Gradient and patterned polymer brushes by photoinitiated “grafting through” approach

Authors

  • Timothy P. Enright,

    1. Department of Chemistry & Biochemistry, University of the Sciences in Philadelphia, Philadelphia, Pennsylvania 19104
    Search for more papers by this author
  • Daniel Hagaman,

    1. Department of Chemistry & Biochemistry, University of the Sciences in Philadelphia, Philadelphia, Pennsylvania 19104
    Search for more papers by this author
  • Maryana Kokoruz,

    1. Department of Chemistry & Biochemistry, University of the Sciences in Philadelphia, Philadelphia, Pennsylvania 19104
    Search for more papers by this author
  • Natalia Coleman,

    Corresponding author
    1. Department of Pharmaceutical Sciences, University of the Sciences in Philadelphia, Philadelphia, Pennsylvania 19104
    • Department of Pharmaceutical Sciences, University of the Sciences in Philadelphia, Philadelphia, Pennsylvania 19104
    Search for more papers by this author
  • Alexander Sidorenko

    Corresponding author
    1. Department of Chemistry & Biochemistry, University of the Sciences in Philadelphia, Philadelphia, Pennsylvania 19104
    • Department of Chemistry & Biochemistry, University of the Sciences in Philadelphia, Philadelphia, Pennsylvania 19104
    Search for more papers by this author

Abstract

More than 2 decades of active investigations in the field of polymer brushes have revealed continuous and growing interest in different aspects of synthesis, properties, and applications of tethered polymers. In this article, we report on our recent advances in brush synthesis. The method we explore is based on the combination of “grafting through” approach with the functional anchoring polymer layer technique. We introduce the photoinitiated “version” of synthesis of polyacrylamide brushes. Both homogeneous depositions and laterally resolved gradient and patterned samples have been prepared by this technique. The results for flat polymer brushes, that is, thickness, stability, and contact angles, are complimented by kinetic parameters as deducted from analysis of gradient samples obtained by the method of a sliding mask. A microscopic shadow mask is used to fabricate patterned brushes. The microscopically patterned brushes demonstrate high lateral resolution limited by optical phenomena. Finally, we have performed a viability assaying of neuronal cell on both flat and patterned brushes. Sufficient restraint of cell adhesion on polyacrylamide photobrushes and very low cytotoxicity of the brush components (polymer brush itself, anchoring layer) make photografting a promising platform to control cell deposition and surface localization. © 2010 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys 48: 1616–1622, 2010

INTRODUCTION

This work was inspired by more then 10 years of our collaboration with Prof. Dr. Manfred Stamm. His interest in different aspects of physics and chemistry of polymers at the interface initialized our work dedicated to polymer brushes on flat surfaces.1 Based on classic works on polymer chains covalently tethered to solid substrate published by De Gennes,2 Milner,3 Boven et al.,4 Ligoure and Leibler,5 Halperin et al.,6 and Rafailvoich and coworkers7 in a series of articles, we investigated details of formation of the polymer brushes and fundamental properties of polymer chains at the interface with solid substrates. End-functionalized polymers can be tethered to substrate surface modified with a corresponding anchoring layer, usually self-assembled monolayer8 or adsorbed functional polymer.9, 10 This “grafting to” approach provides rather low grafting densities of grafted polymers. The so-called “grafting from” approach consists of attachment of the initiator to the substrate surface and the successive propagation of polymerization from the surface.11–13 This method assures high grafting density with more than 100-nm-thick brushes. Both approaches and the wide range of applications of polymer brushes have been in the focus of the recent reviews.14–21

The third approach is “grafting through.” It gained very little attention in research literature. Recently, the approach was further developed by Ruhe and coworkers by application of its principles to flat surfaces.22 It consists of introduction of C[DOUBLE BOND]C bonds onto the substrate surface; the bonds are involved in the propagation step of macromolecular radicals initiated in volume. The method allows for the synthesis of polymer brushes with relatively high grafting densities. The polymer brushes produced in this way reveal high stability, sufficient grafting density even in dry state, and perfect homogeneity (Fig. 1). In several modifications, the approach was used as a convenient method of polymer brush fabrication.23–25

Figure 1.

