Spectroscopic ellipsometry meets AFM nanolithography: about hydration of bio-inert oligo(ethylene glycol)-terminated self assembled monolayers on gold

For the first time, to our knowledge, spectroscopic ellipsometry (SE) has been combined with state-of-the-art AFM diﬀerential height measurements conducted after shaving nano-lithography of ultrathin, soft-matter films for thickness determination. We investigated self-assembled monolayers of SH– (CH 2 )11–EG n –OH molecules on gold, where EG is ethylene glycol units and n = 3 and 6, a prototypical non-fouling system. We performed SE measurements (245–1200 nm) focusing on the changes induced by the formation of the film (difference spectra). SE measurements, analysed by simple models, confirm the formation of the S–Au interface, transparency of the SAMs and provide a sharp picture of the ability of the EG functionality to protect the surface from unspecific adsorption of proteins. A quantitative assessment of the film thickness by SE was carried out ex situ , thanks to the optical contrast between the film and the ambient, and by AFM in liquid. The cross-check between SE and AFM height measurements combined with the comparison between in-liquid and ex situ SE measurements allowed obtaining non-perturbative information about the vertical density profile of the SAM. The in-liquid SE measurements indicate a refractive index matching between the aqueous medium and the outer part of the SAM, consistent with a disordered configuration of OEG and/or the penetration of water amid the OEG strands. A critical discussion provides a detailed insight into the subtle issues and pitfalls related to the thickness determination of soft-matter films to the monolayer limit.


Introduction
The progress of bioanalytical technologies based on the integration of sensors with micro-and nano-electro-mechanical devices requires the development of label-free and non-invasive tools for the accurate characterization of sensing surface elements.In this context, organic or bio-functionalization (for example, by self assembled monolayers, SAMs) of solid supports plays a pivotal role. 1,2In general, biomolecular adsorption on a surface can be determined by measuring the inherent mass transfer, or, alternatively, the inherent increase of thickness.
Light underlies the most gentle analytical methods of organic and biologic ultrathin films and SAMs.Optical measurements can be performed in transparent liquid media such as the aqueous environment that is associated with biological samples, and in a time-resolved fashion.Ellipsometry, as well as SPR measurements, are commonly exploited for monitoring adsorption/desorption processes, 3,4 eventually in a combined fashion as in total internal reflection ellipsometry. 5,68][9] In contrast to SPR, ellipsometry is in principle applicable to a number of solid supports with adequate reflectivity.
Specifically, optical techniques provide indirect measurements of the thickness and other structural parameters of films.Within the frame of Fresnel optics, for ultra-thin films (film thickness d { l), specifically in the range of transparency, the thickness is increasingly correlated to the refractive index, 10,11 which is known only approximately in general.6][17] In shaving nano-lithography, 18 for instance, exploiting the removal of molecules by the high shear force a Dipartimento di Fisica, Universita `di Genova and CNISM, via Dodecaneso 33, Genova, Italy.E-mail: canepa@fisica.unige.itb Elettra Sincrotrone Trieste S.C.p.A., s.s.14 km 163, 5 in Area Science Park, Basovizza, Trieste, Italy c Department of Medical and Biological Sciences -University of Udine, Udine, Italy exerted by the tip during an areal scan, one can perform differential height measurements between fully-covered and fully-depleted, adjacent areas.Shaving is therefore particularly appealing as it allows examining compact SAMs that are prepared under conditions basically similar to those adopted for ellipsometry measurements; furthermore, such kinds of measurements offer an additional complementarity as to the lateral resolution, at the micro-scale (AFM) vs. macro-scale (SE).
In this work, to our knowledge, for the first time, state-ofthe-art spectroscopic ellipsometry (SE) and AFM nanolithography are combined to obtain detailed information on the thickness and optical properties of antifouling SAMs that are typically utilized to reduce the adsorption background of surfacebased biosensors. 2,19e focus on oligo(ethylene glycol)-terminated alkanethiol (for brevity, T-OEG) SAMs on gold surfaces, whose intriguing non-fouling properties were early emphasized in seminal studies. 7,1][22][23][24][25][26] Such an intrinsic and local disorder is believed to play a role in determining the interaction between the OEG moieties and water molecules, effective in preventing the underlying methylene groups from wetting by water. 7There is also an accumulating consensus on the manifold role of the water-OEG interaction, and in particular of the penetration of water into the SAM and the related hydration of the OEG part in establishing the bio-inertness of these SAMs, [27][28][29][30][31][32][33] which also favours their stability in room ambient.
Furthermore, the relatively poor organization of T-OEG SAMs makes simple isotropic models rather adequate to their optical modelling.
On the practical side of SE measurements, we adopted an approach which is based on the analysis of small changes in the ellipsometric response upon formation of the film.][41][42] Our approach helps defining the sources of systematic uncertainty associated with both SE and differential AFM measurements, an aspect which is not adequately addressed in routine applications of the individual methods.The combination of difference ellipsometry spectra and differential AFM measurements considerably strengthens the reliability of the thickness determination proposed in this work.The cross-check between SE and AFM, and the comparison between wet and dry samples allow obtaining non-perturbative information regarding the vertical density profile of the SAM and the penetration of water amid the OEG chains.Last but not least, improving the accuracy in the determination of the thickness opens the door to obtaining more reliable estimates of the refractive index of the films, a piece of information that is elusive for such ultrathin layers.

