JP2011133582A - Semiconductor mach-zehnder type optical modulator and method of manufacturing the same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a semiconductor Mach-Zehnder type optical modulator which is compact, is accurately push-pull driven and has a low loss. <P>SOLUTION: The semiconductor Mach-Zehnder type optical modulator includes: a demultiplexer 11a; a multiplexer 11b; semiconductor optical waveguides 13 and 14; a signal electrode 15; a positive bias electrode 17; and a negative bias electrode 18, wherein the demultiplexer 11a and the multiplexer 11b are connected by two semiconductor optical waveguides 13 and 14, the signal electrode 15 is connected to the upper part of the two semiconductor optical waveguides 13 and 14 and a common high frequency signal can be input to the upper part of the two semiconductor optical waveguides 13 and 14, the positive bias electrode 17 is connected to the lower part of the semiconductor optical waveguide 13, and the negative bias electrode 18 is connected to the lower part of the other semiconductor optical waveguide 14. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a semiconductor Mach-Zehnder optical modulator and a method for manufacturing the semiconductor Mach-Zehnder optical modulator.

  In recent years, ultra-high-speed, large-capacity, long-distance optical communication backbone systems have been developed and put into practical use centering on WDM (Wavelength Division Multiplexing). As a modulation method in such a system, for example, binary phase shift keying (BPSK) is exemplified. The light modulation in BPSK is performed by, for example, a Mach-Zehnder type optical modulator.

  The Mach-Zehnder modulator guides light through two optical waveguides called arms, and performs modulation using interference of the light. In the Mach-Zehnder modulator, the demultiplexing of light to the two optical waveguides and the multiplexing of light from the two optical waveguides are performed by, for example, 2 × 2-MMI (Multimode Interference). The light modulation in BPSK is performed by changing the light amplitude from 1 to -1 while keeping the light phase constant. When the phase of the light fluctuates, wavelength chirping occurs. When this wavelength chirping occurs, dispersion of light by the optical fiber occurs, and the transmission waveform of light breaks down. Therefore, it is important to modulate the light phase while keeping it constant.

  As described above, in order to perform modulation while keeping the light phase constant, for example, a voltage (high frequency signal) is applied to the two optical waveguides of the Mach-Zehnder optical modulator so that the light phase moves in the opposite direction. . At this time, if the high frequency signals applied to both the optical waveguides are V1 and V2, respectively, for example, V1 + V2 is made constant. Such a driving method is called push-pull driving. As a Mach-Zehnder type optical modulator that performs push-pull driving, for example, one that applies two high-frequency signals of opposite phases can be cited (for example, see FIG. 5 of Patent Document 1). An example of this Mach-Zehnder type optical modulator is shown in FIG. In the Mach-Zehnder type optical modulator 30, the guided light incident on the optical waveguide 39a is demultiplexed and multiplexed by 2 × 2-MMIs 31a and 31b and provided separately to the respective arms (optical waveguides 33 and 34). Two high-frequency signals having opposite phases are applied from the signal electrodes 35a and 35b. By modulating the light in this way, the Mach-Zehnder optical modulator 30 is push-pull driven. The signal electrodes 35a and 35b together with the ground electrode 36 form a coplanar transmission line. The combined light is emitted from the optical waveguide 39b.

  Further, as a Mach-Zehnder type optical modulator that performs push-pull drive, for example, an input of in-phase signals has been proposed (see Patent Document 2). In the semiconductor PIN Mach-Zehnder modulator described in Patent Document 2, when the coupled quantum wells in the intrinsic layers of both arms (optical waveguides) are appropriately designed, the refraction of both arms can be achieved when a DC bias is applied (when an electric field is applied). The phenomenon that the rate changes in reverse is used. It is said that push-pull drive can be performed by this coupled quantum well.

  Also, as a Mach-Zehnder type optical modulator that performs push-pull drive, for example, one that is driven by a single driver has been proposed (see Patent Document 3). In the semiconductor Mach-Zehnder modulator described in Patent Document 3, two pin diodes are connected in series, and a constant DC reverse bias is applied to the whole. In this state, push-pull driving is performed by applying a single driver signal to a connection point between one first conductivity type cladding layer and the other second conductivity type cladding layer in both pin diodes.

  As a Mach-Zehnder type optical modulator that performs push-pull drive, for example, there is one that uses a slot line type high-frequency transmission line for two optical waveguides (see Patent Document 4). In the Mach-Zehnder optical modulator described in Patent Document 4, an electric field is generated from the second phase modulation electrode to the high conductivity layer, and at the same time, an electric field is generated from the high conductivity layer to the first phase modulation electrode. By doing so, the directions of the electric fields in the first waveguide and the second optical waveguide are reversed. Thereby, push-pull drive can be performed.

Japanese Patent Laid-Open No. 9-218384 JP 2004-318095 A JP 10-333106 A JP 2004-151590 A

  In the Mach-Zehnder type optical modulator shown in FIG. 3, as described above, two high-frequency signals having opposite phases are applied from the signal electrodes 35a and 35b separately provided in the optical waveguides 33 and 34, respectively. Therefore, in order to suppress the crosstalk between the two high-frequency signals having opposite phases, it is necessary to arrange the optical waveguides with a space therebetween. As a result, it is difficult to reduce the size of the Mach-Zehnder optical modulator shown in FIG.

  In the semiconductor PIN Mach-Zehnder modulator described in Patent Document 2, as described above, the push-pull operation is enabled by the design of the coupled quantum well. However, designing coupled quantum wells that allow accurate push-pull operation is difficult.

  In the semiconductor Mach-Zehnder modulator described in Patent Document 3, the high-frequency electrode (signal electrode) includes a first conductive clad layer (n-type InP clad layer) located below one pin diode, and the other pin diode. It is formed between the second conductivity type cladding layer (p-type InP cladding layer) located on the upper part. As described above, since the high-frequency electrode is formed in the valley between the two pin diodes, it is necessary to separate the pin diodes (between the optical waveguides) to some extent. For this reason, miniaturization is difficult.

  Also in the Mach-Zehnder type optical modulator described in Patent Document 4, the distance between the two optical waveguides is required to be, for example, 8 μm or more (paragraph 0038 of Patent Document 4).

  Accordingly, an object of the present invention is to provide a semiconductor Mach-Zehnder type optical modulator that is small, accurately push-pull driven, and has low loss, and a method for manufacturing the semiconductor Mach-Zehnder type optical modulator.

In order to achieve the above object, a semiconductor Mach-Zehnder optical modulator according to the present invention includes:
A duplexer, a multiplexer, a semiconductor optical waveguide, a signal electrode, a positive bias electrode, and a negative bias electrode;
The semiconductor optical waveguide is two,
The duplexer and the multiplexer are connected by the two semiconductor optical waveguides,
The signal electrode is connected to the upper part of the two semiconductor optical waveguides, and can input a common high-frequency signal to the two semiconductor optical waveguides,
The positive bias electrode is connected to a lower portion of one of the two semiconductor optical waveguides, and the negative bias electrode is connected to a lower portion of the other semiconductor optical waveguide.

The manufacturing method of the semiconductor Mach-Zehnder type optical modulator of the present invention includes:
A duplexer forming process for forming the duplexer;
A multiplexer forming step for forming a multiplexer;
A semiconductor optical waveguide forming step of forming two semiconductor optical waveguides connecting the duplexer and the multiplexer;
Forming a signal electrode connected to the upper part of the two semiconductor optical waveguides and capable of inputting a common high-frequency signal to the two semiconductor optical waveguides;
A positive bias electrode forming step of forming a positive bias electrode connected to a lower portion of one of the two semiconductor optical waveguides;
And a negative bias electrode forming step of forming a negative bias electrode connected to the lower part of the other optical semiconductor optical waveguide.

