JPS63305328A - Optical gate matrix switch - Google Patents

Abstract

PURPOSE:To obtain a miniature and inexpensive optical gate matrix which with high productivity by constituting all of a light demultiplexing circuit, a light multiplexing circuit and a semiconductor laser beam gate of a semiconductor heterojunction structure. CONSTITUTION:The light demultiplexing circuit 11 and the light multiplexing circuit 12 consisting of cores 20 constituted of the same conductive component as a semiconductor monocrystal substrate 15 and having a refractive index larger than the substrate 15 and inhibition band width smaller than that of the substrate 15 are arranged on the substrate 15 and embedded in an embedding layer 16 consisting of a clad having the same component and the same conductive characteristics as the substrate 15 to constitute a heterojunction single mode optical waveguide with embedding structure. An optical gate 14 has a cap layer 18 and an optical waveguide core part 20 into which an active layer 21 having a refractive index larger than that of the core part 20 and inhibition band width smaller than that of the core part 20 is inserted and constitutes a heterojunction optical gate by the active layer 21 and p-n junction based upon the upper optical waveguide core part 20 and the upper face of the active layer 21. Since highly accurate alignment between an optical circuit and an optical gate part can be radically settled, the manufacture is easy and massproductivity is extremely high.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an optical gate matrix switch used in the fields of optical communication and optical information processing.

[Conventional technology]

Conventionally, optical gate matrix switches include an optical branch circuit 1 and an optical convergence circuit constructed of optical fibers or optical waveguides, as in the first conventional example shown in FIG. 2, a lens system 3 coupled to the output end of the optical branching circuit 1 and the input end of the optical combining circuit 2 to convert the light into a parallel beam;
An optical gate 4 constituted by a liquid crystal TN cell arranged between these lens systems 3 and for opening and closing optical signals,
Each type is provided in the form of a matrix.

Here, the operation of the gate type matrix switch shown in FIG. 6 will be explained. The configuration shown in FIG. 6 is for a 4×4 matrix. For example, the light incident from the input port ^l is branched into four by the branch circuit 1, and the light ON/OFF shutter G+ I that constitutes the optical gate 4 is split into four branches. lG211G31
It reaches 1041. Here, when the shutter 631 opens, light is output to the output capacitor 83. When the shutter G2+ opens at the same time, light will also be output to the output cap 7 and B2. As described above, a feature of such gate type matrix switches is that 1:N simultaneous multiple-pair connections are possible.

Now, as shown in FIG. 6, a large number of optical branch circuits 1.
Light combining circuit 2. There is a limit to miniaturization in a structure with a three-dimensional circuit configuration consisting of the lens system 3 and the optical gate 4;
There are problems such as the need for precise alignment between each component.

On the other hand, as a second conventional example, an optical gate matrix switch that uses waveguide optical circuits and is made into a planar circuit with the aim of miniaturization has already been proposed (for example, the Institute of Electronics and Communication Engineers, 1988 National Conference 5ll- 2), this is a hybrid integrated optical switch in which an optical gate using a semiconductor laser is installed in a silica-based optical waveguide circuit. Since this optical switch uses a semiconductor laser as the optical gate, gain can be expected at the gate, and therefore there is a possibility that a lossless optical switch can be obtained by compensating for the waveguide loss that occurs in the optical circuit. A major feature is that

(Problem to be solved by the invention) However, in this case, although it is miniaturized, because it is a hybrid type, high-precision alignment work is required when installing the gate, as in the first conventional example. It was necessary.

Generally speaking, the conventional methods require a high-precision alignment process, resulting in extremely low productivity and high costs.

Therefore, an object of the present invention is to provide a configuration that can fundamentally eliminate the high-precision alignment process that is inevitably associated with conventional optical switches composed of individual parts and hybrid integrated optical switches. An object of the present invention is to provide an inexpensive optical gate matrix switch that is small in size and has excellent productivity.

