JP5497678B2 - Semiconductor optical integrated device - Google Patents

Description

The present invention relates to a semiconductor optical integrated device, and more For details relates to a semiconductor optical integrated device monolithically integrated LD element and high-speed, high-efficiency optical modulator in the reverse mesa direction.

A transmitter used in a high-speed and large-capacity optical communication system usually employs a method of generating an optical signal by combining a laser diode (LD) light source and an external modulator. A typical external modulator used for this type of purpose is an LN modulator fabricated with a LiNbO ₃ (LN) waveguide. Modulation of the refractive index by the electro-optic effect is the basic operation, and there are a light intensity modulator including a Mach-Zehnder interferometer in addition to a simple optical phase modulator.

  Recently, semiconductor optical modulators that are more advantageous than LN modulators in terms of miniaturization have attracted attention. In particular, a semiconductor optical modulator that can be manufactured using the same material as the LD is excellent in that a large-capacity integrated device can be integrated in a small and monolithic manner.

  The semiconductor optical modulator uses a hetero pin junction, and an InP / InGaAsP optical modulator that effectively applies a voltage to the core portion of the waveguide along with light confinement, and further low-voltage driven optical modulation. An npin type semiconductor optical modulator structure in which both InP cladding layers are made n-type and a thin p-type semiconductor layer (p-type barrier layer) is inserted as a barrier layer for suppressing electron current in order to realize a device. It has been proposed (see, for example, Patent Document 1). Since this npin type does not use a p-type cladding layer that causes optical loss, it is possible to use a relatively long waveguide, which is advantageous in lowering the driving voltage. Further, since there is a degree of freedom that the thickness of the depletion layer can be arbitrarily optimized, it is easy to satisfy the matching of the electric impedance and the matching of the electric speed / light speed at the same time, and it is advantageous for speeding up.

  In the pin and npin types, in addition to the primary electro-optic effect (Pockels effect), the semiconductor-specific Franz-Keldish effect (FK effect), and the quantum confined Stark effect due to the core structure having a multiple quantum well structure By using (QCSE) at the same time, a highly efficient optical modulation operation is possible.

Here, a description of FIG. 6 will be given. FIG. 6 shows the refractive index ellipsoid change due to the Pockels effect when the electric field direction as the reverse bias is [100], that is, when the electric field is applied upward of the substrate. Note that the polarization of propagating light is TE polarization. Normally, zinc-blende crystals such as InP and GaAs are isotropic (n ₀ ) regardless of crystal orientation when no electric field is applied ((1) in the figure) When an electric field is applied to the crystal, a refractive index change peculiar to the crystal occurs in accordance with the direction of application ((2) in the figure) When TE polarized light is propagated in the forward mesa direction, TE polarized light is Pockels. On the other hand, when TE polarized light is propagated in the reverse mesa direction, it propagates in a region where the refractive index is small due to the Pockels effect. In the figure, r ₄₁ is an electro-optic constant.

  While the Pockels effect described above varies in refractive index change depending on the light propagation direction, the FK effect and QCSE change only in accordance with the electric field strength applied to the crystal. That is, when a reverse bias is applied, the refractive index change shifts in the direction of increasing as the electric field strength increases regardless of the forward mesa direction or the reverse mesa direction. FIG. 7 is a diagram in which the total sum of the refractive index changes including the increase and decrease of the refractive index due to the Pockels effect is normalized in addition to the FK effect and the QCSE increase due to voltage application. From FIG. 7, in order to increase the refractive index change with applied voltage, that is, to obtain a high-efficiency modulation operation, the large refractive index change characteristic shown in [1] in FIG. 7 in which the FK effect, QCSE, and Pockels effect are combined. Need to be realized.

FIG. 8 shows a cross-sectional structure of an npin type semiconductor optical modulator 10 which is a conventional technique as an example.
As shown in FIG. 8, an n-InGaAs contact layer 2 is disposed on an SI (semi-insulating) -InP substrate 1, and an electrode 3 is disposed on the contact layer 2. The electrodes 11a and 11b, n-InGaAs contact layers 12a and 12b, n-InP layers 13a and 13b, p-InAlAs layers 14a and 14b, i-light confinement layers 15a and 15b, i-core layers 16a and 16b, i A semiconductor optical modulator 10 is configured by providing a semiconductor multilayer structure (optical modulation element) formed by stacking optical confinement layers 17a and 17b and n-InP layers 18a and 18b on a substrate 1 (contact layer 2). Has been.

  In high-speed optical modulation operations in the pin type and npin type, it is necessary to apply a reverse bias so that a high electric field is applied to the core layers 16a and 16b. In this case, a negative bias is applied to the electrode 11a and the electrode 11b. Apply voltage. That is, a bias electric field is applied in the [100] direction, and the refractive index ellipse change due to the Pockels effect is as shown in FIG. On the other hand, as the applied voltage increases, the refractive index change due to the FK effect and QCSE shifts in the direction in which the refractive index increases, so that the Pockels effect is synergized with this to change the refractive index (see [1] in FIG. 7). 6), it is necessary to fabricate the light modulation element (semiconductor multilayer structure) in the forward mesa direction from FIG. On the other hand, when manufactured in the reverse mesa direction, the direction of increase / decrease in the refractive index due to the FK effect and QCSE and the direction of increase / decrease due to the Pockels effect are opposite to each other. [3] in FIG.

