Infrared spectrometer

Abstract

To achieve high spectral performance [Solution] The infrared spectrometer according to the embodiment comprises a diffraction grating structure, a metal thin film, and a current detection unit. The diffraction grating structure includes continuous convex and concave portions. A metal thin film is formed on the surface of the diffraction grating structure so that surface plasmon resonance can be excited when infrared light is incident at a specific incident angle. The current detection unit detects the current generated by the surface plasmon resonance. The pitch of the uneven structure in the diffraction grating structure is 1.52 times or more and 2.17 times or less the width of the convex portion.

Description

[Technical Field]

[0001] This invention relates to an infrared spectrometer. [Background technology]

[0002] Patent Document 1 describes an electromechanical surface plasmon resonance sensor in which an electrode, a silicon semiconductor film, and a plasmon resonance electrode are arranged in this order, and a prism is placed on top of them to measure current or voltage.

[0003] In contrast, a spectroscopic sensor that utilizes surface plasmon resonance generated by a diffraction grating structure can also be considered. In a spectroscopic sensor that utilizes surface plasmon resonance generated by a diffraction grating structure, for example, a diffraction grating structure (a bumpy structure) is included on the silicon surface, and a thin film of gold, which is a plasmon resonance material, is formed on top of it. When light of a specific wavelength is incident at a specific angle, plasmon resonance occurs, and energy is transferred to electrons inside the gold. A current detection unit detects the current generated when this energy crosses the Schottky barrier between silicon and gold.

[0004] There is a certain relationship between the wavelength and angle at which surface plasmon resonance occurs. Therefore, by irradiating a sensor whose angle has been changed by vibration, measuring the current value with respect to the angle, and converting the measured current value into wavelength and the intensity of incident light at that wavelength, spectroscopic analysis can be performed. Specifically, wavelength-incident light intensity data can be obtained by applying the inverse matrix of a matrix called the device characteristic matrix to the angle-current value data obtained during actual measurement. Here, the device characteristic matrix is ​​data created from the wavelength of the incident light, the intensity of the incident light, the angle, and the current value measured in advance, and is device-specific data.

[0005] Here, parameters that affect the quality of the device characteristic matrix include the fabrication accuracy of the diffraction grating structure, the Schottky barrier height accuracy, and also the pitch width and the width of the convex portion of the diffraction grating structure. [Prior art documents] [Patent Documents]

[0006] [Patent Document 1] Patent No. 7313607 [Overview of the project] [Problems that the invention aims to solve]

[0007] However, in diffraction grating-type plasmon resonance spectroscopic sensors, the relationship between the diffraction grating structure and the quality of the device characteristic matrix, as well as the relationship between the diffraction grating structure and the intensity of the plasmon resonance signal, remained unclear, and the structure that yielded high spectral performance was unknown.

[0008] The present invention has been made in view of the above, and provides a diffraction grating type plasmon resonance spectrometer that has high spectral performance. [Means for solving the problem]

[0009] To solve the above-mentioned problems and achieve the objective, the infrared spectrometer of the embodiment comprises a diffraction grating structure, a metal thin film, and a current detection unit. The diffraction grating structure includes continuous convex and concave portions. The metal thin film is formed so that surface plasmon resonance can be excited when infrared light is incident on the surface of the diffraction grating structure at a specific incident angle. The current detection unit detects the current generated by the surface plasmon resonance. The pitch of the uneven structure in the diffraction grating structure is 1.52 times or more and 2.17 times or less the width of the convex portion. [Effects of the Invention]

[0010] According to the infrared spectrometer of the present invention, the plasmon resonance signal caused by second-order diffracted light can be reduced, and high spectral performance can be achieved. [Brief explanation of the drawing]

