Received 23 July 2012, in final form 23 August 2012Published 4 October 2012Online at stacks.iop.org/JMM/22/115024
AbstractWe report development and application of 3D structured nano-microfluidic devices that wereproduced via soft lithography with poly(dimethylsiloxane). The procedure does not rely onhazardous or time-consuming production steps. Here, the nanochannels were created bychannel-spanning ridges that reduce the flow height of the microchannel. Several realizationsof the ridge layout and nanochannel height are demonstrated, depicting the high potential ofthis technique. The nanochannels proved to be stable even for width-to-height aspect ratios of873:1. Additionally, an application of these submicrometer structures is presented with a newtechnique of a dielectrophoretic mobility shift assay (DEMSA). The DEMSA was used todetect different DNA variants, e.g. protein–DNA-complexes, via a shift in(dielectrophoretically retarded) migration velocities within an array of nanoslits.
(Some figures may appear in colour only in the online journal)
1. Introduction
Applications with nanochannels are of considerable interestdue to their favorable size-based characteristics. The surface-to-volume ratio becomes one of the most prominent aspectswhen typical length scales are in the micrometer or nanometerrange. As a consequence, the adsorption rates dramaticallyincrease. This is relevant in applications for chemicaland biological detection, when surface-immobilized receptormolecules are used [1]. In contrast to receptor-based detection,label-free detection of biomolecules is of special interest topreserve biological activity. Here, nanochannels can serveas Coulter counter devices for label-free detection of DNAor proteins bound to DNA strands [2–5]. Therefore, twoseparate regions are connected by a nanochannel where anelectric voltage is applied. The current is monitored and fora biomolecule entering the nanochannel the current changescharacteristically [2–5].
Besides detection of macromolecules, analytical orpreparative separation is another field of research innanofluidic devices. In 2000, Han et al presented separationof DNA based on an entropic ratchet in a nano-microfluidicdevice [6]. Regtmeier et al used an overlapping Debyelayer in a nanofluidic channel to separate nanobeads [7]. Asrecently reviewed, structuring of microfluidic channels with
(nanofluidic) constrictions allows non-invasive manipulationvia dielectrophoresis (DEP), e.g. for analytical separation [8].DEP is the migration of a polarizable particle withinan inhomogeneous electric field �E. The dielectrophoreticpotential WDEP scales quadratically with the electric field,WDEP = − 1
2α�E2, where α denotes the electric polarizability.So in the vicinity of nanochannels with sharp and well-defined edges, the dielectrophoretic potential increasesdramatically, opening the field of new methods for trappingand manipulation of bioanalytes, like small DNA fragmentsor proteins [8]. Here, we present a new technique toinvestigate/detect DNA-complexes by exploiting DEP atnanoslits.
Apart from the analytical and preparative applications,nanochannels yield insights into new physical phenomena.For example the anomalous I–V characteristics, based on theDebye layer overlap in nanochannels, is known for severaldecades but the underlying physics is still poorly understood[9]. Therefore, the wide field of using small nanochannels innano- and microfluidics is of very high and actual interest ofresearch.
Up to now, the precise and reproducible fabrication ofnanochannels is challenging. Most of the nanochannels arecustom made by chemical or physical etching [3]. Here, wepresent a nanofluidic device that is monolithically fabricated
J. Micromech. Microeng. 22 (2012) 115024 M Viefhues et al
(a )(c )
(b)
Figure 1. Device. (a) Masterwafer with SU-8 negative relief structure mold and two layers of 40 μm hard PDMS and 1 mm PDMS (not toscale). (b) Side view of chip with reservoirs and glass coverslip. (c) Zoomed channel layout to the region of multiple ridges used for theDEMSAs. For the DEMSA experiments, the nanoslit had a height of 580 nm over a channel width of 200 μm.
by soft lithography with poly(dimethylsiloxane) (PDMS). Softlithography allows for a fast chip development in one day andsubsequent chip production within a few hours, without theuse of hazardous acids or bases or time-consuming physicaletching like focus-ion-beam etching [10, 11].
