JP2012160518A - Polarization-independent type semiconductor optical amplification device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a polarization-independent type optical semiconductor amplifier having a high optical gain and driven with low power consumption.SOLUTION: A polarization-independent type optical semiconductor amplification device has: an n-type InP clad layer 2 formed above an InP substrate 1; a p-type InP clad layer 7 formed above the InP substrate 1; an optical active layer 5 formed between the n-type InP clad layer 2 and the p-type InP clad layer 7 and to which tensile strain is given; and an n-type guide layer 3 formed between the optical active layer 5 and the n-type InP clad layer 2 so as to have a thickness of 500 nm or more, and formed from a compound semiconductor having an energy bandgap wavelength between that of InP and that of the optical active layer 5.

Description

  The present invention relates to a polarization-independent semiconductor optical amplifier.

  The semiconductor optical amplifier is a compact optical amplifier that can be used in various wavelength regions, and is adopted in various standardization standards such as optical Ethernet (registered trademark) and high-speed access optical network with a total traffic volume of 100 Gb / s. Full-scale application has begun in the network. In addition, various uses such as a booster amplifier, an in-line amplifier, and a preamplifier applied to various photonic networks are being studied for semiconductor optical amplifiers.

  In a semiconductor optical amplifier (SOA) operating in optical communication wavelength bands of 1.3 μm and 1.55 μm, a photoactive layer is formed by forming a GaInAsP or AlGaInAs material on an InP substrate using an epitaxial growth technique. . Then, by injecting a current into the photoactive layer, an optical gain is generated by a stimulated emission effect generated in the photoactive layer.

  A semiconductor optical amplifier applied to an optical communication system may have a characteristic that is not required for a semiconductor laser used as a light source, that is, a characteristic that the optical gain is always constant regardless of the polarization state of the input signal light. desirable. As a technique for realizing such a polarization-independent semiconductor optical amplifier, it is known to form a photoactive layer into which an extension strain is introduced above an InP substrate.

JP 2009-224691 A

  By the way, low power consumption is required for the SOA applied to the photonic network. The power consumption of the SOA is in a trade-off relationship with its optical gain and optical output. In general, in order to reduce the power consumption of the SOA, it is effective to use a thick photoactive layer to increase the optical confinement factor, set the required optical gain to a shorter SOA element length, and reduce the drive current. is there.

  However, in the polarization independent SOA as described above, it is difficult to form the photoactive layer with a preferable thickness because it uses a photoactive layer having a tensile strain. That is, in general, in a technique of forming a semiconductor layer having crystal strain using an epitaxial growth technique, when a certain reference film thickness determined by the applied strain amount, that is, a critical film thickness is exceeded, many crystal dislocations occur and the crystallinity is increased. to degrade. As a result, serious characteristic deterioration occurs in the SOA having the photoactive layer into which the extension strain having a strain amount exceeding the critical film thickness is introduced.

  Therefore, according to the polarization-independent SOA having a stretched strain photoactive layer according to the prior art, it is difficult to reduce the power consumption by increasing the thickness of the photoactive layer while maintaining good amplification characteristics.

  An object of the present invention is to provide a polarization-independent optical semiconductor amplifier that has high optical gain and is driven with low power consumption.

According to one aspect, the n-type InP cladding layer formed above the InP substrate, the p-type InP cladding layer formed above the InP substrate, the n-type InP cladding layer, and the p-type InP cladding. A photoactive layer formed between the layers and subjected to tensile strain, and is formed between the photoactive layer and the n-type InP cladding layer to a thickness of 500 nm or more, and between InP and the photoactive layer. There is provided a polarization-independent semiconductor optical amplifier having an n-type guide layer formed of a compound semiconductor having an energy band gap wavelength of the size of
The objects and advantages of the invention will be realized and attained by means of the elements and combinations particularly pointed out in the appended claims. It is understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not intended to limit the invention.