Schematic presentation of photografting of PAAm brush: the synthetic aspect. The anchoring layer of PGMA is covalently attached to silica surface by reaction of epoxy groups with silanols (a). The approach provides sufficient density of the epoxy groups to be modified with acrylic acid providing C[DOUBLE BOND]C reactive double bond (b). The radical polymerization starts in volume by photoinitiator In (Insert) and propagates while occasionally reaches the surface-attached acrylic groups, thus incorporating them into growing polymer chains R• (c).

Here, we introduce a photoinitiated version of “grafting through” technique. We will show that the method can be used for convenient and reproducible fabrication of polymer brushes on flat surfaces. Furthermore, photografting allows easy synthesis of homogeneous, gradient, or patterned brushes.

EXPERIMENTAL

Materials

The monomers acrylamide (AAm), acrylic acid (AAc), and glycidyl methacrylate (GMA) were purchased from Sigma-Aldrich and used as received. The synthesis of poly(glycidyl methacrylate) (PGMA) was performed by radical polymerization in 2-butanone (Sigma-Aldrich) solution (30%) in the presence of initiator (AIBN) at 60 °C for 5 h. The crude PGMA was precipitated from polymerizate solution in diethyl ether (Sigma-Aldrich); the dissolution–precipitation cycle was repeated six times to obtain neat PGMA. The polymer was stored in Ar atmosphere at 4 °C to prevent crosslinking of glycidyl groups. The 1% PGMA solution in chloroform was used for PGMA adsorption onto substrates.

The substrates silicon wafers (Addison Engineering {100} orientation) and glass slides were cleaned successively in an ultrasonic bath (dichloromethane) for 15 min and an “alkali piranha” bath (25% H2O2 and 25% NH4OH in water; chemical hazard!) for 40 min at 82 °C and then thoroughly rinsed with Millipore water and dried under an argon flow.

Photografting

In a typical experiment, we performed photografting in three consecutive steps. First, an adsorbed layer of PGMA was deposited from 1% solution in chloroform onto the substrate surface. The thickness of the layer was 1.3 nm according to spectroscopic ellipsometry on silicon wafers. The samples were placed in a humidity chamber (85–90%, 25 °C) for 16 h to assure the reaction of the glycidyl groups of the PGMA with the silanol groups of the silicon dioxide. Additionally, crosslinking of glycidyls occurred in the presence of water, improving stability of the anchoring layer.10 Second, the samples were immersed in acrylic acid for 10 min. This step introduces C[DOUBLE BOND]C double bonds onto the surface due to the reaction of the rest of the glycidyl groups exposed on the top of the PGMA layer with the carboxyl groups of the acrylic acid. The samples were then thoroughly rinsed with water. The third step was grafting via the anchoring layer. An aqueous solution of 1% of photoinitiator (2-hydroxy-40-(2-hydroxyethoxy)-2-methylpropiophenone) and 10–50% of monomer (acrylamide, acrylic acid, or their mixture) was deposited on the sample, covered with a quartz slide, and exposed to ultraviolet light (mercury “BlakRay” lamp, 90 cm distance) for 5–10 min. Thorough rinsing with water to remove any unattached polymer was the final step of the protocol. The thickness of prepared brushes was in the range of 5–40 nm, depending on the concentration of the monomers as measured with a spectroscopic ellipsometer (Angstrom Advanced) using flat reference samples.

A shadow mask technique was used to estimate the applicability of the method to create micropatterned brushes. The mask was prepared by vacuum deposition of gold using ion-beam lithography. The brush patterns were investigated using AFM/tapping mode. The grafting gradients were fabricated with a sliding mask. The mask, assembled with computer-controlled step-motor unit, was moving with microscopic accuracy and thus allowed us to control exposition of a particular location on the sample. The brush gradients were measured as thickness versus X position on the sample by a spectroscopic ellipsometer with accuracy 0.1 nm in thickness and 1 mm in X position.

Brush Characterization

Samples with grafted polymers were characterized by several different methods depending on the substrate. Atomic force microscopy (AFM, Innova, Veeco Metrology) in tapping mode was used to measure thickness of the polymer depositions by scratch test. This method was applied to all samples, both wafers and glass slides. WSxM software was used for AFM data analysis.26 A spectroscopic ellipsometer (PhE-102, Angstrom Advanced) was used for measuring the thickness of wafer samples. We applied the standard model for Si wafers according to ref. 27 and transparent polymer layer with refractive index of 1.50–1.55 in the range of 350–850 nm of wavelength of the incident beam. Hydrophilic–hydrophobic changes upon surface modification on different stages were monitored by water contact angle setup (a home-built setup with a CCD camera). The measurements were performed in the sessile droplet mode and repeated at least three times to obtain reproducible values of contact angles.

