JP2007201072A - Semiconductor optical element - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To suppress diffusion of Zn into an undoped layer in the crystalline multilayer structure of an EA modulator. <P>SOLUTION: An InGaAlAs layer 201 is introduced by about 30 nm between an upper side guide layer 104 of undoped InGaAsP and a p-type InP clad layer 105. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a semiconductor optical device that is one of the main components of basic optical transmission having a transmission rate of 2.5 Gbit / s or more, and in particular to control the doping concentration of the semiconductor crystal in order to realize the characteristics. This is a technique related to an optical modulator which is a design parameter. The technical field extends to a semiconductor laser module for optical communication, an optical transmission module, an optical transmission device, and an optical communication system equipped with an optical modulator.

  The prior art will be described using an example of a ridge waveguide semiconductor electroabsorption modulator integrated laser with a transmission speed of 40 Gbit / s wavelength 1.55 μm band for 2 km optical transmission.

  A 40 Gbit / s semiconductor electroabsorption modulator integrated laser (hereinafter referred to as EA (Electro-Absorption) modulator integrated laser) is a DFB laser (Distributed feedback laser) unit driven by a constant current and an EA modulator operated by a modulation voltage. It consists of parts. The EA modulator turns on and off the light of the DFB laser by shifting the EA modulator active layer absorption edge using the quantum confined Stark effect generated by applying a voltage to the EA modulator. It is an EA modulator (see Non-Patent Document 1).

  A method for manufacturing this EA modulator integrated laser will be described below. This EA modulator is described in, for example, “N. Sasada, et.al.,“ 1.55-μm 40-Gbit / selection-absorption modulation Integrated DFB laser module for highly short-reaching EC2 ” EA modulator integrated laser as typified by.

  FIG. 12 is a top view of such an electroabsorption modulator integrated laser. FIG. 13 is a cross-sectional view along the optical axis of the ridge-type optical waveguide (a-a ′ in FIG. 12). 14 is a cross-sectional view in the EA modulator section mesa vertical direction (b-b 'in FIG. 12).

As shown in FIG. 12, as the crystal growth of the EA modulator portion on the n-InP semiconductor substrate 100, an n-type InP buffer layer 101, an InGaAsP lower light guide are formed by a known selective growth method using a metal organic vapor phase method. An undoped strained multiple quantum well layer 103 composed of an InGaAsP well layer as an active layer and a barrier layer, an undoped InGaAsP upper light guide layer 104, and a p-type InP cap layer 105 of about 4 × 10 ¹⁷ cm ⁻³ are formed. .

  Next, after forming a SiN film by plasma CVD, patterning is performed on the SIn film in the region that becomes the EA modulator portion, and the region that becomes the EA modulator portion by dry etching and wet etching using the SiN film as a mask. The EA modulator multilayer structure other than is removed.

  Next, as the second crystal growth, a quantum well comprising an InGaAsP lower optical guide layer 106, an InGaAsP well layer and a barrier layer, which is a multi-layer growth of a laser using a metal organic vapor phase method in addition to the EA modulator region. An active layer 107, an InGaAsP upper light guide layer 108, an InP spacer layer 109, and a diffraction grating layer 110 are sequentially formed. At this time, the EA modulator and the laser unit are optically connected by a known butt joint technique.

  Further, an InGaAsP layer 111 serving as a passive optical waveguide portion is crystal-grown between the EA modulator region and the laser region by a process similar to the procedure in which the EA modulator multilayer other than the EA modulator region is removed. Also at this time, the EA modulator, the passive waveguide, and the laser unit are optically connected by a known butt joint technique.

Next, a diffraction grating is formed in the semiconductor diffraction grating layer by the interference exposure method using a photolithograph on the diffraction grating layer 110 in a region to be a laser part, and further, on the EA modulator, the passive waveguide, and the laser part Then, a p-type InP clad layer 112 of about 6 × 10 ¹⁷ cm ^{−3 serving as a} clad layer, a p-type InGaAsP layer 113 and a p-type InGaAs layer 114 serving as contact layers, and a p-type InP cap layer are crystal-grown. This p-type InP cap layer is removed in an intermediate process and does not remain in the final structure. Here, Zn is used for the p-type dopant. Zn doped in the p-InP cladding layer diffuses into the InGaAsP upper optical guide layer, which is an undoped layer under the EA modulator region, due to the influence of heat during crystal growth. This has been found by SIMS (secondary ion mass spectroscopy) analysis.

  Thereafter, a ridge-type optical waveguide is manufactured by wet etching from the EA modulator to the passive waveguide and the laser portion, and the side of the optical waveguide is formed by passivation film formation 115, polyimide resin 116 application, and etch back. The p-side electrode 117 is formed by EB (Electron Beam) vapor deposition and ion milling in a through-hole process. Then, in the back surface polishing process, the n-type InP substrate is polished to about 100 μm, an n-side electrode is formed 118, and an electrode alloy process is performed to complete the wafer process.