Here, PDMS was used because of its dedicated materialproperties. It is biocompatible, gas permeable and electricallyinsulating and exhibits remarkable mechanical and chemicalstability [12, 13]. Additionally, PDMS is transparent in thevisible spectrum, 300–700 nm [14], allowing monitoring ofdye-labeled molecules within the nano-microfluidic channel.By adding carbon black particles to the PDMS, Hellmich etal could demonstrate a native label-free UV-LIF detection ofamino acids and proteins [15]. Last but not the least, PDMSchips are very cheap with an averaged cost of US$ 0.47 perdevice, with respect to the raw material costs and effort it takesto produce the device [16].
In this paper, first, we present the fabrication ofnanochannels by soft lithography with PDMS, followed bythe characterization of the channel dimensions and subsequentpassability experiments. We could demonstrate the fabricationof stable nanochannels with a maximum width-to-heightaspect ratio of 873:1, which was an improvement by a factorof 14.5 compared to earlier published papers [17].
With these miniaturized devices it became possibleto detect DNA-complexes gel-free and virtually label-freeby DEP in an array of nanoslits. More precisely, twovariants of DNA were subsequently driven through the arrayand dielectrophoretically retarded. By the analysis of themigration velocity, a shift between the DNA variants could bedetermined. The new method was termed as dielectrophoreticmobility shift assay (DEMSA) in analogy to the well-established technique of electrophoretic mobility shift assays[18]. DNA of different lengths and DNA-complexes withmolecules of interest, either the cancer drug actinomycin D
or Escherichia coli RNA polymerase core enzyme (RNAP),were investigated with the DEMSA.
Since our microfluidic methodology does not rely onclassical filtration/separation, shear-force–associated analytedeterioration could be reduced to a minimum. Togetherwith the fact that our approach is inherently independent oftoxic agents, future applications in molecular biotechnologyand medicine can be implemented in more compliant waywith respect to the concerns of safety regulation in drugadministration.
2. Methods
2.1. Fabrication of the device
The microfluidic devices, consisting of a crossinjector and oneor more insulating, channel-spanning ridges (see figure 1),were fabricated via soft lithography. Briefly, a masterwaferwas developed, consisting of the negative relief structure of thechannel design. In order to produce an inverted structure, liquidpolymer was cast over the masterwafer, thermo-cured andpeeled off. Afterward, the structures were cut out and reservoirholes were punched at the channel ends. After assemblingwith a PDMS-coated coverslip, the chips were filled withbuffer [19].
The negative relief structure of the channel layout wasbuild on a silicone wafer (CrysTec, Germany) by two-stepcontact lithography. Before, the silicone wafer was purified inperoxymonosulfuric acid for 5 min twice and rinsed afterwardwith deionized water. In the first lithographic step, the heightof the nanoslit was defined. Therefore, SU-8(50) photoresist(Microresist, solid fraction 69%, Germany) was dilutedwith GBL thinner (Microresist, Germany) due to supplier’sinformation and spin coated (Spincoater ST147, Convac,Germany), with solid fraction and spincoating parametersaccordingly to constriction height (see table 1). After a
2
J. Micromech. Microeng. 22 (2012) 115024 M Viefhues et al
Table 1. SU-8 layer thickness due to the spincoating speed (for 30 s)and solid fraction of photoresist.
pre-exposure bake, the photoresist was illuminated througha chromium photomask (Delta Mask, The Netherlands) ina mask aligner (MJB3, Suss MicroTec, Germany). Afterpostbake, the photoresist was developed in MR-dev 600(Microresist, Germany) for 1 min, rinsed with acetone (p.a.)and isopropyl alcohol (p.a.) (both obtained from VWR,Germany) and dried with nitrogen afterward. The height ofthe structure was determined in a DekTak profilometer 3030ST (Stanford Nanofabrication Facility Equipment, USA).
The second layer of SU-8 photoresist consists of 52%solid fraction for all the devices used; it was spin coatedwith 3000 rpm for 30 s, resulting in about 5 μm structureheight. Again, a pre-exposure bake was performed. Afterward,the photoresist was illuminated (for 7 s) through a secondchromium photomask, carefully aligned in the maskalignerusing positioning crosses as references. The photoresistwas developed for 1 min after post-exposure bake. Thesecond layer defines the layout of the ridge and the freechannel height (see figure 1(a)). After terminal hardbake,the masterwafer was silanized with tridecafluor-1,1,2,2-tetra-hydrooctyltrichlorsilane (Merck, Germany). All baking andillumination procedures (i.e. time and temperature) wereaccording to supplier’s information.