[0011] [Figure 1] Figure 1 shows an example of the structure of a diffraction grating type plasmon resonance sensor according to an embodiment. [Figure 2] Figure 2 illustrates diffraction due to the diffraction grating structure. [Figure 3] Figure 3 illustrates the operating principle of the plasmon resonance sensor according to this embodiment. [Figure 4] Figure 4 illustrates the conversion of wavelength-intensity data and angle-current value data using a device characteristic matrix. [Figure 5] Figure 5 shows an example of the spectral characteristics of a plasmon resonance sensor according to this embodiment. [Figure 6] Figure 6 shows the change in signal intensity for each diffraction order of the plasmon resonance sensor according to the embodiment when the value of d1 (width of the protrusion) / d2 (pitch of the uneven structure) is changed. [Figure 7] Figure 7 is a flowchart illustrating the process for calculating the device characteristic matrix in the plasmon resonance sensor according to this embodiment. [Figure 8] Figure 8 is a flowchart illustrating the process for calculating the device characteristic matrix in the plasmon resonance sensor according to this embodiment. [Figure 9] Figure 9 is a flowchart illustrating the calculation of the device characteristic matrix when d2 (pitch of the uneven structure) = 2600 nm and d1 (width of the protrusions) is varied to 866 nm, 1200 nm, 1300 nm, and 1400 nm. [Figure 10] Figure 10 is a flowchart illustrating the calculation of the device characteristic matrix when d2 (pitch of the uneven structure) = 2600 nm and d1 (width of the protrusions) is varied to 1500 nm, 1600 nm, 1700 nm, and 1950 nm. [Figure 11] Figure 11 is a flowchart illustrating the process for calculating an index value representing the spectral performance of a given diffraction grating structure. [Figure 12] Figure 12 illustrates the calculation of an index value representing the spectral performance of a plasmon resonance sensor according to this embodiment. [Figure 13]FIG. 13 is a flowchart for explaining the reconstruction of wavelength-intensity data when d2 (pitch of the concavo-convex structure) = 2600 nm and d1 (width of the convex portion) is changed to 866 nm, 1200 nm, 1300 nm, and 1400 nm. [Figure 14] FIG. 14 is a flowchart for explaining the reconstruction of wavelength-intensity data when d2 (pitch of the concavo-convex structure) = 2600 nm and d1 (width of the convex portion) is changed to 1500 nm, 1600 nm, 1700 nm, and 1950 nm. [Figure 15] FIG. 15 is a diagram showing the residual for each wavelength when the wavelength-intensity data is reconstructed by changing d1 (width of the convex portion) to 866 nm, 1200 nm, 1300 nm, and 1400 nm at d2 (pitch of the concavo-convex structure) = 2600 nm. [Figure 16] FIG. 16 is a diagram showing the residual for each wavelength when the wavelength-intensity data is reconstructed by changing d1 (width of the convex portion) to 1500 nm, 1600 nm, 1700 nm, and 1950 nm at d2 (pitch of the concavo-convex structure) = 2600 nm. [Figure 17] FIG. 17 is a diagram showing an example of the spectroscopic characteristics of the plasmon resonance sensor according to the embodiment when d2 (pitch of the concavo-convex structure) is changed to 1000 nm, 1100 nm, 1200 nm, 1250 nm, 1300 nm, and 1400 nm. [Figure 18] FIG. 18 is a diagram showing an example of the spectroscopic characteristics of the plasmon resonance sensor according to the embodiment when d2 (pitch of the concavo-convex structure) is changed to 2600 nm, 2800 nm, and 3900 nm.

Mode for Carrying Out the Invention

[0012] (Embodiment) Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[0013] Figure 1 shows an example of the structure of a diffraction grating type plasmon resonance sensor 1 according to an embodiment. The plasmon resonance sensor 1, as an example, consists of a diffraction grating structure 2 including continuous convex and concave portions, a metal thin film 10 formed on the surface of the diffraction grating structure 2, a semiconductor layer 11 having the diffraction grating structure 2, and a current detection unit 12 that detects the current generated by surface plasmon resonance generated by the metal thin film 10.

[0014] The diffraction grating structure 2 includes continuous convex and concave portions and has a grating structure. In Figure 1, the pitch 21(d2) of the convex and concave structure is selected to be, for example, 1250 nm to 3900 nm, and more preferably, 1400 nm to 2800 nm. The width 20(d1) of the convex portions is selected to be, for example, 1250 nm to 3900 nm, and more preferably, 1300 nm to 1600 nm.

