Open Access
Add to BibSonomy
Abstract
Aimed at detecting fluorescent-labeled biological substances sensitively, a sensor that utilizes near-field light has attracted much attention. According to our calculations, a planar structure composed of two dielectric layers can enhance the electric field of UV near-field light effectively by inducing waveguide-mode (WM) resonance. The fluorescence intensity obtainable by a WM chip with an optimized structure is 5.5 times that obtainable by an optimized surface plasmon resonance chip. We confirmed the above by making a WM chip consisting of TiO2 and SiO2 layers on a silica glass substrate and by measuring the fluorescence intensity of a solution of quantum dots dropped on the chip.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Total internal reflection fluorescence microscopy, which illuminates objects by near-field light, is widely used for imaging and detecting a fluorescently labeled cell [1,2] and protein [3]. The microscopy can enhance the electric field of near-field light by utilizing surface plasmon resonance (SPR), which makes it possible to excite fluorescent substances effectively [4,5].
As a fluorescent label for biological substances, research on a semiconductor crystal with a diameter of a few nm, called a quantum dot (QD), has been carried out extensively in recent years, since it emits bright fluorescence and has a strong durability against breaching [6,7]. Many QDs can be excited more effectively by ultraviolet (UV) light than by visible light since their optical absorption coefficients increase toward a shorter wavelength. Another advantage for using a QD as a fluorescent label is that its fluorescence can be easily separated from the excitation light because of a large Stokes shift.
For exciting QDs, the electric field of near-field light should be enhanced effectively in UV region. It is widely known that SPR can be excited by a thin layer of metal such as Au, Ag, and Al. In the visible and near-infrared region, Au and Ag can excite SPR effectively, but not so much in the UV region [8–10]. In contrast, Al can excite SPR over a wide wavelength range from UV to visible [8–12]. Therefore, SPR sensing techniques with an Al thin layer have been used for fluorescent labeling excitable by UV lights [13,14].
In addition, a two-layer structure for waveguide-mode (WM) excitation shown in Fig. 1 has been reported as another candidate [15–18]. We have numerically optimized the structure of WM excitation layers to realize a WM chip suitable for UV near-field fluorescence sensor using a transfer matrix method [19–21]. According to the calculation results, it is expected that an optimized WM chip with a 28-nm TiO2 layer 1 and a 315-nm MgF2 layer 2 can enhance the normalized electric field strength of UV near-field light at a wavelength of 375 nm by 387 times. This enhancement ratio is much higher than the corresponding ratio calculated for a SPR chip [19]. However, when we actually made a WM chip by sputtering, an MgF2 layer was found to have several problems such as its weak adhesiveness and large roughness.
Fig. 1 Schematic optical arrangement of the UV near-field fluorescence sensor with the two-layer WM chip (not to scale).
Instead of MgF2, SiO2 can satisfy the various conditions mentioned in [19] and can be sputtered on TiO2 very smoothly and stably. From such a viewpoint, in this research, we compare the ability of enhancement of electric field of near-field light between a WM chip composed of a TiO2 layer 1 and a SiO2 layer 2 and a SPR chip with an Al layer.
2. Structural optimizations of a TiO2/SiO2 WM chip and a SPR chip for comparison
The optical setup to excite the WM resonance and SPR that we made for this study has similar Kretschmann configurations shown in Fig. 1. Assuming the light path denoted by the solid red line in Fig. 1, the electric field of the evanescent light with a wavelength in vacuum of λ, penetrating in the sample, can be denoted as E(λ, z). Here, on the Cartesian coordinates shown on the right, which is on the plane of light incidence perpendicular to the interface, z is the perpendicular distance from the interface between the chip and the sample. The interface between the chip and the sample is set to z = 0. We define the normalized electric field strength squared as
where E0(λ) denotes the electric field of incident light before the light illuminates the prism [19]. In this paper, subscripts wm and sp are used to represent the excitation of WM resonance and SPR, respectively.We reported that Nwm(λ, z) becomes maximum when the thicknesses of two WM excitation layers, consisting of the layer 1 with the real part of its refractive index n1 and the layer 2 with n2, are set so that phase-aligned reflection of light occurs at each interface on the condition that n1>n2 [19]. In this section, λ is 375 nm and TiO2 and SiO2 are assumed to be the materials for WM excitation layers 1 and 2. Under this condition, the incident angle θ to the layer 1 and the thicknesses of two WM excitation layers are numerically optimized to maximize Nwm (375, 0).
