Porous SnO2 nanoparticles based ion chromatographic determination of non-fluorescent antibiotic (chloramphenicol) in complex samples

Nowadays, there are rising concerns about the extensive use of the antibiotics such as chloramphenicol (CAP), has threatened the human life in the form of various vicious diseases. The limited selectivity and sensitivity of confirmatory techniques (UV and electrochemical) and non-fluorescence property of CAP make its determination a challenging task in the modern pharmaceutical analysis. In order to redeem the selective, sensitive and cost-effective fluorescence methodology, here by the dual role of synthesized porous SnO2 nanoparticles were exploited; (i) a porous sorbent in a µ-QuEChERS based sample preparation and as (ii) a stimulant for the transformation of non-fluorescent analytes namely CAP and p-nitrophenol (p-NP) into their respective fluorescent product. We report a green, simple, selective and cost effective ion chromatographic method for CAP sensitive determination in three complex matrices including milk, human urine and serum. The synthesized sorbent not only selectively adsorbed and degraded the matrix/interferences but also selectively reduced the non-fluorescent antibiotic CAP into a fluorescent species. This developed ion chromatographic method exhibited good selectivity, linearity (r2 ≥ 0.996) and limit of detection (LOD) was in the range 0.0201–0.0280 µg/kg. The inter- and intraday precisions were also satisfactory having a relative standard deviation (RSDs) less than 14.96% and excellent recoveries of CAP in the range of 78.3–100.2% were retrieved in various complex samples.

great pharmacokinetics properties and low price have coerced its illegal use in livestock and aquaculture, especially in the developing countries 16,17 .
A number of chromatographic methodologies have been investigated for CAP determination after extracting it from different complex matrices by applying expensive, time consuming, tedious and non-environmental friendly extraction methods such as solid phase extraction (SPE) 18 , ultrasound-assisted dispersive liquid-liquid microextraction (DLLME) 19 , aqueous two-phase extraction system (ATPES) 20 and QuEChERS sample preparation method 21,22 . However, the previous analytical techniques have various limitations; for examples, gas chromatography coupled with electron capture detector (ECD) is sensitive technique but requires complex chemical derivatization 23 , liquid chromatography coupled to mass spectrometry (LC-MS) is rapid and sensitive techniques but these instruments are too expensive for routine analyses and out of access in developing countries laboratories 5,[24][25][26] . Unlike with UV, the LC coupled with fluorescence detector is a simple, cheap, highly selective and sensitive technique for analytes determination in complex samples but the non-fluorescence property of CAP has made its use limited 27,28 . Therefore, various indirect fluorescence methods i.e. derivatization 29 , florescent probe 30 , fluorescent nanosensor 31 and photochemical induce detection were applied for its selective and sensitive in various complex matrices 32 . However, all above-mentioned techniques are complex, laborious and involved separate rigorous and expensive sample preparation methods. In order to reinstate cheap, selective and sensitive fluorescence techniques for CAP determination in complex matrices, green porous SnO 2 nanoparticles were synthesized and their multiple roles were availed for the automatic transformation of both non-fluorescent CAP and p-NP into their respective fluorescence products, and at same time as a sorbent in a µ-QuEChERS based sample preparation method used for CAP extraction and maximum elimination of organic and inorganic matrix antagonism in three complicated samples.
The sorbent "SnO 2 NPs" is unambiguously a better choice for this purpose as: (a) it need simple method of preparation; (b) require cost-effective starting materials; (c) it is nontoxic and not carcinogenic in nature; (d) it involve green synthesis and (e) its non-destructive nature to analyte of interest and high selectivity to reduce particular functional groups in the presence of other at that particular condition make it unique over other metal NPs 51 .
Therefore, porous SnO 2 NPs were synthesized by following modified method reported by A. Bhattacharjee et al. 52 and used in IC for sensitive and selective determination of CAP in three complex samples. It was observed SnO 2 NPs size, porosity and morphology have a significant effect on their application as a porous sorbent in the µ-QuEChERS based sample preparation for matrices interference (organic and inorganic) elimination and for the transformation of the non-fluorescent CAP into its fluorescent product. The nanoparticles definite morphology and size were harnessed by the addition of an optimized amount of stabilizing (sodium dodecyl sulfate) and a capping agent (glycine) 41,52,53 .
Here, very first time dual application of porous SnO 2 NPs as a sorbent for the matrices (polar interferences, dyes, heavy metals) selective degradation, adsorption and selective transformation of non-fluorescent antibiotic CAP into fluorescent species were exploited, which later cleanly separated and sensitively determined by using IC-FLD in three complex matrices including milk, human serum and urine.

