JPH04367286A - Semiconductor light-emitting element - Google Patents

Abstract

PURPOSE:To provide a double heterotype light-emitting element formed into such a structure that the element has an AlGaAs clad layer, uses an AlGaAs or GaAs layer as an active layer and has a high-speed responsivity while maintaining high light output. CONSTITUTION:The thickness of an AlGaAs or GaAs active layer is formed in a thickness of 0.5mum or thinner. A carrier concentration P2 of the P-type active layer is made higher (n1<P2) than a carrier concentration (n1) of an n-type clad layer. The carrier concentration P2 of the P-type active layer is set in a concentration of 5X10<17>cm<-3> to 3X10<18>cm<-3>.

Description

Description: FIELD OF THE INVENTION This invention relates to an improvement in an AlGaAs double heterostructure light emitting diode. In particular, the purpose is to provide a light-emitting element that is capable of high-speed response and has a large optical output. 2. Description of the Related Art AlGaAs light emitting devices have been in practical use for the longest time. The first double hetero semiconductor laser had a structure in which a GaAs active layer was sandwiched between AlGaAs n-type and p-type cladding layers. There are other semiconductor light emitting devices that use various other mixed crystal semiconductors, but AlGa
The As type has the most proven track record and has a wide range of uses. The problem here is not with lasers but with light emitting diodes. In the case of a laser, it is necessary to have a striped structure because carriers, current, and light must be concentrated in the lateral direction. It is an edge-emitting type, using both end faces as resonators. In the case of a light emitting diode, there is no great need to confine carriers and current in the lateral direction, so a stripe structure is not formed. A double hetero type increases luminous efficiency by confining carriers and light in the thickness direction, so a double hetero type is often used especially for light emitting diodes that require high output. [0004] AlGaAs mixed crystal is Alx Ga1-x
Sometimes it is written in detail, such as As, but sometimes it is abbreviated as the former, omitting the mixed crystal ratio x. AlAs and G
Since the lattice constants of aAs are similar, an AlGaAs mixed crystal thin film of any mixed crystal ratio can be deposited on a GaAs substrate by liquid phase epitaxy, molecular beam epitaxy, or metal-organic pyrolysis method (M
It can be epitaxially grown by OCVD) or the like. The basic double heterostructure consists of a thin active layer
It is sandwiched between a type cladding layer and a p-type cladding layer. It is required that the bandgap of the active layer be narrower than that of the cladding layer, and that the refractive index of the active layer be greater than the refractive index of the cladding layer. The former is a condition for carrier confinement. The latter is a condition for light confinement. AlxGa
In the 1-x As mixed crystal, as x increases, the band gap increases and the refractive index decreases. Therefore, the active layer is made of GaAs.
In this case, the above conditions are automatically satisfied if the cladding layer is made of AlGaAs. Therefore, in both semiconductor lasers and light emitting diodes, the active layer is GaAs and the cladding layer is
It is most often made of AlGaAs, which has a small . Furthermore, since it is desirable that the active layer has few lattice defects, it is often made non-doped. Or use low concentration p-type. However, when the active layer is made of GaAs, the emission wavelength is determined. In order to change the emission wavelength, the active layer itself must be made of a mixed crystal such as Aly Ga1-y As. Even in such a case, if the mixed crystal ratio of the cladding is set to x and the mixed crystal ratio of the active layer is made smaller than this (y<x), the band gap and refractive index conditions described above can be satisfied. I can do it. Nor does the problem of lattice matching between epitaxially grown layers pose any difficulties. [0007] The first requirement of such a light emitting element is that it has high brightness. A light emitting diode having an active layer of AlGaAs is described, for example, in JP-A-64-36089. This makes the active layer non-doped. As already mentioned, it is desirable for the active layer to have few lattice defects, and it is better to have less carrier scattering, so it is generally a non-doped active layer.
It is often This is also in line with conventional thinking. This light emitting device has an AlGaAs active layer sandwiched between p-type AlGaAs and n-type AlGaAs cladding layers having a higher mixed crystal ratio. The carrier concentration of the cladding layer is p1,
When n1, p1 < n1 and 1×1017c
m-3<n1<5×1017 cm-3, 5×1016
