1. Introduction

The first surface plasmon resonance (SPR) microscope was demonstrated by Yeatman and Ash in 1987 [1, 2]. They displayed a high-contrast image of a 2.5-nm-thick WO₃ grid formed on a silver thin film with 25-μm lateral resolution. The most significant feature of this microscope was its ability to create high-contrast images from very weak refractive index changes or thickness variations of samples, which was achieved by making use of the fact that the resonant condition of the SP is susceptible to the existence of dielectric on the metal film. Rothenhäusler and Knoll also reported similar results [3]. They demonstrated visualizing a Langmuir-Blodgett film of cadmium arachidate by using SPR microscopy. They also proposed introducing a combination of the Otto and the Kretschmann configurations to SPR microscopy in order to image surface relief structures of external samples [4]. As an application of SPR microscopy, Yeatman and Caldwell proposed a spatial light modulator using nematic liquid crystal (LC) as the active material [5, 6]. The interaction of an SP with nematic LC has been discussed by Evans et al., and they demonstrated that an SPR microscope can image LC anchoring on a patterned self-assembled monolayer [7].

One of the weaknesses of SPR microscopy, however, is the lack of spatial resolution due to the long traveling length of SPs [8]. This problem was discussed by Berger et al. [9]; as a solution, they proposed the use of shorter-range plasmons by changing the wavelength of the light source. Another solution was proposed by Bruijn et al. [10]. They proposed rotating a sample object to change the object’s orientation relative to the propagation direction of the plasmon, and they obtained a lateral resolution of 1.5 μm without any decrease in sensitivity with respect to index or thickness. We also investigated a laser-scanning SPR microscope that could obtain images with a high spatial resolution of 1.6 μm, while maintaining high sensitivity to thickness and refractive-index distributions [11]. This system used the spatially localized surface plasmon spot produced by using a high-NA microscope objective [12, 13].

In this paper, we apply the laser-scanning SPR microscope technique to simultaneously measure both birefringence and its orientation of an anisotropic material in real-time and with high sensitivity.

2. Optical setup

Figure 1 shows a schematic diagram of the laser-scanning SPR microscope used. This setup is based on the Kretchman-Raether configuration [14]. A silver thin film was evaporated on a high-index cover slip whose refractive index and thickness were 1.78 and 170 μm, respectively. An anisotropic sample was attached on this silver thin film. A He-Ne laser (λ = 632.8 nm, 5 mW) was used as a light source. The linearly polarized laser beam was converted to circularly polarized light using a quarter wave plate in order to generate radially polarized components, and then this laser beam was collimated by a beam expander. The collimated beam was introduced to a microscope objective (oil-immersion, 100× magnification, and NA = 1.65), and was tightly focused on the silver thin film through the cover slip. The focused laser beam excited surface plasmons on the metal surface facing the sample. The laser beam reflected from the silver film was collected by the same objective lens. The Fourier spectrum of the reflected light, which was created at the objective’s exit pupil, was observed by a CCD camera after reflection at a beamsplitter (BS). An interlaced NTSC signal was output from the CCD camera and it was captured by a computer at a rate of 30 frames/sec. A diffuser was introduced in front of the beam expander for reducing coherent noise, such as speckle and unwanted interference fringes.

 

Fig. 1. Schematic diagram of optical setup of laser-scanning surface plasmon resonance microscope.

Download Full Size | PDF

Figure 2 shows the principles of measuring the sample’s birefringence and its orientation. Figure 2(a) depicts the relationship between a dark ring seen in the Fourier image of the reflection beam and the resonance of the SP. If the resonant angle of the SP is inside the angular semi-aperture of the objective lens, a certain component of the reflected light is absorbed by the SP, and as the result, a dark ring is created in the Fourier image. When the sample attached to the silver film is isotropic, the shape of the dark ring is completely circular because of the rotational symmetry of the system. Since the radius of the ring corresponds to the resonant angle of the SP and the resonant angle is directly associated with the refractive index of the sample, the system can measure the absolute value of the sample’s refractive index from this Fourier image. Note that since the SP is excited only by p-polarized light, the incident beam of this system must have radial components.

Figure 2(b) schematically shows the transformation of the dark ring’s shape when an anisotropic material is placed on the silver film. Whereas the incident light coming from point A and reflected to point C experiences the refractive index in the x-z plane, the light coming from point B and reflected to point D, on the other hand, experiences the refractive index in the y-z plane. As a result, the dark ring changes to an elliptic shape according to the birefringence of the material. By measuring the length of the major and minor axes of the elliptical dark ring, the system can measure the absolute value of the birefringence, as well as its orientation.

 

Fig. 2. Optical arrangement for exciting surface plasmons by use of a high-numerical-aperture objective lens. This figure also depicts the creation of a dark ring due to the absorption of the surface plasmon excitation.

Download Full Size | PDF

Note that because this system can measure the birefringence without any mechanically moving optical components, such as a polarizer, normally required in conventional polarization microscopy, it can measure changes in birefringence and changes in orientation of an anisotropic material in real-time. In addition, since the evanescent field associated with the SP is localized at the metal-sample interface, this system can selectively measure the sample’s birefringence only in the vicinity of the interface.

