Langmuir 2007, 23, 12645-12654
12645
Characterization of Supported Membranes on Topographically Patterned Polymeric Elastomers and Their Applications to Microcontact Printing Annapoorna R. Sapuri-Butti, Ravi Chandra Butti, and Atul N. Parikh* Department of Applied Science, UniVersity of California, DaVis, California 95616 ReceiVed June 27, 2007. In Final Form: September 3, 2007 This article describes the fluorescence microscopy and imaging ellipsometry-based characterization of supported phospholipid bilayer formation on elastomeric substrates and its application in microcontact printing of spatially patterned phospholipid bilayers. Elastomeric stamps, displaying a uniformly spaced array of square wells (20, 50, and 100 µm linear dimensions), are prepared using poly(dimethyl)siloxane from photolithographically derived silicon masters. Exposing elastomeric stamps, following UV/ozone-induced oxidation, to a solution of small unilamellar phospholipid vesicles results in the formation of a 2D contiguous, fluid phospholipid bilayers. The bilayer covers both the elevated and depressed regions of the stamp and exhibits a lateral connectivity allowing molecular transport across the topographic boundaries. Applications of these bilayer-coated elastomeric stamps in microcontact printing of lipid bilayers reveal a fluid-tearing process wherein the bilayer in contact regions selectively transfers with 75-90% efficiency, leaving behind unperturbed patches in the depressed regions of the stamp. Next, using cholera-toxin binding fluid POPC bilayers that have been asymmetrically doped with ganglioside Gm1 ligand in the outer leaflets, we examine whether the microcontact transfer of bilayers results in the inversion of the lipid leaflets. Our results suggest a complex transfer process involving at least partial bilayer reorganization and molecular re-equilibration during (or upon) substrate transfer. Taken together, the study sheds light on the structuring of lipid inks on PDMS elastomers and provides clues regarding the mechanism of bilayer transfer. It further highlights some important differences in stamping fluid bilayers from the more routine applications of stamping in the creation of patterned self-assembled monolayers.
Introduction Solid elastomers, such as those formed by cross-linking of poly(dimethylsiloxanes) (PDMS), represent an attractive class of polymeric materials for the surface immobilization of biomolecules for many reasons.1-3 First, PDMS has a low glasstransition temperature and can be conveniently molded from its liquid precursors. As a result, they are practical substrates for creating surface topographic features and microfluidic platforms. They can be fabricated over a broad range of thicknesses4 and can be designed to embed useful functionalities (e.g., fluorescent dyes, nanoparticles, etc.). Second, the surface chemistry of PDMS can be easily controlled.5,6 The surface of as-prepared, solid PDMS elastomer presents low surface free energy (21-22 D/cm) and thus exhibits high hydrophobicity. However, PDMS surfaces can be readily rendered hydrophilic using simple plasma treatments and UV/ozone treatments.6 Subsequent chemical derivatization and surface metallization can also be used to tune its surface properties.7 Third, PDMS is structurally stable, optically transparent, and chemically homogeneous. Fourth, PDMS surfaces are mechanically deformable and easily pierced. These properties facilitate the transport of materials to their surfaces * Corresponding author. E-mail:
[email protected]. Tel: (530) 7547055. Fax: (530) 752-2444. (1) Whitesides, G. M.; Ostuni, E.; Takayama, S.; Jiang, X. Y.; Ingber, D. E. Annu. ReV. Biomed. Eng. 2001, 3, 335-373. (2) Quist, A. P.; Pavlovic, E.; Oscarsson, S. Anal. Bioanal. Chem. 2005, 381, 591-600. (3) Tanaka, M.; Sackmann, E. Nature 2005, 437, 656-663. (4) Schmid, H.; Michel, B. Macromolecules 2000, 33, 3042-3049. (5) Sharpe, R. B. A.; Burdinski, D.; Huskens, J.; Zandvliet, H. J. W.; Reinhoudt, D. N.; Poelsema, B. J. Am. Chem. Soc. 2005, 127, 10344-10349. (6) Chaudhury, M. K.; Whitesides, G. M. Science 1992, 255, 1230-1232. (7) Schmid, H.; Wolf, H.; Allenspach, R.; Riel, H.; Karg, S.; Michel, B.; Delamarche, E. AdV. Funct. Mater. 2003, 13, 145-153.
