Silver nanoparticles embedded in polymer matrices - a FTIR-SE study
solidi
status
pss
physica
phys. stat. sol. (c) 5, No. 5, 1210 – 1214 (2008) / DOI 10.1002/pssc.200777840
c
www.pss-c.com
current topics in solid state physics
` 1 , Severine ´ ´ 2 , Michel Voue´ 1,* , and Joel De Coninck1 Nora Dahmouchene Coppee 1 2
Centre de Recherche en Mod´elisation Mol´eculaire, Universit´e de Mons-Hainaut, Place du Parc, 20, 7000 Mons, Belgium Laboratoire de Physico-Chimie des Polym`eres, Universit´e de Mons-Hainaut, Place du Parc, 20, 7000 Mons, Belgium
Received 11 June 2007, revised 29 October 2007, accepted 17 December 2007 Published online 18 March 2008 PACS 78.20.Ci, 78.67.Bf, 78.68.+m ∗
Corresponding author: e-mail
[email protected], Phone +32-65-373885, Fax +32-65-373881
The optical properties of silver nanoparticles embedded in a poly (vinyl) alcohol matrix have been investigated using Fourier transform infrared spectrocopic ellipsometry (600 and 6000 cm−1 ). An atomic force microscopy analysis showed that the particles were nearly spherical and that their average size was 20 nm. The extinction coefficient spectrum of the annealed films reveals the major peaks associated with polyvinyl alcohol, as well as two additionnal peaks located at 1134 and 1036 cm−1 , suggesting the presence of bonds between Ag and the polymer matrix.
AFM topographic image of the embedded silver nanoparticles (200x200 µm2 , Full scale in Z dimension: 4.77 nm). © 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1 Introduction In material science, composite materials and, in particular, insulating materials with embedded metal nanoparticles are under the focus because of their special structural properties and the specific optical behavior that these confer upon them. These properties, as well as the preparation of such materials have been reviewed by Heilmann [1] and by Murphy and co-workers [2]. Silver and gold nanoparticles are the most significant in researches. The former are responsible for the yellowish color of the stained glass windows of the gothic cathedrals of the Middle Ages, while the use of the latter results in a typically red color. The optical properties of such metal particles are dominated, in the visible spectral domain, by the surface plasmon resonance [3] whose position and width of the plasmon band depend on a variety of factors, such as the size and shape of the particles [4–6], and their dielectric environment. Polymers are excellent host materials for such particles because they prevent their agglomera-
tion and precipitation. Polymer films with embedded metal nanoparticles are generally prepared by coating a substrate by a dispersion of particles in the polymer solution [1] but an interesting alternative has been proposed by Longenberger and Mills [7]. It has recently been adapted by Porel and co-workers [8]: they considered the in situ production of the particles, starting from polymer films doped with silver ions. The formation of silver nanoparticles is controlled by the chemical reduction of the Ag+ cations and the electrons required by the reduction are provided by the polymer matrix. Poly(vinyl alcohol) or PVA has been widely used due its easy processability and high transmittance in the visible spectral domain [9] but also because simple alcohols can act as electron donors. They therefore promote the reduction of the Ag+ ions. The reduction is not fully effective at room temperature and an annealing procedure (at moderate temperature) of the doped films is required.
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Contributed Article phys. stat. sol. (c) 5, No. 5 (2008)
The optical properties of such composite materials have been widely studied by photometry in the visible or infrared spectral domains but the use of spectroscopic ellipsometry received todate only little attention [10,11]. In a recent article [12], we considered the optical properties of PVA films doped with silver nanoparticles. The effects of annealing time were investigated by spectroscopic ellipsometry (SE) in the visible spectral domain. The size of the silver nanoparticle particles, as determined from AFM topographic analysis, and their number respectively decreased and increased as the annealing time of the film increased. The complex refraction index N = n+ik of the composite film showed a localized absorption around 420 nm, which could be attributed to the collective oscillations of the electrons at the surface of the metallic nanoparticles. The intensity of this absorption was linearly correlated to the nearest-neighbor distance between the silver nanoparticles. In this article, we continue these investigations in the mid-infrared spectral domain and use infrared spectroscopic ellipsometry (FTIR-SE) to probe the effects of the film annealing on the molecular vibrations of the composite film. 2 Experimental Unless otherwise stated, all the chemicals and reagent were analytical grade and used as received without a further purification. Water used in the experiments was 18.2 MΩ MilliQ water filtrated on a 0.22 µm MilliPAK membrane (Millipore, USA). 2.1 Preparation of Ag-PVA nanocomposites The Ag-PVA nanocomposite films were prepared according to the work of Porel and coworkers [8]. The procedure is briefly summarized hereafter. AgNO3 (99.99%, Sigma) was added to an aqueous solution of poly (vinyl alcohol) (8% w:w , MW = 85000-24000, hydrolysis: 87-89%) at 80 ◦ C to obtain a silver concentration of 8.4%. The silvermodified polymer solution was spin-coated on 1mm-thick float glass pieces (6000 rpm - 15 s) to reach a final thickness for the dried film in the 1 - 2 µm range. The coatings were allowed to dry in open air atmosphere, before being annealed 60 minutes at 110 ◦ C. Due to the annealing, the coatings became yellowish-brown in color. These films will be referred to as the ’Ag-PVA films’. Pure PVA sample was prepared in the same manner, omitting the addition of the silver nitrate in the initial solution, to obtain reference films. 2.2 Characterisation of PVA-Ag nanocomposites The surface morphology of the films was studied with a Nanoscope III AFM (Digital Instruments, USA), operated in air in tapping mode with commercial 300 kHz tips. Areas of 200 x 200 nm2 were investigated for every sample. The measurements were realized on pure PVA sample and on Ag-PVA films (non-annealed and annealed at 90 ◦ C during 10, 30 and 60 minutes). The radius of curvature of the tips was typically 7–10 nm and the resolution of the images was 512 pixels x 512 pixels.
