JP7342390B2 - Optical filter and its manufacturing method - Google Patents

Description

The present invention relates to an optical filter and a method for manufacturing the same, and more specifically, the present invention has both a function of transmitting only incident light (signal light) in a transmission range and a function of absorbing incident light (noise light) in a stop range. The present invention relates to an optical filter and a method for manufacturing the same.

Optical filters that include alternately laminated films (so-called "one-dimensional photonic crystals") made of two or more types of transparent materials with different refractive indexes generally function as angle-selective filters (ASF). It also functions as a wavelength-selective filter (WSF).
Here, an "angle selection filter" is an optical filter that has the function of transmitting light with an incident angle within a specific range (transmission range) and the function of reflecting light with an incidence angle outside of that range (stopband). means.
A "wavelength selection filter" refers to an optical filter that has the function of transmitting light with wavelengths within a specific range (transmission range) and the function of reflecting light with wavelengths other than that (stopband). Examples of the WSF include a bandpass filter, shortpass filter, and longpass filter. When a plurality of lights having the same wavelength and different angles of incidence are incident on a WSF, the difference in the angle of incidence is equivalent to the difference in wavelength, so the WSF also functions as an ASF.

An optical filter that functions as an ASF is used to improve the S/N of a directional photodetector. ASF improves S/N because signal light incident from a specific direction passes through and reaches the detector, but noise light incident from other directions is reflected and does not reach the detector. It is. It has also been proposed to apply an optical filter that functions as a WSF to light-trapping of photoelectric conversion elements.

For example, Non-Patent Document 1 discloses a Si photoelectric conversion cell in which a bandpass filter is formed on the front surface and a diffuse reflection surface is formed on the back surface.
The same document states that when monochromatic light with a wavelength of 1060 nm is incident on such a Si photoelectric conversion cell at an incident angle of 8 degrees, the incident light is almost completely absorbed even if the thickness of the Si wafer is only 45 μm. points are disclosed.

Furthermore, Non-Patent Document 2 discloses a c-Si photoelectric conversion cell in which a short pass filter is formed on the front surface and a diffuse reflection surface is formed on the back surface.
The same document states that when a c-Si photoelectric conversion cell is irradiated with monochromatic light that changes the incident angle, a short-pass filter is designed to function as a polarization-insensitive angle-selective filter. It is described that it exhibits a high light confinement effect.

Furthermore, Non-Patent Document 3 discloses an angle selective filter (ASF) using a one-dimensional photonic crystal.
The document states that ASF for a specific wavelength can be used to improve the conversion efficiency of photoelectric conversion elements for monochromatic light irradiation such as solar pumped lasers, to reduce noise and improve sensitivity of directional sensors, and to improve directivity of light emitting elements. Are listed.

When an optical filter that functions as an ASF or WSF is used to improve the S/N ratio of a detector, the noise light reflected by the optical filter may be absorbed by components around the detector (for example, when the detector is installed). (e.g., the device housing, optical window, etc.) and becomes stray light, which may reach the detector and cause an increase in noise. In particular, as devices become smaller, they become more susceptible to stray light. However, there has never been an example of an optical filter proposed that can reduce the influence of noise light without interfering with the transmission of signal light.

Furthermore, in order to stably obtain such functions, it is necessary to precisely control the thickness of each layer constituting the optical filter. However, the thickness of each layer constituting the optical filter is approximately several nm, which is extremely thin. Generally, the optical properties of a thin film are significantly affected by a slight deviation in thickness or a slight deviation in the optical properties of the material itself. Therefore, it is not easy to stably produce an optical filter with desired characteristics.

Y. Takeda and T. Ito, Jpn. J. Appl. Phys. 54, 08KD13(2015) Y. Takeda, H. Iizuka, N. Yamada, and T. Ito, Appl. Opt. 56, 5761(2017) Takeda, Iizuka, Ito, Yamada, Ito, Motohiro, Proceedings of the 65th Japan Society of Applied Physics Spring Academic Conference, 19p-C301-15 (March 17-20, 2018)

The problem to be solved by the present invention is to develop a novel optical filter that has both the function of transmitting only incident light (signal light) in the transmission range and the function of absorbing incident light (noise light) in the stopband. It is about providing.
Further, another problem to be solved by the present invention is to provide an optical filter that can stably obtain such a function even when the thickness of each layer constituting the optical filter changes, and the like. The purpose is to provide a manufacturing method.

In order to solve the above problems, the optical filter according to the present invention has the following configuration.
(1) The optical filter is
a filter layer including alternately laminated films made of two or more types of transparent materials (A) having different refractive indexes;
and a light absorption layer formed on the incident surface side of the filter layer,
The filter layer is configured to transmit light (signal light) in a transmission range of the filter layer among the incident light,
The light absorption layer is made of a layer for absorbing light (noise light) that is in a blocking range of the filter layer, out of the incident light.
(2) The light absorption layer is made of a light absorption material that satisfies the following formula (1).
πk _A /n _A ≧1 ... (1)
however,
n _A is the refractive index of the light absorption layer when the light absorption layer is irradiated with the incident light;
k _A is an extinction coefficient of the light absorption layer when the light absorption layer is irradiated with the incident light.

The light absorbing material is
(a) transition metal (A),
(b) a cermet containing a transition metal (B) and a transparent material (B), or
(c) a material capable of irreversibly changing the extinction coefficient by heating;
is preferred.
Further, the light absorbing material is preferably one or more selected from the group consisting of Ge ₂ Sb ₂ Te ₅ , GeSb ₂ Te ₄ , GeSb ₄ Te ₇ , and AgInSbTe-based optical recording materials.

The method for manufacturing an optical filter according to the present invention includes:
A light-absorbing layer containing a light-absorbing material whose extinction coefficient can be irreversibly changed by heating on the entrance surface side of a filter layer including alternately laminated layers of two or more transparent materials (A) having different refractive indexes. A laminate production step of producing a laminate in which
and a heat treatment step of adjusting the filter characteristics of the laminate by heat-treating the laminate to obtain an optical filter according to the present invention.

An alternately laminated film made of two or more types of transparent materials (A) having different refractive indexes has a function as an angle selection filter. When light is incident on a filter layer including such alternately laminated films, the light is repeatedly reflected at the interface between different materials. As a result, the light intensity (or the electric field amplitude of the incident light) within the filter layer is not uniform and varies periodically along the depth direction. Furthermore, the distribution of light intensity also changes depending on the angle of incidence of light.

When a light absorption layer that satisfies predetermined conditions is formed on the entrance surface side of such a filter layer, when the incident light (signal light) in the transmission region is irradiated, the light absorption layer near the entrance surface (that is, the light absorption layer (near the installed position) the light intensity becomes weaker. As a result, a large proportion of the signal light passes through the light absorption layer without being absorbed by the light absorption layer.
On the other hand, when incident light (noise light) in the stop band is irradiated, the light intensity near the incident surface increases. As a result, the proportion of noise light absorbed by the light absorption layer increases.
Therefore, when such an optical filter is applied to a directional photodetector or the like, the influence of noise light can be reduced without significantly interfering with the transmission of signal light.

Furthermore, when producing such an optical filter, if an error occurs in the thickness of the light absorption layer, desired optical characteristics may not be obtained. On the other hand, if a material whose extinction coefficient can be irreversibly changed by heating is used as the light-absorbing material constituting the light-absorbing layer, an error may occur in the thickness of the light-absorbing layer during film formation. Even if it is, the thickness error can be compensated for by heat treatment. Therefore, optical filters having desired characteristics can be manufactured with high yield.

Incident transmittance T(θ) and reflectance R(θ) of a reflective angle selective filter (R-ASF) (Comparative Example 1) consisting of a 28-layer TiO ₂ /SiO ₂ multilayer film that does not include a light absorption layer It is a figure showing angle dependence. FIG. 2(A) shows a light intensity distribution when light is incident on the R-ASF of FIG. 1 at an incident angle θ=9°. FIG. 2(B) is an enlarged view of the vicinity of the entrance plane (depth=0) in FIG. 2(A). FIG. 3(A) shows the light intensity distribution when light is incident on the R-ASF of FIG. 1 at an incident angle θ=24.5°. FIG. 3(B) is an enlarged view of the vicinity of the entrance plane (depth=0) in FIG. 3(A).

FIG. 4A shows the signal light intensity I _signal of an absorption angle selective filter (A-ASF) in which a metal layer is formed on the incident surface under the condition that the noise light intensity I _noise ≦0.2. FIG. 4B shows the noise light intensity I _noise of the A-ASF in which a metal layer is formed on the incident surface under the condition that the signal light intensity I _signal ≧0.75.

An absorption angle selection filter (A-ASF) (Example 1) that combines a Cr layer (2.4 nm) and a reflection angle selection filter (R-ASF) consisting of a 28-layer TiO ₂ /SiO ₂ multilayer film. FIG. 3 is a diagram showing the dependence of transmittance T(θ), reflectance R(θ), and absorption rate A(θ) on the angle of incidence. FIG. 6(A) shows a light intensity distribution when light is incident on the A-ASF of FIG. 5 at an incident angle θ=9°. FIG. 6(B) is an enlarged view of the vicinity of the entrance plane (depth=0) in FIG. 6(A). FIG. 7(A) shows the light intensity distribution when light is incident on the A-ASF of FIG. 5 at an incident angle θ=24.5°. FIG. 7(B) is an enlarged view of the vicinity of the entrance plane (depth=0) in FIG. 7(A).

An absorption angle selection filter (A-ASF) (Example 2) that combines a Cr layer (1.7 nm) and a reflection angle selection filter (R-ASF) consisting of a 28-layer TiO ₂ /SiO ₂ multilayer film. FIG. 3 is a diagram showing the dependence of transmittance T(θ), reflectance R(θ), and absorption rate A(θ) on the angle of incidence. FIG. 9(A) shows a light intensity distribution when light is incident on the A-ASF of FIG. 8 at an incident angle θ=9°. FIG. 9(B) is an enlarged view of the vicinity of the entrance plane (depth=0) in FIG. 9(A). FIG. 10(A) shows the light intensity distribution when light is incident on the A-ASF of FIG. 8 at an incident angle θ=24.5°. FIG. 10(B) is an enlarged view of the vicinity of the entrance plane (depth=0) in FIG. 10(A).

FIG. 11A shows the signal light intensity I _signal of an absorption type angle selective filter (A-ASF) in which a cermet layer is formed on the incident surface under the condition that the noise light intensity I _noise ≦0.2. FIG. 11B shows the noise light intensity I _noise of the A-ASF in which a cermet layer is formed on the incident surface under the condition that the signal light intensity I _signal ≧0.75.

_{An absorption} type angle _{selection filter} ( _A- 3 is a diagram showing the dependence of the transmittance T(θ), the reflectance R(θ), and the absorption coefficient A(θ) on the incident angle of ASF) (Example 3). FIG. FIG. 13(A) shows the light intensity distribution when light is incident on the A-ASF of FIG. 12 at an incident angle θ=9°. FIG. 13(B) is an enlarged view of the vicinity of the entrance plane (depth=0) in FIG. 13(A). FIG. 14(A) shows the light intensity distribution when light is incident on the A-ASF of FIG. 12 at an incident angle θ=24.5°. FIG. 14(B) is an enlarged view of the vicinity of the entrance plane (depth=0) in FIG. 14(A).

An _absorptive angle _{selective filter ( A-} FIG. 4 is a diagram showing the dependence of the transmittance T(θ), reflectance R(θ), and absorption coefficient A(θ) on the incident angle of ASF) (Example 4). FIG. 16(A) shows the light intensity distribution when light is incident on the A-ASF of FIG. 15 at an incident angle θ=9°. FIG. 16(B) is an enlarged view of the vicinity of the entrance plane (depth=0) in FIG. 16(A). FIG. 17(A) shows the light intensity distribution when light is incident on the A-ASF of FIG. 15 at an incident angle θ=24.5°. FIG. 17(B) is an enlarged view of the vicinity of the entrance plane (depth=0) in FIG. 17(A).

This is an absorption type angle selective filter (A-ASF) that combines a reflective angle selective filter (R-ASF) consisting of a Cr layer and a 28-layer TiO ₂ /SiO ₂ multilayer film. Transmittance T(θ), reflectance R(θ), and absorptance A of the Cr layer formed in the fifth layer (FIG. 18(A), Example 5) or the fifth layer (FIG. 18(B), Example 6) FIG. 3 is a diagram showing the incident angle dependence of (θ). This is an absorption angle selective filter (A-ASF) that combines a Ti _0.74 Si _0.26 O _1.91 layer and a reflective angle selective filter (R-ASF) consisting of a 28-layer TiO ₂ /SiO ₂ multilayer film. Transmittance T(θ) and reflection of a Ti _0.74 Si _0.26 O _1.91 layer formed on the third layer (FIG. 19(A), Example 7) or the fifth layer (FIG. 19(B), Example 8) FIG. 3 is a diagram showing the dependence of the rate R(θ) and the absorption rate A(θ) on the incident angle.

Transmittance T of an absorption angle selective filter (A-ASF) (Example 9) that combines a Cr layer and a double resonator band pass filter (2C-BPF) consisting of a 32-layer TiO ₂ /SiO ₂ multilayer film (θ), reflectance R(θ), and absorption rate A(θ) on incident angle dependence. Absorption angle selective filter (A-ASF) (Example 10) combining a Ti _0.74 Si _0.26 O _1.91 layer and a double resonator band pass filter (2C-BPF) consisting of a 32-layer TiO ₂ /SiO ₂ multilayer film ) is a diagram showing the dependence of the transmittance T(θ), the reflectance R(θ), and the absorption coefficient A(θ) on the incident angle.

Figure 22(A) shows the transmittance of an absorption angle selective filter (A-ASF) (Example 11) that combines a Cr layer and a short pass filter (SPF) consisting of a 40-layer TiO ₂ /SiO ₂ multilayer film. FIG. 3 is a diagram showing the dependence of T(θ), reflectance R(θ), and absorptance A(θ) on the incident angle. FIG. 22(B) shows the optical thickness of each layer other than the Cr layer of the A-ASF shown in FIG. 22(A). FIG _{. 23} (A) shows _{an absorption} angle selective filter (A- _ASF ) (Example 12) is a diagram showing the dependence of the transmittance T(θ), the reflectance R(θ), and the absorption coefficient A(θ) on the incident angle. FIG. 23(B) shows the optical thickness of each layer other than the Ti _0.74 Si _0.26 O _1.91 layer of A-ASF shown in FIG. 23(A).

