JPH09222588A - Optical semiconductor device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain an optical modulator which has the α parameter suitable for long-distance transmission and is low in the insertion loss at the time of a transmissible state by constituting the optical modulator with a quantum box for confining three-dimensional quanta. SOLUTION: Diffraction gratings 3 are formed on part of the surface of an n-type InP substrate 1 and an n-type layer 2 is formed on this substrate 1. A distribution feedback type (DFB) laser 14 and the optical modulator 15 are formed thereon by interposing a separating region 16 therebetween. The DFB laser 14 is formed by laminating an n-type clad layer 2, an active layer 4a, a p-type clad layer 5, a p-type layer 6 and a contact layer 7a on the diffraction gratings 3. The optical modulator 15 is formed by laminating a modulation absorptive layer 4b including a quantum box on the n-type layer 2 and laminating p-type layers 5, 6 and p-type contact layer 7b thereon. A polyimide region 9 is formed on the periphery of the p-side electrode 10b and a wiring layer 11 formed thereon is connected to the p-side electrode 10b. Further, an n-side electrode 12 is commonly formed on the rear surface of the substrate 1.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical semiconductor device, and more particularly, to an optical semiconductor device which changes optical absorption intensity by an applied electric field to perform optical modulation.

[0002]

2. Description of the Related Art A quantum well structure can be formed by sandwiching a thin layer of a semiconductor having a narrow band gap with a layer of a semiconductor having a wide band gap. The case where a single semiconductor thin layer forms quantum wells for both electrons and holes is called a type I quantum well. In this case, a semiconductor layer having a narrow bandgap is called a well layer, and a semiconductor layer having a wide bandgap is called a barrier layer. A type I multiple quantum well structure is formed by alternately stacking a narrow bandgap semiconductor layer and a wide bandgap semiconductor layer.

In addition, a quantum well for electrons is formed in the first semiconductor layer, a quantum well for holes is formed adjacently in the second semiconductor layer, and a barrier is formed for electrons and holes. The structure sandwiched by the semiconductor layers of type I
Called I quantum well. In this case, the stacked structure of the first and second semiconductor layers is called a well layer, and the third semiconductor layer is called a barrier layer. When the first semiconductor layer, the second semiconductor layer, and, if necessary, the third semiconductor layer are sequentially stacked, a type II multiple quantum well is formed.

A well layer of a type I or type II quantum well structure is formed of a graded bandgap material, and a graded bandgap quantum well having a pre-bias function is provided. Bonded quantum wells coupled through a tunnelable thin barrier layer, a coupled superlattice in which a superlattice is formed by coupled quantum wells and states common to many well layers and states resulting from only a single well layer are generated. Etc. are also known.

Such a quantum well structure or multiple quantum well structure can be used for various semiconductor devices. For example, a semiconductor laser using a type I multiple quantum well structure has been manufactured by utilizing the fact that electrons and holes are more stable in the well layer.

In the field of optical communication, light sources and modulators are of great importance. Although it is possible to modulate the output light by turning on / off the driving of the laser light source itself, the stability of the output wavelength is poor and so-called chirping occurs. In order to stably hold the light wavelength and perform high-speed modulation, it is preferable to continuously oscillate the laser light source and perform modulation by the modulator. A modulator using a quantum well structure is being developed.

13A and 13B show an example of a type I quantum well structure. FIG. 13A shows the band energy distribution of a type I quantum well as a function of position in the thickness direction.
A p-type layer, a quantum well structure, and an n-type layer are stacked on an InP substrate to form a pin structure. The barrier layers L1 and L3 in the quantum well structure are formed of In _0.73 Ga _0.27 As _0.49 P _0.51 and have an extension strain of −0.31% with respect to the InP substrate. The energy position at the band edge of the conduction band of the barrier layer L1 is −8.173 eV, the energy position at the band edge of the heavy hole in the valence band is −9.228 eV, and the energy position at the band edge of the light hole is -9.2
It is 06 eV.

[0008] well layer L2 is formed in _{In 0.89 Ga 0.11 As 0.65 P 0.} 35 having a + 1.32% of compressive strain, the energy position -8.344eV the band edge of the conduction band, heavy valence band The energy position at the band edge of the hole is -9.
The energy position at the band edge of the light hole is 138 eV and -9.221 eV. The well layer L2 has a thickness of 9.0 nm, for example. The well layer L2 and the regions of the barrier layers L1 and L3 adjacent to the well layer L2 are non-doped.

Although only a single well layer is shown, the well layers L2 and the barrier layers L1 (L3) may be alternately arranged to form a multiple quantum well. With such a structure, a type I quantum well structure in which a quantum well for electrons and a quantum well for holes are formed in the same layer is formed.

With respect to such a pin structure, in the state where no bias electric field is applied, both the wave function of electrons in the conduction band and the wave function of holes in the valence band are distributed around the well layer L2. The peak of the electron wave function and the peak of the hole wave function coincide. When light having a photon energy having an energy corresponding to the band gap of the well layer L2 is incident on the quantum well structure, electron-hole pairs are generated and the light is absorbed.

Further, since the electrons and holes are confined only in the one-dimensional direction, the energy levels near the band edge are distributed almost continuously. Therefore, electron-hole pairs are generated even for light having a photon energy equal to or more than the energy corresponding to the band gap of the well layer L2, and the light absorption intensity is increased.
Therefore, the wavelength of the light incident on the quantum well structure is set to the well layer L
When the wavelength longer than the wavelength corresponding to the bandgap of 2 is gradually shortened, the light absorption intensity is almost 0 at the beginning, and the light rises at the wavelength corresponding to the bandgap and becomes larger for the shorter wavelengths. The light absorption intensity is shown.

FIG. 13B shows the distribution of the wave function when a bias electric field is applied to the quantum well structure shown in FIG. 13A.
The electrons and holes are subjected to a force by the electric field and change their wave function distribution.

In the figure, the distributions of the electron and hole wavefunctions are shown for four cases in which the polarity of the electric field strength and the strength are different in four stages. In the figure, the solid line shows the wave function of the electron, and the dotted line shows the wave function of the hole (heavy hole). As shown in FIG. 13A, this quantum well structure is symmetric with respect to the thickness direction, and the distribution of the wave function shows a symmetrical shape with respect to the positive electric field application and the negative electric field application. .

