JPH05188412A - Optical semiconductor element - Google Patents

Abstract

PURPOSE:To obtain the optical semiconductor with which an insertion loss and propagation loss are negligible, low-voltage driving is possible and a wide wavelength variable range is obtainable by forming the semiconductor into a p-i-n-i-p or n-i-p-i-n structure provided with the waveguides as i layers in such a manner that electric fields are applied independently to two layers of the laminated waveguides. CONSTITUTION:A directional coupler constituted of two layers of the laminated semiconductor waveguides 1, 2 is constituted. The layer constitution including two layers of these semiconductor waveguides 1, 2 is the p-i-n-i-p or n-i-p-i-n structure. Two layers of the semiconductor waveguides 1, 2 are both constituted of the i layer and the voltage impression is independently controlled through the p-i-n structure. The 2nd waveguide 2 where the light to be used is amplified by current implantation and the 1st waveguide 1 where a refractive index change is induced by the voltage impression or current implantation and optical filtering control and optical modulation, etc., are executed are separately formed. The change in the refractive index arising from the refractive index control of the 1st waveguide 1 is, therefore, compensated by using the optical amplification in the 2nd waveguide 2 and always a stable transmission gain is obtd. regardless of wavelengths.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical modulator used for optical transmission, optical switching, optical information processing, optical recording, wavelength division multiplexing (WDM) optical communication, wavelength division multiplexing optical switching, optical arithmetic, etc. The present invention relates to an optical semiconductor device such as a wavelength tunable filter used in.

[0002]

2. Description of the Related Art Conventionally, driving voltage is low and high speed driving is possible.
Further, there is known an optical modulator using a semiconductor that can be easily integrated with other optoelectronic devices such as a semiconductor laser. For example, an absorption-type optical modulator that uses the absorption edge shift of a semiconductor (such as a semiconductor layer having a bulk structure or a quantum well structure) by applying an electric field, and a directional coupler type or total reflection that uses the change in refractive index by applying an electric field. Type modulators are known.

The former is composed of a semiconductor waveguide having a pin structure, and when an electric field is applied, the absorption edge is shifted by the Franz-Keldysh effect or QCSE (quantum confined Stark effect) as shown in FIG. The absorptance changes, and the transmittance of light of a certain wavelength can be controlled (A
applied Physics Lett. 47, p.
1148-1150 (1985)). However, in this type of optical modulator, it is necessary to bring the used wavelength close to the absorption edge in order to improve the extinction ratio, and therefore the transmittance in the transmissive state is low. Therefore, there is a drawback that the insertion loss is large. In addition, there is a problem that the wavelength of light cannot be modulated depending on the absorption edge wavelength and the usable wavelength, that is, the modulated wavelength is limited.

In the latter directional coupler type or total reflection type, as shown in FIGS. 9 (a) and 9 (b), a coupling region (in the case of (a)) or a crossing region ((in the case of (a)) of two waveguides is used. In the case of b)), an electrode is provided, and the refractive index is changed by applying an electric field to the electrode to shift the light wave between the waveguides (see IEICE research report OQE86-39). As a result, the output light from one waveguide emission end is modulated. However, in this type of optical modulator, even though the optical modulation degree is controlled by changing the refractive index, the optical absorption rate inevitably changes at the same time as the refractive index change, so the optical modulation is stable. It had the drawback that it couldn't be applied. Moreover, shorten the element length (that is, shorten the waveguide length).
At the same time, if an attempt is made to obtain a large change in the refractive index with respect to a constant electric field in order to reduce the drive voltage, there is a problem that the wavelength of the modulated light is set close to the wavelength range in which the light absorption is large.

On the other hand, conventionally, in a wavelength division multiplexing system, a demultiplexer has been used as a device for dividing a channel. For example, a wavelength dispersion element such as an interference filter or a grating is used, and demultiplexing is performed by utilizing the fact that transmission / reflection components are separated or the reflection angle is different depending on the wavelength. However, while such a demultiplexer can receive information of several wavelengths at the same time, it divides the information multiplexed in the wavelength region into a spatial region, which leads to an increase in the area of the element, and the relationship with that makes it possible to integrate optical signals. Since the number of detectors is limited, it is difficult to increase the wavelength multiplexing density.