FIG. 24(A) shows the transmittance T of an absorption angle selective filter (A-ASF) (Example 13) that combines a Cr layer and a long pass filter (LPF) consisting of a 40-layer TiO ₂ /SiO ₂ multilayer film. (θ), reflectance R(θ), and absorption rate A(θ) on incident angle dependence. FIG. 24(B) shows the optical thickness of each layer other than the Cr layer of the A-ASF shown in FIG. 24(A). FIG. 25(A) shows an absorption angle selective filter (A-ASF) (Example 14) that combines a Ti _0.74 Si _0.26 O _1.91 layer and a long pass filter (LPF) consisting of a 40-layer TiO ₂ /SiO ₂ multilayer film. ) is a diagram showing the dependence of the transmittance T(θ), the reflectance R(θ), and the absorption coefficient A(θ) on the incident angle. FIG. 25(B) shows the optical thickness of each layer other than the Ti _0.74 Si _0.26 O _1.91 layer of A-ASF shown in FIG. 25(A).

Transmittance T(θ) when monochromatic light with a wavelength of 905 nm, 905.9 nm, or 906.8 nm is incident on an absorption angle selective filter (A-ASF) tuned for a wavelength of 905 nm (Fig. 26(A)), reflectance R(θ) (FIG. 26(B)), and absorption rate A(θ) (FIG. 26(C)).

FIG. 2 is a diagram showing the dependence of the complex refractive index (n+ik) of Ge ₂ Sb ₂ Te ₅ (GST) on heat treatment temperature. _{An absorptive} angle _{selective filter} ( _A -ASF) (Example 21-1) is the transmission spectrum T(θ) when heat treated at 130°C (θ _T1 ~θ _T2 = 6 ~ 12°, θ _B1 ~ θ _B2 =23~26°). This is the reflection spectrum R(θ) of A-ASF in FIG. 28 (θ _T1 to θ _T2 =6 to 12°, θ _B1 to θ _B2 =23 to 26°). This is the absorption spectrum A(θ) of A-ASF in FIG. 28 (θ _T1 to θ _T2 =6 to 12°, θ _B1 to θ _B2 =23 to 26°). This is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF (heat treatment temperature: 130° C.) in FIG. 28 and the optical characteristics (θ _T1 to θ _T2 = 6 to 12°, θ _B1 to θ _B2 =23~26°).

_{An absorptive} angle _{selective filter} ( _A -ASF) (Example 21-2) is the transmission spectrum T(θ) when heat-treated at an appropriate temperature (θ _T1 ~θ _T2 =6~12°, θ _B1 ~θ _B2 = 23~26°). This is the reflection spectrum R(θ) of A-ASF in FIG. 32 (θ _T1 to θ _T2 =6 to 12°, θ _B1 to θ _B2 =23 to 26°). This is the absorption spectrum A(θ) of A-ASF in FIG. 32 (θ _T1 to θ _T2 =6 to 12°, θ _B1 to θ _B2 =23 to 26°). This is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF in FIG. 32 and the appropriate heat treatment temperature (θ _T1 to θ _T2 = 6 to 12°, θ _B1 to θ _B2 = 23 to 26°. ). This is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF (heat treatment temperature: appropriate temperature) in FIG. 32 and the optical characteristics (θ _T1 to θ _T2 = 6 to 12°, θ _B1 to θ _B2 = 23~26°).

_{An absorptive} angle _{selective filter} ( _A -ASF) (Example 21-3) before heat treatment (θ _T1 to θ _T2 = 6 to 12°, θ _B1 to θ _B2 = 23 to 26°). This is the reflection spectrum R(θ) of A-ASF in FIG. 37 (θ _T1 to θ _T2 =6 to 12°, θ _B1 to θ _B2 =23 to 26°). This is the absorption spectrum A(θ) of A-ASF in FIG. 37 (θ _T1 to θ _T2 =6 to 12°, θ _B1 to θ _B2 =23 to 26°). 38 is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF in FIG. 37 and the reflectance R (30°) at an incident angle of 30° before heat treatment (θ _T1 ~θ _T2 = 6 ~ 12 °, θ _B1 ~θ _B2 = 23~26°). 38 is a diagram showing the relationship between the reflectance R (30°) at an incident angle of 30° before heat treatment of A-ASF in FIG. 37 and the appropriate heat treatment temperature (θ _T1 ~θ _T2 =6 to 12°, θ _B1 ~θ _B2 = 23~26°).

Absorption angle selective filter (A-ASF) (Example 22) that combines a Cr layer with different thickness and a double resonator band pass filter (2C-BPF) consisting of a 28-layer TiO ₂ /SiO ₂ multilayer film (θ _T1 to θ _T2 = 6 to 12 degrees, θ _B1 to θ _B2 = 23 to 26 degrees). This is the reflection spectrum R(θ) of A-ASF in FIG. 42 (θ _T1 to θ _T2 =6 to 12°, θ _B1 to θ _B2 =23 to 26°). This is the absorption spectrum A(θ) of A-ASF in FIG. 42 (θ _T1 to θ _T2 =6 to 12°, θ _B1 to θ _B2 =23 to 26°). 43 is a diagram showing the relationship between the error in the thickness of the Cr layer of the A-ASF of FIG. 42 and the optical characteristics (θ _T1 to θ _T2 = 6 to 12 degrees, θ _B1 to θ _B2 = 23 to 26 degrees). FIG.

_{An absorptive} angle _{selective filter} ( _A -ASF) (Example 23) is the transmission spectrum T(θ) when heat-treated at an appropriate temperature (θ _T1 ~θ _T2 = 0 ~ 10°, θ _B1 ~ θ _B2 = 10-30°). This is the reflection spectrum R(θ) of A-ASF in FIG. 46 (θ _T1 to θ _T2 =0 to 10°, θ _B1 to θ _B2 =10 to 30°). This is the absorption spectrum A(θ) of A-ASF in FIG. 46 (θ _T1 to θ _T2 =0 to 10°, θ _B1 to θ _B2 =10 to 30°). This is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF in FIG. 46 and the appropriate heat treatment temperature (θ _T1 to θ _T2 = 0 to 10°, θ _B1 to θ _B2 = 10 to 30°. ). This is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF (heat treatment temperature: appropriate temperature) in FIG. 46 and the optical characteristics (θ _T1 to θ _T2 =0 to 10°, θ _B1 to θ _B2 = 10~30°).

47 is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF in FIG. 46 and the reflectance R (30°) at an incident angle of 30° before heat treatment (θ _T1 ~θ _T2 =0 to 10 °, θ _B1 ~θ _B2 = 10~30°). 47 is a diagram showing the relationship between the reflectance R (30°) at an incident angle of 30° before heat treatment of A-ASF in FIG. 46 and the appropriate heat treatment temperature (θ _T1 ~θ _T2 =0~10°, θ _B1 ~θ _B2 = 10~30°).

Absorption angle selective filter (A-ASF) that combines Ge ₂ Sb ₂ Te ₅ (GST) layers with different thicknesses and a short pass filter (SPF) consisting of 40 TiO ₂ /SiO ₂ multilayer films (Example) 24), the transmission spectra T(θ) when heat treated at appropriate temperatures (θ _T1 to θ _T2 = 6 to 12°, θ _B1 to θ _B2 = 23 to 26°). This is the reflection spectrum R(θ) of A-ASF in FIG. 53 (θ _T1 to θ _T2 =6 to 12°, θ _B1 to θ _B2 =23 to 26°). This is the absorption spectrum A(θ) of A-ASF in FIG. 53 (θ _T1 to θ _T2 =6 to 12°, θ _B1 to θ _B2 =23 to 26°). A-FIG. 53 shows the optical thickness of each layer of the SPF included in the ASF.

This is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF in FIG. 53 and the appropriate heat treatment temperature (θ _T1 to θ _T2 = 6 to 12°, θ _B1 to θ _B2 = 23 to 26°. ). This is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF (heat treatment temperature: appropriate temperature) in FIG. 53 and the optical characteristics (θ _T1 to θ _T2 = 6 to 12°, θ _B1 to θ _B2 = 23~26°). 54 is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF in FIG. 53 and the reflectance R (30°) at an incident angle of 30° before heat treatment (θ _T1 ~θ _T2 = 6 ~ 12 °, θ _B1 ~θ _B2 = 23~26°). 54 is a diagram showing the relationship between the reflectance R (30°) at an incident angle of 30° before heat treatment of A-ASF in FIG. 53 and the appropriate heat treatment temperature (θ _T1 ~θ _T2 =6 to 12°, θ _B1 ~θ _B2 = 23~26°).

Absorption angle selective filter (A-ASF) that combines Ge ₂ Sb ₂ Te ₅ (GST) layers with different thicknesses and a short pass filter (SPF) consisting of 40 TiO ₂ /SiO ₂ multilayer films (Example) 25), the transmission spectra T(θ) when heat treated at appropriate temperatures (θ _T1 to θ _T2 =0 to 30°, θ _B1 to θ _B2 = 30 to 60°). This is the reflection spectrum R(θ) of A-ASF in FIG. 61 (θ _T1 to θ _T2 =0 to 30°, θ _B1 to θ _B2 =30 to 60°). This is the absorption spectrum A(θ) of A-ASF in FIG. 61 (θ _T1 to θ _T2 =0 to 30°, θ _B1 to θ _B2 =30 to 60°). This is the optical thickness of each layer of the SPF included in the A-ASF of FIG. 61.

This is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF in FIG. 61 and the appropriate heat treatment temperature (θ _T1 to θ _T2 = 0 to 30°, θ _B1 to θ _B2 = 30 to 60°. ). This is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF (heat treatment temperature: appropriate temperature) in FIG. 61 and the optical characteristics (θ _T1 to θ _T2 =0 to 30°, θ _B1 to θ _B2 = 30~60°). 62 is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF in FIG. 61 and the reflectance R (45°) at an incident angle of 45° before heat treatment (θ _T1 ~θ _T2 =0 ~ 30 °, θ _B1 ~θ _B2 = 30~60°). 62 is a diagram showing the relationship between the reflectance R (45°) at an incident angle of 45° before heat treatment of A-ASF in FIG. 61 and the appropriate heat treatment temperature (θ _T1 ~θ _T2 =0~30°, θ _B1 ~θ _B2 = 30~60°).

Absorption angle selective filter (A-ASF) (Example 26), which is a combination of Ge ₂ Sb ₂ Te ₅ (GST) layers with different thicknesses and a long pass filter (LPF) consisting of 40 TiO ₂ /SiO ₂ multilayer films. ) are the transmission spectra T(θ) when heat treated at appropriate temperatures (θ _T1 to θ _T2 = 10 to 60°, θ _B1 to θ _B2 = 0 to 10°). This is the reflection spectrum R(θ) of A-ASF in FIG. 69 (θ _T1 to θ _T2 =10 to 60°, θ _B1 to θ _B2 =0 to 10°). This is the absorption spectrum A(θ) of A-ASF in FIG. 69 (θ _T1 to θ _T2 =10 to 60°, θ _B1 to θ _B2 =0 to 10°). A-FIG. 69 shows the optical thickness of each layer of the LPF provided in the ASF.

This is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF in FIG. 69 and the appropriate heat treatment temperature (θ _T1 to θ _T2 = 10 to 60°, θ _B1 to θ _B2 = 0 to 10°. ). This is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF (heat treatment temperature: appropriate temperature) in FIG. 69 and the optical characteristics (θ _T1 to θ _T2 = 10 to 60°, θ _B1 to θ _B2 = 0~10°). This is a diagram showing the relationship between the error in the thickness of the GST layer of A-ASF in FIG. 69 and the reflectance R (5°) at an incident angle of 5° before heat treatment (θ _T1 ~θ _T2 =0 ~ 30 °, θ _B1 ~θ _B2 = 30~60°). This is a diagram showing the relationship between the reflectance R (5°) at an incident angle of 5° before heat treatment of A-ASF in FIG. 69 and the appropriate heat treatment temperature (θ _T1 ~θ _T2 =0~30°, θ _B1 ~θ _B2 = 30~60°).

An embodiment of the present invention will be described in detail below.
[1. Optical filter]
The optical filter according to the present invention has the following configuration.
(1) The optical filter is
a filter layer including alternately laminated films made of two or more types of transparent materials (A) having different refractive indexes;
and a light absorption layer formed on the incident surface side of the filter layer,
The filter layer is configured to transmit light (signal light) in a transmission range of the filter layer among the incident light,
The light absorption layer is made of a layer for absorbing light (noise light) that is in a blocking range of the filter layer, out of the incident light.
(2) The light absorption layer is made of a light absorption material that satisfies the following formula (1).
πk _A /n _A ≧1 ... (1)
however,
n _A is the refractive index of the light absorption layer when the light absorption layer is irradiated with the incident light;
k _A is an extinction coefficient of the light absorption layer when the light absorption layer is irradiated with the incident light.

[1.1. Incident light]
Of the incident light, all or part of the light in the transmission range of the filter layer is transmitted through the filter layer as signal light. On the other hand, all or a part of the light in the blocking zone of the filter layer among the incident light is absorbed by the light absorption layer as noise light.
In the present invention, incident light is not particularly limited. That is, the incident light may be monochromatic light or white light. When the optical filter according to the present invention is applied to an optical device for inputting monochromatic light, the S/N ratio can be improved without significantly reducing the optical characteristics (e.g., sensitivity, signal light intensity, directivity, etc.) of the optical device. be able to.
Here, "monochromatic light" means that the center wavelength is λ ₀ and the ratio of the half width at half maximum (Δλ/2) of the spectrum to the center wavelength λ ₀ (=Δλ/2λ ₀ ) is 0.001 or less. It refers to light. If Δλ/2λ ₀ is 0.001 or less, as shown in FIG. 26, which will be described later, the shift between the transmission region and the blocking region is slight.

