Growth by heat treatment of silver nanorods inside mesostructured silica thin films: Synthesis, colours of thin films, study of some experimental parameters and characterization

Microporous and Mesoporous Materials 139 (2011) 45–51

Contents lists available at ScienceDirect

Microporous and Mesoporous Materials

journal homepage: www.elsevier .com/locate /micromeso

Growth by heat treatment of silver nanorods inside mesostructured silica thin films:Synthesis, colours of thin films, study of some experimental parametersand characterization

Reine Sayah a, Latifa Nadar a, Francis Vocanson a,⇑, Yann Battie a, Stéphanie Reynaud a, Ruben Vera b,Aziz Boukenter a, Nathalie Destouches a,⇑a Université de Lyon, F-42023, Saint-Etienne, France; CNRS, UMR 5516, Laboratoire Hubert Curien, F-42000, Saint-Etienne, France; Université de Saint-Etienne, Jean Monnet,F-42000, Saint-Etienne, Franceb Université de Lyon, F-69003, Lyon, France; Université Lyon I, Centre de diffractométrie Henri Longchambon, 43 boulevard du 11 novembre 1918, F-69622, Villeurbanne Cédex, France

a r t i c l e i n f o

Article history:Received 25 June 2010Received in revised form 29 September2010Accepted 12 October 2010Available online 19 October 2010

Keywords:SynthesisMesoporous silicaSilver nanorodsThin films

1387-1811/$ - see front matter � 2010 Elsevier Inc. Adoi:10.1016/j.micromeso.2010.10.015

⇑ Corresponding authors. Tel.: +33 0 4 77 91 58 22E-mail addresses: francis.vocanson@univ-st-etienn

[email protected] (N. Destouches).

a b s t r a c t

In this paper we report a soft route leading to spherical or rod nanoparticles confined in SBA-15 thin filmsexhibiting a hexagonal mesostructure. For that purpose, these transparent mesoporous silica thin filmsare elaborated in the presence of a non ionic triblock copolymer P123 as structure directing agent. Aftersoaking in a silver salt solution, they are heat treated to grow silver structures. In function of severalexperimental parameters as the immersion time in the silver salt solution, the solvent of the silver saltsolution or the heat treatment temperature, nanorods and/or nano spherical particles can be obtained.These coloured samples are characterized by X-ray diffraction (XRD), ultraviolet–visible (UV–vis) absorp-tion and Fourier transformed infrared (FTIR) spectroscopies, atomic force microscopy (AFM), scanningelectron microscopy (SEM) and transmission electron microscopy (TEM). The presence of the nanorodshas been showed by using polarized UV–vis spectroscopy. The results show that most of silver nanopar-ticles grow inside the film thickness.

� 2010 Elsevier Inc. All rights reserved.

1. Introduction

In recent years, silver nanoparticles have attracted much atten-tion due to their exceptional catalytic and optical properties [1].The latter, which lead to various colors, result from the collectiveoscillation of conduction electrons, when interacting with an elec-tromagnetic wave, which is called surface plasmon resonance(SPR). The shape and the localization of the SPR spectrum stronglydepend on the silver nanoparticles size and shape, the interparticleinteractions, and the dielectric properties of the local environment[2]. In this last case, the k of the SPR band grows with the index ofthe matrix; it is due to the excitation of surface evanescent fieldswhich increase the induced dipole moment of the molecule, themodulation of the plasmon radiation by the oscillating nuclei,and the increasing of the radiation of the induced dipole.

Silver particles incorporation into mesoporous compounds isusually achieved either by adding silver salt precursors to thesol–gel mixture [3] or by soaking already formed mesoporous com-pounds in a silver salt solution [4]. These methods rely on thereduction or the decomposition of the metal salt precursor by

ll rights reserved.

; fax: +33 0 4 77 91 57 81.e.fr (F. Vocanson), nathalie.-

means of thermal treatment in air or in an hydrogen atmosphere[5], irradiation with UV light or gamma rays [6] or chemical reduc-tion by using reducing agents like borohydride [7], sodium citrate[8], ascorbic acid [9], hydrazine dihydrochloride [10], dim-ethylsulfoxyde (DMSO) [11]. Silver nitrate (AgNO3) [7–10] is themost common source of silver ions although silver sulfate (Ag2SO4)[12], silver 2-ethyl-hexagonate [11] and silver perchlorate [13]have also been used.