  In an LD that requires high element reliability, an embedded structure in which the periphery of an active layer serving as a light source is embedded with a cladding layer is often used. However, if an attempt is made to perform burying regrowth on the forward mesa direction LD element, the regrowth also proceeds on a surface other than the (100) surface, which is the upper surface, and it becomes difficult to produce a desired embedded structure. ing. On the other hand, since the growth rate to the (100) plane is dominant for the reverse mesa LD element, a desired buried structure can be manufactured. For this reason, highly reliable LD elements are often fabricated in the reverse mesa direction.

  Therefore, for example, when a semiconductor optical modulation element for high-speed and large-capacity communication manufactured in the forward mesa direction is integrated with an embedded LD as an optical integrated element, a method of hybrid integration of each element, an LD element and an optical A method of monolithically integrating modulation elements arranged orthogonally has been used.

JP 2005-099387 A JP 2009-198881 A Japanese Patent No. 3230785 (Japanese Patent Laid-Open No. 07-135369)

J. Pamulapati et al, "Refractive index and electro-optic effect in compressive and TEnsile strained quantum wells", Journal of Applied Physics, Vol. 69, pp. 4071-4074, (1991).

  However, when the LD element and the light modulation element are hybrid-integrated, the magnitude of the optical coupling loss between the elements becomes a problem. Therefore, a semiconductor optical amplifier (SOA) is provided in front of the light modulation element to compensate for the light loss. (For example, refer to Patent Document 2). When the LD element and the light modulation element are arranged orthogonally and monolithically integrated, the area of the entire element becomes unnecessarily large due to the orthogonal arrangement, and many problems remain from the viewpoint of element mass productivity. Against this background, there are still many problems in fabricating a semiconductor optical integrated device for high-speed and large-capacity communication in a small size and at low cost.

The present invention has been made in view of such a problem, and nip-type or nipn-type semiconductor optical modulation elements capable of high-speed modulation operation in the same reverse mesa direction as LD elements and LD elements are monolithically integrated. Accordingly, an object of the present invention is to provide a high-speed and large-capacity semiconductor optical integrated device in a small size and at a low cost.

The configuration of the semiconductor optical integrated device of the present invention that solves the above problems is as follows.
A semiconductor multilayer including at least an n-type cladding layer, an i-core layer, and a p-type cladding layer on a substrate surface equivalent to the (100) plane of a zinc blende semi-insulating semiconductor crystal substrate in order from the upper layer to the substrate surface An MZ type light modulation element having a structure, wherein the light modulation waveguide of the MZ type light modulation element is formed in a reverse mesa direction which is a [011] plane direction with respect to the (100) plane of the substrate, and A laser element is mounted on the substrate in a reverse mesa direction which is a [011] plane direction with respect to the (100) plane of the substrate.

The configuration of the semiconductor optical integrated device of the present invention is as follows.
On the substrate surface equivalent to the (100) plane of the zincblende semi-insulating semiconductor crystal substrate, at least an n-type cladding layer, an i-core layer, a p-type carrier block layer, and an n-type in order from the upper layer to the substrate surface An MZ type optical modulation element having a semiconductor multilayer structure including a cladding layer is provided, and the optical modulation waveguide of the MZ type optical modulation element is in a reverse mesa direction which is a [011] plane direction with respect to the (100) plane of the substrate In addition, the laser element is mounted on the substrate in a reverse mesa direction that is a [011] plane direction with respect to the (100) plane of the substrate.

The configuration of the semiconductor optical integrated device of the present invention is as follows.
The MZ type optical modulation element is applied with a reverse bias electric field from the upper layer toward the substrate surface, and the amplitude of the signal voltage so that a positive voltage is always applied to the signal electrode of the MZ type optical modulation element. A larger offset voltage is applied.

  Finally, in order to achieve the above object, the semiconductor light modulation device according to the present invention is a light modulation device in the reverse mesa direction with respect to a semiconductor wafer laminated in the order of nip or nipn from the upper layer toward the substrate side. A waveguide is produced. Further, a positive voltage is used as a bias voltage for the signal electrode in order to apply a reverse bias, so that the directions of the refractive index increase / decrease of the FK effect, QCSE, and Pockels effect accompanying the increase in the applied voltage are made equal.

Normally, zinc-blende crystals such as InP and GaAs are isotropic (n ₀ ) regardless of crystal orientation when no electric field is applied ((1) in the figure). When an electric field is applied to the crystal, a refractive index change peculiar to the crystal corresponding to the applied direction occurs ((2) in the figure). When TE polarized light is propagated in the forward mesa direction, the TE polarized light propagates in a region where the refractive index is reduced by the Pockels effect. On the other hand, when TE polarized light is propagated in the reverse mesa direction, it propagates in a region where the refractive index is increased by the Pockels effect. That is, in the case of FIG. 9, the refractive index change in which the FK effect and the QCSE are combined with the Pockels effect can be obtained by producing the light modulation waveguide of the light modulation element in the reverse mesa direction ([[ 1]).

