WO2018038413A1 - Spectrometer and spectrum measurement method utilizing same - Google Patents

Abstract

Disclosed is a spectrometer provided with: a first unit spectral filter for absorbing or reflecting the incoming light of a partial wavelength band of the light spectrum of a target object; a second unit spectral filter for absorbing or reflecting the light of a wavelength band different from the partial wavelength band; a first light detector for detecting a first light spectrum passing through the first unit spectral filter; a second light detector for detecting a second light spectrum passing through the second unit spectral filter; and a processing unit for executing the function of reconstructing the entered light spectrum of the target object from the spectrum of light detected from the first light detector and second light detector.

Description

Spectroscope and spectral measurement method using the same

The present invention relates to a spectroscope, and more particularly, to a spectrometer using a spectroscopic filter having a stopband characteristic and a spectrum measurement method using the same.

Spectrometers using optical filters are used in various wavelength ranges, such as visible light and infrared light. For example, the wavelength band used for infrared spectroscopy can be divided into a near infrared region and a mid infrared region. The prior art will be described below by taking an example of a mid-infrared spectrometer.

The mid-infrared wavelength region (2-20 mm) is a region in which the fundamental vibrational mode of almost all chemical molecules exists and is also called molecular fingerprint band because it shows different infrared absorption spectra according to molecular bonding state. Lost is an important wavelength band. Regardless of solid, liquid, or gas, unknown samples can be discriminated or qualitative and quantitative analysis can be performed with high selectivity for specific target molecules.

The near-infrared region (0.78-2 mm) is also a section in which a mode by overtone and combination of the mid-infrared band basic vibration modes exists and its intensity is low, but the substance can be identified or quantified. In the visible region (0.38-0.78 mm), it can be used not only for the material's unique absorption spectrum but also for the color of the object, the analysis of the phosphor, and the detection of the biomolecule.

Conventionally, benchtop spectroscopy, such as Fourier transform infrared (FTIR) spectroscopy, has been used as an organic-inorganic material analyzer, but gradually control of environmental factors, water quality inspection, industrial and agricultural process line monitoring, process control, residual pesticide detection and country of origin. The demand for development as a miniaturized device for field measurement is increasing with the applicability in various applications such as food, oil oxidation measurement, and medical biotechnology.

As the most effective method of miniaturizing the spectrometer, a method of integrating an optical component, which is responsible for light dispersion, with an optical detector array by manufacturing a band pass filter in an array form instead of a conventional prism and a diffraction grating has been proposed. Unlike the Fourier transform infrared method, there is no need for a moving object, which makes it robust and compact.

A typical band pass filter is a Fabry-Perot filter that uses the optical interference effect of a dielectric resonator placed between two reflective films. However, since the number of filters required for constructing the spectroscope is large, and the lithography step is increased as much as the number of dielectric resonant layers required for the spectroscope, it is suitable for application to a planar spectrometer. In order to solve this problem, a linear variable filter (LVF) has been developed and used.

The linear variable filter (LVF) is an optical filter of a Fabry-Perot resonator structure, and has a structure in which the thickness of the dielectric resonant layer varies linearly in the longitudinal direction. In the linear variable filter, a lower mirror layer and an upper mirror layer are positioned with a dielectric resonance layer interposed therebetween.

Such a linear variable filter has a limitation in process reproducibility due to the linear structure whose thickness varies in the longitudinal direction. In addition, since the resolution of a spectrometer using a linear variable filter is determined by the height-to-length ratio of the linear variable filter, it is difficult to miniaturize the device of the spectrometer. In particular, the linear structure is disadvantageous in terms of productivity due to the lack of process compatibility with the two-dimensional imaging sensor technology.

The distance between the filter and the array of photodetectors existed because the transmission spectra of each linear variable filter consisted of continuous spectral superposition and the integration between the linear variable filter and the photodetector was not monolithic. Due to the stray light effect, there was a disadvantage that the filter performance is reduced.

In addition, due to the characteristics of the interference filter, a multi-transmission mode occurs, which limits the free spectral range and requires a separate device in order to analyze the broadband.

In addition, in the spectroscope utilized in the mid-infrared wavelength band, there is a limitation in the construction of the interference type optical filter because there is a limitation in the material system having high transmittance.

In addition, there was a problem due to interference effects with neighboring cells, and the available materials were also limited.

In order to solve the above problems, an object of the present invention is to provide a spectrometer that is capable of wideband control and advantageous for two-dimensional integration only by controlling the horizontal structure.

On the other hand, it is another object of the present invention to provide a spectroscopic filter array having excellent heat resistance and durability, and a spectrometer using the same.

One aspect of the present invention is a first unit spectroscopic filter that absorbs or reflects light of a portion of the wavelength band of the light spectrum of the incident object, and a light absorbing or reflecting light of a wavelength band different from the wavelength range A two-unit spectroscopic filter, a first photodetector for detecting a first light spectrum passing through the first unit spectroscopic filter, and a second photodetector for detecting a second light spectrum passing through the second unit spectroscopic filter And a processing unit for performing a function of restoring the light spectrum of the object incident from the spectrum of the light detected from the first and second photodetectors.

The unit spectroscopic filters have a stopband characteristic. The stopband characteristic is that the unit spectral filters have a peak of reverse transmittance according to the wavelength so that light of a specific wavelength band cannot be transmitted. In addition, this is another characteristic of the filter having the peak of reverse transmittance according to the wavelength by preventing the light of the specific wavelength band from transmitting by absorbing or reflecting the light of the specific wavelength band corresponding to the central wavelength for each unit filter. Means. On the other hand, when the object spectrum is a peak function, the intensity distribution for each filter sequence measured in the photodetector is observed in the dip function, if the object spectrum is a dip function, the photodetector measurement profile is in the form of a peak function. That is, it may be determined by expressing that the object spectrum and the reverse phase form, or inferred from the intensity distribution in the neighboring wavelength band during the spectral restoration process.

Preferably, in the first and second unit spectroscopic filters, metal patterns having a predetermined shape may be periodically arranged, and the metal patterns of the first and second unit spectroscopic filters may be arranged. Have different periods.

Preferably, the first and second photodetectors are comprised of some photodetecting pixels of the CMOS image sensor.

First, the plasmonic metal may be composed of a material selected from Au, Ag, Al, Cu, or an alloy including at least one of them. For example, AgPd, CuNi alloy, etc. can also be used.

In addition, the metal patterns may be formed of Cr, Ni, Ti, Pt, Sn, Sb, Mo, W, V, Ta, Te, Ge, Si or at least one of the metals having both high absorption and refractive index in the visible and near infrared bands. It may be composed of a material selected from the alloy containing. In addition, at least one metal pattern may be selected from the group consisting of Ta, W, Mo, Ni, Cr, TiN, and TiON in which the optical behavior in the mid-infrared band follows the druid free electron model.

