Fabrication and operation analysis of a surface-plasmon sensor using a non-propagating mode

Our research focuses on the development of a surface-plasmon sensor that uses a stationary surface plasmon, referred to as a “non-propagating mode.” This mode is observed when light is incident perpendicularly on a surface-plasmon sensor based on a metal diffraction grating. We performed a comprehensive analysis of the behavior of the surface-plasmon resonances within this non-propagating mode, employing the rigorous coupled-wave analysis (RCWA) method. Using electron-beam lithography, sputtering, and a lift-off process, we fabricated such a surface-plasmon sensor and evaluated its optical properties rigorously. By combining simulations and experiments, we successfully utilized the non-propagating mode to detect a liquid medium with a refractive index of 1.70. Simulations show that the non-propagating mode arises due to a Fano resonance; i.e., to a resonant interaction between a localized surface plasmon generated at the edge of a metal grating strip during normal incidence and a propagating surface plasmon that occurs at the boundary between the metal diffraction grating and the measurement medium. The present results provide useful information for the advancement of surface-plasmon sensing technologies.


Introduction
A surface-plasmon sensor is a sensor that detects a medium in contact with a metal surface by using the surface-plasmon resonance generated at the interface between the metal and the medium to be measured.Such devices are frequently employed as biosensors and chemical sensors.There are several types of such sensors [1][2][3], including immunoassays, electrochemical sensors, optical sensors, and piezoelectric sensors.Immunological tests are laboratory-based, and they require concentration and puri cation prior to analysis, expensive equipment, and slow measurement times.Electrochemical sensors, optical sensors, and piezoelectric sensors can solve these problems.Electrochemical sensors can perform high-sensitivity and real-time measurements, but the electrochemical reactions may be affected by the time required for the measurement process-which may result in modi cations of the electrode surface or the substance to be measured-so accurate sensing may not be possible.
Piezoelectric sensors also are highly sensitive and capable of real-time measurements.
In contrast, optical sensors can perform non-destructive measurements and can detect multiple analytes, although some sensors have problems such as concentration-range limitations.Surface-plasmon sensors, however-which perform sensing based on the refractive index of the contacted medium-can easily measure the transmittance or re ectance peaks that are sensitive to the medium's refractive index.
They also are capable of detecting target analytes.Refractive-index sensors such as the minimumde ection-angle method, the V-block method, and the Abbe refractometer are superior to other refractiveindex sensors in terms of their high measurement accuracy, ease of use for measurements, wide measurement range, and ability to perform measurements with a small amount of sample.
On the contrary, media with refractive indices as high as 1.5 or higher may contain various harmful substances, and if a sensor can detect such media, it can be employed as a simple sensor for water pollution and soil contamination.In the Kretschmann arrangement, the angle of incidence is large, and measurement is di cult, by using a high-refractive-index medium such as GaP as the substrate, it is possible to detect measurement media with high refractive indices [13].Using this method, it is also possible to perform sensing under conditions that facilitate measurements at low angles.We have recently succeeded in using a metal diffraction grating to fabricate a surface-plasmon sensor that can detect media with refractive indices up to 1.7, and we have used it successfully to perform measurements [20,21].However, this method requires an optical system that can control the angle of incidence precisely, which can be a problem.In the present study, we have therefore focused on normal incidence, where the wavenumber of the surface plasmon is zero, and the Kretschmann arrangement cannot be used for measurements at normal incidence.Localized propagating surface plasmon polaritons (hereafter referred to as non-propagating modes) are generated under the following conditions.It has been shown that anomalous transmission occurs when propagating surface plasmons occur in metal nanostructures [39,40].If we can obtain transmitted light that is su ciently strong compared to that obtained using normal propagating surface plasmons, we can expect not only to improve sensor performance, but also to simplify measurements, due to the use of normal incidence.In addition, since a rotating mechanism is not used, the size of the device can be reduced.
Based on the background discussed above, our goal in this research was to create a surface-plasmon sensor using non-propagating modes.To achieve this goal, we rst performed simulations using the rigorous coupled-wave analysis method (RCWA method).Guided by the results of these simulations, we designed such a sensor and fabricated a sample device using electron-beam lithography and lift-off.We subsequently evaluated the optical characteristics of the fabricated device.In addition, to clarify the behavior of the non-propagating modes, we used the RCWA method to obtain transmittance maps and the distributions of the electric-eld vector.