In the following description, it is assumed that the incident light is monochromatic light, and the "wavelength of the incident light" refers to the center wavelength λ ₀ .
On the other hand, the transmission range/stopping range depends on the wavelength of incident light. When the incident light is non-monochromatic light, when determining the characteristics of the optical filter according to the present invention, it is sufficient to divide it into each wavelength component, that is, to consider it as a superposition of monochromatic light. That is, the transmittance of each monochromatic light component of the incident light is calculated, and the sum of the products multiplied by the ratio of the components is the transmittance of the entire incident light. Reflectance and absorptance are also determined in the same manner. Therefore, as the range of wavelength components widens, the transmittance in the pass band becomes lower and the absorbance in the stop band becomes lower, but the required pass band/stop band and the values of transmittance, reflectance, and absorbance in each region are In some cases, it can be used as an angle selection filter. It is also possible to configure a filter that transmits/absorbs incident light in a specific wavelength range depending on the angle, and transmits (or absorbs) incident light in other wavelengths regardless of the angle.

[1.2. Filter layer]
[1.2.1. Definition]
The filter layer includes alternately laminated films (so-called "one-dimensional photonic crystal") made of two or more types of transparent materials (A) having different refractive indexes. A filter layer including alternately laminated films usually functions as a wavelength selective filter for incident light. Furthermore, the filter layer that functions as a wavelength selection filter usually also functions as an angle selection filter.
Here, a "wavelength selective filter (WSF)" is a filter that has the function of transmitting light with wavelengths in a specific range (transmission range) and the function of reflecting light with wavelengths other than that (stopband). means.
An "Angle Selective Filter (ASF)" is an optical filter that has the function of transmitting light with an incident angle within a specific range (transmission range) and reflecting light with an incident angle other than that range (stopband). means.

Examples of the WSF include a bandpass filter (BPF), a shortpass filter (SPF), and a longpass filter (LPF). In the present invention, any of these can be used for the filter layer depending on the transmission range/inhibition range of the filter.

Here, the term "bandpass filter (BPF)" refers to an optical filter that has the function of transmitting light with wavelengths within the transmission range and the function of reflecting light with wavelengths shorter and longer than the transmission range. .
"Short pass filter (SPF)" refers to an optical filter that has the function of transmitting light with a wavelength shorter than a threshold value and the function of reflecting light with a longer wavelength than a threshold value.
A "long pass filter (LPF)" refers to an optical filter that has a function of transmitting light with a wavelength longer than a threshold value and a function of reflecting light with a wavelength shorter than the threshold value.

[1.2.2. Structure of filter layer]
The filter layer includes alternately laminated films of two or more types of transparent materials (A) having different refractive indexes. At this time, by controlling the material (refractive index) of each layer, the thickness of each layer, the lamination order of each layer, etc., the wavelength of the light to be transmitted can be changed.

The basic components of WSF are:
(a) High refractive index layer ( _H4) whose refractive index is n H4 and whose optical thickness n _H4 d _H4 (d _H4 is the actual thickness) is proportional to λ/4 (λ is the wavelength of the incident light) ,
(b) a high refractive index layer (H2) whose refractive index is n _H2 and whose optical thickness n _H2 d _H2 (d _H2 is the actual thickness) is proportional to λ/2;
(c) a low refractive index layer (L4) whose refractive index is n _L4 (<n _H4 , n _H2 ) and whose optical thickness n _L4 d _L4 (d _L4 is the actual thickness) is proportional to λ/4; as well as,
(d) A low refractive index layer (L2) whose refractive index is n _L2 (<n _H4 , n _H2 ) and whose optical thickness n _L2 d _L2 (d _L2 is the actual thickness) is proportional to λ/2.
It is. In the case of the incident angle θ, the condition for the optical thickness nd of each layer constituting the WSF is nd=(1−sin ² θ/n ² ) ^1/2 ×(λ/4 or λ/2).

The basic structure of the BPF is such that L2 or H2, which determines the resonator length, is sandwiched between Bragg reflectors in which H4 and L4 are alternately stacked. The wavelength λ used in determining the optical thickness of each layer is the center wavelength of the transmission region. The Bragg reflector basically consists of H4 and L4, but the thickness of at least the outermost layer and the layer in contact with the substrate is changed to an optimal value so that ripples in the transmission region are reduced. is preferred.

The basic structure of the SPF is a Bragg reflector in which H4 and L4 are alternately stacked. By controlling the average thickness and number of laminated layers of each layer constituting the Bragg reflector, a filter layer that functions as an SPF can be obtained. A wavelength longer than the threshold wavelength (for example, 1.1 times the threshold wavelength) is used as the wavelength λ when determining the optical thickness of each layer. The thickness of each layer does not necessarily have to be the same, but rather the thickness of the high refractive index layer and/or the low refractive index layer such that the threshold wavelength is the desired wavelength and the ripple in the transmission region is reduced. It is preferable to optimize.

The basic structure of the LPF is a Bragg reflector in which H4 and L4 are alternately stacked, similar to the SPF. By controlling the average thickness and number of laminated layers of each layer constituting the Bragg reflector, a filter layer that functions as an LPF can be obtained. A wavelength shorter than the threshold wavelength (for example, 0.86 times the threshold wavelength) is used as the wavelength λ when determining the optical thickness of each layer. The thickness of each layer does not necessarily have to be the same, but rather the thickness of the high refractive index layer and/or the low refractive index layer such that the threshold wavelength is the desired wavelength and the ripple in the transmission region is reduced. It is preferable to optimize.

[1.2.3. Filter layer material]
The material constituting the WSF is not particularly limited, and various materials can be used depending on the purpose. Further, each high refractive index layer included in the WSF may be made of the same material or may be made of different materials as long as it functions as a WSF. Similarly, each low refractive index layer included in the WSF may be made of the same material or may be made of different materials as long as they function as a WSF.
Examples of low refractive index materials include MgF ₂ (refractive index at a wavelength of 905 nm: approximately 1.36) and SiO ₂ (refractive index at a wavelength of 905 nm: approximately 1.47).
Examples of high refractive index materials include ZnS (refractive index at a wavelength of 905 nm: approximately 2.30) and TiO ₂ (refractive index at a wavelength of 905 nm: approximately 2.299).

[1.3. Light absorption layer]
[1.3.1. Definition]
A "light absorption layer" is a layer that has the function of transmitting incident light (signal light) in the transmission range of the filter layer and the function of absorbing incident light (noise light) in the rejection range of the filter layer. say. The light absorption layer is made of a light absorption material capable of absorbing noise light, and is formed on the incident surface side of the filter layer.

[1.3.2. Complex refractive index of light absorption layer]
In order to efficiently absorb noise light in the light absorption layer, the complex refractive index (n+ik) of the light absorption layer needs to satisfy at least the following formula (1).
πk _A /n _A ≧1 ... (1)
however,
n _A is the refractive index of the light absorption layer when the light absorption layer is irradiated with the incident light;
k _A is an extinction coefficient of the light absorption layer when the light absorption layer is irradiated with the incident light.

When the extinction coefficient k _A becomes excessively small compared to the refractive index n _A , a large proportion of noise light is transmitted as is without being absorbed by the light absorption layer. In order to sufficiently absorb noise light, πk _A /n _A needs to be 1 or more.

[1.3.3. Formation position of light absorption layer]
The light absorption layer is formed on the incident surface side of the filter layer. "Incidence surface side" refers to a position where signal light can be transmitted and noise light can be absorbed, and does not necessarily mean the outermost surface of the filter layer. That is, the light absorption layer may be formed on the outermost surface of the filter layer, or may be formed inside the filter layer as long as it can absorb noise light.

In order to efficiently absorb noise light, it is preferable that the light absorption layer satisfies the following formula (2) regarding its formation position.
0≦Σn _i d _i ≦1.5λ (2)
however,
d _i is the actual thickness of the i-th surface layer (i is an integer greater than or equal to zero) formed on the light absorption layer,
n _i is the refractive index of the i-th surface layer when the i-th surface layer is irradiated with the incident light;
λ is the wavelength of the incident light.

The light absorption layer may be formed on the outermost surface of the filter layer. That is, the actual thickness d _i of the i-th surface layer (or the number i of surface layers) may be zero. In this case, although the reflectance of noise light is low, the light absorption layer is easily damaged. In order to obtain high durability, the total actual thickness Σd _i of the i-th surface layer is preferably 10 nm or more, more preferably 50 nm or more.
On the other hand, when a surface layer is formed on the light absorption layer, the light absorption layer is less likely to be damaged and the thickness of the light absorption layer can be easily controlled. However, if the surface layer becomes too thick, noise light is likely to be reflected. Therefore, Σn _i d _i is preferably 1.5λ or less. Σn _i d _i is preferably 1.1λ or less, more preferably 0.6λ or less.

Note that the material constituting the surface layer may be at least transparent (extinction coefficient k=0). Since the surface layer is also a part of the filter layer, it is preferable to use the same high refractive index material and/or low refractive index material as the filter layer.
Further, the surface layer may be only a high refractive index layer, only a low refractive index layer, or an odd number of layers such as a high refractive index layer/low refractive index layer/high refractive index layer. However, in order to facilitate design, it is preferable that the surface layer is composed of an integral multiple of (high refractive index layer/low refractive index layer).
Further, the actual thickness d _i (or optical thickness n _i d _i ) of each i-th surface layer may be the same as H4 or L4 constituting the WSF, or the light intensity at the light absorption layer The actual thickness d _i of each i-th surface layer may be optimized so that d i is maximized.

[1.3.4. Actual thickness of light absorption layer]
The actual thickness (d _A ) of the light absorption layer affects the transmittance of signal light and the absorption rate of noise light. In order to achieve both high transmittance of signal light and high absorption of noise light, it is preferable that the light absorption layer satisfies the following formula (3) with respect to its actual thickness.
0.005≦πk _A d _A /(n _A・λ)≦0.2 (3)
However, _dA is the actual thickness of the light absorption layer, and λ is the wavelength of the incident light.

If _dA becomes relatively too thin, the absorption rate of noise light (stopping band) will decrease. Therefore, πk _A d _A /(n _A ·λ) is preferably 0.005 or more. πk _A d _A /n _A is preferably 0.007 or more.
On the other hand, if _dA becomes relatively too thick, the transmittance of the signal light (transmission range) decreases. Therefore, πk _A d _A /(n _A ·λ) is preferably 0.2 or less. πk _A d _A /(n _A ·λ) is preferably 0.1 or less, more preferably 0.015 or less.

[1.3.5. Light absorbing material]
The type of light-absorbing material is not particularly limited as long as it satisfies the above-mentioned conditions. Examples of light-absorbing materials that meet the above conditions include:
(a) one or more transition metals (A),
(b) a cermet containing a transition metal (B) and a transparent material (B);
(c) a material whose extinction coefficient can be irreversibly changed by heating (hereinafter also referred to as "phase change material");
and so on.

[A. Transition metal (A)]
When the light absorption layer is made of a transition metal (A), the light absorption layer preferably satisfies the following formulas (11) to (13).
1≦n _A1 ^*・・・(11)
1≦k _A1 ^*・・・(12)
｜n _A1 ^* -k _A1 ^* ｜≦4 ...(13)
however,
n _A1 ^* = 905n _A1 /λ, k _A1 ^* = 905k _A1 /λ, λ is the wavelength (nm) of the incident light,
n _A1 is the refractive index of the light absorption layer when the light absorption layer is irradiated with the incident light;
k _A1 is an extinction coefficient of the light absorption layer when the light absorption layer is irradiated with the incident light.

n _A1 ^* and k _A1 ^* represent values obtained by normalizing the refractive index n _A1 and extinction coefficient k _A1 of the light absorption layer by the wavelength λ of the incident light, respectively. Formulas (11) to (13) represent the range in which the properties are excellent when the light absorption layer is made of a transition metal (A). Among these, equation (12) can be interpreted as a condition that the attenuation of the light intensity within the light absorption layer is sufficiently large. Furthermore, equations (11) and (13) can be interpreted as conditions in which absorption is superior to reflection.

Examples of transition metals (A) that satisfy the above-mentioned conditions include Cr, Ti, W, Ni, and Co. The light absorption layer may contain any one of these transition metals (A), or may contain two or more of these transition metals (A).
Examples of the light absorption layer containing two or more types of transition metals (A) include an alloy consisting of a single phase, an alloy consisting of a mixed phase of two or more types, and the like.

[B. cermet]
When the light absorption layer is made of cermet, the light absorption layer preferably satisfies the following formulas (14) to (15).
2≦n _A2 ^* ≦3.5 ... (14)
(n _A2 ^* -1)/3≦k _A2 ^* ...(15)
however,
n _A2 ^* =905n _A2 /λ, k _A2 ^* =905k _A2 /λ, λ is the wavelength (nm) of the incident light,
n _A2 is the refractive index of the light absorption layer when the light absorption layer is irradiated with the incident light;
k _A2 is an extinction coefficient of the light absorption layer when the light absorption layer is irradiated with the incident light.

n _A2 ^* and k _A2 ^* represent values obtained by normalizing the refractive index n _A2 and extinction coefficient k _A2 of the light absorption layer by the wavelength λ of the incident light, respectively. Formulas (14) to (15) represent the range in which the properties are excellent when the light absorption layer is made of cermet. Among these, n _A2 ^* ≦3.5 in equations (15) and (14) can be interpreted as a condition that the attenuation of the light intensity within the light absorption layer is sufficiently large. Furthermore, 2≦n _A2 ^* in equation (14) can be interpreted as a condition in which absorption is superior to reflection.

Examples of the transition metal (B) constituting the cermet that satisfies the above-mentioned conditions include Cr, Ti, W, Ni, and Co. The light absorption layer may contain any one of these transition metals (B), or may contain two or more of these transition metals (B).
Furthermore, examples of the transparent material (B) constituting the cermet that satisfies the above-mentioned conditions include SiO ₂ , Al ₂ O ₃ , ZrO ₂ , AlN, and the like. The light absorption layer may contain any one of these transparent materials (B), or may contain two or more of these transparent materials (B).

The amount of transition metal (B) contained in the cermet is not particularly limited, and an optimal amount can be selected depending on the purpose. Usually, the amount of transition metal (B) is 10 to 90 vol%.
Moreover, the particle size of the smaller of the transition metal (B) and the transparent material (B) contained in the cermet is preferably 1/10 or less of the wavelength of the signal light. If the smaller particle size becomes larger than this, the effect of light scattering will increase, and the function of the interference filter will no longer be expressed.

