Photophysical Properties of Spirobifluorene-Based o-Carboranyl Compounds Altered by Structurally Rotating the Carborane Cages

9,9′-Spirobifluorene-based o-carboranyl compounds C1 and C2 were prepared and fully characterized by multinuclear nuclear magnetic resonance (NMR) spectroscopy and elemental analysis. The solid-state structure of C1 was also determined by single-crystal X-ray diffractometry. The two carboranyl compounds display major absorption bands that are assigned to π−π* transitions involving their spirobifluorene groups, as well as weak intramolecular charge-transfer (ICT) transitions between the o-carboranes and their spirobifluorene groups. While C1 only exhibited high-energy emissions (λem = ca. 350 nm) in THF at 298 K due to locally excited (LE) states assignable to π−π* transitions involving the spirobifluorene group alone, a remarkable emission in the low-energy region was observed in the rigid state, such as in THF at 77 K or the film state. Furthermore, C2 displays intense dual emissive patterns in both high- and low-energy regions in all states. Electronic transitions that were calculated by time-dependent-DFT (TD-DFT) for each compound based on ground (S0) and first-excited (S1) state optimized structures clearly verify that the low-energy emissions are due to ICT-based radiative decays. Calculated energy barriers that are based on the relative energies associated with changes in the dihedral angle around the o-carborane cages in C1 and C2 clearly reveal that the o-carborane cage in C1 rotates more freely than that in C2. All of the molecular features indicate that ICT-based radiative decay is only available to the rigid state in the absence of structural fluctuations, in particular the free-rotation of the o-carborane cage.


Introduction
To date, closo-ortho-carborane (1,2-C 2 B 10 H 12 ) derivatives, which are well-known icosahedral boron-cluster compounds, have been widely studied for their interesting optoelectronic properties, which are based on a variety of organic/organometallic luminophores . Indeed, these desirable electronic properties are due to the unique characteristics of the o-carborane unit, such as its strongly electron-deficient nature, as well as its highly polarizable σ-aromaticity. In addition, these properties facilitate the formation of various donor-acceptor systems, leading to intriguing intramolecular charge-transfer (ICT) transitions between numerous π-conjugated aromatic groups and o-carborane cages . Hence, these features have become ultimate sources for generating specific luminescence behavior in a variety of o-carborane-based compounds. Importantly, the ICTtransition-based radiative mechanism that was observed in an o-carborane-incorporated system can be fine-tuned through strategic approaches, such as structurally varying its molecular geometry and the introduction of appropriate electron donors. As representative examples, the C-diazaboryl-ocarboranyl compounds that were reported by Fox and co-workers show dual emissive features that are characterized as locally excited (LE) high-energy states and low-energy ICT states [32]. These properties are the results of changes in the molecular structure and the regulation of the free rotations of the aromatic groups bound to the o-carboranyl units. In addition, various o-carborane-containing donor-acceptor systems exhibit intriguing photophysical properties through twisted intramolecular charge-transfer (TICT) processes [43][44][45][46][47][48][49][50][51][52][53]. Along with these previous studies, our group reported significant changes in radiative ICT mechanisms by modulating the planarities of several phenylene groups that were substituted on o-carborane cages [60][61][62]. Accordingly, we clearly demonstrated that the appropriate combination of molecular geometry and donor plays a key role in implementing promising optoelectronic features.
In this context, we selected 9,9′-spirobifluorene as the donor for attachment to the o-carborane in order to investigate the relationship between geometry and optoelectronic behavior in more detail. 9,9′-Spirobifluorene derivatives have been extensively used as prominent versatile materials [63] in a variety of optoelectronic applications, such as organic light-emitting diodes (OLEDs) [64][65][66][67][68][69] and photovoltaic cells [70][71][72]. Herein, we report two spirobifluorene-based o-carboranyl compounds as novel donor-acceptor systems ( Figure 1, C1 and C2). The synthesis procedures, characterization data, and detailed photophysical properties are presented in conjunction with computational results.

