Day‐Long Persistent Luminescence in Intrinsically Integrated Donor‐Acceptor Carbon Dots Enabled by Defect‐Mediated Charge Trapping

ABSTRACT

Long‐persistent luminescence (LPL) materials capable of storing and releasing optical energy over extended timescales are highly desirable for next‐generation photonic technologies, yet structurally stable integrated donor‐acceptor (D‐A) systems capable of day‐scale, color‐tunable LPL remain elusive. Here, we report intrinsically integrated D‐A carbon dots exhibiting continuously tunable LPL from deep blue to yellow‐green, featuring day‐scale persistence of up to 36 h and a naked‐eye visible afterglow exceeding 4 h. Structurally, nitrogen‐doped carbon‐core donor domains are covalently coupled with arylboronic acid derived acceptor moieties through B─N linkages, forming an integrated D‐A framework enriched with intrinsic defect‐related trap states. Upon photoexcitation, intraparticle charge transfer (CT) generates long‐lived charge‐separated states, some of which are stabilized by intrinsic traps. Subsequent thermally activated detrapping releases the stored carriers and drives charge recombination, ultimately giving rise to day‐scale LPL. Furthermore, modulation of the acceptor electronic structure regulates the energy of the emissive CT state, enabling rationally tunable multicolor LPL. Benefiting from its day‐scale persistence and continuously tunable multicolor emission, this system enables potential applications in high‐resolution displays, dynamic anti‐counterfeiting, and intelligent information encryption. More importantly, this work establishes an intrinsically integrated D‐A‐trap design principle for ultralong LPL and a rational acceptor‐engineering strategy for multicolor persistent luminescence.

Carbon dots featuring intraparticle covalently integrated donor–acceptor and trap architectures enable continuously tunable multicolor long‐persistent luminescence with persistence up to 36 h, providing a general molecular design strategy for high‐performance afterglow materials.

1. Introduction

Long‐persistent luminescence (LPL), defined as light emission that continues for seconds to days after the removal of excitation, has attracted extensive interest owing to its wide‐ranging applications in emergency signage, security tagging, bioimaging, and low‐power optical devices [1, 2, 3, 4]. Since the discovery of the benchmark SrAl₂O₄:Eu^{2 +},Dy^{3 +} phosphor in the 1990s [5], a broad family of inorganic afterglow materials has been developed, featuring bright emission and efficient excitation under ambient or solar illumination [6, 7, 8]. However, their practical implementation remains constrained by high‐temperature synthesis, intrinsic brittleness, and potential environmental concerns [9, 10]. Organic LPL materials have recently emerged as a promising alternative, offering structural tunability, solution processability, and intrinsic biocompatibility [11, 12]. However, their persistent emission, typically arising from room‐temperature phosphorescence (RTP), is generally restricted to the millisecond–second timescale because triplet excitons are highly susceptible to nonradiative deactivation [13, 14]. The introduction of exciplex‐forming donor‐acceptor (D‐A) systems has provided an effective strategy to overcome this limitation and extend LPL lifetimes into the hour regime (Figure 1a). A seminal study by Adachi and co‐workers demonstrated that photoinduced charge transfer (CT) within a binary D‐A assembly can generate long‐lived charge‐separated states, whose radiative recombination gives rise to hour‐scale LPL [15]. Inspired by this concept, numerous binary D‐A systems have subsequently been developed through organic–inorganic hybridization, polymer doping, and trap‐state engineering to further enhance LPL performance [16, 17, 18, 19, 20, 21, 22, 23, 24, 25]. However, the physically separated D‐A architecture of these systems inevitably leads to phase instability and environmental sensitivity, compromising their structural robustness and long‐term reliability.

FIGURE 1.

Schematic illustration of the design principles and synthesis of D‐A CDs. (a) Conceptual comparison between conventional binary D‐A LPL systems and the present intrinsically integrated D‐A LPL system. (b) Design strategy for achieving prolonged LPL by synergistically coupling defect‐assisted charge trapping with an optimized intrinsic D‐A structure. Under UV excitation, singlet/triplet CT (CT¹/CT³) states and CS states are generated, while intrinsic trap sites capture a fraction of the electrons. Upon cessation of excitation, thermally released electrons repopulate the CT states and undergo radiative recombination, producing ultralong LPL. (c) Schematic illustration of the one‐pot thermal synthesis of D‐A CDs via urea and arylboronic acid derivatives (BDBA, 2‐biphenylboronic acid (2‐BBA), 1‐naphthylboronic acid (1‐NBA), and 4,4′‐biphenyldiboronic acid (4,4′‐BDBA)), yielding tunable LPL spanning from deep blue to yellow‐green.

