Research on an all-medium two-parameter metasurface sensor based on Fano resonance

Abstract

Compared to metallic metasurfaces, all-dielectric metasurfaces exhibit significantly lower losses and offer superior sensing performance. However, most existing metasurfaces are limited to single-parameter sensing and exhibit sensitivity to the incident angle of the light source. In this paper, we design and analyze an all-dielectric dual-parameter sensor that leverages structural asymmetry. The sensor consists of an array of silicon elements arranged on a quartz substrate. This metasurface excites two distinct Fano resonance peaks within the near-infrared spectrum, both corresponding to magnetic dipole modes. These resonance peaks demonstrate excellent polarization insensitivity and enhanced stability under oblique incidence. The study demonstrates that this design achieves outstanding performance in biosolution sensing, with a maximum refractive index sensitivity of 404.43 nm/RIU and a maximum temperature sensitivity of 51.76pm/°C.

1. Introduction

A metasurface is an artificially engineered two-dimensional material with subwavelength features that can interact with incident light waves to produce resonance peaks [1]. These resonance peaks exhibit high sensitivity to variations in environmental parameters such as refractive index [2]and dielectric constant [3], leading to observable shifts in resonance frequency, intensity changes, and phase alterations. With the rapid advancement of nanotechnology, a variety of optical sensors based on metasurfaces have been developed. These sensors can detect material concentrations [4]and monitor the pathological conditions of cells and tissues [5,6]by sensing changes in refractive index and temperature. Metasurface-based sensors provide enhanced light-capturing capabilities and compact designs, making them ideal for label-free, on-chip integration. Moreover, they have found extensive applications across diverse fields including agriculture [7], industry [8], and biomedicine [9–11].

Fano resonance is an asymmetric resonance phenomenon typically caused by the interference between a narrowband resonance and a broadband background [12,13]. This effect can be traced back to the anomalous phenomena observed by Wood in his studies of diffraction gratings [14]. By breaking the symmetry of nanostructures or composing plasmonic nanoclusters, corresponding transmission or reflection spectral resonance peaks can be generated. Examples include asymmetric nanodisks [15], length- and thickness-asymmetric nanorods [16–18], and nanoclusters [19]. For instance, Guo et al. achieved refractive index sensing by introducing asymmetry in the widths of left and right elongated silicon nanowires, thereby exciting quasi-bound states in the continuum [20].

Dielectric metasurface sensors exhibit lower loss compared to their metal counterparts [21–23], resulting in higher quality factor (Q) and figure of merit (FOM) values. This makes them a focal point in current research. Various dielectric structures have been designed, such as dimer semi-elliptical cylindrical shapes [24] and H-shaped configurations [25]. Based on the aforementioned research, investigators have initiated explorations into multi-parameter sensing using metasurfaces, yielding significant advancements. Yin et al. developed an asymmetric herringbone dielectric metasurface sensor composed of silicon and quartz, achieving refractive index sensing with a sensitivity of 232 nm/RIU and temperature sensing with a sensitivity of 63pm/K [26]. Guo et al. introduced a dielectric metasurface featuring an inclined gap silicon wafer on a silicon nitride substrate, which demonstrated dual Fano resonance in the near-infrared spectrum, attaining a temperature sensing sensitivity of 54pm/°C [27]. Beyond temperature and refractive index sensing, researchers have further expanded the functionality of existing refractive index metasurface sensors by applying gas-sensitive films [28]. Zhang et al. coated a methane-sensitive film, thereby enabling gas volume fraction sensing with a sensitivity of −1.44 nm/% [29].

In this paper, we present a dielectric metasurface sensor designed for dual-parameter biosensing, capable of simultaneously detecting refractive index and temperature with considerable angular tolerance. The sensor features a two-layer structure composed of a silicon wafer atop a quartz substrate. By breaking the structural symmetry, two Fano resonance peaks are excited in the near-infrared region, suitable for sensing applications. Multipole decomposition analysis reveals that these resonance peaks are predominantly generated by magnetic dipole excitations. The refractive index sensing sensitivity reaches up to 404.43 nm/RIU, while the temperature sensing sensitivity can reach up to 51.76pm/°C. These results indicate that the proposed metasurface sensor holds significant potential for complex biological solution sensing, including blood analysis and cancer cell detection. This design offers new insights into multi-parameter sensing, nonlinear optical devices, and optical switches.

