JPH0555713A - Light emitting semiconductor element - Google Patents

Abstract

PURPOSE:To realize a micro-resonator laser which can raise the coupling rate of naturally emitted light with its oscillating resonator mode to a sufficiently high level and has a higher efficiency and lower threshold. CONSTITUTION:In a micro-resonator laser provided with a double heterostructure section composed of a quantum well active layer 31 between the first and second distributed Bragg reflectors 20 and 40 respectively constituted by alternately piling up pluralities of semiconductor films having a larger refractive index and another semiconductor films having a smaller refractive index, the reflector 40 which is on the opposite side of a substrate 10 with respect to the double heterostructure section is formed in a projecting state on the opposite side of the double heterostructure section on the light emitting area of the double heterostructure section.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a semiconductor light emitting device, and more particularly to a highly efficient semiconductor light emitting device in which the effective size of a resonator is set to about the wavelength of operating light.

[0002]

2. Description of the Related Art In recent years, semiconductor lasers have been widely used as a light source for optical communication and optical information processing. Also, OEI
In integrated circuits called C (optical / electronic integrated circuits) and photonic ICs, attempts have been made to monolithically integrate a semiconductor laser with other light receiving elements. Furthermore, array integration of semiconductor lasers has also been studied for the purpose of parallel computing. In such an integrated circuit, it is important to reduce the threshold value of the laser in order to increase the integration degree of the semiconductor laser.

By the way, the efficiency of the semiconductor laser and the ultra-low threshold value are achieved by using a microresonator structure in which the size of the resonator is about the wavelength of the operating light (electromagnetic wave). Is being studied. This is to control the spontaneous emission light and improve the characteristics of the light emitting element.
The principle for controlling the spontaneous emission light will be briefly described below.

As is well known, there are two types of light emission processes, a stimulated emission process and a spontaneous emission process (A. Einstein, Verhandlung der Deutsche Physicalisc.
heGesellschaft, Bd. 18, 318 (1916)). Stimulated emission is
Probability that is proportional to the strength of the resonant electromagnetic waves at that location,
It is the reverse process of absorption. On the other hand, spontaneous emission is said to occur regardless of the incident electromagnetic wave. A laser is an optical device that uses a stimulated emission process, and a light emitting diode uses spontaneous emission light.

Here, the spontaneous emission probability is photon (electromagnetic wave).
The product of the modal density (m (ν)) and the shape of the emission spectrum line (g (ν)) is integrated by the frequency (ν) of the electromagnetic wave. In the above discussion of A. Einstein, we assume vacuum photon modal density. Therefore, when the light emitting medium is in the microresonator having a size of about the emission wavelength, the spontaneous emission probability changes due to the change of the photon mode density.

A laser based on such an idea is called a microcavity laser. Taking a semiconductor laser as an example, further description will be given with reference to FIG. In FIG. 22, the left side shows the characteristics of a normal semiconductor laser and the right side shows the characteristics of a microcavity laser.

In an ordinary semiconductor laser, assuming that the cavity length is L and the refractive index is n ', the wavelength of each mode in the axial direction is defined by λ _m = 2L / n'm where m is an integer. ..
Here, since L is generally several hundred μm, (1 in FIG.
The mode interval is several angstroms as shown in -a).

Since the emission spectrum line width of the semiconductor is several tens of nm as shown in (1-b) of FIG. 22, the spontaneous emission light is shown in (1-c) of FIG. 22 on the wavelength axis. Like That is, at a current below the oscillation threshold,
The injected carriers are recombined at each wavelength within the emission spectrum line width, and the spontaneous emission probability is the same as in the case without the resonator. When the injection current is further increased, (1-d) in FIG.
As shown in FIG. 22, the laser beam oscillates at the wavelength λ _L , and an optical output-current curve having a threshold characteristic as shown in (1-e) of FIG. 22 is obtained.

Now, assuming that the resonator length L is about the peak wavelength λg of the spectrum line (such a resonator is called a microresonator), the spectrum g (λ) is obtained.
It is possible to allow only a few modes, including zero, to exist in a region where is finite. In FIG. 22, (2-
The case where only one mode exists is shown in a). By doing so, the emission spectrum, the spontaneous emission spectrum, and the oscillation spectrum become as shown in (2-b), (2-c), and (2-d) of FIG. Therefore, a zero threshold semiconductor laser can be realized as shown in (2-e) of FIG.

Even if such optical confinement is not perfect, the microresonator can reduce the oscillation threshold value. In general, when the proportion of spontaneous emission light coupled to the oscillation resonator mode is β (β ≦ 1), the oscillation threshold value is approximately inversely proportional to β. Since a normal laser has β to 10 ⁻⁵ , it is expected to reduce the threshold value by about four digits even if the ratio β coupled to the oscillation resonator mode is 10%.

As described above, it is already known that the threshold value of the semiconductor laser and various functional light emitting devices can be realized in the microresonator (Japanese Patent Laid-Open No. 59-50589). However, the microresonator needs to be a resonator having a high reflectance. Otherwise, the optical mode outside the cavity will enter the cavity and the mode density will be the same as that of the vacuum field. The above publication does not describe any specific measure regarding the structure for obtaining a high reflectance.

