JP2011197427A - Semiconductor optical modulator and semiconductor optical transmitter using the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a semiconductor optical modulator capable of obtaining good transmission characteristics even in prechirp transmission, and a semiconductor optical transmitter using the semiconductor optical modulator.SOLUTION: The semiconductor modulator includes: a Mach-Zehnder type optical interferometer including optical waveguides 4a and 4b provided on a substrate 1; an electrode 11a formed on the optical waveguide 4a; an electrode 11b formed on the optical waveguide 4b and different in length from the electrode 11a; a semiconductor mesa structure 7 arranged on the substrate 1 apart from the optical waveguides 4a and 4b by a wavelength of the light to be modulated or more; and an auxiliary electrode 12 formed on the semiconductor mesa structure 7 and electrically connected to the electrode 11b.

Description

  The present invention relates to a semiconductor optical modulator and a semiconductor optical transmitter using the same.

  Conventionally, for long-distance and large-capacity optical communication, a Mach-Zehnder (MZ) interferometer type modulator (hereinafter referred to as an LN-MZ modulator) using the electro-optic effect in the nonlinear optical crystal lithium niobate (LN). Transmitters in combination with semiconductor lasers are mainly used.

  Further, a general silica glass core optical fiber used for optical communication exhibits a very small transmission loss when the wavelength of light is near 1.5 μm in the near infrared. For this reason, 1.5 μm band light is often used for long-distance optical communication. On the other hand, many single-mode optical fibers currently laid in the trunk line system have positive chromatic dispersion in this wavelength band. Therefore, in order to compensate for transmission waveform deterioration due to dispersion, pre-chirp transmission that intentionally gives negative chirp to modulated light is used in long-distance transmission, particularly long-distance transmission exceeding 50 km. That is, in the LN-MZ modulator, a desired negative chirp is imparted by appropriately adjusting the modulation amplitude ratio of the intensity of the electric field applied to the two arms (waveguides) that are the paths of the demultiplexed optical signal. . In particular, in a type of LN-MZ modulator called a Z-cut type, a single electrode formed at an asymmetric position with respect to two waveguides gives electric field changes of different strengths to the two waveguides. The desired negative chirp is given.

  In recent years, research on Mach-Zehnder interferometric modulators using semiconductors (hereinafter sometimes referred to as semiconductor MZ modulators) has also been conducted. The semiconductor MZ modulator has an advantage that it can be miniaturized because the refractive index of the semiconductor waveguide and the rate of change thereof are larger than those of the LN waveguide. Further, since integration with a semiconductor laser is easy, there is an advantage that the number of parts in the transmitter can be reduced.

  Here, the structure of a conventional semiconductor MZ modulator will be described. FIG. 1 is a schematic diagram showing an example of the structure of a conventional semiconductor MZ modulator.

  As shown in FIG. 1A, an input waveguide 52, an optical distributor 53, waveguides 54a and 54b, an optical coupler 55, and an output waveguide 56 are formed on a substrate 51. These are composed of semiconductor waveguides. An electrode 61a is formed on the waveguide 54a, and an electrode 61b is formed on the waveguide 54b. Furthermore, an electrode is formed on the back surface of the substrate 51. A bias DC power supply 71a and an AC power supply 72a are connected in series between the electrode 61a and the ground, and a bias DC power supply 71b and an AC power supply 72b are connected between the electrode 61b and the ground.

  The semiconductor MZ modulator is also required to be pre-chirped like the LN-MZ modulator. Several methods for giving pre-chirp are being studied. The first method is a method in which the amplitude of the modulation voltage applied to the electrode 61a (modulation voltage of the AC power supply 72a) and the amplitude of the modulation voltage applied to the electrode 61b (modulation voltage of the AC power supply 72b) are made asymmetric. . The second method is a method in which the bias voltage applied to the electrode 61a (voltage of the bias DC power supply 71a) and the bias voltage applied to the electrode 61b (voltage of the bias DC power supply 71b) are made asymmetric. The third method is a method of making the length of the electrode 61a and the length of the electrode 61b asymmetric as shown in FIG. Both of these methods are premised on push-pull drive that applies antisymmetric modulation voltages to both arms (waveguides 54a and 54b).

