JP5655412B2 - Semiconductor optical modulator, manufacturing method thereof, and optical communication module - Google Patents

Description

  The present invention relates to a semiconductor optical modulator, a manufacturing method thereof, and an optical communication module.

  A Mach-Zehnder (MZ) type optical modulator is an optical intensity modulator that performs on / off modulation of light according to interference conditions when the light is once split into two and then multiplexed again. By applying a modulation voltage to the electrode provided on the waveguide, the interference condition for multiplexing is changed.

  Please refer to FIG. For example, the demultiplexing unit 1, the multiplexing unit 2, and two semiconductor optical waveguides 3 a and 3 b that connect between these are formed on the surface of the semiconductor substrate. Electrodes 4a and 4b are formed on the optical waveguides 3a and 3b. The optical input input to the duplexer 1 is branched into two by the duplexer 3, transmitted through the first optical waveguide 3 a and the second optical waveguide 3 b, and incident on the multiplexing unit 2. The multiplexer 2 combines the light provided from the first optical waveguide 3a and the second optical waveguide 3b to form an optical output.

  A first modulation electrode 4a and a second modulation electrode 4b are disposed on the first optical waveguide 3a and the second optical waveguide 3b, respectively, and modulation voltages Vr1 and Vr2 are applied thereto. In the case of a semiconductor MZ modulator, a waveguide layer and lower and upper optical confinement layers (sometimes collectively referred to as a core layer) are formed by an i-type layer, and an electric field is efficiently applied to the core layer. A structure in which the upper and lower sides of the core layer are sandwiched between p-type or n-type semiconductor layers doped with a high concentration of impurities is used. The refractive index change characteristics when an electric field is applied vary greatly depending on the semiconductor crystal material (mixed crystal composition ratio), the presence or absence of quantum effects, and the like. In the case of the pin type stacked structure, the bias voltages Vb1 and Vb2 are added to the modulation voltages Vr1 and Vr2 so that the pin junction becomes reverse bias. A common electrode is formed on the back surface of the substrate and grounded.

  The light beams incident on the first optical waveguide 3a and the second optical waveguide 3b are phase-modulated by changes in the refractive index of the waveguide that occur in response to the voltages applied to the first modulation electrode 4a and the second modulation electrode 4b, respectively. receive. The light beams emitted from the first optical waveguide 3 a and the second optical waveguide 3 b are combined by the optical multiplexer 5. The intensity of the emitted light changes according to the phase difference between the light beams generated by the modulation voltage applied to the first modulation electrode 4a and the second modulation electrode 4b.

  A general silica-based optical fiber has a minimum optical transmission loss near the wavelength of 1.5 μm in the near infrared region. For this reason, 1.5 μm band light is often used for long-distance communication. Many of the single-mode optical fibers currently used in the trunk line system have positive chromatic dispersion in the 1.5 μm band, and the waveform deteriorates when transmitting an optical signal with small frequency fluctuation (chirp). To compensate for waveform degradation, pre-chirp transmission that gives the modulated light a negative chirp whose wavelength changes to the high frequency side when the light intensity increases, and conversely, when the light intensity decreases Is used.

  By appropriately making the amount of phase change given to both optical waveguides of the interferometer asymmetric (different values), a suitable negative chirp can be imparted. In a semiconductor optical modulator, in order to introduce an appropriate asymmetry in the phase change amount of both optical waveguides, 1) the amplitude of the modulation voltage applied to the electrodes on both optical waveguides is different. 2) on both optical waveguides. Different means such as making the DC bias voltage applied to the electrodes different and 3) making the electrode lengths on both optical waveguides different are conceivable. The means of 2) and 3) alone have a small degree of freedom. Further, in the case of means 3), if the electrode lengths are simply changed, the high frequency characteristics of both electrodes are different, the eye opening ratio of the modulated light waveform is deteriorated, and the transmission characteristics are deteriorated. The means 1) can be realized simply by using a high-frequency attenuator, but it leads to an increase in the size of the apparatus and an increase in power consumption.