As a means for solving this, there is a wavelength tunable filter. If this is used, one photodetector can sufficiently cope with the wavelength multiplexing system, and if the number of channels of the wavelength tunable filter is further expanded, the wavelength multiplexing degree is obtained. Can be increased. A device that uses a TE-TM mode converter (Applied Physics Lett.
53, 13, (1988)). Device using even-odd mode converter (IEICE research report OQE81-
129, (1981)), a device utilizing SAW (surface acoustic wave) (IEICE research report US88-4).
2, (1988)) and the like are known. But,
Although all of these have a wide wavelength tunable range of 100 Å or more, all of them are devices using LiNbO ₃ , so that the coupling loss with the photodetector becomes a problem. Also, in order to obtain the refractive index by the electro-optic effect (Pockels effect),
A high voltage of several tens of volts to hundreds of tens of volts is required. Further, as a wavelength tunable filter using a compound semiconductor such as GaAs or InP, a DFB (distributed feedback type) laser, a DBR (distributed reflection type) laser, or a Fabry-Perot type laser is used below an oscillation threshold (electronic information communication Academic Research Report OQE88-65, (1988))
It has been known. These have the advantage that they can be integrated with photodetectors and can have gain by current injection. However, the wavelength tunable range is directly determined by the range of change in the refractive index, so at present only values of several tens of tens are obtained.

[0007]

As described above, in the conventional optical modulator using the semiconductor waveguide, the control of the light absorption rate or the control of the refractive index was used. The insertion loss, extinction ratio, degree of freedom of the modulated wavelength, etc. were not satisfactory. Further, even with the filter according to the conventional technique, a wavelength tunable filter having sufficiently satisfactory performance cannot be obtained.

Therefore, in view of the above problems, an object of the present invention is to provide an optical semiconductor device having a structure which can be used as an optical modulator, a wavelength tunable filter or the like which can have a sufficiently satisfactory performance.

[0009]

In an optical semiconductor device that achieves the above object, there is provided a directional coupler composed of stacked two-layer semiconductor waveguides, and the two-layer semiconductor waveguides are included. The layer structure is p-i-n-i-p or n
-I-p-i-n structure, and the two-layer semiconductor waveguides are both composed of i-layers, and are independently controlled through the p-i-n structure.

More specifically, the first and second semiconductor waveguides are composed of waveguides having different refractive indexes and / or layer thicknesses from each other, or the waveguides of these two layers each have a central intensity. A grating is formed in the region where wave modes are mutually coupled.

According to the structure of the present invention, the second waveguide in which the light to be used is amplified by the current injection, and the refractive index change caused by the voltage application or the current injection to perform the optical filtering control or the optical modulation. It is different from the waveguide of 1. Therefore, the change in the absorptance associated with the control of the refractive index of the first waveguide can be compensated by using the optical amplification in the second waveguide, and a stable transmission gain can always be obtained regardless of the wavelength. Further, since the non-selected wavelength does not move to the second waveguide, it cannot receive optical amplification. Therefore, compared to the selected wavelength, the light intensity level can be lowered,
The S / N ratio can be improved.

[0012]

Embodiment 1 FIG. 1 shows an embodiment of a wavelength tunable filter according to the present invention.
First, the principle and configuration will be described. The present embodiment has a laminated type directional coupler, and the two-layer waveguide 1 constituting the directional coupler,
Reference numeral 2 is a so-called vertical asymmetric directional coupler having different refractive indexes and layer thicknesses. This directional coupler satisfies the condition that the 0th and 1st modes propagate, and the 0th mode mainly propagates in the upper waveguide 2 and the 1st mode mainly propagates in the lower waveguide 1. To do.