Results and Discussion
Characterization of SnO 2 nanoparticles. The structure and phase composition of synthesized SnO 2 nanoparticles was characterized by XRD measurements. Figure  Infrared characterization of the porous SnO 2 nanoparticles was also performed in the range of 400 to 4000 cm −1 as depicted in Fig. 1(b). The peaks between 660-610 cm −1 are assigned to the Sn-O-Sn stretching vibration mode of surface bridging of SnO 2 NPs, while the peak at 530-560 cm −1 is attributed due to the terminal Sn-O stretching vibration mode of Sn-OH group. The existence of -OH group is also confirmed by the peak appeared around 3433 cm −1 . A strong peak around 618 cm −1 is an indication of potential conversion of stannous hydroxide into stannous oxide efficient at 600 °C, which is in agreement with the previous studies 52, 54,55 .
Size distribution and morphology study of SnO 2 NPs was carried out by SEM, TEM images and BET method as shown in Fig. 1a(c-h). Both SEM and TEM images show the good dispersion of synthesized SnO 2 nanoparticles with mean diameter of about 11.92 nm (RSD = 2.09%) which was calculated from fifty random NPs (see Fig. S1). The TEM images also show the spherical and polycrystalline nature of SnO 2 NPs which are composed of individual nanocrystallites. The spherical shape SnO 2 nanoparticles under the overlapped nanospheres can be visibly observed, which indicates the rough surface and porosity of NPs, which enhances their matrices adsorption and electron conducting capabilities 56,57 . The SnO 2 NPs showed an average pore size in mesopore range between 6.02 to 7.6 nm at 600 °C, whereas their pore volume about 0.177068 cm³/g was observed as shown in Fig. 1(g). The large pore size help to adsorb and reduce maximum CAP into its fluorescent specie and meanwhile these pore volume also help to adsorb matrix interferences to avoid matrix effect. This phenomenon can be observed from N 2 adsorption/desorption isotherms of SnO 2 NPs, indicating the presence of mesopores of adsorption and desorption as shown in Fig. 1(h).
Scanning of the fluorescent species (CAP) and its IC separation. The reduced CAP solution was collected and subsequently scanned with a fluorescence detector in alkaline media and λ em /λ ex = 232/361 nm was observed as its optimum wavelength for sensitive fluorescence detection as displayed in Fig. S2. A clean isocratic separation of CAP was performed by using a regular anion-exchange IonPac ® AS12 column protected with IonPac ® AG 12 guard column. The mobile phase 35 mM NaOH with only 15% acetonitrile at a normal flow rate of 1.0 mL/min was selected, which elute CAP at 7.1 min as given in Table 1. The isocratic elution of standard solution of CAP (1 mg/kg) with and without treatment of SnO 2 sorbent is presented in Fig. 2(a,b). It shows that that CAP can only be detected after SnO 2 sorbent treatment. The fluorescent property of reduced CAP can be attributed due to the conversion of electron withdrawing -NO 2 into electron donating -NH 2 group, which results in electron delocalization to an aromatic ring. Secondly, the alpha proton also ionized in basic media, which further helps to stabilize the resonating electrons and gave high fluorescence intensity.
The reduction mechanism of CAP into a fluorescent species involved the following two steps: (a) transfer of electrons to the CAP nitro functional group and (b) protonation.
Both CAP and p-NP reduction mechanism involved the transfer of electrons from the donor BH 4 −1 ions, while SnO 2 NPs presence in the solution help to adsorb BH 4 −1 ions and discharge of electrons from BH 4 −1 to the receptor species i.e. CAP or p-NP. In the second step, the aqueous protic solvent facilitates the required amount of H + ion for the maximum reduction of CAP/p-NP into their respective fluorescent species as the detailed mechanism is given in Fig. 3.

