JPH043125A - Nonlinear optical element - Google Patents

Abstract

PURPOSE:To prevent an operating speed from being limited by the relaxation time of electrons or holes by impressing a static electric field to a region consisting of the optical nonlinear material of a light guide region. CONSTITUTION:This optical element has the light guide region having the region of a material exhibiting cubic optical nonlinearity (X<(3)>) and has an electric field impressing means which impresses the static electric field to the region of the material exhibiting the optical nonlinearity of this light guide region. Light to be controlled is made incident from a port A and is bisected by a beam splitter 1. The one of the beams arrives via a mirror 3 at a beam splitter 2. The other beam is reflected by a multilayered dielectric film mirror 4 and is then made incident on a nonlinear waveguide 5 which is one of the nonlinear optical elements, where the beam is subjected to a phase change. The beams are thereafter multiplexed by the beam splitter 2. The multiplexed component appears in the port C or port D by the relative phase difference between the two beams at the time of multiplexing. The relative phase difference is generated while the beams propagate in the nonlinear waveguide. The operating speed is prevented from being limited by the relaxation time of the carriers in the nonlinear optical material to be used.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a nonlinear optical element that enables ultrahigh-speed optical communication and optical signal processing.

[Conventional technology]

In optical information processing and optical communication, it is necessary to perform optical control such as optical modulation to add a signal to light and optical calculation. Currently in practical use,
In these systems, a control method using electrical signals (electro-optical control method) is used. That is, optical modulation and optical calculation are performed by directly modulating a semiconductor laser, which is a light source, with an electric current, or by changing the refractive index or absorption by applying a voltage to a semiconductor or dielectric material.

In this control method using electrical signals, the processing speed is limited by the speed of the element itself and the speed mismatch between the electrical signal and the controlled light, and it is extremely difficult to increase the processing speed to about sub-nanoseconds or higher. Therefore, a so-called light-to-light control method, which performs optical modulation and optical calculation using an optical signal, is being considered as a method to overcome the processing speed limitations of the conventional electric-to-light control method. In this light-light control method,
Since there is no restriction due to the CR time constant, it has the characteristic that it is possible to achieve ultra-high speed.

[Problem to be solved by the invention]

The operating characteristics of an optical element that controls light with light largely depends on the optical nonlinear material used in the optical element. The magnitude of optical nonlinearity of a material is expressed as the value of χfj+ of that material. Characteristics required of optical devices include not only high speed as described above, but also low power such as the ability to operate with low optical power. In order to achieve these, a nonlinear optical material is required that exhibits a short relaxation time τ for optical excitation (high speed) and a large χ0) (low power). However, nonlinear optical materials that generally have a large χt31 exhibit a large τ value. Further, even if the same material is used under different conditions of use such as light wavelength, the physical phenomena involved in the manifestation of optical nonlinearity will differ, but physical phenomena that exhibit a large χ+31 generally also exhibit a large τ. A detailed explanation of this can be found in R.A. Finnyer, ed., Optical Phase.
Consegation (Academic Press, 1983)
See Chapter 10 of 2010).

Among various combinations of nonlinear materials and nonlinear phenomena,
It is known that methods of generating electron/hole plasma by light in semiconductor materials exhibit large χ(3). This is a phenomenon in which the refractive index of the semiconductor changes as a result of exciting electrons in the valence band to the conduction band by control light (band filling effect). However, the relaxation time of this phenomenon is
Even in the case of direct transition type semiconductors such as GaAs, it takes more than 10-9 seconds, which is insufficient for applications such as ultra-high-speed signal processing (Optical Phase Conjugation, Chapter 10 (mentioned above)) .

The purpose of the present invention is to reduce the power by utilizing the (large) The object of the present invention is to provide a nonlinear optical element that is not limited by hole relaxation time.

[Means to solve the problem]

The nonlinear optical element of the present invention has third-order optical nonlinearity (χ
0))), comprising an optical waveguide region having at least a region made of a material exhibiting 0)), and an electric field applying means for applying a static electric field to a region of the optical waveguide region made of an optically nonlinear material. It is said that

Another nonlinear optical element of the present invention includes an optical waveguide region having at least a region made of a third-order nonlinear optical material consisting of a quantum well structure, a multiple quantum well structure, or a large number of parallel quantum well wire structures, A static electric field applying means is provided in which the direction of the electric field applied to the region made of the third-order nonlinear optical material is parallel to the laminated plane of the quantum well or multiple quantum well, or in the axial direction of the quantum well thin wire. It is a feature.

