Plasma Oxidized Polyhydroxymethylsiloxane—A New Smooth Surface

Feb 19, 2010 - A novel substrate for preparation of supported lipid bilayers (SLBs), smooth at the subnanometer scale and of variable thickness from t...
0 downloads 0 Views 3MB Size
pubs.acs.org/Langmuir © 2010 American Chemical Society

Plasma Oxidized Polyhydroxymethylsiloxane;A New Smooth Surface for Supported Lipid Bilayer Formation C. Satriano,*,† M. Edvardsson,‡ G. Ohlsson,‡ G. Wang, S. Svedhem,‡ and B. Kasemo*,‡ †

Department of Chemical Sciences, Catania University, Viale A. Doria, 6, 95125 Catania, Italy, and Department of Applied Physics, Chalmers University of Technology, SE-412 96 G€ oteborg, Sweden



Received October 9, 2009. Revised Manuscript Received January 31, 2010 A novel substrate for preparation of supported lipid bilayers (SLBs), smooth at the subnanometer scale and of variable thickness from ten to several hundred nanometers, was developed by surface oxidation of spin-coated poly(hydroxymethylsiloxane) (PHMS) films. The deposited polymeric thin films were modified by a combination of oxygen plasma and thermal treatment (PHMSox), in order to convert the outermost surface layer of the polymer film to a stable SiO2 film, suitable for SLB formation. The hydrophilic, SiO2-like surfaces were characterized by XPS, wetting angle, ellipsometry, and AFM. Lipid bilayers were formed on this surface using the well-known vesicle adsorptionrupture-fusion process, usually performed on glass or vapor-deposited SiO2. Reproducible formation of homogeneous SLBs of different compositions (POPC, DOEPC, and POPC/DOPS) was demonstrated on the new SiO2 surface by quartz crystal microbalance with dissipation (QCM-D), surface plasmon resonance (SPR), and optical reflectometry measurements. The SLB formation kinetics on the PHMSox-coated sensors showed very similar characteristics, for all investigated PHMS thicknesses, as on reference sensors coated with vapor-deposited SiO2. The good adhesive properties of the PHMS to gold allows for the preparation of thin PHMSox layers compatible with SPR. The much smaller roughness at the nanometer scale of the PHMSox surfaces, compared to standard vapor-deposited SiO2-coated sensors, makes them advantageous for AFM and optical experiments and promising for patterning. To benefit optical experiments with the PHMSox surfaces, it was also investigated how the PHMS film thickness influences the SPR and reflectometry responses upon SLB formation.

1. Introduction The preparation of supported phospholipid bilayers (SLBs) on different kinds of supports (often inorganic surfaces1 or polymer cushions2) has attracted great interest in recent years, both for fundamental research3,4 and for applications in biotechnologies, biophysics and molecular biology,5-7 cell mimicking,4,8,9 in the design of biosensors,10,11 and for the attachment and cultivation of cells.12,13 A common way to prepare SLBs is the vesicle adsorptionrupture-fusion method, where lipid vesicles in the liquid bulk phase, tens to a few hundred nanometers in size, are allowed to adsorb onto a substrate, where they rupture after adsorption and fuse to a coherent bilayer.14-16 Most work to date has focused on *Corresponding authors. [email protected]. Phone: (þ39) 095 7385136. Fax: (þ39) 095 580138. [email protected]. Phone: (þ46) 31 7723370. Fax: (þ46) 31 772 3134. (1) Richter, R. P.; Brat, R.; Brisson, A. R. Langmuir 2006, 22, 3497. (2) Sackmann, E.; Tanaka, M. Trends Biotechnol. 2000, 18, 58. (3) Sackmann, E. Science 1996, 271, 43–48. (4) Tanaka, M.; Sackmann, E. Nature 2005, 437, 656. (5) Reimhult, E.; Kumar, K. Trends Biotechnol. 2008, 26, 82. (6) Anrather, D.; Smetazko, M.; Saba, M.; Alguel, Y.; Schalkhammer, T. J. Nanosci. Nanotechnol. 2004, 4, 1. (7) Zhou, X.; Moran-Mirabal, J. M.; Craighead, H. G.; Mceuen, P. L. Nat. Nanotechnol. 2007, 2, 185. (8) Deng, Y.; Wang, Y.; Holtz, B.; Li, J.; Traaseth, N.; Veglia, G.; Stottrup, B. J.; Elde, R.; Pei, D.; Guo, A.; Zhu, X.-Y. J. Am. Chem. Soc. 2008, 130, 6267. (9) Hone, J.; Kam, L. Nat. Nanotechnol. 2007, 2, 140. (10) Yang, T.; Jung, S.-Y.; Mao, H.; Cremer, P. S. Anal. Chem. 2001, 73, 165. (11) Johnson, J. M.; Ha, T.; Chu, S.; Boxer, S. G. Biophys. J. 2002, 83, 3371. (12) Thid, D.; Bally, M.; Holm, K.; Chessari, S.; Tosatti, S.; Textor, M.; Gold, J. Langmuir 2007, 23, 11693. (13) Oliver, A. E.; Ngassam, V.; Dang, P.; Sanii, B.; Wu, H.; Yee, C. K.; Yeh, Y.; Parikh, A. N. Langmuir 2009, 25, 6992. (14) Nollert, P.; Kiefer, H.; Jahnig, F. Biophys. J. 1995, 69, 1447. (15) Jonsson, M. P.; J€onsson, P.; H€oo€k, F. Anal. Chem. 2008, 80, 7988. (16) Keller, C. A.; Kasemo, B. Biophys. J. 1998, 75, 1397.

