1

e-Polymers 2002, no. 006. http://www.e-polymers.org Feature article:

Surface-directed phase separation in nanometer polymer films: self-stratification and pattern replication a Andrzej Budkowski 1 *, Andrzej Bernasik 2, Piotr Cyganik 1, Jakub Rysz 1, Rüdiger Brenn 3 1 Smoluchowski Institute of Physics, Jagellonian University, Reymonta 4, 30-059 Kraków, Poland; Fax +48-12-633-7086; ufbudkow@cyf-kr.edu.pl 2 Surface Spectroscopy Laboratory, University of Mining and Metallurgy and Joint Centre for Chemical Analysis and Structural Research, Jagellonian University, Reymonta 23, 30-059 Kraków, Poland 3 Fakultät für Physik, Universität Freiburg, H.-Herder-Str. 3, 79104 Freiburg i.Br., Germany (Received: February 1, 2002; published: February 23, 2002)

Abstract: Phase separation occurs in thin films of polymer blends when molecular mobility is promoted by a temperature above the glass transition but inside the two-phase region (temperature quench), or a common solvent added to the polymers (solvent quench). This phenomenon can be altered by a homogeneous surface or pre-patterned substrate, resulting, e.g., in self-stratification or pattern replication, respectively. Such self-organisation processes ordering polymer phases were observed for model polymer blends (deuterated/hydrogenated polystyrene, dPS/hPS, and deuterated/partially brominated PS, dPS/PBrS, both with hPS-polyisoprene diblocks added; dPS/poly(vinylpyridine) and PBrS/PVP) with high-resolution ion beam techniques (Nuclear Reaction Analysis, profiling and mapping mode of dynamic Secondary Ion Mass Spectrometry) and Atomic Force Micro-scopy. The self-stratification process is strongly affected by both the range as well as the strength of the surface/polymer interactions. This is illustrated for the temperature-quenched blends with surface-active copolymer additives tuning the interactions exerted by both external surfaces. The pattern transfer from the substrate to the films is demonstrated for the solvent-quenched blends. Pattern-directed composition variations (SIMS maps) coincide with free surface undulations (AFM images). The most effective pattern replication is achieved for the length scale of phase domain morphology comparable with the pattern periodicity and for carefully adjusted polymer/substrate interactions.

1. Introduction In recent years there has been an increasing interest in understanding interfacial phenomena, such as micro- and macro-phase separation, occurring in thin (submicron) polymer films (for a review, see refs. [1-4]). In particular, the influence of boundary surfaces on the phase domain morphology has been intensively studied.

a Presented at the Workshop on Polymer Dynamics, organised as part of the Excellence Centre of the European Commission in Lódz, Poland, November 15, 2001.

2

For bulk copolymer melts the amphiphilic character of block copolymers, where two or more unlike subchains are covalently linked, leads to the spontaneous creation of ordered domains of well-defined size and uniform spacing, each on the order of tens of nanometers [5]. For thin films this micro-phase separation process is altered by the presence of a homogeneous surface which controls the orientation of the domain morphology [1-3,6]. The domains can be also arranged into patterns reflecting lateral chemical [7] or topographic [8] modifications of the substrate. For bulk homopolymer blends macro-phase separation leads to the formation of an isotropic, disordered phase morphology with a characteristic length scale increasing in the course of the process with no specific equilibrium value [5]. External surfaces, relevant for thin film geometry, can however direct the separation process to create ordered phase domain structures [1,3,4]. For homogeneous surfaces this self-organisation can result in a multilayer or column-like phase domain morphology [3,4,9] with the interfaces between blend phases oriented (in general) parallel or perpendicular to the substrate, respectively (Fig. 1a). Self-stratification of thin films with initial compositional homogeneity, leading to a multilayer ordering, is the first [4,10-13] of two main problems considered in the present paper. Phase separation in polymer mixtures is also altered by the presence of a pre-structured substrate with lateral pattern of surface energy [7,14-21] (Fig. 1b). Pattern replication, i.e. the alignment of phase domains with respect to the patterned substrate (accompanied by free surface undulations), is the second main issue [19] discussed here. Self-assembly of polymer blends, cast as thin films on both homogeneous as well as patterned substrates, is a consequence of surface-directed phase separation [1,3,4,9-13,23-32], discovered [23,27] only 10 years ago and described shortly below. Molecular mobility in the phase-separating films is promoted by the temperature elevated above glass transition (temperature quench applied for partly miscible or immiscible mixtures) or a common solvent added to the polymer blend (solvent quench encountered commonly for immiscible blends).

Fig. 1. Self-organisation of thin polymer blend films. A homogeneous substrate (a) can induce multilayer (self-stratification) or column-like-morphology of phase domains (φ1 and φ2). A patterned substrate (b) can lead to the alignment of both – phase domains and free surface undulations (pattern replication) Studies on self-organisation in thin films composed of polymer blends are motivated by numerous technological applications. In addition to attractive aspects of self-assembly, specific phase domain morphologies obtained result in increased efficiency of various devices. For instance, multilayer morphology is used in gas-separating membranes [33] and polymeric photodiodes [34]. Column-like morphology is present in light emitting polymeric diodes with a variable colour [35] or in anti-

z

0

x,y

φφ1φφ2

a) b)

3

reflection coatings [36]. Patterned substrate is used in all-polymer field transistors to keep apart the phase regions of source and drain [37]. The rest of this work is organised as follows. First, we describe the essential details of the experiments illustrating two main sections. Then, the concept of surface-directed phase separation will be briefly sketched followed by the main sections on self-stratification and pattern replication, where we mainly (but not exclusively) focus on the role of the strength of ordering surface/polymer interactions. 2. Experimental part 2a. Polymers Self-stratification (s.-s.) was studied for two blends composed of: i) protonated hPS1 and deuterated dPS1 polystyrene; ii) deuterated dPS2 and partially brominated PBrS1 polystyrene. A small volume fraction (φPIPS ≤ 0.16) of symmetric polyisoprene-polystyrene PI-hPS diblock copolymer was added to both blends with the binary composition of both homopolymers kept equal to the critical value. To provide contrast necessary for our experimental techniques, some of these ternary system constituents (dPS2 and PI-hPS) were replaced by their isotopic counterparts (hPS2 and symmetric PI-dPS). Pattern replication (p.r.) was studied for two nearly critical blends composed of: i) poly(vinylpyridine) PVP and poly(bromostyrene) PBrS2; ii) PVP and deuterated polystyrene dPS3. Molecular characteristics of these polymers are given in Tab. 1. Polymers hPS1 and dPS1 were obtained from Tosoh Corporation (Japan) and Polymer Laboratories, respectively. Standards dPS2, hPS2, dPS3, PVP and PBrS2 were purchased from Polymer Standards Service, Mainz. Diblock copolymers PI-hPS and PI-dPS were kindly provided by Profs. T. Hashimoto (Kyoto University) and L.J. Fetters (Exxon), respectively. Poly(bromostyrene) PBrS1 was obtained by the bromination of the polystyrene (see ref. [9] and references therein). Tab. 1. Molecular characteristics of the polymers used

