Sensitive and Selective Determination of Sudan I in Food by Molecularly Imprinted Polymer (MIP) Based Fluorescence Resonance Energy Transfer (FRET)

Abstract Sudan I is a cancer-causing chemical synthetic dye that is often illegally added to foods as a colorant. In this manuscript, based on molecularly imprinted polymers coated carbon dots (CDs@SiO2@MIPs), a novel fluorescence sensor is reported for the selective and sensitive determination of Sudan I. The sensor was prepared upon the surface of silica-coated carbon dots (CD@SiO2) by a one-pot sol-gel method using Sudan I as the template, 3-aminopropyl triethoxysilane (APTES) as the monomer, and tetraethyl orthosilicate (TEOS) as the crosslinker. Due to fluorescence resonance energy transfer, quenching occurred when CDs@SiO2@MIPs combined with Sudan I. The established sensor showed excellent recognition and performance for Sudan I with a wide linear range (1 to 120 μM) and low detection limit (0.61 μM). Furthermore, the sensor was successfully applied to determine Sudan I in chili powder and ketchup with recoveries from 89.7 to 101.7%. This work provides broad applications of fluorescence sensors based on CDs@SiO2@MIPs for the rapid analysis of food.


Introduction
Sudan I is a lipophilic azo compound and a synthetic dye.Because of its bright orangered color, Sudan I is widely used in petroleum, oil paint, wax, textile dyeing, shoe polish, and floor polish (Rebane et al. 2010;Zhou et al. 2013).Because this compound is carcinogenic, mutagenic, and allergenic, (Stiborova et al. 2002), Sudan I is prohibited from use in food products in China and the United Kingdom.Sudan I is listed as a Class III carcinogen (Zhang et al. 2022) by the International Agency for Research on Cancer (IARC) and may induce liver, bladder, and spleen tumors in humans.Therefore, for the sake of food safety and consumer health, it is necessary to develop an efficient, sensitive, and accurate method for Sudan I in food.
So far, several methods have been developed to determine Sudan I, including highperformance liquid chromatography (Sricharoen et al. 2017), mass spectrometry (Sciuto et al. 2017), liquid chromatographymass spectrometry (Mazzotti et al. 2008), Raman spectroscopy (Di Anibal et al. 2012), enzyme-linked immunoassay (Wang et al. 2012), flow-injection chemiluminescence (Liu et al. 2007), and electrochemical analysis (Ensafi et al. 2012).Although the above methods have high sensitivity and accuracy for Sudan I, most are cumbersome, time-consuming, and require expensive instruments and professional operators, which restricts their applications.
Fluorescence analysis is a focus of attention due to its simple operation, high sensitivity, fast response speed, and low cost (Ansari and Masoum 2021).Recently, carbon dots (CDs) have attracted attention (Xu et al. 2004).Compared with traditional organic fluorescent dyes and semiconductor quantum dots, CDs have the advantages of photobleaching resistance, photostability, non-toxicity, good biocompatibility, and excellent water solubility (Liu et al. 2016;Liu, Ding, et al. 2019).The fluorescence quantum yield and optical sensing performance of CDs is markedly increased by doping or surface passivation (Walekar et al. 2019).
Lu et al. used L-serine and L-tryptophan as carbon sources to prepare membrane structure nitrogen-doped carbon dots (N-CDs) with high a quantum yield (61.12%) for H þ and Fe 3þ determination (Lu et al. 2014).Hu et al. prepared nitrogen-doped carbon dots (N-CDs) by a hydrothermal method using used tires as a carbon source that provided fluorescence detection of Sudan dyes in chili powder based upon the inner filter effect (Hu and Gao 2020).Although CDs have been widely used as fluorescent sensors, they are affected by environmental conditions, sample concentration, and instrumentation.Therefore, the modification of CDs to improve selectivity, such as molecular imprinting technology (Liu, Ding, et al. 2020), is a topic of interest.
Molecular imprinting is used to prepare three-dimensional crosslinked polymers with specific recognition for template molecules.The resulting polymers are called molecularly imprinted polymers (MIPs) (Ahmad et al. 2019).Because the recognition site matches the size, shape, and functional group of the target, the prepared MIPs exhibit excellent selectivity in complex samples.Due to the strong mass transfer, fast kinetics, and high selectivity, surface molecularly imprinted polymers have received considerable attention (Cui et al. 2021).
Many MIPs-based CDs have been reported which combine the high selectivity of MIPs and sensitivity of fluorescence detection (Xie et al. 2023).Liu et al. prepared MIPs-coated CDs via a sol-gel method, which achieved selective recognition and sensitive fluorescent determination of oxytetracycline (Liu, Ding, et al. 2019).Yang et al. reported the preparation of molecular imprinting polymers coated carbon dots (MIPs@CDs) by reverse microemulsion for the selective and sensitive determination of trace tetracycline in fish meat (Yang et al. 2018).Unfortunately, no applications of these highly selective and sensitive materials have been reported for the determination of Sudan I.
In this study, nitrogen and sulfur-doped CDs (NS-CDs) were functionalized with silanization by the Stob€ er method to prepare CDs@SiO 2 to maintain the fluorescent properties.Next, the CDs@SiO 2 @MIPs were obtained by coating an imprinted layer on the surface of CDs@SiO 2 by the sol-gel method.The CDs@SiO 2 @MIPs enabled selective recognition and fluorescence determination of Sudan I by quenching induced by fluorescence resonance energy transfer (FRET).Hence, a fluorescence sensor based on CDs@SiO 2 @MIPs was established that was employed to determine Sudan I in chili powder and ketchup.The developed sensor provides a new approach for the rapid determination of pollutants in food.
The chili powder and ketchup samples were purchased from markets in Changchun (China).

