JPH07128622A - Semiconductor quantum well structure - Google Patents

Abstract

(57) [Abstract] [Purpose] An object is to provide a semiconductor quantum well structure with low loss and low driving voltage. A semiconductor quantum well structure has a coupled quantum well composed of a first quantum well layer and a second quantum well layer separated by a barrier layer, a wave function in the coupled quantum well, and adjacent to the coupled quantum well. At least a first or a second quantum well in a semiconductor quantum well structure including a coupled quantum well outer barrier layer provided on both sides of the coupled quantum well so as to suppress coupling with a wave function in an adjacent coupled quantum well At least one intra-quantum well layer barrier layer for reducing the existence probability of holes in the first quantum well layer is provided in the layer.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a semiconductor quantum well structure having a low loss and a low driving voltage.

[0002]

2. Description of the Related Art Conventionally, by applying an electric field perpendicular to the surface of a well layer of a quantum well structure, the quantum levels in the conduction band and the valence band are changed, and the shape of the wave function is changed. It is known that the light absorption spectrum of the structure changes with the application of an electric field.

FIGS. 13 and 14 show a potential shape of one unit constituting a conventional multiple quantum well (MQW) consisting of one rectangular quantum well layer and two barrier layers (barrier layers). FIG. 13 shows the wave function of electrons, the wave function of holes, and the transition energy of excitons. FIG. 13 shows a case of no electric field (V = 0) and FIG. 14 shows a case of applying an electric field by voltage V (V ≠ 0). Indicate the case.

In these figures, reference numeral 1 is InA.
A coupled quantum well outer barrier layer (hereinafter, also referred to as a barrier layer) made of 1As and 2 are quantum well layers (hereinafter, also referred to as well layers) made of i-InGaAlAs. In these figures, the wave functions of the electron in the conduction band and the wave function of the hole in the valence band are ψ _el and ψ _hh , respectively, and the energy level of the electron and the hole (in this figure, the heavy hole Energy levels of E _el and E respectively
_hh, and the exciton transition energy in this case is E _ex
And

In the absence of an electric field (v = 0) (FIG. 13), the curves drawn by the wave functions ψ _el and ψ _hh associated with the respective levels of the electron in the conduction band and the hole in the valence band are It has symmetry about the center.

The barrier layer 1 has a role of hindering the coupling between the wave function in one well layer and the wave function in another well layer.
Here, when the thickness W of the well layer 2 is 9 nm and the thickness of the barrier layer 1 is 5 nm, the wavelength of the steep absorption peak due to excitons (hereinafter, also referred to as exciton absorption peak) in the optical absorption spectrum is 1. 44 μm, 1.5
It can be applied to a phase modulator and an optical switch for light in the 5 μm wavelength band.

When an electric field is applied to the well layer surface (V ≠
0), as shown in FIG. 14, the electron wave function ψ _el and the hole wave function ψ _hh gather in the direction of lower potential. At this time, the energy level E _{el of the} electron decreases and the energy level E _{hh of the} hole increases. As a result, the exciton transition energy E _ex when an electric field is applied is lower than that when no electric field is applied. Therefore, the exciton absorption peak wavelength is shifted to the long wavelength side by the electric field application. Figure 15 shows
The amount of exciton absorption peak shift when a negative voltage V is applied to the quantum well (W = 9 nm), and the vertical axis represents the amount of shift from the absorption peak wavelength when no electric field is applied to the absorption peak wavelength when an electric field is applied. The horizontal axis represents the strength of the internal electric field due to the voltage V. As shown in this figure, the exciton absorption peak wavelength shifts from 1.44 μm to the longer wavelength side. This phenomenon is called the quantum confined Stark effect (Quantum Confied
Stark Effect; QCSE).

FIG. 16 shows the relationship between the absorption coefficient α of the quantum well and the wavelength λ of the incident light when there is no electric field (V = 0) and when a negative voltage is applied (V ≠ 0). The holes include a heavy hole (h) that absorbs TE mode light and a light hole (l) that absorbs TM mode light. It is also called a hole. In the figure, H is the peak of the absorption coefficient of heavy holes (h), and L is the peak of the absorption coefficient of light holes. By applying a voltage, the entire spectrum including the absorption peak shifts to the long wavelength side.

FIG. 17 shows a change in absorption coefficient Δα ((a) of FIG. 17) and a change of refractive index Δn ((b) of FIG. 17) of the quantum well and incident light when an electric field is applied (V ≠ 0). It shows the relationship with the wavelength. What is important here is that this change in absorption coefficient Δα may take a negative value or a positive value depending on the wavelength of incident light when an electric field is applied. On the other hand, the refractive index change Δn when an electric field is applied has a relationship of Kramers-Kronig.
That is, the change Δn in the refractive index shown in FIG.

[0010]

[Formula 1] Δn = λ ² / (2π ² ) ∫ Δα / (λ ² −λ ′ ² ) dλ ′ (1) (where λ is the wavelength used and λ ′ is the integral variable).

