Fabricating Nanoscale Chemical Gradients with ThermoChemical

Jun 10, 2013 - School of Chemistry and Biochemistry, Georgia Institute of Technology, 901 Atlantic Drive, Atlanta, Georgia 30332-0400, United States...
0 downloads 0 Views 4MB Size
Article pubs.acs.org/Langmuir

Fabricating Nanoscale Chemical Gradients with ThermoChemical NanoLithography Keith M. Carroll,†,‡ Anthony J. Giordano,§ Debin Wang,∥ Vamsi K. Kodali,⊥ Jan Scrimgeour,†,‡ William P. King,# Seth R. Marder,§ Elisa Riedo,†,§ and Jennifer E. Curtis*,†,‡ †

School of Physics, Georgia Institute of Technology, 837 State Street, Atlanta, Georgia 30332-0430, United States Parker H. Petit Institute for Bioengineering and Bioscience, Georgia Institute of Technology, 315 Ferst Drive, Atlanta, Georgia 30332-0363, United States § School of Chemistry and Biochemistry, Georgia Institute of Technology, 901 Atlantic Drive, Atlanta, Georgia 30332-0400, United States ∥ Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, California 94720, United States ⊥ Pacific Northwest National Laboratory, 902 Battelle Boulevard, Richland, Washington 99352, United States # Department of Mechanical Science and Engineering, University of Illinois at Urbana-Champagne, 1206 West Green Street, Urbana, Illinois 61801-2906, United States ‡

S Supporting Information *

ABSTRACT: Production of chemical concentration gradients on the submicrometer scale remains a formidable challenge, despite the broad range of potential applications and their ubiquity throughout nature. We present a strategy to quantitatively prescribe spatial variations in functional group concentration using ThermoChemical NanoLithography (TCNL). The approach uses a heated cantilever to drive a localized nanoscale chemical reaction at an interface, where a reactant is transformed into a product. We show using friction force microscopy that localized gradients in the product concentration have a spatial resolution of ∼20 nm where the entire concentration profile is confined to sub-180 nm. To gain quantitative control over the concentration, we introduce a chemical kinetics model of the thermally driven nanoreaction that shows excellent agreement with experiments. The comparison provides a calibration of the nonlinear dependence of product concentration versus temperature, which we use to design two-dimensional temperature maps encoding the prescription for linear and nonlinear gradients. The resultant chemical nanopatterns show high fidelity to the user-defined patterns, including the ability to realize complex chemical patterns with arbitrary variations in peak concentration with a spatial resolution of 180 nm or better. While this work focuses on producing chemical gradients of amine groups, other functionalities are a straightforward modification. We envision that using the basic scheme introduced here, quantitative TCNL will be capable of patterning gradients of other exploitable physical or chemical properties such as fluorescence in conjugated polymers and conductivity in graphene. The access to submicrometer chemical concentration and gradient patterning provides a new dimension of control for nanolithography.

A

dynamic range in surface concentration gray scale values, but their spatial resolution is intrinsically limited by optical diffraction.4−7 Similarly, microfluidic approaches afford concentration variation and spatial resolution on the micrometer scale, but arbitrary patterns are difficult to achieve, requiring sophisticated redesign for each pattern.8−10 Existing highresolution patterning techniques that might increase spatial resolution, such as dip pen nanolithography11−13 and micellar block copolymer nanolithography,14,15 are binary by nature and

nalog patterning of chemical concentration is a prevalent feature of natural interfaces. Most established nanopatterning schemes, however, provide only a binary prescription of surface concentration, severely limiting our ability to fabricate sophisticated chemical interfaces. Studies and applications of nanoscale phenomena such as those involving controlled doping levels, chemical kinetics, biomimetic interfaces, and molecular packing would benefit greatly from high-resolution surface gradients in chemical composition.1−3 Of those methods that can produce chemical concentration gradients, most if not all are limited in their spatial resolution and/or their flexibility. Light-based patterning techniques, for example, produce impressive arbitrary patterns with a wide © 2013 American Chemical Society

