3D plasmonic SERS aptasensor for rapid detection of a�atoxin B1 combined with Au@Ag bimetallic nanostars and Fe3O4@MoS2 magnetic nano�owers

As a virulent metabolite, a�atoxin B 1 (AFB 1 ) presented in various cereal grain is tightly implicated in severe human diseases. In this study, 3D plasmonic nanohybirds of Raman molecule 4-mercaptobenzoic acid (4-MBA)-embedded and AFB 1 aptamer-modied bimetallic nanostars as probes bound to magnetic nano�owers were fabricated and demonstrated as a high-performance SERS-active aptasensor to quantitatively analyze AFB 1 . Bimetallic Au@Ag SERS plasmonic nanoprobes with enhanced properties were capable of enhancing discriminative Raman peaks of 4-MBA. Then, the integration of iron tetroxide nanoparticles (Fe 3 O 4 NPs) and molybdenum disul�de nanosheets (MoS 2 NSs) with huge speci�c surface area constituted stable 3D Fe 3 O 4 @MoS 2 plasmonic nano�owers, facilitating the bind of numerous aptamer-based SERS probes via the non-covalent interaction between MoS 2 NSs and aptamer, which were ideal candidates for SERS-active substrates. Additionally, Fe 3 O 4 NPs as magenetic core endowed 3D nanocomposites with speci�c magnetic separation characteristic that caused the collected SERS hotspots to exhibit superior signal response, and further strengthening the sensitivity in a complex food matrix. Aptamer-target AFB 1 speci�c recognition triggered linearly diminished 4-MBA signal intensity (I 4-MBA ) on the substrate to achieve a low detection limit of 58.9 pg/mL. Furthermore, the sensor has the potential to be a promising monitoring tool for trace contaminants.


