. The filtrate was then evaporated under reduced pressure. The residue was purified by column chromatography on silica gel (CH2Cl2) to obtain 8cdd (78 mg, 0.07 mmol) as white block crystals in 67% yield. Mp = 197 °C (decomp.). 1H NMR (400 MHz, CDCl3) δ 7.46–7.42 (m, 3H), 6.96–6.95 (m, 2H), 6.77–6.73 (m, 3H), 6.55–6.53 (m, 1H), 3.19–3.17 (m, 8H), 2.97 (s, 6H), 1.74–1.68 (m, 8H), 1.62–1.58 (m, 4H). 13C NMR (100 MHz, CDCl3) δ 162.3, 162.3, 152.5, 150.8, 148.1, 148.1, 137.9, 137.9, 133.7, 119.5, 119.5, 115.9, 112.4, 110.5, 106.7, 102.9, 100.4, 50.1, 40.6, 25.7, 24.4. UV-vis (CH2Cl2, c = 1.0 × 10–5 M) λmax (ε) 319 (7300), 266 (48,100), 259 (42,300), 254 (28,600). HRMS (ESI, positive mode): m/z calcd for C39H36Br6N3O6: [M + H]+ 1121.7638; found: 1121.7638.
Synthesis of 1aad. To a 50 mL Schlenk tube was added 8aad (100 mg, 0.10 mmol), excess activated Cu Powder (1410 mg, 22.0 mmol), and dry DMF (30 mL) under argon atmosphere. The reaction mixture was refluxed for 18 h. After cooling to ambient temperature, the reaction mixture was filtered through Celite and washed with EtOAc. The organic phase was then washed with H2O and dried over Na2SO4. The solvent was evaporated under reduced pressure, and the residue was purified by flash column chromatography on Al2O3 (CH2Cl2: EtOAc = 30:1 (v/v)) to give 1aad (24 mg, 0.05 mmol) as a reddish solid in 47% yield. Mp = 150 °C (decomp.). 1H NMR (400 MHz, CDCl3) δ 8.01–7.99 (m, 2H), 7.79–7.77 (m, 1H), 7.60–7.55 (m, 2H), 7.40–7.38 (m, 2H), 7.28–7.24 (m, 2H), 6.77–6.74 (m, 1H), 6.60–6.65 (m, 1H), 3.50–3.44 (m, 4H), 1.74–1.68 (m, 4H), 1.58–1.54 (m, 2H). 13C NMR (100 MHz, CDCl3) δ 159.1, 158.8, 158.5, 154.7, 154.5, 151.6, 151.6, 146.7, 146.7, 146.5, 133.1, 133.0, 131.8, 131.6, 129.0, 123.8, 123.7, 117.3, 117.3, 117.1, 117.0, 115.3, 113.8, 113.1, 110.5, 106.4, 98.9, 48.4, 25.5, 24.5. UV-vis (CH2Cl2, c = 1.1 × 10–5 M) λmax (ε) 472 (20,600), 335 (19,200) nm. HRMS (ESI, positive mode): m/z calcd for C32H21NO6 [M]+ 515.1364; found: 515.1363.
Synthesis of 1add. To a 50 mL Schlenk tube was added 8add (100 mg, 0.09 mmol), excess activated Cu Powder (1300 mg, 20.5 mmol), and dry DMF (28 mL) under argon atmosphere. The reaction mixture was refluxed for 18 h. After cooling to ambient temperature, the reaction mixture was filtered through Celite and washed with EtOAc. The organic phase was then washed with H2O and dried over Na2SO4. The solvent was evaporated under reduced pressure, and the residue was purified by flash column chromatography on Al2O3 (CH2Cl2: EtOAc = 30:1 (v/v)) to give 1add (26 mg, 0.04 mmol) as a reddish solid in 47% yield. Mp = 115 °C (decomp.). 1H NMR (400 MHz, CDCl3) δ 8.98–7.96 (m, 1H), 7.78–7.76 (m, 2H), 7.55–7.51 (m, 1H), 7.37–7.35 (m, 1H), 7.25–7.21 (m, 1H), 6.76–6.72 (m, 2H), 6.65–6.64 (m, 2H), 3.49–3.37 (m, 8H), 1.77–1.64 (m, 12H). 13C NMR (100 MHz, CDCl3) δ 159.6, 159.3, 159.0, 154.5, 154.5, 154.3, 154.3, 151.5, 147.2, 146.9, 146.9, 132.6, 131.7, 131.7, 130.7, 123.5, 117.6, 116.8, 113.0, 112.2, 111.2, 110.4, 106.7, 106.6, 99.1, 99.0, 48.5, 25.5, 24.5. UV-vis (CH2Cl2, c = 1.1 × 10–5 M) λmax (ε) 463 (44,900), 357 (14,400) nm. HRMS (ESI, positive mode): m/z calcd for C37H30N2O6: [M + H]+ 599.2177; found: 599.2176.
