Preparation and characterisation of gold nanoparticle ...

Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 225 /239 www.elsevier.com/locate/colsurfa

Preparation and characterisation of gold nanoparticle assemblies on silanised glass plates Oliver Seitz a, Mohamed M. Chehimi a,*, Eva Cabet-Deliry b, Ste´phanie Truong a, Nordin Felidj a, Christian Perruchot c, Steve J. Greaves c, John F. Watts c a

Interfaces, Traitements, Organisation et Dynamique des Syste`mes (ITODYS), Universite´ Paris 7, Denis Diderot, associe´ au CNRS (UMR 7086), 1 rue Guy de la Brosse, Paris 75005, France b Laboratoire d’Electrochimie Mole´culaire, Universite´ Paris 7, Denis Diderot, associe´ au CNRS (UMR 7591), 2 place Jussieu, case 7107, 75251 Paris Cedex 05, France c School of Mechanical and Materials Engineering, University of Surrey, Guildford GU2 7XH, UK Received 24 July 2002; accepted 13 December 2002

Abstract Gold nanoparticles were prepared by the chemical reduction of AuCl4 and attached to aminopropyltrimethoxysilane-treated glass plates. The assemblies of gold nanoparticles on silanised glass were characterised by UV spectroscopy, X-ray photoelectron spectroscopy (XPS), atomic force microscopy (AFM) and surface-enhanced Raman scattering (SERS). The gold nanoparticles had a diameter in the range 409/10 nm as estimated by AFM. Gold was found to be in the metallic state as judged from XPS measurements of the Au 4f7/2 core electron binding energy. AFM showed that the gold nanoparticles experience a self-organisation on the silanised surface in such a way that the final assemblies have a certain degree of roughness and compactness. These characteristics are intimately related to the SERS effect as determined using the molecular probe bi-ethylene-pyridine at very low concentration. A huge enhancement of the Raman signals was observed and assigned to a coupling between gold particles. However, this SERS effect critically depends on the surface treatment of the substrate by the silane coupling agent, a procedure that is necessary for the attachment of the gold nanoparticles. # 2003 Elsevier Science B.V. All rights reserved. Keywords: Gold nanoparticles; Glass; Silane coupling agent; XPS; AFM; SERS

1. Introduction Materials with controlled surface physicochemical properties can be obtained by immobilisation

* Corresponding author. Fax: /33-144276814. E-mail address: [email protected] (M.M. Chehimi).

of catalysts, proteins, colloidal particles, etc. for various purposes such as environmental, biomedical or (electro)chemical analytical applications [1 / 6]. Either as planar or divided substrates, such materials must have not only the desired functionality but also a controlled surface chemistry (specific functional groups) and morphology. For

0927-7757/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0927-7757(02)00594-0

226

O. Seitz et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 225 /239

example, the surface roughness is a very important physical factor that governs surface-enhanced Raman scattering (SERS) activity [7]. The SERS effect leads to a spectacular increase of the Raman scattering cross-section for molecules adsorbed onto suitably roughened surfaces and thus has a large potential in analytical chemistry and for biological applications. This sensitive technique permits to detect organic molecules adsorbed on rough gold, silver or copper surfaces from very diluted solutions (down to 106 M), with a maximum enhancement factors of 106 /107. Two mechanisms are considered to account for the SERS effect. The electromagnetic mechanism arising from the excitation of localised surface plasmons (LSPs) resonance localised on roughened features is responsible for the main enhancement due to an enormous increase in the local electric field. The second mechanism is due to an increase in the polarisability of the adsorbed molecule when a resonant charge transfer can occur between the metal and the adsorbate (chemical effect). Rough metallic surface is of primordial importance to obtain SERS effect. For instance, excitation of LSP occurs on the gold particles (leading to a huge enhancement of the local electric field), the frequency of which depends on the size, the interparticle distance, and the surrounding medium. Unfortunately, when deposited on the bare glass substrate, the particles are loosely bound and therefore the SERS effect is not stable. Moreover, the control of the aggregation state is not possible at all. This is the reason why it is necessary to modify specifically the substrates prior to the immobilisation of the gold nanoparticles. Such a modification can be achieved with bifunctional molecules, one function binding to the substrate and the other at the free surface that is ready to bind to gold nanoparticles. On metallic substrates, usually dithiols are used. One thiol group is grafted on the surface via metalS bond and the second thiol binds a gold nanoparticle [8,9]. Other routes were proposed to bind gold particles to metallic substrates using diisocyanides [10] and SH-terminated aryl diazonium salts deposited by electrochemical reduction of the diazonium [11]. Alternatively, 1,4-aminothiophenol has been proposed to bind gold

