Introduction

Organic delayed luminescent materials have garnered significant attention owing to their diverse optical properties1,2,3,4,5, including prolonged emission lifetimes and high signal-to-noise ratios, which are crucial for applications in various fields like organic light-emitting diodes (OLEDs)6,7,8,9, sensors10,11, anti-counterfeiting12,13,14,15, and time-gated bioimaging16,17,18,19. Among these materials, thermally activated delayed fluorescence (TADF) and room-temperature phosphorescence (RTP) represent two prominent types of delayed luminescence. However, they are often considered competitive, as both processes originate from triplet excitons20,21,22,23,24. Hence, designing materials with TADF and RTP dual-mode afterglow poses significant challenges, as it is difficult to develop a unified platform that can meet their distinct requirements and achieve multifunctional luminescence. Moreover, the limited mechanistic understanding of dual-mode delayed emission further impedes the rational design of multifunctional organic luminophores, despite the compelling advantages that such systems could offer25,26,27,28.

Typically, efficient TADF relies on rapid reverse intersystem crossing (RISC), which is facilitated by spatial separation of the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO), thereby minimizing the energy gap between the first singlet (S1) and triplet (Tn) excited states. A widely adopted molecular design strategy involves the incorporation of twisted donor-acceptor architectures29,30,31,32,33. For RTP, persistent RTP emission relies on two key factors: (i) enhancing spin-orbit coupling (SOC) through the introduction of heteroatoms or heavy halogens to accelerate intersystem crossing (ISC)34,35,36,37, and (ii) suppressing nonradiative decay of triplet excitons by restricting molecular motion, often achieved via crystallization, polymerization, metal-organic frameworks, or host–guest doping strategies38,39,40,41,42,43,44,45,46,47.

Previous studies indicate that tailoring the molecular packing or conformation can modulate the photophysical properties of multifunctional luminescent materials at the single-molecule level48,49,50,51. With the growing importance of host-guest doping strategies in the development of TADF and RTP materials, a deeper understanding of the underlying luminescence mechanisms has become critical for advancing multifunctional smart materials26,52. In host-guest systems, intermolecular charge transfer (CT) or triplet-triplet energy transfer has been identified as key processes governing organic RTP53,54,55,56. However, TADF is rarely observed in organic doped systems, even though doping often facilitates the efficient generation of triplet excitons57,58,59,60,61. It is challenging to develop a doped system that generates both TADF and RTP, not to mention efficient TADF and near-infrared (NIR) RTP for advanced applications in OLED or bioimaging (Fig. 1a).

Fig. 1: Schematic diagram.
figure 1

a Comparison of luminescence mechanism elucidation in doped systems between previous studies and this work. b Strategy for constructing efficient purely organic dual-mode delayed emission materials via host-guest doping.

Full size image

In this study, we selected several conjugated polyaromatic hydrocarbon molecules as guests, and benzophenone (BP), a typical electron acceptor, as the host to construct doped luminescence systems. To our surprise, dual-band delayed emission integrating both TADF and RTP was observed, attributed to CT and charge-recombination (CR) processes (Fig. 1b), as revealed by theoretical calculations and spectroscopic analyses. This strategy enabled long-lived, multicolor-delayed luminescence spanning from blue to near-infrared (NIR) regions. Ultrafast spectroscopy further confirmed the formation of triplet charge-transfer (³CT) states through the simultaneous observation of radical anions and cations, thereby validating the role of CT states as an “energy redistribution hub” that facilitates efficient TADF and red RTP via RISC and internal conversion (IC) processes. Notably, the high-performance red RTP is well retained when the doped systems are converted into nanoparticles, marking an advancement toward bioimaging and related applications.

