Surface Analysis by X-ray Photoelectron Spectroscopy of Sol−Gel

11850

J. Phys. Chem. B 2007, 111, 11850-11857

Surface Analysis by X-ray Photoelectron Spectroscopy of Sol-Gel Silica Modified with Covalently Bound Peptides Sabrina S. Jedlicka,†,‡ Jenna L. Rickus,*,†,‡,§ and Dmitry Y. Zemlyanov*,|,⊥ Department of Agricultural & Biological Engineering, Purdue UniVersity, West Lafayette, Indiana 47907, Physiological Sensing Facility, Bindley Bioscience Center, Purdue UniVersity, West Lafayette, Indiana 47907, Weldon School of Biomedical Engineering, Purdue UniVersity, West Lafayette, Indiana 47907, Birck Nanotechnology Center, Purdue UniVersity, West Lafayette, Indiana 47907, and Materials and Surface Science Institute, UniVersity of Limerick, Limerick, Ireland ReceiVed: June 7, 2007; In Final Form: July 28, 2007

Chemical surface characterization of biologically modified sol-gel derived silica is critical but somewhat limited. This work demonstrates the ability of x-ray photoelectron spectroscopy (XPS) to characterize the surface chemistry of peptide modified sol-gel thin films based on the example of four different free peptidesilanes, denoted RGD, NID, KDI ,and YIG. The N 1s and C 1s peaks were found to be good fingerprints of the peptides, whereas O 1s overlapped with the signal of substrate oxygen and, therefore, the O 1s peak was not informative in the case of the thin films. The C 1s peak was fitted and the contribution of the residual hydrocarbons was sorted out. The curve-fitting procedure of the C 1s peak accounted for the different chemical states of carbon atoms in the peptide structure. The curve-fitting procedure was validated by analyzing free peptides in the powder form and was then applied to the characterization of the peptide-modified thin films. The XPS measured ratio between nitrogen and carbon for the peptide thin film was similar to the corresponding value calculated from the peptide structures. Angle resolved XPS confirmed the surface nature of peptides in modified thin films. The coverage and thickness of the peptides on the thin film surface depended on the peptide sequence. The coverage was in the range of 10% of a monolayer, and the layer thickness varied from 10 to 30 Å. We believe that the different thicknesses and surface coverage are due to the local structure of the peptides, with the RGD and NID peptides taking a globule conformation and the YIG and KDI peptides adopting a more linear structure.

Introduction The integration of cells into engineered devices has considerable potential in implantable biomedical therapeutics and cellbased biosensors. These applications require that cells survive on inorganic or hybrid materials and carry out normal metabolism, proliferation, and differentiation. The microenvironment immediately surrounding the cells substantially impacts cellular processes and ultimately the final fate of the cell. The presentation and characterization of bioactive molecules at the surface of inorganic and hybrid materials, therefore, is a critical materials capability. Sol-gel derived materials produced under biologically benign conditions have demonstrated an ability to serve as substrates for adherent mammalian cells.1,2 The sol-gel method of producing amorphous inorganic or organically modified porous solids from liquid precursors is a functionally diverse technique that allows for simple manipulation of both chemical features as well as nanoscale morphology through processing procedures. Metal alkoxides, such as tetramethoxysilane, are commonly used * To whom correspondence should be addressed. Dmitry Y. Zemlyanov, Purdue University, Birck Nanotechnology Center, 1205 West State Street, West Lafayette, IN 47907-2057. Phone: +1 (765) 496-2457. Fax: +1 (765) 496-8299. E-mail: [email protected] (D.Y.Z). E-mail: rickus@ purdue.edu (J.L.R.). † Department of Biological and Agricultural Engineering. ‡ Bindley Bioscience Center. § Weldon School of Biomedical Engineering. | Birck Nanotechnology Center. ⊥ University of Limerick.

as liquid precursors due to the ease of hydrolysis.3 Organically modified silica (ormosil) can be created by modifying the functional groups on the silane precursors,4 resulting in a material that is decorated with the chemical functionality of choice. Recently, we developed a novel sol-gel method for producing biologically active peptide-modified porous silica matrixes from peptide-silane precursors.5 A variety of materials have taken advantage of known extracellular matrix peptide sequences at the biointerface to enhance cell adhesion. Peptides such as the RGD group from fibronectin type III repeats6-11 and the YIGSR peptide from laminin12-14 enhance cell adhesion of various cell types to twodimensional materials. Previous work has generally presented such bioactive peptides as a self-assembled monolayer immobilized using various chemistries, including thiol attachment,15 covalent assembly using amine linkers,16 or physioadsorption onto the surface of interest.15,17 Controlling the percentage of the peptides at the surface can be tedious, as multilayer organization and environmental factors all play a role in the efficiency of the coupling chemistry, leading to a lack of consistency from surface to surface. In this work, the use of sol-gel chemistry allows for a simple one-pot synthesis, allowing for precise concentration manipulation of added peptides. In addition, the peptide of choice is covalently linked to the resulting film such that the peptides do not leach out of the material matrix. Chemical surface characterization of biologically modified sol-gel derived silica is critical, but somewhat limited. Attenu-

10.1021/jp0744230 CCC: $37.00 © 2007 American Chemical Society Published on Web 09/20/2007

