Chemiluminescence Switching on Peroxidase-Like Fe3O4Nanoparticles for Selective Detection and SimultaneousDetermination of Various PesticidesGuijian Guan,† Liang Yang,† Qingsong Mei,† Kui Zhang,† Zhongping Zhang,*,† and Ming-Yong Han*,†,‡

†Institute of Intelligent Machines, Chinese Academy of Sciences, Hefei, Anhui 230031, China‡Institute of Materials Research and Engineering, A-STAR, 3 Research Link, 117602, Singapore

*S Supporting Information

ABSTRACT: To achieve selectivity in direct chemiluminescence (CL) detection isvery significant and a great challenge as well. Here, we report a novel concept ofdeveloping intrinsically selective CL switching at the surface of Fe3O4 nanoparticles forthe sensitive detection and simultaneous determination of various pesticides. Fe3O4nanoparticles have peroxidase-like catalytic activity and catalyze the decomposition ofdissolved oxygen to generate superoxide anions, so that the CL intensity of luminolwas amplified by at least 20 times. The CL signals can be quenched by the addition ofethanol because ethanol readily reacts with superoxide anions as a radical scavenger.However, the quenching effect can be inhibited through the specific binding of target molecules on Fe3O4 nanoparticles, leadingto CL “turn-on” in the presence of ethanol. The novel CL “switching-on” concept demonstrated unique advantages in thedetection of pesticide residues. Using the surface coordinative reactions, nonredox pesticide ethoprophos were sensitivelydetected with a detection limit of 0.1 nM and had a very wide detection range of 0.1 nM to 100 μM. More importantly, theselectivity of CL switching is tunable through the special surface modification of Fe3O4 nanoparticles, and these Fe3O4nanoparticles with different surface groups can generate unique CL response pattern for the simultaneous determination ofvarious pesticides. Meanwhile, the superparamagnetic properties of Fe3O4 nanoparticles provide a simple magnetic separationapproach to attain interference-free measurement for real detection. The very facile and versatile strategy reported here shouldopen a new window to exploration of selective CL molecular switching and application of magnetic nanoparticles for chemo/biodetection.

In the past decades, chemiluminescence (CL) has widelybeen investigated as a facile, fast, sensitive, and cost-effective

analytical technique for chemical analyses,1 biological assays,2

clinical diagnoses,3 and environmental detections,4 due to theunemployment of the excitation light source and the opera-tional simplification of the equipment.5 Because the classical CLsystems have very low efficiency for transforming the chemicalenergy into light,6 the current systems generally needed tofurther enhance their CL efficiency to give intense emissionintensity for quantitative analysis. Recently, a number ofapproaches have been explored for the enhancement of theefficiency by employing various catalysts including the usualperoxidase enzyme7 and unique surface-active nanomaterialssuch as noble metal nanoparticles (Au, Ag, and Pt)8 and metaloxide nanostructures (TiO2, ZnO2, and ZrO2).

9 Although theCL has been enhanced significantly by several to tens of times,these methods give us a general impression that they originatein the spontaneous reactions in a redox pair10 and sointrinsically lack molecular selectivities to detect specific targets.Nowadays, it remains a great challenge to achieve the selectivityin direct CL detections while improving CL efficiency. Incomparison, the selective detection depends on immunoassay,electrophoresis, solid phase extraction (SPE), and high-performance liquid chromatography (HPLC),11 which com-promise the simpleness and cheapness of CL assays, and the

operating procedures also are time-consuming, tedious, andunavailable in the in-field rapid detections.In this paper, we report an intrinsically selective CL

switching on peroxidase-like Fe3O4 nanoparticles for thesensitive detection and simultaneous determination of variouspesticides. Fe3O4 nanoparticles with 10-nm size have highperoxidase-like catalytic activity12 and can catalyze thedecomposition of dissolved oxygen to generate superoxideanions at their surface. The surface superoxide anions are easilyscavenged by the addition of ethanol; however, the scavengingeffect can be effectively inhibited through the specific binding oftarget molecules on Fe3O4 nanoparticles. On the basis of thisnew finding, a CL switching-on chemosensor was facilelyestablished and would have three unique advantages in CLdetections: (1) The selectivity is determined by the bindingcapacities of analytes on Fe3O4 nanoparticle surfaces; therefore,the novel sensing concept can be used to sensitively detectnonredox targets without any additional techniques; (2) Fe3O4

nanoparticles with different surface groups exhibit differentmolecular selectivity, so the combination of diverse nano-

