JPH11145549A - Multiple quantum well structure, and optical semiconductor device and light modulator having the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To transport hole carriers in an MQW(multiple quantum well) structure to be bulk-like, while the quantum effects of the MQW structure are maintained. SOLUTION: In a semiconductor light modulator in which an n-type InGaAsP guide layer 6, a strained MQW layer 8, a graded tensile-strained InGaAsP SCH layer 12, and a p-InP clad layer 14 are successively formed on an n-type InP substrate 4, the compressive strain and tensile strain of a compressive-strained InGaAsP well layer 9 and a tensile-strained InGaAsP barrier layer 10 are adjusted to 1.5% by having their group III compositions change reversely from the same InGaAsP composition. When the heavy hole level of the well layer 9 and the light hole level of the barrier layer 10 are adjusted to the same height in this way, the transfer of hole carriers in the strained MQW layer 8 is set bulk-like.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multiple quantum well structure and an optical semiconductor device having the same.
The present invention relates to an optical semiconductor device and an optical modulator having a strained multiple quantum well structure in which a reverse strain is applied to a barrier layer.

[0002]

2. Description of the Related Art Conventionally, a bulk active layer has a multi-quantum well structure (hereinafter also referred to as an MQW structure.
Abbreviation for tum Well. By changing to the active layer of (1), the characteristics of the semiconductor laser have been remarkably improved. Further, also in the optical modulator, the quantum confined Stark effect (hereinafter referred to as QCSE.
Abbreviation for Quantum Confined Stark Effect. ) Can be obtained.

However, in the multiple quantum well structure, carrier transport between well layers (hereinafter, well layers) is called.
There are unique problems not found in bulk. In particular, holes having a large effective mass do not sufficiently transport carriers between wells. When a multiple quantum well structure is applied to a semiconductor laser,
As the injection current increases after laser oscillation, carrier non-uniformity between wells tends to increase, which causes a phenomenon that the total carrier density of the active layer increases. This leads to a behavior equivalent to the gain saturation phenomenon,
This gives rise to drawbacks such as an inability to extend the modulation band.

In the optical modulator, hole carriers generated by light absorption are not sufficiently extracted.
This causes a problem that the extinction ratio during high-speed modulation cannot be sufficiently obtained.

[0005] In order to solve such a drawback, it has been proposed to introduce tensile strain into a barrier layer (hereinafter referred to as a barrier layer).

In a semiconductor laser, Japanese Patent Application Laid-Open No.
No. 4184, InGaAs on InP substrate
There is an example of a P-based MQW structure. In an InGaAsP-based material, since the conduction band discontinuity is small and the valence band discontinuity is large, overflow of electrons and uneven injection of holes are serious problems. Therefore, if a tensile strain barrier is introduced, the band discontinuity of the conduction band increases and the band discontinuity of the hole decreases, so that the threshold value can be reduced and the modulation band up to 10 GHz can be improved.

Further, from the viewpoint of strain stress compensation, a tensile strain barrier is often used together with a compressive strain well.
1995, Journal of Applied Physics, Vol. 15, H. Oohashi, et al, “1.3 μm In
AsP compressively strained multiple-quantum-wel
l lasers for high-temperaure operation ”, (J. Appl.
Phys., 15, p. 4119 (1995)) describes a strain-compensated MQW structure of a combination of an InAs _0.52 P _0.48 compressive strain well and an InGaAsP barrier corresponding to a band gap of 1.1 μm at −0.6% strain. ing. Similar distortion M
According to QW, 1995 IEICE Preprints, Ohashi et al., "InP-based Compressed Strain MQW Laser for High Temperature Operation"
Excellent high temperature properties are reported.

On the other hand, in the optical modulator, in 1995,
Electronics Letters, Vol. 31, K. Morito, R.
Sahara, K. Sato, Y. Kotaki and H. Soda, “High pow
er modulator integrated DFB laser incorporating st
rain-compensated MQW and graded SCH modulator for
10Gbit / s transmission ”, (Electronics Letters, 31,
p.975 (1995)) and an optical modulator with a compressive strain well and a tensile strain barrier were reported in Kotaki's "Electro-absorption-type optical modulator integrated DFB laser" at the C-308 Electronics Society Conference of the Institute of Electronics, Information and Communication Engineers in 1996. Have been. As described in this report, it is important in the design of a modulator to reduce the valence band discontinuity so as not to deteriorate the modulation characteristics due to hole pile-up. The valence band discontinuity is determined from MQW to p-In
It is necessary to make it smaller toward P. By designing in this way, the hole carrier withdrawal time can be shortened. Hereinafter, such an integrated optical modulator will be described in detail with reference to the drawings.