  However, in the first method, it is necessary to provide a high-frequency attenuator in the transmitter, which is disadvantageous from the viewpoint of miniaturization of the apparatus and power saving. Only with the second method, the obtained asymmetry is small and a sufficient pre-chirp cannot be imparted. In the third method, a difference occurs in high-frequency characteristics (power reflection, transmission characteristics, etc.) between the electrodes 61a and 61b, and good transmission characteristics cannot be obtained.

JP 2004-102097 A Japanese Patent Laid-Open No. 2000-066156

L. B. Soldano, et al, J. Lightwave Technol., Vol.13, pp.615-627 (1995)

  An object of the present invention is to provide a semiconductor optical modulator capable of obtaining good transmission characteristics even during pre-chirp transmission and a semiconductor optical transmitter using the same.

  In one aspect of the semiconductor optical modulator, a Mach-Zehnder optical interferometer provided with a first optical waveguide and a second optical waveguide provided on a substrate, and the first optical waveguide are formed. A first electrode and a second electrode formed on the second optical waveguide and having a length different from that of the first electrode are provided. Furthermore, a semiconductor mesa structure provided at a distance equal to or greater than the wavelength of the modulated light from the first optical waveguide and the second optical waveguide, and formed on the semiconductor mesa structure, the first electrode and An auxiliary electrode electrically connected to a short one of the second electrodes.

  According to the above semiconductor optical modulator or the like, the high frequency characteristics between the two electrodes can be made uniform, and good transmission characteristics can be obtained even during pre-chirp transmission.

It is a schematic diagram which shows an example of the structure of the conventional semiconductor MZ modulator. It is a figure which shows the result of a 1st simulation. It is a figure which shows the result of a 2nd simulation. Similarly, it is a figure which shows the result of a 2nd simulation. (A) is a schematic diagram which shows the structure of the semiconductor optical modulator based on 1st Embodiment, (b) is a schematic diagram which shows the structure of the semiconductor optical modulator concerning 2nd Embodiment. It is a top view which shows the structure of the semiconductor optical modulator based on 3rd Embodiment. It is sectional drawing which shows the method of manufacturing the semiconductor optical modulator concerning 3rd Embodiment in process order. FIG. 7B is a cross-sectional view illustrating a method of manufacturing the semiconductor optical modulator in the order of steps following FIG. 7A. Similarly, it is sectional drawing which shows the method of manufacturing the semiconductor optical modulator concerning 3rd Embodiment in order of a process. It is a top view which shows the method of manufacturing the semiconductor optical modulator concerning 3rd Embodiment in order of a process. It is a figure which shows the modification of 3rd Embodiment. It is a figure which shows the other modification of 3rd Embodiment. It is a figure which shows the structure of a semiconductor optical transmitter.

  As described above, as shown in FIG. 1B, when the length of the electrode 61a and the length of the electrode 61b are asymmetrical, a difference occurs in the high-frequency characteristics between the electrodes 61a and 61b, resulting in good transmission characteristics. Can't get. This is because the waveform of the modulation voltage that is effectively applied differs between the electrodes 61a and 61b, and the eye opening ratio of the waveform of the modulated light is degraded as compared with the case where the waveforms are equal.

  Here, various simulations performed by the present inventors will be described.

  In the first simulation, two types of low-pass filters with different cutoff frequencies are connected to an AC power source and an electrode in order to simulate the appearance of the RF waveform being degraded by the high-frequency characteristics of the transmission line. The RF modulation voltage waveform (eye pattern) was obtained. Here, the low-pass filter is a fourth-order Butterworth filter often used for frequency characteristic analysis of an optical transmitter. One cutoff frequency was 20 GHz, and the other cutoff frequency was 7.5 GHz. The RF waveform of the input signal was a 10 Gbit / s NRZ signal waveform having a raised cosine waveform with a raised cosine waveform, and both the rise time and the fall time were 50 ps.

  FIG. 2 shows the result of the first simulation (the RF modulation voltage waveform after passing through the low-pass filter). FIG. 2A shows an RF modulation voltage waveform when the cutoff frequency is 20 GHz, and FIG. 2B shows an RF modulation voltage waveform when the cutoff frequency is 7.5 GHz. As shown in FIG. 2, since the cut-off frequencies are different, the RF modulation voltage waveforms are greatly different. This means that RF signals applied to electrodes having different high frequency characteristics (high frequency bands) are also greatly different from each other.

