JPH03174503A - Optical wavelength filter and device using the same - Google Patents

Abstract

PURPOSE:To suppress a side robe by comprising one cycle of a grating of a chevron part that is a high refractive index area and a valley part that is a low refractive index area, and changing a ratio of chevron part to valley part in one cycle. CONSTITUTION:One cycle of the grating formed in either first and second waveguides 4, 6 or their peripheral areas is comprised of the chevron part that is the high refractive index area and the valley part that is the low refractive index area, and the ratio of chevron part to valley part is changed. It is enough to change the ratio so that the intensity of photocoupling between two waveguides can be changed in the progressive direction of light. Thereby, it is possible to suppress the side robe, and to improve the prevention of crosstalk between signal optical wavelength, and to increase the number of channels or wavelength multiplicity storable in a specific band.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to an optical wavelength filter for demultiplexing or multiplexing light waves in a wavelength multiplexing optical information transmission system, or a photodetector using the same. The present invention relates to optical amplifiers and the like, and particularly to optical wavelength filters using directional couplers and equipped with gratings in which the ratio of peaks to valleys changes, and photodetectors and optical amplifiers using the same.

[Prior Art] An optical wavelength filter using a conventional directional coupler is, for example, as described in R. C. Alfernes S et al.: Applied
Physics I, etters, 33. P16
1 (1978), JP-A No. 61-250607, or Miki et al.
Two leading waveguides were formed on the same substrate.

FIG. 21 is a diagram showing the configuration of such a conventional optical wavelength filter. In the figure, two waveguides 191, 1
92 are formed with different line widths or heights W5, Ws, refractive indexes nt, 02, etc. as shown in the figure, so that the wavelength and propagation constant of the guided light propagating through the respective leading waveguides 191 and 192 are different. The dispersion relationship is also different. At this time, the propagation constants of both groove waveguides match for guided light in a constant wavelength range,
Optical coupling occurs between the two waveguides 191 and 192. In other words, only light in this specific wavelength range is selected and the waveguide 1
Optical power can be transferred between 91 and 192.

In addition, a grating that compensates for the difference in propagation constant between the two waveguides is formed in one of the regions where the optical electric field distribution of the two guided lights exists, and as in the conventional example above, only light in a specific wavelength range can be used. An example is also known in which optical power is transferred between waveguides by selecting .

In this way, conventionally, in order to perform multiplexing or demultiplexing between a signal light and a light wave of a specific wavelength by utilizing the optical power transition,
A light wavelength filter was used.

[Problems to be Solved by the Invention] However, with the filter characteristics obtained in the conventional example described above, as shown in FIG. was formed. The existence of side lobes causes crosstalk in wavelength-multiplexed optical signals, and the need to maintain sufficient wavelength spacing to avoid crosstalk, reducing the number of channels that can be communicated. This was a factor that hindered performance improvement.

As a countermeasure, for example, R, C, Alferness
Others: IEEE Journal10f Quantu
m Electronics%QE-14, No. 1
1. P, 843 (1978) (see Figure 23), by tapering the spacing between the waveguides 193 and t94 constituting the directional coupler, the strength of optical coupling can be adjusted in the direction of light propagation. Proposals have been made to suppress the above-mentioned sidelobes by varying the angle along the curve.

However, in this method, as shown in FIG.
3 is difficult to form into a curved shape, and when a laminated type directional coupler is used, the difficulty of manufacturing increases further.

SUMMARY OF THE INVENTION In view of the above-mentioned problems, an object of the present invention is to provide an optical wavelength filter that sufficiently suppresses the above-mentioned side lobes and is relatively simple to manufacture, and an apparatus using the same.

[Means for Solving the Problems] In the present invention that achieves the above object, in an optical wavelength filter that selects the wavelength of a specific light wave from wavelength-multiplexed light, the wavelength-multiplexed light enters and propagates. The device includes a first waveguide, a second waveguide through which the specific light wave propagates, and a grating formed in either of the first and second waveguides and their surrounding areas.

One period of the grating is composed of a peak portion that is a high refractive index region and a valley portion that is a low refractive index region, and the proportion of the peak portion and the valley portion in one period changes.

The way the above ratio changes depends on the configuration or design.
Various configurations are possible, and the point is that, similar to the principle of sidelobe suppression in the conventional example shown in FIG. All you have to do is change. Furthermore, the pitch of the grating does not necessarily need to be constant over the entire grating area.

In the optical wavelength filter having the above configuration, if a photodetector section that performs photoelectric conversion of the light propagating in the second waveguide is added, a photodetector that demultiplexes and detects a specific light wave from the wavelength-multiplexed optical signal is added. If the second waveguide is a light absorption layer that generates carriers according to the intensity of light when it is incident there, this absorption layer will select only a specific light wave from the input light. The second waveguide is an active layer where optical amplification is performed, and the grating is arranged in the same plane parallel to the active layer and the first waveguide at a distance. An optical amplifier using a semiconductor laser structure is constructed by providing a pair of gratings, each having the same selected wavelength, and forming an electrode so that an injection current flows between the pair of gratings. Example] FIG. 1 is a diagram showing the configuration of a first embodiment of the present invention.

First, the structure of this embodiment will be explained.

