O-phenylenediamine, D-glutamic acid, and concentrated hydrochloric acid (12 mol L⁻¹) were purchased from Energy (Shanghai, China). Dulbecco’s modified Eagle’s medium (DMEM) and trypsin were purchased from Gibco (Shanghai, China). Fetal calf serum (FBS) and 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2-H-tetrazolium bromide (MTT) were purchased from Sigma-Aldrich. Luria-Bertani (LB) medium was purchased from MP Biomedical Company (Shanghai, China). HeLa cells were purchased from Beyotime (Shanghai, China). Escherichia coli ATCC 700926 and Staphylococcus aureus ATCC 29213 were purchased from BioBw (Beijing, China). The microtiter plates were purchased from Corning (Shanghai, China). Confocal Petri dishes with glass bottoms were purchased from MatTek Company (MA, United States). All chemical reagents were used as received without any purification.

O-phenylenediamine (50 mg) and D-Glu (25 mg) were dissolved in 25 ml of HCl aqueous solution (1 mol/L). The solution was transformed into a polytetrafluoroethene-lined autoclave and heated at 200°C in an oven for 3 h. After being cooled to room temperature, the products were centrifuged at 3,000 rpm for 10 min to remove large sediments. To remove the unreacted substrate and muriatic acid, the product was dialyzed against deionized water through a dialysis membrane (500 Da) overnight. Finally, the black powder named BAPTCDs was obtained after freeze-drying for 24 h. Some of the CDs were dissolved in deionized water to 10 mg/ml.

Transmission electron microscopy (TEM) images were obtained by using a JEM-2100 transmission electron microscope at an acceleration voltage of 200 kV. Ultraviolet-visible (UV-Vis) absorption and fluorescence spectra were collected using a UV-Vis V3900H spectrometer and an F4600 spectrometer (HITACHI), respectively. The quantum yield was determined with Rhodamine B as a reference. X-ray photoelectron spectroscopy (XPS) was conducted on a PHI Quantera II electron microscope with a correction voltage of 284.6 eV. Fourier transform infrared (FT-IR) spectra were recorded with a HITACHI Z2012-350 infrared spectrometer. Confocal microscopy images were obtained using an SP5 (Leica) confocal microscope. Scanning electron microscopy images were recorded on an S-4800 (HITACHI) scanning electron microscope. The samples were freeze-dried using an ALPHA1-4/LD plus lyophilizer. The temperature was recorded with a FOTRIC 225s infrared thermal imager.

We performed MTT experiments with HeLa cells at different concentrations of CDs to investigate the cytotoxicity of carbon dots. The cells were cultured in a 96-well plate at a density of 5,000 cells per well in an incubator (37°C, 5% CO₂). After being cultured for 24 h, the cell culture medium was replaced with 150 μl of DMEM medium with 10% FBS including carbon dots (0, 50, 100, 300, 500, and 700 μL ml⁻¹). Then, 20 μl of 5 mg ml⁻¹ MTT reagent was added to each well. After incubation for another 4 h, the medium was removed, and 150 μl of DMSO was added to dissolve the MTT. The resulting mixture was shaken for 10 min at room temperature. The optical density of each well was measured with a microplate reader at 490 nm. Cell viability was evaluated using the following formula: Cellular   activity = ODTreat / ODControl × 100 % . (1)

Here, OD_Treat is the OD value in the presence of carbon dots, and OD_Control is the OD value in the absence of carbon dots.

To explore the photothermal properties of BAPTCDs, we prepared BAPTCDs solutions with different concentrations of 100, 200, and 400 μg/ml with PBS and chose PBS only as a control. Here, 1 ml of each solution was placed in a 1.5 ml centrifuge tube, then fixed and continuously irradiated using an 808 nm laser (energy density is 1.5 W/cm²) for 10 min. An infrared thermal imager was used to record the temperature change throughout the process (temperature accuracy of 0.1°C).

