Colloidal N-Doped Graphene Quantum Dots with Tailored Luminescent Downshifting and Detection of UVA Radiation with Enhanced Responsivity.

Luminescent downshifting (LDS) materials are in great demand for their applications in light conversion devices. In this work, by an ingenious chemical approach of in situ doping, N-doped graphene quantum dots (N-GQDs) have been synthesized with tailored green photoluminescence (PL) under ultraviolet (UV) light excitation. The incorporation of N atoms in the form of pyridinic and graphitic C-N bonding into the sp2-hybridized graphitic framework of N-GQDs has led to tailored LDS via PL emissions. The LDS property of synthesized N-GQDs has been advantageously utilized to demonstrate enhanced responsivity (R) of a low-cost commercially available photoconductive cell (PC) for detection of UVA radiation through an indigenous technique. The linear optical responses of samples are optimized by varying the concentration and the dispersing medium. Also the N-GQDs are shown to be photostable in poly(vinyl alcohol) (PVA) hydrogel. A 60% enhancement in photocurrent of the PC-based photodetector under UV radiation has been obtained here by using N-GQDs/PVA as LDS material. Thus, detection of UVA radiation with a high specific detectivity (D*) of 9 × 1013 Jones and responsivity (R) of 3 A W-1 has been demonstrated, which might open the opportunity of using this material in future energy conversion devices.

Introduction

The fascinating optical properties of graphene quantum dots (n class="Chemical">GQDs) have led to new advancements for its photonic applications. GQDs have a number of distinct merits, such as low toxicity, high photostability, and easy synthesis,[1,2] which make them attractive for applications in bioimaging,[3] sensors,[4] energy storage,[5] and optoelectronic devices.[6−10] The optical properties of GQDs are still regarded as an imperative research topic, to reveal the exact origin of their luminescence and to optimize it for light-emitting applications. The photoluminescence (PL) emission mechanisms of GQDs mainly rely on: (i) the formation of molecular fluorophore and their contribution to the strong emission of the resulting sample[3,11] and (ii) the intrinsic (band gap related) emission from sp2 carbon core of GQDs.[6,8] Theoretical calculations and experimental studies have pointed out that the intrinsic emission in GQDs can be controlled by tailoring of the chemical and electronic environments of their conjugated sp2 domains by means of substitutional doping of heteroatoms.[8,12] Among many other heteroatoms, nitrogen (N) serves as a natural dopant for GQDs since it is electron-rich and has a similar atomic size to carbon. Modulation in the photophysics of GQDs and/or carbon quantum dots (CQDs) is possible by incorporating N atoms into their sp2 carbon matrix, and recently, Zhang et al. have shown that the fluorescence quantum yield of nitrogen- and sulfur-doped GQDs is 84 times higher than that of Ru(bpy)3Cl2, which is 9.3-fold higher than that of undoped GQDs.[13]

Li et al. reported for the first time synthesis of n class="Chemical">N-doped GQDs (N-GQDs) from tetrabutylammonium perchlorate and demonstrated the superior electrocatalytic ability of the synthesized materials.[14] Several groups showed that PL intensity of GQDs can be enhanced significantly by N-doping.[14−16] Santiago et al. reported pulsed laser ablation technique to synthesize N-GQDs from commercially available GO, where urea and diethylenetriamine were used separately as N dopants.[16,17] In some recent studies, the hydrothermal/solvothermal method was suggested for the synthesis of N-GQDs by using different precursors[18,19] and solvents[3] for achieving high PL quantum yield (PLQY). However, some of the previously used methods of synthesis of N-doped GQDs are limited by the critical synthesis conditions and/or due to their requirements of very expensive experimental setup. Also the synthesis of luminescence N-GQDs from the same precursor with different levels of N doping is very much essential to understand their PL mechanism.

Apart from their promising biomedical and chemical senn class="Chemical">sing applications, carbon-based emerging environment-friendly phosphors have received significant interest in different lighting applications due to their luminescent downshifting (LDS) property.[20,21] In energy-generation devices as well as in some optoelectronic devices, there is a high demand of LDS materials for increasing their efficiency as well as protecting them from harmful ultraviolet (UV) radiations.[22−24] Recently, Tian et al. have used red- and green-emitting carbon dot phosphors on commercially available blue-emitting InGaN chips for making white light-emitting diodes (LEDs).[10] Performance of the CQD-poly(vinyl alcohol) (PVA) film as a UV radiation protecting layer has also been reported earlier.[25] Because of the excellent UV-light absorption capability, silane-functionalized GQDs have been used as a full-band UV-shielding material,[26] and it gives an idea about their LDS application. However, some of the essential criteria for a luminescent species to be applicable as LDS material are:[27] (i) wide absorption band in the region where the external quantum efficiency (EQE) of the photosensing device is low; (ii) good separation between the absorption and emission bands to minimize the photo-reabsorption losses; and (iii) low cost. Therefore, synthesis of GQDs having full-band UV absorption (200–400 nm) and which can produce a highly intense narrow band emission of light in the visible regions are very much required. Further, as mentioned earlier, the optical properties of GQDs can be tuned by changing the chemical and electronic environments of their conjugated sp2 domains[12] by means of substitutional doping of heteroatoms. Considering these aspects, N doping into the sp2 skeleton of GQDs could be an excellent way of designing a new carbon-based LDS material for photonic applications.