Synthesis of 1ccd. To a 50 mL Schlenk tube was added 8ccd (100 mg, 0.09 mmol), excess activated Cu Powder (1330 mg, 20.5 mmol), and dry DMF (28 mL) under argon atmosphere. The reaction mixture was refluxed for 18 h. After cooling to ambient temperature, the reaction mixture was filtered through Celite and washed with EtOAc. The organic phase was then washed with H2O and dried over Na2SO4. The solvent was evaporated under reduced pressure, and the residue was purified by flash column chromatography on Al2O3 (CH2Cl2: EtOAc = 20:1 (v/v)) to give 1ccd (36 mg, 0.06 mmol) as a reddish solid in 64% yield. Mp = 222 °C (decomp.). 1H NMR (400 MHz, CDCl3) δ 7.79–7.77 (m, 3H), 6.75–6.72 (m, 1H), 6.64–6.63 (m, 1H), 6.60–6.57 (m, 2H), 6.48–6.47 (m, 2H), 3.45–3.39 (m, 4H), 3.10 (s, 12H), 1.72–1.66 (m, 6H). 13C NMR (100 MHz, CDCl3) δ 159.9, 159.9, 159.9, 154.3, 154.2, 154.1, 153.6, 1477, 147.6, 147.4, 131.6, 110.3, 110.2, 110.1, 108.4, 107.1, 106.2, 99.2, 97.2, 77.5, 77.2, 76.8, 48.5, 40.2, 25.5, 24.5. UV-vis (CH2Cl2, c = 1.1 × 10–5 M) λmax (ε) 453 (62,900) nm. HRMS (ESI, positive mode): m/z calcd for C36H31N3O6: [M]+ 601.2207; found: 601.2208.
Synthesis of 1cdd. 50 mL Schlenk tube was added 8cdd (101 mg, 0.09 mmol), excess activated Cu Powder (1300 mg, 20.5 mmol), and dry DMF (27 mL) under argon atmosphere. The reaction mixture was refluxed for 18 h. After cooling to ambient temperature, the reaction mixture was filtered through Celite and washed with EtOAc. The organic phase was washed with H2O and dried over Na2SO4. The solvent was evaporated under reduced pressure, and the residue was purified by flash column chromatography on Al2O3 (CH2Cl2: EtOAc = 20:1 (v/v)) to give 1cdd (28 mg, 0.04 mmol) as a reddish solid in 49% yield. Mp = 125 °C (decomp.). 1H NMR (400 MHz, CDCl3) δ 7.79–7.77 (m, 3H), 6.74–6.72 (m, 2H), 6.64–6.63 (m, 2H), 6.59–6.57 (m, 1H), 6.48–6.47 (m, 1H), 3.44–3.40 (m, 8H), 3.10 (s, 6H), 1.72–1.67 (m, 12H). 13C NMR (100 MHz, CDCl3) δ 159.9, 159.9, 159.8, 154.3, 154.2, 154.1, 153.6, 147.7, 147.4, 147.4, 131.6, 110.4, 110.4, 110.3, 110.2, 108.4, 107.1, 106.2, 99.2, 97.2, 48.5, 40.2, 25.5, 24.5. UV-vis (CH2Cl2, c = 1.1 × 10–5 M) λmax (ε) 453 (55,400) nm. HRMS (ESI, positive mode): m/z calcd for C39H36N3O6: [M + H]+ 642.2599; found: 642.2598.