  Japanese Unexamined Patent Application Publication No. 2009-86111 discloses a thickness of a core layer (lower optical confinement layer + multiple quantum well layer + upper optical confinement layer) to which an electric field of a first modulation signal is applied in a first waveguide, Negative chirp by setting the thickness of the core layer (lower optical confinement layer + multiple quantum well layer + upper optical confinement layer) to which the electric field of the second modulation signal is applied in the waveguide is set to a different thickness. Propose to give properties. By varying the thickness of the core layer, the electric field strength in the waveguide can be varied even when a modulation voltage having the same amplitude is applied.

JP 2009-86111 A

  One object of the embodiment is to provide a semiconductor optical modulator having a novel configuration capable of providing a pre-chirp.

According to one aspect of the embodiment,
A demultiplexing unit formed on a semiconductor substrate and capable of branching input light into two, a first optical waveguide and a second optical waveguide transmitting two lights emitted from the demultiplexing unit, the first optical waveguide, the first optical waveguide, A Mach-Zehnder type optical waveguide structure having a multiplexing portion that receives light propagating through two optical waveguides and emits combined output light;
A first modulation electrode and a second modulation electrode, which are respectively formed above the first optical waveguide and the second optical waveguide and apply a first modulation signal and a second modulation signal;
Have
The first optical waveguide includes a first i-type semiconductor layer including a first waveguide layer, and a first pair of doped semiconductor layers that are disposed above and below the first i-type semiconductor layer and doped with impurities at a high concentration. And
The second optical waveguide includes a second i-type semiconductor layer including a second waveguide layer, and a second pair of doped semiconductor layers disposed on the upper and lower sides of the second i-type semiconductor layer and doped with impurities at a high concentration. And
The first i-type semiconductor layer and the second i-type semiconductor layer are equal in thickness in terms of high frequency characteristics,
The first modulation electrode and the second modulation electrode have equal lengths in terms of high frequency characteristics,
Said first modulation electrode, a first optical waveguide of said second modulation electrode downward, one of the second optical waveguide, has a butt joint joining mutually butting in the middle in the extending direction different waveguide layers of structures, while Does not have a butt joint,
A semiconductor optical modulator is provided.

  Appropriate and different phase change amounts can be given to the modulated light transmitted through both optical waveguides.

  It is not necessary to make the surface height of both optical waveguides different, and the manufacturing process after the waveguide formation can be stabilized.

  There is no need to use a high frequency attenuator.

It is a top view which shows roughly the structure of a semiconductor Mach-Zehnder modulator. When, 2A to 2G are plan views and cross sections schematically showing the configuration of the semiconductor optical modulator according to the first embodiment in which the first waveguide does not have a butt joint and the second waveguide has a butt joint. FIG. 3A and 3B are plan views schematically showing a lumped constant electrode and a traveling wave electrode. 4A and 4B are plan views schematically showing a configuration of a semiconductor optical modulator according to a second embodiment in which both the first and second waveguides have butt joints. When, When, When, When, FIGS. 5A to 5S are plan views and cross-sectional views for explaining main processes of a method for manufacturing a semiconductor optical modulator, taking one mode of the first embodiment as an example. 6A to 6H are plan views showing various forms of the first embodiment and the second embodiment.

  FIG. 2A is a plan view schematically showing the semiconductor modulator according to the first embodiment. The demultiplexing unit 1 and the multiplexing unit 2 have the same configuration as the semiconductor modulator shown in FIG. The modulation electrodes 4a and 4b also extend in the horizontal direction with the horizontal position aligned in parallel. The first optical waveguide 3a and the second optical waveguide 3b below the modulation electrodes 4a and 4b have different configurations. The first optical waveguide 3a does not have a butt joint, and the second waveguide 3b has a butt joint. Hereinafter, the configuration of the optical waveguide will be described. First, the configuration of the first optical waveguide 3a having no butt joint will be described.

  FIG. 2B shows a bulk waveguide in which an i-type layer sandwiched between an n-type region 31 and a p-type region 32 has a waveguide layer 33 and optical confinement layers 34 and 35 disposed on the lower and upper sides thereof. A case of 36 is shown. Since the waveguide layer 33 and the optical confinement layers 34 and 35 have physical properties in the bulk state of the constituent materials, this configuration is referred to as a bulk type. The basic properties of the waveguide are determined by the material of the waveguide layer 33. The optical confinement layers 34 and 35 are selected from materials having a smaller band gap and a lower refractive index than the waveguide layer 33 so as to have an optical confinement effect. By selecting materials for the waveguide layer and the optical confinement layer, the characteristics of the waveguide can be variously changed.

