**2.3 Surface treatment procedure**

The silanization of nat samples was performed in the gas phase with APTES. To this end, nat samples were placed, together with a small vial containing 100 µL APTES, in a 7 dm3 closed recipient under vacuum for 30 min at room temperature. The samples were cured for 1 h at 100°C under vacuum, rinsed and incubated for 6 h in milliQ water to eliminate the excess of non attached silanes. These are called "sil".

Both nat and sil samples were subjected to treatments with the coupling agent (BS) and/or the enzyme (Gox) as detailed below:


After BS or Gox treatment, the samples were rinsed three times with milliQ water and dried under nitrogen gas flow.

In order to clarify the issue of surface contamination, nat samples were further cleaned by UV-ozone treatment (UVO Cleaner, Jelight Co, Irvine, Ca, USA) during 20 minutes. They were then placed in a Petri dish, left in the laboratory environment, and submitted to water contact angle measurements as a function of time.

#### **2.4 X-ray photoelectron spectroscopy**

XPS analyses were performed using a Kratos Axis Ultra spectrometer (Kratos Analytical, UK), equipped with a monochromatized aluminum X-ray source (powered at 10 mA and 15 kV) and an eight channeltrons detector. No charge stabilization device was used on these conductive samples. Analyses were performed in the hybrid lens mode with the slot aperture; the resulting analyzed area was 700 µm ¯ 300 µm. A pass energy of 40 eV was used for narrow scans. In these conditions, the full width at half maximum (FWHM) of the Ag 3d5/2 peak of clean silver reference sample was about 0.9 eV. The samples were fixed on the support using a double-sided adhesive conducting tape. The pressure in the analysis chamber was around 10-6 Pa. The photoelectron collection angle *θ* between the normal to the sample surface and the electrostatic lens axis was 0° or 60°. The following sequence of spectra was recorded: survey spectrum, C 1s, O 1s, N 1s, P 2p, Cr 2p, Fe 2p, Ni 2p, Mo 3d, Na 1s, S 2p, Si 2p and C 1s again to check for charge stability as a function of time, and absence of sample degradation. The binding energy scale was set by fixing the C 1s

Silanization with APTES for Controlling the Interactions


Height (nm)


Height (nm)

Between Stainless Steel and Biocomponents: Reality vs Expectation 107

tendency regarding the C 1s and O 1s peaks was also observed after BS and/or Gox treatment of nat sample, but was less pronounced (data not shown). It may be attempted to

**nat sil** 


Height (nm)

Lateral displacement (µm) Lateral displacement (µm)

0 0.2 0.4 0.6 0.8 1

0 0.2 0.4 0.6 0.8 1

Fig. 1. AFM height images (1¯1 µm2, contact mode, in water; z scale 10 nm) of native (nat), silanized stainless steel (sil), the same after adsorption of glucose oxidase (sil+Gox) and after

Lateral displacement (µm) Lateral displacement (µm)

**sil+Gox sil+BS+Gox** 


Height (nm)

treatment with glucose oxidase subsequent to treatment with the coupling agent (sil+BS+Gox). Cross sections were taken at the place indicated by the dashed lines.

0 0.2 0.4 0.6 0.8 1

0 0.2 0.4 0.6 0.8 1

component due to carbon only bound to carbon and hydrogen at 284.8 eV. The data treatment was performed with the Casa XPS software (Casa Software Ltd., UK). The peaks were decomposed using a linear baseline, and a component shape defined by the product of a Gauss and Lorentz function, in the 70:30 ratio, respectively. Molar concentration ratios were calculated using peak areas normalized according to the acquisition parameters and the relative sensitivity factors and transmission functions provided by the manufacturer.

#### **2.5 Atomic force microscopy (AFM)**

The surface topography was examined using a commercial AFM (NanoScope III MultiMode AFM, Veeco Metrology LLC, Santa Barbara, CA) equipped with a 125 µm ¯ 125 µm ¯ 5 µm scanner (J-scanner). A quartz fluid cell was used without the O-ring. Topographic images were recorded in contact mode using oxide-sharpened microfabricated Si3N4 cantilevers (Microlevers, Veeco Metrology LLC, Santa Barbara, CA) with a spring constant of 0.01 N.m-1 (manufacturer specified), with a minimal applied force (<500 pN) and at a scan rate of 5-6 Hz. The curvature radius of silicon nitride tips was about 20 nm. Images were obtained at room temperature (21-24°C) in milliQ water. All images shown in this paper were flattened data using a third order polynomial. The surface roughness (Rrms) was computed over an area of 1 µm ¯ 1 µm using the Veeco software.

#### **2.6 Water contact angle measurements**

Water contact angles were measured at room temperature using the sessile drop method and image analysis of the drop profile. The instrument, using a CCD camera and an image analysis processor, was purchased from Electronisch Ontwerpbureau De Boer (The Netherlands). The water (milliQ) droplet volume was 0.3 μL, and the contact angle was measured 5 s after the drop was deposited on the sample. For each sample, the reported value is the average of the results obtained on 5 droplets.

#### **3. Results**

AFM images obtained in water on SS samples after different treatments are presented in Figure 1. The nat sample showed the presence of nanoparticles with different sizes (nat, Figure 1, Rrms = 3.2 nm), in agreement with previous results. The formation of nanoparticles, presumably made of ferric hydroxide, resulted from oxidation occurring during the 48 h of immersion subsequent to polishing (Landoulsi et al., 2008a). The surface of silanized SS exhibited particles with a bigger size (sil, Figure 1) and the roughness decreased slightly (sil, Figure 1, Rrms = 2.4 nm). The treatment with Gox, with or without previous treatment with BS, led to the formation of particles with a more uniform size and a higher density, in comparison with nat sample, and no appreciable change of surface roughness (Rrms = 1.7 nm for sil+Gox and 2.5 nm for sil+BS+Gox, Figure 1).

XPS is a suitable method to obtain information regarding the different constituents at the surface (substrate, silane, other organic compounds). The elemental composition of the samples is given in Table 2. Representative C 1s and O 1s peaks recorded on SS surface prior to and after silanization, and after further treatments are given in Figure 2. After BS and/or Gox treatment of silanized SS, a relative increase of the component around 531.2 eV in the O 1s peak was observed (Figure 2). In the C 1s peak, an increase of the components at 286.3 and 288.7 eV was also clear, while the main component remained at 284.8 eV. The same 106 Biomaterials – Physics and Chemistry

component due to carbon only bound to carbon and hydrogen at 284.8 eV. The data treatment was performed with the Casa XPS software (Casa Software Ltd., UK). The peaks were decomposed using a linear baseline, and a component shape defined by the product of a Gauss and Lorentz function, in the 70:30 ratio, respectively. Molar concentration ratios were calculated using peak areas normalized according to the acquisition parameters and the relative sensitivity factors and transmission functions provided by the manufacturer.

The surface topography was examined using a commercial AFM (NanoScope III MultiMode AFM, Veeco Metrology LLC, Santa Barbara, CA) equipped with a 125 µm ¯ 125 µm ¯ 5 µm scanner (J-scanner). A quartz fluid cell was used without the O-ring. Topographic images were recorded in contact mode using oxide-sharpened microfabricated Si3N4 cantilevers (Microlevers, Veeco Metrology LLC, Santa Barbara, CA) with a spring constant of 0.01 N.m-1 (manufacturer specified), with a minimal applied force (<500 pN) and at a scan rate of 5-6 Hz. The curvature radius of silicon nitride tips was about 20 nm. Images were obtained at room temperature (21-24°C) in milliQ water. All images shown in this paper were flattened data using a third order polynomial. The surface roughness (Rrms) was computed over an

Water contact angles were measured at room temperature using the sessile drop method and image analysis of the drop profile. The instrument, using a CCD camera and an image analysis processor, was purchased from Electronisch Ontwerpbureau De Boer (The Netherlands). The water (milliQ) droplet volume was 0.3 μL, and the contact angle was measured 5 s after the drop was deposited on the sample. For each sample, the reported

AFM images obtained in water on SS samples after different treatments are presented in Figure 1. The nat sample showed the presence of nanoparticles with different sizes (nat, Figure 1, Rrms = 3.2 nm), in agreement with previous results. The formation of nanoparticles, presumably made of ferric hydroxide, resulted from oxidation occurring during the 48 h of immersion subsequent to polishing (Landoulsi et al., 2008a). The surface of silanized SS exhibited particles with a bigger size (sil, Figure 1) and the roughness decreased slightly (sil, Figure 1, Rrms = 2.4 nm). The treatment with Gox, with or without previous treatment with BS, led to the formation of particles with a more uniform size and a higher density, in comparison with nat sample, and no appreciable change of surface roughness (Rrms = 1.7 nm

XPS is a suitable method to obtain information regarding the different constituents at the surface (substrate, silane, other organic compounds). The elemental composition of the samples is given in Table 2. Representative C 1s and O 1s peaks recorded on SS surface prior to and after silanization, and after further treatments are given in Figure 2. After BS and/or Gox treatment of silanized SS, a relative increase of the component around 531.2 eV in the O 1s peak was observed (Figure 2). In the C 1s peak, an increase of the components at 286.3 and 288.7 eV was also clear, while the main component remained at 284.8 eV. The same

**2.5 Atomic force microscopy (AFM)** 

area of 1 µm ¯ 1 µm using the Veeco software.

value is the average of the results obtained on 5 droplets.

for sil+Gox and 2.5 nm for sil+BS+Gox, Figure 1).

