Supporting data for "Enhancing luminescence properties of multi-resonance thermally activated delayed fluorescence emitters through gold(I) coordination, extended π-conjugation, and deuteration"

https://doi.org/10.25442/hku.29379365

dataset

posted on 2025-07-07, 04:46 authored by Siyuan Cai

Multi-resonance thermally activated delayed fluorescence (MR-TADF) materials have emerged as promising candidates for high-color-purity organic light-emitting diodes (OLEDs) due to their narrowband emission characteristics. Nevertheless, MR-TADF-based OLEDs generally exhibit significant efficiency roll-off and limited operational lifetimes. These limitations primarily result from inefficient reverse intersystem crossing (RISC) processes caused by weak spin-orbit coupling (SOC) and large single-triplet energy gaps (ΔEST). This thesis focuses on accelerating the RISC of MR-TADF emitters by gold(I) coordination and p-conjugation extension. In addition, the stability of MR-TADF emitters is further enhanced by deuteration.1H NMR, 13C NMR, and 19F NMR spectra were recorded on Bruker AVANCE III 400M, DRX-500, AVANCE III 500M, and AVANCE III 600M with Prodigy Platform. Chemical shifts (δ) were calibrated by an internal standard (TMS) or solvent residue peaks. Matrix-assisted laser desorption ionization time-of-flight mass spectra (MALDI-TOF-MS) were recorded using a Voyager Elite MALDI-TOF mass spectrometer. Trans-2-[3-(4-tert-butylphenyl)-2-methyl-2-propenylidene]malononitrile (DCTB, molecular weight 250.34) was selected as matrix substance of MALDI-TOF-MS. High-resolution electrospray ionization (HR-ESI) mass spectra were obtained from Thermo Scientific Q Exactive mass spectrometer, operated in heated electrospray ionization (HESI) mode, and coupled with Thermo Scientific Ultimate 3000 system. Elemental analyses were recorded on a Vario MICRO analyzer. The thermogravimetric analysis (TGA) measurements were performed on (1) (Chapter III) SDT Q600 instrument at a heating rate of 10 ºC min–1 in the range of 25 ºC to 800 ºC under a nitrogen atmosphere, (2) (Chapter IV and Chapter V) HITACHI STA200 simultaneous thermalanalyzer with a heating rate of 5 ºC min–1 from 30 ºC to 600 ºC at N2 atmosphere.Single crystals suitable for X-ray diffraction analysis were obtained by liquid diffusion between methanol and dichloromethane (or ethyl acetate) over the period of several days. With the help of Dr. Xiaoyong Chang, Dr. Kam-Hung Low, Mr. Shihao Liu, Dr. Haochong Tan, and Dr. Mengyao Chao, data were collected on Bruker D8 VENTURE diffractometer using Mo‒Kα radiation (λ = 0.71073 Å) or Cu-Kα radiation (λ = 1.54178 Å) and PHOTON II CMOS detector. The SADABS (Apex3, Bruker 2019) software package was used for data reduction and empirical absorption correction. Using Olex2[1], structures were solved with the SHELXT[2] structure solution program using Intrinsic Phasing and refined with the SHELXL[3] refinement package using Least Squares minimization. All non-hydrogen atoms were refined anisotropically, while the hydrogen atoms were generated geometrically and refined with isotropic thermal parameters.Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) measurements were conducted using a CH Instruments CHI620E Electrochemical Analyzer. A three-electrode cell configuration was employed, featuring a saturated calomel electrode (SCE) reference, a glassy carbon working electrode, and a platinum wire counter electrode. Samples were prepared in argon-purged DMF containing 0.1 M [nBu4N]PF6 as the supporting electrolyte, with the ferrocene/ferrocenium (Fc/Fc+) couple serving as the internal reference standard.Ultraviolet-visible light (UV-vis) absorption spectra were recorded on a Hewelett-Packard 8453 diode array spectrophotometer. The spectra were generally obtained with 2×10-5 M solutions unless specified. Steady-state emission and excitation spectra of samples in solution, or glassy state