Polymer Mw Mw/Mn Polymer Mw Mw/Mn Bromination index in mol-%

hPS1

dPS1 dPS2 hPS2 dPS3

2.9 M 1.9 M 203 k 207 k 174 k

1.09 1.14 1.03 1.02 1.03

PVP PI-hPS PI-dPS

PBrS1 PBrS2

115 k 26 k 20 k

208 k 171 k

1.02 1.13 1.03 1.02 1.04

22.7 10.5

2b. Films cast on homogeneous and patterned substrates The polymer blend components were dissolved in a common solvent (toluene for s.-s.; tetrahydrofurane for p.r.). Then a drop of the solution was placed on a sub-strate, and the films were prepared through a rapid rotation of the substrate (spin casting) [38,39]. The combination of carefully adjusted total polymer concentration and spin speed resulted in films with average thickness of 400 - 500 (s.-s.) and 50 - 80 nm (p.r.). Homogeneous substrates with different surface energy (≈ 20 - 81

4

mJ/m2) were provided by Si wafers with evaporated Au- or native SiOx-layers (s.-s.), and by Self-Assembled Monolayers (SAM) deposited on Au-covered Si wafers (p.r.). Hydrophobic and hydrophilic SAM substrates were obtained by exposing Au surfaces to ethanol solutions of HS(CH2)15CH3 (CH3-SAM) and HS(CH2)15COOH (COOH-SAM) thiols, respectively [19,40,41].

The patterned substrates used were composed of alternating (with the periodicity λ = 4 µm) 2 µm wide stripes of i) Au and CH3-SAM; ii) CH3-SAM and COOH-SAM. Micro-contact printing procedure [42] was used to create CH3-SAM stripes on the Au surface (Fig. 2). This procedure supplemented by additional immersion in ethanol solution of HS(CH2)15COOH produced the second pattern [16-19].

4 µm

a) b)

c)

d)

PDMS

PDMS

SAM

1.1

0.0

µm

Au

Au

Fig. 2. Micro-contact printing: A patterned elastomer (PDMS) stamp (with a relief revealed by AFM in a) prints the ‘ink’ of HS(CH2)15CH3 molecules in ethanol solution onto the Au surface creating CH3-SAM stripes Blend films with and without PI-dPS (PI-hPS) diblock copolymer additives (s.-s.) were annealed in a vacuum oven at T = 150 ± 1oC (dPS2/PBrS1, hPS2/PBrS1) up to 10 h and at T = 190 ± 1oC (hPS1/dPS1) up to 38 days, and subsequently quenched prior to the examination with experimental techniques. 2c. Experimental techniques revealing blend film morphology The information on the phase domain structure was obtained with three different techniques (Fig. 3): Nuclear Reaction Analysis (NRA) [43,44], profiling [45-47] and mapping [13] mode of dynamic Secondary Ion Mass Spectroscopy (SIMS), and with Atomic Force Microscopy (AFM) combined with selective dissolution [15]. The free surface undulations were determined with AFM [48]. The phase domain structure in the direction (z) perpendicular to the blend surface was traced with two high-resolution (i.e. with a precision better than 10 nm) depth profiling techniques (yielding composition φ vs. depth z profiles φ(z) across thin films). In NRA (Fig. 3a) [43,44] energetic (E ≈ 1 MeV) 3He ions penetrate easily the film until they encounter 2H atoms labelling one of the blend components. The energetic spectrum of the nuclear reaction (non-resonant, 3He(2H, 1H) 4He) products (4He or 1H) provides the information on the profile φ(z) of the deuterated polymer. In profiling mode of dynamic SIMS (Fig. 3b) [45-47] low-energy (E ≈ 5 keV) ions sputter the sample, exposing its successive layers. The created secondary ions (eg., C-, CH-, C2H-, C2

2H-, CN-, Si-, Br-), identified by mass spectrometry, are monitored as a

5

function of the sputtering time. This yields profiles φ(z) of various mixture constituents (characterised by different secondary ions) simultaneously.

AFMtip

Fig. 3. Phase domain structure determined by: (a) NRA (depth profiling composition of deuterated polymers), (b) SIMS (depth profiling or mapping composition of polymers made up of different elements), (c) AFM (yielding topographic images resembling compositional maps of selectively dissolved polymer phase) The 3D and lateral phase domain structure was revealed by the mapping mode of dynamic SIMS (Fig. 3b) [13] and by AFM combined with selective dissolution (Fig. 3c) [15]. In the former method, the focused ion beam is scanned over a microscopic square area. The secondary ions created at different points of this area provide a compositional map. Due to sputtering, such compositional maps can be collected for successive depths z. This novel technique, just introduced [13], has a depth resolu-tion of 20 nm and a lateral (Raleigh) resolution better than 120 nm. 3. Surface-directed phase separation 3a. Temperature quench Typically, phase separation takes place when a polymer mixture is quenched in temperature from the one-phase region of the phase diagram into a point inside the coexistence curve (binodal). Here we focus on spontaneous phase separation (spinodal decomposition), corresponding to the mixture quenched inside the spinodal of the two-phase region (Fig. 4a) [3-5]. Thermally induced composition fluctuations may be presented as composition waves with growing amplitude and wavelength (Fig. 4b). These composition waves have random directions and phases in the bulk of the mixture (Fig. 4c). The surface perturbs this phenomenon: it breaks the symmetry of the system and preferentially attracts one of the blend components. As a result, composition waves with a fixed phase develop normally to the surface (Fig. 4d). In the late stage of this process the domains of coexisting phases are formed, and the growth of phase domain morphology is characterised by a single time-dependent length scale R (Fig. 4e). Morphology coarsening is driven in all its (diffusive, hydrodynamic: viscous and inertial) regimes [49]) by interfacial tension between coexisting phases γ = γ12. This coarsening is ordered by surface/polymer interactions, specified by the difference ∆γ = γ1−γ2 of surface tension between coexisting phases.

4He, 1H z

detector

sample (2H)

MeV 3He

keV Ga

y~10 µm

xz

massspectr.

b)a) c)

6

Tc T

Ta

c)

d)

φ1φ2 φAφ0

spinodal binodal

φAφ0

r

rR(t)

φ1φ2

a) b)

e)