[0011]

However, the conventional semiconductor quantum well structure has a drawback that the change in the refractive index when an electric field is applied is small. This is because, in the above formula (1), since Δα takes both positive and negative values with respect to the wavelength λ ′, the positive and negative Δα cancel each other out,
The obtained Δn has an extremely small value. For example, the experimentally obtained value is Δn = 10 ⁻³ to 10 ⁻⁴ . Therefore, since the change in refractive index is small, the wavelength used (1.55 μm
It is necessary to bring the exciton absorption peak wavelength closer to m) (about 1.44 μm). Furthermore, since the absorption spectrum shifts to the long wavelength side when an electric field is applied, when the above-mentioned conventional quantum well structure is applied to an optical phase modulator or optical switch, the absorption coefficient increases as the refractive index changes, resulting in the insertion of light. The problem of increased loss arises.

Further, in order to realize a desired exciton absorption peak wavelength, InGaAlAs, which is a quaternary mixed crystal formed by further adding Al to InGaAs, which is a ternary mixed crystal, has conventionally been used. There is.

That is, in order to obtain a large QCSE, it is necessary to make the quantum well layer as thick as about 9 nm. In this case, in the quaternary mixed crystal, when the Al content is about 8%, the exciton absorption peak wavelength is 1.44 μm. Three
If the quantum well is expanded by using InGaAs, which is the original mixed crystal, the exciton absorption peak wavelength is set to the long wavelength side of 1.55 μm or more, so that it is necessary to reduce the depth of the quantum well. Therefore, InGaAs which is a ternary mixed crystal
Instead of, a quaternary mixed crystal InGaAlAs quantum well in which Al is added to make the quantum well shallow has been used.

On the other hand, it is known that fluctuations in the absorption peak wavelength of excitons due to application of an electric field are proportional to the fourth power of the quantum well width (IEEE Journal of Qua).
ntum Electron, Physical R
ev. , Vol. B32, pp. 1043-106
0,1985). Therefore, simply shortening the thickness of the quantum well to shorten the wavelength reduces the QCSE, which is not a good measure from the viewpoint of reducing the driving voltage.

Therefore, when a thick quantum well is used, as shown in FIG. 18, the Al composition ratio of the quaternary mixed crystal is adjusted so that a desired exciton absorption peak wavelength can be obtained. At this time, as can be seen from the figure, since the Al composition (%) has a great influence on the setting of the absorption peak wavelength of the exciton, it is necessary to accurately adjust the Al content of the quantum well layer and its overall composition. However, such strict composition control is very difficult, and there is a problem in crystal yield.

Further, the quaternary mixed crystal is inferior in reproducibility of a high quality crystal to the ternary mixed crystal, and further, when it is attempted to realize the exciton absorption peak wavelength at 1.3 μm, InAlA
In the s type, the content of Al, which easily oxidizes and significantly deteriorates the crystal quality, is as large as several tens of percent. as a result,
The experimental fact that no clear exciton absorption peak is observed is also confirmed.

Further, when the wavelength used is 1.55 μm, in the InP system, in order to set the exciton absorption peak wavelength to 1.44 μm, quaternary mixed crystal InGaAsP is used not only in the quantum well layer but also in the barrier layer. . Therefore, there is a problem in yield and reproducibility of crystal growth. When the wavelength used is 1.3 μm, the potential of the quantum well must be made shallower and the band gap between the conduction band and the valence band must be widened, so that no clear exciton absorption peak can be observed. At the same time, there is a drawback that the absorption spectrum is blunted immediately when an electric field is applied.

The conventional quantum well structure described above has a structural unit composed of one quantum well layer and two barrier layers arranged on both sides of the quantum well layer, as shown in FIGS. 19 and 20. The conventional conventional asymmetric coupling quantum well also has problems as described below.

In the figure, reference numeral 11 is a first quantum well layer, 1
Reference numeral 2 is a second quantum well layer, and 13 is a barrier layer between the quantum well layers. The barrier layer 13 adjusts the coupling between the wave function in the first quantum well layer 11 and the wave function in the second quantum well layer 12 depending on its thickness B and its composition. In FIG. 19, the first quantum well layer 11 is different from the second quantum well layer 12 in thickness, and in FIG. 20, the first quantum well layer 11 is different from not only the thickness but also the depth in the second quantum well layer 12. Although these figures show asymmetrically coupled quantum wells having different configurations, in both cases, a large number of holes are present in the second quantum well layer 12 when no electric field is applied. Since it moves to the one quantum well layer 11, a large change in the refractive index can be expected. However, these structures also have the same problems as the above-mentioned conventional example.