  A p-side electrode 41 is formed on the surface of the p-type region 32, and an n-side electrode 42 is formed on the lower surface of the n-type region 31. A case where the demultiplexing unit 1 and the multiplexing unit 2 are configured by multimode interference (MMI) couplers is shown. In the figure, the horizontal length is shown short, but the width (for example, about 10 μm) and the length (for example, about 400 μm) of the MMI coupler are selected so that the branching ratio is 1: 1.

  2C shows an MQW in which an i-type layer sandwiched between an n-type region 31 and a p-type region 32 is composed of optical confinement layers 34 and 35 and a multiple quantum well (MQW) layer 37 sandwiched therebetween. A mold waveguide 38 is shown. The multi-quantum well layer 37 has a configuration in which barrier layers B and well layers W are alternately stacked, and each well layer W is sandwiched between barrier layers B, and a stacked structure of n well layers and (n + 1) barrier layers. It is. In a multiple quantum well layer, a well layer and a barrier layer are formed with dimensions that produce a quantum effect, and exhibit properties different from bulk physical properties. By selecting the material and thickness of the well layer and barrier layer, the characteristics of the multiple quantum well can be variously changed.

  The point that the light confinement layer layers 34 and 35 have a light confinement effect is the same as the configuration of FIG. 2B. Note that. The waveguide layer of the bulk waveguide and the multiple quantum well layer of the MQW waveguide may be collectively referred to as a waveguide layer. Next, the configuration of the second optical waveguide 3b having a butt joint will be described.

  FIG. 2D shows a configuration of the second optical waveguide 3b having a butt joint in which waveguide layers having different configurations are butt-coupled (butt joint). An optical waveguide having a butt joint that is a discontinuous change in configuration may be referred to as a butt joint type optical waveguide. As an example, a case where the bulk waveguide 36 and the MQW waveguide 38 are butt-jointed is shown. The bulk type waveguide 36 and the MQW type waveguide 38 are basically the same as those shown in FIGS. 2B and 2C. The characteristics of the entire waveguide are variously changed depending on the combination.

  In addition, although the case where the waveguide layer 33 of the bulk type waveguide 36 and the MQW layer 37 of the MQW type waveguide 38 have the same thickness is shown, the thickness may not be the same.

  FIG. 2E shows an example in which the thickness of the waveguide layer changes in the second optical waveguide 3b. In the butt joint type waveguide 3 b, the thickness of the MQW layer 38 is set to be larger than the thickness of the waveguide layer 33. However, the optical confinement layer 35 is adjusted to be thick on the waveguide layer 33 and thin on the MQW 37, and the thickness of the entire i-type layer is made equal.

  FIG. 2F shows an example in which a butt joint is formed between bulk-type waveguide layers in the second optical waveguide 3b. In the drawing, the left waveguide layer 33x and the right waveguide layer 33y have different compositions and constitute a butt joint. The optical confinement layers 34 and 35 may have the same composition or different compositions.

  FIG. 2G shows an example in which a butt joint is formed between MQW layers. The combination of the well layer and barrier layer of the left MQW layer 37x in the drawing and the combination of the well layer and barrier layer of the right MQW layer 37y differ in at least one of composition and thickness, and constitute a butt joint. The optical confinement layers 34 and 35 may have the same composition or different compositions.

  Although the case where the first waveguide 3a has no butt joint and the second waveguide 3b has a butt joint has been described, the reverse relationship may be used.

  The two waveguide layers forming the butt joint are selected from materials having different interband transition wavelengths, for example. When a composition with an interband transition wavelength of 1300 nm and a composition with an interband transition wavelength of 1400 nm are selected, the composition material that is generally less detuned with respect to the laser oscillation wavelength of 1550 nm has a phase change when an electric field is applied. It becomes relatively large. Even if the electrode lengths above the two waveguides are equal and the amplitude of the applied voltage is equal, an asymmetric phase change is obtained and a negative chirp operation is obtained.

  A decrease in detuning generally results in an increase in light absorption loss and a decrease in on-level output. From this point of view, when a composition with small detuning is used for the waveguide portion that generates the modulation, in the waveguide region other than the region that generates the modulation (the waveguide region outside the electrode, the demultiplexing unit, the multiplexing unit) It is preferable not to use a material with small detuning.

