Near-Surface Structure of Plasma Polymer Films Affects Surface Behavior in Water and its Interaction with Proteins
- 197 Downloads
Using low pressure plasma polymerization, nano-scaled oxygen-rich plasma polymer films (CO) were deposited onto pristine silicon wafers as well as on nitrogen-containing plasma polymer (CN) model surfaces. We investigate the influence of the nature of the substrate as well as a potential sub-surface effect emerging from the buried CO/CN interface, just nanometers below the surface. X-ray Photoelectron Spectroscopy and Time-of-Flight Secondary Ion Mass Spectrometry revealed two important phenomena that occurred during the deposition of the terminal CO layer: (1) a strong degree of oxidation, already for 1 nm nominal thickness, and (2) a gradual transition in chemical composition between the two layers, clearly indicating that effectively a vertical chemical gradient results, even when a two-step coating process was applied. Such terminal gradient film structures were used to study film stability in aqueous environments. Molecular rearrangements were scrutinized in the top-surface in contact with water and we found that the top-surface chemistry and wetting properties of the oxygen-rich termination layer matched those of thick CO reference coatings. Nevertheless, the adsorption of green fluorescent protein (GFP) was observed to be sensitive to the CO terminal layer thickness. Namely, an enhanced protein adsorption was observed for 1–2 nm thick CO layers on CN, whereas a significantly reduced protein adsorption was seen on ≥ 3 nm thick CO terminal layers. We conclude that both, surface and sub-surface conditions significantly affect protein adsorption as opposed to the traditional consideration of surface properties alone.
KeywordsPlasma deposition Vertical gradient film Chemical depth profiling Surface properties Protein adsorption
Achieving control over the surface properties of materials widens their range of applications. Amongst the many available surface modification methods, polymer coatings are commonly applied to generate new properties at the surface of materials, such as anti-fouling [1, 2, 3], anti-corrosive [4, 5], bio-sensing [6, 7], heavy metal adsorbent [8, 9] or fire-retardant [10, 11] properties. Homogeneous single-layer coatings are often preferred as they provide a single-step treatment—often by self-assembly. In polymer chemistry, this has led to the design of single precursor molecules carrying multiple functional properties, as in the case of antifouling zwitterionic polymers and peptoids that combine negative and positive charges and assemble into brushes [12, 13]. As an alternative, coating processes based on plasma state processing provide additional benefits for applications in diverse fields, such as (1) a high level of control over the process and free choice of the substrate material, (2) the (covalent) surface treatment of materials without compromising on the bulk properties of the same material , (3) the economic use of minimal amounts of substances, and (4) the production of a minimal amount of effluents. Moreover, plasma polymer films (PPFs) are cross-linked and can be characterized by higher stability than conventional polymers [15, 16]. At reduced cross-linking degrees, PPFs can be chemically more functional, but undergo degradation in presence of water . This is an issue in long-term applications in aqueous environments. Indeed, the stability is therefore often antagonistic to its degree of chemical functionality [18, 19, 20, 21]. For many applications, a high density of functional groups is essential. Thus, recent studies were carried out to provide stable and sufficiently functional PPFs. These studies mainly consisted in forming a vertical chemical gradient in the plasma coating, by first depositing a cross-linked (and stable) base layer onto the substrate, terminated by a few-nanometer-thick less cross-linked and more functional top-layer [16, 22, 23, 24]. This way, functional PPF terminations with enhanced stability were obtained.
Two main methods have been reported in the literature to prepare such PPFs. The first method consisted in tuning the plasma polymerization mode from continuous wave (CW) to pulsed power to provide a vertical gradient regarding chemical composition and cross-linking . Indeed, at comparable power input, the CW plasma polymerization caused stronger fragmentation of the monomers and etching processes, reducing the availability of the desired functional groups, thus leading to more cross-linked polymer-like networks . In contrast, the use of a pulsed plasma polymerization mode reduced fragmentation of monomers and enhanced gas phase radical polymerization during the plasma-off time , leading to the retention of a higher level of chemical functional groups [27, 28]. This second method consists of tuning key plasma parameters during the deposition of the film. Indeed, power input, gas flow rates as well as gas flow ratio (for gaseous mixtures, e.g. a monomer mixed with a reactive gas) can be easily changed during plasma deposition . While power input directly affects ion bombardment, which is linked to cross-linking in the PPF, a higher gas flow ratio of reactive gas to monomer gas increases the number of fragments that contain heteroatoms. Accordingly, the number of functional groups in the coating can increase, yielding less cross-linking in the PPF. Regardless of the method used to provide a coating with a vertical gradient, the sub-surface, i.e. the region underneath the outermost ~ 1 nm thick top-surface termination layer, was observed to have an additional influence on the surface properties, as the gradient coating was found to be more stable than a monolayer entirely consisting of the top layer composition—also when stored in water over days . The precise film architecture, i.e. the thickness of the top-layer, however, might be crucial for the observed stabilization effect.