Fig. 4. Bulk (c) and surface-directed (d) spinodal decomposition (a, b, e) of a binary mixture. Coexisting compositions are φ1 and φ2, while the initial concentration is φ0 Surface induced spinodal decomposition has been intensively studied in the last decade (e.g., refs. [1,3,4,9-13,23-32,50,51]). Depending on the strength of inter-actions exerted by both external surfaces, different kinetic pathways of phase coarsening are observed for blend films leading to different final phase domain structures. In general, bilayer and column-like morphologies (see Fig. 1a) are expected for flat surfaces with antisymmetric and symmetric surface fields, respectively [3]. The columns extend across the film up to both symmetric surfaces partially wetted by blend phases [9], or they are terminated at both surfaces by wetting layers of the same phase [13]. This picture is more complex for very thin films composed of blends with high interfacial tension, where the laterally separated phase domains are accompanied by coupled undulations of the free surface [52] as observed very recently [50,53]. 3b. Solvent quench The study of phase separation in temperature quenched binary mixtures might be problematic for polymers with large difference in glass transition temperature Tg, since the lower Tg polymer can degrade for annealing temperatures much higher than Tg of the other blend component [54,55]. One way to perform phase separation studies is to use a ternary system: two polymers dissolved in a common solvent. Such solution is dilute and therefore homogeneous. Phase separation of polymer components can be initiated when the solvent is removed from the solution (Fig. 5a). Such solvent quench [39,54-60] accompanies the spin-casting process (Fig. 5b-f) [38]. There are three consecutive phases of spin coating. First, most (≈90%) of the polymer solution is flung from the rotating substrate leaving a thin uniform film (Fig. 5b-c). Second, the film thickness is decreased due to fluid flow, which is a balance between centrifugal and viscous forces (Fig. 5c-d). It is during this stage that phase separation occurs for immiscible blends. Third, the viscosity of the film becomes so large that the fluid is frozen in place, and further solvent loss is due to solvent evaporation from the film surface (Fig. 5e-f). Some residual solvent present in the film after completion of the spin-casting procedure can be removed by baking. Laterally separated phase domains present in solvent-quenched thin films (and corresponding to column-like morphologies in temperature-quenched samples with flat surfaces) are always accompanied by free surface undulations [39,54-62]. Different theories, referring to various solvent effects, try to account for this pheno-menon [59,61-62]. The most versatile theory relates the surface undulations with a

7

different evaporation rate of the solvent from various blend phases [59-60]. This is a consequence of different solubility of various polymers in their common solvent. The solvent evaporates faster from the phase rich in less soluble polymer, which solidifies earlier. In turn, the phase rich in more soluble polymer is swollen. As the solvent evaporates finally also from this phase, it collapses and forms lower regions of the free surface. This mechanism is claimed to take place during the third stage of spin coating (Fig. 5d-f) [59].

AB

solvent

RT

φ1φ2

a) b)

c)

d)

e) f)φ2 φ1

φ0 Fig. 5. Phase separation (c-d) occurs in the course of the spin-casting process (b-f) for a certain range of decreasing solvent concentration (a) rather than at a fixed point of phase diagram (as in Fig. 4a). Surface undulations (e-f) are formed with the elevated phase rich in the polymer less soluble in a common solvent (φ2 in (e), φ1 in (f)) [59] 4. Self-stratification (s.-s.) Even for symmetric [13,50] (or weakly asymmetric [63]) interactions exerted by both external surfaces on polymer blend films, a transient self-stratification is realised for early and intermediate stages of surface-directed phase separation. However, a full self-stratification pathway leading to equilibrium bilayer morphology is possible only for antisymmetric surface-fields [3]. In both cases self-stratification can occur for blend films thicker that the wavelength of composition waves [64]. 4a. Surface/polymers interactions tuned by surface-active diblock copolymers

While the late stage of phase separation is driven by the interfacial tension γ, the surface-induced ordering of coexisting phases is caused by surface/polymer interactions, specified by the surface tension difference ∆γ between coexisting phases. For homogeneous substrates simple strategies have been frequently used to alter the phase separation process by coarse modification of γ (e.g. by exchanging polymers [51,56]) or ∆γ (e.g. by changing the chemical nature of the substrate [9,11]). A tuneable variation of γ (due to varied temperature [65-66]) or ∆γ (due to the tuned composition of mixed hydrophobic and hydrophilic SAMs covering the substrate [67]) has been accomplished only very recently leading e.g. to the first observation of the simple [65-66] - and reversal [67] - wetting transition (from partial to complete wetting). Approaches using mixed SAMs [67] or end-grafted random copolymer brushes [68], both with variable composition, were used to tune the effective interactions between polymers and the substrate. A controlled variation of both effective surface fields was achieved by adding surface-active copolymers to the mixtures [10,12,69]. The co-

8

polymers segregate predominantly to both surfaces of the blend film and reduce the fraction of the surface area exerting ordering interactions. Here we will discuss briefly this versatile approach to control the self-stratification process [10,12]. In our studies [10,12] polyisoprene-polystyrene PI-PS (PI-hPS and PI-dPS) diblock copolymers were used, added to two different critical mixtures: hPS1/dPS1 and dPS2/PBrS1 (hPS2/PBrS1). Isotopic polymer counterparts were used to provide the contrast necessary for the depth profiling techniques. Isotopic exchange (PI-hPS for PI-dPS and dPS2 for hPS2) does not introduce any discernible thermodynamic modifications [70-72] for the systems studied in this work as the changes of relevant parameters are too small [10,12]. The surface activity of diblock copolymers is illustrated in Fig. 6a for (3% of) PI-dPS additives in the annealed (for 300 min) mixture hPS2/PBrS1. Independently of the four-layer phase domain structure (indicated by the PBrS1 profile) a surface excess Γ of PI-PS copolymers at both external surfaces (peaks of PI-dPS profile at z = 0 and 400 nm) is observed. The PI-PS segregation is driven by two factors: i) the reduction in the surface tension caused by the PI moieties replacing host segments at the surface; ii) the enthalpic penalty paid by PI segments positioned in the bulk (and not at the surface) of the host. These factors are jointly expressed by the adsorption parameter β [4,73,74]. Forces driving the surface segregation of the major components of both blends are at least one order of magnitude weaker [10,12] and therefore preferential adsorption of PI-dPS copolymers from both mixtures is always observed. In fact, in contrast to homo-polymers, the PI-dPS diblocks segregate to both surfaces already during the spin-casting process [12,73,74].

0 200 4000.0

0.4

0.0

0.8

Γ[nm]

*φ

6

00 0.15

Γ <*ΓΓ >*Γ

φPI-dPS

dept z [nm]

Γ

φPBrS1

φPI-PS

Γ >*ΓφPI-PS >*φ

φPI-PS <*φ

d)

c)b)a)

Γ <*Γ

Fig. 6. Surface activity of PI-PS diblocks: (a) PI-dPS depth profile indicating surface excess Γ in the hPS2/PBrS1 blend film (with a four-layer domain structure revealed by PBrS1 profile). (b) The excess of PI-PS at the free surface of two matrix types plotted as a function of bulk copolymer concentration φPI-PS (closed squares for PI-dPS in hPS1 [74], an open square for PI-hPS in dPS1 [10]; closed circles for PI-dPS in hPS2/PBrS1 mixture [12]). The isotherms for hPS1/dPS1 and hPS2/PBrS1 are fitted with β = 3.7(1) · kBT and β = 4.5(1) · kBT, respectively. Surface excess *Γ and con-centration *φ values (dotted lines) correspond to the onset of a continuous layer formed at the surface by overlapping segregated copolymers (c-d)

The excess Γ of PI-PS copolymers is characterised by segregation isotherms (Fig. 6b) determined for the free surfaces of various matrices corresponding to both blends (dashed curve for hPS1/dPS1 and solid curve for hPS2/PBrS1, respectively). Strongly

