Dip-Coating Deposition on Chemically Patterned Surfaces: A

Feb 15, 2014 - The solvent evaporation stages of the processes, that occur subsequent to the capillary bridge failure events, also vary significantly ...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/Langmuir

Dip-Coating Deposition on Chemically Patterned Surfaces: A Mechanistic Analysis and Comparison with Topographically Patterned Surfaces Yan Wang and Thomas J. McCarthy* Polymer Science and Engineering Department, University of Massachusetts, Amherst, Massachusetts 01003, United States S Supporting Information *

ABSTRACT: Chemically patterned surfaces containing hydrophilic features on a hydrophobic background have been used by a number of groups to deposit arrays of particles/crystals/ substances by dip-coating deposition. In this technique, a substrate is simply withdrawn from a solution (or dispersion) of the desired substance, the solution dewets from the hydrophobic region and wets the hydrophilic features, and the particles/ crystals/substances deposit on the hydrophilic features after solvent evaporation. An apparently similar approach, recently described by several groups, involves dip-coating deposition of substances from solutions onto hydrophobic topographic features (arrays of posts on superhydrophobic surfaces) that are separated by air. We report results of dip-coating deposition using chemically patterned surfaces and compare them directly with results from post-containing superhydrophobic surfaces. This comparison involves the analysis of events at receding three-phase contact lines; these events differ significantly in the two approaches with the key difference being tensile (normal to the surface) versus sessile (parallel to the surface) capillary bridge failure. Tensile failure occurs with the post-containing superhydrophobic surfaces and sessile failure with chemically patterned surfaces. The solvent evaporation stages of the processes, that occur subsequent to the capillary bridge failure events, also vary significantly in the two approaches and depend on the receding contact angles of the hydrophobic post tops and the hydrophilic chemically patterned features. These differences, as the adjectives suggest, are pronounced. Controlling the evaporation rate (adjusting the vaporization/condensation equilibrium) by raising the partial pressure of the solvent is identified as a useful variable for chemically patterned surfaces.



INTRODUCTION There have been several recent reports of studies using pillarcontaining “superhydrophobic” surfaces as substrates for the preparation of patterned arrays of substances.1−6 Hatton and Aizenberg1 demonstrated that calcium carbonate deposits from a supersaturated aqueous solution selectively onto the tips of a nanopost array that was hydrophobized using a perfluoroalkylcontaining trifunctional silane. These workers also showed that poly(vinyl alcohol) and poly(L-lysine)1,2 selectively adsorb to and that colloidal particles can be attached to these post tops. Jiang’s group3 described a technique involving “clinging microdroplets” and showed that when puddles of solutions or suspensions were evaporated between parallel surfaces, one of which is a micropost-containing surface, also hydrophobized with a perfluoroalkyl-containing trifunctional silane, an array of microdroplets remained attached to the micropost tops. This technique was used to prepare arrays of NaCl crystals as well as protein, microsphere, and nanoparticle aggregates. Senez et al.4 described the distribution of a nonvolatile optical adhesive and particles (that were dispersed in water) on post tops relative to the direction of the receding contact line. They used a perfluoroalkyl-containing trifunctional silane to lyophobize an © 2014 American Chemical Society

overhanging post (reentrant geometry)-containing PDMS surface and were able to visualize the receding contact line of the cured optical adhesive using scanning electron microscopy. Our group reported studies similar to these5,6 that involved withdrawing superhydrophobic surfaces, containing posts modified with a monofunctional perfluoralkyl-containing silane, from aqueous NaCl and ionic liquid (with negligible vapor pressure) solutions to form controlled size individual crystals or controlled volume (in the femtoliter range) sessile drops. Wang et al.7 characterized these depositions on pillar arrays as “airgrid” surface patterning and pointed out a connection between these studies and those of Levkin’s group8−10 that used “superhydrophobic−superhydrophilic micropatterns” on porous materials to support arrays of sessile drops of aqueous solutions. Wang’s paper7 optimistically discusses the impact and potential impact of patterned arrays of various materials and gives leading references to both applications and preparative Received: January 13, 2014 Revised: February 13, 2014 Published: February 15, 2014 2419

