JP2556288B2 - Semiconductor laser - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a semiconductor laser used as a light source for optical communication and optical information processing.

[0002]

2. Description of the Related Art In recent years, in order to extend optical communication to subscribers, semiconductor lasers serving as light sources are required to have lower thresholds and better temperature characteristics than ever before. Furthermore, if an external modulator is not used for cost reduction, the modulation characteristic will also be 1G.
It is desired that it is about b / s. It has been widely practiced to introduce a multiple quantum well structure into an active layer in order to reduce the threshold current of a semiconductor laser, improve quantum efficiency, enable high temperature operation, and enable high speed modulation. . At this time, when the discontinuity of the conduction band between the quantum well layer and the barrier layer is small, electrons are thermally distributed not only in the quantum well layer but also in the barrier layer, so that various characteristics are improved even if the multiple quantum well structure is introduced. Disappear. Therefore, it has been considered desirable that the difference in the forbidden band between the quantum well layer and the barrier layer is large. Recently, the use of so-called strained quantum wells, which use a semiconductor with a lattice constant larger than that of the substrate for the quantum well layer, has been actively studied. This is because the band structure of the valence band is changed by strain and the positive level of the lowest level is changed. This is because the effective mass of the holes can be reduced. In order to obtain the effect of the strained quantum well, it is necessary to increase the discontinuity of the valence band and increase the energy difference between the lowest level of holes and the excited level.

However, if the forbidden band of the quantum well layer and the barrier layer is large and the discontinuity of the valence band is large, holes are strongly confined in the quantum well.
Holes are localized in wells close to the clad layer, and holes are only slightly injected into wells apart from the cladding layer. Since the well in which holes are not injected cannot contribute to the gain, the characteristics of the laser are not improved. That is, the effect of quantum well and the effect of localization of holes are in a trade-off relationship. In the past, attempts have been made to optimize the forbidden band difference between the quantum well layer and the barrier layer. For example, Takaoka and Kushibe have a lecture number 28p-H-3 (1993) of the 54th Academic Meeting of Applied Physics.
, A 1.3 μm band InGaAsP quantum well laser with an InGaAsP barrier layer composition of 1.1 forbidden band wavelength.
It is reported that the characteristic temperature of the quantum efficiency becomes optimal when the thickness is 3 μm. Ogita et al., Lecture number 28p-H-5
At 1.3 μm band strained InGaAsP quantum well laser
The characteristic temperature is maximized when the composition of the nGaAsP barrier layer is 1.1 μm in the forbidden band, and Kito et al. gave a lecture in the lecture number 28p-H-6 of 1.3 μm band strained InGaAsP quantum well laser. It is reported that the relaxation oscillation frequency is optimized when the composition has a forbidden band wavelength of 1.05 μm.

It has also been attempted to reduce the thickness of the barrier layer in order to suppress the localization of holes, thereby increasing the probability that holes will move to an adjacent well by tunneling. For example, Uomi et al., Lecture number 28p-
H-3 (1991), InGaAs / InG
It has been reported that the relaxation oscillation frequency is approximately doubled by changing the barrier layer thickness from 10 nm to 5 nm in the aAsP quantum well laser. In addition, Yamada et al., At the 54th Annual Meeting of the Applied Physics Academic Lecture No. 28p-H-10 (1993),
It has been reported that in a 1.3 μm band InGaAsP quantum well laser, the barrier layer thickness is changed from 3 nm to 10 nm and the thinner the barrier layer, the larger the differential gain.

In another aspect, as an attempt to improve the characteristics of a laser, a II-VI semiconductor is used for the cladding layer to reduce carrier leakage from the active layer to the cladding layer. -86282 publication (Japanese Patent Application No. 57-
195254) by Hiiragimoto.
89 (Japanese Patent Application No. 62-335868). Further, if a II-VI group semiconductor is used for the clad layer, leakage of carriers from the active layer to the clad layer can be reduced, and at the same time, confinement of light in the active layer can be increased and threshold carrier density can be lowered to improve laser characteristics. Can be expected.

[0006]

However, the above-mentioned optimization sacrifices the effect of the quantum well because the difference in the forbidden band between the quantum well layer and the barrier layer becomes small. Even when the barrier layer is thinned, the electrons become three-dimensional, so that the density of states at the band edge becomes small and the gain decreases, and the band structure of holes also changes, so that the effect particularly when the strained quantum well is formed is obtained. As the size of the quantum well becomes smaller, a trade-off relationship appears between the quantum well effect and the hole localization effect. Therefore, in order to further improve the laser characteristics, it is necessary to eliminate the trade-off relationship between the quantum well effect and the hole localization effect.