  Subsequently, the wafer is divided, and a non-reflective coating is applied to the front end face of the device, and a high reflectivity coat is applied to the rear end face. Thus, an electroabsorption modulator integrated laser is produced.

  In an optical modulator operating at 40 Gbit / s, it is necessary to secure an f3 dB band of at least 35 GHz for its high-speed operation, whereas the parasitic capacity of a 10 Gbit / s EA modulator is about 0.5 pF. The capacitance must be reduced to about 0.1 to 0.2 pF. For this reason, the 40 Gbit / s EA modulator shown here is designed to have a modulator length of 100 μm and shorter than the 10 Gbit / s EA modulator, and in order to prevent diffusion of p-type dopants in the lateral direction of the optical waveguide, it is 10 Gbit / s. The EA modulator of this type employs a ridge waveguide structure in which a waveguide made of resin is buried without embedding with a commonly used Fe-doped InP semiconductor, thereby realizing a reduction in capacitance.

  Furthermore, it is known that the capacity of the EA modulator section is inversely proportional to the undoped layer thickness in the multilayer structure, and the capacity of the EA modulator can be reduced as the undoped layer thickness is increased. Specifically, in this example, the value obtained by subtracting the thickness of the cladding layer from which the Zn diffused from the p-type from the total thickness of the active layer which is an undoped layer and the InGaAsP upper optical guide layer is the undoped layer. It becomes the thickness of. However, if the active layer that is an undoped layer or the upper guide layer is thickened, the f3 dB characteristic is improved, but the electric field strength applied to the undoped layer is reduced, so that holes in the modulator active layer are not swept out of the quantum well. In other words, a so-called pile-up phenomenon occurs, or the extinction characteristics and chirp characteristics, which are heavy parameters during optical fiber transmission, deteriorate.

Therefore, it is necessary to apply a sufficient electric field to the undoped layer including the active layer while ensuring a sufficient thickness of the undoped layer. Ideally, the thickness is about 8 × 10 ¹⁷ cm ⁻³ suddenly from the undoped layer. A structure that becomes a p-doping layer is desirable.

However, not only does the phenomenon that Zn in the InP clad layer added to apply a sufficient electric field to the undoped layer diffuses in the direction of the active layer of the undoped layer, not only changing the undoped layer to p-type, The doping profile is also a profile that gradually changes to a p-type from a steep one that is originally intended for securing electric field strength. Under these circumstances, in the conventional technique, the p-InP layer concentration just above the undoped layer is suppressed to 4 to 6 × 10 ¹⁷ cm ⁻³ in order to suppress excessive diffusion of Zn into the undoped layer.

M.M. Aoki, et. al. , “High-speed (10 Gbit / s) and low-drive-voltage (1V peak to peak) InGaAs / InGaAsP MQW electro-absorbed-in- ter-burned-in-between-DF Lett. , Vol. 28, pp. 1157-1158, 1992.

  As described in the related art, the EA modulator is designed to have a low capacity. However, when considering 43 Gbit / s operation that is a bit rate corresponding to the yield and FEC (forward error correction), etc. It is necessary to improve the characteristics including the f3 dB band.

  On the other hand, the p-type InP layer cladding layer crystal-grown on the EA modulator uses Zn as a dopant, and Zn doped in this p-InP cladding layer due to the influence of heat during crystal growth It has been found by SIMS (secondary ion mass spectroscopy) analysis that it diffuses into the lower undoped layer, the InGaAsP upper light guide layer.

Because of this diffusion, the p-InP layer concentration immediately above the undoped layer is suppressed to about 4 × 10 ¹⁷ cm ^{−3 in} order to secure the undoped layer thickness that determines the band characteristics.

On the other hand, in order to ensure pile-up strength, extinction characteristics, and chirp characteristics, the p-InP layer immediately above the undoped layer should be doped at about 6 × 10 ¹⁷ to 1 × 10 ¹⁸ cm ⁻³ , and abruptly doped from the undoped layer. A profile that increases the concentration is desirable.

  Therefore, Zn diffusion into the undoped layer greatly affects the f3 dB band, which is the main characteristic of the EA modulator, and the extinction characteristic. Therefore, the control of the Zn diffusion amount is a problem of characteristic improvement including the band improvement of the device. It has become.

  If Zn diffusion from the p-type cladding layer to the undoped upper guide layer can be suppressed and controlled, it is possible to stabilize the f3 dB band and the extinction characteristic, which are the aforementioned problems. In order to suppress Zn diffusion, it is possible to introduce an InGaAlAs layer between the undoped InGaAsP upper guide layer and the p-type InP cladding layer. As a result of comparing the characteristics of the EA modulator integrated laser with the case where the p-type cladding layer concentration and the multiple quantum well layer are the same and the InGaAlAs layer is not present in the trial production of the EA modulator integrated laser, the case where the InGaAlAs layer is present The result was that the capacity of the EA modulator was small, the f3 dB band was large, and the extinction ratio characteristic was small.

