JP3406907B2 - Semiconductor light emitting diode - Google Patents

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 diode using a compound semiconductor material, and more particularly to a semiconductor light emitting diode using an InGaAlP-based material for an active layer.

[0002]

2. Description of the Related Art LEDs (Light Emitting Diodes) are used in various fields such as optical communication and optical information processing as a light source with low power consumption, high efficiency and high reliability.
Particularly, in the visible wavelength region, GaP is used as a light emitting layer material.
(Green), GaAsP (yellow, orange, red), GaAlA
Compound semiconductors such as s (red) are used.

However, GaP and GaAsP are indirect transition type semiconductors, and the luminous efficiency thereof is extremely low at about 0.5% or less even if the influence of absorption is eliminated by using a transparent substrate. Further, although GaAlAs has an efficiency of about 8% at 660 nm, the efficiency at 635 nm is about 1% at wavelengths shorter than this due to the influence of indirect transition.

Considering the visibility of the human eye, GaAl
In As system, 3 cd (candela) at 660 nm, 635 n
It corresponds to 1 cd in m. On the other hand, GaP has realized only 0.5 cd or less and GaAsP has 0.3 cd or less, and high-intensity LEDs in the orange, yellow, and green regions.
Was strongly desired.

InGaAlP mixed crystals are nitrides except III.
It has the largest direct transition type energy gap in a mixed crystal of a group-V compound semiconductor, and is attracting attention as a light emitting device material in the 0.5 to 0.6 μm band. In particular, an InGaAlP LED that uses GaAs as a substrate and is lattice-matched to the substrate has a possibility of emitting light of high brightness from green to red. However, even with this type of LED, it cannot be said that the luminous efficiency in the short wavelength region is necessarily sufficiently high.

FIG. 30 is a sectional view showing the device structure of a conventional LED having an InGaAlP light emitting portion. In the figure, 1 is n-G
aAs substrate, 2 is n-InGaAlP cladding layer, 3 is InGaAlP active layer, 4 is p-InGaAlP cladding layer, 5 is p-GaAlAs current spreading layer, and 6 is p-Ga.
As contact layer, 7 is a p-side electrode made of AuZn, 8
Is an n-side electrode made of AuGe.

The mixed crystal composition is set so that the energy gap of the InGaAlP active layer 3 is smaller than that of the clad layers 2 and 4, and the active layer 3 receives light and carriers.
It has a double hetero structure that is confined in. Also, p
-The composition of the GaAlAs current diffusion layer 5 is InGaAlP.
It is set to be substantially transparent to the wavelength of light emitted from the active layer 3.

In the structure of FIG. 30, the active layer 3 is formed of undoped In _0.5 (Ga _{1 -X} Al _X ) with a thickness of 0.2 μm.
_When the conductivity was _0.5 P (x = 0.4), the conductivity type was n-type and the concentration was about 1 to 5 × 10 ¹⁶ cm ⁻³ . At this time, the emission wavelength is 565 nm (green) and the emission efficiency is DC2.
It was about 0.07% at 0 mA. When x = 0.3, the emission wavelength is 585 nm (yellow) and the emission efficiency is D.
C20mA is as low as 0.4% and GaP, GaAsP
The characteristic merit to the system was not always found. On the other hand, when x = 0.2, the emission wavelength is 620n.
m (orange), the light emission efficiency was about 1.5% at 20 mA DC, and the light emission efficiency higher than that of GaAlAs system was obtained without particularly removing the GaAs substrate 1 serving as an absorber for the emission wavelength.

As a result of intensive studies made by the present inventors on the cause of such composition dependence of luminous efficiency, it was found that there is a strong correlation with the mobility in the n-conduction type. That is, the mobility dropped sharply when the Al composition x was about 0.3. It was also observed that the donor level would be deepened accordingly. It is considered that the generation of deep levels in the n-conductivity type causes non-radiative recombination, which leads to a decrease in luminous efficiency. Further, when the Al composition x of the active layer 3 is high and the emission wavelength is short, the energy gap difference between the active layers 3 and the cladding layers 2 and 4 cannot be sufficiently secured, and the electrons having a particularly small effective mass are transferred to the p-type cladding layer 4. It was also considered that the overflow became remarkable, which led to a decrease in luminous efficiency.

On the other hand, in the structure of FIG. 30, the current diffused layer 5 is provided to spread the current injected from the electrode 7. However, when the current diffused layer 5 is not provided, the spread of the injected current is increased due to the following reasons. It becomes smaller, and the light emitting region is only in the vicinity immediately below the electrode. That is, the Al composition of the cladding layers 2 and 4 is the same as that of the active layer 3
It is necessary to make it sufficiently larger than that of the active layer 3 in order to have a band gap difference with In the p-cladding layer 4, if the Al composition is large, the carrier concentration cannot be increased and the resistance becomes large. Therefore, the spread of the current in the p-clad layer 4 is small, and the light emitting region is only in the vicinity immediately below the electrode. In this case, in the structure in which the light is extracted from above, the electrode 7 blocks the light from the light emitting region, resulting in a decrease in the light extraction efficiency.

Further, in the semiconductor light emitting device using the InGaAlP material, if the Al composition of the active layer is increased in order to obtain light emission of a shorter wavelength, the non-radiative centers of the active layer increase and non-radiative recombination occurs. However, there is a problem in that the luminous efficiency is lowered due to the increase.

[0012]

As described above, the conventional I
It has been difficult to obtain highly efficient light emission particularly in yellow and green in the nGaAlP-based LED. In addition, current concentration under the electrode in the light emitting portion occurs, the light extraction efficiency is lowered, and non-radiative recombination in the active layer is increased, so that it is difficult to realize high brightness.

The present invention has been made in consideration of the above circumstances, and an object thereof is to provide a semiconductor light emitting diode capable of emitting light with high efficiency even in a short wavelength region such as yellow and green. .

[0014]

(Structure) In order to solve the above problems, the present invention adopts the following structure.

That is, the present invention provides a GaP substrate, a double heterostructure portion formed on the substrate, in which an active layer made of an InGaAlP material is sandwiched between cladding layers made of an InGaAlP material, and a double heterostructure portion on the double heterostructure portion. And a second electrode selectively formed on a side of the substrate opposite to the double heterostructure portion, the semiconductor light emitting device extracting light from the second electrode side. Die
The active layer has a carrier concentration of 1 × 10 ¹⁷
cm ⁻³ or less, and the active layer has a thickness of 0.15.
To 0.75 μm, the double hetero structure
Of the clad layers forming the part, the p-type clad layer
Range of the rear concentration ^{5 × 10 17 cm -3 ~2 ×} 10 18 cm -3
It is characterized by being set to .

Here, the following are preferred embodiments of the present invention.

[0017]

(2) Of the cladding layers forming the double hetero structure, the carrier concentration of the n-type cladding layer is 1 × 10.
^It must be set in the range of ¹⁶ cm ^{-3 to} 7 × 10 ¹⁷ cm ^-3 .

(3) The active layer is made of InGa having a different Al composition.
A multiple quantum well structure composed of a well layer made of an AlP material and a barrier layer.

(4) A current diffusion layer made of GaAlAs is formed between the double hetero structure and the first electrode, and the film thickness of this current diffusion layer is set in the range of 5 to 30 μm, and the carrier concentration is set. Is set in the range of 5 × 10 ¹⁷ cm ^{−3 to} 5 × 10 ¹⁸ cm ⁻³ .

(5) A current diffusion layer is formed between the double hetero structure and the first electrode so that light can be extracted also from the first electrode side.

(6) The thickness of the current diffusion layer is set in the range of 5 to 30 μm, and the carrier concentration is 5 × 10 ¹⁷ cm ^{−3 to} 5 × 1.
It ^must be set within the range of 0 ¹⁸ cm ^-3 .

