Introduction

Fluoroquinolones (FQs) are widely applied as veterinary and human medicines to fight against many Gram negative and Gram positive organisms [1]. Among fluoroquinolones, norfloxacin (NOR) and ciprofloxacin (CIP) are the most widely used antibiotics [2]. Antibiotics residues have been found in a wide range of environmental surface water and food. Especially, their residues in milk products have reported and drew wide attention. Long-term exposure to such milk products contaminated with FQs residues can lead to an increased allergic reactions, liver damage, gastrointestinal disturbance and drug-resistance of microbial strains [3,4,5,6]. Consequently, monitoring of FQs residues is a crucial issue.

To date, a variety of methods have been applied for determining FQs residues, which include liquid chromatography/post-column derivatization [7], surface enhanced Raman spectroscopy (SERS) detection [8], liquid chromatography with fluorescent detection (LC-FLD) [9], liquid chromatography-electrospray ionization-tandem mass spectrometry (LC-ESI-MS/MS) [10], voltammetric [11], chemiluminescence [12], colorimetry [1], spectrofluorimetric detection [13,14,15]. The widely used HPLC methods or LC-ESI-MS/MS for the determination of NOR and CIP or similar fluoroquinolone in complex matrices involve tedious clean-up steps prior to chromatography, as well requirement of sophisticated instrumentation and time consumption. Also, ordinary spectrometric detection can’t be applied to complex food-based matrices.

Carbon dots (CDs) are considered as one of the most extensively optical nanomaterials in the last few years [16, 17]. They exhibit excellent properties including outstanding water solubility, good biocompatibility, low toxicity, robust chemical inertness and ease of preparation [18,19,20,21]. These advantages make CDs superior to classic fluorescent molecules and the reported fluorescent nanoprobes based on CDs focused on the bioimaging [22], detection of metal ions [23], protein [24], hormone [25], glucose [26] and other small molecules [27, 28]. Theoretically, any phenomenon of fluorescence change (intensity, wavelength, anisotropy, or lifetime) related to the concentration of different analytes demonstrates the potential to be used as probes. The mechanism of assay is significant for fluorescent probes, the conventional mechanisms include fluorescence resonance energy transfer (FRET), twisted intramolecular charge transfer (TICT), electronic energy transfer (EET), intramolecular charge transfer (ICT), and photo-induced electron transfer (PET), photo-induced charge transfer (PCT), inner filter effect (IFE) [29,30,31,32].

In general, spectrofluorimetric detection, because of its simplicity and rapidity, the method has already been described for determination of FQs through sensitization of fluorescence. According to the previous literature [33], in order to accurately recognize target molecule in complex matrix at trace levels, magnetic solid-phase extraction (MSPE) was used as a preconcentration method. In addition, CDs are hydrophilic, which provides a good opportunity to produce hydrogen bonds with analytes [34, 35]. Strong hydrogen bonds only include F- H…F−, O –H…O− and O+ -H…O, which are always a few two-center bonds of short distances and are directional, with energies higher than 41 kJ·mol−1 [36]. Herein, we present a novel strategy to detect NOR or CIP residues by carbon dots fluorescence enhancement. Fluorescent intensity of the S-CDs is enhanced by norfloxacin or ciprofloxacin based on strong hydrogen bond interaction and charge transfer, it’s the theoretical basis in our work. It should be pointed out that this is the first time that there has been report on the use of fluorescent S-CDs for detection of NOR or CIP residues in real sample matrices coupled with magnetic solid-phase extraction. This method exhibits many advantages, for instance, environmentally friendly, improvements in automation, short time requirement, reduced solvent consumption.

