- Email: [email protected]
Pulsed laser deposition of nanostructured Ag films
Applied Surface Science 252 (2006) 4445–4448 www.elsevier.com/locate/apsusc
Pulsed laser deposition of nanostructured Ag films Tony Donnelly, Brendan Doggett, James G. Lunney * School of Physics, Trinity College, Dublin 2, Ireland Received 3 May 2005; accepted 11 June 2005
Abstract Ultra-thin (0.5–5 nm) films of Ag have been prepared by pulsed laser deposition in vacuum using a 26 ns KrF excimer laser at 1 J cm2. The deposition was controlled using a Langmuir ion probe and a quartz crystal thickness monitor. Transmission electron microscopy showed that the films are not continuous, but are structured on nanometer size scales. Optical absorption spectra showed the expected surface plasmon resonance feature, which shifted to longer wavelength and increased in strength as the equivalent film thickness was increased. It is shown that Maxwell Garnett effective medium theory can be used to calculate the main features of optical absorption spectra. # 2005 Elsevier B.V. All rights reserved. Keywords: Pulsed laser deposition; Nanostructured Ag films; Maxwell Garnett theory
1. Introduction There is much current interest in the preparation and characterisation of nanostructured metal films with feature sizes in the range 1–100 nm. This interest is due to the many novel properties (optical, magnetic and catalytic) which arise when the dimensions of a solid material are reduced to the point where the particle contains from a few to thousands of atoms [1,2]. Many of the techniques used for thin film deposition, such as thermal evaporation, sputtering, ion implantation, chemical vapour deposition and pulsed laser deposition (PLD), can be adapted for the preparation of nanostructured thin films. In particular, it has been shown that PLD is a relatively simple and effective nano-fabrication technique. Alfonso et al. [3,4] used PLD with a nanosecond laser, both in vacuum and a gaseous atmosphere, to prepare nanocomposite films of Cu in amorphous alumina. By examining the areal density dependence of the morphology, they concluded that nucleation and growth of the metal nanoparticles is dominated by processes occurring on the substrate rather than in the gas phase. Similarly, Dolbec et al. [5] used a nanosecond laser to investigate PLD of Pt nanoparticles on pyrolitic graphite. They showed that there is a power law dependence of the particle diameter on the nominal thickness and that the shape of the nanoparticles
* Corresponding author. Tel.: +353 1 608 1259; fax: +353 1 671 1759. E-mail address: [email protected] (J.G. Lunney). 0169-4332/$ – see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2005.06.048
depends on the energy of the ablated species. Femtosecond laser pulses have also been used for PLD of nanoparticles. Nanoparticles have been observed in the plume produced by femtosecond ablation [6]; and it is believed that these nanoparticles are formed by supercritical phase decomposition [7,8]. Thus it would seem that, for PLD in vacuum, femtosecond ablation can lead to nanoparticle formation in the ablation plume but nanosecond ablation does not. The lack of nanoparticles in a nanosecond ablation plume seems to be primarily due to laser absorption and heating of the ablated material, much of which is evaporated during the laser pulse. However, it is possible to promote nanoparticle formation in nanosecond ablation plume if a gas is used to confine the expansion of the plume to allow sufficient time for nucleation [9]. In this paper we report the results of a systematic investigation of PLD of nanostructured films of Ag in vacuum. We have sought to correlate the optical absorption properties of the films with transmission electron microscopy (TEM) images of the film morphology. In particular, we have explored how the morphology and optical properties vary with equivalent thickness (areal density divided by solid density), which is controlled by the laser shot number and a quartz crystal monitor. The main feature in the visible/UV optical response of a nanoparticle metal film is an absorption feature due to the surface plasmon resonance (SPR) which is a collective oscillation of free electrons confined to the metal particle. The wavelength and strength of the SPR depends on the bulk dielectric properties of both the nanoparticle and host materials,
4446
T. Donnelly et al. / Applied Surface Science 252 (2006) 4445–4448
but is also sensitive to the size distribution and shape of the nanoparticles [10]. The theoretical description of the optical response of nanostructured metals can be traced back for about 100 years to the work of Mie and Maxwell Garnett. Mie theory considers the interaction of an electromagnetic wave with a homogeneous spherical particle placed in host material where the response to the field of both the particle and the host can be described using bulk dielectric functions [11]. The theory can be extended to nonspherical particles. Effective medium theories, such as that of Maxwell Garnett and Bruggeman, offer a simpler method for the computation of the optical properties of nanostructured metal films; and for this paper we have chosen to use Maxwell Garnett theory. Fig. 1. Ion time of flight signal recorded using a planar ion probe placed beside the substrate. The inset shows the corresponding ion energy distribution. The mean ion energy hEi is 73 eV.