FIG. 5 is a perspective view of the integrated optical modulator. The integrated optical modulator includes a DFB laser unit 1, an electrode separation unit 2, and a modulator unit 3. DFB (Distribute
The d-feed-back laser is also called a distributed feedback laser, and oscillates a single-mode laser beam by the diffraction grating 5 on the n-type InP substrate 4 shown in FIG. The modulator unit 3
It has a function of turning on and off the DFB laser beam and generating a digital signal. The peripheral structure of the integrated optical modulator includes a silicon dioxide film 2 as a protective film.
0, polyimide 21, p-side electrode 2 for injecting power supply current
2, an n-side electrode 23 as a substrate electrode, and a non-reflective AR coating film 24 applied to the reflection surface.

FIG. 8 is a sectional view showing a distortion MQW of this type of conventional optical modulator and its surroundings. 8, as an example of a semiconductor laser, an n-type InP substrate 4, a guide layer 6 of composition n-InGaAsP,
Compressive strain well layer 9 of composition InGaAsP, composition InGa
AsP tensile strain barrier layer 10, composition InGaAsP
SCH layer 111, graded SCH layer 112 of composition InGaAsP, SCH layer 13 of composition InGaAsP,
And a cladding layer 14 having a composition of p-InP.

The InP / InGaAsP epitaxial layer structure is formed by metal organic vapor phase crystal growth (hereinafter referred to as MO-V
Called PE. ) Method. The source gas of the MO-VPE method is trimethylindium (hereinafter, referred to as TMI),
Trimethylgallium (hereinafter, referred to as TMG), trimethylaluminum (hereinafter, referred to as TMAl), arsine (hereinafter, referred to as AsH ₃ ), phosphine (hereinafter, referred to as TMH).
That PH _3. ) Is used. The organic metal is supplied by bubbling hydrogen. For doping, a gas obtained by diluting disilane (hereinafter, referred to as Si ₂ H ₆ ) and dimethyl zinc (hereinafter, referred to as DMZn) with hydrogen is used as appropriate.

The layer structure has a band gap wavelength of 1.15 on an n-type InP substrate 4 having a (100) plane orientation.
μm (hereinafter referred to as a 1.15 μm composition) nI
The nGaAsP guide layer 6 has a thickness of 50 nm to 100 nm, and a strained MQW layer 8 is provided thereon. The strained MQW layer 8 has seven compressive strained InGaAsP well layers 9 and a tensile strain I between them.
It is composed of an nGaAsP barrier layer 10. On the strained MQW layer 8, InGaAsP SC having a composition of 1.15 μm is formed.
Graded InGaAsP SC in which the H layer 111 has a composition changed from 10 nm to 1.15 μm to 1.0 μm.
H layer 112 is made of 50 nm, 1.0 μm composition InGaAsP
SCH layer 13 is 10 nm, p-InP cladding layer 14
There is. SCH stands for Separate Confinement Heter
ostructure (isolated confinement heterostructure)
It has the function of confining guided light.

The strain amount of the compressively strained InGaAsP well layer 9 is + 0.5%, and the thickness is 9 nm. Tensile strain InG
The strain amount of the aAsP barrier layer 10 is -0.3%, and the thickness is 5.1 nm. The oscillation wavelength of the DFB laser is 1.55
μm, the transition wavelength of this strained MQW layer 8 is:
1.47 μm and 1.5 when reverse biased
5 μm light can be absorbed.

FIG. 9 is a band diagram of this layer structure. When a reverse bias is applied to the optical modulator, the light from the DFB laser is absorbed by the strained MQW layer 8 and becomes InGaAs.
Electron carriers and heavy hole carriers are generated in the P well layer 9. Since the effective mass of electrons is light, they are easily transported to the n-type InP substrate 4 by reverse bias. However, since the holes are heavy, it is necessary to design a band lineup such that carriers are not piled up at band discontinuity points.

First, transport of hole carriers in the strained MQW layer 8 will be described. Hole carriers in each well-compressed strained InGaAsP well layer 9 have a plurality of tensile strained InGaP layers.
Over the barrier of the AsP barrier layer 10, InGaAs
It needs to be transported to the PSCH layer 111. In order for this carrier extraction to occur quickly, the band discontinuity of the valence band between the well and the barrier should be small. If the band gap of the barrier is simply reduced to achieve this, the band discontinuity of the conduction band is also reduced, so that QCSE becomes insufficient and a large extinction ratio cannot be obtained. To solve this, as shown in the lower part of FIG. 9, a tensile strain is introduced into the barrier layer 10 to increase the electron barrier and decrease the hole barrier. When tensile strain is applied, the well layer 9, the heavy hole level and the barrier layer 1 are removed.
0 light hole band edge approaches and the above-mentioned distortion MQ
In the example of the W layer 8, the band edge discontinuity on the hole side is 74 me.
V.

