Heavy atom oriented orbital angular momentum manipulation in metal-free organic phosphors

Metal-free purely organic phosphors (POPs) are emerging materials for display technologies, solid-state lighting, and chemical sensors. However, due to limitations in contemporary design strategies, the intrinsic spin–orbit coupling (SOC) efficiency of POPs remains low and their emission lifetime is pinned in the millisecond regime. Here, we present a design concept for POPs where the two main factors that control SOC—the heavy atom effect and orbital angular momentum—are tightly coupled to maximize SOC. This strategy is bolstered by novel natural-transition-orbital-based computational methods to visualize and quantify angular momentum descriptors for molecular design. To demonstrate the effectiveness of this strategy, prototype POPs were created having efficient room-temperature phosphorescence with lifetimes pushed below the millisecond regime, which were enabled by boosted SOC efficiencies beyond 102 cm−1 and achieved record-high efficiencies in POPs. Electronic structure analysis shows how discrete tuning of heavy atom effects and orbital angular momentum is possible within the proposed design strategy, leading to a strong degree of control over the resulting POP properties.


I. Additional experimental details General
All chemicals used were purchased from Millipore Sigma or Fisher Scientific unless specified and used without further purification. hydrazine was purchased from Oakwood Products, Inc.. Deuterated solvents for NMR spectroscopy (nuclear magnetic resonance) were purchased from Cambridge Isotope Laboratories. Phenoxathiine (S-O, 98.0+%) and 9H-thioxanthen-9-one (S-CO, 98.0+%) were purchased from TCI America and used without further purification. 9H-selenoxanthen-9-one (Se-CO) was purchased form Millipore Sigma and used without purification.

Physical measurements
Nuclear Magnetic Resonance (NMR) spectra were collected on Varian MR400 (400 MHz), Varian Vnmrs 500 (500 MHz), or Varian Vnmrs 700 (700 MHz) spectrometer as indicated. Photoluminescence spectra were collected on a Photon Technologies International (PTI) QuantaMaster spectrofluorometer (QM-400) equipped with an integrating sphere (K-Sphere) and a cryostat. The emitters were doped in atactic PMMA matrix for solid-state measurements: quartz substrates (1.5*2.5 cm) were prepared and cleaned by sonication consecutively in soap, deionized water, acetone, isopropyl alcohol, and then proceeded to UVozone treatment for 30 min. Chloroform solution containing 0.025 wt% emitter and 2.5 wt% PMMA was prepared and spin-coated on the cleaned quartz substrates (500 rpm for 5 min). Last, the films were transferred into a glovebox filled with N 2 and baked at 120 o C for 30 min.

Synthesis of prototype molecules
We've designed a series of molecules in this report to provide systematic experimental support for the HAAM concept, as well as to demonstrate the capability of the HAAM concept in creating highly efficient POPs. The synthetic scheme of 10-methyl-10H-phenothiazine (S-N) and 10-methyl-10H-phenoselenazine (Se-N) were adopted and modified from ref 1. The scheme of 10-mesityl-10H-dibenzo[b,e][1,4]thiaborinine (S-B) and 10-mesityl-10Hdibenzo[b,e][1,4]selenaborinine (Se-B) were adopted and modified from ref 2. The scheme of phenoxaselenine (Se-O) was adopted from ref 3. Phenoxathiine (S-O, 98.0+%) and 9Hthioxanthen-9-one (S-CO, 98.0+%) were purchased from TCI America and used without further purification. 9H-selenoxanthen-9-one (Se-CO) was purchased form Millipore Sigma and used without purification.
10H-phenoselenazine (98 mg, 0.4 mmol, 1 eq.) and potassium tert-butoxide (67 mg, 0.6 mmol, 1.5 eq.) were dissolved in anhydrous THF (3 mL) at 0 °C. After stirring for 20 min, methyl iodide (142 mg, 1 mmol, 2.5 eq.) was added to the solution. The mixture was stirred at room temperature overnight. Then, the reaction was quenched with water and the resulting mixture was extracted with CH 2 Cl 2 . The combined organic layers were washed with brine, dried over Na 2 SO 4 , filtered and evaporated to dryness under vacuum. The crude product was further purified with flash column chromatography (Hexane/CH 2 Cl 2 ) to afford 10-methyl-10H-phenoselenazine (Se-N, 1) as a white solid (17.0 mg, 16  (S-N, 2) 10-methyl-10H-phenothiazine 10H-phenothiazine (1 g, 5 mmol, 1 eq.) and potassium tert-butoxide (0.845 g, 7.5 mmol, 1.5 eq.) were dissolved in anhydrous THF (10 mL) at 0 °C. After stirring for 20 min, methyl iodide (0.623 mL or 1.42 g, 10 mmol, 2 eq.) was added to the solution. The mixture was stirred at room temperature overnight. Then, the reaction was quenched with water, and the resulting mixture was extracted with CH 2 Cl 2 . The combined organic layers were washed with brine, dried over Na 2 SO 4 , filtered and evaporated to dryness under vacuum. The crude product was further purified with flash column chromatography (Hexane/CH 2 Cl 2 ) to afford the pure 10-methyl-10H-phenothiazine (S-N, 2) as a white solid (992. 2  2-bromophenyl)hydrazine (2 g, 10.69 mmol, 1 eq.) was dissolved in acetonitrile (25 mL) and added to a stirred suspension of selenium dioxide (1.78 g, 16.04 mmol, 1.5 eq.) in acetonitrile (25 mL). The suspension turned orange and then orange-red after bubble formation, and was then stirred at 60 o C for 1 h. Desired product was observed on TLC after 1 h and the resulting mixture was extracted with hexane to afford a red solution. After concentration under reduced pressure, the crude product was further purified with flash column chromatography (hexane) to afford bis (