Received: March 15, 2013 Published: June 10, 2013 8675

dx.doi.org/10.1021/la400996w | Langmuir 2013, 29, 8675−8682

Langmuir

Article

the reaction, we model the chemical transformation using first order chemical kinetics:

have not shown an ability to create nanoscale chemical gradients so far. To address these limitations, we introduce a scheme to tailor chemical concentration and concentration gradients by controlling nanoscale chemical reactions at the interface. The approach harnesses the powerful capability of thermal probe lithography to precisely position and rapidly vary a spatiotemporal temperature profile at the surface using a thermal cantilever. We show that in the vicinity of a thermal tip used to drive a chemical reaction, concentration gradients are produced with a ∼20 nm spatial resolution and a full gradient reproducibility every ∼125 nm in at least one direction. Modeling of the reaction at the tip−surface interface indicates that the spatial concentration profile will decay as a triple exponential, providing a very high spatial resolution. Previous studies show that thermal lithography enables direct nanoscale physical and chemical changes at the surface of a broad range of materials. Localized heating with a thermal probe has been used to engrave refined topographical features into surfaces,16−18 where for example Pires and co-workers have used local desorption of a glassy organic resist19 and thermal decomposition of self-amplified depolymerization (SAD) polymers20 to gain exquisite 3-D control over topography. In a closely related approach, ThermoChemical NanoLithography (TCNL) transforms materials from one chemical state to another with up to ∼15 nm resolution via chemical reactions localized at the thermal tip.21 TCNL has been used to chemically nanopattern a broad range of materials including polymers,21−23 and block copolymer films,24 as well as to thermochemically fabricate nanostructures of piezoelectric ceramics,25 organic semiconductors,26,27 and graphene nanowires.28 Unlike the 3-D topographical thermal nanopatterning, these TCNL studies focus on binary nanowriting of chemical features where the material is either converted or not. In the present manuscript, we show how to achieve gray scale writing of chemical concentration with quantitative precision and sub-180 nm resolution. This strategy expands the versatility of the TCNL platform, which we illustrate with examples of complex, user-defined chemical concentration patterns and gradients fabricated with unprecedented spatial resolution. While the work here primarily focuses on the patterning of the concentration of chemical (amine) functionality at the surface, the method can easily be extended to the fabrication of gradients of other useful properties already manipulated by TCNL in other studies, like gradients of proteins, 23 conductivity in graphene, 28 fluorescence of conjugated molecules,26 or any other attribute that can be regulated by thermo-activated chemistries.

⎛ −EA ⎞ dP = k(P0 − P), k = A ·exp⎜ ⎟ dt ⎝ RT (x , y) ⎠

(1)

where t is time, P is the product concentration (i.e., the concentration of exposed amines at the given surface location), P0 is the maximum possible concentration of product, k is the rate constant, T is the temperature, EA is the activation energy for the reaction, and A is the Arrhenius constant. EA represents the energetic barrier hindering the dissociation of the protecting group; A is the maximum rate of reaction. EA and A can be determined approximately using thermogravimetric analysis (TGA) data. Detailed calculations are shown in Supporting Information and yield the values EA ≈ 134 ± 32 kJ/mol and A ≈ 1.7 × 1016 s−1. A successful model requires an accurate description of the temperature field driving the localized reaction, T(x,y,z), resulting from a cantilever tip. We assume that the temperature profile instantaneously reaches equilibrium upon tip contact (see Supporting Information) and use the static heat equation. The approximate solution, given TCNL boundary conditions, is: T (x , y) = T0 + (Tpeak − T0)e(−|x|−|y|)/ λ

(2)

where x and y are the x and y distance from the tip surfacecontact, λ is the decay length, T0 is room temperature, and Tpeak is the peak temperature at the surface. The decay length is estimated to be λ ≈ 100 nm (see Supporting Information) by using our proposed model to match the 12 nm resolution achieved on similar polymer surfaces.21 In the current study, we focus on the surface and hence neglect the z-dependence of the temperature, using only the surface profile. A temperature profile with Tpeak = 245 °C is shown in Figure 1a (dotted blue line). By combining eq 1 with eq 2, we obtain that for a surface in contact with a static tip for a dwell time, td, the spatial distribution of deprotected amines is P(x , y) = P0(1 − exp( −Atd exp−EA / RT(x , y)))

(3)

where P0 is the maximal achievable surface concentration of amines. Figure 1a shows the relative concentration, P/P0 for several dwell times at the same temperature. Even for a static cantilever and a time-independent temperature field, a spatially varying concentration is produced, albeit on relatively short length scales. This calculation clearly illustrates how TCNL achieves nanometer resolutions: the concentration profiles have a spatial variation that scales as an exponential of an exponential of an exponential, assuming an exponentially decaying temperature profile. Increasing the dwell time or temperature leads to increased production of the final product. In Figure 1b, the temperature dependence of the relative concentration profile at a fixed dwell time 50 ms is shown. Raising the temperature increases the concentration until a critical temperature is reached where the deprotection saturates yielding the maximum possible concentration, P0. These simulations show that generating higher concentrations results in a sacrifice of spatial resolution (see insets in Figure 1). Achieving a target concentration with maximum possible resolution will be an optimization problem



