JPH03109788A - Superconducting semiconductor optical element - Google Patents

Abstract

PURPOSE:To realize an optical element excellent in light receiving and emitting characteristics by utilizing a phenomenon that a semiconductor behaves like a superconductor due to a proximity effect of the superconductor. CONSTITUTION:A structure having a junction system of a superconductor region and a semiconductor region wherein the junction system soaks a Cooper pair in the superconductor region into the semiconductor region by using a proximity effect and a structure for injecting the cooper pair into the semiconductor region are provided. Within soaked length xsiS and injected length xsi1, a conductor part of the semiconductor region is transferred to be a superconductor, wherein a junction structure is provided and a take-out electrode is provided. Thus a light emitting device can be formed by using the semiconductor as well as an effect due to its being the superconductor can be realized.

Description

[Detailed description of the invention]

[-Industrial Application Field] The present invention relates to an optical device having a light emitting and/or light receiving function, and particularly to a superconducting semiconductor optical device suitable for use in optical fiber communication systems, optical information processing systems, etc.

[Conventional technology]

Conventional light emitting diodes pass current through the pn junction of a semiconductor in the forward direction, injecting holes into the valence band and electrons into the conduction band of the semiconductor, and recombining them directly or through a luminescent center to emit light. It is radiating. Therefore, the energy of the light emitted from the semiconductor light emitting diode is approximately equal to or lower than the band gap (Eg) of the semiconductor material of the active layer region. This situation is the same in the case of semiconductor lasers. In a semiconductor laser, a current is passed in the forward direction through the pn junction of a direct transition semiconductor, injecting holes into the valence band and electrons into the conduction band of the semiconductor, forming population inversion, and converting their recombined light into light. This is amplified as stimulated emission light within the resonator, leading to laser oscillation. Therefore, the energy of light emitted from the semiconductor laser is limited to about the band gap (Eg) of the semiconductor material of the active layer region. In conventional lasers using ■-■ group semiconductor materials, the wavelength can be selected to a certain extent by selecting the semiconductor materials and controlling the bandgap by mixing them, but the energy of the emitted light is between 0.3 and 2. It was about .5 eV. Next, some problems with conventional light emitting and light receiving elements will be listed. The first one concerns laser light with high optical energy. Conventionally, the following two methods have been attempted in order to realize higher energy laser light. One method is to use a semiconductor material with a larger band gap. Specifically, these materials include I
Examples include I-VI group semiconductors (CdTe, Zn5e-ZnS, etc.) and group II raw conductors (SiC, diamond, etc.). Although laser oscillation by electron beam excitation has been confirmed with some of these materials, it is not practical because it requires a high voltage source for electron beam generation and a vacuum atmosphere. A process for obtaining crystals of a quality suitable for use in semiconductor light emitting diodes and semiconductor lasers has not yet been developed, and a doping method has not yet been established. Although there are examples of creating light emitting diodes using these materials. Luminous efficiency is low and impractical. (For example, Shigeo Fujita,
(See Applied Physics Vol. 54, pp. 40-47, 1985). The other method is to irradiate a nonlinear optical crystal with light from a semiconductor laser to generate harmonics due to the nonlinear optical effect. This method uses a plurality of components including a semiconductor laser and a nonlinear optical crystal, so it is not possible to miniaturize the device, and the harmonic generation efficiency is low, making it impossible to obtain large laser power. These have been reported in detail by Sasaki (see Takatomo Sasaki, Applied Physics Vol. 58, pp. 895-903, 1989). The second problem with conventional light-emitting/light-receiving elements is the reduction in threshold current for laser oscillation. In order to efficiently cause laser oscillation with a low threshold current and to narrow the spectral width of the emission wavelength, it is generally necessary to increase the shape of the carrier energy distribution near the band gap. For example, in conventional semiconductor lasers, a quantum well structure is used instead of a semiconductor bulk crystal as an active layer to achieve lower threshold current and spectral width. First, this principle will be explained. We will discuss the distribution of electrons and holes when the same amount of electrons are excited in the bulk case and in the case of a quantum well structure. Figure 2 (,
) and (b) show the dependence of the density of states (horizontal axis) on the energy (vertical axis) of a bulk semiconductor and a quantum well structure semiconductor, respectively. 101 shows the density of states in the conduction band, 102 shows the density of states in the valence band, 103 shows the energy distribution of electrons, and 104 shows the energy distribution of holes. Thanks to the change in structure from the bulk to the quantum well structure, the energy distribution of electrons and holes changes, and the quantum well structure has a larger distribution near the band gap. This explains why the threshold for oscillation of a semiconductor laser using a quantum well structure is low. These have been reported in detail by Okamoto et al., including actual examples (for example, see Okamoto et al., Applied Physics Vol. 52, pp. 843-851 (1983)). Another known method for changing the energy dependence of the density of states is to utilize quantization of the cyclotron motion of carriers in a strong magnetic field (for example, in Japanese Patent Laid-Open No. 6
4-68994, Okamoto et al., Applied Physics Vol. 52, 84
3-851 (1983)). The density of states of the bulk semiconductor shown in Figure 2(a) is as shown in Figure 2(c) in a strong magnetic field.
It splits into Landau levels as shown in . In this case as well, the density of states rapidly rises at the band edge, and an effect similar to that of the quantum well structure has been revealed. however,
It must coexist with a magnetic field generation source such as a superconducting magnet, which is not realistic when considering the obstacles to miniaturization of the device and the adverse effects of the magnetic field on other devices. In the above examples in which the density of states distribution is changed by applying a quantum structure or a magnetic field, although the energy dependence of the density of states at the band edge becomes steep, the carrier density on the high energy side determined by the injected carrier concentration It is not possible to make the distribution steep. As an example with fundamentally different properties from the examples listed above, a laser that utilizes light emission from a pose condensed system of excitons has been proposed (for example, Japanese Patent Application Laid-Open No. 61-218192).
. This principle will be briefly explained. Electrons and holes are accumulated two-dimensionally on both sides of a tunnel barrier layer formed in a Peter junction semiconductor. The band diagram and carrier distribution state in this state are shown in FIG. 2(d). Wide bandgap half-4 body 111,1
A narrow bandgap semiconductor 113,1 surrounded by 12
Electrons 116 and holes 117, which exist in the barrier layer 115 in the barrier layer 115, form excitons through Coulomb interaction. When the thickness of the tunnel barrier layer 115 is optimized, electrons and holes tunnel through the barrier layer and are recombined by light emission. The proposal is that laser oscillation becomes possible by stimulating and amplifying this light emission using an optical resonator. If the exciton system undergoes pause condensation at low temperatures, a superradiant laser will be realized. In this laser, the optical spectral width is 1/10 and the threshold current for laser oscillation is 1/10 that of conventional semiconductor lasers.
It becomes 10,000. However, in this superradiant laser, the electrons and holes to be recombined are spatially separated, and the recombination probability is controlled by the overlap of wave derivatives in the tunnel barrier layer. I can't make it bigger. Furthermore, more essentially this exciton system cannot be made highly concentrated and does not undergo pause condensation at practical temperatures (temperatures of approximately 10°C or higher). Similarly, lasers using superconductors as pose-condensed electron systems are known. There are two operating principles for this type of laser. One method involves injecting electron quasiparticles and hole quasiparticles into a superconductor with a gap of Δ, and emitting 2Δ light when they recombine to cause laser oscillation. 302581, Japanese Patent Publication No. 6
3-263783. ). In this invention,
In superconductors, Cooper pairs are broken from the ground state 2
It is based on the idea of population inversion of quasiparticles with higher energy by Δ. However, it is difficult in principle to invert the ground state and quasiparticle state by injecting electrons and holes while maintaining the superconducting state.
Even if a laser oscillates, the gap energy of currently known superconductors is from several meV to 50 meV or less at most, so the problem arises that the light emitted from the laser is limited to microwaves to infrared rays. In lasers with other operating principles, a Josephson junction is formed by laminating a superconductor and a transparent insulating film.
Applying a potential difference δV between superconductors separated by an insulating film,
There are some that emit photons with an energy of 2qδ■ (for example, JP-A-63-316493.63-
262882. ). However, in the superconductor-body-superconductor structure, the wave derivatives of the Cooper pair cannot be superimposed at the same location in the insulator, and in reality only extremely weak light emission can be obtained. In addition, ultra-thin films (
1-3 nm) It is extremely difficult to uniformly form an insulating layer with a high-quality single crystal, and non-emissive transitions via traps in the insulating layer become dominant, making it impossible to put it into practical use as a light-emitting device. Not yet. The third problem is that the relaxation oscillation frequency (f,) is low. A large f is essential for ultra-high-speed optical communications, but conventional modulation-doped multiple quantum well lasers that operate at high speed (for example, Kazuhisa Uomi Applied Physics Vol. 57, 708-
713 (1988) '), the frequency is 40 GHz, and a normal DH laser has a frequency of 10 GHz. Furthermore, there has been no experimental demonstration of the process of emitting and receiving light from a coherent electronic system, and there are many unknowns regarding the basic process of emitting and receiving light.

