碳点作为纳米温度计的新一代材料：回顾

Mechanism of Thermo-sensing

Up to now, there is no well-established mechanism for explaining the thermo-sensing behavior of carbon dots. Some reports attribute the mechanism to the thermal activation of non-radiative channels of surface (trap/defect) states. The general picture is that the non-radiative channels were not activated at low temperatures, so the excited electrons could emit photons radiatively. On the contrary, as the temperature increases, more non-radiative channels became activated, and excited electrons got back to the ground state by non-radiative processes, leading to the decreasing fluorescence intensity [2, 95, 99, 100, 103]. The mechanism of CDs emissions with heating/cooling is shown in Fig. 10.

Schematic illustration of CDs responding to temperature changes

To better understand the thermodynamics of the CD emission processes, Kalytchuk et al. [89] correlated the radiative (\( {\tau}_r^{-1} \)) and nonradiative ( \( {\tau}_{nr}^{-1} \)) recombination rates of a CD sample with its PL quantum yield. The radiative rate is determined from the PL quantum yield (QY) and the measured recombination rate τ ⁻¹ as \( {\tau}_r^{-1} \) =QY × τ ⁻¹ . The nonradiative relaxation rate \( {\tau}_{nr}^{-1} \) is expressed as \( {\tau}_{nr}^{-1} \) =τ ⁻¹ - \( {\tau}_r^{-1} \). The PL QY of CDs at various temperatures was calculated from their temperature-dependent absorption and integrated PL intensity together with the PL QY determined at room temperature. Both radiative and nonradiative recombination rates derived from time-resolved PL measurement data are plotted as functions of temperature in Fig. 11. The radiative recombination rate is greater than the correspondent nonradiative rate up to 70 °C and does not vary appreciably at temperatures between 2 and 80 °C, remaining in the range of (0.74–0.82) × 10 ⁶ s ⁻¹ . In contrast, there is a pronounced (almost 7-fold, from 0.16 × 10 ⁶ to 1.12 × 10 ⁶ s ⁻¹ ) monotonic increase in the rate of nonradiative recombination by increasing the temperature from 2 to 80 °C. Temperature-dependent crossover of the radiative and nonradiative rates occurs at 70 °C, at which temperature the PL QY is 50%. These results suggest that the temperature activation of PL quenching in their CDs is primarily caused by the activation of nonradiative relaxation channels [89].

Radiative (solid symbols, blue color) and nonradiative (hollow symbols, red color) recombination rates for CDs plotted against the temperature for temperatures of 2–80 °C. (Reproduced from supporting information of reference [89])

Guo et al. [102] ascribed the thermal-quenching of their prepared carbon dots not just to activation of the nonradiative decay process, but also to the occurrence of nonradiative trapping with increasing temperature. They measured the temperature-dependent decay lifetimes of the CDs and shown in Fig. 12a. The data were collected by monitoring emission maximum as a function of the temperature under the 320 nm laser excitation. The result shows that the PL lifetime drops from 15.03 to 11.70 ns with the temperature increasing from 283 to 343 K, which could be ascribed to the occurrence of nonradiative decay processes. Besides, the PL relaxation dynamics of the CDs reveal multi-exponential decay with temperature increasing, which suggests the photoexcited carriers following the complicated relaxation processes. The occurrence of non-radiative trapping will be increased with rising temperature, and this could be quantitatively analyzed by the Arrhenius plot of the integrated PL intensities as:

Temperature-dependent decay curves of CDs solution (a , λex =350 nm, λem =450 nm); the dependence of ln[(I 0/I T)-1] on 1/kT CDs solution (b ）。 (Reproduced from reference [102])

其中 E a is the activation energy, k is the Boltzmann constant, and a is a constant. Figure 12a displays the plotting of the emission intensity with respect to 1/T , where the value of activation energy (E a) is calculated to be 0.329 ± 0.02 eV. In order to probe the reason for thermal quenching of CDs emission process, the radiative (V r) and nonradiative (V nr) recombination rates of CDs were determined from the lifetime (τ*) and quantum yield (QY) as:

The QY of CDs at various temperatures was calculated from their temperature-dependent absorption and integrated PL intensity with the QY determined at room temperature. They have noticed that the radiative rates have a slight decline when the temperature rises from 283 to 343 K; at the same time, the nonradiative recombination rates have gradually increased by about 2-fold. These results further indicate that the temperature-activated PL quenching in CDs is mainly due to the activation of nonradiative relaxation channels [102].

Other groups used microscopic and spectroscopic techniques to understand the mechanism of thermos-sensing of carbon dots.