Only few studies were carried out to investigate the effects of nanometer sub-surface layer of a coating on protein adsorption [27, 29, 30, 31], even if it was observed to play a significant role for its properties. A recent study showed that the sub-surface region of PPFs can energetically modify the protein adsorption process . Indeed, adsorption of bovine serum albumin (BSA) protein was investigated on highly stable plasma polymer films, namely a hydrophilic SiOx, covered by a nanometer-thick hydrophobic layer plasma-polymerized from hexamethyldisiloxane (ppHMDSO) used as terminal coating. If the latter was composed of a 4 nm thick terminal ppHMDSO layer deposited on SiOx, which was found to significantly reduce protein adsorption on the hydrated gradient structure . As adsorption of proteins onto surfaces involves conformational changes and modified interaction energies in presence of water, studies regarding water penetration into these coatings were carried out . It was observed that water penetrated more than 10 nm into the hydrophobic ppHMDSO within 4 h, while hydration water reached a depth of over 30 nm in the hydrophilic (and nanoporous) SiOx PPF. It was reported that water molecules in the sub-surface layer can energetically contribute to the protein adsorption process via long-range forces, demonstrating the potential of tailored sub-surface film architecture .
The present study aims at investigating the near-surface structure of a nano-scaled oxygen-rich PPF deposited onto a nitrogen-containing base film and its behavior in aqueous media. By varying the thickness of the terminal oxygen-containing PPF, the contribution of the nitrogen-containing sub-surface to the coating properties can be modified and studied. The nitrogen-containing CN film is used in this study as base film and has previously been described and characterized . Likewise, the thicker version of the oxygen-rich PPF has previously been studied, e.g. for biomedical applications [32, 33]. However, the top layer (CO) discussed in those previous studies was 5–10 nm thick, essentially masking a possible contribution from the sub-surface interface. Following our hypothesis that the sub-surface contribution distinctly depends on the top-layer thickness, a series of coatings were here prepared by depositing a 1–3 nm thick oxygen-rich PPF (CO) onto a 5 nm thick base layer of amine-containing PPF (CN). This base layer was carefully analyzed regarding: (1) chemical composition by X-ray Photoelectron Spectroscopy (XPS) and Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS) combined with depth profiling technique; and (2) top-surface properties such as wettability, by water contact angle (WCA) measurements, surface charging, by surface zeta potential (SZP) measurements, and surface interactions via adsorption of green fluorescent protein (GFP).
PPFs were deposited onto microscope glass slides (7.6 cm × 2.6 cm) and Si wafers. Substrates were cleaned in an Ar plasma (10 min, 50 W, 10 Pa) prior to plasma deposition. All chemicals were purchased from Sigma Aldrich (Buchs, Switzerland).
Plasma Deposition of the O-rich PPF onto the CN Model Surface
The cylindrical plasma reactor used to perform the PPF deposition consisted of two symmetric plane parallel electrodes (30 cm in diameter) that were separated by a glass ring (5 cm in height). The upper electrode contained a gas showerhead, while the chamber was pumped via a grid through the lower electrode, which was coupled to the RF generator enabling well-defined deposition conditions . A capacitively coupled low pressure discharge with 13.56 MHz radiofrequency excitation was used for the deposition of different PPFs: discharges were carried out in a continuous flow of NH3/C2H4 and CO2/C2H4 mixtures to obtain amine-containing (CN) and carboxylic acid-containing (CO) PPFs, respectively.
