JP2010103430A - Structure and semiconductor optical apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a germanium structure increasing a reduction effect of direct energy gap E<SB>0</SB>by tensile strain. <P>SOLUTION: The germanium structure has a germanium particle and a buried layer covering over the germanium particle for embedding the germanium particle, and the buried layer generates the tensile strain in the germanium particle in each direction of three crystal axes of the germanium particle. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a structure including germanium and an optical semiconductor device in which a light receiving layer or a light emitting layer is formed by the structure.

  As one of the elemental technologies of optical interconnections between semiconductor integrated circuit devices (semiconductor chips) and optical interconnections connecting between boards equipped with semiconductor integrated circuit devices, light emitting elements and light receiving elements on a silicon (Si) substrate There is a need for a technology for integrating elements.

  The light emitting element and the light receiving element are usually formed of a III-V semiconductor (for example, InGaAsP) that absorbs light by direct transition and emits light.

  III-V compound semiconductors are polar semiconductors in which ionic bonds contribute to the bonds between atoms. On the other hand, Si is a nonpolar semiconductor in which bonds between atoms are formed only by covalent bonds. For this reason, when trying to grow a III-V compound semiconductor on a Si substrate, various difficulties are encountered.

  On the other hand, germanium (Ge) is the same nonpolar semiconductor as Si and has high affinity with Si. For this reason, the crystal growth of Ge on the Si substrate is easy. From such a viewpoint, integration of a Ge light receiving element and a light emitting element on a Si substrate is expected.

  However, Ge is an indirect transition semiconductor. Therefore, it is extremely difficult to form a light emitting element with Ge.

  On the other hand, Ge is one of the few group IV semiconductors used for forming a light receiving element.

  The fundamental absorption edge of Ge is at 1.85 μm (at 300 K) corresponding to an indirect energy gap of 0.67 eV between the Γ point that is the top of the valence band and the L point that is the bottom of the conduction band. However, light absorption due to indirect transition is extremely weak, so that light is not substantially absorbed in the vicinity of 1.85 μm.

On the other hand, the direct energy gap E _{0 at} the Γ point of Ge is 0.81 eV. For this reason, Ge strongly absorbs light on a shorter wavelength side than 1.54 μm corresponding to 0.81 eV. For this reason, a light receiver using Ge as a light absorption layer detects light on a shorter wavelength side than 1.54 μm.

  The optical interconnect is a technology that enables short-distance optical communication between semiconductor chips and boards. Therefore, from the viewpoint of reducing transmission loss, it is not always necessary to use a wavelength band (C band, L band, U band, etc.) used in optical fiber communication for communication between semiconductor chips and boards. However, various optical components (such as optical couplers) have been developed according to the wavelength band for optical fiber communication. Therefore, it is preferable to use the same wavelength band as the optical fiber communication in the optical interconnect.

  By the way, as described above, the wavelength band in which Ge absorbs light is substantially on the shorter wavelength side than 1.54 μm, and only covers a part of the C band (1.53 μm−1.57 μm). .

In view of such a situation, an attempt has been made to reduce the direct energy gap E ₀ at the Γ point by generating tensile strain in the Ge thin film (Non-patent Document 1).

When the direct energy gap E ₀ at the Γ point is reduced, the substantial light absorption edge of Ge shifts to a longer wavelength side than 1.54 μm (that is, shifts down). If downshifting of the light absorption edge is realized in this way, the operating wavelength band of the optical receiver using Ge as the light receiving layer is widened, and its usefulness is high, particularly useful as a Ge optical receiver for optical interconnects. Become.

Further, if the direct energy gap E ₀ between the Γ points is reduced to be smaller than the indirect energy gap between the Γ point and the L point, a light emitting element can be formed using Ge.
"Perfectly tetragonal, tensile-strained Ge on Ge1-ySny buffered Si (100)", Appled Physics Letters 90, 061915 (2007). “Theoretical calculations of heterojunction discontinuities in the Si / Ge system”, Physical Review B 34, 5621 (1986).

However, the amount of reduction of the direct energy gap E ₀ expected by the above attempt is not sufficient. For this reason, depending on the germanium thin film according to the above attempt, the light absorption band of Ge (based on direct transition) cannot cover the U band (1.63 μm-1.68 μm) as well as light emission by Ge.

Accordingly, an object of the present invention is to provide a germanium structure that increases the effect of reducing the direct energy gap E ₀ due to tensile strain, and a semiconductor optical device using the germanium structure.

  In order to achieve the above object, the structure includes germanium particles and an embedded layer that embeds the germanium particles so as to cover the periphery of the germanium particles, and the embedded layer includes three crystals of the germanium particles. In each of the axial directions, tensile strain is generated in the germanium particles.

  In the optical semiconductor device, the structure is a light-receiving layer that receives light to generate photocarriers or a light-emitting layer that generates light by injecting current.

According to this structure, in the structure (germanium structure) including germanium, the effect of reducing the direct energy gap E ₀ due to tensile strain can be increased.

  Further, in this optical semiconductor device, since the light receiving layer is formed of the germanium structure, light detection in the optical communication wavelength band (particularly, C band to U band) is possible in the optical semiconductor device having Ge as the light receiving layer. become. Therefore, according to the present optical semiconductor device, an optical semiconductor device formed of Ge having a high affinity with Si can be integrated on the Si substrate to form an integrated circuit device for interconnect.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the technical scope of the present invention is not limited to these embodiments, but extends to the matters described in the claims and equivalents thereof.

(Reduction of _{E 0} due to tensile strain)
First, the method (related technology) and the principle of reducing the direct energy gap E ₀ of the Ge thin film by introducing the tensile strain described above will be described together with the inventors' unique knowledge.

(1) Formation of strained Ge thin film FIG. 1 is a cross-sectional view illustrating the configuration of a Ge thin film in which tensile strain is introduced to directly reduce the energy gap E ₀ . FIG. 1 shows the crystal axis of the Ge thin film and the coordinate axes (x axis, y axis, z axis) used for the definition of strain described later.

In order to directly reduce the energy gap E ₀ by introducing tensile strain into the Ge thin film 2, first, a germanium-tin mixed crystal buffer layer (Ge _1-x Sn _x buffer layer 6) is formed on the Si (100) substrate 4. Is formed. Here, the composition ratio x of Sn is 0.015 to 0.025. Although the lattice constant of Ge _1-x Sn _x is larger than that of Si, the Ge _1-x Sn _x buffer layer 6 is formed in a state in which strain due to lattice mismatch with the Si (100) substrate 4 is almost completely relaxed. The

A Ge thin film 2 having a thickness of 120 to 240 nm is formed on the Ge _1-x Sn _x buffer layer 6.

