JP2007133286A - Wavelength multiplexer/demultiplexer - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a wavelength multiplexer/demultiplexer using a semiconductor material with which the polarization dependency generated due to the error in a manufacturing process or the machining error is compensated without being accompanied by the deterioration in characteristic. <P>SOLUTION: The multiplexer/demultiplexer 100 is composed by providing: an array waveguide lattice which is provided with a plurality of array waveguides 130 having a semiconductor substrate 101, a first semiconductor clad layer 102 provided on the substrate 101, a non-dope semiconductor core layer 103 provided on the clad layer 102, a second semiconductor clad layer 105 provided on the core layer 103, and a half-insulation layer 104 composed of a half-insulated semiconductor and provided between the clad layer 105 and the core layer 103; and electrodes 106 and 107 which apply an electric field upon the array waveguides 130 so as to vary the refractive index for a TE mode without varying the refractive index for a TM mode in the core layer 103. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a wavelength multiplexer / demultiplexer applied to an optical transmission system such as optical communication, optical switching, and optical information processing using an optical wavelength multiplexing system.

  Currently, in order to realize large-capacity communication, research and development of wavelength multiplexing optical communication network technology that multiplexes and transmits a plurality of optical signals having different optical wavelengths (frequencies) on one optical fiber is actively performed. Yes. In a wavelength multiplexing optical communication network, a wavelength multiplexer / demultiplexer that performs multiplexing (multiplexing) and demultiplexing of a plurality of optical signals having different wavelengths (frequencies) has become indispensable.

  A representative example of the wavelength multiplexer / demultiplexer is one using an arrayed waveguide grating (AWG). As shown in FIG. 10, the arrayed waveguide grating includes an input waveguide 910, slab waveguides 920 and 940, an arrayed waveguide 930, an output waveguide 950, and the like (for details, see Non-Patent Document 1 below). Etc.). As such an arrayed waveguide grating, a low-loss and polarization-independent one has been put into practical use with a quartz-based material and is generally widely used as a wavelength multiplexer / demultiplexer 900 or the like.

  Recently, materials using semiconductor materials have been actively researched and developed. This is because when a semiconductor material is used, the relative refractive index difference of the optical waveguide is high, and the waveguide can be bent with a small bending radius, so that it can be very small, and other semiconductor optical devices such as LD and PD. This is because monolithic integration is possible, and there is a great merit that a multifunctional optical circuit that will be required in the future can be realized on a single chip.

Advanced Optical Electronics Series 8, Optical Integrated Device, by Isao Kobayashi, Kyoritsu Publishing Co., Ltd. M. Kohtoku et al., “Polarization-independent InP arrayed waveguide grating filter using deep ridge waveguide structure”, CLEO / Pacific Rim'97, pp.284-285, 1997 M.K.Smit, “PHASAR-based WDM devices: Principles, design and applications”, IEEE Journal of selected topics in quantum electronics, vol.2, No.2, pp.236-250, 1996 J.G.Mendoza-Alvarez, “Analysis of depletion edge translation lightwave modulators”, J.of Lightwave Technology, vol.6, pp.793-808,1988

  In the wavelength multiplexer / demultiplexer, one of important characteristics is that there is no polarization dependence, that is, the filter characteristics do not change between the TE wave and the TM wave. However, when a semiconductor material is used for the wavelength multiplexer / demultiplexer, it becomes difficult to control the polarization dependency, and there is a problem that the yield is lowered. The reason for this will be described below.

In the arrayed waveguide grating, the peak wavelength of the transmission characteristic can be expressed by the following equation (1). Λ is the peak wavelength, n _eq is the equivalent refractive index, ΔL is the difference in length between adjacent arrayed waveguides, and m is the diffraction order.

λ = ΔL · n _eq / m (1)

That is, as is apparent from the above formula (1), the arrayed waveguide grating cannot be made polarization independent unless the equivalent refractive index n _eq is made equal between the polarized waves. In order to achieve such polarization independence, for example, a structure in which the waveguide structure is made the same in the horizontal direction and the vertical direction, that is, a structure in which a square core layer is embedded can be considered.

However, in general, in the semiconductor manufacturing process, although the thickness of the core layer can be precisely controlled in units of several tens of millimeters by epitaxial growth, the width is controlled by dry etching or the like in order to control the width in units of 10 ⁻¹ μm. Processing errors cannot be avoided. This processing error in the width of the waveguide leads to polarization dependence, which reduces the yield.

  Further, for example, Non-Patent Document 2 and the like have proposed means for making the equivalent refractive index the same between polarized waves even when the waveguide structure is asymmetric in the horizontal direction and the vertical direction. Specifically, a high-mesa waveguide having a core layer made of InGaAsP having a bandgap wavelength of 1.05 μm and a thickness of 0.5 μm, and a cladding layer made of InP sandwiching the core layer vertically. It is described that when the width of the waveguide is 2.65 μm, the polarization becomes independent.