2. Model and simulation

Silicon and silica materials exhibit advantages such as low loss, facile processing, and easy integration within the near-infrared wavelength range. The metasurface structure designed in this paper comprises silicon and quartz, with refractive indices of n_Si = 3.45 and n_SiO2 = 1.45, respectively, as referenced in the Palik handbook [30]. Simulations were conducted using CST software based on the Finite Integration Technique (FIT). Periodic boundary conditions were applied in the x and y directions (unit cell), while open boundary conditions were set in the z direction to simulate an infinite extension along this axis. The surrounding environment was modeled as water with a refractive index of n_water = 1.33. A plane wave polarized in the y-direction was incident normally on the metasurface chip along the z-axis. The schematic diagram of the optimized chip structure is shown in Fig. 1, with initial simulation parameters set as p_x = 883 nm, p_y = 900 nm, L = 800 nm, w₁ = 380 nm, w₂ = 200 nm, m = 100 nm, and h = 200 nm. An asymmetry factor δ was introduced, with a value of δ=q. As illustrated in Fig. 2(a), when the structure is symmetric (q = 0 nm), no resonance peaks are observed across the entire wavelength band. When q = 50 nm, two Fano resonance dips, labeled as Dip 1 and Dip 2, are excited at wavelengths of λ = 1478 nm and λ = 1527 nm, respectively. This phenomenon arises from breaking the structural symmetry, which introduces trapped modes. The coupling and interference between these trapped modes and radiative modes result in the formation of Fano resonances.

Fig. 1.

(a) Schematic diagram of the metasurface structure. (b) Diagram of a single unit cell of the structure. (c) Top view of the all-dielectric metasurface unit cell.

Fig. 2.

(a) Transmission spectrum of the symmetric structure (q = 0 nm). (b) Transmission spectrum of the asymmetric structure (q = 50 nm).

3. Result and disscussion

Fano resonance is an asymmetric resonance phenomenon typically caused by the interference between a narrow-band resonance and a broad-band background. Its functional form can be expressed as [31,32]:

$$T(\omega )=T_0+A_0\frac{{[p+2(\omega -{\omega }_0)/\mathrm{\Gamma }]}^2}{1+{[2(\omega -{\omega }_0)/\mathrm{\Gamma }]}^2}$$ (1)

In this expression, T₀ is the scattering coefficient, A₀ is the coupling coefficient, and the parameter Γ denotes the resonance linewidth, specifying the width of the resonant Dips. The parameter p represents the asymmetry parameter of the resonance peak [12]. By modifying the structural parameter q, the asymmetry factor δ can be influenced, thereby altering the mode coupling strength and indirectly affecting the value of p.ω₀ is the resonance frequency. It is defined as 2πc/λ₀, where c denotes the speed of light and λ₀ represents the wavelength corresponding to the position of the resonance peak. Consequently, the values of other parameters can be fitted using the constant λ₀ in the Eq. (1). The Fig. 3(a) and (b) show the Fano fit for two resonance peaks, where the black dots represent the simulation data, and the red curve represents the fitted curve.

Fig. 3.

(a) Simulation results and fitting curve of the Dip 1 transmission spectrum for the asymmetric structure (q = 50 nm). (b) Simulation results and fitting curve of the Dip 2 transmission spectrum for the asymmetric structure (q = 50 nm). (c) Transmission spectra under different symmetry factors δ. (d)2D color map of transmission spectra.

We analyzed the resonance peaks generated by the proposed metasurface. The modulation depth of the spectrum is an important parameter for sensors and filters, defined as [33]:

$$T_d=[(T_{peak}-T_{dip})/(T_{peak}+T_{dip})]\times 100\mathrm{\%}$$ (2)

From our calculations, the modulation depths of peaks Dip 1 and Dip 1 are 74.72% and 99.29%, respectively. As shown in the Fig. 3(c)-(d), the asymmetric parameter δ significantly impacts the transmission peaks. When δ≠0, transmission peaks are excited, and as asymmetry increases, both the modulation depth and full-width at half-maximum (FWHM) of the left resonance peak Dip 1 increase. Meanwhile, the linewidth of the right resonance peak Dip 2 broadens, while its modulation depth remains essentially unchanged. Additionally, the quality factor Q is another critical parameter for evaluating the sensing performance of the metasurface, defined as the ratio of the resonance wavelength λ to the peak’s FWHM [34]. Its expression is:

$$Q=\lambda /FWHM$$ (3)

