JP5494464B2 - Semiconductor device and method for manufacturing semiconductor device - Google Patents

Description

  The present invention relates to a semiconductor element and a method for manufacturing a semiconductor element.

  In order to stably operate the optical semiconductor element at a high temperature, it is desired to suppress carrier leakage from the oscillation level due to thermal excitation. For this purpose, it is effective to use quantum dots having discrete energy levels in the active layer.

  There has been proposed a columnar quantum dot in which quantum dots self-generated in a Stranski-Krastanov (SK) mode are stacked in a height direction via a spacer layer (also called a barrier layer). A plurality of quantum dots stacked in one columnar quantum dot are coupled to each other quantum mechanically. Specifically, the interval between the quantum dots in the height direction is narrower than the spread of the wave function of the carriers. The polarization characteristics of the columnar quantum dots can be controlled by changing the structural parameters such as the number (height) of the quantum dots and the amount of strain in the spacer layer.

  A semiconductor laser is required to have high electro-optical conversion efficiency, and a semiconductor optical amplifier is required to have high optical gain. By increasing the distribution density of the columnar quantum dots, it is possible to improve electro-optical conversion efficiency and optical gain. By stacking columnar quantum dot layers in which columnar quantum dots are distributed in the plane in multiple stages, the volume distribution density of columnar quantum dots can be increased.

JP 2007-157975 A JP 2007-242748 A JP 2008-244235 A Proceedings of the 56th Joint Lecture on Applied Physics in Spring 2009, Yasuoka et al., 2a-G-2

  When the columnar quantum dot layers are stacked, an intermediate layer is disposed between the columnar quantum dot layers. An active layer is formed by the multistage columnar quantum dot layer. If the active layer is thick, a longitudinal higher-order mode is likely to occur, and therefore it is not preferable to make the active layer too thick. In order to prevent the active layer from becoming too thick, it is preferable to make the intermediate layer thin. When the intermediate layer becomes thinner, the strain of the lower columnar quantum dot layer is likely to propagate to the upper columnar quantum dot layer via the intermediate layer. This makes it difficult to grow crystals while maintaining a good crystal state.

  In addition, not only when columnar quantum dot layers are stacked in multiple stages, but also when a semiconductor layer is formed on a single columnar quantum dot layer, a technique for improving the crystal quality of the semiconductor layer is desired.

  An object of the present invention is to provide a semiconductor device suitable for growing a high-quality semiconductor layer above a columnar quantum dot, and a method for manufacturing such a semiconductor device.

  According to an aspect of the present invention, a semiconductor substrate and a columner quantum dot formed above the semiconductor substrate are provided, and the columner quantum dot is stacked on the first quantum dot and the first quantum dot. A first portion formed of a first spacer layer having a tensile strain amount with respect to the semiconductor substrate, a second quantum dot, and stacked on the second quantum dot from the first spacer layer to the semiconductor substrate. There is provided a semiconductor device having a structure in which second portions formed of second spacer layers having a large amount of tensile strain are alternately arranged in the thickness direction.

  The columnar quantum dots in which the first spacer layer having a relatively small tensile strain amount and the second spacer layer having a relatively large tensile strain amount are alternately arranged in the thickness direction provide a tensile strain amount of the spacer layer. Compared with a uniform columnar quantum dot, it is possible to reduce the average strain amount corresponding to a predetermined emission wavelength.

FIG. 1 is a schematic cross-sectional view showing a multi-layered portion of columnar quantum dots included in the semiconductor optical amplifier of the first embodiment. FIG. 2 is an enlarged schematic cross-sectional view showing one columnar quantum dot of the first embodiment. FIG. 3 is a schematic cross-sectional view of the semiconductor optical amplifier according to the first embodiment viewed from the incident direction of the amplified light. FIG. 4 is a schematic cross-sectional view (in the thickness direction) of the semiconductor optical amplifier according to the first embodiment as seen from the side where the light to be amplified enters. FIG. 5 is a graph showing the relationship between the average strain amount of the columnar quantum dots of the first and second examples and the PL emission wavelength. FIG. 6 is a schematic conduction band lineup of quantum dots of the columnar quantum dots of the comparative example and the first example. FIG. 7 is an enlarged schematic sectional view showing one columnar quantum dot of the second embodiment. FIG. 8 is a schematic cross-sectional view showing a multi-layered portion of columnar quantum dots included in the semiconductor optical amplifier of the third embodiment. FIG. 9 is an enlarged schematic cross-sectional view showing one columnar quantum dot of the third embodiment. FIG. 10 is a schematic cross-sectional view of the semiconductor optical amplifier according to the third embodiment viewed from the incident direction of the amplified light. FIG. 11 is a schematic cross-sectional view showing a multi-layered portion of columnar quantum dots included in the semiconductor optical amplifier of the fourth embodiment. FIG. 12 is an enlarged schematic sectional view showing one columnar quantum dot of the fourth embodiment. FIG. 13 is a schematic cross-sectional view showing a columnar quantum dot of a comparative example. FIG. 14 is a graph showing the relationship between the average strain amount of the columnar quantum dots of the comparative example and the PL emission wavelength.

  Before describing semiconductor devices according to embodiments of the present invention, comparative examples will be described first. As a comparative example, an experiment in which a single columnar quantum dot was produced and a photoluminescence (PL) emission wavelength was measured, and an experiment in which multi-layer stacking of columnar quantum dots was attempted were performed.

  FIG. 13 is a schematic cross-sectional view showing a columnar quantum dot of a comparative example. An InGaAsP layer 103 is formed above the InP (001) substrate. On the InGaAsP layer 103, quantum dots 104 and spacer layers 105 are alternately laminated to form columnar quantum dots 106. Quantum dots 104 are arranged on the top of the columnar quantum dots 106, 12 quantum dots 104 are formed, and 11 spacer layers 105 are formed.

  The quantum dots 104 are formed of InAs, and the spacer layer 105 is formed of InGaAsP having a wavelength corresponding to the band gap (hereinafter referred to as a composition wavelength) of 1.0 μm. As a method for forming the quantum dots 104 and the spacer layer 105, metal organic chemical vapor deposition (MOVPE) can be used.

The average strain epsilon _a of columnar quantum dots is defined by the following equation.