Further, when a II-VI group semiconductor is used for the cladding layer, it is still difficult to obtain a p-type semiconductor of II-VI group semiconductor having a resistance sufficiently low as that of a III-V group semiconductor, and the element resistance is still present. However, there is a problem in that the temperature rise due to the increase of the temperature rise and the improvement of the characteristics is prevented. Furthermore, in order to improve carrier confinement and optical confinement, I-VI is used for the cladding layer.
When a group I semiconductor or an insulator is used, it becomes difficult to inject current into carriers. The same problem occurs when using a II-VI group semiconductor, a I-VII group semiconductor, or an insulator, which is a material having a large bandgap, for the barrier layer of the quantum well forming the active layer for the purpose of increasing the quantum effect.

Therefore, an object of the present invention is to use a material having a large bandgap, which enables high carrier and optical confinement and low threshold carrier density operation by the quantum effect, for the barrier layer and the cladding layer of the quantum well forming the active layer. An object of the present invention is to provide a semiconductor laser having various characteristics such as threshold current, quantum efficiency, temperature characteristics, and high-speed modulation by having a structure capable of efficiently injecting carrier current.

[0009]

The semiconductor laser of the present invention has the following features.

1) A group III-V semiconductor is used as an active layer, and a clad layer sandwiching the active layer is I- having a dielectric constant smaller than that of the active layer.
Carriers made of a VII group or a II-VI group semiconductor or an insulator and required for laser oscillation are injected parallel to the stacking direction.

2) The active layer for laser oscillation is III
-V group semiconductor is used as a well layer, and group I-VII or II-
It includes a quantum well having a Group VI semiconductor or an insulator as a barrier layer.

3) The active layer for laser oscillation is III
-V group semiconductor is used as a well layer and group I-VII or II-V
It consists of a quantum well with a group I semiconductor or an insulator as a barrier layer, and the energy of the interband transition of the quantum well is larger than the required photon energy, and the energy of the electron intersubband transition is almost equal to the required photon energy. The laser is oscillated by the intersubband transition of electrons by the mechanism of selectively injecting electrons into the higher order sublevels of the electrons in the quantum well.

4) The semiconductor laser of 3) has an electron injection layer made of an n-type semiconductor having a conduction band edge at an energy higher than a higher sub-level related to laser oscillation of electrons in a quantum well constituting an active layer.

5) In the semiconductor laser of 3), a quantum well having a quantum level at an energy higher than a higher sub-level related to laser oscillation of electrons in a quantum well in which an electron injection layer made of an n-type semiconductor constitutes an active layer. have.

6) In the semiconductor laser of 3), carriers are injected into the active layer by a current, and at the same time, light having an energy larger than the energy of intersubband transition of electrons in the quantum well forming the active layer is injected to increase the intensity. Generates electrons in the next subband.

7) In the semiconductor laser of 3), the photon energy is smaller than the energy of the transition between the lowest subbands of electrons and holes in the quantum well constituting the active layer, and the higher subbands of electrons and holes are separated. Electrons are generated in higher subbands by two-photon absorption using light larger than 1/2 of the transition energy between.

[0017]

1 is a sectional view showing the structure of a first embodiment of the present invention. On the InP substrate 11, non-doped Z
nSe _0.54 Te _0.46 1 μm laminated clad layer 12,
A buffer layer 13 of InGaAsP having a thickness of 40 nm, which is lattice-matched to an undoped substrate having a bandgap wavelength of 1.1 μm, an InGaAsP well layer having a bandgap wavelength of 1.35 μm and a lattice constant 0.5% shorter than that of the substrate, and ZnSe. _0.54 Te _0.46 barrier layers 5 nm and 7 nm respectively
The quantum well active layer 14 having a thickness of 0.24 μm, which is formed by alternately stacking 20 layers each, has an undoped bandgap wavelength of 1.1.
A 40 nm-thick buffer layer 15 made of InGaAsP lattice-matched to the substrate of μm, and ZnSe _0.54 Te _0.46 are used as 1
After each layer of the clad layer 16 laminated by μm is grown by the MBE apparatus, the clad layer 16, the buffer layer 15, the quantum well active layer 14 and the buffer layer 13 are formed with a width of 0.5 μm.
The remaining waveguide is removed by RIBE. A p-type InP p-cladding layer 17 having a Be concentration of 5 × 10 ¹⁷ cm ^-3 and a p-type InGaAsP contact layer 18 having a Be concentration of 2 × 10 ¹⁸ cm ^-3 are provided on one side surface of the waveguide, and a Si concentration is provided on the other side surface. 5 x 10 ¹⁷
cm ⁻³ n-type InP n-clad layer 19 and Si concentration 1 × 1
An n-type InP contact layer 20 of 0 ¹⁸ cm ⁻³ is laminated by selective growth, and an electrode is formed on each.