 (Se-B, 5) 10-mesityl-10H-dibenzo[b,e][1,4]selenaborinine
t BuLi (1.9M pentane solution, 0.303 mL, 0.5766 mmol, 4.42 eq.) was added to a solution of bis(2-bromophenyl)selane (51 mg, 0.1304 mmol, 1 eq.) in diethyl ether (3 mL) at -78 o C. After stirring for 30 min at -78 o C, MesB(OMe) 2 (37.58 mg, 0.1957 µmol, 1.40 eq.) in diethyl ether (1.5 mL) was added to the reaction mixture. The mixture was warmed up to room temperature and stirred overnight before quenched with water. The organic layer was extracted with CHCl 3 and dried over anhydrous Na 2 SO 4 . After removal of the solvent under reduced pressure, the residue was purified with flash column chromatography (hexane) to afford 10-mesityl-10H-dibenzo [ 1. To a solution of oxydibenzene (1.532 g, 9 mmol, 1eq.) in anhydrous hexane (25 mL), a 2.5 M solution of n-BuLi (9 mmol, 3.7 mL) in hexane was added dropwise at 0 o C and the reaction mixture was kept at this temperature for about 3 h with stirring. The reaction mixture was slowly warmed up to room temperature and it was left under stirring overnight.
2. The solvent was removed in vacuum and the remaining oil was dissolved in THF (20 mL). Afterwards, the stoichiometric amount of Se (710.64 mg, 9 mmol, 1eq.) was added and the solution became deep brown.
3. The reaction mixture was stirred for additional 2 h and then water was added (under argon) and subsequently the solution was exposed to air. The solvent was removed in vacuo and the remaining solid was dissolved in DCM (25 mL). The organic phase was washed with water and dried over anhydrous Na 2 SO 4 to afford 1,2-bis(2-phenoxyphenyl)diselane as a red oil which turned into a solid after storage in freezer (2.076 g, 92.93% yield).