Both the prism and the glass substrate are made of silica glass (np = 1.473 and kp = 0.000 [22]) as in the case of our previous work [15–18]. As a sample to be dropped and measured on the chip, water (nw = 1.354, kw = 0.000 [22]) is assumed for simplicity. As shown in Fig. 1, an optical fiber through which the light is transmitted, is attached on the side wall of the prism in parallel to its bottom surface. The two angles ϕ and θ, shown in Fig. 1, satisfy the following equation [16],
Note that θ agrees with the incident angle of the light to the substrate, since the prism and the substrate are of the same material.According to the Snell law, θ is calculated to be 66.8° when the incident angle to the interface between the layer 2 and the water is critical to induce the total internal reflection, using the above-mentioned refractive indices [19]. For increasing Nwm(λ, 0), θ should be as close as possible to but larger than this critical angle to satisfy the total internal reflection condition, even if the refractive index of a sample used in an actual experiment is higher than that of water [19]. Therefore, we decided to use a trapezoid prism with ϕ of 32° for experiments, with which θ is calculated to be 67.1° from Eq. (2).
As a next step, the thicknesses of TiO2 and SiO2 layers must be optimized. For that purpose, we need to know the refractive indices of the TiO2 and the SiO2 layers. Therefore, TiO2 and SiO2 were separately formed on Si single crystal wafers using a radio-frequency magnetron sputter (CFS-4EP-L, Shibaura Mechatronics). The refractive index, n, and the extinction coefficient, k, were measured for TiO2 and SiO2 using a spectroscopic ellipsometer (VASE, J.A. Woollam). As a result, n and k at 375 nm of λ were respectively found to be 2.757 and 0.029 for TiO2 and 1.503 and 0.000 for SiO2.
The WM resonance can be excited by both s- and p-polarized lights, but its electric-field enhancement is higher with s-polarized light than with p-polarized light [19]. Therefore, Nwm(375, 0) was calculated using a transfer matrix method for s-polarized light. Here, the thickness of the TiO2 layer 1 was changed from 0 to 80 nm, while that of the SiO2 layer 2 was changed from 0 to 300 nm. The result is shown in Fig. 2(a) as a function of thicknesses of the two layers. Figure 2(b) is an enlarged view of the area surrounded by a white frame in Fig. 2(a). The color bar, representing the value of Nwm(375, 0), is linearly divided from cold to warm. As shown by the white square in Fig. 2(b), the maximum value of Nwm (375, 0) is 45.3 when the thickness of TiO2 is around 38 nm and that of SiO2 is around 160 nm. The three solid circles in Fig. 2(b) will be mentioned in Subsection 3.1.
Fig. 2 (a) Normalized electric field strength squared, Nwm(375, 0), calculated for the WM chip as a function of thicknesses of the TiO2 layer 1 and the SiO2 layer 2. The base angle ϕ of the prism is set to 32°. Water is put on the layer 2. The color bar is linearly divided from cold to warm, depending on the value of Nwm(375, 0). (b) Enlarged view of the area surrounded by the white frame in Fig. 2(a). The solid red, blue, and black circles denote experimental conditions for the WM chips (i), (ii), and (iii), respectively, while the white square denotes the ideally optimized condition for WM excitation.
For making a comparison with the WM chips, a SPR chip consisting of an Al layer sputtered on a silica glass substrate was examined. Here, the thickness of the Al layer would be changed from 0 to 30 nm and the base angle ϕ of the prism would be changed from 30° to 70°. Further, a naturally grown Al2O3 layer with a thickness of 5 nm is assumed to be formed on the surface of the Al layer [13]. Important parameters, such as the refractive indices of the prism, the substrate, and the water dropped on the chip, are the same as those assumed for the WM chips. The real and the imaginary parts of the complex refractive indices of Al and Al2O3 are assumed to be nAl = 3.467 × 10−1, kAl = 4.535, nAl2O3 = 1.790, and kAl2O3 = 0.000 based on our previous report [13]. The incident light assumed in the calculation is p-polarized light, by which SPR is induced. As shown by the white square in Fig. 3, Nsp(375, 0) reaches its maximum value of 12.6 when the thickness of the Al layer is around 18 nm and ϕ is around 52°. The solid green circle in Fig. 3 will be mentioned in Subsection 3.1.