In the laser having such a structure, the refractive index of the waveguiding layer comprising the quantum well active layer 14 and the buffer layers 13 and 15 is about 3, and the refractive index of the cladding layers 12 and 16 of ZnSe _0.54 Te _0.46 is 2. It is 4. At this time, the optical confinement coefficient is about 0.7 as shown in FIG. 2 when the thickness of the waveguide layer is about 0.3 μm, and ordinary InP is used as the cladding layer and about 10 quantum well layers are used as the active layers. The structure has a confinement coefficient of about 0.1, which is 7 times larger. This means that the threshold gain per quantum well layer of the semiconductor laser of this embodiment is 1/14 of that of a normal 10-layer quantum well laser, and the threshold carrier density is greatly reduced. Therefore, the differential gain of the active layer is increased, and the efficiency and speed are improved. Further, since the Auger current is reduced, the temperature characteristic is improved. When the difference in the refractive index between the waveguiding layer and the cladding layer becomes large, higher-order modes tend to stand up.
As shown in (3), only the fundamental mode can exist when the thickness is 0.34 μm or less.

In this structure, carriers are injected from the clad layers on the side surfaces of the quantum well active layer, so that they are not affected by the electrical characteristics of the upper and lower clad layers. Also, despite the high quantum barrier, carriers are uniformly injected into all quantum wells.

By using a II-VI semiconductor having a large bandgap energy for the barrier layer of the quantum well active layer 14, the electron barrier height is 0.9 eV and the hole barrier height is 0. Since it becomes 7 eV and confinement of carriers in the well layer increases, the optical gain in the quantum well increases due to the quantum effect. Furthermore, since the difference in refractive index between the barrier layer and the well layer becomes large, the binding energy of excitons becomes large, and the enhancement by excitons with respect to the gain becomes remarkable, and the differential gain increases.

In the present embodiment, II-VI group semiconductors are used for the cladding layer and the barrier layer of the quantum well, but CuCl which is an I-VII image semiconductor having a larger band gap energy.
_0.27 I _0.73 , CuBr _0.5 I _0.5 and insulator CaS
It is also possible to use rF ₂ or the like. When these are used for the cladding layer, only the fundamental mode can exist. The confinement coefficient at the waveguide layer thickness is 0.7-0.8, which does not change much, but the waveguide layer thickness for obtaining the same confinement coefficient is set. Since it can be made smaller, the threshold current becomes smaller. Moreover, the quantum effect is further enhanced by using it in the barrier layer of the quantum well, and the optical gain in the quantum well is increased. At the same time, since the difference in the dielectric constant between the barrier layer and the well layer is large, the exciton is bound by the image charge. Energy increases and differential gain due to exciton enhancement also increases.

In this embodiment, the quantum well is used for the active layer, but this is not always necessary for increasing the light confinement due to the large difference in the refractive index between the active layer and the cladding layer, and the threshold carrier density is reduced even if the bulk semiconductor is used. Reduction of the threshold current,
Improvements in quantum efficiency, high speed, temperature characteristics, etc. are realized.

Although InP is used for the substrate and a material that is substantially lattice-matched to InP is used, other III-V group semiconductors such as GaAs and GaSb, II-VI group semiconductors, other semiconductors such as Si, sapphire, etc. It is also possible to use the above insulator as a substrate and a semiconductor on which a good crystal can be obtained as a material for the active layer.

FIG. 4 is a sectional view showing the structure of the second embodiment of the present invention. A clad layer 12 having 1 μm of CaSrF ₂ laminated on an InP substrate 11, a buffer layer 13 made of non-doped InP having a thickness of 40 nm, an InGaAs well layer lattice-matched with InP, and a CaSrF ₂ barrier layer are alternately 6 nm and 4 nm, respectively. 20 layers each with a thickness of 0.2
A quantum well active layer 14 of μm, a buffer layer 15 of non-doped InP having a thickness of 40 nm, and CaSrF ₂ of 1 μm
After growing each of the laminated clad layers 16 by the MBE apparatus, the clad layer 16, the buffer layer 15, the quantum well active layer 14, and the buffer layer 13 are removed by RIBE while leaving a waveguide having a width of 0.3 μm. On the side surfaces of the waveguide, n-type ZnSe _0.54 Te _0.46 n-clad layers 19 and 23 and n-type InP contact layers 18 and 20 were laminated by selective growth, and electrodes were formed on the respective layers.