  According to the present invention, it is possible to produce semiconductor optical modulation elements having the same characteristics as those of npin type and pin type optical modulators conventionally used as high-speed and high-efficiency optical modulators in the reverse mesa direction. As a result, a small, high-speed and large-capacity semiconductor optical integrated device can be manufactured with low loss and mass production by monolithically integrating the LD manufactured in the reverse mesa direction and the semiconductor optical modulation device.

FIG. 6 is a top view showing a semiconductor light modulation element of Reference Example 1 . FIG. 4 is an enlarged cross-sectional view showing a modulation region of a semiconductor light modulation element of Reference Example 1 . The characteristic view which shows the relationship between a bias voltage and a signal voltage when a voltage is applied in a forward bias direction when a modulation signal is applied. The characteristic view which shows the relationship between a bias voltage and a signal voltage at the time of a voltage application to a reverse bias direction at the time of the application of a modulation signal. FIG. 6 is a top view showing a semiconductor light modulation element of Reference Example 2 . FIG. 6 is an enlarged cross-sectional view showing a modulation region of a semiconductor light modulation element of Reference Example 2 . FIG. 3 is a top view showing the semiconductor optical integrated device according to the first embodiment. FIG. 3 is an enlarged sectional view showing a modulation region of the semiconductor optical integrated device according to the first embodiment. FIG. 6 is a top view showing a semiconductor optical integrated device according to a second embodiment. The characteristic view which shows the refractive-index ellipsoid change by the Pockels effect when an electric field is applied toward the board | substrate upper direction. The characteristic view which normalizes and shows the sum total of the refractive index variation | change_quantity including the refractive index increase / decrease by a Pockels effect in addition to the FK effect accompanying voltage application, and the refractive index increase of QCSE. Sectional drawing which shows the cross-section of an example of the npin-type semiconductor optical modulator which is a prior art. The characteristic view which shows the refractive-index ellipsoid change by the Pockels effect when the electric field direction as a reverse bias is applied toward a substrate downward direction.

  Embodiments of the present invention will be described with reference to the accompanying drawings. The embodiments described below are examples of the present invention, and the present invention is not limited to the following embodiments. In addition, in this specification and drawing, the component with the same code | symbol shall show the mutually same thing.

Reference example 1

FIG. 1A is a top view of the semiconductor optical modulation element 100 of Reference Example 1 , and FIG. 1B is a cross-sectional view in the modulation region.
As shown in FIGS. 1A and 1B, an n-buffer layer 102 is disposed on an SI (semi-insulating) -InP substrate 101, and a p-InGaAs contact layer 103 is disposed on the buffer layer 102. The electrode 104 is disposed on the contact layer 103. From the upper layer toward the substrate surface, electrodes 111a and 111b, n-InGaAs contact layers 112a and 112b, n-InP layers 113a and 113b, i-light confinement layers 114a and 114b, i-core layers 115a and 115b, i A semiconductor multilayer structure (light modulation element) formed by stacking optical confinement layers 116a and 116b and p-InP layers 117a and 117b is provided on the substrate 101 (contact layer 103), and the semiconductor light according to Reference Example 1 A modulation element 100 is configured.
As described above, in this example, the semiconductor multilayer structure (light modulation element) is stacked in the nip order from the upper layer toward the substrate surface, and the light modulation waveguide of the light modulation element is formed in the reverse mesa direction. Yes.
Reference numeral 120 denotes an optical waveguide, and 121 denotes an MMI coupler.

  Crystal growth is performed by a metal organic chemical vapor deposition (MOVPE) method suitable for a crystal regrowth process, and a (100) plane orientation substrate is used as a substrate crystal. The core layers 115a and 115b in the light modulation region are non-doped layers, and the band gap wavelengths and structures of the core layers 115a and 115b are determined so that the electro-optic effect works effectively at the operating light wavelength and the light absorption does not become a problem. ing. For example, in the case of a 1.55 μm band device, semiconductor materials having different band gaps such as InGaAlAs and InAlAs are used for the quantum well layer and the barrier layer, but the material is not limited to InGaAlAs and InAlAs. InGaAlAs, InGaAs, InGaAsP, InP, or the like may be used for the layers and barrier layers.

The upper and lower surfaces of the core layers 115a and 115b in the light modulation region have a band gap larger than that of the core layer so that carriers generated by light absorption are not trapped at the heterointerface, such as InGaAsP. Light confinement layers 114a, 116a, 114b, and 116b having a composition are provided. For example, n-type InP cladding layers 113a, 113b, n-type InGaAs 112a, 112b for signal electrode contact, and p-type for ground electrode contact are provided on the lower surfaces of the optical confinement layers 116a, 116b. A type InGaAs 103 and p-type InP cladding layers 117a and 117b are sequentially laminated.
The doping concentration of these layers is 5 × 10 ¹⁷ cm ⁻³ or more for the n-type and p-type InP cladding layers so that a voltage drop is efficiently generated in the non-doped layer, and for electrode contact so that sufficient ohmic contact can be obtained. It is desirable that the n-type and p-type InGaAs electrode contact layers be 1 × 10 ¹⁹ cm ⁻³ or more. For example, the doping concentration of the n-type and p-type InP cladding layers is 1 × 10 ¹⁸ cm ^−3, and the doping concentration of the n-type and p-type InGaAs electrode contact layers is 2 × 10 ¹⁸ cm ⁻³ and 1 × 10 ¹⁹ cm, respectively. ^-3 . Further, for example, an n-type InP layer is stacked as a buffer layer between the semi-insulating InP substrate and the p-type InGaAs layer.
It should be noted that the layer laminated for the electrode contact is not a problem as long as sufficient conductivity can be ensured. Therefore, the semiconductor doped with n-type and p-type impurities is not limited to InGaAs, and for example, InGaAsP may be used. .

