REVIEW

Label-free technologies for quantitative multiparameterbiological analysis

Abraham J. Qavi & Adam L. Washburn &

Ji-Yeon Byeon & Ryan C. Bailey

Received: 17 November 2008 /Revised: 12 January 2009 /Accepted: 20 January 2009 /Published online: 17 February 2009# Springer-Verlag 2009

Abstract In the postgenomic era, information is king andinformation-rich technologies are critically important driv-ers in both fundamental biology and medicine. It is nowknown that single-parameter measurements provide onlylimited detail and that quantitation of multiple biomolecularsignatures can more fully illuminate complex biologicalfunction. Label-free technologies have recently attractedsignificant interest for sensitive and quantitative multipa-rameter analysis of biological systems. There are severaldifferent classes of label-free sensors that are currentlybeing developed both in academia and in industry. In thiscritical review, we highlight, compare, and contrast some ofthe more promising approaches. We describe the funda-mental principles of these different methods and discussadvantages and disadvantages that might potentially helpone in selecting the appropriate technology for a givenbioanalytical application.

Keywords Genomics/proteomics . Bioassays . Biochips/high-throughput screening . Biosensors . Clinical/biomedicalanalysis . Immunoassays/ELISA

Introduction

High-information-content genomic and proteomic technol-ogies, such as capillary sequencing, complementary DNAmicroarrays, two-dimensional polyacrylamide gel electro-

phoresis, and mass spectrometry, have greatly increased thelevel of molecular clarity with which we now understandhuman biology. Perhaps the most critical insight gleanedfrom these continued efforts is the vast interconnectivity ofgene and protein regulatory networks. This in turn leads tothe realization that biological systems are more completelycharacterized as an increasing number of molecularexpression profiles are obtained from a single analysis.Coupled with immortalized cell lines and modern molecularand cell biology techniques, the aforementioned genomicand proteomic tools are well suited and established inresearch laboratories. Unfortunately, many of the samemeasurement approaches are not rigorously quantitativeand also are not ideal for use in the clinic, where samplesizes and specialized training in analytical methods aremore limited.

The greatest challenges in quantitative clinical bio-analysis arise because of the requirement of a label—usually fluorescent or enzymatic. This label may bedirectly tethered either to the biomolecule under interro-

Anal Bioanal Chem (2009) 394:121–135DOI 10.1007/s00216-009-2637-8

A. J. Qavi :A. L. Washburn : J.-Y. Byeon : R. C. Bailey (*)Department of Chemistry,University of Illinois at Urbana-Champaign,600 S. Mathews Ave,Urbana, IL 61801, USAe-mail: baileyrc@illinois.edu

Ryan C. Baileyis an Assistant Professor of Chemis-try at the University of Illinois atUrbana-Champaign and an affiliatein the university’s Institute forGenomic Biology. He won a 2006Camille and Henry Dreyfus Foun-dation New Faculty Award and wasthe recipient of an inaugural USNational Institutes of Health Direc-tor’s New Innovator Award in 2007.His group is developing newmultiparameter biological analysistechnologies based upon high-

density optical biosensor arrays and novel surface patterning strategies.

gation or to a secondary or tertiary recognition elementsuch as in a sandwich assay configuration. In the case ofantibody-based sandwich assays, such as conventionalenzyme-linked immunosorbent assays, the requirementfor a secondary protein capture agent adds significantcost and development time as generation of multiple,high-binding-affinity antibodies that recognize distinctand nonoverlapping target epitopes can be very difficult.Direct labeling has its own challenges. Label incorpora-tion itself can be highly heterogeneous, making anyresulting measurement inherently nonquantitative [1].Furthermore, Sun et al. [2] recently demonstrated thatthe presence of a fluorescent label can have detrimentaleffects on the affinity of an antigen–antibody interaction.Since almost all biosensing methods essentially provide ameasure of surface-receptor occupancy, which is dictatedby the binding affinity between the capture agent and theantigen, this report validates the quite obvious fact thatlabels can, in many cases, negatively impact the limit ofdetection of an assay.

For these reasons, among others, there is great interest indeveloping label-free methods of biomolecular analysis.There are many different classes of label-free biosensors,but all are based upon the measurement of an inherentmolecular property such as refractive index or mass. In thisreview, we focus on examples of label-free biosensors inwhich multiple target analytes are assayed simultaneouslyfrom within the same sample. These transduction methods,which are often based upon micro- and nanotechnologies,therefore also have an advantage of relatively low sampleconsumption, since multiple sample volumes are notrequired for multiple assays.

Clearly it would be impossible to discuss every technol-ogy that fits the label-free multiplexed sensing criteria;therefore, in this review we have attempted to highlightsome of the technologies that we feel are the most promisingat present to make an impact in this rapidly developing field.We have chosen to break the review down into sectionsaccording to the manner in which the presence of thebiological moiety is transduced: plasmonic, photonic, elec-tronic, and mechanical methods. In each section we brieflyintroduce the key aspects of sensor operation, highlightnotable research to date, and comment on the advantagesand/or disadvantages of each technology as it applies to thedirect, multiplexed analysis of complex biological samples. Insome cases we will discuss applicability to detection ofdifferent classes of biomolecules and/or cells. Furthermore,while this review is mainly focused on detection (quantita-tion), we will also highlight several instances where label-freetechniques have been used in multiplexed molecular library(chemical and biological) screening applications, since manyof these demonstrations are highly relevant and potentiallyamenable to multiparameter quantitation.

We have intentionally chosen not to discuss othervaluable analytical metrics such as time-to-result andspecificity. Our reasoning is that while these metrics arecritically important to sensor utility, they can be severelycomplicated by other factors that are not related to thefundamental physical performance of the device. Proper-ties such as the dimensions of the sample chambersurrounding the sensor [3, 4] or the quality of the captureagent can dominate the observed time response andspecificity (and sensitivity, for that matter). We recognizethat these are vitally important facts to consider, but it isnot practical to qualify each literature report in terms ofthe many peripheral factors that affect these performanceattributes.

Plasmonic methods

One promising technique for the multiplexed, label-freedetection of biomolecules is surface plasmon resonanceimaging (SPRI), also referred to as “surface plasmonresonance microscopy.” SPRI is based upon the samefundamental principles as conventional surface plasmonresonance (SPR) spectroscopy; light is coupled to theinterface of a thin metallic film (typically gold forbiosensing applications) via total internal reflection wherepropagating surface plasmon modes are excited, if thephotons are of a particular frequency and incident angle.The evanescent field associated with the plasmon resonancesamples the proximal optical dielectric environment and ishighly sensitive to local changes in refractive index,including those associated with the binding of biomoleculesto receptors presented by the surface [5]. When biomole-cules bind to specific receptors anchored to the metallicfilm, the corresponding changes in the refractive indexmodulate the intensity of light reflected off the surface,which in turn is measured by the detector [6].

SPR spectroscopy was first demonstrated for biosensingapplications by Lundstrom in 1983 [7], and was furtherdeveloped throughout the mid-1980s as a method to monitorimmunochemical reactions [8]. SPRI, which allows multiplebinding events to be monitored simultaneously, was intro-duced by Yeatman and Ash [9] in 1987 and furtherdeveloped by Corn and colleagues through and mid- tolate-1990s [10–12]. In a typical SPRI experiment, shownschematically in Fig. 1, a CCD detector is used to image theintensity of light reflected off the surface, which directlycorresponds to the amount of material bound to the metalfilm at a given image position. In this arrangement, changesin reflected light intensity can be measured down to aresolution of approximately 4 μm, allowing for highlymultiplexed measurements of a variety of biological bindingevents [13]. However, multiplexing comes with a cost; SPRI

122 A.J. Qavi et al.

typically has limits of detection 10–100 times higher (worse)than standard, nonimaging SPR spectroscopy [6].

One major application of SPRI is the readout of massiveprotein microarrays. Work by Shumaker-Parry et al. demon-strated that the change in intensity of reflected light in anSPRI array could be correlated to a change in mass per unitarea for proteins [14]. They successfully utilized this methodto detect the binding of streptavidin to a biotinylated-DNAsubstrate, with a limit of detection of approximately 0.5 pgper 200-µm spot [15]. In later work, highlighted in Fig. 2,the same group demonstrated quantitative measurement ofthe sequence-specific binding of the transcription factor Gal4to double-stranded DNA sequences in a 120-componentarray with similar sensitivity [16].