That is, in the structure shown in FIG. 19, many electrons and holes are present in the second quantum well layer 12 when there is no electric field. Therefore, the exciton absorption peak wavelength depends on the thickness of the second quantum well layer 12. It is determined. Therefore, the thickness H ₂ of the second quantum well layer 12 must be set thin, and accordingly, the thickness H _{1 of} the first quantum well layer 11 must also be set thin. As a result, QCSE becomes smaller. on the other hand,
In the structure shown in FIG. 20, it is necessary to make the composition of the second quantum well layer 12 shorter than that of the first quantum well layer 11. However, as described above, it is very difficult to control the composition, and there are problems in the crystal yield and reproducibility.

Therefore, an object of the present invention is to solve the above-mentioned problems and realize a required quantum well thickness in order to obtain a large QCSE, that is, a large refractive index change, without using a composition control technique which is generally difficult. In addition, it is to provide a low-loss semiconductor quantum well structure having a sharp exciton absorption peak wavelength in a wide wavelength range.

[0022]

In order to solve the above-mentioned problems, a semiconductor quantum well structure according to the present invention is a coupled quantum well composed of a first quantum well layer and a second quantum well layer separated by a barrier layer. And a barrier layer outside the coupled quantum well provided on both sides of the coupled quantum well so as to suppress coupling between the wave function in the coupled quantum well and the wave function in the adjacent coupled quantum well adjacent to the coupled quantum well. In the semiconductor quantum well structure consisting of, at least one quantum well inner barrier layer for reducing the existence probability of holes in the first quantum well layer is provided in at least the first quantum well layer. To do.

Preferably, the barrier layer between the quantum well layers and the barrier layer inside the quantum well layer are made of a semiconductor material having the same band gap as that of the barrier layer outside the coupled quantum well.

Preferably, the first quantum well layer and the second quantum well layer are made of a semiconductor material having the same bandgap.

Preferably, a large number of electrons are present in at least one of the first quantum well layer and the second quantum well layer when no electric field is applied and when an electric field is applied, while holes are more than in the second quantum well layer when no electric field is applied. Many exist in the first quantum well layer, and holes move from the first quantum well layer to the second quantum well layer when an electric field is applied, or electrons are in the first quantum well layer both when there is no electric field and when an electric field is applied. Alternatively, a large number of holes are present in at least one of the second quantum well layers, while holes are more present in the second quantum well layer than in the first quantum well layer when no electric field is applied and from the second quantum well layer when an electric field is applied. Move to the first quantum well layer.

Preferably, by applying a voltage in a direction that cancels the built-in voltage in the semiconductor quantum well structure, electrons do not move and the first quantum well layer to the second quantum well layer or the second quantum well structure moves. The holes move to the first quantum well layer.

[0027]

According to the present invention, since only the wave function ψ _{hh of the} hole largely moves when an electric field is applied, the sign of the absorption coefficient change Δα becomes the same in almost the entire wavelength region, or the exciton absorption occurs when the electric field is applied. The peak wavelength becomes shorter. Therefore, an extremely large refractive index change Δn can be obtained as compared with the conventional quantum well structure, or the absorption of light when an electric field is applied is reduced. Furthermore, big Q
The barrier layer in the quantum well structure is controlled by controlling the thickness of the quantum well layer, which is necessary to realize the CSE, while controlling the crystal growth without changing the crystal growth. The absorption peak of excitons can be obtained with good reproducibility at any wavelength over a wide wavelength range, which has been difficult in the past, by setting the number and the thickness thereof. Furthermore, since the composition of the crystal can be simplified, it is possible to realize a sharp exciton absorption peak without impairing the crystallinity and reproducibility of the crystal.

[0028]

Embodiments of the present invention will now be described in detail with reference to the drawings. In addition, the same reference numerals indicate the same things in each drawing.

Example 1 FIG. 1 and FIG. 2 explain the structure of an example of a semiconductor quantum well structure according to the present invention, potential shape, electron wave function, hole wave function, and exciton transition energy. Therefore, FIG. 1 shows the case when there is no electric field (V = 0), and FIG. 2 shows the case when the electric field is applied (V ≠ 0).

The semiconductor quantum well structure shown in these figures is composed of a first quantum well layer 3, a second quantum well layer 4, and barrier layers 1, 7, and 8. The first quantum well layer 3 and the second quantum well layer 4 are coupled to form a coupled quantum well,
They are separated from each other by a first barrier layer (barrier layer between quantum well layers) 7, and the first quantum well layer 3 is further divided into two sub-quantum well layers by a second barrier layer (barrier layer in the quantum well layer) 8. It is divided into 5 and 6.

Further, in FIG. 1, wave functions of electrons and holes are shown as ψ _el and ψ _hh , respectively.

First quantum well layer 3 and second quantum well layer 4
The material of 3 is not quaternary mixed crystal, but InGaAlAs
The original mixed crystal is InGaAs. Therefore, crystal preparation is easily achieved, and its reproducibility is high.

The material of the barrier layers 1, 7 and 8 is InAlAs, which is easy to prepare crystals.