dx.doi.org/10.1021/la5000149 | Langmuir 2014, 30, 2419−2428

Langmuir

Article

Sections were held near the center of a particular edge (defining the direction of withdrawal relative to the photolithography pattern), positioned so that the plane of the wafer was perpendicular to the liquid−vapor interface, and withdrawn straight up at a rate of ∼1 cm/s. Although simple and unsophisticated, this method is easily performed reproducibly. Aqueous solutions of NaCl and N+S− as well as neat N+S− liquid samples were prepared and used in 40 mL polystyrene vials (Fisher Scientific). Hexane and hexadecane solutions of TIPSDBC (0.16 mg/mL) were prepared and used in 20 mL glass scintillation vials (Fisher Scientific).

methods. The authors identify two main factors required of a patterned substrate for supporting/seeding/nucleating/docking a patterned array of materials: the repeating unit and the separating barrier. They express “excitement” concerning the “unprecedented” reports of using air as a “gas-phase separating barrier.” Their analysis is insightful, but we suggest an additional perspective on these preparative methods that we have found useful, that expands their arguments, and that may further justify their excitement. In contrast to the few recent reports1−6 of gas-phase separated pillars, there is a significant literature that addresses the use of chemically patterned surfaces for formation of patterned arrays of various materials.8−34 This chemical approach is arguably simpler, more generally accessible, and would be more easily scaled to larger areas and production applications than approaches that require the preparation of arrays of pillars. Photolithography and microcontact printing are now widely available, inexpensive tools and are logical choices for creating “repeating units” and “separating barriers”. In this article we compare these two approaches and address a key difference between smooth, chemically patterned and postcontaining surfaces as substrates for dip-coating deposition. Our perspective is that of the dewetting dynamics of bulk liquid during formation of sessile drops/puddles on patterned surfaces. This article focuses on the events that occur at the receding contact line and clearly differentiates tensile capillary bridge failure (normal to the surface), with surfaces containing gas-phase separated posts, and sessile capillary bridge failure (parallel to the surface), which occurs with chemically patterned surfaces.





RESULTS The results that we report in this article concern dip-coating deposition onto smooth, flat, chemically patterned surfaces. We present these first with brief analysis and compare them, in the Discussion section, with results that were recently obtained using superhydrophobic surfaces containing pillars.5,6 An important initial objective in undertaking this work was to develop chemically patterned alternatives to the postcontaining surfaces. In light of the obvious advantages noted in the Introduction, if chemically patterned substrates can be prepared that are functionally equivalent to those based on pillar arrays, the chemical approach would be the obvious choice of most any research or development group. In the course of this work it became clear that there are aspects of the mechanism of this process (dip-coating deposition) that have not been addressed, analyzed, or (perhaps even) considered. The discussion of these mechanistic aspects is the important substance of this article. The preparative aspects of the chemically patterned surfaces used are neither novel nor particularly impressive in quality, but they were required for and are sufficiently reproducible for direct comparison with the pillar-containing substrates. Others have demonstrated8−34 that chemical pattern approaches for dip-coating deposition are functional and useful, at least for demonstrations with specific systems, and some used more sophisticated and more refined lithography than that which we report. The mechanistic analysis that we present, however, identifies rational approaches for controlling process variables that can readily be implemented, should certainly be considered, and that should make chemically patterned dip-coating deposition more general and more practical. We report results from five different types of experiment. Four of these involve dip-coating deposition of either bis(hydroxyethyl)dimethylammonium methanesulfonate (N+S−) or sodium chloride (NaCl) onto a patterned array of features that was chosen (arrived at) by carrying out hundreds of preliminary experiments. The size scale, shape, separation, and orientation (relative to the liquid−vapor interface) of the features, the concentration of the solutions, and the rate of withdrawal were all addressed as variables in initial experiments. Figure 1 shows the structure of N+S−, the chemical pattern of the features on the silicon substrates that were used for the first four experiment types that we describe, and a schematic of the simple dip-coating procedure. N+S− and NaCl were chosen for comparison with reported data.5,6 N+S− was used for dipcoating either as a neat liquid or as a 25 vol % aqueous solution. Aqueous NaCl (0.5M) was used with two experimental protocols. One involved no control of the evaporation rate, and this method generated data that varied significantly with the ambient relative humidity. This fact was not appreciated during initial experimentation. The other method, which is simple in retrospect, but took many months to adopt, involved humidity control and staged evaporation. All reported data