3. Experimental results

In the experiment, we used birefringent liquid crystal (LC) as the anisotropic material. Figure 3 depicts the configuration of the sample. On the silver thin film, a poly(vinylalcohol) (PVA) thin film (n = 1.49 and 120 nm in thickness) was spin-coated and rubbed in one direction with absorbent cotton to serve as an alignment layer. The nematic LC used was Merck ZLI-1132. The birefringent indices of this LC were n_o = 1.4929 and n_e = 1.6332. This LC was dripped on the PVA film and covered with another cover slip on which an aluminum film was evaporated as an electrode for applying an external voltage to the LC. Figure 3(b) shows the alignment of the LC molecules with and without an applied external voltage. The silver thin film for the SP excitation was used as the opposing electrode. PMMA beads of 5 μm in diameter were dispersed between the cover slips to serve as spacers.

 

Fig. 3. The configuration of the liquid crystal cell.

Download Full Size | PDF

Figure 4 shows experimental results. Figure 4(a) depicts the elliptic dark ring when the PVA film was rubbed in the horizontal direction of the photograph and no external voltage was applied. The dark ring was deformed to form a horizontally extending ellipsoid. Figure 4(b) shows an image where the PVA film was rubbed vertically. The deformation of the ring also changes to form a vertically extending ellipsoid. From these photographs, it is clear that the orientation of the elliptic ring changes according to the rubbing direction, which defines the orientation of the LC molecules; also, the photographs reveal that the observed elliptical dark ring represents the orientation of the LC molecules’ birefringence. From Figs. 4(a) and (b), the ellipticity of each dark ring was calculated to be 1.016 and 1.018, respectively.

 

Fig. 4. Experimental results of the observed elliptic dark ring. Liquid crystal (LC) molecules were aligned (a) in the horizontal direction, and (b) in the vertical direction.

Download Full Size | PDF

We examined the dependence of both the orientation of the birefringence and the light propagation direction on the lengths of elliptic ring axes by using a 4×4 matrix method [15]. Figure 5 shows the calculation model and results. We assumed that the light travels only in the x-z plane; incident light comes from point A and is reflected to point C in Fig. 2(b), and the orientations of the LC molecules are parallel to x-, y-, and z-axes, as shown by α, β, and γ in Fig. 5(a). Other conditions were the same as the experimental setup. The incident plane wave comes from a high index material (cover slip, n=1.78). A silver thin film of 55 nm in thickness is placed on this material, and a dielectric thin film (n=1.49, 120 nm in thickness) is coated on the silver film, which corresponds to the alignment layer of the P VA. Anisotropic material (liquid crystal) with n_o=1.492 and n_e=1.632 is then placed on the alignment layer; these values are equivalent to those of ZLI-1132.

Figure 5(b) shows the calculation results. From these results, the calculated SP resonant angles for each configuration were 64.254, 64.614 and 68.155 deg, respectively. By using the former two values, the ellipticity, which is defined as the ratio of the length of the major axis to that of the minor axis, was calculated to be 1.016. This value has good agreement with the above experimental results. In addition, note that under the condition of z-orientation, the resonant angle of 68.155 corresponds to a numerical aperture of 1.652, which exceeds the numerical aperture of the objective lens we used in the experiment. This indicates that in our experimental setup, the dark ring will disappear from the Fourier image when the external voltage is applied to the LC.

 

Fig. 5. Relationship between incident angle and reflectance. (a) Schematic diagram of the numerical simulation model. (b) The results of calculation. Lowest reflectance point corresponds to the surface plasmon’s excitation angle.

Download Full Size | PDF

Figure 6 shows the deformation of the dark ring when the external voltage was instantaneously changed from 0 V to 5 V, and from 5 V to 0 V. Since the distance between the electrodes was regulated to 5 μm by the PMMA beads, the applied electric field was estimated to be 1 MV/m. When the external voltage was applied to the LC, the LC changed from homogeneous alignment to homeotropic alignment and we can see that the dark ring disappeared quickly. This indicates the rapid orientation change of LC molecules (less than 33 ms) when the external voltage is applied. After turning off the external voltage, the LC molecules returned back to homogeneous alignment, and the dark SPR ring also appeared again, taking almost 0.5 s.

 

Fig. 6. (2.5MB Movie) Deformation of the dark ring when external voltage is applied and removed.

Download Full Size | PDF

4. Conclusions

In conclusion, we have demonstrated that a laser-scanning surface plasmon resonance (SPR) microscope can simultaneously measure both the absolute value of the birefringence of an anisotropic material and its orientation. Because this method does not need any moving optical components, such as a polarizer, normally used in conventional polarizing microscopy, we can observe changes in birefringence in real-time. This is extremely advantageous for evaluating the dynamic properties of devices using active birefringent materials, such as LC displays, spatial light modulators, and so on. Note that to obtain a two-dimensional image of the birefringent distribution of the sample, we must scan the laser beam spot two-dimensionally in the lateral plane, which may take some time. In addition, in order to observe the orientation of the birefringence, a Fourier image shown in Fig. 4 for each sample point must be obtained. This also takes time comparing to the normal laser scanning microscope (LSM) such as reflection LSM, confocal LSM and so on, because these LSMs acquire only one voltage signal from photodetector for each sample point. Therefore, our system cannot currently observe 2D images of anisotropic material in real-time, but this problem will be solved by using recently developed high-speed beam scanners, such as resonant galvanometer scanners and high-speed video cameras [16].