Photodiode, light wave detector, frequency down-conversion device, and communication device

Abstract

In order to improve heat radiation performance, response characteristics and photoelectric conversion efficiency of a photodiode, there is provided a photodiode (12) having a layered structure including light-absorbing layers (124a, 124b, 124c, 124d) that absorb light and generate photoelectrons / light holes, and a plurality of periodically arranged lattice walls (121a, 121b, 121c, 121d, 121e). The layered structure is divided into a plurality of active regions by the plurality of lattice walls (121a, 121b, 121c, 121d, 121e). Each of the lattice walls (121a, 121b, 121c, 121d, 121e) is composed of an insulator having a refractive index different from that of the plurality of active regions. The insulator constituting each of the lattice walls (121a, 121b, 121c, 121d, 121e) is composed of a material having a higher thermal conductivity than the light-absorbing layers (124a, 124b, 124c, 124d), for example, SiN or AlN.

Description

Photodiode, light wave detector, frequency down-conversion device, and communication device

[0001] The present application relates to photodiodes, light wave detectors, frequency down-conversion devices, and communication devices.

[0002] Photodiodes (PDs) are widely used in fields such as light detection, optical communications, measurement, spectroscopic analysis, imaging, and environmental technology. In particular, PDs using compound semiconductors are used in a variety of applications because their wavelength sensitivity range can be adjusted by changing their composition ratio. For example, PDs are used in optical power meters that detect light intensity, laser diode (LD) monitors that monitor the output level of LDs, optical time domain reflectometers (OTDRs), optical receivers in optical communication networks, signal converters that convert optical signals into wireless signals, distance measuring devices such as LiDAR, thermometers, flame detectors, moisture meters, gas analyzers, spectrophotometers, and imaging devices.

[0003] Performance indicators of PDs include response speed, photoelectric conversion efficiency, and saturation output. Furthermore, in cases where a high-intensity light wave input is required to obtain a large signal output, thermal damage / destruction of the PD due to the high-intensity light wave input can become a problem. In such cases, the heat dissipation performance of the PD is also an important performance indicator. Generally, response speed is improved by thinning the light absorption layer of the PD. On the other hand, photoelectric conversion efficiency is improved by thickening the light absorption layer of the PD. In other words, there is a trade-off between response speed and photoelectric conversion efficiency. Therefore, there is a need for a method to improve photoelectric conversion efficiency without increasing the thickness of the light absorption layer.

[0004] Non-Patent Document 1 proposes a technology for improving the photoelectric conversion efficiency of PDs by utilizing the guided-mode resonance (GMR) effect. Specifically, the method proposes a heterostructure in which InP (p-doped layer), InGaAs (neutral layer), and InP (n-doped layer) are stacked on an Au mirror, and dielectric gratings are periodically arranged within the p-doped layer. Benzocyclobutene (BCB), which has a refractive index different from that of the heterostructure material, is used for the dielectric grating. Non-Patent Document 1 shows that a guided mode is induced within the heterostructure in response to an optical input having a resonant wavelength, thereby resonantly improving the external quantum efficiency (EQE) of the PD.

[0005] Verdun et al, "Dark current investigation in thin PiN InGaAs photodiodes for nano-resonators", Journal of Applied Physics, 120, 084501 (2016)

[0006] The technology of Non-Patent Document 1 focuses on the photoelectric conversion efficiency of PDs, and application of the technology of Non-Patent Document 1 does not improve the response speed of PDs. Based on this, this application proposes a method for improving both the response speed and photoelectric conversion efficiency of PDs. Note that application of the technology of Non-Patent Document 1 also has the problem of not improving the heat dissipation performance of PDs, and a solution to this problem is proposed in this specification.

[0007] According to a first aspect of the present application, there is provided a photodiode sandwiched between a metal layer (first electrode) and a semiconductor contact layer connected to a second electrode. The photodiode has a stacked structure including a light-absorbing layer that absorbs light and generates photoelectrons and photoholes, and a plurality of periodically arranged lattice walls. The stacked structure is divided into a plurality of active regions by the lattice walls. Each lattice wall is made of an insulator having a refractive index different from that of the active regions. One end of each lattice wall contacts the metal layer, and the other end of each lattice wall contacts the semiconductor contact layer. The stacked structure may also include a carrier collection layer in addition to the light-absorbing layer. The carrier collection layer is sometimes referred to as a "carrier transport layer" in reference to its function of accelerating photoelectrons by an electric field generated therein and transporting the photoelectrons. The light-absorbing layer may also function as a carrier collection layer.

[0008] The photodiode may be a UTC-PD (Uni-Traveling-Carrier Photo-Diode) with a high response speed. In a UTC-PD, the light absorption layer is composed of a p-type doped layer, and electrons (minority carriers) generated by light absorption are diffused to the carrier collection layer. Furthermore, in a UTC-PD, a p-type diffusion prevention layer is provided between the p-electrode (first electrode) and the light absorption layer to prevent back-diffusion of electrons toward the p-electrode (first electrode). In a UTC-PD, photoholes, which are majority carriers in the light absorption layer and have a slow response speed, are collectively extracted to the p-electrode within approximately the dielectric relaxation time and rapidly disappear, so the photoresponse speed is determined by the transit time of photoelectrons. The diffusion prevention layer and the light absorption layer are composed of semiconductor materials such as InGaAs, InAlGaAs, and InGaAsP. Because electrons in InP / InGaAs / InGaAsP have high mobility, a UTC-PD can achieve a higher response speed than other photodiodes. The response speed is related to the cutoff frequency, and the faster the response speed, the higher the cutoff frequency. A higher cutoff frequency also means a wider frequency response band.

[0009] The cutoff frequency of a typical UTC-PD is the electron transit time τ absand the electron transit time τ in the carrier collection layer cc The sum of the area A of UTC-PD UTC-PD (corresponding to the area of ​​the cross section perpendicular to the stacking direction) is smaller, the higher the value. UTC-PD As the electron transit time τ decreases, the junction capacitance decreases, and the RC time constant determined by the sum of the external load resistance and the junction resistance decreases, resulting in a higher cutoff frequency (this will be explained in more detail later with reference to FIG. 14). abs and τ cc depends on the thickness of the light absorption layer and the carrier collection layer. UTC-PD The smaller the cutoff frequency, the higher the response speed and the wider the response frequency band.

