JP2006323137A - Optical combined/branch circuit and optical multiplexing/demultiplexing circuit - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an optical combined/branch circuit that can be formed without a pattern conversion error and makes reflected light less likely to occur. <P>SOLUTION: The optical combined/branch circuit 100 includes: a three-dimensional waveguide having an input waveguide 101 or an output waveguide 103; and two-dimensional propagation area 102 formed wide enough so as not to sideways enclose light signals made incident from the input waveguide 101. Grooves 105 and 106 for enclosing the light signals are formed only on both sides of the input waveguide 101 and the output waveguide 103. The widths of the grooves 105 and 106 are made almost constant and equal to or greater than the depth thereof. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an optical add / drop circuit and an optical add / drop circuit used in an optical integrated circuit.

  In order to realize an optical integrated circuit, it is necessary to realize an optical merging / branching circuit that is easy to manufacture and has low loss. In recent years, a star-type coupler has been proposed as a structure of an optical merging / branching circuit that is easy to manufacture and has low loss (see FIG. 7 of Patent Document 1 and Non-Patent Document 1).

  FIG. 7 shows an example of the structure of a conventional optical merging / branching circuit, so-called star coupler. As shown in this figure, the optical merge branch circuit 700 is formed between one input waveguide 701, six output waveguides 703a to 703f, and the input waveguide 701 and the output waveguide 703. This is a one-input six-output semiconductor optical merging / branching circuit having a two-dimensional propagation region 702.

The operating principle of this optical merging / branching circuit 700 is briefly as follows. An optical signal that has entered the two-dimensional propagation region 702 having a sufficiently wide width from the input waveguide 701 propagates while diffusing into the two-dimensional propagation region 702. The divergence angle θ at this time is expressed as follows according to the spot size at the time of incidence on the two-dimensional propagation region 702, in other words, the waveguide mode spot size w ₀ of the input waveguide 701.
θ = tan ⁻¹ (λ / n / π / w ₀ ) (1)

Here, λ is the wavelength of incident light in vacuum, and n is the refractive index of the medium. Now, assuming that the incident light propagates through the two-dimensional propagation region 702 by a distance z, the spread of the beam (incident light) at that time is as follows from the equation (1): w (z) = 2z · tan θ = 2 · z · λ / n / π / w ₀ (2)
It becomes.

  If the incident electric field has a Gaussian shape, the optical signal spreads while maintaining the Gaussian shape, and this w (z) indicates the spot size of the Gaussian beam when propagating by the distance z. In order for the region 702 to act as a two-dimensional propagation region, the width of the two-dimensional propagation region is such that incident light does not feel the side wall of the waveguide or does not feel lateral confinement in the two-dimensional propagation region. Since it is necessary to be wide, the width is required to be sufficiently wider than w (z) expressed by the equation (2) at a distance z from the incident end.

Here, if the length of the two-dimensional propagation region 702 is L, the beam spread at the end of the two-dimensional propagation region 702 is 2L · λ / n / π / w _{0 according} to the equation (2). Therefore, if the output waveguide 703 is arranged in the range of ± L · λ / n / π / w ₀ from the center of the two-dimensional propagation region 702, the input signal is distributed to the six output waveguides 703a to 703f and branched. Operates as a circuit. As described above, the optical merging / branching circuit of this configuration is called a star coupler because it can be branched into a plurality of waveguides in a star shape at a time via the two-dimensional propagation region 702. On the other hand, when an optical signal is incident from any one of the output waveguides 703a to 703f, the operation is the reverse process of the six-branch operation, and the incident light is coupled to the input waveguide 701 by the ray reverse theorem, and the six inputs Operates as a 1-output coupler.

  As shown in FIG. 8, the input waveguide 701 includes an InP substrate 801 constituting a lower cladding, an InGaAsP core layer 802 formed on the InP substrate 801, and an InP upper portion formed on the InGaAsP core layer 802. A clad 803. In the optical converging / branching circuit 700 having such a structure, the optical confinement in the lateral direction of the waveguide is performed by the refractive index difference between the medium constituting the waveguide core 802 and the clad 801 and 803 and air. Since both sides of the mask can be fabricated by etching deeper than the core 802 of the waveguide, and it is not necessary to strictly control the etching depth like a directional coupler, it is optimal for an optical integrated circuit.

  On the other hand, an example of a conventional 4 × 4 semiconductor array diffraction grating is shown in FIG. As shown in this figure, the semiconductor array diffraction grating 1000 is connected to four input waveguides 1001a to 1001d, a two-dimensional propagation region 1002 connected to these input waveguides 1001a to 1001d, and a two-dimensional propagation region 1002. The seven delay waveguide arrays 1003, a two-dimensional propagation region 1004 connected to the delay waveguide array 1003, and four output waveguides 1005a to 1005d connected to the two-dimensional propagation region 1004 Have. Here, reference numerals 1006 and 1007 are star coupler type optical merging / branching circuits having four inputs and seven outputs and seven inputs and four outputs.

Japanese Patent Laid-Open No. 05-241033 (see FIG. 7) DELeaird, AMWeiner, T.Saida, A.Sugita, K.Okamoto, `` Temporal Response of an Excitation Engineered 1x16 Splitter '', Lasers and Electro-Optics Society, 2003.LEOS 2003.The 16th Annual Meeting of the IEEE, Volume : 2, 2003

  9 is a cross-sectional view taken along arrow IX-IX in FIG. In this figure, reference numeral 901 indicates an InP substrate constituting the lower cladding, reference numeral 902 indicates an InGaAsP core layer, reference numeral 903 indicates an InP upper cladding, and reference numerals 904a and 904b indicate terminations of a two-dimensional propagation region. Yes. As shown in FIGS. 7 and 9, the end portion 704 of the two-dimensional propagation region in the optical merge branch circuit 700 is etched deeper than the waveguide core layer 902 to form an etching mirror. In the case of a semiconductor waveguide, it is known that the reflectance at the semiconductor-air interface is about 30%.

