Decoupling Substrate Surface Interactions in Block Polymer Thin Film

Jun 29, 2015 - We report a highly predictive approach to capturing the major substrate–polymer interactions that can dominate nanoscale ordering and...
0 downloads 10 Views 3MB Size
Download PDF
Article pubs.acs.org/Macromolecules

Decoupling Substrate Surface Interactions in Block Polymer Thin Film Self-Assembly Cameron K. Shelton† and Thomas H. Epps, III*,†,‡ †

Department of Chemical & Biomolecular Engineering and ‡Department of Materials Science and Engineering, University of Delaware, Newark, Delaware 19716, United States S Supporting Information *

ABSTRACT: We report a highly predictive approach to capturing the major substrate−polymer interactions that can dominate nanoscale ordering and orientation in block polymer (BP) thin films. Our approach allows one to create designer BP thin films on modified substrates while minimizing the need for extensive parameter space exploration. Herein, we systematically and quantitatively examined the influence of substrate surface energy components (dispersive and polar interactions) on thin film self-assembly, and our analysis demonstrates that although total surface energy plays a dominant role in substrate wetting, individual contributions from the dispersive and polar components of the surface energy influence the composite through-film behavior. Additionally, long-range forces as described by the Hamaker constant are under-recognized factors in thin film assembly and can alter expected wetting behavior by affecting thermodynamic stability. This more inclusive interpretation of surface energy effects, including the Hamaker constant, on BP thin films was supported by studies of interfacial and through-film behavior as gleaned from temporal island/hole measurements via in situ optical microscopy during thermal annealing. The formalism correctly predicted experimental wetting and hole formation sizes over a wide range of substrate surface energies when employing the appropriate relationships based on decoupled dispersive and polar components. Our results indicate a promising and more universal approach for matching desired BP thin film self-assembly with chemically tailored substrate modifications.

■

ΔγA = |γS − γA|

INTRODUCTION

Block polymers (BPs) have received significant attention because they can self-assemble into nanostructures (e.g., spheres, cylinders, networks, and lamellae) with periodicities from 5 to 100 nm, which are ideal for thin film applications including nanolithographic masks, nanotemplates, nanoporous membranes, and organic solar cells.1−7 Although bulk BP selfassembly is dependent largely on the Flory−Huggins interaction parameter (χ), the degree of polymerization (N), and the block volume fractions ( f), thin film self-assembly is influenced by additional confinement parameters such as film thickness (t), substrate surface interactions, and/or free surface interactions.5,8−12 Furthermore, as the number of potential BP systems continues to increase,13 experimental investigation of the full parameter space in a given system is no longer feasible and must be replaced by predictive tools that incorporate the nuanced effects of surface interactions and film thickness on nanoscale morphology, ordering, and orientation.14 Previous studies suggest that substrate−polymer interactions are some of the most influential factors controlling polymer wetting behavior,15,16 nanostructure orientation,17−20 nanostructure uniformity through the film,21−23 and defect density 24,25 in BP thin films. The substrate−polymer interactions typically are quantified by the interfacial energy (ΔγA) or difference between the substrate (γS) and individual polymer block (γA) surface energies as shown in eq 1. © 2015 American Chemical Society

(1)

Smaller magnitudes of ΔγA indicate more favorable enthalpic interactions between a particular polymer block and the substrate.4,26,27 Although eq 1 often is used for substrate− polymer interaction analyses, Han et al. demonstrated its potential limitations when annealing poly(styrene-b-methyl methacrylate) thin films on poly(styrene-r-methyl methacrylate) brushes modified with either a terminal hydroxyl, sidechain hydroxyl, or side-chain epoxy group.17 Although the total surface energies were similar on all three random copolymer brushes (for a given polystyrene composition in the brush), neutral wetting behavior occurred at different polystyrene compositions due to the differing polarity of functional groups; this effect is not described by eq 1. Originally described by Fowkes, γ can be decoupled into individual dispersive (induced dipole) and polar (permanent dipole) components to better represent intermolecular interactions.28 Using a modified version of Fowkes’ geometric mean approach to define surface wetting, Owens and Wendt calculated the interfacial energy between a substrate (S) and polymer block (A) as Received: April 20, 2015 Revised: June 10, 2015 Published: June 29, 2015 4572

DOI: 10.1021/acs.macromol.5b00833 Macromolecules 2015, 48, 4572−4580

Article

Macromolecules ΔγA = γS + γA − 2 γSDγAD − 2 γSPγAP

total, dispersive, and polar surface energy interactions, we employed a series of chlorosilane monolayers combined with UVO treatments. Resulting substrate−polymer interaction effects were quantified visually by imaging island/hole formations. Additionally, changes in island/hole growth rates as a function of surface energy were studied by thermally annealing films during in situ optical microscopy. Experimental results were paired with comparable predictive equations utilizing total and/or decoupled surface energy components to determine the impact of individual substrate−polymer interactions on substrate wetting behavior and through-film self-assembly.