An example of embodying a microresonator with a semiconductor material is described in the literature (Yoshihisa Yamamoto, Masahito Ueda, The Institute of Electronics, Information and Communication Engineers, 72 (9) pp.1014-1020 (1989) and 73 (2) pp.161. -16
6: Y.Yamamoto, "Cavity Quantum Electrodynamics & Qu
antum Computing "; 1stInternational Forum on the Fr
ontier of Telecommunications Technology, SESSION 3
) Is reported.

An example of this is shown in FIGS. 23 (a) and 23 (b). In order to make a resonator with high reflectance, AlGaAs /
Distributed Bragg reflectors (λ / 4 dielectric multilayer films) 2 and 3 made of AlAs multilayer films are used, and the active layer 1 has a quantum well structure. A reflector in which a plurality of semiconductor films having a large refractive index and a semiconductor film having a small refractive index are alternately laminated is called a distributed Bragg reflector. In addition, by setting the interval surrounded by the two distributed Bragg reflectors 2 and 3 to be λ (FIG. 23A) or λ / 2 (FIG. 23B), the antinode and the node 4 of the mode are respectively separated. In this case, it is designed to be located in the active layer. In the former case, the spontaneous emission intensity is increased, and in the latter case, the spontaneous emission intensity is suppressed.

In the above-mentioned example of the microresonator, the optical mode is formed only in the stacking direction, and the light traveling in the oblique direction or the layer direction with respect to the stacking direction is difficult to be coupled to the resonator mode. It becomes or it does not combine. This is a factor for reducing the ratio (β) of coupling the spontaneous emission light with the oscillation resonator mode. Therefore, in the semiconductor microresonator, not only in the vertical direction (thin film crystal growth direction) but also in the horizontal direction, the resonator length is shortened to about the wavelength, and the entire surface of the resonator is covered with a light reflection film (or a multilayer film). Is desirable.

However, although it is easy to form the distributed Bragg reflector in the vertical direction by the thin film crystal growth technique, it is not easy to form the distributed Bragg reflector also in the horizontal direction. That is, in the formation of a light-reflecting film by the coating of a normal semiconductor laser, sputter vapor deposition, electron beam vapor deposition, or the like is used, and the surface to be coated is vapor-deposited in the direction in which the vapor-deposition species come in. Surfaces other than the surface are also vapor-deposited due to the inclusion of vapor-deposition species. The thickness of the vapor deposition layer due to the wraparound is generally thinner than the thickness of the vapor deposition on the surface to be originally coated. Therefore, it is impossible to apply the same multi-layer film coating in multiple directions such as the vertical direction and the horizontal direction.

Further, even when attempting to monolithically integrate such a semiconductor light emitting device, the conventional coating technology has a problem in that the adjacent microresonators interfere and the vapor deposition species are not evenly distributed over the entire surface. There is also.

[0017]

As described above, in the conventional semiconductor light emitting device using the microresonator, in order to increase the ratio of coupling of spontaneous emission light to the oscillation resonator mode, not only the vertical direction but also the horizontal direction. It is desirable to form a resonator for light propagating in the direction. However, it is difficult to form the light reflection film for forming the resonator in the horizontal direction. For this reason, spontaneous emission cannot be sufficiently controlled, and in the case of a light emitting element, the luminous efficiency is reduced.

The present invention has been made in consideration of the above circumstances, and an object thereof is to sufficiently increase the rate of coupling of spontaneous emission light to the oscillation resonator mode, resulting in higher efficiency and lower efficiency. It is to provide a threshold semiconductor light emitting device.

[0019]

The essence of the present invention is that the light reflecting surface of the microresonator is composed of two sets of distributed Bragg reflectors, and at least one of the two sets of distributed Bragg reflectors is a single unit. It consists of a distributed Bragg reflector which is not a plane. Here, a surface that is not a single plane means a surface or a curved surface that is composed of a plurality of planes.

That is, according to the present invention (claim 1), a first distributed Bragg reflector in which a plurality of semiconductor films having a large refractive index and semiconductor films having a small refractive index are alternately laminated, and a structure similar to that of the reflector are provided. In a semiconductor light emitting device in which a semiconductor double heterostructure is formed between a second distributed Bragg reflector and at least one of the distributed Bragg reflectors on the light emitting region of the double heterostructure. It is characterized in that it is formed in a convex shape on the opposite side.

According to the present invention (claim 2), a semiconductor double film is provided between a distributed Bragg reflector and a metal reflector in which a plurality of semiconductor films having a large refractive index and semiconductor films having a small refractive index are alternately laminated. In a semiconductor light emitting device having a heterostructure part, at least one of a distributed Bragg reflector and a metal reflector is formed in a convex shape on the light emitting region of the double heterostructure part on the side opposite to the double heterostructure part. It is characterized by

According to the present invention (claim 3), a first distributed Bragg reflector in which a plurality of semiconductor films having a large refractive index and semiconductor films having a small refractive index are alternately laminated, and a structure similar to the reflector. In a semiconductor light emitting device in which a semiconductor double heterostructure section is formed between the second distributed Bragg reflector and the second distributed Bragg reflector, both of the distributed Bragg reflectors have a curved surface having a recess on the active layer side of the double heterostructure section. And the active layer of the double heterostructure is located at the focal point of the resonator mode defined by the Bragg reflector.