Fig. 3 Nsp(375, 0), calculated for the SPR chip, as a function of thickness of the Al layer and base angle ϕ of the prism. Water is on a 5-nm Al2O3 layer put on the Al layer. The solid green circle denotes the experimental condition for the SPR chip, while the white square denotes the ideally optimized condition for SPR excitation.
3. Experimental methods
3.1 Preparation of the WM and SPR chips
To demonstrate experimentally the above-mentioned electric-field enhancement of the WM chip, we prepared three WM chips (i), (ii), and (iii) by sputtering TiO2 first as the layer 1 and then SiO2 as the layer 2 on a silica glass plate having dimensions of 18 mm long by 14 mm wide and 1 mm thick. Here, the sputtering rate of TiO2 was confirmed to be 0.063 nm/s under an argon atmosphere of 0.5 Pa and an RF (13.56 MHz) power of 200 W. Therefore, the thickness of TiO2 sputtered for each WM chip was 36.2 nm when its sputtering time was set to be 9 min 35 s. Note that the thickness is shown with an accuracy of 0.1 nm, although such accuracy may not be guaranteed. The reason for describing the thickness with the 0.1-nm accuracy is that we have to discuss the relation between the thickness and the fluorescence intensity of a QD solution with three or four significant digits below in this section and in Section 4.
To analyze the thickness dependence of the electric field enhancement effect, the sputtering time of SiO2 for each chip was changed. The sputtering rate of SiO2 was confirmed to be 0.023 nm/s under the argon atmosphere of 0.5 Pa and the RF power of 400 W. Each sputtering time and the resultant thickness of SiO2 are as follows: 12 min 35 s and 168.4 nm for the WM chip (i), 12 min 50 s and 171.7 nm for (ii), and 13 min 5 s and 175.0 nm for (iii). The thicknesses of TiO2 and SiO2 for the chips (i), (ii), and (iii) are respectively shown by the solid red, blue, and black circles in Fig. 2(b).
In order to make a SPR chip for comparison, Al was sputtered on the same silica glass substrate as that used for the WM chip. The sputtering rate of Al was confirmed to be 0.246 nm/s under the argon atmosphere of 0.5 Pa and the RF power of 200 W. The thickness of the sputtered film was 18.2 nm. Assuming that the thickness of the Al2O3 layer naturally grown on the Al layer was 5.0 nm, the thickness of the Al layer was estimated to be 13.2 nm. The prism, on which the SPR chip was put, has ϕ of 42°. Under these conditions, Nsp(375, 0) was calculated to be 7.8 as shown by the solid green circle in Fig. 3.
3.2 Samples and methods of fluorescence measurements
The intensity of fluorescence appearing in water with or without QDs was measured using the WM and the SPR chips. In Fig. 1, light from an LED light source (M375F2, Thorlabs) illuminates a prism through an optical fiber, a collimator lens, and a polarizer (SPF-30C-32, Sigmakoki). Note that the incident light is s-polarized for the WM chip and p-polarized for the SPR chip. Then, the light reaches the excitation layer of a chip put on a prism. Near-field light with enhanced electric field is generated on the top surface of the chip when WM resonance or SPR is induced, which excites fluorescence in a sample put on the chip. The fluorescence, which passes through an objective lens (M Plan Apo, Mitsutoyo) and an optical filter (FELH0700, Thorlabs), is detected using a CMOS image sensor (BPU-30, Bitran).
In the experiment, the WM or the SPR chip was put on the prism. Then, a sheet of silicone rubber, around 1.0 mm thick with a through hole of around 10 mm in diameter, was put on each chip. De-ionized water of 100 μL was dropped on the surface of the chip, through the hole in the silicone rubber sheet, and the fluorescence intensity was measured by the image sensor with an exposure time of 1 min. Note that the optical filter passes only the light with wavelengths longer than about 700 nm, which prevents the incident light from reaching the image sensor. As a QD sample, a Qdot 705 streptavidin conjugate (Q10163MP, Thermo Fisher Scientific) [23] was diluted with water to make a QD solution with a concentration of 1.0 nM (M = mol/L). Here, the QD emits bright fluorescence with a peak wavelength of 705 nm when excited by UV light.
4. Experimental results
The s- and p-polarized light intensities of the LED before reaching the prism were measured by a spectroscope (USB4000, Ocean Optics). The broken red line in Fig. 4 shows the intensity of the s-polarized light, Is(λ), while the solid blue line shows that of the p-polarized light, Ip(λ). The intensities of the s- and p-polarized lights show exactly the same at any wavelength.