(2)

in which the superscripts D and P refer to dispersive and polar components, respectively.29 By decoupling the total surface energy into its dispersive and polar components, this equation separates the long-range and attractive interactions between polar/dispersive moieties from the short-range and repulsive total surface energy effects at contacting surfaces. However, the additional parameters in the surface wetting equation necessitate a more comprehensive analysis to determine which components (total, dispersive, and/or polar interactions) are significant for a given system. Decoupled interfacial energy equations to describe the qualitative wetting behavior of BP thin films have been applied in the literature,15,16,19 but to the authors’ knowledge, no studies to date cover a suitable range of distinct total, dispersive, and polar surface energies to fully define the individual component contributions. Typically, substrate modification methods such as random copolymer brushes,5,25,30 chlorosilane monolayers,16,19,31 or X-ray or ultraviolet ozone (UVO) exposure15,32,33 are used to alter the substrate surface energy for analysis. Random copolymer brushes are effective in creating neutral substrates but normally cover only the range of surface energies between the two polymer blocks.34 Additionally, random copolymers can require tedious synthesis depending on the BP system.13 As an alternative, chlorosilane monolayers are small-molecule mimics that cover a range of different total (≈20−45 mJ/m2), dispersive, and polar surface energies depending on the functional groups of the chosen chlorosilane.31,35 Finally, subjecting polymer or chlorosilane monolayers to X-ray or UVO treatments increases the total surface energy of the substrate by chemically modifying the monolayer with oxygen-functionalized moieties (carboxyl, carbonyl, etc.); however, this technique predominantly modifies only the polar component.33,36−38 As a visible complement to direct surface energy measurements, literature has shown that surface relief structures, more commonly referred to as “islands” and “holes”, can provide detailed information about substrate surface effects, both at the substrate−polymer interface and through the film thickness.15,16,33,38−40 Islands/holes are thicker/thinner regions on the film that form at the free surface to reduce thermodynamically unfavorable stretching and compression of polymer chains attempting to conform to an incommensurate film thickness.41 In a diblock copolymer, if the same polymer block wets the substrate and free surface (symmetric wetting), commensurate film thicknesses occur at integer multiples (n) of the polymer domain spacing (L0).3,4 If a different polymer block wets the substrate and free surface (asymmetric wetting), commensurate film thicknesses occur at half-integer multiples (0.5 + n) of L0.3,4 Consequently, because the lower surface energy block is known to wet the free surface,5 the topography at integer and half-integer L0 thicknesses only depends on which polymer block wets the substrate surface.15 With regard to through-film effects, Peters et al. used n-octadecyltrichlorosilane monolayers exposed to X-ray radiation at various dose rates to show qualitatively that the size of island/hole formations on a poly(styrene-b-methyl methacrylate) film was related directly to the substrate interfacial interaction, as larger incompatibilities between substrate−polymer surface energies resulted in larger island/hole formations.38 Herein, to achieve the widest range and most detailed analysis of substrate surface effects resulting from changes to

■

EXPERIMENTAL SECTION

Substrate Surface Modification. Benzyldimethylchlorosilane (benzyl silane), 2-acetoxyethyldimethylchlorosilane (aceto silane), 3methacryloxypropyldimethylchlorosilane (methacryl silane), and nbutyldimethylchlorosilane (n-butyl silane) (Gelest Inc.) were used as received. Silicon wafers (bare silica) (Wafer World Inc.) were rinsed with toluene and dried with compressed nitrogen gas (Keen Compressed Gas Co.), processed in a UVO cleaner (model 342, Jelight Co., Inc.) for ≈1 h, and rerinsed with toluene before use. Chlorosilane monolayers were deposited on cleaned silicon substrates using liquid or vapor deposition techniques as appropriate. Liquid deposition was carried out by pipetting chlorosilanes directly onto substrate surfaces with a Pasteur pipet. The chlorosilane was allowed to react for 2 h. Vapor deposition was performed in a glass chamber containing cleaned substrates and Teflon reservoirs to hold pure chlorosilanes.31 Dynamic vacuum was pulled on the glass chamber for 5 h to allow the chlorosilane enough time to react with the substrate surface. After chlorosilane deposition, substrates were rinsed with toluene to remove unreacted chlorosilane and dried with compressed nitrogen gas. Modified substrates were stored in a desiccator until film casting or UVO treatment depending on the sample. Substrates with benzyl silane and n-butyl silane monolayers were subjected to UVO treatments to investigate additional substrate surface energies. Exposure to UVO radiation for various times altered the measured water and diiodomethane contact angles in a predictable fashion.12,33 Surface Energy Measurements. Substrate surface energy was determined by depositing 3 μL droplets of diiodomethane (99% stabilized, Acros Organic) and water (purified with a Milli-Q reagent water purification system) onto substrates and measuring the contact angle (θ) using a goniometer (First Ten Ångstroms FTÅ 125). Angle analysis was performed using FTÅ software and the drop shape method applied after the drop shape had stopped changing (typically 0.1 s for water and 0.3 s for diiodomethane). The Owens and Wendt method, an extension of the Good and Girifalco geometric mean approximation, was utilized to calculate surface energy dispersive (γDS ) and polar (γPS ) terms from contact angle measurements.29 The Owens and Wendt method was chosen over other suitable surface energy methods (Wu, Good−Girifalco−Fowkes, etc.) for two reasons.42,43 First, the Owens and Wendt method requires only two contact angle fluids for measurements, which can simplify analysis especially for cases in which modified substrate surfaces adversely interact with (or are dissolved by) common contact angle fluids. Second, surface energy values calculated via the Owens and Wendt method generally were in very good agreement with the literature results examined herein.37 Using the Owens and Wendt method, eqs 3 and 4 were solved simultaneously. D (1 + cos θ)γL,water = 2( γSDγL,water + D (1 + cos θ)γL,diio = 2( γSDγL,diio +

P γSPγL,water )

(3)

P γSPγL,diio )

(4)

In eqs 3 and 4, γL refers to the liquid surface energy (mJ/m ), with dispersive (γDL ) and polar (γPL) components, for water or diiodomethane (diio). The liquid surface energy values were γL,water = 72.8 2

