JPH06237011A - Optical semiconductor element - Google Patents

Abstract

PURPOSE:To realize an optical semiconductor element having high output power by specifying the optical confinement coefficient per quantum of a quantum well active layer, the total optical confinement coefficient of the active layer and the value of the reflectivity of an edge, respectively. CONSTITUTION:When the reflectivity of an edge is close to zero and a laser oscillating mode is not present, the expression I, wherein the gain, which is generated during the propagation in a waveguide is Gs, holds true. In this expression, GAMMA is the optical confinement coefficient, g0 is the gain coefficient, etai is the inner quantum efficiency, J is the density of the injected current, (d) is the thickness of an active layer, alpha is the waveguide-loss coefficient and 1 is the length of the waveguide. At this time, the active layer has the multiple quantum well structure, the inner quantum coefficient etai is increased, the gain coefficient g0 is made large and the confinement per quantum well for the confinement coefficient is made to be 0.05 or less. The gain saturation is not generated in the transition between base levels in this range. As the entire active layer, the single peak condition of the distribution of the light is satisfied when the confinement coefficient is made to be 0.3 or more and 0.6 or less. Thus, the optical semiconductor element having the high output power can be realized.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical semiconductor device, and more particularly to a high power super luminescent diode as a light source for multimode optical fiber communication and optical measurement.

[0002]

2. Description of the Related Art Conventionally, a super luminescent diode (hereinafter referred to as an SLD) has a conventional method of providing a low reflection coating on both end faces of a normal gain waveguide type semiconductor laser, or a stripe width with respect to a normal width of 5 μm. 30-40
There is a method of increasing the output by increasing the width to μm and inclining the angle of the stripe with respect to the resonator end surface by 5 degrees so that the end surface reflectance is 10 ⁻⁴ or less.

For example, FIG. 5 shows a schematic structure of a "high-power super luminescent diode" published in 1988, Gerard et al., Quantum Electronics Division, Journal of the Institute of Electrical and Electronics Engineers, Volume 12, pages 2454-2457 (A. Gerard, et al: "High-
Power SuperluminescentDiodes "IEEE, J,
Quantum, Electrics, vol24, No.12, pp2454-245
7 (1988)).

Reference numeral 1 in the figure is an n-type GaAs substrate.
On this substrate 1, n-type Al _{0.4Ga 0.6} As cladding layer 2, Al _0.06 Ga _0.94 As active layer 3, p-type Al _0.4
A Ga _0.6 As cladding layer 4, an n-type GaAs current blocking layer 5, and an insulating film 6 made of SiO ₂ are sequentially formed. The insulating film 6 is provided with a concave portion (stripe) 6a having an inclination angle (θ) with respect to the longitudinal direction of the substrate 1. A p-side electrode 7 connected to the current block layer 5 via the recess 6a is formed on the insulating film 6. An n-side electrode 8 is formed on the back surface of the substrate 1. In addition, 9 (hatched portion) in the drawing is Z formed in the current block layer 5 and the cladding layer 4 immediately below the stripe.
It is an n diffusion region.

In the diode having such a structure, n ₁
Where is the equivalent refractive index when the propagation constant of light around the active layer immediately below the stripe is converted into a refractive index, and n ₂ is the refractive index of the active layer portion other than the stripe, the tilted stripe 6a
The critical angle θ _c in which the Fabry-Perot mode exists is expressed by the following equation. sin θ _c = {1- (n ₂ / n ₁ ) ² } ^0.5

In the case of the prototype gain-guided laser of Gerrard et al., The refractive index difference is about 5 × 10 ⁻³ ,
θ _c becomes 3.13 degrees. By design, when both end faces are anti-reflection coated, the reflectance ρ (θ) in the stripe direction is 6 × 10 ⁻⁵ when the inclination angle θ of the stripe matches θ _c.
However, the Fabry-Perot mode cannot be completely suppressed. For angles larger than this critical angle θ _c , ρ
(Θ) becomes zero. In their prototype, θ = 5 degrees, resonator length = 400 μm, stripe width 5 μm (effective width is about 30 μm due to carrier diffusion), and maximum output is 2
It is 8 mW, and the SLD is obtained although the laser oscillation mode is superimposed. In other words, the spectrum of light is LED
Like, there is no resonance mode like laser, and I-L
The characteristics of (current and light output) become super linear.