In this example, a 0.5 μm thick substrate is placed on a GaAs substrate l.
buffer layer 2 of GaAs with a thickness of 1.5 μm;
A cladding layer 3 of l a, s Gao, s As,
GaAs and A l O, 4G a. A multi-quantum well (MQW) with a thickness of 0.1 μm formed by alternately stacking aAs and
m first waveguide 4. A lo,s with a thickness of 0.8 um
Cladding layer 5 which is Gao, sAs, GaAS and AI
.

A second waveguide 6 with a thickness of 0.4 μm in which zGaa and aAs are alternately laminated to form an MQW is grown in order by molecular beam epitaxial (MBE) method, and then photolithography using a photoresist is performed. After fabricating a resist mask by the method, reactive ion beam etching (R
TBE), a part of the upper surface of the second waveguide 6 has a depth of 0.0
A grating 7 made of 7 μm collage is formed.

In this embodiment, in order to suppress the sidelobes of the filter spectrum, the coupling coefficient of the grating 7 is changed approximately symmetrically from the central portion toward the light incident side and the light propagation side along the light traveling direction. There is. In other words, as shown in Figure 1, the pattern of the grating 7 is such that the pitch A is constant, and the ratio of peaks and valleys (or lines and spaces) that make up the grating 7 changes along the direction of light travel. . The peaks are areas with a high refractive index n2, and the valleys are areas with a low refractive index n2.
This is part 1.

Subsequently, after forming the grating 7, SiOt 8 is deposited on the grating 7 by a sputtering method. Next, photoresist is applied again, and stripes for lateral confinement of the guided light are patterned. R.I.B.
The wafer is etched using the E method until the GaAs substrate 1 is exposed. At this time, the tips of the stripes are cut obliquely with respect to the direction in which the light travels in order to couple the light only to the first waveguide 4 at the time of light input.

In this way, next, Ala, s Gao, and s As are grown by the liquid phase epitaxial (LPE) method, stripes are embedded, and then 5tO2 is deposited over the entire surface again. As a result, a structure in which two waveguides 4.6 are stacked as shown in FIG. 1 is obtained.

In this way, in the optical wavelength filter of this example, two waveguides (i-th waveguide 4, second waveguide 6) are laminated in the layer direction,
A directional coupler is constituted by these. Since each waveguide 4.6 is formed to have a different thickness and composition, the propagation constant of light guided through each waveguide will be different.

The grating 7 formed in the second waveguide 6 is for selecting the wavelength to be optically coupled, and the selected wavelength range is determined by changes in the pitch and the ratio of peaks and valleys.

FIG. 2 shows the optical electric field distribution in the waveguide mode of this example. The vertical axis shows the electric field intensity distribution of light, and the horizontal axis shows the distance in the stacking direction with respect to the top surface of the second waveguide 6. In this way, the waveguide mode in the waveguide 4.6 of this embodiment is the odd mode 11 established around the first waveguide 4 and the second waveguide 6.
There is an even mode 12 that is established around There is.

The operation of this embodiment will be explained below.

Now, input light 14 wavelength-multiplexed over wavelengths from 0.8 am to 0.86 μm is coupled into the first waveguide 4 as shown in FIG. As mentioned above, the modes established in the two waveguides 4.6 include the even mode 12 and the odd mode 11, but the input light 14 manually applied to the first waveguide 4
propagates as an odd mode 11 centered on the first waveguide 4. At this time, in the region where the grating 7 does not exist, the odd mode 11 and the even mode 12 have different propagation constants, so they are hardly coupled and operate independently. However, in the region where the grating 7 is present, the propagation constant β of the odd mode 11 propagates in a close manner. . If the following relationship holds between the propagation constant βmvan of the even mode 12 and the propagation constant βmvan of the even mode 12, a shift in optical power occurs.

β, v1(eh)-〇. , 6 (yield) = 2π/A...
(1) Here, E is the wavelength of light, and A is the pitch of the grating 7.

When the optical power shift as described above occurs, the guided light of the odd mode 11 coupled with the input light 14 is converted to the guided light of the even mode 12. Therefore.

The input light 14 finally becomes a light wave propagating through the second waveguide 6 and is output as a selected output light 15. Light of other wavelengths is output as is from the first waveguide 4 as non-selected output light 16.

Incidentally, at this time, there is a coupler 11g as a coefficient representing the strength of coupling between the light waves propagating through the two waveguides 4.6. This coupling coefficient g is expressed by the following formula.

g” I even (X) ” A+ (X)
・εaaa (X)dx・ ... (2) Here, Cevens C. aa is even mode 12,
It represents the electric field distribution of light in odd mode 11, and A1 is a component corresponding to the one-character diffracted light of the Fourier coefficient of grating 7. Assuming a rectangular wave grating 7, A+(x) is expressed by the following equation.

・ ・ ・ ・ (3) Here, g = 1 (because the coupling between βeven and β is performed by the first-order diffracted light), and n2 and nl are the refraction of the material constituting the peaks and valleys of the grating 7, respectively. where A is the pitch of the grating 7 and W is the width of the valley.

