Highly efficient regulation strategy of fluorescence emission wavelength via designing the structure of carbon dots

Multicolor carbon dots (CDs) possess tremendous potential applications, especially in optoelectronic devices. However, further applications of multicolor light-emitting diode (LED) have been constrained due to the very limited researches concerning the wavelength control mechanism of multicolor CDs. In this work, through theoretical calculation and experimental verification, the regulatory effects of sp2 conjugated domain on the fluorescence wavelength of CDs were explored. Through the regulation of the structure size and the introduction of amide bonds, four kinds of structures were designed in density functional theory (DFT) calculation to explore the influence of sp2 conjugated domain on the fluorescence wavelength of CDs. Using thiourea and p-phenylenediamine as the precursors and regulating the reaction solvents, multicolor CDs with blue, green, and red fluorescence emission were prepared to experimentally verify the emission mechanism. It was confirmed that the increasing structure size and the introduction of amide bond would induce an increasing size of the sp2 conjugated domain, leading to the red shift of the CDs fluorescence wavelength. Furtherly, in order to suppress the self-quenching performance, the CDs@polyvinylpyrrolidone (CDs@PVP) fluorescent film possessing bright solid-state fluorescence was constructed for a better application in light-emitting diodes. The approach provided an effective strategy to realize the programmed regulation on the fluorescence wavelength of CDs, offering us wide potential applications of CDs in the photoelectric device fields.

In recent years, researchers have proposed a variety of synthetic routes or reaction methods to satisfy CDs' emission [33][34][35][36][37][38]. Kwon and his coworkers [39] have reported a synthetic method and prepared a range of the GQDs (graphite phase quantum dots) with certain size distributions via amidative cutting of tattered graphite, which could achieve a size range of 2 to over 10 nm for GQDs by simply regulating the amine concentration. Accompanied with the increasing size was the narrowed down energy gaps in the synthesized GQDs, leading to a red shift of the CDs with colorful photoluminescence from blue to brown. Jin et al. [22] developed a green hydrothermal method to obtain three emission colors of CDs using l-tyrosine (for blue CDs), o-phenylendiamine (for green CDs), and l-tyrosine/o-phenylendiamine mixture (for orange-red CDs), which also proved the dependence of CDs PL properties on their surface group and the excitontrapping functions of the surface functional groups. However, the regulation mechanism on the emission wavelength of CDs still needs to be verified through theoretical calculations and related experiments.
In this work, through theoretical and experimental verifications, the regulatory effects of sp 2 conjugated domain on the fluorescence wavelength of CDs were explored. Firstly, by changing the structural size and introducing amide groups in DFT theoretical calculation, four kinds of CDs structures were designed to investigate the structural effect of sp 2 conjugated domain on wavelength regulation theoretically. Secondly, with thiourea and p-phenylenediamine as the precursors, multicolor CDs emitting blue, green, and red fluorescence were prepared by regulating the reaction solvents. The analysis results confirmed that the increasing structure size and the introduction of amide groups would induce an increase of sp 2 conjugated domain, leading to the red shift of the CDs fluorescence wavelength. Finally, the obtained multicolor CDs were applied for LED device in the form of CDs@PVP films. The schematic illustration of multicolor CDs is shown in Fig. 1.

Experimental
Materials and measurements were shown in SI.

DFT simulation calculation
The ground-state geometry was optimized using the equation of DFT, and the excited states were calculated with the equation of linear-response time-dependent DFT (TD-DFT) at the optimized ground-state geometry [40]. All calculations are performed with the Gaussian 16 package (Rev. C.01) using the CAM-B3LYP functional and the 6-311G* basis set. Grimme's D3BJ dispersion correction was used to improve the calculation accuracy. The energy gap between HOMO and LUMO was calculated by Eq. 1.

Synthesis of CDs
The d-CDs were obtained via a facile one-step hydrothermal strategy. A total of 0.1-g thiourea and 0.1-g pphenylenediamine were dissolved in 10-mL DMF, and then, the mixture was dispersed by ultrasonic unit for 5 min. The obtained uniform solution was sealed in a Teflon autoclave and heated at 200 °C for 8 h. After that, the suspension was filtered by a 0.22-μm pore diameter microporous membrane. The purified solution was drying at 60 °C to obtain solid-state CDs.
The synthetic method of e-CDs and m-CDs was similar, except that the reaction solvents were changed to ethanol and methanol, respectively. In addition, to further verify the effect of sp 2 conjugated domain on the fluorescence wavelength of CDs, two kinds of CDs were prepared with the same process except that the reaction solvents were changed to formamide (f-CDs) and DMA (a-CDs).

