Characterization of material and investigation of two-photon property
Amino-N-GQDs were synthesized from graphene oxide sheets through ultrasonic shearing according to the modified Hummers method  (Additional file, Fig. S1, Table S1 and Scheme S1a). The as-prepared amino-N-GQDs with homogeneous oxygen (O) and N distributors exhibiting high crystallinity and uniformity were sieved by membranes with pores of various sizes. Low-magnification (Additional file, Figs. S2a–d) high-resolution transmission electron microscopy (HR-TEM; Additional file, Figs. S2e-h) was used to characterize the amino-N-GQDs. The mean lateral sizes of the sorted dots were set at 9.1 ± 0.2 nm (amino-N-GQD 9.1), 9.9 ± 0.2 nm (amino-N-GQD 9.9), 11.1 ± 0.3 nm (amino-N-GQD 11.1), and 12.0 ± 0.3 nm (amino-N-GQD 12.0). Other characterizations helped successful preparation of sorted amino-N-GQDs (Additional file, Figs. S3–S5).
Recombination of zigzag edge sites, localized electron–hole pairs, and quantum effect are used to induce intrinsic-state emission of GQD-based materials, whereas the defect effect (energy traps) triggers defect-state emission [2, 5, 6]. To demonstrate the effect, Fig. 1a displays sorted amino-N-GQDs dispersions, various levels of PL emission (gray-level images), dots with slight variation in sizes, and wavelengths encompassing the NIR-I window at 630 nm. The laser system’s x-y focal point and z-axis resolution (full width at half maximum, FWHM) were set at approximately 0.45 and 0.90 µm, respectively (Fig. 1b). Satisfactory TPA in the NIR-II window was measured using a self-developed femtosecond Ti-sapphire laser optical system, as displayed in Scheme 1; for details of the system, please refer to the Materials section), with an approximately 960 nm extension in subsequent experiments (Fig. 1c). With the application of the most efficient excitation wavelength, the materials can considerably advance relevant two-photon properties applied in bioimaging with TPE . Figure 2a displays the TPL spectra of sorted amino-N-GQDs, with peaks of amino-N-GQD 9.1, amino-N-GQD 9.9, amino-N-GQD 11.1, and amino-N-GQD 12.0 at approximately 719, 772, 810, and 862 nm in the NIR region under TPE (222.7 nJ pixel− 1 with 20 or 170 scans, approximately 0.13 or 1.11 s of total effective exposure times; Ex: 960 nm). XPS revealed that as the number of carbonyl groups increased (Additional file, Fig. S3), larger electron redistribution appeared, which eventually decreased energy gaps and TPL red shifts . Given that the quadratic dependence of TPL intensifies with the TPE power in the process , Fig. 2b confirms the existence of a two-photon process with exponent 2.00 ± 0.02 for sorted dots and conventional fluorophore (e.g., Rhodamine B and Fluorescein; Fig. 2b).
Determination of EWI-TPL phenomenon
The sorted amino-N-GQDs with homogeneous O and N functionalities can be used for investigating intrinsic electronic properties related to optical behavior with quantum confinement, leading to EWI-TPL under TPE. Furthermore, sorted amino-N-GQDs exhibited two-photon stability, which could be attributed to limited photobleaching because of dots’ post-TPE TPL intensity (Fig. 2c), whereas that of Rhodamine B and Fluorescein’s fluorescence demonstrated poor robustness against photobleaching on TPE exposure (222.7 nJ pixel− 1 with 20, 100, or 170 scans, approximately 0.13, 0.65, or 1.11 s of total effective exposure times). Furthermore, ultraviolet photoelectron spectroscopy revealed that n-state levels were fixed at almost the same energetic positions (6.6–6.8 eV; Additional file, Fig. S6), irrespective of the size determined through TEM and Raman spectroscopy (Additional file, Figs. S2 and S5), which confirmed the highest occupied orbital level of sorted dots. The quantum confinement resulting from the particle size regulated the wavelengths of radiative transitions. Furthermore, EWI-TPL emissions from sorted amino-N-GQDs implied the absence of trap states between the n-state and π* energy levels. A change in the particle size did not cause any disturbance in the n-state level. The EWI-TPL of the sorted dots could be attributed to π*→n recombination triggering electron transition and phonon scattering. Measurements revealed that the absolute fluorescence QY  of the materials ranged from approximately 0.39 (for amino-N-GQD 9.1) to 0.48 (for amino-N-GQD 12.0). Desirable yields were achieved because of the electron-donating species of the sorted amino-N-GQDs structure. XPS revealed that the high percentage of C–N configurations functioned as electron-donation species and improved QY through the inhibition of nonradiative transitions (Additional file, Fig. S3). By contrast, low