We have prepared optically birefringence materials consisting of an isotropic core of metal nanoparticle and an anisotropic shell of amorphous oxide. The sample shows an enhanced optical birefringence in a wavelength-selective way. The sample was prepared by depositing amorphous iron oxide thin films on top of the silver nanoparticles using the oblique deposition technique. This results in ellipsoidal shell of amorphous iron oxide surrounding a silver nanoparticle. The form birefringence appears because of the anisotropic shape of shells; the refractive index for the light polarized whose polarization is parallel to the elongation direction of ellipsoid is different from that for the light polarized perpendicularly. Moreover, the rotation of polarization plane is significantly enhanced at around the wavelength of localized surface plasmon resonance (LSPR). The difference in refractive index between two optical axes is as large as 0.34 for a 600 nm light, which is more than twice of typical birefringence crystal calcite (0.14 for visible light). It is speculated that the anisotropic shell induces the dependence of LSPR wavelength on the polarization direction of the incident light, which causes the polarization dependence of refractive index through the Kramers-Kronig relation.
© 2011 OSA
Optical birefringence refers to the difference in refractive index between different polarizations of light traveling through a material. The phenomenon is important to photonics because it can be utilized to control the polarization state of light. Some naturally available materials, e.g., calcite, rutile, and mica, show optical birefringence because of the anisotropic crystal structures. For example, calcite has a difference in refractive index between ordinary and extraordinary lays of 0.14 for visible light. Instead of anisotropy in crystal structures, it is possible to induce optical birefringence via anisotropy on a scale much larger than the atomic and molecular scale, i.e., form birefringence. Technological development of fabricating anisotropic nanostructures has allowed to produce materials that have a difference in refractive index larger than that of naturally available birefringence materials. For example, Muskens et al. reported that the vertically aligned nanowires of gallium phosphide (GaP) show a large difference between the in-plane and out-of-plane refractive indices of 0.8 . Also, Motoyoshi et al. developed a technique of glancing angle deposition that enables the production of highly porous and anisotropic thin films which exhibit significant form birefringence .
Metal nanostructures offer a route to modulate the polarization states of light in an enhanced way through the strong coupling between photons and surface plasmons. Experimentally, a polarization-dependent optical response has been observed in an array of elliptical metal nanoparticles on glass substrate [3,4], elongated nanoparticles dispersed inside glass [5–8], and elliptical or rectangular hole arrays on noble metal films [9,10]. Reyes-Esqueda et al. prepared elliptical silver (Ag) nanoparticles dispersed in glass matrix by irradiating spherical Ag nanoparticles embedded in glass with a silicon (Si) ion beam, which showed the refractive index difference as high as 0.15 at the wavelength of 532 nm. Sung et al. used the two-dimensional arrays of L-shaped Ag nanoparticles and found that the retardation between lays polarized along two orthogonal blanches of L-shaped Ag reaches to as high as 30° at the wavelength of 770 nm although the thickness of Ag structure is only 30 nm.
There are many interesting aspects in using the arrays of metal nanoparticle as birefringence materials, even though there are some drawbacks such as absorption. One advantage of using metal nanostructure is wavelength selectivity in modulation of light. The scattering strength of the metal nanoparticles attains the maximum at around the wavelength of localized surface plasmon resonance (LSPR), and their refractive indices are highly dispersive in this wavelength range. As a result, the nanoparticle arrays have a large birefringence within a particular wavelength range. Moreover, the resonance wavelength can be tuned in a systematic way by changing the size and surroundings of the nanoparticles.
In those metallic systems mentioned above, birefringence comes from the anisotropic metal nanostructures; because of the geometric anisotropy of metallic materials, the resonance wavelength of LSPR varies with the polarization direction of the incident light. This induces the variation in real part of refractive index via the Kramers-Kronig relation. The same phenomenon occurs when an isotropic metal sphere is embedded in an anisotropic matrix [11,12]. We have recently found several examples of this phenomenon; in a system consisting of Ag nanoparticles in glass matrix, the birefringence of matrix is enhanced by the metal nanoparticles at around the wavelength of LSPR, although the values of difference between refractive indices are less than 10−4.