For decades, development of innovative strategies[28−30] for sensing of ultraviolet (UV) radiation has drawn extensive attention due to wide range of applications in civilian and military areas, such as water sterilization, optical communication, missile plume detection, and so on.[31] Responsivity, detectivity, response time, and linear response with radiation intensity are regarded as critical performance parameters of a UV detector in all application areas. For example, UV detectors made of photoconductive materials are used in flame detection with a responsivity of 0.1 A W–1, high detectivity of the order of 1010 Jones, and a short response time of 1–3 s.[32] Besides, the ability of room-temperature operation, large photosensitive area, and low cost of assembling are some of the essential advantages of a photoconductor-type UV detector for practical applications. Perhaps for these reasons, journey of CdS started in the 1960s, as a photoconductive cell (PC) to detect optical radiation.[30] Although PCs have a much higher sensitivity than a photodiode, they exhibit a nonlinear photocurrent response upon incidence of light.[33] More importantly, the detection of UV radiation by using a commercially available CdS (band gap = 2.43 eV) PCs is not possible as it is hardly sensitive to UV photons.[34] On the other hand, there are several reports on the use of LDS materials as a light-shifting component to improve the efficiency of solar cells.[22−24] Bella et al. proposed the use of photocurable LDS fluoropolymers for improving the efficiency and stability of perovskite solar cells.[22] Similarly, a GQD-based LDS polymer layer for enhancing the efficiency of a crystalline silicon solar cell was reported by Lee et al.[35] Therefore, the aforementioned LDS property of GQDs could be used for boosting up the UV-light-sensing performance of a PC.

Here, we have reported the hydrothermal synthesis of green luminescent n class="Chemical">N-GQDs. In this work, p-phenylenediamine (p-PDA) has been used as a source of carbon and urea for doping of N atoms. The PL emission intensity of the samples has been found to be varied with the concentration of N dopant, and a 6-fold enhancement in PL intensity of N-GQDs has been demonstrated. Further, we have demonstrated here a novel approach for improvement in responsivity of a commercial CdS PC in UVA region by using N-GQDs as LDS material. X-ray photoelectron spectroscopy (XPS) and Fourier transform infrared (FTIR) studies are carried out to reveal the successful insertion of N atoms into the sp2-hybridized carbon core in the form of pyridinic and graphitic N, leading to the formation of a new energy state within the band gap of the material. The emission of visible green light in N-GQDs under UV irradiation has been used to improve the photosensitivity of a CdS PC, which otherwise shows poor photoconductivity under UV irradiation. To improve the photocurrent responsivity of PC in the UV region, N-GQD/PVA (N-GQDs in PVA hydrogel) gel has been used as the LDS material. Furthermore, responsivity (R) and specific detectivity (D*) of the designed photodetector are found to be highly sensitive to the concentration of N dopant in N-GQDs. The photodetection circuit thus designed has been found to be very much effective in the detection of UVA radiation with an impressive linear photocurrent response unlike the bare CdS PC and other wide-band-gap semiconductor photodetectors. Moreover, synthesized N-GQDs may find applications in live cell microscopy.[36]

Results and Discussion

Formation and Structural Features of N-GQDs

In this work, N-GQDs has been hydrothermally synthen class="Chemical">sized by using p-PDA as a source of carbon and urea for N-doping in GQDs.[37] The doping concentration of N atoms into GQDs has been varied by taking various concentrations of urea during the hydrothermal process. The transmission electron microscopy (TEM) image of N-GQDs (viz. for N-GQD3) shown in Figure a depicts the presence of spherical carbon nanostructures having an average size of ∼6 nm. The inset of Figure a shows the high-resolution TEM (HRTEM) image of a single N-GQD of spherical shape with a highly crystalline core. The enlarged view of region I (inset of Figure a) is shown in Figure b. Figure b clearly depicts the formation of zone boundary (yellow dotted curved line), which may have arisen due to the formation of large graphitic core of the N-GQDs from tiny π conjugated carbon structures. Moreover, the enlarged view of rectangular region II shown in the inset (bottom right) of Figure b, confirms the formation of hexagonal honeycomb graphitic lattice with zigzag edges. On the basis of the image intensity profiles (Figure c) derived from single-pixel line traces 1 and 2 in (parallel to the zigzag chains and crossing the centers of the carbon hexagons), the lattice constant has been estimated to be 0.26 nm, which corresponds to the (1120) plane of graphene.[38] The Raman spectrum of the N-GQD3 sample has been collected at a laser excitation wavelength of 532 nm, and it is shown in Figure d along with the fitted curves. It is found that the Raman spectrum consists of two bands, one having center ∼1390 cm–1, corresponding to D band and the other at ∼1550 cm–1 due to the G band of the disordered graphitic structures.[39] However, the finite crystalline size in the presence of defects in N-GQD3 might have caused the broadening in the D and G bands of the Raman spectrum.[40] Dimension of the sp2 core has also been calculated to be ∼7 nm[41] by using the data of intensity ratio ID/IG = 0.68 of D band (ID) and G band (IG) Raman peaks. The value of ID/IG in the present study is large enough compared to that of the pure graphitic structures (∼0.365),[42] and it is due to the presence of zone boundary-related defect states as mentioned earlier. We have also measured the Raman spectrum of N-GQD4 as shown in Figure S2 along with the fitted curves. In this case, the G band is shifted from 1550 cm–1 in N-GQD3 to ∼1620 cm–1 in N-GQD4, due to the increase in lattice defect in sp2 core of N-GQDs.[43] Due to the increase in defect levels, the ID/IG ratio has been found to increase to 1.86 in N-GQD4.

Figure 1

(a) TEM image of N-GQD3 sample, and the inset (top left) shows the HRTEM image of a single N-GQD. (b) Enlarged view of the rectangular region I of the inset of (a). The inset (bottom) of (b) shows the enlarged view of rectangular region II, demonstrating orientation of zigzag edges and the corresponding fast Fourier transform pattern. (c) Image intensity profile of HRTEM image of (b) in the regions along the lines 1 and 2. This is used to find the lattice constant. (d) Raman spectra of the N-GQD3 sample at an excitation wavelength of 514.6 nm.