The use of the combination of the compressive strain well / tensile strain barrier also has an advantage in that, in order to compensate for stress, it is difficult to introduce defects in crystal growth.

In the case of the modulator, not only the carrier transport in the MQW but also the carrier extraction from the SCH layer to the p clad cannot be ignored. 1.15 μm composition InGaAsP
SCH layer 111 and 1.0 μm composition InGaAsP
Between the SCH layers 13, graded InGaAsP S
The insertion of the CH layer 112 eliminates valence band discontinuity of about 100 meV. What remains is InGaAsP
The valence band discontinuity between the SCH layer 13 and the p-InP cladding layer 14 is as small as 66 meV and p-doped on the InP side, so that hole carriers can cross this heterogap. .

In the integrated optical modulator using such a modulator structure, the front end face of the modulator section 3 is coated with an anti-reflection coating 24 so that the return light to the DFB laser section 1 does not cause oscillation wavelength fluctuation. Is applied. In order to obtain a sufficient extinction ratio, the modulator length is set to 250 μm. In the integrated optical modulator having such a design, a sufficient extinction ratio of 15 dB can be obtained with a bias of 2 V even at an output of 10 mW.

[0019]

However, the first problem is that the hole modulator generated by the light absorption is not extracted sufficiently quickly in the optical modulator. The point is that the extinction ratio at the time cannot be made sufficiently large. At the time of DC bias, even if the extinction ratio at the time of 2 V bias is 15 dB, the dynamic extinction ratio is degraded to 10 dB.

That is, carrier transport of the MQW structure can be considered. This is because the bulk modulator has a relatively small difference between the DC extinction ratio and the dynamic extinction ratio, whereas M
It can be inferred from the fact that a difference that cannot be ignored occurs in the modulator having the QW structure. In the conventional MQW structure described above, the band discontinuity of holes is reduced by the distortion compensation type, but it is still not as thorough as the graded p-side SCH layer. In addition, since the hole carrier has to overcome many barrier layers, the barrier controls the hole carrier extraction.

A modulator length of 250 μm and an extinction ratio of 15 dB
In this case, there is light absorption of 3 dB at 50 μm on the DFB laser side of the modulator. 10mW power from laser
Then, since a light of 5 mW is absorbed at a length of 50 μm, a photocurrent of 4 mA is generated there. If such a large generated carrier does not come out of the MQW smoothly,
Since the electric field for extracting the carriers is weakened by the space charge effect, the carriers are considered to stay in the well for a considerably long time, so that the extraction of the hole carriers is slow and the erasing ratio is reduced.

A second problem is that the modulation band cannot be sufficiently extended in the laser having the MQW structure. This is because the MQW structure has a problem that a bulk laser requires a carrier-to-well transport time, which is not a problem of a bulk laser, and furthermore, the non-uniformity of the carrier density between wells becomes worse as the current injection increases. This is because a substantial gain saturation phenomenon occurs. The modulation characteristic of the laser when the gain saturation coefficient is large is such that lifting due to relaxation oscillation is suppressed and the 3 dB band is reduced. Therefore, uneven carrier injection leads to deterioration of the laser band.

From this fact, if it is attempted to use high-speed direct modulation exceeding 10 Gb / s in short-range optical communication in the 1.3 μm band where wavelength chirping does not become a problem, such a modulation band cannot be used. Slow growth is a major problem.

An object of the present invention is to completely eliminate the problem of carrier transport between wells that occurs in a multiple quantum well structure.
This is to make the carrier transport of the MQW structure, particularly holes, bulk-like. As a result, in a semiconductor laser, non-uniform injection of hole carriers is prevented from being performed, and in an optical modulator, generated hole carriers are quickly extracted.

As described above, in the semiconductor laser, the strain MQW
Reduction of threshold current of semiconductor laser, improvement of slope efficiency,
It is an object of the device characteristics of the semiconductor laser to improve the high-output and high-temperature characteristics and to improve the modulation band. Another object of the optical modulator is to improve the extinction ratio by shortening the time for extracting generated hole carriers.

[0026]

According to the present invention, there is provided a multiple quantum well structure comprising: (1) a heavy hole level of a well layer of a compressive strain well layer in a multiple quantum well structure comprising a compressive strain well layer and a tensile strain barrier layer; (2) In the multiple quantum well structure, the strain amount of the compressive strain well layer and the strain amount of the tensile strain barrier layer are opposite signs. (3) Also, the multiple quantum well structure of each of the above items (1) and (2) is made of a group III-V compound semiconductor. (4) The multiple quantum well structure is characterized in that the compressive strain well layer and the tensile strain barrier layer have the same Group V composition and differ only in the Group III composition.