When an electric field is applied along the stacking direction, holes move in the direction of the electric field and electrons move in the direction opposite to the electric field. For this reason, the wave functions of electrons and holes, whose peak positions coincided with each other when there was no applied electric field, moved in the opposite directions as the electric field strength increased, and their overlap decreased. A decrease in the overlap of the wave functions means a decrease in the light absorption intensity. The square of the overlap of the wave functions is proportional to the transition probability, that is, the oscillator strength, and represents the light absorption intensity.

The band structure shown in FIG. 13A is inclined by the application of an electric field. When the band of the well layer L2 is inclined downward to the right, for example, electrons are distributed to the right and holes are distributed to the left. Therefore, the energy required to generate electron-hole pairs is reduced by the amount of tilt of the band, and the wavelength of absorbed light moves to the longer wavelength side (red shift).

FIG. 14 shows an example of the optical absorption spectrum of the type I quantum well structure shown in FIG. 13A. The horizontal axis represents wavelength in nm, and the vertical axis represents attenuation of incident light in dB. That is, the vertical axis corresponds to the light absorption intensity of the quantum well structure.

When no electric field is applied, the light absorption intensity has a wavelength of about 1
It has a maximum at 480 nm and decreases as the wavelength becomes longer. The light absorption intensity becomes almost zero in the wavelength region of 1560 nm or more. Even if the wavelength becomes shorter than 1480 nm, a relatively large light absorption intensity is maintained.

When an electric field having a strength of -50 kV / cm is applied, the absorption peak shifts to the long wavelength side, and the wavelength is about 1495.
A weak absorption peak appears near nm. The application of the electric field reduces the overlap of the electron and hole wavefunctions, and thus the height of the absorption peak becomes low. Also, by applying an electric field, a wavelength of about 1
In the region of 500 nm or more, the light absorption intensity is higher than that when no electric field is applied. Light modulation can be performed by utilizing this change in light absorption intensity.

Next, the relationship between the α parameter and the maximum transmission distance of the optical pulse signal will be described with reference to FIG. The α parameter of the optical pulse source is

[0020]

## EQU1 ## It is defined as α = Δθ / ΔAm = Δn / ΔAb. Where Δθ is the phase change of the optical pulse, Δ
Am is the change in amplitude, Δn is the change in the refractive index of the optical modulator, Δ
Ab represents a change in the light absorption intensity of the optical modulator.

FIG. 15 shows the relationship between the α parameter and the maximum transmission distance. The horizontal axis represents the α parameter, and the vertical axis represents the transmission distance in the unit of km. Dispersion coefficient of transmission medium is 17 pS /
nm / km optical fiber, frequency of optical pulse signal is 10
Gb / s, optical pulse waveform was Gaussian waveform, and power penalty was 1 dB.

When the α parameter is about -1, the maximum transmission distance shows the maximum value. When the α parameter is increased in the positive direction, the maximum transmission distance is sharply shortened, and when it is increased in the negative direction, the maximum transmission distance is gradually shortened. Therefore, in order to perform long-distance transmission, it is preferable to make the α parameter negative, and
More preferably, it is set to 0.

FIG. 16 shows the wavelength dependence of the light absorption intensity change, refractive index change, and α parameter of the quantum well structure having the characteristics shown in FIG. The horizontal axis represents wavelength in nm, the left vertical axis represents light absorption intensity change ΔAb on an arbitrary scale, and the right vertical axis represents refractive index change Δn and α parameter. The solid line in the figure indicates the change in light absorption intensity ΔAb, the symbol ▪ indicates the change in refractive index Δn, and the symbol ◯ indicates the α parameter. The refractive index change Δn is calculated using the Kramers-Kronig equation.

For long-distance transmission, it is preferable that the optical absorption intensity of the optical modulator is substantially zero in the light transmitting state in order to reduce the insertion loss of the optical modulator. Therefore,
The optical modulator using the quantum well structure having the characteristics shown in FIG. 14 is suitable for modulating light in the wavelength range of 1560 nm or more from the viewpoint of light absorption intensity. However, in FIG. 16, the α parameter is positive in the wavelength range of 1560 nm or more. From the viewpoint of the α parameter, this wavelength range is not suitable for long-distance transmission.

Further, since the α parameter becomes negative in the wavelength region of 1500 to 1530 nm, from the viewpoint of the α parameter, this optical modulator is suitable for modulating light in the wavelength region of 1500 to 1530 nm. However, as can be seen from FIG. 14, since the light absorption intensity in this wavelength range is relatively large, this optical modulator is 1500
It is not suitable for modulating light in the wavelength range of ˜1530 nm. As described above, the optical modulator using the type I quantum well structure that performs one-dimensional confinement does not have a wavelength range in which both the optical absorption intensity and the α parameter are suitable values for long-distance transmission.

17A and 17B show a type II quantum well. FIG. 17A shows the band energy distribution of a type II quantum well as a function of position in the thickness direction. In the figure, a barrier layer M1, a well layer M2 for electrons, a well layer M3 for holes, and a barrier layer M4 are stacked to form a type II quantum well structure. The p-type layer is continuous on the side of the barrier layer M1 and the n-type layer is continuous on the side of the barrier layer M4.

A layer M2 forming a well layer for electrons and a layer M3 forming a well layer for holes are arranged adjacent to each other, and barrier layers M1 and M4 are arranged on both sides thereof. In the case of the multi-well structure, the layers M1 and M
4 is composed of the same layer.

For example, this quantum well structure is formed on an InP substrate, and the layer M1 (M4) is made of In _0.73 Ga _0.27 As.
It is a layer with a tensile strain of -0.31% formed by _0.49 P _0.51 and the energy position at the band edge of the conduction band is -8.173e.
The energy position at the band edge of the heavy hole in V and the valence band is −9.228 eV, and the energy position at the band edge of the light hole in the valence band is −9.206 eV.

The well layer M2 for electrons is formed of, for example, InAs _0.43 P _0.57 having a thickness of 9.0 nm, and has a compressive strain +
With 1.37%. The energy position at the band edge of the conduction band is −8.295 eV, the energy position at the band edge of the heavy hole in the valence band is −9.220 eV, and the energy position at the band edge of the light hole in the valence band is −9.307 eV.
It is.

Layer M3 forming a quantum well for holes
Is, for example, 12.0 nm thick In _0.53Ga_0.47As
_0.71P_0.29It has a tensile strain of -1.0%. This
The energy position at the band edge of the conduction band of the layer is −8.169.
eV, energy level at band edge of heavy hole in valence band
The position is -9.178 eV, the valence band light hole van.
The energy position at the terminal is −9.105 eV.