In a polarization-independent semiconductor optical amplifier having a photoactive layer to which a tensile strain is applied, an n-GaInAsP guide layer having a thickness of 500 nm or more is inserted between the n-InP cladding and the photoactive layer, and a band gap For the wavelength, a material is selected in which the guide layer is between InP and the photoactive layer. This induces an effect of anisotropically reducing the internal loss in the TE polarization mode and the TM polarization mode, and can reduce the amount of stretch distortion necessary for polarization independence.
This makes it possible to increase the photoactive layer thickness more than before without degrading crystal quality due to strain accumulation. Furthermore, compared with the conventional polarization-independent semiconductor optical amplifying device, the optical confinement of the photoactive layer is larger, and a high-efficiency optical gain can be realized with a short element length.

FIG. 1A is a cross-sectional view (part 1) illustrating a process of forming a polarization-independent semiconductor optical amplifier according to the first embodiment. FIG. 1B is a cross-sectional view (part 2) illustrating the process of forming the polarization-independent semiconductor optical amplifier according to the first embodiment. FIG. 1C is a cross-sectional view (part 3) illustrating the process of forming the polarization-independent semiconductor optical amplifier according to the first embodiment. FIG. 1D is a cross-sectional view (part 4) illustrating the step of forming the polarization-independent semiconductor optical amplifier according to the first embodiment. FIG. 2 is a cross-sectional view in the light traveling direction showing the polarization-independent semiconductor optical amplifier according to the first embodiment. FIGS. 3A and 3B are intensity distribution diagrams of the optical waveguide mode of the polarization-independent semiconductor optical amplifier according to the first embodiment. 4A and 4B are intensity distribution diagrams of the optical waveguide mode of the polarization-independent semiconductor optical amplifier according to the comparative example. FIG. 5 is a diagram showing the relationship between the strain amount of the photoactive layer and the material gain in the polarization-independent semiconductor optical amplifier. FIG. 6 is a diagram illustrating the relationship of internal loss depending on the presence or absence of the light guide layer in the polarization-independent semiconductor optical amplifier according to each of the first embodiment and the comparative example. FIG. 7 is a diagram showing the difference in the optimum strain amount of the photoactive layer by adopting either the presence or absence of the guide layer in the layer structure of the polarization-independent semiconductor optical amplifier according to the first embodiment. . FIG. 8 is a diagram illustrating the relationship between the thickness of the guide layer and the internal loss in the polarization-independent semiconductor optical amplifier according to the first embodiment. FIG. 9 is a diagram illustrating the relationship between the TE / TM polarization internal loss difference and the thickness of the guide layer in the polarization-independent semiconductor optical amplifier according to the first embodiment. FIG. 10A is a cross-sectional view (part 1) illustrating the process of forming the polarization-independent semiconductor optical amplifier according to the second embodiment. FIG. 10B is a cross-sectional view (part 2) illustrating the step of forming the polarization-independent semiconductor optical amplifier according to the second embodiment. FIG. 10C is a cross-sectional view (part 3) illustrating the process of forming the polarization-independent semiconductor optical amplifier according to the second embodiment. FIG. 10D is a cross-sectional view (part 4) illustrating the step of forming the polarization-independent semiconductor optical amplifier according to the second embodiment. FIGS. 11A and 11B are intensity distribution diagrams of the optical waveguide mode of the polarization-independent semiconductor optical amplifier according to the second embodiment. 12A and 12B are intensity distribution diagrams of the optical waveguide mode of the polarization-independent semiconductor optical amplifier according to the comparative example. FIG. 13 is a diagram illustrating the relationship of internal loss depending on the presence or absence of the light guide layer in the polarization-independent semiconductor optical amplifier according to each of the second embodiment and the comparative example. FIG. 14 is a diagram showing the difference in the optimum strain amount of the photoactive layer by adopting either the presence or absence of the guide layer in the layer structure of the polarization-independent semiconductor optical amplifier according to the second embodiment. .

  Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings. In the drawings, similar components are given the same reference numerals.

(First embodiment)
1A to 1D are cross-sectional views illustrating a manufacturing process of a polarization-independent semiconductor optical amplifying device (SOA) having a buried heterojunction current confinement (BH) structure according to a first embodiment of the present invention. .

  Next, steps required until a structure shown in FIG. 1A is formed will be described. Note that elements such as gallium, indium, phosphorus, and arsenic that represent the compound semiconductor layer are represented by element symbols Ga, In, P, As, and the like.

First, an n-type (n-) InP clad layer 2 having a thickness of about 200 nm and a guide layer 3 having a thickness of 500 nm or more, for example, about 1000 nm are formed in order on the n-type InP substrate 1. The guide layer 3 is formed of a Ga _x In _1-x As _y P _1-y (where 0 <x <1, 0 <y <1) layer.

  Further, on the guide layer 3, a first undoped (i-) GaInAsP light / carrier separation confinement (SCH) layer 4 having a thickness of about 100 nm and an i-GaInAsP stretched strain multiple quantum well (MQW) having a thickness of about 110 nm are formed. ) The photoactive layer 5 is formed in order. The photoactive layer 5 may be formed of an AlGaInAs layer.

  On the photoactive layer 5, a second i-GaInAsP • SCH layer 6 having a thickness of about 100 nm, a p-type (p-) InP cladding layer 7 having a thickness of about 2000 nm, and a p-type having a thickness of about 300 nm. A GaInAs contact layer 8 is formed in order. The n-InP cladding layer 2 to the p-GaInAs contact layer 8 are epitaxially grown using, for example, metal organic vapor phase epitaxy (MOVPE).

  In the materials of the SCH layers 4 and 6, the composition ratio of the group III element and the group V element, that is, Ga, In, As, and P, is adjusted so that the energy band gap wavelength is 1.3 μm. Further, the SCH layers 4 and 6 are not limited to a single layer structure, and may be formed in a multi-layer structure in which the energy band gap becomes wider as the composition ratio moves away from the tensile strain MQW photoactive layer 5.

Further, the extension strain MQW photoactive layer 5 has a plurality of quantum well layers having a small energy band gap and a plurality of quantum well layers having a large energy band gap sandwiching the quantum well layers individually. In the extension strain MQW photoactive layer 5, either the well layer or the barrier layer has extension strain by controlling the composition of the group III element and the group V element. The amount of tensile strain is adjusted to 0.8% when strain is introduced into the barrier layer, for example, but is not limited thereto.

  The composition of Ga, In, As, and P in the n-GaInAsP guide layer 3 is adjusted so that the energy band gap wavelength is intermediate between the extension strain MQW photoactive layer 5 and the n-InP cladding layer 2. The energy band gap wavelength of the material of the guide layer 3 is set to 1.05 μm, for example.

  After the contact layer 8 is formed, the n-InP substrate 1 is transferred to a CVD apparatus, and a silicon oxide film having a thickness of about 200 nm is formed as a dielectric film 9 on the upper surface of the contact layer 8 in the chamber. Further, after the n-InP substrate 1 is taken out from the CVD apparatus, a photoresist is applied on the dielectric film 9, and this is exposed and developed to form a resist pattern 10 having a stripe optical waveguide shape.

Next, steps required until a structure shown in FIG. 1B is formed will be described.
First, the silicon oxide film 9 is etched using the resist pattern 10 as a mask to form a hard mask 9a made of the silicon oxide film 9. For the etching of the silicon oxide film 9, for example, dry etching such as reactive ion etching (RIE) using a fluorine-based gas may be employed, or wet etching using hydrofluoric acid may be employed.

  Subsequently, the contact layer 8 in the region exposed from the hard mask 9a to the n-InP cladding layer 2 is dry-etched to form a waveguide mesa structure and to form buried holes 1a on both sides thereof. As the dry etching method, for example, inductively coupled plasma (ICP) RIE is used. Further, a chlorine-based gas is used as an etching gas.