Langmuir 2010, 26(8), 5715–5725

silicon dioxide (SiO2), glass (essentially SiO2), mica, and titania. All these surfaces exhibit a strong (SiO2) or moderate (TiO2) degree of vesicle-surface interaction and hydrophilicity. Other than the hydrophilicity, the lipid bilayer formation is strongly dependent on the precise surface chemistry of the support, and its specific interaction with the lipid membrane.1,17,18 The detailed kinetics of the SLB formation process depends on a combination of several interactions, where the strength of the vesicle-surface interaction is of prime importance and where also the ability of the vesicle to deform (related to the vesicle-surface interaction) and, in some cases, the vesicle-vesicle interaction on the surface play a role (for a recent overview, we refer to Dimitrievski and Kasemo,17 and refs therein). Important factors are the substrate surface chemistry,18,19 the lipid vesicle composition and size,1,17,20 pH, temperature, osmotic pressure,20-22 ionic strength, and the presence of specific ions such as mono- or divalent cations.23 A key technique to characterize the SLB formation from vesicles on solid supports has been the quartz crystal microbalance (17) Dimitrievski, K.; Kasemo, B. Langmuir 2009, 25, 8865. (18) Rossetti, F. F.; Bally, M.; Michel, R.; Textor, M.; Reviakine, I. Langmuir 2005, 21, 6443. (19) Keller, C. A.; Glasm€astar, K.; Zhdanov, V. P.; Kasemo, B. Phys. Rev. Lett. 2000, 84, 5443. (20) Reimhult, E.; H€oo€k, F.; Kasemo, B. Langmuir 2003, 19, 1681. (21) Seantier, B.; Breffa, C.; Felix, O.; Decher, G. J. Phys. Chem. B 2005, 109, 21755. (22) Giger, K.; Lamberson, E. R.; Hovis, J. S. Langmuir 2009, 25, 71–74. (23) Seantier, B.; Kasemo, B. Langmuir 2009, 25, 5767. (24) Munro, J. C.; Frank, C. W. Polymer 2003, 44, 6335. (25) Reimhult, E.; Larsson, C.; Kasemo, B.; H€oo€k, F. Anal. Chem. 2004, 76, 7211. (26) Richter, R. P.; Brisson, A. R. Biophys. J. 2005, 88, 3422. (27) Viitala, T.; Hautala, Y. T.; Vuorinen, J.; Wiedmer, S. K. Langmuir 2007, 23, 609.

