Introduction

As a versatile carbon nanomaterial, carbon dots (CDs) have many fascinating properties such as non-toxicity, chemical stability, and unique fluorescence [1, 2]. Thus they have shown great potential in bioimaging [35], photocatalysis [6], electrocatalysis [7], and chemical sensing [811]. In 2004, CDs were first discovered accidentally during the separation and purification of single-walled carbon nanotubes [12]. Since then, this new class of carbon nanomaterials has garnered much focus of researchers, and various synthesis methods have been demonstrated for the preparation of CDs. Up to now, the main synthesis methods for the preparation of CDs consist of hydrothermal or solvothermal treatment [13], chemical ablation [3, 14], microware irradiation [15], ultrasonic treatment [16], laser ablation [17], and electrochemical carbonization [18]. However, all the approaches above-mentioned suffer from several drawbacks such as drastic processes, time consuming, and expensive precursors [2], which limit the large-scale preparation and wide application of CDs. Therefore, novel moderate and facile methods for synthesizing CDs from green and inexpensive precursors still need to be explored.

Ionic liquids (ILs) as an attractive “green solvent” have many unique properties, including high polarity, negligible vapor pressure, a wide electrochemical window, and good thermal stability. These properties make ILs have many advantages in the fabrication of the nanomaterials [19]. In recent years, ILs have been used to synthesize nano-structured metal materials [20] and other inorganic nanomaterials [21]. In 2009, Lu and co-authors reported the fabrication of nanoparticles from graphite electrode using ionic liquid-assisted electrochemical exfoliation [22]. Size-controlled CDs were synthesized through adjusting the applied voltages on the electrodes in the mixed electrolyte containing ILs and water [23]. Besides, ILs can also be used as precursors [24] and capping agent [25] for the preparation of CDs. However, the application of acidic ionic liquid in the preparation of CDs remains less studied until now. Acidic ionic liquid, which combined the advantages of mineral acid and ionic liquid, is an environmentally benign catalyst for many organic reactions, such as esterification, polymerization, and carbonylation [26]. This catalyst may also be useful in the synthesis of CDs and makes the preparation process become mild.

Herein, 1-butyl-3-methylimidazolium chloride ([Bmim]Cl) as a solvent and SO3H-functionalized acidic ionic liquid (SO3H-IL) as a catalyst were used for the preparation of CDs from microcrystalline cellulose (MCC) in this work. Because of the presence of acidic ionic liquid, cellulose could be carbonized at relatively low temperature and atmospheric pressure. The obtained CDs have excellent water solubility and photostability. In addition, the CDs had fluorescent response to Hg2+ and could be used as “turn-off” fluorescent probe for monitoring Hg2+ in aqueous solution.

Materials and methods

Materials

Microcrystalline cellulose (MCC, average particle size of 50 μm) and 1-butyl-3-methylimidazolium chloride ([Bmim]Cl, CP, >99 %) were purchased from J&K Chemical Company (Beijing, China). CaCl2, Hg(NO3)2·H2O, MgCl2, CuCl2, Co(NO3)2, BaCl2, CdCl2, and NiCl2 were purchased from Beijing Chemical Corp. MnCl2, Pb(NO3)2, and FeCl3 were purchased from Tianjin Chemical Corp. All the chemicals are of analytical grade and used as received without further purification. And all the aqueous solutions were prepared using doubly deionized water.

Synthesis of CDs

SO3H-functionalized acidic ionic liquid (1-(1-propylsulfonic)-3-methylimidazolium hydrogen sulfate) was synthesized according to the previously reported method [27]. For the preparation of CDs, MCC (0.3 g) and [Bmim]Cl (4.0 g) were added into a 50-mL round-bottom flask and stirred for 1 h at 100 °C until the transparent and homogeneous solution was obtained. Then the as-prepared SO3H-IL (0.4 g) was added to the reaction system and reacted for 4 h under stirring at 80 °C. The obtained dark sticky mixture was diluted with doubly deionized water and centrifuged for 10 min at 2000 rpm to remove the black precipitates. Then the supernatant was filtered through a 0.22-µm microporous membrane, and the obtained dark yellow CDs dispersion was dialyzed with a molecular weight cutoff of 500 Da for 24 h to remove ILs and other small molecular compounds.