An incredible breadth of SPRI applications in proteomicshave been described over the past decade. This literature isfar too broad to cover in its entirety here, and we direct thoseinterested to two outstanding reviews on SPRI technologiesfor biomolecular interaction monitoring [5, 13]. Thoughmany screening demonstrations have focused on only alimited number of components, SPRI systems are capable ofhigher levels of multiplexing—perhaps allowing for 10,000or more parallel measurements [17].

There has also been significant effort focused on utilizingSPRI for quantitative nucleic acid analysis; work pioneeredby the group of Corn. Nelson et al. monitored hybridizationof DNA and RNA onto microarrays at concentrations aslow as 10 nM for short oligonucleotides (18-mers) anddown to 2 nM for longer oligonucleotides (1,500 bases)[18]. Goodrich and coworkers [19, 20] used an enzymaticapproach to further extend the detection limit to approx-imately 1 fM. These reports took advantage of the abilityof RNase H to selectively cleave RNA from DNA:RNA

heteroduplexes—allowing the rate of duplex hydrolysis tobe correlated to the amount of bound target DNA. Lee etal. [21] developed a related enzymatic amplificationscheme using Exonuclease III, an enzyme that selectivelycleaves DNA:DNA homoduplexes. In this work, theauthors were able to detect down to 10 pM target DNA.Wolf et al. [22] demonstrated the use of SPRI as a tool forscreening small molecule–DNA interactions by observingthe interaction between actinomycin D with a multitude of

Fig. 2 a SPRI image of a 120-element double-stranded DNA array. bDifference image and c line scan after incubation of the array from awith the transcription factor Gal4. Specific protein binding is observedas a positive change in the reflected light image. (Adapted from [16])

Fig. 1 A surface plasmon resonance imaging (SPRI) instrumentalconfiguration. Biomolecular binding events are transduced as a changein reflected light intensity, and multiplexing is accomplished by imaginga large portion of the substrate using the CCD. (Adapted from [5])

Label-free technologies for quantitative multiparameter biological analysis 123

DNA sequences. While all of the aforementioned demon-strations of nucleic acid analysis used a limited number ofarray components, this technology clearly could be scaledto much higher levels of multiplexing.

Two variations to the traditional prism-based SPRtechniques that have attracted considerable attention formultiplexed biosensing are grating-coupled SPR (GCSPR)and waveguide-coupled SPR (WCSPR). GCSPR utilizes anoptical grating incorporated into the sensor surface togenerate high diffracted orders that couple photons to thesurface and in turn launch propagating surface plasmons[23–25]. This technique provides a number of advantagesover traditional methods of SPRI. Prisms or index-matchingfluids are not required for the generation of surfaceplasmons, providing flexibility in the experimental layouts.Furthermore, the sensors can be mass produced withrelative ease and low cost. However, the sensitivity ofGCSPR is generally lower than that of prism-based SPRmeasurements [6]. WCSPR involves the incorporation ofan optical fiber or other waveguide into the sensor as ameans to generate surface plasmons. While the incorpora-tion of a waveguide allows for miniaturization of thesensor, the coupling of the light from the waveguide to thesurface plasmons is heavily dependent on the polarizationof the incident light, which is sensitive to deformations inthe waveguide geometry, thus limiting the general utility ofthe technique.

The widespread utility of SPRI as a biomoleculartransduction technology is, in part, reflected in the numberof companies offering commercial instrumentation. Bia-core, owned by GE Healthcare, is the largest maker of SPRIinstruments, offering a variety of models for different scalesof analysis [26]. GWC Technologies currently manufac-tures an SPRI instrument with the ability to screen over 25different analytes simultaneously [27]. IBIS Technologiesoffers a versatile SPRI system that provides both fixed andscanning angle measurements for increased sensitivity [28].Toyobo, manufacturer of the MultiSPRinter, packages amicroarray spotter with its SPR imager, integrating theentire fabrication process within a single product [29].GenOptics has developed commercial SPRI instrumentscapable of interrogating arrays having more than 1,000different components [30].

SPRI is a robust technology that has proven to be avaluable tool for the label-free detection and analyses ofbiomolecules. The technique possesses the ability tosensitively detect a wide range of biomolecules, includingnucleic acids, proteins, and carbohydrates. On account ofthe relatively large area of individual sensing elements,SPRI has a relatively poor limit of detection in terms ofabsolute bound mass, which might be a drawback in somesample-limited applications. While SPRI has not beenwidely utilized for multiparameter quantitation (examples

focused almost exclusively on nucleic acid detection), themany successful demonstrations of multiplexed interactionscreening make it a promising technology for suchapplications.

Another plasmonic-based biosensing platform that hasrecently emerged is based on the phenomenon of extraor-dinary optical transmission through periodic, subwave-length nanoholes in metallic thin films [31]. The intensityof light transmitted through these substrates is significantlyhigher than predicted by classical theory and has beenshown to be mediated by surface plasmons [32]. Theperiodic holes act as a high-order diffraction grating thatlaunches propagating plasmons linking the front and backsides of the metal thin film. The propagating plasmons thendecouple from the substrate by emission of a new photonfrom the back side of the film [33]. Since propagatingplasmons are sensitive to changes in the local refractiveindex, similar to SPR, nanohole arrays are responsive tobiomolecular surface binding events. However, in thissystem biomolecular binding is transduced as a shift in thewavelength of light maximally transmitted by the nanoholearray. Several recent reports of nanohole-based biosensorshave emerged out of the Larson group [34, 35]. The authorshave rigorously defined the factors that effect devicesensitivity and demonstrated the sensitive detection of anti-glutathione S-transferase antibodies down to concentrationsas low as 10 nM on arrays that are capable of supporting atleast 25 simultaneous measurements, as shown in Fig. 3.Advantages of nanohole arrays for biomolecular detectioninclude the extremely small footprint of the active sensingarea (down to 1 μm2) and the batch fabrication potential ofthe substrates, both of which should greatly facilitate highlevels of sensor multiplexing (yet to be demonstrated).

Photonic techniques

In addition to SPRI and nanohole arrays, several non-plasmonic optical biosensors currently show promise forhigh-throughput, multiparameter analysis—we term these“photonic techniques.” Photonic-based, label-free biosens-ing is not a new concept. As early as 1937, Langmuir andSchaefer [36] described a method for evaluating thethickness of adsorbed monolayers of biomolecules on ametal surface by observing the colors generated byreflective interference. In the late-1960s, Vroman andAdams [37] demonstrated the use of ellipsometry formeasuring immunoadsorbed molecules on a surface. Fol-lowing a timeline similar to that for SPR spectroscopy,optical biosensors based on technologies such as opticalwaveguides and reflective interferometry began to emergein the 1990s [38, 39]. Examples that we will discuss hereinclude photonic crystals, optical microcavity resonators,

124 A.J. Qavi et al.

reflective interferometry, and imaging ellipsometry. Typi-cally, these methods take advantage of microscale fabrica-tion methods and incorporate imaging or rapid scanning tointerrogate a large number of sensors simultaneously or innear real time.

Cunningham et al. [40] have pioneered a novel photoniccrystal biosensing platform that has proven to be effectivefor multiplexed screening and detection. Photonic crystalsare engineered to selectively reflect a narrow bandwidth oflight, and the wavelength at which this maximal reflectionoccurs is sensitive to the refractive index environmentsurrounding the substrate. As a result, bound biomoleculescause a measurable shift in the reflected wavelength (seeFig. 4) [40, 41]. The wavelength shift directly correspondsto the amount of bound biomolecule, and thus can be usedfor quantitation. With use of this principle of detection,photonic crystal biosensors have demonstrated the ability todetect adsorbed protein down to approximately 1 pg/mm2

on the surface [42].For performing multiplexed measurements, plastic-

molded photonic crystal structures have been incorporated

into a microplate format for use with conventionalbiological assays, and multiple wells can be analyzedsimultaneously using optical imaging techniques. Photoniccrystal sensors are well suited for multiparameter detectionand quantitation of biological analytes [40, 42, 43];however, many of the most relevant demonstrations to datehave focused on multiplexed biomolecular interactionscreening. For example, by monitoring the density of cellswithin a microplate well, Chan et al. [44, 45] havedemonstrated the ability to rapidly screen compounds foreffects on rates of cancer cell proliferation and apoptosis.They have also recently shown that a microplate photoniccrystal format can be used to screen thousands ofcompounds for their ability to inhibit specific protein–DNA interactions [46]. Choi and Cunningham [47] incor-porated 11 different microfluidic channels onto a 96-wellmicroplate system to allow parallel determination ofrelative binding affinities between protein A and sevendifferent IgGs using a single photonic crystal substrate.