Experimental section

Reagents and materials

Poly(4-styrenesulfonic acid-co-maleic acid) (PSMA), norfloxacin (NOR), ciprofloxacin (CIP), ofloxacin (OFL), difloxacin (DIF), fleroxacin (FLE), pefloxacin (PEF), enoxacin (ENO) and levofloxacin (LEV) were purchased from Aladdin Chemistry Co. Ltd. (http://www.aladdin-e.com/). Citric acid (CA), dibasic sodium phosphate (Na2HPO4), sodium hydroxide (NaOH), hydrochloric acid (HCl), sodium dodecyl sulfate (SDS), 2-(5-bromo-2-pyridylazo)-5-(diethylamino)phenol (5-Br-PADAP) were purchased from Sinopharm Chemical Reagent Co., Ltd. (http://www.sinoreagent.com/). Ferric chloride (FeCl3•6H2O), ammonium ferrous sulfate ((NH4)2Fe(SO4)2•6H2O), ammonia solution (NH3•H2O), methanol and ethanol were supplied by Tianjin Fengchuan Chemical Reagent Co., Ltd.(http://www.tjhxsj.cn/). Standard stock solutions of NOR and CIP were prepared by double distilled water containing 10 mM HCl. They were all stored at 4 °C in the dark and can be stable for 2 months, the working solutions were obtained by appropriate dilution of the stock solutions with double distilled water. All Chemicals used were at least of analytical reagent grade and were used without further purification.

Apparatus

The UV-Vis absorption spectrum of CDs was recorded by a UV-2600 UV-Vis spectrophotometer (SHIMADZU, Japan). Fluorescence measurements were carried out on a G9800A Cary Eclipse fluorescence spectrophotometer (Agilent Technologies, USA). The images of particle CDs were obtained by a Tecnai G2 F30 transmission electron microscope (TEM) (FEI, USA) at 200 kV. The functional groups on the CDs were measured with a TENSOR27 Fourier transform infrared spectroscopy (FT-IR) spectrometer (Bruker, Germany). Other instruments including: a D8-advance X-ray diffractometer (XRD) (Bruker, Germany), a Thermo Scientific K-Alpha X-photoelectron spectroscopy (XPS) (Thermo Fisher Scientific Inc. U.S.A.), a vortex mixer (Hanuo Instrument Co., Ltd., XH-B, Shanghai, China), high speed centrifuge (Shanghai Surgical Instrument Factory, 80–2, Shanghai, China), vacuum drying oven BPZ-6033 (Shanghai, China), a digital display temperature control water-bath Model XMTB (Central Experimental Furnace Co., Ltd., Tianjin, China), a pH meter pHS-3C (Shanghai Leici Instruments Factory, China) .

Preparation of S-CDs

The hydrothermal synthesis of S-CDs is described as follows: Poly(4-styrenesulfonic acid-co-maleic acid) (3.0 g) was dissolved in deionized water (100 mL). The solution was transferred into a teflon-lined stainless steel autoclave (150 mL) and heated at 220 °C for 5 h. After the reaction, the autoclave was taken out from the oven and cooled to room temperature naturally. The yellow aqueous solution was purified by filtration through 0.22 μm filter and then centrifuged at 10,000 rpm for 15 min to remove large particles, stored at 4 °C before used (quantum yield measurement details is displayed in Electronic Supplementary Material).

Preparation of milk samples

Bovine raw milk was purchased from a local farm in Kunming City (Yunnan Province, China). Preliminary analyses showed that no FQs residues were present to disturb the preparation of matrix-matched calibration standards and spiked samples for the validation of the method. For the precipitation of proteins and extraction of residual drugs, 50 mL of bovine raw milk sample was heated to 40 °C in a water bath. 0.5 M HCl was added slowly into the sample described above under continuously stirring, the pH of the sample solution should be adjusted to 4.7. The resulting suspension cooled to room temperature naturally, then centrifuged at 4000 rpm for 5 min to remove precipitation of proteins. Sample solution was stored at 4 °C before used.

MSPE procedure

In a typical process, magnetic nanoparticles (MNPs) were synthesized by the in situ chemical coprecipitation (see details in Electronic Supplementary Material). The MSPE procedure was carried out as follows: 120 μL of the above MNP suspension was added to 5 mL sample solution. The mixture was vortex-mixed for 120 s to make NOR or CIP fully adsorbed by the MNPs. Then, MNPs were separated quickly from the sample solution by using an external magnet, the supernatant was decanted and the MNPs were eluted by 50 μL of methanol. Finally, the eluate was isolated from MNPs by using an external magnet and combined with S-CDs for subsequent spectrofluorimetric detection analysis.