[0015] The pitch 21(d2) of the uneven structure is selected to be, for example, 1.52 to 2.17 times the width 20(d1) of the protrusions, and more preferably, 1.62 to 2.0 times. As an example, twice the width 20(d1) of the protrusions (when measured including the thickness of the metal thin film) is selected as the pitch 21(d2) of the uneven structure. Alternatively, twice the width 22 of the protrusions (when measured without including the thickness of the metal thin film) may be selected as the pitch 21(d2) of the uneven structure. The appropriate ranges for the pitch 21 of the uneven structure, the width 20 of the protrusions, and the ratio of the pitch 21 of the uneven structure to the width 20 of the protrusions will be described later.

[0016] The grating structure of the reflective diffraction grating structure 2 causes interference in the reflected light. Figure 2 illustrates the diffraction caused by the diffraction grating structure 2. Consider the case where monochromatic incident light 30 is incident on the diffraction grating structure 2 at an incident angle 36 (angle θ), the incident light 30 is reflected by the diffraction grating structure 2, and reflected light 31 is produced. At this time, the reflected light 31 is also diffracted light with diffraction order m=0. In addition, the diffraction grating structure 2 produces +1st order diffracted light 33 with diffraction order m=+1, +2nd order diffracted light 35 with diffraction order m=+2, diffracted light 32 with diffraction order m=-1, and diffracted light 34 with diffraction order m=-2, etc.

[0017] Here, for example, as in the diffracted light 35 in Figure 2, if the reflection of a specific order of diffraction is in the direction of the surface of the diffraction grating structure, i.e., horizontal in Figure 2, then surface plasmon resonance occurs for reflected light of a specific wavelength, depending on the type of atoms constituting the metal thin film 10. Specifically, the following equation (1) holds true.

[0018]

number

[0019] Here, θ is the angle of incidence, ε is the dielectric constant of the metal thin film 10, n is the refractive index of air, d2 is the pitch 21 of the uneven structure, λ is the wavelength of the incident light, and m is the order of diffraction. Since the dielectric constant ε of the metal thin film 10, the refractive index n of air, and the pitch 21 of the uneven structure are fixed values ​​once the device structure is determined, this equation defines the relationship between the wavelength λ of the incident light and the angle of incidence θ at which the metal thin film 10 undergoes surface plasmon resonance.

[0020] Returning to Figure 1, the metal thin film 10 is formed on the surface of the diffraction grating structure and is made of a metal capable of exciting surface plasmon resonance when infrared light is incident at a specific incident angle. Typically, gold (Au) is selected as the metal thin film 10, but the embodiment is not limited to this, and silver (Ag), aluminum (Al), copper (Cu), platinum (Pt), palladium (Pd), zinc (Zn), sodium (Na), etc. may also be used. In addition to metals, other materials for the metal thin film 10 include metal nitrides such as titanium nitride (TiN), and metal oxides such as ITO (Indium tin oxide) and FTO (Fluorine-doped tin oxide). The thickness of the metal thin film 10 is, for example, 50 nm.

[0021] The semiconductor layer 11 is formed of a semiconductor such as silicon. When sufficient energy is transferred to the electrons of the metal thin film 10 through surface plasmon resonance to overcome the Schottky barrier between the metal thin film 10 and the semiconductor layer 11, the electrons from the metal thin film 10 move to the semiconductor layer 11, generating an electric current. This current is detected by the current detection unit 12.

[0022] The current detection unit 12 is composed of, for example, a metal layer and a sensor 13 such as a current sensor or potential sensor connected thereto, and detects electrons that have moved from the metal thin film 10 to the semiconductor layer 11 due to energy exceeding the Schottky barrier being provided by surface plasmon resonance.

[0023] The operating principle of the plasmon resonance sensor 1 according to this embodiment will be explained using Figure 3. First, let's consider the case where the plasmon resonance sensor 1 according to the embodiment is incident with monochromatic light of a wavelength that causes surface plasmon resonance at the incident light angle θ. For example, when incident light 40a of monochromatic light of a wavelength that causes the greatest intensity of surface plasmon resonance at the incident light angle θ1 is incident, the sensor output when the incident light angle θ is changed will have the shape shown in curve 40b. Similarly, for example, when incident light 41a and 42a of monochromatic light of wavelengths that cause the greatest intensity of surface plasmon resonance at the incident light angles θ2 and θ3, respectively, the sensor output when the incident light angle θ is changed will have the shapes shown in curves 41b and 42b, respectively.