particles to metallic substrates, thiol being linked to the substrate and the amine coordinating the particles [12]. More complex treatment of glass substrates was achieved with DNA oligomers that bind gold nanoparticles functionalised with the complementary DNA oligomer [13]. Organosilanes constitute a broad class of ‘‘molecular glues’’ which can be used for the attachment of gold particles on metal oxide substrates. amine-terminated Thiol-terminated [14 /16], [14,16 /18], alkyl-terminated [14] and phenyl-terminated silanes [16] were employed in this regard. The gold nanoparticles attached to silanised surfaces can serve as seeds for their further enlargement using in situ reduction of gold salt solutions [19], or for the stepwise construction of gold colloid multilayers [20]. Traditionally, silane coupling agents are used (i) as fibre-matrix adhesion promoters in composite materials [21], (ii) for preparing grafted silica for selective interactions with analytes in liquid chromatography [22], and (iii) to pretreat metallic surfaces for structural adhesive bonding [23]. As far as we are concerned, we modified the surface of quartz plates [24,25] and glass beads [25] by aminopropyltriethoxysilane (APTES) to achieve the immobilisation of dendrimers and polystyrene microspheres, both bearing surface aldehyde groups. We have also reported on the effect of APTES treatment of silica gel particles in the preparation of novel, conducting polypyrrolesilica composite materials [26,27]. Glass slides were also modified by APTES prior to coating by precipitating the conducting polypyrrole [28]. The resulting polymer coating was found to have a much better adhesion to APTES-treated glass than alkyl silanetreated glass. The results were interpreted in terms of hydrophobic and acid /base interactions at the glass /polypyrrole interface. This work is a follow-up to our ongoing research projects on the attachment of macromolecules and colloidal particles to silanised substrates [24 /29] and deals with the synthesis, deposition, organisation, and SERS activity of gold colloidal assemblies on glass plates. Aminopropyltrimethoxysilane (APTMS or APS) was used to obtain amine-functionalised glass plate


surfaces for the attachment of gold colloidal nanoparticles in aqueous media. The amount and morphology of the gold nanoparticle assemblies are monitored as a function of the APS treatment of glass plates. The uncoated and gold-decorated plates were characterised by UV /visible absorption spectroscopy for measuring the LSP resonance, and X-ray photoelectron spectroscopy (XPS) for monitoring the change in the surface chemical compositions of the untreated and silanetreated glass plates before and after gold particle deposition. Surface morphology of the uncoated and gold-decorated glass plates was observed by atomic force microscopy (AFM), and Raman spectroscopy was used to evaluate the SERS activity of the assemblies. These Raman measurements were assessed using trans -1,2-bis(4-pyridyl)ethylene (BPE), a common molecular probe to demonstrate the SERS activity of the resulting gold-decorated substrates.

2. Experimental

2.1. Preparation of gold nanoparticles Tetrachloroauric acid (0.5 M, AuHCl4, Sigma product) and sodium tricitrate (Prolabo) solutions (0.5 M) were first prepared. Forty microlitres of 0.5 M AuHCl4 solution was added to 10 ml of distilled water and the mixture was refluxed with stirring. The resulting AuHCl4 solution has a concentration of 2/103 M. One hundred and sixty microlitres of solution of sodium tricitrate (0.5 M) in water is added to 10 ml of water, resulting in a concentration of 8/10 3 M. This solution (10 ml) is added to the boiling AuHCl4 solution. The mixture of AuHCl4 and sodium tricitrate was refluxed for 1 h. The resulting grey mixture turned slowly to a light red colour. After boiling for 1 h, the solution was left to cool at room temperature. The gold sols prepared by this procedure were wine red. Such particles are well known to exhibit an extinction band located at

227

approximately 520 nm, due to excitation of LSP on the particles.