Preparation of CDs@PVP fluorescent film
A total of 1-g PVP was dissolved in 20-mL filtered CDs solution, coated on the template after stirred evenly, and then solidified at room temperature for 24 h. Finally, the

Structure design and DFT theoretical calculation
In order to investigate the mechanism of CDs wavelength regulation, by changing the structural size and by introducing amide groups in the DFT theoretical calculation, four CD structures were designed to explore the effect of sp 2 conjugated domain size on the wavelength regulation theoretically [41]. As shown in Fig. 2a, the basic structure from 1 to 3, except the structural size, was similar, which could effectively verify the effect of structural size on the fluorescence wavelength of CDs. Furthermore, when the − OH group on the surface of structure 3 was replaced by − CONH − to give structure 4, the effect of amide groups on the fluorescence wavelength of CDs could be verified. These structures could be used to explore the influence of sp 2 conjugated domain on the HOMO-LUMO energy gap of CDs.
DFT theoretical calculations were carried out for energy gap between HOMO and LUMO of the four mentioned structures. Detailed calculation process was described in Sect. 2.1 [42,43]. From structure 1 to structure 3, in Fig. 2b, a trend of gradual decrease in turns (6.18 eV, 4.92 eV, 4.30 eV) in the energy gap was depicted, which indicated that the increase of the structural size led to a red shift fluorescence wavelength of CDs (the decrease of the energy gap would lead to the red shift of the CDs fluorescence wavelength) [44]. In Fig. 2c, the energy level gap of structure 3 and structure 4 was 4.30 eV and 4.28 eV, respectively, which indicated that the introduction of amide groups could induce the red shift fluorescence wavelength. Through two measures in theoretical calculation, it was improved that the increasing size of sp 2 conjugated domain could lead to the red shift of CDs fluorescence wavelength.