The notation of the third-order optical nonlinear constant χj31, which is based on the photoexcitation of electrons from the valence band to the conduction band in a C('l”m) band semiconductor, is given by D, A, B, Miller et al. (D , A. E. Miller et al., Optics Communication, Vol. 35, 2.
2, pp. 221-226, 1980). Equation (11) in this document has χ13', but there is clearly a tights error in this notation. Therefore, if we calculate χt31 using essentially the same method as in the same document, we get Re [χ"(ω5,' (tJp+ -ωp,
ωs)'3. Here, e is the electron charge, h is a blank constant, and h=h/(2π)2m. is the mass of the electron, snore is the dipole moment between the valence band and the conduction band, ω1-E, /
h, E, are the band gap, μ is the reduced mass relative to the effective mass of electrons and holes, and TI and T2 are the longitudinal and transverse relaxation times of the electron. Further, here, the analysis of the above-mentioned literature is extended to consider the case where the angular frequency ω2 of the pump light (control light) and the angular frequency ω of the signal light (controlled light) are different. In addition, since nonlinear refractive index changes are considered here, only the real part of )
Here, n is the linear refractive index, and n2I is the change in the refractive index due to the stiffness.

This n2 is called a nonlinear refractive index, and has a relationship with the above-mentioned χ13' (unit: cgs).

From the above, the nonlinear refractive index change caused by photoexcitation of electrons from the valence band to the conduction band in a semiconductor can be interpreted as follows. When a semiconductor is irradiated with light of a wavelength whose energy is greater than the band gap, many electrons are excited from the valence band to the conduction band as the light is absorbed. However, as electrons accumulate in the conduction band, excitation from the valence band to the conduction band becomes more difficult (band filling effect). This effect is expressed as χ13]. The real part of χ(3) is related to the refractive index of the semiconductor through equations (2) and (3). Here, the speed of conduction ♂X excitation from the valence band due to light absorption is extremely fast (several hundreds of fs).

In other words, it can be considered that the nonlinear refractive index appears the moment light enters the semiconductor. When this nonlinear phenomenon is applied to an optical device, the operating speed of the optical device is limited because the already excited carriers do not disappear immediately even after the light irradiation is stopped. This depends on the size of T (longitudinal relaxation time or carrier recombination time) in equation (1). T
, is more than 1 Ons even in a direct transition semiconductor, so χ
(3゛ cannot follow fluctuations in light intensity of 0.1 GHz or more.

Focus on what you do. The speed limit of this type of nonlinear phenomenon is usually considered to be the carrier recombination time as described above. However, if the carriers (N electrons or holes) are swept out of the optical path of the signal light by some method, the signal light will no longer be affected by the carriers, so non-linearity will occur depending on the carrier sweep time. Disappear. To sweep the carrier spatially,
A direct current electric field (or a magnetic field may be used) may be applied to the semiconductor.

Taking high purity undoped GaAs as an example,
When the electric field strength is several KV/am, the drift velocity of electrons can be 10'cm/s or more even at room temperature. If the optical path width is about 1 μm, the time required to sweep the electrons is about 10 ps. In this way, the optical device can operate at about 100 GHz.

What should be noted here is that if the speed of the optical device is increased in this way, the value of χ'3' will decrease accordingly, and as a result, the optical power required for the operation of the device will also increase. This is because sweeping the electrons out of the optical path is approximately equivalent to decreasing T1 in equation (1). Therefore, in order to increase the speed of the − layer, it is necessary to increase the drift speed and at the same time increase χ131. For this purpose, the bulk GaAs described above is replaced with a quantum well structure, a multiple quantum well structure, or a quantum wire structure.

By introducing such a structure, the electron wave function, which normally spreads three-dimensionally, can be confined two-dimensionally or one-dimensionally. As a result of the reduction in the dimension of the spread, the number of modes of electrons in the direction of the reduced dimension is greatly reduced, and thus the scattering probability of electrons is significantly reduced.

As a result, electron mobility and drift speed are greatly improved (for example, Physics and Applications of Semiconductor Superlattices 1, edited by the Physical Society of Japan, Chapter 10).

What must be noted here is that there is significant anisotropy in carrier mobility in such a structure. In other words, in the case of quantum wells, although the carrier drift speed in the direction parallel to the stacked layers increases, it is extremely difficult for current to flow in the direction perpendicular to the layers. The same is true for quantum well thin wires, and only the drift speed in the axial direction of the thin wires is improved.

