JPH0831657B2 - Optical amplifier - Google Patents

Description

TECHNICAL FIELD The present invention relates to a semiconductor laser (LD) type optical amplifier used in fields such as optical communication and optical switching.

(Prior Art) An optical amplifier is an indispensable device for the purpose of long-distance, large-capacity optical communication, large-scale optical switching system, and the like.
As the optical amplifier, one using non-linear scattering in the optical fiber is possible, but the semiconductor laser (LD) type is superior because of its advantages such as small size, high efficiency, and integration with other semiconductor optical devices. LD type optical amplifier has an internal gain of 20-30
The value of dB and the gain between the optical fibers with the optical fiber connected to the input and output ends is about 20 dB. In addition, due to the progress of anti-reflection (AR) coating technology on the end face in recent years, the saturated light output and the gain wavelength band have been greatly expanded, and the device has become close to practical use.

(Problems to be Solved by the Invention) However, the conventional LD optical amplifier has a problem that its characteristics greatly depend on the polarization state of incident light. In a normal use state, the polarization state of incident light is not preserved in the long-distance single-mode fiber, and the polarization state of propagating light largely changes due to external temperature, pressure and the like. Therefore, when the LD optical amplifier is inserted in the middle of the SMF, the output light intensity fluctuates greatly unless some polarization control means is used together.

There are three possible causes for the characteristics of the LD optical amplifier to depend on the incident polarization.

(1) Polarization dependence of gain itself (2) Difference of confinement coefficient to active layer due to polarization (3) Polarization dependence of end facet reflectivity In LD optical amplifier of ordinary double hetero (DH) structure, gain itself is not No polarization dependence occurs. Also, due to the isotropic waveguide structure of the active layer and the reduction of facet reflectivity, (2)
(3) can be solved. However, the first Optoelectronics Conference
ics Conference Post-Deadline Papers Technica
l Digest) B11-2, pages 12-13 (Tokyo, July 1986), according to a paper by Saito et al., on both end faces of a buried hetero (BH) structure LD with an isotropic waveguide structure, Reflectance R =
Traveling-wave LD with an extremely good AR coating of 0.04%
Even in optical amplifiers, a gain difference of more than 10 dB is observed between TE and TM polarizations. In other words, it was difficult to reduce the polarization dependence of the characteristics of the LD optical amplifier only by making the waveguide structure isotropic and reducing the facet reflectivity.

One way to solve this problem is to use a combination of polarization controllers. However, it is difficult to realize a small-sized, low-voltage (current) polarization controller with a semiconductor material, so that monolithic integration is difficult, and a complicated optimum control system must be used.

An object of the present invention is to provide an optical amplifier which can be configured monolithically with a semiconductor material, eliminates the above problems, does not require a complicated control system, and has reduced incident polarization dependence of characteristics.

(Means for Solving the Problem) According to the present invention, a quantum well layer made of a quantum well structure is subjected to in-plane tensile stress due to lattice mismatch, and is light as a sub-band next to the base of the conduction band. A first active layer region having an energy value between the sub-bands of the ground band next to the ground band, which is smaller than an energy value between the sub-bands of the ground band next to the conduction band and the sub band of the heavy hole band; The first active layer region has substantially the same effective band gap as that of the first active layer region and has a second active layer region formed of a bulk semiconductor or a quantum well structure that is not stressed, and an input / output end face for coupling input / output optical signals. It is a semiconductor laser type optical amplifier characterized by the above.

Further, the optical amplifier is characterized in that the first active layer region and the second active layer region are laminated in a layer thickness direction.

Further, the first active layer region and the second active layer region of the optical amplifier may be optically cascaded.

(Operation) The LD type optical amplifier according to the present invention is provided with a region for strongly amplifying the TM mode and a region for strongly amplifying the TM mode inside the active layer to reduce the above-mentioned incident polarization dependency as a whole. is there.

