lable at ScienceDirect

Trends in Food Science & Technology 46 (2015) 252e263

Contents lists avai

Trends in Food Science & Technologyjournal homepage: ht tp: / /www.journals.e lsevier .com/trends- in- food-science-

and-technology

Review

Application of microfluidic “lab-on-a-chip” for the detection ofmycotoxins in foods*

Lujia Guo a, 1, Jinsong Feng a, 1, Zecong Fang b, Jie Xu c, Xiaonan Lu a, *

a Food, Nutrition and Health Program, Faculty of Land and Food Systems, The University of British Columbia, Vancouver, British Columbia V6T 1Z4, Canadab Department of Mechanical Engineering, Washington State University, Vancouver, 98686, United Statesc Department of Mechanical & Industrial Engineering, University of Illinois at Chicago, Chicago, IL 60607, United States

a r t i c l e i n f o

Article history:Received 19 August 2014Received in revised form24 July 2015Accepted 26 September 2015Available online 27 October 2015

Keywords:MycotoxinsMicrofluidic “lab-on-a-chip”Agricultural and food safetyBiosensorsFabrication methods

* Submitted to Trends in Food Science and Technol* Corresponding author.

E-mail address: xiaonan.lu@ubc.ca (X. Lu).1 Equal contribution as co-first author.

http://dx.doi.org/10.1016/j.tifs.2015.09.0050924-2244/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t

Background: Various foods are susceptible to contamination and adulteration with mycotoxins, pre-senting serious health risks to humans. Microfluidic “lab-on-a-chip” devices could integrate and mini-aturize versatile functions from sample preparation to detection, showing great potential in rapid,accurate, and high-throughput detection of mycotoxins.Scope and approach: This review focuses on the application of microfluidic “lab-on-a-chip” platforms todetect mycotoxins in foods. Fabrication processes and major components of microfluidic devices, as wellas separation and detection methods integrated with “lab-on-a-chip” systems are summarized anddiscussed. Finally, challenges and future research directions in the development of microfluidic devicesto detect mycotoxins are highlighted.Key findings and conclusions: Microfluidic “lab-on-a-chip” devices have a great potential for accurate andhigh-throughput detection of mycotoxins in agricultural and food products.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Mycotoxins are secondary metabolites of fungi and the majorfungal genera producing them include Aspergillus spp., Fusariumspp. and Penicillium spp. These molds produce various types ofmycotoxins, such as aflatoxins (AFs), deoxynivalenol (DON), zear-alenone (ZEA), fumonisin B1 (FB1), ochratoxin A (OTA) and citrinin(CIT), almost all of which are toxic to humans (Ar�evalo, Granero,Fern�andez, Raba, & Z�on, 2011; Zheng, Richard, & Binder, 2006).Representative mycotoxins widely identified in different foodmatrices are listed in Table S1 (Richard, 2007; Stoloff, 1976; vanEgmond, Schothorst, & Jonker, 2007). Mycotoxin contaminationcan occur throughout the entire food chain, from processing totransportation and storage (O'Brien & Dietrich, 2005). Besides,mycotoxin in feed could also lesion in animal origin food, exposingpotential high risks to consumers (Zain, 2011). For example, AFs arethe major mycotoxins which account for almost 93% of mycotoxincontamination in foodstuffs and beverage, resulting in carcinogenic

ogy.

cases in consumers (Petroczi, Nepusz, Taylor, & Naughton, 2011).Studies on AFs showed the LD50 for ducklings, rats and sheep were0.4, 1, and 500 mg/kg, respectively (Hussein & Brasel, 2001). OTA istoxic as nephrotoxic. Besides, due to possible occurrence of BalkanEndemic Nephropathy (a renal tumor), it is considered as carcin-ogen (Frenette et al., 2008; Pfohl-Leszkowicz, Petkova-Bocharova,Chernozemsky, & Castegnaro, 2002). In addition, ZEA has beenassociated with human cervical cancer (Shim, Dzantiev, Eremin, &Chung, 2009). Due to the potential carcinogenic, teratogenic, andmutagenic effects of mycotoxins as well as their wide existence inagricultural and food products, rapid, high-throughput andportable methods for sensitive detection are needed.

Conventional methods for the detection of mycotoxins in theenvironment and agricultural products are primarilychromatographic-based techniques, including thin-layer chroma-tography (TLC), high performance liquid chromatography (HPLC),gas chromatography coupled with mass spectrometry (GCeMS)(Lehotay & Haj�slov�a, 2002; Sforza, Dall'Asta, & Marchelli, 2006).However, all these methods require extensive sample preparationprocedures and they are time consuming and need highly trainedpersonnel. In addition, large amount of hazardous regents andsolvents are often required during analysis. Commercially availablemethods for the detection of mycotoxins are mainly

L. Guo et al. / Trends in Food Science & Technology 46 (2015) 252e263 253

immunological-based techniques, which are on the basis of specificinteraction between monoclonal and/or polyclonal antibodies andthe toxins. These techniques can be further divided into immu-noaffinity column (IAC)-based analysis and enzyme-linked immu-nosorbent assay (ELISA) (Magan & Olsen, 2004). Compared withchromatographic-based methods, immuno-based methods havehigher selectivity, but the high expense for antibody screening andpoor limit of detection inhibit the application of immuno-basedmethods. Due to the wide distribution of mycotoxins and analyt-ical complexity of food matrices, a rapid, in-field, high-throughput,and lab-independentmethod to recordmycotoxin contamination ishighly demanded. Further, the developed methods should behighly sensitive to meet the legislative LOD of mycotoxins in foods(van Egmond et al., 2007). Table S2 summarizes the regulations forsome typical mycotoxins that are presented in European Commu-nity (Egmond & Jonker, 2004). Taken together, a great effort hasbeen devoted to ultra-fast and ultra-accurate determination ofextremely low levels of mycotoxins in foods, and microfluidicsdevices have emerged as a promising alternative as modernanalytical platform.

2. Microfluidic device: major principles and components

The idea of microfluidic analytical platform derives from theconcept of Total Analysis System (TAS), which aims to shrink andintegrate all necessary steps for chemical analysis of a sample ontoa single device. The whole system mainly include types of drivingapparatus (e.g., pumps and reactors) and processes patterns (e.g.,sample preparation, filtration, dilution, reaction, and detection)(Connelly et al., 2012). While microfluidic analytical platform, alsoknown as Micro Total Analysis Systems (mTAS), further expands itsapplication, making the whole setup of a laboratory onto a singlechip in micro-meter level (Dittrich, Tachikawa, & Manz, 2006;Kovarik et al., 2013). As its name indicates, microfluidics dealswith controlling fluids of tiny amount (typically in nanoliters) inmicroscale channels (Squires & Quake, 2005). The characteristicchannel size of microfluidics analytical devices ranges from 10 mmto 200 mm and in some cases down to 1 mm (Bayraktar & Pidugu,2006). Fluid flow in these microchannels behaves quite differ-ently from those in macroscale channels. For example, surfaceforces are dominating over volume forces at the microscale. As oneof the most important forces affecting flow behaviors at themacroscale, the gravitational force in the fluid flow is usuallynegligible on a microfluidic platform (Mark, Haeberle, Roth, vonStetten, & Zengerle, 2010). Moreover, strong viscous forces usu-ally limit the flow to laminar regime, making molecular transport achallenging task in many microfluidics sensing experiments(Squires & Quake, 2005). Since chemical and biological analysisusually involves multiple steps of fluid manipulation (e.g., dilution,mixing, separation, aliquoting etc.), for these applications, micro-fluidics originated in these fields since 1980s. Indeed, the devel-opment of early microfluidics was mainly driven by molecularanalysis and molecular biology (Whitesides, 2006). Motivated byrapid development of biomedical and cell biology in the recentyears, microfluidics has witnessed extraordinary advancements inthe past decade. Many microfluidics branches have been estab-lished to utilize unique behaviors of fluidic flow in microscale, suchas droplet-based microfluidics (Seemann, Brinkmann, Pfohl, &Herminghaus, 2012; Teh, Lin, Huang, & Lee, 2008; Xu & Attinger,2008; Zhang, Betz, Qadeer, Attinger, & Chen, 2011), bubble-basedmicrofluidics (Ahmed et al., 2013; Ahmed, Mao, Shi, Juluri, &Huang, 2009; Chen & Lee, 2014; Hashmi et al., 2012; Xu et al.,2013), paper-based microfluidics (Martinez, Phillips, Whitesides,& Carrilho, 2009), and inertial microfluidics (Di Carlo, 2009).

Compared to traditional fluidic platform, microfluidic analytical

platforms provide tremendous advantages, such as low sample andreagent consumption, low fabrication costs, flexible design withmore functions, fast analysis and response time, high throughputscreening, precise process control (i.e. hydrodynamic parameterand temperature), and easy to carry which facilitate in-fielddetection (Atalay et al., 2011; Sackmann, Fulton, & Beebe, 2014).As mycotoxins often disperse from a small area to a wholefoodsystems and usually require on-site and in-field detection, theseadvantages of microfluidics devices fully cater the requirement formycotoxins detection.