That is, in this case, the coupling coefficient g depends on the proportion of peaks and valleys in the pitch of the grating 7. Figure 3 shows
7 is a graph showing a change in the coupling coefficient g when the ratio of peaks and valleys of the grating 7 is changed in the directional coupler type optical wavelength filter having the configuration of this embodiment. In this illustrated example, the strength of the bond reaches its maximum when the ratio of peaks to valleys is around 0.75:0.25, and the strength of the bond decreases as the ratio increases or decreases. However, the ratio of peaks and valleys at which the strength of this coupling is maximum varies depending on the configuration of the waveguide, and in this example, the ratio was as described above.

Therefore, in this embodiment, first, in order to perform wavelength filtering using light of 0.83 μrn as center wavelength light, (1)
From the formula, the pitch A is set to 7° and 8 μm, and the ratio of peaks and valleys forming the grating 7 changes as shown in FIG. That is, the ratio of peaks to valleys is set to 0.7:0.3 at the center of the grating region, and the ratio of valleys gradually increases toward the periphery, becoming 0.1:0.9 at both ends.

In this way, by changing the ratio of peaks and valleys from the center of the grating to the periphery, the coupling coefficient, that is, the strength of coupling, is varied in the same manner as in the conventional example shown in FIG. 23.

Here, the bond length or complete bond length will be explained and a specific example of the bond strength distribution will be shown.

The traveling direction of the guided light is expressed as two directions, and the complete coupling length is
If the coupling coefficient that changes in the z direction is G (z), then 1L72 G (z) = π/2 ・ ・ ・ ・ (
4).

Here, the function that distributes G (z) is called the taper relationship, and if we set this as F (z), then G (z) = Go - F (z
)... (5) (Go: constant). However, F (z) is normalized by L. In other words, l""F (z)dz=L...(6) From equations (4), (5), and (6), the complete bond length is determined, and GoL"π/2,', L = π/2 G.

Since Go is 21.5 cm-' in this example, the bond length is L=730 (um).

Therefore, in this example, the bond length is 730 gm, and the taper function F(Z) representing the bond strength distribution is the Hamming function shown in FIG. 4. This Hamming function is F (z) = 1 + 0.852 ・cos ( 2πz/L
).

As a result of the above configuration, in this embodiment, the filter spectrum of the selected output light 15 output from the second waveguide 6 out of the light 14 input to the first waveguide 4 is as shown in FIG.

For comparison, FIG. 6 shows a spectrum of an example in which the ratio of peaks and valleys of the grating is constant and the strength of coupling is constant over the grating region. From a comparison between FIG. 5 and FIG. 6, it can be seen that the configuration of this embodiment can sufficiently suppress side lobes. Looking at the filter spectrum in Figure 5, the full width at half maximum is 55.

The ratio of the transmitted light intensity at the center wavelength of the main lobe to the transmitted light intensity at a wavelength 1oOAil away from it is approximately 30.
dB.In addition, in the configuration of Fig. 1,
An anti-reflection coating made of Z r Oa is applied to the input and output end faces for the purpose of suppressing the decrease in efficiency and the generation of ripples due to flil.

FIG. 7 shows a second embodiment. In the second embodiment, in the same manner as in the first embodiment, a layer with a thickness of 0.5
buffer layer 22 of am GaAs, 1.5 μm thick
A 1 of.

s Gao, @ A s cladding layer 23, Ga
A first waveguide layer 24 with a thickness of 0.15 μm made of MQW in which As, A10.4 Gao, and As are alternately laminated.
, 0.9um thick A1. , sGa a, s A S
The cladding layer 25 is made of GaAs and Ala, iGao.
, S As
．． A second waveguide layer 26 with a thickness of 5 μm is grown using the MBE method. Next, using photoresist and the RIBE method, the second waveguide layer 26 is formed into a grating (not shown) with a depth of 0.3 am.
Then cut into A 1 o, s G by LPE method.
ao, s A S 27 is regrown to fill in the valleys of the grating. Next, GaAs is grown as a cap layer 28, and after forming stripes of SiO*, impurities such as Zn (or SL) are doped on both sides of the stripes to perform lateral optical confinement. Diffuse heat. As a result, both sides of the first waveguide polishing 24 and the second waveguide M26 are disordered, and there is a region 29 with a low refractive index.
is formed, and the optical wavelength filter shown in FIG. 7 is obtained.

The guided light in the first waveguide layer 24 and the second waveguide layer 26 is confined in the lateral direction by the low refractive index regions 29 formed on both sides, reducing loss due to diffraction spread of the guided light, resulting in high efficiency. A light wavelength filter is obtained.

For the lateral confinement, various methods other than the methods of the first and second embodiments, such as a method of forming a ridge and a loading method, can be applied.

In the second example, the ratio of peaks (high refractive index portions) and valleys (low refractive index portions) that make up the grating is varied from 0.6:0.4 to l:0 from the center to the periphery. It is increasing.

With this configuration, the change in the coupling coefficient in the grating region has a Blackman function shape as shown in FIG. The Blackman function is F (z) = t + i, 19.
cos (2zz/L)+0.19 ・cos 14x
z/L).

The filter spectrum of the second embodiment is as shown in FIG. 9, and it can be seen that side lobes are suppressed. The full width at half maximum is 73, and the ratio of the transmittance at the center wavelength of the main lobe to the transmittance at a wavelength lOO people away from it is 30~
It is 35dB or more.

