Extended Conjugation Attenuates the Quenching of Aggregation‐Induced Emitters by Photocyclization Pathways

Abstract Herein, we expose how the antagonistic relationship between solid‐state luminescence and photocyclization of oligoaryl alkene chromophores is modulated by the conjugation length of their alkenyl backbones. Heptaaryl cycloheptatriene molecular rotors exhibit aggregation‐induced emission characteristics. We show that their emission is turned off upon breaking the conjugation of the cycloheptatriene by epoxide formation. While this modification is deleterious to photoluminescence, it enables formation of extended polycyclic frameworks by Mallory reactions. We exploit this dichotomy (i) to manipulate emission properties in a controlled manner and (ii) as a synthetic tool to link together pairs of phenyl rings in a specific sequence. This method to alter the tendency of oligoaryl alkenes to undergo photocyclization can inform the design of solid‐state emitters that avoid this quenching mechanism, while also allowing selective cyclization in syntheses of polycyclic aromatic hydrocarbons.


General Methods
Materials: All reagents were purchased from commercial suppliers (Sigma-Aldrich, Acros Organics, or Alfa Aesar) and used without further purification. meta-Chloroperbenzoic acid (mCPBA) was purchased and used as a mixture of >77% purity, where the remainder is m-chlorobenzoic acid and water. in Hertz (Hz). 13 C NMR Experiments were proton decoupled. Assignment of 1 H and 13 C NMR signals were accomplished by two-dimensional (2D) NMR spectroscopy (COSY, NOESY, HSQC, HMBC). NMR spectra were processed using MestReNova version 11. Data are reported as follows: chemical shift; multiplicity; coupling constants; integral and assignment. Low-resolution ASAP-MS were performed using a Waters Xevo QTOF equipped with an Atmospheric Solids Analysis Probe (ASAP). High-resolution electrospray (HRESI) and ASAP (HRASAP) mass spectra were measured using a Waters LCT Premier XE high resolution, accurate mass UPLC ES MS (also with ASAP ion S3 source). Melting points were recorded using a Gallenkanp (Sanyo) apparatus and are uncorrected.

Instrumentation and Analytical
The X-ray single crystal data have been collected at 120.0(2)K using λMoKα radiation (λCuKα for compound asym-phenPh5C7H; λ =0.71073 and 1.54178Å respectively) on a Bruker D8Venture  impurities. This fraction was further purified by recrystallization by slow evaporation of a 5:3 CH2Cl2-EtOH solution, yielding a third batch of asym-phenPh5C7H and sym-phenPh5C7H as a colorless solid (181 mg, 0.29 mmol, 26%). The three batches were combined, giving asym-spectroscopic analysis ( Figure S11) showed the two isomers were present in a 2:1 ratio of asym-phenPh5C7H to sym-phenPh5C7H. This ratio corresponds to a statistical mixture, as asym-phenPh5C7H is present as a racemic mixture of Rand S-stereoisomers. M.P.: 288 -290 °C.  (51 mg, 20 μmol, 2.1 equiv) were added to a 7 mL quartz tube, which was fitted with a septum and purged with N2(g).