[Problem to be solved by the invention]

The problems that the present invention aims to solve are: 1. Efficiently emit and receive light with energy higher than the band gap, which cannot be achieved with conventional semiconductor and superconductor light-emitting/light-receiving devices; 2. Even when using a quantum well structure. 3. A semiconductor optical device having a relaxation oscillation frequency f approximately exceeding 100 GHz; 4. Semiconductor light-emitting device characterized by a generation and reception process that reflects the coherence factor φ possessed by a coherent electronic system (corresponding to one phase φ in the case of a superconductor); 5. Light-receiving device with an extremely wide band The purpose of the present invention is to realize the following using a light-emitting and light-receiving element based on a new principle that utilizes the proximity effect of superconductors/semiconductors as described below.

[Means to solve the problem]

Recently, research on the proximity effect between superconductors and semiconductors has progressed, and the mechanism by which Cooper pairs seep into semiconductors is becoming clear (for example, Solid State Physics Vol. 1.23 No. 3 pp1
53-162 1988; superconducting proximity effect; see Tanaka, Tsukada and references therein). However, these research results are related to the proximity effect between a metal superconductor and a semiconductor, and no research has been reported on the proximity effect in a high-temperature superconductor/semiconductor junction where the transition temperature Tc is several tens of degrees. The inventors have developed a high-temperature superconductor (with a transition temperature c of 90 K).
YBCO: Oxide high temperature superconductor YBa2Cu3
The abbreviation 0t is hereinafter written as YBCO. )/Au(5
0nm) /M o (10nm) /W (40nm
) / n-type semiconductor; GaAs (or p-type semiconductor; G
We measured the distance ξS of the Cooper pair at the aAs junction to the semiconductor side and the superconducting gap Δ1 in the semiconductor, and obtained the following results. The experimental results are shown in Table 1. To summarize the results, Table 1 1. The seepage distance ξS is from 300 nm to 500 nm, 2
．． The average size Δav of the superconducting gap Δ□ of the oozing Cooper pair is 20 meV to 30 meV. Here, the reason why a buffer metal (A u / M o / W) is inserted between the high temperature superconductor and GaAs is that the oxide superconductor is placed on top of GaAs in an oxygen atmosphere. Au and GaAs are oxidized, and the insulating film is sandwiched, As is removed, and the oxide superconductor does not exhibit superconductivity, so Au is difficult to oxidize.
An oxide superconductor was formed on top of the oxide superconductor. However, as is well known, Au cannot be stably deposited directly onto GaAs. Also, A, u and Ga
As starts to react at 470°C, Au diffuses into GaAs. Furthermore, in the process of depositing YBCO, 40
To set the substrate temperature from 0'C to 550°C, Ga
Furthermore, in order to prevent As from reacting with the buffer metal,
In order to prevent the diffusion of u, W, a highly heat-resistant metal that does not easily react with Au, was used as the buffer metal. Furthermore, on W, A,
It was difficult to form +i with good adhesiveness, so Mo was inserted in order to improve the adhesiveness. The reason for inserting such a buffer metal is that the high temperature superconductor is
This is because we want to make ohmic contact with the (semiconductor) metal superconductors N b , P b , N b , G
It goes without saying that a buffer metal is not required in cases such as e, Nb3Sn, etc. The inventors further advanced this idea and calculated the leakage distance ξS of the Cooper pair toward the semiconductor side at the high-temperature superconductor/buffer metal/semiconductor junction and the superconducting gap Δ in the semiconductor.
, was measured systematically. FIG. 4 shows experimental results for n-type GaAs and n-type GaAs at a plurality of doping levels. However, the specific contact resistance ρC is the specific contact resistance at the interface between the buffer metal and the semiconductor, and a semiconductor layer with a thickness of 30 nm and a varying concentration is inserted. In the figure, No and NA are the donor and acceptor concentrations of the oozing semiconductor layer, respectively. No= 101g/cm, NA: 10"/cm'
FIG. 5 shows the investigation of temperature changes in the seepage distance ξS and the superconducting gap Δ in the semiconductor for S'n, Be-doped GaAs, respectively. In FIGS. 4 and 5, the solid line indicates n-type GaAs, and the broken line indicates n-type GaAs. The following general conclusions were drawn from the measurement results. 1. The seepage distance ξS can be controlled from Onm to a maximum of 700n+n. 2. The seepage distance ξS depends very sensitively on the doping level. 3. The seepage distance ξS does not depend much on the specific contact resistance ρC at the interface between the buffer metal and the semiconductor. 4. The average size Δav of the superconducting gap Δ1 of the oozing Cooper pair largely depends on the specific contact resistance ρC and the doping level. In particular, there are large restrictions on ρC, which represents the degree of ohmic characteristics. 5. The average size Δav of the oozing distance ξS and the superconducting gap Δ of the oozing Cooper pair is Tc (in this case,
At temperatures below half of 90K), there is no significant change. 6. If the semiconductor is not degenerate, the seepage distance ξS is 2
It becomes less than nm. Next, the physical image of this close system will be explained. FIG. 1(a) shows a Cooper vs. potency diagram at the interface. Let ξS be the distance that Cooper pairs seep into the semiconductor side through the bulk Cooper pair potential depletion layer in the superconductor, and let Δav be the average Cooper pair size.
shall be. In this way, this junction system can be divided into a Cooper pair supply layer and a superconducting semiconductor layer. When the semiconductor is an n-type semiconductor, an energy band diagram near the conduction band of a degenerate semiconductor that has become superconducting and the corresponding state density are shown in FIG. 1(b). Fermi level EF is located at a position qVc from the conduction band edge Ec, and a superconducting gap 2Δ is opened. In the figure, the shaded area from the conduction band edge Ec is a condensed region of Cooper pairs. In the conceptual diagram for the energy E of the corresponding density of states, there is no state between the superconducting gaps 2Δ, and the originally existing states are distributed above and below it. When the semiconductor is p-type, the valence band of the degenerate semiconductor that has become superconducting
ceBand) and the corresponding density of states are shown in FIG. 1(c). Fermi level Ep is located at a position qVv from the valence band edge Ev, and a superconducting gap 2Δ is opened. In the figure, the shaded area from the valence band edge Ev is a condensed region of Cooper pairs. In the conceptual diagram for the energy E of the corresponding density of states, there are no states between the superconducting gaps 2Δ, and the originally existing states are distributed above and below it. The seepage of Cooper pairs into a P-type semiconductor is intuitively understandable, but if we consider that electrons are packed up to the Fermi Level and there is a vacant space for electrons from the valence band edge Ev to the Fermi Level. ,
easy to understand. Moreover, such a picture is similar to that of ordinary metal superconductors (Nb and Pb).