XPS of Sol-Gel Silica Modified with Peptides ated total reflection (ATR) FT-IR is challenging due to the large peak derived from the Si-O bonds, typically covering up the fingerprint region of many organic molecules. Atomic force microscopy provides information on the material topography and can give limited chemical interaction information via phase interactions.18,19 Electron microscopy gives structural information and limited chemical information as well.20 The presentation of peptides at the surface of materials requires a precise technique to analyze the chemical nature of the surface of the material for available surface chemistry for critical cellular interactions, such as integrin receptor binding. Here we demonstrate the application of X-ray photoelectron spectroscopy (XPS) analysis to the surface of peptide modified silica solgel derived materials. XPS is a surface sensitive technique that has become widely used for studying properties of atoms, molecules, solids, and surfaces.21-34 Generally speaking, intensities of core level photoelectron peaks are used for quantitative analysis, and binding energies of core level photoelectrons exhibit chemically induced shifts. The main success of the XPS technique is associated with studies of the physical and chemical phenomena on the surface of solids (see for instance refs 35 and 36 and refs therein). These investigations were limited to relatively simple inorganic reactions, and few biologically relevant problems have been approached using XPS. Impartial reasons exist for the low involvement of XPS into investigations of biologically related objects. First, organic chemistry samples often exhibit high vapor pressure and, therefore, degas poorly in vacuum creating incompatibilities with the XPS technique. Second, X-rays might cause radioactive damage of a sample. Third, the C 1s region, which is most informative for organic chemistry samples, is narrow and the photoemission peaks can overcrowd the region. At this point, NMR and mass spectrometry provide better data on the chemical composition and molecule structure. The real advantage of XPS, however, is in the study of surface chemistry, and therefore XPS is a potential technique for the characterization of biologically modified sol-gel surfaces. In this work, we report on the systematic XPS investigation of four free peptide-silanes and peptide-silane derived organically modified silica thin films. We compared spectral characteristics of the free peptide-silanes and peptide presenting solgel surfaces. Coverage and thickness of the peptide-silane layer were calculated. This work demonstrates a use for XPS to characterize biologically inspired surfaces, providing critical information on peptide coverage on the surface of the materials. Experimental Section Reagents and Materials. All Fmoc-protected amino acids, synthesis resin, and O-benzotriazole-N,N,N′,N′-tetramethyluronium-hexafluoro-phosphate (HBTU) were purchased from Anaspec, Inc. Tetramethylorthosilicate (TMOS) and 3-aminopropyltrimethoxysilane (APTMS) were obtained from Fluka (Sigma Aldrich). All solvents and remaining peptide synthesis chemicals were obtained from Sigma. Peptide-Silane Synthesis. Free peptides and conjugated peptide silanes were prepared as previously described.5 Peptides were purified using ether precipitation, followed by at least five ether washes prior to peptide drying, to remove a majority of the protective groups from the peptide-silanes prior to solgel synthesis. Sol-Gel Synthesis. Tetramethylorthosilicate (3.8 mL) (TMOS) was hydrolyzed under acid catalysis with 850 µL of purified H2O (18 MΩ) and 55 µL of 0.04 N HCl. The sol was then

J. Phys. Chem. B, Vol. 111, No. 40, 2007 11851 filtered through a 0.2 µm Whatman filter and aged overnight at 4 °C to ensure near complete hydrolysis. Peptide-silanes were solubilized in 0.02 M phosphate buffer (pH 6.0) by sonication at a molar ratio of 0.1% to silicon dioxide. TMOS sol was combined with the buffer containing peptide-silanes at a 30: 70 ratio, with an additional 10% of the final volume of methanol added to slow condensation. Thin films were dipped onto a piranha-cleaned (H2SO4/H2O2, 4:1) WPI coverglass at a rate of 35 mm/s. The peptide-silica films were briefly allowed to gel and then transferred to a closed container in the dry state to reduce surface contamination. X-ray Photoelectron Spectroscopy. XPS data were obtained by a Kratos Ultra DLD spectrometer using monochromatic Al KR radiation (hν ) 1486.58 eV). The survey and high-resolution spectra were collected at a fixed analyzer pass energy of 160 and 20 eV, respectively. The spectra were collected at a 0° and 60° angle with respect to the surface normal. The atomic concentrations of the chemical elements in the near-surface region were estimated after the subtraction of a Shirley-type background, taking into account the corresponding Scofield atomic sensitivity factors and inelastic mean free pass (IMFP) of photoelectrons as a standard procedure of the CasaXPS software. The peak areas without the IMFP correction were used in eqs 1 and 3 because IMFP attenuation is included to the formulas. The binding energy (BE) values, referred to the Fermi level, were corrected using the C 1s 284.80 eV; the standard deviation of the peak position associated with the calibration procedure was (0.05 eV. A commercial Kratos charge neutralizer was used to achieve a resolution of 1.0-1.2 eV measured as a full width at half maximum (fwhm) of the C 1s deconvoluted peaks. The XPS spectra were fitted by CasaXPS software assuming the line shape to be a Gaussian-Lorentzian function. No X-ray damage of the samples was detected: the spectra did not degrade with time under X-ray beam, and the sample color was also unchanged. Results Free Peptide-Silane Analysis and XPS Peak CurveFitting Procedure. Before characterization of the peptide thin film surface, free peptides were investigated to determine reference data and to validate the XPS approach and analysis. Four different free peptide-silanes, RGD, NID, KDI, and YIG, were characterized by XPS. The structures of the peptides are shown in Figure 1. Peptides were prepared as described in the Experimental Section. Fluoride, oxygen, carbon, silicon, and sulfur were detected by XPS on the surfaces of all samples. As an example, the survey and high-resolution spectra obtained from the free RGD peptide are represented in Figure 2. Fluoride is an impurity derived from trifluoroacetic acid cleavage, and its trace amounts remain in the peptide-silane postsynthesis and after purification. Complete removal of this impurity would require lypholization from water or acetic acid after ether purification. The peptide molecules consist of oxygen, nitrogen, and carbon. The O 1s spectra typically show a broad featureless peak of ∼1.6 eV fwhm at ∼531.5 eV with a high BE shoulder. The N 1s peak is usually at ∼400 eV and is also featureless. The C 1s peak is the most promising for XPS characterization. The shape of the C 1s should vary depending on the free peptide-silane structure. As can be seen in Figure 1, carbon atoms in the peptide structure can be in one of six chemical environments. The RGD structure with marked carbon positions is shown in the Supporting Information (Figure 1-SI). The first