Cell Culture and Assaying on PAAm Brushes

A neural cell line N2a (ATCC No. CCL-131) was used in this study. The cells were routinely propagated using Eagle's minimum essential medium (EMEM ATCC), supplemented with 10% fetal bovine serum (ATCC) and 1% penicillin/streptomycin (Gibco) in 100 cm2 Petri dishes at 37 °C in 5% CO2. The cells were plated at a concentration of 20,000 cells/well on a 96-well open-bottom plate. The brush samples and the reference samples were attached to the plates' bottoms providing specific cell culture substrates. The MTT assay (ATCC) was performed according to the manufacturer's instructions.28 The following determinants were optimized for this cell line: plating cell concentration; incubation time with MTT reagent, and incubation time with MTT detergent reagent. Absorbance was recorded at 570 nm by a microtiter plate reader Victor3 (Perkin-Elmer). The number of surviving cells is directly proportional to the level of the formazan product created. The statistical analysis was performed using the t-test and Origin 8.0 software.

RESULTS AND DISCUSSION

Homogeneous Photografting

The experimental photografting protocol reflects both the fundamentals of the “grafting through” approach of polymer brush growth and the specific features used in this investigation. Figure 1 illustrates two main steps of photografting: (i) surface preparation consists of chemisorption of PGMA [Fig. 1(a)] providing significant density of functional epoxy groups exposed at the exterior. They can be easily converted into C[DOUBLE BOND]C double bonds by reaction with acrylic acid [Fig. 1(b)]. Noteworthy, epoxy polymers used as the anchoring layers can be applied for tethering polymer brushes onto both inorganic oxides (SiO2 and TiO2) and noble metals (Au and Pt) as well;10, 29, 30 (ii) the process of photografting is initiated by free radicals from photoinitiator dissolved in polymerization solution upon illumination with UV light (Fig. 1, Insert). The propagating macroradicals occasionally reach C[DOUBLE BOND]C double bonds at the substrate surface, thus incorporating them into growing polymer chains [Fig. 1(c)].

Qualitatively, the brush growth can be conveniently monitored by contact angle measurements. The contrast in hydrophobicity of the PGMA layer and the PAAm brush allows us to precisely control the completion of the surface deposition or conversion. Table 1 summarizes the contact angles and provides the direct proof of successful PAAm brush fabrication by photografting “through” method. The hydrophilic character of the clean Si wafer surface (θ < 10°) is altered by deposition of PGMA. Its chemisorbed layer of 1.3 nm in thickness reveals relatively hydrophobic behavior with a contact angle the same as in bulk (θ = 62°). Further modification with acrylic acid only slightly changes the contact angle (θ = 57°). The surface becomes very hydrophilic (θ < 15°) upon brush growth. Even a small amount of PAAm grafted to the surface (5-nm-thick brush) provides complete coverage of the surface as revealed by contact angle measurements (θ = 15°).

Table 1. Contact Angle θ (Sessile Droplet Method) Measurements at Different Stages of Sample Preparation
inline image

A classic quantitative analysis of polymer brush formation includes thickness measurements as a function of the monomer and initiator concentration. These results are shown in Figure 2. The monomer (AAm) concentration is limited by AAm solubility in water (about 50 g/dL), providing up to 30-nm-thick PAAm brush [Fig. 2(a)]. The brush thickness increases nonlinearly with monomer concentration. The analysis with the power function gives an order of 2. This observation sets the “grafting through” apart from the “grafting from” approach. In the latter case, the grafted amount of tethered polymer chains in brush regime depends linearly on the chain length and therefore on the monomer concentration.3 Analysis of the dependency brush thickness–monomer concentration reveals a power function with the power coefficient of about 2. We suggest the following scenario. The increase in chain length of tethered chains caused by increase of monomer concentration results in increased amount of grafted polymer. In the brush regime, such effect has linear character.3 Besides this apparent effect, the C[DOUBLE BOND]C sites on the surface involved in “grafting through” mechanism are more available for longer chains of growing macroradicals. In other words, the longer the macroradicals can grow the more they can reach the surface and react with the anchors. It provides more drastic increase in brush thickness. The same phenomenon may have an opposite effect on thickness versus initiator concentration. It demonstrates an apparent decrease [Fig. 2(b)] when shorter chains have a lesser probability of surface attachment.