2. Experimental setup The nanostructured films were prepared using a KrF excimer laser (248 nm, 26 ns, 10 Hz) to ablate a rotating Ag target in a vacuum chamber at a pressure of 4 105 mbar (4 m Pa). Part of the laser beam was imaged onto the target, at an incident angle of 458, to give a spot of 0.22 0.06 cm2 and an average fluence of 1 J cm2. Films were deposited on polymer and glass substrates placed 9.5 cm directly in front of the target. The polymer substrates were 150 nm thick Formvar films on Cu TEM support grids. A planar Langmuir ion probe of area 0.18 cm2, biased at 32 V, was placed beside the substrate facing the target to measure the flux and energy distribution of the ion flow in the ablation plume. In addition the ion probe also ensures reproducibility of plasma parameters. A quartz crystal monitor was used to measure the equivalent thickness of the deposited material. The number of laser shots per deposition was varied between 1000 and 10,000 to change the equivalent thickness. The morphology of the deposits on the polymer films was examined by TEM. Optical transmission spectra of films on polymer and glass were measured with a dual beam UV–vis spectrophotometer. An uncoated substrate was placed in the reference arm of the spectrophotometer. 3. Results and discussions The pulse of ion current measured by the negatively biased Langmuir probe [12] at 9.5 cm in front of the target is shown in Fig. 1. To good approximation, the ion velocity v is given by the probe-target distance d divided by the time-of-flight t. The peak of the ion signal corresponds to an ion energy of 95 eV. The complete ion energy distribution function dN/dE is obtained from the ion current I(t) using: dN t3 IðtÞ ¼ dE Aemd 2
(1)
where m the ion mass. The ion energy distribution is shown in the inset in Fig. 1; the average energy is 73 eV and the most probable energy is 22 eV. Thus it can be seen that the deposition plasma is sufficiently energetic to cause significant self-sputtering of the growing film [13–15]. The deposition per laser
shot measured by the quartz crystal monitor was 6 104 nm (or 3 1016 atoms m2). Integrating the ion signal gives an ion dose of 2.3 1016 ions m2 per shot. Thus it can be seen that the PLD process is a highly energetic deposition process similar to that of Dolbec et al. [5]. Fig. 2 shows TEM images of silver films deposited on Formvar with 1000, 3000 and 10,000 laser shots where the equivalent thicknesses are 0.53 nm, 1.37 nm and 4.4 nm, respectively. The 0.53 nm and 1.37 nm films show wellseparated particles that have diameters less than 7 nm. In the 4.4 nm film it can be seen that some of the particles have coalesced to form oblong structures. Fig. 2 also shows the absorbance spectra of these films; the absorption feature due to the SPR is clearly observed. Fig. 3 shows the absorbance spectra of films prepared under the same conditions but on glass instead of Formvar. Again the SPR is clearly observed and the magnitude of the peak absorption is similar to the value for deposition on Formvar, though there are some differences in the detailed structure of the two sets of absorption spectra, perhaps arising from some difference in the nature of surface diffusion on glass and Formvar. In particular it can be seen that the 4.4 nm film on Formvar shows two peaks at 465 nm and 580 nm, while only a single peak at 490 nm is observed for the deposition on glass. The second peak at 580 nm for the films on Formvar is probably due to the appearance of the oblong structures which show different resonant frequencies associated with the E-field along short and long dimensions [16]. For both Formvar and glass substrates the wavelength and absorbance of the SPR increases as the equivalent film thickness is increased. The values of peak absorbance are nearly the same for the two substrates. However, the peak wavelength of the SPR is slightly higher for films on glass as compared to Formvar. This indicates some difference in the growth mechanisms between the two cases. It is useful to explore to what extent it is possible to obtain a quantitative theoretical description of the absorption spectra of nanostructured metal films. Of the wide variety of theoretical methods which are available we have chosen to use the effective medium theory of Maxwell Garnett [11]. This formalism is more
T. Donnelly et al. / Applied Surface Science 252 (2006) 4445–4448
4447
Fig. 2. TEM images of nanostructured Ag films on Formvar for different equivalent thicknesses: (a) 0.53 nm; (b) 1.37 nm and (c) 4.4 nm. The dark regions correspond to Ag. Also shown are the corresponding absorbance spectra.
