Magneto-Optical Biosensing Platform Basedon Light Scattering from Self-AssembledChains of Functionalized Rotating MagneticBeadsSang Yoon Park,†,§ Hiroshi Handa,‡,§ and Adarsh Sandhu†,‡,§,*†Quantum Nanoelectronics Research Center, Tokyo Institute of Technology, Tokyo 152-8552, Japan, ‡IntegratedResearch Institute, Tokyo Institute of Technology, 2-21-1 Ookayama, Meguro-ku, Tokyo, Japan, and §Tokyo TechGlobal COE Program on Evolving Education and Research Center For Spatiotemporal Biological Network,Tokyo, 152-8552, Japan

ABSTRACT We describe a simple protocol for the rapid, highly sensitive, and quantitative measurement of the concentration ofbiomolecules in a solution by monitoring light scattered by self-assembled chains of functionalized superparamagnetic beads (SBs)rotating in the solution. A rotating external field (Hex) applied to an aqueous solution containing 250 nm diameter biotinylated SBsproduced linear chains of SBs rotating in phase with Hex due to magnetically induced self-assembly. At constant Hex, the addition ofavidin to the solution led to the formation of longer SB-chains than without the presence of avidin. The generation of longer SB-chainswas revealed by increases in the amplitude of the oscillating optical transmittance signal of the magnetic colloid solution. Monitoringchanges in the amplitude of the optical transmittance of the solution enabled quantitative determination of the concentration ofavidin added to the solution with a sensitivity of 100 pM (6.7 ng/mL) and a dynamic range of at least 3 orders of magnitude. Therotating chains acted as biomolecule probes and micromagnetic mixers, enabling detection of biomolecular recognition in less than30 s. This approach offers a rapid, highly sensitive, inexpensive, and homogeneous means for detecting biorecognition processes.

KEYWORDS Biosensing, magnetic colloids, magnetic nanoparticles, bioassay, point of care (POC) diagnostics

A wide range of biosensors based on biomolecularrecognition are being developed for testing safety offood, monitoring environmental pollutants, and clini-

cal diagnostics. Practical protocols must be highly sensitiveand molecule-selective, and have measurement times ofminutes and not hours. For example, enzyme-linked immu-nosorbent assays (ELISA) is an established procedure basedon the detection of changes in fluorescence when a probebinds to the desired target.1 However, this “inhomogeneous”approach is time-consuming because of multiple steps suchas amplifying and washing procedures, and not suitable forpoint for care applications. Recently, “magnetic-tags” com-posed of nano- and micrometer-sized superparamagneticbeads (SBs) with specific biomolecules immobilized ontotheir surfaces have been used for fast immunoassaying,2

DNA hybridization,3 and high-yield mRNA capture.4 Bio-sensing platforms based on the detection of magnetic-tagsinclude the use of Hall effect devices,5-7 giant magnetore-sistance spin-valves,8 magnetic tunnel junctions,9 and SQUIDmagnetometer.10,11 These electronic biosensing devices candetect several tens of picomolars of biomolecules within

several minutes.12 Now, in addition to high sensitivity,biosensing systems must be simple to operate, and inex-pensive to fabricate. However, the aforementioned “solidstate” based biosensors require expensive fabrication equip-ment such as electron lithography technology, and biologicalmolecules must be selectively immobilized onto function-alized surfaces of the solid state devices. Furthermore,complicated peripheral electronics is needed to extractextremely weak magnetic signals from environmental noise,and it is not possible to detect small densities of sub-250 nmdiameter magnetic-tags, which is important for acceleratingthe recognition rate, and improving accuracy of the results.