In addition, the metal patterns are formed of at least a double layer, and a low loss high reflectivity metal material and a light absorbing metal material are laminated, and the low loss high reflectivity metal material includes Ag, Au, Al, Mg and at least one of them. The light-absorbing metal material may be selected from among alloys. The light-absorbing metal material may be selected from among Cr, Ni, Ti, Pt, Sn, Sb, Mo, W, V, Ta, Te, Ge, and Si. And silicides, carbides, nitrides, or sulfides comprising alloys thereof and these metals.

Meanwhile, the metal patterns of the first unit spectroscopic filter and the metal patterns of the second unit spectroscopic filter preferably have the same duty cycle.

Preferably, a period of the metal patterns of the first and second unit spectroscopic filters is between 100 nm and 800 nm. This is a preferred range for Si-based photosensitive device operating band or visible-near-infrared region (380-1100 nm) spectrometer construction. The preferred period of the metal patterns in the InGaAs or Ge based near infrared photosensitive device operating band (800-1700 nm) is between 0.6 um and 1.5 um. In the mid-infrared band (2-15 um) preferably the period of the metal nanopatterns is between 0.8 um and 8 um.

The first unit spectrometer and the second unit spectroscopic filter may further include a passivation layer. The passivation layer is made of a material selected from HfO ₂ , ZrO ₂ , ZnO, ZnSe, TiO ₂ , Al ₂ O ₃ , SiO _x , SOG, or an alloy consisting of at least two of them.

The first unit spectroscopic filter and the second unit spectroscopic filter may further include a protective layer, and the protective layer may include a low refractive index silicon oxide, a silicon nitride film, magnesium fluoride, calcium fluoride, a low molecular resin, or a polymer material. desirable.

Preferably, the process unit, the step of calculating the intensity of light absorbed or light reflected by the first unit spectroscopic filter from the spectrum of the light of the first photodetector, the second unit spectroscopic filter from the spectrum of the light of the second photodetector Calculating an intensity of light absorbed or reflected by the light; and restoring a light spectrum of the incident object from the intensity of light absorbed or reflected by the first and second unit spectroscopic filters.

According to another aspect of the present invention, the light spectrum of the object is incident on the first and second unit spectroscopy filter and the first unit spectroscopy light absorbs or reflects light of some wavelength band, the second unit spectroscopy The filter absorbs or reflects light in a wavelength band different from the wavelength band, and a first photodetector detects a first light spectrum passing through the first unit spectroscopic filter, and a second photodetector detects the first light spectrum. A spectroscope comprising detecting a second light spectrum passing through the two-unit spectroscopic filter and restoring the light spectrum of the object incident from the spectra of the light detected by the first and second photodetectors. It provides a spectrum measurement method used.

Preferably, the restoring the light spectrum of the object may include calculating an intensity of light absorbed or reflected by the first unit spectroscopic filter from the first light spectrum of the first light detector and the second light detector. Calculating the intensity of the light absorbed or reflected by the second unit spectroscopic filter from the second light spectrum of < RTI ID = 0.0 > and / or < / RTI > Restore the spectrum of light incident from it.

In the step of restoring the spectrum of the object, a direct readout or regularization technique is preferably used.

The reconstructing of the spectrum of the object may be performed by substituting the transmission spectrum f _i (λ) of the individual filter and the spectral sensitivity function d _i (λ) of the photodetector into the following equation and measuring the detected detection signal r _i. Can be derived. Here, D _i (λ) is f _i (λ) d _i (λ).

Here, the spectrum of the object to be analyzed is s (λ), the transmission functions of the individual filters F are f _i (λ), the noise is n _i , and the sensitivity function of the photodetector PD is d _i (λ). In this case, the detection signal r _i generated when the spectrum of the object passes through the filter and reaches the photodetector.

According to the invention as described above, there is an effect that it is possible to provide a spectrometer that is easy to integrate two-dimensional in a low cost simple process.

In addition, it is possible to implement a wide-range operating range that can cover the wavelength range from the visible light to near infrared, infrared.

In addition, by adopting a spectrometer using a single stop band, the design and manufacturing process of the spectral filter is simplified, the metal material selectivity is extended, and the filter function can be easily measured and analyzed, thereby improving the signal restoration ability.

In addition, there is an effect that can improve the reliability of the device by applying a material system that is excellent in the long-term thermal stability and durability in the mid-infrared region, while being able to express plasmonic properties in the mid-infrared band.

1 is a spectroscopic filter array according to an embodiment of the present invention, Figure 2 is a view showing a part of the spectrometer using the same, Figure 3 is a block diagram of a spectrometer according to an embodiment of the present invention.

4 is a plan view illustrating an array of spectral filters according to an embodiment of the present invention, and FIG. 5 is a diagram illustrating the metal patterns of FIG. 4.

6 to 8 are diagrams illustrating spectral filters according to another embodiment.

9, 10 and 11 are graphs showing simulation results of transmission characteristics of stop bands for several metal materials and patterns according to an exemplary embodiment of the present invention.

12 is a scanning electron micrograph showing the thermal stability according to the selection of the metal material system constituting the nanostructure array.

13 is a simulation result showing that the free spectrum range of the metal nanostructure array type stopband filter according to the present invention can be extended not only to the near infrared band but also to the visible wavelength range.

14 shows an example of constructing a spectrometer through a one-dimensional linear array coupling between the filter array and the photodetector of the present invention.

15 shows an example of constructing a spectrometer through a two-dimensionally arranged coupling between the filter array and the photodetector of the present invention.

16 is a flowchart for explaining the spectroscopic method of the present invention.

FIG. 17 illustrates a digital signal processing process for reconstructing an object spectrum for spectroscopic operation using an infrared optical filter according to an embodiment of the present invention.

18 and 19 are graphs for comparing the transmission band type filter array and the stop band type filter array in the visible-near-infrared wavelength band.

FIG. 20 is a schematic diagram for explaining gain in terms of detection limits of spectral signals compared to transmission band filter arrays when using a stopband filter array according to the present invention.

21 and 22 show graphs of the function reconstructed by varying the duty cycle in the metal nanodisk array filter.

FIG. 23 is a graph comparing spectral changes according to nanodisk shapes in a filter of a nanodisk array structure having a circle and a hexagonal lattice structure, and FIG. 24 shows a spectroscopic filter array by mixing into disk arrays having two or more shapes. The constructed e is shown.

FIG. 25 is a graph showing the distribution of the absorption coefficient versus the refractive index of optical constant dispersion characteristics of metals, dielectrics, and semiconductor materials.

FIG. 26 is a light transmittance spectrum of a hexagonal lattice-structured nanodisk array using Cr and Ti and calculated in the visible-near-infrared band.

27 is a graph of light transmittance and light reflectivity of a nanodisk array calculated using tungsten (W).

FIG. 28 is a graph illustrating spectral recovery performance by fabricating a tungsten nanodisk array of FIG. 26 using a stopband filter array and applying a digital signal processing algorithm.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, embodiments of the present invention illustrated below may be modified in many different forms, and the scope of the present invention is not limited to the embodiments described below. The embodiments of the present invention are provided to more completely explain the present invention to those skilled in the art.