Here, epsilon _QDi is distortion of i-th quantum dots 104 from below, epsilon _SBi is the amount of strain i-th spacer layer 105 from the bottom. The strain amounts ε _a , ε _QDi , and ε _SBi are expressed in units of%, for example.

d _QDi and d _SBi are the thicknesses of the i-th quantum dot 104 and the spacer layer 105 from the bottom, respectively. The thickness of the quantum dot 104 and the spacer layer 105 means the thickness of the film when it is assumed that a film having a uniform thickness is formed by the raw material supplied during the growth. The thicknesses d _QDi and d _SBi are expressed in units of _monolayer (ML), for example.

The strain amount ε _QDi of the quantum dots 104 is calculated by the following equation.

The strain amount ε _SBi of the spacer layer 105 is calculated by the following equation.

Here, a _InAs , a _InGaAsP , and a _InP are the lattice constants of InAs, InGaAsP, and InP, respectively. The lattice constant of InGaAsP can be calculated by obtaining the lattice constant from the composition ratio of each element of InGaAsP and using the obtained lattice constant. The “lattice constant” means a lattice constant when it is assumed that no strain is generated in the semiconductor material.

  Compressive strain is positive and tensile strain is negative. InAs forming the quantum dots 104 has a lattice constant larger than that of InP of the substrate and has compressive strain. InGaAsP forming the spacer layer 105 has a smaller lattice constant than InP of the substrate and has tensile strain. The spacer layer 105 compensates for compressive strain caused by the quantum dots 104.

  Hereinafter, the tensile strain amount of the spacer layer 105 is indicated by omitting a minus sign, and the larger the absolute value of the tensile strain amount, the larger (or higher) the tensile strain amount, and the smaller the absolute value of the tensile strain amount. Expressed as small (or low) tensile strain. The average strain amount of columnar quantum dots is similarly defined in the examples described later. The average strain amount of the columnar quantum dots may be simply referred to as “average strain amount”.

  In the columnar quantum dots 106 of the comparative example, the tensile strain amount is set uniformly in all the stacked spacer layers 105. A plurality of samples in which the columnar quantum dots 106 were formed by changing the tensile strain amount of the spacer layer 105, that is, changing the average strain amount, were prepared, and the PL emission wavelength of these samples was measured.

  The tensile strain amount of the spacer layer 105 of each sample was 0.4%, 0.6%, 0.8%, 1.0%, 1.2%, and 1.4%. Note that the amount of compressive strain of the InAs quantum dots 104 is 3.2%. As the tensile strain amount of the spacer layer 105 increases, the compressive strain due to the quantum dots 104 is compensated, and the average strain amount (the compressive strain amount remaining without compensation) decreases.

  FIG. 14 is a graph showing the relationship between the average strain amount of the columnar quantum dots of the comparative example and the PL emission wavelength. The horizontal axis is the average strain amount expressed in units of%, and the vertical axis is the PL emission wavelength expressed in units of μm. As the average strain amount decreases, that is, as the tensile strain amount of the spacer layer increases, the PL emission wavelength becomes shorter. For example, the average strain amount for obtaining PL emission at a wavelength of about 1.55 μm included in the C band (1.53 μm to 1.565 μm) is 0.775% (the spacer layer tensile strain amount is 0.6%).

  Next, assuming a desired PL emission wavelength of 1.55 μm, an average strain amount of 0.775% was adopted to attempt multi-layer stacking of columnar quantum dots. When we tried to stack three columnar quantum dots through an InGaAs intermediate layer with a thickness of 40 nm, the compressive strain caused by the columnar quantum dots propagated in the thickness direction, making it difficult to grow in a good crystalline state. Met.

  On the other hand, even when the average strain amount is reduced to 0.654% (spacer layer tensile strain amount is increased to 0.8%), the stack of three-stage columnar quantum dots is formed through the 40 nm thick InGaAs intermediate layer. Attempts were made to grow while maintaining a good crystalline state.

  Thus, it was found that it is desirable to reduce the average strain amount in order to satisfactorily stack columnar quantum dots. However, as the average strain amount is reduced to 0.654%, for example, the PL emission wavelength becomes 1.529 μm, which is shorter than the desired wavelength. A columnar quantum dot structure is desired in which the emission wavelength is less likely to shift to the short wavelength side even when the average distortion amount is reduced, or the average distortion amount corresponding to the desired emission wavelength can be reduced.

  Next, the columnar quantum dot according to the first embodiment and a semiconductor optical amplifier using the same will be described.

  FIG. 1 is a schematic cross-sectional view showing a multi-layered portion of columnar quantum dots included in the semiconductor optical amplifier of the first embodiment.

  FIG. 2 is an enlarged schematic cross-sectional view showing one columnar quantum dot of the first embodiment.

The description will proceed along the manufacturing process with reference to FIGS. MOVPE can be used as the crystal growth method. Trimethylindium (TMIn) and triethylgallium (TEGa) can be used as group III element materials, and arsine (AsH ₃ ) and phosphine (PH ₃ ) can be used as group V element materials. Monosilane (SiH ₄ ) can be used as an n-type dopant material, diethyl zinc (DEZn) can be used as a p-type dopant material, and hydrogen (H ₂ ) can be used as a carrier gas. The growth pressure is, for example, 50 Torr.

The n-type InP (001) substrate 1 is equipped in a MOVPE growth furnace, and the growth temperature (growth surface temperature) is raised to 630 ° C. in a PH ₃ atmosphere.

Next, TMIn and SiH ₄ are also supplied, and the n-type InP buffer layer 2 having a doping concentration of 5.0 × 10 ¹⁷ cm ⁻³ is formed to a thickness of 500 nm.

The supply of TMIn and SiH ₄ is stopped, and the growth temperature is lowered to 430 ° C. in a PH ₃ atmosphere. Next, the group III gas is TMIn and TEGa, the group V gas is AsH ₃ and PH ₃ , and the non-doped InGaAsP optical confinement layer 3 having a composition wavelength of 1.1 μm and a strain amount of 0% is formed on the buffer layer 2. A thickness of 30 nm is formed.