  ADVANTAGE OF THE INVENTION According to this invention, the manufacturing method of a semiconductor Mach-Zehnder type optical modulator and a semiconductor Mach-Zehnder type optical modulator which are small, accurately push-pull drive, and low loss can be provided.

(A) is a top view which shows the structure of an example in Embodiment 1 of the semiconductor Mach-Zehnder type | mold optical modulator of this invention. (B) is sectional drawing seen in the II direction of Fig.1 (a). It is sectional drawing which shows the structure of an example in Embodiment 2 of the semiconductor Mach-Zehnder type | mold optical modulator of this invention. It is a top view which shows the structure of an example of the conventional Mach-Zehnder type | mold optical modulator.

  Hereinafter, the problems of the background art described above will be described in more detail. The inventors of the present invention focused on these points, and as a result of intensive studies, reached the present invention.

  First, in the Mach-Zehnder type optical modulator shown in FIG. 3, it is necessary to suppress crosstalk as described above. In order to suppress crosstalk when the length of both optical waveguides 33 and 34 is about 3 mm from the relationship between the change in refractive index with respect to voltage application and the light absorption loss, etc., between the optical waveguide 33 and the optical waveguide 34, the crosstalk is suppressed. For example, an interval of the order of 100 μm is necessary. In order to separate the optical waveguides by 100 μm, for example, the length of the S-shaped waveguide between the optical waveguides, the duplexer, and the multiplexer needs to be about 200 to 300 μm. If it is made shorter than this, the curvature of the S-shape becomes large, and the radiation loss of guided light cannot be ignored. In this case, the insertion loss as a Mach-Zehnder type optical modulator is unacceptable. However, as described above, in order to suppress crosstalk, the two optical waveguides cannot be brought close to each other, and the S-shaped waveguide cannot be omitted or shortened. As a result, it is difficult to reduce the size of the Mach-Zehnder optical modulator shown in FIG.

  In the semiconductor PIN Mach-Zehnder modulator described in Patent Document 2, as described above, the refractive index of both arms changes inversely when a DC bias is applied (when an electric field is applied) by appropriately designing a coupled quantum well. However, push-pull drive is possible. However, it is quite difficult to design a coupled quantum well so that the refractive index change with respect to applied electric field is always exactly reversed over a range of electric field strengths. Therefore, depending on the phase change of the light, the push-pull drive cannot be performed accurately due to the error of the refractive index change, which may cause wavelength chirping and transmission error.

  In the semiconductor Mach-Zehnder modulator described in Patent Document 3, one first conductivity type cladding layer (n-type InP cladding layer) and the other second conductivity type cladding layer (p-type InP cladding layer) in both the pin diodes are provided. Since a single driver signal is applied to the connection point, there is no crosstalk problem between the two pin diodes. However, as described above, since the high-frequency electrode is formed in the valley between the two pin diodes, it is necessary to separate the two pin diodes to some extent, and it is difficult to reduce the size.

  Further, in the Mach-Zehnder type optical modulator described in Patent Document 4, since a slot line type high-frequency transmission line is used for two optical waveguides, the two optical waveguides are relatively close to each other. However, considering impedance matching, a gap of at least 3 μm is required between the electrodes of the two optical waveguides. In order to ensure the gap between the electrodes, the distance between the two optical waveguides must be, for example, 8 μm or more. The length (MMI length) of the 2 × 2-MMI used in the duplexer that branches the two optical waveguides and the multiplexer that joins the two optical waveguides is approximately the width (MMI width). It is proportional to the square. For this reason, when the distance between the two optical waveguides that are input to and output from the 2 × 2-MMI is as wide as 8 μm or more as described above, the MMI length becomes too long. To avoid this, an S-shaped optical waveguide is required. Thus, the length of the MMI length and the necessity of the S-shaped optical waveguide are in a trade-off relationship. As a result, even the Mach-Zehnder optical modulator described in Patent Document 4 is difficult to downsize. In order to overcome the trade-off relationship, it is desirable to further reduce the distance between the optical waveguides.

  Hereinafter, the semiconductor Mach-Zehnder type optical modulator of the present invention will be described in detail with examples. However, the present invention is not limited to the following embodiments.

(Embodiment 1)
The semiconductor Mach-Zehnder type optical modulator of this embodiment is a semiconductor Mach-Zehnder type optical modulator using an InP system as a constituent material.

  FIG. 1 shows a configuration of an example of a semiconductor Mach-Zehnder optical modulator according to this embodiment. FIG. 1A is a plan view of a semiconductor Mach-Zehnder optical modulator according to this embodiment. FIG.1 (b) is sectional drawing seen in the II direction of Fig.1 (a). As shown in FIGS. 1A and 1B, the semiconductor Mach-Zehnder optical modulator 10 includes a 2 × 2-MMI 11a that is a duplexer, a 2 × 2-MMI 11b that is a multiplexer, and 2 The semiconductor optical waveguides 13 and 14, the signal electrode 15, the positive bias electrode 17, and the negative bias electrode 18 are included. The 2 × 2-MMI 11 a (demultiplexer) and the 2 × 2-MMI 11 b (multiplexer) are connected by two semiconductor optical waveguides 13 and 14. The 2 × 2-MMI 11a is connected to the input optical waveguide 19a. The 2 × 2-MMI 11b is connected to the output optical waveguide 19b. A p-type contact layer 134 is formed on the semiconductor optical waveguide 13. An n-type contact layer 144 is formed on the semiconductor optical waveguide 14. The signal electrode 15 is formed on the p-type contact layer 134 and the n-type contact layer 144. The positive bias electrode 17 is electrically connected to the lower part of the semiconductor optical waveguide 13. The negative bias electrode 18 is electrically connected to the lower part of the semiconductor optical waveguide 14.

  In the present invention, the “positive bias electrode” is an electrode to which a voltage higher than the “negative bias electrode” is applied, and is not limited to an electrode to which a positive voltage is applied. The “negative bias electrode” is an electrode to which a voltage relatively lower than that of the “positive bias electrode” is applied, and is not limited to an electrode to which a negative voltage is applied.

  Both the semiconductor optical waveguides 13 and 14 are directly connected to the 2 × 2-MMI 11a which is a duplexer and the 2 × 2-MMI 11b which is a multiplexer without using an S-shaped optical waveguide. Both the 2 × 2-MMIs 11a and 11b have a width of, for example, 11 μm and a length of, for example, 200 μm. However, the widths and lengths of both 2 × 2-MMIs 11a and 11b are not limited to this example, and may be different sizes according to, for example, generally known MMI design theory.

  In the semiconductor Mach-Zehnder optical modulator of this embodiment, the signal electrode 15 is formed on the p-type contact layer 134 and the n-type contact layer 144 as described above. For this reason, in the semiconductor Mach-Zehnder type optical modulator of this embodiment, a common high-frequency signal can be input from the signal electrode 15 to the upper portions of both the semiconductor optical waveguides 13 and 14. Does not need to consider the problem of crosstalk. As a result, both the semiconductor optical waveguides 13 and 14 can be brought close to each other, and as described above, both the semiconductor optical waveguides 13 and 14 are connected to both the 2 × 2-MMIs 11a and 11b without using the S-shaped optical waveguide. Can do. As described above, the semiconductor Mach-Zehnder type optical modulator of this embodiment is small because the two semiconductor optical waveguides are close to each other and the S-shaped optical waveguide is omitted. Furthermore, since the S-shaped optical waveguide is omitted, there is no insertion loss such as radiation loss or internal loss due to the S-shaped optical waveguide, and the loss is low.

