JPH04282875A - Light emitting diode - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a double-hetero type light emitting diode formed on a Zn-doped p-type InP substrate.

0002

[Prior Art] Light emitting diodes are manufactured using III-V compound semiconductor substrates such as GaAs and InP.
It is made by epitaxially growing multiple layers of group mixed crystal semiconductors. A well-known light emitting diode is n-type G.
A double heterostructure is formed by epitaxially growing a thin film of AlGaAs mixed crystal and GaAs on an aAs substrate. An n-type cladding layer, a non-doped (or low concentration p-type) active layer, a p-type cladding layer, a p-type contact layer, etc. are laminated on an n-type substrate. The n-type has a lower specific resistance than the p-type, and the n-type substrate was more convenient for reducing the resistance caused by the substrate. Also, it is easier to form ohmic junction electrodes on n-type substrates.

[0003] GaAs-based light emitting elements have a thickness of 0.7 to 0.8μ.
It emits light with a wavelength of m, but in order to generate light with a longer wavelength of about 1.3 μm, an InP-based light emitting element is used. Engineers who originally developed GaAs-based light emitting devices developed InP.
Since we decided to proceed with the development of light-emitting devices based on n-type In
Thin films of InGaAsP and InP were laminated on a P substrate to form a pn junction. Even now, light emitting diodes using n-type InP substrates are mainstream.

0004

Problems to be Solved by the Invention Unlike the case of n-type InP substrates, new problems arise with p-type InP substrates. Although there is active research on buried layers and active layers, sufficient consideration has not been given to the substrate itself. Since electron mobility is 20 to 30 times higher than hole mobility, n-type substrates have low resistance and require a small amount of dopant. However, if a p-type InP substrate is used, it is necessary to add a large amount of dopants to make it low resistance. Examples of p-type dopants for InP include Zn, Mg, and Be. Here, we will focus on those using Zn as a dopant. Since the substrate is much thicker than the epitaxial film, it must have a particularly low resistance.

[0005] In the case of a light emitting diode, since light is not emitted from the substrate, there should be no problem of light absorption, and it was thought that the only problem was the electrical characteristics of the substrate. For this reason, substrate characteristics have been evaluated solely by carrier concentration. When Zn is doped, the characteristics of the substrate are specified, such as the carrier concentration. For example, an InP substrate with a carrier concentration of 5×10 18 cm −3 is specified. The carrier concentration can be easily measured by Hall measurement, so it is easy to confirm. However, even if the carrier concentration is similarly defined, the characteristics of light emitting diodes fabricated on the same substrate vary. Some emit light well, while others have low luminous efficiency. Although this is not the case with ordinary light emitting diodes that emit light from the epitaxial layer, the variation in characteristics is large in the case of light emitting diodes that emit light from the substrate side.

[0006] This means that there are unknown variables that are not controlled. After repeating various experiments, the present inventor found that even if the carrier concentration (hole concentration) is constant, the concentration of Zn, which is a dopant, often varies. Why is this? Conventionally, it was thought that the dopant concentration and carrier concentration were the same in InP as in the case of Si doping. However, when an InP substrate is doped with Zn, the dopant concentration and the carrier concentration are the same up to a certain point, but beyond this point, the carrier concentration becomes saturated. This was discovered for the first time by the inventor.

Of course, anyone could have expected that if a large amount of dopants were added, they would not be activated and some would not release holes, but this is true for the substrates actually used. It was thought that there would be no problem with the dopant concentration. However, that was not the case. Saturation of the carrier concentration actually occurs for low dopant concentrations that dope the substrate. Conventionally, it has been thought that carrier concentration alone is sufficient to specify the characteristics of a substrate, so the actual Zn concentration has not been measured. Another reason was that it would be difficult to carry out because it involves destructive testing such as atomic emission spectroscopy and SIMS.

Therefore, the present inventor fabricated various p-type InP substrates by changing the amount of Zn doped, and measured the dopant concentration in addition to the carrier concentration. As a result, we found that something unexpected was happening.

FIG. 1 shows the Z
The measurement results of Zn atomic concentration and carrier (hole) concentration in an n-doped InP single crystal are shown. The horizontal axis represents the Zn atomic concentration, which is measured by atomic emission spectrometry. The vertical axis represents the hole concentration of the same single crystal, measured by Hall measurement. Zn atomic concentration is 2
．． In the case of 4 x 1018 cm-3, 3.8 x 1018 cm-3, the hole concentration is 2.4 x 1018 cm-3, 3.7
×1018 cm-3.