Characterization of the CO/CN Coatings
X-ray Photoelectron Spectroscopy (XPS)
The chemical composition of the surfaces deposited onto Si wafer samples was analyzed by XPS measurements using a Scanning XPS Microprobe (PHI VersaProbe II spectrometer, Physical Electronics) with monochromatic Al Kα radiation (1486.6 eV). The operating pressure of the XPS analysis chamber was below 5 × 10−7 Pa during the measurements. The spectra were collected at photoemission take off angles of 45° (with respect to the sample surface). Survey scan spectra (0–1100 eV) were acquired with an energy step width of 0.8 eV, acquisition time of 160 ms per data point and analyzer pass energies of 187.85 eV. For each sample a randomly chosen spot was analyzed using a micro-focused, scanned X-ray beam with a diameter of 100 µm (operated at a power of 25 W at 15 kV). The 180° spherical capacitor energy analyzer was operated in the fixed analyzer transmission mode (FAT). In order to compensate possible sample charging, dual beam charge neutralization with a flux of low energetic electrons (1.4 eV) combined with very low energy positive Ar-ions (10 eV) was used. Obtained spectra were shifted relatively to the aliphatic carbon (C–C) at 285.0 eV. Curve fitting (least-squares fit routines) with CasaXPS software version 2.3.16 was employed to calculate the atomic concentrations. A mixed Gaussian–Lorentzian product function (constant ratio of 70% Gaussian and 30% Lorentzian) was used to de-convolute the XP spectra and a linear type background was subtracted from the XPS peak areas. For quantification, PHI sensitivity factors  were corrected for our system’s transmission function and spectrometer geometry (asymmetry function).
Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS) and Depth Profiling
ToF-SIMS measurements (ToF-SIMS.5 instrument, IONTOF, Germany) were performed to analyze the chemical composition of the CO/CN coatings (deposited onto Si wafer pieces of 1 × 1 cm2) at the molecular level and more precisely, to identify the molecular composition between the two layers. 50 keV Bi3+ primary ions in the high mass resolution mode were used to analyze the PPF topmost surface (first few monolayers) on areas of 100 × 100 μm2. To determine the chemical depth profile through the gradient, a layer-by-layer abrasion 250 eV Cs+ sputtering beam (700 × 700 μm2, concentric to the analysis area) was used in combination with Bi3+. Note that such low energy sputtering yields a standard in-depth resolution of typically ≤ 2 nm. All negatively charged secondary ions from mass 1–550 atomic mass units were detected in parallel (typical sensitivity in the ppm range) and followed over time while abrading the plasma coating in order to reveal the two layers. After inspection of surface spectra from the reference single-layer samples and assignment of relevant intensity peaks, the following 14 characteristic fragments for both layers and substrate were considered for the analysis of the depth profiles: C2HO− (m/z = 41.00), C3N− (m/z = 50.01), C3O− (m/z = 51.99), C3H2N− (m/z = 52.02), C3HO− (m/z = 53.00), C3H3O− (m/z = 55.01), SiO2− (m/z = 59.97), C5N− (m/z = 74.00), C6HN− (m/z = 87.01). All the profiles were calibrated to convert the sputter time into depth (nm), based on independent prediction of the respective layers thickness. For both reference films, CN and CO, as well as for all the CO/CN combinations, three depth profiles have been performed at randomly selected locations to assess reproducibility and film homogeneity.
Water Contact Angle Measurements
Static WCA were determined for two samples (Si wafer pieces of 0.5 × 2 cm2) of each coating stored for 30 min in different environments, namely air or distilled water at 20 °C. In addition, the WCA was measured after 1 day of water storage to give a first hint on the stability of the plasma coatings. Samples were then dried by air blowing right before measurements, to remove only superficial water. WCA was measured by depositing at least 3 drops (2 μL) of water (CHROMASOLV®, for HPLC, Sigma Aldrich) onto the surface of each coated sample, and the software controlled drop shape analyser (DSA4, Krüss) was used to determine the contact angle.
Dynamic water contact angles were determined on 3 different samples of each coating after 20 min of air storage. Advancing and receding angles were also measured for freshly prepared PPFs by increasing and then decreasing the volume of the drop. To ensure good and reliable measurements  the following protocol was followed using the automatic mode of the DSA25 Krüss device: a drop of 5 μL is first deposited onto the surface—the needle being very close to the surface in the middle of the drop. Then, the volume of the drop was increased to 50 μL at a rate of 30 μL/min and finally decreased until the drop reached a volume of 3 μL. When the drop volume increases, a constant WCA can be obtained after few seconds: this is the advancing WCA. When the drop volume decreases, a constant WCA can also be obtained after few seconds: this is the receding WCA. For each drop, the arithmetic mean of the advancing and of the receding WCA values was reported.