Received: August 14, 2012Accepted: October 2, 2012Published: October 2, 2012

Article

pubs.acs.org/ac

© 2012 American Chemical Society 9492 dx.doi.org/10.1021/ac302341b | Anal. Chem. 2012, 84, 9492−9497

particles would generate a unique pattern to a special pesticide,which should provide a facile method for determination; (3)the superparamagnetic properties of Fe3O4 nanoparticlesprovide a simple magnetic separation approach to attaininterference-free measurement for real detection.

■ EXPERIMENTAL SECTIONChemicals and Materials. Ferric chloride (FeCl3·6H2O),

ferrous chloride (FeCl2·4H2O), H2O2 (30 wt %), hydrochloricacid (36−38 wt %), ammonium hydroxide (25−28 wt %), citricacid, β-cyclodextrin (β-CD), methanol, superoxide dismutase(SOD), and ethanol were purchased from Sinopharm ChemicalReagent Co., Ltd. (Shanghai, China). 3-Aminopropyltriethox-ysilane (APTS), luminol, and the typical pesticides includingethoprophos (EP), profenofos (PF), dylox (DL), 2,4-dichlorophenoxyacetic acid (2,4-D), parathion-methyl (PM),nicosulfuron (NS), and endosulfan (ES) were supplied fromSigma-Aldrich. All of these reagents were used without furtherpurification. Ultrapure water (18.2 MΩ cm) was producedusing the Millipore water purification system. Luminol solutionin CL reactions was freshly prepared by diluting the mothersolution (50 mM, keeping in dark for at least 1 week) in 0.1 MNaOH solutions.Synthesis of Fe3O4 Nanoparticles with 10-nm Size.

Fe3O4 nanoparticles were prepared via the modified copreci-pitation method.12 Typically, 2.4 g of ferrous chloride and 4.6 gof ferric chloride were first dissolved in 20 mL of oxygen-freeultrapure water (pH 2) and subsequently were filtered into 80mL of ultrapure water. After purging with nitrogen gas for 30min, 10 mL of ammonium hydroxide was dropwise added intothe above mixed solution under vigorous stirring at roomtemperature in a nitrogen atmosphere. The black colloidalsolution was stirred at 40 °C for 0.5 h and 85 °C for 2 h,respectively. Finally, the Fe3O4 nanoparticles were separatedand purified by magnetic separation and washed with deionizedwater and ethanol 3 times, respectively, and then redispersed inwater for future use.CL Detection of Pesticide. A 10−2 M EP solution in

acetonitrile was prepared and then diluted with water to thedesired concentrations. For detecting pesticide, 30 μL of Fe3O4colloid, 30 μL of EP solution, and 30 μL of ethanol were firstadded into the well of a 96-well plate. After shaking for 10 s, 90μL of luminol solution was added, followed by collecting theCL signals. The blank experiments were similarly carried outonly that the EP solution was replaced with the same volume ofaqueous solution. The CL response was determined bycalculating the difference of CL intensities between the sampleand blank. The detection selectivity was further investigatedusing various kinds of pesticides as controls.Real Detection of CL Analysis. Grape juice and green tea

were selected as the test samples to demonstrate the procedureof detecting pesticide residues in real samples. The Fe3O4nanoparticles were first dispersed in sample solution spikedwith EP pesticide to form a dark homogeneous suspension.After the mixture was incubated on a rocking table with shakingfor 10 min, the Fe3O4 nanoparticles were drawn to the wall ofthe vial through adding a small magnet. After discarding thesupernatant, the magnet was taken away and the vial wasshaken gently after adding a certain amount of water/ethanol(v/v, 2:1) mixing solution. The black aggregates were againredispersed well, and the colloidal solution was used to measureCL emission. The above separation/redispersion procedure canbe repeated until the noninterfering detection was obtained.