Figure 2.

Thickness of PAAm brush as a function of the monomer (a) and the photoinitiator (b) concentration.

Further discussion of the results showed in Figure 2 is limited by the assumption that the thickness of PAAm brush has been developed in each experiment while maintaining standard grafting conditions, that is, PGMA layer morphology and content,10 light intensity, and exposition time. In the case of low monomer and PI concentrations, the brush could be underdeveloped for standard conditions, thus resulting in significantly lower values on Figure 2(a,b). Apparently, kinetics measurement can provide more comprehensive data on the brush formation. Recently, Harris and Metters explored the method of sliding mask for kinetic study of photoinitiated brush synthesis by photoiniferter-mediated “grafting from.”31 Therefore, we designed a computer-controlled sliding screen device shown in Figure 3. It allowed us 100 μm accuracy lateral control and 1-s range exposition precision in fabrication of gradient brush samples for kinetic measurements. Each position on the sample corresponds to certain exposure time, thus giving the kinetics plots [Fig. 3(a)].

Figure 3.

The method of sliding mask (top) and resulting gradient brush at different sliding speeds of 40 μm/s (open circles) and 80 μm/s (black squares) to control the gradient steepness (a). The kinetic profiles obtained by sliding mask method for different concentration of AAm in polymerizate solution: 10% (crosses), 20% (circles), 30% (triangles), and 40% (squares) (b).

Gradient Brushes

A sliding mask technique is a convenient tool to control the gradient steepness. Apparently, the slower the mask speed, the smaller the distance between starting point (no exposition) and a fully developed brush. Figure 3(a) illustrates this concept. The results for gradient samples obtained with different monomer concentration are shown on Figure 3(b). The saturation levels of gradient samples for 10, 20, 30, and 40% of the monomer are in quantitative agreement with homogeneous samples [Fig. 2(a)]. Evidently, chosen exposition time used for homogeneous samples (420–540 s) covers the exposition range to develop the brush thickness.

The photografting reveals three distinctive kinetic regimes [Fig. 3(a,b)]. It starts with an initial period of no grafting, which can be associated with the presence of a small amount of oxygen and other inhibitors. After 2–3 min upon staring illumination, fast accumulation of grafted PAAm occurs. Depending on the monomer and PI concentrations, we observed up to 0.5 nm/s linear grafting rate. In 3–10 min, photografting sharply stops, whereas ungrafted polymer still accumulates. Therefore, we concluded that grafting amount is limited by the surface density of sites available for “grafting through,” that is, C[DOUBLE BOND]C double bonds. This availability can be limited either by the presence of the surface sites or access of the sites to the growing macroradicals. Recently, the latter model has been suggested by Ruhe and coworkers.22 Although it is difficult to further increase the grafting site density, one can easily decrease it in controlled manner. For instance, in a series of samples, we changed the ratio of acrylic and acetic acids for PGMA modification. Assuming equal reactivity of CH3COOH and CH2[DOUBLE BOND]CHCOOH with epoxy groups of PGMA, the ratio CAAc/CHAc is equivalent to the ratio of CH3[BOND] and C[DOUBLE BOND]C[BOND] groups on substrate surface. The brush thickness depends linearly on this ratio (Fig. 4). This observation clearly proves that grafting density is determined by the density of available grafting sites.

Figure 4.

Effect of decreased fraction of acrylic groups (C[DOUBLE BOND]C double bonds) on the brush thickness. The double bond depletion is achieved by adding acetic acid (inactive component) to acrylic acid for PGMA modification.

Patterned Brushes

Photoinitiated version of “grafting through” method also allows straightforward patterning of brushes. To demonstrate applicability of the method, we performed both microscale and macroscale (submillimeter) patterns. The results of micropatterning are shown in Figure 5. The shadow mask of 20-nm Au on glass was deposited using ion-beam lithography. Next, the mask was deposited on top of polymerization solution and exposed to UV light for 600 s. The 3D presentation of the real AFM image of the mask is shown on the scheme together with AFM image and corresponding profiles of the PAAm brush. The brush profiles along and across a shadow beam-like feature reveal high lateral resolution of 1–2 μm. This is the typical resolution limit for optical methods of patterning. The microscopic lateral resolution of the patterned brush in combination with sufficient brush thickness makes this method very promising for many industrial applications.