www.pss-c.com
1211
Spectroscopic ellipsometry (SE) is a powerful tool for the characterization of optical properties of thin films or bulk materials [15]. This method is commonly applied in process control or in the investigation of thin polymer films. Spectroscopic ellipsometry is based in the measurement of changes in the state of polarization of linearly polarized light which is reflected by a planar surface or a planar layered system. The complete information of the sample system is contained in the two measured ellipsometric parameters Ψ and ∆, that are defined by the complex ratio of the p− and s−polarized reflectance coefficients rp and rs (rp and rs denote parallel and perpendicular polarization to the plane of incidence, respectively): ρ=
rp = tan Ψ ei∆ . rs
(1)
The infrared measurements were made using a SOPRA GES5-FTIR spectroscopic ellipsometer. This system operates in the mid-IR spectral domain (550 - 7000 cm−1 ). The principles of FTIR-SE remains the same as the one of UV-visible SE [13–15]. The equipment is also similar but requires different optical components. The IR part of the instrument uses a commercial FTIR spectrometer with globar as source, grid polarizers and a liquid-nitrogen cooled MCT detector (Nicolet, USA). The two ellipsometric parameters tan Ψ and cos ∆ are calculated using Fourier analysis and 64 to 512 spectra are accumulated for each of the analyzers positions to improve signal noise ratio. The FTIR ellipsometer was operated at 70 ◦ of incidence and the spectral resolution was 4 cm−1 . The spectral domain was limited to 600-6000 cm−1 . According to the classical Lorentzian oscillator model, non-interacting oscillators were used to describe the molecular vibrations in the infrared range. The real and imaginary parts of the dielectric function are respectively given by r = ∞ +
j
i =
j
2 Fj (˜ νoj − ν˜2 ) , − ν˜2 )2 + Γj2 ν˜2
2 (˜ νoj
Fj Γj ν˜ , 2 −ν (˜ νoj ˜2 )2 + Γj2 ν˜2
(2)
(3)
where ν˜ denotes the wavenumbers. ν˜oj is the resonance wavenumber of the jth oscillator, Fj its oscillator strength and Γj its half-width. ∞ is the dielectric constant at high frequency, i.e. in the visible domain. A non-linear leastsquares regression analysis was performed to fit the experimental data to the model. The differences between the model and the experimental data were expressed by the mean square error (MSE). The calculations were carried out on the Fourier coefficient α = − cos 2Ψ and β = sin 2Ψ cos ∆ because their values are bounded by [-1,+1] and do not favor one of the two spectra in the nonlinear fitting procedures. The smaller the MSE, the better
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
solidi
status
physica
pss
c
1212
N. Dahmouchène et al.: Nanoparticles in polymer films
the agreement of experimental and model generated results. The optical calculations were carried out using the WINELLI analysis software (SOPRA, France).