appropriate than the Bruggeman formalism when the volume fraction of one component is significantly less than the other. The Maxwell Garnett expression for the average dielectric function eeff of a composite medium is given by Eq. (2) eeff ¼ em
Fig. 3. Absorbance spectra of silver films on glass for various equivalent thicknesses.
3 f ððe em Þ=ðe þ 2em ÞÞ 1þ 1 f ððe em Þ=ðe þ 2em ÞÞ
(2)
where em is the dielectric function of the host medium, e the bulk dielectric function for the nanoparticle material and f is the volume fraction of nanoparticles in the composite film [11]. The factor 2 is appropriate for spherical particles. The effective dielectric function was used to find the effective refractive index of the composite medium and the absorbance was
4448
T. Donnelly et al. / Applied Surface Science 252 (2006) 4445–4448
4. Conclusion In conclusion we have used PLD to prepare nanostructured Ag films and investigated how the film morphology and optical absorption spectrum depends on equivalent thickness. TEM showed that the films are composed of nanoparticles of diameter less than 7 nm, though these particles begin to coalesce when the equivalent thickness is increased above about 4 nm. The expected SPR was clearly observed; the wavelength and strength increased as the equivalent thickness was increased. It was also shown that Maxwell Garnett effective medium theory may be used to calculate the main features of the optical absorption of these nanostructured Ag films. Fig. 4. Calculated absorbance of a 1.37 nm equivalent thickness Ag film on Formvar for different volume fractions.
calculated using the expression for the optical transmission of a thin absorbing film on a transparent substrate [17]. Fig. 4 shows a Maxwell Garnett calculation of the absorbance spectra for a film of spherical Ag particles where the equivalent thickness of fully dense Ag is 1.37 nm and the volume fraction is changed from 0.1 to 0.3. Following the work of Xu et al. [18] em was taken to be 1.625, the mean value of the dielectric constants of glass and air, to account for the fact that both glass and air constitute the host medium of the nanoparticles. The refractive index of the substrate was taken to be 1.5, and for comparison with the measurement the calculated transmission was divided by 0.96 to take account of the reflection loss at the rear surface of the substrate. The wavelength of maximum absorption moves to longer wavelength as the volume fraction is increased; a value of around f = 0.2 gives the same resonant wavelength as the 1.37 nm Ag film on polymer (Fig. 2b). The width of the calculated resonance is less than the measured value, which may due to variation of the Ag volume fraction through the depth of the nanoparticle film; near the substrate all particle sizes will contribute to the volume fraction while close to the top only the largest particles will do so. It should be noted that the Maxwell Garnett theory is only valid for f 0.5 [19] and the Bruggeman formalism may be more appropriate, particularly for films showing sign of coalescence. It will be of interest to compare the predictions of the Maxwell Garnett and Bruggeman formalisms and see which gives the better description of the optical absorption spectra of nanostructured Ag films.