RESULTS AND DISCUSSION When a heated tip comes into contact with a substrate, the surface locally heats, which can induce a chemical change that transforms the substrate’s initial state, X, to a new state P. To quantify this transformation, we focus on measuring and modeling reactions on a known polymer containing tetrahydropyran-2-ylcarbamate protected amine groups,23 wherein thermal activation of the polymer leads to cleavage of tetrahydropyran and carbon dioxide (here referred to as deprotection), exposing reactive amines. The final surface concentration of the exposed amines is determined by the temperature profile at the tip−surface interface, the integrated time of exposure, and the activation energy for deprotection of the amine. Based upon the likely unimolecular mechanism for 8676

dx.doi.org/10.1021/la400996w | Langmuir 2013, 29, 8675−8682

Langmuir

Article

Figure 2. (a) Illustration of surface heating at two neighboring locations situated 50 nm apart. When the distance between the first and second contact of a static tip are sufficiently close, the spatially extended regions of the two chemical reactions will overlap. The final resultant surface concentration will depend on the integrated effect of both events rather than just the superposition of the independent profiles. This is shown in (b), which illustrates a simulation of two independent static events (left/red and right/blue) and the corresponding full deprotection profile (black) across the entire region due to the convolution of the two events. (T = 270 °C, td0 = td1 = 50 ms, Δx = (x1 − x0) = 50 nm, Δy = 0 nm.).

Figure 1. (a) Temperature profile (blue) decays from a peak at 245 to 50 °C over a distance of 200 nm. The dwell time of a static thermal tip strongly influences the width and the maximum concentration of the deprotected region, as shown by three curves corresponding to td = 1 ms, 25 ms, and 50 ms. For the selected temperature, the feature created using td = 50 ms is saturated (outermost curve, green), showing full deprotection of the amines over a width of ∼50 nm. (b) Concentration profiles of the deprotected amines versus applied tip− interface peak temperatures 120, 170, 195, 220, and 270 °C at fixed dwell time td = 50 ms. (Insets) Spatial resolution characterized by the full width at half max (fwhm) decreases with dwell time and temperature, where the black portion of the curves indicate 100% concentration.

image shown in Figure 3a is the result of these amine concentration gradients. We note that a higher concentration of amines (higher T) is related to a higher friction force, in agreement with the expectation that more hydrophilic and reactive regions give rise to higher friction forces. Figure 3c shows a zoom in and the corresponding friction and topography cross-section of the gradient area. The zoom was chosen in a region where the topography is flat, because in order to guarantee that the friction force is solely related to surface chemistry, the topography in that region needs to be extremely smooth and no convolution between topography and friction should be present. Figure 3c clearly indicates that this requirement is satisfied; furthermore it demonstrates that the change in concentration occurs with a spatial resolution of ∼16 nm, defined as the distance between three discrete measurement points along an extended path of 100 nm. Hence, chemical gradients produced by TCNL have sub-20 nm spatial resolution, with reproducibility every ∼125 nm in at least one direction. This is the highest spatial resolution reported for chemical gradients, to the best of our knowledge.1,6 The sharp spatial concentration profile of functional groups produced at the surface enables one to develop strategies to design high-resolution, large scale patterns with prescribed varying surface concentrations. We define the local concentration value at a given location to be c(xi,yi), where c is the peak value of the spatial profile and (xi,yi) are the coordinates of the peak. The pixel size of the resultant pattern will vary slightly

in finding the ideal combination of the applied temperature and dwell time. To measure the spatial resolution of TCNL in fabricating chemical gradients, a high resolution technique is necessary since TCNL has the potential to produce features on the nanoscale. Friction force microscopy is known to be very sensitive to chemical changes on surfaces and can achieve ∼1 nm resolution.29 Although the relationship between chemical concentration on the surface and friction force magnitude is complex and likely nonlinear, it is well-known that nanoscale friction is sensitive to even a minimal change in surface chemistry. For these reasons, friction force microscopy was chosen to detect surface chemistry changes related to changes of amine concentration on the polymer used in this work. In particular, TCNL was used to write alternate lines with a spacing of 125 nm in at high and low temperatures (corresponding to cantilever powers 7.8 mW, 4.8 mW), respectively. As shown in Figure 1b and Figure 2, two lines written at two different temperatures will give rise to two different overlapping spatial chemical gradients in the region around the respective tip-polymer contact areas. The friction 8677

dx.doi.org/10.1021/la400996w | Langmuir 2013, 29, 8675−8682

Langmuir

Article

To test the model’s prediction, we examined the relative deprotection of amine groups at the polymer surface versus temperature maintained at the tip-interface by writing an array of micrometer sized squares, each at a fixed temperature. Figure 4a shows the results of a typical experiment, where the numbers