3. Results and Discussion
The synthetic route to
C3 symmetrical compounds
1A–
E is depicted in
Scheme 1. The central benzene core
6 was efficiently prepared from mesitylene (
2) in five steps, according to the literature procedures [
32,
33,
34,
35]. Compound
6 was reacted with the corresponding 2-bromophenol derivatives in the presence of K
2CO
3 in acetone to afford precursors
8A–
E in good yields. Finally, the targeted molecular triskelions
1A–
E were obtained via an intramolecular Ullmann coupling reaction. The
C3 symmetric derivatives were fully characterized using
1H and
13C NMR spectroscopy and high-resolution mass spectrometry (HRMS) [
28,
29].
Subsequently, we employed the synthetic routes for
1A–
E (
Scheme 2) for the synthesis of asymmetric triskelions. First,
6 and 1.0 or 2.0 eq. of
7d were treated with K
2CO
3 in acetone to give mono-substituted
9 and di-substituted
10 in 19% and 50% yield, respectively. Next, asymmetric precursors
8aad,
8add,
8ccd, and
8cdd were obtained in 52–74% yield using 2.0 eq. of
7a or
7c for
9, and 1.0 eq. of
7a or
7c for
10. Owing to the nucleophilic reactivity of
6, it was difficult to introduce three different-type phenyl units into the central benzene core. The intramolecular Ullmann coupling of
8aad,
8add,
8ccd, and
8cdd in the presence of excess copper in DMF afforded the desired asymmetric triskelions
1aad,
1add,
1ccd, and
1cdd, which were analogous to
1A–
E. Although the
1H and
13C NMR spectra of these derivatives were highly complex due to their
C1 symmetry, the asymmetric triskelions structures were identified (
Figures S13–S20). In this coupling condition, several debrominated compounds of
8, mono- and bis-coupling products were observed as the byproducts, however, they could not be isolated.
The molecular structures of
1A–
D were confirmed using X-ray crystallography [
28,
29]. As expected, the molecular framework of
1 presented a helical geometry arising from intramolecular steric repulsion. This result was in good agreement with the corresponding optimized structures derived from quantum chemical calculations. Similarly,
1aad,
1add,
1ccd, and
1cdd were optimized as twisted propeller-shaped structures (
Figure 2). Furthermore, in the case of these asymmetric triskelions, left-handed
M,M,M- and right-handed
P,P,P-formers were the most stable conformation regardless of the substituted groups.
The maximum transition state energy for the isomerization between (
P,P,P) and (
M,M,M)-
1A was estimated to be 9.46 kcal/mol at the ωB97-XD/6-31G(d) level of theory. Even in the case of
1C, bearing
N-containing donor groups, the corresponding energy was 8.7 kcal/mol. This indicates that these derivatives underwent continuous flipping in diluted solutions. Furthermore, all triskelion-shaped compounds adopted a Janus-type structure and presented two distinct molecular π-surfaces (
Figure 3). Based on the helical triskelion-shaped framework, the electrostatic potential mapping of these compounds demonstrated that the outer phenyl groups on the front surface of
1 were positively charged (blue) and the three electron-withdrawing carbonyl groups on the back face were negatively charged (red), regardless of the molecular symmetry. Thus, these molecules exhibit out-of-plane anisotropy via the central benzene ring from the positive area on the front to the negative area on the back face. The estimated dipole moments of the derivatives bearing donor groups reached 7.83 Debye, which is higher than those of bowl-shaped sumanene [
36] and Janus-type subphthalocyanine [
37]. Interestingly, the dipole moments of asymmetrical derivatives
1aad,
1add,
1ccd, and
1cdd were also showed large owing to the triskelion-shaped framework. We believe that this Janus-type anisotropy is beneficial for nanomaterials applicable in molecular recognition that operates through molecular rearrangement control [
38].