Fig. 4 Light intensity of the LED, measured as a function of λ. Broken red:Is (λ) for s-polarized light, solid blue: Ip(λ) for p-polarized light. Note that the two curves overlap each other.
Figure 5 shows the fluorescence intensities of water or the QD solution measured using the WM chips (i) to (iii) and the SPR chip. Here, the CMOS image sensor used in this study divides the fluorescence intensity to 16-bit levels. The average of relative fluorescence intensities obtained in the whole area of the image, which is 2.5 mm long by 2.0 mm wide, is shown in dimensionless form. The average μw measured for water is shown by the open red circle for the WM chip (i), by the open blue circle for the chip (ii), by the open black circle for the chip (iii), and by the open green circle for the SPR chip. The standard deviation of μw, which is denoted as σw, cannot be seen, because it is smaller than each open circle. The average of fluorescence intensities measured for the QD solution, μq, is also shown in Fig. 5 by the solid red triangle for the WM chip (i), by the solid blue triangle for the chip (ii), by the solid black triangle for the chip (iii), and by the solid green triangle for the SPR chip. The standard deviation of each μq is denoted by σq and is shown by the vertical bar with the same color.
Fig. 5 Fluorescence intensities measured using the WM chips (i) – (iii) and the SPR chip. The solid triangles represent the average of fluorescence intensities from QD solutions μq, while the open circles represent the average from water μw. The standard deviations of fluorescence intensities, represented by vertical bars, are larger for the QD solution than for the water.
The thermal noise by the image sensor and the autofluorescence of the chips become a background fluorescence of water μw during the measurement of fluorescence from biological substances. Here, the ratio μq / μw can be an indicator of a good chip. While μq / μw is 4.5 for the WM chip (i), 3.3 for the chip (ii), and 2.9 for the chip (iii), it is only 1.6 for the SPR chip. In addition, the effective fluorescence intensity of the QD solution after the removal of the background is μq - μw and its standard deviation is , according to the law of propagation of errors [24]. In Fig. 5, μq - μw is calculated as 3820 ± 420, 2520 ± 290, and 2070 ± 330 for the WM chips (i), (ii), and (iii), respectively, and 590 ± 150 for the SPR chip. This means that μq - μw is 6.5 times higher when it is measured by the WM chip (i) than when measured by the SPR chip.
5. Discussion
The above-mentioned experimental values of effective fluorescence intensity of the QD solution, μq - μw, were evaluated by numerical calculations of electric fields. First, N(λ, 0) was calculated using the transfer matrix method as a function of λ between 360 and 400 nm for the WM and SPR chips actually made. In calculations, the incident light was assumed to be s-polarized for the WM chip and p-polarized for the SPR chip. The thickness of each layer was as mentioned in Subsection 3.1. The result is shown in Fig. 6. The solid red, dashed dotted blue, and dashed double-dotted black curves denote respectively Nwm(λ, 0) values calculated for the WM chip (i), (ii), and (iii), while the broken green curve denotes Nsp(λ, 0) calculated for the SPR chip. The peaks in Nwm(λ, 0) arise due to most efficient coupling into the WM at the peak wavelengths.
Fig. 6 Nwm(λ, 0) and Nsp(λ, 0) calculated as a function of λ. Solid red: WM chip (i), dashed dotted blue: WM chip (ii), dashed double-dotted black: WM chip (iii), broken green: SPR chip.
The intensities of the s- and p-polarized lights emitted from the LED, Is(λ) and Ip(λ), vary as a function of λ as shown in Fig. 4. Since the fluorescence efficiency of the QD solution also depends on excitation wavelength λ [23], η(λ) is defined as the relative fluorescence intensity per unit width of λ in nm at a certain λ, measured when one photon illuminates the solution. Then, the relative total fluorescence intensity F of the QD solution in the wavelength range from 360 to 400 nm can be calculated by the equation,
where c is a constant and d(λ) denotes the penetration depth of near-field light, at which the electric field strength decreases to 1/e. For the WM and SPR chips, d(λ) can be calculated by the equations [25],andwhere εAl and εw in Eq. (4b) denote the relative permittivities of Al and water, respectively. The absorption of near-field light by QDs is ignored in Eq. (3), since the absorption coefficient of the QD solution is negligibly small compared to d(λ) [26]. Note that the refractive index of WM excitation layer 2 does not appear in Eq. (4a) as long as all the surfaces of layers 1 and 2 and substrate are parallel.For calculating the integrand in Eq. (3), we used N(λ) shown in Fig. 6, I(λ) shown in Fig. 4, and η(λ) reported in [23] where I(λ) was normalized by its maximum Ip(374) at λ = 374 nm and η(λ) was normalized by its maximum at λ = 360 nm.