After depositing the nip layer, the p-type InGaAs electrode contact layer 103 is exposed by dry etching using, for example, a SiO ₂ mask in which a waveguide shape is simulated in the reverse mesa direction so that the semiconductor light modulation element functions as an optical waveguide. Then, the mesa structure is fabricated. Thereby, the light modulation waveguide of the light modulation element is formed in the reverse mesa direction. At this time, the region that does not contribute to voltage application is removed by dry etching including the p-type InGaAs electrode contact layer 103 to expose the semi-insulating InP substrate 101. After removing the SiO ₂ mask, an Au electrode 104 is formed on the p-type and n-type InGaAs electrode contact layer 103 as shown in FIG.
Note that the signal electrode used in the semiconductor optical modulation element 100 preferably has a traveling wave electrode structure, but other than this, for example, a lumped constant type and a resonance type electrode structure may be used.

In order to function as an optoelectronic waveguide (light modulation waveguide), for example, as shown in FIG. 1A including two MMI couplers including a mesa structure having a cross section as shown in FIG. A Mach-Zehnder (MZ) waveguide structure is used, and a reverse bias voltage and a signal voltage are input to the electrodes 111a and 111b in a state where light is propagated through the waveguide formed in the reverse mesa direction. That is, a Mach-Zehnder (MZ) type optical waveguide 120 and an MMI coupler 121 are provided on the substrate 101 (contact layer 103), an optical modulation element is connected to the optical waveguide 120, and a reverse bias voltage is applied. In order to apply a high electric field to the non-doped layer, it is necessary to apply a voltage in the reverse bias direction. In the case of the nip layer structure, as shown by the arrow in FIG. 1B, applying a bias voltage downward on the substrate side corresponds to the reverse bias voltage. In addition, if a voltage is applied in the forward bias direction ([100] direction in FIG. 1B) when a modulation signal is applied, not only a desired voltage drop does not occur, but also carriers are generated with respect to the non-doped layer. Since it is injected, it does not function as a light modulation element in the first place (FIG. 2A). Therefore, in order to prevent an electric field from being generated in the forward bias direction due to the amplitude of the signal voltage, the modulation operating point is shifted by a reverse bias larger than that, and modulation is always performed with the electric field applied to the reverse bias applied (see FIG. 2 (b)). For example, when the amplitude of the signal voltage is driven at ± 2V, the reverse bias voltage is applied at 2V or more.
Further, in order to reduce the power consumption and suppress the modulation chirping, the MZ optical modulation element is required to perform a push-pull drive modulation operation. For this reason, as modulation signals, signal voltages having the same intensity and the opposite direction of the electric field displacement are input to the electrodes 101a and 101b at the same timing.
In addition, it is clear that the present reference example is useful even when it does not have an MZ waveguide structure, for example, only a linear waveguide structure is used as a phase modulation element.

  As described above, according to the present invention, since the electric field is applied as the reverse bias downward, the Pockels effect can be combined with the light modulation element manufactured in the reverse mesa direction in addition to the FK effect and the QCSE. Therefore, it is possible to obtain a large refractive index change and good modulation characteristics shown in [1] in FIG.

Here, it is clear that the usefulness of the substrate material does not change even when GaAs, GaP, ZnS, ZnSe, for example, has the same structure other than InP.
In this reference example , the semiconductor light modulation element corresponding to the 1.55 μm wavelength band is used, but the element corresponding to the 1.3 μm wavelength band may be used.
For example, if GaAs is used, it can respond to a wavelength range of 0.6 to 1.3 μm.

Reference example 2

A semiconductor light modulation device 100a according to Reference Example 2 will be described in which the semiconductor layer in the light modulation region described in Reference Example 1 is changed from the nip structure to the nipn structure with the aim of reducing light propagation loss. FIG. 3A is a top view of the semiconductor light modulation device 100a of this reference example , and FIG. 3B is a cross-sectional view of the light modulation region.
As shown in FIGS. 3A and 3B, an n-InGaAs contact layer 105 is disposed on the SI (semi-insulating) -InP substrate 101, and an electrode 104 is disposed on the contact layer 105. . From the upper layer toward the substrate surface, electrodes 111a and 111b, n-InGaAs contact layers 112a and 112b, n-InP layers 113a and 113b, i-light confinement layers 114a and 114b, i-core layers 115a and 115b, i A semiconductor multilayer structure (light modulation element) formed by stacking optical confinement layers 116a and 116B, p-InAlAs layers 121a and 121b, and n-InP layers 122a and 122b is provided on the substrate 101 (contact layer 105). The semiconductor light modulation device 100a according to Reference Example 2 is configured.
As described above, in this example, the semiconductor multilayer structure (light modulation element) is stacked in the order of nipn from the upper layer toward the substrate surface, and the light modulation waveguide of the light modulation element is formed in the reverse mesa direction. Yes.
Reference numeral 120 denotes an optical waveguide, and 121 denotes an MMI coupler.
Here, since the configuration conditions of the non-doped core layer and the light confinement layer in Reference Example 1 are the same, the structure of the doping layer in the light modulation region will be described below.