[0023]

As described above, in a microresonator in which two sets of semiconductor distributed Bragg reflectors are opposed to each other, light confinement is effective only in the one-dimensional direction, and spontaneous emission in oblique and lateral directions also occurs. In addition, the ratio of spontaneous emission light coupling with the oscillation resonator mode is low. On the other hand, like the present invention,
By making the distributed Bragg reflector (or metal reflective film) convex when viewed from the resonator, it is possible to act as a reflector even for light traveling in a direction other than the vertical direction, and oscillation resonance of spontaneous emission light. It is possible to increase the rate of coupling to the vessel mode.

Further, in the present invention, since the distributed Bragg reflector has a structure formed by a plurality of planes, it is possible to confine light not only in one direction but also in a plurality of directions, and the oscillation resonator of spontaneous emission light is provided. The rate of coupling to the mode can be increased. Furthermore, since the structure of the present invention can be collectively formed by using the semiconductor dry etching or the crystal growth technique, it can be formed in a large area with good reproducibility and is easily integrated.

[0025]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a sectional view showing a schematic structure of the first embodiment of the present invention. This example is an ultralow threshold microcavity laser with enhanced spontaneous emission.

N-type GaAs substrate (carrier concentration 2 × 10
A laminated structure of semiconductors is formed on ¹⁸ cm ⁻³ ) 10 by using, for example, a low pressure MOCVD method, and electrodes 11 and 12 are formed on the uppermost layer and the back surface of the substrate, respectively.

In the laminated structure of semiconductors, 21, 23,
25, 27, 41, 43, 45, 47 are Al _0.8 Ga
_0.2 As layer, 22, 24, 26, 28, 42, 44, 4
6, 48, 32, and 33 are Al _0.3 Ga _0.7 As layers, and 50 is a p-type GaAs contact layer (carrier concentration 5
× 10 ¹⁸ cm ^-3 ). 41 to 48 are p-type and impurity-doped to a carrier concentration of 2 × 10 ¹⁸ cm ⁻³ , and 21 to 28 are n-type.
The mold is impurity-doped to a carrier concentration of 2 × 10 ¹⁸ cm ^-3 .

A junction of semiconductors having different compositions or composition ratios is called a heterojunction, and a structure having two heterojunctions is called a double heterostructure. Reference numerals 31 to 33 form a double heterostructure portion, and reference numeral 31 is an active layer in which electrons and holes are optically recombined. In this embodiment, the active layer 31 is made of InG having a thickness of 7 nm.
31-33 is called a quantum well structure because it is an aAs layer and its thickness is about the de Broglie wavelength of electrons.

Reference numerals 21 to 28 and 41 to 48 constitute distributed Bragg reflectors 20 and 40, respectively, and the thickness of each layer (in the case of a curved surface, the shortest distance between hetero interfaces) is λ / It is 4. Here, λ is the emission peak wavelength of the quantum well. In the following description, the distance using the symbol λ represents the optical distance (the actual length multiplied by the effective refractive index) unless otherwise specified. In general, the thickness of each layer forming the distributed Bragg reflector may be an odd multiple of λ / 4.

The InGaAs quantum well active layer 31 is the first
From the top layer 21 of the distributed Bragg reflector 20 of λ / 2. The bottom layer 41 of the second distributed Bragg reflector 40
Has a hemispherical shape with a radius of λ / 2 at the center,
The active layer 31 is located in the diameter portion of this hemisphere. More specifically, the active layer 31 is formed in a circular shape with a radius of λ / 2 in the central portion of the spacer layer 33, and the spacer layer 32 on the active layer 31 has a hemispherical shape with a radius of λ / 2. A distributed Bragg reflector 40 is grown and formed on these spacer layers 32 and 33. Further, the electrode 11 has a circular opening in the central portion as shown in the plan view of the device of this embodiment in FIG.

With this structure, since the second distributed Bragg reflector 40 is formed in a convex shape when viewed from the resonator, the distributed Bragg reflector 40 is adapted to light traveling in directions other than the vertical direction. However, it can be made to act as a reflector, and the rate of coupling of spontaneous emission light to the oscillation resonator mode can be increased. In this embodiment, the ratio β of spontaneous emission light coupling with the oscillation resonator mode can be increased to about 10%, and the oscillation threshold value is reduced by about 4 digits.

Next, a method of manufacturing the laser of this embodiment will be described with reference to FIG.
Will be described.

First, as shown in FIG. 3A, n-type Ga
The distributed Bragg reflector 20 is laminated on the As substrate 10 by the low pressure MOCVD method, and each layer 33, 31, 32 forming the double hetero structure portion is further laminated thereon. At this time, the spacer layer 32 grows thicker. Then, the mask 60 is formed by an electron beam exposure method and wet etching.
The (silicon oxide film) is formed so as to be convex upward.

Next, the spacer layer 32 and the active layer 31 are etched until they reach the spacer layer 33 by directional dry etching using chlorine plasma. At this time, since the mask 60 is also etched at a slower rate than the spacer layer 32, the mask 60 shown in FIG.
As shown in (b), the spacer layer 32 is left in a convex shape.

Before carrying out the directional dry etching using chlorine plasma, the active layer 3 is masked by using the mask 60.
You may perform the implantation of the proton which reaches 1. By doing this, the implantation region functions as a high resistance region,
It is possible to more effectively inject a current into the active layer 31.

