Stress–Structure Correlation in PS–PMMA Mixed Polymer Brushes

Article pubs.acs.org/Macromolecules

Stress−Structure Correlation in PS−PMMA Mixed Polymer Brushes Jannis W. Ochsmann,†,‡ Sebastian Lenz,† Philipp Lellig,†,‡ Sebastian G.J. Emmerling,† Ali A. Golriz,†,‡ Peter Reichert,† Jichun You,† Jan Perlich,§ Stephan V. Roth,§ Rüdiger Berger,† and Jochen S. Gutmann*,†,‡,⊥ †

Max Planck Institute for Polymer Research, Experimental physics of interfaces, Ackermannweg 10, D-55128 Mainz, Germany Department of Chemistry and Center for Nanointegration Duisburg-Essen (CeNIDE), University Duisburg-Essen, Universitätsstr. 5, D-45117 Essen, Germany § HASYLAB at DESY, Notkestr. 85, D-22603 Hamburg, Germany ⊥ Deutsches Textilforschungszentrum Nord-West e.V., Adlerstr. 1, D-47798 Krefeld, Germany ‡

S Supporting Information *

ABSTRACT: The ability to alter surface properties such as morphology and surface energy upon external stimuli makes switchable polymer surfaces a promising field of research. Mixed polymer brushes consisting of two different homopolymers covalently attached to a surface are one system in which surface properties can be switched. In this work the correlation between the change in structure and the resulting surface stress in thin poly(methyl methacrylate)−polystyrene mixed polymer brush film upon exposure to selective solvents is investigated. By measuring the forces acting inside the film, we are able to achieve a deeper understanding of the observed structural changes. To obtain a thorough understanding of the film’s morphology, the structure is analyzed by scanning probe microscopy, X-ray reflectivity, and grazing incidence small-angle X-ray scattering (GISAXS). Upon exposure to acetic acid, a selective solvent for PMMA, the film showed a dimple-like structure. This is linked to collapsed domains of polystyrene covered by PMMA chains. Bending experiments resulted in tensile stress, pointing to attractive forces acting inside the polymer film. After exposure to dichloromethane, a good solvent for both polymers, bending experiments revealed a decreased but still high tensile stress, indicating that the microdomains are still present. The results of the experiments enable us to further explain the domain memory effect typically found in these kinds of mixed polymer brush systems.

■

INTRODUCTION Switchable polymer surfaces are a promising field of research.1 The ability to switch surface characteristic such as morphology and surface energy upon external stimuli make them an interesting research topic. In addition, the possibility to move nano-objects on top of a surface by reversible switching of the polymer surface is a subject of current research.2 Mixed polymer brushes consisting of two different homopolymers covantly attached to a surface are one candidate for these types of polymer surfaces.3−9 Extensive theoretical research has been done in the past couple of years.10 Studying the film morphology in relation to the polymer incompatibilities,11 solvent selectivity,12 composition,11 and chain length asymmetry.13 Fluctuations in local grafting-point positions and its correlations have been investigated. With the help of “singlechain-in-mean-field” simulations it was shown that small © 2012 American Chemical Society

fluctuations in grafting point positions influence domain locations and prevent long-range order.14,15 Although these simulations resemble the experimental findings,16 the applied mean-field approach neglects chain conformation. Microphase separation of chemical bond polymer chain requires a large amount of stretching of indivual chains. These conformational restrains and the resulting stress inside the polymer film are not accounted for in the mean-field approach. The stress inside the polymer film is transduced into the substrate. To measure such forces, nanomechanical cantilever sensors (NCS) have become a valuable tool in recent years.17−21,22 By this means surface stress changes due to Received: November 18, 2011 Revised: February 11, 2012 Published: March 20, 2012 3129

dx.doi.org/10.1021/ma2025286 | Macromolecules 2012, 45, 3129−3136

Macromolecules

Article

Figure 1. Schematic diagram of the steps necessary for the preparation of PS−PMMA mixed polymer brushes on the amine-functionalized silica surface. First a thin film of PFP−PMMA is applied to an amine-functionalized surface. By heating the film above the glass transition temperature of PMMA, the active ester function at the end of the polymer react with the amine groups. The PMMA is therefore covalently bound to the surface. After extraction of excess polymer, a thin film of PFP−PS is applied to the same surface. By heating the system above the glass transition temperature of the PS, the polystyrene reacts with the vacant amine groups.