The relationship between the metasurface quality factor and the asymmetry factor can be mathematically expressed as Q∝A/δ² [24]. In this metasurface design, we sacrificed the quality factor Q to enhance the coupling between the radiative and trapped modes. When δ is 10 nm, the quality factor Q of the designed metasurface reaches 1710.87. When δ is 50 nm, the quality factors of Dip1 and Dip2 are 170.4 and 332.25, respectively.

To further investigate the mechanism behind the formation of the two resonance peaks, we conducted an electromagnetic field analysis on the metasurface. As shown in the Fig. 4, Fig. 4(a)-(c) depict the electric field distribution in the x-y plane for the symmetric structure (δ=0 nm) and the asymmetric structure (δ=50 nm), corresponding to the two resonance peaks, Dip 1 and Dip 2. Figure 4(d)-(f) illustrate the magnetic field distribution in the y-z plane for the respective configurations. The black and white arrows indicate the directions of the electromagnetic field vectors. In the symmetric structure, the electromagnetic field is primarily influenced by electric dipoles. After introducing the asymmetry factor, the original electromagnetic field configuration is disrupted. At the two square holes above and below the notch, the electric field vector direction forms circular rings in the x-y plane (shown in Fig. 4(b) and (c)). Consequently, two magnetic dipoles are formed in the y-z plane of the metasurface (shown in Fig. 4(e) and (f)).

Fig. 4.

(a) Electric field distribution in the x-y plane for the symmetric structure. (b), (c) Electric field distribution in the x-y plane for the asymmetric structure (q = 50 nm) at two resonance wavelengths. (d)-(f) Corresponding magnetic field distribution in the y-z plane.

Subsequently, we conducted a multipole decomposition of the electromagnetic field within the structure using a Cartesian coordinate system. We considered only the five commonly used types of multipoles: electric dipole (ED), magnetic dipole (MD), toroidal dipole (TD), electric quadrupole (EQ), and magnetic quadrupole (MQ) [35–39]. Contributions from higher-order multipoles were neglected [24,36]. The corresponding formula is as follows:

$$ED:P=\frac{1}{i\omega }\int j{d}^{3}r$$ (4)

$$MD:M=\frac{1}{2c}\int (r\times j)d^{3}r$$ (5)

$$TD:T=\frac{1}{10c}\int [(r\cdot j)r-2r^{2}j]d^3$$ (6)

$$EQ:Q_{\alpha \beta }^{(e)}=\frac{1}{2i\omega }\int [(r_{\alpha }{j}_{\beta }+r_{\beta }{j}_{\alpha })-\frac{2}{3}(r\cdot j){\delta }_{\alpha ,\beta }]d^{3}r$$ (7)

$$MQ:Q_{\alpha \beta }^{(m)}=\frac{1}{3c}\int [{(r\times j)}_{\alpha }{r}_{\beta }+{(r\times j)}_{\beta }{r}_{\alpha }]d^{3}r$$ (8)

Here, ω represents the angular frequency, c is the speed of light, j denotes the displacement current density, α,β=x,y, and r is the position vector. The far-field radiated power from different multipoles can be calculated using the following formula:

$$I=\frac{2{\omega }^4}{3c^3}|P{|}^2+\frac{2{\omega }^4}{3c^3}|M{|}^2+\frac{{\omega }^6}{5c^5}|Q_{\alpha \beta }^{(e)}{|}^2+\frac{{\omega }^6}{20c^5}|Q_{\alpha \beta }^{(m)}{|}^2+\frac{2{\omega }^6}{3c^5}|T{|}^2$$ (9)

Based on the aforementioned theory and combined with simulation data, the results shown in Fig. 5 can be obtained, which represent the far-field scattering energy of different multipoles. It is evident that the primary contributors to the resonance peaks Dip 1 and Dip 2 are magnetic dipoles, consistent with the previous analysis.

Fig. 5.

(a) Multipole decomposition of the symmetric structure. (b) and (c) Multipole decomposition of the asymmetric structure (q = 50 nm) at the two resonance wavelengths.