  In such a waveguide structure, the change in the equivalent refractive index difference between the polarized waves with respect to the width of the waveguide is gradual and has a relatively wide fabrication tolerance. Since it depends on how far the shift of the peak wavelength of the transmission characteristic is tolerated, it is still necessary to strictly control the width of the waveguide in order to realize more complete polarization independence.

  Further, for example, in Non-Patent Document 3 and the like, a method of inserting a waveguide for compensating dispersion between polarized waves into an arrayed waveguide, or using a polarization splitter is proposed. However, all the various means described in Non-Patent Document 3 and the like require high production accuracy.

  Furthermore, in all the means described above, since the presence or absence of polarization dependence is determined at the stage of production, if there is polarization dependence after production, the correction, so-called trimming, cannot be performed.

  The present invention has been made in view of such a problem, and an object of the present invention is to generate a polarization caused by an error during a manufacturing process or a processing error in a wavelength multiplexer / demultiplexer using a semiconductor material. By making it possible to compensate the dependence without accompanying characteristic deterioration such as an increase in light propagation loss, the yield can be remarkably improved.

  A wavelength multiplexer / demultiplexer according to a first invention for solving the above-described problem includes a semiconductor substrate, a first semiconductor cladding layer provided on the semiconductor substrate, and the first semiconductor cladding layer. A non-doped semiconductor core layer provided on the semiconductor core layer, a second semiconductor clad layer provided on the semiconductor core layer, and between the first semiconductor clad layer and the semiconductor core layer and the second semiconductor clad layer. An arrayed waveguide grating having a plurality of arrayed waveguides having a semi-insulating layer made of a semi-insulating semiconductor provided between at least one of the semiconductor core layers, and the array of the arrayed waveguide gratings An electric field is applied to each arrayed waveguide so as to change the refractive index for the TE mode without changing the refractive index for the TM mode in the semiconductor core layer of the waveguide. Characterized in that it comprises a field applying means for applying.

  The wavelength multiplexer / demultiplexer according to the second invention is the wavelength multiplexer / demultiplexer according to the first invention, wherein the electric field applying means is located on the semi-insulating layer of the arrayed waveguide of the arrayed waveguide grating. A first electrode disposed on the waveguide and applied with a voltage; and a second electrode provided on the first semiconductor cladding layer or the semiconductor substrate of the arrayed waveguide grating and grounded. The first electrode is disposed between the arrayed waveguides so as to communicate between the plurality of arrayed waveguides of the arrayed waveguide grating, and the axial length of the arrayed waveguides is reduced. The arrayed waveguide is formed so as to be longer toward the other side than the one side in the arrangement direction.

  The wavelength multiplexer / demultiplexer according to a third aspect of the present invention is the wavelength multiplexer / demultiplexer according to the first aspect, wherein the electric field applying means is located on the semi-insulating layer of the arrayed waveguide of the arrayed waveguide grating. A first electrode disposed on the waveguide and applied with a voltage; and a second electrode provided on the first semiconductor cladding layer or the semiconductor substrate of the arrayed waveguide grating and grounded. The first electrodes are respectively disposed in the arrayed waveguides so as to intermittently connect the plurality of arrayed waveguides of the arrayed waveguide grating, and the axial lengths of the arrayed waveguides are respectively It is characterized by being equal.

  The wavelength multiplexer / demultiplexer according to a fourth aspect of the invention is the second or third aspect of the invention, wherein the arrayed waveguide of the arrayed waveguide grating has a linear portion that is linear in the axial direction, The first electrode is disposed on the straight portion of the arrayed waveguide of the arrayed waveguide grating.

A wavelength multiplexer / demultiplexer according to a fifth invention is the wavelength multiplexer / demultiplexer according to the fourth invention, wherein the semiconductor substrate of the arrayed waveguide grating has a (100) plane orientation, and the linear portion of the arrayed waveguide is These are characterized by being parallel to (1) orientation or (2) orientation shown in Table 1 below.

  A wavelength multiplexer / demultiplexer according to a sixth invention is the wavelength multiplexer / demultiplexer according to any one of the first to fifth inventions, wherein the first semiconductor clad layer and the second of the arrayed waveguide of the arrayed waveguide grating. The semiconductor clad layer is n-type.

  A wavelength multiplexer / demultiplexer according to a seventh invention is characterized in that, in any one of the first to sixth inventions, the arrayed waveguide of the arrayed waveguide grating has a high mesa structure.

  According to the wavelength multiplexer / demultiplexer according to the present invention, since the polarization-dependent trimming mechanism is provided, the polarization dependency caused by an error during the manufacturing process or a processing error can be prevented from deteriorating characteristics such as an increase in optical propagation loss. Compensation can be performed without this, and the yield can be significantly improved.

<Overview>
The shift of the peak wavelength of the transmission characteristics between the polarizations in the arrayed waveguide grating, that is, the polarization dependence is caused by the difference in the equivalent refractive index between the polarizations, as is apparent from the above-described equation (1). It is. For example, in the case of an equivalent refractive index difference of 0.0001, a peak wavelength shift of about 0.08 nm (10 GHz) occurs. Therefore, in the present invention, the polarization dependency can be compensated by providing the arrayed waveguide grating with a mechanism capable of adjusting the equivalent refractive index difference between the polarized waves after fabrication.

