Top-Down N-Doped Carbon Quantum Dots for Multiple Purposes: Heavy Metal Detection and Intracellular Fluorescence

In the present study, we successfully synthesized N-doped carbon quantum dots (N-CQDs) using a top-down approach, i.e., hydroxyl radical opening of fullerene with hydrogen peroxide, in basic ambient using ammonia for two different reaction times. The ensuing characterization via dynamic light scattering, SEM, and IR spectroscopy revealed a size control that was dependent on the reaction time, as well as a more pronounced -NH2 functionalization. The N-CQDs were probed for metal ion detection in aqueous solutions and during bioimaging and displayed a Cr3+ and Cu2+ selectivity shift at a higher degree of -NH2 functionalization, as well as HEK-293 cell nuclei marking.


Introduction
Carbon-based nanomaterials, such as zero-dimensional fullerenes, 1D carbon nanotubes, and 2D graphene, have received considerable attention as useful materials for various applications in electronics, optoelectronics, photovoltaics, and sensing [1][2][3]. Carbon quantum dots (CQDs) not only inherit the excellent optical properties of traditional semiconductor quantum dots [4,5] but also compensate for the deficiencies of the traditional materials in terms of biocompatibility, cytotoxicity, and biohazards [6,7]. Other relevant aspects of these nanomaterials are their excellent solubility in an aqueous environment, chemical stability, photobleaching resistance, large-scale preparation, and ease of surface functionalization [8]. Due to quantum confinement effects and localized surface states, CQDs exhibit a variable photoluminescence emission [9,10]. The subsequent interaction of ions or molecules can modify the localized surface states, thus causing a photoluminescence quenching or enhancement effect. As a consequence, these materials are promising candidates for replacing the metal-based quantum dots in various applications, such as bioimaging [11][12][13], biosensing [14,15], drug delivery [16,17], adjuvant selection in vaccines [18], and photocatalysis [19]. The working mechanism in sensing and imaging applications is related to the presence of functional groups, which can coordinate the metals being detected, thus affecting the optical properties. The coordinating role is often played by oxygenated functional groups, such hydroxyls and carboxylates, that are attached to CQDs, but additional N-doping is sought for shifting selectivity parameters or spectral features [20].
was probed in the application as fluorescent tools for metal ion detection in water and cell imaging. We found that longer reaction times for the N-CQDs' preparation favored the introduction of a larger amount of -NH 2 . The concomitant effects were a shift in selectivity toward Cu 2+ and Cr 3+ metal detection and more efficient imaging of cell nuclei.