11852 J. Phys. Chem. B, Vol. 111, No. 40, 2007

Jedlicka et al.

Figure 1. Peptide-silanes synthesized and characterized.

type of carbon is bonded only to carbon and/or hydrogen. The second type of carbon is coordinated with one nitrogen atom along with carbon and/or hydrogen atoms. Type 3 is a carbon atom with a single bond to oxygen atom. Type 4 is a carbon from an amide group. Type 5 is a carbon atom from a carboxyl group. Type 6 is carbon coordinating with three nitrogen atoms, as in arginine. The carbon atoms in these six different positions should have different chemical shifts, and the intensity of the

individual components should be proportional to the population of the particular carbon type; therefore, different peptide-silanes are predicted to show the C 1s spectrum with different shape. For convenience, the number of carbon atoms in the different position and their expected binding energies for different peptide-silanes are summarized in Table 1. The characteristic BE values of the C 1s peaks observed in this study are summarized in Table 2.

XPS of Sol-Gel Silica Modified with Peptides

Figure 2. The survey and narrow-region spectra obtained from the free RGD peptide at 0° collection angle.

The validation of this curve-fitting technique was verified with the free peptide-silane samples. Figure 3 shows the C 1s peak obtained from the free RGD peptide-silane at 0° collection angle and the curve-fitting analysis according to the technique described above. The relative ratio between six C 1s components was fixed according to the number of the corresponding atoms in the certain peptide structure. It should be noted that no angledependence was expected, and the spectra obtained from the free peptide-silane samples did not change with the collection angle. Extra components were added to the curve-fitting analysis to accommodate the peaks originating from residual hydrocarbons and the CFx species. Also, in order to simplify the curvefitting, the C-O and C-N components were simulated by one peak. This curve-fitting procedure of the C 1s spectra demonstrated consistent results for all four peptides: RGD, NID, KDI, and YIG. The C 1s spectra obtained from NID, KDI, and YIG peptides are shown in the Supporting Information in Figure 2-SI. The typical level of the residual hydrocarbons was around 1015% of the total carbon signal, and the contribution of the CFx species was less than 2%. The fit was further validated by comparison of the expected ratio of oxygen/carbon and nitrogen/ carbon with the XPS measured values. The results are represented in Table 3, and one can see that the XPS measurements and the expected values demonstrate remarkably good agreement. Therefore, we conclude that XPS can be used for reliable characterization of peptide-silanes. XPS Characterization of the Peptide Thin Film Surface. The next step was XPS identification of the RGD, NID, KDI, and YIG thin-film bound peptide-silanes, which were produced using 1 mol % final peptide-silane concentration to silicon. XPS detected the presence of sodium, fluoride, oxygen, carbon, phosphorus, silicon, sulfur, and potassium on the surfaces of all samples. Sodium, potassium, phosphorus, and sulfur are all byproducts of TMOS thin film preparation. Specifically, the sodium, potassium, and phosphorus are all components of the phosphate buffer used during the condensation reaction of the sol-gel silica films. The sulfur is a contaminant from dimethyl sulfoxide, used to enhance the solubility of the peptide-silanes. Fluoride is an impurity derived from trifluoroacetic acid cleavage, and trace amounts remain in the peptide-silanes postsynthesis and after purification. As an example, the C 1s spectrum of the KDI thin film is shown in Figure 4. The C 1s spectra and the curve-fitting results for RGD, NID, and YIG films are shown in Figure 5. The K 2p3/2 and K 2p1/2 contributions were added to the curve-fitting profile compared with the fitting profile for free peptides. The level of residual hydrocarbon was 25-30% of the total carbon signal, which is

