JPS60173519A - Semiconductor optical switch - Google Patents

Abstract

PURPOSE:To employ a large crossing angle of an optical waveguide, and to obtain a large extinction ratio by providing a power source for injecting a current to a switch part, and an optical incident means for making waveguide light incident at an angle THETA to the switch part through a waveguide layer. CONSTITUTION:An InGaAsP layer 22 of non-dope is brought to epitaxial growth on an n-InP substrate 21, and InGaAsP waveguides 11, 12 of a rib structure are formed by executing an etching in a shape of a pattern to its opitaxial layer 22. Subsequently, an SiO2 film 31 is formed on the whole surface on the epitaxial layer 22 by spattering, and thereafter, a crystal surface of the InGaAsP layer is exposed by making a window 32 on one half part 18 of a crossing part of the optical waveguides 11, 12. On the InGaAsP layer exposed by a selected liquid phase epitaxial growth, a P-InP clad layer 23 and a P-InGaAsP electrode layer 24 are grown successively and continuously and a double hetero-structure is realized. Thereafter, electrodes 25, 26 are formed and alloyed.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an optical switch constructed from a semiconductor optical waveguide.

<Prior Art> An optical switch that has the function of electrically selecting between two optical paths in an optical integrated circuit is an essential circuit element in constructing an optical matrix circuit. A total internal reflection optical switch is known as one such optical switch. Consider the case where two optical waveguides 11 and 12 intersect at an angle of 2θ as shown in FIG. Port 1 at one end of the leading waveguide 11
Light entering from port 3 travels to port 14 at the other end of optical waveguide 11, and light entering from port 15 at one end of optical waveguide 12 travels to port 16 at the other end of optical waveguide 12. In the area where these two optical waveguides 11 and 12 intersect, the refractive index of the area 18 on one side of the bisector 17 of the intersection angle 2θ of both optical waveguides is locally determined as indicated by diagonal lines in FIG. When the relationship θ〈90°−aif' (1+△n/n) (where n is the refractive index of the optical waveguide) is satisfied, the refractive index change amount is △n. The light incident from 13 satisfies the total internal reflection condition at the boundary of the refractive index changing region 18, and the light does not proceed in the direction of port 14 but proceeds to port 16.

In other words, this light is subjected to a light switch action.

A total internal reflection optical switch is constructed based on the principle described above. Usually, in order to produce such a local refractive index change, the electro-optic effect of applying a high electric field to a ferroelectric material such as LiNb0a is used, or the electro-optic effect of applying a high electric field to a ferroelectric material such as LiNb0a, or GaAS
The electro-optic effect, which applies a reverse bias voltage to a semiconductor such as P, is mainly employed.

However, the amount of change in refractive index caused by the electro-optic effect is usually extremely small, on the order of 10<-4 >. For this reason, the intersection angle 2θ that satisfies the total reflection condition is about 1° or less, and there is a problem that the extinction ratio between the switch-on state and the switch-on state becomes extremely poor. Furthermore, since the intersection angle 2θ is small, it is difficult to construct a matrix circuit by arranging a large number of these switching elements. Furthermore, in the case of the electro-optic effect, since the refractive index changes greatly depending on the polarization direction of incident light, it is necessary to input light having a specific polarization plane.

<Purpose of the Invention> In order to solve these problems, the present invention creates a change in refractive index that is about two orders of magnitude larger than the conventional one than the electro-optic effect caused by the plasma effect inherent in semiconductors, and increases the crossing angle of optical waveguides. The purpose of the present invention is to provide a waveguide type optical switch that can be used to obtain a large extinction ratio.

(3) <Embodiment> FIGS. 2 and 3 show an embodiment of the present invention.
An epitaxial layer 22 having the same conductivity type and smaller than the forbidden band (band gap) of the substrate 21 is formed on the nP substrate 21, that is, a layer 22 that is lattice matched to the substrate 21.

This layer 22 is, for example, an epitaxial layer of non-doped nGaA or P layer.

This epitaxial layer 22 is etched in a pattern to form lip-structured optical waveguides 11 and 12 that intersect at an angle of 20 degrees. For example, the thickness d of the epitaxial layer 22 is approximately 1 μm, the valley width W of the optical waveguide 11.12 is 8 μm, and the height h thereof is approximately 5000 μm. In this way, an optical waveguide 11.12 made of nG and A5P quaternary elements with a forbidden band wavelength of 1.2 μm is constructed using InP as a substrate, and IoGaA
Light having a wavelength (>13 μm) that is transparent to the sP layer 22 can be propagated along the leading wavepaths 11 and 12.