Surface Zeta Potential Measurements
Surface zeta potential (SZP) was measured on coated Si wafers of 3 × 4 mm2 size using a Nano ZSP (Malvern Instruments) with a surface zeta potential dip cell (ZEN1020)  and the following tracer particles: micromer® PEGylated polystyrene particles [micromod Partikeltechnologie GmbH; monodispersed; size: 1 μm; concentration: 5 µg/mL; at pH: 7.00 ± 0.05 (phosphate buffer)]. The intrinsic zeta potential of the particles is − 14 ± 1 mV. As the surface charge has an influence on the mobility of those tracer particles at smaller distance from the surface, the determination of the surface zeta potential was carried out by measuring the velocity of the tracer particles at different distances from the surface (i.e. at 125, 250, 375 and 500 μm). Indeed, close to the surface, the tracer mobility is dominated by electro-osmotic surface flow, which essentially depends on the zeta potential of the surface, and, at larger distances the tracer mobility is dominated by electrophoretic motion; the zeta potential of the tracer particles was thus determined at 1000 μm from the surface of the sample. Surface zeta potential was then calculated from the velocity of standard tracer particles at different distances from the surface of the sample and the known zeta potential of these particles .
Atomic Force Microscopy Measurements
Surface roughness of the plasma polymer films (deposited onto Si wafer pieces of 1 × 1 cm2) was determined by atomic force microscopy (AFM) just after plasma deposition and after 1 day in distilled water. To perform the AFM analyses, a scanning probe microscope FlexAFM V5 (Nanosurf AG, Liestal, Switzerland) equipped with a C3000 controller and the associated software was used. Areas of 1 × 1 μm2 were scanned using a pyramidal silicon tip with a resonance frequency of 190 kHz and a spring constant of 48 N/m (BudgetSensors, Tap190Al-G). Measurements were carried out in tapping mode. AFM images were then analyzed and the RMS roughness was calculated using the open-source software Gwyddion 2.45.
Protein Preparation and Characterization
Recombinant Protein Production
The GFPuv version of the green fluorescent protein (GFP) was used for protein adsorption studies, whereas a yellow fluorescent negatively charged protein (negYFP) was used as reference in zeta potential measurement. Protein negYFP was obtained by site-directed mutagenesis (T214Y) of the negatively supercharged GFP using plasmid pET-6xHis-(-30)GFP. pET-6xHis-(-30)GFP was a gift from David Liu (Addgene plasmid # 62936) . Fluorescent proteins were overexpressed in E. coli and purified by immobilized metal affinity chromatography as previously described . E. coli BL21 (DE3) chemocompetent cells were transformed with plasmid pPP1 [40, 41]. Cultivations were carried out in LB medium supplemented with 0.1 mg/mL ampicillin and 1 mM IPTG (final concentration), while shaking at 180 rpm. Harvested by centrifugation, cells were lysed by sonication, the cell-free extract clarified by ultracentrifugation, and the His-tagged fluorescent protein separated by affinity chromatography using a Histrap column (column volume = 5 mL) on an ÄKTA Protein Purification Systems (GE Healthcare Life Sciences) using 100 mM potassium phosphate buffer pH 7.00 ± 0.05.
Biochemical Protein Characterization and Zeta Potential in Solution
The purity of the protein preparations and the size of the molecules were determined by SDS PAGE. Absorbance and fluorescence spectra of GFP were recorded in 100 mM potassium phosphate buffer at pH 7.00 ± 0.05. The absorbance and fluorescence properties of GFP were analysed by UV–Vis and fluorescence spectroscopy in 100 mM potassium phosphate pH 7 solutions. GFP concentrations were measured by UV–Vis absorption using a calculated ε0.1% of 0.8, at 280 nm. Measurements of zeta potential of the proteins were performed in filtered McIlvaine buffer (Na2HPO4, citric acid) in a pH range from 2.2 to 8 using a Malvern Zetasizer at 25 °C.
Protein Structure Analyses
The software Pymol  (The PyMOL Molecular Graphics System, Version 1.7.4 Schrödinger, LLC) was used to visualize the three-dimensional structure of GFP (PBS ID: 1eme) as retrieved from the Protein Database . Surface potential was visualized using the Adaptive Poisson-Boltzmann Solver (APBS) using default charge values and a solvent dielectric constant of 80 (water at pH 7) .
Protein Adsorption onto the CO/CN Coatings
Visual analysis of Protein Adsorption
Before treatment, plasma-treated surfaces deposited on glass slides were incubated overnight in deionized water to neutralize the still-living radicals [45, 46]. Protein adsorption was assessed by depositing 200 μL of a solution containing the green fluorescent protein, GFP in 100 mM potassium phosphate pH 7.00 ± 0.05 at a protein concentration of 2 mg/ml onto the coated surfaces and incubated at 22 °C for 30 min in the dark under moist conditions. Samples were then rinsed for 2 min in distilled water (3 times) before they were gently dried in air, while covered by aluminum foil, before characterization. Samples were prepared in triplicate to ensure repeatability.
