Theobromine and direct arylation: a sustainable and scalable solution to minimize aggregation caused quenching

Aggregation caused quenching (ACQ) of fluorescence is commonly observed in organic luminophores. 1 While there are a number of reasons why ACQ can take place, in the case of organic luminophores with extended p conjugation, such as pyrene 1b and perylene diimide, 2 the molecules can interact with each other via p orbital overlap and form non-fluorescent excimers. Aggregation causes non-radiative thermal decay to occur, dissipating energy into heat instead of light emission. This limits their efficiency in luminescent applications, such as organic light emitting diodes (OLEDs) and organic lasers (OLs). 1 To date, numerous efforts have been invested into suppressing ACQ in organic luminophores including blending, 3 co-crystallization, 4 and covalently attaching functional moieties. 5 In this paper, we focus on advancing the chemical methods, whose products can be solution-processed, which is crucial for future scale-up and low-cost device fabrication. 6 Huang et al. suppressed ACQ by attaching bulky oligo-alkylfluorene moieties onto luminophores (Fig. 1). 7 Once installed, the fluorene modifiers were able to act as spacers, preventing luminophore aggregation and thus improving the solid-state photoluminescence quantum yields (PLQYs). Moreover, because the oligo-alkylfluorenes are conjugated, charge mobilities were maintained after modification, in sharp contrast to insulating spacers. 8 Tang et al. discovered that ACQ can be suppressed by attaching tetraphenylethene (TPE) onto luminophores triggering the mechanism of aggregation induced emission (AIE). 1b Impressive results have been obtained by applying these types of materials to OLED and OL devices.

However, these methods bring up environmental concerns: they heavily rely on Suzuki coupling, which uses air-sensitive and explosive organolithium compounds to prepare boronic ester precursors. 9 Should these materials be prepared at a commercial scale, a method avoiding organolithium would substantially diminish workplace safety risks. Moreover, organolithium decreases the functional group tolerance in the synthesis, as it is strongly nucleophilic.

Direct arylation offers a green and atom-efficient method to form C-C bonds between aromatic building blocks. This direct formation of C-C bond allows us to bypass the extra step of precursor synthesis, where organolithiums are usually involved. 10 Therefore, the risks of fire and explosion are reduced, while cost and productivity are optimized. Moreover, direct arylation is compatible with large scale production. 11 Major pharmaceutical companies, such as Merck 12 and Pfizer, 13 are applying it into commercial production of medicines; flow chemistry has been successfully implemented enhancing productivity further. 14 Additionally, substrate compatibility is improved because of the absence of highly reactive intermediates. However, because the C-H bond on benzene has poor reactivity under the concerted metalation-deprotonation (CMD) mechanism, 15 it is not feasible to utilize direct arylation to attach oligofluorenes or TPE onto luminophores to minimize ACQ. A new moiety needs to be identified to pair with direct arylation to develop a green method to suppress ACQ.

In this paper, alkylated theobromine is introduced onto luminophores via direct arylation to suppress ACQ. The mild reaction conditions allow for wide functional group tolerance. In this strategy, theobromine has the following advantages: 1) it is a natural product originally from cacao plants that is now produced in industrial scale and thus readily available and inexpensive (see Table S1 for price comparison for starting materials on each route); 2) the imidazole C-H bond is highly reactive for direct arylation; 16 3) the N-methyl group on imidazole will repulse the luminophore and induce a large dihedral angle, creating steric hinderance preventing the luminophores from interacting with each other; 4) the lactam group is applied widely in the design of organic semiconductors with high electronic performance and one can easily tune the processability of the final product by introducing different solubilizing chains onto it. 17 Pyrene was chosen as the luminophore to test the effectiveness of this method. Pyrene is highly fluorescent in solution but completely quenched in the solid state, 1b, 18 making it a good candidate to examine our approach. Three theobromine-pyrene derivatives, with differing theobromine:pyrene ratios, were successfully synthesized and characterized.

The synthetic route to PT1, PT2 and PT4 is straightforward, beginning with the N-alkylation of theobromine, as shown in Scheme 1. The relatively long octyl chains were introduced to increase the hydrophobicity of the final products, improving solubility in common solvents and enabling solution processing. The key intermediate Theo8 was obtained in near quantitative yield. Subsequently, Theo8 and pyrene was cross-coupled by direct arylation, forming PT1, PT2, and PT4 with increasing equivalents of Theo8. Notably, these materials are made in just two steps from commercially available starting materials. In contrast, it would take two extra steps to synthesize these materials via Suzuki coupling: converting brominated pyrenes to pyrene boronic esters and brominating Theo8. Moreover, as for products with high functionalities such as PT4, their overall yields would be significantly reduced with each step added to the overall route due to relatively higher incomplete conversion of each synthetic step.

As hypothesized, these three Please do not adjust margins Please do not adjust margins compounds show high PLQYs in both solid state and solution.

Notably PT1 shows the highest PLQYs of all, approaching 100% as thin films. Increasing the ratio of Theo8 leads to a decrease of PLQYs as thin films, as shown in Table 1, consistent with the photoluminescence lifetime measurements in Fig. S13.

