JP2002344013A - Semiconductor light-emitting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a semiconductor light-emitting device, suitable for the emission of lights in an infrared wavelength region, where the light reception sensitivity of a photodiode using silicon is high capable of improving the cutoff frequency, without generating the deterioration of the emitted light output. SOLUTION: A light-emitting laminated structure is arranged on a supporting substrate constituted of first semiconductor materials. This light-emitting laminated structure is constituted of a quantum well layer of second semiconductor materials, a pair of carrier closed layers constituted of third semiconductor materials whose band gap is larger than that of the second semiconductor materials with the quantum well layer being interposed, and a pair of clad layers constituted of fourth semiconductor materials, whose band gap is larger than that of third semiconductor materials with the quantum well layers and the pair of carrier closed layers interposed. In this case, the third semiconductor materials and the thickness of the quantum well layer are selected, so that the difference between the energy level of the lower part of the conductive band of the carrier closed layer and the base level of electrons in the quantum well layer can be set, so that is not less than 100 meV.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a semiconductor light emitting device, and more particularly to a semiconductor light emitting device which is suitable for emitting light in a wavelength region where the light receiving sensitivity of a photodiode using silicon is high, and can increase both the light emission output and the cutoff frequency. The present invention relates to a simple semiconductor light emitting device.

[0002]

2. Description of the Related Art As an element for outputting light in an infrared wavelength region (infrared region having a wavelength of 920 nm or less) having high light receiving sensitivity of a photodiode using silicon, Zn-doped GaAs is used.
There is known a double hetero light-emitting diode in which an s layer is sandwiched between AlGaAs layers. By increasing the Zn concentration in the GaAs layer, the cutoff frequency of the light emitting diode can be increased. However, when the Zn concentration is increased, the luminous output decreases.

[0003]

SUMMARY OF THE INVENTION An object of the present invention is to provide a photodiode using silicon which is suitable for emitting light in the infrared wavelength region where the light receiving sensitivity is high, and to improve the cutoff frequency without lowering the emission output. It is to provide a semiconductor light emitting device capable of achieving the above.

[0004]

According to one aspect of the present invention, there is provided a support substrate made of a first semiconductor material, and a light-emitting laminated structure disposed on the support substrate, wherein the light-emitting laminate structure is made of a second semiconductor material. A pair of carrier confinement layers made of a third semiconductor material having a band gap larger than that of the second semiconductor material, sandwiching the quantum well layer, and the quantum well layer and the pair of carrier confinement layers. And a pair of cladding layers made of a fourth semiconductor material having a band gap larger than that of the third semiconductor material, with the energy level at the lower end of the conduction band of the carrier confinement layer interposed therebetween. The light-emitting stacked structure in which the thickness of the second and third semiconductor materials and the quantum well layer is selected so that the difference between the ground level of electrons in the well layer is 100 meV or more; Inject carrier into luminescent stack The semiconductor light emitting device and an electrode for is provided.

The difference between the energy level at the bottom of the conduction band of the carrier confinement layer and the ground level of the electrons in the quantum well layer is 100 m.
When it is eV or more, luminescent recombination in the carrier confinement layer can be prevented, and the cutoff frequency can be increased.

According to another aspect of the present invention, there is provided a support substrate made of a first semiconductor material, and a light emitting laminated structure disposed on the support substrate, wherein the quantum well layer is made of a second semiconductor material; A pair of carrier confinement layers made of a third semiconductor material having a band gap larger than that of the second semiconductor material, and a layer of the quantum well layer and a pair of carrier confinement layers sandwiching the quantum well layer. A pair of cladding layers made of a fourth semiconductor material having a band gap larger than that of the third semiconductor material, and when current is injected into the light emitting laminated structure, electrons and holes are generated in the quantum well layer. The second and third semiconductor materials, and the thicknesses of the quantum well layer and the carrier confinement layer are selected so that light emission recombination does not occur in the carrier confinement layer. Said light emitting laminate structure; The semiconductor light emitting device and an electrode for injecting carriers is provided in the serial light emitting stack structure.

[0007] By preventing luminescent recombination in the carrier confinement layer, the cutoff frequency can be increased.

[0008]

FIG. 1 is a sectional view of a semiconductor light emitting device according to an embodiment of the present invention. High concentration layer 2 and low concentration layer 3
The support substrate 4 is composed of two layers. Support substrate 4
Is formed of Zn-doped p-type Al _0.26 Ga _0.74 As. The Zn concentration of the high concentration layer 2 is 1 × 10 ¹⁸ c
m ⁻³ and the Zn concentration of the low concentration layer 3 is 5 × 10 ¹⁷ cm ^−3.
It is. Further, the thickness of the high concentration layer 2 is 40 μm, and the thickness of the low concentration layer 3 is 110 μm.