[0010] In the first aspect, the stacked structure is divided into a plurality of active regions by a plurality of lattice walls made of an insulator. Therefore, the response speed of the photodiode according to the first aspect is determined by the response speed of each active region. Although the thickness of each layer constituting one active region does not change due to the division, the area A of one active region Active (corresponding to the area of ​​the cross section perpendicular to the stacking direction) is area A UTC-PD . Therefore, the response speed in each active area is higher than when it is not divided (i.e., a general UTC-PD). In this way, according to the first aspect, it is possible to improve the response speed and expand the response frequency band without reducing the photoelectric conversion efficiency. Note that although the description is given here using a UTC-PD as an example, the technology according to the first aspect is also widely applicable to p-i-n photodiodes and the like, and that even in such cases, it is possible to improve the response speed and expand the response frequency band without reducing the photoelectric conversion efficiency.

[0011] In the first aspect, since a plurality of lattice walls having a refractive index different from that of the active region are periodically arranged, a waveguide mode is induced in each active region in response to an input light wave having a resonant wavelength, and the light intensity in the light absorption layer is locally enhanced, thereby achieving a resonantly high photoelectric conversion efficiency. The arrangement period of the lattice walls (i.e., the width of each lattice wall and the width of each active region) is preset to a value that provides the GMR effect.

[0012] To achieve high heat dissipation performance, the insulators constituting the multiple lattice walls are preferably made of a material with a higher thermal conductivity than the light absorption layer, such as SiN or AlN. In a typical UTC-PD, the diffusion prevention layer, light absorption layer, and carrier collection layer are the main heat sources. Semiconductor materials such as InGaAs, InAlGaAs, and InGaAsP constituting the diffusion prevention layer and light absorption layer have relatively low thermal conductivity, making it difficult for generated heat to be transferred to the metal layer and for heat to escape to the outside of the device. On the other hand, in the first aspect, each lattice wall contacts the entire stacked structure, and one end of each lattice wall contacts the metal layer (first electrode). Therefore, when each lattice wall is made of a material such as SiN or AlN, heat is quickly transferred to the metal layer via each lattice wall, allowing the generated heat to be quickly dissipated to the outside. This configuration achieves high heat dissipation performance and reduces the risk of thermal damage / destruction due to high-intensity light wave input.

[0013] The insulator lattice in Non-Patent Document 1 is composed of BCB. In this case, the thermal conductivity of the insulator lattice is low, and the heat dissipation effect of the insulator lattice cannot be expected. Furthermore, the insulator lattice in Non-Patent Document 1 is low in height and is embedded inside the p-type doped layer. In this case, heat from the heat source is not directly transmitted to the insulator lattice, so even if the material of the insulator lattice is replaced from BCB to SiN or AlN, a sufficient heat dissipation effect cannot be obtained. In contrast, in the first aspect, each lattice wall is in contact with each layer, and therefore, by constructing the lattice wall from a material with high thermal conductivity, high heat dissipation performance is achieved.

[0014] In one implementation example of the first aspect, the plurality of grating walls include only a plurality of first grating walls arranged at equal intervals in a first direction. The first direction is any direction perpendicular to the stacking direction. For example, if the first direction is the Y direction and the stacking direction is the Z direction, the arrangement of the grating walls in this implementation example is an arrangement in which a plurality of first grating walls having wall surfaces parallel to the ZX plane are arranged at regular intervals. In the following description, a structure having such an arrangement of grating walls may be referred to as a "one-dimensional GMR structure."

[0015] In another implementation example of the first aspect, the plurality of grating walls include a plurality of first grating walls arranged at equal intervals in a first direction and one or more second grating walls arranged at equal intervals in a second direction different from the first direction. The first direction and the second direction are perpendicular to the stacking direction, and the second direction is different from the first direction. For example, if the first direction is the Y direction, the stacking direction is the Z direction, and the second direction is the X direction, the first grating wall is an insulating wall having a wall surface parallel to the Z-X plane, and the second grating wall is an insulating wall having a wall surface parallel to the Z-Y plane, and the combination of these forms a grid pattern. Each active region is surrounded by the first grating wall and the second grating wall. In the following description, a structure having such a grating wall arrangement may be referred to as a "two-dimensional GMR structure."

[0016] For the same photodiode area and the same number of first lattice walls, the area of ​​each active region in a 2D GMR structure is smaller than that in a 1D GMR structure. This is advantageous from the viewpoints of improving response speed and expanding the response frequency band. On the other hand, the 1D GMR structure has a simpler structure than a 2D GMR structure, which is advantageous from the viewpoints of ease of design and manufacturing. For the same wall thickness, when comparing the area of ​​the total active region relative to the area of ​​the photodiode (hereinafter referred to as the fill factor), the fill factor of a 1D GMR structure is greater than that of a 2D GMR structure. A higher fill factor can be advantageous in terms of photoelectric conversion efficiency. On the other hand, a lower fill factor means that there are more heat dissipation paths formed by the lattice walls, so a lower fill factor is advantageous in terms of heat dissipation performance.

[0017] According to a second aspect of the present application, there is provided a lightwave detector. This lightwave detector has a structure in which the photodiode of the first aspect and a high electron mobility transistor (HEMT) are monolithically integrated, and the HEMT is used instead of the semiconductor contact layer. The photodiode is sandwiched between a source electrode and the HEMT. The lightwave detector is configured to read out a photovoltage signal generated by a lightwave incident on the photodiode from a gate electrode or a drain electrode disposed on the HEMT. By applying the above-mentioned high-performance photodiode, a lightwave detector with excellent response and detection performance can be realized.

[0018] According to a third aspect of the present application, there is provided a frequency down-conversion device. This frequency down-conversion device has a structure in which the photodiode of the first aspect and a high electron mobility transistor (HEMT) are monolithically integrated, and the HEMT is used instead of the semiconductor contact layer. The photodiode is sandwiched between a source electrode and the HEMT. The frequency down-conversion device is configured to output a data signal having a frequency in the radio frequency band or baseband from a gate electrode or a drain electrode disposed on the HEMT in response to input of an optical data signal and an optical subcarrier signal having different frequencies in the infrared optical frequency band. By applying the above-mentioned high-performance photodiode, a frequency down-conversion device with excellent response performance and conversion gain can be realized.

[0019] According to a fourth aspect of the present invention, there is provided a communication device. The communication device includes the frequency down-conversion device according to the third aspect, an optical processing unit that inputs an optical data signal and an optical subcarrier signal transmitted through an optical fiber to the frequency down-conversion device, and a radio processing unit that converts the data signal output from the frequency down-conversion device into a radio signal and transmits the radio signal via an antenna. By applying the above-mentioned high-performance frequency down-conversion device, a communication device with high processing performance can be realized.

[0020] According to the present invention, the response characteristics and photoelectric conversion efficiency of the photodiode are improved.

