Cationic photoredox catalysts, 2,4,6-triphenylpyrylium tetrafluoroborate (TPP+) and 2,4,6-tris(methoxyphenyl)pyrylium tetrafluoroborate (MeOTPP+) were used as electron acceptors28–30 and semiconducting host molecules 3,3’-di(9H-carbazol-9-yl)biphenyl (mCBP)31 and 1,3,5-tris(1-phenyl-1H-benzimidazol-2-yl)benzene (TPBi)32 were used as electron donors (Fig. 1c). 4,4’,4’’-Tri(9-carbazoyl)triphenylamine (TCTA)33 was used as a hole-trap material. Analytical results of these materials by UV–visible absorption, fluorescence, and phosphorescence spectra are shown in Figure S1. Energy levels of the lowest singlet excited state (1LE) and triplet excited states (3LE) were estimated from the onset of fluorescence and the phosphorescence spectra, respectively. LUMO levels were obtained from the first reduction peaks of the cyclic voltammograms (CV) (Figure S2), and the HOMO levels were estimated from the optical gap that was calculated from the absorption edge (Fig. 1d, Table S1). The HOMO level of the TCTA was obtained from the CV, and the LUMO level was estimated from the optical energy gap because of the electric window of the solvent.
LPL films with a 1:99 molar ratio of acceptor:donor system were fabricated by conventional melt casting11. Steady-state photoluminescence (PL) and LPL spectra, emission decay profiles, and PL quantum yields (FPL) were obtained under nitrogen gas. The LPL duration was defined as the time until the emission intensity dropped below 1 pW after 60 s of excitation. The triplet charge-transfer excited state (3CT) level was assumed from the singlet charge-transfer excited state (1CT) level that was obtained from the onset of the PL spectrum because most LPL systems have a small energy gap between the 1CT and 3CT34.
When the films were excited by 365-nm light, which can only be absorbed by the acceptors, the TPP+/TPBi, TPP+/mCBP, and MeOTPP+/mCBP films exhibited a long LPL emission in which the decay profiles followed a power-law decay (Fig. 2a)35,36. This power-law emission decay indicates the generation of intermediate charge-separated states and successive gradual charge recombination, which led to LPL. The emission spectra of these films was attributed to the CT excited states between the donors and acceptors16,17.
The TPP+/TPBi, TPP+/mCBP, and MeOTPP+/mCBP films form 1CT excited states (Fig. 2b) and the 1CT energy levels are lower than those of the locally excited states of the donors and acceptors (3LED and 3LEA). Therefore, the LPL that was caused by the recombination of accumulated charges occurs from the CT states. The TPP+/TPBi film exhibits a CT emission at 603-nm with a shoulder peak at ~ 555-nm. The shoulder peak decreased with an increase in TPP+ concentration because of the self-absorption of TPP radicals at 500–600 nm (Figure S3a)17,37,38 and it disappeared at the higher TPP+ concentration because of a strong self-absorption of the TPP radical (Figure S3b). The TPP+/TPBi film showed the longest LPL duration of 1435 s because of the highest FPL of 10.2%. The LPL duration decreased with an increase in TPP+ concentration because the charge recombination probability increased at the higher TPP+ concentration (Figure S3c).
The TPP+/mCBP film exhibited a broad NIR emission at 731-nm because of the smaller energy gap between the TPP+ LUMO and mCBP HOMO. Because of the low FPL of the CT emission and low NIR sensitivity of the photodiode for detection, the LPL duration was 19 s. The MeOTPP+/mCBP films exhibited a CT emission at 624-nm with an LPL duration of 605 s. In contrast, the MeOTPP+/TPBi film did not form a CT excited state although the HOMO and LUMO levels were appropriate. The TPBi may not act as a donor because of the closed HOMO levels between them. Instead, this film exhibited the fluorescence and RTP of the MeOTPP+ (Figure S4c). These results indicate that the formation of the lowest 1CT state is important for efficient LPL emission in p-type OLPL systems.