One half is (indicated by diagonal lines in FIG. 2) of the intersecting region of the optical waveguides 11 and 12 has a layer structure as shown in FIG. 4A. That is, on one half of the intersection of the optical waveguides 11 and 12, there is a cladding layer of an epitaxial layer having a conductivity opposite to that of the optical waveguides 11 and 12 and having a forbidden band larger than that of the optical waveguide, in this example, a P engraving. A P cladding layer 23 is formed, and a gold layer electrode, such as p-type ■nG, AsP
An electrode layer 24 is formed, and the P 10-type InGaA3P electrode layer 2
A P-type electrode 25 is formed on 4. n-InP substrate 2
An n-electrode 26 is formed on the opposite side of the first epitaxial layer 22. A power source 27 is connected between the electrodes 25 and 26, which is capable of passing a forward current. In this way, non-doped In with a long band gear rough wavelength (small forbidden band)
The GaAsP layer 22 is n-type with a large forbidden band.
A so-called double heterostructure switch section 28 is formed between the P layer 21 and the P type InP layer 23.

Because of this double heterostructure, carriers injected by flowing a forward current through the p-n junction of the switch section 28 are
5P waveguide layer 11.12 is confined in portion 29. As a result, the refractive index of the InGaAsP layer portion 29 of the double heterostructure shown in FIG. 4 decreases due to the plasma effect. This plasma effect is explained, for example, in T.'I'amir [■ integrated Qptics J244 page. It is given by Eq.

△n=-e"λ2△N/8π2nεm*C2 where e is the charge amount of the carrier, λ is the wavelength of the incident light, and n is the optical waveguide 1
The refractive index is 1.12, ε0 is the permittivity of vacuum, m is the effective mass of the carrier, and C is the speed of light dust. For incident light with wavelength λ=15 μm, ΔN=1 xi 018. -8 and △n=
5.6xlO is obtained.

This refractive index change amount Δn is one order of magnitude larger than that of the conventional example 10-4. Therefore, the optical waveguide 11.1
If the intersection angle 2θ of 2 is formed, total reflection will occur and an optical switch can be realized. In conventional optical switches, the crossing angle 2θ was at most about 1° or less as mentioned above, but in this invention, by increasing △n, the crossing angle 2θ can be increased to 8° at maximum (・θ=4°). It is possible to do this to a certain extent.

Even if a local double heterostructure as shown in FIG. 4 is formed, the injected carriers actually diffuse laterally and the refractive index may not change sharply. Critical angle θ when the refractive index changes exponentially with diffusion distance L=3 μm. FIG. 5 shows the result of calculating the relationship between (maximum angle at which total reflection can be obtained) and the amount of change in refractive index Δn. When the refractive index gradually changes numerically between indices (L=3μ
m), the critical angle of incidence θC is slightly smaller than when the refractive index change is steep (L−0 μm). However, steepness is not a necessary condition. In order to obtain a steep change, the area outside the double heterostructure region shown in FIG. 4 may be irradiated with proton ions to make it semi-insulating.

The semiconductor optical switch of the present invention can be manufactured, for example, as shown in FIG. That is, as shown in FIG. 6A, n
-l. A non-doped InGaA double layer 22 is epitaxially grown on a P substrate 21, and the epitaxial layer 22 is etched in a pattern (7) to form a rib-structured 1° GaA P waveguide 11° (12). Form. Next, as shown in FIG. 6B, 810
After a □ film 31 is formed on the entire surface of the epitaxial layer 22 by sputtering, a window 32 is opened in the single part 18 at the intersection of the optical waveguides 11 and 12 to expose the entire crystal plane of the InGaAsP layer. As shown in FIG. 6C, P-I,
A P cladding layer 23 and a P-InGaA8P electrode layer 24 are successively grown to realize a double heterostructure. Thereafter, electrodes 25, 26 are formed and alloyed as shown in FIG.

Or as shown in Figure 7 AvC, the n-InP substrate 21
A groove 33 having a depth of several 1000' is formed in the shape of an optical waveguide pattern by etching, and then, as shown in FIG. 7B, non-doping (2) is applied. GaA8P epitaxial layer 2
2, P-I. P cladding@23 and P±■oGaA5P electrode layer 24 are epitaxially grown continuously. Groove 33
Non-dope inside■. Rib-shaped optical waveguides 11 and 12 are constituted by the G and A5P layers. Next, by selective etching, as shown in FIG.
The P cladding layer 23 and the oGaAsP electrode layer 24 are removed. In this way a limited double heterostructure can be realized.