Figure 3. (a) Force friction microscopy image of a TCNL patterned surface with alternating lines corresponding to high and low temperature spaced 125 nm apart. (b) Friction measurement of the fwhm of lines produced at increasing temperatures. (c) Zoom into region where the topography is flat. In this region, the friction varies significantly indicating a local gradient in concentration. Data was gathered every 8 nm.

since it depends on the corresponding fwhm of each concentration feature. Modeling (Figure 1b) and friction force measurements confirm this fwhm is weakly dependent on temperature (peak concentration), and the FFM data (Figure 3b) shows the fwhm is ∼180 nm at most. Production can be realized either by consecutive static contact events between the tip and the surface, or more efficiently, by translating the tip along the surface while maintaining contact. In both cases, a convolution of the chemical reaction between neighboring areas will result in the resultant product concentration as illustrated in Figure 2. The final concentration Ptot at a given location can be predicted by assessing the history of the surface treatment in the vicinity of that point. For a tip moving at a constant speed v in the x-direction, the temperature at a fixed location versus time will vary as the tip travels toward that position, on top of, and beyond it. Assuming the temperature profile of a translating tip is represented by eq 2, the time dependence of the temperature at a given spot can be described as a function of position using the relation. Making this substitution into the result for consecutive static TCNL treatments (see Supporting Information for details) and changing the summation to an integral over time, the concentration at a position (x0,y0) treated by a tip moving at constant speed will be: ⎛ ⎛ −A Ptot(x0 , y0 ) = P0⎜1 − exp⎜ ⎝ v ⎝

Figure 4. (a) Array of patterned square areas using incrementally increasing powers (temperatures) in the range from 5.0 to 11.0 mW. The product of the chemical reaction, thermally deprotected amines, are labeled with a NHS fluorescent dye for visualization. Areas fabricated at 7.8 mW or less are difficult to visualize. Scale bar: 10 μm. (b) Taking the average intensity of each patterned area and normalizing with respect to the maximum measured intensity, the relative intensity and hence the relative concentration realized by the chemical reaction is plotted versus power and peak temperature. The line represents an excellent fit of the model in eq 4 to the data. The bars are standard deviations.

⎞⎞

∫ dxte−E /RT(x ,y ,x ,y )⎠⎟⎠ A

0 0

t t ⎟

(4)

below each square correspond to the constant dissipated power (in mW) selected to fabricate each square. To analyze the results, we fluorescently labeled the exposed functional amine groups using a NHS fluorescent dye and imaged the surface with fluorescence microscopy. An increase in concentration is clearly depicted by the increasing intensities (see Supporting Information for detailed discussion of artifacts). To compare the experimental results with the model, we measure relative

As expected, when the tip speed increases, this result shows that the final concentration Ptot decreases since the tip is in contact with surface for a shorter period of time. We have assumed a static profile because given the length scales, the time scale to reaching a static temperature would be much faster than the time scale for the given speeds. A more detailed explanation is provided in the Supporting Information. 8678

dx.doi.org/10.1021/la400996w | Langmuir 2013, 29, 8675−8682

Langmuir

Article

Figure 5. (a) Normalized image of the user-defined pattern to be produced using TCNL with a rastered thermal cantilever. This image is input into the algorithm to calculate the local power that should be applied at the surface. (b) Calculated power map corresponding to (a) required to generate a user-defined pattern. (c) Experimentally produced user-defined pattern corresponding to (a). Scale bar: 20 μm. (d) Comparison of the experimental concentration profiles in (c) and the user-defined profiles in (a) shows good agreement.