  By the way, means capable of changing the refractive index include the Pockels effect, the Franz-Keldish effect, the quantum confined Stark effect (QCSE) effect seen in the multiple quantum well structure, and the current injection, etc. A plasma effect, a temperature change, and the like due to a change in carrier density can be given.

  Here, except for the Pockels effect, the polarization dependency of the refractive index change is not or small, so that the peak wavelength of the transmission characteristic can be changed with the change of the refractive index by the above formula (1). Since the polarization shifts the wavelength in the same direction at the same time, it is not appropriate for compensation of polarization dependency. Furthermore, since the plasma effect is accompanied by absorption of free carriers when the carrier density is increased, there is a problem that the loss of light propagation increases.

On the other hand, the Pockels effect has a large polarization dependence of the refractive index change. For example, as described in Non-Patent Document 4, when the electric field E _b (= V _b / core thickness) in the direction perpendicular to the substrate ([100] orientation) is applied, the refractive index change with respect to the TE mode is as _follows. (2). Note that n ₀ is the refractive index of the material, and γ ₄₁ is a Pockels constant, which is a negative value.

^{_{Δn 0 = ± (n 3/}} 2) γ 41 E b ··· (2)

  In addition, due to the limitation of the document description method in the electronic application system, each direction is described as shown in Table 2 below.

  The + sign indicates the case where light propagates parallel to the [0AA] direction, and the − sign indicates the case where light propagates parallel to the [01A] direction. For this reason, when the arrayed waveguide portion to which the electric field is applied is arranged in parallel to the [0AA] direction, the refractive index change is negative by the application of the electric field, while the arrayed waveguide portion to which the electric field is applied is in the [01A] direction. When arranged in parallel, the change in the refractive index is reversed by application of the electric field, but no change in the refractive index with respect to the TM mode occurs.

  Therefore, in the present invention, by utilizing this Pockels effect, only the equivalent refractive index for the TE mode is appropriately changed to efficiently compensate for the polarization dependence.

  In the above description, the case of the (100) substrate is taken as an example, but the same applies to the case of the (001) substrate. That is, in the case of the (001) substrate, the [100] orientation may be replaced with the [001] orientation, the [0AA] orientation with the [110] orientation, and the [01A] orientation with the [A10] orientation.

<First embodiment>
A first embodiment of a wavelength multiplexer / demultiplexer according to the present invention will be described below with reference to FIGS. 1 is a plan view showing a schematic configuration of the wavelength multiplexer / demultiplexer, FIG. 2 is a sectional view showing a schematic configuration of the wavelength multiplexer / demultiplexer of FIG. 1, and FIG. 3 is a diagram of the wavelength multiplexer / demultiplexer of FIG. FIG. 4 is a band diagram when an electric field is applied to the wavelength multiplexer / demultiplexer of FIG. 1, and FIG. 5 is an adjacent array waveguide of the wavelength multiplexer / demultiplexer of FIG. FIG. 6 is a graph showing the relationship between the width of the arrayed waveguide and the equivalent refractive index difference, and FIG. 7 is the relationship between the equivalent refractive index change of the arrayed waveguide and the refractive index change of the semiconductor core layer. It is a graph showing.

  As shown in FIG. 1, on a semiconductor substrate 101 made of InP and having a (100) plane orientation, a plurality of input waveguides 110 having one end positioned on the peripheral edge of the substrate 101 are formed. The other end side of these input waveguides 110 is connected to one end side of the slab waveguide 120 formed on the substrate 101. On the other end side of the slab waveguide 120, a plurality of array conductors formed on the substrate 101 so as to have a linear portion 130a parallel to the [0AA] direction or the [01A] direction between the curved portions 130b. One end sides of the waveguides 130 are connected to each other.

  The other end sides of the arrayed waveguides 130 are connected to one end sides of the slab waveguides 140 formed on the substrate 101, respectively. One end side of a plurality of output waveguides 150 formed on the substrate 101 is connected to the other end side of the slab waveguide 140, and the other end side of the output waveguide 150 is connected to the periphery of the substrate 101. Located at each end.

  As shown in FIGS. 2 and 3, each of the waveguides 110, 120, 130, 140, and 150 has a first semiconductor clad layer 102 made of n-type InP provided on the substrate 101, and the clad layer. A semiconductor core layer 103 (band gap wavelength: 1.05 μm, thickness: 0.5 μm) made of non-doped InGaAsP is provided on 102, and a semi-insulation made of semi-insulating (SI) type InP is formed on the core layer 103. An (SI) layer 104 is provided, and a second semiconductor clad layer 105 made of n-type InP is provided on the SI layer 104 to form a high mesa structure.