Characterization. CL signals were recorded in a 96 wellpolypropylene microtiter plate using a Berthold LB 960microplate luminometer. The structures of the Fe2+ complexeswere determined with a ProteomeX-LTQ mass spectrometeremploying a regular electrospray ionization (ESI) source setup.The morphologies of Fe3O4 nanoparticles were examined bytransmission electron microscopy (TEM, JEOL 2010), and thecrystalline structure was examined by X-ray diffraction on aMAC Science Co. Ltd. MXP 18 AHF X-ray diffractometer withmonochromatized Cu Kα radiation (λ = 1.540 56 Å). Themagnetization measurements of the ferromagnetic Fe3O4nanoparticles at room temperature were performed with aQuantum Design superconducting quantum interference device(SQUID) magnetic properties measurement system. Theinfrared spectra were recorded with Nicolet Nexus-670 FT-IRspectrometer.

■ RESULTS AND DISCUSSIONIn this work, we prepared the Fe3O4 nanoparticles with a 10-nm size by the coprecipitation method (see Figure S1A in theSupporting Information),13 considering both the high catalyticactivity and good monodispersity. The obtained nanoparticlesexhibit high crystalline quality and only have the face-centered-cubic magnetite (JCPDS no. 01-1111) (see Figure S1B in theSupporting Information). Moreover, the Fe3O4 nanoparticlesare superparamagnetic (see Figure S1C in the SupportingInformation) and have high magnetic response ability (insetimage).14 These characteristics will endow Fe3O4 nanoparticleswith versatility and power as a CL enhancer, sensing probe, andseparation tool. When adding the Fe3O4 nanoparticles into theluminol solution, the CL intensity was amplified by at least 20times (see Figure 2S in the Supporting Information). Throughcomparing the CL intensities from luminol in various mediumsincluding Fe2O3 colloidal solution, the leaching solution fromFe3O4 nanoparticles and nitrogen-saturated Fe3O4 colloidalsolution, it can be concluded that the strong light emission isattributed to Fe2+ ions in the Fe3O4 nanoparticles and dissolvedoxygen.12 Meanwhile, the optimization of reaction conditionsshows the optimized pH value of Fe3O4 colloidal solution was 2for yielding the highest light emission when the pH value ofluminol solution was fixed at 13 (see Figure S3 in theSupporting Information). These results would suggest the CLmechanism of the luminol−Fe3O4 system.

15 As demonstratedin Figure S4 in the Supporting Information, dissolved oxygenfirst decomposes into superoxide anions at the surface of Fe3O4nanoparticles, which mainly results from the peroxidase-likecatalytic activity of Fe3O4 nanoparticles in acidic conditions.12

The resultant superoxide anions further oxidize luminol in basicmedia to produce the strong CL emission.Table 1 demonstrates the unique advantages of the luminol−

Fe3O4 CL system by using pesticide ethoprophos (EP) as thetarget analyte. EP is a widely used organophosphorus pesticidebut is not detected through the traditional CL reaction of theluminol−H2O2 whether adding aqueous solution or ethanolsolution. Similarly, there also is no CL response of the EPaqueous solution in the luminol−Fe3O4 system. However, if EPethanol solution is added, a significant CL enhancement can beclearly observed. For understanding the function of ethanol, wefurther studied the effect of water and ethanol on CL emission.As shown in Table 1, water do not cause the CL change butethanol can lead to a significant CL quenching of the luminol−Fe3O4 system. The quenching process follows a nonlinearbehavior and shows a saturation concentration of ethanol above

Analytical Chemistry Article

dx.doi.org/10.1021/ac302341b | Anal. Chem. 2012, 84, 9492−94979493

1 M (see Figure S6 in the Supporting Information). Ethanol iswell-known as a radical scavenger and readily reacts withsuperoxide anions,16 leading to the quenching of CL emissiondue to the loss of superoxide anions.As shown in the inset of Figure 1A, the different CL

responses of ethanol and the mixture of EP and ethanol on

luminol−Fe3O4 system provide a CL switching-on way forsensitively detecting EP molecules. The CL kinetic character-istics under different EP concentrations are shown in Figure S7in the Supporting Information. With the addition of EPsolution to the luminol−Fe3O4−ethanol system, the CL willcontinuously enhance with the increase of EP concentration.About a 16-fold CL enhancement was measured when theconcentration of EP reached 60 μM. Such high enhancing

