Driven Pattern Formation in Organic Thin Film Materials: Complex

CF3CF2CF2O(CF(CF3)CF2O)nCF(CF3)CONHCH2CH2CH2Si(OCH3)3 (Krytox SA, DuPont) was deposited on a SiO2 surface by both spin-casting and contact ...
0 downloads 0 Views 2MB Size
8776

J. Phys. Chem. B 2001, 105, 8776-8784

Driven Pattern Formation in Organic Thin Film Materials: Complex Mesoscopic Organization in Microcontact Printing on Si/SiO2 via the Spontaneous Dewetting of a Functionalized Perfluoropolyether Ink† Martin K. Erhardt‡ and Ralph G. Nuzzo*,‡,§ Departments of Chemistry and of Materials Science and Engineering, and the Frederick Seitz Materials Research Laboratory, UniVersity of Illinois at Urbana-Champaign, Urbana, Illinois 61801 ReceiVed: March 13, 2001; In Final Form: July 14, 2001

Polyfluoropolyether (PFPE) films have long been used as lubricant coatings for magnetic recording media. In this paper, we demonstrate that the unique wetting properties of PFPEssmore specifically, the dynamical organizations that result from spontaneous dewettingscan be harnessed to generate mesoscopically patterned features of these materials on SiO2. In this work, a functionalized PFPE amphiphile with the formula CF3CF2CF2O(CF(CF3)CF2O)nCF(CF3)CONHCH2CH2CH2Si(OCH3)3 (Krytox SA, DuPont) was deposited on a SiO2 surface by both spin-casting and contact printing. Both methods produce complex surface structures comprised of beaded domains and depletion regions (and in the case of spin-casting, also thin films) that result from dewetting processes. Spontaneous dewetting was used to generate self-organizing PFPE bead patterns by microcontact printing. The wetting transitions in this latter case occur directly on the printing tool and, via the bias provide by the topography of the stamp, provide a means for generating and transferring complex organizations of adsorbate domains to the substrate. The combination of contact printing with spontaneous dewetting of the PFPE enabled us to produce high-fidelity patterns of discrete, micron-scale beads from printing tools with continuous line shapes without any alteration to the original mask pattern. The patterned beads typically had radii and characteristic separation lengths on the micron-scale, and heights on the nanoscale. These length scales appear to be governed by the combined influences of solvent-mediated nucleation processes and coarsening. Characterization was performed by optical microscopy, atomic force microscopy (AFM), secondary ion mass spectrometry (SIMS), and X-ray photoelectron spectroscopy (XPS). These data suggest general design strategies that can be exploited to control the contours and structural profiles that result from the dewetting of the PFPE ink on a stamp used for contact printing.

Introduction The use of perfluorinated polyethers (PFPEs) as lubricants for magnetic recording media has prompted considerable research on their surface chemical and wetting properties on a variety of substrates, including amorphous carbon and SiO2,1-17 as well as various metals.18,19 The significant improvements that PFPEs have enabled in magnetic recording media, in turn, have sparked an interest in developing new uses for PFPE materials as agents for modifying surfaces. Recently, this class of compounds has attracted attention for its wear resistance and water barrier properties, and for potential applications of PFPEs as ultrathin barrier layers to diffusion and corrosion on solid surfaces.20-35 For example, derivatives of commercial PFPE lubricants have been shown to chemisorb on SiO2 and amorphous carbon substrate surfaces, opening up the possibility of using amphiphilic PFPE derivatives as stable ultrathin Teflonlike polymeric coatings for surfaces.6,29-35 Nanoscale thin films of these materials have been formed on SiO2 using commercially available PFPEs, including those available under the trade names Krytox (DuPont), Fomblin (Montedison), and Demnum (Daikin), and their derivatives.6,30-35 †

Part of the special issue “Royce W. Murray Festschrift”. * To whom correspondence should be addressed. Email: r-nuzzo@ uiuc.edu. Phone: 217-244-0809. Fax: 217-244-2278. ‡ Department of Chemistry. § Department of Materials Science and Engineering.

The PFPE derivative used in this work is one such material: a monofunctional silyl amide of Krytox (Figure 1a), known as Krytox SA (DuPont), with the formula CF3CF2CF2O(CF(CF3)CF2O)nCF(CF3)CONHCH2CH2CH2Si(OCH3)3 (Figure 1b). Krytox SA appeared to us to be an attractive candidate as an etchresistant coating for the thermal oxide thin films commonly encountered in microfabrication, given the possibility for strong chemisorption between the trialkoxysilane end group and the SiO2 surface, a well understood reaction that has been described in detail for similar systems.36-38 Related monolayer resists have been effectively patterned (and applied in fabrication) using microcontact printing (µCP), a soft-lithographic patterning method.38-49 The utility of PFPE compounds as either resists or inks for controlled pattern formation via contact printing, has not been reported, however. Given the great promise of contact printing for applications in microelectronic device fabrication,40-41,49-50 rapid biological assaying,40-42,51-54 and selective surface catalysis and film growth,40-41,55-63 the development of new ink chemistries is highly desirable. Typically, contact printing is used to achieve uniform, continuous mass transfer of an adsorbate from the printing tool to the substrate. In this work, however, we demonstrate that simply by changing the ink used, a single printing tool with continuous feature line shapes can effect both a uniform mass transfer of the adsorbate molecules comprising the ink and a transfer of complex patterns of discrete beads delineating the