  The semiconductor Mach-Zehnder type optical modulator of the present invention is small in size, so that, for example, the chip size can be reduced. For this reason, for example, the probability that one chip hits a certain defect is lowered, and the chip yield is improved. As a result, for example, the chip cost can be reduced.

  The semiconductor Mach-Zehnder optical modulator of the present invention preferably has a structure in which the S-shaped optical waveguide is omitted as shown in FIG. 1, for example, but the present invention is not limited to this example. For example, the semiconductor Mach-Zehnder type optical modulator of the present invention may have a structure in which the S-shaped optical waveguide is shortened as much as possible by shortening the distance between the two optical waveguides. Even in such an embodiment, the two optical waveguides are brought close to each other and the S-shaped optical waveguide is shortened, so that the size can be reduced. Furthermore, since the S-shaped optical waveguide is shortened, the insertion loss due to the S-shaped optical waveguide can be suppressed and the loss can be reduced.

  As described above, the S-shaped optical waveguide has a trade-off relationship between miniaturization and low loss. That is, if the S-shaped optical waveguide is downsized, the curvature increases and radiation loss and the like increase. On the other hand, if an attempt is made to suppress radiation loss and the like as much as possible, the S-shaped optical waveguide needs to be gently bent, so that it becomes very long. Since the semiconductor Mach-Zehnder type optical modulator of the present invention can omit or shorten the S-shaped optical waveguide as described above, there is a great effect that such a trade-off relationship can be eliminated or greatly improved.

  The semiconductor optical waveguides 13 and 14 are formed on the semi-insulating InP substrate 12. In the semiconductor waveguide 13, an n-type InP clad layer 131, an SCH-multiple quantum well layer (i-type) 132, and a p-type InP clad layer 133 are arranged on the semi-insulating InP substrate 12 from the bottom to the top (FIG. 1 ( b) in the vertical direction) having a conductive structure (pin structure) stacked in the above order. The SCH-multiple quantum well layer (i-type) 132 is a core layer of the semiconductor optical waveguide 13. The SCH is an abbreviation for Separate Confinement Heterostructure, and is a heterojunction structure for guiding light confinement. The semiconductor waveguide 14 has a p-type InP clad layer 141, an SCH-multiple quantum well layer (i-type) 142, and an n-type InP clad layer 143 on the semi-insulating InP substrate 12 from the bottom to the top (FIG. 1 ( b) in the vertical direction) having a conductive structure (pin structure) stacked in the above order. The SCH-multiple quantum well layer (i-type) 142 is a core layer of the semiconductor optical waveguide 14. In this semiconductor Mach-Zehnder type optical modulator 10, the conductive type structures in the vertical direction of the semiconductor optical waveguides 13 and 14 are structures that are vertically inverted.

  The layer thicknesses of the n-type InP cladding layer 131 and the p-type InP cladding layer 141 are, for example, 1 ± 0.7 μm, respectively. The two layers are formed so as to cover the upper surface of the left part and the right part (left and right direction in FIG. 1B) of the semi-insulating InP substrate 12 with an interval of 3 ± 2 μm, for example. ing. The SCH-multiple quantum well layer (i-type) 132 and the p-type InP cladding layer 133 are on the side facing the p-type InP cladding layer 141 on the upper surface of the n-type InP cladding layer 131 (the right side in FIG. 1B). The semiconductor optical waveguide 13 is formed together with the n-type InP cladding layer 131 by being laminated at the edge. The SCH-multiple quantum well layer (i-type) 142 and the n-type InP cladding layer 143 are on the side facing the n-type InP cladding layer 131 on the upper surface of the p-type InP cladding layer 141 (left side in FIG. 1B). The semiconductor optical waveguide 14 is formed together with the p-type InP clad layer 141 on the edge. As described above, the p-type contact layer 134 is formed on the upper surface of the p-type InP cladding layer 133 (on the semiconductor optical waveguide 13). As described above, the n-type contact layer 144 is formed on the upper surface of the n-type InP cladding layer 143 (on the semiconductor optical waveguide 14).

  In the present invention, “on top” may be in a state of being in direct contact with the upper surface or a state in which other components are disposed therebetween unless otherwise specified. Unless otherwise specified, “under” may be in a state of being in direct contact with the lower surface or in a state in which other components are disposed therebetween. Further, “on the top surface” means a state in which the top surface is in direct contact unless otherwise specified. “On the bottom surface” means in direct contact with the bottom surface unless otherwise specified.

  A semi-insulating InP buried layer 151 a is formed between the semiconductor optical waveguide 13 and the p-type contact layer 134 and the semiconductor optical waveguide 14 and the n-type contact layer 144. The semi-insulating (SI: Semi-Insulating) InP buried layer 151a is a layer having a high resistance by doping Fe or Ru, for example. The buried layer 151a may be a buried layer formed of a resin such as polyimide or BCB (benzocyclobutene resin) instead of InP. If polyimide or BCB is used as the material for forming the buried layer, the refractive index difference between the buried layer and both SCH-multiple quantum well layers (i-type) 132 and 142 can be increased. Thereby, confinement of guided light in both SCH-multiple quantum well layers (i-type) 132 and 142 can be strengthened, and coupling of guided light between both semiconductor optical waveguides 13 and 14 can be reduced. As a result, the distance between the two semiconductor optical waveguides 13 and 14 can be made narrower.

  The widths of the SCH-multiple quantum well layer (i-type) 132 and the SCH-multiple quantum well layer (i-type) 142 are, for example, 1.3 ± 0.8 μm, respectively. In order not to couple the guided lights of both semiconductor optical waveguides 13 and 14, the gap between both SCH-multiple quantum well layers (i-type) 132 and 142 is preferably 1 μm or more, and more preferably 2 μm or more. . Of the p-type InP cladding layer 141, the portion immediately below the SCH-multiple quantum well layer (i-type) 142 preferably has a low impurity concentration in order to reduce light absorption loss.

  The layer thickness of the p-type InP cladding layer 133 is, for example, 2 ± 1 μm. The same applies to the thickness of the n-type InP cladding layer 143.

  The p-type contact layer 134 can be formed from, for example, InGaAsP or InGaAs. Similarly to the p-type contact layer 134, the n-type contact layer 144 can also be formed from InGaAsP, InGaAs, or the like.

  On the upper surface of the n-type InP clad layer 131, a positive bias electrode 17 is formed on a part of the edge opposite to the semiconductor optical waveguide 13 (left side in FIG. 1B). The positive bias electrode 17 is in contact with (connected to) the n-type InP cladding layer 131, that is, the lower portion of the semiconductor optical waveguide 13. On the upper surface of the p-type InP cladding layer 141, a negative bias electrode 18 is formed on a part of the edge opposite to the semiconductor optical waveguide 14 (on the right side in FIG. 1B). The negative bias electrode 18 is in contact with (connected to) the p-type InP cladding layer 141, that is, the lower portion of the semiconductor optical waveguide 14.