A double layer of hard PDMS (h-PDMS) [20],3.4 g vinylmethylsiloxane-dimethylsiloxane-trimethylsiloxyterminated copolymer, 18 μl platin-divinyltetramethyl-siloxane (both from ABCR, Germany)), 50 μl 2, 4, 6,8-tetramethyl-2,4,6,8-tetravinylcyclo-tetrasiloxane (Sigma-Aldrich,Germany), and Sylgard 184 (PDMS), 6.8 g ofpolymer and 0.68 g of linker (both from Dow Corning, USA),turn out to be successful for stable molding of the device. Theh-PDMS was spin coated at 1500 rpm for 10 s and placed ona hot plate (HAT-303D, AVT-Technology Munich, Germany)at 65 ◦C for 15 min. Afterward, the liquid PDMS was pouredon the wafer and cured at 65 ◦C for 40 min. The channelstructures were cut out after peel-off and reservoir holes werepunched at the channel ends. After cleaning successivelyin acetone, ethanol and deionized water, the PDMS slabsand PDMS-coated coverslips were dried with nitrogen andoxidized in a home-made plasma chamber [21]. The PDMSslabs were assembled with the coverslips subsequent to theplasma oxidization. After 30 min, the chip could be filled withrunning buffer. We used 1 mM phosphate buffer containingn-dodecyl-β-D-maltoside and methyl cellulose (DDM-MC)(Sigma, Germany) [22], and incubated for 30 min before theexperiments were performed.
DEP is referred to as the migration of a polarizableparticle within an inhomogeneous electric field. Thenecessary inhomogeneous electric fields can be generatedat insulating structured elements (obstacles), inter alia, alsocalled insulating or electrodeless dielectrophoresis (eDEP)[8, 23–26]. The resulting force acting on a particle withpolarizability α can be written as �FDEP = α(�E · ∇)�E, withthe electric field �E. For a spherical, homogeneous particle, thepolarizability scales cubical to the particle radius [8]. As aconsequence, manipulation of nano-objects, like small DNAor proteins via eDEP, requires the development of challengingdevices with a rather large electric field gradient. Here, weused insulating ridges (figure 1(c)) creating nanoslits with aheight of 580 nm in order to generate adequate electric fieldgradients for successful manipulation of small DNA molecules(less than 7.0 kbp). The electric field at a defined regiondepends in a distinct manner on the channel width and height.Thus, for fixed microchannel dimensions, the width-to-heightaspect ratio allows a characterization of the dielectrophoreticmanipulation at a nanoslit.
Experimentally, a superposition of dc and ac voltages,U (t) = Udc + Uaccos(ωt), was used to move the DNAmolecules electrophoretically through a dielectrophoreticpotential landscape. The dc voltages move the DNA dueto electrophoresis, whereas the ac voltage generates adielectrophoretic potential at each nanoslit. For appropriateac and dc voltages the dielectrophoretic force overcomes theelectrophoretic motion, i.e. the molecules are (temporarily)dielectrophoretically trapped [21, 27] and retarded. The acvoltages were set such that the trapped molecules could escapedue to thermal energy. In consequence, the migration velocitywas decreased.
Theoretically, the mean lifetime τ within a DEP trappotential can be described by the inverse Kramers rate k, τ =1k ∝ exp
(WDEPkBT
)(for details refer to [21]). The dielectrophoretic
potential WDEP = − 12α�E2 is proportional to the polarizability
of the DNA. For DNA molecules, the polarizability dependson the molecules length (see [28] for details) as well as oncharged molecules that intercalate or bind to the DNA strand[29]. Accordingly, the mean trapping time τ varies for differentDNA lengths and DNA-complexes. So, for the investigationof the migration velocity of different DNA variants in an arrayof insulating obstacles a shift could be detected between thevariants (see figure 4). We termed this new technique DEMSA.The DEMSA was exploited to distinguish between differentspecies of DNA or DNA-complexes. Therefore, analyteswere subsequently driven through an array of ridges and themigration velocities were determined by the time needed forpassing a defined length for each analyte of interest.