Next, as shown in FIG. 3 (c), the reduced pressure M
Distributed Bragg reflector 40 and GaAs by OCVD method
The contact layer 50 is laminated. Finally, the electrodes 11 and 12 are formed by ordinary photolithography and vapor deposition,
By etching a part of the electrode 11 and the contact layer 50, the structure shown in FIG. 1 is obtained.

There may be various cases in the shapes of the quantum well active layer and the distributed Bragg reflector shown in FIG. 1, and the positional relationship between the active layer and the distributed Bragg reflector. For example, a plurality of pairs having a thickness of λ / 4 (a high refractive index layer and a low refractive index layer) may be added to the distributed Bragg reflector. The distance between the quantum well active layer and the distributed Bragg reflector may be an integral multiple of λ / 2. A so-called GRIN-SCH structure may be used in which the forbidden band width increases as the spacers 32, 33 move away from the quantum well active layer 31.

As shown in the perspective view of FIG. 4, the distributed Bragg reflector 40 may be formed in a semi-cylindrical shape. Even if the distributed Bragg reflector 40 of FIG. 1 is not spherical, if β is convex upward, the increase of β is similarly expected. In addition to the quantum well, the active layer may be a quantum wire or a quantum box as shown in FIG. In FIG. 5, 70
Indicates the quantum wire active layer. Further, although the distributed Bragg reflector 40 has a hemispherical shape in FIG. 1, the distributed Bragg reflector 20 may have a hemispherical shape, or both of the distributed Bragg reflectors 20 and 40 may have a spherical shape. Also, 4
Although 1 surrounds the active layer 31, it does not matter even if the structure is not surrounded (for example, the Al composition of 41 is the Al composition of 32).

FIG. 6 is a sectional view showing the schematic arrangement of the second embodiment of the present invention. This example is an example in which a microcavity laser can be produced by using the dependence of the crystal growth rate on the growth plane orientation.

Reference numeral 120 is an n-type distributed Bragg reflector similar to 20 in FIG. 1, and 140 is a p-type distributed Bragg reflector similar to 40 in FIG. However, the distributed Bragg reflector 140 has a regular tetrahedral structure. A plan view of the device of this example as seen from above is shown in FIG. A of this FIG.
A cross section of -A 'is shown in FIG.

The thickness of each layer of the distributed Bragg reflector 140 is set to λ / 4. Referring to FIG. 6, in the distributed Bragg reflector 140, the layer spacing on the left side seems thicker than the layer spacing on the right side, but as can be seen from FIG. 7, the layer on the left side has the correct layer spacing. Is not shown, and only looks thick. Further, 131 is a GaAs active layer, 151 is a silicon oxide film, and 132 and 133 are Al _0.3.
Ga _0.7 As layer (spacer layer), 111 and 112 are electrodes.

Next, the outline of the method of manufacturing the element of the second embodiment will be described.

A (111) plane GaAs substrate 110 is used for crystal growth. Distributed Bragg reflector 1 by MOCVD method
After growing 20 and the spacer layer 133, a silicon oxide film 151 is deposited thereon. Then, patterning is performed so that each side of the equilateral triangle becomes [110], [011], and [101], and the equilateral triangle portion of the silicon oxide film 151 is removed by etching. The length of one side of the equilateral triangle is about 0.
It is 1 μm. In addition, references (T. Fukui, S. Ando,
Y. Tokura, and T. Todhiyama: Extended Abstractsof th
e 22nd (1990 International) Conference on Solid St
ate Devices andMeterials, Sendai, 1990, pp.99-10
Using a selective growth method similar to that described in 2),
The active layer 131, the spacer layer 132, and the distributed Bragg reflector 140 are sequentially stacked. Due to the plane orientation dependence of crystal growth, structures with tetrahedral symmetry are stacked.

When the method of the present embodiment is used, since the symmetry of the crystal is utilized, the patterning of the silicon oxide film does not have to be an equilateral triangle. Regular tetrahedron structures are stacked. The silicon oxide film is
It functions as a selective growth mask when the device is formed, but it functions as a current confinement layer in the formed device.

FIG. 8 is a sectional view showing the schematic arrangement of the third embodiment of the present invention. In the second embodiment, the active layer is three-dimensionally surrounded by the distributed Bragg reflector, but in the present embodiment, it is an example in which it is two-dimensionally surrounded. This embodiment is also an example that can be created by using the dependence of the crystal growth rate on the growth plane orientation. It is to be noted that FIG. 9 is a plan view of the device of this example as seen from above. FIG. 8 shows a cross section taken along the line BB 'in FIG.

Reference numerals 220 and 240 are n-type and p-type distributed Bragg reflectors, and 231 is an InGaAs active layer, 2
51 is a silicon oxide film. In this embodiment, the GaAs substrate 2
As (10), the (001) plane is used. Silicon oxide film 251
In the reference (T.
Fukui, S. Ando, H. Saito: Extended Abstractsof the 2
2nd (1990 International) Conference on Solid State
Devices and Materrials, Sendai, 1990, pp.753-756)
Perform selective growth in the same manner as described in,
233, 231, 232 and 240 are sequentially laminated. Although this structure cannot make β as large as that of the second embodiment, it has an advantage that the lateral width of the active layer 231 can be controlled to be narrow even if the stripe of the silicon oxide film having a relatively wide width is formed. Is easy.

FIG. 10 is a sectional view showing a schematic structure of the fourth embodiment of the present invention. The present embodiment is an example in which the p-type distributed Bragg reflector 140 of the second embodiment is replaced with a metal reflective film which also serves as an electrode.

