Surface modification on microfluidic devices with 2-methacryloyloxyethyl phosphorylcholine polymers for reducing unfavorable protein adsorption

Surface modification of polymer materials for preparing microfluidic devices including poly(dimethyl siloxane) (PDMS) was investigated with phospholipids polymers such as poly(2-methacryloyloxylethyl phosphorylcholine(MPC)-co-n-butyl methacrylate) (PMB) and poly(MPC-co-2-ethylhexyl methacrylate-co-2-(N,N-dimethylamino)ethyl methacrylate) (PMED). The hydrophilicity of every surface on the polymer materials modified with these MPC polymers increased and the value of zeta-potential became close to zero. The protein adsorption on the polymer materials with and without the surface modification was evaluated using a protein mixture of human plasma fibrinogen and serum albumin. Amount of proteins adsorbed on these polymeric materials showed significant reduction by the surface modification with the MPC polymers compared to the uncoated surfaces ranging from 56 to 90%. Furthermore, we successfully prepared PDMS-based microchannel which was modified by simple coating with the PMB and PMED. The modified microchannel also revealed a significant reduction of adsorption of serum albumin. We conclude that the MPC polymers are useful for reducing unfavorable protein adsorption on microfluidic devices.     

Adsorption of lipid-functionalized poly(ethylene glycol) to gold surfaces as a cushion for polymer-supported lipid bilayers

Inclusion of a polymer cushion between a lipid bilayer membrane and a solid surface has been suggested as a means to provide a soft, deformable layer that will allow for transmembrane protein insertion and mobility. In this study, the properties of a heterofunctional, telechelic PEG lipopolymer (1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-poly(ethylene glycol)-2000-N- [3-(2-(pyridyldithio)propionate]) (DSPE-PEG-PDP) adsorbed from ethanol and water solutions onto gold surfaces were studied using a variety of surface-sensitive techniques. X-ray photoelectron spectroscopy showed that the PEG molecules are tethered to the gold surface via thiolate bonds. When adsorbed from water, ethanol, or their mixtures, reflection-absorption infrared spectroscopy showed that amorphous PEG layers with disordered DSPE alkyl chains were formed, independent of adsorption time or solution concentration. On the basis of advancing and receding water and hexadecane contact angles on the lipopolymer films, the DSPE lipid groups appear to segregate from the PEG layer and become exposed at the surface of the polymer films. Swelling observed in surface plasmon resonance experiments and the large contact angle hysteresis observed indicate that highly swellable, mobile films capable of molecular rearrangements are formed. The self-assembling and amorphous properties of these PEG layers make them ideal candidates as polymer cushions for polymer-supported lipid bilayers. The DSPE surface concentration can be controlled, to a limited degree, by varying the adsorption time of DSPE-PEG-PDP from ethanol. A more effective strategy is to coadsorb DSPE-PEG-PDP with a non-lipid-functionalized PEG-PDP from an ethanol/water mixture, which allows the PEG thickness and density to remain constant while decreasing the density of DSPE groups.     

Mobility of Thermomyces lanuginosus lipase on a trimyristin substrate surface

We have studied the mobility of active and inactive Thermomyces lanuginosus lipase (TLL) on a spin-coated trimyristin substrate surface using fluorescence recovery after photobleaching (FRAP) in a confocal microscopy setup. By photobleaching a circular spot of fluorescently labeled TLL adsorbed on a smooth trimyristin surface, both the diffusion coefficient D and the mobile fraction f could be quantified. FRAP was performed on surfaces with different surface density of lipase and as a function of time after adsorption. The data showed that the mobility of TLL was significantly higher on the trimyristin substrate surfaces compared to our previous studies on hydrophobic model surfaces. For both lipase variants, the diffusion decreased to similar rates at high relative surface density of lipase, suggesting that crowding effects are dominant with higher adsorbed amount of lipase. However, the diffusion coefficient at extrapolated infinite surface dilution, D0, was higher for the active TLL compared to the inactive (D0 = 17.9 x 10(-11) cm2/s vs D0 = 4.1 x 10(-11) cm2/s, data for the first time interval after adsorption). Moreover, the diffusion decreased with time after adsorption, most evident for the active TLL. We explain the results by product inhibition, i.e., that the accumulation of negatively charged fatty acid products decreased the diffusion rate of active lipases with time. This was supported by sequential adsorption experiments, where the adsorbed amount under flow conditions was studied as a function of time after adsorption. A second injection of lipase led to a significantly lower increase in adsorbed amount when the trimyristin surface was pretreated with active TLL compared to pretreatment of inactive TLL.     

