Design of Conformationally Distorted Donor-Acceptor Dyads Showing Efficient Thermally Activated Delayed Fluorescence

A highly potent donor-acceptor biaryl TADF dye is accessible by a concise two-step sequence, employing twofold Ullmann arylation and a sequentially Pd-catalyzed Masuda borylation-Suzuki arylation (MBSA). Photophysical investigations show efficient TADF at ambient temperature due to the sterical hindrance between the donor and acceptor moieties. The photoluminescence quantum yield amounts to  PL = 80 % in toluene and 90 % in PMMA arising from prompt and delayed fluorescence with decay times of 21 ns and 30 µs, respectively. From an Arrhenius plot, the energy gap  E(S 1 -T 1 ) between the lowest excited singlet S 1 and triplet T 1 state was determined to 980 cm -1 (120 meV). A new procedure is proposed that allows us to estimate the intersystem crossing time to ≈ 10

could be overcome. Accordingly, efficient OLEDs could be developed. Meanwhile, they have reached a stage of commercial application from small smart phone to large tv displays. The emitters usually applied so far are based on phosphorescent materials. 1,2,3 Currently, iridium(III) complexes are the best emitter materials to achieve this task. 4,5 However, they are not well suited as stable blue-light emitters resulting in a low device stability. Moreover, these nobel metal complexes are rare and expensive. [1] Thus, intense research activities have been initiated, either towards the cheaper Cu(I) or Ag(I) complexes 4,6,7,8 or metal-free organic materials 9,10,11,12,13,14 , both classes employing the TADF (thermally activated delayed fluorescence) effect as a rationally designed singlet harvesting mechanism. 15 Employing TADF (or "E-type emission") 16 for exciton harvesting, almost 100 % internal OLED quantum yield can be obtained. 17 For achieving high TADF efficiency, a small energy gap between the excited singlet and the triplet state E(S1-T1) is a prerequisite for an effective thermal up-ISC (or reverse ISC, RISC) from T1 to S1. 6, 18,19 Twisted donor-acceptor systems are very promising lead structures with respect to small HOMO-LUMO overlap and thus, small E(S1-T1). 17,20,21,22 Based upon our program to develop sequential palladium-catalyzed one-pot methods for synthesizing functional organic chromophores, 23 we decided to use the Masuda borylation-Suzuki arylation (MBSA) one-pot process 24 as a key step to access predicted donor-acceptor biaryl structures. The MBSA sequence represents an efficacious, fast and economical way to synthesize functional bi(hetero)aryl directly from the halogenated precursors. Here, we communicate a concise synthesis of a specific donor-acceptor biaryl structure and prove the TADF concept by TD-DFT calculations and detailed photophysical investigations. The resulting biaryl molecule represents an efficient and bright green emitting TADF compound.
The structure of the title compound 1 (Scheme 1) was proposed by using TD-DFT calculations with special regard to the energy gap E(S1-T1) to achieve efficient TADF (see below). Therefore, a structure with an ortho-methyl group adjacent to the aryl-aryl-bond between the donor and acceptor parts was deduced, creating a twisting angle to obtain a reasonably small energy gap of the charge transfer (CT) states of singlet and triplet character.
Based upon transition metal catalyzed transformations, the target structure 1 was retrosynthetically analyzed by a cross-coupling of the bromotriarylamine 2 and 2-iodo terephthalicdinitrile 3 employing the MBSA sequence (Scheme 1). Bromotriarylamine 2 should be accessible by twofold Ullmann phenylation of the commercially available 4-bromo-3methylaniline 4. Scheme 1. Retrosynthetic analysis of the calculated twisted biaryl TADF chromophore 1 by MBSA and Ullmann coupling.
Starting from commercially available 4-bromo-3-methylaniline 4 as the corner stone, the Ullmann reaction at 135 °C with an excess of iodobenzene furnished the bromotriarylamine 2 in a yield of 77 % (Scheme 2). This bromotriarylamine was then subjected to the MBSA sequence by first performing the Pd-catalyzed borylation with pinacolyl borane at 80 °C in 1,4dioxane in the presence of triethylamine for 3 h (monitored by GC-MS). After complete conversion of the bromide 2 to the corresponding pinalcolyl boronate, in the same reaction vessel methanol was added to quench excessive pinacolyl borane and for the Suzuki arylation 2-iodo terephthalic dinitrile 3 and cesium carbonate was added and the reaction mixture was heated to 100 °C for 11 h to furnish after chromatography the donor-acceptor biaryl 1 in 50% yield.
Frequently, a theoretical approach based on DFT and TD-DFT calculations give a good indication, whether a specific molecule will be a TADF candidate. Therefore, we carried out TD-DFT calculations for a DFT optimized molecule of compound 1. The calculations were done at the B3LYP/6-31G(d,p) level of theory 25,26 for molecule 1 in a cavity of a toluene solvent and using the polarizable continuum model 27  Indeed, TD-DFT calculations indicate a relatively small energy separation of E( 1 CT(S1) -3 CT(T1)) = 840 cm 1 (0.1 eV). Both the S1 and T1 states result essentially from the HOMO→LUMO transition and thus, HOMO and LUMO correspond well to natural transition orbitals, i. e. to "hole" and "electron", respectively. . The dihedral angle between the donor and acceptor planes is marked. At the right hand side the contour plots of natural transition orbitals [28] for the ground state S0 1 CT(S1) excitation are displayed. "Hole" and "electron" correspond essentially to HOMO and LUMO of the non-excited molecule, respectively.