By the way, the lattice constant of Ge is smaller than that of the Ge _1-x Sn _x mixed crystal. For this reason, biaxial tensile strain is generated in the Ge thin film 6. As will be described later, this tensile strain directly reduces the energy gap E ₀ .

FIG. 2 is a diagram for explaining the atomic arrangement at the interface 10 between the Ge thin film 2 and the Ge _1-x Sn _x buffer layer 6. As shown in FIG. 2, the atoms (Ge atoms 18) constituting the Ge thin film 2 and the atoms (Ge atoms 20 and Sn atoms 22) constituting the Ge _1-x Sn _x buffer layer 6 have continuous atomic arrangement. As shown in FIG.

As is well known, the crystal structure of Ge is a diamond structure. On the other hand, the crystal structure of Ge _1-x Sn _x is a structure in which one Ge atom or one Sn atom is arranged at each lattice point of the diamond structure.

Therefore, the arrangement of the atoms (Ge18) constituting the Ge thin film 2 and the atoms (Ge20 or Sn22) constituting the Ge _1-x Sn _x buffer layer 6 are the same (being focused on the arrangement of the atoms, Don't focus on the difference.) In the structure described with reference to FIG. 1, the Ge thin film 2 is formed on the Ge _1-x Sn _x buffer layer 6 by epitaxial growth.

Therefore, the atoms constituting the Ge thin film 2 (Ge atoms 18) and the atoms constituting the Ge _1-x Sn _x buffer layer 6 (Ge atoms 20 and Sn atoms 22) are arranged as shown in FIG. Are bonded at the interface 10 so that they are continuous.

Thus, since the Ge thin film 2 and the Ge _1-x Sn _x buffer layer 6 are connected so that their atomic arrangements are continuous at the interface 10, the lattice constants of both coincide in a direction parallel to the interface.

By the way, in the related technique described with reference to FIG. 1, the Ge _1-x Sn _x buffer layer 6 is formed sufficiently thicker than the Ge thin film 2. For this reason, the lattice constant of the Ge _1-x Sn _x buffer layer 6 hardly changes. On the other hand, the lattice constant of the Ge thin film 2 changes in a direction parallel to the interface 10 so as to coincide with the lattice constant of the Ge _1-x Sn _x buffer layer 6.

By the way, the lattice constant of Ge _1-x Sn _x is larger than that of Ge. Accordingly, tensile strain (biaxial tensile strain) is generated in the Ge thin film 2 in a direction parallel to the interface 10.

However, if the difference in lattice constant between the Ge thin film 2 and the Ge _1-x Sn _x buffer layer 6 becomes too large, crystal defects are generated in the Ge thin film 2 and the tensile strain is relaxed. Therefore, there is an upper limit to the strain that can be generated in the Ge thin film 2 by the structure described with reference to FIG. This upper limit is 0.0025 (0.25%) if expressed by a strain ε (= ε _xx = ε _yy ) described later.

(2) Change in E ₀ due to strain FIG. 3 is a diagram for explaining the band structure in the wave number space of Ge. The right half of the horizontal axis is the wave number along the Δ axis from the Γ point (wave number k = 0) toward the X point at the end in the <100> direction. On the other hand, the left half of the horizontal axis is the wave number along the Λ axis from the Γ point (wave number k = 0) to the L point at the end in the <111> direction. The vertical axis represents the energy of electrons with respect to the top 8 of the valence band.

  As shown in FIG. 3, the top 8 of the valence band of Ge and the bottom 9 of the conduction band are at the Γ point and the L point, respectively. For this reason, in order for electrons occupying the top 8 of the valence band of Ge to absorb (or emit) light and make a transition to the bottom 9 of the conduction band, the wave number conservation law must be satisfied by interposing phonons. I must. Further, in order for electrons occupying the bottom 9 of the conduction band of Ge to recombine with holes occupying the top 8 of the valence band, the law of conservation of wave number must be satisfied with phonons.

  In this way, transitions between levels having different wave numbers, that is, indirect transitions, require phonon intervention, so the transition probabilities are several compared to direct transitions that do not require phonon intervention (transitions that do not involve wave number changes). An order of magnitude lower. For this reason, Ge hardly absorbs light (however, the fundamental absorption) near the (short wavelength side) 1.85 μm fundamental absorption edge (corresponding to the energy difference between the bottom 9 of the conduction band and the top 8 of the valence band). The wavelength of 1.85 μm at the end is the value when the temperature is 300K).

However, the transition between the bottom 11 of the conduction band and the top 8 of the valence band at the Γ point is a direct transition, and the transition probability is high. Therefore, Ge strongly absorbs light on the shorter wavelength side than the wavelength of 1.54 μm corresponding to the energy gap based on the direct transition, that is, the direct energy gap E ₀ .

By the way, when distortion occurs in the semiconductor, the band structure changes. For this reason, when distortion occurs, the energy gap E ₀ also changes directly.

The change ΔE _c ^Γ of the direct energy gap E ₀ due to strain is generally calculated by numerical calculation in consideration of the deformation potential (Non-Patent Document 2).

Except in special cases, it is difficult to explicitly describe ΔE _c ^Γ . However, it is possible to approximate ΔE _c ^Γ by the following equation.

Where a _c ^Γ is the rate of change of the direct energy gap E _{0 at} the Γ point with respect to strain. As shown in FIG. 1, when the coordinate axes (x axis, y axis, z axis) and the crystal axes ([100], [010], [001]) are matched, in the case of Ge, a _c ^Γ is −9.48 eV. become.

On the other hand, ε _xx , ε _yy , and ε _zz are distortion components and are defined by the following equations.

  Here, x, y, and z are the position coordinates of the point P inside the solid.

On the other hand, u _x , u _y , and u _z are displacements of the point P induced by the elastic deformation of the solid. In other words, by elastic deformation, the solid internal point P is displaced from the coordinate (x, y, z) to the coordinate _{(x + u x, y +} u y, z + u z). When the deformation of the solid is uniform, ε _xx , ε _yy , and ε _zz are constants rather than position functions.

In the following description, each of ε _xx , ε _yy , and ε _zz may be considered as a constant.

(3) Principle of E ₀ reduction By the way, the lattice constant of Ge _1-x Sn _x is larger than the lattice constant of Ge.

Therefore, as described above, biaxial tensile strain is generated in the Ge thin film 2 formed on the Ge _1-x Sn _x buffer layer 6. That is, in the Ge thin film 2, strains ε _xx and ε _yy are generated in the x-axis direction and the y-axis direction parallel to the main surface (001) of the Si substrate 4.