Adsorption of cationic cellulose derivative/anionic surfactant complexes onto solid surfaces. II. Hydrophobized silica surfaces

The effect of the anionic surfactant SDS (sodium dodecyl sulfate) on the adsorption behavior of cationic hydroxyethyl cellulose (Polymer JR-400) and hydrophobically modified cationic cellulose (Quatrisoft LM-200) at hydrophobized silica has been investigated by null ellipsometry and compared with the previous data for adsorption onto hydrophilic silica surfaces. The adsorbed amount of LM-200 is found to be considerably larger than the adsorbed amount of JR-400 at both surfaces. Both polymers had higher affinity toward hydrophobized silica than to silica. The effect of SDS on polymer adsorption was studied under two different conditions: adsorption of polymer/SDS complexes from premixed solutions and addition of SDS to preadsorbed polymer layers. Association of the surfactant to the polymer seems to control the interfacial behavior, which depends on the surfactant concentration. For the JR-400/SDS complex, the adsorbed amount on hydrophobized silica started to increase progressively from much lower SDS concentrations, while the adsorbed amount on silica increased sharply only slightly below the phase separation region. For the LM-200/SDS complex, the adsorbed amounts increased progressively from very low SDS concentrations at both surfaces, and no large difference in the adsorption behavior was observed between two surfaces below the phase separation region. The complex desorbed from the surface at high SDS concentrations above the critical micelle concentration. The reversibility of the adsorption of polymer/SDS complexes upon rinsing was also investigated. When the premixed polymer/SDS solutions at high SDS concentrations (>5 mM) were diluted by adding water, the adsorbed amount increased due to the precipitation of the complex. The effect of the rinsing process on the adsorbed layer was determined by the hydrophobicity of the polymer and the surface.     

Fourier-transform Carr-Purcell-Meiboom-Gill NMR experiments on polymers in colloidal dispersions: how many polymer molecules per particle?

Fourier transform relaxation NMR has been used to study how the mobility of poly(ethylene oxide) is affected by its adsorption onto colloidal silica particles of various sizes. Novel results have been obtained which illustrate the unexploited potential of this method for the study of interfacial species in complex systems. The results quantify how polymer mobility varies along an adsorption isotherm. When the particles are in excess, the polymer is strongly adsorbed and hence has a large spin-spin magnetic relaxation rate constant, R(2). The value of R(2) in this region increases with particle size, because the associated reduction in particle surface curvature results in a reduction in the mobility of the adsorbed polymer. This is accompanied by a reduction in the signal intensity, as a higher fraction of the polymer is adsorbed in the form of train segments too immobile to detect using the Carr-Purcell-Meiboom-Gill pulse sequence. When the polymer concentration reaches approximately 0.5 mg m(-2), the initial region of high affinity adsorption ends and so the polymer solution concentration increases. This is accompanied by a reduction in R(2), which then approaches the value for a simple polymer solution in the absence of particles. The results are corroborated by comparison with rheological measurements and molecular dynamics simulations of an analogous particle-polymer system.     