intensity at the surface, I/I0, where I is the average intensity associated with a given square and I0 is the maximum average intensity from the brightest square. The ratio I/I0 is equivalent to the relative surface concentration if the average intensity of the bound dye is linearly proportional to the average concentration of the underlying amines, a reasonable assumption for the low concentrations of dye used. Figure 4b shows a plot of the extracted relative product concentration versus temperature. At lower temperatures, the relative concentration is constant or increases slowly, while at higher temperatures, the relative concentration saturates at P/ P0 =1, indicating that the maximal deprotection of the surface, P0, has been achieved. To examine our understanding the chemical reaction at the surface, and to extract a full calibration without extensive control experiments, we fit the proposed model (eq 4) to the data. The theory and the experiment agree quite well, as shown by the solid curve in Figure 4b. A leastsquares fit was used with two free parameters, the activation energy EA and the so-called tip efficiency, η, a surrogate for the peak temperature (see Method and Materials). The fit gave parameter values EA = 125 kJ/mol (which agrees well with the EA obtained by TGA) and η = 0.296. These data verify that it is possible to smoothly vary the product’s surface concentration from essentially zero to the maximum possible value using TCNL, where the resultant concentration depends nonlinearly on the applied temperature (see Supporting Information). Our experimentally validated model provides the calibration needed to pattern sophisticated chemical surfaces in an efficient and predictable manner. Here we integrate our results to create user-defined patterns of unprecedented chemical complexity and spatial resolution. Figure 5 shows several TCNL-fabricated concentration profiles with spatial changes on the micrometer scale: a sinusoidal variation, a linear variation, and a cubic variation in concentration. Quantitative comparison of the userdefined pattern with the experimental results shows good agreement (Figure 5d), including our ability to vary the relative

concentration linearly and nonlinearly by a factor of 3 over submicrometer length scales that are typically difficult or impossible to access using most other techniques.1 The sinusoidal pattern requires variation in concentration from minimum to maximum over just 6.25 μm. Decreasing the wavelength further, Figure 6 shows a sinusoidal pattern with varying the concentration by a factor of 4 over just two micrometers. Next, to explore the versatility of the TCNL platform, we fabricated chemical representations of complex images whose intensity variations require realization of high-resolution

Figure 6. Rapidly varying sinusoidal concentration profile with peak to minimum concentration separated by just ∼2 μm. Blurring due to diffraction makes analysis of such images difficult to accurately compare with the intended pattern. The average fluorescence in the minimum is increased due to this diffraction, as expected, since the intensities at positions nearby the minimum are larger. (Inset) Image of the sinusoidal surface pattern. The plotted data corresponds to peaks 2−4 in the image. Scale bar: 20 μm. 8679

dx.doi.org/10.1021/la400996w | Langmuir 2013, 29, 8675−8682

Langmuir

Article

users to generate nonbinary nanopatterns of a wide range of useful material properties.

chemical gradients. Figure 7 shows the results, where on the left are the original images and on the right are fluorescently



CONCLUSIONS The work described herein demonstrates the powerful capabilities of TCNL to fabricate nanoscale chemical gradients. Friction force microscopy experiments show that the inherent chemical gradients produced by a thermal cantilever have a high spatial resolution of ∼20 nm. Further implementation of TCNL in combination with an accurate model of the chemical reaction kinetics of a given material enables the systematic design and creation of variable chemical concentration patterns with sub-180 nm spatial resolution. Accordingly this technique provides a powerful new tool for nanoscale fabrication and should enable a wide range of previously inaccessible experiments and applications in fields as diverse as bioengineering, optoelectronics, nanophotonics, nanoelectronics, separations, and sensing.



METHOD AND MATERIALS

Polymer and Surface Preparation. Polymer films have been prepared as previously reported.23 For extensive details on the preparation of these surfaces, see the Supporting Information of that work. Thermochemical Nanolithography. We use an Agilent 5600 LS atomic force microscope (AFM) equipped with thermal cantilevers30,31 to achieve chemical patterning. The thermal cantilever is placed in series with a reference resistor (2 kΩ) biased with an applied voltage, V. We use the reference resistor to monitor the current and hence the dissipated power using a data acquisition device (National Instruments 6216-USB DAQ), where the dissipated power is adjusted through the bias voltage. Each tip is calibrated as previously reported31 in order to extract the linear relation between the heater temperature and the dissipated power. The peak temperature Tpeak at the tip− surface interface is adjusted through the cantilever heater temperature, Theater, by controlling the current level through the semiconducting cantilever, which leads to controlled resistive heating. The loss in thermal transport from the heater region to the surface is quantified by the efficiency parameter, η, given by η = (Tpeak − T0)/(THeater − T0), where T0 is the ambient surface temperature far from the tip. Further details regarding specific parameters used to realize each pattern in Figures 4−7 are described in the Supporting Information. The thermal cantilevers were fabricated in house by the King lab. Friction Force Microscopy. All FFM measurements were performed on an Agilent 5600 LS AFM. The scans were performed at scan rates of 8 μm/s and loads of ∼15−30 nN. The scans were performed with the same thermal tip that made the patterns. All AFM images were processed with Gwyddion and Matlab Software. Fluorescent Labeling, Fluorescence Microscopy, and Image Analysis. Patterned samples were fluorescently labeled using 100 μL of 25 nM NHS-DyLight 633 (ThermoScientifc) diluted in DMSO for approximately 1 h. The samples were then rinsed in DMSO and thoroughly washed with alternating cycles of PBS and distilled water for 5−10 min. Low dye concentrations were used to ensure linearity with concentration, to avoid quenching effects, and to minimize nonspecific binding. All samples were left in solution at room temperature for a fixed period of time (1 h); by fixing the concentrations and time period, we have removed any overlap with the chemical kinetics associated with the dye reacting with the exposed functional groups. Figures 4−7 were obtained using a TE2000 Nikon microscope equipped with a 40× 1.3 NA oil immersed objective and a Roper Scientific CoolSnap CCD camera (Figures 4−6) and an Andor iXon camera (Figure 7). In order to relate the fluorescence signal in the squares (Figure 4a) to a relative concentration, the fluorescence images were first corrected for uneven illumination. Then the average intensity for each square was measured from a centered, smaller square area (∼3 μm × 3 μm)