(Means for Solving the Problems) In order to achieve such an object, in the present invention, the three components of the optical gate matrix switch, the optical branching circuit, the optical combining circuit, and the semiconductor laser optical gate, are all integrated. In view of the fact that it can be constructed with a semiconductor heterojunction structure, the gate matrix switch is of a monolithic integrated construction, unlike the conventional example.

That is, the present invention provides a semiconductor single crystal substrate, a core portion having a larger refractive index and a smaller forbidden band width than the single crystal substrate, and a cladding made of the same material as the single crystal substrate. an optical branch circuit and an optical convergence circuit respectively constituted by a heterojunction single mode optical waveguide structure; The optical branching circuit, the optical gate, and the optical converging circuit are coupled in this order.

〔Example〕

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

FIG. 1 shows a perspective view of a 2×2 gate matrix switch as an embodiment of the present invention. Here, the semiconductor single crystal substrate 15 has the same conductivity type as the substrate 15 and a larger refractive index than the substrate 15, and the guided light is directed to the substrate 15.
The optical branching circuit 11 and the optical combining circuit 12 are arranged such that the optical branching circuit 11 and the optical combining circuit 12 are made of a core having a composition having a smaller forbidden band width than the substrate 15 so as to be transparent to the substrate 15. These optical branching circuits 11 and optical combining circuits 12 are embedded with a cladding having the same composition and the same conductivity type as the substrate 15, as shown in FIG. It is embedded within layer 16 and constitutes a buried structure heterojunction single mode optical waveguide. The refractive index and shape of the core portion 20 of such a waveguide are set to satisfy a single mode condition so that multidirectional switching is possible.

Four semiconductor laser light gates 14 are installed at each of the two output ends of the optical branching circuit 11 and the two input ends of the optical merging circuit 12, for a total of four connections. This light gate 14
The cross section taken along the Y2 line along the y-axis is shown in Figure 3 (a).
FIG. 3(b) shows a cross section taken along line 21 along z4ith. The optical gate 14 includes a cap layer 18;
An optical waveguide core section 20 is disposed directly below the optical waveguide core section 20 and has an active layer 21 interposed therein. That is, the core part 2 of the optical waveguide
Active layer 2 with a refractive index greater than 0 and a small forbidden band width
1 is sandwiched between the core portions 20. 22 is a power source for exciting the optical gate 14, that is, injecting a current into the active layer 21; 23 is a switch for turning the power source 22 on and off.

Here, the position of the active layer 21 in the X-axis direction is shown in FIG. 3(b).
As shown in FIG. 3, the optical guide is arranged so that the center of the electric field distribution of the guided light in the waveguide coincides with the center of the electric field distribution of the guided light of the optical gate 14 where the active layer 21 is located, and the coupling efficiency between both waveguides is maximized. It is set to be located at the center of the core portion 20 of the wave path.

The upper part of the active layer 21, that is, FIGS. 3(a) and (b)
The region surrounded by the dotted line is of the second conductivity type in order to inject current into the active layer 21, and the lower part of this region,
That is, a pn junction is formed by the upper optical waveguide core portion 20 and the upper surface of the active layer 21, and a heterojunction optical ON/OFF gate is formed between the active layer 21 and this pn junction.

A cap layer 18 is provided above the active layer 21 to obtain a low resistance ohmic contact. moreover,
An insulator 1]i17 is provided in the upper part of the optical switch except for the cap layer 18 for surface stabilization and to facilitate power supply wiring.

As shown in FIG. 1, an antireflection film 19 is provided at the input and output ends of the optical switch in order to suppress oscillation when current is injected into the optical gate 14.

In such a configuration, consider the case where light with a wavelength matching the forbidden band width of the active layer 21 is incident on the optical gate 14.