[0024] In actual measurements, the incident light 43 input to the plasmon resonance sensor 1 is typically white light, not monochromatic light. When there is no object to be measured 45, the sensor output when the incident light angle θ is changed takes the shape shown, for example, by curve 44a. When there is an object to be measured 45, a portion of the spectrum of the incident light 43 is absorbed by the object 45, so the sensor output when the incident light angle θ is changed takes the shape shown, for example, by curve 44b. By taking the difference between curve 44a and curve 44b, spectroscopy can be performed with respect to the object to be measured 45.

[0025] Next, we will explain the device characteristic matrix. When light consisting of a superposition of multiple wavelengths of light is incident, the angle-current value data I is expressed by the following equation (2), using the device characteristic matrix R and the wavelength-intensity data P.

[0026]

number

[0027] Here θ SPRi θ represents the i-th angle of incidence, and λ j 'j' represents the j-th wavelength. Equation (2) shows that applying the device characteristic matrix R to the wavelength-intensity data P yields angle-current value data I. That is, for example, as shown in Figure 4, applying the device characteristic matrix R to data 50, which is the wavelength-intensity data P, yields data 51, which is the angle-current value data I. Conversely, applying the inverse of the device characteristic matrix R to data 51, which is the angle-current value data I, yields data 50, which is the wavelength-intensity data P. In other words, the plasmon resonance sensor 1 according to the embodiment can obtain data 50, which is the wavelength-intensity data P of incident light, by applying the inverse of the device characteristic matrix R to data 51, which is the angle-current value data I obtained from the data obtained from the current detection unit 12, and perform spectroscopy.

[0028] Next, we will explain the background of the embodiment. In the diffraction grating type plasmon resonance sensor 1, the factors that determine the structure of the diffraction grating structure 2 include the pitch 21(d2) of the uneven structure of the diffraction grating structure and the width of the protrusions (d1). As already mentioned, the pitch 21(d2) of the uneven structure of the diffraction grating structure is included on the right side of equation (1), so the pitch 21(d2) of the uneven structure of the diffraction grating structure is directly important in the plasmon resonance sensor 1, but we also focus on the importance of the width of the protrusions (d1). This will be explained using Figures 5 and 6.

[0029] Figure 5 plots the wavelengths at which the intensity of surface plasmon resonance increases when the incident angle is varied, assuming that the pitch 21 of the uneven structure of the diffraction grating structure 2 is d2 = 2600 nm. The wavelengths of infrared light targeted for measurement generally range from approximately 1000 nm to 2500 nm. Here, curves 60, 61, 62, 63, and 64 show the surface plasmon resonance peaks corresponding to diffraction orders 1, 2, 3, 4, and 5, respectively. As can be seen from this figure, when the incident angle is fixed, multiple plasmon resonance signals from different diffraction orders coexist at a single angle. However, the coexistence of multiple plasmon resonance signals from different diffraction orders is expected to increase errors in the device characteristic matrix R, leading to increased numerical errors during the inverse matrix calculation of the device characteristic matrix R and a decrease in spectral accuracy.

[0030] Therefore, we hypothesized that a plasmon resonance sensor 1 with excellent spectral characteristics could be obtained by designing the pitch 21 (d2) and the width of the protrusions (d1) of the uneven structure of the diffraction grating structure in such a way that surface plasmon resonances of multiple diffraction orders are less likely to coexist. In particular, we focused on the ratio of the pitch 21 (d2) and the width of the protrusions (d1) of the uneven structure of the diffraction grating structure and found that when d1:d2 = 1:n, that is, when d2 is n times d1, the surface plasmon resonance of nth-order diffraction is reduced.

[0031] Figure 6 shows the results of such a simulation. Figure 6 shows the simulation results of calculating the magnitude of the surface plasmon resonance signal for each diffraction order m by changing the value of d1 / d2. From these results, it can be seen that when d2 is n times d1, the surface plasmon resonance of the nth order diffraction is reduced. For example, when d1 / d2 = 0.5, that is, when d2 is twice d1, the surface plasmon resonance signal of diffraction order m=2 is reduced. Also, when d1 / d2 = 0.33, that is, when d2 is three times d1, the surface plasmon resonance signal of diffraction order m=3 is reduced. In particular, when m=2, that is, when d2 is around twice d1, it was predicted that the surface plasmon resonance signal of diffraction order m=2 would be reduced and the spectral accuracy would improve, so the optimal ratio of d1 and d2 was investigated.