As described in the operation principle, the spread of light at the end portion 704 of the two-dimensional propagation region in the optical merging / branching circuit 700 extends from the center of the two-dimensional propagation region 702 to a range of ± L · λ / π / w _0. . Therefore, the optical signal that cannot be coupled to the output waveguides 703a to 703f at the terminal portion 704 of the two-dimensional propagation region existing in this range is reflected at the semiconductor / air interface.

  In addition, as described in the operation principle, the intensity distribution of light propagating through the two-dimensional propagation region 702 has a Gaussian shape, and therefore, an output waveguide existing at a position away from the center, for example, the output waveguide 703a or the output In order to couple a sufficient power to the waveguide 703f, it is necessary to make L long and widen the optical signal sufficiently. At this time, the optical signal spreads further to the outside of the output waveguides 703a and 703f, and all the optical signals in this portion are reflected at the semiconductor / air interface 904a existing at the terminal portion 704 of the two-dimensional propagation region 702. End up.

  Further, when an optical signal is incident from any one of the output waveguides 703a to 703f, the incident light is coupled to the input waveguide 701 as described in the operation principle, but only one input waveguide 701 is present. Since it does not exist, most of the power is reflected at the semiconductor / air interface 904b. If an optical amplifier is connected to the input waveguide 701, oscillation may occur due to this reflection, and the optical circuit may not operate normally. In particular, when used as a 6-input 1-output merging circuit, most of the input power is reflected at the semiconductor-air interface 904b as described above, so that optical amplifiers are connected to the input / output waveguides 701 and 703. If is connected, oscillation may occur and the optical circuit may not operate normally.

  Here, a dimensional error in manufacturing the waveguide will be briefly described. When a waveguide structure such as the optical junction branch circuit 700 shown in FIG. 7 is manufactured, the shape of the photomask or reticle used for manufacturing the waveguide etching mask is metal only in the waveguide regions 701, 702, and 703. No metal is formed outside the waveguide region. Therefore, when exposure is performed to produce an etching mask, part of the UV light that has passed through the photoresist and reached the substrate surface is scattered on the substrate surface and emitted in all directions. Since the scattered light sensitizes the photoresist in the shadow portion of the photomask metal, the width of the developed photoresist becomes narrower than the width defined by the photomask or reticle. This is called a pattern conversion error.

  If there is a pattern conversion error during the formation of the photoresist, the width of the resulting waveguide will differ from the design even if the waveguide is etched using the photoresist as a mask. In the structure of the conventional optical merging / branching circuit 700 as described above, the width in which the metal is formed is about 2 μm in the input / output waveguide portions 701 and 703 and about 20 μm in the two-dimensional propagation region 702. The non-waveguide region where the metal is not formed, or the interval between the waveguides reaches several hundred μm. Pattern conversion errors that occur during photoresist exposure are determined by the amount of UV light scattered on the substrate surface.

  Therefore, in the conventional structure, a pattern conversion error occurs due to scattered light from a region that is several to 100 times or more times larger than the shielding portion, which causes a large error. This causes a decrease in device yield, and also causes performance degradation due to a deviation from the design. In an optical integrated circuit, a waveguide enters in a complicated manner. Therefore, a dense region and a sparse region are formed depending on the location, and the pattern density changes depending on the location. This means that the ratio of the area of the waveguide region to the non-waveguide region varies depending on the location, and the pattern conversion error varies depending on the location, as the intensity of the UV scattered light varies depending on the location. Become. As a result, there has been a problem that the waveguide width changes depending on the location, and it becomes difficult to produce an integrated device as designed.

  As described above, the star coupler type optical converging / branching circuit having the conventional structure has a problem that a large amount of power is essentially reflected at the terminal portion of the two-dimensional propagation region. In addition, there has been a drawback that pattern conversion errors are likely to occur when a star coupler type optical merging / branching circuit having such a conventional structure is manufactured.

  The conventional semiconductor array diffraction grating also has the same problems as the conventional optical merging / branching circuit. That is, reflection at the end of the two-dimensional propagation region has been a problem. In particular, when an optical amplifier or the like is integrated, there is a possibility that oscillation occurs in a form including the semiconductor array diffraction grating in the cavity. Further, in the semiconductor array diffraction grating, the delay waveguide arrays are densely arranged near the two-dimensional propagation region and sparsely arranged away from the two-dimensional propagation region. Therefore, the location dependence of the pattern conversion error leads to variations in the delay amount in the delay waveguide array, and the wavelength demultiplexing characteristic due to the phase error of the semiconductor array diffraction grating leads to variations in the delay amount in the delay waveguide array. There has been a problem that wavelength demultiplexing characteristics are deteriorated due to the phase error of the diffraction grating.

  Therefore, the present invention has been made in view of the above problems, and a star coupler type optical combining / branching circuit and optical combining / demultiplexing circuit which can be formed without pattern conversion errors and hardly generate reflected light. The purpose is to provide.

  In the present application, the outline of typical inventions among the disclosed inventions will be briefly described as follows.

The optical converging / branching circuit according to the first aspect of the present invention for solving the above-described problems is formed with a three-dimensional waveguide and a wide width so that an optical signal incident from the three-dimensional waveguide does not feel lateral confinement. An optical merging / branching circuit having a two-dimensional propagation region, wherein grooves for confining an optical signal are formed only on both sides of the three-dimensional waveguide.
Examples of the three-dimensional waveguide include an optical waveguide in which an optical signal is confined in the in-plane direction of the substrate and the direction perpendicular to the substrate surface.
Examples of the two-dimensional waveguide include an optical waveguide in which an optical signal is confined in a direction perpendicular to the substrate surface, but an optical signal is not confined in the in-plane direction of the substrate.