polymer brush swelling,23−26 polyelectrolyte adsorption,27 and stress upon thermal treatment18−20,28 could be measured at extremely high resolution. Mixed polymer brushes can be obtained by the “graftingfrom” as well as “grafting-to” technique. The “grafting-from” approach usually involves the preparation of a layer containing two different kinds of initiators (e.g., ATRP-, NMRP-initiator).7 By this means mixed polymer brushes with high grafting densities may be obtained. In the “grafting-to” approach the two homopolymers are attached to the same surface subsequently. Stamm et al. utilized a reaction in melt to graft acid-functionalized polymers to an epoxide-functionalized surface. Variation of the temperature and duration of the “grafting-to” reaction allowed Stamm et al. to alter the composition of the mixed polymer brushes.3,29−31 Both “grafting-from” and “grafting-to” mixed polymer brushes could be switched between different morphologies by exposing them either to selective solvents for one of the polymers, good solvents for both polymers, or thermal annealing.4,8 In recent years a domain memory effect has been experimentally observed in mixed polymer brush films. Santer et al. introduced a quantified measurement for domains forming at the same position, size, and shape upon cyclical change.16,32,33 The effects of chemical structure and fluctuations of the grafting points have been investigated. The authors attribute the domain memory effect to small fluctuation on grafting points, nucleating the domains in the mixed brush. While this explains the static structure formation, the reason for the domain stability over time remains unsolved. In this paper we use an amine-functionalized silica surface, active ester-functionalized polystyrene, and poly(methyl methacrylate) to obtain PS−PMMA mixed polymer brushes. We investigate the switching behavior of the mixed polymer brushes after they have been immersed in a good solvent for both polymers (dichloromethane) and a selective solvent for only the PMMA (acidic acid). The change in structure between the two equilibrium states is investigated by X-ray reflectivity (XRR), scanning probe microscopy (SPM), contact angle measurement, and μ-focused grazing incidence small-angle Xray scattering (μ-GISAXS). Using these methods enables us to obtain a comprehensive picture about the change in polymer confirmation before and after switching. Via bending experiments on NCS coated with the PS−PMMA mixed brushes, the change in surface stress was measured. With the help of a phase shifting interferometer36 (PSI) the 3d topography of the NCS

is determined. This allows us to obtain the bending profile of each individual cantilever, enabling us to calculate the absolute values of surface stress acting inside the film. Measuring these forces gives a deeper understanding of the network inside the polymer brush and the surface memory effect.