(a) TEM image of N-GQD3 samn class="Chemical">ple, and the inset (top left) shows the HRTEM image of a single N-GQD. (b) Enlarged view of the rectangular region I of the inset of (a). The inset (bottom) of (b) shows the enlarged view of rectangular region II, demonstrating orientation of zigzag edges and the corresponding fast Fourier transform pattern. (c) Image intensity profile of HRTEM image of (b) in the regions along the lines 1 and 2. This is used to find the lattice constant. (d) Raman spectra of the N-GQD3 sample at an excitation wavelength of 514.6 nm.

X-ray photoelectron spectroscopy (XPS) measurement has been carried out to investigate the changes in the chemical environment and composition of n class="Chemical">N-GQDs before and after addition of N-dopant during the synthesis. Figure a shows the full-scan XPS images of the N-GQDs, which consist of three peaks at ∼284, 400, and 530.2 eV, due to the presence of C 1s, N 1s, and O 1s, respectively. A significant enhancement of N-to-C ratio from 12 to 25% has been observed when the concentration of urea is increased from 0 to 1.6 mM.

Figure 2

(a) Full survey spectra of N-GQD1 and N-GQD3 samples. (b, c) High-resolution XPS image (corresponding to N 1s peak) of N-GQD1 and N-GQD3, respectively. (d) FTIR spectra of N-GQD1 and N-GQD3 samples.

(a) Full survey spectra of N-GQD1 and n class="Chemical">N-GQD3 samples. (b, c) High-resolution XPS image (corresponding to N 1s peak) of N-GQD1 and N-GQD3, respectively. (d) FTIR spectra of N-GQD1 and N-GQD3 samples.

The value of atomic ratio N/C in n class="Chemical">N-GQD3 (25%) is larger than that of graphene-based materials reported earlier by some other studies,[14,44] and it indicates the successful incorporation of N atoms into the sp2 carbon matrix by the present method. Also the value of O/C is found to be increased in the case of N-GQD3 sample due to the presence of oxygen in urea and oxidation of p-PDA during the hydrothermal process. Furthermore, to determine the C–N configuration before and after addition of N-dopant, the high-resolution N 1s spectra of the synthesized N-GQDs samples have been deconvoluted, as shown in Figure b,c. The deconvoluted N 1s spectrum of N-GQD1 in Figure b clearly depicts the presence of two components centered at 398.2 and 400.6 eV, which correspond to pyridinic N (sp2-hybridized N atom with two carbon atoms) and pyrrolic N (sp3-hybridized N atom with five carbon atoms).[16] These pyridinic N and pyrrolic N are also present in the N-GQD3 sample, as shown in Figure c. However, it has been observed that the dominant pyrrolic N in N-GQD1 (40%) decreases with the concentration of N-dopant and it becomes 9% in the case of N-GQD3 sample (Figure b,c). On the other hand, the amount of pyridine N has been enhanced from 60 to 84% with increase in dopant concentration. Additionally, the deconvoluted N 1s spectrum of N-GQD3 (Figure c) depicts the existence of a new component at 401.5 eV due to the presence of graphitic N (substitution of one carbon atom in the hexagonal graphitic lattice by N atom). The high density of pyridinic N with the appearance of graphitic N in N-GQD3 sample indicates the high possibility of radiative recombination[12] compared to that in the N-GQD1 sample. However, these results suggest that the use of urea as a base and N-dopant in the reaction environment can lead to the incorporation of N atoms into the graphitic matrix of N-GQDs in the form of pyridinic N and graphitic N, which can contribute extra electrons to the π-electron system. Therefore, formation of a new energy state occurs near the Fermi levels with electron-rich density distribution (Nπ* states) and change in the valance band structure may occur.[45] Hence, PL intensity of N-GQDs is found to be increased (as presented later) with the increased concentration of N-dopant.

The Fourier transform infrared (FTIR) spectra of synthesized n class="Chemical">N-GQD1 and N-GQD3 samples are shown in Figure d. The IR absorption peaks at ∼3389 and 3269 cm–1 are due to the stretching mode of vibration of O–H and N–H groups,[46] respectively. Most notably, the sharp absorption due to the presence of O–H groups in N-GQD1 becomes broader in N-GQD3, indicating its high degree of carbonization.[47] Also the appearance of absorption band at ∼1560 cm–1 (due to the stretching vibration of C=C bonds) in N-GQD3 suggests the formation of N-GQDs with large sp2-hybridized graphitic core in the presence of urea. However, the absorption band due to the stretching vibration of C=C bond is completely absent in the N-GQD1 sample. Additionally, the sharp IR absorption band at ca. 1450–1370 cm–1 is associated with the stretching and bending modes of vibration of −N bonds[48] in the N-GQD3 sample and is hardly present in those samples synthesized without using urea, i.e., N-GQD1. Consistently, XPS measurement also revealed the presence of these C–N bondings, as discussed earlier.

On the basis of our experimental observations as discussed, herein, the formation mechanism of luminescent n class="Chemical">N-GQDs from p-PDA in the presence of urea has been described by using the reaction pathway shown in Scheme . In the absence of urea at elevated temperature, p-PDA immediately forms dimers and trimers. The formation of rigid polymers containing phenazine units occurred[4] after polymerization of these oligomers, and thereafter N-GQDs (N-GQD1) are formed via the carbonization process. Under hydrothermal condition, the polymerization is mostly due to the formation of radical cations and radical coupling,[4] leading to the synthesis of N-GQDs containing several amino/hydroxyl functional groups on their surfaces. During the high-temperature treatments, the pyridinic and pyrrolic N functionalities are formed in fewer amounts replacing the oxygen-containing groups on the carbon.[49] The formation of N-GQDs may have occurred in two intermediate steps: (i) At elevated temperatures, urea can gradually release amide and/or ammonia[50] and (ii) those continually react with the oxygen functional groups of N-GQD1, which is highly favorable for doping of N atoms inside the graphitic skeleton.[51] Importantly, in the present study, we have been able to control the level of nitrogen doping by varying the concentration of urea during the synthesis of N-GQDs from p-PDA and thus the PL emission intensity of the synthesized N-GQDs has been tuned as shown later.

