Optimized Synthesis of Solution‐Processable Crystalline Poly(Triazine Imide) with Minimized Defects for OLED Application

Abstract Poly(triazine imide) (PTI) is a highly crystalline semiconductor, and though no techniques exist that enable synthesis of macroscopic monolayers of PTI, it is possible to study it in thin layer device applications that are compatible with its polycrystalline, nanoscale morphology. We find that the by‐product of conventional PTI synthesis is a C−C carbon‐rich phase that is detrimental for charge transport and photoluminescence. An optimized synthetic protocol yields a PTI material with an increased quantum yield, enabled photocurrent and electroluminescence. We report that protonation of the PTI structure happens preferentially at the pyridinic N atoms of the triazine rings, is accompanied by exfoliation of PTI layers, and contributes to increases in quantum yield and exciton lifetimes. This study describes structure–property relationships in PTI that link the nature of defects, their formation, and how to avoid them with the optical and electronic performance of PTI. On the basis of our findings, we create an OLED prototype with PTI as the active, metal‐free material.

= 1 = = ρ= electrical resistivity σ=electrical conductivity U=bias I=current A=surface area w=channel width t=film thickness L=channel length For Figure S9b the LED was not focused with a collimator and the irradiance was determined by calculating it in dependence on the distance between substrate and LED light source.
Film thickness measurements: For extraction of conductivity film thickness values have been obtained with an Olympus LEXT laser scanning microscope.

Fourier transform infrared (FT-IR):
Spectra were recorded from solid on a Thermo Scientific Nicolet iS5 spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) in the wavenumber range of 4000-600 cm-1 with resolution of 4 cm-1.
Raman: UV Raman spectra were recorded with a Horiba T64000 spectrometer in single-grating mode. The excitation source was provided by a diode-pumped solid-state laser from CryLas at a wavelength of 266 nm and power of 4 mW. The light was focused on the sample with a Thorlabs LMU-40x-UVB objective (backscattering geometry). A notch filter with cut-off at 220 cm -1 was used to filter out the elastically scattered light. The acquired Raman spectra were calibrated by comparison with the spectrum of a Ga2O3 crystal by using a quadratic calibration curve. This procedure allows for reduction of the experimental error at approximately 5 cm -1 , which is below the spectral resolution of 8 cm -1 .
Powder X-ray: Structural analysis of the prepared PTI-MX was performed with a Bruker D2 Phaser X-Ray powder diffractometer (XRD) in Bragg-Brentano geometry. X-rays were generated by a Cu Kα1+2 source at 30 kV operating voltage and collected with a LynxEye detector.
Photoluminescence, Photoluminescence excitation, quantum yield, Lifetime: Edinburgh Instruments FLS 980 spectrometer. Photoluminescence, photoluminescence excitation and quantum yield were measured using a Xe lamp. The quantum yield was determined in a direct excitation set up with an integrating sphere. Films were drop-casted from ethanol dispersions on quartz substrates PGO 10x10x0.5 mm. Further, it is important to note that the presented samples were prepared and characterised on the same day. Diffusion of protons in PTI-LiBr has not been studied yet. It is possible that crystals of PTI-LiBr are not fully protonated because the system has had not enough time to equilibrate. This might explain why the luminescence of the basic Li-defect is still present in acidic solutions. Lifetime measurements were conducted with an Edinburgh Instruments 375 nm pulsed laser set to a 50 ns pulse period. Evaluation was conducted with the instrument software choosing a two exponential decay function. .032) For determination of the secondary electron cut off (SECO) a negative bias of 10 V was applied to clear the analyser work function. All measurements were performed at room temperature. The obtained values were rounded to one decimal place. The work function was determined by linear fitting of the background and the linear region of the SECO, and reading of the point of intersection. The hole injection barrier was determined by linear fitting of the background and the valence band onset, and reading of the point of intersection. The error of the measurement is ±0.1 eV. Samples for UPS were prepared by dispersing 6 mg/mL of PTI-LiBr in MeOH (Roth, ≥99%) by sonication for 10 min and subsequently dropcasting 50 µL of the dispersion at 60 °C on ITO coated glas substrates (1x1cm).

SEM:
The sample was dispersed in Methanol (~0.6 mg/mL) with a micropipette and 0.5-1 µL were dropcast in the middle of a holy carbon TEM grid. SEM images have been recorded with a GeminiSEM 500 electron microscope (Carl Zeiss GmbH, Germany).

Low-dose, high-resolution transmission electron microscopy:
The sample was dispersed in Methanol (~0.6 mg/mL) with a micropipette and 0.5-1 µL were dropcast in the middle of a holy carbon TEM grid. The TEM grids were loaded into a cryogenic transfer holder (Gatan 914, Gatan, Munich, Germany) at room temperature and transferred to the TEM. Once in the TEM, the holder was cooled down with liquid nitrogen and imaging was performed with a low dose acquisition scheme using SerialEM (doi: 10.1016/j.jsb.2005.07.007) on JEM-2100 (JEOL GmbH, Eching, Germany) operated at 200 kV and equipped with a 4 k × 4 k CMOS digital camera (TVIPS TemCam-F416). HRTEM images were acquired at a magnification of 500,000×, corresponding to a pixel size of 0.23 Å at the specimen level, while keeping the total electron dose below 20 e Å -2 . All imaging was carried out at temperatures around 90 K.
PTI-LiBr HCl titration: 26 mg PTI-LiBr were suspended in 30 mL dest. water by sonication. The pH value changed from 5.5 to 9.4 after suspension of PTI-LiBr. A 0.1 M HCl solution was used as titrant and added in 50 µL then 100 µL steps while stirring with a magnet stirrer. The pH was monitored with a pH glas electrode. The experiment was conducted at room temperature.PTI-LiBr dispersion for OLED preparation: 30 mg of PTI-LiBr and 5 mL chlorobenzene (Sigma-Aldrich, anhydrous, 99.8%) are added into a 50 mL falcon tube. The falcon tube is placed in an ice bath and the sonotrode has been immersed about 3-5 mm into the dispersion. Dispersions are sonicated for 3 h, centrifuged at 95 g for 5 min. The supernatant is transferred into a new falcon tube and the procedure is repeated at 857 g for 5 min two times. Sediments were re-dispersed via sonication (10 min) in a sonication bath prior to usage.
OLED preparation: ITO-coated glass substrates (sheet resistance = 20 Ω per square) were cleaned by sequential sonication (10 minutes) in (i) acetone and (ii) isopropanol followed by drying via a nitrogen gun. The substrates were then treated via O2 plasma (partial pressure 1.2 × 10 −1 mbar) for 15 minutes at 10.2 W. 50 nm PEDOT:PSS (Osilla) films were spin coated as hole injection layer and heated to 220 °C for 10 min. 40 μL PTI-LiBr dispersion were drop cast at 60 C from the chlorobenzene dispersion onto the pixel areas of the PEDOT:PSS covered substrate (ca. 150 mm 2 ). 5 nm Calcium and 200 nm Aluminium were evaporated in a PVD chamber at 10 -5 mbar. Finally the OLEDs were encapsulated with UV-curable resin (Osilla) and a glass slide. Current density-voltage-luminance characterization was performed with a Keithley 2612B source meter and a Konica Minolta LS-160 luminance meter in a purpose-built setup. Electroluminescence spectra were taken with a CS2000 spectrometer (Ocean Optics) using OceanView software.