Surface Modification of Polymers with Self-Assembled Molecular

Biomacromolecules 2000, 1, 139-148

139

Surface Modification of Polymers with Self-Assembled Molecular Structures: Multitechnique Surface Characterization† Connie S. Kwok,‡ Pierre D. Mourad,§ Lawrence A. Crum,§ and Buddy D. Ratner*,‡ Department of Bioengineering and University of Washington Engineered Biomaterials and Applied Physics Laboratory, University of Washington, Seattle, Washington 98195 Received January 20, 2000

A simple, one-step procedure for generating ordered, crystalline methylene chains on polymeric surfaces via urethane linkages was developed. The reaction of dodecyl isocyanate with surface hydroxyl functional groups, catalyzed by dibutyltin dilaurate, formed a predominantly all-trans, crystalline structure on a crosslinked poly(2-hydroxyethyl methacrylate) (pHEMA) substrate. Allophanate side-branching reactions were not observed. Both X-ray photoelectron spectrocopy and time-of-flight secondary ion mass spectrometry show that the surface reaction reached saturation after 30 min at 60 °C. Unpolarized Fourier transform infrared-attenuated total reflection showed that, after 30 min, the stretching frequencies, νCH2,asym and νCH2,sym, decreased and approached 2920 and 2850 cm-1, indicative of a crystalline phase. The distance between two hydroxyl groups is roughly 4 Å. A tilt angle of 33.5° ( 2.4° was estimated by dichroic ratios measured in polarized ATR according to the two-phase and Harrick thin film approximations. The findings reported here are significant in that the possibilities for using structures similar to self-assembled monolayers (SAMs) are expanded beyond the rigid gold and silicon surfaces used through most of the literature. Thus, SAMs, biomimetics for ordered lipid cell wall structures, can be applied to real-world biomedical polymers to modify biological interactions. The terminal groups of the SAM-like structure can be further functionalized with biomolecules or antibodies to develop surface-based diagnostics, biosensors, or biomaterials. 1. Introduction Self-assembled monolayers (SAMs) with different terminal groups have stimulated interest in several areas of bioengineering, including biosensors, biomaterials, and biomimetics. Their well-defined structures and intrinsic stability are conducive to precision molecular engineering and quantitative analysis. These monolayers have been formed mostly on metal surfaces, such as on gold using thiol chemistry 1,2 or on oxidized glass/silicon surfaces using a siloxy linkage.3,4 Others have focused on derivatives of fatty acids or bifunctional amphiphiles using the LangmuirBlodgett (LB) technique.5,6 Although these studies provide excellent models for exploring the self-assembly mechanisms of various long alkyl chain molecules, few studies report on using polymeric surfaces as the substrates. Because a large proportion of medical devices are made from polymeric materials, successful transition of SAM structures from model surfaces to real-world polymers is desirable. To date, there is only a limited number of publications addressing the construction of SAMs on polymers. Sagiv7 first described the covalent binding of octadecyl trichlorosilane onto spincast poly(vinyl alcohol) and later reported on synthesizing corresponding multilayers8,9 based on the previous successful * To whom correspondence should be addressed. Tel.: 206-685-1005. Fax: 206-616-9763. E-mail: [email protected]. † Presented in part at the 25th annual meeting of the Society for Biomaterials at Providence, RI, 1999. ‡ Department of Bioengineering and University of Washington Engineered Biomaterials. § Applied Physics Laboratory, University of Washington.

results. Bohme et al.10 used bola-amphiphile molecules for the preparation of LB monolayers on the surface of spincast poly(allylamine hydrochloride) film. Whitesides and coworkers systematically attempted to form SAMs on oxidized poly(dimethyl siloxane) and polyethylene slab surfaces pretreated with an oxygen plasma.11-13 Their strategy to modify these surfaces involved multistep reactions. Comparing data from X-ray photoelectron spectroscopy (XPS) and contact angle measurements with the data from a model gold surface, they concluded that dense and ordered monolayers were formed with alkyl chains in an all-trans configuration on these oxidized surfaces. To expand the repertoire of methods for preparing SAMs on real-world polymers used in medical devices, we sought to construct SAM structures on a polymeric biomaterial, poly (2-hydroxyethyl methacrylate) (pHEMA),14,15 widely used in contact lenses. Because of the presence of a hydroxyl group in the side chain of the polymer, various modifications of pHEMA using its primary alcohol are possible and provide a wide range of pHEMA derivatives for various biomedical applications.16,17 In particular, we were interested in surface derivatizing the polymer using isocyanate chemistry to form urethane linkages catalyzed by dibutyltin dilaurate (Figure 1). This catalyst is believed to minimize the undesirable allophanate side reaction.18 A more complete overview of isocyanate chemistry is described elsewhere.19,20 The onestep procedure developed here is not limited to molecularly smooth surfaces and can be applied to many hydrogels and functional polymers. In this paper, we report that a SAM

10.1021/bm000292w CCC: $19.00 © 2000 American Chemical Society Published on Web 03/14/2000

140

Biomacromolecules, Vol. 1, No. 1, 2000

Kwok et al.