**2.6 Water contact angle measurements** 

**3. Results** 

tendency regarding the C 1s and O 1s peaks was also observed after BS and/or Gox treatment of nat sample, but was less pronounced (data not shown). It may be attempted to

Fig. 1. AFM height images (1¯1 µm2, contact mode, in water; z scale 10 nm) of native (nat), silanized stainless steel (sil), the same after adsorption of glucose oxidase (sil+Gox) and after treatment with glucose oxidase subsequent to treatment with the coupling agent (sil+BS+Gox). Cross sections were taken at the place indicated by the dashed lines.

Silanization with APTES for Controlling the Interactions

ether and ester).

**nat**

**sil**

Intensity (a.u.)

Intensity (a.u.)

Intensity (a.u.) 538 536 534 532 530 528 526 Binding Energy (eV)

538 536 534 532 530 528 526 Binding Energy (eV)

**sil+Gox**

538 536 534 532 530 528 526 Binding Energy (eV)

N 1s peak is overlapped with a Mo 3p3*/*2 contribution.

Between Stainless Steel and Biocomponents: Reality vs Expectation 109

The O 1s peak was decomposed in three components (Landoulsi et al., 2008a). The first one, at 529.7 eV, is due to inorganic oxygen in metal oxides (M–O) (NIST Database). The FWHM of the two other components were arbitrarily imposed to be equal. The component at about 531.2 eV may be due to oxygen making a double bond with carbon (C=O including amide and carboxyl group) and to oxygen of carboxylate. Contributions of metal hydroxides (M– O–H) as well as oxygen bound to silicon [Si–O] in silane are overlapping with this component (NIST Database; Genet et al., 2008). The last component, at 533.1 eV, is attributed to oxygen making single bonds with carbon (C–O–H of alcohol and carboxyl, C–O–C of

**O 1s N 1s C 1s**

Mo 3p3/2

404 402 400 398 396 394 392 390 Binding Energy (eV)

404 402 400 398 396 394 392 390 Binding Energy (eV)

Binding Energy (eV)

Fig. 3. Decomposition of O 1s, N 1s and C 1s peaks recorded on: native stainless steel (nat), silanized stainless steel (sil) and the same after adsorption of glucose oxidase (sil+Gox). The

404 402 400 398 396 394 392 390 Binding Energy (eV) 538 534 530 526 404 400 394 392 294 290 286 282

294 292 290 288 286 284 282 280 Binding Energy (eV)

294 292 290 288 286 284 282 280 Binding Energy (eV)

294 292 290 288 286 284 282 280 Binding Energy (eV)

extract chemical information by careful decomposition of the peaks. This requires to impose reasonable constraints (number of components, full width at half maximum FWHM) in order to insure reliable comparisons, and to check the chemical relevance of the results by examining correlations between spectral data of different natures (Genet et al., 2008; Rouxhet & Genet, 2011). In previous studies (Landoulsi et al., 2008a; Landoulsi et al., 2008b), we have demonstrated the usefulness of this approach, even when the evolution of the C 1s and O 1s peak shape is weak, in order to obtain information on the amount and the nature of organic and inorganic constituents on SS surfaces.

Fig. 2. O 1s and C 1s peaks of native (nat), of silanized stainless steel (sil), of the same after treatment with the coupling agent (sil+BS), after adsorption of glucose oxidase (sil+Gox) and after treatment with glucose oxidase subsequent to treatment with the coupling agent (sil+BS+Gox).

Figure 3 presents typical O 1s, N 1s and C 1s XPS peaks recorded on native SS (nat), silanized (sil) and the same after Gox treatment (sil+Gox). For the decomposition of these peaks, reasonable constraints were applied, based on our experience with the XPS analysis of biosurfaces (Genet et al., 2008, Rouxhet & Genet, 2011). The C ls peak was decomposed in four components, the FWHM of which were imposed to be equal: (i) a component at 284.8 eV due to carbon only bound to carbon and/or hydrogen [C–(C,H)]; (ii) a component at about 286.3 eV due to carbon making a single bond with oxygen and/or nitrogen [C–(O,N)] in alcohol, amine, or amide; (iii) a component at 287.8 eV due to carbon making one double bond or two single bonds with oxygen (C=O, O–C–O) and (iv) a component at 288.7 eV attributed to carboxyl or ester functions [(C=O)–O–R].

extract chemical information by careful decomposition of the peaks. This requires to impose reasonable constraints (number of components, full width at half maximum FWHM) in order to insure reliable comparisons, and to check the chemical relevance of the results by examining correlations between spectral data of different natures (Genet et al., 2008; Rouxhet & Genet, 2011). In previous studies (Landoulsi et al., 2008a; Landoulsi et al., 2008b), we have demonstrated the usefulness of this approach, even when the evolution of the C 1s and O 1s peak shape is weak, in order to obtain information on the amount and the nature

> Intensity (a.u.)

Fig. 2. O 1s and C 1s peaks of native (nat), of silanized stainless steel (sil), of the same after treatment with the coupling agent (sil+BS), after adsorption of glucose oxidase (sil+Gox) and after treatment with glucose oxidase subsequent to treatment with the coupling agent

Figure 3 presents typical O 1s, N 1s and C 1s XPS peaks recorded on native SS (nat), silanized (sil) and the same after Gox treatment (sil+Gox). For the decomposition of these peaks, reasonable constraints were applied, based on our experience with the XPS analysis of biosurfaces (Genet et al., 2008, Rouxhet & Genet, 2011). The C ls peak was decomposed in four components, the FWHM of which were imposed to be equal: (i) a component at 284.8 eV due to carbon only bound to carbon and/or hydrogen [C–(C,H)]; (ii) a component at about 286.3 eV due to carbon making a single bond with oxygen and/or nitrogen [C–(O,N)] in alcohol, amine, or amide; (iii) a component at 287.8 eV due to carbon making one double bond or two single bonds with oxygen (C=O, O–C–O) and (iv) a component at 288.7 eV

294 292 290 288 286 284 282 280 Binding Energy (eV)

**C 1s**

294 292 290 288 286 284 282 280 Binding Energy (eV)

of organic and inorganic constituents on SS surfaces.

**O 1s**

sil+BS+Gox

sil+Gox

sil+BS

sil

Intensity (a.u.)

nat

(sil+BS+Gox).

538 536 534 532 530 528 526 Binding Energy (eV)

538 536 534 532 530 528 526 Binding Energy (eV)

attributed to carboxyl or ester functions [(C=O)–O–R].

The O 1s peak was decomposed in three components (Landoulsi et al., 2008a). The first one, at 529.7 eV, is due to inorganic oxygen in metal oxides (M–O) (NIST Database). The FWHM of the two other components were arbitrarily imposed to be equal. The component at about 531.2 eV may be due to oxygen making a double bond with carbon (C=O including amide and carboxyl group) and to oxygen of carboxylate. Contributions of metal hydroxides (M– O–H) as well as oxygen bound to silicon [Si–O] in silane are overlapping with this component (NIST Database; Genet et al., 2008). The last component, at 533.1 eV, is attributed to oxygen making single bonds with carbon (C–O–H of alcohol and carboxyl, C–O–C of ether and ester).

Binding Energy (eV)

Fig. 3. Decomposition of O 1s, N 1s and C 1s peaks recorded on: native stainless steel (nat), silanized stainless steel (sil) and the same after adsorption of glucose oxidase (sil+Gox). The N 1s peak is overlapped with a Mo 3p3*/*2 contribution.

Silanization with APTES for Controlling the Interactions

primary amide and ester functions. Accordingly,

relative contribution of the constituents of the organic adlayer.

al., 2008 ; Landoulsi et al., 2008a).

by:

(Table 2).

**4. Discussion** 

order of 1 to 1.5.