  On the upper surface of the n-type InP cladding layer 131, a semi-insulating InP buried layer is interposed between the SCH-multiple quantum well layer (i-type) 132, the p-type InP cladding layer 133 and the p-type contact layer 134 and the positive bias electrode 17. Layer 151b is embedded. On the upper surface of the p-type InP clad layer 141, a semi-insulating InP buried is interposed between the SCH-multiple quantum well layer (i-type) 142, the n-type InP clad layer 143 and the n-type contact layer 144, and the negative bias electrode 18. Layer 151c is embedded. Thus, from the semi-insulating InP buried layer 151b, the p-type contact layer 134 on the semiconductor optical waveguide 13, the semi-insulating InP buried layer 151a, the n-type contact layer 144 on the semiconductor optical waveguide 14, and the semi-insulating InP buried layer. Over 151c, the top of these layers is flattened. The semi-insulating InP buried layers 151 b and 151 c are covered with a dielectric film 161. The material for forming the semi-insulating InP buried layers 151b and 151c may be the same as that of the InP buried layer 151a, for example. That is, the InP buried layers 151b and 151c are, for example, high resistance layers by doping Fe or Ru. For example, instead of semi-insulating InP, polyimide or BCB (benzocyclobutene, that is, benzocyclobutene resin) It may be a buried layer formed of a resin such as.

  Further, as described above, the signal electrode 15 is formed on the p-type contact layer 134 and the n-type contact layer 144. The signal electrode 15 is a traveling wave electrode. The signal electrode 15 is formed in contact with the upper surfaces of the p-type contact layer 134 and the n-type contact layer 144, thereby being connected to the upper portions of the two optical waveguides 13 and 14, and both the semiconductor optical waveguides 13 and 14. A common high-frequency signal can be input. Ground electrodes 16 a and 16 b are provided on the dielectric film 161 on both sides of the signal electrode 15. The ground electrodes 16a and 16b together with the signal electrode 15 constitute a coplanar transmission line. This coplanar transmission line is designed to be impedance matched with a high frequency circuit (not shown) for supplying a high frequency signal. Specifically, for example, the characteristic impedance of the coplanar transmission line is designed to be 50Ω. As shown in FIG. 1A, the coplanar transmission line is drawn to the end of the chip (the lower end in FIG. 1A) and electrically connected to an external 50Ω high frequency circuit (not shown). Connected.

  In the semiconductor Mach-Zehnder optical modulator of the present invention, the “upper part” of the semiconductor optical waveguide refers to a portion above the core layer (layer that becomes the optical path) of the semiconductor optical waveguide. That is, in the semiconductor Mach-Zehnder optical modulator of the present invention, the signal electrode is connected to a portion of the two semiconductor optical waveguides above the core layer, and is connected to the two semiconductor optical waveguides. A common high-frequency signal can be input (in the portion above the core layer of the two semiconductor optical waveguides). In the semiconductor Mach-Zehnder optical modulator of the present invention, the “lower part” of the semiconductor optical waveguide refers to a part below the core layer (layer that becomes the optical path) of the semiconductor optical waveguide. That is, the positive bias electrode is connected to a portion of one of the two semiconductor optical waveguides below the core layer. The negative bias electrode is connected to a portion of the other semiconductor optical waveguide below the core layer. In the present invention, the “connection” may be in direct contact or may be connected through another component. For example, the signal electrode may be in direct contact with the upper part of the two semiconductor optical waveguides, or may be connected via another component (for example, a contact layer). The positive bias electrode and the negative bias electrode may be in direct contact with the lower portions of the two semiconductor optical waveguides, or may be connected via other components (for example, a contact layer). Good. The semiconductor Mach-Zehnder optical modulator of the present invention needs to be able to apply a bias to the two semiconductor optical waveguides by the positive bias electrode and the negative bias electrode. For this reason, the positive bias electrode needs to be electrically connected to one lower part of the two semiconductor optical waveguides. Similarly, the negative bias electrode needs to be electrically connected to the lower part of the other semiconductor optical waveguide.

  In the semiconductor Mach-Zehnder type optical modulator of this embodiment, in order to input a common high-frequency signal to the upper part of both semiconductor optical waveguides, the signal electrode is formed on both semiconductor optical waveguides, and from the core layer of both semiconductor optical waveguides Also has a contact (connected) at the top. For this reason, as described above, the signal electrode can be a traveling wave electrode.

  In the semiconductor Mach-Zehnder optical modulator of the present invention, the signal electrode is not particularly limited, but it is preferable to use a traveling wave electrode, for example, as in this embodiment. For example, in the semiconductor Mach-Zehnder type optical modulator of Patent Document 3, since the signal electrode (high frequency electrode) is formed between the valleys of two arms, it is difficult to form a traveling wave electrode. Become. When the signal electrode is a lumped constant electrode, for example, at a very high symbol rate such as 25 Gsps, the time during which the guided light passes under the signal electrode (time shift) cannot be ignored. That is, if the signal electrode is a lumped constant type electrode and the symbol rate of the signal by the signal electrode is very high, the modulation waveform may be lost due to the time lag and a transmission error may occur. On the other hand, if a traveling wave electrode is used as the signal electrode, the high-frequency propagation speed that travels through this electrode (in this embodiment, the coplanar transmission line) matches the propagation speed of guided light that propagates through both semiconductor optical waveguides. Thus, the signal electrode can be designed. For this reason, it is possible to suppress a transmission error resulting from the collapse of the modulation waveform due to the time shift.

  In the semiconductor Mach-Zehnder type optical modulator of this embodiment, as described above, the portion where the signal electrode (coplanar transmission line) is formed is flattened. For this reason, for example, it is not necessary to contact the signal electrode (coplanar transmission line) with a layer below the core layer of the semiconductor optical waveguide. As a result, for example, when using a traveling wave electrode (transmission line) with a step contacted with a layer below the core layer, scattering, loss, reflection, etc. generated when a 25 Gsps high-frequency signal propagates are reduced. Effect is obtained.

  If the signal electrode 15 can input a common high-frequency signal to the p-type contact layer 134 and the n-type contact layer 144, the signal electrode 15 has a gap between the p-type contact layer 134 and the n-type contact layer 144. There may be gaps.

  The semiconductor Mach-Zehnder optical modulator of the present invention preferably further has a ground electrode (ground electrode) as in the present embodiment, for example.

  Next, the operation of the semiconductor Mach-Zehnder optical modulator of this embodiment will be described.

  The light input from the input optical waveguide 19 a is demultiplexed by the 2 × 2-MMI 11 a that is a demultiplexer, and is guided by both the semiconductor optical waveguides 13 and 14. A constant DC voltage is applied between the negative bias electrode 18 and the positive bias electrode 17. On the other hand, a common (in-phase) high-frequency signal is applied from the signal electrode 15 to the upper portions of both the semiconductor optical waveguides 13 and 14. Thereby, the refractive index of the SCH-multiple quantum well layers (i-type) 132 and 142 as the core layer is changed. Here, in the semiconductor Mach-Zehnder type optical modulator of this embodiment, since the vertical conductive type structures (pin structures) of the semiconductor optical waveguides 13 and 14 are vertically inverted with respect to each other, it is possible to accurately output a high-frequency signal having a reverse phase. Can be added. For this reason, the phase of the light guided through both the semiconductor optical waveguides 13 and 14 can be accurately moved in the opposite direction. As a result, in the semiconductor Mach-Zehnder type optical modulator of this embodiment, there is no variation in speed matching between the two semiconductor optical waveguides, and the push-pull drive is performed accurately. Thereby, for example, control of wavelength chirping is greatly improved.

  In the semiconductor Mach-Zehnder optical modulator of the present invention, as described above, the high-frequency signal applied to both arms (the two semiconductor optical waveguides) is applied by a common transmission line (signal electrode). For this reason, there is no variation in speed matching between both arms, and as described above, there is no problem of crosstalk in the first place. Therefore, according to the present invention, it is possible to provide a semiconductor Mach-Zehnder optical modulator that realizes accurate push-pull driving.