Similar to FIG. 13B, FIG. 17B shows the movement of the wave function with respect to the applied electric field. In the type II quantum well, the position of the electron quantum well and the position of the hole quantum well are different even when no electric field is applied, so that the wave functions of the electron and the hole exist separately. When the thickness of both quantum well layers is equal to or less than the spread of the wave function, the wave function penetrates the barrier layer and is distributed.

When an electric field in which electrons and holes are separated from each other is applied, as shown in the upper two graphs in FIG. 17B,
The electron and hole wavefunctions are distributed further apart. On the other hand, when an electric field in the direction in which electrons and holes approach each other is applied, the wave functions of electrons and holes are biased to the boundary side of both well layers and further enter into the barrier layer, and the overlap of wave functions increases.

Further, in FIG. 17A, when an electric field is applied such that the band structure is downward-sloping to the left, the wave function of electrons and the wave function of holes are distributed apart from each other. in this case,
The transition energy decreases by an amount corresponding to the band slope, and the transition wavelength moves to the long wavelength side (red shift).

When an electric field in which the electron wave function and the hole wave function come close to each other (to the lower right) is applied, the band gap at the boundary does not change, but the energy at the band edge at the bottom of the well layer has a slope. As a result, the distribution of the wave function is gathered near the boundary and the effective well width is reduced. The transition energy increases slightly. This increase in transition energy results in a shorter transition wavelength (blue shift).

Conventionally, when the quantum well structure is used in an optical modulator, the red shift of the transition wavelength due to the application of an electric field is used. Light having a wavelength slightly longer than the light absorption edge energy when no electric field is applied is used as the input light. When the red shift occurs due to the application of the electric field, the light absorption edge wavelength moves to the long wavelength side. As a result, the wavelength of the input light in the transparent region becomes the wavelength of the absorption region. Thus, when the transition wavelength is red-shifted by applying the electric field and the light modulation is performed by utilizing the light absorption when the electric field is applied, the α parameter becomes positive, and this method is not suitable for long-distance transmission.

Since the wavelength region on the shorter wavelength side than the light absorption edge energy when no electric field is applied becomes an absorption region both when no electric field is applied and when an electric field is applied, light in this wavelength range cannot be used as modulated light. Not preferable.

[0037]

As described above, in the optical modulator using the quantum well structure for one-dimensional confinement, the transition wavelength is red-shifted by applying the electric field, and the optical absorption when the electric field is applied is utilized. Since the optical modulation is performed, the α parameter becomes positive, which is not suitable for long-distance transmission.

An object of the present invention is to provide an optical semiconductor device having an α parameter suitable for long distance transmission.

[0039]

According to one aspect of the present invention, a light emitting means for emitting light and a three-dimensional quantum for electrons and holes are arranged in the optical path of the light emitted from the light emitting means. An optical semiconductor device having a quantum box for confining and an electric field applying means for applying an electric field to the quantum box is provided.

In the quantum box, energy levels are discretely distributed in the conduction band and the valence band. When light having energy equal to the difference between the energy level of the conduction band and the energy level of the valence band is incident, electron-hole pairs are generated and light absorption occurs. The absorption peaks of the optical absorption spectrum also appear discretely corresponding to the energy levels distributed discretely. When an electric field is applied to the quantum box, the energy difference between energy levels fluctuates, and the absorption peak shifts to the long wavelength side or the short wavelength side.

When light in the wavelength range within the absorption peak shift range is incident on the quantum box and the electric field intensity is changed, the light absorption intensity for the incident light changes. Light modulation can be performed by utilizing this change in light absorption intensity.

According to another aspect of the present invention, when no electric field is applied, the spatial coordinates giving the peaks of the wavefunctions of the electrons and holes confined in the quantum box are three-dimensionally coincident, and the light emitting means is provided. Provided is an optical semiconductor device in which light absorption intensity by the quantum box when an electric field is applied to the quantum box is smaller than light absorption intensity when an electric field is not applied with respect to light having a wavelength emitted from the quantum box. It

When the electric field is not applied, the peaks of the wavefunctions of the electrons and holes coincide with each other, so that the transition probability increases and the light absorption intensity increases. When an electric field is applied, the peaks of the electron and hole wavefunctions move in opposite directions, so that the wavefunction overlaps decrease and the transition probability decreases. Therefore, the light absorption intensity decreases. At the same time, since the band edges of the conduction band and the valence band are inclined, the energy gap is narrowed and the absorption peak moves to the long wavelength side.

When light having a wavelength range in which the light absorption intensity when the electric field is applied is smaller than the light absorption intensity when the electric field is not applied is the modulated light, the non-transmission state occurs when the electric field is not applied and the transmission state occurs when the electric field is applied. .

According to another aspect of the present invention, the wavelength of the light emitted by the light emitting means is the wavelength corresponding to the peak on the longest wavelength side among the peaks of the light absorption spectrum by the quantum box. Provided is an optical semiconductor device having a wavelength shorter than the wavelength obtained by adding 1/2 of the half width.

In order to use the optical modulator as an optical signal generation source for long distance transmission, it is desired to reduce the insertion loss in the transmission state of the optical modulator. Since discrete absorption peaks appear in the optical absorption spectrum of the quantum box, a wavelength region having a sufficiently small optical absorption peak exists on the shorter wavelength side than the absorption peak on the longest wavelength side.

Further, for light in a wavelength range shorter than the wavelength corresponding to the absorption peak plus 1/2 of the half-value width of the peak, the absorption peak is shifted to the long wavelength side to absorb light. The strength can be made sufficiently small. Therefore,
Light in this wavelength range can be used as modulated light.

According to another aspect of the present invention, the wavelength of the light emitted by the light emitting means is half of any peak of the optical absorption spectrum of the quantum box when an electric field is not applied to the quantum box. Provided is an optical semiconductor device having a wavelength within a value range.

It is possible to obtain a relatively large light absorption intensity when no electric field is applied to light within the range of the half-value width of the absorption peak when no electric field is applied. The absorption peak shifts when an electric field is applied. By appropriately selecting the electric field intensity, the light absorption intensity can be made smaller than that when no electric field is applied.

According to another aspect of the present invention, when no electric field is applied, the spatial coordinates giving the peaks of the wavefunctions of the electrons and holes confined in the quantum box are the electric fields applied by the electric field applying means. The light absorption intensity by the quantum box when an electric field is applied to the quantum box is different from that of the light emitted from the light emitting means when the electric field is not applied. An optical semiconductor device is provided that is greater than strength.

Since the peak positions of the wave functions of electrons and holes are deviated when no electric field is applied, the transition probability is relatively low and the light absorption intensity is also small. When an electric field is applied in the direction in which the peak position of the wave function approaches, the transition probability increases and the light absorption intensity also increases.