Complexation of polycations to anionic liposomes: composition and structure of the interfacial complexes

Poly(N-ethyl-4-vinylpyridinium bromide) (a polycation with a degree of polymerization of 1100) was adsorbed onto liposomes composed of egg lecithin with a 0.05-0.20 molar fraction (nu) of anionic headgroups provided by cardiolipin (a doubly anionic lipid). According to electrophoretic mobility data, this led to total charge neutralization of the liposomes, whereupon the liposomes adopted a positive charge as additional polymer continued to adsorb. Although the liposomes aggregated at the charge-neutralization point, they disassembled into individual liposomes after becoming positively charged. The degree of polymer adsorption was shown to reach a limit. Thus, by measuring the free polymer content in a liposome suspension, it was possible to determine the polymer concentration at which the liposome surface became saturated with polymer. Beyond this point, an electrostatic/steric barrier at the surface suppressed further adsorption. Dynamic light scattering studies of liposomes with and without adsorbed polymer allowed calculation of the polymer film thickness which ranged from 22 to 35 nm as the molar fraction of cardiolipin (nu) increased from 0.05 to 0.20. The greater the content on the anionic lipid in the bilayer, the thicker the polymer film. The maximum number of polymer molecules adsorbed onto the liposomes was estimated: 1-2 molecules for nu = 0.05; 3 molecules for nu = 0.1; 4- molecules for nu = 0.15; and 6 molecules for nu = 0.2. The polymer appears to lie on the liposome surface, rather than embedding into the bilayer, because addition of NaCl easily dislodges the polymer from the liposome into the bulk water.     

Adsorption and activity of Thermomyces lanuginosus lipase on hydrophobic and hydrophilic surfaces measured with dual polarization interferometry (DPI) and confocal microscopy

The adsorption and activity of Thermomyces lanuginosus lipase (TLL) was measured with dual polarization interferometry (DPI) and confocal microscopy at a hydrophilic and hydrophobic surface. In the adsorption isotherms, it was evident that TLL both had higher affinity for the hydrophobic surface and adsorbed to a higher adsorbed amount (1.90 mg/m²) compared to the hydrophilic surface (1.40–1.50 mg/m²). The thickness of the adsorbed layer was constant (3.5 nm) on both surfaces at an adsorbed amount >1.0 mg/m², but decreased on the hydrophilic surface at lower surface coverage, which might be explained by partially unfolding of the TLL structure. However, a linear dependence of the refractive index of the adsorbed layer on adsorbed amount of TLL on C18 surfaces indicated that the structure of TLL was similar at low and high surface coverage. The activity of adsorbed TLL was measured towards carboxyfluorescein diacetate (CFDA) in solution, which upon lipase activity formed a fluorescent product. The surface fluorescence intensity increase was measured in a confocal microscope as a function of time after lipase adsorption. It was evident that TLL was more active on the hydrophilic surface, which suggested that a larger fraction of adsorbed TLL molecules were oriented with the active site facing the solution compared to the hydrophobic surface. Moreover, most of the activity remained when the TLL surface coverage decreased. Earlier reports on TLL surface mobility on the same surfaces have found that the lateral diffusion was highest on hydrophilic surfaces and at low surface coverage of TLL. Hence, a high lateral mobility might lead to a longer exposure time of the active site towards solution, thereby increasing the activity against a water-soluble substrate.     

Rapid development of hydrophilicity and protein adsorption resistance by polymer surfaces bearing phosphorylcholine and naphthalene groups

In order to provide a protein adsorption resistant surface even when the surface was in contact with a protein solution under completely dry conditions, a new phospholipid copolymer, poly (2-methacryloyloxyethyl phosphorylcholine (MPC)- co-2-vinylnaphthalene (vN)) (PMvN), was synthesized. Poly(ethylene terephthalate) (PET) could be readily coated with PMvN by a solvent evaporation method. Dynamic contact angle measurements with water revealed that the surface was wetted very rapidly and had strong hydrophilic characteristics; moreover, molecular mobility at the surface was extremely low. When the surface came in contact with a plasma protein solution containing bovine serum albumin (BSA), the amounts of the plasma protein adsorbed on the dry surface coated with PMvN and that adsorbed on a dry surface coated with poly(MPC-co-n-butyl methacrylate) (PMB) were compared. Substantially lower protein adsorption was observed with PMvN coating. This is due to the rapid hydration behavior of PMvN. We concluded that PMvN can be used as a functional coating material for medical devices without any wetting pretreatment.     