This indicates that Zn diffusion from the p-type cladding layer to the undoped upper guide layer is suppressed by the InGaAlAs layer in the case of the structure having the InGaAlAs layer. Further, in the results of evaluating the Zn concentration profile in the crystal stacking direction by SIMS analysis, as a result of comparing the Zn diffusion from the p-type cladding layer to the undoped upper guide layer, the case where there is an InGaAlAs layer and the case where there is no InGaAlAs layer In comparison, when compared with a Zn concentration of 1 × 10 ¹⁷ cm ⁻³ or more, it was confirmed that about 20 nm of Zn was suppressed from being diffused toward the undoped InGaAsP upper guide layer. This is due to the fact that in a compound semiconductor, the diffusion rate of Zn is slower in the order of higher saturation concentration, so that InGaAlAs has a slower diffusion rate of Zn than InP.

  Therefore, by introducing an InGaAlAs layer between the undoped InGaAsP upper guide layer and the p-type InP cladding layer, it becomes possible to suppress Zn diffusion into the undoped layer, and to improve characteristics such as stabilization of the f3 dB band and extinction characteristics. It becomes possible.

The InGaAlAs layer preferably has a carrier concentration of 1 to 10 × 10 ¹⁷ cm ⁻³ , a composition wavelength of 0.92 to 1.4 μm, and a thickness of 10 to 50 nm.

  According to the present invention, it is possible to improve the ability of Zn diffusion to the undoped layer in the crystal multilayer structure of the EA modulator portion to be improved, and it is possible to realize a broad band of the EA modulator and stabilization of main characteristics. In other words, by using the optical element to which the present invention is applied, it is possible to realize a broadband and high-speed operation of a semiconductor laser module, which is a main component of optical communication, and an optical transmission module at a low cost, and construction of a large-capacity optical network. Realize.

<Example 1>
The embodiment will be described by using an example of a ridge waveguide semiconductor electroabsorption modulator integrated laser having a transmission speed of 40 Gbit / s wavelength 1.55 μm band for 2 km optical transmission as in the prior art.

  1-3 show the 40 Gbit / s EA modulator integrated laser of this embodiment. 1 is a top view of the element, FIG. 2 is a cross section in the direction of the optical axis of the element (cross section taken along line cc ′ in FIG. 1), and FIG. 3 is a cross-sectional view in the vertical direction of the EA modulator section. Is shown.

  First, a method for manufacturing this EA modulator integrated laser will be described below.

On the n-InP semiconductor substrate 100, as the crystal growth of the EA modulator portion, an InP buffer layer 101, an InGaAsP lower light guide layer 102, and an InGaAlAs which is an active layer are formed by a known selective growth method using a metal organic vapor phase method. An undoped quantum well layer 103 composed of a well layer and a barrier layer, an undoped InGaAsP upper light guide layer 104, a p-type InGaAlAs diffusion prevention layer 201, and a 6 × 10 ¹⁷ cm ⁻³ p-type InP cap layer 105 are formed.

Here, unlike the conventional technique, a p-InP clad is formed by growing an InGaAlAs layer 201 having a composition wavelength of 1.15 μm p-doped to 6 × 10 ¹⁷ cm ⁻³ by 30 nm, thereby performing crystal growth in the subsequent steps. Diffusion from layer 111 is prevented.

  Thereafter, as described in the prior art, the laser multilayer and passive waveguide are integrated by the butt joint technology, and the semiconductor in the region to be the laser portion is subjected to the interference exposure method using a photolithograph. A diffraction grating is formed in the diffraction grating layer.

Furthermore, an 8 × 10 ¹⁷ cm ⁻³ p-type InP cladding layer 111, a p-type InGaAsP layer 112, a p-type InGaAs contact layer 113, and a p-type InP cap are simultaneously formed on the EA modulator, the passive waveguide, and the laser portion. Crystallize the layer.

  Here, Zn is used for the p-type dopant. Zn doped in the p-InP cap layer 105 and the p-InP clad layer 111 is diffused in the InGaAlAs layer 201 which is a diffusion prevention layer thereunder, despite the influence of heat during growth. It is suppressed, and Zn diffusion to the undoped upper light guide layer below it is suppressed. This was confirmed by SIMS analysis.

  Thereafter, a 40 Gbit / s EA modulator integrated laser was completed through the same process as the conventional technique.

  In the 40 Gbit / s EA modulator integrated laser of this example, since the InGaAlAs layer 30 nm for preventing Zn diffusion is inserted in the waveguide of the EA modulator portion, Zn diffusion into the undoped InGaAsP upper guide layer is prevented. It is suppressed. For this reason, the undoped layer thickness can be controlled to a desired thickness without variation for each lot, and the capacity of the EA modulator section can be reduced.