In this embodiment, since the difference in the refractive index between the waveguide layer and the cladding layer is large, the laser confinement coefficient is large and the laser oscillation occurs with a small threshold gain as described in the description of the first embodiment. Further, as shown in FIG. 5, in the quantum well active layer of this embodiment, the energy difference between the first sub-level and the second sub-level of electrons is 0.8 eV corresponding to a wavelength of 1.55 μm.
Is. Further, since the energy edge of the conduction band of ZnSe _0.54 Te _0.46 of the n-clad layer is higher by 0.14 eV than the second sublevel of electrons in the quantum well, the electrons are selectively injected into the second sublevel of the quantum well. To be done. First
Since the energy difference between the sub-level and the second sub-level is large, the first to the second sub-levels of electrons due to phonon emission
The probability of relaxation to the sub-level is small, and the injected electrons are accumulated in the second sub-level, so that the population inversion is realized between the first sub-level and the second sub-level, and optical gain occurs. As described above, in the laser of this embodiment, the gain due to the intersubband transition of electrons is used, and the radiative recombination of electrons and holes is not used as in the ordinary semiconductor laser. However, the conductivity type of the cladding layer may be only n-type. In the current, electrons are injected from the n-clad layer into the second sub-level of the quantum well, and light is emitted from the first sub-level.
It is caused by falling to the sub-level and flowing out from the other n-clad layer. Here, the effective masses of two sub-level electrons in the layer plane are almost equal, so the gain is proportional to the product of the dipole efficiency and the population inversion between sub-levels, and the coupling state density and distribution like ordinary interband transitions. There is no need to integrate the product of the functions. Therefore, the optical gain does not depend on the carrier distribution, that is, the temperature, and a laser having excellent temperature characteristics is realized. In fact, the temperature characteristics are dominated by the leakage of electrons from the second sublevel to the n-clad layer. The transition energy between the lowest sub-levels of electrons and holes in the quantum well active layer is about 1.1 eV, and the laser oscillation light is not lost due to band-to-band absorption.

The n-clad layer in this embodiment may be composed of a quantum well or a superlattice. For example, the n-clad layer of the second embodiment described above may be In _0.52 Al _0.48 on one side only.
As and ZnSe _0.54 Te _0.46 in two atomic layers (about 1.2 nm)
When the superlattices are stacked one by one, the subband edges of the electrons of the superlattice almost coincide with the second sublevel of the quantum well, so that electrons are resonantly and efficiently injected.

In addition, GaAs is used as a substrate and a clad layer 2 in which CaF ₂ is laminated to a thickness of 1 μm, a buffer layer 13 made of non-doped GaAs and having a thickness of 40 nm, a GaAs well layer and a CaF ₂ barrier layer are 4.65 nm and 5 nm, respectively. Approximately 0.2 μm thick quantum well active layers 14 each having 20 layers alternately, a buffer layer 15 made of non-doped GaAs having a thickness of 40 nm, and a clad layer 16 having 1 μm of CaF ₂ laminated.
N-type AlAs n-clad layers 19 and 23 and n-type GaAs contact layers 18, which are laminated on the respective sides of the waveguide,
Even in the combination of 20, the energy difference between the first sublevel and the second sublevel of electrons is 0.8 eV corresponding to a wavelength of 1.55 μm, but the electrons from the n-clad layer are injected into the first sublevel. Therefore, an InGaAs / GaAs strained quantum well laser structure that oscillates 0.98 μm light is integrated so as to be coupled to a laser that oscillates at a transition between sublevels. The light of 0.98 μm excites the electrons from the first sub-level to the second sub-level to realize the population inversion and cause laser oscillation. The temperature characteristic at this time is mainly determined by the temperature characteristic of the InGaAs / GaAs strained quantum well laser for excitation, and a high characteristic temperature of about 150 K can be obtained.