[0032] In the following simulation, d2 = 2600 nm was selected as the pitch 21 of the uneven structure. When the pitch 21 of the uneven structure is small, surface plasmon resonance signals of other diffraction orders are less likely to appear in the measurement range, but the angle of incidence relative to the target infrared wavelength region becomes larger, and the measurement angle tends to become larger. On the other hand, when the pitch 21 of the uneven structure is large, surface plasmon resonance signals of other diffraction orders are more likely to appear in the measurement range, but the angle of incidence relative to the target infrared wavelength region becomes smaller, and a smaller measurement angle is sufficient. Therefore, considering the balance between the effect of suppressing surface plasmon resonance signals of other diffraction orders and the size of the measurement angle, d2 = 2600 nm was selected as the pitch 21 of the uneven structure.

[0033] The calculation of the device characteristic matrix R will be explained using Figures 7 and 8. As preparation for investigating the influence of the diffraction grating structure 2 on the spectral characteristics of the plasmon resonance sensor 1, the device characteristic matrix R was calculated for a given structure of the diffraction grating structure 2. Figure 7 is a flowchart illustrating the process of calculating the device characteristic matrix R in the plasmon resonance sensor 1 according to this embodiment.

[0034] First, in step S110, a Finite-difference time-domain (FDTD) simulation was performed for monochromatic light of a constant wavelength to numerically simulate the electromagnetic field and generate the angle-intensity data 70 shown in Figure 8. The incident light wavelength was calculated in 100 nm increments from 1000 nm to 2500 nm, and the incident angle was calculated in 0.2 degree increments from 0 degrees to 50 degrees. In addition, the width of the convexity (d1) was calculated for values ​​of d1 = 866 nm, 1200 nm, 1300 nm, 1400 nm, 1500 nm, 1600 nm, 1700 nm, and 1950 nm. In the case where d2 = 2600 nm, d1 = 1300 nm corresponds to the case where the ratio of the width 20 (d1) of the protrusions including the thickness of the metal thin film 10 to the pitch 21 (d2) of the uneven structure is 1:2, and d1 = 1400 nm corresponds to the case where the ratio of the width 22 of the protrusions excluding the thickness of the metal thin film 10 to the pitch 21 (d2) of the uneven structure is 1:2.

[0035] Furthermore, in the simulation, there is a correlation between the signal intensity of surface plasmon resonance in the metal thin film 10 and the intensity of transmitted light passing through the metal thin film 10. For simplicity, we decided to calculate the intensity of transmitted light passing through the metal thin film 10 and treat this as the quantity corresponding to the signal intensity of surface plasmon resonance.

[0036] This yields angle-intensity data 70, as shown in Figure 8, for example. However, while FDTD simulation is a precise calculation method, it is computationally expensive, and the number of data points that can be calculated is limited. Therefore, the device characteristic matrix R was calculated by interpolating data points by performing the processes in steps S120 to S150 for the wavelength and angle of incidence of the incident light.

[0037] Specifically, in step S120, the angle-intensity data 70 generated in step S110 was interpolated with respect to wavelength to generate angle-intensity data 71. Subsequently, in step S130, some of the prominent peaks in the angle-intensity data 71 were fitted with a predetermined function using linear interpolation or a support vector machine to obtain angle-intensity data 71. Subsequently, in step S140, the fitting obtained in step S130 was performed to interpolate data in the angular direction to obtain data 73, which is angle-specific data with data interpolation in the angular direction calculated for light of a certain wavelength. Subsequently, in step S150, data 73 was concatenated in the wavelength direction to obtain data 74, which is data corresponding to the device characteristic matrix R.