■

MATERIALS AND METHODS

Polymer Synthesis. Active Ester-Functionalized Chain Transfer Agent. Pentafluorophenyl-(4-phenylthiocarbonylthio-4-cyanovalerate) was synthesized according to a literature procedure.34 The required di(thiobenzoyl) disulfide was synthesizes according to a different route because no red solid could be obtained after the reaction of dithiobenzoic acid with iodine. The di(thiobenzoyl) disulfide was therefore obtained by dissolving dithiobenzoic acid (5 g, 0.032 mol) and sodium hydroxid (1.28 g, 0,032 mol) in 100 mL of deionized water. Potassium ferricyanide(III) (10.67 g, 0.032 mol) dissolved in 50 mL of deionized water was added dropwise and stirred for several minutes. The violet precipitate was filtered and washed. The product was recrystallized from ethanol.35 Pentafluorophenol−Poly(methyl methacrylate) (PFP−PMMA). The polymer was synthesized as described here.34 The reaction yielded 2.8 g (69%) of pink polymer with a molecular weight Mn(GPC) = 16 196 g/mol with a PDI (GPC) = 1.08. Pentafluorophenol−Polystyrene (PFP−PS). 5.63 mL (48.8 mmol) of freshly distilled styrene, 11.1 mg of AIBN, 3 mL of toluene, and 120.8 mg of pentafluorophenyl (4-phenylthiocarbonylthio-4-cyanovalerate) were added to a Schlenk tube under an argon atmosphere. The mixture was degassed by three freeze−pump−thaw cycles. The polymerization was carried out at 110 °C for 4 days. The polymer was precipitated three times in methanol from THF. The reaction yielded 2.2 g (40%) of pink polymer with a molecular weight Mn(GPC) = 13 477 g/mol with a PDI (GPC) = 1.15. Altering the experimental conditions did not result in higher molecular weight polymer with reduced PDIs. Nanomechanical Cantilever Arrays (NCS) and Silicon Wafers. NCS arrays (Octosensis) were purchased from Micromotive Mikrotechnik, Germany. Each array consisted of eight individual cantilevers with an area of 750 × 90 μm2, a thickness of 2 μm, and a pitch of 250 μm. Silicon wafers were obtained from Si-Mat Silicon Materials, Germany. The polished silicium wafers with an (100) orientation were cut into 2.5 × 2.5 cm2 large substrates for sample preparation. Wafer Preparation. The schematic diagram of the preparation route can be found in Figure 1. Silicon wafer substrates were put in dichloromethane and placed in an ultrasonic bath for 15 min and dried in a stream of nitrogen. The wafers were immersed in a solution of ammonia, hydrogen peroxide, and distilled water (1:1:5) and heated to 3130