Simulations and Experimental Method
We rst performed simulations using the RCWA method.Figure 1 shows a schematic diagram of the sensor structure used in the calculations.It consists of a one-dimensional Au diffraction grating on a SiO 2 -glass substrate, and this metal diffraction grating is in contact with the liquid medium to be measured.We assumed light to be incident from the side of the SiO 2 -glass substrate of the device, and we calculated both the dependence of the transmittance and re ectance on the angle of incidence and the wavelength dependence at normal incidence.The calculation parameters were the thickness h of the Au lm; the widths W = P/2 of the Au strips, where P is the period of the diffraction grating; the angle of incidence θ of the light; the wavelength of the incident light; and the refractive index of the liquid medium.In addition, in order to analyze the sensing operation in the non-propagating mode, we obtained transmittance maps, the electromagnetic eld distributions, and the distributions of the electric-eld vector from the transmittance-calculation results.
We fabricated a sample surface-plasmon sensor using electron-beam lithography and lift-off.We rst applied hexamethyldisilazane, an electron-beam resist (Zeon Corporation, ZEP-520A), and an antistatic agent to the cleaned SiO 2 -glass substrate, and we then drew the desired grating structure using an electron-beam lithography system (JEOL Ltd., JBX-6300).After drawing this structure, we developed it to create the diffraction-grating pattern in the resist.After coating the pattern with Au using a sputtering device (Sanyu Electron Co., Ltd, SC-701MC), we stripped off the electron-beam resist using N-N dimethylacetamide, and removed the unwanted Au using lift-off.In this way, we fabricated a dimensional diffraction-grating pattern on the substrate.
To perform measurements with this device, we placed the fabricated sample surface-plasmon sensor in contact with the liquid medium to be measured.For this medium, we used a liquid refractive-index standard consisting of a mixed solution of 1-iodonaphthalene and bromine with a refractive index of 1.7.We measured the transmittance using a laser beam with a wavelength of 670 nm to irradiate the device through the glass substrate.Irradiating the device with light from a tungsten lamp, we measured the angular dependence of the transmittance, and we measured the wavelength dependence of the transmittance at normal incidence using spectroscopy.