Next, TMIn and AsH ₃ are used as the supply gas, and InAs quantum dots 4 are formed on the optical confinement layer 3. The growth conditions of the InAs quantum dots 4 are, for example, a supply amount of 2.5 ML, a growth rate of 0.045 μm / h, and a V / III ratio of 30.

Under these conditions, quantum dots 4 having a lateral dimension of about 20 nm in the [110] direction and about 30 nm in the [−110] direction and an in-plane density of about 7.5 × 10 ¹⁰ cm ⁻² are formed. The quantum dots 4 grow in the SK mode, and the wetting layer 4a remains outside the quantum dots 4.

Next, TMIn, TEGa, AsH ₃ , and PH ₃ are used as the supply gas, and the composition wavelength is 1.0 μm and the tensile strain amount is relatively low on the optical confinement layer 3 so as to cover the quantum dots 4. A low tensile strain InGaAsP spacer layer 5L having a composition of% is formed. The growth conditions of the low tensile strain InGaAsP spacer layer 5L are, for example, a supply amount of 2.5 ML, a growth rate of 0.1 μm / h, and a V / III ratio of 1600.

  For convenience of explanation, as shown in FIG. 1, the spacer layer may be expressed as “spacer layer 5” without distinguishing the relative magnitude of the tensile strain amount. In FIG. 2, the spacer layer is shown by distinguishing the relative magnitude of the tensile strain amount.

  In this way, the low tensile strain quantum dot portion QDL in which the low tensile strain spacer layer 5L is laminated on the quantum dot 4 is formed. The quantum dot 4 on which the low tensile strain spacer layer 5L is stacked may be referred to as a quantum dot 4L.

Next, the supply gas is changed to TMIn and AsH ₃ again, and InAs quantum dots 4 are formed on the low tensile strain spacer layer 5L. The growth conditions of the InAs quantum dots 4 are the same as those of the quantum dots 4 formed first, for example. However, for the second and subsequent quantum dots 4 from the bottom, the supply amount is reduced to, for example, 1.2 ML.

Next, the supply gas is changed to TMIn, TEGa, AsH ₃ , and PH ₃ again, covering the quantum dots 4 and on the low tensile strain spacer layer 5L, the composition wavelength is 1.0 μm and the tensile strain amount is relatively high. A high tensile strain InGaAsP spacer layer 5H having a composition of 2.4% is formed. The growth conditions of the high tensile strain InGaAsP spacer layer 5H are the same as those of the low tensile strain InGaAsP spacer layer 5, for example.

  In this way, the high tensile strain quantum dot portion QDH in which the high tensile strain spacer layer 5H is laminated on the quantum dot 4 is formed on the low tensile strain quantum dot portion QDL. The quantum dots 4 on which the high tensile strain spacer layer 5H is stacked may be referred to as quantum dots 4H.

  Thereafter, the process of forming the low tensile strain quantum dot portion QDL and the high tensile strain quantum dot portion QDH as one set is repeated four times.

  Next, the low tensile strain quantum dot portion QDL is formed on the uppermost high tensile strain quantum dot portion QDH. Finally, InAs quantum dots 4 are formed on the uppermost low tensile strain spacer layer 5L. In this way, columnar quantum dots 6 in which 12 layers of quantum dots 4 and 11 layers of spacer layers 5 are alternately stacked are formed.

  The columnar quantum dot 6 of the first embodiment has a structure in which the low tensile strain spacer layer 5L and the high tensile strain spacer layer 5H are alternately stacked, that is, the low tensile strain quantum dot portion QDL and the high tensile strain quantum. It has a structure in which the dot portions QDH are alternately stacked one by one.

Thereafter, the supply gas is TMIn, TEGa, AsH ₃ , and PH ₃ , the first columnar quantum dots 6 are covered, the InGaAsP intermediate layer 7 having a composition wavelength of 1.1 μm and a strain amount of 0% is formed. Form 30 nm.

  The process of forming columnar quantum dots 6 and intermediate layer 7 as a set is then repeated twice. In this way, a multi-layer stacked structure of columnar quantum dots is formed in which the three columnar quantum dot layers 6A, 6B, 6C are stacked via the intermediate layer 7. Note that the uppermost InGaAsP layer formed is called the optical confinement layer 3, not the intermediate layer 7.

  Lower optical confinement layer 3, first columnar quantum dot layer 6A, lower intermediate layer 7, second columnar quantum dot layer 6B, upper intermediate layer 7, third columnar quantum dot 6C, The active layer 8 is formed including the upper optical confinement layer 3.

Thereafter, the growth temperature is raised to 630 ° C. in a PH ₃ atmosphere, TMIn, PH ₃ , and DEZn are supplied, and the doping concentration is 5.0 × 10 ¹⁷ cm ⁻³ on the upper optical confinement layer ^3. The p-type InP cladding layer 9 is formed to a thickness of 200 nm.

  Further, the description will be made with reference to FIGS.

  FIG. 3 is a schematic cross-sectional view of the semiconductor optical amplifier according to the first embodiment viewed from the incident direction of the amplified light.

  FIG. 4 is a schematic cross-sectional view (in the thickness direction) of the semiconductor optical amplifier according to the first embodiment as seen from the side where the light to be amplified enters.

  As shown in FIG. 3, a dielectric mask MSK extending in the [110] direction and having a length of 100 μm and a width of 1.5 μm is formed on the p-type cladding layer 9, and dry using the dielectric mask MSK. By etching, the outer portion of the dielectric mask MSK is removed to a halfway depth of the substrate 1 to form a stripe mesa structure.

  Next, with the dielectric mask MSK formed, the p-type InP buried layer 10, the n-type InP block layer 11, and the p-type InP layer 12 are grown on both sides of the mesa to form a current confinement structure. For forming the p-type InP buried layer 10 to the p-type InP layer 12, a known semiconductor buried growth technique can be used. Thereafter, the dielectric mask MSK is removed.

  As shown in FIG. 4, a p-type InGaAs contact layer 13 is further formed on the p-type InP cladding layer 9, a p-side electrode 14 p is formed on the p-type InGaAs contact layer 13, and the back surface of the n-type InP substrate 1. An n-side electrode 14n is formed thereon, and an antireflection film 15 is formed on the end surfaces on the light incident side and the light emission side. In this way, the semiconductor optical amplifier of the first embodiment is formed.