Results

Ten representative polyaromatic hydrocarbons (PAHs) with strong fluorescence—benzo[c]phenanthrene (BecPh), benzo[a]phenanthrene (BePh), fluoranthene (FlAn), benz[a]anthracene (BeAn), picene (Pi), dibenz[a,h]anthracene (BeTe), pentaphene (DBePh), dibenzo[g,p]chrysene (DBeCh), benzo[ghi]perylene (BePe), and coronene (Co)—were selected as guest molecules, with benzophenone (BP) serving as the host (Fig. 2a), based on theoretical calculations (Supplementary Fig. 1 and Supplementary Table 1). As expected, these ten molecules have abundant excited-state properties, endowing the potential of delayed luminescence. Excited-state energy and SOC calculations revealed a notably high SOC value in BP (20.45 cm−1; for S1 → T2), promoting efficient ISC and triplet generation. In contrast, the PAHs showed negligible SOC values, indicative of their intrinsically low ISC efficiency. Given the well-aligned energy levels and potential for RTP, ten donor-acceptor doped systems (BecPh-BP, BePh-BP, FlAn-BP, BeAn-BP, Pi-BP, BeTe-BP, DBePh-BP, DBeCh-BP, BePe-BP, and Co-BP) were proposed for further investigation. Although the large energy gap between the HOMO of the PAHs molecules and the LUMO of BP hindered intermolecular charge transfer at the ground state (Supplementary Fig. 2), the electronic cloud distribution of the frontier molecular orbitals of host-guest molecule pairs showed that intermolecular CT in the excited state between the guest and host is feasible (Supplementary Figs. 3, 4, and Supplementary Table 2). Accordingly, all ten doped materials were successfully prepared using a simple and efficient melt-casting method, with a host-to-guest molar ratio of 100:1.

Fig. 2: Photophysical properties of the doped systems.
figure 2

a Molecular structure of the host and the ten selected guest molecules. b Photographs of the doped powders under daylight, 365 nm UV light, and after cessation of UV excitation. Guest names are shown above each sample, with visually observable afterglow durations indicated below. c Steady-state and delayed PL spectra of ten PAH–BP doped systems (delay time: 1 ms). d Time-resolved decay curves of phosphorescent emission peaks for the ten doped powders at room temperature. (excitation wavelength: 365 nm; doping ratio: 100:1).

Full size image

First of all, the photophysical properties of the host-guest systems were systematically studied. Under 365 nm UV irradiation, all doped powders exhibited fluorescence-RTP dual emission as shown in Fig. 2b, the afterglow emission can be observed from 3 s (FlAn-BP, BeAn-BP, DBePh-BP) to 28 s (Co-BP). Photoluminescence (PL) spectra revealed well-defined emissions for the ten doped powders closely corresponding to the fluorescence emission peaks of the guest powders (Supplementary Fig. 5). Additionally, delayed PL spectra recorded at varying delay times further demonstrated their afterglow characteristics (Fig. 2c and Supplementary Fig. 6). All doped systems exhibited two afterglow bands spanning 380–500 nm and 500–800 nm. The first band overlapped with the steady-state emission, and its intensity exhibited a temperature-dependent increase followed by a decrease, confirming its origin as TADF (Supplementary Fig. 7). The second emission band (500–800 nm) matched well with the phosphorescence spectra of the respective guests in 2-methyltetrahydrofuran at 77 K, indicating that the RTP originated from the triplet excited states of the guest molecules (Supplementary Fig. 8). The lifetimes of ten host-guest doped powders’ phosphorescent emission peaks at room temperature are 961, 598, 379, 211, 546, 678, 308, 620, 401 and 1814 ms (Fig. 2d and Supplementary Table 3). Among them, Co-BP and BecPh-BP exhibited notably longer lifetimes than other systems, which may be attributed to the more stability of the triplet state of Co and BecPh. Furthermore, the decay profiles of both the TADF and phosphorescence peaks indicated that the RTP lifetimes in the doped systems are significantly longer than those of TADF (Fig. 2d, Supplementary Fig. 9 and Supplementary Table 3).

To gain deeper insight into the luminescent mechanism of the doped systems, femtosecond transient absorption (fs-TA) and nanosecond transient absorption (ns-TA) spectroscopy were employed to study the excited state evolution of BePe and BePe-BP as representative systems. For BePe, two distinct excited-state evolution processes were observed. Initially, the two main excited state absorption (ESA) peaks at 425  and 522 nm rapidly increased, reaching their maxima at 112 ps. Subsequently, these peaks gradually decayed, accompanied by the emergence of a new ESA peak at 470 nm. The ESA peak at 425 nm corresponds to the characteristic peak of the excited singlet state of BePe (Supplementary Fig. 10). The ESA intensity at 522 nm, including the shoulder peak at 552 nm, increased with the BePe concentration, indicating the absorption peak of the BePe’s excimer (Supplementary Fig. 11).