[0038] Figures 9 and 10 show the calculation results. Figure 9 is a flowchart explaining the calculation of the device characteristic matrix when d1 (width of the protrusions) is changed to 866 nm, 1200 nm, 1300 nm, and 1400 nm, with d2 (pitch of the uneven structure) = 2600 nm and the thickness of the metal thin film being 50 nm. The top section, similar to Figure 5, plots the wavelength of incident light corresponding to the surface plasmon resonance signal for each diffraction order when the incident angle is changed, and superimposes the transmittance data obtained from FDTD simulation. It can be seen that the transmittance data obtained from FDTD simulation is well superimposed on the angle-wavelength plot data, and it can be seen that it is generally reasonable to think that the surface plasmon resonance signal data is calculated by calculating the transmittance data. The middle section is the angle-signal intensity data obtained from FDTD simulation. Simulations were performed for the wavelength of incident light from 1000 nm to 2500 nm in 100 nm increments. The lower section shows the data obtained in step S150 after data concatenation processing, and corresponds to the device characteristic matrix R. There is no discrepancy between the data in the middle section and the data in the lower section, suggesting that a generally reasonable device characteristic matrix R has been calculated.

[0039] Figure 10 shows the same calculations as in Figure 9, but with d1 (width of the convex portion) set to 15000 nm, 1600 nm, 1700 nm, and 1950 nm, respectively.

[0040] Next, we will explain how to calculate the index value of the spectral performance using Figures 11 and 12. Figure 11 is a flowchart illustrating the process for calculating the index value representing the spectral performance of a given diffraction grating structure.

[0041] Here, we will briefly explain the method for calculating the spectral performance index value. Given a diffraction grating structure 2, the device characteristic matrix R is calculated by the process already shown in Figure 7. In this state, multiple trial data of wavelength-intensity data are randomly generated. Angle-current value data is created by multiplying this wavelength-intensity data by the device characteristic matrix R, and wavelength-intensity data is created again by multiplying the angle-current value data by the inverse matrix of the device characteristic matrix R. At this time, mathematically, the data should be the same as the obtained wavelength-intensity data, but in reality, due to numerical instability of the device characteristic matrix R such as rounding errors, it will deviate slightly from the original wavelength-intensity data. By calculating an index that shows how much the recalculated wavelength-intensity data deviates from the original wavelength / intensity data, it is possible to evaluate the stability of the spectral performance of the device characteristic matrix R, and by extension, the plasmon resonance sensor 1 according to the embodiment.

[0042] First, in step S210, trial data of the incident light spectrum (wavelength-intensity data) is created by changing the incident light spectrum (incident light intensity at each wavelength). For example, the wavelength-intensity data 80 shown in Figure 12 is created as trial data. Here, the trial data may be randomly generated wavelength-intensity data, or it may be trial data that simulates the spectrum of the incident light actually incident and is generated to be close to the spectrum of the incident light actually incident. Multiple sets of trial data are created, and the spectral performance is evaluated by, for example, averaging the indices created for each set of trial data.

[0043] Next, in step S220, the device characteristic matrix R is applied to the created wavelength-intensity trial data to create angle-current value data. For example, angle-current value data 82 is obtained by multiplying the trial data of wavelength-intensity data 80 created in step S210 by the device characteristic matrix R. Next, in step S230, the wavelength-intensity data is recreated by applying the inverse of the device characteristic matrix R to the angle-current value data. For example, the recreated wavelength-intensity data is obtained by applying the inverse of the device characteristic matrix R to the angle-current value data 82 created in step S230.

[0044] Next, in step S240, the index value is calculated based on the incident light spectrum created in step S210 and the incident light spectrum created in step S230. For example, in Figure 12, the index value is calculated based on the wavelength-intensity data 85 shown in step S210 (shown by the solid line) and the intensity data 86 shown in step S240. The calculated index value is R 2 Examples include R-squared (coefficient of determination) and RSS (Residual Sum of Squares).

[0045] In Figure 13, when d2 (pitch of uneven structure) = 2600 nm, and d1 (width of protrusions) is 866 nm, 1200 nm, 1300 nm, and 1400 nm, the index value is R. 2 The data for calculating the coefficient of determination is shown. Figure 14 also shows the data obtained when the same process was performed with d1 (width of the convex portion) set to 1500 nm, 1600 nm, 1700 nm, and 1950 nm. 2 The closer the coefficient of determination is to 1, the better the spectral characteristics of the device are considered to be.