(Operation) According to the present invention, since GaP is used as the element formation substrate and InGaAlP is used as the active layer, light from the active layer is not absorbed by the substrate, and light is emitted from the substrate side. It can be effectively taken out. In addition, an active layer made of InGaAlP is provided with a low concentration p-type (carrier concentration 1 × 10 ¹⁷ cm ⁻³ or less p-type) or a low concentration n-type (carrier concentration 5 × 10 ¹⁶ cm ⁻³ or less n-type).
By so doing, it becomes possible to reduce the influence of the non-radiative recombination described above. In addition, the thickness of the active layer is optimized (0.15 to 0.75 μm), and the carrier concentration of the p-type cladding layer is optimized (5 × 10 ¹⁷ cm ⁻³ to 2 × 1).
0 ¹⁸ cm ^-3 ) makes it possible to obtain highly efficient light emission.

[0024]

DETAILED DESCRIPTION OF THE INVENTION The details of the present invention will be described below with reference to the illustrated embodiments.

(First Embodiment) FIG. 1 shows a first embodiment of the present invention.
3 is a cross-sectional view showing a schematic structure of a semiconductor light emitting diode according to the embodiment of FIG. In the figure, 11 is an n-GaAs substrate, and on this substrate 11, an n-InGaAlP clad layer 12, a p-InGaAlP active layer 13 and a p-InGa.
A double heterostructure portion composed of the AlP cladding layer 14 is grown and formed. P-Ga is formed on the double hetero structure.
An AlAs current diffusion layer 15 is grown and formed, and a GaAs contact layer 16 is grown and formed on a part of the current diffusion layer 15. Then, the p-side electrode 17 made of AuZn is formed on the contact layer 16, and Au is formed on the lower surface of the substrate 11.
An n-side electrode 18 made of Ge is formed.

The mixed crystal composition is set so that the energy gap of the p-InGaAlP active layer 13 is smaller than that of the cladding layers 12 and 14, and a double hetero structure for confining light and carriers in the active layer 13 is formed. There is. The composition of the p-GaAlAs current diffusion layer 15 is set to be substantially transparent to the emission wavelength from the active layer 13. Although the substrate side has the n-type conductivity type in the present embodiment, the same concept can be applied to a structure in which the conductivity type is reversed.

In the structure shown in FIG. 1, the thickness of each layer,
The carrier concentration is set as follows. n-Ga
As substrate 11 (250 μm, 3 × 10¹⁸cm^-3), N-
InGaAlP clad layer 12 (1 μm, 5 × 10¹⁷c
m^-3), P-InGaAlP active layer 13 (0.5 μm,
5 x 10¹⁶cm^-3), P-InGaAlP cladding layer 1
4 (1 μm, 7 × 10¹⁷cm^-3), P-GaAlAs
Flow diffusion layer 15 (7 μm, 3 × 10¹⁸cm^-3), P-Ga
As contact layer 16 (0.1 μm, 3 × 10 ¹⁸c
m^-3),

The p-side electrode 17 has a circular shape with a diameter of 200 μm. When the composition of the active layer 13 and the clad layers 12 and 14 is 0.4 and 0.7, respectively, with the notation of In _0.5 (Ga _1-X Al _X ) _0.5 P, the emission wavelength is 565 nm.
(Green), the luminous efficiency was about 0.7% at 20 mA DC, which was 10 times higher than the conventional efficiency.

FIG. 2 shows the dependence of the luminous efficiency on the active layer carrier concentration when the thickness of the active layer 13 is 0.5 μm. The p-type was Zn-doped, the n-type of 5 × 10 ¹⁶ cm ⁻³ or less was undoped, and the n-type higher than that was Si-doped. At this time, in the n-type, the luminous efficiency decreased as the concentration thereof increased. It is considered that this is due to the increase in non-radiative recombination and the decrease in the diffusion length of holes, which are minority carriers, due to the generation of deep levels.

On the other hand, in the p-type, when the concentration was increased from 1 × 10 ¹⁷ cm ^-3 , the luminous efficiency was remarkably reduced. It is considered that this is because the center that causes non-radiative recombination was formed by the large amount of Zn doping. In the characteristic curve shown in FIG. 2, even if the film thickness of the active layer 13 was changed, it was only shifted in the vertical direction, and the range in which the luminous efficiency was high remained almost unchanged. Therefore, as the active layer 13, a p-type with a carrier concentration of 1 × 10 ¹⁷ cm ⁻³ or less or a carrier concentration of 5 × 1
An n-type of 0 ¹⁶ cm ^-3 or less is desirable.

In FIG. 3, the carrier concentration of the active layer 13 is 5 ×.
The ^graph shows the dependence of the luminous efficiency on the thickness of the active layer when it is set to 10 ¹⁶ cm ⁻³ (p-type). Although the thickness of the active layer 13 depends on the carrier concentration of the p-clad layer 14, it ranges from 0.15 to 0.15.
A high luminous efficiency was obtained in the range of 0.75 μm.

Although not shown in FIG. 3, the n-type active layer having a carrier concentration of 5 × 10 ¹⁶ cm ⁻³ or more has a lower luminous efficiency than the Zn-doped p-type active layer at the same film thickness. And the peak value of luminous efficiency was present on the thin film side. It is considered that the decrease in efficiency in the thin film is caused by the fact that the injection carrier density of the active layer is determined by the injection current density / active layer thickness, so that overflow into the cladding layer due to the high injection density occurs remarkably. When the active layer 13 is a p-type, the majority carriers are holes, and light is emitted by the injection of electrons. Therefore, it is considered that the influence due to the overflow is more difficult than in the undoped case. On the other hand, in the thick film, when the diffusion length of injected carriers is larger than that in the active layer, the effect due to the double hetero structure is reduced, and light emission due to absorption in the layer and relative non-radiative recombination at low carrier density is increased. It is thought that a decrease in efficiency will occur.

As described above, according to this embodiment, InGa
By optimizing the carrier concentration and the film thickness of the active layer 13 made of AlP, the influence of non-radiative recombination can be reduced and the luminous efficiency can be improved. Therefore, it is possible to realize a semiconductor light emitting diode capable of emitting light with high efficiency even in a short wavelength region such as yellow and green.

In the above embodiment, the Al composition x of the active layer 13 is set to 0.4.
The same effect was obtained with the composition of 1 l. Here, although the emission wavelength changes when the Al composition x is changed regardless of the carrier concentration, as shown in FIG.
Is set to be shorter and the emission wavelength is made shorter, the luminous efficiency is lowered. Also in this case, the carrier concentration is 1 × 10 ¹⁷
If cm ^-3 or less of p-type or carrier concentration of 5 × 10 ¹⁶ cm ^-3 or less of n-type, high luminous efficiency is obtained.

(Second Embodiment) FIG. 5 shows a second embodiment of the present invention.
3 is a cross-sectional view showing a schematic structure of a semiconductor light emitting diode according to the embodiment of FIG.

In the structure shown in FIG. 5, 31 is a first conductivity type semiconductor substrate, and on this substrate 31, a first conductivity type buffer layer 32, a reflection layer 33 and a transparent buffer layer 34.
Is being grown and formed. On the transparent buffer layer 34, a lower clad layer (first conductivity type clad layer) 35, an active layer 3
6 and an upper clad layer (second conductivity type clad layer) 37, a double heterostructure portion is formed.

A second conductivity type current diffusion layer 38 is grown and formed on the double hetero structure portion, and a second conductivity type contact layer 39 is formed on a part of the current diffusion layer 38.
Further, a current blocking layer 40 of the first conductivity type is formed at a position corresponding to the contact layer 39 at the interface between the double hetero structure portion and the current diffusion layer 38. An upper electrode 41 is formed on the contact layer 39, and a lower electrode 42 is formed on the lower surface of the substrate 31.

Next, the substrate, carrier concentration and film thickness of each layer, and other conditions in the above device will be described. Note that this condition can be similarly applied to the first embodiment described above.

The first conductivity type semiconductor substrate 31 is n-GaAs.
And its carrier concentration is 2 × 10 ¹⁸ cm ⁻³ to 4 ×
It is 10 ¹⁸ cm ^-3 . This is because a substance having a low defect density cannot be obtained at a low concentration and a layer made of InGaAlP formed on the substrate 31 cannot be favorably formed at a high concentration. The thickness is 50 to 450 μm. This is because it is difficult to handle a thin film on a wafer and warps due to a slight difference in lattice constant between the substrate and the layer formed on it, making it difficult to form a pellet into a thick film. It depends.