Each layer from the AlGaAs buffer layer 5 to the GaAs contact layer 12 is formed on the surface of the low concentration layer 3 by metal organic chemical vapor deposition (MOCVD). The buffer layer 5 is made of p-type Al doped with Zn.
It is formed of _0.26 Ga _0.74 As, has a thickness of 0.2 μm,
Its Zn concentration is 1 × 10 ¹⁸ cm ⁻³ .

The lower cladding layer 6 is made of Zn-doped p-type Al _0.32 Ga _0.68 As, has a thickness of 0.5 μm, and a Zn concentration of 1 × 10 ¹⁸ cm ^-3 .
The lower carrier confinement layer (CCL layer) 7 is formed of Al _0.18 Ga _0.82 As not intentionally doped with an impurity,
Its thickness is between 2 and 190 nm. The background concentration of Zn in the lower CCL layer 7 is 5 × 10 ¹⁶ to 1 × 1.
It was 0 ¹⁷ cm ^-3 .

The strain quantum well layer 8 is formed of InGaAs, has a thickness of 2.4 to 15 nm, and has an In composition ratio of 0.1.
2 to 0.25. The upper carrier confinement layer 9 is formed of Al _0.18 Ga _0.82 As not intentionally doped with an impurity, and has a thickness of 2 to 190 nm. The background concentration of Si in the upper carrier confinement layer 9 is 5 ×
It was 10 < ¹⁶ > ^-1 * 10 < ¹⁷ > cm < ^-3 >. Upper cladding layer 1
0 is n-type Al _0.32 Ga _0.68 As doped with Si
The thickness is 5.5 μm and the Si concentration is 1
× 10 ¹⁸ cm ^-3 .

The current diffusion layer 11 is made of n-doped Si.
Type Al _0.18 Ga _0.82 As, the thickness of which is 4.
5 μm, and its Si concentration is 1 × 10 ¹⁸ cm ⁻³ . The contact layer 12 is formed of n-type GaAs doped with Si, the thickness is 0.1 μm, and the Si concentration is 2 ×
It is 10 ¹⁸ cm ^-3 .

On the contact layer 12, N
An n-side electrode 15 in which an i-layer, a Ge layer, and an Au layer are stacked is formed. The n-side electrode 15 has, for example, an X-shaped planar shape by a lift-off method. On the surface of the high concentration layer 2 constituting a part of the support layer 4, A
A p-side electrode 16 in which a u layer and an AuZn alloy layer are stacked is formed. The p-side electrode 16 has, for example, a honeycomb shape by a lift-off method.

Next, with reference to FIG. 2, a method of manufacturing the base substrate 4 used in the semiconductor light emitting device according to the above embodiment will be described.

A temporary substrate 1 made of GaAs shown in FIG. 2A is prepared. The main surface of the temporary substrate 1 is a (100) plane of GaAs. The temporary substrate 1 is doped with Zn to have p-type conductivity, and its concentration is 2 to 2.
It is 5 × 10 ¹⁹ cm ⁻³ .

On the main surface of the temporary substrate 1, a high-concentration layer 2 having a thickness of 40 μm and a low-concentration layer 3 having a thickness of 150 μm made of Al _0.26 Ga _0.74 As are sequentially grown by liquid phase epitaxy (LPE). . These two layers become the support substrate 4. LPE mainly includes a temperature difference method and a slow cooling method.
Here, the temperature difference method is employed. By employing the temperature difference method, the Al composition ratio of the support substrate 4 can be made substantially uniform. As the growth apparatus, for example, a slide boat type apparatus can be used. The high-concentration layer 2 and the low-concentration layer 3 have Zn concentrations of 1 × 10 ¹⁸ cm ⁻³ and 5 ×, respectively.
Zn is doped during growth to 10 ¹⁷ cm ^-3 .

The growth solution used was GaAs in a Ga solvent.
s, Al and Zn are dissolved. The vertical temperature gradient of the growth solution filled in the melt bath is about 5
° C / cm, and the temperature of the lower part of the growth solution in contact with the seed crystal is 830 to 850 ° C. The temperature and the temperature gradient at the lower part of the growth solution are kept almost constant during the growth.

The steps up to the state shown in FIG. 1B will be described. The temporary GaAs substrate 1 shown in FIG. 1 is removed by etching. As a result, only the support substrate 4 remains.
The temporary substrate 1 made of GaAs can be etched using an etching solution in which ammonia water and hydrogen peroxide solution are mixed at a volume ratio of 20: 1. The concentration of ammonia water was 28% by weight, and the concentration of hydrogen peroxide solution was 3%.
1% by weight.