4573

DOI: 10.1021/acs.macromol.5b00833 Macromolecules 2015, 48, 4572−4580

Article

Macromolecules

Table 1. Average Water and Diiodomethane Contact Angle Measurements along with Corresponding Surface Energies for Bare Silica and Chlorosilane Monolayer Substrates contact anglea (deg) substrate bare silica aceto silane benzyl silane methacryl silane n-butyl silane

water 6.4 69.5 81.6 80.7 91.6

± ± ± ± ±

0.3 2.0 0.6 0.7 1.2

surface energy (mJ/m2)

diiodomethane 33.5 49.0 43.1 49.9 60.2

± ± ± ± ±

total

1.0 0.2 0.4 0.4 0.8

77.4 44.0 41.3 38.7 30.6

± ± ± ± ±

dispersive 0.5 0.8 0.3 0.3 0.6

42.7 34.8 38.0 34.3 28.4

± ± ± ± ±

0.4 0.1 0.2 0.2 0.5

polar 34.7 9.1 3.3 4.4 2.2

± ± ± ± ±

0.3 0.8 0.2 0.3 0.3

a

The uncertainty in the contact angle represents one standard deviation of the data from repeated measurements and was propagated through calculations to determine surface energy uncertainty.

mJ/m2, γDL,water = 21.8 mJ/m2, γPL,water = 51 mJ/m2, γL,diio = 50.8 mJ/m2, γDL,diio = 50.8 mJ/m2, and γPL,diio = 0 mJ/m2.44 Reported contact angle measurements were averaged overall multiple spots (≈10) for each sample. One standard deviation of the data from repeated measurements was taken as the experimental uncertainty for a given sample. The experimental uncertainty was propagated through interfacial energy equations in the final analyses. Polymer Film Preparation. A poly(methyl methacrylate-b-n-butyl acrylate) (PMMA-PnBA) BP was purchased from Polymer Source Inc. and used as received. The polymer molecular weight was 69 kg/mol, dispersity was 1.1, and PMMA block volume fraction was 0.33 as determined by size exclusion chromatography (Viscotek, GPCmax VE2001) and proton nuclear magnetic resonance (Bruker, AVX400). An L0 of 41 nm was measured using atomic force microscopy (AFM; Veeco Dimension 3100) and thickness differences between successive commensurate thicknesses. A hexagonally perforated lamellae (HPL) morphology was suggested for PMMA-PnBA thin films as AFM revealed perpendicular cylinders at the free surface where islands/holes were present, a phenomenon caused by HPL structures.19,45 Additionally, a bulk lamellar morphology and increase in L0 to 45 nm was measured with small-angle X-ray scattering (SAXS); a similar increase in L0 for an HPL to lamellar morphology shift is reported in the literature.46 To cast BP films, PMMA-PnBA was dissolved in tetrahydrofuran (Fisher Scientific, uninhibited) to generate a 2.1 wt % polymer solution. The solution was cast onto bare silica and modified substrates using a flow coating setup to generate either uniform or gradient thickness films.47 Film thickness was measured using a spectral reflectometer (Filmetrics, F20-UV). Hole sizes were analyzed on films at incommensurate thicknesses (88−90 nm for symmetric wetting films and 60−64 nm and 106−108 nm for asymmetric wetting films). At these thicknesses, micron-scale bicontinuous surface structures gave way to individual hole formations. Films were subjected to dynamic vacuum (20 mTorr) at room temperature to remove residual solvent before annealing under vacuum at 175 °C for 24 h to promote island/ hole formation. After 24 h, films were removed from the vacuum oven and quenched on a metal plate to bring the films to room temperature and kinetically trap the microstructures. Tracking Growth of Islands/Holes. BP thin films were imaged using an optical microscope equipped with a Nikon Eclipse LV100 5 MP CCD camera. Island/hole formations in optical micrographs were sized using ImageJ software. The procedure to analyze images was standardized into four steps: enhance contrast, remove noise, convert image to black and white, and measure feature sizes. Resolution dictated a minimum measurable feature size of 0.3 μm2. Additionally, circularity (related to the ratio of area to perimeter and equivalent to one for a circle) was set to 0.75 or greater to remove formations that bridged during image analysis. The average area, standard deviation, and number of formations were recorded for use in data analysis. In situ optical microscopy was performed using an airtight Linkam thermal stage with a viewing window. Samples were prepared by placing polymer films on the Linkam stage while operating inside an argon glovebox and then sealing the stage to isolate it from the atmosphere. The stage was removed from the glovebox and then placed under the optical microscope. The stage was heated from room

temperature to 175 °C at a heating rate of 3 °C/min, similar to the heating rate of the vacuum oven. The in situ anneal was monitored for 24 h with images captured every 5−15 min initially and every 30 min once the island/hole sizes began to plateau. After 24 h, the temperature of the Linkam stage was decreased to room temperature, and the film was removed from the viewing stage.

