1. Introduction

A quantum of a charge-density wave of free electrons on a metal surface is called a surface-plasmon polariton (SPP), which is excited by coupling with light waves. SPPs can be applied to sensors [1], solar cells [2], plasmonic lasers [3,4], near-field optical microscopy [5], Raman spectroscopy with high spatial resolution [6], optical nanocircuits [7], photodynamic cancer cell treatments [8], and nanolens magnifiers [9]. In some of these applications, the resonant-frequency dependence of SPPs is utilized. Recently, we reconstructed holograms in color with white-light illumination by exploiting the optical frequency dependence of SPPs [10]. Colors were extracted from white light by using this frequency dependence. In this paper, we report on the details of color selectivity and spatial color uniformity. When the color isolation is not sufficiently large, colors become mixed, and the resultant reconstructed image is whitish. When the color selectivity is spatially uneven, color unevenness occurs in a reconstructed image.

2. Color Extraction by SPPs

Figure 1 shows the optical geometry for the excitation of SPPs. Silver film is illuminated by light from the glass side. $k_{\mathrm{gx}}$ is the wavenumber of the incident light in the direction of the $x$ axis and is given by

 

Fig. 1. Excitation of SPPs by light waves. Light is illuminated at the incident angle $\theta$, which is larger than the critical angle. When the wave vector of SPPs ($k_{\mathrm{spp}}$) equals that of incident light in the $x$ component ($k_{\mathrm{gx}}$), SPPs are excited. $n_a$, $n_m(\omega )$, and $n_g$ are the refractive indices of air, silver, and glass, respectively.

Download Full Size | PDF

3. Color-Selectivity Tuning

To improve color selectivity, the silver surface must be covered with an ${\mathrm{SiO}}_2$ layer, as shown in Fig. 2(a). Figure 2(b) is a calculated relationship between the incident angle $\theta$ and the color of the SPPs. The calculations are based on Fresnel’s equations. When the ${\mathrm{SiO}}_2$ thickness $t_d$ is 0 nm, incident-angle separation between red and blue is less than 3°, which is not sufficiently large to separate colors, and the obtained reconstructed images are whitish. When the ${\mathrm{SiO}}_2$ thickness is 25 nm, which is the most appropriate thickness for color holograms in the calculations, the incident angle $\theta$ for red, green, and blue broadens to 45.6°, 48.6°, and 54.1°, respectively. If ${\mathrm{SiO}}_2$ thickness is 50 nm or more, blue SPPs cannot be excited. These color-selectivity responses are caused by the effective value of $n_a$, which is affected by ${\mathrm{SiO}}_2$ thickness [13,14].

 

Fig. 2. (a) Excitation of SPPs on silver film covered with ${\mathrm{SiO}}_2$ film. (b) Relationship between the incident angle of white light and color of SPPs to be excited when $t_m$ is 55 nm. The blackness of the band indicates reflectance at each wavelength. Low reflectance (dark part) corresponds to excitation of SPPs. When $t_d$ is 0 nm, the incident angle $\theta$ for red, green, and blue SPP excitation is 42.8°, 43.7°, and 45.2°, respectively. When $t_m$ is 25 nm, the incident angle $\theta$ for red, green, and blue SPP excitation is 45.6°, 48.6°, and 54.1°, respectively. When ${\mathrm{SiO}}_2$ thickness is 50 nm or more, blue SPPs cannot be excited.

Download Full Size | PDF

Figure 3(a) shows a color selectivity identical to that shown in Fig. 2(b), except that the silver-film thickness is thinner (35 nm). Dark bands corresponding to the excitation of SPPs broaden, and hence, the color selectivity gets worse. When the silver thickness is thicker, the bandwidth does not broaden, but the excitation efficiency of SPPs decreases, as shown in Fig. 3(b).

 

Fig. 3. (a) Relationship between the incident angle of white light and SPP color when $t_m$ is 35 nm. When bandwidth was broadened, color selectivity became worse compared with Fig. 2(b). (b) Relationship between the incident angle of white light and the SPP color when $t_m$ is 75 nm. SPP-excitation efficiency decreased compared with Fig. 2(b).