Materials
Gold substrates and SAM preparation.AFM measurements were performed on the so-called Ulman ultra-flat gold substrates, 43 prepared following a modified version of the stripping procedure described in detail in previous papers. 44The stripped substrates presented a typical surface roughness of 0.2 nm as measured by AFM on 5 Â 5 mm 2 areas.The samples were then soaked in a 300 mM solution of T-OEG in ethanol for about 24 hours, to make it ready for the nanoshaving procedure.
For optical measurements we used commercial (Arrandee) gold films deposited on glass with a Cr primer. 34All the substrates were treated by chemical etching (60% MilliQ water, 30% ammonia solution and 30% hydrogen peroxide at 100 1C for 10 min), rinsed in MilliQ water and fluxed with N 2 and characterized by ex situ SE.The optical properties of substrates were consistent with previous determinations. 34,45A subset of substrates was also characterized by SE in TE buffer.After incubation in the T-OEG solution in ethanol (4 1C, 18 h) the samples were extracted and rinsed with ethanol, dried under N 2 flow and immediately characterized by ex situ SE.For subsets of samples, ethanol was substituted by TE buffer or MilliQ water for in situ SE characterization and test of the resistance properties.A representative set of T-OEG SAMs were exposed to a 1 mM YCC-Milli-Q solution for 2 h at 4 1C.The molecular solution was then gradually substituted with pure Milli-Q water and the sample was characterized by in situ and ex situ SE, after being dried under N 2 flow.

Methods
Spectroscopic ellipsometry.Principles of SE are described at length in books. 10,11,46Recent reviews were oriented to nanoscience applications. 47,48The so-called standard ellipsometry investigates the coefficient r = r ˜p/r ˜s = tan C exp(iD), where r ˜p and r ˜s are the complex Fresnel reflection coefficients for p-and s-polarization, respectively.Data are analysed by comparison with simulations based on an optical and morphological model of the layered system.The value of convenient adjustable parameters is derived by fitting experimental spectra and simulations, exploiting fit quality estimators such as the mean squared error, MSE, function. 49,50he measurements were obtained using a rotating compensator spectroscopic ellipsometer (M-2000, J.A. Woollam Co. Inc.) equipped This journal is © the Owner Societies 2015 with a 75 W Xe lamp.Ex situ spectra of dry samples have been collected at 65-701 of incidence and are reported in the 245-1200 nm interval where difference spectra showed a satisfactory signal-to-noise ratio.In-liquid measurements were performed in a home-made Teflon cell described in ref. 51 (angle of incidence 651) and data are reported in the range of high transparency of the medium.The spot size on the sample was of the order of a few mm 2 .][39] dC and dD will denote the difference between the data obtained for SAM-covered and pristine substrates (dC = C film+Au À C Au and dD = D film+Au À D Au ).Difference spectra are presented after averaging over not less than five samples and typically over four different zones of each sample.
Atomic force microscopy: nanoshaving and height measurements.AFM experiments were carried out using XE100 (Park Instruments) and MFP-3D (Asylum Research) instruments in custom-designed liquid cells.For nanoshaving, sketched in the upper part of Fig. 1, we operated in contact mode in pure ethanol with commercial AFM tips (NSC19 Mikromasch, k = 0.6 N m À1 ).During the shaving process a high load (100 nN) was applied to a small (0.5 Â 0.5-1 Â 1 mm 2 ) scanning area.The lateral movement of the tip during the scanning at high force allows the local displacement of the molecules linked to the surface (Fig. 1A and B).The molecules will then diffuse into the solution, leaving the gold surface exposed in the shaved area (Fig. 1C).
The depth of the shaved areas (h SAM ) provided an accurate estimate of the film thickness in a liquid.h SAM was measured from both line profiles and height histogram analyses as the difference between the top of the molecular layer and the bottom of the shaved area (Fig. 1C).Regarding line profiles, we took the average value of the height of the SAM surface and subtracted the average value of the height of the gold surface, being careful to avoid the edges of the shaved area as suggested by shading in Fig. 1F and G.The larger surface roughness of the nanoshaved areas with respect to the surrounding SAM carpet is due to the nanoshaving process that can slightly damage the surface and/or leave the residual molecules on the depleted surfaces.The height histograms were calculated for every image independently, by analysing the 256 Â 256 or 512 Â 512 z-coordinate values in the topographic maps.Height histograms have been aligned setting arbitrarily the zero at the SAM surface.The two methods provided comparable results.We will focus in Section 3.1 on the height histogram analysis.
Topographic images were acquired in soft contact mode in saline buffer (10 mM Tris-HCl, 1 mM EDTA, 1 M NaCl, pH = 7.2) using softer commercial cantilevers (CSC38 Mikromasch, k = 0.03 N m À1 ) at the minimum force detectable (0.1 nN).This minimum force allows a soft contact on the surface and maximally reduces the effect of the applied load of the tip on the molecules.Piezoscanners and cantilevers were carefully calibrated before the nanoshaving sessions, and the non-linear effects are corrected through a feedback loop system (internal to the AFM setups) in both the X-Y plane and the Z-direction.Images and data were analyzed using XEI (Park Instruments), Igor Pro (Wavemetrics Inc.) and Gwyddion (Gwyddion.net)softwares.The measurements on dried samples (images not shown) were affected by stray adsorption in the depleted regions, even after careful washing in pure ethanol.However, while the depleted areas were contaminated, the SAM surfaces preserved the roughness and conformation measured in a liquid.