Imaging of Adsorbed Proteins on the Coatings
Surfaces treated with GFP were imaged with a LS Reloaded Microarray Scanner (Tecan, Switzerland) using a 488 nm laser, a FITC filter (535 nm), a 140 gain, and a 10 µm resolution with a small pinhole. Images were then analyzed with the software provided by Tecan (Switzerland) to determine the fluorescence intensity averaged over a grid of nine circles (Ø = 2 mm). Fluorescence intensity values obtained from surfaces incubated with protein-free buffer solutions under identical conditions of temperature and time were used as control and subtracted from the values obtained.
Results and Discussion
A base CN coating was initially prepared as reference model surface with a nominal thickness of 5 nm to ensure highly stable, amine-functional plasma polymer films . The CO reference coating was deposited onto pristine (and plasma-cleaned) Si wafers with a nominal film thickness of 6 nm, to ensure full coverage of the substrate. In addition, the CO film of variable thickness (1–3 nm) was deposited onto 5 nm of CN according to a sequential process. For all CO/CN coatings, the nominal thickness values of the terminal layer were based on the previously determined deposition rate of the plain plasma polymer film (i.e. CO reference), at a rate of ~ 1.2 nm/min for CO. As plasma polymerization occurs through both deposition and etching processes, a mixture of nitrogen- and oxygen-containing chemical groups may form at the interface between the two layers leading to a vertical chemical nano-gradient through the nano-film. Moreover, with a thickness of the top-layer of about 1–2 nm, chemical groups from the CN layer are expected to have an influence on surface properties, i.e. the chemical gradient layer might extend to the outermost surface.
Characterization of O-rich PPF Deposited onto CN Model Surface
A similar [O]/[C] ratio, i.e. around 25–27%, is observed for the coatings CO and CO/CN (independent of the thickness of the CO top-layer on the CN layer), whereas lower [O]/[C] ratios were obtained for the CN coatings of around 10.3%. Regarding the [N]/[C] ratios, the same trend can be noticed. Indeed, from XPS analyses, no nitrogen content is detected in the CO coating. Furthermore, a decreased amount of nitrogen was observed with the increase of the terminal CO layer thickness: [N]/[C] ratio was 5.6, 3.3, and 1.8% for a CO layer thickness of 1, 2, and 3 nm, respectively. As expected, the CN coating contained a higher amount of nitrogen ([N]/[C] = 18.5%). Since the top 6 nm were analyzed by XPS under these experimental conditions, both the top layer and the base layer contributed to the measurement of the chemical composition, also involving the entire chemical gradient.
Evidence for added oxidation of the CN layer, i.e. formation of a vertical chemical gradient, can be given by considering the detail spectra of C1s and N1s (Fig. 2b, d). Indeed, the C1s photoelectron signal peaks (Fig. 2b, c) are broader, indicating the presence of an additional oxidation state, since a layer of CO (> 1 nm) was deposited onto the CN layer. This can also be supported by the finding that C1s peaks of CO and CO/CN coatings were best de-convoluted with 4 peaks, whereas 3 peaks were adequate to de-convolute the C1s peak of the CN coating alone. As one might expect, a clear decrease of the nitrogen content was observed by increasing the thickness of the CO layer on top of the CN base layer (Fig. 2d, e). The [N]/[C] ratio decreased from 18.5 to 5.6% by adding 1 nm of CO onto the CN layer (CO/CN 1:5), whereas it decreased further from 5.6 to 3.25% and from 3.25 to 1.80% by increasing the CO layer thickness to 2 and 3 nm (CO/CN 2:5 and CO/CN 3:5), respectively (Fig. 2a, d). Thus, a significant oxidation of the interface with the CN layer occurred during the deposition of the first 1 nm of terminal CO induced by the high CO2/C2H4 gas ratio used yielding chemical (oxidative) etching. For thicker CO/CN coatings, the decrease of N1s peak corresponds to the expected signal attenuation of photoelectrons across a thicker CO top-layer.
