JPH0943557A - Polarization independent optical element and its production - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a polarization independent optical element having high efficiency and high gain and a process for producing the optical element. SOLUTION: An active layer 1 which consists of a square or trapezoidal shape in sectional shape and current element layers 3, 4 of a triangular prism shape which are arranged on both sides of the active layer 1 and have (111)B as slants are formed on a substrate 10. The element has two gate switches which are connected in series or parallel. The first gate switch consists of the active layer having the length in the direction horizontal to the substrate larger than that of the second gate switch and the higher TE gain. The second gate switch may be composed of the active layer having the length in the direction perpendicular to the substrate larger than that of the first gate switch and the higher TE gain.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a polarization-independent optical element whose operation characteristics do not change depending on the polarization direction of incident light, and a method for manufacturing the same.

[0002]

2. Description of the Related Art The plane of polarization of signal light emitted from an optical fiber changes with time due to changes in pressure and temperature. Therefore, in order to handle the light that has passed through the optical fiber, an optical element whose operation characteristics do not depend on the polarization of the input light is required. For optical devices that require polarization independence,
There are gate type semiconductor optical switches, semiconductor optical amplifiers, semiconductor optical modulators, optical receivers, semiconductor optical waveguides, and the like. In these optical devices, the polarization independent method is common. Therefore, in the following description, the means for making the polarization independent will be described taking the gate type semiconductor optical switch as an example.

The gate type semiconductor optical switch is an optical switch based on a semiconductor optical amplifier. When the current is not injected, the light is blocked (OFF state), and when the current is injected, the amplified light is passed (ON state). In order to make the gate-type semiconductor optical switch polarization-independent, a bulk semiconductor having no polarization-dependent gain is used as the active layer, and the cross section of the active layer is made square to confine TE and TM waves. The means of equalizing the coefficients is used.

FIG. 5 is a sectional view of a gate type semiconductor optical switch. This optical switch has a polarization-independent property by having an active layer made of a bulk semiconductor. In the figure, reference numeral 25 is a bulk semiconductor active layer (1.3
μm composition InGaAsP layer), 26 is a buffer layer (n−
InP), 27 is the upper cladding layer (p-InP), 28
Is a cap layer (p ⁺ -InGaAsP), 29 is a current blocking layer (p-InP / n-InP from the bottom), and 30 is SiN.
x insulating film layer, 31 is a p-electrode, 32 is a substrate (n-In
P) and 33 are n electrodes. The current blocking layer 29 effectively blocks the leakage current and increases the injection efficiency into the active layer 15.

Gain is generated when carriers are injected into the active layer. The signal gain G (dB) is expressed by the following equation as in the traveling wave type semiconductor laser amplifier.

[0006]

## EQU1 ## G = 10.L..GAMMA. (G-.alpha.). Log (e) dB (1) where .GAMMA. Is an optical confinement coefficient and g is a gain coefficient (c
m ⁻¹ ), α is the waveguide loss (cm ⁻¹ ), L is the element length (c
m), and log is the common logarithm.

Unlike the semiconductor quantum well, the gain coefficient g of the bulk semiconductor has no polarization dependence, and the main waveguide loss α is free carrier absorption having no polarization dependence. Therefore, from equation (1), equalize the optical confinement factor Γ with respect to the TE wave and TM wave (make Γ independent of polarization).
Thus, it is clear that the polarization dependence of the signal gain G is eliminated. In order to make the optical confinement factor Γ independent of polarization, a means has been taken to make the cross section of the active layer square so as to make the refractive index distributions felt by TE waves and TM waves equal (when an optical guide layer is provided, , It is necessary to make it square including the light guide layer).

In order to obtain a high gain with a small injection current, the active layer thickness (vertical dimension) needs to be narrower than 1.0 μm. Therefore, in order to make the cross section (horizontal dimension) of the active layer square, the width of the active layer is 1.0 μm.
It is necessary to make it narrower than m.