2.1. Basic physics of microfluidics

The development of microfluidics depends on abroad of disci-plines, expanding from fluid mechanics, thermodynamics, elec-trostatics, chemistry, to material science (Bayraktar & Pidugu,2006; Squires & Quake, 2005; Stone, Stroock, & Ajdari, 2004). Inthe recent years, the interplay among microflow, microstructures(Chen et al., 2011; Chen, Lam, & Fu, 2012; Lam, Sun, Chen, & Fu,2012) and nanomaterials (Mao & Koser, 2006; Zhang & Wang,2013) have been extensively explored for “lab on a chip” (LOC)applications. Understanding of microfluidics physics facilitates thedesign of chips to realize different functions. Several dimensionlessnumbers are used to parameterize competitions among a variety offorces. (1) Reynolds number is the most important dimensionlessnumber in microfluidics, which determines the ratio between in-ertial and viscous forces. Since characteristic length of microfluidicsdevices is at microscale, Reynolds number is usually very small,making inertial forces irrelevant. (2) Peclet number relates toconvective and diffusive transport. If Peclet number is smaller thanunity, diffusion would dominate over convection and tracers influid would flow side-by-side, making fluid transport and targetdetection very slow and inefficient (Squires, Messinger, & Manalis,2008). (3) Capillary number relates to viscous and interfacial forcesand concerns multiphase flow in a microfluidics device. Surfacetension plays an important role in the dynamics of interfaces be-tween different fluids, creating all types of interesting phenomenain droplet-based or bubble-based microfluidics (Squires & Quake,2005). (4) Weber number is associated with inertial and interfa-cial forces, another critical parameter concerning multiphasemicrofluidics. For example, drops or bubbles will undergo sub-stantial deformation if the flow rate is high enough to induce highWeber number in a microfluidics device (Xu, Vaillant, & Attinger,2010).

Microfluidic devices are mainly composed of actuators andsensors. Due to different functions, actuators can be categorized asmicrovalves, micropumps, and micromixers, while sensors aremainlymolecular and cellular detectors (Beebe, Mensing,&Walker,2002). Microvalves are used for flow control in the microfluidicssystems for purposes, such as separation and timing. Basically,there are two kinds of microvalves, naming as active and passivemicrovavels. More specifically, according to different actuationforce, microvalves also can be classified as mechanical, pneumat-ical, electrokinetical, magnetical, or capillary microvalves (Oh &Ahn, 2006; Pan, McDonald, Kai, & Ziaie, 2005). One type of flowcontrol using active microvalves are driven by active pressure, andin order to achieve this, the chip structures will be more compli-cated compared to passive microvalves, which open to forwardpressure and derive flow. A representative micromechanical flapmicrovalve used to control biochemical reactions of two com-pounds is shown in Fig. S1. When the pressure outside the flapmicro-valve was higher than that inside channel, the valve wouldbe forced to open and then chemical reactions happen (Au, Lai,Utela, & Folch, 2011; Voldman, Voldman, Gray, & Schmidt, 2000).Although a variety of microvalves have been developed in the

L. Guo et al. / Trends in Food Science & Technology 46 (2015) 252e263254

recent years, the proper integration of microvalves into a specificmicrofluidics system for flexible liquid handling is still challenging.

The critical functions of micropumps in microfluidic device aretransporting flow and controlling flow rate. By transmission mode,micropumps can be classified into two categories, namely recip-rocating pumps and continuous flow pumps (Chen, Lee, Choo, &Lee, 2008). By actuation mechanism, micropumps can be namedas piezoelectric micropump, localized phase change micropump,electrowetting micropump, electrostatical micropump, electro-hydrodynamic micropump, electroosmotic micropump, magneto-hydrodynamic micropump and surface tension micropump (Laser& Santiago, 2004). Maximum flow rate, maximum differentialpressure generated and the overall package size are factors to beconsidered when selecting micropumps. Continuous flow micro-pumps are mostly used in microfluidic devices; they can generatenon-pulsatile flows which meet the requirement for chemicalanalysis and biological cultivation. Further, their structures are easyto fabricate (Pennathur, 2008). A schematic illustration of micro-pump and its operation are shown in Fig. S2. Themicropump can beoperated using the push-pins of a Braille display, which is a tactiledevice used by the blind to read computer message. The movablepins are controlled by computer and can act as valves. Threesequential valves can be used as a peristaltic pump. All the pumpsand valves can generate versatile fluid flow (Futai, Gu, Song, &Takayama, 2006; Gu et al., 2004). Although the material formicrofluidic device are high erosion resistant to acid or base,crystallization of high salt buffers may block microchannel or erodenon-plastic parts which subsequently cause leakage, it is necessaryto change pumps periodically (Seidel& Niessner, 2008). Due to tinyscale of microfluidic channel, Reynolds number and Peclet numberin a microfluidics device are low enough to keep flow at laminarregime, and molecular transport in flow is mainly via diffusionrather than convection. Although, low frequency molecular trans-port will benefit reagent to keep concentration during transmissionit results a low mixing efficiency and slow molecule transmission(Squires& Quake, 2005; Stone et al., 2004). That's main stimulationfor micromixers innovation in microfluidics area. Micromixers canbe classified into passive and active micromixers (Nguyen & Wu,2005). Passive micromixers utilize specific geometrical design ofmicrofluidics channels to improve mixing efficiency; the limitationis hard to control the position and time of the mixing events. Incontrast, the utilization of active micromixers have better controlover the mixing of reagents. But mixing force derives from extra-neous module which increase the difficulty for fabrication (Strobl,Schneider, Wixforth, Sritharan, & Guttenberg, 2006). In themicrofluidics platforms, the required channel for completing mix-ing depends upon flow velocity, diffusivity of the sample speciesand the size of the channel. Mixing time and efficiency should beconsidered when selecting micromixers (Hashmi and Xu, 2014).

The detectors are the most important component for micro-fluidic devices. The requirements for detectors are higher than thatof on conventional fluidic platform. In order to be integrated withmicrofluidic devices, the detector should be miniaturization.Further, since the amount of analytic target on microfluidic devicesare usually quite tiny, detector need to be ultrasensitive withrelatively low detection limits (Roman & Kennedy, 2007). Still avariety of conventional detection techniques can be implementedinto the microfluidics devices by simple integration. For example,mass spectrometry (MS), electrochemical and optical baseddetection techniques are widely used in microfluidic platform asreadout (Newman, Giordan, Copper, & Collins, 2008). Electro-chemical detection includes potentiometry, amperometry, andconductometry while optical detection includes fluorescence,chemilumimescence and absorbance (Atalay et al., 2011). Generally,due to the high surface-to-volume ratio at microscale, the smaller

the sensing element, the higher sensitivity the sensor can obtain.With the development of innovative sensing material, such asgraphene, carbon nanotubes (CNT), nanowires and nanoparticles, itis believed that the sensitivity of microfluidic devices will befurther improved.

3. Application of microfluidic devices in the detection ofmycotoxins

3.1. Fabrication material and process

Microfluidic devices are commonly fabricated using glass, sili-con, polymers and even paper based material (Martinez et al.,2009). Compared with silicon and glass, the feasibility ofpolymer-based materials enables the production of more sophis-ticated design of microfluidics systems that can be sealed thermallyor chemically (Zhou, Ellis, & Voelcker, 2010). In which, the mostwidely used is polydimethylsiloxane (PDMS) (Lim, Kouzani, &Duan, 2010). Compared with glass, PDMS has much lower ther-mal conductivity and higher hydrophobicity. Lower thermal con-ductivity can ensure a relatively low thermal noise during theprocess, which increases the accuracy of detection (Giannitsis &Min, 2010). Further, PDMS has good transparency and biocompat-ibility, making it an excellent substrate for optical detection andbiological cultivation. Soft lithographic techniques are frequentlyapplied for PDMS based chip fabrication (Unger, Chou, Thorsen,Scherer, & Quake, 2000; Xia & Whitesides, 1998). The soft litho-graphic techniques could mold a high resolution pattern on PDMSlayer. This PDMS layer can then be bonded to another material withspecific characteristics. The modification of hydrophobicity ofPDMS-based microfluidic devices makes its application wide.Several surface modification methods are usually utilized in thefabrication process, including plasma treatment, silanization,chemical vapor deposition (CVD), surfactant treatment, proteinadsorption and also the application of nanocomopsite materials(Zhou, Khodakov, Ellis, & Voelcker, 2012). Although less widelyused than PDMS, paper is an alternative biocompatible material forthe fabrication of microfluidic devices. Besides, due to its porosity,paper-based material has its own advantages over PDMS and othernonporous materials. Fig. 1 describes the fabrication of microfluidicdevices from omniphobic paper by embossing technique (Thuoet al., 2014). By compressing mold-A and mold-B together on thepaper with pressure (about 0.2 kg/cm2), “T”-shaped channels areembossed. This three-dimensional system of microchannel canwork as droplet generator and phase separator. The gas perme-ability of the paper also enables the contact and exchange of fluid inthe microchannel with its environment around. For example, thefluid with pH indicator in the microchannel could sense HCl or NH3gases in surrounding environment. However, enclosed paper basedmicrofluidic devices could eliminate evaporation and improvewicking rate. For example, Schilling, Lepore, Kurian, and Martinez(2012) used four layers of tone to seal microPADs. The plasticlayer in device could perfectly seal the channels and inhibit thediffusion of flow to other layers. Fig. 2 shows a typical fabricationprocess for a microfluidics device to detect aflatoxin B1 (AFB1) incorn extract solution (Hu, Deng, & Zou, 2013). To be specific, thedevice is fabricated by bonding transparent stencil onto a micro-scope glass slide. Then, the smectite-polyacrylamide (PAM) nano-composite has sensitive adsorption to AFB1 which is assembledonto glass surface as adsorbent agent. By removing the stencil andattaching PDMS flow layer which contains a pre-made serpentinechannel to the surface of the glass slide, the microfluidics sensordevice is completed. The short transit distances and times for flowtransmission in microfluidic system enable direct delivery of ana-lytes to sensing elements, decreasing non-specific bonding

Fig. 1. Fabrication of microfluidic devices from omniphobic paper by embossing technique (Thuo et al., 2014).