  In the arrayed waveguide grating configured as described above, as shown in FIGS. 1 and 3, a part of the arrayed waveguide 130 on the linear portion 130a is connected to a power source and applied with a voltage. The electrodes 106 are arranged between the arrayed waveguides 130 so as to communicate between the linear parts 130 a of the arrayed waveguides 130, and the electrodes 106 are connected to the linear parts of the arrayed waveguides 130. The length of the axial direction of 130a is formed in a trapezoidal shape so as to be longer toward the other side in the arrangement direction (lower side in FIG. 1) than one side (upper side in FIG. 1) of the arrayed waveguide 130 in the arrangement direction. ing.

  That is, in the arrayed waveguides 130 arranged in a plurality, the electrodes 106 have the lengths in the axial direction of the linear portions 130a corresponding to the lengths of the arrayed waveguides 130.

  A second electrode 107 that is grounded is provided on the cladding layer 102. A part of the clad layer 105 is removed by etching so as to electrically insulate between the place where the electrode 106 is provided and the place where the electrode 106 is not provided on the linear portion 130a of the arrayed waveguide 130. Separation grooves 130c are respectively formed.

  That is, by operating the power supply, the same electric field can be efficiently applied only to the core layer 103 in the portion where the electrodes 106 of all the arrayed waveguides 130 are disposed. In the present embodiment, an electric field applying means is constituted by the electrodes 106 and 107, the power source and the like.

In the wavelength multiplexer / demultiplexer 100 according to this embodiment, when a negative bias _Vb is applied to the electrode 106 by the power supply, the SI layer 104 blocks the electrons e as shown in FIG. Therefore, no current flows and an electric field is efficiently applied to the core layer 103. For this reason, in the core layer 103, a refractive index change due to the Pockels effect as shown in the equation (2) is induced with respect to the TE mode.

Here, as shown in FIG. 5, in the i-th arrayed waveguide 130, the length of the linear portion 130a where the electrode 106 is located is Le _{, i,} and the electrode 106 is located above. The lengths of the straight portion 130a and the curved portion 130b that are not located are L _{0, i,} and the length of the straight portion 130a in which the electrode 106 is located in the upper portion of the arrayed waveguide 130 located at the i + 1th position is L. _{e, i + 1} , the length of the straight portion 130a and the curved portion 130b where the electrode 106 is not located above is L _{0, i + 1,} and the straight portion 130a where the electrode 106 is located above. the equivalent refractive index of the TE mode when the electric field is applied to the n ₁ ^TE, the equivalent refractive index of the TM mode when an electric field is applied in the linear portion 130a that the electrode 106 to the upper position and n ₁ ^TM at, before the top The equivalent refractive index of the TE mode in the linear portion 130a and the curved portions 130b electrodes 106 is not positioned as n ₀ ^TE, the equivalent refractive the TM mode in the linear portion 130a and the curved portion 130b the electrode 106 on the top is not positioned When the rate is n ₀ ^TM , the phase matching condition for the TE mode of the wavelength multiplexer / demultiplexer 100 can be expressed by the following general formula (3). The phase matching condition for the TM mode of the wavelength multiplexer / demultiplexer 100 is Can be represented by the following general formula (4). It should be noted that the linear portion 130a on which the electrode 106 is positioned has a TE mode equivalent refractive index of n ₀ ^TE and a TM mode equivalent refractive index of n ₀ ^TM when no electric field is applied. It is.

ΔL _e · n ₁ ^TE + ΔL ₀ · n ₀ ^TE = mλ (3)
ΔL _e · n ₁ ^TM + ΔL ₀ · n ₀ ^TM = mλ (4)

At this time,
ΔL _e = L _{e, i + 1} −L _{e, i} ,
ΔL ₀ = L _{0, i + 1} −L _{0, i} ,
ΔL _e + ΔL ₀ = ΔL (where ΔL> 0)
If the left side of equation (3) is equal to the left side of equation (4) as a polarization-independent condition, the following equation (5) is obtained.

(N ₁ ^TE −n ₁ ^TM ) / (n ₀ ^TE −n ₀ ^TM ) = − ΔL ₀ / ΔL _e
= 1− (ΔL / ΔL _e ) (5)

Further, as described above, there is no change in the refractive index with respect to the TM mode due to the Pockels effect, and the band gap wavelength (1.05 μm) of the core layer 103 is a wavelength band generally used in optical communication (1.55 μm band). Therefore, the equivalent refractive index n ₁ ^TM and the equivalent refractive index n ₀ ^TM can be made equal to each other. Therefore, if ΔL _e = αΔL (where α is an arbitrary real number other than 0), the above equation (5) can be further expressed as the following equation (6).

n ₁ ^TE −n ₀ ^TE = − (1 / α) (n ₀ ^TE −n ₀ ^TM ) (6)

That is, polarization independence can be realized by adjusting the magnitude of the applied voltage V _b so that the equivalent refractive index n ₁ ^TE satisfies the above formula (6).

  More specific description will be given below. As can be seen from FIG. 6, the arrayed waveguide 130 becomes independent of polarization when the width is near 2.65 μm, and when the width is wider than around 2.65 μm, the equivalent refractive index of the TE mode is equivalent to that of the TM mode. When the refractive index is larger than the refractive index and the width is narrower than around 2.65 μm, the TE mode equivalent refractive index becomes smaller than the TM mode equivalent refractive index.