**4.1 Passive film composition** 

Between Stainless Steel and Biocomponents: Reality vs Expectation 111

concentration of carbon in oxidized form, *Cox*, in consistency with alcohol, primary amine,

where the name of an element in italic designates its concentration and the number in subscript designates the binding energy of the peak component. Errors would occur in case of a high concentration of polysaccharides (*Cox*/*O* = 6/5) or carboxyl (*Cox*/*O* = 1/2) (Genet et

The sum of the concentrations of the elements present in organic compounds is then given

For sake of uniformity, all spectral data involved in correlations below are ratioed to *∑org*

The concentration of the main elements or functions due to organic compounds, obtained at photoelectron collection angle *θ* = 0°, is plotted in Figure 4 as a function of the same quantity obtained at *θ* = 60°. A 1:1 relationship is obtained for all elements or functions and all samples, indicating no significant effect of the photoelectron collection angle *θ* on the

Table 2 shows that the apparent concentrations of metal elements varied only slightly according to the surface treatment. The main change concerned the decrease of the Feox concentration for sil+BS, sil+Gox and sil+BS+Gox samples. However, no change in the shape of the Fe 2p3/2 peak was observed (data not shown). For these samples, a significant decrease of the molar concentration of *O529.7* was also noticed (Table 2). It appears that the oxide layer of SS passive film, after incubation in the aqueous medium for 48 h (nat sample), was mainly constituted with a mixture of Fe and Cr oxides/hydroxides and small amounts of partially oxidized Ni and Mo. This is in agreement with a previous study (Landoulsi et al., 2008a), however in the latter, the stoichiometry of the passive film was not computed. The O 1s component at 529.7 eV is due to metal oxides. By considering that Moox is in the form of MoO3 (Landoulsi et al., 2008a), the difference between the *O529.7* concentration and three times the Moox concentration should be due to Fe and Cr oxides. Figure 5 presents the relation between this difference and the sum of Feox and Crox concentrations. All data show reasonable linear regressions. The shift of the dots along the line when the photoelectron collection angle changes from 0° to 60° is due to the presence of the organic constituents on top of stainless steel. The average ratio between the *y* and *x* scales is 0.80 and 0.96 at *θ* = 0 and 60°, respectively; the slope of the regression lines is 1.03 (s.d. 0.23) and 1.42 (s.d. 0.12), respectively. Thus, the ratio oxide/metal ions in chromium and iron oxyhydroxides is of the

The evaluation of the quantity of hydroxide associated to Fe and Cr is complex due to the multiple chemical functions overlapping in the O531.2 component. Niox is in the form of Ni(OH)2 (component at ~855.6 eV (Briggs & Seah, 1990, Zhou et al., 2006), spectra not shown). The amount of oxygen associated to silane depends on the products of APTES

*Oorg = Cox - N = C286.3 + C287.8 + C288.7 – Ntot* (1)

*∑org = Ctot + Oorg + Ntot + Siorg = Ctot + Cox + Siorg* (2)

The N 1s peak showed a main component at 400.0 eV attributed to amide or amine (N–C). An additional component appeared clearly near 401.6 eV in silanized SS samples (sil and sil+Gox, Figure 3), indicating the presence of protonated amines. The contributions at lower binding energies in the N 1s spectral window are due to Mo 3p3*/*2 components (Olefjord & Wegrelius, 1996); the reliability of their quantification was checked by comparison with the Mo concentration deduced from the Mo 3p3/2 and Mo 3d peaks (Landoulsi et al., 2008a). The N 1s contribution of sil shows a shape which is in agreement with spectra reported in the literature for APTES-modified surfaces (Suzuki et al., 2006; Xiao et al., 1997). It was not found justified to decompose it in three components attributed to amine, amide and protonated amine, respectively, as done in (Suzuki et al., 2006). The surface concentrations (mole fraction) associated with the components of C 1s, O 1s and N 1s peaks are given in Table 2.


Table 2. Surface concentration (mole fraction (%) computed over the sum of all elements except hydrogen) of elements determined by XPS (*θ* = 0°) on stainless steel samples.

The Si 2p peak was decomposed in two components, at 99.3 and 101.8 eV, attributed to non oxidized silicon in SS (Si0) and to silicon of silane (Siorg), respectively. It was not decomposed in Si 2p3/2 and Si 2p1/2 contributions because these are very close in energy. The decomposition procedure for Fe 2p3/2, Cr 2p, Ni 2p3/2 , and Mo 3d peaks was described before (Landoulsi et al., 2008a). The Fe concentration may be underestimated due to the procedure used to treat the complex baseline of the Fe 2p peak. The distinction between contributions of oxidized (Mox) and nonoxidized (Mmet) metal elements was easily made. The concentrations obtained are also given in Table 2.

The concentration of oxygen present in organic compounds *Oorg* may not directly be deduced from the O 1s peak owing to the overlap with inorganic hydroxide. However, for biological systems the sum of O and N concentrations may be evaluated by the concentration of carbon in oxidized form, *Cox*, in consistency with alcohol, primary amine, primary amide and ester functions. Accordingly,

$$CO\_{org} = C\_{ox} \cdot N = C\_{286.3} + C\_{287.8} + C\_{288.7} - N\_{tvt} \tag{1}$$

where the name of an element in italic designates its concentration and the number in subscript designates the binding energy of the peak component. Errors would occur in case of a high concentration of polysaccharides (*Cox*/*O* = 6/5) or carboxyl (*Cox*/*O* = 1/2) (Genet et al., 2008 ; Landoulsi et al., 2008a).

The sum of the concentrations of the elements present in organic compounds is then given by:

$$
\Sigma org = \mathbb{C}\_{tot} + O\_{org} + N\_{tot} + Si\_{org} = \mathbb{C}\_{tot} + \mathbb{C}\_{ox} + Si\_{org} \tag{2}
$$

For sake of uniformity, all spectral data involved in correlations below are ratioed to *∑org* (Table 2).

The concentration of the main elements or functions due to organic compounds, obtained at photoelectron collection angle *θ* = 0°, is plotted in Figure 4 as a function of the same quantity obtained at *θ* = 60°. A 1:1 relationship is obtained for all elements or functions and all samples, indicating no significant effect of the photoelectron collection angle *θ* on the relative contribution of the constituents of the organic adlayer.

### **4. Discussion**

110 Biomaterials – Physics and Chemistry

The N 1s peak showed a main component at 400.0 eV attributed to amide or amine (N–C). An additional component appeared clearly near 401.6 eV in silanized SS samples (sil and sil+Gox, Figure 3), indicating the presence of protonated amines. The contributions at lower binding energies in the N 1s spectral window are due to Mo 3p3*/*2 components (Olefjord & Wegrelius, 1996); the reliability of their quantification was checked by comparison with the Mo concentration deduced from the Mo 3p3/2 and Mo 3d peaks (Landoulsi et al., 2008a). The N 1s contribution of sil shows a shape which is in agreement with spectra reported in the literature for APTES-modified surfaces (Suzuki et al., 2006; Xiao et al., 1997). It was not found justified to decompose it in three components attributed to amine, amide and protonated amine, respectively, as done in (Suzuki et al., 2006). The surface concentrations (mole fraction) associated with the components of C 1s, O 1s and N 1s peaks are given in

Ni ox Ni met **Ni tot** Fe ox Fe met **Fe tot** Cr ox Cr met **Cr tot** Mo ox Mo met **Mo tot Si 0**

C 288.7 C 287.8 C 286.3 C 284.8 **C tot** O 533.1 O 531.2 O 529.7 **O tot Oorg N401.6 N400 Ntot Si org ∑org**

**nat** 1.23 0.73 **1.95** 7.59 1.57 **9.16** 4.69 0.33 **5.02** 0.34 0.12 **0.46 0.38 +BS** 0.81 0.54 **1.35** 8.32 1.89 **10.21** 5.96 0.40 **6.36** 0.46 0.12 **0.57 0.29 +Gox** 0.59 0.51 **1.10** 8.63 1.70 **10.33** 5.53 0.33 **5.86** 0.37 0.12 **0.49 - +BS+Gox** 0.66 0.39 **1.05** 8.90 1.43 **10.33** 4.56 0.30 **4.87** 0.32 0.12 **0.44 0.36**

**sil** 0.47 0.33 **0.80** 7.89 1.22 **9.10** 4.59 0.27 **4.86** 0.31 0.10 **0.40 0.37 sil+BS** 0.47 0.43 **0.91** 6.91 1.45 **8.36** 5.24 0.33 **5.57** 0.35 0.11 **0.45 0.32 sil+Gox** 0.60 0.38 **0.98** 6.14 1.29 **7.42** 4.95 0.29 **5.23** 0.32 0.12 **0.44 0.26 sil+BS+Gox** 0.54 0.32 **0.86** 6.22 1.16 **7.39** 4.58 0.24 **4.82** 0.34 0.09 **0.42 0.30**