J. Phys. Chem. B, Vol. 111, No. 40, 2007 11853 higher than the signal for the free peptide-silanes. Regardless, we were able to reliably identify the surface presented peptides by XPS. The ratio between nitrogen and carbon calculated from XPS data was close to the ideal ratio for the certain peptidesilane structures as shown in Table 4. Inconsistently high O/C ratios can be explained by oxygen contributions from silica. As shown in Figure 6, the curve-fitting resolves two components. The high BE component at 532.5 eV is assigned to silica, whereas the low BE peak at ∼531 eV is due to peptides. Unfortunately, the strong broad feature of the silica oxygen results in poor fitting of the peptide component and this causes inconsistency in the O/C ratio. To verify the surface nature of the peptide, the XPS data were acquired at the collection angle of 60° with respect to the surface normal. The data collected at 60° are more surface sensitive, whereas at a 0° angle, the bulk contribution dominates. As shown by this experiment, the concentration of carbon increases with the angle indicating the presence of peptide at the surface. A short sputtering by Ar+ also resulted in the disappearance of peptide features. This also indicates that peptides are surface species. Discussion Many factors could potentially affect the surface presentation of the peptides. First, upon thin film formation, the pores of the silica collapse resulting in remarkably different nanoscale morphology compared to traditional silica monoliths. This collapse will likely affect the presentation of the surface peptides. In addition, surface integrated peptides could interact with one another, increasing the local concentration of peptides in the final material. Finally, sample degradation and chemical contamination on the surface of the materials could potentially occur; altering the final presentation of surface-presented peptides. Peptide-silane concentration and structure can alter the final fate of cells existing at the material interface. Understanding how the peptide-silanes are present on the surface of the thin film peptide-silane materials is critical in designing biointerfaces to modulate cellular processes. Determining the impact of each of the above factors on surface presentation, therefore, is essential in the final design of peptide functionalized sol-gels for cell-based devices. The ability of XPS to analyze chemical bonding profiles through binding energy changes can detect peptide degradation and surface presentation of intact peptide-silanes, as well as the presence of chemical contamination. Below, we demonstrate XPS application for quantification of surface peptide-silanes and for understanding the nature of their surface presentation. Carbon and nitrogen can serve as fingerprints of the surface peptides, whereas oxygen is not a good choice due to a strong signal from the silica sol-gel matrix as shown in Figure 6. The thickness of the peptide layer can be estimated using the intensity of the C 1s and N 1s peaks through the equation

IA,k 0 IA,k

(

d ) 1 - exp λA,k cos θ

)

(1)

where IA,k is the intensity (area) of the k electron level peak (1s) of atom A (in our case, carbon or nitrogen) from a certain 0 layer thickness; IA,k is the intensity from an “infinitely” thick layer; d is the thickness of the layers in Å; λA,k is the inelastic mean free pass (IMFP) of the photoelectrons emitted from the k level inside of the covering layer in Å; and θ is the takeoff angle, which is between the surface normal and the central axis of the electron optics (direction of photoemission collection).

11854 J. Phys. Chem. B, Vol. 111, No. 40, 2007

Jedlicka et al.

TABLE 1: The Number of Carbon Atoms in the Different Position and Expected BE Range (Free Peptides/Peptide-Silanes)

binding energy expected, eV NID silane KDI silane RGD silane YIG free peptide

C-C type 1

C-N type 2

C-O type 3

OdC-N (amide) type 4

OdC-OH (carboxyl) type 5

C-N3 type 6

284.6-285a

285.5-286.5b

286.4-287c

287.4-288.4d

288.5-289.6e

289.0-289.5f

18 20 20 18

12 13 13 14

7 7 2

13 12 12 9

3 2 2 1

1 1 1

a References 28, 32, 33, and 43-51. b References 28, 52, and 53. c References 24, 33, 44, 46, 47, 49, 51, and 53-55. d References 28, 53, 54, and 56. e References 24, 44, 46, 47, 51, 52, and 54. f Reference 55.

TABLE 2: The Characteristic C 1s Peak Positions Observed in This Study peptide peptide silane

C-C, eV

C-N, eV

C-O, eV

OdC-N, eV

OdC-OH, eV

C-N3, eV

284.75 ( 0.04 284.74 ( 0.07

286.16 ( 0.09 286.33 ( 0.32

286.16 ( 0.09 286.33 ( 0.32

288.05 ( 0.13 288.19 ( 0.36

288.75 ( 0.22 289.17 ( 0.40

289.54 ( 0.24 289.63 ( 0.59

TABLE 3: Oxygen to Carbon and Nitrogen to Carbon Ratios Measured by XPS and Expected from the Peptide Structure

NID silane KDI silane RGD silane YAVTGRGDSPAS free peptide YIG free peptide

total carbon amount, at %

Cpeptide amount, at %

N amount, at %

O amount, at %

ratio between N and Cpeptide ideal/XPS

ratio between O and Cpeptide ideal/XPS

52.4 ( 0.1 58.3 53.5.0 ( 0.2 48.2.0 ( 1.3 56.6 ( 4.8

41.9 ( 0.3 51.8 50.6 ( 0.1 43.0 ( 1.5 37.4 ( 0.8

14.1 ( 0.0 13.7 14.6 ( 0.0 13.0 ( 0.3 21.7 ( 1.0

22.7 ( 0.1 20.7 25.3 ( 0.2 25.1 ( 0.7 11.4 ( 1.1

0.310/0.337 ( 0.0029 0.278/0.265 0.313/0.289 ( 0.00019 0.291/0.304 ( 0.0032 0.31/0.306 ( 0.036

0.429/0.403 ( 0.0047 0.333/0.335 0.375/0.440 ( 0.0087 0.400/0.473 ( 0.032 0.333/0.456 ( 0.1856

0 In our case, θ was 0° and 60°. The value of IA,k (I0C1s and I0N1s) can be obtained from the free peptide measurements assuming the same roughness for the powder samples and the surface peptide thin films. In order to calculate IMFP, λA,k (in nm), in polymer compounds, Cumpson proposed37 the following equation:

Nrings is the number of aromatic six-member rings in the molecule or polymer repeat unit considered, 0χ(V) is the zerothorder valence connectivity index of Kier and Hall (evaluated by Bicerano’s method in the case of a polymer), Nnon-H is the number of atoms in the molecule or polymer repeat unit, excluding hydrogen atoms, EA,k kin is the kinetic energy (in keV) of photoelectrons emitted from the k level. The procedure for calculating the 0χ(V) index is given in ref 37. For full and detailed explanations and justification for the steps involved, see the books by Bicerano38 or Kier and Hall.39