9

segregating diblocks may overlap forming, for concentrations φPI-PS higher than the onset value *φ (equal to 0.025 and 0.011 for hPS1/dPS1 and hPS2/PBrS1, respectively), a continuous layer covering both surfaces of blend films (Fig. 6c-d). This occurs for surface excess values Γ > *Γ, that is when each diblock copolymer occupies an area smaller than that described by the block radius of gyration [10,12]. The onset volume fraction *φ separates two regimes of separated copolymer islands (φPI-PS < *φ) and a continuous diblock copolymer layer (φPI-PS > *φ), both reflected in experimentally detected characteristics such as segregation properties [4,73,74] or free surface morphology (prior to annealing) [11]. The concentrations of PI-PS co-polymer additives were too low for the critical micelle concentration transition to occur [4,73-75]. 4b. Ordering induced by short-range surface forces Surface-directed spinodal waves, observed for the hPS1/dPS1 blend with PI-PS additives annealed at 190oC (shallow quench with reduced temperature [3] ε = 0.5), originate mainly at the free surface while the field exerted by the SiOx substrate was concluded to be very weak [10]. Initially (up to 20 days) the thickness R' of the layer adjacent to the free surface (determined for φPI-PS = 0.005) grows following the power law R' ~ t1/3 (characteristic of the diffusive coarsening regime [3,49]). At later times R'(t) seems to level off at a plateau value corresponding to the effective range of the region where surface-driven phase separation dominates over its isotropic bulk-counterpart. The role of modified surface/polymer interactions in the self-stratification process was investigated [10] for samples annealed long enough to correspond to the situation with saturated R'(t). Typical results of these experiments are shown in Fig. 7.

R

R'

Fig. 7. Depth profiles (a-c) of deuterated and protonated components of the dPS1/hPS1 blend with varied composition φPI-PS of PI-hPS additives (see text for details). Self-stratification is progressively (a, d; b, e; c, f) destroyed for increased content φPI-PS, when the bulk isotropic mode of phase separation (f) prevails over its surface-directed (d) counterpart

RR

RRRR R

RRRRRRR

R

R

RRRRRRRR

RRRRRRRRRRRRRRRR

RR RRRRRRRRRR

RRRRRRRRRRRRRR

RRRR

R

RRRRRR

R

R

RRRRRRR

RRRRRRRRRRRRRRRRRRRRRRR

RRRRRRRRRRRRR

RRRRRRRR

RRRRR

R RR

φφ m

RRRRRRR RRR

RRRRRRR

RRRRRRRRR

R

RRRRRRRRRRRRR RRRR

RRRRRRRRRRRRRRRRRR

RRRRRRR

RRRRRRRRR

RRRRRRRR

RRRRRRRRRRRRRRRR

RRRRRR

RRRRRRRRRR R RRRRRRRRRRR

RRRRRRRRR

de pth z [ nm ]

R

RRRRRRRR

R

R

RRRRRRRR

RRRRRR

RRRRRRRRRRRRRRRRRR

RRRRRRRRRRRRRRRRRRRRRRRRRR

R

RR

RRR

R

R

RRRRR

RRRRRRRRRRRRRR

RRRR

RRRRRRRRRRRRRRRR

RRRRRRRRRRRRRR

RRRRRRRR

0 100 200 300 4000.0

0.2

0.4

0.6

0.8

1.0

φφdPS1,φφhPS1&PI-hPS

R' φPI-hPS=0

hP

S1

laye

r

φφM

RR

φPI-hPS=0.019 φPI-hPS=0.055

a) b) c)

hPS1 rich

dPS1 rich

z

d) e) f)

z z

10

Annealed samples were covered here with additional reference hPS1 layers prior to the measurements. Strong oscillations, propagating throughout the thin film, are visible in dPS1 and hPS1 profiles obtained for a pure isotopic mixture (Fig. 7a). They are much damped for φPI-PS = 0.019 (Fig. 7b) and completely exhausted at φPI-PS = 0.055 (Fig. 7c), i.e. for a diblock concentration φPI-PS higher than the onset value *φ. To present this behaviour more quantitatively, we have plotted as a function of φPI-PS (Fig. 8a) the relative amplitude (φM-φm)/<φdPS1> of the surface maximum φM, measured with respect to the mid-layer minimum φm in the oscillating dPS1 profile (see Fig. 7b). This would measure the strength of surface-directed spinodal waves. We read from the graph in Fig. 8a that this strength decreases monotonically with increasing φPI-PS and zeroes exactly at the onset value *φ of diblock concentration. All annealing experiments were performed for the unstable region of the ternary phase diagram [10]. Thus, spinodal decomposition is advocated for all the samples (e.g. Fig. 7f) and only its surface-directed mode (Fig. 7d), leading to self-stratification, is reduced to zero for φPI-PS > *φ. We conclude also that the effective surface/polymer interactions, leading to anisotropic composition waves, must be of short-range character.

a)

0.0

0.5

1.0

0.0 0.10

(φM-φm)/<φdPS1>

φPI-hPS*φ *φ φPI-hPS0.0 0.02

φ s

0.5

0.8

b)

Fig. 8. Strength (for definitions see text) of surface-directed composition waves in the blends dPS1/hPS1 (a) and dPS2/PBrS1 (b) plotted as a function of PI-PS copolymer concentration φPI-PS. The values plotted here were constant for time ranges: t ≥ 20 days for dPS1/hPS1 and 70 min ≤ t ≤ 360 min for dPS2/PBrS1. The onset PI-PS values *φ (cf. Fig. 6b) are marked by dashed lines. The amplitude of surface-directed composition waves zeroes for φPI-PS > *φ and short-range surface forces 4c. Ordering induced by long-range surface forces Phase separation, observed for the dPS2/PBrS1 blend with PI-PS additives annealed at 150oC (deep quench with ε = 5.2), is directed by two surfaces with competing fields: dPS2 and PBrS1 are attracted to the free surface and the Au substrate, respectively (Fig. 9a). Hence, a bilayer morphology is expected for equilibrium. The self-stratification process leading to this morphology was observed [12] to consist of two stages. First, growth of two surface lamellae leads to the formation of a four-layer phase structure (each lamella is followed by a subsurface layer depleted in the phase enriched at the surface, see Fig. 9a). Second, the four-layer phase domain structure is transformed into the final bilayer morphology (Fig. 10). It was observed for the first stage that the amplitude of the coarsening surface-directed composition waves is reduced for higher PI-PS concentrations φPI-PS. This is illustrated in Fig. 9a-d for blend films annealed long enough (360 min and 300 min for

11

Fig. 9a and b, respectively) to form a four-layer structure. Such behaviour was observed also for earlier times. In fact, for each PI-PS content the amplitude of spinodal waves (extending from the free surface) was found to be constant after 70 min of annealing. These constant values, expressed in terms of dPS2 surface composition φs (the sum of the amplitude and average dPS2 composition in the film <φdPS2> ≈ 0.5), are plotted in Fig. 8b as a function of φPI-PS. As previously (cf. Fig. 8a) the measure φs of the compositional oscillations is reduced for increased φPI-PS values. However, in contrast to the previous case, it does not vanish for diblock concentrations much higher than the onset value *φ. We conclude that the effective surface/polymer interactions must be of long-range character.