Upon introducing a new technique, a comparison withestablished techniques is required. Thus, the DEMSA wascompared with electrophoretic mobility shift assay (EMSA),which is the most used technique to investigate DNA-complex formation and determine dissociation rates. InEMSA, electrophoretic migration velocities are investigated todetect DNA-complex formation or investigate binding kinetics[18]. In contrast to the new technique of DEMSA, where
3
J. Micromech. Microeng. 22 (2012) 115024 M Viefhues et al
the analytes are subsequently driven, for EMSA experimentstypically two lanes in an agarose gel are used, one with pureDNA as a reference and the second lane with the DNA andthe anticipated binding partner. Due to higher retention ofone compound a shift is observed, i.e. one analyte migratesmore slowly and a complex formation is assumed [18]. Theselectivity criterion is the pore size of the agarose gel that isadjustable only preliminary to the experiment by the agaroseconcentration, whereas for the DEMSA the selectivity isadjustable by the applied ac voltages and is therefore easy toadapt. According to EMSAs, DEMSAs can be used to detectcomplex formation or binding constants.
2.3. Experimental procedure
The passability for various nanoslit heights and layouts (seethe results) was checked. Additionally, DEMSAs as well asEMSAs were performed with DNA and DNA-complexes. Inorder to check for unspecific adsorption of polystyrene beadsor DNA to the nanoslit surface, the nanoslits were inspectedafter experiments. The inspection revealed that no analyteadsorbed to the surface of the nanoslit, supporting the conceptto prevent unspecific adsorption by DDM-MC [22].
For the passability experiments we used fluorescentlabeled, carboxylated polystyrene beads (Invitrogen, USA)with diameter of 500 nm for nanoslit heights of 510 nmto 670 nm and beads with diameter of 100 nm for nanoslitheights of 180 nm and 125 nm. The DEMSA experimentswere performed with 6.0 kbp DNA (MBBL GmbH, Germany)and 2.686 kbp DNA (homemade). YOYO-1 (Invitrogen, USA)was used to fluorescently label the DNA molecules, 1 YOYO-1per ten DNA base pairs. For the DNA-complexes actinomycinD (MW 384 kDa, ratio of one ACTD per five DNA base pair) orE. coli RNAP core Enzyme (Epicentre Biotechnologies, USA,MW 389 kDa) were added (0.1 mg DNA, 0.5 mg polymerase)and incubated on a Vortexer (0-level) for 2 h. For all chipexperiments the DNA was diluted to 10 pM concentrationwith running buffer.
An inverted fluorescent microscope (Axiovert 200, Zeiss,Germany) was used with a CCD sensicam qe (PCO, Germany).A 100-fold oil immersion objective (Zeiss, Plan Neofluar,NA 1.3) was used for experiments with DNA and 100 nmpolystyrene beads, whereas for experiments with 500 nmpolystyrene beads a 40-fold objective was used. The chip wasmounted on a motorized x/y-stage (99S008, Ludl ElectronicProducts, USA). DaVis 6.2 was used for image acquisition(10 frames s−1, eightfold binning) and ImageJ 1.43 for imagepost-processing. For electric contacting and to enlarge thereservoirs a plexiglass with reservoirs and platinum electrodeswas placed on the chip. The dc voltages were generated withpower supplies from FUG (HCL 14-12500, Germany). ACvoltages were generated with a DS 345 function generator fromStanford Research Systems, USA, and 100-fold amplified withan AMS-1B30 amplifier (Matsusada Precision, Japan). Theapplied voltages and frequencies were controlled via custom-made LabView 6i programs.
For the passability experiments the polystyrene beadswere electrophoretically driven by applying dc voltages, about
Table 2. Aspect ratios of different ridge layouts (cf figure 3) due tonanoslit width and height. All channels proved to be passable bynanobeads of appropriate size (cf figure 3).