The plane orientation of the n-GaAs substrate 31 is (10
It is inclined in the range of 0) or (100) to 25 degrees.
This is because the composition of the active layer 36 for obtaining the same wavelength is different depending on the direction and angle of the inclination, the defect density is different, and the like, which changes the crystallinity and affects the light emission efficiency. FIG. 6 shows the inclination angle of the substrate surface and 590.
It is a characteristic diagram showing the relationship with the Al composition x for obtaining the luminescence of nm. From this figure, it is understood that the Al composition x can be reduced to 0.3 when the tilt angle is 7 degrees or more in the [011] direction.

FIG. 7 is a characteristic diagram showing the relationship between the angle inclined in the [011] direction and the full width at half maximum of photoluminescence. From this figure, the smallest half-value width is obtained when the tilt angle is 10 to 20 degrees. Therefore, the desirable plane orientation of the substrate 31 is the case of inclining from (100) to the [011] direction by 7 degrees to 25 degrees, and the particularly desirable plane orientation is from (100) to the [011] direction. 10 degrees to 2
This was the case when tilted at 0 degrees.

The buffer layer 32 is for preventing generation of defects due to contamination of the surface of the substrate and transmission to the light emitting layer, and is made of, for example, n-GaAs. Its carrier concentration is 4 × 10 ¹⁷ cm ⁻³ and its thickness is 0.6 μm. In this case, if the material satisfies the above purpose, n-GaAs
However, for example, n-InGaP, n-G
aAlAs, n-InGaAlP, or n-GaAs
Superlattice structure of these materials including (including atomic layer order of 100 atomic layer order and total thickness 10
0 nm) or the like. In addition, the buffer layer 32 may not be provided if it does not adversely affect the light emitting layer.

The reflection layer 33 prevents the light from reaching the substrate 31 and the buffer layer 32, which are light absorbers, for the emission wavelength.
The purpose is to reflect the emitted light in the direction opposite to the substrate 31. For this purpose, it is desirable to stack two kinds of semiconductor layers having different refractive indices η with respect to the emission wavelength λ with a thickness of λ / (4n).

For example, for an emission wavelength of 590 nm, n
-GaAs (η = 3.94; thickness 37.5 nm), n-
Ga _0.3 Al _0.7 As (η = 3.425; thickness 43n
By alternately laminating 20 or more layers of m), about 30% of the light traveling to the substrate side can be reflected before reaching the absorber such as the substrate, and high luminous efficiency can be obtained. The doping amount at this time may be set so that an operating voltage rise due to a large voltage drop does not occur during energization, and here it is set to 4 × 10 ¹⁷ cm ⁻³ . At this time, n-GaAs in the reflective layer 33 is an absorber for the emission wavelength, but the effect of reflection due to the effect of the refractive index is sufficient.

Besides this, n-InGaAlP, n-I
nAlP, n-InGaP, n-InGaAsP, n-
High luminous efficiency can be obtained by forming the reflective layer in the same manner by using a combination of materials such as GaAlAs and n-GaAs and the difference in composition thereof. In addition, when the light absorption on the substrate side does not cause a large demerit to the luminous efficiency, the reflection layer 33 may not be provided.

The transparent buffer layer 34 is made of a transparent material having a sufficiently small absorption coefficient with respect to the emission wavelength, and contributes to the effective extraction of the emission from the side surface. This is because when the transparent buffer layer 34 is sufficiently thick, as shown in FIG. 8, for light emitted toward the substrate side and light emitted toward the opposite side of the substrate and traveling toward the substrate again due to total reflection on the pellet surface, This is because the rate at which the light can be emitted from the side surface of the pellet increases before reaching the light absorber such as the substrate.

For example, if the pellet size is 300 μm
Angle, when the emission wavelength is 590 nm, n-Ga _0.3 Al
_By providing the _0.7 As transparent buffer layer 34 with a thickness of 30 μm, the amount of light that can be extracted from the side surface increases, and
A relatively high luminous efficiency can be obtained. It is considered that this is because the light emission in the vicinity of the side surface of the pellet can be effectively taken out in the range of the distance about the same as the thickness of the transparent buffer layer 34.

The doping amount of the transparent buffer layer 34 may be set so that the operating voltage does not increase due to a large voltage drop during energization, and here, it is 4 × 10 ¹⁷ cm.
^-3 . In addition, the thickness of the transparent buffer layer 34 needs to be significantly larger than the size of the pellet in the region near the side surface of the pellet where light emission can be effectively taken out. Become.

For a 300 μm square chip, for example, as shown in FIG. 9, when the thickness of the transparent buffer layer 34 is set to about 15 μm or more, a sufficient improvement in luminous efficiency was observed. At this time, the thicker the transparent buffer layer 34 is, the higher the light emission efficiency can be obtained. However, when the thickness is 100 μm or more, the light emission efficiency is lowered due to defects and in-plane distribution generated by the warp of the wafer.

By combining the transparent buffer layer 34 and the reflective layer 33, as shown in FIG.
Needless to say, the total reflection on the surface of the pellet is repeated in a zigzag pattern so that the light emitted to the side surface can be effectively extracted. In addition, such an effect is obtained by the transparent buffer layer 3
It is also possible to obtain 4 by forming a layer structure of two or more kinds of materials and compositions having different refractive indexes. In this case, it is important to set the layer having a low refractive index on the substrate side.

Here, as the transparent buffer layer 34, n-
GaAlAs was used, but in addition to this, n-InGaA
It goes without saying that the same effect can be obtained with a material transparent to the emission wavelength such as 1P or n-InAlP. The reflective layer 33 may be in the transparent buffer layer 34.
Note that the transparent buffer layer 34 may not be provided when the light absorption on the substrate side does not cause a large demerit to the luminous efficiency.

The lower cladding layer 35 is for obtaining high luminous efficiency by confining the carriers injected into the active layer 36 in the active layer 36. That is, the active layer 3
When the conductivity type of 6 is n-type and the carrier concentration is lower than the injected carrier density (when the minority carriers are holes), the active layer 3
When the conductivity type of 6 is p-type and the carrier concentration is lower than the injected carrier density (when the minority carriers are electrons), or when the injected carrier density is higher than the carrier concentration of the active layer 36 (in the so-called double injection state), In either case, there is an effect of suppressing the carriers injected into the active layer 36 from diffusing into the lower cladding layer 35. In this case, the lower clad layer 35 needs to have a larger energy gap than the active layer 36.

Here, as the lower cladding layer 35, n
—In _1-Y (Ga _1-X Al _X ) _Y P composition notation (x,
In y), x = 0.7 and y = 0.5. This is because even if the emission wavelength of the active layer 36 is shortened to 555 nm, an energy gap with which a sufficient carrier confinement effect can be obtained can be obtained. Also, Ga
This is because a good crystal can be obtained because the lattice constant matches that of the As substrate 31. The composition range of such a clad layer was 0.6 ≦ x ≦ 1 in the above composition notation.
This is a range in which n-InGaAlP has an almost indirect transition type energy gap. Since the energy gap in this region increases with the composition x, the carrier confinement effect is large in the composition with a large x.

Further, in the InGaAlP-based material, formation of an ordered structure of atomic arrangement is likely to occur, whereby the energy gap changes even with the same composition. The energy gap in a random state with no ordered structure gives the maximum energy gap in the same composition.
The absence of an ordered structure is advantageous for carrier confinement. Such a disordered structure could be easily formed in the substrate 31 by using the above-mentioned inclined orientation.

On the other hand, the carrier concentration of the lower cladding layer 35 was set to 1 × 10 ¹⁷ cm ^-3 . FIG. 11 shows the n-clad layer 35.
2 shows the relationship between the carrier concentration and the luminous efficiency. As shown in this figure, the desirable carrier concentration range was from 1 × 10 ¹⁶ cm ⁻³ to 7 × 10 ¹⁷ cm ⁻³ from the viewpoint of luminous efficiency. Due to the relationship with the resistivities of the upper clad layer 37, the current diffusion layer 38, etc., in order to spread the current to a place other than directly below the electrode where light cannot be extracted by the electrode 41, the resistivity is high and the carrier concentration is low. It is desirable to set. Further, dopants such as Si and Se form a deep level in InGaAlP according to the concentration, and carriers injected into the active layer 36 pass through these levels in the n-InGaAlP lower cladding layer 35 near the active layer. It was thought that the reason was non-radiative recombination.