  As shown in FIG. 2, a sample in which columnar quantum dots in which low tensile strain quantum dot portions QDL and high tensile strain quantum dot portions QDH are alternately stacked one by one is formed, The PL emission wavelength was measured.

  The tensile strain amount of the low tensile strain spacer layer 5L is kept constant at 0.4%, and the tensile strain amount of the high tensile strain spacer layer 5H is changed to 1.4% and 2.4% to prepare two types of samples. did. A sample having a low tensile strain amount of 0.4% and a high tensile strain amount of 1.4% has an average strain amount of 0.61% (compression strain), and a low tensile strain amount of 0.4% and a high tensile strain amount. The sample of 4% has an average strain amount of 0.33% (compression strain).

  FIG. 5 is a graph showing the relationship between the average strain amount of the columnar quantum dots of the first embodiment and the PL emission wavelength. The horizontal axis is the average strain amount expressed in units of%, and the vertical axis is the PL emission wavelength expressed in units of μm. The results of the first example are shown as triangular plots. What is indicated by a circle plot is the result of the comparative example similar to that shown in FIG.

  The sample with an average strain amount of 0.61% (low tensile strain amount 0.4% / high tensile strain amount 1.4%) in the first embodiment emits light having a wavelength of about 1.52 μm, which is almost the same as the comparative example. Indicated. The sample having an average strain amount of 0.33% (low tensile strain amount 0.4% / high tensile strain amount 2.4%) in the first example has a longer wavelength than the sample of the comparative example having the same average strain amount. The emission of about 1.50 μm wavelength was exhibited.

  As described above, it was found that the columnar quantum dots of the first example are suppressed from shortening the emission wavelength when the average strain amount is reduced as compared with the comparative example. Or it turned out that the average distortion amount when obtaining a desired light emission wavelength (for example, 1.50 micrometers) can be made small compared with a comparative example.

  Next, the reason why the columnar quantum dots of the first example showed long-wavelength PL emission with an equal average strain amount as compared with the comparative example will be considered. The reason for this is not clear, and the following discussion shows one way of thinking.

The magnitude of the tensile strain amount of the spacer layer of the comparative example is ε _0, and the magnitude of the strain amounts of the low tensile strain spacer layer and the high tensile strain spacer layer of the first embodiment are ε ₁ and ε ₂ , respectively. When the average strain amount is equal, the magnitudes of these tensile strain amounts satisfy ε ₁ <ε ₀ <ε ₂ .

The quantum dot in which the conduction band energy level of the quantum dot of the comparative example is E ₀ and the quantum dot 4L in which the low tensile strain spacer layer 5L is laminated and the high tensile strain spacer layer 5H in the first embodiment are laminated. Let the conduction band energy levels of 4H be E ₁ and E ₂ , respectively.

As can be seen from the comparative example, the PL emission wavelength tends to be shorter as the tensile strain amount of the spacer layer increases. From this, it is presumed that as the tensile strain amount of the spacer layer increases, the energy gap of the quantum dots covered by the spacer layer increases and the conduction band energy level of the quantum dots increases. Therefore, it is considered that E ₁ <E ₀ <E ₂ with respect to the conduction band energy level of the quantum dots.

  FIG. 6 is a schematic conduction band lineup of quantum dots of the columnar quantum dots of the comparative example and the first example. The conduction band lineup when the average strain is equal is shown side by side. The left side is a comparative example, and the right side is a first embodiment. “InGaAsP SCH” indicates an optical confinement layer (or intermediate layer).

The coupled quantum level formed in columnar quantum dot of the comparative example and E _u. It is formed on the columnar quantum dots of the first embodiment and E the coupled quantum level of the quantum dots 4L _m, the binding quantum level of the quantum dot 4H and E _n. Quantum dots 4L by binding quantum level E _m is strongly affected by the quantum dots 4L low energy level of E _1, believed to be less susceptible state the influence of high energy level E ₂ quantum dots 4H.

As a result, the coupled quantum level E _m has an energy level that is lower by ΔE than the coupled quantum level E _u , and the columnar quantum dot of the first example has a columnar quantum of the comparative example having an equal average strain amount. It is thought that light is emitted at a longer wavelength than the dots.

  As shown in FIG. 5, even in the sample of the first example, the sample having an average strain amount of 0.61% (low tensile strain amount 0.4% / high tensile strain amount 1.4%) is a comparative example. And emitted light at a wavelength of about the same level as the above. This is considered to be because when the strain amount of the high tensile strain spacer layer is small, the conduction band offset between the quantum dots 4L and 4H is small and the energy level is not lowered due to the coupling of wave functions.

  Next, the columnar quantum dot according to the second embodiment and a semiconductor optical amplifier using the same will be described. Differences from the first embodiment will be mainly described. The second embodiment is different from the first embodiment in the stacking order of the low tensile strain spacer layer 5L and the high tensile strain spacer layer 5H in the columnar quantum dots 6.

  The multi-layered structure of columnar quantum dots and the structure as a semiconductor optical amplifier are the same as those of the first embodiment. The structure of the columnar quantum dots 6 is replaced with that of the second embodiment, and FIGS. And FIG. 4 is also used to refer to the second embodiment. Further, in order to reduce the complexity of assigning reference numerals, reference numerals are used for members that clearly correspond to the first embodiment. The formation methods of the quantum dots 4, the low tensile strain spacer layer 5L, the high tensile strain spacer layer 5H, and the like are the same as in the first embodiment.

  FIG. 7 is an enlarged schematic sectional view showing one columnar quantum dot 6 of the second embodiment. In the same manner as in the first embodiment, an n-type InP buffer layer 2 and an InGaAsP light confinement layer 3 are formed from below on an n-type InP substrate 1, and InAs quantum dots 4 are formed on the InGaAsP light confinement layer 3, A low tensile strain InGaAsP spacer layer 5L is formed on the InAs quantum dots 4. In this way, first, a low tensile strain quantum dot portion QDL is formed.

  Next, InAs quantum dots 4 are formed on the low tensile strain spacer layer 5L. As in the first embodiment, the supply amount of the second and subsequent quantum dots 4 from the bottom is reduced to, for example, 1.2 ML.

