pubs.acs.org/Langmuir © 2009 American Chemical Society
End Terminal, Poly(ethylene oxide) Graft Layers: Surface Forces and Protein Adsorption Paul Hamilton-Brown,†,§ Thomas Gengenbach,† Hans J. Griesser,‡ and Laurence Meagher*,†,§ †
CSIRO Molecular and Health Technologies, Bag 10, Clayton South, Victoria 3169, Australia, ‡Ian Wark Research Institute, University of South Australia, Mawson Lakes, South Australia 5095, Australia, and § Vision CRC, University of New South Wales, Kensington, New South Wales 2052, Australia Received February 27, 2009. Revised Manuscript Received May 26, 2009
Covalently grafted poly(ethylene oxide) coatings have been widely studied for use in biomedical applications, particularly for the reduction of protein and other biomolecule adsorption. However, many of these studies have not characterized the hydrated structure of the coatings. This new study uses a combination of silica colloid probe interaction force measurements using atomic force microscopy (AFM) and X-ray photoelectron spectroscopy (XPS) in order to determine the grafting density and hydrated layer structure of monomethoxy poly(ethylene oxide) aldehyde layers, covalently grafted onto amine plasma polymer surfaces, and their interactions with silica surfaces. For high grafting densities, purely repulsive interactions were measured as expected for densely grafted polymer brushes. These interactions could be described by theoretical expectations for compression of one polymer brush layer. However, at lower grafting densities, attractive interactions were observed at larger separation distances, originating from bridging interactions due to adsorption of the PEO chains on the surface of the silica colloid probe. This is a new finding indicating that the coupled PEO molecules have sufficient conformational freedom to interact strongly with an adjacent surface or, for example, protein molecules for which there is an affinity. The attractive interactions could be removed by grafting an additional PEO layer onto the silica colloid probe. Protein adsorption measurements confirmed that at high grafting densities, the amount of adsorbed protein on the PEO grafted surfaces was greatly reduced, to the order of the detection limit for the XPS technique.
Introduction Polymeric materials have been used for many years to fabricate biomedical devices such as ocular devices, catheters, heart valves, and other implantable devices. Adverse biological responses to these biomedical devices are often caused by the irreversible accumulation (fouling) of biological deposits (protein and lipid adsorption, bacterial adhesion, platelet deposition) onto the surfaces of these devices.1 This is a continuing problem as suitable methods (i.e., commercially applicable) of controlling bioadhesion have yet to be implemented; unfavorable effects range from inconvenient (fouling of contact lenses) to life threatening (thrombosis). The current hypothesis is that the formation of deposits is triggered by an initial stage in which various proteins and lipids rapidly adsorb onto the polymer surface; further biological molecules and ultimately cellular entities (e.g., bacteria) may then attach onto the initial adsorbed layer.2-4 Adsorption can be caused by attractive electrostatic or van der Waals interactions, or more specific interactions such as hydrogen bonding or those between ligands and receptors. *Author for correspondence. E-mail:
[email protected]; phone: +61 3 9545 8121; fax: +61 3 9545 2446. (1) Ratner, B. D.; Hoffman, A. S.; Schoen, F. J.; J.E., L., Biomaterials Science: An Introduction to Materials in Medicine; Academic Press: New York, 1996. (2) Horbett, T. A. Colloids Surf., B 1994, 2, 225–240. (3) Nakanishi, K.; Sakiyama, T.; Imamura, K. J. Biosci. Bioeng. 2001, 91, 233– 244. (4) Zhu, B.; Eurell, T.; Gunawan, R.; Leckband, D. J. Biomed. Res. 2001, 56, 406–416. (5) McArthur, S.; McLean, K.; Kingshott, P.St; John, H.; Chatelier, R.; Griesser, H. Colloids Surf., B 2000, 17, 37–48. (6) Osterberg, E.; Bergstr€om, K.; Holmberg, K.; Riggs, J. A.; Van Alstine, J. M.; Schuman, T. P.; Burns, N. L.; Harris, J. M. Colloids Surf., A 1993, 77, 159–169. (7) Perrino, C.; Lee, S.; Choi, S. W.; Maruyama, A.; Spencer, N. D. Langmuir 2008, 24, 8850–8856.