The switch 23 is open, and the power supply 12 is connected to the active layer 21.
When no current is injected from the optical gate 4 and the optical gate 4 is in the OFF state, the incident light is absorbed by the active layer 21 and the optical gate 1
On the other hand, the switch 23 is closed, and current is injected from the power supply 22 into the active layer 21, and the active layer 2
1 is in a population inversion state and turns on, the incident light becomes
As shown in FIG. 3(b), the light passes through the optical gate 14 with a gain corresponding to the amount of the injected current.

The above configuration can be realized if a semiconductor heterojunction can be formed and there is a relationship between the refractive index and the forbidden band width such that the smaller the forbidden band width, the larger the refractive index.

The material corresponding to this is InP-I described below.
There are nGaAsP systems and GaAs-GaAlAs systems. In the InP-InGaAsP system, InP may be used as the substrate 15 and the buried layer 16, and InGaAsP may be used as the optical waveguide core 20 and the active layer 21;
A GaAs substrate with AlAs attached is buried in the buried layer 1.
6 and the optical waveguide core 20, and GaAs may be used as the active layer 21.

[Example 1] Application of the structure of the present invention to InP-InGaAsP system 2
An example of a ×2 gate matrix optical switch will be described.

First, a tin (Sn)-doped n-type (100) InP substrate with a carrier concentration of 2xlO"/cc is used as a single-crystal plate 15, and a Sn-doped n-type film of 2x10"/cc is placed on this substrate 15 by the LPE method. I with a film thickness of 0.7 μm at a carrier concentration of cc
no, 97Gao, 03^so, oaF'o,
94 film, non-doped, 0.2 μm thick Ino, t+
Gao, 29AS0.86P0.34 m, and Z
Ino, 97GaO, a3^so, n-doped p-type with a carrier concentration of 2X1017/cc and a film thickness of 0.5 μm,
As shown in FIG. 4(a), the oapo, 94 films were layered and grown in order to form a leading wave circuit core section 20, an active layer 21, and an upper optical waveguide core section 20 so as to be lattice matched.

Next, using the photoresist as a mask, the upper optical waveguide core section 20 and the active layer 21 other than where the optical gate is to be provided are selectively etched with an etching solution, H(fL+HsPO4 solution,
It was removed by sequentially immersing it in 3HNOs+HF+2H20 solution (FIG. 4(b)).

Next, after removing the photoresist, the optical waveguide core portion 20
As, non-doped Ino, 97 (iao, 03
ASO,. 1 lpo, 94 membranes to a thickness of 0.9 μm
It was grown over the entire surface using the PE method (Fig. 4(C)).

Next, using a photoresist waveguide circuit pattern with a width of 4 μm as a mask, 2Cr207+ )IBr + CH3C
After forming a waveguide circuit by immersing it in an O0H solution and etching it until it reaches the substrate 15, non-doped InP, non-doped Ino, 76[iao, 2] was formed using the LPE method.
48so, 5spo, and 4m cap layers 18 were sequentially grown to embed the circuit (FIG. 4(d)).

Next, using photoresist as a mask, 3) INO3+
After removing the cap layer 18 other than the gate part by immersing it in HF+2) 120 liquid, an insulating film 17 is formed on the part other than the cap layer 18 to a thickness of 400 mm by plasma CVD.
5i384 film was deposited (FIG. 4(e)).

Finally, using the insulating film 17 as a mask, the P-type Ino, 97 (iao, 03) deposited during the first LPε growth is
/1sO,06po, 94 After Zn is diffused until it reaches the layer 20, the back surface of the substrate 15 is polished and ^u+ is added to the p side.
A Zn electrode and an Au+Sn electrode were formed on the n side, and the ends were cleaved, and an SiO film was deposited on the cleaved surface to a thickness of 1800 mm to provide anti-reflection 1ii19.

Through the above process, the waveguide width is 2 μm and the waveguide thickness is 2 μm! , 6μ
The thickness of the single mode optical branching circuit 11 and optical combining circuit 12 with m1 relative refractive index difference of 0.6%, and the active layer 21 is 0.2 μm.
An optical gate 14 having an active layer 21 width of 2 μm, a gate length of 300 μm, and a reference wavelength of 1.34 μm was monolithically formed on an InP substrate 15. The total switch length was set to 3 mm in order to suppress excessive branching and merging losses to 3 dB or less.