1 is a view showing a part of the spectroscopic filter array according to an embodiment of the invention, Figure 2 is a view showing a portion of the spectrometer using the spectroscopic filter array shown in FIG.

The spectroscopic filter array 10 according to the embodiment of the present invention includes a plurality of unit spectroscopic filters F ₁ and F ₂ . The plurality of unit spectroscopic filters F ₁ and F ₂ are configured to filter light of different wavelengths. A plurality of unit spectral filters means at least two unit spectral filters. The spectrometer 20 according to the embodiment of the present invention includes the spectroscopic filter array 10 and each of the photodetection areas PD ₁ and PD corresponding to each of the plurality of unit spectroscopic filters F ₁ and F ₂ . _And a photodetector array 210 comprising ₂ ). The plurality of unit filters F ₁ and F ₂ are configured to filter light of different wavelengths, and each of the unit filters F ₁ and F ₂ corresponds to each of the photodetection areas PD ₁ and PD ₂ . do. The plurality of unit filters mean at least two unit filters. 2 illustrates an example in which the unit filters F ₁ and F ₂ correspond to the photodetection areas PD ₁ and PD ₂ with the substrate 110 interposed therebetween. F ₁ and F ₂ and corresponding light detection regions PD ₁ and PD ₂ mean that the unit filters F ₁ and F ₂ and the light detection regions PD ₁ and PD ₂ are directly connected to each other. It may include a situation in which it is in contact, in addition to that it is to be understood to include a separate modular module form, or a form in which an optical system such as a relay lens is separately inserted between the two.

The unit spectroscopic filters F ₁ and F ₂ have a main characteristic configuration using a stop band. For example, the unit spectroscopic filters F ₁ and F ₂ are formed of a plasmonic filter that forms a metal pattern 120 periodically to enable filtering.

The unit spectroscopic filters F ₁ and F ₂ of FIG. 1 illustrate the implementation of the stop band through a structure in which the protruding metal patterns 120 having a predetermined shape (embossing) are periodically arranged.

The metal patterns can form an array of metal nanostructures with a periodic lattice structure, and the extraordinary light absorption to light reflection enhanced in a specific wavelength band by coupling of the localized surface plasmon and the lattice mode. Indicates a phenomenon. As a result, the spectrum of light passing through the array of metal nanostructures forms a dip curve in which the transmittance decreases rapidly in the selective wavelength band where specific light absorption to light reflection is enhanced. It acts as a stopband when based on transmitted light, and its spectral shape depends on the choice of metal and the geometry, such as the period and particle size of the nanostructure array, especially its central wavelength by the lattice period. Has a predominantly determined characteristic.

Conventionally, a metal nanohole array structure showing a transmission band has been utilized as a plasmonic filter. The metal nanohole array structure exhibits an extraordinary optical transmission (EOT) phenomenon in which light transmittance is increased at a specific wavelength by coupling between a surface plasmon wave and a lattice mode that travels along the metal thin film surface. In addition, since the metal nanohole array structure is based on coupling between traveling waves, unlike the metal nanodisk array structure, various modes exist and are not defined as a single transmission band. The presence of this multimode spectrometer in the filter array method In operation, distortion may occur in the process of processing signal wavelengths incident on the respective light detection regions. In addition, since the generation of attenuated surface plasmon waves traveling along the metal thin film surface is required, it is difficult to use light-absorbing metals and there is a limitation that the material system is limited to low loss high reflectivity noble metals such as Ag and Au. . Moreover, these precious metal materials also have a disadvantage in that structural design for forming a transmission band filter based on a metal nanohole array has a lot of limitations because the imaginary constant of the dielectric constant increases in the mid-infrared band.

On the other hand, since the array of metal patterns made of highly reflective metal material shows the reflectance peak curve in a specific wavelength band by coupling with the lattice mode, it is used for the limited use such as reflective color filter or decorative coating using reflected light. come. The present invention provides a spectroscopic technique in which the arrangement of the metal nanostructures with the photodetector array in the form of a stopband filter is not a reflective structure, but a transmissive structure.

The present inventors have a relatively wide spectral free spectral range when using the stop band formed by the arrangement of the plasmonic nanostructures in the visible wavelength region and the infrared wavelength band, so that the visible and infrared wavelengths are larger than those of the transmission band filter. It was confirmed that there is an advantage of covering the entire band.

In addition, the inventors of the present invention have confirmed that the phenomenon of deterioration occurs especially when the spectrometer is composed of a transmission band in the infrared wavelength band, but in the case of the stopband filter, this problem can be solved. In this case, the spectroscope of the mid-infrared band can be defined as operating in the wavelength band of 2 to 15 m, more preferably the spectrometer covers the mid-infrared of 2.5 m to 12 m.

In order to configure the spectrometer, a processing unit 330 (see FIG. 3) is separately provided, and the processing unit is incident by using the optical signal detected from the photodetector array including the photodetection areas PD ₁ and PD ₂ . It performs the function of reconstructing the spectrum of light. It will be described later in detail.

The photodetector array in the mid-infrared wavelength band may be a one-dimensional array type infrared photodetector using a pyroelectric, thermopile, volometer, photoconductive and photovoltaic type photodetector elements or an infrared image sensor in the form of a two-dimensional array. In the visible and near-infrared wavelength bands, one-dimensional photodetector arrays or two-dimensional CMOS image sensors using Si, Ge, InGaAs-based photodetectors may be used.

Meanwhile, the periodic metal patterns 120 are formed on a separate substrate 110 and then optically coupled with the photodetector array, or directly monolithic with a buffer layer (not shown) interposed on the photodetector array. Can be formed. The buffer layer (not shown) is preferably an optically transparent dielectric layer serving as a protective layer for each pixel of the photodetector, and may be a SiNx or SiO ₂ layer. In addition, it is possible to further include an interfacial adhesion layer such as Ti, Cr, and transition metal oxide in order to enhance adhesion with the upper metal pattern layer.

The substrate 110 may be various kinds without being particularly limited, and may be a flexible light-transmissive substrate such as glass or a polymer, Ge, GeSe, ZnS, ZnSe, sapphire, CaF ₂ , MgF _2, or the like. The flexible light transmissive substrate is preferably composed of a transparent or translucent polymer having appropriate adhesion and shock absorption. Specific examples of the polymer include, but are not limited to, polystyrene (PS), expandable polystyrene (EPS), polyvinyl chloride (PVC), styrene acrylonitrile copolymer (SAN), polyurethane (PU: Polyurethane), Polyamide (PA: Polyamide), Polycarbonate (PC: Polycarbonate) Modified Polycarbonate, Poly (vinyl butyral), Polyvinyl acetate, Acrylic It can be resin (Acrylic Resin), epoxy resin (EP: Epoxy Resin), silicone resin (Silicone Resin), unsaturated polyester (UP: Unsaturated Polyester), polyimide, polyethylene naphtalate, polyethylene terephtalate, etc. Can be used. On the other hand, the silicon wafer is preferably in the mid-infrared wavelength band, but is not limited thereto. When manufacturing a spectrometer that operates effectively in the infrared region, it may be inappropriate for the substrate itself to generate a lot of absorption in the infrared band.