Fig. 9. Four-layer structure, formed in a dPS2/PBrS1 blend film with varied PI-hPS content, revealed by depth profiles and illustrated by schematic cartoons (a, c; b, d). Hydrodynamic regime of phase separation evidenced by: (e) nearly linear temporal evolution of free surface lamella, (f) mass flow channels (dark) extending from the free surface (AFM image of the film with dissolved upper dPS region). See text for details While the amplitude of the first stage of self-stratification was reduced for increased PI-PS concentration, the kinetics of this process (measured by the rate of the surface lamella growth, see Fig. 9e) was unchanged. In addition, this temporal evolution is nearly linear R' ~ t indicating a hydrodynamic flow mechanism of phase domain coarsening [3,49]. In this domain-coarsening regime, the phase wetting the surface flows from the bulk region of the film through continuous pathways in the subsurface depletion layer. Such mass flow channels were in fact observed with AFM combined with selective dissolution of the phase rich in dPS2. For instance, darker regions in Fig. 9f correspond to dPS2 mass flow channels present in the blend film (with φPI-PS = 0.015 annealed for 300 min). Diblock copolymers heavily modify the kinetics of the second stage of phase evolution when the four-layer structure is transformed into the equilibrium bilayer morphology. This is demonstrated in Fig. 10, where dPS2 depth profiles of the left and right column correspond to φPI-PS = 0.005 and 0.029, respectively. We notice at once that the transformation process is for higher diblock content (i.e. for weaker effective

φPI-PS=0.5%

a)

c)R' de pt h [nm]40020004002000

φdPS2,φPBrS1

0.0

1.0

0.5

φPI-PS=0.005 φPI-PS=0.029

b)

d)R

' [nm

]

50

00 200 time [min]

e)

φPI-PS=1.5% ,2.9%

5 µm

f)

12

surface fields) completed earlier! This apparently surprising result is easily explained (see also schematic cartoons in Fig. 10) when the hydrodynamic flow mechanism is taken into account [12]. Blends with low φPI-PS value are exposed to a strong surface field. As a result, surface-directed compositional waves with large amplitudes are created which coarsen forming a four-layer structure with well-defined almost continuous lamellae (channels supplying the material to the surface layers are allowed). Extensive rupturing of the two interior lamellae (Fig. 10a-b) out of four layers is necessary to develop perforations acting as channels for the mass flow leading to the final bilayer morphology. For blends with higher φPI-PS concentrations, weaker effective surface fields result in surface layers with more fragmented phase domains and reduced amplitudes of the surface-directed composition waves. The created phase plates of the four-layer structure are already very much fragmented (Fig. 10e) and no additional rupturing process is necessary to allow the hydro-dynamic flow to both surfaces confining the films. Hence the transformation process is expected to be faster.

Fig. 10. Second stage of the self-stratification process in dPS2/PBrS1 blends with varied content φPI-PS of PI-PS copolymers added: Transformation of the four-layer structure into the final bilayer morphology is completed earlier (e-g) for higher diblock content, i.e. for weaker effective surface fields 5. Pattern replication (p.r.) The transfer of the pattern from the substrate (with alternating stripes of equal width w and different surface energy, e.g., ∆γ and -∆γ) to the phase-domain structure of a

dPS 2

PBrS

1

strong effective surface field

a)

b)

c)

d)

φPI-PS=0.029φPI-PS=0.005

e)

f)

g)

t=300min t=300min

t=390min t=390min

t=555min t=555min

t=1000min

weak effective surface field

φdPS2 φdPS2

0.0

0.8

0.0

0.80.0

0.8

0.0

0.8

0.0

0.8

0.0

0.8

0.0

0.8

4002000

4002000de pt h [n m]

13

thin polymer blend film (Fig. 11) depends on a wide variety of parameters [7,14-22]. The parameters important for self-stratification, such as interfacial tension γ and surface tension difference ∆γ between coexisting phases, are relevant also here. The pattern replication process is controlled mainly (in addition to film thickness d) by the ratio R/w of substrate pattern periodicity λP = 2w and characteristic length scale 2R of the coarsening phase domain morphology [15,17-21]. The final morphology depends on the results of the competition between surface and interfacial energy [20-21]. Generally, a phase domain structure matching the substrate motif (Fig. 11c, eviden-ced by refs. [15,16,19]) is expected only for ∆γ w >> d γ, i.e. when substrate/polymer interactions dominate over the interfacial tension term (a free surface is usually assumed to be neutral). In the other limit, ∆γ w << d γ, there is no pattern transfer into the blend film and only local alignment close to the substrate with very strong surface fields is eventually observed [7].

Fig. 11. Pattern replication for polymer blend films with dominant substrate/polymer interactions (∆γw >> dγ) [21]. ‘Checkerboard’ morphology (b), equivalent to com-position waves ordered by a homogeneous surface (cf. Fig. 4d), is absent for films thinner that the wavelength of composition oscillations [16] A few numerical studies of pattern replication have been focused on phase domain evolution in polymer blend films confined by neutral - and patterned - flat surfaces [16,20,21] (Fig. 11). They have shown that two surface-directed phase separation processes are responsible for the pattern replication [16,20,21]: First, surface-directed composition waves (out-of-plane phase separation) are ordered by and originated at different substrate regions (a mode visible in Fig. 11b). Second, pattern-directed lateral (in-plane) bulk-like phase separation is driven by the reduction of interfacial and surface energy (e.g. relevant for Fig. 11c). 5a. Substrate patterns echoed by domain morphology and surface undulations The alignment of blend phases with respect to the patterned substrate has been observed for several mixtures with molecular mobility promoted by elevated temperature [14,16-18] or by a common solvent added to the polymer blend [7,14,15,19]. Here, pattern replication is illustrated by the results obtained for two solvent-quenched film blends: PVP/PBrS2 (Fig. 12) and PVP/dPS3 (Fig. 13). To understand the pattern replication process it is instructive to consider first the phase domain morphology of blend films cast on two homogeneous substrates identical with the stripes of surface energy modulations. For the PVP/PBrS2 blend cast on Au surface (Fig. 12a-c) the bilayer morphology is deduced (from AFM and SIMS data) with PBrS2- and PVP-rich lamellae adjacent to smooth free and substrate surfaces, respectively. Phase evolution is however not completed here as indicated by the PVP/PBrS2 interface much wider than expected for the thermodynamic

γγ

∆∆ γγ ((−−∆∆γγ ))λλP=2w

da) b) c)

14

equilibrium [19]. A drastically changed morphology is observed for the same blend coating the CH3-SAM substrate (Fig. 12d-f). Corrugated free surface topography (with vertical distance amplitude of ≈26 nm in Fig. 12d) and an effective thickness ∆z different for both phases (determined from SIMS profiles in Fig. 12e with ∆z(PVP) ≈ 75 nm and ∆z(PBrS2) ≈ 53 nm) suggest the formation of laterally disordered phase domains, which are higher and more narrow for PVP than for PBrS2. These results indicate that the more polar PVP is favoured by the Au surface, while no preferential attraction takes place for the CH3-SAM substrate.