We analyzed the impact of different structural parameters on the transmission peaks. After calculating the numerical shifts of the metasurface resonance peaks under varying structural parameters, we summarize the fitting relationship between them in Table 1. As shown in Fig. 6(a), when the period p_x increases, the position of the Dip 1 remains relatively unchanged, while the Dip2 exhibits a redshift. Furthermore, as the width w₂ of the silicon inner layer increases, the transmission spectrum shows an overall linear blueshift (Fig. 6(b)). Additionally, as the width w₁ of the silicon outer layer, the long edge L, the period p_y, and the silicon height h increase, the transmission spectrum demonstrates an overall linear redshift (Fig. 6(c)–(f)). By varying these structural parameters, the modulation depth, resonance position, and linewidth of the transmission spectrum all exhibit changes to varying extents, while the overall linearity of the resonance peaks remains consistent. From this analysis, it can be concluded that this metasurface structure exhibits significant tunability and a certain level of fabrication tolerance.

Table 1. Linear Fitting Results of Different Structural Parameters and Resonance Peak Positions (y = kx + b).

	px	py	w1	w2	h	L
R2Dip 1	0.9253	0.9869	0.9989	0.9904	0.9971	0.9970
conclusion	Linearity	Linearity	Linearity	Linearity	Linearity	Linearity
R2Dip 2	0.9998	0.9661	0.9996	0.9992	0.9992	0.9992
conclusion	Linearity	Linearity	Linearity	Linearity	Linearity	Linearity

Fig. 6.

(a) The effect of changing the period p_x on the transmission spectrum curve of the Dips.(b) The effect of changing the width w₂ on the transmission spectrum curve of the Dips.(c)-(f) The effect of changing w₁, L, p_y, and h on the transmission spectrum curve of the Dips.

When the refractive index or temperature of the environment in which the metasurface is located undergoes a slight change, the corresponding transmission spectrum also changes. The metasurface chip designed in this paper can be used for refractive index and temperature sensing. We performed a detailed numerical simulation to analyze the sensing performance of the metasurface. When the temperature around the sensing chip changes, the properties of its constituent materials (such as refractive index) will also undergo slight changes. Therefore, in sensing applications of metasurfaces in complex environments, both temperature and refractive index need to be detected simultaneously.

First, we investigated the scenario where the metasurface is used solely for refractive index sensing. It is assumed that the refractive index of different gases surrounding the metasurface chip ranges from 1.00 to 1.10 [40]. Figure 7(a) and 7(b) show the transmission spectra of the metasurface with various surrounding gases, along with the fitting analysis results of the refractive index change and the corresponding resonance peak shift. The refractive index sensitivity S_n of the metasurface sensor can be defined as the change in resonance wavelength per unit refractive index change, expressed as S_n =Δλ/Δn [41]. The gas refractive index sensing sensitivities of the two resonance peaks of the sensor are S_Dip1 = 267.70 nm/RIU and S_Dip2 = 214.76 nm/RIU, respectively. According to relevant literature [42,43], we increased the background refractive index from 1.3 to 1.4 to characterize the refractive index sensing performance of the metasurface for different biological liquids. The refractive indices of plasma, white blood cells, hemoglobin, and red blood cells are assumed as 1.35, 1.36, 1.38, and 1.40, respectively. The transmission spectra and fitting results are shown in the Fig. 8. The liquid refractive index sensing sensitivities of the two resonance peaks of the sensor are S_Dip1 = 404.43 nm/RIU and S_Dip2 = 341.14 nm/RIU, respectively.

Fig. 7.

(a) Variation in the transmission spectrum as the background refractive index ranges from 1.0 to 1.1. (b) Wavelength shift under different air refractive indices.

Fig. 8.

(a) Variation in the transmission spectrum as the background refractive index ranges from 1.3 to 1.4. (b) Wavelength shift under different biological solution refractive indices.