When applying the refractive index change using the plasma effect caused by injected carriers to the optical switching phenomenon as described above, there is a concern that light absorption will increase due to an increase in the number of carriers. However, the paper by E. Pinkas et al. IEEE.

J, Quantum EI electronics QE-
9, p. 281 (1973), the number of carriers ~ 2
It has been reported that no significant increase in absorption loss is observed below Xl 018.1lI8. In this invention, a sufficient refractive index change Δn can be obtained with the number of injected carriers ΔN2X10'8α-8, so there is no problem of absorption loss.

Furthermore, in the case of this invention, the light is transmitted to the region 2 where carriers have accumulated.
Since total reflection occurs at the boundary of 9 (Figure 4), even if 2x10
1. There is no increase in absorption even when operating with an imaginary number of carriers.

Here ■. Although the optical switch is constructed using a GaAsP r IMP-based semiconductor, an optical switch having a similar structure can also be constructed using a GaAs-based semiconductor or the like.

Furthermore, in order to configure an optical switch for switching a 2×2 optical path, two refractive index changing regions are required not only in one half of the intersecting region of the optical waveguides 11 and 12 but also in the half of the pond, as shown in FIG. , that is, forming the switch part 34 with a double heterostructure,
It is sufficient if current can be independently injected into these switch sections 28 and 34. For example, protons may be injected into the gap between the switch parts 28 and 34 to prevent carrier diffusion.

In addition, in the above description, the optical waveguides 11 and 12 having a lip structure are configured, and the extension direction of the optical waveguides is set at the switch portions 28 and 34.
The light input means is configured to intersect the light at an angle θ and enter the light at an angle θ into the switch portion 28, 34. However, the optical waveguides 11 and 12 do not have a lip structure, that is, the epitaxial layer 22 is a planar optical waveguide, and light is incident on this optical waveguide using a diffraction grating coupler, and the light is transmitted to the switch section 28.34. The light incident means may be configured so that the light enters at an angle θ.

<Effects> As explained above, since this invention uses the plasma effect unique to semiconductors, it is possible to obtain a large refractive index change, and as a result, the intersection angle is large and the extinction ratio (crosstalk is It becomes possible to construct an optical switch with a high cost (low). The response speed of an optical switch is determined by the carrier recombination time, and is expected to be several hundred Mn2. Furthermore, it has an extremely small dependence on the plane of polarization of light.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the basic configuration of a total internal reflection optical switch, FIG. 2 is a plan view showing an embodiment of the present invention, and FIG.
4 is a sectional view taken along line BB in FIG. The figure is a sectional view showing an example of the process for manufacturing the semiconductor optical switch of the invention, and FIG. 8 is a perspective view showing an embodiment of the pond of the invention. 11.12, 22: Optical waveguide, 26: N electrode, 3: Refractive index decreasing region, 28, 34: Switch section, (11) 21: InP substrate, 23: I, P clarad layer, 24
: IoGaA, P electrode layer, 25: P electrode, 26: n electrode, 28, 34: switch section, 29: refractive index decreased region. Patent applicant: Representative of Nippon Telegraph and Telephone Public Corporation Takashi Kusano (12) +171 Figure left 20';

Claims (5)

[Claims] (1) A semiconductor substrate of one conductivity type with an electrode formed on the − surface, and an epitaxial substrate formed on the surface of the semiconductor substrate, which is of the same conductivity type as the substrate, and whose forbidden band is smaller than that of the substrate. a cladding layer formed on a limited part of the waveguide layer and consisting of an epitaxial layer having a conductivity type opposite to that of the substrate and having a forbidden band larger than that of the waveguide layer; A switch part having a double heterostructure consisting of an upper metal electrode and including the waveguide layer in the thickness direction, a power supply for injecting current into the switch part, and an angle θ to the switch part through the waveguide layer. 1. A semiconductor optical switch comprising a light input means for inputting guided light. (2) The semiconductor substrate is made of n-type InP, the waveguide layer is n-type nGaA8P, and the cladding layer is a p-type In2 layer. Semiconductor optical switch described in Section 1. (3) The semiconductor optical switch according to claim 1 or 2, wherein the light input means is configured as a linear optical waveguide in which the waveguide layer has a rib structure. (4) The semiconductor optical switch according to claim 1, 2, or 3, wherein the angle θ is 4° or less. (5) The linear optical waveguide is at an angle of 2θ at the bottom of the switch section.
5. The semiconductor optical switch according to claim 3, wherein the semiconductor optical switch intersects at .