2.2. Silanisation of glass plates Glass plates (1 /2 cm2) were immersed in chromic acid and boiled for 4 h. They were then left to cool at room temperature in the acidic solution overnight. They were thoroughly rinsed with distilled water and left to dry for 1 h at 100 8C. The plates were then immersed in methanol (95%, Prolabo) solutions of aminopropyltrimethoxysilane (Aldrich) for 4 h at room temperature. The initial concentrations of APS were 0.5, 1, 2, 5, and 10% (v v1). After soaking in APS solutions, the plates were rinsed with a copious amount of methanol to remove any physisorbed APS. The rinsed plates were then treated ultrasonically in methanol and left to dry for 1 h at 100 8C. Note that it is very important to remove the excess of APS from the silanised glass to prevent gold particle aggregation in the colloidal suspension during the next step.

2.3. Deposition of colloidal gold particles on silanised glass plates Each silanised glass plate was dipped in a tube containing 2 ml of the freshly prepared gold colloidal suspension for a given time in the range 2 min to 4 h (see text). The plates were then removed from the suspension and left to dry for 10 min in an oven at 100 8C.

2.4. UV /visible absorption spectroscopy A UV /visible Perkin/Elmer spectrometer (model Lambda 2) was used. The 300/900 nm wavelength region was scanned as it includes the absorbance of the gold nanoparticles. The UV beam was relatively wide (ca. 6 mm diameter) thus leading to an average absorbance of the gold nanoparticles coated on glass.

228


2.5. Raman spectroscopy: assessment of SERS activity Raman spectra were recorded on a Dilor XY spectrometer equipped with 1800 l mm 1 holographic grating with a double foremonochromater used in subtractive mode (i.e. with no dispersion) and the dispersion completed by the third-stage spectrometer. High signal-to-noise ratio spectra were obtained using a thermoelectrically cooled Jobin-Yvon CCD matrix as a multichannel detector. Using 200 mm slit widths, the spectral resolution was less than 6 cm 1. Excitation was provided by a Spectra Physics model 207B helium /neon laser (632.8 nm, 25 mW power at the laser head). We used BPE as a molecular probe at very low concentration (10 6 M) to test the SERS activity of the attached colloidal particles. 2.6. Atomic force microscopy Untreated, silanised, and gold-coated silanised glass plates were imaged by a Nanoscope III Digital Instrument in the tapping mode using an Si3N4 tip cantilever. The cantilever oscillation frequency was set at 300 kHz. Tips of the cantilever were characterised by the radius of their curvature equal to 79/2 nm. No computer filtering procedure was used to treat the images. Tapping mode imaging was recorded with 256 pixels per line with a scan rate of 2.0 Hz. Surface roughness (Ra) was determined at the same scale (1 mm) for each sample.

3. Results

3.1. UV /visible absorption spectroscopy UV spectroscopy was used to characterise the absorbance of the gold particle suspension since they exhibit strong absorption bands. The absorption cross-section of the particles increases dramatically at the resonance due to surface plasmon excitations, whose frequencies depend on the refractive index of the surrounding medium. For gold particles, this maximum occurs at 520 nm (Fig. 1). In a first approximation, the spectral position of this maximum does not depend on the shape and the surrounding dielectric constants. From previous work, the size distribution of the same kind of gold colloids was deduced using discrete dipole approximation (DDA), displaying a low dispersion of size (radius ranges between 10 and 20 nm) in agreement with histogram distribution determined from TEM micrographs [30]. Fig. 2 exhibits UV spectra obtained for APStreated glass plates covered with gold particles. Attachment of gold particles has been obtained by dipping for 2 h the plates treated with APS at indicated initial concentration. When deposited on the glass substrate, in the presence of the coupling agent, the UV /visible spectra display a broadening of the band profile

2.7. X-ray photoelectron spectroscopy A Thermo VG Scientific SIGMA PROBE spectrometer equipped with a monochromatic Al Ka X-ray source (1486.6 eV) was used at a spot size of 400 mm. The pass energy was set at 100 and 50 eV for the survey and the high-resolution spectra, respectively. The step size was 1 eV for the survey spectrum and 0.2 eV for the highresolution spectra. Charge compensation was achieved with a flood gun of 6 eV electrons using standard procedures for this spectrometer. The surface composition was determined using the manufacturer’s sensitivity factors.