Mesostructured porous silica (e.g. MCM-41, SBA-15) [14] is aparticularly attractive host matrix due to its highly porous inter-penetrating channel system without any capillary effect that couldprevent the species diffusion, its surface area that can reach1000 m2 g�1, its high thermal and mechanical stability and thepossibility of tuning pore size between 2 nm and 10 nm [15].

Recently, the formation of well-ordered spherical silver nano-particles in mesoporous silica films, impregnated with silver saltand chemically treated, was reported [16]. In the present contribu-tion, we report on the silver nanorods preparation into transparentsilica mesoporous SBA-15 thin films deposited on glass substratesusing the dip-coating method. The preparation and characteriza-tion of mesoporous silica thin films as well as and the in situ for-mation of silver nanorods in these films by thermal treatmentare described. The influence of the soaking time in the silver solu-tion, the solvent used, the temperature and reduction time, on the

46 R. Sayah et al. / Microporous and Mesoporous Materials 139 (2011) 45–51

silver nanorod formation are studied, giving in function of theparameters, spherical or rod silver nanostructure. By using polar-ized UV–vis spectroscopy, the presence and the type of silver nano-particles (spherical or rod) have been showed.

2. Experimental

2.1. General procedures

SBA-15 synthesis and dip-coating were conducted in a cleanroom. Tetraethyl orthosilicate (TEOS, P99%) was purchased fromFluka and poly-(ethylene oxide)-poly(propylene oxide)-poly(ethyl-ene oxide) block copolymer (Pluronic 123, MW: 5000) from Al-drich. Concentrated chlorhydric acid (HCl, 37%) was obtainedfrom Roth and ethanol (EtOH, absolute) from Carlo Erba. Silver ni-trate (AgNO3, 99%) was purchased from Acros and ammoniumhydroxide (NH4OH, 5 N) from Sigma–Aldrich.

2.2. Characterizations

Low-angle X-ray powder diffraction (XRD) data were acquiredon a Bruker-AXS ‘‘D8 Advance” diffractometer using Cu Ka mono-chromatic radiation (k = 1.5418 Å). Scanning electron micrographs(SEM) were obtained using a FEI NovananoSEM microscope (SEM–FEG) operating at an acceleration voltage of 10–15 kV. TEM imageswere acquired with a TOPCON EM002B TEM operating at 200 kV.Samples for TEM characterization were prepared by scrappingthe films, fragments being then deposited directly on a copper gridcoated with holey carbon film. AFM pictures were recorded with anAgilent Technologies 5500 instrument using a Si tip in acoustic ACmode. UV–vis and FTIR absorption spectra were respectively re-corded on a Perkin Elmer Lambda 900 UV/Vis/NIR spectrophotom-eter and on a Perkin Elmer Spectrum One spectrophotometer. Thefilm thickness was determined using a Veeco Dektak 3 ST surfaceProfiler with a diamond stylus.

2.3. Synthesis of SBA-15 mesoporous silica films

Mesoporous SBA-15 type silica was used as host matrix and wasprepared by the acidic catalyzed, non ionic assembly pathway de-scribed by Alberius et al. [17]. First, tetraethyl orthosilicate (10.4 g)was diluted in ethanol (12.0 g) before addition of a water solution(5.4 g) at pH = 2. This solution was stirred for 20 min at room tem-perature. A separate solution was prepared, containing poly(ethyl-ene oxide)-poly(propylene oxide)-poly(ethylene oxide) blockcopolymer pluronic P123 (EO20PO70EO20) (2.75 g) dissolved in eth-anol (8.0 g). The two solutions were combined, stirred for 1 h andthen aged for 2 h. Thin films were deposited by dip-coating at awithdrawal speed of 7 cm/min on glass slides before respectivelycleaning with pure water, acetone and ethanol. The films wereair dried at ambient temperature with a humidity rate of around45% for 24 h to increase the extent of silica cross linking. Finally,the structure directing agent (pluronic P123) was removed fromthe films by calcination at 450 �C during 4 h with a heating rateof 1 �C/min.