Synthesis of C1
Excess Et 2 S (2.5 equiv.) was added at room temperature to a toluene solution (15 mL) of decaborane (B 10 H 14 , 1.53 mmol) and E1 (0.40 g, 1.18 mmol). After heating to reflux, the reaction mixture was further stirred for 72 h. The solvent and volatiles were removed under vacuum and methanol (10 mL) was added. The resulting solid was filtered and re-dissolved in toluene. The solution was purified by passing through a basic alumina column and the solvent was removed in vacuo. The product was purified by column chromatography on basic alumina (eluent: ether/n-hexane = 1/5, v/v) to yield C1 as a white solid, 0.34 g (yield = 62.2%). 1

UV/Vis Absorption and Photoluminescence (PL) Experiments
Solution UV/Vis absorption and PL experiments were performed in degassed solvents (toluene, THF, and DCM) in 1-cm quartz cuvettes (30 µM) at 298 K. PL was also investigated in THF solution at 77 K and the film state (5 wt% doped on PMMA) for each closo-o-carborane compound on 1-mm-thick 1.5 × 1.5 cm quartz plates at 298 K. Absolute photoluminescence quantum yields (PLQYs, Φ em ) for the solutions and films were obtained while using an absolute PL quantum yield spectrophotometer (FM-SPHERE, 3.2-inch internal integrating sphere on FluoroMax-4P, HORIBA) at 298 K.

X-Ray Crystallography
A single crystal of C1 suitable for X-ray diffractometry was grown from a DCM/n-hexane mixture. The single crystal was coated with Paratone oil and then mounted in a glass capillary.
Crystallographic experiments were performed on a Bruker D8 QUEST CCD area detector diffractometer with graphite-monochromated Mo-Kα radiation (λ = 0.71073 Å). Structures were solved by direct methods and all of the non-hydrogen atoms were subjected to anisotropic refinement while using the full-matrix least-squares method on F 2 in the SHELXTL/PC package to obtain X-ray crystallographic data in CIF format (CCDC 1947010). Hydrogen atoms on the carbon and boron atoms were placed at their geometrically calculated positions and refined as riding on the corresponding carbon and boron atoms with isotropic thermal parameters. Detailed crystallographic data are provided in Tables S1 and S2

Computational Studies
The geometries of C1 and C2 in their ground (S 0 ) and first-excited (S 1 ) states in THF were optimized at the B3LYP/6-31G(d,p) [73] level of theory. Vertical excitation energies at the optimized groundand first-excited-state geometries were calculated while using the time-dependent density functional theory (TD-DFT) method [74] at the same level of theory. Solvent effects were included while using the conductor-like polarizable continuum model (C-PCM) [75,76]. We constructed one dimensional potential energy surfaces (PESs) as function of dihedral angle (Ψ = C1-C2-C3-C4) by rotating the o-carborane cage, vide infra ( Figure 5) in the 0-180 • range in 30 • steps in order to determine the most stable geometries. Among the 49 initial conformations for each compound, conformations that show unphysical atomic overlaps were excluded from further geometry optimization. The dihedral angle was fixed, while the other geometrical variables were fully relaxed during geometry optimizations and energy calculations of the resultant initial conformations. All geometry optimizations and energy calculations were performed while using the Gaussian 16 program [77]. The frequency-checks for all of the optimized structures were performed and no imaginary frequency was confirmed. The percent contribution of a group in a molecule to each molecular orbital was calculated while using the GaussSum 3.0 program [78]. GaussView 6 was used to visualize molecular properties [79]. Figure 1 shows the overall routes for the syntheses of the spirobifluorene-based carboranyl compounds C1 and C2, with o-carborane cages substituted with fluorene moieties at their C2 and C4 positions, respectively. The Sonogashira-coupling reaction between ethynyltrimethylsilane and 2-or 4-bromo-9,9 -spirobifluorene produced (9,9 -spirobi[fluoren]-2-ylethynyl)trimethylsilane (S1) and (9,9 -spirobi[fluoren]-4-ylethynyl)trimethylsilane (S2) in high yields (75% for S1 and 80% for S2). The protonated compounds E1 and E2 were obtained by the treatment of S1 and S2 with a weak base (K 2 CO 3 ), after which the 2-and 4-o-carborane-substituted spirobifluorenes C1 and C2 were respectively prepared from E1 and E2 while using decaborane-promoted (B 10 H 14 -promoted) cage-forming reactions in the presence of Et 2 S (Figure 1) [61][62][63]. All of the compounds were fully characterized by multinuclear ( 1 H, 13 C, and 11 B{ 1 H}) NMR spectroscopy (Figures S1-S6) and elemental analysis. The 1 H and 13 C NMR spectra of C1 and C2 show resonances that correspond to the spirobifluorene moieties as well as the o-carboranyl groups, as expected. In particular, the peaks at~66 ppm that were observed in the 13 C NMR spectra correspond to the carbon atoms of the tetragonal centers of the spirobifluorene moieties. In addition, two signals are observed at around 77 and 60 ppm, which are assignable to the C atoms of the o-carboranyl groups. Six broad 11 B NMR signals that were observed between −2 and −13 ppm clearly confirm the presence of the closo-carborane cages in C1 and C2. Furthermore, the solid-state molecular structure of C1 was determined by single-crystal X-ray diffractometry, as shown in Figure 2; detailed data are available in Table S1, while selected bond lengths and angles are listed in Table S2. The structure clearly exhibits a tetrahedral carbon center (average bond angle at C15 = 109.57 • , Table S2) in the 9,9 -spirobifluorene moiety that was attached at the C2 position of the o-carborane cage.