II. Computational Details
The RAS-SF method is programmed in the Q-Chem 5.0 software package, 9 and the SOC computations are implemented in a development version of Q-Chem. All RAS-SF calculations were performed with the polarized, triple-zeta def2-TZVP basis set 10,11 and the RIMP2-cc-pVTZ auxiliary basis. 11 RAS-SF hole, particle calculations with 4 electron in 4 orbital active spaces were carried out with RAS1 and RAS3 subspaces including all occupied and virtual orbitals, respectively. Unless otherwise stated, the core electrons were kept frozen. Reference orbitals for RAS-SF were obtained from restricted open-shell density functional theory (RODFT) using the B3LYP functional in the triplet state. Geometries of the molecules were optimized at the ground state using ωB97X-D functional 12,13 and the def2-TZVP basis set 10,11 . Calculations of SOC constants utilize general libraries developed for SOC calculations within EOM-CC. 14 Spin-orbit NTOs were computed and analyzed using the libwfa library. 15 The NTOs with the largest singular values, for each compound, were plotted using Gabedit program. 16 Figure S1. Reduced SOCME in the selected orientations between S 0 and T 1 states, for Se-N with varying dihedral angle (showing the modulus).     The emission peak at 433 nm (298 K) is from delayed fluorescence since it stayed the same relative intensity in the delayed spectrum, and disappeared at 78 K.   The general trend where or SOCMEs decrease with decreasing dihedral angle is clearly observed in the Se compounds. In the sulfur series, the general trend in still follows prediction in that a sharp drop in dihedral angle from Se-O/N to Se-B/CO has led to obvious reduction in . This leaves just one case, the of S-CO, which is higher than that of S-B and therefore does not fit the trend. We presume two plausible reasons. First, sulfur is not a heavy atom and due to the intrinsically lower SOCME, S derivatives will be more prone to non-radiative decay than their Se counterparts. This deactivation pathway comes from both the compound itself and its interaction with the matrices. Thus, the varying non-radiative decay rate from S-O, S-N to S-B and S-CO could lead to the deviation from the predicted T 1 -S 0 SOCME trend.

IV. Emission/excitation spectra
Second, experimental dihedral angles of the two CO compounds are likely shifted from their calculated value. This is very important particularly for planar molecules since a little distortion from complete planarity could lead to large gain in T1-S0 SOCME. For instance, the predicted T 1 -S 0 SOCME for Se-CO is 0.046 cm -1 and 1.04e -4 cm -1 for S-CO, which should both lead to very long intrinsic phosphorescence lifetime or very low . However, the experimental results in PMMA at 78K suggested otherwise. For instance, Se-CO has a fastthan-predicted of 4.09 ms. Thus, we hypothesized that the experimental dihedral angle or more precisely, Δ in the CO series is non-negligible. Hence, the actual phosphorescence rate should be larger than the predicted value. To demonstrate these in experiments, we've compared the lifetimes of 8 compounds studied in dilute toluene solution with those in PMMA at 78K.
According to Figure S9, changing the matrix has inevitably changed the lifetime at 78K. This change is extremely large for S-CO which had of 8.36 1/s in toluene v.s. 11.61 1/s in PMMA. S-CO and S-B now have similar lifetimes in toluene. We suspect that either nonradiative decay or dihedral angle change both due to interaction with the matrix may have led to the great lifetime change. Due to similar reasons, surprisingly, S-O presented similar large change from PMMA to toluene. However, the overall lifetime trend is very obvious as predicted in the Se compounds, and hence the decreasing (Se derivative)/ (S derivative) as shown in Figure S9c-d. It is less pronounced in the S compounds, but follows the general reducing trend from S-N/O to S-B/CO. Figure S10. Photoluminescence decay of S-CO, S-B, S-N, and S-O in PMMA matrix measured at 78K. The fitting curve was included, which was done by Origin Lab software with its builtin exponential decay function. The raw decay data was recorded from 0 to 250 ms, while fitting was done using the data from 10 to 250 ms. The insets show the fitting in the head (1-30 ms). The head, body, and tail of the decay were all well-fitted, indicating good fitting quality. Figure S11. Photoluminescence decay of Se-CO, Se-B, Se-N, and Se-O in PMMA matrix measured at 78K. The fitting curve was included, which was done by Origin Lab software with its built-in exponential decay function. The raw decay data was recorded from 0 ms, while fitting was done using the data from 0.2 ms. The insets show the fitting in the head (0-2 ms). The head, body, and tail of the decay were all well-fitted, indicating good fitting quality. Figure S12. Photoluminescence decay of S-CO, S-B, S-N, and S-O in toluene measured at 78K. The fitting curve was included, which was done by Origin Lab software with its built-in exponential decay function. The raw decay data was recorded from 0-500 ms, while fitting was done using the data from 10 ms. The insets show the fitting in the head (0-60 ms). The head, body, and tail of the decay were all well-fitted, indicating good fitting quality. Figure S13. Photoluminescence decay of Se-CO, Se-B, Se-N, and Se-O in toluene measured at 78K. The fitting curve was included, which was done by Origin Lab software with its built-in exponential decay function. The raw decay data was recorded from 0 ms, while fitting was done using the data from 0.2 ms. The insets show the fitting in the head (0-2 ms). The head, body, and tail of the decay were all well-fitted, indicating good fitting quality.