Further, the optical semiconductor device of the present invention is characterized by having the multiple quantum well structure of the above (1) to (4).

Further, the optical modulator of the present invention has the above (1) to (4).
(4) It has a multiple quantum well structure or an optical semiconductor device. Furthermore, the optical modulator of the present invention optically transmits the DFB laser unit having a diffraction grating, the modulator unit having the multiple quantum well structure, and the DFB laser unit and the modulator unit. And an electrode separation part.

[Operation] In a group III-V compound semiconductor, when strain is applied, the valence band energy is split into heavy holes and light holes. When the group III composition is changed while the group V composition is constant, the heavy hole level increases, the light hole level decreases, and the conduction band edge decreases under compressive strain. The opposite is true for tensile strain. Therefore, when the group V composition is made equal in the well and the barrier, and the group III composition is changed to respectively compressive strain and tensile strain, electrons and heavy holes have quantum confinement levels in the well layer, but light In the hole, a confinement level is formed on the barrier layer side.

The heavy hole level in the well layer and the light hole level in the barrier layer can be made equal with an appropriate amount of strain and layer thickness. At this time, since the transition between the heavy hole in the well layer and the light hole in the barrier layer can be made without any exchange of phonon energy, the hole carriers can move around without any barrier in the MQW structure, Substantially equivalent to carrier transport in the bulk.

In carrier transport, holes become bulk-like, but are quantum confined in terms of wave function. The heavy hole in the well layer has a quantum confined eigenstate using the heavy hole level of the barrier layer as a barrier. For this reason, the quantum confined Stark effect at the time of light absorption is alive. When an electron-hole pair is generated by receiving light energy, it functions as a quantum well, and once the carriers are generated, the behavior of the hole becomes bulk-like.

For example, InP which is a material for long-wave optical communication
In the InGaAsP system on the substrate, the strain-free InGaAsP
When the well layer and the Ga-rich layer having a strain amount of -1.5% are defined as the well layer and the barrier layer, the energy level of the heavy hole and the light hole is as follows. It splits by about 100 meV, and the heavy hole band edge of the well layer and the light hole band edge of the barrier layer can be made substantially equal. As for the quantum confinement of the heavy hole, the heavy hole band edge of the barrier layer is about 100 meV with respect to the heavy hole band edge of the well layer.
, A sufficient quantum confinement effect can be obtained. On the other hand, the band discontinuity on the electron side is 200 to 300 m.
Because of the eV, a sufficient quantum confinement effect can be obtained even for an electron having a small effective mass.

Now, as an example, it is assumed that the well layer and the barrier layer have V
Although the group composition was fixed, this is not always necessary. However, with a strain amount of 1.5%, the critical film thickness is small, and there is a high risk of introducing crystal defects during crystal growth. Therefore,
By imposing the condition that the composition of the group V is constant, it is possible to easily switch the interface of crystal growth and to prevent the introduction of crystal defects.
In fact, in MO-VPE growth, As re-evaporation from the susceptor holding the substrate becomes a problem, and it is quite difficult to switch the group V between the well and the barrier. If the composition of the well and the barrier is selected from those having the same V-group composition, the V-group switching becomes unnecessary, and such a problem can be eliminated.

In the MQW structure thus obtained,
Since carrier transport has a bulk-like effect and interaction with light has a quantum effect, the characteristics that semiconductor lasers and optical modulators lost when they evolved from bulk to the MQW structure can be recovered while enjoying the excellent characteristics obtained. You can continue to do.

The disadvantages when moving from bulk to MQW are the non-uniformity of the carrier, and this non-uniformity causes a delay in carrier transport time and an effective gain saturation. For this reason, there are rather deteriorated characteristics such as a dynamic extinction ratio in an optical modulator and an extension of a modulation band in a semiconductor laser. The multiple quantum well structure of the present invention fundamentally solves such a problem for the reasons described above.

[0036]

DESCRIPTION OF THE PREFERRED EMBODIMENTS [First Embodiment] Next, an embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a sectional view of a multiple quantum well structure of an optical modulator according to the present invention. FIG. 5 is a perspective view of an integrated optical modulator incorporating the multiple quantum well of the optical modulator.