Materials and Equipment
Fullerene C60 flakes and hydrogen peroxide solution (30%) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Ammonium hydroxide solution (28%) was ordered from VWR Chemicals (VWR International, Radnor, PA, USA). The water used in all the experiments was doubly distilled and purified using a Milli-Q system (Millipore, Milford, MA, USA). All metal salts (Cd(NO 3 ) 2 ·5H 2 O, Cu(NO 3 ) 2 , AgNO 3 , CrCl 3 ·6H 2 O, SnCl 2 ·6H 2 O, ZnCl 2 , AlCl 3 ) were purchased from Merck KGaA (Darmstadt, Germany). The metal salts solutions were prepared at the concentration of 10,000 µM. Infrared spectra were taken with a Shimadzu Prestige-21 FT-IR instrument (Shimadzu Corp., Nakagyo-ku, Kyoto, Japan), which was equipped with an attenuated total reflectance (ATR) diamond crystal (Specac Golden Gate, Specac Ltd., Orpington, Kent, UK), in the range 400-4000 cm −1 , with a resolution of 4 cm −1 . A layer of N-CQDs was deposited on a clean Al foil using dropcasting and the deposits were dried in air before measurements were taken. SEM images were taken with a Zeiss Auriga Field Emission-Scanning Electron Microscope instrument (Carl Zeiss Microscopy, Cambridge, UK) operating at 7 kV on N-CQDs diluted 1:100 with deionized water and deposited on a clean Si wafer surface. Measurements were taken upon complete solvent evaporation. The size distribution was measured using a Nano ZS90 (Malvern, UK) at room temperature using disposable plastic cuvettes. Fluorescence measurements were performed using a laboratory setup for photoluminescence, which was equipped with a 200 W Hg (Xe) continuous discharge lamp (Oriel Corp., Stratford, CT, USA) with an excitation 25 cm monochromator (Photon Technology International, Inc., Birmingham, NJ, USA) and an emission 25 cm monochromator (Cornerstone 260, Oriel Instruments, Stratford, CT, USA). The spectral response of the setup was calibrated over the wavelength range of interest using reference fluorophore solutions of quinine sulfate and a certified spectral fluorescence standard kit (Sigma-Aldrich) [40]. The samples were investigated in rectangular fused silica cuvettes with an optical length of 10 mm. The PL signal was collected at the conventional 90 • geometry by quartz lenses with appropriate rejection filters. All the emission spectra were generally recorded using the excitation wavelength of λ exc = 360 nm with a spectral bandwidth of approximately 3 nm. UV-Vis absorption spectra were recorded using a Cary 50 spectrophotometer (Varian Inc., Palo Alto, CA, USA). All the PL and UV-Vis spectra were elaborated with the Origin Pro program. The PL spectra were deconvolved using Gaussian functions to determine the peak position and the higher full width at half maximum (FWHM). Human embryonic kidney 293 (HEK-293) cells were cultured in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 2 mM glutamine and 10% fetal bovine serum (FBS) maintained in a humidified incubator at 10% CO 2 and periodically tested to ensure the absence of mycoplasma contamination. Afterward, the cells were incubated with 2.2 mM N-CQDs-40 min or N-CQDs-15 h in DMEM for 30 min, 1 h, or 2 h. Nuclei of the cells were labeled with DAPI (4 ,6-diamidin-2-fenilindolo) fixed in 4% paraformaldehyde for 15 min and then washed three times with PBS. Images were taken with an ApoTome System (Zeiss) that was connected to an AxioObserver Z1 inverted microscope (Carl Zeiss Microscopy GmbH, Jena, Germany).

Synthesis of N-CQD-Based Nanomaterials
The synthesis of N-CQD-based nanomaterials was performed using a hydroxyl-radical related method to open and cut down fullerene into N-CQDs. Ariga and co-workers reported the decomposition of H 2 O 2 into hydroxyl radicals at a high NaOH concentration [41,42]. In order to increase the optical properties and electrocatalytic activities of the CQDs, heteroatom doping was evaluated to improve the properties and enhance the Nanomaterials 2021, 11, 2249 4 of 14 selectivity of these materials toward the detection of specific heavy metal ions. For this reason, NH 4 OH was used instead of NaOH during the synthetic process. In a three-neck round-bottom flask, 20 mL of fullerene that was dispersed in ultrapure water (5 mg/mL) was sonicated for 1 h 30 min. Afterward, N-CQDs were prepared by loading 4 mL of NH 4 OH (28%) to the fullerene dispersion and slowly adding 16 mL of H 2 O 2 (30%). Subsequently, the resulting solution was heated in an oil bath at 120 • C for 40 min or for 15 h to produce N-CQDs. The final products were centrifuged at 4500 rpm for 10 min, followed by centrifugation at 15,000 rpm for 15 min. The precipitate was then removed and the supernatant of each solution was used for the subsequent analyses without any further treatment. The achieved samples were named N-CQDs-40 min and N-CQDs-15 h. The reaction scheme is shown in Figure 1.