The dry etching method is often used for forming a mesa stripe of a semiconductor laser because of its excellent processing accuracy. However, when an attempt is made to realize the above-mentioned active layer cross section by using dry etching, a problem that sufficient gain cannot be obtained is encountered. That is, when the width of the active layer becomes narrower than 1 μm, the threshold current of the semiconductor laser increases due to damage introduced into the active layer during dry etching. This is 1 when the dry etching method is used.
It shows that the gain decreases when the active layer width is narrower than μm (in particular, when the active layer width is 0.5 μm or less and the end face is not coated, laser oscillation does not occur.
At this time, the gain is less than 5 dB, which corresponds to the mirror loss). There is also a method of removing the damaged region by wet etching, but it is not practical because the control of the active layer width is difficult and the device characteristics are not constant.

Therefore, as a method for producing an active layer having a width of less than 1 μm and having a sufficient gain, a method of producing by selective growth using a SiN _x film as a selective mask has been proposed.

FIG. 6 is a sectional view showing the structure of a conventional gate type optical switch in which a bulk semiconductor active layer is produced by selective growth. In the figure, reference numeral 34 is an active layer (1.3 μm composition p
-InGaAsP), 35 is a buffer layer (n-In)
P), 36 is a cover layer (p-InP), 37 is an upper cladding layer (p-InP), and 38 is a cap layer (p ⁺ -In).
GaAsP), 39 is a current blocking layer (SI-InP), 4
0 is a SiN _x insulating film layer, 41 is a p electrode (AuZnNi /
Au), 42 is a substrate (n-InP), and 43 is an n-electrode (A).
uGeNi / Au). In this method, a lateral waveguide structure of the active layer is simultaneously formed by crystal growth. Therefore, since dry etching is not necessary, it is possible to manufacture an element in which the gain is not lowered due to damage (damage-free) even with an active layer width of 1 μm or less.

To date, the width is 0.4 μm and the thickness is 0.25 μm.
m active layer was produced by selective growth, and a device with a polarization dependent less than 1 dB and a gain of 20 dB was reported (S. Kita
mura, K .; Komatsu and M.K. Kita
mura, "Polarization-Insen"
sitive Semiconductor Opti
cal Amplifier Array Grow
n by Selective MOVPE, IEEE
Photon Tech. Lett. , Vol. 6,
pp. 1783-175, 1994).

By the way, in the conventional polarization-independent gate type optical switch shown in FIG. 6, when the injection current density is increased to obtain a high gain, the applied voltage becomes p-InP37 and n.
A large amount of injection current starts to flow through the built-in potential of the pn junction formed between -InP 42, so that the current injected into the active layer 34 does not increase and the gain saturates.

In view of this, Japanese Patent Laid-Open No. 6-90065 discloses an element in which the active layer is surrounded by SI-InP 46 as shown in FIG. 7 to prevent such leakage current. In the figure, reference numeral 44 is an active layer (1.3 μm composition p-InG).
aAsP), 45 is a buffer layer (n-InP), 46 is a side cladding layer (SI-InP), and 47 is 1.3 μm.
m composition p-InGaAsP layer, 48 n-InP layer, 4
9 is an upper clad layer (p-InP), 50 is a cap layer (p ⁺ -InGaAsP), 51 is a current blocking layer (SI-).
InP), 52 is a SiN _x insulating film layer, 53 is a p-electrode (A
uZnNi / Au), 54 is a substrate (n-InP), 55
Is an n electrode (AuGeNi / Au). With the introduction of the semi-insulating side cladding layer 46,
Even if high injection is performed, no leak current is generated and high gain is obtained. This element is manufactured by the following procedure. First,
Side-side cladding layer (SI-InP) by selective growth method
And then the buffer layer 45 is formed again by the selective growth method.
And the active layer 44 is grown. At this time, layers 48 and 47 having the same composition as the buffer layer 45 and the active layer 44 are also grown on the side clad layer 46. Then, the upper clad layer (p-InP) 49 and the cap layer (p ⁺ -InGa).
AsP) 50 is grown, unnecessary portions are removed by dry etching, and then current blocking layer 51 is regrown. As a result, this element structure can prevent the generation of leakage current.

