JPS61151627A - Optical switch - Google Patents

Abstract

PURPOSE:To make an optical switch very small in size and, at the same time, to operate the optical switch under low-voltage conditions, by using a gain guiding type waveguide which utilizes a gain difference caused by carrier injection into a semiconductor material as one of plural channel waveguides. CONSTITUTION:A buffer layer 4, activating layer 5, cladding layer 6, and gap layer 7 are piled up on an N-InP substrate 3 and P-diffused areas 9a and 9b are formed through an SiO2 mask 8. The electric current injecting part of the activating layer 5 acts as a channel waveguide and, when electric currents are injected into electrodes 11a and 11b, incident light 31a advances straight on and becomes outgoing light 31b. When the current injection into the electrode 11b on the opposite side of the incident side is stopped, a level difference in refractive index is produced at the central part of the intersecting section and the light 31a is totally reflected and becomes another outgoing light 31c. Therefore, when the current injection into the P-side electrode on one side is turned on/off, this optical switch makes switching operations and the voltage variation required for making the switching operations is about 2V. Accordingly, an optical switch which is very small in size and low in voltage can be obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an optical switch that switches optical signals as they are.

(Prior art and its problems) With the full-scale commercialization of optical communication systems in recent years, systems that provide new functions and services not available in the past are being considered. An example of a device required in such a system is an optical switch that switches connections between a large number of optical transmission lines at high speed.

Conventionally, so-called mechanical switches that move prisms, lenses, or the optical transmission line themselves have been widely used as such optical switches. Considering the demand, it is thought that non-mechanical switches that can be integrated will become mainstream in the future.

One possible method for such an optical switch is to convert an optical signal into an electrical signal, perform switching using an electrical switching circuit, and then convert the electrical signal back into an optical signal. Considering its application to future wavelength division multiplexing systems, etc., an optical switch of the type that switches optical signals up to t of light is desirable. Known optical switches for this purpose include directional coupler type optical switches and channel waveguides that form crossovers or Y-branches and perform switching by utilizing changes in refractive index due to electro-optic effects.

Among these, those using crossed waveguides are described in the magazine I.E.E. Journal of Quantum Electronics (I FiEE Journal of
Quantum Electronics) QB-
Volume 14, 1978, pp. 513-517.
8, Tsuhui (C.

S, Tsai) are described in detail in other papers.

Regarding those using branched waveguides, the magazine Applied Physics Letters (Applied Phys.
cs Letters) Volume 29, 1976,
W, on pages 790-792, Parnes (W,
(Burns) et al. These switches generally have the advantage of being smaller and more suitable for integration than directional coupler diffuser switches, and are suitable for multi-channel and
It is suitable for realizing a switch. Here, the structure and operation of an optical switch using crossed waveguides will be explained. FIG. 3 shows the structure of a conventionally known optical switch using crossed waveguides. Here, a switch in which a waveguide is formed by diffusing the most common two-plate LiNbO5KTi will be explained as an example. 2-plate LiNbO
5. 92 waveguides 22a are formed by diffusion of KTi on the substrate 21.
, 22b are formed to intersect with each other. In addition, an electrode 23a and a ground electrode 23 are placed on the waveguide via a suitable buffer layer.
b is formed.

When no voltage is applied to the electrode 23a, the light incident on the waveguide 22b travels straight due to the straightness of the light. When a voltage is applied to the electrode 23g, the electro-optic effect causes the electrode 2
The refractive index of the lower part of 3a changes, and a refractive index change occurs at the center of the intersection. Therefore, if the crossing angle is set shallow, the condition for total reflection of light is satisfied due to this refractive index change, and the light is reflected to the crossing waveguide 22a. Switch operation is obtained in this way, but since the refractive index change obtained by the electro-optic effect is extremely small, the voltage required for switch operation is several tens of thousands.
V or more, and it was not practical.

On the other hand, in a switch using a branching four-wave path, switching is performed by controlling the refractive index distribution at the branching portion using an electro-optic effect, but similarly, the voltage required for the switching operation is extremely high.