[0021] FIG. 1 is a schematic diagram of an optical semiconductor device according to an embodiment of the present application. FIG. 2 is a schematic diagram for explaining a stacked structure of a photodiode according to an embodiment of the present application. FIG. 3 is a schematic diagram for explaining a first structure (one-dimensional GMR structure) of a photodiode according to an embodiment of the present application. FIG. 4 is a schematic diagram for explaining a second structure (two-dimensional GMR structure) of a photodiode according to an embodiment of the present application. FIG. 5 is a schematic diagram for explaining an implementation example (optoelectronic fusion transistor) of a photodiode according to an embodiment of the present application. FIG. 6 is a diagram showing absorbance characteristics of a photodiode according to an embodiment of the present application. FIG. 7 is a schematic diagram for explaining a first simulation model (example: high GR). FIG. 8 is a schematic diagram for explaining a second simulation model (comparative example: no GR). FIG. 9 is a schematic diagram for explaining a third simulation model (comparative example: low GR). FIG. 10 is a diagram showing simulation results of heat dissipation characteristics (example 1). FIG. 11 is a diagram showing simulation results of heat dissipation characteristics (example 2). FIG. 12 is a diagram showing simulation results of heat dissipation characteristics (comparative example A). FIG. 13 is a diagram showing simulation results of absorbance and heat dissipation characteristics. FIG. 14 is a schematic diagram for explaining an RC time constant. FIG. 15 is a diagram comparing simulation results of cutoff frequencies. 1 is a schematic diagram for explaining a frequency down-conversion device according to an embodiment of the present application;FIG. 2 is a schematic diagram for explaining a communication device according to an embodiment of the present application;FIG.

[0022] Hereinafter, an embodiment of the present application (hereinafter, the present embodiment) will be described with reference to the accompanying drawings. In this specification and the drawings, elements having substantially the same functions are designated by the same reference numerals, and redundant description may be omitted.

[0023] (Optical Semiconductor Device) An optical semiconductor device 10 will be described with reference to Fig. 1. The optical semiconductor device 10 is an example of an optical semiconductor device to which the photodiode according to this embodiment can be applied.

[0024] 1, an optical semiconductor device 10 includes a metal layer (first electrode) 11, a photodiode 12, a second electrode 13, a semiconductor contact layer 14, and a semiconductor substrate 15. The photodiode 12 is disposed so as to be sandwiched between the metal layer 11 and the semiconductor contact layer 14 connected to the second electrode 13.

[0025] The first electrode 11 is a cathode electrode. The second electrode 13 is an anode electrode. Light waves are input to the optical semiconductor device 10 from the semiconductor substrate 15 side. In response to the light wave input, photoelectrons and photoholes are generated in the photodiode 12, and photoelectrons flow out to the second electrode 13 via the semiconductor contact layer 14, while photoelectrons flow in from the first electrode 11.

[0026] The configuration of the optical semiconductor device 10 shown in Fig. 1 is merely an example for the purpose of explanation, and the application scope of the technology according to this embodiment is not limited to the example in Fig. 1. This embodiment is characterized by the configuration of the photodiode 12, and any optical semiconductor device to which the photodiode 12 can be applied can be included in the application scope of the technology according to this embodiment. The optical semiconductor device 10 can be applied to, for example, an optical power meter, an LD monitor, an OTDR, an optical receiver, a signal conversion device, a distance measuring device, various measurement devices, an imaging device, and the like.

[0027] Fig. 2 shows a cross-sectional view of the photodiode 12 taken along the line II-II in Fig. 1. The layered structure of the photodiode 12 will be described with reference to Fig. 2. For convenience of explanation, Fig. 2 also shows a cross section of the metal layer 11.

[0028] 2, the photodiode 12 includes photo-absorption layers 124a...124d that absorb light and generate photoelectrons, carrier collection layers 125a...125d that collect the photoelectrons, and a plurality of periodically arranged grating walls 121a...121e. In the example shown in FIG. 2, diffusion block layers 123a...123d are provided on the photo-absorption layers 124a...124d, and p-type contact layers 122a...122d are provided on the diffusion block layers 123a...123d. The semiconductor contact layer 14 (see FIG. 1) located below the p-type contact layers 122a...122d is an n-type contact layer. The "carrier collection layers" are sometimes called "carrier traveling layers" in relation to their function of accelerating photoelectrons by an electric field generated therein and causing the photoelectrons to travel.

[0029] As can be seen from Figure 2, unlike a typical PD (e.g., a typical pin-type PD or UTC-PD), in this embodiment, a stacked structure including a light absorption layer and a carrier collection layer is divided into multiple active regions by multiple lattice walls. In the example of Figure 2, multiple lattice walls 121a, ..., 121e divide the light absorption layer and the carrier collection layer into four light absorption layers 124a, ..., 124d and four carrier collection layers 125a, ..., 125d. In this example, the p-type contact layer 122a, the diffusion prevention layer 123a, the light absorption layer 124a, and the carrier collection layer 125a constitute one active region (hereinafter referred to as active region A).

[0030] 2, the p-type contact layer 122b, the diffusion prevention layer 123b, the light absorption layer 124b, and the carrier collection layer 125b form an active region B, the p-type contact layer 122c, the diffusion prevention layer 123c, the light absorption layer 124c, and the carrier collection layer 125c form an active region C, and the p-type contact layer 122d, the diffusion prevention layer 123d, the light absorption layer 124d, and the carrier collection layer 125d form an active region D. Here, an example is shown in which the stacked structure is divided into four active regions A to D, but the division method (such as the number of divisions) is not limited to this example.

[0031] The photodiode 12 can be formed, for example, by stacking a p-type contact layer, a diffusion prevention layer, a light absorption layer, and a carrier collection layer on the metal layer 11, etching out regions corresponding to the lattice walls, and filling the removed regions with a lattice wall material. The lattice wall material is an insulating material with a refractive index different from that of the active region. As an example, the lattice wall material can be an insulating material with a higher thermal conductivity than the light absorption layer and the diffusion prevention layer, such as SiN or AlN.

[0032] As can be seen from FIG. 2 , one end of each lattice wall contacts the metal layer 11. Furthermore, each lattice wall contacts each layer in the active region. In the example of FIG. 2 , the main heat sources are the diffusion prevention layer, the light absorption layer, and the carrier collection layer. Materials such as InAlGaAs, InGaAs, and InGaAsP, which are often used for the diffusion prevention layer and the light absorption layer in UTC-PDs, have low thermal conductivity, and heat is not easily transferred to the metal layer 11 through these materials. However, by using an insulating material with high thermal conductivity, heat can be efficiently transferred from the heat source sandwiched between the multiple lattice walls to the metal layer 11 through each lattice wall. This suppresses excessive temperature rise in the active region due to light wave input, significantly reducing the risk of thermal damage or destruction.