Next, the surface of the low concentration layer 3 is ground to reduce irregularities. Further, after the ground surface is polished to remove processing damage, final finishing by chemical mechanical polishing (CMP) is performed. Generally, a semiconductor layer grown by the temperature difference method has poorer surface flatness than a semiconductor grown by the slow cooling method. By performing the final finish by CMP,
The flatness of the surface can be improved. By the steps so far, the support substrate 4 is obtained.

By applying a forward bias between the n-side electrode 15 and the p-side electrode 16 and injecting carriers into the strained quantum well layer 8, the infrared region (wavelength 800 to 920 nm) is used.
m) can be emitted.

In the above embodiment, the support substrate 4 made of AlGaAs becomes a substrate having a physical supporting force and a seed crystal for MOCVD. GaAs as substrate material
Since no s is used, the contact layer 12 shown in FIG.
Light can be extracted not only from the side but also from the support substrate 4 side. The contact layer 12 made of GaAs
Is not removed by acid treatment in the subsequent chip forming process, and does not hinder light extraction. When the wavelength at which the peak of the emission spectrum of the strained quantum well layer 8 gives a shorter wavelength than the wavelength corresponding to the band gap of the semiconductor material forming the temporary substrate 1, the effect of removing the temporary substrate 1 is particularly high.

The lower carrier confinement layer 7, strained quantum well layer 8, and upper carrier confinement layer 9 are formed by MOCVD. For this reason, compared with the case where these layers are formed by LPE, the uniformity of the film thickness can be improved, and high luminous efficiency can be realized. Note that molecular beam epitaxial growth (MBE) may be used instead of MOCVD.

Next, preferred conditions for improving the light emission intensity and operation speed of the semiconductor light emitting device according to the above embodiment will be described.

FIG. 3 shows an energy band structure on the conduction band side from the lower cladding layer 6 to the upper cladding layer 10.
Lz the thickness of the strained quantum well layer 8, Ee ₁ the ground level of electrons in strained quantum well layer 8, DerutaEe ₁ the energy difference between the conduction band and the ground level Ee ₁ of the strained quantum well layer 8, Lower and upper conduction band of lower and upper carrier confinement layers 7 and 9 and ground level E of electrons
Let ΔEb be the energy difference between e ₁ and ΔEc be the energy difference between the conduction band lower ends of the lower and upper carrier confinement layers 7 and 9 and the strained quantum well layer 8.

Electrons injected from the upper cladding layer 10 into the upper carrier confinement layer 9 form a potential barrier B _{1 at} the interface between the lower cladding layer 6 and the lower carrier confinement layer 7 and the upper cladding layer 10 and the upper carrier confinement layer. Is confined in three layers from the lower carrier confinement layer 7 to the upper carrier confinement layer 9 by the potential barrier B _{2 at} the interface with the confinement layer 9. The confined electrons are trapped by the ground level Ee ₁ of the electrons in the strained quantum well layer 8 and recombine with holes in the valence band, thereby emitting light.

When the band gap of the carrier confinement layers 7 and 9 is narrowed and the energy difference ΔEb is reduced, light emission occurs due to recombination of electrons and holes in the carrier confinement layers 7 and 9. This light emission has a slower response speed than the light emission in the strained quantum well layer 8. In order to increase the response speed of the light emitting device, it is necessary to prevent light emission from the carrier confinement layers 7 and 9.

When the band gap of the carrier confinement layers 7 and 9 is widened to increase the energy difference ΔEb,
The potential barriers B ₁ and B ₂ are reduced. Therefore, the effect of confining electrons is reduced. Therefore, it is considered that the energy difference ΔEb greatly affects the luminous efficiency and the response speed.

Hereinafter, the energy difference ΔEb (= ΔEc−ΔE)
The method for obtaining e ₁ ) will be described. Strain DerutaEe _n the energy difference between the n-th level Ee _n the bottom of the conduction band and the strained quantum well layer 8 of the quantum well layer 8, the electron effective mass of the strained quantum well layer 8 me ₁ ^*, carrier Assuming that the effective mass of the electrons in the confinement layers 7 and 9 is me ₂ ^* and the Planck constant is h, when the order n is an odd number, the following equation is established. Note that it is assumed that the carrier confinement layers 7 and 9 are sufficiently thicker than the depth of penetration of the electron wave function into the carrier confinement layers 7 and 9 (depending on the well structure, but about several nm). ing.

[0029]

[Number 1] (αLz / 2) tan (αLz / 2) = (βLz / 2) (me 1 * / me 2 *) α 2 = 2 (me 1 *) ΔEe n / (h / 2π) 2 β 2 = ₂ (me 2 ^*) to _{(ΔEc-ΔEe n) / (} h / 2π) 2 ··· (1) in general, in _1-x Ga x band gap Eg ₁ of As,

[0030]

Eg ₁ = 1.422−1.53 (1-x) +0.45 (1-x) ^2. (2), and the effective mass me ₁ ^* of electrons in the conduction band is free. Let the rest mass of the electron be me ₀ ,

[0031]

## EQU3 ## This is given by: me ₁ ^* / me ₀ = 0.0225 (1-x) + 0.0665x (3) Equation (2) is derived without considering the strain amount of the strained quantum well layer 8.