As described above, in the LD optical amplifier, it is difficult to reduce the incident polarization dependence unless the gains for the TE mode and the TM mode are adjusted. In general, a normal DH laser has no polarization dependence of gain. In a quantum well (QW) laser using a semiconductor quantum well as an active layer, heavy hole subbands and light hole subbands are separated, and the gain when the same carrier flows is the transition between electron-heavy hole subbands. Because TE is the main
It is known that the gain for the mode is larger than that for the TM mode (Yamanishi et al., Japanese Journal of Applied Physics, Vol. 23, page L35, (M.Yamanishi et al., Jan.J.Appl. Phy 23 , L3
Five)).

In the present invention, the valence band of a semiconductor subjected to biaxial tensile stress is split, and the fact that the axial hole band is located above the heavy hole band in terms of energy is used to obtain a gain for the TM mode. By mainly using the transition between electron and light hole which is about 4 times larger at the band edge than the gain for TE mode,
The polarization dependence of the gain itself is generated to reduce the polarization dependence of the LD optical amplifier. FIG. 4 (a) shows the energy dependence of the gain of an optical amplifier with a stressed quantum well structure, and FIG. 4 (b) shows the energy dependence of the gain of an optical amplifier with an ordinary quantum well structure. Modes and TM are shown. In particular, in order to prevent the generation of dislocations due to the influence of strain, the thickness of the semiconductor layer which is subjected to strain is made small by using the quantum well structure. In such cases, the subband energies of each ground order of the light and heavy holes are calculated as the level in the potential well formed by the connection of each band edge in the bulk split by strain. Therefore, if the value of biaxial stress is selected so that the magnitude of strain is appropriate, that is, if the degree of lattice mismatch of the quantum well layer is selected, the transition energy between the subbands of electrons and light holes is calculated as follows. -Being smaller than the transition energy between heavy holes, the transition at the time of carrier injection is mainly the transition between subbands of electron-light holes, and the gain of TM mode is increased as shown in Fig. 4 (a). It is possible to Here, according to the quantum size effect, the sub-band energy of the ground hole of the heavy hole is smaller than the sub-band energy of the ground hole of the light hole, which is contrary to the above-mentioned tendency due to the effect of strain. This is not a problem if the thickness of the quantum well layer is increased and the degree of lattice mismatch is increased.

In principle, it is possible to eliminate the gain difference between the TE and TM modes by using the strained superlattice structure as described above and adjusting the magnitude of strain. But actually T
It is necessary to set the magnitude of strain very precisely in response to the great confinement coefficient in the active layer of both E and TM modes, which makes fabrication difficult. However, increasing the gain for the TM mode compared to the TE mode can be easily realized by setting the amount of distortion to some extent or more.

The design of such a potential profile may be performed as follows. First, the band gap of the semiconductor A, which is a quantum well, is E _q ^a , the lattice constant is a _o ^a , and the elastic constants are C ₁₁ ^a and C ₁₂
^a, a ^a (hydrostatic pressure term) the deformation potential, and b ^a (shear stress term), in the semiconductor B a barrier layer surrounding the quantum well, ^{_{respectively, E q b, a o b}} , C 11 b, C Let ₁₂ ^b , a ^b , b ^b .
Consider a case where these layers are stacked on a semiconductor thick film having a lattice constant a _o ^s .

Here, a function of a _o ^a <a _o ^s <a _o ^b is required between the lattice constants. Further, in order to consider the change in band gap due to the effect of strain, the conduction band discontinuity amount between the semiconductor A and the semiconductor B is E _c , the valence band discontinuity amount is E _V, and a ^a = a _c ^a
+ A _v ^a , a ^b = a _c ^b + a _v ^b (a _c ⁱ and a _v ⁱ are hydrostatic pressure terms of deformation potentials for the conduction band and valence band, respectively). In that case, since the potential quantum well structure is generated, here, Need a relationship. Where equation (1) is the conduction band,
Formula (2) is a condition for forming a potential quantum well structure in a light hole band.

Here, the valence band discontinuity for the heavy hole band is And the relationship of E _v ^{hh ′} <E _v ^lh holds.