Adopted from a previously published protocol (Yu et al., 2021), we tested the antibacterial activity of BAPTCDs via a standard plate counting method. Briefly, E. coli ATCC 700926 and S. aureus ATCC 29213 were chosen as representative Gram negative and Gram positive bacteria. First, 100 μl of bacterial (10⁸ CFU/ml) suspension and 400 μL of BAPTCD solution (200 μg/ml) were transferred to a centrifuge tube and cultured for 3 h (37°C, 200 rpm) for the full reaction of BAPTCDs and the bacteria. Then, the mixture was immediately either treated with NIR (808 nm, 1.5 W/cm²) for 10 min or with no irradiation. The BAPTCD solution was replaced with PBS in the control groups. The number of bacteria was counted by plating 100 μl of 10-fold serial dilutions onto the LB agar plates. All plates were cultured at 37°C for 16 h, and the bacterial viability was calculated according to the equation: Bacterial   viability = N e / N c × 100 % , (2) where N_c is colonies of bacteria treated by PBS, and N_e is colonies of bacteria treated by different methods.

To examine the binding ability of BAPTCDs and the morphology change of bacteria after the treatment, confocal imaging and SEM were conducted. The bacterial solution in the centrifuge tube was centrifuged and washed three times with sterilized PBS, and then, the bacteria were transferred to confocal Petri dishes to observe their fluorescence with a confocal microscope. Subsequently, 2.5% glutaraldehyde solution was added to the bacterial solution for 2 h to fix the bacteria. After being washed for three times, the bacteria were dehydrated by adding 200 μl of different concentrations of ethanol in a gradient (30%, 50%, 70%, 90%, and 100%) for 15 min each time. After the samples were dried, the bacterial morphology was observed by SEM.

We investigated the recent research on antibacterial carbon dots and found that few photothermal carbon dots can be targeted to bacteria, which will pose great challenges to their application in humans (Table 1). Therefore, we chose D-Glu and o-phenylenediamine as precursors and synthesize a new type of nitrogen-doped carbon dots (BAPTCDs) via a one-step solvothermal method. The manufacturing process is shown in Figure 1. We optimized the experimental conditions to improve the quantum yield to 38% (reactant ratio of 1:2, 200°C, and reaction time of 3 h; Supplementary Tables S1, S2).

To analyze the morphology of the BAPTCDs we performed TEM and DLS. Figure 2A shows the TEM image: the carbon dots were spherical and uniformly distributed with a size of 3 nm. DLS showed that the BAPTCD size distribution was normal with the peak between 3 and 5 nm (Figure 2B). DLS was performed in the solution, and the particles will stretch and deform after absorbing water; thus, the DLS data are consistent with TEM.

FT-IR and XPS measurements were performed to investigate the BAPTCD composition. Figure 3 shows the FT-IR spectrum of the prepared CDs. The broad absorption peak at 3,444 cm⁻¹ is attributed to N-H and O-H, indicating its excellent solution properties. The small peaks at 1,677 cm⁻¹, 1,574 cm⁻¹, 1,503 cm⁻¹, and 1,131 cm⁻¹ were assigned to C=O, C=O, C=C, and C-O stretching vibrations. The absorption peak at 675 cm⁻¹ corresponds to the stretching vibration of the functional group C-H. These functional groups indicated strong nitrogen doping. The XPS survey spectra of BAPTCDs (Figure 4A) showed the presence of C, N, and O with percentages of 57.2%, 18.4%, and 24.4%, respectively; the corresponding C 1s, N 1s, and O 1s peaks were located at 284.7, 401.1, and 531.3 eV, respectively. The C 1s spectrum of BAPTCDs (Figure 4B) is composed of three peaks at 284.7, 401.1, and 531.3 eV, which suggests the presence of C-C/C=C, C-O/C-N, and O-C=O, respectively. The N 1s spectrum of BAPTCDs (Figure 4C) can be divided into two obvious peaks at 399.2 and 401.2 eV, which is consistent with the chemical bonds of graphitic N and pyrrolic N. The O 1s spectrum of BAPTCDs (Figure 4D) can also be deconvoluted into two distinct peaks at 531.2 and 532.4 eV, consistent with C=O and C-O, respectively. These structural features suggest the successful doping of N; the functional groups were remarkably consistent with the FT-IR spectrum.

