Modulating photoluminescence (PL) of light-responsive materials has aroused great interest because of its potential applications in super-resolution microscopy, displays, and optical memories^[1–7]. These materials are typically constructed by the combination of a high PL material and a function-switchable material that can alternate between two states, with only one state capable of quenching the PL. One approach is the combination of quantum dots (QDs) with diarylethene (DAE) molecular photoswitches to modulate PL through alternating UV and visible light irradiation^[8–16]. QDs as the PL materials feature excellent properties such as high PL quantum yields (PLQYs), flexible emission wavelength tunability, and narrow full width at half-maxima (FWHM) bandwidths^[17–19]. DAE molecules can be interconverted between the ring-open isomer (o-DAE) and ring-closed isomer (c-DAE) upon alternating irradiation with UV and visible light^[20–22]. DAE consists of two aryl groups connected by an ethene bridge, which can undergo a reversible photochemical reaction when exposed to UV and visible light. A cyclization process, where the ethene bridge forms a closed-ring structure, occurs upon UV light irradiation. The closed-ring form absorbs visible light, triggering a reverse reaction that breaks the ring and reverts it to the original open-ring structure. Their strong thermal stability and good reversibility make DAE molecules ideal switching materials for QDs^[23]. Optically switching the PL of QDs can be achieved through Förster resonance energy transfer (FRET), triplet energy transfer^[11] (TET), or charge transfer (CT), facilitated by spectral overlap or energy level alignment between QDs and c-DAEs, but not o-DAEs. Many optically switching PL systems have been reported by DAEs with CdSe/ZnS QDs^[24], CdSe/CdS/ZnS QDs^[10,25], CsPbX3 perovskite QDs^[9], and PbS QDs^[11]. However, the inherent toxicity of the above QDs limits their further application in biological and environmental detection. Therefore, it is essential to develop non-toxic PL switchable systems with a high PL on/off ratio (the PL intensity ratio between on and off states) and excellent reversibility under alternating irradiation cycles of different wavelengths.

Carbon quantum dots (CQDs) of less than 10 nm diameter, consisting of carbon atoms, have garnered significant attention in recent years as a non-toxic alternative to traditional semiconductor and perovskite QDs for bioimaging^[26–33]. CQDs exhibit unique quantum confinement and edge effects, leading to the emission of visible light when excited. Moreover, functional groups on the surface of CQDs allow for tuning the PL range, binding to targets, and enhancing their stability and solubility in various solvents^[34]. Although there are few reports on QD and DAE combinations for optically switching PL with a reasonable PL on/off ratio (up to 35)^[12,35], synthesized CQDs exhibit a low PLQY ($\sim 13\%$) compared with conventional QDs. The covalent-linked DAEs to the surface of CQDs further reduce the PLQY (from $\sim 12.8\%$ to 8.1%)^[35]. Importantly, the mechanism of optically switching PL of CQDs is far from being thoroughly investigated and fully understood. The claimed dynamic quenching such as FRET between CQDs and c-DAEs starkly contradicts the observed nearly unchanged PL lifetime in the PL on and off states. Dynamic quenching is a process in which the energy transfer or charge transfer occurs between a PL donor and a PL acceptor, with the ratio of decrease in PL lifetime being equivalent to the ratio of decrease in PL intensity^[36]. Therefore, it is highly desired to develop a simple yet efficient optically switching PL system based on CQDs and DAEs with a well-understood mechanism.

Herein, we demonstrate a novel optically switchable PL material consisting of commercially available CQDs and DAEs via a simple mixing approach, as shown in Fig. 1(a). Efficient PL on/off switching is observed when the hybrid material is alternatively illuminated with 310 and 515 nm lights. The PLQY of CQDs is as high as 75% and unaffected in the presence of o-DAEs, whereas it can be significantly quenched by c-DAEs, achieving a PL on/off ratio up to 500. Our design also exhibits strong reversibility under 20 cycles of switching between UV and visible light irradiation. The mechanism of our efficient approach was investigated by PL lifetime measurement, temperature-dependent PL spectroscopy, and density functional theory (DFT) calculations. For the first time, to the best of our knowledge, we demonstrate that static quenching and inner filter (IF) effects between CQDs and c-DAE are the predominant processes enabling efficient PL switching.

1,2-bis(2,4-dimethyl-5-phenyl-3-thienyl)-3,3,4,4,5,5-hexafluoro-1-cyclopentene (DAE) was purchased from TCI company. CQDs were purchased from Nanjing Muke Nanotechnology Co., Ltd. Both DAEs and CQDs had no ulterior depuration before use. Analytical reagent grade ethanol was purchased from AOPUSHENG chemical. All photophysical measurements were carried out in ethanol using 10 mm path quartz cuvettes.