Scheme 1

Schematic Illustration of the Reaction Pathway for the Formation of N-GQDs

Optical Properties of N-GQDs and the Role of N Doping

The optical properties of all as-synthesized samn class="Chemical">ples have been investigated by UV–vis absorption and PL spectroscopies, and the results are shown in Figure . The UV–vis absorption spectra of all N-GQD samples are shown in Figure a, which also depicts the presence of a sharp absorption peak at ∼264 nm due to the π → π* transition of electrons in sp2-hybridized graphitic domains. It is found that there exist two more absorption peaks at ∼332 and 362 nm in the longer-wavelength (low-energy) region of 300–400 nm, as shown in the inset of Figure a. The absorption band at ∼362 nm is related to n → π* electronic transition due to the presence of conjugated C–N/C=N bonds.[52] Furthermore, to confirm the origin of absorption at ∼362 nm, we have also performed the reduction experiment for all kinds of N-GQDs in the presence of NaBH4, and almost unaltered absorption characteristics in lower-energy region have been observed (see Figure S3, Supporting Information). On the other hand, the absorption band ∼332 nm is associated with n → π* electronic transitions of C=C and C=O bonds in the synthesized N-GQDs.[3,53] It is found that 332 nm absorption band is almost absent in the N-GQD1 sample, but those two peaks clearly appeared in other samples and are most intense in the N-GQD4 sample. This observation suggests that some amount of oxygen-containing functional groups may have been attached onto the surface of N-GQDs during their synthesis in the presence of urea. We have also observed that the intensity ratio of absorption peaks at ∼264 and 362 nm increases with dopant concentration (Figure b), and this result indicates higher sp2 C content (C=C) within the N-GQDs in the presence of N-dopant (as discussed earlier). Moreover, the enhancement in UV absorption intensity of the N-GQD samples with high dopant concentration (Figure a, inset) gives a notion to their use as an LDS material in photovoltaic applications.[27] However, compared to the N-GQD3 sample, N-GQD4 exhibits quenching in absorption intensity, which is a consequence of reduction in the amount of sp2 content within the N-GQD4 sample[54] and is due to excessive insertion of graphitic N inside the core.

Figure 3

(a) UV–vis absorption spectra of the synthesized N-GQD samples; the inset shows absorption characteristics of the samples in lower-energy region (320–400 nm). (b) Variation of intensity ratio of absorption bands at 264 and 362 nm (α264/α362) with dopant concentration. (c) Excitation-dependent PL emission spectra of N-GQD3 sample are collected at a concentration of 1 mg mL–1 in aqueous medium. (d) Variation of PLQY of N-GQDs with dopant concentration and the digital image of N-GQD samples under 365 nm UV lamp are shown in the inset. (e) Photoluminescence excitation (PLE) spectra of the N-GQD3 sample at different concentrations (0.1–0.025 mg mL–1) are collected at a fixed emission wavelength of 520 nm. Schematic of energy band diagram showing the possible highest occupied molecular orbital to lowest unoccupied molecular orbital (LUMO) electronic transition of the N-GQDs in (f) low and (g) high concentrations. (h) Sketch of N-GQDs at low and high concentrations. Top: at low concentration, nanosized clusters separated into single N-GQDs; bottom: at high concentration, a number of single N-GQDs agglomerated to form nanosized clusters; the insets show the corresponding TEM images of N-GQDs at low and high concentrations.

(a) UV–vis absorption spectra of the synthesized n class="Chemical">N-GQD samples; the inset shows absorption characteristics of the samples in lower-energy region (320–400 nm). (b) Variation of intensity ratio of absorption bands at 264 and 362 nm (α264/α362) with dopant concentration. (c) Excitation-dependent PL emission spectra of N-GQD3 sample are collected at a concentration of 1 mg mL–1 in aqueous medium. (d) Variation of PLQY of N-GQDs with dopant concentration and the digital image of N-GQD samples under 365 nm UV lamp are shown in the inset. (e) Photoluminescence excitation (PLE) spectra of the N-GQD3 sample at different concentrations (0.1–0.025 mg mL–1) are collected at a fixed emission wavelength of 520 nm. Schematic of energy band diagram showing the possible highest occupied molecular orbital to lowest unoccupied molecular orbital (LUMO) electronic transition of the N-GQDs in (f) low and (g) high concentrations. (h) Sketch of N-GQDs at low and high concentrations. Top: at low concentration, nanosized clusters separated into single N-GQDs; bottom: at high concentration, a number of single N-GQDs agglomerated to form nanosized clusters; the insets show the corresponding TEM images of N-GQDs at low and high concentrations.