According to another aspect of the present invention, the wavelength of light emitted by the light emitting means is such that the peaks of the light absorption spectra of the quantum boxes adjacent to each other when no electric field is applied to the quantum boxes. An optical semiconductor device having an intermediate wavelength is provided.

When an electric field is applied to the quantum box in such a direction that the peak positions of the electron and hole wavefunctions approach each other, the height of the absorption peak increases and the central wavelength of the absorption peak shifts to the short wavelength side. If the wavelength of the modulated light is set to an intermediate wavelength between two adjacent absorption peaks when no electric field is applied, the modulated light is in a transmissive state when no electric field is applied. When the absorption peak is shifted to the wavelength range of the modulated light by applying an electric field of appropriate strength, the non-transmission state is set.

[0054]

1 is a partially cutaway perspective view of a modulator integrated distributed feedback laser according to an embodiment of the present invention. n
The diffraction grating 3 is formed on a part of the surface of the type InP substrate 1. An n-type layer 2 is grown on the substrate 1, and a distributed feedback (DFB) laser 14 and an optical modulator 15 are formed on the n-type layer 2.
Are formed with the isolation region 16 interposed.

The DFB laser 14 is provided on the diffraction grating 3 with n
The clad layer 2, the active layer 4a, the p-type clad layer 5, the p-type layer 6 and the contact layer 7a are formed by being laminated. The optical modulator 15 is configured by laminating a modulation absorption layer 4b including a quantum box on the n-type layer 2, and laminating p-type layers 5, 6 and a p-type contact layer 7b on the modulation absorption layer 4b.

The isolation region 16 is realized by a structure in which the intermediate region between the contact layers 7a and 7b is removed from the upper surface of the laminated structure. A mesa structure reaching a substrate having a width of about 1 to 3 μm is formed by etching along the alignment direction of the DFB laser 14 and the optical modulator 15, and the side surface of the mesa structure is backfilled with the semi-insulating buried hetero region 8. Has been done. Also,
The p-side electrode 1 is formed on the contact layer 7a of the DFB laser 14.
0a is formed, and the p-side electrode 10b is formed on the contact layer 7b of the optical modulator 15.

Further, the polyimide region 9 is the p-side electrode 10b.
The wiring layer 11 formed on the periphery and formed on the periphery is connected to the p-side electrode 10b. An n-side electrode 12 is commonly formed on the back surface of the substrate 1.

Such a structure can be produced by using techniques such as selective epitaxial growth on a substrate, selective etching, mesa burying growth, polyimide coating, substrate polishing, electrode layer deposition, patterning, and cleavage.

In such a configuration, the DFB laser 1
4 continuously oscillates light of a single wavelength. The optical modulator 15 is
The light emitted from the DFB laser 14 is selectively absorbed to generate modulated output light.

Next, an optical modulator according to the first embodiment of the present invention will be described with reference to FIGS. Figure 2A
1 schematically shows one of a plurality of quantum boxes included in the modulation absorption layer 4b of the optical modulator according to the first embodiment. The quantum box has a rectangular parallelepiped shape in which each surface is perpendicular to the x axis, the y axis, and the z axis. 2A shows a quantum box having a rectangular parallelepiped shape, other quantum boxes such as a cylindrical box, a spherical box, and an elliptic box may be used.

FIG. 2B shows an energy level diagram in the x-axis, y-axis and z-axis directions of the quantum box. In each axial direction, the lowest energy level of the conduction band in the quantum box is lower than the lowest energy level around it, forming a quantum well for electrons. Further, the highest energy level of the valence band in the quantum box is higher than the highest energy level of its surroundings, forming a quantum well for holes. That is, the quantum box performs three-dimensional quantum confinement on electrons and holes, and when the electric field is not applied, the peak of the wave function of the confined electrons and holes is 3
The spatial coordinates of the axes all match. This is a structure similar to a type I quantum well layer.

FIG. 2C shows the optical absorption spectrum of the quantum box. Since the energy levels in the quantum box are distributed discretely,
The absorption peak of the light absorption intensity is also discretely distributed with respect to the wavelength of the incident light. The absorption peak at wavelength λ ₁ corresponds to the minimum energy gap of the quantum box, and no peak appears on the longer wavelength side. Absorption peaks corresponding to the second and third energy gaps appear on the shorter wavelength side than the wavelength λ ₁ .

If all of the plurality of quantum boxes forming the optical modulator have the same size and good crystallinity, the light absorption intensity between the absorption peaks adjacent to each other of the light absorption intensity becomes sufficiently small.

For reference, FIG. 2D shows an optical absorption spectrum of a quantum well layer that performs one-dimensional quantum confinement for electrons and holes and a quantum wire that performs two-dimensional quantum confinement.
In this case, since the energy levels are distributed almost continuously with a constant width in the vicinity of the band edge, no discrete peak appears and the light absorption intensity changes stepwise. Therefore, in order to reduce the insertion loss of the optical modulator, the wavelength of the modulated light needs to be longer than the wavelength at the absorption edge.

In the case of FIG. 2C, since the light absorption intensity is small on both sides of the absorption peak of the light absorption spectrum,
The wavelength of the modulated light may be set not only on the long wavelength side of each absorption peak but also on the short wavelength side. Later, when the optical absorption spectrum of the quantum box produced by the method described with reference to FIGS. 6 to 8 was measured, it was confirmed that the light of the wavelength corresponding to each of the absorption peak on the longest wavelength side and the second absorption peak was detected. The energy difference is about 60 meV, and the half-width of each absorption peak is about 1
It was about 0 meV. Therefore, there is a wavelength range in which the light absorption intensity is sufficiently small between the two absorption peaks. Light in this wavelength range can be used as modulated light.

When the wavelength becomes shorter, the energy difference between the absorption peaks adjacent to each other becomes smaller, and on the shorter wavelength side than the wavelength corresponding to the band gap of the material around the quantum box,
Since the energy levels exist almost continuously, no sharp absorption peak appears. Therefore, it is preferable to use light having a wavelength near the absorption peak on the long wavelength side as the modulated light.
By reducing the size of the quantum box, it is possible to increase the energy difference of light having wavelengths corresponding to absorption peaks adjacent to each other in the light absorption spectrum.