Figure 7. (a) (Left) Original image of the Mona Lisa, scaled and pixelated for input into the model in order to extract a power map. (Right) Experimental rendition of the Mona Lisa with a total width of just ∼32 μm produced by consecutive touchdowns of a static tip with a varying power at each position (Δx, Δy = 125 nm). Scale bar: 10 μm. (b) (Left) Rose and Driftwood, 1932, photograph by Ansel Adams, copyright 2012 The Ansel Adams Publishing Rights Trust. (Right) TCNL reproduction of the photograph. Scale bar: 10 μm.

labeled reproductions of Leonardo da Vinci’s Mona Lisa and Ansel Adam’s photograph Rose and Drif twood. The images were fabricated using a sequential treatment of the surface with a static thermal tip every 125 nm while applying the appropriate temperature determined by the model (see Supporting Information). The resultant chemical images illustrate the exquisite control achieved with TCNL technique in varying the surface concentration of functional groups. Prescribed average concentration differences at sub-500 nm scales should be attainable, although verification is complicated by optical diffraction and the convolution of friction with topography. We reason that the spatial resolution of this socalled effective concentration gradient is defined approximately as the distance between the centers of two neighboring patterned points of unique concentrations, where each spot’s individual spatial resolution is defined by its fwhm. Mathematically this is ∼(w1 + w2)/2 where w1 and w2 are the fwhm of the two features. Friction measurements show a range of fwhm from ∼100 to 180 nm depending on the applied temperature (see Figure 3b). Hence, the spatial resolution should be on the order of ∼180 nm or better. The approach developed in this work to vary amine surface concentration on the nanoscale should be straightforward to extend to other chemistries and more broadly to material properties other than concentration, such as fluorescence of conjugated polymers26 or conductivity in graphene.28 We envision that TCNL’s versatile ability to drive chemical reactions on many different classes of materials will enable 8680

dx.doi.org/10.1021/la400996w | Langmuir 2013, 29, 8675−8682

Langmuir

Article

to avoid artifacts at the edges (see Supporting Information for discussion about artifacts). Background correction was performed by subtracting the average intensity measured from the nonpatterned region. Lastly, to find the relative intensity, I/I0, the intensity value of the fitted curve at the plateau in Figure 4b was used to normalize each intensity. Image processing was performed in Image J. The extracted data set was then fit using a least-squares algorithm with the efficiency, η, and the activation energy EA left as fitting parameters (see Supporting Information). The concentration profiles shown in Figures 5d and 6 were extracted by averaging the intensity over the center 3 μm region and dividing by the maximal intensity I0 discerned from independently patterned squares to the left of the profiles (see Figure 5c) produced using the saturation temperature determined from the calibration curve.