−15 V, to channels 1, 3 and 4; channel 2 was grounded (seefigure 1(c)). Fluorescence image sequences were taken at theridge monitoring the beads passing the nanoslit.
A pinched injection protocol [30] was used for theDEMSAs. First, dc voltages at channels 1, 2 and 3 wereset such that the DNA molecules migrate from reservoir 3to channel 4 without leaking into channels 1 and 2. In thesecond step an ac voltage is superimposed by a dc voltage, bothapplied at channel 1, to drive a defined volume of DNA solutioninto channel 2 and to create selective dielectrophoretic trapsat the ridges simultaneously. Fluorescence image sequencesof molecules passing several ridges were taken and thefluorescence intensity was monitored in a defined regionof interest between two ridges. Afterward, the fluorescenceintensity was plotted versus the time to detect a shift betweentwo different sorts of molecules (see also figure 4).
For the EMSA experiments, the sample preparation wasthe same as for the chip experiments, excluding the sampledilution. The gel electrophoresis was performed in a 1.2%agarose gel with a onefold TAE buffer (40 mM Tris (99%,Roth, Germany), 10 mM NaCl (Riedel-de-Haen, Germany)and 1 mM EDTA). DNA of 100 ng were pipetted with thesame amount of loading buffer (TE/G buffer, MBBL GmbH,Germany) into the wells and 333 V m−1 (Power PAC 3000,BIO RAD, USA) were applied for 3 h. Thereafter, the gelwas placed into an ethidium bromide (Merck, Germany) bathfor 5 min and then in a water bath for 10 min. Finally, thegel was photographed under UV light (3UV-38 3UV LAMP,UVP, Cambridge, UK).
3. Results and discussion
As already mentioned in section 2.1 the thickness of theSU-8 layer varies with the amount of solid fraction and theparameters of the spincoating procedure. In table 1, the layerthickness is shown for various parameters. The layer heightwas controlled with a profilometer (see figure 2). For the firstSU-8 layer, an error of ±15 nm was determined. So the heightof the nanoslit could be defined by an accuracy of 15 nm.For the second layer SU-8 of 52% solid fraction was spincoated at 3000 rpm for 30 s for all devices. The thickness ofthe second layer was 5 μm, whereas the total structure heightvaries between 5.7 and 5.2 μm due to the height of the firstlayer. The resulting aspect ratios of the produced nanochannelsare summarized in table 2.
The whole developing process was done by hand, so thesecond photomask was not always perfectly aligned to the first
4
J. Micromech. Microeng. 22 (2012) 115024 M Viefhues et al
(a ) (b) (c)
(d ) (e) (f ) (g )
Figure 2. Control of the production process via profilometer scans (a)–(c) and photographs of the masterwafer with 20-fold magnification,all channels with the width of 200 μm (d)–(g). (a) 126 nm structure height of the first layer. (b) 510 nm structure height of the first layer.(c) 5.6 μm total structure height after terminal back at 200 ◦C for 20 min. (d) and ( f ) Perfectly aligned structures with arc and s-shapedridges. (e) and (g) Structures with lateral displacement. The second layer is slightly shifted to the left.
layer resulting in a lateral displacement or tilting. In figures 2(e) and (g) a displacement of the two layers is visible.For most of the produced masterwafers, the second layerfitted perfectly to the first, see figures 2(d) and ( f ). Here,all nanochannels in a specific device had the same height.But even structures of higher complexity can be built sinceadditional contact lithographic steps are possible, as well. Thesubsequent monolithic soft lithographic production would notalter anyway. The masterwafer was silanized after the terminalhard bake (see section 2.1) to prevent adhesion of PDMS tothe SU-8 structures. After silanizing the masterwafer once, noadhesion of polymer was observed, so a high reliability wasachieved.