QY was due to the presence of a large amount of electron-withdrawing carbonyl functional groups acting as nonradiative trap centers (Additional file, Fig. S3). Characterization of sorted amino-N-GQDs revealed successful preparation causes the GQDs to exhibit EWI-TPL characteristics. However, a large cross-section is preferred in the monitoring of molecular actions. Sorted amino-N-GQDs exhibited large absolute TPE cross-section ranging approximately from 55946 to 60728 Goeppert-Mayer (GM; with 1 GM = 10− 50cm4s photon− 1), which was more than 2900 times the magnitude of the Fluorescein (~ 19.2 GM; Table 1). The absolute TPE cross section for amino-N-GQD 9.1, amino-N-GQD 9.9, amino-N-GQD11.1, and amino-N-GQD12.0 were approximately 55946, 57332, 59051, and 60728 GM, respectively (Table 2). For detailed calculation, refer to the Materials section. This difference indicates that a high ratio of the energy absorption to the energy input in biospecimens. This phenomenon is highly favorable in two-photon imaging (TPI) .
Observation of EM and TPI images during the process of two-photon PDT in the NIR-II window
Because the self-developed optically inverted microscopy system is not suitable for investigating in vivo assay processes, the biological environment was mimicked by embedding the Escherichia coli (E. coli.; 3.98 ± 1.37 µm in length and 0.98 ± 0.34 µm in width calculated from 400 counts of bacteria) strain in a collagen matrix . The TPI action occurred at a specimen depth of 180 µm under TPE (222.7 nJ pixel− 1 with 20 scans, total effective illumination, ~ 0.13 s; Ex: 960 nm; scan rate, 6.53 ms scan− 1; scan area, 200×200 µm2. For details of calculation, refer to the Materials section; Figs. 3a-b). Bacteria was observed clearly under TEM (inset image of Fig. 3a) but was undetectable in the TPL images (Fig. 3a). Lipopolysaccharide (LPS) is the major component of the outer membrane of E. coli. The physiologically stable and biocompatible sorted-amino-N-GQDs (Table 4 and Additional file, Fig. S7; the selected concentration of 0.75 µg mL− 1 material was used in sequential experiments conducted in the dark) were coated with anti-LPS antibody (AbLPS) through electrostatic interaction to improve efficiency and specificity (Additional file, Scheme S1b), which resulted in the absorption of a substantial amount of sorted dots-AbLPS on the surface of the bacteria. No exceptional morphology (inset images of Figs. 3b) was observed on the surface of the bacteria. By contrast, when the GQD size was increased from amino-N-GQD 9.1 to amino-N-GQD 12.0, highly fluorescent QY and larger cross-section were detected in the TPL images (Figs. 3b). However, all the photoexcited material–AbLPS-treated bacteria were severely damaged when the power was increased to 222.7 nJ pixel− 1 with 100 or 170 scans (with a total effective illumination of ~ 0.65 or ~ 1.11 s), which resulted in abnormal morphology, as observed in TEM (inset images of Figs. 3c and d). TPL decreased after 100 scans (Figs. 3c) and became undetectable after 170 scans (Figs. 3d). For unlabeled bacteria, intrinsic fluorophores’ two-photon autofluorescence (TPAF) could not be observed easily under TPE with the same power (Fig. 3e). By contrast, TEM images revealed limited attachment and nonspecific binding for the sorted dot–treated E. coli (without coating antibody; 222.7 nJ pixel− 1 and 20 scans; Ex: 960 nm; inset images of Figs. 3f). Sequentially, TPI also revealed almost no TPL emission at 180 µm (Figs. 3f). Therefore, all the photoexcited sorted dot–treated E. coli exhibited no unusual morphology even after photoexcitation (222.7 nJ pixel− 1 and 170 scans; Ex: 960 nm; inset images of Figs. 3g). Under the same conditions, clear TPI without TPL emission were observed for bacteria without antibody-coated materials (Figs. 3g). However, the images captured at depth > 180 µm contained spherical aberrations, which severely degraded image quality. Such aberrations were caused by the mismatch between the refractive indexes of the aqueous sample and maximal z depth of the optical laser system, in addition to the influence of the set objective, detection efficiency, and maximal z depth of the optical laser system used . Therefore, TPI was not detected at a 200-µm depth for all the sorted dots (Fig. 4). In this study, the maximum z depth for the detection of TPL emissions with the specific laser optical system was 180 µm because of the detection efficiency and set objective, which was, therefore, set as the optimal depth affording the best resolution for the examination of amino-N-GQDs used as a two-photon contrast probe, particularly for sorted amino-N-GQDs with large lateral size.