FIG. 3 shows the optical absorption space of the quantum box shown in FIG. 2A.
The calculation result of the absorption peak on the longest wavelength side of the koutor is shown.
The size of the quantum box is a cube with sides of 15 nm,
Material of Ga_0.415In_0.585As_0.890P_0.110When
The material around the quantum box to Ga _0.146In_0.854As
_0.318P_0.682And In addition, the peak shape has a half width of 20
/ E^TwoA Gaussian waveform of [meV] is assumed. here
Where e represents the base of the natural logarithm.

When no electric field is applied, an absorption peak centering on a wavelength of 1550 nm appears. When an electric field is applied in the z-axis direction in FIG. 2A and its strength is gradually increased, the central wavelength of the absorption peak shifts to the long wavelength side (red shift) and the height of the absorption peak decreases. The red shift is due to the smaller band gap of the quantum box. Further, the height of the absorption peak is lowered because the overlap of the wave functions of electrons and holes is reduced.

FIG. 4 shows the applied electric field from -150 kV / cm to -100 kV in the light absorption spectrum shown in FIG.
The change in the light absorption intensity and the α parameter when changed to / cm are shown. The horizontal axis represents wavelength in nm, the left vertical axis represents changes in light absorption intensity on an arbitrary scale, and the right vertical axis represents α parameter. The solid line in the figure indicates the light absorption intensity, and the symbol ■
Indicates the α parameter.

The wavelength is 1580 nm or less and 1640 to 1
The α parameter is negative in the 700 nm region. For example, if the wavelength of the modulated light is 1550 nm, the intensity-
A transparent state is obtained when an electric field of 150 kV / cm is applied, and a non-transparent state is obtained when no electric field is applied. The α parameter at this time is about −1.1 as shown in FIG.

FIG. 5 shows changes in the light absorption intensity and the α parameter with respect to the electric field intensity when the wavelength of the modulated light is 1550 nm. The horizontal axis represents the electric field intensity in kV / cm, the left vertical axis represents the light absorption intensity on an arbitrary scale, and the right vertical axis represents the α parameter. The solid line in the figure indicates the light absorption intensity,
The symbol ■ indicates an α parameter. The α parameter is
Since it is determined by the change of the light absorption intensity and the refractive index when the electric field intensity is changed, it is plotted at the center position of two points of the calculated electric field intensity.

As described with reference to FIG. 4, the electric field strength is −15.
Α when changing from 0 kV / cm to −100 kV / cm
The parameter is negative. The light absorption intensity at this time is relatively small, and for example, when the applied electric field is -150 kV / cm, it can correspond to the transmission state. Α when the electric field strength changes from −50 kV / cm to 0 kV / cm
The parameter is about +0.8. At this time, the light absorption intensity is large, and it is possible to correspond to the non-transmissive state when no voltage is applied, for example.

As described above, since the α parameter in the transmission state becomes negative, it is suitable for long-distance transmission. Further, the light absorption intensity in the transmitted state is very small as shown in FIG. 3 and is suitable for long-distance transmission. It should be noted that the α parameter in the non-transmissive state becomes positive, but since no optical pulse signal is output at this time, there is no problem in signal transmission.

In the case of an optical modulator composed of a quantum well layer or a quantum wire, the wavelength of the modulated light needs to be longer than the absorption edge of the optical absorption spectrum in order to reduce the insertion loss as described above. . On the other hand, in the case of the optical modulator configured by the quantum box, for example, as described in FIG. 3, the wavelength of the modulated light is the absorption edge (about 165 in FIG. 3) when no electric field is applied.
The wavelength can be shorter than 0 nm).

Further, the optical modulator composed of the quantum boxes is in the non-transmissive state when the light absorption intensity is maximum, that is, when the overlap of the wave functions of electrons and holes is maximum. On the other hand, when a type I quantum well layer that performs one-dimensional quantum confinement is used, the state in which an electric field is applied to the quantum well layer, that is, the overlapping of the electron and hole wave functions is reduced, and the optical absorption intensity is reduced. The state where is small is the non-transmissive state.
Therefore, by using the quantum box, the amount of attenuation in the non-transmissive state can be increased as compared with the case of using the quantum well layer. Therefore, the optical modulator can be downsized.

When the quantum box is used as a laser diode, it is a serious problem that it takes a relatively long time for the electrons and holes to be captured by the quantum box. But,
Since the process of generating electron-hole pairs by electric field absorption is sufficiently fast, the operating speed is not limited by the trapping time as in the case of the laser diode.

In the above example, the wavelength of the modulated light is set to 1550.
However, other wavelengths may be used. For example, it may be selected from a wavelength range in which the light absorption intensity when an electric field is applied is smaller than the light absorption intensity when no electric field is applied. FIG.
For example, in the wavelength range of 1600 nm or less, the light absorption intensity when an electric field having a strength of −150 kV / cm is applied is smaller than the light absorption intensity when no electric field is applied.
As shown in FIG. 4, even if the α parameter becomes negative or positive in this wavelength range, it becomes relatively small at 0.2 or less, which is suitable for long-distance transmission.

Further, the wavelength of the modulated light is shorter than the wavelength corresponding to the absorption peak on the longest wavelength side of the light absorption spectrum when no electric field is applied, plus one half the half-value width of the absorption peak. You may select from a wavelength range. Taking the case of FIG. 3 as an example, this corresponds to a wavelength range of 1585 nm or less.

FIG. 3 shows the case where the red shift of the absorption peak on the longest wavelength side of the light absorption spectrum is used, but the red shift of other absorption peaks may be used.
In this case, in order to secure a sufficient difference in light absorption intensity between the transmission state and the non-transmission state, it is preferable to set the wavelength of the modulated light to a wavelength within the range of the half-value width of the absorption peak when no electric field is applied. .

Next, a method of manufacturing the quantum box shown in FIG. 2A will be described with reference to FIGS. FIG. 6 shows a semiconductor laminated structure including a quantum box. Ga on the GaAs substrate 31
As buffer layer 32, InGaAs layer 33 having quantum dots 35 arranged in a scattered manner, and GaAs cap layer 3
4 are epitaxially grown. In the semiconductor laminated structure including the quantum boxes, the layer having the quantum boxes 35 (hereinafter, referred to as “quantum box layer”) 33 is the buffer layer 3 in this manner.
2 and the cap layer 34.

FIG. 7 schematically shows a low pressure MOCVD (metal organic chemical vapor deposition) apparatus for forming the semiconductor laminated structure shown in FIG. However, the crystal growth is performed by alternately supplying the raw materials. Alternatively, ALE (atomic layer epitaxial growth) may be used. A gas channel 45 opens below the reaction vessel 40, and a reaction gas is introduced into the reaction vessel 40 from the gas channel 45. The reaction gas introduced into the reaction container 40 is exhausted to the outside from a gas exhaust pipe 44 provided above the reaction container 40.