[C. Phase change material]
The light absorption layer may include a phase change material. When a phase change material is used as the light absorption material, even if an error occurs in the thickness of the light absorption layer during film formation, the thickness error can be compensated for by heat treatment.

Examples of phase change materials include Ge ₂ Sb ₂ Te ₅ , GeSb ₂ Te ₄ , GeSb ₄ Te ₇ , and AgInSbTe-based optical recording materials. Examples of AgInSbTe-based optical recording materials include Ag _3.4 In _3.7 Sb _76.4 Te ₁₆ , Ag _3.5 In _3.8 Sb _75.0 Te _17.7 and Ag ₅ In ₅ Sb ₇₀ Te ₂₀ . The light absorption layer may be made of any one of these materials, or may contain two or more of these materials. All of these materials are amorphous with a small extinction coefficient immediately after film formation, but are crystallized by heat treatment and the extinction coefficient increases. Therefore, these are suitable as materials constituting the light absorption layer.

[1.4. substrate]
Filter layers and light absorbing layers are typically very thin and therefore formed on a substrate. The substrate is not particularly limited as long as it can support the filter layer and the light absorption layer and can transmit signal light.
Examples of the substrate include a silica glass plate, a borosilicate glass plate, and a sapphire plate.

[2. Manufacturing method of optical filter]
[2.1. Optical filter whose light absorption layer is made of transition metal (A) or cermet]
An optical filter whose light absorption layer is made of a transition metal (A) or a cermet can be obtained by forming thin films having a predetermined composition on a substrate in a predetermined order. The method for forming the thin film is not particularly limited, and any known method can be used.

[2.2. Optical filter whose light absorption layer is made of a phase change material]
An optical filter whose light absorption layer is made of a phase change material is
A laminate in which a light absorption layer is formed on the incident surface side of the filter layer is produced,
It can be manufactured by adjusting the filter characteristics of the laminate by heat treating the laminate.

[2.2.1. Laminate production process]
First, a laminate in which a light absorption layer is formed on the incident surface side of a filter layer is produced (laminate production step). The filter layer includes alternating layers of two or more types of transparent materials (A) having different refractive indexes. Further, the light absorption layer includes a light absorption material whose extinction coefficient can be irreversibly changed by heating. The method of forming each layer is not particularly limited, and any known method can be used.

[2.2.2. Heat treatment process]
Next, the filter characteristics of the laminate are adjusted by heat-treating the laminate (heat treatment step). Thereby, an optical filter according to the present invention is obtained.
When a phase change material is used as the light absorption layer, even if an error occurs in the thickness of the light absorption layer during film formation, the thickness error can be compensated for by heat treatment. Therefore, optical filters having desired characteristics can be manufactured with high yield.
The heat treatment method is not particularly limited as long as desired optical properties can be obtained. Specifically, the heat treatment method includes the following methods.

[A. First method]
The first method is to perform heat treatment while monitoring the transmission/absorption characteristics of the laminate. If the transmission/absorption characteristics deviate from the desired characteristics, check the extinction coefficient while monitoring the transmission/absorption characteristics (e.g., transmittance, reflectance, and absorption of light with a specific angle of incidence). The heat treatment is performed under gently changing conditions, and when the desired transmission/absorption properties are achieved, the heat treatment is terminated by rapid cooling.

[B. Second method]
The second method is to measure the thickness of the light absorption layer and perform heat treatment under conditions depending on the thickness of the light absorption layer. The transmission/absorption properties of the light absorption layer also depend on the thickness of the light absorption layer. Therefore, once the thickness of the light absorption layer is known, it is possible to uniquely estimate the heat treatment conditions necessary to obtain desired transmission/absorption characteristics.
In addition, if it is difficult to non-destructively measure the thickness of the light absorption layer, it is sufficient to sample a predetermined number of optical filters from the same lot and actually measure the thickness of the light absorption layer.

[C. Third method]
A third method is to measure the optical properties of the laminate and perform heat treatment under conditions depending on the optical properties. Once the structure of the optical filter is determined, its optical characteristics (for example, the intensity of signal light, the intensity of noise light, etc.) can be estimated by numerical calculation. Therefore, if the optical characteristics of an optical filter are actually measured and the deviation from the theoretical value is known, it is possible to uniquely estimate the heat treatment conditions necessary to obtain the desired optical characteristics.

[3. Effect]
An alternately laminated film made of two or more types of transparent materials (A) having different refractive indexes has a function as an angle selection filter. When light is incident on a filter layer including such alternately laminated films, the light is repeatedly reflected at the interface between different materials. As a result, the light intensity (or the electric field amplitude of the incident light) within the filter layer is not uniform and varies periodically along the depth direction. Furthermore, the distribution of light intensity also changes depending on the angle of incidence of light.

When a light absorption layer that satisfies predetermined conditions is formed on the entrance surface side of such a filter layer, when the incident light (signal light) in the transmission region is irradiated, the light absorption layer near the entrance surface (that is, the light absorption layer (near the installed position) the light intensity becomes weaker. As a result, a large proportion of the signal light passes through the light absorption layer without being absorbed by the light absorption layer.
On the other hand, when incident light (noise light) in the stop band is irradiated, the light intensity near the incident surface increases. As a result, the proportion of noise light absorbed by the light absorption layer increases.
Therefore, when such an optical filter is applied to a directional photodetector or the like, the influence of noise light can be reduced without significantly interfering with the transmission of signal light.

Furthermore, when producing such an optical filter, if an error occurs in the thickness of the light absorption layer, desired optical characteristics may not be obtained. On the other hand, if a material whose extinction coefficient can be irreversibly changed by heating is used as the light-absorbing material constituting the light-absorbing layer, an error may occur in the thickness of the light-absorbing layer during film formation. Even if it is, the thickness error can be compensated for by heat treatment. Therefore, it becomes possible to manufacture optical filters having desired characteristics at a high yield.

(Examples 1 to 4, Comparative Example 1: Design of optical filter)
[1. Introduction]
In order to improve the signal-to-noise ratio of a directional photodetector, an angle selective filter (ASF) that transmits only light incident from a specific direction is effective. A wavelength selection filter (WSF) made of a dielectric multilayer film selectively transmits only a specific wavelength from white light or light containing multiple wavelengths, and reflects the others, and usually transmits light in a nearly perpendicular direction. It is used with the angle of incidence. On the other hand, from the perspective of phase difference, oblique incidence is equivalent to a long wavelength (Reference 1), so the WSF also functions as an ASF for a predetermined wavelength (References 2 to 4).

However, when a dielectric multilayer film is used, the light reflected by the ASF may be reflected by the casing or optical window of the device equipped with the detector, becoming stray light, reaching the detector and causing noise. . In particular, as devices become smaller, they become more susceptible to stray light. Therefore, if an absorption-type angle-selective filter (A-ASF) that absorbs light at an incident angle outside the transmission range instead of reflecting it is realized, the restrictions on implementation will be reduced.
Therefore, we designed the A-ASF, which is based on the conventional reflection-type angle-selective filter (R-ASF) and adds a light absorption function to it.

[2. A-ASF design]
[2.1. A-ASF expression mechanism]
[2.1.1. overview]
Considering the case where the signal light is incident on the photodetector at an incident angle of 6 to 12 degrees, and the noise light generated by multiple reflection by the housing and other optical elements is incident on the photodetector at an incident angle of 23 to 26 degrees, the transmission range is 6 degrees. We design an A-ASF with ~12° and a stopband of 23-26°. The characteristics of other incident angles are not considered. The wavelength of the incident light is 905 nm.

The signal light intensity (I _signal ) and the noise light intensity (I _noise ) are defined as shown in the following equations (21) and (22), respectively. Here, T(θ) and R(θ) are the transmittance and reflectance depending on the incident angle θ, respectively, and w is the ratio of light reflected by the ASF reaching the detector.

First, consider the case where w=0, that is, the reflected light has no adverse effect at all. In this case, a _{2 -} cavity bandpass filter (2C -BPF), the R-ASF characteristics can be obtained (Reference 5).
The basic structure of a BPF is that a Bragg reflector consisting of a stacked high refractive index layer/low refractive index layer with an optical thickness of 1/4 wavelength is sandwiched between a layer with an optical thickness of 1/2 wavelength that determines the cavity length. be. When two 1/2 wavelength layers are provided, the transmission range (wavelength or incident angle range) is expanded depending on the thickness of the black reflecting mirror between them.

Here, based on the 28-layer 2C-BPF,
(a) Each optical thickness of the high refractive index layer/low refractive index layer constituting the Bragg reflector,
(b) resonator length (optical thickness of 1/2 wavelength layer), and
(c) Treat a total of five optical thicknesses of the outermost surface layer and the layer in contact with the substrate (silica glass) as independent parameters, and find a value that reduces I _noise under the condition of I _signal ≧ 0.95. Ta. Adjustment of the thickness of the outermost layer and the layer in contact with the substrate is necessary to reduce ripples in the transmission region.

[2.1.2. Comparative example 1]
The structure of an R-ASF (Comparative Example 1) consisting of a 28-layer TiO ₂ /SiO ₂ multilayer film that does not include a light absorption layer is shown below. The optical thickness of each layer is shown in quarter wavelengths.
0.352SiO ₂ /(1.003TiO ₂ /1.038SiO ₂ /) ³
1.003TiO ₂ /1.96SiO ₂ /(1.003TiO ₂ /1.038SiO ₂ /) ⁶
1.003TiO ₂ /1.96SiO ₂ /(1.003TiO ₂ /1.038SiO ₂ /) ²
_0.983TiO2 /Substrate
FIG. 1 shows the transmittance T(θ) and reflectance R(θ) of the R-ASF (Comparative Example 1) having the above configuration. At this time, I _signal =0.95, I _noise =0.00475+w.

In the transmission range of 9°≦θ≦12°, a high value of T(θ) is maintained. When θ increases beyond 12°, T(θ) rapidly decreases and R(θ) increases. As a result, I _signal =0.95, I _noise =0.00475, and S/N=I _signal /I _noise =200 (however, w=0). However, in this case, since the light absorption is zero, when w is finite, I _noise =0.00475+w, and S/N≈1/w after all. Therefore, w is replaced with its maximum value, 1, and I _noise is defined again as shown in the following equation (23). According to this definition, the maximum value of S/N by R-ASF is 1.

Consider a configuration in which the S/N ratio is increased by combining this with a light absorption function. As shown later, T(θ)≒0 in the stopband can be easily obtained (see Figure 5, etc.), so T(θ) in the passband is made high and R(θ) in the stopband is made low. That is the issue. FIG. 2 shows the light intensity distribution when θ=9°, which is the center of the transmission region. As a result of multiple interference, there is a maximum within each SiO ₂ layer and a minimum within each TiO ₂ layer. The maximum value is small at the entrance surface of the filter (depth = 0), the interface with the substrate, and roughly in the middle in the thickness direction, and becomes large between them.

In contrast, in the case of θ=24.5°, which is the center of the stopband, as shown in FIG. 3, the maximum value of the light intensity is large near the incident surface, and then it attenuates rapidly from there. Therefore, if a layer of light-absorbing material is formed at an appropriate position near the entrance surface of the filter, absorption will be large in the stop band, resulting in a low R(θ), whereas in the transmission band, the effect of the light-absorbing layer will be It is predicted that the value of T(θ) will be small and a high value of T(θ) will be maintained.

[2.2. Selection of light-absorbing material]
As can be seen from FIGS. 2 and 3, the light intensity varies greatly within the range of the thickness of each layer λ/(4n)≈100 to 150 nm (n is the refractive index). Therefore, a light absorption layer that is sufficiently thinner than this must be formed at a location where the light intensity is high, and the incident light in the stop zone must be absorbed while passing through this layer. Therefore, the complex refractive index n+ik of the light-absorbing material is required to satisfy the condition πk/n≧1. n+ik of dielectric materials with absorption cannot satisfy this in many cases. For example, since n+ik=4.5+i0.18 for Ge and n+ik=4.7+i0.7 for PbS (Reference 6), excellent A-ASF characteristics cannot be obtained. Therefore, the design uses metal materials and cermets as light-absorbing materials.

Most metal materials satisfy the above condition of πk/n≧1. However, in the case of noble metals such as Au, Ag, and Cu, the dielectric loss (imaginary part of the dielectric constant) is small (Reference 7), so there is a possibility that the light is not absorbed but rather reflected. On the other hand, transition metals such as Cr, Ti, Mo, W, Ni, and Co have large dielectric loss (i.e., low reflectance) (References 8, 9), so they are expected to function as light-absorbing materials. Ru.

Another candidate material is cermet, in which fine metal particles are dispersed in a dielectric material (Reference 10). Due to the structure in which transition metals such as Cr, Ti, Mo, W, Ni, and Co are dispersed in SiO ₂ and Al ₂ O ₃ , the condition of πk/n≧1 is satisfied (References 11 to 14). ).

[3. Results and discussion]
Assuming a structure in which the light absorption layer described in Section 2.2 is formed on the incident surface of a 2C-BPF consisting of a 28-layer TiO ₂ /SiO ₂ multilayer film, find the appropriate value for the optical thickness of each layer. Ta.

[3.1. When using metal materials]
[3.1.1. Complex refractive index]
First, consider the case where a metal layer is formed on the incident surface. Appropriate values for the thickness of each layer were determined for each value of the complex refractive index n+ik=1+i to 5+i6. However, the optical thickness of the high refractive index layer/low refractive index layer constituting the Bragg reflector was the same, and the resonator length was limited to twice this. This value, the thickness of the metal material layer, and the optical thickness of each layer in contact with the metal material layer and the substrate are four independent parameters.

FIG. 4A shows the results of determining the thickness of each layer such that the I _signal becomes large under the condition of I _noise ≦0.2. Even if πk/n≧1, if n≪k (noble metal), the penetration length of light into the metal layer is small, but the light is reflected rather than absorbed, so it is not suitable as an A-ASF. You won't get great features. Conversely, a region where k<<n is inappropriate because the condition πk/n≧1 is not satisfied. Except for these extreme cases, I _signal has a nearly constant value that is almost independent of n and k, but rather fluctuates because the thickness of each layer is not completely optimized. However, when n≈k, there is a tendency for a large I _signal to be obtained.