Optimization of factors affecting the fluorescence intensity of CAP.
There are various parameters which directly or indirectly influenced the fluorescence intensity of CAP. Effect of these parameters were investigated and optimized such as optimization of sorbent amount, optimization of NaBH 4 concentration, optimization of sonication time, effect of agitation temperature and evaluation of extracting solvent.
The flexibility of µ-sample preparation is very important for the analysis of particular analyte of interest in various matrices. Therefore, herein synthesized porous SnO 2 nanoparticles were used as sorbent and their dual application were exploited in µ-sample preparation for the transformation of the non-fluorescent CAP into the fluorescent product and to retain and degrade matrices/interferences at the same time. It also helped to transfer the analyte of interest in the organic phase.
During µ-sample preparation and optimization, it was observed that different amount of sorbent has a significant influence on interferences elimination and analyte recovery in different extracts. Therefore, different amount of sorbent was varied from (10, 20, 30, 40, 50 and 60 µg) and analytes recoveries were investigated. Herein, the fluorescent peak of analyte starts appearing at 30 µg sorbent and reached to maximum at 40 µg. Afterward, no significant increase was observed in fluorescence intensity of CAP in standard solution as shown in Fig. 4(a,b). Interestingly, during different samples analysis, it was observed the maximum recovery of analytes and low matrix peak appeared at 50 µg of the sorbent amount as it can be observed in sample chromatogram Fig. 4(a).
Hence, in order to utilize, dual application of sorbent 50 µg was chosen and used for further experiments to have maximum fluorescence quantum yield and recoveries of CAP in three samples.
Sodium borohydride is a reducing agent and its various advantages were exploited in this study. For example, (a) decomposition of various organic matrices pigments and dyes particularly, (b) reductive transformation of the non-fluorescent CAP into a fluorescent product and (c) enhancement of the extraction recovery by reducing   Fig. 4(b). It shows that 0.1 M NaBH 4 is enough to reduce maximum CAP into a fluorescent specie. HCl was also used in place of NaBH 4 for the reduction of the non-fluorescent CAP into a fluorescent specie. But the peak area of the fluorescent CAP was reduced significantly in the presence of HCl, this might be due to the fluorescence quenching effect of Cl −1 ions. Effect of both the reducing agents on CAP peak area can be observed from the Fig. 4(d).
In this µ-extraction ultra-sonic agitation is immensely vital for extraction of analytes of interest from various matrices, especially when its multi-purpose uses were exploited: for conversion of the non-fluorescent CAP into a fluorescent specie (chemical reduction process), extraction of CAP from various complex matrices and for acceleration of analyte mass transfer into the organic phase. The ultra-sonication time was varied and it was observed the CAP fluorescent peak appear after 30 minutes and continues to increase with the passage of time and became maximum at 75 minute, which reflects the maximum conversion of CAP into a fluorescent species, as shown in Fig. 4(c). This elongated time also enough to overcome the need for extra time which is normally required for analytes extraction from various complex matrices. Therefore, CAP maximum recoveries were obtained in the range 88.6-100.2% in various matrices as given in Table 2. This time also appearsed large enough for adsorption and degradation of different matrices interferences and impurities under sunlight 52 .
Contrary to the other extraction methods, the temperature is regarded as an important parameter in this particular µ-QuEChERS based sample preparation for the chemical reduction of CAP into a fluorescent specie. It was observed that more efficient conversion of the non-fluorescent CAP into a fluorescent CAP occurs above 40 °C and gave a high fluorescent peak at 60 °C. However, with further increase of temperature abnormal peak start appearing at R t 3.1 minute. It shows the possible chances of decomposition of CAP into more than one fluorescent products and this abnormal peak started to become high with the increase of temperature as shown in Fig. S3(a). Therefore, more suitable temperature 50-55 °C was chosen for further experiments.
It is also extremely vital to select an appropriate solvent in order to have a maximum recovery of desire analyte and to guarantee the minimal co-extraction of impurities and interferences. Various solvents including methanol, acetonitrile, acetone and dichloromethane were evaluated for CAP extraction from complex samples. Among these extraction solvents, acetonitrile exhibited desired results as its interaction toward organic analytes of interest to dissolve them quickly and its ability of back extraction of dissolved polar interfering impurities and matrices into aqueous phase via "salting-out" effect 54,58 . While methanol not only have dissolved the matrices and impurities but also was unable to form two layers. Therefore, huge positive matrix effect was observed. The other used solvents exhibited a poor recovery of the analyte. The effect of these four solvents on CAP recovery can be observed from Fig. S3(b).