Characterization

Ultraviolet–visible (UV–Vis) absorption and photoluminescence (PL) spectra were measured with a TU-1900 UV–vis spectrophotometer (Beijing Purkinje) and FP-6500 fluorescence spectrophotometer (Jasco, Japan), respectively. Fourier transform infrared (FTIR) spectrum of CDs was recorded on a Nicolet 6700 spectrophotometer (Nicolet) using KBr pressed disks. The high-resolution transmission electron microscopy (HRTEM) image was obtained with a JEOL 2010 electron microscope at an accelerating voltage of 200 kV. The X-ray photoelectron spectroscopy (XPS) measurements were made using a Thermo Scientific K-Alpha electron energy spectrometer using Al Ka (1486.6 eV) as the X-ray excitation source. Binding energy for C1s at 284.6 eV is standard. The absolute quantum yield (QY) and fluorescence lifetime of CDs were measured by FLS980 (Edinburgh Instruments Ltd., British) fluorescence spectrometer.

Detection of Hg2+

Detection of Hg2+ in pure water was performed in phosphate buffer solution (PBS, 10 mM, pH 7.0) at room temperature. In a typical assay, 40 μL of CDs solution was added to 1 mL of PBS. Then a calculated amount of Hg2+ was mixed together with the aqueous solution containing the same amount of CDs and reacted for 20 min before spectral measurements. To confirm the selectivity for Hg2+, the PL intensities of CDs with other metal ions including Cu2+, Co2+, Ca2+, Mn2+, Ni2+, Mg2+, Cd2+, Ba2+, Fe3+, and Pb2+ were inspected in the similar way.

Results and discussion

In the present reaction system, for the low-cost preparation of CDs, cellulose (the most abundant biorenewable material on Earth) was selected as a raw material. However, cellulose is difficult to be dissolved in water and common organic solvents [28]. [Bmim]Cl, which can form hydrogen-bonding networks between hydroxyl groups of cellulose and chloride ions [29], is a good solvent for the dissolution of cellulose. SO3H-IL is a benign catalyst for hydrolysis of cellulose owing to high catalytic activity for the fracture of glycosidic bonds. Additionally, SO3H-IL is also an excellent catalyst for many organic reactions, such as esterification, polymerization, carbonylation, aldol condensation, etc. [26]. Hence, these two ILs were selected as solvent and catalyst in this study and the preparation process of CDs is illustrated in Fig. 1. The potential mechanism for the formation of CDs may involve two steps: Firstly, hydrolysis of cellulose took place in the presence of SO3H-IL after dissolution of cellulose in [Bmim]Cl, generating sugars and some oligomers, some of which subsequently dehydrated to acids, aldehydes, phenols, etc. [30, 31]. Secondly, these products reacted with each other under the catalysis of SO3H-IL to produce soluble polymers which further developed into CDs after the consecutive processes of polymerization, aromatization, and ultimately generation and growth of carbogenic nuclei [32, 33]. For comparison, the experiments without adding SO3H-IL were carried out. There was no distinct carbonation phenomenon for cellulose even after reaction for 4 h at 150 °C.

Fig. 1
figure 1

The diagram of preparation process for CDs

To find the optimal reaction condition, we investigated the fluorescence spectra of resultant CDs under different reaction ratios and temperatures. The results indicated that reaction ratio and temperature have no obvious impact on the fluorescent properties of the resultant CDs in our experiments, but they have an effect on carbonation speed of cellulose. In order to use ILs effectively and synthesize CDs in short time and at low temperature, a mass ratio of [Bmim]Cl/cellulose/SO3H-IL of 4.0/0.3/0.4 and a reaction temperature of 80 °C were selected as the optimal reaction condition.

TEM image (Fig. 2a) shows that most of the obtained CDs exhibit well-dispersed near-spherical morphology. The diameters of most of the obtained CDs are distributed in the range of 3–13.5 nm (Fig. 2b) and with an average diameter value of 8.0 nm (100 random nanoparticles were selected for the sample). The high-resolution TEM image (HRTEM) reveals that CDs have crystalline structure with lattice fringe of 0.32 nm (the inset image of Fig. 2a).

Fig. 2
figure 2

a TEM image of CDs, inset: HRTEM image of single CD; b the particle size distribution histogram of CDs

In the FTIR spectrum (Fig. 3a), there is a broad and intense peak at around 3380 cm−1 which corresponds to the stretching vibration of O–H and N–H. The peaks at 2961, 2875, and 1401 cm−1 indicate the existence of C–H bond. The peaks at 1573 and 1123 cm−1 are due to the stretching vibration of C=C and C–O–C, respectively.