To fundamentally address these limitations, intrinsically integrated D‐A architectures have recently emerged, where CT, charge separation (CS), and charge recombination (CR) are confined within a unified molecular or polymeric framework (Figure 1a). Tang et al. pioneered single‐molecule organic LPL materials based on covalently integrated D‐A architectures [26], and subsequent studies further extended this concept to polymeric systems [27, 28], establishing intramolecular CT and CR as a viable strategy for structurally stable LPL materials. Although covalent integration D‐A effectively eliminates phase separation and improves structural stability, the shortened D‐A distance inevitably accelerates CR following CS, limiting carrier‐storage lifetime and LPL duration. Therefore, developing structurally stable integrated D‐A systems that simultaneously suppress rapid CR and enable day‐scale persistent luminescence remains a major challenge.

Carbon dots (CDs) have recently emerged as an attractive platform for constructing metal‐free LPL systems owing to their low toxicity, excellent biocompatibility, superior photostability, and scalable synthesis [29, 30, 31, 32, 33, 34, 35, 36, 37, 38]. Unlike conventional molecular emitters, the electronic structures of CDs can be rationally engineered through precursor selection, heteroatom incorporation, and surface‐state regulation [39, 40, 41]. Such structural tunability provides a unique opportunity to simultaneously introduce donor‐like carbon domains, acceptor‐like surface moieties, and defect‐related trap states within a single nanoparticle. While D‐A interactions facilitate intraparticle CT and CS, defect‐related states can serve as long‐lived carrier‐storage reservoirs analogous to trap centers in inorganic persistent phosphors [42, 43, 44]. Therefore, we envisioned that integrating trap‐assisted carrier storage into an intrinsically integrated D‐A framework could effectively prolong CS by capturing photoinduced carriers and retarding CR, thereby enabling ultralong LPL (Figure 1b). Nevertheless, the deliberate construction of intrinsically integrated D‐A–trap architectures for day‐scale persistent luminescence remains largely unexplored.

Herein, we report intrinsically integrated D‐A CDs synthesized via a one‐pot melt reaction of urea and 1,4‐benzenediboronic acid (BDBA) (Figure 1c). The resulting D‐A CDs exhibit deep‐blue LPL centered at 422 nm, a visually detectable afterglow exceeding 4 h, and an ultralong persistence time of up to 36 h (instrumentally measured). Within this architectural, nitrogen‐doped carbon cores are covalently linked to surface BDBA chains through B─N bonds, forming an intrinsically integrated D‐A scaffold. Upon UV excitation, photoinduced electron transfer from the carbon‐core donor domains to the BDBA‐derived acceptor moieties generates long‐lived CS states, a fraction of which are further stabilized by intrinsic defect‐related trap states within the D‐A CDs. Following excitation cessation, the trapped carriers are gradually released through thermally activated detrapping and subsequently undergo CR to regenerate emissive CT states, thereby giving rise to ultralong persistent luminescence. Moreover, systematic modulation of the BDBA‐derived acceptor electronic structure enables continuous tuning of the LPL color from deep blue to yellow‐green, establishing a rational acceptor‐engineering strategy for predictable multicolor persistent luminescence. Together with its day‐scale LPL and continuously tunable multicolor emission, this system provides a versatile platform for advanced photonic technologies, including high‐resolution displays, dynamic anti‐counterfeiting, and intelligent information encryption.

2. Results and Discussion

2.1. Structural Design and Characterization of D‐A CDs

As a proof of concept, D‐A CDs were synthesized through a one‐step melt reaction by thoroughly grinding urea and BDBA prior to heating. The precursor selection was guided by a rational molecular‐design strategy aimed at integrating strong D‐A interactions with abundant defect‐mediated trap states within a single CDs framework. Urea was chosen for its high amine density and capability to introduce nitrogen into the carbon core, which imparts multiple critical functionalities: (i) it renders the carbon core a potent electron donor, (ii) generates diverse intrinsic defect states that serve as efficient electron‐trapping sites essential for sustaining ultralong LPL, and (iii) enhances intersystem crossing via heteroatom‐induced spin–orbit coupling, thereby increasing the triplet population necessary for persistent emission. Concurrently, BDBA provides electron‐deficient boronic acid moieties that covalently anchor to the CDs surface, functioning as robust acceptors. This deliberate structural integration constructs an intrinsic D‐A framework coupled with trap states, establishing a rational foundation for achieving ultralong‐lived LPL. Systematic optimization showed that heating at 230°C for 1 h with a urea‐to‐BDBA mass ratio of 150:1 affords the highest LPL intensity and longest lifetime (Figures S1–S3), and this condition was used in all subsequent experiments.