A transmission electron microscope (TEM, Tecnai G2, the Netherlands) was used to confirm the morphology of CQDs. Fourier-transform infrared (FTIR) spectra were acquired using KBr pellets on a Nicolet 6700 spectrometer. UV–visible absorption spectra were measured on a Cary 60 UV–vis NIR spectrophotometer. Steady-state and temperature-dependent PL spectra were recorded on a Cary Eclipse fluorescence spectrophotometer. PL lifetime measurements were carried out using time-correlated single photon counting (TCSPC), which was excited by a 400 nm laser diode (PicoQuant) and recorded by an MCP-PMT detector (10 000 counts, 2048 channels). For light irradiation, a UV analytic lamp (4.5 mW) was used for 310 nm light irradiation, and an LED light source (Zolix MLED 4-3) was used for 515 nm light irradiation (12 mW).

The PLQY of CQDs was determined by comparison of the absorption and PL spectra of fluorescein as a standard reference. Upon excitation at the same wavelength, the PLQY ($\varphi$) can be calculated based on the following equation^[37]: $$\varphi ={\varphi }_r\frac{I}{I_r}\frac{A_r}{A},$$(1)where $I$ and $I_r$ are the integrated PL intensity of the tested sample and fluorescein, respectively, and $A$ and $A_r$ are the absorbances at the excitation wavelength of the tested sample and fluorescein, respectively. Both CQDs and fluorescein are measured in ethanol with ${\varphi }_r=0.87$ for the fluorescein.

DFT calculations of o-DAE and c-DAE were performed using the Gaussian 16 software package. Full optimization of the ground-state structure was performed by the hybrid functional B3LYP and basis set 3-21G(d).

The properties of selected CQDs and DAEs were first investigated individually. A TEM image of CQDs reveals a uniform spherical morphology with an average diameter of 3–6 nm [Fig. 2(a)]. FTIR spectroscopy was used to detect the carboxyl functional group on the surface of the CQDs, and the spectrum is shown in Fig. 2(b). The peak at $1647{\mathrm{cm}}^{-1}$ is ascribed to C = C stretching, $3400{\mathrm{cm}}^{-1}$ is ascribed to OH stretching of the -COOH group, $2920{\mathrm{cm}}^{-1}$ along with a sub-peak is ascribed to a C–H bond, and the stretching vibration of -C = O is located at $1750{\mathrm{cm}}^{-1}$. Figure 3(a) shows the UV–vis absorption (black) and PL spectra (red) of CQD in ethanol. CQDs exhibit a strong absorption band at 546 nm, which is attributed to $\pi -\pi *$ transitions of the aromatic ${\mathrm{sp}}^2$ system containing -C = O bonds^[38]. A small absorption peak at 350 nm is usually assigned to the $n-\pi *$ transition of conjugated -C = O. The PL spectrum of CQDs shows an intense emission peak centered at 570 nm upon 440 nm excitation with a PLQY of 75%. Figure 3(b) shows the absorption spectra of DAEs in ethanol. o-DAE molecules (black) absorb only in the UV region with a UV cutoff wavelength around 350 nm. Upon 310 nm UV light irradiation for 16 s, c-DAEs can be formed with the appearance of a visible absorption band with the peak maximum at 570 nm (red). The subsequent reverse switching from c-DAEs to o-DAEs occurs after being irradiated by 515 nm visible light for 5 min, evidenced by the recovery of the initial absorption spectrum.

The switching properties of DAEs were retained after mixing with CQDs. Figure 4(a) shows the absorption spectra of a solution of 2.5 µg/mL CQDs mixed with 50 µM DAEs in ethanol. Upon 310 nm UV irradiation, the appearance of absorption peaks in the visible region clearly indicates the formation of c-DAEs (red). Following irradiation at 515 nm, the absorption spectrum recovered to the initial state of o-DAEs (black). Adding o-DAEs into the solution of CQDs does not affect the PL characteristics, maintaining a PL peak at 570 nm and a PLQY of 75%, identical to CQDs alone. The PL of the mixed solution (2.5 µg/mL CQDs and 550 µM DAEs, approximately with the CQD/DAE mass ratio of 120) can be alternately switched between on and off states by exposure to UV and visible light, as shown in Fig. 4(b). Upon UV light irradiation, the PL intensity of CQDs is significantly reduced, achieving a 99.8% quenching efficiency and a PL on/off ratio of 500 (the normalized PL intensity at 570 nm compared to that of the closed-form 0.002). Irradiation at 515 nm fully restores the PL spectrum of CQDs. The PL on/off ratio in Fig. 4(b) is estimated to be 500, which outperforms all previously reported combinations of QDs and DAEs^{[9–11,25,35]}. To assess the reversibility of our design, repeated cycles of light irradiation alternating between 310 and 515 nm were performed. The PL intensity at 570 nm was monitored after each light irradiation step, as shown in Fig. 4(c). No noticeable fatigue after 20 switching cycles demonstrates the excellent reversibility of our design. The modulation efficiency of the system (the PL intensity ratio of the difference between on and off states to the on state) reached 99.8%.