10.1021/jp010946a CCC: $20.00 © 2001 American Chemical Society Published on Web 08/17/2001

Microcontact Printing on Si/SiO2

J. Phys. Chem. B, Vol. 105, No. 37, 2001 8777

Figure 2. Optical micrographs of domains formed by spin-casting Krytox SA solution onto the native oxide surface of a silicon wafer.

Figure 1. Ball and stick representations of (a) “Krytox”, and (b) “Krytox SA” PFPEs (DuPont), shown here with one repeat unit (n ) 1).

identical feature shape. This novel form of pattern transfer suggests a new method for modifying the effective feature sizes of existing patterning tools, one appropriate for generating complex arrays of discrete surface structures with nanoscale vertical dimensions. Experimental Section Samples of Krytox SA were supplied by Dr. Jon Howell of E. I. du Pont de Nemours and Company. Krytox SA (MW = 1200-1500) solutions were prepared using 1,1,2-trichlorotrifluoroethane (TTFE) as a solvent in a ratio of 1 mg Krytox SA to 1 mL TTFE, a concentration of approximately 10-4 M. Spin-cast films were formed by dropwise application of a 1 mg/ mL solution of Krytox SA in TTFE to the substrate surface, followed by spinning at 4000 rpm for 30 s. Printed patterns of Krytox SA were formed by applying a 1 mg/mL solution of Krytox SA in TTFE dropwise to the patterned surface of a silicone (poly(dimethylsiloxane) (PDMS), Sylgard 184, Dow Corning) stamp, followed by 30 s of drying in air, and subsequent contact printing. The contact printing process and fabrication methods used to prepare patterned silicone stamps have been described elsewhere.16,39-41 Characterization was performed by optical microscopy, atomic force microscopy (AFM), secondary ion mass spectrometry (SIMS), and X-ray photoelectron spectroscopy (XPS). The AFM instrument used was a Digital Instruments Dimension 3100 operated in both contact and tapping modes, with a maximum scan size of 90 µm, scanning at 1 Hz. The AFM images were collected simultaneously in height and amplitude mode (deflection mode for contact mode images). Height mode images were used for quantitative analysis, while amplitude and deflection mode images, which are more easily visualized, are shown in the figures. SIMS data acquisition was performed using a Cameca ims Sf secondary ion mass spectrometer equipped with a resistive anode encoder for position-sensitive ion counting. The ion beam used for the analysis was a 14.5 keV Cs+ beam. Negative secondary ions were detected. XPS data acquisition was performed using a Kratos Axis Ultra

instrument equipped with a monochromatic Al X-ray source, and operating at a voltage of 15 keV with a beam current of 15 mA. XPS fluorine maps were acquired at a binding energy of 688 eV. Results and Discussion The focus of this work is on the character of the mass transfer obtained in the microcontact printing of a novel ink based on a trialkoxysilane derivative of a PFPE telomer. As we show below, remarkable properties attend the printing-based pattern transfer realized. To provide points of comparison, we first describe the results obtained for two nonprinting based methods used to deposit Krytox SA onto Si substrates. The first deposition method involved a simple immersion of an oxidized Si(100) substrate into an adsorbate containing solution. After a 20minute immersion in a 1 mg/mL solution of Krytox SA in TTFE, a coherent film with an ellipsometrically determined mass coverage of ∼30 Å was obtained. This film was strongly bound to the substrate, resisting removal by repeated rinsing steps carried out with a variety of solvents. The wetting properties of this treated sample were consistent with those expected for a surface bearing a monolayer of the Krytox SA (i.e., the surfaces are extremely hydrophobic). The results obtained by spin-casting the adsorbate were somewhat more complex in terms of the structures of the layers produced. As illustrated by the micrograph shown in Figure 2, this deposition method lead to the formation of a complex pattern of beads on the surface. Interestingly, these beads were found to be arrayed in a patchwork of irregular polygonal domains with mesoscale dimensions ranging from tens to hundreds of microns. The domains are well-defined, with clear boundaries and straight edges. The resulting appearance of this surface, from a macroscopic perspective, was similar to that of shattered glass. In the micrograph, one also notes that the interior regions of the domains bear a large number of more randomly arrayed beads of the Krytox. These beads, if allowed to remain on the sample surface in the laboratory ambient atmosphere, became very difficult to remove in essentially any solvent. We take this later result as being reflective of the cross-linking of the Krytox, an expected reaction of the alkoxysilane groups with water vapor. Taken together, the data suggest that the PFPE molecules poorly wet this substrate (see below) and coalesce into more macroscopic domains. The factors that drive their organization

8778 J. Phys. Chem. B, Vol. 105, No. 37, 2001

Erhardt and Nuzzo

Figure 3. Optical micrographs of spin-cast Krytox SA films at a 4000 Å thermal oxide step.