[0033] As shown in FIG. 2, the multiple lattice walls 121a, ..., 121e are arranged periodically. In the example of FIG. 2, the wall surface of each lattice wall is perpendicular to the Y direction and parallel to the Z-X plane. The thickness (size in the Y direction) of each lattice wall is w1, and the height (size in the Z direction) of each lattice wall is H. The thickness (size in the Y direction) of the active region is w2, and the stacking direction is the Z direction. Thus, in the example of FIG. 2, the multiple lattice walls 121a, ..., 121e are arranged side by side at a constant period (w1 + w2) in the Y direction.

[0034] The refractive indices of the layers in the active region are similar, and each region surrounded by the metal layer 11 and the multiple grating walls can be considered a single waveguide. The multiple grating walls are an insulating grating with a different refractive index from the active region. When a light wave is input in the direction perpendicular to the grating plane (the Z direction), the insulating grating functions as a grating coupler (GC) for the light wave component with the resonant wavelength. This excites guided light within the waveguide. This guided light propagates in the ±Y direction and has an electric field in the Z direction due to the diffraction effect of the GC. The input light wave (or the light wave reflected by the metal layer 11) propagates in the Z (or −Z) direction and has an electric field in the Y direction. The wavelength of the guided light is determined by the grating period. When the grating period is set appropriately, the light absorption layer absorbs both the guided light and the optical wave (GMR effect). The GMR effect increases the absorbance in the light absorption layer, resulting in increased photoelectric conversion efficiency.

[0035] In general, photoelectric conversion efficiency can be improved by thickening the light absorption layer. However, as the light absorption layer becomes thicker, photoelectrons generated by the input light wave diffuse widely within the light absorption layer, increasing the time it takes for the photoelectrons to reach the carrier collection layer and reducing the response speed. The method using the GMR effect described above increases photoelectric conversion efficiency by using multiple lattice walls without changing the height H of the active region. This has the advantage of not reducing the response speed.

[0036] The higher the response speed of a PD, the higher the cutoff frequency. Generally, the cutoff frequency is defined as the frequency at which the output intensity is reduced to half compared to the low frequency limit. A high cutoff frequency also means a wide response frequency band. For example, in the case of a typical UTC-PD, the cutoff frequency is determined by the electron transit time τ in the light absorption layer. abs and the electron transit time τ in the carrier collection layer cc The sum of the area A of UTC-PD UTC-PD (corresponding to the area of ​​the XY cross section of the entire UTC-PD) is smaller, the higher the value. UTC-PD As the electron transit time τ decreases, the junction capacitance decreases, and the RC time constant determined by the sum of the external load resistance and the junction resistance decreases, resulting in a higher cutoff frequency (this will be explained in more detail later with reference to FIG. 14). abs and τ cc mainly depends on the thickness of the light absorbing layer and the carrier collecting layer. UTC-PD The smaller the value, the higher the cutoff frequency, resulting in a faster response speed and a wider response frequency band.

[0037] In the photodiode 12, the laminated structure is divided into a plurality of active regions by a plurality of insulating lattice walls 121a, ..., 121e. Since each active region is electrically separated from the other active regions by the lattice walls, the cutoff frequency of the photodiode 12 is determined by the area A of each active region. Active (corresponding to the area of ​​the XY cross section of each active area) Area A Active is the area of ​​the entire photodiode 12 (the above A UTC-PD Therefore, the photodiode 12 has a high cutoff frequency, a wide response frequency band, and a high response speed. A method for calculating the cutoff frequency will be described in detail later.

[0038] (One-Dimensional GMR Structure) Next, with reference to FIG. 3 , the first structure (one-dimensional GMR structure) of the photodiode 12 will be described. When only a plurality of lattice walls 121a, ..., 121e (first lattice walls) having wall surfaces parallel to the Z-X plane are arranged side by side, the top view of the photodiode 12 will be as shown in FIG. 3 . For ease of explanation, a PD structure having such a lattice arrangement will be referred to as a "one-dimensional GMR structure." FIG. 3 is a schematic diagram showing the photodiode 12 viewed from above (in the -Z direction) with the metal layer 11 placed on the photodiode 12 removed. In the case of a one-dimensional GMR structure, lattice walls are arranged to sandwich each active region from both sides. The depth D of each lattice wall corresponds to the size of the photodiode 12 in the X direction.

[0039] (Two-Dimensional GMR Structure) Next, with reference to FIG. 4, a second structure (two-dimensional GMR structure) of the photodiode 12 will be described. FIG. 4 shows a top view of the photodiode 12. The grid-like two-dimensional lattice 121′ shown in FIG. 4 is a structure that combines multiple lattice walls 121a, ..., 121e (first lattice walls) having wall surfaces parallel to the Z-X plane and multiple lattice walls (second lattice walls) having wall surfaces parallel to the Y-Z plane. The second lattice walls refer to the portions of the two-dimensional lattice 121′ other than the first lattice walls (see FIG. 3). In the example of FIG. 4, the two-dimensional lattice 121′ includes multiple second lattice walls, but the number of second lattice walls may be one. Alternatively, the two-dimensional lattice 121′ may be integrally formed. Hereinafter, for ease of explanation, a PD structure having such a lattice structure will be referred to as a "two-dimensional GMR structure."

[0040] In a two-dimensional GMR structure, the stacked structure is further divided by the second lattice walls. In the example of the one-dimensional GMR structure shown in FIG. 3, the stacked structure is divided into four active regions, but in the example of the two-dimensional GMR structure shown in FIG. 4, the stacked structure is divided into eight active regions (FIG. 4 shows p-type contact layers 122a', ..., 122h' of the eight active regions). When there are multiple second lattice walls, the second lattice walls are also arranged periodically. In the example of FIG. 4, the width of the second lattice wall is w3, and the width of each active region in the X direction is w4. In this case, the period of the lattice walls in the X direction is (w3 + w4). Furthermore, the period of the lattice walls in the Y direction is (w1 + w2).

[0041] When the area of ​​the photodiode 12 and the number of first lattice walls are the same, the area of ​​each active region in a 2D GMR structure is smaller than that in a 1D GMR structure. This is advantageous from the viewpoints of improving response speed and expanding the response frequency band. On the other hand, the 1D GMR structure has a simpler structure than the 2D GMR structure, which is advantageous from the viewpoints of ease of design and manufacturing. When the wall thickness is the same, when comparing the area of ​​the active region relative to the area of ​​the photodiode 12 (filling factor), the filling factor of the 1D GMR structure is greater than that of the 2D GMR structure. A higher filling factor can be advantageous in terms of photoelectric conversion efficiency. On the other hand, a lower filling factor means that there are more heat dissipation paths formed by the lattice walls, so a lower filling factor is advantageous from the viewpoint of heat dissipation performance.