  When the InGaAlAs layer is inserted as in this embodiment, the EA modulator capacity has been confirmed to be reduced by about 0.05 pF compared to the lot without insertion, and it is possible to obtain an f3 dB band of 40-45 GHz. is there.

Furthermore, in the extinction cost characteristics, the concentration of the p-type InP cap layer 105 and the p-type InP cladding layer 111 is set to 6 × 10 ¹⁷ cm ⁻³ and 8 × 10 ¹⁷ cm ⁻³ by inserting the InGaAlAs layer, respectively. Further, the electric field strength applied to the active layer is increased, and the extinction characteristics, the chirp characteristics, and the pile-up strength characteristics can be improved.

  In this embodiment, the thickness of the InGaAlAs layer is set to 30 nm, but the present invention is not limited to this. If diffusion from the p-InP cladding layer can be suppressed, the same effect can be obtained even with a thickness of less than 30 nm. However, if it is too thin, it causes a quantum effect, and if it is too thick, holes tend to accumulate, so it is desirable that the thickness be in the range of 10 to 50 nm.

<Example 2>
Although the first embodiment has been described with the structure in which the EA modulator and the DFB laser are integrated on the same InP substrate, the present invention can also be applied to the case of the EA modulator alone.

  FIG. 4 shows the configuration. FIG. 4A is a top view. FIG. 4B is an e-e ′ cross-sectional view.

  As described in the first embodiment, the effect of inserting the InGaAlAs layer, which is the main point of the present invention, can be obtained in the EA modulator alone.

<Example 3>
The present embodiment will be described using an example of a 10 Gbit / s optical transmission wavelength 1.55 μm band ridge waveguide semiconductor electroabsorption modulator integrated laser.

  FIG. 5 is a vertical cross-sectional view of the EA modulator mesa of the electroabsorption modulator integrated laser in the present embodiment. Note that the top view is the same as FIG.

  This example is described in the literature “S. Makino,“ Wide Temperature Range (0 to 85 ° C.), 40-km SMF Transmission of a 1.55-mm, 10 Gbit / s InGaAlAs Electrobular Modulation Incorporation Modulation ”. This is an example of an EA modulator that operates in a wide temperature range as shown in “Conference (OFC'05), 2005, Postdeline-paper, PDP14”.

As the crystal growth of the EA modulator portion on the n-InP semiconductor substrate 300, an n-type InP buffer layer 301, an n-type InGaAsP lower light guide layer 302, an InGaAlAs well are formed by a known selective growth method using a metal organic vapor phase method. An undoped quantum well active layer 303 comprising a layer and a barrier layer, an undoped InGaAsP upper light guide layer 304, a p-type InGaAlAs diffusion prevention layer 305 having a composition wavelength of 1.15 μm and 6 × 10 ¹⁷ cm ⁻³ , and 6 × 10 6 A p-type InP cap layer 306 doped to ¹⁷ cm ⁻³ is formed.

Again, by growing a 30 nm thick InGaAlAs layer with a composition wavelength of 1.15 μm p-doped to 6 × 10 ¹⁷ cm ⁻³ , crystal diffusion is performed in the subsequent steps, preventing diffusion from the p-InP cladding layer 307 To do.

  Next, after forming a SiN film by plasma CVD, patterning is performed on the SiN film in the region to be the EA modulator portion, and EA other than the region to be the EA modulator portion is formed by wet etching or the like using this SiN film as a mask. Remove the modulator multilayer structure.

  Even if this etching method is dry etching, the effect of the present invention is not impaired. Thereafter, similarly to the first embodiment, the laser multilayer and the passive waveguide are integrated by the butt joint technique, and the semiconductor diffraction grating is formed by the interference exposure method using the photolithograph on the semiconductor of the region to be the laser portion. A diffraction grating is formed in the layer.

Thereafter, similarly to the first embodiment, the 8 × 10 ¹⁷ cm ⁻³ p-type InP cladding layer 307, the p-type InGaAsP contact layer 308, and the p-type InGaAs are simultaneously formed on the EA modulator, the passive waveguide, and the laser portion. Crystal growth is performed on the contact layer 309 and the p-type InP cap layer, and ridge waveguide formation, electrode formation, and the like are performed.

  Here, Zn doped in the p-type InP cap layer 306 and the p-InP clad layer 307 is prevented from being diffused by the underlying InGaAlAs layer 305, and the undoped upper light guide layer therebelow. Zn diffusion into the metal is suppressed.

  For this reason, the undoped layer thickness can be controlled to a desired thickness without variation for each lot, and the capacity of the EA modulator section can be reduced.