■

RESULTS AND DISCUSSION For this surface energy study, PMMA-PnBA was chosen as a model material because the individual polymer blocks had similar dispersive but different polar and total surface energies (PMMA: γtotal = 41.1 mJ/m2, γD = 29.6 mJ/m2, γP = 11.5 mJ/ m2; PnBA: γtotal = 33.7 mJ/m2, γD = 30.4 mJ/m2, γP = 3.3 mJ/ m2 at 20 °C).44 Additionally, the change in dispersive component surface energy as a function of temperature (dγ/ dT) was nearly identical between the two blocks (PMMA = −0.076 mJ/(m2 K); PnBA = −0.070 mJ/(m2 K)), thereby maintaining the difference in surface energies between PMMA and PnBA at the annealing temperature (175 °C) herein.44 By minimizing dispersive component differences between the two blocks and altering polar interactions through substrate modification, total and decoupled surface energy effects could be explored in a systematic manner. Modified substrates were prepared on bare silica surfaces using chlorosilane monolayers combined with UVO treatments to span a total substrate surface energy range of 30.6−77.4 mJ/m2, dispersive surface energy range of 28.4−42.7 mJ/m2, and polar surface energy range of 2.2−34.7 mJ/m2. The total, dispersive, and polar surface energy components of the bare silica and chlorosilanemodified substrates used in this study are listed in Table 1 and match well with reported measurements found in the literature.15,16,31,48 Gradient thickness PMMA-PnBA films were coated onto modified substrates via flow coating to determine commensurate and incommensurate thicknesses.47 Substrate Wetting Behavior Determined by Total Surface Energy. To promote the formation of islands/holes, films were thermally annealed under vacuum for 24 h at 175 °C. Film thicknesses were kept below 110 nm (2.7L0) to allow the substrate surface effect to propagate to the free surface.22 Wetting behavior was determined by locating island/hole formations at known film thicknesses using optical microscopy. For all substrates, PnBA was expected to wet the free (air) surface, as it had a lower total surface energy (33.7 mJ/m2) than PMMA (41.1 mJ/m2).5 Thus, symmetric or asymmetric wetting occurred when PnBA or PMMA wet the substrate surface, respectively. As shown in Figure 1, films coated on bare silica, aceto silane, and benzyl silane substrates were featureless at a half-integer L0 thickness and showed evidence of islands/holes at an integer L0 thickness. Films coated on methacryl silane substrates showed 4574

DOI: 10.1021/acs.macromol.5b00833 Macromolecules 2015, 48, 4572−4580

Article

Macromolecules

Initially, the experimental wetting behavior was compared to predictions from the standard decoupled (eq 2) and total (eq 1) surface energy equations. This comparison is shown in Figure 2. The decoupled surface energy equation matched the experimental results for all substrates except benzyl silane, for which the prediction was symmetric wetting, as opposed to the experimental result of asymmetric wetting. The total surface energy equation matched the experimental results for all substrates except methacryl silane, for which the prediction was asymmetric wetting, as opposed to the experimental result of symmetric wetting. Shaded quadrants on the plot indicate the discrepancies between predicted (top axis) and experimental (right axis) behavior. To eliminate inconsistencies between the experimental and predicted results, the Hamaker constant was introduced to define the long-range attractive or repulsive interactions. The Hamaker constant describes the attractive or repulsive nature of van der Waals interactions between two or more surfaces and is directly proportional to the total surface energy.49 Positive Hamaker constants indicate a long-range, attractive force between the substrate and polymer, and negative Hamaker constants indicate a long-range, repulsive force between the substrate and polymer.50−52 For the three-layer system at the substrate−polymer interface (substrate, PMMA block, PnBA block) the Hamaker constant becomes negative if the middle layer’s surface energy is the maximum, i.e., ΔγPMMA < ΔγPnBA and γPMMA > γS.49,53−56 The remaining polymer domains are constrained to the layering behavior driven by the preferential substrate−polymer wetting. Including the Hamaker constant in total surface energy analyses led to three cases that completely explained experimental substrate wetting behavior. The description of each case and comparison with experimental results are shown in Figure 3. Case 1 describes when the substrate interfacial energy of PMMA is less than that of PnBA, and the Hamaker constant is positive. Therefore, the PMMA block segregates to the substrate surface, resulting in asymmetric wetting. In case 2, the PMMA has a lower substrate interfacial energy than PnBA, but the Hamaker constant is negative for the system. Therefore, PnBA segregates to the substrate interface to reduce overall repulsive interactions in the

Figure 1. Optical microscopy images of commensurate (tC) and incommensurate film thickness regions for bare silica and chlorosilane monolayers as indicated by presence (or lack) of islands/holes. Bare silica, aceto silane, and benzyl silane substrates constrained PMMAPnBA to asymmetric wetting as commensurate thicknesses were noted at a half-integer L0 thickness. Films cast on methacryl- and n-butyl silane substrates exhibited symmetric wetting as a half-integer L0 thickness showed island/hole formations. The location of island/hole formations indicated bare silica, aceto silane, and benzyl silane substrates were preferential for PMMA, and methacryl silane and nbutyl silane substrates were preferential for PnBA. The scale bar represents 10 μm and applies to all micrographs.

islands/holes at a half-integer L0 thickness and were featureless at an integer L0 thickness (see Figure 1). Distinct island/hole formations were not present on 100 nm or thicker films on nbutyl silane substrates most likely due to the substrate/free surface competition,21 but thinner films (62 nm) on n-butyl silane substrates exhibited clear island/hole formations at incommensurate thicknesses. Therefore, experimental results indicated bare silica, aceto silane, and benzyl silane substrates were preferential for PMMA, while methacryl silane and n-butyl silane substrates were preferential for PnBA.