  In the second simulation, when an RF signal having the same waveform is applied to two electrodes to perform push-pull driving, and when an RF signal having a different waveform is applied to two electrodes to perform push-pull driving. The waveform (eye pattern) of the modulated light was obtained. Here, the semiconductor MZ modulator operates an MZ interferometer having a semiconductor multiple-quantum well (MQW) as a waveguiding layer by a symmetric push-pull drive, and in each arm of the MZ interferometer. The change in refractive index was expressed by a quadratic function with respect to the applied voltage. Also, the amplitude of the RF signal applied to both arms of the interferometer is the same, the electrode length of each arm is asymmetric as 1: 3, and the bias voltage is also asymmetric. When applying RF signals with the same waveform, the high frequency bands of the electrodes on both arms are aligned at 7.5 GHz regardless of the length of the electrodes. When applying RF signals with different waveforms, The high frequency band of the long electrode is 7.5 GHz, and the high frequency band of the short electrode is 20 GHz. The waveform of the modulated light was obtained before and after transmission using an optical fiber. In the calculation of the waveform after transmission, the time waveform is Fourier transformed, multiplied by the transfer function of the optical fiber on the frequency axis, and inverse Fourier transformed to return to the time domain, resulting in waveform distortion due to group delay dispersion in the fiber. Evaluated. In this simulation, only the secondary dispersion was considered and a dispersion amount of 1600 ps / nm was given. This is a dispersion amount corresponding to transmission of 80 km to 100 km with a standard optical fiber.

3 and 4 show the result of the second simulation (the waveform of the modulated light). FIG. 3A shows a waveform before transmission (waveform of output light from the modulator) when RF signals having the same waveform are applied, and FIG. 3B shows the output light having the waveform shown in FIG. The waveform after transmission is shown. 4A shows a waveform before transmission (waveform of output light from the modulator) when RF signals having different waveforms are applied, and FIG. 4B shows output light having the waveform shown in FIG. The waveform after transmission is shown. Comparing FIG. 3 (a) and FIG. 4 (a), when the high frequency band is asymmetric, the ratio between the on-level and the off-level of the waveform before transmission (dynamic extinction ratio), compared to the symmetric case. It can be seen that is deteriorated. The length of the arrow in FIG. 3A and FIG. 4A corresponds to the dynamic extinction ratio. It can also be seen from the part indicated by the arrow in FIG. 4B that the waveform after transmission is asymmetrically distorted and the eye opening is degraded. In order to perform error-free transmission with a code error rate of 1 × 10 ⁻¹¹ or less, such deterioration of the eye opening is not allowed and cannot be used as it is.

  As a result of these simulations, the inventors of the present application have found that if the high-frequency bands of the electrodes on the two waveguides are symmetrical, good transmission characteristics can be obtained even when these high-frequency bands are low. Hereinafter, embodiments based on such knowledge will be specifically described with reference to the accompanying drawings.

(First embodiment)
First, the first embodiment will be described. FIG. 5A is a schematic diagram showing the structure of the semiconductor optical modulator according to the first embodiment. The first embodiment is an example of a semiconductor optical modulator provided with a lumped constant type electrode.