Acknowledgements This work is supported by the PRTLI programme funded by the Higher Education Authority in Ireland and by the EU under the DESYGN IT project.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]
F. Rosei, J. Phys.: Condens. Matter 16 (2004) S1373–S1436. W. Eberhardt, Surf. Sci. 500 (2002) 242. C.N. Alfonso, et al. Appl. Surf. Sci. 127–129 (1998) 339–343. C.N. Alfonso, et al. Appl. Phys. A 69 (1999) S201–S207. S. Dolbec, et al. Phys. Rev. B 70 (2004) 201406. S. Amoruso, et al. Appl. Phys. Lett. 84 (22) (2004) 4502. S. Amoruso, et al. Phys. Rev. B 71 (3) (2005) 33406. S. Eliezer, et al. Phys. Rev. B 69 (2004) 144119. D.B. Geohegan, A.A. Puretzky, G. Duscher, S.J. Pennycook, Appl. Phys. Lett. 72 (23) (1998) 2987–2989. J.J. Mock, et al. J. Chem. Phys. 116 (15) (2002) 6755. C.F. Bohren, D.R. Huffman, Absorption and Scattering of Light by Small Particles, Wiley Interscience, 1998. B. Toftmann, J. Schou, T.H. Hansen, J.G. Lunney, Phys. Rev. Lett. 84 (17) (2000) 3998. S. Fahler, K. Strum, H.U. Krebs, Appl. Phys. Lett. 75 (24) (1999) 3768. R. Jordan, et al. Appl. Surf. Sci. 86 (1995) 24–28. P.C. Zalm, J. Vac. Sci. Technol. B 2 (2) (1984) 151. U. Kreibig, M. Vollmer, Optical Properties of Metal Clusters, vol. 25, Springer-Verlag, New York, 1995, pp. 306–307. O.S. Heavans, Optical Properties of Thin Solid Films, Dover, New York, 1965. G. Xu, et al. Appl. Phys. A 80 (2005) 1535–1540. T.C. Choy, Effective Medium Theory: Principles and Applications, Clarendon Press, Oxford, 1999.
Pulsed laser deposition of nanostructured Ag films Tony Donnelly, Brendan Doggett, James G. Lunney * School of Physics, Trinity College, Dublin 2, Ireland Received 3 May 2005; accepted 11 June 2005
Abstract Ultra-thin (0.5–5 nm) films of Ag have been prepared by pulsed laser deposition in vacuum using a 26 ns KrF excimer laser at 1 J cm2. The deposition was controlled using a Langmuir ion probe and a quartz crystal thickness monitor. Transmission electron microscopy showed that the films are not continuous, but are structured on nanometer size scales. Optical absorption spectra showed the expected surface plasmon resonance feature, which shifted to longer wavelength and increased in strength as the equivalent film thickness was increased. It is shown that Maxwell Garnett effective medium theory can be used to calculate the main features of optical absorption spectra. # 2005 Elsevier B.V. All rights reserved. Keywords: Pulsed laser deposition; Nanostructured Ag films; Maxwell Garnett theory
1. Introduction There is much current interest in the preparation and characterisation of nanostructured metal films with feature sizes in the range 1–100 nm. This interest is due to the many novel properties (optical, magnetic and catalytic) which arise when the dimensions of a solid material are reduced to the point where the particle contains from a few to thousands of atoms [1,2]. Many of the techniques used for thin film deposition, such as thermal evaporation, sputtering, ion implantation, chemical vapour deposition and pulsed laser deposition (PLD), can be adapted for the preparation of nanostructured thin films. In particular, it has been shown that PLD is a relatively simple and effective nano-fabrication technique. Alfonso et al. [3,4] used PLD with a nanosecond laser, both in vacuum and a gaseous atmosphere, to prepare nanocomposite films of Cu in amorphous alumina. By examining the areal density dependence of the morphology, they concluded that nucleation and growth of the metal nanoparticles is dominated by processes occurring on the substrate rather than in the gas phase. Similarly, Dolbec et al. [5] used a nanosecond laser to investigate PLD of Pt nanoparticles on pyrolitic graphite. They showed that there is a power law dependence of the particle diameter on the nominal thickness and that the shape of the nanoparticles
* Corresponding author. Tel.: +353 1 608 1259; fax: +353 1 671 1759. E-mail address: [email protected] (J.G. Lunney). 0169-4332/$ – see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2005.06.048