In this example, the dimensions of each layer were as follows.

Sub-quantum well layer 5 thickness W ₁ 3.9 n
m Sub-quantum well layer 6 thickness W ₂ 3.9 nm Second quantum well layer 4 thickness W ₃ 3.6 nm First barrier layer 7 thickness B ₁ 2.4 nm Second barrier layer 8 thickness B ₂ 1.2 nm Therefore, the first quantum well layer 3 (thickness: W ₁ + B ₂ + W ₂ ) is thicker than the second quantum well layer 4 (thickness: W ₃ ). As a result, many electrons and holes are present in the first quantum well layer 3 even though the intra-quantum well barrier layer 8 is present in the first quantum well layer 3. In this case, the exciton peak wavelength depending on the electron energy level E _el and the hole energy level E _hh is 1.39 μm.

When an electric field is applied (V ≠ 0), the electron wave function ψ _el is almost the same as when there is no electric field (V = 0), but the hole wave function ψ _hh is the thick first The quantum well layer 3 shifts to the side of the second quantum well layer 4 having a small thickness. Hereinafter, this will be specifically described.

FIG. 2 shows the wave functions ψ _el and ψ _hh when an electric field of 12 kV / cm is applied.

Since the effective mass of the electron is small, the shape of the wave function ψ _el of the electron hardly changes even if the potential is tilted. However, since holes have a large effective mass and the second barrier layer 8 exists in the first quantum well layer 3, holes cannot exist as stably as electrons in the first quantum well layer 3. This means that the wavefunction ψ that represents the existence probability of holes is
It is clear from the fact that _hh has a constriction near the second barrier layer and the existence probability in this portion is low. Therefore, when a small electric field of about 12 kV / cm is applied, most of the wave function ψ _hh of the electric field moves from the first quantum well layer 3 to the second quantum well layer 4 side.

The exciton absorption coefficient α is proportional to the square of the absolute value of the overlap integral of the electron and hole wave functions. That is, the formula

[0040]

[Expression 2] α∝ | ∫φ _el (z) φ _hh (z) dz | ² (2) In the formula, z is a coordinate in the thickness direction of the quantum well layer.

The shift amount of the exciton absorption peak at the time of applying an electric field (0 to 30 kV / cm) obtained based on the equation (2) and the absolute value of the overlap integral of the wave functions of electrons and holes are squared. The obtained values are shown in FIG. In the figure,
The solid line is the value of the exciton absorption peak shift amount, and the broken line is the curve obtained by plotting the squared value of the absolute value of the overlap integral.

As can be seen from this figure, when there is no electric field,
The value of the square of the absolute value of the overlap integral, which was about 93, decreases to about 0.05 when an electric field is applied of 12 kV / cm.
This indicates that the absorption coefficient α is significantly reduced. In this case, as can be seen from the calculation result of the exciton absorption peak shown in FIG. 3, the shift amount (red shift) on the long wavelength side of the exciton absorption peak wavelength is 0.005 μm.
It is extremely small as m or less.

Therefore, the absorption spectra when no electric field is applied (V = 0) and when an electric field is applied (V ≠ 0) are as shown in FIG. In this figure, the vertical axis is the absorption coefficient α, and the horizontal axis is the wavelength λ (μm) of the incident light. When there is no electric field (V = 0)
Is shown by a solid line, the absorption spectrum when an electric field is applied (V ≠ 0) is shown by a broken line, the absorption peak of a light hole is shown by L, and the absorption peak of a heavy hole is shown by H.

FIG. 5 shows the relationship between (a) absorption coefficient change Δα and (b) refractive index change Δn and wavelength λ (μm) when an electric field is applied (V ≠ 0). As is clear from this figure, in the quantum well structure of the present embodiment, FIG.
Unlike the case of the conventional example shown in, Δ
The sign of the value of α never changes. Therefore, FIG.
Substituting the value of Δα in (a) into equation (1), the change in refractive index Δn
Then, the value becomes extremely large as shown in FIG. 5B (improved by about one digit compared with the conventional one).

As described above, in the semiconductor quantum well structure of this embodiment, the change in refractive index Δn can be made extremely large as compared with the conventional one, so that the exciton absorption peak at the time of no electric field (V = 0). Compared to the case of the conventional quantum well structure, it becomes possible to sufficiently separate the used wavelength to the short wavelength side. That is, when the wavelength used is 1.55 μm, the absorption peak wavelength of excitons was about 1.44 μm in the conventional example, but it can be set to 1.40 μm or less in this embodiment. Therefore, it is possible to suppress the absorption of light by the excitons when there is no electric field (V = 0), and because the absorption is extremely small when an electric field is applied, it is possible to reduce the loss at the wavelength used, contrary to the conventional one. Become.