**nat** 2.45 2.29 3.92 39.62 **48.28** 2.36 19.02 12.60 **33.98 7.92** 0.00 0.73 **0.73 0.44 57.37 +BS** 2.89 2.08 3.88 34.87 **43.71** 2.48 18.96 13.86 **35.30 7.64** 0.00 1.21 **1.21 0.43 52.99 +Gox** 2.03 2.16 4.78 35.56 **44.53** 2.97 20.96 11.27 **35.20 7.50** 0.13 1.34 **1.48 - 53.50 +BS+Gox** 2.55 2.60 5.89 30.02 **41.06** 3.96 22.01 12.57 **38.54 9.46** 0.23 1.35 **1.58 0.89 52.99 sil** 1.15 2.37 4.57 36.13 **44.23** 1.95 21.22 12.28 **35.45 5.93** 0.63 1.54 **2.17 2.29 54.61 sil+BS** 2.03 3.30 6.48 32.72 **44.53** 1.07 23.16 9.71 **33.94 8.49** 0.66 2.65 **3.31 2.13 58.47 sil+Gox** 1.18 5.97 10.39 25.08 **42.62** 4.74 21.09 8.86 **34.70 11.78** 0.59 5.16 **5.75 1.74 61.90 sil+BS+Gox** 1.46 4.89 8.31 30.39 **45.05** 2.87 22.24 8.52 **33.63 9.86** 0.50 4.30 **4.80 2.01 61.72** Table 2. Surface concentration (mole fraction (%) computed over the sum of all elements except hydrogen) of elements determined by XPS (*θ* = 0°) on stainless steel samples.

The Si 2p peak was decomposed in two components, at 99.3 and 101.8 eV, attributed to non oxidized silicon in SS (Si0) and to silicon of silane (Siorg), respectively. It was not decomposed in Si 2p3/2 and Si 2p1/2 contributions because these are very close in energy. The decomposition procedure for Fe 2p3/2, Cr 2p, Ni 2p3/2 , and Mo 3d peaks was described before (Landoulsi et al., 2008a). The Fe concentration may be underestimated due to the procedure used to treat the complex baseline of the Fe 2p peak. The distinction between contributions of oxidized (Mox) and nonoxidized (Mmet) metal elements was easily made.

The concentration of oxygen present in organic compounds *Oorg* may not directly be deduced from the O 1s peak owing to the overlap with inorganic hydroxide. However, for biological systems the sum of O and N concentrations may be evaluated by the

The concentrations obtained are also given in Table 2.

Table 2.

#### **4.1 Passive film composition**

Table 2 shows that the apparent concentrations of metal elements varied only slightly according to the surface treatment. The main change concerned the decrease of the Feox concentration for sil+BS, sil+Gox and sil+BS+Gox samples. However, no change in the shape of the Fe 2p3/2 peak was observed (data not shown). For these samples, a significant decrease of the molar concentration of *O529.7* was also noticed (Table 2). It appears that the oxide layer of SS passive film, after incubation in the aqueous medium for 48 h (nat sample), was mainly constituted with a mixture of Fe and Cr oxides/hydroxides and small amounts of partially oxidized Ni and Mo. This is in agreement with a previous study (Landoulsi et al., 2008a), however in the latter, the stoichiometry of the passive film was not computed.

The O 1s component at 529.7 eV is due to metal oxides. By considering that Moox is in the form of MoO3 (Landoulsi et al., 2008a), the difference between the *O529.7* concentration and three times the Moox concentration should be due to Fe and Cr oxides. Figure 5 presents the relation between this difference and the sum of Feox and Crox concentrations. All data show reasonable linear regressions. The shift of the dots along the line when the photoelectron collection angle changes from 0° to 60° is due to the presence of the organic constituents on top of stainless steel. The average ratio between the *y* and *x* scales is 0.80 and 0.96 at *θ* = 0 and 60°, respectively; the slope of the regression lines is 1.03 (s.d. 0.23) and 1.42 (s.d. 0.12), respectively. Thus, the ratio oxide/metal ions in chromium and iron oxyhydroxides is of the order of 1 to 1.5.

The evaluation of the quantity of hydroxide associated to Fe and Cr is complex due to the multiple chemical functions overlapping in the O531.2 component. Niox is in the form of Ni(OH)2 (component at ~855.6 eV (Briggs & Seah, 1990, Zhou et al., 2006), spectra not shown). The amount of oxygen associated to silane depends on the products of APTES

Silanization with APTES for Controlling the Interactions

0.0 0.1 0.2 0.3 0.4

hydroxide in the passive film may be given as follows:

where *a* can take the values of 0, 0.5, 1, 1.5 or 2 (Figure 6).

0.0

0.1

(O529.7

(,).

stratification.

− 3Mo

ox)/Σorg

0.2

0.3

0.4

Between Stainless Steel and Biocomponents: Reality vs Expectation 113

0.0

(Feox + Cr ox)/Σorg (Feox + Cr ox)/Σorg

Fig. 5. Relations between molar concentrations ratioed to the sum of organic elements (Σorg) measured by XPS at θ = 0° (data from Table 2) and θ = 60° on native (open symbols) or silanized stainless steel (closed symbols), as such (¡,) or further treated with coupling agent BS (S,U), glucose oxidase (z,{) or coupling agent followed by glucose oxidase

reaction. The different possibilities, corresponding to the relative importance of grafting with respect to polymerization, are shown in Figure 6 and characterized by a, defined as the sum, over oxygen atoms which are not bound to a metal element, of the inverse of the number of bonds oxygen forms with silicon. Accordingly the concentration of inorganic

*OHinorg = Otot – Oorg – O529.7 – a* 

Depending on the value of *a*, the *y* : *x* ratio varies from 1.1 (*a* = 0) to 0.8 (*a* = 2).

Figure 7 shows the plot of the concentration of hydroxide which should be associated with Fe and Cr, considering different values of *a*. Taking silicon into account (a≠0) brings the silanized substrates better in line with the non-silanized substrates and the correlation improves as *a* increases, indicating that silane is polymerized and not just grafted.

These observations indicate that the stoichiometry of the Cr and Fe oxyhydroxide at the surface is close to (Fe,Cr)OOH. Many studies have reported a stratification in the passive film and the presence of Fe and Cr oxyhydroxides in the outermost layer (Le Bozec et al., 2001). In our case, no significant effect of the photoelectron collection angle appears in the *OHinorg/O529.7* ratio (Figure 8), whatever the value selected for *a* to evaluate *OHinorg*. However it must be kept in mind that the surface roughness revealed by Figure 1, which is of the order of the inelastic mean free path of photoelectrons in oxides, may mask the effect of a

×

0.0 0.1 0.2 0.3 0.4

 *Siorg* (3)

0.1

0.2

0.3

0.4

**θ = 0° θ = 60°**

Fig. 4. Plots of molar concentrations ratioed to the sum of organic elements (Σorg) measured by XPS at θ = 60° *vs* θ = 0° on native (open symbols) or silanized stainless steel (closed symbols), as such (¡,) or further treated with coupling agent BS (S,U), glucose oxidase (z,{) or coupling agent followed by glucose oxidase (,). The dashed lines represent a *y/x* ratio of 1:1.

**C284.8 / Σorg**

0.0 0.2 0.4 0.6 0.8 1.0

at θ = 0°

0.00

0.00

0.02

0.04

0.06

**C286.3 / Σorg Siorg / Σorg**

0.00 0.04 0.08 0.12

0.00 0.02 0.04 0.06 at θ = 0°

0.04

0.08

0.12

**Ntot / Σorg C287.8 / Σorg**

Fig. 4. Plots of molar concentrations ratioed to the sum of organic elements (Σorg) measured by XPS at θ = 60° *vs* θ = 0° on native (open symbols) or silanized stainless steel (closed symbols), as such (¡,) or further treated with coupling agent BS (S,U), glucose oxidase (z,{) or coupling agent followed by glucose oxidase (,). The dashed lines represent a

at θ = 0°

0.00 0.05 0.10 0.15 0.20

0.00 0.04 0.08 0.12

0.0

0.2

0.4

at

θ = 60°

0.6

0.8

1.0

*y/x* ratio of 1:1.