were recorded on Hamamatsu Quantaurus-QY Absolute PL quantum yields measurement system C11347 or a SPEX Fluorolog 3 spectrofluorometer. For measurements in solution at room temperature, samples were placed in two-compartment cells consisting of a 10 mL Pyrex bulb and a quartz curvette with 1 cm path length. The cells were sealed from the atmosphere with Rotaflo stopcocks. Solutions were degassed in a high-vacuum line by five freeze thaw-pump cycles. Glassy solutions in 2-methyltetrahydrofuran (10−6–10−5 M) were placed in 5 mm quartz tubes. For low-temperature (77 K) measurements, the quartz tubes containing the samples were placed in a quartz-walled Dewar filled with liquid nitrogen. Thin-film samples were prepared by drop-casting solutions of a samples in chlorobenzene with polymethyl methacrylate (PMMA) onto clean quartz plates. Solvents were evaporated at 60 °C. Delayed florescence lifetimes (τ) were measured with a Quanta Ray GCR 150-10 pulsed Nd:YAG laser system (pulse λexc = 355 nm). Prompt Fluorescence lifetimes were record on Hamamatsu Quantaurus-Tau Fluorescence lifetime spectrometer C16361 (Chapter III) or FLS1000 Photoluminescence Spectrometer (Edinburgh Instruments) equipped with time-correlated single-photon counting (TCSPC) system. The measurements were performed using an LED lamp excitation at 365 nm (for C16361, pulse width: < 200 ps) or EPL375 (for FLS1000, typical pulse width: 60 ps).Chapter III: Femtosecond transient absorption (fs-TA) and time-resolved fluorescence (fs-TRF) measurements utilized a commercial Ti:Sapphire regenerative amplifier (800 nm, 120 fs, 1 kHz, 3.5 mJ/pulse). For both techniques, excitation was at 267 nm, generated as the third harmonic of the fundamental 800 nm pulse. In fs-TA, samples were probed with a white light continuum generated in sapphire by the 800 nm beam. A computer-controlled delay line varied the probe-pump timing, and signals were dispersed by a monochromator onto an air-cooled CCD. The instrument response function (IRF) was around 200–400 fs, exhibiting slight wavelength dependence. For fs-TRF, the 800 nm pulse (200 mW) acted as the gate pulse and the 267 nm pulse (10 mW) as the pump. Sample fluorescence was mixed with the gate pulse in a BBO crystal to generate a sum frequency signal; broadband spectra were recorded by scanning the crystal angle and detecting with the CCD. Solutions (1 mL, 2 mm path length) had an absorbance of 0.5–1.5 at 267 nm during experiments. Separately, nanosecond time-resolved emission (ns-TRE) was measured using a laser flash photolysis system (LP920), exciting samples at 355 nm (6–8 ns pulse, third harmonic of a Nd:YAG laser). Data were processed using L900 software on a PC.Chapter IV and Chapter V: The fs-TA, fs-TRF, and ns-TRE measurements were performed using the same instruments as above. For fs-TA, excitation wavelengths were 420 nm (Chapter IV) and 360/580 nm (Chapter V), generated as the second harmonic of the fundamental 800 nm pulse. For fs-TRF, the output 800 nm laser pulse (200 mW) was used as gate pulse while the 400 nm laser pulse (10 mW) (second harmonic) was used as the pump laser. For the present experiments, the compound in solution was excited by a 430/435/465 nm pump beam (the second harmonic of the fundamental 800 nm from the regenerative amplifier). The solutions were studied in a 2 mm path-length cuvette with an absorbance of 2 at specific excitation wavelength throughout the data acquisition.Temperature-dependent time-resolved photoluminescence (TD-TRPL) spectra and corresponding PL spectra were recorded on FLS1000 Photoluminescence Spectrometer (Edinburgh Instruments) using microflash lamp and xenon lamp as light sources while PMT980 as detector. Ground state (S0) geometry optimizations were carried out using DFT/TPSSh[5] or DFT/PBE0[6] functionals with 6-311G(d)[7] basis set for main-group elements and the Stuttgart relativistic pseudopotential