Incidentally, in addition to the Taber function of the coupling coefficient in the first and second embodiments, there are the following distributions.

That is, the raised cosine function, F (z)=l+cos (2xz/L)---
(7) Kaiser function, F (z) == [y/s i nh (y)]
r, (y(1-(2z/L)") "")
... Function distributions such as (8) can similarly suppress side lobes.

In the above equation, L is the complete coupling length, Z represents the position along the waveguide, -L/2≦2≦L/2, γ is an arbitrary number, and ■. is a zero-order Bessel function of the first kind.

Further, in the above embodiment, the grating is connected to the second waveguide 6.
26, but the position where the grating is made is based on the optical electric field distribution of the guided light (even mode 12, odd mode 1
Any location is acceptable as long as 1) also exists. However, since the coupling efficiency varies accordingly.

It is necessary to adjust the bond length.

In addition, in each of the above examples, G a A s /A
It was constructed from lGaAs-based material.

It is obvious that it is also possible to construct it from other compound semiconductors such as oil-free, InGaAs/InGaP, glass-based materials such as S t Ot / T i O2, optical crystals such as LLNbOx, LiTaOz, BSO, etc. .

FIG. M10 shows the configuration of a third embodiment of the present invention.

This example is realized by integrating a photodetector having wavelength selectivity by using the structure of the optical wavelength filter according to the present invention.

First, the structure of this embodiment will be explained.

On the substrate 31 which is n-GaAS, a 05 μm thick n
- a buffer layer 32 of GaAs, with a thickness of 1.5 μm;
-A cladding layer 33 made of 1o, s Gao, sAs
．． n-A10, s Gao, y As with a thickness of 0.2 μm
The first waveguide layer 34 is n −A l with a thickness of 0.84 m.
A cladding layer 35.o, sGao, sAs.
i-GaAS and i-A 1a, 4Gao, s As
are alternately stacked to form a multi-quantum well with a thickness of 0.4
The second waveguide layer 36 with a thickness of μm is formed by molecular beam epitaxial (M B
E) Grow sequentially using the method. After that, using a photolithography method, a depth of 0.05 μm and a pitch of 7
．． A grating 37 made of a collage with a thickness of 7 um and whose ratio of peaks and valleys changes according to the Kaiser function distribution (formula (8) above) along the direction of light propagation is placed on the upper surface of the 24th wave layer 36 with a length of 1. It is formed over a length of .277 mm.

Next, a cladding layer 38 of i A 1 o, s Gao, s A S and a cap layer 39 of i-GaAS with a thickness of 0.5 gm (1) are formed on this using a liquid phase epitaxial (LPE) method. regrow. Thereafter, the cladding layer 38 and second waveguide layer 36 in the region adjacent to the grating 37 are removed by etching. Further, in this recessed part, an absorption layer 40 of 1-GaAs with a thickness of 0.1 LLm, p-A1°, s Gao, with a thickness of 1.2 μm are formed.
s A cladding layer 41 with a thickness of 0.5 μm
The cap layer 42, which is ``''-GaAs, is regrown by LPE, and then Cr/GaAs is grown on the cap layer 42.
An electrode 43 made of Au is formed on the back surface of the substrate 31.
An electrode 44 made of uG e /A u is formed.

In this embodiment, out of the light 46 input to the first waveguide layer 34, only the light having the wavelength selected by the optical wavelength filter is coupled to the second waveguide layer 36 and is detected by the photodetector. It is absorbed by a certain absorbent layer 40. The photodetector is p-1-n
A reverse bias is applied between the electrodes 43 and 44. Therefore, carriers generated by absorption are detected as a current signal.

FIG. 11 is a diagram showing the wavelength characteristics of signal light extracted as electricity in this embodiment.

The -3 dB bandwidth, or full width at half maximum, which is a wavelength selection characteristic, is approximately 78 inputs, and the ratio of the output current at the center wavelength of the main lobe to the output current at a wavelength 100 wavelengths away from this (i.e., inter-wavelength crosstalk) is approximately 30dB,
It can be seen that the side lobes are sufficiently suppressed.

FIG. 12 is a diagram showing the configuration of a fourth embodiment of the present invention.

In this embodiment, a buffer layer 52 of n-GaAs with a thickness of 0.56 m is formed on a substrate 51 of n''-GaAs, and n-ALo, s Gao, s A S with a thickness of 1.5 μm are formed on a substrate 51 of n''-GaAs.
The cladding layer 53 is a multiple quantum well (MQW) with a thickness of 0.2 μm, in which n-GaAs with a thickness of 30 μm and AI a, sGaa, and sAS with a thickness of 70 μm are alternately laminated.
), a waveguide layer 54 with a thickness of 0.7 μm n −A 1
cladding layer 55 that is as G no, @ A S
, a light absorption layer 56 made of 1-GaAs having a thickness of 0.4 μm was grown in sequence using the molecular beam epitaxial (MBE) method.

Next, a resist mask is created by a photolithography method using a photoresist, and the upper surface of the light absorption layer 56 is etched using ammonia and hydrogen peroxide solution.
The grating ink 57, which has a depth of 0°2 μm and a pitch of 5.5 μm, and in which the area ratio of peaks and valleys of the grating 57 is a collage of the lens docosine function distribution of equation (7) above, is lengthened along the direction of light propagation. Form each other into LOOum.