The donor-acceptor compound 1 shows a bright green emission with max(300 K) = 492 nm at a photoluminescence quantum yield of PL = 80 % as determined for a degassed toluene solution at ambient temperature (PMMA (poly(methyl methacrylate) with ≈ 1 % doping concentration of compound 1: PL = 95 %). Figure 2a reproduces the steady state (cw) emission spectrum exhibiting a halfwidth of about 3400 cm -1 (420 meV). Such a broad band width is expected for a donor to acceptor transition of CT character as predicted by the TD-DFT calculations. Moreover, the observed red shift of the emission spectrum with increasing solvent polarity in the series of toluene, tetrahydrofurane to dichloromethane from max(300 K) = 492 to 592 nm by ≈ 100 nm (≈ 3400 cm -1 ) supports the CT-assignment of the lowest excited states. This indicates already a suitable situation for the occurrence of a small energy separation between the 1 CT(S1) and 3 CT(T1) states, being a condition for achieving TADF behavior. transfer state 1 CT(S1) to the electronic ground state S0. Interestingly, this short decay time observed for compound 1 is significantly longer than found for typical *→ emitters. For these, the prompt fluorescence decay times usually lie in the range of 1 to 5 ns 29 given by a significant allowedness of the corresponding transitions. The weaker allowedness (longer radiative decay time) of the 1 CT(S1)→S0 transition found for compound 1 can be tracked back to the small HOMO-LUMO overlap ( Figure 1) and hence, the resulting small transition dipole moment. 6a,30 On the other hand, the long-lived component of 30 s decay time is assigned as TADF emission. An alternative interpretation assigning this emission to result from triplet-triplet annihilation 31 can be excluded due to the very low concentration of the compound and the very low excitation intensities applied during measurements. Further support for the TADF classification is given by its temperature dependence (see below).  (Figure 2b), while the delayed spectrum is measured from t = 10 µs with a time window of ∆t = 90 s (Figure 2c). It is an important finding that both spectra shown in Figure 3 overlap almost completely. Exactly, this is expected for emissions that stem from the same 1 CT(S1) state, independently of prompt or delayed population. Thus, the delayed spectrum is clearly classified as TADF. An alternative assignment as phosphorescence can be excluded, since (i) a phosphorescence spectrum should appear at lower energy due to the singlet-triplet splitting of 980 cm -1 (120 meV) (see below) and (ii) the phosphorescence decay time, measured in toluene at T = 77 K with (phos) = 1.3 s (PL(total) = 100 %) is more than four orders of magnitude longer than the observed decay of (TADF) = 30 s. The relatively fast TADF decay path (compared to the T1 → S0 paths) prevails competitive non-radiative quenching, hence, a high quantum yield of PL = 80 % is found, even at ambient temperature and in solution.  (Figure 4) In this range, toluene is liquid and the decay behavior is largely monoexponential. (Figure 4, inset) With temperature increase, the decay time continuously decreases from 1 ms (200 K) to 30 s (300 K), while the emission intensity is not changing significantly. Therefore, the decrease of the emission decay time is mainly attributed to an increase of the radiative rate with increasing temperature, which is induced by the activation of the decay path via the 1 CT state, which represents the TADF process.
According to the mono-exponential decay of the long-living component, thermal equilibration between the involved 1 CT and 3 CT states can be assumed to be established. This is valid at least after about 1 s, i.e. after a time being much longer than the processes of prompt  (1) Herein k(T1) is the decay rate of the triplet emission and kB is the Boltzmann constant. k(S1, ISC) is the effective in-series-rate of the processes of ISC with k(ISC) and prompt S1→S0 emission with k(S1). An expression for k(S1, ISC) is obtained by solving the corresponding rate An alternative approach is mentioned in ref 34.
Eq. (1) can drastically be simplified if applied to the emission properties of compound 1. In particular, the energy separation ∆E(S1-T1) >> kBT. Below, it will be shown that ∆E(S1-T1) amounts to 980 cm -1 , while kBT at 300 K is 210 cm -1 . Further, k(T1) << k(S1, ISC). k(T1) and k(S1, ISC) are of the order of 1 s -1 and 10 7 s -1 , respectively (see below). Accordingly, eq. (1) simplifies in the logarithmic form to an Arrhenius-type equation This simple expression with a = ln (k(S1, ISC)/3) is applied to the experimental data as displayed in Figure 4. The slope of the straight line gives the energy separation of ∆E(S1-T1) = 980 cm -1 (120 meV). This value is in good accordance to the calculated TD-DFT result of 840 cm -1 (104 meV) and fits well to the described TADF behavior. For completeness, it is mentioned that the process of ISC between electronic states of the same configuration, i.e. between pure 1 CT and 3 CT states is formally forbidden, since these states do not exhibit spin-orbit coupling (SOC) between each other. 36,37,38 But higher lying singlet and triplet state admixtures can slightly modify the character of the 1,3 CT states and thus, can lead to a slight allowedness for ISC. According to our TD-DFT calculations, adequate states of other configuration are found at about 0.5 eV higher energy than the 1,3 CT states.
Moreover, as recently discussed in detail, also vibronic coupling can induce distinct ISC. 39,40 However, a quantitative approach is not in the scope of this investigation.