Figure 4. Optical micrographs of an IDA line pattern: (a) a contact printed pattern on a native oxide surface and (b) a corner of the stamp used to print the pattern in (a).

along domain boundaries, however, cannot be established on the basis of these data alone and will be discussed later. Deposition by spin-casting was also performed on thermally grown SiO2 surfaces. The optical micrographs presented in Figure 3 show a complex silicon substrate that had been treated in this way. A thermal oxide was grown on this sample and then half of it was further treated in a wet etchant to strip that layer. The image was collected in the vicinity of the step created by the etching; the native oxide region is pictured in the upper part of both images. The step height between these two regions as measured by profilometry was found to be 4000 Å. Two aspects of this image are particularly striking. First, the domains on the thermal oxide surface tend to be larger and more sparsely populated with PFPE beads than is found on the smoother native oxide surface. Second, there appears to be a depletion region formed on either side of the step where little PFPE is visible. Significantly, the edge beads that form the boundaries of the depletion regions are substantially larger than those found elsewhere on the surface, suggesting that PFPE beads that would otherwise have formed in the depletion region have spontaneously coarsened to form the larger edge beads displaced from the step. These data thus suggest that the deposition process is sensitive to subtle details of the surface topography, a conclusion borne out by profilometry data (not shown) and SIMS data, to which we now turn.

We performed a further investigation of the surface composition present in the depletion region boundary revealed by the micrograph in Figure 3 using SIMS. A SIMS fluorine map (data not shown) of the boundary region and step clearly revealed that this region is depleted in PFPE. The Krytox SA coverage appears to stop abruptly at the boundary visible in the optical micrographs (Figure 3). These data strongly suggest that the regions away from the step actually bear a thin film of Krytox SA, one that supports the beads seen in the optical micrographs. The spin-casting thus appears to generate a thin film of the adsorbate, albeit one that has a more complex structure associated with it than that found for a sample formed by immersion in an adsorbate containing solution. The fact that topological features (such as a step) can be used to induce edge bead formation with PFPE materials suggests a rich capability for extending microcontact printing. The sidewalls of patterned printing stamps might be expected to effectively act as steps and thus serve to induce a similar effect were the ink to poorly wet the PDMS stamp surface. This hypothesis was completely validated by experiment. Figures 4-6 show optical micrographs of two different PFPE bead patterns formed on Si native oxide surfaces by contact printing. These patterns were chosen to explore the geometric sensitivities one expects in aggregation and depletion phenomena that are mediated by nucleation and mass transfer dynamics. Figure 4a

Microcontact Printing on Si/SiO2

J. Phys. Chem. B, Vol. 105, No. 37, 2001 8779

Figure 5. (a) An AFM micrograph of printed Krytox SA lines like those pictured in Figure 4a and (b) an optical micrograph of a silicone printing stamp with the 15 µm IDA line pattern following wetting by Krytox SA (but prior to printing).

Figure 6. Optical micrographs of a circle pattern: (a) a contact printed pattern on a native oxide surface and (b) a section of the stamp used to print the pattern in (a).

shows a collinear array of Krytox beads. The lines derive from a stamp based on an interdigitated array (IDA) electrode pattern with a 15 µm line-to-line spacing; a portion of this stamp is shown in Figure 4b. The morphology of the bead pattern is much more readily examined by AFM. A representative image of the pattern transfer obtained using the fine line (IDA) pattern is shown in Figure 5a. The importance of depletion effects in the micrcontact printing based patterning of PFPE inks is readily deduced from the characteristics of this micrograph. Figure 5a shows an AFM (deflection mode) image of the lines printed from an IDA stamp pattern with a line-to-line spacing of 15 µm. From the image, however, it is apparent that the line-to-line distances of the bead arrays vary alternately by an amount that is displaced consistently by 3-5 µm from the expected 15 µm design rule. This

suggests that the beads of the PFPE ink form along a depletion zone that is uniformly displaced from each edge of the line features present on the printing tool. Microscopy directly confirms this hypothesis. Figure 5b shows an optical micrograph of a printing stamp based on the IDA line pattern after the application of the Krytox SA solution, but prior to printing. The image, focused on the upper (i.e., the contact) features of the stamp, clearly shows the presence of edge bead arrays recessed from each sidewall. Although blurry in this image due to the limitations of the depth of focus, similar bead arrays are also seen in the recessed regions of the stamp. It seems clear that the sidewalls of the printing tool features induce the formation of a depletion region and edge bead structure, one similar to that seen at the thermal oxide step shown in Figure 3. These mesoscopic patterns are very rapidly formed (their formation