depends on the energy of the ablated species. Femtosecond laser pulses have also been used for PLD of nanoparticles. Nanoparticles have been observed in the plume produced by femtosecond ablation [6]; and it is believed that these nanoparticles are formed by supercritical phase decomposition [7,8]. Thus it would seem that, for PLD in vacuum, femtosecond ablation can lead to nanoparticle formation in the ablation plume but nanosecond ablation does not. The lack of nanoparticles in a nanosecond ablation plume seems to be primarily due to laser absorption and heating of the ablated material, much of which is evaporated during the laser pulse. However, it is possible to promote nanoparticle formation in nanosecond ablation plume if a gas is used to confine the expansion of the plume to allow sufficient time for nucleation [9]. In this paper we report the results of a systematic investigation of PLD of nanostructured films of Ag in vacuum. We have sought to correlate the optical absorption properties of the films with transmission electron microscopy (TEM) images of the film morphology. In particular, we have explored how the morphology and optical properties vary with equivalent thickness (areal density divided by solid density), which is controlled by the laser shot number and a quartz crystal monitor. The main feature in the visible/UV optical response of a nanoparticle metal film is an absorption feature due to the surface plasmon resonance (SPR) which is a collective oscillation of free electrons confined to the metal particle. The wavelength and strength of the SPR depends on the bulk dielectric properties of both the nanoparticle and host materials,
4446
T. Donnelly et al. / Applied Surface Science 252 (2006) 4445–4448
but is also sensitive to the size distribution and shape of the nanoparticles [10]. The theoretical description of the optical response of nanostructured metals can be traced back for about 100 years to the work of Mie and Maxwell Garnett. Mie theory considers the interaction of an electromagnetic wave with a homogeneous spherical particle placed in host material where the response to the field of both the particle and the host can be described using bulk dielectric functions [11]. The theory can be extended to nonspherical particles. Effective medium theories, such as that of Maxwell Garnett and Bruggeman, offer a simpler method for the computation of the optical properties of nanostructured metal films; and for this paper we have chosen to use Maxwell Garnett theory. Fig. 1. Ion time of flight signal recorded using a planar ion probe placed beside the substrate. The inset shows the corresponding ion energy distribution. The mean ion energy hEi is 73 eV.
2. Experimental setup The nanostructured films were prepared using a KrF excimer laser (248 nm, 26 ns, 10 Hz) to ablate a rotating Ag target in a vacuum chamber at a pressure of 4 105 mbar (4 m Pa). Part of the laser beam was imaged onto the target, at an incident angle of 458, to give a spot of 0.22 0.06 cm2 and an average fluence of 1 J cm2. Films were deposited on polymer and glass substrates placed 9.5 cm directly in front of the target. The polymer substrates were 150 nm thick Formvar films on Cu TEM support grids. A planar Langmuir ion probe of area 0.18 cm2, biased at 32 V, was placed beside the substrate facing the target to measure the flux and energy distribution of the ion flow in the ablation plume. In addition the ion probe also ensures reproducibility of plasma parameters. A quartz crystal monitor was used to measure the equivalent thickness of the deposited material. The number of laser shots per deposition was varied between 1000 and 10,000 to change the equivalent thickness. The morphology of the deposits on the polymer films was examined by TEM. Optical transmission spectra of films on polymer and glass were measured with a dual beam UV–vis spectrophotometer. An uncoated substrate was placed in the reference arm of the spectrophotometer. 3. Results and discussions The pulse of ion current measured by the negatively biased Langmuir probe [12] at 9.5 cm in front of the target is shown in Fig. 1. To good approximation, the ion velocity v is given by the probe-target distance d divided by the time-of-flight t. The peak of the ion signal corresponds to an ion energy of 95 eV. The complete ion energy distribution function dN/dE is obtained from the ion current I(t) using: dN t3 IðtÞ ¼ dE Aemd 2
(1)
where m the ion mass. The ion energy distribution is shown in the inset in Fig. 1; the average energy is 73 eV and the most probable energy is 22 eV. Thus it can be seen that the deposition plasma is sufficiently energetic to cause significant self-sputtering of the growing film [13–15]. The deposition per laser