  In the first embodiment, as shown in FIG. 5A, the input waveguide 2, the optical distributor 3, the waveguides 4a and 4b, the optical coupler 5, and the output waveguide 6 are formed on the substrate 1. ing. These are composed of semiconductor waveguides. The structure of the semiconductor waveguide is not particularly limited, and is, for example, a p-i-n layer structure, a p-i-p layer structure, or a n-i-n layer structure. In the case of integration on the same substrate as the semiconductor laser, a pin layer structure is preferable among these. The input waveguide 2, the optical distributor 3, the waveguides 4a and 4b, the optical coupler 5, and the output waveguide 6 constitute a Mach-Zehnder interferometer. Further, a semiconductor mesa structure 7 extending in parallel with the waveguides 4a and 4b is formed on the substrate 1 independently of the waveguides 4a and 4b at a position sandwiching the waveguide 4b with the waveguide 4a. The width of the semiconductor mesa structure 7 is equal to the width of the waveguides 4a and 4b, and the sectional structure (layer structure) of the semiconductor mesa structure 7 is the same as the sectional structure (layer structure) of the waveguides 4a and 4b. For example, the thickness of the i layer (first depletion layer) included in the waveguides 4 a and 4 b is equal to the thickness of the i layer (second depletion layer) included in the semiconductor mesa structure 7. The semiconductor mesa structure 7 is provided at a position that does not affect the distribution of the electromagnetic field of the light guided through the waveguides 4a and 4b. That is, the distance between the semiconductor mesa structure 7 and the waveguides 4a and 4b is equal to or greater than the wavelength of the light (modulated light) guided through the waveguides 4a and 4b. Note that the width of the semiconductor mesa structure 7 does not need to match the width of the waveguides 4a and 4b, but even if it does not match, the width of the semiconductor mesa structure 7 is the width of the waveguides 4a and 4b. It is preferably 90% to 110%, more preferably 95% to 105%. Similarly, the thickness of the i layer included in the semiconductor mesa structure 7 does not need to match the thickness of the i layer included in the waveguides 4a and 4b. The thickness of the i layer included in 7 is preferably 90% to 110%, more preferably 95% to 105% of the thickness of the i layer included in the waveguides 4a and 4b.

  An electrode 11a is formed on the waveguide 4a, and an electrode 11b is formed on the waveguide 4b. Further, the auxiliary electrode 12 is formed on the semiconductor mesa structure 7, and the auxiliary electrode 12 and the electrode 11 b are connected to each other through the connection portion 13. The electrode 11a, the electrode 11b, the auxiliary electrode 12, and the connection part 13 are comprised from the same material. The electrode 11a is longer than the electrode 11b, and the total length of the electrode 11b and the auxiliary electrode 12 preferably matches the length of the electrode 11a, but these do not need to match. Even if they do not match, the total length of the electrode 11b and the auxiliary electrode 12 is preferably 90% to 110%, more preferably 95% to 105% of the length of the electrode 11a. Further, the length of the semiconductor mesa structure 7 is not less than the length of the auxiliary electrode 12. Note that the length of the interferometer (dimension in the direction in which the waveguides 4a and 4b extend) is considerably larger than the width of the interferometer (dimension in the direction in which the waveguides 4a and 4b are arranged), and the length of the electrode 11a is For example, it is about 1.5 mm, and the interval between the waveguides 4a and 4b is, for example, about 50 μm. Moreover, the space | interval between the semiconductor mesa structure 7 and the waveguide 4b is 5 micrometers, for example. This interval is equal to or greater than the wavelength of light (modulated light) guided through the waveguides 4a and 4b. Furthermore, an electrode is formed on the back surface of the substrate 1. A bias DC power supply 21a and an AC power supply 22a are connected between the electrode 11a and the ground, and a bias DC power supply 21b and an AC power supply 22b are connected between the electrode 11b and the ground.

  In the first embodiment configured as described above, when light having a constant intensity (CW: continuous wave light) is incident on the input waveguide 2 and an RF signal is applied to the electrodes 11a and 11b, the semiconductor MZ light modulation is performed. Operates as a container. At this time, since the lengths of the electrodes 11a and 11b are asymmetrical, a desired negative chirp operation can be obtained even if push-pull driving is performed with an RF signal having the same amplitude. Therefore, pre-chirp transmission suitable for long-distance transmission is possible.

  The high-frequency characteristics (high-frequency band) of the electrode to which the RF signal is applied mainly depend on the capacity and length per unit length between the electrode to which the RF signal is applied and the electrode provided on the back surface of the substrate 1. Depends on. Further, the capacitance per unit length between the electrode to which the RF signal is applied and the electrode provided on the back surface of the substrate 1 is substantially determined by the thickness and width of the i layer positioned therebetween. In the present embodiment, the RF signal from the AC power supply 22a is applied to the electrode 11a, and the RF signal from the AC power supply 22b is applied to the electrode 11b, the connection unit 13, and the auxiliary electrode 12. Here, the length of the electrode 11a is equivalent to the sum of the length of the electrode 11b and the length of the auxiliary electrode 12, and the widths thereof are also equivalent. Furthermore, the width of the semiconductor mesa structure 7 is equal to the width of the waveguides 4a and 4b, and the cross-sectional structure of the semiconductor mesa structure 7 is the same as the cross-sectional structure of the waveguides 4a and 4b. Therefore, the characteristic impedances of the electrodes 11a and 11b are equal, and the high frequency characteristics are uniform. For this reason, as apparent from the above-described simulation, the modulation voltage waveforms of both RF signals are aligned, it is possible to suppress undesirable distortion of the modulated light waveform, and the extinction ratio and the eye opening ratio of the post-transmission waveform are excellent. Can be. That is, good transmission characteristics can be obtained even in high-capacity communication on the trunk line.