0.00

0.00

0.05

at

θ = 60°

0.10

0.15

0.20

0.04

at

θ = 60°

0.08

0.12

Fig. 5. Relations between molar concentrations ratioed to the sum of organic elements (Σorg) measured by XPS at θ = 0° (data from Table 2) and θ = 60° on native (open symbols) or silanized stainless steel (closed symbols), as such (¡,) or further treated with coupling agent BS (S,U), glucose oxidase (z,{) or coupling agent followed by glucose oxidase (,).

reaction. The different possibilities, corresponding to the relative importance of grafting with respect to polymerization, are shown in Figure 6 and characterized by a, defined as the sum, over oxygen atoms which are not bound to a metal element, of the inverse of the number of bonds oxygen forms with silicon. Accordingly the concentration of inorganic hydroxide in the passive film may be given as follows:

$$\rm OH\_{inorg} = O\_{bt} - O\_{org} - O\_{529.7} - a \times Si\_{org} \tag{3}$$

where *a* can take the values of 0, 0.5, 1, 1.5 or 2 (Figure 6).

Figure 7 shows the plot of the concentration of hydroxide which should be associated with Fe and Cr, considering different values of *a*. Taking silicon into account (a≠0) brings the silanized substrates better in line with the non-silanized substrates and the correlation improves as *a* increases, indicating that silane is polymerized and not just grafted. Depending on the value of *a*, the *y* : *x* ratio varies from 1.1 (*a* = 0) to 0.8 (*a* = 2).

These observations indicate that the stoichiometry of the Cr and Fe oxyhydroxide at the surface is close to (Fe,Cr)OOH. Many studies have reported a stratification in the passive film and the presence of Fe and Cr oxyhydroxides in the outermost layer (Le Bozec et al., 2001). In our case, no significant effect of the photoelectron collection angle appears in the *OHinorg/O529.7* ratio (Figure 8), whatever the value selected for *a* to evaluate *OHinorg*. However it must be kept in mind that the surface roughness revealed by Figure 1, which is of the order of the inelastic mean free path of photoelectrons in oxides, may mask the effect of a stratification.

Silanization with APTES for Controlling the Interactions

0.0 0.5 1.0 1.5 2.0 2.5

at θ = 0°

**4.2 Chemical speciation of the organic adlayer** 

oxidase (,). The dashed lines represent a *y/x* ratio of 1:1.

0.0

component attribution.

0.5

1.0

at θ = 60°

1.5

2.0

2.5

Between Stainless Steel and Biocomponents: Reality vs Expectation 115

*a* **= 0** *a* **= 2**

0.0

0.0 0.5 1.0 1.5 2.0 2.5

at θ = 0°

0.5

1.0

at

Fig. 8. Plots of OHinorg / O529.7 ratio measured by XPS at θ = 60° *vs* θ = 0° on native (open symbols) or silanized stainless steel (closed symbols), as such (¡,) or further treated with coupling agent BS (S,U), glucose oxidase (z,{) or coupling agent followed by glucose

Table 2 reveals an increase of *Ntot* concentration as a result of surface treatments, and an increase of *Siorg* concentration for samples prepared with APTES treatment. However, the concentration of carbon is high and remains almost unchanged, suggesting that organic contaminants, mainly hydrocarbon-like compounds, are always dominating in the organic adlayer. If nitrogen was exclusively due to amide functions (N-C=O) as in the peptide link of proteins, and if the C 1s component at 287.8 eV was exclusivley due to amide, a 1:1 correlation would be found between the concentrations of *C287.8* and *Ntot*. This is indeed observed (Figure 9a) for the set of samples involving the silanized substrate. As nitrogen is partly in the form of silane, relevant alternatives for the abscissa scale may be the concentration of *N400* or the difference between the concentrations of *Ntot* and *Siorg*. If polysaccharides were present with protein, the *C287.8* concentration should be corrected by subtracting the contribution of acetal and thus replaced by *[C287.8 − (C286.3 − N400)/5]* (Ahimou et al., 2007, Landoulsi et al., 2008a) or [*C287.8 − (C286.3 − Ntot + Siorg)/5].* A comparison between different plots in Figure 9 shows that the dots representative of samples prepared with nonsilanized substrate remain clustered. The shift of the cluster along the ordinate scale according to the plot indicates that *C287.8* concentration is higher than what can be attributed to amide. On the other hand, the samples prepared with silanized substrate preserve a unit slope whatever the plot, with much higher values of the coordinates for samples exposed to the enzyme, with or without the linker. This reveals an excellent agreement between the increases of concentrations of nitrogen and of carbon attributed to peptidic links (N−C=O), which result from the Gox treatment. It also validates the C 1s peak decomposition and

The meaning of the surface composition appears more clearly if it is summarized in terms of concentration of model molecular compounds. This approach was already used for microbial surfaces (Dufrêne & Rouxhet, 1996; Tesson et al., 2009), for food products (Rouxhet et al., 2008) and for stainless steel aged in different conditions (Landoulsi et al.,

θ = 60°

1.5

2.0

2.5

Fig. 6. Possible products of APTES reaction. "M" designates metal elements of stainless steel.

Fig. 7. Relations between molar concentrations ratioed to the sum of organic elements (Σorg) measured by XPS at θ = 0° (data from Table 2) on native (open symbols) or silanized stainless steel (closed symbols), as such (¡,) or further treated with coupling agent BS (S,U), glucose oxidase (z,{) or coupling agent followed by glucose oxidase (,).

O − M …

O − Si … O − M … *a* **= 0.5**

O − Si …

O − Si … O − Si … *a* **= 1.5**

O − Si …

O − Si … *a* **= 2**

OH

0.0 0.1 0.2 0.3 0.4

*R2* = 0.53

*R2* = 0.80

0.0 0.1 0.2 0.3 0.4

O − M … *a* **= 0** H2N − (CH2)3 − Si −

Fig. 6. Possible products of APTES reaction. "M" designates metal elements of stainless steel.

0.0

0.4

0.0

(Fe ox + Cr ox) / Σorg (Fe ox + Cr ox) / Σorg

Fig. 7. Relations between molar concentrations ratioed to the sum of organic elements (Σorg)

measured by XPS at θ = 0° (data from Table 2) on native (open symbols) or silanized stainless steel (closed symbols), as such (¡,) or further treated with coupling agent BS (S,U), glucose oxidase (z,{) or coupling agent followed by glucose oxidase (,).

0.1

0.2

0.3

0.1

0.2

0.3

0.4

H2N − (CH2)3 − Si −

H2N − (CH2)3 − Si −

*a* **= 1**

*a* **= 2**

H2N − (CH2)3 − Si −

H2N − (CH2)3 − Si −

H2N − (CH2)3 − Si −

*a* **= 0**

*a* **= 1.5**

(OHinorg – 2Ni

0.0

0.1

0.2

0.3

0.4

(OHinorg – 2Ni

0.0

0.1

0.2

0.3

0.4

ox)/Σorg

ox)/Σorg O − M …

O − M …

O − M …

O − Si … O − Si … *a* **= 1**

O − M …

0.0 0.1 0.2 0.3 0.4

*R2* = 0.02

*R2* = 0.68

0.0 0.1 0.2 0.3 0.4

O − Si … *a* **= 1.5**

OH

Fig. 8. Plots of OHinorg / O529.7 ratio measured by XPS at θ = 60° *vs* θ = 0° on native (open symbols) or silanized stainless steel (closed symbols), as such (¡,) or further treated with coupling agent BS (S,U), glucose oxidase (z,{) or coupling agent followed by glucose oxidase (,). The dashed lines represent a *y/x* ratio of 1:1.

#### **4.2 Chemical speciation of the organic adlayer**

Table 2 reveals an increase of *Ntot* concentration as a result of surface treatments, and an increase of *Siorg* concentration for samples prepared with APTES treatment. However, the concentration of carbon is high and remains almost unchanged, suggesting that organic contaminants, mainly hydrocarbon-like compounds, are always dominating in the organic adlayer. If nitrogen was exclusively due to amide functions (N-C=O) as in the peptide link of proteins, and if the C 1s component at 287.8 eV was exclusivley due to amide, a 1:1 correlation would be found between the concentrations of *C287.8* and *Ntot*. This is indeed observed (Figure 9a) for the set of samples involving the silanized substrate. As nitrogen is partly in the form of silane, relevant alternatives for the abscissa scale may be the concentration of *N400* or the difference between the concentrations of *Ntot* and *Siorg*. If polysaccharides were present with protein, the *C287.8* concentration should be corrected by subtracting the contribution of acetal and thus replaced by *[C287.8 − (C286.3 − N400)/5]* (Ahimou et al., 2007, Landoulsi et al., 2008a) or [*C287.8 − (C286.3 − Ntot + Siorg)/5].* A comparison between different plots in Figure 9 shows that the dots representative of samples prepared with nonsilanized substrate remain clustered. The shift of the cluster along the ordinate scale according to the plot indicates that *C287.8* concentration is higher than what can be attributed to amide. On the other hand, the samples prepared with silanized substrate preserve a unit slope whatever the plot, with much higher values of the coordinates for samples exposed to the enzyme, with or without the linker. This reveals an excellent agreement between the increases of concentrations of nitrogen and of carbon attributed to peptidic links (N−C=O), which result from the Gox treatment. It also validates the C 1s peak decomposition and component attribution.