Synthesis of N-CQD-Based Nanomaterials
The synthesis of N-CQD-based nanomaterials was performed using a hydroxyl-radical related method to open and cut down fullerene into N-CQDs. Ariga and co-workers reported the decomposition of H2O2 into hydroxyl radicals at a high NaOH concentration [41,42]. In order to increase the optical properties and electrocatalytic activities of the CQDs, heteroatom doping was evaluated to improve the properties and enhance the selectivity of these materials toward the detection of specific heavy metal ions. For this reason, NH4OH was used instead of NaOH during the synthetic process. In a three-neck round-bottom flask, 20 mL of fullerene that was dispersed in ultrapure water (5 mg/mL) was sonicated for 1 h 30 min. Afterward, N-CQDs were prepared by loading 4 mL of NH4OH (28%) to the fullerene dispersion and slowly adding 16 mL of H2O2 (30%). Subsequently, the resulting solution was heated in an oil bath at 120 °C for 40 min or for 15 h to produce N-CQDs. The final products were centrifuged at 4500 rpm for 10 min, followed by centrifugation at 15,000 rpm for 15 min. The precipitate was then removed and the supernatant of each solution was used for the subsequent analyses without any further treatment. The achieved samples were named N-CQDs-40 min and N-CQDs-15 h. The reaction scheme is shown in Figure 1.

Results
The synthesized samples were characterized with different techniques. In particular, the average size of the quantum dots was estimated using DLS and SEM imaging. The presence of functional groups on the surface of the quantum dots was assessed using infrared spectroscopy. Subsequently, the optical and fluorescence properties and metal detections were determined using absorption and fluorescence spectroscopies and the bioimaging was achieved using fluorescence microscopy.

Characterization of the N-CQDs-Based Nanomaterials
The average size of the N-CQDs was estimated using DLS. The size distribution is reported in Figure 2a,b for the two samples. It was 4.19 ± 0.02 nm for the 40 min preparation, decreasing to 2.81 ± 0.02 nm for the 15 h preparation, with a size distribution range of ±0.8 nm in both cases. This indicated the efficacy of the hydroxyl-radical-opening procedure, as well as the erosion of the quantum dots' edges as an effect of the reaction time due to the continuous action of the reaction mixture.

Results
The synthesized samples were characterized with different techniques. In particular, the average size of the quantum dots was estimated using DLS and SEM imaging. The presence of functional groups on the surface of the quantum dots was assessed using infrared spectroscopy. Subsequently, the optical and fluorescence properties and metal detections were determined using absorption and fluorescence spectroscopies and the bioimaging was achieved using fluorescence microscopy.