Figure 5.

Micropatterned PAAm brush. Top: the shadow mask used for this experiment (3D presentation of AFM image of the mask). Bottom: AFM image of the PAAm brush with the corresponding profiles across the beam features (green line) and along the beam (blue line). The profiles reveal both high lateral resolution (about 1–2 μm) and good thickness contrast (0–14 nm).

Cell Adhesion and Cell Viability on Photografted PAAm Brushes

This part of our work is inspired by recent advances in development biononfouling surfaces, that is, surfaces with low or zero adhesion for proteins and cells. To the date, the most common approach for fabrication nonfouling surfaces is modification of glass or plastic substrates with poly(ethylene oxide) (PEO). Brushes of PEO or poly(meth)acrylates with short PEO side chains or sugar substituents are considered to be the most successful candidates.32, 33 We found that PAAm brushes can sufficiently restrain cell adhesion as can be observed by an optical microscope. To investigate the behavior of neuronal cells on PAAm brushes, we seeded mouse N2a cells on 96-well open-bottom plate with different substrates attached, that is, poly-D-lysine, PGMA, PAAm brush, and patterned PAAm brush at a concentration of 20,000 cells per well. The brush samples and the reference samples were attached to the plates' bottom providing specific cell culture substrates. The micrographs of neuronal cells on different substrates after 24-h culture period in growth medium are shown in Figure 6. Both control substrate (poly-D-lysine, not shown) and chemisorbed PGMA layer are adhesive to the cells. The cells are strongly attached to the substrates and form dense layers [Fig. 6(a)]. The cell morphology on PAAm brush is completely different. They form rare large 3D clusters weakly attached to the substrate [Fig. 6(b)], while most of the substrate surface area remains free of the cells. We also cultured the cells onto patterned PAAm brush. The brush was prepared using shadow mask with 125-μm round spots and the spot-to-spot distance of 500 μm (square mesh). The cells occupied shadowed spots without brush (i.e., covered with PGMA) and escaped from surface covered with PAAm brush. Noteworthy, the PAAm brush thickness of 20 nm (dry brush revealed by AFM) was sufficient enough to effectively restrain cell adhesion.

Figure 6.

The culture of mouse N2a cells on a chemisorbed PGMA layer (a), PAAm brush (b), and patterned PAAm brush (c). The pattern parameters are as follows: round shadow spots of 125 μm in diameter and spot-to-spot distance is 500 μm (square mesh).

The viability of neuronal cells culturing on different substrates was evaluated using the standard MTT assay (Fig. 7). MTT (3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyltetrazoliumbromide) is a well-documented and widely used cell viability assay. The MTT substance is incorporated into the cell by endocytosis and is reduced by mitochondrial enzymes in living cells to form a blue-colored formazan precipitate. The absorption of dissolved formazan in the visible region correlates with the number of intact alive cells.28 The results of the MTT assay demonstrate that both PGMA and PAAm brush are noncytotoxic substrates with cell viability compared with the standard cell culturing substrates, such as poly-D-lysine.

Figure 7.

The MTT assay reveals high viability of the mouse N2a cells cultured on different substrates with glass as control. The patterned PAAm surface is the same as used for Figure 6. The bar shows standard deviation of the assay.

CONCLUSIONS

We developed a method for photoinitiated synthesis of PAAm brushes using “grafting through” approach and functional anchoring layer of PGMA. The method allows for fabrication of either homogeneous, or gradient, or micropatterned brushes in reliable and simple manner. The basic parameters of the brushes (thickness, kinetics of formation, and gradient or pattern characteristics) were determined. The brushes combine low cytotoxicity and biononfouling behavior, which makes the method a promising platform for cell adhesion control and spatial localization.

Acknowledgements

The financial support of the Center for Drug Design and Delivery (KISK grant of the Department of Community and Economic Development, Commonwealth of Pennsylvania) is greatly acknowledged. The authors acknowledge helpful discussions with Dr. Nikolai Zhitenev and Dr. Suyong Jung, National Institute of Standards and Technology. T. P. Enright acknowledges financial support of his fellowship at NIST (SURF Award).