  The constant voltage is, for example, −5 ± 4V. With this applied voltage, the reverse current in the structure of the p-type InP layer 141 / semi-insulating InP buried layer 151a / n-type InP layer 131 can be suppressed sufficiently low. The structure of the p-type InP layer 141 / semi-insulating InP buried layer 151a / n-type InP layer 131 preferably has a capacitance of a certain value or less for high-speed modulation. Here, as described above, the thicknesses of both the n-type InP layer 131 and the p-type InP layer 141 are set to 1.7 μm or less, and the n-type InP layer 131 and the p-type InP layer 141 are separated by 1 μm or more, A desired capacity is achieved by applying a reverse bias of 1 V or more. Further, by setting the thicknesses of both the n-type InP layer 131 and the p-type InP layer 141 to 0.3 μm or more, the resistance value can be suppressed to a desired value or less. A preferable value of the resistance value is, for example, 10Ω or less.

  In the semiconductor Mach-Zehnder type optical modulator of this embodiment, the vertical conductive structure of both the semiconductor optical waveguides 13 and 14 is a pin structure, and the refractive index of the core layer is changed by applying a voltage to the pin structure. . However, the present invention is not limited to this example. The vertical layer structure of the semiconductor optical waveguide is, for example, n-type / p-type / i-type / n-type, n-type / semi-insulating type / i-type / n-type, or n-type / semi-insulating type / i. Type / semi-insulating type / n-type. More specifically, in the semiconductor Mach-Zehnder type optical modulator of the present invention, the two semiconductor optical waveguides are respectively p-type / i-type / n in the vertical direction from the negative bias side toward the positive bias side. Type order, n-type / p-type / i-type / n-type order, n-type / semi-insulating type / i-type / n-type order, or n-type / semi-insulating type / i-type / semi-insulating type It is preferable to have a stacked structure of / n type order. In the n-type / semi-insulating type / i-type / semi-insulating type / n-type stacked structure, at least one of the semi-insulating semiconductor layers may be replaced with a p-type semiconductor layer. . In the present invention, the positive bias electrode and the negative bias electrode are connected to the lower portions of the two semiconductor optical waveguides as described above. Therefore, “in the vertical direction from the negative bias side to the positive bias side” means “from the bottom to the top” in the semiconductor optical waveguide connected to the negative bias electrode, and is connected to the positive bias electrode. In a semiconductor optical waveguide, it means “from top to bottom”. Further, for example, the conductive structure in the vertical direction of the two semiconductor optical waveguides may be a structure in which the two are turned upside down. In the semiconductor Mach-Zehnder type optical modulator of FIG. 1, the conductive type structures in the vertical direction of the two semiconductor optical waveguides are vertically inverted as described above.

  As described above, the semiconductor Mach-Zehnder type optical modulator of this embodiment is a semiconductor Mach-Zehnder type optical modulator on an InP substrate, but the present invention is not limited to this example. For example, the semiconductor Mach-Zehnder type optical modulator of the present invention may be formed of a semiconductor capable of forming a pn junction. Specifically, for example, a semiconductor Mach-Zehnder optical modulator on a GaAs substrate or a semiconductor Mach-Zehnder optical modulator on a Si substrate may be used.

  The manufacturing method of the semiconductor Mach-Zehnder type optical modulator of the present invention is not particularly limited, but is preferably manufactured by the manufacturing method of the present invention. In the production method of the present invention, the order in which the steps are performed is not particularly limited, and may be any order, and may be sequential or simultaneous. Hereinafter, an example of a manufacturing method of the Mach-Zehnder optical modulator of the present embodiment will be described with reference to FIG.

  First, an undoped InP layer is grown on the entire surface of the semi-insulating InP substrate 12. In this undoped InP layer, an n-type region, a p-type region, and a semi-insulating region are formed by ion implantation or diffusion, for example. The region from the semiconductor optical waveguide 13 to the region where the positive bias electrode 17 is formed is defined as the n-type region (n-type InP cladding layer 131) by ion implantation of n-type impurities, for example. The region from the semiconductor optical waveguide 14 to the region where the negative bias electrode 18 is formed is defined as the p-type region (p-type InP cladding layer 141) by, for example, diffusion of p-type impurities. At this time, it is preferable to adjust, for example, a diffusion mask so that the p-type impurity concentration is lowered in a region corresponding to the SCH-multiple quantum well layer (i-type) 142 in the semiconductor optical waveguide 14. A region extending from the region where both 2 × 2-MMIs 11a and 11b are formed to the input waveguide 19a and the output waveguide 19b extending toward the end face of the semi-insulating InP substrate 12 is made semi-insulating by, for example, proton implantation. . By doing so, the n-type region and the p-type region can be separated with high resistance.

  Next, an SCH-multiple quantum well layer is grown on the entire surface of the InP layer on which the n-type InP cladding layer 131 and the p-type InP cladding layer 141 are formed. A p-type InP cladding layer 133 and a p-type contact layer 134 are formed in this order on the region (SCH-multiple quantum well layer 132) including the semiconductor optical waveguide 13 of the SCH-multiple quantum well layer. An n-type InP cladding layer 143 and an n-type contact layer 144 are formed in this order on the region (SCH-multiple quantum well layer 142) including the semiconductor optical waveguide 14 of the SCH-multiple quantum well layer. In order to perform different crystal growth in the two regions in this way, for example, a method similar to the method described in Patent Document 3 is used. That is, first, a p-type InP layer and a p-type contact layer are grown on the SCH-multi-quantum well layer in the order described above. Next, an etching mask is formed in a region other than the region including the semiconductor optical waveguide 14 of the SCH-multiple quantum well layer. In this state, selective etching is performed to remove the p-type InP layer and the p-type contact layer. Next, an n-type InP layer and an n-type contact layer are grown in the above order with the etching mask left. After this growth, both the semiconductor optical waveguides 13 and 14, 2 × 2-MMI 11a (demultiplexer) and 11b (multiplexer), and the portions other than the portions that become the input optical waveguide 19a and the output optical waveguide 19b are selected. The SCH-multiple quantum well layer is removed using an etchant. Further, additional etching is performed between the semiconductor optical waveguides 13 and 14 until reaching the semi-insulating InP substrate 12. As described above, the semiconductor optical waveguide forming step, the duplexer forming step, the multiplexer forming step, and the like can be performed simultaneously (collectively). The three steps may be performed separately. Further, after the first undoped InP layer is grown, for example, when the region corresponding to between the semiconductor optical waveguides 13 and 14 is made semi-insulating by proton implantation, this additional etching is unnecessary.

  Next, an Fe-doped or Ru-doped InP crystal is grown between the semiconductor optical waveguides 13 and 14 to form a semi-insulating InP buried layer 151a. At the same time, an Fe-doped or Ru-doped InP crystal is grown on the side surface of the semiconductor optical waveguide 13 opposite to the semiconductor optical waveguide 14 side (the side surface on the left side in FIG. 1B). A buried layer 151b is formed. At the same time, an Fe-doped or Ru-doped InP crystal is grown on the side surface of the semiconductor optical waveguide 14 opposite to the semiconductor optical waveguide 13 side (the side surface on the right side in FIG. 1B). A buried layer 151c is formed. Here, when the semi-insulating InP buried layers 151a, 151b, and 151c are formed, the p-type contact layer 134, the semi-insulating InP buried layer 151a on the semiconductor optical waveguide 13, the semi-insulating InP buried layer 151b, Over the n-type contact layer 144 and the semi-insulating InP buried layer 151c on the semiconductor optical waveguide 14, the uppermost portion of these layers is made flat.