At E _v ^{hh '} ≤ 0, the sub-band of the ground order due to the heavy holes does not exist or is generated in the barrier layer, so that the electron −
The radiative recombination between the heavy holes is small and has no significant effect on the amplification of light. In this case, the film thickness can be freely set.

For E _v ^{hh '} > 0, it is necessary to design the film thickness so that the sub-band of the ground order due to the light holes is energetically above the sub-band of the ground order due to the heavy holes. Here, if the shift amount from each band edge of each subband of the base of light holes and heavy holes is ΔE _lh and ΔE _hh , the following relations are required between these two values. Becomes

If this condition is satisfied, the sub-band of the ground order due to the light holes is energetically above the sub-band of the ground order due to the heavy holes, so that the required condition is satisfied.
Here, ΔE _lh and ΔE _hh need to be calculated and calculated from the effective mass and the film thickness of each layer.

As a material system that enables the above design, GaAs−In _x Al _{(1-x) y} Ga _{(1-x) (1-y)} As, In _x Ga _1-x As−In _y Al ₁ There are material systems such as _-y As, In _x Ga _1-x As-InAs _y P _1-y , In _x As _1-x Sb-In _y As _1-y Sb.

The present invention utilizes this fact, in the thickness direction of the active layer or in the optical axis direction, a bulk semiconductor that strongly amplifies TE mode or a gain medium having a strain-free superlattice structure, and a strained superlattice for amplifying TM mode. By introducing a gain medium, the polarization dependence of the LD type optical amplifier is reduced as a whole.

Example 1 An example of the present invention will be described below with reference to the drawings. FIG. 1 (a) is a perspective view of a first embodiment of an LD type optical amplifier according to the present invention, and FIG. 1 (b) is a band diagram of its active layer. Here, the quantum size effect is most prominent.
The case of using a GaAs / (In) AlGaAs material will be described.

First, the structure of the optical amplifier shown in FIG. 1A will be described together with its manufacturing method. On the n-GaAs substrate 101, an n-Al _0.4 Ga _0.6 As / n-GaAs multiple quantum well (MQW) layer 102 serving as a buffer layer, n-In _x (Al _0.4 Ga _0.6 ) _(1-x) As ( x is 0
→ Change to 0.1) Clad layer 103, MQW active layer 104, p-In
_0.1 Al _0.45 Ga _0.45 As Intermediate layer 105, p-In _0.1 Al _0.36 Ga _0.54 A
s clad layer 106, p-In _0.1 Al _0.1 Ga _0.8 cap layer 107
It grows continuously by the MBE method. Next, using photolithography and chemical etching, stripes of n-GaAs are formed.
Etching is performed to reach the substrate 101. Next, by the LPE method,
This stripe is formed with p-Al _0.38 Ga _0.62 As layer 108 and n-Al.
_0.38 Ga _0.62 As embedded by the layer 109. At this time, the intermediate layer 105
Due to the presence of the pn junction, the pn junction position of the buried layers 108 and 109 is automatically determined under the active layer 104. This structure is B
Known as the CM structure, the details of this growth method are described in Proceedings of the 59th General Conference of the Institute of Electronics and Communication Engineers, 1984, No. 1016 (1984).

The MQW active layer used here has two regions 104 in the layer thickness direction.
It consists of a and 104b. FIG. 1 (c) is a diagram showing the MQW active layer region. In FIG. 1 (c), the first region 104a is GaAs.
The region well layer 110 and the In _0.2 Ga _0.32 Al _0.48 As barrier layer 111 are alternately laminated for three periods, and the thickness of each layer is
There are 150Å and 50Å. The GaAs quantum well layer 110 is In _0.1 Al
_0.36 Ga _0.54 As Due to the lattice mismatch with the cladding layer 103,
It is subject to tensile strain of 0.7% and is therefore the first
As in the band diagram shown in FIG. 2B, the light hole band is above the heavy hole band by about 50 mV in energy. And in that case
The subband 201 of the ground next to the light holes also comes above the subband 202 of the ground next to the energy due to the heavy holes, and the transition at the time of carrier injection is between the electron-light hole subbands Is the main.