8780 J. Phys. Chem. B, Vol. 105, No. 37, 2001

Figure 7. An AFM micrograph of a segment of a printed circle like those pictured in Figure 6a.

occurs faster than the time it takes to produce a focused image and is likely limited only by the rate at which the ink’s solvent evaporates after its application to the stamp). The printing, then, merely serves to transfer the complex fluid patterns that nucleate and coarsen on the stamp surface. These depletion effects thus act as a means for reducing pattern feature sizes without the need to refabricate the patterning tool. Perhaps more striking, though, is the very high symmetry of the bead assemblies that result. The stamp edge contours appear to completely dominate the nucleation processes that control the subsequent coarsening of bead domains. We will return to this point later in the paper. Figure 6a shows an array of discrete circles (and more complex crosses) generated from the stamp pattern pictured in Figure 6b; the circles are 200 µm in diameter. In addition to highlighting the variety of shapes that can be formed, Figure 6a illustrates the contrast sensitivities that result when the dominant area of contact for the stamp is the area outside the pattern (i.e., outside of the circles). The patterns revealed in these images can be used to develop a qualitative understanding of the dynamics of the depletion effects, which can be used in turn to devise ways to selectively manipulate the bead coarsening behavior. Of interest in this latter example are the length scales seen in the domain coarsening (the regions from which the PFPE domains can accrete material are larger in this case than was true for the IDA design). Figure 7 shows an AFM micrograph of a segment of a circle like those pictured in Figure 6. This deflection mode image illustrates the typical sizes and spacing of the PFPE droplets obtained by contact printing. One of the most striking aspects of the image shown in Figure 7 is the relatively large size and spacing of the droplets formed at the pattern edges as compared with the smaller and more randomly distributed droplets formed in the interior of a patterned region. This contrast suggests that the organization of bead domains arises from a complex interplay of nucleation dynamics and mass-transfer-driven coarsening. For the ink loadings examined in this work, these processes generate domains showing two distinct length scales of organization. The edge mediated coarsening generates structures with characteristic spacings in the micron range, while the nucleation dynamics elsewhere drives the formation of beads

Erhardt and Nuzzo with nanoscale lateral dimensions. In tandem, these factors produce beads with vertical dimensions in the nanometer size range. We currently do not understand why the nanoscale beads in the interior regions do not coarsen further. We note that Krytox SA is a polymerizable amphiphile. The initially liquid domains transferred by the printing step do age quickly in the laboratory ambient, presumably due to the presence of water vapor. A loss of droplet fluidity due to cross-linking should limit the extent to which coarsening occurs. We believe, however, that this reaction is slow compared to the initial rate of film collapse on the stamp that produces the droplets upon solvent evaporation (a point to which we will return later). Figure 8 shows photoelectron images of patterned lines with fine spacings, ones in which individual PFPE beads are more clearly discernible. Figure 8a is a representative XPS survey scan of a patterned region of a native oxide layer on a Si substrate. The strong fluorine peak at 688 eV was used as the reference signal for collecting the fluorine map images shown in the lower portion of the figure. Figure 8b shows a corner of an area patterned with lines separated by 30 µm spacings, and Figure 8c shows a high magnification image of those lines. XPS fluorine maps of the 15 µm IDA line pattern shown in Figure 4a were also acquired, an example of which is shown in Figure 8d. Though the clarity of the latter image is limited by the resolution of the instrument, the individual beads comprising the lines are clearly discernible. These data serve to strongly contrast the printing results from those obtained for solution or spin-casting depositions. The XPS data establish that the bead structures transferred by contact printing, at least for the cases of the fine design rules of the line structures shown here, do not sit on top of an underlying PFPE thin film. These bead structures are thus likely to result from true depletion effects in a nucleation-limited aggregation process occurring on the PDMS stamp, a point we address more directly via the microscopy data presented below. The region of the substrate contacted by the printing stamp (circular pattern) can be seen clearly in the XPS images shown in Figure 9. The PFPE beads formed by this pattern (except for at the line edges) are too small to be fully resolved by the XPS instrument. The gradient characteristics of the F 1s image remain instructive all the same. As expected, the F1s signal intensity is highest at the edges of the patterned region. The high magnification image shown in Figure 9b explicitly illustrates the radial symmetry of the segregation and coarsening of domains seen at the domain boundaries. This latter image further confirms the role of the stamp topology in defining the characteristics of this pattern forming process. We should note that the shapes of the transferred PFPE beads reflect the wetting properties of the printing system. The PFPE fluid used as the printing ink, therefore, must have a finite contact angle on PDMS. Without this, discrete droplets will not form on the stamp. Using a surface oxidation step to modify the wetting properties of the stamp may lead to a different outcome. We should note here, though, that the solvent used to formulate this ink does in fact swell the PDMS and thus might reconstruct a modified surface. This brings us to a point of significant importance that we have yet to consider. Even though the obvious notion of a transport-limited aggregation mechanism appears to correlate generally with the types of patterns seen, other aspects of the contact printing process might also need to be considered. Most notable in this regard is the manner in which the ink is applied to the stamp. A volatile solvent is used in the inking process. This solvent is lost by evaporation during the preparation of the stamp for printing. The dewetting

