Persistent luminescence, long-lived emission from inorganic or organic materials after the cessation of excitation, receives considerable attention in the field of optoelectronics. Despite great achievements in the past decades, the performance of organic materials still lags behind their inorganic counterparts, which have thousands of years of history. This is largely caused by the limited understanding of the mechanisms involved in organic materials. Here we report trap-induced persistent luminescence (TIP) in organic host-guest materials, with controllable trap depths from 0.11 to 0.56 eV and tunable afterglow emission at wavelengths from 507 to 669 nm via energy level engineering. The TIP phenomenon in a typical TN@TPBi film lasts for more than 24 h, with additional energy stored at room temperature for over 1 week. It is found that the trap depth in TIP is probably determined by the energy gap between the lowest unoccupied molecular orbitals of the radical anions of the host and guest molecules, matching well with density functional theory calculations. TIP was also observed after electrical excitation, demonstrating the potential of exploiting the semiconductor features of the organic hosts. These results provide a fundamental principle to design metal-free organic emitters of persistent luminescence, thereby expanding their applications in fields such as medical delivery identification, semiconductor devices, and imaging techniques.

Materials and characterizations: TN was synthesized in two steps including dehydration reaction and palladium catalyzed Suzuki cross-coupling reaction, followed by comprehensive characterizations of nuclear magnetic resonance (NMR) spectroscopy (Advance III 500 MHz NMR, Bruker), mass spectrometry (FTICR MS, Bruker), and element analysis (Vario EL III, Elementar).

TN@TPBi film fabrication. A mixture of TN and TPBi (total 1 mmol, around 60 mg) was heated up to 350 °C on a quartz substrate in a glovebox. After melting, the molten liquid of materials was stirred thoroughly, then cooled rapidly to RT and encapsulated under a nitrogen atmosphere using ultraviolet-cured epoxy resin and glass covers. The other blend films were fabricated with the same method.

Photocurrent device fabrication: Si-SiO₂ substrate was washed with deionized water, acetone, and isopropanol sequentially, and dried in drying oven at 100 °C. The gold as parallel electrodes were evaporated on the substrate at 3.5 Å/s when the vacuum degree of the evaporating cave reached 1×10⁻⁴ Pa. Then the gold electrode deposited substrate was washed with above solutions again and finally treated by oxygen plasma for 30 min. 1 wt% TN@TPBi melt-casting film was prepared via a melt-cooling assisted crystallization method on the substrate and encapsulated with glass covers in a glovebox.

Photophysical properties measurement. Emission spectra were collected in a fluorescence spectrometer (FLS980, Edinburgh and QE-Pro, Ocean Optics). The absorption spectra were measured by an ultraviolet/visible/near-infrared spectrophotometer (UV-3600Plus, Shimadzu). The cyclic voltammetry curves and electrochemical deposition operation were carried out by electrochemical workstation (CHI660E, CH Instruments). The radical cation and anion of host/guest were obtained from electrical oxidation or reduction reaction when the host/guest were dissolved in dichloromethane containing 0.1 M TBAPF₆. We obtained ESR spectra using the X-band EPR spectrometer (EMX-10/12, Bruker). A device was fabricated by depositing two Au electrodes on a glass substrate and then melting the organic emitter to cover electrodes, as shown in the inset of Fig. 3g. The photocurrents of the device with ultraviolet light (365 nm) on and off (at zero bias) were measured using a high-precision electrometer (6517B, Keithley). For persistent luminescence photocurrent measurements, the device was illuminated under 365 nm light for 300 s, and after the illumination removal, the photocurrent decay versus time (300 s) was recorded. Five cycles were applied. For dark current recording, the device was first thermally bleached at 400 K to fully eliminate trapped charge carriers. Photographs and Videos of the samples were taken with a digital camera (EOS 5D Mark II, Canon and α7SIII, Sony).

TIP measurements. TL spectra, persistent luminescence spectra, and decay profiles were obtained using the TIP measurement system shown in Supplementary Fig. 24. The samples were placed in a cryostat controlled by a cooling-heating stage (THMS600E, Linkam Scientific Instruments), and excited by a 365-nm LED with a 5 mW/cm² excitation power density for 5 min. A filter-attached photomultiplier tube (PMT, R928P, Hamamatsu photonics), a luminance meter (LM-5, Evenfine), a multimeter (2400, Keithley), and a high voltage power supply (HVC1800, Zolix) simultaneously monitored the persistent luminescence intensity (including TL emission and TIP decay profile) after ceasing the excitation source. The TIP emission spectra were recorded at the same time by using a multichannel spectrometer (QE-Pro, Ocean Optics) during TL measurements. In a typical TL measurement, the sample was firstly cooled to 100 K and irradiated by the excitation source for 5 min. Twenty seconds after creasing the 365-nm LED source, the sample was heated to 400 K at a certain heating rate (50, 20, 10, 5, and 2 K/min) to record the emission intensity. The above-mentioned homemade measurement systems were driven by LabVIEW-based PC programs. An NIR laser at 980 nm (~200 mW/cm²) was projected onto the surface of the sample as a photo-stimulation source.

Computational details. All molecules were optimized at the B3LYP/def2-SVP level for the equilibrium geometries of ground state (S₀) using Gaussian 09 program, and subsequently, the frontier molecular orbitals based on the S₀-geometry were depicted for the close-shell and open-shell systems.