Here, when the linear portion 130a of the arrayed waveguide 130 is arranged in parallel to the [0AA] direction, the refractive index change of the core layer 103 is negative with respect to the negative bias _{Vb according} to the equation (2). It becomes. That is, in the arrayed waveguide 130, when the equivalent refractive index n ₁ ^TE is smaller than the equivalent refractive index n ₀ ^TE (n ₁ ^TE <n ₀ ^TE ), polarization dependence can be compensated. It is.

If the width of the arrayed waveguide 130 is intentionally set in advance so as to be slightly wider than the polarization-independent condition, as can be seen from FIG. 6, the equivalent refractive index n ₀ ^TE is equal to the equivalent refractive index n. _It becomes larger than ₀ ^{TM and} the right side of the equation (6) becomes positive. Further, since the equivalent refractive index n ₁ ^TE is smaller than the equivalent refractive index n ₀ ^TE , the left side of the equation (6) is negative. Therefore, by satisfying α> 0 so that the relationship of the formula (6) is satisfied, the polarization dependency can be compensated. That is, according to the arrayed waveguide 130 becomes longer by [Delta] L, is the length of the straight portion 130a of the electrode 106 to the upper position may be lengthened by [Delta] L _e.

To explain with a simple example, when α = 1, that is, ΔL _e = ΔL, n ₁ ^TE = n ₀ ^TM (<n ₀ ^TE ), so that the equivalent refractive index of the TE mode is equivalent to that of the TM mode. It is only necessary to adjust (lower the refractive index) to be equal to the refractive index.

If the width of the arrayed waveguide 130 is intentionally set in advance so as to be slightly narrower than the polarization-independent condition, as can be seen from FIG. 6, the equivalent refractive index n ₀ ^TE is equal to the equivalent refractive index n. _It becomes smaller than ₀ ^{TM and} the right side of the equation (6) becomes negative. Further, since the equivalent refractive index n ₁ ^TE is smaller than the equivalent refractive index n ₀ ^TE , the left side of the equation (6) is negative. Therefore, by satisfying α <0 so that the relationship of the formula (6) is satisfied, the polarization dependence can be compensated.

On the other hand, when the linear portion 130a of the arrayed waveguide 130 is arranged in parallel to the [01A] direction, the refractive index change of the core layer 103 is positive with respect to the negative bias _{Vb according} to the equation (2). Become. Even in this case, compensation can be made in the same manner as the above-described theory in which the linear portion 130a of the arrayed waveguide 130 is arranged in parallel to the [0AA] direction. The only difference between this case and the case described above is that the sign of α is reversed. As can be seen from the equation (6), when α is increased, the refractive index change can be reduced, that is, the applied electric field can be reduced.

  The above conditions are summarized as shown in Table 3 below.

Next, a simple verification will be described. As described above, the processing error of the waveguide width is about 10 ⁻¹ μm. This magnitude is about 0.00015 when converted to the amount of change in the equivalent refractive index of the waveguide structure of the present invention (see FIG. 6). In other words, polarization independence can be achieved by changing the equivalent refractive index by about 10 ⁻⁴ .

Although the Pockels constant of InGaAsP is not exactly known, the Pockels constant of GaAs of the same material is about −1.4 × 10 ⁻¹² m / V, so that it is estimated that InGaAsP has the same value. . Since the core layer of InGaAsP has a refractive index of about 3.25, it is possible to induce a refractive index change of about 10 ^{−4 with} respect to a voltage of about 2 V from the above formula (2).

Further, as can be seen from FIG. 7, it can be said that the equivalent refractive index change of the arrayed waveguide 130 is affected by about half of the refractive index change of the core layer 103. Therefore, in order to change the equivalent refractive index of the arrayed waveguide 130 by about 10 ⁻⁴ , a voltage of about 4 V may be applied, and by appropriately setting the doping density and thickness of Fe or the like in the SI layer 104. Since it can have a sufficient breakdown voltage, it can be handled without any problem.

  As described above, since the wavelength multiplexer / demultiplexer 100 according to the present embodiment does not use the p-type layer, the optical loss is not increased, and the polarization dependence is efficiently compensated by applying the voltage even after the fabrication. can do. Therefore, according to the wavelength multiplexer / demultiplexer 100 according to the present embodiment, polarization independence can be easily realized even with a semiconductor material, so that the yield can be improved and the cost can be reduced. Can be achieved.

<Second Embodiment>
A first embodiment of the wavelength multiplexer / demultiplexer according to the present invention will be described below with reference to FIGS. FIG. 8 is a plan view showing a schematic configuration of the wavelength multiplexer / demultiplexer, and FIG. 9 is a diagram for explaining the correlation between adjacent arrayed waveguides of the wavelength multiplexer / demultiplexer of FIG. In addition, about the part similar to the case of 1st embodiment mentioned above, the code | symbol similar to the code | symbol used in description of 1st embodiment mentioned above is used in this embodiment, and 1st mentioned above. Description overlapping with that in the second embodiment is omitted.