and promise to be useful in designing lab-on-chip-type biomimetic devices (e.g., biosensors and microfluids8,9). PDMS elastomers displaying topographic relief patterns are also routinely employed in soft lithography.10,11 In a typical microcontact printing application, topographically patterned PDMS is “inked” with the molecule of interest and brought into contact with a planar surface. The ability of the PDMS surface to make a conformal contact affords site-selective transfer of the molecules to the desired parts of the substrate, ultimately resulting in patterned depositions of molecules. The variety of surface films that have been patterned using the PDMS transfer method includes a wide range encompassing (1) short-chain monolayerforming molecules such as alkanethiols on coinage metals and alkoxysilanes on glass, (2) polymeric and polyelectrolyte molecules, (3) colloidal particles, and (4) many biomolecules. Although most of the structures above were formed under nominally dry conditions, the ability of PDMS stamps to generate reliably patterned relief structures in aqueous environments has also been demonstrated.12 Recently, elastomeric stamps have been used to create patterns of fluid consisting of phospholipid bilayers in aqueous environments.13-15 In one study, Hovis and Boxer14 have shown that oxidized PDMS can be used to selectively blot regions of preformed phospholipid bilayers, which can subsequently be (8) McDonald, J. C.; Whitesides, G. M. Acc. Chem. Res. 2002, 35, 491-499. (9) Delamarche, E.; Bernard, A.; Schmid, H.; Michel, B.; Biebuyck, H. Science 1997, 276, 779-781. (10) Kumar, A.; Abbott, N. L.; Kim, E.; Biebuyck, H. A.; Whitesides, G. M. Acc. Chem. Res. 1995, 28, 219-226. (11) Xia, Y. N.; Whitesides, G. M. Ann. ReV. Mater. Sci. 1998, 28, 153-184. (12) Xia, Y. N.; Whitesides, G. M. J. Am. Chem. Soc. 1995, 117, 3274-3275. (13) Hovis, J. S.;Boxer, S. G. Langmuir 2001, 17, 3400-3405. (14) Hovis, J. S.; Boxer, S. G. Langmuir 2000, 16, 894-897. (15) Jung, S. Y.; Holden, M. A.; Cremer, P. S.; Collier, C. P. ChemPhysChem 2005, 6, 423-426.
10.1021/la701920v CCC: $37.00 © 2007 American Chemical Society Published on Web 11/03/2007
12646 Langmuir, Vol. 23, No. 25, 2007
used to deposit patterns of a bilayer on a clean, planar surface. In another study,13 these authors have shown that the bilayers formed on elastomeric substrates via vesicle fusion techniques can be stamped under water onto clean hydrophilic surfaces, thereby providing spatially patterned bilayers on a variety of hydrophilic substrates. Because these methods are simple to implement and provide one of the few simple means to confine the translational fluidity of phospholipid membranes in predetermined spatial patterns without modifying the substrate surface,16,17 they are gaining considerable attention as a means to design membrane compositional patterns, to compartmentalize membrane fluidity, and in developing parallel or high-throughput assays of ligand-receptor interactions.15,18-22 Despite these achievements and the progress that these studies portend, the structures of lipid bilayers at PDMS surfaces and the mechanism of membrane soft lithography remain incompletely understood. Membrane soft lithography differs from typical microcontact printing applications in many ways. It is now well appreciated that the phospholipid vesicles “ink” the PDMS surface by self-assembling into an oriented phase. Native PDMS surfaces (-O-Si(CH3)2- groups) are hydrophobic and facilitate vesicle fusion to form single phospholipid monolayers. A popular strategy to render PDMS surfaces hydrophilic involves the generation of surface silanol groups (-Si-OH) via plasma oxidation.23 These oxidized PDMS elastomers support the formation of oriented bilayers at their aqueous interfaces.22 The factors that control this preorganization of the lipid ink at the PDMS/water interface and their transfer during the stamping process are not fully understood. For instance, existing reports in the literature do not address how topographic steps on the PDMS surface influence bilayer organization, any lateral reorganization or phase separation of membrane components, and whether a continuous fluid bilayer across these steps is formed. Furthermore, the transfer of the oriented ink may result in the inversion of lipid leaflets: the distal leaflet of the bilayer on PDMS should become the inner leaflet proximal to the substrate if