) and a degenerate semiconductor such as n(p) type GaAs or Si, the physical picture is the same, only the value of the superconducting gap is different. From these experimental facts, in a superconductor/semiconductor junction, the semiconductor itself changes into a superconductor with a superconducting gap Δ within the distance ξS where the Cooper pair penetrates toward the semiconductor side, and the superconductor becomes a superconductor. It can be considered as a Cooper pair supply layer to semiconductors. Furthermore, the inventors have obtained the following knowledge regarding the injection of Cooper pairs. A buffer metal layer and a superconductor film similar to those described above were formed on a GaAs nip junction and pin junction shrimp wafer, and an experiment was conducted in which the Cooper couple that seeped into the n or pi was further injected into the i layer. went. An example in which Cooper pairs are implanted from the n-layer will be described below. A band diagram in a state of thermal equilibrium is shown in FIG. 1(d). Ni from buffer metal 2/n-GaAs3 interface
The distance to the depletion layer at the interface is sufficiently shorter than ξS. That is, all of the n-type GaAs 3 is superconducting when no bias is applied. That is. Cooper pairs are ubiquitous in n-type GaAs. A band diagram when a forward bias is applied to this nip junction is shown in FIG. 1(e). The injection distance ξI of the Cooper pair into the i-layer up to the forward bias voltage Va at which the band becomes flat is ξ1=ξ5oexP ('1 (V Va)/k
It was found that it is expressed as T). In this way, with forward bias up to Va, the density of states and carrier distribution in the region where Cooper pairs seep out in the i layer are shown in Figure 1(f).
Shown below. A hole pseudo-Fermi level (Erh) determined by the hole density injected from the pm is formed in the valence band, and an electron pseudo-Fermi level (Ere) determined by the electron density injected from the n-layer is formed in the conduction band. be done. However, near the pseudo-Fermi level of electrons, the i-layer semiconductor becomes a superconductor due to the leakage of Cooper pairs, and the gap 2Δ2 opens.
Levels that should originally be within the gap are distributed above and below the gap. FIG. 5 shows the results of examining the dependence of the implantation distance into the IF on the contact resistance ρC and the doping level No. of the average gap size Δ2av. The following conclusions were drawn from the measurement results. 1. The implantation distance ξI can be controlled from Onm to a maximum of 1100n. 2. The implantation distance ξ■ depends very sensitively on the doping level. 3. The implantation distance ξI does not depend very much on the specific contact resistance ρC at the interface between the buffer metal and the semiconductor. 4. The average size Δ2av of the superconducting gap Δ2 of the implanted Cooper pair strongly depends on the specific contact resistance ρC and the doping level. In particular, ρ, which represents the degree of ohmic characteristics,
C has major limitations. 5. The injection distance ξI and the average size Δ2aV of the superconducting gap Δ2 of the injected Cooper pairs are Tc (in this case,
At temperatures below half of 90K), there is no significant change. Superconductor gap Δ. (T) is approximately Δ when T≦0.8Tc. (0) and Δ. Since (0) is about several times kTc, 2Δ is about the same as kT when T≦0.8Tc. In this state density distribution, the statistical distribution of quasiparticle electrons and quasiparticle holes follows the usual Fermi-Dirac statistics. However, when looking at the entire conduction band, the energy dependence of the density of electrons (including Cooper pairs) changes much more rapidly than in the case without Cooper pair speckles. In the method of changing the density of states by introducing a quantum structure or applying a strong magnetic field as shown in Figures 2(b) and 2(c), the density of states distribution on the band edge side of the semiconductor becomes sharp. It's a contrast. Next, we will discuss the relaxation process of the injected Cooper pair. The Cooper pair was stable at nil. When holes are injected into the conduction band of the i-layer, if the holes are injected into the valence band of the i-layer, the electrons forming Cooper pairs and the holes recombine and relax. . There are three types of recombination process: 1. A Cooper pair recombines with two holes injected into the valence band at the same time. 2. One of the electrons forming the Cooper pair recombines with one hole injected into the valence band. 3. A thermally resolved Cooper pair, that is, a quasiparticle electron, recombines with a hole injected into the valence band. The recombination process includes radiative recombination and non-radiative recombination.
Here we consider only radiative recombination. In the first transition process, one photon may be emitted, or two photons with the same energy may be emitted. When a Cooper pair and two holes recombine by emitting one photon, the energy of the photon is approximately 2 of the band gap Eg.
Double. In other cases, the energy of the photon is comparable to Eg. However, the light emitted in step 2 has an energy lower by 2Δ2 than the light that is numerically compared in step 3, reflecting the bound state of the Cooper pair. In the above example, the Cooper pair supply layer is provided only on one side of the semiconductor nip junction, but in the next example, a case will be described in which the Cooper pair supply layer is provided on both sides of the nip junction. Similarly to the above, if the distance from the superconductor to the i-layer is shorter than ξS, then superconductor / n - Ga
As/i-Ga As/p-Ga
The band diagram of the structure represented by A s /superconductor in a forward bias state is shown in FIG. 1(g). Electrons are injected from nl into the conduction band of the i-layer, and Cooper pairs are also injected accordingly. Similarly, holes are injected from pm into the valence band of the i-layer, and Cooper pairs are also injected along with the holes. The state density and carrier distribution state of the i-layer in the Cooper pair injection state are shown in the 11th m (h). There are gaps Δ2 and Δ at the positions of the electron pseudo-Fermi level and the hole pseudo-Fermi level, respectively.
opens. Next, we will discuss the relaxation process of the injected Cooper pair. As in the previous example, the Cooper pair, which was stable in the n layer, is
When holes are injected into the conduction band of the layer, if holes are injected into the valence band at the same position in the i-layer, the electrons and holes forming a Cooper pair recombine and relax. There are three types of recombination process: 1. A hole injected into the valence band by a Cooper pair in the conduction band 2
recombine at the same time. 2. One of the electrons forming a Cooper pair in the conduction band recombines with one hole injected into the valence band or with a quasiparticle hole in the valence band. 3. Quasiparticle electrons in the conduction band recombine with holes injected into the valence band or quasiparticle holes in the valence band. In the case of this structure as well, when a Cooper pair and two holes emit one photon and recombine, the energy of the photon becomes approximately twice the band gap Eg. However, there are some differences from the structure described above. Reflecting the injection of Cooper pairs into the valence band, the hole state density existing in the valence band is shown in Figure 1 (
h), resulting in a state density distribution unique to superconductors. The forward bias current of this element oscillates for one constant voltage. Along with this, the emission intensity also oscillates with the same period and the same phase. Forward bias voltage■, current J, emission intensity!
Similar to the AC Josephson effect, the relationship can be expressed as J"Jcsin(φC-φh) 2qV=6d(φC-φh)/dtIcI:J. Here, φC and φh are respectively,