Jablonski and Powell40 suggested that elastic scattering in a solid can also lead to the disappearance of a photoelectron from the cone of analysis, and therefore they proposed to use the value of an average electron attenuation length, Lave, instead of IMFP, λA,k. An average electron attenuation length can be calculated using NIST SRD 82.41 This estimation method requires values for such quantities as the band gap and the density, in the case of peptide-silanes for which accurate measurements are not available. The parameters and the data used for λA,k and Lave calculation are provided in the Supporting Information. λA,k is in the range of 33-34 Å for C 1s and 3031 Å for N 1s, whereas Lave varies less and is equal to ∼33.6 and ∼31 Å for C 1s and N 1s, respectively. As one can see, λA,k and Lave values are very close. The discrimination between λA,k and Lave is not the goal of this paper and, moreover, which method is used is not practically important. Equation 1 demonstrates that in practice the ratio d/λA,k is calculated, and the change of λA,k for Lave should not qualitatively change the results. Most likely, the accuracy of the thickness measurements is determined by XPS technique limitations and sample differ-

Figure 3. The C 1s spectrum obtained from the free RGD peptide at 0° collection angle.

Figure 4. The C 1s spectrum obtained from the KDI surface peptidesilane at 0° collection angle.

λA,k )

3.117(0χ(V)) + 0.4207Nrings 0.79 + 1.104(EA,k (2) kin ) Nnon-H

XPS of Sol-Gel Silica Modified with Peptides

J. Phys. Chem. B, Vol. 111, No. 40, 2007 11855 In ref 42, Fadley proposed that for a nonattenuating adlayer approximation, the intensity overlayer/substrate ratio can be described as

dσl Ω0(El)A0(El)D0(El) d soverl dΩ ) dσk subst ssubst Nk(θ) Ω0(Ek)A0(Ek)D0(Ek) Λe (Ek)cos θ dΩ Nl(θ)

Figure 5. The C 1s spectra obtained from YIG, RGD, and NID silanes at the takeoff angle of 0° with respect to the surface normal.

( )

where Nl(θ) and Nk(θ) are the peak intensity of the overlayer and substrate, respectively; Ω0 is the acceptance solid angle of the electron analyzer; A0 the effective area of specimen over which Ω0 * 0; D0 is the instrument detection efficiency; θ is the angle between the surface normal and the electron emission direction; dσk/dΩ is the differential cross-section, which can be calculated using the tabulated Scofield cross-sections and the Reilman asymmetric parameter; Λsubst (Ek) is the IMFP of e the substrate photoelectron in the substarate; soverl is the mean surface density of atoms in which peak l originates in cm-2; ssubst is the mean surface density of substrate atoms in cm-2; soverl/ssubst is the fractional monolayer coverage of the atomic species in which peak l originates; d is the mean separation between layers of density s in the substrate. We modified eq 3 and estimated the peptide coverages using the nonattenuating adlayer approximation,42 which gives coverage in the modified form:

dσSi2p NC1s ΛSiO2 (KESi2p) cos θ dΩ e Coverage ) dσC1s NSi2pd dΩ

Figure 6. The O 1s spectrum obtained from the KDI surface peptidesilane at 0° collection angle.

ences (peptide thin film versus free peptide in the form of powder). To calculate the thickness of the peptide layer, we used IMFP and λA,k. The results are summarized in Table 5. Total C 1s intensity was used for the calculation without subtraction of residual carbon contributions. Peptide coverage is also valuable information, which can be extracted from the XPS data. The intensities of C 1s/N 1s and Si 2p can be used to calculate peptide coverage. The thickness of the peptide layers measured using the C 1s peak varies from 10 to 30 Å depending on the peptide (Table 5). According to NIST SRD 82,41 a 10 and 30 Å layer attenuate the substrate signal to 75% and 40%, respectively. These attenuations are within the boundary region between nonattenuating adlayer and attenuating adlayer approximations.42 We calculated coverages using the nonattenuating adlayer approximation because (i) adsorbed layers should not be dense, and therefore the apparent attenuation should be less than one predicted by NIST SRD 82,41 (ii) the solution for the nonattenuating adlayer approximation is analytical. We thoroughly verified this approach with a number of adsorption systems, for instance such as a selfassembled monolayers on Au, GaAs, InP, GaP, etc. surfaces.

(3)

(4)

where NC1s and NSi2p are the intensity of the C 1s and Si 2p 2 peaks, respectively; ΛSiO e (KESi2p) is the IMFP of the Si 2p photoelectron in SiO2; d is the mean distance between the layers of Si atoms in silica. As discussed in ref 40, IMFP can be replaced by the electron attenuation length for quantitative analysis, QEAL, and is calculated by NIST SRD 8241 to be equal to 33.39 Å. The mean distance d depends on the structure of SiO2, and because the exact structure of substrate is unknown, d was assumed to be equal to 2.5 Å, which is the “average” value for the different SiO2 structures. The peptide coverages calculated using eq 4 are summarized in Table 5. The values were corrected for the real peptide concentration; the contribution of residual hydrocarbons was subtracted. The coverage is measured in monolayer, ML, which is the ratio between the number of peptide molecules and the number of surface silicon atoms. It should be noted that the surface coverages calculated based on the C 1s and N 1s signals are consistent. The N 1s signal originates exclusively from the peptide, whereas the C 1s features of the peptide were obtained from the curve-fitting procedure. The consistency between two values unambiguously validates the curve-fitting protocol and also validates our coverage calculations presented by eq 4. As seen in Table 5, however, the thickness calculated based on the N 1s peak was much lower than those estimated using the C 1s peak. For instance, the KDI thickness is 29.9 and 16.8 Å as calculated from the C 1s and N 1s peaks, respectively. For the RGD thin film, these values are 15.0 and 3.7 Å (Table 5). Equation 1 indirectly assumes that the roughness of free peptides and surface peptides is the same. Different surface roughness, however, between the free and surface bound

11856 J. Phys. Chem. B, Vol. 111, No. 40, 2007

Jedlicka et al.