Referring to FIG. 5, the integrated optical modulator comprises a DFB laser unit 1 having a length of 400 μm, an electrode separation unit 2 having a length of 50 μm, and a modulator unit 3 having a length of 250 μm. These are formed on the n-type InP substrate 4 having the (100) plane orientation. In the DFB laser unit 1, the n-type InP substrate 4
A diffraction grating 5 having a pitch of 243 nm is formed thereon.
There is a function of oscillating light having a wavelength of 55 μm in a single mode.
The peripheral structure of the integrated optical modulator includes a silicon dioxide film 20 as a protective film, a polyimide 21, a p-side electrode 22 for injecting power supply current, and an n-side electrode 2 as a substrate electrode.
3. Non-reflective AR coating film 24 applied to the reflection surface
Is provided.

Next, referring to FIG. 1, the layer structure of the semiconductor crystal of the modulator section 3 is as follows.

First, on the n-type InP substrate 4, a thickness of 2
There are an n-InGaAsP guide layer 6 having a thickness of 0 to 200 nm, a tensile strained InGaAsP SCH layer 7 having a thickness of 5 to 50 nm, and a strained MQW layer 8. The number of the strained MQW layers 8 is three to three.
10 compressive strain InGaAsP well layer 9 and tensile strain I
Stress strain compensation type strain M composed of nGaAsP barrier layer 10
It is a QW layer. On the strained MQW layer 8, a tensile strained InGaAsP having a bandgap increasing in order.
SCH layer 11, graded tensile strain InGaAsP
SCH layer 12, InGaAsP SCH layer 13, p-
The InP cladding layer 14 is laminated, and the total thickness of the SCH layer is 30 to 200 nm.

FIG. 2 is a band diagram corresponding to this layer structure. A first feature that the strained MQW layer of the present invention must have at least is that the heavy hole level in the well layer is equal to the light hole level in the barrier layer. Desirably, the second feature that the heavy hole band edge of the well layer 9 and the light hole band edge of the barrier layer 10 are equal is provided. The first feature and the second feature can be made compatible by appropriately selecting the thickness of the well layer 9 and the thickness of the barrier layer 10.

If the light hole band edge of the well layer 9 and the heavy hole band edge of the barrier layer 10 are made substantially equal to each other, hole carriers at the barrier level are also transported in the strained MQW layer 8. This has the advantage of being easy.

As a method of obtaining such a band lineup of the strained MQW layer 8, the well layer and the barrier layer are changed from the unstrained InGaAsP having the same composition only in the group III composition, and the opposite strains are respectively changed by the same amount. Good to put on. At this time, the MQW has almost the above-described characteristics. An appropriate amount of distortion is 1.2% to 1.8%. The lower limit of the amount of strain is determined by the barrier height at which good quantum confinement can be achieved, and the upper limit of the amount of strain is determined by conditions for preventing crystal defects.

Next, the manufacturing method according to the embodiment of the present invention will be described in detail with reference to FIGS.

First, the plane orientation of the surface is (100).
A groove parallel to the (011) plane has a pitch of 243n on the 400 μm long DFB laser unit 1 on the n-type InP substrate 4 of the plane.
The diffraction gratings 5 arranged in m are formed. Next, a stripe mask of two pairs of silicon dioxide films 15 having a gap of 1.8 μm is formed. The width of the silicon dioxide film 15 is
It is 10 μm in the DFB laser unit 1 and 4 μm in the modulator unit 3, and the end of the modulator unit 3 on the opposite side to the DFB laser unit 30 μm
In the part m, the shape is such that the two pairs of silicon dioxide films 15 are closed. The length of the modulator section 3 is 250 μm, and a transition area where the width of the silicon dioxide film 15 changes continuously between the modulator section 3 and the DFB laser section 1 is 50 μm.
m. This silicon dioxide film 15
Is a mask for preventing growth in MO-VPE, and the transition level of MQW can be changed by changing the mask width in the DFB laser unit 1 and the modulator unit 3.

In this selective growth of MO-VPE, InP
And InGaAsP are epitaxially grown. TMI, TMG, AsH ₃ and PH ₃ are used as source gases, and the organic metal is supplied by bubbling hydrogen. For doping, a gas obtained by diluting Si ₂ H ₆ and DMZn with hydrogen is used as appropriate. The growth pressure is 100 Torr.
r.

After the formation of the silicon dioxide film 15, the above-described FIG.
MO-VPE selective growth is performed with the layer structure of FIG. The transition wavelength of the MQW structure is 1.56 μm in the DFB laser unit 1 and 1.47 μm in the modulator unit 3. In MO-VPE, (1
11) Since the growth rate of the B surface is low, the (111) B surface is generally formed on the selective growth side surface, but the (111) A surface is formed on the side surface of the window portion 16 of the exit portion of the modulator 3. It is formed.