The metal material constituting the metal patterns 120 may be appropriately selected according to the wavelength band. This will be described in detail.

First, a low loss high reflectance metal material widely used as a plasmonic metal may be suitably used in the visible-near-infrared wavelength band, the mid-infrared wavelength band, and the like. The metal patterns may include Au, Ag, Al, Cu or at least one of the plasmonic metals. It may be an alloy comprising two alloys or at least one of them and other elements.

Next, in the visible region and the near-infrared band, Cr, Ni, Ti, Pt, Sn, Sb, Mo, W, V, Ta, Te, Ge, Si or at least two alloys thereof, which are metals having a high light absorption and refractive index, It may be an alloy containing at least one of these and other elements.

In the mid-infrared band, the optical behavior may be at least one selected from the group consisting of Ta, W, Mo, Ni, Cr, TiN, and TiON following the druid free electron model.

3 is a block diagram of a spectrometer according to another embodiment of the present invention. Referring to FIG. 3, the spectrometer 30 includes a spectroscopic filter array 310, a photodetector array 320, and a processing unit 330. The spectral filter array 310 includes a plurality of unit spectral filters F ₁ and F ₂ for filtering light in different wavelength regions, and the photodetector array 320 corresponds to each of the plurality of spectral filters. The photodetection regions PD ₁ and PD ₂ are provided, and the processing unit 330 performs a function of reconstructing the spectrum of incident light using the optical signal detected from the photodetector array 320. The plurality of unit spectroscopic filters F ₁ and F ₂ are filters having stop band characteristics as described above. The processing unit 330 according to the present invention serves as a spectrometer for restoring the object spectrum by applying a subsequent digital signal processing algorithm, and it becomes possible to implement a filter array-based spectrometer. It will be described later in detail.

4 is a plan view illustrating a spectral filter array according to an embodiment of the present invention. It is understood that a cross section taken along the line II ′ in FIG. 4 is shown as in FIG. 1. 5 exemplarily shows metal patterns of the spectral filter array. As can be seen in FIG. 5, both the linear lattice structure and the two-dimensional lattice structure are applicable. The two-dimensional lattice structure may be a square lattice or a hexagonal lattice. The shape of the metal nanostructure may be a variety of shapes, such as rectangular disk, circular disk, polygonal structure, nano-bar unit structure, cross bar.

On the other hand, each of the unit spectroscopic filters (F ₁ , F ₂ ) are implemented to have the same duty cycle or filling rate of the nanostructure. That is, when D ₁ / P ₁ is a duty cycle in the unit spectroscopic filter F ₁ , this value is preferably 30% to 80%. If the duty cycle is less than 30%, the transmittance dip is very small and if it is more than 80%, too broad a dip curve tends to be generated.

The duty cycle of D ₂ / P ₂ in the spectral filter unit (F ₂₎ is kept equal to the duty cycle of the spectral filter unit (F _1). However, the period of the unit spectroscopic filter F ₁ and the unit spectroscopic filter F ₂ is changed.

The element that determines the resonant wavelength of the unit spectroscopic filters F ₁ and F ₂ is determined by the period, the shape of the metal structure, the thickness of the metal structure, the duty cycle, and the like. The main factor determining the resonant wavelength here is the period. When manufacturing a spectrometer using unit spectroscopic filters F ₁ and F ₂ , one of the advantages is that the resonance wavelength can be easily obtained by changing a relatively simple element.

In the case of the visible light wavelength range to the near infrared wavelength band, the period of the metal patterns is preferably determined between 0.1 μm and 1.5 μm, and when targeting the mid infrared range, it is determined between 0.8 μm and 8 μm. desirable.

For example, when using Si as a substrate for the mid-infrared band and having a large refractive index, the period of the metal patterns is preferably determined between 0.8 µm and 4 µm. The thickness of the metal patterns is preferably 5 nm to 500 nm, more preferably 10 nm to 300 nm. If it is smaller than 5 nm, the ratio of free electrons scattered on the surface is increased to act as a large factor of plasmon attenuation, and if it is more than 500 nm, multipole resonance may occur due to a volume increase effect.

6 is a diagram illustrating an example of a spectral filter according to another embodiment. Referring to FIG. 6, a low reflection coating layer 180 is additionally formed on a lower portion of the substrate 110 opposite to the upper portion of the substrate 110 on which the unit spectroscopic filters F ₁ and F ₂ are formed.

The low reflection coating layer 180 may be coated with a thin film layer having a refractive index that satisfies graded index conditions between the substrate 110 and a neighboring medium or may be formed of a nanocon structure having a motheye shape. In this case, the coupling with the lower photodetection area may be configured such that the periodic metal patterns face the lower photodetector areas so that light may enter the surface of the low reflection coating layer. For example, since the refractive index of the substrate materials showing high transmittance in the mid-infrared wavelength band is high except for some materials such as CaF ₂ , the opposite side of the upper part where the periodic metal patterns 120 are formed to reduce the reflection loss at the interface. A structure in which the low reflection coating layer 180 is additionally formed below the phosphorus substrate 110 may be effective.

7 is a diagram illustrating an example of a spectral filter according to another embodiment. Referring to FIG. 7, the metal patterns of FIGS. 1 and 2 are formed as the double layer 130. The double layer 130 may be configured as a double layer of a low loss high reflectance metal material 134 and a light absorbing metal material 132. The low loss high reflectivity metal material may be selected from Ag, Au, Al, Mg and alloys thereof, and the light absorbing metal material may include Cr, Ni, Ti, Pt, Sn, Sb, Mo, W, V, Ta, Te, Ge, Si, and the like are included, and alloys therebetween, and silicides, carbides, nitrides, sulfides, etc. including these metals are also possible.

When using light-absorbing metal materials, nanodisk arrays can be constructed in the form of double layers of relatively low-loss metals and light-absorbing metal materials for the purpose of improving the modulation depth of the stopband curve or improving the line width. In this case, as shown in FIG. 7, the light absorption type metal material 132 is disposed on the low loss high reflectivity metal material 134. On the contrary, it is possible to have a structure in which a low loss high reflectance metal material is disposed on the light absorbing metal material. It is also possible to construct a single layer in the form of an alloy between two materials instead of a double layer structure.

 8 is a diagram illustrating an example of a spectral filter according to another embodiment. Referring to FIG. 8, in the case of constructing the spectral filter array 10 using the metal nanodisks, the metal patterns 120 are formed by forming the passivation layer 140 in a conformal manner for protecting the metal nanodisk layer. An example is shown. At this time, the passivation layer 140 is preferably formed to a thickness of less than 10 nm, more preferably several nm or less in order to minimize the effect on the optical characteristics of the nanodisk array filter.