The lower limit of the desirable carrier concentration is the lowest concentration necessary for preventing an increase in the operating voltage due to a large voltage drop during energization, and the upper limit is determined from the increase in non-radiative recombination. is there. Further, a particularly desirable carrier concentration range is 1 × 10 ¹⁶ cm ⁻³ to 2 × 10 ¹⁷
It was cm ^-3 . It was considered that the upper limit in this case was due to the decrease in the current spreading effect due to the lower resistance.

When the lower clad layer 35 and the buffer layer 32 or the reflective layer 33 are directly joined, n-InGaA
A heterojunction between IP and n-GaAs is formed.
At this time, when injecting electrons from the GaAs side having a small energy gap, an excessive voltage drop was required to overcome the barrier due to the heterojunction. For this reason, the current uniformly flows through the interface throughout the pellet to reduce the current density and prevent voltage drop, resulting in a large current spread. This is n-InG
greatly depends on the carrier concentration of aAlP, and is 1 × 10 ¹⁷ c
It was remarkable at m ⁻³ or less.

The thickness of the lower cladding layer 35 was 1 μm. This is because the carrier injected into the active layer 36 has an effect of suppressing diffusion to the lower clad layer 35, and needs to have a thickness of about 0.5 μm or more. Is about 2 μm or less. In addition, the lower clad layer 35
The same material and composition as the transparent buffer layer 34 may be used. That is, n-GaAlAs or the like may be used.
Further, in the case where the carriers injected into the active layer have a single hetero structure or a homojunction structure and the diffusion of the carriers to the substrate side is not suppressed, the lower cladding layer 35 is formed if the luminous efficiency does not become a large disadvantage. It may not be provided.

The active layer 36 is made of In _1-Y (Ga _1-x Al _x ).
_The energy gap is determined by the composition x and y of the YP and the state of the ordered structure, and when the injected carriers recombine radiatively, light is emitted at a wavelength corresponding to the energy gap. As shown in FIG. 4, Al composition x
The emission wavelength was shortened with the increase of, and the emission efficiency was decreased accordingly. This is because, as the Al composition increases, the difference between the direct transition type energy gap and the indirect transition type energy gap becomes smaller, particularly the deep level generation becomes remarkable in the n-type, and it becomes difficult to obtain a good crystal. It is thought that it depends on things.

There are several possible methods for avoiding these problems. As much as possible to obtain the same wavelength, Al
As a method of realizing a low composition, there is a method of forming the active layer 36 in a disordered structure. Such a chaotic structure
The substrate 31 could be easily formed by using the above-mentioned inclined orientation. Further, by adopting the multiple quantum well structure (MQW), the wavelength can be shortened even if a layer having a small Al composition is used as a light emitting layer due to energy level quantization.

For example, the composition (x, y) = (0.3, 0.
5) is a well layer and (0.7, 0.5) is a barrier layer, and each has a thickness of 2.5 to 5 nm and 10 cycles to 100
By using an active layer provided with a period of about,
, (0.4,0.5) to (0.5,0.
The wavelength (57) equivalent to that of the active layer made of the bulk of composition 5)
5 to 555 nm) was obtained.

Further, the frame of lattice matching with GaAs is removed, that is, the lattice constant of the active layer InGaAlP is set to GaAs.
By making it smaller, the composition of Al can be reduced. Conversely, the lattice constant of the active layer InGaAlP is Ga
By making it larger than As, the difference between the energy gap of the direct transition type and the energy gap of the indirect transition type can be increased. In all cases, the effect is remarkable when the lattice mismatch is about 0.5% or more, and the upper limit is due to an increase in non-radiative recombination due to the occurrence of transition. The occurrence of such a transition depends on the thickness, and is about 1% at about 0.3 μm.
The limit has appeared.

As a method of avoiding this, it was effective to use a layer having a shifted lattice constant as the well layer of the active layer having the MQW structure. Since the thickness of the well layer is very thin, as shown in FIG. 13, the degree of lattice mismatch that limits the occurrence of dislocation is about 3% or more, and a very wide composition range could be used. The conductivity type of the active layer 36 is the n-type of 5 × 10 ¹⁶ cm ⁻³ or less or the p-type of 1 × 10 ¹⁷ cm ⁻³ or less (see FIG. 2) as in the previous embodiment. Thus, high luminous efficiency was obtained. Further, the thickness of the active layer 36 is the same as in the previous embodiment, and the pellet size is 300 μm.
As for the square type, high luminous efficiency was obtained from 0.15 μm to 0.75 μm (see FIG. 3).

The upper clad layer 37 is the lower clad layer 3.
5, carriers injected into the active layer 36 are transferred to the active layer 3
This is to obtain high luminous efficiency by confining it in the interior of No. 6. That is, in either case where the minority carriers are holes, when the minority carriers are electrons, or in the double injection state, it is possible to suppress the diffusion of the carriers injected into the active layer 36 to the upper cladding layer 37. There is. In this case, the upper cladding layer 37 needs to have a larger energy gap than the active layer 36.

Here, p-In _1-Y (Ga _1-X A
In the composition notation (x, y) of l _X ) _Y P, x = 0.
7, y = 0.5. This is because even if the emission wavelength of the active layer 36 is shortened to 555 nm, an energy gap with which a sufficient carrier confinement effect can be obtained can be obtained. Also, since the GaAs substrate lattice constants match, a good crystal can be obtained. like this,
The composition range of the p-clad layer 37 is as shown in FIG.
In the above composition notation, 0.6 ≦ x ≦ 1. This is a range in which p-InGaAlP has an almost indirect transition type energy gap. Since the energy gap in this region increases with the composition x, the carrier confinement effect is large in the composition with a large x.

Further, in the InGaAlP-based material, formation of an ordered structure of atomic arrangement is likely to occur, whereby the energy gap changes even with the same composition. The energy gap in a random state with no ordered structure gives the maximum energy gap in the same composition.
The absence of an ordered structure is advantageous for carrier confinement. Such a disordered structure could be easily formed by using the above-mentioned inclined orientation in the substrate 31 and by performing sufficiently high doping.

On the other hand, the carrier concentration of the upper cladding layer 37 was set to 7 × 10 ¹⁷ cm ^-3 . The p-clad layer 37 is shown in FIG.
2 shows the relationship between the carrier concentration and the luminous efficiency. As shown in this figure, from the viewpoint of luminous efficiency, the desirable carrier concentration range was 5 × 10 ¹⁷ cm ⁻³ to 2 × 10 ¹⁸ cm ⁻³ . This is because of the relationship with the resistivity of the lower clad layer 35, the current diffusion layer 38, etc., the resistivity is low and the carrier concentration is high in order to spread the current to a place other than directly under the electrode where light cannot be extracted by the electrode 41. It is desirable to set it, because by performing high doping of 5 × 10 ¹⁷ cm ⁻³ or more, a disordered structure can be easily formed and the energy gap can be increased, so that carrier confinement can be effectively performed. .

The thickness of the upper clad layer 37 was 1 μm. This is because the carrier injected into the active layer 36 has an effect of suppressing diffusion to the upper clad layer 37, and needs to have a thickness of about 0.5 μm or more. Is about 2 μm or less. In addition, the upper clad layer 37
May have the same material and composition as the current spreading layer 38 and the like. That is, p-GaAlAs or the like may be used.
Also, in the case of a single hetero structure or homojunction structure, if the carriers injected into the active layer are not suppressed from diffusing to the side opposite to the substrate side, but there is no major disadvantage to the luminous efficiency, the upper cladding layer Need not be provided.

The current diffusion layer 38 is for spreading the current to a position other than directly below the electrode where light cannot be extracted by the electrode 41. The current diffusion layer 38 is made of a transparent material having a sufficiently small absorption coefficient for the emission wavelength. For example, when the size of the pellet is 300 μm square, the electrode is circular with a diameter of about 150 μm, and the emission wavelength is 590 nm, p-Ga _0.3 Al _0.7 As with a carrier concentration of 3 × 10 ¹⁸ cm ⁻³ has a thickness of 1
By providing 5 μm, the spread of the current spreads over the entire chip, and a luminous efficiency about 30 times higher than that without this was obtained.