Figure 1. Schematic of proposed surface derivatization of hydroxyl groups on a pHEMA surface with dodecyl isocyanate to form an ordered, assembled monolayer.

structure was successfully bound to a pHEMA slab surface via covalent urethane bonding, verified by XPS, time-offlight secondary ion mass spectrometry (TOF-SIMS) and Fourier transform infrared-attenuated total reflection (FTIRATR). Furthermore, the absence of the undesirable allophanate side reaction, resulting in layer branching, was assessed by infrared (IR) and TOF-SIMS analysis. Last, the molecular orientation of the methylene chains, described by the tilt angle, was estimated by both the two-phase and Harrick thin film approximations. 2. Materials and Methods 2.1. Materials. 2-Hydroxyethyl methacrylate (HEMA, no. 04675) monomer with a purity of more than 99.5% and tetraethylene glycol dimethacrylate (TEGDMA, no. 02654) were purchased from Polysciences Inc., Warrington, PA. Ethylene glycol (no. 32,455-8), sodium metabisulfite (no. 16,151-9), ammonium persulfate (no. 24,861-4), anhydrous tetrahydrofuran (THF, no. 40,175-7), dodecyl isocyanate (C12 isocyanate, no. 29,123-4), and dibutyl tin dilaurate (no. 38,906-4) were received from Aldrich, Inc. All chemicals were used as received. Glass plates and glass apparatus for synthesis were soaked in 2% RBS-35 detergent (no. 27950, Pierce) overnight and rinsed with Millipore purified water prior to the experiments. 2.2. Sample Preparation and Synthesis. 2.2.1. Preparation of the Polymeric Substrate (pHEMA). Cross-linked hydrogel slabs were synthesized from HEMA using the procedures previously described.21,22 Briefly, 5 g of HEMA monomer and 0.2 g of the TEGDMA cross-linking agent were added to a water/ethylene glycol mixed solvent (1 g/1.5 g) with 1 mL of 15% sodium metabisulfite and 40% ammonium persulfate as redox initiators to begin the radical polymerization. The mixture was allowed to polymerize between two clean glass plates with a Teflon gasket of thickness 0.025 in. Although the gel set within an hour, the film was allowed to stand overnight. The pHEMA film was released from the glass plates and soaked in distilled water for a few days to leach out unreacted monomers, initiators, and oligomer residues. To speed the leaching process, later films were soaked in water for only 1 day, consistent with the method used by Brynda et al.23 who reported that most impurities were washed out within the first few hours. After leaching, the pHEMA film was cut into smaller specimens

for surface modification with C12 isocyanate. The pHEMA samples must be vacuum-dried prior to surface derivatization because water molecules easily terminate the urethanelinkage reaction between the hydroxyl group on the pHEMA surface and the isocyanate on the C12 compound. 2.2.2. Preparation of Ordered Methylene Chains on pHEMA. In a three-necked round-bottom flask connected to a nitrogen gas line, ≈4.5 mL of C12 isocyanate (5%) and 0.18 mL of dibutyl tin dilaurate (as the catalyst, 0.3%) were added to 90 mL of anhydrous tetrahydrofuran containing the dry polymer films. In this case, the choice of the reaction medium is important. Tetrahydrofuran is a poor swelling solvent for pHEMA which prevents polymer swelling and optimizes the surface immobilization reaction of hydrocarbon chains to the gel slab. To further optimize the reaction conditions, temperature and reaction time were studied. The reaction was performed under a nitrogen atmosphere at 40, 50, or 60 °C in an oil bath. At each temperature, the reaction was allowed to run for 5, 15, 30, 45, and 60 min. At each time point, one pHEMA sample was retrieved from the reaction flask and sonicated (43 kHz, L&R model T21) in fresh THF for 5 min to remove physically adsorbed C12 isocyanate. Following sonication, the surface-derivatized films were blown dry with nitrogen for surface characterization. 2.3. Surface Characterization. Surface-derivatized films were examined by a number of surface characterization techniques. XPS was used to measure the chemical composition and functional groups on the surface,24,25 TOF-SIMS to study the molecular fragments that were chemically bonded to the surface,26-29 FTIR-ATR to investigate the chain order and crystalline structure, and polarized ATR to estimate the molecular chain orientation.30-34 2.3.1. X-ray Photoelectron Spectroscopy. XPS, also known as electron spectroscopy for chemical analysis (ESCA), was performed with an S-Probe surface analysis system (Surface Science Instruments, Mountain View, CA) using a monochromatic Al KR1,2 X-ray source to stimulate photoemission. The system consists of a 30° solid angle acceptance lens, a hemispherical analyzer, and a positionsensitive detector. All polymer samples were analyzed at a 55° takeoff angle, probing the uppermost 50-80 Å of the surface. The takeoff angle was defined as the angle between the surface normal and the axis of the analyzer acceptance lens. Survey scans (0-1000-eV binding energy) were run