We next investigated the optical properties of BAPTCDs. The UV-vis spectrum of BAPTCDs showed two absorption peaks centered at 560 and 612 nm (Figure 5A). This might be because the high temperature increases oxidation on the surface of the carbon dots and decreases the band gap between the lowest unoccupied molecular orbital (LUMO) and the highest occupied molecular orbital (HOMO), thus causing a red-shift in the carbon dots (Song et al., 2015). The light blue BAPTCD aqueous solution emitted light red fluorescence under visible light and emitted bright red fluorescence when irradiated with a 365 nm lamp (Figure 5 inset). Figure 5B shows the fluorescence emission spectra of BAPTCDs. Although the excitation wavelength increased from 380 to 600 nm, and the optimal emission wavelength remained at 639 nm. Excitation-development photoluminescence behaviors common in carbon dot materials were not observed, which may be due to the relatively uniform particle size and special surface functional groups of BAPTCDs (Li et al., 2010; Zhang et al., 2016).

PTT is known for its deep penetration and effective photothermal conversion ability which can cause irreversible damage to biological tissues and cells and quickly kill bacteria. An 808-nm near-infrared laser was used to test the photothermal performance of BAPTCDs. Figure 6A shows that the temperature of the BAPTCD solution with different concentrations rapidly increased and stabilized within 10 min. The maximum temperature of BAPTCD solutions with concentrations of 100, 200, and 400 μg/ml under laser irradiation increased to 52.2°C, 70.3°C, and 90.7°C, respectively. However, there was no obvious temperature change in the PBS solution, which indicates that BAPTCDs have a fast and efficient ability to convert near-infrared light energy into heat. Figure 6B shows the temperature change of the BAPTCD solution (200 μg/ml) exposed to an 808-nm laser at various laser power densities (1.0, 1.5, 2.0, and 2.5 W/cm²). The results showed that a higher power density can heat the BAPTCD solution to a higher temperature. Considering that high temperatures can damage the host cells, we used a concentration of 200 μg/ml and the power density of 1.5 W/cm² to test the CDs’ antibacterial activity.

We chose HeLa cells to test the toxicity of BAPTCDs to host cells with an MTT assay. As shown in Supplementary Figure S1, the MTT assay revealed that the average cell viability is higher than 85% even with BAPTCD concentration reaching 400 μg/ml. These results suggested that BAPTCDs have low toxicity to HeLa cells at a relatively high concentration. Supplementary Figure S2 showed that the BAPTCDs cannot bind to HeLa cells. This may be due to their chiral structure (Kouidmi et al., 2014; Xin et al., 2017; Zhang et al., 2018b; Zhang et al., 2020), which contributes to the low toxicity of BAPTCDs to cells. Thus, BAPTCDs will be a very safe material for human use.

To test the antibacterial activity of BAPTCDs, we treated the bacteria with carbon dots and lasers and evaluated the antibacterial efficiency via a standard plate counting method. Figure 7 shows that 80.33% of E. coli and 89.27% of S. aureus were killed by BAPTCDs without NIR, which may contribute to its inhibition of the MurD protein. When irradiated with NIR (808 nm, 1.5 W/cm²), only 3.67% of E. coli survived and all of the S. aureus were killed due to the rapidly rising temperature.

To further explore whether BABTCDs bind to bacteria as we designed, we performed fluorescence imaging to observe the position of BAPTCDs. BAPTCDs showed a strong affinity for bacteria (Figure 8) as all bacteria were wrapped by carbon dots and emitted bright fluorescence under laser irradiation. The affinity increases the spatial accuracy of the antibacterial material and reduces the damage to host cells.

To assess the effect of cell wall damage by BAPTCDs, the morphological changes of bacteria after the treatment were observed under SEM. The SEM images (Figure 9) showed that bacteria without BAPTCD treatment were in good condition with a complete cell wall while the cavities occurred on the cell surface when treated with BAPTCDs even without NIR irradiation. This was consistent with a previous study in which the researchers proposed that carbon dots with a chiral structure may penetrate the bacterial cell wall and specifically bind to cytoplasmic proteins, resulting in damage to the cell wall (Xin et al., 2017). When exposed to NIR, obvious lethal effects occurred. The cell walls were badly damaged, leading to bacterial death. This approach mainly uses physical action to sterilize to avoid the generation of drug resistance. Thus, it may be an effective substitute for antibiotics.