By introducing a quantum structure such as a multiple quantum well as described above, the drift velocity of carriers increases, and a higher speed nonlinear element becomes possible. In addition, by introducing a quantum well structure, excitons J- are generated even at room temperature, and therefore a thick dipole moment accompanying excitons is generated. As is clear from equation (1), χ(3も) is proportional to the fourth power of the dipole moment p, so when p increases by 10 times, χ131 increases by 101 times. The optical power decreases to 1/10' when the operating speed remains the same, and even if the operating speed is increased by 104, the increase in optical power remains at about 102.

〔Example〕

Next, the nonlinear optical element of the present invention will be explained in detail using the drawings.

FIG. 2 is a schematic configuration diagram of an example in which the nonlinear optical element of the present invention is applied to a light modulation device using five-light-light control. In the figure, controlled light with a wavelength of 1455 μm is incident from boat ①, is split into two by beam splitter 1, and one reaches beam splitter 2 via mirror 3. After being reflected by the dielectric multilayer mirror 4, the other beam enters the nonlinear waveguide 5, which is one of the nonlinear optical elements of the present invention, and after experiencing a phase change, the beam splitter 2 It is combined with Then, due to the relative phase difference between the two beams at the time of combining, the combined component appears in the ◎ boat or the ■ boat. The relative phase difference occurs during propagation in the nonlinear waveguide. In other words, control light with a wavelength of 0.81 μm is incident on the nonlinear waveguide 5 from the boat (2) and is absorbed (the dielectric multilayer mirror 4 has a wavelength of 0.81 μm).
, 81 μm, the reflectance is zero. ). As a result, the refractive index of the nonlinear waveguide 5 changes as shown in equations (1) to (3) in the section on effects.

The details of the structure of a nonlinear waveguide, which is an embodiment of the nonlinear optical element of the present invention, are shown in an enlarged scale in FIG. This is a schematic cross-sectional view of the same waveguide structure.

This nonlinear waveguide is fabricated on a substrate made of GaAs 51 with Aρ
orGao, aA S 52 with a thickness of 3 μm, then Aj12As53 with a thickness of 2 μm, and then GaAs
-AρAs54 multiple quantum well structure (MQW) with 05μ
m thickness, and finally A n O, 2G a o, sA
S55 was grown to a thickness of 0.5 μm by molecular beam epitaxy. A, n O, 2G a
The o8A three layer 55 is then etched into a ridge shape with a width of 0.5 μm as shown in FIG.

The regions 56° and 57 have become n+ type by diffusing Si. Gold electrodes 58°59 on the surfaces of regions 56 and 57
is formed by vapor deposition, and a DC power source 6 is connected to this electrode 58,59.
0 is connected. 54 AfAs-GaAs MQ
W is obtained by growing 25 cycles of 100 GaAs layers and 100 AAAs layers.

The light input and output ends of the nonlinear waveguide 5 are coated with a film that prevents light reflection. In the nonlinear waveguide, the ridge portion 55 and the MQW layer 54 around it form a waveguide portion. This is because the refractive index of the MQW layer 54 is approximately 3.1 and the refractive index of the Clara F layer 53 is approximately 2.85, so that vertical confinement is performed.

The lateral binding is performed by the ridge portion 55. In other words, since the refractive index of the ridge is as high as 3.22, the light is
It is confined within the ridge portion and the MQW layer 54 around it without spreading laterally within the QW layer.

Here, the cross section of the ridge portion 55 is 0.5 μm x 0.5 μm.
.

As a result of setting the thickness of the MQW layer 54 to 0.5 μm, approximately 50% of the optical power is distributed to the bridge portion, and the remaining 50% is distributed to the MQW layer 54 . Further, the lateral spread of light within the MQW layer is about 1 μm.

Ridge portion 55. MQW layer 54. The hand gear of the cladding layer 53 has a voltage of 1.73 eV, 1.53 eV, and 2.
It is about 76v. Therefore, the optical absorption coefficient for the controlled light having a wavelength of 155 μm is almost negligible. For control light with a wavelength of 0.81 μm, the MQW layer 54
Only one part shows strong absorption, and the absorption in other parts can also be ignored.

The control light absorption in the MQW layer 54 has a wavelength of 081
It is particularly large because μm corresponds to the exciton wavelength of the same MQW layer. Here, in equation (1), e-4,8Xl O”
esu, c=3×10”aIl/s, 'A=1.05
4×10-'7ergSeCrma=0-99X 10
(=;g, mri=0-62ma+ mAA-0,62m
.

The absorption coefficient α is approximately 5000 cm-' (the factor 3.0 in p2 is the bulk value p2=1.0 of the dipole moment regarding exciton absorption in the MQW structure).
The correction factor for 33 x m o x E.