(Objective of the Invention) An object of the present invention is to eliminate the above-mentioned problems and provide an optical switch that is extremely compact and capable of low voltage operation.

(Structure of the Invention) According to the present invention, a plurality of channel waveguides form a branch or intersection, and an optical switch performs optical switching between the channel waveguides forming the branch or intersection. An optical switch is obtained in which at least one of the plurality of channel waveguides is made of a semiconductor material and is a gain waveguide type waveguide that utilizes a gain difference due to carrier injection.

(Principle of the Invention) First, the waveguide principle of a gain waveguide type waveguide using a gain difference due to carrier injection using a semiconductor material used in the present QIK will be explained. Generally speaking, when considering an optical waveguide in which light is confined only in one direction, the electric field intensity distribution Elx of light in the direction in which light is confined (here, the X direction)
l is determined by the following wave equation using the complex permittivity distribution ε((transition) in this direction.

(−' + k” g (x) ) Ft(x)=β”
g(ddx” where = 2π/λ is the wave number, β is the propagation constant in the propagation direction of light, and ε(xi is the refractive index n(xl, local gain G(
x), it can be written as ε(X) = (n-density) + j G(→/2k)''.For optical waveguides that have distributions in both the refractive index and gain, semiconductor lasers Many studies have been conducted in connection with the stabilization of the horizontal transverse mode of
In 1974, the paper by Suematsu and Yamada, "Oscillation state and state control of striped semiconductor lasers," published in 434-440, describes the case where the refractive index n (→2) has a rectangular distribution of local gain (x). A detailed study has been carried out.

According to these results, not only when there is no change in the refractive index n between the waveguide layer and the cladding layer sandwiching it, but also when the waveguide layer has a lower refractive index than the cladding layer. However, it has been shown that when the waveguide layer has a higher local gain than the cladding layer, there is a region where stable waveguide modes exist. This means that if you apply gain to a part of a one-dimensional active waveguide made of a semiconductor material (specifically, a planar double heterostructure semiconductor laser, etc.) using a button such as current injection, that part becomes a two-dimensional active waveguide. This shows that it can be used as a waveguide. This point will be explained in more detail below using the drawings. FIG. 4 is a diagram showing the structure of a normal planar stripe type semiconductor laser used to explain the waveguide effect due to the difference in gain. Regarding such lasers, the magazine Japan Kushi Journal op Abride-Physics (Japan Journal of
Applied Physics) Volume 12, 1973
Year, 1585-1592-! - Yonezu (H, Y) published in
onezu) AA in other papers! The case where GaAs-based materials are used is described in detail.

Here, a case where InGaAsP/InP-based material is used is shown as an example. n-In on the n-4nP substrate 3
n-InGaAsP active layer 5 via P buffer layer 4,
A P-InP cladding layer 6 and an n-InGaAsP cap layer 7 are formed. 5i on this epitaxial layer side
A P region 9 is formed by selectively diffusing Zn, Cd, etc. through a 0.02 mask 8 until it reaches the P-InP cladding layer 6.

Ohmic electrodes 10 and 11 are formed of Au-Ge-Ni metal on the substrate side and Au-Zn or Ti-Pt-Au metal on the epitaxial layer side.

In such a structure, since the InGaAsP active layer 5 has a higher refractive index than the InP layers above and below it, a strong refractive index waveguide is formed in the layer thickness direction.

On the other hand, there is no particular refractive index distribution in the direction parallel to the layers. When this element is forward biased and a current is injected, the current is selectively injected into the active layer 5 from the p region 9. Figure 5 shows the stacked structure near the active layer (a) and the distribution of carrier density (b) in the X direction, gain (C), and refractive index (d) in the active layer near the threshold value and above. This is what is shown. Due to lateral diffusion, the density of the middle Yaria decreases toward the periphery, showing an upwardly convex distribution as shown in FIG. 5(b). Gain is approximately proportional to carrier density, and refractive index is inversely proportional to carrier density, so their respective distributions are shown in Figure 5 (c
l, (di). In the X direction, the refractive index alone does not have a waveguiding effect. However, above the threshold value, there is a large gain difference between the center and the periphery, so a stable waveguiding mode exists. The stabilization mechanism of the transverse mode in a planar stripe laser was due to such a gain waveguide effect. In use, the same can be expected if the wavelength of the injected light is near the center of the wavelength distribution of gain.Furthermore, in this way, the optical waveguide effect can be controlled by the presence or absence of gain, that is, the presence or absence of current injection. Without current injection, it is possible to create a state where there is no optical waveguide effect at all.Therefore, if such a gain waveguide type waveguide is used in an optical switch, the refractive index distribution of the waveguide will change due to the electro-optic effect. This makes it possible to significantly increase the change in the waveguide state compared to when perturbation is applied.The present invention utilizes this principle, and provides an optical switch that is extremely compact and capable of low voltage operation. can get.