The thickness of the current diffusion layer 38 is desirable so that the spread of current can be increased. The effective current spreading effect was obtained at a thickness of 5 μm or more, as shown in FIG. On the other hand, when the thickness is 30 μm or more, the effect of current spreading is saturated, and the luminous efficiency is reduced due to the diffusion of impurities from the upper cladding layer 37 and the like into the active layer 36 due to long-time growth, and the like. Rather, the luminous efficiency of the device decreased.

The carrier concentration of the current diffusion layer 38 is shown in FIG.
5 × 10 to obtain sufficient current spread, as shown in FIG.
¹⁷cm^-3The above is desirable. On the other hand, 5 × 10¹⁸
cm ^-3With more doping, the current spread is sufficient
However, the absorption for the emission wavelength becomes large, and eventually the
The luminous efficiency as a child decreased.

The Al composition x of the current diffusion layer 38 is the emission wavelength.
The absorption coefficient is 100 cm^-1As low as possible, as low as possible
Is better. At this time, as shown in FIG.
Ga _1-XAl_XAs the composition x of As is larger, at a certain wavelength
Has a smaller absorption coefficient. Light emission wave by InGaAlP
For the long region, x should be 0.6 or more.
Good On the other hand, GaAlAs with large x has good crystallinity.
It was difficult to obtain, and the absorption coefficient increased rather. For this reason,
It was desired that x be 0.8 or less.

At the junction of the current diffusion layer 38 and the upper cladding layer 37, a heterojunction of p-InGaAlP and p-GaAlAs is formed. At this time, holes are injected from the p-GaAlAs side where the energy gap is small. In that case, an excessive voltage drop was required to overcome the barrier due to the heterojunction. For this reason, the current uniformly flows through the interface throughout the pellet to reduce the current density and prevent voltage drop, resulting in a large current spread. This largely depends on the carrier concentration of p-InGaAlP and was remarkable at 1 × 10 ¹⁸ cm ⁻³ or less.

Here, p-Ga is used as the current diffusion layer 38.
AlAs was used, but in addition to this, p-InGaAl
Needless to say, a material that is transparent to the emission wavelength such as P or p-InAlP has the same effect. In addition, when the current diffusion layer 38 is not provided and the light emission efficiency is not largely disadvantageous, the current diffusion layer 38 may not be provided.

The contact layer 39 is a layer for easily making ohmic contact with the electrode 41, and is, for example, p-.
GaAs was used. Specifically, as shown in FIG. 19, after the p-GaAs layer 39 is grown and formed on the current diffusion layer 38, the upper electrode 41 is formed, and the resist 41 is used as a mask to select the electrode 41 and the p-GaAs layer 39. It was formed by etching and finally removing the resist 44.

The concentration of the contact layer 39 is preferably 1 × 10 18 cm ^-3 or more in order to easily lower the contact resistance. In addition, it was important that the thickness was 50 nm or more in order to make a contact with good reproducibility. The contact layer 39 does not necessarily need to be transparent to the emission wavelength. This is because the absorption effect can be avoided by selectively removing the portion other than the electrodes. In this case, there is also an advantage that a damaged portion caused by alloying or the like at the time of electrode formation can be removed at the same time.

In this case, if the contact layer 39 is thick, the contact layer at the end of the electrode 41 disappears in the removal process and the electrode 41 easily peels off. Is 150 nm
It is desirable that it is not more than the level. In addition, the contact layer 3
If 9 is sufficiently thin, absorption can be reduced, so it is not always necessary to remove it at portions other than the electrodes.

Here, p-G is used as the contact layer 39.
Although aAs was used, p-InGaP, p-
InGaAlP, p-GaAlAs, or the like may be used. Further, the electrode may be directly formed on the current diffusion layer without providing the contact layer 39.

The upper electrode 41 is made of AuZn / Au,
A current is injected into the pellet, and it also serves as a pad for wire bonding. The electrode 41 is effective in spreading the current over the entire pellet, it is important not to block the light emission.

The current blocking layer 40 is for spreading the current to a position other than directly below the electrode where light cannot be extracted by the electrode 41. That is, a layer that functions to prevent current injection is inserted into a part or all of the electrode 41 immediately below to avoid invalid current injection directly below the electrode. The current blocking layer 40 may be provided between the electrode 41 and the active layer 36. For example, as shown in FIG. 5, between the upper clad layer 37 and the current diffusion layer 38, and as shown in FIG. 20A, inside the current diffusion layer 38, between the current diffusion layer 38 and the contact 39, Further, as shown in FIG.
It does not matter whether it is between 9 and the electrode 41. At this time,
In order to reduce the current flowing directly under the current blocking layer 40, it is desirable that the current blocking layer 40 be located as close to the active layer 36 as possible.

The current blocking layer 40 is a semiconductor having a conductivity type opposite to that of the upper clad layer 37, for example, n-GaA.
s, n-GaAlAs, n-InGaP, n-InGa
AlP, n-InAlP, etc., or high-resistance semiconductor,
For example, GaAs, GaAlAs, InGaP, InGa
AlP, InAlP, etc. can be used. Also,
A semiconductor layer having the same conductivity type as that of the upper clad layer 37 and having a structure requiring an excessive voltage drop for overcoming a barrier due to a heterojunction between the upper clad layer 37 and a layer in contact therewith, for example, p-GaAs / p-InGaAlP, p -GaAs / p
-InAlP, p-GaAlAs / p-InGaAl
P, like p-GaAlAs / p-InAlP, or insulator can be used as long such as _{SiO 2, Al 2 O 3,} Si 3 N 4.

When the current blocking layer 40 is made of a material transparent to the emission wavelength, the emitted light can be more effectively taken out. That is, by providing the current blocking layer 40 that is transparent to the emission wavelength in the region including immediately below the electrode 41, the reactive current to the immediately below the electrode is reduced, and the light emission due to the current that circulates immediately below the current blocking layer 40 is absorbed. It can be taken out without being affected.

In this embodiment, for example, when the pellet size is 300 μm square, the electrode is circular with a diameter of 150 μm, and the emission wavelength is 590 nm, n-In _{1 -Y} (Ga _{1 -X} A
In the composition notation (x, y) of l _X ) _Y P, x = 0.
7, y = 0.5 was inserted as a current blocking layer 40 between the upper clad layer 37 and the current diffusion layer 38. Current blocking layer 4
The concentration of 0 was 1 × 10 ¹⁸ cm ⁻³ and the thickness was 150 nm. The shape was set to a diameter of 180 μm including the electrode 41. At this time, almost no current was injected directly below the electrode 41, and even if the light emission here was hidden by the electrode 41, it did not affect the light emission efficiency. In addition, the light emission due to the current flowing right under the current blocking layer 40 was taken out to the outside through the transparent current blocking layer 40, so that it was not a reactive current. Note that the current blocking layer 40 may not be provided unless the insertion of the current blocking layer 40 causes a large demerit to the luminous efficiency.

With the structure set as described above, an element pellet having a pellet size of 300 μm square and an electrode having a circular shape with a diameter of 150 μm is molded with resin so as to have a spread half-value width of about 8 degrees, thereby forming an element. By doing In
The composition of _1-Y (Ga _1-X Al _X ) _Y P active layer 35 is x =
By setting 0.2 and y = 0.5, the emission wavelength 61
0 nm orange emission (luminance 10 candela) is x
= 0.3 and y = 0.5, the emission wavelength is 5
Yellow emission of 90 nm (luminance 7 candela) was observed at x = 0.
4, by setting y = 0.5, an emission wavelength of 570n
The yellowish green emission of m (luminance 3 candela) is x = 0.5,
By setting y = 0.5. Green light emission (luminance: 1 candela) having an emission wavelength of 555 nm was obtained.
In other words, by optimizing the carrier concentration and film thickness of each layer, selecting the plane orientation of the substrate, and the like, high luminous efficiency could be obtained even with light emission of a short wavelength.