Introduction
A atoxin B 1 (AFB 1 ) derived from fungi species is difuran-coumarin compound with the most poisonous activities (Marchese et al. 2018).Over the last few decades, AFB 1 contamination in a wide variety of common foodstuffs has become an unavoidable global concern, containing grains (corn, wheat and peanuts) and snack foods (jerky, pistachios and almonds) (Mishra et al. 2022;Pleadin et al. 2015;Wu et al. 2020).In particular, for peanuts, the AFB 1 detection rate reached 33.8% from four major production regions in China (Yang et al. 2020).In Mexico, AFB 1 residues were found in 80% of investigated samples, 26% of the samples exceeded the EU maximum AFB 1 limit value (5 μg/kg) (Zuki-Orozco et al. 2018).
AFB 1 contents measurement more than 182.28 μg/kg in Burkina Faso accounted for 41.50% of peanut samples (Bandé et al, 2022), suggesting an elevated contamination rate.Meanwhile, conventional heating treatments are hard to su ciently degrade AFB 1 toxins with stable biochemical properties, even pasteurization.Prolonged exposure to low doses of the insidious AFB 1 poses a critical and potential threat to humans, and has been closely linked via epidemiological studies to chronic diseases consisting of malnourishment, growth disorders, and immunosuppression (Rushing and Selim 2019).
Analytical approaches depending on large-scale instruments have been explored to accurately measure AFB 1 , such as thin layer chromatography (TLC) (Var et al. 2007), high-performance liquid chromatography (HPLC) (Mochamad and Hermanto 2017) and liquid chromatography-tandem mass spectrometry (LC-MS) (Janik et al. 2021).However, certain downsides of poor response, low e ciency, and excellent professionalism limit the application of methods in rapid food safety testing.Thus, designing simple, quick, and ultrasensitive detection methods based on signal ampli cation strategies of nanomaterials is conducive to real-time monitoring of trace amounts of toxicants.
Surface-enhanced Raman scattering (SERS) is an advanced spectroscopy analysis technology, which has received great attention in food safety control, environmental monitoring, and disease diagnosis (Deng et al. 2022;Lu et al. 2020;Song et al. 2017).With the additional excellent properties of easy manipulation, resistance to discoloration and uorescence, and fast response, sensitive detection of various mycotoxins and biomarkers is achieved (Pettine et al. 2020;Turan et al. 2022;Zhao et al. 2020).
Since the SERS signal ampli cation on precious metal surfaces was elucidated by Jeanmaire in 1977 (Jeanmaire and Van Duyne 1977), the enhancement mechanism has developed into dominant electromagnetic enhancement (EM) excited by local surface plasmon resonance (LSPR) at present (Ding et al. 2017).Thus, SERS-based biosensing strategies mainly consist of target-guided direct detection, and probe-mediated indirect detection by virtue of the ability to generate molecular ngerprint information via EM.Although direct AFB 1 detection possesses a simple preparation process, it requires a combination of complex stoichiometry to distinguish the characteristic peaks to achieve quanti cation, showing low sensitivity (Shao et al. 2021).These issues are addressed via a designed AFB 1 labeled sensor based on Raman signal molecules with low background noise and signi cant characteristic peaks, realizing e cient indirect AFB 1 determination.In this sensor, the combination of Raman molecules, metal NPs (various morphology containing triangular, cubic, and star) (Wang and Guo 2020;Yang et al. 2020), and tailored aptamer (single-stranded oligonucleotides binding the target with strong speci city) (Pan et al. 2022) constitutes highly sensitive probes, such as Au@4-MBA@Ag NRs (Lin et al. 2020) and Au@4-MBA@Si NPs The components of the Au-Ag alloy and the tip effect of the anisotropic NPs enable signi cant enhancement of the SERS signal, improving the sensing sensitivity of the SERS-based approach.
In recently reported magnetic substrates, Fe 3 O 4 NPs are widely applied in various biosensors, due to the advantages of low toxicity, low cost, excellent biocompatibility, and especially superparamagnetic properties (Guo et al. 2021;Tajik et al. 2021).But bare Fe 3 O 4 NPs exposed to air are susceptible to oxidation and agglomeration.The introduction of speci c carriers is considered by researchers as an effective measure.Notably, MoS 2 NSs, layered two-dimensional (2D) materials, possess distinct bene ts of large speci c surface area, and multiple active sites, which are favorable for binding with SERS probes (Karaman et al. 2021;Rani et al. 2020).Thus, the complexes of MoS 2 NSs wrapped around Fe 3 O 4 NPs as magnetic substrates in combination of SERS aptasensor have the capability to quickly collect and concentrate SERS signals from complex systems, which greatly simpli es the test procedures, and improves sensitivity for trace analyses.
In this work, a single-response SERS aptasensor was rationally prepared to detect AFB 1 for the rst time with Fe 3 O 4 @MoS 2 NFs/Au@Ag NSs-AFB 1 apt nanohybirds.Firstly, Au-4MBA@Ag NSs-AFB 1 apt act not only as SERS enhancer of 4-MBA signal by su cient SERS hot spots from sharp edges and tips but also as speci c capture probes of AFB 1 in the presence of other interfering substances.Meanwhile, Fe 3 O 4 @MoS 2 NFs serve as the enrichment and ampli cation substrates of SERS signal when AFB 1 aptfunctionalized SERS probes were immobilized on MoS 2 NSs covered on Fe 3 O 4 NPs via π-π interaction, exhibiting the strongest SERS signals.With the assistance of external magnets, 3D magnetic compound has the ability to purify the detection signal in a short time.The target AFB 1 was e ciently recognized by AFB 1 apt-modi ed SERS probes.Owing to the dissociation of the probes, the effective abatement of 4-MBA signal intensity in the substrate was showed.Therefore, the SERS sensor achieved ultrasensitive and good selective measure toward AFB 1 .