For example, n-type InP cladding layers 113a and 113b and n-type InGaAs signal electrode contact layers 112a and 112b are sequentially stacked on the upper surfaces of the optical confinement layers 114a and 114b. Further, for example, an n-type InGaAs ground electrode contact layer 105, n-type InP clad layers 122a and 122b, and p-type InAlAs carrier block layers 121a and 121b are sequentially stacked on the lower surfaces of the optical confinement layers 116a and 116b, thereby operating state. In the applied voltage range used in step 1, the entire region of p-type InAlAs and the partial region or the entire region of n-type InP are depleted.
The doping concentration profile of these layers and the thickness of p-type InAlAs are determined so that the potential change of the band in such a depletion region becomes sufficiently large, that is, to induce a sufficient potential barrier against electrons. . The doping concentration of these layers is preferably 2 × 10 ¹⁷ cm ⁻³ or more for the n-type layer and 1 × 10 ¹⁸ cm ⁻³ or more for the p-type layer. For example, the doping concentration of the n-type layer is 5 × 10 ¹⁷ cm ⁻³ , the doping concentration of the p-type layer is 1 × 10 ¹⁸ cm ⁻³ , and the film thickness is 50 nm.
Further, the p-type layer for functioning as a carrier block layer is not limited to InAlAs, and may be InGaAlAs, InP, or the like, for example.
Note that a layer stacked for electrode contact is not a problem as long as sufficient conductivity can be ensured. Therefore, a semiconductor doped with an n-type impurity is not limited to InGaAs, and for example, InGaAsP may be used.

After the nnip layer is deposited, the n-type InGaAs electrode contact layer 105 is exposed by dry etching using, for example, a SiO ₂ mask in which the waveguide shape is simulated in the reverse mesa direction so that the semiconductor light modulation element functions as an optical waveguide. Then, the mesa structure is fabricated. Thereby, the light modulation waveguide of the light modulation element is formed in the reverse mesa direction. At this time, the region that does not contribute to voltage application is removed by dry etching including the n-type InGaAs electrode contact layer 105 to expose the semi-insulating InP substrate 101. After removing the SiO ₂ mask, an Au electrode 104 is formed on the n-type InGaAs electrode contact layer 105 as shown in FIG.
Note that the signal electrode used in the semiconductor light modulation device 100a preferably has a traveling-wave electrode structure, but other than this, for example, a lumped constant type and a resonance type electrode structure may be used.

In order to function as an optoelectronic waveguide (light modulation waveguide), for example, as shown in FIG. 3A including two MMI couplers including a mesa structure having a cross section as shown in FIG. A Mach-Zehnder (MZ) waveguide structure is used, and a reverse bias voltage and a signal voltage are input to the electrodes 111a and 111b in a state where light is propagated through the waveguide formed in the reverse mesa direction. That is, a Mach-Zehnder (MZ) type optical waveguide 120 and an MMI coupler 121 are provided on the substrate 101 (contact layer 105), an optical modulation element is connected to the optical waveguide 120, and a reverse bias voltage is applied. In order to apply a high electric field to the non-doped layer, it is necessary to apply a voltage in the reverse bias direction. In the case of the nipn layer structure, as shown by the arrow in FIG. 3B, applying a bias voltage downward on the substrate side corresponds to the reverse bias voltage. Further, if a voltage is applied in the forward bias direction ([100] direction in FIG. 3B), a desired voltage drop does not occur, so that the function as a light modulation element is not achieved in the first place (FIG. 2 ( a)). Therefore, in order to prevent an electric field from being generated in the forward bias direction due to the amplitude of the signal voltage, the modulation operating point is shifted by a reverse bias larger than that, and modulation is always performed with the electric field applied to the reverse bias applied (see FIG. 2 (b)). For example, when the amplitude of the signal voltage is driven at ± 2V, the reverse bias voltage is applied at 2V or more.
Further, in order to reduce the power consumption and suppress the modulation chirping, the MZ optical modulation element is required to perform a push-pull drive modulation operation. Therefore, a signal voltage having the same intensity and the opposite direction of the electric field displacement is input to the electrodes 111a and 111b at the same timing as the modulation signal.
In addition, it is clear that the present reference example is useful even when it does not have an MZ waveguide structure, for example, only a linear waveguide structure is used as a phase modulation element.

As described above, according to this reference example , since an electric field is applied as a reverse bias in the downward direction of the substrate, the Pockels effect can be synergized in addition to the FK effect and the QCSE even for a modulation element manufactured in the reverse mesa direction. Therefore, it is possible to obtain a large refractive index change and good modulation characteristics shown in [1] in FIG. Furthermore, since the loss of light propagating in the optical waveguide is reduced as compared with the semiconductor optical modulation element having the nip layer structure shown in Reference Example 1 , the modulation element length can be made longer. Therefore, for example, by using a traveling wave type electrode as a signal electrode, a good modulation characteristic can be obtained even at high speed operation.