Transmission electron microscopy (TEM) reveals that the D‐A CDs are well‐dispersed spherical nanoparticles with an average diameter of 7.6 nm (Figure S4), whereas atomic force microscopy (AFM) shows a corresponding height of ∼5.5 nm (Figure S5), confirming their uniform nanoscale morphology. High‐resolution TEM shows no discernible lattice fringes, reflecting the intrinsically poor crystallinity of the particles, a feature that can be attributed to the abundant structural defects formed during the melt reaction. Consistently, x‐ray diffraction (XRD) exhibits a broad peak centered at ∼25°, characteristic of amorphous carbon, in sharp contrast to the well‐defined diffraction peaks of the urea and BDBA precursors (Figure S6).

To elucidate the chemical structure of the D‐A CDs, Fourier transform infrared (FTIR) and x‐ray photoelectron spectroscopy (XPS) analyses were conducted. The FTIR spectrum features ‒NH₂ and ‒OH stretching vibrations at 3464 and 3037 cm^{− 1} [45], alongside characteristic C─N and C═C bands at 1462 and 1400 cm^{− 1}. Peaks at 1356 and 1052 cm^{− 1} arise from B─O and B─C bonds, while a distinct band at 1591 cm^{− 1} corresponds to B─N stretching [46], providing direct spectroscopic evidence for covalent coupling between urea‐derived carbon cores and BDBA moieties (Figure S7). XPS analysis confirms the presence of C (39.7%), N (31.8%), B (4.6%), and O (23.9%). Deconvolution of the C 1s spectrum reveals peaks at 288.9, 285.8, 284.8, and 283.8 eV, corresponding to C═O, C─N, C─C/C═C, and C─B bonds, respectively (Figures S8–S11) [47]. The O 1s spectrum displays components at 532.9 eV (C═O), 531.8 eV (B–O) [48]. The N 1s spectrum comprises N─H (401.4 eV), N─C (400.7 eV), and N–B (399.8 eV) contributions, while the B 1s spectrum contains peaks at 192.4, 191.2, and 190.5 eV, assignable to B─O, B─N, and B─C bonds, respectively (Figure S11) [49]. The formation of B–N linkages is further substantiated by ^{1 1}B nuclear magnetic resonance (NMR) spectroscopy, which shows resonances from −10 to 40 ppm for B─O/B–C species and a pronounced peak near −20 ppm attributable to electron‐rich B–N bonds (Figure S12). To further verify the D‐A CDs structure, a physically mixed binary system consisting of urea‐derived CDs and BDBA (U‐CDs@BDBA) was prepared as a control. In contrast to the D‐A CDs, the characteristic B─N stretching vibration is absent in the FTIR spectrum of the physically mixed sample (Figure S13), confirming that the B─N linkages are generated during the synthesis process rather than arising from simple physical mixing. Collectively, the FTIR, XPS, ¹¹B NMR, and control‐experiment results consistently demonstrate that the D‐A CDs consist of nitrogen‐doped carbon‐core domains covalently connected to BDBA‐derived moieties through B–N linkages (Figure 1c). Such an atomically integrated carbon‐core/BDBA architecture establishes the structural foundation for the proposed D‐A framework and provides the basis for the CT and LPL behaviors discussed below.

2.2. Photophysical Properties and Mechanism Investigations of Day‐Long LPL

The photophysical behavior of D‐A CDs was comprehensively investigated to elucidate the processes responsible for their day‐scale persistent luminescence. Steady‐state photoluminescence (PL) measurements reveal a dominant fluorescence band at 383 nm with a short lifetime of 1.44 ns and negligible excitation dependence (Figures 2a, S14, and S15), confirming that prompt emission arises from singlet excitons within the D‐A CDs. In contrast, delayed PL spectra collected with a 10 ms delay exhibit four additional emission bands centered at 387, 487, 520, and 550 nm (Figures 2a and S16). These bands display lifetimes ranging from 235.3 to 1828.8 ms (Figure S17; Table S1), and all undergo progressive quenching with increasing temperature (Figure S18), establishing their phosphorescent origin. Among them, the 487–550 nm bands match the RTP signatures of BDBA, indicating triplet emission from surface‐anchored BDBA domains, while the 387 nm band originates from the intrinsic triplet emission of the carbon cores (Figure S19). To further verify this assignment, U‐CDs synthesized in the absence of BDBA were investigated as a control. U‐CDs exhibit a delayed‐emission band centered at ∼386 nm, which closely matches the corresponding phosphorescence band observed in D‐A CDs (Figures S20–S22). This result confirms that the ∼387 nm emission originates from the intrinsic carbon‐core triplet state and contributes to the short‐lived phosphorescence component of the D‐A CDs.