The meaning of the surface composition appears more clearly if it is summarized in terms of concentration of model molecular compounds. This approach was already used for microbial surfaces (Dufrêne & Rouxhet, 1996; Tesson et al., 2009), for food products (Rouxhet et al., 2008) and for stainless steel aged in different conditions (Landoulsi et al.,

Silanization with APTES for Controlling the Interactions

concentration specific to each compound (Table 3).

and concentration of carbon) and their densities.

photoelectron collection angle *θ* = 0° and 60°.

available on the ExPASy molecular biology server (http://us.expasy.org).

Between Stainless Steel and Biocomponents: Reality vs Expectation 117

The experimental concentration ratios (*CX/Ctot*) can be converted into weight percentages of model compounds (Genet et al., 2008, Rouxhet & Genet, 2011), using the carbon

**Proteins** 0.312 \* 0.273 \* 43.5 1.4 **Silane** 0.5 0.333 27.3 0.9 **Polysaccharides** 0.833 - 37.0 1.5 **Hydrocarbons** - - 71.4 0.9

\* Data computed from the amino acid sequence of glucose oxidase on the basis of ProtParam tool

Table 3. Chemical composition of organic compounds (molar concentration ratio of elements

Results, presented in Table 4, show the often dominating presence of hydrocarbon-like compounds and polysaccharides, for all samples. Both compounds are due to adventitious contamination which may originate from adsorption from air, as always observed for high surface energy solids (Caillou et al., 2008; Mantel et al., 1995), but also from aqueous media (Landoulsi et al., 2008a). They may be taken as a global way to reflect the amount of compounds which contain hydrocarbon chains and oxygen, such as esters, but excluding proteins and silane. The concentration of silane deduced for non-silanized samples is non negligible but highly variable, and may also be attributed to contamination. A drastic

increase of silane concentration is observed for all silanized samples (Table 4).

**nat** 2.9 5.8 29.5 61.7 3.6 2.2 **+BS** (8.5) 6.1 28.3 57.1 3.2 2.1 **+Gox** 16.3 0.0 24.7 59.0 3.2 2.3 **+BS+Gox** (7.3) 12.3 35.5 44.9 3.3 2.2

**sil** 0.0 29.0 22.2 48.8 3.8 2.7 **sil+BS** (10.8) 25.5 25.2 38.5 3.9 2.8 **sil+Gox** 35.2 20.0 24.3 20.5 3.9 2.9 **sil+BS+Gox** 24.4 22.9 22.5 30.2 4.1 2.8

\* The data between brackets are protein equivalent of BS products and have no physical meaning. Table 4. Chemical composition (weight %) of the organic adlayer present on stainless steel samples, as deduced from XPS data and expressed in terms of classes of molecular compounds\*. Thickness of the organic adlayer deduced from measurements at

The Gox treatment leads to a marked increase of the protein concentration on nat sample and a much stronger increase on sil sample. This may be attributed to physical adsorption

**O/C N/C Carbon concent.** 

**Pr Sl PS HC** at θ = 0° at θ = 60°

**(wt. %) Thickness (nm)**

**(mmol.g-1)** 

**Density (g.cm-3)** 

2008b), considering proteins (Pr), polysaccharides (PS), and hydrocarbon-like compounds (HC) which represent mainly lipids in biological systems. The novelty here is to take silane (Sl) into account in addition to the three classes of biochemical compounds; the chemical composition and density considered for these model compounds are listed in Table 3. Accordingly, the proportion of carbon atoms due to each model compound X (*CX/Ctot*) can be computed by solving the following system of equations:

$$(\text{N}\_{\text{tot}} - \text{Si}\_{\text{org}}) \% \text{C}\_{\text{tot}} = 0.273 \times (\text{C}\_{\text{Py}} \% \text{C}\_{\text{tot}}) \tag{4}$$

$$\rm{O}\_{org}/\rm{C}\_{tot} = 0.312 \times (\rm{C}\_{P\eta}/\rm{C}\_{tot}) + 0.833 \times (\rm{C}\_{P\circ}/\rm{C}\_{tot}) \tag{5}$$

$$\text{Si}\_{org} \text{/} \text{C}\_{\text{tot}} = 0.333 \times \text{(C}\_{\text{S}} \text{/} \text{C}\_{\text{tot}}\text{)}\tag{6}$$

$$1 = \left(\mathbf{C}\_{P\eta} / \mathbf{C}\_{\text{tot}}\right) + \left(\mathbf{C}\_{P\xi} / \mathbf{C}\_{\text{tot}}\right) + \left(\mathbf{C}\_{S\xi} / \mathbf{C}\_{\text{tot}}\right) + \left(\mathbf{C}\_{\text{HC}} / \mathbf{C}\_{\text{tot}}\right) \tag{7}$$

Fig. 9. Relations between molar concentrations ratioed to the sum of organic elements (Σorg) measured by XPS at θ = 0° (data from Table 2) on native (open symbols) or silanized stainless steel (closed symbols), as such (¡,) or further treated with coupling agent BS (S,U), glucose oxidase (z,{) or coupling agent followed by glucose oxidase (,).The dashed lines represent a *y/x* ratio of 1:1.

2008b), considering proteins (Pr), polysaccharides (PS), and hydrocarbon-like compounds (HC) which represent mainly lipids in biological systems. The novelty here is to take silane (Sl) into account in addition to the three classes of biochemical compounds; the chemical composition and density considered for these model compounds are listed in Table 3. Accordingly, the proportion of carbon atoms due to each model compound X (*CX/Ctot*) can

×

×

*1 = (CPr/ Ctot) + (CPS/ Ctot) + (CSl/ Ctot) + (CHC/ Ctot)* (7)

 *(CPr/ Ctot) + 0.833* 

×

0.00 0.05 0.10 0.15

Ntot / Σorg

0.00

[C287.8 – (C286.3 – Ntot +

Fig. 9. Relations between molar concentrations ratioed to the sum of organic elements (Σorg)

measured by XPS at θ = 0° (data from Table 2) on native (open symbols) or silanized stainless steel (closed symbols), as such (¡,) or further treated with coupling agent BS (S,U), glucose oxidase (z,{) or coupling agent followed by glucose oxidase (,).The

Siorg)/5]/ Σorg

0.00 0.05 0.10 0.15 (Ntot – Siorg) / Σorg

0.05

0.10

0.15

**c**

 *(CPr/Ctot)* (4)

 *(CSl/ Ctot)* (6)

 *(CPS/ Ctot)* (5)

*(Ntot – Siorg)/Ctot = 0.273* 

×

*Siorg/ Ctot = 0.333* 

be computed by solving the following system of equations:

*Oorg/ Ctot = 0.312* 

0.00

0.05

C287.8 / Σorg

0.00 0.05 0.10 0.15 N400 / Σorg

dashed lines represent a *y/x* ratio of 1:1.

0.00

[C287.8 – (C286.3 – Ntot +

Siorg)/5]/ Σorg

0.05

0.10

0.15

**b**

0.10

0.15

**a**

The experimental concentration ratios (*CX/Ctot*) can be converted into weight percentages of model compounds (Genet et al., 2008, Rouxhet & Genet, 2011), using the carbon concentration specific to each compound (Table 3).


\* Data computed from the amino acid sequence of glucose oxidase on the basis of ProtParam tool available on the ExPASy molecular biology server (http://us.expasy.org).

Table 3. Chemical composition of organic compounds (molar concentration ratio of elements and concentration of carbon) and their densities.

Results, presented in Table 4, show the often dominating presence of hydrocarbon-like compounds and polysaccharides, for all samples. Both compounds are due to adventitious contamination which may originate from adsorption from air, as always observed for high surface energy solids (Caillou et al., 2008; Mantel et al., 1995), but also from aqueous media (Landoulsi et al., 2008a). They may be taken as a global way to reflect the amount of compounds which contain hydrocarbon chains and oxygen, such as esters, but excluding proteins and silane. The concentration of silane deduced for non-silanized samples is non negligible but highly variable, and may also be attributed to contamination. A drastic increase of silane concentration is observed for all silanized samples (Table 4).


\* The data between brackets are protein equivalent of BS products and have no physical meaning.

Table 4. Chemical composition (weight %) of the organic adlayer present on stainless steel samples, as deduced from XPS data and expressed in terms of classes of molecular compounds\*. Thickness of the organic adlayer deduced from measurements at photoelectron collection angle *θ* = 0° and 60°.

The Gox treatment leads to a marked increase of the protein concentration on nat sample and a much stronger increase on sil sample. This may be attributed to physical adsorption

Silanization with APTES for Controlling the Interactions

increased rate of transfer (Caillou et al., 2008).

of or a substitution by new compounds in the adlayer.