(Second Embodiment)
Next, a second embodiment will be described. FIG. 5B is a schematic diagram showing the structure of the semiconductor optical modulator according to the second embodiment. The second embodiment is an example of a semiconductor optical modulator provided with traveling wave type electrodes.

  In the second embodiment, as shown in FIG. 5B, the input pad 14a is provided at the end of the electrode 11a on the input waveguide 2 side, and the output pad 15a is provided at the end of the output waveguide 6 side. The input pad 14b is provided at the end of the electrode 11b on the input waveguide 2 side, and the output pad 15b is provided at the end of the auxiliary electrode 12 on the output waveguide 6 side. A bias DC power supply 21a and an AC power supply 22a are connected between the input pad 14a and the ground, and a bias DC power supply 21b and an AC power supply 22b are connected between the input pad 14b and the ground. A termination resistor 23a is connected between the output pad 15a and the ground, and a termination resistor 23b is connected between the output pad 15b and the ground. Other configurations are the same as those of the first embodiment.

  In the second embodiment including the traveling wave type electrode configured as described above, an appropriate pre-chirp can be imparted and good transmission characteristics can be obtained as in the first embodiment. it can.

  The bias DC power supply 21a may be connected in series with the termination resistor 23a, and the bias DC power supply 21b may be connected in series with the termination resistor 23b.

(Third embodiment)
Next, a third embodiment will be described. FIG. 6 is a top view showing the structure of the semiconductor optical modulator according to the third embodiment. The third embodiment is an example of a semiconductor optical modulator provided with traveling wave type electrodes.

  In the third embodiment, as shown in FIG. 6, an input waveguide 202, an optical distributor 203, waveguides 204a and 204b, an optical coupler 205, and an output waveguide 206 are formed on a substrate. These are composed of semiconductor waveguides. The structure of the semiconductor waveguide is not particularly limited, and is, for example, a p-i-n layer structure, a p-i-p layer structure, or a n-i-n layer structure. In the case of integration on the same substrate as the semiconductor laser, a pin layer structure is preferable among these. The input waveguide 202, the optical distributor 203, the waveguides 204a and 204b, the optical coupler 205, and the output waveguide 206 constitute a Mach-Zehnder interferometer. Further, a semiconductor mesa structure 207 extending in parallel with the waveguides 204a and 204b is formed on the substrate independently of the waveguides 204a and 204b at a position sandwiching the waveguide 204b with the waveguide 204a. The width of the semiconductor mesa structure 207 is equal to the width of the waveguides 204a and 204b, and the cross-sectional structure (layer structure) of the semiconductor mesa structure 207 is the same as the cross-sectional structure (layer structure) of the waveguides 204a and 204b. is there. For example, the thickness of the i layer (first depletion layer) included in the waveguides 204a and 204b and the thickness of the i layer (second depletion layer) included in the semiconductor mesa structure 207 are approximately the same. . The width of the waveguide 204a, the waveguide 204b, and the semiconductor mesa structure 207 is 1.5 μm, for example, and the distance between the semiconductor mesa structure 207 and the waveguide 204b is 5 μm, for example. This interval is equal to or greater than the wavelength of light (modulated light) guided through the waveguides 204a and 204b.