Published on Web 02/19/2010

DOI: 10.1021/la903826d

5715

Article

with dissipation monitoring (QCM-D)19,21-30 due to its unique ability to distinguish between intact vesicle adsorption and SLB formation on the surface (see further below). Complementary information to the QCM-D results has in various studies been obtained by optical and electrochemical methods25,29 as well as by atomic force microscopy (AFM)1,31-33 and Monte Carlo simulations.17 The optical techniques used are mainly surface plasmon resonance (SPR),25 ellipsometry,1 fluorescence microscopy,28,33 reflectrometry,29 and optical waveguide light mode spectroscopy.34 Bulk substrates like glass or oxidized Si wafer surfaces are conveniently used for microscopy and ellipsometry. Several methods require thin film samples, like SPR (ca. 1 h), followed by rinsing with water, drying with nitrogen, and UVozone treatment (30 min). The measurements were carried out at 22 C. After stabilization of the baseline, 0.5 mL of a temperaturestabilized vesicle solution (100 μg/mL) was injected over the sensor surface. Additional QCM-D experiments were performed to probe the amount of hydroxyl (-OH) groups on the different surfaces by using their reactivity with the silane APTES. These experiments consisted of 30 min exposure of the different surfaces to APTES in aqueous solution (1% v/v), followed by multiple rinsing steps with water. Frequency and dissipation shifts were measured at the third overtone. Frequency shifts were normalized to the fundamental frequency of the sensor (5 MHz) by division by 3. 2.6.2. Surface Plasmon Resonance (SPR). The SPR instrument used in this study is designed in the Krestchmann configuration. A laser beam is applied, through a prism, to a thin Au film, such that a surface plasmon is excited and decays exponentially into the bulk liquid at the other side of the metal film. The characteristic decay length, ldecay, of the evanescent field is 25-50% of the wavelength of the applied light, and depends as well on the dielectric properties of the medium near the metal surface and the metal itself:54,55 ldecay

λ ¼ 2π

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ε þ εmetal ε2

ð4Þ

where λ is the wavelength of the incident light, ε (= n2) is the dielectric constant of the medium in the sensed volume (which in our study includes the PHMS layer precoated on the gold surface as well as lipid mass adsorbed on top of the PHMS and the buffer), and εmetal is the real part of the complex dielectric function of the metal (which is dependent on wavelength, film thickness, and method of deposition56). The SPR response, i.e., the angle at which the surface plasmon arises, depends on the effective refractive index near the surface, neff, and decays from the surface with a characteristic sensing depth being half that of the decay length of the evanescent field, ldecay.54,57 The effective refractive index, neff, is the refractive index integrated over the entire volume above the Au-surface,54 according to neff ¼

2 ldecay

Z

¥

0

nðzÞ 3 e -2z=ldecay dz

ð5Þ

where n(z) is the refractive index as a function of the distance z above the surface. Refractive index changes caused by, e.g., adsorption of biomolecules at the surface, will be detectable via monitoring of shifts of the plasmon angle, ΔΘ, and the response will be proportional to the difference between the effective refractive indices at the start of the measurement, generally that of bulk liquid, and neff as the measurement (54) Jung, L. S.; Campbell, C. T.; Chinowsky, T. M.; Mar, M. N.; Yee, S. S. Langmuir 1998, 14, 5636. (55) Schasfoort, R. B. M.; Tudos, A. J. Handbook of Surface Plasmon Resonance; Royal Society of Chemistry, 2008. (56) Stenberg, E.; Persson, P.; Roos, H.; Urbaniczky, C. J. Colloid Interface Sci. 1991, 143, 513. (57) Liedberg, B.; Lundstr€om, I.; Stenberg, E. Sens. Actuators B 1993, 11, 63.

5718 DOI: 10.1021/la903826d

proceeds:54 ΔΘ ¼ mðneff -n0eff Þ ¼ mðn -n0 Þ 3 ð1 -e -2d=ldecay Þ

ð6Þ

where the 0 denotes the initial condition, that is, the bare surface in contact with ambient medium (under the present conditions buffer solution), n and d are the refractive index and the thickness, respectively, of an adsorbed layer, and m is the sensitivity factor of the instrument in response units per refractive index change. The latter is preferably calibrated by performing a series of measurements on solvents with known refractive indices. If the metal surface is precoated at the start of the measurement (in the present study by a PHMS layer), onto which, e.g., a biomolecular film is adsorbed (here, a lipid bilayer), eq 6 will be tranformed into ΔΘ ¼ mðnbilayer -n0 Þ 3 ð1 -e -2dbilayer =ldecay Þ 3 e -2dPHMS =ldecay