Fig. 1. UV /visible extinction spectra of various gold colloidal suspensions prepared in the same conditions.


229

Fig. 2. Effect of APS initial concentration on the UV absorbance of silanised glass plates soaked for 2 h in the gold colloidal suspension.

and a slight shift of the maximum towards the high wavelengths. This modification, with respect to the absorption profile of the suspension, is due to electromagnetic interaction between the particles. Note that UV absorption between 600 and 700 nm is particularly strong, giving rise to a large enhancement of the local electric field and is located within the excitation wavelength (633 nm) for obtaining SERS spectra. The comparison between plates dipped into the various APS solutions and for the same soaking time does not show any significant effect of the organosilane APS initial concentration. Indeed, the peak intensities around 550 nm are comparable. In contrast, the plates treated with the same APS concentration and dipped for various times in the colloidal suspension show a substantial increase in the optical density with soaking time due to a larger amount of deposited particles (Fig. 3). Therefore, the dipping time is an important factor controlling the formation of gold nanoparticle assemblies at the surface of silanised glass plates. Indeed, one can note that the coupling and aggregation of the particles are considerable even after only 10 min of soaking the APS-treated

plates in the colloidal suspension, as revealed by the development of the wide shoulder at approximately 650 nm, as confirmed by AFM investigations (see following section). 3.2. Atomic force microscopy AFM imaging of the gold-coated plates was used to characterise the coating in terms of particle size, shape, and organisation at the silanised glass surface. The APS coated on glass plates (figure not shown) exhibited an island-like structure, a result that we have already reported elsewhere [31]. These nanoclusters are roughly spherical in shape with a diameter of approximately 100 nm, higher than 159/5 nm diameter obtained previously using KOH-cleaned glass plates [28]. Fig. 4 depicts AFM images of gold-decorated silanised glass plates obtained after 10 min and 2 h soaking time in the colloidal suspension. The gold particles have a spherical shape, and clearly a close packing of the gold particles needs a few hours to occur in agreement with the literature [16]. Fig. 4(a) suggests that first, colloids tend to adsorb on the surface directly without forming any aggregates. Then, as the time of soaking increases, one

230


Fig. 3. Effect of soaking time in the gold colloidal suspension on the UV absorbance of gold-decorated silanised glass substrates. Initial APS concentration was 5%.

obtains a closer packing of the gold particles (Fig. 4(b)). A cross-section of the AFM profile allowed us to estimate the diameter of the particles by measuring the height of the spherical particle deposited on the substrate. We found by this approach an average diameter of 409/10 nm. The gold nanoparticle submonolayers have also been obtained by Lyon et al. [32] on evaporated gold surface for up to 90 min soaking time, however, with a lower coverage than that shown in Fig. 4(a). It is only on a glass substrate treated by a silane coupling agent (2-mercaptoethylamine) that a close packing of gold nanoparticles (size: 12 nm) was achieved [33]. In addition to the imaging and estimation of the gold nanoparticle diameter, we have determined the average roughness (Ra) of the various plates, the values of which are reported in Table 1. For silanised glass, Ra is very close to 0.755 nm, a value we previously established for a slightly different protocol of surface treatment [28] but for a slightly different protocol of glass cleaning and APS treatment. It is interesting to note that for silicon wafers with native silica overlayers, Balladur et al.

[34] obtained a very comparable roughness of 0.8 nm for 3-aminopropylmonoethoxydimethylsilane coated from a 2% (v v1) solution of anhydrous toluene. Table 1 clearly shows that even if the silanised glass plate is dipped for only 10 min, it roughens readily due to the deposition of gold particles. It is very important to point out at this stage that this contrasts very well with the case (not shown here) of untreated glass plates dipped for a longer time in the colloidal suspension and for which no change in roughness occurred for the reason it did simply not attach any particles. Ra increases for soaking time up to 2 h but then decreases after this time. This is due to the filling of the interparticle space when soaking time increases. Then, the surface reappeared relatively flat.

3.3. X-ray photoelectron spectroscopy XPS was used to characterise the chemical structure and composition of the untreated and silane-treated glass plate surfaces before and after gold decoration. In the latter case, the silanised


231

Fig. 4. AFM image of a silanised glass plate (5%, v v 1) in Au suspension with soaking times of 10 min (a) and 2 h (b).