Figure 2. Comparison between predicted (dotted line intersecting x-axis) and experimentally determined (dotted line intersecting y-axis) wetting behavior of PMMA−PnBA on different substrate surfaces using decoupled (black circles) and total (white squares) surface energy approaches. Shaded quadrants represent incongruities between predicted and experimental results. Negative interfacial energy differences predict that the PnBA block has a lower interfacial energy with the substrate and will wet the substrate surface, while positive interfacial energy differences predict that the PMMA block has a lower interfacial energy and will wet the substrate surface. The PnBA block was expected to wet the free surface in all cases due to its lower total surface energy. 4575

DOI: 10.1021/acs.macromol.5b00833 Macromolecules 2015, 48, 4572−4580

Article

Macromolecules

Figure 3. Scheme depicting wetting behavior in BP thin films using three cases. In case 1, PMMA has a lower interfacial energy than PnBA, and a net long-range, attractive force is present, resulting in asymmetric wetting. In case 2, long-range, repulsive interactions exist at the substrate−PMMA interface as the higher surface energy PMMA layer separates the lower surface energy substrate and PnBA layers. Therefore, PnBA segregates to the substrate, stabilizing the substrate−polymer interactions and resulting in symmetric wetting. In case 3, PnBA has a lower interfacial energy, there is a net attractive interaction in the system, and symmetric wetting is predicted. Aligning each case to the appropriate system, as shown in the plot, accurately matched predicted and experimental results.

Figure 4. Hole sizes on aceto silane, benzyl silane, methacryl silane, and n-butyl silane surfaces (black circles), UVO-treated benzyl silane and n-butyl silane surfaces (gray circles), and bare silica (white circles) substrates. Two equations were utilized to compare with hole size: (a) eq 5, in which the attractive forces were described using the total surface energy, and (b) eq 2, in which the attractive forces were decoupled. The dotted red line represents a linear fit to the data. Though the two fits have similar R2 values (0.92 for the total surface energy equation and 0.96 for the decoupled surface energy equation), the total surface energy equation did not fully capture the expected trend in hole size, with the major discrepancy noted at lower interfacial energy differences, likely leading to the somewhat lower R2 value. On the other hand, the decoupled surface energy equation (R2 = 0.96) more accurately captured the experimental trend over the entire data range.

system. This effect results in symmetric wetting. In case 3, PnBA has a lower substrate interfacial energy, and the Hamaker constant is positive, leading to symmetric wetting. From surface energy values calculated for each modified substrate, case 1 applied to bare silica, aceto silane, and benzyl silane substrates, case 2 applied to methacryl silane substrates, and case 3 applied to n-butyl silane substrates. By including Hamaker constant stabilities, each case accurately predicted the experimental wetting behavior on all surfaces. This result suggests that the

inclusion of both short- and long-range repulsive interactions is necessary to describe the complete surface energy/wetting behavior. Island/Hole Size Determined by Decoupled Surface Energy. At film thicknesses greater than the contacting substrate−polymer layer, the short-range and long-range forces define substrate−polymer interactions with larger incompatibilities, creating stronger driving forces for nanostructure ordering and defect annihilation.28,38 To examine the impact 4576

DOI: 10.1021/acs.macromol.5b00833 Macromolecules 2015, 48, 4572−4580

Article

Macromolecules

Figure 5. (a) Schematic of in situ anneal setup with a sealed Linkam thermal stage under an optical microscope. The Linkam thermal stage heated the film to 175 °C using a 3 °C/min heat ramp for consistency with vacuum oven annealed samples. (b) Size measurements from holes produced on bare silica and benzyl silane substrates over 24 h showed expected nucleation and growth kinetics, followed by island/hole stabilization at equilibrium sizes, that match those from oven annealed samples. The scale bar represents 20 μm and can be applied to all micrographs.

methacryl silane, and n-butyl silane substrates for which total vs decoupled surface energy equations separately predicted dissimilar substrate−polymer block preferences. For example, benzyl silane substrates had a total surface energy close to PMMA but a polar surface energy close to PnBA. Based on these preferences, each equation predicted different hole sizes for films on benzyl silane substrates, but only the decoupled equation’s predictions aligned with experimental results. Thus, this decoupling of the surface energy into dispersive and polar components was necessary to describe through-film selfassembly mechanisms as the long-range, attractive forces were not considered properly when lumped into a total surface energy alone. In Situ Study of Island/Hole Formations. BP thin film self-assembly was examined by thermally annealing films during in situ optical microscopy to ensure previously mentioned substrate effects were thermodynamic surface energy phenomena and not kinetic limitations. Previous work on BP thin film kinetics indicates that islands/holes nucleate and grow to an equilibrium size, after which individual islands/holes can coalesce.41,57−59 Therefore, using a 24 h oven anneal for experiments had the following constraints: (1) that the time frame was sufficient for the surface structures to reach a relatively stable size, (2) that the time frame was not long enough to lead to excessive coalescence that would obscure island/hole size measurements, and (3) that the wetting behavior did not change over the time frame due to film expansion, high-temperature, or kinetic effects. Thermal annealing during in situ optical microscopy of films coated on bare silica and benzyl silane substrates was accomplished using a sealed Linkam heating stage (Figure 5a). Bare silica and benzyl silane substrates were used because films on these substrates produced large hole features for accurate tracking. Images and sizing from in situ measurements

of decoupled dispersive and polar components on the throughfilm driving forces, films were cast at incommensurate and holeforming thicknesses (≈88−90 nm for asymmetric wetting and ≈64 or ≈108 nm for symmetric wetting) and annealed under vacuum for 24 h at 175 °C. Film thicknesses that formed holes were selected because holes have larger equilibrium sizes in comparison to islands, thereby making them easier to measure accurately.33 The effect of individual dispersive and polar components was resolved from experimental results using two models: the Owens and Wendt model (eq 2) and the Good and Girifalco model (eq 5).43 ΔγA = γS + γA − 2 γS + γA