Next, we analyzed the performance of the metasurface when used solely for temperature sensing. Given the small size of the designed metasurface, only the thermo-optic coefficient of the material is considered, while its thermal expansion coefficient is neglected. The relationship between the material's refractive index and temperature can be expressed as [44]:

$$n(T)=n(T_0)-\eta (T-T_0)$$ (10)

where n(T) is the refractive index at temperature T, n(T₀) is the refractive index at the reference temperature T₀, and η is the thermo-optic coefficient. Assuming the reference temperature is 20°C, the background environment for the analysis is set as water.the thermo-optic coefficients for silicon, quartz and water are given η_Si = 1.84 × 10⁻⁴/K [45],η_SiO2 = 8.6 × 10⁻⁶/K [46]and η_water=–1.02 × 10⁻⁴/K [27]. Figure 9 shows the temperature sensing performance of the metasurface between 20°C and 80°C. As the temperature increases, the resonance peaks Dip1 and Dip2 experience redshifts of 2.91 nm and 3.08 nm, respectively. The temperature sensitivity of the metasurface sensor S_T can be defined as S_T =Δλ/ΔT, where ΔT represents the change in temperature. The temperature sensing sensitivities are 47.33pm/°C and 51.76pm/°C, demonstrating that this all-dielectric sensor exhibits excellent temperature sensing performance.

Fig. 9.

(a) Transmission spectrum shift of Dip1. (b) Transmission spectrum shift of Dip2.

When the metasurface is used for dual-parameter sensing of refractive index and temperature, it is necessary to establish a sensitivity matrix to account for the refractive index and temperature at different wavelength shifts. The sensitivity matrix can be expressed as:

$$S=\left(\begin{array}{cc}{S}_{n_1}& S_{T_1}\\ S_{n_2}& S_{T_2}\end{array}\right)$$ (11)

where S_n1 and S_T1 are the refractive index sensitivity and temperature sensitivity for Dip1, respectively, and S_n2 and S_T2 are the refractive index sensitivity and temperature sensitivity for Dip2, respectively.

Therefore, the relationship between the wavelength shift of the resonance peak and the changes in refractive index and temperature can be expressed as:

$$\left[\begin{array}{l}\mathrm{\Delta }{\lambda }_1\\ \mathrm{\Delta }{\lambda }_1\end{array}\right]=A\left[\begin{array}{l}\mathrm{\Delta }n\\ \mathrm{\Delta }T\end{array}\right],A=\left[\begin{array}{l}404.43nm\cdot RI{U}^{-1}47.33pm{\cdot }^{\circ }{C}^{-1}\\ 314.14nm\cdot RI{U}^{-1}51.76pm{\cdot }^{\circ }{C}^{-1}\end{array}\right]$$

After the inverse operation, the dual-parameter variations of the refractive index and temperature in the measured environment can be simultaneously obtained. Next, we analyze and verify the decoupling error of the sensitivity matrix when the metasurface is used for dual-parameter sensing of temperature and refractive index based on the condition number. In Table 2, seven sets of data are selected for error analysis. Here, Δn and ΔT denote the initial refractive index and temperature changes that we set. Δλ₁ and Δλ₂ represent the resonance peak wavelength shifts for Dip1 and Dip2, respectively, as calculated through CST simulation software. Δn_A and ΔT_A correspond to the changes predicted based on sensitivity matrix theory. The refractive index and temperature errors, δ_n and δ_T, can be defined as [47]:

$${\delta }_n=[(\mathrm{\Delta }{n}_A-\mathrm{\Delta }n)/\mathrm{\Delta }n]\times 100\mathrm{\%},{\delta }_T=[(\mathrm{\Delta }{T}_A-\mathrm{\Delta }T)/\mathrm{\Delta }T]\times 100\mathrm{\%}$$

Table 2. Comparison of Errors Between Sensitivity Matrix Results and Predicted Values.

Δn	ΔT	Δλ1	Δλ2	ΔnA	ΔTA	δn/%	δT/%
0.02	20	11.09	9.54	0.0202	19.938	1.00	−0.31
0.02	40	12.38	11.09	0.0191	39.902	−4.50	−0.25
0.04	20	20.83	17.62	0.0403	19.904	0.75	−0.05
0.04	60	22.53	19.53	0.0399	59.065	−0.50	−0.22
0.05	60	27.07	23.08	0.0509	59.083	1.80	−0.02
0.07	40	34.83	29.22	0.0691	39.055	−1.26	−0.36
0.07	60	35.56	30.12	0.0684	59.033	−2.29	−0.28

As shown, the discrepancy between the values predicted by the sensitivity matrix theory and the set changes is relatively small, thereby confirming the accuracy of the proposed model structure.