Wang at el. used TEM and UV–Vis spectra to study the temperature-responsive PL behavior of prepared CDs. As shown in Fig. 13a, the CDs display no change in the UV–Vis spectra upon increasing the temperature from 20 to 80 °C. However, it was found that the average diameter of CDs increased from 2.6 ± 0.2 nm at room temperature to 4.4 ± 0.2 nm at 80 °C (Fig. 13b). Thus, increasing the temperature, the aggregation of as-prepared CDs occurred which caused the obvious fluorescence quenching [1].

一 UV–Vis absorption spectra of CDs in aqueous solution under 20 and 80 °C. b the TEM image of CDs in aqueous solution (a ) at room temperature, the average size was 2.6 ± 0.2 nm (b ) at 80 °C, and the size increased up to 4.4 ± 0.2 nm. (Reproduced from reference [1])

He et al. [101] reported that the hydration particle size of their CDs emerges as larger with the increase in the temperature (Fig. 14a), which indicates that the temperature rise gives rise to the aggregation of the CDs, eventually results in the fluorescence quenching. Nonetheless, with the decline in the temperature, the hydration particle size of CDs starts declining (Fig. 14b), which indicates that the cooling has the potential of causing CDs to depolymerize [101].

Change of hydrated particle size of carbon dots during heating (a ) and cooling (b ）。 (Reproduced from reference [101])

Another group such as Cui et al. also attributed the fluorescence quenching to the aggregation of CDs. They also tried to apply the undoped CDs synthesized using only acrylic acid as a precursor in temperature sensors. Unfortunately, undoped CDs possessed weaker quenching effects under the same temperature elevation than doped CDs [92].

杨等人。 [94] in their work proposed two key factors concerning the temperature-dependent PL property of the N-CDs, including (i) surface functional groups and (ii) hydrogen-bonding interaction. To examine the effect of the first factor, the surface O-containing groups, another control experiment was conducted by treating N-CDs (4.0 mL) with a strong reducing agent NaBH4 (1.0 mL, 0.1 mol L ⁻¹ ) to remove C=O species on carbon dots surface. The obtained reduced N-CDs are denoted as r-N-CDs for brevity. Compared with N-CDs, the r-N-CDs exhibit weaker fluorescence intensity (Fig. 15a). Besides, the fluorescence intensity of r-N-CDs only decreases by 13% with temperature increasing from 20 to 80 °C (the inset in Fig. 15a) that gives a much lower temperature sensitivity, which is ascribed to the decreased O-containing groups [94].

一 PL spectra (excitation wavelength, 400 nm) of N-CDs (black trace) and r-N-CDs (red trace). The inset shows I /I 0−T of reduced N-CDs. b PL spectra (excitation wavelength, 400 nm) of N-CDs dispersed in C2H5OH at various temperatures. c A schematic mechanism for the temperature-dependent fluorescence intensity of N-CDs. (Reproduced from reference [94])

The second factor, the effect of hydrogen bonding with the solvent on fluorescent behavior of N-CDs was explored. N-CDs solution (1.0 mL) was dropped on a filter paper and left to dry in the air to obtain a solid sample, which still emits bright fluorescence. However, no obvious change of fluorescence intensity of the solid N-CDs with temperature increase was observed. They also measured the fluorescence of N-CDs dispersed in ethanol. The fluorescence intensity of N-CDs in C2H5OH is lower than that in water and little variation of the PL intensity is observed with temperature increasing from 20 to 80 °C (Fig. 15b). Hence, the strong hydrogen bonds play a key role in the temperature-dependent PL property of the N-CDs. Figure 15c is a schematic mechanism for the temperature-dependent fluorescence intensity of N-CDs [94].

However, our group used the same experimental strategy as Yang group; in both cases, the r-CDs and e-CDs emissions were quenched linearly with increasing temperature, in the same way as the original results of their CDs (Fig. 16). Thus, our results ruled out the synergistic effects of abundant oxygen-containing functional groups and hydrogen bonds [77].

一 , b Fluorescence spectra of reduced CDs (r-CDs) and CDs in ethanol (e-CDs) at temperatures (20 to 60 °C). c , d Linear correlation between fluorescence intensity and temperature (°C) for r-CDs and e-CDs respectively. (Reproduced from the supplementary information of reference [77])

Bioimaging in Living Cells (Thermal Imaging)

In literature, only a few articles explored the temperature-responsive fluorescent properties of CDs in biological imaging. Prior to such experimentations, in vitro cytotoxicity analysis is crucial for CDs because they make it possible to estimate the CD’s toxicity in living subjects. In vitro cytotoxicity analysis evaluates the effect or influence of the nanomaterial on cultured cells [89].