HRASAP-MS
Anhydrous THF (6 mL) was deoxygenated (3 × freeze-pumpthaw cycles under argon). The mixture was sparged with N2(g) for 10 min followed by the addition of propylene oxide (0.5 mL) and sparging with N2(g) for a further 5 min. The reaction mixture was then irradiated by 4 × 9W 254 nm bulbs for 3 h, while being sparged with N2(g). A saturated aqueous solution of Na₂S₂O₃ (2 mL) was added and the resulting biphasic mixture was stirred for 2 min. The reaction mixture was diluted with CHCl3 (5 mL) and the organic layer was separated and then washed with a saturated aqueous solution of Na₂S₂O₃ (2 × 5 mL) and brine (5 mL), before being dried over MgSO4, filtered and evaporated to dryness to give a dark solid. The crude solid was purified by column chromatography (Teledyne Isco CombiFlash Rf+ system, 12 g SiO2, hexanes-CH2Cl2, 0 -30% gradient elution). The title compound was isolated as a colorless S14 crystalline solid (54 mg, 90 μmol, >99% and iodine (13 mg, 51 µmol, 1.1 equiv) were added to a 10 mL quartz tube, which was fitted with a septum and sparged wit h N2(g). Anhydrous THF (8 mL) was degassed through 3 × freezepump-thaw cycles and added. The mixture was sparged with N2(g) for 10 min followed by the addition of propylene oxide (1 mL) and sparging with N2(g) for a further 5 min. The reaction mixture was then irradiated by 4 × 9W 254 nm bulbs for 2 h, while being sparged with N2(g). A saturated aqueous solution of Na₂S₂O₃ (5 mL) was added and the resulting biphasic mixture was stirred for 2 min. The reaction mixture was diluted with CHCl3 (5 mL) and the layers were separated.
The organic layer was washed with a saturated aqueous solution of Na₂S₂O₃ (2 × 10 mL) and brine (10 mL), before being dried over MgSO4, filtered and evaporated to dryness to give a yellow solid.

Variable-Temperature (VT) NMR Spectroscopy
We performed VT NMR measurements to determine the energy barrier to rotation of phenyl rings in sym-phenPh5C7H. As we were not able to unambiguously assign all 13 C resonances based on 2D NMR spectra, we have assigned ( Figure S25) the phenyl group that experiences the highest energy barrier to rotation as ring c by analogy to our investigation of Ph7C7H reported previously. 3 VT 13 C NMR spectra were acquired to facilitate analysis using a two-spin system model in the WinDNMR 4 software package. In order to obtain a solution of sym-phenPh5C7H with a low freezing point and of sufficiently high concentration (20 mg in 0.8 mL) for 13 C NMR analysis, a mixture of CS2 (0.7 mL) with CD2Cl2 (0.1 mL) was used as solvent. A series of spectra ranging from 24 to -98 °C were recorded. Figure S26. Partial 13 C VT NMR spectra of sym-phenPh5C7H. Peaks corresponding to the carbon pairs A/A′ and X/X′ are observed as individual, averaged signals in the fast exchange regime, but appear as distinct signals at low temperature in the slow exchange regime. Only three of the four signals can be distinguished at low temperature on account of overlapping signals; however, only one pair is needed to perform lineshape analysis.

S44
Analysis of the 13 C spectra reveals that within the temperature range studied, signals corresponding to ring c carbons broaden, merge into the baseline, and then re-emerge as four separate peaks as the temperature is decreased. There are two pairs of resonances in slow exchange below -60 °C. Only three resonances can be observed as distinct peaks as the fourth overlaps with other signals. In order to determine which of the three re-emerged peaks correspond to a pair of exchanging sites, a HSQC experiment was performed at -98 °C ( Figure S27). Both 1 H NMR and 13 C NMR spectra had been recorded from 24-98 °C. Analysis of the HSQC spectrum reveals that the three 13 C signals that have re-emerged from the baseline correlate with two doublets and a signal that appears as an apparent triplet in the 1 H NMR.
The signals correlating to the two doublets must be the pair of exchanging nuclear environments.

S45
The doublet at 6.63 ppm looks broad, but this is due to peaks under the doublet and this has been confirmed by integration. Thus, we can assign ( Figure S28) the labels A+A′ to these resonances.
We selected eight spectra for further analysis, choosing temperatures close to the transition of ring c resonances from fast to the slow exchange regimes. Lineshape analysis was performed to derive rate constants by comparison to model spectra produced using WinDNMR. Figure S28. 13 C VT NMR spectra of sym-phenPh5C7H used for lineshape analysis. The dashed blue lines illustrate the resolution of a single peaks into a pair of peaks as temperature decreases.
Activation energy barriers for the rotation of ring c were calculated for each of the eight temperatures using equation S1:

Equation S1. A variation of the Eyring equation.
Where kr is the measured rate constant, kB is the Boltzmann constant, T is temperature, h is Planck's constant, and R is the ideal gas constant. A line was fitted ( Figure S28) to a plot of ΔG ‡ vs T. The slope of the line corresponds to -ΔS ‡ and the y-axis intercept to ΔH ‡ . The entropy of activation ΔS ‡ was calculated to be -27.9 J·mol -1 ·K -1 and enthalpy of activation ΔH ‡ was calculated to be 36.2 kJ·mol -1 . We also performed VT NMR measurements to determine the energy barrier to rotation of phenyl rings in Ph7C7H-O, using the same approach. Based on 2D NMR, we were able to assign ring d ( Figure S30) as the phenyl group that experiences the highest energy barrier to rotation. A 20 mg

T / K
S47 sample of Ph7C7H-O was dissolved in a mixture of CS2 (0.7 mL) with CD2Cl2 (0.1 mL) as the NMR solvent and a series of 13 C NMR spectra ranging from 24-87 °C were recorded. Figure S30. Partial 13 C VT NMR spectra of Ph7C7H-O. Peaks corresponding to the carbon pairs A/A′ and X/X′ are observed as individual, averaged signals in the fast exchange regime, but appear as distinct signals at low temperature in the slow exchange regime. Only three of the four signals can be distinguished at low temperature on account of overlapping signals; however, only one pair is needed to perform lineshape analysis.
Analysis of the 13 C spectra reveals that, within the temperature range studied, signals corresponding to ring d carbons broaden, merge into the baseline, and then re-emerge as four separate peaks as the temperature is decreased made up of two pairs of resonances in slow exchange. Only three resonances are observed as distinct peaks because the fourth overlaps with other signals. In order to determine which of the three re-emerged peaks are a pair a HSQC spectrum was acquired ( Figure   S30) at -87 °C. Both 1 H NMR and 13 C NMR spectra had been recorded from 24 °C to -87 °C. By matching the J coupling patterns of the 1 H signals following the approach described above for sym-phenPh5C7H, we can assign ( Figure S32) the 13 C signals of Ph7C7H-O corresponding to X and Xʹ. We selected five 13 C NMR spectra for further analysis, choosing temperatures close to the transition of ring d resonances from fast to the slow exchange regimes. As above, line shape analysis was performed to derive rate constants by comparison to model spectra produced using WinDNMR and activation energy barriers were calculated using equation S1.  A line was fitted ( Figure S29) to a plot of ΔG ‡ vs T. The entropy of activation ΔS ‡ was measured to be +27.0 J mol -1 K -1 and enthalpy of activation ΔH ‡ was measured as 48.5 kJ mol -1 .

X-Ray Crystallographic Analysis
Analysis of all crystal structures and their packing are shown ( Figure S34-47) with the crystal system, space group, unit cell parameters, bond lengths, bond angles and dihedral angles reported below. Direct comparisons of important bond lengths and centroid-centroid distance between the phenyl ring bound to the tertiary sp 3 -centre and the opposite phenyl ring are outlined in Table S3 below. Notably, the C5-C6 distance is 1.47-1.49 Å for the triene compounds but 1.50-1.51 Å for the epoxides. This increase of 2-4 pm is characteristic of the increased single-bond character in the nonconjugated epoxides. Figure S34. Solid-state structure of Ph7C7H viewed side-on to the seven-membered ring showing the centroid-centroid distance on the two rings associated with through-space dimers.