[0042] As described above, the photodiode 12 according to this embodiment has multiple lattice walls, thereby achieving high photoelectric conversion efficiency, high response speed, and a wide response frequency band. Below, we will explain an optoelectronic fusion transistor as an implementation example of the above-described photodiode 12, and show simulation results for its heat dissipation characteristics and response characteristics.

[0043] (Implementation Example: Optoelectronic Fusion Transistor) With reference to FIG. 5, an implementation example (optoelectronic hybrid transistor 20) of the photodiode according to this embodiment will be described.

[0044] As shown in FIG. 5 , the photonic-electronic fusion transistor 20 includes a first electrode 21, a UTC-PD 22, a second electrode 23, a HEMT 24, an InP substrate 25, and a third electrode 26. The photonic-electronic fusion transistor 20 has a monolithic integrated structure of the UTC-PD 22 and the HEMT 24. Therefore, the UTC-PD 22 is sandwiched between the first electrode 21 and the HEMT 24. The first electrode 21 is a source electrode, the second electrode 23 is a drain electrode, and the third electrode 26 is a gate electrode. Note that the electrode structure of the photonic-electronic fusion transistor 20 is not limited thereto. For example, the interdigital electrode structure described in Japanese Patent No. 7,440,988 may also be applied.

[0045] The UTC-PD 22 corresponds to the photodiode 12 described above, and has the layer structure shown in FIGS. 2 to 4 (see FIGS. 2 and 3 for the one-dimensional GMR structure / see FIGS. 2 and 4 for the two-dimensional GMR structure). The HEMT 24 has, for example, a layer structure as shown in Table 1 below. However, the layer structure shown in Table 1 is merely an example for the purpose of explanation, and the material, composition ratio, and layer thickness of each layer may be changed as appropriate.

[0046]

[0047] (Absorbance Characteristics of Photodiode) Next, the absorbance characteristics of the photodiode according to this embodiment will be described with reference to FIG.

[0048] The absorbance characteristics of the UTC-PD 22 were investigated assuming the above-described photonic-electronic fusion transistor 20. Specifically, a simulation was performed assuming a photonic-electronic fusion transistor 20 with a one-dimensional GMR structure (see Figures 2 and 3), and the absorbance distribution was obtained for the lattice wall thickness w1 and the active region thickness w2. The simulation results are shown in Figure 6. In this simulation, the first electrode 21 was made of Au, and the material of each lattice wall in the UTC-PD 22 was SiN (refractive index: 2.4629). Other layer configurations and parameters are as shown in Table 2 below. The refractive index in Table 2 is the refractive index for light waves with a wavelength of 1550 nm. Periodic boundary conditions were applied assuming a case where a sufficient number of lattice walls are provided.

[0049]

[0050] When 1550 nm CW (Continuous Wave) light was incident on the bottom side (HEMT 24 side) of the UTC-PD 22, the maximum absorbance exceeded 46% (0.46 in the figure), as shown in Figure 6. This value is approximately 2.42 times the maximum absorbance (approximately 19%) obtained for a conventional UTC-PD without lattice walls under the same conditions. Furthermore, from the shape of the absorbance peak distribution, it can be seen that the improvement in maximum absorbance described above is due to the GMR effect. While the above results are for the UTC-PD 22 alone, the photonics-electronics fusion transistor 20 combined with the HEMT 24 has a photoelectric conversion gain of approximately 5.86 times (=2.42 times) compared to the conventional method. 2 ) (See Japanese Patent No. 7440988 for the conversion gain amplification effect of photonic-electronic fusion transistors).

[0051] (Simulation Results of Heat Dissipation Characteristics / Response Characteristics) Next, the simulation results regarding the heat dissipation characteristics and response characteristics will be described.

[0052] First, a first simulation model (embodiment: high GR) will be described with reference to FIG. 7. The first simulation model corresponds to a photonics-electronics fusion transistor including a UTC-PD with a high lattice wall (high GR) as shown in FIG. 2. As shown in FIG. 7, the first simulation model includes a pad electrode 51, a contact electrode 52, a lattice wall 53, a diffusion prevention layer 54, a light absorption layer 55, a carrier collection layer 56, a HEMT 57, insulator layers 58a and 58b, and an InP substrate 59.

[0053] The width w1 of the lattice walls, the width w2 of the active region, and the period (w1 + w2) of the lattice walls used in the simulation are as shown in Table 3 below. The materials and thicknesses of the pad electrode 51, contact electrode 52, UTC-PD (diffusion prevention layer 54, light absorption layer 55, and carrier collection layer 56), and HEMT 57 used in the simulation are as shown in Table 4 below. The material of the lattice walls 53 and the insulator layers 58a and 58b is SiN or AlN. The width of the insulator layers 58a and 58b is 100 nm. It is assumed that the InP substrate 59 has a width and thickness of 600 μm, and that a 5 μm-thick Au electrode is placed on the pad electrode 51.

[0054]

[0055]

[0056] The thermal conductivity parameters of the materials used in the first simulation model are shown in Table 5 below. For reference, Table 5 also lists the thermal conductivity parameters of BCB. As is clear from Table 5, the SiN and AlN used for the lattice walls have higher thermal conductivity than InAlGaAs, InGaAs, and InAlAs, which may be used in the active region of a UTC-PD. On the other hand, the BCB used for the insulator lattice in Non-Patent Document 1 has a significantly lower thermal conductivity than SiN and AlN, and also has a lower thermal conductivity than InAlGaAs, InGaAs, and InAlAs. This also explains why the insulator lattice in Non-Patent Document 1 does not contribute to heat dissipation.

[0057]

[0058] Next, a second simulation model (comparative example: without GR) will be described with reference to Fig. 8. As shown in Fig. 8, the second simulation model includes a pad electrode 61, a contact electrode 62, a diffusion prevention layer 63, a light absorption layer 64, a carrier collection layer 65, a HEMT 66, insulator layers 67a and 67b, and an InP substrate 68. In other words, the second simulation model is obtained by replacing the UTC-PD of the first simulation model with a UTC-PD without a lattice wall (without GR).

[0059] The materials and thicknesses of the pad electrode 61, contact electrode 62, UTC-PD (diffusion prevention layer 63, light absorption layer 64, and carrier collection layer 65), and HEMT 66 used in the simulation are as shown in Table 4 above. The material of the insulator layers 67a and 67b is SiN or AlN. The width of the insulator layers 67a and 67b is 100 nm. It is assumed that the InP substrate 68 has a width and thickness of 600 μm, and that a 5 μm-thick Au electrode is placed on the pad electrode 61. The thermal conductivity parameters of the materials used in the second simulation model are also as shown in Table 5 above.

