Photoresponsive oxidase-like phosphorescent carbon dots in colorimetric Hg²⁺ detection

2.1 Materials and instrumentations

The details of materials and instruments were shown in Supporting Information.

2.2 Synthesis of P-NCDs

P-NCDs were synthesized hydrothermally. Into a beaker, first 0.79 g of TTHA and then 30 mL of ultrapure water were added and fully mixed under ultrasound. The miscible solution was transferred to a polytetrafluoroethylene reactor. Reaction proceeded in an air dry oven at 180 °C for 5 h, and then the system cooled down to room temperature. The resulting light yellow solution was filtrated through a microporous membrane (0.22 μm) and then dialyzed via a dialysis membrane (MWCO:1000 Da) for 24 h, with ultrapure water changed every 4 h. Finally, the solution in the dialysis membrane was collected and freeze-dried to form white solid products, which were exactly P-NCDs.

2.3 Photoresponsive oxidase-like activity of P-NCDs

The photoresponsive oxidase-like activity of P-NCDs was studied through direct oxidation of TMB. First, 50 μL of an acetic acid-sodium acetate (HAc-NaAc) buffer solution (0.2 M, pH 5.4), 100 μL of a TMB solution (10 mM), and 50 μL of the P-NCDs solution (20 mg/mL) were successively added to a cuvette, which was diluted to 1 mL with ultrapure water. Then the mixture was irradiated under a 365 nm UV lamp for 30 min (2 min on/2 min off). Finally, the oxidase-like activity of P-NCDs was investigated by detecting the absorbance at 652 nm. Each experiment was repeated three times.

2.4 Steady-state dynamics analysis of P-NCDs

The steady-state dynamics of P-NCDs were analyzed under the optimal conditions in a system containing 1 mg/mL P-NCDs, and the HAc-NaAc buffer solution (10 mM, pH 5.4) with different concentrations of TMB. Based on the dynamics data, the Michaelis-Menten parameter was calculated from the Lineweaver-Burk plots: $$V=V_{\mathit{max}}[S]/(K_m+[S])$$

where V and [S] are the initial velocity and TMB concentration respectively, K_m is the Michaelis-Menten constant, and V_max is the maximum reaction velocity.

2.5 Effects of scavenger type on catalytic activity of P-NCDs

To each sample solution, a specific species of active oxygen trapping agent was added to observe its effect on the activity of P-NCDs. The sample solutions were: HAc-NaAc buffer solution (10 mM, pH 5.4), P-NCDs (1 mg/mL), TMB (1 mM), and a specific trapping agent, such as mannitol (5 mM), tryptophane (2.5 mg/mL), catalase (200 U/mL), superoxide dismutase (300 U/mL) and ethanol (10 % vt). Then the mixed solutions were irradiated under a 365 nm UV lamp for different periods of time.

2.6 Colorimetric detection of Hg²⁺

First, a Cys solution (60 μL, 400 μM) and Hg²⁺ solution of a specific concentration were added to 50 μL of an HAc-NaAc buffer solution (0.2 M, pH 5.4). The mixed solutions reacted at room temperature for 10 min. Then P-NCDs (50 μL, 20 mg/mL) and TMB (100 μL, 10 mM) were added to each of the resulting mixtures above. Each mixed solution was diluted to 1 mL with ultrapure water and further reacted for 10 min. The resulting solutions were irradiated under the 365 nm UV lamp for 15 min, followed by observation of UV–vis spectra.

2.7 Analysis of real samples

Tap water (from our laboratory) and lake water (from a lake in our campus) were sampled and purified through 0.22 μm microporous membranes. The resulting samples were used for analysis. Urine was collected from a healthy volunteer and diluted 100 times, without any other pretreatment.

3.1 Synthesis and characterization of P-NCDs

With TTHA as the only raw material, P-NCDs with a dense core and some low-polymer long-chains on the surface were prepared via a one-step hydrothermal method (180 °C, 5 h), including high-temperature dehydration, condensation, crosslinking, and polymerization carbonization (Fig. 2A). The morphology and structure of the P-NCDs were observed via TEM and XRD. The P-NCDs are spherically shaped and evenly dispersed (Fig. 2B) in average grain diameter of ∼ 3.25 nm (Fig. 2C) and without obvious lattice fringes, indicating the surface of P-NCDs may have a polymer structure. The XRD patterns of P-NCDs show an evident broad peak at 22 ° (Fig. 2D), which further proves the existence of polymer carbon-dot and indicates the P-NCDs have an amorphous structure. The hydrodynamic size of P-NCDs was measured using dynamic light scattering (DLS) to be 90 nm (Fig. S1A), indicating the P-NCDs have high water solubility. The hydration particle size of P-NCDs is far larger than the particle size measured by TEM, which may be because the surface low-polymer long chains of P-NCDs are significantly hydrated and thereby extended. The zeta potential of P-NCDs is 8.03 mV (Fig. S1B).