The passivation layer 140 may be at least one selected from HfO ₂ , ZrO ₂ , ZnO, ZnSe, TiO ₂ , Al ₂ O ₃ , SiO _x , SOG, and the like, and may be formed using a metal surface oxidation method or an atomic layer deposition method. It is possible.

If necessary, the passivation layer 140 may further include a protective layer 150 such as a light transmissive polymer and a dielectric layer. In addition, a configuration including only the upper passivation layer 150 without the passivation layer 140 is possible.

The protective layer 150 may be silicon oxide, silicon nitride, magnesium fluoride, calcium fluoride, low molecular resin, or a polymer material. Examples of polymer materials include polymers including poly (dimethyl siloxane), polycarbonate, poly (vinyl phenyl sulfide), poly (methyl metharcylate) poly (vinyl alcohol), poly (vinyl butyral), poly (methyl acrylate) and air It is a polymer material which is not particularly limited including the coalescing.

9, 10 and 11 are graphs showing simulation results of transmission characteristics of stop bands for several metal materials and patterns according to an exemplary embodiment of the present invention. Hexagonal lattice structure was assumed as the metal nanostructure array formed on the Si substrate, and the period was calculated at 40 nm intervals from 1 um to 2.92 um. The transmission characteristics of the stop band were simulated in the mid-infrared region of 2 μm to 10 μm.

In order to implement the plasmonic filter array, the type of material forming the metal nanostructure array may be an important factor. In general, alkali and precious metal materials such as Al, Ag, Au, and Cu have been used as metal materials for causing surface plasmon resonance. In the present invention, in addition to these metals, transition metals such as Ta, W, Mo, Ni, Cr, and metal nitrides such as TiN and TiON, which have optical behaviors in the mid-infrared region, follow the druid free electron model. To be effective.

In particular, these materials are particularly preferred as mid-infrared wavelength band materials because of their excellent thermal and mechanical stability. In addition, these materials have the advantage that additional adhesive layers are not necessary because of their excellent adhesion to the substrate. Metal nitrides have an advantage that additional control of optical properties is possible through composition control.

9 and 10 show that the nanodisk having a duty cycle of 50% using Au and Ta as a metal pattern, respectively, and its transmittance dip curves are shown through theoretical computational calculations. FIG. 11 illustrates the results of calculating the optical transmittance curve by simulation after forming a nano-pattern having a duty cycle of 60% using Ni and W as a metal pattern.

9, 10, and 11 have slight differences in the spectral shape depending on the type of metal in the mid-infrared band of about 3 μm to 10 μm, but single-stop bands vary in their center wavelength continuously. You can see that it covers. Among the transition metals, Ta, W, and Mo, which are classified as heat-resistant metals, have the characteristics of optical constant dispersion in the mid-infrared band very similar to Au.

On the other hand, Ni and Cr, but the real value of the refractive index shows a relatively high value in the short wavelength, but is characterized by maintaining a lower value than the heat-resistant metal toward the longer wavelength. Therefore, it is not suitable for the nanohole array structure using the excitation of the surface plasmon wave traveling along the surface of the thin film, but it is very suitable as the nanodisc type stopband filter coupled with the lattice mode and using specific light absorption or light reflection phenomenon. It can be usefully used.

Alkali and noble metals such as Al, Ag, Au, etc., which are generally used as plasmonic metal materials, may use a heat light source and may have a problem in that heat resistance is somewhat insufficient in an infrared region where thermal excitation is expected due to the plasmon resonance effect.

12 is a scanning electron micrograph showing the thermal stability according to the selection of the metal material system constituting the nanostructure array. In order to confirm the heat resistance, 50 nm-thick Ta and Au nanodisk array patterns were formed on a Si substrate, and heat treatments were performed to compare shape changes. It was produced using the same template, but the shape after vacuum heat treatment at 900 ^° C. for 30 minutes showed that the Ta nanodisks remained unchanged, whereas in the case of Au, spherical particles were formed by self diffusion. Ag, which has a lower melting point and higher atomic mobility than Au, is expected to have stability problems even at a much lower temperature.

The fundamental vibration of the molecule shows the light absorption mode by harmonics and combinations in the near infrared region band (0.78-2 um). These harmonic and combination vibration modes have the disadvantages of low intensity and broad absorption line width compared to the mid-infrared band, but are well developed light sources and detectors, and thus become very effective spectroscopy areas. The infrared stopband filter according to the present invention can also operate as a near infrared spectrometer by shifting the wavelength band toward the near infrared region.

13 is a simulation result showing that the free spectrum range of the metal nanostructure array type stopband filter according to the present invention can be extended not only to the near infrared band but also to the visible wavelength range.

When the 50 nm-thick Al nanodisk array has a hexagonal lattice structure, the transmission spectrum is calculated by varying the period from 200 nm to 1500 nm at 100 nm intervals. Here, the duty cycle was fixed at 50%. It can be seen that it shows a single stopband characteristic that continuously varies from 0.35 um to 2 um band.

The spectral filter array of the present invention can be formed on the substrate to be manufactured as a separate spectral filter module. The substrate may be used as long as it is a transparent material in each wavelength band in the operating wavelength, and may be glass or a polymer as described above. As an example, a light-transmissive film is used as the substrate, and the light-transmissive film is preferably composed of a transparent or semitransparent polymer having appropriate adhesive force and shock absorbency. The spectral filter module is manufactured in a form in which the photodetector array is not integrated, and thus can be attached to the photodetector array module and used in actual use. When the spectral filter module is attached to the photodetector in actual use, for example, the optical filter module may be used by combining the optical filter module in front of the lens of the camera.

FIG. 14 shows an example of configuring the spectrometer 10000a through a one-dimensional linear array coupling between the spectroscopic filter array 1000a and the photodetector array 2000a of the present invention.

The schematic diagram shown in FIG. 14 shows a spectrometer 10000a including a spectral filter array 1000a composed of M spectral filters F and a photodetector array 2000a composed of M photodetection regions PD. do. The period of each unit spectral filter F may be determined to match the period of the photodetection area PD of the combined one-dimensional linear array photodetector array or to match the size of the plurality of photodetection areas PD. Can be. That is, the coupling between the unit spectroscopic filter and the photodetection pixel may be 1: 1 or 1: N (N is 2 or more) coupling.

15 shows an example of configuring the spectrometer 10000b through a two-dimensional coupling between the spectroscopic filter array 1000b and the photodetector array 2000b of the present invention. Compared to the spectrometer 10000a of the one-dimensional coupling, it is advantageous for integration, and it is advantageous for coupling with a conventional CMOS image sensor, a thermal imaging camera, and the like.

Meanwhile, the spectrometers 10000a and 10000b of FIGS. 14 and 15 include the spectroscopic filter arrays 1000a and 1000b and the photodetector arrays 2000a and 2000b of the present invention. The spectrometers 10000a and 10000b may be spectrometer chips.