However, the Zn concentration is 4×1018 cm−3
If the value exceeds , the hole concentration will deviate greatly from the Zn concentration. When the Zn concentration is 4.8 x 1018 cm-3, the hole concentration (carrier concentration) is 4.4 x 1018 cm-3, and the latter is lower by 0.4 x 1018 cm-3. Z
When the n concentration is 5.4 x 1018 cm-3, the hole concentration is 4.8 x 1018 cm-3. In this case 0.6×1
There is a difference of 0.018 cm-3. Zn concentration is 7.4×101
In the case of 8 cm-3, the carrier concentration is 5.1 x 1018
cm-3, and there is a difference of 2.3×1018 cm-3.

[0011] When looking at this only from the carrier concentration,
p=4.4×1018cm-3, p=5.1×1018
cm-3, the carrier concentration is 0.7 x 1018 cm
Although the difference is only -3, the Zn atomic concentration is 4.8×101
8cm-3 and 7.4x1018cm-3, which means that the latter contains 1.5 times as much Zn as the former.

[0012] Therefore, the present inventor has determined that [carrier concentration]/[Z
n concentration] is termed the activation rate, and the activation rate is shown in FIG. 2 as a function of the Zn atom concentration for the same data as in FIG. When the Zn concentration is below 3 x 1018 cm-3, the activation rate is 1.
It is 00%. When the Zn concentration is 3 to 5 x 1018 cm-3, the activation rate is about 97 to 95%. However, when the Zn concentration exceeds 5.times.10.sup.18 cm.sup.-3, the activation rate rapidly decreases. When [Zn] = 5.4 x 1018 cm-3, the activation rate is 85%, [Zn] = 6.2 x 1018 cm-3
When , the activation rate is 77%, [Zn] = 7.4 × 1018
At cm-3, the activation rate is 67%.

The reason for this is thought to be as follows. Zn
When the concentration becomes low, Zn does not replace In sites in the InP lattice, but instead occupies interatomic positions. In this case, it is considered that Zn does not have the effect of attracting electrons and does not generate a single hole. In other words, Zn becomes dormant and does not generate carriers. It is fine if the Zn is just sleeping, but if Zn is present in the substrate at a high concentration, it absorbs light. Particularly in the case of a light emitting diode that extracts light from the substrate side, the absorption of light by Zn atoms cannot be ignored. For example, even if the substrates are selected by specifying p=5×10 18 cm −3 , some will have a high Zn concentration and some will have a low Zn concentration, and if the Zn concentration is high, the light absorption will be large and the luminous efficiency will be reduced.

What can be seen from these facts is that in order to specify the quality of a Zn-doped p-type InP substrate, not only the carrier concentration (p) must be used, but also the Zn concentration must be used. . Another is that the Zn concentration should not be too high. The Zn concentration must be limited to an appropriate range in order to prevent light absorption by Zn and to lower the electrical resistance. Based on these discoveries, we developed a light emitting diode formed on a p-InP substrate with Zn as a dopant, which operates well even at high temperatures due to less absorption of light by Zn. It is an object of the present invention to provide.

[0015]

Means for Solving the Problems The light emitting diode of the present invention includes a p-InP first cladding layer, a p-GaInAsP active layer, an n-InP second cladding layer, and an n-InP first cladding layer on a p-InP substrate. - A light emitting diode having a GaInAsP contact layer, characterized in that the atomic concentration of zinc, which is a p-type dopant in the p-InP substrate, is 3 to 7×10 18 cm −3 . More preferably, the atomic concentration of zinc, which is a p-type dopant in the p-InP substrate, is set to 4 to 5×1.
018 cm-3.

To define other structures in more detail, the light emitting diode of the present invention has a p-I layer on a p-InP substrate.
nP first cladding layer, p-GaInAsP active layer, n-
In a light emitting diode having an InP second cladding layer and an n-GaInAsP contact layer, and having electrodes on the p-InP substrate and the n-GaInAs contact layer, the p-type
- The atomic concentration of zinc, which is the p-type dopant of the InP substrate, is 3 to 7 x 1018 cm-3, the thickness of the p-InP first cladding layer is 1.5 μm or less, and the p-type carrier concentration is 2 x
1017~1×1018 cm-3, p-GaIn
The AsP active layer has a thickness of 0.1 to 2.0 μm and a p-type carrier concentration of 2 x 1017 to 3 x 10 cm,
The n-type carrier concentration of the n-InP second cladding layer is 2×1
018 cm-3 or less, and the position of the p-n junction is p-
It is characterized by being located at the interface between the GaInAsP active layer and the n-InP second cladding layer.