  Next, the low tensile strain InGaAsP spacer layer 5L is formed again on the low tensile strain spacer layer 5L so as to cover the quantum dots 4. In this way, the low tensile strain quantum dot portion QDL is overlaid on the low tensile strain quantum dot portion QDL.

  Next, the third InAs quantum dots 4 from the bottom are formed on the second low tensile strain spacer layer 5L from the bottom. Next, the high tensile strain InGaAsP spacer layer 5H is formed on the low tensile strain spacer layer 5L so as to cover the third quantum dot 4 from the bottom, and the high tensile strain quantum dot portion QDH is formed. In the second embodiment, a structure in which two layers of low tensile strain quantum dot portions QDL are laminated and one layer of high tensile strain quantum dot portions QDH is laminated thereon is thus formed.

  The process of forming the low tensile strain quantum dot portion QDL, the low tensile strain quantum dot portion QDL, and the high tensile strain quantum dot portion QDH as one set is then repeated twice.

  Two lower tensile strain quantum dot portions QDL are stacked on the uppermost high tensile strain quantum dot portion QDH. Finally, the quantum dots 4 are formed on the uppermost low tensile strain spacer layer 5L. In this way, columnar quantum dots 6 in which 12 layers of quantum dots 4 and 11 layers of spacer layers 5 are alternately stacked are formed.

  The columnar quantum dot 6 of the second embodiment has a structure in which two low tensile strain spacer layers 5L and one high tensile strain spacer layer 5H are alternately laminated, that is, two low tensile strain quantum dots. Part QDL and one high tensile strain quantum dot part QDH are alternately stacked.

  Thereafter, as in the first embodiment, three columnar quantum dot layers 6A to 6C are stacked through the InGaAsP intermediate layer 7 to form the active layer 8, and the p-type InP cladding layer 9 is formed. The structure shown in is obtained.

  Further, a stripe mesa structure is formed in the same manner as described in the first embodiment with reference to FIGS. 3 and 4, and the p-type InGaAs contact layer 13, the p-side electrode 14p, the n-side electrode 14n, and the reflection are formed. The prevention film 15 is formed. In this way, the semiconductor optical amplifier of the second embodiment is formed.

  As shown in FIG. 7, a sample in which one columnar quantum dot in which two layers of low tensile strain quantum dot portions QDL and one layer of high tensile strain quantum dot portions QDH are alternately stacked is formed. The PL emission wavelength was prepared and measured.

  The tensile strain amount of the low tensile strain spacer layer 5L is kept constant at 0.4%, and the tensile strain amount of the high tensile strain spacer layer 5H is changed to 1.4%, 2.4%, and 3.4%. A seed sample was made. A sample having a low tensile strain of 0.4% and a high tensile strain of 1.4% has an average strain of 0.74% (compression strain), and a low tensile strain of 0.4% and a high tensile strain of 2. The 4% sample has an average strain of 0.56% (compression strain), and the low tensile strain of 0.4% and the high tensile strain of 3.4% sample has an average strain of 0.39% (compression). Distortion).

  Refer to FIG. 5 again. FIG. 5 also shows the relationship between the average strain amount of the columnar quantum dots of the second embodiment and the PL emission wavelength. The results of the second example are shown as square plots.

  The sample having an average strain amount of 0.74% (low tensile strain amount 0.4% / high tensile strain amount 1.4%) of the second embodiment emits light having a wavelength of about 1.542 μm, which is almost the same as the comparative example. Indicated. A sample having an average strain amount of 0.56% (low tensile strain amount 0.4% / high tensile strain amount 2.4%) and an average strain amount of 0.39% (low tensile strain amount 0. 4% / high tensile strain amount 3.4%) each emitted light having a longer wavelength and a wavelength of about 1.538 μm and a wavelength of about 1.526 μm than the sample of the comparative example having the same average strain amount. .

  Also in the columnar quantum dots of the second example, as in the first example, the shortening of the emission wavelength when the average strain amount is reduced is suppressed as compared with the comparative example. The inclination of shortening the emission wavelength due to the decrease in the average strain amount tends to be gentler in the second embodiment than in the first embodiment.

  In the second embodiment, light emission with a longer wavelength can be obtained with the same average strain amount than in the first embodiment. It can be said that when obtaining a desired long-wavelength emission wavelength, the average strain amount can be further reduced.

  Next, the reason why the columnar quantum dots of the second example exhibited longer wavelength PL emission with the same average strain amount than that of the first example will be considered. The reason for this is not clear, and the following discussion shows one way of thinking.

  In the first embodiment, low tensile strain quantum dot portions QDL and high tensile strain quantum dot portions QDH are alternately stacked one by one. The quantum dots 4L included in the low tensile strain quantum dot portion QDL are affected by the low tensile strain spacer layer 5L covering the quantum dots 4L, and are also somewhat effective from the high tensile strain spacer layer 5H of the high tensile strain quantum dot portion QDH immediately below. Receive. Therefore, it is considered that the action received from the low tensile strain spacer layer 5L and the action received from the high tensile strain spacer layer 5H are somewhat averaged. This also applies to the quantum dots 4H. Due to this, it is considered that the energy level difference between the quantum dots 4L and 4H decreases.

  On the other hand, in the second embodiment, two layers of low tensile strain quantum dot portions QDL and one layer of high tensile strain quantum dot portions QDH are alternately stacked. The quantum dots 4L (which will be referred to as quantum dots 4LU) included in the upper low tensile strain quantum dot portion QDL of the two layers of the low tensile strain quantum dot portion QDL overlapped with each other, and Both of the spacer layers immediately below the low tensile strain spacer layer 5L. For this reason, it is considered that the quantum dots 4LU are not easily affected by the high tensile strain spacer layer 5H, and the energy level is easily kept low. Therefore, in the second embodiment, it is considered that the energy level difference between the quantum dots 4L and 4H is less likely to decrease than in the first embodiment, and a low coupled quantum level is likely to be obtained.