Download Full Size | PDF

4. Hologram Reconstruction with Spatially Uniform Color

In SPP holography, the hologram to be reconstructed should be located in the near field of the metal surface. In Fig. 4(a), an image hologram [15] is on a flat silver surface on which SPPs are excited. SPPs propagating on a silver surface are diffracted by the hologram [16]. The diffracted waves focus an image. In this hologram, color separation is broadened by a hologram layer instead of by an ${\mathrm{SiO}}_2$ film. Figure 4(b) shows a reconstructed image. The center of the image is green, although collimated white light illuminates the hologram to reconstruct a red image. This partial color shift is caused by hologram thickness that is thinner than the surrounding area, as shown in Fig. 4(c). Figure 4(d) shows the relationship between the hologram thickness and the wavelength of the light exciting the SPPs. Red SPPs were excited in the thick hologram area, while green SPPs were excited in the thin area. The uneven thickness of the hologram is mainly caused by the intensity distribution of the exposure light. In this case, the intensity of the center area is larger than that of the surrounding area.

 

Fig. 4. (a) Dielectric hologram on a silver surface on which SPPs are excited. SPPs propagating on the silver surface are diffracted by the hologram. The diffracted waves focus an image. (b) Reconstructed image with color shift in the center. (c) Cross section of a hologram that reconstructs images with uneven color such as in (b). The center of the hologram is thinner than both ends. (The modulation pitch is several hundred nanometers and the depth is several dozen nanometers in actual SPP holograms.) (d) Relationship between the effective thickness of the hologram and the color of SPPs (53° illumination angle).

Download Full Size | PDF

To suppress the color unevenness, a silver-relief SPP hologram [Fig. 5(a)] is needed. Fringes of the hologram consist of corrugations of silver film on which SPPs are excited and diffracted. To create this hologram, we first coated the glass surface with the photoresist and then exposed it to the reference and object light beams. Next, we developed the exposed photoresist and deposited the silver onto the hologram. As a last step, we deposited ${\mathrm{SiO}}_2$ onto the silver surface for tuning the color selectivity. In this hologram, the spatially uneven photoresist layer recorded is below the silver film, and the ${\mathrm{SiO}}_2$ layer above it can be uniformly deposited by sputtering. Thus, the color shift is eliminated. Figure 5(b) shows a red bar reconstructed from the silver-relief SPP hologram. A color shift does not occur, and the center of the image, in which exposure intensity is high, is brighter than both ends of the image. Figure 5(c) shows the distribution of image brightness.

 

Fig. 5. (a) Silver-relief hologram. SPPs are excited on the hologram made of silver-film corrugations. SPPs are diffracted by the corrugations. The diffracted waves focus an image. (b) Image reconstructed from the silver-relief SPP hologram. No color shifts occur. [Compare it to Fig. 4(b).] (c) Normalized brightness of the image, where the center is brighter than both ends.

Download Full Size | PDF

5. Conclusion

The thickness of an ${\mathrm{SiO}}_2$ film on a silver surface controls color selectivity of SPP holograms reconstructed with white-light illumination. The appropriate ${\mathrm{SiO}}_2$ thickness is 25 nm and needs to be uniform since its unevenness causes a partial color shift. The thickness of the silver film also affects the color selectivity. The optimal thickness of the silver is 55 nm.

Although gold is well known as a plasmonic material, we use silver because it covers the entire visible range, while gold covers only the range from yellow to red, owing to the resonance frequency of SPPs.

In terms of color uniformity, holograms made of silver-film corrugations are more appropriate than dielectric holograms on flat silver film because the thickness of the hologram layer is uneven, which causes a partial color shift. Although a periodical structure, such as a hologram, also shifts the color of SPPs excited by white light, the depth of our hologram is up to 25 nm, and hence, the color shift is negligible [17,18]. When we suppress an image blur by using the incident angle dependence of the SPP excitation [19], a silver-relief SPP hologram is preferred because it offers spatially stable excitation conditions.

When the hologram is recorded in our experiments, a thin photoresist film is exposed with interference fringes of the object light and the reference beam, and then it is developed. Therefore, this is a traditional process to record the thin holograms, and it is possible to make fringes of computer-generated holograms in some way. In addition, even though the SPP holograms described here are such thin holograms based on Raman–Nath diffraction, a color image with white-light illumination can be reconstructed.

When the refractive index $n_g$ decreases, the incident angle $\theta$ increases, according to Eq. (3). Thus, the hologram is illuminated at a larger incident angle when a low refractive-index substrate glass is used. Such holograms, with large incident-angle illumination, are called “edge-illuminated holograms”; these have potential for use in slate-type devices and improve the usability of holography [20–22]. An SPP hologram supports slate-type devices based on edge and back illumination, since the opaque metal film prevents stray light from emerging on the observer’s side, while typical back-lit holograms are illuminated through transparent materials.