This means that the absorption coefficient is approximately tripled). Here, the reason why T+=i00 ps is set is that a DC power source of 1 cm is connected between the electrodes 58 and 59. Here, the electrode spacing is about 10 it m, and as a result, M
An electric field of IKV/cm is maintained in the layer direction of the QW layer 54. This electric field causes the light guide section to have a wavelength of λ=0.81 μm.
The carriers generated by the control light are swept out of the waveguide section in about 100 ps, which is approximately equal to the longitudinal relaxation time T.
1 to 100 ps. Therefore, the operating speed of this element is 10 GHz. If no electric field is applied, the original T of the GaAs-based material is l On
s or more, the maximum operating speed of the element is 0.1 GHz or less.

Here, the control power required for the operation of this device is as follows (the power of the controlled light is arbitrary). In order for the device shown in FIG. 2 to operate, it is required that the phase shift of the signal light (controlled light) accompanying nonlinear waveguide propagation differs by π radians depending on the presence or absence of the control light. This relative phase shift Δφ is given by Δφ. Here, α is the absorption coefficient λ of the pump light,
The wavelength of the signal light, n is the refractive index (23.0), and c is 3 x
At 1010an/s, is the pump light intensity (W/n()
It is jin. If Δφ=π and the above value is applied to equation (4) without an operator, it becomes 3.2×10”W/n(.The size of the waveguide mode is approximately 1 μm×1 μm, so the required control light power is approximately It becomes 32mW.

The above estimate actually overestimates the power considerably. This is because it is assumed that the semiconductor is in a weak absorption saturation state in the pressure induction process of equation (1). This assumption is a good assumption in a wavelength region where the absorption coefficient is not so large. However, when using an exciton wavelength with extremely high light absorption as in this example,
A large error occurs in the direction where the required power becomes considerably large. This is because at the exciton wavelength, a remarkable absorption saturation effect is exhibited at an optical power of 1/10 or less compared to the bulk. Therefore, χ0' is much larger than the magnitude of equation (1), and the optical absorption coefficient is also saturated at low power.

Therefore, αa in equation (4) is effectively reduced. For these reasons, in the configuration of this embodiment], 10G at about mW
Hz operation is possible.

This means that by increasing the voltage of the DC power supply and increasing the drift speed of carriers, the optical power required for higher-speed operation does not increase so much.

Although the nonlinear optical element of the present invention has been described in detail using Examples, the present invention is not limited to the Examples. In this embodiment, a waveguide type element is used, but it may also have a solid structure in which light is incident on the substrate in the perpendicular direction. Furthermore, although a GaAs-based material is used in this embodiment, it is obvious that other materials such as InP-based materials may be used depending on the wavelength. Furthermore, various modifications can be considered to the method of applying the electric field, such as directly connecting the electrode to the semiconductor structure (without intervening a heavily doped layer).

〔Effect of the invention〕

As described above, the operating speed of the nonlinear optical element of the present invention is
It is not limited by the carrier relaxation time of the nonlinear optical material used.

[Brief explanation of the drawing]

FIG. 1 is a schematic diagram of a nonlinear optical element of the present invention. FIG. 2 is a schematic diagram of an example of the use of the nonlinear waveguide shown in FIG. 1.2 is a half mirror, 13 is a mirror, 4 is a dielectric multilayer mirror, 5 is a nonlinear waveguide, 51 is a GaAs substrate, 52
is AI! 0.20 a o-s As buffer layer, 53 is ApAs cladding layer, 54 is GaAs-Af
fAs MQW layer, 55 is Al1.2 Ga o,
*A S ridge portion, 56.57 is a Si-doped n+ region, 58.59 is a gold electrode, and 60 is a DC power source. Agent Patent Attorney Susumu Uchihara 55A16αA5 Takuzu? figure

Claims (2)

[Claims] (1) An optical waveguide region having at least a region made of a material exhibiting third-order optical nonlinearity (χ^(^3^)),
A nonlinear optical element comprising electric field applying means for applying a static electric field to a region of the optical waveguide region made of a material exhibiting optical nonlinearity. (2) An optical waveguide region having at least a region made of a third-order nonlinear optical material consisting of a quantum well structure, a multi-quantum well structure, or a large number of parallel quantum well wire structures, A nonlinear optical element comprising a static electric field applying means in which the direction of the electric field applied to the region is parallel to the laminated plane of the quantum well or multi-quantum well, or in the axial direction of the quantum well wire.