TABLE 4: Oxygen to Carbon and Nitrogen to Carbon Ratios Measured by XPS and Expected from the Surface Peptide-Silane Structure

NID silane KDI silane RGD silane YIG silane

total carbon amount, at %

Cpeptide amount, at %

N amount, at %

O amount, at %

ratio between N and Cpeptide ideal/XPS

ratio between O and Cpeptide ideal/XPS

10.2 17.7 19.0 ( 1.6 13.3.0 ( 0.4

5.3 11.1 10.8 ( 1.4 6.2 ( 0.4

1.3 3.0 3.5 ( 0.4 1.5 ( 0.1

61.2 53.1 51.3 ( 2.6 58.7 ( 0.3

0.29/0.30 0.26/0.27 0.31/0.32 ( 0.01 0.29/0.24 ( 0.03

0.48/1.15 0.38/0.54 0.42/0.23 ( 0.05 0.33/0.18 ( 0.03

TABLE 5: Peptide Coverage and Peptide Layer Thicknessa height, Å NID silane KDI silane RGD silane YIG silane

coverage, molecules per Si atoms on the surface

from C 1s

from N 1s

from C 1s

from N 1s

10.3 ( 1.9 29.9 15.0 ( 0.9 23.2 ( 0.5

4.0 ( 0.7 16.8 3.7 ( 0.3 12.6 ( 1.0

0.08 0.14 0.14 ( 0.02 0.07 ( 0.01

0.05 0.13 0.14 ( 0.02 0.06 ( 0.01

a The error represents the standard deviation obtained by averaging through a few samples.

peptides cannot explain the mismatch in the N 1s and C 1s calculations; as this would lead to proportional numerical differences. A variety of morphological and chemical interactions, therefore, may be associated with the calculated difference between the N 1s and C 1s peptide thickness. In addition to the morphology of the thin film structure, several other factors may contribute to the increased peptide concentration and variable molecular height at the surface of the peptide-silane thin films. A primary factor to consider is the likelihood of peptide folding. Each of the peptide-silanes exhibits a variety of chemical characteristics that contribute to hydrophobicity, molecular “bends”, and interfacial structure. The nature of peptides to adopt a structure based on hydrophobic and electrostatic interactions may also contribute to the conformation of the peptide molecules on the surface. One can conclude that due to different peptide conformations, N and C are located at different depths in the free peptide and in the surface peptide-silane. Because of peptide folding, nitrogen atoms are hidden inside of the peptide structure and therefore are closer to the silica surface, while carbon atoms are more exposed. This also implies that peptide configuration is different in the powder form and in the surface state. This conformation potentially explains the larger attenuation of the N 1s electrons compared to the C 1s electrons and results in the apparent lower thickness of the nitrogen layer in comparison with the carbon layer. The importance of the peptide conformation can be demonstrated with the following example. As seen in Table 4, the BE of the peptide shifts slightly from the expected value when the peptide-silane is incorporated into the sol-gel matrix. This shift is very likely due to peptide folding characteristics on the surface of the thin films. Additionally, the NID and RGD peptides have C/N and C/O ratios that are somewhat more skewed from the expected, given the spectral characteristics of the free peptide-silanes. These peptides are of a length sufficient to induce minor secondary structure formation, which can affect both the presentation of particular carbon bonds through masking effects as well as increase the possibility of hydrogen bonding leading to slight BE shifts. This hypothesis is supported by the relatively low measured height of the NID and RGD peptides. Summary and Conclusions Four different free peptide-silanes, RGD, NID, KDI, and YIG, and their thin film were characterized by XPS. This work