In the figure, 331 is an InGaAs active layer, 332
Is a p-type GaAs spacer layer, and 312 is a metal reflective film that also serves as an electrode. The spacer layer has a thickness of about λ / 5 to λ / 4. As the metal reflection film 312, A which is usually used
A laminated film of uZn / Au may be used, but Ag or Ag is used in order to prevent a decrease in reflectance due to alloying at the interface with the semiconductor.
A method using a metal film such as / Au or Ti / Pt / Au is desirable. The active layer 331 and the spacer layer 3 of this embodiment
During 32, the p-type distributed Bragg reflector 1 of the second embodiment
A structure in which 40 layers are stacked is also possible. In this case, there is an advantage that a high reflectance can be obtained even if the number of pairs of p-type distributed Bragg reflectors is reduced.

Next, fifth to seventh embodiments of the present invention will be described. In the above-described first and second embodiments, the planar-concave resonator is used as one of the resonator structures that simultaneously realizes the ease of manufacturing the one-dimensional resonator structure and the large β equivalent to that of the closed resonator. There is. In such a resonator structure, the resonance mode has a single wavelength over almost the entire solid angle, so that a value of β close to 1 can be obtained if an appropriate design is made. In addition, since the resonator structure can be formed by epitaxial growth, the manufacturing process is relatively simple.

The probability of spontaneous emission to the resonator mode is proportional to the square of the amplitude of the normalized resonator mode at the position where the light emitter exists. Therefore, in order to increase β, ideally, the mode amplitude at the position of the active layer is increased (that is, the Q value of the resonator is increased), and the active layer is ideally located at the antinode of the mode. Is. But first
In the structure of the second embodiment, since the flat reflective film is the distributed reflective film, the light emitted from the active layer penetrates deep into the reflective film. Therefore, if the curvature of the concave mirror is too large, part of the emitted light is emitted to the outside of the resonator, and a sufficiently large Q value cannot be obtained.

That is, in the first and second embodiments, FIG.
As shown in FIG. 1, the luminescence component reflected at the interface from the active layer 31 with respect to the center of curvature P of the upper spherical reflective film (distributed Bragg reflector 40) follows a stable resonance path like the optical path A, The light-emission component reflected at the interface closer to the substrate 10 than P escapes to the outside of the resonator through the path shown in B. In order to avoid this, the curvature of the spherical reflecting film is limited, and the center of curvature P is set to the lower flat reflecting film (distributed Bragg reflector 20).
It is necessary to move it closer to the substrate 10 side. However, if the curvature is reduced to that extent, the electric field amplitude of the resonator mode in the vicinity of the active layer 31 cannot be made sufficiently large as compared with the conventional one-dimensional resonator. Furthermore, since the surface on which the active layer 31 is present does not coincide with the equiphase surface of the resonator mode, the entire surface of the active layer 31 cannot coincide with the antinode portion of the mode. That is, in the structures of the first and second embodiments, the advantage of using the spherical reflecting film may not be fully utilized.

Therefore, in the following embodiments, in order to further reduce the loss of the resonator and to locate the active layer at the antinode portion of the oscillation resonator mode, a pair of concave reflecting films facing each other or optically with the concave reflecting films. By the transparent plane-concave mirror reflective film,
An optical resonator having a resonator length about the emission wavelength of the active layer is formed.

FIG. 12 is a perspective view showing a schematic configuration of the fifth embodiment of the present invention, and FIG. 13 is a sectional view taken along the line AA 'in FIG. In this embodiment, an n-type distributed reflection film 520, an n-type clad layer 533, an active layer 531, a p-type clad layer 532, a p-type distributed reflection film 540, and a p-type electrode 511 which are stacked on an n-type GaAs substrate 510. And the n-type electrode 512 on the back surface of the substrate are the basic constituent elements. The light emitted from the active layer 531 is extracted to the outside through the window W opened in the electrode 511 through the spherical distributed reflection film 540.

The distributed Bragg reflector films 520 and 540 are composed of a spherically curved multilayer film as shown in FIG. 13, and are 10 nm thick GaAs quantum well active layers 531 and Al _0.3 Ga _0.7 As.
A resonator structure is formed to face each other with the cladding layers 532 and 533 sandwiched therebetween. Each of these multilayer films is formed of AlAs (521,
541) and Al _0.18 Ga _0.82 As (522,542) are alternately laminated. The thickness of each layer is λ / 4 in terms of optical length (actual thickness multiplied by the refractive index). Here, λ is the emission wavelength of the active layer 531 in vacuum.

The clad layers 532 and 533 are processed into a lens shape having a diameter of about 0.5 μm by the method described later. The distance between the top and bottom of the double heterostructure including the clad layers 532 and 533 and the active layer 531 is λ in optical length as shown in FIG. The symbol F in the figure indicates the amplitude of the electric field along the resonator axis in the resonator mode, and the active layer 531 is located at the antinode portion of this resonator mode. Clad layer
The outside of the lens-shaped portions of 532 and 533 has a high resistance by ion implantation due to current constriction (indicated by a region I).