2.4. Preparation of mesoporous silica films with silver nanoparticles

The calcined films were immersed in three different solutions ofsilver precursor at 0.5 M for different durations varying from 40 to90 min. The three solutions were prepared by dissolving the silversalt in water (first solution), in a mixture of ethanol/water (1/1)(second solution) or by adding NH3�H2O in the solution with etha-nol/water until a clear solution was obtained (third solution). Aftersoaking, the films were rinsed in water, dried at room temperature

and heat-treated in air between 100 �C and 400 �C for several times(2 h or 6 h).

Table 1 gives the experimental conditions applied for the elab-oration of each sample. It must be red as follows: the SBA-15@Ag-40 min was obtained after soaking a thin film in a solution of½AgðNH3Þ2�

þNO�3 at 0.5 M in EtOH–H2O (1:1 v/v) for 40 min andheat-treating it at 200 �C for 2 h to decompose the silver salt intosilver nanoparticles. The complex solution was obtained by addingNH3�H2O (5 N) to AgNO3 in EtOH until obtaining a clear solution.Ultra pure water was then added to complete the proportions be-tween the EtOH and the aqueous phase.

3. Results and discussion

3.1. Characterization of mesoporous silica films

The common starting point for all studies described herein isthe mesoporous silica thin film. Before and after calcination, allthe coatings are transparent and crack free over the whole surface.After calcination, the film thickness is around 360 nm and FTIRanalysis discloses the absence of organic functions confirmingthe efficiency of the heat treatment.

Fig. 1A and B show small-angle XRD patterns of the SBA-15 thinfilms, respectively collected before and after calcination. In bothcases, two well resolved peaks in the [0.6�–3�] 2h-range are ob-served, and can be attributed to the (1 0 0) and (2 0 0) reflexionsof a hexagonal mesostructural ordering. The absence of the(1 1 0) reflexion is likely to be due to the hexagonal unit cell orien-tation with the c and a axes parallel to the substrate plane, as wasexplained by Zhao et al. [18]. This orientation is privileged becauseof preferential interaction between the copolymer PEO moities andthe silicon oxide of the glass slide. A decrease of the unit cellparameter after calcination is noted; the d spacing obtained variesfrom 85.6 Å for the as-synthesised SBA-15 film, to 55.0 Å for thecalcined film. This variation is due to the anisotropic shrinkageoccurring normally to the film plane, i.e. to the mesostructure com-pression in the direction perpendicular to the film surface as de-scribed by Stucky [17].

AFM images of the calcined thin films confirm that the channelsof the hexagonal structure lie in the plane parallel to the substrate(Fig. 2). In this condition, the pore-to-pore distance is estimated to9.7 nm after calcination confirming that the contraction of themesostructure is making along the direction perpendicular to thesubstrate.

3.2. Preparation and characterization of silver nanoparticles withinmesoporous silica films

Silver nanorods were prepared within the silica thin films de-scribed above as following: the SBA-15 thin films were soaked insilver precursor solutions at 0.5 M and then heat treated. This con-centration of the precursor solution was optimized after prelimin-ary experiments by using UV–vis analysis; it led to a greaterconcentration of nanorods. After heat treatment, the colorless filmsturned pale yellow, pale pink, gray, brown or green (Fig. 3). Smallangle X-ray diffraction patterns showed that the different treat-ments accomplished on the SBA-15 thin films do not affect thehexagonal mesostructure.

The colored films were characterized by UV–visible absorptionspectroscopy to identify the nanoparticle size and shape. Thesespectral measurements were carried out under normal incidenceand under an incidence angle of 50� with a linearly polarized light.The polarization angle was rotated to be perpendicular to the inci-dence plane (TE polarization) or in the incidence plane (TM polar-ization). Measurements under normal incidence give information

Table 1Experimental conditions, sample notations and first absorbance wavelength.