response demonstrates that EP molecules are efficient to inhibitthe scavenging of superoxide anions. Figure 1A shows the plotof the CL enhancement coefficient (I/I0) vs the concentrationof EP. Even at a concentration as low as 0.1 nM, CLenhancement can still be clearly observed. In addition, it can beseen that there are two consecutive linear detecting ranges (seeFigure S8 in the Supporting Information), which correspondwith a single molecular process (SMP) and a double molecularprocess (DMP), respectively (see below). The evolution of CLintensity is suitable for the detection of pesticide EP within avery wide range of 0.1 nM to 100 μM.The detecting selectivity was further investigated by

comparing the CL responses of Fe3O4 nanoparticles tostructure-analogous pesticides. As shown in Figure 1B, onlyEP and PF organophosphorus esters (containing PO bond)with a phosphorus−sulfur bond (P−S bond) are able to lightup the CL in the presence of ethanol, and the EP molecule withtwo P−S bonds had much higher enhancement than the PFmolecule with one P−S bond. Although the addition of dylox(DL) without a P−S bond can also enhance CL intensity in asmall quantity, the CL response can still be negligible comparedwith that of EP. In addition, 2,4-dichlorophenoxyacetic acid(2,4-D) did not nearly cause any change in CL intensity fromthat of blank samples. Therefore, the luminol−Fe3O4 systemshows very high specificity for the detection of organo-phosphorus esters with a P−S bond.To understand the mechanism of CL switching, we detailedly

studied the interaction between EP molecules and Fe2+ ions byusing FeCl2 as source materials. Figure S9 in the SupportingInformtion shows the electrospray ionization-mass spectrome-try (ESI-MS) of the mixtures of EP with Fe2+ ions in theethanol/water mixture solution. The protonated (EP)2FeCl2H

+

peak at m/z = 610.99−612.46 was clearly observed, suggestingthe formation of the (EP)2−Fe(II)−Cl2 complex. In thiscomplex, the central iron(II) atom is thought as hexa-coordinated in an octahedral structure.17 For determining thecoordinative groups, we introduced theoretical energy calcu-lations, as shown in Table 2. According to the principle ofenergy minimum, the conformation was optimized as thecoordination of sulfur atom, but not oxygen atom, besidedoubly linked oxygen and chlorine (see the inset of Table 2 andFigure S9 in the Supporting Information). This result isunderstandable because sulfur is more polarizable than oxygenand thus has a stronger coordinative bond formation. Thesecalculations and discussions imply that EP molecules will havethe strongest coordinative ability on Fe2+ ions, among the fourdifferent pesticides. Therefore, it is reasonable to suppose thatthe CL enhancement is attributed to the coordination of EPmolecules on surface Fe2+ ions that inhibits the scavenging ofsuperoxide anions.

Table 1. Pesticide Ethoprophos (EP) and the MediumsProduce the CL Response on Luminol (LUM)-Based CLSystemsa

CL systems samples CL response

LUM + H2O2 EP aqueous solutionb no changeEP ethanol solutionb no change

LUM + Fe3O4 NPs EP aqueous solution no changeEP ethanol solution CL enhancing

LUM + Fe3O4 NPs waterc no changeethanolc CL quenching

aThe detailed experiment and results were provided in Figure S5 ofSupporting Information. bThe reference CL signals are the CLintensities only that the EP solution is replaced with the same volumeof water or ethanol. cThe reference signals are the CL intensitieswithout the addition of water or ethanol.

Figure 1. CL switching-on chemosensor for sensitively and selectivelydetecting pesticide EP. (A) CL enhancement of the luminol−Fe3O4−ethanol system with increasing the concentration of EP (inset showsthe CL “on-off-on” process). (B) The CL response selectivity tovarious pesticides (0.1 mM): ethoprophos (EP), profenofos (PF),dylox (DL), and 2,4-dichlorophenoxyacetic acid (2,4-D). (I0 and I arethe CL intensity of the luminol−Fe3O4−ethanol system in the absenceand presence of pesticide, respectively).