shot measured by the quartz crystal monitor was 6 104 nm (or 3 1016 atoms m2). Integrating the ion signal gives an ion dose of 2.3 1016 ions m2 per shot. Thus it can be seen that the PLD process is a highly energetic deposition process similar to that of Dolbec et al. [5]. Fig. 2 shows TEM images of silver films deposited on Formvar with 1000, 3000 and 10,000 laser shots where the equivalent thicknesses are 0.53 nm, 1.37 nm and 4.4 nm, respectively. The 0.53 nm and 1.37 nm films show wellseparated particles that have diameters less than 7 nm. In the 4.4 nm film it can be seen that some of the particles have coalesced to form oblong structures. Fig. 2 also shows the absorbance spectra of these films; the absorption feature due to the SPR is clearly observed. Fig. 3 shows the absorbance spectra of films prepared under the same conditions but on glass instead of Formvar. Again the SPR is clearly observed and the magnitude of the peak absorption is similar to the value for deposition on Formvar, though there are some differences in the detailed structure of the two sets of absorption spectra, perhaps arising from some difference in the nature of surface diffusion on glass and Formvar. In particular it can be seen that the 4.4 nm film on Formvar shows two peaks at 465 nm and 580 nm, while only a single peak at 490 nm is observed for the deposition on glass. The second peak at 580 nm for the films on Formvar is probably due to the appearance of the oblong structures which show different resonant frequencies associated with the E-field along short and long dimensions [16]. For both Formvar and glass substrates the wavelength and absorbance of the SPR increases as the equivalent film thickness is increased. The values of peak absorbance are nearly the same for the two substrates. However, the peak wavelength of the SPR is slightly higher for films on glass as compared to Formvar. This indicates some difference in the growth mechanisms between the two cases. It is useful to explore to what extent it is possible to obtain a quantitative theoretical description of the absorption spectra of nanostructured metal films. Of the wide variety of theoretical methods which are available we have chosen to use the effective medium theory of Maxwell Garnett [11]. This formalism is more
T. Donnelly et al. / Applied Surface Science 252 (2006) 4445–4448
4447
Fig. 2. TEM images of nanostructured Ag films on Formvar for different equivalent thicknesses: (a) 0.53 nm; (b) 1.37 nm and (c) 4.4 nm. The dark regions correspond to Ag. Also shown are the corresponding absorbance spectra.
appropriate than the Bruggeman formalism when the volume fraction of one component is significantly less than the other. The Maxwell Garnett expression for the average dielectric function eeff of a composite medium is given by Eq. (2) eeff ¼ em
Fig. 3. Absorbance spectra of silver films on glass for various equivalent thicknesses.
3 f ððe em Þ=ðe þ 2em ÞÞ 1þ 1 f ððe em Þ=ðe þ 2em ÞÞ
(2)
where em is the dielectric function of the host medium, e the bulk dielectric function for the nanoparticle material and f is the volume fraction of nanoparticles in the composite film [11]. The factor 2 is appropriate for spherical particles. The effective dielectric function was used to find the effective refractive index of the composite medium and the absorbance was
4448
T. Donnelly et al. / Applied Surface Science 252 (2006) 4445–4448
4. Conclusion In conclusion we have used PLD to prepare nanostructured Ag films and investigated how the film morphology and optical absorption spectrum depends on equivalent thickness. TEM showed that the films are composed of nanoparticles of diameter less than 7 nm, though these particles begin to coalesce when the equivalent thickness is increased above about 4 nm. The expected SPR was clearly observed; the wavelength and strength increased as the equivalent thickness was increased. It was also shown that Maxwell Garnett effective medium theory may be used to calculate the main features of the optical absorption of these nanostructured Ag films. Fig. 4. Calculated absorbance of a 1.37 nm equivalent thickness Ag film on Formvar for different volume fractions.
