Main

Upon exposure to ultraviolet or visible light, some substances absorb the excitation energy and release it as a differently coloured light either quickly, as fluorescence, or slowly, in the form of either phosphorescence or long persistent luminescence. Exploiting the slowest emission process, LPL materials have been widely commercialized as glow-in-the-dark paints for watches and emergency signs1, and are being explored for application in in vivo biological imaging because their long-lived emission makes it possible to take time-resolved images long after excitation5. In the mid-1990s, a highly efficient LPL system was developed that uses SrAl2O4 doped with europium and dysprosium, and this inorganic system forms the basis of most commercial glow-in-the-dark paints because of its long emission (more than ten hours) and high durability2. However, this system requires not only rare elements for long-lived emission but also very high fabrication temperatures of more than 1,000 °C. Moreover, the manufacturing of paints from the insoluble SrAl2O4 requires many steps, including grinding of the compounds into micrometre-scale powders for dispersion into solvents or matrices, and light scattering by the powders prevents the formulation of a transparent paint1. The realization of LPL from organic molecules would solve many of these problems.

Long-lived emission from organic molecules is known as phosphorescence, but this lasts for less than one second in most cases and for just a few minutes in the longest cases6, which does not compare favourably with inorganic LPL emitters. Thus, conventional radiative transitions from the photo-excited states of organic molecules—fluorescence, phosphorescence and delayed fluorescence—are not suitable for realizing OLPL. However, very long lifetimes can be achieved when light is used to create ionized states (photo-induced ionized states) and charge-separated states (photo-induced charge-separated states)3.

The photo-ionization of organic molecules was first reported in 1942 (ref. 7) and was the result of successive two-photon absorption, with absorption of the first photon inducing an excited intermediate state (generally a triplet or singlet) and absorption of the second photon producing an ionized state whose energy exceeds the ionization potential8. Subsequent study of the two-photon ionization of organic guest molecules dispersed in polymer matrices—which provide a rigid environment that stabilizes the radical cations of the guests—found that this system could generate LPL that lasted for more than ten hours at 20 K (refs 3, 9, 10). In this process, electrons from the photo-ionized guest molecules accumulate in the polymer matrix and slowly recombine with the radical cations to generate excited states of the guest molecule in a ratio of 25% singlets to 75% triplets (a ratio determined by spin statistics). In this way, both fluorescence and phosphorescence of the guest molecules are observed after recombination3. However, this particular two-photon ionization process requires an intense excitation source and very low temperatures.

Blends of electron-donating and electron-accepting molecules—often used in organic photovoltaics—offer a way of forming charge-separated states even under weak photo-irradiation11,12,13,14. After an electron is photo-excited from the highest occupied molecular orbital (HOMO) to the lowest unoccupied molecular orbital (LUMO) of the electron-donating molecule, the electron is transferred from the LUMO of the donor to the LUMO of the acceptor to form a charge-transfer state before separating into a charge-separated state. A long-lived charge-separated state has even been reported for a single molecule by using a twisted donor–acceptor structure15. However, OLPL under weak photo-irradiation has still not been realized.

Here we use a blend of organic donor and acceptor molecules to demonstrate low-power excitation of OLPL at room temperature. The emission, which lasts for more than one hour, is generated through the charge recombination of long-lived intermediate charge-separated states. For this process, the photo-induced charge separation—as well as the emission from the recombined charges—must be efficient, because even the slow radiative process of phosphorescence (Fig. 1a) normally causes the excited states of organic molecules to deactivate within a few minutes. To achieve long-lived charge-separated states, we selected the strong electron-donating molecule N,N,N′,N′-tetramethylbenzidine (TMB), which has a very stable radical cation10, and the strong electron-accepting molecule 2,8-bis(diphenylphosphoryl)dibenzo[b,d]thiophene (PPT), which has a high triplet energy and provides a rigid amorphous environment to help suppress non-radiative deactivation (see Fig. 1b for chemical structures). Unlike in organic photovoltaics, where the charges are swept out of the device, the photo-generated radical cations and anions must accumulate in our blend without losing their energy through recombination.

