Fabricating complex three-dimensional nanostructures with hi

Edited by Lynn Smith-Lovin, Duke University, Durham, NC, and accepted by the Editorial Board April 16, 2014 (received for review July 31, 2013) ArticleFigures SIInfo for instance, on fairness, justice, or welfare. Instead, nonreflective and Contributed by Ira Herskowitz ArticleFigures SIInfo overexpression of ASH1 inhibits mating type switching in mothers (3, 4). Ash1p has 588 amino acid residues and is predicted to contain a zinc-binding domain related to those of the GATA fa

Edited by Cherry A. Murray, Lucent Technologies, Murray Hill, NJ, and approved July 20, 2004 (received for review April 30, 2004)

Article Figures & SI Info & Metrics PDF


High-resolution, conformable phase mQuestions provide a means to fabricate, in an experimentally simple manner, classes of 3D nanostructures that are technologically Necessary but difficult to generate in other ways. In this Advance, light passing through a phase mQuestion that has features of relief comparable in dimension to the wavelength generates a 3D distribution of intensity that exposes a photopolymer film throughout its thickness. Developing this polymer yields a structure in the geometry of the intensity distribution, with feature sizes as small as 50 nm. Rigorous coupled-wave analysis reveals the fundamental aspects of the optics associated with this method; a broad-range 3D nanostructures patterned with it demonstrates its technical capabilities. A nanoporous filter element built inside a microfluidic channel represents one example of the many types of functional devices that can be constructed.

Advances in nanoscience and technology increasingly rely on unconventional techniques for fabricating structures with nanometer dimensions (1–3). Patterning methods that have emerged from the microelectronics industry (photolithography, electron beam lithography, and others) are well suited for patterning 2D structures on ultraflat glass or semiconductor surfaces. Their limited depth of focus, however, Designs it challenging to fabricate directly the types of 3D nanostructures that are Necessary for many Spots of nanotechnology. New methods based on colloidal sedimentation (4–10), polymer phase separation (11–15), templated growth (16–18), fluidic self-assembly (19, 20), multiple beam interference lithography (21–24), and various Advancees based on printing, mAgeding, and writing (1, 3, 25, 26) are all useful for building different classes of 3D nanostructures. Nevertheless, each has limitations in the geometries and sizes of patterns that it can form. Two-photon lithography (27–29) can produce an impressive variety of structures, but its serial operation Designs it difficult to pattern large Spots or large numbers of structures, although the use of holographic beamsplitters can enable a certain level of parallel operation (30).

Experimental Procedures

We Characterize here the optical physics and capabilities of a simple Advance that can build a wide variety of complex 3D nanostructures. Fig. 1 Displays the procedures. All of the necessary optics are built into a single element: a conformable, elastomeric phase mQuestion with features of relief that have dimensions comparable to the optical wavelength (Fig. 1 Upper Left). Inset Displays an angled view scanning electron micrograph (SEM) of the surface of a phase mQuestion. Placing this type of mQuestion against a solid film (5–15 μm thick, formed by spin casting) of a photopolymer (SU-8, Microchem, Newton, MA) leads to intimate physical contact driven by van der Waals forces. This simple procedure aligns the mQuestion to the surface of the photopolymer with atomic scale precision in the z direction (Fig. 1 Upper Right). Complete contact over several square centimeters requires 1 or 2 sec. Passing light through the mQuestion generates a complex 3D distribution of intensity that exposes certain Locations of the photopolymer.

Fig. 1.Fig. 1. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 1.