FIG. 4B shows the results of determining the thickness of each layer such that I _noise is small under the condition of I _signal ≧0.75. In this case as well, although the change in I _noise is small except in the regions where n<<k and k<<n, it is more appropriate to approximately approximate n≒k.
Note that even in Fig. 4(B), the values fluctuate because the thickness of each layer constituting the A-ASF has not been completely optimized.

Metals that satisfy the condition n≒k include Cr (3.3155+i3.5271, reference 9), Ti (3.2985+i3.9614, reference 9), Mo (3.0074+i3.4853, reference 10), W (3.2940+i2.9973, Reference 10), etc. Further, Ni and Co also have relatively close values of n and k. Therefore, using Cr as the metal material, appropriate values for each optical thickness of the high refractive index layer/low refractive index layer constituting the Bragg reflector and each resonator length were determined independently.

[3.1.2. Example 1]
The structure of an A-ASF (Example 1) including a Cr layer (2.4 nm) and an R-ASF consisting of a 28-layer TiO ₂ /SiO ₂ multilayer film is shown below. The optical thickness of each layer other than the Cr layer is shown in units of 1/4 wavelength.
Cr(2.4nm) _/0.952SiO2/ ( _1.060TiO2 / _0.952SiO2 /) ³
1.060TiO ₂ /1.962SiO ₂ /(1.060TiO ₂ /0.952SiO ₂ /) ⁶
1.060TiO ₂ /1.962SiO ₂ /(1.060TiO ₂ /0.952SiO ₂ /) ²
_1.060TiO2 /Substrate
5 to 7 show the results of A-ASF (Example 1) having the above configuration under the condition of I _noise ≦0.2. At this time, I _noise =0.2, I _signal =0.701, and S/N = 3.51.

As seen in FIG. 5, the tendency of change in T(θ) is similar to that in the case without optical absorption (FIG. 1), but the value in the transmission region becomes lower. On the other hand, the increase in R(θ) in the stopband is suppressed, and the absorption rate A(θ) increases instead. As a result, I _signal =0.701 and S/N=3.51, that is, an S/N that is 3.51 times higher than in the case of R-ASF without optical absorption is obtained.
The light intensity distribution when θ=9° shown in FIGS. 6(A) and 6(B) is similar to that in the case of R-ASF with no light absorption (FIGS. 2(A) and 2(B)). They are very similar, and the light intensity at the Cr layer is small. On the other hand, when θ=24.5°, unlike FIGS. 3(A) and 3(B), by increasing the thickness of the second layer (SiO ₂ layer), the light intensity reaches its maximum just at the Cr layer. Become. As a result, an A-ASF function is obtained in which light absorption is small at θ=9° and large at θ=24.5°.

[3.1.3. Example 2]
The structure of an A-ASF (Example 2) including a Cr layer (1.7 nm) and an R-ASF consisting of a 28-layer TiO ₂ /SiO ₂ multilayer film is shown below. The optical thickness of each layer other than the Cr layer is shown in units of 1/4 wavelength.
Cr(1.7nm) _/1.022SiO2/ ( _0.981TiO2 / _1.022SiO2 /) ³
0.981TiO ₂ /2.032SiO ₂ /(0.981TiO ₂ /1.022SiO ₂ /) ⁶
0.981TiO ₂ /2.032SiO ₂ /(0.981TiO ₂ /1.022SiO ₂ /) ²
0.932TiO ₂ /Substrate
8 to 10 show the results of A-ASF with the above configuration under the condition of I _signal ≧0.75. At this time, I _noise =0.325, I _signal =0.75, and S/N = 2.31.

Example 2 has similar trends in T(θ), R(θ), A(θ), and light intensity distribution as compared to Example 1. However, in Example 2, since the Cr layer is thin, T(θ) and R(θ) are high and A(θ) is low. Therefore, although I _signal is large, it remains at a large value of I _noise =0.325, so S/N = 2.31. Compared to the first embodiment, I _signal =0.701, I _noise =0.2, and S/N=3.51, the S/N of the second embodiment is inferior. If only the optical S/N is a problem, it is advantageous to make the Cr layer thicker to reduce both I _signal and I _noise . However, when the I _signal is large, there is an advantage that the influence of electrical noise of the photodetector and its signal processing becomes relatively small. In this way, the optimal values of I _signal and I _noise depend on the entire system, so here we will only show the design guidelines for A-ASF.

[3.2. When using cermet material]
[3.2.1. Complex refractive index]
Next, consider a case where a cermet layer is formed on the incident surface. Appropriate values for the thickness of each layer were determined for each value of the complex refractive index n+ik=2+i0.1 to 3.5+i1.5. However, as in Section 3.1, this is limited to the case where twice the optical thickness of the high refractive index layer/low refractive index layer constituting the Bragg reflector is equal to the resonator length.

FIG. 11A shows the results of determining the thickness of each layer such that the I _signal becomes large under the condition of I _noise ≦0.2. When k≦0.2, light absorption is small, so I _noise ≦0.2 cannot be obtained. On the other hand, in the range of 0.2<k<0.8, a larger I _signal can be obtained when n is smaller while satisfying I _noise ≦0.2. When k≧0.8, I _signal ≒0.68 to 0.69, and the dependence on n is small, and rather the value fluctuations are noticeable because the thickness of each layer is not completely optimized. . However, there is a slight tendency that the smaller n is, the larger the I _signal is. Therefore, within the calculated range, it is determined that a material with a small n and a large k is appropriate.

FIG. 11B shows the results of determining the thickness of each layer such that I _noise becomes small under the condition of I _signal ≧0.75. In this case as well, the smaller n is and the larger k is, the smaller I _noise can be obtained.
Note that even in FIG. 11(B), the values fluctuate because the thickness of each layer constituting the A-ASF has not been completely optimized.

The values of n+ik have been reported so far for SiO ₂ +Ti (Reference 14), SiO ₂ +Mo (Reference 12), Al ₂ O ₃ +Co (Reference 11), and Al ₂ O ₃ +W (Reference 13). Among them, relatively small n and large k can be obtained with SiO ₂ +Ti, and when the composition is Ti _0.74 Si _0.26 O _1.91 , n+ik=2.6+i1.4. Using this material, appropriate values for each optical thickness and cavity length of the high refractive index layer/low refractive index layer constituting the Bragg reflector were determined independently.

[3.2.2. Examples 3-4]
The following shows the structure of an A-ASF (Example 3) comprising an R-ASF consisting of a Ti _0.74 Si _0.26 O _1.91 layer (optical thickness 0.089) and a 28-layer TiO ₂ /SiO ₂ multilayer film. . The optical thickness of each layer is shown in quarter wavelengths.
0.089Ti _0.74 Si _0.26 O _1.91 /
0.833SiO ₂ /(1.125TiO ₂ /0.916SiO ₂ /) ³
1.125TiO ₂ /1.874SiO ₂ /(1.125TiO ₂ /0.916SiO ₂ /) ⁶
1.125TiO ₂ /1.874SiO ₂ /(1.125TiO ₂ /0.916SiO ₂ /) ²
_1.103TiO2 /Substrate
12 to 14 show the results of A-ASF with the above configuration under the condition of I _noise ≦0.2. At this time, I _noise =0.2, I _signal =0.704, and S/N = 3.52.

The following shows the structure of an A-ASF (Example 4) comprising an R-ASF consisting of a Ti _0.74 Si _0.26 O _1.91 layer (optical thickness 0.067) and a 28-layer TiO ₂ /SiO ₂ multilayer film. . The optical thickness of each layer is shown in quarter wavelengths.
0.067Ti _0.74 Si _0.26 O _1.91 /
0.895SiO ₂ /(1.000TiO ₂ /1.053SiO ₂ /) ³
1.000TiO ₂ /1.958SiO ₂ /(1.000TiO ₂ /1.053SiO ₂ /) ⁶
1.000TiO ₂ /1.958SiO ₂ /(1.000TiO ₂ /1.053SiO ₂ /) ²
1.010TiO ₂ /Substrate
15 to 17 show the results of A-ASF with the above configuration under the condition of I _signal ≧0.75. At this time, I _noise =0.299, I _signal =0.75, and S/N = 2.51.

Under the condition of I _noise ≦0.2, I _signal =0.704 and S/N = 3.52 are obtained. Further, under the condition of I _signal ≧0.75, I _noise =0.299 and S/N = 2.51 are obtained. The characteristics shown in FIGS. 12 to 17 are almost the same as those of the metal materials shown in FIGS. 5 to 10, and the values of I _signal and I _noise obtained as a result are not significantly different.

Since n and k of the cermet material vary depending not only on its composition but also on the size of the dispersed metal particles, the process is more stable than the formation of a pure metal layer as described in the previous section. Desired. On the other hand, since it is thicker than a pure metal layer (an optical thickness of 0.089 is equivalent to 7.7 nm), the thickness may be easier to control.

[4. summary]
By forming a layer made of a light-absorbing material such as Cr or Ti _0.74 Si _0.26 O _1.91 cermet on the light incident surface of a dual-resonator bandpass filter made of a TiO ₂ /SiO ₂ multilayer film, the amount of light incident on the filter can be reduced. We have realized an absorption-type angle-selective filter that transmits light with an incident angle of 6 to 12 degrees (transmission range) and absorbs light with an incidence angle of 23 to 26 degrees (stopping range). It was found that when this was installed in front of the photodetector, the S/N was improved by 2.3 to 3.5 times compared to the case of a conventional reflective angle selection filter.

Here, a case where the wavelength of incident light is 905 nm is treated. In the case of other wavelengths λ, if the refractive index n ^* =905n/λ normalized by λ and the extinction coefficient k ^* =905k/λ are within the appropriate ranges shown in FIG. 4 or FIG. Similar performance can be obtained. In addition, in the case of other different transmission bands and stop bands, the function of the absorption angle selection filter can be enhanced by designing based on an appropriate wavelength selection filter (short pass filter, long pass filter, etc.). It can be realized.

[Reference 1] HA Macleod, Thin-Film Optical Filters, 4th ed. (2010) CRC Press, Boca Raton.
[Reference 2] Y. Takeda, H. Iizuka, S. Mizuno, K. Hasegawa, T. Ichikawa, H. Ito, T. Kajino, and T. Motohiro, J. Appl. Phys. Vol. 116, No. 1(2014)014501
[Reference 3] Y. Takeda and T. Ito, Jpn. J. Appl. Phys. Vol. 54, No. 8S1(2015)08KD13
[Reference 4] Y. Takeda, H. Iizuka, N. Yamada, and T. Ito, Appl. Opt. Vol. 56, No. 20(2017)pp. 5761-5767
[Reference 5] Y. Takeda, T. Ito, N. Yamada, K. Hasegawa, S. Mizuno, T. Ichikawa, HN Luitel, H. Iizuka, K. Higuchi, H. Ito, A. Ichiki, and T Motohiro, Jpn. J. Appl. Phys. Vol. 57, No. 8S3(2018)08RF05

[Reference 6] ED Palik, ed., Handbook of optical constants of solids (Academic Press, London, 1985)
[Reference 7] S. Barbar and JH Weaver, Appl. Opt. Vol. 54, No. 3(2015)pp. 477-481
[Reference 8] PB Johnson and RW Christy, Phys. Rev. B Vol. 9, No. 12(1974)pp. 5056-5070
[Reference 9] MA Ordal, RJ Bell, RW Alexander, LA Newquist, and MR Querry, Appl. Opt. Vol. 27, No. 6(1988)pp. 1203-1209
[Reference 10] F. Cao, K. McEnaney, G. Chen, and Z. Ren, Energy Environ. Sci. Vol. 7, No. 5(2014)1615-1627

[Reference 11] GA Niklasson and CG Granqvist, Solar Energy Mater. Vol. 7, No. 4(1983)pp. 501-510
[Reference 12] L. Zheng, F. Gao, S. Zhao, F. Zhou, JP Nshimiyimana, and X. Diao, Appl. Surf. Sci. Vol. 280(2013)pp. 240-246
[Reference 13] L. Rebouta, A. Sousa, P. Capela, M. Andritschky, P. Santilli, A. Matilainen, K. Pischow, NP Barradas, and E. Alves, Solar Energy Mater. Solar Cells Vol. 137 (2015)pp. 93-100
[Reference 14] E.-T. Hu, X.-X. Liu, Q.-Y. Cai, Y. Yao, K.-Y. Zang, K.-H. Yu, W. Wei, X. -X. Xu, Y.-X. Zheng, S.-Y. Wang, R.-J. Zhang, and L.-Y. Chen, Opt. Mater. Express Vol. 7, No. 7(2017)pp .2387-2395

(Examples 5 to 8: Position where light absorption layer is formed)
[1. overview]
As in Examples 1 to 4, consider the case where the transmission range is 6 to 12° and the rejection range is 23 to 26°. Even if the light absorption layer is not located at the incident plane, the A-ASF function can be obtained as long as it is located close to the incident plane.

[2. result]
The structure of an A-ASF (Example 5) including a Cr layer (5.8 nm) and an R-ASF consisting of a 28-layer TiO ₂ /SiO ₂ multilayer film is shown below. The Cr layer is formed as the third layer from the surface. The optical thickness of each layer other than the Cr layer is shown in units of 1/4 wavelength.
0.978SiO ₂ /1.075TiO ₂ /Cr(5.8nm)/1.052SiO ₂ /
(1.075TiO ₂ /1.052SiO ₂ /) ² 1.075TiO ₂ /1.792SiO ₂ /
(1.075TiO ₂ /1.052SiO ₂ /) ⁶ 1.075TiO ₂ /1.792SiO ₂ /
(1.075TiO ₂ /1.052SiO ₂ /) ² 1.075TiO ₂ /Substrate
FIG. 18A shows the results of determining the thickness of each layer such that the I _signal becomes large under the condition of I _noise ≦0.2 in the A-ASF having the above configuration. At this time, I _signal =0.704, I _noise =0.2, and S/N = 3.52.