  In addition, in 2nd Example, although it was set as the alternately laminated structure of the low tensile strain quantum dot part QDL of 2 layers, and the high tensile strain quantum dot part QDH of 1 layer, the low tensile strain quantum dot part QDL is three or more layers. It is also considered that two or more high tensile strain quantum dot portions QDH can be stacked. However, the stack of the high tensile strain quantum dot portion QDH disposed between the stack portion L1 having the low tensile strain quantum dot portion QDL and the stack portion L2 of the low tensile strain quantum dot portion QDL disposed immediately above this portion. The portion has a thickness within a range in which the quantum dots 4L included in the stacked portion L1 and the quantum dots 4L included in the stacked portion L2 can be coupled to each other quantum mechanically.

  As described above, as described in the first and second embodiments, columnar quantum dots having a structure in which a plurality of low tensile strain quantum dot portions QDL and high tensile strain quantum dot portions QDH are alternately arranged in the thickness direction are formed. By doing so, compared to the columnar quantum dots having a uniform tensile strain amount of the spacer layer, it is possible to suppress the emission wavelength from being shortened due to the decrease in the average strain amount, or to obtain long wavelength light emission with the same average strain amount. It becomes easy.

  When obtaining a desired emission wavelength, the average strain can be reduced, so that it becomes easy to stack columnar quantum dots in a multi-stage while maintaining good crystallinity. Note that, since the average strain amount can be reduced, the crystal quality of the semiconductor layer formed above the columnar quantum dots can be improved not only when the columnar quantum dots are stacked in multiple stages.

  Further, as described in the second embodiment, by stacking a plurality of low tensile strain quantum dot portions QDL (between a certain high tensile strain quantum dot portion QDH and a high tensile strain quantum dot portion QDH immediately above it) By arranging a plurality of stacked low tensile strain quantum dot portions QDL), it is easy to make the emission wavelength at a predetermined average strain amount longer.

  In the first and second embodiments, an n-type InP (001) substrate is used. However, a p-type InP (001) substrate, a high resistance (SI) InP (001) substrate, and the like can also be applied. . In that case, the conductivity type, electrode arrangement, embedded structure, and the like can be changed as appropriate.

  Furthermore, it is considered that a semiconductor substrate other than the InP substrate can be used as the substrate on which the columnar quantum dots are formed.

  Next, formation of columnar quantum dots on a GaAs substrate will be described as third and fourth embodiments. In order to reduce the complexity of assigning reference signs, reference signs are used for members and the like that clearly correspond to the first and second embodiments.

  A columnar quantum dot according to a third embodiment and a semiconductor optical amplifier using the same will be described.

  FIG. 8 is a schematic cross-sectional view showing a multi-layered portion of columnar quantum dots included in the semiconductor optical amplifier of the third embodiment.

  FIG. 9 is an enlarged schematic cross-sectional view showing one columnar quantum dot of the third embodiment.

  The description will be given along the manufacturing process with reference to FIGS. As the crystal growth method, MOVPE can be used as in the first and second embodiments, and the same raw material as in the first and second embodiments can be used.

The n-type GaAs (001) substrate 1 is equipped in a MOVPE growth furnace and the growth temperature is raised to 630 ° C. in an AsH ₃ atmosphere.

Next, TEGa and SiH ₄ are also supplied to form the n-type GaAs buffer layer 2 with a thickness of 500 nm.

The supply of TEGa and SiH ₄ is stopped, and the growth temperature is lowered to 430 ° C. in an AsH ₃ atmosphere. Next, the supply gas is TEGa and AsH ₃ , and the non-doped GaAs optical confinement layer 3 is formed to a thickness of 40 nm on the buffer layer 2.

  Next, in the same manner as in the first example, InAs quantum dots 4 (4L) are formed on the optical confinement layer 3, and a low tensile strain InGaAsP spacer layer 5L is formed on the quantum dots 4 (4L), thereby reducing the low tension. An InAs quantum dot 4 (4H) is formed on the strained InGaAsP spacer layer 5L, and a high tensile strain InGaAsP spacer layer 5H is formed on the quantum dot 4 (4H).

  The conditions for forming the InAs quantum dots 4 are the same as in the first embodiment, for example. The low tensile strain InGaAsP spacer layer 5L has, for example, a composition wavelength of 0.75 μm and a tensile strain of 0.4%, and the high tensile strain InGaAsP spacer layer 5H has, for example, a composition wavelength of 0. The composition is such that the tensile strain amount is 1.0% at 75 μm.

  In this way, a structure in which the high tensile strain quantum dot portion QDH is stacked on the low tensile strain quantum dot portion QDL is formed.

  Thereafter, the process of forming the low tensile strain quantum dot portion QDL and the high tensile strain quantum dot portion QDH as one set is repeated three times. Next, the low tensile strain quantum dot portion QDL is formed on the uppermost high tensile strain quantum dot portion QDH. Finally, InAs quantum dots 4 are formed on the uppermost low tensile strain spacer layer 5L. In this way, columnar quantum dots 6 in which ten quantum dots 4 and nine spacer layers 5 are alternately stacked are formed. The columnar quantum dots 6 of the third embodiment are expected to emit light at a wavelength of about 1.06 μm.

  The columnar quantum dots 6 of the third embodiment have a structure in which low tensile strain quantum dot portions QDL and high tensile strain quantum dot portions QDH are alternately stacked one by one, as in the first embodiment.

Thereafter, the supply gas is set to TEGa and AsH _3, and the GaAs intermediate layer 7 is formed to a thickness of 40 nm, for example, covering the first columnar quantum dots 6.

  The process of forming columnar quantum dots 6 and intermediate layer 7 as a set is then repeated three times. In this way, a multi-layer stacked structure of columnar quantum dots is formed in which the four columnar quantum dot layers 6 </ b> A to 6 </ b> D are stacked via the intermediate layer 7. Note that the uppermost GaAs layer is called the light confinement layer 3, not the intermediate layer 7.

  Lower optical confinement layer 3, first columnar quantum dot layer 6A, first intermediate layer 7 from the bottom, second columnar quantum dot layer 6B, second intermediate layer 7 from the bottom, third step An active layer 8 is formed including the columnar quantum dot layer 6C, the third intermediate layer 7 from the bottom, the fourth columnar quantum dot layer 6D, and the upper optical confinement layer 3.