Detection of NOR and CIP

As shown in Fig. 1a, determination of NOR and CIP was performed in phosphate buffer (PB) (30 mM, pH 6.0) at room temperature. The standard stock solution of NOR and CIP was prepared by dissolving the appropriate amount of norfloxacin (NOR) and ciprofloxacin (CIP). In a typical run, 25 μL S-CD solution, 500 μL of PB and different amounts of NOR or CIP were added into a 10.0 mL colorimetric tube, and the mixture was diluted to 4.0 mL with ultrapure water and vortex-mixed for 1 min to mix completely before spectral measurements. The fluorescence spectra were collected with excitation/emission maxima at 324/412 nm, and the slit widths of the excitation and emission were both 5 nm. In order to assess the influence of co-existing interferences on the fluorescence of S-CDs and evaluate the selectivity for NOR or CIP, the similar procedure was performed for the detection of other fluoroquinolones: ofloxacin (OFL), difloxacin (DIF), fleroxacin (FLE), pefloxacin (PEF), enoxacin (ENO), levofloxacin (LEV).

Fig. 1
figure 1

a Processing diagram for the norfloxacin and ciprofloxacin determination process. b Schematic illustration of carbon dots fluorescence enhancement

Full size image

Results and discussion

Choice of materials

Carbon dots (CDs) are promising in fluorescent assays, but they are limited in various areas of applications [37]. Particularly, heteroatom-doping results in a significant improvement in electronic and fluorescence properties of CDs. Several strategies have been reported in the literature on the doping of CDs with various heteroatoms such as N, S, B or P [25, 38]. According to the previous literature [39], in order to broaden their applications in analysis and probes, S-CDs were synthesized from S-containing polymer precursors. The fluorescence capability of the S-CDs was related to the specific configuration of the heteroatoms. S-CDs do not differ significantly in the surface chemistry from other CDs, and sulfonic acids acting as chromophores played the most important role. Sulfonyl plays a pivotal role in S-CDs acting as sensitive and selective assay towards NOR and CIP. Consequently, poly(4-styrenesulfonic acid-co-maleic acid) was selected to be polymer precursors.

Characterization of S-CDs

The transmission electron microscopy (TEM) technique was used to explore the morphology and particle size distributions of S-CDs. As shown in Fig. 2a and Fig. S1, the S-CDs had high dispersity and uniform spherical shapes, and with a size distribution within the range of 3–10 nm and an average diameter of about 5.4 nm.

Fig. 2
figure 2

a The TEM image of the S-CDs drop casted on copper grid (inset: size distributions). b FT-IR spectra of S-CDs

Full size image

The surface chemistry of S-CDs was studied using Fourier transform infrared (FT-IR) spectroscopy. The typical FT-IR spectra of S-CDs was shown in Fig. 2b. As can be seen, a broad peak at 3456 cm−1 reveals the existence of O–H stretching vibrations of hydrogen bond. Several sharp peaks at 1180 cm−1, 1122 cm−1, 1042 cm−1 correspond to –C(=S)– stretching vibrations, respectively. Characteristic absorption bands can be observed for the –COCH2– bending vibration mode at about 1405 cm−1. Also, the peak appearing at 1581 cm−1 may be caused by the asymmetric and symmetric stretching vibration of COO−. They can be assigned to the existence of carboxyl bending. In addition, the absorptions at 2939 and 618 cm−1 display C–H stretching vibrations, these results indicate the generation of S-CDs.

The XPS spectrum provided more convincing evidence for elemental contents and surface groups of S-CDs. As shown in Fig. S2a, the full scan XPS spectra presented distinct peaks locating at 197.6, 220.0, 283.2 and 533.6 eV, which are attributed to S2P, S2S, C1s and O1s, respectively. This further corroborated that the S-CDs mainly contained carbon, sulfur, and oxygen. The high resolution spectrum of C1s (Fig. S2b) displayed four different types of surface components, corresponding to C–C/C=C at binding energies of 283.0 eV, C–S at 283.8 eV, and O–C=O at 287.0 eV. In addition, the high resolution spectrum of S2P (Fig. S2c) was fitted with two peaks at 167.2 eV and 168.3 eV, which are ascribed to S=O and S–C bonds. The high resolution O1s spectrum of S-CDs (Fig. S2d) is fitted with three peaks at 530.4 eV, 531.2 eV and 534.8 eV, which are ascribed to C–O, C=O and S=O bonds. The surface components of S-CDs determined by the XPS were in agreement with that of FT-IR results. According to the FT-IR and XPS results, sulfur was co-doped in CDs, and their related functional groups, including carboxyl, hydroxyl and sulfonyl existed on their surface (X-ray diffraction (XRD) pattern details is displayed in Electronic Supplementary Material).