Fig. 3
figure 3

a FTIR spectrum and b XPS full-scan spectrum of CDs

The surface composition and chemical state of the as-prepared CDs were further investigated by X-ray photoelectron spectroscopy (XPS). The full-scan XPS spectrum (Fig. 3b) has four peaks at 532.41, 401.61, 284.83, and 167.89 eV, corresponding to C1s, O1s, N1s, and S2p, respectively. The elemental contents of sample are 71.51 % (C), 22.47 % (O), 4.33 % (N), and 1.69 % (S). The results reveal that besides abundant carbon and oxygen, the surface of obtained CDs also contains a small amount of nitrogen and sulfur. The carbon and oxygen may mainly come from cellulose, while nitrogen and sulfur may come from the ILs adsorbed on the CDs’ surface. The high-resolution XPS spectra of the C1s, O1s, N1s, and S2p are shown in Fig. 4. In the high-resolution C1s spectrum (Fig. 4a), the signals of five distinct carbon states C–C/C=C, C–S, C–N, C–O, and C=O present at 284.444, 284.8, 285.188, 286.245, and 287.551 eV, respectively [14]. The O1s spectrum shown in Fig. 4b has two peaks at 531.901 and 532.615 eV, which are attributed to C=O and C–OH/C–O–C groups, respectively. The high-resolution N1s spectrum (Fig. 4c) can be deconvoluted into two peaks at 399.5 and 401.635 eV, which represent N1s states in C=N and C–N [34]. The deconvoluted peaks at 167.529, 168.223, and 168.933 eV in the S2p spectrum (Fig. 4d) indicate that sulfur exists mostly in the form of –C–SO x – (x = 2, 3, 4) [9]. The analysis of FTIR and XPS spectra shows that the surface of the as-synthesized CDs has multiple oxygenated groups like –COOH and –OH.

Fig. 4
figure 4

High-resolution XPS spectra of a C1s, b O1s, c N1s, and d S2p

To further explore their optical properties, the UV–Vis absorption and photoluminescence (PL) spectra of CDs were measured and are shown in Fig. 5a. Two obvious absorption peaks centering at 210 and 276 nm were seen in the UV–Vis absorption spectrum, which should be ascribed to n–π* transition of C=O and π–π* transition of C=C [35]. The synthesized CDs have excitation-dependent emission behavior (Fig. 5b), and the intensity of PL emission peak at 450 nm is the highest when excitation wavelength is 360 nm. The photographs (inset in Fig. 5a) show that the aqueous solution of CDs emitted powder-blue fluorescence under 365 nm UV light (right) while appearing as yellow under room light (left). The absolute quantum yield (QY) of CDs was measured to be 4.7 % using an integrating sphere attached to an FLS980 fluorescence spectrometer.

Fig. 5
figure 5

a The UV–Vis absorption, PL excitation, and emission spectra of CDs, insets: photographs of CDs under room light (left) and 365 nm UV lamp (right) of CDs aqueous solution; b PL emission spectra of CDs aqueous solution at an excitation wavelength from 360 to 450 nm

The photochemical stability of CDs at different conditions was also investigated by recording the PL intensities of CDs at different ionic strengths, pH values, and irradiation times under UV lamp. As shown in Fig. 6a, the PL intensities remain unchanged at different ionic strengths regulated by concentrations of NaCl. Therefore, the CDs have excellent stability when exposed to environments with high ionic strengths. The effect of UV irradiation time on PL intensity is shown in Fig. 6b. A slight decrease of the PL intensity of CDs is observed during continuous exposure of 100 min, which demonstrates that the surface states of CDs are unstable when exposed to UV lamp during this time. However, CDs have resistance to photobleaching in some degree, because PL intensities are unchanged with the irradiation time from 100 to 180 min. The PL intensities of CD in various pH values from 1 to 13 are given in Fig. 6c. Overall, PL intensity is stronger in the acidic condition than under alkaline conditions and reach the highest at pH 7. This result is similar to that reported in the literature [36], which is perhaps owing to the variation of functional groups on the surface of CD at different pH values. The PL intensity of CDs is higher in the pH range from 3 to 7 than in other pH values, indicating that they have more application potential in acidic and neutral conditions.