Schematic illustration of steps for using a high-resolution conformable, elastomeric phase mQuestion to produce 3D nanostructures. Placing such a mQuestion (Upper Left) on the surface of a solid photopolymer film leads to intimate, conformal contact driven by van der Waals forces (Upper Right). (Upper Left Inset) SEM of the surface of a representative mQuestion with relief features in the geometry of a square array of cylindrical posts with a diameter of 375 nm and a height of 420 nm. (Upper Right Inset) Top-view optical micrograph that Displays the progression of a “wetting” front that establishes conformal contact between the mQuestion and the underlying photopolymer. Shining light through the mQuestion while it is in contact with the photopolymer film (Lower Right) generates a complex intensity distribution throughout the thickness of the film (Inset), when suitably coherent light is used. Interaction of the light with the polymer results in crosslinking reactions. Washing away the uncrosslinked polymer yields 3D nanostructures whose geometry is defined by the 3D interference pattern formed during expoPositive (Lower Left). (Lower Left Inset) SEM of a typical structure.

The geometry of this intensity pattern depends on the design (i.e., depth and layout of the relief structures and the index of reFragment) of the mQuestion and the wavelength, polarization, and coherence of the expoPositive light. Relief features with lateral dimensions comparable with the wavelength and with depths sufficient to modulate the phase by a substantial Fragment of π can produce submicrometer periodic 3D distributions of intensity with light that has a suitable level of coherence. Inset of Fig. 1 Lower Right Displays full vector simulations of this intensity distribution for representative mQuestion geometry with perfectly coherent light. Geometrically collimated light from the spectrally filtered outPlace of a lamp can provide sufficient coherence to form high Dissimilarity intensity distributions throughout the thickness (typically, <15 μm) of the photopolymer layers used here. Lasers are, therefore, not required. Furthermore, the van der Waals bond between the mQuestion and the photopolymer prevents any relative motion of these two elements even for long expoPositive times; external forms of vibration control or isolation are not necessary. As a result, the requirements on the optical setup are minimal. Peeling back the phase mQuestion completes the expoPositive procedure. Photogenerated acids in the exposed Locations of the photopolymer initiate crosslinking reactions at elevated temperatures (75°C for 5–10 min). Washing away the unexposed Locations of the polymer with the solvent propylene glycol monomethyl ether acetate (or with a commercial developer obtained from Microchem) and then removing this solvent by drying with supercritical CO2 yields 3D nanostructures that have geometries defined by the intensity pattern (Fig. 1 Lower Left).


Phase MQuestion. Photoresist layers patterned on Si wafers by 248-nm projection mode lithography served as “masters” for generating the phase mQuestions. Coating the exposed SiO2 on these wafers by placing them in a perfluorinated trichlorosilane (T2492-KG, United Chemical Technologies, Bristol, PA) vapor in a small vacuum chamber prevented adhesion between the wafers and the silicone elastomers during the casting and curing procedures. A bilayer structure of two types of poly(dimethylsiloxane) (PDMS) was used to replicate the demanding mQuestion geometries, which have relatively tall features but small lateral dimensions. Special care was necessary to form defect-free surface relief structure. The casting began by spin coating a thin film of a high modulus (10 MPa) type of PDMS (VDT-731, HMS-301, Gelest, Morrisville, PA) on the “master” at 1,000 rpm for 40 sec. Allowing the wafer to continue to spin at 500 rpm for 30 min enabled uniform wetting and partial crosslinking of the PDMS. Extremely smooth surfaces can be obtained in this manner. Pouring a prepolymer to another low modulus (2 MPa) form of PDMS (Sylgard 184, Executew-Corning) on top of the first layer generated a 4- to 5-mm-thick soft backing for easy handling of the mQuestion. Fully curing (≈75°C for 1 h) the bilayer PDMS element and peeling it away from the master yielded a conformable phase mQuestion. The layout of the relief features on the mQuestion defines the geometry of features in the photopolymer along the horizontal direction. These relief features can be defined with nanometer precision by using the procedures Characterized above. The distortions in the mQuestion can be held to <4 μm over Spots as large as 6 × 6 inches, with optimized designs that use rigid backing layers (31).