ð7Þ

where the final exponential, as well as ldecay, takes into account a PHMS coating of thickness dPHMS and refractive index nPHMS, and where dbilayer and nbilayer refer to the thickness and the refractive index, respectively, of a lipid bilayer on top of the PHMS layer. To convert the experimental data into mass, the thickness of the adsorbed layer, dbilayer, is extracted from eq 7, and then multiplied with the appropriate density, Fbilayer. The SPR measurements were performed using a BIAcore 2000 instrument (Biacore AB, Uppsala, Sweden) with PHMSox coated Au sensor chips (section 2.3). The setup operates with a laser of wavelength λ = 760 nm, and includes automated handling of the buffer, vesicle, and protein solutions, flow rate, and temperature during measurement. The running buffer was PBS. Upon mounting of the sensor chip, four measurement cells are formed, and SPR responses can be recorded from four spots on the chip, one in each of these four chambers. All of the data presented here were recorded at a flow rate of 5 μL/min and in multichannel mode. In the SPR instrument used in this work, the plasmon angle shifts, ΔΘ, are reported in resonance units, RU (1000 RU = 0.1 deg).55 The range of operation is 1.33 < n < 1.36, and neff > 1.36 will not be measurable.55 The theoretical prediction of the SPR responses was made using eq 7, where nbilayer = 1.48 and dbilayer = 5 nm,29 n0 = 1.33 (water), and dPHMS = 12, 25, 70, 180, and 280 nm. The sensitivity of the system, m, was calculated to 9.6(( 0.4)  105 RU/RIU using a known experimental response of bilayer formation in RU, divided by the theoretical change in effective refractive index, Δneff, when going from bare surface to surface with bilayer as calculated by eq 5. The decay length of the evanescent field in buffer, ldecay, was modeled for each PHMS thickness according to eq 4, with λ = 760 nm, n0 = 1.33 (water), nPHMS = 1.4,58 and εAu = -24.3,56 first calculating the decay in the PHMS layer of thickness dPHMS, and then calculating the height, h, above the PHMS surface where the evanescent field strength has reached e-1 ≈ 37%, finally letting ldecay = dPHMS þ h. Furthermore, neff was calculated for a bilayer as a function of PHMS thickness to determine the thickness where neff > 1.36, where the allowed Biacore refractive index interval would be exceeded. 2.6.3. Reflectometry. Briefly, in optical reflectometry, a beam of monochromatic light, in our case emitted by a laser diode at λ = 635 nm, is polarized by a linear polarizer and directed to the solid surface via a right-angle prism. The two outgoing beams are separated into p- and s-polarized light with a cubic polarizing beam splitter. The intensities Ip and Is of the reflected beams are monitored by four photodiodes. The angle of incidence of the probing beam is fixed. The measured signal is the intensity ratio S = Ip/Is, which changes upon adsorption to the surface. The optical response, ΔR, corresponds to relative changes of this (58) Hirayama, M. K. N.; Caseri, W. R.; Suter, U. W. Appl. Surf. Sci. 1999, 143, 256. (59) Azzam, R. M. A.; Bashara, N. M. Ellipsometry and Polarized Light; North Holland Publishing Company: Amsterdam, 1987.

Langmuir 2010, 26(8), 5715–5725

Satriano et al.

Article

intensity ratio and is related to the amount of material adsorbed to the surface, according to the following equations:29 ΔR ¼ ðS -S0 Þ=S0

ð8Þ

ΔR ¼ d 3 ðn -n0 Þ 3 A

ð9Þ

where the subscript 0 denotes the initial condition, that is, bare surface in ambient medium, A is the sensitivity factor, and d and n represent the thickness and the refractive index of the adsorbed layer. The sensitivity factor was obtained by using a matrix method,57 based on optical models which represent the structure of the surface and the specific measurement conditions (e.g., the angle of incidence and the wavelength of the probing beam). The adsorbed mass, that is, the optical mass mreflect, was obtained by using De Feijter0 s formula mreflect ¼ d 3 ðn -n0 Þ=ðdn=dcÞ ¼ ðΔR=AÞ 3 ½1=ðdn=dcÞ

ð10Þ

where dn/dc is the refractive index increment for increasing bulk concentrations of the adsorbed material (dn/dc = 0.169 mL/g in the case of lipids.29,60 The combined QCM-D/reflectometry measurements were performed with a new experimental setup provided by Q-Sense AB, G€ oteborg, Sweden. Note that the currently used setup differs from a conventional one in that it uses two additional light beams, reflected from the prism surface facing the liquid, and two additional photodiodes to detect these beams. This arrangement is used in order to correct for possible interference and errors from matter adsorbing on the prism surface from the bulk solution. The details of this setup, allowing simultaneous QCM-D and reflectometry measurements on one and the same surface, were published recently.29,61 Experimental conditions were as described above for QCM-D, except that the flow rate was increased to a flow rate of 100 μL/min.