Atomic force microscopy
Typical AFM images and the corresponding height profiles of shaved areas (in liquid) in T-OEG3 and T-OEG6 SAMs are reported in Fig. 1(D-F) and (E-G), respectively.In panel H, reporting the height histograms relative to the specific images of panels D (T-OEG3) and E (T-OEG6), we can observe two peaks, the one (centered at 0 nm) corresponding to the surface of the SAM (brighter areas in panels D and E), and the broader one corresponding to the shaved regions (darker areas).We performed a fit of the height distribution with two Gaussian curves, whose inter-distance, h SAM , can be considered the AFM thickness of the layer.The errors on h SAM have been evaluated taking into account the sum of standard deviation values derived from the fit relative to the SAM surface and the shaved area.In the specific case reported the values of h SAM are 1.6 AE 0.6 nm for T-OEG3 and 2.2 AE 0.4 for T-OEG6.From the statistical analysis of the h SAM values of different shaved areas

Spectroscopic ellipsometry
The ex situ SE data on T-OEG films are shown in Fig. 2. The reported simulations were calculated according to the simplest model (TF|S) in which a transparent film (TF) is sharply stacked onto the substrate (S).The film was modelled through a simplified Cauchy dispersion formula: 10,11,46 Simulations are presented for an indicative range of n TF as a function of the film thickness d TF .The coefficient B was set to 0.01.They reproduce the shape of experimental spectra with the relevant exception of the negative dC values in the high reflectivity region of gold (EÀ0.11 at 651 and l = 750 nm for both SAMs) comparable to those found for other thiolate SAMs. 34,36No intrinsic molecular optical absorption (observed as sharp dips in difference spectra of other SAMs on Au 38,40 ) was discernible.
For ultrathin transparent films of given n TF the proportionality between dD and d TF is used in routine checks of thickness, which usually assume some convenient values of the refractive index and ignore the interface effects. 7,54The thickness of T-OEG3 and T-OEG6 SAMs that can be extrapolated from panels A and C appears slightly larger than the values obtained from AFM.
Simulated dC curves (panels B and D) at low l, before the evident step-like transition, are proportional to the thickness as well.The thickness is much less dependent on the value of n TF and one can easily extrapolate d TF values in agreement with AFM measurements.
The apparent discrepancy between the d TF values indicated by the disjunct analysis of dD and dC data as well as the negative dC values witnesses the limitations of the TF|S model.dC negative values of thiolate films were indeed attributed, after optical reflectivity studies, 55 to the film/substrate interface. 34This interpretation found support in our recent work on selenide SAMs where effects induced by the formation of the Se-Au bond were even stronger than in the S-Au case. 37herefore a more precise evaluation of the film thickness needs the inclusion of interface effects in the optical model. 37ithin the Fresnel approach, a quick way to account for the interface is the addition of a so-called transition layer between the substrate and the SAMs (TF|I|S model).Following a wellknown approach, 56 the dielectric function of the transition layer is approximated through the Bruggeman effective medium approximation (BEMA): where e i and f i are the dielectric functions and volume fractions of the mixing constituents, respectively.The BEMA formula is frequently used to account for the micro-roughness at the vacuum-solid interface. 46Analogously, in our context, the BEMA transition layer could be viewed as a kind of interface roughness.The BEMA model depends on two parameters, i.e. the fraction f SAM and the interface thickness d I .In the simplest approximation f SAM was set to 0.5.In the case of gold substrates the introduction of the BEMA transition layer effectively introduces negative dC values in the NIR, depending crucially on the value of d I . 37The BEMA layer also affects, to a minor extent, the dC UV region, and the dD spectra as well as it is detailed in Fig. 3, where the effect of d TF , d I and f SAM is analysed separately.In panels A and B, the increase of d I values from 0 to 0.4 nm induces the progressive formation of a negative minimum of dC at about 600 nm and a parallel quasi-rigid downward shift of dD related to the increment of the total d TF + d I thickness.In panels C and D, upon increasing d TF from 1.0 to 3.0 nm, one can note the downward shift of dD spectra related to the increase of thickness, and the corresponding increase of dC in the UV limit.The increase of d TF does not significantly affect the negative dC data.The effect of f SAM at fixed d TF (2 nm) and d I (0.2 nm) is then illustrated in panels E and F.An increasing f SAM leads to a broader and broader Only small adjustments of interface parameters were observed; the dispersion parameter B showed a poor variability in the 0.005-0.006range, values which appear fully comparable to determinations on thicker organic transparent films. 