Analysis of milk, human urine and serum samples. Applications of this developed method was inves-
tigated for the analysis of antibiotic CAP in three diverse complex matrices including human urine, milk, and serum. The clean ion chromatographic based isocratic separation and extremely sensitive fluorescence detection of CAP in complex samples is displayed in Fig. 5(a-c).

Validation of Method.
The IC-FLD chromatograms of blank and spiked milk, human serum and urine samples are shown in Fig. 5(a-c), which exhibit good selectivity as there is no interfering peak at the retention    The intra-and inter-day precisions are represented in term of relative standard deviations (RSDs). Both including intra-day recoveries and precisions were determined by anatomizing all three fortified samples in quintet per day at two fortification levels 0.48 and 5.0 mg/kg, while the inter-day recoveries and precisions were obtained by running the fortified samples for two successive days. At both fortified levels, RSDs of intra-and inter-day were in the range of 0.55-14.7% and 4.8-15.98%, respectively. The recoveries of CAP in all three samples were in the range of 78.2-100.2% as shown in Table 3, while matrix effect (n = 5) was observed between 92.20 and 101.35% in all three complex samples as given in Table 2.   Table 3. Contents and recovery studies of CAP in milk, human serum and urine samples at two fortification level 0.48 mg/kg and 5 mg/kg. (n = 5).
SCiEnTifiC RePORtS | (2018) 8:12327 | DOI:10.1038/s41598-018-29922-5 The significance of the Method. The developed method is very attractive, cheap, green and facile as: (a) It avoids the potential health risk imparted by pre/ post-column, electrochemical and other complex derivatization methods (avoids the need of such as, UV reactor, extra pump and reagents etc.); (b) green and cost effective as it minimized the use of reagents and chemicals; (c) utilizing the dual role of sorbent; (d) the method proved more selective and sensitive than previously developed method as shown in comparison Table 4; (e) involve µ-sample pretreatment; (f) inception of a new research area for determination of non-fluorescent compounds like CAP and p-nitophenol (p-NP). The fluorescent spectra of reduced non-fluorescent p-NP is given in Fig. 4(e).

Conclusions
This work introduces a novel dual role of synthesized porous SnO 2 nanoparticles in IC for the transformation of non-fluorescent antibiotic CAP and p-nitro phenol into their respective fluorescent products and as porous sorbent in the µ-QuEChERS based sample preparation method for riddance and degradation of interference and matrices in various complex samples (milk, human serum and urine samples).

Instrumentation and Methods.
Instrumentation. The IC-FLD separation and analysis were performed by using an Ion Chromatographic system (Thermo Fisher (ICS-1500 Waltham, MA, USA). An isocratic separation was conducted by using Ion Pac ® AS12A column (250 mm × 4 mm i.d; 13 mm particle size) protected by a guard column Ion Pac ® AG12A (50 mm × 4 mm i.d; 13 mm particle size) and sensitive detection was conducted by using a fluorescence detector (Ultimate 3000 RS, Dionex, Sunny vale, CA, USA). Furthermore, output signals were accumulated and subsequently processed with Chromeleons 7.2 software. The ultrasonic cleaner (SB-5200DT, Scientz Biotechnology Co. Ltd., Ningbo, China) was utilized for agitation, sonication and cleaning of apparatus and samples. The various characterization of nanoparticles were performed by using powder X-ray diffraction (XRD) method using Phillips X'Pert PRO diffractometer, scanning electron microscopy (SEM), transmission electron microscopy (TEM) images were obtained by using Hitachi SU-70 (Hitachi Japan), and H-7650 (Hitachi Japan), respectively. Fourier transform infrared (FTIR) spectroscopy was performed on SGE/Agilent 6890/Nicolet 5700 spectrometer (USA).
The selected real milk sample was purchased from the local market wall mart and stored in amber glass bottles. The real samples of human urine and serum were collected from the healthy volunteer from the local Zhejiang University medical hospital (yuquan branch). All the samples vials and bottles were wrapped in aluminum foil and kept at 4 o C in the laboratory refrigerator. The informed consent of all volunteers being obtained for analysis of CAP in their samples. Both types of experiments (i) synthesis of SnO 2 and (ii) analysis of CAP in biological samples were precisely performed in compliance of relevant local government and institution guidelines (Hangzhou Municipal Centre for Disease Control and Prevention Ethical Review Committee, China). Finally, all experimental protocols were approved by the analytical chemistry department of Zhejiang University.
Synthesis of porous SnO 2 NPs. The simple and completely green chemical precipitation method was used for the SnO 2 NPs synthesis by following A. Bhattacharjee et al. 52 scheme with slight modifications. Briefly, 0.02 M aqueous solution of SnCl 2 .2H 2 O and glycine were individually prepared in respective 100 ml flask and were subsequently mixed into a 400 or 500 ml clean beaker. The prepared 40 ml of 60 mmole SDS surfactant was added dropwise in the SnCl 2 .2H 2 O and glycine mixture with regular stirring at ambient temperature. Afterwards, the mixture was heated at 100 °C for 3-4 hours with continuous stirring until it turned yellow. Next day, the obtained precipitates were centrifuged, rinsed thrice with deionized water, dried at 60 °C and subsequently calcined at 600 °C for two hours. The obtained dry NPs powder was characterized and kept in a bottle for further use in the research work.