  As shown in FIG. 8, a first electrode 206 connected to a power source and applied with a voltage is intermittently connected between the arrayed waveguides 130 at a part of each of the linear portions 130 a of the arrayed waveguides 130. The electrodes 206 are equal in length in the axial direction of the linear portion 130a of the arrayed waveguide 130.

  That is, in the wavelength multiplexer / demultiplexer 100 according to the first embodiment described above, the wavelength multiplexer / demultiplexer 100 is disposed between the arrayed waveguides 130 so as to communicate with the linear portions 130a of the arrayed waveguides 130. At the same time, a single first trapezoidal shape is formed so that the length in the axial direction of the linear portion 130a of the arrayed waveguide 130 is longer than the one side in the array direction of the arrayed waveguide 130. One electrode 106 is applied so that the same electric field can be applied by electrically short-circuiting between the straight portions 130a. In the wavelength multiplexer / demultiplexer 200 according to the present embodiment, each array waveguide 130 is provided. And applying a plurality of first electrodes 206 having the same length in the axial direction of the linear portion 130a of the arrayed waveguide 130, respectively. It had to be able to apply independently of the electric field electrically separated respectively between lines portion 130a.

Here, as shown in FIG. 9, the length of the straight portion 130a of the array waveguide 130 where the electrode 206 on the top is positioned and L _e, the straight portion 130a and the said electrode 206 on the top is not positioned The equivalent refractive index of the TE mode in the curved portion 130b is n ₀ ^TE , the equivalent refractive index of the TM mode in the straight portion 130a and the curved portion 130b where the electrode 206 is not located above is n ₀ ^TM, and the i th In the arrayed waveguide 130 positioned at, the length of the straight portion 130a and the curved portion 130b where the electrode 206 is not located at the upper portion is L _{0, i,} and the length of the straight portion 130a where the electrode 206 is located at the upper portion. voltage V _b, the equivalent refractive index of the TE mode when _i is applied to the n ₁ ^TE, the voltage V _b at the straight portion 130a of the electrode 206 to the upper _{position, i} The equivalent refractive index of the TM mode at the time of pressurizing and n ₁ ^TM, in the arrayed waveguide 130 located at the i + 1 th, the length of the straight portion 130a and the curved portion 130b the electrode 206 on the top is not positioned L _{0 , i + 1,} and the equivalent refractive index of the TE mode when the voltage V _{b, i + 1 is} applied to the linear portion 130a where the electrode 206 is located above is n _{i + 1} ^TE, and the electrode 206 is located above. If the equivalent refractive index of the TM mode when the voltage V _{b, i + 1 is} applied to the linear portion 130a is n _{i + 1} ^TM (where L _{0, i + 1} −L _{0, i} = ΔL (> 0)) The phase matching condition of the TE mode of the wavelength multiplexer / demultiplexer (arrayed waveguide grating) 200 can be expressed by the following general formula (7). The TM mode phase matching condition can be expressed by the following general formula (8). The It should be noted that the linear portion 130a on which the electrode 206 is positioned has a TE mode equivalent refractive index of n ₀ ^TE and a TM mode equivalent refractive index of n ₀ ^TM when no electric field is applied. It is.

_Le · (n _{i + 1} ^TE −n _i ^TE ) + ΔL · n ₀ ^TE = mλ (7)
_Le · (n _{i + 1} ^TM −n _i ^TM ) + ΔL · n ₀ ^TM = mλ (8)

  Therefore, assuming that the left side of equation (7) is equal to the left side of equation (8) as a polarization-independent condition, the following equation (9) is obtained.

{(N _{i + 1} ^TE −n _i ^TE ) − (n _{i + 1} ^TM −n _i ^TM )} / (n ₀ ^TE −n ₀ ^TM )
= -ΔL / L _e (9)

Similarly to the case of the first embodiment described above, since there is no change in the refractive index of the TM mode and n _{i + 1} ^TM = n _i ^TM , the above equation (9) is further expressed by the following equation ( 10).

n _{i + 1} ^TE −n _i ^TE = − (ΔL / L _e ) (n ₀ ^TE −n ₀ ^TM )
= Constant (10)

That is, the applied voltage V _{b, i + 1} (<0) to the i ₊ 1th array waveguide 130 is set to a certain value ΔV _{b with} respect to the applied voltage _{Vb, i} (<0) of _{the ith} array waveguide 130. By making adjustment so that = V _{b, i + 1} −V _{b, i} , polarization independence can be realized.

  More specific description will be given below. In order to make the polarization independent regardless of the width of the arrayed waveguide 130, when the straight portion 130 a of the arrayed waveguide 130 is arranged in parallel to the [0AA] direction, Thus, the refractive index change of the core layer 103 becomes negative.