  An optical merging / branching circuit according to a second invention for solving the above-mentioned problem is the optical merging / branching circuit according to the first invention, wherein the three-dimensional waveguide inputs an optical signal to the two-dimensional waveguide. Or an output waveguide that outputs an optical signal propagated through the two-dimensional waveguide.

  An optical merging / branching circuit according to a third invention for solving the above-described problem is the optical merging / branching circuit described in the first or second invention, wherein the groove width of the groove is in the two-dimensional propagation region. It is substantially constant except for the vicinity to be connected, and is equal to or more than the depth of the groove. As a result, it is possible to provide a star coupler type optical merging / branching circuit having no pattern conversion error and less reflection as designed with good reproducibility.

  An optical merging / branching circuit according to a fourth invention that solves the above-described problem is the optical merging / branching circuit described in any one of the first to third inventions, wherein the outermost groove among the grooves is In the vicinity of connecting to the two-dimensional propagation region, it is tapered.

  An optical merging / branching circuit according to a fifth aspect of the present invention that solves the above-described problem is the optical merging / branching circuit described in any of the first to fourth aspects, wherein the end portion of the two-dimensional propagation region is A region sandwiched between a plurality of three-dimensional waveguides is inclined with respect to the propagation direction of the optical signal.

The optical multiplexing / demultiplexing circuit according to the sixth invention for solving the above-described problems is formed with a three-dimensional waveguide and a wide width so that an optical signal incident from the three-dimensional waveguide does not feel lateral confinement. An optical multiplexing / demultiplexing circuit having a two-dimensional propagation region, wherein grooves for confining an optical signal are formed only on both sides of the three-dimensional waveguide.
Examples of the three-dimensional waveguide include an optical waveguide in which an optical signal is confined in the in-plane direction of the substrate and the direction perpendicular to the substrate surface.
Examples of the two-dimensional waveguide include an optical waveguide in which an optical signal is confined in a direction perpendicular to the substrate surface, but an optical signal is not confined in the in-plane direction of the substrate.
As a result, it is possible to provide an optical multiplexing / demultiplexing circuit (array diffraction grating) having no pattern conversion error and less reflection as designed with good reproducibility.

  An optical multiplexing / demultiplexing circuit according to a seventh invention for solving the above-described problem is the optical multiplexing / demultiplexing circuit according to the sixth invention, wherein the three-dimensional waveguide inputs an optical signal to the two-dimensional waveguide. Or an output waveguide that outputs an optical signal propagated through the two-dimensional waveguide, or a delay waveguide array. Thereby, it is possible to provide an optical multiplexing / demultiplexing circuit (array diffraction grating) with less reflection.

  An optical multiplexing / demultiplexing circuit according to an eighth invention for solving the above-described problem is the optical multiplexing / demultiplexing circuit according to the sixth or seventh invention, wherein the groove width of the groove is in the two-dimensional propagation region. It is substantially constant except for the vicinity to be connected, and is equal to or more than the depth of the groove. Thereby, it is possible to provide an optical multiplexing / demultiplexing circuit (array diffraction grating) with less reflection.

  An optical multiplexing / demultiplexing circuit according to a ninth invention for solving the above-described problem is the optical multiplexing / demultiplexing circuit described in any of the sixth to eighth inventions, wherein the outermost groove among the grooves is In the vicinity of connecting to the two-dimensional propagation region, it is tapered. Thereby, it is possible to provide an optical multiplexing / demultiplexing circuit (array diffraction grating) with less reflection.

  An optical multiplexing / demultiplexing circuit according to a tenth aspect of the present invention that solves the above-described problem is the optical multiplexing / demultiplexing circuit according to any one of the sixth to ninth aspects, wherein the end portion of the two-dimensional propagation region is A region sandwiched between a plurality of three-dimensional waveguides is inclined with respect to the light propagation direction.

  An optical multiplexing / demultiplexing circuit according to an eleventh invention for solving the above-mentioned problem is connected to the optical merging / branching circuit described in any of the first to fifth inventions, and to the optical merging / branching circuit, And a delay waveguide composed of a waveguide.

  According to the present invention, it is possible to eliminate the dependence of the waveguide width on the location, and to uniformly form an optical merging / branching circuit without pattern conversion error, and to provide an optical waveguide with less reflection as designed. It becomes. Further, when applied to an optical multiplexing / demultiplexing circuit, not only the reflection is suppressed, but also a phase error generated in the delay waveguide array can be removed, and the chromatic dispersion characteristic can be improved.

  The best mode for carrying out the optical coupling / branching circuit and the optical multiplexing / demultiplexing circuit according to the present invention will be specifically described below on the basis of examples.

  Hereinafter, the optical junction branch circuit according to the first embodiment of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a top view of an optical merging / branching circuit according to a first embodiment of the present invention, and FIG. 2 is a sectional view taken along the line II-II in FIG.