  3A and 3B are schematic plan views showing forms of the modulation electrode and the modulation circuit. FIG. 3A shows a lumped constant electrode. The electrodes 4a and 4b are designed as a lumped constant type, and the wiring from the modulation circuit is connected to one place of the electrode. The modulation circuit includes a high frequency modulation voltage source Vr and a bias voltage source Vb. FIG. 3B shows a traveling wave electrode. A wire from the modulation circuit is connected to one end of the electrodes 4a and 4b, and the other end is grounded via a load. In this mode, the high-frequency signal travels through the electrodes 4a and 4b. The bias voltage source Vb may be connected to the termination side.

  The total thickness of the i-type layer is set so as to be considered the same from the viewpoint of high-frequency characteristics. The high frequency characteristics of the electrode are substantially determined by the thickness and area of the i-type layer below the electrode, and do not depend on the composition of the i-type layer. Therefore, if the thickness of the i-type layer is made equal in both waveguides, the high frequency characteristics of the electrode can be easily controlled. It is from this point of view that the thicknesses of the i-type layers are equal, and the physically exact same thickness is not necessarily required. This is called the same or equal thickness in terms of high frequency characteristics. When a high-frequency modulation signal having the same amplitude is applied by push-pull from the modulation circuit, the i layers of the two waveguides have the same thickness, so that the same electric field strength is generated. The two waveguides have different configurations, and the lengths of the electrodes above the two waveguides are preferably set to be the same from the viewpoint of high frequency characteristics even for the same electric field. The lengths may be regarded as equal in terms of high frequency characteristics, and physically identical lengths are not required. 2A to 2G, the electrodes above the two optical waveguides are arranged at equivalent or symmetrical positions. From the viewpoint of light modulation, the modulation electrodes above the two optical waveguides do not necessarily have to be arranged at equivalent positions. In the first embodiment, one optical waveguide does not have a butt joint, and the other optical waveguide has a butt joint.

  4A and 4B show a semiconductor optical modulator according to a second embodiment in which both optical waveguides have butt joints. In order to realize negative chirp, a butt joint type optical waveguide having a different configuration is formed below the modulation electrode. It is assumed that two optical waveguides 3a and 3b extending in parallel with the start point and the end point being in equivalent positions.

  In FIG. 4A, the modulation electrodes 4a and 4b are arranged above the optical waveguides 3a and 3b so that the positions in the extending direction are aligned. The optical waveguides 3a and 3b include first optical waveguide portions 3-1a and 3-1b having the same cross-sectional structure, and second optical waveguide portions 3-2a and 3-2b having the same cross-sectional structure. In the optical waveguide 3a, the first optical waveguide portion 3-1a and the second optical waveguide portion 3-2a form a butt joint in the right region below the modulation electrode 4a. In the optical waveguide 3b, the first optical waveguide portion 3-1b and the second optical waveguide portion 3-2b form a butt joint in the left region below the modulation electrode 4b. Butt joints are formed at different positions below the two modulation electrodes to enable pre-chirping.

  In FIG. 4B, the butt joints are formed at different positions below the modulation electrodes 4a and 4b as in FIG. 4A by aligning the lateral positions of the butt joints and making the horizontal positions of the modulation electrodes 4a and 4b different. . Since the lateral position of the butt joint at the semiconductor substrate level is aligned, the etching process and the subsequent crystal growth process in the manufacturing process can be simplified and stabilized.

  In FIGS. 4A and 4B, the two waveguide layers forming the butt joint may be bulk waveguide layers, MQW waveguide layers, or a bulk waveguide layer-MQW waveguide layer. Hereinafter, in the configuration of FIG. 4B, taking the case where the first optical waveguide unit 3-1 is a bulk type waveguide and the second optical waveguide unit 3-2 is an MQW type waveguide as an example, the manufacturing process will be described. explain.

  5A to 5S are plan views and cross-sectional views illustrating a manufacturing process of the semiconductor optical modulator, taking the form shown in FIG. 4B as an example.