Close

Fig. 2

(A) Schematic diagram about formation of P-NCDs. Characterization of P-NCDs: (B) TEM; (C) particle size distribution; (D) XRD spectra; (E) FT-IR spectra; (F) XPS spectra; (G) C 1 s, (H) N 1 s, (I) O 1 s high-resolution spectra.

Export to PPT

FTIR and XPS demonstrate the surface structure and chemical composition of P-NCDs. The peak at 3442 cm⁻¹ represents the surface O—H and N—H of P-NCDs. The peaks at 3022, 1728, 1632, 1398, and 1224 cm⁻¹ correspond to the stretching vibrations of C—H, COOH, C⚌C/C—N, C—O—C, and C-OH respectively (Fig. 2E). XPS further verifies the FT-IR results. P-NCDs show three characteristic peaks of C 1 s (286.0 eV), N 1 s (402.0 eV), and O 1 s (532.0 eV) (Fig. 2F). The high-resolution C 1 s spectrum splits into three peaks at 284.6, 285.9, and 288.2 eV (Fig. 2G), which correspond to C—C/C⚌C, C—N/C—O, and COOH, respectively. The high-resolution N 1 s spectrum shows two peaks at 399.3 and 401.5 eV (Fig. 2H), which stand for pyrrole nitrogen and N—H respectively. The O 1 s spectrum splits into two peaks at 531.2 and 533.1 eV (Fig. 2I), indicating the surface of P-NCDs may contain C⚌O and C-OH. The above results indicate the surface of P-NCDs is rich in O- and N-containing groups.

3.2 Optical properties and mechanism of P-NCDs

During the irradiation under the 365 nm UV lamp, the P-NCDs water solution emitted blue fluorescence (inset in Fig. 3A), without phosphorescence, and the fluorescence was excitation-dependent (Fig. S2A). The fluorescence attenuation curve was fitted using a double-exponential function, which showed the average lifetime was 7.17 ns (Fig. S2B). These photoinduced fluorescent properties indicate P-NCD may have multiple luminescence centers/channels. The absolute fluorescent quantum yield under excitation at 345 nm is 15.04 %.

Close

Fig. 3

(A) Normalized UV–vis spectrum (Abs), fluorescence (FL) optimal excitation and emission spectra of the P-NCDs solution. (B) fluorescence (FL) and phosphorescence (Phos.) spectra of P-NCDs powder under the 365 nm UV lamp. Inset: photos of P-NCDs powder in the environment with the 365 nm UV lamp on or off. (C) normalized optimal phosphorescence (Phos.) excitation and emission spectra of P-NCDs powder. (D) 3D phosphorescence spectrum of P-NCDs powder. (E) phosphorescence attenuation spectrum and fitted curve of P-NCDs powder.

Export to PPT

The P-NCDs powder were white in the environment, but emitted blue fluorescence under the ultraviolet lamp. After the ultraviolet radiation was stopped, the P-NCDs powder showed obvious green RTP (inset in Fig. 3B). The offset of the fluorescence from the phosphorescence emission center is 80 nm (Fig. 3B), indicating the energy gap between the singlet state and the triplet state is about 0.45 eV (Deng et al., 2013). In addition, the source of RTP emission from P-NCDs was studied according to the UV–vis and excitation spectra of P-NCDs (Fig. 3A and 3C). The UV–vis absorption spectrum of P-NCDs shows wide bands at 275–400 nm, which may be ascribed to the n-π* transition of C⚌O/C⚌N (Deng et al., 2013; Jiang et al., 2020; Li et al., 2016; Nie et al., 2021; Tao et al., 2018; Wu et al., 2022). The presence of C⚌O/C⚌N can improve spin–orbit coupling and thereby promote inter-system crossing (ISC), forming more triplet-state excitons and producing RTP (Jiang et al., 2018b; Nie et al., 2021; Wang et al., 2021). The wider phosphorescent excitation band at 250–450 nm and the narrower fluorescent excitation band at 250–400 nm are largely overlapped with the absorption band of P-NCDs at 275–400 nm, indicating the fluorescent and phosphorescent emission of P-NCDs may come from the C⚌O/C⚌N of P-NCDs. Fig. 3D shows the 3D phosphorescence spectrum of P-NCDs powder. Compared with the fluorescence spectrum, the RTP emission evidently red-shifts. The P-NCDs powder also shows the excitation dependence as the P-NCDs solution. The phosphorescence attenuation spectrum of P-NCDs was fitted using a three-exponent function, and the lifetime was 140.07 ms (27.48 %), 627.21 ms (55.08 %) and 6086.33 ms (17.44 %) (Fig. 3E), suggesting the phosphorescence irradiation of P-NCDs originated from multiple luminescence centers. The average attenuation lifetime of P-NCDs was calculated to be 1.45 s.