Microcontact Printing on Si/SiO2

J. Phys. Chem. B, Vol. 105, No. 37, 2001 8781

Figure 8. (a) An XPS survey scan of a native oxide surface patterned with Krytox SA. (b) A high magnification XPS fluorine map depicting an internal region of patterned lines separated by 30 µm. (c) A medium magnification XPS fluorine map depicting a corner of a region patterned with lines separated by 30 µm.(d) A high magnification XPS fluorine map depicting an edge of a region patterned with lines separated by 15 µm.

Figure 9. XPS fluorine maps of the circle pattern pictured in Figure 7a at (a) low magnification and (b) medium magnification.

dynamics, as a result, must be more complicated than what one might presume from considerations based solely on a simple two-phase interfacial system (PDMS and a PFPE fluid). The loss of solvent, then, must provide a very significant force that drives at least some aspects of the pattern formation seen in this system. Data on the size and spacing distributions of the edge beads were obtained from AFM micrographs (not shown here) measured around the circumference of a complete circle (represented by the segment shown in Figure 7). These data

are shown in the form of histograms in Figure 10. Bead radii (Figure 10a) ranged from 0.9 to 2.3 µm, with over 80% of beads having radii between 1.2 and 1.8 µm, and 90% between 1.0 and 2.0 µm. Bead height and spacing data were also collected and are shown in histogram form in parts b and c of Figure 10, respectively. Over 95% of the beads had heights between 150 and 350 nm, with over 85% being between 150 and 300 nm tall. The beads also exhibit some degree of local uniformity. For example, over a lateral distance of 100 µm from a given bead, over 85% of beads in that region had radii within 0.5 µm

8782 J. Phys. Chem. B, Vol. 105, No. 37, 2001

Erhardt and Nuzzo

Figure 10. Histograms of Krytox SA bead dimensions as measured by AFM from a printed circle pattern like those pictured in Figure 7: (a) individual bead radii, (b) individual bead heights, and (c) spacing between adjacent beads.

Figure 11. Histograms of Krytox SA bead dimensions as measured by AFM along a 1000 µm line segment of a printed IDA line pattern like those pictured in Figure 5a: (a) individual bead radii, (b) individual bead heights, and (c) spacing between adjacent beads.

of the given bead’s radius, and over 90% of beads in that region had heights within 100 nm of the given bead’s height. Bead spacing distances ranged from 3.0 to 10.5 µm, with over 80% within 4-8 µm of their nearest neighbor. Beads in the IDA line pattern pictured in Figure 4a had radial dimensions on the same order as those in the circle pattern, but with larger variations in bead height and spacing. Figure 11 shows a graphical representation of bead dimension measurements made by AFM along a 1000 µm segment of one representative line pictured in Figure 5a. The measured bead

radii (Figure 11a) ranged from 1 to 4 µm, with over 75% between 1.5 and 3.0 µm. Bead heights (Figure 11b) ranged from 21.6 to 149.4 nm, with over 80% between 40 and 100 nm. Bead separation lengths (Figure 11c) ranged from 4.2 to 29.2 µm, with over 90% lying between 5 and 20 µm. The breadth of the distributions seen in the data shown in Figure 11 are, at first glance, somewhat surprising, given the more uniform nature of the distribution obtained on a circular edge. Both patterns produce a depletion region, with comparable local symmetry. The distributions seen in Figure 11 are not

Microcontact Printing on Si/SiO2

J. Phys. Chem. B, Vol. 105, No. 37, 2001 8783

Figure 12. AFM micrographs showing examples of pattern-induced coarsening: (a) a hairpin pattern printed from an edge of the IDA pattern (see Figure 4b) and (b) an alignment cross printed from the circle pattern shown in Figure 6b.