  In the formation of the waveguide mesa structure, the depth of the buried hole 1a is not particularly limited when the stretched strain MQW photoactive layer 5 is etched to form the buried hole 1a, and in particular, the n-GaInAsP guide layer 3, The etching depth of the n-InP clad layer 2 is not limited.

  However, in the case of adopting a width taper structure in which the widths of both the input and output ends of the SOA are tapered, the n-GaInAsP guide layer 3 is etched to a depth that reaches at least the n-InP cladding layer 2, and FIG. The waveguide is processed into the mesa structure shown. As a result, the performance and uniformity of the light spot size conversion at both the input and output ends of the SOA are improved, and a higher performance SOA can be manufactured.

  Subsequently, as shown in FIG. 1C, a high-resistance semi-insulating (SI) -InP current blocking layer 11 heterojunctioned on both sides of the waveguide mesa structure with the hard mask 9a left is buried by a MOVPE method in the buried hole 1a. A current confinement structure is formed by selectively embedding in the inside. Here, InP does not substantially grow on the hard mask 9a made of a silicon oxide film.

  Thereafter, the hard mask 9a is removed by, for example, buffered hydrofluoric acid, and then the back surface of the n-InP substrate 1 is polished. Thereafter, as shown in FIG. 1D, a cathode electrode 12 having a laminated structure of, for example, gold, germanium, and gold is formed on the back surface of the n-InP substrate 1, and then the p-GaInAs contact layer 8 and the SI-InP current blocking layer are formed. A laminated structure of, for example, titanium, platinum, and gold is formed on 11 as the anode electrode 13.

  Thus, the wafer process is completed. Thereafter, the InP wafer is cleaved in an array, and if necessary, anti-reflection (AR) coating with a dielectric multilayer film is applied to each of the input end face and output end face, which are cleavage faces of the element, and chips are separated. Thereby, a polarization-independent semiconductor optical amplifier is completed.

  The cross section seen from the light output end or the light input end of the polarization-independent semiconductor optical amplifier is as shown in FIG. 1D, and the cross section in the light traveling direction is as shown in FIG. In FIG. 2, the alternate long and short dash line indicates the light intensity distribution of the TE waveguide mode and the TM waveguide mode, and the broken line indicates the distribution of the refractive index n in the stacking direction.

  In the above polarization-independent semiconductor optical amplifier, current is injected into the photoactive layer 5 made of GaInAsP or AlGaInAs formed on the InP substrate 1 using an epitaxial growth technique, and light is input to the optical input end. By entering the light, an optical gain due to the stimulated emission effect is generated in the photoactive layer 5.

  When light is output from the polarization-independent semiconductor optical amplifier, the electric field intensity distribution in the TE waveguide mode is as shown in FIG. 3A, and the electric field intensity distribution in the TM waveguide mode is As shown in FIG. As a comparative example for this, the electric field intensity distribution in the TE waveguide mode and the electric field intensity distribution in the TM waveguide mode of the polarization-independent semiconductor optical amplifier without the n-type guide layer 3 are shown in FIGS. Shown in (b).

By the way, in a semiconductor optical amplifier applied to an optical communication system, the optical gain is always constant regardless of the polarization state of the input signal light, that is, the polarization difference (Polarization Dependent
Gain: PDG) needs to be sufficiently small.

The optical gain G and the inter-polarization gain difference PDG in the semiconductor optical amplifier are generally expressed by the following equations (1) and (2). In these equations, Γ, Γ _TE and Γ _TM are optical confinement coefficients for each polarization mode, and g, g _TE and g _TM are the material gain coefficients of the photoactive layer for each polarization mode, and α, α _TE and α _TM are the internal loss factors of the waveguide for each polarization mode, and L is the SOA element length.