In the spectrometers 10000a and 10000b of the present invention, the spectral filter arrays 1000a and 1000b may be formed of a plurality of unit spectral filters F. Details of the spectroscopic filters F have been described above. The plurality of spectral filters F may form a stopband filter array structure by continuously forming a stopband characteristic of varying a center wavelength.

The photodetector array 2000a is arranged such that the plurality of photodetection regions PD correspond to the plurality of unit spectral filters F, and is installed to detect light passing through each unit spectral filter. The photodetector array 2000a is disposed to be spaced apart from the spectroscopic filter array 1000a by a predetermined distance. In other variations, the photodetector array 2000a may be manufactured to be in direct contact with the spectral filter array 1000a. The photodetection area PD may be a unit pixel.

The spectrometers 10000a and 10000b according to the present invention can output subtractive intensity of light incident through a stopband filter in which the center wavelength is moved slightly by position in one direction of the spectroscopic filter in measuring an object spectrum. Let's do it. As a result, the intensity distribution according to the wavelength of light is shown to be inversely related to the case of a conventional transmission band filter array-based spectrometer, and serves as a spectrometer that restores an object spectrum by applying a subsequent digital signal processing algorithm. It is possible to implement a spectrometer based. The processing unit 330 of FIG. 3 performs an integral function of reconstructing the spectrum of incident light using the optical signal detected from the photodetector array.

Hereinafter, a spectroscopic method using a spectrometer according to an embodiment of the present invention will be described. 16 is a flowchart of a spectroscopic method according to an embodiment of the present invention.

First, the object spectrum is incident on the spectrometer (S100). The spectrometer includes a photodetector array having photodetection regions corresponding to each of the spectroscopic filter array and the unit spectroscopic filters.

The spectra of the incident object may selectively generate light reflection or light absorption through the spectroscopic filter array (S110). This characteristic is a characteristic of the above-mentioned "stop band" filter, the unit spectral filters have a characteristic that the transmittance according to the wavelength has a peak in the reverse direction so that light of a specific wavelength band does not transmit.

Next, the light spectrum signal transmitted through the spectroscopic filter array is detected by the photodetector array (S120). Then, the spectrum of the object is restored by the signal restoration algorithm (S130).

Hereinafter, the signal restoration principle in the spectrometer based on the stopband filter array according to the present invention will be described mathematically with reference to FIG.

If the spectrum of the object to be analyzed is s (λ), the transmission function of the individual filters F is f _i (λ), and the sensitivity function of the photodetector PD is d _i (λ), the spectrum of the object is filtered. The detection signal r _i generated when passing through the photodetector is expressed by the following relation (1), and can be developed by the determinant such as equation (2).

(One)

(2)

In general, since the number M of filters is smaller than the number N of wavelength samplings, the linear algebraic expression of Equation (2) results in an ill-posed problem. Since there is no explicit inverse of D (λ) with MXN (M <N) size, pseudo inverse can be used to recover the spectral signal, but it is very susceptible to small fluctuations or system noise, making it unstable. Results are shown.

The regularization technique is used to obtain more effective and numerically stable solutions. The most representative method is Tikhonov regularization. This method recovers the spectrum of the object to be analyzed by determining the solution S _α that minimizes the sum of residual norm and side constraint norm as shown in Eq. (3). Here, α is a regularization factor that determines the weight of side constraint minimization versus minimization of residual norm, and there is an optimal value to obtain a robust solution.

(3)

Singular value decomposition (SVD) and L-curve analysis can be used to adapt the system to determine the optimal regularization factor for itself and to enable real-time spectrum recovery.

The L-curve method solves the Tikhonov regularization equation when substituting and increasing the value of α and reconstructs the residual norm

And Solution norm

After substituting in, and plotting on the log scale coordinate axis, L-curve-shaped graph is obtained.

The method of calculating corner values is to take the log scale values of residual norm and solution norm as variables and determine a with the smallest radius of curvature. Substitute this value in Tikhonov regularization

By recovering the object spectrum can be restored.

Using this regularization technique, it is possible to restore the spectral spectrum with a relatively high resolution while using a non-ideal filter array having a wide half width. The signal restoration algorithm is not limited to the illustrated regularization technique and can be applied to various techniques.

17 shows an example of a calculation result for explaining a signal restoration principle of a spectrometer using a plasmonic stopband filter array according to the present invention.

The subject spectrum to be analyzed was assumed to have two separate peaks as shown in the image on the top left. When this object spectrum passes through the stopband filter array shown in the upper right, the intensity distribution measured by the photodetector array through each filter is determined by Equation (1) and is distorted or unclear due to the filter function. Indicates the distribution of. In this case, after substituting the information of the transmission spectrum f _i (λ) of the individual filter and the spectral sensitivity function d _i (λ) of the photodetector into equation (2), the digital signal processing algorithm of equation (3) is performed. As shown in the lower left, it is possible to accurately restore the object spectrum. The transmission spectrum f _i (λ) of the individual filter is determined by using an optical system previously measured for each filter area (for example, by using a spectroscopic microscope for each filter area) or by using a photodetector. The spectral sensitivity function (d _i (λ)) can be measured using the value provided by the manufacturer or by measuring the ratio of detector output intensity to wavelength intensity of each light source using a monochromator. Alternatively, the D _i (λ) value, which is the intensity distribution reaching the photodetector region through each filter region by varying the wavelength of incident light through a monochromator in the combined or integrated state with the photodetector array, is obtained. It is also possible to use directly measured.

Since the intensity distribution measured in the actual photodetector array includes system noise and the like, it is preferable to add a noise canceling algorithm for stabilizing the restored signal.

Since the plasmonic filter can change the resonant wavelength in the wide-range range only by adjusting the two-dimensional horizontal structure without changing the vertical structure, the highly integrated band stop having different spectroscopic characteristics even with a low-cost single layer process using photolithography or nanoimprint, etc. The advantage is that an array of filters can be formed.

 18 and 19 are graphs for comparing the transmission band type filter array and the stop band type filter array in the visible-near-infrared wavelength band. A calculation example for explaining the difference between the signal acquisition and the spectral restoration process in a spectrometer using a transmission band filter array and a stopband filter array is shown.

 As the transmission band type and the stop band type filter, Al nanohole array and Al nanodisk array were selected, and the transmission spectrum according to the lattice period variation was calculated by FDTD computer simulation. In both cases, the hexagonal lattice was assumed, and the period was varied from 200 nm to 700 nm at 5 nm intervals. A glass substrate was used and the Al thickness was the same at 50 nm and the duty cycle was fixed at 50%.

18 shows the filter function light transmission spectrum of the transmission band filter array composed of the Al metal nano hole array. It can be seen that the transmission band due to the EOT phenomenon is continuously changed according to the period. When a transmission band type filter array is used, the intensity signal at a specific wavelength of the object spectrum is determined from the intensity of light detected through a filter forming a transmission band at that wavelength. If the half width of the transmission band filter is very narrow, such as a delta function, the spectrum of the object may be reproduced by directly measuring the intensity distribution detected for each center wavelength of the transmission band of the filter array.