  An electrode 211a is formed on the waveguide 204a, and an electrode 211b is formed on the waveguide 204b. Further, one auxiliary electrode 212 is formed at two locations on the semiconductor mesa structure 207, and the two auxiliary electrodes 212 and the electrode 211 a are connected to each other via the connection portion 213. The electrode 211a is longer than the electrode 211b, and the sum of the length of the electrode 211b and the length of the two auxiliary electrodes 212 is equal to the length of the electrode 211a. For example, the length of the electrode 211a is 1.5 mm, the length of the electrode 211b is 0.5 mm, and the total length of the auxiliary electrode 212 is 1.0 mm. Further, the length of the semiconductor mesa structure 207 is equal to or greater than the sum of the lengths of the two auxiliary electrodes 212. The length of the interferometer (dimension in the direction in which the waveguides 204a and 204b extend) is considerably larger than the width of the interferometer (dimension in the direction in which the waveguides 204a and 204b are arranged), and the interval between the waveguides 204a and 204b. Is, for example, about 50 μm. Further, an electrode is formed on the back surface of the substrate.

  For the semiconductor optical modulator configured as described above, as in the second embodiment, for example, a bias DC power supply and an AC power supply are connected between the input pad 214a and the ground, and the input pad 214b and the ground are connected. Are connected to another bias DC power source and an AC power source. Further, a termination resistor is connected between the output pad 215a and the ground, and another termination resistor is connected between the output pad 215b and the ground. Note that the protective film 110 covers a region where no electrode, input pad, or output pad is provided.

  According to the third embodiment, the same effect as that of the second embodiment can be obtained.

  Next, a method for manufacturing the semiconductor optical modulator according to the third embodiment will be described. 7A and 7B are cross-sectional views showing a method of manufacturing the semiconductor optical modulator according to the third embodiment in the order of steps. FIG. 8 is also a cross-sectional view showing a method of manufacturing the semiconductor optical modulator according to the third embodiment in the order of steps. FIG. 9 is a top view showing the method of manufacturing the semiconductor optical modulator according to the third embodiment in the order of steps. 7A and 7B show a cross section taken along the line II in FIGS. 6 and 9, and FIG. 8 shows a cross section taken along the line II-II in FIGS.

  First, as shown in FIG. 7A (a), an n-type buffer layer 102, a multiple quantum well (MQW: multi-well) are formed on an n-type substrate 101 by, for example, metal organic chemical vapor deposition (MOCVD). -quantum wells) layer 103, cladding layer 104, p-type cladding layer 105, and p-type contact layer 106 are laminated in this order. For example, an n-type InP substrate is used as the n-type substrate 101. As the n-type buffer layer 102, for example, an n-type InP layer having a thickness of about 100 nm is formed. As the cladding layer 104, for example, an i-type InP layer to which an impurity having a thickness of about 100 nm is not added is formed. As the p-type cladding layer 105, for example, a p-type InP layer having a thickness of about 1500 nm is formed. As the p-type contact layer, for example, a p-type InGaAs layer having a thickness of about 150 nm is formed. As the MQW layer 103, for example, a stacked body in which InGaAsP well layers and InP barrier layers are alternately stacked is formed. For example, the thickness of the well layer is set to 10 nm, the thickness of the barrier layer is set to 10 nm, and these are repeated 20 cycles. The MQW layer 103 functions as a waveguide layer. The composition of the well layer is determined so that the transition wavelength at the band edge is 1.4 μm, for example. In this case, when the wavelength of the modulated light is about 1.55 μm, a practically sufficient refractive index change can be obtained by the quantum confined Stark effect (QCSE).

  Next, a silicon oxide film is formed on the p-type contact layer 106 by, for example, chemical vapor deposition (CVD). Thereafter, a resist pattern is formed on the silicon oxide film to cover a region where the semiconductor waveguide is to be formed and a region where the semiconductor mesa structure 207 is to be formed. Then, the silicon oxide film is wet etched using the resist pattern as a mask. As a result, the shape of the resist pattern is transferred to the silicon oxide film, and as shown in FIG. 7A (b), a region covering the region where the semiconductor waveguide is to be formed and the region where the semiconductor mesa structure 207 is to be formed are covered. A pattern 107 is formed.

  Subsequently, the p-type contact layer 106, the p-type cladding layer 105, the cladding layer 104, the MQW layer 103, and the n-type buffer layer 102 are dry-etched using the mask pattern 107 as a mask. This dry etching is performed by, for example, reactive ion etching (RIE). In this dry etching, the n-type substrate 101 is slightly over-etched. As a result, as shown in FIGS. 7A (c) and 9 (a), the mesa stripe structure of the waveguide 204a, the waveguide 204b, and the semiconductor mesa structure 207 is obtained.