dx.doi.org/10.1021/ma2025286 | Macromolecules 2012, 45, 3129−3136

Macromolecules

Article

70 °C for 20 min. The wafer were then rinsed with distilled water and dried in a stream of nitrogen. The cleaned wafers were put into a Schlenk tube filled with argon gas. Two drops of (3-aminopropyl)dimethylethoxysilane (APDES) were added, and the Schlenk tube was evacuated for ∼1 s. The Schlenk tube was placed into a vacuum oven at 120 °C for 14 h. Subsequently, the wafers were extracted with dichloromethane in a Soxhlet extractor for at least 5 h. A 2 wt % toluene solution of PFP−PMMA was spin-coated on to the before prepared wafer for 30 s at 1000 rpm. Afterward, the film was heated to 120 °C for 90 min. The wafers were cooled down to room temperature and extracted with dichloromethane for at least 5 h. A 2 wt % toluene solution of PFP−PS was spin-coated on to the PMMA film for 30 s at 1000 rpm. The film was afterward heated to 120 °C for 12 h. The wafers were cooled down to room temperature and extracted with dichloromethane for at least 5 h. Cantilever Preparation. The cantilever arrays were immersed into a solution of ammonia, hydrogen peroxide, and distilled water (1:1:5) and heated to 70 °C for 20 min. Afterward, the NCS were rinsed with distilled water and carefully dried in a stream of nitrogen. To ensure one-sided coatings of the cantilever, the backside of the NCS was passivated by evaporation of a 30 nm layer of gold via thermal evaporation (0.1 nm/s, p ≈ 2 × 10−5 mbar) (BALTEC MED 020, BALTEC, Balzers, Lichtenstein). The NCS were put into an argon gas filled Schlenk tube. Two drops of (3-aminopropyl)dimethylethoxysilane were added, and the Schlenk tube was then evacuated for ∼1 s. The tube was placed into a vacuum oven and heated to 120 °C for 14 h. The NCS were cooled to room temperature and extracted in dichloromethane for at least 5 h. A 2 wt % toluene solution of PFP−PMMA were selectively applied to a single cantilever. The solution was deposited using a piezocontrolled nanoliter pipet using a Nano-Plotter NP 2.0 device (GeSim mbH, Germany). The cantilevers were printed by the array dispense procedure (ADP), dispensing a column of 15 drops 150 μm apart from each other. Using the program SpotFrontEnd, the cantilevers were aligned properly and a high reproducibility was obtained. Usually four cantilevers were functionalized using this printing technique. Next the NCS was heated to 120 °C for 90 min, cooled down to room temperature, and extracted in dichloromethane for at least 5 h. Afterward, a 2 wt % toluene solution of PFP−PS was printed onto two cantilevers previously coated with PFP−PMMA and two cantilevers with no previously deposited polymer. The cantilever array was subsequently heated to 120 °C for 12 h, cooled to room temperature, and extracted in dichloromethane for at least 5 h. By this means we obtained a cantilever array containing two cantilevers coated with mixed PS−PMMA brushes, two cantilevers coated with polystyrene and poly(methyl methacrylate) for comparison, and two cantilevers with no polymer brushes as reference (Figure 2). At the end immersion of the cantilever in a KI/I2 solution and rinsing with Millipore water removed the protective gold layer. Switching. For the selective solvent treatment, the wafers/ cantilever were immersed into concentrated acetic acid and heated to 45 °C for 30 min. The sample was then quickly rinsed with distilled water several times and dried in a stream of nitrogen gas. The procedure was repeated three times. For the nonselective treatment the wafers/cantilever arrays were immersed in dichloromethane for at least 90 min at 45 °C and subsequently dried in a nitrogen stream. As reported in the literature,13 switching of the morphology takes several minutes; also, separate tests using solvent vapors showed a fast brush response (see Supporting Information). Since the solvent concentration is even higher in liquids, the chosen experimental conditions should guarantee a complete switching of the brushes. Evaporation of the solvent on the other hand only takes several seconds due to the nanometer film thickness. The brush morphology is therefore frozen during the drying process. After sample preparation the sample are stored under ambient conditions. For both polymers this is well below its glass transition, preserving the film morphology over time. Methods. X-ray Reflectivity. X-ray reflectivity (XRR) measurements on wafers were conducted on a XRD 3003 TT, Seifert Ltd. GB diffraction system. Monochromatic and

Figure 2. SEM Image of a nanomechanical cantilever sensor array (top) and schematic representation of the used cantilever design (bottom). Two cantilevers are coated with PMMA brushes, two with PMMA−PS mixed brushes, two with PS brushes, and two cantilevers with no polymers as references. collimated X-rays were obtained from a Cu anode with a wavelength of λ = 0.154 nm. The obtained curves were fitted using PARRATT32. The starting values for scattering length densities used in the fit as well as the densities obtained by the fitted scattering length densities were calculated using the scattering length density calculator provided by NIST.37 Scanning Probe Microscopy. Scanning probe microscopy images were obtained in tapping mode using an Dimension 3100 scanning probe microscope, Veeco Instruments GmbH. OMCL AC 160 TS (Olympus, Japan) cantilevers were used with a resonance frequency of ∼300 kHz and a spring constant of 42 N/m as specified by the manufacturer. Contact Angle. Contact angle measurements were conducted on a Krüss, DSA10-MK2. The static contact angle was measured in air under ambient conditions using Milli-Q water. μ-GISAXS. The μ-GISAXS measurements were conducted at the beamline BW4, HASYLAB at DESY, Hamburg, Germany. The beam had a wavelength of λ = 0.138 nm. The used beam had a size of 32 × 17 μm2 (horizontal × vertical). The schematic representation of the used μ-GISAXS geometry can be found in Figure 3. To obtain transverse qII scans at the Yoneda maximum, the 2D images were cut vertically at the Yoneda maximum (αf = αc). The qII scans were fitted with the unified fit model.38 The model allows the analysis of qII scans for lateral structural information on thin films containing weakly correlated and polydisperse structural components.28,39,40 It uses structural levels, which allows the modeling of scattering from small (d ≈ 10 nm) up to larger structures (d ≈ 350 nm). The model allows size and structural determination with an error of less than 20%. Further details can be found here.41 Initially, the incident angle (αi) was set to αi = 0.4°. This avoids the overlap of specular and off-specular scattering in the Yoneda peak regime, which is necessary for the previously described analysis. NCS Bending Experiments. A home-built interferometric setup was used to determine the bending of each nanomechanical cantilever. The setup consists of a laser with a wavelength of λ = 785 nm. The recorded image had an area of 1.5 × 1.5 mm2. All eight cantilevers and part of the supporting chip were simultaneously measured. Custommade software allowed the simultaneous recording of different interferometric images and its analysis. To obtain curvature data from each individual cantilever, five images at different mirror positions were taken. From these images a 3D profile was calculated. For each cantilever four topography profiles were taken along its xaxis. These profiles were averaged to obtain a mean length/height profile of each cantilever. The averaged profile was then fitted with a parabolic function.42 z(x) = a0 + a1x + 3131