(7) Belisle, J. M.; Correia, J. P.; Wiseman, P. W.; Kennedy, T. E.; Costantino, S. Patterning Protein Concentration Using Laser-Assisted Adsorption by Photobleaching, Lapap. Lab Chip 2008, 8, 2164−2167. (8) Dertinger, S. K. W.; Chiu, D. T.; Jeon, N. L.; Whitesides, G. M. Generation of Gradients Having Complex Shapes Using Microfluidic Networks. Anal. Chem. 2001, 73, 1240−1246. (9) Mosadegh, B.; Huang, C.; Park, J. W.; Shin, H. S.; Chung, B. G.; Hwang, S.-K.; Lee, K.-H.; Kim, H. J.; Brody, J.; Jeon, N. L. Generation of Stable Complex Gradients across Two-Dimensional Surfaces and Three-Dimensional Gels. Langmuir 2007, 23, 10910−10912. (10) Wu, H.; Huang, B.; Zare, R. N. Generation of Complex, Static Solution Gradients in Microfluidic Channels. J. Am. Chem. Soc. 2006, 128, 4194−4195. (11) Ginger, D. S.; Zhang, H.; Mirkin, C. A. The Evolution of DipPen Nanolithography. Angew. Chem. Int. Edn 2004, 43, 30−45. (12) Piner, R. D.; Zhu, J.; Xu, F.; Hong, S.; Mirkin, C. A. “Dip-Pen” Nanolithography. Science 1999, 283, 661−663. (13) Mirkin, C. A. The Power of the Pen: Development of Massively Parallel Dip-Pen Nanolithography. ACS Nano 2007, 1, 79−83. (14) Spatz, J. P. Nano- and Micropatterning by Organic−Inorganic Templating of Hierarchical Self-Assembled Structures. Angew. Chem., Int. Ed. 2002, 41, 3359−3362. (15) Arnold, M.; Hirschfeld-Warneken, V. C.; Lohmüller, T.; Heil, P.; Blümmel, J.; Cavalcanti-Adam, E. A.; López-García, M. ; Walther, P.; Kessler, H.; Geiger, B.; et al. Induction of Cell Polarization and Migration by a Gradient of Nanoscale Variations in Adhesive Ligand Spacing. Nano Lett. 2008, 8, 2063−2069. (16) Gotsmann, B.; Duerig, U.; Frommer, J.; Hawker, C. J. Exploiting Chemical Switching in a Diels−Alder Polymer for Nanoscale Probe Lithography and Data Storage. Adv. Funct. Mater. 2006, 16, 1499− 1505. (17) Coulembier, O.; Knoll, A.; Pires, D.; Gotsmann, B.; Duerig, U.; Frommer, J.; Miller, R. D.; Dubois, P.; Hedrick, J. L. Probe-Based Nanolithography: Self-Amplified Depolymerization Media for Dry Lithography. Macromolecules 2009, 43, 572−574. (18) Milner, A. A.; Zhang, K.; Prior, Y. Floating Tip Nanolithography. Nano Lett. 2008, 8, 2017−2022. (19) Pires, D.; Hedrick, J. L.; De Silva, A.; Frommer, J.; Gotsmann, B.; Wolf, H.; Despont, M.; Duerig, U.; Knoll, A. W. Nanoscale ThreeDimensional Patterning of Molecular Resists by Scanning Probes. Science 2010, 328, 732−735. (20) Knoll, A. W.; Pires, D.; Coulembier, O.; Dubois, P.; Hedrick, J. L.; Frommer, J.; Duerig, U. Probe-Based 3-D Nanolithography Using Self-Amplified Depolymerization Polymers. Adv. Mater. 2010, 22, 3361−3365. (21) Szoszkiewicz, R.; Okada, T.; Jones, S. C.; Li, T.-D.; King, W. P.; Marder, S. R.; Riedo, E. High-Speed, Sub-15 nm Feature Size Thermochemical Nanolithography. Nano Lett. 2007, 7, 1064−1069. (22) Duvigneau, J.; Schönherr, H.; Vancso, G. J. Nanoscale Thermal AFM of Polymers: Transient Heat Flow Effects. ACS Nano 2010, 4, 6932−6940. (23) Wang, D.; Kodali, V. K.; Underwood, W. D., II; Jarvholm, J. E.; Okada, T.; Jones, S. C.; Rumi, M.; Dai, Z.; King, W. P.; Marder, S. R.; et al. Thermochemical Nanolithography of Multifunctional Nanotemplates for Assembling Nano-Objects. Adv. Funct. Mater. 2009, 19, 3696−3702. (24) Duvigneau, J.; Schönherr, H.; Vancso, G. J. Atomic Force Microscopy Based Thermal Lithography of Poly(Tert-Butyl Acrylate) Block Copolymer Films for Bioconjugation. Langmuir 2008, 24, 10825−10832. (25) Kim, S.; Bastani, Y.; Lu, H.; King, W. P.; Marder, S.; Sandhage, K. H.; Gruverman, A.; Riedo, E.; Bassiri-Gharb, N. Direct Fabrication of Arbitrary-Shaped Ferroelectric Nanostructures on Plastic, Glass, and Silicon Substrates. Adv. Mater. 2011, 23, 3786−3790. (26) Wang, D.; Kim, S.; Underwood, W. D., II; Giordano, A. J.; Henderson, C. L.; Dai, Z.; King, W. P.; Marder, S. R.; Riedo, E. Direct Writing and Characterization of Poly(P-Phenylene Vinylene) Nanostructures. Appl. Phys. Lett. 2009, 95, 233108−3.

