1. Introduction

Surface Plasmon Resonance (SPR) can occur at a metal-dielectric interface when the incident light electromagnetic energy is coupled to the plasmon mode. This enables harmonic free electrons density fluctuations within the metal layer, leading to enhanced electromagnetic fields at the metal-dielectric boundary, as well as increased transparency of the metal layer. Coupling incident light into a surface plasmon strongly depends on the optical properties of the metal and dielectric layers, the interface morphology, as well as the homogeneity of both materials [1]. Since the SPR wave vector is always larger than the incident light wave vector in the dielectric material [2], the excitation of SPR is only possible by using phase-matching techniques, such as evanescent waves from total internal reflections, also known as the Kretschmann configuration [3], and diffracted light from gratings. Both of these techniques permit an equality in the SPR/incident light wave vector magnitudes [4]. Furthermore, the SPR excitation becomes especially effective when the incident light is linearly polarized and its wave vector projection onto the metal film plane is equal to that of the surface plasmon [5].

In 1988, W. Knoll and B. Rothenhausler were the first to suggest using SPR for microscopy [6]. While the large majority of the existing SPR imaging devices [7–10] are built using the Kretschmann configuration, very few are based on metallic gratings, which are significantly less bulky. SPR excitation with gratings is advantageous because it permits the use of collinear optics, thus significantly reducing the overall size of the device, and the SPR wavelength can be precisely tuned via the grating pitch, which is set during its fabrication process.

Even though the phase-matching condition for exciting a SPR theoretically yields a single light wavelength, experimentally, a narrow bandwidth of SPR wavelengths is observed due to dispersion in materials. When polychromatic light is used to excite a SPR, a positive peak is observed in transmission, while a negative peak is observed in reflection due to the conservation of energy. When using diffraction gratings to excite plasmons, the theoretical SPR wavelength can be computed based on the refractive index of the dielectric, the real part of the dielectric permittivity of the metal, the grating pitch, the light incidence angle and finally the light polarization [11]. Changing any one of those parameters, such as the refractive index (RI), while keeping the others fixed, induces a measurable wavelength shift in the SPR spectrum, from which the precise shift in the refractive index of the dielectric medium can be computed.

The possibility of effective excitation of SPR and its influence on many photo-physical processes occurring on the surface of investigated materials prompted a wide range of applications of this phenomenon in scientific laboratories and industry. SPR sensors have been used in optical spectroscopy [12], non-linear optics [13], high-resolution microscopy [14], light modulators [15], biosensors technology development [16], and other areas. However, some of these SPR devices have limitations due their bulkiness, limited resolution, and integrability in practical electronics.

The principal constituting element of the device presented in this work is a Crossed Surface Relief Grating (CSRG) fabricated by the orthogonal superposition of two sinusoidal surface relief diffraction gratings inscribed sequentially on an azobenzene molecular glass thin film and coated with a thin layer of silver [17]. The gratings are formed in the azobenzene layer upon exposure to a laser interference pattern produced by a Lloyd mirror interferometer [18]. This occurs due to the unique photoisomerization properties of the azobenzene moieties, in which a photomechanical effect is observed through the migration of the thin-film’s molecules from high to low laser irradiance regions [19]. The inscribed gratings are stable over time and under ambient temperature, and subsequent high-intensity laser irradiations with different interference patterns yields superimposed gratings that permit the retention of the diffraction properties from each individual laser exposure. The final surface relief pattern on the azobenzene film is subsequently coated with a metal layer which takes the same shape of the underlaying nanostructures.

When linearly polarized light is incident on a metallic CSRG, a SPR is excited by the grating having its vector aligned with the incident light polarization. At normal incidence, a SPR standing wave is created [20]. The energy in this plasmonic standing wave interacts with the second orthogonal grating in a way that the SPR energy is re-emitted by the second grating following the reverse principle, although in a polarization state that is orthogonal to that of the incident light. Hence, a polarization conversion of light occurs through momentum transfer in between the crossed gratings [17]. When placed in between crossed polarizers, this plasmonic energy exchange in CSRGs provides a remarkable opportunity for observing the pure SPR spectral response, without any residual light from a polychromatic light source, hence with a very high signal-to-noise ratio. This provides an enhanced detection sensitivity at the metal-dielectric interface since the SPR signal is nominally measured in transmission as a narrowband signal of interest that can tuned to the desired wavelength range. This approach has been used in the past by our group to successfully detect proteins [21] and bacteria [22] in water-based solutions placed over the gratings through multispectral SPR interrogation. However, in these previous studies, the multispectral white light only allowed the detection of changes in the average RI of the solutions placed over the entire CSRG surface. Furthermore, a spectrometer was needed to analyze the spectral composition of light to detect any shift in the SPR resonance.