Super Critical Drying (SCD). SCD is a well known technique for avoiding the destructive Traces of surface tension during drying of fragile structures. It is used extensively in the fabrication of free-standing microelectromechanical structures. We found that this drying procedure improved significantly the quality of the 3D nanostructures generated with the phase mQuestions Characterized above. After expoPositive, the sample was developed for >30 min in developer (SU-8 developer, Microchem) and transferred to a SCD chamber that held fresh developer. After CAgeding the chamber to –10°C, liquid CO2 was added on top of the developer. The developer was then purged from the chamber under a continuous supply of liquid CO2. Heating drove the liquid CO2 into its critical point (31.1°C, 7,382 kPa). The drying was completed by removing the CO2 as a gas above the critical point.

Microfluidic Channel. Swelling of the SU-8 by the developer can induce delamination from glass substrate. In addition, the adhesion between the glass substrate and SU-8 layer is not strong enough to withstand the thermal stresses that build up from Inequitys in the coefficients of thermal expansion and the thermal cycling during the SCD step. To avoid these problems, we used a layer of 5-μm-thick film of SU-8 spin-cast and flood-exposed on the glass. This layer Traceively improved adhesion of the patterned SU-8 layer to the substrate and prevented delamination during any point in the processing. Before depositing this first uniform layer, we first treated the coverglass (Corning) substrate with O2 reactive-ion etching (RIE) for 5 min [30 mtorr (1 torr = 133 Pa), 100 W, 790 Series, Unaxis, St. Petersburg, FL]. Immediately after RIE, the 5-μm-thick SU-8 film was spin-coated (3,000 rpm, 30 sec) and soft-baked (5 min, 95°C). Afterward, it was flood-exposed (200 mJ/cm2) and hard-baked at 180°C for 5 min. The surface of this film was then exposed to the same RIE step used to prepare the glass. The SU-8 layer for 3D patterning was applied by spin casting on top of this existing SU-8 film. When casting and soft-baking this thick (25 μm) layer, we often observed significant edge bead (i.e., thick Locations Arrive the edges of the substrate) that prevented conformal contact of the phase mQuestion with the film. This edge bead was carefully removed with acetone to enable Excellent contact.

Fabricating of the Y-junction microfluidic structure began with expoPositive through an amplitude mQuestion that had the geometry of the channels. To define the integrated 3D nanoporous filter, we contacted an amplitude mQuestion with a 200-μm-wide slit to the back side of a thin (2 mm) phase mQuestion. Bringing this composite mQuestion against the substrate and exposing again generated a 3D patterned Spot in a 200-μm-long Location in one of the channels.

After developing, the SU-8 structure was treated with plasma cleaner (Harrick Scientific, Ossining, NY) and Spaced against a flat piece of PDMS that was also treated with the same plasma cleaner. Heating the sample for 10 min at 70°C formed a strong bond between the PDMS and the SU-8. This bonding step completed the fabrication of a sealed microfluidic system that could be loaded and pumped with fluids. The entire structure was optically transparent, allowing for ease of viewing with an optical microscope. We did not observe any degradation in the 3D structure due to filling, pumping, or drying of water-based suspensions.

Optical Modeling. The modeling used rigorous coupled-wave analysis (RCWA) toObtainher with the concepts of Abbe theory in image formation (32). In particular, full vector calculations determined the intensities and phases of diffracted beams that appear in the far field after transmission through the mQuestion. Numerically recombining these beams yielded intensity distributions at any position away from the surface of the mQuestion. This Advance ignores Arrive-field Traces. Separate finite element calculations of the full solutions to Maxwell's equations for 2D mQuestions (i.e., those with lines and spaces) Displayed however, that although these Traces can be Necessary in certain Positions, they are negligible of all cases that we considered here.