broad in a simple Gaussian sense, though. Rather, the data appear to show a very long period of roughly 600 µm. We believe this is due to dynamics unrelated to the coarsening of simple fluid domains. More likely, length scales of this magnitude result from the mechanics of manual contact printing over a wide area. We are continuing to investigate refinements to this technique that might serve to increase the uniformity of bead size and spacing. The observed variation in bead dimensions and spacing suggest a use for techniques that use contour shapes and dimensions to selectively influence the dimensions of individual PFPE beads. From an applications perspective, control over the height of individual beads would allow control over intersubstrate spacing in a multisubstrate stack, as a substrate stacked on top of a bead pattern will only contact the tallest beads in the pattern if both substrates are planar. Such concerns commonly arise in MEMS devices,64 and more recently bead structures have been considered for applications in high-density data storage systems, including ultralightweight hard disk drives. The results presented above demonstrate a directed method for generating such polymeric microstructures using a softlithographic patterning method. We note that while the sizes of the beads formed are not monodispersed, control over the size and height of every bead in a pattern may be less important for such an application than would controlling the location and height of the large, coarsened beads at critical support points. We have found that one method by which the vertical and lateral dimensions of individual beads generated from PFPE inks can be controlled is via an explicit pattern-induced coarsening, as described below. Figure 12 shows AFM micrographs in which the coarsening of the PFPE domains has resulted in striking (and reproducible) variations in bead size. Figure 12a shows a section of the IDA pattern (see Figure 4b) involving a hairpin turn with two right angles. The contrast between the inside and outside corners of the hairpin pattern illustrates the way in which the pattern geometry and feature dimensions can be manipulated to tailor the bead size. For example, the relatively large size of the beads formed at the inner corners demonstrate how coarsening can be selectively accentuated through the use of orthogonal junctions in close proximity, even when the dimensions of parallel features would not otherwise be expected to result in

coarsening at this length scale. We see, in marked contrast to the relatively small distance separating the two larger beads at the inner corners, that the beads connecting the outer corners span a distance of over 40 µm, and show no sign of coarsening. The image in Figure 12b, an AFM micrograph of a cross pattern, shows that coarsening behavior can be controlled not only by the proximity of other corners, but also by proximity to areas of high-density coverage outside of the patterned region. In this example, the inner and outer corners of the pattern are equidistant from their counterparts, but the dominant coarsening only occurs at the eight outer corners, which are bordered on three sides by areas of high-density PFPE coverage. The four inner corners, on the other hand, are bordered by regions of high-density PFPE coverage on one side only, and do not show evidence of coarsening. The effects seen in this latter example closely parallel trends seen in electroplating, in which transport effects frequently lead to an inhomogeneous coverage of high aspect ratio features (such as the “knee” that forms at the outer corners of surface steps). This parallel notwithstanding, many of the mechanistic underpinnings of the complex pattern formation seen in these latter examples remain poorly understood. The cross-shaped pattern shown in Figure 12b presents an especially difficult challenge in this regard. The cross-shaped pattern shown above corresponds to a fiduciary marker on the original mask used to construct the stamp with the circular patterns. The shape is a negative contrast image, meaning that it is the outer region of the feature that actually contacts the substrate during printing. The limiting dimensions of the shape are easily understood in the context of the edge reflection mechanism discussed earlier. The centerto-center separation between the “dominant” drops along the inner perimeter that form the arms of the cross is ∼12 µm. This value is ∼2 µm larger than the dimension of that feature on the patterning tool. As with the IDA lines, the edge contours both reflect and direct bead nucleation at a similar distance from the line edge (here ∼1 µm). Unlike the organization seen in either the IDA lines or the circles, printing the cross-shaped pattern yielded a complex array of secondary droplets bracketing those on the inner perimeter with remarkably high symmetry. The large area that feeds the droplet coarsening in this case is clearly an important parameter in defining the formation of the printed pattern. Further work will be needed to develop a predictive