This is the phase of Cooper pairs that seep into the conduction band and valence band. The value is the blank constant divided by 2π, and q is the elementary charge of the electron. When this element is driven with a constant voltage, the current oscillates at a frequency of νv=2qV/h. The energy of the light emitted during the elementary process is the band gap Eg of the single-layer semiconductor and 2Eg, which is twice that.If the frequencies are νEg and 2vEg, the emission intensity of the light is modulated at the frequency νV. In addition to the fundamental frequency, the effectively emitted light has the following frequencies: Eg-vl/, VEg+v, 2Eg-v. This is equivalent to containing light with a frequency of 2Eg+YV. The frequencies of these satellite lights can be changed in accordance with the Iasumi pressure. The new luminescence phenomenon discovered by the present invention is essentially different from the AC Josephson effect. While the AC Josephson effect is caused by a tunnel effect between two superconductors via an insulator barrier, the luminescence phenomenon discovered in this invention is caused by the proximity effect of n This is a new physical phenomenon of light emission between Cooper pairs in n-type and p-type semiconductors, in which a p-type semiconductor behaves as if the bandgap were an insulator through the band gap of the semiconductor. In particular, the superconducting phase difference φ of a semiconductor whose emission spectrum is made into a superconductor
It is essentially new in that optical and electrical vibrations combined with superconducting current appear, reflecting C-φh. Furthermore, an example in which a quantum structure is applied to the i-layer will be explained. When a single quantum well is introduced into the i-layer, the carrier distribution state in the injection state is as shown in FIG. 1(i). 1st
The difference from FIG. 1(g) is that the light emitting semiconductor layer 4 is as shown in FIG. 1(i).
The point here is that it is formed of a quantum well consisting of a semiconductor 402 with a small band gap sandwiched between semiconductors 401 and 403 with a large band gap. The density of states changes rapidly on the band edge side due to the quantum structure and on the quasi-Fermi level side due to the superconducting gap. The behavior of the injected Cooper pair in this structure is almost the same as in the previous case. From the perspective of applying this phenomenon to optical devices, the semiconductor itself can be regarded as a superconductor within the seepage distance ξS and the injection distance ξi, and various superconducting devices using semiconductors (i.e., This will pave the way for the realization of superconducting semiconductor optical devices. In particular, the seepage distance ξS can be increased from lnm to 700nm by carefully selecting the carrier concentration of the semiconductor.
Since the injection distance ξl can be controlled by the forward bias voltage, various crystal structures can be designed within the distance ξS of this Cooper pair to the semiconductor side, which cannot be achieved with conventional semiconductor optical devices. It is possible to realize a superconducting semiconductor optical device based on a new principle. Also,
All phenomena and technologies realized in conventional semiconductor optical devices can be applied. That is, by using this light emission phenomenon, light emitting diodes, laser diodes, and even light receiving elements can be realized, and various techniques that have been developed to improve the performance of these elements can all be applied. The device of the present invention described above is a superconducting semiconductor optical device that utilizes the proximity effect of a superconductor to transform a semiconductor that does not originally exhibit superconductivity into a superconducting semiconductor optical device that utilizes the fact that the semiconductor behaves like a superconductor. It is. Moreover, the present invention is disclosed in Japanese Patent Application Laid-Open No. 63-3025.
8 and JP-A-63-262882, the technology is fundamentally different from the technology disclosed in JP-A No. 63-262882, in which a semiconductor that does not exhibit superconductivity (at high temperatures) is partially converted to superconductivity using the above-mentioned proximity effect, thereby creating an optical device. (Unexamined Japanese Patent Publication No. 63-30)
258 and the invention of JP-A-63-262882 use the superconductor Josephson junction itself). In particular, the crystallinity of semiconductor crystals such as GaAs is different from that of other materials (
Compared to the above-mentioned degenerate semiconductor materials that exhibit superconductivity and the insulating materials that have been used for JJ devices up until now, it is significantly superior in terms of material quality (perfect crystallinity, no defects, no grain boundaries, etc.). ing. Furthermore, it is possible to make the semiconductor superconducting by interposing high-quality buffer metals (such as Au / Mo / W Si; which usually do not exhibit superconductivity, or even if they do, the temperature is extremely low) that has been used in GaAs LSI. Therefore, unlike conventional Josephson junction lasers, there is no reduction in luminous efficiency due to grain boundaries of the insulating layer material. Furthermore, it has the great advantage of being able to utilize many parts of the GaAs LSI process technology and device circuit technology that have been developed up to now for device fabrication. Another advantage is that it can be easily combined or integrated with existing GaAs (hetero) devices. Furthermore, the present invention does not need to be limited to GaAs-based semiconductors; direct transition type semiconductor InG can be used as a light emitting element.
aAs, InGaAsP, GaSb, InP', Zn5
It goes without saying that ordinary light emitting device semiconductor materials such as E, ZnS, and CdTe may be used. As the light receiving element, the present invention can be realized not only by direct transition type semiconductors but also by non-direct transition type semiconductors such as Si and Ge. (Effect 1) As described above, in the superconducting semiconductor optical device of the present invention, a light-emitting device is formed using a semiconductor (GaAs, etc.) with extremely good crystallinity, extremely high film thickness, and extremely high impurity concentration. Therefore, not only can all conventional high-performance technologies be applied, but also the effects of superconducting can be realized.Such effects are based on the following effects of the present invention: 1. Semiconductor conduction There is a process in which a Cooper pair injected into the band and two holes injected into the valence band simultaneously recombine, and the entire energy is released in one photon.At this time, the band gap is Light with an energy of 2Eg is emitted from a superconducting semiconductor of Eg. 2. In the operating state of a semiconductor (laser) diode, when Cooper pairs are injected on both sides of the conduction band and valence band, In the operating state, the injected current and emission intensity oscillate. This oscillation behaves similar to the AC Josephson effect. 3. When a Cooper pair is injected into the band of a semiconductor, a superconducting gap is created at the Fermi level. 4. Reflecting the steepness of the density of states, relaxation can be applied. The vibration frequency becomes very large. 5. Parts made of semiconductors can inherit all conventional semiconductor technology in principle. Therefore, any desired structure can be created by any desired method. Various. MBE of heterojunctions, quantum well structures, and modulation doped structures
Using crystal growth techniques that can control monoatomic layers, such as molecular beam epitaxy (Molecular Beam Epitaxy) and MOCVD (Metalorganic Pyrolysis), it is possible to form extremely uniform and accurate film thicknesses with good reproducibility. Furthermore, device technologies used in conventional semiconductor lasers, such as waveguide and resonator configurations, are also applicable. For example, a resonator configuration such as a distributed feedback (DFB) laser or a buried (BH) laser can be applied. 6. The superconducting semiconductor optical device of the present invention is an existing semiconductor device (GaAs MESFET, 2DEGFET (two-dimensional electron gas FET: HEMT)