Figure 1: A mechanism for organic long persistent luminescence.
figure 1

a, The emission mechanism of long-lived phosphorescence. A molecule is photo-excited (i) from its ground state (S0) to its singlet excited state (S1), from which it can return to its ground state by emitting fluorescence; alternatively, it can cross (ii) to its triplet excited state (T1), which relaxes to the ground state by phosophorescence (iii). Non-radiative decay may also occur. b, Chemical structures of TMB (the donor molecule) and PPT (the acceptor). c, Emission mechanism of OLPL. Left, during photo-excitation, electrons (black circles) are transferred from the HOMO of the donor molecules to the HOMO of the acceptor molecules to form charge-transfer states (i, ii). Then, the acceptor radical anions diffuse to isolate the donor radical cations from the acceptor radical anions, forming charge-separated states (iii). Right, gradual recombination of the radical anions and radical cations (iv) generates exciplex emission (transitions from the LUMO of the acceptor to the HOMO of the donor; v). The HOMO and LUMO energies were determined by photoelectron spectroscopy and ultraviolet/visible absorption spectroscopy. d, Triplet harvesting in exciplex emission. Charge recombination of the PPT radical anions (minus symbol) and TMB radical cations (plus symbol) produces exciplexes in a ratio of 25% singlet exciplexes to 75% triplet exciplexes. A small energy gap between the lowest singlet excited state (S1) and the lowest triplet excited state (T1) of exciplex systems enables reverse intersystem crossing (ISC) via thermal activation.

PowerPoint slide

Full size image

We fabricated amorphous films with varying amounts of TMB dispersed into PPT by a conventional melt-casting method, in which blends of the two materials were heated over the melting point of PPT (250 °C) under a nitrogen atmosphere and then rapidly cooled to room temperature. During photo-excitation, charge-transfer states are formed between TMB and PPT (Fig. 1c, i, ii). Then, the generated PPT radical anions diffuse by the hopping of charges among the PPT molecules to isolate the TMB radical cations from the counter radical anions of PPT and form stable charge-separated states (Fig. 1c, iii) Gradual recombination of the PPT radical anions and TMB radical cations generates exciplex emission (Fig. 1c, iv, v), which continues long after photo-excitation ceases.

Figure 2a shows the ultraviolet/visible absorption, fluorescence and phosphorescence spectra of TMB and of PPT in 10−5 M toluene solution. Because of their short π-conjugation, both TMB and PPT show absorption below 350 nm. The fluorescence maxima of TMB and PPT were located at 394 nm and 346 nm, respectively. The emission peaks of phosphorescence spectra obtained at 77 K were located at 486 nm and 520 nm for TMB, and 425 nm and 457 nm for PPT. By contrast, a 1 mol% TMB/PPT blend exhibited a broad emission spectrum with a maximum at 526 nm (Fig. 2a), which is substantially redshifted compared with the fluorescence and phosphorescence of the individual molecules. This photoluminescence spectrum indicates the formation of an exciplex between TMB and PPT. The absorption spectrum of the blended film was slightly extended up to 450 nm and agrees well with the excitation spectrum of the exciplex emission. This extended absorption might result from the charge-transfer absorption band or from the presence of a small amount of TMB radical cations, which should absorb in the longer-wavelength regions.

Figure 2: Photophysical properties of TMB, PPT, and a TMB/PPT blend.
figure 2

a, Ultraviolet/visible absorption and photoluminescence spectra of TMB and PPT in toluene and of a 1 mol% TMB/PPT film. We measured the fluorescence and phosphorescence spectra in toluene at room temperature and 77 K, respectively. We measured the photoluminescence, LPL and excitation (at 530 nm) spectra of the 1 mol% TMB/PPT film at room temperature. ε is the molar absorption coefficient. b, Logarithmic plot of the emission decay profile of a 1 mol% TMB/PPT film (excitation power, 500 μW; excitation time, 60 s; sample temperature, 300 K). c, Transient absorption spectrum of a 1 mol% TMB/PPT film. Here, the change in absorbance is the absorbance spectrum after photo-excitation minus that before photo-excitation. d, Photographs of LPL from a 1 mol% TMB/PPT film. a.u., arbitrary units.

PowerPoint slide

Full size image

After turning off the photo-excitation of the blend film, we observed LPL with a spectrum identical to that of photoluminescence, indicating that the LPL emission is also derived from exciplexes between TMB and PPT. The emission decay profile of a 1 mol% TMB/PPT film is shown in Fig. 2b. Surprisingly, this LPL continues for more than one hour at 300 K. By contrast, the phosphorescence lifetimes calculated from the transient emission decay profiles of TMB and of PPT in toluene at 77 K (Extended Data Fig. 1) are only 2.20 s and 0.63 s, respectively. Although exciplexes often exhibit delayed emission through thermally activated delayed fluorescence, such emission typically lasts for less than one second and follows a multi-exponential decay1. But for the 1 mol% TMB/PPT film, the LPL intensity decay follows an inverse-power function of time t−1, and does not follow a single exponential decay (Extended Data Fig. 2). This dependence corresponds to an order-of-magnitude decrease in intensity for each order-of-magnitude increase in time.