Subsequently, the excited-state evolution of doped BePe-BP was investigated using fs-TA and ns-TA spectroscopy. As shown in Fig. 3a, the initial transient absorption spectrum of BePe-BP closely resembled that of pristine BePe, featuring a peak at 425 nm characteristic of BePe excitation and a peak at 522 nm attributed to BePe’s excimer. Significant changes were observed in the fs-TA spectra from 53.6 ps to 1.32 ns. The ESA peak at 425 nm gradually decreased, while the peak at 520 nm increased rapidly, and a new ESA peak emerged at 470 nm. After 1.32 ns, the 425 nm peak continued to decay, the 520 nm peak decreased sharply, and the 470 nm peak increased significantly. The peak at 520 nm was identified as the ESA of the BePe cation radical, as its intensity was rapidly quenched upon the addition of NaI (Supplementary Fig. 12)62,63, demonstrating an excited state CT between BePe and BP.

Fig. 3: Complete excited state evolution in the doped system.
figure 3

a Fs-TA and (b) ns-TA spectra of BePe-BP in DCM under 365 nm excitation. c Ns-TA decay profiles of BePe–BP at 470 nm under N2 and O2. d Decay curves at 520 nm from fs-TA and ns-TA spectra of BePe–BP in DCM. e Ns-TA decay curves at 520 nm for BePe–BP under N2 and O2 atmospheres. (BePe: 0.01 M, BP: 0.1 M).

Full size image

The fs-TA spectra of BePe-BP revealed a strong ESA peak until 7 ns, prompting further investigation using ns-TA to capture the complete excited state evolution process (Fig. 3b). Fortunately, the ns-TA spectra following excitation at 365 nm detected a new broad ESA band at 778 nm after expanding the detection wavelength range. Literature comparisons confirmed that this ESA peak is characteristic of the BP anion radical64,65, further supporting the presence of excited-state CT. These findings are exciting as they are crucial for understanding the impact of CT states on luminescent properties and comprehensively understanding the evolution of excitons in doped systems. Beginning at 8 ns, the ESA peaks at 520 nm and 778 nm decayed rapidly, while the 470 nm peak continued to rise reaching its maximum at 273 ns. Importantly, a minor shoulder peak at 425 nm was also observed, influenced by the dominant 470 nm peak, although partially obscured by the dominant 470 nm band. These results suggest that during the CR process, the CT state can relax not only to the excited triplet state of BePe but also to its singlet excited state—a key mechanistic origin of TADF in the doped system. Subsequently, all ESA peaks decayed rapidly and approached baseline levels by 17.4 us.

Kinetic analysis of the 470 nm ESA decay under nitrogen and oxygen atmospheres (Fig. 3c) revealed a markedly faster decay in the presence of oxygen, indicating its triplet-state nature. Bi-exponential fitting of the decay curves (Supplementary Fig. 13) yielded a long-lived component of 2.68 μs under nitrogen, which decreased sharply to 0.16 μs under oxygen, further corroborating the assignment of the 470 nm ESA band to a triplet excited state. Particularly, it is noteworthy that the intensity of the 470 nm ESA signal intensity in the BePe-BP system was significantly stronger than that in pristine BePe (Supplementary Fig. 14), indicating that CT states markedly enhance triplet exciton generation in BePe. The decay dynamics at 520 nm were also analyzed using combined fs-TA and ns-TA data (Fig. 3d), revealing four distinct processes. The first component, τ1 (24.4 ps), was assigned to the formation of the BePe radical cation, corresponding to a singlet charge-transfer state (¹CT). The second, τ2 (394 ps), represents the ISC from ¹CT to the triplet charge-transfer state (³CT). In the ns-TA, τ1’ (10.9 ns) was assigned to either the transition from ³CT to a localized triplet state of BePe or the RISC from ³CT to the BePe singlet state. This component also showed strong oxygen sensitivity (Fig. 3e), further supporting its assignment to ³CT. Moreover, both τ2 and τ1′ increased with solvent polarity, indicating that the ¹CT state is more stabilized in polar media and thus exhibits slower decay via CR. The final component, τ2’, corresponds to the long-lived decay of the ³CT state (Supplementary Fig. 15). Similar fs-TA behaviors were observed for other doped systems, including BecPh–BP, DBeCh–BP, BeAn–BP, BeTe–BP, and Co–BP (Supplementary Figs. 16–25), indicating a generalizable role of CT states in mediating the transition between singlet and triplet excited states in PAHs.

