8,8′-(Benzo[c][1,2,5]thiadiazole-4,7-diyl)bis(quinolin-4(1H)-one): a twisted photosensitizer with AIE properties

A new benzothiadiazole (BTZ) luminogen is prepared via the Suzuki–Miyaura Pd-catalysed C–C cross-coupling of 8-iodoquinolin-4(1H)-one and a BTZ bispinacol boronic ester. The rapid reaction (5 min) affords the air-, thermo-, and photostable product in 97% yield as a yellow precipitate that can be isolated by filtration. The luminogen exhibits aggregated-induced emission (AIE) properties, which are attributed to its photoactive BTZ core and nonplanar geometry. It also behaves as a molecular heterogeneous photosensitizer for the production of singlet oxygen under continuous flow conditions.

S3 S1 Experimental Section S1.1 General methods and materials All chemicals were commercially available except those whose synthesis is herein described. Anhydrous MgSO4 was used for drying organic extracts and all volatiles were removed under reduced pressure. All reaction mixtures and column eluents were monitored by TLC using commercial glass backed thin layer chromatography (TLC) plates (Merck Kieselgel 60 F254). The plates were observed under UV light at 254 and 365 nm. The technique of flash chromatography was used throughout for all non-TLC scale chromatographic separations using Merck Silica Gel 60 (less than 0.063 mm). Melting points were determined using a Stuart SMP10 digital melting point apparatus. Small scale (μL) liquid handling measurements were made using variable volume (1.00-5000.00 μL) single channel Gilson PIPETMAN precision micropipettes. Solvents used for recrystallisation are indicated after the melting point. IR spectra were recorded on a Thermo Scientific Nicolet iS5 FTIR spectrometer with iD5 ATR accessory and broad, strong, medium and weak peaks are represented by b, s, m and w, respectively. 1 H and 13 C NMR spectra were recorded on a Bruker AVANCE III HD machine (at 400 and 100 MHz, respectively). An AVANCE III 300 MHz NMR Spectrometer was also used for reaction monitoring. The benchtop NMR that was used for flow reaction monitoring was a NANALYSIS corp. NMRReady 60 Benchtop 1 H NMR. Chemical shifts (δ) are expressed in ppm and coupling constants J are given in Hz. Data are represented as follows: chemical shift, multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet and/or multiple resonances, br s = broad singlet). Deuterated solvents were used for homonuclear lock and the signals are referenced to the deuterated solvent peaks. For the acquisition of mass spectra the samples were prepared as detailed below and analysed by positive ion nanoelectrospray (nES) using a Thermo Scientific™ LTQ Orbitrap XL™ ETD Hybrid Ion Trap-Orbitrap Mass Spectrometer. The absolute PLQY of of each sample was measured using a calibrated spectrofluorometer (Edinburgh Instruments, FLS920), equipped with an integrating sphere (Jobin-Yvon). Flow reactions were carried out with a commercial E-series Photochem reactor by Vapourtec Ltd, with two V-3 peristaltic pumps and using an LED module emitting at 390-440 nm with 7.44 W light output (EPILED, Future Eden Ltd.). For the collection of LC-MS data a Shimadzu LC-2040C 3D Plus instrument was used. Iodophenyl)amino]methylene}-2,2-dimethyl-1,3-dioxane-4,6-dione 1 To a stirred solution of 2-iodoaniline (3.00 g, 13.6 mmol), in anhydrous acetonitrile (50 mL) at ca. 20 °C was added in one portion Meldrum's acid (2.15 g, 15.0 mmol), followed by triethyl orthoformate (2.41 g, 16.3 mmol). The resulting mixture was then heated to ca. 95 °C (reflux) for 2 h until the 2-iodoaniline was consumed (by TLC). The resulting mixture was then allowed to cool to ca. 20 °C, evaporated to dryness under vacuum, and the residue recrystallised by hot ethanol (EtOH) to afford the title compound 1 (

S1.5 Preparation of 5H2 2+ from 5
To a 5 mL glass vial was added compound 5 (20.0 mg, 0.047 mmol) in 1 mL of TFA. The reaction mixture was then allowed to stir at ca. 20 °C for 10 minutes at which point all of the solid was dissolved. The contents of the vial were then transferred into a 50 mL conical flask containing 10 mL of distilled water, leading to the formation of a powdery dark yellow precipitate, which was isolated by filtration, washed thoroughly using cold distilled water (2 × 5 mL) and acetone (2 × 5 mL), and then dried under a stream of air, to afford the title compound 5H2 2+ (25.0 mg, 82%) as a fine yellow powder. Crystals of 5H2 2+ were grown by dissolving a small amount (~5 mg) in a 1:1 mixture (1 mL

