JP7569294B2 - Semiconductor optical modulator - Google Patents

Description

The present invention relates to a semiconductor optical modulation element that functions as a high-speed optical modulator applied to optical communication systems.

As optical communication capacity increases, there is a demand for smaller, less power-consuming optical devices used in optical communication systems. Mach-Zehnder (MZ) optical modulators are considered an important device in optical communication systems, and much research and development has been conducted on them. In recent years, in order to achieve even smaller size and lower power consumption, attention has shifted from optical modulators made of lithium niobate (LN) to optical modulators made of compound semiconductor materials such as InP.

To operate a compound semiconductor optical modulator, the core layer that confines the light is sandwiched between p-type and n-type cladding layers as a non-doped layer, and a reverse bias voltage is applied to utilize the strong interaction between light and electricity. Compared to n-type semiconductors, p-type semiconductors have a material electrical resistivity and optical absorption rate that are about an order of magnitude higher, making it a major challenge to increase the speed of modulation operation and reduce optical loss. In addition, if the contact area between the p-type semiconductor and the electrode is small, there is the issue of increased contact resistance leading to bandwidth degradation.

As a result, the n-p-i-n structure or n-i-p-n structure (see, for example, Patent Document 1) has been proposed as the device structure of a compound semiconductor optical modulator. By replacing the p-type semiconductor doping layer with an n-type semiconductor doping layer, it is possible to achieve both wideband characteristics and low driving voltage characteristics.

Patent No. 6458143 JP 2018-6590 A

Y. Ogiso et al., “80-GHz Bandwidth and 1.5-V Vπ InP-Based IQ Modulator”, Journal of Lightwave Technology Volume: 38, Issue: 2, Jan.15, 15 2020.

Even when most of the cladding is made of n-doped layers, such as in the n-i-p-n structure described above, it is necessary to insert a small p-doped layer to ensure voltage resistance (current blocking) when a reverse bias voltage is applied. In the case of semiconductor devices stacked on an InP substrate, Zn, Be, C, etc. are used as dopants for the p-doped layer. In particular, in crystal growth using the MOVPE method, which is suitable for mass production, Zn is almost always used as the dopant. However, Zn dopants are known to have a very large diffusion coefficient, and how to suppress Zn diffusion is also cited as an important issue in device design such as laser diodes and optical modulators (see, for example, Patent Document 2).

Figure 1 shows a band diagram when a voltage is applied in a conventional n-i-p-n optical modulator, and Figure 2 shows the electric field distribution. These are schematic diagrams showing the change in hand end energy and electric field strength when the applied voltage is changed. In an optical modulator, as shown in Figure 1, Zn is diffused into a non-doped layer including an MQW (multiple quantum well) region, and when a reverse bias voltage is applied, the hole concentration increases in a part of the non-doped layer, and no electric field drop occurs in the MQW region. In other words, as shown in Figure 2, the electric field strength on the P layer side of the non-doped layer becomes weak. The MQW region is gradually depleted by increasing the bias voltage, but the bias voltage required to obtain the desired depletion layer thickness is higher than in the case without Zn diffusion.

The problem here is that while no voltage drop occurs in the MQW region, a large electric field drop occurs in the non-doped region on the substrate side. In order to obtain high electro-optical modulation, optical modulators made of InP use the quantum confined Stark effect (QCSE) with an MQW structure. The QCSE effect changes the material refractive index and light absorption characteristics according to the electric field strength. In other words, the voltage value required to obtain the depletion layer thickness required to ensure high frequency characteristics is increased by Zn diffusion, while a higher electric field is applied than expected in areas not affected by Zn diffusion. Therefore, if an MQW structure is present in an area not affected by Zn diffusion, light absorption will increase more than expected.

This is a problem that prevents both high frequency characteristics and optical characteristics from being achieved at the same time, and unless Zn diffusion into the non-doped layer is suppressed, it becomes difficult to introduce the optical modulator into a practical system from the viewpoint of signal-to-noise (SN) ratio. It is also possible to avoid applying a higher electric field strength to the MQW region by shifting the relative height of the MQW structure in the stacked structure toward the p-layer side of the non-doped layer. However, moving the MQW region closer to the p-layer side increases optical absorption, which ends up degrading the device characteristics instead.

As a method for suppressing Zn diffusion, a method such as inserting a material with a high solid solubility limit concentration, such as InGaAsP, has been adopted. When trying to suppress Zn diffusion to a concentration of 1E+16 cm ^-3 or less, this method alone may not be sufficient when the concentration of the p-doped layer is high.

