Surface morphology

Figure S1 (Supporting Information File 1) presents the surface morphology of the CQD samples. Figure S1a shows a TEM micrograph of CQDs. The average diameter of these dots is 4 ± 1 nm. A top-view AFM image of CQDs is presented in Figure S1b (Supporting Information File 1). TEM and AFM images show that the CQDs are spherical. Statistical analysis conducted on more than 20 AFM images in Gwyddion software showed that more than 80% of the CQDs had a diameter between 2 and 5 nm while their height was 2.6 nm (Figure S1c,d in Supporting Information File 1). Figure 1 shows the surface morphology of neat PU and a CQDs/PU sample. The RMS roughness values of these samples are 4.45 and 14.04 nm, respectively.

Figure 1: (a) Top-view AFM image of neat PU and (b) top-view AFM image of a CQDs/PU composite.

FTIR, UV–vis, and PL spectra of CQDs and CQDs/PU

To study the chemical and optical properties of CQDs, FTIR, UV–vis, and PL spectra were measured. A FTIR spectrum of the CQDs is presented in Figure S2a (Supporting Information File 1). The spectrum contains many peaks associated with the following bonds: Peaks at 3634 and 3448 cm−1 stem from O–H stretching vibrations. A peak at 3367 cm−1 could be assigned to N–H stretching vibrations of aliphatic primary amines whereas the peaks at 2880 and 2943 cm−1 originate from C–H stretching vibrations. A peak at 1759 cm−1 stems from C=O stretching vibrations (carboxylic acid) whereas the peaks at 1696 and 1593 cm−1 could be assigned to C=N stretching and N–H bending vibrations, respectively. The peaks at 1521 and 1061 cm−1 stem from N–O stretching and C–O stretching vibrations, respectively. The peaks at 961, 815, and 759 cm−1 could be assigned to C=C bending, C-H bending, and C=C bending vibrations, respectively [30]. An et al. reported that XPS analysis of CQDs prepared from o-phenylenediamine showed the presence of sp2 domains predominantly in the carbon core structure of the CQDs, which contributed to the formation of polyaniline fluorophores [31]. Figure S2b (Supporting Information File 1) shows a UV–vis spectrum of CQDs. We can observe that the toluene solution of CQDs has a strong broad absorption band at 248 nm with a shoulder at 224 nm, which represents the π–π* transition of C=C bonds. Apart from this broad band, there is a shoulder peak at 330 nm corresponding to the n–π* transition of C=O [32]. In the visible light region, at 444 nm, there is a weak absorption peak caused by surface state of CQDs [33,34]. Figure S2c (Supporting Information File 1) shows PL spectra of CQDs. It is obvious that the emission of CQDs does not depend on the excitation wavelength. Regardless of the excitation wavelengths the emission is located at 545 nm. Thus, CQDs prepared from o-phenylenediamine emit green light and their photoluminescence is excitation-independent. The highest PL intensity can be observed with an excitation wavelength of 490 nm.

The PL of CQDs can be tuned by modifying different factors affecting the structure of CQDs, namely surface states, precursors, preparation methods, and heteroatom doping [9]. Previous investigations showed that possible mechanisms of the CQD photoluminescence are radiative recombination of electron–hole pairs in the sp2 domains inside the sp3 matrix as well as the effect of zig-zag edges [6,18]. Apart from this, surface defects can cause a redshift of the PL emission [35]. Based on the recorded PL spectra, we can conclude that the PL of these dots is dominantly governed by the core states in the conjugated π domains and the quantum confinement effect. Similar to other semiconducting quantum dots (QDs) of nanometer scale, the CQD edges influence the electronic structure of the conjugated sp2 domains [35,36]. Figure 2 shows FTIR, UV–vis and PL spectra of CQDs/PU composite samples. It is obvious from Figure 2a that there are some additional peaks in the CQDs/PU FTIR spectrum compared to that of neat PU. The peaks at 3463 and 3160 cm−1 belong to O–H stretching vibrations whereas a peak at 1612 cm−1 stems from C=C stretching vibrations. A peak at 1496 cm−1 could be assigned to N–O stretching vibrations whereas the peaks at 1295 and 1153 cm−1 belong to C–O stretching vibrations. The peaks at 715 and 1015 cm−1 stem from C=C bending vibrations [30]. All other peaks identified in both neat PU and CQDs/PU composite samples originate from PU.

Figure 2: (a) FTIR spectra of PU and CQDs/PU; (b) UV–vis spectrum of CQDs/PU; (c) PL spectra of CQDs/PU.