Based on the above analyses, the RTP mechanism of P-NCDs can be predicted. First, the high-temperature dehydration, condensation, crosslinking and polymerized carbonization of triethylenetetramine hexaacetic acid form specific sub-fluorescent functional groups (C⚌O/C⚌N) responsible for phosphorescence emission, including the dense core and some low-polymer long-chains (Fig. 2A). The dense core and long chains of P-NCDs stably excite the triplet state and inhibit nonradiative transition, facilitating the phosphorescence emission of P-NCDs (Jiang et al., 2018a; Zhu et al., 2015). Then, TEM and XRD show the amorphous P-NCDs may contain a certain type of intertwined polymers, which act as the substrate to fix the luminescence center (Gao et al., 2020; Jiang et al., 2018b). FTIR and XPS demonstrate the surface of P-NCDs is rich in –OH and –COOH, which provide effective binding sites for forming the hydrogen bond and hydrogen bond frames of P-NCDs. These changes effectively prevent the triplet state excitons from being quenched by atmospheric water and oxygen and increase the probability of phosphorescent emission. In addition, TTHA, the only raw material, has both a carbon source and a doped nitrogen source. The introduction of N facilitates n-π* transition and effectively promotes intersystem crossing, leading to an increase in effective triplet excitons and promoting phosphorescence emission (Gao et al., 2018b; Li et al., 2016). In all, the dense core, low-polymer long-chains, rigid hydrogen bonds, and nitrogen doping jointly promote the phosphorescent emission of P-NCDs.

3.3 Photoresponsive oxidase-like activity of P-NCDs

The photoresponsive oxidase-like activity of P-NCDs was studied with TMB as the coloration substrate. Under irradiation with the 365 nm UV lamp, the P-NCDs catalytically oxidized colorless TMB into blue TMB cation free radical (ox-TMB), and an intense peak appeared at 652 nm (Fig. 5A), which prove the oxidase-like activity of P-NCDs. In comparison, without P-NCDs or irradiation, no blue ox-TMB was formed, and no peak appeared. The above results confirm that P-NCDs have photoinduced oxidase-like activity. To make P-NCDs achieve the optimal oxidase activity and to maximize the Hg²⁺ colorimetric sensing performance, we studied the effects of pH, P-NCDs concentration, TMB concentration, reaction temperature, and irradiation time on the catalytic activity of P-NCDs. Under irradiation, the absorbance of the pure TMB solution (curve a) first increased (the solution was obviously blue at pH 3.8–4.2) and then decreased (the solution did not change in color at pH 4.6–5.8) with the rise of pH (Fig. 4A). After P-NCDs were added to the TMB solution (curve b), the absorbance first rose and then dropped with the increment of pH. The difference of absorbance before and after the addition of P-NCDs (curve b-a) was calculated and found to be maximized at pH 5.4, and the solution before the addition of P-NCDs was colorless, indicating the errors due to coloration can be avoided at pH 5.4. The catalytic activity of P-NCDs was maximized at the concentration of 1.0 mg/mL (Fig. 4B) and at 35 °C (Fig. 4D). As the TMB concentration or irradiation time was increased, the catalytic activity of P-NCDs was further enhanced (Fig. 4C, E). After comprehensive consideration of detection sensitivity and analytical speed, we determined the optimal reaction conditions to be: room temperature, pH 5.4, 1 mg/mL P-NCDs, 1 mM TMB, 365 nm UV lamp irradiation for 15 min. Then the stability of P-NCDs was tested under the optimal conditions. Even after 60 days of storage, the P-NCDs still maintained high catalytic ability (Fig. S3).

Close

Fig. 4

Effects of (A) pH, a: TMB + Light, b: TMB + P-NCDs + Light; (B) P-NCDs concentration; (C) TMB concentration; (D) reaction temperature, and (E) irradiation time on catalytic activity of P-NCDs. Inset: photos of samples.

