1. Introduction

We have many nanostrucutres in nature [1–4]. A wing of cicada (C. orni) has aligned nanoposts, which gives hydrophobicity [5]. The wings of Morpho rhetenor butterflies have a periodically ordered photonic structure that shows structural color of blue [6, 7]. Superhydrophobicity is also observed in butterfly wings due to its nannostructure [8]. The mosquito (C. pepiens) has compound eyes of nipple array for antiforgging properties [9]. The surface of lotus (Nelumbo nucifera) leaves shows strong hydrophobicity due to microscopic protuberances with nanoscale secondary roughness composed of wax crystalloids [10–12]. Many nanomaterials have been reported mimicking the surface of lotus leaves to enhance its hydrophobicity and self-cleaning [13–20].

In our previous paper, we report that a 10-nm thick gold-thin film over a lotus leaf shows great light absorption over the visible wavelengths [21]. It stems from makaroni-like nanostructures of the surface of a lotus leaf, at which gold-coated nanorods are randomly oriented. The reflectivity from the gold-covered surface was less than 1% over the visible spectral range. Therefore it can be used for blackbody or light absorber.

The surface of a taro (Colocasia esculenta) leaf is also water-repellent as shown in Fig. 1(a), although the surface nanostructure of taro leaves completely differs from that of lotus leaves. We coated the surface of a taro leaf, and found that the gold-coated surface of the taro leaf is almost black, similar to the gold-coated lotus leaf. The gold-coated taro-leaf surface has reflectivity below 1% over the visible spectral range. The finite-difference time-domain (FDTD) calculations on a model based on the scanning electron microscope (SEM) image of the taro leaf surface. The experimental observation is supported by the calculated results.

 

Fig. 1 Photo images of (a) a taro leaf and (b) gold-coated taro leaf.

Download Full Size | PDF

2. Experimental

The samples were prepared by the following procedure: a taro tuber was sowed in water. After a few weeks, the cotyledon was nipped off and fixed on a glass slide. A thin gold-film coat was made on the leaves by sputtering in air at low pressure, using an E-1030 sputtering coater (Hitachi).

The reflection spectra were recorded with a MCPD-3000 spectrometer (Otsuka Electronics) using a halogen lamp as a light source. For the reflectivity measurements, the light was conveyed to the sample with a Y-type optical fiber (400 μm in diameter) and the reflected light was collected by it. The light was irradiated at normal incidence. For the scattering measurements, the light from the halogen lamp was likewise conveyed by an optical fiber to the sample at normal incidence. The back-scattered light was collected by another optical fiber and transferred to the spectrometer. As a reference, an SRS-99 diffuse reflectance standard (Labsphere) was used. The scattering angle was approximately 30° with respect to the surface normal. SEM observations were performed with an S-4500 SEM (Hitachi).

3. Results and discussion

The photographic image of the gold-coated taro leaf is shown in Fig. 1(b). The gold coating was done by sputtering and the thickness of the gold is 30 nm. The leaf was fixed by adhesive tapes, which looks lustrous and metallic. The surface of the leaf is slightly-brownish dark, and is not as black as the gold-coated lotus leaf [21]. Some stripes are seen due to veins of the leaf. When the thickness of gold is 10 nm, the surface looks green, in contrast to our previous report that the 10-nm thick gold-coat is enough for the lotus leaf to be black. This suggests that the gold-coated taro leaves have different mechanism of the reduction in reflectivity.

Figure 2(a) shows reflection spectra of the gold-coated taro leaf (curve I, red line) and that of the taro leaf without gold coating (curve II, black line). In the wavelengths from 400 to 700 nm, the reflectivity is almost similar. The reflectivity spectrum from the gold-coated leaf shows gradual increase with increasing the wavelengths. In the spectrum of the uncoated taro leaf, a broad peak is observed at 540 nm. This peak is attributed to the multiple scattering in the leaf due to absence of absorption by Chlorophyll at this wavelength range [22]. The peak is absent in the reflection spectrum of the gold-coated taro leaf. The difference in reflectivity is found in the wavelengths longer than 700 nm, where Chlorophyll has absorption. In this wavelength range, the reflectivity from the bare leaf-surface is much higher than that of the gold-coated surface. This behavior is more significant in the scattering spectra shown in Fig. 2(b), in which the scattering intensity S is normalized by that from the reference (SRS-99 diffuse reflectance standard), S₀.

 

Fig. 2 (a) Reflection spectra from the gold-covered taro leaf (I) and bare taro leaf without coating (II). (b) Scattering intensity spectra S/S₀ from the gold-covered taro leaf (I) and taro leaf with no gold-coating (II).

Download Full Size | PDF

Figures 3(a)–(d) show SEM images of the gold-coated taro leaf. At the low magnification image (a), microscopic protuberances of the size of about 10–20 μm are observed in image (a). Each microscopic protuberance is partitioned by the surface convex structures. Image (b) is the magnified image of the center of image (a). There are three kinds of structures, denoted A–C. Region A is the surfaces other than the protuberances. They are covered with thin rectangular plates that are vertically set on edge, as shown in image (c) of Fig. 3. The thickness of the plate is approximately 50 nm, according to the magnified image (inset of Fig. 3(c)). The thickness is much lower than the gold thin films deposited both sides (30 nm + 30 nm). The reason for the thickness lower than expected is that the plate is vertically ordered and the sticking efficiency of evaporated gold to the vertical surface of the plate is much lower than that expected. Thus it is likely that sputtered gold is deposited on the upper edge of the plate. This vertical plate structure is also observed on the protuberances in region B, as shown in image (d), whereas they are tilted or disordered. At the surface of some of the protuberances, the plates are absent and the surface is rather smooth (region C). The region C is, however, too narrow to contribute to the optical response.