Surface Modification of Polymers with SAMs

at an analyzer pass energy of 150 eV (resolution 4) with an X-ray spot size of 1000 × 1700 µm to determine the elemental composition of each surface. High-resolution O(1s), C(1s), and N(1s) scans were obtained at a pass energy of 50 eV (resolution 2). The high-resolution spectra were resolved into individual Gaussian peaks using a least-squares fitting routine in the SSI software. The chemical composition of each surface was determined from the peaks resolved in the high-resolution scans. All binding energies (BEs) were referenced by setting the maximum of the resolved C(1s) peak, corresponding to carbon in a hydrocarbon environment (CHx), to 285.0 eV. When the binding energy referencing was performed in the same manner, the primary O(1s) peak was found to be shifted to 532.8 eV, the expected value for oxygen in an ether environment in polymers. A 5-eV electron flood gun was used to minimize surface charging. Typical pressures in the analysis chamber during spectral acquisition were 10-9 Torr. XPS analysis was carried out on all coated and derivatized samples immediately after reaction, unless otherwise specified. 2.3.2. Time-of-Flight Secondary Ion Mass Spectroscopy. TOF-SIMS was performed on a model 7200 Physical Electronics PHI instrument (Eden Prairie, MN). An 8-keV Cs+ ion source was used for surface impingement, a lowenergy pulsed electron flood gun for charge neutralization and a chevron-type multichannel plate (MCP) for secondary ion detection. Data were acquired over a mass range from m/z ) 0-1000 for both positive and negative secondary ions. The mass scale for the negative secondary ions was calibrated using peaks originating from HOCHdCHO- (m/z ) 59-), CH3C(CH2)COO- (m/z ) 85-), and C13H26NO2- (m/z ) 228-). The first two were fingerprints originating from the pHEMA substrate while the third one resulted from the C12urethane fragment of the modified surface (see Figure 6). Both spectroscopic and imaging modes were employed. Images and spectra were analyzed with TOF-Pak software packages provided by Physical Electronics. 2.3.3. Fourier Transform Infrared Spectroscopy-Polarized Attenuated Total Reflection Mode. FTIR-ATR measurements were performed in the mid-IR frequency range (4000-400 cm-1) on a Digilab FTS-60A spectrometer equipped with a liquid nitrogen-cooled mercury-cadmiumtelluride detector. The system was purged with water and carbon dioxide free air. An internal reflection cell with a variable-angle holder (model 300, SpectraTech CT) was used to measure FTIR-ATR spectra. The incidence angle for internal reflection was calculated using the procedure described in the model 300 manual and was set to 45°. Either unpolarized or s- and p-polarized light was used. For polarized spectra, the KRS-5 grid polarizer (SpectraTech) was set to 0° and 90°, giving absorbance perpendicular (s) and parallel (p) to the plane of incidence, respectively. The IR beam was aligned and calibrated prior to measurements. The background spectrum for all ATR measurements was the single-beam spectrum of the clean KRS-5 element (parallelogram, 50 × 10 × 3 mm, 45°) and the absorbance spectra of the modified films were taken as a ratio against the background spectrum of a clean KRS-5 element. All data were taken at 4-cm-1 resolution with a total of 64 scans to

Biomacromolecules, Vol. 1, No. 1, 2000 141

Figure 2. XPS C(1s) of pHEMA surface-derivatized with TFAA. (a) Schematic of the reaction; (b) pHEMA after reaction; (c) poly(vinyl alcohol) positive control; (d) polystyrene negative control.

achieve adequate signal/noise. Spectra were baseline-corrected, and both asymmetric (2850 cm-1) and symmetric (2920 cm-1) CH2 stretching frequencies were normalized against internal CdO (aliphatic ester at 1720 cm-1) to obtain the polarized dichroic ratio (i.e., Ap/As). The ratio was later used to calculate the molecular orientation of methylene chains using the two-phase approximation and the Harrick thin film approximation. 3. Results 3.1. XPS Analysis of Surface Chemical Composition and Functional Groups. Prior to performing the dodecyl isocyanate reaction, we asked if surface hydroxyl groups on the dry pHEMA film were available for reaction. A straightforward method is to use trifluoroacetic anhydride (TFAA) as a chemical marker following the protocol developed by Chilkoti et al.28 TFAA reacts with hydroxyl groups according to the reaction scheme in Figure 2a. The XPS C(1s) spectrum in Figure 2b shows that there were surface hydroxyl groups available for reaction and that they readily reacted with TFAA stoichiometrically. Poly(vinyl

142


Figure 3. Temperature effect on C12-isocyanate-pHEMA reaction depicted by XPS C/O ratio (a) and N% (b) from XPS surface analysis.

alcohol) and polystyrene were used as positive and negative controls (Figure 2c,d). With the encouraging results from the TFAA experiments, dry pHEMA films were reacted with dodecyl isocyanate at various temperatures (40, 50, and 60 °C) for various times (0 f 60 min) to explore conditions for maximum surface coverage. Dibutyl tin dilaurate was used as a catalyst to accelerate the reaction. At each temperature, the C/O ratio increased with increasing reaction timesthe ratios reached plateaus after 30 min for 50 and 60 °C (Figure 3a). This suggested that, at these two temperatures, all surface hydroxyl groups reacted and a longer reaction time did not allow increased binding. Another alternative was to use nitrogen as an indicator of the reactions, as nitrogen can be attributed to the postreaction urethane linkage. Consistent with the C/O ratio in Figure 3a, the fraction of nitrogen on the surface also increased with increasing reaction time and plateaued after 30 min for both 50 and 60 °C (Figure 3b). Typical XPS C(1s) and O(1s) spectra of the modified surfaces with increasing reaction times are shown in Figure 4. Because of the methylene chain addition via urethane bonding, Figure 4a shows an increased CHx peak amplitude, a decreased C-OH/C-O peak (because of the reaction of hydroxyl groups with isocyanate), and broadening of the O-CdO peak (because of the incorporation of urethane NH-O-CdO into the substrate ester O-CdO). Added significance for this observation was gleaned from the corresponding O(1s) spectra. Figure 4b clearly illustrates that the broad but symmetrical O(1s) peak commonly observed for pHEMA split into two distinct subpeaks represented by the two types of oxygens in the urethane environment. The pronounced cleavage at 533.1 eV undoubtedly resulted from the disappearance of hydroxyl groups after reaction. 3.2. Molecular Fragments Identified by TOF-SIMS. Both positive and negative scans were performed on uncoated

Kwok et al.