the transfer process does not involve molecular rearrangements. Whether such inversions of leaflets occur during the stamping of asymmetric bilayers is also not established. Using a combination of imaging ellipsometry and fluorescence (epi- and confocal) microscopy measurements, we examine the organization of lipid ink at the PDMS surface and its subsequent transfer during stamping. A unique feature of the present work is our focus on the characterization of bilayers on PDMS prior to patterning and a systematic investigation of any leaflet inversion during microcontact printing. Our results confirm the formation of single fluid phospholipid bilayers on a topographically patterned, UV/ozone-oxidized PDMS surface. We find that facile molecular transport occurs across the topographic boundaries, establishing long-range lateral contiguity. Stamping of these preformed bilayers from the elastomeric surfaces results in the tearing of the bilayer sheets at the topographic edges of the PDMS elastomers. In the manual implementation of stamping (16) Yee, C. K.; Amweg, M. L.; Parikh, A. N. J. Am. Chem. Soc. 2004, 126, 13962-13972. (17) Yee, C. K.; Amweg, M. L.; Parikh, A. N. AdV. Mater. 2004, 16, 1184. (18) Yamazaki, V.; Sirenko, O.; Schafer, R. J.; Nguyen, L.; Gutsmann, T.; Brade, L.; Groves, J. T. BMC Biotechnol. 2005, 5. (19) Burridge, K. A.; Figa, M. A.; Wong, J. Y. Langmuir 2004, 20, 1025210259. (20) Sapuri, A. R.; Baksh, M. M.; Groves, J. T. Langmuir 2003, 19, 16061610. (21) Kung, L. A.; Kam, L.; Hovis, J. S.; Boxer, S. G. Langmuir 2000, 16, 6773-6776. (22) Lenz, P.; Ajo-Franklin, C. M.; Boxer, S. G. Langmuir 2004, 20, 1109211099. (23) Hillborg, H.; Gedde, U. W. IEEE Trans. Dielectr. Electr. Insul. 1999, 6, 703-717.
Sapuri-Butti et al.
that we used, we find an incomplete transfer of the lipid molecules from the elevated regions of the elastomer, which make conformal contact with the support surface. The remainder of the bilayer in the depressed regions of the elastomer retains its lateral fluidity but does not reveal any macroscopic spreading or reorganization. Using a protein-binding assay for an asymmetrically distributed probe lipid, we also examine whether the microcontact transfer of the lipids resulted in the inversion of the lipid leaflets. Our data indicate a more complex transfer process, involving bilayer reorganization and possible re-equilibration of at least some of the molecules during (or upon) transfer, than previously assumed. Materials and Methods Materials. 1-Palmitoyl-2-oleyl-sn-glycero-3-phosphocholine (POPC) and monosialoganglioside (brain, ovine-ammonium salt) (Gm1) were purchased from Avanti Polar Lipids (Birmingham, AL). Texas red 1,2-dihexadecanoyl-sn-glycero-3-phosphatidylethanolamine, triethylammonium salt (TR-DHPE), and BODIPY FL C5ganglioside Gm1 probes were purchased from Molecular Probes (Eugene, OR). The fluorescein isothiocyanate-derivatized B subunit of cholera toxin, FITC-CTB, was purchased from Sigma (St. Louis, MO). All lipids were suspended and stored in chloroform or a chloroform/methanol mixture in the freezer (-20 °C) until use. Hydrogen peroxide (30% v/v) and sulfuric acid were purchased from J. T. Baker (Phillipsburg, NJ) and Fisher Chemicals (Fairlawn, NJ), respectively. All organic solvents were HPLC grade. All chemicals were used without further purification. Organic-free deionized (DI) water of high resistivity (approximately 18.0 mΩ cm) was obtained by processing water through a Millipore water system (model ZD40-11595, Bedford, MA). Phosphate buffer saline (PBS, pH 7.4, 150 mM NaCl, 1.54 mM KH2PO4, and 2.71 mM Na2HPO4) was obtained from Gibco-Life Technology (Rockville, MD). Corning glass coverslips (nos. 11/2 and 2, 22 and 18 mm2, respectively, Fisher HealthCare, Houston, TX) were used as substrates unless noted otherwise. Silicon substrates with native oxide overlayers (Silicon Sense, Nashua, NH) were used for imaging ellipsometry measurements. Methods. Stamp Fabrication. PDMS stamps were prepared by adapting the now well-established replica molding method.10,24 The process begins by first preparing silicon masters displaying desired patterns in a photoresist coating using standard photolithography processes. Briefly, 4 in. silicon wafers were first prebaked on hot plate for ∼10 min at 110 °C and were vapor primed with hexadimethyldisilazane (HMDS). The wafers were then coated with Shipley 1813 positive photoresist