L. Guo et al. / Trends in Food Science & Technology 46 (2015) 252e263 255

between analytes and substrate which usually occur in macroscalesystems (deMello, 2006). After adsorbed by smectite-polyacrylamide (PAM) nano-composite, AFB1 would release

Fig. 2. Fabrication process of a representati

fluorescence signal. By detecting fluorescence intensity derivedfrom AFB1 and correlating the peaks information from sample withreference, the AFB1 concentration in the test solution was

ve microfluidics chip (Hu et al., 2013).

L. Guo et al. / Trends in Food Science & Technology 46 (2015) 252e263256

quantified (Hu, Garcia-Uribe, Deng, & Zou, 2012).

3.2. Separation methods: signal enrichment and collection

Mycotoxins are usually involved in complicated food matrix,separation is necessary before detection. Appropriate samplepreparation can increase detecting accuracy and recovery rate.Solvent extraction, liquideliquid extraction (LLE), solid-phaseextraction (SPE) and immunoaffinity SPE have already beenapplied for sample pretreatment before determination of myco-toxins by capture agents (Lattanzio, Solfrizzo, Powers, & Visconti,2007; Monbaliu et al., 2009).

Among these methods, solvent extraction is one of the mostfrequently used methods for the preliminary separation of myco-toxins from different food matrices. Based upon the inherentstructural features of mycotoxins (lipo-soluble or water-soluble),different solvents such as methanol, acetone and acetonitrile areused. Polyethylene (PEG) and phosphate buffer saline (PBS) arefrequently applied to improve the extraction efficiency. Then,capturing agents are employed to capture and consequently facil-itate complete separation of targeted mycotoxins. Currently, anti-body, aptamer, molecularly imprinted polymers (MIPs) are threepromising elements which can be integrated with microfluidic LOCdevices to conduct molecule-specific capture.

Antibodies are the only immunological based separation ele-ments have been applied into microfluidics devices. The advantageof antibodies based immunoassay is high specificity. Integration ofdifferent antibodies into a multi-channel microfluidic device canlead to quantitative analysis of mycotoxins in a high-throughputmanner. The typical microfluidics immunosensors used for thedetermination of mycotoxins are capillary electromigrationmicrochip (CE chip) and lateral flow test strip (LFTS) that are drivenby capillary electromigration and capillary force, respectively (P. Li,Zhang, et al., 2012; Li, Zhou, et al., 2012). For example, a competitiveimmunoassay-microfluidics device has been developed for theseparation and detection of ZEA. The schematic introduction isshown in Fig. 3 (Herv�as, L�opez,& Escarpa, 2011b). The electric fieldscan drive the fluids at different chambers and enable reactionsoccurred between different chambers. First, magnetic beads coatedwith protein G were used as immobilization substrate for thecapture of the anti-ZEA antibody. Subsequently, a competitiveimmunoassay took place in the immunological reaction chamber

Fig. 3. Layout of a microfluidics device and immunoassay for the separation and detection of(Herv�as et al., 2011b).

where ZEA and horse radish peroxidase-labeled derivatives (HRP-ZEA) competed for the binding sites in the magnetic beads thatcovered with protein. Then, enzymatic substrate was added andreacted with the derivatives in the enzymatic reaction chamber. Byevaluating the signal released by the electrochemical reaction be-tween enzyme and substrate, the activity of the enzyme tracer(HRP-ZEA) as well as the concentration of mycotoxin can bededuced.

MIPs is another efficient separation method that has been in-tegrated into the microfluidics LOC system. By applying interactionand conjugation among templates, functional monomers andcross-linking agents, MIPs can form unique cavities that containspecific and well-orientated functional groups complementary tothe structure of analyte molecules, resulting in highly selectivemolecular binding sites (Mosbach& Ramstrom,1996). The workingprinciple of MIPs (also termed as “artificial antibody”) and subse-quent disposable microfluidics biochip with on-chip molecularlyimprinted biosensors for the optical detection of anesthetic pro-pofol is shown in Fig. 4 (Hong, Chang, Lin, & Hong, 2010). SyntheticMIPs was deposited onto a chamber of the microfluidics chip,which can capture and separate targeted analytes within the MIPs-microfluidics chip. The application of MIPs-microfluidics LOC de-vice in the separation of biomolecules from complicated matrices isshown in Fig. S3 (Huang et al., 2006). The availability of differentMIPs films in different regions enabled separation and detection ofmultiple molecules simultaneously. The integration of “spider-web” micropumps and microvalves facilitated automated sampleintroduction process and precisely controlled flow rates insidemicro-channels, which was essential to determine the absorptionamount of the molecules in the MIPs chips (Huang et al., 2006).Another research group (Weng, Yeh, Ho, & Lee, 2007) developedMIPs films toward morphine (template molecule) and used elec-trochemical reaction to detect morphine in clinical samples. MIPsfilm as separation elements were placed in the upstream of themicrochannel to avoid the interference of ascorbic acid in humanblood. Moreover, the integration of micropumps andmicrovalves inthe microfluidics LOC device enabled the automation of sampleinjection and continuous measurement of morphine. The pumpingrate was increased to largely reduce the sample transportationtime, resulting in rapid and precise detection of morphine in blood.

Aptamers are single-stranded oligonucleotides can recognizeand bind to target molecules by folding into a unique secondary or

zearalenone (IRC: immunological reaction chamber; ERC: enzymatic reaction chamber)

Fig. 4. Schematic illustration of the synthesis of molecularly imprinted polymers (MIPs) and its application for molecular sorting; (a) the integration of microfluidic system; (b)structure of microfluidic chip; (c) the principle of MIPs for molecular sorting (Hong et al., 2010).

Fig. 5. Scheme of capturing targeted analyte by aptamers integrated in the micro-channels (Sheng et al., 2012).

L. Guo et al. / Trends in Food Science & Technology 46 (2015) 252e263 257

tertiary structure (Lou et al., 2009). Fig. 5 illustrates capture andselection of cancer cells by aptamers in a microfluidics LOC device(Sheng et al., 2012). To be specific, avidin was plated onto the glasssurface as aptamers substrate and then biotinylated aptamers wereimmobilized via biotin-avidin interaction onto avidin. Then, thetarget cancer cells were captured due to the specific binding be-tween cell surface receptors and aptamers. Xu et al. (2009) inte-grated aptamers in a microfluidic device for the detection ofmultiple cancer cells from blood samples. The aptamer-basedmicrofluidics device enabled a 135-fold enrichment of cells in asingle run as the height of microchannels was on the order of a celldiameter. This device required no sample pretreatment, subse-quently eliminating artificial influence from pretreatment processand enhancing the efficiency of cell selection. Besides, it can couplewith various detection readouts, largely expanding its applicationpossibility. A research group (He et al., 2011) utilized aptamer-based surface-enhanced Raman spectroscopy (SERS) in the detec-tion of ricin in liquid foods. The aptamer was conjugated onto thesurface of silver dendrite. After capturing the ricin B chain from

L. Guo et al. / Trends in Food Science & Technology 46 (2015) 252e263258

food matrices, the sliver dendrite below the aptamer will amplifyRaman signal which improve the limit of detection. The comparisonamong antibody, MIPs and aptamer in terms of separation speed,sensitivity, specificity, stability, reusability and cost efficiency asseparation elements integrated into microfluidics devices aresummarized in Table 1. Although the aforementioned separationmethods have not yet been widely used in the recognition andseparation of mycotoxins, these techniques have been used for theseparation and enrichment of other food chemical hazards, such aspesticides, herbicides and antibiotic residues. Therefore, these el-ements have the potential to be successfully integrated into themicrofluidics LOC platform for a complete separation of mycotoxinsfrom foods before a reliable detection.