An alternative and more straightforward approach wasrecently reported by Baudry et al.13 who measured thedifferences in the intensity of light transmitted throughsolutions containing functionalized magnetic particles andtarget molecules, before and after the application of auniform external magnetic field. The detection limit of thismagneto-optical-based method is several tens of picomolar,where biomolecular recognition is detected by monitoringthe optical transmittance of so-called “magnetic bead dou-blets” due to linkage between ligands and receptors on thesurfaces of the magnetic beads in magnetic colloids. Con-firmation of biorecognition by this method consists of twosteps. First, magnetic beads are aggregated, which is ac-celerated by application of an external magnetic field (Hex)

* To whom correspondence should be addressed. E-mail: sandhu.a.aa@m.titech.ac.jp. Telephone number: 81-03-5734-2807. Fax number: 81-03-5734-2807.Received for review: 09/14/2009Published on Web: 12/28/2009

pubs.acs.org/NanoLett

© 2010 American Chemical Society 446 DOI: 10.1021/nl9030488 | Nano Lett. 2010, 10, 446-451

(typically, 20 mT for 5 min). The final step is detection oflight scattered by the magnetic bead doublets (formed viabiomolecular linkages) in the absence of the external fieldbecause turning the field-off causes the dissociation ofmagnetic beads, which do not have any biomolecular link-ages. Therefore, sufficient time must be allowed for diffusivemotion of probe magnetic beads and biomolecules in thesolution after the external field is removed. Thus, in orderto reduce the time taken for biomolecular recognition (the“recognition rate”), a large volume fraction (ø) of magneticbeads (ø ) 0.03%)13 and relatively large external magneticfields are necessary because the characteristic time requiredfor doublet formation is inversely proportional to ø and Hex

2,which results in the use of large concentrations of costlyprobe magnetic beads.14 Notably, if the concentration ofmagnetic beads is too large then light scattered from non-aggregated beads adds to the background noise, therebyreducing the detection sensitivity. Furthermore, the tempo-ral evolution of the average size of clusters follows a power-law and the exponent has a very small value of a fewminutes when ø is below a critical value because it takes alonger time for SBs to enter the ‘ballistic aggregation vol-ume’, wherein magnetic dipolar interaction with other SBsoccurs. Thus such diffusion-limited reaction rates are im-portant issues to resolve for not only practical microfluidicbiomoleculear recognition systems15,16 but also magneticcolloid based optical devices such as optical modulators,17

optical switches,18 tunable gratings.19 Therefore, improvingthe probability of particle-to-particle interaction in ultradi-luted magnetic solutions is essential to overcome the limitsof the slow diffusion-limited process.20

In this letter, we report on an alternative approach forhomogeneous, magneto-optical-based biomolecular rec-ognition protocol utilizing “rotating” chains of self-as-sembled superparamagnetic bead (SBCs) for enhancingthe recognition rate, thereby reducing the detection time,even in assays containing very small concentrations offunctionalized magnetic beads. First, we investigated themagneto-optical transmittance (MT) characteristics of asolution containing rotating chains of beads as a functionof the angle between the orientation of SBCs and incidentlight, and angular frequency (ω ) 2πf) of the rotating SBC,and quantified the relationship between the amplitude ofoptical transmittance and structure of SBCs. On the basisof our MT experiments, we demonstrate that subnano-molar concentrations of avidin can be detected by rotatingSBC functionalized with biotinylated beads within 30 s.In our protocol, the effect of mixing due to the rotatingSBCs enhances the probability that the biotinylated mag-netic beads interact with and capture the avidin added tothe solution.21,22 Thus, in our approach, SBC-induced mixingyielded extremely rapid detection times, using extremely smallquantities of nanometer sized magnetic probes.

Figure 1a illustrates the principle of the MT effect dueto changes in the structure and rotation of SBCs. In the

presence of Hex, magnetic dipolar interaction causes SBsto aggregate and form chains. The light scattered bynanometer-sized particles in colloids is well-described bythe Rayleigh-Gans theory. The intensity of light transmit-ted through the magnetic solution, depends on the sizeand the concentration of SBs. In particular, Rayleigh-Gansscattering for a cylindrical rigid objects depends on theangle between cylinder direction and light direction,cylinder length, and radius.23,24 Therefore, the light trans-mitted by SBCs depends on the angle (θ) between thedirection of SBC and the light, while isotropic SB singletsdo not give rise to angle-dependent light transmission. Wedefine the MT effect in terms of MT ratio (%) ) (Tmax -Tmin)/Tmax × 100. Here, Tmax and Tmin are the opticaltransmittance at θ ) 90 and 0°, respectively. An MT ratioof zero implies that SBCs behave as an optically isotropicmedium. Figure 1b shows the MT effect of SBCs under acontinuously rotating field, as used for our biomolecularrecognition protocol. A magnetic field rotating at a low-frequency, f, of a few hertz causes the SBCs to rotate atthe same frequency owing to the action of the magnetictorque. The rotation of SBCs is counteracted by thehydrodynamic force due to the viscosity of the solution,