κ 2 x 2

(1)

dx.doi.org/10.1021/ma2025286 | Macromolecules 2012, 45, 3129−3136

Macromolecules

Article

Figure 3. μ-GISAXS geometry. Under the assumption that the cantilever is bend uniformly, we can calculate the absolute curvature κ

κ=

d2z(x) (dx 2)

(2)

For a typical measurement 100 averaged profiles were recorded over time (0.33 measurements/s) and analyzed. The surface stress σ was calculated with Stoney’s formula.43 σ=

ESitNCS2 κ 6(1 − νSi)

(3)

For the calculations ESi = 165 GPa, tNCS = 2 μm, and a Poisson ratio of νSi = 0.28 were used.

■

RESULTS AND DISCUSSION Film Preparation. To obtain mixed polymer brushes on top of the amine-functionalized silicium surface, a “grafting-to” approach was selected. The amine-functionalized surface was obtained by the self-assembling gas phase reaction of (3aminopropyl)dimethylethoxysilane. A thin film of PMMA carrying an active ester end group was deposited on top of the functionalized silicium surface. By heating the sample above the glass temperature (Tg) of the polymer, the active ester end groups increase their mobility and react with the amine end groups on the surface. After this solidstate reaction the polymer is covalently bound to the surface. By varying the temperature and time of heating partial saturation of the amine groups can be achieved. After extraction of the excess polymer, a thin PS film is deposited on the sample. Heating the sample above the Tg allows the active ester groups at the end of the polystyrene to react with the remaining free amine groups. In the end the excess polystyrene is again extracted. The preparation of the mixed polymer brush film was followed by X-ray reflectivity (XRR) to verify the proposed mechanism. The obtained reflectivity curves are shown in Figure 4. The treatment of the freshly cleaned siliceous wafer with (3-aminopropyl)dimethylethoxysilane leads to the formation of a 0.8 nm thick film. This corresponds nicely to the f o r m a t i o n o f a m o n o l ay er o f (3 - am in o p r o p y l) dimethylethoxysilane covalently bond to the silica surface. The analysis of the first prepared PMMA film yielded a 3.6 nm thick film with a relative low roughness of σrms = 0.3 nm. The following spin-coating of the polystyrene solution led to an increase in thickness. The resulting film yielded a thickness of 5.1 nm and an increase in roughness to σrms = 0.5 nm. These consecutive increase in film thickness supports the proposed

Figure 4. X-ray reflectivity curves of the monolayer of APDES (a), APDES/PMMA (b), and APDES/PMMA−PS after treatment with dichloromethane (c) and acetic acid (d). Hollow symbols represent the experimental data. Straight lines represent the Parratt fit. The curves are shifted along the vertical axis for improved visualization.

mechanism of step-by-step “grafting-to” of polymer chains to the same functionalized surface. The SPM images (Figure 5 bottom) at the end of the assembling process also supports the formation of a continuous and smooth polymer film.