57The best reproduction of the negative part of dC required values of d I of about 0.25 for both molecules, with f SAM of E60%, comparable to values found on other thiolate films. 34n a refining step, we restricted n TF between 1.35 and 1.4.The best-fit is presented in Fig. 4. The corresponding values of the adjustable parameters are reported in Table 1.The reproduction of dC and dD data was very accurate, even in the IR region, for both molecular moieties and both angles of incidence.
Considering the fit uncertainty, the mentioned index vs.thickness correlation, and the dispersion of experimental data, we have estimated an overall uncertainty of AE0.2 nm for d TF , for both dry SAMs.
The TF|I|S model assumption of a uniform film could appear oversimplifying in view of the conceivable different morphologies of the lower and upper regions of the T-OEG film. 24However, attempts to refine the model by introducing a grading of n collided with severe problems of mathematical correlation between parameters, without leading to a significant reduction of MSE.Nevertheless, the approach proved useful in the interpretation of in-liquid (TE buffer) SE measurements, which are presented in Fig. 5.
In these measurements the medium/film optical contrast drastically decreased, leading to values of dC and dD definitely   smaller than those found ex situ.Taking into account the increased relative uncertainty, the thickness difference between the two SAMs would appear even questionable.This finding at the first glance appears in striking contrast to both AFM and ex situ SE results.And indeed the simple model used in the case of dry samples is not able to fit the SE data in liquid.In this respect we note that several authors claimed the penetration of water molecules in the film. 21,31,32,58Partial hydration of the film could contribute to render the surface of the SAM less defined, thus smoothing the optical contrast between the brush-like part of the film and the ambient.The modelling of complex SAM/water interphases requires an increasing number of parameters to account for e.g. the conceivable grading of water density.The interplay between the mathematical correlation between the fit parameters 14 and the relatively large, experimental uncertainty makes it hard to achieve a robust, quantitative assessment of the degree of water penetration on the refractive index of the SAM part and, consequently, a robust determination of the thickness.However, an appealing interpretation can be advanced just inspecting the data.Indeed, we find here a situation typical for Immersion Ellipsometry experiments (SIE), which were exploited on semiconductor interfaces. 59,60The index matching between the ambient and the outer part of the film tends to optically remove the latter, leading to an enhancement of the sensitivity to the inner more packed part of the film and to the film/substrate interface as well.
The similarity between T-OEG3 and T-OEG6 results would then reflect a similar organization and density of the lowest-lying part of the film, especially the C11 part, common to the two molecules.Note that the negative dC values above 550 nm are well visible, quantitatively similar to those determined ex situ; thus the in-liquid measurements provide an elegant confirmation of the existence of a transition layer optically distinct from the film and the substrate.
Finally we present ellipsometry measurements aimed to check the resistance of T-OEG SAMs to unspecific adsorption.Representative data, for T-OEG6, are shown in Fig. 6; results obtained for T-OEG3 SAMs were equivalent.There are no significant differences between SE spectra taken after and before exposure to YCC (dC = C YCC+TOEG6/Au À C TOEG6/Au , dD = D YCC+TOEG6/Au À D TOEG6/Au ).Instead, in the case of octadecanethiol (C18) SAM well-defined difference spectra are visible with evident features at about 400 nm indicating the Soret band of YCC. 39Note that for the YCC/C18 case dC values in the NIR are substantially null, as expected for the sharp interface related to unspecific adsorption.These findings can also be compared with previous studies reporting on the adsorption of YCC onto bare Au 39 and silicon oxide substrates, 40 where the YCC-induced signal was definitely significant and neat fingerprints of the optical absorption associated with the heme group were detected.