  FIG. 5A is a plan view schematically showing the configuration of the semiconductor optical modulator. The laser diode active layer region 100 is driven by a forward current supplied from the electrode 102 and emits a 1.55 μm band laser beam. The InGaAsP waveguide layer region 110 continuing to the laser diode active layer region 100 constitutes a demultiplexing unit, demultiplexes the laser light emitted from the laser diode active layer region 100 into two lights, and guides the waveguides WG1, WG2. To supply. The MQW waveguide layer region 120 is continuous with the InGaAsP waveguide layer region 110 via a butt joint, and forms waveguides WG1 and WG2. Above the waveguide WG1, an electrode 130 is disposed on a region including the butt joint, and a modulation signal from the modulation circuit 135 is applied. Above the waveguide WG2, an electrode 140 is disposed above the MQW waveguide region 120 that does not include a butt joint, and a modulation signal from the modulation circuit 145 is applied. The optical signals propagated through the waveguides WG1 and WG2 are multiplexed at the multiplexing section formed by the InGaAs waveguide layer region 110 and output.

  The butt joints of the waveguides WG1 and WG2 are arranged with their lateral positions aligned, and the electrodes 130 and 140 are arranged with their lateral positions shifted (asymmetrically). Hereinafter, a manufacturing process having the configuration shown in FIG. 5A will be described.

  As shown in FIG. 5B, an n-type InGaAsP layer 12 is grown to a thickness of 50 nm on an n-type InP substrate 11 by metal organic chemical vapor deposition (MOCVD). As shown in FIG. 5C, a diffraction grating G suitable for obtaining laser oscillation with a wavelength of 1550 nm is patterned on the InGaAsP layer 12 by dry etching using an electron beam resist pattern.

  As shown in FIG. 5D, the diffraction grating G is embedded to form a laser active layer. First, an n-type InP cladding layer 13 is grown on a flat portion by 80 nm in thickness by MOCVD to embed a diffraction grating, and an i-type InP lower optical confinement layer 14 having a thickness of 100 nm, multiple quantum well layers MQW1, An i-type layer including an i-type InP upper optical confinement layer 17 having a thickness of 100 nm is grown by MOCVD. The multiple quantum well layer MQW is formed by repeating an InGaAsP barrier layer having a thickness of 10 nm and an InGaAsP well layer having a thickness of 5 nm, includes 7 barrier layers and 6 well layers, and has a total thickness of 100 nm. The total thickness of the i-type layer is 300 nm. The composition of the InGaAsP mixed crystal is selected so that the interband transition wavelengths of the well layer and the barrier layer are 1550 nm and 1200 nm, respectively. On the i-type InP upper optical confinement layer 17, a p-type InP cladding layer 18 is grown by MOCVD with a thickness of 150 nm.

As shown in FIG. 5E, a silicon dioxide (SiO ₂ ) layer 19 is formed on the p-type InP cladding layer 18 by CVD, and a mask covering the laser active layer is patterned. Wet etching is performed using the SiO ₂ layer 19 as a mask to remove the semiconductor layer up to the upper surface of the n-type InP layer 13 and leave the active layer only in the region to be the laser.

As shown in FIG. 5F, the bulk waveguide layer is grown with the SiO ₂ mask 19 left. An i-type layer having a total thickness of 350 nm including an i-type InP light confinement layer 21 having a thickness of 80 nm, an InGaAsP waveguide layer 22 having a thickness of 150 nm, and an i-type InP light confinement layer 23 having a thickness of 120 nm is grown by MOCVD. The interband transition wavelength of the InGaAsP waveguide layer 22 is 1300 nm. A p-type InP cladding layer 24 having a thickness of 100 nm is grown on the i-type layer by MOCVD.

Crystal growth does not occur on the SiO ₂ mask 19. The upper surface of the p-type InP cladding layer 18 and the upper surface of the p-type InP cladding layer 24 are flush with each other, facilitating subsequent processes. The waveguide upper surfaces of the i-type layers 21, 22, and 23 need not coincide with the waveguide upper surfaces of the i-type layers 14 and MQWs 1, 17. Thereafter, the SiO ₂ mask 19 is removed by wet etching.

As shown in FIG. 5G, a new SiO ₂ layer 20 is formed by CVD and patterned to form a SiO ₂ mask 20 that covers the laser part, the demultiplexing part, the bulk waveguide part, and the multiplexing part. Using the SiO ₂ mask 20 as an etching mask, wet etching of the semiconductor waveguide layer up to the upper surface of the n-type InP layer 13 is performed.