Results and Discussion
We measured the dependence of the transmittance on the angle of incidence at each incident wavelength (620 nm, 680 nm, 740 nm) when using a liquid medium with a refractive index of 1.70.The fabricated sensor sample consisted of a one-dimensional Au diffraction-grating structure with a structural period P = 400 nm and an Au lm thickness h = 40 nm. Figure 2 shows the measured dependences.For the diffraction-grating structure shown in Fig. 1, we found that surface plasmon resonances occur at two interfaces: the interface between the Au and the liquid medium and the interface between the Au and the SiO 2 -glass substrate [21,40].From the dispersion relations for the surface plasmons at each interface, the red circles in Fig. 2 identify the transmittance peaks due to the excitation of surface plasmon polariton (SPP) at the Au-liquid-medium interface, and the black circles identify the peaks of the SPPs at the Au-silica-glass-substrate interface; these are the transmittance peaks caused by the excitations.When the wavelength of the incident light is changed, the angles change for both types of transmittance peaks.At a wavelength of 680 nm, we found that the transmittance peak due to the excitation of SPPs at the Au-liquid-medium interface occurs at an angle of incidence of 0°. Figure 3 shows a color map of the transmittance.The horizontal axis represents the angle of incidence, the vertical axis represents the incident wavelength, and the color bar indicates the transmittance.Four yellow trajectories can be seen in Fig. 3; from the dispersion relation for the surface plasmons, they are the traces of the transmittance peaks generated by the surface-plasmon resonance at the Au-silica-glass-substrate interface and the Au-liquid-medium interface.Of these four trajectories, the two that intersect at an angle of incidence of 0° are the transmittance peaks generated by the surface-plasmon resonance at the Au-liquid-medium interface.The two remaining trajectories are those associated with the transmittance peaks generated by the surface-plasmon resonance at the substrate interface.In this gure, a straight line parallel to the horizontal axis corresponds to the incident wavelength, so the peak angle of the transmittance at that incident wavelength can be obtained from the intersection of this straight line and each of the four trajectories that represent the transmittance peaks.The peak angle of transmittance at each incident wavelength shown in Fig. 2 coincides with the peak angle found in Fig. 3. Furthermore, we con rmed that at an incident wavelength of 680 nm, there is a resonance point at an angle of incidence of 0°, and the highest transmittance is obtained there.From the dispersion relation for surface plasmons, this transmittance peak is due to a localized SPP with a wavenumber equal to zero (hereafter referred to as a non-propagating surface plasmon polariton).Figure 4 shows the results of simulations of the relationship between the angle of incidence and the transmittance on the surface-plasmon resonance line at the Au-liquid-medium interface in the transmittance map shown in Fig. 3.This con rms the sharp peak of the transmittance in the surface-plasmon resonance of the non-propagating mode, which has a resonance point at an angle of incidence of 0° (at a wavelength of 680 nm). Figure 3 shows that, at an incident wavelength of 680 nm, the two resonance lines of the Au-liquid-medium interface intersect at an angle of incidence of 0°.The transmittance of the non-propagating mode is thus superior to that of the propagating surface-plasmon resonance at other angles of incidentce.
In order to verify these simulation results experimentally-and the considerations obtained from them-we fabricated a one-dimensional Au diffraction grating, placed it in contact with a liquid medium having a refractive index of 1.7, irradiated it, and measured the spectra at various angles of incidence.Figure 5 shows a scanning electron microscope (SEM) image of the surface of the fabricated one-dimensional Au diffraction grating.This gure shows that both the structural period and the widths of the thin Au strips were fabricated as designed.Figure 6 shows a measurement of the dependence of the transmittance on the angle of incidence at an incident wavelength of 670 nm when using a liquid medium with a refractive index of 1.70 in contact with the fabricated Au one-dimensional diffraction-grating structure.Similar to the simulation results, these measurements con rmed the existence of peaks in the transmittance due to the surface-plasmon resonance at the Au-glass-substrate interface (around ± 10°) and at the Au-liquidmedium interface (around ± 1°).Furthermore, Fig. 7 shows measurements of the wavelength dependence of the transmittance when light is incident normally on the fabricated Au one-dimensional diffractiongrating structure in contact with a liquid medium having a refractive index of 1.70.Similar to the simulation results, these measurements con rmed the existence of peaks in the transmittance due to the surface-plasmon resonance at the Au-glass-substrate interface (around 585 nm) and at the Au-liquidmedium interface (around 680 nm).These results clearly demonstrate that this surface-plasmon sensor operates with a non-propagating mode for a medium with a refractive index of 1.70.
In order to clarify the behavior of the non-propagating mode, we used the RCWA method to obtain the electric-eld distribution and the distribution of the direction of the electric-eld vector near the metal diffraction grating, and we considered the behavior of the non-propagating mode.Figure 8 shows the distribution of the electric-eld vector and the electric-eld strength obtained using the RCWA method.The red arrows in the gure indicate the direction of the electric-eld vector, and the color bars indicate the electric-eld strength.Figure 8© is an enlarged view of part of Fig. 8(a).Figures 8(a) and (b) both show strong electric-eld concentrations at the edges of the Au diffraction-grating strips, and the electric-eld vector is distributed mainly in the x-direction in the medium between the Au strips.Conversely, in the medium below the Au diffraction grating, the electric-eld vector is distributed mainly in the vertical direction.We found that after the radiation passes through the metal diffraction grating, the intensity of the electric eld was stronger at an angle of incidence of 0° than at 9.9°, and the standing wave was localized with respect to time.On the contrary, at an angle of incidence of 9.9°, we found that the electric eld propagated along the interface between the metal and the medium.This shows that non-localized surface plasmons are generated at normal incidence and are non-propagating modes.
Based on these results, we considered the following model for the behavior of the non-propagating modes shown in Fig. 9. First, the x-direction electric eld of the incident TM-polarized light generates an electric dipole (the orange arrow) between the edges of adjacent Au strips, which excites localized surface plasmon polaritons.Next, the electric dipole generated at the edges of the Au strips generates an electric eld in the z-direction (blue arrows), which excites the surface plasmon polariton.The z-direction electric eld generated by this localized surface plasmon resonates with the propagating SPP; this phenomenon is a Fano resonance.Furthermore, if anomalous transmission occurs due to coupling with the propagating surface plasmon for which the Fano resonance occurs, light is emitted into the liquid medium.In the non-propagating mode, the normal incidence enhances the Fano resonance, which promotes resonance between the localized SPP and the propagating SPPs.The presence of excited propagating SPPs results in higher transmittance than is the case for propagating surface-plasmon resonances at angles of incidence other than 0°.
Next, we investigated the dependence of the transmission spectrum on the refractive index of the medium when using a non-propagating mode.Figure 10 shows a map of the refractive-index dependence of the transmittance at normal incidence.The horizontal axis shows the refractive index, the vertical axis shows the wavelength, and the color bar shows the transmittance.The peak wavelength of the transmittance due to the surface-plasmon resonance at the Au-medium interface changes with the refractive index of the medium.It is thus clear that the refractive index of a medium can be sensed using the change in the peak wavelength of the transmittance at the Au-medium interface, even for non-propagating modes.

Conclusions
We have designed and fabricated a surface-plasmon sensor using a non-propagating mode and have used transmittance measurements to demonstrate the operation of this sensor.In addition, from transmission maps and the distribution of the electric-eld vector, we showed that the enhancement of the electromagnetic eld by a Fano resonance leads to high transmission in the non-propagating mode.

Declarations Figures
Schematic Relationship between the angle of incidence and the transmittance for the SPP resonance line at the Au/medium interface.
Surface SEM image of the fabricated Au diffraction grating.
Dependence of the transmittance on the angle of incidence at an incident wavelength of 670 nm.
Page 16/18 Wavelength dependence of the transmittance at normal incidence.
Physical model of the non-propagating mode.
Figure 10 diagram of the surface-plasmon sensor.