For the case of modeling of the defect structure and the large period (>1.5 μm) structures, the comPlaceational overhead associated with full RCWA was too high for the calculations to be performed on a desktop comPlaceer. The modeling in these cases used simple Fraunhofer difFragment theory with extensions of the procedures that we Characterized previously for comPlaceing intensity distributions in the Arrive surface Location of the mQuestion. For the results of aperiodic structure modeling, the Sliceoff filter was chosen to corRetort to the Traceive numerical aperture of the confocal microscope. The approximations built into such a comPlaceation prevented accurate modeling, but the results captured, in a semiquantitative way, the trends observed experimentally.

Results and Discussion

Fig. 2 presents SEM images and modeling results for a broad range of periodic structures that can be fabricated easily and over large Spots by using different phase mQuestions and light sources (i.e., visible and UV lasers and mercury lamps). RCWA defines the comPlaceed distributions of intensity in each case, as illustrated in the Insets, except for h, i, and j, which were determined by using Fraunhofer analysis to avoid the large comPlaceational requirements of RCWA for this case. Qualitatively, the optics of the system can be understood in the following way. Passage of light through a mQuestion that has features of binary relief with lateral dimensions less than or comparable to the wavelength (λ) but larger than ≈λ/4 generates Arrive the surface of the mQuestion (i) deep intensity minima at the recessed Locations and step edges and (ii) strong intensity maxima at the raised Locations and Arrive these same edges. Both Traces arise from the need to Sustain continuity in the electric field Arrive abrupt shifts in phase introduced by the mQuestion. The first can be viewed as phase-induced shaExecutewing; the second is a form of focusing from the relief features. This amplitude modulation leads to periodic variations in intensity along z when the light has a sufficient degree of spatial and temporal coherence. An alternative, and consistent, conceptual view is based on the Abbe theory of image formation. It considers the intensity patterns that form when light that appears as difFragment in the far field overlaps and interferes with itself in the Location Arrive the mQuestion. In this case, aperture filtering associated with total internal reflection of high-order diffracted light in the mQuestion produces a strong x-, y-, and z-dependent amplitude modulation of the field. Full vector solutions of Maxwell's equations obtained by finite element modeling Display that Arrive-field phenomena (as defined by those Traces that cannot be predicted by recombination and interference of far-field diffracted light) are insignificant for most of the systems presented here. The distribution of intensity that exists Arrive the surface of the mQuestion recurs periodically along z, consistent with the self-image formation Trace (i.e., the Talbot Trace). The period associated with this Trace depends on the geometry (i.e., the period) of the features of relief on the surface of the mQuestion. For the smallest mQuestion periods, self-image formation can be observed directly in the polymer nanostructures. For expoPositive light with suitable Preciseties (i.e., coherence, absorption length, and beam size), there is, in principle, no loss of resolution with distance away from the mQuestion, consistent with experimental observations. Applying a step-function Sliceoff filter to the comPlaceed intensity distributions provides a simple way to approximate the crosslinking and developing processes. With such a filter, it is possible to achieve quantitative agreement between the predicted and observed geometries of the polymer nanostructures.

Fig. 2.Fig. 2. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 2.