, heterojunction FET, heterobipolar transistor, 2DEGHBT, etc.) or optical devices (semiconductor laser, LED, various light emitting and light receiving elements) and can be easily formed monolithically on the same substrate. Furthermore, 2DEGHBT (for example, T, Usagawa and 6 others: [EDM 87Tech
nical Digest pp78-81 or Japanese Patent Application No. 60-164128, Japanese Patent Application No. 60-164126,
Patent Application No. 61-40244, etc.) and resonant tunnel FETs (e.g., A. R. Bonnefoi and 2 others; IEEE
Trans, Electron Devices,
It is also possible to incorporate the superconducting semiconductor Josephson junction element of the present invention and an existing semiconductor device into the same element as in EDL-6 (1985) 636). Such integration with a semiconductor device is a 2DEGHBT, which improves device performance at low temperatures (e.g. liquid nitrogen temperature; low temperature where thermal fluctuations of the superconductor do not essentially contribute). Especially when combined with 2DEGFET, low-temperature bipolar, low-temperature Bi CMO3, etc., it contributes to improving the performance of such ICs and LSIs. [Example] Hereinafter, the present invention will be explained in more detail through Examples. Example 1 An example of a superconducting semiconductor light emitting diode is shown in FIG.
). p- containing 3 x 10"/cm3 of Zn
Using MBE (molecular beam epitaxy) on the aAs substrate 90, p+A containing Be I x 10"/cm3 was formed.
lxGa, -xAsJl 5 (x == 0, 15
: Usually selected in the range of 0.1≦X≦0.45) is 500 nm, and the undoped 1-GaAs layer 4 is 10
nm, n containing 3 x 10"/cm3 of Si
+GaAs layer 31 4OnLIl, Sn 9X
10"/cm" containing n+÷GaAs layer 32 is 1
After forming the SNM, an insulating film 40 such as SiO□ is formed, and an electrode pattern is formed using ordinary lithography and etching techniques. Typically, the n+GaAs layer 31 and the n++GaAs layer 32, depending strongly on the doping level,
The film thickness is selected to be in the range of 21Snm to 70nm. In addition, the film thickness of layer 4, which is the light emitting layer, is from Snm to 50nm.
Often selected within the range. Next, as a buffer metal, 30 nm of W54, 10 nm of Mo53,
80 nm of Au 52 was deposited using a sputtering method. Furthermore, YBCO film 1 was deposited at 40% by reactive vapor deposition.
A thickness of 0 nm was formed. In this case as well, the total thickness of the buffer metal is often selected in the range of 30 nm to 150 nm. In general, n-type GaAs is difficult to achieve non-alloy ohmic properties. Therefore, in order to further improve the ohmic contact, instead of the Sn-doped layer 32, for example, (T, N1t
tono and 3 othersJapanese Journal
of Applied Physics Vol. 27, N009, September, 1988.
As shown in pp. 1718-1722), n÷InyGa−yAs layers may be sandwiched in which the In composition y is graded. However, unlike the case of T and N1ttono, I
The thickness of the nGaAs film is usually selected to be about 15 nm. That is, the In composition is graded within this 15 nm film thickness. By doing this, the specific contact resistance ρC between the buffer metal and the semiconductor layer is 5 x 10+9 Ωcm
The film thickness can be designed more effectively so that the penetration depth of the Cooper pair reaches up to J-type GaAs. Next, electrical isolation between the electrodes and isolation between the elements is performed, and 30% of Au is used as the lead-out electrode 50 of the diode.
00 nm was formed. On the back side of the wafer, there is a ring-shaped ordinary Cr/Au ohmic contact for p-GaAs.
1 was formed, and a window for light extraction was created by etching. In this embodiment, p-type AlxGa1-xAs5 was used in order to easily open a window in the p-type GaAs substrate 90 by selective etching and to prevent the emitted light from being absorbed. The emission spectrum of the superconducting semiconductor light emitting diode of the present invention when operated at liquid nitrogen temperature (77K) is shown in FIG. 3(b). When the injection current Ii reaches 60 AICI1+2, light having a spectrum having a peak at the band gap (Eg) of the active layer semiconductor is emitted. Injection current is 8OA/c
m" or more, light having a peak at 2Eg is also emitted. When the injection current is 120 A/cm" or more, the light of 2Eg becomes dominant. This is because the injection current increases,
This is because the injection density of Cooper pairs increases, and the process of simultaneous recombination of Cooper pairs and two holes becomes dominant. The external quantum efficiency of light emission when the driving current was 12 OA/am was 5%. On the Motohiro side, YBCO was used as a superconductor.
However, in principle, any material that exhibits superconductivity may be used. For example, TI, Ba, Ca, Cu, 01. (Hereinafter, T
Abbreviated as BCCO; Tc=125K), Bi2Sr, C
a, Cu30□. (abbreviated as B5CC0; Tc = 1
15K), Nd,s,Ceal,Cu04(N
A high-temperature superconductor such as CCO (abbreviated as Tc = 24K), or a thin film of a normal metal superconductor such as Nb, Nb3Ge, Nb3Sn, or pb may be used.Although a GaAs-based pn junction is shown as an example, an InP-based It goes without saying that a similar device can be realized by using a pn junction in a semiconductor, an n-vr group semiconductor, or the like. For example, depending on the desired emission wavelength, CdTe, Zn5e, ZnS, etc.
- VI group group group conductor, GaP, GaN, GaAsP,
Similar devices can be realized using m-v group raw conductors such as GaInAsP and InGaAs. The following examples describe examples with specific materials, but the comments herein are equally applicable. Also, as a buffer metal between high temperature superconductors and compound semiconductors! Although the example of j/Mo/Au was shown, WSi/M. /Au or Ag can also be used instead of Au. Example 2 Another example of a superconducting semiconductor LED is shown in FIG. 3(c). Differences from Example 1 will be mainly explained. In a device with this structure, Cooper pair supply layers were provided on both sides of the semiconductor pin junction. By using MBE (molecular beam epitaxy) on a semi-MA edge GaAs substrate 91, Be is
X 10”/cn+3 containing p+AlxGa1-y
As layer 5 of 50 nm, undoped 1-GaAs layer 4 of 10 r+m, n+GaAs layer 31 containing SL of 3 x 10"/cm' of 40 nm, Sn of 9 x 1
An n++ GaAs layer 32 containing 0''/crm was formed in 15 layers. In this example, the A1 composition of the p+AlxGa, -yAs layer 5 was set to X=0.15, but normally 0.1≦X≦0.45.
Often selected within the range. Thereafter, an electrode pattern was formed through the same process as in Example 1, and then buffer metal layers 52, 53, 54, superconductor layer 1, and contact electrode layer 50 were formed. In order to form the back side @ electrode, the wafer is etched from the back side in a stripe pattern down to the p+A1xca knee xAs layer 5 using a normal etching technique, and after forming the electrode pattern, the buffer metal layer 521.53' is etched.
, 54', superconductor layer 1' and contact electrode layer 50
The specific contact resistance pc of a WSx 54' formed with p+AlxGa1-xAs at such a concentration is 3
X to -'Ωcn is about 12, and does not change at liquid nitrogen temperature or liquid helium temperature. Furthermore, to improve ohmic contact, use MOMBE (gas source MBE).
A C (carbon)-doped P+
By inserting 10 nm of ÷GaAs (doping level: 10"/cm3), the specific contact resistance can be reduced to I x 10"
00m2 (for example, Usa use and 9 other PN)
Study on non-alloy emitter electrode of p-type 2DEG-HBT; Sodo Technique ED88-134 pp87-pp92:1
January 19, 989). In this case, p+AlxGa, -
The thickness of the yAs layer 5 is reduced by 10 nm. The emission spectrum of the superconducting semiconductor light emitting diode of the present invention when operated at liquid nitrogen temperature (77 K) is shown in FIG. 3(d). In this structure, light is reflected from the end face, and light with a spectrum having a peak at the band gap (Eg) of the active layer semiconductor is emitted until the injection current Ii reaches 6 OA/am2. When the injection current is 80A/cm2 or more, 2