FIGURE 1. Schematic illustration of the process of magneto-opticalbiomolecular detection by rotating superparamagnetic bead chainsin solution. (a) The light incident on the SBSs is scattered andthe intensity of transmitted light depends on the angle (θ) betweenthe SBC direction and light propagation path, which we refer toas the magneto-optical transmittance (MT) effect. (b) The averagelength of a rotating SBCs at a given hydrodynamics force (FH)depends only on the magnetic dipolar force between the beads (Fm).Incorporation of avidin results in stronger and longer biotinylatedSBCs due to the larger avidin-biotin conjugation force (FABC). Acombination of Brownian motion and the micromixing action of themagnetic chains leads to a larger amplitude of the oscillating signalof the MT effect; the longer the chains the larger the peak-to-valleyratio of the signal oscillation of rotating SBCs.

© 2010 American Chemical Society 447 DOI: 10.1021/nl9030488 | Nano Lett. 2010, 10, 446-451

eventually leading to deformation and rupture of SBCs.The dynamics of rotating SBCs in liquids can be describedin terms of the dimensionless Mason number (Mn),defined by

where FH is the hydrodynamic force, Fm is the magneticdipolar force, η is the viscosity of surrounding fluid, ω isthe angular velocity of the field, µ0 is the permeability offree space, and � is the magnetic susceptibility of SBs.25-28

The average number of SBs in a stable chain (N) isdescribed by N ≈ (1/Mn)0.5.23 Therefore, without avidin inthe solution, the MT ratio corresponding to the averagelength of biotinylated SBCs depends only on the magneticdipolar interaction for a given FH. When avidin is introducedinto the solution containing the SBCs, it is necessary toconsider not only Fm, but also the force due to avidin-biotininduced conjugation (FABC) of SBCs because avidin and biotinhave a high affinity for each other. Therefore, the Masonnumber becomes Mn ≡ FH/(Fm + FABC). Notably, experi-ments show that FABC is much greater than Fm betweennanometer- or micrometer-sized SBs.29 Therefore, this ad-ditional avidin-biotin conjugation contributes significantlyto strengthening the bead-to-bead linkages in the chains,which leads to a decrease in Mn, and consequently theaverage length of SBCs increases due to aggregation of chain-to-chain and/or chain-to-isolated SBs in the vicinity of rotat-ing chains, even if a small concentration of avidin is addedinto SBC solution. Furthermore, mixing, the other role of theSBCs, enables avidin to attach homogenously and quicklyonto biotinylated SBCs. Thus, quantitative bioanalysis isrealized by detecting the change in the MT ratio correspond-ing to enhanced avidin-biotin conjugation due to the addi-tion of avidin into the solution.

We examined the variation of the optical transmittanceof magnetic colloid solutions with θ, as shown in Figure 2a.The average diameter of biotinylated SBs was about 250 (50 nm, and they were covered with ∼6 × 106 biotin-groups/bead. A highly diluted SB solution of 19.85 pmol/L was usedin this study (see Materials in Supporting Information). Therotating Hex was generated by two pairs of coils, and the fieldstrength and frequency was adjusted by a two-channelfunction generator (see Experimental Setup in SupportingInformation). The wavelength of the incident collimated lightwas kept constant at 785 nm because wavelengths between700 to 800 nm exhibited the highest signal-to-noise ratio forour experimental configuration (see Figure S2 in SupportingInformation). The rotation frequency, f, and Hex were 0.1 Hzand 0.95 kA/m, respectively. Transmittance measurementswere taken for a period (t) of 15 s immediately after applyingHex. The optical transmittance of SBC solution oscillated, and