  Next, portions of the p-type contact layer 134 on the semiconductor optical waveguide 13 and the p-type contact layer 144 on the semiconductor optical waveguide 14 other than the portion that contacts the signal electrode 15 are removed by etching. At this time, the p-type InP clad layer 133 and the n-type InP clad layer 143 exposed by etching are made semi-insulating, for example, by proton implantation. By doing so, for example, leakage of high frequency from the signal electrode 15 can be prevented.

  Next, the portion where the positive bias electrode 17 and the negative bias electrode 18 are formed is etched using a selective etchant up to the SCH-multiple quantum well layer. In this state, a dielectric film 161 is formed on the crystal. Of the dielectric film 161, only the contact portion with the signal electrode 15 on both the semiconductor optical waveguides 13 and 14, the portion for forming the positive bias electrode 17 and the portion for forming the negative bias electrode 18 are opened. In this state, an electrode is formed on the entire surface of the crystal. The formed electrodes are patterned to form the signal electrode 15, the ground electrodes 16a and 16b, the positive bias electrode 17 and the negative bias electrode 18. In this way, the signal electrode forming step, the positive bias electrode forming step, and the negative bias electrode forming step can be performed.

  Further, this crystal is cleaved and an antireflection coating is applied to the input / output end faces (left and right end faces in FIG. 1A). In this way, the semiconductor Mach-Zehnder type optical modulator of this embodiment is obtained. Note that the method of manufacturing the semiconductor Mach-Zehnder optical modulator of the present embodiment is not limited to this example.

  In the above-described manufacturing method, the semi-insulating InP buried layer described above is formed in portions other than both semiconductor optical waveguides, and the uppermost portions thereof are flattened, and the flattened uppermost portion is connected to the signal electrode 15. Adopting a contact process. For this reason, a coplanar transmission line composed of the signal electrode 15 and the ground electrodes 16a and 16b is formed on a flat surface. As a result, high-frequency scattering, loss, reflection, and the like can be sufficiently suppressed.

(Embodiment 2)
Next, another embodiment of the present invention will be described. The semiconductor Mach-Zehnder type optical modulator of this embodiment is a semiconductor Mach-Zehnder type optical modulator using an InP system as a constituent material.

  FIG. 2 shows a configuration of an example of the semiconductor Mach-Zehnder optical modulator according to this embodiment. In this figure, the same parts as those in FIG. As shown in FIG. 2, the semiconductor Mach-Zehnder optical modulator 20 includes semiconductor optical waveguides 23 and 24 instead of the semiconductor optical waveguides 13 and 14 in the semiconductor Mach-Zehnder optical modulator 10 described above. An n-type contact layer 236 is formed on the semiconductor optical waveguide 23 instead of the p-type contact layer 134. An n-type contact layer 246 is formed on the semiconductor optical waveguide 24 instead of the n-type contact layer 144. The signal electrode 15 is formed on the n-type contact layer 236 and the n-type contact layer 246. The positive bias electrode 17 is electrically connected to the lower part of the semiconductor optical waveguide 23. The negative bias electrode 18 is electrically connected to the lower part of the semiconductor optical waveguide 24.

  The semiconductor optical waveguides 23 and 24 are formed on the semi-insulating InP substrate 12. The structures of the semiconductor optical waveguides 23 and 24 in the vertical direction (the vertical direction in FIG. 2) are the same. The semiconductor waveguide 23 includes an n-type InP cladding layer 231, a semi-insulating InP cladding layer 232, an SCH-multiple quantum well layer (i-type) 233, a semi-insulating InP cladding layer 234, on the semi-insulating InP substrate 12. The n-type InP clad layer 235 has a vertical structure laminated in the above order. The SCH-multiple quantum well layer (i-type) 233 is a core layer of the semiconductor optical waveguide 23. The semiconductor waveguide 24 includes an n-type InP cladding layer 241, a semi-insulating InP cladding layer 242, an SCH-multiple quantum well layer (i-type) 243, a semi-insulating InP cladding layer 244, on the semi-insulating InP substrate 12. The n-type InP cladding layer 245 has a vertical structure laminated in the above-described order. The SCH-multiple quantum well layer (i-type) 243 is a core layer of the semiconductor optical waveguide 24. The positive bias electrode 17 is in contact with (connected to) the n-type InP cladding layer 231. The negative bias electrode 18 is in contact with (connected to) the n-type InP cladding layer 241. In both the semiconductor optical waveguides 23 and 24, the layer thickness of each layer other than both SCH-multiple quantum well layers (i-type) 233 and 243 is, for example, 1 ± 0.7 μm.

  For example, the n-type InP clad layer 231 and the n-type InP clad layer 241 are spaced apart by 3 ± 2 μm from the left and right parts of the semi-insulating InP substrate 12 (in FIG. ) So as to cover the upper surface of each. The semi-insulating InP cladding layer 232, the SCH-multiple quantum well layer (i-type) 233, the semi-insulating InP cladding layer 234, and the n-type InP cladding layer 235 are formed on the n-type InP cladding layer 231 on the n-type InP cladding layer 231. The semiconductor optical waveguide 23 is formed together with the n-type InP clad layer 231 by being laminated on the edge on the side facing 241 (the right side in FIG. 1B). The semi-insulating InP cladding layer 242, the SCH-multiple quantum well layer (i-type) 243, the semi-insulating InP cladding layer 244, and the n-type InP cladding layer 245 are n-type InP cladding layers on the n-type InP cladding layer 241. The semiconductor optical waveguide 24 is formed together with the n-type InP clad layer 241 by being laminated on the edge on the side facing the H.231 (left side in FIG. 1B). As described above, the n-type contact layer 236 is formed on the upper surface of the n-type InP cladding layer 235 (on the semiconductor optical waveguide 23). As described above, the n-type contact layer 246 is formed on the upper surface of the n-type InP cladding layer 245 (on the semiconductor optical waveguide 24). As described above, the signal electrode 15 is formed on the n-type contact layer 236 and the n-type contact layer 246. The signal electrode 15 is a traveling wave electrode. The signal electrode 15 is formed in contact with the upper surfaces of the n-type contact layer 236 and the n-type contact layer 246, thereby being connected to the upper portions of the two semiconductor optical waveguides 23 and 24, and the semiconductor optical waveguides 23 and 24. A common high-frequency signal can be input.

  Other configurations are the same as those of the semiconductor Mach-Zehnder optical modulator 10 described above.

  As described above, the semiconductor optical waveguide 23 and the semiconductor optical waveguide 24 have the same structure. However, since the vertical structures of both the semiconductor optical waveguides 23 and 24 are n-type / semi-insulating type / i-type / semi-insulating type / n-type as described above, the current can be obtained even if the polarities are not reversed. Does not flow and operates as a phase modulator. As a result, the semiconductor Mach-Zehnder optical modulator of this embodiment can obtain the same effects as those of the semiconductor Mach-Zehnder optical modulator 10 described above. Further, according to the structure of the semiconductor optical waveguides 23 and 24, no current flows even if the direction of the bias is reversed, and the semiconductor optical waveguides 23 and 24 operate as a phase modulator. That is, although the electrode 17 has been described as a positive bias electrode, in the present embodiment, a negative bias may be applied thereto and used as a negative bias electrode. Similarly, although the electrode 18 has been described as a negative bias electrode, in the present embodiment, a positive bias may be applied thereto and used as a positive bias electrode. Thus, according to the semiconductor optical waveguide structure of the present embodiment, the semiconductor Mach-Zehnder optical modulator of the present invention has an effect that the degree of freedom in the bias application direction is increased.