Light having a peak in layer g is also emitted. This is because as the injection current increases, the injection density of Cooper pairs into one layer increases, and the process of simultaneous recombination of Cooper pairs and two holes becomes dominant. At an injection current of 120 A/Cm" or more, the light from the second layer g becomes dominant, and furthermore,
When the forward bias voltage is V, 2nd layer g+qV, 2nd layer g-
The emission of satellite light of qV, Eg+qV, and Eg-qV was also measured in wt. These satellite lights are originally generated because the intensity of the energy light of Eg and the second layer g, which is emitted by elementary processes, is modulated at a frequency of 2qV/h. Example 3 An example of a double heterojunction superconducting semiconductor laser is shown in FIG. 3(e). MB on a semi-M edge GaAs substrate 91
Using E (molecular beam epitaxy), Be is 5X 1
01'/cm3 Contains p+AlxGa, -xAs!
(x≧0.4) 21 at 300 nm, Be at 5X
p-AlyGa, -yAs containing 10”/cm3
Layer (y = 0.2 to 0.3) 22 to 100 nm, undoped AlyGa, -yAs (z = o to 0.1)
n-AlyGal-yAs containing 3 nm of 23 and 5X10”/cm3 of Si 40 nm of 24, Si
n-AlxGa, -x containing 5 x 10"/amff
As 25 at 40 nm, and Sn at I x 102
0/cm3 containing n++GaAsM26 of 10nm
, formed. After that, a width of 1 mm is used as a laser resonator.
A 0 μm stripe is left using conventional gluing and etching techniques. Non-doped 1-AlxGa for embedding
t-xAs (x≧0.5 is often used) 27 is selectively grown by liquid phase epitaxy, NOCVD, or MOMBE4m. After forming an electrode pattern using the same method as in Example 1, buffer metals 54, 53, 52, superconducting layer 1, and contact electrode 50 were formed. In order to form backside electrodes, the wafer is etched in stripes to form an electrode pattern, and then a buffer metal layer 54 is etched.
', 53', 52', superconductor layer 1', and contact electrode layer 50' were formed. The threshold current (Ith) of the superconducting semiconductor double hetero BH laser created in this way at liquid nitrogen temperature (77K) is I
OA/cm". This value was 750 A/cm" for the conventional bulk DH laser and 250 A/cm" for the MQlj laser.
am" is much lower than that. Figure 3 (g)
shows the relationship between the injection current and optical output of this laser. When the laser is operated at an injection current value of Ith to about twice Ith, a laser beam of 862 nm corresponding to the bandgap Eg of the active layer is emitted. When the current density is further increased, the substantial optical gain increases even for light corresponding to 2Eg, and a laser beam of 431 nm is emitted. At this time, the oscillation of light corresponding to Eg is suppressed to an extremely low level. The relaxation oscillation frequency during modulation of laser oscillation is 4 GHz for a DH laser and 20 GHz for a modulation doped multiple quantum well (MQld) laser when operating at 100 mW, but in the superconducting semiconductor DH laser of the present invention, 8
62nn+ light reached 100 GHz, 431 nm light reached 50 GHz, and modulation between 862 nm light and 431 nm light reached 500 GHz. The optical spectrum during modulation also oscillated in a completely single mode, reflecting the extremely limited state density distribution of the laser of the present invention. Furthermore, when emitting 2Eg light, the intensity of the light is modulated at a frequency proportional to the applied voltage, effectively Eg-qV,
Light of Eg+qV, 2Eg-qv, and 2Eg+qV is emitted. The intensity of these satellite lights is extremely strong at the basic emission line, so it is not very noticeable, but emission that is about four orders of magnitude weaker is observed. During operation of a superconducting semiconductor laser, qv is almost equal to Eg, so 0, Eg, 2Eg
, 3Eg satellite light was emitted, and what was actually observable was light corresponding to 3Eg at 287 nm. In the above embodiments, the device structure in which the electrodes are taken from the back side of the substrate has been described, but the electrodes can also be taken from the front side. Next, this case will be explained. Example 4 A third example of a planar structure of a superconducting semiconductor LED
Shown in Figure (h). Differences from Example 2 will be mainly explained. The element with this structure differs from Example 2 in that the superconducting electrode connected to the ρ-type semiconductor layer 5 has a planar structure. semi-insulating G
Using MBE (molecular beam epitaxy) on an aAs substrate 91, p+Al containing Be I x 10"/cm'
xGa1-xAsAsO20nm, undoped 1-Ga
A5J! f 4 is 10 nm, Si is 3×10
18/cm” n+GaAs layer 31 containing 40
A 15 nm thick n++ GaAs layer 32 containing 9XIO1 g/cI113 of Nu and Sn was formed. In this example, the p+AlxGa1-xAsAsO21 composition was set to X=015, but it is usually selected in the range of 0.1≦X≦0.45. Thereafter, through the same process as in Example 1, after forming an electrode pattern, the buffer metal layer 52, 53, 54, superconductor layer 1 and contact electrode layer 5 are formed.
0 was formed. At this time 1, the semiconductors N4, 31 and 32 are selectively removed using a normal etching technique, and the buffer metal layers 52', 53', 54' and the superconductor layer 1'' are removed.
A contact electrode layer 50' was then formed. At this time, both electrodes are insulated with a thin RM99 (thick film) in between. At this time, due to the proximity effect of the superconductor 1' connected to the p-type semiconductor layer 5, the tItAm layer 99 (
It is desirable that the p-type semiconductor 5 corresponding to the layer thickness 9) be made into a superconductor. To this end, the Cooper pair seepage distance ξL may be made larger than the thickness of the insulating layer 99. Further, it is necessary to design the shrimp structure so that the Cooper pair seeping into the n-type semiconductor layers 32 and 31 reaches the p-type semiconductor layer 5 at least when the forward bias to be used is applied. Such a planar structure can be similarly applied to the laser of the third embodiment. Further, the technique for improving the ohmic characteristics of the P-type semiconductor layer 5 is the same as in Example 2.3. In addition, when the present invention is utilized as a light receiving element, it is possible to realize a light receiving element having the structure shown in FIGS. 3(a) and 3(C) described in Example 1.2, for example. According to such a light receiving element according to the present invention, since a portion of the semiconductor is converted into a superconductor, the density of states is extremely steep and the absorption band is extremely narrow. That is, FIG. 3(a),
In the case of (c), by applying a reverse bias to the junction portion, it is possible to operate as a light receiving element with a narrow band. [Effects of the Invention] The present invention uses semiconductors that do not exhibit conventional superconductivity by themselves (Si, GaAs, InGaAsP, In
By simply forming a superconducting electrode on a semiconductor (such as GaAs), the semiconductor can be converted to a superconducting state, making it possible to realize an optical element with extremely excellent light-emitting and light-receiving characteristics. Furthermore, by using a semiconductor junction that has been converted to a superconducting state, we discovered a new physical phenomenon of light emission that is coupled to a superconducting current that reflects the phase difference Δφ of the superconductor in the light emission spectrum. Further, the optical device of the present invention has the advantage that conventional semiconductor devices can be easily integrated monolithically. It goes without saying that the present invention becomes particularly practical by using a high-temperature superconductor whose transition temperature Tc is higher than the liquid nitrogen temperature.