Figure 5. SPM topography images of the PMMA−PS mixed polymer brush film after acetic acid and dichloromethane treatment. 3132

dx.doi.org/10.1021/ma2025286 | Macromolecules 2012, 45, 3129−3136

Macromolecules

Article

From the measured film thickness d and the molecular weight Mn of the unbound homopolymer measured by GPC the grafting density was calculated using the formula

Γ=

dρ NA Mn

The density ρ of the PMMA and the mixed PMMA−PS brushes was determined by the fit of the X-ray reflectivity. For the PMMA brush the value was 1.19 g/cm−3 equal to the literature44 value of an unbound chain. For the PMMA−PS mixed brush case the value was 1.25 g/cm−3, slightly above the literature values of PMMA and PS.45 The resulting grafting densities are ΓPMMA = 0.16 chains/nm2 and ΓPMMA−PS = 0.26 chains/nm2. From the measured film thicknesses the brush composition can be estimated. If we neglect small differences in density, the composition of the mixed polymer brush can be calculated by dividing the film thickness of the PMMA film by the value of the PMMA−PS film. To account for the APDES layer, 0.8 nm is subtracted from both film thicknesses. The PMMA−PS brush films has a composition of ∼65% PMMA and ∼35% PS. Switching. First the change in structure caused by the switching upon different solvent exposure of the PS−PMMA mixed brushes was examined. Therefore, mixed polymer samples were prepared on silicum wafers and investigated before and after exposition to acetic acid, a selective solvent for PS, and dichloromethane, a good solvent for PS and PMMA, respectively. Figure 5 shows an SPM height image of the same dichloromethane and acidic acid treated film. The immersion of the polymer film in acetic acid leads to the formation of nanodomains, showing a dimple-like structure. This change in structure is accompanied by a large increase in roughness from σrms = 0.23 nm to σrms = 0.65 nm (1 × 1 μm2). Multiple line cuts were performed to measure the distance between nanodomains. The analysis resulted in an average center-tocenter distance of 27.5 ± 8.8 nm of the nanodomains. The change in structure can also be seen by XRR (Figure 4). The fitted reflectivity curve also confirmed the increase in roughness by a factor of ∼3, which is also clearly visible by vanishing of all but one of the Kiessig fringes. The change in structure also leads to an increase in overall thickness from 5.1 to 6.7 nm. To investigate the composition at the surface of the polymer film, contact angle measurements were performed on acidic acid as well as dichloromethane treated films. The dichloromethane treated film showed a contact angle of 76°. Treating the film with acetic acid led to a decrease in contact angle to 67°, when compared to the dichloromethane film. The decrease in water contact angle can be attributed to enrichment of the more hydrophilic PMMA at the film−air interface. Grazing incidence small-angle X-ray scattering was performed to obtain a more averaged and quantified information on the films morphology. The incident angle was set to α = 0.4°, which is above the critical angle of both the polymer film and the silicon substrate. Therefore, structural information from the surface as well as beneath was obtained. The horizontal GISAXS data (Figure 6) obtained from the acetic acid treated film showed a large and broad interference peak. Fitting of the data resulted in domains with an Rg = 6.5 ± 1.3 nm and a center-to-center distance of 19.4 ± 3.9 nm. These results can be related to the domains observed from the SPM images.

Figure 6. μ-GISAXS horizontal detector cuts of the PS−PMMA mixed polymer brush films after dichloromethane (bottom) and acidic acid (top) treatment. Hollow symbols represent the experimental data. Straight lines represent the unified fit. The upper curve is shifted along the vertical axis for improved visualization. Inset shows the 2D scattering image of the acetic acid treated film.