the maxima (Tmax) and the minima (Tmin), were observed forthe normal (θ ) nπ/2, n ) odd number) and parallelconfigurations (n ) even number), respectively. Here, wedefine ∆T as the difference between Tmax and Tmin. Video ofoptical microscopic images were used to quantify the dy-namics of SBC under rotating fields. Microscopic imageswere obtained for higher concentrations of SPB (83.1 pmol/L) because it was difficult to clearly image chain motion ofSBCs for lower concentrations using our microscope. Imageswere captured for the same f and Hex. Figure 1 panels b andc are images of rotating SBCs at θ ) π/4 and π/2, respec-tively. The direction of all the chains is parallel to Hex. Wedid not observe the deformation or rupture of SBC underthese conditions. We also investigated the temporal evolu-tion of the transmittance of a biotinylated SBC solution in

Mn ≡FH

Fm≈ 16ηω

µ0�2H2(1)

FIGURE 2. (a) The angular dependence of the light transmittedthrough the magnetic colloid solution at f ) 0.1 Hz, where thehighest (Tmax) and the lowest transmission (Tmin) occurred at θ ) nπ/2(n ) odd) and n ) even, respectively. (b) Representative opticalmicroscope images of SBCs at (c) θ ) π/4, and (d) π/2. (c) Temporalevolution of magneto-optical transmittance as a function of rotationfrequency (f).

© 2010 American Chemical Society 448 DOI: 10.1021/nl9030488 | Nano Lett. 2010, 10, 446-451

rotating fields as a function of the rotation frequency (Figure2d). The applied Hex was 0.011 kA/m in region I and III, and0.95 k Am1- in region II. In region I, there was no significantdifference of the optical transmittance. In region II, themagnitude of ∆T gradually increased over time. The averagelength of the SBCs (l) in diluted magnetic colloids is governedby the population of SBs inside the capture volume of otherSBs or SBCs, which is defined by the ratio of the maximummagnetic attraction energy for alignment of SBs to that of

the thermal energy.14 When Hex was increased to 0.95 kA/m, the chain length increased dramatically due to a largerFm, resulting in the capture of SBs within the capture volume.Thus, the magnitude of ∆T, which is proportional to averagelength of SBCs, increased rapidly. Then, when Hex wasreduced 0.011 kA/m in region III, the transmittance returnedto approximately the same magnitude as the initial value inregion I. Notably, these results imply that the sedimentationof the SBCs and SB was negligible. The Peclet number (Pe)given by Used a/D, where a is the diameter of a magnetic beadand D is diffusion coefficient, gives the ratio betweensedimentation effects and Brownian motion. In our experi-ments, Pe was approximately 0.018, indicating that Brown-ian motion dominated over sedimentation. The other inter-pretation of these findings is that biotinylated SBC solutionunderwent reversible aggregation and dispersion processes,and that attractive forces that were not related to themagnetic field, such as electrostatic effects, were negligibleand thus the dynamics of biotinylated SBCs can indeed bedescribed by the Mason number, composed of hydrody-namic and magnetic dipolar forces.

Figure 3 shows the variation of the MT ratio on therotation frequency for magnetic fields from 0.24 to 0.95 kA/m. The MT ratio decreased gradually with increasing f (Figure3a). According to previous theoretical and experimentalstudies, flexible linear SBCs in rotating magnetic fieldsdeform into S-shape structures with increasing f because ofthe balance required between FH and Fm.25-28 Incrementsin Hex cause Fm to strengthen, which gives rise to a linearincrease in the number of beads forming the chain, N. Figure3b shows the MT ratio versus Mn at f ) 0.1 Hz. Notably, MTratio was linearly proportional to (1/Mn)0.5, which impliesthat the MT ratio is also proportional to N. This resultindicates that the Mn model is valid for describing the MTratio of SBCs. Therefore, lengthened SBCs correspond to areduction of the value of Mn, which results in an increase inthe MT ratio. The slope of the graph of Figure 3 was about

FIGURE 3. The dependence of the MT effect on the rotationfrequency. (a) Plot of MT ratio vs f for several values of Hex. The MTratio decreases with increasing f for all values of Hex. (b) Plot of MTratio vs (1/Mn)0.5. N is proportional to (1/Mn)0.5.