In the above-described second embodiment, an example in which n-GaAs is used as the first conductivity type semiconductor substrate 31 has been shown. However, p-GaAs having different conductivity types may be used. At this time, it is necessary to reverse the conductivity type of each semiconductor layer. In this case, the thickness of each semiconductor layer is preferably set to be the same as that when the n-GaAs substrate 31 is used.

On the other hand, regarding the carrier concentration, the optimum value is n-GaAs in consideration of current spread and luminous efficiency.
Different from the case of the substrate 31. For the clad layer, n−
In the case of the GaAs substrate 31, the lower clad layer 35 has an n-I
nGaAlP, the upper clad layer 37 is p-InGaAl
However, when a p-GaAs substrate is used, the upper clad layer is n-InGaAlP and the lower clad layer is p-InGaAlP. At this time, n-InGaA
It can be considered that the settings for lP and p-InGaAlP are the same.

The buffer layer 32, the reflective layer 33, the transparent buffer layer 34, the current diffusion layer 38, the contact layer 39, and the current blocking layer 40 may have the same conductivity type and the same carrier concentration. The active layer 36 is preferably set similar to that of n-GaAs regardless of the conductivity type of the substrate.

The advantage of using the p-type conductivity type substrate is that
The current spread can be expanded more easily than in the case of the n-type. This is n-I in comparison with p-InGaAlP and p-GaAlAs in the above composition range.
nGaAlP and n-GaAlAs have a higher mobility for the same composition and carrier concentration, and a layer having a low resistivity is easily obtained. By making this layer on the upper electrode side with respect to the active layer, the current is easily spread, This is because the light emitting portion can be easily enlarged. In particular, when n-GaAlAs is used for the current diffusion layer 38, the resistivity becomes 1/2 of that when p-GaAlAs is used. Therefore, as shown in FIG. 21, its thickness (thickness on the lower limit side). Even if the ratio was set to 1/2, substantially the same current spread and eventually luminous efficiency were obtained. This was effective in shortening the crystal growth time and, in turn, improving productivity.

Further, as the first conductivity type semiconductor substrate 31, n
When -GaAs or p-GaAs is used, the substrate acts as a light absorber for the emission wavelength, as described in detail in the embodiment. Since the light absorbed by the substrate 31 cannot be extracted to the outside, about 1/2 of the light emitted from the active layer 36 does not contribute to the light emission efficiency. As a method of taking out this, FIG.
As shown in FIG. 2, it is effective to remove the substrate 31 and the buffer layer 32 or the like that serves as an absorber for the emission wavelength.

In this case, in order to maintain the mechanical strength of the pellet, the thickness of the transparent buffer layer 34 and the current diffusion layer 38 may be set thick, and the wafer thickness after the substrate removal may be set to about 50 μm or more. is important. At this time, the lower electrode 4
2 is formed on the transparent buffer layer 34. By the lower electrode 42 or the reflector formed below this during assembly,
The light emitted to the lower part can be taken out without being absorbed.

Further, in the above description, n-GaAs, p-GaAs as the first conductivity type semiconductor substrate 31 and an example in which this is removed have been shown. In addition to this, GaP, Zn
A semiconductor such as Se, ZnS, Si or Ge, or a mixed crystal including GaAs and the like may be used as the substrate.
In this case, the concept of the conductivity type described above is directly applied, and the conductivity type may be n-type or p-type. Ga
When P, ZnSe, ZnS, GaAsP, InGaP or the like is used as the substrate, the substrate can be made transparent to the emission wavelength, and there is an advantage that the substrate need not be removed as described above.

However, in these cases, unlike GaAs, the active layer material and the substrate may not be lattice matched. In this case, it is important to suppress the generation of dislocations in the active layer and reduce non-radiative recombination by providing a lattice constant gradient layer having the same conductivity type as the substrate. In this case, the lattice constant of the gradient constant layer is gradually changed from the substrate lattice constant to the active layer lattice constant from the substrate side to the active layer side. At this time, the material of the lattice constant gradient layer is InG.
A transparent material such as aAlP for the emission wavelength is desirable.
When the lattice constant of the active layer is made equal to that of GaAs, by forming GaAlAs as the transparent buffer layer on the lattice constant gradient layer, a layer having the same lattice constant as that of the active layer is formed, thus reducing dislocations. It is also effective. The configuration in this case is shown in FIG. In the figure, 51 is a transparent substrate, and 52 is a lattice constant gradient layer.

Although Si and Ge are opaque with respect to the emission wavelength, a substrate crystal of good quality can be obtained, so that they are desirable substrate crystals from the viewpoint of mass productivity and productivity. In the case of Si, the lattice matching between the active layer material and the substrate cannot be achieved. It is important to suppress generation of dislocations in the active layer and reduce non-radiative recombination by providing a lattice constant gradient layer having the same conductivity type as the substrate. In this case, the lattice constant of the gradient constant layer is gradually changed from the substrate lattice constant to the active layer lattice constant from the substrate side to the active layer side. At this time, the material of the lattice constant gradient layer is GaP, AlP, GaA.
A mixed crystal of s, InP or the like, or a superlattice is used. Ge
In the case of, there is no problem in lattice matching.

In any case, forming GaAlAs as the transparent buffer layer is effective for reducing dislocations. Further, it is effective to remove these substrates after crystal growth, as in the case of GaAs.

Even when these crystals other than n-GaAs are used for the substrate, the effect of spreading the current due to the current diffusion layer and the like, the effect of reducing the reactive current due to the current blocking layer, the carrier concentration and the thickness of the active layer and the cladding layer, etc. The effects on the luminous efficiency such as the degree of disordering are the same as in the above-described embodiment.

(Third Embodiment) FIG. 24 is a sectional view showing an element structure of a semiconductor light emitting diode according to a third embodiment of the present invention. In the figure, 61 is an n-GaAs substrate, and the main surface of this substrate 61 is [011] from the (100) plane.
It is inclined 15 degrees in the direction. On the substrate 61, n-In
_0.5 (Ga _1-X1 Al _X1 ) _0.5 P cladding layer 62, In
_0.5 (Ga _1-X2 Al _X2 ) _0.5 P active layer 63, p-In
_{0.5 (Ga 1-X3 Al X3} ) 0.5 P cladding layer 64 _{(Zn-doped), p-In 0.5 Ga 0.5} P intermediate band gap layer 65, p-GaAs contact layer 66 are sequentially formed. Then, a p-side electrode 67 made of AuZn is formed on the contact layer 66, and AuG is formed on the back surface of the substrate 61.
An n-side electrode 68 made of e is formed.

FIG. 25 is a schematic diagram showing a current distribution and a light emitting region in the light emitting element shown in FIG. In the figure, the current distribution 72 in the light emitting element is indicated by a broken line arrow, and
No. 1 is indicated by each node. The Al composition of each InGaAlP layer is set to X2 ≦ so that high luminous efficiency can be obtained.
X1, X2 ≦ X3 are satisfied. For example, X1 =
X3 = 0.7 and X2 = 0.3.

In the structure shown in FIGS. 24 and 25,
The thickness and carrier concentration of each layer are set as shown in parentheses below. n-GaAs substrate 61 (80 μm, 3
× 10 ¹⁸ cm ^-3 ), n-InGaAlP cladding layer 62
(1 μm, 5 × 10 ¹⁷ cm ⁻³ ), InGaAlP active layer 63 (0.5 μm, undoped), p-InGaAlP
Cladding layer 64 (2 μm, 2 × 10 ¹⁸ cm ⁻³ ), p-I
nGaP intermediate bandgap layer 65 (0.05 μm, 1
× 10 ¹⁸ cm ⁻³ ). The p-side electrode 67 has a diameter of 2
The circular shape was 00 μm.

The structure of this embodiment is different from the conventional structure in that
The main growth surface of the GaAs substrate 61 on which the element structure is laminated is set to (1
The surface is inclined by 15 degrees from the (00) plane in the [011] direction. The superiority of this structure will be described below.