Figure 2b shows a UV–vis spectrum of the CQDs/PU composite sample. The first peak at 317 nm is due to the π–π* transition of C=C. A peak at 350 nm is due to n–π* transition of C=O whereas a very wide band at 585 nm stems from surface states of the CQDs. All of these peaks can be identified also in the UV–vis spectrum of pure CQDs as well, but when the CQDs are encapsulated inside the polymer matrix, these peaks are redshifted because of the crosslinking of CQDs with polymer chains. Figure 2c shows the PL intensity of the CQDs/PU composite sample. The PL intensity depends on the excitation wavelength while the PL intensity of neat CQDs showed wavelength-independent behavior. After encapsulation of the CQDs into a polymer matrix, PL emission spectra of CQDs are blueshifted compared to the PL emission of neat CQDs, and the highest PL emission intensity is measured at 430 nm for an excitation wavelength of 380 nm (CQDs/PU samples emit blue light). In this way, the PL of CQDs/PU samples is affected by the quantum confinement effect.

Reactive oxygen species production Singlet oxygen generation

The ability to produce reactive oxygen species (ROS) is a very important parameter for the determination of antibacterial activity of certain material. First, we examined the ROS generation of CQDs (Figure S2d, Supporting Information File 1). From this figure it is obvious that CQDs do not generate singlet oxygen. Further, we examined the ROS production of neat PU (control) and CQDs/PU composite samples by three methods, namely EPR, luminescence at 1270 nm, and UV–vis probe measurement. Figure 3a presents the intensity of EPR signals of control and CQDs/PU composite samples. The figure shows that the CQDs/PU composite samples do not generate singlet oxygen. Second, luminescence measurements were carried out at 1270 nm in different atmospheres, namely air, vacuum, and oxygen (Figure 3b,c). The results show that there is no singlet oxygen generation in any atmosphere. The singlet oxygen production of the CQDs/PU composite sample was additionally investigated through an established photochemical procedure based on the use of 1,3-diphenylisobenzofuran (DPBF) as an efficient quencher of 1O2 (Figure S3, Supporting Information File 1). This figure shows that the absorption of DPBF solution in both vials is nearly identical. This means that the CQDs/PU sample does not produce singlet oxygen, which confirms the previous results obtained by EPR and luminescence at 1270 nm. Ge et al. reported earlier that graphene quantum dots generate singlet oxygen through energy transfer to molecular oxygen [21]. Chong et al. claimed that superoxide anions are involved in the generation of singlet oxygen, implying that electron transfer is an intermediate step for the generation of singlet oxygen by photoexcited graphene quantum dots [20]. In nitrogen-doped graphene, depending on the doping procedure, the nitrogen moieties include graphitic N together with pyrrolic and pyridinic nitrogen and amino groups [37-39]. Bianco et al. reported recently that pyridine nitrogen can be a reactive center and activates other reactive centers at the adjacent carbon atoms in functionalized C–N bonds for additional post reactions such as oxidations [40]. Obtained FTIR and EDS results indicate that in the CQDs synthesized from o-phenylenediamine, NH2 groups are dominantly bonded to the basal plane and the edges of the CQDs whereas pyrrolic and pyridinic nitrogen play only a minor role. Furthermore, during the hydrothermal synthesis of CQDs from o-phenylenediamine, the used precursor was able to form slowly a thermodynamically stable polyaniline and further conjugated sp2 domains with NH2 groups. Thus, the formed CQDs do not have reactive centers to generate singlet oxygen or any other reactive oxygen species. In our previous works, we used polyoxyethylene−polyoxypropylene−polyoxyethylene Pluronic 68 (PF68) as precursor to synthesize hydrophobic CQDs [18,19]. These quantum dots produce high levels of singlet oxygen, and their chemical composition differs from that of dots synthesized from o-phenylenediamine. They contain aromatic rings bonded with oxygen functional groups, which are distributed over the basal plane and edges, but they do not have any NH2 groups or pyrrolic and pyridinic nitrogen. According to Ge et al., CQDs prepared from polyoxyethylene−polyoxypropylene−polyoxyethylene Pluronic 68 generate singlet oxygen through energy transfer to molecular oxygen [21]. But CQDs prepared from o-phenylenediamine do not generate singlet oxygen or OH radicals through energy or electron transfer, because the condensation process of these dots includes NH2 groups in their structure whereas the presence of pyrrolic and pyridinic nitrogen is really minor. Thus reaction centers for ROS generation (dominantly pyridinic N) do not exist in o-phenylenediamine CQDs [40].

Figure 3: (a) Intensity of the EPR signal of control (left) and CQDs/PU composite sample (right); (b) luminescence of the CQDs/PU composite at 1270 nm in different atmospheres; (c) singlet oxygen luminescence of the CQDs/PU composite in oxygen atmosphere calculated as the difference between traces in oxygen atmosphere and vacuum.

Hydroxyl radical production

To examine the production of hydroxyl radicals, two measurements at excitation wavelengths of 365 and 405 nm were conducted. In Figure 4, PL spectra of hydroxyterephthalic acid (h-TA) at different times under 365 nm excitation with 6 mW/cm2 intensity and PL spectra of h-TA at different times under 405 nm excitation with 40 mW/cm2 intensity are presented.