Instruments
Transmission electron microscope (TEM) images of all nanomaterials were acquired by the JEM-21OOF microscope, operating at 200 kV (Tokyo, Japan).The UV-Vis absorption spectra were recorded by the UV-1800 spectroscopy (China).The Raman spectra of 4-MBA in the assemblies were measured by DXR2xi microscope (U.S.A.) equipped with 50× microscope lens and 632.8 nm laser excitation.And the origin software was used to smooth and correct the acquired Raman spectra for better data analysis.X-Ray Diffractomer (XRD) patterns of Fe 3 O 4 @MoS 2 NFs were obtained by D2 PHASER analyzer using Cu-Kα radiation (Switzerland).Zeta potential was determined using Zetasizer Nano ZS analyzer (Melvin, UK).
To prepare functionalized nanoprobes, 250 μL of Au-4MBA@Ag NSs was resuspended in 0.05% Tween-20 solution.5 μL of AFB 1 apt (50 μM, activated with TCEP solution in equal volume for 1 h) was added, and incubated for 1 h.NaCl solution was dropped every 30 min to complete aging (with a nal concentration of 0.25 M).Afterward, the mixture maintained overnight at 37°C.The free nucleic acid was discarded under centrifugation to produce Au-4MBA@Ag NSs-AFB 1 apt.
Preparation of Fe 3 O 4 @MoS 2 NFs Take 1.35 g of FeCl 3 •6H 2 O powder into 18 mL of ethylene glycol solution.Then, 1.62 g of CH3COONa and 0.45 g of polyethylene glycol were successively added into the above mixture.The solution was thoroughly mixed well under sonicating for 30 min, and reacted at 200°C for 8 h in a high temperature reactor.The reaction products were washed three times with ethanol and ultrapure water by magnetic separation, and dried under vacuum at 60°C for 6 h to collect Fe 3 O 4 NPs.20 mg of the obtained Fe 3 O 4 NPs powder was dispersed well in 10 mL of ultrapure water by ultrasonication.Then, 0.112 g of ammonium heptamolybdate and 0.365 mg of thiourea were sequentially added by vigorous sonicating for 30 min.Under 200°C, the mixture was placed in a high temperature reactor for 8 h.The resulting products were washed repeatedly with ethanol and ultrapure water.Finally, Fe 3 O 4 @MoS 2 NFs powder was obtained under vacuum drying at 60°C for 10 h.
Next, AFB 1 standard solutions with different amounts ( nal concentrations of 0.1 ng/mL, 1 ng/mL, 10 ng/mL, 100 ng/mL, and 1000 ng/mL, respectively) were reacted with 200 μL of the above solution at room temperature for 1 h.The complex was magnetically separated and washed twice to ensure signal drop.After that, 8 μL of Au-4MBA@Ag NSs-based 3D nanohybirds solutions were dropped onto aluminum foil and dried at 25℃.The SERS measure was performed at an excitation wavelength of 633 nm.The logarithmic value of AFB 1 concentration was used as the horizontal coordinate, and the Raman intensity of 4-MBA at 1581 cm -1 (I 1581 ) was employed as the vertical coordinate to determine the standard curve of the method.

Selectivity evaluation
To assess the selectivity of the SERS aptasensor, the control experiments were performed with interfering toxins including AFM 1 , OTA and FB 1 and AFB 1 under the same experimental conditions.The concentrations of the above toxins were set as 100 ng/mL.The comparison of SERS intensity directly re ected the speci city of aptamer-based sensor.

Real samples detection
Fresh peanuts were selected to analyze the utility of SERS aptamer sensors to detect AFB 1 in real samples.For pretreatment, the peanut samples were fully ground to powder rstly. 2 g of powder was dissolved in a mixture of methanol/water (5.6 mL/2.4 mL) under sonicating for 30 min to aid the extraction performance (Jing et al. 2009).The supernatant was collected by centrifuging (5000 rpm, 15 min).Lastly, AFB 1 standards were added at different nal concentrations of 0.5 ng/mL, 5.0 ng/mL, and 50.0 ng/mL, respectively.After the above procedure analysis, the recovery results were calculated.

Detection strategy of the aptasensor for AFB 1
The working principle of designed 3D plasmonic SERS aptasensor for AFB 1 detection was illustrated in scheme 1. Bimetallic Au@Ag plasmonic nanostars were acquired by modifying 4-MBA on anisotropic Au NSs via Au-SH bond, further reducing AgNO 3 to form Ag outer shell.The enhanced EM excited by a uent hot spots of sharp tips signi cantly ampli ed the SERS response of 4-MBA embedded on bilayer Au-Ag.The speci c core-shell nanostructure gave effectively protective effect on signal, improving the detection stability.Then, the coupled AFB 1 apt on Au@Ag stars employed as SERS probes could be capable of capturing target AFB 1 with high sensitivity and high speci city.Meanwhile, Fe 3 O 4 @MoS 2 plasmonic nano owers provided large number of active binding sites for SERS probes, generating the 3D magnetic SERS plasmonic substrates-aptamer-SERS plasmonic probes (Fe 3 O 4 @MoS 2 -AFB 1 apt-Au-4MBA@Ag NSs), which owned both SERS activities and magnetic features.Combined with the chemical enhancement effect of MoS 2 NSs on the multilayer plasma nanostructures, SERS signal was further magni ed to the maximum.When AFB 1 appeared, the probes were separated from the nano owers owing to the speci cal capture of aptamers, causing an obvious decrease of I 4-MBA .Thus, I 4-MBA was inversely related to the AFB 1 concentration, enabling sensitive detection in the peanuts.