calculated using the expression for the optical transmission of a thin absorbing film on a transparent substrate [17]. Fig. 4 shows a Maxwell Garnett calculation of the absorbance spectra for a film of spherical Ag particles where the equivalent thickness of fully dense Ag is 1.37 nm and the volume fraction is changed from 0.1 to 0.3. Following the work of Xu et al. [18] em was taken to be 1.625, the mean value of the dielectric constants of glass and air, to account for the fact that both glass and air constitute the host medium of the nanoparticles. The refractive index of the substrate was taken to be 1.5, and for comparison with the measurement the calculated transmission was divided by 0.96 to take account of the reflection loss at the rear surface of the substrate. The wavelength of maximum absorption moves to longer wavelength as the volume fraction is increased; a value of around f = 0.2 gives the same resonant wavelength as the 1.37 nm Ag film on polymer (Fig. 2b). The width of the calculated resonance is less than the measured value, which may due to variation of the Ag volume fraction through the depth of the nanoparticle film; near the substrate all particle sizes will contribute to the volume fraction while close to the top only the largest particles will do so. It should be noted that the Maxwell Garnett theory is only valid for f 0.5 [19] and the Bruggeman formalism may be more appropriate, particularly for films showing sign of coalescence. It will be of interest to compare the predictions of the Maxwell Garnett and Bruggeman formalisms and see which gives the better description of the optical absorption spectra of nanostructured Ag films.
Acknowledgements This work is supported by the PRTLI programme funded by the Higher Education Authority in Ireland and by the EU under the DESYGN IT project.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]
F. Rosei, J. Phys.: Condens. Matter 16 (2004) S1373–S1436. W. Eberhardt, Surf. Sci. 500 (2002) 242. C.N. Alfonso, et al. Appl. Surf. Sci. 127–129 (1998) 339–343. C.N. Alfonso, et al. Appl. Phys. A 69 (1999) S201–S207. S. Dolbec, et al. Phys. Rev. B 70 (2004) 201406. S. Amoruso, et al. Appl. Phys. Lett. 84 (22) (2004) 4502. S. Amoruso, et al. Phys. Rev. B 71 (3) (2005) 33406. S. Eliezer, et al. Phys. Rev. B 69 (2004) 144119. D.B. Geohegan, A.A. Puretzky, G. Duscher, S.J. Pennycook, Appl. Phys. Lett. 72 (23) (1998) 2987–2989. J.J. Mock, et al. J. Chem. Phys. 116 (15) (2002) 6755. C.F. Bohren, D.R. Huffman, Absorption and Scattering of Light by Small Particles, Wiley Interscience, 1998. B. Toftmann, J. Schou, T.H. Hansen, J.G. Lunney, Phys. Rev. Lett. 84 (17) (2000) 3998. S. Fahler, K. Strum, H.U. Krebs, Appl. Phys. Lett. 75 (24) (1999) 3768. R. Jordan, et al. Appl. Surf. Sci. 86 (1995) 24–28. P.C. Zalm, J. Vac. Sci. Technol. B 2 (2) (1984) 151. U. Kreibig, M. Vollmer, Optical Properties of Metal Clusters, vol. 25, Springer-Verlag, New York, 1995, pp. 306–307. O.S. Heavans, Optical Properties of Thin Solid Films, Dover, New York, 1965. G. Xu, et al. Appl. Phys. A 80 (2005) 1535–1540. T.C. Choy, Effective Medium Theory: Principles and Applications, Clarendon Press, Oxford, 1999.