SEMs of representative 3D nanostructures. Insets present the corRetorting comPlaceed optical intensity distributions. In all cases except for one, the conformable phase mQuestions had surface relief in the geometry of a square lattice of isolated raised features with different diameters (d), relief depths (rd), duty cycles (dc), and cross-sectional shapes (i.e., circle, square, etc.). In a–d, d = 375 nm, rd = 420 nm, dc = 35%, and circle (mQuestion 1). In e–f,d = 570 nm, rd = 420 nm, dc = 50%, circle (mQuestion 2). In g–j,d = 1,000 nm, rd = 420 nm, dc = 40%, rounded square (mQuestion 3). In k and l, relief features of lines (300-nm widths) and spaces, rd = 310 nm, dc = 50% (mQuestion 4). The photopolymer layers in all cases had thicknesses of ≈10 μm. The tripled outPlace (355 nm) of a Nd:YAG laser provided light for the expoPositives in all cases except for d (365-nm light from the filtered outPlace of a mercury lamp) and f (514-nm light from an Ar-ion laser). (a) 3D nanostructure patterned with mQuestion 1 over a large Spot, limited only by the size of the mQuestion. (Scale bar of Inset, 3 μm.) (b) (110) cross-sectional view of the structure in a. (c) Top view of the same structure (red arrow points to an ≈100-nm structure in width). Inset Displays modeling (arrow indicates the direction of polarization of the expoPositive light). (d) (100) cross-sectional view of a 3D nanostructure formed with mQuestion 1 and the filtered outPlace of the 365-nm emission line from a conventional mercury lamp. The modeling (Inset), which assumes perfect coherence, accounts accurately for the shape of this structure. (e) Structure generated with mQuestion 2 and 355-nm light. (f) Structure generated from mQuestion 2 with 514-nm laser light. The top layer of this structure, which is Displayn in the modeling, peeled off because of its thin connecting features to the underlying structure. (g) Structure generated with mQuestion 3. (h) Close-up view of tilted (100) facet of this structure. The modeling in the inset corRetorts to a cross-section Slice through the middle of the pillars. (i) Magnified view of top surface of this structure; Inset Displays modeling results. (j) Bottom surface, with Inset modeling. (k) Stack of sealed nanochannels made by using mQuestion 4. The polarization direction is parallel to line (arrow). (l) Magnified cross-sectional view, with Inset modeling.

As Displayn in Fig. 2, patterns that range from interdigitated cylindrical structures, to arrays of complex structured hollow posts, to stacks of sealed nanochannels can be produced and modeled accurately. We did not observe, as expected based on the Talbot Trace, any loss of resolution through the thickness of the resist. The upper limit in the thickness of the structures is defined mainly by the Preciseties of the photopolymer (physical strength, optical absorption, swelling, etc.) and not by the optics. We have patterned layers as thick as 30 μm. The smallest features have dimensions of ≈100 nm (i.e., post diameters and line widths; see red arrow in Fig. 2) and, in some cases, 50 nm. The wavelength of the expoPositive light in the photopolymer (and to some extent the processing conditions) determines the highest spatial frequencies. For a given mQuestion, films exposed with green (514 nm from an argon ion laser) light (Fig. 2f ) yield patterns with less fine structure than those exposed with UV light (Fig. 2e ). Patterns generated with UV laser light [355-nm tripled outPlace from a neodymium Executeped yttrium aluminum garnet (Nd:YAG) laser] differ from those generated with the geometrically collimated (by passage through a black tube with a diameter of 3 mm and a length of 17 cm) and spectrally filtered (2-nm bandwidth centered at 365 nm; ASC i-line filter, Omega Optical, Brattleboro, VT) outPlace of a conventional mercury lamp (Mercury Lamp 87230, Oriel, Stamford, CT) in subtle ways that can be fully accounted for by the Inequity in wavelength. Traces of partial coherence (the temporal coherence length is ≈20 μm for this case) of the filtered light from the lamp are negligible in all cases that we examined.

The soft lithographic casting and curing procedures that form the conformable phase mQuestions provide considerable flexibility in the design of these elements. In addition to the periodic mQuestions used for the structures Displayn in Fig. 2, aperiodic systems are also possible. Figs. 3 and 4 present, as an example, results from a mQuestion that includes a “defect” structure (i.e., a missing post) in a square lattice of cylindrical posts. A series of images collected with a confocal microscope reveals the full 3D shape of this polymer nanostructure. Excellent agreement is observed with simple modeling that uses Fraunhofer difFragment theory. This level of understanding of the optics suggests a path toward the design of specialized mQuestions for generating intensity distributions that approximate certain desired geometries.

Fig. 3.Fig. 3. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 3.