Characterization of the N-CQDs-Based Nanomaterials
The average size of the N-CQDs was estimated using DLS. The size distribution is reported in Figure 2a,b for the two samples. It was 4.19 ± 0.02 nm for the 40 min preparation, decreasing to 2.81 ± 0.02 nm for the 15 h preparation, with a size distribution range of ±0.8 nm in both cases. This indicated the efficacy of the hydroxyl-radical-opening procedure, as well as the erosion of the quantum dots' edges as an effect of the reaction time due to the continuous action of the reaction mixture.
SEM images are reported in Figure 2c,d for diluted samples of N-CQDs-40 min and N-CQDs-15 h, respectively, and indicate the presence of CQDs in a size range that was in line with the DLS measurements. The shapes and boundaries, though difficult to estimate at these large magnifications, were irregular.
The introduction of functional groups on the unfolded moieties was evaluated using ATR-FT-IR spectroscopy, especially for determining the simultaneous presence of oxygen and nitrogen derivatives. ATR-FT-IR spectra of the QDs dispersions were registered after drop-casting the aqueous solutions on clean aluminum foils and subsequently drying the deposits in air.
In Figure 3, the ATR-FT-IR spectrum of pristine fullerene powder is compared with the spectra of N-CQDs-40 min and N-CQDs-15 h. SEM images are reported in Figure 2c,d for diluted samples of N-CQDs-40 min and N-CQDs-15 h, respectively, and indicate the presence of CQDs in a size range that was in line with the DLS measurements. The shapes and boundaries, though difficult to estimate at these large magnifications, were irregular.
The introduction of functional groups on the unfolded moieties was evaluated using ATR-FT-IR spectroscopy, especially for determining the simultaneous presence of oxygen and nitrogen derivatives. ATR-FT-IR spectra of the QDs dispersions were registered after drop-casting the aqueous solutions on clean aluminum foils and subsequently drying the deposits in air.
In Figure 3, the ATR-FT-IR spectrum of pristine fullerene powder is compared with the spectra of N-CQDs-40 min and N-CQDs-15 h.   The introduction of functional groups on the unfolded moieties was evaluated using ATR-FT-IR spectroscopy, especially for determining the simultaneous presence of oxygen and nitrogen derivatives. ATR-FT-IR spectra of the QDs dispersions were registered after drop-casting the aqueous solutions on clean aluminum foils and subsequently drying the deposits in air.
In Figure 3, the ATR-FT-IR spectrum of pristine fullerene powder is compared with the spectra of N-CQDs-40 min and N-CQDs-15 h.  Both N-CQDs samples appeared quite different with respect to the pristine fullerene powder since they showed broad absorptions in the region between 3500-2500 cm −1 , as well as evident peaks below 1700 cm −1 . In contrast, pristine fullerene was characterized by only four vibration modes located at 522 cm −1 , 572 cm −1 , 1180 cm −1 , and 1427 cm −1 due to radial displacements of the carbon atoms (lower wavenumber peaks) and tangential modes of the carbon atoms (absorptions above 1000 cm −1 ) [43]. During the hydroxylradical-induced opening of fullerene, the surface of the obtained CQDs was modified by several functional groups, such as hydroxyl (-C-OH), carboxyl (-CHO, -COOH), ether and/or epoxy (-C-O-C-), amine (-C-NH 2 ), and amide (-CO-NH x , x = 1, 2) moieties. The FT-IR spectra of the two samples of unfolded fullerene showed four broad absorptions between 3500 cm −1 and 2500 cm −1 due to the overlap of hydroxyl -OH stretching centered at 3440 cm −1 with -NH x stretching (3199 cm −1 ), -CH=CH-(3044 cm −1 ) aromatic stretching, and -CH bond stretching of sp 3 carbons (2860 cm −1 ). The most prominent peaks in this region were the -NH x stretching and -C=C-aromatic stretching motions, indicating the formation of amine/amide group functionalized aromatic carbon particles. The weak but quite sharp peak registered at 1753 cm −1 could be assigned to -C=O carboxylic group stretching. The broad, weak signal centered around 1614 cm −1 conveyed signals from -OH bending, -C=C-aromatic stretching, -NH 2 scissoring, and/or carboxylic acid -C=O stretching, thus hampering a one-to-one assignment [44]. The very intense and broad peak located at 1414 cm −1 could be related to -C-N-stretching of primary amides and to -C-O-stretching of hydroxyl groups in fullerenols [43,45]. Both N-CQD samples showed a prominent broad and intense absorption around 1319 cm −1 , which could be ascribed to the -C-N-stretching vibration of secondary amines bound to fullerene moieties [43], as well as to the -C-O-stretching of the carboxylic acid bonds. The weak and broad absorption registered at 1093 cm −1 could be ascribed to the -C-O-C-ether group stretching and the weak and sharp signal at 1042 cm −1 derived from -C-C-O-asymmetric stretching of hydroxyl groups. Two sharp absorptions were observed in both samples below 1000 cm −1 and might have been related to the symmetric -C-C-O-hydroxyl stretching (827 cm −1 ) or -C-O-C-bending of isolated epoxy moieties on the CQDs' surfaces [46]. Fingerprint wagging motions of -NH 2 containing functional groups (amines, amides) were observed as a sharp absorption centered at 716 cm −1 . After 15 h of reaction time, the intensities ascribed to -NH x stretching (3236 cm −1 ) and -C-N-stretching of the amide groups (1408 cm −1 ) tended to increase. The largest increase of features occurred to the fingerprint -NH 2 wagging, thus reasonably indicating a larger introduction of amine and/or primary amide groups over 15 h of synthesis. In addition, the peak assigned to epoxy group formation located at 827 cm −1 gained intensity in the N-CQDs-15 h sample, which might be ascribed to a more favorable/efficient cutting of the fullerene molecules. The summary of the IR bands of the samples and corresponding assignments is reported in Table 1.