Photonic crystal biosensors have enabled rapid, label-free monitoring of protein binding and cell growth,

Fig. 3 a Sample introductiononto a nanohole array biosensor.b Nanohole array instrumentalsetup. c CCD image of 30 setsof nanohole arrays having dif-ferent geometries. d, e Scanningelectron micrographs showing atop and a side view of a 9×9nanohole array. (Adapted from[34])

Label-free technologies for quantitative multiparameter biological analysis 125

demonstrating the utility of these sensors for high-through-put screening applications. Scalable manufacturing methodshave facilitated the commercialization of this technology, asit is currently available from SRU Biosystems [48]. To date,photonic crystal biosensors have not been rigorouslyutilized for quantitative concentration determination incomplex solutions; however, their promise for multiplexeddetection points towards future applications in these areas.

Optical microcavity sensors of various geometries havealso shown promise for highly sensitive label-free detection[49]. These sensors are based on the refractive indexsensitivity of cavity modes supported by microfabricatedwaveguide structures that satisfy the constructive interfer-ence condition: ml ¼ 2prneff , where m is an integer value,λ is the wavelength of light, r is the radius of the cavity,and neff is the effective refractive index that is sampled bythe optical mode. The wavelength at which resonanceoccurs is extremely narrow owing to precise fabrication ofthe optical cavity (high-Q cavities). Because the resonantwavelength is a function of the refractive index environ-ment surrounding the optical cavity, sensing is accom-plished by measuring the change in resonant wavelength inresponse to biomolecular binding at the microcavitysurface. The narrow resonance bandwidth of high-Qmicrocavities, amongst other factors, helps make small

shifts resolvable, which translates into low detection limitsfor biomolecular binding events.

Using microtoroidal resonators, Armani et al. [50]demonstrated single-molecule detection of the cytokineinterleukin-2 binding to an antibody-modified microcavity.Suter et al. [51] utilized liquid-core optical ring resonators(LCORRs) to detect DNA at a surface density of 4 pg/mm2,and Zhu et al. [52] demonstrated virus detection withLCORRs at 2.3×10–3 pfu/mL. White et al. [53, 54] alsointroduced a multiplexed LCORR array incorporating up toeight antiresonant reflecting optical waveguides coupled toa glass capillary allowing interrogation of multiple opticalcavities. In addition to containing multiple sensing ele-ments, the active resonator element, the capillary, integratesfluid handling into the sensing system.

Using silicon-on-insulator (SOI) microring resonators,Ramachandran et al. [55] measured bacteria down toconcentrations of 105 cfu/mL. Mandal et al. [56, 57]demonstrated an optofluidic system based on microfabri-cated photonic crystal cavities that are coupled to awaveguide bus on a patterned SOI substrate. Though theyhave demonstrated a 20-component-capable chip, they havenot yet performed actual biosensing with their system.

Work by De Vos et al. [58] and Ramachandran et al. [55]has illustrated the potential for using SOI microring

Fig. 4 Photonic crystal biosen-sors transduce biomolecularbinding events by measuring theshift in wavelength of lightreflected by the substrate.Shown here is a 384-well-plateconfiguration of a photoniccrystal sensing platform, whichcan be interrogated using alight-emitting diode and a sim-ple spectrometer. This exampledemonstrates the screening ofsmall-molecule libraries forinhibiting a specific DNA–pro-tein binding event. (Adaptedfrom [46])

126 A.J. Qavi et al.

resonators for multiplexed biosensors. Notably, standardsemiconductor processing should allow multiple sensors tobe integrated onto a single chip, as shown in Fig. 5.Furthermore, fabrication of both waveguides and micro-cavities on the same SOI substrate may offer significantadvantages in terms of baseline noise, compared withcoupling via even the most efficient free-standing extrudedoptical fiber approaches [59]. Our own unpublished workhas demonstrated that biomolecular binding to arrays of 2430-µm-diameter SOI microring resonators can be simulta-neously monitored with a sensitivity to surface-boundprotein of approximately 1 pg/mm2 (A.L. Washburn and R.C. Bailey, unpublished results). Given the very small totalsurface area of SOI microring structures, this corresponds todetection limits of less than 100 ag of total bound protein.

Thus far, optical microcavity resonators have shownpromise for highly sensitive, label-free biosensing. Inaddition, the small size of the microcavities makes thesebiosensors more sensitive to the absolute mass of surface-adsorbed biomolecules compared with techniques which uselarger sensing elements or surface areas. Multiplexed sensing

with optical microcavity resonators appears promising, andliterature demonstrations are expected in the near future.

Other photonic detection techniques such as ellipsometryand reflective interferometry are also candidates for label-free multiplexed biosensing applications. In general, thesetechniques sensitively measure small changes in opticalthickness on a surface. In 1995, Jin et al. [60] firstdemonstrated that imaging ellipsometry could be used tomeasure arrays of adsorbed biomolecules on a surface. Thetechnique is identical to traditional ellipsometry—themeasurement of polarization changes in light reflected offa surface—except that a CCD imaging detector is utilizedto simultaneously interrogate thousands of discrete loca-tions on a functionalized surface. Standard imagingellipsometry can measure coatings of biomolecules approx-imately 0.1 nm thick (average optical thickness) on asurface [61]. Using oblique-incidence reflectivity differencemicroscopes, Landry et al. [62] reported sensitivity down tooptical thicknesses of approximately 0.01 nm for adsorbedprotein on a surface and Wang and Jin [63] reported asimple array-based multiplexed analysis of common proteins(bovine serum albumin, human serum albumin, IgG, andfibrinogen). Subsequent examples have demonstrated 48-element arrays with the ability to simultaneously detect humanIgG, fibrinogen, and five protein markers for hepatitis [61].Landry et al. [62, 64] have also shown that oblique-incidencereflectivity difference microscopy can be used to analyzelarge arrays having hundreds of sensing elements. However,these spots were largely redundant, containing only one ortwo unique types of capture probes such as antibodiesagainst IgG and albumin or nucleic acid arrays presentingonly complements to a single DNA target sequence.

Currently, imaging ellipsometry allows measurement oflarge arrays of biomolecular interactions on a single surfacewith sensitivity comparable to that of SPRI [64, 65]. Recentapplications of the technique have shown the ability ofimaging ellipsometry to detect multiple biomoleculessimultaneously from complex samples. Though it is not asfully developed as SPRI, recent refinements and improve-ments making the technique more “user friendly” (see thereview by Jin [61]) may lead to an increased application ofthe technique for multiparameter analyses.

Reflective interferometric platforms have also shown theability to measure small changes in optical thickness of abiomolecular layer on a surface. Rather than measuringchanges in polarization, as in the case of ellipsometry,reflective interferometric techniques utilize optical interfer-ence between incident photons reflecting off a thin film.Changes in reflectance can thus be correlated withdifferential interference that occurs as biomolecules attachto a reflective surface. The sensitivity of reflectiveinterferometric techniques enables measurement of changesin surface thickness of about 1 pm, corresponding to

Fig. 5 Photograph of a five-ringed silicon-on-insulator microringresonator array used to detect biological binding events. In thisexample, the microrings are accessed by on-chip waveguides that aretapered off-chip to conventional fiber optics. (From [55])

Label-free technologies for quantitative multiparameter biological analysis 127

approximately 1 pg/mm2 of adsorbed biomolecule on thesurface [66, 67]. Gauglitz and coworkers have appliedreflective interferometric spectroscopy (RIfS) to the multi-plexed measurement of biomolecules in 96- and 384-wellplates [68–70]. Using a backlit configuration with CCDimaging, the authors were able to perform real-timeanalyses of biological binding events. This technology hasbeen demonstrated for multiplexed assays aimed at screeningantibodies against triazine libraries (four antibodies against36 different compounds) [69] as well as for epitope mappingof the enzyme transglutamase [70]. The same group hasalso utilized RIfS for measuring nucleic acid duplex melting[71] and cell binding [72]. A commercial RIfS platform iscurrently being developed by Biametrics [73].