By the way, the sign of tensile strain is positive. Therefore, ε _xx and ε _yy are positive values.

On the other hand, the strain ε _zz in the direction perpendicular to the main surface (001) of the Si substrate 4 takes a negative value due to the Poisson effect. Specifically, it is expressed by the following equation.

Here, C ₁₁ and C ₁₂ are elastic constants (elastic compliance constants) and are components of the elastic modulus tensor [C].

C ₁₁ and C ₁₂ of Ge are 13.15 × 10 ¹¹ dyn / cm ² and 4.13 × 10 ¹¹ dyn / cm ² , respectively.

Therefore, the change ΔE _c ^Γ of the direct energy gap E _{0 in} the Ge thin film 2 is expressed as follows.

By the way, in biaxial strain, ε _xx and ε _yy are equal. Therefore, when ε _xx and ε _yy are expressed by ε (= ε _xx = ε _yy ), ΔE _c ^Γ is expressed by the following equation.

That is, when the Ge thin film 2 is grown on the Ge _1-x Sn _x buffer layer 6 having a larger lattice constant than Ge, the energy gap E ₀ is directly reduced according to the equation (5).

(4) Problem As described above, the upper limit of ε in the related technology described with reference to FIG. 1 is 0.0025. Therefore, according to equation (5), ΔE _c ^Γ is only 32.5 meV at most. The direct energy gap E ₀ (0.7723 eV) at this time corresponds to a wavelength of 1.61 μm.

  However, with such a downshift, the C band (1.53 μm−1.57 μm) can be covered, but the L band (1.57 μm−1.63 μm) and the U band (1.63 μm−1.68 μm). ) Cannot be covered.

Therefore, in each of the following embodiments, a Ge structure or the like in which the reduction width of the direct energy gap E ₀ is increased is provided.

(Embodiment 1)
The present embodiment relates to a structure in which Ge particles are embedded with a buried layer made of Ge _1-x Sn _x .

(1) Configuration FIG. 4 is a cross-sectional view illustrating a configuration of a structure according to the present embodiment. On the lower side of FIG. 4, the crystal axis of Ge is shown.

  As shown in FIG. 4, structure 12 according to the present embodiment includes germanium particles 14.

  In addition, the structure 12 includes a buried layer 16 that covers the periphery of the germanium particles 14 and embeds the germanium particles 14.

  In the structure 12, the buried layer 16 causes the germanium particles 14 to generate tensile strain in each of the three crystal axis directions of the germanium particles 14.

Here, the buried layer 14 is a semiconductor single crystal having a lattice constant larger than that of Ge, for example, Ge _1-x Sn _x (where 0 <x <1), or Si _y Ge _1-xy Sn _x (where 0 <x, y <1).

  FIG. 5 is a view for explaining an atomic arrangement at the interface 26 between the Ge particles 14 and the buried layer (semiconductor single crystal) 16 according to the present embodiment.

  As shown in FIG. 5, in this structure 12, the atoms 24 forming the Ge particles 14 and the buried layer 16 are bonded at the interface 26 so that the respective atomic arrangements are continuous. Further, the periphery of the Ge particles 14 is covered with a buried layer 16. Accordingly, the Ge particles 14 are not biaxial tensile strains, but tensile strains are generated in the three crystal axis directions.

(2) Principle Here, when the volume of the buried layer 16 is sufficiently larger than the volume of the Ge particles 14, the semiconductor single crystal forming the buried layer 16 is hardly distorted, and the Ge particles 14 have three crystals. Each axis is distorted equally.

This strain is ε (= ε _xx = ε _yy = ε _zz ), and this strain ε is substituted into the equation (1). Then, as shown below, a calculation formula representing the change ΔE _c ^Γ of the direct energy gap E ₀ is obtained.

As is apparent from a comparison between this equation and equation (5), the change in the direct energy gap E _{0 in} the structure 12 according to the present embodiment is the change in E ₀ of the Ge thin film with respect to the same tensile strain ε. It becomes about 2.19 times.

Therefore, according to this structure 12, a large E ₀ change can be realized with a slight strain. For example, the U band can be covered with a tensile strain of only 0.0019. This value is smaller than the upper limit 0.0025 of the strain that can be formed in the Ge thin film.

This is because, in the Ge thin film, ε _zz becomes a negative value (see Equation (3)), and the E ₀ downshift induced by ε _xx and ε _yy is reduced. This is because, in the Ge particles of the form, ε _zz also has a positive value ε (= ε _xx = ε _yy = ε _zz ) and contributes to the downshift of E ₀ (see formula (1)).

According to a highly accurate calculation result using the deformation potentiometer, in order to cover the U band, the sum of the strains in each crystal axis direction (ε _xx + ε _yy + ε _zz ) only needs to be 0.0054 or more. I understand that. However, it is not preferable that the absolute value of any one of ε _xx , ε _yy , and ε _zz is larger than 0.02 because crystal defects are generated in the Ge particles. (
That is, in structure 12 according to the present embodiment, it is preferable that the sum of strains in each crystal axis direction is 0.0054 or more and the absolute value of strains in each crystal axis direction is 0.02 or less.

As described above, according to the present embodiment, in the structure including Ge, the reduction width of the direct energy gap E ₀ of Ge due to tensile strain can be easily increased.

(3) Manufacturing Method Next, details of structure 12 according to the present embodiment will be described in accordance with a manufacturing procedure.

  FIG. 6 is a cross-sectional view illustrating an example of structure 12 according to the present embodiment. FIG. 7 is a flowchart illustrating the manufacturing procedure of the structure according to the present embodiment.

(I) Formation of gradient composition layer (step S1)
First, the Ge (001) substrate 28 is mounted on a gas source molecular beam growth apparatus and heated to, for example, 300 to 700 ° C.

Next, germane (GeH ₄ ) and tin deuteride (SnD ₄ ) are supplied to the heated Ge (001) substrate 28. Here, the supply ratio of SnD _{4 to} GeH ₄ is gradually increased.

  By this step, the graded composition GeSn layer 30 in which the Sn composition ratio gradually increases is formed. The graded composition GeSn layer 30 is formed to have a thickness of 1 μm, for example. Note that the notation of only the element symbol such as “GeSn” does not specify the composition of the mixed crystal semiconductor (that is, GeSn does not mean germanium-tin of Ge: Sn = 1: 1). ).

In addition, highly deuterated tin deuteride (SnD ₄ ) is easily obtained. Therefore, in the present embodiment, deuterated tin (SnD ₄ ) is used as the Sn source gas.