[0046] In addition, in FIG. 15, when d2 (pitch of the concavo-convex structure) = 2600 nm and d1 (width of the convex portion) is 866 nm, 1200 nm, 1300 nm, and 1400 nm, data when calculating RSS (root mean square) as an index value is shown. Also, in FIG. 16, data when performing the same processing when d1 (width of the convex portion) is 1500 nm, 1600 nm, 1700 nm, and 1950 nm is shown. The closer RSS (root mean square) is to 0, the better the spectral characteristics of the device are evaluated.

[0047] Summarizing the above results, it becomes as shown in Table 1 and Table 2 below.

[0048]

Table 1

[0049]

Table 2

[0050] Regarding the result of R 2 (coefficient of determination) shown in Table 1, when d2 (pitch of the concavo-convex structure) = 2600 nm , when d1 (width of the convex portion) = 1400 nm, R 2 becomes the maximum value of 0.99955, and then when d1 (width of the convex portion) = 1600 nm, R 2 becomes 0.99944. Also, in all ranges of d1, R 2 is 0.99 or more, and in the range where d1 is 1300 nm to 1600 nm, R 2 is 0.999 or more.

[0051] Regarding the result of RSS (root mean square) shown in Table 2, when d2 (pitch of the concavo-convex structure) = 2600 nm, when d1 (width of the convex portion) = 1400 nm, RSS becomes the minimum value of 0.124, and then when d1 (width of the convex portion) = 1600 nm, R 2 becomes 0.18783. Also, in the range where d1 is 1200 nm to 1700 nm, RSS is less than 0.55.

[0052] From the above, when d2 (pitch of the uneven structure) = 2600 nm, the width of the protrusions is preferably in the range of 1200 to 1700 nm, and even more preferably in the range of 1300 to 1600 nm. Converting this to the ratio of the pitch of the uneven structure to the width of the protrusions in the diffraction grating structure, the pitch of the uneven structure in the diffraction grating structure is preferably in the range of 1.52 to 2.17 times the width of the protrusions, and even more preferably in the range of 1.62 to 2.0 times.

[0053] Furthermore, since generally good results were obtained when d2 (pitch of the uneven structure) was around twice d1 (width of the protrusions), it can be seen that it is preferable for the pitch of the uneven structure to be approximately twice the width of the protrusions when measured including the thickness of the metal thin film. In particular, the best results were obtained when d1 (width of the protrusions) = 1400 nm. Here, when d1 = 1400 nm, if we consider only the silicon portion as the width of the protrusions, without including the width of the metal thin film, the width of the protrusions is approximately 1300 nm, and in this case, the pitch of the uneven structure is approximately twice the width of the protrusions 22 (when measured without including the thickness of the metal thin film). Therefore, it can be seen that it is preferable for the pitch 21 of the uneven structure to be approximately twice the width of the protrusions 22 when measured without including the thickness of the metal thin film.

[0054] The reason why it is preferable for d2 (pitch of the uneven structure) to be an integer multiple of d1 (width of the protrusions) is, as mentioned earlier, that when d2 (pitch of the uneven structure) is n times d1 (width of the protrusions), the intensity of the surface plasmon resonance signal with diffraction order n is reduced. In particular, when n=2, when d2 (pitch of the uneven structure) is twice d1 (width of the protrusions), the intensity of the surface plasmon resonance signal with diffraction order 2 is reduced. As a result, the numerical stability of the device characteristic matrix R is improved, and the spectral characteristics of the plasmon resonance sensor 1 are improved.

[0055] The preferred range for d2 (pitch of the uneven structure) will be explained using Figures 17 and 18. Figure 17 is a diagram showing the spectral characteristics of the plasmon resonance sensor 1 according to the embodiment, calculated by changing d2 (pitch of the uneven structure) to 1000 nm, 1100 nm, 1200 nm, 1250 nm, 1300 nm, and 1400 nm in order to investigate the lower limit of the preferred range for d2 (pitch of the uneven structure). Assuming that infrared light in the range up to approximately 2500 nm should be detected, the wavelength of light that causes surface plasmon resonance at an incident angle of 90 degrees will be 2500 nm or more when d2 (pitch of the uneven structure) is 1250 nm or more. Therefore, it is desirable that d2 (pitch of the uneven structure) be 1250 nm or more.