Spectroscopic properties of S-CDs

To explore the optical properties of the S-CDs, UV-Vis absorption spectra and fluorescence spectra of the S-CDs in aqueous solution were investigated as shown in Fig. 3. It was observed that a characteristic UV-Vis strong absorption peak was located at 262 nm (Fig. 3a). The optical absorption peak was typically assigned to the n–π* transition and the contribution of surface moieties. The electron-rich nature of unsaturated groups with sulfur atoms changed the charge density of S-CDs, resulting in the energy transfer to the excited π* state of the sp2 cluster of S-CDs. As shown in Fig. 3b, the maximum emission intensity is centered at 412 nm with an excitation wavelength of 324 nm (QY = 12%). The S-CD solution is faint yellow under daylight and emit a brighter blue fluorescence under the excitation of 365 nm UV light. Moreover, the S-CDs exhibited excitation-dependent fluorescence behavior. As shown in Fig. S4, when the excitation wavelength increased from 264 nm to 404 nm, the fluorescence emission peak was obviously shifted to longer wavelength with variable fluorescence intensity. It may be attributed to different sized carbon nanoparticles in the S-CD solution, these excitation tunable emission spectra are considered to be the versatile characteristic of S-CDs. This property may be due to quantum confinement, size distribution or the presence of emissive traps on the surface.

Fig. 3
figure 3

a UV-Vis absorption spectra of the (a) S-CD solution black, (b) S-CDs + CIP mixture, (c) S-CDs + NOR mixture. The strong absorption peak was located at 262 nm. b Fluorescence excitation and emission spectra of S-CDs, the excitation/emission maxima at 324/412 nm. The insert is a photograph of S-CD solution under sunlight (left) and 365 nm UV light (right), respectively

Full size image

Optimization for the magnetic solid-phase extraction conditions

Several parameters, including the amount of MNPs, the vortex extraction time for adsorption, sample solution pH, and desorption conditions were investigated to achieve the best extraction efficiency. All the experiments were performed in triplicate and the means of the results were used for optimization. In summary, based on the all optimization results, the addition of 120 μL MNP suspension, vortex for 2 min for NOR and 1.5 min for CIP, pH 6.0, elution with only 50 μL methanol for NOR and CIP were employed for the following studies.

Analysis of fluorescence enhancement mechanisms

The fluorescence enhancement mechanisms may attribute to the valuable phenomenon of hydrogen bond interaction and charge transfer between S-CDs and NOR or CIP. On the one hand, taking into account the structural similarity of NOR and CIP, as well their structure-reactivity relationship toward S-CDs. As shown in Fig. 1b, the fluorescent probe can be associated with surface states. In accordance with characterization of S-CDs, the unique assay properties of S-CDs arise from the combination of material and dimensionality. The large surface area-to-volume ratios of S-CDs, and it contains abundant carboxyl, hydroxyl and sulfonyl groups. These advantages provide aqueous solubility and opportunities for combination with NOR or CIP based on hydrogen bond interaction. The special and strong hydrogen bond may occur between the carboxyl of S-CDs and piperazine ring, as well between the hydroxyl and carboxyl or fluorine of NOR/CIP. Besides, the carbon on the aromatic ring also has a relatively strong electron-withdrawing ability, the Ar-H…O hydrogen bond also can be formed. On the other hand, charge transfer also plays an important role. It can be also related to the electron accepting nature of the –SO3H groups, and the electron donating character of the target FQs. The conjugation system of NOR/CIP can be connected to S-CDs by charge transfer. The introduction of NOR/CIP to S-CDs may promote the polymerization of C=C bonds. As shown in Fig. 3a, Fig. S5 and Fig. S6, compared with S-CDs blank, the UV-Vis absorption and fluorescence intensity of S-CDs + NOR/CIP are markedly enhanced but the wavelength is almost no changed. The synergy effect of the hydrogen bond and charge transfer can promote the generation of larger chromophores and fluorophores. The conjugates produce more intense emission under same excitation [29, 31, 36, 40].