Preparation of CDs@SiO 2
Based on our previous work, NS À CDs with blue fluorescence were prepared by a onestep hydrothermal method.The synthesized NS-CDs have good dispersion with a particle size from 2 to 3 nm and an absolute quantum yield of 73% (Chen et al. 2021).The CDs@SiO 2 was prepared according to the literature (Fang et al. 2019) with modifications.5.0 mg of NS-CDs were dispersed in 20 mL ethanol and stirred for 15 min.40 lL of APTES were added and stirred for an additional 10 min followed by an additional 40 lL of TEOS with stirring for 30 min.40 lL ammonia were added and stirred for 6 h.The obtained solid was washed three times with ultra-pure water and ethanol, dried at 60 C to obtain CDs@SiO 2 , and evenly ground.

Preparation of CDs@SiO 2 @MIPs
The CDs@SiO 2 @MIPs were prepared by improved sol-gel surface imprinting (Cai et al. 2019).30 mg of Sudan I were ultrasonically dissolved in 40 mL of 95% ethanol.80 lL of APTES were introduced and magnetically stirred for 30 min.10 mg of CDs@SiO 2 were added with stirring for 10 min, followed by 250 lL of TEOS and 100 lL of ammonia with stirring for 24 h at room temperature.After the sol-gel process, the crude product was centrifuged, washed with ethanol several times, dried, and Soxhlet extracted with a 9:1 methanol:acetic acid to remove the Sudan I template in the supernatant using a pipette.The presence of Sudan I in the supernatant was determined by a spectrophotometer.The Soxhlet extraction was discontinued once the Sudan I peak was no longer present.
The non-imprinted polymers coated carbon dots (CDs@SiO 2 @NIPs) were prepared by the same procedure but without the Sudan I template.

Fluorescence measurements
Equal volumes of Sudan I standard or sample and 600 mg/L of CDs@SiO 2 @MIPs were sonicated for 3 min.The fluorescence spectrum was recorded using an excitation voltage of 700 V and a scanning speed of 3000 nm min À1 .The excitation and emission slits were 5 nm with a response time of 0.4 s.An excitation wavelength of 360 nm was set to record the emission spectrum from 380 nm to 650 nm.All measurements were in performed in triplicate.

Determination of Sudan I in chili powder and ketchup
The chili powder and ketchup were processed as reported in the literature (Hu and Gao 2020).60 mg of sample were dispersed in 3 mL ethanol, stirred for 10 min, sonicated for 30 min, allowed to stand for 20 min, and centrifuged.The supernatant was passed through a 0.22 lm filter membrane diluted to 10 mL.CDs@SiO 2 @MIPs (1 mL, 600 mg/L) were added into the above extract (1 mL) and sonicated for 3 min to obtain a homogeneous solution.The Sudan I content was determined by monitoring the fluorescence intensity.All measurements were performed in triplicate.