The composition of the strained MQW layer 8 is as follows.
When the V-group compositions of the aAsP well layer 9 and the tensile strained InGaAsP barrier layer 10 are fixed,
At the time of MO-VPE selective growth of the strained MQW layer 8, AsH ₃ ,
PH ₃ is supplied at a constant flow rate, and only the group III flow rate may be switched at the well / barrier interface. By adopting such a growth method, it is not necessary to consider a problem such as re-evaporation of AsH ₃ from the susceptor, and it is possible to grow a strain-compensated MQW having a strong strain. The fact that a good crystal can be obtained when the group V composition is kept constant is described by CNET in 1993 at the 5th International Conference on Indium Phosphide and Related Materials, paper number TuBl (5th Intern).
ational Conference on Indium Phosphide and Related
Materials, paper TuBl (1993)).

Next, referring to FIG. 7, the silicon dioxide film 15 of 1 μm on both sides of the selective growth ridge portion is removed. In the window portion 16, the silicon dioxide film 15 is removed to the same position. Then, the p-InP buried layer 1
7, p-InGaAsP contact layer 18, p-InG
An aAs contact layer 19 is grown. Then, the epitaxial layer in the flat growth portion other than the ridge portion is removed.

Next, the p-InGaAsP of the electrode separation portion 2
Contact layer 18, p-InGaAs contact layer 19
Is removed, and a silicon dioxide film 20 opened only on the selective growth upper surfaces of the DFB laser unit 1 and the modulator unit 3 is formed as shown in FIG.

Next, a polyimide 21 is formed so that the upper surface of the selective growth is exposed, and a p-side electrode 22 is formed on the DFB laser unit 1 and the modulator unit 3. Next, after the back surface is polished to reduce the thickness of the wafer to 120 μm, the n-side electrode 23 is formed.

Finally, the end face is cleaved, and as shown in FIG.
An AR coating film 24 made of a silicon nitride film is applied to the end face on the modulator section 3 side. The end face on the side of the DFB laser section 1 is coated with three layers of silicon dioxide film / amorphous silicon film / silicon dioxide film so as to obtain a facet reflectivity of 75%.

The reflectance of the end face of the modulator 3 can be reduced to 0.04% or less by the structure of the window 16 and the AR coating film 24 formed by the above steps.

In the above embodiment, the InP substrate on the InP substrate
Of course, the present invention can be applied to a GaAsP well / InGaAlAsP barrier and other general III-V compound semiconductors.

[0055]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, embodiments of the present invention will be described in detail with reference to the drawings.

Referring to FIG. 1, according to the embodiment of the present invention, the epi structure of the modulator section 3 is such that an n-InGaAsP guide layer 6 having a thickness of 100 nm and a composition of 1.0 μm is formed on an n-type InP substrate 4. And tensile strain In having a thickness of 15 nm each.
GaAsP SCH layer 7 and tensile strained InGaAsP
There is a strained MQW layer 8 sandwiched between the SCH layers 11, on which a graded tensile strained InGaAsP SCH
Layer 12, InGaAs having a thickness of 20 nm and a composition of 1.0 μm
A PSCH layer 13 and a p-InP cladding layer 14 are stacked.

The strained MQW layer 8 has five layers of compression strained InGaAs.
The strain MQW is a stress-strain-compensated strain MQW including the sP well layer 9 and the tensile strained InGaAsP barrier layer 10. Tensile strained InGaAsP SCH layer 7 and tensile strained InGaAs
sP barrier layer 10 and tensile strained InGaAsP SCH
The composition of the layer 11 is equal, and InGaAs having a composition of 1.4 μm
In the case of P, only the group III composition was changed and 1.5% tensile strain was applied. The compressively-strained InGaAsP well layer 9 is obtained by applying a reverse strain to InGaAsP having the same composition of 1.4 μm to have a compressive strain of 1.5%. (Hereinafter, the notation that only the group III composition is changed when strain is applied is omitted.) The thickness of the compressively strained InGaAsP well layer 9 is 4.5.
nm, and the thickness of the tensile-strained InGaAsP barrier layer 10 is 10 nm. By adopting such a structure, both the heavy hole level in the well layer 9 and the light hole level in the barrier layer 10 are 20 to 2 from the band edge.
At the position of 5 meV, these energy levels can be made substantially equal.

Graded tensile strain InGaAsP
The SCH layer 12 is obtained by combining 15 nm each of a layer obtained by applying 0.9% tensile strain from InGaAsP having a composition of 1.21 μm and a layer obtained by applying 0.5% tensile strain from InGaAsP having a composition of 1.11 μm.

In the optical modulator employing such a band lineup, an extinction ratio of 14 dB at 10 Gb / s operation and 2 V bias, and sufficient characteristics for optical communication were obtained.