Özkumur et al. [67] have demonstrated a spectralreflectance imaging biosensor that measures small changesin interference due to the change in optical pathlength fromsurface-adsorbed molecules. With use of a CCD camera, 200different spots can be measured simultaneously, allowingreal-time kinetic measurements to be made by monitoringthe binding regions through a glass microfluidic substrate.To date, they have demonstrated the interaction of a proteinwith a surface-bound capture agent at a detection limit of19 ng/mL. A commercialized version of this technology isbeing developed by Zoiray Technologies [74].

Zhao et al. [75] reported a similar reflectometric systemfor immunosensing applications called the “biologicalcompact disc” (commercialized by Quadraspec [76]). Thistechnique has been expanded for parallel analyses and isreferred to as “molecular interferometric imaging” (MI2).MI2 has been used to measure prostate-specific antigen(PSA) as well as interleukin-5 (IL-5) with detection limitsof 60 and 50 pg/mL, respectively, on a surface presenting128 functionalized areas [77, 78]. Currently, however, MI2has only demonstrated the ability to measure the concen-trations of two different proteins simultaneously. Mace etal. [79] recently utilized an array-based sensor for measur-ing three different cytokines simultaneously and demon-strated a limit of detection of 25 pg/mL for interferon-γ;however, the sensitivity and reproducibility was limited bylarge variation in spot morphology as a result of needing adry surface for analysis.

Though they are widely variable in experimental setup(front- versus back-illuminated, fluid handling, etc.), reflec-tive interferometric techniques have proven to be extremelysensitive to biomolecules adsorbed on a surface. Quantita-tive multiplexed assays based upon reflectometric techni-ques are still at an early stage of development, butmolecular library screening applications are more mature.Future efforts in multiplexed analyses will likely lead toimprovements in limits of detection for proteins andbroaden the classes of target analytes to which thetechnology can be applied.

Electrical detection

Electrical detection methods have been suggested asattractive alternatives to optical readouts owing to theirlow cost, low power consumption, ease of miniaturization,and potential multiplexing capability [80, 81]. The basis ofthese detection systems stems from a binding-inducedchange in some electrical property of the circuit, of whichthe sensor is a vital component. Electronic biosensingplatforms that we will discuss here include semiconductingsilicon nanowires (SiNWs), carbon nanotubes (CNTs), andelectrochemical impedance spectroscopy (EIS).

The most common mode of biomolecular detectionusing semiconductor nanowires is based on the principlesof the field effect transistor (FET). In normal transistoroperation, a semiconducting element is attached to a sourceand a drain electrode, and current flowing through theelement is modulated by changing the voltage applied to agate electrode [82]. In a FET-based nanowire biosensorconfiguration, the SiNW, functionalized with appropriatereceptors such as single-stranded DNA oligomers or anti-bodies, is connected to a source and drain electrode.Binding of target biomolecules changes the dielectricenvironment around the nanowires and plays a role similarto that of the gate electrode. Thus, molecular binding can bedirectly quantitated as a change in the conductivity of thenanowires (see Fig. 6) [83]. Compared with the relativelylarge dimensions of earlier planar FET sensors [84], thesmall size of nanowires means that individual bindingevents result in a more significant change in the electricalproperties of the circuit. This unique feature of nanowireFETs provides ultrahigh sensitivity down to the single-molecule level [85]. Since the first report of biosensingapplications [86], SiNWs have been shown to be broadlyapplicable to a wealth of analytical challenges, including thelabel-free detection of ions [86], small molecules [87],proteins [86, 88, 89], nucleic acids [90–92], viruses [85],and neuronal signals [93].

The first example of a multiplexed array of SiNWsensors was reported in 2004 by Patolsky et al. [85].Surfaces of two p-type SiNWs were modified withmonoclonal antibodies specific for influenza A and adeno-virus particles, and selective binding and unbindingresponses for each virus were detected in parallel at thesingle-particle level. In 2005, Zheng et al. [88] demonstrat-ed the multiplexed detection of three cancer markerproteins, free PSA, carcinoembryonic antigen, and mucin-1, in undiluted serum at femtomolar concentrations.Notably, in this work the serum sample was desalted toreduce the solution ionic strength.

Nanowire sensing technology holds much promise formultiparameter biological detection; however, there aresome significant challenges to be addressed before routine

128 A.J. Qavi et al.

operation of higher-order multiplexed analyses can berealized. One challenge that is inherent to FET-baseddetection systems is that they suffer reduced sensitivitywhen operated at physiological ionic strengths (approxi-mately 0.15 M). Ions in solution gate the FET similarly tothe biomolecular target, and thus the device can experiencea much diminished response to the binding event [94]. Thiscan be avoided by desalting the sample prior to analysis,but introduces an additional preparative step prior toanalysis [82, 92]. Though it is only partially related to theabsolute mass sensitivity of an individual nano-FET, thenanometer-size regime of these sensing elements alsomakes their response highly dependent upon sample cellgeometry and related parameters [3, 4].

A second significant challenge relates to the integrationof nanowires on substrates with reproducibility anduniformity [95]. Most SiNW biosensors reported to datehave used nanowires fabricated via the vapor–liquid–solid(VLS) method, which gives high yields of uniform nano-wires that have very advantageous properties for electronicand bioelectronic applications. Large numbers of VLSnanowires are grown simultaneously from individualcatalyst particles and subsequent positioning and alignmenton a sensor surface represents a significant hurdle.Recently, Fan et al. [95] introduced a novel method thataddresses this challenge, achieving approximately 95%directional alignment of VLS nanowires via contact

printing, which may greatly help facilitate integration ofhigh-density VLS nanowire sensing arrays.

An alternative method of preparing SiNW arrays wasreported—originally for nonanalytical applications —byMelosh et al. [96] in 2003. This technique, termed “super-lattice nanowire pattern transfer” (SNAP), utilizes a noveltemplate and shadow mask approach to create ultrahighdensity arrays of precisely aligned SiNWs on standard SOIsubstrates. Using these SNAP nanowires, shown in Fig. 7,Bunimovich et al. [92] demonstrated label-free detection ofsubnanomolar (approximately 100 pM) DNA concentra-tions. Stern et al. [89] have used CMOS-compatibleSiNWs, though fabricated at considerably lower densities,for the detection of proteins (antibodies) below 100 fM. Acombination of the described scalable fabrication methodsand ultrasensitive device operation may provide an attrac-tive method for measuring the concentrations of manydifferent biomolecules simultaneously.

CNT-based devices have also been widely investigatedas biosensors owing to their unique structure-dependentelectronic and mechanical properties [97]. In particular,single-walled CNTs can be metallic, semiconducting, orsemimetallic depending on the tube diameter and thechirality [98–100]. As a result, they have been used in awide range of applications, including, but not limited to, theFET transduction principle. CNT-based biosensors havebeen demonstrated for detection of small molecules [101,

Fig. 6 a A silicon nanowire-based field effect transistor de-vice configured as a sensor withantibody receptors (green),where binding of a protein withnet positive charge (red) yields adecrease in the conductance. bCross-sectional diagram andscanning electron microscopyimage of a single silicon nano-wire sensor device, grown viathe vapor–liquid–solid method,and a photograph of a prototypenanowire sensor biochip withintegrated microfluidic sampledelivery. (Adapted from [83])

Label-free technologies for quantitative multiparameter biological analysis 129

102], DNA (approximately 50 nM) [100, 103, 104],hepatitis C viral RNA (0.5 pM) [105], and IgE (250 pM)[106]. Multiparameter detection of biomolecules using aFET operation modality has yet to be realized, but reportsof multiplexed gas detection point towards the possibility[107]. Using AC voltammetry and multiwalled CNT arrays,Koehne et al. [104] demonstrated a readily scalableapproach to DNA detection. A novel electrochemicaletching and surface passivation scheme that should allowmultiplexing was described, but it was only applied tosingle-parameter detection at moderate sensitivity (approx-imately 100 nM). CNT arrays have shown potential forbiological detection; however, widespread utility has beenlimited to date by difficulties in controlling the physicalparameters relevant to biosensing: length, diameter, andchirality [98]. These issues are particularly significant formultiplexed sensing, in which uniform and reproducibleperformance of each sensor element is essential.