The object of the present invention is to provide a semiconductor optical modulation element that has a layer structure in which Zn diffusion is suppressed and is capable of high-speed optical modulation operation.

To achieve this object, one embodiment of the present invention is a semiconductor optical modulation element having a refractive index modulation region that modulates the refractive index of light guided to an optical waveguide, characterized in that a first n-type clad layer, a non-doped clad core layer, a second n-type clad layer, a p-type clad layer, and a third n-type clad layer are stacked from the top toward the substrate surface on a substrate surface equivalent to the (100) surface of a semiconductor crystal substrate 10.

According to the present invention, by inserting a p-type cladding layer, Zn diffusion is suppressed, and high-speed optical modulation operation can be achieved by efficiently depleting the entire non-doped layer when a reverse bias voltage is applied, and by achieving a phase modulation effect that suppresses optical absorption.

FIG. 1 is a diagram showing a band diagram when a voltage is applied to a conventional optical modulator with an nipn structure. FIG. 1 is a diagram showing the electric field intensity distribution when a voltage is applied in a conventional optical modulator with an nipn structure. 1 is a diagram showing a layer structure of a semiconductor optical modulation element according to one embodiment of the present invention; FIG. 2 is a diagram showing a band diagram when a voltage is applied to the semiconductor optical modulation element of the present embodiment. 4 is a diagram showing an electric field intensity distribution when a voltage is applied in the semiconductor optical modulation element of the present embodiment. FIG. 1 is a plan view showing a semiconductor optical modulation element according to one embodiment of the present invention; A cross-sectional view of the refractive index modulation region (A-A') of the semiconductor optical modulation element of this embodiment. A cross-sectional view of the input/output region (B-B') of the semiconductor optical modulation element of this embodiment.

The following describes in detail an embodiment of the present invention with reference to the drawings. It is known that Zn diffusion is suppressed when Zn hardly diffuses from the p-layer side to the n-layer side between pn junctions due to an internal electric field. For this reason, a pn-buried type laser diode is known in which high-concentration p-layers and n-layers are stacked to block leakage current. This principle is then applied to provide a new semiconductor optical modulation element having an n-i-n-p-n type heterostructure. This structure incorporates a so-called npn thyristor structure into part of the modulation region.

Conventionally, most semiconductor optical elements made of InP are active elements that utilize carrier recombination, and take advantage of the behavior of carriers, including photocarriers. Therefore, structures that include heterojunctions or pn junctions that may trap carriers in the electric field application region have only been used in elements that handle carrier injection or photocarrier generation, such as laser diodes and electroabsorption optical modulators.

On the other hand, in a phase modulator where light absorption is very low and attention needs only to be paid to the behavior of the electric field strength near the MQW region, the properties of the pn junction can be fully utilized. That is, in addition to the electric field application action by carrier blocking, it can also function as a Zn diffusion blocking layer. This makes it possible to suppress the diffusion of p-type dopants into the non-doped layer and apply a uniform electric field to the non-doped layer. As a result, it is possible to achieve both efficient depletion of the entire non-doped layer when a reverse bias voltage is applied and phase modulation action with suppressed light absorption, thereby realizing high-speed optical modulation operation.

Figure 3 shows the layer structure of a semiconductor optical modulation element according to one embodiment of the present invention. The semiconductor optical modulation element of this embodiment has an n-i-n-p-n type heterostructure and is stacked on a semiconductor crystal substrate 10. From the top, the layers are a first n-type cladding layer 16, a non-doped cladding core layer 15, a second n-type cladding layer 14, a p-type cladding layer 13, a third n-type cladding layer 12, and a second n-type contact layer 11. The non-doped cladding core layer 15 is made of a non-doped core layer 52 that becomes an MQW region inserted into the non-doped cladding layers 51 and 53. The stacking order of the second n-type contact layer 11 and the third n-type cladding layer 12 does not affect the usefulness of this embodiment.

The relative center position of the non-doped core layer 52 in the laminate structure is shifted toward the first n-type clad layer 16 from the center position of the entire non-doped clad-core layer 15. As described below, according to the layer structure of this embodiment, a high electric field is not applied to the non-doped core layer 52, which becomes the MQW region, so optical absorption can be suppressed by shifting the MQW region to the opposite side from the p-type clad layer 13.

Figure 4 shows a band diagram when a voltage is applied in the semiconductor optical modulation element of this embodiment, and Figure 5 shows the electric field intensity distribution. Each is a schematic diagram showing the change in hand end energy and electric field intensity when the applied voltage is changed. As shown in Figure 5, when there is no bias (0 V applied), no built-in voltage is applied to the non-doped core layer 52 consisting of the MQW region. In other words, by suppressing the application of unnecessary electric fields to the waveguide core, the effect of band edge optical absorption in the MQW can be minimized. Therefore, the optical propagation loss in the semiconductor optical modulation element can be reduced.