demonstrates the functionality of XPS to characterize the surface chemistry of biologically inspired sol-gel hybrids. Peptides consisting of oxygen, nitrogen, and carbon can be characterized by the O 1s, N 1s, and C 1s signals. The N 1s and C 1s peaks were found to be good fingerprints of the peptides, whereas O 1s overlapped with the signal of substrate oxygen and, therefore, the O 1s peak was not informative in the case of the thin films. The N 1s peak originates solely from the peptide but the C 1s signal contains also the contribution of the residual hydrocarbons, which appeared during sample transfers. In order to separate the contributions of the residual carbon and the peptides, the curve-fitting analysis was employed. The peptide constraint was constructed to account for the different chemical states of carbon atoms in the peptide structure. The curve-fitting procedure was validated by analyzing free peptides, RGD, NID, KDI, and YIG, in powder forms. The XPS measured ratios between nitrogen and carbon and between oxygen and carbon were very close to those predicted from the peptide structures. This finding supports the appropriateness of the curve-fitting procedure, which was then applied for thin film studies. The XPS measured ratio between nitrogen and carbon for the peptide thin film was very close to the corresponding value calculated from the peptide structures, whereas the XPS ratio between oxygen and carbon was different due to substrate oxygen contribution. The thin films of peptides also contained the residual hydrocarbons in the amount of approximately 30% of total carbon content. The characteristic binding energies are shown in Table 4. The coverage and thickness of the peptide-silane thin films were calculated. The results are shown Table 5. The coverage of the peptide-silanes is in the range of 10% of a monolayer (1 monolayer ) 1 ML ) the number of presented molecules is equal to the number of substrate atoms). The difference in the peptide film thickness calculated from the N 1s and C 1s peaks was assigned to peptide folding. On the basis of the chemical structural characteristics of the individual peptides and the experimental XPS data, it can be reasonably assumed that the RGD and NID peptides adopt a globular conformation on the surface of the peptide ormosils, while the YIG and KDI peptides are more linear in nature. Future work will include the calibration of surface coverage of samples with differing concentrations of peptide-silanes incorporated into the matrix and also the investigation of morphology of the peptide-silane layer. Acknowledgment. The authors would like to acknowledge Jianjie Huang and Don Bergstrom for use of the peptide synthesis equipment. S.S.J. was supported by the Purdue Research Foundation. Supporting Information Available: The chemically different positions of carbon marked by numbers for the RGD peptide; the curve-fitting and the C 1s spectra obtained from the free NID, KDI, and YIG peptides at 0° collection angle; the parameters for the calculation of λA,k using eq 2. This

XPS of Sol-Gel Silica Modified with Peptides material is available free of charge via the Internet at http:// pubs.acs.org. References and Notes (1) Jedlicka, S. S.; McKenzie, J. L.; Leavesley, S. J.; Little, K. M.; Webster, T. J.; Robinson, J. P.; Nivens, D. E.; Rickus, J. L. J. Mater. Chem. 2006, 16, 3221. (2) Zolkov, C.; Avnir, D.; Armon, R. J. Mater. Chem. 2004, 14, 2200. (3) Brinker, C. J.; Scherer, G. W. Sol-Gel Science: The Physics and Chemistry of Sol-Gel Processing; Academic Press: Boston, MA, 1990. (4) Cruz-Aguado, J. A.; Chen, Y.; Zhang, Z.; Brook, M. A.; Brennan, J. D. Anal. Chem. 2004, 76, 4182. (5) Jedlicka, S. S.; Little, K. M.; Nivens, D. E.; Zemlyanov, D.; Rickus, J. L. J. Mater. Chem., in press. (6) Biltresse, S.; Attolini, M.; Marchand-Brynaert, J. Biomaterials 2005, 26, 4576. (7) Boxus, T.; Touillaux, R.; Dive, G.; Marchand-Brynaert, J. Bioorg. Med. Chem. 1998, 6, 1577. (8) Collier, T. O.; Anderson, J. M. J. Biomed. Mater. Res. 2002, 60, 487. (9) Cook, A. D.; Hrkach, J. S.; Gao, N. N.; Johnson, I. M.; Pajvani, U. B.; Cannizzaro, S. M.; Langer, R. J. Biomed. Mater. Res. 1997, 35, 513. (10) De Giglio, E.; Sabbatini, L.; Colucci, S.; Zambonin, G. J. Biomater. Sci., Polymer Ed. 2000, 11, 1073. (11) Deng, C.; Tian, H. Y.; Zhang, P. B.; Sun, J.; Chen, X. S.; Jing, X. B. Biomacromolecules 2006, 7, 590. (12) He, W.; Bellamkonda, R. V. Biomaterials 2005, 26, 2983. (13) Massia, S. P.; Holecko, M. M.; Ehteshami, G. R. J. Biomed. Mater. Res., Part A 2004, 68A, 177. (14) Santiago, L. Y.; Nowak, R. W.; Rubin, J. P.; Marra, K. G. Biomaterials 2006, 27, 2962. (15) Faucheux, N.; Schweiss, R.; Lutzow, K.; Werner, C.; Groth, T. Biomaterials 2004, 25, 2721. (16) Kim, T. G.; Park, T. G. Biotechnol. Prog. 2006, 22, 1108. (17) Betancor, L.; Lopez-Gallego, F.; Hidalgo, A.; Alonso-Morales, N.; Mateo, G.; Fernandez-Lafuente, R.; Guisan, J. M. Enzyme Microb. Technol. 2006, 39, 877. (18) Stroh, C. M.; Ebner, A.; Geretschlager, M.; Freudenthaler, G.; Kienberger, F.; Kamruzzahan, A. S. M.; Smith-Gill, S. J.; Gruber, H. J.; Hinterdorfer, P. Biophys. J. 2004, 87, 1981. (19) Wicikowski, L.; Kusz, B.; Murawski, L.; Szaniawska, K.; Susha, B. Vacuum 1999, 54, 221. (20) Yao, J.; Radin, S.; Leboy, P. S.; Ducheyne, P. Biomaterials 2005, 26, 1935. (21) Adolphi, B.; Jahne, E.; Busch, G.; Cai, X. D. Anal. Bioanal. Chem. 2004, 379, 646. (22) Berchmans, S.; Yegnaraman, V.; Sandhyarani, N.; Murty, K.; Pradeep, T. J. Electroanal. Chem. 1999, 468, 170. (23) Brizzolara, R. A.; Beard, B. C. Surf. Interface Anal. 1999, 27, 716. (24) Casaletto, M. P.; Kaciulis, S.; Mattogno, G.; Mezzi, A.; Ambrosio, L.; Branda, F. Surf. Interface Anal. 2002, 34, 45. (25) Cecchet, F.; Fioravanti, G.; Marcaccio, M.; Margotti, M.; Mattiello, L.; Paolucci, F.; Rapino, S.; Rudolf, P. J. Phys. Chem. B 2005, 109, 18427. (26) Cecchet, F.; Pilling, M.; Hevesi, L.; Schergna, S.; Wong, J. K. Y.; Clarkson, G. J.; Leigh, D. A.; Rudolf, P. J. Phys. Chem. B 2003, 107, 10863. (27) Charlier, J.; Cousty, J.; Xie, Z. X.; Vasset-Le Poulennec, C.; Bureau, C. Surf. Interface Anal. 2000, 30, 283.