  When the InGaAlAs layer is inserted as in the present embodiment, the EA modulator capacity is confirmed to be reduced by about 0.05 pF compared to the lot without insertion, and the f3 dB band is also applied to the lot to which the InGaAlAs layer is applied. 15-18 GHz and good. Furthermore, since the p-doping concentration on the undoped layer can be increased as compared with the conventional case, the electric field strength applied to the active layer is increased, and the extinction characteristic, the chirp characteristic, and the pile-up strength characteristic can be improved.

  When this device was used, an excellent optical fiber transmission result of 40 km was confirmed in a laser temperature range of 0 to 80 ° C.

  Also in this embodiment, the example of the light source in which the EA modulator and the DFB laser are integrated is described, but it goes without saying that the effect of the present invention can be obtained even in the case of the EA modulator alone.

<Example 4>
This embodiment will be described using an example of a 10 Gbit / s optical transmission wavelength 1.55 μm band Fe buried waveguide type semiconductor electroabsorption modulator integrated laser.

  FIG. 6 is a cross-sectional view in the vertical direction of the EA modulator mesa of the Fe buried waveguide type semiconductor electroabsorption modulator integrated laser. The top view is substantially the same as FIG.

First, a method for manufacturing this EA modulator integrated laser will be described below. As in Example 1, as the crystal growth of the EA modulator portion on the n-InP semiconductor substrate 100, the InP buffer layer 400, the InGaAsP lower light guide layer 401, by a known selective growth method using a metal organic vapor phase method, An undoped quantum well active layer 402 composed of an InGaAlAs well layer and a barrier layer, an undoped InGaAsP upper optical guide layer 403, a composition wavelength of 1.15 μm p-doped 6 × 10 ¹⁷ cm ⁻³ , a p-type InGaAlAs diffusion prevention layer 404, 30 nm, and 6 A p-type InP cap layer 405 of × 10 ¹⁷ cm ⁻³ is formed.

  Thereafter, similarly to the first embodiment, the laser multi-layer and the passive waveguide are integrated by the butt joint technology, and the semiconductor diffraction grating is formed by the interference exposure method using the photolithograph on the semiconductor of the region to be the laser portion. A diffraction grating is formed in the layer.

Furthermore, an 8 to 10 × 10 ¹⁷ cm ⁻³ p-type InP cladding layer 406, a p-type InGaAsP layer 407, a p-type InGaAs contact layer 408, and a p-type are simultaneously formed on the EA modulator, the passive waveguide, and the laser portion. Crystal growth of the InP cap layer is performed.

After that, a mesa stripe-like SiO ₂ film pattern is formed from the EA modulator to the passive waveguide and the laser part, and the mesa stripe-like optical waveguide is dried using an ICP etcher or the like using this SiO ₂ film as a mask. It is formed by etching and wet etching.

Further, using an SiO ₂ film mask, burying growth is performed with an Fe-doped InP crystal and MOCVD is performed to form an Fe—InP burying layer 409 on both sides of the mesa stripe optical waveguide. At this time, since the InGaAlAs layer that is easy to oxidize is provided as the semiconductor of the optical waveguide, a process that removes the Al oxide film during the optical waveguide etching and Fe-InP regrowth, for example, Fe-InP buried growth The previous chlorine addition treatment or the like is necessary to ensure device reliability.

  Subsequently, a passivation film 410 is formed, and thereafter, an Fe buried waveguide type semiconductor electroabsorption modulator integrated laser is obtained through a manufacturing process similar to the conventional technique.

  Here, the Zn doped in the p-InP cladding layer is suppressed from being diffused in the InGaAlAs layer 404 which is the diffusion prevention layer below it, and the Zn diffusion to the undoped upper light guide layer below is suppressed. . For this reason, the undoped layer thickness can be controlled to a desired thickness without variation for each lot, and the capacity of the EA modulator section can be reduced.

  In this example, since the Al-based material is formed by mesa stripes, the process until the burying growth with Fe-InP requires more technology than in Example 1 or 2, but in the Fe-InP layer By embedding and protecting the optical waveguide, the optical waveguide becomes physically strong, which may be an advantage in terms of reliability.

  In general, Fe-InP buried regrowth involves growth of several tens of minutes or more at a temperature of about 500 to 600 ° C., so that a large amount of Zn diffuses from the p-InP cladding layer to the undoped upper light guide layer. However, the introduction of the InGaAlAs layer of the present invention makes it possible to suppress the Zn diffusion even when the Zn diffusion is likely to occur as compared with the case without the InGaAlAs layer.

<Example 5>
An example in which the present invention is applied to a semiconductor optical device module will be described with reference to FIG.

  The 40 Gbit / s optical transmission EA modulator integrated laser 501 described in the first embodiment is mounted on a chip carrier 502 with a termination resistor and a high frequency design.