Another electrical transduction method that has shownpromise for multiparameter biological detection is EIS. InEIS, sensing is accomplished by measuring changes in the

resistance and/or capacitance of the electrode–solutioninterface upon binding of a target molecule to a receptor-functionalized surface [80, 81]. Compared with otherelectrical measurements, such as amperometry and voltam-metry, EIS does not require the use of enzymes to amplifybinding events by generating faradaic readout signals.This is very significant because sensor crosstalk due todiffusion of enzymatic products would be a fatal problemfor multiplexed detection applications [108]. Improve-ments to basic EIS operation have been reported that utilizealternative electrode materials such as polymers [109,110] and nanoparticles [111, 112]. Furthermore, electrodegeometry has been shown to be extremely important, witharrays of interdigitated electrodes providing greater devicesensitivity [81, 113, 114]. Various types of biologicalspecies have been detected using EIS, including nucleicacids [113, 115, 116] down to 10 fM [117] and bacteria[114, 118] as low as 10 cfu/mL [119]. Impedance-basedsensing systems have also been applied to monitor protein–carbohydrate [120] and protein–protein [121, 122] inter-actions. Recently, a demonstration of multiplexed proteininteraction monitoring was reported by Yu et al. [108]where an array of gold electrodes was used to probe forfour antibodies that recognized proteins immobilized on theelectrodes.

At this stage, EIS appears to be promising for multi-plexed biosensing and several commercialized systems arealready available for cell-based screening [123] (fromACEA Bioscience [124] and Applied BioPhysics [125]).In these systems, cells grown in electrode-containing wellscan be assayed for proliferation [126], adhesion andspreading [127, 128], and cytotoxicity [126, 128–130]. Asthe technology continues to mature, it will be necessary todevelop a greater understanding of the exact mechanismsunderlying the binding-induced change in impedance. Thisinsight will then allow for better a priori design ofexperimental conditions, circuit modeling, and fitting ofthe resulting data [81].

Mechanical sensors

Mechanical sensors are another promising tool for themultiplexed, label-free detection of biomolecules. Likeother label-free sensing methods, the specificity of thesesensors is determined by the topmost coating layer, which canrange from a modified gold surface to a variety of polymers[131–133]. One key advantage of this class of sensors is thatit is amenable to a wide range of surface coatings.

Well-known acoustic wave biosensors, including quartzcrystal microbalance and integrated surface acoustic wavetechnologies, are based on mechanical transduction andthus do not require labels. Both approaches utilize a

Fig. 7 A diagram (a) and a scanning electron micrograph (b) of threegroups of ten, 20-nm-wide silicon nanowires used for label-free DNAdetection. With use of the superlattice nanowire patterning scheme,large numbers of precisely aligned nanowires can be fabricated for useas biosensors. (From [92])

130 A.J. Qavi et al.

piezoelectric quartz crystal connected to an oscillatingexternal circuit that is able to measure the resonantoscillatory frequency of the system. Binding of moleculesto the surface of the sensor is measured as a shift in theresonance frequency of the device [134]. These sensorshave been utilized to detect a wide range of biomolecules,including nucleic acids [135–137], proteins [138–140], andlipids [141]. While acoustic wave devices are extremelysensitive towards binding events, there are a number offactors that limit their effective utilization for quantitative,multiplexed biosensing [142, 143]. Most importantly,considerations such as viscoelasticity and hydration leadto nonlinearities in frequency shifts accompanying biomo-lecular binding to the quartz crystal microbalance surface[144–146].

A second class of mechanical sensors that has recentlyattracted considerable attention as a multiplexed, label-freebiosensing platform is microcantilevers, as highlighted inFig. 8. Binding events on the cantilever surface aretransduced via one of three methods: deflection of thecantilever [147], a change in the resonance frequency ofthe cantilevers [148], or a change in the stress exerted onthe cantilever, which in turn generates an electric current inan attached piezoelectric element [133].

The simplest transduction event to monitor is the changein deflection, measured by reflecting a laser beam off theback of a cantilever and measuring the position with a splitphotodiode. The binding of an analyte to the surface of thecantilever exerts a torque, meaning that the location on thecantilever at which the molecule binds affects the amountof deflection. Because each molecule does not generate thesame amount of deflection, the position of bound moleculesmust be considered during deflection measurements (par-ticularly important for large molecules). Furthermore, therequired number of molecules bound to a surface to exert adetectable deflection is significantly higher than with theother two transduction methods [149] (discussed below).

Owing to its ease of implementation, cantilever deflec-tion has resulted in a wide range of biological moleculesand even entire microorganisms being detected, illustratedschematically in Fig. 9. In 2000, Fritz et al. [150]demonstrated the detection of a single base-pair mismatchwithin a 12-mer sequence of DNA. Following this,McKendry et al. [151] demonstrated the detection ofDNA to concentrations as low as 75 nM with an eight-component array. A number of protein systems have alsobeen studied with this technique [152–154]. Of specialinterest is the work by Wu et al. [155] in which PSA wasdetected at clinically relevant levels as low as 0.2 ng/mL inserum containing 1 mg/mL albumin and plasminogen.Recently, an interesting report out of the same laboratorydemonstrated sensitive protein detection using very largearrays of up to 960 individually readable microcantilevers

Fig. 8 Two-dimensional microcantilever array chip used to monitorprotein–protein interactions. a A reaction well. There were multiplecantilevers in each reaction well. Laser light reflected off a cantilever’send pad was used to monitor the deflection of cantilevers. b A chipsoaked in deionized water. c A scanning electron micrograph of threecantilevers in a reaction well. (Adapted from [156])

Fig. 9 The principle of deflection-based microcantilever biosensing.(From [155])

Label-free technologies for quantitative multiparameter biological analysis 131

[156]. Notably, though only a single protein was detected,PSA, the sensitivity was quite good, 1 ng/mL, and thesensor array was shown to have minimal response tononspecific proteins at much higher concentration.

Resonance-based transduction of microcantileversinvolves monitoring the change in the resonance frequencyupon binding of an analyte [148]. This is typicallyaccomplished using one of a number of interferometricschemes. While the readout equipment for this method issignificantly more complex than that for deflection-basedmethods, it is by far the most sensitive [147]. Onelimitation of resonance-based sensing is that the oscillationsof microcantilevers are highly susceptible to dampeningeffects in liquids, which vary with solution composition.

Using the more sensitive resonance-based measurementstrategy, involving the measurement of cantilever resonancefrequency shifts, Ilic et al. [157] demonstrated the detectionof a single strand of DNA 1,587 base pairs in length,having a mass of 1.7 ag. A number of protein systems havealso been explored, with detection limits as low as 10 pg/mLfor PSA [158, 159]. Studies focusing on the detection ofmicrobes have also pushed the limits of detection down tosingle cells and virus particles [160–162].

The incorporation of piezoelectric materials into micro-cantilevers can also be used to probe the presence ofbiomolecules [133]. In the piezoelectric method, thebinding of an analyte causes the cantilever to deflect,subsequently depolarizing the material and generating acurrent. It should also be noted that piezoelectric-modifiedmicrocantilevers can be used for both deflection-based and

resonance-based measurements. While the piezoelectricmethod is less sensitive than the resonance method, it doesnot require the extensive optical layouts used in the othertechniques and allows for incorporation of additionalelectronic components in the sensor. Even though detectionsystems have been demonstrated utilizing the piezoelectricreadout method, including examples of nucleic acid (limitof detection 10 nM for a 12-mer) and protein (limit ofdetection 5 ng/mL) analysis [163–165], significant advan-ces are still needed to lower the limits of detection tocompete with resonance-based methods.