  At this time, the width of the waveguide 204a, the waveguide 204b, and the semiconductor mesa structure 207 is, for example, 1.5 μm. This is because the waveguide mode is single at the wavelength of the modulated light. Further, the interval between the semiconductor mesa structure 207 and the waveguide 204b is, for example, 5 μm. Further, as the optical distributor 203 and the optical coupler 205, an MMI (multi-mode interference) coupler is provided. As a result, light can be branched to both arms (waveguides 204a and 204b) of the interferometer, and light from both arms can be multiplexed. The width and length of the MMI coupler are designed so that the branching ratio becomes 1: 1, for example, in accordance with the ratio of the refractive index of the waveguide section to the surrounding refractive index.

  Next, as shown in FIG. 7B (d), for example, buried growth by MOCVD is performed to form a buried layer 108 around the mesa stripe structure. As the buried layer 108, for example, a high resistance (SI: semi-insulated) InP layer is formed. Thereafter, the mask pattern 107 is removed using BHF (Buffered hydrofluoric acid).

  Subsequently, a resist pattern 109 is formed on the buried layer 108 and the p-type contact layer 106 as shown in FIGS. 7B (e) and 8 (a). The resist pattern 109 is formed so as to cover the p-type contact layer 106 in a region where electrodes are to be formed and to expose the p-type contact layer 106 in other regions. Then, wet etching is performed using the resist pattern 109 as a mask, and the portion of the p-type contact layer 106 exposed from the resist pattern 109 is removed. Next, as shown in FIG. 9B, the resist pattern 109 is removed.

  Thereafter, a protective film 110 is formed on the entire surface, and a resist pattern 111 is formed thereon. For example, a silicon oxide film is formed as the protective film 110. Further, as the resist pattern 111, a pattern that exposes the protective film 110 in a region where an electrode on the mesa is to be formed is formed. Then, the protective film 110 is wet etched using the resist pattern 111 as a mask. As a result, as shown in FIGS. 7B (f) and 8 (b), the p-type contact layer 106 in the region where the electrode is to be formed is exposed.

  Subsequently, as shown in FIG. 7B (g) and FIG. 8 (c), for example, a seed metal film 112 is formed on the resist pattern 111 and the p-type contact layer 106. In the formation of the seed metal film 112, for example, vacuum deposition of Au, vacuum deposition of Zn, and vacuum deposition of Au are performed in this order.

  Next, the resist pattern 111 is removed together with the seed metal film 112 thereon. As a result, the seed metal film 112 remains only on the p-type contact layer 106 as shown in FIGS. 7B (h) and 8 (d). That is, the seed metal film 112 is formed by a lift-off method.

  Thereafter, a resist pattern 113 exposing the seed metal film 112 is formed on the protective film 110 as shown in FIGS. 7B (i) and 8 (e). Then, an Au film 114 is formed on the seed metal film 112 by performing Au plating using the resist pattern 113 as a mask.

  Subsequently, as shown in FIGS. 7B (j) and 8 (f), the resist pattern 113 is removed. Next, a seed metal film 115 and an Au film 116 are formed in this order on the back surface of the n-type substrate 101. In forming the seed metal film 115, for example, AuGe vacuum deposition and Au vacuum deposition are performed in this order. The Au film 116 can be formed by plating, for example.

  As a result of such processing, the semiconductor optical modulator shown in FIG. 6 is obtained.

  Note that the relationship among the electrode 211b, the auxiliary electrode 212, and the semiconductor mesa structure 7 is not limited to that shown in FIG. For example, as shown in FIG. 10A, the semiconductor mesa structure 7 may be separated from the end face of the substrate. Further, as shown in FIG. 10B, the auxiliary electrode 212 may be connected to only one end of the auxiliary electrode 212. Further, as shown in FIG. 10C, an auxiliary electrode 222 similar to the auxiliary electrode 212 is connected to the electrode 211a, and a semiconductor mesa structure 227 similar to the semiconductor mesa structure 207 is formed under the auxiliary electrode 222. It may be provided. However, the lengths of the electrodes 211a and 211b are different from each other, and the total length of the electrode 211a and the auxiliary electrode 222 is equal to the total length of the electrode 211b and the auxiliary electrode 212. As shown in FIG. 10D, the auxiliary electrode 212 may be located between the waveguides 204a and 204b in plan view.