The absorption and fluorescence spectra of asymmetrical triskelion-shaped coumarin dyes
1aad,
1add,
1ccd, and
1cdd were recorded in toluene to investigate their photophysical properties and compare them with those of
C3 symmetrical dyes
1A,
1C, and
1D (
Figure 4). As shown in
Figure 4a, in the absorption spectra of the symmetrical dyes, a single peak was observed at 342 nm for
1A and at 443 nm for
1C and
1D. In contrast, two peaks were observed in the absorption spectra of
1aad (335 and 472 nm) and
1add (357 and 463 nm). On the other hand, the absorption spectra of
1ccd and
1cdd only contained one peak at 443 nm, similar to the results for the symmetrical dyes. These results indicated that the absorption maxima of
1aad (λ
max: 472 nm) and
1add (λ
max: 463 nm) exhibited a bathochromic shift compared to that of
1A (λ
max: 342 nm). This shift stems from the narrow gap between the highest occupied molecular orbital (HOMO)–lowest unoccupied molecular orbital (LUMO) resulting from reduced molecular symmetry (
Figure 5). Conversely, the absorption bands of
1ccd and
1cdd at 443 nm resembled that of
1C (λ
max: 443 nm). Although there is not much difference among
1aad,
1add,
1ccd, and
1cdd, the HOMO and LUMO levels of
1ccd and
1cdd were similar to those of
N-substituted
C3 triskelions due to the presence of strong donor groups, despite their
C1 molecular symmetry.
To further elucidate the photochemical behavior of the asymmetrical triskelion-shaped coumarin dyes, their excited states were examined using quantum chemical calculations.
Figure 6 shows the absorption spectra simulated using quantum chemical calculations and the main components of the natural transition orbitals (NTOs) for each electronic excitation. The simulated absorption spectra are in good agreement with the experimental results (
Figure 4a), wherein two peaks were observed in the spectra of
1aad and
1add, and a single peak for
1ccd and
1cdd. In the case of
1aad, the NTO analysis shown in
Figure 6 indicates that the S
0 → S
2 transition, the peak of which is observed in the longer-wavelength region, mainly involves the
N-substituted moiety in the molecule (
d unit), whereas the S
0 → S
4 transition, the peak of which appears in the shorter wavelength region, mainly involves two unsubstituted moieties (
a units). Similarly, for
1add, the S
0 → S
3 transition with a peak in the long-wavelength region involves two
N-substituted moieties (
d units), whereas the S
0 → S
4 transition with a peak in the short wavelength region involves one non-substituent moiety (
a unit). On the other hand, in the cases of
1ccd and
1cdd, the S
0 → S
n (
n = 2, 3, and 4) transitions all involve intramolecular
N-substituted moieties, and their transition energies do not differ significantly; thus, they are observed as a single peak in the long wavelength region of the absorption spectra. Another interesting point here is that the oscillation intensity of S
0 → S
4 transition is smaller than that of S
0 → S
2 in the case of
1aad and
1add, but larger in the case of
1ccd and
1cdd. In general, the transition probability depends on the overlap between the molecular orbitals involved in the electronic excitation. Indeed, as shown in
Figure 6, the overlap of the hole-particle pairs of NTOs for the S
0 → S
4 transition is small in the case of
1aad and
1add, but large in the cases of
1ccd and
1cdd. That is, due to the combination of the substitute positions, this framework was found to be designable for controlling the frontier orbitals and electronic transitions. In contrast, the NTO analysis of
C3 symmetrical derivatives
1A–
1E suggested that their corresponding S
0 → S
2 or S
0 → S
3 transitions were composed of two pairs with large eigenvalues (
Figure S24). This is due to the degenerated molecular orbitals resulting from their
C3 symmetry. Therefore, distribution and overlapping from the hole to particle cover the whole of the molecular framework.