Structure of Protein Layers during Competitive Adsorption

The formation of protein layers during competitive adsorption was studied with ellipsometry. Single, binary, and ternary protein solutions of human serum albumin (HSA), IgG, and fibrinogen (Fgn) were investigated at concentrations corresponding to blood plasma diluted 1/100. As a model surface, hydrophobic hexamethyldisiloxane (HMDSO) plasma polymer modified silica was used. By using multiambient media measurements of the bare substrate prior to protein adsorption the adsorbed amount as well as the thickness and refractive index of the adsorbed protein layer could be followedin situand in real time. Under conditions used in these experiments neither IgG nor fibrinogen could fully displace serum albumin from the interface. The buildup of the protein layer occurred via different mechanisms for the different protein systems. Fgn adsorbed in a rather flat orientation at low adsorbed amounts, while at higher surface coverage the protein reoriented to a more upright orientation in order to accommodate more molecules in the adsorbed layer. IgG adsorption proceeded mainly end-on with little reorientation or conformational change on adsorption. Finally, for HSA an adsorbed layer thickness greater than the molecular dimensions was observed at high concentrations (although not at low), indicating that aggregates or multilayers formed on HMDSO plasma polymer surfaces. For all protein mixtures the adsorbed layer structure and buildup indicated that Fgn was the protein dominating the adsorbed layer, although HSA partially blocked the adsorption of this protein. At high surface concentration, HSA/Fgn mixtures show an abrupt change in both adsorbed layer thickness and refractive index suggesting, e.g., an interfacial phase transition of the mixed protein layer. A similar but less pronounced behavior was observed for HSA/IgG. For IgG/Fgn and HSA/IgG/Fgn a buildup of the adsorbed layer similar to that displayed by Fgn alone was observed.     

H/D isotopic exchange in water interactions with the ferroelectric copolymer: Poly(vinylidene fluoride-trifluoroethylene) (70%:30%)

For both water and heavy water adsorption and absorption on crystalline poly(vinylidene fluoride with trifluoroethylene (30%)), P(VDF-TrFE 70:30), two distinctly different adsorption sites have been identified by thermal desorption spectroscopy. One adsorbed water species resembles ice and there is also an absorbed water species that interacts more strongly with the polymer thin film, and in addition, there is a polymer surface (polymer to ice interface) water species. We find that there is H/D exchange between the water or heavy water molecules and the ferroelectric polymer (largely -(CH2-CF2)-), particularly at the polymer surface.     

Adsorption of PEO-PPO-PEO triblock copolymers with end-capped cationic chains of poly(2-dimethylaminoethyl methacrylate)