As illustrated in Table 3, the metasurface designed in this study is compared with the state-of-the-art in published metasurface sensor research [48]. This design achieves simultaneous sensing of temperature and refractive index, demonstrating superior sensing performance.

Table 3. The comparison between the metasurface introduced in this paper and conventional sensors.

Sensor Types	RI Detection range	Temperature Detection range	Sensitivity to the RI	Sensitivity to the temnperature	Qmax	Ref.
Hollow cylindrical metasurface	1.33-1.36	0-50°C	540 nm/RIU	30pm/°C	1024	[7]
Dimer semi-elliptical cylindrical metasurface	1-1.4	none	405 nm/RIU	none	319	[24]
Asymmetric dimerall-dielectrie metastructure	1-1.4	293.15-323.15K	279 nm/RIU	60pm/K	9352	[25]
Hollow herringbone metastructure	1.33-1.4	5-50K	232 nm/RIU	63 pm/K	5366	[26]
Asymmetric dimer square metasurface	0-70%salinity	20-70°C	327.85 nm/RIU	50pm/°C	107	[48]
An Eight-shaped Metasurface	1-1.4	20-80°C	404.43 nm/RIU	51.76 pm/°C	1710	This work

In the experiment, the incident light on the chip often deviates from strict vertical alignment due to the choice of light source and the setup of the experimental optical path. This can lead to issues such as distorted transmission spectra and weakened resonance peak intensities for sensing. We analyzed the stability of the transmission spectrum with respect to light source angle tolerance by numerical simulation analysis. The initial configuration assumes that the light source excites the transmission spectrum along the z-axis, where θ is the angle between the incident light and the z-axis. Simulation results demonstrate consistent behavior for both negative and positive angles. Therefore, only the results for positive angles are presented in Fig. 10(a). Figure 10(b) and (c) display the relationships between the transmission peak position, full width at FWHM, and transmission intensity T. For the Dip1, a ± 5° angular deviation results in a maximum shift of 1.2 nm in the peak position, a maximum change of 0.3 nm in the FWHM, and a maximum modulation depth change ΔT_max of 3.33%. For the Dip2, as the angle θ increases, the modulation depth gradually approaches 100%. At θ = 8°, the FWHM changes by up to 20.1 nm. Despite experiencing a more significant change in FWHM compared to the left-side peak, no new spurious non-sensing resonant peaks are generated in the band corresponding of Dip2. Figure 11 and Fig. 12 present the 2D color maps of the transmission spectrum and the scattering intensity contributions of each dipole at different incident angles θ for Dip1 and Dip2, respectively. It is evident that within an 8° angle range, the magnetic dipoles MD of both resonant peaks remain dominant, and this dominance persists over time. As the incident angle increases, the MD dipole contribution for Dip1 remains at 90%, leading to minimal variation in the overall trend of this resonant peak. For Dip2, however, the MD dipole contribution decreases to 82%, and this reduction affects the coupling relationship between the metasurface's bright and dark modes, thereby resulting in an increase in the linewidth of Dip2. This analysis aligns with the observations in Fig. 10. Consequently, this structure demonstrates good stability with respect to incident light angle tolerance for sensing applications.

Fig. 10.

(a) Variation in the transmission spectrum with different incident angles θ. (b) and (c) Changes in the full width at FWHM, resonance wavelength, and modulation depth for the two resonance Dips.

Fig. 11.

(a) 2D color map of Dip1 transmission spectra at various angles θ. (b) Comparison of the scattering intensity contributions from Dip1 poles at different angles θ.

Fig. 12.

(a) 2D color map of Dip2 transmission spectra at various angles θ. (b) Comparison of the scattering intensity contributions from Dip2 poles at different angles θ.