In the present study, we have fabricated anisotropic nanostructures based on core-shell nanoparticles and examined the optical birefringence of the composites. The structure consists of an isotropic metallic core and an anisotropic shell, which was fabricated by oblique deposition of an amorphous oxide on top of the Ag nanoparticles that assembled beforehand on a substrate. The composite shows the birefringence because of the anisotropy in shell deposited obliquely. The difference in refractive index along two orthogonal principle axes (Δn) shows the maximum at around the wavelength of LSPR, and the value of Δn reaches to 0.34.
2.1 Sample preparation
Ag thin film (thickness ~30 nm) was grown on SiO2 glass substrates by using electron beam deposition. The film thus obtained was heat-treated at 300°C in air for 5 min to convert the Ag film into assemble of Ag nanoparticles . By using pulsed laser deposition (PLD), iron oxide thin films were grown on top of the Ag nanoparticles deposited on the SiO2 glass substrate. A KrF excimer laser (248 nm, 2 Hz) was focused on sintered α-Fe2O3 ceramic target at a fluence of 1.8 J/cm2. The substrate temperature was maintained at room temperature and the oxygen pressure was 1.0 × 10−4 Pa. The distance between the target and the substrate was 2.5 cm, and the typical growth rate of the film was 70 nm / h. The surface of the substrate (typical area 7.5 mm × 7.5 mm) was tilted with respect to the surface of target by an angle α, and the direction of tilt on the substrate that makes an angle α with respect to the holder, which was parallel to the target, was defined as t (see Fig. 1 ).
The optical extinction spectra were measured with a spectral photometer (JASCO V-570). Field emission scanning electron microscopy (FE-SEM; JEOL JSM-6700F) was utilized to evaluate the nanostructure of samples. X-ray diffraction (XRD) measurement was carried out to identify the crystalline phases of the Ag nanoparticles and an iron oxide thin layer on the substrate.
Optical birefringence was evaluated by use of the polarization rotation of linearly polarized light. A magneto-optical effect evaluation system (JASCO K-250) was utilized at zero magnetic field for the measurement. The setup is schematically shown in Fig. 2 . The sample was set on the rotation stage, and linearly polarized light with electric field oscillating along the vertical axis was normally incident on the sample surface with a square spot size of ~2 mm × 2 mm, and the polarization rotation angle ξ of transmitted light was detected by using a polarization modulation technique: The polarization plane of incident light was modulated by a frequency f and transmitted light was detected by a photomultiplier through a liner polarizer. An angle of the linear polarizer was tuned while the intensity of the transmitted light was monitored, and the angle which made the frequency of signal 2f was recorded as ξ. The sample was rotated around an axis normal to the surface, and the angle between t on the substrate and the oscillation direction of the incident electric field is defined as azimuth Φ.
Figure 3(a) shows the SEM image of the heat-treated Ag thin film deposited on the SiO2 glass substrate. Most of the Ag nanoparticles are spherical, and no directional orientation is observed for the arrangement of Ag nanoparticles . Figures 3(b) and 3(c) show the SEM images of the samples with α = 40° and 60°, respectively. Arrows in the images indicate the direction of tilt t during the deposition. Ellipsoidal nanoparticles can be observed, and the anisotropy is obvious especially when α is 60°; compared to the structure with α = 40°, the ellipsoids become more elongated along the direction parallel to t. XRD measurement confirms the amorphous nature of the iron oxide layer (not shown). Figure 3(d) shows the SEM image of the iron oxide film deposited on a bare SiO2 substrate at a tilt angle of α = 60°. No ellipsoidal structure can be observed, indicating that the deposition of Ag nanoparticles is necessary to develop ellipsoidal structures.
Figure 4 shows the typical absorption spectrum of the sample (α = 60°) and that of the heat-treated Ag thin film. The heat-treated Ag film shows absorbance at around 420 nm, which corresponds to the LSPR of the Ag nanoparticles [Fig. 4(a)]. Deposition of iron oxide causes the red shift of the LSPR to around 600 nm due to the increment of refractive index surrounding Ag NPs (refractive index of crystalline Fe2O3 is 2.9 at a wavelength of 632.8 nm). Figure 4(b) shows the dependence of LSPR peak for the film with α = 60° on the polarization of incident light. Due to the anisotropic structure of the shell, the peak position of the LSPR is shifted depending on the polarization direction of incident linearly polarized light. When the polarization direction of incident light is perpendicular to t, the LSPR peak is 613 nm, while that is 587 nm for the polarization parallel to t. This indicates that the refractive index of the film is higher for the linearly polarized light with electric field oscillating perpendicular to t than for the light parallel to t. The difference in absorbance between two polarizations is also plotted in the same figure. The difference is the largest at 650 nm.
