An invention according to an embodiment of the present invention relates to a semiconductor light modulation element, and more particularly to an optical coupling technique between a light modulation element and an optical fiber.
In recent years, light modulators using compound semiconductor materials have been actively researched and developed in the context of a reduction in size and an increase in speed of light modulators. In particular, a light modulator using InP as a substrate material is capable of highly efficient modulation operation by utilizing the quantum-confined Stark effect or the like in a communication wavelength band, and thus InP has attracted attention as a promising modulator material in place of conventional ferroelectric materials.
An InP-based light modulator is able to obtain highly efficient light modulation characteristics due to the quantum-confined Stark effect (QCSE), known as an electrooptical effect, by using a multipleb quantum well structure (MQW) for an optical waveguide core, and therefore many of the InP modulators employ a structure in which the MQW is taken as a core. The MQW of the InP-based modulator used for a communication wavelength (in particular, C-band, near 1.55 μm) is broadly classified into two types of material bases. One of them is an MQW containing Al atoms (hereinafter, referred to as an Al-based MQW) in which InAlAs or InGaAlAs is a barrier layer and InGaAlAs is a well layer, and the other one is an Al atom-free MQW (hereinafter, referred to as a P-based MQW) in which InP or InGaAsP is a barrier layer and InGaAsP or InGaAs is a well layer.
In general, since a conductor band offset (ΔEc) is larger in the Al-based MQW than that in the P-based MQW, a sharper band absorption edge can be obtained, thereby making it possible to perform highly efficient light modulation by QCSE. Therefore, the Al-based MQW has been adopted, in many cases, the high speed modulators in recent years.
Next, a light input/output section of the light modulation element will be described. In the InP-based light modulators, as described above, since the MQW is used as a core layer in many cases, a mode field MDF of light is determined approximately by the width and height of the MQW.
Since it is preferable for optical devices to be smaller in spot size in many cases, it is necessary to form a spot size converter (SSC) on an optical waveguide as needed in order to increase a spot size only in a localized area. Various structures and manufacturing methods for SSCs configured to locally expand spot sizes in optical waveguide-type optical devices have been present.
There are mainly two types of mechanisms for spot size conversion. One of them is an approach, as illustrated in
In PTL 1, a dry etching device with high accuracy is necessary in order to effectively leverage a micro loading effect and obtain a uniform etching depth in the wafer surface, which raises a problem that the manufacturing environment is limited. Further, it may be difficult to stably obtain a depth on the order of several nanometers even in a case in which the depth is controlled by an internal monitor or the like during dry etching. Furthermore, because a hard mask of SiO2 or the like is typically used in dry etching, there is a problem of an increase in manufacturing process to be additionally carried out, such as processing of SiO2 or the like.
Thus, in order to solve the above problems, an object of an invention according to an embodiment of the present invention is to provide a spot size converter that is manufactured by a simple manufacturing apparatus, is stably etched with the accuracy on the order of several nanometers, and is able to shorten the manufacturing process.
An invention according to an embodiment of the present invention is conceived to provide a semiconductor light modulation element serving as a semiconductor light modulator including an InP-based compound semiconductor, wherein a waveguide core layer of the semiconductor light modulator includes an etching stop layer containing a P element, and a multiple quantum well structure located on the etching stop layer and containing an Al element, a bather layer is located over the etching stop layer and is provided in the multiple quantum well structure, and an energy band gap of the etching stop layer is smaller than a band gap of the bather layer.
The etching stop layer is inserted into a desired position (a position at which the etching is expected to be stopped) in a modulator core layer (including an MQW). When an MQW containing an Al element is used, it is desirable for the etching stop layer containing a P element. For example, in the case of an MQW in which InAlAs is a barrier layer and InGaAlAs is a well layer, InP or InGaAsP that can be lattice-matched with the above layers is used for the etching stop layer. It is desirable for the stop layer to have a smaller band gap than the barrier layer.
By using the invention according to an embodiment of the present invention, it is possible to manufacture an optical spot size converter provided in a light input/output section of a semiconductor light modulation element in a shorter time and with higher accuracy (the overall layer thickness can be controlled on the order of several nanometers) than the conventional art, without impairing light transmittance and light modulation characteristics (quenching characteristics) of the overall light modulation element.
Embodiments of the present invention will be described with reference to the accompanying drawings. The embodiments described below are embodiments of the present invention, and the present invention is not limited to the following embodiments.