Characterization and optimization of Au@Ag NSs-AFB 1 apt
The TEM images for the morphology of the nanoparticles were shown in Fig. 1.Au NSs had great dispersion and showed sharply star-shaped morphology, which consisted of spherical core sized 34.53 nm ± 5.29 nm and rich sharp tips with a size of around 31.13 nm ± 6.71 nm (Fig. 1A).As shown in Fig. 1B, after the reduction of Ag ions was induced under ammonia-adjusted alkaline conditions, the Ag atoms were uniformly deposited to form Ag shells on the surface of monolayer Au NSs, which was due to the extremely similar lattice edge widths of both Au and Ag (Rodriguez-Gonzalez et al. 2005).The results showed that moderate amount of silver nitrate had little effect on the tip sharpness of the nanostars.In the UV-Vis spectra of Fig. 1C, as the formation of Au NSs prepared by Au seed growth method, the resonance plasmonic absorption peak of Au NPs located at 521 nm redshifted to around 691 nm.
Thereby the clear characteristic absorption peak of Ag NPs appeared at 407 nm, indicating the successful modi cation of Ag shell outside Au NSs.The results showed the bimetallic Au@Ag plasmonic nanostars exhibited two absorption peaks at 407 nm and 549 nm in optical properties.Furthermore, the SERS properties of bimetallic nanomaterials were characterized by Raman spectrometer in Fig. 1D.There were no obvious Raman peaks on Au NSs.In contrast, Au-4MBA NSs occurred effectively enhanced the SERS signals of 4-MBA, which was owing to the multiple SERS "hot spots" at the tips.It was worth noting that Au@Ag NSs plasmonic structure-ampli ed 4-MBA signal was 4 times stronger than that of monolayer precious metal Au, which was in agreement with previous study (Jing et al. 2020).The above results clearly veri ed the successful preparation of Au-4MBA@Ag NSs for the further SERS detection of AFB1.
In addition, the effect of the addition amount of 4-MBA and AgNO 3 on the plasmonic SERS probes was investigated.As shown in Fig. S1, the SERS intensity of 4-MBA was the strongest with the addition of 4-MBA (1 mM) up to 0.8 μL, suggesting the 4-MBA adsorption on the surface of Au NSs reached saturation.
Based on the optimized addition value of Raman molecular, the optical features of materials modi ed with different amounts of AgNO 3 (1 M) were characterized by UV-Vis spectra shown in Fig. S2C.Along with the increase of the AgNO 3 volume to 4 μL, the resonance plasmonic absorption peak of Au@Ag NSs underwent a gradual blue shift, which was closely related to the aspect ratio of the star-shaped tip.Meanwhile, a new absorption peak at around 407 nm appeared and enhanced, proving the successful deposition of Ag shells and a gradual increase in thickness.All these results were tightly dependent on the dielectric properties around the material (Han et al. 2017).In Fig. S2 (A-B), the SERS intensity of 4-MBA at 1078 cm -1 showed an obvious increase in the range of 0.5 μL to 4 μL.Although the SERS signal was still enhanced when the modi cation amount exceeded 2 μL, the stability of signal measured by repeated tests signi cantly dropped, which strongly in uenced the detection sensitivity.Moreover, the TEM image of Fig. S2D characterized the morphology of bimetallic Au@Ag NSs at the AgNO3 addition amount of 3 μL.It was found that the nanocomposites had completely tended to be spherical owing to the modi cation of extremely thick silver shells, which was also the direct cause of above signal instability.The results showed that the thickness of the Ag shell in Au NSs tips manifested positive correlation with the enhancement effects, while the change of morphology was signi cantly related to signal stability, showing that appropriate Ag shell thickness was a vital factor for stable detection (Mott et al. 2012).Thus, 0.8 μL of 4-MBA (1 mM) and 2 μL of AgNO 3 (0.1 M) were chosen as the optimal addition amounts for the preparation of SERS probes.
The coupled AFB 1 apt on the surfaces of Au@Ag NSs was illustrated by UV-Vis spectroscopy and zeta potential measurements in Fig. S3.For Au@Ag NSs-AFB 1 apt, the absorbance of the supernatant decreased greatly, and the average zeta potential decreased to -33.9 mV, which was due to the characteristic absorption peak at 260 nm (Wu et al. 2012), and the presence of a negatively charged phosphate group of the introduced ssDNA (Zhu et al. 2021).Thus, the functionalized bimetallic nanoprobes were prepared successfully.