When the nanoslit height is in the submicrometer range,the polymer to mold the structure should posses sufficientmechanical properties. Such as certain elastic stiffness tominimize deformation and collapse of the nanoslit [31] onthe one hand. On the other hand, it had to be flexible to allowpeel-off of the masterwafer. A double layer of h-PDMS [20]and PDMS proved to be sufficient. To check if the ridges sag oreven close the nanochannel, passability tests were performedwith appropriate nanobeads for different layouts, heights andlengths of the ridge, see also figure 3. For nanochannels ofheight larger or equal to 520 nm, 500 nm beads were used.The beads were driven electrophoretically through the channeland passed the nanoslit, over the whole channel width, withoutbeing stopped. So the flow through height of the nanoslit couldbe assumed to be more than 500 nm, when the beads could notpass otherwise. For 180 and 126 nm nanoslit heights 100 nmbeads were used. Here, the beads could pass the nanoslit over
the whole channel width only for the 180 nm nanoslit height.The 126 nm nanoslit (width-to-height aspect ratio of 1122:1)collapsed and was not passable (not shown). The ridges werearranged perpendicular to the long channel axis reducing theflow through height over the full microchannel width. Butnanochannel layouts with reduced width and height, relativeto the microchannel, are possible as well, i.e. one or severalnanochannels connect two microchannels.
Based on these results, we have shown stable aspectratios up to 873:1 that is 14.5 times better than the previouslypublished results of Mao with an aspect ratio of about 60:1,using PDMS [17], and 3.5 times better than the results ofMao et al with an aspect ratio of 250:1 for using glass–silicone structures [32]. In figure 3, the beads are depicted aslongish structures when passing the nanoslit, especially for thetrapezoidal ridge. This is due to the fast migration accordingto the high electric field within the nanoslit.
One possible application of the three-dimensionalinsulating structures is for dielectrophoretic manipulation [8].Here, the miniaturized structures were used to detect DNA-complexes in a gel-free application without specific matrices.Therefore, the analyte of interest (DNA) was injected into thestructured channel via pinched injection as described above.In figure 4 the fluorescence intensity is plotted versus time fortwo different DNA fragments and two DNA-complexes. For allthree experiments, a clear peak-shift is visible, i.e. the analytesmigrate with different velocities. In analogy to EMSA,this indicates two different species or complex formation,respectively. It stands out that for both complexes the pureDNA exhibits a higher velocity, i.e. the maximum fluorescence
5
J. Micromech. Microeng. 22 (2012) 115024 M Viefhues et al
Figure 3. Passability test of nanochannels. Time-lapse fluorescence images of nanobeads passing ridges are depicted. Beads of 500 nm wereused, except for 180 nm nanoslit height, where 100 nm beads were used. The direction of flow was from the left to the right.
(a ) (b) (c)
Figure 4. DEMSAs. The fluorescence intensity in a defined region was plotted versus time (see section 2.3). For all assays a clear shift isvisible within less than 30 s, indicating a complex formation of DNA with actinomycin D (ACTD) or E. coli RNAP, respectively. (a) 6.0 kbpand 2.686 kbp DNA at 350 Hz, 325 V ac voltage. The 6.0 kbp DNA migrates more slowly than the 2.686 kbp DNA (peak maxima at 13.2and 11.2 s, respectively). (b) 6.0 kbp DNA and 6.0 kbp DNA/ACTD complex, at 550 Hz, 350 V ac voltage. The complexed DNA migratesmore slowly (peak maxima at 10.4 and 9.6 s, respectively). (c) 6.0 kbp DNA and 6.0 kbp DNA/RNAP complex at 300 Hz, 475 V ac voltage.Again the complex migrates more slowly (peak maxima at 10.6 and 9.4 s, respectively).
intensity appears earlier than for the complex. This isequivalent to larger polarizabilities of the DNA-complexes,also shown in [29]. Although manipulation and separationby dielectrophoresis of cells or DNA was presented earlier[7, 25, 26], only the miniaturization toward nanoslits allowed
determination between pure DNA and DNA-complexes asdemonstrated with the DEMSA.
In figure 5, the corresponding EMSA experiments aredepicted. Clear shifts could be detected for all three analytes.The EMSA had a runtime of 3 h [18], whereas the time to
6
J. Micromech. Microeng. 22 (2012) 115024 M Viefhues et al
(a ) (b) (c)
Figure 5. Electrophoretic mobility shift assays. (a) Shift between 6.0 and 2.686 kbp DNA. (b) Shift between 6.0 kbp DNA/ACTD complexand 6.0 kbp DNA. (c) Shift of 6.0 kbp DNA/RNAP complex and 6.0 kbp DNA.
detect a shift was in all DEMSA experiments about 30 s. Evenwhen taken into account the time for plotting the molecules’distribution, the DEMSA is much faster than the EMSA.