When strain is introduced into the InGaAs well layer,
The energy gap changes. The change ΔEg ₁
Is

[0033]

ΔEg ₁ = [− 2a (C ₁₁ −C ₁₂ ) / C ₁₁ + b
(C ₁₁ + 2C ₁₂ ) / C ₁₁ ] ε. Where a and b are deformation potentials, C
₁₁ and C ₁₂ are elastic stiffness constant, epsilon is an elastic strain caused by the lattice mismatch between the well layer and the carrier confinement layer. The elastic strain epsilon, Aw lattice constant of _{In 1-x Ga x As,} Al z Ga
_If the lattice constant of _1-z As is Ab, it is represented by the following equation.

[0034]

Ε = (Aw−Ab) / Ab where, lattice constants Aw, Ab, deformation potentials a, b
The elastic stiffness constants C ₁₁ and C ₁₂ are given by the following equations.

[0035]

Aw = 0.56533x + 0.60584 (1-x) [nm] Ab = 0.565533 (1-z) + 0.015z [nm] a = -5.8 (1-x) -9.8x [eV] b = -1.8 (1-x) -1.76x [eV] _{C 11 = 0.833 (1-x} ) + 1.188x [× 10 ¹² dyn / cm ^2] C ₁₂ = 0.432 (1-x) + 0.532x [× 10 ¹² dyn / cm ² ] The band gap Eg ₂ of Al _z Ga ₁ -z As is

[0036]

Eg ₂ = 1.425 + 1.444z (7) The effective mass me ₂ ^* of electrons in the conduction band is given by

[0037]

[Expression 8] me ₂ ^* / me ₀ = 0.0665 (1−z) + 0.15z (8)

When the carrier confinement layers 7 and 9 are formed of AlGaAs and the strain quantum well layer 8 is formed of InGaAs, the band gap difference ΔEg between them is Eg ₂ −
(Eg ₁ + ΔEg ₁ ), and the energy difference ΔE
c is

[0039]

ΔEc = 0.57ΔEg The carrier confinement layers 7 and 9 are formed of GaAs, and the strain quantum well layer 8 is formed of InGaAs.
s, the energy difference ΔEc is

[0040]

## EQU10 ## ΔEc = 0.62ΔEg. The carrier confinement layers 7 and 9 are made of AlGaAs, and the strain quantum well layer 8 is made of GaAs.
s, the energy difference ΔEc is

[0041]

ΔEc = 0.62ΔEg

The semiconductor light emitting device shown in FIG. 1 was manufactured with various combinations of the In composition ratio (1-x) and the Al composition ratio z. The energy difference ΔEb of each semiconductor light emitting device was obtained by calculation, and the cutoff frequency was actually measured. Energy difference Δ
Eb can be calculated from equation (1) and ΔEb = ΔEc-ΔEe ₁ relationship. The cutoff frequency was measured by the following method.

An amplitude of 10 mA is applied to a forward direct current of 50 mA.
The current obtained by superimposing the alternating current is supplied to the semiconductor light emitting device. The frequency at which the amplitude of the light output is −3 dB with respect to the amplitude of the light output when the frequency of the alternating current is f ₀ is called a cutoff frequency fc. Here, f ₀ = (1/100) fc.

FIG. 4 shows the relationship between the energy difference ΔEb and the cutoff frequency. The horizontal axis represents the cutoff frequency in the unit “MHz”, and the vertical axis represents the energy difference ΔEb in the unit “meV”.
In the figure, diamond symbols, circle symbols, and triangle symbols indicate semiconductor light emitting devices in which the thickness of the strained quantum well layer 8 is 3 nm, 5 nm, and 15 nm, respectively.

As the energy difference ΔEb becomes smaller, the cutoff frequency decreases. In particular, when the thickness of the quantum well layer 8 is 5 nm, it can be seen that the slope of the graph changes when the energy difference ΔEb is 100 meV. Note that, regardless of whether the thickness of the quantum well layer 8 is 3 nm or 15 nm, the inclination of the graph tends to change when the energy difference ΔEb is 100 meV.

In each case, the energy difference ΔEb is 10
When the voltage is 0 meV or less, the cutoff frequency tends to decrease. It is considered that this is because recombination of electrons and holes began to occur in the carrier confinement layers 7 and 9 due to the reduced energy difference ΔEb.