Synthesis, basic morphology, and structure
Based on the results of DFT theoretical calculation, the effect of sp 2 conjugated domain on the emission wavelength of CDs was expected to be further testified in the experiments. Thiourea and p-phenylenediamine were selected as precursors; three kinds of CDs, emitting red (d-CDs), green (m-CDs), and blue (e-CDs), were successfully synthesized by one-step hydrothermal reaction with DMF, methanol, and ethanol as the reaction solvents, respectively. In order to optimize the experimental conditions, CDs were synthesized at 160 °C, 180 °C, 200 °C, and 220 °C in each reaction solvent. Accompanied with the increasing reaction temperature, it could be seen that the fluorescence intensity firstly increased and then decreased (Fig. S1a). When the reaction temperature was 200 °C, the fluorescence intensity was the highest. In order to explore the optimal reaction time, CDs were synthesized under the reaction time of 3 h, 5 h, 8 h, 10 h, and 12 h, respectively. When reaction time was extended, the fluorescence intensity showed a trend of firstly increasing and then decreasing (Fig. S1b) and reached the highest fluorescence intensity at 8 h. The molar ratio of p-phenylenediamine to thiourea was set to be 0.5:1, 0.7:1, 1:1, and 1.2:1 for the synthesized CDs. The fluorescence intensity diagram of the CDs obtained was shown in Fig. S1c in the supplementary information. The results showed that when the molar ratio of p-phenylenediamine to thiourea was in the range of 0.5:1-1.2:1, the molar ratio affected the fluorescence intensity of the CDs slightly, which could be neglected. Based on this inquiry, the reaction conditions were optimized as follows: thiourea and pphenylenediamine were reacted in ethanol, methanol, and DMF for 8 h at 200 °C, respectively. The quantum yield of these CDs was tested to be 21.1% (e-CDs), 16.5% (m-CDs), and 8.3% (d-CDs), respectively. The fluorescence intensity basically remained unchanged after being stored for 5 months, indicating that the CDs could be stored for at least 5 months at room temperature (Fig. S1d). And the pH stability, salt resistance, was described in Fig. S1e-f, which combined to indicate that the synthesized CDs presented remarkable conjugate structure (UV absorption in Fig. S2) and had excellent stability in various environments. For a better demonstration of the CDs structural size, the transmission electron microscope (TEM), the X-ray diffraction (XRD), and Raman spectra were carried out. As seen in the TEM images (Fig. 3a-c), all three kinds of CDs presented uniform distribution of morphology and size. The average size of the synthesized CDs was 7.068 nm (d-CDs), 6.098 nm (m-CDs), and 3.451 nm (e-CDs) (Fig. 3d-f). Considering that the CDs emitting red, green, and blue in turns, it could be concluded that the fluorescence wavelength red shifts took place with the particle size increasing, which was probably caused by the increased size of sp 2 conjugated domain. The XRD test was carried out to verify whether the CDs have obvious conjugated structures (Fig. 3g). The peak positions located at 22°, 24°, and 25° for the three kinds of CDs belonged to (002) crystal plane of graphite phase carbon. This indicated that the three kinds of CDs had obvious conjugated structures [33]. The Raman spectra of the three CDs were presented in Fig. 3h. All the three kinds of CDs showed obvious D-band (1353 cm −1 ) and G-band (1561 cm −1 ). The I G /I D of d-CDs, m-CDs, and e-CDs was 1.1, 1.03, and 0.88, respectively, which confirmed that the order of the decreased size of the sp 2 conjugated domain was d-CDs, m-CDs, and e-CDs [45].
In order to explore the surface groups of the three kinds of CDs, zeta potential, Fourier transform infrared spectroscopy (FT-IR), and X-ray photoelectron spectroscopy (XPS) were carried out. The zeta potential results were shown in Fig. 3i; these CDs showed obvious peak at − 15.5 mV, − 16.1 mV, and − 23.2 mV. It was revealed that the surface states of the three CDs were different. From the FTIR spectra shown in Fig. 4a, these CDs had the same peaks at 820 cm −1 , 1395 cm −1 , 1510 cm −1 , and 3127 cm −1 . Specifically, 820 cm −1 was the characteristic peak of C-H of p-benzene, 1395 cm −1 could be attributed to the characteristic absorption peaks of C = S, and 1510 cm −1 could be attributed to the characteristic absorption peaks of C = C in aromatic structure. And 3127 cm −1 could be regarded as the stretching vibration of N-H. By comparing the peak positions of reaction materials and synthesized CDs, infrared absorption peak in 820 cm −1 (− C-H of p-benzene) and 1395 cm −1 (− C = S) appeared simultaneously in synthesized CDs, and the strength of absorption peak was obviously different from that of reactants, indicating that the CDs were successfully synthesized. The results proved that d-CDs, m-CDs, and e-CDs were the product of the joint synthesis of thiourea and p-phenylenediamine. The characteristic vibration peak of d-CDs was at 1620 cm −1 , related to the stretching vibration caused by C = O of amide, which was not found in m-CDs and e-CDs. Therefore, it could be speculated that amide groups were successfully introduced into the structure of d-CDs.
The full spectrum-specific results of XPS were shown in Fig. 4b. The peaks of the four constituent elements C1s, N1s, O1s, and S2p were observed at about ~ 284.1 8 eV, ~ 398.93 eV, ~ 531.81 eV, and ~ 162.23 eV for the CDs [38,46] (the specific peaks of the three CDs were shown in Table S1), respectively. In the high-resolution XPS spectra of C1s, four types of carbon signals at about 284.0 eV, 284.4 eV, 285.3 eV, and 287.4 eV could be found in the high-resolution C1s for d-CDs (Fig. 4c), which was in corresponding with C = C/C-C, C-O, C-N, and − CONH − bonds, respectively. There were three types of carbon signals at 283.7 eV, 284.4 eV, and 285.3 eV for e-CDs and m-CDs (Fig. 4d, e), corresponding to C = N, C-O, and C-N bonds, respectively. The XPS results proved that amide groups were successfully introduced into the structure of d-CDs.
Moreover, the contents of C = C/C = N in the three kinds of CDs were 31.25% (e-CDs), 37.11% (m-CDs), and 46.84% (d-CDs), respectively (Table 1). N1s spectra of the fluorescent CDs simulated three characteristic peaks at 397.5 eV, 398.8 eV, and 400.0 eV, which were corresponded to pyridine type N, graphite type N, and pyrrole type N, respectively (Figs. S3, S4, S5). The content of graphite type N was 32.01% (e-CDs), 45.99% (m-CDs,) and 49.83% (d-CDs). It was shown that the sp 2 conjugated domain was increasing in turns, which furtherly proved the connections of amide group introduction and the increasing size of sp 2 conjugated domain (Table 1).

