Open Access
Add to BibSonomy
Abstract
Photo thermal refractive glasses of the type Na2O-SiO2-Al2O3-K2O-CaO-CaF2-ZnO doped with CeO2, Ag2O, SnO2, Sb2O3 were prepared using different concentrations of KBr. A UV irradiation followed by a thermal treatment leads to the formation of Ag-nanoparticles, indicated by the appearance of a plasmon resonance peak. This optical resonance position shifts with increasing KBr concentrations to higher wavelengths caused by the formation of an AgBr shell. The Mie theory with the aid of the optical dispersion of AgBr together with the measured dispersion of the used glasses was successfully applied to describe the optical relation of particle size and AgBr core shell thickness within the given glass. The results were compared with UV-vis-NIR spectroscopy.
© 2017 Optical Society of America
1. Introduction
Optical properties of glasses can be modified by the formation of metal colloids inside the glass matrix [1]. In particular, the photosensitization by metallic silver is of great interest for optical applications [2]. The nucleation and growth kinetics of silver particles in a photo process show the requirement of UV light and that even small amounts of cerium sensitize this process [3]. Stookey developed a photo-thermo-induced crystallization process in glass wafers already in 1949 [4]. These photo-thermo-refractive (PTR) glasses are usually silicate glasses doped with Ce, Ag, Sn, Sb and Br [5]. This was the beginning for modern volume phase hologram writing using refractive index modulation [6]. UV exposure leads to a redox reaction where Ce3+ is oxidized (Ce3+ + hυ → Ce4+ + e-). The released electrons are trapped by various trapping centers, e.g. Sn4+, Sb5+ and Ag+. The desired equilibrium of the photosensitivity is controlled by the Sb3+/Sb5+ and Sn2+/Sn4+ redox pairs. The Ag+ ions are finally reduced to neutral silver atoms Ag0 (Ag+ + e- → Ag0). During a subsequent heat treatment step slightly above Tg, the silver particles grow by aggregation of silver atoms. A certain temperature is necessary to reach a sufficiently large diffusion coefficient for a reasonable growth velocities. These Ag cluster act as nuclei for (earth) alkaline fluoride nano crystallization in a second heat treatment step at a higher temperature [5, 7, 8]. In the case that the Ag particles are small compared to the wavelength of light, the colour of the glass is determined by size, shape, volume fraction of the colloidal metal particles as well as the dispersion properties of the glass matrix [9, 10]. Also the optical properties of core-shell structures within the nano scale range are described in the literature [11]. The changes in the localized surface plasmon resonance are easy accessible by optical absorption spectroscopy.
In photo-thermo-refractive glass, the nucleation and growth of the silver particles is affected by the dosage of UV exposure and by the heat treatment temperature [12, 13]. This affects not only the shift of the surface plasmon resonance (SPR) band maximum, but also the full width at half maximum [12]. In conventional NaF PTR glass, a linear increase of the Ag absorption band position with the UV dosage is observed [13]. Moreover, the Ag band maximum shifts to longer wavelengths with increasing bromide concentration [14]. It is proposed, that this may be associated with the appearance of silver bromide particles shell around the Ag cluster which results in an additional shift of the plasmon resonance peak to longer wavelengths [15]. More likely is that the occurrence of a shell with high refractive index around the silver nanoparticles is responsible for this behavior [16]. The addition of low quantities of AgBr in a fluorophosphate glass matrix leads to a shift of the plasmon absorption band of silver nanoparticles after heat treatment to higher wavelengths. Using simulations according to Mie theory showed that a ~0.5 nm AgBr shell formed on the surface of the metallic Ag0n cluster might explain this behavior [17].
Glasses within the system Na2O/K2O/CaO/CaF2/Al2O3/ZnO/SiO2 were doped with Ce, Ag, Sn, Sb and Br which show photosensitivity and refractive index change after UV exposure and heat treatment due to the tailored nanoscale precipitation of CaF2 [8, 18]. This paper provides a study on the effect of KBr additions on the plasmonic resonance of metallic silver in the Na2O/K2O/CaO/CaF2/Al2O3/ZnO/SiO2 glass system.