The morphology obtained after exposure to the selective solvent acetic acid is in good agreement with theoretical studies done by Müller et al.15 Using self-chain in mean-field (SCMF), they predicted structures for different solvent qualities and brush compositions. In their calculations a dimple conformation is expected for mixed brushes exposed to a solvent, which is a good solvent for the polymer majority component in the film. As calculated by XRR the majority component of the prepared layer is PMMA (∼65%), which has a good solubility in acetic acid. Lateral microphase separation results in the formation of different domains, mainly depending on the brush composition. Highly asymtretric brushes form dimple-like structures. The effect is amplified by the acetic acid, which causes the polystyrene to collapse and migrate to the substrate. After the collapse the dimples are covered by PMMA. From the center-to-center distance and the extracted radius of gyration and average number of chain per domain can be estimated. As visible from the wide peak in the GISAXS scan, the domains have a broad size distribution. A single domain consists of roughly 40−70 single polymer chains. The broad size distribution is probably caused by the preparation method. Each step of the “grafting-to” process is a complete statistical process. A statistical distribution of the different polymer and a large variance in dimple size is therefore anticipated. The large error in the center−center distance extracted from the line cuts of the SPM images is another evidence of this. Also, the composition of the mixed brushes might play a role. Theoretical calculation predicts a transition between dimples to ripple morphology for an increase in the minor component. Mixed morphology would lead to an increase in polydispersity and uncertainty in the radius of gyration. The fitted power law exponent of P = 2.4 is also another evidence of a mixed morphology, proposing an particle arrangement in a arbitrary three-dimensional mass fractal.41 The GISAXS image also reveals a lag of long-range order in case of the acetic acid treated film. As reported in the literature, this is caused by 3133

dx.doi.org/10.1021/ma2025286 | Macromolecules 2012, 45, 3129−3136

Macromolecules

Article

action. These attractive forces are commonly attributed to interchain entanglements.46,47 For interchain entanglements to be present, the brushes have to be close enough with respect to the radius of gyration of an unperturbed polymer chain. The PMMA brushes have a grafting density of d = 0.31 chains/nm2 and the PS brushes of d = 0.45 chains/nm2. If we calculate the interchain distance from these values and compare them to the radii of gyration of the unperturbed chain, one can see that an overlap is present. For the mixed polymer brushes much higher tensile stress was measured. Structural information suggests that for the acetic acid treat film microphase separation is the reason for this. As discussed earlier, the polystyrene chain collapses and forms dimple close to the substrate surface. Multiple chains have to stretch and migrate to a common location. Since they are chemical bound to the surface, this happens under loss of entropy. Additional stress is induced through the PMMA covering the PS coil. To cover the coil each chain has to stretch out of their natural configuration. This induces a tensile stress, i.e., the cantilever bends toward the polymer layer. After treatment with the good solvent dichloromethane the tensile stress in the cantilever coated with the mixed brushes decreases. However, compared to the homopolymer brush films, the tensile stress is higher. The contact angle measurement suggests an increase in PS at the surface of the film. The measured tensile stress indicates that microphase separation is still present inside the mixed brush layer. This means that the PS and PMMA domains are still present in its previous composition. Treatment with dichloromethane causes the polystyrene to swell, increasing its surface coverage. At the same time, stress caused by PMMA bending over the PS is released. The additional surface stress in the acetic acid treated film is mainly a result of an increased stretching of the chains caused by the collapse of the PS brushes and the forced reshaping of the PMMA brush domains. In summary, the change in stress and structure can be described as follows (Figure 8). Microphase separation inside the mixed polymer brush film induces tensile stress in both acetic acid and dichloromethane treated films. The therefore required stretching of the chain induces a force causing the cantilever to bend upward. Treatment with acetic acid collapses the polystyrene domains; the therefore increased stretching of the PS and PMMA chains covering the polystyrene domains causes the increase in tensile stress. With our results, the domain memory effect as reported by Santer et al.16,32,33 can be explained. The fluctuations in local composition and grafting density results in domain formation at specific places. Since domain formation involves a high amount of stretching PS and PMMA domains form most likely at relatively high local grafting densities of their respective species, minimizing the amount of stretching. As measured by the bending experiments, treatment with the good solvent dichloromethane does change the surface structure from a dimple to a flat but does not dissolve the domains itself. The flattening of the surface is due to the swelling of the polystyrene domain caused by the exposure to the good solvent. Treatment with acetic acid causes the polystyrene domain to collapse, leading to the reappearence of the dimple structure. Since no PS and PMMA domains are dissolved during the switching, the domains are formed at the same place and in the same size as before. The change in surface structure from flat to dimple structure is mainly caused by the collapsed and swelling of the polystyrene domains and the therefore necessary rearrange-