3. Results and Discussion First, we describe the characterization of the PHMS and the PHMSox films with respect to thickness (by ellipsometry), surface free energy (by water contact angle measurements), chemical composition (by XPS and through PHMSox reactivity to silanes), and surface topography (by AFM). Note that two types of oxidized PHMS layers were used. They are both referred to as PHMSox, independent of whether a single plasma oxidation step (PHMSp) or a three-step plasma oxidation f (thermal) curing f plasma oxidation sequence (PHMSpcp) was used. In the second part of this section, the performance of the PHMSox films with respect to SLB formation will be described based on the results from the three employed surface sensitive techniques, the electroacoustic technique QCM-D and the two optical techniques SPR and reflectometry. 3.1. Physicochemical Properties of Oxidized PHMS Surfaces. 3.1.1. PHMS Film Thickness after Spin-Coating. To obtain good control over the spin-coating procedure, different solvents and PHMS concentrations were tried. In Figure 1, the thicknesses, measured by ellipsometry, of one series of PHMS films on Si-wafers are shown, before and after plasma oxidation. For the thickness determination, we used a fixed refractive index of n = 1.398 þ 0.000 72/λ2, where λ is the wavelength in micrometers (μm). This refractive index yielded a good fit to all the layers with different thicknesses in ellipsometric modeling. (60) Salamon, Z.; Tollin, G. Biophys. J. 2001, 80, 1557. (61) Wang, G.; Rodahl, M.; Edvarsson, M.; Svedhem, S.; Ohlsson, G.; H€oo€k, F.; Kasemo, B. Rev. Sci. Instrum. 2008, 79, 0751071.

Langmuir 2010, 26(8), 5715–5725

Figure 1. PHMS films thickness versus bulk solution concentration for PHMS untreated (unt) spin coated;either on silicon or titanium oxide;as well as for oxidized films, both PHMSp (p) and PHMSpcp (pcp). C0 is the concentration of the as-received PHMS solution.

The thickness varied from almost 600 nm, for an undiluted PHMS solution, down to one or a few tens of nanometers after dilution of the original polymer solution before spin-coating. No significant differences were measured for PHMS films spincoated on Si or TiO2 substrates. For the highest dilutions, i.e., for the thinnest films (ca. 10-200 nm), a linear relationship was obtained between film thickness and PHMS concentration in the solution (see Figure 1). Thereafter, the thickness increases more slowly with increasing concentration. Both oxidative treatments described above produce the same final film thickness for both PHMSp and PHMSpcp, which shows that the oxygen radiofrequency plasma treatments have negligible etching effects on the polymer film thickness41 for the used plasma conditions. 3.1.2. Wetting Angle. Since the success of SLB formation on a certain substrate is significantly dependent on the exact surface properties, including the degree of contamination, it can often, for SiO2 surfaces, be predicted by measuring of the surface wettability.1,18,33,42 For the present oxidized PHMS surface, one would expect that incomplete oxidation of the surface would result in higher contact angles than for complete oxidation to SiO2, because incomplete oxidation would leave residual hydrophobic methyl groups at the surface. Measured water contact angle (WCA) values and the corresponding surface free energy (γ) terms for the different PHMS surfaces are shown in Figure 2, as measured before and after the UV-O3 cleaning step (see section 2.4). In particular, Figure 2a shows that the untreated PHMS surfaces are hydrophobic, which is due to the predominant exposure of methyl groups at the air interface.62,63 For the surface oxidized films, it is evident that both oxidation procedures (yielding PHMSp or PHMSpcp) produce hydrophilic surfaces. However, PHMSpcp surfaces are a little less hydrophilic than PHMSp surfaces, likely due to surface contamination during thermal curing. Upon storage, the PHMSox layers again become less hydrophilic (Figure S1 in Supporting Information), i.e., they “age” (62) Satriano, C.; Carnazza, S.; Guglielmino, S.; Marletta, G. Langmuir 2002, 18, 9469. (63) Licciardello, A.; Satriano, C.; Marletta, G. In Secondary Ion Mass Spectrometry SIMS XII; Benninghoven, A., Bertrand, P., Migeon, H. N., Eds.; Elsevier, 2000; p 889.