Langmuir 2009, 25(16), 9149–9156
Surface modification with polysaccharides,5-7 phospholipids,8-10 poly (ethylene oxide)11,12 (PEO) and other molecules has been used in an attempt to control bioadhesion. Of these, possibly the most effective method has been modification with PEO, although there are issues of oxidation on exposure to air,13 which could limit its application in biological environments. Covalently grafted PEO has been used extensively as a low-fouling coating due to its ability to significantly reduce protein adsorption onto surfaces.12,14,15 The ability of PEO coatings to reduce protein adsorption has been attributed to the molecule’s high mobility, which results in steric repulsion, while its neutral charge minimizes electrostatic interactions.16-21 Although there have been numerous studies published in the literature investigating the factors that determine the ability of grafted PEO layers to reduce protein adsorption, not all agree and at times appear contrary. (8) Feng, W.; Brash, J. L.; Zhu, S. P. Biomaterials 2006, 27, 847–855. (9) Feng, W.; Nieh, M. P.; Zhu, S.; Harroun, T. A.; Katsaras, J.; Brash, J. L. Biointerphases 2007, 2, 34–43. (10) Lewis, A. L. Colloids Surf., B 2000, 18, 261–275. (11) Kingshott, P.; Griesser, H. J. Curr. Opin. Solid State Mater. Sci. 1999, 4, 403–412. (12) Kingshott, P.; Thissen, H.; Griesser, H. J. Biomaterials 2002, 23, 2043– 2056. (13) Han, S.; Kim, C.; Kwon, D. Polymer 1997, 38, 317–323. (14) Pasche, S.; De Paul, S. M.; Voros, J.; Spencer, N. D.; Textor, M. Langmuir 2003, 19, 9216–9225. (15) Unsworth, L. D.; Sheardown, H.; Brash, J. L. Langmuir 2008, 24, 1924– 1929. (16) Halperin, A. Langmuir 1999, 15, 2525–2533. (17) Jeon, S. I.; Andrade, J. D. J. Colloid Interface Sci. 1991, 142, 159–166. (18) Jeon, S. I.; Lee, J. H.; Andrade, J. D.; de Gennes, P. G. J. Colloid Interface Sci. 1991, 142, 149–158. (19) Leckband, D.; Sheth, S.; Halperin, A. J. Biomater. Sci.;Polym. Ed. 1999, 10, 1125–1147. (20) Lee, J. H.; Lee, H. B.; Andrade, J. D. Prog. Polym. Sci. 1995, 20, 1043–1079. (21) Szleifer, I. Phys. A 1997, 244, 370–388.
Published on Web 06/17/2009
DOI: 10.1021/la900703e
9149
Article
The most significant factors in the reduction of protein adsorption appear to be grafting density and molecular weight,11,12,14,17-19,21,22 as well as hydration.15 The thickness of the neutral PEO layer must be enough so that electrostatic interactions from the underlying substrate are effectively screened by the surrounding ionic solution. Also, the mobility, and hence excluded volume, of the molecules is related to the size of the molecules and the distance between them, as suggested by de Gennes.23-25 As such, higher molecular weight, densely packed molecules should produce larger “steric barriers” to proteins adsorption than lower molecular weight molecules. Of particular importance is that the grafting density of the molecules should be high enough that the conformation of the PEO molecules is that of a polymer brush (i.e., where the distance between the molecules is less than the radius of gyration or Flory radius)26 in order to achieve a significant reduction in the amount of protein adsorbed. A vital piece of information lacking in many of these studies is direct measurement of the structure of the PEO layer (i.e., to determine whether the molecules are in the brush regime). This can only be achieved using techniques where the molecules are hydrated and where the measurement is sensitive to the structure of the PEO layer (i.e., techniques such as direct interaction force measurement and neutron reflectivity). There have been a number