杨等人。 [94] verified that the CDs could be used as an effective thermometer in living cells; HeLa cells were washed with PBS after treatment with CDs for 6 h. Bright-blue fluorescence of the CDs in HeLa cells is observed when the temperature is 25 °C as shown in Fig. 17a. With the temperature increasing to 37 °C, the blue fluorescence becomes weaker (Fig. 17b), the fluorescence of N-CDs is recovered when the temperature was decreased to 25 °C (Fig. 17c). As a thermos-imaging in vivo, the fluorescent images of mice were collected immediately after being injected with N-CDs at different temperatures. By setting the fluorescence intensity at 28 °C as the reference (I o), I /I o of the area where CDs were injected varies from 1.0 to 0.87 with the increase of temperature from 28 to 34 °C (Fig. 17d, e). With temperature further increasing to 43 °C, I /I o declines to 0.52 and the fluorescence becomes nearly undetectable (Fig. 17f). And I /I o can be reversibly enhanced back from 0.66 to 1.0 with the temperature decreases from 39 to 28 °C (Fig. 17g–i). All of these results indicated that N-CD could be used as an effective in vitro and in vivo nanothermometer [94].

a–c Fluorescent images of a single Hela cell at 25, 37, and 25 °C after treatment with N-CDs, respectively. d –f Fluorescent photographs of a mouse given an injection of N-CDs at increasing temperatures. g –我 Fluorescent photographs of a mouse given an injection of N-CDs at decreasing temperatures. ( Reproduced from reference [94])

In an exploratory experiment, Kalytchuk et al. tested the capacity of CDs for intracellular temperature monitoring in human cervical cancer HeLa cells. Figure 18 shows the measured intracellular temperatures of HeLa cells incubated with CDs (500 μg/mL). The CDs’ PL decay curves at each temperature were highly reproducible and could be fitted with a single-exponential function at all recorded temperatures. In Fig. 18a, the recorded PL decay curves for temperatures between 25 and 50 °C (with a step size of 5 °C) are indicated by symbols, while the corresponding single-exponential fits are represented by solid lines. They were able to confirm that the PL signal in these PL lifetime measurements was derived exclusively from the CDs for the CD concentration of ≥ 100 μg/mL. The intracellular temperature in each measurement was determined from the calibration curve between PL lifetimes as the temperature increases from 2 to 80 °C. The temperatures determined in this way (T meas) are plotted as functions of the PL lifetime in Fig. 18b. Independently, the temperature of the cell solution was determined using a calibrated reference thermometer (shown in Fig. 18b as Tset). The temperatures reported by the luminescent CD probe and the reference detector are in good agreement. These results show that the PL lifetime of nanoprobes based on CDs can be reliably used to measure intracellular temperatures [89].

In vitro intracellular PL lifetime thermal sensing using CDs. 一 PL emission decays of HeLa cells incubated with CDs (500 μg/mL) at different temperatures (T set). b Temperatures determined using the calibration (T meas) and set temperatures (T set) plotted against the PL lifetime. c –f Applicability of CDs for long-term remote intracellular temperature monitoring. c PL lifetimes extracted from PL transients recorded every 15 min for 24 h of HeLa cells incubated with CDs (500 μg/mL). d Temperatures determined using the calibration curve. e Temperatures measured with a reference thermometer (T ref). f Histogram showing the distribution of temperature differences between the obtained and reference temperatures; the solid line is the distribution curve. (Reproduced from reference [89])

To further evaluate the potential of luminescent CD nanoprobes for long-term real-time temperature monitoring, PL decay profiles of HeLa cells incubated with CDs (500 μg/mL) every 15 min for 24 h were recorded. The PL lifetimes extracted from the measured PL decay values during this period were then plotted as functions of time, as shown in Fig. 18c. Using these results, the temperature variation over time was calculated using the calibration curve, as shown in Fig. 18d. In addition, the sample’s temperature at each measurement point was determined using a reference thermometer with a temperature reproducibility of 60 mK. The temperature determined with the reference thermometer is plotted in Fig. 18e, which shows that there was excellent agreement between the measured and reference temperatures. The high accuracy of the PL-based temperature measurements is further demonstrated by statistical analysis of the differences between the measured and real (reference) temperatures (Fig. 18f). Using these data, the absolute average accuracy of temperature detection by the presented method was calculated to be 0.27 °C. This experiment confirms the potential of CD-based thermal probes in biological systems [89].