ASSOCIATED CONTENT

* Supporting Information S

Description of assumed temperature profile, further details of the chemical kinetics model, a description of TCNL patterning and artifacts, materials descriptions, and calculations of EA and A. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Author Contributions

K.M.C. performed the experiments. K.M.C., J.S., E.R., S.R.M., and J.E.C. conceived, designed, and analyzed the experiments. A.J.G. fabricated the polymer surfaces. W.P.K. fabricated thermal cantilevers. D.W. and V.K.K. ran initial experiments. All authors wrote the manuscript. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS Thanks to Yadong Zhang for materials synthesis. This work has been supported by the National Science Foundation CMDITR program DMR 0120967 (S.R.M., K.M.C., J.E.C., E.R.), MRSEC program DMR 0820382 (E.R.), CMMI 1100290 (E.R.), PHYS 0848797 (J.E.C.), the Office of Basic Energy Sciences DOE DE-FG02-06ER46293 (E.R.), a National Defense Science and Engineering Graduate Fellowship (A.J.G.), a NSF graduate research fellowship DGE-0644493 (A.J.G.), and a COPE fellowship (K.M.C., D.W.).



REFERENCES

(1) Morgenthaler, S.; Zink, C.; Spencer, N. D. Surface-Chemical and -Morphological Gradients. Soft Matter 2008, 4, 419−434. (2) Fuierer, R. R.; Carroll, R. L.; Feldheim, D. L.; Gorman, C. B. Patterning Mesoscale Gradient Structures with Self-Assembled Monolayers and Scanning Tunneling Microscopy Based Replacement Lithography. Adv. Mater. 2002, 14, 154−157. (3) Chaudhury, M. K.; Whitesides, G. M. How to Make Water Run Uphill. Science 1992, 256, 1539−1541. (4) Holden, M. A.; Cremer, P. S. Light Activated Patterning of DyeLabeled Molecules on Surfaces. J. Am. Chem. Soc. 2003, 125, 8074− 8075. (5) Scrimgeour, J.; Kodali, V. K.; Kovari, D. T.; Curtis, J. E. Photobleaching-Activated Micropatterning on Self-Assembled Monolayers. J. Phys.: Condens. Matter 2010, 22, 194103. (6) Waldbaur, A.; Waterkotte, B.; Schmitz, K.; Rapp, B. E. Maskless Projection Lithography for the Fast and Flexible Generation of Grayscale Protein Patterns. Small 2012, 8, 1570−1578. 8681

dx.doi.org/10.1021/la400996w | Langmuir 2013, 29, 8675−8682

Langmuir

Article

(27) Fenwick, O.; Bozec, L.; Credgington, D.; Hammiche, A.; Lazzerini, G. M.; Silberberg, Y. R.; Cacialli, F. Thermochemical Nanopatterning of Organic Semiconductors. Nat. Nanotechnol. 2009, 4, 664−668. (28) Wei, Z.; Wang, D.; Kim, S.; Kim, S.-Y.; Hu, Y.; Yakes, M. K.; Laracuente, A. R.; Dai, Z.; Marder, S. R.; Berger, C.; et al. Nanoscale Tunable Reduction of Graphene Oxide for Graphene Electronics. Science 2010, 328, 1373−1376. (29) Leggett, G. J.; Brewer, N. J.; Chong, K. S. L. Friction Force Microscopy: Towards Quantitative Analysis of Molecular Organisation with Nanometre Spatial Resolution. Phys. Chem. Chem. Phys. 2005, 7, 1107−1120. (30) Jungchul, L.; Beechem, T.; Wright, T. L.; Nelson, B. A.; Graham, S.; King, W. P. Electrical, Thermal, and Mechanical Characterization of Silicon Microcantilever Heaters. J. Microelectromech. Syst. 2006, 15, 1644−1655. (31) Nelson, B. A.; King, W. P. Temperature Calibration of Heated Silicon Atomic Force Microscope Cantilevers. Sens. Actuators, A 2007, 140, 51−59.

8682

dx.doi.org/10.1021/la400996w | Langmuir 2013, 29, 8675−8682