      G = exp [(Γg−α) L] (1)

_{PDGα [(Γ TE g TE -α} TE) - (Γ TM g TM -α TM)] (2)

In order to obtain a high light output and a low element electric resistance, a general SOA photoactive layer often adopts a flat shape whose width is significantly wider than the thickness. In this case, the optical confinement factor Γ for each polarization mode is not the same, and the optical confinement factor Γ _TE for the TE polarization mode is larger than the optical confinement factor Γ _TM for the TM polarization mode. That is, Γ _TE > Γ _TM . On the other hand, when a photoactive layer made of a crystal lattice-matched to the substrate is used, the material gain α (α _TE , α _TM ) is isotropic, so that the PDG shown in Equation (2) becomes large.

In order to avoid this, there is a method of forming a photoactive layer made of a material to which a uniaxial extension strain is applied, and adopting a structure that induces a material gain anisotropy relationship g _TE <g _TM . As a result, the polarization dependence of the above-described optical confinement Γ _TE > Γ _TM can be compensated to obtain a relationship of Γ _TE g _TE ≈Γ _TM g _TM , which is an effective technique for realizing a small PDG. is there.

On the other hand, the absolute value of the anisotropy g _TE / g _TM of the material gain coefficient of TE and TM waveguide modes indicating the anisotropy generated in the material gain coefficient of the photoactive layer in the TE and TM waveguide modes is shown in FIG. As shown, it increases as the strain amount of the photoactive layer increases. Further, as shown in FIG. 5, the magnitude relationship between the gain coefficients g _TE and g _TM of the TE waveguide mode and the TM waveguide mode is reversed between the strain in the in-plane direction of the growth film and the compression strain. In this case, in the polarization-independent SOA, a region having an extension strain in which the relationship of g _TE <g _TM is obtained as described above is employed.

By the way, low power consumption is required for the SOA applied to the photonic network. However, the power consumption of the SOA is in a trade-off relationship with its optical gain and optical output. In general, in order to reduce the power consumption of the SOA, it is necessary to increase the optical confinement factor Γ by using a thicker photoactive layer and to adopt a shorter element length L in the equation (1) of the optical gain G. A measure for realizing a high optical gain G and reducing the drive current is effective.

  However, as described above, in a polarization-independent SOA using a stretched strain photoactive layer, there is a limit to increasing the layer thickness of the photoactive layer. That is, in general, in a semiconductor layer having a crystal strain grown using an epitaxial growth technique, for example, a tensile strain, when the film thickness is increased beyond the critical thickness, crystal dislocation occurs frequently and the crystallinity deteriorates. If the is applied to the SOA, serious characteristic deterioration occurs.

  On the other hand, in this embodiment, by forming the guide layer 3 having a thickness of 500 nm or more between the first SCH layer 4 and the n-InP clad layer 2 under the photoactive layer 5, the tensile strain MQW The problem is solved by reducing the amount of strain required for the photoactive layer 5. In this case, the guide layer 3 is formed of a material having an energy band gap that is a size between the material of the first SCH layer 4 and InP. The first SCH layer 4 is formed of a material having a larger energy band gap than the material of the photoactive layer 5.

  By adopting such a structure, the n-GaInAsP guide layer 3 has a size between the refractive indexes of the InP cladding layer 2 and the SCH layer 4 as shown in the distribution schematic diagram of the refractive index n in FIG. Of the refractive index. That is, the refractive index profile of the SOA having the n-GaInAsP guide layer 3 is asymmetrical in the vertical direction with respect to the photoactive layer 5, so that the electric field intensity distribution of the signal light is dragged by the guide layer 3. Close to the cladding layer 2 side. However, the guide layer 3 in the present embodiment is formed in a range in which a higher-order waveguide mode in the stacking direction does not occur.

  In the SOA having such a guide layer 3, the electric field intensity profiles of the optical signals in the TE waveguide mode and the TM waveguide mode at the light exit end are as shown in FIGS. 3 (a) and 3 (b). Thus, the photoactive layer 5 is asymmetrical in the vertical direction toward the guide layer 3 side. In FIGS. 3A and 3B, the strength is indicated by contour lines.