Export to PPT

Close

Fig. 5

(A) UV–vis spectra of samples under different conditions, including TMB/P-NCDs + Light, TMB/ P-NCDs, TMB + Light, and TMB. Inset: photos of samples. (B) continuous on-and-off of the light source, and ladder-like behavior of oxidizing activity of P-NCDs. (C) Michaelis-Menten curve and (D) Lineweaver-Burk curve of catalytic activity of P-NCDs with TMB as the substrate.

Export to PPT

When the lamp was repeatedly turned on or off, the oxidase-like activity of P-NCDs was enhanced in a ladder-like way (Fig. 5B), indicating P-NCDs have the optically controlled oxidase activity. The photocontrolled oxidase activity of P-NCDs was further studied via steady-state dynamics analysis. Fig. 5C shows the Michaelis-Menten curve and Lineweaver-Burk double-reciprocal curve (Fig. 5D). With the Lineweaver-Burk function, the Michaelis constant K_m and the maximum reaction velocity V_max were calculated to be 0.19 mM and 4.67 × 10^-8 M s⁻¹ respectively. Compared with some natural enzymes and artificial enzymes, the K_m is lower and V_max is larger (Table S1), indicating P-NCDs as an oxidase mimic are more affinitive to and catalytically active over TMB.

3.4 Mechanism on oxidase-like activity of P-NCDs

As for the catalysis mechanism of P-NCDs, it is hypothesized that the photo-activated oxidase-like activity of P-NCDs is related to the produced singlet oxygen. To validate this mechanism, we conducted a series of experiments. First, compared with air-saturated solutions, the absorbance in the solution blown in with nitrogen for 1 min obviously decreased (Fig. 6A), indicating the dissolved oxygen in the solution was involved in the photoactivated oxidation. Then chemical scavengers were used to clarify the active oxygen produced from oxidation. Mannitol, catalase, superoxide dismutase, tryptophane, and ethanol were chosen to scavenge •OH, H₂O₂, O₂^•-, ¹O₂, and h⁺ respectively. The oxidase-like activity was obviously weakened after the addition of tryptophane (Fig. 6B), indicating the photooxidation was mediated by ¹O₂. EPR further verifies the generation of ¹O₂ (Fig. 6C). The ¹O₂ producing ability of P-NCDs was assessed using 1,3-diphenylisobenzofuran (DPBF). As the irradiation time of the 365 nm UV lamp was extended, the absorbance at 418 nm of the P-NCDs and DPBF mixed solution gradually decreased (Fig. 6D). This result indicates the formation of active ¹O₂ in the reaction system.

Close

Fig. 6

(A) Effect of N₂ on oxidase-like activity of P-NCDs. (B) Effect of scavengers on TMB oxidation catalyzed by P-NCDs under 365 nm UV lamp irradiation. (C) EPR spectrum of P-NCDs (UV lamp irradiation time, 5 min) with the presence of TEMP (TEMP is a special spin trap of ¹O₂). (D) absorption spectra of the P-NCDs (0.01 mg/mL) and DPBF (60 μM) water mixed solutions after different duration of irradiation. (E) mechanism on the photoinduced oxidase activity of P-NCDs.

Export to PPT

Based on the above results, we propose the mechanism underlying the photoinduced oxidase activity of P-NCDs (Fig. 6E): P-NCDs under ultraviolet illumination absorb photons, inducing electron transition from the ground state (S₀) to the excited singlet state (S₁). In addition to the fluorescence emission to recover the ground state in S₁, the electrons of S₁ also can be transferred via intersystem crossing to the excited triplet state (T₁) (Celli et al., 2010; Wang et al., 2016), so T₁ has longer lifetime than S₁, and is attenuated via phosphorescence emission to the ground state. During the collision between the excited triplet-state P-NCDs and the triplet-state dissolved oxygen, energy was transferred to the surrounding oxygen molecules, forming ¹O₂ (the scheme shown below) (DeRosa and Crutchley, 2002; Miao et al., 2021; Plaetzer et al., 2009), which oxidized its substrate -- TMB to blue ox-TMB.

where P, photosensitizer; S₀, singlet ground state, S₁, the first excited singlet state; T₁, the first excited triplet state; k_ISC, intersystem crossing rate constant; k_en, energy transfer rate constant; ³O₂, ground-state triplet oxygen; ¹O₂, singlet oxygen.