On the other hand, the second region 104b of the MQW active layer is formed by alternately stacking In _0.4 Ga _0.9 As quantum well layers 120 and In _0.1 Al _0.36 Ga _0.54 As barrier layers 121 for three periods, and the thickness of each layer is 100
It is Å, 50Å. The layers forming the second region 104b are p-I
n _0.1 Al _0.45 Ga _0.45 As It is lattice-matched with the intermediate layer 105 and is not strained. The effective band gaps of the two regions 104a and 104b are substantially equal.

Therefore, in the first region 104a of the active layer, the TM mode is
In the region 104b, the TE mode is amplified more strongly.

Here, the MQW active layer is 3 in both the first and second regions.
Although a multiple quantum well structure with a period is used, a single quantum well structure may be used.

Next, a SiO ₂ stripe 112 for current confinement is formed on the p side, and then electrodes 113 and 114 are formed on the n side and the p side, respectively. SiN and AR coating (not shown in FIG. 1) films were respectively formed by plasma CVD on the input and output end faces 115a and 115b formed by cleavage to form a traveling wave LD optical amplifier.

FIG. 2 is a view for explaining the operation of this embodiment, and shows a cross-sectional view of the embodiment shown in FIG. 1 along the optical axis and in a plane perpendicular to the substrate. Figure 2 shows the AR coating film
116a and 116b are shown. In this prototype sample, the development threshold after AR boot was> 100 mA. A spherical fiber 11 for coupling incident light into the active layer 104 and for extracting an optical signal.
7a and 117b are used. When a forward bias is applied between the electrodes 113 and 114, the gain in the active layer is increased and the amplification function is obtained.

In this example, the TE mode and TM mode gains were measured by changing the device length. As a result, when the device length was about 500 μm, the gain difference between the two modes was 1 dB or less, and the incident polarization dependence of the gain was greatly reduced.

(Embodiment 2) FIG. 3 is a sectional view showing an embodiment of the second invention of the present application. Here, a cross-sectional view taken along a plane including the optical axis and perpendicular to the active layer is shown. For simplicity, a planar structure will be described below as an example.

First, the structure of the optical amplifier shown in FIG. 3 will be described together with its manufacturing method. On the n-GaAs substrate 301, n-Al _0.4 Ga _0.6 As / n-GaAs multiple quantum wells (MQW) to be a buffer layer
Layer 302, n-In _x (Al _0.4 Ga _0.6 ) _(1-x) As (x changes from 0 to 0.1) cladding layer 303, MQW active layer 304a, p-In _0.1 Al
MBE of _0.45 Ga _0.45 As intermediate layer 305, p-In _0.1 Al _0.36 Ga _0.54 As clad layer 306, p-In _0.1 Al _0.1 Ga _0.8 As cap layer 307
It grows continuously by the method.

Next, by photolithography, n-In _x (Al _0.4 G
_0.6 ) _(1-x) As clad layer 303 is etched and M is deposited on it.
N-In _x (Al _0.4 G _0.6 ) _(1-x) As clad layer 30 by BE method
3, MQW activation method 304b, p-In _0.1 Al _0.45 G _0.65 As intermediate layer 30
5, p-In _0.1 Al _0.36 Ga _0.54 As Clad layer p-In _0.1 Al _0.1
A Ga _0.8 As cap layer 307 is selectively embedded and grown. MQ at this time
In order for the positions of the W active layers 304a and 304b to match in the layer thickness direction,
The growth layer thickness was controlled.