[0015]

However, since the p-InGaAsP layer 47 having the 1.3 μm composition has the same composition as that of the active layer 44, it absorbs light leaked from the active layer and increases the internal loss. There is. In order to eliminate this increase in internal loss, it is difficult to manufacture the device even if the layer 47 is removed. This is a 1.3 μm composition p-InGaA
This is because the cap layer 49 is simultaneously etched even if the sP layer 47 is removed by wet etching after the dry etching in the above steps. Normally, the cap layer 49 has a thickness of 1.3 μm to reduce the contact resistance.
The composition has a wavelength longer than m. Since the etching rate is higher for longer wavelength compositions, p-InGaAsP with 1.3 μm composition is used.
47 layers are etched faster than the cap layer 50,
As a result, if the p-InGaAsP layer 47 is to be completely removed, the cap layer (p ⁺ -InGaAsP) 50 is erased or the width thereof is significantly reduced. For this reason, the p-electrode cannot be manufactured, and an element in which the p-InGaAsP layer 47 is removed from the element structure shown in FIG. 7 cannot be realized.

An object of the present invention is to solve the above problems and provide a highly efficient and high gain polarization-independent optical element and a method for producing the optical element.

[0017]

In order to solve the above-mentioned problems, a polarization-independent optical element according to the present invention comprises an active layer having a square or trapezoidal cross section, and both sides of the active layer. In addition, a triangular prism-shaped current element layer having a slope of (111) B is formed on the substrate. Preferably, it has two gate switches connected in series or in parallel, and the first gate switch has a larger horizontal length with respect to the substrate than the second gate switch.
The second gate switch is composed of an active layer having a high gain, and the second gate switch is composed of an active layer having a high TM gain having a length longer than that of the first gate switch in the direction perpendicular to the substrate.

Further, the method of manufacturing a polarization-independent optical element according to the present invention is <01 on a substrate having a (100) plane orientation.
Stacking an insulating film for selective growth having two windows extending in the 1> direction, an active layer having a square or trapezoidal cross section and disposed on both sides of the active layer, and (111) B being a slope. And a step of forming a triangular pole current element layer.

[0019]

BEST MODE FOR CARRYING OUT THE INVENTION A polarization-independent optical element according to the present invention comprises an active layer having a square or trapezoidal cross section,
It is characterized in that a triangular prism-shaped current element layer which is arranged on both sides of the active layer and has a (111) B slope is formed on the substrate. Therefore, the leakage current does not occur due to the action of the current blocking layers on both sides of the active layer, and high gain can be achieved. In addition, since the current blocking layer is a triangular columnar region having a (111) B slope which is significantly slower than the plane orientation (100) of the substrate, the current blocking layer is formed on the current blocking layer during growth of the active layer after the current blocking layer is formed. The growing light absorbing layer becomes negligible.

Further, the polarization-independent optical element according to the present invention preferably has two gate switches connected in series or in parallel, and the first gate switch is on the substrate rather than the second gate switch. On the other hand, the second gate switch is composed of an active layer having a large horizontal length, and the second gate switch is composed of an active layer having a larger vertical length with respect to the substrate than the first gate switch. Therefore, elements having high gains of TE and TM polarized waves are connected in series or in parallel one by one, and the gain of the entire element thus connected is a multiplication or addition of the respective gains (magnification display instead of dB display). . Since the polarization dependence can be freely set by changing the shape of the active layer, in any connection, it is easy to make the gain of TE polarization and the gain of TM polarization of the entire connected elements equal. become.

Furthermore, the method of manufacturing a polarization independent optical element according to the present invention is <0 on a substrate having a (100) plane orientation.
A step of laminating an insulating film for selective growth having two windows extending in the 11> direction, an active layer having a square or trapezoidal cross section and arranged on both sides of the active layer, and (111) B
And a step of forming a triangular column-shaped current element layer having an inclined surface. Therefore, the (111) B plane having a slow growth rate serves as the slope of the current blocking region, and the triangular column current blocking layer is automatically formed.

[0022]

Embodiments of the present invention will be described below in detail with reference to the drawings.