[Second Embodiment] Next, a second embodiment of the present invention will be described with reference to the drawings.

Here, an example of application to a 1.3 μm band semiconductor laser will be described. FIG. 3 shows a layer structure around the active layer.

The layer structure is an n-type I having a (100) surface plane orientation.
Tensile strain n-InGaAsP SC is formed on the nP substrate 4.
The H layer 106, the strained MQW layer 8 having five well layers sandwiched between the tensile strained InGaAsP SCH layer 7 and the tensile strained InGaAsP SCH layer 11, and the tensile strain I
It has a structure in which an nGaAsP SCH layer 113 and a p-InP clad layer 14 are stacked.

The tensile strained n-InGaAsP SCH layer 106 and the tensile strained InGaAsPSCH layer 113
Each of the InGs has a thickness of 50 nm and a composition of 1.13 μm.
This is obtained by applying -0.9% strain from aAsP.

The tensile strain InGaAsP S on the inside
CH layer 7 and tensile strained InGaAsP SCH layer 11
Is InGaAsP having a thickness of 10 nm and a composition of 1.3 μm.
And 1.5% tensile strain. Compression strain I
The nGaAsP well layer 9 and the tensile strained InGaAsP barrier layer 10 are made of InGaAsP having a composition of 1.3 μm.
Conversely, the distortions were + 1.5% and -1.5%, respectively.

FIG. 4 is a band diagram of the present structure. By setting the thickness of the well layer 9 to 3.4 nm and the thickness of the barrier layer 10 to 7 nm, both the heavy hole level of the well layer 9 and the light hole level of the barrier layer 10 become about 35 meV from the band edge. And energy levels are almost equal. Therefore, the tensile strained InGaAsP SCH layer 7 and the tensile strained InG
In the strained MQW layer 8 sandwiched between the aAsP SCH layers 11, the heavy hole state and the light hole state transition with each other, so that hole carrier transport becomes barrier-free and bulk-like.

Further, the structure is such that the hole carriers can be easily transported from the SCH layer to the MQW. Tensile strain I
The hole carriers injected into the nGaAsP SCH layer 113 first become write holes having low energy for the holes. The energy level of this light hole is
Light hole level of the compressively strained InGaAsP well layer 9;
Since the level is almost equal to the heavy hole level of the tensile strained InGaAsP barrier layer 10, hole carriers are easily injected into the strained MQW layer 8.

On the other hand, on the electron side, the energy difference between the electron level in the compressively strained InGaAsP well layer 9 and the conduction band edge of the tensile strained InGaAsP barrier layer 10 is about 120 meV, and electrons can be sufficiently confined.

The transition wavelength of the strained MQW layer 8 is 1.31
If the coating has a cavity length of 300 μm and a coating of 30% to 75%, a Fabry-Perot laser with a threshold current of 5 mA and a slope efficiency of 0.6 W / A can be obtained.

In such an MQW structure, an ordinary MQW
Non-uniform injection of hole carriers as in the structure does not occur, and no effective gain saturation occurs. Also,
Transport time between each well is also negligibly short. In this way, since the modulation response is not deteriorated due to the carrier transport of the MQW structure, the block layer is made of Fe-doped In.
If it is composed of a semi-insulating epitaxial layer such as P and has a low capacity, high-speed digital modulation of 10 Gb / s to 40 Gb / s becomes possible. Such an ultra-high-speed directly modulated semiconductor laser is effective for optical communication over a short distance where wavelength chirping can be ignored.

[0070]

According to the present invention, when the above-described semiconductor laser is applied to the above-mentioned optical modulator, the extinction ratio at the time of high-speed modulation is improved from 10 dB to 14 dB. This was achieved by reducing the difference between the DC extinction ratio and the modulation extinction ratio.

That is, in the MQW, while the state of the heavy hole or the light hole is changing, the hole carrier can move without the energy barrier, so that the hole carrier generated by the light absorption is quickly pulled out. . At the same time, the heavy hole level of the well forms a quantum level due to the barrier of the heavy hole level of the barrier, which causes a quantum confined Stark effect, so that the efficiency of light absorption does not deteriorate. Rather, by adopting the structure of the present invention, quantum confinement of electrons becomes sufficiently deep, so that the efficiency of light absorption is improved.

Further, according to the present invention, in a normal MQW, hole carriers having a large effective mass are non-uniformly injected into each well. However, in the MQW structure of the present invention, the hole carriers behave like a bulk. Since such a phenomenon does not occur, there is no gain saturation due to uneven injection,
Further, the carrier transport time between the wells can be neglected. Further, when the present quantum well structure is applied to a semiconductor laser, the modulation response band is extended, and the modulation band can be improved.