Changes in the bacterial cell walls or oxidation were detected. The deterioration of surrounding biological surface substrates was attributable to the reactive oxygen species (ROS), which were through PDT under TPE. These changes may cause bacterial atrophy, morphological damage, and distortion (inset images of Figs. 3c and d) due to amino-N-GQD desorption from the surface of the bacteria (Figs. 3c and d). The LIVE/DEAD kit  was used to investigate the green fluorescence of living bacteria (the selected amino-N-GQD 12.0 was used to conduct this experiment; viability > 99%; Additional file, Fig. S8a).The results revealed that the bacteria were almost completely undamaged by exposure to laser treatment (222.7nJ pixel− 1 with 170 scans; with a total effective illumination of ~ 1.11 s; Ex: 960 nm) with 3h of additional incubation. Photoexcited amino-N-GQD 12.0-AbLPS-treated bacteria without incubation were also nearly undamaged (viability > 99%; Additional file, Fig. S8b). After 3h of additional incubation, the same panel revealed that dead bacteria were somewhat distinguishable (represented by red fluorescence in Additional file, Fig. S8c). Then, bacterial viability was quantified for further antimicrobial testing, which revealed the almost complete elimination of amino-N-GQD 12.0-AbLPS-treated bacteria (elimination > 99%; Additional file, Fig. S8d; corresponding to Fig. 3d) and efficient antibacterial effect of amino-N-GQDs in PDT. Thus, no other photochemical activity (e.g., photothermal effect) appeared after photoexcitation. Furthermore, bacteria treated without antibody-coated materials exhibited almost no antimicrobial effect under the same condition (Additional file, Figs. S8e–h; corresponding to Fig. 3g).
ROS assays and bactericidal capability under two-photon PDT
ROS could induce DNA damage, inactivate enzymes, and oxidize amino acid, causing bacterial injury. ROS plays a crucial role in PDT, such as the superoxide radical anion (O2.−) and singlet oxygen (1O2), which are detectable. Despite the generation of a considerable amount of 1O2 and O2.−, false-positive ROS signals may occur  perhaps due to interactions among the sorted amino-N-GQDs, SOSG, t-MVP, XTT, and GSH; such signals could affect results (Additional file, Tables S2 and S3). ROS generated by the sorted amino-N-GQD-Ab-treated bacteria was monitored (Additional file, Tables S4 and S5) and appeared to be consistent with the 1270-nm 1O2 phosphorescence signals emitted from the sorted amino-N-GQDs (Fig. 5). The material without antibody coating (Additional file, Tables S6 and S7) generated lower amount of ROS than that produced by the AbLPS-coated material. Furthermore, 1O2 QY (ΦΔ)  was estimated to be approximately 0.26, 0.28, 0.31, and 0.34 for amino-N-GQD 9.1, amino-N-GQD 9.9, amino-N-GQD 11.1, and amino-N-GQD 12.0, respectively. The study revealed that the material’s antimicrobial potential using PDT against E. coli. The bactericidal capability of the dots was investigated at a low dose of 0.75 µg mL− 1 in the dark, with 222.7 nJ pixel− 1 (20 scans) of TPE (Ex: 960 nm). No significant difference was in the viability of each panel (Additional file, Figs. S9a and b; corresponding to Fig. 3b). With 100 scans, TPE still exhibited no bactericidal effect on bacteria alone and the material in the panel continued to exhibit considerable biocompatibility with the sorted dots-Ab-treated bacteria without TPE (Additional file, Figs. S9c and d). However, under TPE, sorted amino-N-GQDs displayed excellent bactericidal capability (approximately 89%, 93%, 98%, and 100% elimination for the amino-N-GQD 9.1-AbLPS, amino-N-GQD 9.9-AbLPS, amino-N-GQD 11.1-AbLPS, and amino-N-GQD 