Table 1 Average roughness for glass, silanised glass, and gold-coated silanised glass plates Materials

Glass

Glass /APS5%

Gold-coated silanised glass plates

Soaking times Roughness (nm)

0.388

0.785

2 min 4.65

1h 4.56

2h 5.53

4h 3.17

232


Fig. 5. XPS survey spectra of (a) a bare glass substrate; (b) APS (5%)-treated glass substrate; and (c) a glass substrate after a soaking time of 2 h in a gold colloidal suspension.

glass plates were obtained by treatment in 5% (v v1) APS. Fig. 5 shows XPS survey spectra of untreated glass, APS-treated glass, and gold-coated silanised glass. For untreated glass plate, the spectra show a sharp O 1s peak and Si 2p, Si 2s, and Na 1s peaks (Fig. 5(a)). There is some degree of surface

contamination by adventitious hydrocarbons, sulphur, and nitrogen. Sulphur originates from the sulfochromic acid used to clean the plates before APS treatment and gold particle attachment, whereas the origin of nitrogen is unknown. These contaminating species were removed at the APS treatment stage (see Fig. 5(b)). For APS-treated


233

Fig. 5 (Continued)

plate, the most significant peaks are attributed to oxygen, silicon, and carbon. The Na 1s and Na KLL peaks from the glass substrate are attenuated markedly. It is interesting to note that the uptake of APS leads to a more intense C 1s peak by comparison to glass. Fig. 5(c) shows a survey scan of the APS-treated glass substrate decorated by gold particles. This spectrum has a completely different structure by comparison to the spectra obtained for glass (Fig. 5(a)) and silanised glass (Fig. 5(b)). Indeed, gold is detected by its sharp Au 4f7/2 5/2, Au 4d5/2 3/2, and Au 4p3/2 peaks. O 1s and Si 2p peaks experience a fairly strong attenuation due to the immobilisation of gold nanoparticles on the silanised substrates. Fig. 6 shows the key XPS core-level regions for the untreated glass, glass /APS, and glass /APS/ Au specimens. In the case of gold particle-coated plates, it is very important to note that Au 4f7/2 peak is centred at 84 eV (Fig. 6(a)), which is a binding energy characteristic of gold in the metallic state [35]. For glass /APS substrate, the C 1s peak (Fig. 6(b)) is fitted with four components centred at 285, 286.1, 287.3, and 288.8 eV, assigned to C /C/C /H, C /N/C /O, C /O, and O /C /O groups, respec-

tively. Note that the C /N and C /O bond contributions are lumped into the same component for the sake of simplicity. The C 1s region of glass /APS5% /Au 4h (Fig. 6(c)) is fitted with five components centred at 285, 285.8, 287.0, 288.6, and 289.0 eV, assigned to the C /C/C /H, C /N/C /CO2, C /O, C /O, and O /C / O/O /C /O species, respectively. The latter has a contribution of carboxylates from the citrate anions, which are counterbalanced by Na cations, detected by the fairly intense Na 1s peak in Fig. 5(c). Actually, the Na 1s peak reappears by comparison to the silanised glass plate (Fig. 5(b)) and is connected with the gold colloidal nanoparticle surface. Gold nanoparticle surfaces are actually known to be negatively charged due to the existence of citrate anions [17]. Therefore, the charge balance is brought by Na at the free surface of the attached nanoparticles. It is worth noting that the main peak around 285 eV for glass /APS5% /Au 4h (Fig. 6(c)) is wider than that of glass /APS5% (Fig. 6(b)) due to the contributions of C /CO2 and C /OH from the citrate anions. The C /CO2 carbon type undergoes what is commonly called ‘‘b-shift’’ [36], that is the effect of oxygen (or fluorine) in b-position of the

234


Fig. 6. XPS core-level regions for key elements in the glass /APS and glass /APS /Au specimens: (a) Au 4f region from glass / APS5% /Au 4h; (b) C 1s region from glass /APS5%; (c) C 1s region from glass /APS5% /Au 4 h; and (d) N 1s region from glass / APS5% /Au 4 h.