Photophysical Properties
The photophysical properties of the spirobifluorene-based o-carboranyl compounds C1 and C2 were investigated by UV/Vis-absorption and PL spectroscopies ( Figure 3 and Table 1). C1 and C2 display major absorption bands in the 270-322 nm region with structureless vibronic features. These bands are assigned to spin-allowed π−π* local transitions on each spirobifluorene moiety and typical intramolecular charge-transfer (ICT) transitions between each o-carborane unit and spirobifluorene group (see TD-DFT results, vide infra). The emissive properties of both compounds were examined by PL experiments under a variety of conditions ( Figure 3 and Table 1). The emission spectrum of C1 in THF at 298 K exhibits an intense emission in the high-energy region at λem = 349 nm due to π−π* transitions that are based on the spirobifluorene moiety. The emission band was consistently maintained in a variety of solvents of different polarity (λem = 349-350 nm, Table 1 and Figure S7a), which further confirms the locally excited (LE) emission. Meanwhile, the PL spectra of C1 in the rigid state (THF at 77 K and the film state (5 wt% doped on poly(methyl methacrylate) (PMMA)) show intriguing dual emissive patterns, each with an intense high-energy LE emission and a significantly broad low-energy emission (λem = 476 nm in THF at 77 K and λem = 490 nm in the film) that tailed to 600 nm in the low-energy region. We conclude that these low-energy emissions are closely associated with intramolecular charge-transfer (ICT) transitions between the o-carborane cage and the spirobifluorene moiety based on the computational results (vide infra).
Interestingly, such dual emissive behavior is also observed in the PL spectra of C2 under all conditions (in THF at 298 K and 77 K, and in the film state, Figure 3). The ICT character of the lowenergy emission from C2 was confirmed by the solvatochromism that was observed in various solvents of different polarity (Table 1 and Figure S8b). These interesting features suggest that the ICTbased radiative decay processes in C1 and C2 are significantly enhanced in rigid molecular states that effectively prevent structural fluctuations, such as the C−C bond variations that occur in o-carborane cages [8,35,55,61,62]. Indeed, the DFT-optimized structures of the ground (S0) and first-excited singlet (S1) states of both C1 and C2 confirm such structural fluctuations. The bond lengths (2.38 Å in C1 and 2.45 Å in C2) in their S1-optimized structures are much longer than those (≈1.7 Å) of their S0-optimized structures. In particular, the bond length in C1 in the S1 state is also considerably longer than that (1.662 Å, Table S2) determined by X-ray diffractometry. Importantly, the ICT-based emission in THF at 298 K was only observed for C2, which indicated that C2 undergoes smaller structural variations in solution than C1, which is supported by additional relative free-energy calculations in which the dihedral angles around the o-carboranyl cages in C1 and C2 were altered (vide infra).
The absolute PL quantum efficiencies (Φem) and decay lifetimes (τobs) of C1 and C2 were determined in dilute THF solutions and film states at 298 K to gain insight into the relationship between structure and the radiative decay mechanism for the ICT-based emission (Table 1 and Figures S8 and S9). The emission decay lifetimes, which were measured to be 1.2-1.5 ns for both compounds, reveal that these emissions are fluorescent. The Φem values of C2 in THF at 298 K and