FIGURE 2.

Photophysical properties and LPL behavior of D‐A CDs. (a) Prompt and delayed PL spectra (10 ms and 20 s delays) of D‐A CDs under 310 nm excitation. (b) LPL decay profile monitored at 422 nm following 310 nm excitation (gray line: instrumental noise). (c) Time‐resolved LPL emission spectra collected at delay times ranging from 20 s to 36 h. (d) Photographs of D‐A CDs, C‐D‐A CDs, G‐D‐A CDs, and YG‐D‐A CDs (top to bottom) under 310 nm UV irradiation and after excitation cessation. (e) Logarithmic plot of LPL decay profiles of YG‐D‐A CDs, G‐D‐A CDs, and C‐D‐A CDs. (f) Energy gaps of BDBA, 2‐BBA, 1‐NBA, and 4,4’‐BDBA derived from HOMO and LUMO energy levels based on CV measurements.

A distinct long‐lived emission centered at 422 nm emerges when the delay time extends to 20 s after excitation. Remarkably, this emission remains detectable for up to 36 h, and its deep‐blue afterglow is visible to the naked eye for more than 4 h under ambient conditions (Figures 2c,d and S23, S24). The decay profile follows a Debye–Edwards power‐law:

$$I\left(t\right)\propto t^{-m}$$ (1)

where I is the emission intensity at 422 nm and t is the delay time. The characteristic kinetic behavior widely associated with CR–mediated persistent luminescence [25]. Double‐logarithmic analysis reveals an initial exponential decay (∼0–20 s), followed by a multi‐regime power‐law decay (Figure 2b). The extracted exponents (m = 1.22, 0.92, and 1.15 for 20–50 s, 50–1500 s, and 1500–100 000 s, respectively; Figure S25) indicate temporal evolution in the recombination kinetics [17]. These kinetic signatures are commonly associated with long‐timescale CR processes, which may involve trap‐depth‐dependent carrier dynamics. Notably, D‐A CDs display a broad absorption shoulder spanning 250–350 nm. This low‐energy absorption feature can be partially contributed by n→π* transitions of surface heteroatomic groups, and is predominantly attributed to intramolecular CT transition between the urea‐derived electron‐donating carbon core and covalently bonded BDBA acceptor moieties (Figure S26). Solvent‐polarity‐dependent PL further corroborates strong CT interactions within D‐A CDs (Figure S27), consistent with the formation of intramolecular CT states that contribute to the observed LPL emission. Interestingly, the U‐CDs@BDBA system also exhibits an LPL band centered at ∼418 nm (Figure S28), which closely matches the ∼422 nm LPL emission observed in D‐A CDs. This observation demonstrates that the characteristic LPL emissive channel emerges upon introduction of the BDBA component and is therefore closely associated with D‐A interactions between the U‐CDs and BDBA. However, the afterglow duration of the physically mixed system is limited to approximately 15 min (Figures S29 and S30), substantially shorter than that of the D‐A CDs. Collectively, these observations indicate that the 422 nm emission originates from a D‐A CT‐related emissive state formed between the carbon‐core donor and BDBA acceptor. The PL quantum yield (QY) of D‐A CDs was measured to be 47.5%. An estimated LPL efficiency was derived from the delayed‐emission analysis, and the corresponding calculation procedures, assumptions, and limitations are provided in the Supporting Information (Figures S31, S32). The D‐A CDs exhibit exceptional long‐term durability, maintaining nearly unchanged LPL properties across various atmospheric environments and even after 1 year of ambient storage (Figures S33–S35).