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
400
200
0
0
9000
6000
3000
12000
Particle area (Å2) Figure 2 Perimeter versus area relation for the silver nanoparticles (Plain line: P = µAα , with log(µ) = 0.698 ± 0.055 and α = 0.529 ± 0.016; Dashed line : same with log(µ) = 0.797 ± 0.009 and an imposed value of α = 0.50).
cm−1 at an incidence angle Φ = 70 ◦ . The simulated data were calculated from a Cauchy optical model associated with independent Lorentzian oscillators (plain line) to describe the effects of the molecular vibrations in the infrared range. The calculated ellipsometric data are in good agreement with the experimental ones (χ2 10−3 ). The thickness of the pure PVA film and of the annealed Ag-PVA film is, respectively, 1.33 µm and 1.65 µm. The optical indices n and k are represented in Fig. 5. At high frequency, n tends to 1.5, in agreement with previously measured values. The extinction coefficient spectrum, calculated from the optimized ellipsometric data, is showed and it clearly reveals the major peaks associated with polyvinyl alcohol. Although the fine structure of the peaks is not entirely resolved because we wanted to limit the number of parameters in the optical model, the main
tan Ψ
3 Results and discussion 3.1 AFM characterization of the films The topography of pure PVA films as well as (non-)annealed films was analyzed by AFM. The images (200 x 200 nm2 ) were recorded in taping mode. The phase images are given in Fig. 1. Fig. 1A shows that the pure PVA film is homogeneous. Its root-mean-square roughness, rrms , is 0.23 nm. Let us now consider the effect of adding the Ag nanoparticles in the polymer matrix (Fig. 1B) and the effect of annealing the films at 60 ◦ C during 90 min (Fig. 1C). These two images show that the annealing of the film results in the onset of the silver nanoparticles. The roughness also increases: from 0.32 nm (non-annealed film) to 0.56 nm (annealed film). Although the curvature radius of the tip was 7–10 nm, it has only a small effect on the final size of the detected particles. For particles laying on a surface, this effect could not be neglected and the final image would have resulted from the convolution of the true image by the shape of the tip. In our case, the nanoparticles do not appear at the surface but in the volume of the polymer matrix and the convolution effect can be neglected. The low roughness values of the (non-)annealed films support this hypothesis. The characteritics of the particles can be extracted by a flooding method of the image. Neglecting the particles whose radius was less than 0.54 nm, the link between the particle area A and its perimeter P can be experimentally determined (Fig. 2). The perimeter scales as Aα , with 2α being the autosimilarity of the particle. A value of α = 0.53 ± 0.02 was found, suggesting that the particles do not have a fractal dimensionality. A further check can be performed by comparing the fitted curve with the curve generated with α = 0.53 (Fig. 2, dashed lines). Both curves cannot be distinguished from each other for particles whose area is less ˚ 2 . The average first neighbor distance between than 4000 A the particles is 17.9 nm. 3.2 Optical properties of the Ag-PVA films. Let us now consider the optical behavior of the pure PVA films (Fig. 3) and of the annealed Ag-PVA films (Fig. 4)in the mid-infrared spectral domain. The ellipsometric parameters tan Ψ and cos ∆ were measured between 600 and 6000
600
1.6
1
1.2
0.5
cos ∆
Figure 1 Phase images (200 x 200 nm2 ) of the films. A: pure PVA, B: Non-annealed Ag-PVA, C: Annealed Ag-PVA.
Particle perimeter (Å)
800
0.8
0
-0.5
0.4
-1
0
1000 2000 3000 4000 -1
Wavenumbers (cm )
1000 2000 3000 4000 -1
Wavenumbers (cm )
Figure 3 FTIR-SE data (symbols) and regression (plain line) for the pure PVA in the infrared range.
www.pss-c.com
Contributed Article phys. stat. sol. (c) 5, No. 5 (2008)
1213
0.5
cos ∆
1.2
tan Ψ
0.3
Pure PVA 30 min 45 min 60 min
(a)
(b1)
0.2
(b2)
0.1
0 1600
1400
1200
1000
800
Wavenumbers (cm-1)
Figure 6 Extinction coefficient k(λ) for the pure PVA and the annealed Ag-PVA films. Influence of the annealing time. (a) C-O stretching at 1100 cm−1 , (b1) and (b2) Additional peaks at 1036 and 1134 cm−1 .
1
1.6
0.8
0
-0.5
0.4
-1
0
1000 2000 3000 4000
1000 2000 3000 4000
Wavenumbers (cm-1)
Wavenumbers (cm-1)
Figure 4 FTIR-SE data (symbols) and regression (plain line) for the annealed Ag-PVA film (60 min at 110◦ ).