Microcantilever-based sensors currently represent anattractive method for sensitive, label-free detection ofbiomolecules. The detection limits are comparable if notbetter than those for SPRI and the flexibility of operationand ease with which the cantilevers can be functionalizedallows for virtually any system to be studied. Currently,several companies are developing commercial microcanti-lever-based systems, including Cantion [166], BioScale[167], and Concentris [168]. The advances stemming fromboth industry and several academic groups are rapidlyadvancing microcantilever technology for multiparameterbioanalytical applications.

Conclusions and outlook

Taking values from a selection of literature sources anddiscussions with experts in the respective fields, we havecompiled a table comparing many of the label-free

Table 1 Comparison of several promising label-free, multiparameter biosensing technologiesa

Technology Reported limit of detection(lowest values found in the literature)

Multiplexing Commercialproduct?

Plasmonic Surface plasmon resonanceimaging

∼1 fM for nucleic acids [19, 20]; 1,000+ elements demonstrated [17] Yes [26–30]0.5 pg per element for proteins [15, 16]

Nanohole arrays 10 nM for proteins [35] 25 measurements demonstrated [35] NoPhotonic Photonic crystals 1 pg/mm2 [40] 1,536-well plate assays [48] Yes [48]

High-Q microcavities Microtoroids: single-molecule(zeptogram) [50]

Two-component [169], larger arrayspossible

No

Silicon-on-insulator microrings: 80 agb

Imaging ellipsometry <1 ng/mL for protein [61] ∼10-component assays demonstrated [61] NoReflective interferometry 1 pg/mm2 [66, 67] 384-well plate assays [68] Yes [74, 75]

Electronic Silicon nanowires 10 fM for DNA [90, 170];∼3 pM for protein[88]

3 parameters demonstrated [88] No

Carbon nanotubes 0.5 pM for RNA strand [105]; Single parameter demonstrated Yes [171]250 pM for a protein [106]

Electrochemical impedancespectroscopy

∼25 pM for protein [123] 4 proteins detected [108] Yes [124, 125]

Mechanical Microcantilevers 1.7 ag [157] 320+ demonstrated [156] Yes [166–168]

a Compiled to the best of the authors’ knowledge at the time of submission. In addition to our own expertise and searching, additional informationwas sought from experts in the respective fields. The authors apologize for any unintentional oversights.b A.L. Washburn and R.C. Bailey, unpublished results

132 A.J. Qavi et al.

biosensing technologies discussed in this review (Table 1),highlighting reported limits of detection, degree of multi-plexing demonstrated in the literature, and status ofcommercialization. This is not meant to be a standaloneselection guide, as the specific requirements of the assay(s)should play a critical role in which technology best suits theapplication at hand. For example, absolute mass sensitivitymay be of extreme importance in sample-limited applica-tions and therefore a smaller surface area sensor, providedthere is no loss in “bulk” sensitivity, might be advanta-geous. Another application might require very immediateanalysis because of sample degradation, for example. Inthis situation, nanowire sensors requiring sample desaltingmight not be the best choice. Having small sensing featuresmay allow for construction of higher-density sensor arraysfor more highly multiplexed applications. However, asignificant discrepancy may exist between theoretical and“functional” sensor densities, which may be limited moreby the method of sensor derivitization than the dimensionsof the sensing elements themselves. While it is certain thateach technology will have specific advantages and dis-advantages for a given application, each of the modalitiesdescribed may be the best option for a targeted multipa-rameter bioanalysis situation.

There are more specific applications and many addition-al compelling reasons that motivate the development ofnew and improvement of current label-free multiparameterbiodetection technologies. The next several years promiseto be an exciting time in this rapidly advancing field, whichis poised to impact clinical diagnosis and disease manage-ment in the very near future.

Acknowledgements We acknowledge financial support for our ownefforts in developing new quantitative, label-free multiparameterbiomolecular analysis methods from the following agencies: the NIHDirector’s New Innovator Award Program, part of the NIH Roadmapfor Medical Research, through grant number 1-DP2-OD002190–01;the Camille and Henry Dreyfus Foundation, through a New FacultyAward; and the US National Science Foundation through the Scienceand Technology Center of Advanced Materials for the Purification ofWater with Systems (WaterCAMPWS, CTS-0120978). A.L.W.acknowledges support via a National Science Foundation GraduateResearch Fellowship.

References

1. Kodadek T (2001) Chem Biol 8:1052. Sun YS, Landry JP, Fei YY, Zhu XD, Luo JT, Wang XB, Lam

KS (2008) Langmuir 24:13399–134053. Sheehan PE, Whitman LJ (2005) Nano Lett 5:8034. Squires TM, Messinger RJ, Manalis SR (2008) Nat Biotechnol

26:4175. Smith EA, Corn RM (2003) Appl Spectrosc 57:320a6. Homola J (2008) Chem Rev 108:462

7. Liedberg B, Nylanderand C, Lundstrom I (1983) Sens Actuators4:299

8. Kooyman RPH, Kolkman H, Van Gent J, Greve J (1988) AnalChim Acta 213:35

9. Yeatman E, Ash EA (1987) Electron Lett 40:109110. Jordan CE, Corn RM (1997) Anal Chem 69:144911. Jordan CE, Frutos AG, Thiel AJ, Corn RM (1997) Anal Chem

69:493912. Thiel AJ, Fructos AG, Jordan CE, Corn RM, Smith LM (1997)

Anal Chem 69:494813. Campbell CT, Kim G (2007) Biomaterials 28:238014. Shumaker-Parry JS, Campbell CT (2004) Anal Chem 76:90715. Shumaker-Parry JS, Zareie MH, Aebersold R, Campbell CT

(2004) Anal Chem 76:91816. Shumaker-Parry JS, Aebersold R, Campbell CT (2004) Anal

Chem 76:207117. Boozer C, Kim G, Cong S, Guan H, Londergan T (2006) Curr

Opin Biotechnol 17:40018. Nelson BP, Grimsrud TE, Liles MR, Goodman RM, Corn RM

(2001) Anal Chem 73:119. Goodrich TT, Lee HJ, Corn RM (2004) Anal Chem 76:617320. Goodrich TT, Lee HJ, Corn RM (2004) J Am Chem Soc

126:408621. Lee HJ, Li Y, Wark AW, Corn RM (2005) Anal Chem 77:509622. Wolf LK, Gao Y, Georgiadis RM (2007) J Am Chem Soc

129:1050323. Beusink JB, Lokate A, Besselink G, Pruijin G, Schasfoort R

(2008) Biosens Bioelectron 23:83924. Brockman J, Fernandez SM (2001) Am Lab 33:3725. Baggio R, Carven GJ, Chiulli A, Palmer M, Stern LJ, Arenas JE

(2005) J Biol Chem 280:418826. General Electric (2009) Biacore Life Sciences. http://www.

biacore.com/27. GWC Technologies (2009) GWC Technologies. http://www.

gwctechnologies.com/28. IBIS Technologies (2009) IBIS Technologies. http://www.ibis-spr.nl29. TOYOBO (2009) TOYOBO. http://www.toyobo.co.jp/30. GenOptics (2006) GENOPTICS—Genoptics. http://www.genop-

tics-spr.com31. Ebbesen TW, Lezec HJ, Ghaemi HF, Thio T, Wolff PA (1998)

Nature 391:66732. Gao HW, Henzie J, Odom TW (2006) Nano Lett 6:210433. Dintinger J, Degiron A, Ebbesen TW (2005) MRS Bull 30:38134. Yang J-C, Ji J, Hogle JM, Larson DN (2008) Nano Lett 8:271835. Ji J, O’Connell JG, Carter DJD, Larson DN (2008) Anal Chem

80:249136. Langmuir I, Schaefer VJ (1937) J Am Chem Soc 59:140637. Vroman L, Adams AL (1969) Surf Sci 16:43838. Lukosz W, Clerc D, Nellen PM, Stamm C, Weiss P (1991)

Biosens Bioelectron 6:22739. Gauglitz G, Brecht A, Kraus G, Nahm W (1993) Sens Actuators