fluctuations in the grafting point positions caused by the step by step preparation method.14 Theoretical SCMF calculations suggest that microphase separation is still present after the film has been exposed to a good solvent for both polymers. The increase in the contact angle to 76° suggests that both PMMA and PS are at the polymer−air interface. The previously collapsed PS chain swell due to the osmotic pressure, causing the polystyrene domains to rise to the surface. The SPM images show no contrast in phase mode. This is probably due to the small difference in elastic modulus between PS and PMMA. The horizontal GISAXS analysis also shows no structural peak at an incidence angle of α = 0.4°. Above the critical angle of both PS and PMMA, the beam penetrates the whole film, giving raise to structural information on the whole film. Since each polymer chain is chemically bound to the substrate, microphase separation can only occur at the surface, with a mixed PS/ PMMA layer underneath it. Because of the low contrast in electron density between PMMA and PS, no peak is visible in the GISAXS image. In the SPM images no phase differences is visible due to the similarities in elastic modulus. NCS Analysis. For surface stress analysis of the PS−PMMA brushes NCS bending experiments were conducted. One NCS array possessing eight independent cantilevers was single-sided coated with three different polymer films. Two cantilevers were coated with PS−PMMA mixed brush films, two cantilevers with a homo-PMMA brush film, and two cantilevers with a homo-PS brush film. The other two cantilevers were left uncoated and used as references (Figure 2). The reference cantilevers allow us to deduct and account for all nonspecific effects on the cantilever. The cantilevers only coated with PMMA and PS, on the other hand, allow us to take any polymer specific forces into account. NCS bending experiment were conducted before and after with dichloromethane and acetic acid, respectively. Figure 7

Figure 7. Absolute values of surface stress corrected by reference cantilever.

displays the absolute surface stress values for PMMA−PS mixed polymer brush coated cantilever and the cantilever only coated with one of the homopolymers. Both cantilevers coated with homopolymer brushes as well as the one coated with mixed polymer brushes show a tensile stress upon the exposure with the selective solvent acetic acid. Bending experiments give information about forces acting laterally inside the polymer brush. For the homopolymer case, the tensile stress can therefore be attributed to attractive polymer−polymer inter3134

dx.doi.org/10.1021/ma2025286 | Macromolecules 2012, 45, 3129−3136

Macromolecules

Article

Figure 8. Schematic representation of the proposed structural change of the PMMA−PS mixed polymer brush film upon exposure to dichloromethane and acetic acid, respectively. Upon exposure to acetic acid the polystyrene collapses to the surface and gets covered by poly(methyl methacrylate) in an attempt to maximize the favorable PMMA−PMMA contacts. Upon treatment with dichloromethane the polystyrene microdomains swell.

Society (Institutsübergreifende Forschungsinitiative FRM II), the DFG (SPP-1369 GU 771/3-1, BE 3286/1-1; SPP-1181 GU 771/2, GU 771/6), the graduate school MAINZ, and the Chinese Academy of Science.

ment of the surrounding PMMA. Even under optimal conditions only a surface reconstruction of