The synthesized n class="Chemical">N-GQDs exhibit usual excitation wavelength-dependent (λex-dependent) PL emission behavior. As shown in Figure c, PL emission intensity is found to increase with the excitation wavelength, becomes the highest at an excitation wavelength of 375 nm, and then decreased. Notably, a red shift in PL emission center (λem) from 522 to 547 nm is also observed when excitation wavelength was varied from 375 to 425 nm. Such excitation-dependent PL emission of N-GQDs can help to control the efficiency of an LDS layered photovoltaic device just by doing a small variation of incident optical wavelength. Moreover, a good separation between the absorption and emission bands of N-GQDs is beneficial for their LDS applications as it can minimize losses due to reabsorption.[27] The least possibility of reabsorption in the case of the synthesized LDS material, i.e., N-GQDs, has been confirmed by the calculation of emission overlap (EO) integral as discussed later. Interestingly, in this work, we have envisioned to tune the UV to visible light conversion efficiency of the synthesized N-GQDs by controlling the incorporation of N atoms into the sp2 carbon matrix. Figure S4a (Supporting Information) depicts the PL emission spectra (λex = 375 nm) of all samples synthesized with different concentrations of N dopant (0–3.2 mM), and PL intensity of the N-GQD3 sample is found to be enhanced by 6 times compared to that of the N-GQD1 sample. Moreover, the PLQY of the N-GQD sample has been calculated (Supporting Information) in aqueous medium by using rhodamine 6G as a reference, and it is found to be 20% for N-GQD3, which is the highest among those of all other samples (Figure d). The digital photograph presented in the inset of Figure d clearly demonstrates the perceptible change in the intensity of visible green luminescence of N-GQD samples with dopant concentration when exposed to 365 nm UV irradiation. Such an emission of green light from the synthesized samples under 365 nm UV lamp (6 W) renders its LDS capability under UV excitation. On the other hand, PL emission center of the synthesized N-GQD samples is found to be red-shifted by an amount of 20 nm with increase in the concentration of N-dopant, as depicted in Figure S4b (Supporting Information). On the basis of this observation of red shift in the PL emission center of N-GQDs, we believed that the presence of urea in precursor solution could act as a base to facilitate dehydration reaction[55] during the hydrothermal process to form a larger π-conjugated graphitic core in N-GQDs as well as more incorporation of nitrogen atoms into the sp2 carbon matrix occurs. Moreover, the narrowing in the line shape of PL emission spectrum of N-GQDs (Figure S4c, Supporting Information) has been observed with increasing concentration of N-dopant, due to the effective improvement in the possibility of radiative recombination of electron–hole pairs. Hence, we have achieved a highly intense PL emission in the case of the N-GQD3 sample, which is 6 times higher than that of the N-GQD1 sample. Such an enhancement in PL emission due to improved radiative recombination of electron–hole pair could be correlated to the increased density of pyridinic N atoms into the graphitic backbone of the N-GQD3 sample, which has been confirmed by XPS data analyses (as discussed earlier). However, quenching in PL intensity with further increase in N-dopant concentration, as observed in N-GQD4 sample, may be due to the instability of chemisorbed N on C at high N radical concentration.[45] As discussed earlier, excessive insertion of graphitic N inside the graphitic core also caused the reduction in sp2 content in N-GQDs. Therefore, if graphitic N enters the graphite structure, it will change sp2 to sp3, which will destroy the conjugated structure. Thus, too much graphitic N is not good for the green PL emission.

Furthermore, herein, we have also studied the PL emission behavior of the N-GQD3 sample with varying concentrations, to (i) find the optimum concentration of the synthesized sample to achieve highly intense PL emission for LDS applications and (ii) reveal the possible origin of excitation-dependent PL emissive behavior of N-GQDs samples. As shown in Figure S4d (Supporting Information), the PL emission intensity (λex = 375 nm) of N-GQD3 is found to be enhanced with concentration and becomes the highest at a concentration of 1 mg mL–1. When the concentration of N-GQD is further increased, quenching in PL emission takes place, mainly due to aggregation of nanostructures, which is known as “aggregation-caused quenching”.[56] In addition, high optical density or turbidity at higher concentration may also lead to reduction in PL intensity in fluorescent carbon nanostructures.[57] Thus, the concentration of N-GQDs used should be carefully maintained to investigate the PL mechanism and light conversion applications.

When N-GQDs are illuminated by UV excitation, the absorbed photons may cause electronic transitions of ∼4.7 eV (264 nm, π → π* of C=C), 3.73 eV (332 nm, n → π* of C=O), and 3.4 eV (362 nm, n → π* of C=N) producing three absorption bands in the UV–vis spectrum (as discussed earlier). The excited electrons can be deactivated by two means: (i) through band-to-band recombination (Cπ* → π transition) after vibration relaxation, producing PL in high-energy region; (ii) by undergoing intersystem crossing (Cπ* → Nπ* and Nπ* → Oπ*), followed by vibration relaxation, and finally radiative recombination (interstate to band transitions; N,Oπ* → π) occurs in the longer-wavelength region.[44] The latter is more effective at high concentration of fluorophore molecule.[53] To find the emissive states that are responsible for excitation-dependent PL emission in N-GQDs, we have collected the PL excitation (PLE) spectra (λem = 520 nm) of the N-GQD3 sample at different concentrations, as shown in Figure e. PLE spectra of N-GQD3 at higher concentration exhibits existence of three emission bands at ∼230, 330, and 377 nm. The transition at ∼230 nm (5.39 eV) corresponds to the electronic transition from σ orbital to the lowest unoccupied molecular orbital (LUMO), i.e., Cπ* state, and it arises due to the presence of carbene-like zigzag sites at the edges of the synthesized graphitic carbon nanostructures[58] (as evident from HRTEM observations). However, the transitions at ∼330 nm (3.76 eV) and 377 nm (3.29 eV) are due to N,Oπ* → π transitions due to the presence of C=O and conjugated C–N/C=N bonds, respectively. Interestingly, a new PLE band at ∼260 nm (4.77 eV) has aroused when the N-GQD sample was diluted to lower concentration. The PLE peak at ∼260 nm, corresponding to Cπ* → π transitions, originates from sp2 core, and it is known as band-to-band transition. As depicted in Figure f, at lower concentration (0.05–0.125 mg mL–1) of N-GQDs, Cπ* → π transition is allowed due to the separation of individual N-GQDs and N,Oπ* → π transitions are less allowed in this case. However, interstate transitions are more probable when N-GQDs are in high concentration (1 mg mL–1) and is due to aggregation of individual nanoparticles. TEM images shown in Figure h are taken at a concentration of 0.025 mg mL–1 (top inset) and 1 mg mL–1 (bottom inset), which confirms the separation of individual N-GQDs at low concentration and their aggregation at higher concentrations, respectively. Furthermore, it has been observed that all N-GQDs samples (i.e., N-GQD2-4) exhibit a lower level of λex-dependent emission characteristics at the lowest concentration of 0.025 mg mL–1 as shown in the excitation-dependent PL emission spectra in Figure S5 (Supporting Information). However, the maximum amount of λex-dependent shift in PL emission center (Δλ in nm) has been observed in N-GQD4 at a concentration of 1 mg mL–1 and is equal to 42 nm (Figure S5c-i, Supporting Information). From these experimental observations, we have concluded that two types of emissive centers are present within the synthesized N-GQDs. The low level of λex-dependent emission characteristics of N-GQDs at very low concentration is due to the PL emission from intrinsic sp2 carbon core (Cπ* → π; band-to-band transition). However, comparatively, a large λex-dependent shift in PL emission center of the synthesized sample at higher concentration is due to PL emission from interstate to band transitions (N,Oπ* → π).[53] Additionally, the polar groups on the surface of the particles are of particular importance to determine the emissive property of carbon nanostructures as demonstrated earlier by us.[59] For higher concentrations of samples, the surface moieties of N-GQDs help to form nanoclusters, as shown in Figure h (bottom inset). As the amount of surface moieties is increased with dopant concentration, the value of Δλ is found to be the highest in the case of N-GQD4 compared to that of other samples. On the other hand, the as-formed nanoclusters are separated and redispersed into single N-GQD when the aqueous solution of N-GQDs is diluted to low concentration (TEM image in Figure h, top inset). This leads to weakening of the van der Waals attraction in between N-GQDs and the emission in the longer-wavelength region due to the disappearance of interstate transitions.[53] Such investigations on concentration-dependent PL emission behavior of the N-GQDs will provide a way to tune the emission wavelength and offer new insights into the PL emission mechanism in luminescent carbon nanostructures, which will promote diverse potential applications of the synthesized material in the near future.