(Ii) Formation of strain relaxation layer (step S2)
Next, the supply ratio of SnD _{4 to} GeH ₄ is fixed, and a Ge _1-x Sn _x layer having a constant composition is grown, for example, to 500 nm. Here, x is the composition of Sn at the top of the gradient composition GeSn layer 30 (where 0 <x <1). A specific value of x will be described later.

A semiconductor layer in which lattice mismatch with the Ge (001) substrate 28 is eliminated is formed by the graded composition GeSn layer and the Ge _1-x Sn _x layer formed by step S1 and this step.

That is, the uppermost lattice constant of the Ge _1-x Sn _x layer 32 is the original lattice constant of Ge _1-x Sn _x .

The Ge _1-x Sn _x layer 32 is hereinafter referred to as a strain relaxation GeSn layer.

(Iii) Formation of Ge particles (Step S3)
Next, the supply of SnD ₄ is stopped, only GeH ₄ is supplied, and Ge corresponding to 3 to 10 ML (atomic layer) is supplied.

  At this time, Ge grows by an SK (Stranski-Krastanov) mode, and Ge fine particles 34 (Ge particles) are formed. The dimensions of the Ge fine particles 34 are 10 nm or more and 100 nm or less in all directions. That is, Ge quantum dots grow.

(Iv) Formation of coating layer (step S4)
Next, the supply of SnD ₄ is restarted, and the coating layer 36 made of Ge _1-x Sn _x is formed.

  The thickness of the coating layer 36 is not less than 10 and not more than 100 nm. Here, the coating layer 36 covers the periphery of the Ge fine particles 34 and embeds the Ge fine particles 34. The composition ratio x of the coating layer 36 is the same as that of the strain relaxation GeSn layer 32 described above. The coating layer 36 functions as a spacer layer for Ge fine particles 34, that is, Ge quantum dots.

  Here, the Ge fine particles 34 are covered with a strain relaxation GeSn layer 32 and a GeSn coating layer 36. That is, the strain relief GeSn layer 32 and the GeSn coating layer 36 cover the periphery of the Ge fine particles 34 and form the buried layer 16 in which the Ge fine particles 34 are embedded.

(V) Multi-layering (step S5)
Next, Step S3 and Step S4 are repeated a plurality of times to form the structure 12 shown in FIG. However, this step may not be performed. In that case, Ge fine particles are formed and only one Ge fine particle layer is formed.

  However, the external quantum efficiency can be increased when the structure 12 is applied to an optical semiconductor device by multilayering the Ge fine particle layer.

  In the above example, the Ge fine particles 34 are formed by a self-forming method using the SK mode. However, the Ge fine particles 36 may be formed by processing the Ge layer by lithography and etching.

(4) Operation In the structure 12 formed as described above, the constituent atoms of the Ge _1-x Sn _x (0 <x <1) buried layer 16 covering the Ge fine particles 34 and the periphery thereof are formed. , Each atomic arrangement is bonded at the interface so that they are continuous.

Incidentally, the lattice constant of Ge is less than _{Ge 1-x} Sn x. Therefore, the buried layer 16 generates tensile strains in the Ge fine particles 34 in the three crystal axes ([100], [010], [001]) in the Ge fine particles 34 (on the lower side of FIG. 6). , The crystal axes of the Ge microparticles and their orientations are shown.)

  Therefore, the direct energy gap of the Ge microparticles 34 is downshifted according to the above-described equation (6).

By the way, in the structure 12 according to the present embodiment, the Ge fine particles 34 are overwhelmingly smaller in volume than the buried layer 16. Therefore, the lattice constant of the Ge fine particles 34 matches the lattice constant of Ge _1-x Sn _x forming the buried layer 16. Therefore, the strain ε (= ε _xx = ε _yy = ε _zz ) generated in the Ge fine particles 34 is expressed as follows.

Here, a _GeSn is a lattice constant of Ge _1-x Sn _x forming the buried layer 16. On the other _{hand, a Ge} is such have distortion occurs Ge natural lattice constant (5.679nm).

On the other hand, the lattice constant a _GeSn of Ge _1-x Sn _x is expressed as follows.

  Here, x is the composition ratio of Sn.

Therefore, according to the equations (6) to (8), the composition x of Ge _1-x Sn _x that forms the desired downshift ΔE _c ^Γ is obtained. For example, in order to cover the U band, x may be 0.013 or more (however, an increase in energy level in the quantum dot is not taken into consideration).

However, if the strain is too large, crystal defects occur in the Ge fine particles. Therefore, it is desirable that the absolute value of the strain (ε _xx , ε _yy , ε _zz ) is 0.02 or less.

By the way, the embedded layer 16 may be formed of Si _y Ge _1-xy Sn _x (where 0 <x, y <1). Here, the lattice constant a _SiGeSn of Si _y Ge _1-xy Sn _x is expressed as follows.

According to this equation, and equations (6) and (7), the composition x, y of Si _y Ge _1-xy Sn _x forming the desired downshift ΔE _c ^Γ is obtained (provided that the equation ( In 7), a _GeSn is replaced with a _SiGeSn .)

  For example, in order to cover the U band, 0.013 <x−0.306y may be satisfied. However, in order to prevent crystal defects from occurring in the Ge fine particles, the absolute value of the strain generated in the Ge fine particles is preferably 0.02 or less.

  When the buried layer 16 is separated from each other, if the Ge particles are displaced from the center of the buried layer, the strain generated in the Ge particles changes depending on the position. Further, by inserting a material having a different lattice constant into a part of the buried layer 16, the strain of the Ge particles changes depending on the position.

(Embodiment 2)
The present embodiment relates to an optical semiconductor device (optical receiver) having the structure according to the first embodiment as a light receiving layer.

(1) Configuration and Manufacturing Method Hereinafter, the configuration of the optical semiconductor device (optical receiver) according to the present embodiment will be described according to a manufacturing procedure.

  FIG. 8 is a cross-sectional view illustrating the configuration of the optical semiconductor device (optical receiver 37) according to the present embodiment. FIG. 9 is a flowchart illustrating the manufacturing procedure of the optical semiconductor device (optical receiver 37) according to the present embodiment.

(I) Formation of SGOI substrate (step S1)
First, an SOI substrate 44 (silicon on insulator substrate; not shown) in which a SiO ₂ layer 40 and a Si single crystal layer (not shown) are sequentially formed on a Si substrate 38 is prepared. The SOI substrate 44 can be formed, for example, by bonding a Si single crystal layer to a Si substrate on which a SiO ₂ film is formed.