Macairan et al. [29] displayed that the prepared CDs can be used for thermal sensing inside cells using intensity and ratiometric approaches. HeLa cells treated with CDs were allowed to equilibrate at 32, 37, and 42 °C (Fig. 19). Excitation at 640 nm was used to selectively monitor the red fluorescence of the CDs in the cells. The thermal changes could be due to changes in intracellular concentration or not correlate with a change in intensity (λex + 640 nm). This could be due to changes in intracellular concentration or localization of the CDs at higher temperatures. Thus, simply relying on changes in fluorescence intensity leads to accurate intracellular thermal sensing.

Fluorescence microscopy images of CD-treated HeLa cells. Fluorescence signals from the CDs (λex =640 nm; left and 405 nm; right) fluorescence ratios are 1.8 at 32 °C, 2.0 at 37 °C, and 2.3 at 42 °C. The control shows untreated HeLa cells at 42 °C with no fluorescence signal as expected. (Reproduced from reference [29])

In contrast, these limitations are excluded using the ratiometric approach. The CDs maintain dual blue and red fluorescence in cells following excitation at 405 nm, as previously observed for the colloidal dispersions (Fig. 19). The red-to-blue ratio increases with increasing temperature, with values of 1.8 at 32 °C, 2.0 at 37 °C, and 2.3 at 42 °C. The ratiometric relationship of the red-to-blue fluorescence of the CDs highlights the advantage of ratiometric temperature sensing in the development of fluorescent nano-thermometry probes. The relative red-to-blue emission ratio remains unaffected, regardless of the amount of CDs taken up by the cells, which can be affected by various factors such as confluency and is not concentration-dependent. Lastly, the CDs have shown fluorescence reversibility with respect to changes in intracellular temperature. Following incubation in HeLa cells, they were subjected to a heating/cooling cycle from 32/42/32 °C. This emphasizes the robustness of the proposed CD-nanothermometer and these findings further demonstrate the fluorescence reversibility [29].

Confocal laser scanning microscopy was used to thermal image colon cancer cell HT-29 using N, S, and I-doped CDs, as shown in Fig. 20 [100]. The fluorescent spots were temperature-dependent as shown in Fig. 20g–i, as the most intense is at 15 °C, while the weakest point was at 35 °C. Interestingly, the fluorescence spots were reversible, and the spots were very photostable after 20 min of continuous excitation [100]. Shin et al. [96] also used confocal laser scanning microscopy for Hela cells, shown in Fig. 20e–g.

a–d Are confocal microscopy images of N, S, I-CDs-colon cancer cell HT-29 with corresponding fluorescence field at 15, 25, 35, and 15 °C, respectively. (Reproduced from reference [100]). e –g Confocal microscopy images of N, S-CDs-stained cells with corresponding fluorescence field at 25, 35, and 25 1C, respectively. (Reproduced from reference [96])

李等人。 103 prepared a nanocomposite composed of MnOx-CDs to be used as a nano thermos responsive fluorophore for biological environments. HepG2 cells were incubated with MnOx-CDs at different culture temperatures. As illustrated by confocal laser scanning microscopy (excited at 405 nm), the blue luminescence of the MnOx-CDs in the HepG2 cells is weak when the environmental temperature was 40 °C (Fig. 21a), and the blue luminescence of the MnOx-CDs in HepG2 cells enhances as the temperature decreased to 30 °C (Fig. 21b) even to 20 °C (Fig. 21c). Due to the temperature-responsive properties, the as-synthesized MnOx-CDs can be readily applied in the biomedical fields like bioimaging and photothermal therapy in cancer treatment [103].

Confocal laser scanning microscopy images (excited at 405 nm) of MnOx-CDs in HepG2 cells at 40 °C (a ), at 30 °C (b ), and 20 °C (c ）。 (Reproduced from reference [103])

Conclusions and Future Perspective

Carbon nanodots exhibit unique properties to be exploited for nanothermometry, such as thermal-sensitivity, low-cost, and photostability. Flexible surface modification and facile preparation will pave the way to establish an enormous number of thermal sensitive nanomaterials for a variety of applications. The overall trends in thermo-sensing nanomaterials are aimed at enhancing photostability and thermal-resolution with using low-cost and safe materials. CDs can be classified as a new generation of thermometer that can fulfill these requirements and can be used for biomedical thermometry applications, such as temperature monitoring during hyperthermia treatment. Facile-preparation protocols, biocompatibility, and easy functionalization of CDs are promising criteria which make the CDs alternative next-generation nanothermometer materials. More efforts are required to promote basic research in this field. Limitations should be overcome to produce carbon dot-based nanothermometers comprising enhancing thermal sensitivity, and working in a broader range of temperature. A better understanding of the fluorescence thermal-sensing mechanism is another key issue to be able to design and manipulate the structure of CDs and enhance thermal resolution. More experiments and theoretical modeling are necessary to understand the correlation between the methods of fabrication of CDs with their thermal behavior.