(5)

Both models include the total surface energy to account for repulsive interactions at the interface, and both models incorporate long-range, attractive interactions using a geometric mean approach. However, only the Owens and Wendt model (eq 2) separates the attractive interaction into dispersive and polar components. Measured hole sizes were matched with the corresponding interfacial energy differences calculated from each model (Figure 4). A linear trend line was fit to the data because literature suggests island/hole size should be directly proportional to surface energy.57 Based on the fits, the total surface energy model (Figure 4a) failed to describe the data on all substrates, especially for those substrates for which the polymer−substrate interfacial energy difference was relatively small. On the other hand, the decoupled surface energy model (Figure 4b) accurately described the data on all substrate surfaces. Differences in the model fits to the data highlight the importance of incorporating decoupled surface energy components when predicting substrate−polymer interactions. The major deviation between total and decoupled surface energy equation fits was noted on aceto silane, benzyl silane, 4577

DOI: 10.1021/acs.macromol.5b00833 Macromolecules 2015, 48, 4572−4580

Article

Macromolecules followed the expected nucleation and growth behavior and are shown in Figure 5b. Time-lapse video of hole growth can be found in the Supporting Information. Key results from the in situ studies supported findings from static experiments by ensuring the use of a 24 h anneal highlighted the thermodynamic effects in feature formation. First, within the 24 h time frame, holes grew to an equilibrium size that matched well with oven-annealed samples (Figure 4). Films cast on bare silica and annealed in the vacuum oven and the Linkam stage for 24 h formed holes with average sizes of 8.4 ± 0.9 and 8.1 ± 0.9 μm2, respectively. Similarly, films cast on benzyl silane and annealed in the vacuum oven and the Linkam stage for 24 h formed holes with average sizes of 1.5 ± 0.1 and 1.4 ± 0.8 μm2, respectively. Second, coalescence was not excessive after an equilibrium size was reached. Third, no temporal changes in wetting behavior (i.e., from asymmetric to symmetric) were captured during the testing of any individual substrates. Stable wetting behavior over time suggested that kinetic and high temperature effects were not responsible for the wetting behavior discrepancies noted in Figure 2. Instead, thermodynamic effects described by incorporation of a Hamaker constant accounted for the necessary repulsive interactions that explained wetting behavior (Figure 3). Furthermore, in situ studies supported formation and growth kinetics for the larger hole formations on higher surface energy substrates (Figure 5). This outcome indicated that increases in substrate interfacial energy differences directly impacted through-film self-assembly by enhancing polymer stretching/compression driving forces visible via larger equilibrium hole sizes (Figure 4). Combining the wetting behavior and through-film predictive equations provides the full picture of substrate surface selfassembly effects as seen in Figure 6. The total surface energy (eq 1) and Hamaker constant define the three possible wetting behavior regimes: PMMA preferential, PnBA preferential, and neutral (nonpreferential) substrates. The neutral, or nonpreferential region, represents substrate surface energies at which the total surface energy and Hamaker constant effects result in a different block being preferential. In some instances, neither set of interactions dominate and a true neutral surface is achieved; both polymer blocks segregate to the substrate and islands/holes do not form. In other instances (e.g., PMMA− PnBA on methacryl silane), one set of interactions dominates and drives a particular wetting behavior. As is shown in our work, the latter case is more common when the total surface energies of the two polymer blocks are significantly different. The gradient across the plot represents changes in the substrate self-assembly driving force predicted using eq 2 to describe through-film effects. Overall, these results emphasize the ability to control the selfassembly of BP thin film nanostructures by tuning the substrate surface energy and can be applied to most polymer systems to predict wetting behavior and substrate surface field strength. By examining the effect of the substrate surface energy on BP thin film self-assembly in detail, we have generated more universal guidelines to produce desired interfacial and through-film behavior.

Figure 6. Composite prediction of surface wetting and through-film behavior to pinpoint necessary substrate surface energy conditions to achieve prescribed substrate wetting behavior. The dotted lines separate the plot into three possible substrate wetting behaviors (PMMA preferential, PnBA preferential, and neutral) using the total surface energy and Hamaker constant analysis. The neutral region indicates a nonpreferential substrate region as the total surface energy, and Hamaker constant calculations result in a different block being preferential, although the strength of either of these parameters can dominate and force a preferential substrate (e.g., methacryl silane). The gradient across the plot represents the change in the substrate self-assembly driving force calculated using the decoupled surface energy.

influence nanostructure ordering and orientation. Our results indicate that the repulsive nature of the total surface energy dominated wetting behavior thermodynamics at the substrate− polymer interface, while attractive contributions from individual dispersive and polar components significantly impacted through-film interactions. Additionally, we found that our improved surface energy analysis accurately predicted wetting behavior and hole formation sizes on all substrates. Lastly, thermal annealing of films under in situ optical microscopy suggested that the substrate surface energy, not kinetic or high temperature effects, was responsible for exhibited substrate− polymer interfacial and through-film effects. This work represents an improved interpretation of the dominant substrate−polymer interactions affecting BP thin film selfassembly and an alternative method to screen chemicallymodified substrates for desired orientation and ordering control using predictive equations with the appropriately decoupled surface energy components.

■

ASSOCIATED CONTENT

S Supporting Information *

Time-lapse videos showing the formation and growth of holes on bare silica and benzyl silane substrates over a 24 h period. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.5b00833.