4. Conclusion
We presented a fabrication process of nanochannels with thelack of chemical or physical etching by using soft lithographywith poly(dimethylsiloxane). The nanochannels were formedby ridges that reduce the height of a microfluidic channel.Passability experiments were performed to prove the stabilityof the nanochannels. Different ridge layouts and width-to-height aspect ratios were investigated. Stable nanochannelswith height down to 180 nm and width-to-height aspect ratioup to 873:1 could be fabricated reproducible.
Additionally, by exploiting these devices, the newtechnique of dielectrophoretic mobility shift assays (DEMSA)was introduced and proved sufficient to detect DNA-complexes. Therefore, the migration of pure DNA and DNA-complexes was monitored in an array of nanochannels. Withinless than 30 s a clear shift could be detected, indicating twospecies of molecules. Together with the fact that this approachis inherently independent of toxic agents, future applicationsin molecular biotechnology and medicine can be implementedin more compliant way with respect of the concerns of safetyregulation in drug administration.
With the presented technique, fast prototyping of newnanofluidic devices is possible. Enabling novel insides intotheory of nanochannels and development of three-dimensionallayouts for wide fields of bioanalysis and purification methods,e.g. based on DEP [23–26, 33] as presented with the DEMSA.
Acknowledgments
This work was supported by the German Research Foundation(DFG) within the collaborative research center SFB 613(project D2).
References
[1] Chen Z, Hsu F-C, Battigelli D and Chang H-C 2006 Captureand release of viruses using amino-functionalized silicaparticles Anal. Chim. Acta 569 76–82
[2] Spiering A, Getfert S, Sischka A, Reimann Pand Anselmetti D 2011 Nanopore translocation dynamics ofa single DNA-bound protein Nano Lett. 11 2978–82
[3] Fanzio P, Mussi V, Manneschi C, Angeli E, Firpo G,Repetto L and Valbusa U 2011 DNA detection with apolymeric nanochannel device Lab Chip 11 2961–6
[4] Saleh O A and Sohn L L 2001 Quantitative sensing ofnanoscale colloids using a microchip Coulter counter Rev.Sci. Instrum. 72 4449–51
[5] Saleh O A and Sohn L L 2003 An artificial nanopore formolecular sensing Nano Lett. 3 37–38
[6] Han J and Craighead H G 2000 Separation of long DNAmolecules in a microfabricated entropic trap array Science288 1026–9
[7] Regtmeier J, Kasewieter J, Everwand M and Anselmetti D2011 Continuous-flow separation of nanoparticles byelectrostatic sieving at a micro-nanofluidic interface J. Sep.Sci. 34 1180–3
[8] Regtmeier J, Eichhorn R, Viefhues M, Bogunovic Land Anselmetti D 2011 Electrodeless dielectrophoresis forbioanalysis: theory, devices and applicationsElectrophoresis 32 2253–73
[9] Chang H-C and Yossifon G 2009 Understandingelectrokinetics at the nanoscale: a perspectiveBiomicrofluidics 3 12001
[10] Xia Y and Whitesides G M 1998 Soft lithography Annu. Rev.Mater. Sci. 28 153–84
[11] Unger M A, Chou H-P, Thorsen T, Scherer A and Quake S R2000 Monolithic microfabricated valves and pumps bymultilayer soft lithography Science 288 113–6
[12] Chen W, Lam R H W and Jianping F 2012 Photolithographicsurface micromachining of polydimethylsiloxane (PDMS)Lab Chip 12 391–5
[13] Becker H and Gartner C 2008 Polymer microfabricationtechnologies for microfluidic systems Anal. Bioanal. Chem.390 89–111
[14] Zhang X and Haswell S J 2006 Materials matter inmicrofluidic devices MRS Bull. 31 95–9