Concluding remarks
Let us resume the information obtained by just inspecting the difference SE spectra or resorting to the comparison with simulations built on simple models.The measurements confirmed the absence of any intrinsic optical absorption of the T-OEG molecules in the probed wavelength range.Well-defined spectroscopic features in the NIR spectral region, shared by all the thiol/gold systems investigated in our previous studies, specifically provided information on the formation of the S-Au interface.The measurements in liquid showed a limited   sensitivity to the OEG part of the film, indicating a decrease of the refractive index going from the more packed C11 part to the part consisting of flexible OEG chains, eventually penetrated by water.The broadband data provided a sharp picture of the high resistance of T-OEG SAMs to unspecific adsorption of proteins.][63] Regarding the quantitative determination of film thickness the comparison of SE and AFM nanolithography, exploited here for the first time, revealed complementary points of strength.AFM nanolithography displayed the highest potential for in-liquid measurements.Conversely, in-liquid SE measurements proved problematic for a precise determination of the thickness; nevertheless they provided interesting clues about the vertical morphology of the films, which is not accessible to the AFM probe, and is difficult to obtain with other more perturbative methods.Furthermore, the in-liquid SE measurements indicate a refractive index matching between the aqueous medium and the outer part of the SAM, consistent with a disordered configuration of the OEG and/ or the penetration of water amid the OEG strands, factors which are expected to determine the bio-inertness of the film.5][66] Several studies explicitly suggested the trapping of water molecules into the SAM, eventually correlated to the packing density, as a necessary prerequisite for the resistance properties of OEG. 23,29 quantitative assessment of thickness by SE was instead possible on dry samples, thanks to the increased ambient/film optical contrast.This could support previous claims 24 on the fact that after removal from the fluid the water molecules would remain absorbed mainly at the terminal groups of the SAM.The thickness difference between T-OEG6 and T-OEG3 SAMs derived from the T|I|S model applied to SE data on dry samples is in nice agreement with AFM height measurements.Regarding absolute values (Table 1), the values of 1.9 AE 0.2 and 2.6 AE 0.2 for T-OEG3 and T-OEG6 SAMs, respectively, are slightly higher than the AFM-based determinations and appear even higher taking the interface layer thickness (E0.3 nm) into account.Note that conceivable factors may lead to underestimating the absolute thickness such as the imperfect thiol depletion of shaved areas as well as a slight compression of the soft layer under the tip pressure. 67On the other hand, the refractory character of T-OEG SAMs cannot rule out some adventitious adsorption from the room ambient, which can lead to a slight overestimation of thickness in ex situ SE measurements.Thickness values are fully compatible with other determinations based on ellipsometry and more perturbative methods such as XPS. 7,20,54The resulting T-OEG6 film thickness (d TF + d I E 2.9) appears slightly smaller than the estimated length of the molecule (E3-3.3nm), 68 which is consistent with a relative SAM disorder.
The good agreement between AFM and SE regarding the film thickness strengthens the spectroscopic ellipsometry analysis, helping to attenuate the uncertainty related to the thickness/ index correlation.The refractive index of the T-OEG films, in the 1.35-1.40range, could appear somewhat lower than those commonly assumed for dense alkanethiol or even T-OEG SAMs. 7,34,68In this respect it is worth recalling that the T|I|S model provides an average between the indices of more (C11) and less densely packed (OEG) parts of the film.
The results of our work demonstrate the high potential of the method of cross-checking AFM nanolithography and SE in the analysis of ultrathin organic and biologic films, providing a detailed insight into the subtle issues and pitfalls involved in the thickness determination of soft-matter films, especially transparent ones, down to the limit of a monolayer.