Table 2. Optimized Conformations of the Complexes of (EP)2FeCl2 Based on the Theoretical Calculation of the BindingEnergies (ΔE) for Different Coordination Formats

Analytical Chemistry Article

dx.doi.org/10.1021/ac302341b | Anal. Chem. 2012, 84, 9492−94979494

Similarly, we study the interaction between EP molecules andFe3+ ions by using FeCl3 as the source material. From the ESI-MS of the mixture, a (EP)−Fe(III)−Cl3 complex can beformed through the tetra-coordination of the central iron(III)atom in the tetrahedral structure. On the basis of thisconformation, the binding energy of EP on Fe3+ ion wascalculated as 132.39 kcal/mol, which is much smaller than thebinding energy of EP on the Fe2+ ion of 310.37 kcal/mol(Table S1 in the Supporting Information). Therefore, we havereason to think that the CL enhancement is because of thecoordinative binding of EP molecules on the Fe2+ ion, not onthe Fe3+ ion.The mechanism of CL switching is illustrated in Figure 2.

Superoxide anions are generated from the decomposition of

dissolved oxygen under the intrinsic peroxidase-like catalyticactivity of Fe3O4 nanoparticles, and the resultant superoxideanions will absorb at the surface of the Fe3O4 nanoparticles dueto the high surface-to-volume ratio of the nanoparticles (1).Then, ethanol reacts with superoxide anions through the radicalscavenging process (2),16 so there is no light emission after theaddition of luminol (3). As another way, before adding ethanol,EP molecules are first added into the Fe3O4 colloidal solution.The EP molecules would firmly bind at the surface of the Fe3O4nanoparticles through the coordinative reaction with the surfaceFe2+ ions, and their branched chains will encircle the surfacesuperoxide anions (4). This molecular structure can effectivelyinhibit the scavenging of superoxide anions from ethanol (5);therefore, the addition of luminol still produces a strong lightemission even if in the presence of ethanol (6). The two CLresponse formats are combined to form a CL “switching-on”chemosensor (7).The art of the CL switching-on method is that the EP

molecules have much stronger binding capacity at the Fe3O4nanoparticle surface than ethanol, whereas luminol moleculessimilarly have strong coordinative ability on Fe3O4 nano-particles through the interactions between amino groups andFe2+ ions.18 Therefore, EP molecules can obstruct theapproaching of ethanol to superoxide anions, but luminol canreach the surface of Fe3O4 nanoparticles through thecompetitive adsorption with EP molecules and is oxidized bysuperoxide anions to yield light emission. When theconcentration of EP is lower (smaller than 1 μM), only one

EP molecule can bind on one surface Fe2+ ion. In this state(Figure 1A, SMP), the inhibition of radical scavenging is notcomplete, so that the CL enhancements are slow with theincrease of EP concentration. Once the concentration of EP ishigher than 1 μM, the double molecular process will occur. Thecoordination of two EP molecules on one Fe2+ ion caneffectively inhibit the scavenging of free radicals, resulting in themuch stronger enhancement of CL intensity. Furthermore, wealso investigated the validity of other radical scavengers such asmethanol and superoxide dismutase (SOD). The experimentresults indicated that the CL switching can still be achievablewith methanol or SOD, only their sensitivities are lower thanthat of using ethanol. Therefore, ethanol was used as a radicalscavenger in the detection of pesticide EP.According to the mechanism demonstrated above, the

selectivity of CL switching is tunable through the surfacemodifications of Fe3O4 nanoparticles. Herein, we selected fourdifferent functionalized Fe3O4 nanoparticles (containing bareFe3O4 nanoparticles, NPs) as examples to represent thealterability of selectivity. Typically, the citric acid monomerand β-cyclodextrin (β-CD) molecules were grafted at thesurface of Fe3O4 nanoparticles,19 leading to the formation ofcarboxyl-capped nanoparticles (CNPs) and β-CD-modifiednanoparticles (CDNPs), respectively. Meantime, amine-cappednanoparticles (ANPs) were prepared by the surface silanecoupling reaction with 3-aminopropyltriethoxysilane mono-mer.20 The existence of functional groups was evidenced byFourier transform-infrared spectroscopy (see Figure S10 in theSupporting Information). For the differentiability test, fourtypical pesticides including EP, parathion-methyl (PM),nicosulfuron (NS), and endosulfan (ES) (see Figure S11 inthe Supporting Information) were chosen as sensing targets. Asillustrated in Figure 3, direct addition of pesticide solution