Photophysical Properties
The photophysical properties of the spirobifluorene-based o-carboranyl compounds C1 and C2 were investigated by UV/Vis-absorption and PL spectroscopies ( Figure 3 and Table 1). C1 and C2 display major absorption bands in the 270-322 nm region with structureless vibronic features. These bands are assigned to spin-allowed π−π* local transitions on each spirobifluorene moiety and typical intramolecular charge-transfer (ICT) transitions between each o-carborane unit and spirobifluorene group (see TD-DFT results, vide infra). The emissive properties of both compounds were examined by PL experiments under a variety of conditions ( Figure 3 and Table 1). The emission spectrum of C1 in THF at 298 K exhibits an intense emission in the high-energy region at λ em = 349 nm due to π−π* transitions that are based on the spirobifluorene moiety. The emission band was consistently maintained in a variety of solvents of different polarity (λ em = 349-350 nm, Table 1 and Figure S7a), which further confirms the locally excited (LE) emission. Meanwhile, the PL spectra of C1 in the rigid state (THF at 77 K and the film state (5 wt% doped on poly(methyl methacrylate) (PMMA)) show intriguing dual emissive patterns, each with an intense high-energy LE emission and a significantly broad low-energy emission (λ em = 476 nm in THF at 77 K and λ em = 490 nm in the film) that tailed to 600 nm in the low-energy region. We conclude that these low-energy emissions are closely associated with intramolecular charge-transfer (ICT) transitions between the o-carborane cage and the spirobifluorene moiety based on the computational results (vide infra).
Interestingly, such dual emissive behavior is also observed in the PL spectra of C2 under all conditions (in THF at 298 K and 77 K, and in the film state, Figure 3). The ICT character of the low-energy emission from C2 was confirmed by the solvatochromism that was observed in various solvents of different polarity (Table 1 and Figure S8b). These interesting features suggest that the ICT-based radiative decay processes in C1 and C2 are significantly enhanced in rigid molecular states that effectively prevent structural fluctuations, such as the C−C bond variations that occur in o-carborane cages [8,35,55,61,62]. Indeed, the DFT-optimized structures of the ground (S 0 ) and first-excited singlet (S 1 ) states of both C1 and C2 confirm such structural fluctuations. The bond lengths (2.38 Å in C1 and 2.45 Å in C2) in their S 1 -optimized structures are much longer than those (≈1.7 Å) of their S 0 -optimized structures. In particular, the bond length in C1 in the S 1 state is also considerably longer than that (1.662 Å, Table S2) determined by X-ray diffractometry. Importantly, the ICT-based emission in THF at 298 K was only observed for C2, which indicated that C2 undergoes smaller structural variations in solution than C1, which is supported by additional relative free-energy calculations in which the dihedral angles around the o-carboranyl cages in C1 and C2 were altered (vide infra).
The absolute PL quantum efficiencies (Φ em ) and decay lifetimes (τ obs ) of C1 and C2 were determined in dilute THF solutions and film states at 298 K to gain insight into the relationship between structure and the radiative decay mechanism for the ICT-based emission (Table 1 and Figures S8  and S9). The emission decay lifetimes, which were measured to be 1.2-1.5 ns for both compounds, reveal that these emissions are fluorescent. The Φ em values of C2 in THF at 298 K and the film were determined to be 7% and 41%, respectively; however, the value for C1 could only be measured in the film state, where it was found to be 2%. In the film state, the radiative decay constant (k r , 2.7 × 10 8 s −1 , Table 1) for the ICT-based emission from C2, as calculated from the Φ em and τ obs values, is five-times larger than that (0.5 × 10 8 s −1 ) in THF, whereas the nonradiative decay constant (k nr ) was calculated to be 3.9 × 10 8 s −1 , which is much smaller than that (6.6 × 10 8 s −1 ) determined in THF. These results reveal that restricting structural variations can lead to more radiative decay and fewer non-radiative decay ICT transitions. Interestingly, the k r value (1.7 × 10 7 s −1 ) of C1 in the film is significantly (15-times) lower than that of C2, but its k nr value (8.3 × 10 8 s −1 ) is much larger than that of C2. These findings demonstrate that structural fluctuations in C1, especially the free rotation of the o-carboranyl cage, are more severe than those in C2.
Molecules 2019, x, x 7 of 14 the film were determined to be 7% and 41%, respectively; however, the value for C1 could only be measured in the film state, where it was found to be 2%. In the film state, the radiative decay constant (kr, 2.7 × 10 8 s −1 , Table 1) for the ICT-based emission from C2, as calculated from the Φem and τobs values, is five-times larger than that (0.5 × 10 8 s −1 ) in THF, whereas the nonradiative decay constant (knr) was calculated to be 3.9 × 10 8 s −1 , which is much smaller than that (6.6 × 10 8 s −1 ) determined in THF. These results reveal that restricting structural variations can lead to more radiative decay and fewer nonradiative decay ICT transitions. Interestingly, the kr value (1.7 × 10 7 s −1 ) of C1 in the film is significantly (15-times) lower than that of C2, but its knr value (8.3 × 10 8 s −1 ) is much larger than that of C2. These findings demonstrate that structural fluctuations in C1, especially the free rotation of the o-carboranyl cage, are more severe than those in C2.