The integrand in Eq. (3) was calculated as a function of λ for the WM and SPR chips as shown in Fig. 7. Here, the curves in Fig. 7 correspond to the same chips as those in Fig. 6. The reason for the large difference in the product value among the three WM chips shown in Fig. 7 can be understood by the fact that the WM chip (i) has the highest Nwm at around 374 nm, at which the LED intensity becomes maximum as shown in Fig. 4.
Fig. 7 Integrand in the right side of Eq. (3), d(λ)N(λ, 0)I(λ)η(λ), as a function of λ. To calculate the relative total fluorescence intensity F, each curve is integrated with respect to λ. Solid red: WM chip (i), dashed dotted blue: WM chip (ii), dashed double-dotted black: WM chip (iii), broken green: SPR chip. Only in this figure, λ is treated as dimensionless.
Each curve in Fig. 7 was then integrated with respect to λ from 360 to 400 to calculate F for each chip. Here, λ was treated as dimensionless and c was assumed to be 1 for simplicity. As a result, F was found to be 9.47 × 104 for the WM chip (i), 6.93 × 104 for (ii), 5.17 × 104 for (iii), and 7.63 × 103 for the SPR chip. Figure 8 shows the relation between F and μq - μw. Solid circles denote the values of μq - μw calculated for the WM and SPR chips using the data shown in Fig. 5. The error bars show the above-mentioned standard deviations. The open red and light blue squares will be mentioned later.
Fig. 8 Effective fluorescence intensity, μq - μw, for the WM chips (i) – (iii) and the SPR chip as a function of F. Vertical bars show the standard deviations. Solid circles: measured, open squares: estimated for the ideally optimized chips.
The dotted red line in Fig. 8 is a linearly approximate line obtained by a least square method for the WM chips (i) – (iii), which passes through the origin, since both F and μq - μw should be zero when no near field exists on the surface of the chip. It is clear that μq - μw is in proportion to F almost perfectly. The dotted light blue line is a linearly approximate line for the SPR chip obtained by assuming a linear relationship between μq - μw and F.
Then, μq - μw is estimated for the ideally optimized WM and SPR chips by taking account of the above-mentioned approximate lines. As shown in Fig. 2(b), the optimized WM structure is composed of a 38-nm TiO2 layer 1 and a 160-nm SiO2 layer 2. In this case, F is calculated to be 13.3 × 104 using Eqs. (3) and (4a). Therefore, μq - μw for the ideally optimized WM chip can be estimated to be 5090 as shown by the open red square in Fig. 8.
As for the SPR chip, we can have the optimized structure if the Al layer is 18 nm thick and ϕ is 52° as shown in Fig. 3. Therefore, μq - μw for the optimized SPR structure with F = 1.20 × 104, the value calculated using Eqs. (3) and (4b), becomes 920 as shown by the open light blue square in Fig. 8. That is, μq - μw or the enhancement of the electric field of UV near-field light is 5.5 (= 5090/920) times higher by the optimized WM chip than by the optimized SPR chip. Note that the inclination of the approximate line for WM (dotted red line) is smaller than that for SPR (dotted light blue line). That is, the fluorescence obtained experimentally using the WM chips cannot be enhanced effectively despite the fact that the WM chips have high F. This is probably because it is difficult for the WM chip, which needs two thick layers compared to the SPR chip, to obtain ideally flat surfaces.
In the experiments, the prism made of silica glass with ϕ of 32° was used for the WM chip, while the one with ϕ of 42° was used for the SPR chip. This difference in ϕ changes the light intensity per unit area in the projection of light generated on the surface of the chip by 10%. However, it is much smaller than the above-mentioned ratio, 5.5, of the effective fluorescence intensity. As for the substances used for the prism and substrate, their refractive indices affect the electric field enhancement. For example, when BK 7 (n = 1.534 and k = 0.000 at 375 nm [27]) is used for the prism and substrate as is often the case in SPR sensors, the maximum normalized electric field strength squared is calculated to be Nwm(375, 0) = 46.8 and Nsp(375, 0) = 17.7 by the transfer matrix method, compared to the present values of Nwm(375, 0) = 45.3 and Nsp(375, 0) = 12.6.