Characterization and optimization of the fabricated Fe 3 O 4 @MoS 2 NFs
In order to obtain magnetic substrates, Fe 3 O 4 @MoS 2 NFs was synthesized by secondary hydrothermal methods.As shown in Fig. 2A, Fe 3 O 4 NPs were mostly homogeneous spherical morphology with a particle size of around 231.77 nm ± 17.97 nm.The magnetic response feature of Fe 3 O 4 NPs was further veri ed by the separation test of the external magnetic eld.The illustration of Fig. 2A showed that effective separation and enrichment by magnets were achieved within 15 s, indicating good paramagnetic properties of Fe 3 O 4 NPs.Fig. S4A revealed that 2D MoS 2 NSs (carbon-based nanomaterials) were highly transparent nanosheets with numerous folds.After another high temperature reaction, MoS 2 NSs were successfully combined around spherical Fe 3 O 4 NPs as revealed in Fig. 2B, forming multi-functional 3D core-shell Fe 3 O 4 @MoS 2 magnetic nano owers.Besides, the time of good magnetic separation was 30s for Fe 3 O 4 @MoS 2 NFs in the illustration of Fig. 2B, indicating that the load of moderate MoS 2 NSs had a weak in uence on the magnetic effect, which was suitable to be SERS magnetic substrate in subsequent analysis.The combination of plasmonic Fe 3 O 4 NPs with good magnetic effect and MoS 2 NSs with large surface area was conducive to rapid enrichment of SERS signal in the complex detection system.Moreover, the crystalline morphology of materials was con rmed by XRD in Fig. 2C.The typical diffraction peaks of Fe 3 O 4 NPs at 30. 6°, 35.9°, 43.5°, 53.9°, 57.4° and 62.9°w ere ascribed to (220), ( 331), ( 400), ( 422), ( 511) and (440) planes, which was consistent with their standard cards .Then, the diffraction peaks at 18.1° (002), 36.1°(100), 43.9° (103) and 58.1° (110) proved the fabrication of MoS 2 NSs nanostructure.The characteristic peaks of Fe 3 O 4 NPs were measured for XRD spectra of Fe 3 O 4 @MoS 2 NFs, indicating that the crystal morphology and phases of Fe 3 O 4 NPs were preserved.The results were similar with those of Lu et al (Lu et al. 2021).As shown in Fig. 2D, the Zeta potential of the nal composite was -19.5 mV as a result of the MoS 2 NSs with negative charge (-32.3 mV) as shells wrapping around the positive charged Fe 3 O 4 NPs cores (18.9 mV).
Therefore, all the above results characterized the successful preparation of 3D magnetic Fe 3 O 4 @MoS 2 nano owers, which provided more active sites for the connection of the aptamers to facilitate the construction of next detection system.

Feasibility and optimization of 3D SERS aptasensor for AFB 1 detection
To obtain the designed assemblies to analyze AFB 1 , the TEM images were employed to characterize the surface morphology of SERS sensor.From Fig. 3A, the anisotropic bimetallic nanostars were successfully bound around 3D Fe 3 O 4 @MoS 2 NFs core to build assemblies, due to the aptamers were combined to the outer shell MoS 2 NSs through non-covalent bonds.In SERS spectra of Fig. 3B, Fe 3 O 4 @MoS 2 magnetic substrates showed no Raman peaks in the range from 900 cm -1 to 1800 cm -1 .However, compared with Au-4MBA@Ag NSs, the stronger 4-MBA SERS signal was observed on the plasmonic Fe 3 O 4 @MoS 2 NFs-AFB 1 apt-Au-4MBA@Ag NSs assemblies, owing to the fact that chemical enhancement of MoS 2 NSs synergistically enhanced SERS performance of the assemblies.After adding target AFB 1 , 4-MBA response signi cantly reduced, due to the decrease of SERS probes on the assemblies.Moreover, TEM images of the assemblies without and with AFB 1 were clearly shown in Fig. S5(A-B), indicating that the presence of AFB1 induced the dissociation of nanostars from the substrates due to the forming of Au@Ag NSs-AFB1apt/AFB 1 composites.Thus, all the above results demonstrated the feasibility of SERS sensor for quantitatively sensing AFB 1 based on signi cant changes in SERS signal peak intensity caused by changes in the composition of assembler.
To further obtain highly sensitive SERS aptasensor, Fe 3 O 4 @MoS 2 NFs as magnetic substrates and Au-4MBA@Ag NSs-AFB 1 apt as SERS probes were mixed in different volume ratios at room temperature.As obtained from Fig. 3C, the SERS intensity of the assemblies at 1581 cm -1 reached the strongest at the volume ratio of 1:2, which was chosen as the optimal addition ratio.Besides, in order to evaluate the reproducibility and stability of the constructed AFB 1 SERS sensor, the original Raman spectra of random 15 points shown in Fig. 3D (waterfall plot) were collected on the assemblies.And the RSD value of Raman peak intensity (1581 cm -1 ) was calculated to be 7.6% (Fig. S6A), manifesting the good uniformity of signals on the SERS assemblies.Different storage times (1, 3, 5, 7, 9, 11, and 13 days) of 3D aptasensor were also studied (Fig. S6B).SERS signal intensity remained 86.5% on the around thirteenth day, revealing the good time-dependent stability of sensor.