If the propagation constant of the 0th-order mode is β ₀ and the propagation constant of the 1st-order mode is β ₁ , the propagation constants β ₀ and β ₁ are greatly different due to the asymmetry between the waveguides 1 and 2. .. At this time, the upper waveguide has a higher refractive index than the lower waveguide,
In addition, if the layer thickness is made thin, there is a wavelength at which the values of the effective refractive index are the same, although the two propagation constants have different wavelength dispersions as shown in FIG. That is, at this wavelength λ _C ,
Coupling of the 0th and 1st order modes occurs.

However, when the upper waveguide has a higher refractive index and a larger layer thickness than the lower waveguide, wavelengths having the same propagation constant do not occur. Therefore, in this directional coupler, a grating 3 for compensating for the difference in propagation constant is formed in the upper waveguide 2. Assuming that the period of the grating 3 is Λ and the incident wavelength is λ, at a wavelength λ that satisfies β ₀ (λ) −β ₁ (λ) = 2π / Λ (1), as shown in FIG. And first-order mode coupling occurs.

With the above structure, the light 4 incident on the lower waveguide 1 becomes a first-order mode, and is coupled to the 0th-order mode at a specific wavelength λ to cause a transition to the upper waveguide 2. At other wavelengths, no coupling with the 0th-order mode occurs, and the light propagates through the lower waveguide 1 as it is.

In the tunable filter of the first embodiment, the directional coupler is made of GaAs / AlGaAs, the upper waveguide 2 is in a carrier non-doped state, that is, an i (intrinsic) layer, and A p-layer in which the cladding layer 5 and the contact layer 6 above the waveguide 2 are p-type doped, an n-layer in which the intermediate cladding layer 7 between the upper and lower waveguides 1 and 2 is n-type, and a lower waveguide layer 1 is an i layer and the clad layer 8 below the lower waveguide 1 is a p-type p-type doped layer. That is, the two-layer waveguide (i
The layers form a p-i-n-i-p structure as a whole.

According to this structure, the substrate-side p-side electrode 9 and n
When a reverse electric field, that is, a negative voltage is applied to the p-side electrode 9 between the electrodes (not shown) of the mold intermediate cladding layer 7, the electric field is concentratedly applied to the lower waveguide 1 which is the i-layer. Here, the lower waveguide 1
Is a structure including MQW (multiple quantum well structure), this reverse electric field causes QCSE (quantum confined Stark effect) and changes the absorptance of the lower waveguide 1. Therefore,
At the same time, the refractive index changes, as expressed by the Kramers-Kronig relationship.

As a result, 1 which propagates mainly through the lower waveguide 1
The propagation constant β ₁ of the next mode changes, and the value of the wavelength λ that satisfies the above expression (1) changes (that is, the wavelength change width is not directly determined by the refractive index change width).

The wavelength λ satisfying the expression (1) is converted into the 0th-order mode and is transferred to the upper waveguide 2. wavelengths other than λ,
That is, the light of the non-selected wavelength continues to propagate through the lower waveguide 1, but since the lower waveguide 1 has increased absorption due to QCSE, it undergoes considerable attenuation.

However, according to this structure, the electrode of the n-type intermediate cladding layer 7 and the electrode 1 on the upper p-type contact layer 6 are formed.
If a forward electric field, that is, a positive voltage is applied to the p-side electrode 10 during 0, carriers are injected into the upper waveguide 2 that is the i-layer. As a result, the upper waveguide 2 obtains an optical gain at a desired wavelength. Therefore, the light of the selected wavelength can be transferred to the upper waveguide 2 and undergo optical amplification, and the attenuation received in the lower waveguide 1 can be compensated to maintain a constant transmission gain even when the wavelength is varied. ..

In the present embodiment, a wavelength tunable filter is constructed which transfers a signal of an arbitrary wavelength from the wavelength-multiplexed optical signals of a plurality of wavelengths to the upper waveguide 2 and outputs it. A wavelength filter that uses an asymmetric directional coupler and a grating has a filter bandwidth that uses mode conversion by a grating, as compared with a wavelength filter that uses mode dispersion of a waveguide that constitutes the asymmetric directional coupler. A narrow filter characteristic can be obtained.