As observed on Fig. 3a for the fragments C3N−, the intensities of the nitrogen-containing fragments were higher for the CN layer than for the CO coating. Moreover, CO/CN coatings exhibited a gradual transition (over a range of approximately 2 nm) between the top-layer CO and the base layer CN (at different depth according to the thickness of the CO layer), even if double layer coatings were produced by a subsequent (two-step) process. Thus, formation of a vertical chemical gradient occurred during the coating process. The formation of the gradient structure is likely due to the fact that the terminal layer deposition, i.e. the CO layer, was prepared under plasma conditions that promote moderate chemical etching of the underlying CN layer. The used gas flow ratio CO2/C2H4 of 6:1 is known to enhance etching phenomena during plasma deposition [47, 48]. Consequently, at the early stage of deposition of the terminal CO layer, (ion-assisted) oxygen etching occurs, yielding oxidation and formation of radical sites as well the release of some nitrogen-containing species into the plasma zone. These species are able to react with other species in the plasma zone and can also be re-deposited onto the surface, leading—together with the oxidation processes—to the formation of a vertical chemical gradient.
According to expectations, the oxygen-containing fragment C3HO− was counted higher or lower from the pure CO or CN coatings, respectively. Anyway, the evolution of the intensity of the C3HO− fragments (Fig. 3b) in the depth of the CO and CN reference coatings followed the same tendency: (1) an important decrease of the intensity counts occurs in the first 0.5 nm of both coatings, and is followed by (2) a regular increase of the intensity counts until the PPF/substrate interface is reached. The first 0.5 nm actually corresponds to the oxidation layer at the extreme surface of all coatings (< 1 nm thick), and is formed due to post-plasma oxidation process. The increase in intensity observed close to the interface between the PPF layer and the Si wafer is actually due to a so-called matrix effect (the proximity of the oxygen from the substrate promotes the ionization potential of secondary ions) in addition to a possible recombination of some substrate oxygen atoms with polymer fragments. As coatings have very small thicknesses in this study, these effects hinder the precise behavior of the oxygen-containing fragments in the depth of the coatings. Anyway, differences in between the CO/CN coatings can be observed on Fig. 3b. Indeed, higher intensity count values were found at the top surface, highlighting the post-plasma oxidation process that occurs when exposed to ambient air after plasma deposition. When 1 nm of CO was deposited onto 5 nm of CN, an immediate transition from the oxidized layer to the CN layer was observed: the gradient was thus extended to the outermost surface. Note that such result was already obtained when creating an oxygen-containing coating with a lateral chemical gradient: the gradient was also extended to the outermost surface . When 2 or 3 nm of CO was deposited onto 5 nm of CN, a new transition was observed in the evolution of the intensity of oxygen-containing fragments. Indeed, below the oxidized top-layer, a new layer with a higher oxygen-content compared to the CN layer, was detected. In addition, a small transition (at 3 or 4 nm in the depth of coatings CO/CN 2:5 and 3:5, respectively) indicates another layer below this oxygen-containing layer: as a slight decrease of the intensity counts was observed, a lower oxygen-content is expected in this base layer corresponding to the CN layer. Figure 3b reveals that the oxygen content in the CN base layer seems to increase with the CO terminal layer thickness. Actually, the oxygen content should be the same in the bulk of all CN base layers, and this increase is due to the matrix effect, as previously mentioned. This is the reason why the transition in between the CO and CN layers is less sharp, and thus slightly more difficult to observe. Regardless, the position of this transition in the depth clearly depends on the thickness of the CO layer, showing the interface in between CN and CO layers.
Thus, in-depth profiling ToF-SIMS measurements revealed some variation in the chemical composition of the CO layers, depending on the substrate on which this layer was deposited. Indeed, when the CO layer was deposited onto the CN model surface, a mixture of oxygen- and nitrogen-containing fragments were observed in the top-layer, showing a gradual chemical composition (with a thickness of ~ 4 nm) different from the CO layer deposited onto Si wafer. The formation of such a gradual interface structure was explained by etching and (re-)deposition processes occurring at the same time during plasma deposition of the second layer. Moreover, the CN model surface had the same chemical composition, even after the deposition of the CO layer, but the final film thickness (calculated thanks to the previously determined etching rate) was smaller after CO layer deposition, showing the etching process occurring during the deposition of the CO layer.
Such analyses on plain and gradient plasma polymer films demonstrated that the surface properties, e.g. wettability and surface charge, were very similar even for a chemical gradient extending to the outermost surface (as for a terminal CO layer thickness close to 1 nm). Since it was already recognized that the single layers and gradient-containing coatings aged according to the same processes in air, yet differently in water , it becomes thus interesting to investigate the aging behavior of the CO/CN coatings in water.