Results and discussion
Synthesis of CDs@SiO 2 @MIPs Figure 1 shows a schematic of the preparation of the CDs@SiO 2 @MIPs.The NS-CDs were prepared by a one-step hydrothermal method, using L-cysteine and L-tryptophan as precursors.The CDs@SiO 2 were prepared by the Stob€ er method.The transparent silica ensures the fluorescent properties of NS-CDs, improves the biocompatibility, dispersibility, and surface functionalization of NS-CDs, and promotes further polymerization (Jalili and Amjadi 2015).
The CDs@SiO 2 @MIPs were prepared by sol-gel surface imprinting with APTES (functional monomer), TEOS (crosslinker), and ammonia (catalyst).During the polymerization, the amino groups on the surface of APTES interact with the hydroxyl groups on the surface of Sudan I by hydrogen bonding.After eluting Sudan I with methanol-acetic acid, the structure of CDs@SiO 2 @MIPs includes imprinting sites with the Sudan I structure.APTES has a non-covalent interaction with the template Sudan I.If the quantities of functional monomers are low, they do not fully interact with the template.Otherwise, nonspecific adsorption occurs due to the self-polymerization of monomers (Qi et al. 2021;Zhang et al. 2012).As shown in Figure S1, 80 lL of APTES were deemed to be optimum for the synthesis of CDs@SiO 2 @MIPs.

Characterization of CDs@SiO 2 @MIPs
The morphologies and sizes of CDs@SiO 2 @MIPs and CDs@SiO 2 @NIPs were characterized by TEM and SEM.As shown in Figure 2a and b, the particles of CDs@SiO 2 were uniform with an average particle size of 7.1 nm.After the sol-gel imprinting, the resulting CDs@SiO 2 @MIPs particles were enlarged with a particle size of 37.8 nm.In addition, the TEM image of CDs@SiO 2 @MIPs shows that the CDs were coated in the CDs@SiO 2 @MIPs particles (the red dotted circle in Figure 2b).
The SEM images of CDs@SiO 2 @MIPs and CDs@SiO 2 @NIPs in Figure 2c and d show that both are aggregated with sizes from 30 to 50 nm due to the same synthesis.Compared with the previously prepared NS-CDs (Chen et al. 2021), the sizes of CDs@SiO 2 and CDs@SiO 2 @MIPs increased, indicating the successful synthesis of silica imprinted layer upon the surface of NS-CDs.
The XRD pattern of CDs@SiO 2 @MIPs in Figure 3a shows that CDs@SiO 2 @MIPs had a broad peak at 2h ¼ 23 corresponding to the C (002) crystal plane.The broad diffraction peak in the XRD spectrum indicates the formation of amorphous silica (Guo et al. 2018).Figure 3b shows the FT-IR spectra of NS-CDs, CDs@SiO 2 , and CDs@SiO 2 @MIPs.In the NS-CDs, the wide band at 3413 cm À1 is due to the stretching vibrations of O-H and N-H, indicating the presence of hydrophilic groups.The peaks at 1615 and 1375 cm À1 are due to C ¼ N and C-N stretching, respectively.The peaks at 1106 and 646 cm À1 are caused by the stretching vibrations of C ¼ S and C-S, respectively.The peaks at 1520, 1240 and 750 cm À1 indicate the presence of C ¼ C, ¼C-Oand C-H (Zeng et al. 2015), showing the preparation of NS À CDs.In the CDs@SiO 2 and CDs@SiO 2 @MIPs FT-IR spectra, the peak at 1065 cm À1 corresponds to the stretching of Si-O-Si.The peak at 795 cm À1 is due to the absorption by Si-OH (Fang et al. 2019).Hence, the FT-IR results demonstrate the incorporation of silica onto the surface of NS-CDs and the successful preparation of the CDs@SiO 2 @MIPs.
The surface functional groups and elemental compositions of CDs@SiO 2 @MIPs were characterized by XPS. Figure 3c includes four peaks in the XPS full spectrum scan of Figure 2. Field-emission transmission electron microscope images of (a) CDs@SiO 2 and (b) CDs@SiO 2 @MIPs.Field-emission scanning electron microscope images of (c) CDs@SiO 2 @MIPs and (d) CDs@SiO 2 @NIPs.