Next, p-A1 was
A cladding layer 58 of 0, s Gao, i As is regrown, and a cap layer 59 of p''-GaAS is grown.
grow. Finally, a contact layer (not shown) made of Au-Ge and an electrode 60 made of Au are formed on the back surface of the substrate 51, and a contact layer (not shown) made of Cr and an electrode made of Au are formed on the top surface of the cap layer 59. 61 is formed into a film. A photodetector which is a p-1-n type photodiode was fabricated in this way.The photodetector of this embodiment is constructed as described above, and the waveguide layer 54 laminated in the layer direction. The light absorption layer 56 forms a directional coupler. Since the waveguide layer 54 and the light absorption layer 56 have different compositions and layer thicknesses, the propagation constants of light guided through each layer also differ. The grating 57 formed on the upper surface of the light absorption layer 56 selects light to be directionally coupled by changing the grating pitch and the ratio of peaks to each region.

Next, the operation of this embodiment will be explained.

In this example, a reverse bias is applied between the electrodes 60 and 61, and the wavelength is 0. Signal light 64 consisting of light having a wavelength of 0.78 μm to 0.88 μm in Olum increments is made incident on the waveguide layer 54 using end face coupling. Input coupled signal light 6
4 is a contingency that is established in this photodetector as shown in FIG.
Among the odd modes 12 and 11, the odd mode 11 has a central intensity in the waveguide layer 54 and propagates. The optical field strength distribution 11 of this odd mode 11 hardly extends to the light absorption layer 56 as shown in the second figure, so the propagation loss due to absorption in the absorption layer 56 is extremely small.As mentioned above, at a specific wavelength ( If the relationship of equation 1) is satisfied, the light in the odd mode 11 is converted to the even mode 12, and the center intensity is transferred to the light absorption layer 56. In the case of this example, the grating pitch A is 5.5 μm
, and a wavelength of 0.83 μm is detected. The guided light thus transferred to the light absorption layer 56 is absorbed, generates electrons and a genuine tag, and is detected externally as a photocurrent. FIG. 13 is a diagram showing the wavelength distribution of detected light.

With a full width at half maximum of 64, it can be seen that the sidelobes are sufficiently suppressed and sharp wavelength selection is performed.

In this example, the area of the grating 57 was set at a length of 10 ogm, which did not reach the complete coupling length (the length of the coupling region where the coupling efficiency is maximized) of 262 μm as a directional coupler using the grating 57. However, this is based on consideration of the responsivity of the photodetector, and if the deterioration of the responsivity due to the increase in the light-receiving area is acceptable, the light absorption efficiency can be improved by making the length of the grating region closer to the perfect coupling length. will further increase.

Furthermore, if a plurality of elements according to this embodiment are connected in series by changing the grating pitch and peak-to-valley ratio, an integrated photodetector capable of simultaneously detecting signal light having multiple wavelengths can be created. can do.

FIG. 14 is a diagram showing the configuration of a fifth embodiment of the present invention.

In this embodiment, photodetection is performed using a horizontally formed p-1-n structure.

In the structure of this embodiment, a buffer layer 7i made of 1-GaAs with a thickness of 0.5 μm is placed on a substrate 7i made of semi-insulating GaAs.
2. i-A 1 a, s Ga with a thickness of 1.5 μm
o, s A cladding layer 73, a 1-GaAs layer with a thickness of 50 and Al a, s Gao, +
i A S is alternately laminated MQW thickness 0.2
μm waveguide layer 74, 0.75 μm thick 1-ALo,s
Cladding layer 75 of Gao,s As, thickness 0.3
A light absorption layer 76 made of 1-GaAs with a thickness of μm is sequentially formed. Next, by a process similar to that of the fourth embodiment shown in FIG. 12, a collage-like grating 77 with a tapered coupling strength of 0.05 μm in depth was created on the upper surface of the light absorption layer 76. do. Collagation pitch is 46
The length of the grating region was 2° Oum.

After this, i-A 1 o, s Ga with a thickness of 1.5 μm
A cladding layer 78 made of a, i AS is grown, and a protective layer 79 made of 5iiN4 is further formed.

Next, Zn is deposited on the upper surface of the protective layer 79 at intervals of 2 μm in width.
and Si on both sides to form an n-type region 80. An n-type region 81 is formed.

After that, a cap layer 82 made of p′″-GaAS and an electrode 83 made of Cr/Au are formed on the top of the n-type region 80.
A cap layer 84 made of n''-GaAs and an electrode 85 made of Au-Ge/Au are formed on the n-type region 81.

With a reverse bias applied to the lateral p-1-n structure thus fabricated, the wavelength characteristics of the detected intensity with respect to input light were observed in the same way as in the fourth example. As a result, good wavelength selectivity was obtained as in the fourth example.

Since the structure of this embodiment uses a semi-insulating substrate 71,
Electrical separation from other elements is easy, and it is advantageous when integrating a plurality of photodetectors, or when integrating a detection amplifier, a light emitting element, or a control driver.

FIG. 15 is a diagram showing the configuration of a sixth embodiment of the present invention.

In this embodiment, in addition to the wavelength demultiplexing detection function, an amplification function using a FET III structure is added.

First, the structure will be explained.