[0017]

[Function] In the light emitting diode of the present invention, p-type InP
Rather than specifying the hole concentration in the substrate, the Zn concentration is
~7x1018cm-3. As shown in FIG. 2, when the Zn concentration is 3 to 7×10 18 cm −3 , the activation rate is 70% or more. Also, the carrier concentration is 4.9
x1018 cm-3 or less. 7×1018cm-3
This is the upper limit at which Zn does not excessively increase light absorption in the substrate. 3×10 18 cm −3 is the lower limit to prevent the electrical resistance of the p-type substrate from becoming too high. Therefore, the substrate used in the present invention is an optimal substrate with low electrical resistance and little absorption of light. As already mentioned, even at a fairly low doping amount, Zn
The equation of [Zn concentration] = [hole concentration], which was vaguely believed in the doped InP substrate, does not hold. Even if it is defined based on the hole concentration, it does not mean that the Zn concentration is correctly defined.

Conventional p-type (Zn-doped) InP substrates have low hole mobility (1/20 to 1/30 of electrons) and tend to have high resistance, so it was attempted to lower the resistance by doping as much Zn as possible. was. For example, p=5.0×1018cm-3
A p-type InP substrate is used. If the manufacturing method is the same as the inventor's method, the Zn concentration will be 7.2×
This corresponds to 1018 cm-3 or more. Since the hole concentration is saturated above this, the Zn concentration is 8 to 10 x 1018 cm-3.
Maybe it is. In this way, when the Zn concentration is high, Z
This is undesirable because a large amount of light is absorbed by n, which deteriorates the light emitting characteristics.

[0019]

[Example] A light emitting diode was fabricated using a wafer as a substrate, which was obtained by thinly slicing a p-type InP single crystal doped with Zn and pulled using the liquid-encapsulated Czochralski method (LEC), and then etching, polishing, etc. . In order to examine the effects of the present invention, substrates with different Zn concentrations were prepared.

Double hetero type light emitting diodes were fabricated on five types of p-type InP substrates having different Zn concentrations by liquid phase epitaxy. Figure 3 shows the structure of a light emitting diode. This is a device that emits light from the substrate and is not widely used. In this way, the light emitting part and the tip of the optical fiber (not shown) can be brought very close to each other, resulting in good incidence efficiency. The substrate is Zn-doped
-InP. On top of this is a p-InP cladding layer, an InP cladding layer,
GaAsP active layer, n-InP cladding layer, n-InG
An aAsP contact layer is provided by epitaxial growth as described below.

[0021] p-InP cladding layer (first cladding layer) Zn concentration 6 to 7 x 1017 cm-3, thickness 0
．． 5μm ■ p-InGaAsP active layer (In0.73Ga0.27As0.57P0.43)
Zn concentration 1 to 2 x 1018 cm-3, thickness 1.
5μm ■ n-InP cladding layer (second cladding layer) Si concentration 8 to 10 x 1017cm-3
, 0.5 μm thick n-InGaAsP contact layer, 0.5 μm thick (In0.79Ga0.21As0
．． 44P0.56). The epitaxial layer is removed on both sides in a mesa shape to concentrate current and light. The p-type electrode on the bottom surface of the substrate is a ring electrode. This is to allow light to pass through the center. The p-side electrode portion is fixed in the hole of the package by a solder layer. The light emitted from the active layer passes through the substrate and is emitted to the outside through the hole in the package. Therefore, the absorption of light by the substrate must be reduced.

[0022] It is generally desirable that the p-InP first cladding layer (①) has a thickness of 1.5 μm or less. The carrier concentration is 2 x 1017 cm-3 to 1 x 1018 cm-3. The p-GaInAsP active layer of (2) has a p-type carrier concentration of 2 x 1017 cm-3 to 3 x 1018 c.
m-3, and the thickness is 0.1 to 2.0 μm. ■n
-InP second cladding layer has n-type carrier concentration of 2×10
It should be 18cm-3 or less. The Zn concentration of the p-type InP substrate is (a) 2 x 1018 cm-3 (b) 3 x
1018cm-3 (c) 4.5×1018cm-3
(d) 6.5 x 1018 cm-3 (e) 7.5
x1018 cm-3.

The ambient temperature was varied from 20° C. to 80° C., and the current and light output characteristics at each temperature were measured. Among these, the one with the best properties is (c) [Zn] = 4.5×
It was a light emitting diode using a 1018 cm-3 substrate. The next best is (d) [Zn] = 6.5 x 1018 cm-
This is a light emitting diode using a substrate of No. 3, and the third best one is one using a substrate of (b) [Zn] = 3 x 1018 cm-3. The fourth one is (a) [Zn] = 2×
Of the light emitting diodes fabricated on a 1018 cm-3 substrate, the one with (e) [Zn]=7.5×1018 cm-3 was the worst. This is considered to be due to an increase in light absorption by Zn in the substrate. From these results, it can be seen that it is best for the Zn concentration of the p-type InP substrate to be 4 to 5 x 1018 cm-3, and that sufficient performance can be achieved if it is 3 to 7 x 1018 cm-3.