We study the adsorption of a symmetric triblock copolymer of ethylene oxide, EO, and propylene oxide, PO, end-capped with quarternized poly(2-dimethylaminoethyl methacrylate), DMAEMA (DMAEMA(24)-EO(132)PO(50)EO(132)-DMAEMA(24)). Light scattering and tensiometry are used to measure the relative size of the associated structures and surface excess at the air-liquid interface. The adsorbed amount, the amount of coupled water, and the viscoelasticity of the adsorbed polymer layer are measured on hydrophobic and hydrophilic surfaces (polypropylene, cellulose, and silica) by using quartz crystal microgravimetry (QCM) and surface plasmon resonance (SPR) at different ionic strengths and temperatures. The results of the experiments are compared with those obtained after adsorption of the uncharged precursor copolymer, without the cationic end-caps (EO(132)PO(50)EO(132)). DMAEMA(24)-EO(132)PO(50)EO(132)-DMAEMA(24) possesses higher affinity with the negatively charged silica and cellulose surfaces while the uncharged copolymer adsorbs to a larger extent on polypropylene surfaces. In this latter case, adsorption increases with increasing solution ionic strength and temperature. Adsorption of EO(132)PO(50)EO(132) on silica surfaces has little effect on the water contact angle (WCA), while adsorption of DMAEMA(24)-EO(132)PO(50)EO(132)-DMAEMA(24) increases the WCA of silica to 32°, indicating a large density of exposed PPO blocks upon adsorption. After adsorption of EO(132)PO(50)EO(132) and DMAEMA(24)-EO(132)PO(50)EO(132)-DMAEMA(24) on PP, the WCA is reduced by ≈14° and ≈28°, respectively, due to the exposed hydrophilic EO and highly water-soluble DMAEMA segments on the surfaces. The extent of surface coverage at saturation at the polypropylene/liquid interfaces (≈31 and 40 nm(2)/molecule obtained by QCM and SPR, respectively) is much lower, as expected, when compared with results obtained at the air/liquid interface, where a tighter packing is observed. The percentage of water coupled to the adsorbed cationic polymer decreases with solution ionic strength. Overall, these observations are ascribed to the effects of electrostatic screening, polymer hydrodynamic size, and solvency.     

Chitosan-N-poly(ethylene oxide) brush polymers for reduced nonspecific protein adsorption

The possibility of using a novel comb polymer consisting of a chitosan backbone with grafted 44 units long poly(ethylene oxide) side chains for reducing nonspecific protein adsorption to gold surfaces functionalized by COOH-terminated thiols has been explored. The comb polymer was attached to the surface in three different ways: by solution adsorption, covalent coupling, and microcontact printing. The protein repellent properties were tested by monitoring the adsorption of bovine serum albumin and fibrinogen employing surface plasmon resonance and imaging null ellipsometry. It was found that a significant reduction in protein adsorption is achieved as the comb polymer layer is sufficiently dense. For solution adsorption this was achieved by adsorption from high pH solutions. On the other hand, the best performance of the microcontact printed surfaces was obtained when the stamp was inked either at low or at high pH. For a given comb polymer layer thickness/poly(ethylene oxide) density, significant differences in protein repellent properties were observed between the different preparation methods, and it is suggested that a reduction in the mobility of the comb polymer layer generated by covalent attachment favors a reduced protein adsorption.     

Fibronectin and bovine serum albumin adsorption and conformational dynamics on inherently conducting polymers: a QCM-D study

Quartz crystal microbalance with dissipation monitoring (QCM-D) was employed to characterize the adsorption of the model proteins, bovine serum albumin (BSA) and fibronectin (FN), to polypyrrole doped with dextran sulfate (PPy-DS) as a function of DS loading and surface roughness. BSA adsorption was greater on surfaces of increased roughness and was above what could be explained by the increase in surface area alone. Furthermore, the additional mass adsorbed on the rough films was concomitant with an increase in the rigidity of the protein layer. Analysis of the dynamic viscoelastic properties of the protein adlayer reveal BSA adsorption on the rough films occurs in two phases: (1) arrival and initial adsorption of protein to the polymer surface and (2) postadsorption molecular rearrangement to a more dehydrated and compact conformation that facilitates further recruitment of protein to the polymer interface, likely forming a multilayer. In contrast, FN adsorption was independent of surface roughness. However, films prepared from solutions containing the highest concentration of DS (20 mg/mL) demonstrated both an increase in adsorbed mass and adlayer viscoelasticity. This is attributed to the higher DS loading in the conducting polymer film resulting in presentation of a more hydrated molecular structure indicative of a more unfolded and bioactive conformation. Modulating the redox state of the PPy-DS polymers was shown to modify both the adsorbed mass and viscoelastic nature of FN adlayers. An oxidizing potential increased both the total adsorbed mass and the adlayer viscoelasticity. Our findings demonstrate that modification of polymer physicochemical and redox condition alters the nature of protein-polymer interaction, a process that may be exploited to tailor the bioactivity of protein through which interactions with cells and tissues may be controlled.     