By suppressing the diffusion of Zn and the increase in the concentration in the non-doped clad core layer 15, as shown in FIG. 5, a more uniform electric field is applied to the non-doped clad core layer 15 when a reverse bias voltage is applied. This prevents a high electric field from being applied in a concentrated manner to the MQW region, reducing the optical absorption caused by QCSE and enabling the desired depletion layer thickness to be obtained.

In addition, the diffusion of Zn replaces part of the non-doped layer with a p-layer, increasing the effective thickness of the p-layer. In terms of the electrical circuit, this is equivalent to inserting a high-resistance resistor of a certain thickness, and is the main cause of increased electrical loss in high-frequency lines. By suppressing the diffusion of Zn, it becomes possible to design wideband high-frequency circuits without having to consider high-resistance resistors.

Next, we will explain some specific examples of semiconductor optical modulation elements.

Figure 6 shows a semiconductor optical modulation element according to one embodiment of the present invention. The semiconductor optical modulation element is a Mach-Zehnder type optical modulator, and two arm waveguides 31a and 31b are connected between an optical splitter and an optical multiplexer formed in an optical waveguide 31. A capacitance-loaded electrode structure is formed in the arm waveguides 31a and 31b, and a secondary electro-optic effect is caused in the core layer by application of an electric signal, forming a refractive index modulation region that modulates the refractive index of the core layer. The capacitance-loaded electrode structure is composed of a signal electrode 18 and a ground electrode 19, and a signal source 33 is connected to one end and a termination resistor 34 is connected to the other end.

Figure 7 shows a cross section of the refractive index modulation region (A-A') of the semiconductor optical modulation element of this embodiment. A second n-type contact layer 11, a third n-type cladding layer 12, a p-type cladding layer 13, a second n-type cladding layer 14, a non-doped cladding core layer 15, and a first n-type cladding layer 16 are sequentially laminated on a substrate surface equivalent to the (100) surface of a zinc-blende semi-insulating semiconductor crystal substrate 10. Dry etching and wet etching processes are used to separate the regions where the arm waveguides 31a and 31b are formed, and the regions that do not contribute to modulation are removed. In addition, from the viewpoint of reducing optical loss, the removed portions are backfilled with a semi-insulating InP buried layer 20.

Furthermore, by etching the first n-type cladding layer 16 in a direction equivalent to the [011] plane direction into the shape of the arm waveguides 31a, 31b, the cladding formed in a ridge shape in the reverse mesa direction becomes an optical modulation waveguide. A signal electrode 18 and a ground electrode 19 are formed on the upper surface of the first n-type cladding layer 16 via the first n-type contact layer 17. The signal electrode 18 and the ground electrode 19 are traveling wave electrodes having a coplanar strip line structure.

For example, the second n-type contact layer 11 is made of InGaAs with a carrier concentration of 5E+18 cm ^-3 , and the third n-type cladding layer 12 and the first n-type cladding layer 16 are made of InP with a carrier concentration of 1E+18 cm ^-3 . The carrier concentration of the p-type cladding layer 13 is set to 5E+17 to 1E+18 cm ^-3 in consideration of the optical absorption coefficient and electrical resistivity, and a material with a band gap larger than that of InP, such as InAlAs, is used to enhance the electron carrier blocking effect.

If the second n-type cladding layer 14 has a high concentration or a large thickness, it will prevent efficient application of an electric field to the core layer due to the thyristor structure, so it is necessary to set the minimum concentration and thickness to suppress Zn diffusion. Specifically, the maximum layer thickness is estimated to be 50 nm when the concentration is 1E+17 cm ^-3 , and approximately 20 nm when the concentration is 2E+18 cm ^-3 . In this embodiment, InGaAsP is used with a concentration of 1E+18 cm ^-3 and a layer thickness of 20 nm.

The crystals were grown on a semi-insulating InP (100) substrate 10 by metal organic vapor phase epitaxy (MOVPE). The band gap wavelength of the non-doped core layer 52 is determined within a range that effectively operates the electro-optic effect at the operating light wavelength with high efficiency and does not cause optical absorption problems. For example, in the case of the 1.55 μm band of the communication wavelength band, the emission wavelength of the non-doped core layer 52 is set to about 1.4 μm. From the viewpoint of high efficiency modulation, the non-doped core layer 52 is preferably formed with a multiple quantum well structure of InGaAlAs/InAlAs. Note that the usefulness of this embodiment is not lost even if the multiple structure is InGaAsP/InP, InGaAsP/InGaAsP. In addition, the composition of the contact layer and the cladding layer is not limited to the above, and for example, an InGaAsP composition may be used.