In this work, a camera was used for the first time to image and detect two-dimensional refractive index variations in solutions of deionized water mixed with either silicon oil or sucrose and placed over the plasmonic CSRGs. Monochromatic red light was used instead of the usual polychromatic light to excite the SPR. Since the use of a spectrometer is no longer needed, this would be beneficial for transitioning the use of these SPR sensors towards cheaper and more portable devices. The technique presented in this work enabled the measurement of small changes in the transmitted light intensity through the metallic CSRG, which was placed in between crossed polarizers. As a consequence, the change in the transmitted light intensity was correlated to changes in the refractive index of a control water-sucrose solution. Subsequently, a camera was used for the first time to successfully image the two-dimensional refractive index distribution in the water and silicon oil solution through the SPR polarization-contrast conversion phenomenon.

2. Experimental procedure

CSRGs were fabricated on thin films of azobenzene molecular glass as described elsewhere [21]. For this work, the gratings’ pitch was selected so that when pure water is placed on top of the CSRG, the SPR signal intensity from incident polychromatic white light would be maximized near the wavelength of 645 nm in transmission. Both orthogonal gratings’ pitches were kept identical in order to fulfill the polarization conversion condition for normally incident light [17]. When varying the refractive index of the water-based solution, the maximum SPR spectrum wavelength shift will be at the highest slope value of the SPR transmission spectrum. In this case, the maximum SPR spectrum slope was optimized to be at a light wavelength of 632.8 nm. This was done through slight variations of the CSRG pitch and modulation depth during the fabrication process. Therefore, when using incident monochromatic light at 632.8 nm, instead of the polychromatic white light, a change in the SPR signal intensity should occur in response to a change in the tested dielectric RI. Various water-based solutions were made by mixing glycerol or sucrose and deionized water at different concentrations and the obtained solutions’ RIs were measured using an Abbe Refractometer. The bulk sensitivity of the SPR device was obtained by illuminating it with polychromatic white light from a halogen lamp and the SPR spectrum of the solutions was measured using a fiber optic CCD spectrometer in transmission mode.

For imaging, the CCD spectrometer was substituted by a high-dynamic-range monochromatic camera (Edmund Optics Model EO-04138M MONO HDR) and a narrowband 632.8-nm filter was added in front of the white light source. As illustrated in Fig. 1(a), the white light passed through a square aperture and a variable iris. Then, it went through the narrowband 632.8-nm filter and was polarized by a horizontally-aligned linear polarizer. The light was focused on the metal CSRG surface, using a convex lens, forming a spot of approximately 1 mm in diameter at normal incidence. After passing through the sample, the light went through a second linear polarizer with its axis perpendicular to the first one in order to eliminate all residual non-SPR light from the halogen lamp. Then, it was focused again with a second lens onto the CCD spectrometer or the camera, which had a 762 × 576 pixels sensor matrix and a single pixel size of 10 × 10 µm².

 

Fig. 1. a) The optical setup used for the SPR imaging. b) Side-view illustration of the CSRG device.

Download Full Size | PDF

To contain the liquid dielectric solution over the CSRG, an opening was cut in a 2-mm thick slab of polydimethylsiloxane (PDMS) prepared beforehand. The PDMS slab was placed atop of the CSRG area, surrounding the entire CSRG quarter-disk, it was filled with the test solution and covered with a microscope cover glass to avoid leakage and the lensing effect. A side-view of a device ready for testing is illustrated in Fig. 1(b).

3. Results and discussion

The fabrication of a metallic CSRG, with a SPR spectrum slope maximized at a wavelength of 632.8 nm, involved optimizing the gratings’ pitch and modulation depth, which are both controlled during the CSRG laser inscription process. Similar plasmonic CSRG optimizations had been done previously in our group for different purposes and the results were published in the literature [22–24]. During those optimizations, it was found that laser irradiation times of 120 secs for the primary grating and 60 secs for the secondary grating, with a 532-nm laser at an irradiance of 428 mW/cm², provided CSRGs deep enough to generate the highest possible SPR polarization conversion in transmission, while avoiding SPR sensing resolution decrease typically associated with SPR photonic band gaps in deep gratings, as explained elsewhere [25].