3.3. Detection methods: signal transduction and detection

Current methods integrated into microfluidics devices for thedetection of mycotoxins are mainly optical-based methods,electrochemical-based methods as well as label-free methods.Among different optical methods, fluorescence has been widelyused in the detection of mycotoxins due to its excellent sensitivity,selectivity and applicability in micro- or nano-scale. A researchgroup detected CIT in rice samples using a microfluidics devicecoupled with both fluorescence and electrochemical methods(Ar�evalo et al., 2011). As CIT has a planar conjugated chemicalstructure and natural fluorescence, it is feasible to quantify CITusing a fluorometer. The results obtained from fluorescencedetection can bewell correlated with the result from amperometricimmunosensor, validating these two methods for CIT detection.Another research group (Anfossi et al., 2010) designed a lateral flowimmunoassay test strip to quantify fumonisins in maize products.By measuring the color intensity derived from test strip, quantifi-cation of fumonisins in samples could be achieved. As chem-iluminescence detection does not require either the external lightsources or the excitation of fluorescent labels, it can significantlyreduce the complexity of integrated optical components, leading tomore portable and convenient devices than fluorescence micro-arrays to some extent (Novo et al., 2011; Novo, Moulas, Prazeres,Chu, & Conde, 2013; Roda et al., 2006; Soares et al., 2014; Yang,Sun, Kostov, & Rasooly, 2011). In addition, little background signalalso contributed to relatively high sensitivity of chemiluminescencemicroarrays. A research team (Sauceda-Friebe et al., 2011) devel-oped a fully automated flow through device with chem-iluminescence readout system to detect OTA in green coffee extract.Peptide-linked OTA and biotin conjugates were synthesized andimmobilized in an array of 4� 6microspots and used for an indirectcompetitive immunoassay. By using chemiluminescence detectionsystem, only a CCD camera was required for signal collection. Theminiaturization and automation of devices enabled the high con-jugate coupling density and a flow-through format, largelyincreasing mass transport and reducing detection time.

Electrochemical-based detection method has its own

Table 1Comparison of antibody, molecularly imprinted polymers (MIPs) and aptamer usedas separation elements integrated into microfluidics devices.

Antibody MIPs Aptamer

Separation speed þ þ* þ þ þ þ þ þSensitivity þ þ þ þ þ þSpecificity þ þ þ þ þ þ þ þ þStability þ þ þ þ þ þ þ þReusability þ þ þ þ þ þ þCost efficiency þ þ þ þ þ þ

*Note: ‘þ’ stands for the capacity.

advantages because its response is not limited by optical pathlength or sample turbidity (Hervas, Lopez, & Escarpa, 2012). Be-sides, its inherent facility for miniaturization without loss of per-formance, high compatibility and sensitivity together makeselectrochemical-based detection method an excellent techniqueto be incorporated into the microfluidics LOC devices (Wang et al.,2008; Yeh, Chen, Lin, Lin, & Chang, 2009). Compared to opticalmicroarray readout systems, electrochemical microarray readoutsystems are more applicable for simultaneous detection of multipleanalytes. One research group developed an electrochemical basedassay in a 96-well screening printed microplate to detect aflatoxinB1 (AFB1) from corn samples. Alkaline phosphatase was used as thelabel enzyme to detect the toxins in several different samples,enabling the multichannel read-out to be performed simulta-neously (Piermarini, Micheli, Ammida, Palleschi,&Moscone, 2007).Recently, a research team (Herv�as, L�opez, & Escarpa, 2011a) suc-cessfully detected ZEN in infant food products by electrochemicaldetector. They integrated immunoassay into double-T layoutmicrochannels and both channels were used as immunological andenzymatic reaction chambers, respectively. The application ofelectrokinetic motion could easily manipulate the fluids in differentdirections within the microchannels as well as stop flows. Bymeasuring the responsive current, the amount of ZEN could bedetermined. Parker and his colleagues (Parker, Lanyon, Manning,Arrigan, & Tothill, 2009) also developed an electrochemical basedmicroarray for the detection of aflatoxin M1 (AFM1) in milk. HRPwas used as the enzyme label and the tetramethylbenzidine (TMB)(please double check tetramethylbenzidine is the right full name)/H2O2 electrochemical detection system could achieve a detectionlimit of 8 ng L�1. The use of gold microelectrodes in this studyenhanced the mass transport and hemispherical diffusion layersformed at such electrodes provided a steady cyclic voltammograms,resulting in improved response time, greater sensitivity and signal-to-noise ratio.

Label-free method, such as MS analysis combined with micro-fluidics immunosensor, is another method for the development ofmycotoxin assays. As MS detection system is attractive for its highspeed and sensitivity, the combination of this instrument withmicro-fabricated devices can achieve a low limit of detection. Jiangand his colleagues (Jiang, Wang, Locascio, & Lee, 2001) establisheda plastic microfluidics immunosensor for the detection of AFB1. Bymanipulating the molar binding ratios of AFB1 antibody to eachaflatoxins, AFB1 was captured and concentrated by the onlinecoupling of a specific affinity zone, and was then directly identifiedby electrospray ionization mass spectrometry (ESI-MS). Liu andcoauthors (Liu, Lin, Chan, Lin,& Fuh, 2013) investigated a chip-nanoliquid chromatography interface/triple quadrupole MS (chip-nanoLC/QqQ-MS) system to determine aflatoxins in peanut prod-ucts. Solvent extraction followed by immunoaffinity SPE samplepretreatment enhanced the sensitivity for the quantification of AFs,while the utilization of two column design chip-based LC tech-niques reduced matrix interference and online sample pre-concentration, enabling rapid and simultaneous determination ofindividual AFs. As another label-free and non-destructive detectiontool, Raman spectroscopy can also be applied for the detection offood chemical hazards (Knauer, Ivleva, Liu, Niessner, & Haisch,2010). A SERS platform embedded into a microfluidics channel forthe determination of OTA contamination was studied (Galarreta,Tabatabaei, Guieu, Peyrin, & Lagugn�e-Labarthet, 2013). The re-searchers used aptamer as the capturing agent and SERS could beapplied to sense the conformational change of oligonucleotide ofthe aptamer sequence before and after its interaction with OTA.Faint Raman signal can be significantly enhanced by surface plas-mon resonance derived from the interaction of incident laser andthe patterned metallic nanostructures, and a high signal-to-noise

L. Guo et al. / Trends in Food Science & Technology 46 (2015) 252e263 259

ratio of the enhanced Raman spectra could be obtained. Thefabrication of a SERS platform on glass layer and the assembly of aPDMS layer are shown in Fig. S4 (Galarreta, Hart�e, Marquestaut,Norton, & Lagugn�e-Labarthet, 2010). First, the SERS platform wasinscribed onto the glass layer by electron beam lithography. PDMSlayer coated with aluminum was then aligned at and bonded toglass layer under pressure at 90 �C. Thus, SERS substrate was in-tegrated into the microfluidics channel.

Recent studies of microfluidics-based analytical tools for thedetection of mycotoxins have emphasized the combinational use ofmicrofluidics devices with other modern technologies such asnanotechnology and immunoassays. One of the most advantageousfeatures over the traditional immunoassays performed in micro-wells is the reduced diffusion distance due to the high surface-to-volume ratio of the miniaturized microfluidics devices, whichsignificantly enhances analytical sensitivity and reduces analyticaltime (Hervas et al., 2012). One research group developed a micro-fluidics immunosensor combined with integrated micro-fabricatedhydrogenated amorphous silicon photodiodes andchemiluminescence-based indirect competitive ELISA for thequantification of OTA in red wine and beer (Novo et al., 2013). Theutilization of a two-channel U-shaped microfluidics device for thesimultaneous analysis of a reference solution and a solution con-taining OTA reduced measurement errors and thus improved thelimit of detection of OTA. A typical immunochromatographic teststrip was developed for the detection of FBs in food samples, asshown in Fig. S5 (Li, Zhang, et al., 2012; Li, Zhou, et al., 2012). Oncethe strip was dipped into the sample solution, the goldnanoparticle-monoclonal antibody (McAb) probe would be dis-solved and flowed along with sample solution. The analyte insamples competed with immobilized FB1 to bind to gold-McAbprobe. As the mobile complex (gold-McAb probe-FB1) can becaptured by the secondary antibody (control line) but cannot becaptured by chicken egg ovalbumin (OVA)-FB1, the analyte insamples can be visually analyzed by the degree of color density oftest line, which was negatively correlated to the concentration ofFB1. Table 2 summarizes the recent application of microfluidicsdevices for the detection of mycotoxins in different agricultural andfood products.

4. Challenges and further efforts

Microfluidics LOC is a multidisciplinary field that integrates thestudy of micro-/nano-technology, physics, chemistry and engi-neering. Apart from a few publications on separation based on-chipanalysis, microfluidics-based analytical devices are not yet exten-sively applied in the detection of mycotoxins. Challenges remainedand further effort is required for the development of microfluidicsanalytical devices to determine mycotoxins in agricultural and foodproducts.

One of the major challenges is the complexity of food matrices(Asensio-Ramos, Hern�andez-Borges, Rocco, & Fanali, 2009). Due tothe random distribution of mycotoxins in foods and potential co-existence of several types of mycotoxins in the same food prod-uct, the accuracy of determination could be a challenge. Further,microfluidic systems can be easily blocked by small particles orvapor bubbles. Thus, adequate sample pretreatment is essential foraccurate detection and minimized error (Künnemeyer, Revermann,Karst, & G€otz, 2008). Currently, almost all the microfluidics devicesfor the detection of mycotoxins are conducted with off-chip samplepretreatment. Further efforts are desired for the development ofon-chip sample pretreatment.