FIG. 5H shows the planar shape of the SiO ₂ mask 20. A broken line shows the cross-sectional position of FIG. 5G. A laminated structure is left and right.

As shown in FIG. 5I, an i-type InP optical confinement layer 25 having a thickness of 50 nm, a multiple quantum well layer MQW2, and a thickness of 100 nm are formed on the n-type InP layer 13 in a region not covered with the SiO ₂ mask 20 by MOCVD. The i-type InP optical confinement layer 26 is grown. The multiple quantum well layer MQW2 is formed by repeating an InP barrier layer having a thickness of 5 nm and an InGaAsP quantum well layer having a thickness of 10 nm, and includes a barrier layer 14 and a quantum well layer 13 and has a thickness of 200 nm. The mixed crystal composition is selected so that the interband transition wavelength of the InGaAsP quantum well layer is 1400 nm. The total thickness of the i-type layer is 350 nm, which matches the i-type layer thickness of the region grown in the previous process. The transition wavelength of the quantum well layer is 1400 nm, the detuning from the laser oscillation wavelength of 1550 nm is small, and the refractive index change is enhanced by the quantum confined Stark effect (QCSE), so that the refractive index change of the InGaAsP waveguide layer when an electric field is applied A larger index change occurs.

Further, a p-type InP cladding layer 27 having a thickness of 100 nm is grown on the i-type InP optical confinement layer 26 by MOCVD. Thereafter, the SiO ₂ mask 20 is removed.

As shown in FIG. 5J, a p-type InP clad layer 28 is grown on the p-type InP clad layers 18, 24, 27 by MOCVD to a thickness of 1500 nm and a p-type InGaAs contact layer 29 is grown to a thickness of 150 nm. For the waveguide structure etching, a SiO ₂ layer 30 is formed on the p-type InGaAs contact layer 29 by CVD.

As shown in FIG. 5K, the SiO ₂ layer 30 is wet etched into a waveguide shape to form a hard mask. The demultiplexing unit 150 and the multiplexing unit 160 are, for example, patterns having a horizontal length of about 400 μm and a vertical width of about 10 μm so as to form an MMI coupler. The waveguide section 170 has a pattern with a width of 1 μm to 2 μm, for example, a width of about 1.5 μm so as to have a single waveguide mode.

As shown in FIG. 5L, the semiconductor layer exposed from the SiO ₂ hard mask 30 is subjected to reactive ion etching (RIE) until reaching the InP substrate 11 to form a waveguide mesa. The figure shows a cross section of a portion where the waveguides indicated by broken lines in FIG. 5K are parallel to each other.

As shown in FIG. 5M, a semi-insulating (SI) InP layer 51 is grown by MOCVD using the SiO ₂ mask 30 as a growth mask, and a mesa stripe is embedded. The SiO ₂ mask 30 is removed by wet etching. The contact layer 29 other than the electrode formation region is removed by etching using a resist mask. Thereafter, a new SiO ₂ film is formed by CVD for surface protection.

As shown in FIG. 5N, on the SiO ₂ film 52, a photoresist pattern 53, the SiO ₂ film 52 exposed in the opening for the electrode contact is removed by etching. An Au / Zn / Au film 54 is vacuum deposited.

  As shown in FIG. 5O, by removing the photoresist pattern 53, the Au / Zn / Au film 54 on the photoresist pattern 53 is lifted off.

FIG. 5P is a plan view showing a pattern of the Au / Zn / Au film 54 on the surface. The region other than the Au / Zn / Au film 54 is covered with the SiO ₂ film 52.

  As shown in FIG. 5Q, a photoresist pattern 55 is formed, and an electrode forming opening is formed above the Au / Zn / Au film 54. Au plating is performed to form an Au electrode 56 in the opening of the photoresist pattern 55. Thereafter, the photoresist pattern 55 is removed.

  FIG. 5R shows a top view. An Au electrode 56 is disposed on the laser part and the first and second waveguides. The electrodes 56 on the first and second waveguides are arranged at positions shifted asymmetrically with respect to the horizontal direction in the figure.