  Next, a SiGe layer is formed on the SOI substrate by epitaxial growth.

Next, the SiGe film is oxidized. At this time, Si is selectively oxidized and a SiO ₂ film is formed on the surface of the SiGe layer. On the other hand, Ge diffuses into the SiGe layer to increase the Ge composition of the SiGe layer.

Next, the SiO ₂ film is removed by etching.

  This oxidation and etching process is repeated a plurality of times to form a SiGe layer 42 having a thickness of 10 to 300 nm and a Ge composition ratio of 0.5 to 1.0.

  In order to form a SiGe layer having such a high Ge composition by epitaxial growth, the SiGe layer must be grown thick. However, according to the above process, a SiGe layer having a high Ge composition can be formed only by growing a thin GeSi layer.

  By this step, an SGOI substrate 44 (silicon-germanium on insulator substrate) is formed.

(Ii) Formation of strain relaxation GeSn layer (step S2)
Next, a strain relaxation layer 46 made of GeSn is formed on the SGOI substrate 44 in accordance with step S2 of the first embodiment.

However, the strain relaxation layer 46 made of GeSn is formed to be 3 to 10 times thicker than the thickness of the SiGe layer 42. At this time, almost no strain is generated in the strain relaxation layer 46 formed thicker than the SiGe layer 42, and the SiGe layer 42 is distorted. Alternatively, the SiGe slips at the SiGe / SiO ₂ interface and the strain is released.

The GeSn strain relaxation layer 46 is doped with a p-type impurity such as boron (B). The concentration of the p-type impurity is, for example, 1 × 10 ¹⁹ cm ⁻³ .

(Iii) Formation of light receiving layer (step S3)
Next, an undoped GeSn layer 48 is formed.

  Next, Ge fine particles 34 and a clothing layer 36 made of GeSn are formed according to substantially the same procedure as Step S3 and Step S4 of the first embodiment.

  Here, the buried layer 16 is formed by the undoped GeSn layer 48 and the coating layer 36.

  By this step, the light receiving layer 49 is formed.

(Iv) Formation of n-type GeSn layer (step S4)
Next, the n-type GeSn layer 50 is grown, for example, by 100 to 300 nm. At this time, an n-type impurity such as phosphorus (P) is doped. The impurity concentration is, for example, 1 × 10 ¹⁹ cm ⁻³ .

(V) Formation of electrode (step S5)
Next, the n-type GeSn layer 50 to the undoped GeSn layer 48 are partially removed by lithography and etching, and the surface of the strain relaxation layer 46 made of p-type GeSn is exposed.

  A p-electrode 58 and an n-electrode 60 are formed on the exposed surface and the surface of the n-type GeSn layer 50 to complete the optical receiver 37.

(2) Operation The Sn composition ratio of each GeSn layer forming the light receiver 37 is set so that the direct gap of the Ge fine particles 34 covers the U band in consideration of the quantum effect.

  Therefore, for example, when incident light 39 is incident from the back surface of the Si substrate 38, electrons and holes (photocarriers) are generated in the Ge fine particles 34 in response to a wide wavelength range including the U band.

  These electrons and holes are accelerated toward the p-electrode 46 and the n-electrode 60 by the electric field applied between the p-electrode 46 and the n-electrode 60, respectively, and become photocurrents.

  This photocurrent is measured by a measuring instrument connected to the light receiver 37, and the incident light 39 is detected.

  Therefore, according to the light receiver 37, it is possible to detect a wide range of light including the U band. If an integrated circuit device is formed on the Si substrate 38 on which the photoreceiver 37 is formed, the photoreceiver 37 receives and photoelectrically converts an optical signal generated by the integrated circuit device formed on another Si substrate. Signals can be passed to integrated circuit devices on the same substrate.

  As is clear from the above description, according to the present embodiment, it is possible to increase the wavelength by the tensile strain of the detection limit wavelength in the light receiver in which the absorption layer is formed of Ge.

(Embodiment 3)
The present embodiment relates to an optical semiconductor device (light emitting diode) having the structure according to the first embodiment as a light emitting layer.

(1) Configuration and Manufacturing Method Hereinafter, the configuration of the optical semiconductor device (light emitting diode) according to the present embodiment will be described in accordance with a manufacturing procedure.

  FIG. 10 is a flowchart illustrating the manufacturing procedure of the optical waveguide layer according to the present embodiment. FIG. 11 is a cross-sectional view illustrating the configuration of the optical semiconductor device (light emitting diode 52) according to the present embodiment.

(I) Formation of SGOI substrate (step S1)
First, an SGOI substrate is formed by the same procedure as in step S1 of the second embodiment.

(Ii) Formation of strain relaxation layer (step S2)
Next, the GeSn strain relaxation layer 46 is formed by substantially the same procedure as in step S2 of the second embodiment.

However, the concentration of the p-type impurity is, for example, 1 × 10 ¹⁸ cm ⁻³ . The composition ratio of the strain relaxation layer 46 is set to be the same as the composition of the undoped GeSn layer 48 and the coating layer 36 described later.

(Iii) Formation of lower SCH layer (step S3)
Next, an undoped SiGeSn layer 54 having a thickness of 10-100 nm is formed.

  Here, the composition ratios y and x of Si and Sn are, for example, 0.03 and 0.023. The refractive index of the SiGeSn layer 54 is smaller than that of Ge.

On the other hand, the direct energy gap E ₀ (at the Γ point) of the SiGeSn layer 54 is larger than the direct energy gap E _{0 of} Ge. Therefore, the SiGeSn layer 54 functions as a lower SCH (separate confinement heterostructure) layer.

(Iv) Formation of light emitting layer (step S4)
Next, the undoped GeSn layer 48, the Ge fine particles 34, and the coating layer 36 are formed by substantially the same procedure as in step S3 of the second embodiment. However, the composition of the undoped GeSn layer 48 and the coating layer 36 is set so that the magnitude of the direct energy gap at the Γ point and L point of the Ge fine particles 34 is reversed, as will be described later.

  By this step, the light emitting layer 56 is formed.

(V) Formation of upper SCH layer (step S5)
Next, an undoped SiGeSn layer 58 is formed. The thickness and composition of the SiGeSn layer 58 are the same as those of the undoped SiGeSn layer 54.

  Similar to the SiGeSn layer 54, the SiGeSn layer 58 also functions as an SCH layer.

(Vi) Formation of n-type GeSn layer (step S6)
Next, the n-type GeSn layer 50 is formed by substantially the same procedure as in step S4 of the second embodiment.