The MQW active layer used here is 304a, and the GaAs quantum well layer 110 is
In _0.2 Ga _0.32 Al _0.18 As barrier layers 111 are alternately laminated for three cycles, and the film thickness of each layer is 150Å and 50Å, respectively. The GaAs quantum well layer 110 is subjected to tensile strain of about 0.7% due to lattice mismatch with the In _0.1 Al _0.36 Ga _0.54 As clad layer 303. Therefore, as shown in the hand diagram of FIG. 1 (b), the light hole band is above the heavy hole band by about 50 mV in terms of energy. In that case, the subband 201 of the ground order due to the light holes also comes above the subband 202 of the ground order due to the heavy holes in energy, and the transition at the time of carrier injection is electron-light hole Those between sub-bands are the main ones.

On the other hand, 304b of the MQW active layer is composed of In _0.1 Ga _0.9 As quantum well layers and I
n _0.1 Al _0.36 Ga _0.54 As A barrier layer is formed by alternately laminating three cycles, and the film thickness of each layer is 100Å and 50Å, respectively. The layers constituting the active layer 304b have a lattice match and are not strained. The effective band gaps of the active layers 304a and 304b are substantially equal.

Therefore, the TM mode is strongly amplified in the active layer 304a and the TE mode is strongly amplified in 304b. Moreover, the active layer 304
Since TM and TE modes also have gain in a and 304b, there is no large attenuation in that part, and overall, both TE and TM modes are amplified. The same BH structure as in Example 1 is formed, and electrodes 313 and 314 are formed on the p-side and the n-side, respectively, and AR coating films 316a and 315a are formed on the input / output end faces 315a and 315b formed by cleavage.
316b was formed to make a traveling wave type optical amplifier. In this example, when the element length is 500 μm (each of the two regions is 250 μm) and the internal gain is 20 dB, the gain difference between the TE and TM motes is 1 dB.
It was below.

Although the present embodiment has been described using the GaAs / (In) AlGaAs-based material, it is obvious that the present invention can be applied to any material system that can obtain the quantum size effect. The device structure is not limited to the BCM structure shown in the embodiment, and the transverse mode control structure used in a normal LD may be adopted.

(Effect of the Invention) According to the present invention, it is possible to obtain an optical amplifier whose gain has very little dependency on the polarization of incident light.

[Brief description of drawings]

FIG. 1 (a) is a perspective view of an optical amplifier according to an embodiment of the present invention, FIG. 1 (b) is a band diagram of its active layer, and FIG.
It is a figure which shows a MQW active layer, FIG. 2 is a figure for demonstrating operation | movement of this Example, and FIG. 3 is sectional drawing of the optical amplifier by the 2nd Example of this invention. FIG. 4 (a) is a graph showing the energy dependence of the gain of an optical amplifier having a conventional quantum well structure in a strained state and FIG. 4 (b). In the figure, 101,301 substrate 102,302 multiple quantum well (MQW) layer 103,106,303,306 clad layer, 104,304a, 304b ...... MQW active layer 105,305 …… intermediate layer 107,307 …… cap layer 108,109 …… buried layer 110… … GaAs quantum well layer 111 …… In _0.2 Al _0.32 Al _0.48 As barrier layer 112 …… SiO ₂ stripe 113,114,313,314 …… electrode 115a, 115b, 313,314 …… input / output end face 116a, 116b, 313,314 …… AR coating film 117a, 117b …… Spherical fiber 201 …… The subband of the ground order due to light holes 202 …… The subband of the ground order due to the heavy holes.

Claims (3)

[Claims] 1. A quantum well layer having a quantum well structure is subjected to an in-plane tensile stress due to lattice mismatch, and a sub-band of the conduction band and a sub-band of the light hole band A first active layer region having an energy value between subbands smaller than an energy value between the subbands of the conduction band and the heavy hole band is effective. Of a bulk semiconductor or quantum well structure that has substantially the same band gap and is not stressed
2. A semiconductor laser type optical amplifier having an active layer region and an input / output end face for coupling input / output optical signals. 2. The optical amplifier according to claim 1, wherein the first active layer region and the second active layer region are laminated in a layer thickness direction. 3. The optical amplifier according to claim 1, wherein the first active layer region and the second active layer region are optically cascaded.