(Embodiment 1) FIG. 1 is a sectional structural view of a gate type optical switch showing a first embodiment according to the present invention.
In the figure, reference numeral 1 is an active layer (1.3 μm composition p-InG).
aAsP), 2 is a buffer layer (n-InP), 3 is a first
N-InP layer forming the current blocking layer, 4 is a p-InP layer forming the first current blocking layer, 5 is an upper cladding layer (p-InP), and 6 is a cap layer (p ⁺ -InGaAs).
P), 7 is the second current blocking layer (SI-InP), 8 is S
iN _x insulating film layer, 9 is a p electrode (AuZnNi / Au),
10 is a substrate (n-InP) having a plane orientation (100), and 11 is an n electrode (AuGeNi / Au).

Next, a method of manufacturing the gate type optical switch having the above structure will be described. FIG. 2A is a cross-sectional structure diagram after the current blocking layer is manufactured. In the figure, reference numeral 12 is an insulating film for selective growth (SiN _x film) having two stripe-shaped windows 101 extending in the <011> direction. First, the p-InP layer 4 is formed using the selective growth insulating film (SiN _x film) 12 as a mask.
And n-InP layer 3 were continuously grown. The length of the base of the current blocking layer having a triangular cross section was 2 μm. The structure composed of the p-InP layer 4 and the n-InP layer 3 is
It acts as a current blocking layer when a forward voltage is applied. Next, after the insulating film for selective growth (SiN _x film) 12 is removed, the buffer layer 2, the active layer 1, the upper clad layer 5, and the cap layer 9 are continuously formed (shown in FIG. 2B). Reference numeral 13 is a 1.3 μm composition p-InGaAsP layer, and 14 is a buffer layer (n-InP). The layers 13 and 14 are layers grown simultaneously when the buffer layer (n-InP) 2 and the active layer 1 are manufactured. Since the slope of the current blocking layer (p-InP layer 4 and n-InP layer 3) is (111) B, the growth rate of this surface is low, and the n-I on this slope is small.
The growth of nP and 1.3 μm composition p-InGaAsP layer is negligible. Then, the current blocking layers (3 and 4) and the active layer 1 are formed by normal photolithography and dry etching.
Form a mesa for embedding so that 3 is included inside, and SI
After re-growing the InP layer 7, p-electrode 9 and n-electrode 1
1 and 1 were formed.

Now, the gate type optical switch shown in FIG. 1 will be described. Unlike the prior art shown in FIG. 7, the light absorption layer having the same composition as the active layer does not exist in the device of FIG. Therefore, the gate type optical switch shown in FIG. 1 does not have the problem of increased internal loss. The element of this example is
A structure in which current injection is blocked by the second current blocking layer (SI-InP) and the first current blocking layer (p-InP layer 4 and n-InP layer 3) and only flows into the active layer 1 is provided. I am taking it. Therefore, even if the injection current density is high, no leakage current is generated, so that the injection density into the active layer 1 can be increased. As a result, the gate type optical switch of FIG. 1 can have a high gain. The active layer 1 has a thickness of 0.3 m and the central portion of the trapezoid has a thickness of 0.45 μm.
(Upper bottom is 0.3 μm, lower bottom is 0.6 μm). That is, the aspect ratio of the active layer cross section is 1: 1.5, and the active layer cross section is almost square. Furthermore, since the active layer is a bulk semiconductor, the signal gain of the TM mode and the signal gain of the TE mode are made equal.

As described above, the gate type optical switch of this embodiment is characterized by high gain, high efficiency and polarization independence.

When signal light was actually incident on the gate type optical switch, the polarization gain difference between the TE mode and the TM mode was 0.5 dB or less, and a gain of 5 dB was obtained at a driving current of 10 mA.

As the second current blocking structure, SI-InP is used.
Layer 7 was used, but p-InP / n-InP / p-InP
If a layer is used to apply reverse bias, current confinement can be similarly performed. Further, instead of the SI-InP layer 7, p-
Even if InP is used, a homojunction is formed with n-InP of the substrate, and current confinement can be performed although it is only in the low voltage region.