Figure 5 shows the typical results of the optical rotation. Both heat-treated Ag film and iron oxide film deposited on a bare SiO2 substrate show very small optical rotations as illustrated in Fig. 5(a). In contrast, when α = 0°, the sample shows optical rotation much larger than that shown in Fig. 5(a). It is also found that the optical rotation shows wavelength-dependence; the rotation is largely enhanced at around the wavelength of LSPR (indicated as dashed line in the same figure). The enhancement becomes larger when α = 60°, as shown in Fig. 5(c). The difference in the magnitude of ordinate-scale between Figs. 5(b) and 5(c) should be noted. It is also shown in Figs. 5(b) and (c) that the optical rotation ξ is changed with a variation of azimuth Φ. We also prepared the samples by using titanium dioxide instead of iron oxide as a dielectric shell. A similar optical rotation is observed, although the rotation angle is less than that observed for iron oxide.
Figure 5(d) illustrates the variation of ξ at λ = 650 nm as a function of Φ. The variation of ξ with Φ can be expressed by the relation14]. The results of fitting are also shown in Fig. 5(d), where calculated curves are superimposed on the data denoted by solid and open circles. The agreement between experimental data and calculated curve is reasonably good. The fact that the period of oscillation is 180° suggests that the phenomenon comes from the birefringence . ξ becomes zero at every 90°, where the polarization direction of incident light is parallel to the one of the principle axes of the sample. The value of A corresponds to a maximum angle of rotation, and is related to the value of Δn by 2A = 2πΔnd/λ (d: sample thickness). Figure 6(a) illustrates the dependence of A, deduced from the fit, on the value of α. The value of A is increased with the increment of α, reaches a maximum of 10.0 at 60°, and decreases with a further increment in α.
In contrast to the value of A, the value of B does not largely vary with Φ (within the range of ± 15° for all the samples). B refers to the azimuth of principle axes, i.e., when Φ = B, ξ = 0° and the direction of polarization of incident light is parallel and perpendicular to two orthogonal principle axes. Since we defined Φ = 0° when the electric field is parallel to t, the result indicates that the principle axes are nearly parallel and perpendicular to t.
Figure 6(b) shows the dependence of A on deposition time for a fixed α of 60°. The typical sample thickness is 100 nm for the deposition time of 60 min. Τhe value of Δn is also plotted in the right ordinate of the same figure. Δn is the largest ( = 0.34) when the deposition time is 30 min and monotonically decreases with an increase in deposition time, while the value of A attains a maximum when the deposition time is 60 min.
4.1 Formation of anisotropic structures by oblique deposition
The technique of controlling the surface morphology of thin films by depositing source materials at an oblique angle has been applied to many systems, and unique morphologies have been prepared including inclined columnar structures, zigzag columns, and spirals . They have potential applications in photonic crystals, magnetic storage media, and other optoelectronic devices. It is claimed that the growing process of columnar morphologies is mainly dominated by geometrical shadowing effect. Ablation of the target with an excimer laser creates a flux of ablated species called laser plume that impinging on the substrate. The species are bombarded on the surface of the substrate and diffuse to find the site for nucleation and growth, which creates a microscopic surface roughness. When the flux impinging on the surface has an angle with respect to the surface of the substrate, sites of a lower height receive less incoming flux due to the geometrical blocking, i.e., shadowing effect. As a consequence, the un-shadowed positions which are exposed to the incoming flux grow preferentially, leading to highly underdense columnar microstroctures. Substrate patterning prior to the deposition predetermines the nucleation sites and provides a means to control the nanostructures. For example, nanospheres such as SiO2 and polystylene colloids have been used as templates; preferential nucleation and growth of the ablated species on top of the sphere result in the growth of nanocolumns at predefined positions [15,16]. In the present case, we have used the random array of Ag nanoparticles as a template to control the nucleation and growth sites of iron oxide to make an elongated shell. In the absence of pre-patterning, no ellipsoidal structure appears as indicated by the SEM image [Fig. 3(d)]. The shadowing effect is obvious especially when the incident angle is large, since a larger α means a larger ratio of the shadowed area to the un-shadowed area on the surface. As a result, the SEM image of the sample with α = 60° [Fig. 3(c)] manifests the ellipsoidal structure much more elongated than that with α = 40° [Fig. 3(b)]. Also, a larger α means a smaller projection area of the substrate to the plane parallel to the target. This causes a less incoming flux impinging on the substrate, resulting in a thinner film.