8784 J. Phys. Chem. B, Vol. 105, No. 37, 2001 model for the pattern formation seen in this case, especially as regards the significant complexity that must attend the stampsolvent interactions. The current study demonstrates a simple example of how driven wetting dynamics can be used to generate complex mesoscale patterns in a PFPE thin film material. This method amplifies patterns transferred by contact printing and is capable of directing the mesoscopic organization of fluid domains with micron-scale and nanoscale dimensions. With better control of the coarsening behavior, we believe that this procedure could provide an interesting new approach to forming complex structures. Acknowledgment. This work was supported by the National Science Foundation (CHE-9626871), the Department of Energy (DEFG02-91-ER45439), and the Defense Advanced Research Projects Agency (N66001-98-1-8915). AFM, XPS, SIMS, optical microscopy, and surface profilometry studies were carried out in the Center for the Microanalysis of Materials, University of Illinois, which is supported by the U. S. Department of Energy under grant DEFG02-91-ER45439. Rick Haasch was very helpful with XPS data acquisition and analysis, as was Nancy Finnegan with AFM data collection, and Judy Baker with SIMS data collection. We are especially grateful to Dr. Jon Howell of DuPont for supplying us with a sample of Krytox SA. Special thanks also goes to Noo Li Jeon for fabricating masters for the mold pattern. M.K.E. acknowledges a fellowship from the Department of Chemistry. Supporting Information Available: Figure S1: Optical micrographs of a linked-square pattern. Figure S2: XPS fluorine maps of the linked-square pattern in Figure S1. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Bhushan, B. Tribology and Mechanics of Magnetic Storage DeVices, 2nd ed.; Springer-Verlag: New York, 1996; Chapter 8. (2) Bhushan, B.; Israelachvili, J. N.; Landman, U. Nature 1995, 374, 607. (3) Jhon, M. S.; Phillips, D. M.; Vinay, S. J.; Messer, C. T. IEEE Trans. Magn. 1999, 35, 2334. (4) Kajdas, C.; Bhushan, B. J. Inf. Storage Process Syst. 1999, 1, 303. (5) Kim, M. C.; Phillips, D. M.; Ma, X.; Jhon, M. S. J. Colloid Interface Sci. 2000, 228, 405. (6) Ru¨he, J.; Kuan, S.; Novotny, V. J.; Blackman, G.; Clarke, T.; Street, G. B. ACS Symp. Ser. 1991, 485, 156. (7) Saperstein, D. D.; Lin, L. J. Langmuir 1990, 6, 1522. (8) Schulz, K. J.; Viswanathan, K. V. IEEE Trans. Magn. 1991, 27, 5166. (9) Toney, M. F.; Mate, C. M.; Leach, K. A.; Pocker, D. J. Colloid Interface Sci. 2000, 225, 219. (10) Tyndall, G. W.; Waltman, R. J. Mater. Res. Soc. Symp. Proc. 1998, 517, 403. (11) Viswanathan, K. V.; Schulz, K. J. Ceram. Trans. 1991, 19, 197. (12) Vurens, G.; Zehringer, R.; Saperstein, D. ACS Symp. Ser. 1992, 485, 169. (13) Waltman, R. J. Chem. Mater. 2000, 12, 2039. (14) Wu, J. H.; Mate, C. M. Langmuir 1998, 14, 4929. (15) Yanasigawa, M. Tribol. Trans. 1994, 37, 629. (16) Yanasigawa, M. Tribol. Trans. 1993, 36, 484. (17) Yanasigawa, M. Wear 1993, 168, 167. (18) Herrera-Fierro, Pilar; Jones, William, R., Jr.; Pepper, Stephen, V. J. Vac. Sci. Technol. A 1993, 11, 354. (19) Shogrin, B.; Jones, W. R., Jr.; Herrera-Fierro, P. Lubr. Eng. 1996, 52, 712. (20) Abe, M. Curr. Opin. Colloid Interface Sci. 1999, 4, 354. (21) Bell, G. A.; Howell, J.; Del Pesco, T. W. Chem. Ind. 1999, 77, 215. (22) Clarke, M. J.; Harrison, K. L.; Johnston, K. P.; Howdle, S. M. J. Am. Chem. Soc. 1997, 119, 6399.