(Embodiment 2) FIG. 3 is a sectional structural view of a gate type optical switch showing a first embodiment according to the present invention.
In the drawing, reference numeral 15 is an active layer (1.3 μm composition p-In
GaAsP), 16 is a buffer layer (n-InP), 17
Is a first current blocking layer (SI-InP), 18 is an upper cladding layer (p-InP), 19 is a cap layer (p ⁺ -In).
GaAsP), 20 is the second current blocking layer (SI-In)
P), 21 is a SiN _x insulating film layer, 22 is a p-electrode (AuZ
nNi / Au), 23 is a substrate (n-InP) having a plane orientation (100), and 24 is an n electrode (AuGeNi / Au).
It is.

In this device, the first current blocking layer is SI-In.
Only the point P is different from the element of Example 1. The same operating characteristics as those of Example 1 were obtained.

(Embodiment 3) When the width of the active layer is wide, the optical confinement coefficient in the lateral direction is large and the TE polarization becomes large. On the other hand, when the width of the active layer is narrowed, the optical confinement coefficient in the lateral direction is reduced, and a gate switch with a high TM gain can be realized.

FIG. 4 is a schematic diagram of a switch in which a gate type optical switch having a high TE polarization and a gate type optical switch having a high TM polarization are cascade-connected, showing a third embodiment of the present invention. In the figure, reference numeral 1a is an active layer having a high TE polarized wave gain, and 1b is an active layer having a high TM polarized wave gain. The active layer 1a having a high TE polarized wave gain has a higher optical confinement coefficient in the horizontal direction than that in the vertical direction, and the active layer 1b having a high TM polarized wave gain.
Has a higher optical confinement coefficient in the vertical direction than in the horizontal direction. The two active layers are connected in a dependent manner, and the lengths of the active layers 1a and 1b are adjusted so that the gains match. In addition, polarization independence can be realized by individually adjusting the currents of the active layers 1a and 1b.

When signal light is incident on the gate type optical switch having this structure, the polarization gain difference between TE mode and TM mode becomes 0.5 dB or less, and a gain of 5 dB can be obtained with a drive current of 20 mA. , Low polarization, low drive current, high efficiency, high gain optical gate switch was realized.

Although the current blocking layers 3 and 4 are used here, it goes without saying that the current blocking can be performed by using the SI-InP layer instead. It goes without saying that polarization independence can also be achieved by connecting in parallel a gate type optical switch having a high TE gain and a gate type optical switch having a high TM gain. In addition, monolithic integration with other passive and active elements is possible.

Although the gate type semiconductor optical switch has been described as an example in the embodiments, similar effects can be obtained in other optical devices such as a semiconductor optical amplifier, a semiconductor optical modulator, an optical receiver and a semiconductor optical waveguide. Is obtained.

[0036]

As described above, the polarization-independent optical element and the method of manufacturing the same according to the present invention can achieve the following requirements that cannot be satisfied at the same time by the conventional technology.

(1) Since the active layer is surrounded by the current blocking layer, leakage current does not occur even at high injection.

(2) Since the crystal growth rate of the (111) B plane is much smaller than that of the (100) plane, a layer having the same composition as the active layer is hardly formed. Therefore, the internal loss is small.

Therefore, by satisfying the above (1) and (2) at the same time, it is possible to realize a polarization-independent optical element with high efficiency and high gain.

In addition, by adjusting the active layer structure, a gate switch having a high TE gain and a gate switch having a high TM gain can be manufactured, and a gate switch having a high TE gain and a gate switch having a high TM gain are connected in series or. There is also an effect that polarization independence can be realized by parallel connection.

[Brief description of drawings]

FIG. 1 is a sectional view showing a schematic structure of a gate type optical switch which is an embodiment of a polarization independent optical element according to the present invention.

FIG. 2 is a sectional view showing a schematic structure of a gate type optical switch which is an example of a polarization independent optical element according to the present invention,
(A) And (b) is a figure corresponding to each process at the time of forming an active layer.

FIG. 3 is a sectional view showing a schematic structure of a gate type optical switch which is an example of a polarization independent optical element according to the present invention.

FIG. 4 shows a polarization-independent optical device according to the present invention,
FIG. 3 is a perspective view of an optical switch in which a gate type optical switch with high TE polarization and a gate type optical switch with high TM polarization are connected in cascade.