Figure 4. Typical XPS C(1s) (a) and O(1s) (b) spectra of derivatized C12-pHEMA with different reaction times at 60 °C. Control # 1 is a negative control in which pHEMA samples were put in the reaction mixture without the addition of catalyst.

and derivatized samples. In this case, the negative ion spectra gave relatively more interesting results. Typical negative ion spectra at different reaction times at 60 °C are shown in Figure 5. Because of the many carbon-oxygen bonds in pHEMA, a preferred C-O cleavage was responsible for stable negative ion fragmentation (Figure 6). The m/z ) 228negative ion appeared after 5 min of reaction, while the intensity of ions from the pHEMA backbone decreased with increasing reaction time. Because the absolute SIMS peak intensities are not useful for performing quantitative studies, we must normalize the peak intensity, assuming that the relationship between the numbers of molecules or structures at the surface and the relevant ions detected in the spectra is linear.27,35 In this report, relative intensities were calculated by normalizing individual peak intensities with respect to selected total ion intensities. The total ion intensity was computed by summing up the peak intensities of selected ions that were relevant to the pHEMA substrate and modified surface (mainly nitrogen-containing fragments). Negative peak ions of m/z ) 59-, 85-, 113-, 127-, and 139- were selected27,35-40 for the pHEMA. In particular, m/z ) 59- is a fingerprint negative ion for pHEMA polymer39 while others are common to many methacrylate polymers. For the alkyl isocyanate-modified surface, mostly negative nitrogencontaining peaks were included. They were CN- (26-), CH2N- (28-), CNO- (42-), OCON(CH2)- (72-), and C12urethane (228-) fragments along with some simple atomic peaks such as Si- (28-) and O2- (32-). The plot of relative peak ion intensities versus reaction time is illustrated in Figure 7. It shows that the relative intensities of pHEMArelated negative ions decreased with increasing reaction time while the opposite trend held for nitrogen-containing ions from the modified surface. Relative peak intensities in both part a and b of Figure 7 reached plateaus (see trend lines in



Figure 5. Typical negative ion SIMS spectra of uncoated and derivatized pHEMA surfaces at different reaction times at 60 °C.

Figure 6. Chemical structure of derivatized pHEMA with C12 isocyanate via a urethane linkage. Mass/charge ratios for some selected negative molecular fragments are shown.

the graphs) after 30 min, consistent with XPS findings, suggesting that maximum surface coverage was achieved after 30 min. To summarize the results, surface coverage as a function of reaction time was studied. It is common to take the ratio of the peak intensity of fragments from a modified surface with that from an uncoated control surface.27,35,36 The C12urethane molecular ion at m/z ) 228- was a good indicator

for the derivatized surface because it was the urethane linkage that bound the C12 methylene chains to the polymer surface. On the other hand, the peaks at m/z ) 59- and 85-, which were pertinent to the pHEMA substrate, were taken as reference peaks because they were most intense in the spectrum of uncoated pHEMA as well as in the spectra of modified pHEMA. Figure 8 summarizes the relationship between the surface coverage, represented by the peak intensity ratio of 228/(59 + 85), with the reaction time, and recapitulates the consistent findings from both XPS and SIMS. To visualize the spatial uniformity of C12-urethane fragments on the modified surface, a study using SIMS in the imaging mode was employed and is shown in Figure 9. Negative C12-urethane ions were mapped and assigned as white color. This visibly demonstrates the densely packed C12 chains at 30-min reaction time, compared to the uncoated pHEMA control. Preliminary studies were also performed to check the possibility of allophanate formation during the surface reaction. Allophanate is formed by an attack on the active

144


Kwok et al.

Figure 9. TOF-SIMS images of C12-urethane negative fragments (indicated as white dots) observed on the derivatized sample surface (b). No C12-urethane ions were found on the uncoated pHEMA control (a). The reaction was performed at 60 °C for 30 min.

Figure 7. Normalized peak intensities of various representative negative molecular ions originating from (a) substrate pHEMA and (b) derivatized C12-pHEMA surface-nitrogen-containing fragments. Various reaction times at 60 °C were studied.

Figure 10. Unpolarized IR absorbance spectra of the derivatized C12-pHEMA sample treated for various reaction times at 60 °C. (a) Low-frequency region: CdO, C-NH, and CH2 methylene scissoring vibration frequencies. (b) High-frequency region: OH and NH and CH2 (asym, sym) stretching frequencies. Figure 8. Summary of the effect of reaction time on C12-urethane formation at 60 °C.

hydrogen in the urethane bond with a free isocyanate molecule.41 Allophanate formation would lead to disorganized, multilayer films. If this reaction had taken place, a peak at m/z 439.3899476 (C26H51N2O3-) might be observed. However, as no such peak was identified (data not shown) in the mass range m/z ) 0-1000, we believe the allophanate side reaction was repressed. This lends strength to an ordered model for the nature of this surface modification. 3.3. Molecular Orientation Provided by Polarized FTIR-ATR. 3.3.1. Unpolarized ATR. Figure 10 shows the unpolarized IR spectra of dry, derivatized-pHEMA films at different reaction times. Consistent with the conclusion from SIMS, characteristic IR bands for allophanate were not observed for the derivatized samples. Allophanates are

normally characterized by a triplet of intense bands at 1220, 1280, and 1310 cm-1 associated with the skeletal vibrations of the allophanate group as well as unique NH bands at 3298, 3267, and 3233 cm-1.18 The absence of these bands in the IR spectra demonstrated that undesirable allophanate formation had not occurred. Furthermore, the absence of the NCO band at 2270 cm-1 suggested that free isocyanates were not physisorbed to the surface after the 5-min sonication. In the IR spectra shown in Figure 10a, several important features were noted. First, the absorbance of the CNH stretching peak at 1530 cm-1, due to the secondary urethane,42 increased with increasing reaction time and so did the methylene scissoring vibration frequency at 1453 cm-1. Second, two carbonyl stretching frequencies in the range 1705-1750 cm-1 showed a reversal of intensities: the IR absorbance of the aliphatic ester (O-CdO; 1720 cm-1) from