[ ] [ ]

**4.4. Thickness of the organic adlayer** 

following equation:

inorganic oxygen close to (*Otot*

above. For Ox

Between Stainless Steel and Biocomponents: Reality vs Expectation 119

It must be emphasized that outgassing a material of high surface energy is not suitable to prevent adventitious contamination of the surface. Materials cleaned with UV-ozone treatment and showing a water contact angle below 5° reached an appreciable contact angle (20° for silica, 40° for stainless steel and gold) after a stay of 5 minutes in the vacuum of the spectrometer chamber. The high rate of contamination may be due to evacuation itself, owing to the increased proportion of organic compounds in the residual gas and to their

No variation of the relative concentrations of organic compounds was revealed by angleresolved XPS analysis (Figure 4) but an effect of stratification may be masked by the surface roughness. The evaluation of the adlayer thickness may clarify whether the increase of the silane or protein concentration resulting from the respective treatments reflects an addition

If the surface is considered as atomically smooth and the organic layer is continuous with a constant thickness, the apparent concentration ratio [*C*]/[*Cr*] may be computed using the

> C C Org Cr C C

<sup>⎡</sup> ⎛ ⎞ <sup>−</sup> <sup>⎤</sup> <sup>⎢</sup> <sup>−</sup> ⎜ ⎟⎥ ⎢⎣ ⎝ ⎠⎥⎦ <sup>=</sup> ⎛ ⎞ <sup>−</sup> ⎜ ⎟ ⎝ ⎠

Cr Cr Org

Cr

 *Oorg*), owing to the low concentration of silane. The

(8)

Cr λ were 3.79 and 3.04 nm in

Cr C ; between

Org Org

Cr i <sup>σ</sup> <sup>t</sup> <sup>λ</sup> C exp <sup>λ</sup> cos<sup>θ</sup>

<sup>t</sup> <sup>λ</sup> C 1 exp <sup>C</sup> <sup>i</sup> <sup>σ</sup> <sup>λ</sup> cos<sup>θ</sup>

iC and iCr are the relative sensitivity factors of C and Cr, respectively, provided by the spectrometer manufacturer. The photoionization cross sections σ are 1 for C 1s and 11.7 for Cr 2p (Scofield, 1976). The superscripts Org and Ox designate the organic adlayer and the

0.015 and 0.020 mol.cm-3, depending on the sample), was determined on the basis of the above discussion indicating that the oxide layer is constituted with FeOOH, CrOOH, Ni(OH)2 and MoO3 (*section 4.1*). Note that this is in agreement with a concentration of

concentration of carbon in the organic adlayer was determined on the basis of the surface composition modeled as detailed above, using the densities given in Table 3. The density of silane, was taken as the average between the densities of 3-aminopropyl(trimethylsilane) (0.8 g.cm-3) and 3-aminopropyl(trimethoxysilane) (1.0 g.cm-3). The electron inelastic mean free paths (IMFP) were calculated using the Quases program (http://www.quases.com ) and the TPP2M formula (Tanuma et al., 1997), considering the matrix composition deduced

respectively (energy gap 2.3 and 1.6 eV, respectively). Considering an energy gap of 6 eV for

hydrocarbon-like compounds [(CH2)n], 3.59 nm and 2.89 nm in protein, 3.67 and 2.00 nm in

The thickness of the organic adlayer deduced for photoelectron collection angles *θ* = 0° and 60° is given in Table 4. The difference between the values computed for the two

polysaccharides [(C6(H2O)5)n], and 3.68 and 2.95 nm in silane [H2N(CH2)3SiO1.5].

Cr λ , values of 2.04 and 1.99 nm were computed for FeOOH and CrOOH,

<sup>C</sup> λ and Org

C Cr Ox Ox

passive oxide layer, respectively. The concentration of Cr in the oxide layer ( Org

<sup>−</sup>

the organic compounds, respective values of Org

and to the fact that protonated amine of silane favors adsorption by electrostatic attraction (Jasienak et al., 2009). The pH of the Gox solution (6.8) is indeed higher than the isoelectric point of glucose oxidase (4.9).

In the above computation, the concentration of the BS coupling agent could not be evaluated. The reaction of BS with NH2 transforms an amine function into amide. If only one end of BS reacts with silane, the *N400* concentration should be doubled, which is consistent with the increase found in Table 2. However, converting it into protein-equivalent gives a number with no physical meaning. When the second end of the coupling agent reacts with the protein, no additional nitrogen is incorporated, the evaluation of the protein concentration is correct but the suberate (CH2)6 chain is counted in the HC concentration. This has no important impact owing to the high concentration of hydrocarbon-like compounds. Despite the limitations regarding the accuracy of the data in Table 4, it is clear that prior silanization increases markedly the concentration of glucose oxidase (sil+Gox and sil+BS+Gox compared to nat+Gox); however the treatment with the coupling agent does not increase the amount of immobilized enzyme (sil+BS+Gox compared to sil+Gox).

#### **4.3 State of stainless steel surface**

It appears in Table 4 that the SS surface prior to and after silanization or enzyme immobilization is bearing a high amount of organic contaminants. It may be argued that this is due to improper cleaning protocols, inappropriate sample manipulation or contamination in the XPS spectrometer. Actually, a clean stainless steel surface is getting quickly contaminated in contact with the surrounding atmosphere, as revealed by water contact angles, which can be measured quickly in the same environment. A nat sample showed a water contact angle of 44° which increased to about 60° within a delay of a few hours (Figure 10). When a nat sample was further treated with UV-ozone to oxidize organic compounds, the water contact angle was lowered down to 12°. However it increased rapidly in contact with the surrounding atmosphere (Figure 10) to reach values above 40° in a few hours. Similar results were obtained with 304L stainless steel. Wet cleaning essentially standardizes the surface contamination; further cleaning leaves a material with a high surface energy, which adsorbs quickly significant amounts of contaminants (Caillou et al., 2008).

Fig. 10. Contact angle measurements as a function of incubation time in ambient atmosphere performed on (z) native stainless steel and ({) after UV-ozone treatment.

It must be emphasized that outgassing a material of high surface energy is not suitable to prevent adventitious contamination of the surface. Materials cleaned with UV-ozone treatment and showing a water contact angle below 5° reached an appreciable contact angle (20° for silica, 40° for stainless steel and gold) after a stay of 5 minutes in the vacuum of the spectrometer chamber. The high rate of contamination may be due to evacuation itself, owing to the increased proportion of organic compounds in the residual gas and to their increased rate of transfer (Caillou et al., 2008).

#### **4.4. Thickness of the organic adlayer**

118 Biomaterials – Physics and Chemistry

and to the fact that protonated amine of silane favors adsorption by electrostatic attraction (Jasienak et al., 2009). The pH of the Gox solution (6.8) is indeed higher than the isoelectric

In the above computation, the concentration of the BS coupling agent could not be evaluated. The reaction of BS with NH2 transforms an amine function into amide. If only one end of BS reacts with silane, the *N400* concentration should be doubled, which is consistent with the increase found in Table 2. However, converting it into protein-equivalent gives a number with no physical meaning. When the second end of the coupling agent reacts with the protein, no additional nitrogen is incorporated, the evaluation of the protein concentration is correct but the suberate (CH2)6 chain is counted in the HC concentration. This has no important impact owing to the high concentration of hydrocarbon-like compounds. Despite the limitations regarding the accuracy of the data in Table 4, it is clear that prior silanization increases markedly the concentration of glucose oxidase (sil+Gox and sil+BS+Gox compared to nat+Gox); however the treatment with the coupling agent does not

increase the amount of immobilized enzyme (sil+BS+Gox compared to sil+Gox).

It appears in Table 4 that the SS surface prior to and after silanization or enzyme immobilization is bearing a high amount of organic contaminants. It may be argued that this is due to improper cleaning protocols, inappropriate sample manipulation or contamination in the XPS spectrometer. Actually, a clean stainless steel surface is getting quickly contaminated in contact with the surrounding atmosphere, as revealed by water contact angles, which can be measured quickly in the same environment. A nat sample showed a water contact angle of 44° which increased to about 60° within a delay of a few hours (Figure 10). When a nat sample was further treated with UV-ozone to oxidize organic compounds, the water contact angle was lowered down to 12°. However it increased rapidly in contact with the surrounding atmosphere (Figure 10) to reach values above 40° in a few hours. Similar results were obtained with 304L stainless steel. Wet cleaning essentially standardizes the surface contamination; further cleaning leaves a material with a high surface energy, which adsorbs quickly significant amounts of contaminants (Caillou et al.,

Fig. 10. Contact angle measurements as a function of incubation time in ambient atmosphere

0 4 8 12 Square root of time (h1/2)

performed on (z) native stainless steel and ({) after UV-ozone treatment.