12.0-AbLPS, respectively), amounting to approximately 7.73–7.75 log10 reduction (Additional file, Figs. S9c and d; corresponding to Fig. 3c). By contrast, bacterial viability was higher for materials without antibody coating (over 98% viability) than that for materials with the coating (Additional file, Figs. S9c and d). However, although antimicrobial capabilities were still not apparent (approximately 6%,8%, 9%, and 11% elimination for the amino-N-GQD 9.1-, amino-N-GQD 9.9-, amino-N-GQD 11.1-, and amino-N-GQD 12.0-without coating antibody, respectively) sorted dots exhibited 100% antimicrobial efficacy with the number of scans increasing to 170 for all the sorted dots-AbLPS–treated E. coli under TPE (Additional file, Figs. S9e and f; corresponding to Fig. 3d). The results were attributed to sorted amino-N-GQDs functioning as two-photon photosensitizer (PS) to generate ROS that was involved in PDT action. The results also revealed the effectiveness of antibody coating in enhancing the functions of materials.
Effects of nitrogen dopant and amino functionalization
Amino-N-GQDs exhibited remarkable quantum confinement, and their edge effect could be altered to enhance their electrochemical, electrocatalytic, and photochemical activities [4, 6]. The effect of strong electron donation and large π-conjugated system reportedly improves charge transfer efficiency in amino-N-GQDs , which resulted in favorable TPA, post-TPE stability, TPE cross-sections, and TPL, as well as a higher ratio radiative and nonradiative decay rates (amino-N-GQD 9.1: 0.64; amino-N-GQD 9.9: 0.69; amino-N-GQD 11.1: 0.82; amino-N-GQD 12.0: 0.92; for calculation, refer the Materials section; Table 2). The results indicate that the material passed mainly through the radiative pathway, as fluorescence QY increased (amino-N-GQD 9.1: 0.39; amino-N-GQD 9.9: 0.41; amino-N-GQD 11.1: 0.45; and amino-N-GQD 12.0: 0.48) and lifetime decreased (from 1.13 to 0.93 ns; Fig. 6, Tables 2 and 3). Radiative electron–hole pair recombination was induced by N dopants and amino groups on the surface of GQD-based material, which boosted intrinsic-state emission. However, as suggested previously, for N dopants and amino groups, the maximum occupied molecular orbital energy of graphene flakes can be increased with the presence of edge amine groups . Thus, the narrowing of orbital band gap boosting PL QY may be caused by resonance between delocalized π orbital and the primary amine’s molecular orbital. Furthermore, XPS revealed that C-O, C = O, and amide groups, which induced localized electron–hole pairs’ nonradiative recombination and prevented intrinsic-state emission , were favorable for small materials (Additional file, Figs. S3 and S4). As the particle size increased, so did PL QY. Moreover, chemical modifications strongly affected the electronic properties of amino-N-GQDs, enabling strong electron donation in primary amine molecules known as amino-group functionalization. Singlet-triplet splitting of amino-N-GQDs results in intersystem crossing and a high triplet state yield. The efficiency of this splitting is sufficient for it to compete with the internal conversion between multiplicity-identical states, which results in the creation of ROS for involvement in PDT [4, 20]. Furthermore, as the number of edge sites increased, the amount of C-N, pyridinic-, amino-, and pyrrolic-N groups increased (Additional file, Figs. S3 and S4). Similarly, as the size of the amino-N-GQDs (Additional file, Figs. S3 and S4) enlarged (Additional file, Figs. S2 and S5), their antibacterial ability increased and so did the amount of generated ROS (Fig. 5; Additional file, Tables S2-S7), which lead to highly efficient PDT action.