[0007]

However, the conventional technique has the following problems. The output of the SLD when the antireflection coating is provided on both end faces of the conventional semiconductor laser is about 5 mW at maximum. Also,
The SLD proposed by Gerrard et al., In which the stripe direction is inclined 5 degrees with respect to the resonator direction, has a high output of 28 mW at the maximum output, but since the stripe direction is inclined, the far field pattern (FFP: Far Field Pattern) is The pattern is separated from the center by 15 degrees horizontally and vertically, which is a problem in assembling the device. Further, in the case of tilting by 5 degrees, theoretically the Fabry-Perot mode (laser oscillation mode) is completely suppressed, but in reality, the diffusion effect of carriers in the lateral direction with respect to the junction surface, and Since the spectrum spread is remarkably wider than that of the LD, the antireflection coat cannot cover all wavelengths, end face reflection occurs, a laser mode exists, and an FFP side mode exists. In the system using the reflected light, the returned light is mode-converted in the optical element, and the time phase information and the intensity information cannot be used.

The present invention has been made in view of the above circumstances, and the optical confinement coefficient per unit quantum of the quantum well active layer is 0.05 or less, and the optical confinement coefficient of the entire active layer is 0.3 to.
It is an object of the present invention to provide a high-output optical semiconductor device by setting the reflectivity at the end face to be 0.6 or less and 10 ⁻⁴ or less.

[0009]

According to the present invention, a mesa-shaped second layer is formed on a substrate by sequentially interposing a first cladding layer and a quantum well active layer.
An optical semiconductor device of a double-heterojunction type refractive index difference waveguide, in which a clad layer is provided and the second clad layer is embedded with a current block layer,

The optical confinement coefficient per unit quantum of the quantum well active layer is set to 0.05 or less, the optical confinement coefficient of the entire active layer is set to 0.3 or more and 0.6 or less, and the reflectance of the end face is 10 or less. ^-4 is an optical semiconductor device characterized by being set to ⁴ or less.

In the present invention, the quantum well active layer is n
The structure is sandwiched between the optical output waveguide layers doped in the mold (see FIG. 2), or at least the central quantum barrier layer of the multiple quantum well structure is 2 to 5 times as thick as the other quantum barrier layers. A thick p-type or n-type doped structure is used (see FIG. 4), or the active layer is disordered near the end face for higher output, and current injection is suppressed to become a region. Different configuration (Fig. 3
See) is possible.

In the present invention, in order to improve heat dissipation by assembling the joint surface on the heat sink surface side in assembling, the stripe pattern shape is inclined on the back surface with respect to the waveguide direction in the beam emitting direction, and the center of the beam and the assembly center are arranged. Are matched.

[0013]

When the end face reflectance is close to zero and the laser oscillation mode does not exist, the gain generated when propagating in the waveguide is G
_{If s} , then the following equation holds. Gs = exp [Γ {g ₀ · η _i · (J / d) −α} l] (1)

Here, Γ is the optical confinement coefficient, g ₀ is the gain coefficient, η _i is the internal quantum efficiency, J is the injection current density, d is the thickness of the active layer, α is the waveguide loss coefficient, and l (ell) is The waveguide length is shown. If the spontaneous emission light is P _s , the output of the SLD is P _s × G _s . To increase the output, G _{s is set} to 10
^Four It is necessary to be above. In order to increase the value of equation (1) above, the optical confinement coefficient is increased and the internal quantum efficiency η
It is necessary to increase _i and reduce the waveguide loss coefficient. To increase (J / d), increase the injection current,
It is necessary to make the active layer small, but if d is made smaller,
The optical confinement factor Γ decreases. Further, when the injection current density is increased, the device efficiency is smaller than that of the laser, which causes a problem in heat dissipation. From the above viewpoints, the active layer has a multiple quantum well structure, the internal quantum efficiency η _i is increased, the gain coefficient g ₀ is increased, and the confinement coefficient is such that the confinement per single quantum well is 0.
It is set to be 05 or less so that the gain saturation does not occur in the transition between the base levels. The confinement coefficient is 0.
If it is 3 or more and 0.6 or less, the unimodal condition of the light distribution is satisfied. As the number of quantum wells increases, the quantum barrier layer is p-typed for the purpose of efficiently injecting carriers into each quantum well.
By performing type doping, holes having a large effective mass are excessively present in the well to increase the radiative recombination coefficient, or by providing optical output waveguides above and below the active layer to dope n-type, the quantum efficiency is improved. Allow electrons to be injected into the well.