Further, when the conventional quantum well structure is applied to a directional coupling type switch (thickness of multiple quantum well core 0.4 μm, electrode length (ie, interaction length) 1.2 mm), optical switching occurs. 110kV /
An electric field of about cm is required to be applied. However, in this embodiment, a large refractive index change Δα can be obtained by applying an electric field of only 12 kV / cm. Therefore, the semiconductor quantum well structure of the present invention can significantly reduce the drive voltage of the waveguide type optical switch.

When it is desired to make the absorption peak wavelength of excitons in the absence of an electric field (V = 0) longer, the thickness of each quantum well layer (particularly, the sub-quantum wells 5 and 6) is increased or each barrier is increased. The layer thickness may be reduced. On the contrary, when it is desired to shorten the absorption peak wavelength of the excitons, the thickness of each quantum well layer (particularly, the sub-quantum wells 5 and 6) may be reduced or the thickness of each barrier layer may be increased. Such modifications will be described below.

FIG. 6 is a view for explaining an example of the semiconductor quantum well structure and the potential shape of the conduction band.
Has the same structure as the semiconductor quantum well structure shown in FIG. 1, but the thickness of each layer is different from that of the embodiment shown in FIG. That is, the thickness of the sub-quantum well layer 5 W ₁ 2.0 nm, the thickness of the sub-quantum well layer 6 W ₂ 3.5 nm, the thickness of the second quantum well layer 4 W ₃ 2.5 nm, the thickness of the first barrier layer 7. B ₁ 2.5 nm is the thickness B ₂ X nm of the second barrier layer 8 (X is a variable). Then, the thickness B ₂ of the second barrier layer 8
The exciton absorption peak wavelength (μm) was calculated according to the above, and the result was plotted in FIG. 7 (curve drawn by a solid line in the figure).

FIG. 6B shows a case where two quantum well inner barrier layers 8 and 10 are present in the first quantum well layer 3, and the dimensions of each layer are the same as those of the sub quantum well layer 5. Thickness W ₁ 2.0 nm Thickness of sub quantum well layer 6 W ₂ 3.5 nm Thickness of sub quantum well layer 9 W ₄ 2.5 nm Thickness of second quantum well layer 4 W ₃ 3.0 nm First barrier The thickness of the layer 7 was B ₁ 2.0 nm, the thickness of the second barrier layer 8 was B ₂ X nm, and the thickness of the third barrier layer 10 was B ₃ X nm (that is, B ₂ = B ₃ ).

Then, the absorption peak wavelength (μ) of the exciton corresponding to the thicknesses B ₂ and B ₃ of the second and third barrier layers 7 and 8
m) was calculated and the result was plotted in FIG. 7 (in the figure,
Curve drawn with a broken line).

As is apparent from FIG. 7, by increasing the thickness of the second barrier layer 8, the exciton absorption peak wavelength can be shortened. In addition, a plurality of barrier layers (second and third barrier layers 8,
By providing 10), the wavelength can be further shortened, and the exciton absorption peak wavelength in a wide wavelength range can be easily realized as compared with the case where there is one barrier layer. In order to shorten the wavelength, the thickness of the sub quantum well layers 5 and 6 and the thickness of the second quantum well layer 4 may be reduced, or the thickness of the second barrier layer 8 may be further increased.

<Embodiment 2> FIG. 8 is a structural example of a semiconductor quantum well structure according to the present invention, a potential shape in the absence of an electric field (V = 0), an electron wave function, a hole wave function, and And the transition energy of excitons.

The semiconductor quantum well structure shown in this figure is the same as that of the first embodiment, and the material of each layer is also the same. But,
The thickness of the second quantum well layer 4 is different. That is, in this embodiment, the dimensions of each layer are as follows (the values in the parentheses are the values of the first embodiment).

Sub-quantum well layer 5 thickness W ₁ 3.9 nm (3.9 nm) Sub-quantum well layer 6 thickness W ₂ 3.9 nm (3.9 nm) Second quantum well layer 4 thickness W ₃ 4.5 nm (3.6 nm) Thickness B ₁ of the first barrier layer 7 2.4 nm (2.4 nm) Thickness B ₂ of the second barrier layer 8 1.2 nm (1.2 nm) As in Example 1. Although the first quantum well layer 3 (thickness: W ₁ + B ₂ + W ₂ ) is thicker than the second quantum well layer 4 (thickness: W ₃ ), The thickness ratio of the quantum well layer 4 in Example 2 is slightly larger than that in Example 1. Therefore, the existence probability of holes in the first quantum well layer 3 at the time of no electric field (V = 0) is much larger than that in the first embodiment. In this example, the exciton absorption peak wavelength depending on the difference between the electron energy level E _el and the hole energy level E _hh is 1.39 μm.

When the electric field is applied (V ≠ 0), the electron wave function ψ _el is almost the same as when there is no electric field (V = 0), but the hole wave function ψ _hh is the second thin function. It shifts from the quantum well layer 4 to the side of the thick first quantum well layer 3. Hereinafter, this will be specifically described.