Beyond the D‐A CDs system, three additional D‐A CDs were synthesized by replacing BDBA with different arylboronic acid derived acceptors (2‐BBA, 1‐NBA, and 4,4′‐BDBA, respectively). As shown in Figures 2d and S36–S41, the resulting LPL colors can be continuously tuned from cyan and green to yellow‐green (denoted as C‐D‐A CDs, G‐D‐A CDs, and YG‐D‐A CDs). Notably, all of these systems exhibit pronounced ultralong LPL behavior with afterglow lifetimes extending to the hour scale (Figure 2e), demonstrating the generality of the D‐A structural design strategy (Figures S42–S44). More importantly, systematic electronic‐structure analyses reveal that the emission color evolution is closely correlated with the electronic characteristics of the acceptor units. As the electron‐accepting ability gradually increases, the calculated LUMO levels of the acceptors progressively decrease, leading to a continuous narrowing of the D‐A CT energy gap (Figures 2f and S45–S48; Table S2). Correspondingly, the LPL emission undergoes a gradual bathochromic shift from cyan to yellow‐green. These results not only highlight the versatility of the integrated D‐A architecture for constructing multicolor LPL CDs, but also provide further support for the CT‐mediated origin of the persistent luminescence.

To elucidate the mechanistic origin of the exceptional day‐long LPL in D‐A CDs, we systematically probed the excited‐state dynamics and the role of CS states in sustaining persistent emission. The dependence of LPL on excitation duration and intensity was first examined. As shown in Figure 3a, extending UV irradiation markedly prolongs the LPL: a brief 10 s excitation yields emission lasting ∼3000 s, whereas 5 min of irradiation extends the afterglow lifetime beyond 90 000 s. Further increases in irradiation time produce negligible enhancement, indicating saturation of charge‐carrier accumulation. Similarly, increasing UV intensity enhances both LPL duration and emission intensity at 422 nm, which follows a power‐law dependence on excitation intensity,

$$I_{Em}\propto I_{Ex}^n$$ (2)

where n = 0.46, consistent with CT‐based LPL systems (Figures 3b,c and S49) [17]. These observations directly indicate that photoinduced generation and accumulation of charge carriers within the D‐A CDs underlie the observed ultralong LPL emission. Electron paramagnetic resonance (EPR) measurements further confirm radical formation under UV irradiation, consistent with the formation of long‐lived CS species (Figure S50).

FIGURE 3.

Photophysical dynamics and electronic structure of D‐A CDs. (a, b) LPL decay profiles of D‐A CDs as a function of excitation time (a) and excitation power (b), plotted on a logarithmic scale. (c) Excitation‐power dependence of LPL emission intensity with corresponding linear fits. (d) fs‐TA spectra of D‐A CDs at selected pump–probe delay times. (e) fs‐TA spectral map of D‐A CDs under 320 nm excitation. (f) fs‐TA kinetic traces monitored at 570 nm. (g) Calculated electron and hole distribution of the D‐A CDs. (h) Electrostatic potential surface analysis of D‐A CDs.

Ultrafast excited‐state dynamics were investigated via femtosecond transient absorption (fs‐TA) spectroscopy. Notably, fs‐TA herein aims to resolve the initial ultrafast excited‐state evolution rather than the macroscopic minute‐scale LPL decay that involves slow charge trapping/detrapping. Upon 320 nm excitation, a negative signal at 420–520 nm appears within 1 ps, assigned to ground‐state bleaching and/or stimulated emission of intrinsic transitions of CDs (Figure 3d,e). Concurrently, a positive transient band centered at ∼570 nm emerges within ∼0.3 ps, which is significantly red‐shifted relative to the intrinsic absorption and is characteristic of CT‐related excited states rather than surface‐ or defect‐associated states typically located near the band edge. Global kinetic analysis further resolves three distinct temporal components: (i) an ultrafast 0.3 ps process corresponding to CT state formation; (ii) an 87.2 ps process attributed to relaxation toward a stabilized CT manifold; and (iii) a 2.3 ns long‐lived component assigned to the generation of CS species (Figure 3f). The combination of pronounced spectral red shift and well‐resolved sequential kinetics indicates that the dominant transient response originates from the CT/CS‐related pathway, while minor contributions from surface or defect states may coexist but do not govern the observed dynamics [19, 28, 50]. Importantly, the characteristic ∼570 nm transient band is absent in the control U‐CDs (Figure S51), indicating that this transient species originates from D‐A CT interactions introduced by BDBA rather than from intrinsic carbon‐core or defect‐related excited states. To further examine the electronic characteristics of the proposed D‐A architecture, time‐dependent density functional theory (TD‐DFT) calculations were performed on a representative structural motif constructed according to the experimentally identified carbon‐core/BDBA framework. The HOMO is predominantly distributed over the nitrogen‐doped carbon‐core/conjugated region, whereas the LUMO is preferentially localized on the BDBA‐derived moieties, indicating D‐A electronic polarization within the D‐A CDs (Figure S52). More importantly, natural transition orbital (NTO) analysis reveals that the hole density is mainly localized on the carbon‐core region, while the electron density is predominantly distributed over the BDBA‐derived units (Figure 3g), providing direct excited‐state evidence for photoinduced CT from the carbon‐core donor to the BDBA acceptor. Electrostatic potential mapping further shows electron‐deficient BDBA moieties and an electron‐rich carbon‐core region, consistent with their respective acceptor and donor roles (Figure 3h). These calculations support the CT characteristics of the experimentally established D‐A framework and are consistent with the proposed D‐A‐mediated excited‐state evolution in D‐A CDs.