Here, it is clear that the usefulness of the substrate material does not change even when GaAs, GaP, ZnS, ZnSe, for example, has the same structure other than InP.
In this reference example , the semiconductor light modulation element corresponding to the 1.55 μm wavelength band is used, but the element corresponding to the 1.3 μm wavelength band may be used.
For example, if GaAs is used, it can respond to a wavelength range of 0.6 to 1.3 μm.

FIG. 4A is a top view of the semiconductor optical integrated device 200 of this embodiment. Crystal growth is performed by a metal organic chemical vapor deposition (MOVPE) method suitable for a crystal regrowth process, and a (100) plane orientation substrate is used as a substrate crystal. 4B is a cross-sectional view of the LD region 150 integrated with the semiconductor light modulation device 100 of Reference Example 1. FIG. The semiconductor substrate 101 is preferably manufactured using a semi-insulating InP substrate from the viewpoint of designing a high-frequency circuit used for light modulation.

The semiconductor optical integrated device 200 of Example 1, on the SI-InP substrate 101 includes a semiconductor multilayer structure of Reference Example 1 (light modulation element), moreover, the reverse mesa direction optical waveguide and the laser device of the light modulation element It is a monolithic integrated device formed in the above. That is, the MZ type light modulation region 130 of the semiconductor optical integrated device 200 has the same configuration as the semiconductor light modulation device 100 of Reference Example 1 .
In the LD region 150, an n-InGaAs contact layer 158 is disposed on the substrate 101, and an electrode 151, a p-InGaAs contact layer 152, a p-InP layer 153, i-light confinement are disposed on the substrate 101 (contact layer 158). A semiconductor multi-layer structure in which a layer 154, an i-active layer 155, an i-light confinement layer 156, and an n-InP layer 157 are stacked is provided.

  The active layer 155 in the LD region 150 has a multiple quantum well structure, and for example, non-doped InGaAsP and InP are used as the quantum well layer and the barrier layer, respectively. Optical confinement layers 154 and 156 corresponding to a band gap approximately between the active layer 155 and the clad are laminated on the upper and lower sides, respectively. A distributed feedback LD capable of oscillating at a desired single wavelength is manufactured by providing a diffractive structure in the optical confinement region. A p-type InP cladding layer 153 and a p-type InGaAs electrode contact layer 152 are sequentially stacked on the upper surface of the optical confinement layer 154, and an n-type InGaAs electrode contact layer 158 and an n-type InP cladding layer 157 are sequentially stacked on the lower surface. A pin structure is produced (see, for example, Patent Document 3). The oscillation wavelength of the LD is 1.55 μm.

After producing the pin layer structure, after the optical modulation element to remove crystal layer deposited by, for example, inductively coupled plasma (ICP) dry etching using an SiO ₂ mask was having an area made, as in Reference Example 1 A nip layer structure with the same conditions is laminated by crystal regrowth.

After the pin layer in the LD region 150 and the nip layer in the light modulation region 130 are deposited, in order to make the LD element and the light modulation device function as an optical waveguide, a waveguide shape is simulated in the reverse mesa direction, for example, using a SiO ₂ mask Then, processing is performed by dry etching until the n-type and p-type InGaAs electrode contact layers 103 and 158 are exposed, thereby producing a mesa structure. Thereby, a light modulation waveguide is formed in the reverse mesa direction. At that time, the regions not contributing to voltage application are removed by dry etching including the n-type and p-type InGaAs electrode contact layers 103 and 158 to expose the semi-insulating InP substrate 101.
Thereafter, for the LD element, both sides of the mesa structure are filled with semi-insulating InP by crystal re-growth, and again a desired electrode contact region is imitated with a SiO ₂ mask, and dry etching is performed, and FIG. The n-type InGaAs electrode contact layer 158 as shown in FIG.
Then, Au electrodes 159 and 104 via Ti, for example, are formed on the p-type and n-type InGaAs electrode contact layers 158 and 103 in the LD region 150 and the light modulation region 130.
The signal electrode used in the semiconductor light modulation element preferably has a traveling wave type electrode structure. However, for example, a lumped constant type and a resonance type electrode structure may be used.

In order to function as an optoelectronic waveguide (light modulation waveguide), for example, as shown in FIG. 4A including two MMI couplers including a mesa structure having a cross section as shown in FIG. With a Mach-Zehnder (MZ) waveguide structure, a reverse bias voltage and a signal voltage are applied to the electrodes 111a and 111b in a state where light oscillated from the same reverse mesa LD element shown in FIG. input. Here, the injection current value of the LD is 50 mA. In order to apply a high electric field to the non-doped layer, it is necessary to apply a voltage in the reverse bias direction. In the case of the nip layer structure, as shown by the arrow in FIG. 1B, applying a bias voltage downward on the substrate side corresponds to the reverse bias voltage. Further, if a voltage is applied in the forward bias direction ([100] direction in FIG. 1B), not only a desired voltage drop does not occur, but also carriers are injected into the non-doped layer. In the first place, it does not function as a light modulation element (FIG. 2A). Therefore, in order to prevent an electric field from being generated in the forward bias direction due to the amplitude of the signal voltage, the modulation operating point is shifted by a reverse bias larger than that, and modulation is always performed with the electric field applied to the reverse bias applied (see FIG. 2 (b)). For example, when the amplitude of the signal voltage is driven at ± 2V, the reverse bias voltage is applied at 2V or more.
Further, in order to reduce the power consumption and suppress the modulation chirping, the MZ optical modulation element is required to perform a push-pull drive modulation operation. Therefore, as the modulation signal, signal voltages having the same intensity and the opposite direction of the electric field displacement are input to the electrodes A and B at the same timing.
It is obvious that the present embodiment is useful even in the case where the MZ waveguide structure is not provided, for example, only a linear waveguide structure is used as the phase modulation element.