Optical Properties of N-CQDs-Based Nanomaterials
The optical properties of the N-CQDs-40 min and N-CQDs-15 h were probed using UV-Vis and fluorescence spectroscopies. The prepared N-CQDs displayed slight differences in the absorption spectra. More in detail, the N-CQDs-40 min showed a typical absorption shoulder at 275 nm, whereas the N-CQDs-15 h presented a pronounced peak at 300 nm, which could be assigned to the π-π* transition of aromatic −C=Cbonds in the sp 2hybridized domain of the graphitic core ( Figure 4).

Optical Properties of N-CQDs-Based Nanomaterials
The optical properties of the N-CQDs-40 min and N-CQDs-15 h were probed using UV-Vis and fluorescence spectroscopies. The prepared N-CQDs displayed slight differences in the absorption spectra. More in detail, the N-CQDs-40 min showed a typical absorption shoulder at 275 nm, whereas the N-CQDs-15 h presented a pronounced peak at 300 nm, which could be assigned to the π-π* transition of aromatic −C=C-bonds in the sp 2 -hybridized domain of the graphitic core ( Figure 4). In addition, both samples showed a peak at 350 nm, which was more evident in the N-CQDs-40 min sample and could be assigned to the n-π* transition of −C=O, -C-N-, or −C-OH bonds in the sp 3 -hybridized domains. This transition could be related to hydroxyl (-OH), carboxylic (−COOH), or amine (−NH2) groups on the surface of CQDs [47][48][49][50].
In Figure 5, the PL spectra of the N-CQDs-40 min and N-CQDs-15 h excited at 360 nm are reported. Both samples displayed a blue-green emission in the visible spectrum with rather similar PL profiles. More in detail, fairly broad PL bands were observed with an FWHM of roughly 140 nm that peaked at 457 nm and 451 nm for N-CQDs-40 min and N-CQDs-15 h, respectively, highlighting a peak displacement that was compatible with the average size variation. The photograph in the inset of Figure 5 shows the emission of the N-CQDs upon excitation at 360 nm.
The measured values of the fluorescence quantum efficiency of the two compounds were Φ = 4.5 ± 0.5% and Φ = 10 ± 1% for the syntheses over 40 min and 15 h, respectively. These values compared well with those found in carbon dots prepared through bottomup strategies (though in the lower efficiency range). However, they were significantly higher than the value previously reported for undoped fullerene-derived quantum dots [51]. In this regard, the appreciable increase observed for 15 h of synthesis was consistent In addition, both samples showed a peak at 350 nm, which was more evident in the N-CQDs-40 min sample and could be assigned to the n-π* transition of −C=O, -C-N-, or −C-OH bonds in the sp 3 -hybridized domains. This transition could be related to hydroxyl (-OH), carboxylic (−COOH), or amine (−NH 2 ) groups on the surface of CQDs [47][48][49][50].
In Figure 5, the PL spectra of the N-CQDs-40 min and N-CQDs-15 h excited at 360 nm are reported. Both samples displayed a blue-green emission in the visible spectrum with rather similar PL profiles. More in detail, fairly broad PL bands were observed with an FWHM of roughly 140 nm that peaked at 457 nm and 451 nm for N-CQDs-40 min and N-CQDs-15 h, respectively, highlighting a peak displacement that was compatible with the average size variation. The photograph in the inset of Figure 5 shows the emission of the N-CQDs upon excitation at 360 nm.
The measured values of the fluorescence quantum efficiency of the two compounds were Φ = 4.5 ± 0.5% and Φ = 10 ± 1% for the syntheses over 40 min and 15 h, respectively. These values compared well with those found in carbon dots prepared through bottom-up strategies (though in the lower efficiency range). However, they were significantly higher than the value previously reported for undoped fullerene-derived quantum dots [51]. In this regard, the appreciable increase observed for 15 h of synthesis was consistent with the larger introduction of amine and/or primary amide groups indicated by the FT-IR spectra discussed above (see Figure 3) and agreed with the hypothesis that nitrogen introduces new surface states that trap electrons, thus easing radiative recombination [9,20]. The photostability of the samples was also good, with a fluorescence signal at 99.6% and 99.8% of the initial value for N-CQDs-40 min and N-CQDs-15 h, respectively, after 1 h of irradiation at 360 nm.
with the larger introduction of amine and/or primary amide groups indicated by the FT-IR spectra discussed above (see Figure 3) and agreed with the hypothesis that nitrogen introduces new surface states that trap electrons, thus easing radiative recombination [9,20]. The photostability of the samples was also good, with a fluorescence signal at 99.6% and 99.8% of the initial value for N-CQDs-40 min and N-CQDs-15 h, respectively, after 1 h of irradiation at 360 nm.