CDs@SiO 2 @MIPs at 293. 65, 407.65, 539.65, and 111.65 eV, corresponding to C 1s, N 1s, O 1s, and Si 2p, respectively.Figure S2a shows the peaks at 284.6, 285.1, and 285.9 eV correspond to C-C/C ¼ C, C-O/C-N, and C ¼ O groups, respectively (Zhao et al. 2019).Figure S2b shows N 1s includes peaks at 399.5 and 400.5 eV that correspond to C-N and C ¼ N (Zhao et al. 2019).The peak fitting of O 1s (Figure S2c) shows peaks at 531.4, 532.2 and 532.9 eV, indicating the presence of C ¼ O, C-OH, and Si-O.The fitting of Si 2p in Figure S2d includes peaks at 103.5 and 104.4 eV, corresponding to Si-O/SiO 2 and Si-C, respectively (Liu, Ding, et al. 2020).The above XPS characterization was consistent with the FT-IR results, further illustrating the successful incorporation of silicon and preparation of the CDs@SiO 2 @MIPs.
Optical performance of the CDs@SiO 2 @MIPs Figure 3d shows the absorption and fluorescence spectra of CDs@SiO 2 @MIPs with the absorption spectrum of Sudan I.The absorption spectrum shows that CDs@SiO 2 @MIPs Figure 3. (a) X-ray diffraction pattern of the CDs@SiO 2 @MIPs.(b) Fourier transform infrared spectra of the NS-CDs, CDs@SiO 2 , and CDs@SiO 2 @MIPs.(c) Full x-ray photoelectron spectrum of the CDs@SiO 2 @MIPs.(d) Absorption spectra of the (1) CDs@SiO 2 @MIPs and (2) Sudan I with the (3) fluorescence excitation and (4) emission spectra of the CDs@SiO 2 @MIPs.
had a peak at 265 nm due to the p-p Ã transition of C ¼ C (Pei et al. 2015).The fluorescence spectra show that the maximum excitation and emission wavelengths of CDs@SiO 2 @MIPs were 346 and 416 nm, respectively, and the fluorescence excitation and emission spectra of CDs@SiO 2 @MIPs overlapped with the absorption spectrum of Sudan I (Hu and Gao 2020), suggesting that the quenching mechanism may be FRET (Liao et al. 2021).
The change in the fluorescence intensity of CDs@SiO 2 @MIPs was investigated under natural light for 7 d to characterize their stability.Figure S3a shows the fluorescence intensity remained nearly constant.The CDs@SiO 2 @MIPs were also irradiated with ultraviolet light for 60 min (Figure S3b) with no change.The above results show that CDs@SiO 2 @MIPs had suitable stability for fluorescence sensing.
Reusability is a significant figure of merit for sensors.Figure 4 shows the sensor may be reused at least five times.
Fluorescence determination of Sudan I by CDs@SiO 2 @MIPs The pH, the concentration of CDs@SiO 2 @MIPs, and the equilibration time were optimized as shown in Figure S4.The optimal pH was 7, the optimal concentration of CDs@SiO 2 @MIPs was 300 mg/L, and the optimum equilibration time was 3 min.Various concentrations of Sudan I were added to the CDs@SiO 2 @MIPs and CDs@SiO 2 @NIPs systems using optimal conditions as shown in Figure S5.The fluorescence intensities of CDs@SiO 2 @MIPs and CDs@SiO 2 @NIPs decreased with the Sudan I concentration.The quenching was described by the Stern-Volmer equation: F0=F ¼K sv C q þ1 where F 0 and F are the fluorescence intensities in the absence and presence of Sudan I, K sv is the Stern-Volmer constant, and C q is the concentration of Sudan I.When the concentration of Sudan I was between 1 lM and 120 lM, the parameter F 0 /F-1 was linearly related to the concentration of Sudan I by F 0 /F-Figure 4. Reusability of the CDs@SiO 2 @MIPs for the determination of Sudan I.
The limit of detection (LOD) was determined to be 0.61 lM using the relationship LOD ¼ 3r/k where r is the standard deviation of 11 blank measurements and k is the slope of the calibration relationship.The imprinting factor (IF) ¼ K sv, MIPs /K sv, NIPs was determined to be 1.63, indicating that the imprinting layer improves the sensitivity for Sudan I.The developed protocol has a wide linear range and a low detection limit with a simple, environmentally friendly, and low-cost approach.