the pHEMA substrate decreased and that of secondary urethane amide (1705 cm-1) from the C12-derivatized surface increased with increasing reaction time. Third, the strong, broad OH band at 3300 cm-1 also narrowed and exhibited similarity to an NH stretching peak. All lines of evidence suggested that a successful attachment of C12 methylene chains via covalent urethane bonding was achieved. Another important observation is associated with the frequency change of the CH2 stretching frequencies at 2850-2920 cm-1. Figure 10b clearly demonstrates that the frequencies of the νCH2,asym and νCH2,sym bands decreased and approached 2920 and 2850 cm-1 with increased amplitudes at longer reaction times (0 f 30 min). After a 30-min reaction time, both the νCH2,asym and νCH2,sym stretching frequencies were found to be stabilized at 2920 and 2851 cm-1, typical peak frequencies of CH2 units in the trans state,31 indicating the immobilized methylene chains ordered themselves in a crystalline structure.43-45 The increased amplitudes of the CH2 IR peaks along with other observations consistently pointed to increased surface coverage at longer reaction times. 3.3.2. Polarized ATR. s-Polarized and p-polarized ATR spectra of a dry, 30-min-derivatized pHEMA film on a KRS-5 element are shown in Figure 11 over the region 3050-2750 cm-1. CH2,asym and CH2,sym frequencies at 2920 and 2850 cm-1 are easily distinguished in the spectra. To determine the orientation of the axis of the methylene chain with respect to the normal to the ATR crystal, that is, the tilt angle (θ), the dichroic ratio (Ap/As) can be calculated from the polarized spectra according to linear dichroism theory,31 where Ap and As are the absorbances of the bands obtained with infrared radiation polarized parallel and perpendicular to the plane of incidence, respectively. Because absorbance strongly depends on the pressure with which the sample is flattened to the ATR crystal and it may vary from experiment to experiment, the aliphatic ester (O-CdO) stretching frequency at 1720 cm-1 was chosen as the internal reference and used to normalize the observed absorbances of all CH2 stretching frequencies. The calculated dichroic ratios for CH2,asym (2920 cm-1) and CH2,sym (2850 cm-1) are 1.004 and 0.967, respectively. 4. Discussion An organized, ordered structure of methylene chains, similar to a self-assembled monolayer (SAM), was successfully synthesized on pHEMA cross-linked gel surfaces. Although several research groups had shown promising results on constructing SAMs on polymeric surfaces, they were limited by the alkyl siloxane chemistry on poly(dimethyl siloxane) substrates12,46,47 or they required multistep derivatization procedures involving plasma pretreatment.11,12 The procedure reported here was a simple, one-step modification of a polymer surface using hydroxyl groups on pHEMA as the reactive sites and dibutyl tin dilaurate as a catalyst. Using a stannous catalyst,18 we demonstrated that it could prevent the allophanate branching side reaction. Anhydrous tetrahydrofuran, a poor solvent for pHEMA yet an excellent organic medium for reaction, was also chosen

to prevent swelling of the hydrogel during reaction. A steady increase in methylene chain surface coverage with increasing reaction time is indicated by an increasing C/O ratio in XPS and a relative increase in the C12-urethane fragment in SIMS. By 30 min, most, if not all, surface hydroxyls reacted to form the C12-chain overlayer. Molecular orientation of the ordered methylene chains, commonly described by their tilt angle, can be estimated with the dichroic ratio measured in the polarized FTIR-ATR measurements.30-32,48 Others have also used a combination of grazing angle IR/unpolarized ATR5,49 or grazing angle IR/transmission IR techniques50 to study the chain orientation of SAM and Langmuir-Blodgett films. In this polarized ATR assay, the dichroic ratio, Ap/As, from the 30-min derivatized sample was used. Because the dichroic ratios for both νCH2,asym and νCH2,sym frequencies were about the same (Table 1), the uniaxial orientation distribution model would be a good approximation for the chain orientation of the C12-urethane layer.32 In polarized ATR, three major formalisms for calculating the mean-square electric field amplitudes were used throughout the literature to estimate the tilt angle. They are the twophase approximation, the Harrick thin film approximation, and the thickness- and adsorption-dependent approximation.30 The two-phase approximation states that, for very thin films such as supported lipid monolayers or bilayers, the optical constants of the monolayers/bilayers are highly perturbed by the optical properties of the adjacent media and that the optical properties of the adsorbed film should be neglected. The electric field amplitudes of the x, y, and z directions at such interfaces are given by51 Ex )

4 cos2 θI(sin2 θ1 - n212)

2

(1 - n212)[(1 + n212) sin2 θ1 - n212) Ey2 )

Ez )

4 cos2 θI (1 - n212)

4 cos2 θI sin2 θI

2

(1 - n212)[(1 + n212) sin2 θI - n212)

(1)

(2)

(3)

where θI is the incidence angle in the ATR element and equal to 45°, n21 ) n2/n1 where n2 and n1 are real parts of refractive indices of phase 1 (KRS-5 crystal) and phase 2 (pHEMA substrate) However, when the thickness of the monolayer becomes very small, yet not insignificant compared to the depth of penetration, dp, defined by dp )

λ 2πn1(sin θI - n212)1/2 2

(4)

(where dp is wavelength-dependent) the Harrick thin film approximation should be followed. While the mean electric field amplitudes in the x and y directions are the same as those in the two-phase approximation, Ez is modified by taking the refractive index of the thin adsorbed film into the calculation. The modified equation

146


Kwok et al.