[0060] 9, a reference model (low GR) in which the UTC-PD of the first simulation model is replaced with a UTC-PD having a low lattice wall (low GR) will be discussed. The reference model shown in FIG. 9 includes a pad electrode 71, a contact electrode 72, a lattice 73, a diffusion prevention layer 74, a light absorption layer 75, a carrier collection layer 76, a HEMT 77, insulator layers 78a and 78b, and an InP substrate 79.

[0061] The width of the lattice 73 is the same as the width w1 of the lattice wall 53 (see Table 3 above), and the spacing between the lattices 73 is the same as the width w2 of the active region in the first simulation model (see Table 3 above). The materials and thicknesses of the pad electrode 71, contact electrode 72, UTC-PD (diffusion prevention layer 74, light absorption layer 75, and carrier collection layer 76), and HEMT 77 are assumed to be as shown in Table 4 above. The materials of the lattice 73 and insulator layers 78a and 78b are SiN, AlN, or BCB. The width of the insulator layers 78a and 78b is 100 nm. The InP substrate 79 is assumed to have a width and thickness of 600 μm, and a 5 μm-thick Au electrode is placed on the pad electrode 71.

[0062] The reference model is similar to the first simulation model, but as shown in FIG. 9 , the lattice 73 is embedded in the diffusion prevention layer 74, and its structure is clearly different from that of the lattice wall 53. In particular, the lattice 73 is not in contact with the light absorption layer 75 and the carrier collection layer 76, which are heat sources. Therefore, even if the lattice 73 is made of SiN or AlN, which have high thermal conductivity, it contributes very little to heat dissipation. Of course, if BCB, which has low thermal conductivity, is used as the material for the lattice 73, heat dissipation performance may be degraded.

[0063] However, in the reference model, the gratings 73 are periodically arranged, so that the GMR effect can be obtained. However, the gratings 73 are not structured to divide the stacked structure of the UTC-PD (the diffusion prevention layer 74, the light absorption layer 75, and the carrier collection layer 76) into multiple active regions. Therefore, the cutoff frequency of the UTC-PD is determined by the area A of the entire UTC-PD. UTC-PD Therefore, the configuration of the reference model does not provide the effects of improving the response speed and widening the response frequency band. This means that even if the technology of Non-Patent Document 1 is applied to a photonics-electronics fusion transistor, the special effects of the technology of this embodiment cannot be obtained.

[0064] (Simulation Results of Heat Dissipation Characteristics) Hereinafter, the simulation results of heat dissipation characteristics will be described.

[0065] Fig. 10 shows the results of a simulation of heat dissipation characteristics (Example 1). Fig. 10 shows the temperature distribution when AlN is used as the material for the lattice walls 53 using the first simulation model. In the simulation, the temperature at a location sufficiently distant from the UTC-PD (the InP substrate 59 and the end of the Au electrode) is kept at 30°C, and the UTC-PD acts as a heat source and generates a constant heat density (10 16 W / m 3 = 10 mW / μm 3 10 shows that the maximum temperature increase ΔT is 27°C near the center of the carrier collection layer 56, and the temperature decreases with increasing distance from the center. In particular, it is clear that heat is transferred to the metal layer via the lattice walls 53.

[0066] Fig. 11 shows the results of a simulation of heat dissipation characteristics (Example 2). Fig. 12 shows the temperature distribution when the first simulation model is used and SiN is used as the material for the lattice walls 53. In the simulation, as in the case of Fig. 10, the temperature at a location sufficiently distant from the UTC-PD (the InP substrate 59 and the end of the Au electrode) is kept at 30°C, and the UTC-PD acts as a heat source with a constant heat density (10 16 W / m 3 = 10 mW / μm 3 11 shows that the maximum temperature increase ΔT is 32°C near the center of the carrier collection layer 56, and the temperature decreases with increasing distance from the center. In particular, it is clear that heat is transferred to the metal layer via the lattice walls 53.

[0067] 12 shows the results of a simulation of heat dissipation characteristics (Comparative Example A). FIG. 12 shows the temperature distribution when the second simulation model is used. In the simulation, as in the case of FIG. 10, the temperature at a location sufficiently distant from the UTC-PD (the InP substrate 68 and the edge of the Au electrode) is kept at 30°C, and the UTC-PD acts as a heat source with a constant heat density (10 16 W / m 3 = 10 mW / μm 3 ) was assumed to dissipate heat. Looking at the results in Figure 12, it can be seen that the maximum temperature rise ΔT was 59°C near the center of the carrier collection layer 65, and that the UTC-PD and HEMT were generally at high temperatures. It can also be seen that the temperature difference between the metal layer and the UTC-PD was large, and the heat from the UTC-PD was not sufficiently transferred to the metal layer, making it difficult for the heat to be dissipated to the outside.

[0068] The above simulation results are summarized in Table 6 below. Example 1 corresponds to the results in FIG. 10, Example 2 corresponds to the results in FIG. 11, and Comparative Example A corresponds to the results in FIG. 12. As is clear from Table 6 below, there is a significant difference in heat dissipation performance depending on whether or not the lattice wall 53 is present. In Example 1, the temperature rise is suppressed to about 45% of that in Comparative Example A. Similarly, in Example 2, the temperature rise is suppressed to about 54% of that in Comparative Example A. In this way, high heat dissipation performance is achieved by applying the technology of this embodiment.

[0069]

[0070] For reference, the results of a simulation performed assuming the configuration described in Non-Patent Document 1 are shown in Table 7 below. This simulation assumes a heterostructure in which InP (p-type doped layer), InGaAs (neutral layer), and InP (n-type doped layer) are stacked on an Au mirror. The simulation results were compared for Comparative Example B, a model without an insulator lattice, Comparative Example C, a model in which a BCB insulator lattice is embedded within the p-type doped layer, and Comparative Example D, a model in which an AlN insulator lattice is embedded within the p-type doped layer. The conditions were the same except for the model structure. As shown in Table 7, the maximum temperature rise was large for all models, and the difference between the models was small. This result indicates that the low GR structure (i.e., the embedded insulator lattice) contributes little to suppressing temperature rise. Note that the maximum temperature rise is larger when BCB is used than when no GR is used, which is thought to be due to the reduced heat dissipation efficiency caused by the presence of BCB, which has particularly low thermal conductivity.

[0071]

[0072] It should be noted that the excellent heat dissipation characteristics of Examples 1 and 2 described above can be achieved while maintaining the excellent absorbance characteristics obtained by applying this embodiment. As an example, a range of w1 and w2 where the absorbance is 0.5 (i.e., the light absorption rate is 50%) or more was identified from the simulation results of the absorbance characteristics shown in FIG. 6 , and the simulation results of the absorbance characteristics and heat dissipation characteristics within that range are shown in FIG. 13 . The simulation conditions are the same as those of Example 2 described above (where the GR material is AlN). Graph (a) in FIG. 13 shows the absorbance characteristics, and graph (b) shows the characteristics of the maximum temperature rise ΔT.