Further, the thickness of the quantum barrier layer in the central portion is set to 2 to 5 times the thickness of the other barrier layers, and n-type doping is performed in this region to efficiently inject electrons into the quantum well.
Since electrons are efficiently injected into the quantum well in this way,
Although the internal quantum efficiency increases and the injection current per unit quantum well is small, the gain coefficient obtained increases, so even if the number of quantum wells is increased, the injection current density does not increase significantly, and light emission between the ground levels occurs. The gain is not saturated and the device efficiency is comparable to that of a laser.

Light generated in the semiconductor layer in the vicinity of the end face is accompanied by a remarkable increase in light density at the end face portion and a temperature rise at the end face due to non-radiative recombination due to dangling bonds of crystals or dislocation vacancies. For the purpose of preventing optical damage (COD) caused by absorption, the window region is formed and the injection current is suppressed from being injected into this region to suppress the non-radiative recombination probability as much as possible.

Further, by using the quantum well structure as the active layer, the waveguide absorption loss coefficient α due to the deformation of the band structure becomes 1/2 to 1/3 of that of the bulk crystal, so that the efficiency is increased. In assembly, when the stripe direction is tilted 5 degrees with respect to the direction perpendicular to the end face, the output beam is n ₁ · sin θ ₁ = n ₂ · sin θ ₂ (2) according to Snell's law.

However, in the above formula (2), n ₁ is the equivalent refractive index in the stripe, θ ₁ is the incident angle, n ₂ is the refractive index of the emitting medium, θ ₂ is the outgoing angle, and n ₁ is about 3 .5, θ ₁ is 5 degrees, n ₂ = 1 and θ ₂ is 17.8 degrees.

Since the emitted beam is inclined 17.8 degrees from the center, a part of the output cannot be taken out from the stem window in the usual semiconductor laser assembling method. After polishing the back surface, use the cleavage surface of the wafer to
By patterning with an inclination of 7.8 degrees, the emitted beam is corrected to the center position using this pattern.

[0020]

Embodiments of the present invention will be described below with reference to the drawings. (Example 1)

Referring to FIG. 11 in the figure is an n-type Ga
It is an As substrate. On this substrate 11, n-type Al _0.45 G
a _0.55 As first cladding layer 12, multiple quantum well active layer 13,
P-type Al _0.45 Ga _0.55 As second cladding layer 14 having a mesa portion 14 a, and p ⁺ Type GaAs cap layer 15 is formed. Here, the multiple quantum well active layer 13 is
With a 10-period structure of aAs quantum well width of 10 nm and Al _0.3 Ga _0.7 As quantum barrier width of 5 nm, the optical confinement per unit quantum well is about 3%, but the optical confinement of the entire active layer is about 50%. . The thickness of both sides 14b of the mesa portion 14a of the second cladding layer 14 is 0.2 to 0.3 μm.
And the stripe width is 3 μm. The second cladding layer 14 is composed of an n-type InGaP first current blocking layer 16 and an n-type GaA layer stacked on the first current blocking layer 16.
s It is embedded by the second current blocking layer 17. A p-side electrode 18 is formed on the cap layer 15, a part of the first current blocking layer 16 and the second current blocking layer 17,
An n-side electrode 19 is formed on the back surface side of 11. In addition, 20 in the figure indicates a back stripe. In addition, this Example 1
In, the striped waveguide is inclined 5 degrees from the vertical direction with respect to the end face.