  As shown in FIG. 1, the optical merge branch circuit 100 according to the first embodiment of the present invention is a 1-input 6-output semiconductor star coupler type optical merge branch circuit. In other words, the optical combining branch circuit 100 is wide enough that one input waveguide 101, six output waveguides 103a to 103f, and an optical signal incident from the input waveguide 101 do not feel lateral confinement. And a two-dimensional propagation region 102 formed in a width. The input waveguide 101 is connected to the two-dimensional propagation region 102, and the terminal portion 104 of the two-dimensional propagation region is connected to the output waveguide 103. Only along the side walls of both the input waveguide 101 and the output waveguides 103a to 103f are formed grooves 105 and 106 having substantially constant widths for confining optical signals. The grooves 105 and 106 are formed in a shape inclined with respect to the propagation direction of the optical signal in a region sandwiched between the plurality of output waveguides 103 in the vicinity of the terminal portion 104 of the two-dimensional propagation region. Reference numeral 107 in FIG. 1 denotes a slab region, and here, the slab region 107 and the two-dimensional propagation region are connected. Reference numeral 108 in FIG. 1 indicates the terminal end of the outermost groove of the input / output waveguides 101 and 103. Here, the width of the input waveguide 101 and the output waveguides 103a to 103f is set to 2 μm, the distance between the centers of the adjacent output waveguides is set to 3 μm, and the length of the two-dimensional propagation region 102, that is, the input The distance between the end of the waveguide 101 and the end of the output waveguide 103 was set to 60 μm.

  The operation principle of the optical merging / branching circuit 100 having such a structure is the same as that of the conventional optical merging / branching circuit 700 shown in FIG. At this time, the spot size in the input waveguide 101 is about 0.8 μm, the equivalent refractive index of the two-dimensional propagation region 102 is 3.3, and the incident signal light wavelength (optical signal wavelength) is 1.55 μm. Considering this, the spread of the beam according to the equation (2) is ± 11.2 μm. It can be seen that this value is sufficient to distribute the optical signal to the six waveguides at a pitch of 3 μm.

  As shown in FIG. 2, the optical merging / branching circuit 100 described above includes an InP substrate 201 that constitutes a lower cladding portion, InGaAsP core layers 202 a to 202 c formed on the InP substrate 201, and an upper portion of the InGaAsP core layer 202. InP upper clads 203a to 203c formed on the first and second grooves, and grooves 204a and 204b formed by etching deeper than the InGaAsP core layers 202a to 202c along both side walls of the waveguide. The optical confinement in the lateral direction of the waveguide is performed by the difference in refractive index between the medium constituting the waveguide core 202 and the clads 201 and 203 and air. That is, in the input / output waveguides 101 and 103, the optical signal is confined in the direction in the substrate plane and in the direction perpendicular to the substrate surface. In the two-dimensional propagation region 102, the optical signal is confined in the direction perpendicular to the substrate surface. However, no optical signal is confined in the in-plane direction of the substrate.

  The widths of the grooves 105 and 106 formed on the side wall of the waveguide were set as follows. The optical merge branch circuit 100 is produced by dry etching. This is because, as shown in FIG. 1, since the etching surface faces various directions with respect to the crystal surface, wet etching that depends on the surface orientation cannot be used.

  When dry etching is used, the etching rate and the etching shape depend on the aperture ratio of the etched portion (the ratio of the area of the etched portion or the portion where the etching mask does not exist to the unetched portion or the portion where the etching mask exists). Is generally known to be different. When the groove is etched deeper than the core layer 202b of the waveguide as in this embodiment, the etching does not proceed even if the etching time is extended beyond the depth limited by the width of the groove having a small opening width. Such a phenomenon may occur.

In order to prevent such a phenomenon, the width of the groove is desirably equal to or greater than the etching depth. In general, as a waveguide structure, considering that the thicknesses of the core layers 202a to 202c and the upper cladding layers 203a to 203c are 0.3 μm to 0.5 μm and 1.5 μm to 2.5 μm, respectively, the etching depth Needs to be 3 μm or more. Accordingly, the groove width is desirably 3 μm or more, which is not less than the groove depth. In this embodiment, considering that the groove width is several times the depth, both W ₁ and W ₂ are set to 15 μm.

  In this embodiment, the branching and merging operations are the same as those of the conventional optical merging / branching circuit 700 shown in FIG. 7, but the reflection occurring in the conventional example is suppressed as follows. That is, in this embodiment, as shown in FIG. 1, in the terminal portion 104 of the two-dimensional propagation region, the shape of the terminal portion 108 of the outermost groove of the input / output waveguide is tapered and the two-dimensional propagation region 102 is formed. Since the lateral direction and the outside of the input / output waveguide are connected to the slab region 107 having substantially no lateral confinement or a sufficiently wide lateral width, in the conventional example, at the end of the two-dimensional propagation region, The reflected optical signal can be guided and propagated to the slab region, and reflection that goes back to the input waveguide does not occur.

  This means that even when an optical signal is input from the input waveguide 101 and used as a 1-input 6-output optical branch circuit, the optical signal is input from any of the output waveguides 103a to 103f, and 6-input 1-output The same holds true when used as an optical shunt circuit. Further, as shown in FIG. 1, when an optical signal enters from the input waveguide 101 and is used as a 1-input 6-output optical branch circuit, the terminal portion 104 of the two-dimensional propagation region includes a plurality of output guides. The region sandwiched between the waveguides has a shape that is not perpendicular to the propagation direction of the optical signal. Even if the reflection of the optical signal occurs here, the reflected light is guided to the slab region 107 and the input waveguide 101. This is equivalent to the case where reflection is substantially suppressed.

Further, in this embodiment, the following measures are taken in order to produce the pattern shape for preventing reflection without pattern conversion error. The photomask or reticle used when producing the waveguide pattern of this embodiment has a shape in which metal is formed in portions other than the regions of the grooves 204a and 204b shown in FIG. Whereas the waveguide width W _0, which is formed of metal is set to 2 [mu] m, the width W ₁ and W ₂ of the transparent region is not formed of metal was set to 15 [mu] m, the width W ₀ of the waveguide region The ratio with the groove widths W ₁ and W ₂ is at most about 15 times.