B Chem 11:2140. Cunningham B, Li P, Lin B, Pepper J (2002) Sens Actuators B

Chem 81:31641. Lee M, Fauchet PM (2007) Opt Express 15:453042. Cunningham BT, Laing L (2006) Expert Rev Proteomics 3:27143. Cunningham BT (2008) Proc SPIE 6959:69591044. Chan LL, Gosangari SL, Watkin KL, Cunningham BT (2007)

Apoptosis 12:106145. Chan LL, Gosangari SL, Watkin KL, Cunningham BT (2008)

Sens Actuators B Chem 132:41846. Chan LL, Pineda M, Heeres JT, Hergenrother PJ, Cunningham

BT (2008) ACS Chem Biol 3:43747. Choi CJ, Cunningham BT (2007) Lab Chip 7:55048. SRU Biosystems (2008) SRU Biosystems. http://www.srubio-

systems.com/

Label-free technologies for quantitative multiparameter biological analysis 133

49. Vollmer F, Arnold S (2008) Nat Methods 5:59150. Armani AM, Kulkarni RP, Fraser SE, Flagan RC, Vahala KJ

(2007) Science 317:78351. Suter JD, White IM, Zhu H, Shi H, Caldwell CW, Fan X (2008)

Biosens Bioelectron 23:100352. Zhu H, White IM, Suter JD, Zourob M, Fan X (2008) Analyst

133:35653. White IM, Oveys H, Fan X, Smith TL, Zhang J (2006) Appl

Phys Lett 89:19110654. White IM, Oveys H, Fan X, Smith TL, Zhang J (2007) Proc

SPIE 6475:64750555. Ramachandran A, Wang S, Clarke J, Ja SJ, Goad D, Wald L,

Flood EM, Knobbe E, Hryniewicz JV, Chu ST, Gill D, Chen W,King O, Little BE (2008) Biosens Bioelectron 23:939

56. Mandal S, Akhmechet R, Chen L, Nugen S, Baeumner A,Erickson D (2007) Proc SPIE 6645:66451J

57. Mandal S, Erickson D (2008) Opt Express 16:162358. De Vos KM, Bartolozzi I, Bienstman P, Baets R, Schacht E

(2007) Proc SPIE 6447:64470K59. Michael CP, Borselli M, Johnson TJ, Chrystal C, Painter O

(2007) Opt Express 15:474560. Jin G, Tengvall P, Lundstrom I, Arwin H (1995) Anal Biochem

232:6961. Jin G (2008) Phys Status Solidi A Appl Mater Sci 205:81062. Landry JP, Sun YS, Guo XW, Zhu XD (2008) Appl Opt

47:327563. Wang ZH, Jin G (2003) Anal Chem 75:611964. Zhu X, Landry JP, Sun Y-S, Gregg JP, Lam KS, Guo X (2007)

Appl Opt 46:189065. Chamritski I, Clarkson M, Franklin J, Li SW (2007) Aust J

Chem 60:66766. Hanel C, Gauglitz G (2002) Anal Bioanal Chem 372:9167. Ozkumar E, Needham JW, Bergstein DA, Gonzalez R, Cabodi

M, Gershoni JM, Goldberg BB, Unlu MS (2008) Proc Natl AcadSci USA 105:7988

68. Birkert O, Gauglitz G (2002) Anal Bioanal Chem 372:14169. Birkert O, Tunnernann R, Jung G, Gauglitz G (2002) Anal Chem

74:83470. Kroger K, Bauer J, Fleckenstein B, Rademann J, Jung G,

Gauglitz G (2002) Biosens Bioelectron 17:93771. Mohrle BP, Kumpf M, Gauglitz GN (2005) Analyst 130:

163472. Mohrle B, Kohler K, Jaehrling J, Brock R, Gauglitz G (2006)

Anal Bioanal Chem 384:40773. biametrics (2009) biametrics—biomolecular interaction analysis.

http://www.biametrics.com/74. Zoiray Technologies (2009) Home. http://www.zoiray.com/75. Zhao M, Nolte D, Cho W, Regnier F, Varma M, Lawrence G,

Pasqua J (2006) Clin Chem 52:213576. Quadraspec (2007) Quadraspec—diagnostic testing for medical

research. http://www.quadraspec.com/77. Zhao M, Wang X, Lawrence GM, Espinoza P, Nolte DD (2007)

IEEE J Sel Top Quantum Electron 13:168078. Zhao M, Wang X, Nolte DD (2008) Opt Express 16:710279. Mace CR, Striemer CC, Miller BL (2008) Biosens Bioelectron

24:33480. Lisdat F, Schafer D (2008) Anal Bioanal Chem 391:155581. Daniels JS, Pourmand N (2007) Electroanalysis 19:123982. Patolsky F, Zheng G, Lieber CM (2006) Anal Chem 78:426083. Patolsky F, Lieber CM (2005) Mater Today 8:2084. Janata J (2004) Electroanalysis 16:183185. Patolsky F, Zheng G, Hayden O, Lakadamyali M, Zhuang X,

Lieber CM (2004) Proc Natl Acad Sci USA 101:1401786. Cui Y, Wei Q, Park H, Lieber CM (2001) Science 293:128987. Wang WU, Chen C, Lin KH, Fang Y, Lieber CM (2005) Proc

Natl Acad Sci USA 102:3208

88. Zheng G, Patolsky F, Cui Y, Wang WU, Lieber CM (2005) NatBiotechnol 23:1294

89. Stern E, Klemic JF, Routenberg DA, Wyrembak PN, Turner-Evans DB, Hamilton AD, LaVan DA, Fahmy TM, Reed MA(2007) Nature 445:519

90. Hahm J, Lieber CM (2004) Nano Lett 4:5191. Gao Z, Agarwal A, Trigg AD, Singh N, Fang C, Tung CH, Fan

Y, Buddharaju KD, Kong J (2007) Anal Chem 79:329192. Bunimovich YL, Shin YS, Yeo WS, Amori M, Kwong G, Heath

JR (2006) J Am Chem Soc 128:1632393. Patolsky F, Timko BP, Yu GH, Fang Y, Greytak AB, Zheng GF,

Lieber CM (2006) Science 313:110094. Stern E, Wagner R, Sigworth FJ, Breaker R, Fahmy TM, Reed

MA (2007) Nano Lett 7:340595. Fan Z, Ho JC, Jacobson ZA, Yerushalmi R, Alley RL, Razavi H,

Javey A (2008) Nano Lett 8:2096. Melosh NA, Boukai A, Diana F, Gerardot B, Badolato A, Petroff

PM, Heath JR (2003) Science 300:11297. Wang J (2005) Analyst 130:42198. Sinha N, Ma J, Yeow JT (2006) J Nanosci Nanotechnol 6:57399. Kojima A, Hyon CK, Kamimura T, Maeda M, Matsumoto K

(2005) Jpn J Appl Phys 44:1596100. Dong XC, Fu DL, Xu YP, Wei JQ, Shi YM, Chen P, Li LJ

(2008) J Phys Chem C 112:9891101. Kerman K, Vestergaard M, Tamiya E (2007) Anal Chem 79:6881102. Besteman K, Lee JO, Wiertz FGM, Heering HA, Dekker C

(2003) Nano Lett 3:727103. Tang X, Bansaruntip S, Nakayama N, Yenilmez E, Chang YL,

Wang Q (2006) Nano Lett 6:1632104. Koehne JE, Chen H, Cassell AM, Ye Q, Han J, Meyyappan M,

Li J (2004) Clin Chem 50:1886105. Dastagir T, Forzani ES, Zhang R, Amlani I, Nagahara LA, Tsui

R, Tao N (2007) Analyst 132:738106. Maehashi K, Katsura T, Kerman K, Takamura Y, Matsumoto K,

Tamiya E (2007) Anal Chem 79:782107. Pengfei QF, Vermesh O, Grecu M, Javey A, Wang O, Dai HJ,

Peng S, Cho KJ (2003) Nano Lett 3:347108. Yu X, Xu D, Lv R, Liu Z (2006) Front Biosci 11:983109. Darain F, Park D-S, Park J-S, Shim Y-B (2004) Biosens