by spinning at 3000 rpm to obtain an approximately 1.6-µm-thick layer. The photoresist-coated wafers were vacuum baked on a hot plate for 10 min at 110 °C and allowed to cool under ambient conditions. The masters were then exposed to ultraviolet light (λ ) 385 and 408 nm) through photomasks displaying the desired pattern of chrome on glass. Photomasks were either obtained from commercial sources (Photoscience, Inc. Torrance, CA) or were fabricated in the microfabrication facility of the Northern California Nanotechnology Center at UC Davis. The exposed silicon wafers were then developed in MF-319 for ∼1 min and rinsed in DI water and dried in a stream of N2. The development step dissolved the exposed photoresist. The unexposed photoresist was further stabilized by curing the photoresist pattern on a hot plate for 10 min at 110 °C. The process resulted in relief patterns defined by the masks. Silicon masters so prepared were further primed with HMDS vapors. The latter step renders the master surface hydrophobic, allowing the facile separation of PDMS stamps from the master. Next, we poured a 9:1 mixture of Sylgard 182 (Dow Corning) elastomer and its curing agent onto the silicon masters, followed by curing at 75 °C for 2 h. Cured PDMS stamps were then separated from the silicon masters by gently peeling off. The surfaces of PDMS stamps so derived are hydrophobic. The stamp surfaces were then rendered hydrophilic by exposing the textured stamp surfaces to (24) Xia, Y. N.; Whitesides, G. M. Angew. Chem., Int. Ed. 1998, 37, 551-575.
Membranes on Patterned Polymeric Elastomers
Langmuir, Vol. 23, No. 25, 2007 12647
ozone-generating short-wavelength (184-257 nm) ultraviolet radiation for approximately 90 s. Typically, 1 × 1 or 2 × 2 mm2 PDMS stamps displaying elevated grids and depressed square features were used. The smallest feature size used was 20 µm × 20 µm2 separated by 20 µm edge-to-edge feature separation. Other sizes included square patterns of 50, 100, and 250 µm linear dimensions. Preparation of Vesicles and Glass Substrates. Supported phospholipid bilayers were formed using a previously reported vesicle fusion and rupture method.25,26 Briefly, the desired lipids (in chloroform) were first mixed in predetermined mole ratios in a glass vial. The solvent was subsequently evaporated using a stream of nitrogen. Subsequently, the vial was maintained under house vacuum for ∼45 min to ensure the removal of any trapped or residual solvent, leaving behind a film of lipid at the bottom of the vial. The dried lipid film was then hydrated using DI water and kept overnight at 4 °C. The total lipid concentration in these stock solutions was 2 mg/mL. To prepare small unilamellar vesicles, a portion of the stock solution was sonicated for ∼4 min and extruded through 0.1 µm polycarbonate membrane filters (Whatman, Inc., Newton, MA) using a commercial syringe miniextruder (Avanti Polar Lipids, Alabaster, AL). Sonication and extrusion occurred at temperatures at least 10 °C above the transition temperature of the lipid or the lipid mixtures used. The lipid solution was passed through the extruder at least 19 times. SUVs so formed27 were stored at 4 °C and used within 2 days. Inking and Microcontact Printing. PDMS stamps were UV oxidized for approximately 90 s to render the PDMS surface hydrophilic. Within 15 min of UV-assisted surface oxidation, the PDMS stamps were incubated with a solution of 1:1 lipid vesicles and PBS (150 mM NaCl) for ∼ 2 min. Subsequently, the stamps were rinsed in DI water multiple times to remove any unbound lipids or vesicles from the aqueous environment. PDMS stamps inked in this manner were kept under water at all times. Microcontact printing of bilayers was achieved by contacting the inked stamps with freshly cleaned glass substrates using a light weight (6-13 g depending on pattern geometry) for 15-20 s in water. The stamps were then separated from the glass slide under water, and the slide was stored for further use. Patterns so formed were tested for lateral fluidity by fluorescence recovery after photobleaching (FRAP). Adventitious contaminants28 were first removed from the substrates (silicon oxide wafers or Corning glass coverslips) to be used for microcontact printing by oxidation in a freshly prepared 4:1 (v/v) mixture of sulfuric acid and hydrogen peroxide for a period of 4 to 5 min at a temperature of ∼100 °C (Caution! This mixture reacts