Sample name Imp solvent Basea Imp time (min)b T� decomp (�C)c Time decomp (h)d kmax absorb (nm)e

SBA-15@Ag-40 min EtOH/H2O Yes 40 200 2 489SBA-15@Ag-60 min EtOH/H2O Yes 60 200 2 447SBA-15@Ag-90 min EtOH/H2O Yes 90 200 2 480SBA-15@Ag–EtOH/H2O EtOH/H2O No 60 200 2 441SBA-15@Ag–H2O H2O No 60 200 2 456SBA-15@Ag-100 �C EtOH/H2O Yes 60 100 2 451SBA-15@Ag-150 �C EtOH/H2O Yes 60 150 2 455SBA-15@Ag-250 �C EtOH/H2O Yes 60 250 2 477SBA-15@Ag-300 �C EtOH/H2O Yes 60 300 2 455SBA-15@Ag-400 �C EtOH/H2O Yes 60 400 2 418SBA-15@Ag-200 �C-6 h EtOH/H2O Yes 60 200 6 439

a Base = NH3�H2O.b Soaking time.c Decomposition temperature.d Decomposition time.e kmax absorbance (1st band).

Fig. 1. X-ray diffraction patterns of mesostructured silica thin films: (A) assynthesised (B) after calcination at 450 �C.

Fig. 2. AFM images of a SBA-15 silica film before heat treatment; (a) 2D, and (b) 3D.The layer roughness has been filtered in order to better visualize the structure ofinterest.

Fig. 3. images of thin films including silver nanoparticles; (a) SBA-15@Ag-40 min,(b) SBA-15@Ag-60 min, (c) SBA-15@Ag-90 min, (d) SBA-15@Ag–EtOH/H2O, (e) SBA-15@Ag–H2O, (f) SBA-15@Ag-200 �C-6 h (g) SBA-15@Ag-100 �C, (h) SBA-15@Ag-150 �C, (i) SBA-15@Ag-250 �C, (j) SBA-15@Ag-300 �C, and (k) SBA-15@Ag-400 �C.

R. Sayah et al. / Microporous and Mesoporous Materials 139 (2011) 45–51 47

on the nanorod width and length whereas the nanorod thickness isalso probed under oblique incidence.

Several parameters were varied in order to obtain specificallysilver nanorods. We report hereafter the optical absorption spectrameasured under normal incidence, of SBA-15 thin films obtainedwith different immersion times in the silver precursor solution, dif-ferent solvents of the silver precursor, different heat treatmenttemperatures and decomposition times. Under normal incidencethe absorption spectra do not depend on the polarization orienta-tion since the channels of the mesostructure, and then the silvernanorods, are randomly oriented in the film plane over the illumi-nated area (about 1 cm2). Therefore, only the measurements re-corded with a TE polarization are shown. In addition thin filmshave been studied by electron microscopy (SEM or TEM).

Fig. 4a shows that the immersion time in the silver precursorsolution strongly influences the nanoparticle shape and size. Aftera 40 min-long immersion of the mesoporous thin film in the silverprecursor solution and a heat-treatment at 200 �C for 2 h, a singleabsorption band lies in the sample spectrum, centered at 489 nm(Table 1). This band corresponds to the SPR band of spherical silvernanoparticles and its position depends on the mean particle size.As for SBA-15@Ag-60 min sample, it clearly shows two absorptionbands, at 447 nm and 742 nm, which can be attributed to the pres-ence of silver nanorods. The first band is mainly linked to the par-ticle short axis (half-width b; Fig. 5) whereas the second one shiftsto the red when the particle long axis (half-length a; Fig. 5) in-creases. Dispersion in the nanorod length can therefore lead toan inhomogeneous widening of this second absorption band. This

Fig. 4. UV–vis spectra of SBA-15 thin films containing silver nanoparticles under normal incidence for different (a) soaking times in the silver precursor solution, (b) silverprecursor solvents, (c) decomposition temperatures, and (d) decomposition times.

c

b

a x

y

z

Fig. 5. Scheme of an ellipsoidal particle. a is to the nanorod half-length, b thenanorod half-width and c the nanorod half-thickness.

Fig. 6. SEM micrographs of (a) SBA-15@Ag-40 min, (b) SBA-15@Ag-90 min, and (c)SBA-15@Ag-60 min.

48 R. Sayah et al. / Microporous and Mesoporous Materials 139 (2011) 45–51

widening may be the cause of the high level of absorbance in thered and IR ranges going with the absence of a well-defined secondabsorption band after 90 min of soaking.