  Subsequently, the Peltier substrate 503, the lens 504, the monitor photodiode 505, the thermistor 506, the isolator 507, and the optical fiber 508 are mounted in the same package.

  Regarding the f3 dB band, which is an important characteristic of the semiconductor optical element module, the high frequency design of this module is good, so that the f3 dB band of the EA modulator integrated laser is almost the same as the f3 dB band of the element module.

  When an EA modulator integrated laser manufactured by applying the present invention is used, a low-capacity modulator without introducing Zn as a p-type dopant into an undoped layer such as an upper guide layer by introducing an InGaAlAs layer Can be stably produced.

  For this reason, an element module equipped with this has an f3 dB band characteristic of 40-45 GHz. For this reason, the optical waveform was a good eye pattern with a large margin with respect to the ITU-T standard mask.

  For example, when mounted on a CAN-type package that does not have a temperature adjustment function instead of the component configuration as in the present embodiment, f3 dB and extinction ratio characteristics are obtained, and similar effects are obtained.

<Example 6>
A highly reliable optical transmission device with stable characteristics can be realized by incorporating the optical transmission module to which the present invention is applied, which is shown in the fifth embodiment, into an optical transmission device or a router.

<Example 7>
The case where the present invention is applied to a 105 Gbit / s optical transmission wavelength 1.55 μm band Mach-Zehnder optical modulator will be described with reference to FIG.

  FIG. 8A is a top view. FIG. 8B is a cross-sectional view in the f-f ′ direction of FIG.

  This device is represented by, for example, the document “H. Sano,“ High-speed InGaAs / InAlAs MQW Mach-Zehnder-type modulator ”, OFC / IOOC '93, ThK5, San Jose, CA, Feb. 1993. It is.

An InAlAs layer 701 n-doped to 5 × 10 ¹⁸ cm ⁻³ is crystal-grown on the n-InP substrate 700, and then a multiple quantum well layer comprising 20 periods of a 6 nm undoped InGaAs well layer and a 6 nm undoped InAlAs barrier layer 702 is produced.

Next, an InAlAs clad layer in which a p-InGaAlAs diffusion prevention layer 703 having a composition wavelength of 1.15 μm p-doped to 6 × 10 ¹⁷ cm ⁻³ is stacked in a thickness of 30 nm, and 8 × 10 ¹⁷ cm ⁻³ is p-doped with Zn. Crystal growth of 704 to 1.5 μm and p-InGaAs contact layer 705 is performed.

  After that, the optical waveguide 706 of the Mach-Zehnder modulator shown in FIG. 8 is processed by dry etching using a silicon oxide film as a mask, embedded with a passivation film 707 and a polyimide resin 708, and an electrode is formed. Complete.

  Also in this modulator, it is possible to secure a sufficient band for 10G optical transmission due to the effect of the p-InGaAlAs diffusion prevention layer 703.

<Example 8>
The case where the present invention is applied to a directional coupler having a wavelength of 1.55 μm for 10 Gbit / s optical transmission will be described with reference to FIG.

  FIG. 9A is a top view. FIG. 9B is a cross-sectional view in the g-g ′ direction of FIG.

  This device is described in, for example, the document “JE Zucker,“ Compact directive coupler switches using quantum well electrorefraction ”. APPl. Phys. Lett. Co. 1989, 27 November. Modulator ", IQEC and CLEO / PR 2005, CWJ1-5-INV, Japan, Jul. 2005." is an example of an optical element using Al-based MQW on an n-InP substrate.

An InP layer 801 n-doped to 5 × 10 ¹⁸ cm ⁻³ is crystal-grown on the n-InP substrate 800, and then a multi-quantum well layer 802 composed of 10 periods using undoped InGaAlAs is fabricated. Next, an InP clad layer in which a p-InGaAlAs diffusion prevention layer 803 having a composition wavelength of 1.15 μm p-doped to 6 × 10 ¹⁷ cm ⁻³ is stacked in a thickness of 30 nm, and 8 × 10 ¹⁷ cm ⁻³ is p-doped with Zn. Crystal growth of 804 to 1.5 μm and p-InGaAs contact layer 805 is performed.

  Thereafter, the optical waveguide 806 of the directional coupler shown in FIG. 9 is processed by dry etching using a silicon oxide film as a mask, embedded with a passivation film 807 and a polyimide resin 808, and an electrode is formed to complete the device. To do.

  Also in this directional coupler, Zn diffusion from the p-InP cladding layer to the undoped multiple quantum well layer can be suppressed by the effect of the p-InGaAlAs diffusion preventing layer 803. For this reason, a sufficient electric field strength can be applied to the undoped layer, and it is possible to secure a quenching characteristic and a band sufficient for 10G optical transmission.

  In this example, an example using an n-InP substrate has been described. However, even when a device is manufactured using a semi-insulating substrate doped with Fe, for example, the structure above the multiple quantum well layer is equivalent. In some cases, similar effects can be obtained.