Computational Chemistry and Orbital Analyses
To elucidate the nature of the electronic transitions and to analyze the orbitals in C1 and C2, their S0-and S1-optimized structures were subjected to TD-DFT calculations while using the B3LYP functional ( Figure 4 and Table 2); the geometries were optimized from initial structures based on the X-ray crystal structure of C1. To include the effects of the THF solvent [75,76], a conductor-like polarizable continuum model was also used. The computational data for the S0 state show that HOMO → LUMO transitions are the lowest-energy electronic transitions in C1 and C2. The HOMO of each compound is entirely localized on the bifluorene moiety (>99%, Tables S4 and S6), whereas  1 c = 30 µM in THF. 2 c = 30 µM, observed at 298 K. 3 Measured in the film state (5 wt% doped on PMMA) at 298 K. 4 Absolute PL quantum yield. 5 Measured for the ICT-based emissive band. 6 k r = Φ em /τ. 7 k nr = k r (1/Φ em −1). 8 Not observed due to weak emission.

Computational Chemistry and Orbital Analyses
To elucidate the nature of the electronic transitions and to analyze the orbitals in C1 and C2, their S 0 -and S 1 -optimized structures were subjected to TD-DFT calculations while using the B3LYP functional ( Figure 4 and Table 2); the geometries were optimized from initial structures based on the X-ray crystal structure of C1. To include the effects of the THF solvent [75,76], a conductor-like polarizable continuum model was also used. The computational data for the S 0 state show that HOMO → LUMO transitions are the lowest-energy electronic transitions in C1 and C2. The HOMO of each compound is entirely localized on the bifluorene moiety (>99% , Tables S4 and S6), whereas the orbital contribution of the o-carborane unit to each LUMO is slightly higher, at >15%. These results indicate that the lowest-energy absorptions of C1 and C2 are attributable to the π−π* transitions in the bifluorene moieties, with minor contributions from ICT transitions between the o-carborane and fluorene groups. All of the calculated results based on the optimized S 0 structures are in good agreement with the experimentally observed UV/Vis absorption spectra.
On the other hand, the calculated results for the S 1 states of C1 and C2 show that the major transitions associated with the low-energy emissions involve both HOMO → LUMO and HOMO−1 → LUMO+1 transitions ( Figure 4 and Table 2). While the LUMO of each compound is significantly localized in the o-carborane moiety (∼82%, Tables S4 and S6), each HOMO is predominantly located in the fluorene group (>99%). These results strongly suggest that the experimentally observed emissions in the low-energy regions mainly originate from ICT transitions between the o-carborane and aryl moieties. In addition, each HOMO−1 and LUMO+1 are mostly located on the bifluorene group (>92%, Tables S4 and S6), which strongly suggests that the intense emissions observed in the high-energy region centered at~349 nm for C1 and~356 nm for C2 originate from π−π* transitions in the bifluorene moieties; i.e., LE-based emissions. Consequently, the electronic transitions that occur in each o-carboranyl compound were precisely predicted while using computational methods. the orbital contribution of the o-carborane unit to each LUMO is slightly higher, at >15%. These results indicate that the lowest-energy absorptions of C1 and C2 are attributable to the π−π* transitions in the bifluorene moieties, with minor contributions from ICT transitions between the o-carborane and fluorene groups. All of the calculated results based on the optimized S0 structures are in good agreement with the experimentally observed UV/Vis absorption spectra.
On the other hand, the calculated results for the S1 states of C1 and C2 show that the major transitions associated with the low-energy emissions involve both HOMO → LUMO and HOMO−1 → LUMO+1 transitions ( Figure 4 and Table 2). While the LUMO of each compound is significantly localized in the o-carborane moiety (∼82%, Tables S4 and S6), each HOMO is predominantly located in the fluorene group (>99%). These results strongly suggest that the experimentally observed emissions in the low-energy regions mainly originate from ICT transitions between the o-carborane and aryl moieties. In addition, each HOMO−1 and LUMO+1 are mostly located on the bifluorene group (>92%, Tables S4 and S6), which strongly suggests that the intense emissions observed in the high-energy region centered at ~349 nm for C1 and ~356 nm for C2 originate from π−π* transitions in the bifluorene moieties; i.e., LE-based emissions. Consequently, the electronic transitions that occur in each o-carboranyl compound were precisely predicted while using computational methods.   1 Singlet energies for vertical transitions were calculated while using optimized S1 geometries.

DFT Energy-Barrier Calculations for Rotational Motion of the o-Carboranyl Cage
The energies of C1 and C2 in their ground states (S0) were calculated as functions of the dihedral angles associated with their o-carboranyl cages (Ψ: C1-C2-C3-C4, Figure 5) using the B3LYP functional