The SEM pictures of Fig. 6 support the previous reasoningsshowing the shape and the size of silver nanoparticles lying onthe samples surface. They show rather short nanoparticles onSBA-15@Ag-40 min sample (Fig. 6a) and nanorods of variouslengths on the surface of both other samples. The nanorod lengthmainly varies between about 10 and 70 nm on the SBA-15@Ag-60 min film (Fig. 6c) and is more heterogeneous on the SBA-15@Ag-90 min film (Fig. 6b). TEM images of the SBA-15@Ag-60 min sample (Fig. 7a and b) give similar information to theSEM pictures but prove in addition that silver nanorods also growin the film thickness. These silver nanorods trapped in the filmmesostructure have a width ranging from 7.8 nm and 8.9 nm anda length varying from 17 nm to 70 nm.

The high resolution TEM picture of Fig. 7c, displays (1 1 1) lat-tice fringes of cfc (cubic face centered) silver and proves that thenanorods are polycrystalline silver particles. It can be concludedfrom these results that increasing the soaking time in the silversolution from 40 min to 90 min increases the silver content inthe film and promotes the growth of a larger number of nanorods

and of longer nanorods. A soaking time of 60 min will be used inthe next experiments. Contrary to what was observed under nor-mal incidence, the absorption spectra of the samples measured un-der oblique incidence depend on the polarization state of the

Fig. 7. TEM images of SBA-15@Ag-60 min thin film with silver nanorod. (a and b):top-views, and (c): high resolution.

R. Sayah et al. / Microporous and Mesoporous Materials 139 (2011) 45–51 49

incident light. Because of the film shrinkage occurring during theheat treatment, the thickness of the channels has been observedto be lower than their width (XRD, SEM and AFM data). The spec-troscopic measurements under oblique incidence prove that thesilver nanorods that grow in the channels of the mesostructureare also less thick than wide and can be modelled as 3-axis ellip-soids, as sketched in Fig. 5.

As it will be calculated in the next section, such particles pres-ent three plasmon resonances corresponding to the collectiveoscillation of electrons along each axe. Fig. 8 shows the experimen-tal spectra of SBA-15@Ag-60 min measured under oblique inci-dence for both TE and TM polarizations. Both spectra have SPRbands centered at around 447 nm and 742 nm that a third SPRband located at 333 nm that shows that the nanorod thicknessreally differs from their width.

The role of the silver precursor solvent on the quantity of silverparticles reduced in the film is studied in Fig. 4b. Regardless of thesolvent used, two absorption peaks are observed, resulting fromthe formation of silver nanorods within the silica films. But, theabsorbance level is lower without base than with base. The SBA-15 silanol surface becomes charged negatively in the presence ofthe basic precursor solution, which improves the attraction of thesilver complex ½AgðNH3Þ2�

þNO�3 charged positively and thus in-creases the silver content in the film. The addition of ethanol inwater also clearly increases the absorbance level. This can be ex-

Fig. 8. UV–vis spectra of SBA-15@Ag-60 min recorded under an incidence angle of50� in TE and TM polarizations.

plained by the higher diffusion of the silver salt in the mesoporousstructure due to the density difference between the two solvents[19]. Despite the lower absorbance level of the spectra measuredon samples soaked in silver precursor solutions without base, thered-shift of their second absorption band is likely to result froma higher mean nanorod length than in the SBA-15@Ag-60 min sam-ple. The TEM image of SBA-15@Ag–H2O (Fig. 9) shows essentiallyspherical particles, with diameter comprised between 7.6 nm and8.6 nm, and some nanorods that have the same width than thechannels and a length varying between 12 nm and 30 nm. Thisobservation confirms the UV–vis spectrum and especially the pres-ence of a low intensity band at the large wavelengths.