carbon atom considered. b-Shifts can be as high as /0.7 eV relative to the normal C /C/C /H C 1s peak located at 285 eV. The contribution of functionalised carbon atoms (binding energy /285 eV) is only 20.1% in the case of glass /APS and increases significantly

in the case of gold-decorated plates to reach 45.9% for glass /APS5% /Au 4 h, for example. This change in the structure in the C 1s spectrum is attributed to the citrate anions. The N 1s peak (Fig. 6(d)) for APS5% /Au 4 h (similar to that of glass /APS5%) has two compo-


235

Fig. 6 (Continued)

nents centred at 400.39/0.3 and 402.39/0.3 eV assigned to free and quaternised amine from APS, respectively, and this is in agreement with XPS studies of APS-treated quartz [37] and glass [28] plates. The existence of protonated form of APS is certainly a key in the attachment of the nanoparticles to the glass plate surface as will be discussed below.

The surface compositions (in at.%) of the various plates are reported in Table 2. O/Si atomic ratio is always higher than 2 due to the different oxides from glass, especially Na2O, and to the citrate ions in the case of gold-coated plates. In the case of the silanised glass plates, the C/N ratio far exceeds the theoretical value of 3 due to the adventitious hydrocarbon contamination.

236


Table 2 Surface chemical composition (at.%) as determined by XPS for glass, silanised glass, and gold-decorated silanised glass Materials

Cleaned glass Glass /APS0.5% Glass /APS1% Glass /APS2% Glass /APS5% Glass /APS10% Glass /APS /Au 1 h Glass /APS /Au 2 h Glass /APS /Au 4 h

Surface composition (at.%) Na

O

N

C

Si

Au

S

Cl

3.27 1.91 1.07 1.91 1.13 2.20 7.05 6.11 2.53

61.0 48.7 44.6 46.1 46.2 53.9 38.6 45.5 47.2

2.33 2.19 2.18 1.62 2.32 2.12 0.77 0 2.11

13.4 29.9 35.1 34.3 32.5 24.1 35.7 24.5 28.3

15.5 16.6 16.1 14.8 16.9 17.6 9.31 9.78 17.0

0 0 0 0 0 0 7.69 13.0 2.89

4.46 0 0 0 0 0 0 0 0

0 0.76 0.93 1.26 0.86 0 0.87 1.17 0

For Glass /APS /Au plates, the initial concentration of APS was 5% (v v 1).

The APS grafted layer is very thin since the N/Si atomic ratio is much smaller than 1, the theoretical ratio for a polyaminosiloxane. For the silanised glass plates, using N/Si atomic ratio, we obtained a high-affinity adsorption isotherm (not shown) of APS. Again, saturation is obtained for 1% APS initial concentration or less [26,28,37]. It is surprising to note that the surface composition in the case of the gold-coated substrates is apparently not very rich in gold although this element leads to sharp peaks in the survey scans. Chloride species resulting from the reduction of AuHCl4 were not always removed from the goldcoated plates (see Table 2). Silicon and nitrogen are detected at a fairly high extent because the gold particles do not form multilayers on the silanised substrates (see Fig. 4(a)). 3.4. Raman measurements The Raman characterisation of the gold-decorated APS-treated glass plates allowed us to compare the SERS activity of the assemblies as a function of various parameters, e.g. APS initial concentration and dipping time. The SERS effect was monitored using the peak area of the signal centred at 1200 cm 1 for 10 6 M BPE aqueous solution. Fig. 7 shows Raman spectra of BPE at indicated soaking time, with a concentration of 106 M BPE solution, in interaction with the gold-decorated silanised glass. The threshold of

the SERS effect is obtained for a 10 min soaking time of the silanised glass plates in the colloidal suspension. It is worth to note that the massive increase in SERS effect was obtained for a concentration three orders of magnitude lower than 1 mM used by Grabar et al. [16] for similar gold nanoparticle size (ca. 60 nm).