We investigated the polarization characteristics of the proposed metasurface sensor. Figure 13 illustrates the variations in the transmission spectrum of the metasurface at various polarization angles. As the polarization angle, Φ, varies from 0° to 90°, corresponding to the transition of the incident light from y-polarized to x-polarized, the intensity of the Dip 2 gradually decreases until it vanishes, while the intensity of the Dip1remains nearly constant. We analyze the electric field distribution in the x-y plane of the two formant peaks at z = 100 nm (Fig. 14 and Fig. 15). As the polarization angle increases, the electric field intensity of Dip1 changes little. At the same time, the electric field strength of Dip2 gradually decreases, which means that the energy confined to the metasurface gradually decreases. When Φ=75°, the intensity of Dip2 electric field reaches the minimum, and the formant disappears. Therefore, Dip1 exhibits a certain degree of polarization insensitivity. Dip2, on the other hand, holds potential for applications as a polarization-dependent optical switch in the corresponding wavelength range. By altering the polarization state of the incident light, the generation and disappearance of Dip2 can be controlled, thereby functioning as an optical path switch and closure at the corresponding wavelength.

Fig. 13.

Variation of the Transmission Spectrum with Different Polarization Angles Φ.

Fig. 14.

Electric field distribution of Dip1 in the x-y plane at various polarization angles Φ (z = 100 nm).

Fig. 15.

Electric field distribution of Dip2 in the x-y plane at various polarization angles Φ (z = 100 nm).

We performed a feasibility analysis of the fabrication process for the proposed metasurface sensor chip. The overall production process is illustrated in Fig. 16. First, the quartz substrate is cleaned using an acetone solution to ensure that the surface is free from impurities. A silicon film is deposited on the quartz substrate via Low Pressure Chemical Vapor Deposition (LPCVD), and the film thickness is verified with an ellipsometer to ensure it meets the design specifications. Given that the minimum linewidth of the designed structure is 50 nm, AR-PC conductive glue and ZEP520 photoresist are applied, followed by electron beam lithography (EBL) to transfer the pattern of the silicon pillars. After development, reactive ion etching (RIE) is performed using a CF₄ and SF₆ gas mixture to etch the silicon pillars. Finally, the sample is heated and immersed in an acetoDne solution to remove the photoresist and clean the sample, thereby obtaining the desired metasurface structure. After fabricating the metasurface, it is necessary to employ a Scanning Electron Microscope (SEM) to observe and measure whether parameters such as the roughness of the silicon layer and the sidewall angle conform to the design specifications.

Fig. 16.

Metasurface Preparation Flow Chart.

4. Conclusions

In summary, we have theoretically and numerically investigated an asymmetric metasurface sensor composed of silicon plates and the quartz substrate. After introducing symmetry breaking, two Fano resonance peaks are excited in the near-infrared band. Through multipole analysis, it is found that the resonance peaks at 1478 nm and 1527 nm are both generated by MD. In this design, we sacrificed the Q factor of the metasurface and increased the asymmetry parameters to achieve stronger coupling between the bright and dark modes. By adjusting the structural parameters, the position and linewidth of the resonance peaks can be flexibly tuned, making the metasurface highly tunable. Near-infrared spectral simulations show that the design exhibits a certain degree of polarization insensitivity and tolerance to light source angle, achieving resonance peak stability at a z-axis incident angle of ±8°. In terms of applications, this design can be used for dual-parameter sensing of refractive index and temperature. For temperature sensing, the sensitivity can reach up to 51.76pm/°C, and for refractive index sensing, the sensitivity can reach 404.43 nm/RIU. This metasurface is compatible with current semiconductor chip fabrication processes, offering a new design method for high-sensitivity multi-parameter sensors and optical switches. Additionally, it expands the potential applications of metasurface in areas such as biological solution identification, substance concentration measurement, and environmental monitoring.

Funding

Capital construction funds in Jilin Province ( 2022C026); Changchun science and technology development plan project ( 22SH03); Jilin province and Chinese Academy of Sciences Science and Technology Cooperation High Tech Special Fund project ( 2022SYHZ0008, 2023SYHZ0047, 2023SYHZ0020); Jilin Province Science and Technology Development Plan Project ( 20220204079YY, 20220201060GX, 20230204095YY, 20230508038RC, 20230201045GX, 20240402029GH, 20240601051RC, YDZJ202401310ZYTS); Tianjin Municipal Science and Technology Program 10.13039/501100019065 ( 23YFYSHZ00300); Scientific and Technological Innovation Project of Black Land Protection and Utilization ( XDA28050201); National Natural Science Foundation of China 10.13039/501100001809 ( 42377037); Provincial and local cooperation Fund - Changchun Joint Fund for Innovation and Development ( YDZJ202401310ZYTS).

Disclosures

The authors declare no conflicts of interest.

Data availability

Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.