Influence of Water Aging on the Chemical Composition of the CO/CN Structures
The chemical and structural stability of ultra-thin plasma polymer films is of high interest that motivated different approaches to reduce aging phenomena over recent years, also including gradient structures . Therefore, static water contact angles were measured after 30 min storage of the PPFs in air or water, and also after 1 day storage in water. Regarding the CN model surface, similar WCA were measured for both storage conditions after 30 min, e.g. a WCA of 55 ± 3° after air storage and of 52 ± 5° after water storage. In addition, after 1 day in water, similar water contact angle values were obtained, e.g. 55 ± 1°, indicating the high stability of the CN model surface.
When the CO layer was deposited onto the Si wafer, WCA of 45 ± 1° after air storage and 43 ± 1° after water storage were measured. After 1 day in water, the same coating displayed a WCA value of 46 ± 2°, showing the equally high stability of this coating being slightly more hydrophilic than CN. When CO is deposited onto the CN model surface, a WCA of 47 ± 2° was measured after 30 min of storage in air, regardless of the thickness of the top-layer. This value is remarkably similar to the one obtained for the CO reference, e.g. 45 ± 1°. After 30 min of storage in distilled water, different WCA values were obtained—this time depending on the thickness of the oxygen-containing termination: for CO/CN 1:5, a WCA value of 51 ± 2° was obtained, which was closer to the WCA value as obtained for the CN coating, indicating that small restructuring and degradation effects of the CO termination layer may have occurred due to hydrolysis of ester groups . When the CO layer had a thickness of 2 or 3 nm, significantly smaller water contact angle values of 41 ± 2° and 44 ± 1°, respectively, were measured. These findings are again close to the WCA value obtained for the CO reference, indicating that, even if some chemical changes of the top-layer had occurred, the top-layer largely remained intact over the immersion time of 30 min in distilled water. After 1 day in water, CO/CN 2:5 and 3:5 displayed WCA values of 41 ± 3° and 43 ± 2°, respectively, showing merely minor changes limited to the outermost molecular layer.
The chemical composition in the plasma polymer films was determined by XPS after 1 day in water (Figure S1). The CN model surface is already known to retain its chemical composition when immersed in water over 1 day , with [O]/[C] and [N]/[C] ratios of ~ 10 and ~ 18%, respectively. In this study, the chemical composition of the CO layer was also observed to be constant after 1 day in water, with a [O]/[C] ratio of 27.7% (compared to the initial [O]/[C] ratio of 27.9%). When a 1 nm layer of CO is deposited onto the CN model surface, the [O]/[C] ratio was found to be almost similar (25.4% after 1 day in water compared to the initial 26.0%), whereas the [N]/[C] ratio was observed to increase from 5.6 to 9.5%. With the CO termination thickness of 2 or 3 nm minor changes in the chemical composition were observed: for CO/CN 2:5, the [O]/[C] ratio increased from 26.4 to 28.7% and the [N]/[C] ratio slightly increased from 3.3 to 3.8%; and for CO/CN 3:5, the [O]/[C] ratio remained the same with 26.5%, and the [N]/[C] ratio slightly increased from 1.8 to 2.6%. Thus, it can be assumed that some fragments from the CO layer (when deposited onto the CN model surface) were dissolved during water aging which, however, represent less than 1 nm of the CO coating.
As a conclusion, it was observed based on the change in WCA, the variation of the chemical composition and of the surface topography that the top of the CO layer slightly degraded in water. The release of dissolved fragments seems to occur over 1 day of immersion in water and affects only the top ≤ 1 nm of the coatings, which could be demonstrated by using the adjusted film architectures. The outermost surface of functional PPFs might thus form some low molecular weight organic compounds that can be dissolved , which was, however, limited by the design of the CO/CN gradient structure. In accordance, it was recently shown that about 1 nm of a functional CO layer stabilized on a cross-linked CO base layer (termination thickness > 2 nm) can be released when immersed in water for 1 h, while a 1 nm CO termination showed less surface reconstructions . Hence, water interaction with the near-surface structure of the PPFs induced small surface rearrangement effects depending on the exact chemical gradient structure acting as a stabilizing factor. The high cross-linking degree, however, prevented the film from swelling .