  FIG. 5S shows a cross-section at the position of the broken line in FIG. 5R. In the two waveguides, the laminated structure in the mesa is symmetrical (same configuration) except for the contact layer 29. An Au electrode is arranged on the second waveguide shown on the right side, and no Au electrode is arranged on the first waveguide shown on the left side. As shown in FIG. 5A, the electrode 130 of the first waveguide WG1 has a butt joint under the region, and the electrode 140 of the second waveguide WG2 does not have a butt joint under the region. It is arranged on the waveguide. The dimensions of the electrodes 130 and 140 are about 1.5 mm in length and about 5 μm in width on the waveguide.

  In the above, the manufacturing process has been described by taking as an example one form that has a butt joint under one of the two waveguides and no butt joint under the other electrode. The manufacturing process will be self-evident. The configuration of the waveguide can be variously modified.

  FIGS. 6A, 6B, 6C, and 6D are plan views showing a form example having a butt joint under one electrode and not having a butt joint under the other electrode.

  In FIG. 6A, the first waveguide WG1 does not have a butt joint, and is formed of the same first semiconductor waveguide layer SWG1 as the demultiplexing unit and the multiplexing unit. A part of the second waveguide WG2 is formed of a second semiconductor waveguide layer SWG2 having a different configuration. From another viewpoint, a waveguide structure SWG2 different from the surrounding is formed in a part of the second waveguide WG2, and the electrode 140 is disposed so as to cover one butt joint.

  In FIG. 6B, the demultiplexing part / multiplexing part is formed by the semiconductor waveguide layer SWG1, and the first and second waveguides are the semiconductor waveguide layer SWG1 following the demultiplexing part / multiplexing part, and a different semiconductor. The waveguide layer SWG2 and both waveguides have butt joints. The electrode 130 above the first waveguide WG1 is disposed so as to cover one butt joint, and the electrode 140 above the second waveguide WG2 is disposed only above the second semiconductor waveguide layer SWG2.

  FIG. 6C shows the configuration shown in FIG. 5A.

  FIG. 6D shows an electrode arrangement similar to FIG. 6B, but on the right side of the electrodes 130 and 140 (region following the multiplexing portion) in the drawing, the second semiconductor waveguide layer SWG2 is followed by the third semiconductor. A waveguide layer SWG3 is disposed. There are three types of semiconductor waveguide layer configurations.

  6E to 6H show a configuration in which the electrode 130 on the first waveguide and the electrode 140 on the second waveguide both cover the butt joint.

  In FIG. 6E, in the first and second waveguides, the second semiconductor waveguide layer SWG2 is formed in an asymmetric arrangement, and the remaining portion is formed by the first semiconductor waveguide layer SWG1. The electrodes 130 and 140 are disposed symmetrically, and the butt joint below the electrodes 130 and 140 is formed in an asymmetric position.

  In FIG. 6F, the first semiconductor waveguide layer SWG1 forms the demultiplexing portion and the asymmetric position in the middle of the first and second waveguides, and the first and second portions from this position to the multiplexing portion are formed. The waveguide and the multiplexing part are formed by the second semiconductor waveguide layer SWG2. The position of the butt joint is asymmetric, and the electrodes 130 and 140 are arranged symmetrically.

  In FIG. 6G, both the waveguides are symmetrically formed in the first semiconductor waveguide layer SWG1 up to the middle of the demultiplexing portion and the first and second waveguides, and the remaining portions of the first and second waveguides are formed. The portion and the combining portion are formed by the second semiconductor waveguide layer. The position of the butt joint is symmetric. The electrodes 130 and 140 are asymmetrically arranged, and the arrangement of the butt joint under the electrodes is asymmetrical.

  In FIG. 6H, the second semiconductor waveguide layer is disposed at a symmetrical position between the first waveguide and the second waveguide, and the remaining portion is formed of the first semiconductor waveguide layer. The electrodes 130 and 140 are arranged asymmetrically. One butt joint is formed under the electrode 130, and two butt joints are formed under the electrode 140.

  As mentioned above, although this invention was demonstrated along the Example, these structures do not have a restrictive meaning at all. Various changes, substitutions, combinations, improvements and the like may be possible.