  In the semiconductor Mach-Zehnder type optical modulator of the present invention, for example, as in the present embodiment, any semiconductor optical waveguide is vertically n-type / semi-insulating type / i from the negative bias side toward the positive bias side. It is preferable to have a stacked structure in the order of type / semi-insulating type / n type. Further, for example, at least one of the semi-insulating semiconductor layers in the laminated structure of the order of n-type / semi-insulating type / i-type / semi-insulating type / n-type is replaced with a p-type semiconductor layer. Also good. That is, the vertical structure of the semiconductor optical waveguide is, for example, a laminated structure in the order of n-type / p-type / i-type / p-type / n-type, n-type / semi-insulating type / i-type / p-type / n. It may be a laminated structure in the order of molds or a laminated structure in the order of n-type / p-type / i-type / semi-insulating type / n-type. According to these semiconductor optical waveguide structures, as described above, the degree of freedom in the bias application direction is increased, and as described later, the semiconductor optical waveguide structure can be manufactured easily and at low cost.

  In the semiconductor Mach-Zehnder optical modulator of this embodiment, the upper and lower layer structures of both SCH-multiple quantum well layers (i-type) 233 and 243 are the same as described above. Therefore, for example, crystal growth from the n-type InP layer to the n-type contact layer on the semi-insulating InP substrate can be performed at once. Except for this point, the semiconductor Mach-Zehnder optical modulator of this embodiment can be manufactured in the same manner as the manufacturing method described in the first embodiment. As described above, the semiconductor Mach-Zehnder type optical modulator of this embodiment is described in, for example, Patent Document 3 as compared with the case where the vertical structure (conductivity type structure) of both semiconductor optical waveguides is inverted. The manufacturing process of the entire surface growth, selective etching, and selective growth can greatly reduce the manufacturing process and can be manufactured at a lower cost. However, the method of manufacturing the semiconductor Mach-Zehnder optical modulator of the present embodiment is not limited to this example.

  In the manufacturing method of the present invention, as shown in the present embodiment, for example, in the semiconductor optical waveguide forming step, the semiconductor layer is n-type / semi-insulating / i-type / semi-insulating from bottom to top. N-type / n-type, n-type / semi-insulating / i-type / p-type / n-type, n-type / p-type / i-type / semi-insulating / n-type, or n-type / It is preferable to laminate in the order of p-type / i-type / p-type / n-type, and simultaneously form the two semiconductor optical waveguides having these laminated structures. In this way, as described above, the semiconductor Mach-Zehnder optical modulator of the present invention can be manufactured more simply and at low cost.

  The semiconductor Mach-Zehnder optical modulator of the present invention is preferably used for, for example, differential quadrature phase shift keying (DQPSK) as described below. However, the following description does not limit or limit the present invention.

  In an ultra-high-speed, large-capacity, long-distance optical communication backbone system, for example, multi-leveling is considered effective for further improving the wavelength utilization efficiency. When multi-level processing is performed, the bit rate is symbol rate × multi-level. Specifically, for example, according to a differential quadrature phase shift keying (DQPSK) with a symbol rate of 25 Gsps (symbols per second, the number of symbols propagated per second), 50 Gbps (bits per second). A bit rate of the number of bits propagated per second) is realized, and the transmission capacity per wavelength is improved. That is, in the case of the above-described BPSK, the bit rate = symbol rate, but in the case of the above-described DQPSK (multilevel: 2), the bit rate = symbol rate × 2. In the DQPSK, light is orthogonally decomposed into two signals I and Q whose phases are different by 90 °. It can be considered that the signal I and the signal Q are independently performing the above-described BPSK communication.

  In the DQPSK, when the Mach-Zehnder type optical modulator that applies the above-described two anti-phase high-frequency signals is used, in order to perform push-pull driving, a total of four high-frequency signals I, -I, Q, and -Q are used. Are input to two Mach-Zehnder optical modulators. Here, considering that the DQPSK modulator is integrated on, for example, one InP substrate, at least two Mach-Zehnder optical modulators are arranged in parallel. In the case where two DQPSK modulations are integrated in parallel, S-shaped optical waveguides for demultiplexing and multiplexing are required for the two Mach-Zehnder type optical modulators. The distance between the two optical waveguides that must be separated by the S-shaped optical waveguide is as wide as about 200 μm, for example. In order to take this interval, the length of the S-shaped optical waveguide needs to be about 400 μm, for example, if it is gently bent so that the radiation loss can be ignored. After all, when two Mach-Zehnder modulators are arranged in parallel, for example, a total length of about 6 mm is required. Considering a phase adjustment electrode for orthogonalizing two BPSK-modulated lights, a Mach-Zehnder type optical modulator for RZ (Return to Zero), etc., an InP optical semiconductor device has an unusually large chip. It becomes size. Thus, as the Mach-Zehnder type optical modulator is integrated, the area of the S-shaped optical waveguide becomes enormous and the S-shaped optical waveguide becomes excessively long. For this reason, the chip size is increased, and for example, the chip cost is increased.

  Further, when the optical waveguide is extended to this extent, for example, insertion loss becomes a problem. Furthermore, there is a problem in manufacturing the S-shaped optical waveguide itself. For example, when the optical waveguide is an embedded waveguide made of InP, if the optical waveguide is inclined obliquely from the crystal orientation, crystal fogging may occur during embedded growth. This causes voids and may cause scattering loss of guided light. For example, in the case of a ridge waveguide, when the mesa is formed with a wet selective etchant, the mesa side surface is inclined from 90 ° according to the oblique angle of the optical waveguide. This complicates the optical waveguide design. In any case, in a large-scale S-shaped optical waveguide, there are more opportunities to cause loss, and it is difficult to stably achieve low insertion loss.

  In addition, in the DQPSK modulator integration, when the Mach-Zehnder type optical modulator that applies the above-described two anti-phase high-frequency signals is used, four high-frequency transmission lines must be wired. For this reason, the wiring is crowded. As a result, it is difficult to consider that there is no time delay while suppressing crosstalk between the signals. Also, for example, at an ultrahigh frequency such as 25 Gsps, if the coplanar line is routed long, the loss or reflection of the electric signal cannot be ignored. Further, for example, at an ultra-high frequency such as 25 Gsps, even if there is a slight variation in the wiring length of four high-frequency transmission lines for inputting each high-frequency signal (an error in the execution length of the transmission line), the high-frequency signals may be between each high-frequency signal. There will be a time delay that cannot be ignored. As a result, V1 + V2 deviates from the constant condition, and an error from accurate push-pull driving occurs. At this time, wavelength chirping occurs and a transmission error occurs.

  That is, when integrating a Mach-Zehnder type optical modulator, particularly an InP-based Mach-Zehnder type optical modulator, for example, the following problems occur. First, since the area of the S-shaped optical waveguide becomes enormous as the integration is increased, the chip size is increased and the chip cost is increased. Second, the more integrated, the longer the S-shaped waveguide becomes, and the larger the insertion loss due to the radiation loss and internal loss of the S-shaped optical waveguide. Thirdly, the more integrated, the longer the transmission line becomes, and the transmission line gets crowded. For this reason, errors from accurate push-pull driving occur due to errors in the effective length of the transmission line, crosstalk, high-frequency loss and reflection.