3.5 Feasibility and mechanism of colorimetric Hg²⁺ detection

The experiments above prove the photoresponsive oxidase-like activity of P-NCDs, the ability to convert colorless TMB to blue ox-TMB, and the evident characteristic peaks (Fig. 5A). Fig. 7A illustrates the UV–vis spectra of three reaction systems. In the presence of Cys, the absorbance at 652 nm obviously decreased, and the solutions gradually turned from dark blue to light blue. In the coexistence of Cys and Hg²⁺, the solutions became dark blue again, and the absorbance at 652 nm was recovered. The above results indicate this sensor is feasible for colorimetric detection of Hg²⁺.

Close

Fig. 7

(A) UV–vis spectra of TMB/P-NCDs, TMB/P-NCDs + Cys, and TMB/P-NCDs + Cys + Hg²⁺ systems under the 365 nm UV lamp. Inset: photos of samples. (B) mechanism of colorimetric Hg²⁺ detection. (C) UV–vis spectra of detection systems in response to different concentrations of Cys. Inset: relevant photos. (D) relationship between absorbance at 652 nm and Cys concentration. Inset: linear calibration curve of Cys detection. (E) UV–vis spectra of detection systems in response to different concentrations of Hg²⁺. (F) linear fitting between absorbance at 652 nm and Hg²⁺ concentration. Inset: photos of samples.

Export to PPT

The mechanism of Hg²⁺ detection is shown in Fig. 7B. Under the 365 nm UV lamp irradiation, the P-NCDs photosensitized dissolved oxygen to form ¹O₂, which oxidized TMB to ox-TMB, so the solution turned blue. Cys added to the sensing system can prevent the formation of cation free radicals, so TMB cannot be converted to ox-TMB, 'turning off' the sensor and making the solution colorless. When Hg²⁺ existed in the sensing system, owing to its specific complexation with Cys, TMB can be converted to ox-TMB, 'turning on' the sensor and making the solution blue, which facilitated the colorimetric detection of Hg²⁺.

3.6 Colorimetric detection of Hg²⁺

As the Cys concentration rose, the absorption at 652 nm decreased (Fig. 7C). The reducible Cys effectively inhibited the TMB oxidation and color change by P-NCDs. Hence, the solution gradually turned from blue to light blue and even colorless (inset). The absorbance at 652 nm was linearly related to the Cys concentration (Fig. 7D). When Hg²⁺ existed, its specific complexation with Cys weakened the reducibility of Cys, so more colorless TMB was oxidized to blue ox-TMB. As the Hg²⁺ concentration rose, the absorbance intensity gradually increased (Fig. 7E), and the solution became dark blue (inset in Fig. 7F). The absorbance at 652 nm is well linearly related to the Hg²⁺ concentration, with a range of 0.01–14 μM, R₂ = 0.9933 (Fig. 7F). The limit of detection (LOD) was calculated to be 3.1 nM (LOD = 3σ/S, where σ is the standard deviation of the absorbance of the blank sample (n = 11), and S is the slope of the standard curve). These results are similar to other Hg²⁺ detection methods (Table S2), indicating this colorimetric sensor can rapidly and sensitively detect Hg²⁺.

3.7 Selectivity in Hg²⁺ detection

To further assess the selectivity of the colorimetric sensor over specific metal ions, we studied its response to common metal ions. The effects of all metal ions on this sensor were low or negligible (such as Na⁺, Mn²⁺, Ag⁺, Pb²⁺, Cr³⁺ only result in weak growth of absorbance, and some ions such as K⁺, Zn²⁺, Mg²⁺, Cd²⁺, Al³⁺, Ca²⁺, Ba²⁺, Ni²⁺ even caused a slight decrease in absorbance), except Cu²⁺ (Fig. S4A). The significant enhancement of absorbance at 652 nm with the presence of Cu²⁺ may be ascribed to its high binding capacity. Hence, the potential interference from Cu²⁺ was eliminated (Fig. S4B). Then the ethylenediamine tetraacetic acid (EDTA) and Cl^- mixture (molar ratio = 2: 1) was used as a masking agent (Lu et al., 2021), so the formation of a stable complex fully eliminated or avoided the interference from Cu²⁺. In summary, this colorimetric sensor shows high selectivity in Hg²⁺ detection.

3.8 Analysis of real samples

To further explore whether this photoresponsive colorimetric sensing system was feasible for detecting Hg²⁺ in real samples, we conducted spiked recovery experiments by detecting Hg²⁺ in urine, tap water, and lake water. The spiked recovery rates are within 92.2 %-109.5 % (Table S3), suggesting this method can successfully detect Hg²⁺ in real samples.