4.2 The origin of wavelength-selective optical rotation
The anisotropic structure in a scale slightly smaller than the optical wavelength gives form birefringence. In the present samples, anisotropic shells surrounding Ag nanoparticles induce the form birefringence. The importance of anisotropic shell structure is evident since no rotation is observed for the heat-treated Ag thin film and the iron oxide thin film on a bare SiO2 substrate with α = 60° [Fig. 5(a)]. Also, XRD patterns clarify that the iron oxide layer is amorphous, which excludes the possibility that an anisotropy due to crystalline orientation of iron oxide causes the birefringence. In addition, strong wavelength dependence of optical rotation around the wavelength of LSPR indicates the great contribution of LSPR to this phenomenon. As shown in absorption spectra in Fig. 4(b), the LSPR wavelength for the polarization parallel to t is different from that perpendicular to t because of the anisotropic nature of the shell. The difference in absorption is directly related to the difference in real part of refractive index through the Kramers-Kronig relation. Consequently, the birefringence is developed in a wavelength-selective way as shown in Figs. 5(b) and 5(c).
Based on the above discussions, the results in Fig. 6 can be interpreted as follows. In Fig. 6(a), with the increment of α from 0°, A becomes larger until α = 60°, and then becomes smaller with further increase in α. The increase of A is ascribed to the development of anisotropic ellipsoids, and the decrease is primarily due to the reduction of the projection area which lowers the amount of iron oxide deposited on the substrate. The enhanced optical birefringence is observed even when α = 0°. We speculate that the incoming flux does not impinge on the substrate at normal incidence but has a component parallel to the substrate because of the relatively short distance between the target and substrate (2.5 cm), which would create the anisotropy of iron oxide shell. It is worth noting that the birefringence property depends on the shape of the plume during the deposition of iron oxide. When we defocus the laser to increase a spot size on the target, the plume becomes less directional. The resultant samples exhibit birefringence less than those shown here, i.e., the samples prepared by tightly focusing the laser. As shown in Fig. 6(b), the value of A shows a maximum when the deposition time is 60 min. Further deposition causes the decrease in A, indicating that a less anisotropic structure is developed when the deposition time is much longer. Namely, the environment around the metal particle loses its anisotropic property as the deposition layer becomes thicker.
To confirm the close relation between the wavelength of LSPR and that of enhanced optical rotation, we fabricated a film using a different heat treatment condition of Ag thin film to modify the LSPR wavelength. We heated the Ag film at 500 °C for 5 min, and iron oxide was deposited at α = 60 °. The SEM observation indicates that the mean size of the particle is 100 nm. This value is larger than the size of particle made by heating at 300 °C for 5 min (40 nm). As shown in Fig. 7(a) , the film shows the LSPR peak at 700 nm, which is red-shifted by 100 nm compared to the sample where the Ag film is heated at 300°C for 5 min [Fig. 4(b)]. The enhanced rotation appears at around 700 nm, corresponding to the wavelength of LSPR. We also plotted the dependence of ξ on Φ in Fig. 7(b), which shows the oscillation with a period of 180°. The results demonstrate that tuning of wavelength of birefringence is possible by controlling the wavelength of LSPR.