Table 1. FTIR-ATR Values of Normalized p- and s-Polarized Absorbances, Dichroic Ratios, Mean-Square Electric Field Amplitudes at 2850 and 2920 cm-1, Refractive Indices, and Tilt Angle (θ) Calculated for Dry pHEMA Surface Derivatized with C12 Isocyanate (30 min Reaction) two phasea cm-1

2850 sym measurements normalized Apc normalized Asc dichroic ratio, Ap/As calculations θI

Harrick thin filmb cm-1

cm-1

2920 asym

2920 cm-1 asym

2850 sym

0.4064 0.4203 0.967

0.6096 0.6069 1.004

0.4064 0.4203 0.967

0.6096 0.6069 1.004

45

45

45

45

nKRS-5 npHEMA substrate nC12-urethane (1.4 or 1.5)d

2.37 1.5119

2.37 1.5119

2.37 1.5119

2.37 1.5119

E x2 E y2 E z2

1.0582 3.3724 5.6867

θ (deg)

35

a

1.4

1.5

1.4

1.5

1.0582 3.3724 5.6867

1.0582 3.3724 7.7340

1.0582 3.3724 5.8701

1.0582 3.3724 7.7340

1.0582 3.3724 5.8701

36

30

34

31

35

b

c

θ was calculated using eqs 1-3 and 6. θ was calculated using eqs 1-2 and 5-6. Absorbance was normalized with respect to O-CdO ester carbon at 1720 cm-1. d Two common values of 1.4 and 1.5 were used to examine the effect of refractive index on calculated tilt angles.

for Ez in the Harrick thin film approximation is described by eq 5,30 Ez )

4n324 cos2 θI sin2 θI

2

(1 - n312)[(1 + n312) sin2 θI - n312)

(5)

where n31 ) n3/n1, n32 ) n3/n2, phase 1 is the KRS-5 crystal, phase 2 is the C12-urethane monolayer, and phase 3 is the pHEMA substrate. Values of the real refractive index of SAMs between 1.4 and 1.5 have been used in the literature30,32 and both values were used here to study the effect of the refractive index on tilt angle computation. The third method is the thickness- and adsorptiondependent approximation.30 This requires more complicated calculations by taking into account the number of layers, their complex refractive indices, and their thicknesses. Because Picard et al. concluded that the Harrick thin film approximation was acceptably accurate in estimating the tilt angle if the film thickness was less than 200 Å (true in this case; C12-urethane monolayer ∼ 20 Å , dp ∼ 5500-55000 Å), we calculated the average tilt angle using only the twophase and Harrick film approximations. Finally, θ was calculated according to eq 6.31

θ ) sin-1

[

]

Ap 2Ex2 - 2 Ey2 As Ap Ex2 - Ey2 - 2Ez2 As

1/2

(6)

Table 1 summarizes the normalized dichroic ratios, calculated mean electric amplitudes, and average tilt angles at 2850 and 2920 cm-1 using both the two-phase and the Harrick thin film approximations. The tilt angles calculated in both cases ranged from 30° to 36°, which fell into the commonly observed tilt angle range of 30°-40° for SAM molecules in an equilibrium state.52,53 The tilt angles calculated with two νCH2,asym and νCH2,sym stretching frequencies differed by only

1°, indicating either peak intensity could be used in estimating the tilt angle without significant errors. Furthermore, it appears that both theories gave similar average tilt angles except when the refractive index of C12-urethane ) 1.4 was used. The tilt angles increased from 30° to 34° for n2 ) 1.4 at 2850 cm-1 and from 31° to 35° for n2 ) 1.5 at 2920 cm-1 in the Harrick thin film approximation. This magnitude of discrepancy was also observed by Picard et al. Because the absolute values of refractive indices of many SAM films are unknown, this discrepancy is probably the uncertainty we have to accept. In fact, for tilt angles 30 min), substantial m/z ) 59peak intensity is still seen. The IR data point to considerable crystallinity and orientation within the hydrocarbon chains. However, the ATR sampling depth is 1-5 µm. The fact that, at 30 min, the ATR signals associated with the hydrocarbon chains are intense suggests the reacted zone may be microns thick. If a slab of the C12-derivatized pHEMA is cut in half, a clear zone is seen in the center of the material, surrounded by a translucent “halo.” On the basis of all data, a model such as that illustrated in Figure 12 is proposed. A surface zone, the thickness of which is governed by reaction time, consists of stoichiometrically reacted C12 chains. The chains are sufficiently close to each other to permit crystallization or organization, as has been observed in several polymers with long alkyl side chains.55 The tilt angle of ∼30° noted here is associated with the most efficient packing of the zigzag, all-trans methylene units, as reported by Ulman,1 to maximize van der Waal interactions. In his review article, Ulman pointed out that, with a molecular spacing of ∼5.0 Å, the long n-alkyl chains preferentially tilt at such angle so that the optimal distance between the chains is reached at ∼4.4 Å and the fit of “bulges into depressions” is perfect. To further support the model that our derivatized pHEMA samples have ordered structural organization in the surface zone, we can approximate the surface density of the hydroxyl groups available on the pHEMA surface. This was done by using the XPS composition data from the TFAA derivatization experiment described earlier. For complete stoichio-