This filter was manufactured by forming a p-GaAs buffer layer (included in 11), a p-Al _0.5 Ga _0.5 As clad layer 8 and an undoped MQW (GaAs / Al on a p ⁺ -GaAs substrate 11 in this order. Lower waveguide 1 (0.2 μm thickness) made of _0.4 Ga _0.6 As), n-Al _0.5 G
a _0.5 As intermediate cladding layer 7 (0.7 μm thick), upper waveguide 2 (0.25 μm thick) made of undoped MQW (GaAs / Al _0.2 Ga _0.8 As), p-Al
_0.2 Ga _0.8 As Grating layer 3 (0.1 μm thickness)
Are grown by the MBE (molecular beam epitaxy) method. Then, a grating is formed by resist patterning, and a corrugated grating having a period of 9 μm is formed on the grating layer 3 by RIBE (Reactive Ion Beam Etching).

After removing the resist, the p-Al _0.5 Ga _0.5 As cladding layer 5 and p ⁺ are formed by LPE (liquid phase epitaxy) or MOCVD (metal organic chemical vapor deposition).
-The GaAs contact layer 6 was grown.

Subsequently, a stripe pattern is formed in a direction orthogonal to the grating 3 with a resist, and the wafer is etched in stripes up to the n-AlGaAs intermediate cladding layer 7 with a sulfuric acid-based etchant, and then n-AlG.
An n-GaAs contact layer (not shown) was grown on the aAs intermediate cladding layer 7. AuGe / Au 9 and 10 were vapor-deposited on the back surface of the substrate 11 and the p-contact layer 6, and AuCr / Au (not shown) was vapor-deposited on the n-contact layer on the intermediate cladding layer 7, and alloying was performed.

For the filter manufactured as described above, p-
When a reverse voltage is applied between the electrode 9 and the n-electrode on the intermediate cladding layer 7, an electric field is applied to the lower waveguide 1 made of undoped MQW, and the refractive index changes due to QCSE.

In FIG. 2, the voltages are 1.0 V, 1.5 V and 2.0.
An example in which V is applied shows the coupling efficiency from the lower waveguide 1 to the upper waveguide 2. It can be seen that the transmission band of the bandpass filter shifts to the short wavelength side due to the change in the applied voltage.

As described above, since the transmission band shifts according to the applied voltage of the lower waveguide 1 and the transmittance changes at the same time, in the present invention, the upper waveguide 2 changes according to the applied voltage of the lower waveguide 1. It is characterized in that the amount of current injected into the device is controlled to stabilize the transmission gain of the selected wavelength. As a result, wavelength tunable characteristics with stable transmission gain were obtained as shown in FIG. This allows filtering of channels of one wavelength from the multiplexed channels.

Of course, depending on the specifications required for the filter, there may be cases where the variation in transmittance as shown in FIG. However, according to the present invention, since the transmission gain is always stable, it is possible to provide a filter that can satisfy the specifications in a wide range.

It is also possible to use the waveguide as a bulk layer and to cause a shift in the transmission band in the filter characteristic by causing a refractive index change by the Franz-Keldysh effect by applying a reverse voltage. However, the element length, the grating period, the depth, etc. need to be set appropriately.

In the above example, the GaAs series is explained, but of course other semiconductor materials such as InP / InGaAsP may be used.

Example 2 The optical semiconductor device according to the present invention can control the transmittance at a specific wavelength and may be configured as an optical modulator.

In this embodiment, since a directional coupler having a sharp wavelength selectivity by adding the grating 3 is used,
A slight change in the transmission band causes a large change in coupling efficiency. That is, as shown in FIG. 4, it is possible to obtain ON / OFF of the light with a difference of only about 1 V with respect to the modulated light having the specific wavelength λ _S.