[0073] As shown in FIG. 13 , the maximum temperature rise ΔT decreases as the active region width w2 decreases (i.e., as the grating wall width w1 increases). Under these simulation conditions, for example, the absorbance is 0.56 (i.e., the light absorption rate is 56%) or greater when the active region width w2 is in the range of approximately 285 nm to approximately 340 nm. Even when the active region width w2 is approximately 340 nm, the maximum temperature rise ΔT is 38°C, significantly lower than the maximum temperature rise values ​​of Comparative Examples A to D described above. Furthermore, when the active region width w2 is approximately 285 nm, the maximum temperature rise ΔT drops to 29°C. Similar results are obtained under the conditions of Example 1, where the GR material is SiN. Thus, by applying this embodiment, it is possible to improve heat dissipation characteristics while maintaining a constant light absorption rate.

[0074] (Simulation of response characteristics) Next, the results of a simulation of the response characteristics will be described. As already mentioned, there is a correspondence between the response speed and cutoff frequency of a PD, and the higher the response speed, the higher the cutoff frequency. Furthermore, a higher cutoff frequency means a wider response frequency band. The cutoff frequency of a UTC-PD is defined as the frequency at which the output intensity drops to half compared to the low-frequency limit. Below, we will describe a model for calculating the cutoff frequency (hereinafter referred to as the cutoff frequency calculation model), and then explain the results of a simulation of the response characteristics using this model.

[0075] UTC-PD output intensity P out is expressed by the following formula (1). In formula (1), I ph represents the DC photocurrent, and τ RC represents the RC time constant. abs represents the electron transit time in the light absorption layer, and τ cc represents the electron transit time in the carrier collection layer.

[0076] …(1)

[0077] From the above equation (1) and the definition of the cutoff frequency, the cutoff frequency f T is expressed as the following equation (2): As can be seen from the following equation (2), the cutoff frequency fT is the sum of the electron transit times in the light absorption layer and the carrier collection layer (τ abs +τ cc ) and the RC time constant τ RC The electron transit time τ abs and τ cc The RC time constant τ depends on the thickness of the light absorption layer and the carrier collection layer. RC is the area A of UTC-PD UTC-PD Therefore, according to the following equation (2), the area A of the UTC-PD UTC-PD The smaller the cutoff frequency f T becomes higher.

[0078] …(2)

[0079] RC time constant τ RC can be calculated using the model shown in Figure 14. In Figure 14, R L is the load resistance, and R j is the junction resistance, and C j is the junction capacitance. Using this model, the RC time constant τ RC is expressed as the following equation (3).

[0080] …(3)

[0081] Junction resistance R j is determined by the resistance of the light absorption layer (p-type doped InGaAs) and is expressed as the following equation (4): In equation (4), e is the elementary charge (e = 1.6 × 10 -19 [C]), and μ abs is the electron mobility in the light absorption layer (100 [cm 2 / Vs]), and p abs is the average hole density (3 × 10) in the light absorption layer. 8 [1 / cm 3 ]).

[0082] …(4)

[0083] Junction capacitance C jis determined by the parallel plate capacitance of the carrier collection layer (InP) sandwiched between the light absorption layer (p-type doped InGaAs) and the contact layer (n-type doped InGaAs), and is expressed by the following equation (5). cc is the relative dielectric constant (12.5) of the carrier collection layer (InP), and ε is the dielectric constant of vacuum (8.85×10 -12 [F / m]).

[0084] …(5)

[0085] Based on the above cutoff frequency calculation model, the parameter τ abs = 0.2 [ps], τ cc = 0.2 [ps], the thickness of the light absorption layer d abs = 50 [nm], the thickness of the carrier collection layer d cc =100 [nm], R L The simulation was performed with a setting of Ω = 50 [Ω]. In this simulation, the structural design values ​​for the UTC-PD were lattice wall widths w1 = w3 = 190 [nm] and active region widths w2 = w4 = 280 [nm]. For the one-dimensional GMR structure (see Figures 2 and 3), the number of lattice walls was set to 4 and the number of active regions was set to 3. For the two-dimensional GMR structure (see Figures 2 and 4), the number of active regions was set to 6. In this case, the ratio of the total area of ​​all active regions to the area of ​​the UTC-PD (filling factor) was 59.6% for the one-dimensional GMR structure and 35.5% for the two-dimensional GMR structure. The results of the simulation based on these conditions are shown in Figure 15.

[0086] FIG. 15 shows three simulation results C1, C2, and C3. C1 is a simulation result for a two-dimensional GMR structure (see FIGS. 2 and 4). C2 is a simulation result for a one-dimensional GMR structure (see FIGS. 2 and 3). C3 is a simulation result for a UTC-PD without GR (see FIG. 8) or with low GR (see FIG. 9). As already explained with reference to FIG. 9, in a low GR structure, the presence or absence of an insulator lattice does not divide the stacked structure, so in this case, the presence or absence of an insulator lattice does not affect the cutoff frequency. Therefore, the same simulation result C3 is obtained for the no GR and low GR structures.

[0087] 15, it can be seen that when the area of ​​the UTC-PD is the same, the cutoff frequency of C2 is higher than that of C3, and the cutoff frequency of C1 is higher than that of C2. For example, when the area of ​​the UTC-PD is 10 μm 2 In this case, the cutoff frequency of the 2D GMR structure reaches 354 GHz, which is an improvement of over 100 GHz compared to UTC-PDs without GR or with low GR. Similarly, the cutoff frequency of the 1D GMR structure reaches 309 GHz, which is an improvement of over 60 GHz compared to UTC-PDs without GR or with low GR.

[0088] From the above results, by applying the technology according to this embodiment, not only can the photoelectric conversion efficiency and heat dissipation efficiency be improved, but also the cutoff frequency can be significantly improved, i.e., the response speed and the response frequency band can be significantly expanded. The effect of improving the cutoff frequency can be obtained even if the material of the lattice wall has low thermal conductivity. Therefore, in cases where heat dissipation performance is not important, such as when the intensity of the input light wave is relatively low, a PD with excellent response characteristics can be realized even if a material with relatively low thermal conductivity is used for the lattice wall. Such a configuration also falls within the technical scope of this embodiment.

[0089] (Frequency Down-Conversion Device) Next, a frequency down-conversion device 1100 will be described with reference to Fig. 16. The frequency down-conversion device 1100 is an example of a frequency down-conversion device according to this embodiment. As shown in Fig. 16, an optical data signal S1 having a carrier frequency of f1 and an optical subcarrier signal S2 having a frequency f2 are input to the frequency down-conversion device 1100. In addition, an LO (Local Oscillation) signal S4 having a frequency f4 is input to the frequency down-conversion device 1100 from a local oscillator (not shown).