Fig. 6
figure 6

a The PL intensity of CDs at different ionic strengths regulated by concentrations of NaCl from 0 to 1 M; b the PL intensity of CDs at different irradiation times with an UV lamp; c the effect of pH value on PL intensity of CDs; and d the fluorescence lifetime of CDs

At present, although the actual PL mechanism of CDs is still not yet entirely clear, there is mounting evidence that emission arises from both an intrinsic band gap resulting from confined sp2 conjugation in the core of CDs and the surface state that can be excited either directly or by energy transfer from an intrinsic band [37]. The luminescence decay profile of the CDs excited at 360 nm was recorded using a time-correlated single-photon counting technique (Fig. 6d). The data of fluorescence lifetime are very well fitted to a triple exponential, and the average fluorescence lifetime was calculated to be 3.24 ns. Such a short lifetime demonstrates that the possible luminescent mechanism is the radioactive recombination of excitations [38].

To evaluate the feasibility sensitivity of the CDs as sensors monitoring Hg2+ in aqueous solutions, the PL intensities of CDs with different concentrations of Hg2+ in PBS (pH 7) were studied. The PL emission spectra of CDs with different concentrations of Hg2+ in the range of 0–80 μM (Fig. 7a) reveal that PL intensity of CDs at 450 nm decreases gradually with the increase of concentration of Hg2+. This suggests that Hg2+ can effectively quench the fluorescence of CDs, which may be ascribed to the non-radiative electron transfer from the excited state to the d-orbital of Hg2+ [39]. The relative PL intensities (F/F 0) displayed a good linear relationship versus the concentration of Hg2+ in the range of 6–80 μM (inset in Fig. 7a) with a linear equation of F/F 0 = −0.00301c + 0.98315 (R 2 = 0.9863), where c is the concentration of Hg2+, and F and F 0 are FL intensities at 450 nm in the presence and absence of Hg2+, respectively. Furthermore, the relative standard deviation (RSD) was 1.02 % for five repeated measurements of 60 μM Hg2+, showing an excellent reproducibility of these fluorescent sensors. The limit of detection (calculated according to a signal-to-noise ratio of S/N = 3) was estimated to be 1.6 μM, which is higher than the acceptable value mandated by the EPA for drinking water. However, the sensors can be useful in detecting inorganic mercury in samples of biological products, drugs, fish, and also in samples where regulations for mercury are less stringent [40].

Fig. 7
figure 7

a PL emission spectra of CDs with different concentrations of Hg2+ (0, 5, 6, 9, 20, 30, 50, 60, 70, and 80 μM) in PBS (10 mM, pH 7) at an excitation of 355 nm, inset: relative fluorescence intensity (F/F 0) versus concentration of Hg2+ (from 6 to 80 μM); b the PL response of CDs to various metal ions including Cu2+, Co2+, Ca2+, Mn2+, Ni2+, Mg2+, Hg2+, Cd2+, Ba2+, Fe3+, and Pb2+ in PBS (10 mM, pH 7). F and F 0 correspond to the fluorescence intensities of the CDs with and without 60 mM of different metal ions, respectively

We have also evaluated the selectivity of this analytical system. The obtained CDs have more obvious fluorescence quenching in the presence of Hg2+ than other metal ions including Cu2+, Co2+, Ca2+, Mn2+, Ni2+, Mg2+, Cd2+, Ba2+, Fe3+, and Pb2+ (Fig. 7b). This phenomenon indicates that the CDs have selectivity for Hg2+, which may due to the factor that the carboxylic and hydroxyl on the surfaces of the CDs have stronger affinity to Hg2+ than to other metal ions.

Conclusion

We developed a moderate method for the preparation of fluorescent CDs from cellulose (green, low-cost, and commonly available biomass) using [Bmim]Cl as a solvent and SO3H-functionalized acidic ionic liquid as a catalyst. The obtained CDs have good water solubility, excitation-dependent emission behavior, excellent stability in high ionic strength, and resistance to photobleaching. The as-prepared CDs, with any surface functionalization, have a certain degree of sensitivity and selectivity for Hg2+, indicating that the CDs have potential to be used as “turn-off” fluorescent sensors for the detection of Hg2+ in aqueous solution.