According to the optical modulator of the present invention, the lower waveguide 1
By controlling the refractive index of (1), the transmission wavelength band (coupling wavelength band) is shifted to modulate the modulated light, and at the same time, by controlling the gain of the upper waveguide 1, the modulated light can also be modulated. Therefore, an optical modulator having an extremely high extinction ratio and a deep modulation degree can be obtained regardless of the wavelength of the modulated light.

The optical modulator of this embodiment is the same as that of the first embodiment. However, the lower waveguide 1 is made of undoped MQW, and the energy gap between electrons and heavy holes is 1.50.
eV, that is, the wavelength is 824 nm, and the absorptance greatly changes with respect to the wavelength 835 nm of the modulated light, but at the same time, the refractive index also changes greatly.

The upper waveguide 2 is made of undoped Al _0.06 Ga.
_It is composed of _0.94 As, the peak wavelength of the optical gain is around 835 nm, and a sufficient optical gain can be obtained with a low injection current amount.

In this embodiment described above, 2 GHz is used as the input.
When the output intensity of the CW input light (continuous light) to which the NRZ signal was applied was observed, an extremely good following waveform was obtained.

Since the optical modulator of the present embodiment can compensate the scattering loss due to the propagation of the waveguide and the absorption loss at the time of controlling the refractive index, the insertion loss can be eliminated. Furthermore, by controlling the transmission band, the transmitted light intensity can be made extremely low,
Moreover, the intensity of the transmitted light can be sufficiently increased by independently controlling the optical gain. Therefore, the extinction ratio of the modulated light can be easily increased to 30 dB or more. Further, since the setting wavelength is free, it is possible to provide an optical modulator effective over a wide wavelength range.

In the above-mentioned first and second embodiments, the refractive index control is obtained by QCSE by applying a reverse voltage. However, it is needless to say that the Franz-Keldysh effect in a bulk crystal is utilized or carrier injection is performed by applying a forward voltage. A change in the refractive index due to a change in the amount of carriers such as a plasma effect and a band filling effect may be used.

According to the principle of the present invention, a sufficient degree of modulation can be obtained through the relationship of the above equation (1) by a slight change in the refractive index (β ₀ (λ), β _{1 in the} equation (1). (Λ of (λ) changes), and the heat generation incidental to the carrier injection can be suppressed.

Embodiment 3 FIG. 5 shows a third embodiment of the optical modulator. In the third embodiment, the grating is formed not in the upper waveguide as in the above embodiment but in the middle portion of the upper and lower waveguides. This example is manufactured as follows.

The n ⁺ -GaAs substrate 51 is formed by the MBE method.
On top, an n-GaAs buffer layer (not shown) (n = 2 × 1)
0 ¹⁸ cm ⁻³ ) with a thickness of 0.5 μm and n-Al _0.5 Ga _0.5 A
s clad layer (n = 1 × 10 ¹⁷ cm ⁻³ ) 52 with a thickness of 1.5 μm
m-thick, undoped MQW (GaAs / Al _0.4 Ga _0.6
As) The lower waveguide layer 53 has a thickness of 0.1 μm and is made of p-Al.
_0.5 Ga _0.5 As intermediate cladding layer (p = 1 × 10 ¹⁷ c
m ^-3 ) 54 with a thickness of 0.4 μm and GaAs (thickness 50Å,
Well) / Al _0.5 Ga _0.5 As (100 Å thickness, barrier)
The p-MQW layer ^{(p = 1 × 10 17 m} -3) 55 made of grown 0.1μm thick.

Then, by the photolithography method,
The MQW layer 55 is formed into a grating with a period of 8 μm,
Again, p-Al _0.5 Ga _0.5 As intermediate cladding layer (p-1 ×
10 ¹⁷ cm ⁻³ ) 56 with a thickness of 0.4 μm and undoped Al
_0.06 Ga _0.94 As upper waveguide layer (non-doped) 57 with a thickness of 0.3 μm, n-Al _0.5 Ga _0.5 As cladding layer (n = 1 × 10 ¹⁷ cm ⁻³ ) 58 with a thickness of 1.5 μm, and n ⁺
-GaAs contact layer (n = 1 × 10 ¹⁸ cm ^-3 ) 59
Was grown to a thickness of 0.5 μm by MOCVD. Furthermore,
Similar to the process of the first embodiment, the p-electrode was taken out from the p-AlGaAs intermediate cladding layer to form an optical modulator.