However, the concentration of the n-type impurity is, for example, 1 × 10 ¹⁸ cm ⁻³ . The composition of the GeSn layer 50 is the same as that of the coating layer 36 and the undoped GeSn layer 48.

(Vii) Formation of electrodes (step S7)
Next, the n-type GeSn layer 50 to the undoped GeSn layer 48 are partially removed by lithography and etching, and the surface of the strain relaxation layer 46 made of p-type GeSn is exposed.

  A p-electrode 58 and an n-electrode 60 are formed on the exposed surface and the surface of the n-type GeSn layer 50, thereby completing the light-emitting diode 52.

  The n electrode 60 is formed in a ring shape.

(2) Operation The Sn composition ratio of each GeSn layer forming the light emitting diode 52 is set so that the direct energy gap at the Γ point of the Ge fine particles 34 is smaller than the direct energy gap at the L point. The Accordingly, the bottom of the conduction band and the top of the valence band of the Ge fine particles 34 are both Γ points.

  Therefore, when a signal source is connected between the electrode 60 and the n-electrode 58 and current is injected into the light emitting layer 56, electrons and holes are concentrated on the bottom of the conduction band and the top of the valence band of the Ge fine particles 34, respectively. An inversion distribution is formed. After forming the population inversion, electrons and holes are combined and disappear. At this time, the electrons directly transition from the bottom of the conduction band to the top of the valence band to generate light.

  This light is extracted through the inside of the ring-shaped n-electrode 60 and becomes emitted light 62.

  By the way, since the present light emitting diode 52 includes the SCH layers 54 and 58, it can be used after being processed into a waveguide type light emitting diode.

(3) Principle Next, the reason why Ge particles become a direct transition semiconductor by introducing strain will be described.

  As described above, the top 8 of the valence band of Ge is at the Γ point (see FIG. 3). On the other hand, the bottom 9 of the conduction band exists at the L point. Here, the energy difference between the Γ point 11 and the L point in the conduction band is 0.15 eV.

By the way, the direct energy gap E _{0 at} the Γ point changes with strain as described above. The amount of change ΔE _c ^Γ can be approximated by the following equation (see equations (1) and (6)).

On the other hand, the direct energy gap at point L also changes due to strain. The change amount ΔE _c ^L is approximated by the following equation.

Here, ε = ε _xx = ε _yy = ε _zz . Also, a _c ^Γ = −9.48 (eV) and a _c ^L = −2.78 (eV).

  As is apparent from a comparison between Equation (10) and Equation (11), the change amount of the direct energy gap with respect to strain is much larger at the Γ point.

  Therefore, as the strain ε is increased, the energy difference of 0.15 eV between the L point and the Γ point (in the conduction band) is eventually eliminated, and finally the Γ point becomes the bottom of the conduction band. Therefore, by introducing a predetermined strain, the Ge particles can be made a direct transition type semiconductor.

  It can be seen that the strain ε required to make the Ge particles a direct transition type semiconductor is an equation (10), an equation (11), and an energy difference between the Γ point and the L point from 0.15 eV to 0.008.

  On the other hand, the strain ε necessary for making the Ge thin film described with reference to FIG. 1 into a direct transition type semiconductor is expressed by the equations (5), (3) and (11), and the Γ and L points. It can be seen that the energy difference is 0.15 eV to 0.018.

  That is, according to the present embodiment, the direct transition type of Ge is facilitated. Therefore, according to the present embodiment, it is possible to form a light emitting device using Ge as a light emitting layer. Further, if an integrated circuit device is formed on the Si substrate 38 on which the light emitting diode 52 is formed, an electrical signal generated by the integrated circuit device is photoelectrically converted by the light emitting diode 52 and formed on another Si substrate. Optical communication can be performed with the integrated circuit device.

(Embodiment 4)
This embodiment relates to a structure embedded with a buried layer of metal or oxide.

(1) Configuration FIG. 12 is a cross-sectional view illustrating a configuration of structure 64 according to the present embodiment. As shown in FIG. 12, structure 64 according to the present embodiment includes germanium fine particles 34 (germanium particles).

  In addition, the structure 64 includes the embedded layer 16 that covers the periphery of the germanium fine particles 34 (germanium particles) and embeds the germanium fine particles 34 (germanium particles).

  In the structure 64, the embedded layer 16 generates tensile strains in the germanium fine particles 34 (germanium particles) in the three crystal axis directions of the germanium fine particles 34 (germanium particles). Note that the orientation of the crystal axis of the germanium fine particles 34 is shown on the lower side of FIG.

  Further, in this structure 64, the buried layer 16 is in contact with the bottom surface of the germanium fine particles 34 (germanium particles) and is in close contact with the semiconductor single crystal 66 having a larger lattice constant than germanium and the top and side surfaces of the germanium fine particles 34 (germanium particles). The metal film 68 (or oxide film) having a thermal expansion coefficient smaller than that of germanium is used.

  In the structure 64, the atoms forming the germanium fine particles 34 (germanium particles) and the semiconductor single crystal 66 are bonded at the bottom surface of the germanium fine particles 34 (germanium particles) so that the respective atomic arrangements are continuous. .

  The metal film 68 (or oxide film) is formed so as to be in close contact with the germanium fine particles 34 (germanium particles) at a temperature higher than the use temperature of the structure 64.

(2) Principle In the structure 64, as in the structure 12 according to the first embodiment, the embedded layer 16 causes the germanium particles 34 to generate tensile strains in the three crystal axis directions of the germanium particles 34. . Therefore, the energy gap E ₀ is directly reduced as in the structure 12 according to the first embodiment.

  However, tensile strain is generated in the Ge fine particles 34 of the structure 64 due to a difference in lattice constant from the semiconductor single crystal 66 in contact with the bottom surface in two crystal axis directions parallel to the bottom surface. On the other hand, tensile strain is generated by the metal film 68 (or oxide film) in the direction perpendicular to the bottom surface.

  Here, the tensile strain formation process due to the difference in lattice constant from the semiconductor single crystal 66 is the same as the tensile strain formation process described in the first embodiment. Therefore, a description of the tensile strain formation process due to the difference in lattice constant from the semiconductor single crystal 66 is omitted.

  Next, the tensile strain generated in the Ge fine particles 34 by the metal film 68 (or oxide film) will be described.

  As described above, the metal film 68 (or oxide film) is formed so as to be in close contact with the germanium fine particles 34 at a temperature higher than the use temperature of the structure 64.

In general, the crystal structure of a metal (or oxide) is different from that of a semiconductor or a special group IV semiconductor. Therefore, when the metal film 68 (or oxide film) is formed on the Ge fine particles 34, both atoms are bonded only when the distance is sufficiently close. At this stage, no strain (strain ε _{zz in the} direction perpendicular to the bottom surface) is generated in the Ge fine particles 34.