Selectivity
The fluorescent quenching of CDs@SiO 2 @MIPs toward several azo compounds (Figure 5c) was evaluated to evaluate the selectivity.Figure 5d shows CDs@SiO 2 @MIPs underwent stronger fluorescence quenching with Sudan I compared to the analogues.The results confirm that CDs@SiO 2 @MIPs formed specific recognition sites for Sudan I.
Sugars, food additives, vitamins, salts, metal ions, and amino acids that may be present in food were investigated to characterize interferences with Sudan I determination as shown in Figure S6.Cu 2þ was able to quench the fluorescence intensity of CDs@SiO 2 @MIPs.The quenching of CDs@SiO 2 @MIPs due to Cu 2þ may be due to an aggregation-caused mechanism (Shamsipur et al. 2023) or electron transfer of cupric amine complexes (Zhang et al. 2023).However, the Cu 2þ content in the samples is relatively low and its interference was eliminated by EDTA masking.Other substances had little influence upon the fluorescence intensity of CDs@SiO 2 @MIPs.The results show that CDs@SiO 2 @MIPs had suitable selectivity for the determination of Sudan I in food.

Quenching mechanism
The mechanism of CDs@SiO 2 @MIPs fluorescence quenching with Sudan I was investigated.As shown in Figure 3d, the excitation and emission spectra of CDs@SiO 2 @MIPs overlapped significantly with the absorption spectrum of Sudan I, suggesting that the quenching mechanism may involve non-radiative FRET (Feng et al. 2021) in which the donor transfers energy to the acceptor through dipole-dipole interactions between molecules (Feng et al. 2021).Fluorescence resonance energy transfer occurs when there is overlap between the donor's emission spectrum (CDs@SiO 2 @MIPs) and the acceptor's absorption spectrum; a reduction in the fluorescence lifetime of the donor occurs following the addition of the acceptor; and the distance between the donor and acceptor is from 1 to 10 nm (Wang et al. 2020).
Figure 6a shows the fluorescence decay of the CDs@SiO 2 @MIPs in the absence and presence of Sudan I.The calculation of fluorescence lifetime (s) is described in the Supplementary Material.The fluorescence lifetime results are summarized in Table S1.The fluorescence lifetime of CDs@SiO 2 @MIPs decreased from 6.71 ns to 5.16 ns with the Sudan I concentration (Shi et al. 2021;Wang et al. 2020).
The TEM image in Figure 6b shows that many CDs@ SiO 2 particles were embedded upon the surface of CDs@SiO 2 @MIPs.The molecular imprinting on the surface of CDs@SiO 2 @MIPs allows FRET quenching to occur due to the close contact between Sudan I and CDs.

Analysis of chili powder and ketchup
The developed method was employed to determine Sudan I in chili powder and ketchup.No detectable level of Sudan I was present in the samples, so spike recovery experiments were carried out.Table 1 shows the recoveries of Sudan I ranged from 89.7 to 101.7% and the relative standard deviation (RSD) was less than 3.3%.
Table 2 compares the analytical figures of this work with literature fluorescence methods for Sudan I.The reported sensor presented provides a wide linear range, comparable limit of detection and recovery, and excellent stability.The results demonstrate that the fluorescent sensor based upon CDs@SiO 2 @MIPs was suitable for the determination of Sudan I in chili powder and ketchup.TEM image of the CDs@SiO 2 @MIPs.The black dots inside the red circles in the inset show CDs@SiO 2 on the surface of the CDs@SiO 2 @MIPs.

Conclusions
NS-CDs were synthesized by a one-pot hydrothermal method using L-cysteine and L-tryptophan as precursors and employed as the fluorescent unit for silicon modification to obtain CDs@SiO 2 .Then the CDs@SiO 2 @MIPs were prepared by a sol-gel method to prepare a molecularly imprinted layer on the CDs@SiO 2 surface for Sudan I determination.Following the addition of Sudan I, the fluorescence intensity of CDs@SiO 2 @MIPs was quenched due to FRET.The decrease in fluorescence intensity was proportional to the concentration of Sudan I. Therefore, a fluorescence sensor was established and employed to determine Sudan I in chili powder and ketchup.This report is the first application of fluorescence molecular imprinting for the selective determination of Sudan I.The established sensor has the advantages of good precision, fast response, low cost, wide linear range, and a low detection limit, and offers broad application prospects for practical analysis.

Figure 5 .
Figure 5. Calibration curves of (a) CDs@SiO 2 @MIPs and (b) CDs@SiO 2 @NIPs for the determination of Sudan I. (c) Structures of Sudan I, Sudan II, Sudan B, Sudan G, azophloxine and Orange II; (d)Fluorescence response (F 0 /F-1) of CDs@SiO 2 @MIPs and CDs@SiO 2 @NIPs for Sudan I, Sudan II, Sudan B, Sudan G, azophloxine and Orange II.

Figure 6 .
Figure 6.(a) Fluorescence decay of the CDs@SiO 2 @MIPs in the absence and presence of Sudan I. (b)TEM image of the CDs@SiO 2 @MIPs.The black dots inside the red circles in the inset show CDs@SiO 2 on the surface of the CDs@SiO 2 @MIPs.

Table 1 .
Accuracy of the fluorescence determination of Sudan I in chili powder and ketchup (n ¼ 3).

Table 2 .
Comparison of the reported CDs@SiO 2 @MIPs fluorescent method with the literature for the determination of Sudan I.