In the optical semiconductor device having such a structure, the carriers injected from the p-side electrode 18 are the second current blocking layer 17
And the first current blocking layer 16 intensively implants only in the mesa stripe portion. Here, the carriers intensively injected into the mesa stripe portion recombine in the active layer to generate spontaneous emission light. In the first embodiment, since the striped waveguide is inclined 5 degrees from the vertical direction with respect to the end face as described above, the light guided in this waveguide is reflected by the end face and returns to the inside of the waveguide again. It is emitted from the end face without (optical feedback amplification). Here, the light generated by the recombination of carriers spreads horizontally in the horizontal direction with respect to the bonding surface, but since the refractive index of the first current blocking layer 16 is smaller than the mesa portion, the light once spread in the horizontal direction. Is converged to the mesa part. (Example 2)

Referring to FIG. However, the same members as those in FIG. In the figure, 21 and 22 are n-type Al _0.3 formed above and below the multiple quantum well active layer 13, respectively.
Ga _0.7 As first optical output waveguide, n-type Al _0.3 Ga
_0.7 As Second optical output waveguide. The thickness of each of the optical output waveguides 21 and 22 is 50 nm, Si-doped, and the impurity concentration is 1 × 10 ¹⁷ cm ⁻³ . According to the optical semiconductor device having such a structure, the active layer is formed by the existence of the first and second optical output waveguides 21 and 22 formed above and below the multiple quantum well active layer 13, respectively.
Confine the light concentrated in 13 to the first and second optical output waveguides 2
It can be expanded to 1, 22 and relaxed, and higher output and higher efficiency can be achieved as compared with the first embodiment. (Example 3)

Referring to FIG. However, the same members as those in FIG. 31 in the figure is p-type Ga
It is an As substrate. On this substrate 31, p-type Al _0.45 G
a _0.55 As first cladding layer 32, multiple quantum well active layer 33,
N-type Al _0.45 Ga _0.55 As second cladding layer 34 having a mesa portion 34 a, and n ⁺ Type GaAs cap layer 35 is formed. Here, the multiple quantum well active layer 33 is
With a 10-period structure of aAs quantum well width of 10 nm and Al _0.3 Ga _0.7 As quantum barrier width of 5 nm, the optical confinement per unit quantum well is about 3%, but the optical confinement of the entire active layer is about 50%. . The thickness of both sides 34b of the mesa portion 34a of the second cladding layer 34 is 0.2 to 0.3 μm.
And the stripe width is 3 μm. The second clad layer 34 includes a p-type InGaP first current blocking layer 36 and a p-type GaA layer stacked on the first current blocking layer 36.
s It is embedded by the second current blocking layer 37. An n-side electrode 38 is formed on the cap layer 35, a part of the first current blocking layer 36 and the second current blocking layer 37, and the substrate is formed.
A p-side electrode 39 is formed on the back surface side of 31. In the figure, 40 indicates the front stripe and 41 indicates the back stripe.

The cap layer 35, the first and second current blocking layers 36 and 37, and the second cladding layer 34 along the surface stripe 40 have a dose amount of 20 to 30 μm from the end of these layers. 4 × 10 ¹⁴ cm ^-2 Ga ⁺ Ga ⁺ ion-implanted An ion implantation region 42 is formed so as to reach the active layer 33. This Ga ⁺ This is because the ion-implanted region 42 suppresses optical damage due to an increase in light density near the end face and further increases the light output. In the third embodiment, the stripe-shaped waveguide is arranged in the direction 5 perpendicular to the end face.
Being inclined.

According to the third embodiment having such a configuration, the G
a ⁺ Due to the existence of the ion-implanted region 42, the multiple quantum well structure is disordered, and the effective band gap is changed to Al _0.1 Ga.
_Make a mixed crystal of _0.9 As. Therefore, the bandgap of the window region is widened by 90 meV, and the emitted light has a full width at half maximum of 40 meV or less, so that it is not absorbed (the full width at half maximum is narrower than that of the light emitting diode.
Below), the optical output is obtained.