(Example) The present invention will be explained in detail below using examples.

FIG. 1 is a diagram showing a first embodiment of an optical switch according to the present invention, in which (a) is a top view of the switch excluding the p-side electrode, (bl), (c) is an end face and B- A cross-sectional view at B' is shown. Note that (bl (c) also shows the electrodes. First, the method for manufacturing such an optical switch will be explained using InGaAsP/InP-based materials as an example. is also B- like a normal semiconductor laser.
Kn-InP buffer layer 4 on InP substrate 3, n-I
nGaAsP active layer5. It is manufactured using a double heterostructure wafer on which an n-InGaAsP two-rad layer 6 and a P-InGaAsP cap layer 7 are grown. First, diffusion regions 9a, 9b of Zn, Ca, etc. are formed on a double heterostructure wafer through an 8i0* mask 8, except for the central portion 9c of the intersection, so as to reach the n-1nGa&sP cladding layer 6.

On top of that, the P [ItK pole is divided at the straight line -A' part to form two P diffusion regions 9g and 9b so that current can be injected independently, and the substrate side is made of An-Ge-Ni etc. Electrode 1
form 0. Then, it is cleaved on a plane perpendicular to the straight line so as to include the intersection near the center to form an end face to form a switch. Next, the operation of this switch will be explained. As explained above, the portion of active layer 5 into which current is injected functions as a channel waveguide. Therefore, the electrodes 11a, 11
When current is injected simultaneously with bK, intersecting waveguides are formed in the active layer 5. Therefore, the light 31a+ that entered one waveguide from the end face travels straight, and the output light 31b
is extracted as. Therefore, when the current injection to the P@ electrode 11b on the opposite side to the light incident side is stopped, the waveguide in that part disappears, and a large step in the refractive index occurs at the center of the intersection along the straight line. Therefore, the incident light 31a is totally reflected and extracted as the output light 31c. In other words, one P
A switch operation is obtained by turning on and off the current injection into the side electrodes. The voltage change required to turn on and off the current injection is determined by the built-in potential of the semiconductor and is very low, about 2V at most. Moreover, the advantage of small size, which is a characteristic of switches using crossed waveguides in general, is maintained. Therefore, a very compact and low voltage switch can be obtained, and a multi-channel switch can be easily obtained. In the structure of the embodiment, if the width of the portion 9c in which the P region is not formed at the center of the intersection is too wide, the portion where no waveguide is formed will be wide, which will adversely affect switch operation and loss, and if it is too narrow, current will be injected independently. becomes impossible. Therefore, it is desirable to form the portion 9e not only by simple selective diffusion but also by inactivating it by proton implantation or by forming a slit by dry etching. Also, since a normal semiconductor laser has a resonator configuration, the transmission characteristics of light injected from the outside have strong wavelength selectivity, but this type of optical switch does not require a resonator configuration, so it uses a traveling wave type configuration. In that case, switching can be performed over a wide wavelength range within the gain band of the medium. Even in such a case, the gain characteristics of a traveling wave optical amplifier can be expected, so the insertion loss of the optical switch will be extremely small.

FIG. 2 shows a second embodiment of the present invention, which is applied to a switch using a Y-branch waveguide.