The structure of an A-ASF (Example 6) including a Cr layer (14.9 nm) and an R-ASF consisting of a 28-layer TiO ₂ /SiO ₂ multilayer film is shown below. The Cr layer is formed as the fifth layer from the surface. The optical thickness of each layer other than the Cr layer is shown in units of 1/4 wavelength.
1.194SiO ₂ /1.000TiO ₂ /1.012SiO ₂ /1.000TiO ₂ /
Cr(14.9nm) _/1.012SiO2/
(1.000TiO ₂ /1.012SiO ₂ /) ¹ 1.000TiO ₂ /2.000SiO ₂ /
(1.000TiO ₂ /1.012SiO ₂ /) ⁶ 1.000TiO ₂ /2.000SiO ₂ /
(1.000TiO ₂ /1.012SiO ₂ /) ² 1.120TiO ₂ /Substrate
FIG. 18B shows the results of determining the thickness of each layer such that the I _signal becomes large under the condition of I _noise ≦0.2 in the A-ASF having the above configuration. At this time, I _signal =0.693, I _noise =0.2, and S/N = 3.46.

The configuration of an A-ASF (Example 7) including an R-ASF consisting of a Ti _0.74 Si _0.26 O _1.91 layer (optical thickness 0.22) and a 28-layer TiO ₂ /SiO ₂ multilayer film is shown below. . The light absorption layer is formed as the third layer from the surface. The optical thickness of each layer is shown in quarter wavelengths.
0.943SiO ₂ /0.987TiO ₂ /0.22Ti _0.74 Si _0.26 O _1.91 /0.972SiO ₂ /
(0.987TiO ₂ /0.972SiO ₂ /) ² 0.987TiO ₂ /2.086SiO ₂ /
(0.987TiO ₂ /0.972SiO ₂ /) ⁶ 0.987TiO ₂ /2.086SiO ₂ /
(0.987TiO ₂ /0.997SiO ₂ /) ² 0.997TiO ₂ /Substrate
FIG. 19(A) shows the results of determining the thickness of each layer such that the I _signal becomes large under the condition of I _noise ≦0.2 in the A-ASF having the above configuration. At this time, I _signal =0.699, I _noise =0.2, and S/N = 3.50.

The configuration of an A-ASF (Example 8) comprising an R-ASF consisting of a Ti _0.74 Si _0.26 O _1.91 layer (optical thickness 0.42) and a 28-layer TiO ₂ /SiO ₂ multilayer film is shown below. . The light absorption layer is formed as the fifth layer from the surface. The optical thickness of each layer is shown in quarter wavelengths.
1.068SiO ₂ /1.077TiO ₂ /0.921SiO ₂ /1.077TiO ₂ /
0.42Ti _0.74 Si _0.26 O _1.91 /0.921SiO ₂ /
(1.077TiO ₂ /0.921SiO ₂ /) ¹ 1.077TiO ₂ /1.964SiO ₂ /
(1.077TiO ₂ /0.921SiO ₂ /) ⁶ 1.077TiO ₂ /1.964SiO ₂ /
(1.077TiO ₂ /0.921SiO ₂ /) ² 1.282TiO ₂ /Substrate
FIG. 19(B) shows the results of determining the thickness of each layer such that the I _signal becomes large under the condition of I _noise ≦0.2 in the A-ASF having the above configuration. At this time, I _signal =0.603, I _noise =0.2, and S/N = 3.02.

Since the light-absorbing layer is thin, there is a concern that if it is placed on the outermost surface, it may be affected by scratches, but if it is formed inside, this problem will be resolved. Furthermore, since the thickness of the light absorption layer is thicker than that at the outermost surface, the thickness can be easily controlled. However, the functionality of A-ASF tends to be slightly inferior. Furthermore, the process is complicated because the TiO ₂ /SiO ₂ layer is formed again after the metal layer or cermet layer is formed.

(Examples 9 to 14: Configuration according to transmission region and rejection region)
[1. overview]
It is necessary to select an appropriate configuration of the TiO ₂ /SiO ₂ multilayer film depending on the transmission region and the rejection region. When the transmission band is θ _T1 to θ _T2 and the stop band is θ _B1 to θ _B2 , the signal light intensity (I _signal ) and noise light intensity (I _noise ) are rewritten as the following equations (24) and (25). Define.

Consider the case where θ _T1 =0°, θ _T2 =θ _B1 =10°, and θ _B2 =30°. Unlike the cases in Examples 1 to 4, this is a case in which light is not reflected by a specific surface and becomes noise light, but light reflected from many surfaces becomes background noise. In this case as well, an A-ASF can be constructed based on the 2C-BPF, but since the transmission region and the blocking region are in contact, it is necessary to make changes such as T(θ) at the boundary more steeply. A TiO ₂ /SiO ₂ multilayer film having a larger total number of layers than in Examples 1 to 4 is required (Reference Document 5).

[2. result]
[2.1. A-ASF based on 2C-BPF (Examples 9-10)]
The structure of an A-ASF (Example 9) including a Cr layer (2.9 nm) and an R-ASF consisting of a 32-layer TiO ₂ /SiO ₂ multilayer film (2C-BPF) is shown below. The optical thickness of each layer other than the Cr layer is shown in units of 1/4 wavelength.
Cr(2.9nm) _/0.914SiO2/ (0.980TiO2 _{/ 1.015SiO2} /) ³
0.980TiO ₂ /2.036SiO ₂ /(0.980TiO ₂ /1.015SiO ₂ /) ⁷
0.980TiO ₂ /2.036SiO ₂ /(0.980TiO ₂ /1.015SiO ₂ /) ²
_1.009TiO2 /Substrate
FIG. 20 shows the results of determining the thickness of each layer such that the I _signal becomes large under the condition of I _noise ≦0.2 in the A-ASF having the above configuration. At this time, I _signal =0.666, I _noise =0.2, and S/N = 3.33.

Below, an A-ASF (Example 10) comprising an R-ASF consisting of a Ti _0.74 Si _0.26 O _1.91 layer (optical thickness 0.109) and a 32-layer TiO ₂ /SiO ₂ multilayer film (2C-BPF) is shown. ) shows the configuration of The optical thickness of each layer other than the Cr layer is shown in units of 1/4 wavelength.
0.109Ti _0.74 Si _0.26 O _1.91 /0.893SiO ₂ /
(1.010TiO ₂ /0.884SiO ₂ /) ³ 1.010TiO ₂ /2.150SiO ₂ /
(1.010TiO ₂ /0.884SiO ₂ /) ⁶ 1.010TiO ₂ /2.150SiO ₂ /
(1.010TiO ₂ /0.884SiO ₂ /) ² 0.970TiO ₂ /Substrate
FIG. 21 shows the results of determining the thickness of each layer such that the I _signal becomes large under the condition of I _noise ≦0.2 in the A-ASF having the above configuration. At this time, I _signal =0.668, I _noise =0.2, and S/N = 3.35.

The reason why the I _signal is smaller and the S/N is lower than the results described in Examples 1 to 4 is that the transmission band and stop band are in contact with each other, so in order to reduce the I _noise , the I _signal must also be reduced. This is because it is inevitable. T(θ) and A(θ) are approximately constant until θ≦10°, but change within the range of θ=10 to 15°, and then become approximately constant again. If the number of TiO ₂ /SiO ₂ layers is increased, changes in T(θ) and the like at the boundary between the transmission region and the stop region become more steep, so that a larger I _signal and a higher S/N can be obtained. However, higher process stability is required during fabrication.
Further, as in Examples 1 to 4, when noise light is incident from a specific direction, assuming, for example, θ _B1 =15°, the configuration using the Cr layer shown in FIG. 20 allows I _noise =0. 161, excellent characteristics of S/N=4.13 are obtained.

[2.2. A-ASF based on SPF (Examples 11-12)]
When the transmission range is wider, better characteristics can be obtained by using a shortpass filter (SPF) instead of the 2C-BPF used so far. For the cases of θ _T1 =0°, θ _T2 =θ _B1 =30°, and θ _B2 =60°, the thickness of each layer was determined so that the I _signal would be large under the condition of I _noise ≦0.2.

A-ASF has
(a) A combination of a Cr layer (4 nm) and an SPF consisting of a 40-layer TiO ₂ /SiO ₂ multilayer film (Example 11), and
(b) A combination of Ti _0.74 Si _0.26 O _1.91 layer (optical thickness: 0.131×1/4 wavelength) and SPF consisting of 40 layers of TiO ₂ /SiO ₂ multilayer film (Example 12)
was used.

FIG. 22(A) shows the dependence of the transmittance T(θ), reflectance R(θ), and absorptance A(θ) of the A-ASF of Example 11 on the incident angle. FIG. 22(B) shows the optical thickness of each layer other than the Cr layer of the A-ASF shown in FIG. 22(A). At this time, I _signal =0.589, I _noise =0.2, and S/N = 2.95.
FIG. 23(A) shows the dependence of the transmittance T(θ), reflectance R(θ), and absorptance A(θ) of the A-ASF of Example 12 on the incident angle. FIG. 23(B) shows the optical thickness of each layer other than the Ti _0.74 Si _0.26 O _1.91 layer of A-ASF shown in FIG. 23(A). At this time, I _signal =0.594, I _noise =0.2, and S/N = 2.97.

Since the transmission range is wider than in the past, reducing polarization dependence and suppressing ripples are more important. Therefore, there is a large difference in thickness between the TiO ₂ layer and the SiO ₂ layer, and the thickness changes from the center to the ends of the multilayer film (Reference Document 4). As in the previous case, the transmission region/stopping region are in contact with each other, so the I _signal is smaller and the S/N is lower than the results described in Examples 1 to 4, but this is due to the TiO ₂ /SiO ₂ This can be improved by increasing the total number of layers. Further, when θ _B1 =35°, the configuration using the Cr layer shown in FIG. 22 provides excellent characteristics of I _noise =0.142 and S/N=4.14.

[2.3. A-ASF based on LPF (Examples 13-14)]
Contrary to what has been done so far, it is also possible to block components that are close to normal incidence and allow components with a large angle of incidence to pass through. For this purpose, a longpass filter (LPF) is used instead of the 2C-BPF and SPF. For the cases of θ _B1 =0°, θ _B2 =θ _T1 =10°, and θ _T2 =60°, the thickness of each layer was determined so that the I _signal would be large under the condition of I _noise ≦0.2.

A-ASF has
(a) A combination of a Cr layer (2.5 nm) and an LPF consisting of a 40-layer TiO ₂ /SiO ₂ multilayer film (Example 13), and
(b) A combination of a Ti _0.74 Si _0.26 O _1.91 layer (optical thickness: 0.097×1/4 wavelength) and an LPF consisting of a 40-layer TiO ₂ /SiO ₂ multilayer film (Example 14)
was used.

FIG. 24(A) shows the dependence of the transmittance T(θ), reflectance R(θ), and absorptance A(θ) of the A-ASF of Example 13 on the incident angle. FIG. 24(B) shows the optical thickness of each layer other than the Cr layer of the A-ASF shown in FIG. 24(A). At this time, I _signal =0.656, I _noise =0.2, and S/N = 3.28.
FIG. 25(A) shows the dependence of the transmittance T(θ), reflectance R(θ), and absorptance A(θ) of the A-ASF of Example 14 on the incident angle. FIG. 25(B) shows the optical thickness of each layer other than the Ti _0.74 Si _0.26 O _1.91 layer of A-ASF shown in FIG. 25(A). At this time, I _signal =0.656, I _noise =0.2, and S/N = 3.28.

As the transmission range is wide, suppression of ripples becomes more important, and the characteristics are improved by increasing the total number of TiO ₂ /SiO ₂ layers, similar to the case based on SPF shown earlier. However, the way the thickness of the TiO ₂ layer and the SiO ₂ layer change is different. Furthermore, if θ _T1 =20°, the configuration using Cr shown in FIG. 24 provides excellent characteristics of I _signal =0.724, I _noise =0.2, and S/N =3.62. .

(Example 15: Wavelength of incident light)
[1. overview]
We investigated the selection angle when light of different wavelengths were incident on an optical filter with a structure tuned for the wavelength of incident light of 905 nm. The optical filter was the same as in Example 11 (A-ASF based on SPF, using Cr as the light absorption layer).

[2. result]
26(A) to 26(C) show the transmittance T when monochromatic light with a wavelength of 905 nm, 905.9 nm, or 906.8 nm is incident on an A-ASF tuned for a wavelength of 905 nm. (θ), reflectance R(θ), and absorption rate A(θ) on incident angle dependence. From FIG. 26, it can be seen that even if the structure of the optical filter is the same, the selection angle changes as the wavelength of the incident light changes.

(Examples 21 to 26: Optical filter with phase change material)
[1. Introduction]
In order to improve the signal-to-noise ratio of a directional photodetector, an angle-selective filter (ASF) that transmits only light incident from a specific direction is effective. When the light incident on the detector is monochromatic, a wavelength-selective filter (WSF) made of a dielectric multilayer film can be used as the ASF. WSF selectively transmits only a specific wavelength from white light or light containing a plurality of wavelengths, and reflects the others, and is usually used with light incident at an angle close to perpendicular. On the other hand, from the perspective of phase difference, grazing incidence is equivalent to a long wavelength (Reference 1), so these filters function as an ASF for a predetermined wavelength (References 4, 5, 24).

However, when a dielectric multilayer film is used, the light reflected by the ASF is reflected by the casing and optical window of the device equipped with the detector and becomes stray light, which may reach the detector and cause noise. It can be. In particular, as devices become smaller, they become more susceptible to stray light.

Therefore, a layer of a light-absorbing material with a large extinction coefficient, such as a transition metal or cermet, is formed on or near the light-incidence surface of the ASF, so that the electric field amplitude of the incident light at the position of the light-absorbing layer becomes large in the transmission region. By adjusting the thickness of each layer constituting the ASF so that it is smaller in the stopband, an absorption angle selective filter is created in which light in the transmission range is transmitted, but most of the light in the stopband is absorbed. -type angle-selective filter, A-ASF) (Examples 1 to 15 described above). This reduces implementation constraints.