Based on the fs-TA spectral results and supporting experimental evidence, the exciton transition in host-guest materials can be systematically elucidated. For PAHs molecules, owing to their rigid π-conjugated structures and π-π* electronic transitions, exhibit strong fluorescence under UV excitation. (Fig. 4a, Supplementary Fig. 26 and Supplementary Fig. 27) UV-Vis absorption spectrum, confirm that no ground-state CT occurs in the doped systems, as evidenced by the absence of band broadening or new tail absorption features. (Fig. 4b and Supplementary Fig. 28) Upon UV excitation, excited-state intermolecular CT proceeds much faster than fluorescence emission, resulting in nearly complete fluorescence quenching in solution (Fig. 4a). However, PL spectra of the doped powders display two distinct afterglow bands. Taking the DBeCh-BP system as an example (Fig. 4c), the first emission band (400–500 nm) matches that neat DBeCh, with a slight blue shift attributed to TADF, consistent with the results shown in Fig. 2c. The second band (550–750 nm) aligns with the phosphorescence of DBeCh in dilute solution at 77 K, indicating emission from the triplet excited state. Fortunately, transient absorption measurements further allow us to track the evolution of both singlet and triplet excitons. As shown in the fs-TA spectra, at 7 ns, the ESA peaks corresponding to the singlet (528 nm) and triplet (451 nm) states are significantly stronger in the DBeCh–BP system than in the pristine DBeCh. (Fig. 4d) This indicates that intermolecular CT of the doped system not only promotes the formation of the excited triplet state of guests but also its singlet state. More notably, the triplet state formation is substantially enhanced, as evidenced by the ns-TA decay curves (Fig. 4e), where the triplet ESA signal at 451 nm in the doped system is nearly three times stronger than that in the guest-only sample. Together, these findings reveal the dual-emission mechanism in the doped systems. Upon UV excitation, rapid intermolecular CT between guest and host generates a stabilized ³CT state that efficiently stores excitation energy. During the CR process, the ³CT state may undergo RISC to produce singlet excitons, giving rise to TADF, or IC to the guest’s triplet state, leading to highly efficient RTP (Fig. 4f and Supplementary Table 3). More importantly, according to Marcus theory, the rate constant k for an electron transfer process (or charge recombination process) can be described as Eq. 1:

$$k={{\rm{A}}}{{\rm{\cdot }}}\exp \left[-\frac{{(\lambda+\Delta {{\rm{G}}})}^{2}}{4\lambda \,{k}_{B}{{\rm{T}}}}\right]$$
(1)

Where: kB is the Boltzmann constant, and T is temperature; A is a pre-exponential factor related to electronic coupling. The transition rate between excited states—such as RISC or IC from the intermediate ³CT state—is governed by the reorganization energy (λ) and the Gibbs free energy change (ΔG), the latter of which is closely related to the energy gap (ΔE) between ³CT and S1 or T1. In our doped systems, the rigid host BP matrix and the delocalized π-systems of PAHs help maintain a low λ, further enhancing the efficiency of both ISC and RISC depending on the energy alignment. Additionally, the ³CT state serves as an excitation energy redistribution hub, mediating decay toward either the singlet or triplet state of the guest. When ΔE between ³CT and S1 is small and λ is low, RISC to S1 is facilitated, enabling TADF; conversely, larger ΔE favors IC to T1, giving rise to RTP. Thus, the energetic positioning of the ³CT state and its coupling with the emissive states of the guest dictate the balance between TADF and RTP, offering a tunable pathway for dual-mode organic afterglow emission.

Fig. 4: Mechanistic insights into dual-mode delayed luminescence under ambient conditions.
figure 4

a Luminescence photographs (left) and emission spectra (right) of DBeCh and DBeCh–BP in DCM solution, and DBeCh–BP in the solid state under 365 nm UV irradiation. b UV-Vis absorption and (c) normalized steady-state PL spectra of DBeCh, BP and DBeCh-BP powders. d Fs-TA spectra of DBeCh and DBeCh-BP recorded at 7 ns. e Ns-TA spectra decay curves of peak in 451 nm of DBeCh and DBeCh-BP in N2 and O2. f Proposed mechanism of dual delayed luminescence integrating TADF and RTP of this host−guest system.