4. Discussion and conclusion The results reported in this work clearly indicate that the surface treatment of glass plates by aminopropylsilane coupling agent (APS) is very effective in the immobilisation of gold particles. No such attachment occurred on bare, untreated glass plates. Therefore, APS acts as an important adhesion promoter for gold nanoparticles. Adhesion of gold particles is most probably due to interfacial electrostatic interactions between the positively charged, protonated terminal amino groups in APS and the citrate anions attached to the gold nanoparticles. These functional groups were simultaneously detected by high-resolution XPS. Raman spectroscopy brings evidence for the SERS activity of these assemblies attached to silanised glass by using bi-ethylene-pyridine as a model molecular probe. However, this SERS activity requires a minimal soaking time of 10 min (see Fig. 7) for the gold particles to aggregate at the surface of silanised glass as shown by optical


237

Fig. 7. Effect of soaking time in a colloidal suspension on the Raman spectra of BPE (10 6 M) at the surface of gold-coated silanised glass plates: (a) 1 /4 h exposure (APS initial concentration was 10% (v v 1)) and (b) up to 30 min exposure (APS initial concentration was 5% (v v 1)).

analysis. This corroborates the AFM study which permitted to connect the surface roughness of the substrates to their SERS activity. The highest SERS activity was obtained for 2 h soaking

corresponding to the highest surface average roughness (Table 1). Although an important change of roughness was obtained for 10 min soaking time, the SERS effect was very poor or nil

238


for the corresponding plates because it did not exhibit gold particle aggregation (characterised by a UV band around 650 nm, see Fig. 3) which is a prerequisite for SERS activity. As far as APS treatment is concerned, XPS results explicitly showed that saturation occurred for all surface treatments. This is in close connection with the observation that the SERS activity was depending mainly on the soaking time, APS being necessary only for attaching the particles to the surface. Therefore, an APS initial concentration of 0.5 /1% (v v 1) is ideal for gold particle attachment. Despite the homogeneity of the glass surface treatment, silicon was unambiguously detected by XPS from the underlying silanised substrate although the particle size is 40/65 nm, thus, far higher than the analysis depth probed by XPS (10 nm maximum). The gold particles assemble in a random way at the surface of glass (as revealed by AFM) to form submonolayers and thus leaving some unoccupied surface sites from the glass plates. Nevertheless, the so-assembled nanoparticles exhibit a certain degree of compactness and roughness which, obviously, yielded an important SERS activity for a BPE concentration as low as 106 M by comparison to a previously published literature [16]. To summarise, this study clearly shows that the combination of bulk and surface analytical techniques, namely UV absorption spectroscopy, AFM, XPS, and Raman spectroscopy permits to correlate the chemical, morphological, and optical properties of gold nanoparticle assemblies immobilised on silanised glass plates. More importantly, the simple method of silanisation (at low silane initial concentration) presented in this work is very effective in attaching gold nanoparticles via electrostatic interactions, as evidenced by XPS, and to induce SERS active surfaces for short soaking time of the plates in the gold nanoparticle suspension.

References [1] K. Tamaru, Dynamic Processes on Solid Surfaces, Plenum Press, New York, 1995. [2] E.K. Pye, B. Chance, Meth. Enzymol. 44 (1976) 357.