S1.6 Heterogeneous flow mediated photocatalytic generation of singlet oxygen with in-line NMR analysis
The procedure for the heterogeneous flow photosensitisation of 5 to produce singlet oxygen was adapted from Thomson et. al, 2 and was carried out as follows: To a transparent borosilicate glass column fixed bed reactor (6.6 mm ID, 100 mm length) was added a mixture of photocatalyst 5 (1.1 mg, 0.0025 mmol, 1 mol%) and 500 mg chromatography grade (60A, 40 -63 μm particle size) SiO2. The dry-packed mixture was then attached to the Vapourtec flow machine and it was primed with CDCl3 (10 mL) for 5 min by flowing solvent through the column at 5 mL/min. To a dry 20 mL round bottom flask loaded with a magnetic stirrer bar, was added α-terpinene (34.1 mg, 40.1 μL, 0.25 mmol) and CDCl3 (5 mL) before sealing with a septum. The flask was connected to the Vapourtec flow machine and then covered to prevent photosensitisation events from ambient light. The reaction solution was pumped at 5 mL/min using peristaltic pumps (Vapourtec V-3) to a Tjunction where it was mixed with O2 gas from a cylinder which was being pumped concurrently at 5 mL/min by a second peristaltic pump. The heterogeneous liquid-gas mixture formed a slug flow at the T-junction and was pumped through the fixed bed column reactor placed in a reflective housing, irradiated by an LED array (390 -440 nm, 7.44 W). After exiting the photoreactor, the slug flow was then passed through a back pressure regulator which kept the pressure at 2 bar before passing through a 60 MHz Nanalysis-60e benchtop 1 H NMR instrument which monitored the reaction progression in real time. Following that, the slug flow was returned to the initial reaction flask to be continuously cycled for 25 min. The catalyst bed was permitted to form a 'mixed bed' regime (hybrid between a packed bed and fluidised bed) to prevent pressure build up.
After the reaction was complete, as indicated by the bench-top NMR, the resulting solution was evaporated in vacuo to afford an oily residue. 1 H NMR (300 MHz) analysis of the residue revealed the complete disappearance of the α-terpinene alkene protons (δΗ 5.6-5.7 ppm) and the appearance of the corresponding ascaridole alkene protons (δΗ 6.4-6.6 ppm), confirming the bench-top NMR data and the complete transformation of the starting material.
For the cycling experiments, the loaded packed bed reactor was left in place, 15 mL of fresh CDCl3 was flowed through the aforementioned set-up and it was then discarded. A new cycle could then be initiated by following the protocol above.

S1.7 Photoluminescence quantum yield (PLQY) calculations
The FLS920 spectrofluorometer was equipped with an extended red-sensitive single-photon counting photon multiplier (Hamamatsu, R2658P, 200-1010 nm), which was used to measure all spectra. A Xenon lamp, centered to 320 nm, was used as the excitation source.
The PLQY is defined by equation 1. 3 It is determined by dividing the number of photons emitted (Lsample), by the number of photons absorbed. The later is calculated by measuring the intensity of excitation light (Ereference) and subtracting it from the intensity of excitation light not absorbed by the sample (Esample). Comparison of the calculated structures with the structure determined from the TFA-crystal of 5 indicates that the match in conformation and geometry is high. It appears structurally there is little difference in the protonation of 5 on distances or angles. This is highlighted by the low RMSD and maximum deviation numbers when comparing the molecular geometries (Table S1 and Figure S2). Table S1. Comparison of the geometry parameters from the X-ray structure and the optimised structures in the gas phase at the RB3LYP/6-311G(d,p) level of theory.

S2.3 Computational data
The geometries of the molecule 5 in neutral and protonated form were fully optimized at the DFT RB3LYP/6-311G(d,p) level of theory and analytical second derivatives were computed using vibrational analysis to confirm each stationary point to be a minimum by yielding zero imaginary frequencies. TD-DFT calculations were performed also at the RB3LYP/6-311G(d,p) level of theory to obtain the vertical excitation energies. All the above computations were performed using the Gaussian 03 suite of programs. 5 X-ray crystallography showed that molecule 5 exists in the solid state as the phenolic prototautomeric form 5'' cocrystallised with molecules of TFA that participate via various H-bonding interactions ( Figure S5). In the absence of TFA or in the gas or solution phase the structure of compound 5 can differ. As two possible symmetrical prototautomers 5' and 5'' exist, and one unsymmetrical, 5''', all were optimised at the RB3LYP/6-311G(d,p) level of theory in gas phase. The quinolone tautomer 5' was computed to be 42.3 kJ·mol -1 more stable than prototautomer 5'' and 14.6 kJ·mol -1 more stable than 5'''.

S12
The main vertical excitations for molecule 5'' were calculated using TD-DFT at the B3LYP/6-311G(d,p) level of theory (Table S2), which indicated that the longest wavelength absorption corresponds to a HOMO → LUMO transition; observed as a shoulder in the theoretical and the experimental spectrum, blue shifted in the latter by ca. 51 nm). The two other absorptions observed theoretically at 322 and 312 nm, have a main contribution from a HOMO-2 → LUMO and a HOMO → LUMO+1 transition, respectively.  Figure S7. The molecular orbitals associated with the vertical excitations responsible for the absorption profile of tautomer 5'' revealed that the HOMO → LUMO transition has significant charge transfer character from the quinolinol moieties to the benzothiadiazole core.

HOMO → LUMO+1 (85%)
As supported by X-ray crystallography, in TFA we expect the quinolinols to be in the protonated form, and as such the dication in the presence and absence of coordinated TFA molecules were studied computationally. In the absence of explicitly added TFA molecules in the cationic form the absorption is significantly red shifted owing to an intramolecular hydrogen bond between the NH of the S13 protonated quinolinol and the nitrogen of the thiadiazole ring. This leads to increased planarity of the system, and, as such, the absorption was shifted to longer wavelengths presumably owing to greater conjugation ( Figure S8).

Figure S8. Comparison of the experimental absorption spectrum for compound 5 and the theoretical absorption spectra for the neutral and dication species of 5.
The UV-vis absorption of the neutral molecule and the pyridinium form in the presence of coordinated anionic and neutral TFA molecules, however, is in good agreement with the experimental UV-vis absorption in TFA as solvent. The more TFA molecules incorporated into the calculation, the closer is the absorption to the experimental data in TFA.