[15] Hellmich W, Greif D, Pelargus C, Anselmetti D and Ros A2006 Improved native UV laser induced fluorescence
J. Micromech. Microeng. 22 (2012) 115024 M Viefhues et al
detection for single cell analysis in poly (dimethylsiloxane)microfluidic devices J. Chromatogr. A 1130 195–200
[16] Kuo J S and Chiu D T 2011 Disposable microfluidicsubstrates: transitioning from the research laboratory intothe clinic Lab Chip 11 2656–65
[17] Mao P 2005 Fabrication and characterization of nanofluidicchannels for studying molecular dynamics in confinedenvironments PhD Thesis Department of MechanicalEngineering, Massachusetts Institute of Technology, USA
[18] Hellman L M and Fried M G 2007 Electrophoretic mobilityshift assay (EMSA) for detecting protein-nucleic acidinteractions Nature Protoc. 2 1849–61
[19] Duong T T, Kim G, Ros R, Streek M, Schmid F, Brugger J,Ros A and Anselmetti D 2003 Size-dependent free solutionDNA electrophoresis in structured microfluidic systemsMicroelectron. Eng. 67-8 905–12
[20] Odom T W, Love J C, Wolfe D B, Paul K Eand Whitesides G M 2002 Improved pattern transfer in softlithography using composite stamps Langmuir 18 5314–20
[21] Regtmeier J, Duong T T, Eichhorn R, Anselmetti D and Ros A2007 Dielectrophoretic manipulation of DNA: separationand polarizability Anal. Chem. 79 3925–32
[22] Viefhues M, Manchanda S, Chao T-C, Anselmetti D,Regtmeier J and Ros A 2011 Physisorbed surface coatingsfor poly(dimethylsiloxane) and quartz microfluidic devicesAnal. Bioanal. Chem. 401 2113–22
[23] Lapizco-Encinas B H and Rito-Palomares M 2007Dielectrophoresis for the manipulation of nanobioparticlesElectrophoresis 28 4521–38
[24] Pethig R 2010 Review article—dielectrophoresis: status of thetheory, technology and applications Biomicrofluidics4 022811
[25] Pethig R, Menachery A, Pells S and De Sousa P 2010Dielectrophoresis: a review of applications for stem cellresearch J. Biomed. Biotechnol. 2010 182581
[26] Srivastava S K, Gencoglu A and Minerick A R 2011DC insulator dielectrophoretic applications inmicrodevice technology: a review Anal. Bioanal. Chem.399 301–21
[27] Viefhues M, Eichhorn R, Fredrich E, Regtmeier Jand Anselmetti D 2012 Continuous and reversible mixingor demixing of nanoparticles by dielectrophoresis Lab Chip12 485–94
[28] Regtmeier J, Eichhorn R, Bogunovic L, Ros Aand Anselmetti D 2010 Dielectrophoretic trapping andpolarizability of DNA: the role of spatial conformationAnal. Chem. 82 7141–9
[29] Viefhues M, Regtmeier J and Anselmetti D 2012 Fast andcontinuous-flow separation of DNA-complexes andtopological DNA variants in microfluidic chip formatsubmitted
[30] Ermakov S V, Jacobson S C and Ramsey J M 2000 Computersimulations of electrokinetic injection techniques inmicrofluidic devices Anal. Chem.72 3512–7
[31] Huang Y Y, Zhou W, Hsia K J, Menard E, Park J-U,Rogers J A and Alleyne A G 2005 Stamp collapse in softlithography Langmuir 21 8058–68
[32] Mao P and Han J 2005 Fabrication and characterization of20 nm planar nanofluidic channels by glass–glass andglass–silicon bonding Lab Chip 5 837–44
[33] Ros A, Hellmich W, Regtmeier J, Duong T-Tand Anselmetti D 2006 Bioanalysis in structuredmicrofluidic systems Electrophoresis 27 2651–8
![High efficiency dielectrophoretic ratchet - PureHigh efficiency dielectrophoretic ratchet. Physical Review E, 86(4), 1-9. [041106]. ... theoretical upper limit corresponding to the](https://static.documents.pub/doc/80x56/5e48381b49401c3bfa26d20c/high-efficiency-dielectrophoretic-ratchet-pure-high-efficiency-dielectrophoretic.jpg)