However, in the case of using a non-ideal filter array having a wide half-width as shown in the left graph of FIG. 19, the signal distribution measured by the photodetector is significantly distorted out of the object spectrum due to overlap of transmission bands between neighboring filters. .

Assuming that the object spectrum is composed of two Gaussian peak functions separated from each other as shown in the right graph of FIG. 18, the intensity distribution for each filter measured in the photodetector array through the filter function of FIG. 18 is as shown in the center graph of FIG. 18. appear. Although there is a distortion of the spectrum, the characteristic of the transmission band filter reflects the shape of the peak function of the object spectrum.

The photodetector measurement signal is substituted into Equation (2), and the object spectrum is restored by finding a solution using a regularization technique. In the right graph of FIG. 18, the reconstructed spectrum is illustrated as the object spectrum. The two curves are nearly identical, indicating that the spectral recovery is very good.

19 shows a signal restoration process using a stopband filter array. The left graph of FIG. 19 shows the filter function of the stopband filter array. The intensity distribution observed in the photodetector array through the stopband filter array for the same object spectrum as in FIG. 18 is the same as the center curve in FIG. 19. In contrast to the case where the transmission band filter is used, it can be seen that the peak function of the object spectrum appears in the photodetector in the form of reverse dip curve. That is, the signal measured by the photodetector through the stopband filter array is characterized in that the intensity distribution in the reverse phase form as opposed to the case of the transmission band filter array. The graph on the right side of FIG. 19 confirms that the spectrum recovery by the digital signal processing algorithm is performed well even in the case of the stopband filter array.

FIG. 20 is a schematic diagram for explaining gain in terms of detection limit of the spectral signal compared to the transmission band filter array in the wavelength range where the intensity of the light source and the sensitivity index of the photodetector decrease when using the stopband filter array according to the present invention; FIG. to be.

For ease of explanation, FIG. 20 shows the wavelength-specific sensitivity index of a typical Si-CMOS image sensor. Due to the energy band structure of Si semiconductors, it can be seen that the quantum efficiency drops rapidly toward the near-infrared wavelength band where the natural vibration mode of chemical molecules can be observed. Therefore, when the transmission band filter array is used in this section, there is a disadvantage in that spectrum analysis becomes very difficult due to the detection limit of the detection device.

On the other hand, because the stopband blocks the light of the designed central wavelength band and transmits the light of the remaining bands, the object signal information on the corresponding wavelength has a characteristic of being traced back from the transmission intensity distribution in the neighboring wavelength band. Therefore, even in the near-infrared region where the sensitivity index of the detection element itself is very low, such as the Si-CMOS image sensor, the signal analysis is inferred from the light intensity distribution in the other wavelength bands, not the light intensity detected in the wavelength band. There is a big gain in terms of band scalability and detection limit.

FIG. 21 shows the calculated filter function when the duty cycle is reduced to 30% and the number of filters is reduced to 50 in the Al nanodisk array filter shown in FIG. 19. FIG. 22 shows a graph of the restored function when the duty cycle is 30% and 50% in the Al nanodisk array filter.

Referring to FIG. 21, it can be seen that as the duty cycle decreases, the modulation depth of the stopband dip becomes shallow while the line width becomes narrow and sharp. In this case, it can be seen that the overlapping degree of the stopband curve between neighboring filters decreases. In the case of 50% duty cycle, even when the number of filters is 50, the overlap between neighboring filters is largely generated by 80% or more due to the effect of widening the band line width. In both cases, spectral restoration was performed using the spectrum of the white LED light source as an object. As a result, when the duty cycle is 30% with low overlap, the restored spectrum is accompanied by a lot of noise signals, and when the duty cycle is 50%, no noise is obtained. It can be confirmed that the spectral restoration has been satisfactorily achieved. In filter array-based spectrometers, spectral recovery resolution is known to improve as the bandwidth of the filter used is narrower and the number of filters increases. However, even if the bandwidth of the filter function is narrow, if the overlap between neighboring filters is small, it can be seen that a situation in which signal restoration becomes disadvantageous compared with the case of using a filter having a wider bandwidth.

FIG. 23 is a graph comparing spectral changes according to nanodisk shapes in a filter of a nanodisk array structure having a hexagonal lattice structure. FIG. FIG. 23 shows only the calculation results for the circular and hexagonal disk structures, but when the disk shape is isotropic symmetrical structure and the duty cycle is similar, the polygonal and cross-shaped disks show almost similar filter spectrum regardless of the shape. And it was found.

 This is because when the metal nanodisk array forms a periodic lattice structure, its center wavelength and optical spectrum are dominantly dependent on the lattice period. As a result, it is possible to use a single-shaped disk structure in the filter array configuration process or to mix and configure a disk array having two or more shapes as shown in FIG. FIG. 24 shows an example in which a spectroscopic filter array is formed by mixing into two or more disk arrays. Such a mixed configuration has an effect of providing process convenience such as reducing process complexity for manufacturing a circular disk and shortening process time when manufacturing a filter array.

FIG. 25 is a graph showing the distribution of the absorption coefficient versus the refractive index of optical constant dispersion characteristics of metals, dielectrics, and semiconductor materials. The complex optical constant values from the infrared band to the near infrared band of 1300 nm are shown on two-dimensional coordinates. From the features of the optical constant combination, it can be divided into three zones. First, the region labeled I is characterized by an absorption rate of less than one and converging to zero, which is the imaginary term of the complex refractive index, most of which is an optically transparent dielectric. In addition, semiconductor materials may also belong to this group in the wavelength region where the wavelength of light is smaller than the band gap.

Zone II is characterized in that the refractive index value is less than 1 or close to 0, as opposed to zone I, and is a low loss high reflectivity precious metal material widely used as a plasmonic metal. On the other hand, region III is a region in which both the refractive index and the absorptance exhibit a value of a certain size or more, which corresponds to most light absorbing metals and semiconductor materials in the wavelength band below the band gap. Materials of these III zones may be utilized as a material for constructing a nanodisk array for implementing a stopband type filter array according to the present invention. Light absorbing metal materials include Cr, Ni, Ti, Pt, Sn, Sb, Mo, W, V, Ta, Te, Ge, Si, and the like, and alloys therebetween and silicides, carbides, nitrides containing these metals. Also, sulfide, sulfide, etc. can be used without distinction if the distribution of refractive index and absorption rate satisfies the conditions of zone III in the operating wavelength band.

FIG. 26 is a light transmittance spectrum of a hexagonal lattice-structured nanodisk array using Cr and Ti and calculated in the visible-near-infrared band. In addition to the plasmonic material, which is represented by a common low-loss noble metal material, a hexagonal lattice-type nanodisk array is formed using Cr and Ti, which are light-absorbing metal materials, and computed in the visible-near-infrared band using computer simulation. The transmittance spectrum was obtained. Compared to the Al nanodisk array of FIG. 17, the modulation depth is smaller and the line width is wider. However, since the stopband formation is clear and the wavelength variability according to the lattice period is continuous and clear, it can be used as a filter array for the spectroscope. Do.