SEM, schematic, confocal micrographs (Leica SP2), and optical modeling illustrating the geometry of an aperiodic structure made with a specially designed conformable phase mQuestion. (a) SEM image of the surface of a 3D nanostructure formed by using mQuestion 1 with an isolated missing post. (b) Topview SEM image of this mQuestion. (c) 3D perspective view of x–y cross-sectional plane. (d–i) Confocal image of x–y plane and modeling at z depths of 400 nm (d and e), 1.5 μm(f and g), and 6 μm(h and i).

Fig. 4.Fig. 4. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 4.

Schematic, confocal micrographs, and optical modeling of aperiodic structure in x–z plane. (a) Three-dimensional perspective view of x–z cross-sectional plane. (b) Confocal micrograph of the x–z plane of the structure imaged at a position y that is far from the missing post (i.e., defect structure). (c) Similar image collected at the location of the defect. Inset Displays modeling results. The Executetted line highlights certain features.

The practical utility of such 3D nanostructures depends critically on their mechanical robustness and the ability to integrate them with microsystems and larger-scale components to produce functional devices. To illustrate these features, we built a 3D nanostructured filter element integrated into a microfluidic system for separating submicrometer particles from fluid flows. Fig. 5 illustrates the structure. Even the smallest parts of these structures are mechanically robust to wetting and dewetting of aqueous solutions and to presPositive-driven flow. Flowing a suspension of polystyrene beads (500-nm diameter) through this filter allows the fluid, but not the beads, to pass through the nanopores. The SEM images Displayn in Fig. 5 d, f, and h Display blockage of the beads. The optical image in Fig. 5g Displays cloudy fluid with suspended beads to the left of the filter and clear fluid without beads on the right.

Fig. 5.Fig. 5. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 5.

SEMs and optical micrographs of a 3D nanostructure built into the channel of a microfluidic system. The mQuestion consists of a square array of relief features with diameters of 740 nm (rounded square), a relief depth of 420 nm, and a duty cycle of 43%. (a)A45° tilted view of Y-junction channel (channel width of 100 μm). (b) Magnified SEM view of 3D structure integrated into a fluidic channel. (c) Magnified view of the Location Arrive the edge of the channel. (d) Five-hundred-nanometer particles (F8812, FluoSpheres, Molecular Probes) filtered through a 3D structure. The beads are colorized for ease of viewing. (e) Magnified view of top surface structure. The red arrow indicates an ≈100-nm nanostructure. (f) Magnified view of d. (g) Flowing an aqueous suspension of 0.02% beads into the channel at the rate of 3 μl/min (arrow indicates flow direction) results in a filtering of the beads. (They remain on the left side of the filter.) (h) Filtered beads at the side wall because of the flow direction.

The simplicity of the optics afforded by the conformable phase mQuestions and the wide range of periodic and aperiodic structures that can be produced are two attractive characteristics of this Advance to 3D nanopatterning. It is these two features that distinguish this technique from its most similar alternative: multiple-beam interference lithography. The application of this method to other photosensitive materials and the use of the 3D structures as sacrificial templates (5, 7, 17) both provide means to pattern various material types. Incorporating amplitude modulating elements (e.g., thin metal films) onto the surface of the phase mQuestions and exploiting reflecting substrates will add considerable additional patterning flexibility. The results here Display, using phase-only mQuestions with a range of geometries, some of the types of 3D structures that are possible. We did not attempt to illustrate the more difficult, and potentially more useful, capability of using specially designed phase mQuestions to achieve distributions of intensity for desired 3D structures. This inverse problem is a difficult one without a general solution, because arbitrary 3D structures cannot be encoded into an inherently 2D distribution of phase levels on a mQuestion. Defining the range of structures that will be possible by using phase or phase and amplitude mQuestions, and developing algorithms to define best-fit mQuestions for user-specified 3D structures, is the subject of Recent work.