0

20

40

Contact angle (°)

60

80

point of glucose oxidase (4.9).

**4.3 State of stainless steel surface** 

2008).

No variation of the relative concentrations of organic compounds was revealed by angleresolved XPS analysis (Figure 4) but an effect of stratification may be masked by the surface roughness. The evaluation of the adlayer thickness may clarify whether the increase of the silane or protein concentration resulting from the respective treatments reflects an addition of or a substitution by new compounds in the adlayer.

If the surface is considered as atomically smooth and the organic layer is continuous with a constant thickness, the apparent concentration ratio [*C*]/[*Cr*] may be computed using the following equation:

$$\frac{\mathbf{[C]}}{\mathbf{[Cr]}} = \frac{\mathbf{i\_{cr}}}{\mathbf{i\_{c}}} \frac{\sigma\_{\rm c}}{\sigma\_{\rm c}} \frac{\sigma\_{\rm c}}{\sigma\_{\rm c}} \frac{\left[1 - \exp\left(\frac{-\mathbf{t}}{\lambda\_{\rm c}^{\alpha\_{\rm v}} \mathrm{cos}\theta}\right)\right]}{\lambda\_{\rm c}^{\alpha\_{\rm c}} \mathbf{C}\_{\rm cr}^{\alpha\_{\rm cr}} \exp\left(\frac{-\mathbf{t}}{\lambda\_{\rm c}^{\alpha\_{\rm v}} \mathrm{cos}\theta}\right)}\tag{8}$$

iC and iCr are the relative sensitivity factors of C and Cr, respectively, provided by the spectrometer manufacturer. The photoionization cross sections σ are 1 for C 1s and 11.7 for Cr 2p (Scofield, 1976). The superscripts Org and Ox designate the organic adlayer and the passive oxide layer, respectively. The concentration of Cr in the oxide layer ( Org Cr C ; between 0.015 and 0.020 mol.cm-3, depending on the sample), was determined on the basis of the above discussion indicating that the oxide layer is constituted with FeOOH, CrOOH, Ni(OH)2 and MoO3 (*section 4.1*). Note that this is in agreement with a concentration of inorganic oxygen close to (*Otot* <sup>−</sup> *Oorg*), owing to the low concentration of silane. The concentration of carbon in the organic adlayer was determined on the basis of the surface composition modeled as detailed above, using the densities given in Table 3. The density of silane, was taken as the average between the densities of 3-aminopropyl(trimethylsilane) (0.8 g.cm-3) and 3-aminopropyl(trimethoxysilane) (1.0 g.cm-3). The electron inelastic mean free paths (IMFP) were calculated using the Quases program (http://www.quases.com ) and the TPP2M formula (Tanuma et al., 1997), considering the matrix composition deduced above. For Ox Cr λ , values of 2.04 and 1.99 nm were computed for FeOOH and CrOOH, respectively (energy gap 2.3 and 1.6 eV, respectively). Considering an energy gap of 6 eV for the organic compounds, respective values of Org <sup>C</sup> λ and Org Cr λ were 3.79 and 3.04 nm in hydrocarbon-like compounds [(CH2)n], 3.59 nm and 2.89 nm in protein, 3.67 and 2.00 nm in polysaccharides [(C6(H2O)5)n], and 3.68 and 2.95 nm in silane [H2N(CH2)3SiO1.5]. The thickness of the organic adlayer deduced for photoelectron collection angles *θ* = 0° and

60° is given in Table 4. The difference between the values computed for the two

Silanization with APTES for Controlling the Interactions

Vol. 64, No. 6, pp. 827-837.

pp. 116-122.

No. 6, pp. 1591-1600.

Vol. 27, No. 21, pp. 4158-4165.

Vol. 23, No. 19, pp. 4019-4027.

*Science*, Vol. 324, No. 1-2, pp. 118-126.

*of Microbiology*, Vol*.* 42, pp. 548-556.

*Spectroscopy* (Chichester, UK, Wiley).

**7. References** 

Between Stainless Steel and Biocomponents: Reality vs Expectation 121

Ahimou, F., Boonaert, C.J.P., Adriaensen, Y., Jacques, P., Thonart, P., Paquot, M. & Rouxhet,

Asenath Smith, E. & Chen, W. (2008) How To Prevent the Loss of Surface Functionality Derived from Aminosilanes, *Langmuir*,Vol. 24, No. 21, pp. 12405-12409. Balasundaram, G., Sato, M. & Webster, T.J. (2006) Using hydroxyapatite nanoparticles and

Bartouilh de Taillac, L., Porté-Durrieu, M.C., Labrugère, C., Bareille, R., Amédée, J. &

Boucher, C., Liberelle, B.t., Jolicoeur, M., Durocher, Y. & De Crescenzo, G. (2009) Epidermal

Caillou, S., Gerin, P.A., Nonckreman, C.J., Fleith, S., Dupont-Gillain, C.C., Landoulsi, J.,

Charbonneau, C., Liberelle, B., Hébert, M.-J., De Crescenzo, G. & Lerouge, S. (2011)

Cho, Y. & Ivanisevic, A. (2004) SiOx Surfaces with lithographic features composed of a TAT peptide, *The Journal of Physical Chemistry B*, Vol. 108, No. 39, pp. 15223-15228. Chuang, Y.-J., Huang, J.-W., Makamba, H., Tsai, M.-L., Li, C.-W. & Chen, S.-H. (2006)

Davis, D.H., Giannoulis, C.S., Johnson, R.W. & Desai, T.A. (2002) Immobilization of RGD to

Dekeyser, C.M., Buron, C.C., Mc Evoy, K., Dupont-Gillain, C.C., Marchand-Brynaert, J.,

Doh, J. & Irvine, D.J. (2006) Immunological synapse arrays: Patterned protein surfaces that

Dufrêne, Y.F. & Rouxhet, P.G. (1996) Surface composition, surface properties, and

*National Academy of Sciences*, Vol. 103, No. 15, pp. 5700-5705.

phosphorylation, *Bioconjugate Chemistry*, Vol. 20, No. 8, pp. 1569-1577. Briggs, D. & Seah, M.P. (1990) *Practical Surface Analysis, Auger and X-ray Photoelectron* 

*Journal of Colloid and Interface Science*,Vol. 309, No. 1, pp. 49-55.

with RGD, *Biomaterials*,Vol. 27, No. 14, pp. 2798-2805.

P.G. (2007) XPS analysis of chemical functions at the surface of Bacillus subtilis,

decreased crystallinity to promote osteoblast adhesion similar to functionalizing

Baquey, C. (2004) Grafting of RGD peptides to cellulose to enhance human osteoprogenitor cells adhesion and proliferation, *Composites Science and Technology*,

growth factor tethered through coiled-coil interactions induces cell surface receptor

Pancera, S.M., Genet, M.J. & Rouxhet, P.G. (2008) Enzymes at solid surfaces: Nature of the interfaces and physico-chemical processes, *Electrochimica Acta*, Vol. 54, No. 1,

Stimulation of cell growth and resistance to apoptosis in vascular smooth muscle cells on a chondroitin sulfate/epidermal growth factor coating, *Biomaterials*, Vol. 32,

Electrophoretic mobility shift assay on poly(ethylene glycol)-modified glass microchips for the study of estrogen responsive element binding, *Electrophoresis*,

<1 1 1> silicon surfaces for enhanced cell adhesion and proliferation, *Biomaterials*,

Jonas, A.M. & Rouxhet, P.G. (2008) Oligo(ethylene glycol) monolayers by silanization of silicon wafers: Real nature and stability, *Journal of Colloid and Interface* 

modulate immunological synapse structure formation in T cells, *Proceedings of the* 

adhesiveness of Azospirillum brasilense - variation during growth, *Canadian Journal* 

photoelectron collection angles is due to the approximation of a smooth surface while the analyzed surface is rough at the scale of the inelastic mean free paths. The real organic adlayer thickness should be between the values given in Table 4. Comparison between nat and sil samples suggests that silane just adds up to the contaminants. A 3.0 nm thick adlayer containing 25 wt.% silane corresponds to 4.5 molecules.nm-2. This value is consistent with a monolayer of silane, however the retained silane is mixed with a much larger amount of contaminants. The protein treatment of silanized substrates (compare sil+Gox and sil+BS+Gox to sil) led to a significant decrease of the amount of hydrocarbon-like compounds, while the adlayer thickness did not change appreciably. This suggests that the protein adsorption caused the displacement of part of contamination present on the silanized stainless steel surface in the form of hydrocarbon-like compounds.