Thereafter, a pattern of an arm waveguide made of SiO ₂ formed in a direction equivalent to the [011] plane direction is formed, and a ridge-shaped optical waveguide is formed by dry etching and wet etching.

Benzocyclobutene (BCB) is applied as an insulating film to flatten the unevenness of the waveguide, and after removing the BCB from the first n-type contact layer 17, a capacitive-loaded electrode structure is formed by gold plating. Note that there is no problem in using an insulating low-refractive index material such as polyimide instead of BCB.

Figure 8 shows a cross section of the input/output region (B-B') of the semiconductor optical modulation element of this embodiment. In order to apply a bias voltage to the third n-type cladding layer 12, dry/wet etching is performed to expose a part of the second n-type contact layer 11 and form a DC bias electrode 32.

To operate the fabricated semiconductor optical element as an optical modulator, a predetermined bias is applied to the DC bias electrode 32 so that a reverse electric field is applied to the pn junction (between the p-type cladding layer 13 and the second n-type cladding layer 14), and a high-frequency signal is fed from the signal source 33 to the signal electrode 18. Note that in this embodiment, a single-phase signal is fed from the signal source 33 to operate the element as an optical modulator, but from the viewpoint of reducing power consumption, each electrode may be terminated with a termination resistor to feed a differential signal.

REFERENCE SIGNS LIST 10 Semiconductor crystal substrate 11 Second n-type contact layer 12 Third n-type cladding layer 13 P-type cladding layer 14 Second n-type cladding layer 15 Non-doped cladding core layer 16 First n-type cladding layer 17 First n-type contact layer 18 Signal electrode 19 Ground electrode 20 Buried layer 31 Optical waveguides 31a, 31b Arm waveguide 32 DC bias electrode 33 Signal source 34 Termination resistors 51, 53 Non-doped cladding layer 52 Non-doped core layer

Claims (6)

A semiconductor optical modulation element having a refractive index modulation region that modulates the refractive index of light guided to an optical waveguide,
a first n-type cladding layer, a non-doped cladding core layer, a second n-type cladding layer, a p-type cladding layer, and a third n-type cladding layer are laminated on a substrate surface equivalent to a (100) surface of a semiconductor crystal substrate from the top toward the substrate surface ;
The semiconductor optical modulation element , wherein the second n-type cladding layer has an n-type doping concentration in the range of 1E+17 cm ^-3 to 2E+18 cm ^-3 and a layer thickness in the range of 20 nm to 50 nm . A semiconductor optical modulation element having a refractive index modulation region that modulates the refractive index of light guided to an optical waveguide,
a first n-type cladding layer, a non-doped cladding core layer, a second n-type cladding layer, a p-type cladding layer, and a third n-type cladding layer are laminated on a substrate surface equivalent to a (100) surface of a semiconductor crystal substrate from the top toward the substrate surface;
A semiconductor optical modulation element characterized in that the non-doped clad-core layer is composed of a non-doped core layer inserted into a non-doped clad layer, and the relative center position of the non-doped core layer in the laminate structure is located higher than the center position of the entire non -doped clad-core layer. A semiconductor optical modulation element having a refractive index modulation region that modulates the refractive index of light guided to an optical waveguide,
A semiconductor optical modulation element comprising, on a substrate surface equivalent to a (100) surface of a semiconductor crystal substrate, a first n-type cladding layer, a non-doped cladding core layer, a second n-type cladding layer, a Zn-doped p-type cladding layer, and a third n-type cladding layer, stacked from the top toward the substrate surface. The semiconductor optical modulation element according to claim 1, 2 or 3, characterized in that an electrode is formed on each of the first n-type cladding layer and the third n-type cladding layer via a contact layer, and a reverse electric field is applied to the pn junction. The semiconductor optical modulation element according to any one of claims 1 to 4, characterized in that the optical splitter and the optical multiplexer formed in the optical waveguide are connected by two arm waveguides, and a Mach-Zehnder type optical modulator is formed in which a capacitively loaded electrode structure is formed in the arm waveguide. The semiconductor optical modulation element according to claim 5, characterized in that the first n-type cladding layer is etched in a direction equivalent to the [011] plane direction to the shape of the arm waveguide, so as to have a ridge shape in the reverse mesa direction.

Images