Fig. 12. Overall phase domain morphology (and free surface topography) for the PVP/PBrS2 blend films cast at identical conditions on homogeneous substrates Au (a-c) and CH3-SAM (d-f) and on the substrate patterned with alternating stripes of Au and CH3-SAM (g-j). AFM images of the films as cast (a, d, g, h) and after selective dissolution of PVP (i; the same spot as in h). The insets in d) and g) are the results of Fourier analysis of the AFM images. The effective thickness ∆z of PVP and PBrS2 can be determined from SIMS profiles (b, e; with 1 sputtering cycle ≈ 1 nm) Now we are ready to discuss pattern replication in the PVP/PBrS2 blend cast on the heterogeneous substrate with Au and CH3-SAM stripes. AFM images (Fig. 12g-h) show relatively good ordering of surface undulations, resembling the substrate pattern (cf. Fig. 2). Narrow and elevated strips alternate with wider regions located lower. The results obtained for a homogeneous substrate (Fig. 12a-f) suggest that the linear protrusions correspond to the phase rich in PVP. This structural feature is confirmed by the AFM image (Fig. 12i), which was taken after selective dissolution of the PVP-rich phase domains. The overall morphology concluded from the measure-ments is presented in Fig. 12j. We postulate that protruded linear PVP domains are located on Au stripes, while the PBrS2-rich phase is displaced onto CH3-SAM sub-

φPBrS2,φPVP

AuPV

PPB

rS2

Au

CH3-SAM

PBrS

2PVP

PVPPBrS2

Au

b) c)

d)

a)

e) f)

g)h)

j)30nm

0

1

0 50depth z [sputt. cycles]

depth z [sputt. cycles]

φPBrS2,φPVP

0

1

0 50∆z(PVP)

∆z(PBrS2)

20µ m

20µ m

12µm

i)

0

20µm

CH3-SAM

15

strate regions with no preferential attraction. We note that preferential attraction of one blend component to one substrate region is the main force driving pattern replication in many experiments [7,14-19]. In contrast, it is usually assumed in numerical studies that both blend components preferentially segregate to alternating stripes of the heterogeneous substrate [16,20,21]. 5b. Correspondence between domain morphology and surface undulations Patterned phase-domain structures can be reflected in the deformations of a flexible free surface, as it is observed for solvent-quenched blend films (Fig. 12g-j) [15,19] and often concluded for temperature-quenched ultra-thin film blends with high interfacial tension [16-18]. The overall coincidence between surface topography and phase domain structure has been concluded based on the results of scanning probe microscopy and related techniques (e.g. lateral force microscopy [16-18] or AFM combined with selective dissolution [15,19]; e.g. Fig. 12h-i). These techniques cannot however yield the exact composition maps of, say, blend component A at various depths z, but only images (prone to eventual artefacts [12,39,59]) resembling the lateral structure of the A-rich phase. The application of the mapping mode of dynamic SIMS [13] and Fourier analysis of composition- and topography-maps have enabled us to examine this issue again for a solvent-quenched blend.

Fig. 13. SIMS images (a-c) of the PVP maps determined at the distance z ≈ 18 nm (a), ≈ 55 nm (b) and ≈ 92 nm (c) from the free surface of the PVP/dPS3 blend film cast on the substrate with alternating Au and CH3-SAM stripes. AFM image (d) of free surface undulations (different spot). Power spectra Pa(k) (e) calculated for Fourier transforms of compositional (dashed line, for Fig. 13c) and topographic (solid line, for Fig. 13d) maps. See text for details The PVP/dPS3 blend cast on the heterogeneous substrate with Au and CH3-SAM stripes is characterised by PVP lateral distributions determined by mapping SIMS mode at subsequent depths z ≈ 18 nm (Fig. 13a), 55 nm (Fig. 13b) and 92 nm (Fig. 13c) as well as by an AFM image of free surface undulations (Fig. 13d). The com-

a) b) c)

d) e)20µm

P a(k

) [a.

u.]

0

40

0.0 1.0

AFM dataSIMS data

wave vector k [1/µm]

16

parison of both data types proves that the overall phase domain morphology of Fig. 12j (with PBrS2 exchanged by dPS3) is preserved. To examine more quantitatively the topographic and compositional maps we have analysed their two-dimensional fast Fourier transforms (FFT). The spectra of FFT amplitudes, computed for the blends cast on homogeneous (CH3-SAM) substrates, resemble isotropic diffusive rings (see e.g. inset to Fig. 12d). The radial average of the FFT ring exhibits a maximum at wave vector k = 1/(2R). Hence the length scale 2R of the phase domain morphology (e.g., Fig. 12d-f) can be determined. The value 2R is characteristic of given spin-casting conditions for both homogeneous (CH3-SAM) and heterogeneous (striped) substrates. FFT spectra calculated for the blends on patterned substrates exhibit isotropic (diffuse ring) and anisotropic components (series of sharp peaks distributed along one line, see e.g. inset to Fig. 12g). The latter were used to compute the power spectra Pa(k) analysed below. An exact coincidence between topography and morphology of PVP/dPS3 blends cast on stripes of Au and CH3-SAM is confirmed by the results of the Fourier analysis (Fig. 13e). The power spectrum Pa(k) of the AFM image determined for the free surface topography (Fig. 13d, solid line in Fig. 13e) is in accordance with Pa(k) of the PVP domain (SIMS) chemical map recorded for the region adjacent to the substrate (Fig. 13c, dashed line in Fig. 13e). This result shows also that Fourier characteristics of surface topography of very thin films replicating a substrate pattern can, in principle, be used as a measure of related phase domain morphology. A similar procedure is often used for blend films cast on homogeneous substrates [17-19,39]. 5c. Optimum conditions for pattern replication The conditions necessary to obtain high quality transfer of the substrate pattern are addressed in recent works [7,15,17-19]. They focus almost exclusively on the role of the length scale 2R of coarsening phase domain morphology (discussed with respect to the periodicity λP = 2w of the substrate pattern for regions: R ≤ w [15,17,18], R ≥ w [19] and R >> w [7]). This problem is illustrated below for R ≥ w [19]. In addition, we discuss also the issue of substrate/polymer interactions. AFM images (Fig. 14a-b) of the blend PVP/dPS3 cast on the Au/CH3-SAM stripes indicate drastic changes accompanying the modification of the phase domain scale from 2R = 4.1(1) µm (Fig. 14a) to 2R = 5.3(1) µm (Fig. 14b), which was obtained by varying the total polymer concentration in a common solvent. Apparently too large phase domains (Fig. 14b) developing during the solvent quench are not very sus-ceptible to periodic variations of substrate interactions. Effective pattern replication (Fig. 14a) is achieved for the scale length 2R which is commensurate with the periodicity λP = 2w = 4 µm. Carefully adjusted substrate/polymer interactions specify also the optimum conditions of pattern transfer. To demonstrate this we have exchanged: i) one of two polymer blend components (non-polar dPS3 for slightly polar PBrS2, see Figs.14a and c); ii) one of two alternating stripes forming the patterned substrate (Au for polar COOH-SAM, see Figs.14a and d). The first substitution worsens the quality of the created patterns. Linear protrusions, well separated for PVP/dPS3 (Fig. 14a), are sporadically connected for PVP/PBrS2 (Fig. 14c). The polymer exchange hardly affects the compatibility of blend com-ponents (solubility parameters of dPS3 and PBrS2 are very similar [19]). Therefore we attribute the change in pattern replication to the reduced difference in surface tension

17

∆γ between blend components positioned on Au stripes (Au may slightly attract polar PBrS2 but not dPS3).