  However, when the metal film 68 (or oxide film) is formed at a high temperature and then the temperature of the structure 64 is lowered, the Ge particles 34 are distorted due to the difference in thermal expansion coefficient between the metal film 68 (or oxide film) and the Ge particles 34. Will occur.

  Here, it is assumed that the thermal expansion coefficient of the metal film 68 (or oxide film) forming the buried layer 16 is smaller than that of germanium. In this case, when the temperature of the structure 64 decreases, the Ge fine particles 34 tend to shrink more than the metal film 68 (or oxide film). For this reason, tensile strain is also generated in the Ge fine particles 34 in the direction perpendicular to the bottom surface.

Here, the thermal expansion coefficient of _Ge is K _Ge and the thermal expansion coefficient of the metal film 68 (or oxide film) is K _Matrix . Further, the lattice constant of Ge at the use temperature (for example, room temperature) of the structure 64 is a _Ge , and the difference between the use temperature of the structure 64 and the temperature at the time of forming the metal film 68 (or oxide film) is ΔT. And

Then, the strain ε _zz generated in the Ge fine particles 34 is expressed by the following equation.

Here, _let us consider a case where the strain generated in the Ge fine particles 34 is constant regardless of the direction of the crystal axis (ε _xx = ε _yy = ε _zz ). In such a case, in order to cover the L band (1.57 μm−1.63 μm), the strain ε (= ε _xx = ε _yy = ε _zz ) must be 0.001 or more.

  Examples of the metal film satisfying this condition include a tungsten (W) film formed at a temperature higher than the use temperature by 600K or more and a molybdenum (Mo) film formed at a temperature higher than the use temperature by 1040K or more.

  Even in the structure according to the first embodiment, the buried layer is formed at a temperature higher than the use temperature. Therefore, the strain seems to be affected by the difference in thermal expansion coefficient between the buried layer and Ge. However, in the structure according to the first embodiment, since the buried layer is also formed of a semiconductor, there is little difference in the thermal expansion coefficient between Ge and the buried layer. For this reason, the difference between the thermal expansion coefficients of the buried layer and Ge has little effect on the strain generated in the Ge particles.

  In the first embodiment, at the time of forming the buried layer, the constituent atoms of the buried layer and the Ge particles are bonded at the interface so as not to disturb each atomic arrangement. And the structure in which Ge particle was formed is cooled and used, maintaining this state. That is, the semiconductors forming the buried layer with the Ge particles are regularly bonded so that the constituent atoms do not disturb each other's atomic arrangement regardless of the temperature.

  Therefore, in the structure according to the first embodiment, attention should be paid only to the strain generated due to the difference in lattice constant between Ge at the operating temperature and the semiconductor forming the buried layer.

(3) Manufacturing Method Next, the structure of structure 64 according to the present embodiment will be described according to a manufacturing procedure.

  FIG. 13 is a flowchart illustrating the manufacturing procedure of the structure according to the present embodiment.

(I) Formation of gradient composition layer, strain relaxation layer, and Ge particles (steps S1 to S3)
According to steps S1 to S3 described in the first embodiment, a GeSn graded composition GeSn layer 30, a GeSn strain-relaxed GeSn layer 32 (semiconductor single crystal 66), and Ge are formed on the Ge (001) substrate 28. Fine particles 34 are formed.

  At this time, tensile strain in the biaxial direction parallel to the bottom surface of the Ge fine particles 34 is generated.

(Ii) Formation of metal film (step S4)
Next, the structure 64 being manufactured is heated to 627 ° C.

  Next, a W film (metal film 68) having a thickness of 100 to 300 nm is formed by, for example, vacuum deposition so as to be in close contact with the top and side surfaces of the Ge fine particles 34.

  Thereafter, the temperature of the structure 64 is lowered to room temperature (27 ° C.). At this time, tensile strain is also generated in the Ge fine particles 34 in the direction perpendicular to the bottom surface. As a result, tensile strain is generated in the Ge fine particles 34 in all three crystal axes.

  The structure 64 is completed by the above procedure.

As is clear from the above description, according to the present embodiment, the reduction width of the direct energy gap E ₀ of Ge due to tensile strain in the structure including Ge has been described with reference to FIG. The Ge thin film can be made wider.

In the above example, the film that adheres to the top and side surfaces of the Ge particles is a metal, but an oxide film such as SiO ₂ may be used.

(Embodiment 5)
The present embodiment relates to an optical semiconductor device (optical receiver) having a structure according to the fourth embodiment as a light receiving layer.

(1) Configuration and Manufacturing Method Hereinafter, the configuration of the optical semiconductor device (optical receiver) according to the present embodiment will be described according to a manufacturing procedure.

  FIG. 14 is a cross-sectional view illustrating the configuration of the optical semiconductor device (optical receiver 70) according to the present embodiment. FIG. 15 is a flowchart illustrating the manufacturing procedure of the optical semiconductor device (optical receiver 70) according to the present embodiment.

(I) SGOI substrate formation and strain relaxation layer formation (steps S1 to S2)
First, the SGOI substrate 44 and the GeSn strain relaxation layer 46 are formed by substantially the same procedure as in steps S1 and S2 described in the second embodiment.

(Ii) Formation of light receiving layer (step S3)
Next, an undoped GeSn layer 48 is formed.

  Next, Ge fine particles 34 are formed according to substantially the same procedure as step S3 of the fourth embodiment. Here, the width of the bottom of the Ge fine particles 34 is 10 to 100 nm, and the height of the Ge fine particles 34 is also 10 to 100 nm.

  Next, a coating layer 72 made of n-type GeSn covering the Ge fine particles 34 is formed. However, the coating layer 72 may be omitted.

(Iii) Formation of metal film (step S4)
Next, a metal film 68 made of W is formed by the same procedure as step S4 of the fourth embodiment.

(Iv) Formation of electrodes (step S5)
Next, the n-type GeSn coating layer 72 to the undoped GeSn layer 48 are partially removed by lithography and etching to expose the surface of the p-type GeSn strain relaxation layer 46.

  A p-electrode 58 is formed on the exposed surface, and the optical receiver 70 is completed.

  In the light receiver 70, the metal film 68 serves as an n electrode.

(3) Operation The light receiver 70 operates according to the light receiver 37 of the second embodiment.