The calculation method we used in this study, namely the transfer matrix method, does not take into account the condition of occurrence of SPR directly. Regarding this, we calculated the angle, at which the reflectance of an Al SPR chip becomes minimum by the transfer matrix method. As a result, the angle agrees quite well with the one derived by the ω - k dispersion relation [25]. Therefore, the effect of SPR is incorporated into our calculations substantially.
To summarize, the major reason for the high fluorescence intensity obtained by the WM chip compared to the SPR chip is the electric field enhancement effect of the WM chip, as is demonstrated by the comparison between Fig. 2(b) and Fig. 3. In addition, from Eqs. (4a) and (4b), dwm(λ) is calculated to be 520 – 695 nm, while dsp(λ) is calculated to be 128 – 162 nm. That is, the penetration depth for the WM chip is longer than that for the SPR chip and more QDs can be excited using the WM chip. For these reasons, we have successfully shown that our TiO2/SiO2 WM chip can enhance the electric field of UV near-field light effectively.
6. Conclusions
A WM chip suitable for an ultraviolet near-field fluorescence sensor, which consists of TiO2 and SiO2 layers, has been developed. An Al SPR chip was made for comparison. As a result, the fluorescence intensity from a QD solution dropped on the chip, excited by an LED light with wavelengths from 360 to 400 nm, is about 6.5 times when the WM chip is used compared to the SPR chip. By extrapolating the experimental results, the ratio of fluorescence intensity obtainable by the optimized WM chip to that obtainable by the optimized SPR chip is calculated to be 5.5 times. These findings indicate clearly that WM chips are superior to SPR chips.
Funding
Japan Society for the Promotion of Science (JSPS) Grant-in-Aid for JSPS Fellows (16J10336).
References and links
1. D. Axelrod, “Total internal reflection fluorescence microscopy in cell biology,” Traffic 2(11), 764–774 (2001). [CrossRef] [PubMed]
2. L. Liang, J. Li, Q. Li, Q. Huang, J. Shi, H. Yan, and C. Fan, “Single-particle tracking and modulation of cell entry pathways of a tetrahedral DNA nanostructure in live cells,” Angew. Chem. Int. Ed. Engl. 53(30), 7745–7750 (2014). [CrossRef] [PubMed]
3. H. Yamamura, Y. Suzuki, and Y. Imaizumi, “New light on ion channel imaging by total internal reflection fluorescence (TIRF) microscopy,” J. Pharmacol. Sci. 128(1), 1–7 (2015). [CrossRef] [PubMed]
4. R.-Y. He, C.-Y. Lin, Y.-D. Su, K.-C. Chiu, N.-S. Chang, H.-L. Wu, and S.-J. Chen, “Imaging live cell membranes via surface plasmon-enhanced fluorescence and phase microscopy,” Opt. Express 18(4), 3649–3659 (2010). [CrossRef] [PubMed]
5. E. Chung, Y.-H. Kim, W. T. Tang, C. J. R. Sheppard, and P. T. C. So, “Wide-field extended-resolution fluorescence microscopy with standing surface-plasmon-resonance waves,” Opt. Lett. 34(15), 2366–2368 (2009). [CrossRef] [PubMed]
6. I. L. Medintz, H. T. Uyeda, E. R. Goldman, and H. Mattoussi, “Quantum dot bioconjugates for imaging, labelling and sensing,” Nat. Mater. 4(6), 435–446 (2005). [CrossRef] [PubMed]
7. U. Resch-Genger, M. Grabolle, S. Cavaliere-Jaricot, R. Nitschke, and T. Nann, “Quantum dots versus organic dyes as fluorescent labels,” Nat. Methods 5(9), 763–775 (2008). [CrossRef] [PubMed]
8. M. W. Knight, N. S. King, L. Liu, H. O. Everitt, P. Nordlander, and N. J. Halas, “Aluminum for plasmonics,” ACS Nano 8(1), 834–840 (2014). [CrossRef] [PubMed]