FIG. 5 shows a carrier confinement layer (Al _z Ga ₁ -z A).
s) Emission spectra of various samples having different Al composition ratios z of 7 and 9 are shown. The horizontal axis indicates the emission wavelength in the unit of “nm”.
, And the vertical axis represents the normalized emission intensity. The normalized light emission intensity is a value obtained by standardizing the maximum light emission intensity as 1. Curves a to d in the figure show that the Al composition ratio z is 0.09,
The emission spectra at 0.13, 0.18 and 0.26 are shown. In each of the samples, the I
The n composition ratio is 0.12 and the thickness is 5 nm. Further, the Al composition ratio of the current diffusion layer 11 is set to 0.32 so that the light emission is not absorbed by the current diffusion layer 11.

As the Al composition ratio z of the carrier confinement layers 7 and 9 decreases, the energy difference ΔEb decreases.
Further, the emission wavelength shifts to the longer wavelength side. Note that the curve a
Ｄd are 80 me
V, 107 meV, 143 meV, and 202 meV.

A sub-peak is observed around the wavelength 800 nm of the curve a and around 770 nm of the curve b. Observation of a sub-peak on the shorter wavelength side than the main peak indicates that light is emitted from the carrier confinement layers 7 and 9. When the energy difference ΔEb increases, no sub-peak is observed. Therefore, in order to obtain a high cutoff frequency, the energy difference ΔEb is preferably set to 100 meV or more, and further, the energy difference ΔEb is set to 110 meV.
Then, a light-emitting element having a high cutoff frequency in which no subpeak is observed can be obtained. Energy difference ΔEb
Can be substantially specified if the thickness of the strained quantum well layer 8 and the semiconductor materials of the strained quantum well layer 8 and the carrier confinement layers 7 and 9 are determined as shown in the above equation (1).

FIG. 6 shows the relationship between the thickness of the strained quantum well layer 8 and the cutoff frequency in comparison with the energy difference ΔEb. The horizontal axis represents the thickness of the strained quantum well layer 8 in units of “nm”, the left vertical axis represents the cutoff frequency in units of “MHz”, and the right vertical axis represents the energy difference ΔEb in units of “meV”. The sample to be evaluated has a strain quantum well layer 8 having a thickness of 5 nm, 10 nm,
And 15 nm. In each sample, the In composition ratio of the strained quantum well layer 8 is 0.12, and the Al composition ratio of the carrier confinement layers 7 and 9 is 0.18.

The square symbols in FIG. 6 indicate the energy difference ΔEb. Triangle symbols and diamond symbols indicate cutoff frequencies when the bias current is 100 mA and 50 mA, respectively. Each sample has a square shape with a side length of 300 μm.

When the strained quantum well layer 8 is made thicker, the cutoff frequency is reduced despite the increase in the energy difference ΔEb. For example, in order to realize a communication speed of 100 Mbps in the NRZ communication method, the cutoff frequency needs to be about 70 MHz.
must be greater than or equal to z. If the thickness of the strained quantum well layer 8 is 15 nm or less, the cutoff frequency can be made 70 MHz or more by setting the bias current to 100 mA.

FIG. 7 shows the relationship between the thickness of the strained quantum well layer 8 and the emission wavelength for five types of In composition ratios (1-x) of the strained quantum well layer 8. The horizontal axis represents the thickness of the strained quantum well layer 8 in units of “nm”, and the vertical axis represents the emission wavelength in units of “nm”. The numerical value attached to the solid line in the figure is the In composition ratio.
Note that the Al composition ratio z of the carrier confinement layers 7 and 9 is 0.1.
Eighteen.

It can be seen that the emission wavelength becomes longer when the In composition ratio of the strained quantum well layer 8 is increased, and the emission wavelength becomes longer even when the strained quantum well layer 8 is made thicker. In order to receive light with a photodiode using silicon, the emission wavelength must be 800 n.
m to 920 nm. Strained quantum well layer 8
In order to make the In composition ratio of O.sub.2 larger than 0.25 and the emission wavelength to be 800 nm to 920 nm, the strained quantum well layer 8 must be thinner than about 3 nm. It is difficult to form such a thin strained quantum well layer with good reproducibility. When the In composition ratio is large, the emission wavelength is 8
Since the slope of the graph is steep in the range of 00 to 920 nm, the light emission wavelength greatly changes due to a slight variation in the thickness of the strained quantum well layer 8. Therefore, the In composition ratio of the strained quantum well layer 8 is preferably set to 0.25 or less.

If the In composition ratio is reduced, light with a wavelength of 800 nm to 920 nm can be obtained even if the strained quantum well layer 8 is relatively thick. However, when the strained quantum well layer 8 is made thicker and its In composition ratio is made smaller, FIG.
As described above, the cutoff frequency is reduced.