Thereafter, the growth temperature is raised to 630 ° C. in an AsH ₃ atmosphere, TEGa, AsH ₃ , and DEZn are supplied, and a p-type GaAs cladding layer 9 is formed on the upper optical confinement layer 3 with a thickness of, for example, 1 Form 5 μm.

  The description will be continued with reference to FIG. FIG. 10 is a schematic cross-sectional view of the semiconductor optical amplifier according to the third embodiment viewed from the incident direction of the amplified light. A p-type GaAs contact layer 13 is formed on the p-type GaAs cladding layer 9. Next, the portion other than the p-side electrode formation region is etched to a depth halfway of the p-type cladding layer 9. A p-side electrode 14p is formed on the remaining p-type contact layer 13, and an n-side electrode 14n is formed on the back surface of the n-type GaAs substrate 1 to form a ridge structure.

  Furthermore, FIG. 4 is also referred to and the description is advanced. A schematic cross-sectional view (in the thickness direction) of the semiconductor optical amplifier according to the third embodiment, viewed from the side where the light to be amplified enters, is the same as the semiconductor optical amplifier according to the first and second embodiments. 4 and so on. Antireflection films 15 are formed on the end surfaces on the light incident side and the light emission side. In this way, the semiconductor optical amplifier of the third embodiment is formed.

  A columnar quantum dot according to a fourth embodiment and a semiconductor optical amplifier using the same will be described. The fourth embodiment is different from the third embodiment in the stacking order of the low tensile strain spacer layer 5L and the high tensile strain spacer layer 5H in the columnar quantum dots 6. The structure as a semiconductor optical amplifier is the same as that of the third embodiment, and FIG. 11 and FIG. 4 are used to refer to the fourth embodiment. The formation methods of the quantum dots 4, the low tensile strain spacer layer 5L, the high tensile strain spacer layer 5H, and the like are the same as in the third embodiment.

  FIG. 11 is a schematic cross-sectional view showing a multi-layered portion of columnar quantum dots included in the semiconductor optical amplifier of the fourth embodiment.

  FIG. 12 is an enlarged schematic sectional view showing one columnar quantum dot of the fourth embodiment.

  Similarly to the third embodiment, an n-type GaAs buffer layer 2 and a GaAs light confinement layer 3 are formed on the n-type GaAs substrate 1 from below, and an InAs quantum dot 4 is formed on the GaAs light confinement layer 3. A low tensile strain InGaAsP spacer layer 5L is formed on the InAs quantum dots 4.

  An InAs quantum dot 4 (4L) is formed on the low tensile strain spacer layer 5L, and a low tensile strain InGaAsP spacer layer 5L is formed again on the quantum dot 4 (4L).

  InAs quantum dots 4 (4H) are formed on the second low tensile strain spacer layer 5L from the bottom, and high tensile strain InGaAsP spacer layers 5H are formed on the quantum dots 4 (4H).

  In this manner, a structure in which two layers of low tensile strain quantum dot portions QDL are stacked and one layer of high tensile strain quantum dot portions QDH is stacked thereon is formed.

  The process of forming the low tensile strain quantum dot portion QDL, forming the low tensile strain quantum dot portion QDL, and forming the high tensile strain quantum dot portion QDH as one set is performed once thereafter.

  Two lower tensile strain quantum dot portions QDL are stacked on the uppermost high tensile strain quantum dot portion QDH. Finally, the quantum dots 4 are formed on the uppermost low tensile strain spacer layer 5L. In this way, columnar quantum dots 6 in which nine quantum dots 4 and eight spacer layers 5 are alternately stacked are formed. The columnar quantum dots 6 of the fourth embodiment are expected to emit light at a wavelength of about 1.2 μm to 1.3 μm.

  As in the second embodiment, the columnar quantum dots 6 of the fourth embodiment have a structure in which two layers of low tensile strain quantum dot portions QDL and one layer of high tensile strain quantum dot portions QDH are alternately stacked. Have.

  Thereafter, as in the third embodiment, four columnar quantum dot layers 6A to 6D are stacked through the GaAs intermediate layer 7 to form the active layer 8, and the p-type GaAs cladding layer 9 is formed. The structure shown in is obtained.

  Furthermore, the semiconductor optical amplifier of the fourth embodiment is formed through the same steps as those described with reference to FIGS. 10 and 4 in the third embodiment.

  In the third and fourth embodiments, an n-type GaAs substrate is used. However, a p-type GaAs substrate, a high resistance (SI) GaAs substrate, and the like can also be applied. In that case, the conductivity type, electrode arrangement, and the like can be changed as appropriate.

  The structural parameters, materials, device structures, etc. are not limited to those described in the first to fourth embodiments.

  For example, the supply amount of InAs quantum dots, the composition wavelength, thickness, and strain of the InGaAsP spacer layer are not limited to the values in the first to fourth embodiments. It is not essential that the composition wavelengths of the low tensile strain spacer layer and the high tensile strain spacer layer be equal.

Further, for example, it is considered that InSb, InGaAs, InGaAsSb, etc. can be applied as the quantum dot material in addition to InAs. If the composition is expressed explicitly, it is considered that In _x Ga _1-x As _y Sb _1-y (0 <x ≦ 1, 0 ≦ y ≦ 1) can be used as the material of the quantum dots. However, InAs is preferable from the viewpoint that the in-plane symmetry of the quantum dots is good.

  Further, for example, it is considered that AlGaInAs can be applied as a material for the spacer layer in addition to InGaAsP. However, when the growth temperature of the quantum dots is low, InGaAsP is more preferable from the viewpoint of improving the crystallinity of the spacer layer.

  Further, for example, the semiconductor optical amplifier is shown as the semiconductor element of the first to fourth embodiments. However, it is possible to manufacture a semiconductor laser by cleaving or coating a highly reflective film on both end faces in the same manufacturing process.

  In the first to fourth embodiments, the uppermost part of the columnar quantum dots is structured such that the quantum dots are not covered with the spacer layer. However, the uppermost part of the columnar quantum dots is covered with the spacer layer. It can also be a structure.

  Although the present invention has been described with reference to the embodiments, the present invention is not limited thereto. It will be apparent to those skilled in the art that various modifications, improvements, combinations, and the like can be made.