Moreover, the photostability of a luminescent material is an important criterion for its different applications, especially as an LDS material.[27] The photostability of synthen class="Chemical">sized N-GQDs has been studied under 1 h continuous exposure of UV radiation after dissolving them in deionized (DI) water and PVA hydrogel. As shown in Figure S6a (Supporting Information), the relative PL intensity of the N-GQD sample is found to be remarkably constant in PVA hydrogel compared to DI water. In Figure S6b,c (Supporting Information), we have shown a comparison of fluorescence stability of N-GQDs in DI water and PVA hydrogel, and it is found that when N-GQDs are dispersed in PVA hydrogel, they remain stable for up to 12 days. Moreover, in our previous study, we have also reported that[60] the transmittance of the PVA consisting of carbon nanostructures has high transmittance in the visible region and is comparable to bare PVA. On the basis of these observations and discussions, in this study, we have used PVA hydrogel as a suitable matrix to demonstrate the LDS application of synthesized N-GQDs.

Application of N-GQDs as Luminescent Downshifter

The fascinating LDS response of N-GQDs under UV excitation has been judiciously utilized for enhancement in responsivity of a PC, and thereby detection of UV (UVA) radiation has been demonstrated by using an indigenously designed experimental setup, as shown in Figure a. A very low-cost commercially available 0.5 cm CdS PC has been used as a light sensor, and a transistor (Q1) operating in common emitter mode is used to amplify the photocurrent across the PC. Photoresponse of a PC strongly depends on the band gap of photosensitive material (CdS) and hence on photon wavelength. We have also measured the spectral response of the CdS PC at five different wavelengths for calibration, as shown in Figure b, and the responsivity of PC is found to be 5 times higher under green light (532 nm) compared to that of UV light of wavelength 395 nm. This observation confirms that the PC, which has been used as a photosensor, is less sensitive to UV radiation compared to visible light and hence motivates us to use the as-synthesized N-GQDs LDS to improve its photocurrent response.

Figure 4

(a) Schematic diagram of the experimental setup for the detection of UVA radiation by using N-GQDs as an LDS material, where UV blind PC acts as a light sensor. (b) Responsivity of the CdS PC operating under five different wavelengths. (c) Polar plot of green emission intensity from N-GQDs inside the cuvette under UV excitation in different direction. (d) Emission spectra of the output red LED at different intensities of UV radiation in input, and the corresponding digital images of the red LED are shown in the inset.

(a) Schematic diagram of the experimental setup for the detection of UVA radiation by using n class="Chemical">N-GQDs as an LDS material, where UV blind PC acts as a light sensor. (b) Responsivity of the CdS PC operating under five different wavelengths. (c) Polar plot of green emission intensity from N-GQDs inside the cuvette under UV excitation in different direction. (d) Emission spectra of the output red LED at different intensities of UV radiation in input, and the corresponding digital images of the red LED are shown in the inset.

A schematic of the experimental setup for CdS PC-based light detection system is shown in Figure a. Briefly, light from UV light source has been focused (spot area: 0.3 cm2) on the first surface of a four-n class="Chemical">side-polished quartz cuvette (1 cm) containing N-GQDs dispersed in PVA hydrogel (Figure a). Thereafter, the visible green light emitted by N-GQDs under UVA excitation has been collected through PC. A digital photograph of the experimental setup is shown in Figure S7a (Supporting Information). The intensity of the UVA radiation has been varied by changing the biasing voltage across the UVA source LED. The intensity response of the source LED with biasing voltage is given in Figure S7b (Supporting Information). Figure c shows that the intensity of green light emitted from N-GQDs is the maximum in a plane perpendicular to the first surface (incident plane of UV radiation) of the cuvette. Therefore, the PC has been placed in a plane perpendicular to the first surface of the cuvette. The photocurrent of the detector circuit has been measured by varying the intensity of the incident UVA radiation. Furthermore, the intensity of light emission from a red LED kept at the output of the detector circuit also changes with the intensity of incident radiation at input, as shown in Figure d.