of studies investigating the interactions between PEO coatings, in particular the interaction of two PEO coatings;27-33 however, few of these coatings were covalently grafted to the substrate, a condition which is mandatory for biomedical applications. Most studies observed purely repulsive interactions as the PEO layers were compressed. More importantly, there has been only one study where one PEO layer and a different surface was investigated. In this case the PEO layer was not covalently grafted to the substrate, and the ionic strength of the solution was not physiologically relevant.28 In in vivo situations, it would be unlikely that a PEO surface layer would interact with another PEO layer. Therefore, it is of great importance to understand how grafted PEO layers interact with other surfaces and materials, particularly where there might be an affinity between them. For example, PEO molecules have a demonstrated affinity with human serum albumin when in solution.34 Therefore, this study aims to determine the factors affecting grafted PEO surface density and layer thickness during grafting reactions with different molecular weight PEOs in order to produce samples with defined surface grafting density, and to directly measure the interactions between such PEO layers so that an indication of the molecular conformation of the layers and/or affinity for other surfaces was obtained. This was achieved by (22) Unsworth, L. D.; Sheardown, H.; Brash, J. L. Langmuir 2005, 21, 1036– 1041. (23) Alexander, S. J. Phys. 1977, 38, 983–987. (24) de Gennes, P. G. J. Phys. 1976, 37, 1445–1452. (25) de Gennes, P. G. Adv. Colloid Interface Sci. 1987, 27, 189–209. (26) de Gennes, P. G. Macromolecules 1980, 13, 1069–1075. (27) Golander, C.-G.; Herron, J. N.; Lim, K.; Claesson, P.; Stenius, P.; Andrade, J. D., Properties of Immobilized PEG Films and the Interaction with Proteins. In Poly(ethylene glycol) Chemistry: Biotechnical and Biomedical Applications, Harris, J. M., Ed.; Plenum: New York, 1992; pp 221-246. (28) Heuberger, M.; Drobek, T.; Spencer, N. D. Biophys. J. 2005, 88, 495–504. (29) Kenworthy, A.; Hristova, K.; Needham, D.; McIntosh, T. J. Biophys. J. 1995, 68, 1921–1936. (30) Kuhl, T. L.; Leckband, D. E.; Lasic, D. D.; Israelachvili, J. N. Biophys. J. 1994, 66, 1479–1488. (31) McLean, S. C.; Lioe, H.; Meagher, L.; Craig, V. S. J.; Gee, M. L. Langmuir 2005, 21, 2199–2208. (32) Pasche, S.; Textor, M.; Meagher, L.; Spencer, N. D.; Griesser, H. J. Langmuir 2005, 21, 6508–6520. (33) Sheth, S. R.; Efremova, N.; Leckband, D. E. J. Phys. Chem. B 2000, 104, 7652–7662. (34) Ragi, C.; Sedaghat-Herati, M. R.; Ouameur, A. A.; Tajmir-Riahi, H. A. Biopolymers 2005, 78, 231–236.
9150 DOI: 10.1021/la900703e
Hamilton-Brown et al.
following surface grafting reactions by both X-ray photoelectron spectroscopy (XPS) and colloid probe, direct interaction force measurements using atomic force microscopy (AFM). The use of colloid-probe AFM in force mode allowed us to directly measure the magnitude of the forces between two surfaces and the distances over which they act. This in turn gives an indication (for purely repulsive interactions) of the layer thickness, grafting density, and structure. This information can be used to directly correlate the magnitude and range of the measured force to grafted surface densities and the range of the measured forces to coating thicknesses, thus providing information on the layer structure. Finally, we carried out protein adsorption measurement on some of the surfaces prepared to ensure that the surfaces prepared did indeed reduce protein adsorption significantly.