Although this is a type of light emitting diode that emits light from the substrate side, the present invention can also be applied to ordinary type light emitting diodes that emit light from the epitaxial layer side as shown in FIG. can do. This involves growing a p-InP cladding layer, an n-InGaAsP active layer, an n-InP cladding layer, and an n-InGaAsP contact layer on a p-InP substrate, attaching a p-side electrode to the entire surface of the substrate, and
-A ring-shaped n-side electrode is attached to the InGaAsP contact layer. Light emerges from the transparent insulating layer.

[0025]

[Effect of the invention] Light-emitting diodes are manufactured using p-type InP substrates, but since quality evaluation was conventionally performed based on carrier concentration, hidden parameters were overlooked, and devices with certain characteristics could be manufactured with good reproducibility. That was difficult. In the present invention, when an InP substrate is doped with Zn above a certain level, the carrier concentration becomes saturated with respect to the Zn concentration, and even if the carrier concentration is specified near the saturation concentration, the concentration of Zn atoms is unique. It has been made clear that this is not specified. Furthermore, since the Zn concentration is 3 to 7.times.10.sup.18 cm.sup.-3, the absorption of light by Zn is not large, and the electric resistance of the substrate is also small, making it possible to provide a light emitting diode with extremely high luminous efficiency.

[Brief explanation of the drawing]

FIG. 1 is a graph showing measured values of Zn atomic concentration and carrier concentration doped into an InP substrate.

FIG. 2 is a graph showing measured values of Zn atom concentration and activation rate when an InP substrate is doped with Zn.

FIG. 3 is a schematic cross-sectional view of a light emitting diode using a p-type InP substrate that emits light from the substrate and mounted in a package.

FIG. 4 is a schematic cross-sectional view of a light emitting diode using a p-type InP substrate in which light is emitted from an epitaxial layer.

Claims (4)

[Claims] 1. On a p-InP substrate, a p-InP first cladding layer, a p-GaInAsP active layer, an n-InP
In a double heterostructure light emitting diode having a second cladding layer and an n-GaInAsP contact layer,
A light emitting diode characterized in that the p-InP substrate has an atomic concentration of zinc, which is a p-type dopant, of 3 to 7 x 1018 cm-3. 2. On a p-InP substrate, a p-InP first cladding layer, a p-GaInAsP active layer, an n-InP
In a double heterostructure light emitting diode having a second cladding layer and an n-GaInAsP contact layer,
A light emitting diode characterized in that the atomic concentration of zinc, which is a p-type dopant, in a p-InP substrate is 4 to 5 x 1018 cm-3. 3. On the p-InP substrate, a p-InP first cladding layer, a p-GaInAsP active layer, an n-InP
In a double-hetero type light emitting diode having a second cladding layer and an n-GaInAsP contact layer, and electrodes on the p-InP substrate and the n-GaInAs contact layer, the p-type dopant of the p-InP substrate The atomic concentration of a certain zinc is 3 to 7 x 1018 cm-3, the thickness of the p-InP first cladding layer is 1.5 μm or less, the p-type carrier concentration is 2 x 1017 to 1 x 1018 cm-3, and p-
When the thickness of the GaInAsP active layer is 0.1 to 2.0 μm, p
Mold carrier concentration is 2 x 1017 to 3 x 1018 cm-3
and the n-type carrier concentration of the n-InP second cladding layer is 2 x 1018 cm-3 or less, and the pn junction is located at the interface between the p-GaInAsP active layer and the n-InP second cladding layer. A light emitting diode characterized by: 4. On the p-InP substrate, a p-InP first cladding layer, a p-GaInAsP active layer, an n-InP
In a light emitting diode having a second cladding layer and an n-GaInAsP contact layer, and having electrodes on the p-InP substrate and the n-GaInAs contact layer, the p-InP
The atomic concentration of zinc, which is a p-type dopant in the substrate, is 3 to 7
1018 cm-3, the thickness of the p-InP first cladding layer is 1.5 μm or less, and the atomic concentration of zinc is 2×1017~
1×1018 cm−3, the thickness of the p-GaInAsP active layer is 0.1 to 2.0 μm, and the atomic concentration of zinc is 2×
1017~3×1018 cm-3 and n-InP second
The n-type carrier concentration in the cladding layer is 2 x 1018 cm-3
and the position of the p-n junction is p-GaInAsP
A light emitting diode characterized in that it is located at an interface between an active layer and an n-InP second cladding layer.