Fluorescence properties of CDs
CDs emitting red (d-CDs, Em = 620 nm), green (m-CDs, Em = 520 nm), and (e-CDs, Em = 445 nm) were presented at Fig. 4f. Excitation-emission maps of e-CDs, m-CDs, and d-CDs were shown in Fig. 5a, b, and c. The e-CDs exhibited multiple luminescence centers covering from 400 to 480 nm, with the main emission center in blue region at 445 nm (Fig. 5a). The m-CDs exhibited a main green emission center at 520 nm (Fig. 5b). The d-CDs exhibited multiple luminescence centers covering from 605 to 635 nm, with the main emission center in red region at 620 nm (Fig. 5c). Combined with the above analyses, it could be  In order to further verify the above-mentioned mechanism, another two kinds of CDs (f-CDs, a-CDs) were prepared using thiourea and p-phenylenediamine as the precursors and formamide/DMA as the reaction solvents. The normalized fluorescence emission diagram was shown in Fig. 5d. In a comparison of the e-CDs and f-CDs, amide group was introduced in f-CDs, the fluorescence wavelength of f-CDs increased from 445 to 605 nm, which proved that the introduction of amide group on CDs could lead to the red shift of CDs fluorescence wavelength. Furthermore, we used DMA as the reaction solvent to introduce amide bond into the system. In a comparison of the e-CDs and a-CDs, amide group was introduced in a-CDs, the fluorescence wavelength of a-CDs increased from 445 to 650 nm, which proved that the introduction of amide group on CDs could lead to the red shift of CDs fluorescence wavelength. It could be verified again that the enlarged size of sp 2 conjugated domain could cause the red shift of the fluorescence wavelength of CDs.

Luminescent properties of LEDs based on CDs@ PVP fluorescent films
The aggregation quenching effect of CDs, caused by direct π-π interaction or excessive resonance energy transfer, may seriously limit the further application of CDs. To avoid this phenomenon, CDs were doped into PVP (polyvinylpyrrolidone) to suppress the aggregation induced luminescence quenching. The abundant surface chains of PVP could be able to prevent the graphitizing cores from π-π interactions. Thus, the CD particles embedded in the PVP kept an appropriate distance from each other, which effectively avoided the consequent fluorescence quenching of solid-state CDs. For a better application in LED devices, CDs@PVP fluorescent films were prepared, as shown in Fig. 6. The CDs@PVP fluorescent film under the irradiation of 365-nm ultraviolet lamp was shown in the Fig. 6a-c. The bright blue, green, and red fluorescence could be observed with naked eyes under the irradiation of 365 UV lamp, which indicated that aggregation quenching effect could be effectively suppressed. The morphology of CDs@PVP  Fig. 6d-f. The surface of the film prepared by this method was homogeneous smooth and flat without bubble wrinkles, greatly reducing the light loss in the light transmission process, which was conducive to further application in LED. Furthermore, the CDs@PVP solution was uniformly coated on commercial UV chips (365 nm) to prepare multicolor CDs@PVP-based LED. The corresponding CIE color coordinates of the CDs@ PVP-based LEDs (e-CDs@PVP in (0.17, 0.14), m-CDs@ PVP in (0.36, 0.51), d-CDs@PVP in (0.59, 0. 4)) were described in Fig. 6g-i. The results indicated that the synthesized CDs could be used in multicolor LED and would have excellent development in the field of photoluminescence.

Conclusions
In summary, through DFT theoretical calculation and experiments, it was found that the size of sp 2 conjugated domain could regulate the wavelength of CDs. Meanwhile, multicolor CDs (emitting red (620 nm), green (520 nm), and blue (445 nm) fluorescence) with the designed structures, in accordance with the DFT theoretical calculation results, were successfully prepared by the hydrothermal reaction. The CDs@PVP solid-state fluorescent films could be furtherly applied to multicolor LEDs. The approach provided an effective and novel strategy to realize the programmed regulation on the fluorescence wavelength of CDs, offering us full of potentials for the applications of CDs in the photoelectric device fields.
Author contribution Haiyan Bai, methodology, investigation, and writing -original draft. Xilang Jin, supervision, project administration, and writing -review and editing. Zhao Cheng, methodology and data curation. Hongwei Zhou, methodology and data curation. Haozhe Wang, data curation and formal analysis. Jiajia Yu, methodology and investigation. Jialing Zuo, methodology and data curation. Weixing Chen, supervision, validation, and methodology. Data availability Data available within the article or its supplementary materials.