FIGURE 4. Change in the MT ratio of rotating SBCs (f ) 0.1 Hz) due to incorporation of avidin in the magnetic colloid under an external fieldof Hex ) 0.95 kA/m. (a) Temporal evolution of magneto-optical transmittance in the absence of avidin (red line) and that related to the additionof 0.15 nM of avidin (upper figure) and for an avidin concentration of 2.54 nM (lower figure). (b) Temporal evolution of the MT ratio in region-II. (c) The difference of the MT ratio after 15 min as a function of avidin concentration.

© 2010 American Chemical Society 449 DOI: 10.1021/nl9030488 | Nano Lett. 2010, 10, 446-451

0.7, which provides quantitative information about thedegree to which changes in N due to Fm contribute to theMT ratio.

On the basis of our aforementioned results, the MT ratiomeasurements for the magnetic colloid solutions with vari-ous concentrations of avidin were carried out for 18 min.First, we prepared two samples with the same SB concentra-tion (Sample-A, and Sample-B). After optical transmittancemeasurements of both samples, different concentrations ofavidin were added to solutions of SBs; the avidin concentra-tions were 0.15 nM (Sample-A-a), and 2.54 nM (Sample-B-a). The experimental procedure was the same as that shownin Figure 2d. As expected, the magnitudes of ∆T for Sample-Aand Sample-B were identical (∆T ) 1.14 ( 0.014% at t )900 s) even though the absolute values of both transmit-tances were slightly different (Sample-A: T ≈ 8.1%, Sample-B: T ≈ 8.0%), and they increased slowly during the 15 minin region II. And in region III, the MT ratio returned to thesame value as that in region I. It should be noted that in thecase of Sample-A-a and Sample-B-a, the value of ∆T overthe whole of region II was higher than that of nonavidincounterpart samples. As shown in Figure 5b, the MT ratiosof avidin-incorporative samples (blue) increased faster andwere larger than those of corresponding nonavidin solutions(red) over longer time intervals. Here, we define thesechanges in the MT ratio, as the ∆MT ratio. We also examinedmagnetic colloid solutions with different concentrations ofavidin (Figure 4c). In this experiment, the ∆MT ratio wasproportional to the concentration of avidin in the range of afew hundreds pM to several tens of nanomolars. Notably,the noise level was very much lower than signal level, eventhough the absolute value of the ∆MT ratio was not excep-tionally large. These results clearly show the existence ofbiomolecular recognition processes. The ∆MT ratios areexplained by a modified Mn model. The force, FABC due tothe incorporation of avidin onto biotinylated SBCs causedthe length of the chains, l to increase, leading to an enhancedMT ratio. The effect of FABC on l can be estimated from theslope in Figure 3b. The increment in l (∆l ) a∆N ≈ a(1/Mn)0.5

) a∆MT ratio(%)/0.7(%/µm)), where a is the diameter ofSBs, at an ∆MT ratio ) 4.61% (Sample-B-a) was determinedto be about 1.64 µm. This difference is too small to beobserved by conventional optical microscopy. To estimatethe detection range in a bioassay, we also conducted experi-ments for higher concentrations of avidin (see Figure S3 inSupporting Information). We observed not only higher a∆MT ratio (∼46%) but also a decrease in the optical trans-mittance. The larger volume of SBCs gives rise to increasedRayleigh scattering, leading to lower optical transmittance.Therefore, we believe that high concentrations of avidincontributed not only to lengthening the SBCs but also toenlarging the average size of agglomerated SBs. In regionIII, an oscillation of transmittance was observed, but nosuch signal was seen in region I. This result implies that asufficient FABC due to biotin-avidin biomolecular recognition

enables SBC to be sustained even in a very small magneticfield of 0.011 kA/m.

To verify the possible effects of nonspecific aggregationon the MT ratio signal, we used 0.1 mM of bovine serumalbumin (BSA) as blocking agents for nonspecific-back-ground tests. (see Figure S4 in Supporting Information).Importantly, we did not observe any significant change inthe MT ratio due to the presence of BSA in tests solutionscompared to ones without BSA. In contrast, after addingavidin (42 nM) into a SB-BSA solution, the MT ratio changeddramatically. These results imply that the effect nonspecificaggregation on the MT ratio was negligibly small, and thatenhanced MT ratio was due to avidin-mediated SBC (specificaggregation).