FIG. 9 shows the wave functions ψ _el and ψ _hh when an electric field of 5.3 kV / cm is applied.

Since the effective mass of electrons is small, the shape of the wave function thereof hardly changes even if the potential is tilted, as in the first embodiment. On the other hand, holes have a large effective mass, and due to the applied electric field (5.3 kV / cm), the potential of the second quantum well layer 4 is lower than that of the first quantum well layer 3 where the existence location is narrow, As shown in FIG. 9, the hole wave function ψ _hh shifts from the second quantum well layer 4 to the first quantum well layer 3 side.

FIG. 10 shows the applied electric field (0 to 30 kV / c).
It is a plot of the absorption peak shift amount of excitons at time m) and the value obtained by squaring the absolute value of the overlap integral of the wave functions of electrons and holes. In the figure, the solid line is the curve of the exciton absorption peak shift amount, and the broken line is the curve obtained by plotting the square of the absolute value of the overlap integral. As can be seen from this figure, the squared value of the absolute value of the overlap integral, which was about 0.2 when there was no electric field, increases to about 0.87 when the electric field is applied at 5.3 kV / cm.
On the other hand, the absorption peak wavelength of excitons, unlike the embodiment shown in FIG. 3, first shifts to the short wavelength side (blue shift) when an electric field is applied, and then shifts to the long wavelength side as the electric field strength increases.

Therefore, the wavelength dependence of the absorption coefficient α in the present embodiment when there is no electric field (V = 0) and when an electric field is applied (V ≠ 0) is as shown in FIG. In the figure, the absorption spectrum when there is no electric field (V = 0) is a solid line, and when an electric field is applied (V ≠
The absorption spectrum of (0) is shown by a broken line, and the absorption peak of a light hole is shown by L and the absorption peak of a heavy hole is shown by H.
As a result, the change Δα in absorption coefficient is as shown in FIG.

FIG. 12 shows (a) absorption coefficient change Δα and (b) refractive index change Δn when an electric field is applied (V ≠ 0). As is clear from this figure, the refractive index change Δα is positive in almost all wavelength regions, and the sign is not inverted. Therefore, as in the first embodiment, FIG.
Substituting the value of Δα in 2 (a) into equation (1), the change in refractive index Δ
When n is calculated, the value becomes extremely large as shown in FIG. 12B (improved by about one digit as compared with the conventional one).

In this embodiment, since the absorption peak wavelength of excitons shifts to the long wavelength side when an electric field is applied,
There is a concern that the loss will increase in the operating wavelength range (here, 1.55 μm), but as described above, the change Δn in the refractive index can be made extremely large as compared with the conventional one, so that there is no electric field (V The exciton absorption peak of (= 0) can be sufficiently separated from the used wavelength to the short wavelength side as compared with the case of the conventional quantum well structure. That is, when the wavelength used is 1.55 μm, the exciton absorption peak wavelength was about 1.44 μm in the conventional case, but it can be set to 1.38 μm or less in this embodiment. Therefore, even when an electric field is applied (V ≠ 0), absorption of light by excitons can be suppressed.

Note that FIG. 15 (case of the conventional example) and FIG.
As is clear from comparison with (in the case of this embodiment), absorption can be caused by applying an electric field of lower intensity in this embodiment compared to the conventional quantum well structure, so that the exciton of On the contrary, if the absorption peak is set to a wavelength close to the used wavelength (for example, 1.50 μm), it is possible to realize an intensity modulator having an extremely low driving voltage.

As described above, the feature of the present invention is that only the wave function of holes is significantly shifted when an electric field is applied. Therefore, this is when it is realized, for example, the thickness B ₁ of the first barrier layer 7 is equal to the thickness B ₂ of the second barrier layer, or the thickness W ₁ and the sub-quantum well sub-quantum well layer 5 The thickness of the first quantum well layer, the thickness of the sub-quantum well layer, the thickness of the second quantum well layer, and the thickness of each barrier layer are limited by making the thickness W ₂ of the layer 6 different. Without this, the thickness of each layer can be arbitrarily set.

In the first embodiment, both electrons and holes were present in the first quantum well layer when there was no electric field, and the holes moved to the second quantum well layer when an electric field was applied. On the other hand, in the second example, the electrons existed in the first quantum well layer and the holes existed in the second quantum well layer when there was no electric field, and the holes moved to the first quantum well layer when the electric field was applied. Furthermore, if the second quantum well layer is made slightly wider, both electrons and holes are present in the second quantum well layer when there is no electric field, and holes can move to the first quantum well when an electric field is applied. . Alternatively, when there is no electric field, electrons exist in the second quantum well layer and holes exist in the first quantum well layer, and it is possible to move the holes to the second quantum well layer by applying a voltage. In all of these cases, the change Δα of the absorption coefficient in almost all wavelength regions has the same sign, and the sign is not inverted.
Therefore, it is possible to obtain a large change in the refractive index.