Fibronectin displacement at polymer surfaces

The interactions of fibronectin with thin polymer films are studied in displacement experiments using human serum albumin. Fibronectin adsorption and exchange on two different maleic anhydride copolymer surfaces differing in hydrophobicity and surface charge density have been analyzed by quartz crystal microbalance and laser scanning microscopy with respect to adsorbed amounts, viscoelastic properties, and conformation. Fibronectin is concluded to become attached onto hydrophilic surfaces as a "softer", less rigid protein layer, in contrast to the more rigid, densely packed layer on hydrophobic surfaces. As a result, the fibronectin conformation is more distorted on the hydrophobic substrates together with remarkably different displacement characteristics in dependence on the adsorbed fibronectin surface concentration and the displacing albumin solution concentration. While the displacement kinetic remains constant for the strongly interacting surface, an acceleration in fibronectin exchange is observed for the weakly interacting surface with increasing fibronectin coverage. For displaced amounts, no change is determined for the hydrophobic substrate, in contrast to the hydrophilic substrate with a decrease of fibronectin exchange with decreasing coverage leading finally to a constant nondisplaceable amount of adsorbed proteins. Furthermore, the variation of the albumin exchange concentration reveals a stronger dependence of the kinetic for the weakly interacting substrate with higher rates at higher albumin concentrations.     

INTERACTIONS BETWEEN POLY(ETHYLENE OXIDE) COATED SURFACES AND BETWEEN SUCH SURFACES AND PROTEINS

Surfaces coated with poly(ethylene oxide) containing nonionic polymers are of interest in medical applications due to, among other things, the low adsorption of proteins on such surfaces. In this paper we have studied the interfacial properties of surfaces coated with PEO by measuring the forces acting between two such surfaces in water and across a protein solution as well as between one such surface and a surface carrying adsorbed proteins. One type of surface coating was a graft copolymer of poly(ethylene imine) and poly(ethylene oxide) where the cationic poly(ethylene imine) group anchored the polymer to negatively charged mica surfaces. Three different ways to prepare this coating was used and compared. It was found that this coating was not stable in the presence of lysozyme, a small positively charged protein, when the PEO graft density was low. The other type of coating was obtained by adsorbing ethyl(hydroxyethyl)-cellulose onto hydrophobised mica surfaces. The driving force for adsorption is in this case the hydrophobic interaction between nonpolar segments of the polymer and the surface. The EHEC coating was stable in the presence of lysozyme and the interactions between adsorbed layers of lysozyme and EHEC coated surfaces are purely repulsive due to long-range steric forces.     

Application of ion mobility spectrometry in study of surfactants adsorbed on different dish surfaces

Sodium lauryl sulfate (SLS) and sodium lauryl ether sulfate (SLES) are commonly used in many dishwashing liquids. These chemicals are adsorbed on the dish surface during the washing process and then transferred to food or drink in the cooking process. In this work, the adsorption of SLS and SLES on different dish surfaces in aqueous solution was studied. Stainless steel, copper, aluminum, Pyrex, Teflon and arcopal china ware were used in this study. The adsorbed chemical remained on the surface after rinsing was measured by thermal desorption using an ion mobility spectrometer as the detector. Although arcopal china ware showed the maximum amount of adsorption, and Pyrex and stainless steel dishes showed the minimum amount of residual chemical, the results showed that the amount of adsorbed chemicals on dish surfaces is less than 428 ng.cm^?2, which is well below the health risk dosage. The released SLS and SLES from dish surfaces into cold or hot water were also measured and compared for different dishes.     