The photocurrent (Iph) response of n class="Chemical">PC with intensity of UVA radiation has been investigated at different concentrations of N-GQDs in PVA hydrogel, and the results are shown in Figure a. However, the value of responsivity R has been estimated from the slope of the linear variation of photocurrent with the optical power of incident UVA radiation (Figure S7c, Supporting Information). Notably, the responsivity of PC under UVA irradiation is found to be doubled when the concentration of the N-GQDs in PVA hydrogel was increased to 0.6 mg mL–1. Such amplification in responsivity of PC is due to enhancement in the intensity of green emission from N-GQDs with its concentration in the presence of UVA irradiation. In addition, the detectivity D* (in units of Jones) is commonly reported as an important parameter to characterize the sensitivity of a photodetector[30] and is defined asandwhere IL and Id are currents in the presence of light and dark, respectively, q is the electronic charge, Aeff is the active area (0.03 cm2) of the exposed surface of PC participating in absorption of photon emitted by N-GQDs, and Ei is the intensity of light incident on the exposed surface of PC. The detectivity of PC-based UV detector circuit has been calculated under the light illumination of 0.8 mW cm–2 and is observed to be increasing with concentration of N-GQDs, as shown in Figure b.

Figure 5

(a) Photocurrent response of PC-based UVA detector by using different concentrations of N-GQDs as a light conversion material. (b) Variation of responsivity (R) and specific detectivity (D*) of the detector with concentration of N-GQDs. (c) Photocurrent response of PC-based UVA detector by using all of the synthesized N-GQD samples (0.6 mg mL–1) in PVA hydrogel as LDS material. (d) Variation of R and D* of the detector with concentration of N-dopant used to synthesize N-GQDs.

(a) Photocurrent response of PC-based UVA detector by un class="Chemical">sing different concentrations of N-GQDs as a light conversion material. (b) Variation of responsivity (R) and specific detectivity (D*) of the detector with concentration of N-GQDs. (c) Photocurrent response of PC-based UVA detector by using all of the synthesized N-GQD samples (0.6 mg mL–1) in PVA hydrogel as LDS material. (d) Variation of R and D* of the detector with concentration of N-dopant used to synthesize N-GQDs.

To find the effectiveness of N-dn class="Chemical">oping in N-GQDs on its LDS application, the photocurrent response of the PC has also been studied in the presence of all synthesized N-GQD samples (0.6 mg mL–1) and the results are shown in Figure c,d. The linear variation of photocurrent with intensity of input UVA irradiation in the presence of different N-GQD samples is shown in Figure c. By using N-GQD3 as a converter of UV light, the highest photocurrent of 0.035 mA is achieved at an irradiation of 0.8 mW cm–2. From the linear variation of photocurrent with the optical power of incident UVA radiation (Figure S7d, Supporting Information), the value of responsivity R has been estimated. In the present study, we have observed that the values of R and D* are strongly dependent on the amount of N-doping in N-GQDs. As depicted in Figure d and Table , by using the N-GQD3 sample (N-dopant concentration, 1.6 mM) as an optical wavelength converter at the input of the PC-based detection circuit, maximum achievable values of R and D* are found to be 3 A W–1 and 9 × 1013 Jones, respectively.

Table 1

Figures of Merit Value of the Luminescent N-GQDs as LDS Material and Performance Parameters of the LDS UVA Detector Based on N-GQDs

photosensor	LDSa material	PLQYb (%)	EOc (%)	ESMd (%)	Re (A W–1)	D*f (×1013 Jones)
CdS PC	no LDS material			10	1.40 ± 0.009	4.54
	N-GQD1	4	1.30	35	1.50 ± 0.004	5.40
	N-GQD2	13	0.61	37	1.60 ± 0.087	6.00
	N-GQD3	20	0.55	40	3.00 ± 0.010	9.00
	N-GQD4	10	0.92	38	1.90 ± 0.006	7.00

Luminescent downshifting.

Photoluminescence quantum yield.

Emission overlap.

Emission spectral matching.

Responsivity.

Detectivity.

The performance of a luminescent species as an LDS material depends on the optical properties, viz., on the PLQY, separation between the absorption and emission bands and how well its PL emission band spectrally matched with the external quantum efficiency (EQE) band of the photosensor.[22] By considering these aspects, here we have also calculated the value of EO and emission spectral matching (ESM) integral. EO gives the possibility of photo-reabsorption quantitatively, and it should be zero ideally for LDS material with no overlap between the absorption and emission spectra. The value of EO has been evaluated in terms of a normalized emission profile (Iem(λ)) and absorption profile (Iabs(λ)) of the LDS material, i.e., N-GQDs in solution phase[23] by using eq (Figure a)However, ESM characterizes the overlap between the emission band of LDS material and the EQE profile of the photosensor. The value of ESM has been estimated in terms of the normalized emission profile (Iabs(λ)) of the LDS material and the normalized EQE band (ηb(λ)) of the bare photosensor by eq (Figure b). For an LDS material having an emission profile perfectly overlapping with the EQE band of the photosensor, ESM would be nearly unity. ESM is defined asThe values of EO and ESM have been calculated for all of the synthesized N-GQD samples, and the results are presented in Table . As shown in Figure a, the absorption and emission bands of N-GQD3 sample are found to be well separated with the EO value of 0.55%, which is the minimum among all other synthesized samples. However, the maximum value of ESM is found to be 40% for the N-GQD3 sample, and the comparative data for all samples are given in Table . Moreover, the values of EQ and ESM are better than those of Q Switch 5 and IR-26 dye.[23]

Figure 6

Schematic representation of figures of merit of N-GQD3 used as an LDS material. (a) Emission overlap (EO) integral, (b) emission spectral matching (ESM) integral, (c) cartoon of PC-based UVA detector by using N-GQD fluorophore as LDS material. (i) UV light is directly collected near the surface of the bare PC (gray box) and giving low efficiency. (ii) With the addition of the LDS material, UV light is absorbed by the N-GQD/PVA gel (green box) and reemitted in the visible region where it is collected in the active region of the PC with improved detection efficiency.