Building upon these findings, we further investigated how the separated charges are stabilized and subsequently recycled to sustain day‐long LPL. To this end, we next examined the LPL behaviors of the samples under controlled thermal‐stimulation conditions. The D‐A CDs exhibit pronounced thermally activated LPL enhancement: emission intensity at 393 K increases nearly sevenfold relative to 293 K (Figure S53). Moreover, the LPL lifetime progressively declines with increasing temperature, indicative of a typical thermally activated detrapping process (Figure S54). Even after extensive natural decay, moderate thermal stimulation efficiently reactivates the emission, providing direct evidence for the existence of long‐lived trapped charge carriers. For instance, a sample pre‐irradiated for 5 min and stored in the dark at 293 K for 2 h exhibits only weak residual luminescence; however, upon gradual heating from 313 to 423 K, the emission intensity increases dramatically and reaches a maximum at approximately 373–393 K (Figure 4a). Notably, the temperature at which the emission peaks depend on the initial resting conditions, indicating a dynamically evolving population of trap states with distinct depths and activation energies (Figure 4a). Such pronounced thermal enhancement is a hallmark of thermally assisted detrapping processes that replenish emissive CT states, thereby sustaining the LPL. Quantitatively, after decaying for 1 h at 293 K, subsequent heating to 353, 373, and 393 K enhances the afterglow intensity by approximately 4.3‐, 9.9‐, and 13.8‐fold, respectively (Figure 4b), in good agreement with a trap‐release–mediated emission model. Near‐infrared (NIR, 980 nm) irradiation further corroborates the presence of trapped carriers: NIR stimulation accelerates detrapping and immediately boosts emissive recombination (Figure 4c). EPR measurements reveal that D‐A CDs retain strong radical signals even after 100 h following photoactivation, far exceeding those of unactivated controls (Figure 4d) [51, 52].

FIGURE 4.

Trap characteristics and mechanistic investigation of LPL in D‐A CDs. (a) Afterglow photographs of D‐A CDs recorded 2 h after excitation at different temperatures (left), corresponding thermally activated afterglow images (middle), and TL emission intensities at 422 nm after 5 min dark storage (right). (b) LPL decay curves under programmed thermal stimulation (293 K → 353/373/393 K, three times). (c) LPL decay profiles recorded with and without 980 nm NIR stimulation. (d) Electron paramagnetic resonance (EPR) spectra collected after UV excitation followed by dark storage for 15–100 h. (e) TL glow curves measured after 310 nm excitation and storage for durations ranging from 10 min to 100 h. (f) TL glow curves obtained after pre‐excitation at different temperatures, reflecting the temperature‐dependent trap population behavior. (g) Initial‐rise analysis of the TL glow curves used to extract trap activation energies. (h) Derived trap‐depth distribution of D‐A CDs. (i) Schematic illustration of the complete photophysical cycle of D‐A CDs, encompassing photoinduced CS, charge trapping, thermally assisted detrapping, and subsequent CR that collectively give rise to the LPL. (j) Jablonski diagram illustrating localized and CT excited states in D‐A CDs, highlighting long‐lived charge trapping and CT‐state recombination pathways responsible for day‐scale LPL.

This observation demonstrates that deep traps introduced by heteroatom doping within the carbon cores possess exceptionally high charge‐capturing capability, providing a structural and mechanistic basis for ultralong LPL exceeding 36 h under ambient conditions. To further map the trap‐energy landscape, thermoluminescence (TL) measurements were performed. A broad TL band spanning 300–500 K was observed (Figures 4e and S55), consistent with a continuous distribution of trap depths. Remarkably, TL signals persist even after 100 h of room‐temperature aging following UV irradiation (Figure 4e), in excellent agreement with the EPR results and the inferred longevity of CS states. Moreover, systematic TL measurements under different excitation temperatures reveal a gradual decrease in TL intensity accompanied by a shift of the TL maximum toward higher temperatures, indicative of hierarchically distributed shallow and deep traps (Figure 4f) [25]. Using the initial‐rise method, trap depths were quantitatively extracted [53]:

$$I_T=C_0·\exp \left(-\frac{E}{k_B·T}\right)$$ (3)

where C ₀ is a frequency‐dependent constant, E represents the trap depth, k _B is the Boltzmann constant, and T is the absolute temperature. By plotting ln(I) versus 1/T, the shallowest occupied trap depth for each TL segment can be derived from the linear slope (Figure 4g) [23]. Integration of TL intensities enable estimation of relative trap densities, revealing a broad, continuous trap‐energy distribution ranging from 0.105 to 0.629 eV (Figure 4h). This wide trap landscape underpins the gradual release of carriers, sustaining ultralong LPL emission through slow, thermally assisted radiative recombination. For comparison, identical TL measurements and trap‐depth calculations were also performed on the other three multicolor D‐A CDs. The extracted trap‐depth ranges are determined to be 0.11–0.483 eV for C‐D‐A CDs, 0.104–0.501 eV for G‐D‐A CDs, and 0.103–0.512 eV for YG‐D‐A CDs (Figures S56–S58). Notably, a clear quantitative correlation is observed between trap energetics and LPL performance: deeper trap distributions correspond to higher activation energy barriers, slower carrier detrapping kinetics, and consequently longer LPL durations.

To further elucidate the origin of the trap states responsible for the day‐long LPL, identical TL measurements were performed on the U‐CDs@BDBA system. In contrast to the D‐A CDs, U‐CDs@BDBA exhibits essentially no detectable TL signal under identical conditions (Figure S59), indicating the absence of an effective trap‐state distribution. These results suggest that D‐A interactions alone are insufficient to establish long‐lived charge‐storage capability in the absence of effective trap states. Considering that the most significant structural difference between these two systems is the formation of B–N linkages in D‐A CDs, the results further suggest that the covalently integrated B–N‐linked architecture plays a critical role in generating trap‐associated charge‐storage sites. Therefore, while D‐A interactions are responsible for the formation of the CT emissive state, the introduction of B–N‐associated trap states enable long‐term charge storage and delayed recombination, ultimately leading to the observed day‐long persistent luminescence. Integrating these findings with the previously established CS dynamics, a coherent mechanistic model for ultralong LPL in D‐A CDs is proposed (Figure 4i) [54, 55]:

1. Excitation and CS:

   $$\mathrm{D}-\mathrm{A}\mathrm{CDs}+hv\to {\mathrm{D}}^{·+}+{\mathrm{A}}^{·-}$$
2. Electron trapping by defect sites:

   $${\mathrm{A}}^{·-}\to \mathrm{A}+{\mathrm{e}}_{\mathrm{trap}}^{-}$$
3. Thermally stimulated detrapping:

   $${\mathrm{e}}_{\mathrm{trapped}}^{-}\stackrel{\mathrm{\Delta }}{\to }{\mathrm{A}}^{·-}$$
4. CR and delayed emission:

   $${\mathrm{A}}^{·-}+{\mathrm{D}}^{·+}\to \mathrm{D}-\mathrm{A}\mathrm{CD}{\mathrm{s}}^{\ast }\to \mathrm{D}-\mathrm{A}\mathrm{CDs}+h{v}_{\mathrm{LPL}}$$

Upon UV excitation, electrons transfer from carbon‐core donors (D) to BDBA acceptors (A), generating interfacial CT states that further evolve into long‐lived CS species (D^{• +}/A^{• −}). A fraction of the photoexcited electrons is subsequently captured by intrinsic defect traps within the carbon framework, leading to spatially separated trapped charge states, while the corresponding D^{• +} holes remain confined on the carbon cores. Upon thermal stimulation, the trapped electrons are gradually released and migrate back to recombine with D^{• +}, regenerating emissive CT states. Radiative decay of the singlet CT (¹CT) state to the ground state gives rise to the observed ultralong LPL (Figure 4j) [22]. This stepwise photophysical process, integrating intramolecular D‐A CT states, defect‐mediated electron trapping, and thermally assisted CR, provides a coherent structural and mechanistic framework that accounts for the efficient and stable day‐scale LPL observed in D‐A CDs.