That is, for example, when the width of the arrayed waveguide 130 becomes wider than the polarization-independent condition due to an error during manufacturing, the equivalent refractive index n ₀ ^TE is larger than the equivalent refractive index n ₀ ^TM. (N ₀ ^TE > n ₀ ^TM ) (see FIG. 6), the right side of the equation (10) is negative, so that the equivalent refractive index n _i ^TE is greater than the equivalent refractive index n _{i + 1} ^TE . Therefore, the applied voltage should be adjusted so that (n _{i + 1} ^TE <n _i ^TE ) increases. That is, as the arrayed waveguide 130 becomes longer by ΔL, the bias value is increased (| V _{b, i + 1} |> | V _{b, i} |, ΔV _b <0).

When the width of the arrayed waveguide 130 becomes narrower than the polarization-independent condition, the equivalent refractive index n ₀ ^TE is smaller than the equivalent refractive index n ₀ ^TM (n ₀ ^TE <n ₀ ^TM ) for (see FIG. 6), the equivalent refractive index n _i ^TE may be adjusted to the to be smaller than the effective refractive index ^{_{n i + 1 TE (n i}} + 1 TE> n i TE) applied voltage . That is, as the arrayed waveguide 130 becomes longer by ΔL, the bias value may be reduced (| V _{b, i + 1} | <| V _{b, i} |, ΔV _b > 0).

On the other hand, when the linear portion 130a of the arrayed waveguide 130 is arranged in parallel to the [01A] direction, the refractive index change of the core layer 103 becomes positive with respect to a negative bias according to the equation (2). Even in this case, compensation can be made in the same manner as the above-described theory in which the linear portion 130a of the arrayed waveguide 130 is arranged in parallel to the [0AA] direction. The only difference between this case and the case described above is that the sign of ΔV _b is reversed with respect to the actual deviation of the width of the arrayed waveguide 130 from the polarization-independent condition. As can be seen from the equation (10), when _{Le is} increased, the equivalent refractive index change n _{i + 1} ^TE −n _i ^TE can be reduced, that is, the absolute value of ΔV _b can be reduced. .

  The above conditions are summarized as shown in Table 4 below.

As described above, since the wavelength multiplexer / demultiplexer 200 according to the present embodiment does not use the p-type layer, the optical loss is not increased, and the polarization dependence is efficiently compensated by applying the voltage even after fabrication. be able to. Therefore, according to the wavelength multiplexer / demultiplexer 200 according to the present embodiment, as in the case of the wavelength multiplexer / demultiplexer 100 according to the first embodiment described above, even if it is a semiconductor material, it is polarization independent. Can be realized easily, so that the yield can be improved and the cost can be reduced, as well as the width of the waveguide actually deviates from the polarization independent condition. Even in this case, it can be easily compensated by changing the sign of ΔV _b .

<Other embodiments>
In the first and second embodiments described above, the upper layer of the core layer 103 is n-type InP. However, as another embodiment, for example, the upper layer of the core layer 103 is p-type InP. Even in the so-called pin structure, in principle, the same operation and effect as in the case of the present embodiment can be obtained. However, the p-type layer has a large light absorption due to the transition between valence band bands, and when an optical field is applied, it causes an increase in propagation loss. This is not very desirable for relatively large devices. Furthermore, if the core layer (i layer) is made thick in order to reduce light absorption, the electric field strength will be weakened and the efficiency will be deteriorated. Therefore, it is preferable to avoid it if possible.

  In the first and second embodiments described above, n-type InP is used for the clad layer 105. However, as another embodiment, for example, the SI layer 104 is not exposed to an optical field on the clad layer 105. If the thickness of the clad layer 105 is increased, p-type InP can be used for the clad layer 105.

  In the first and second embodiments described above, the SI layer 104 is provided between the core layer 103 and the clad layer (second semiconductor clad layer) 105. For example, even if an SI layer is provided between the semiconductor core layer and the first semiconductor clad layer, the same effects as those of the present embodiment can be obtained. However, in this case, it is necessary to reverse the direction of the applied electric field as in the case of the first and second embodiments described above. In addition, it is possible to provide SI layers both between the semiconductor core layer and the first semiconductor clad layer and between the semiconductor core layer and the second semiconductor clad layer.

  In the first and second embodiments described above, the case where the core layer 103 is bulk has been described. However, as another embodiment, for example, the core layer 103 may have a quantum well structure. is there.

  In the first and second embodiments described above, the electrodes 106 and 107 and the layers 102 and 105 are in direct contact. However, as another embodiment, for example, the electrodes 106 and 107 and the layers 102 and 105 are in contact with each other. It is also possible to provide a contact layer made of InGaAs, InGaAsP or the like between the layers 102 and 105 so as to lower the contact resistance.

  Further, in the first and second embodiments described above, the second electrode 107 to be grounded is provided on the first semiconductor clad layer 102. However, as another embodiment, for example, n is provided on the semiconductor substrate. If a mold is used, a second electrode to be grounded can be provided on the back surface of the semiconductor substrate.