EXPERIMENTAL SECTION

Silicon wafers (100 mm diameter) were obtained from International Wafer Service (100 orientation, P/B doped, resistivity from 20 to 40 Ω cm). Photolithography masks were prepared using AutoCAD and printed on polyester film by CAD/Art Services, Inc. A Shipley S1813 photoresist and a Microposit MF-321 developer were obtained from Microchem Corp. UV exposure was carried out using a 500 W (365 nm) OAI UV lamp. Tridecafluoro-1,1,2,2-tetrahydrooctyldimethylchlorosilane (RFSiMe2Cl, Gelest), sodium chloride (Fisher Scientific), potassium chloride (EM Science), n-hexadecane (Acros Organics), hexanes (Fisher), and bis(hydroxyethyl)dimethylammonium methanesulfonate (N+S−, Evonik) were used without further purification. 7,14-Bis(triisopropylsilylethynyl)dibenzo[b,def ]chrysene (TIPS-DBC) was prepared using a literature method35 and recrystallized from hexane. Water was deionized using a Millipore Milli-Q system that involves reverse osmosis and filtering steps (18 MΩ). Scanning electron microscopy (SEM) images were recorded using a JEOL Neoscope, and optical microscopy images were acquired with an Olympus BX60 system. Viscosity measurements were conducted using an Advanced Rheometry AR2000 parallel plate rheometer. Sections of silicon wafers (∼1 cm × ∼3 cm) were cleaned using a Harrick plasma cleaner operated at 300 mTorr (O2) and 30 W for 15 min prior to being spin-coated with S1813 photoresist (4000 rpm for 60 s). After heating at 115 °C for ∼90 s to remove solvents (soft bake), the wafers were exposed to UV through a photolithography mask for 12 s. Development comprised submersion in a MF-321 developer for 50 s, rinsing with deionized water for 1 min and drying in a nitrogen stream. The wafer was then treated with oxygen plasma (300 mTorr, 30 W, 5 min) to remove contaminants from the developed areas before being allowed to react with RFSiMe2Cl vapor in a closed container at 80 °C for 72 h. After this surface modification reaction, the remaining photoresist was removed by rinsing with toluene, acetone, and water (in this order). Dip-coating deposition involved carefully withdrawing silicon wafer sections from solutions using hand-held stainless steel tweezers. 2420

dx.doi.org/10.1021/la5000149 | Langmuir 2014, 30, 2419−2428

Langmuir

Article

Figure 1. (a) Schematic depiction of the dip-coating deposition procedure. An initially submerged silicon wafer section is withdrawn from the liquid (using tweezers) with the surface of the wafer perpendicular to the liquid−vapor interface. (b) Pattern of Si/SiO2 features (30°/60°/90° triangular spots) on a perfluoroalkyl monolayer background. Wafer sections containing this pattern were withdrawn in the direction pointing up in the figure for all experiments reported using this pattern. (c) Structure of TIPS-DBC. (d) Structure of N+S−.

involved withdrawing the silicon wafer section by hand, using tweezers at a rate of ∼1 cm/s, at orientations of