Here, as shown in FIG. 15, the eigenmode of the resonator in which two spherical reflecting mirrors are opposed to each other is represented by a "Gaussian beam". In this figure, the solid line represents the position where the electric field strength is 1 / e of the central axis.
As shown by the alternate long and short dash line in the figure, the equiphase surface of the beam becomes smaller in curvature as it approaches the minimum beam spot, and becomes a plane at the minimum beam spot. Therefore, by aligning the position of the active layer with the minimum beam spot, the active layer coincides with the equiphase surface of the cavity mode. Therefore, if the resonator length is set appropriately, the entire active layer is located in the antinode of the mode, and the ratio of spontaneous emission emitted to the resonator mode can be maximized.

Even if the reflection film is not a perfect spherical surface, such an effect can be similarly obtained as long as the shape is not significantly deviated from the spherical surface. In this embodiment, a spherical distributed Bragg reflecting film is used as the spherical reflecting mirror. In addition, the curvature radii R ₁ , R ₂ , and R ₂ of the reflecting mirror are set so as to satisfy the relationship of 0 ≦ (1-L / R ₁ ) (1-L / R ₂ ) ≦ 1.
By setting the resonator length L, optical loss in the lateral direction is suppressed. As a result, a resonator structure having a much higher Q value is realized as compared with a Fabry-Perot resonator structure in which plane reflecting mirrors are opposed to each other.

Since the spontaneous emission probability is proportional to the Q value (S.Ha
roche and D. Kleppner, PhysicsToday, p24 Jan. 198
9) The probability of spontaneous emission from the active layer increases both when there is no resonator and when it is placed in a Fabry-Perot resonator (typically, there is no resonator). About 10 times). This means that the luminous efficiency is substantially increased. Because the luminous efficiency is
It is determined by the balance between the non-radiative recombination probability and the radiative recombination probability. The former is constant regardless of the structure of the resonator, whereas
This is because the latter increases due to the effect of the curved resonator of this embodiment, and the proportion of injected carriers that contribute to light emission increases.

Generally, in a semiconductor light emitting medium, the probability of non-radiative recombination is higher than the probability of radiative recombination at room temperature, and the luminous efficiency for spontaneous emission is extremely low (about several percent). However, this increase in the radiative recombination probability makes it possible to increase the luminous efficiency for spontaneous emission to several tens of percent. This is an important point in realizing a microcavity laser that operates at room temperature. By the synergistic action of the above two effects,
A remarkable threshold value reducing effect can be obtained.

The semiconductor light emitting device according to this embodiment is
The spontaneous emission light itself has a relatively narrow spectral line width of several GHz to several tens of GHz, and the spontaneous emission life is 1 ns at room temperature.
Because of its short length, it can be used in a manner similar to a laser even with an injection current below the oscillation threshold. For example, it can be used as a high-efficiency and high-speed LED operating at 1 GHz with an injection current of about several nA per element.

FIG. 16 is a sectional view showing a schematic structure of the sixth embodiment of the present invention. In this embodiment, the electrode 511 to the active layer
Up to 531 is common to the fifth embodiment. The difference is that the bottom surface of the cladding layer 533 is flat, the substrate 510 directly below the lens-shaped portion of the cladding layer 532 is removed by etching, and the metal reflective film / electrode 513 is directly deposited on the cladding layer 533. Is. The reflective film / electrode 513 is made of GaAs in order to obtain ohmic contact while maintaining high reflectance.
A metal such as Ag that does not cause alloying with is used. In this embodiment, it is essential to use this metal reflecting film as a plane reflecting mirror in order to obtain a high Q value. The positional relationship between the position of the active layer 531 and the cladding layers 532, 533 and the reflective film / electrode 513 in this embodiment is the mode amplitude (F) of the resonator.
On the other hand, it is shown in FIG. Clad layer 532
The optical distance between the top of the and the reflective film / electrode 513 is 3λ /
4, the active layer 531 is located in the mode antinode of the resonator in this case as well.

Even if the spherical reflecting mirror and the plane reflecting mirror are combined, a resonator equivalent to the resonator described in the fifth embodiment can be obtained. FIG. 18 shows a schematic diagram thereof. this is,
A structure in which a spherical reflector is folded back around a plane reflector,
That is, the structure is optically equivalent to a symmetrical spherical resonator having a resonator length of 2L. At this time, the condition for a stable resonator is R> L, where R is the radius of curvature of the spherical reflecting mirror. In this embodiment, the spherical reflecting mirror is the distributed reflecting film 540 and the plane reflecting mirror is the metal reflecting film 513. However, in this resonator structure, the equiphase surface becomes flat on the plane reflection film, so that the active layer cannot exactly match the equiphase surface. However, since the active layer is located on a plane that is very close to the quarter-wavelength to one-wavelength from the plane reflection film, there is substantially no problem. Of course, even in this structure, a significant reduction in the threshold value is achieved for the reasons described above.

FIG. 19 is a sectional view showing the schematic construction of the seventh embodiment of the present invention. In this embodiment, the outer surface of the lens-shaped portion of the cladding layer 532 is covered with a silicon dioxide film 534, and the distributed reflection film 540 is grown using this as a selective growth mask. In this embodiment, since the silicon dioxide film 534 also functions as a current confinement layer, the ion implantation region I is unnecessary. Although the manufacturing process of the present embodiment is more complicated than that of the sixth embodiment, the Q value of the resonator becomes larger and the threshold value becomes larger because the solid angle that allows the reflection film 540 to be seen from the active layer 531 is large. Can be lowered.