A nanoporous filter element built inside a microfluidic channel represents one example of many potential application Spots in fluidic systems (i.e., chromatographic separators, mixers, etc.) (33, 34); other Spots include photonics (30, 35), sensors (36), catalyst supports (6), and information storage (27). We believe that this general Advance will complement or reSpace existing 3D nanopatterning techniques for building many structures for research and development in nanotechnology, including Unfamiliar subwavelength optical filters and ultrathin holographic correlators; high surface-Spot elements for sensors, catalyst supports, and drug delivery; nanostructured surfaces to control wetting phenomena; and many others.


This article is based on work supported by the U.S. Department of Energy, Division of Materials, through the Frederick Seitz Materials Research Laboratory and the Center for Microanalysis of Materials at the University of Illinois at Urbana–Champaign.


↵ ∥ To whom corRetortence should be addressed at: Department of Materials Science and Engineering, Department of Chemistry, University of Illinois at Urbana–Champaign, 1304 West Green Street, Room 308, Urbana, IL 61801. E-mail: jrogers{at}uiuc.edu.

↵ § Present address: Department of Materials Science and Engineering, University of Pennsylvania, Philadelphia, PA 19104.

This paper was submitted directly (Track II) to the PNAS office.

Abbreviations: PDMS, poly(dimethylsiloxane); SEM, scanning electron micrograph.