In the present work, we have prepared the nanostructured materials consisting of an isotropic core and an anisotropic shell by using an oblique deposition of the iron oxide on the top of the Ag nanoparticles, and observed the enhanced optical birefringence at around the wavelength of LSPR. Birefringence is induced via the anisotropic dielectric shell and is enhanced by the LSPR of Ag nanoparticles. The maximum value of Δn is 0.34, which is more than twice as large as that of calcite (= 0.14 for visible light). Also, the wavelength of the maximum rotation can be modified by tuning the wavelength of LSPR. We think that the combination of an isotropic metallic core with an elliptical shell is a good alternative to the oriented arrays of metallic ellipsoids as form birefringence materials showing a wavelength-selectivity, taking into account the simple and robust fabrication process. For applications the absorption by metals needs to be avoided. One possible way of minimizing the absorption is to use the samples at infrared region where the absorption is small and they still show birefringence.
This study was supported by the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) of Japan through the Grant-in-Aid for Young Scientist (B, No. 22760512).
References and links
1. O. L. Muskens, M. T. Borgstrom, E. P. A. M. Bakkers, and J. Gómez Rivas, “Giant optical birefringence in ensembles of semiconductor nanowires,” Appl. Phys. Lett. 89(23), 233117 (2006). [CrossRef]
3. J. Sung, M. Sukharev, E. M. Hicks, R. P. Van Duyne, T. Seideman, and K. G. Spears, “Nanoparticle spectroscopy: birefringence in two-dimensional arrays of L-shaped silver nanoparticles,” J. Phys. Chem. C 112(9), 3252–3260 (2008). [CrossRef]
6. A. Berger, “Prolate silver particles in glass surfaces,” J. Non-Cryst. Solids 163(2), 185–194 (1993). [CrossRef]
7. J. A. Reyes-Esqueda, C. Torres-Torres, J. C. Cheang-Wong, A. Crespo-Sosa, L. Rodríguez-Fernández, C. Noguez, and A. Oliver; “Large optical birefringence by anisotropic silver nanocomposites,” Opt. Express 16(2), 710–717 (2008). [CrossRef] [PubMed]
8. M. Kaempfe, G. Seifert, K.-J. Berg, H. Hofmeister, and H. Graener, “Polarization dependence of the permanent deformation of silver nanoparticles in glass by ultrashort laser pulses,” Eur. Phys. J. D 16(1), 237–240 (2001). [CrossRef]
9. J. Elliott, I. I. Smolyaninov, N. I. Zheludev, and A. V. Zayats, “Wavelength dependent birefringence of surface plasmon polaritonic crystals,” Phys. Rev. B 70(23), 233403 (2004). [CrossRef]
10. R. Gordon, A. G. Brolo, A. McKinnon, A. Rajora, B. Leathem, and K. L. Kavanagh, “Strong polarization in the optical transmission through elliptical nanohole arrays,” Phys. Rev. Lett. 92(3), 037401 (2004). [CrossRef] [PubMed]
11. S. Murai, R. Hattori, T. Matoba, K. Fujita, and K. Tanaka, “Enhancement of optical birefringence in tellurite glasses containing silver nanoparticles induced via thermal poling,” J. Non-Cryst. Solids 357(11-13), 2259–2263 (2011). [CrossRef]
12. S. Murai, R. Hattori, K. Fujita, and K. Tanaka, “Optical birefringence in tellurite glass containing silver nanoparticles precipitated through thermal process,” Appl. Phys. Express 2(10), 102001 (2009). [CrossRef]
13. H. Nakashima, H. Omoto, and H. Wakabayashi, “Formation of a random array of fine silver particles from a silver film: preparation of the frequency selective screen,” J. Appl. Phys. 95(12), 7790–7797 (2004). [CrossRef]
14. K. Konishi, T. Sugimoto, B. Bai, Y. Svirko, and M. Kuwata-Gonokami, “Effect of surface plasmon resonance on the optical activity of chiral metal nanogratings,” Opt. Express 15(15), 9575–9583 (2007). [CrossRef] [PubMed]
15. A. Dolatshahi-Pirouz, D. S. Sutherland, M. Foss, and F. Besenbacher, “Growth characteristics of inclined columns produced by glancing angle deposition (GLAD) and colloidal lithography,” Appl. Surf. Sci. 257(6), 2226–2230 (2011). [CrossRef]
16. S. V. Kesapragada and D. Gall, “Two-component nanopillar arrays grown by glancing angle deposition,” Thin Solid Films 494(1-2), 234–239 (2006). [CrossRef]