Violently with organic materials and must be handled with extreme care.) The substrates were then withdrawn using Teflon tweezers, rinsed immediately with copious amounts of deionized water, and dried in a stream of nitrogen. All cleaned, oxidized substrates were used within several hours of pretreatment. Incorporation of Gm1 and Gm1-Cholera into Toxin-Binding Assay. To determine the distribution of Gm1 in patterned bilayer samples, the Gm1-CTB binding assay was performed. Two different methods were used for the incorporation of Gm1 in the bilayers. First, symmetric mixed vesicles containing 0.5-2 mol % Gm1 in POPC were prepared by drying the two molecular populations together before hydration and extrusion. For each sample composition, two sets of samples were prepared. One set was used to examine the CTB interaction on PDMS stamps whereas the second set was used for microcontact printing on the coverglass. The CTB binding experiments were performed by incubating 20 µL of 0.1 mg/mL FITC-CTB in 1× PBS, making the final concentration of protein 1 × 10-8 M. After approximately 15 min of incubation that ensured near-equilibrium binding, the samples were rinsed several times in PBS. In the second method, two samples of POPC bilayers were first formed on the PDMS surface. These preformed bilayers were
incubated with Gm1 solution (40 µL of a 0.26 mM sol in PBS was added to achieve a Gm1 concentration of 2.5 × 10-6M) for 20 min. These samples were rinsed several times with DI water to remove any excess Gm1 in solution. (Note that this method does not allow us to estimate a priori the exact amount of Gm1 in the bilayer.) One of the two samples is used to stamp the POPC on glass. Subsequently, both of these samples, lipid on a stamp and a stamped pattern, are treated with FITC-CTB as in the first method described above. Epifluorescence Microscopy. A Nikon eclipse TE2000-S inverted fluorescence microscope (Technical Instruments, Burlingame, CA) equipped with an ORCA-ER (model LB10-232, Hamamatsu Corporation, Bridgewater, NJ) or Retiga-1300 CCD camera (Technical Instruments, Burlingame, CA) and a Hg lamp as the light source was used to visualize all fluorescent samples. Two filter wheels, one containing a set of excitation and the other a set of emission filters, were mounted in front of the light source and the CCD camera, respectively. An extra triple-band emitter was installed in the dichroic mirror cube to aid in focusing through the eyepiece. Typically, images taken were using a Plan 10× (NA, 0.25) or a Plan Fluor ELWD 20× (NA 0.45) objective (Nikon, Japan). Highresolution images were obtained using 60× (NA 0.7) objectives. Images were stored and processed using simple PCI software (Compix, Inc., Cranberry Township, PA) augmented with a quantitative dynamic intensity analysis module. Fluorescence images taken with the Texas red filter set were assigned the color red, and the images acquired with the FITC filters were assigned the color green. Excitation and emission maxima for the probes used were 583/601 nm for TR-DHPE and 496/519 nm for FITC-CTB. To characterize the membrane fluidity, a simple method to assess fluorophore mobility within the membrane media was employed. We used microscopy-based fluorescence recovery upon photobleaching (FRAP) measurements by adapting the circular spot photobleaching method.29,30 A circular region of the fluorescent bilayer sample, ∼30-50 µm in diameter, was continuously illuminated at high power at the excitation wavelength for the fluorophore through a Plan Fluor ELWD 60× (NA 0.70) objective for ∼2 min. The exposure bleaches a dark spot on the bilayer by photoexciting the fluorophore, which results in an irreversible chemical transformation effected by a reaction with oxygen dissolved in the ambient buffer. After photobleaching, the illumination path was replaced by a low-power observation beam through a 10× objective to record wide-field images of fluorescence recovery in the bleached area at 30 s intervals. The subsequent lateral motion of unperturbed fluorophore lipids from the unbleached background into the bleached spot (and vice versa) is recorded as a time series of images. It has been previously established that the precise shape of the recovery curve can be used to qualitatively characterize the nature of the fluorophore motion.30 Furthermore, for diffusion-like motions, the time required for the fluorescence intensity to