  Therefore, the amount of UV light reaching the substrate surface is also greatly reduced as compared with the conventional optical merging / branching circuit 700 shown in FIG. 7, and the amount of UV light scattered on the substrate surface is also compared with the conventional structure. Then, since it is about 1/10 or less, it is possible to greatly suppress pattern conversion errors caused by scattered light. As a result, it is possible to minimize deterioration of characteristics caused by a deviation from the design.

  In the case of the present embodiment, the width of the transparent region where the metal is not formed is kept substantially constant, so that the waveguide portion where the metal is formed even when the waveguide enters in a complicated manner like an optical integrated circuit. The ratio of the area of the transparent region where no metal is formed does not change greatly. Therefore, the pattern conversion error does not change depending on the location. As a result, the problem of the conventional example that the waveguide width changes depending on the location is solved, and a resist mask as designed can be formed. Therefore, the waveguide width etched using the resist mask can be manufactured as designed, and a high-performance integrated device with less reflection can be manufactured.

  As described above, the input waveguide 101 or the output waveguide 103 is a three-dimensional waveguide, and an optical signal incident from the three-dimensional waveguide is formed to have a wide width so as not to feel lateral confinement. The optical converging / branching circuit 100 having the two-dimensional propagation region 102 is formed with grooves 105 and 106 having substantially constant widths for confining optical signals only on both sides of the three-dimensional waveguide. The grooves 105 and 106 are tapered in the vicinity of being connected to each other, and are inclined with respect to the propagation direction of the optical signal in a region sandwiched between the plurality of output waveguides 103 in the vicinity of the terminal portion 104 of the two-dimensional propagation region. By adopting such a shape, it becomes possible to provide an optical converging / branching circuit with less reflection with good reproducibility.

  In the present embodiment, the 1-input 6-output optical combining / branching circuit 100 has been described. However, the present invention is not limited to such an optical combining / branching circuit, and 1-input N-output (N is an integer) is also used. The same operational effects as the 1-input 6-output optical confluence branch circuit 100 are obtained.

  Hereinafter, an optical merge branch circuit according to the second embodiment of the present invention will be described in detail with reference to the drawings.

  FIG. 3 is a top view of the optical merge branch circuit according to the second embodiment of the present invention.

  As shown in FIG. 3, the optical merge branch circuit 300 according to the second embodiment of the present invention is a 6-input 6-output semiconductor star coupler type optical merge branch circuit. In other words, the optical merging / branching circuit 300 has six input waveguides 301a to 301f, six output waveguides 303a to 303f, and an optical signal incident from the input waveguide 301 that does not feel lateral confinement. And a two-dimensional propagation region 302 having a wide width. The input waveguide 301 is connected to the two-dimensional propagation region 302, and the end portion 304 of the two-dimensional propagation region is connected to the output waveguide 303. Only along the side walls of both the input waveguides 301a to 301f and the output waveguides 303a to 303f, grooves 305a to 305f and 306a to 306f having substantially constant widths for confining optical signals are formed, respectively. The grooves 305a to 305f and 306a to 306f are formed in a shape inclined with respect to the propagation direction of the optical signal in a region sandwiched between the plurality of output waveguides 303 in the vicinity of the terminal portion 304 of the two-dimensional propagation region. A reference numeral 307 in FIG. 3 indicates a slab region. Here, the slab region 307 and the two-dimensional propagation region 302 are connected. Reference numeral 308 in FIG. 3 indicates a terminal portion of the outermost groove of the input / output waveguides 301 and 303. Here, the width of the input waveguides 301a to 301f and the output waveguides 303a to 303f is set to 2 μm, the distance between the centers of the adjacent output waveguides is set to 3 μm, and the length of the two-dimensional propagation region 302 is 100 μm. Set to.

  The operation principle of the optical converging / branching circuit 300 having such a structure is that there are six input waveguides, and the optical signal is input from any one of the input waveguides 301a to 301f. This is the same as the optical merging / branching circuit 700 of FIG. As for the arrangement of the input / output waveguides, the output waveguide 303 is arranged so as to fall within the range of the beam spread in the two-dimensional propagation region given by equation (2) regardless of where the light enters from 301a to 301f. It ’s fine. In this embodiment, the spot size in the input waveguide 301 is about 0.8 μm, the equivalent refractive index of the two-dimensional propagation region 302 is 3.3, and the incident signal light wavelength (optical signal wavelength) is 1.55 μm. Therefore, the beam spread according to the equation (2) is ± 18.7 μm. It can be seen that this value is sufficient to distribute the optical signal to the six waveguides at a pitch of 3 μm.

  The cross-sectional structure of the waveguide of the optical merge branch circuit 300 shown in FIG. 3 in the A-A ′ cross section is the same as that shown in FIG. 2 except that the number of waveguides is six. Also in this example, the groove width was set to 15 μm for the same reason as in the optical merge branch circuit 100 according to the first example.

  In this embodiment, the reflection suppression method for suppressing the reflection of the optical signal is the same as that of the optical merging / branching circuit 100 described in the first embodiment. That is, in this embodiment, as shown in FIG. 3, the shape of the terminal portion 308 of the outermost groove of the input / output waveguide is tapered at the terminal portion 304 of the two-dimensional propagation region, and the two-dimensional propagation region 302 is formed. Since the lateral direction and the outside of the input / output waveguide are connected to the slab region 307 having substantially no lateral confinement or having a sufficiently wide lateral width, the end portion of the two-dimensional propagation region in the conventional example The optical signal reflected by the optical waveguide can be guided and propagated to the slab region, and reflection that goes back to the input waveguide does not occur. Further, as shown in FIG. 3, the end portion 304 of the two-dimensional propagation region is an optical signal propagation in the region sandwiched between the plurality of input waveguides 301 and the region sandwiched between the plurality of output waveguides 303. Since it has a shape that is inclined with respect to the direction (not perpendicular), even if the reflection of the optical signal occurs here, the reflected light is guided to the slab region 307 and does not travel backward through the input waveguide 301, and is substantially reflected. Is equivalent to being suppressed.