J. Phys. Chem. B, Vol. 111, No. 40, 2007 11857 (28) Cho, Y.; Ivanisevic, A. J. Phys. Chem. B 2005, 109, 12731. (29) Cossement, D.; Pierard, C.; Delhalle, J.; Pireaux, J. J.; Hevesi, L.; Mekhalif, Z. Surf. Interface Anal. 2001, 31, 18. (30) Curulli, A.; Cusma, A.; Kaciulis, S.; Padeletti, G.; Pandolfi, L.; Valentini, F.; Viticoli, M. Surf. Interface Anal. 2006, 38, 478. (31) Fan, L. J.; Yang, Y. W.; Tao, Y. T. J. Electron Spectrosc. Relat. Phenom. 2005, 144, 433. (32) Sinapi, F.; Delhalle, J.; Mekhalif, Z. Mater. Sci. Eng., C 2002, 22, 345. (33) Sinapi, F.; Naji, A.; Delhalle, J.; Mekhalif, Z. Surf. Interface Anal. 2004, 36, 1484. (34) Song, S. H.; Park, K. M.; Chang, S. M.; Nakamura, C.; Miyake, J.; Kim, W. S. Mol. Cryst. Liq. Cryst. 2003, 407, 531. (35) Surface Analysis by Auger and X-ray Photoelectron Spectroscopy; Briggs, D., Grant, J. T., Eds.; IM Publication and SurfaceSpectra Limited: Chicchester, U.K., 2003, pp 899. (36) Longo, L.; Vasapollo, G.; Guascito, M. R.; Malitesta, C. Anal. Bioanal. Chem. 2006, 385, 146. (37) Cumpson, P. J. Surf. Interface Anal. 2001, 31, 23. (38) Bicerano, J., Ed. Prediction of Polymer Properties, 2nd ed., revised and expanded, Marcel Dekker: New York, 1996. (39) Kier, L. B.; Hall, L. H. Molecular ConnectiVity in Structure-ActiVity Analysis, Chemometrics Series, 9; Wiley, New York, 1986. (40) Jablonski, A.; Powell, C. J. Surf. Sci. Rep. 2002, 47, 33. (41) NIST Electron Effective-Attenuation-Length Database, NIST Standard Reference Database 82, version 1.1; National Institute of Standards and Technology: Gaitherburg, MD, 2003. (42) Fadley, C. S. Electron Spectrosc.: Theory, Tech. Appl. 1978, 2, 1. (43) Chen, M.; Nie, M. Y.; Li, H. L. J. Colloid Interface Sci. 1999, 209, 421. (44) Farah, A. A.; Voicu, R.; Barjovanu, R.; Bensebaa, F.; Tufa, K.; Faid, K. Appl. Surf. Sci. 2006, 252, 5158. (45) Killampalli, A. S.; Ma, P. F.; Engstrom, J. R. J. Am. Chem. Soc. 2005, 127, 6300. (46) Lee, C. Y.; Gong, P.; Harbers, G. M.; Grainger, D. W.; Castner, D. G.; Gamble, L. J. Anal. Chem. 2006, 78, 3316. (47) Pimanpang, S.; Wang, P. I.; Wang, G. C.; Lu, T. M. Appl. Surf. Sci. 2006, 252, 3532. (48) Shin, H. Y.; Jung, J. Y.; Kim, S. W.; Lee, W. K. J. Ind. Eng. Chem. 2006, 12, 476. (49) Sinapi, F.; Deroubaix, S.; Pirlot, C.; Delhalle, J.; Mekhalif, Z. Electrochim. Acta 2004, 49, 2987. (50) Sinapi, F.; Forget, L.; Delhalle, J.; Mekhalif, Z. Appl. Surf. Sci. 2003, 212, 464. (51) Steffens, G. C. M.; Nothdurft, L.; Buse, G.; Thissen, H.; Hocker, H.; Klee, D. Biomaterials 2002, 23, 3523. (52) Fabris, L.; Antonello, S.; Armelao, L.; Donkers, R. L.; Polo, F.; Toniolo, C.; Maran, F. J. Am. Chem. Soc. 2006, 128, 326. (53) Popat, K. C.; Swan, E. E. L.; Desai, T. A. Langmuir 2005, 21, 7061. (54) Clochard, M. C.; Betz, N.; Goncalves, M.; Bittencourt, C.; Pireaux, J. J.; Gionnet, K.; Deleris, G.; Le Moel, A. Nucl. Instrum. Methods Phys. Res., Sect. B 2005, 236, 208. (55) Saprigin, A. V.; Thomas, C. W.; Dulcey, C. S.; Patterson, C. H.; Spector, M. S. Surf. Interface Anal. 2005, 37, 24. (56) Wagner, A. J.; Wolfe, G. M.; Fairbrother, D. H. Appl. Surf. Sci. 2003, 219, 317.