To investigate the mechanism of our well-performed optically switchable PL system, PL intensity and lifetimes of CQDs with various concentrations of DAEs were carried out. Since PL quenching occurs specifically between CQDs and c-DAEs, our discussion below only focuses on the quenching behavior of c-DAEs. The concentration of CQDs in the mixed solution was kept constant (2.5 µg/mL), while the concentration of c-DAEs varied from 0 to 100 µM, and the corresponding PL spectra and lifetimes are shown in Figs. 5(a) and 5(b). The PL intensity of CQDs decreases with increasing c-DAEs, whereas the PL lifetimes of CQDs only exhibit a slight decrease from 2.9 to 2.8 ns. Although the spectra overlap between PL of CQDs and the absorption of c-DAEs allow for dynamic quenching such as energy transfer to occur for the PL quenching. Dynamic quenching is a time-dependent collisional process, where the excited state of the PL donor encounters the PL acceptors and relaxes to the ground state through a non-radiative pathway^[37]. Based on the PL lifetime measurements, the dynamic quenching contribution $E_{\mathrm{dyn}}$ in the quenching process can be estimated by $$E_{\mathrm{dyn}}=1-\tau /{\tau }_0,$$(2)where $\tau$ and ${\tau }_0$ represent the PL lifetimes of the CQD/DAE hybrid and CQDs alone, respectively. From the PL lifetime measurement, the dynamic quenching contributes only 3.5% to the overall quenching process.

The spectral overlap between QDs and c-DAEs can also result in the IF effect. The IF effect can manifest in two ways: 1) primary IF where the excitation light is absorbed by components within the sample, reducing the detected PL intensity; 2) secondary IF where the emitted light is reabsorbed by components within the sample, further decreasing the PL intensity. None of the two IF processes can change the PL lifetime of the donor. Since c-DAE has nearly no absorbance at the excitation wavelength of 440 nm, the first process of the IF effect can be ignored. The IF effect causing the emission photon reduction in PL intensity $E_{\mathrm{IF}}$ of our design can be calculated from $$E_{\mathrm{IF}}=1-\frac{I}{I_0}=1-{10}^{(-\frac{\frac{A_{570}}{2}}{2})},$$(3)where $I$ is the transmitted light intensity, $I_0$ is the incident light intensity, and $A_{570}$ is the absorbance at the PL peak of 570 nm in a 1 cm cuvette. The IF effect causing a reduction in PL intensity $E_{\mathrm{IF}}$ at 570 nm of CQDs mixed with 100 µM c-DAEs is 43.9%. However, 100 µM c-DAEs induced 82.8% PL quenching of CQDs as shown in Fig. 5(a), indicating that additional interactions occur between CQDs and DAEs that enable the non-radiative process.

We hypothesized that the high quenching efficiency observed in CQDs and c-DAEs is also due to the static quenching. Static quenching occurs when the PL donors and acceptors bind in their ground states and form non-emissive complexes, and the static quenching cannot change the PL lifetime of the donor. Static and dynamic quenching can be distinguished by their behaviors at different temperatures. Higher temperatures increase the kinetic energy of the molecules, reducing the likelihood of the stable complex formation while increasing the likelihood of molecular collisions. Therefore, static quenching is less favorable at higher temperatures compared to dynamic quenching, such as FRET. PL spectra of QDs and c-DAEs were further investigated at low and high temperatures (0°C and 70°C). Figure 6(a) shows the Stern–Volmer plots at 0°C and 70°C with varied c-DAE concentrations. The quenching rate constant $k_q$ can be calculated from $$\frac{I_0}{I}=1+{\tau }_0{k}_q[Q],$$(4)where $I_0$ and $I$ are the PL intensities of CQDs without and with c-DAE, respectively, and $[Q]$ is the concentration of c-DAE. The quenching rate constants $k_q$ at 0°C and 70°C are $1.44\times {10}^{13}$ and $1.21\times {10}^{13}{\mathrm{M}}^{-1}{\mathrm{s}}^{-1}$, respectively, indicating a less efficient quenching process at higher temperatures, consistent with the characteristics of static quenching^[39,40]. The $\pi$-electron-rich surface of CQDs can form strong $\pi -\pi$ stacking interactions with planar aromatic compounds, leading to complex formation. The $\pi$-conjugation of o-DAE is localized only in the thiophene rings, whereas c-DAE features a $\pi$-conjugation that is delocalized throughout the entire molecule^[41]. DFT calculations indicate that c-DAE adopts a more planar structure compared to o-DAE, with the height of the DAE molecule decreasing from 0.7 to 0.4 nm upon isomerization from o-DAE to c-DAE [Fig. 6(b)]. Consequently, the PL off state can be achieved through static quenching by the formation of CQD/c-DAE complexes via $\pi -\pi$ stacking, while o-DAEs allow the CQDs to retain their PL properties for the on state.

In conclusion, we have demonstrated an efficient optically switchable PL material consisting of commercially available CQDs and DAEs through a simple mixing approach. The PL of CQDs can be alternately switched on and off upon UV and visible light irradiation. The high PLQY of CQDs remains unaffected in the on state when DAEs are in the open form, while significant PL quenching occurs in the off state when DAEs convert to the closed form under UV light irradiation. Our design achieves a PL on/off ratio of up to 500 with excellent reversibility. Importantly, we investigated and revealed the mechanism behind this high performance. The PL quenching in the off state is primarily due to a strong IF effect and static quenching, with minimal contribution from dynamic quenching. These findings hold the potential of developing light responsive materials toward smart optoelectronics devices and imaging.