  Furthermore, the relationship between these electrodes, auxiliary electrodes, and semiconductor mesa structure can also be applied to a semiconductor optical modulator having a lumped constant type electrode as shown in FIG. The structures shown in FIGS. 11A to 11D correspond to the structures shown in FIGS. 10A to 10D, respectively, and the structure shown in FIG. 11E corresponds to the structure shown in FIG.

  Also, as shown in FIG. 12, a semiconductor optical transmitter can be obtained by integrating the semiconductor optical modulator and the semiconductor laser 120 on a single substrate. This semiconductor optical transmitter is suitable for trunk-system large-capacity communication.

  Furthermore, the semiconductor material constituting the waveguide is not particularly limited. For example, a group II element such as Zn, Cd, Hg, Al, N, Sb, S, Te, a group III element, a group V element, or a group VI A mixed crystal containing an element may be used as a material for the waveguide layer (active layer). For example, a quantum well layer using AlGaInAs as the material of the well layer and AlInAs as the material of the barrier layer may be used. Moreover, you may use common couplers, such as a directional coupler and a Y branch waveguide, as an optical coupler and an optical splitter. In the above embodiment, the optical waveguide has an SI-BH (buried heterostructure) structure, but it may be a PNPN type buried waveguide. Further, a high-mesa waveguide structure that does not embed using a semiconductor may be provided, or a ridge-type waveguide structure in which only the upper cladding layer is etched may be provided.

  Further, the width of the semiconductor mesa structure does not necessarily match that of the waveguide, and the thickness of the i layer included in the semiconductor mesa structure does not necessarily match that of the waveguide. . However, especially when performing modulation with a high-frequency RF signal of several GHz or more, the frequency characteristics of the electrode are affected not only by the overall capacitance but also by the length of the electrode and the capacity per unit length. It is desirable that they match. Also, if they match, manufacturing is easier than if they do not match.

1: Substrate 4a, 4b: Waveguide 7: Semiconductor mesa structure 11a, 11b: Electrode 12: Auxiliary electrode 13: Connection part 204a, 204b: Waveguide 207, 227: Semiconductor mesa structure 211a, 211b: Electrode 212, 222 : Auxiliary electrode 213: Connection part

Claims (5)

A Mach-Zehnder type optical interferometer provided with a first optical waveguide and a second optical waveguide provided on a substrate;
A first electrode formed on the first optical waveguide;
A second electrode formed on the second optical waveguide and having a length different from that of the first electrode;
A semiconductor mesa structure provided apart from the first optical waveguide and the second optical waveguide by at least the wavelength of the modulated light;
An auxiliary electrode formed on the semiconductor mesa structure and electrically connected to a short one of the first electrode and the second electrode;
A semiconductor optical modulator comprising: The first optical waveguide and the second optical waveguide have a first depletion layer;
The semiconductor mesa structure has a second depletion layer;
The thickness and width of the second depletion layer are 90% to 110% of the thickness and width of the first depletion layer,
The sum of the length of the auxiliary electrode and the length of the short one of the first electrode and the second electrode is 90% to 110% of the length of the long one of the first electrode and the second electrode. The semiconductor optical modulator according to claim 1, wherein: The layer structure of the semiconductor mesa structure is equal to the layer structure of the first optical waveguide and the second optical waveguide,
The width of the semiconductor mesa structure is 90% to 110% of the width of the first optical waveguide and the second optical waveguide,
The sum of the length of the auxiliary electrode and the length of the short one of the first electrode and the second electrode is 90% to 110% of the length of the long one of the first electrode and the second electrode. The semiconductor optical modulator according to claim 1, wherein:   4. The semiconductor optical modulator according to claim 1, wherein each of the first optical waveguide and the second optical waveguide has a waveguide layer having a semiconductor multiple quantum well structure. 5. A semiconductor optical modulator according to any one of claims 1 to 4,
A semiconductor laser integrated with the semiconductor optical modulator on the substrate;
A semiconductor optical transmitter characterized by comprising:

Images