metric reaction of one pHEMA monomer unit + one TFAA, a theoretical composition of 20.0%, 53.3%, and 26.6% fluorine, carbon, and oxygen, respectively, is expected. In the XPS survey scan, experimental values of 20.4% fluorine, 54.0% carbon, and 25.6% oxygen were obtained, indicating a complete surface reaction. Because three fluorine atoms are associated with each -OH, the density of the reactive hydroxyl groups on the original pHEMA was then backcalculated by dividing the experimental fluorine percentage by a factor of 3, yielding a hydroxyl group composition of 6.81%. Furthermore, on the basis of the density (1.073 g/cm3) and molecular weight (130.1 g/mol) of HEMA along with the analyzed volume (1000- × 1700-µm, 50-80-Å deep) in the XPS at a takeoff angle of 55°, it was estimated that there is 0.25 hydroxyl molecule per Å2. Assuming the hydroxyl groups are distributed evenly, then there is ≈1 hydroxyl group per every 4 Å. This spatial distance agrees well with the effective packing of alkyl chains at the ∼30° tilt angle, described earlier by Ulman. In addition, the area per molecule of long, close-packed alkyl chain thiols on Au(111) is reported to be 21.7 Å2 (square root ) 4.7 Å). The optimal s,s spacing for an all-trans conformation is 4.7 Å, also in agreement with the 5-Å spacing reported by Ulman and Camillone et al.1,56 Because our hydroxyl spacing is roughly 4 Å and our calculated average tilt angle is ∼33.5°, we believe that we have established a closed-packed, multilevel all-trans structure, within the polymeric slab surface, with similarities to SAM structures on gold surfaces. 5. Conclusions A simple, one-step procedure for generating ordered, crystalline methylene chains on polymeric surfaces via urethane linkages was developed. Using dibutyl tin dilaurate as a catalyst, a well-ordered, assembled layer on pHEMA was achieved without allophanate side reaction branching. Both XPS and TOF-SIMS analysis indicated that maximum surface coverage of C12-urethane was obtained at 60 °C for 30 min, while unpolarized ATR gave evidence of an alltrans, crystalline structure. The dichroic ratio from polarized ATR measurements yielded an average chain tilt angle of 33.5° ( 2.4° between the molecular axis and the surface normal. On the basis of XPS stoichiometry and the intensity of the IR absorbance, this ordered structure in the surface zone is proposed to be multilayered, possibly microns in thickness. Unpublished studies suggest that this film provides an excellent barrier for controlled release devices. The transport data also point to a relatively homogeneous and dense ordered surface. The surface structure described here is novel and may have an impact in areas of bioengineering that have, to date, been limited to model SAMs on rigid surfaces. By using blocking groups to protect reactive terminal amines or hydroxyls on n-alkyl isocyanates, we may be able to prepare multilayer surface structures that can be further functionalized with biomolecules or antibodies to yield biorecognition surfaces for use in medical devices and sensors. Acknowledgment. The authors gratefully acknowledge the generous funding from the UWEB (University of

148


Washington Engineered Biomaterials) NSF Engineering Research Center and NIH funded NESAC/BIO to support this research work. The authors also would like to thank Professor David Allara, Dr. Sheng Pan, Dr. Vickie Pan, Dr. Felix Simonovsky, and Dr. Esmaeel Naeemi for their helpful discussions, and Matt Wagner and Dan Graham for their assistance in converting SIMS files. References and Notes (1) Ulman, A. Chem. ReV. 1996, 96, 1533-54. (2) Bain, C. D.; Troughton, E. B.; Tao, Y.-t.; Evall, J.; Whitesides, G. M.; Nuzzo, R. G. J. Am. Chem. Soc. 1989, 111, 321-35. (3) Sagiv, J. J. Am. Chem. Soc. 1980, 102, 92-98. (4) Maoz, R.; Netzer, L.; Gun, J.; Sagiv, J. J. Chim. Phys. 1988, 85, 1059-65. (5) Song, Y. P.; Petty, M. C.; Yarwood, J.; Feast, W. J.; Tsibouklis, J.; Mukherjee, S. Langmuir 1992, 8, 257-61. (6) Charych, D. H.; Bednarski, M. D. MRS Bull. 1992, 17, 61-66. (7) Sagiv, J. Isr. J. Chem. 1979, 18, 339-46. (8) Maoz, R.; Sagiv, J. Langmuir 1987, 3, 1034-44. (9) Maoz, R.; Sagiv, J.; Degenhardt, D.; Helmuth, M.; Quint, P. Supramol. Sci. 1995, 2. (10) Bohme, P.; Vedantham, G.; Przybycien, T.; Belfort, G. Langmuir 1999, 15, 5323-28. (11) Chaudhury, M. Biosens. Bioelectron. 1995, 10, 785-88. (12) Ferguson, G. S.; Chaudhury, M. K.; Biebuyck, H. A.; Whitesides, G. M. Macromolecules 1993, 26, 5870-75. (13) Whitesides, G. M.; Ferguson, G. S. Chemtracts-Org. Chem. 1988, 1, 87. (14) Montheard, J.-P.; Chatzopoulos, M.; Chappard, D. J. Macromol. Sci. ReV. Macromol. Chem. Phys. 1992, C32, 1-34. (15) Ratner, B. D.; Hoffman, A. S. Synthetic Hydrogels for Biomedical Applications; Andrade, J. D., Ed.; American Chemical Society: Washington D.C., 1976; pp 1-36. (16) Montheard, J.-P.; Chatzopoulos, M.; Chappard, D. Homopolymers and Copolymers of 2-Hydroxyethyl Methacrylate for Biomedical Applications; Reza, A., Ed.; American Chemical Society: Washington D.C., 1997; pp 699-711. (17) Kahovec, J.; Coupek, J. React. Polym. 1988, 8, 105-11. (18) Stovbun, E. V.; Kuzaev, A. I.; Baturin, S. M. Polym. Sci., Ser. A 1996, 38, 691-95. (19) Arnold, R. G.; Nelson, J. A.; Verbanc, J. J. Chem. ReV. 1957, 57, 47-76. (20) Saunders, J. H.; Slocombe, R. J. Chem. ReV. 1948, 43, 203-18. (21) Ratner, B. D.; Miller, I. F. J. Biomed. Mater. Res. 1973, 7, 353-57. (22) Kwok, C. S.; Ratner, B. D. Modification of Polymer Surfaces with Self-assembled Monolayers; Society of Biomaterials: Providence, RI, 1999; p 73. (23) Brynda, E.; Stol, M.; Chytry, V.; Cifkova, I. J. Biomed. Mater. Res. 1985, 19, 1169-79. (24) Hart, D. E.; DePaolis, M.; Ratner, B. D.; Mateo, N. B. CLAO J. 1993, 19, 169-73. (25) Castner, D. G.; Ratner, B. D.; Hirao, A.; Nakahama, S. Surf. Sci. Spectra 1997, 4, 14-20. (26) Schamberger, P. C.; Gardella, J. A., Jr.; Grobe, G. L., III; Valint, P. L., Jr. Polym. Prepr. (Am. Chem. Soc., DiV. Polym. Chem.) 1993, 34, 58-59.