Increasing the amount of Theo8 in the molecules increases the size of their conjugated systems. This could lead to the increase in intermolecular π overlap in solid state, facilitating ACQ and thus lowering solid-state PLQYs. 1b Notably, we spincoated thin films from PT1, PT2 and PT4 that have been stored under ambient conditions for over 1 year. The resultant films showed unvaried PLQYs compared to what they were 1 year ago (see Fig. S14).

Molecular packing has significant impacts on solid-state PLQYs. 19 We thus investigated how the molecules are arranged through single crystal X-ray diffraction (SCXRD) (Fig. 2). The dihedral angles between theobromine and pyrene are relatively large: 63°, 80° and 47° for PT1, PT2 and respectively. These large dihedral angles are caused by the steric repulsion between pyrene and the adjacent methyl group on theobromine. This steric repulsion forces these molecules to adopt a twisted conformation prevents the p-p interaction between pyrene cores. Fig. 2 shows the crystalline molecular arrangement of PT1, PT2 and PT4. While each compound has different packing patterns, in all cases theobromine drives selfassembly, due to strong polar interactions between the lactam groups. Meanwhile, the pyrenes and alkyl chains are arranged according to the positioning of the theobromine moieties. As a result, adjacent pyrenes are separated from each other, effectively suppressing pyrene's ACQ in the solid state. To further confirm the efficacy of theobromine, we replaced the octyl on PT1 with methyl and synthesized PC (Scheme S1). PC and PT1 both demonstrated PLQYs around 95% as thin film, justifying theobromine is the origin for ACQ suppression.

As shown in Table 1, increasing the theobromine ratios leads to spectral redshifts in absorbance and emission, which implies that theobromine is conjugated to the luminophore it attaches to. This presents an apparent contradiction with the large dihedral angles observed by SCXRD, which would limit any such conjugation. It is proven feasible and widely applied using gas phase density function theory (DFT) simulations as close approximations for spincoated films considering their amorphous nature. 20 Molecular orbitals (MOs) were simulated based on B3LYP functional and 6-31g (d) basis set to investigate the apparent conjugation between theobromine and pyrene despite the large dihedral angles. Notably, there is an unusual distribution in the ground state (S0) MOs of both PT1 and PT2, as circled in Fig. 3. Despite the large dihedral angle between the pyrene and theobromine moieties, 63° for PT1 and 80° for PT2, the orbital symmetries allow for MO overlap between the two moieties, intertwining around the single bond linkage. On the other hand, with a relatively smaller dihedral angle of 47°, the MOs of PT4 extend through the molecule as commonly seen in other planar conjugated structures.

In all three molecules, the MOs are spread out among the whole molecule in S0, while in the first singlet excited state (S1) As shown in Table 1, these pyrene-theobromine compounds possess high solid-state PLQYs and large Stokes shifts, which make them potential candidates for OL application. 23 We therefore studied their waveguiding properties in the solid state. When photopumping of the film is intense enough, spontaneously emitted photons are waveguided through the gain medium and amplified by stimulated emission, resulting in amplified spontaneous (ASE). To measure their ASE thresholds, a 375 nm laser beam was focused through a cylindrical lens into a stripe and used to photoexcited the thin film at a normal angle. Emission was collected from the edge of the film (see experiment set up in Fig. S5). Fig. 4a shows fluence dependent emission spectra of a 90 nm thin film of PT2 on a glass substrate. Its output spectrum significantly narrowed as the excitation fluence was raised above 20 μJ/cm 2 . This spectral narrowing was accompanied by a sudden increase in output intensity vs. excitation fluence curve (Fig. 4b). Taken together, Fig. 4a & 4b indicate that the ASE threshold for PT2 is 20 μJ/cm 2 . In contrast, we did not observe ASE in PT1 and PT4 films of ~50 nm. Although emission was waveguided through the excitation area, PL emission intensity increased with pump intensity only in a linear fashion over several orders of magnitude (Figure S6 & S7). We were not able to obtain smooth films of 90 nm from PT1 and PT4 because of the poor film-forming ability of small molecules in general. 24 It is well-known that ASE threshold is dependent on film thickness, 25 and Bradley et al. calculated that a film thickness of 40 -70 nm, depending on the compound, is prerequisite for ASE. 26 We did not observe ASE in the 50 nm films of PT1 and PT4, likely because their thickness cut-offs are located at the upper end of 40 -70 nm range. We next investigated the theoretical thickness for lowest ASE threshold in PT2 films based on the method reported by Anni et al., 25a using optical constants determined via a previously reported transfer matrix method (Fig. S8). 27 Briefly, we assume that the ASE threshold thickness dependence is governed by two factors: (1) the spatial overlap between the pump electric field and the guided mode (pump mode overlap, PMO in Fig. 4c) and ( 2) the fraction of the guided mode that exists within the organic film (mode confinement, MC in Fig. 4c). We simulated the waveguiding of 0-1 emission, resolved from low-temperature PL measurements (Fig. S9) in a SiO2-PT film-air slab waveguide. Fig. 4c shows MC and PMO in the waveguide as a function of film thickness. The ASE threshold, which is inversely proportional to the product of PMO and MC, 2525a reached the minimum around 100 nm. This is consistent with our thickness dependent ASE measurement, where ASE thresholds of PT2 decreased from 50, 30 to 20 μJ/cm 2 as the film thickness increased from 50, 60 to 90 nm. No ASE was observed in PT2 film of 30 nm, in agreement with the calculation by Bradley et al. 26