The gratings’ pitch was adjusted so that when pure water is placed on top of the grating, the theoretical SPR wavelength ${\lambda _{sp}}$ would be given by the following equation [26]

 

Fig. 2. Normalized transmitted SPR signal by a CSRG optimized for obtaining a maximum SPR spectrum shift at a wavelength of 632 nm. a) The normalized signal obtained when the CSRG is covered with pure water. b) Signal redshift caused by the dielectric RI change.

Download Full Size | PDF

Once the plasmonic CSRG depth and pitch optimizations were completed, a series of tests were carried out in order to obtain the bulk sensitivity of the device with respect to the dielectric medium RI change, as well as to observe how the brightness of the two-dimensional camera images would correlate with the RI change. To simulate the RI change, several homogeneous glycerol/water solutions of different concentrations were prepared with refractive indices reported in Table 1.

Table 1. Glycerol-water solutions refractive index.

View Table

The SPR spectra were obtained for all the solutions listed in Table 1 using the same CSRG device, after careful cleaning of the PDMS well. All these plots were normalized to their respective maxima. In Fig. 2(b), the normalized SPR spectra are shown to redshift with varying solution concentration, under polychromatic light illumination. Since the redshift is small when compared to the spectrum bandwidth, the raw spectrum plot is zoomed around the 632.8 nm area in order to give a better visual representation of the spectrum shift. From Fig. 2(b), it can be concluded that concentration increments of 5%, increased the RI by ∼0.0055, and caused ∼0.84 nm redshift on average and ∼4.2% normalized signal strength change.

Under monochromatic 632-nm light illumination, the camera capability to capture the dielectric’s bulk RI change was tested using aqueous sucrose solutions of concentrations from 0 to 25% with an increment of 5% for the bulk RI change simulation. Raw images with no software filtering and a fixed brightness adjustment were recorded in the BMP format with 8-bit, 256-level greyscale. Figure 3 has the resulting images with the corresponding dielectric’s RIs. As seen in these pictures, the changes in the visual brightness of the images correlates very well with the RI change, making this imaging technique a great candidate for further research on SPR microscopy imaging of inhomogeneous mixtures.

 

Fig. 3. Image brightness change as a function of the water-sucrose solution’s RI increase, observed at a wavelength of 632.8 nm.

Download Full Size | PDF

In the final part of the work, polarization-contrast images of inhomogeneous water-based solutions were recorded and analyzed when the silver-coated CSRGs were irradiated with incoherent light at a wavelength of 632.8 nm. Deionized water-based silicone oil mixtures were prepared with varying concentrations from 1 to 5% (w/w). Figure 4(a) represents a time-lapse sequence of polarization-contrast images extracted from a video observation of the silicone oil particles floating up in the water during the mixture sedimentation over the CSRG’s surface. The dark-and-white patterns indicate different RIs within the same mixture due to the water-oil inhomogeneity. The movement of selected oil regions is indicated by the yellow circles, seen moving upward with time through sedimentation, noting that the image orientation is reversed by the focusing lens. Prior to placing the water-oil mixture in the PDMS reservoir over the CSRG, a polarization-contrast image was captured with pure water only. This image was used as a baseline for obtaining the difference image of the silicone oil regions in the mixture, after the sedimentation had completely stopped. The result of the simple mathematical subtraction of the pixel intensity between the two images was made using the Origin software and is shown in Fig. 4(b).

 

Fig. 4. Polarization-contrast SPR imaging of an inhomogeneous water-oil mixture. a) A silicon region sedimentation is observed with time. b) A silicone mixture difference image is obtained 30 min after the sedimentation has completed.

Download Full Size | PDF

4. Conclusion

Two-dimensional crossed surface relief gratings were fabricated in azobenzene thin films and they were used to obtain SPR polarization-contrast images of various homogeneous and inhomogeneous aqueous mixtures. When monochromatic light was incident on the test device, which was placed in between crossed polarizers, variations in the image brightness were correlated to a change in the solutions’ RI, by means of the SPR polarization conversion in between the two superimposed gratings. Therefore, the one- or two-dimensional spatial distribution of the SPR signal intensities were recorded using a high-dynamic-range camera and analyzed in terms of the RI change.