2. Experimental part
Glasses with the mol% composition (100-x) (60.18% SiO2 • 12.12% CaO • 9.32% Na2O • 8.46% CaF2 • 5.44% K2O • 3.76% Al2O3 • 0.72% ZnO) (x KBr) with x = 0.0%, 1.0%, 2.0% were doped with 0.02% Ag2O, 0.02% CeO2, 0.02% SnO2, 0.04% Sb2O3 and melted from analytical grade raw materials in 300 g batches. The batch was melted in an inductive furnace for 4 h at a temperature of 1430 °C in a Pt crucible. Stirring was applied to homogenize the melt. After stirring, the melt was cast into a preheated steel mold. The obtained glass was placed in a muffle furnace which was then switched off to enable a slow cooling to room temperature (RT) with a rate of approximately 3 K/min. The glass was cut into pieces of 10 x 10 x 1 mm3 size and the surfaces were polished.
The final concentration of bromide was measured by X-ray fluorescence (XRF) analysis according to DIN 51001. Nominal Br concentrations and measured Br concentrations (wt%) are listed in Table 1.
Table 1. Comparison of nominal and measured Br concentration of the PTR glasses
UV exposure was carried out with a high pressure Xe arc lamp (LOT Oriel). The thermal treatments were performed in a muffle furnace (Nabertherm N11/H) applying a heating rate of 5 K/min.
Optical absorption spectra were recorded at RT using wavelengths from 200 to 1200 nm by a commercial double beam UV-vis-NIR spectrophotometer (3102PC, Shimadzu) with air as the reference; the error in the wavelength is about ± 1 nm.
Refractive indices were measured at 20 °C with a Pulfrich refractometer PR2 equipped with VoF5 prism at certain wavelengths according to the Fraunhofer lines (D, d, e, F, g, and h).
The dispersion of the glass was calculated using the Wemple equation [19].
where E is the light energy ( = hν), E0 is the energy of the effective dispersion oscillator and Ed is the dispersion energy. According to Eq. (1), a plot of measured refractive indices as 1/(n2-1) and the square of photon energy E2 enables the calculation of E0/Ed and 1/(E0Ed) from intercept and slope of a simple linear regression. Finally, the optical dispersion of the glass is available for a further adaption. The “MiePlot” software ver. 4.6 (supplied by Laven [20]) based on the BHMIE and BHCOAT algorithm [10] was used to calculate the optical behavior of Ag nano particles with different sizes in a glass matrix. The required dielectric function of Ag is based on data provided by Johnson and Christy [21]. The required optical dispersion for AgBr was taken from Schröter [22].The plasmon resonance wavelengths of the calculated and measured spectra were determined by fitting the peaks with a Gaussian function and locating of the center of gravity.
3. Results and discussion
The obtained glasses were homogeneous, colorless and transparent. Samples from the glass with the composition (mol%) 59.58 SiO2 • 12.0 CaO • 9.22 Na2O • 8.38 CaF2 • 5.39 K2O • 3.72 Al2O3 • 0.71 ZnO • 1.0 KBr doped with Ag2O, CeO2, Sb2O3 and SnO2 were irradiated and heat treated. The influence of the UV irradiation and heat treatment on the silver nano particle formation was investigated. Samples were UV irradiated at the same intensity for different times with a Xe lamp and were subsequent heat treated at 530 °C for 1 h. After heat treatment, the glass samples had a yellow colour due to the presence of colloidal silver nanoparticles. Optical absorption spectra exhibit the characteristic silver plasmon resonance. In previous studies, the photosensitivity of the Ag nano particle formation in this glass matrix has already been shown [8]. The formation of the Ag nanoparticles was studied in detail by high resolution TEM imaging [23]. The results showed that TEM studies in high resolution with high acceleration voltages or comparatively high currents, as needed for EDXS generate artefacts. In particular, an EDXS mapping was not possible due to the high damage of the electron beam. Furthermore, the evidence of Ag nanoparticles with electron microscopy techniques was not possible due to the resolution limit of the method which is 0.5 at%.
Furthermore, samples of the glass were irradiated and heat treated at 450, 480, 500, 515 and 530 °C for 1 h, respectively. The respective absorption band maximum wavelengths of the plasmon resonance peaks after irradiation and subsequent heat treatment are shown in Fig. 1.
Fig. 1 Comparison of Ag plasmon resonance absorption band maximum wavelengths for: ▼ samples irradiated (180 min, Xe lamp) and heat treated at temperatures of 450 and 530 °C for 1 h, ■ samples irradiated at respective time and heat treated at 530 °C for 1 h.