[0056] However, considering the range of motion of the sensor as a MEMS (Micro Electro Mechanical Systems) mechanism, the practically measurable incident angle range is thought to be approximately -50 to 50 degrees. Therefore, considering the d2 (pitch of the uneven structure) at which the wavelength of light that causes surface plasmon resonance reaches 2500 nm at an incident angle of 50 degrees, it is even more desirable for d2 (pitch of the uneven structure) to be 1400 nm or greater.

[0057] Figure 18 shows the spectral characteristics of the plasmon resonance sensor according to the embodiment, calculated by varying d2 (pitch of the uneven structure) to 2600 nm, 2800 nm, and 3900 nm in order to investigate the upper limit of the preferred range of d2 (pitch of the uneven structure). If signals of various diffraction orders are mixed within the measurement range, the accuracy of the device characteristic matrix decreases. Therefore, it is undesirable for signals of various diffraction orders to be mixed within the measurement range. As shown in Figure 18, when d2 is around 3900 nm, the peak of the 7th order diffraction is at a detection wavelength of around 1000 nm when the incident angle is 50 degrees, and when d2 is around 2800 nm, the peak of the 5th order diffraction is at a detection wavelength of around 1000 nm. Therefore, by setting d2 (pitch of the uneven structure) to 3900 nm or less, the detection wavelength of the 7th order diffraction peak can be set to a wavelength outside the measurement range when the incident angle is 50 degrees, and by setting d2 (pitch of the uneven structure) to 2800 nm or less, the detection wavelength of the 5th order diffraction peak can be set to a wavelength outside the measurement range when the incident angle is 50 degrees.

[0058] Based on the above, the range of d2 (pitch of the uneven structure) is preferably 1200 nm to 1700 nm, and even more preferably 1300 nm to 1600 nm.

[0059] In the examples described so far, we have mainly described cases where the pitch of the uneven structure in the diffraction grating structure is 1.52 times or more and 2.17 times or less the width of the convex portion, that is, cases where the pitch of the uneven structure in the diffraction grating structure is about twice the width of the convex portion. However, the embodiments are not limited to this, and the pitch of the uneven structure in the diffraction grating structure may be an integer multiple (3 or more) of the width of the convex portion. When d2 (pitch of the uneven structure) is n times d1 (width of the convex portion), the intensity of the surface plasmon resonance signal with diffraction order n is reduced, which improves the numerical stability of the device characteristic matrix R and has the effect of improving the spectral characteristics of the plasmon resonance sensor 1.

[0060] As described above, the infrared spectrometer according to this embodiment can reduce plasmon resonance signals caused by diffracted light of a specific order, thereby achieving high spectral performance. [Explanation of symbols]

[0061] 2...Diffraction grating structure, 10...Metal thin film, 12...Current detection unit, 21...Pitch of the uneven structure, 20...Width of the protrusions, 22...Width of the protrusions when measured without including the thickness of the metal thin film

Claims

1. A diffraction grating structure including continuous convex and concave portions, A thin metal film is formed on the surface of the diffraction grating structure such that surface plasmon resonance can be excited when infrared light is incident at a specific angle of incidence. A current detection unit for detecting the current generated by the surface plasmon resonance, Equipped with, An infrared spectrometer in which the pitch of the uneven structure in the diffraction grating structure is 1.52 times or more and 2.17 times or less the width of the protrusions.

2. The infrared spectrometer according to claim 1, wherein the pitch of the uneven structure is 1.62 times or more and 2.0 times or less the width of the protrusions.

3. The infrared spectrometer according to claim 1, wherein the pitch of the uneven structure is twice the width of the protrusions when measured including the thickness of the metal thin film.

4. The infrared spectrometer according to claim 1, wherein the pitch of the uneven structure is twice the width of the protrusions when measured without including the thickness of the metal thin film.

5. The infrared spectrometer according to claim 1, wherein the pitch of the uneven structure is 1250 nm or more and 3900 nm or less.

6. The infrared spectrometer according to claim 1, wherein the pitch of the uneven structure is 1400 nm or more and 2800 nm or less.

7. The infrared spectrometer according to claim 1, wherein the width of the protrusion is 1200 nm or more and 1700 nm or less.

8. The infrared spectrometer according to claim 1, wherein the width of the protrusion is 1300 nm or more and 1600 nm or less.

Images