In the fifth to seventh embodiments described above, the thickness of the cladding layers 532 and 533 and the position of the active layer 531 are the same as those of the active layer 5.
It can be arbitrarily changed within the range that satisfies the condition that 31 is located in the antinode of the resonator mode. The examples are shown in FIGS. 14 and 17. Symbols H and L in the figure respectively indicate that the refractive index is higher and lower than the adjacent layers (excluding the active layer). 14 (a) and 17
14B is an example in which the stacking order of the distributed Bragg reflector films is reversed from that in the fifth and sixth embodiments, and FIG. 14C shows only the stacking order of the distributed reflector films in the reverse order of the fifth embodiment. Here is an example. Further, FIG. 17C shows an example in which a dielectric film 515 such as silicon dioxide is inserted between the cladding layer 533 and the reflective film / electrode 513 in order to reduce the absorption loss due to the metal film.

Next, a method of manufacturing the device of this embodiment will be described. FIG. 20 is a cross-sectional view showing the manufacturing process of the fifth embodiment element.

First, as shown in FIG. 20A, a silicon nitride film M1 is formed on an n-type substrate 510 by a plasma CVD method or the like. A resist R is applied thereon and an opening having a diameter of 0.5 μm is formed by electron beam exposure. Then, the surface of the mask M1 is etched using hydrofluoric acid or a dilute solution of hydrofluoric acid to form a recess in the opening of the resist. At this time, the etching of the mask M1 is stopped before the etching reaches the substrate surface.

Then, as shown in FIG. 20B, the resist R is removed and chlorine-based reactive ion etching (R
IE), reactive ion beam etching (RIBE) or the like is used to etch the mask M1 having a depression and the substrate 510. Since the etching depth of the substrate 510 becomes larger as the mask M1 becomes thinner, a recess is also formed on the substrate 510. This shape is the mask shape and the mask and Ga.
It depends on the difference in the etching rate of As. Usually Ga
Since the etching rate of As is slightly higher than the etching rate of the mask, the curvature of the depression on the substrate is larger than that of the mask.

Then, metal organic chemical vapor deposition (MOCVD)
Method and molecular beam epitaxy (MBE) method.
0 (c), the n-type distributed Bragg reflection film 520, n
The type clad layer 533, the active layer 531, and the p-type clad layer 532 are laminated. Then, a resist is applied on the cladding layer 532, a circular pattern having a diameter of 0.5 μm is formed by electron beam exposure, and a Ga ion beam of about 100 keV is implanted to form a high resistance region I for current confinement. After that, while rotating the substrate 510, several hundred eV is obliquely applied.
Then, a lens-shaped mask M2 made of resist is formed.

Next, RI using a mixed gas of oxygen and chlorine is used.
The cladding layer 532 is etched together with the mask M2 by E or RIBE to form lens-like protrusions on the cladding layer 532 as shown in FIG. A p-type distributed Bragg reflection film 540 is formed on the protrusion by MOCVD or MBE.
To stack. Finally, an electrode is vapor-deposited on the back surface of the substrate and the distributed Bragg reflection film 540, and a window W (see FIG. 11) for taking out output light is formed to complete the process.

The fifth and fifth manufacturing methods of the elements of the sixth and seventh embodiments are also described.
According to the example element of. However, as a matter of course, in these embodiments, the step of forming the recess in the substrate 510 and the distributed reflection film 52
The step of stacking 0 is unnecessary. Instead, the substrate 510
Is added until the cladding layer 533 is reached. In order to stop the etching accurately on the surface of the clad layer 533, selective etching using a mixed etchant of ammonia water and hydrogen peroxide solution may be performed.

In the seventh embodiment, the cladding layer 532
After providing a protrusion on the upper surface of the substrate, a silicon dioxide film 534 is laminated. As shown in FIG. 19, the silicon dioxide film 534 on the upper surface of the protrusion is removed by lithographic etching using an electron beam and hydrofluoric acid, and then selective growth is performed by the MOCVD method.
A type distributed Bragg reflection film 540 is laminated. The p-type electrode 511 is provided so as to be in contact with the distributed Bragg reflection film 540.

FIG. 21 shows an example in which the elements of the fifth embodiment are integrated. The elements are separated from each other by trenches deeper than the active layer. This groove is made of insulating polyimide 58
It is embedded with 0 and is flattened. The p-type electrode 511 is individually provided for each element and can be individually controlled. A microlens 590 is provided on the upper part of each element to correct the output light radially emitted from the element into a parallel beam. These microlenses 590 are CC
It is made of a thermosoftening resin such as that used for the condenser lens of the D array, or a photocurable resin. Further, in order to reduce crosstalk between elements, the lenses are separated by grooves.

According to the above fifth to seventh embodiments, the micro-resonator constituted by the spherical distributed Bragg reflection film or the flat metal reflection mirror and the spherical distributed Bragg reflection film facing each other is used. Therefore, the active layer can be kept parallel to the cavity mode equiphase plane. Therefore, as compared with the conventional structure, the ratio of spontaneous emission light coupling with the resonator mode can be increased, and a semiconductor light emitting device having a lower threshold value can be realized.

The present invention is not limited to the above embodiments. In the embodiment, GaAs / AlGaA
Although the s type and the InGaAs / AlGaAs type have been described as examples, the present invention need not be limited to these material types, for example, InP / InGaAsP type and InGaP type.
It can also be widely applied to / InAlP system and the like. Further, in the examples, although it was intended to enhance spontaneous release, it is possible to suppress spontaneous release as well. Further, it can be applied to optical functional elements such as an optical bistable element, an element for finely separating light, and a wavelength conversion element.