Erhardt and Nuzzo (23) Elsner, P.; Wiggeralberti, W.; Pantini, G. Dermatology 1998, 197, 141. (24) Huang, L. J.; Hung, Y.; Chang, S. IEEE Trans. Magn. 1997, 33, 3154. (25) Johnson, G.; Meijs, G. F.; Laycock, B. G.; Griffith, M. G.; Chaouk, H.; Steele, J. G. J. Biomater. Sci., Polym. Ed. 1999, 10, 217. (26) Kasai, P. H. J. Appl. Polym. Sci. 1995, 57, 797. (27) Tonelli, C.; Gavezotti, P.; Strepparola, E. J. Fluorine Chem. 1999, 95, 51. (28) Yoshino, N.; Hamano, K.; Omiya, Y.; Kondo, Y.; Ito, A.; Abe, M. Langmuir 1995, 11, 466. (29) Bunyard, W. C.; DeSimone, J. M. Polym. Prepr. 1999, 40, 827. (30) Barth, G.; Cormia, R. D.; Teasley, L. A. Solid State Technol. January 1989, 119. (31) Kasai, P. H.; Spool, A. M. J. Phys. Chem. B 1998, 102, 7331. (32) Kasai, P. H.; Wass, A.; Yen, B. K. J. Inf. Storage Process Syst. 1999, 1, 245. (33) Ru¨he, J.; Novotny, V.; Clarke, T.; Street, G. B. J. Tribology 1996, 118, 663. (34) Ru¨he, J.; Blackman, G.; Novotny, V. J.; Clarke, T.; Street, G. B.; Kuan, S. J. Appl. Polym. Sci. 1994, 53, 825. (35) Xu, C. B.; Frank, C. W.; Tang, W. T.; Terrill, C. J. Adhes. 1998, 67, 195. (36) Arkles, B. Chemtech 1977, 7, 766. (37) Bascom, W. D. J. Colloid Interface Sci. 1968, 27, 789. (38) St. John, P. M.; Craighead, H. G. J. Vac. Sci. Technol. B 1996, 14, 69. (39) Kumar, A.; Whitesides, G. M. Appl. Phys. Lett. 1993, 63, 2002. (40) Xia, Y. N.; Rogers, J. A.; Paul, K. E.; Whitesides, G. M. Chem. ReV. 1999, 99, 1823. (41) Xia, Y.; Whitesides, G. M. Angew. Chem., Int. Ed. Engl. 1998, 37, 550. (42) Mrksich, M.; Whitesides, G. M. Trends Biotechnol. 1995, 13, 228. (43) Jeon, N. L.; Finnie, K.; Branshaw, K.; Nuzzo, R. G. Langmuir 1997, 13, 3382. (44) Xia, Y.; Zhao, X.-M.; Whitesides, G. M. Microelectron. Eng. 1996, 32, 255. (45) St. John, P. M.; Craighead, H. G. Appl. Phys. Lett. 1996, 68, 1022. (46) Xia, Y.; Mrksich, M.; Kim, E.; Whitesides, G. M. J. Am. Chem. Soc. 1995, 117, 9576. (47) Jeon, N. L.; Nuzzo, R. G.; Xia, Y.; Mrksich, M.; Whitesides, G. M. Langmuir 1995, 11, 3024. (48) Jeon, N. L.; Clem, P. G.; Payne, D. A.; Nuzzo, R. G. Langmuir 1996, 12, 5350. (49) Jeon, N. L.; Clem, P.; Jung, D. Y.; Lin, W. B.; Girolami, G. S.; Payne, D. A.; Nuzzo, R. G. AdV. Mater. 1997, 9, 891. (50) Deng, T.; Goetting, L. B.; Hu, J. M.; Whitesides, G. M. Sens. Actuators, A 1999, 75, 60. (51) Yang, Z. P.; Belu, A. M.; Liebmann-Vinson, A.; Sugg, H.; Chilkoti, A. Langmuir 2000, 16, 7482. (52) Bieri, C.; Ernst, O. P.; Heyse, S.; Hofmann, K. P.; Vogel, H. Nature Biotech. 1999, 17, 1105. (53) Urban, G. Sens. Actuators, A 1999, 74, 219. (54) St. John, P. M.; Davis, R.; Cady, N.; Czajka, J.; Batt, C. A.; Craighead, H. G. Anal. Chem. 1998, 70, 1108. (55) Kim, N. Y.; Jeon,N. L.; Choi, I. S.; Takami, S.; Harada, Y.; Finnie, K. R.; Girolami, G. S.; Nuzzo, R. G.; Whitesides, G. M.; Laibinis, P. E. Macromolecules 2000, 33, 2793. (56) Jeon, N. L.; Choi, I. S.; Whitesides, G. M.; Kim, N. Y.; Laibinis, P. E.; Harada, Y.; Finnie, K. R.; Girolami, G. S.; Nuzzo, R. G. Appl. Phys. Lett. 1999, 75, 4201. (57) Kind, H.; Bonard, J. M.; Forro, L.; Kern, K.; Hernadi, K.; Nilsson, L. O.; Schlapbach, L. Langmuir 2000, 16, 6877. (58) Kind, H.; Geissler, M.; Schmid, H.; Michel, B.; Kern, K.; Delamarche, E. Langmuir 2000, 16, 6367. (59) Kind, H.; Bonard, J. M.; Emmenegger, C.; Nilsson, L. O.; Hernadi, K.; Maillard-Schaller, E.; Schlapbach, L.; Forro, L.; Kern, K. AdV. Mater. 1999, 11, 1285. (60) Pitois, C.; Vukmirovic, S.; Hult, A.; Wiesmann, D.; Robertsson, M. Macromolecules 1999, 32, 2903. (61) Jeon, N. L.; Lin, W. B.; Erhardt, M. K.; Girolami, G. S.; Nuzzo, R. G. Langmuir 1997, 13, 3833. (62) Hidber, P. C.; Nealey, P. F.; Helbig, W.; Whitesides, G. M. Langmuir 1996, 12, 5209. (63) Hidber, P. C.; Helbig, W.; Kim, E.; Whitesides, G. M. Langmuir 1996, 12, 1375. (64) For example, see: Garabedian, R.; Gonzalez, C.; Richards, J.; Knoesen, A.; Spencer, R.; Collins, S. D.; Smith, R. L. Sens. Actuators A 1994, 43, 202.