The carrier concentration of the cladding layer is set as cm-3<p1<3×1017 cm-3. The reason why the carrier concentration n1 of the n-type cladding is greater than the carrier concentration p1 of the p-type cladding is that a thyristor structure is not generated due to the diffusion of Zn. A p-type cladding layer and an active layer are laminated on a p-type substrate by liquid phase epitaxy. Zn is used as a p-type dopant. Zn has a particularly large diffusion coefficient at high temperatures. Liquid phase epitaxy is continuously performed to grow an n-type cladding layer and an n-type contact layer. For example, Te, Si, etc. are used as the n-type dopant. Liquid phase epitaxy is 700℃
Since the process is carried out at a high temperature of ~800°C, Zn enters the active layer and the n-type cladding layer during this time. This may create a p-type inversion layer above the n-type cladding layer. If that happens pn
This results in a thyristor with a pn structure. To avoid this, the n-type cladding is heavily doped so that p1 < n1
It is said that This creates an inversion layer. Also, n1
, p1 is based on experimental results and seems to have been selected to increase the brightness. The basis is not clear. Although the active layer is non-doped, Zn
Due to the diffusion of , it becomes p-type. Of course, the p-type carrier concentration p2 in the active layer is lower than p1. p2 < p1
<n1. [0009] In the past, light emitting diodes were often used as light sources such as point light sources for displays and printers, and vehicle lamps, and therefore were strongly required to have high brightness. However, light emitting diodes may also be used for signal transmission. Particularly in recent years, the field of application that requires high-speed response performance has expanded, such as for data links and spatial transmission. Conventional high-intensity LEDs have not always been satisfactory in terms of high-speed response. It is an object of the present invention to provide a light emitting device (LED) having high output and high speed response. Means for Solving the Problems The semiconductor light emitting device of the present invention uses a p-type GaAs substrate or a p-type Alx1Ga1-x1A substrate.
s substrate and p-type A epitaxially grown on the substrate.
lx2Ga1-x2As mixed crystal cladding layer and p-type Alx3G epitaxially grown on the p-type cladding layer.
A light emitting device with a double heterostructure including an a1-x3As mixed crystal active layer and an n-type Alx4Ga1-x4As mixed crystal cladding layer epitaxially grown on the active layer, the active layer having a thickness g of 0.5 μm. and the carrier concentration p2 of the p-type active layer is 5×1017 cm-3≦ p
2 ≦3×10 18 cm −3 and is characterized in that the carrier concentration p2 of the p-type active layer is larger than the carrier concentration n1 of the n-type cladding layer (p2 > n1). [Operation] FIG. 1 shows a schematic diagram of a semiconductor light emitting device to which the present invention is directed. This is a GaAs substrate or an AlG substrate in order from the bottom.
aAs grown film, p-type cladding layer (AlGaAs), active layer (AlGaAs or GaAs), n-type cladding layer (A
1GaAs). This is the basic form. Further above the n-type cladding layer, a highly doped n-type contact layer of GaAs is provided, and an AuGeNin side ohmic electrode is attached thereon. A p-side ohmic electrode is attached below the GaAs substrate. First, the thickness of the active layer has been found to have a significant effect on optical output and response time. FIG. 2 is a graph showing the results of the inventor's measurement of variations in optical output by varying the thickness of the active layer. The horizontal axis is the active layer thickness, and the vertical axis is the relative value of optical output. The thinner the active layer, the greater the light output. If it exceeds 1 μm, it will drop to about 70% or less of the value in the case of 0.5 μm. The thickness of the active layer and response time were also investigated. Figure 3 shows the results. The horizontal axis is the active layer thickness, and the vertical axis is the response time. In order to shorten the response time and achieve high-speed response, it is better for the active layer to be thinner. It can be seen that the response time increases as the active layer becomes thicker. For this reason, in the present invention, the active layer thickness is set to 0.5 μm or less. However, there have been a number of light emitting diodes that have active layers with a thickness of 0.5 μm or less, and this is not the only new feature. It has been found that the carrier concentration in the active layer also affects the response time. As already mentioned, conventional active layers have often been non-doped. Because carriers are injected into the active layer by the action of an externally applied current, holes are injected from the p-type layer and electrons from the n-type layer, so the active layer itself does not require its own carriers. be. The electrical resistance must not increase, but since the active layer is thin, the overall resistance hardly increases even if it is non-doped. Furthermore, there is a fear that scattering due to defects caused by doping with impurities will increase. For this reason, the active layer has conventionally been non-doped. Even if it is non-doped, the active layer becomes p-type because Zn diffuses from the p-type layer. The inventor of the present invention wondered what effect the carrier concentration in the active layer would have on the response time of the device. There is a report that the higher the carrier concentration in the active layer, the faster the response speed (NEC Research