Fig. 4c shows the crucial role of the heat treatment temperatureon the silver nanoparticles formation. The conditions used forthese experiments are the following: a 60 min-long soaking inEtOH/H2O solvent with NH3�H2O. The absorption spectrum ofFig. 4c and SEM observations (Fig. 10a) show that silver nanorodsgrow in the film mesostructure from 100 �C. At 150 �C and200 �C, the density of particles is higher (Fig. 10b) resulting in ahigher absorption (Fig. 4c). At 250 �C, a large number of particles,with a circular section from the top and a diameter ranging from5.4 nm 30.5 nm, appear on the film surface to the detriment ofnanorods that become scarce (Fig. 10c). This trend leads to the de-crease of the second absorption band and to an increase of the firstone (more spherical-like silver nanoparticles) on the UV–visiblespectrum (Fig. 4c). At 300 �C, the absorbance level at resonance islower than at 150 �C and only the first absorption band remains(Fig. 4c). The particles are hardly distinguishable on the film sur-face by SEM (Fig. 10d) and they look small. Therefore, at this tem-perature most particles having grown during the temperatureincrease seem to be partially reoxidized, their size and their num-ber decreasing. At 400 �C, the oxidation intensifies and the maxi-mum absorbance level decreases (Fig. 4c). The SEM pictures alsoshow that the mesostructure is preserved, whatever the experi-mental conditions, after the thermal growth of the nanoparticlesin the films. After 250 �C, the UV–vis spectra with the diminutionof the absorbance level, the colour of the thin films (Fig. 3i, j and k)and the SEM images show the disappearing of the silver particlesdue to the temperature as reported by Seal [20]. The silver colloidsare get oxidized while being heat-treated, and eventually diffuseaway from the surface [20] and toward the glass substrate [21].

Finally, Fig. 4d shows the influence of the decomposition timeon the nanoparticle growth. A higher level of absorbance at reso-nance is reached and a red-shift of the second SPR band is obtainedafter a longer heat treatment at 200 �C. In opposition with the ther-mal treatment parameter, the treatment duration does not influ-ence the formation of spherical particles in the surface of thethin film as shown in Fig. 11.

3.3. Numerical simulations

The presence of three distinct SPR bands in the UV–visible spec-trum of the silver nanorods can be interpreted using the Gans the-ory [22]. This theory allows to estimate the extinction cross sectionof small ellipsoidal nanoparticles, such that one described in Fig. 5,

Fig. 9. TEM image of SBA-15@Ag–H2O.

Fig. 10. SEM micrographs of (a) SBA-15@Ag-100 �C, (b) SBA-15@Ag-150 �C, (c) SBA-15@Ag-250�C and (d) SBA-15@Ag-300 �C.

Fig. 11. SEM micrographs of (a) SBA-15@Ag-60 min and (b) SBA-15@Ag-200 �C-6 h.

0

500

1000

1500

2000

2500

300 500 700 900 1100Wavelength (nm)

Extin

ctio

n cr

oss-

sect

ion

(nm

²)a=8.5nma=15nma=20nma=25nm

0

20

40

60

80

100

120

9 29 49 69

Length (nm)

Num

ber o

f par

ticle

s abc

0

50

100

150

200

250

300

350

300 500 700 900 1100Wavelength (nm)

Aver

aged

ext

inct

ion

cros

s-se

ctio

n (n

m²)

abc

(a)

(b)

(c)

Fig. 12. Numerical simulations of the spectral variations of the extinction cross-section of ellipsoidal nanoparticles. (a) Single NP of different lengths 2a. (b)Theoretical length histograms of Gaussian shape used to calculate the averagedspectra shown in (c). (c) Theoretical averaged spectra of ensembles of nanoparticlesof different lengths (2a) and fixed width and thickness. The nanoparticle half-widthb and half-thickness c are 4.5 nm and 2.5 nm, respectively.

50 R. Sayah et al. / Microporous and Mesoporous Materials 139 (2011) 45–51

when they have a small aspect ratio. Our nanorods may have a highaspect ratio when their length increases, but this theory gives agood qualitative description of the spectra evolution as a functionof the nanorod length and allows to validate the conclusions previ-ously drawn from the experimental results. According to the Ganstheory, the extinction cross-section rext, can be written as a func-tion of the incident wavelength k, the dielectric functions of themetallic nanoparticle [23] em and of the host matrix e as follows:

rext;i ¼8p2abc

3kImðemÞe

½eð1� LiÞ þ LiRðemÞ�2 þ ½LiImðemÞ�2i ¼ x; y; z ð1Þ

when the particle is excited by an electric field parallel to the x, y orz axes, respectively. For small nanoparticles, this extinction cross-section is nearly equal to the absorption cross-section. The geomet-rical factors Li, calculated along each direction i, depend only on theparameters a, b, c of the ellipsoid and are given by Eq. (2):