Further, Ga ⁺ Due to the existence of the ion-implanted region 42, the second cladding layer 34 and the cap layer 35 have high resistance (resistivity increases by four digits), and current injection can be suppressed. Furthermore,
It is possible to prevent optical damage caused by heat generation due to non-radiative recombination caused by dangling bonds near the end face of the optical element or lattice defects. (Example 4)

Referring to FIG. In the second embodiment, the case where n-type doping is performed on the first and second optical output waveguides has been described as an efficient electron injection method into each quantum well region of carriers. The fourth embodiment, as shown in FIG.
The central quantum barrier layer is 20 nm, which is four times as large as the other quantum barrier layers, and this region is n-type doped (Si-doped, impurity concentration 1 × 10 ¹⁷ cm ⁻³ ). Thereby, the recombination probability in the quantum well can be increased and the efficiency can be improved. In addition, only the quantum barrier layer was p-typed into the p-type (the front and rear two periods were undoped, carbon and beryllium were 1 × 10 ¹⁸ cm ⁻³ ), and excess acceptor
Efficient recombination with a small amount of injected electrons.

[0029]

As described above, the present invention has the following effects. The active layer is increased compared to the number of quantum wells of a normal semiconductor laser, the optical confinement coefficient in each quantum well is lowered, and gain saturation for injection current is suppressed (this is the same as the semiconductor laser). In order to increase the gain generated by the propagation of the waveguide, the optical confinement of the entire active layer is increased (in the direction opposite to that of the semiconductor laser) to increase the output. In addition, for the purpose of end face protection, current injection is suppressed at the same time as the window structure, aiming at high efficiency and high output. Further, in order to stabilize the gain in each quantum well as the number of quantum wells increases, only the quantum barrier layer is p-type doped, only the optical output waveguide layer is n-type doped, and the central quantum barrier is The layer is 2 to 5 times thicker than the other quantum barrier layers, and is n-type doped to increase the internal quantum efficiency in each quantum well and to increase the operating current associated with the increase in the number of quantum wells. It suppresses the increase. In the above embodiment, the stripe direction is tilted 5 degrees with respect to the vertical direction of the end face to make the reflectance 10 ⁻⁴ or less. However, it is effective with a multilayer dielectric coat and a window structure of the end face in the stripe direction of a normal semiconductor laser. Reflectance of 1
Even if it is ^0-4 or less, the same effect is obtained. According to the present invention, the quantum efficiency can be expected to be 20% or more, which is close to that of a semiconductor laser, and a stable SLD with a maximum output of 50 mW or more can be realized.

[Brief description of drawings]

FIG. 1 is a schematic perspective view of an optical semiconductor device according to a first embodiment of the invention.

FIG. 2 is a schematic perspective view of an optical semiconductor device according to a second embodiment of the present invention.

FIG. 3 is a schematic perspective view of an optical semiconductor device according to a third embodiment of the invention.

FIG. 4 is a band diaphragm diagram of an optical semiconductor device according to a fourth embodiment of the present invention.

FIG. 5 is a schematic perspective view of a conventional high-power superluminescent diode.

[Explanation of symbols]

11, 31, ... GaAs substrate, 12, 14, 32, 34 ...
Cladding layer, 13, 33 ... Multiple quantum well active layer, 15,
35 ... Cap layer, 16, 17, 36, 37 ... Current blocking layer,
18, 39 ... P-side electrode, 19, 38 ... N-side electrode,
21, 22 ... Optical output waveguide layer, 42 ... Ga ⁺ Ion implantation area.

Claims (1)

[Claims] 1. A refractive index difference waveguide type structure in which a mesa-shaped second cladding layer is provided on a substrate through a first cladding layer and a quantum well active layer in this order, and the second cladding layer is embedded with a current blocking layer. In a double heterojunction optical semiconductor device, the optical confinement coefficient per unit quantum of the quantum well active layer is 0.05 or less, and the optical confinement coefficient of the entire active layer is 0.3 or more and 0.6 or less, and An optical semiconductor device having a reflectance of 10 ⁻⁴ or less on an end face.