Sensitivity of AFB 1 detection by SERS aptasensor
To verify the detection capability of the aptasensor under optimized experimental conditions, a series of AFB 1 addition amounts were reacted with the SERS system at the nal concentration from 0.1 ng/mL to 1000 ng/mL.As shown in Fig. 4A, the SERS signal intensity of 4-MBA molecular gradually decreased with the increase of AFB 1 concentration, which was owing to the release of a large amount of 4MBA-based SERS probes induced by AFB 1 from the magnetic substrates.The SERS peak intensity of 4-MBA at 1581 cm -1 (I 1581 ) as quantitative value showed negatively linearly correlation with the logarithmic value of AFB 1 concentration (log 10 C AFB1 ).After the linear t in Fig. 4B, the regression equation was Y = 8541.55-2348.69X(R 2 = 0.986).The associated LOD was as low as 58.9 pg/mL calculated by the formula (LOD = 10 (Y+3SD-A)/B .Where Y and SD are the mean and standard deviation of the blank sample signal, respectively.A and B represent the intercept and slope of the curve, respectively.And signal-to-noise ratio was set as 3:1).The detection sensitivity of 3D plasma SERS nanoassemblies satis ed the EU requirements for the measure of the maximum limit of AFB 1 in cereal.Additionally, Table S1 gave a comparison between our work for AFB 1 detection and other previously reported analytical methods.The results showed that this study had obtained lower detection limits and wider detection range relative to the work of some researchers, but the sensitivity still needed to be further improved.

Selectivity, reproducibility, and practicality evaluation
To validate the selectivity of the aptasensor, AFM 1 , OTA and FB 1 toxins were selected to detect under the same experiment.In Fig. 5(A-B), I 1581 of the three interfering toxins changed weakly compared to that of the blank sample, while that of AFB 1 (100 ng/mL) decreased signi cantly, showing that the speci c aptamer conferred good selectivity to the SERS sensor.To evaluate the reproducibility of the proposed sensor, AFB 1 (10 ng/mL) was analyzed repeatedly on the same batch and different six batches.It can be seen from the SERS intensity bar chart in Fig. S7 that this strategy presented relatively low and acceptable RSD values of I 1581 , which were 5.1% and 6.4%, respectively.To verify the practicality of the method in real samples, the peanuts at AFB 1 spiked concentrations of 0.5 ng/mL-50.0ng/mL were selected for the recovery experiments.The satisfactory recoveries ranging from 97.9% to 99.4% with RSDs of 4.3%-9.3%were obtained as shown in Table 1, indicating that the assay was suitable for the detection of peanut, and had bright application prospect in food safety analysis.

Conclusions
In all, a rational SERS aptasensor for quantitative detection of AFB 1 has been contrasted based on 3D plasmonic magnetic SERS assemblies (Au-4MBA@Ag NSs-AFB 1 apt-Fe 3 O 4 @MoS 2 NFs).The designed SERS-active platform has the following excellent performance: (1) rapid magnetic separation features, (2) good stability of substrates for long-term storage, (3) signal ampli cation function provided from abundant hot spots, (4) synergistic enhancement effect of bimetal Au-Ag and MoS 2 NSs on the SERS signal of 4-MBA.AFB 1 toxin was sensitively detected from 0.1 ng/mL to 100 ng/mL, achieving the LOD as low as 58.9 pg/mL.The high selectivity under interfering toxins, and good reproducibility within different batches measurement of this sensor were further determined.Moreover, the strategy can be capable of detecting AFB 1 in peanut samples with good recoveries of 97.9%-98.7%.Therefore, the proposed SERS sensor as effective detection technique has the potential to be applied to the monitoring of various hazards in the eld of food safety.

Figures Figure 1
Figures