In this example, as in the fifth example, a buffer layer 92 of 1-GaAs with a thickness of 0.5 μm is formed on a substrate 91 of semi-insulating GaAs, and an i-A 1 o layer with a thickness of 1.5 gm is formed on a substrate 91 of semi-insulating GaAs.
, s Gaa, s As, and a cladding layer 93 having a thickness of 0.2 μm and having the same structure as the waveguide layer 74 of the fifth embodiment.
waveguide layer 94. t −A lo,s with a thickness of 0.6 μm
A cladding layer 95 made of As is sequentially formed using the MBE method. On this upper surface, a collage-like grating 96 tapered to increase the bonding strength is provided as in the fourth embodiment.
Then, after regrowing a light absorption layer 97 of n-GaAs (doping concentration ~I x 1017 cm-'') with a thickness of 0.4 μm, a layer of Si, N with a thickness of 0.3 μm is formed by sputtering. , an insulating film 98 is formed.

Next, as shown in the figure, a source electrode lO is placed on the light absorption layer 97.
O, create a gate electrode 101 and a drain electrode 102,
It has an FET structure.

The source electrode lOO and the drain electrode 102 are Au-G
e is an Au electrode with an underlying layer, and the gate electrode 10t
is formed by AI.

The operation of this embodiment is similar to that of the previous embodiment, and the light incident on the waveguide layer 94 is mode-converted in the grating 96 region and absorbed by the light absorption layer 97. Carriers generated as a result of absorption are amplified and detected as drain current.

In this embodiment, in addition to the wavelength detection function, an amplification function using the FET structure is added, so that a photodetector with excellent detection sensitivity can be obtained.

Also in this embodiment, the layers on which the grating 96 is formed are the light absorption layers 97. Any region where the waveguide modes (even mode 12 and odd mode 11) established around the waveguide layer 94 overlap may be used.

FIG. 16 is a diagram showing the configuration of the seventh embodiment of the present invention.
Figure 7 is a cross-sectional view taken along line AA'.

In this example, n-
Buffer layer 112 of GaAs, 1.5 um thick n
- First cladding layer 11 that is A10, s Gao, AS
3. Non-doped GaASqA1o, s Gao, s A
A waveguide 114 with a thickness of O, 2 gm and an n-AL formed by alternately stacking s and s to form a multiple quantum well (MQW).

s G ao, s A s second cladding layer 115
, non-doped GaAS and A10.4 Gao,a AS
An active layer 116 having a thickness of 0.4 .mu.m and having a thickness of 0.4 .mu.m is grown in order by forming an MQW layer by alternately stacking the active layers. Although the crystal growth method used up to this point has been the metal organic chemical vapor deposition method (MO-CVD method), a molecular beam epitaxial method (MBE method) may also be used. After forming the active layer 116, two gratings 117 and 118 having the same selected wavelength suitable for the wavelength of the signal light to be optically amplified are formed on the upper surface thereof by photolithography at intervals. Third cladding layer 119 of p-L10, s Gao, s As with a thickness of 1.5 um, p"-GaA with a thickness of 0.2 um
A cap layer 120° and an insulating layer 121 are formed in this order, and a p-type electrode 122 is formed on the upper surface of the cap layer 120 in a portion corresponding to between gratings 117 and 118, and an n-type electrode 122 is formed on the back surface of the substrate 111. An electrode 123 is provided. This third cladding layer 119 and cap layer 120
Although the liquid phase epitaxial method (LPE method) was used for the production, the 1M0-CVD method may also be used.

Next, in order to make the waveguide 114 a three-dimensional structure, both sides of the waveguide 114 are removed by wet etching down to the first cladding layer 113, as shown in FIG. o,s G ao,s
A buried layer 125 of As and n-A 1 o, a G
A buried layer 126 made of ao, s As is grown by the LPE method to form a buried structure.

As described above, the two-layer waveguide (waveguide 114° active! 116
) in this example.

Manually coupled light 128 in which a plurality of laser beams with wavelengths from 0.8 gm to 0.86 gm in steps of 0.001 μm are superimposed is manually coupled to the waveguide 114 . Here, in this embodiment, 0.
Since the center wavelength of light is 83 um and wavelength filtering is performed, from equation (1), A = 9 μm and the complete coupling length is (6
) is set to 250 μm, and the ratio of the gI region between peaks and valleys changes along the direction of propagation of light according to the Blackman function distribution shown in FIG. 8. The signal light transferred from the waveguide 114 to the active layer 116 by the grating 117 whose coupling has changed in this manner is amplified during propagation because the active layer 116 below the electrode 122 is a laser amplification section having gain.

The signal light propagates through the active layer 116 and is amplified.
As described above, the light is coupled to the waveguide 114 again by the grating 118 formed on the waveguide 6, and is output from the input/output waveguide 114.

In this way, grating 117 and grating l18
By injecting current only during this period, the region of the active layer 116 other than this section becomes an absorption waveguide, which eliminates unnecessary signal power for the amplified signal light and eliminates signal wavelengths other than those generated from the amplifier. Spontaneous emission light can be removed. Alternatively, the same effect can be obtained by removing the gap between gratings 117 and 118 by etching.

FIG. 18 and FIG. i9 are diagrams showing the configuration of an eighth actual flow example of the present invention.