The substrate uses the semi-insulating InP substrate 101 doped with Fe, for example, as a compound semiconductor crystal of a zinc blende type. An n-type contact clad layer, a non-doped core clad layer, and a p-type clad contact layer are laminated in that order from the substrate 101 surface by epitaxial growth. The n-type contact clad layer corresponds to a lower clad layer 102, the non-doped core clad layer corresponds to a multiple quantum well (MQW) structure 204a, and the p-type clad contact layer corresponds to an upper p-type clad layer 105a. As illustrated in
A multiple quantum well (MQW) structure 204b (PL wavelength: 1.4 μm) constituted of a period of InGaAlAs/InAlAs was used in the core layer in order to efficiently use a refractive index change due to the electrooptical effect with respect to a 1.5 μm band wavelength.
When an MQW containing an Al element is used, it is desirable for the etching stop layer containing a P element. For example, in the case of an MQW in which InAlAs is a barrier layer and InGaAlAs is a well layer, InP or InGaAsP that can be lattice-matched with the above layers is used for the etching stop layer. It is desirable for the stop layer to have a smaller band gap than the barrier layer.
As illustrated in
A heterostructure of a p-i-n type from above is used in the present embodiment, but the invention according to the embodiment of the present invention exhibits its effects when the waveguide includes etching stop layers in the MQW structure, and therefore it is apparent that even a heterostructure in which, for example, n-i-p, n-p-i-n, and n-i-p-n are laminated in that order from above causes no problem. The clad layer was composed of InP having a lower refractive index than the core layer, for example, and InGaAs being lattice-matched with InP and having a small energy band gap was used for the p-type contact layer. The doping concentrations of the n-type clad layer and the p-type clad layer were both 1×1018 cm−3, and the doping concentration of InGaAs was 1×1019 cm3.
It is only required that the compositions of the core and the clad each have a relative refractive index difference, and therefore InGaAlAs and the like having different compositions may be used in the core clad layer, the n-type clad layer, and the p-type clad layer, for example.
The wavelength is not limited to the 1.5 μm band, and even when a 1.3 μm band is used, the usefulness of the invention according to the embodiment of the present invention will not be lost.
To form electro-separation between electrodes and an SSC structure, the p contact layer and the p clad layer in the regions other than the light modulation region are removed by dry etching and chemical etching. Subsequently, photoresist patterning is performed in the SSC regions 202a and 202b by using a first mask pattern (opening) 301a to form a first MQW 304a of a first step (the uppermost step of the four steps). Thereafter, the first MQW 304a is wet-etched down to the first etching stop layer 314a. An etchant containing hydrogen peroxide and having a high etching rate difference between Al and P elements was used as an etching liquid. Subsequently, in a similar manner, after forming a resist pattern by using a second mask pattern (opening) 301b, etching of the first etching stop layers 314a and a second MQW 304b is performed. The opening of the second mask pattern 301b is smaller than the opening of the first mask pattern 301a. A hydrochloric acid-based etchant was used for the etching stop layer. Finally, similar processing is performed with respect to a third mask pattern 301c so as to pattern a third MQW 304c. The opening of the third mask pattern 301c is smaller than the opening of the second mask pattern 301b. Consequently, only the third etching stop layer 314c and a fourth MQW 304d are left. The film thickness of the fourth MQW 304d is controlled by crystal growth, which makes it possible to control the thickness on the order of nanometers. In the present example, the thickness of the fourth MQW 304d was set to be 100 nm, for example, but it is unnecessary to limit the thickness to 100 nm because the thickness thereof differs depending on the desired mode field. It is possible to obtain the spot size converter (SSC) 202b in
After the processing of the MQW in the SSC region is completed, a non-doped clad layer 105 (here, InP was used) is deposited by crystal regrowth, for example. An Fe-doped clad layer may be used instead of the non-doped clad layer.
When the light modulation region and the light input/output region are configured to have different cores, a complicated manufacturing process needs to be additionally performed, and therefore it is preferable to have the MQW cores constituted of the same composition.
Subsequently, the Mach-Zehnder (MZ) interference waveguide 201 is formed by dry etching using a SiO2 mask, as illustrated in
As illustrated in
Chemical wet etching is used to simplify the processing apparatus and shorten the processing.
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/JP2019/025431 | 6/26/2019 | WO |