The position of the plasmon resonance maximum is shifted very slightly to longer wavelengths with increasing irradiation dosage (see Fig. 1). The overall shift of 2.3 nm is very small and is near to the limit of the error of the measurement. A shift of 2.3 nm corresponds to an increase in size of 3 nm at the mentioned wavelength range. Hence, at a certain value of irradiation dosage Ag particles with larger sizes are formed. The dosage of irradiation does affect the number of the formed Ag particles. The position of the Ag plasmon resonance peak maximum is shifted to higher wavelengths with increasing heat treatment temperature. The peak maximum is shifted from a wavelength at 407 nm to 423 nm. The Ag nanoclusters grow with increasing heat treatment temperature and change the plasmon resonance frequency. Moreover, the respective absorption band intensities vs. irradiation time are shown in Fig. 2. The plasmon resonance absorption band intensity is increasing with irradiation dosage up to a certain value and is then decreasing again. The slope of the absorption band intensity with irradiation time is linear. The Ag particles grow with increasing irradiation dosage. With higher UV dosage, more Ce3+ is reduced and consequently more electrons are released to reduce Ag+ to Ag0 which can agglomerate in a subsequent heat treatment. At a certain size of the Ag cluster the plasmon resonance intensity decreases and the band is broadened [1].
Fig. 2 Ag plasmon resonance absorption band intensity vs. irradiation time; samples irradiated and heat treated at 530 °C for 1 h.
As mentioned the position of the plasmon resonance peak is dependent on the particle size and the refraction index of the matrix surrounding it. The simulation of the Ag plasmon resonance peak position for an Ag particle for the glass matrices with the mol% composition (100-x) (60.18% SiO2•12.12% CaO•9.32% Na2O•8.46% CaF2•5.44% K2O•3.76% Al2O3•0.72% ZnO) (x KBr) with x = 0.0%, 1.0%, 2.0%, was performed and is shown in Fig. 3. The calculation was performed using the measured dispersion of the glasses containing different KBr concentrations.
Fig. 3 Simulated plasmon resonance peak positions for a single Ag particle in the glass matrix containing 0 mol %, 1 mol % or 2 mol % KBr, in the initial batch composition.
The calculated resonance wavelength increases from 407 nm for a particle of 10 nm diameter up to 457 nm for a particle of 60 nm in diameter. A comparison of the simulated plasmon resonance peak wavelengths with measured plasmon resonances (see Fig. 1), it is considered that Ag particles of 10 nm in diameter are grown at a heat treatment temperature of 450 °C and the Ag particle diameter increases up to 34 nm at a heat treatment temperature of 530 °C. For particles below around 10 nm, quantum confinement effects cause to an additional size dependence of complex dielectric function of the silver [24].
With increasing Br content, a difference in resonance wavelength has almost not occurred (see also the inset with higher magnification shown in the lower right part of the the image). Hence, the simulation of the plasmon resonance peak position shows that the plasmon resonance frequency is very sensitive to the particle size, but not to the Br concentration in the matrix. The influence of the bromide, especially in the measured concentrations, on the refractive index of the glass is much too small to cause a significant influence.
In Ref [14], it was shown that a long wavelength shift of the plasmon resonance is caused by the incorporation of bromide in the glass matrix. Hence in this study, samples were irradiated (180 min, Xe lamp) and heat treated at 530 °C for 1 h and optical absorption spectra were measured. Figure 4 (left) shows the optical absorption spectra of the native glass samples with different bromide concentrations. Furthermore, the respective absorption spectra of irradiated and heat treated samples are shown. The absorption spectra of the as melted glasses are similar for different bromide concentrations in the initial batch composition.
Fig. 4 left: UV-vis absorption spectra of native glass samples and respective irradiated, heat treated samples with different bromide concentrations; right: plasmon resonance peak wavelength position of irradiated and heat treated samples vs. measured bromide concentration (XRF- analyses) of respective samples.
After irradiation and heat treatment, the intensity of the absorption peak maximum at 310 nm, related to the Ce3+ [18], is decreasing with increasing bromide concentration. This is an effect of the matrix composition on the redox equlibrium of cerium ions [25]. The position of the silver plasmon resonance absorption peak is shifted to longer wavelengths, from 414 to 433 nm by adding bromide to the initial batch composition. The intensity of the plasmon resonance absorption maximum is decreasing with increasing bromide concentration.
Figure 4 (right) shows the plasmon resonance peak wavelength maximum vs. the measured bromide concentration of the respective samples. The measured bromide concentrations by XRF analyses confirm the high volatility of the bromide (see section 2,). But as already shown in Fig. 3, the influence of the measured Br content to the glass matrix cannot be responsible for the resonance shift. Hence, the shift of the plasmon resonance absorption to longer wavelengths by the addition of bromide indicates the formation of larger Ag particles or by the formation of an AgBr shell around the Ag particles. The formation of such a dielectric shell around silver nanoparticles with a high refractive index leads to a notable wavelength shift of the plasmon absorption band as proposed in Ref [16].