resulted in a variety of unique CL response patterns due totheir diverse binding abilities on functionalized Fe3O4 nano-particles, which can be used to determine the kind of targetanalytes.In addition, Figure 3 also suggests another superiority of the

CL switching-on concept. The given molecules of interest canbe sensitively detected through the special surface functional-ization of Fe3O4 nanoparticles (Figure S12 in the SupportingInformation). For example, the bare Fe3O4 nanoparticles arenot sensitive to organophosphorothioate pesticide PM, whichhas been frequently used for the protection of crops and itsresidues have threatened human health.21 For improving thesensitivity, we modified Fe3O4 nanoparticles with β-CDmolecules. It can be expected that an inclusion complex with

Figure 2. Mechanism of CL switching at the surface of Fe3O4nanoparticles: (1) dissolved oxygen transfers into superoxide anionsat the surface of Fe3O4 nanoparticles; (2) the resultant superoxideanions are scavenged by ethanol, so that (3) there is no light emissionafter adding luminol. Alternatively, (4) pesticide EP molecules are firstadded into Fe3O4 colloidal solution and bind onto Fe3O4 nanoparticlesthrough the coordinative reaction with surface Fe2+ ions; (5) thebound EP molecules inhibit the scavenging of superoxide anions, sothat (6) there still is a strong CL emission after adding luminol. (7) Ifusing the quenched CL intensity as the reference signal, a novel CL“turn-on” concept is formed.

Figure 3. Different response pattern to the special pesticide generatedby different surface-functionalized Fe3O4 nanoparticles for thedetermination. The concentration of pesticides was 1 μM. (PM,parathion-methyl; NS, nicosulfuron; ES, endosulfan).

Analytical Chemistry Article

dx.doi.org/10.1021/ac302341b | Anal. Chem. 2012, 84, 9492−94979495

guest molecules can be efficiently formed between aromatic PMmolecules and cyclodextrin;19 therefore, β-CD-modified nano-particles substantially increased the enhancement coefficientfrom 1.38 to 2.54. This universal and sensitive strategy might bethe breakthrough in CL assays for the simultaneous detectionand determination of multiple target analytes.Another outstanding and unique advantage of CL switching

at the surface of Fe3O4 nanoparticles is that the super-paramagnetic properties provide a simple magnetic separationapproach to attain interference-free measurement for realdetection.14 In this work, green tea was used as the test sampleand spiked with a certain amount of EP molecules. Green tea isa buffer solution with pH values at 6, and thus the directaddition of green tea yields no light emission for any luminol-based CL systems. Therefore, to detect analytes in green tea,the general CL techniques either show undetectable or needcomplicated sample pretreatments such as SPE and HPLC. TheFe3O4 nanoparticles with a high saturation magnetization couldbe quickly separated from a suspension under an externalmagnetic field, providing a rapid method to overcome theinterference of coexisting substances. As shown in Figure 4A,the Fe3O4 nanoparticles were first dispersed in green tea spikedwith EP molecules. Through the coordinative reaction of EP atthe surface of the Fe3O4 nanoparticles, most of the EPmolecules will bind on the Fe3O4 nanoparticles. When a smallmagnet was put near the vial, the nanoparticles with EP weredrawn to the wall of the vial. After discarding the green tea, theblack aggregates were again redispersed in a certain amount ofwater/ethanol mixing solution for measuring the CL signals.The above separation/redispersion procedure was repeated twotimes, and the noninterfering detection was obtained. Figure 4Bshows that the CL enhancement was obviously observed andproduced an enhancement coefficient of 2.11. In addition, wealso tested spiked grape juice (pH = 4.5), and the enhancementcoefficient can be up to 2.39 after the separation/redispersionwas done for one time (see Figure S13 in the SupportingInformation). Through controlling the amount of Fe3O4nanoparticles and the volume of redispersion solution, thedetection sensitivity can further be improved. Detailedexperiments have revealed that both of the detection limitscan be as low as approximately nanomolar.The above results also suggest that we can calculate the total