In Examples 1 and 2 above, InAlAs
Although the system has been described, the present invention is not limited to this, and other InP, InGaAsP, or G
It is possible to use aAs system, AlGaA system, or the like.
Further, only the heavy hole that modulates the TE mode light has been described, but the present invention is not limited to this, and it is also possible to modulate the TM mode light by using the absorption peak of the light hole. Further, in the example described here, when the wave function of holes moves from the thick first quantum well layer to the thin second quantum well layer when an electric field is applied, the exciton absorption peak When the wavelength becomes longer and the wave function of the hole moves from the thin second quantum well to the thick first quantum well, the exciton absorption peak wavelength becomes shorter. However, depending on the thickness of the first quantum well layer, the sub-quantum well layer, the second quantum well layer, the first barrier layer, the second barrier layer, etc. Of course, there may be cases where the opposite of the above example occurs. Further, by providing a barrier layer in the second quantum well layer, it is possible to shorten the absorption peak wavelength of excitons or lower the voltage for moving holes.

Furthermore, for example, in the first embodiment, the thickness of the second quantum well layer is made substantially equal to the thickness of the sub quantum well layer of the first quantum well (W ₁ = W ₂ = W ₃ ). The holes can be moved by a built-in voltage inside the semiconductor (for example, about 0.8 V in the case of InP type). In this case, since the holes have already moved as shown in FIG. 2 when the crystal grows in the semiconductor quantum well structure, the square value of the absolute value of the overlap integral is small. Therefore, when a voltage is applied from the outside in the direction of canceling the built-in voltage, holes return to the first quantum well 3 as shown in FIG. 1, and as a result, the squared absolute value of the overlap integral increases. In this case, the absorption coefficient change Δα becomes positive in almost all wavelength regions, and a large change in the refractive index can be caused.

As described above, when the semiconductor quantum well structure is crystal-grown by the built-in voltage, holes are caused to move in advance, and then a voltage is applied from the outside in the direction of canceling the holes. The method of returning to the original position to obtain the refractive index change Δn can be applied to the case of the second embodiment of the present invention. In this case, the holes change from the state of FIG. 9 to the state of FIG. 8, and the absorption coefficient change Δα has a negative value in almost the entire wavelength region, and it is possible to cause a large change in the refractive index. In the above description,
Although both holes and electrons have been described by using the transition between the ground levels, the same effect can be obtained by considering the transition between the higher levels.

[0068]

As described above, according to the semiconductor quantum well structure according to the present invention, only the wave function of holes largely moves when an electric field is applied, and therefore, the sign of the absorption coefficient change Δα in almost the entire wavelength range. Are equal. As a result, it is possible to obtain an extremely large change in refractive index Δn as compared with the conventional quantum well. That is, since the exciton absorption peak can be far away from the used wavelength to the short wavelength side, the light absorption loss in the absence of an electric field is low. Even if a barrier layer is provided in the first quantum well layer, the total thickness of the first quantum well is QC.
Dominate the SE. Therefore, while ensuring the thickness of the quantum well required to realize a large QCSE, the thickness in the quantum well is controlled by controlling the thickness that is extremely controllable than the composition without changing the crystal composition. By setting the thickness and number of layers, it is possible to reproducibly obtain an exciton absorption peak at an arbitrary wavelength over a wide wavelength range, which has been difficult in the past. Furthermore, since the composition of the crystal can be simplified, without impairing the crystallinity and reproducibility of the crystal,
It has the effect of achieving a sharp exciton absorption peak.

[Brief description of drawings]

FIG. 1 is a diagram illustrating a potential shape, an electron wave function, a hole wave function, and exciton transition energy in an example of a semiconductor quantum well structure according to the present invention (when no electric field is applied).

FIG. 2 is a diagram illustrating a potential shape, an electron wave function, a hole wave function, and exciton transition energy in an example of a semiconductor quantum well structure according to the present invention (when an electric field is applied).

FIG. 3 is a diagram showing the voltage dependence of the exciton absorption peak shift amount and the electric field dependence of the value obtained by squaring the absolute value of the overlap integral of the wave functions of electrons and holes.

FIG. 4 shows no electric field (V = 0) and no electric field applied (V ≠).
It is a figure showing the absorption spectrum of 0).

FIG. 5 is a diagram showing (a) absorption coefficient change Δα and (b) refractive index change Δn when an electric field is applied (V ≠ 0).

6 is a view for explaining a modification of the semiconductor quantum well structure shown in FIG. 1 and a potential shape of a conduction band,
FIG. 3A is a diagram showing one quantum well layer barrier layer in the first quantum well layer, and FIG. 6B is a diagram showing two quantum well layer barrier layers in the first quantum well layer. Is.

7 is a diagram in which an absorption peak wavelength (μm) of excitons according to the thicknesses of the second and third barrier layers in FIG. 6 is calculated, and the result is plotted.