According to Ref [17]. where the formation of a dielectric shell of AgBr around the Ag particles is proposed, a simulation of the plasmon resonance peak positions was performed for a single Ag particle surrounded by a dielectric shell of AgBr. The hole core-shell structure was additionally embedded in the glass matrix described by the measured dispersion of the glass. The as calculated positions of plasmon resonance peak maxima for different initial Ag particle diameters are shown in Fig. 5.
Fig. 5 Simulated plasmon resonance peak positions for single Ag particle with different diameters surrounded by a dielectric AgBr shell of different thickness; in the glass matrix of the current system.
The positions are shown for particles with an Ag core with diameters of 15, 25, 35 and 45 nm and for different AgBr shell thickness in the range from 0.0 to 3.0 nm. The silver core diameter was kept constant and the AgBr shell was added in a certain thickness. Without a dielectric shell, the plasmon resonance peak maximum is, as expected, shifted to longer wavelengths with increasing Ag core diameter. With increasing thickness of the AgBr shell, the plasmonic resonance is shifted to longer wavelength. The slope of the band positions with increasing AgBr shell thickness is nearly the same for particles with a Ag core size of 25, 35 and 45 nm and is nearly linear. If the AgBr shell reaches a thickness of 2.0 nm, the positions of the plasmon resonance peak maxima are slightly different for particles with a Ag core of 25 and 35 nm. Moreover, the plasmon resonance peak maximum positions of a particle with a Ag core of only 15 nm in diameter shows a recognizable larger slope with increasing shell thickness probably caused by the ratio of core diameter to shell thickness. The plasmon resonance peak maximum position increases from a wavelength of 405 nm for an Ag particle without an AgBr shell to a wavelength of 470 nm for an Ag particle with a 3.0 nm AgBr shell and a core diameter of 15 nm. The high refractive index of AgBr (n = 2.2) leads for very thin shell dimension to the observed notable shift of the plasmon resonance peak position towards higher wavelengths.
Similar glass compositions as well as similar UV and thermal treatment processes leads to similar Ag particle sizes. Hence the strong shift caused by adding KBr to the glass composition, can only be explained by forming a dielectric shell around a Ag particle.
4. Conclusion
The photoinduced formation of Ag nanoparticles in an oxyfluoride glass was investigated. The number of the Ag particles is affected by the irradiation dosage. The heat treatment of irradiated samples lead to the occurrence of a plasmon resonance peak. This peak is shifted to longer wavelengths with increasing temperature of the subsequent thermal treatment. A simulation of the plasmon resonance peak position of a Ag particle with different diameters was performed for this particular glass matrix using Mie theory. Both, experimental and simulation results show, that the plasmon resonance frequency is very sensitive to the particle size. The addition of bromide to the initial batch composition has a significant effect on the spectral position of the plasmon resonance band. The simulation of the plasmon resonace peak position for Ag particles surrounded by a shell of AgBr show the most noticeable spectral red shift for small Ag particles and is further increased with increasing AgBr shell thickness.
Funding
Bundesministerium für Bildung und Forschung, Germany (Wachstumskern Brightlas, 03WKCF3E)
References and links
1. M. Quinten, Optical Properties of Nanoparticle Systems: Mie and Beyond (Wiley-VCH, 2011).
2. F. Gonella, “Silver doping of glasses,” Ceram. Int. 41(5), 6693–6701 (2015).
3. U. Kreibig, “Small Silver Particles in Photosensitive Glass - Their Nucleation and Growth,” Appl. Phys. (Berl.) 10(3), 255–264 (1976).
4. S. D. Stookey, “Photosensitive Glass - a New Photographic Medium,” Ind. Eng. Chem. 41(4), 856–861 (1949).
5. V. A. Borgman, L. B. Glebov, N. V. Nikonorov, G. T. Petrovskii, V. V. Savvin, and A. D. Tsvetkov, “Photothermorefractive Effect in Silicate-Glasses,” Dokl. Akad. Nauk SSSR 309(2), 336–339 (1989).
6. L. B. Glebov, N. V. Nikonorov, E. I. Panysheva, G. T. Petrovskii, V. V. Savvin, I. V. Tunimanova, and V. A. Tsekhomskii, “New Potentialities of Photosensitive Glasses for Volume Phase Hologram Recording,” Opt. Spektrosk. 73(2), 404–412 (1992).