FIG. 8 shows the relationship between the cutoff frequency and the energy difference ΔEb. The horizontal axis represents the cutoff frequency in the unit “MHz”, and the vertical axis represents the energy difference ΔEb in the unit “meV”.
The diamond symbol in the figure indicates the case where a strained quantum well layer with an In composition ratio of 0.12 is used, and the square symbol indicates the case where a GaAs well layer is used. As the energy difference ΔEb increases, I
It can be seen that the cutoff frequency in the case of using the nGaAs strained quantum well layer is increased. Moreover, a higher cutoff frequency can be obtained by using InGaAs than by using GaAs for the strained quantum well layer. Therefore, in order to keep the cutoff frequency high, the strained quantum well layer 8 should be made of In ₁ -xG
a _x As (0 <x <1) is preferably used, and the In composition ratio is more preferably set to 0.05 or more.

FIG. 9 shows the relationship between the cutoff frequency and the bias current for a plurality of samples having different Al composition ratios of the carrier confinement layers 7 and 9. The horizontal axis represents the bias current in the unit of “mA”, and the vertical axis represents the cutoff frequency in the unit of “MHz”. Curves a to d in the figure each have an Al composition ratio of 0.1.
The cut-off frequencies for 09, 0.13, 0.18, and 0.26 are shown. Note that the In composition ratio of the strained quantum well layer 8 is 0.12, and its thickness is 5 nm. When the Al composition ratio is 0.09 and 0.13, the bias current is 1
It shows a maximum value near 00 mA, and even if the bias current is further increased, the cutoff frequency does not increase but saturates. A
It can be seen that if the 1 composition ratio is larger than 0.13, the cutoff frequency is unlikely to be saturated when the bias current is increased. If the bias current is set to 40 mA or more, Al
It is possible to obtain a cutoff frequency of 70 MHz or more under the condition that the composition ratio is 0.13 or more.

FIG. 10 shows that the carrier confinement layers 7 and 9 have Al
The relationship between the bias current and the light emission output of a plurality of samples having different composition ratios will be shown in comparison with the normalized differential quantum efficiency. The horizontal axis represents the bias current in the unit of “mA”, the left vertical axis represents the light emission output in the unit of “mW”, and the right vertical axis represents the normalized differential quantum efficiency. In this case, the differential quantum efficiency means the slope of a straight line connecting adjacent measurement points in a graph showing the relationship between the bias current (I) and the light emission output (p), and was obtained from the following equation. .

[0059]

(Dp / dI) _n = (p _n −p _n−1 ) / (I _n −I _n−1 ) where n is an integer from 1 to the number of measurement points.

The normalized differential quantum efficiency is obtained by standardizing the maximum value of the differential quantum efficiency as 1. Solid lines a to e in the figure show light emission outputs when the Al composition ratios of the carrier confinement layers 7 and 9 are 0.09, 0.13, 0.18, 0.26, and 0.32, respectively. Dotted lines a ′ to e ′ indicate that the Al composition ratio of the carrier confinement layers 7 and 9 is 0.0
9 shows normalized differential quantum efficiencies for 9, 0.13, 0.18, 0.26, and 0.32.

As the bias current increases, the light emission output also increases, and the normalized differential quantum efficiency corresponding to the slope decreases. When the Al composition ratio is reduced,
The optical output tends to increase, and the normalized differential quantum efficiency also increases. When the Al composition ratio is 0.32, which is the same as the Al composition ratio of the cladding layer, a sharp decrease in the normalized differential quantum efficiency is observed, and the increase rate of the light emission output is small even when the bias current is larger than 150 mA. It is considered that this is because the potential barriers B ₁ and B ₂ shown in FIG. 3 were lowered, and the carrier confinement effect was reduced.

From FIG. 10, it can be considered that it is preferable that the Al composition ratio of the carrier confinement layers 7 and 9 is less than the Al composition ratio of the cladding in order to obtain a sufficient output. In particular, in the case of an infrared light emitting element used for optical communication, a high cutoff frequency and a large light emission output are required in order to increase the communication speed and extend the communicable distance. In order to obtain a larger light emission output, a light emitting element for communication is often used with a large current of 200 mA or more which is not used in a visible light LED used for general lighting. In order to stably exhibit a high normalized differential quantum efficiency even when such a large current flows, the Al composition ratio is preferably less than 0.26. At this time, the band gap difference ΔEg between the carrier confinement layers 7 and 9 (Al composition ratio 0.26) and the cladding layer (Al composition ratio 0.32) is 84 meV. ΔEc =
Using the relational expression of 0.65ΔEg, the height of the potential barriers B ₁ and B ₂ is 55 meV. That is, it is preferable that the height of the potential barrier at the interface between the cladding layer and the carrier confinement layer be 55 meV or more.