FIG. 8 shows the configuration of the second embodiment of the semiconductor quantum well structure according to the present invention, and the potential shape, electron wave function, hole wave function, and exciton transition energy when there is no electric field (V = 0). It is a figure for explaining.

FIG. 9 Wave functions ψ _el and ψ _hh when an electric field is applied
FIG. 6 is a diagram for explaining the behavior of the.

FIG. 10 is a diagram showing the electric field dependence of the exciton absorption peak shift amount and the electric field dependence of the value obtained by squaring the absolute value of the overlap integral of the wave functions of electrons and holes.

FIG. 11 shows no electric field (V = 0) and no electric field applied (V ≠).
It is a figure for demonstrating the wavelength dependence of absorption coefficient (alpha) of 0).

12 is a diagram showing (a) absorption coefficient change Δα and (b) refractive index change Δn when an electric field is applied (V ≠ 0) in FIG. 11.

FIG. 13 is a diagram for explaining a potential shape, a wave function of an electron, a wave function of a hole, and a transition energy of excitons of a conventional MQW structural constituent unit (in the case of no electric field).

FIG. 14 is a diagram for explaining a potential shape, a wave function of an electron, a wave function of a hole, and a transition energy of excitons of a conventional MQW structural unit (when an electric field is applied).

FIG. 15 is a diagram showing the electric field dependence of the exciton absorption peak shift amount of a conventional quantum well (W = 9 nm).

FIG. 16 is a diagram when no electric field is applied (V = 0) and when an electric field is applied (V ≠).
It is a figure which shows the relationship between the absorption coefficient (alpha) of the conventional quantum well and incident light wavelength in (0).

FIG. 17 shows (a) a change in absorption coefficient Δα and (b) a change in refractive index Δ of a conventional quantum well when an electric field is applied (V ≠ 0).
It is a figure which shows the relationship between n and the wavelength of incident light.

FIG. 18 is a diagram for explaining the relationship between the Al content and the absorption peak wavelength of excitons in a conventional quantum well.

FIG. 19 is a diagram for explaining an example of a conventional asymmetrically coupled quantum well.

FIG. 20 is a diagram for explaining a second example of a conventional asymmetrically coupled quantum well.

[Explanation of symbols]

 1 Coupling quantum well outer barrier layer made of InAlAs 2 Quantum well layer made of i-InGaAlAs 3 First quantum well layer 4 Second quantum well layer 5 Sub-quantum well layer 6 Sub-quantum well layer 7 Barrier layer between quantum well layers (first One barrier layer) 8 Quantum well layer barrier layer (second barrier layer) 9 Sub-quantum well layer 10 Quantum well layer barrier layer (third barrier layer)

Claims (6)

[Claims] 1. A coupled quantum well comprising a first quantum well layer and a second quantum well layer separated by a barrier layer, a wave function in the coupled quantum well and an adjacent coupled quantum well adjacent to the coupled quantum well. In a semiconductor quantum well structure consisting of a coupled quantum well outer barrier layer provided on both sides of the coupled quantum well so as to suppress coupling with a wave function in the inside, at least in the first quantum well layer, the first quantum well layer, A semiconductor quantum well structure, wherein at least one barrier layer in a quantum well layer is provided for reducing the probability of existence of holes in one quantum well layer. 2. The semiconductor quantum well structure according to claim 1, wherein the barrier layer between the quantum well layers and the barrier layer inside the quantum well layer are made of a semiconductor material having the same band gap as that of the outside barrier layer of the coupled quantum well. Semiconductor quantum well structure characterized by. 3. The semiconductor quantum well structure according to claim 1, wherein the first quantum well layer and the second quantum well layer are made of a semiconductor material having the same bandgap. Construction. 4. The semiconductor quantum well structure according to claim 1, wherein electrons are present in at least the first quantum well layer or the second quantum well layer both in the absence of an electric field and in the application of an electric field. There are many holes on one side, more holes are present in the first quantum well layer than the second quantum well layer when no electric field is applied, and the holes are present from the first quantum well layer when an electric field is applied. A semiconductor quantum well structure, wherein the semiconductor quantum well structure moves to the second quantum well layer. 5. The semiconductor quantum well structure according to claim 1, wherein electrons are present in at least one of the first quantum well layer and the second quantum well layer when no electric field is applied and when an electric field is applied. There are more holes on one side, and more holes are present in the second quantum well layer than in the first quantum well layer when no electric field is applied, and from the second quantum well layer when the electric field is applied. A semiconductor quantum well structure characterized by moving to a well layer. 6. The semiconductor quantum well structure according to claim 1, wherein electrons are not moved by applying a voltage in a direction that cancels a built-in voltage in the semiconductor quantum well structure.
A semiconductor quantum well structure, wherein holes move from the first quantum well layer to the second quantum well layer or from the second quantum well structure to the first quantum well layer.