To get a better understanding about properties of microfluidics,such as surface tension and electrostatic forces conducted at themicro-scale, is another challenge (Squires & Quake, 2005). For

example, the limit of detection will be affected by surface proper-ties of microfluidic substrate material. Although the high surface-to-volume ratio of microfluidics devices has its advantages, itmay also lead to technical challenges. Specifically, the effect of evenlittle contamination to devices will be significantly higher than thatof on a macro-scale platform, leading to unreliable detection result.Even worse, analyte molecules may be absorbed onto the wallsbefore reaching detection area. Three feasible methods have beenused to overcome this technique barrier, including the develop-ment of new materials (e.g. polyethylene glycol) to create non-fouling surfaces, using an “inert” protein (e.g. albumin or casein)to block the surfaces, and flooding the device before sample anal-ysis (Folch, 2012).

Future research work should focus on the improvement of thesensitivity of detection and development of self-calibration assaywhich will expand the application of microfluidic platform (Novoet al., 2011). In the meanwhile, the development of automaticmicrofluidics devices will be a future direction which can beapplied in industry for the analysis of mycotoxins. Programcontrolled operation will eliminate the interference from artificialoperation, increasing the reliability of analysis result (Hervas et al.,2012).

Integrating other separation and detection techniques into themicrofluidics systems is also a future research direction. For in-stances, nanotechnology, such as nanomaterials and nano-structures, in the microfluidics systems can be employed toimprove analytical performance (Atalay et al., 2011). Based uponnanotechnology, innovative labels (e.g. quantum dots) may improvethe sensitivity of assay, allowing for detecting ultra-trace level ofmycotoxins (Li, Zhang, et al., 2012; Li, Zhou, et al., 2012). As fluo-rescence detection is a simple method in terms of excitation,detection schemes as well as its suitability for low volume samples,it is still a promising candidate to be integrated into the micro-fluidics platform to assist in the detection of mycotoxins (Baker,Duong, Grimley, & Roper, 2009). Electro-analytical instrumenta-tion, such as electronic readout instruments, can also be applied tothe microfluidics systems for the detection of mycotoxins. Recently,two promising trends in fluorescence detection are with the aid ofsmart-phone camera and miniaturized electrochemical detectionrunning on portable batteries (Xu, 2014).

From biological perspective, cell based assay show great po-tential to combine with microfluidic system for the detection ofmycotoxins. Cell based assay utilize living microorganisms (pro-karyotic or eukaryotic cell) as biological sensor (Liu et al., 2014). Bydetermining response of microorganisms to stimulation, such asfluorescence generation or quench, the amount of mycotoxins canbe determined (Bouaziz et al., 2013; Stocker et al., 2003). Besides,cell based assay only respond to active toxins rather than invalidones, which could reflect the actual toxicity of a contaminatedproduct (Hong, Young, Tepp, Johnson, & Beebe, 2013). The majoradvantage of cell based assay is that the response of biosensor tothe mycotoxins is recorded that can be correlated to the physio-logical performance of the victims when the amount of mycotoxinsis evaluated. This information can provide insightful prediction intotoxic effect, which in turn contributes to not only detection but alsotherapeutic method for vulnerable individuals afflicted to myco-toxin contamination (Liu et al., 2014). The characteristics ofmicrofluidic systems can be perfectly compatible to the cell basedsensors. The biocompatible materials for chip fabrication andcontinuous flow distribution provide suitable environment for cellcultivation. Recently, several groups successfully incorporated cellcultivation into microfluidic environment to mimic organismfunction (Lee, Oh, & Park, 2015; Witters et al., 2011). Coupled withoptical detection, researchers have also observed physiologicalresponse of encapsulated cells to stimulations (Itle, Zguris, &

Table 2Representative application of microfluidics lab-on-a-chip devices in the detection of mycotoxins.

Mycotoxin types Target analyte/food matrix Sample pretreatment Capture agents Detection devices Characteristics ofdevice and analysis

Reference

Aflatoxins(including AFB1,AFB2, AFG1, andAFG2)

Peanuts,peanut powder,peanut butter

Solvent extraction(MeOH) andimmunoaffinity solid-phase extraction (SPE)

Chip-based nanoliquid chromatograph(LC)

Triple quadrupole MSsystem

Linear range:0.048e16 ng/g;LOD: 0.004e0.008 ng/g;Recovery:90.8%e100.4%.

(Liu et al., 2013)

Zearalenone(ZEN)

Infant foods Solvent extraction in anultrasonic bath

A competitive enzyme-linked immunosorbentassay (ELISA)

Electrochemicaldetection

LOD: 0.4 mg/L;Recovery: 103% for solidsamples and 101% for liquidsamples;Immunoassay time:<15 min.

(Herv�as et al.,2011b)

Citrininn (CIT) Rice Solvent extraction(ACN, aqueous solutionof KCl)

A competitive ELISA Microfluidicselectrochemicaldetection (usingamperometricmeasurements)

LOD: 0.1 ng/mL;LOQ: 0.5 ng/mL;Detection time:<2 min;Total assay time: <45 min.

(Ar�evalo et al.,2011)

Ochratoxin A(OTA)

Red wine; beer Liquideliquidextraction

A competitive ELISA Chemiluminescencedetection

LOD (ng/mL):0.85 in pure PBS, 0.1 in beerand 2 in red wine.

(Novo et al.,2013)

Ochratoxin A(OTA)

Red wine; beer No samplepretreatment

An indirect ELISA Chemiluminescencedetection

LOD (ng/mL): 0.5 in wineand beer

(Novo et al.,2012)

Ochratoxin A(OTA)

White wine No samplepretreatment

An indirect competitiveELISA

Chemiluminescencedetection

LOD (ng/mL): 0.5 in puresolutions and 1 in whitewine

(Novo et al.,2011)

Ochratoxin A(OTA)

Green coffee Solvent extraction(methanol/aqueoussodium bicarbonatesolution).

An indirect competitiveELISA

Chemiluminescencedetection integrated ina regenerable glassmicrofluidicsimmunosensor

LOQ: 7 mg/kg in green coffeeextract;Analysis time: 12 min;allow for at least 20 assay-regeneration cycles of thebiochip surface

(Sauceda-Friebeet al., 2011)

Fumonisin B(includes FB1

and FB2)

Maize Aqueous extractionbuffer

Lateral flowimmunoassay

Optical method (colorintensity)

LOD: 120 mg/L (Anfossi et al.,2010)

Fumonisin B(includes FB1,FB2 and FB3)

Maize Solvent extraction(methanol/water)

A paper-basedcompetitive ELISA teststrip

Optical method (colorintensity)

LOD: 2.5 ng/mL;Total immunoassayanalytical time: <15 min.

(Li, Zhang, et al.,2012; Li, Zhou,et al., 2012)

Zearalenone(ZEN)

Corn Solvent extraction(PBS)

A monoclonal antibodybased gold nanoparticleimmune-chromatographic assay

Optical method (colorintensity)

LOD: 2.5 ng/mL and 30 mg/kg for the standard solutionand spike samples;Immunoassay analyticaltime: <15 min.

(Shim et al.,2009)

Aflatoxin M1

(AFM1)Milk Centrifugation of milk

samples (no extraction)An indirect competitiveELISA

Electrochemicaldetection

LOD: 8 ng/L (Parker et al.,2009)

Ochratoxin A(OTA)

/ No samplepretreatment

Aptamers Surface enhancedRaman spectroscopy(SERS)

Successfully detected2.5 mMOTA

(Galarreta et al.,2013)

L. Guo et al. / Trends in Food Science & Technology 46 (2015) 252e263260

Pishko, 2004). Since the signal from cell derives from toxicresponse, translation of signal to suitable quantitative readout willbe critical to cell based microfluidic system.

Combination of microfluidics technology with immunoassayswill still be a trend in the detection of mycotoxins. Paper-basedmicrofluidics immunosensors and regenerable microfluidicsimmunosensors will definitely be popular in the determination ofmycotoxins (Fu et al., 2011; Ge et al., 2013; Lu, Lin, & Qin, 2011).Based upon these techniques, multiplex analytical methods can bedeveloped to simultaneously detect different targeted mycotoxins(Sauceda-Friebe et al., 2011). Multiplex analytical method is criticalto the analysis of mycotoxins because several types of mycotoxinsmay exist in the same food product (Ngundi, Shriver-Lake, Moore,Ligler, & Taitt, 2006; Tothill & Saeger, 2011). For example, barcodemicrofluidics chips were developed as multiplex analytical plat-form (Fan et al., 2008). Ngundi et al. (2005; 2006) developed a rapidcompetitive immunoassay-based array biosensor to simulta-neously detect OTA and DON in different types of foods without anyclean-up procedures or additional preparation after simple sampleextraction. Integrating fluidic systems and multiplex analyticalparameters are essential to the acceptance of analytical microarrays

in the detection of mycotoxins (Seidel & Niessner, 2008). Takentogether, the development of miniaturized, portable and inexpen-sive on-chip microfluidics systems for the detection of mycotoxin isstill in progression.

In conclusion, microfluidics LOC platform provides an inno-vative method to convert complicated traditional methods intosimplified and efficient micro-scale devices. The feature ofportability enables the development of in-field devices. Fullintegration of all the elements onto the disposable and/or reus-able microfluidics systems is the ultimate goal for commercialapplication in the detection of mycotoxins in agricultural and foodproducts. Great efforts are also requested to develop the devices tobe user-friendly, reducing potential risk to both operators and theenvironment (Herv�as et al., 2011a; Li, Zhang, et al., 2012; Li, Zhou,et al., 2012).