concentration of the pesticides in real samples. If there areseveral kinds of pesticides, we can first measure the amount ofeach pesticide after magnetic separation of the special surface-functionalized nanoparticles. Then, the total concentration ofpesticides naturally becomes the addition of each concen-

tration. Obviously, the sensitivity, selectivity, versatility, andpracticability of the CL switching-on chemosensor should meetthe requirements for detection of pesticide residues in complexproducts.

■ CONCLUSIONSIn summary, we have developed an intrinsically selective CLswitching-on chemosensor by the employment of peroxidase-like Fe3O4 nanoparticles and radical scavenger ethanol. Thesensing mechanism is based our new finding that the specificbinding of target molecules at the surface of Fe3O4 nano-particles inhibits the scavenging of surface radicals, leading tothe CL “turn on” in the presence of ethanol. The CL “turn-on”chemosensor was used to sensitively detect nonredox moleculeswith a high specificity and has the ability to detect molecules ofinterest in complexed and real samples after a simple magneticseparation. More importantly, the selectivity of CL switchingcan facilely be tuned through the special surface modificationaof Fe3O4 nanoparticles, and these CL switches with differentselectivity can generate unique CL response pattern for thesimultaneous discrimination of analytes. The very simple andversatile strategy reported here should open a new window ofinterest in the application of magnetic nanoparticles and thedevelopment of CL chemosensors.

■ ASSOCIATED CONTENT*S Supporting InformationPartial experiments, characterization of Fe3O4 NPs, optimiza-tion of the reaction conditions, CL mechanism, massspectrometry, linear correlations, and infrared spectra. Thismaterial is available free of charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Author*E-mail: zpzhang@iim.ac.cn (Z.Z.); my-han@imre.a-star.edu.sg(M.-Y.H.).NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis project was financed by Natural Science Foundation ofChina (Grant Nos. 21277145, 21275145, 30901008, 21077108,61071055, and 20925518) and the National Science &Technology Pillar Program (Grant 2013BAJ24B02-3). Wethank Prof. Shuhu Du and Mr. Jiawei Zhao from NanjingMedical University for assistance with the Gaussian 09 software.

Figure 4. CL detection of pesticide EP in real samples. (A) Schematic drawing of the magnetic separation/redispersion process for overcoming theinterfering effect in green tea. (B) CL response resulted from EP molecules in green tea after carrying out the separation/redispersion process twotimes (the CL signals resulted from three different batches). The enhancement coefficient (EC) is the ratio of the CL intensities of sample withpesticide to that of without pesticide.