DOI: 10.1021/la903826d

5719

Article

Satriano et al. Table 1. Average Atomic Composition and Atomic Ratios from XPS Analysis, as Measured after the UV-O3 Cleaning Step of the Various PHMS Surfaces as well as Reference SiO2-Coated Sensorsa C 1s (at. %) O 1s (at. %) Si 2p (at. %) O/Si C/Si PHMS

7.7 59.9 (19.5) (49.2) PHMSpcp 10.2 58.4 (18.9) (51.4) PHMSp 8.7 59.3 (16.6) (52.8) 7.4 58.9 SiO2 standard a In the bracket, the corresponding values surfaces, before the UV-O3 treatment.

32.5 1.8 (31.3) (1.6) 31.5 1.8 (29.8) (1.7) 32.0 1.8 (30.6) (1.7) 33.7 1.8 for modified and

0.2 (0.6) 0.3 (0.6) 0.3 (0.5) 0.2 aged

Scheme 1. Schematic Representation of the Silane Reaction with the as-Deposited (PHMS) (a) or Oxidized (PHMSox) (b) Surfaces

Figure 2. Water contact angle (WCA) and surface free energy (SFE) for PHMS untreated (unt), PHMS oxidized, both PHMSp (p) and PHMSpcp (pcp), and SiO2 films, as measured before and after the UV-O3 cleaning (-c) step. (a) Advancing and static WCAs. (b) Total SFE (TOT), Lifshitz van der Waals (LW) and acid-base (AB) components. (c) Lewis base electron donor (-) and Lewis acid electron-acceptor (þ) components.

slowly, either due to surface contamination and/or due to reorganization of methyl groups in the uppermost part of the surface layer. A UV-ozone exposure of the surface transforms both PHMSp and PHMSpcp back to a very hydrophilic state (WCA < 5), even more hydrophilic than standard PVD-coated SiO2 samples (WCA ∼ 10) (plasma oxidation could also be used for this step). For PHMSp and PHMSpcp surfaces that are not used directly after preparation, either a UV-ozone exposure or a short plasma oxidation step is therefore recommended. The PVDcoated surfaces are significantly rougher than the PHMSox ones (see below). To further analyze the surface free energy components, contact angles using several different solvents were measured. These results are displayed in Figure 2b, which shows very comparable trends for the UV-ozone treated surfaces of PHMSp, PHMSpcp, and standard SiO2. In fact, all these surfaces have a predominant acid-base character, i.e., about twice the value of the corresponding dispersive Lifshitz van der Waals component. This finding suggests similar surface termination of oxidized PHMS and SiO2 surfaces, mainly consisting of Si-OH groups. 3.1.3. XPS Measurements and OH Titration by APTES. Another valuable characterization tool is XPS, since it can reveal the chemical state of the SiOx surface. Particularly important in the present context is that XPS can tell if the surface is completely (x = 2) or incompletely oxidized (x < 2), and also can reveal the presence of impurities, contaminants either from the environment or diffusing to the surface from the underlying PHMS layer. The XPS analysis of the PHMS surface chemical composition (see Table 1) confirms the picture proposed from the contact angle data. In particular, for both the oxidized surfaces, PHMSp and PHMSpcp, the carbon atomic percentage, which is mostly related to methyl groups and background hydrocarbon contamination, 5720 DOI: 10.1021/la903826d

drops down to about one-half of that for the untreated surface, i.e., to about 10%. This finding indicates the creation of a SiO2-ε(Cε) surface layer, with a very low (ε f 0) carbon content.41 However, the XPS technique has, at the used experimental conditions, not enough resolution (of the O1s peak) to spectroscopically quantify the increase of hydroxyl groups at the surface of oxidized PHMS when compared with untreated PHMS. To get more information on this issue, the ability of the surfaces to react with silanes was investigated by monitoring the reaction between the cationic silane APTES and -OH sites on the surfaces (Scheme 1).64 QCM-D measurements of the amount of irreversibly bound APTES, after APTES exposure and rinsing, yielded Sauerbrey masses of about 50 ng/cm2 on untreated PHMS, and about 120 ng/cm2 on PHMSp, corresponding to an average thickness for the APTES coverage of 1.2 nm on oxidized PHMSp and of 0.5 nm on unmodified PHMS. This is in line with an increased number of -OH groups on the surface after oxidation. It is of note that the thickness of such oxidized surface films (i.e., the SiO2 like film) on the PHMS coating, after plasma oxidation, is about 200 nm;41 thus, for almost all the PHMS thicknesses investigated here, the complete conversion of the polymer coating to a homogeneous SiO2-like layer is obtained. 3.1.4. AFM and Surface Roughness. Roughness and morphology of the native and oxidized PHMS surfaces at the nanometer scale were investigated by AFM. The surface topographies of differently coated QCM-D sensors is shown in Figure 3: first, as a reference, (a) a gold coated sensor, and (b) a standard PVD SiO2 coated sensor, and then sensors with (c) unmodified and (d) oxidized PHMS-coatings. (64) Demirel, G.; C-aykara, T.; Akaoglu, B.; C-akmak, M. Surf. Sci. 2007, 601, 4563.