Materials and Methods Materials. Cleaned (see Supporting Information), ultra-flat single crystal, silicon wafers (Æ100æ, 1 cm20.5 mm thick, M.M.R.C Pty Ltd., Melbourne, Australia) were used as substrates for the deposition of radio frequency glow discharge (RFGD) thin films. All glassware used was cleaned in RBS surfactant solution and thoroughly rinsed in Milli-Q water. N-Heptylamine (HA) (99% purity) was obtained from Sigma Aldrich (Castle Hill, NSW, Australia) and used as received. Sodium chloride (AnalaR), sodium hydroxide (AnalaR), sodium dihydrogen orthophosphate, disodium hydrogen orthophosphate, and potassium sulfate were obtained from Merck Pty Ptd (Kilsyth, VIC, Australia). Monomethoxy poly(ethylene oxide) aldehyde (mPEOald) with molecular weights of 5000 and 20 000 g mol-1 and polydispersities in the range between 1.01 and 1.05 were purchased from Nektar Therapeutics (Huntsville, Al, USA) (see Table 1). Human serum albumin (HSA, fatty acid free, 99%) and fibrinogen (fraction 1, human plasma, 51% protein, 49% buffer salts) were purchased from Sigma-Aldrich (Castle Hill, NSW, Australia) and were used as received. All solutions were prepared using Milli-Q purified water. RFGD Thin Film Deposition. Radio frequency glow discharge (RFGD, also known as plasma polymerization) onto Si wafers was performed using a RFGD struck in the vapor of N-heptylamine (often referred to as a monomer) in a capacitively coupled custom-built reactor described elsewhere.36 Clean Si wafers were placed on the lower electrode in the vacuum chamber that was then evacuated to a base pressure of 0.13 Pa or less, in preparation for thin film deposition. Hereafter, these thin films will be referred to as HApp. The parameters used were: frequency = 200 kHz, load power = 20 W, and initial monomer pressure was 20 Pa. Treatment times were 25-30 s, resulting in HApp films of approximately 25 nm thickness.37 PEO Grafting Reactions. The covalent grafting of PEO molecules to amine-modified surfaces was performed via reductive amination under cloud-point conditions.12 The HApp modified surfaces were immersed in the buffered PEO solutions (0.5-4 mg/mL, pH 6.2, 11% w/v K2SO4) immediately following HApp deposition followed by addition of NaCNBH3 (3 mg/mL). Grafting of mPEOald to the surface was then carried out in a temperature-controlled water bath, at the cloud-point temperatures of the 5000 and 20 000 g mol-1 mPEOald solutions; at 60 and 37 C respectively. For experiments where the grafting time (35) Fleer, G. J.; Cohen Stuart, M. A.; Scheutjens, J. M. H. M.; Cosgrove, T.; Vincent, B. Polymers at Interfaces, 1st ed.; Chapman and Hall: London, 1993. (36) Griesser, H. J. Vacuum 1989, 39, 485. (37) Hartley, P. G.; Thissen, H.; Vaithianathan, T.; Griesser, H. J. Plasmas Polym. 2000, 5, 47–60.
Langmuir 2009, 25(16), 9149–9156
Hamilton-Brown et al.
Article
Table 1. Physical Parameters, Molecular Weight (MW), Radius of Gyration (Rg), Flory Radius (RF) and Contour Length (CL) of mPEOald Polymers Used in This Studya MW (Kg mol-1)
No. of monomers (N)
Rg (nm)
RF (nm)
CL (nm)
5.0 ( 0.25 114 ( 6.0 1.5 ( 0.2 6.0 ( 0.2 39.7 ( 2.0 20.0 ( 1.0√ √ 454 ( 23 3.0 ( 0.4 13.8 ( 0.4 158.9 ( 7.9 a Rg = (a N)/ 6, RF = aN3/5 and CL = aN where a is the monomer size (0.35 nm for ethylene oxide) and N is the number of monomers.35 b Values were calculated based on the manufacturers specification for polydispersity (e1.05). b
b
was varied, the surfaces were allowed to incubate between 1 and 45 h in a PEO solution of containing 2 mg/mL. For experiments where the solution concentration was varied, surfaces were incubated in solutions ranging from 0.5 to 4 mg/mL of PEO for 26 h. In some instances, a PEO layer was also grafted onto the colloid probe so that symmetric interaction force measurements could be carried out. After the grafting reactions were complete, the surfaces were rinsed by shaking in PBS for 3 days (PBS solution was changed 3 times) and then in Milli-Q water overnight to remove salts from the PBS solution. The surfaces were then dried in a nitrogen stream and used immediately or stored in MilliQ water until required. XPS Surface Analysis. XPS analysis was performed on all samples using an AXIS HSi spectrometer (Kratos Analytical Ltd., UK) equipped with a monochromated Al KR source at a power of 144 W. Typical chamber pressure for all experiments was of the order of 10-6 Pa. Surfaces were dried in a nitrogen