A susceptor 41 is arranged in the reaction container 40, and the substrate 31 is held at a position facing the opening of the gas passage 45. A high frequency coil 42 is arranged around the reaction container 40 so as to surround the susceptor 41, and the susceptor 41 and the substrate 31 can be heated by high frequency.

A carrier gas and H ₂ gas as a purge gas are supplied from the gas flow path 50 to the gas supply system. H ₂ gas branched from the gas flow path 50, and the gas flow path H ₂ gas branched bubbling TMIDMEA (trimethylindium dimethylethylamine adduct) from the gas passage 50 through the mass flow controller MFC, respectively 5
1 is supplied.

[0084] Similarly, H ₂ gas branched from the gas flow path 50, and TMG H ₂ gas was bubbled (trimethyl gallium) is supplied to the gas passage 52 through the mass flow controller MFC, respectively.

Gas flow paths 51 and 52 are gas switching valves
It is connected to the input side of 55. Gas switching valve 5
The output side of 5 is for supplying the reaction gas to the reaction vessel 40.
Connected to the gas flow channel 45 and the exhaust gas flow channel 57 of
You. Use the gas switching valve 55 to switch the gas flow path.
As a result, H containing TMIDMEA in the gas flow channel 45 _Two
H containing gas or TMG_TwoGas, or both
Can be supplied. In addition, exhaust both gases
It can also be exhausted to the flow passage 57.

The H ₂ gas branched from the gas flow passage 50 is supplied to the gas flow passage 53 via the mass flow controller MFC. The H ₂ gas branched from the gas flow path 50 and AsH ₃ (arsine) are supplied to the gas flow path 54 through the mass flow controller MFC.

The gas flow paths 53 and 54 are connected to the input side of the gas switching valve 56. Gas switching valve 5
The output side of 6 is connected to the gas flow passage 45 and the exhaust gas flow passage 58, similarly to the gas switching valve 55. By switching the gas flow path with the gas switching valve 56, the purging H ₂ gas or AsH ₃ gas is supplied to the gas flow path 45.
A mixed gas of H ₂ and H ₂ can be supplied.

Next, a method of forming the laminated structure shown in FIG. 6 by using the low pressure MOCVD apparatus shown in FIG. 7 will be described. First, the GaAs substrate 31 is held on the susceptor 41. The susceptor 41 is heated by the high frequency coil 42 while flowing H ₂ gas as a purge gas, and the substrate temperature is 460 ° C.
And The exhaust amount is controlled so that the pressure inside the reaction container 40 becomes 2000 Pa.

When the substrate temperature reaches 460 ° C., TM
The flow rate of H ₂ gas containing G is 25 sccm, AsH ₃ and H ₂
Is supplied at a flow rate of 100 sccm to deposit a 100 nm thick GaAs buffer layer 32 by MOCVD.

Next, TMIDMEA, TMG and AsH
₃ is supplied by switching the time, and InGa with a thickness of 7 nm is supplied.
Deposit the As quantum box layer 33. FIG. 8 shows TMIDME
A, shows a time chart of the supply of TMG and AsH _3. First, H ₂ gas is flowed for 0.5 seconds to purge the inside of the reaction vessel 40. Subsequently, the H ₂ gas bubbling with TMIDMEA was flowed at a flow rate of 200 sccm for 1.0 second, and then the H ₂ gas bubbling with TMG was flowed at a rate of 35 sccc.
Supply for 0.1 second under the condition of m. H ₂ gas is supplied for 0.5 seconds to purge the inside of the reaction vessel 40. Then, AsH ₃
A mixed gas of H ₂ and H ₂ is supplied for 10 seconds at a flow rate of 400 sccm.

The above gas supply sequence is set as one cycle, and gas supply is repeated for 12 cycles. Then, GaAs
A GaAs cap layer 34 having a thickness of 100 nm is deposited under the same conditions as the formation of the buffer layer 32. It is also possible to use another source gas as the organometallic material.

InG which is normally formed when TMIDMEA, TMG, and AsH ₃ are divided in time and supplied.
aAs has an In composition specified by the supply time of the source gas of each constituent element and the like.

When InGaAs is deposited on the GaAs buffer layer, InGaAs with a uniform In composition is formed on the entire surface.
If a layer is formed, the InGaA due to lattice mismatch
In-plane strain occurs in the s layer. InGaA
Ga of the underlying layer is spread over a wider area than strain is generated on the entire surface of the s layer.
It is considered that the strain energy becomes lower when the region having a small In composition, which has a lattice constant substantially matching that of the As buffer layer, is formed, and the regions having a large In composition are locally formed in a scattered manner. Further, it is considered that the strain energy becomes lower as the region having a large In composition becomes closer to a spherical shape.

From the above reasons, it is considered that island-like regions having a large In composition are formed in a scattered manner in the region having a low In composition as shown in FIG. When the laminated structure was formed under the above conditions, island-shaped regions having a diameter of about 10 nm were formed. The island-like region having a large In composition is formed in the surrounding region having a low In composition, the GaAs buffer layers above and below, and GaA.
Since the energy gap is smaller than that of the s cap layer,
It acts as a three-dimensional quantum well for carriers.

Although FIG. 8 shows the case where the In raw material, the Ga raw material, and the As raw material are supplied in this order, they may be supplied in another order. For example, the order of supplying the Ga raw material and the In raw material may be reversed. Further, the As raw material may be further supplied between the supply of the In raw material and the supply of the Ga raw material.

In addition, in FIG.
Although the case where the hydrogen gas for purging is supplied at the time of switching to the supply of the raw material and at the time of switching from the supply of the As raw material to the supply of the In raw material has been shown, even if the hydrogen gas for purging is supplied at the time of switching the other raw materials, Good.

The number of gas supply cycles is not limited to 12 times. The gas supply cycle is preferably repeated 6 times or more, more preferably 10 to 24 times.

Although the case where the growth temperature of the quantum box layer is set to 460 ° C. has been described, other temperatures may be set. The preferred growth temperature range is 250 ° C to 600 ° C, and the more preferred range is 420 ° C to 500 ° C.

Next, an optical modulator according to the second embodiment of the present invention will be described with reference to FIGS. FIG. 9A
2 schematically shows one of a plurality of quantum boxes included in the modulation absorption layer 4b of the optical modulator according to the second embodiment. A quantum box Qe that performs three-dimensional quantum confinement for electrons and a quantum box Qh that performs three-dimensional quantum confinement for holes are arranged side by side in the z-axis direction. FIG. 9A shows a case where the quantum boxes Qe and Qh have a rectangular parallelepiped shape, but may have other shapes.