The decay profile follows the Debye–Edwards law16 (t−m, with m = 1) and is similar to that reported for two-photon-ionized TMB in a polymer matrix3,9,10; however, the photoluminescence and LPL intensities in our system depend linearly on excitation power (Extended Data Fig. 3) with a slope of one, indicating a single-photon process. The Debye–Edwards law describes long-range electron transfer and recombination and was originally derived to explain the emission of directly ionized phosphors frozen in solution by modelling the process as the diffusion of ejected electrons through the medium to radical phosphors that act as sinks at which emission occurs. In the model, an m of unity corresponds with a low concentration of recombination sites. Thus, the t−1 decay profile is consistent with the presence of intermediate charge-separated states produced by photo-irradiation.

While two-photon ionization of TMB in a polymer matrix results in both fluorescence and phosphorescence, the TMB/PPT blend film exhibited only exciplex emission. Exciplex systems can harvest triplet excitons as delayed emission via reverse intersystem crossing because of a small energy gap between the lowest singlet excited state and lowest triplet excited state17,18. Therefore, this OLPL system may also be able to harvest all excitons generated by recombination if the non-radiative decay processes are completely suppressed (Fig. 1d). However, we could not obtain direct evidence for the presence of delayed fluorescence resulting from reverse intersystem crossing, because the single exponential decay of any delayed fluorescence was masked by the LPL emission decay.

The LPL emission was quenched in air because the photo-generated radical anions and cations are unstable in the presence of water or oxygen. Thus, encapsulation is necessary, and the development of paints that can be used without external barrier layers will require new techniques to effectively achieve encapsulation within the paint itself. The emission quantum yield (ΦPL) obtained using an integrating sphere was 21% ± 3% in nitrogen (which mimics encapsulation), but dropped to 7% ± 2% in air. Given that the measurement system for ΦPL captures the emission under photo-excitation, this value does not represent just LPL emission, but it nonetheless provides a rough estimate of 14% ± 3% for the triplet and charge-separation-related component of emission. Once encapsulated, the OLPL system is stable for more than one month in dark conditions, and we expect that stabilities similar to those of encapsulated organic photovoltaics19 and organic light-emitting diodes—which can operate for a couple to tens of years—should be achievable.

Figure 2d and Supplementary Video 1 show images of the OLPL at 300 K, recorded using a charge-coupled-device camera. The emission could be detected for more than 30 min and is most intense around the cracks caused by the rapid cooling of the melt-cast film. Because efficient charge separation occurs even under weak photo-irradiation, the OLPL system can be excited using a standard white LED lamp that does not have emission below 400 nm (Extended Data Fig. 4 and Supplementary Video 2). To confirm the formation of a charge-separated state under such illumination, we obtained the transient absorption spectrum after photo-excitation of a TMB/PPT film (Fig. 2c). The photo-excited film exhibited a broad absorption located between 600 nm and 1,400 nm, which corresponds to the absorption of the TMB radical cation20 and clearly indicates the formation of TMB radical cations by photo-irradiation.

The duration of the LPL depends on the excitation power and time, the sample temperature, and the TMB doping concentration. As shown in Fig. 3a, the LPL duration in a 1 mol% TMB/PPT film at 300 K, excited by a 340-nm LED for 60 s, increased with increasing excitation power but was still observable at excitation powers as low as 10 μW (even at 300 K). The LPL duration also increased with excitation time at a constant excitation power of 10 μW (Fig. 3b). Although the detected emission duration was only 200 s after an excitation time of 1 s, LPL lasting for more than 5,000 s was achieved with an excitation time of 180 s. This excitation time dependence is clear evidence for the generation and accumulation of charge carriers by weak photo-irradiation. The LPL intensity and duration start to saturate at high excitation powers and long excitation times because the number of accumulated charge carriers begins to reach its upper limit (Extended Data Fig. 5).

Figure 3: LPL decay profiles for TMB/PPT films.
figure 3

a, Decay profiles showing the dependence of LPL on excitation power (1 mol% TMB/PPT; excitation time, 60 s; sample temperature, 300 K). b, Decay profiles showing how LPL depends on excitation time (1 mol% TMB/PPT; excitation power, 10 μW; sample temperature, 300 K). c, Decay profiles showing how LPL depends on temperature (1 mol% TMB/PPT; excitation power, 500 μW; excitation time, 60 s). d, Decay profiles showing the dependence of LPL on the concentration of TMB (excitation power, 500 μW; excitation time, 60 s; sample temperature, 300 K). For all measurements, we used a 340-nm LED as the excitation source.