However, the thickness of the light absorption layer used in A-ASF is as thin as several nanometers, and furthermore, a slight deviation in this thickness or a slight variation in the optical properties of the material itself significantly affects the properties of A-ASF. Therefore, it is not easy to stably produce A-ASF with desired characteristics. Therefore, we propose a manufacturing method in which a phase change material such as Ge ₂ Sb ₂ Te ₅ (GST) or AgInSbTe is used for the light absorption layer, and a heat treatment is performed after forming a multilayer film to adjust its optical properties.

[2. Absorption angle selective filter (A-ASF) using phase change material]
Phase change materials such as GST are used in optical recording materials such as DVDs (References 25-27). Normally, after the film is formed, it is amorphous and has a small extinction coefficient, but when it is heated, it crystallizes and the extinction coefficient increases. When recording on a DVD, recording marks in a crystalline state are formed by laser irradiation and heating. Since the light reflectance of this portion is low, it can be distinguished from the amorphous (unrecorded) portion, which has a high reflectance.

This material is used for the light absorption layer of A-ASF. After film formation, since the extinction coefficient of GST is small, the transmittance of the A-ASF in the transmission region is high, but the reflectance in the stop region is also high. If heat treatment is then performed, GST will gradually crystallize and its extinction coefficient will increase, so that both the transmittance in the transmission region and the reflectance in the stop region will decrease. While monitoring the optical properties of A-ASF, the heat treatment is stopped when the desired properties are achieved.

In the case of a DVD, it is heated to a high temperature of 400 to 500° C. in order to cause crystallization (recording) within several tens of nanoseconds (References 26, 27). However, if the heat treatment is for a long time of about 20 minutes, crystallization occurs even at 160°C, and an increase in the extinction coefficient is already observed at 100°C (Reference Document 28). When forming a dielectric multilayer film by vapor deposition, the temperature of the sample usually rises naturally to about 200°C, so if the temperature is then heat-treated to 160°C, there will be little effect on the substrate or dielectric multilayer film. it is conceivable that.

[3. Design of absorption angle selection filter (A-ASF)]
We designed A-ASF using GST. The wavelength of the incident light was set to 905 nm. A TiO ₂ (refractive index: 2.299)/SiO ₂ (refractive index: 1.47) multilayer film and then a GST film were formed on a silica glass substrate. The optimal value of the thickness of each layer was determined so that desired optical characteristics could be obtained when this was subjected to heat treatment at 130° C. for 20 minutes. In this case, it is shown that even if the thickness of the GST layer deviates from the optimum value, desired optical characteristics can be obtained if heat treatment is performed at an appropriate temperature.

When a signal enters the photodetector at an angle of incidence of θ _T1 to θ _T2 , and noise light generated by being reflected by the device housing or other optical elements enters the photodetector at an angle of incidence of θ _B1 to θ _B2 , the signal is transmitted. Assume that an A-ASF with a range of θ _T1 to θ _T2 and a stop band of θ _B1 to θ _B2 is installed in front of a photodetector. The characteristics of incident angles outside these ranges are not considered.
As described above, the signal light intensity (I _signal ) and the noise light intensity (I _noise ) are defined as in the following equations (24) and (25), respectively.

Here, T(θ) and R(θ) are transmittance and reflectance depending on the incident angle θ, respectively. T(θ)+R(θ) in equation (25) is based on the assumption that all the light reflected by the A-ASF reaches the photodetector as stray light. In reality, not all of the reflected light reaches the photodetector, the proportion of which reaches the photodetector depends on the device. However, as will be shown later, since T(θ)≈0 in the stopband, the relative change in I _noise when changing the thickness of each layer or heat treatment conditions cannot be correctly evaluated using equation (25). can.

FIG. 27 shows the heat treatment temperature dependence of the complex refractive index (n+ik) of Ge ₂ Sb ₂ Te ₅ (GST). For the heat treatment temperature dependence of the complex refractive index (n+ik) of GST, the values shown in FIG. 27 estimated from Reference 28 are used. In order to obtain the A-ASF function, light must pass through a light absorption layer that is sufficiently thinner than the thickness of each layer of the dielectric multilayer film, λ/(4n) ≒ 100 to 150 nm (λ is the wavelength and n is the refractive index). In between, the incident light in the stopband must be absorbed. Therefore, it is required that the complex refractive index of the absorbing material satisfies πk/n≧1. The values in FIG. 27 satisfy this condition.

[4. Results and discussion]
It is necessary to select an appropriate structure of the TiO ₂ /SiO ₂ multilayer film depending on the transmission range and rejection range of A-ASF.

[4.1. Use of double resonator bandpass filter]
[4.1.1. When θ _T1 to θ _T2 = 6 to 12°, θ _B1 to θ _B2 = 23 to 26°]
First, consider the case where θ _T1 to θ _T2 = 6 to 12 degrees and θ _B1 to θ _B2 = 23 to 26 degrees. Thus, when the transmission band is narrow and close to 0° (normal incidence), it is appropriate to use a 2-cavity bandpass filter (2C-BPF). The basic structure of a BPF is a Bragg reflector made of laminated high and low refractive index layers with an optical thickness of λ/4, sandwiching a layer with an optical thickness of λ/2 that determines the cavity length (Reference Reference 1). When two λ/2 layers are provided, the transmission range (wavelength or incident angle range) is expanded depending on the thickness of the Bragg reflector between them (Reference Document 5).

Here, based on the 28-layer 2C-BPF,
(a) Each optical thickness of TiO ₂ /SiO ₂ constituting the Bragg reflector,
(b) resonator length (optical thickness of 1/2 wavelength layer),
(c) the optical thickness of the layer in contact with the light entrance surface layer and the substrate (silica glass), and (d) the thickness of the GST layer on the light entrance surface (birefringence is the value of heat treatment at 130 ° C.),
A total of six parameters were treated as independent parameters, and a value was determined that would increase I _signal under the condition of I _noise ≦0.2. Adjustment of the thickness of the light entrance surface layer and the layer contacting the substrate is necessary to reduce ripples in the transmission region.

[A. A-ASF with GST layer and 2C-BPF (Example 21)]
The structure of A-ASF (Example 21), which is a combination of a GST layer (X nm) and a 2C-BPF consisting of a 28-layer TiO ₂ /SiO ₂ multilayer film, is shown below. Here, the thickness of each layer of the TiO ₂ /SiO ₂ layer is indicated by the optical thickness in units of 1/4 wavelength.
Ge ₂ Sb ₂ Tb ₅ (Xnm)/0.725SiO ₂ /
(1.269TiO ₂ /0.725SiO ₂ /) ³ 1.269TiO ₂ /1.828SiO ₂ /
(1.269TiO ₂ /0.725SiO ₂ /) ⁶ 1.269TiO ₂ /1.828SiO ₂ /
(1.269TiO ₂ /0.725SiO ₂ /) ² 1.269TiO ₂ /substrate

When the transmission region and the rejection region are within the above ranges and the heat treatment temperature is 130° C., the optimum value of the thickness X of the GST layer is 3.1 nm. In the following, the cases where the thickness X of the GST layer is the optimum value, the optimum value -30%, and the optimum value +30% are examined, respectively.

[A. 1. Optical properties after heat treatment at 130°C]
28 to 30 respectively show the transmission spectra T( θ), reflection spectrum R(θ), and absorption spectrum A(θ). FIG. 31 shows the relationship between the error in the thickness of the GST layer of this A-ASF (heat treatment temperature: 130° C.) and the optical characteristics.

When the thickness of the GST layer is at the optimum value, I _signal =0.70 and I _noise =0.20 are obtained, as shown in FIG. However, when the thickness of the GST layer becomes thicker than the optimum value (3.1 nm) for 130°C heat treatment, R(θ) becomes smaller and I _noise decreases, but T(θ) also becomes smaller and I _signal also decreases. . Conversely, if the thickness of the GST layer becomes thinner than the optimum value, the I _signal will improve, but the I _noise will also increase.

[A. 2. Optical properties after heat treatment at appropriate temperature]
Figures 32 to 34 show cases in which A-ASF, which is a combination of GST layers and 2C-BPF with different thicknesses, is heat-treated at an appropriate temperature (see Figure 35) (Example 21-2) The transmission spectrum T(θ), the reflection spectrum R(θ), and the absorption spectrum A(θ) are shown. FIG. 35 shows the relationship between the error in the thickness of the GST layer of this A-ASF and the appropriate heat treatment temperature. FIG. 36 shows the relationship between the error in the thickness of the GST layer of A-ASF (heat treatment temperature: appropriate temperature) and the optical characteristics.

When the thickness of the GST layer varies and heat treatment is applied to each layer at an appropriate temperature, as can be seen from FIGS. 32 to 34, T(θ) and R(θ) and A(θ) are obtained. As a result, as shown in FIGS. 35 and 36, even if the thickness of the GST layer varies by ±30%, I _signal ≧0.68, I _noise = 0.20 (at the optimal thickness, I _signal = 0.70, I _noise =0.20). Since the heat treatment temperature is in the range of 100 to 160° C., it is thought that the effect on the substrate and the TiO ₂ /SiO ₂ multilayer film is small, as mentioned above.

[A. 3. Optical properties before heat treatment]
37 to 39 respectively show the transmission spectrum T(θ) and reflection spectrum R(θ ), and the absorption spectrum A(θ) is shown. FIG. 40 shows the relationship between the error in the thickness of the GST layer of this A-ASF and the reflectance R (30°) at an incident angle of 30° before heat treatment. FIG. 41 shows the relationship between the reflectance R (30°) of this A-ASF at an incident angle of 30° before heat treatment and the appropriate heat treatment temperature.

It can be seen from FIGS. 38 and 29 that R(θ) in the stopband (23° to 26°) and a region close thereto is sensitively influenced by the thickness of the GST layer. Therefore, it is appropriate to monitor this value and stop the heat treatment when the desired characteristics are achieved.
Alternatively, as shown in FIGS. 40 and 41, characteristics before heat treatment such as R (30°), thickness of the GST layer, and appropriate heat treatment temperature can be associated in advance. In this case, if the characteristics or the thickness of the GST layer before heat treatment are known, an appropriate heat treatment temperature can be uniquely determined.

[B. A-ASF with Cr layer and 2C-BPF (Example 22)]
For comparison, consider an example in which Cr (n+ik=3.32+i3.53 (Reference 8)) is used as the light-absorbing material.
The structure of an A-ASF (Example 22) in which a Cr layer (X nm) and a 2C-BPF consisting of a 28-layer TiO ₂ /SiO ₂ multilayer film is combined is shown below. Here, the thickness of each layer of the TiO ₂ /SiO ₂ layer is indicated by the optical thickness in units of 1/4 wavelength.
Cr(Xnm) _/0.952SiO2/ ( _1.060TiO2 /0.952SiO2 _/ ) ³
1.060TiO ₂ /1.962SiO ₂ /(1.060TiO ₂ /0.952SiO ₂ /) ⁶
1.060TiO ₂ /1.962SiO ₂ /(1.060TiO ₂ /0.952SiO ₂ /) ²
_1.060TiO2 /Substrate

When the transmission region and the rejection region are in the above ranges, the optimum value of the thickness X of the Cr layer is 2.4 nm. In the following, the cases where the thickness X of the Cr layer is the optimum value, the optimum value -30%, and the optimum value +30% are respectively investigated.

42 to 44 respectively show the transmission spectrum T(θ), reflection spectrum R(θ), and absorption spectrum of A-ASF (Example 22), which combines a Cr layer with different thickness and 2C-BPF. Spectrum A(θ) is shown. FIG. 45 shows the relationship between the error in the thickness of the Cr layer of this A-ASF and the optical characteristics.
Similar to Figures 28 to 31 (GST heat treatment temperature: 130°C), excellent characteristics are obtained when the thickness of the Cr layer is at the optimum value (2.4 nm), but when it deviates from this value, the characteristics deteriorate. .

[4.1.2. When θ _T1 to θ _T2 = 0 to 10°, θ _B1 to θ _B2 = 10 to 30°]
Next, consider the case where θ _T1 to θ _T2 =0 to 10° and θ _B1 to θ _B2 =10 to 30°. In this case as well, A-ASF can be configured using 2C-BPF. As in the previous case, the optimum thickness of each layer was determined so that the I _signal would be large under the condition of I _noise ≦0.20.

The structure of A-ASF (Example 23), which is a combination of a GST layer (X nm) and a 2C-BPF consisting of a 28-layer TiO ₂ /SiO ₂ multilayer film, is shown below. Here, the thickness of each layer of the TiO ₂ /SiO ₂ layer is indicated by the optical thickness in units of 1/4 wavelength.
Ge ₂ Sb ₂ Te ₅ (Xnm)/0.828SiO ₂ /
(1.013TiO ₂ /0.863SiO ₂ /) ³ 1.013TiO ₂ /2.174SiO ₂ /
(1.013TiO ₂ /0.863SiO ₂ /) ⁶ 1.013TiO ₂ /2.174SiO ₂ /
(1.013TiO ₂ /0.863SiO ₂ /) ² 0.912TiO ₂ /Substrate

When the transmission region and the rejection region are in the above ranges and the heat treatment temperature is 130° C., the optimum value of the thickness X of the GST layer is 3.7 nm. In the following, the cases where the thickness X of the GST layer is the optimum value, the optimum value -30%, and the optimum value +30% are examined, respectively.

Figures 46 to 48 show the results of A-ASF (Example 23), which is a combination of GST layers and 2C-BPF with different thicknesses, each subjected to heat treatment at an appropriate temperature (see Figure 49). The transmission spectrum T(θ), the reflection spectrum R(θ), and the absorption spectrum A(θ) are shown. FIG. 49 shows the relationship between the error in the thickness of the GST layer of this A-ASF and the appropriate heat treatment temperature. FIG. 50 shows the relationship between the error in the thickness of the GST layer of A-ASF (heat treatment temperature: appropriate temperature) and the optical characteristics.

When the thickness of the GST layer is the optimum value (3.7 nm) for 130° C. heat treatment, I _signal =0.66 and I _noise =0.20. In this case, since the transmission region and the stop region are in contact with each other, it is inevitable that the I _signal will be smaller than in the case where the transmission region and the stop region are separated from each other. Moreover, even if the thickness varies by ±30%, comparable values of I _signal ≧0.65 and I _noise =0.20 are secured by heat treatment at an appropriate temperature.