Sensitivity and selectivity measurements

Under the optimized condition, the fluorescence emission spectra of S-CDs + NOR/CIP system with different concentrations of NOR and CIP added were measured. As shown in Fig. S5 and Fig. S6, the fluorescence intensity sharply increased along with the increasing concentration of NOR and CIP from 0 to 2.0 μM, but had no effect on the shape of fluorescence spectrum, which revealed the assay was very sensitive to NOR and CIP.

In order to prove the selectivity of the method, the selectivity of extractions by the MNPs was investigated. In conclusion, some potential interferences such as some metal ions, tetracycline and other organic pollutants (or hormone) can hardly be adsorbed. As far as we know, no literature reports that this magnetic nanoparticles was used to adsorb these potential interferences. It proves the practicability of adsorption method. Furthermore, to investigate the selectivity of S-CDs toward different FQs residues, including ofloxacin (OFL), difloxacin (DIF), fleroxacin (FLE), pefloxacin (PEF), enoxacin (ENO), levofloxacin (LEV), norfloxacin (NOR), and ciprofloxacin (CIP). The fluorescent properties of S-CD aqueous solution under diverse FQs residues were examined. The influence of other FQs residues on the fluorescent behavior of S-CDs is evaluated as depicted in Fig. 4.The fluorescence intensity remained nearly unchanged when the foreign substances was added. And fluorescence enhancement only occurred in the system containing NOR or CIP. The possible reasons of high selectivity is described as follows: FQs is a fluorinated 4-quinolone ring containing carboxylic acid, fluorine and a piperazine ring substitution at position 3, 7 and 8. Strong hydrogen bond occurs between the carboxyl of S-CDs and amine of the piperazine ring preferentially [41]. Also, this strong hydrogen bond plays a major role in all hydrogen bond “bridge”. Besides, quinolone ring is the essential conjugation system of FQs, charge transfer occurs between the quinolone ring and S-CDs. However, as shown in Table S1, the difference between all molecular structures was proposed. For most FQs, the amine of the piperazine ring was substituted by methyl, which leads to the strong hydrogen bond can’t be formed. It is negative for charge transfer and further hinders the production of sensitization. In addition, the benzene ring in the quinolone ring was replaced by a pyridine ring. The conjugation system and electron donating character of FQs were weakened, such as ENO. Thus, the system allows for the specific determination of NOR and CIP even in the presence of the potentially interfering substances mentioned above.

Fig. 4
figure 4

The fluorescence enhancement effect of different FQs towards S-CDs with different concentrations: 0.2 μM (black column) and 2.0 μM (red column). From left: 1. S-CDs blank, 2. OFL, 3. DIF, 4. FLE, 5. PEF, 6. ENO, 7. LEV, 8. CIP, 9. NOR. The data was acquired at excitation/emission peaks of 324/412 nm

Full size image

Optimization for the determination of CIP & NOR

The fluorescence enhancement efficiency of the assay system closely relied on some experimental conditions. In this work, we optimized the pH value of aqueous solutions, and incubation time of the assay system. As shown in Fig. S7, the fluorescence intensity of S-CDs (with excitation/emission maxima at 324/412 nm) was investigated under various pH solutions (2.0 to 12.0). An increase in pH from 2.0 to 6.0 results in the increased fluorescence enhancement efficiency of the established system, whereas a further increase in pH from 6.0 to 12.0 leads to a gradual decrease. Such an observation suggests that the fluorescence intensity of the S-CDs strongly depends on the pH value of the system. In the experiments, the carboxyl, hydroxyl and sulfonyl groups may be protonated in acidic solutions, which will lead to the aggregation of S-CDs, resulting in the fluorescence quenching at low pH value. On the other hand, the results may be attributed to neutralization reaction in alkaline solution [42]. This phenomenon further confirms that the carboxyl, hydroxyl and sulfonyl groups are present on the surface of S-CDs. These functional group endow the S-CDs with reasonable hydrophilic properties, stability and dispersibility in water. Consequently, we selected 6.0 as the optimal pH for our study.