recover to halfway (t1/2) between its immediate post-bleach value and its long-time asymptotic value was used to estimate the diffusion coefficient, D, which is a measure of the fluidity of the lipid environment. Specifically, we approximate the fluorescence recovery process in terms of the evolution of the initial Gaussian profile of the photobleached spot into a modified series of Bessel functions as a solution of the 2D lateral diffusion equation. Here, the experimental fluorescence intensity versus time data is replotted as reduced intensity versus time. The reduced intensity is given by I(red) ) [I(∞) - I(t)]/[I(∞) - I(0)], where I(t), I(o), and I(∞) correspond to fluorescence intensity at time t after photobleaching and immediately after photobleaching and the long-time asymptotic recovery values, respectively. The plot is then used to estimate t1/2, which in turn is used to calculate D ) 0.22ro2/t1/2 where ro refers to the initial size of the photobleached spot. The reported error was determined by considering three sources of systematic uncertainties: (1) The time at which recovery begins is estimated as half of
(25) Tamm, L. K.; McConnell, H. M. Biophys. J. 1985, 47, 105-113. (26) Boxer, S. G. Curr. Opin. Chem. Biol. 2000, 4, 704-709. (27) Mayer, L. D.; Hope, M. J.; Cullis, P. R. Biochim. Biophys. Acta 1986, 858, 161-168. (28) Frantz, P.; Granick, S. Langmuir 1992, 8, 1176-1182.
(29) Koppel, D. E.; Axelrod, D.; Schlessinger, J.; Elson, E. L.; Webb, W. W. Biophys. J. 1976, 16, 1315-1329. (30) Axelrod, D.; Koppel, D. E.; Schlessinger, J.; Elson, E.; Webb, W. W. Mobility Measurement by Analysis of Fluorescence Photobleaching Recovery Kinetics. Biophys. J. 1976, 16, 1055-1069.
12648 Langmuir, Vol. 23, No. 25, 2007 the bleaching time (2-4 min). (2) The starting radius of the spot is estimated to be the difference between the first measured bleached radius and a linear extrapolation back to the radius at the beginning of the bleach (for experiments where the bleaching illumination was not measured). (3) The error in the amplitude of recovery is estimated to be the standard deviation of the Gaussian fit (the square root of diagonal elements of the covariance matrix). This includes the error introduced by approximating the recovering profile as a Gaussian. Because the time required in photobleaching a spot in some instances compounds the effects of diffusion with the initial bleach, care is taken to fit the observed intensity profile directly. In our case, isotropic diffusion retains the Gaussian profile as expected for an initial Gaussian distribution of intensities. The analysis of the mobile fraction (mobility) begins by fitting the recovery amplitude to an offset exponent of the form Y(t) ) F + G exp(-((t - H)/I)). Here, F-I are the constants being fitted, and t is time. The mobility, measured as fractional recovery, is the difference between F (corresponding to Y at infinite time) and the initial post-bleach recovery amplitude divided by the difference between the prebleach value of that region and the initial amplitude. The error in the mobility measurements is estimated as the standard deviation of the fit. The nonlinear least-squares fitting function that we used was based on Levenberg-Marquardt nonlinear regression. Confocal Fluorescence Microscopy. Laser scanning confocal fluorescence microscopy was used to acquire fluorescence emission images with spatial resolution along the substrate normal. We used the Zeiss LSM 510 META spectral imaging system (Carl Zeiss Microimaging Inc., Thornwood, NY) equipped with an AxioCam digital camera and Ar and He-Ne lasers. We used a 63× C-APOCHROMAT (NA 1.2 W Korr) objective to acquire 13 sections forming a z stack traversing the focal plane from above the highest and below the lowest planes of the PDMS stamp. Fluorescence images were taken with a He-Ne laser with a rhodamine filter set (Ex 550/Em 573). Excitation and emission maxima for the probes used were 583/601 for TR-DHPE. Subsequent image restoration and analysis were done using LSM510 VisArt and Physiology software packages (version 3.2; Carl Zeiss Inc., Jena, Germany). Imaging Ellipsometry. Ellipsometric angle measurements31 and spatially resolved ellipsometric contrast images32-34 were acquired using a commercial Elli2000 imaging system (Nanofilm Technologie, Go¨ttingen, Germany). Our ellipsometer employed a frequencydoubled Nd:YAG laser (adjustable power