[Brief explanation of drawings]

Figure 1 (a), (b), (c), (d), (e), (f
), (g), (h), (i) and (j) are the principle of the present invention, and a diagram to explain the energy band diagram and density of states etc. for explaining the principle, Fig. 2 (a) ,(b),C
, Q) and (d) are diagrams for explaining the density of states and energy bands of conventional optical devices, and Figures 3 (a) and (b)
, (c), (d). (e), (f), (g) and (h) are diagrams for explaining the device according to the present invention, and Figures 4 and 5 summarize experimental data on high temperature superconductor/semiconductor proximity effect. It is something. Explanation of symbols l, 1'... Oxide high temperature superconductor or superconductor, 2.2'... Buffer metal, 4...
...Low doping level light-emitting and receiving layer, 5...
・N-type degenerate semiconductor (GaAs, etc.), 3...N-type degenerate semiconductor (GaAs, etc.), 101.103...
The density of states of electrons in the region where Cooper pairs seep out, 101 is the excited state 103 corresponding to the ground state, 102...Density of states of holes, 104...The density of states in the degenerate region Hole, 1021.104'...The density of states of the hole in the region where the Cooper pair seeps out, 104' is the excited state 102' corresponds to the ground state, 401.403.111.112 ... Semiconductor layer with a large band gap forming a quantum well, 402.113.114 ... Semiconductor layer with a small band gap forming a quantum well, 115 ... Quantum well Barrier layer inside, 54.52・
...Buffer metal, 50.51...A
u, one electrode metal such as Ag, 5.21□...-p+A
lxGa1-xAs. 4...Undoped 1-GaAs, 90...
・・p-type GaAs substrate, 22− ・=p+A1yGal-yAs, 31−−n+
GaAs, 32.26-- n++GaAs or n+InGaAs
, 91... semi-insulating GaAs substrate, 23...
...Undoped AlzGa1-zAs layer, 24.25
--n-AlxGa1-xAs, 27...Undoped AIGaAsM. Tea 1 Figure (a) Figure 1 (b) Figure ((5) 1 hole obtained L-& Figure (d,), f Figure Ce) 1st < D d person left! 'J N Broom 1 diagram <-/,) (O'tomo! Fangshaku 1 diagram (j) Answer 2 den (0-) Ba 20 (b) 2nd UA (d) Tei 3 diagram (b) Lty, E Kirkey (ε〆) Graduation 3 figure (σ) No. 3 figure (item) Ekiru of the earth! - (ev) Figure 3 (C) 3rd (") Current +!! le <A/l:, rhinoceros) (A'2%) /L11I7 LiYfrA (=B- of "E, +/ -1'l':
AR1 (Scream W) S' Ritachipipe Tsucho Hit -0 (/-1 (A2w) /L17 1 > Qp Envy Seven pj-
, +I-Otsu 1cho 9 (%4ν) s'(