27 is a graph of light transmittance and light reflectivity of a nanodisk array calculated using tungsten (W). Tungsten (W) is a material commonly used in semiconductor processes. The optical transmittance and the light reflectivity of the nanodisk array calculated using tungsten (W) show that, unlike the low-loss plasmonic metal materials, the reflectance peak curve is greatly attenuated by the light loss of the material itself, while the stopband on the transmission curve is shown. The curves can be seen to be characterized by relatively very pronounced by the increased light absorption effect. This feature may be advantageous in terms of suppressing unnecessary noise elements caused by reflected light when implementing a spectrometer chip.

FIG. 28 is a graph illustrating spectral recovery performance by fabricating a tungsten nanodisk array of FIG. 26 using a stopband filter array and applying a digital signal processing algorithm. The spectral reconstruction performance was tested using a digital signal processing algorithm using a stopband type filter array composed of 100 filters in the lattice period from 200 nm to 700 nm. Assuming a white LED spectrum as an object and calculating a filter function, it can be seen that the spectral restoration is relatively superior as in the case of using the Al nanodisk array filter of FIG. 20. This proves that the light-absorbing metal material can be used as the stopband filter material of the nanodisc array structure for the transmission-on-chip spectrometer.

Although preferred embodiments of the spectrometer according to the present invention have been described above, the present invention is not limited thereto, and various modifications can be made within the scope of the claims and the detailed description of the invention and the accompanying drawings. This also belongs to the present invention.

Claims (20)

1. A first unit spectroscopic filter that absorbs or reflects light of a part of a wavelength band of an incident light spectrum of the object;

   A second unit spectroscopic filter that absorbs light or reflects light in a wavelength band different from the wavelength band;

   A first photodetector for detecting a first light spectrum passing through the first unit spectroscopic filter;

   A second photodetector for detecting a second light spectrum passing through the second unit spectroscopic filter; And

   And a processing unit performing a function of restoring a light spectrum of the object incident from the spectra of light detected from the first and second photodetectors.
2. According to claim 1,

   And the first unit spectroscopic filter and the second unit spectroscopic filter are arranged in a metal pattern having a predetermined shape periodically.
3. The method of claim 2,

   And the metal patterns of the first unit spectroscopic filter and the metal patterns of the second unit spectroscopic filter have different periods.
4. The method of claim 2,

   And a first photodetector and a second photodetector comprising some photodetecting pixels of the CMOS image sensor.
5. The method of claim 2,

   The metal pattern is a spectrometer, characterized in that composed of a material selected from the alloy containing at least one of Au, Ag, Al, Cu plasmonic metal.
6. The method of claim 2,

   The metal patterns include Cr, Ni, Ti, Pt, Sn, Sb, Mo, W, V, Ta, Te, Ge, Si, or at least one of these metals having both high absorption and refractive indexes in the visible and near infrared bands. Spectrometer, characterized in that consisting of a material selected from the alloy.
7. The method of claim 2,

   And at least one of the metal patterns is selected from the group consisting of Ta, W, Mo, Ni, Cr, TiN, and TiON, whose optical behavior follows the druid free electron model in the mid-infrared band.
8. The method of claim 2,

   The metal pattern is composed of at least a double layer, the spectrometer, characterized in that the low-loss high reflectivity metal material and the light-absorbing metal material laminated.
9. The method of claim 8,

   The low loss high reflectivity metal material may be selected from Ag, Au, Al, Mg, and alloys thereof, and the light absorbing metal material may include Cr, Ni, Ti, Pt, Sn, Sb, Mo, W, And a silicide, carbide, nitride, or sulfide comprising V, Ta, Te, Ge, Si, and alloys therebetween and these metals.
10. The method of claim 2,

    And the metal patterns of the first unit spectroscopic filter and the metal patterns of the second unit spectroscopic filter have the same duty cycle.
11.  The method of claim 2,

    And a period of the metal patterns of the first and second unit spectroscopic filters is between 100 nm and 800 nm.
12. The method of claim 2,

    And the first unit spectroscopic filter and the second unit spectroscopic filter further comprise a passivation layer.
13. The method of claim 11,

    The passivation layer is spectrometer, characterized in that consisting of a material selected from the alloy consisting of HfO ₂ , ZrO ₂ , ZnO, ZnSe, TiO ₂ , Al ₂ O ₃ , SiO _x , SOG or at least two of them.
14. The method of claim 2,

    And the first unit spectroscopic filter and the second unit spectroscopic filter further comprise a protective layer.
15. The method of claim 12,

    The protective layer is a spectrometer, characterized in that the low refractive index silicon oxide, silicon nitride film, magnesium fluoride, calcium fluoride, low molecular resin, or a polymer material.
16. The method of claim 1, wherein the process unit,

    Calculating the intensity of light absorbed or reflected by the first unit spectroscopic filter from the spectrum of light of the first photodetector;

    Calculating the intensity of light absorbed or reflected by the second unit spectroscopic filter from the spectrum of light of the second photodetector; And

    The spectroscope of claim 1, wherein the spectrometer recovers the light spectrum of the incident object from the intensity of light absorbed or reflected by the first and second unit spectroscopic filters.
17. Incident light spectra of the object onto the first and second unit spectroscopic filters;

    The first unit spectroscopic filter absorbs or reflects light of a portion of the wavelength band, and the second unit spectroscopic filter absorbs or reflects light of a wavelength of a wavelength band different from that of the partial wavelength band;

    A first photodetector detecting a first light spectrum passing through the first unit spectroscopic filter and a second photodetector detecting a second light spectrum passing through the second unit spectroscopic filter; And

    And reconstructing the light spectrum of the object incident from the spectra of the light detected from the first and second photodetectors.
18. The method of claim 17,

    Restoring the light spectrum of the object,

    Calculating an intensity of light absorbed or reflected by the first unit spectroscopic filter from the first light spectrum of the first photodetector;

    Calculating an intensity of light absorbed or reflected by the second unit spectroscopic filter from the second light spectrum of the second photodetector; And

    And a spectrometer for recovering the spectrum of the light incident from the intensity of the light absorbed or reflected by the restored first and second unit spectroscopic filters.
19. The method of claim 18,

    Restoring the spectrum of the object is a spectrum measurement method using a spectroscope, characterized in that a direct readout or regularization technique is used.
20. The method of claim 18,

    The restoring of the light spectrum of the object may be performed by substituting the transmission spectrum f _i (λ) of the individual filter and the spectral sensitivity function d _i (λ) of the photodetector in the following equation and measuring the detected signal r A spectrum measurement method using a spectroscope, which can be derived using _i .

    

    Where D _i (l) is f _i (l) d _i (l),

    If the spectrum of the object to be analyzed is s (λ), the transmission function of each filter F is f _i (λ), the noise is n _i , and the sensitivity function of the photodetector PD is d _i (λ), r _i is the detection signal generated when the spectrum of the object passes through the filter and reaches the photodetector

Images