In the laser described above, 0.98 μm is obtained by using non-doped AlAs or an insulator in the cladding layer on the side surface of the waveguide.
Laser oscillation is also possible only with light from an InGaAs / GaAs strained quantum well laser that oscillates m 2 of light.
At this time, the light of the InGaAs / GaAs strained quantum well laser for excitation directly excites the second subband of the electrons of the GaAs / CaF ₂ quantum well by two-photon absorption. GaA
The transition energy between the lowest sublevels of electrons and holes in the s / CaF ₂ quantum well is 1.7 eV, and the light of the pumping laser is 1
It is not absorbed by photons. The temperature characteristic at this time is also determined mainly by the temperature characteristic of the InGaAs / GaAs strained quantum well laser for excitation, and a high characteristic temperature of about 150 K can be obtained.

[0029]

As described above, the effects of the present invention can be summarized as follows. A material having a large bandgap that enables high carrier and optical confinement and low threshold carrier density operation by the quantum effect is used as a barrier of a quantum well forming an active layer. A semiconductor laser having excellent characteristics such as threshold current, quantum efficiency, temperature characteristics, and high-speed modulation can be obtained by having a structure capable of injecting carrier current by using it as a layer or a clad layer.

[Brief description of drawings]

FIG. 1 is a sectional view showing a structure of a first embodiment of the present invention.

FIG. 2 is a diagram showing the relationship between the thickness of the waveguide layer and the confinement coefficient in the laser structure of the present invention.

FIG. 3 is a diagram showing the relationship between the refractive index difference between the waveguide layer and the cladding layer and the thickness of the waveguide layer at which the first-order transverse mode is cut off in the laser structure of the present invention.

FIG. 4 is a sectional view showing the structure of the second embodiment of the present invention.

FIG. 5 is a band diagram of the second embodiment of the present invention.

[Explanation of symbols]

Reference Signs List 11 substrate 12 clad layer 13 buffer layer 14 quantum well active layer 15 buffer layer 16 clad layer 17 p clad layer 18 contact layer 19 n clad layer 20 contact layer 21 electrode 22 electrode 23 n clad layer 24 waveguide 141 well layer 142 barrier layer 40 InGaAs positive electrons of the first sub-level 42 InGaAs / CaSrF ₂ second sublevel 43 InGaAs the valence band edge 44 InGaAs / CaSrF ₂ quantum well of the quantum well of electrons in the conduction band edge 41 InGaAs / CaSrF ₂ quantum well First sub-level of holes 45 ZnSe _0.54 Te _0.46 conduction band edge 46 ZnSe _0.54 Te _0.46 valence band edge

Claims (7)

(57) [Claims] 1. A III-V semiconductor is used as an active layer, and a clad layer sandwiching the active layer is I-VI having a smaller dielectric constant than the active layer.
A semiconductor laser comprising a group I or group II-VI semiconductor or an insulator, and injecting carriers necessary for laser oscillation in parallel to the stacking direction. 2. An active layer for performing laser oscillation is III-V.
Group I semiconductor is used as a well layer, and group I-VII or II-VI is used.
A semiconductor laser comprising a quantum well having a group semiconductor or an insulator as a barrier layer. 3. An active layer for performing laser oscillation is III-V.
A quantum well having a group I semiconductor as a well layer and a group I-VII or II-VI semiconductor or an insulator as a barrier layer, the energy of interband transition of the quantum well is larger than a required photon energy, and an electron subband The energy of the inter-transition is set to be almost equal to the required photon energy, and laser oscillation is generated by the inter-subband transition of electrons by the mechanism of selectively injecting electrons into the higher sublevels of the electrons in the quantum well. And a semiconductor laser. 4. An electron injection layer comprising an n-type semiconductor having a conduction band edge at an energy higher than a high-order sublevel related to laser oscillation of electrons in a quantum well constituting an active layer. The semiconductor laser described. 5. The electron injection layer made of an n-type semiconductor has a quantum well having a quantum level at an energy of a higher sub-level or higher relating to laser oscillation of electrons in a quantum well constituting an active layer. The semiconductor laser according to claim 3. 6. A carrier is injected into the active layer by a current, and at the same time, light having an energy larger than the energy of intersubband transition of electrons in the quantum well forming the active layer is injected to generate electrons in a higher subband. The semiconductor laser according to claim 3, wherein 7. The photon energy is smaller than the transition energy between the lowest subbands of electrons and holes in the quantum well constituting the active layer, and 1/2 of the transition energy between the higher order subbands of electrons and holes. 4. The semiconductor laser according to claim 3, wherein electrons are generated in a higher-order subband by two-photon absorption using larger light.