Further, in the embodiment, an example using an n-type substrate is shown, but of course a p-type substrate may be used. In this case, the conductivity type of each layer may be reversed. Further, the active layer is not limited to the quantum well structure, and a quantum wire or a quantum box structure can be used. Further, the distributed Bragg reflection film does not need to be limited to semiconductors only, and dielectric,
Alternatively, it may be composed of a combination of a dielectric and a semiconductor. In addition, various modifications may be made without departing from the scope of the present invention.

[0077]

As described above in detail, according to the present invention, at least one of the first distributed Bragg reflector and the second distributed Bragg reflector (or the metal film) sandwiching the double hetero structure portion is provided. , Since it is formed in a convex shape on the side opposite to the double hetero structure portion on the light emitting region of the double hetero structure portion, it is necessary to increase the ratio of spontaneous emission light coupling to the oscillation resonator mode, which was insufficient in the conventional structure. Therefore, it is possible to realize a semiconductor light emitting device that enables an ultra-low threshold value.

[Brief description of drawings]

FIG. 1 is a sectional view showing a schematic configuration of a semiconductor laser according to a first embodiment of the present invention,

FIG. 2 is a plan view of the first embodiment element seen from above,

FIG. 3 is a cross-sectional view showing the manufacturing process of the first embodiment element;

FIG. 4 is a perspective view showing a first modification of the first embodiment,

FIG. 5 is a sectional view showing a second modification of the first embodiment,

FIG. 6 is a cross-sectional view showing a schematic configuration of a second embodiment element,

FIG. 7 is a plan view of the element of the second embodiment as seen from above,

FIG. 8 is a cross-sectional view showing a schematic configuration of a third embodiment element,

FIG. 9 is a plan view of the element of the third embodiment as seen from above,

FIG. 10 is a cross-sectional view showing a schematic configuration of a fourth embodiment element,

FIG. 11 is a schematic diagram for explaining a reflected light path when a plane-concave resonator is used,

FIG. 12 is a perspective view showing a schematic configuration of an element of the fifth embodiment;

13 is a cross-sectional view taken along the line AA ′ in FIG.

FIG. 14 is a schematic view showing a structural example in the vicinity of an active layer of a fifth embodiment element,

FIG. 15 is a schematic diagram showing an outline of a resonator mode in a fifth embodiment,

FIG. 16 is a cross-sectional view showing a schematic configuration of a sixth example element,

FIG. 17 is a schematic view showing a structural example in the vicinity of an active layer of a sixth embodiment element,

FIG. 18 is a schematic diagram showing an outline of a resonator mode in a sixth embodiment,

FIG. 19 is a sectional view showing a schematic configuration of a seventh embodiment element,

FIG. 20 is a cross-sectional view showing the manufacturing process of the device of Example 5;

FIG. 21 is a sectional view showing an example in which the fifth embodiment element is integrated;

FIG. 22 is a schematic diagram showing spontaneous emission in a resonator and light output characteristics of a semiconductor laser;

FIG. 23 is a cross-sectional view showing the structure of a main part of a conventional element.

[Explanation of symbols]

10 ... Substrate, 20, 120, 220, 320 ... N-type distributed Bragg reflector, 40, 140, 240 ... P-type distributed Bragg reflector, 31, 131, 231 ... Quantum well active layer, 32, 33, 132, 133 , 232, 233, 332
... Spacer layer, 50 ... Contact layer, 60 ... Silicon oxide film (mask) 70 ... Quantum well wire active layer, 331 ... Quantum box active layer.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Mitsuhiro Kushibe 1 Komukai Toshiba-cho, Sachi-ku, Kawasaki-shi, Kanagawa Stock company Toshiba Research Institute

Claims (3)

[Claims] 1. A first distributed Bragg reflector in which a plurality of semiconductor films having a large refractive index and a semiconductor film having a small refractive index are alternately laminated, and a second distributed Bragg reflector having the same structure as the reflector. In the semiconductor light emitting device in which a semiconductor double heterostructure is formed, at least one of the distributed Bragg reflectors has a convex shape on the light emitting region of the double heterostructure opposite to the double heterostructure. A semiconductor light-emitting device characterized by being formed in. 2. A semiconductor light emitting device in which a semiconductor double heterostructure is formed between a distributed Bragg reflector and a metal reflector in which a plurality of semiconductor films having a large refractive index and semiconductor films having a small refractive index are alternately laminated. In the device, at least one of the distributed Bragg reflector and the metal reflector is formed in a convex shape on the light emitting region of the double hetero structure portion on the side opposite to the double hetero structure portion. Light emitting element. 3. A first distributed Bragg reflector in which a plurality of semiconductor films having a large refractive index and a semiconductor film having a small refractive index are alternately laminated, and a second distributed Bragg reflector having the same structure as the reflector. In the semiconductor light-emitting device in which a semiconductor double heterostructure part is formed, both of the distributed Bragg reflectors are formed into a curved surface having a recess on the active layer side of the double heterostructure part, and A semiconductor light emitting device, characterized in that the active layer of the heterostructure is located at the focal point of the resonator mode defined by the Bragg reflector.