PowerPoint slide

Full size image

The duration of LPL from the 1 mol% TMB/PPT film depends on the sample temperature because of the presence of non-radiative deactivation caused by molecular vibrations (Fig. 3c). The LPL duration was nearly constant at temperatures below 300 K and decreased at temperatures above 300 K, indicating that non-radiative deactivation by molecular vibrations is nearly suppressed even at room temperature but increases at higher temperatures. This dependence may also be related partly to improved charge mobility at higher temperatures, increasing the probability of charge recombination. The photoluminescence and LPL spectra are practically identical at all temperatures, and no phosphorescence was observed even at low temperatures (Extended Data Fig. 6).

The LPL duration also depends strongly on the doping concentration of TMB. When the concentration of TMB was increased from 1 mol% to 50 mol%, the emission spectra redshifted slightly, from 526 nm to 550 nm (Extended Data Fig. 6), and the emission duration time decreased drastically from 104 s to 102 s, with the same t−1 decay profile (Fig. 3d). This can be explained by a reduced distance between TMB and PPT at higher TMB concentrations, leading to an increase in the recombination probability of TMB radical cations and PPT radical anions. The observed trends are all consistent with the generation of intermediate charge-separated states by the weak photo-irradiation, as proposed in Fig. 1c.

In conclusion, we have demonstrated LPL that lasts for more than one hour at room temperature by using a simple mixture of two organic molecules. Unlike existing inorganic systems, our organic LPL system is free from rare elements, transparent, and easy to fabricate and process; it also has the potential, with further development, to be flexible and colour-tunable. These properties will open new applications for LPL in fabrics, windows, and large-area and flexible paints, and the ability to functionalize the donor and acceptor molecules makes OLPL promising for in vivo bioimaging. The emission colour and absorption wavelengths can be easily modified by selecting appropriate electron-donating and electron-accepting molecules. We are now working to improve the stability of the OLPL in air by applying and further developing encapsulation techniques similar to those used for inorganic systems and by controlling the density of the films using semiconducting polymers. Further study of the long-lived charge-separated state formed here by weak photo-irradiation is expected to advance our understanding of not only a wide variety of organic semiconducting devices but also artificial photosynthesis4.

Methods

Materials

TMB was obtained from TCI Chemicals (Tokyo) and purified by recrystallization. PPT was synthesized according to ref. 21. All compounds were purified by sublimation and stored in a glovebox.

Film fabrication

For the optical measurements, we placed a mixture of TMB and PPT (total 1 mmol, 50–60 mg) on a quartz substrate and heated it up to 250 °C for 10 s in the glovebox. After melting, the substrate was cooled rapidly to room temperature and encapsulated under a nitrogen atmosphere using ultraviolet-cured epoxy resin and glass covers.

Optical measurements

We measured absorption spectra using an ultraviolet/visible/near-infrared spectrophotometer (LAMBDA 950, Perkin Elmer). We measured photoluminescence spectra using spectrofluorometers (FluoroMax, Horiba Jobin Yvon; and QE-Pro, Ocean Photonics). The absolute photoluminescence quantum yields (ΦPL) were measured using an integrating sphere with a photoluminescence measurement unit (Quantaurus-QY, Hamamatsu Photonics). We obtained the transient absorption spectrum using spectrofluorometers (200–1,000 nm: QE-Pro, Ocean Photonics; 900–1,400 nm: C9913GC, Hamamatsu Photonics), with the sample excited by a 340-nm LED (M340L4, Thorlabs) and probed with deuterium/tungsten–halogen lamps (L10290, Hamamatsu Photonics). Work functions were determined using photoelectron spectroscopy in air (AC-3, Riken Keiki).

LPL measurements

LPL spectra and decay profiles were obtained using the measurement system shown in Extended Data Fig. 7. The sample was placed in a cryostat (PS-HT-200, Nagase Techno-engineering) connected to a turbo molecular pump (HiPace80, Pfeiffer Vacuum), and excited by a 340-nm LED (as for the optical measurements) with a bandpass filter (340 ± 5 nm) for a certain time (1, 5, 10, 30, 60 or 180 s) with different excitation powers (10–500 μW). The emission spectra during (photoluminescence) and after (LPL) the excitation were recorded using a multichannel spectrometer (QE-Pro, Ocean Photonics). We obtained emission decay profiles using a Silicon photomultiplier (C13366-1350GA, Hamamatsu photonics) connected with a multimeter (34461A, Keysight).

Data availability

The data that support the findings of this study are available from the corresponding authors on reasonable request.