Selectivity of the N-CQDs as a Probe for Heavy Metal Detection
Selectivity is an important parameter to evaluate the performance of the N-CQDs as a fluorescent probe for heavy metal detection. Therefore, the fluorescence intensities of the N-CQDs were analyzed in the presence of various metal ions including Cu 2+ , Cd 2+ , Ag + , Zn 2+ , Al 3+ , Cr 3+ , and Sn 2+ at the same concentration. Different studies [51][52][53] demonstrated that metal ions can interact with carbon quantum dots to induce quenching of the PL signal. In this scenario, the present study showed how the aqueous solutions of N-CQDs synthetized for different reaction times exhibited a significant response to two heavy metal ions: Cu 2+ and Cr 3+ but with different behaviors. In Figure 6a,b, the PL spectra of N-CQDs-40 min and N-CQDs-15 h upon interaction with the different metal ions are reported. N-CQDs-40 min displayed a fluorescence variation that depended on the metal in the solution. This fluorescence quenching was largest for Cr 3+ and Cu 2+ with intensity decreases of 25% and 32%, respectively (Figure 6a). At variance with this, N-CQDs-15 h showed comparatively very small fluorescence intensity variations with the various metals (Figure 6b), with the exception of Cr 3+ and Cu 2+ , which caused significant quenching, both in comparison with the other metals and the interaction with N-CQDs-40 min. More in detail, the fluorescence intensity decreased by 44% and 60% for Cr 3+ and Cu 2+ , respectively, whereas it was lower or negligible for the other ions. In summary, both the sensitivity and selectivity of N-CQDs-15 h were significantly higher as compared to N-CQDs-40 min. All these results are summarized in the histograms plot in Figure 6c, where F0 and F correspond to the fluorescence intensities of the N-CQDs at 452 nm in the absence and presence of metal ions, respectively. Assuming a linear behavior of the fluorescence quenching effect with decreasing ion concentrations, as is generally observed in fluorescent CQDs (see [28][29][30][31]50]), LODs of 2 µM and 1.5 µM could be estimated for Cr 3+ and Cu 2+ , respectively. The LOD for Cr 3+ just equals the current limit of 2 µM suggested by WHO for drinking water, whereas the LOD for Cu 2+ is well below the current limit of 30 µM.