  Therefore, the light receiver 70 can detect light in a wide wavelength range including the L band. If an integrated circuit is formed on the Si substrate 38 on which the light receiver 70 is formed, the light receiver 70 receives an optical signal generated in the integrated circuit formed on another Si substrate and photoelectrically converts the signal. It can be transferred to an integrated circuit on the same substrate.

  As is clear from the above description, according to the present embodiment, the detection wavelength limit due to tensile strain can be lengthened in the light receiver in which the light absorption layer is formed of Ge.

  The above embodiment is summarized as follows.

(Appendix 1)
Germanium particles,
Covering the periphery of the germanium particles and comprising an embedded layer in which the germanium particles are embedded;
The embedded layer generates tensile strain in the germanium particles in each of the three crystal axis directions of the germanium particles.
Structure.

(Appendix 2)
In the structure according to appendix 1,
The germanium particles are quantum dots,
Characteristic structure.

(Appendix 3)
In the structure according to appendix 1 or 2,
The sum of strains in each crystal axis direction is 0.0054 or more,
And that the absolute value of strain in each crystal axis direction is 0.02 or less,
Characteristic structure.

(Appendix 4)
In the structure according to any one of appendices 1 to 3,
The buried layer is formed of a semiconductor single crystal having a larger lattice constant than germanium;
The atoms forming the germanium particles and the semiconductor single crystal are bonded at the interface between the germanium particles and the semiconductor single crystal so that the respective atomic arrangements are continuous.
Characteristic structure.

(Appendix 5)
In the structure according to appendix 4,
The semiconductor single crystal is
Ge _1-x Sn _x (where 0 <x <1)
That is formed by
Characteristic structure.

(Appendix 6)
In the structure according to appendix 4,
The semiconductor single crystal is
Si _y Ge _1-xy Sn _x (where 0 <x, y <1)
That is formed by
Characteristic structure.

(Appendix 7)
In the structure according to any one of appendices 1 to 3,
The buried layer comprises:
A semiconductor single crystal in contact with the bottom surface of the germanium particles and having a larger lattice constant than germanium;
Adhering to the top and side surfaces of the germanium particles, the thermal expansion coefficient is formed by a metal film or oxide film smaller than germanium,
The atoms forming the germanium particles and the semiconductor single crystal are bonded at the bottom so that the respective atomic arrangements are continuous,
The metal film or the oxide film is formed so as to adhere to the germanium particles at a temperature higher than the use temperature of the structure.
Characteristic structure.

(Appendix 8)
The structure according to any one of appendices 1 to 7,
A light-receiving layer that generates light by receiving light, or a light-emitting layer that generates light by injecting current,
Optical semiconductor device.

It is a cross-sectional view illustrating a structure of a Ge thin film by reducing the energy gap E ₀ directly by introducing tensile strain. The interface between the Ge film and _{Ge 1-x} Sn _x buffer layer is a diagram illustrating an in atomic arrangement. It is a figure explaining the band structure in the wave number space of Ge. FIG. 3 is a cross-sectional view illustrating a configuration of a structure according to the first embodiment. It is a figure explaining the atomic arrangement in the interface of Ge particle and a buried layer (semiconductor single crystal). 5 is a cross-sectional view illustrating an example of a structure body according to Embodiment 1. FIG. FIG. 6 is a flowchart illustrating a manufacturing procedure for a structure according to the first embodiment. It is sectional drawing explaining the structure of the optical semiconductor device (light receiver) according to Embodiment 2. FIG. 10 is a flowchart illustrating a manufacturing procedure of an optical semiconductor device (light receiver) according to the second embodiment. It is sectional drawing explaining the structure of the optical semiconductor device (light emitting diode) according to Embodiment 3. FIG. 11 is a flowchart illustrating a manufacturing procedure of an optical semiconductor device (light emitting diode) according to the third embodiment. It is sectional drawing explaining the structure of the structure according to Embodiment 4. It is a flowchart explaining the manufacturing procedure of the structure according to Embodiment 4. It is sectional drawing explaining the structure of the optical semiconductor device (light receiver) according to Embodiment 5. FIG. 10 is a flowchart illustrating a manufacturing procedure of an optical semiconductor device (light receiver) according to the fifth embodiment.

Explanation of symbols

2 ... Ge thin film 4 ... Si (100) substrate 6 ... Ge _1-x Sn _x buffer layer 8 ... Top of valence band 9 ... Bottom of conduction band 10 ... (Ge Interface 11 between thin film and Ge _1-x Sn _x buffer layer ... Bottom of conduction band 12 at Γ point 12 Structure (Embodiment 1) 14 ... Germanium particles 16 ... Buried layer 18, 20... Ge atom 22... Sn atom 24... Atom 26... (Interface between Ge particles and buried layer) 28... Ge (001) substrate 30. Ge _1-x Sn _x layer (strain relaxed GeSn layer)
34 ... Ge fine particles 36 ... Coating layer 37 ... Optical receiver (second embodiment) 38 ... Si substrate 39 ... incident light 40 ... SiO ₂ layer 42 ... SiGe layer 44 ... SGOI substrate
46 ... strain relaxation layer 48 ... undoped GeSn layer 49 ... light receiving layer 50 ... n-type GeSn layer 52 ... light emitting diode 54, 58 ... undoped SiGeSn layer 56 ... light emitting layer 58 ... P electrode 60 ... n electrode 62 ... emitted light 64 ... structure (Embodiment 4) 66 ... semiconductor single crystal 68 ... metal film 70 ... receiver (implementation) Form 5)
72 ... Cover layer (made of n-type GeSn)

Claims (4)

Germanium particles,
Covering the periphery of the germanium particles and comprising an embedded layer in which the germanium particles are embedded;
The embedded layer generates tensile strain in the germanium particles in each of the three crystal axis directions of the germanium particles.
Structure. The structure of claim 1,
The sum of strains in each crystal axis direction is 0.0054 or more,
And that the absolute value of strain in each crystal axis direction is 0.02 or less,
Characteristic structure. The structure according to claim 1 or 2,
The buried layer comprises:
A semiconductor single crystal in contact with the bottom surface of the germanium particles and having a larger lattice constant than germanium;
Adhering to the top and side surfaces of the germanium particles, the thermal expansion coefficient is formed by a metal film or oxide film smaller than germanium,
The atoms forming the germanium particles and the semiconductor single crystal are bonded at the bottom so that the respective atomic arrangements are continuous,
The metal film or the oxide film is formed so as to adhere to the germanium particles at a temperature higher than the use temperature of the structure.
Characteristic structure. The structure according to any one of claims 1 to 3,
A light-receiving layer that generates light by receiving light, or a light-emitting layer that generates light by injecting current,
Optical semiconductor device.

Images