[Brief description of the drawings]

FIG. 1 is a sectional view showing an embodiment of a multiple quantum well structure of an optical modulator according to the present invention.

FIG. 2 is a band diagram showing an embodiment of the multiple quantum well structure of the optical modulator according to the present invention.

FIG. 3 is a sectional view showing an embodiment of the multiple quantum well structure of the semiconductor laser of the present invention.

FIG. 4 is a band diagram showing an embodiment of the multiple quantum well structure of the semiconductor laser of the present invention.

FIG. 5 is a perspective view showing an integrated optical modulator according to the present invention and a conventional example.

FIG. 6 is a process chart of a manufacturing process of the integrated optical modulator according to the present invention.

FIG. 7 is a process drawing following FIG. 6 according to the present invention.

FIG. 8 is a cross-sectional view showing a multiple quantum well structure of a conventional optical modulator.

FIG. 9 is a band diagram illustrating a multiple quantum well structure of a conventional optical modulator.

[Explanation of symbols]

REFERENCE SIGNS LIST 1 DFB laser unit 2 electrode separation unit 3 modulator unit 4 n-type InP substrate 5 diffraction grating 6 n-InGaAsP guide layer 7 tensile strained InGaAsP SCH layer 8 strained MQW layer 9 compression strained InGaAsP well layer 10 tensile strained InGaAsP barrier layer 11 Strained InGaAsP SCH layer 12 Graded tensile strained InGaAsP SCH
Layer 13 InGaAsP SCH layer 14 p-InP clad layer 15 silicon dioxide film 16 window part 17 p-InP buried layer 18 p-InGaAsP contact layer 19 p-InGaAs contact layer 20 silicon dioxide film 21 polyimide 22 p-side electrode 23 n-side electrode 24 AR coating film 106 Tensile strain n-InGaAsP SCH layer 111 InGaAsP SCH layer 112 Graded InGaAsP SCH layer 113 Tensile strain InGaAsP SCH layer

Claims (14)

[Claims] 1. A multiple quantum well structure comprising a compressive strain well layer and a tensile strain barrier layer, wherein the heavy hole level of the compressive strain well layer is equal to the light hole level of the tensile strain barrier layer. Multiple quantum well structure. 2. The multiplex quantum well structure according to claim 1, wherein the strain amount of the compressive strain well layer and the strain amount of the tensile strain barrier layer are equal in absolute value with opposite signs. Quantum well structure. 3. The multiple quantum well structure according to claim 1, wherein said multiple quantum well structure is made of a group III-V compound semiconductor. 4. The multiple quantum well structure according to claim 3, wherein said compressive strain well layer and said tensile strain barrier layer have the same Group V composition and differ only in the Group III composition. 5. In the multiple quantum well structure, InGaAsP of the compressive strain well layer and the tensile strain barrier layer I
The composition of As and P in group V of nGaAsP is the same, and III
5. The multiple quantum well structure according to claim 3, wherein only the composition of In and Ga of the group is different. 6. An optical semiconductor device having a multiple quantum well structure comprising a compressive strain well layer and a tensile strain barrier layer, wherein the heavy hole level of the compressive strain well layer is equal to the light hole level of the tensile strain barrier layer. An optical semiconductor device, comprising: 7. The light according to claim 6, wherein in the multiple quantum well structure, the strain amount of the compressive strain well layer and the strain amount of the tensile strain barrier layer are opposite signs and have the same absolute value. Semiconductor device. 8. The optical semiconductor device according to claim 6, wherein said multiple quantum well structure is made of a group III-V compound semiconductor. 9. The optical semiconductor device according to claim 8, wherein in the multiple quantum well structure, the compressive strain well layer and the tensile strain barrier layer have the same Group V composition and differ only in the Group III composition. 10. In the multiple quantum well structure, the composition of V-group As and P in the compressive strain well layer InGaAsP and the tensile strain barrier layer InGaAsP is equal to II.
The optical semiconductor device according to claim 8, wherein only the composition of In and Ga of Group I is different. 11. An optical modulator having the multiple quantum well structure according to claim 1. Description: 12. The optical modulator according to claim 11, further comprising: a DFB laser unit having a diffraction grating; a modulator unit having the multiple quantum well structure; and the DFB laser unit and the modulator unit. An optical modulator, comprising: an electrode separator that optically communicates. 13. An optical modulator comprising the optical semiconductor device according to claim 6. Description: 14. The optical modulator according to claim 13, further comprising: a DFB laser unit having a diffraction grating; a modulator unit having the multiple quantum well structure; and the DFB laser unit and the modulator unit. An optical modulator, comprising: an electrode separator that optically communicates.