and its accompanying basis set (ECP60MWB) for Au[8]. DFT-D3 dispersion correction[9] with Becke-Johnson(BJ) damping[10] were applied in geometry optimizations. Subsequent frequency calculations were performed on the optimized structure to confirm the absence of imaginary frequencies. No solvent effect was included in calculations since MR-TADF emitters exhibited minimal solvent effect. All geometry optimizations and frequency calculations were performed using the Gaussian 16 version C.01 program[11].Time-dependent density functional theory (TD-DFT) within the Tamm–Dancoff approximation (TDA-DFT)[12] with double-hybrid density functionals (DHDFs) was performed using ORCA[13] version 5.0.3 program. Unless otherwise stated, def2-TZVP[14] was used as basis sets for all TDA-DFT calculations, while the same basis set was used as the auxiliary basis set in RI approximation (RIJCOSX). The natural transition orbital (NTO) analyses of the calculation results were performed using the Multiwfn[15] program and visualized by VMD[16] version 1.9.3.DLPNO-STEOM-CCSD calculations were performed on the ORCA[13] version 5.0.3 program. The calculations were based on PBE0/6-311G(d)–optimized ground state geometries. Unless otherwise stated, def2-TZVP(-f) was used as basis sets for all DLPNO-STEOM-CCSD calculations, while def2-TZVP was used as the auxiliary basis set in RI approximation (RIJCOSX). The convergence conditions were limited by tightSCF and normalPNO. The first five singlet and triplet states (nroots=5) were calculated for sufficient molecular orbitals in the active space. The active space selection was controlled by OThresh and VThresh (0.001). TCutPNOSingles was defined as 1×10−11. Davidson’s dimension was set as 200 with a tolerance of 1×10−5. All DLPNO-STEOM-CCSD calculations were performed on the large-memory nodes of the Taiyi Super-computing Cluster at the Southern University of Science and Technology (SUSTech). The large-memory node was composed of eight Xeon Platinum 8160 CPUs (24 cores for one CPU, a total of 192 cores for one node) and 6 TB memories. Each DLPNO-STEOM-CCSD task requested 24 cores along with 50 GB memories per core (a total of 1.2 TB memories). For the single task, the peak memory usage was ~960 GB, with ~16 TB of hard disk usage. Most of the DLPNO-STEOM-CCSD tasks could be finished within 7 days.Organic light-emitting diodes (OLEDs) were fabricated via thermal evaporation within a Kurt J. Lesker SPECTROS deposition system (base pressure: 10−8 mbar). Organic layers were sequentially deposited at ~0.1 nm s−1, with the emitting layer doped using co-deposition. Subsequently, a Liq layer (2 nm, 0.03 nm s−1) and an Al cathode (100 nm, 0.2 nm s−1) were thermally evaporated. Film thicknesses were monitored in situ using calibrated quartz crystal monitors.For solution-processed OLEDs, PEDOT:PSS were spin-coated onto the cleaned ITO-coated glass substrate and baked at 120 °C for 20 min to remove the residual water solvent in a clean room. Blends of host and emitter(s) were spin-coated from chlorobenzene atop the PEDOT:PSS layer inside a N2-filled glove box. The thickness for all EMLs was about 60 nm. Afterwards, all devices were annealed at 110 °C for 10 min inside the glove box and subsequently transferred into a Kurt J. Lesker SPECTROS vacuum deposition system without exposing to air. Finally, organic functional layers, Yb (1 nm), and Ag (100 nm) were deposited in sequence by thermal evaporation at a pressure of 10−8 mbar.Device characterization including electroluminescence (EL) spectra, luminance, CIE coordinates, external quantum efficiency (EQE), and current efficiency (CE) employed a Keithley 2400 source meter coupled with a Hamamatsu Photonics C9920-12 external quantum efficiency system. All encapsulated devices [200 nm Al2O3 via atomic layer deposition (ALD) in a Kurt J. Lesker SPECTROS ALD system] were characterized at room temperature.