stream prior to being mounted on XPS sample holders. Samples were analyzed at an emission angle of 0 wrt the surface normal. Elements present on the surface were identified from survey spectra (320 eV pass energy), and spectra for individual peaks were recorded for further analysis and quantification (80 eV pass energy). High-resolution scans were also recorded using 40 eV pass energy. Atomic concentrations were calculated from the integral peak intensities of O 1s, N 1s, C 1s, and Si 2p spectra, using a linear background and applying sensitivity factors supplied by the instrument manufacturer (average of two positions for the PEO grafting optimisation study and an average of six positions (three spots on two identical samples) for the protein adsorption study). Statistical analysis of the data was carried out via a two-tailed t test with equal variances. A value of 285 eV for the binding energy (BE) of the main C 1s component (aliphatic CHx) was used to correct for charging of specimens under irradiation.38 Five components38,39 were fitted to high-resolution C 1s spectra. Component C1 at the lowest binding energy (BE = 285.0 eV) was assumed to represent aliphatic hydrocarbons (“neutral” carbon). A second component (C2) at a slightly higher BE was included to account for all C 1s photoelectrons that display a secondary BE shift. Component C3 at 286.3-286.6 eV represents C-N and C-O based groups (e.g., amines, ethers and alcohols), C4 at 287.9-288.2 eV accounts for all CdO and O-C-O based groups (e.g., carbonyls, amides), and C5 at 288.9-289.3 eV represents O-CdO based groups (e.g., acids or esters). Further details of the XPS fitting procedures can be found in the Supporting Information. The thickness of the dry PEO graft layers was estimated from the attenuation of the N 1s XPS signal from the HApp substrate (38) Beamson, G.; Briggs, D., High Resolution XPS of Organic Polymers: The Scienta ESCA300 Database, 1st ed.; John Wiley & Sons Ltd: New York, 1992. (39) Gengenbach, T.; Chatelier, R. C.; Griesser, H. J. Surf. Interface Anal. 1996, 24, 271–281. (40) Tielsch, B. J.; Fulghum, J. E. Surf. Interface Sci. 1994, 21, 621–630.
Langmuir 2009, 25(16), 9149–9156
by the PEO overlayer.40 The calculated thickness was converted to mass assuming a FPEO of 1.12 gcm-3 and then to molecules per unit area using the molecular weight of the polymer chains (5000 and 20 000 g mol-1). The area per molecule was estimated assuming a circular cross sectional area. AFM Colloid Probe Force Measurements. Interaction force measurements were carried out using a Nanoscope III atomic force microscope (Digital Instruments Inc., Santa Barbara, CA, USA). The measurement of interaction forces between grafted PEO surfaces or HApp surfaces and silica was performed in a liquid cell using solutions of physiological pH and ionic strength, using the colloid probe method developed by Ducker et al.41,42 The silica colloid particles were prepared using a modified Stober process.43,44 The diameters of the spheres used were measured using light microscopy and ranged between 4 and 5 μm in diameter. The cantilevers’ spring constants were determined using the resonance method established by Cleveland et al.45 Values of 0.0916 and 0.330 N m1-, with 18% variation, were calculated for 200 and 100 μm long cantilevers, respectively (average of 14 cantilevers). Cleaning procedures used for silica spheres, fluid cells, etc., are described fully in the Supporting Information. After injection of a 0.15 M NaCl solution, the surfaces were left to equilibrate (20 min) before force curve acquisition. Approximately 15 force curves were obtained from each of 5-6 regions on each surface. At least five force curves were scaled using custom designed software,46 and a representative, scaled force curve was chosen for inclusion in Figures. Fitting of repulsive forces arising from the compression of one PEO layer was carried out using eqs 1 and 2, which are derived from the theory of Milner et al.29,31,32,47,48 This theory includes a parabolic segment density profile for grafted polymer brushes and is therefore more applicable than the scaling theory of de Gennes,23,24 which assumes a uniform segment density as a function of distance from the surface. Z FðDÞ ¼ 2πEðDÞ ¼ -2π pðDÞ dD R " # 2 5 L0 D D 9 þ ¼ 4πP0 L0 L0 5 D
kTN π2 P0 ¼ 2 12
!1=3
a4=3 s10=3
ð1Þ
ð2Þ