 

Fig. 3 SEM images of the gold-covered taro leaf. (a): low-magnification image. (b): the magnified image of the center of the image (a). (c): the magnified image of region A and the fine image of the plates (inset). (d): the magnified image of region B.

Download Full Size | PDF

The gold-covered taro leaf is dark and is not lustrous, similar to the gold-covered lotus leaf as reported previously, although the surface structure of the taro leaf is completely different from that of a lotus leaf [21]. We performed FDTD (Lumerical Solutions Inc.) calculation to reveal the mechanism of the low reflection of the gold-covered taro surface. The models for the calculation were produced as follows. The SEM image (2 μm × 2 μm) in region A shown in Fig. 4(a) was imported into the FDTD software The image was binarized to black-and-white to generate the top-view image of the nanostructure as shown in Fig. 4(b). We produced the model for calculation on the basis of the image. The perspective view is illustrated in Fig. 4(c). It is divided into three layers. The top layer is a 30-nm thick gold film on the nanostructure of rectangular plates corresponding to the plate at the taro-leaf surface. The top view of the nanostructure is the second layer and of the 100 nm-thick dielectric material. The thickness 100 nm is taken from the height of the plate determined from the cross-sectional SEM image. The refractive index of the dielectric material was set to be 1.5. The bottom layer is binary: the nanostructure area is dielectric and other area is a 30-nm thick gold film, which is deposited through the nanostructures in the actual sample. The three layers are stacked in the model for calculation. We also made the model for region B using the SEM image Fig. 4(d). The binarized image and the perspective view are shown in Figs. 4(e) and 4(f), respectively. In the FDTD calculation, the space was divided into meshes of 5 nm ×5 nm ×5 nm and the perfectly matched layers (PML) were used for the boundary. The total-field/scattered-field (TFSF) was used for light illumination at normal incidence to the surface.

 

Fig. 4 (a) SEM image of region A, (b) the image of the binary-processed SEM image in region A, and (c) the model for calculation of the region A. (d) SEM image, (e) the binary-processed SEM image, and (f) the model for calculation of the region B.

Download Full Size | PDF

The calculated spectra of reflectivity R, transmittance T, absorption efficiency A and the scattering efficiency S are shown in Fig. 5(a) for the model in region A. The reflectivity is 0.1 −0.4 over the visible spectral range and the absorption is larger than the others. The scattering efficiency and the transmittance are low ~0.2. Thus the low reflectivity is mainly due to the large absorption of the nanostructures. The spectra for region B are also shown in Fig. 5(b). They are similar to those for region A, in spite of some structural differences observed in the magnified SEM images.

 

Fig. 5 Calculated spectra of reflectivity R, transmittance T, absorption efficiency A and the scattering efficiency S for the model of region A (a) and for the model of region B (b), using the FDTD method.

Download Full Size | PDF

Figure 6 shows calculated spectra of reflectivity R, transmittance T and absorption efficiency A for a flat 30-nm thick gold film using the Transfer Matrix method [23]. Since the film is plannar with no surface structures, the scattering is absent. The reflectivity of a 30-nm thick gold thin film is higher than 0.7 at wavelengths longer than 600 nm, where the reflectivity of the taro-leaf model is much lower (0.25 – 0.4). The absorbance of a 30-nm thick gold thin film is less than 0.1 in the wavelengths longer than 600 nm, whereas the taro-leaf model has the reflectivity greater than 0.5. Thus these great differences stem from the nanostructure of the taro-leaf model. Accordingly it is concluded that the observed low reflectivity stems from the nanostrucutres of the taro leaf surface.

 

Fig. 6 Calculated reflectivity R, transmittance T and absorption efficiency A for a flat 30-nm thick gold thin film using the Transfer Matrix method [23].

Download Full Size | PDF

Although qualitative interpretation was given with the FDTD calculation, there still exists discrepancy in the observed (~0.01) and calculated (0.1–0.4) reflectivity at the visible wavelengths. One reason for the discrepancy is that the calculated R by the FDTD method involves contribution of backscattering. The other reason is the plannar surface of the upper edge of the plate in the model, whereas the edges of the actual plates of the taro leaf are more complicated, as shown in the SEM images of Figs. 3(c) and 3(d). Such gold-coated nanostructures at the plate edges will show strong localized surface plasmon resonance which causes greater absorption, as observed in the taro-leaf samples. The FDTD calculations taking account of such nanostrucutres are in progress and will be reported elsewhere.

4. Conclusion

We have reported a dark metamaterial that a gold thin films is deposited on a taro-leaf surface that was used as a template. In spite of gold coating over the taro leaf, the surface shows dark and light-absorption over the visible wavelengths. These results are qualitatively supported by the FDTD calculations using the model produced on the basis of the SEM image of the taro leaf surface.