FIG. 51 shows the relationship between the error in the thickness of the GST layer of A-ASF of Example 23 and the reflectance R (30°) at an incident angle of 30° before heat treatment. FIG. 52 shows the relationship between the reflectance R (30°) of A-ASF of Example 23 at an incident angle of 30° before heat treatment and the appropriate heat treatment temperature.
As shown in FIGS. 51 and 52, if the relationship between the characteristics before heat treatment such as R (30°), the thickness of the GST layer, and the appropriate heat treatment temperature is understood in advance, the characteristics before heat treatment Alternatively, an appropriate heat treatment temperature can be determined uniquely from the thickness of the GST layer.

[4.2. Use of short pass filter]
[4.2.1. When θ _T1 to θ _T2 = 6 to 12°, θ _B1 to θ _B2 = 23 to 26°]
A-ASF characteristics can also be obtained using a shortpass filter (SPF). The first A-ASF (Example 24) in the case of θ _T1 to θ _T2 = 6 to 12 degrees and θ _B1 to θ _B2 = 23 to 26 degrees is made from a 40-layer TiO ₂ /SiO ₂ multilayer film. It was designed based on the SPF. As before, the thickness of each layer was determined so that the I _signal would be large under the condition of I _noise ≦0.20.

When the transmission region and the rejection region are within the above ranges and the heat treatment temperature is 130° C., the optimum value of the thickness X of the GST layer is 3.1 nm. In the following, the cases where the thickness X of the GST layer is the optimum value, the optimum value -30%, and the optimum value +30% are examined, respectively.

53 to 55 show the transmission spectra T( θ), a reflection spectrum R(θ), and an absorption spectrum A(θ). FIG. 56 shows the optical thickness of each layer of the SPF included in this A-ASF. FIG. 57 shows the relationship between the error in the thickness of the GST layer of this A-ASF and the appropriate heat treatment temperature. FIG. 58 shows the relationship between the error in the thickness of the GST layer of A-ASF (heat treatment temperature: appropriate temperature) and the optical characteristics.

Even if the thickness of the GST layer fluctuates by ±30% from the optimum value (3.1 nm) of 130°C heat treatment, the thickness of the 2C-BPF-based case (Fig. 32~ I _signal =0.70 to 0.71 and I _noise =0.20, which are close to those shown in FIG. 36), are obtained. Compared to using 2C-BPF, it has the advantage that it is easier to maintain high transmittance in the transmission region even if the thickness of TiO ₂ /SiO ₂ changes, but the disadvantage is that the number of layers in the multilayer film increases. be.

FIG. 59 shows the relationship between the error in the thickness of the GST layer of A-ASF of Example 24 and the reflectance R (30°) at an incident angle of 30° before heat treatment. The relationship between the reflectance R (30°) at an incident angle of 30° before heat treatment of A-ASF and the appropriate heat treatment temperature is shown.
As shown in FIGS. 59 and 60, if the relationship between the characteristics before heat treatment such as R (30°), the thickness of the GST layer, and the appropriate heat treatment temperature is understood in advance, the characteristics before heat treatment Alternatively, an appropriate heat treatment temperature can be determined uniquely from the thickness of the GST layer.

[4.2.2. When θ _T1 to θ _T2 = 0 to 30°, θ _B1 to θ _B2 = 30 to 60°]
Next, consider the case where θ _T1 to θ _T2 =0 to 30° and θ _B1 to θ _B2 =30 to 60°. In order to realize such a wide transmission band, it is difficult to use a multi-resonator BPF, and it is necessary to use an SPF. Based on the SPF consisting of a 40-layer TiO ₂ /SiO ₂ multilayer film, the thickness of each layer was determined so that the I _signal would be large under the condition of I _noise ≦0.20 (Example 25).

When the transmission region and the rejection region are in the above ranges and the heat treatment temperature is 130° C., the optimum value of the thickness X of the GST layer is 4.1 nm. In the following, the cases where the thickness X of the GST layer is the optimum value, the optimum value -30%, and the optimum value +30% are examined, respectively.

Figures 61 to 63 show the transmission spectra of A-ASF (Example 25), which is a combination of GST layers and SPF with different thicknesses, when heat treated at appropriate temperatures (see Figure 65). T(θ), reflection spectrum R(θ), and absorption spectrum A(θ) are shown. FIG. 64 shows the optical thickness of each layer of the SPF included in this A-ASF. FIG. 65 shows the relationship between the error in the thickness of the GST layer of A-ASF of Example 25 and the appropriate heat treatment temperature. FIG. 66 shows the relationship between the error in the thickness of the GST layer of A-ASF (heat treatment temperature: appropriate temperature) of Example 25 and the optical characteristics.

Since the transmission range is wider than in the past, reducing polarization dependence and suppressing ripples are more important. Therefore, there is a large difference in thickness between the TiO ₂ layer and the SiO ₂ layer, and the thickness changes from the center to the ends of the multilayer film (Reference Document 4). The optimal thickness of the GST layer in the case of 130° C. heat treatment is 4.1 nm, and in this case, I _signal =0.59 and I _noise =0.20. Even if the thickness of the GST layer varies by ±30%, by selecting an appropriate heat treatment temperature, the variation can be compensated for, and I _signal ≧0.57 and I _noise =0.20 can be obtained.

FIG. 67 shows the relationship between the error in the thickness of the GST layer of A-ASF of Example 25 and the reflectance R (45°) at an incident angle of 45° before heat treatment. FIG. 68 shows the relationship between the reflectance R (45°) of A-ASF of Example 25 at an incident angle of 45° before heat treatment and the appropriate heat treatment temperature.
In this case, since R (45°) in the stop band is sensitively influenced by the thickness of the GST layer, this value is appropriate for monitoring optical properties during heat treatment.

[4.3. Use of long pass filter]
Contrary to what has been done so far, it is also possible to block components that are close to normal incidence and to transmit components that have a large angle of incidence. For this purpose, a longpass filter (LPF) is used instead of 2C-BPF and SPF. For the case of θ _T1 to θ _T2 = 10 to 60° and θ _B1 to θ _B2 = 0 to 10°, I _noise ≦0.2 based on an LPF consisting of 40 TiO ₂ /SiO ₂ multilayer films. The thickness of each layer was determined so that the I _signal becomes large under the conditions (Example 26).

When the transmission region and the rejection region are in the above ranges and the heat treatment temperature is 130° C., the optimum value of the thickness X of the GST layer is 3.7 nm. In the following, the cases where the thickness X of the GST layer is the optimum value, the optimum value -30%, and the optimum value +30% are examined, respectively.

Figures 69 to 71 show the transmission spectra of A-ASF (Example 26), which is a combination of GST layers and LPFs with different thicknesses, when heat treated at appropriate temperatures (see Figure 73). T(θ), reflection spectrum R(θ), and absorption spectrum A(θ) are shown. FIG. 72 shows the optical thickness of each layer of the LPF provided in this A-ASF. FIG. 73 shows the relationship between the error in the thickness of the GST layer of A-ASF of Example 26 and the appropriate heat treatment temperature. FIG. 74 shows the relationship between the error in the thickness of the GST layer of A-ASF (heat treatment temperature: appropriate temperature) of Example 26 and the optical characteristics.

Since the transmission range is wide, suppression of ripples becomes more important, as in the case based on SPF shown above, but the way the thicknesses of the TiO ₂ layer and the SiO ₂ layer change is different. The optimal thickness of the GST layer when heat-treated at 130° C. is 3.7 nm, and in this case, I _signal =0.66 and I _noise =0.20. In this case as well, even if the thickness of the GST layer varies by ±30%, comparable values of I _signal ≧0.66 and I _noise =0.20 can be maintained by selecting an appropriate heat treatment temperature. .

FIG. 75 shows the relationship between the error in the thickness of the GST layer of A-ASF of Example 26 and the reflectance R (5°) at an incident angle of 5° before heat treatment. FIG. 76 shows the relationship between the reflectance R (5°) of A-ASF of Example 26 at an incident angle of 5° before heat treatment and the appropriate heat treatment temperature.
In this case, the inhibition zone is 0-10°. However, since a half mirror or the like is required to measure the reflectance at normal incidence, it becomes complicated. Therefore, it is practical to monitor the heat treatment at R (5°).

[5. summary]
A layer of a phase change material such as Ge ₂ Sb ₂ Te ₅ is formed on the light incidence surface of a wavelength selective filter (WSF) made of TiO ₂ /SiO ₂ , so that light at an incident angle in the transmission range is transmitted, and light at an incident angle in the stop band is transmitted. We proposed an absorption angle selective filter (A-ASF) that absorbs most of the light at the corners. Since the phase change material has a light absorption function, desired optical properties can be obtained by performing heat treatment at an appropriate temperature while monitoring the optical properties after forming the multilayer film. Therefore, even if the thickness (several nanometers) of the phase change material changes, the advantage is that the effect can be compensated for by changing the heat treatment conditions.

[Reference 1] Macleod, HA, Thin-Film Optical Filters, 4th ed. (2010) CRC Press, Boca Raton.
[Reference 4] Takeda, Y., Iizuka, H., Yamada, N., and Ito, T., "Light trapping for photovoltaic cells using polarization-insensitive angle-selective filters under monochromatic illumination," Appl. Opt. Vol. 56, No. 20(2017)pp. 5761-5767
[Reference 5] Takeda, Y., Ito, T., Yamada, N., Hasegawa, K., Mizuno, S., Ichikawa, T., Luitel, HN, Iizuka, H., Higuchi, K., Ito , H., Ichiki, A., and Motohiro, T., "Crystalline silicon photovoltaic cells used for power transmission from solar-pumped lasers: I. Light-trapping concepts," Jpn. J. Appl. Vol. 57, No. 8S3(2018)08RF05

[Reference 24] Yamada, N., Ito, T., Ito, H., Motohiro, T., and Takeda, Y., "Crystalline silicon photovoltaic cells used for power transmission from solar-pumped lasers: I. Prototype for proof of the light trapping concept," Jpn. J. Appl. Vol. 57, No. 8S3(2018)08RF07
[Reference 25] Yamada, “Phase change materials used in DVD-RAM”, Applied Physics Vol. 71, No. 5 (2002) pp.562-565
[Reference 26] Matsunaga, Yamada, "High-speed phase change optical recording material - Relationship between crystal/amorphous structure and high-speed recording rewriting/recording stability maintenance", Materia Vol. 44, No. 1 (2005) pp.17-23

[Reference 27] Matsunaga, Kojima, Yamada, Obara, Takada, “Analysis of atomic arrangement and phase change process of high-speed, high-density phase change recording materials”, Materia Vol. 52, No. 2 (2013) pp. 49 -57
[Reference 28] Xu, Z., Chen, C., Wang, Z., Wu, K., Chong, H., and Ye, H., "Optical constants acqusition and phase change properties of Ge2Sb2Te5 thin films based on spectroscopy," RSC Adv. Vol. 8, No. 37(2018)pp.21040-21046
[Reference 8] Johnson, PB and Christy, RW, "Optical constants of transition metals: Ti, V, Cr, Mn, Fe, Co, Ni, and Pd," Phys. Rev. B Vol. 9, No. 12 (1974)pp. 5056-5070

Although the embodiments of the present invention have been described in detail above, the present invention is not limited to the above embodiments, and various modifications can be made without departing from the gist of the present invention.

The optical filter according to the present invention can be used for photodetectors and the like.

Claims (8)

Optical filter with the following configuration:
(1) The optical filter is
a filter layer including alternately laminated films made of two or more types of transparent materials (A) having different refractive indexes;
and a light absorption layer formed on the incident surface side of the filter layer,
The filter layer is configured to transmit light (signal light) in a transmission range of the filter layer among the incident light,
The light absorption layer is made of a layer for absorbing light (noise light) that is in a blocking range of the filter layer, out of the incident light.
(2) The light absorption layer is made of a light absorption material that satisfies the following formula (1).
πk _A /n _A ≧1 ... (1)
however,
n _A is the refractive index of the light absorption layer when the light absorption layer is irradiated with the incident light;
k _A is the extinction coefficient of the light absorption layer when the light absorption layer is irradiated with the incident light;
The light-absorbing material includes a material whose extinction coefficient can be irreversibly changed by heating. The optical filter according to claim 1, wherein the filter layer has a function as a wavelength selection filter for the incident light. The optical filter according to claim 1 or 2, wherein the filter layer has a function as a bandpass filter, a shortpass filter, or a longpass filter. The optical filter according to any one of claims 1 to 3, wherein the formation position of the light absorption layer satisfies the following formula (2).
0≦Σn _i d _i ≦1.5λ (2)
however,
d _i is the actual thickness of the i-th surface layer (i is an integer greater than or equal to zero) formed on the light absorption layer,
n _i is the refractive index of the i-th surface layer when the i-th surface layer is irradiated with the incident light;
λ is the wavelength of the incident light. The optical filter according to any one of claims 1 to 4, wherein the light absorption layer satisfies the following formula (3).
0.005≦πk _A d _A /(n _A・λ)≦0.2 (3)
However, _dA is the actual thickness of the light absorption layer, and λ is the wavelength of the incident light. Any one of claims 1 to 5 , wherein the light absorbing material includes one or more selected from the group consisting of Ge ₂ Sb ₂ Te ₅ , GeSb ₂ Te ₄ , GeSb ₄ Te ₇ , and AgInSbTe-based optical recording materials. The optical filter according to item 1 . A light-absorbing layer containing a light-absorbing material whose extinction coefficient can be irreversibly changed by heating on the entrance surface side of a filter layer including alternately laminated layers of two or more transparent materials (A) having different refractive indexes. A laminate production step of producing a laminate in which
A method for manufacturing an optical filter, comprising a heat treatment step of adjusting filter characteristics of the laminate by heat-treating the laminate to obtain the optical filter according to any one of claims 1 to 6 . The heat treatment step includes (a) performing heat treatment while monitoring the transmission/absorption characteristics of the laminate;
(b) a step of measuring the thickness of the light absorption layer and performing heat treatment under conditions according to the thickness of the light absorption layer, or
8. The method for manufacturing an optical filter according to claim 7 , comprising the step of: (c) measuring the optical properties of the laminate and performing heat treatment under conditions according to the optical properties.

Images