The following additional notes are further disclosed regarding the embodiment including the first to fourth examples described above.
(Appendix 1)
A semiconductor substrate;
Columnar quantum dots formed above the semiconductor substrate,
The columnar quantum dot is
A first portion formed of a first quantum dot and a first spacer layer stacked on the first quantum dot and having a tensile strain with respect to the semiconductor substrate;
A second portion formed of a second quantum dot and a second spacer layer stacked on the second quantum dot and having a tensile strain larger than that of the first spacer layer with respect to the semiconductor substrate;
A semiconductor element having a structure arranged alternately in the thickness direction.
(Appendix 2)
The semiconductor element according to appendix 1, wherein the columnar quantum dot has a structure in which a plurality of the first parts are stacked between the second part and the second part immediately above the second part.
(Appendix 3)
3. The semiconductor element according to appendix 1 or 2, wherein the columnar quantum dot has a structure in which a plurality of the first portions and a single layer of the second portion are alternately stacked.
(Appendix 4)
The semiconductor element according to any one of appendices 1 to 3, wherein the semiconductor substrate is made of InP.
(Appendix 5)
The semiconductor element according to any one of appendices 1 to 3, wherein the semiconductor substrate is made of GaAs.
(Appendix 6)
Wherein the first quantum dot and the second quantum _dot, either _{In x Ga 1-x As y} Sb 1-y (0 <x ≦ 1,0 ≦ y ≦ 1) is formed by Appendix 1-5 The semiconductor element as described in one.
(Appendix 7)
The semiconductor element according to any one of appendices 1 to 6, wherein the first spacer layer and the second spacer layer are made of InGaAsP or AlGaInAs.
(Appendix 8)
A semiconductor substrate;
Columnar quantum dots stacked on a plurality of stages via an intermediate layer above the semiconductor substrate,
Each of the plurality of columnar quantum dots stacked is
A first portion formed of a first quantum dot and a first spacer layer stacked on the first quantum dot and having a tensile strain with respect to the semiconductor substrate;
A second portion formed of a second quantum dot and a second spacer layer stacked on the second quantum dot and having a tensile strain larger than that of the first spacer layer with respect to the semiconductor substrate;
A semiconductor element having a structure arranged alternately in the thickness direction.
(Appendix 9)
further,
A first conductivity type first semiconductor layer disposed between the semiconductor substrate and the plurality of stacked columnar quantum dots;
A second semiconductor layer of a second conductivity type disposed above the columnar quantum dots stacked in a plurality of stages and opposite to the first conductivity type;
Item 9. The semiconductor element according to appendix 8, which is a semiconductor optical amplifier or a semiconductor laser.
(Appendix 10)
A first quantum dot is formed above the semiconductor substrate, a first spacer layer having a tensile strain with respect to the semiconductor substrate is stacked on the first quantum dot, and the first quantum dot is formed on the first quantum dot. Forming a first portion in which one spacer layer is laminated;
A second quantum dot is formed above the first spacer layer, and a second spacer layer having a tensile strain larger than that of the first spacer layer is stacked on the semiconductor substrate on the second quantum dot. And forming a second portion in which the second spacer layer is laminated on the second quantum dots,
The process of forming the first part and the process of forming the second part are repeated so that a structure in which the first part and the second part are alternately arranged in the thickness direction is formed. A method for manufacturing a semiconductor device.

DESCRIPTION OF SYMBOLS 1 Substrate 2 Buffer layer 3 Optical confinement layer 4 Quantum dot 4L Quantum dot 4H laminated with low tensile strain spacer layer Quantum dot 5 laminated with high tensile strain spacer layer 5 Spacer layer 5L Low tensile strain spacer layer 5H High tensile strain spacer Layer QDL Low tensile strain quantum dot portion QDH High tensile strain quantum dot portion 6 Columnar quantum dots 6A, 6B, 6C, 6D Columnar quantum dot layer 7 Intermediate layer 8 Active layer 9 Cladding layer 10 p-type InP buried layer 11 n-type InP block Layer 12 p-type InP layer 13 contact layer 14p, 14n electrode 15 antireflection film

Claims (5)

A semiconductor substrate;
Columnar quantum dots formed above the semiconductor substrate,
The columnar quantum dot is
A first portion formed of a first quantum dot and a first spacer layer stacked on the first quantum dot and having a tensile strain with respect to the semiconductor substrate;
A second portion formed of a second quantum dot and a second spacer layer stacked on the second quantum dot and having a tensile strain larger than that of the first spacer layer with respect to the semiconductor substrate;
A semiconductor element having a structure arranged alternately in the thickness direction.   2. The semiconductor element according to claim 1, wherein the columnar quantum dot has a structure in which a plurality of the first parts are stacked between a certain second part and the second part immediately above the second part. A semiconductor substrate;
Columnar quantum dots stacked on a plurality of stages via an intermediate layer above the semiconductor substrate,
Each of the plurality of columnar quantum dots stacked is
A first portion formed of a first quantum dot and a first spacer layer stacked on the first quantum dot and having a tensile strain with respect to the semiconductor substrate;
A second portion formed of a second quantum dot and a second spacer layer stacked on the second quantum dot and having a tensile strain larger than that of the first spacer layer with respect to the semiconductor substrate;
A semiconductor element having a structure arranged alternately in the thickness direction. further,
A first conductivity type first semiconductor layer disposed between the semiconductor substrate and the plurality of stacked columnar quantum dots;
A second semiconductor layer of a second conductivity type disposed above the columnar quantum dots stacked in a plurality of stages and opposite to the first conductivity type;
The semiconductor device according to claim 3, which is a semiconductor optical amplifier or a semiconductor laser. A first quantum dot is formed above the semiconductor substrate, a first spacer layer having a tensile strain with respect to the semiconductor substrate is stacked on the first quantum dot, and the first quantum dot is formed on the first quantum dot. Forming a first portion in which one spacer layer is laminated;
A second quantum dot is formed above the first spacer layer, and a second spacer layer having a tensile strain larger than that of the first spacer layer is stacked on the semiconductor substrate on the second quantum dot. And forming a second portion in which the second spacer layer is laminated on the second quantum dots,
The process of forming the first part and the process of forming the second part are repeated so that a structure in which the first part and the second part are alternately arranged in the thickness direction is formed. A method for manufacturing a semiconductor device.