Fig. 14. Pattern replication modified by: i) the change in the length scale 2R of phase domain morphology (a-b), ii) the variation of substrate/polymer interactions (a, c and a, d). AFM images of the blend films PVP/dPS3 (a, d with 2R = 4.1 ± 0.1 µm; b with 2R = 5.3 ± 0.1 µm) and PVP/PBrS2 (c with 2R = 3.9 ± 0.1 µm), cast on the substrate stripes: Au/CH3-SAM (a-c) and COOH-SAM/CH3-SAM (d), both alternating with the periodicity λP = 4 µm The exchange of the striped regions of Au for polar COOH-SAM results in a larger surface energy difference with respect to the other stripes composed of non-polar CH3-SAM. Surprisingly this step does not improve but rather worsens the situation for the PVP/dPS3 blend films with 2R ≅ λP (cf. Figs. 14a and d). Most probably the interactions (driving the lateral order in the film) between polar PVP and a dipole moment induced in the Au surface are more favourable than those between PVP and the COOH-terminated SAM layer. 6. Conclusions Surface-directed phase separation in nanometer polymer films can lead to self-stratification and pattern replication, both involved in present and future technological applications. A novel strategy to modify surface/polymer interactions (with interfacial tension γ kept constant and varied surface tension difference ∆γ between coexisting phases) was introduced and demonstrated for the self-stratification of model temperature-quenched blend films. In contrast to similar methods established earlier, surface active diblock copolymers allow us to tune the interactions at both surfaces confining the blend film. As a result the amplitudes of the composition waves, directed by both surfaces, can be varied. This strategy is effective for short- and for long-range surface forces. Surface active diblocks can suppress the composition

a) b)

c) d)20

µ m

18

waves for the short-ranged fields, while for the long-ranged forces they can induce earlier completion of the final phase domain morphology. The pattern transfer from the substrate to: i) the phase domain structure (in the film interior) and ii) the free surface topography was demonstrated for model solvent-quenched blend films. For the first time the exact correspondence between the former and the latter was concluded based on Fourier analysis of compositional (mapping mode of dynamic SIMS) and topographic (AFM) maps. The effective pattern transfer can be achieved for i) a length scale of phase domain morphology commensurate with the pattern periodicity, and for ii) carefully adjusted substrate/ polymer interactions. Acknowledgement: This work was supported by the Polish Committee of Scientific Research (grant no 2 P03B 098 19), the Deutsche Forschungsgemeinschaft, Graduiertenkolleg “Structure formation in macromolecular systems”, Grant of the Institute of Physics, by Reserve of the Rector of the Jagellonian University. One of us (J.R.) is grateful to the Foundation for Polish Science for support. The authors also thank Professor M. Szymonski for the access to the AFM microscope at the Regional Laboratory for Physical and Chemical Analysis of the Jagellonian University. [1] Krausch, G.; Mater. Sci. Eng. 1995, R14, 1, and references therein. [2] Russell, T.P.; Curr. Opin. Colloid Interface Sci. 1996, 107, 1. [3] Binder, K.; Adv. Polym. Sci. 1999, 138, 1, and references therein. [4] Budkowski, A.; Adv. Polym. Sci. 1999, 148, 1, and references therein. [5] Bates, F.S; Science 1991, 251, 898. [6] Mansky, P.; Huang, E.; Liu, Y.; Russell, T. P.; Hawker, C.; Science 1997, 275, 1458. [7] Rockford, L.; Liu, Y.; Mansky, P.; Russell, T. P.; Yoon, M.; Mochrie, S. G. J.; Phys. Rev. Lett. 1999, 82, 2602. [8] Fasolka, M. J.; Harris, D. J.; Mayes, A. M.; Yoon, M.; Mochrie, S. G. J.; Phys. Rev. Lett. 1997, 79, 3018. [9] Bruder, F.; Brenn, R.; Phys. Rev. Lett. 1992, 69, 624. [10] Rysz, J.; Bernasik, A.; Ermer, H.; Budkowski, A.; Brenn, R.; Hashimoto, T.; Jedlinski, J.; Europhys. Lett. 1997, 40, 503. [11] Rysz, J.; Ermer, H.; Budkowski, A.; Lekka, M.; Bernasik, A.; Wróbel, S.; Brenn, R.; Lekki, J.; Jedlinski, J.; Vacuum 1999, 54, 303. [12] Rysz, J.; Ermer, H.; Budkowski, A.; Bernasik, A.; Lekki, J.; Juengst, G.; Brenn, R.; Kowalski, K.; Camra, J.; Lekka, M.; Jedlinski, J.; Eur. Phys. J. E 2001, 5, 207. [13] Bernasik, A.; Rysz, J.; Budkowski, A.; Kowalski, K.; Camra, J.; Jedlinski, J.; Macromol. Rapid Commun. 2001, 22, 829. [14] Krausch, G.; Kramer, E.J.; Rafailovich, M.H.; Sokolov, J.; Appl. Phys. Lett. 1994, 64, 2655. [15] Böltau, M.; Walheim, S.; Mlynek, J.; Krausch, G.; Steiner, U.; Nature (London) 1998, 391, 877.

19

[16] Karim, A.; Douglas, J. F.; Lee, B. P.; Glotzer, S. C.; Rogers, J. A.; Jackman, R. J.; Amis, E. J.; Whitesides, G. M.; Phys. Rev. E 1998, 57, R6273. [17] Ermi, B.D.; Nisato, G.; Douglas, J.F.; Rogers, J.A.; Karim, A.; Phys. Rev. Lett. 1998, 81, 3900. [18] Nisato, G.; Ermi, B.D.; Douglas, J.F.; Karim, A.; Macromolecules 1999, 32, 2356. [19] Cyganik, P.; Bernasik, A.; Budkowski, A.; Bergues, B.; Kowalski, K.; Rysz, J.; Lekki, J.; Lekka, M.; Postawa, Z.; Vacuum 2001, 63, 307. [20] Kilehorn, L.; Muthukumar, M.; J. Chem. Phys. 1999, 111, 2259. [21] Shou, Z.; Chakrabarti, A.; Polymer 2001, 42, 6141. [22] Seok, C.; Freed, K.F.; Szleifer, I.; J. Chem. Phys. 2000, 112, 6452. [23] Ball, R. C.; Essery, R. L. H.; J. Phys. Condens. Matter 1990, 2, 10303. [24] Marko, J.F.; Phys. Rev. E 1993, 48, 2861. [25] Puri, S.; Binder, K.; Frisch, H.L.; Phys. Rev. E 1997, 56, 6991. [26] Chen, H.; Chakrabarti, A.; Phys. Rev. E 1997, 55, 5680. [27] Jones, R. A. L.; Norton, L. J.; Shull, K. R.; Kramer, E.; Bates; F. S.; Wiltzius, P.; Phys. Rev. Lett. 1991, 66, 1326. [28] Krausch, G.; Dai, C.-A.; Kramer, E.J.; Bates, F.S.; Phys. Rev. Lett. 1993, 71, 3669. [29] Krausch, G.; Dai, C.-A.; Kramer, E.J.; Marko, J.F.; Bates, F.S.; Macromolecules 1993, 26, 5566. [30] Krausch, G.; Kramer, E.J.; Bates, F.S.; Marko, J.F.; Brown, G.; Chakrabarti, A.; Macromolecules 1994, 27, 6768. [31] Jandt, K.D.; Heier, J.; Bates, F.S.; Kramer, E.J.; Langmuir 1996, 12, 3716. [32] Heier, J.; Kramer, E.J.; Revesz, P.; Battistig, G.; Bates, F.S.; Macromolecules 1999, 32, 3758. [33] Bikson, B.; Nelson, J.K.; Muruganandam, N.; J. Membr. Sci. 1994, 94, 313. [34] Halls, J.J.M.; Walsh, C.A.; Greenham, N.C.; Marseglia, E.A.; Friend, R.H.; Moratti, S.C.; Holmes, A.B.; Nature (London) 1995, 376, 498. [35] Berggren, M.; Inganäs, O.; Gustafsson, G.; Rasmusson, J.; Andersson, M.R.; Hjertberg, T.; Wennerström, O.; Nature (London) 1994, 372, 444. [36] Walheim, S.; Schäffer, E.; Mlynek, J.; Steiner, U.; Science 1999, 283, 520. [37] Sirringhaus, H.; Kawase, T.; Friend, R.H.; Shimoda, T.; Inbasekaran, M.; Wu, W.; Woo, E.P.; Science 2000, 290, 5499. [38] Lawrence, C.J.; Phys. Fluids 1988, 31, 2786. [39] Cyganik, P.; Budkowski, A.; Raczkowska, J.; Postawa, Z.; Surface Science 2002, in press. [40] Larsen, N.B.; Biebuyck, H.; Delamarche, E.; Michel, B.; J. Am. Chem. Soc. 1997, 119, 3017. [41] Bergues, B.; Lekki, J.; Budkowski, A.; Cyganik, P.; Lekka, M.; Bernasik, A.; Rysz, J.; Postawa, Z.; Vacuum 2001, 63, 297.