We also investigated the response time for detectingbiorecognition processes. As stated earlier, we predicted thatrotating SBCs could act as micromixers and reduce thebiorecognition time. The ∆MT ratios for 0.89 nM of avidinwere also taken for f from 0.05 to 1 Hz (Figure 5a). All the∆MT ratios increased and did not saturate over the measure-ment time. However, for the initial 3 min, the ∆MT ratioincreased quicker with increasing f as shown in the inset ofFigure 5a. The characteristics of ∆MT ratio were studiedquantitatively by fitting the related exponent formula

FIGURE 5. (a) ∆MT ratio of SBCs solution containing 0.89 nM ofavidin as a function of the rotation frequency of the magnetic field.Inset shows ∆MT during the initial 3 min. (b) The behavior of SBCsas micromixers for a characteristic time (τ) vs f. The magnitude of τwas determined by fitting to eq 2. Lower values of τ indicate shorterresponse time for detecting the change in MT ratio associated withavidin-biotin conjugation.

© 2010 American Chemical Society 450 DOI: 10.1021/nl9030488 | Nano Lett. 2010, 10, 446-451

where, ∆MT0 is the magnitude at t ) 30 s (starting point inregion II), ∆MTi and ∆MTf are the values during the initialinterval and during over long periods, respectively. τi and τf

are the characteristic time for these processes. All the valuesare tabulated in Table-S1 (see Supporting Information). Thefirst and the third terms over all values of f are similar, butthe parameters obtained in the second term are different.These results indicate that the frequency used for rotatingthe SBCs affects the initial response time, where the τi at f) 0.05 Hz was determined to be about 117 s. Interestingly,τi was drastically reduced to about 15 s by increasing f to 1Hz (Figure 5b). The values of τi represent the biomolecularrecognition time accelerated by the rotating SBCs mixers.These results are proof that rotating SBCs act as micromixersleading to significant reductions in the time required forcapturing avidin molecules in solution.

In summary, we demonstrated “nonsolid state” basedmagneto-optical bioassay read-out method utilizing self-assembled SBCs in ultradiluted magnetic colloid. The rela-tionship between the MT ratio and the mean length of SBCswas understood in terms of the Mason number. The effectof avidin-biotin conjugation on the MT ratio and the meanlength of SBC was described by the modified Mason model.On the basis of the study of the dynamics of SBCs, weshowed that the MT effect of biotinylated SBCs is a rapid andhighly sensitive bioanalysis protocol for molecular recogni-tion. Rotating the SBC reduced the time for avidin to reactwith functionalized magnetic beads that formed the SBC. Inthis way, we detected a wide range of avidin concentrationsfrom sub-nanomolar to several tens nanomolars within 30 s.

Acknowledgment. This work has been partly supportedby Ministry of Education, Culture, Sports, Science, andTechnology, Japan.

Supporting Information Available. Supporting Informa-tion is provided to describe detailed materials and methodsused in this work. This material is available free of chargevia the Internet at http://pubs.acs.org.

REFERENCES AND NOTES(1) Murakami, A.; Nakaura, M.; Nakatsuji, Y.; nagahara, S.; Tran-

Cong, Q.; Makino, K. Nucleic Acids Res. 1991, 19, 4097–4102.(2) Hayes, M. A.; Polson, N. A.; Phayre, A. N.; Garcia, A. A. Anal.

Chem. 2001, 73, 5896–5902.(3) Fan, Z. H.; Mangru, S.; Granzow, R.; Heaney, P.; Ho, W.; Dong,

Q.; Kumar, R. Anal. Chem. 1999, 71, 4851–4859.(4) Pamme, N. Lab Chip 2006, 6, 24–38.(5) Sandhu, A.; Kumagai, Y.; Lapicki, A.; Sakamoto, S.; Abe, A.;

Handa, H. Biosens. Bioelectron. 2007, 22, 2115–2120.(6) Sandhu, A.; Sanbonsugi, H.; Shibasaki, I.; Abe, M.; Handa, H. Jpn.