  If the semiconductor Mach-Zehnder type optical modulator of the present invention is applied to the Mach-Zehnder type optical modulator in the above-described DQPSK, first, in each Mach-Zehnder type optical modulator, the two optical waveguides are brought close to each other as described above, and the S-shaped optical Waveguides can be omitted or shortened. For this reason, each Mach-Zehnder type optical modulator can be made into a Mach-Zehnder type optical modulator that is small, low loss, and accurately push-pull driven. Furthermore, since the individual Mach-Zehnder type optical modulators are miniaturized, the S-shaped optical waveguide for connecting them can be omitted or shortened. As described above, the present invention not only contributes to the miniaturization of one Mach-Zehnder type optical modulator, but also reduces the size of one Mach-Zehnder type optical modulator, thereby reducing the S at the time of integration in parallel. The letter-shaped optical waveguide can also be reduced in size. Thus, according to the present invention, the greater the integration of the Mach-Zehnder type optical modulator, the greater the effect that the contribution will spread. As a result, for example, high integration of a Mach-Zehnder type optical modulator is facilitated.

  When the Mach-Zehnder optical modulator that applies the two high-frequency signals of the opposite phases is used as the Mach-Zehnder optical modulator in the DQPSK, in order to perform push-pull driving, as described above, I, − A total of four high-frequency signals I, Q, and -Q are input to the two Mach-Zehnder optical modulators. For this reason, in this case, four high-frequency transmission lines must be wired. If the semiconductor Mach-Zehnder type optical modulator of the present invention is applied to the above-described Mach-Zehnder type optical modulator in DQPSK, the high-frequency signal necessary for performing DQPSK push-pull drive is reduced to, for example, I and Q. can do. For this reason, if the semiconductor Mach-Zehnder type optical modulator of the present invention is applied to the Mach-Zehnder type optical modulator in DQPSK, the number of high-frequency transmission lines can be reduced to two. As a result, for example, high integration of a Mach-Zehnder type optical modulator is facilitated, and accurate push-pull driving is possible.

  The semiconductor Mach-Zehnder type optical modulator of the present invention is preferably used for an InP-based Mach-Zehnder type optical modulator, for example, as follows. However, the following description does not limit or limit the present invention.

Lithium niobate (LiNbO ₃ ) is mainly used as the material system of the Mach-Zehnder type optical modulator used in the transmitter of the latest ultrahigh-speed, large-capacity long-distance optical communication backbone system. On the other hand, InP-based Mach-Zehnder optical modulators are considered promising for miniaturization, but are not currently used in the highest-level communication system.

The compound semiconductor optical waveguide on the InP substrate has a lower coupling efficiency with the optical fiber than the LiNbO ₃ optical waveguide. By increasing the degree of integration of the InP system, this problem can be minimized and the advantage that the InP system is small can be maximized. Thereby, it is conceivable to counter LiNbO ₃ . However, although the InP-based Mach-Zehnder type optical modulator is small, the unit price (chip cost) per area is higher than that of LiNbO ₃ . For this reason, considering high integration, it is desired that the InP-based Mach-Zehnder type optical modulator be further miniaturized as much as possible.

If the present invention is applied to an InP-based Mach-Zehnder type optical modulator, the size can be further reduced, so that it is possible to compete with LiNbO ₃ in terms of cost.

  As described above, the semiconductor Mach-Zehnder type optical modulator of the present invention is small in size and low in loss and can be accurately driven by push-pull. Accordingly, the application of the semiconductor Mach-Zehnder type optical modulator of the present invention includes, for example, an optical modulator of an optical communication transmitter used for a trunk line system, a metro system, an access system, etc., and its application is limited. It can be applied to a wide range of fields.

10, 20 Semiconductor Mach-Zehnder optical modulator 11a 2 × 2-MMI (demultiplexer)
11b 2 × 2-MMI (Multiplexer)
12 Semi-insulating InP substrates 13, 14, 23, 24 Semiconductor optical waveguide 15 Signal electrodes 16a, 16b Ground electrode 17 Positive bias electrode 18 Negative bias electrode 19a Input optical waveguide 19b Output optical waveguide 30 Conventional Mach-Zehnder optical modulator 31a, 31b 2 × 2-MMI
33, 34, 39a, 39b Optical waveguides 35a, 35b, Signal electrode 36 Ground electrodes 131, 143, 231, 235, 241, 245 n-type InP cladding layers 132, 142, 233, 243 SCH-multiple quantum well layers (i-type) )
133, 141 p-type InP clad layer 134 p-type contact layers 144, 236, 246 n-type contact layers 151a, 151b, 151c Semi-insulating InP buried layer 161 Dielectric films 232, 234, 242, 244 Semi-insulating InP cladding layers

Claims (9)

A duplexer, a multiplexer, a semiconductor optical waveguide, a signal electrode, a positive bias electrode, and a negative bias electrode;
The semiconductor optical waveguide is two,
The duplexer and the multiplexer are connected by the two semiconductor optical waveguides,
The signal electrode is connected to the upper part of the two semiconductor optical waveguides, and can input a high frequency signal common to the two semiconductor optical waveguides,
The positive bias electrode is connected to a lower portion of one of the two semiconductor optical waveguides, and the negative bias electrode is connected to a lower portion of the other semiconductor optical waveguide. Mach-Zehnder type optical modulator. 2. The semiconductor Mach-Zehnder type optical modulator according to claim 1, wherein the conductive type structures in the vertical direction of the two semiconductor optical waveguides are vertically inverted from each other. The two semiconductor optical waveguides are respectively in the order of p-type / i-type / n-type and n-type / p-type / i-type / n-type in the vertical direction from the negative bias side to the positive bias side. And a stacked structure in the order of n type / semi-insulating type / i type / n type or n type / semi-insulating type / i type / semi-insulating type / n type. 3. A semiconductor Mach-Zehnder optical modulator according to 1 or 2. The two semiconductor optical waveguides are stacked in the order of n-type / semi-insulating type / i-type / semi-insulating type / n-type in the vertical direction from the negative bias side to the positive bias side. The semiconductor Mach-Zehnder optical modulator according to claim 3, wherein At least one of the semi-insulating semiconductor layers in the stacked structure of the n-type / semi-insulating type / i-type / semi-insulating type / n-type is replaced with a p-type semiconductor layer. The semiconductor Mach-Zehnder type optical modulator according to claim 3 or 4. 6. The semiconductor Mach-Zehnder optical modulator according to claim 1, wherein the signal electrode is a traveling wave electrode. A duplexer forming process for forming the duplexer;
A multiplexer forming step for forming a multiplexer;
A semiconductor optical waveguide forming step of forming two semiconductor optical waveguides connecting the duplexer and the multiplexer;
Forming a signal electrode connected to the top of the two semiconductor optical waveguides and capable of inputting a common high-frequency signal to the two semiconductor optical waveguides;
A positive bias electrode forming step of forming a positive bias electrode connected to a lower portion of one of the two semiconductor optical waveguides;
And a negative bias electrode forming step of forming a negative bias electrode connected to a lower portion of the other semiconductor optical waveguide. A method of manufacturing a semiconductor Mach-Zehnder optical modulator, comprising: In the semiconductor optical waveguide forming step, the semiconductor layers are arranged in the order of n-type / semi-insulating type / i-type / semi-insulating type / n-type from bottom to top, n-type / semi-insulating type / i-type. / P-type / n-type, n-type / p-type / i-type / semi-insulating / n-type, or n-type / p-type / i-type / p-type / n-type The manufacturing method according to claim 7, wherein the two semiconductor optical waveguides having the laminated structure are simultaneously formed. 9. The manufacturing method according to claim 7, wherein the duplexer forming step, the multiplexer forming step, and the semiconductor optical waveguide forming step are simultaneously performed.

Images