  Furthermore, in this embodiment, the pattern conversion error is suppressed for the same reason as in the optical merging / branching circuit 100 according to the first embodiment, so that the waveguide can be manufactured as designed, and the reflection is high. It became possible to fabricate high performance integrated devices.

  As described above, the input waveguide 301 and the three-dimensional waveguide that is the output waveguide 303 and the optical signal incident from the three-dimensional waveguide are formed to have a wide width so as not to feel lateral confinement. The optical converging / branching circuit 300 having the two-dimensional propagation region 302 is formed with grooves 305 and 306 having substantially constant widths for confining optical signals only on both sides of the three-dimensional waveguide. The grooves 305 and 306 are tapered in the vicinity of being connected to each other, and in the region sandwiched between the plurality of input waveguides 301 or the plurality of output waveguides 302 in the vicinity of the end portion 304 of the two-dimensional propagation region, By adopting a shape that is inclined with respect to the signal propagation direction, it is possible to provide the optical converging / branching circuit 300 with less reflection with good reproducibility.

  In this embodiment, the 6-input 6-output optical merge branch circuit 300 has been described. However, the present invention is not limited to such an optical merge branch circuit, and M inputs and N outputs (M and N are integers). The same operational effects as those of the 6-input 6-output optical converging / branching circuit 300 can be obtained.

  As described above, regarding the description of the optical combining branch circuits 100 and 300 according to the first and second embodiments, the waveguide structure has sidewalls on both sides of the optical waveguide region as shown in FIG. 4A rather than the core layer 402a. The case of using a so-called high-mesa waveguide that has been etched deeply has been described. However, if the waveguide structure is such that the cladding thickness on both sides of the optical waveguide region is thinner than the cladding thickness of the optical waveguide region, the same effect is obtained. There is an effect. That is, other examples of the optical waveguide region in the optical merging / branching circuit are shown in FIGS. 4 (b), 4 (c), and 4 (d). In FIG. 4, reference numerals 401a to 401d denote InP substrates which are lower cladding portions, reference numerals 402a to 402d denote InGaAsP core layers, reference numerals 403a to 403d denote upper InP claddings, and reference numerals 404a to 404d denote optical electric fields. Show. The structure shown in FIG. 4B is a so-called ridge-type waveguide in which the clads on both sides of the optical waveguide region are thinned, and the structure shown in FIG. 4C has zero clads on both sides of the optical waveguide region. The structure shown in FIG. 4D is a modification of the ridge-type waveguide in which both sides of the optical waveguide region are cut down to the core layer.

  In all of these structures, as indicated by the alternate long and short dash line in the figure, a part of the light electric field 404a to 404d is above the position of the interface between the medium and the air constituting the waveguides on both sides of the optical waveguide region. Because it exists, you will feel the influence of air. Since the present invention is a structure for suppressing the reflection of the medium / air interface that constitutes the waveguide at the terminal end of the two-dimensional propagation region, the light electric field feels the refractive index of air. In this structure, the same effects are obtained.

In the above-described embodiment, the end portions 104 and 304 of the two-dimensional propagation region are in the region sandwiched between the plurality of input waveguides or the region sandwiched between the plurality of output waveguides with respect to the propagation direction of the optical signal. Inclined (not vertical) shape. That is, the shape inclined with respect to the traveling direction of the optical signal has various shapes as long as it is not a plane perpendicular to the propagation direction λ _a of the optical signal as shown in FIG. It is possible to use. For example, as shown in FIG. 5B, a shape having an inclined plane with respect to the propagation direction λ _b of the optical signal, or a V-shape formed by two planes intersecting as shown in FIG. Or a polygonal shape, or a shape having a curved surface as shown in FIG. In FIG. 5, reference numerals 501a to 501d denote input / output waveguides, reference numerals 502a to 502d denote two-dimensional propagation areas, and reference numerals 503a to 503d denote end portions of the two-dimensional propagation areas.

  Regarding all the above embodiments, there are no particular restrictions on the composition of the core layer and the upper and lower cladding layers used in this configuration, and this configuration is applied to the core layers and cladding layers of the waveguides of all the structures that are normally used. By taking the above, the effects as described above are obtained. In other words, the semiconductor can be applied to any material such as InGaAsP, GaAs, AlGaAs, InGaAs, GaInNAs, Si, and other than the semiconductor, it can be applied to an amorphous material such as quartz glass, an organic material, or the like. Further, regarding the structure of the core layer and the clad layer, the same effects can be obtained regardless of the bulk, the quantum well structure (MQW), the quantum wire, and the quantum dot. Even if, for example, a separate confinement structure (SCH structure) as used in a normal laser is formed between the core layer and the clad layer, the same effect is obtained.

  The optical multiplexing / demultiplexing circuit according to the third embodiment of the present invention will be specifically described below with reference to the drawings.

  FIG. 6 is a top view of an optical multiplexing / demultiplexing circuit according to the third embodiment of the present invention.