Acknowledgment

This study was supported by funds awarded to X.L. by NaturalSciences and Engineering Research Council of Canada( NSERCRGPIN-2014-05487, NSERC EGP-485163-15, NSERC EGP-479494-15,

L. Guo et al. / Trends in Food Science & Technology 46 (2015) 252e263 261

NSERC EGP-467970-14, and NSERC EGP-463727-14).

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.tifs.2015.09.005.

References

Ahmed, D., Chan, C. Y., Lin, S. S., Muddana, H. S., Nama, N., Benkovic, S. J., et al.(2013). Tunable, pulsatile chemical gradient generation via acoustically drivenoscillating bubbles. Lab on a Chip, 13, 328e331.

Ahmed, D., Mao, X., Shi, J., Juluri, B. K., & Huang, T. J. (2009). A millisecond micro-mixer via single- bubble-based acoustic streaming. Lab on a Chip, 9, 2738e2741.

Anfossi, L., Calderara, M., Baggiani, C., Giovannoli, C., Arletti, E., & Giraudi, G. (2010).Development and application of a quantitative lateral flow immunoassay forfumonisins in maize. Analytica Chimica Acta, 682, 104e109.

Ar�evalo, F. J., Granero, A. M., Fern�andez, H., Raba, J., & Z�on, M. A. (2011). Citrinin (CIT)determination in rice samples using a micro fluidic electrochemical immuno-sensor. Talanta, 83, 966e973.

Asensio-Ramos, M., Hern�andez-Borges, J., Rocco, A., & Fanali, S. (2009). Food anal-ysis: a continuous challenge for miniaturized separation techniques. Journal ofSeparation Science, 32, 3764e3800.

Atalay, Y. T., Vermeir, S., Witters, D., Vergauwe, N., Verbruggen, B., Verboven, P., et al.(2011). Microfluidic analytical systems for food analysis. Trends in Food Science& Technology, 22, 386e404.

Au, A. K., Lai, H. Y., Utela, B. R., & Folch, A. (2011). Microvalves and micropumps forBioMEMS. Micromachines, 2, 179e220.

Baker, C. A., Duong, C. T., Grimley, A., & Roper, M. G. (2009). Recent advances inmicrofluidic detection systems. Bioanalysis, 1, 967e975.

Bayraktar, T., & Pidugu, S. B. (2006). Characterization of liquid flows in microfluidicsystems. International Journal of Heat and Mass Transfer, 49, 815e824.

Beebe, D. J., Mensing, G. A., & Walker, G. M. (2002). Physics and application ofmicrofluidics in biology. Annual Review of Biomedical Engineering, 4, 261e286.

Bouaziz, C., Bouslimi, A., Kadri, R., Zaied, C., Bacha, H., & Abid-Essefi, S. (2013). Thein vitro effects of zearalenone and T-2 toxins on vero cells. Experimental andToxicologic Pathology, 65, 497e501.

Chen, W., Lam, R. H. W., & Fu, J. (2012). Photolithographic surface micromachiningof polydimethylsiloxane (PDMS). Lab on a Chip, 12, 391e395.

Chen, Y., & Lee, S. (2014). Manipulation of biological objects using acoustic bubbles:a review. Integrative and Comparative Biology, 1e10.

Chen, L., Lee, S., Choo, J., & Lee, E. K. (2008). Continuous dynamic flow micropumpsfor microfluid manipulation. Journal of Micromechanics and Microengineering,18, 013001.

Chen, X., Wu, J., Ma, R., Hua, M., Koratkar, N., Yao, S., et al. (2011). Nanograssedmicropyramidal architectures for continuous dropwise condensation. AdvancedFunctional Materials, 21, 4617e4623.

Connelly, J. T., Kondapalli, S., Skoupi, M., Parker, J. S., Kirby, B. J., & Baeumner, A. J.(2012). Micro-total analysis system for virus detection: microfluidic pre-concentration coupled to liposome-based detection. Analytical and Bio-analytical Chemistry, 402, 315e323.

deMello, A. J. (2006). Control and detection of chemical reactions in microfluidicsystems. Nature, 442, 394e402.

Di Carlo, D. (2009). Inertial microfluidics. Lab on a Chip, 9, 3038e3046.Dittrich, P. S., Tachikawa, K., & Manz, A. (2006). Micro total analysis systems. Latest

advancements and trends. Analytical Chemistry, 78, 3887e3907.Egmond, H. P. v., & Jonker, M. A. (2004). Worldwide regulations for mycotoxins in

food and feed in 2003. Food and Agriculture Organization of the United Nations,81.

Fan, R., Vermesh, O., Srivastava, A., Yen, B. K. H., Qin, L., Ahmad, H., et al. (2008).Integrated barcode chips for rapid, multiplexed analysis of proteins in micro-liter quantities of blood. Nature Biotechnology, 26, 1373e1378.

Folch, A. (2012). Introduction to bioMEMS. CRC Press.Frenette, C., Paugh, R. J., Tozlovanu, M., Juzio, M., Pfohl-Leszkowicz, A., &

Manderville, R. A. (2008). Structure-activity relationships for the fluorescenceof ochratoxin A: insight for detection of ochratoxin A metabolites. AnalyticaChimica Acta, 617, 153e161.

Fu, Z., Shao, G., Wang, J., Lu, D., Wang, W., & Lin, Y. (2011). Microfabricatedrenewable beads-trapping/releasing flow cell for rapid antigen-antibody reac-tion in chemiluminescent immunoassay. Analytical Chemistry, 83, 2685.

Futai, N., Gu, W., Song, J. W., & Takayama, S. (2006). Handheld recirculation systemand customized media for microfluidic cell culture. Lab on a Chip, 6, 149e154.

Galarreta, B. C., Hart�e, E., Marquestaut, N., Norton, P. R., & Lagugn�e-Labarthet, F.(2010). Plasmonic properties of Fischer's patterns: polarization effects. PhysicalChemistry Chemical Physics: PCCP, 12, 6810.

Galarreta, B. C., Tabatabaei, M., Guieu, V., Peyrin, E., & Lagugn�e-Labarthet, F. (2013).Microfluidic channel with embedded SERS 2D platform for the aptamerdetection of ochratoxin A. Analytical and Bioanaltical Chemistry, 405, 1613e1621.

Ge, S., Ge, L., Yan, M., Song, X., Yu, J., & Liu, S. (2013). A disposable immunosensordevice for point-of-care test of tumor marker based on copper-mediatedamplification. Biosensors & Bioelectronics, 43, 425.

Giannitsis, A. T., & Min, M. (2010). Fabrication methods for microfluidic lab-on-

chips. In Electronics conference (BEC), 2010 12th Biennial Baltic (pp. 69e72).Gu, W., Zhu, X., Futai, N., Cho, B. S., Takayama, S., & Whitesides, G. M. (2004).

Computerized microfluidic cell culture using elastomeric channels and brailledisplays. Proceedings of the National Academy of Sciences of the United States ofAmerica, 101, 15861e15866.

Hashmi, A., Heiman, G., Yu, G., Lewis, M., Kwon, H. J., & Xu, Jie (2012). Oscillatingbubbles in teardrop cavities for microflow control. Microfluidics and Nano-fluidics, 14, 591e596.

Hashmi, A., & Xu, Jie (2014). On the quantification of mixing in microfluidics. Journalof Laboratory Automation, 19, 488e491.

He, L. L., Lamont, E., Veeregowda, B., Sreevatsan, S., Haynes, C. L., Diez-Gonzalez, F.,et al. (2011). Aptamer-based surface-enhanced Raman scattering detection ofricin in liquid foods. Chemical Science, 2, 1579e1582.

Herv�as, M., L�opez, M. A., & Escarpa, A. (2011a). Integrated electrokinetic magneticbead-based electrochemical immunoassay on microfluidic chips for reliablecontrol of permitted levels of zearalenone in infant foods. The Analyst, 136, 2131.

Herv�as, M., L�opez, M. A., & Escarpa, A. (2011b). Integrated electrokinetic magneticbead-based electrochemical immunoassay on microfluidic chips for reliablecontrol of permitted levels of zearalenone in infant foods. The Analyst, 136,2131e2138.

Hervas, M., Lopez, M. A., & Escarpa, A. (2012). Electrochemical immunosensing onboard microfluidic chip platforms. Trends in Analytical Chemistry, 31, 109.

Hong, C.-C., Chang, P.-H., Lin, C.-C., & Hong, C.-L. (2010). A disposable microfluidicbiochip with on-chip molecularly imprinted biosensors for optical detection ofanesthetic propofol. Biosensors and Bioelectronics, 25, 2058e2064.

Hong, W. S., Young, E. W., Tepp, W. H., Johnson, E. A., & Beebe, D. J. (2013).A microscale neuron and Schwann cell coculture model for increasing detectionsensitivity of botulinum neurotoxin type A. Toxicological Sciences, 134(1), 64e72.