DFT Energy-Barrier Calculations for Rotational Motion of the o-Carboranyl Cage
The energies of C1 and C2 in their ground states (S 0 ) were calculated as functions of the dihedral angles associated with their o-carboranyl cages (Ψ: C 1 -C 2 -C 3 -C 4 , Figure 5) using the B3LYP functional and the 6-31G(d) basis set to gain insight into the relationship between the radiative mechanism for the ICT-based emissions observed for the spirobifluorenyl compound and the structure of its o-carboranyl moiety. The thermodynamic stability of each system is reported as the relative energy (∆E/kcal·mol −1 ) against that of the S 0 -optimized structure of C1 in THF (i.e., the thermal energy of the optimized structure of C1 was assigned to be: E = 0 kcal·mol −1 ). The relative energy was calculated at each Ψ value in the 0-180 • range in 30 • steps, with the resulting energy diagrams for C1 and C2 displayed in Figure 5. The Ψ values for the S 0 -optimized structures of C1 and C2 were 30.2 • and 152.4 • , respectively, which correspond to their lowest relative energies. Importantly, the relative energy barrier for C1 does not exceed 0.35 kcal·mol −1 over the entire Ψ range ( Figure 5, squares), while the low energy region of C2 is mainly centered at Ψ = 150 • (Figure 6, circles). In addition, the energy gap between the minimum and maximum points for C2 is nearly 1.2 kcal·mol −1 . Based on the thermal energy (0.59 kcal·mol −1 calculated as k B T (k B = Boltzmann constant, 1.98 cal·K −1 ·mol −1 )) at 298 K, these results strongly indicate that the o-carborane cage in C1 can freely rotate at room temperature, whereas the o-carborane cage in C2 is fixed in all states. Consequently, these results verify that the structural rigidity of C2 induces efficient radiative decay that is based on ICT transition associated with the o-carborane moiety. Furthermore, the ICT-based emission from C2 in the solution state (30 µM in toluene) was observed to gradually decrease with increasing temperature, while the LE emission was essentially unchanged. These findings also confirm that the structural changes associated with the o-carborane cages can distinctly turn off radiative decay that is based on ICT transitions ( Figure 6).
Molecules 2019, x, x 9 of 14 and the 6-31G(d) basis set to gain insight into the relationship between the radiative mechanism for the ICT-based emissions observed for the spirobifluorenyl compound and the structure of its ocarboranyl moiety. The thermodynamic stability of each system is reported as the relative energy (ΔE/kcal·mol −1 ) against that of the S0-optimized structure of C1 in THF (i.e., the thermal energy of the optimized structure of C1 was assigned to be: E = 0 kcal·mol −1 ). The relative energy was calculated at each Ψ value in the 0-180° range in 30° steps, with the resulting energy diagrams for C1 and C2 displayed in Figure 5. The Ψ values for the S0-optimized structures of C1 and C2 were 30.2° and 152.4°, respectively, which correspond to their lowest relative energies. Importantly, the relative energy barrier for C1 does not exceed 0.35 kcal·mol −1 over the entire Ψ range ( Figure 5, squares), while the low energy region of C2 is mainly centered at Ψ = 150° (Figure 6, circles). In addition, the energy gap between the minimum and maximum points for C2 is nearly 1.2 kcal·mol −1 . Based on the thermal energy (0.59 kcal·mol −1 calculated as kBT (kB = Boltzmann constant, 1.98 cal·K −1 ·mol −1 )) at 298 K, these results strongly indicate that the o-carborane cage in C1 can freely rotate at room temperature, whereas the o-carborane cage in C2 is fixed in all states. Consequently, these results verify that the structural rigidity of C2 induces efficient radiative decay that is based on ICT transition associated with the o-carborane moiety. Furthermore, the ICT-based emission from C2 in the solution state (30 μM in toluene) was observed to gradually decrease with increasing temperature, while the LE emission was essentially unchanged. These findings also confirm that the structural changes associated with the o-carborane cages can distinctly turn off radiative decay that is based on ICT transitions ( Figure 6).

Conclusions
We prepared and characterized two spirobifluorenyl o-carborane compounds, C1 and C2, with the solid-state structure of C1 determined by single-crystal X-ray diffractometry. In particular, C2 showed clear and intense ICT-based emissions that involve the o-carborane moiety in all states; however, C1 only displayed an LE-based emission that was attributable to π−π* transitions centered on the spirobifluorenyl group in solution at 298 K, while C2 exhibited additional ICT-based emissions in the rigid state. Energy-barriers that were determined from calculated relative free-energies of rotated o-carborane cages in both C1 and C2 decisively show that the o-carborane cage in C1 more freely rotates than that in C2. Consequently, all of the characterization and computational results definitively show that large structural variations involving the o-carborane cage, namely its freerotation, hinder ICT-based radiative decay that is associated with the o-carborane. This finding provides new decisive evidence that clarifies the relationship between structure and the efficiency of the ICT-based radiative decay that is associated with the o-carborane.