Analytical Chemistry Article

dx.doi.org/10.1021/ac302341b | Anal. Chem. 2012, 84, 9492−94979496

■ REFERENCES(1) (a) Huang, X. Y.; Li, L.; Qian, H. F.; Dong, C. Q.; Ren, J. C.Angew. Chem., Int. Ed. 2006, 45, 5140−5143. (b) Liu, M. L.; Lin, Z.;Liu, J. M. Anal. Chim. Acta 2010, 670, 1−2.(2) (a) Aslan, K.; Geddes, C. D. Chem. Soc. Rev. 2009, 38, 2556−2564. (b) Gill, R.; Polsky, R.; Willner, I. Small 2006, 8−9, 1037−1041.(c) Zhang, S. S.; Zhong, H.; Ding, C. F. Anal. Chem. 2008, 80, 7206−7212.(3) (a) Bourgarit, A.; Carcelain, G.; Martinez, V.; Lascouc, C.;Delcey, V.; Gicquel, B.; Vicaut, E.; Lagrange, P. H.; Sereni, D.; Autran,B. AIDS 2006, 20, 1−7. (b) Schirpenbach, C.; Seiler, L.; Maser-Gluth,C.; Beuschlein, F.; Reincke, M.; Bidlingmaier, M. Clin. Chem. 2006, 52,1749−1755.(4) (a) Gamiz-Gracia, L.; Garcia-Campana, A. M.; Soto-Chinchilla, J.J.; Huertas-Perez, J. F.; Gonzalez-Casado, A. TrAC, Trends Anal. Chem.2005, 24, 927−942. (b) Tsunoda, M.; Imai, K. Anal. Chim. Acta 2005,541, 13−23.(5) Adam, W.; Kazakov, D. V.; Kazakov, V. P. Chem. Rev. 2005, 105,3371−3387.(6) Lowry, M.; Fakayode, S. O.; Geng, M. L.; Baker, G. A.; Wang, L.;McCarroll, M. E.; Patomay, G.; Warner, I. M. Anal. Chem. 2008, 80,4551−4574.(7) Marquette, C. A.; Blum, L. J. Anal. Bioanal. Chem. 2006, 385,546−554.(8) (a) Zhang, Z. F.; Cui, H.; Lai, C. Z.; Liu, L. J. Anal. Chem. 2005,77, 3324−3329. (b) Duan, C. F.; Cui, H. Chem. Commun. 2009, 18,2574−2576. (c) Guo, J. Z.; Cui, H.; Zhou, W.; Wang, W. J. Photochem.Photobiol. A 2008, 193, 89−96.(9) (a) Yang, L.; Guan, G. J.; Wang, S. H.; Zhang, Z. P. J. Phys. Chem.C 2012, 116, 3356−3362. (b) Na, N.; Zhang, S. C.; Wang, S. A.;Wang, X. R. J. Am. Chem. Soc. 2006, 128, 14420−14421.(10) Xie, Z. H.; Zhang, F.; Pan, Y. S. Analyst 1998, 123, 273−275.(11) (a) Lin, J. H.; Ju, H. X. Biosens. Bioelectron. 2005, 20, 1461−1470. (b) Miao, W. J. Chem. Rev. 2008, 108, 2506−2553. (c) Liu, W.;Zhang, Z. J.; Liu, Z. Q. Anal. Chim. Acta 2007, 539, 187−192.(d) Zhang, Q. L.; Lian, M.; Liu, L. J.; Cui, H. Anal. Chim. Acta 2005,537, 31−39.(12) Gao, L. Z.; Zhuang, J.; Nie, L.; Zhang, J. B.; Zhang, Y.; Gu, N.;Wang, T. H.; Feng, J.; Yang, D. L.; Perrett, S.; Yan, X. Y. Nat.Nanotechnol. 2007, 2, 577−583.(13) Wei, H.; Wang, E. K. Anal. Chem. 2008, 80, 2250−2254.(14) Liu, B. H.; Han, M. Y.; Guan, G. J.; Wang, S. H.; Liu, R. Y.;Zhang, Z. P. J. Phys. Chem. C 2011, 115, 17320−17327.(15) Xie, J. X.; Huang, Y. M. Anal. Methods 2011, 3, 1149−1155.(16) Rose, A. L.; Waite, T. D. Anal. Chem. 2001, 73, 5909−5920.(17) Fernandes, C., Jr.; A., H.; Vieira-da-Motta, O.; De Assis, V. M.;Rocha, M. R.; Mathias, L. S.; Bull, E. S.; Bortoluzzi, A. J.; Guimaraes, E.V.; Almeida, J. C. A.; Russell, D. H. J. Inorg. Biochem. 2010, 104, 1214−1223.(18) Taylor-Pashow, K. M. L.; Rocca, J. D.; Xie, Z. G.; Tran, S.; Lin,W. B. J. Am. Chem. Soc. 2009, 131, 14261−14263.(19) He, S. H.; Shi, W. B.; Zhang, X. D.; Li, J.; Huang, Y. M. Talanta2010, 82, 377−383.(20) Liu, Y. P.; Yu, F. Q. Nanotechnology 2011, 22, 145704.(21) Zhang, K.; Mei, Q. S.; Guan, G. J.; Liu, B. H.; Wang, S. H.;Zhang, Z. P. Anal. Chem. 2010, 82, 9579−9586.

Analytical Chemistry Article

dx.doi.org/10.1021/ac302341b | Anal. Chem. 2012, 84, 9492−94979497