h & Development No. 51, p. 6
9 (1978)), but it is not clear what concentration should be used. Therefore, we fabricated a number of light emitting diodes having active layers with different carrier concentrations and investigated their response characteristics. As a result, we found the following. The higher the carrier concentration in the active layer, the shorter the response time. In other words, it provides a fast response. The reason for this is not necessarily clear, but as impurities increase in the active layer, they become trapped, making it easier for transitions between the impurity level and the conduction band to occur, which induces interband transitions, which increases the response speed. I think so. The light emitted by the transition between the acceptor and the conduction band has a slightly shorter wavelength than the light from the interband transition, but this is not very important. However, if the carrier concentration in the active layer is made too high, the crystal properties deteriorate and the number of defects increases, resulting in a decrease in optical output. The carrier concentration of the active layer is 5×10, which can maintain high performance in both optical output and response speed.
It was found to be 17 cm-3 to 3 x 1018 cm-3. This is a fairly high doping level. In the above-mentioned Japanese Patent Application Laid-Open No. 64-36089, the active layer is non-doped and the carriers in the cladding layer are defined, but the value is one order of magnitude lower than the value of the present invention. After growing the active layer, an n-type cladding layer is subsequently grown at a high temperature. During this time, impurities in the active layer and the p-type cladding are thermally diffused and enter the n-type cladding. Zn is often used as a p-type impurity, and Zn is particularly easy to diffuse. Furthermore, n-type impurities may diffuse from the n-type cladding into the active layer. P-type impurity (Zn) from the p-type region to the n-type region
) diffusion is well known and can lead to the parasitic generation of thyristor structures. In order to prevent this, the above-mentioned Japanese Patent Application Laid-Open No. 64-36089 sets p1 < n1. In other words, if the carrier concentration in the p-type cladding is made lower than the carrier concentration in the n-type cladding, even if Zn is diffused, n-type carriers will be dominant and no inversion layer will be formed. The present invention does not take such a phenomenon into consideration, but instead transforms the active layer (p2) from the n-type cladding (n1) to the active layer (p2).
The problem is diffusion in the opposite direction. If n1 > p2
If the n-type impurity diffuses vigorously, the dopant in the active layer is canceled out, and p-type carriers are effectively lost. On the other hand, if n1 <p2, no matter how much diffusion occurs, all the dopants in the active layer will not be canceled out. Therefore, in the present invention, n1
The p-type dopant concentration in the active layer is increased so that <p2. This point is a very unique feature of the present invention. The conditions characterizing the present invention can be abbreviated as follows. EXAMPLE Five types of AlGaAs mixed crystal light emitting diodes were fabricated by liquid phase epitaxy on a GaAs substrate with different parameters. The light output and response time of each sample were then measured. The mixed crystal ratio of the p-type cladding layer (x of AlxGa1-x As) is 0.35 to 0.1
Similarly, the mixed crystal ratio of the n-type cladding layer is 0.35 to 0.
．． It is 1. Since the active layer is also made of AlGaAs mixed crystal, the mixed crystal ratio is 0.07. Since the mixed crystal ratio is low, the bandgap of the active layer is narrower than that of the cladding layer. The refractive index of the active layer is also higher than that of the cladding layer. The emission wavelength is 830 nm. The p-type dopant is Zn,
Te was used as the n-type dopant. The thickness g of the active layer was varied within the range of 0.3 to 0.8 μm. The carrier concentration p2 in the active layer is 3×
It was varied between 1017 cm-3 and 4×1018 cm-3. The carrier concentration n of the n-type cladding layer is 2×1017
It was varied between cm-3 and 9 x 1017 cm-3. A summary of the elements is listed below. Growth method: Liquid phase epitaxial method Emission wavelength: 830nm