20

[42] Kumar, A.; Whitesides, G.M.; Appl. Phys. Lett. 1993, 63, 2002. [43] Chaturvedi, U.K.; Steiner, U.; Zak, O.; Krausch, G.; Schatz, G.; Klein, J.; Appl. Phys. Lett. 1990, 56, 1228. [44] Kerle, T.; Scheffold, F.; Losch, A.; Steiner, U.; Schatz, G.; Klein, J.; Acta Polym. 1997, 48, 548. [45] Coulon, G.; Russell, T. P.; Deline, V. R.; Green, P. F.; Macromolecules 1989, 22, 2581. [46] Schwarz, S. A.; Wilkens, B. J.; Pudensi, M. A. A.; Rafailovich, M. H.; Sokolov, J.; Zhao, X.; Zhao, W.; Zheng, X.; Russell, T.P.; Jones, R. A. L.; Mol. Phys. 1992, 76, 937. [47] Bernasik, A.; Rysz, J.; Budkowski, A.; Kowalski, K.; Camra, J.; Jedlinski, J.; in: ‘ECASIA 97’; Olefjord, I.; Nyborg, L.; Bryggs, D.; Eds.; John Wiley & Sons, New York 1997. [48] Marti, O.; in “STM and SFM in Biology”, chapter 1, Academic, New York 1993. [49] Lookman, T.; Wu, Y.; Alexander, F.J.; Chen, S.; Phys. Rev. E 1996, 53, 5513. [50] Wang, H.; Composto, R.J.; Europhys. Lett. 2000, 50, 622. [51] Geoghegan, M.; Ermer, H.; Juengst, G.; Krausch, G.; Brenn, R.; Phys. Rev. E 2000, 62, 940. [52] Keblinski, P.; Kumar, S.K.; Maritan, A.; Koplik, J.; Banavar, J.R.; Phys. Rev. Lett. 1996, 76, 1106. [53] Ermi, B.D.; Karim, A.; Douglas, J.F.; J. Polym. Sci., Part B: Polym. Phys. 1998, 36, 191. [54] Dalnoki-Veress, K.; Forrest, J.A.; Stevens, J.R.; Dutcher, J.R.; J. Polym. Sci., Part B: Polym. Phys. 1996, 34, 3017. [55] Dalnoki-Veress, K.; Forrest, J.A.; Stevens, J.R.; Dutcher, J.R.; Physica A 1997, 239, 87. [56] Gutmann, J.S.; Müller-Buschbaum, P.; Stamm, M.; Faraday Discuss. 1999, 112, 258. [57] Affrossman, S.A .; Henn, G.; O ’Neill, S.A.; Pethrick, R.A.; Stamm, M.; Macromolecules 1996, 29, 5010. [58] Affrossman, S.A .; O ’Neill, S.A.; Stamm, M.; Macromolecules 1998, 31, 6280. [59] Walheim, S.; Böltau, M.; Mlynek, J.; Krausch, G.; Steiner, U.; Macromolecules 1997, 30, 4995. [60] Walheim, S.; Ramstein, M.; Steiner, U.; Langmuir 1999, 15, 4828. [61] Takahara, A.; Nakamura, K; Tanaka, K.; Kajiyama, T.; Macromol. Symp. 2000, 159, 89. [62] Ton-That, C.; Shard, A.G.; Teare, D.O.H.; Bradley, R.H.; Polymer 2001, 42, 1121. [63] Straub, W.; Bruder, F.; Brenn, R.; Krausch, G.; Bielefeldt, H.; Kirsch, A.; Marti, O.; Mlynek, J.; Marko, J. F.; Europhys. Lett. 1995, 29, 353. [64] Sung, L; Karim, A; Douglas, J.F.; Han, C.C.; Phys. Rev. Lett. 1996, 76, 4368.

21

[65] Rysz, J.; Budkowski, A.; Bernasik, A.; Klein, J.; Kowalski, K.; Jedlinski, J.; Fetters, L.J.; Europhys. Lett. 2000, 50, 35. [66] Rysz, J.; Budkowski, A.; Scheffold, F.; Klein, J.; Fetters, L.J.; Bernasik, A.; Kowalski, K.; Macromol. Symp. 2000, 149, 277. [67] Genzer, J.; Kramer, E.J.; Phys. Rev. Lett. 1997, 78, 4946. [68] Mansky, P.; Liu, Y.; Huang, E.; Russell, T.P.; Hawker, C.; Science 1997, 275, 1458. [69] Huang, E.; Russell, T.P.; Harrison, C.; Chaikin, P.M.; Register, R.A.; Hawker, C.J.; Mays, J.; Macromolecules 1998, 31, 7641. [70] Budkowski, A.; Rysz, J.; Scheffold, F.; Klein, J.; Fetters, L.J.; J. Polym. Sci., Part B: Polym. Phys. 1998, 36, 2691. [71] Budkowski, A.; Klein, J.; Eiser, E.; Steiner, U.; Fetters, L.J.; Macromolecules 1993, 26, 3858. [72] Scheffold, F.; Eiser, E.; Budkowski, A.; Steiner, U.; Klein, J.; Fetters, L.J.; J. Chem. Phys. 1996, 104, 8786. [73] Budkowski, A.; Klein, J.; Steiner, U.; Fetters, L.J.; Macromolecules 1993, 26, 2470. [74] Budkowski, A.; Klein, J.; Fetters, L.J.; Macromolecules 1995, 28, 8571. [75] Budkowski, A.; Steiner, U.; Klein, J.; Fetters, L.J.; Europhys. Lett. 1992, 20, 499.