The conductivity of a conventional CdS PCs is usually enhanced after absorbing vin class="Chemical">sible photons and hence exhibits a better photocurrent response after absorbing only those visible photons having energy (Eph) equal or around the band gap energy (Eg) of photosensitive material. However, if Eph is larger than Eg (as in the case of UV light), the photons are absorbed, but the excess energy (Eph – Eg) is not used effectively for generation of photocurrent, and it causes electron thermalization.[35] As discussed earlier, the value of ESM characterizes the overlap between the emission band of LDS material and the EQE band of the used PC. In the present study, the value of ESM for the green luminescent N-GQD3 is found to be 4 times higher than that of direct UVA radiation (Table and Figure b). Therefore, under the visible emission from N-GQDs, the production of large number of photoelectrons within PC is responsible for enhancement in its conductivity compared to direct UVA radiation. As a consequence, the presented light detection circuit under UVA excitation is able to show an advanced photoresponsive behavior by using N-GQDs as LDS material. Additionally, Figure S8 shows the variation of photocurrent gain of the detector circuit using different N-GQDs samples as LDS material for a particular intensity of incident UVA radiation (0.8 mW cm–2). Clearly, 3.5 times enhancement in the value of photocurrent gain has been observed by using N-GQD3/PVA gel as the LDS material. Moreover, the photocurrent enhancement (ΔIph in %) in the PC-based photodetection circuit in the presence of the N-GQD/PVA LDS gel has been estimated by eq where Iph and Iph are the values of photocurrent in the presence and absence of LDS at input, respectively. Thus, the value of ΔIph is found to be 60%, which is the maximum when the N-GQD3/PVA gel is used as a luminescent downshifter. Notably, the calculated values of R and D* in the present study are comparable to those of UV photodetectors, which have been reported earlier.[30,61−64] Thus, the present work could be an addition to extend the state of the art in fabrication and designing of UV detectors. In this report, we have also proposed a benchtop design of PC-based UV detector, as shown in Scheme , where N-GQD/PVA gel could be used as an LDS material.

Scheme 2

Proposed Benchtop Design of PC-Based UV Detector by Using N-GQD/PVA Gel as LDS Material

To get information about the photoresponse time of PC-based UV detector, which has been designed here, we have measured its photoswitching behavior under pulsating UV light (395 nm, 0.8 mW cm–2) of frequency 200 Hz (Figure S9, Supporting Information). A clear saturation is observed in the transient signal with a rise time (τrise) of 9 ms and a fall time (τfall) of 8 ms in the UV detector system by using N-GQD3/PVA gel as LDS material.

Conclusions

In conclusion, here we have demonstrated the synthen class="Chemical">sis of N-doped GQDs by an in situ chemical approach. The synthesized samples are characterized using standard characterization tools, like TEM, XPS, Raman, and FTIR spectroscopies, and the optical properties are determined by using UV–vis absorption and PL emission spectroscopies. The synthesized samples are found to exhibit tailored green PL emission due to N doping under UV excitation, and thus the samples show LDS property. The PLQY values of N-GQD samples have been calculated, and a maximum value of 20% has been achieved. The origin of observation of tunability of PL emissions and its quenching under higher N dopant concentration in the samples have been discussed. The LDS property of the synthesized samples has further been fruitfully utilized to achieve an enhancement in the responsivity of a low-cost PC for detection of UVA radiation through an indigenously developed technique. A 60% enhancement in photocurrent of the PC-based photodetector has been achieved by using N-GQDs/PVA gel as LDS material. The different characteristics of the designed photodetector have been studied in relation to the detection of UVA radiation, and a relatively large detectivity of 9 × 1013 Jones and responsitivity of 3 A W–1 have been reported. This work might open a new opportunity of using N-GQDs as LDS material in future for photonic and energy conversion devices.

Experimental Section

Chemicals

All chemicals were of analytical grade. para-Phenylenediamine (n class="Chemical">p-PDA, 97%) and urea (99.5%) were used as a precursor for synthesis of N-GQDs purchased from Loba Chemie and Merck, respectively. Silica gel (60–20 mesh) and acetone used for column chromatography were purchased from Merck. Poly(vinyl alcohol) (PVA, Merck) was used for making transparent hydrogel solution of N-GQDs. Deionized (DI) water was used in all experiments.

Instrumentation

A highly intense LED (1 mW cm–2 at a biasing 6 V) having emission wavelength at 395 nm has been procured from Probots Techno Solutions (Bangalore, India) is used as a source of UVA radiation. The 5 mm CdS PC (GL55 series with metal housing) for photosensing and an NPN transistor (BC547) for photocurrent amplification were purchased from a local market. The emission spectrum of the UVA LED and green light emitted by N-GQDs under UV excitation was measured using an optical fiber spectrometer (AvaSpec-2048). The photocurrent (mA) response of the detector circuit was measured using a digital nanoammeter.

Preparation of N-GQDs

The method of preparation of N-GQDs is described below briefly. In short, 1.8 mM n class="Chemical">p-PDA and 3.2 mM urea (N dopant) were dissolved in 50 mL of DI water; then, the solution was transferred into a Teflon-lined stainless steel autoclave of 100 mL capacity. The precursor solution is heated at 170 °C for 8 h and then it is allowed to cool at room temperature. The as-obtained colloidal dark brownish suspension is then purified by silica column chromatography using acetone as eluent. It is worth noting that, to study the effect of nitrogen doping on the fluorescence intensity of GQDs, the molar concentration of N dopant is varied from 0 to 3.2 mM. The samples thus synthesized by using 0, 0.8, 1.6, and 3.2 mM concentrations of N dopant are symbolized as N-GQD1, N-GQD2, N-GQD3, and N-GQD4, respectively. The details of the experimental techniques for characterization of the synthesized samples are discussed in the Supporting Information.