As described above, according to the present invention, since LD elements can be integrated in a small and monolithic manner in the same reverse mesa direction as the semiconductor optical modulation element 100 shown in Reference Example 1 , light between the LD element and the semiconductor optical modulation element can be integrated. Problems such as coupling loss and unnecessary substrate area consumption due to orthogonal arrangement can be solved. As a result, for example, a semiconductor optical integrated device equipped with an MZ light modulation element and an LD element, which are indispensable for large-capacity communication, can be manufactured in a small size and in mass production.

Here, it is clear that the usefulness of the substrate material does not change even when GaAs, GaP, ZnS, ZnSe, for example, has the same structure other than InP.
In the present embodiment, the semiconductor optical modulation element corresponding to the 1.55 μm wavelength band is used, but one corresponding to the 1.3 μm wavelength band may be used.
For example, if GaAs is used, it can respond to a wavelength range of 0.6 to 1.3 μm.

The semiconductor optical integrated device described in Example 1 is provided with the semiconductor optical modulator of the nipn layer structure described in Reference Example 2 for the purpose of further reducing light propagation loss and high-speed modulation operation. The semiconductor optical integrated device 200a will be described with reference to FIG. Crystal growth is performed by a metal organic chemical vapor deposition (MOVPE) method suitable for a crystal regrowth process, and a (100) plane orientation substrate is used as a substrate crystal. As the semiconductor substrate, an n-type doped InP substrate or a semi-insulating InP substrate can be used. From the viewpoint of designing a high-frequency circuit used for optical modulation, a semi-insulating InP substrate is preferably used. I do.

The semiconductor optical integrated device 200a of the second embodiment, on the SI-InP substrate 101 includes a semiconductor multilayer structure of Reference Example 2 (light modulation element), moreover, the reverse mesa direction optical waveguide and the laser device of the light modulation element It is a monolithic integrated device formed in the above. That is, the MZ type light modulation region 130a of the semiconductor optical integrated device 200a has the same configuration as that of the semiconductor light modulation device 100a of Reference Example 2 .
In the LD region 150, similarly to the first embodiment, an n-InGaAs contact layer 158 is disposed on the substrate 101, and an electrode 151, a p-InGaAs contact layer 152, and a p-InP are disposed on the substrate 101 (contact layer 158). A semiconductor multilayer structure in which layers 153, i-light confinement layers 154, i-active layers 155, i-light confinement layers 156, and n-InP layers 157 are laminated is provided.

  The active layer 155 in the LD region 150 has a multiple quantum well structure, and for example, non-doped InGaAsP and InP are used as the quantum well layer and the barrier layer, respectively. Optical confinement layers 154 and 156 corresponding to a band gap approximately between the active layer 155 and the clad are laminated on the upper and lower sides, respectively. A distributed feedback LD capable of oscillating at a desired single wavelength is manufactured by providing a diffractive structure in the optical confinement region. A p-type InP cladding layer 153 and a p-type InGaAs electrode contact layer 152 are sequentially stacked on the upper surface of the optical confinement layer 154, and an n-type InGaAs electrode contact layer 158 and an n-type InP cladding layer 157 are sequentially stacked on the lower surface. A pin structure is produced (see, for example, Patent Document 3).

After producing the pin layer structure, after the optical modulation element has with SiO ₂ mask was having an area made remove crystal layer deposited by an inductively coupled plasma (ICP) dry etching for example, as in Reference Example 2 A nipn layer structure with the same conditions is stacked by crystal regrowth.

After deposition of the nipn layer in pin layer and the light modulation region 130a in the LD region 150, in order to function LD element and the light modulator element as an optical waveguide, waveguide shape was typical in the reverse mesa direction, for example, a SiO ₂ mask used Then, the n-type InGaAs electrode contact layers 105 and 158 are processed by dry etching until a mesa structure is produced. Thereby, a light modulation waveguide is formed in the reverse mesa direction. At that time, the region not contributing to voltage application is removed by dry etching including the n-type InGaAs electrode contact layers 105 and 158 to expose the semi-insulating InP substrate 101. Note that in an optical modulation element having a nipn layer with a small light propagation loss, it is easy to increase the element length in order to realize low consumption driving. Plan It is obvious that the present embodiment is useful even when the element length is 5 mm or less.
Thereafter, both sides of the mesa structure are filled with semi-insulating InP by crystal regrowth for the LD element, and dry etching is performed again to expose the n-type InGaAs electrode contact layer.
Then, Au electrodes 159 and 104 via Ti, for example, are formed on the n-type InGaAs electrode contact layers 158 and 105 in the LD region 150 and the light modulation region 130a.
The signal electrode used in the semiconductor light modulation element preferably has a traveling wave type electrode structure. However, for example, a lumped constant type and a resonance type electrode structure may be used.