In the figure, (a) is a diagram showing the top surface of the friend device excluding the electrodes, (
b) and lc) show cross sections along C-σ and DD'. This embodiment is also a normal double heterostructure wafer (n-InP buffer layer 4 on n-InP substrate 3).
g n-InGaAsP active layer5. P-InGa
AsP cladding layer 6゜n-InGm layer 7
grown. ). The P diffusion regions are separated as 9j, 9e, and 9f through the K 8 i02 mask 8 on the epitaxial layer side of the double heterostructure wafer, and are formed so as to have a Y-branch type as a whole. On top of that, pg
The ml electrode is formed by dividing it, and the mouth electrode is formed entirely on the substrate side. The operation of the second switch is also essentially the same as in the previous embodiment. First, when current is injected into all of the P diffusion regions, since a gain waveguide type Y branch waveguide is formed in the active layer 5, the light 31a incident on the non-branched waveguide is branched. , are output as 31b and 31c, respectively.

If current injection into the P diffusion region 9f is stopped at this point, the waveguiding effect disappears, so the output light 31c disappears. Similarly, P
If the current injection into the diffusion region 9e is stopped, the output light 31b can disappear, and the switch operation becomes possible. In this case as well, as mentioned above, a small, low-voltage switch can be obtained, and by integrating a large number of switches, a multi-channel optical switch or the like can be easily realized.

In the above embodiments, all the waveguides used in the switch are gain waveguide type waveguides, but it is clear that some of the waveguides may be passive index waveguide type guides. Also, for the sake of convenience, all materials used up until now have been InGaAs.
It goes without saying that the P/InP-based material is not limited to the gold material used in the example.

(Effects of the Invention) As described in detail above, according to the present invention, a switch suitable for integration that is compact and can be driven at low voltage can be obtained, and can be used in multi-channel optical switches, various optical integrated circuits, and optical systems. This will greatly contribute to the realization of this.

[Brief explanation of the drawing]

Figure 1 (at, (b), (c), Figure 2 (al
* (b) - (c) are diagrams showing an embodiment of the optical switch according to the present invention, Figure 3 is a diagram for explaining a conventional crossed waveguide type optical switch, and Figures 4 and 5 are diagrams showing the embodiment of the optical switch according to the present invention. FIG. 2 is a diagram illustrating the principle of a gain waveguide used in FIG. In the figure, 3, 4, 5, 6, 7 are semiconductors, 10°11, l
la, llb, llc, 23a, 23b are electrodes, 9a,
9b, 9d, 9e, and 9f are P diffusion regions, 9c is a non-diffusion region, and 31a, 31b. 31c is light, 21 is LiNbO3 substrate, 22a,
22b is a waveguide, and 8 is 5i02. Closed 1 figure (b) (C) 3: n-r
nP group jl 4:n-1nGaAxPt(
-t7y layer 5: n4GaAsP 4''u@/s:
P-1n/'Kura'7do7: 11-1*GaAsF
"Fr-y7♂/91 fj:5j (hna, ub
, to: tgaya, qb: ps slope rafG: ,
F, blame! Notii 4'Ill carp k 31α, 3tb
, 3tc: Figure 2(b) (S3: η-1 false P basis 5: 71-1n Ga As P pickled t'EMrid
9e Ta f: P JE 5jiζ4to or 3rd
Figure 2f 77 anti-Li N b 03 substrate 23α electric 1 23b specialization'! ! , fm Figure 4 3 n-1flP group JgL 4 'n-InF bar sofa layer 5' r14nQaAsP night property 1 6 F-1nP Gra-donohe 1'?
'n-IMGaAsP+Yap 18 S;Ox Mask 9 F Buddha Huan @ Thief Figure 5 7l-1qxGaAsP:, & Small% 36 F
4 Bamboo CraAsP Khufu Shi17 n-1fIGaA
sP jF Yahoo! 9 P 絋ia@area

Claims (1)

[Claims] In an optical switch in which a plurality of channel waveguides form branching or crossing portions, and light is switched between the branching or crossing channel waveguides, at least one of the plurality of channel waveguides is connected. An optical switch characterized by having a gain waveguide type waveguide that utilizes a gain difference caused by carrier injection through a semiconductor material.