<Example 9>
The present embodiment will be described using an example of an electroabsorption modulator using an electrode structure with a transmission speed of 40 Gbit / s semiconductor impedance control for optical transmission.

  FIG. 10A is a top view. FIG. 10B is a cross-sectional view taken along the line a-a ′ in FIG.

  The 40 Gbit / s semiconductor electroabsorption modulator (hereinafter referred to as EA (electro-absorption) modulator) has the same principle of turning on and off the light as the EA modulator described in the prior art 1.

  This embodiment is different in the method of applying a high-frequency electric field to the modulator. For example, “M. shirai, et. Al. Proc. 28th European Conference on Optical Communications (ECOC2002), 9.5.4,“ IMPEDANCE- CONTROL-ELECTRODE (ICE) SEMICONDUCTOR MODURATORS FOR 1.3-μm-40-Gbit / s TRANCEIVERS “.” Is an EA modulator using an impedance-controlled electrode structure.

  The impedance-controlled electrode structure realizes higher-speed operation because the CR time constant of the entire modulator does not limit the operating frequency by matching the traveling direction of the microwave with the traveling direction of light. is there.

In this device, an n-InP buffer layer 902 is formed on a Fe-doped InP semiconductor substrate 901 using a metal organic vapor phase method, and an undoped quantum well comprising an InGaAsP lower light guide layer 903, an InGaAlAs well layer and a barrier layer. Active layer 904, undoped InGaAsP upper light guide layer 905, 6 × 10 ¹⁷ cm ⁻³ p-doped composition wavelength 1.15 μm p-type InGaAlAs diffusion prevention layer 906 30 nm, and 6 × 10 ¹⁷ cm ⁻³ p-type InP cap Layer 907 is formed.

  Subsequently, a mask is formed in a region to be the EA modulator optical waveguide 908, and the active layer in other regions is removed by dry etching.

  Thereafter, in order to form a passive optical waveguide 909 before and after the EA modulator waveguide by performing a cleaning process, a passive optical waveguide layer 910 including an InGaAsP layer is formed by a metal organic vapor phase method, and the EA modulator part activity A butt joint connection is made so as to be optically continuous with the layer 904.

  Next, a p-InP clad layer 911 and a contact layer 912 are formed by a crystal growth process, and the semiconductor crystal in the high impedance line portion 913 is completely removed to the n-InP buffer layer 902 by etching, and an Fe-InP substrate is formed. To expose. Subsequently, a 2 μm-wide optical waveguide composed of the EA modulator optical waveguide 908 and the passive optical waveguide 909 is formed by dry etching. Furthermore, since this device has a structure in which both the p-type and n-type electrodes are taken from the element surface, the Fe—InP buried layer in the n-electrode 914 formation region is removed by dry etching to expose the n-InP buffer layer. .

  Thereafter, after forming a passivation film 915, forming a through hole, and forming a p-type electrode 916 and an n-type electrode 914 formed immediately above the high impedance line portion 913 and the EA modulator portion optical waveguide 908, as shown in FIG. An impedance-controlled EA modulator is obtained.

  Here, in the high impedance line portion 913, the waveguide width and the distance to the n-type electrode that is the ground are optimized in consideration of impedance matching with the driver IC.

  As shown in FIG. 11, the above-described element 920 is mounted on a chip carrier 921 designed for high frequency, and a high frequency substrate 922 including a termination resistor, a Peltier 923, a lens 924, and an input / output optical fiber 925 are mounted in the same package. A semiconductor element module is configured.

  This module also includes a driver IC 926 for driving the EA modulator. This is because the driver IC output amplitude is transmitted to the EA modulator without being attenuated as much as possible by shortening the distance from the driver IC to the EA modulator. However, if the driver IC output amplitude is sufficient, there is no need to incorporate a driver IC, and the same effect is expected.

  In the EA modulator mounted on the EA modulator module, an InGaAlAs layer 30 nm for preventing Zn diffusion is inserted into the waveguide of the EA modulator portion, so that Zn diffusion into the undoped InGaAsP upper guide layer is prevented. It is suppressed. For this reason, the undoped layer thickness can be controlled to a desired thickness without variation among lots, the capacity of the EA modulator can be reduced, and the effects shown in the first embodiment can be obtained.

In the present embodiment, the EA modulator alone is described, but the same effect can be obtained also in an impedance control type EA modulator integrated laser integrated with a DFB laser.
The EA modulator optical element module is configured by mounting the EA modulator, the Peltier, the lens, the optical fiber, and the driver IC for driving the modulator in the same package.