Fig. 1 (
Fig. 1 (A-C) The nanoshaving process: applying a high load the AFM tip locally removes the SAM molecules.The process produces depleted areas ready for height measurements.(D and E) AFM micrographs and (F and G) representative height profiles across shaved regions in T-OEG SAMs.The edges of the shaved areas have been excluded from the process of height evaluation as suggested by the shading in panels F and G. (H) Height histograms for nanoshaved areas in T-OEG3 and T-OEG6, with the corresponding Gaussian fits (dotted lines).Histograms have been aligned setting arbitrarily the zero at the SAM surface.

Fig. 2
Fig. 2 Open circles (grey tones): ex situ spectroscopic ellipsometry data (difference spectra: dC = C film+Au À C Au , dD = D film+Au À D Au ) obtained for the T-OEG SAMs on Au at (A and B) 651 and (C and D) 701 angles of incidence.Thin vertical bars show representative uncertainty in different spectral regions.Shaded areas illustrate simulations based on the TF|S model, for three values of the SAM thickness (d TF = 1.0, 2.0 and 3 nm) and provide a graphical representation of the index/thickness correlation.For dD patterns (left) the top and bottom borders of shading correspond to n TF = 1.35 and 1.55, respectively; for dC spectra (right panels) the order is reversed.For a given n TF , simulated dD curves are proportional to d TF .The TF|S model reproduces the shape of both dD and dC spectra with the exception of the negative dC values above 550 nm.

Fig. 3
Fig. 3 Simulations of SE data (difference spectra) based on the TF|I|S model, in which a transition layer (BEMA approximation) is sandwiched between the substrate and the transparent film.All the simulations were obtained for an indicative value of n TF (1.4).(A and B) The effect of the variation of the interface thickness d I from 0 to 0.4 nm, at fixed d TF (2 nm) and f SAM (0.5); note the negative dC values for l 4 550 nm.Instead, the NIR dD values depend on the total thickness d TF + d I .(C and D) The effect of the variation of d TF from 1 to 3 nm, having set d I to 0.2 nm and f SAM to 0.5; note that NIR dC values are not affected by d TF .(E and F) The effect of the variation of the interface parameter f SAM , from 0.35 and 0.65, having set d TF to 2 nm and d I to 0.2 nm.

Fig. 4
Fig. 4 Comparison between SE data (symbols) for T-OEG3 and T-OEG6 films, and best-fit simulations (lines) at 651 (A and B) and 701 (C and D) angles of incidence.Simulations have been calculated according to the TF|I|S model.The fitting procedure was performed on averaged data, the two species of SAMs being considered independently.The corresponding best-fit values of the adjustable parameters are reported inTable 1.

Fig. 5
Fig. 5 Ellipsometry difference spectra (dC = C film+Au À C Au , dD = D film+Au À D Au ) of T-OEG SAMS in TE buffer.Measurements were obtained at 651 angle of incidence and are shown in the range of high transparency of the medium.Thin vertical bars show representative uncertainty in different spectral regions.SAM-induced dC and dD data are definitely smaller (about a factor of two) than for ex situ measurements.This is mainly due to the limited contrast between the refractive index of the medium (water) and the refractive index of the SAMs, especially the OEG part.

Fig. 6
Fig. 6 Ex situ measurements.Dark grey circles: ellipsometry difference spectra (dC = C TOEG6+Au À C Au , dD = D TOEG6+Au À D Au ) obtained for TOEG SAMs (651 angle of incidence; same as Fig. 2).Red curve: ellipsometry difference spectra (dC = C YCC+TOEG6/Au À C TOEG6/Au , dD = D YCC+TOEG6/Au À D TOEG6/Au ) obtained after exposure of the SAMs to YCC (651 angle of incidence).Difference spectra are practically null for YCC.Adsorption of YCC onto T-OEG therefore appears absolutely negligible.Difference spectra related to the adsorption of YCC on a C18 SAM are also shown (light grey circles) for comparison.

Table 1
Best fit value and fit uncertainty for the adjustable parameters of the TF|I|S model (for details, see the text); the MSE value for each best fit is also shown.The fitting procedure was performed on averaged data and the two species of SAMs were considered independently