In the conventional structure, that is, the structure in which the (100) plane is used as the principal growth surface of the GaAs substrate 61, p-In is used.
The current spread in the GaAlP cladding layer 64 is small due to the high resistivity of the p-cladding layer 64. p-cladding layer 6
It is conceivable to increase the current spread by increasing the film thickness of No. 4, but this InGaAlP material system has poor thermal conductivity, and increasing the film thickness lowers the crystal structure and adversely affects the upper layer. It is not preferable because it appears. Further, the growth rate of the InGaAlP-based material is limited due to its crystal quality, and when growing a thick film, it is necessary to extend the growth time. This is p
-When impurities having a high diffusibility are used as the impurities of the clad layer 64, the diffusion of the impurities into the active layer 63 occurs and the device characteristics are deteriorated. Therefore, p-InGaA
It is difficult to grow the IP cladding layer 64 thick.

In addition, in this InGaAlP-based material, when Zn is doped, the activation rate is low, so that
It is difficult to obtain a high carrier concentration. In addition to this,
Zn is an impurity having a high diffusibility, and it causes diffusion into the active layer 63 due to high-concentration doping, which also causes deterioration of device characteristics. The clad layers 62 and 64 have a large band gap difference from the active layer 63, and
It is desirable to confine light and carriers to the
For that purpose, the Al composition of the cladding layers 63 and 64 must be increased. However, when Zn is used as the impurity of the p-clad layer 64, there is a problem that the carrier concentration decreases and the resistance increases as the Al composition increases.
Therefore, in the conventional structure, the injection current could not be widened depending on the element structure on the GaAs substrate 61, and the light was emitted only directly under the electrode, and the light extraction efficiency was very small.

On the other hand, as shown in FIGS. 24 and 25, when the growth main surface of the GaAs substrate 61 is a surface inclined from the (100) surface in the [011] direction by 5 to 15 degrees, p Even if the Al composition of the cladding layer 64 is high, Zn can be doped at a high concentration, and the p-InGaAlP cladding layer 64 with low resistance can be formed.

FIG. 26 shows InG with an Al composition of 0.7.
Saturated hole concentration by Zn for aAlP and GaA
The relationship with the inclination angle from the (100) plane of the s substrate to the [011] direction is shown. From this figure, the saturated hole concentration increases as the tilt angle increases, and it is 1 x 1 especially at tilt angles of 5 degrees or more.
It was found that a sufficient saturated hole concentration of 0 ¹⁸ cm ⁻³ or more was obtained, and a sufficient emission intensity of 3 times or more was obtained as compared with the case where a substrate not tilted was used.

Therefore, by adopting the element structure of the present embodiment, the current injected from the electrode 67 has a low resistance p.
The -InGaAlP cladding layer 64 can be spread over a wide area, and light can be emitted over a wide area other than directly below the electrode 67 as shown by the current distribution in FIG. G tilted by 15 degrees from the (100) plane used in this embodiment in the [011] direction
p-InGaAlP of the device formed on the aAs growth main surface
The resistivity of the clad layer 64 and the p-InGaAlP clad layer of the device formed on the GaAs growth main surface of the conventional (100) surface at the same carrier concentration at the above carrier concentration is 0. 2 Ω · cm, (1
It is 2 Ω · cm on the (00) plane.

Further, when the active layer 64 of the device produced in this embodiment and the conventional example was evaluated by photoluminescence (PL) or the like, it was found that the device produced in this embodiment had higher luminous efficiency. 27, the Al composition of the active layer of 0.
The relationship between the PL emission intensity and the tilt angle in the case of 3 is shown. Thus, the PL emission intensity increased as the tilt angle increased, and at 5 degrees or more, a sufficient emission intensity of 3 times or more was obtained. Further, FIG. 28 shows the case where the inclination angles of the Al composition and the PL emission intensity dependency are 0 degree, that is, the (100) plane and 15 degrees. The effect of increasing the emission intensity by the tilted substrate was recognized in a wide Al composition.

From this, it is possible to increase the brightness in this embodiment. In addition, from the (100) plane [0
InG grown on a GaAs substrate tilted in the [11] direction
It is known that the aAlP mixed crystal suppresses the generation of natural superlattice and has a band gap larger than that of the crystal grown on the (100) plane. Due to the effect of increasing the band gap, it is possible to use an active layer having a smaller Al composition to obtain a certain emission wavelength, and it is possible to achieve high brightness in a short wavelength region.

In _0.5 (Ga _{1 -X 2} A) having the above-mentioned laminated structure
l _X2 ) _0.5 P An element was formed by using 0.3 for the Al composition X2 of the active layer 63, and when a voltage was applied in the forward direction to pass a current, the current distribution shown in FIG. Light emission having a peak wavelength at 600 nm was obtained from a wide area on the surface of the device excluding. The same effect can be obtained by providing a carrier concentration distribution in the growth direction in the Zn-doped p-cladding layer on the light emitting portion made of InGaAlP.

As described above, according to this embodiment, GaAs
By using the surface of the substrate 61 inclined by 15 degrees from the (100) surface in the [011] direction as the main growth surface, the light emitting portion In
The resistivity of the p-InGaAlP clad layer 64 on the GaAlP active layer 63 can be reduced, the current injected from the electrode 67 can be spread over a wide range by the p-clad layer 64, and light is emitted to a region other than directly under the electrode. The area can be expanded. Therefore, it is possible to improve the efficiency of extracting light from the periphery of the electrodes, and thus it is possible to achieve high brightness.

(Fourth Embodiment) FIG. 29 is a sectional view showing an element structure of a semiconductor light emitting diode according to a fourth embodiment of the present invention. The same parts as those in FIG. 24 are designated by the same reference numerals, and detailed description thereof will be omitted.

This embodiment differs from the third embodiment described above in that the p-InGaAlP clad layer is
This means that two layers having different l compositions were used. That is, the active layer 6
3 as the first p-clad layer 64 adjacent to the No. 3 p-In having a high Al composition necessary for confining light and carriers.
_0.5 (Ga _0.3 Al _0.7 ) _0.5 P layer is formed and
A second p-In _0.5 (Ga _0.6 Al _0.4 ) _0.5 P clad layer 73 for further diffusing a current is formed on the p-clad layer 64 of FIG. The band gap layer 65 and the contact layer 66 are formed.

The first p-clad layer 64 and the second p-
The film thickness and carrier concentration of the clad layer 73 are respectively 1 μm and 2 × 10 ¹⁷ cm for the first p-clad layer 64.
^-3 , the second p-cladding layer 73 has a thickness of 3 μm, 6 × 10 ^18.
cm ^-3 . The other layer structures are the same as those in the third embodiment.

The present embodiment is different from the conventional structure in that the second p-clad layer 73 having a lower Al composition than the first p-clad layer 64 is formed on the first p-clad layer 64. As described above, in Zn-doped p-InGaAlP, the carrier concentration thereof largely depends on the Al composition.
When the supply amount of is constant, the higher the Al composition, the lower the carrier concentration and the higher the resistivity. For this reason,
With the configuration of this embodiment, the second p-cladding layer 73 (Al composition 0.4) has a smaller resistivity than the first p-cladding layer 64 (Al composition 0.7). can do.

In addition, when the growth main surface of the GaAs substrate 61 is inclined by 15 degrees from the (100) surface in the [011] direction, the carrier concentration can be increased and the resistivity can be decreased. The difference in resistivity between the first p-clad layer 64 and the second p-clad layer 73 can be further increased. In this way, the p-clad layers 64, 73
Since the resistivity distribution can be widened within the region, the current injected from the electrode 67 can be spread over a wide area by the p-clad layer 73 having a low resistivity, and light emission can be achieved over a wide area other than directly below the electrode. .

In the first p-cladding layer 64 and the second p-cladding layer 73 used in this embodiment, the resistivity at the above carrier concentration is 2 Ω · cm and 0.2 Ω · cm, respectively.
Has become. Because of the large difference in resistivity,
The current injected from the electrode is spread widely in the second p-clad layer 73 before reaching the first p-clad layer 64. Therefore, the current distribution is similar to that of the third embodiment,
It is possible to obtain light emission from a wide area of the device surface excluding the p-side electrode. Further, from the element formed by the laminated structure of the present embodiment, light emission having a peak wavelength at 600 nm was obtained as in the third embodiment.