In the fluorescence spectra of
1aad,
1add,
1ccd, and
1cdd in toluene, the emission maxima were observed at 619, 606, 593, and 596 nm, respectively, with relatively large Stokes shifts of 5700–5800 cm
−1 (
Figure 4b). These emission maxima were as high as that of
1D (599 nm) or slightly bathochromically shifted. The fluorescence quantum yields of these compounds were low (
ΦF = 0.003–0.007). This inefficiency is possibly caused by thermal non-radiative transitions arising from their helical inversion, in addition to the internal conversion from S
n (
n = 2~4) to S
1. These values are lower than those of obtained for the
C3 symmetrical compounds
1C (
ΦF = 0.14) and
1D (
ΦF = 0.13) in toluene. The orbital degeneracy can be resolved by varying the substitution style on the molecular triskelion. As a result, the orbital overlap between the ground and excited states of the asymmetric derivatives would be reduced owing to their partial localization as shown in NTO analysis.
Next, we investigated the solvent dependence of the photophysical properties of
1aad,
1add,
1ccd, and
1cdd (
Figure S25 and Table S1). All of the compounds demonstrated a gradual bathochromic shift in their absorption spectra depending on the solvent polarity. Conversely, the asymmetrical dyes, expect for
1ccd, exhibited a blue-shift in their fluorescence spectra in highly polar solvents such as DMF and DMSO. The THF and EtOAc solutions of these compounds gave rise to broad emission bands containing shoulder peaks. This tendency was observed in the case of
1C and
1D [
29]. This indicates that asymmetrical compounds undergo at least two radiative decay processes. Intriguingly,
1ccd exhibited a redshift in highly polar solvents. For triskelion-shaped scaffolds, the charge separation structures in the excited states may be controllable depending on the combination of the three substituents. Based on the photophysical data, we evaluated the intramolecular charge transfer (ICT) characteristics of
1aad,
1add,
1ccd, and
1cdd by using Lippert-Mataga plots (
Figure 7) [
39,
40]. Compounds
1aad,
1add, and
1cdd exhibited a negative correlation, whereas
1ccd showed a minimally increasing trend. However, linearity was poor in all cases, indicating that the ICT characteristics of these asymmetrical derivatives depend minimally on solvent polarity, in contract to typical coumarin dyes. It is considered that the resonance structures of
1 in the excited state are destabilized with the loss of aromaticity of the central benzene ring. The fluorescence quantum yields in the tested solvents were considerably low. The cause of low efficiency under diluted solutions is presumed to occur due to the dominant non-radiative process resulting from the helical inversion. We reasoned that fluorescence efficiency can be improved by suppression of molecular inversion in the aggregates, as in the case of AIE-gens [
3].
To investigate the AIE enhancement (AIEE) properties of
1aad,
1add,
1ccd, and
1cdd, we first prepared THF/H
2O solutions (water fraction (Fw): 10–80%) of
1aad,
1add,
1ccd, and
1cdd, and measured their absorption (
Figure S26). For all the compounds, the absorption maxima gradually decreased as the Fw increased to 50–60%. At Fws of ≥60%, the absorbance decreased drastically, and scattering absorption bands were observed in the long-wavelength region. This result indicated that these compounds formed nanoaggregates in the THF/H
2O mixture. Moreover, the DLS results indicated the existence of nanoaggregates with constant particle sizes in the range of 2~700 nm (
Figure S27). As shown in
Figure 8, the fluorescence spectral peaks of
1aad,
1add,
1ccd, and
1cdd at Fws of 10–50% were very low-intensity. Conversely, as the Fw increased to ≥60%, the intensity increased drastically owing to the formation of nanoaggregates. The maximum intensity was observed at Fw = 70%. This suggests that the AIEE characteristics of these asymmetrical triskelions arise due to molecular motions, such as vibrations and inversion, which are restricted in the nanoaggregates. Interestingly, these compounds exhibited a bathochromic shift of approximately 650 nm. Newly formed nanoaggregates of
1aad,
1add,
1ccd, and
1cdd may be affected by the reorientation of H
2O in their excited state. However, at an Fw of 80%, the intensity decreased owing to reprecipitation. Thus, these asymmetrical derivatives are potentially applicable as components of effective AIE-gens owing to their twisted geometry, similar to
C3 symmetrical triskelion-shaped dyes. The introduction of several substituents to the triskelion scaffold enabled the tuning of their orbital levels and molecular packing. It is interested in their emission behavior in the solid-state, such as powder, crystalline, and thin films.