Luminogens for Aggregation-Induced Emission via Titanium-Mediated Double Nucleophilic Addition to 2,5-Dialkynylpyridines: Formation and Transformation of the Emitting Aggregates.

Abstract New luminogens for aggregation‐induced emission (AIE), which are characterized by a branched cross‐conjugated 2,6‐bis(1,2,2‐triarylvinyl)pyridine motif, have been synthesized exploiting the one‐pot Ti‐mediated tetraarylation of 2,6‐bis(arylethynyl)pyridines. Thin layer solid‐state emitters were prepared by spin‐coating of the luminogens, while AIE‐colloidal dispersions were investigated in terms of optical density and scattering behaviour. This has given insight into particle size distributions, time evolution of the aggregation and the influence of different functionalization patterns on the luminescence of molecular aggregates. In particular, a combination of extinction spectroscopy and dynamic light scattering is being proposed as a powerful method for investigating the dynamic aggregation process in AIE‐type colloids.

2,6-bis((4-fluoro-3-methylphenyl)ethynyl)pyridine, pale yellow solid, 1.2 g, 3.5 mmol, Yield: 56%. 2,6dibromopyridine (1.5 g, 6.3 mmol); 4-ethynyl-1-fluoro-2-methylbenzene (1.9 g, 1.9 mL); CuI (0.10 g); PPh3 (0.13 g); PdCl2(PPh3)2 (0.18 g); NEt3 (13 mL). Room temperature, 5 days. 1      Color key for the time evolution: green (t 0 ) to red (t=4 days). Data are not significant for the 50% v/v H 2 O mixture due to insufficient signal-to-noise ratio (A). In the 70% v/v H 2 O mixture particles grow over the time approaching an average particle size as big as 530 nm; precipitation then occurs, leaving in suspension only a minor fraction of smaller particles (160 nm ca., red curves). In the 90% v/v H 2 O mixture the average particle size grows moderately from 130 nm to 160 nm over time resulting in an overall stable colloid. The dispersions were kept at rest inside quartz cuvettes under light exclusion until the measurements were completed. Bottom right: 1 after addition of glycerol to the final ratio DMSO/glycerol = 1:8 mixture.   As discussed in the main text, the scattering background can be determined as the difference between the extinction spectra (Figures S4, S10-14) and the absorbance spectra shown in this Figure     PL and Extinction Evolution of 5

X-ray crystal structure determinations
Crystal data and details of the structure determinations are compiled in Table S1. Full shells of intensity data were collected at low temperature with a Bruker AXS Smart 1000 CCD diffractometer (Mo-K radiation, sealed X-ray tube, graphite monochromator, compound 4·C6D6) or an Agilent Technologies Supernova-E CCD diffractometer (Mo-or Cu-Ka radiation, microfocus X-ray tube, multilayer mirror optics, all other compounds).
Detector frames (typically -, occasionally -scans, scan width 0.4...1°) were integrated by profile fitting. S1 Data were corrected for air and detector absorption, Lorentz and polarization effects S2,S3 and scaled essentially by application of appropriate spherical harmonic functions. S4-S6 Absorption by the crystal was treated with a semiempirical multiscan method (as part of the scaling process), and augmented by a spherical correction, S5,S6 or numerically (Gaussian grid). S6,S7 For datasets collected with the microfocus tubes an illumination correction was performed as part of the numerical absorption correction. S6 The structures were solved by ab initio dual space methods (compound 3·n-pentane: SHELXD, S8 compound 4·C6D6 pentane: VLD procedure S9 ) or by the charge flip procedure S10 and refined by full-matrix least squares methods based on F 2 against all unique reflections. S11 All non-hydrogen atoms were given anisotropic displacement parameters. Hydrogen atoms were input at calculated positions and refined with a riding model. S12 Split atom models were used to refine disordered groups and/or solvent molecules. When found necessary, suitable geometry and adp restraints were applied. S12,S13 Due to severe disorder and fractional occupancy, electron density attributed to solvent of crystallization was removed from the structures of 3 (n-pentane) and 5 (ethyl acetate) with the BYPASS procedure, S14 as implemented in PLATON (squeeze/hybrid). S15 Partial structure factors from the solvent masks were included in the refinement as separate contributions to Fcalc.
To establish comparabiltity with the other structures, the previously published S16 structure of 1 was re-refined with hydrogen atoms at calculated positions (refined riding). CCDC 1965537 and 1963833 -1963835

Wavelength-dependent scattering
While light scattering is typically investigated as function of scattering angle, S17 its wavelength dependence is much less explored. However, this wavelength dependence is of interest, as it contributes to the extinction spectra and is thus readily accessible experimentally. The wavelength dependent scattering can be described by the Mie theory. S18 The mathematical framework of this theory is rather complex so that various approximations have been derived. Probably the most well-known approximation is the Rayleigh approximation S19 , which describes the scattering coefficient as function of wavelength (σ(λ)) according to equation 1: (1) where D is the particle diameter,  is the sphere density and n and n0 are the refractive indices inside and outside the sphere (the scattering coefficient is found by dividing the Rayleigh scattering cross-section by the sphere mass). Importantly, equation 1 strictly only holds for spheres with diameters D</10. When considering molecular solutions with D<0.5 nm, the scattering coefficient over the entire spectral region is very small (<10 -5 L/g -1 m -1 with =1000 kg/m 3 , n=1.5, n0=1.3) and its contribution to extinction spectra can be neglected. However, this is not the case for colloids.
For larger particles with D > λ/π, the van Hulst approximation S20 applies which can also be expressed in the form of the scattering efficient (derivation see S21 ) according to equation 2: Both approximations show that the wavelength dependent scattering coefficient can be described as a powerlaw with exponents of -4 in the case of particles with diameters D</10 and -2 for D > λ/π. Accordingly, taking a wavelength of 800 nm, we would expect a scattering exponent, -m, of 4 for spheres < 80 nm and 2 for > 250 nm.
Importantly, many colloids, such as these investigated here, have intermediate sizes where neither approximation is applicable. As elaborated recently for platelet-like colloids S21 , this intermediate particle size regime is characterized by exponents 2<-m<4 with a strong dependence of the exponent on the nanoparticle projected area. As such, the scattering exponent which is accessible from extinction spectra ( Figure S5) can be used as a measure for the particle size of the colloids.