FIG. 11 shows the relationship between the bias current and the cutoff frequency for a plurality of samples in which the thicknesses of the carrier confinement layers 7 and 8 are different. The horizontal axis represents the bias current in the unit “m”.
A, and the vertical axis represents the cutoff frequency in the unit “MHz”. Curves a to c in the figure show the cases where the thicknesses of the carrier confinement layers 7 and 9 are 30 nm, 50 nm, and 120 nm, respectively. As the bias current increases, the cutoff frequency increases, but the thickness of the carrier confinement layers 7 and 9 is reduced by 12
In the case of 0 nm, it is found that the bias current is saturated when the bias current is 100 mA or more. Therefore, it is preferable that the thickness of the carrier confinement layers 7 and 9 be less than 120 nm. In order to exhibit a sufficient carrier confinement effect, the thickness of the carrier confinement layers 7 and 9 is preferably set to 10 nm or more.

In the above embodiment, as shown in FIG. 1, the number of strain quantum well layers 8 is one, but may be two or more.

FIG. 12 shows that the number of strained quantum well layers is 1, 2,
The relationship between the bias current and the normalized cutoff frequency is shown for the samples Nos. 3, 5, and 10. The horizontal axis represents the bias current in the unit of “mA”, and the vertical axis represents the normalized cutoff frequency. In each sample, the thickness of the strained quantum well layer was 5 nm, the In composition ratio was 0.12, the thickness of the carrier confinement layer was 50 nm,
The Al composition ratio is 0.18. When a plurality of strained quantum well layers are provided, the thickness of the barrier layer disposed between the well layers is 1
0 nm, and the Al composition ratio of Al in the carrier confinement layer is
The composition ratio is the same. Note that the chip size is 400 μm
× 400 μm.

It can be seen that the smaller the number of strained quantum well layers, the higher the normalized cutoff frequency. When the bias current is increased, the light output saturates at a certain intensity. Increasing the number of strained quantum well layers can increase the light output saturation value. Therefore, the number of strained quantum well layers should be selected in view of the required cutoff frequency and light output.
In order to achieve a cutoff frequency of about 70 MHz in order to achieve a transmission speed of 100 Mbps, it is preferable that the number of strained quantum well layers is one or two.

FIG. 13 shows the distribution of the cutoff frequency and the light emission output between the conventional light emitting diode and the light emitting diode according to the above embodiment. The horizontal axis represents the cutoff frequency in the unit “MHz”, and the vertical axis represents the light emission output in the unit “mW”. The circle symbol in the figure corresponds to the light emitting diode according to the embodiment, and the triangle symbol uses Zn-doped GaAs as the quantum well layer.
This corresponds to a conventional light emitting diode using AlGaAs as the carrier confinement layer.

In the conventional light emitting diode, the cutoff frequency can be increased as the Zn concentration in the quantum well layer is increased, but the light emission output is reduced as the cutoff frequency is increased. Further, it was difficult to increase the cutoff frequency to 60 MHz or more. On the other hand, in the light emitting diode according to the embodiment, the cutoff frequency can be set to 60 MHz or more, and the light emission output does not decrease even if the cutoff frequency is increased.

The present invention has been described in connection with the preferred embodiments.
The present invention is not limited to these. For example, it will be apparent to those skilled in the art that various modifications, improvements, combinations, and the like can be made.

[0070]

As described above, according to the present invention,
By using InGaAs for the strained quantum well layer and using AlGaAs or GaAs for the carrier confinement layer, the emission wavelength can be set to 800 nm to 920 nm, and both prevention of reduction in emission output and improvement of cutoff frequency can be achieved.

[Brief description of the drawings]

FIG. 1 is a schematic sectional view of a semiconductor light emitting device according to an embodiment of the present invention.

FIG. 2 is a cross-sectional view illustrating a method of manufacturing a support substrate of a semiconductor light emitting device according to an embodiment of the present invention.

FIG. 3 is a view showing a band structure on a conduction band side of a laminated structure from a lower cladding layer to an upper cladding layer of the semiconductor light emitting device according to the embodiment.

FIG. 4 shows an energy difference ΔE of the semiconductor light emitting device according to the embodiment.
6 is a graph showing a relationship between b and a cutoff frequency.

FIG. 5 is a graph showing emission spectra of various samples in which the Al composition ratio of the carrier confinement layer of the semiconductor light emitting device according to the example is different.

FIG. 6 is a graph showing a relationship between a thickness of a strained quantum well layer and a cutoff frequency of a semiconductor light emitting device according to an example.

FIG. 7 is a graph showing the relationship between the thickness of a strained quantum well layer and the emission wavelength of a semiconductor light emitting device according to an example.