We sequentially optimized the incubation time. The effect of incubation time on the fluorescence intensity of the system is shown in Fig. S8. The fluorescence intensity with excitation/emission maxima at 324/412 nm was recorded from 0 to 30 min. No significant changes in fluorescence intensity were observed after an incubation time of 10 min. To ensure the consistency of the whole experiment, it is important to record the stable fluorescence signal. Thus, 10 min is conservatively chosen as the optimum incubation time.

Norfloxacin (NOR) and ciprofloxacin (CIP) detection

Under the optimized condition, different concentrations of NOR and CIP solution were added to the S-CD solution, respectively. The fluorescence emission intensity at 412 nm was measured to study the sensitivity of the S-CDs. With the increase of NOR and CIP, the fluorescence intensity of the S-CDs at 412 nm increased gradually. As shown in Fig. 5a and b, the value of log (F/F0) increased gradually with increasing NOR or CIP concentration. The linear calibration in the range from 0.02 to 1.25 μM and 0.02 to 1.0 μM, log (F/F0) = 0.87 CNOR + 0.27 and log (F/F0) = 1.01 CCIP + 0.24 were obtained, respectively, where F0 and F were the fluorescence intensities of S-CDs at 412 nm in the absence and presence of NOR or CIP. And the limit of detection was estimated to be 4.6 nM (NOR) and 6.7 nM (CIP) based on three times signal-to-noise ratio. The results demonstrate that the method can be employed as an excellent alternative for trace analysis of NOR or CIP.

Fig. 5
figure 5

The relationship between log (F/F0) and the concentration of (a) NOR (calibration curve showing the linear dependence of log (F/F0) values on the concentration from 0.02 to 1.25 μM. (n = 3)). (b) CIP (calibration curve showing the linear dependence of log (F/F0) values on the concentration from 0.02 to 1.0 μM. (n = 3)). The data was acquired at excitation/emission peaks of 324/412 nm

Full size image

Real sample detection

In order to prove the practical applicability in real sample analysis, this system was applied to norfloxacin and ciprofloxacin residues determination in bovine raw milk samples. The corresponding results of fluorescence measurements before and after the standard addition method were listed in Table 1. The average recoveries of norfloxacin and ciprofloxacin residues in bovine raw milk samples were 96.2%~105.2% and 92.3%~102.5%, respectively. Relative standard deviations (RSDs) of the two compounds of samples were both found lower than 3.7%.

Table 1 Determination of NOR and CIP residues in bovine raw milk samples using the fluorescence enhancement method (n = 3)
Full size table

Moreover, the same samples were analyzed simultaneously by fluorescence enhancement detection without MSPE for comparison. As listed in Table 1, in terms of average recoveries and relative standard deviations (RSDs), it can be concluded that the analytical performance of the method was obviously better than fluorescence enhancement detection without MSPE. It proves MSPE is a necessary part in this method. Compared with the published methods for the determination of FQs, our method shows higher sensitivity, lower LODs (Table 2).Therefore, it is expected that the results summarized in Table 1 further demonstrated the practical value of this method for norfloxacin and ciprofloxacin residues in real samples.

Table 2 The comparison of different analytical methods for the determination of FQs
Full size table

Conclusion

In conclusion, a novel assay for detection of norfloxacin (NOR) and ciprofloxacin (CIP) was constructed by the combination of fluorescence enhancement technology and magnetic solid-phase extraction (MSPE). Based on the strong hydrogen bond interaction and charge transfer between S-CDs and NOR/CIP, the fluorescence probe exhibits excellent selectivity and sensitivity towards analytes. Moreover, the analytical method was successfully used to detect NOR or CIP in real milk samples with a detection limit of 4.6 nM (NOR) and 6.7 nM (CIP). To the best of our knowledge, this is the first time that there has been report on the use of fluorescent S-CDs for detection of NOR or CIP residues coupled with magnetic solid-phase extraction. We believe that the property of distinguishing NOR from CIP in real samples by CDs probe will be promising in wider scope.