Ph3C3·HCl2
Crystals of Ph3C3·HCl2 suitable for X-ray diffraction were grown by slow cooling of a saturated MeCN solution.

sym-phenPh5C7H
Crystals of sym-phenPh5C7H suitable for X-ray diffraction were grown by slow cooling of a saturated MeCN solution of pure sym-phenPh5C7H.

asym-phenPh5C7H
A single crystal of asym-phenPh5C7H suitable for X-ray diffraction were grown by slow cooling of a saturated MeCN solution of a 2:1 mixture of asym-phenPh5C7H and sym-phenPh5C7H. Figure S40. Solid-state structure of asym-phenPh5C7H viewed (a) side-on to the cycloheptatriene and (b) from above the cycloheptatriene. Selected atoms are labelled numerically, the planes of the carbocyclic rings are labelled by italicized uppercase letters. As central ring A is puckered, individual planes α n are defined by a carbon vertex n and its two nearest neighbors within the ring, e.g., A 5 is the plane defined by atoms 4, 5, and 6.

sym-phenPh5C7H-O
A single crystal of sym-phenPh5C7H-O suitable for X-ray diffraction were grown by slow cooling of a saturated MeCN solution.

asym-phenPh5C7H-O
A single crystal of asym-phenPh5C7H-O suitable for X-ray diffraction were grown by slow cooling of a saturated MeCN solution.

sym-phen2Ph3C7H-O
A single crystal of sym-phenPh5C7H-O suitable for X-ray diffraction were grown by slow cooling of a saturated MeCN solution.

UV-Vis Absorption Spectra
All spectra show ( Figure S48)  where I is the measured intensity, c is the concentration of sample, and l is path length of the cuvette.

Variable-Temperature Fluorescence
Samples for VT fluorescence were prepared using 200 M THF stock solutions. The desired quantity was measured into a vial, evaporated to dryness, then the solid residue diluted to 20 M and 2 M

Phenanthrene
Fluorescence spectra were acquired for molecular phenanthrene. A concentration study (Figure S57a Table S4. Lifetimes from the exponential fits for the days of molecular rotors in 2-MeTHF solutions. Assosiated IRF set labeled, after rotor name in each case. IRF data outlined in Table S5. a Single exponentional fit y = y0 + A1*exp(-(x-x0)/t1) used.

Computational Details
Ground state geometry optimization of sym-phen2Ph3C7H-O was conducted in vacuo employing with SCS-MP2 method (spin-component-scaled Møller-Plesset perturbation theory of second order) 13 along with an SVP basis set. The S1 excited-state geometry optimization was performed with SCS-ADC(2) method (spin-component-scaled adiabatic diagrammatic construction up to second order) 14 with an SVP basis set. All electronic transitions were calculated with SCS-ADC (2) using a TZVP basis set, on the S0 or S1 geometries obtained with SCS-MP2/SVP or SCS-ADC(2)/SVP. All calculations were performed with the Turbomole 7.4.1 program package. 15 To further investigate the shape of the S1 emission band, we included non-Condon effects in our calculations by using the nuclear ensemble approach. 16 We optimized the S1 geometry of sym-phen2Ph3C7H-O using linear-response time-dependent density functional theory (LR-TDDFT) within the Tamm-Dancoff approximation, the ωB97X-D functional, and a 6-31G* basis set. The minimum-energy geometry obtained on S1 shows a similar electronic character as the one calculated with SCS-ADC (2). At the LR-TDDFT/TDA S1 geometry, LR-TDDFT/TDA/ωB97X-D/6-31G* gives a vertical S1/S0 energy of 3.98 eV, while SCS-ADC(2)/TZVP (on the same geometry) indicates a transition energy of 3.64 eV. Frequencies were obtained at the same level of theory and used to calculate an approximated Wigner distribution for the lowest vibrational states of the S1 electronic state. 100 geometries were sampled from this Wigner distribution. For each of these geometries, S1/S0 emission energy and oscillator strength were calculated at the LR-TDDFT/TDA/ωB97X-D/6-31G* level of theory. An emission spectrum accounting for non-Condon effects is then obtained by grouping all the transitions, each broadened by a Gaussian with a width of 0.1 eV. The obtained emission spectrum ( Figure S63) reproduces the width of the