■

CONCLUSIONS We systematically studied the effects of total, dispersive, and polar substrate surface energy values on BP thin film interfacial and through-film self-assembly and developed a predictive understanding of the key substrate−polymer interactions that

■

AUTHOR INFORMATION

Corresponding Author

*E-mail [email protected] (T.H.E.). 4578

DOI: 10.1021/acs.macromol.5b00833 Macromolecules 2015, 48, 4572−4580

Article

Macromolecules Notes

(21) Xu, T.; Hawker, C. J.; Russell, T. P. Macromolecules 2005, 38, 2802−2805. (22) Ji, S.; Liu, C.-C.; Liao, W.; Fenske, A. L.; Craig, G. S. W.; Nealey, P. F. Macromolecules 2011, 44, 4291−4300. (23) Welander, A. M.; Craig, G. S. W.; Tada, Y.; Yoshida, H.; Nealey, P. F. Macromolecules 2013, 46, 3915−3921. (24) Nagpal, U.; Müller, M.; Nealey, P. F.; de Pablo, J. J. ACS Macro Lett. 2012, 1, 418−422. (25) Edwards, E. W.; Stoykovich, M. P.; Müller, M.; Solak, H. H.; de Pablo, J. J.; Nealey, P. F. J. Polym. Sci., Part B: Polym. Phys. 2005, 43, 3444−3459. (26) Mansky, P.; Russell, T. P.; Hawker, C. J.; Mays, J.; Cook, D. C.; Satija, S. K. Phys. Rev. Lett. 1997, 79, 237−240. (27) Knoll, A.; Magerle, R.; Krausch, G. J. Chem. Phys. 2004, 120, 1105−1116. (28) Fowkes, F. M. Ind. Eng. Chem. 1964, 56, 40−52. (29) Owens, D. K.; Wendt, R. C. J. Appl. Polym. Sci. 1969, 13, 1741− 1747. (30) Han, E.; Stuen, K. O.; La, Y.-H.; Nealey, P. F.; Gopalan, P. Macromolecules 2008, 41, 9090−9097. (31) Albert, J. N. L.; Kim, J. D.; Stafford, C. M.; Epps, T. H., III Rev. Sci. Instrum. 2011, 82, 065103. (32) Ashley, K.; Sehgal, A.; Amis, E. J.; Raghavan, D.; Karim, A. Mater. Res. Soc. Symp. Proc. 2001, 700, S4.7.1−S4.7.6. (33) Smith, A. P.; Sehgal, A.; Douglas, J. F.; Karim, A.; Amis, E. J. Macromol. Rapid Commun. 2003, 24, 131−135. (34) Bang, J.; Jeong, U.; Ryu, D. Y.; Russell, T. P.; Hawker, C. J. Adv. Mater. 2009, 21, 4769−4792. (35) Peters, R. D.; Yang, X. M.; Kim, T. K.; Nealey, P. F. Langmuir 2000, 16, 9620−9626. (36) Roberson, S. V.; Fahey, A. J.; Sehgal, A.; Karim, A. Appl. Surf. Sci. 2002, 200, 150−164. (37) Janssen, D.; De Palma, R.; Verlaak, S.; Heremans, P.; Dehaen, W. Thin Solid Films 2006, 515, 1433−1438. (38) Peters, R. D.; Yang, X. M.; Nealey, P. F. Macromolecules 2002, 35, 1822−1834. (39) Kim, S.; Bates, C. M.; Thio, A.; Cushen, J. D.; Ellison, C. J.; Willson, C. G.; Bates, F. S. ACS Nano 2013, 7, 9905−9919. (40) Smith, A. P.; Douglas, J. F.; Meredith, J. C.; Amis, E. J.; Karim, A. Phys. Rev. Lett. 2001, 87, 015503. (41) Coulon, G.; Collin, B.; Ausserre, D.; Chatenay, D.; Russell, T. P. J. Phys. (Paris) 1990, 51, 2801−2811. (42) Wu, S. J. Polym. Sci., Polym. Symp. 1971, 34, 19−30. (43) Girifalco, L. A.; Good, R. J. J. Phys. Chem. 1957, 61, 904−909. (44) Brandrup, J.; Immergut, E. H.; Grulke, E. A. Polymer Handbook, 4th ed.; Wiley: New York, 2004. (45) Park, I.; Park, S.; Park, H.-W.; Chang, T.; Yang, H.; Ryu, C. Y. Macromolecules 2006, 39, 315−318. (46) Loo, Y.-L.; Register, R. A.; Adamson, D. H.; Ryan, A. J. Macromolecules 2005, 38, 4947−4949. (47) Stafford, C. M.; Roskov, K. E.; Epps, T. H., III; Fasolka, M. J. Rev. Sci. Instrum. 2006, 77, 023908. (48) Park, D.-H. Nanotechnology 2007, 18, 355304. (49) Hiemenz, P. C.; Rajagopalan, R. Principles of Colloid and Surface Chemistry, 3rd ed.; Marcel Dekker: New York, 1997. (50) Cai, X.-J.; Genzer, J.; Spontak, R. J. Langmuir 2014, 30, 11689− 11695. (51) Oron, A.; Davis, S. H.; Bankoff, S. G. Rev. Mod. Phys. 1997, 69, 931−980. (52) Brochardwyart, F.; Dimeglio, J. M.; Quere, D.; Degennes, P. G. Langmuir 1991, 7, 335−338. (53) Bucknall, D. G. Prog. Mater. Sci. 2004, 49, 713−786. (54) Chaudhury, M. K. Mater. Sci. Eng. 1996, 16, 97−159. (55) Sferrazza, M.; Heppenstall-Butler, M.; Cubitt, R.; Bucknall, D.; Webster, J.; Jones, R. A. L. Phys. Rev. Lett. 1998, 81, 5173−5176. (56) Reiter, G. Langmuir 1993, 9, 1344−1351. (57) Grim, P. C. M.; Nyrkova, I. A.; Semenov, A. N.; ten Brinke, G.; Hadziioannou, G. Macromolecules 1995, 28, 7501−7513.