9B and 9C show the quantum box Qe, respectively.
3A and 3B show energy level diagrams in the x-axis and y-axis directions at positions crossing Qh and Qh. FIG. 9D shows an energy level diagram in the z-axis direction. Since the energy level of the conduction band of the quantum box Qe is lower than the energy levels around it, electrons are confined in the quantum box Qe. Since the energy level of the valence band of the quantum box Qe is lower than the energy levels of its surroundings (the energy level is higher for holes), the quantum box Qe
Holes are not confined in e. Similarly, holes are confined and electrons are not confined in the quantum box Qh.

As shown in FIG. 9D, the quantum boxes Qe and Qh are displaced in the z-axis direction. Therefore, the quantum box Q
The wave function of the electron confined in e and the wave function of the hole confined in the quantum box Qh have different peak positions in the z-axis direction. The quantum boxes Qe and Qh are similar to the type II quantum well structure in the z-axis direction.

FIG. 10 shows the calculation result of the absorption peak on the longest wavelength side of the optical absorption spectrum when an electric field in the z-axis direction is applied to the quantum box of FIG. 9A. That is, it is an absorption absorption peak corresponding to the transition between the lowest energy level of the conduction band of the quantum box Qe and the highest energy level of the valence band of the quantum box Qh. An absorption peak corresponding to the transition between other energy levels appears on the shorter wavelength side than this absorption peak.

The quantum boxes Qe and Qh have a length in the x-axis and y-axis directions of 15 nm, and the quantum box Qe has a length in the z-axis direction of 9 nm.
nm, the length of the quantum box Qh in the z-axis direction is 4.5 nm,
The material of the quantum box Qe is Ga _0.084 In _0.916 As _0.652 P
_0.348 , quantum box Qh material is Ga _0.708 In _0.292 As
_0.769 P _0.231, the surrounding of the quantum box material Ga _0.573 an In
_0.427 As _0.774 P _0.226 . Further, the shape of the absorption peak was assumed to be a Gaussian waveform with a half width of 20 / e ² meV, as in the case of FIG.

Since the second absorption peak appears at a position sufficiently distant from the first absorption peak shown in FIG. 10, there is a wavelength region in which the light absorption intensity is sufficiently small on the short wavelength side of the first absorption peak.

When no electric field is applied, a weak absorption peak with a wavelength of about 1625 nm appears. The central wavelength of the absorption peak becomes short when the electric field in the direction in which the overlap of the wave functions of electrons and holes confined in the quantum box becomes large (the positive direction of the z-axis) and its strength is gradually increased. It shifts to the wavelength side (blue shift) and its height increases.

The height of the absorption peak is increased because the overlap of the wave functions of electrons and holes is increased and the transition probability is increased. The central wavelength of the absorption peak is blue-shifted because the band edges of the conduction band and the valence band are tilted and the energy gap becomes large.

Since there is a wavelength region in which the light absorption intensity is sufficiently small on the short wavelength side of the first absorption peak when no electric field is applied,
Light in this wavelength range, for example, light having a wavelength of 1550 nm can be used as the modulated light.

FIG. 11 shows that the wavelength of the modulated light is 1550 nm.
Then, the light absorption intensity and the α parameter will be shown with respect to the electric field intensity. The horizontal axis represents the electric field intensity in kV / cm, the left vertical axis represents the light absorption intensity on an arbitrary scale, and the right vertical axis represents the α parameter. The solid line in the figure indicates the light absorption intensity,
The symbol ■ indicates an α parameter.

The electric field strength is changed from 0 kV / cm to 50 kV / c.
The α parameter when changed to m is about −0.5. The light absorption intensity at this time is relatively small, and for example, the state when no electric field is applied can correspond to the transmissive state. The α parameter is about +0.2 when the electric field strength is changed from 100 kV / cm to 150 kV / cm. At this time, the light absorption intensity is relatively high, for example, an intensity of 150 kV /
A state in which an electric field of cm is applied can correspond to a non-transmissive state.

As described above, since the α parameter in the transmission state becomes negative, it is suitable for long-distance transmission. In the case of the optical modulator according to the second embodiment, unlike the case of the optical modulator according to the first embodiment, it is in a non-transmissive state when an electric field is applied. Therefore, electron-hole pairs are generated in the quantum box when an electric field is applied. The generated electrons and holes move by the applied electric field and are quickly removed from the quantum box. As described above, in the optical modulator according to the second embodiment, the generated electron-hole pairs can be efficiently removed from the quantum box.

In FIG. 11, the wavelength of the modulated light is 1550n.
Although m is shown, other wavelengths may be used.
For example, the wavelength range may be in the middle of absorption peaks adjacent to each other when no electric field is applied. When an electric field having an appropriate strength is applied, the absorption peak is blue-shifted to increase the light absorption intensity for the modulated light, resulting in a non-transmission state.

In the second embodiment, the case where the blue shift of the absorption peak on the longest wavelength side was used was explained.
The blue shift of other absorption peaks may be used.
In order to utilize the blue shift of the second and subsequent absorption peaks, it is preferable to make the energy difference between the second and subsequent absorption peaks and the next absorption peak as large as possible.

In order to increase the energy difference between absorption peaks, the size of the quantum box may be reduced. But,
In FIG. 9A, when the length of the quantum box in the z-axis direction is shortened, the blue shift amount of the absorption peak when an electric field is applied decreases. Therefore, it is preferable to reduce the size in the xy plane.

Next, referring to FIGS. 12A to 12C, FIG.
A method of manufacturing the quantum box shown in A will be described. In Figure 12A
As shown, on the surface of the InP substrate 61, for example, MOC
By VD, barrier layer 62, hole quantum box layer 63, electron
The quantum box layer 64 and the barrier layer 65 are deposited. Barrier layer 6
2 and 65 are Ga_0.573In_0.427As_0.774P_0.226
It is a layer consisting of. The quantum box layer 63 for holes is Ga_0.708
In_0.292As_0.769P_{0.23 1}Consisting of 4.5 nm
Layers. The quantum box layer 64 for electrons is Ga_0.084In
_0.916As_0.652P_0.348A 9 nm thick layer of
You.

A mask pattern 66 made of SiO ₂ is formed on the surface of the barrier layer 65 by electron beam exposure. The mask pattern 66 has, for example, a side length of 10 to 10.
It has a rectangular shape of 50 nm.