2.3. Photonic Device Applications

By virtue of their long afterglow lifetime, color‐tunable persistent emissions, and thermally activated luminescence, D‐A CDs offer a versatile platform for a broad spectrum of functional applications (Figure 5). Integrated with a programmable light‐emitting diode (LED) array, D‐A CDs powders enable an energy‐independent optoelectronic signage module (Figures 5a and S60) that autonomously displays a bright and long‐lasting blue “EXIT” symbol upon power interruption, demonstrating their suitability for safety and emergency displays. Through offset printing, D‐A CDs powders can be deposited onto transparent adhesive substrates to produce high‐resolution afterglow patterns, such as a “NO SMOKING” label, that remain clearly visible for over 4 h (Figures 5b and S61), illustrating their potential for durable, high‐definition afterglow printing and anti‐counterfeiting. The intrinsic lifetime differences among multicolor D‐A CDs further allow temporal–spectral multiplexing (Figure 5c), where programmed optical patterns evolve in a predictable sequence from fluorescence to persistent deep blue, cyan, green, and yellow‐green emissions, enabling multistage information encoding and a lifetime‐gated Morse‐code encryption scheme that reveals distinct messages (“NTTT,” “NEET,” and the final encrypted output “TE”) at progressively delayed time windows (Figure 5d). Moreover, leveraging the pronounced thermally activated luminescence of D‐A CDs, a temperature‐programmable 3×3 coding matrix was constructed (Figures 5e and S62), in which spatially selective heating drives dynamic transformations of the optical output, from “‐21” at 333 K to “84” at 353 K and ultimately to “0” at 373 K, demonstrating the feasibility of thermally programmable encryption, multidimensional optical logic operations, and intelligent information processing enabled by D‐A CDs.

FIGURE 5.

Demonstration of LPL for photonic device applications. (a) Schematic illustration of the afterglow display of a safety sign (left) and a photograph of a LED device showing the “EXIT” sign after turning off the power source (right). (b) Schematic illustration of the printing and display process of the “NO SMOKING” pattern. (c) Spatiotemporal anti‐counterfeiting enabled by color‐tunable CDs. The pattern integrates D‐A CDs with distinct afterglow lifetimes: D‐A CDs (lower‐right petals), C‐D‐A CDs (upper‐right branches), G‐D‐A CDs (lower‐left branches), and YG‐D‐A CDs (upper‐left petals). (d) Optical information encryption using a Morse code encoder–decoder. Decoded images at different delay times illustrate the conversion of afterglow signals into Morse code, enabling retrieval of preloaded information. (e) Temperature‐dependent message reading and decryption based on D‐A CDs, where encrypted information is selectively decoded at different temperatures through thermally activated luminescence.

3. Conclusion

In summary, we have developed an intrinsically integrated D‐A CDs system via a straightforward one‐pot melt reaction between urea and BDBA. The resulting D‐A CDs exhibit deep‐blue LPL centered at 422 nm with an ultralong persistence of up to 36 h as measured instrumentally and a naked‐eye visible afterglow exceeding 4 h. Structurally, nitrogen‐doped carbon‐core donor domains are covalently linked to surface BDBA‐derived acceptor moieties via B─N bonds, forming an intrinsic D‐A framework. Upon photoexcitation, electrons transfer from carbon‐core donors to BDBA acceptors, generating long‐lived CS states (D^{• +}/A^{• −}), a fraction of which are stabilized by defect‐mediated traps. After excitation ceases, thermally assisted electron release enables gradual recombination, producing day‐scale LPL. Furthermore, systematic modulation of the arylboronic acid derived acceptor electronic structure regulates the energy of the emissive CT state, enabling continuous color tuning from deep blue to yellow‐green. Benefiting from its day‐scale persistence and tunable multicolor emission, the D‐A CDs show strong potential for applications in high‐resolution displays, dynamic anti‐counterfeiting, and information encryption. More importantly, this work establishes an intrinsically integrated D‐A‐trap design principle for ultralong persistent luminescence and a rational acceptor‐engineering strategy for multicolor LPL modulation, providing a general framework for the development of next‐generation persistent luminescent materials and intelligent photonic technologies.

Author Contributions

Hao Qiu: writing – review and editing, writing – original draft, software, data curation, formal analysis, validation, investigation, visualization. Heng Zhou: software, validation. Youquan Yan: software, formal analysis. Jinyang Li: software, formal analysis. Jialei Cui: software, validation. Zhenhua Gao: software, methodology. Zifei Wang: conceptualization, methodology, investigation, supervision, funding acquisition, project administration, resources, writing – original draft, writing – review and editing, visualization, formal analysis. Fanglong Yuan: writing – review and editing, project administration, supervision, funding acquisition, formal analysis.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: anie73529‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information of this article.