In the fourth embodiment, the composition of the second p-clad layer 73 is In _0.5 (Ga _0.6 Al _0.4 ).
_{Although 0.5} P is used, if the Al composition is lower than that of the first p-cladding layer 64, the resistance can be reduced, and the composition has a bandgap energy larger than that of the active layer 63, this composition is It is not limited. A method of increasing the carrier concentration in the second p-cladding layer 73 with the same composition may be used. Further, the second p-clad layer 73 is not limited to one layer, and if the InGaAlP layer has a smaller Al composition than the first p-clad layer 64 and is larger than the band gap of the active layer 63, the composition of the second p-clad layer 73 is not limited. The same effect can be obtained in a structure in which two or more different layers are laminated. Further, a composition gradient layer in which the Al composition is gradually reduced from the first p-clad layer 64 may be formed.

Further, in the third and fourth embodiments, Ga
A surface obtained by inclining the growth principal surface of the As substrate from the (100) plane by 15 degrees in the [011] direction was used, but the inclination angle is not limited to this, and as described above, the inclination angle is 5 degrees or more. It goes without saying that similar effects can be obtained if they exist. Furthermore, the layer structure of the light emitting portion including the active layer is not limited to the double hetero structure, and may be a single hetero structure or a homojunction. In addition, various modifications can be made without departing from the scope of the present invention.

[0117]

As described above in detail, according to the present invention, in a semiconductor light emitting diode in which a double hetero structure having an active layer made of InGaAlP is formed on a semiconductor substrate, a GaP substrate is used as the substrate and the substrate side Since the light is extracted from the substrate, the light from the active layer can be effectively extracted to the outside without being absorbed by the substrate.
This makes it possible to realize a high-luminance light emitting diode. Further, by making the active layer made of InGaAlP a low-concentration p-type and having a thickness of 0.25 μm to 0.75 μm, a semiconductor light emitting diode capable of emitting light with high efficiency even at a short wavelength can be realized.

[Brief description of drawings]

FIG. 1 is a sectional view showing an element structure of a semiconductor light emitting diode according to a first embodiment of the present invention.

FIG. 2 is a characteristic diagram showing a relationship between carrier concentration and luminous efficiency.

FIG. 3 is a characteristic diagram showing a relationship between active layer thickness and luminous efficiency.

FIG. 4 is a characteristic diagram showing the relationship between wavelength and luminous efficiency.

FIG. 5 is a cross-sectional view showing an element structure for explaining the second embodiment of the present invention.

FIG. 6 is a characteristic diagram showing the relationship between the tilt angle of the substrate and the Al composition and half-value width.

FIG. 7 is a characteristic diagram showing the relationship between the tilt angle of the substrate and the Al composition and half-value width.

FIG. 8 is a sectional view for explaining the function of the transparent buffer layer.

FIG. 9 is a characteristic diagram showing the relationship between the film thickness of the transparent buffer layer and the luminous efficiency.

FIG. 10 is a sectional view for explaining the function of the transparent buffer layer.

FIG. 11 is a characteristic diagram showing the relationship between the carrier concentration of the n-clad layer and the luminous efficiency.

FIG. 12 is a characteristic diagram showing the relationship between the thickness of a well layer and wavelength.

FIG. 13 is a characteristic diagram showing the relationship between the lattice mismatch of the well layer and the luminous efficiency.

FIG. 14 is a characteristic diagram showing the relationship between the Al composition of the cladding layer and the energy gap.

FIG. 15 is a characteristic diagram showing the relationship between the clad layer carrier concentration and luminous efficiency.

FIG. 16 is a characteristic diagram showing the relationship between the film thickness of the current diffusion layer and the luminous efficiency.

FIG. 17 is a characteristic diagram showing the relationship between the carrier concentration of the current diffusion layer and the luminous efficiency.

FIG. 18 is a characteristic diagram showing the relationship between the emission wavelength and the absorption coefficient of the current diffusion layer.

FIG. 19 is a sectional view showing an upper electrode forming step.

FIG. 20 is a cross-sectional view showing a formation position of a current blocking layer.

FIG. 21 is a characteristic diagram for explaining a modified example of the present invention, showing the relationship between the film thickness of the current diffusion layer and the luminous efficiency.

FIG. 22 is a cross-sectional view of an element structure showing a modified example of the present invention in which the substrate is removed.

FIG. 23 is a cross-sectional view of an element structure showing a modified example of the present invention using a transparent substrate.

FIG. 24 is a sectional view showing an element structure of a semiconductor light emitting diode according to a third embodiment of the present invention.

FIG. 25 is a schematic diagram showing a current distribution and a light emitting region in the element according to the third embodiment.

FIG. 26 is a characteristic diagram showing the relationship between the inclination angle and the saturated hole concentration of Zn.

FIG. 27 is a characteristic diagram showing a relationship between a tilt angle and PL emission intensity.

FIG. 28 is a characteristic diagram showing the relationship between Al composition and PL emission intensity.

FIG. 29 is a sectional view showing an element structure of a semiconductor light emitting diode according to a fourth embodiment of the present invention.

FIG. 30 is a sectional view showing an element structure of a conventional semiconductor light emitting diode.

[Explanation of symbols]

11, 31 ... n-GaAs substrate 12, 35 ... n-InGaAlP clad layer 13, 36 ... p-InGaAlP active layer 14, 37 ... p-InGaAlP clad layer 15, 38 ... p-GaAlAs current spreading layer 16, 39 ... p-GaAs contact layer 17, 41 ... P-side electrode 18, 42 ... n-side electrode 32 ... n-GaAs buffer layer 33 ... Reflective layer 34 ... n-GaAlAs transparent buffer layer 40 ... n-GaAs current blocking layer

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Yukie Nishikawa 1 Komukai Toshiba-cho, Saiwai-ku, Kawasaki-shi, Kanagawa Toshiba Research Institute Co., Ltd. (72) Inventor Mariko Suzuki 1 Address, Toshiba Research Institute, Inc. (72) Inventor, Kazuhiko Itaya, Komukai Toshiba Town, Komukai-shi, Kawasaki-shi, Kanagawa Address, Toshiba Research Institute, Ltd. (56) Reference JP-A-1-243482 (JP, A) 50-19382 (JP, A) JP-A-58-222577 (JP, A) JP-A-61-102786 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) H01L 33 / 00

Claims (5)

(57) [Claims] 1. A GaP substrate and a substrate formed on the substrate.
An active layer made of InGaAlP-based material is formed as InGaAlP
A double heterostructure part sandwiched between cladding layers made of a system material, a first electrode formed on the double heterostructure part, and a first electrode selectively formed on the opposite side of the substrate from the double heterostructure part. it comprises a and a second electrode, a semiconductor light-emitting diodes out takes light from the second electrode side
The active layer is a p-type with a carrier concentration of 1 × 10 ¹⁷ cm ⁻³ or less.
And the active layer has a thickness of 0.15 to 0.75 μm.
Set in the range, of the clad layers constituting the double heterostructure part,
The carrier concentration of the p-type cladding layer is 5 × 10 ¹⁷ cm ⁻³ to 2
A semiconductor light emitting diode characterized by being set in a range of × 10 ¹⁸ cm ^-3 . 2. A current diffusion layer is formed between the double hetero structure portion and the first electrode, and light is extracted also from the first electrode side. Semiconductor light emitting diode. 3. The current diffusion layer has a thickness set in a range of 5 to 30 μm and a carrier concentration of 5 × 10 ¹⁷ cm ^{−3 to} 5 × 1.
The semiconductor light emitting diode according to claim 2, wherein the light emitting diode is set in a range of 0 ¹⁸ cm ^-3 . 4. The carrier concentration of an n-type cladding layer among the cladding layers constituting the double hetero structure portion is 1 × 10.
^{16 cm -3 ~7 × 10 17 cm} -3 semiconductor light emitting diode according to any one of claims 1 to 3 is set in the range of characterized by comprising. 5. The active layer is made of InGa having a different Al composition.
The semiconductor light emitting diode according to any one of claims 1 to 4 by well layer and a barrier layer made of AlP material is characterized in that a multiple quantum well structure composed.