In this embodiment, in the seventh embodiment, the buried layer 125.12
The lateral light confinement that was performed in the seventh embodiment is achieved by etching both sides of the light confinement.The structure of this embodiment is almost the same as that of the seventh embodiment, The items are given the same number.

In this example, after forming the third cladding layer 119, the second cladding layer 115 (the first cladding layer 113 ) to form a three-dimensional waveguide 116 as shown in the figure.Next, thermal diffusion of p-type impurity, which has the opposite conductivity type to the third cladding layer 119, is performed on the etched end surface. An impurity diffusion layer 131 is formed, and similar thermal diffusion is performed to form an active layer 116 as shown in FIG.
An impurity diffusion layer 132 is formed on the input/output end faces of the impurity diffusion layer 132.

The impurity diffusion layers 131 and 132 aim to disorder the active layer l16 and both ends of the waveguide 114. The reason for this will be explained below. When a three-dimensional waveguide as in this example is formed, since countless interface states exist at both ends of the active layer 116, injected carriers recombine via the interface states, and invalid injected carriers increase. At the same time, the waveguide 1
The signal light transferred from 14 is absorbed.

In this embodiment, the impurity diffusion layer 131 causes the superlattice structure to be disordered in the direction perpendicular to the waveguide light of the active layer 116 and the waveguide 114, and the impurity diffusion layer 132 causes the waveguide light of the active layer 116 to become disordered. Since the superlattice structure is disordered in parallel directions,
At the same time, unnecessary light is no longer input to the active layer 116.
Spontaneous emission light having a wavelength other than the signal wavelength generated in the amplification region can be scattered, and this can be prevented from being emitted to the outside together with the optically amplified signal light.

In this way, in this example, crystal growth can be reduced compared to that in the seventh example, and since the waveguide width is also controlled by the thermal diffusion time, subtle changes can be made. It becomes possible to perform control.

FIG. 20 is a diagram showing the configuration of the embodiment of FIG. 9 of the present invention.

This embodiment differs from the seventh and eighth embodiments in that the active layer 1
46, the substrate 141, the buffer layer 142, the first cladding layer 143, the waveguide 144 and the second cladding layer 145
All of the films were formed using non-doped GaAs and AlGaAs without doping any impurities. After film formation, a grating (not shown) is formed on the active layer 146, and n
The third cladding layer 147 and cap layer 148 are regrown.

After that, a 5i31'J4 film was formed as a diffusion mask, and stripes with a width of about 6 μm were formed by photolithography, and an n-GaAs cap layer 1 was selectively formed using an etchant consisting of aqueous ammonia and hydrogen peroxide.
48 was etched. Next, ZnAs and the sample were vacuum-sealed onto caps H148 on both sides, and heated at 650°C.
Thermal diffusion is performed for 2.5 hours to form an impurity diffusion layer 149. The diffusion front at this time is assumed to reach the first cladding layer 143, and both the waveguide 144 and the active layer 146 are disordered by Zn to form a three-dimensional waveguide. After that, the diffusion mask is removed and the cap layer 14 is removed.
After removing the p-type layer diffused in 8 by etching and forming the p-type electrode 151, an insulating film (SiO□) is formed and a through hole is formed by photolithography to form the n-type electrode 1.
52 is formed.

The device performance of this embodiment is not much different from that of the seventh and eighth embodiments described above. However, in this embodiment, since the layers up to the active layer 146 are non-doped layers, the degree of freedom in device design increases when manufacturing optical integrated circuits and the like.

In this way, in this example, 2
By manufacturing two waveguides (active layer 146 and waveguide 144) in the film-forming direction, the waveguide spacing can be set accurately, and the area ratio of peaks and valleys can be specified along the direction of light propagation. Since a grating whose function distribution is changed is also used, crosstalk between the waveguide 144 into which incident light is input and the active layer 146 is reduced, and crystal growth design and grating design are facilitated. As described above, in the seventh to ninth embodiments,
Although the input/output waveguide and the active layer are made of MQW having a superlattice structure, they may be made of a conventional thin film waveguide. It is also possible to have a structure in which the input/output waveguide is provided above the active layer.

Also, in order to perform optical amplification for multiple wavelengths,
A large number of pairs of gratings having a plurality of different periods and tapered coupling strengths may be provided (each set having a different selected wrinkle length).

[Effects of the Invention] Since the present invention is configured as described above, it produces the effects described below.

First, by suppressing side lobes, it is possible to improve crosstalk prevention between signal light wavelengths and increase the number of channels or wavelength multiplexing that can be accommodated in a specific band. Therefore, the number of information devices that can be transmitted within the same time can be increased.

In addition, this manufacturing process is based on film formation and is suitable for integration, so it is suitable for optical ICs, OEICs (optoelectronic ICs), and optical communications such as optical information transmission, trunk lines, subscriber systems, and optical LAN. , it can be widely applied to optical computing.

Further, according to the present invention according to claims 5 and 6, it is possible to provide a photodetector with a wavelength selection function, and it is possible to realize a photodetector that has excellent wavelength resolution and is suitable for integration. effective. Further, the device according to the present invention is suitable as a photodetector applied to a wavelength division multiplexing communication system, and can be easily integrated with other functional devices.