In order to function as a photoelectron waveguide (light modulation waveguide), for example, as shown in FIG. 4A including two MMI couplers including a mesa structure having a cross section as shown in FIG. With a Mach-Zehnder waveguide structure, reverse bias voltage and signal voltage are input to the electrodes 111a and 111b in a state where light oscillated from the same reverse mesa direction LD element shown in FIG. 4C is propagated. Here, the injection current value of the LD is 50 mA. In order to apply a high electric field to the non-doped layer, it is necessary to apply a voltage in the reverse bias direction. In the case of the nipn layer structure, as shown in FIG. 3B, applying a bias voltage downward on the substrate side corresponds to the reverse bias voltage. If a voltage is applied in the forward bias direction ([100] direction in FIG. 3B), a desired voltage drop does not occur, so that the function as a light modulation element is not achieved in the first place (FIG. 2A). ). Therefore, in order to prevent an electric field from being generated in the forward bias direction due to the amplitude of the signal voltage, the modulation operating point is shifted by a reverse bias larger than that, and modulation is always performed with the electric field applied to the reverse bias applied (see FIG. 2 (b)). For example, when the amplitude of the signal voltage is driven at ± 2V, the reverse bias voltage is applied at 2V or more.
Further, in order to reduce the power consumption and suppress the modulation chirping, the MZ optical modulation element is required to perform a push-pull drive modulation operation. Therefore, a signal voltage having the same intensity and the opposite direction of the electric field displacement is input to the electrodes 111a and 111b at the same timing as the modulation signal.
It is obvious that the present invention is useful even when the phase modulation element has only a linear waveguide structure without having a Mach-Zehnder waveguide structure.

As described above, according to the present invention, the LD elements can be integrated in a small and monolithic manner in the same reverse mesa direction as the semiconductor optical modulation element 100a shown in the reference example 2 , so that the light between the LD element and the semiconductor optical modulation element can be integrated. Problems such as coupling loss and unnecessary substrate area consumption due to orthogonal arrangement can be solved.
Further, since it is possible to lengthen the semiconductor optical modulation element, which is a problem in the optical integrated element including the semiconductor optical modulation element having the nip layer structure described in the first embodiment , for example, a traveling wave electrode is used as a signal electrode. By using it, it is possible to obtain good modulation characteristics even at higher speeds. As a result, for example, it is possible to manufacture a terabit-class transmitter for huge capacity transmission, which is conventionally performed by an LN external modulator, in a smaller and more mass-produced semiconductor optical integrated device.

Here, it is clear that the usefulness of the substrate material does not change even when GaAs, GaP, ZnS, ZnSe, for example, has the same structure other than InP.
In the present embodiment, the semiconductor optical modulation element corresponding to the 1.55 μm wavelength band is used, but one corresponding to the 1.3 μm wavelength band may be used.
For example, if GaAs is used, it can respond to a wavelength range of 0.6 to 1.3 μm.

100, 100a Semiconductor optical modulation element 101 Substrate 102 Buffer layer 103 Contact layer 104 Electrode 105 Contact layer 111a, 111b Electrode 112a, 112b Contact layer 113a, 113b Clad layer 114a, 114b Optical confinement layer 115a, 115b Core layer 116a, 116b Optical confinement Layer 117a, 117b Clad layer 121a, 121b Carrier block layer 122a, 122b Clad layer 130, 130a MZ type light modulation region 150 LD region 200, 200a Semiconductor optical integrated device

Claims (3)

A semiconductor multilayer including at least an n-type cladding layer, an i-core layer, and a p-type cladding layer on a substrate surface equivalent to the (100) plane of a zinc blende semi-insulating semiconductor crystal substrate in order from the upper layer to the substrate surface An MZ type light modulation element having a structure, wherein the light modulation waveguide of the MZ type light modulation element is formed in a reverse mesa direction which is a [011] plane direction with respect to the (100) plane of the substrate, and A semiconductor optical integrated device, wherein a laser device is mounted on a substrate in a reverse mesa direction which is a [011] plane direction with respect to a (100) plane of the substrate. On the substrate surface equivalent to the (100) plane of the zincblende semi-insulating semiconductor crystal substrate, at least an n-type cladding layer, an i-core layer, a p-type carrier block layer, and an n-type in order from the upper layer to the substrate surface An MZ type optical modulation element having a semiconductor multilayer structure including a cladding layer is provided, and the optical modulation waveguide of the MZ type optical modulation element is in a reverse mesa direction which is a [011] plane direction with respect to the (100) plane of the substrate A semiconductor optical integrated device, wherein the semiconductor optical integrated device is formed and mounted on the substrate in a reverse mesa direction which is a [011] plane direction with respect to a (100) plane of the substrate. The MZ type optical modulation element is applied with a reverse bias electric field from the upper layer toward the substrate surface, and the amplitude of the signal voltage so that a positive voltage is always applied to the signal electrode of the MZ type optical modulation element. 3. A semiconductor optical integrated device according to claim 1, wherein a larger offset voltage is applied.

Images