  In the first and second embodiments described above, the case where an InGaAsP / InP-based semiconductor material is used has been described. However, the present invention is not limited to such a material, and other semiconductor materials are used. However, it can be applied in the same manner as in the first and second embodiments described above.

  The wavelength multiplexer / demultiplexer according to the present invention can compensate for polarization dependency caused by errors in manufacturing processes and processing errors without causing deterioration of characteristics such as an increase in optical propagation loss. When applied to an optical transmission system such as optical communication, optical exchange, and optical information processing using the method, it can be used extremely effectively.

1 is a plan view showing a schematic configuration of a first embodiment of a wavelength multiplexer / demultiplexer according to the present invention. FIG. 2 is a cross-sectional view illustrating a schematic configuration of the wavelength multiplexer / demultiplexer of FIG. 1. FIG. 2 is a cross-sectional view illustrating a schematic configuration of a main part of the wavelength multiplexer / demultiplexer in FIG. 1. 2 is a band diagram when an electric field is applied to the wavelength multiplexer / demultiplexer of FIG. FIG. 3 is a correlation explanatory diagram of adjacent arrayed waveguides of the wavelength multiplexer / demultiplexer of FIG. 1. It is a graph showing the relationship between the width | variety of an arrayed waveguide, and an equivalent refractive index difference. It is a graph showing the relationship between the equivalent refractive index change of an arrayed waveguide, and the refractive index change of a semiconductor core layer. It is a top view showing schematic structure of 2nd embodiment of the wavelength multiplexer / demultiplexer which concerns on this invention. FIG. 3 is a correlation explanatory diagram of adjacent arrayed waveguides of the wavelength multiplexer / demultiplexer of FIG. 1. It is a top view showing schematic structure of an example of the conventional wavelength multiplexer / demultiplexer.

Explanation of symbols

100, 200 Wavelength multiplexer / demultiplexer 101 Semiconductor substrate 102 First semiconductor clad layer 103 Semiconductor core layer 104 Semi-insulating (SI) layer 105 Second semiconductor clad layer 106, 206 First electrode 107 Second electrode 110 Input Waveguide 120, 140 Slab waveguide 130 Array waveguide 130a Straight line part 130b Curved part 130c Separation groove 150 Output waveguide

Claims (7)

A semiconductor substrate;
A first semiconductor cladding layer provided on the semiconductor substrate;
A non-doped semiconductor core layer provided on the first semiconductor cladding layer;
A second semiconductor cladding layer provided on the semiconductor core layer;
Semi-insulating made of a semi-insulating semiconductor provided between at least one of the first semiconductor clad layer and the semiconductor core layer and between the second semiconductor clad layer and the semiconductor core layer. An arrayed waveguide grating comprising a plurality of arrayed waveguides having layers; and
Electric field applying means for applying an electric field to the arrayed waveguide so as to change the refractive index for the TE mode without changing the refractive index for the TM mode in the semiconductor core layer of the arrayed waveguide of the arrayed waveguide grating; A wavelength multiplexer / demultiplexer characterized by comprising: In claim 1,
The electric field applying means is
A first electrode disposed on the arrayed waveguide to be positioned on the semi-insulating layer of the arrayed waveguide grating and applied with a voltage;
A second electrode provided on the first semiconductor clad layer or the semiconductor substrate of the arrayed waveguide grating and grounded,
The first electrode is disposed between the arrayed waveguides so as to communicate between the plurality of arrayed waveguides of the arrayed waveguide grating, and the axial length of the arrayed waveguides is reduced. A wavelength multiplexer / demultiplexer characterized in that the other side of the arrayed waveguide is longer than the other side in the arrangement direction. In claim 1,
The electric field applying means is
A first electrode disposed on the arrayed waveguide to be positioned on the semi-insulating layer of the arrayed waveguide grating and applied with a voltage;
A second electrode provided on the first semiconductor clad layer or the semiconductor substrate of the arrayed waveguide grating and grounded,
The first electrodes are respectively disposed in the arrayed waveguides so as to intermittently connect the plurality of arrayed waveguides of the arrayed waveguide grating, and the axial lengths of the arrayed waveguides are respectively Wavelength multiplexer / demultiplexer characterized by equality. In claim 2 or claim 3,
The arrayed waveguide of the arrayed waveguide grating has a linear portion that is linear in the axial direction;
The wavelength multiplexer / demultiplexer, wherein the first electrode is disposed on the linear portion of the arrayed waveguide of the arrayed waveguide grating. In claim 4,
The semiconductor substrate of the arrayed waveguide grating has a (100) plane orientation;
The wavelength multiplexer / demultiplexer, wherein the linear portion of the arrayed waveguide is parallel to (1) direction or (2) direction shown in Table 1 below.

In any one of Claims 1-5,
The wavelength multiplexer / demultiplexer, wherein the first semiconductor clad layer and the second semiconductor clad layer of the arrayed waveguide of the arrayed waveguide grating are n-type. In any one of Claims 1-6,
The wavelength multiplexer / demultiplexer, wherein the arrayed waveguide of the arrayed waveguide grating has a high mesa structure.

Images