7. L. B. Glebov, N. V. Nikonorov, Y. I. Panysheva, G. T. Petrovskii, V. V. Savvin, I. V. Tunimanova, and V. A. Tsekhomskii, “Multichromatic Glasses - a New Material for Recording of Volumetric Phase Holograms,” Dokl. Akad. Nauk SSSR 314(4), 849–853 (1990).
8. M. Stoica, G. N. B. M. de Macedo, and C. Rüssel, “Photo induced crystallization of CaF2 from a Na2O/K2O/CaO/CaF2/Al2O3/SiO2 glass,” Opt. Mater. Express 4(8), 1574–1585 (2014).
9. W. A. Weyl, Coloured Glasses (The Society of Glass Technology, 1951).
10. C. F. Bohren and D. R. Huffman, Absorption and scattering of light by small particles (Wiley, 1983).
11. D. Manikandan, S. Mohan, and K. G. M. Nair, “Annealing-induced metallic core-shell clusterization in soda-lime glass: an optical absorption study - experiment and theory,” Physica B 337(1–4), 64–68 (2003).
12. A. P. Nacharov, N. V. Nikonorov, A. I. Sidorov, and V. A. Tsekhomskii, “Influence of ultraviolet irradiation and heat treatment on the morphology of silver nanoparticles in photothermorefractive glasses,” Glass Phys. Chem. 34(6), 693–699 (2008).
13. J. Lumeau, L. Glebova, and L. B. Glebov, “Influence of UV-exposure on the crystallization and optical properties of photo-thermo-refractive glass,” J. Non-Cryst. Solids, 354(2-9), 425–430 (2008).
14. L. Glebova, J. Lumeau, M. Klimov, E. D. Zanotto, and L. B. Glebov, “Role of bromine on the thermal and optical properties of photo-thermo-refractive glass,” J. Non-Cryst. Solids, 354(2-9), 456–461 (2008).
15. J. Lumeau, A. Sinitskii, L. Glebova, L. B. Glebov, and E. D. Zanotto, “Spontaneous and photo-induced crystallisation of photo-thermo-refractive glass,” Phys. Chem. Glasses-B 48(4), 281–284 (2007).
16. N. V. Nikonorov, A. I. Sidorov, V. A. Tsekhomskii, and K. E. Lazareva, “Effect of a dielectric shell of a silver nanoparticle on the spectral position of the plasmon resonance of the nanoparticle in photochromic glass,” Opt. Spectrosc. 107(5), 705–707 (2009).
17. V. A. Aseev, P. A. Burdaev, E. V. Kolobkova, and N. V. Nikonorov, “Fluorophosphate glasses activated by rare-earth ions and AgBr,” Glass Phys. Chem. 38(4), 366–372 (2012).
18. M. Stoica, A. Herrmann, J. Hein, and C. Rüssel, “UV-vis spectroscopic studies of a CaF2 Photo-Thermo-Refractive Glass,” Opt. Mater. 62, 424–432 (2016).
19. S. H. Wemple, “Refractive-Index Behavior of Amorphous Semiconductors and Glasses,” Phys. Rev. B 7(8), 3767–3777 (1973).
20. P. Laven, “Simulation of rainbows, coronas and glories using Mie theory and the Debye series,” J. Quant. Spectrosc. Ra. 89(1–4), 257–269 (2004).
21. P. B. Johnson and R. W. Christy, “Optical Constants of Noble Metals,” Phys. Rev. B 6(12), 4370–4379 (1972).
22. H. Schröter, “On the refractive indices of some heavy-metal halides in the visible and calculation of interpolation formulas for dispersion,” Z. Phys. 67(1,2), 24–36 (1931).
23. M. Stoica, C. Patzig, C. Bocker, W. Wisniewski, M. Kracker, T. Höche, and C. Rüssel, “Structural evolution of CaF2 nanoparticles during the photoinduced crystallization of a Na2O-K2O-CaO-CaF2-Al2O3-ZnO-SiO2 glass,” J. Mater. Sci. 52(23), 13390–13401 (2017).
24. U. Kreibig and P. Zacharia, “Surface Plasma Resonances in Small Spherical Silver and Gold Particles,” Z. Phys. 231(2), 128–143 (1970).
25. E. I. Panysheva, I. V. Tunimanova, and V. A. Tsekhomskii, “The influence of the Matrix Composition of a Polychromatic Glass on its Properties,” Fiz. Khim. Stekla 17(6), 891–989 (1991).