1 demultiplexer,
2 multiplexing section,
3 waveguide,
4 modulation electrodes,
11 n-type InP substrate,
12 n-type InGaAsP layer,
13 n-type InP cladding layer,
14, 17 i-type InP optical confinement layer,
18 p-type InP cladding layer,
21, 23 i-type InP optical confinement layer,
22 i-type InGaAsP waveguide layer,
24 p-type InP cladding layer,
25, 26 i-type InP optical confinement layer,
27, 28 p-type InP cladding layer,
29 p-type InGaAs contact layer,
31 n-type region,
32 p-type region,
33 Waveguide layer,
34, 35 optical confinement layer,
36 Bulk waveguide
37 multiple quantum well layers,
38 MQW waveguide,
41, 42 electrodes,
54 Au / Zn / Au layer,
56 Au layer,
100 LD active layer,
110 InGaAsP waveguide layer,
120 MQW waveguide layer,
130 (first) modulation electrode,
140 (second) modulation electrode,

Claims (4)

A demultiplexing unit formed on a semiconductor substrate and capable of branching input light into two, a first optical waveguide and a second optical waveguide transmitting two lights emitted from the branching unit, the first optical waveguide, and the second optical waveguide A Mach-Zehnder type optical waveguide structure having a multiplexing section that receives light propagating through the optical waveguide and emits combined output light;
A first modulation electrode and a second modulation electrode, which are respectively formed above the first optical waveguide and the second optical waveguide and apply a first modulation signal and a second modulation signal;
Have
The first optical waveguide includes a first i-type semiconductor layer including a first waveguide layer, and a first pair of doped semiconductor layers that are disposed above and below the first i-type semiconductor layer and doped with impurities at a high concentration. And
The second optical waveguide includes a second i-type semiconductor layer including a second waveguide layer, and a second pair of doped semiconductor layers disposed on the upper and lower sides of the second i-type semiconductor layer and doped with impurities at a high concentration. And
The first i-type semiconductor layer and the second i-type semiconductor layer are equal in thickness in terms of high frequency characteristics,
The first modulation electrode and the second modulation electrode have equal lengths in terms of high frequency characteristics,
Said first modulation electrode, a first optical waveguide of said second modulation electrode downward, one of the second optical waveguide, has a butt joint joining mutually butting in the middle in the extending direction different waveguide layers of structures, while Does not have a butt joint,
A semiconductor optical modulator. Interband transition of the than interband transition wavelength of the first waveguide layer or said second waveguide layer having a butt joint, wherein the first waveguide layer without being perforated with the butt joint or said second waveguide layer 2. The semiconductor optical modulator according to claim 1, wherein the wavelength is shorter. further,
  3. The semiconductor optical modulator according to claim 1, further comprising: a first modulation power source, a second modulation power source, a first load, and a second load connected to the first modulation electrode and the second modulation electrode, respectively. A semiconductor laser unit formed on a semiconductor substrate, a demultiplexing unit capable of branching input light emitted from the semiconductor laser unit into two, and a first optical waveguide and a second optical waveguide that transmit two lights emitted from the demultiplexing unit A semiconductor laminated structure having a waveguide, and a multiplexing unit that receives light propagated through the first optical waveguide and the second optical waveguide and emits combined output light;
A first modulation electrode and a second modulation electrode, which are respectively formed above the first optical waveguide and the second optical waveguide and apply a first modulation signal and a second modulation signal;
A modulation circuit for supplying a push-pull modulation signal to the first modulation electrode and the second modulation electrode;
Have
The first optical waveguide includes a first i-type semiconductor layer including a first waveguide layer, and a first pair of doped semiconductor layers that are disposed above and below the first i-type semiconductor layer and doped with impurities at a high concentration. And
The second optical waveguide includes a second i-type semiconductor layer including a second waveguide layer, and a second pair of doped semiconductor layers disposed on the upper and lower sides of the second i-type semiconductor layer and doped with impurities at a high concentration. And
The first i-type semiconductor layer and the second i-type semiconductor layer are equal in thickness in terms of high frequency characteristics,
The first modulation electrode and the second modulation electrode have equal lengths in terms of high frequency characteristics,
Said first modulation electrode, a first optical waveguide of said second modulation electrode downward, one of the second optical waveguide, has a butt joint joining mutually butting in the middle in the extending direction different waveguide layers of structures, while Does not have a butt joint,
An optical communication module.

Images