[3] C.A.C. Sequeira, in: R.G. Linford (Ed.), Electrochemical Science and Technology, vol. 2, Elsevier, London, 1990, p. 319. [4] B. Miksa, M. Wilkzynska, C. Cierniewsky, T. Basinska, S. Slomkowski, J. Biomater. Sci. Polym. 7 (1995) 503. [5] M. Moskovits, Rev. Mod. Phys. 57 (1985) 783. [6] M.D. Musick, D.J. Pena, S.L. Botsko, T.M. McEvoy, J.N. Richardson, M.J. Natan, Langmuir 15 (1999) 844. [7] R. Cardona, G. Guntherodt (Eds.), Light Scattering in Solids, Springer, Berlin, 1984. [8] E. Hutter, S. Cha, J.-F. Liu, J. Park, J. Yi, J.H. Fendler, D. Roy, J. Phys. Chem. B 105 (2001) 8. [9] M.D. Musick, C.D. Keating, L.A. Lyon, S.L. Botsko, D.J. Pena, W.D. Holliway, T.M. McEvoy, J.N. Richardson, M. Natan, Chem. Mater. 12 (2000) 2869. [10] M.A. Ansell, E.B. Cogan, C.J. Page, Langmuir 16 (2000) 1172. [11] J.A. Harnisch, A.D. Pris, M.D. Porter, J. Am. Chem. Soc. 123 (2001) 5829. [12] T. Zhu, Z.H. Zhu, J. Wang, Y.C. Wang, Z.F. Liu, Mol. Cryst. Liq. Cryst. 337 (1999) 237. [13] C.M. Niemeyer, B. Ceyhan, S. Gao, L. Chi, S. Peschel, U. Simon, Colloid. Polym. Sci. 279 (2001) 68. [14] S.L. Westcott, S.J. Oldenburg, T. Randall Lee, N.J. Halas, Langmuir 14 (1998) 5396. [15] M.S. Fleming, D.R. Walt, Langmuir 17 (2001) 4836. [16] K.C. Grabar, K.J. Allison, B.E. Baker, R.M. Bright, K.R. Brown, R.G. Freeman, A.P. Fox, C.D. Keating, M.D. Musick, M. Natan, Langmuir 12 (1996) 2353. [17] J. Zheng, Z. Zhu, H. Chen, Z. Liu, Langmuir 16 (2000) 4409. [18] Y. Jin, X. Kang, Y. Song, B. Zhang, G. Cheng, S. Dong, Anal. Chem. 73 (2001) 2843. [19] K.R. Brown, M. Natan, Langmuir 14 (1998) 726. [20] (a) M.D. Musick, C.D. Keating, M.H. Keefe, M. Natan, Chem. Mater. 9 (1997) 1499; (b) M.D. Musick, C.D. Keating, M.H. Keefe, M. Natan, J. Phys. Chem. B 130 (1999) 5826. [21] K.L. Mittal (Ed.), Silanes and Other Coupling Agents, VSP, Utrecht, 1992. [22] G. Gu¨bitz, Selective sample handling and detection in high performance liquid chromatography, in: R.W. Frei, K. Zech (Eds.), Journal of Chromatography Library, vol. 39A (Chapter 3), Elservier, Amsterdam, 1988. [23] M.L. Abel, J.F. Watts, R.P. Digby, Int. J. Adhes. Adhesives 18 (1998) 179. [24] B. Miksa, S. Slomkowski, M.M. Chehimi, M. Delamar, J.P. Majoral, A.M. Caminade, Colloid. Polym. Sci. 277 (1999) 58. [25] S. Slomkowski, B. Miksa, M.M. Chehimi, M. Delamar, E. Cabet-Deliry, J.P. Majoral, A.M. Caminade, React. Funct. Polym. 41 (1999) 45. [26] C. Perruchot, M.M. Chehimi, M. Delamar, F. Fievet, Surf. Interf. Anal. 26 (1998) 689. [27] C. Perruchot, M.M. Chehimi, D. Mordenti, M. Briand, M. Delamar, J. Mater. Chem. 8 (1998) 2185.

O. Seitz et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 218 (2003) 225 /239 [28] C. Perruchot, M.M. Chehimi, M. Delamar, E. CabetDeliry, B. Miksa, S. Slomkowski, A.M. Khan, S.P. Armes, Colloid. Polym. Sci. 278 (2000) 1139. [29] M.M. Chehimi, A. Azioune, S. Bousalem, A.Ben Slimane, A. Yassar, S. Slomkowski, Colloid Polymers: Preparation and Biomedical Application, H. Elaissaric (Ed.), Marcel Dekker Inc., New York, in press. [30] N. Fe´lidj, J. Aubard, G. Levi, J. Chem. Phys. 11 (1999) 1195. [31] W. Yuan, W.J. van Ooij, J. Colloid. Interf. Sci. 185 (1997) 197. [32] L.A. Lyon, D.J. Pena, M. Natan, J. Phys. Chem. B 130 (1999) 5826.

239

[33] K.R. Brown, L.A. Lyon, A.P. Fox, B.D. Reiss, M. Natan, Chem. Mater. 12 (2000) 314. [34] V. Balladur, A. Theretz, B. Mandrand, J. Colloid. Interf. Sci. 194 (1997) 408. [35] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, J. Chastain (Eds.), Handbook of X-ray Photoelectron Spectroscopy, Perkin /Elmer, Eden Prairie, MN, 1992, p. 182. [36] G. Beamson, D. Briggs (Eds.), High Resolution XPS of Organic Polymers, The Scienta ESCA300 Database, Wiley, Chichester, 1992. [37] D. Kowalczyk, S. Slomkowski, M.M. Chehimi, M. Delamar, Int. J. Adhes. Adhesives 16 (1996) 227.