  As shown in FIG. 6, an optical multiplexing / demultiplexing circuit (semiconductor array diffraction grating) 600 according to the third embodiment of the present invention is a semiconductor array diffraction grating with 4 inputs and 4 outputs. That is, the optical multiplexing / demultiplexing circuit 600 includes four input waveguides 601a to 601d, seven delay waveguide arrays 603, four output waveguides 605a to 605d, and light incident from the input waveguide 601. A two-dimensional propagation region 602 formed so wide that the signal does not feel lateral confinement, and an optical signal incident from the delay waveguide array 603 formed so wide that it does not feel lateral confinement. And a two-dimensional propagation region 604. The input waveguide 601 is connected to the two-dimensional propagation region 602, and the terminal portion of the two-dimensional propagation region 602 is connected to the delay waveguide array 603. The delay waveguide array 603 is connected to the two-dimensional propagation region 604, and the terminal portion of the two-dimensional propagation region 604 is connected to the output waveguide 605. Grooves 609, 610, and 611 for confining optical signals are formed only along the side walls of both of the input waveguides 601a to 601d, the delay waveguide array 603, and the output waveguides 605a to 605d. Here, in FIG. 6, reference numerals 606 and 607 denote a star coupler type optical merging / branching circuit having four inputs and seven outputs and seven inputs and four outputs. Reference numeral 608 in FIG. 6 indicates a slab region. Here, the slab region 608 and the two-dimensional propagation regions 602 and 604 are connected.

  As the star coupler type optical merging / branching circuits 606 and 607, the optical merging / branching circuits 100 and 300 of the first or second embodiment shown in FIGS. 1 and 3 are used. That is, the input waveguide 601, the output waveguide 605, or the three-dimensional waveguide that is the delay waveguide array 603, and the optical signal incident from the three-dimensional waveguide are formed to have a wide width so as not to feel lateral confinement. The optical multiplexing / demultiplexing circuit (semiconductor array diffraction grating) 600 having the two-dimensional propagation regions 602 and 604 formed on the both sides of only the three-dimensional waveguide has substantially constant width grooves 609, 610, 611 are formed, and the outermost grooves among the grooves 609, 610, 611 are tapered in the vicinity of being connected to the two-dimensional propagation regions 602, 604, and further, in the vicinity of the terminal portions of the two-dimensional propagation regions 602, 604 In the region sandwiched between the plurality of input waveguides 601a to 601d, the plurality of output waveguides 605a to 605d, and the delay waveguide array 603, And a shape inclined to the propagation direction of the signal. Further, the widths of the grooves 609, 610, and 611 were formed to be equal to or greater than the depth of the grooves.

  Thus, for the same reason as described in the first or second embodiment shown in FIGS. 1 and 3, an optical multiplexing / demultiplexing circuit 600 with reduced pattern conversion error and suppressed reflection as designed, so-called It has become possible to provide a semiconductor array diffraction grating.

  Besides the suppression of reflection, the effect of this embodiment is that grooves 609, 611, and 610 having substantially constant widths are formed on both sides of the input waveguide 601, the output waveguide 605, or the delay waveguide array 603. By adopting the configuration, the pattern conversion error at the time of fabrication is reduced as described above, and variations in the pattern conversion error depending on the location are suppressed. As a result, the delay amount of the delay waveguide array 603 was defined as designed, and the phase error was greatly suppressed. Thereby, the transmission characteristics based on the phase error of the optical multiplexing / demultiplexing circuit 600 (semiconductor array diffraction grating), in other words, the chromatic dispersion characteristics are greatly improved.

  Also in the present embodiment, all the explanations in the first and second embodiments are equivalent. That is, as the waveguide structure, a case where a waveguide having a so-called high mesa structure in which the side walls on both sides of the optical waveguide region are etched deeper than the core layer 402a as shown in FIG. 4A has been described. However, if the waveguide structure is such that the clad thickness on both sides of the optical waveguide region is thinner than the cladding thickness of the optical waveguide region, the same effect can be obtained.

  That is, as shown in FIG. 4B, a so-called ridge-type waveguide structure in which the clad on both sides of the optical waveguide region is thinned, and as shown in FIG. 4C, the clad on both sides of the optical waveguide region is zero. A structure in which the ridge-type waveguide is modified, as shown in FIG. 4D, a structure in which the ridge-type waveguide is deformed by cutting down to the core layers on both sides of the optical waveguide region, and the like. Since a part of 404d exists above the position of the interface between the medium constituting the waveguides on both sides of the optical waveguide region and the air, the influence of air is felt. The present invention suppresses the reflection of the medium / air interface that constitutes the waveguide at the end of the two-dimensional propagation region. Thus, in all structures where the electric field of light feels the refractive index of air, The same effects as the embodiment are obtained.

In the above-described embodiment, the terminal portions of the two-dimensional propagation regions 602 and 604 are regions sandwiched between the plurality of input waveguides 601, regions sandwiched between the plurality of output waveguides 605, and a plurality of delay waveguide arrays. In the region surrounded by 603, the shape is inclined with respect to the propagation direction of the optical signal. That is, any shape other than the plane perpendicular to the propagation direction λa of the optical signal as shown in FIG. 5A may be used. For example, the propagation direction of the optical signal as shown in FIG. A shape having an inclined plane with respect to λ _b , a V-shaped shape in which two planes intersect as shown in FIG. 5C, a polygonal shape, or a shape shown in FIG. As shown, it may be a shape having a curved surface.

  Furthermore, with respect to all the above-described embodiments, there are no particular restrictions on the composition of the core layer and the upper and lower cladding layers used in this configuration. By taking this configuration, the above-described effects can be obtained. In other words, the semiconductor can be applied to any material such as InGaAsP, GaAs, AlGaAs, InGaAs, GaInNAs, Si, and other than the semiconductor, it can be applied to an amorphous material such as quartz glass, an organic material, or the like. Further, regarding the structure of the core layer and the clad layer, the same effects can be obtained regardless of the bulk, the quantum well structure (MQW), the quantum wire, and the quantum dot. Even if, for example, a separate confinement structure (SCH structure) as used in a normal laser is formed between the core layer and the clad layer, the same effect is obtained.