Li ¼abc2

Z 1

0

dq

ðs2i þ qÞ½ðqþ a2Þðqþ b2Þðqþ c2Þ�1=2

ð2Þ

in which si equals a, b or c depending on the orientation of the inci-dent electric field. In order to take into account the fact that thethree axes of the nanorods can be excited by the TM incident fieldimpinging on a sample under oblique incidence, we estimate theextinction cross-section by calculating the sum rext,x + rext,y + rext,z

(Fig. 12a). These simulated data show that the third SPR band (thewidest located around 900 nm) rapidly shifts to the red when thenanorod length increases, whereas the two left-hand side bands (lo-cated around 420 nm and 350 nm, respectively) slightly shift to theUV. In order to better estimate the experimental spectra, we calcu-late the weighted average of the spectra obtained by assuming aGaussian length distribution for the particles (Fig. 12b). ThreeGaussian distributions are assumed: (a) is centered on 15 nm andhas a half-width at maximum/e of 20 nm, (b) is centered on30 nm and has a half-width at maximum/e of 20 nm and (c) is cen-tered on 15 nm and has a half-width at maximum/e of 30 nm. Thesespectra show a broad third band whose amplitude is strongly atten-uated compared to that of the two first ones whose shape does notchange a lot with the weighted average. The third band shifts to thered when the length distribution shifts to longer values and whenthe length distribution broadens. It can be noted that the two firstbands are thinner than the experimental ones. The broader experi-mental bands probably also result from a slight dispersion in thenanorods width and thickness. Even if the Gans theory is theoreti-cally not suitable for such long ellipsoid, the last spectra are in goodagreement with the experimental ones. These simulations confirm

R. Sayah et al. / Microporous and Mesoporous Materials 139 (2011) 45–51 51

the growth of flattened nanorods of various lengths in the filmmesostructure.

4. Conclusion

In summary, the paper describes a way to grow silver nanorodsinside SBA-15 mesoporous silica thin films. Our approach is basedon the soaking of the calcined hexagonal mesoporous silica films ina silver precursor solution and their thermal treatment to decom-pose the salt and form metal particles. The experimental condi-tions were varied to obtain the filling in of pore channels of themesoporous silica structure with nanorods. The soaking timestrongly influences the nanorods length; the presence of a base(NH3�H2O) increases the formation of metallic nanostructures in-side the mesoporous framework, as well as the use of EtOH/H2Osolvent; silver nanorods start growing from 100 �C, their densityincreases with the heat-treatment temperature until 200 �C and,at higher temperatures, oxidation of the silver nanoparticles oc-curs. Computational simulations of the extinction cross section of3-axes ellipsoidal particles confirm the presence of flattened silvernanorods in the film through the presence of three distinct SPRbands, which can be observed by UV–visible absorption spectro-copy under oblique incidence. These colored Ag–SiO2 nanocompos-ites described in this work are expected to find importantapplications for the fabrication of sensors. They could used aspolarizers before being tested as chromatic filter.

Acknowledgment

The authors thank Dr. F. Chassagneux from the Laboratoire Mul-timatériaux et Interfaces (University of Lyon I, France) and Dr. S.Sao-Joao from Ecole Nationale Supérieure des Mines de Saint-Eti-enne (France) for gathering the TEM data as well as for fruitfull dis-cussions. They are grateful to I. Anselme from Microscopy Center(University of Lyon, University Jean Monnet, Saint-Etienne, France)and Y. Lefkir from Laboratoire Hubert Curien (University Jean Mon-net) for some TEM images. This work was carried out in the frame-work of the POMESCO project supported by the Agence Nationalede la Recherche.

References

[1] L. Suber, I. Sondi, E. Matijevic, D.V. Goia, J. Colloid Interface Sci. 288 (2005) 489;M. Ullah, K. Il, C.-S. Ha, Mater. Lett. 60 (2006) 1496;K. Mallick, M. Witcomb, M. Scurrell, Mater. Chem. Phys. 97 (2006) 283.