9. K. Diest, V. Liberman, D. M. Lennon, P. B. Welander, and M. Rothschild, “Aluminum plasmonics: optimization of plasmonic properties using liquid-prism-coupled ellipsometry,” Opt. Express 21(23), 28638–28650 (2013). [CrossRef] [PubMed]
10. D. V. Nesterenko and Z. Sekkat, “Resolution estimation of the Au, Ag, Cu, and Al single-and double-layer surface plasmon sensors in the ultraviolet, visible, and infrared regions,” Plasmonics 8(4), 1585–1595 (2013). [CrossRef]
11. I. Gryczynski, J. Malicka, Z. Gryczynski, K. Nowaczyk, and J. R. Lakowicz, “Ultraviolet surface plasmon-coupled emission using thin aluminum films,” Anal. Chem. 76(14), 4076–4081 (2004). [CrossRef] [PubMed]
12. A. Ono, M. Kikawada, R. Akimoto, W. Inami, and Y. Kawata, “Fluorescence enhancement with deep-ultraviolet surface plasmon excitation,” Opt. Express 21(15), 17447–17453 (2013). [CrossRef] [PubMed]
13. H. Ashiba, Y. Sugiyama, X. Wang, H. Shirato, K. Higo-Moriguchi, K. Taniguchi, Y. Ohki, and M. Fujimaki, “Detection of norovirus virus-like particles using a surface plasmon resonance-assisted fluoroimmunosensor optimized for quantum dot fluorescent labels,” Biosens. Bioelectron. 93, 260–266 (2017). [CrossRef] [PubMed]
14. C. Kuroda, R. Iizuka, Y. Ohki, and M. Fujimaki, “Development of a dielectrophoresis-assisted surface plasmon resonance fluorescence biosensor for detection of bacteria,” Jpn. J. Appl. Phys. (to be published).
15. X. Wang, M. Fujimaki, T. Kato, K. Nomura, K. Awazu, and Y. Ohki, “Optimal design of a spectral readout type planar waveguide-mode sensor with a monolithic structure,” Opt. Express 19(21), 20205–20213 (2011). [CrossRef] [PubMed]
16. M. Fujimaki, X. Wang, T. Kato, K. Awazu, and Y. Ohki, “Parallel-incidence-type waveguide-mode sensor with spectral-readout setup,” Opt. Express 23(9), 10925–10937 (2015). [CrossRef] [PubMed]
17. K. Nomura, T. Lakshmipriya, N. Fukuda, X. Wang, and M. Fujimaki, “Fluorescence enhancement by a SiO2-based monolithic waveguide structure for biomolecular detection,” J. Appl. Phys. 113(14), 143103 (2013). [CrossRef]
18. C. Kuroda, Y. Ohki, H. Ashiba, M. Fujimaki, K. Awazu, and M. Makishima, “Design of a sedimentation hole in a microfluidic channel to remove blood cells from diluted whole blood,” Jpn. J. Appl. Phys. 56(3), 037201 (2017). [CrossRef]
19. C. Kuroda, Y. Ohki, and M. Fujimaki, “Optimization of a waveguide-mode sensing chip for an ultraviolet near-field illumination biosensor,” Opt. Express 25(21), 26011–26019 (2017). [CrossRef] [PubMed]
20. O. Arnon and P. Baumeister, “Electric field distribution and the reduction of laser damage in multilayers,” Appl. Opt. 19(11), 1853–1855 (1980). [CrossRef] [PubMed]
21. M. Born and E. Wolf, Principles of Optics. Electromagnetic Theory of Propagation, Interference and Diffraction of Light. Sixth Edition (Cambridge University, 1997).
22. E. D. Palik, G. Ghosh, and T. M. Cotter, Handbook of Optical Constants of Solids (Academic, 1998).
23. Thermo Fisher Scientific HP, https://www.thermofisher.com/order/catalog/product/Q10163MP, (Retrieved on Nov. 8, 2017).
24. J. E. Turner, D. J. Downing, and J. S. Bogard, Propagation of Error (Statistical Methods in Radiation Physics, 2012), pp. 199–214.
25. A. B. Dahlin, Advances in Biomedical Spectroscopy, Plasmonic Biosensors (IOS, 2012).
26. Thermo Fisher Scientific HP, https://www.thermofisher.com/jp/ja/home/references/molecular-probes-the-handbook/tables/extinction-coefficients-of-qdot-streptavidin-conjugates-at-common-wavelengths.html, (Retrieved on Dec. 16, 2017).
27. Schott HP, http://www.schott.com/advanced_optics/japanese/abbe_datasheets/schott-datasheet-n-bk7.pdf, (Retrieved on Feb. 21, 2018).