Bioelectron 19:1245110. Wu Z-S, Li J-S, Deng T, Luo M-H, Shen G-L, Yu R-Q (2004)

Anal Biochem 337:308111. Chen X, Wang Y, Zhou J, Yan W, Li X, Zhu J-H (2008) Anal

Chem 80:2133112. Li C-Z, Liu Y, Luong JHT (2005) Anal Chem 77:478113. Dharuman V, Grunwald I, Nebling E, Albers J, Blohm L,

Hintsche R (2005) Biosens Bioelectron 21:645114. Yang L, Li Y, Erf GF (2004) Anal Chem 76:1107115. Long Y-T, Li C-Z, Kraatz H-B, Lee JS (2003) Biophys J 84:

3218116. Kafka J, Panke O, Abendroth B, Lisdat F (2008) Electrochim

Acta 53:7467117. Li X, Lee JS, Kraatz HB (2006) Anal Chem 78:6096118. Varshney M, Li YB (2007) Biosens Bioelectron 22:2408119. Maalouf R, Fournier-Wirth C, Coste J, Chebib H, Saikali Y,

Vittori O, Errachid A, Cloarec JP, Martelet C, Jaffrezic-RenaultN (2007) Anal Chem 79:4879

120. La Belle JT, Gerlach JQ, Svarovsky S, Joshi L (2007) AnalChem 79:6959

121. Du Y, Li BL, Wei H, Wang YL, Wang EK (2008) Anal Chem80:5110

122. Evans D, Johnson S, Laurenson S, Davies AG, Ko Ferrigno P,Walti C (2008) J Biol 7:3

123. Cooper MA (2006) Drug Discov Today 11:1068124. ACEA Biosciences (2008) ACEA Biosciences—xCELLigence

system,cell index,E-plate. http://www.aceabio.com/

134 A.J. Qavi et al.

125. Applied BioPhysics (2008) Applied BioPhysics, ECIS, cellmigration instrument. http://www.biophysics.com/

126. Heidel JD, Liu JY, Yen Y, Zhou B, Heale BS, Rossi JJ, BartlettDW, Davis ME (2007) Clin Cancer Res 13:2207

127. Sapper A, Reiss B, Janshoff A, Wegener J (2006) Langmuir22:676

128. Glamann J, Hansen AJ (2006) Assay Drug Dev Technol 4:555129. Zhu J, Wang XB, Xu X, Abassi YA (2006) J Immunol Methods

309:25130. Chanana M, Gliozzi A, Diaspro A, Chodnevskaja I, Huewel S,

Moskalenko V, Ulrichs K, Galla HJ, Krol S (2005) Nano Lett5:2605

131. Gimzewski JK, Gerber C, Meyer E, Schlittler RR (1994) ChemPhys Lett 217:589

132. Guo Z, Zhang Q, Dong F, Chen D, Xiong Z, Miao Z, Li C, JiaoB, Wu X (2007) Sens Actuators A Phys 137:13

133. Ivanov T, Gotszalk T, Grabiec P, Tomerov E, Rangelow IW(2003) Microelectron Eng 67–68:550

134. Lange K, Rapp BE, Rapp M (2008) Anal Bioanal Chem391:1509

135. Bardea A, Dagan A, Ben-Dov I, Amit B, Willner I (1998) ChemCommun 839

136. Okahata Y, Kawase M, Niikura K, Ohtake F, Furusawa H, EbaraY (1998) Anal Chem 70:1288

137. Wang J, Jiang M, Nilsen TW, Getts RC (1998) J Am Chem Soc120:8281

138. Ijiro K, Ringsdorf H, Birch-Hirschfeld E, Hoffmann S, SchilkenU, Strube M (1998) Langmuir 14:2796

139. Muratsugu M, Ohta F, Miya Y, Hosokawa T, Kurosawa S, KamoN, Ikeda H (1993) Anal Chem 65:2933

140. Niikura K, Nagata K, Okahata Y (1996) Chem Lett 10:863141. Sato T, Ishii M, Ohtake F, Nagata K, Terabayashi T, Kawanishi

Y, Okahata Y (1999) Glycoconj J 16:223142. Zampetti E, Pantalei S, Macagnano A, Proietti E, Di Natale C,

D’Amico A (2008) Sens Actuators B Chem 131:159143. Zhang B, Mao Q, Zhang X, Jiang T, Chen M, Yu F, Fu W (2004)

Biosens Bioelectron 19:711144. Ebara Y, Okahata Y (1993) Langmuir 9:574145. Fawcett NC, Craven RD, Zhang P, Evans JA (1998) Anal Chem

70:2876146. Su H, Thompson M (1995) Biosens Bioelectron 10:329147. Lavrik NV, Sepaniak MJ, Datskos PG (2004) Rev Sci Instrum

75:2229148. Ilic B, Czaplewski D, Zalalutdinov M, Craighead HG, Neuzil P,

Campagnolo C, Batt C (2001) J Vac Sci Technol B 19:2825149. Ziegler C (2004) Anal Bioanal Chem 379:946

150. Fritz J, Baller MK, Lang HP, Rothuizen H, Vettiger P, Meyer E,Guntherodt HJ, Gerber C, Gimzewski JK (2000) Science288:316

151. McKendry R, Zhang JY, Arntz Y, Strunz T, Hegner M, Lang HP,Baller MK, Certa U, Meyer E, Guntherodt HJ, Gerber C (2002)Proc Natl Acad Sci USA 99:9783

152. Arntz Y, Seelig JD, Lang HP, Zhang J, Hunziker P, RamseyerJP, Meyer E, Hegner M, Gerber C (2003) Nanotechnology14:86

153. Backmann N, Zahnd C, Huber F, Bietsch A, Pluckthun A, LangHP, Guntherodt HJ, Hegner M, Gerber C (2005) Proc Natl AcadSci USA 102:14587

154. Lam Y, Abu-Lail NI, Alam MS, Zauscher S (2006) Nano-medicine 2:222

155. Wu GH, Datar RH, Hansen KM, Thundat T, Cote RJ, MajumdarA (2001) Nature Biotechnol 19:856

156. Yue M, Stachowiak JC, Lim H, Datar R, Cote R, Majumdar A(2008) Nano Lett 8:520

157. Ilic B, Yang Y, Aubin K, Reichenbach R, Krylov S, CraigheadHG (2005) Nano Lett 5:925

158. Hwang KS, Lee JH, Park J, Yoon DS, Park JH, Kim TS (2004)Lab Chip 4:547

159. Lee JH, Hwang KS, Park J, Yoon KH, Yoon DS, Kim TS (2005)Biosens Bioelectron 20:2157

160. Gupta A, Akin D, Bashir R (2004) Appl Phys Lett 84:1976161. Johnson L, Gupta ATK, Ghafoor A, Akin D, Bashir R (2006)

Sens Actuators B Chem 115:189162. Park K, Jang J, Irimia D, Sturgis J, Lee J, Robinson JP, Toner M,

Bashir R (2008) Lab Chip 8:1034163. Capobianco JA, Shih WY, Yuan QA, Adams GP, Shih WH

(2008) Rev Sci Instrum 79:076101164. McGovern JP, Shih WY, Rest R, Purohit M, Pandya Y, Shih WH

(2008) Analyst 133:649165. Mukhopadhyay R, Lorentzen M, Kjems J, Besenbacher F (2005)

Langmuir 21:8400166. Cantion (2006) Cantion. http://www.cantion.com/167. BioScale (2008) BioScale:: biomolecular detection. http://www.

bioscale.com/168. Concentris (2009) Concentris—nanomechanical cantilever arrays

and sensors. http://www.concentris.ch/169. Vollmer F, Arnold S, Braun D, Teraoka I, Libchaber A (2003)

Biophys J 85:1974170. Gao Z, Agarwal A, Trigg AD, Singh N, Fang C, Tung CH, Fan

Y, Buddharaju KD, Kong J (2007) Anal Chem 79:3291171. Nanōmix (2007) Nanōmix—breakthrough detection solutions with

the nanoelectronic sensation technology. http://www.nano.com/

Label-free technologies for quantitative multiparameter biological analysis 135