Langmuir 2010, 26(8), 5715–5725

Satriano et al.

Article

Figure 4. QCM-D shifts of frequency (left-hand side axis, solid symbols) and dissipation (right-hand side axis, open symbols) for the third overtone (n = 3) during the exposure of vesicles of (a) POPC, (b) DOEPC, and (c) POPC/DOPS to standard SiO2 (square symbols), PHMSp (circle symbols), and PHMSpcp (triangle symbols) surfaces. The two arrows indicate time points for the exchange of liquid in the cell to vesicle solution (left arrow) and PBS solution, respectively. Figure 3. Tapping mode AFM images in air of (a) a gold electrode on a QCM-D sensor, (b) standard SiO2 PVD-coated sensor; (c) as deposited PHMS on a QCM-D sensor, and (d) same as (c) but after oxidation, heat treatment, and new oxidation, i.e., PHMSpcp. Note the much higher smoothness in (d) and (c) compared to (a) and (b).

Two main findings are evident from the height and section images: (i) the spin-coated PHMS films (Figure 3c, PHMS, Rq ∼ 0.3 nm; Figure 3d, PHMSox, Rq ∼ 0.4 nm), ∼30-nm-thick, are much smoother than both the unmodified Au surface (Figure 3a, Rq ∼ 1.6 nm) and the PVD-coated SiO2 surface (Figure 3b, Rq ∼ 1.3 nm), (ii) the surface modification treatments by combined thermal treatment and plasma oxidation (PHMSpcp), including additional exposures to UV-ozone (Figure 3d), do not significantly affect the smoothness of the spin-coated PHMS film, as already observed for pure oxygen plasma treatment of PHMS (PHMSp).41 In other words, the spin-coated PHMS films are very smooth and the spin-coating by PHMS of initially rough surfaces tend to smooth the latter. 3.2. Lipid Bilayer Formation on PHMSox Surfaces. With the physicochemical characterization described above as a basis, lipid vesicle adsorption and lipid bilayer formation were investigated and compared for the PHMSox and SiO2 reference surfaces, using QCM-D, SPR, and reflectometry. QCM-D and the optical techniques make use of different sensing principles and thus provide different and complementary information about the kinetics and lipid structures adsorbed on the surface. The optical data will be described and discussed separately (see section 3.3), where it is also shown how the design of the sensor film is more critical, for quantification and sensitivity, for the optical techniques than for QCM-D. QCM-D is a well-established technique to characterize SLBs, and the performance of such experiments on the PHMSox Langmuir 2010, 26(8), 5715–5725

surfaces allowed for the evaluation of how well the PHMSox layers behave compared to the PVD-deposited SiO2 layers. For the present purpose, one unique aspect of the QCM-D technique is particularly important: as shown previously, the technique is extremely valuable in the context of SLB formation from vesicles through its ability to distinguish between intact lipid vesicles and lipid bilayers attached to a surface.19 This originates from the fact that lipid vesicles contain considerable amounts of trapped water, which (i) is measured as part of the attached mass (via Δf) and (ii) gives rise to a high D because adsorbed vesicles form a viscous, quite soft, and dissipative adlayer. In contrast, planar lipid bilayers yield very small dissipation and the frequency shift is essentially caused by, and proportional to, the attached lipid mass. Consequently, a larger than normal frequency shift and dissipation shift at the end of the SLB formation signals incomplete conversion of vesicles to bilayer, i.e., lower quality of the SLB.20 Three different kinds of vesicles were used for the SLB experiments, to show that bilayers form equally well from differently charged vesicles. The used vesicles were POPC vesicles, which are neutral (slightly negative), DOEPC vesicles, which are positively charged, and POPC/DOPS (4:1) vesicles, which are negatively charged. These vesicles form, after adsorption on the surface, SLBs via different pathways. The lipid bilayer formation experiments investigated by QCM-D were performed with all three types of vesicles both on standard SiO2-coated and on PHMSoxcoated sensors. Two series of measurements were performed on both freshly prepared surfaces and after restoration, immediately before use, of the highly hydrophilic character (i.e., WCA < 10) by UV-ozone treatment. Figure 4 shows QCM-D curves (frequency and dissipation shifts) recorded when the different types of vesicles were exposed DOI: 10.1021/la903826d