Full size image

Based on the inspiring dual delayed emission characteristics of the host-guest doped systems, their potential for time-dependent data encryption was further explored. As shown in Fig. 5a, three representative doped crystalline powders were assembled to form the emblem of Shantou University. Under ambient light, the pattern appears white; upon 365 nm UV irradiation, it displays a vivid multicolor image. Following cessation of UV excitation, the emission colors evolve with time: the afterglow of the “auspicious clouds” and “phoenix” motifs remain visible for 3 s, while the “phoenix” persists for up to 10 s. These distinct time- and color-resolved luminescent features highlight the potential of these materials for advanced multilevel data encryption applications.

Fig. 5: Demonstration of dual-mode delayed luminescence materials for data encryption and bioimaging applications.
figure 5

a Luminescence images of patterns composed of dual delayed emission doped materials under daylight, 365 nm UV excitation, and at various time intervals after UV removal. b Fabrication process of host-guest@F127 NPs via a top-down approach. c Diameter distribution of DBeCh-BP@F127 NPs. Inset: TEM image (scale bar: 100 nm). d Delayed PL spectra (delay time: 1 ms) and (e) phosphorescent decay curves of DBeCh-BP@F127 NPs in H2O (5 mg mL−1) under the 365 nm excitation. f IVIS bioluminescence imaging after cessation of 365 nm UV irradiation (15 mW cm−2). DBeCh–BP@F127 NPs (black dashed circle) were injected into the right abdomen and blank F127 NPs (red dashed circle) served as control on the left side. Signal-to-background ratio for phosphorescence imaging in live mice. Data are presented as mean ± SD (n = 3 independent experiments).

Full size image

Furthermore, the good tissue penetration capability of the red afterglow, with a maximum emission peak at 635 nm, renders it highly suitable for bioimaging applications. To this end, biocompatible nanoparticles (NPs) were fabricated using DBeCh–BP as the core and PEG-b-PPG-b-PEG (F127) as the encapsulating matrix via a top-down approach (Fig. 5b). Dynamic light scattering (DLS) and transmission electron microscopy (TEM) analysis revealed that the prepared NPs exhibited a uniform spherical morphology with an average size of approximately 200 nm (Fig. 5c). Notably, the red afterglow of DBeCh-BP@F127 NPs was clearly observable in an aqueous solution, with a long lifetime of 0.47 s at 635 nm (Fig. 5d, e). This long-lasting red RTP from purely organic NPs makes them an ideal candidate for bioimaging. Consequently, DBeCh-BP@F127 NPs were employed for in vivo imaging of live mice. Subcutaneously injecting DBeCh-BP@F127 NPs into Balb/c mice, the afterglow bioimaging images were successfully captured using the IVIS instrument in bioluminescence mode, following 2 min of exposure to a handheld 365 nm UV lamp (Fig. 5f). The subcutaneous phosphorescence signal exhibited a high signal-to-background ratio (SBR) of 170, demonstrating that the red RTP emission can be detected without continuous excitation, thereby eliminating interference from biological autofluorescence. These results highlight the great potential of purely organic RTP NPs for biological imaging and related applications.

Discussion

In summary, we have successfully employed a strategy of constructing CT states to achieve efficient dual delayed emission combining TADF and RTP in host-guest doping systems. Upon UV excitation, intermolecular CT occurs between the guest and host molecules, forming stable CT states that act as excitation energy reservoirs. The 3CT states function as excitation energy redistribution hubs, mediating exciton transitions within the doped systems. During the CR process, ³CT states can convert into the triplet or singlet excited states of the guest molecules via IC or RISC. Importantly, the direct observation of ³CT states and the complete exciton evolution pathway using fs-TA and ns-TA spectroscopy provides critical mechanistic insight into the origin of dual-mode delayed luminescence in these systems. Owing to their exceptional luminescent performance, these materials were further applied in time-dependent multilevel data encryption. Moreover, DBeCh–BP@F127 NPs, fabricated via a top-down approach, exhibited bright red RTP at 635 nm with a lifetime of 0.47 s, enabling high-contrast in vivo imaging. This work not only provides a general and versatile design strategy for purely organic dual-emission materials but also broadens their application scope in information security and biomedical imaging.

Methods

Material science

Benzophenone (BP, 99%), benzo[c]phenanthrene (BecPh, 95%), benzo[a]phenanthrene (BePh, 99%), fluoranthene (FlAn, 98%), benz[a]anthracene (BeAn, 98%), picene (Pi, 99%), dibenz[a,h]anthracene (BeTe, 98%), pentaphene (DBePh, >95%), dibenzo[g,p]chrysene (DBeCh, 98%), benzo[ghi]perylene (BePe, 98%), and coronene (Co, >94%) were obtained from J&K Scientific Ltd, all materials are purchased from commercial sources and purified by recrystallization.