Selectivity of the N-CQDs as a Probe for Heavy Metal Detection
Selectivity is an important parameter to evaluate the performance of the N-CQDs as a fluorescent probe for heavy metal detection. Therefore, the fluorescence intensities of the N-CQDs were analyzed in the presence of various metal ions including Cu 2+ , Cd 2+ , Ag + , Zn 2+ , Al 3+ , Cr 3+ , and Sn 2+ at the same concentration. Different studies [51][52][53] demonstrated that metal ions can interact with carbon quantum dots to induce quenching of the PL signal. In this scenario, the present study showed how the aqueous solutions of N-CQDs synthetized for different reaction times exhibited a significant response to two heavy metal ions: Cu 2+ and Cr 3+ but with different behaviors. In Figure 6a,b, the PL spectra of N-CQDs-40 min and N-CQDs-15 h upon interaction with the different metal ions are reported. N-CQDs-40 min displayed a fluorescence variation that depended on the metal in the solution. This fluorescence quenching was largest for Cr 3+ and Cu 2+ with intensity decreases of 25% and 32%, respectively (Figure 6a). At variance with this, N-CQDs-15 h showed comparatively very small fluorescence intensity variations with the various metals (Figure 6b), with the exception of Cr 3+ and Cu 2+ , which caused significant quenching, both in comparison with the other metals and the interaction with N-CQDs-40 min. More in detail, the fluorescence intensity decreased by 44% and 60% for Cr 3+ and Cu 2+ , respectively, whereas it was lower or negligible for the other ions. In summary, both the sensitivity and selectivity of N-CQDs-15 h were significantly higher as compared to N-CQDs-40 min. All these results are summarized in the histograms plot in Figure 6c, where F 0 and F correspond to the fluorescence intensities of the N-CQDs at 452 nm in the absence and presence of metal ions, respectively. Assuming a linear behavior of the fluorescence quenching effect with decreasing ion concentrations, as is generally observed in fluorescent CQDs (see [28][29][30][31]50]), LODs of 2 µM and 1.5 µM could be estimated for Cr 3+ and Cu 2+ , respectively. The LOD for Cr 3+ just equals the current limit of 2 µM suggested by WHO for drinking water, whereas the LOD for Cu 2+ is well below the current limit of 30 µM.  The different behaviors of N-CQDs-40 min and N-CQDs-15 h with respect to Cr 3+ and Cu 2+ can be correlated with the larger presence of amine groups on N-CQDs-15 h, which may have given a more specific interaction, i.e., a larger affinity toward N-derivatives as compared to the other probed ions, such as in Reinecke's salt (NH 4 [Cr(NCS) 4 (NH 3 ) 2 ]·H 2 O) or other stable Cr 3+ complexes with N-derivatives, which can form with multiple bonds [54]. Similarly, Cu 2+ may undergo a substitution of oxygen-dented ligands bonded to Cu 2+ by nitrogen-dented ones [55], thus shifting the selectivity as well.
The comparison with previously reported studies is limited because there are very few investigations on top-down syntheses of CQD for the detection of Cu 2+ and we did not find any for the detection of Cr 3+ . Top-down syntheses of CQDs for Cu 2+ detection were achieved using nitric or sulfonitric attacks of activated carbon or graphite fibers, respectively [56,57]. They both displayed a better LOD than N-CQDs-15 h, but the QY was not reported and no full assessment could be made. It must be added that the selectivity of the graphite-fiber-derived CQDs was achieved only via the addition of biothiol cysteine to the metal-CQDs water solution. A sulfonitric attack of carbon dusk in the presence of aminophenylboronic acid yields CQDs with a better LOD and worse QY as compared to our samples [58]. As for Cr 3+ , we made a comparison with bottom-up synthesized CQDs. In this case, we achieved both a better LOD and QY [59]. A summary of the data is reported in Table 2, along with the main synthesis details.