where F is the force between the surfaces as a function of the distance (D), R is the radius of the spherical surface, L0 is the equilibrium layer thickness, P0 is the pressure between the surfaces, k is the Boltzmann constant, T is the temperature, N is the number of segments in the PEO chain, a is the size of the ethylene oxide segment (0.35 nm), and s is the distance between the grafted chains. (41) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Nature 1991, 353, 2239–2241. (42) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Langmuir 1992, 8, 1831–1836. (43) Barder, T. J.; DuBois, P. D. US Patent No. 4,983,369, 1991. (44) Stober, W.; Fink, A.; Bohn, E. J. Colloid Interface Sci. 1968, 26, 62. (45) Cleveland, J. P.; Manne, S.; Bocek, D.; Hansma, P. K. Rev. Sci. Instrum. 1993, 64, 1. (46) Chan, D. Y. C.; Ip, L.; Venters, S. AFM Analysis, Version 2; Department of Applied Mathematics and Statistics: University of Melbourne, 1994. (47) Milner, S. T.; Witten, T. A.; Cates, M. E. Macromolecules 1988, 21, 2610– 2619. (48) Milner, S. T.; Witten, T. A.; Cates, M. E. Europhys. Lett. 1988, 5, 413–418.
DOI: 10.1021/la900703e
9151
Article
Hamilton-Brown et al.
To study end-grafted mPEOald layers with different densities, grafting was performed under different conditions to produce a range of samples for subsequent study. As the grafting density could not be predicted, XPS was used to determine the grafting densities obtained under the various conditions. XPS data derived from analysis of samples with grafted 5000 and 20 000 g mol-1 PEO layers after various grafting reaction times are presented in Table 2. For both PEO molecular weights, the O atomic % increased and the N atomic % decreased with increasing grafting reaction time. These changes in atomic composition are consistent with the presence of a PEO overlayer that increased in thickness with reaction time (PEO has an O/C ratio of 0.5 and contains no nitrogen). This observation is reflected in the calculated overlayer thickness values (TOL) also presented. A calculated estimation of the distance between grafting sites (sOL) is also given, which has the converse trend (i.e., the sOL value decreased with increasing grafted amount) indicating that the two parameters are inter-related as expected. Addition of coupled molecules of the same molecular weight can only result in a reduced distance between grafting sites. Differences between the atomic compositions obtained for the two different molecular weights were as follows. The O atomic %, values were higher for the 20000 g mol-1 PEO samples, and the N atomic % decreased by a larger amount, indicating that more mass of PEO was coupled in the 20 000 g mol-1 PEO case by the end of the grafting reaction. This is reflected by larger TOL values. In addition, the calculated sOL values were lower, as might be expected for reactions where larger molecules were coupled to surfaces. The C 1s high resolution spectra obtained from analysis of these samples are presented in Figure 1 for (a) 5000 and (b) 20 000 g mol-1 PEO, respectively. In both cases the relative intensity of the C3 component at a binding energy of ∼286.5 eV increased with increasing coupling reaction time, consistent with the coupling of PEO to the surface (component C3 represents C-O functionality). These data reflect the changes observed in the atomic composition, that is, more PEO was coupled onto the surfaces as the reaction time was increased. The observed change in chemical surface composition was more pronounced in the case of coupling of the 20 000 g mol-1 PEO (see Figure 1b), confirming that more PEO was coupled to the surface in the case of the 20 000 g mol-1 PEO. The changes observed in Figure 1 are quantified in Table 3, which presents the high resolution C 1s
component quantification obtained from curve-fitting all the spectra obtained. Presented in Figure 2 are interaction force measurements (F normalized by the radius of the silica particle as a function of apparent separation distance) obtained in 0.15 M NaCl solution between silica colloidal particles and grafted 5000 (a) and 20 000 (b) gmol-1 PEO layers, respectively, as a function of grafting reaction time. At this ionic strength, electrostatic interactions are virtually completely screened (Debye length, κ-1 =0.8 nm). For the lower molecular weight case, the forces measured at lower grafting reaction times (1, 2 h) were attractive at larger separations and became repulsive at shorter separations (