As shown in FIG. 12B, the mask pattern 6
Using 6 as an etching mask, the laminated structure from the barrier layer 65 to the barrier layer 62 is etched. Etching is
For example, reactive ion etching (RIE) using a mixed gas of C ₂ H ₆ and H ₂ or a mixed gas of CH ₄ and H _{2 as} an etching gas is performed. Thus, the mesa 67 including the barrier layer 62a, the hole quantum box 63a, the electron quantum box 64a, the barrier layer 65a, and the mask pattern 66 is formed.

As shown in FIG. 12C, the MOP does not grow on the surface of the mask pattern 66.
Under the condition that the barrier layer 62 grows only on the surface of the substrate 61,
A barrier layer 67 having the same composition as a and 65a is deposited. The upper surface of the barrier layer 67 is at least higher than the upper surface of the electron quantum box 64a.

In this way, the quantum box shown in FIG. 9A surrounded by the barrier layers 62a and 65a on the upper and lower sides and the barrier layer 67 on the side can be formed. Although the present invention has been described with reference to the embodiments, the present invention is not limited thereto. For example, it will be apparent to those skilled in the art that various modifications, improvements, combinations, and the like can be made.

[0119]

As described above, according to the present invention,
By configuring the optical modulator with a quantum box that performs three-dimensional quantum confinement, it has an α parameter suitable for long-distance transmission,
Moreover, it is possible to obtain an optical modulator with a small insertion loss in the transmissive state. An optical semiconductor device suitable for long-distance transmission can be manufactured by combining a light source that continuously emits light with this optical modulator.

[Brief description of drawings]

FIG. 1 is a partially cutaway perspective view of a modulator integrated distributed feedback laser according to an embodiment of the present invention.

FIG. 2A is a diagram schematically showing a quantum box constituting the optical modulator according to the first embodiment of the present invention, FIG. 2B is an energy level diagram of the quantum box, and FIG. 2C is a quantum box. 2D is a graph showing the light absorption spectrum of the conventional quantum well layer, and FIG. 2D is a graph showing the light absorption spectrum of the conventional quantum well layer.

FIG. 3 is a graph showing an absorption peak on the longest wavelength side of the optical absorption spectrum of the quantum box shown in FIG. 2A.

FIG. 4 is a graph showing changes in light absorption intensity and α of the quantum box shown in FIG. 2A.
It is a graph showing a parameter regarding wavelength.

5 is a graph showing the optical absorption intensity and the α parameter of the quantum box shown in FIG. 2A with respect to the electric field intensity.

FIG. 6 is a cross-sectional view showing a semiconductor laminated structure having a quantum box.

7 is a diagram showing a low pressure MOCVD apparatus for forming the semiconductor laminated structure shown in FIG.

8 is a supply time chart of a reaction gas for forming the semiconductor laminated structure shown in FIG.

FIG. 9A is a diagram schematically showing a quantum box that constitutes an optical modulator according to a second embodiment of the present invention, and FIGS.
Is the energy level diagram of the quantum box.

10 is a graph showing the absorption peak on the longest wavelength side of the optical absorption spectrum of the quantum box shown in FIG. 9A.

FIG. 11 is a graph showing the optical absorption intensity and α parameter of the quantum box shown in FIG. 9A with respect to electric field intensity.

FIG. 12 is a partially cutaway perspective view of the laminated structure for explaining the method of manufacturing the quantum box shown in FIG. 9A.

FIG. 13 is an energy band diagram and a graph showing a wave function for explaining a conventional type I quantum well structure.

FIG. 14 is a graph showing an optical absorption spectrum of a conventional type I quantum well structure.

FIG. 15 is a graph showing the relationship between the α parameter of the optical pulse signal generation source and the maximum transmission distance.

FIG. 16 is a graph showing light absorption intensity change, refractive index change, and α parameter of a conventional type I quantum well structure with respect to wavelength.

FIG. 17 is an energy band diagram for explaining a conventional type II quantum well structure, and a graph showing a wave function.

[Explanation of symbols]

 1 semiconductor substrate 2 n-type region 3 distributed feedback diffraction grating 4a laser active layer 4b modulation absorption layer 6, 7 p-type region 8 semi-insulating buried region 9 polyimide region 10 p-side electrode 11 wiring 12 n-side electrode 14 DFB laser 15 Optical modulator 16 Separation region 31 Substrate 32 Buffer layer 33 Quantum box layer 34 Cap layer 35 Quantum box 40 Reaction container 41 Susceptor 42 High frequency coil 44 Gas exhaust pipe 45, 50, 51, 52, 53, 54 Gas flow channel 55, 56 gas switching valve 57, 58 exhaust gas flow path 61 InP substrate 62, 65, 67 barrier layer 63 hole quantum box layer 64 electron quantum box layer 66 mask pattern

Claims (6)

[Claims] 1. A light emitting means for emitting light, a quantum box arranged in the optical path of the light emitted from the light emitting means to perform three-dimensional quantum confinement for electrons and holes, and an electric field for the quantum box. An optical semiconductor device having an electric field applying means for applying a voltage. 2. The light having the wavelength emitted from the light emitting means is such that the spatial coordinates giving the peaks of the wavefunctions of the electrons and holes confined in the quantum box are three-dimensionally coincident when no electric field is applied. 2. Regarding the above, the optical semiconductor device according to claim 1, wherein the light absorption intensity by the quantum box when an electric field is applied to the quantum box is smaller than the light absorption intensity when no electric field is applied. 3. The wavelength of the light emitted by the light emitting means is set to a wavelength corresponding to the peak on the longest wavelength side among the peaks of the light absorption spectrum by the quantum box, and half of the half width of the peak is set. The optical semiconductor device according to claim 1, wherein the wavelength is shorter than the added wavelength. 4. The wavelength of the light emitted by the light emitting means is within a range of the half value width of any peak of the optical absorption spectrum of the quantum box when an electric field is not applied to the quantum box. The optical semiconductor device according to claim 1 or 2. 5. The spatial coordinates giving the peaks of the wavefunctions of electrons and holes confined in the quantum box when no electric field is applied are shifted with respect to the direction of the electric field applied by the electric field applying means, The light absorption intensity by the quantum box when an electric field is applied to the quantum box is larger than the light absorption intensity when an electric field is not applied, with respect to light having a wavelength emitted from the light emitting means. The optical semiconductor device according to. 6. The wavelength of the light emitted by the light emitting means is a wavelength between the peaks adjacent to each other of the light absorption spectrum of the quantum box when an electric field is not applied to the quantum box. 5. The optical semiconductor device according to item 5.