Huang, S.-C., Lee, G.-B., Chien, F.-C., Chen, S.-J., Chen, W.-J., & Yang, M.-C. (2006).A microfluidic system with integrated molecular imprinting polymer films forsurface plasmon resonance detection. Journal of Micromechanics and Micro-engineering, 16, 1251e1257.

Hu, H., Deng, Y., & Zou, J. (2013). Microfluidic smectite-polymer nanocompositestrip sensor for aflatoxin detection. IEEE Sensors Journal, 13, 1835e1839.

Hu, H., Garcia-Uribe, A., Deng, Y., & Zou, J. (2012). Layer-by-layer assembledsmectite-polymer nanocomposite film for rapid fluorometric detection ofaflatoxin B. Sensors and Actuators B: Chemical, 166, 205e211.

Hussein, H. S., & Brasel, J. M. (2001). Toxicity, metabolism, and impact of mycotoxinson humans and animals. Toxicology, 167, 101e134.

Itle, L. J., Zguris, J. C., & Pishko, M. V. (2004). Cell-based bioassays in microfluidicsystems. Optics East, International Society for Optics and Photonics, 9e18.

Jiang, Y., Wang, P. C., Locascio, L. E., & Lee, C. S. (2001). Integrated plastic microfluidicdevices with ESI-MS for drug screening and residue analysis. Analytical Chem-istry, 73, 2048e2053.

Knauer, M., Ivleva, N. P., Liu, X., Niessner, R., & Haisch, C. (2010). Surface-enhancedRaman scattering-based label-free microarray readout for the detection ofmicroorganisms. Analytical Chemistry, 82, 2766.

Kovarik, M. L., Ornoff, D. M., Melvin, A. T., Dobes, N. C., Wang, Y., Dickinson, A. J.,et al. (2013). Micro total analysis systems: fundamental advances and applica-tions in the laboratory, clinic, and field. Analytical Chemistry, 85, 451e472.

Künnemeyer, J., Revermann, T., Karst, U., & G€otz, S. (2008). Quantitative analysis bymicrochip capillary electrophoresis - current limitations and problem-solvingstrategies. The Analyst, 133, 167.

Lam, R. H. W., Sun, Y., Chen, W., & Fu, J. (2012). Elastomeric microposts integratedinto microfluidics for flow-mediated endothelial mechanotransduction anal-ysis. Lab on a Chip, 12, 1865e1873.

Laser, D. J., & Santiago, J. G. (2004). A review of micropumps. Journal of Micro-mechanics and Microengineering, 14, 35e64.

Lattanzio, V. M. T., Solfrizzo, M., Powers, S., & Visconti, A. (2007). Simultaneousdetermination of aflatoxins, ochratoxin A and Fusarium toxins in maize byliquid chromatography/tandem mass spectrometry after multitoxin immu-noaffinity cleanup. Rapid Communications in Mass Spectrometry, 21, 3253e3261.

Lee, S. H., Oh, E. H., & Park, T. H. (2015). Cell-based microfluidic platform formimicking human olfactory system. Biosensors and Bioelectronics. http://dx.doi.org/10.1016/j.bios.2015.06.072.

Lehotay, S. J., & Haj�slov�a, J. (2002). Application of gas chromatography in foodanalysis. Trends in Analytical Chemistry, 21, 686e697.

Lim, Y. C., Kouzani, A. Z., & Duan, W. (2010). Lab-on-a-chip: a component view.Microsystem Technologies, 16, 1995e2015.

Liu, H.-Y., Lin, S.-L., Chan, S.-A., Lin, T.-Y., & Fuh, M.-R. (2013). Microfluidic chip-basednano-liquid chromatography tandem mass spectrometry for quantification ofaflatoxins in peanut products. Talanta, 113, 76.

Liu, Q., Wu, C., Cai, H., Hu, N., Zhou, J., & Wang, P. (2014). Cell-based biosensors andtheir application in biomedicine. Chemical Reviews, 114, 6423e6461.

Li, P., Zhang, Z., Zhang, Q., Zhang, N., Zhang, W., Ding, X., et al. (2012). Currentdevelopment of microfluidic immunosensing approaches for mycotoxindetection via capillary electromigration and lateral flow technology. Electro-phoresis, 33, 2253e2265.

Li, Y. S., Zhou, Y., Lu, S. Y., Guo, D. J., Ren, H. L., Meng, X. M., et al. (2012). Devel-opment of a one-step test strip for rapid screening of fumonisins B1, B2 and B3in maize. Food Control, 24, 72e77.

Lou, X., Qian, J., Xiao, Y., Viel, L., Gerdon, A. E., Lagally, E. T., et al. (2009). Micro-magnetic selection of aptamers in microfluidic channels. Proceedings of theNational Academy of Sciences of the United States of America, 106, 2989e2994.

Lu, Y., Lin, B., & Qin, J. (2011). Patterned paper as a low-cost, flexible substrate for

L. Guo et al. / Trends in Food Science & Technology 46 (2015) 252e263262

rapid prototyping of PDMS microdevices via “liquid molding”. AnalyticalChemistry, 83, 1830.

Magan, N., & Olsen, M. (2004). Mycotoxins in food: detection and control (vol. 113).Sawston: Woodhead Publishing.

Mao, L., & Koser, H. (2006). Towards ferrofluidics for m-TAS and lab on-a-chip ap-plications. Nanotechnology, 17, 34e47.

Mark, D., Haeberle, S., Roth, G., von Stetten, F., & Zengerle, R. (2010). Microfluidiclab-on-a-chip platforms: requirements, characteristics and applications.Chemical Society Reviews, 39, 1153e1182.

Martinez, A. W., Phillips, S. T., Whitesides, G. M., & Carrilho, E. (2009). Diagnosticsfor the developing world: microfluidic paper-based analytical devices. Analyt-ical Chemistry, 82, 3e10.

Monbaliu, S., Van Poucke, C., Van Peteghem, C., Van Poucke, K., Heungens, K., & DeSaeger, S. (2009). Development of a multi-mycotoxin liquid chromatography/tandem mass spectrometry method for sweet pepper analysis. Rapid Commu-nications in Mass Spectrometry: RCM, 23, 3e11.

Mosbach, K., & Ramstrom, O. (1996). The emerging technique of molecularimprinting and its future impact on biotechnology. Nature Biotechnology, 14,163e170.

Newman, C. I. D., Giordan, B. C., Copper, C. L., & Collins, G. E. (2008). Microchipmicellar electrokinetic chromatography separation of alkaloids with UV-absorbance spectral detection. Electrophoresis, 29, 803e810.

Ngundi, M. M., Shriver-Lake, L. C., Moore, M. H., Lassman, M. E., Ligler, F. S., &Taitt, C. R. (2005). Array biosensor for detection of ochratoxin A in cereals andbeverages. Analytical Chemistry, 77, 148e154.

Ngundi, M. M., Shriver-Lake, L. C., Moore, M. H., Ligler, F. S., & Taitt, C. R. (2006).Multiplexed detection of mycotoxins in foods with a regenerable array. Journalof Food Protection, 69, 3047e3051.

Nguyen, N.-T., & Wu, Z. (2005). Micromixersda review. Journal of Micromechanicsand Microengineering, 15, R1eR16.

Novo, P., Moulas, G., Prazeres, D. M. F., Chu, V., & Conde, J. P. (2011). Lab-on-a-Chipochratoxin a detection using competitive ELISA in microfluidics with integratedphotodiode signal acquisition. Procedia Engineering, 25, 1205e1208.

Novo, P., Moulas, G., França Prazeres, D. M., Chu, V., & Conde, J. P. (2013). Detectionof ochratoxin A in wine and beer by chemiluminescence-based ELISA inmicrofluidics with integrated photodiodes. Sensors and Actuators B: Chemical,176, 232e240.

O'Brien, E., & Dietrich, D. R. (2005). Ochratoxin a: the continuing enigma. CriticalReviews in Toxicology, 35, 33e60.

Oh, K. W., & Ahn, C. H. (2006). A review of microvalves. Journal of Micromechanicsand Microengineering, 16, 13e39.

Pan, T., McDonald, S. J., Kai, E. M., & Ziaie, B. (2005). A magnetically driven PDMSmicropump with ball check-valves. Journal of Micromechanics and Micro-engineering, 15, 1021e1026.

Parker, C. O., Lanyon, Y. H., Manning, M., Arrigan, D. W. M., & Tothill, I. E. (2009).Electrochemical immunochip sensor for aflatoxin M1 detection. AnalyticalChemistry, 81, 5291.

Pennathur, S. (2008). Flow control in microfluidics: are the workhorse flowsadequate? Lab on a chip, 8, 383e387.

Petroczi, A., Nepusz, T., Taylor, G., & Naughton, D. P. (2011). Network analysis of theRASFF database: a mycotoxin perspective. World Mycotoxin Journal, 4, 329e338.

Pfohl-Leszkowicz, A., Petkova-Bocharova, T., Chernozemsky, I. N., & Castegnaro, M.(2002). Balkan endemic nephropathy and associated urinary tract tumours: areview on aetiological causes and the potential role of mycotoxins. Food Addi-tives and Contaminants, 19, 282e302.

Piermarini, S., Micheli, L., Ammida, N. H. S., Palleschi, G., & Moscone, D. (2007).Electrochemical immunosensor array using a 96-well screen-printed micro-plate for aflatoxin B1 detection. Biosensors and Bioelectronics, 22, 1434e1440.