The optical switch manufactured in this way has a gate current of 80
mA, a sufficient gain was obtained in the optical gate 14, and a loss of O dB was obtained by compensating for branching and merging losses.

(Example 2) In Example 1, the case of a 2×2 gate matrix optical switch was described, but the scale can be expanded as follows. For example, to obtain a 4×4 switch, it is sufficient to construct it as shown in FIG. That is, 2× shown in Example 1
Four 2x2 switches are arranged in parallel, and these 2x2 switches are arranged so that the branch and merging waveguides of the two branches intersect with each other.

As the scale increases, branching and merging losses theoretically increase, and propagation losses also increase due to the length of the waveguide, making it difficult to compensate for these with the gain of a single optical gate. come. In that case, as shown on the right side of FIG.
4 may be added, and the output from the optical combining circuit 12 may be coupled to these optical gates 24. In this example, it is sufficient to simply design the mask for patterning for 4x4, and this can be easily realized without increasing the number of manufacturing steps and n8 degrees of the 2x2 switch shown in Example 1. It is something.

〔Effect of the invention〕

As explained above, the present invention monolithically configures the three elements of an optical gate matrix switch, an optical branching circuit, an optical combining circuit, and a semiconductor laser optical gate, using a heterojunction structure on a common semiconductor substrate. Since the problem of high-precision alignment between the optical circuit and the optical gate section in the conventional example is fundamentally solved, the optical gate matrix switch has the advantage of being easy to manufacture and having extremely high mass productivity.

Moreover, the optical switch of the present invention has excellent performance. That is, since a semiconductor laser optical gate that can obtain a high gain is used, branching and merging losses and propagation losses can be compensated for, thereby realizing a lossless optical switch. Furthermore, according to the present invention, even when expanding the scale of the matrix, expansion can be easily achieved by inserting loss compensation gates at various locations.

In addition to the above-mentioned advantages, the optical switch of the present invention also has advantages such as 1:N simultaneous multi-pair connection and high-speed switching on the order of NSI. It is essential for building an exchange system.

[Brief Description of the Drawings] Fig. 1 is a perspective view showing an embodiment of the optical gate matrix switch of the present invention; Fig. 2 is a sectional view taken along the Y line of the leading wave circuit section in the optical switch shown in Fig. 1; Figures (a) and (b) are cross-sectional views taken along the Y2 line and Zl line, respectively, of the optical gate section in the optical switch shown in the first figure. FIG. 5 is an explanatory diagram of an embodiment in which a 4×4 matrix is constructed using the optical switch of the present invention, and FIG. 6 is an explanatory diagram of a conventional optical gate matrix switch. 1.11... Optical branching circuit, 2.12... Optical combining circuit, 3... Lens system, 4.14.24... Optical gate, 15... Semiconductor single crystal substrate, 16... - Buried cladding layer, 17... Insulating film, 18... Cap layer, 19... Antireflection film, 20... Optical waveguide core portion, 21... Active layer. 20 core pot lowered 11 epinephrine Y1 brocade 1 ku'tJ concave Fig. 2 (G) (b)

Claims (1)

[Scope of Claims] A semiconductor single crystal substrate, and a core portion having a larger refractive index and a smaller forbidden band width than the single crystal substrate, and a cladding made of the same material as the single crystal substrate, on the semiconductor single crystal substrate. an optical branch circuit and an optical convergence circuit each constructed of a buried-structure heterojunction single mode optical waveguide; and an optical branching circuit and an optical convergence circuit each constructed of a buried structure heterojunction single mode optical waveguide; and a heterojunction optical gate having a pn junction constituted by the active layer and the core part, and the optical branching circuit, the optical gate, and the optical combining circuit are coupled in this order. Optical gate matrix switch.