The element top view explaining Example 1 by the present invention Element optical axis direction sectional view explaining Example 1 by the present invention EA modulator section mesa vertical sectional view for explaining the first embodiment of the present invention Element top view explaining Example 2 by this invention, and EA modulator part mesa vertical direction sectional drawing EA modulator section mesa vertical sectional view illustrating Embodiment 3 according to the present invention EA modulator section mesa vertical sectional view illustrating Embodiment 4 of the present invention Element sectional drawing explaining Example 5 by this invention Element sectional drawing explaining Example 7 by this invention Device sectional view illustrating Example 8 according to the present invention Device sectional drawing explaining the Example 9 by this invention, and a top view Semiconductor device module diagram illustrating Example 9 according to the present invention Element top view explaining conventional technology Cross-sectional view in the optical axis direction of element ridge type optical waveguide explaining conventional technology EA modulator section mesa vertical direction sectional view explaining conventional technology

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 ... n-InP board | substrate, 101 ... n-type InP buffer layer, 102 ... InGaAsP lower side light guide layer, 103 ... Quantum well active layer, 104 ... InGaAsP upper side light guide layer, 105 ... p-type InP cap layer, 106 ... Laser Part InGaAsP lower light guide layer 107 107 Laser part quantum well layer 108 Laser part InGaAsP upper light guide layer 109 InP spacer layer 110 Diffraction grating layer 111 Optical waveguide part InGaAsP layer 112 P type InP cladding layer, 113 ... InGaAsP contact layer, 114 ... InGaAs contact layer, 115 ... passivation film, 116 ... polyimide resin, 117 ... p-side electrode, 118 ... n-side electrode,
201 ... p-type InGaAlAs diffusion prevention layer,
301 ... n-type InP buffer layer, 302 ... InGaAsP lower light guide layer, 303 ... undoped quantum well active layer, 304 ... InGaAsP upper light guide layer, 305 ... p-type InGaAlAs diffusion prevention layer, 306 ... p-type InP cap layer, 307 ... p-type InP cladding layer, 308 ... InGaAsP contact layer, 309 ... InGaAs contact layer, 310 ... passivation film, 311 ... polyimide resin, 312 ... p-side electrode, 313 ... n-side electrode,
400 ... n-type InP buffer layer, 401 ... InGaAsP lower light guide layer, 402 ... undoped quantum well active layer, 403 ... InGaAsP upper light guide layer, 404 ... p-type InGaAlAs diffusion prevention layer, 405 ... p-type InP cap layer, 406 ... p-type InP cladding layer, 407 ... InGaAsP contact layer, 408 ... InGaAs contact layer, 409 ... Fe-InP buried layer, 410 ... passivation film, 411 ... p-side electrode, 312 ... n-side electrode,
502 ... chip carrier, 503 ... Peltier substrate, 504 ... lens, 505 ... monitor photodiode, 506 ... thermistor, 507 ... isolator, 508 ... optical fiber 700 ... n-InP substrate, 701 ... n-doped InAlAs layer, 702 ... multiple quantum Well layer, 703... P-InGaAlAs diffusion prevention layer, 704... P-doped InAlAs cladding layer, 705... P-InGaAs contact layer, 706. ... p-side electrode, 710 ... n-side electrode 901 ... Fe-doped InP semiconductor substrate, 902 ... n-InP buffer layer, 903 ... InGaAsP lower light guide layer, 904 ... undoped quantum well active layer, 905 ... undoped InGaAsP upper light guide layer 906 ... p-type InGaAlAs diffusion prevention layer, 907 ... p-type InP cap layer, 908 ... EA modulator optical waveguide, 909 ... passive optical waveguide, 910 ... passive optical waveguide layer, 911 ... p-InP cladding layer, 912 ... and contact Layer, 913... High impedance line portion, 914... N electrode, 915... Passivation film, 916 .. p-type electrode, 920... EA modulator 921 according to this embodiment ... chip carrier, 922 ... high frequency substrate, 923 ... Peltier, 924. Lens, 925 ... Input / output optical fiber, 926 ... Driver IC

Claims (7)

A semiconductor optical device comprising a compound semiconductor laminated body having an undoped active layer containing Al and a compound semiconductor layer containing zinc on an InP substrate,
A semiconductor optical device, wherein an InGaAlAs layer is laminated between the undoped active layer containing Al and the compound semiconductor layer containing zinc. In claim 1,
The semiconductor optical device, wherein the InP substrate is an n-type or semi-insulating InP substrate. In claim 1,
The semiconductor optical device, wherein the InGaAlAs layer is doped p-type. The semiconductor optical device according to any one of claims 1 to 3, wherein
A semiconductor optical device comprising a semiconductor electroabsorption modulator, a semiconductor Mach-Zehnder modulator, or any of these integrated.   5. A semiconductor element module on which the semiconductor optical element according to claim 4 is mounted.   An optical transmission module or an optical transmission / reception module on which the semiconductor element module according to claim 5 is mounted. 7. An optical transmission device comprising the optical transmission module or the optical transmission / reception module according to claim 6.

Images