Fluorescence of the N-CQDs-HEK293 Cell Cultures
The cultured HEK-293 cells, which were stained with DAPI and incubated for 30 min with either N-CQDs-40 min or N-CQDs-15 h, are shown in Figure 7a,b. In both cases, the HEK-293 cells could be clearly visualized after the internalization of the N-CQDs and exhibit a bright green fluorescence due to the particle endocytosis [60][61][62]. In addition, the fluorescence of N-CQDs-15 h was much more marked as compared to the N-CQDs-40 min. This could be attributed to a better piercing due to the smaller size of the N-CQDs-15 h, the -NH 2 rich functionalization, and the higher quantum yield, which was twofold higher in the case of N-CQDs-15 h as compared to N-CQDs-40 min, or to a combination of the three factors. More importantly, the overlap of blue (nuclei) and green (N-CQDs related) fluorescence revealed the penetration of the N-CQDs in the nuclei membrane, although a slight diffusion in the cell cytosol was also visible. Typically, CQDs are cell markers of the cell cytoplasm [63,64]. min. This could be attributed to a better piercing due to the smaller size of the N-CQDs-15 h, the -NH2 rich functionalization, and the higher quantum yield, which was twofold higher in the case of N-CQDs-15 h as compared to N-CQDs-40 min, or to a combination of the three factors. More importantly, the overlap of blue (nuclei) and green (N-CQDs related) fluorescence revealed the penetration of the N-CQDs in the nuclei membrane, although a slight diffusion in the cell cytosol was also visible. Typically, CQDs are cell markers of the cell cytoplasm [63,64]. Only recently, cases of nuclei penetration and marking were reported [65], such as the red-emission CQDs synthesized from hydrothermal treatment of p-phenylenediamine and doped with Ni, which showed the labeling of A549 cells nuclei, or the three-weeklong electrochemical synthesis of CQDs from citric acid, followed by dialysis and used to stain the HEK-293 cell nuclei [66]. Comparatively, the present synthesis provided nuclei permeating N-CQDs without using potentially toxic material, such as Ni, and in relatively short time spans. The employment of N-CQDs-40 min or N-CQDs-15 h was also relatively safe since their incubation with HEK-293 for 1 or 2 h was characterized by cell viability of 99% and 98%, respectively, regardless of the type of quantum dots used.

Conclusions
In the present study, we successfully performed a one-step top-down synthesis of Ndoped carbon quantum dots via hydroxyl-radical fullerene opening with H2O2 and NH4OH with different reaction times. The obtained N-CQDs were then characterized and probed regarding metal ion detection and bioimaging. We found a positive correlation between the reaction time, size, functionalization, and quantum yield. In particular, a longer reaction time (15 h) determined the formation, on average, of smaller quantum dots with a larger presence of -NH2 groups. This resulted in a metal ion selectivity shift toward the detection of Cr 3+ and Cu 2+ in an aqueous solution with a fluorescence intensity decrease by 44% and 60%, respectively, at 100 µM.
As far as the bioimaging properties are concerned, N-CQDs-15 h displayed a significant capability of penetrating and marking cell nuclei, which is a property that has been detected only for a limited number of CQDs so far. Only recently, cases of nuclei penetration and marking were reported [65], such as the red-emission CQDs synthesized from hydrothermal treatment of p-phenylenediamine and doped with Ni, which showed the labeling of A549 cells nuclei, or the three-weeklong electrochemical synthesis of CQDs from citric acid, followed by dialysis and used to stain the HEK-293 cell nuclei [66]. Comparatively, the present synthesis provided nuclei permeating N-CQDs without using potentially toxic material, such as Ni, and in relatively short time spans. The employment of N-CQDs-40 min or N-CQDs-15 h was also relatively safe since their incubation with HEK-293 for 1 or 2 h was characterized by cell viability of 99% and 98%, respectively, regardless of the type of quantum dots used.

Conclusions
In the present study, we successfully performed a one-step top-down synthesis of N-doped carbon quantum dots via hydroxyl-radical fullerene opening with H 2 O 2 and NH 4 OH with different reaction times. The obtained N-CQDs were then characterized and probed regarding metal ion detection and bioimaging. We found a positive correlation between the reaction time, size, functionalization, and quantum yield. In particular, a longer reaction time (15 h) determined the formation, on average, of smaller quantum dots with a larger presence of -NH 2 groups. This resulted in a metal ion selectivity shift toward the detection of Cr 3+ and Cu 2+ in an aqueous solution with a fluorescence intensity decrease by 44% and 60%, respectively, at 100 µM.
As far as the bioimaging properties are concerned, N-CQDs-15 h displayed a significant capability of penetrating and marking cell nuclei, which is a property that has been detected only for a limited number of CQDs so far.