[0090] In the frequency down-conversion device 1100, the optical data signal S1 and the optical subcarrier signal S2 are mixed together, and a difference frequency Δf between the frequencies f1 and f2 is generated. opt (Δf opt =|f1-f2|) is generated (optical mixing 1101). Furthermore, in the frequency down-conversion device 1100, the optical mixed signal S3 and the LO signal S4 are mixed, and a signal component having a frequency Δf opt and f4 difference frequency f5 (f5 = |Δf opt -f4|) is generated (RF mixing 1102). In the frequency down-conversion device 1100, this signal component is read out from the drain electrode or gate electrode as the IF / BB signal S5.

[0091] Taking the photonic-electronic fusion transistor 20 shown in Figure 5 as an example, optical mixing 1101 occurs inside the UTC-PD 22, mixing the optical data signal S1 and the optical subcarrier signal S2. Then, RF mixing 1102 occurs inside the HEMT 24, mixing the LO signal S4 input from the third electrode 26 with the optical mixed signal S3 supplied from the UTC-PD 22 to the HEMT 24. After that, the IF / BB signal S5 obtained by the RF mixing 1102 is read out from the second electrode 23. For a gate readout method in which the IF / BB signal S5 is read out from the gate electrode, see Japanese Patent No. 7440988.

[0092] The frequencies f1 and f2 are in different infrared optical frequency bands. The frequency f4 of the LO signal S4 is the difference frequency Δf between the frequencies f1 and f2. optThe difference frequency Δf is set to a frequency in the sub-terahertz to terahertz band, which is close to opt and frequency f4 |Δf opt Since the frequency corresponding to -f4| becomes the frequency f5 of the IF / BB signal S5, the frequency f4 is set so that the frequency f5 becomes the desired frequency (intermediate frequency or baseband frequency). By setting it in this way, it is possible to convert an optical data signal having a frequency in the infrared optical frequency band into an IF / BB signal having an intermediate frequency or baseband frequency.

[0093] (Communication Device) Next, a communication device 1200 will be described with reference to Fig. 17 . The communication device 1200 is an example of a communication device according to this embodiment. As shown in Fig. 17 , the communication device 1200 includes an optical processing unit 1201, a frequency down-conversion device 1202, a local oscillator 1203, and a radio processing unit 1204. The frequency down-conversion device 1202 is the frequency down-conversion device 1100 shown in Fig. 16 .

[0094] The optical processing unit 1201 receives an optical data signal and an optical subcarrier signal transmitted through an optical fiber, and inputs them to the frequency down-conversion device 1202. The frequency down-conversion device 1202 receives an LO signal from a local oscillator 1203 as input.

[0095] In the frequency down-conversion device 1202, an optical mixed signal is generated by optically mixing the optical data signal and the optical subcarrier signal, and an IF / BB signal is generated by RF mixing of the optical mixed signal with the LO signal. The IF / BB signal is read out from the drain electrode or gate electrode of the frequency down-conversion device 1202 and output to the radio processing unit 1204. The radio processing unit 1204 generates a radio signal from the IF / BB signal and transmits the generated radio signal via the antenna 1205.

[0096] The communication device 1200 may further include a control device (not shown) for controlling the operations of the optical processing unit 1201, the frequency down-conversion device 1202, the local oscillator 1203, and the radio processing unit 1204. This control device has a processor and a memory. The processor may be a central processing unit (CPU), a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or the like. The memory may be a storage medium such as a read only memory (ROM), a random access memory (RAM), a hard disk drive (HDD), a solid state drive (SSD), or a flash memory.

[0097] While the preferred embodiments of the present application have been described above with reference to the accompanying drawings, the present application is not limited to these examples. It is clear that a person skilled in the art can conceive of various modifications or alterations within the scope of the claims, and these modifications also naturally fall within the technical scope of the present application.

[0098] 10 Optical semiconductor device 11, 21 Metal layer (first electrode) 12, 12' Photodiode 13, 23 Second electrode 14 Semiconductor contact layer 15 Semiconductor substrate 22 UTC-PD 24 HEMT 25 InP substrate 26 Third electrode 121a, 121b, 121c, 121d, 121e Lattice wall 122a, 122b, 122c, 122d P-type contact layer 123a, 123b, 123c, 123d Diffusion prevention layer 124a, 124b, 124c, 124d Light absorption layer 125a, 125b, 125c, 125d Carrier collection layer 1100, 1202 Frequency down-conversion device 1200 Communication device 1201 Light processing section 1203 Local oscillator 1204 Radio processing unit 1205 Antenna

Claims

1. A photodiode sandwiched between a metal layer and a semiconductor contact layer, comprising: a laminated structure including a light-absorbing layer that absorbs light and generates photoelectrons; and a plurality of periodically arranged lattice walls, the laminated structure being divided into a plurality of active regions by the lattice walls, each lattice wall being made of an insulator having a refractive index different from that of the plurality of active regions.

2. The photodiode according to claim 1, wherein one end of each lattice wall contacts the metal layer and the other end of each lattice wall contacts the semiconductor contact layer.

3. The photodiode according to claim 1, wherein the insulator has a higher thermal conductivity than the light absorption layer.

4. The photodiode according to claim 3, wherein the insulator is SiN or AlN.

5. The photodiode according to claim 1, wherein the plurality of grating walls include a plurality of first grating walls arranged at equal intervals in a first direction, and one or more second grating walls arranged at equal intervals in a second direction different from the first direction.

6. A lightwave detector having a structure in which the photodiode according to claim 1 and a high electron mobility transistor (HEMT) are monolithically integrated, wherein the HEMT is used as the semiconductor contact layer, the photodiode is sandwiched between a source electrode and the HEMT, and the lightwave detector is configured to read out a photovoltage signal generated by a lightwave incident on the photodiode from a gate electrode or a drain electrode arranged on the HEMT.

7. A frequency down-conversion device having a structure in which the photodiode according to claim 1 and a high electron mobility transistor (HEMT) are monolithically integrated, wherein the HEMT is used as the semiconductor contact layer, the photodiode is sandwiched between a source electrode and the HEMT, and the frequency down-conversion device is configured to output a data signal having a frequency in the radio wave band or baseband from a gate electrode or a drain electrode arranged on the HEMT in response to input of an optical data signal and an optical subcarrier signal having different frequencies in the infrared optical frequency band.

8. A communications device comprising: a frequency down-conversion device according to claim 7; an optical processing unit that inputs optical data signals and optical subcarrier signals transmitted via optical fiber to said frequency down-conversion device; and a radio processing unit that converts the data signals output from said frequency down-conversion device into radio signals and transmits said radio signals via an antenna.

Images