Kwok et al. (27) Chilkoti, A.; Lopez, G. P.; Ratner, B. D.; Hearn, M. J.; Briggs, D. Macromolecules 1993, 26, 4825-32. (28) Chilkoti, A.; Ratner, B. D.; Briggs, D. Chem. Mater. 1991, 3, 51. (29) Collett, J. H.; Spillane, D. E.; Pywell, E. J. Polym. Prepr. (Am. Chem. Soc., DiV. Polym. Chem.) 1987, 28, 141-42. (30) Picard, F.; Buffeteau, T.; Desbat, B.; Auger, M.; Pezolet, M. Biophys. J. 1999, 76, 539-51. (31) Jang, W.-H.; Miller, J. D. J. Phys. Chem. 1995, 99, 10272-79. (32) Park, S. Y.; Franses, E. Langmuir 1995, 11, 2187-94. (33) Ahn, D. J.; Franses, E. Thin Solid Films 1994, 244, 971-76. (34) Ahn, D. J.; Franses, E. J. Phys. Chem. 1992, 96, 9952-59. (35) Lub, J.; van der Wel, H.; van Vroonhovem, F. C. B. M.; Benninghoven, A. Recl. TraV. Chim. Pays-Bas. 1990, 109, 367-74. (36) Lub, J.; van Velzen, P. N. T. Quantification of NegatiVe Time-ofFlight Secondary Ion Mass Spectra of Poly(alkylmethacrylate) Surfaces; Benninghoven, A., Ed.; John Wiley & Sons: Munster, Federal Republic of Germany, 1989; Vol. IFOS 4, pp 23-36. (37) Brown, A.; Vickerman, J. C. Surf. Interface Anal. 1986, 8, 75-81. (38) Lopez, G. P.; Chilkoti, A.; Briggs, D.; Ratner, B. D. J. Polym. Sci., Part A: Polym. Chem. 1992, 30, 2427-41. (39) Hearn, M. J.; Briggs, D. Surf. Interface Anal. 1988, 11, 198-213. (40) Briggs, D.; Hearn, M. J. SIMS of Acrylic Polymers: A Detailed Study of Ion Formation from Thick Film; Benninghoven, A., Ed.; John Wiley & Sons: Munster, Federal Republic of Germany, 1989; Vol. IFOS 4, pp 37-42. (41) Antonucci, J. M.; Brauer, G. M.; Termini, D. J. J. Dent. Res. 1980, 59, 35-43. (42) Socrates, G. Infrared Characteristic Group Frequencies, 2nd ed.; John Wiley and Sons: West Sussex, England, 1994. (43) Allara, D. L.; Nuzzo, R. G. Langmuir 1985, 1, 52-66. (44) Snyder, R. G.; Strauss, H. L.; Elllger, C. A. J. Phys. Chem. 1982, 86, 5145-50. (45) Snyder, R. G.; Maroncelli, M.; Strauss, H. L.; Hallmark, V. M. J. Phys. Chem. 1986, 90, 5623-30. (46) Cheng, S. S.; Scherson, D. A.; Sukenik, C. N. Langmuir 1995, 11, 1190-95. (47) Silver, J. H.; Hergenrother, R. W.; Lin, J.-c.; Lim, F.; Lin, H.-b.; Okada, T.; Chaudhury, M. K.; Cooper, S. L. J. Biomed. Mater. Res. 1995, 29, 535-48. (48) Menikh, A.; Saleh, M. T.; Gariepy, J.; Boggs, J. M. Biochemistry 1997, 36, 15865-72. (49) Song, Y. P.; Yarwood, J.; Tsibouklis, J.; Feast, J.; Cresswell, J.; Petty, M. C. Langmuir 1992, 8, 262-66. (50) Gu, Y.; Shi, Z.; Nie, C.-s. Appl. Spectrosc. 1998, 52, 855-62. (51) Harrick, N. J. Principles of Internal Reflection Spectroscopy, 1st ed.; Harrick Scientific Corp.: Ossining, NY, 1967; p 27. (52) Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. J. Am. Chem. Soc. 1987, 109, 3559. (53) Nuzzo, R. G.; Dubois, L. H.; Allara, D. L. J. Am. Chem. Soc. 1990, 112, 558. (54) Lopez, G. P.; Castner, D. G.; Ratner, B. D. Surf. Interface Anal. 1991, 17, 267-72. (55) Plate, N. A.; Shibaev, V. P.; Petrukhin, B. S.; Zubov, Yu. A.; Kargin, V. A. J. Polym. Sci., Part A-1 1991, 9, 2291-2298. (56) Camillone, N., III; Chidsey, C. E. D.; Ling, Y.; Pietvinski, T. M.; Scoles, G. J. Chem. Phys. 1991, 94, 8493.

BM000292W

Surface Modification of Polymers with Self-Assembled Molecular

Recommend Documents