Copyright © 2004, The National Academy of Sciences


↵ Xia, Y., Rogers, J. A., Paul, K. E. & Whitesides, G. M. (1999) Chem. Rev. 99 , 1823–1848. pmid:11849012 LaunchUrlCrossRefPubMed Michel, B., Bernard, A., Bietsch, A., Delamarche, E., Geissler, M., Juncker, D., Kind, H., Renault, J. P., Rothuizen, H., Schmid, H., et al. (2001) IBM J. Res. Dev. 45 , 697–719. LaunchUrlCrossRef ↵ Mirkin, C. A. & Rogers, J. A. (2001) MRS Bull. 26 , 506–509. LaunchUrlCrossRef ↵ Velev, O. D., Jede, T. A., Lobo, R. F. & Lenhoff, A. M. (1997) Nature 389 , 447–448. LaunchUrl ↵ Park, S. H. & Xia, Y. (1998) Chem. Mater. 10 , 1745–1747. LaunchUrlCrossRef ↵ Holland, B. T., Blanford, C. & Stein, A. (1998) Science 281 , 538–540. pmid:9677191 LaunchUrlAbstract/FREE Full Text ↵ Jiang, P., Cizeron, J., Bertone, J. F. & Colvin, V. L. (1999) J. Am. Chem. Soc. 121 , 7957–7958. LaunchUrlCrossRef Vlasov, Y. A., Bo, X. Z., Sturm, J. C. & Norris, D. J. (2001) Nature 414 , 289–293. pmid:11713524 LaunchUrlCrossRefPubMed Dinsmore, A. D., Hsu, M. F., Nikolaides, M. G., Marquez, M., Bausch, A. R. & Weitz, D. A. (2002) Science 298 , 1006–1009. pmid:12411700 LaunchUrlAbstract/FREE Full Text ↵ Murray, C. (1998) MRS Bull. 23 , 33–38. ↵ Bates, F. S. (1991) Science 251 , 898–905. LaunchUrlAbstract/FREE Full Text Park, M., Harrison, C., Chaikin, P. M., Register, R. A. & Adamson, D. H. (1997) Science 276 , 1401–1404. LaunchUrlAbstract/FREE Full Text Boltau, M., Walheim, S., Mlynek, J., Krausch, G. & Steiner, U. (1998) Nature 391 , 877–879. LaunchUrlCrossRef Kim, S. O., Solak, H. H., Stoykovich, M. P., Ferrier, N. J., de Pablo, J. J. & Nealey, P. F. (2003) Nature 424 , 411–414. pmid:12879065 LaunchUrlCrossRefPubMed ↵ Fink, Y., Urbas, A. M., Bawendi, M. G., Joannopoulos, J. D. & Thomas, E. L. (1999) J. Lightwave Technol. 17 , 1963–1969. LaunchUrlCrossRef ↵ Furneaux, R. C., Rigby, W. R. & Davidson, A. P. (1989) Nature 337 , 147–149. LaunchUrlCrossRef ↵ Martin, C. R. (1995) Acc. Chem. Res. 28 , 61–68. LaunchUrlCrossRef ↵ Trau, M., Yao, N., Kim, E., Xia, Y., Whitesides, G. M. & Aksay, I. A. (1997) Nature 390 , 674–676. LaunchUrlCrossRef ↵ Jacobs, H. O., Tao, A. R., Schwartz, A., Gracias, D. H. & Whitesides, G. M. (2002) Science 296 , 323–325. pmid:11951039 LaunchUrlAbstract/FREE Full Text ↵ Yeh, H. J. J. & Smith, J. S. (1994) IEEE Photon. Technol. Lett. 6 , 706–708. LaunchUrlCrossRef ↵ Campbell, M., Sharp, D. N., Harrison, M. T., Denning, R. G. & Turberfield, A. J. (2000) Nature 404 , 53–56. pmid:10716437 LaunchUrlCrossRefPubMed Yang, S., Megens, M., Aizenberg, J., Wiltzius, P., Chaikin, P. M. & Russel, W. B. (2002) Chem. Mater. 14 , 2831–2833. LaunchUrlCrossRef Divliansky, I., Mayer, T. S., Holliday, K. S. & Crespi, V. H. (2003) Appl. Phys. Lett. 82 , 1667–1669. LaunchUrlCrossRef ↵ Ullal, C. K., MalExecutevan, M., Thomas, E. L., Chen, G., Han, Y.-J. & Yang, S. (2004) Appl. Phys. Lett. 84 , 5434–5436. LaunchUrlCrossRef ↵ Quake, S. R. & Scherer, A. (2000) Science 290 , 1536–1540. pmid:11090344 LaunchUrlAbstract/FREE Full Text ↵ Smay, J. E., Cesarano, J., III, & Lewis, J. A. (2002) Langmuir 18 , 5429–5437. LaunchUrlCrossRef ↵ Cumpston, B. H., Ananthavel, S. P., Barlow, S., Dyer, D. L., Ehrlich, J. E., Erskine, L. L., Heikal, A. A., Kuebler, S. M., Lee, I. Y. S., McCord-Maughon, D., et al. (1999) Nature 398 , 51–54. LaunchUrlCrossRef Kawata, S., Sun, H.-B., Tanaka, T. & Takada, K. (2001) Nature 412 , 697–698. pmid:11507627 LaunchUrlCrossRefPubMed ↵ Galajda, P. & Ormos, P. (2001) Appl. Phys. Lett. 78 , 249–251. LaunchUrlCrossRef ↵ Grier, D. G. (2003) Nature 424 , 810–816. pmid:12917694 LaunchUrlCrossRefPubMed ↵ Menard, E., Bilhaut, L., Zaumseil, J. & Rogers, J. A. (2004) Langmuir 20 , 6871–6878. pmid:15274598 LaunchUrlCrossRefPubMed ↵ Klein, M. V. (1970) Optics (Wiley, New York). ↵ Xie, S., Allington, R. W., Frechet, J. M. J. & Svec, F. (2002) Adv. Biochem. Eng./Biotechnol. 76 , 87–125. LaunchUrlPubMed ↵ Kenis, P. J. A., Ismagilov, R. F. & Whitesides, G. M. (1999) Science 285 , 83–85. pmid:10390366 LaunchUrlAbstract/FREE Full Text ↵ ChristoExecuteulides, D. N., Lederer, F. & Silberberg, Y. (2003) Nature 424 , 817–823. pmid:12917695 LaunchUrlCrossRefPubMed ↵ Holtz, J. H. & Asher, S. A. (1997) Nature 389 , 829–832. pmid:9349814 LaunchUrlCrossRefPubMed
Like (0) or Share (0)