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This paper was prepared under cooperative agreement 70NANB12H239 from NIST, U.S. Department of Commerce. The statements, findings, conclusions, and recommendations are those of the authors and do not necessarily reflect the views of NIST or the U.S. Department of Commerce. This work also was partially supported by the National Science Foundation (NSF DMR-1207041). We thank Prof. T. Beebe, Jr., Department of Chemistry and Biochemistry, University of Delaware (UD), for use of the contact angle goniometer and the UD W. M. Keck Microscopy Facility for use of the AFM. Additionally, we thank Dr. S. M. Mastroianni for assistance with SAXS data collection.

■

ABBREVIATIONS BP, block polymer; PMMA, poly(methyl methacrylate); PnBA, poly(n-butyl acrylate); UVO, ultraviolet ozone; benzyl silane, benzyldimethylchlorosilane; aceto silane, 2-acetoxyethyldimethylchlorosilane; methacryl silane, 3-methacryloxypropyldimethylchlorosilane; n-butyl silane, n-butyldimethylchlorosilane; diio, diiodomethane; AFM, atomic force microscopy; HPL, hexagonally perforated lamellae; SAXS, small-angle X-ray scattering.

■

REFERENCES

(1) Bates, F. S.; Fredrickson, G. H. Phys. Today 1999, 52, 32−38. (2) Bates, F. S. Science 1991, 251, 898−905. (3) Fasolka, M. J.; Mayes, A. M. Annu. Rev. Mater. Res. 2001, 31, 323−355. (4) Segalman, R. A. Mater. Sci. Eng. 2005, 48, 191−226. (5) Albert, J. N. L.; Epps, T. H., III Mater. Today 2010, 13, 24−33. (6) Bates, C. M.; Maher, M. J.; Janes, D. W.; Ellison, C. J.; Willson, C. G. Macromolecules 2013, 47, 2−12. (7) Hu, H.; Gopinadhan, M.; Osuji, C. O. Soft Matter 2014, 10, 3867−3889. (8) Knoll, A.; Horvat, A.; Lyakhova, K. S.; Krausch, G.; Sevink, G. J. A.; Zvelindovsky, A. V.; Magerle, R. Phys. Rev. Lett. 2002, 89, 035501. (9) Jung, J.; Park, H.-W.; Lee, S.; Lee, H.; Chang, T.; Matsunaga, K.; Jinnai, H. ACS Nano 2010, 4, 3109−3116. (10) Forrey, C.; Yager, K. G.; Broadaway, S. P. ACS Nano 2011, 5, 2895−2907. (11) Matsen, M. W. J. Chem. Phys. 1997, 106, 7781−7791. (12) Epps, T. H., III; DeLongchamp, D. M.; Fasolka, M. J.; Fischer, D. A.; Jablonski, E. L. Langmuir 2007, 23, 3355−3362. (13) Bates, F. S.; Hillmyer, M. A.; Lodge, T. P.; Bates, C. M.; Delaney, K. T.; Fredrickson, G. H. Science 2012, 336, 434−440. (14) Luo, M.; Epps, T. H., III Macromolecules 2013, 46, 7567−7579. (15) Peters, R. D.; Yang, X. M.; Kim, T. K.; Sohn, B. H.; Nealey, P. F. Langmuir 2000, 16, 4625−4631. (16) Albert, J. N. L.; Baney, M. J.; Stafford, C. M.; Kelly, J. Y.; Epps, T. H., III ACS Nano 2009, 3, 3977−3986. (17) Han, E.; Stuen, K. O.; Leolukman, M.; Liu, C.-C.; Nealey, P. F.; Gopalan, P. Macromolecules 2009, 42, 4896−4901. (18) Khanna, V.; Cochran, E. W.; Hexemer, A.; Stein, G. E.; Fredrickson, G. H.; Kramer, E. J.; Li, X.; Wang, J.; Hahn, S. F. Macromolecules 2006, 39, 9346−9356. (19) Luo, M.; Seppala, J. E.; Albert, J. N. L.; Lewis, R. L.; Mahadevapuram, N.; Stein, G. E.; Epps, T. H., III Macromolecules 2013, 46, 1803−1811. (20) Khaira, G. S.; Qin, J.; Garner, G. P.; Xiong, S.; Wan, L.; Ruiz, R.; Jaeger, H. M.; Nealey, P. F.; de Pablo, J. J. ACS Macro Lett. 2014, 3, 747−752. 4579

DOI: 10.1021/acs.macromol.5b00833 Macromolecules 2015, 48, 4572−4580

Article

Macromolecules (58) Heier, J.; Kramer, E. J.; Groenewold, J.; Fredrickson, G. H. Macromolecules 2000, 33, 6060−6067. (59) Green, P. F.; Limary, R. Adv. Colloid Interface Sci. 2001, 94, 53− 81.

4580

DOI: 10.1021/acs.macromol.5b00833 Macromolecules 2015, 48, 4572−4580