Richard, J. L. (2007). Some major mycotoxins and their mycotoxicosesdAn over-view. International Journal of Food Microbiology, 119, 3e10.

Roda, A., Mirasoli, M., Guardigli, M., Michelini, E., Simoni, P., & Magliulo, M. (2006).Development and validation of a sensitive and fast chemiluminescent enzymeimmunoassay for the detection of genetically modified maize. Analytical andBioanalytical Chemistry, 384, 1269e1275.

Roman, G. T., & Kennedy, R. T. (2007). Fully integrated microfluidic separationssystems for biochemical analysis. Journal of Chromatography A, 1168, 170e188.

Sackmann, E. K., Fulton, A. L., & Beebe, D. J. (2014). The present and future role ofmicrofluidic in biomedical research. Nature, 507, 181e189.

Sauceda-Friebe, J. C., Karsunke, X. Y. Z., Vazac, S., Biselli, S., Niessner, R., & Knopp, D.(2011). Regenerable immuno-biochip for screening ochratoxin A in green coffeeextract using an automated microarray chip reader with chemiluminescencedetection. Analytica Chimica Acta, 689, 234e242.

Schilling, K. M., Lepore, A. L., Kurian, J. A., & Martinez, A. W. (2012). Fully enclosedmicrofluidic paper-based analytical devices. Analytical Chemistry, 84, 1579.

Seemann, R., Brinkmann, M., Pfohl, T., & Herminghaus, S. (2012). Droplet basedmicrofluidic. Reports on Progress in Physics, 75, 016601.

Seidel, M., & Niessner, R. (2008). Automated analytical microarrays: a critical re-view. Analytical and Bioanalytical Chemistry, 391, 1521e1544.

Sforza, S., Dall'Asta, C., & Marchelli, R. (2006). Recent advances in mycotoxindetermination in food and feed by hyphenated chromatographic techniques/mass spectrometry. Mass Spectrometry Reviews, 25, 54e76.

Sheng, W., Chen, T., Kamath, R., Xiong, X., Tan, W., & Fan, Z. H. (2012). Aptamer-

enabled efficient isolation of cancer cells fromwhole blood using a microfluidicdevice. Analytical Chemistry, 84, 4199e4206.

Shim, W.-B., Dzantiev, B. B., Eremin, S. A., & Chung, D.-H. (2009). One-step simul-taneous immunochromatographic strip test for multianalysis of ochratoxin aand zearalenone. Journal of Microbiology and Biotechnology, 19, 83e92.

Soares, R. R. G., Novo, P., Azevedo, A. M., Fernandes, P., Aires-Barros, M. R., Chu, V.,et al. (2014). On-chip sample preparation and analyte quantification using amicrofluidic aqueous two-phase extraction coupled with an immunoassay. Labon a Chip, 14, 4284e4294.

Squires, T. M., Messinger, R. J., & Manalis, S. R. (2008). Making it stick: convection, re-action and diffusion in surface-based biosensors. Natuer Biotechnology, 26,417e426.

Squires, T. M., & Quake, S. R. (2005). Microfluidics: fluid physics at the nanoliterscale. Reviews of Modern Physics, 77, 977e1026.

Stocker, J., Balluch, D., Gsell, M., Harms, H., Feliciano, J., Daunert, S., et al. (2003).Development of a set of simple bacterial biosensors for quantitative and rapidmeasurements of arsenite and arsenate in potable water. Environmental Scienceand Technology, 37, 4743e4750.

Stoloff, L. (1976). Occurrence of mycotoxins in foods and feeds. Mycotoxins and otherfungal related food problems, advances in chemical series (vol. 149, pp. 23e50).

Stone, H. A., Stroock, A. D., & Ajdari, A. (2004). Engineering flows in small devices:microfluidics toward a lab-on-a-chip. Annual Review of Fluid Mechanics, 36,381e411.

Strobl, C., Schneider, M., Wixforth, A., Sritharan, K., & Guttenberg, Z. (2006).Acoustic mixing at low Reynold's numbers. Applied Physics Letters, 88, 054102-054102-054103.

Teh, S., Lin, R., Huang, L., & Lee, A. P. (2008). Droplet microfluidics. Lab on a Chip, 8,198e220.

Thuo, M. M., Martinez, R. V., Lan, W. J., Liu, X., Barber, J., Atkinson, M. B., &Whitesides, G. M. (2014). Fabrication of low-cost paper-based microfluidic de-vices by embossing or cut-and-stack methods. Chemistry of Materials, 26,4230e4237.

Tothill, I., & Saeger, S. d (2011). Emerging bio-sensing methods for mycotoxin analysis.Determining mycotoxins and mycotoxigenic fungi in food and feed (pp. 359e384).Cambridge, UK: Woodhead Publishing Ltd..

Unger, M. A., Chou, H. P., Thorsen, T., Scherer, A., & Quake, S. R. (2000). Monolithicmicrofabricated valves and pumps by multilayer soft lithography. Science, 288,113e116.

van Egmond, H. P., Schothorst, R. C., & Jonker, M. A. (2007). Regulations relating tomycotoxins in food: perspectives in a global and European context. Analyticaland Bioanalytical Chemistry, 389, 147e157.

Voldman, J., Voldman, J., Gray, M. L., & Schmidt, M. A. (2000). An integrated liquidmixer/valve. Journal of Microelectromechanical Systems, 9, 295e302.

Wang, H., Meng, S., Guo, K., Liu, Y., Liu, B., Yang, P., et al. (2008). Microfluidicimmunosensor based on stable antibody-patterned surface in PMMA micro-chip. Electrochemistry Communications, 10, 447e450.

Weng, C. H., Yeh, W. M., Ho, K. C., & Lee, G. B. (2007). A microfluidic system utilizingmolecularly imprinted polymer films for amperometric detection of morphine.Sensors and Actuators B-Chemical, 121, 576e582.

Whitesides, G. M. (2006). The origins and the future of microfluidics. Nature, 442,368e373.

Witters, D., Vergauwe, N., Vermeir, S., Ceyssens, F., Liekens, S., Puers, R., et al. (2011).Biofunctionalization of electrowetting-on-dielectric digital microfluidic chipsfor miniaturized cell-based applications. Lab on a Chip, 11, 2790e2794.

Xia, Y., & Whitesides, G. M. (1998). Soft lithography. Annual Review of MaterialsScience, 28, 153e184.

Xu, J. (2014). Microfluidics “lab-on-a-chip” system for food chemical hazarddetection. Food Chemical Hazard Detection: Development and Application of NewTechnologies, 263e289.

Xu, J., & Attinger, D. (2008). Drop on demand in a microfluidic chip. Journal ofMicromechanics and Microengineering, 18, 065020.

Xu, Y., Hashmi, A., Yu, G., Lu, X., Kwon, H., Chen, X., et al. (2013). Microbubble arrayfor on-chip worm processing. Applied Physics Letters, 102, 023702.

Xu, Y., Phillips, J. A., Yan, J., Li, Q., Fan, Z. H., & Tan, W. (2009). Aptamer-basedmicrofluidic device for enrichment, sorting, and detection of multiple cancercells. Analytical Chemistry, 81, 7436e7442.

Xu, J., Vaillant, R., & Attinger, D. (2010). Use of a porous membrane for gas bubbleremoval in microfluidic channels: physical mechanisms and design criteria.Microfluidics and Nanofluidics, 9, 765e772.

Yang, M., Sun, S., Kostov, Y., & Rasooly, A. (2011). An automated point-of-care systemfor immunodetection of staphylococcal enterotoxin B. Analytical Biochemistry,416, 74e81.

Yeh, C.-H., Chen, W.-T., Lin, Y.-C., Lin, H.-P., & Chang, T.-C. (2009). Development of animmunoassay based on impedance measurements utilizing an antibody-nanosilver probe, silver enhancement, and electro-microchip. Sensors & Actu-ators: B.Chemical, 139, 387e393.

Zain, M. E. (2011). Impact of mycotoxins on humans and animals. Journal of SaudiChemical Society, 15, 129e144.

Zhang, H., Betz, A., Qadeer, A., Attinger, D., & Chen, W. (2011). Microfluidic formationof monodispersed spherical microgels composed of triple-network cross-linking. Journal of Applied Polymer Science, 121, 3093e3100.

Zhang, M., & Wang, Z. (2013). Nanostructured silver nanowires-graphene hybrids

L. Guo et al. / Trends in Food Science & Technology 46 (2015) 252e263 263

for enhanced electrochemical detection of hydrogen peroxide. Applied PhysicsLetters, 102, 213104.

Zheng, M. Z., Richard, J. L., & Binder, J. (2006). A review of rapid methods for theanalysis of mycotoxins. Mycopathologia, 161, 261e273.

Zhou, J., Ellis, A. V., & Voelcker, N. H. (2010). Recent developments in PDMS surfacemodification for microfluidic devices. Electrophoresis, 31, 2e16.

Zhou, J., Khodakov, D. A., Ellis, A. V., & Voelcker, N. H. (2012). Surface modificationfor PDMS-based microfluidic devices. Electrophoresis, 33, 89.