According to the eighth aspect of the present invention, there is an effect that the signal light of a specific wavelength contained in the input light can be optically amplified with high wavelength selectivity and optical coupling efficiency. Since optical amplification in this case is performed within the same element, there is no need to combine it with other parts, and optical axis adjustment, which was conventionally performed, is no longer necessary. In addition, since the active layer and the waveguide are not formed on the same plane, it is possible to use a crystal growth method that allows control when growing them on the order of one atomic layer, and the thickness, composition, and amount of impurities in each waveguide layer can be controlled. Since the type and type can be controlled, it has extremely high precision and excellent reproducibility, and has the effect of facilitating manufacturing.

In addition to the above-described effects, the ninth aspect of the present invention has the effect of further preventing unnecessary light from being mixed with the amplified signal light.

In the thing according to claim 10, in addition to the above effects,
This has the effect of being able to amplify signal lights of multiple wave numbers.

[Brief explanation of the drawing]

FIG. 1 is a plan view showing the first embodiment of the present invention, and FIG. 2 is a top view showing the first embodiment of the present invention.
Figure 3 is a diagram showing the waveguide mode optical electric field distribution of the example. Figure 3 is an example diagram showing the ratio of the valley width to the pitch of the grating and the function of the perfect coupling length or coupling coefficient.
FIG. 5 is a diagram showing the spatial variation of the coupling coefficient of the grating of the example. FIG. 5 is a diagram showing the filter spectrum of the first example. FIG. 6 is a diagram showing the filter spectrum of the conventional grating for comparison. 8 is a diagram showing the spatial variation of the coupling coefficient of the grating in the second embodiment. FIG. 9 is a diagram showing the filter spectrum of the second embodiment. The figure is a side view showing the third embodiment, FIG. 11 is a diagram showing the filter spectrum of the third embodiment, FIG. 12 is a side view showing the fourth embodiment, and FIG.
The figure shows the filter spectrum of the fourth embodiment.
4 shows the fifth embodiment, FIG. 15 shows the sixth embodiment, FIG. 16 shows a side view of the seventh embodiment, and FIG. 17 shows the seventh embodiment. A-A' line sectional view, 1st
8 is a front view showing the eighth embodiment, FIG. 19 is a side view showing the eighth embodiment, FIG. 20 is a front view showing the ninth embodiment, FIG. 21 is an explanatory diagram of the conventional example, and FIG. The figure shows a filter spectrum of a conventional example, and FIG. 23 is an explanatory diagram of another conventional example. 1.21.31.51.71.91% 111141・
... Board, 4.24.34.54.74.94.1
14.144...First waveguide, 6.26.36.
...Second waveguide, 7.27.37.57.77.9
7゜1 17゜ grating, 40. 56. 76. 97 ・・Light absorption layer. 16, 46 ・Active layer

Claims (1)

[Claims] 1. In an optical wavelength filter that selects the wavelength of a specific light wave from wavelength-multiplexed light, a first waveguide into which the wavelength-multiplexed light enters and propagates; The grating includes a second waveguide for propagation, and a grating formed in either the first or second waveguide or a peripheral area thereof, and the grating has a peak portion that is a high refractive index region and a low refractive index region. 1. An optical wavelength filter characterized in that one period is composed of a valley portion which is a ratio region, and the ratio of the peak portion and the valley portion to one period is changed. 2. The optical wavelength filter according to claim 1, wherein the ratio of the peak portions to the valley portions in one period changes approximately symmetrically from the central portion of the grating toward the light incidence side and the light propagation side. 3. The optical wavelength filter according to claim 1, wherein the first and second waveguides are formed by stacking semiconductor layers. 4. The optical wavelength filter according to claim 1, wherein the first and second waveguides have different refractive indexes, thicknesses, etc. so that the waveguide modes centered on each waveguide are different. 5. An optical signal comprising the optical wavelength filter according to claim 1, 2, 3, or 4 and a photodetector that performs photoelectric conversion of the light propagating in the second waveguide, and that is wavelength-multiplexed by the photodetector. A photodetector characterized by separating and detecting a specific light wave from within. 6. The optical wavelength filter according to claim 1, 2, 3 or 4, wherein the second waveguide is a light absorption layer that generates carriers according to the intensity of light when it is incident thereon. A photodetector, characterized in that it selectively detects only a specific light wave out of input light wavelength-multiplexed by the light absorption layer. 7. Photodetection according to claim 6, wherein a source electrode, a gate electrode, and a drain electrode are formed on the light absorption layer to provide an FET configuration, and carriers generated in the light absorption layer are amplified and detected as a drain current. vessel. 8. The optical wavelength filter according to claim 1, 2, 3 or 4, wherein the second waveguide is an active layer in which optical amplification is performed, and the grating is connected to the active layer and the first waveguide.
A pair of gratings are provided at a distance in the same plane parallel to the waveguide, each having the same selection wavelength, and further provided with an electrode formed so that an injection current flows between the pair of gratings. An optical amplifier using a semiconductor laser structure characterized by the following. 9. Claim 8, wherein the terminal end of the active layer is disordered.
The optical amplifier described. 10. The optical amplifier according to claim 8, wherein a plurality of sets of the pair of gratings are formed, and each set has a different selection wavelength.