Eluotropic strength in non-aqueous liquid chromatography with porous graphitic carbon

Protein adsorption can be either endothermic or exothermic depending upon the protein, the sorbent and process conditions. In the case of protein adsorption onto ion-exchange surfaces exothermic adsorption heats are usually characterized as representing the electrostatic interaction between two oppositely charged surfaces. Endothermic adsorption heats are typically characterized as representing protein reconfiguration and/or repulsive interactions between adsorbed molecules. In certain segments of the literature surface dehydration and solution non-idealities have been suggested as possible sources of endothermic heats of adsorption. Each of these phenomena was investigated during studies concerning the adsorption of bovine serum albumin and ovalbumin onto an anion-exchange sorbent. The results demonstrated that electrostatic repulsive interactions between adsorbed molecules appears to be a larger contributor to endothermic heats of adsorption than surface dehydration or solution non-idealities. The presence of mobile phase cations can reduce the magnitude of endothermic adsorption heats by screening repulsive interactions between adsorbed molecules. Although water release was not found to be a major contributor to endothermic adsorption heats, it is likely to be a contributor to the entropic driving force associated with the adsorption of bovine serum albumin.     

Cooperativity in the adsorption of cellulose ethers and fibrinogen at liquid-solid interfaces

The possibility of reducing fibrinogen adsorption to solid surfaces by competitive adsorption of cellulose ethers (EHEC, HEC) was investigated. The surface concentration of fibrinogen adsorbed onto hydrophilic and hydrophobic (methylized) glass was measured with an enzyme-linked immunosorbent assay. The immunoassay was calibrated by ellipsometry, using the initial mass transport limitation of adsorption for calculations of the maximum amount of adsorbed protein.At a hydrophobic surface, the presence of cellulose polymers caused a decrease of the adsorption of fibrinogen. The hydrophobic EHEC (cloud point 35°C) was most efficient and abolished surface-adsorption of the protein.At a hydrophilic surface, positive cooperativity was seen between fibrinogen and cellulose polymers. The hydrophilic HEC (cloud point >100°C) gave the most prominent effect.The results indicate that the affinity between a water soluble polymer or protein and a solid surface is not the only factor determining surface adsorption. The finding that there may be both positive and negative cooperativity in the adsorption of polymers shows the importance of polymer compatibility in layers of adsorbed polymers.     

Inorganic surface nanostructuring by atmospheric pressure plasma-induced graft polymerization

Surface graft polymerization of 1-vinyl-2-pyrrolidone onto a silicon surface was accomplished by atmospheric pressure (AP) hydrogen plasma surface activation followed by graft polymerization in both N-methyl-2-pyrrolidone (NMP) and in an NMP/water solvent mixture. The formation of initiation sites was controlled by the plasma exposure period, radio frequency (rf) power, and adsorbed surface water. The surface number density of active sites was critically dependent on the presence of adsorbed surface water with a maximum observed at approximately a monolayer surface water coverage. The surface topology and morphology of the grafted polymer layer depended on the solvent mixture composition, initial monomer concentration, reaction temperature, and reaction time. Grafted polymer surfaces prepared in pure NMP resulted in a polymer feature spacing of as low as 5-10 nm (average feature diameter of about 17 nm), an rms surface roughness range of 0.18-0.72 nm, and a maximum grafted polymer layer thickness of 5.5 nm. Graft polymerization in an NMP/water solvent mixture, however, resulted in polymer feature sizes that increased up to a maximum average feature diameter of about 90 nm at [NMP] = 60% (v/v) with polymer feature spacing in the range of 10-50 nm. The surface topology of the polymer-modified silicon surfaces grafted in an NMP/water solvent mixture exhibited a bimodal feature height distribution. In constrast, graft polymerization in pure NMP resulted in a narrow feature height distribution of smaller-diameter surface features with smaller surface spacing. The results demonstrated that, with the present approach, the topology of the grafted polymer surface was tunable by adjusting the NMP/water ratio. The present surface graft polymerization method, which is carried out under AP conditions, is particularly advantageous for polymer surface structuring via radical polymerization and can, in principle, be scaled to large surfaces.