Preparation of the host-guest doping dual-mode delayed luminescence powders

Different polyaromatic hydrocarbons (PAHs) as the guest and BP as the body were mixed at a molar ratio of 1:100 and then heated to 70 °C. After complete mixing, the mixture was cooled and crystallized, and the corresponding doped powder was obtained after grinding.

UV-vis absorption experiments

The UV-Vis absorption spectra of the samples were recorded by a spectrometer (PE Lambda950) from the PERKINELMER company.

Photoluminescence (PL) experiments

Prompt and delayed emission spectra, quantum efficiencies, as well as lifetimes, were recorded on an Edinburgh FLS1000 luminescence spectrometer.

Femtosecond transient absorption (fs-TA) experiments

Fs-TA measurements were performed with an apparatus and methods detailed previously and only a brief description is provided here66. The fs-TA measurements were done by using a femtosecond regenerative amplified Ti: sapphire laser system in which the amplifier was seeded with the 120 fs laser pulses from an oscillator laser system. The laser probe pulse was produced with ~5% of the amplified 800 nm laser pulses to generate a white-light continuum (350–800 nm) in a CaF2 crystal and then this probe beam was split into two parts before traversing the sample. One probe laser beam went through the sample while the other probe laser beam went to the reference spectrometer to monitor the fluctuations in the probe beam intensity. After that, 365 nm was selected as the excitation light, and the excitation power was 0.2 mW. The single-wavelength kinetics fitting was used based on Eq. (2). The τ was referred to as delay time, and τ0 means zero time. The τp and A represented the instrument response value and the proportion of species respectively.

$${{\boldsymbol{S}}}\left({{\boldsymbol{t}}}\right)={{{\boldsymbol{e}}}}^{-{\left(\frac{{{\boldsymbol{t}}}-{{{\boldsymbol{t}}}}_{{{\mathbf{0}}}}}{{{{\boldsymbol{t}}}}_{{{\boldsymbol{p}}}}}\right)}^{{{\mathbf{2}}}}}{{\boldsymbol{*}}}\mathop{\sum}\limits_{{{\boldsymbol{i}}}}{{{\boldsymbol{A}}}}_{{{\boldsymbol{i}}}}{{{\boldsymbol{e}}}}^{-\frac{{{\boldsymbol{t}}}-{{{\boldsymbol{t}}}}_{{{\mathbf{0}}}}}{{{{\boldsymbol{t}}}}_{{{\boldsymbol{i}}}}}}$$
(2)

Theoretical calculations

The theoretical calculations are carried out using the Gaussian 16 software packag(es67. B3LYP is selected as the function, and 6–31 g(d) is selected as the base set to optimize the structures and analyze the energy level. The structures of the host-guest complexes are shown in Multiwfn software68.

Preparation of nanoparticles

To 1 mL of the aqueous solution of F127 (10 mg), the DBeCh-BP powders (1 mg) were added. The mixture was then sonicated by a microtip-equipped probe sonicator (Branson, S-250D) for 10 min. The resultant suspension was filtered through a 0.45 μm syringe-driven filter to afford a solution of nanoparticles. And then the resultant solution was concentrated.

Animals

Female BALB/c mice (4–6 weeks old) were group-housed (5 per cage) under a 12 h light/12 h dark cycle (lights on at 07:00 h, lights off at 19:00 h) in a temperature- (22 ± 2 °C) and humidity-controlled (50 ± 10 %) environment. Food and water were provided ad libitum. At the conclusion of each experiment, mice were euthanized by CO2 inhalation followed by cervical dislocation, in accordance with institutional animal care guidelines.

Bioimaging measurement

The amphiphilic copolymer PEG-b-PPG-b-PEG (F127) was purchased from Aladdin Ltd. Transmission electron microscopy (TEM) images were acquired from a JEM-2010F transmission electron microscope with an accelerating voltage of 200 kV. Dynamic light scattering (DLS) was measured on a 90-plus particle size analyzer. In vitro and in vivo phosphorescence imaging was performed by the IVIS® Lumina II imaging system. All animal experiments (approval #A2024193) were performed in accordance with the guidelines for the care and use of laboratory animals and approved by the Animal Ethics Committee of Shanghai Jiao Tong University (#A2024193).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.