Claims (1)

[Claims] 1. It has a junction system of multiple regions including a superconductor region and a semiconductor region, and this junction system infiltrates Cooper pairs in the superconductor region into the semiconductor region using a proximity effect. and a structure for implanting the Cooper pair into the semiconductor region, and within a penetration length ξ_S and an implantation length ξ_I in the semiconductor region, a conductor portion of the semiconductor region is converted to a superconductor. A superconducting semiconductor optical device characterized in that a junction structure is disposed within the penetration length ξ_S and the injection length ξ_I into the semiconductor region, and has one or more lead-out electrodes. 2. A superconducting semiconductor optical device characterized in that the superconducting semiconductor optical device according to claim 1 and a normal semiconductor device are monolithically formed. 3. A superconducting semiconductor optical device according to claim 1 or 2, wherein the semiconductor region is a degenerate semiconductor. 4. In the superconducting semiconductor optical device according to claim 1, 2, or 3, the superconductor region and the semiconductor region are bonded via a buffer metal. Features of superconducting semiconductor optical devices. 5. The superconducting semiconductor optical device according to claim 1, wherein the junction between the buffer metal and the semiconductor region is an ohmic junction. 6. A superconducting semiconductor optical device according to claim 1, 2, or 3, wherein the superconductor region is an oxide high-temperature superconductor. Semiconductor optical device. 7. In the superconducting semiconductor optical device according to claim 4, the superconductor region is an oxide high temperature superconductor, and the buffer metal is a protective film of the semiconductor region (superconductor region / a first metal that acts as a barrier for total interdiffusion between the buffer metal and the semiconductor region) and a first metal that becomes a conductor if the oxide high temperature superconductor is not oxidized or the oxidized material becomes a conductor. and a third metal for adhering the first and second metals formed on the second metal. 8. In the superconducting semiconductor optical device according to claim 1, the superconductor region is made of a metal superconductor, and the superconductor region is formed directly on the semiconductor region,
A superconducting semiconductor optical device 9 characterized in that the junction is an ohmic junction, and the superconducting semiconductor optical device according to claim 5 or 8, wherein the specific contact resistance of the ohmic junction is A superconducting semiconductor optical device characterized in that the superconducting semiconductor optical device is formed to have a resistance of 5×10^-^6 Ωcm^2 or less. 10. The superconducting semiconductor optical device according to any one of claims 1 to 9, in which the semiconductor region is a degenerate p-type semiconductor, a degenerate n-type semiconductor, or a degenerate p-type semiconductor. At least one lead-out electrode is formed in the superconductor region, and another ohmically connected electrode is formed on the side of the semiconductor region remote from the superconductor region. A superconducting semiconductor optical device characterized in that an electrode is formed. 11. In the superconducting semiconductor optical device according to any one of claims 1 to 10, the semiconductor region includes a degenerate first semiconductor portion, a degenerate second semiconductor portion, and A superconducting semiconductor characterized by having a non-degenerate third semiconductor portion having a larger electron affinity than these semiconductors or a smaller sum of electron affinity and band gap than these semiconductor portions. optical element. 12. The superconducting semiconductor optical device according to claim 11, wherein the third semiconductor portion does not substantially contain impurities. 13. A superconducting semiconductor optical device according to claim 1, wherein the semiconductor layer region has one or more quantum wells. 14. The superconducting semiconductor optical device according to item 13, wherein the quantum well does not substantially contain impurities. 15. A superconducting semiconductor optical device according to claim 13, wherein the quantum well is selectively doped with an impurity. In the superconducting semiconductor optical device according to any one of Items 1 to 10, the semiconductor region includes a degenerate fourth semiconductor portion, a degenerate fifth semiconductor portion, and a portion between these semiconductor portions. A superconducting semiconductor optical device comprising a sixth semiconductor portion having a refractive index higher than those of these semiconductors. 17. The superconducting semiconductor optical device according to claim 16, wherein the sixth semiconductor portion does not substantially contain impurities. 18. The superconducting semiconductor optical device according to any one of claims 1 to 17, wherein the semiconductor region is made of GaAs, AlGaAs, AlAs, InGaAs,
A superconducting semiconductor optical device characterized in that it is formed of at least one selected from the group of compound semiconductors consisting of InP and InGaAsP. 19. A superconducting semiconductor region and a means electrically connected to a Cooper pair in the semiconductor region, which allows electrons constituting the Cooper pair in the semiconductor region to regenerate with holes in the semiconductor region. A superconducting semiconductor optical device that emits or receives light by bonding or generating these particles. 20. The superconducting semiconductor optical device according to claim 19, wherein the semiconductor region has a p-type region and an n-type region. 21. The superconducting semiconductor optical device according to claim 19, wherein the semiconductor region has a plurality of semiconductor regions with a large energy gap and a semiconductor region with a small energy gap disposed between these semiconductor regions. Conductive semiconductor optical device. 22. A superconducting semiconductor optical device according to claim 19, wherein the superconducting semiconductor optical device has means electrically connected to the hole. 23. The superconducting semiconductor optical device according to claim 19, wherein the semiconductor region is joined to a superconductor via a metal region. 24. The superconducting semiconductor optical device according to claim 19, wherein the semiconductor region is arranged between a plurality of superconducting regions. 25. The superconducting semiconductor optical device according to claim 23, wherein the superconductor is an oxide superconductor. 26. A superconducting semiconductor optical device that emits or receives light inside a superconducting semiconductor using distributed carriers due to a superconducting gap formed inside the superconducting semiconductor.