material
: p-type cladding layer x=0.35~0
．． 1

(Alx Ga1-x As)

Active layer x=0.07
N-type cladding layer x=0.35
~0.1 Dopant: p type
Zn

n-type Te

Table 1 shows the parameters, light output, response time, and other measurement results for the five types of samples. ■ is an example of the present invention, and ■~■
is a comparative example. [0023] The optical output and response time are shown as relative values with respect to the results for the embodiments of the present invention as 1. The higher the light output, the better, and the shorter the response time, the better. Comparative example ■ has an active layer thickness of 0.8 μm and g≦0.5.
The condition of μm is not satisfied. This is undesirable because the optical output is low and the response time is long. Since Comparative Example (2) and Example (2) differ only in the thickness of the active layer, it is clear that the thickness of the active layer has a strong influence on the optical output and response time. Comparative Example (3) has a thin active layer, but the carrier concentration p2 of the active layer is within the range defined by the present invention (1×
1017 cm-3 to 3 x 1018 cm-3). The condition n1 < p2 is satisfied. Although this element has an optical output of 1.2 and higher brightness, the response speed is significantly lower. This is a comparative example that has a low doping concentration and is closest to a conventional light emitting diode. In Comparative Example (2), on the contrary, the carrier concentration in the active layer was increased. This is excellent in that it shortens response time. However, the drawback is that the light output is weak. Comparative example ■ has n1 > p2,
This does not satisfy the condition n1 < p2 of the present invention. This increases the response time by 1.8 times, resulting in poor response characteristics. The light output is 1.1 times higher and is superior in terms of high brightness. As described above, the optical device of the present invention is excellent in both optical output and response time. Effects of the Invention The light emitting device of the present invention has a p-type active layer doped with
Since the punt concentration is made high (p2 > n1), the response time can be shortened while keeping the optical output high to some extent. Since it can be made into a light-emitting element with high-speed response, it is ideal for light-emitting elements for data links, spatial transmission, etc.

[Brief explanation of the drawing]

FIG. 1 is a schematic configuration diagram of a light emitting device of the present invention.

FIG. 2 is a graph showing measurement results of the active layer thickness and optical output of an AlGaAs light emitting device.

FIG. 3 is a graph showing measurement results of the active layer thickness and response time of an AlGaAs light emitting device.

Claims (1)

[Claims] [Claim 1] P-type GaAs substrate or p-type Alx1G
an a1-x1As substrate, a p-type Alx2Ga1-x2As mixed crystal cladding layer epitaxially grown on the substrate, and a p-type Alx2Ga1-x2As mixed crystal cladding layer epitaxially grown on the p-type cladding layer.
type Alx3Ga1-x3As mixed crystal active layer and n-type Alx4Ga1-x epitaxially grown on the active layer.
4As mixed crystal cladding layer, the thickness g of the active layer is 0.5 μm or less, and the carrier concentration p2 of the p-type active layer is 5×1017 cm-3≦p2≦3. ×1018cm
-3, and has a carrier concentration p2 of a p-type active layer larger than a carrier concentration n1 of an n-type cladding layer (p2 > n1).