FIG. 8 is a graph showing a relationship between a cutoff frequency and an energy difference ΔEb of the semiconductor light emitting device according to the example.

FIG. 9 is a graph showing a relationship between a bias current and a cutoff frequency for various Al composition ratios of a carrier confinement layer of a semiconductor light emitting device according to an example.

FIG. 10 is a graph showing the relationship between the bias current and the normalized optical output for various Al composition ratios of the carrier confinement layer of the semiconductor light emitting device according to the example, in comparison with the normalized differential quantum efficiency.

FIG. 11 is a graph showing the relationship between the bias current and the cutoff frequency for various thicknesses of the carrier confinement layer of the semiconductor light emitting device according to the example.

FIG. 12 is a graph showing a relationship between a bias current and a normalized cutoff frequency when the number of quantum wells of the semiconductor light emitting device according to the example is variously changed.

FIG. 13 is a graph showing distributions of cut-off frequency and light emission output between the optical semiconductor device according to the example and the conventional optical semiconductor device.

[Explanation of symbols]

 Reference Signs List 1 temporary p-type GaAs substrate 2 high-concentration p-type AlGaAs layer 3 low-concentration p-type AlGaAs layer 4 support substrate 5 p-type AlGaAs buffer layer 6 p-type AlGaAs lower cladding layer 7 p-type AlGaAs lower carrier confinement layer 8 InGaAs strain quantum well Layer 9 n-type AlGaAs upper carrier confinement layer 10 n-type AlGaAs upper cladding layer 11 n-type AlGaAs current diffusion layer 12 n-type GaAs contact layer 15 n-side electrode 16 p-side electrode

Continuing on the front page (72) Inventor Satoshi Sakura 2-9-13 Nakameguro, Meguro-ku, Tokyo Stanley Electric Co., Ltd. (72) Inventor Shotaro Tomita 2-9-13 Nakameguro, Meguro-ku, Tokyo Stanley Electric Inside (72) Inventor Keizo Kawaguchi 2-9-13 Nakameguro, Meguro-ku, Tokyo Stanley Electric Co., Ltd.F-term (reference) 5F041 AA02 AA04 CA05 CA34 CA49 CA58 CA65 CA74 CA85 CA92 FF16 5F045 AA04 AB10 AB17 AF04 AF05 AF13 CA10 DA53 DA55 DA63 HA14

Claims (7)

[Claims] 1. A support substrate made of a first semiconductor material, and a light-emitting laminated structure disposed on the support substrate,
A quantum well layer made of a second semiconductor material, a pair of carrier confinement layers sandwiching the quantum well layer, and made of a third semiconductor material having a band gap larger than that of the second semiconductor material; And a pair of cladding layers made of a fourth semiconductor material having a larger band gap than the third semiconductor material, sandwiching the three layers with the carrier confinement layer of FIG. The thickness of the second and third semiconductor materials and the thickness of the quantum well layer are selected such that the difference between the level and the ground level of electrons in the quantum well layer is 100 meV or more. A semiconductor light emitting device comprising: a light emitting laminated structure; and an electrode for injecting carriers into the light emitting laminated structure. 2. The semiconductor light emitting device according to claim 1, wherein a band gap of the first semiconductor material is larger than an energy corresponding to a wavelength giving a peak of an emission spectrum of the quantum well layer. 3. A support substrate made of a first semiconductor material, and a light-emitting laminated structure disposed on the support substrate,
A quantum well layer made of a second semiconductor material, a pair of carrier confinement layers sandwiching the quantum well layer, and made of a third semiconductor material having a band gap larger than that of the second semiconductor material; And a pair of cladding layers made of a fourth semiconductor material having a band gap larger than that of the third semiconductor material, sandwiching the three layers with the carrier confinement layer. The second and third semiconductor materials, and the quantum well layer and the quantum well layer, so that emission recombination of electrons and holes occurs in the quantum well layer and emission recombination does not occur in the carrier confinement layer. A semiconductor light-emitting device comprising: the light-emitting stacked structure in which a thickness of a carrier confinement layer is selected; and an electrode for injecting carriers into the light-emitting stacked structure. 4. The semiconductor light emitting device according to claim 1, wherein said first, third and fourth semiconductor materials are GaAs or AlGaAs, and said second semiconductor material is InGaAs. 5. The relationship between the thickness of the quantum well layer and the In composition ratio of the second semiconductor material, wherein an In composition ratio of the second semiconductor material is 0.05 to 0.25. 5. The semiconductor light emitting device according to claim 4, wherein is selected to be 800 to 920 nm. 6. The semiconductor light emitting device according to claim 4, wherein an Al composition ratio of the third semiconductor material is 0.4 or less. 7. The carrier confinement layer has a thickness of 120 nm.
The semiconductor light emitting device according to any one of claims 1 to 6, wherein