No. Samples
Extraction methods Analytical techniques LOD Ref.  µ-sample preparation method. A novel µ-QuEChERS based sample preparation method was utilized for the transformation of the non-fluorescent antibiotic CAP into a fluorescent specie and its extraction from milk, human urine and serum samples. The optimized µ-sample preparation extraction method involved conditioning of porous 60 µg SnO 2 sorbent in 4.0 ml centrifuge tube; meanwhile, 300 µl of each NaBH 4 and sample (milk, human urine and serum) were mixed and fortified with 0.48 mg/kg CAP in another similar 4.0 ml centrifuge vial. The resulting mixture was vigorously shaken for a moment before it was poured into sorbent tube. The mixture was sonicated for about 60 minutes at a temperature in the range of 55-60 °C to convert non-fluorescent CAP into its fluorescent form and for maximum decomposition and adsorption of matrices/ interferences (organic, inorganic and polar matrixes) on the porous sorbent. Afterward, the sonicated mixture was precisely centrifuged at speed of 9500 RPM for 5 minutes to halt the reaction. Then, a mixture of 140 µg MgSO 4 , 70 µg NaCl and later 900 µl HPLC grade ACN was added and put on sonication bath for another 10 minutes in order to obtain maximum reduced CAP back-extraction in the acetonitrile solvent and to disperse polar and inorganic impurities/ matrices in the aqueous layer. Finally, it was again centrifuged at the same speed for 12 minutes to settle down the used nanoparticles as a sorbent. In between two layers appeared, the upper organic layer containing CAP was precisely poured into another clean vial and it was dried with the help of gentle inert N 2 gas stream. Finally, dried residue was reconstituted to 900 µl with mobile phase and injected into the IC-FLD system with 1 mL syringe having 0.22 µm membrane filter for CAP sensitive fluorescence analysis. The same procedure was adopted for all selected samples, except real human serum which was first diluted 10 times with acetonitrile in order to homogenize the human serum and to precipitate the protein. Systematic SnO 2 nanoparticles synthesis and its dual role in the µ-sample preparation and IC-FLD analysis is illustrated in Fig. 6.

Method validation.
To investigate the liability of the proposed novel method, it was completely validated in term of selectivity, linearity, detection limit (LOD), quantification limit (LOQ), matrix effect, accuracy and intra-and inter-day precision. Selectivity of the method was investigated by observing any interfering peaks at the retention time of CAP. Five points calibration curve (0.48, 1.5, 3 and 5 mg/kg) was plotted for linearity and estimation of matrix effect. The calibration curve in the pure solvent was prepared by spiking the CAP respective concentration in 50% ACN: H 2 O, and then went through µ-sample treatment, while matrix-matched calibration curve was prepared by spiking CAP into 300 µl milk, human serum and urine sample and followed the same step as pure solvent spiked calibration. The matrix effect was evaluated by comparing CAP fluoresce response obtained in ACN solvent and respective matrices. ME was calculated by applying the following formula: where each A and B are the peak areas of CAP in matrix extract and pure solvent, respectively. Both LOD and LOQ were estimated based on the injected concentration at which signal to noise ratio (S/N) equal to 3 and 10-fold, respectively. Furthermore, intra-and inter-day precision were calculated within one and over three consecutive days and every injection was repeated in the quintet.