5721

Article

to the three different oxide surfaces: PHMSp, PHMSpcp, and, for comparison, standard, PVD-deposited, SiO2 surfaces. It is evident that oxidized PHMS surfaces, both PHMSp and PHMSpcp, behave very similarly to the standard SiO2-coated crystals for SLB formation. For the POPC (Figure 4a) and the POPC/DOPS (4:1) (Figure 4c) vesicle solutions, the characteristically peaked curves of frequency and dissipation shifts (Δf and ΔD) are clearly evident. During the first 1.5-3 min, vesicles adsorb until the minimum in frequency shift and the maximum in dissipation shift, corresponding to a critical coverage of intact vesicles on the surface, have been reached, where vesicles begin to rupture and autocatalytically induce bilayer formation.16,19,65 Then, as more vesicles arrive to the surface, vesicles continue to adsorb on empty surface and rupture and fusion continues for about 3-4 min. Finally, there is the saturation region, which corresponds to a coherent (saturated) bilayer over the whole surface, with frequency and dissipation shifts of Δf = -26.6 ((0.2) Hz and ΔD = 0.4 ((0.1)  10-6, for POPC, and Δf = -28 ((2) Hz and ΔD = 0.5 ((0.2)  10-6, for POPC/DOPS, respectively, as obtained after PBS rinsing. These values indicate the formation of a rigid and stable lipid bilayer, in agreement with literature data.19,29 For the positively charged DOEPC (Figure 4b), the vesicles break as soon as they make contact with the substrates, due to stronger (electrostatic) surface interaction in this case,17 and a lipid bilayer is completed within 5 min after the beginning of the adsorption process. The values measured after the PBS rinsing step are Δf = -23.0 ((1.0) Hz and ΔD = þ0.2 ((0.1)  10-6, respectively. The slightly lower value of the final frequency shift compared to the value for SLBs formed from zwitterionic vesicles is in agreement with literature data.66 Note that the time values quoted above are for the specific vesicle concentration and sizes used in the present case. The time scale for the experiment is close to inversely proportional to the vesicle concentration.16 There are some small differences in the POPC bilayer formation kinetics, between PHMSox and PVD-deposited SiO2 in Figure 4a in that the completion of the bilayer formation is somewhat faster on the PHMSox surfaces. This and the shallower minimum in Δf and the lower peak maximum in ΔD could signal a somewhat stronger surface interaction on the latter surfaces. However, the statistics are not enough to draw conclusions about these small differences, and these details are beyond the scope of this work. Polymer oxidation performed by simple thermal treatments followed by further treatment by UV-ozone exposure, i.e., without any plasma oxidation step, also allows bilayer formation (see Figure S2 in Supporting Information). Although this is an alternative preparation path of the SLB forming surface, it does not give quite as good results as those discussed above in terms of reproducibility and quality of the bilayer formation. However, this path could probably be optimized further and is of interest for laboratories lacking plasma treatment equipment. The storage/aging properties of the PHMSox surfaces are of interest for, e.g., sensor applications, but also for basic work. In a previous study, we have shown that storage of PHMSox films for two weeks, either in water or in air, produced surfaces with distinctly different, but in both cases stable, wetting angles of 0-5 and 20, respectively.41 In the present study, we checked the storage properties of PHMSox-coated sensors, with respect to (65) Dimitrievski, K.; Reimhult, E.; Kasemo, B.; Zhdanov, V. P. Colloids Surf. B: Biointerfaces 2004, 39, 77. (66) Wikstr€om, A.; Svedhem, S.; Sivignon, M.; Kasemo, B. J. Phys. Chem. B 2008, 112, 14069.

5722 DOI: 10.1021/la903826d

Satriano et al.

wetting angle and SLB formation, by using PHMSox surfaces aged in air for up to several months, and then treated by UVozone just before the SLB formation experiments. Hydrophobic recovery was unavoidable upon storage in air (WCA was rising up to ∼60 for PHMSp and ∼35 for PHMSpcp), but wettability measurements demonstrated that the highly hydrophilic character of freshly prepared PHMSox surfaces, i.e., WCA