J. Appl. Phys. 2004, 43, 868.(7) Togawa, K.; Sanbonsugi, H.; Sandhu, A.; Abe, M.; Narimatsu, H.;

Nishio, K.; Handa, H. Jpn. J. Appl. Phys. 2005, 44, 1494.(8) Graham, D. L.; Ferreira, H. A.; Freitas, P. P.; Cabral, J. M. S.

Biosens. Bioelectron. 2003, 18, 483–488.(9) Grancharov, S. G.; Zeng, H.; Sun, S.; Wang, S. X. J. Phys. Chem. B

2005, 109, 13030–13035.(10) Lee, S.; Myers, W. R.; Grossman, H. L.; Cho, H. M.; Chemla, Y. R.;

Clarke, J. Appl. Phys. Lett. 2002, 81, 3094–3096.(11) Stromberg, M.; Goransson, J.; Gunnarsson, K.; Nilsson, M.;

Svedlindh, P.; Strømme, M. Nano Lett. 2008, 8, 816–821.(12) Xu, L.; Yu, H.; Akhras, M. S.; Han, S.-J.; Osterfeld, S.; White, R. L.;

Pourmand, N.; Wang, S. X. Biosen. Bioelectron. 2008, 24, 99–103.(13) Baudry, J.; Rouzeau, C.; Boubault, C.; Robic, C.; C-Tannoudji;

Koenig, A.; Bertrand, E.; Bibette, J. Proc. Natl. Acad. Sci. U.S.A.2006, 103, 16076–16078.

(14) Promislow, J. H. E.; Gast, A. P.; Fermigier, M. J. Chem. Phys. 1995,102, 5492–5498.

(15) Liu, R. H.; Ryu, K. S.; Liu, C. J. Microelectromech. Syst. 2002, 11,462–469.

(16) Yang, Z.; Matsumoto, S.; Goto, H.; Matsumoto, M.; Maeda, R.Sens. Actuators, A 2001, 93, 266–272.

(17) Chieh, J. J.; Yang, S. Y.; Horng, H. E.; Hong, C.-Y.; Yang, H. C.Appl. Phys. Lett. 2007, 90, 133505–133507.

(18) Horng, H. E.; Chen, C. S.; Fang, K. L.; Yang, S. Y.; Chieh, J. J.;Hong, C.-Y.; Yang, H. C. Appl. Phys. Lett. 2004, 85, 5592–5594.

(19) Pu, S.; Chen, X.; Chen, L.; Liao, W.; Chen, Y.; Xia, Y. Appl. Phys.Lett. 2005, 87, No. 021901-021903.

(20) Fermigier, M.; Gast, A. P. J. Colloid Interface Sci. 1992, 154, 522.(21) Calhoun, R.; Waskowsky, R.; Phelan, P.; Carcia, A.; Hayes, M.;

Vuppu, A. Lab Chip 2005, 5, 1075–1082.(22) Rida, A.; Gijs, M. A. M. Anal. Chem. 2004, 76, 6239–6246.(23) van de Hulst, H. C. Light Scattering by Small Particles; Dover: New

York, 1981.(24) Lednei, M. F.; Pinkevich, I. P.; Yu, V.; Sluckin, T. J. J. Mol. Liquid.

2001, 92, 139–146.(25) Melle, S.; Martin, J. E. J. Chem. Phys. 2003, 118, 9875–9881.(26) Cebers, A.; Javaitis, I. Phys. Rev. E 2004, 69, No. 021404-021411.(27) Petousis, I.; Homburg, E.; Derks, R.; Dietzel, A. Lab Chip 2007, 7,

1746–1751.(28) Klingenberg, D. J.; Ulicny, J. C.; Golden, M. A. J. Rheol. 2007, 51,

883–893.(29) Wong, J.; Chilkoti, A; Moy, V. T. Bio. Eng. 1999, 16, 45–55.

∆MT ratio ) ∆MT0 + ∆MTi(1 - e-t/τi) +

∆MTf(1 - e-t/τf) (2)

© 2010 American Chemical Society 451 DOI: 10.1021/nl9030488 | Nano Lett. 2010, 10, 446-451