FIG. 5 is a sectional view for explaining a schematic structure of a conventional optical amplifier (gate type optical switch) using a bulk semiconductor as an active layer.

FIG. 6 is a cross-sectional view for explaining a schematic structure of a conventional optical amplifier (gate type optical switch) in which a bulk semiconductor active layer is produced by selective growth.

FIG. 7 is a cross-sectional view for explaining a schematic structure of a conventional optical amplifier (gate type optical switch) in which a current element layer is produced by selective growth and a bulk semiconductor active layer is produced by selective growth.

[Explanation of symbols]

1 active layer (1.3 μm composition p-InGaAsP) 1a TE high polarization gain region 1b TM high polarization gain region 2 buffer layer (n-InP) 3 current blocking layer (n-InP) 4 current blocking layer ( p-InP) 5 upper cladding layer (p-InP) 6 cap layer (p ⁺ -InGaAsP) 7 current blocking layer (SI-InP) 8 SiN _x insulating film layer 9 p electrode (AuZnNi / Au) 10 substrate (n-) InP) 11 n electrode (AuGeNi / Au) 12 Insulating film for selective growth (SiN _x film) 13 1.3 μm composition p-InGaAsP layer 14 Buffer layer (n-InP) 15 Active layer (1.3 μm composition p-InGaAsP) 16 buffer layer (n-InP) 17-side cladding layer (SI-InP) 18 upper cladding layer (p-InP) 19 cap layer (p ⁺ -ing AsP) 20 current blocking layer (SI-InP) 21 SiN _x insulating layer 22 p electrode (AuZnNi / Au) 23 substrate (n-InP) 24 n electrode (AuGeNi / Au) 25 active layer (1.3 .mu.m composition p- InGaAsP) 26 buffer layer (n-InP) 27 upper cladding layer (p-InP) 28 cap layer (p ⁺ -InGaAsP) 29 current blocking layer (p-InP / n-InP from the bottom) 30 SiN _x insulating film layer 31 p electrode (AuZnNi / Au) 32 substrate (n-InP) 33 n electrode (AuGeNi / Au) 34 active layer (1.3 μm composition p-InGaAsP) 35 buffer layer (n-InP) 36 cover layer (p-InP) 37 upper cladding layer (p-InP) 38 cap layer (p ⁺ -InGaAsP) 39 current blocking layer (SI-InP) 0 SiN _x insulating layer 41 p electrode (AuZnNi / Au) 42 substrate (n-InP) 43 n electrode (AuGeNi / Au) 44 active layer (1.3 .mu.m composition p-InGaAsP) 45 buffer layer (n-InP) 46 Side clad layer (SI-InP) 47 1.3 μm composition p-InGaAsP layer 48 n-InP layer 49 Upper clad layer (p-InP) 50 Cap layer (p ⁺ -InGaAsP) 51 Current blocking layer (SI-InP) 52 SiN _x insulating film layer 53 p electrode (AuZnNi / Au) 54 substrate (n-InP) 55 n electrode (AuGeNi / Au)

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Katsuaki Song 1-6, Uchisaiwaicho, Chiyoda-ku, Tokyo Nihon Telegraph and Telephone Corporation

Claims (3)

[Claims] 1. An active layer having a square or trapezoidal cross section, and (111) B disposed on both sides of the active layer.
A polarization-independent optical element, comprising: a current element layer having a triangular prism shape with an inclined surface formed on a substrate. 2. An active device having two gate switches connected in series or in parallel, wherein the first gate switch has a larger horizontal length with respect to the substrate and a higher TE gain than the second gate switch. 2. The layer according to claim 1, wherein the second gate switch comprises an active layer having a higher TM gain and a length longer in a direction perpendicular to the substrate than the first gate switch. Polarization independent optical device. 3. <01 on a substrate having a (100) plane orientation.
Stacking an insulating film for selective growth having two windows extending in the 1> direction, an active layer having a square or trapezoidal cross section and disposed on both sides of the active layer, and (111) B being a slope. And a step of forming a triangular pole current element layer.