In this embodiment, the lower waveguide (i layer) 53 whose refractive index and absorptance are changed by the application of an electric field and the grating 55 which couples the 0th mode and the 1st mode are separated. .. Therefore, even if the distribution of the first-order mode propagating in the lower waveguide 53 does not reach the upper waveguide 57, the coupling with the 0th-order mode (mode propagating around the upper waveguide 57) in the grating portion 55 is performed. Happens. Therefore, the rate at which the first-order mode propagating through the lower waveguide 53 is absorbed by the upper waveguide 57 can be designed to be lower than that in the previous embodiment.

By the way, the grating formation positions in the above embodiments are merely examples, and the positions may be any positions where both modes propagating centering on both waveguides are coupled. For example, the same effect can be obtained by forming the lower waveguide.

[0045]

As described above, according to the present invention, 2
P-i-n-i-p or n-i-p- with the waveguide as an i-layer so that an electric field is independently applied to the laminated waveguide.
By adopting the in-n structure, insertion loss or propagation loss can be ignored, low voltage driving is possible, and optical modulation suitable for wavelength division multiplexing optical communication, optical switching, optical calculation, etc., which can obtain a wide wavelength tunable range. Optical semiconductor devices such as containers and filters are realized. Also, such devices are suitable for integration with other photodetectors, optoelectronic devices such as lasers.

Further, the wavelength tunable filter can have a short element length with a low crosstalk between wavelengths, and the optical modulator can have a high degree of modulation and a high extinction ratio and a high degree of freedom in the wavelength to be modulated.

[Brief description of drawings]

FIG. 1 is a sectional view of a first embodiment of the present invention.

FIG. 2 is a diagram showing a transmittance with respect to an applied voltage.

FIG. 3 is a diagram showing a transmittance with respect to an applied voltage.

FIG. 4 is a diagram illustrating a second embodiment.

FIG. 5 is a sectional view showing a third embodiment.

FIG. 6 is a diagram showing the principle of an optical semiconductor device of the present invention.

FIG. 7 is a diagram showing the principle of an optical semiconductor device of the present invention.

FIG. 8 is a diagram showing an optical absorption spectrum of a GaAs / AlGaAs multiple quantum well structure.

9A and 9B are diagrams showing a schematic shape of a conventional optical modulator.

[Explanation of symbols]

 1, 53 ... Lower waveguide 2, 57 ... Upper waveguide 3, 55 ... Grating 5, 7, 8, 52, 54, 56, 58 ... Clad layer 6, 59 ... Contact layer 11, 51 ... Substrate 9, 10, 60, 61 ... Electrode

Claims (6)

[Claims] 1. A directional coupler having two stacked semiconductor waveguides, wherein a layer structure including the two semiconductor waveguides has a p-i-n-i-p structure and a p-i-n-i-p structure. nip
An optical semiconductor device having one of an -i-n structure, and the two-layer semiconductor waveguides are both i-layers. 2. The optical semiconductor device according to claim 1, wherein the two-layer semiconductor waveguides are waveguides having at least one of a refractive index and a layer thickness different from each other. 3. The optical semiconductor device according to claim 2, wherein a grating is formed in a region where waveguide modes each having a central intensity are mutually coupled in the two-layer waveguides. 4. The light according to claim 1, wherein each of the two-layer semiconductor waveguides formed by the i-layer is configured to be able to independently apply an electric field from the outside through a pin structure. Semiconductor device. 5. The optical semiconductor element according to claim 1, wherein the optical semiconductor element is configured as a wavelength tunable filter. 6. The optical semiconductor element according to claim 1, wherein the optical semiconductor element is configured as an optical modulator.