[2] S. Effrima, H. Metiu, J. Chem. Phys. 70 (1979) 1602;P. Aravind, H. Metiu, Chem. Phys. Lett. 74 (1980) 301;J. Gersten, A. Nitzan, J. Chem.Phys. 73 (1980) 3023;D.-S. Wang, H. Chew, M. Kerker, Appl. Opt. 19 (1980) 2256;C. Mirkin, M. Ratner, Annu. Rev. Phys. Chem. 101 (1997) 1593;M. Rampi, O. Schueller, G. Whitesides, Appl. Phys. Lett. 72 (1998) 1781.

[3] R. Mulukutlz, K. Asakura, T. Kogure, S. Namba, Y. Iwasawa, Phys. Chem. Chem.Phys. 1 (1999) 2027.

[4] M.H. Haung, A. Choudrey, P. Yang, Chem. Commun. (2000) 1063.[5] V. Logvinenko1, O. Polunina, Yu. Mikhailov, K. Mikhailov, B. Bokhonov, J.

Therm. Anal. Cal. 90 (2007) 813.[6] H. Liu, X. Ge, Y. Ni, Radiat. Phys. Chem. 61 (2001) 89.[7] J. Creighton, C. Blatchford, M. Albrecht, J. Chem. Soc., Perkin Trans. 2, 75 (1979)

790.[8] K. Caswell, C. Bender, C. Murphy, Nano Lett. 3 (2003) 667.[9] K. Velikov, G. Zegeres, A. van Blaaderen, Langmuir 19 (2003) 1384.

[10] U. Nickel, A. Castell, K. Pöppl, N. Shirtcliffe, Langmuir 16 (2000) 9087.[11] G. Rodríguez-Gattorno, D. Díaz, L. Rendón, G. Hermández-Segura, J. Phys.

Chem. B 106 (2002) 2482.[12] P. Lee, D. Meisel, J. Phys. Chem. 86 (1982) 3391.[13] D. Van Hyning, C. Zukoski, Langmuir 14 (1998) 7034.[14] D. Zhao, Q. Huo, J. Feng, B.F. Chmelka, G.D. Stucky, J. Am. Chem. Soc. 120 (1998)

6020;D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Frederickson, B.F. Chmelka, G.D.Stucky, Science 279 (1998) 548;C. Kresge, M. Leonowicz, W. Roth, J.C. Vartuli, J. Beck, Nature 359 (1992) 710.

[15] D. Grosso, J. Galo, F. Babonneau, C. Sanchez, P.-A. Albouy, A. Bruneau, A.Balkenende, A. Ruud, Adv. Mater. 13 (2001) 1085;E. Crepaldi, J. Galo de Soler-Illia, D. Grosso, F. Cagnol, F. Ribot, C. Sanchez, J. Am.Chem. Soc. 125 (2003) 9770.

[16] L. Bois, F. Bessueille, F. Chassagneux, Y. Battie, N. Destouches, C. Hubert, A.Boukenter, S. Parola, Eng. Aspects 325 (2008) 86;Y. Battie, N. Destouches, L. Bois, F. Chassagneux, N. Moncoffre, N. Toulhoat, D.Jamon, Y. Ouerdane, S. Parola, A. Boukenter, J. Nanopart. Res. 12 (2010) 1073.

[17] P. Alberius, K. Frindell, R. Hayward, E. Kramer, G. Stucky, B. Chmelka, Chem.Mater. 14 (2002) 3284.

[18] D. Zhao, P. Yang, N. Melosh, J. Feng, B.F. Chmelka, G.D. Stucky, Adv. Mater. 10(1998) 1380.

[19] M. Huang, A. Choudrey, P. Yang, Chem. Commun. (2000) 1063.[20] W. Li, S. Seal, E. Megan, J. Ramsdell, K. Scammon, G. Lelong, L. Lachal, K.A.

Richardson, J. Appl. Phys. 93 (2003) 9553.[21] L. Bois, F. Chassagneux, S. Parola, F. Bessueille, Y. Battie, N. Destouches, A.

Boukenter, N. Moncoffre, N. Toulhoat, J. Solid State Chem. 182 (2009) 1700.[22] C.F. Bohren, D.R. Huffman, Absorption and Scattering of Light by Small

Particles, Wiley-Interscience, New York, 1983.[23] E.D. Palik., Handbook of Optical Constants of Solids, vol. I & II., Academic Press,

1985/1991.