Figure 4: (a) PL spectra of h-TA at different times under 365 nm excitation with 6 mW/cm2 intensity; (b) PL spectra of h-TA at different times under 405 nm excitation with 40 mW/cm2 intensity.

The obtained results indicate a low level of PL intensity after all measured time periods (0‒300 min) and (0‒440 min). This fact shows that there is no production of hydroxyl radicals in the CQDs/PU composite samples at the two excitation wavelengths (365 and 405 nm). In the previous section we established that the CQDs do not generate singlet oxygen. Both obtained results related to production of ROS indicate that there are no reactive centers in the structure of the carbon core that could enable the production of ROS.

Antibacterial testing

The antibacterial activity of CQDs differs from the antibacterial mechanism of commercially available antibiotics. The main parameters that determine the antibacterial action of CQDs are generation of reactive oxygen species, cytoplasm leakage due to DNA binding, and gene expression modulation [28]. In addition, the CQD surface charge affects very much the antibacterial activity of CQDs [41]. Recently, Bing et al. observed that CQDs with different surface charges had different antibacterial activities. Positively charged CQDs damaged the membrane of E. coli completely whereas negatively charged CQDs interacted only weakly with the bacterial membrane [42]. Uncharged CQDs did not show any antibacterial activity against E. coli and B. suptilis. In this study, antibacterial testing of all samples was conducted against two bacterial strains, namely S. aureus and E. coli. The results presented in Table S1 (Supporting Information File 1) showed that CQDs/PU composites prepared from o-phenylenediamine did not exhibit any antibacterial activity against E. coli or S. aureus even after treatment under blue light for 360 min.

These results agree with the results presented in the sections above. The CQDs did not generate any type of ROS. They are uncharged as well. The presence of NH2 groups on their surface can possibly contribute to antibacterial activity. NH2 groups adsorb onto the bacterial membrane and molecules bearing this functional group can diffuse into the cell interior, where the disruption of the cytoplasmic membrane finally leads to cell death [25,43]. The dots synthesized from o-phenylenediamine did not disrupt the cytoplasmic membrane.

Cytotoxicity testing

Low cytotoxicity is one of the mandatory requirements for biomedical applications. In this paper, we performed cell viability tests by applying the MTT assay toward MRC5 human lung fibroblast cells. Lung fibroblasts are very important for maintaining the integrity of the alveolar structure by proliferating and repairing injured areas [44]. MRC5 cells have normal karyotype and are commonly used for genetic, cytotoxicity, viral infection, and other fibroblast-based assays [45]. These cells produce hepatocyte growth factor (HGF), express α-smooth muscle actin, and are used to study the regulation of HGF production and the pathogenesis of tissue fibrosis [46-49]. Figure 5 presents cell viability measurements of individual samples with different extract concentrations. The results are presented as percentage of the control (untreated cells), which was arbitrarily set to 100%. As it can be seen from this figure, none of the tested samples (control and CQDs/PU) showed any cytotoxicity against MRC5 cells regardless of the extract concentration. During the measurements, three different extract concentrations were used, and neat PU control and CQDs/PU composite samples were tested with and without blue irradiation. We established that both MRC5 cells and tested bacteria (S. aureus and E. coli) exhibited almost equal resistance to CQDs/PU composites. Our previous research showed that CQDs/PU composites had different effectiveness on bacteria and tested cells (adenocarcinomic human alveolar basal epithelial cells-A549 and mouse embryonic fibroblast cell line-NIH/3T3) [18]. By comparing the results of these two investigations, we concluded that the surface chemistry of the encapsulated CQDs in polymer composites has a crucial effect on the properties of CQDs/PU composites.

Figure 5: Viability of MRC5 after treatment with PU control (a) and CQDs/PU composite samples (b) using different extract concentrations. No cytotoxicity effect of 50% of extract in the cultivation medium of the CQDs/PU composite on MRC5 cells was detected. All tests were carried out in triplicate.

Cellular uptake

Photobleaching limits the use of hydrophobic probes (e.g., Nile red) [50]. CQDs can be used as probes for bioimaging because of the tuneable strong photoluminescence and high resistance to photobleaching. In order to test internalization of CQDs, Hela cells were treated for 48 h with a concentration of 200 µg/mL CQDs. As shown in Figure 6b, fluorescence imaging demonstrated that CQDs penetrated Hela cells well (compared to Hela cells treated with vehicle control, shown in Figure 6a) and were mainly in the cytoplasm region (Figure 6b). The obtained results indicate that these dots can be used for visualization of different cell lines because of the low cytotoxicity and poor antibacterial activity.

Figure 6: Fluorescence images of (a) HeLa cells treated with vehicle control and (b) HeLa cells treated with CQDs (200 µg/mL) for 48 h. Scale bar is 100 µm.