  In the above description, the optical coupling / demultiplexing circuit 600 including the optical coupling / branching circuits 602 and 604 and the delay waveguide array 603 has been described. However, instead of the delay waveguide array 603, the optical coupling / branching circuits 602 and 604 are connected. An optical multiplexing / demultiplexing circuit having a delay waveguide composed of a plurality of waveguides may be used, and the same effect as the optical multiplexing / demultiplexing circuit 600 is achieved.

  The present invention can be used for an optical add / drop circuit and an optical add / drop circuit used in an optical integrated circuit.

It is a top view of the optical confluence branch circuit which concerns on 1st Example of this invention. It is II-II arrow sectional drawing in FIG. It is a top view of the optical confluence branch circuit which concerns on 2nd Example of this invention. It is explanatory drawing of the optical waveguide in the optical confluence | merging branch circuit based on the 1st and 2nd Example of this invention. It is explanatory drawing of the termination | terminus part vicinity of the two-dimensional propagation area | region which the optical confluence | merging branch circuit based on the 1st and 2nd Example of this invention has. It is a top view of the optical multiplexing / demultiplexing circuit according to the third example of the present invention. It is a top view of the conventional optical confluence branch circuit. It is VIII-VIII arrow sectional drawing in FIG. It is IX-IX arrow sectional drawing in FIG. It is a top view of a conventional 4 × 4 optical multiplexing / demultiplexing circuit.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Optical confluence | merging branch circuit 101 Input waveguide 102 Two-dimensional propagation area | region 103a-f Output waveguide 104 Termination part 105,106a-f groove | channel 107 Slab area | region 108 Termination part of the outermost groove | channel of an input-output waveguide 201 InP substrates 202a-c InGaAsP core layers 203a-c InP upper clad 204a, b Groove 300 Optical converging branch circuit 301a-f Input waveguide 302 Two-dimensional propagation region 303a-f Output waveguide 304 Termination part 305 of two-dimensional propagation region , 306 Groove 307 Slab region 308 Termination portions 401a to d of the outermost grooves of the input / output waveguides InP substrates 402a to d InGaAsP core layers 403a to d Upper InP claddings 404a to d Light electric fields 501a to d Input / output waveguides 502a ~ D Two-dimensional propagation region 503a-d End portion of two-dimensional propagation region 00 optical multiplexing and demultiplexing circuit 601a~d input waveguide 602 2-dimensional propagation region 603 delays waveguide array 604 two-dimensional propagation region 605a~d output waveguides 606 and 607 light confluent branch circuit 608 slab region 609, 610, 611 groove

Claims (11)

An optical merging / branching circuit having a three-dimensional waveguide and a two-dimensional propagation region formed so wide that an optical signal incident from the three-dimensional waveguide does not feel lateral confinement,
An optical converging / branching circuit, wherein grooves for confining an optical signal are formed only on both sides of the three-dimensional waveguide. An optical merge branch circuit according to claim 1,
The optical merging / branching circuit, wherein the three-dimensional waveguide is an input waveguide that inputs an optical signal to the two-dimensional waveguide or an output waveguide that outputs an optical signal propagated through the two-dimensional waveguide. An optical converging / branching circuit according to claim 1 or 2,
The optical merging / branching circuit characterized in that the groove width of the groove is substantially constant except for the vicinity connected to the two-dimensional propagation region, and is equal to or greater than the depth of the groove. An optical merging / branching circuit according to any one of claims 1 to 3,
The outermost groove of the grooves is formed in a tapered shape in the vicinity of being connected to the two-dimensional propagation region. An optical merging / branching circuit according to any one of claims 1 to 4,
An optical merging / branching circuit, wherein a terminal portion of the two-dimensional propagation region is inclined with respect to a propagation direction of an optical signal in a region sandwiched between a plurality of three-dimensional waveguides. An optical multiplexing / demultiplexing circuit having a three-dimensional waveguide and a two-dimensional propagation region formed so as to have a width such that an optical signal incident from the three-dimensional waveguide does not feel lateral confinement,
An optical multiplexing / demultiplexing circuit, wherein grooves for confining an optical signal are formed only on both sides of the three-dimensional waveguide. An optical multiplexing / demultiplexing circuit according to claim 6,
The three-dimensional waveguide is an input waveguide that inputs an optical signal to the two-dimensional waveguide, an output waveguide that outputs an optical signal propagated through the two-dimensional waveguide, or a delay waveguide array, Optical multiplexing / demultiplexing circuit. An optical multiplexing / demultiplexing circuit according to claim 6 or 7,
An optical multiplexing / demultiplexing circuit characterized in that the groove width of the groove is substantially constant except for the vicinity connected to the two-dimensional propagation region and is equal to or greater than the depth of the groove. An optical multiplexing / demultiplexing circuit according to any one of claims 6 to 8,
An optical multiplexing / demultiplexing circuit, wherein an outermost groove of the grooves is formed in a tapered shape in the vicinity of being connected to the two-dimensional propagation region. An optical multiplexing / demultiplexing circuit according to any one of claims 6 to 9,
An optical multiplexing / demultiplexing circuit, wherein a terminal portion of the two-dimensional propagation region is inclined with respect to a propagation direction of an optical signal in a region sandwiched between a plurality of three-dimensional waveguides. An optical multiplexing / demultiplexing circuit comprising: the optical combining / branching circuit according to any one of claims 1 to 5; and a delay waveguide connected to the optical combining / branching circuit and including a plurality of waveguides. circuit.

Images