Because a blend film of donor and acceptor molecules has donor, acceptor, and charge-transfer absorption bands, OLPL systems can be excited by various wavelengths, and is a major advantage over inorganic LPL systems, which are limited mostly to UV to blue excitation wavelengths. The excitation spectra of the TPP+/TPBi and MeOTPP+/mCBP films indicate that these films can be excited by a wavelength from 300-nm to 600-nm (Figure S5). To confirm the excitation wavelength dependence of the LPL emission, these films were excited by 365-nm, 400-nm, 455-nm, 500-nm, 550-nm, and 600-nm LEDs. The LPL emission was observed at all excitation wavelengths, although the LPL duration decreased, which correlates with the absorption intensity (Figs. 2c and S5c). The 600-nm photoexcitation and NIR LPL emission, which corresponds to the biological window, is expected to be used for bio-imaging39.
The LPL performance improved sixfold with hole-trapping material addition to the p-type OLPL system. Previously, we reported that the LPL duration could be enhanced by doping electron-trapping materials into an n-type OLPL system13. The dopants can receive electrons from the acceptor molecules because the LUMO of the dopant is lower than that of the acceptor. By tuning the appropriate LUMO level, the trapped electrons remained for more than one week40. In the p-type OLPL system, TCTA was doped into the TPP+/TPBi system because the HOMO level of the TCTA (–5.3 eV) was shallower than that of the TPBi (–5.9 eV) (Fig. 1d and Figure S2). The PL and LPL spectra of the TPP+/TPBi/TCTA film with a 1:99:1 molar ratio were identical to that of the TPP+/TPBi film (Fig. 3a) but the LPL duration improved sixfold to 9045 s (Fig. 3b). This result indicates that the TCTA acts as hole traps to the charge-separated state (Fig. 3c, Figure S6a).
To confirm the hole trapping by TCTA, absorption spectra of the TPP+/TPBi and TPP+/TPBi/TCTA films were obtained40. After photoexcitation, a clear broad absorption above 800-nm was observed in the TPP+/TPBi/TCTA film (Fig. 3d). This peak corresponds to the radical cation of TCTA in dichloromethane (DCM) under electrical oxidation, although a wavelength shift can be observed because of the polarization effect in the solid film (Fig. 3d). The hole trapping was confirmed by the temperature dependence of the LPL duration41. Because hole detrapping was endothermic, the LPL intensity increased by increasing the temperature in the TPP+/TPBi/TCTA film (Figure S6b). In contrast, the LPL duration of the TPP+/TPBi film decreased gradually with an increase in temperature because the nonradiative process was enhanced at a high temperature (Figure S6c). Thus, detrapping in the TPP+/TPBi/TCTA film for the LPL intensity overwhelmed the nonradiative process at the measured temperatures.
Optical analysis of the TPP+/TPBi, MeOTPP+/mCBP, and TPP+/TPBi/TCTA films was carried out in the air to confirm the air stability of the p-type OLPL systems because the LUMO levels of the TPP+ (–4.0 eV) and MeOTPP+ (–3.8 eV) are lower than the reduction potential of oxygen (–3.5 eV). Although the reported n-type OLPL system of the m-MTDATA/PPT did not show LPL emissions in air (Figure S7a), all p-type OLPL films exhibited LPL in air (Fig. 4a, b, Figure S7b). The observed LPL durations of all films in air were shorter than in nitrogen. Although the reaction between acceptor radicals and oxygen can be prevented by deeper acceptors’ LUMO levels22, energy transfer from the triplet excited state of the emitters to the molecular oxygen with a triplet ground state (triplet quenching) cannot be prevented42. The charge recombination in LPL emissions generates 1CT and 3CT excited states43, and the 3CT excited states are quenched by oxygen. In contrast, the emission spectra did not change because of no 3LE contribution (Fig. 4c, d). The LPL duration of the TPP+/TPBi/TCTA in air was extended to 1421 s, which is almost the same as the TPP+/TPBi film in nitrogen although the FPL of both films is almost identical (Supplementary Video 1). These results indicate that the p-type OLPL system with deep HOMO levels can emit LPL in air but cannot prevent triple quenching by oxygen. Future developments of the CT excited state with a rapid reverse intersystem crossing (RISC) that is faster than the energy transfer to oxygen or advanced encapsulation techniques to prevent oxygen are required to obtain efficient LPL emissions in air.