Protein Characterization and Protein Adsorption onto CO/CN Structures
To investigate protein adsorption, the different plasma coatings were incubated for 30 min in a GFP-rich solution (2 mg/mL) and were rinsed three times with distilled water and intermittent air-drying before analysis. The surfaces were imaged with a microarray scanner (resolution = 5–10 µm), and the fluorescence signal was quantified by averaging the fluorescence intensity over nine smaller areas (Figure S3). Note that the intensity of the fluorescence signal would be affected by denaturation of GFP. In addition, since the drying process is known to have an influence on the protein conformation [55, 56] and thus on the fluorescence intensity, the samples were dried all under the same conditions. The drying process should therefore not have an influence on the relative differences in fluorescence intensity measured between reference and CO/CN structures. Finally, the same washing procedure was applied to all surfaces to remove the loosely bounded GFP, leading to a decrease of the fluorescence signal.
A similar fluorescence value for all surfaces with similar zeta potential and WCA values (i.e. CO reference and CO/CN gradients) would have been expected, if protein adsorption were solely determined by top surface properties alone. The fluorescence values for the CO/CN samples comprising a near-surface vertical gradient structure, however, were found to noticeably differ from the value obtained for the CO reference, suggesting an influence of the underlying CN layer. As it is well-known, the interaction of proteins such as GFP with surfaces is affected by both surface energy (hydrophobicity) and charge–charge-interactions . It can thus be hypothesized that also a gradient in surface energy and zeta potential extending over a range of a few nanometers of the near-surface structure has a pronounced impact on protein adsorption susceptible to the exact chemical gradient structure. Hence, protein adsorption could be relatively enhanced on such vertical chemical CO/CN nano-gradients with 1–2 nm terminal CO layer.
Nano-scaled oxygen-rich CO layers were deposited onto Si wafer and amine-containing CN model surfaces to investigate the influence of the nature of the substrate and the sub-surface onto the surface properties. The elemental and chemical composition was assessed using XPS and ToF-SIMS techniques. A chemical mixing zone was found forming a vertical chemical nano-gradient between the CO and the CN plasma polymer films. In addition, it was shown that the CO layer deposited onto the CN model surface contained slightly more nitrogen and less oxygen than the CO layer deposited onto the Si wafer also indicating the near-surface chemical gradient structure. Regardless of its thickness, however, the thin CO termination layers showed similar chemical surface composition, wettability, hydration properties and surface charge as the thicker CO reference coating.
After 1 day in water, restructuring and partial degradation of the CO layer was observed, but it was limited to the outermost molecular layers as evidenced on the gradient CO/CN coatings. Accordingly, a higher RMS roughness was detected for all samples. In addition, an increase of the [N]/[C] ratio was observed mainly for the CO/CN 1:5 coating comprising the thinnest termination layer of 1 nm. The actual surface conditions are thus slightly changing in aqueous conditions partly exposing the CN model surface, allowing nitrogen-containing functional groups to be available.
Finally, protein adsorption was observed to be distinctly different for the examined CO surfaces (as deposited onto Si wafer or onto CN base layer) and distinctly depending on the thickness of the top-layer. The green fluorescent protein (GFP) adsorbed in highest amount onto the double layer structure with a 1 nm-thick terminal CO layer, whereas the amount of adsorbed protein was found to be significantly reduced already with a 3 nm-thick top-layer thickness. While it was not the aim to assess the overall performance of the examined PPFs regarding high or low protein adsorption, water interaction and related small surface rearrangements were found to affect molecule-surface interaction such as protein adsorption, which cannot be explained adequately by surface properties alone. Proteins can thus react sensitively to the structure and composition of plasma polymer films, including the range of a few nanometers below the surface.
M. V. and D. H. gratefully acknowledge the Swiss National Science Foundation (SNSF, Bern) that funded this study under Grant No. IZ73Z0_152661 (SCOPES). A. C. and G. G.-G. like to acknowledge support by the Swiss National Science Foundation as part of the NCCR Molecular Systems Engineering as well as Prof. Wolfgang Meier. G. G.-G. thanks the German Academic Exchange Service (DAAD) for a postdoctoral fellowship. G. F. also would like to thank Erik Mailand for the technical assistance in the expression and purification of GFP.
- 35.Moulder F, Stickle WF, Sobol PE, Bomben KD (1995) Handbook of X-ray photoelectron spectroscopy. Physical Electronics Inc., Eden PrairieGoogle Scholar
- 42.Lerner MG, Carlson HA (2006) APBS plugin for pymol. University of Michigan, Ann ArborGoogle Scholar
- 54.Ward WW (1998) Biochemical and physical properties of green fluorescent protein. In: Chalfie M, Kain S (eds) Green fluorescent protein: properties, applications and protocols. Wiley-Liss, New York, pp 45–75Google Scholar