0.2
Extinction coefficient
1.7
Refractive index
0.4
Extinction coefficient k(λ)
absorption bands can be identified. At high energy, the spectrum is dominated by the broad OH stretching band around 3500 cm−1 and the CH stretching band, slightly below 3000 cm−1 . Intramolecular as well as intermolecular hydrogen bonds are expected to occur among PVA chains due to high hydrophilic forces. As PVA used in our experiments is obtained by incomplete (i.e. 87%) hydrolysis of polyvinyl acetate, we expect the residual carbonyl groups to appear in the extinction coefficient spectrum. The very sharp band at approximately 1720 cm−1 corresponds to the stretching of the C=O bond. At lower frequency, two bands clearly appear: the C-O stretching band ( 1100 cm−1 ) and the (unresolved) C-H bending and C-C stretching bands ( 1390 cm−1 ). This extinction coefficient k was obtained by describing the peaks related to the molecular vibrations as 6 Lorentzian oscillators in case of the pure PVA film. Two
1.6
1.5
4 Conclusion The complex index of refraction of poly-(vinyl) alcohol films containing dispersed silver nanoparticles has been investigated using FTIR-SE. The extinction coefficient spectrum of the annealed films reveals the major peaks associated with PVA, as well as two additionnal peaks located at 1134 and 1036 cm−1 . These suggest the presence of bonds between Ag clusters and PVA molecules.
0.1
0
1.4
1000
3000
5000 -1
Wavenumbers (cm )
1000
3000
5000 -1
Wavenumbers (cm )
Figure 5 Refractive index n and extinction coefficient k of the pure PVA films (dashed line) and of the annealed Ag-PVA films (plain lines).
www.pss-c.com
additional peaks were necessary to describe the optical behaviour of the annealed Ag-PVA films. These two peaks are respectively located at 1134 and 1036 cm−1 . This clearly suggests the presence of bonds between Ag and PVA. As those peaks are located in the fingerprint region, their precise modeling is far from being simple and a direct extraction of the optical properties is more suitable to study the effect of the annealing time. To that purpose, the ellipsometric data were processed as follows: the thickness of the films was determined from the data above 2500 cm−1 , as the C-H stretching and the O-H stretching can easily be calculated and require only a limited number of Lorentzian bands. Knowing the film thickness, the complex refractive index was calculated by direct extraction. These results are reported in Fig. 6 for the pure PVA and for the annealed Ag-PVA films. The bands at 1036 and 1134 cm−1 , respectively labelled (b1) and (b2), are present and do not vary much with the annealing time.
Acknowledgements This work was partially supported by the Minist`ere de la R´egion Wallonne and by the National Funds for Scientific Research of Belgium (FNRS). Thanks are due to C. Defranoux and O. Dulac (SOPRA) for helpful discussions.
References [1] A. Heilmann, Polymer films with embedded metal
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
solidi
status
physica
pss
c
1214
N. Dahmouchène et al.: Nanoparticles in polymer films
nanoparticles, Springer Series in Material Science (Springer, Berlin, 2002). [2] C. J. Murphy, T. K. San, A. M. Gole, C. J. Orendorff, J. X. Gao, L. Gou, S. E. Hunyadi, and T. Li, J. Phys. Chem. B 109, 13857 (2005). [3] U. Kreibig and M. Vollmer, Optical Properties of Metal Clusters (Springer, Berlin, 1995). [4] C. J. Murphy and N. R. Jana, Adv. Mater. 14, 80 (2002). [5] A. I. Kirkland, D. A. Jefferson, D. G. Duff, P. P. Edwards, I. Gameson, B. F. G. Johnson, and D. J. Smith, Proc. Roy. Soc. A - Math. Phys. Eng. Sci. 440, 589 (1993). [6] J. J. Mock, M. Barbic, D. R. Smith, D. A. Schultz, and S. Schultz, J. Chem. Phys. 116, 6755 (2002). [7] L. Longenberger and G. Mills, J. Phys. Chem. 99, 475 (1995). [8] S. Porel, S. Singh, S. S. Harsha, D. N. Rao, and T. P. Radhakrishnan, Chem. Mater. 17, 9 (2005). [9] J. E. Mark (Ed.), Polymer Data Handbook (Oxford University Press, Oxford, 1999). [10] K. C. See, J. B. Spicer, J. Brupbacher, D. J. Zhang, and T. G. Vargo, J. Phys. Chem. B 109, 2693 (2005). [11] T. W. H. Oates and E. Christalle, J. Phys. Chem. C 111, 182 (2007). [12] N. Dahmouch`ene, M. Vou´e, and J. De Coninck, Optical properties of poly (vinyl) alcohol films with embedded silver nanoparticles: a spectro-ellipsometry study, submitted to Langmuir (2007). [13] A. R¨oseler, Infrared Spectroscopic Ellipsometry (Akademie Verlag, Berlin, 1988). [14] M. Schubert, Infrared Ellipsometry on Semiconductor Layer Structures (Springer, Heidelberg, 2004). [15] H. G. Tompkins and E. A. Irene (eds.), Handbook of Ellipsometry (Springer, Heidelberg, 2005).
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.pss-c.com