  On the other hand, in the SOA having no guide layer 3, the refractive index profile is vertically symmetric about the photoactive layer 5, and the electric field intensity distribution of the signal light is also vertically symmetric about the photoactive layer 5. . The electric field intensity profiles of the optical signals of the TE waveguide mode and the TM waveguide mode of the SOA are as shown in FIGS. 4 (a) and 4 (b). In FIGS. 4A and 4B, the strength is indicated by contour lines.

  Semiconductor lasers having a guide layer are described in, for example, Nagashima et.al, proc. Of IEEE semiconductor laser conference 2004 and Japanese Patent Application Laid-Open No. 5-243669. In these semiconductor lasers, by concentrating the electric field intensity distribution toward the n-type cladding, light confinement in the p-type cladding layer having a high absorption coefficient is reduced, and the waveguide light loss is reduced.

  On the other hand, the present inventor has found that the waveguide optical loss reduction effect due to the insertion of the guide layer 3 has a strong polarization dependence, and thereby the distortion of the photoactive layer 5 necessary to obtain a small PDG. By reducing the amount, it is possible to increase the critical film thickness of the photoactive layer.

  Specifically, as shown in the optical waveguide mode shape in the case where the guide layer 3 is not provided, in the SOA having a flat photoactive layer structure, the TM waveguide mode photoactive layer 5 is compared with the TE waveguide mode. The light confinement is weak. In this case, the optical electric field distribution of FIG. 4B is more strongly oozed out by the upper and lower cladding layers 2 and 7 than the optical electric field distribution of FIG.

  On the other hand, according to the SOA in which the n-GaInAsP guide layer 3 is inserted, an asymmetric electric field strength distribution effect is generated, and a strong asymmetry works particularly in the TM waveguide mode. That is, in the optical waveguide mode optical electric field intensity distribution in the SOA having the guide layer 3, the optical electric field intensity distribution of the TM waveguide mode shown in FIG. 3B is larger than that in the TE waveguide mode shown in FIG. Since it is strongly close to the n-GaInAsP guide layer 3 side, it can be seen that the seepage of the optical electric field distribution to the highly absorbing p-InP cladding layer 7 is reduced. Therefore, the internal light loss (α) reduction effect of the TM waveguide mode is greater than the internal light loss (α) reduction effect of the TE waveguide mode.

Next, FIG. 6 shows the result of comparison of the internal optical loss α for each of the SOA waveguide structure with and without the guide layer 3. In FIG. 6, by adopting a structure in which the guide layer 3 is inserted into the SOA, the internal optical loss α in the TE polarization mode and the TM polarization mode is reduced. Further, the magnitude relationship between the internal optical losses α _TE and α _{TM in} the TE polarization mode and the TM polarization mode is reversed by inserting the guide layer 3, and the internal loss difference (α _TE −α _TM ) in each polarization mode. Changes in the positive direction, ie, α _TM is about 2 cm ⁻¹ smaller than α _TE .

According to the PDG equation (2), the PDG changes depending on the internal optical loss difference (α _TE −α _TM ) of each polarization mode. When (α _TE −α _TM ) shifts in the positive direction, the effective gain difference (Γ _TE g _TE −Γ _TM g _TM ) is obtained in order to obtain a polarization-independent gain, that is, a PDG = 0 relationship. Must be shifted in the positive direction.

Here, the anisotropy of the strain amount ε and the material gains g _TE and g _TM of the photoactive layer 5 having an extensional strain has the relationship shown in FIG. If Γ _TE and Γ _TM do not change, in order to shift (Γ _TE g _TE −Γ _TM g _TM ) in the positive direction, the strain amount ε of the extension strain is reduced and the g _TE with respect to g _TM is reduced. The value of the material gain coefficient ratio (g _TE / g _TM ) needs to be close to 1 by increasing the value.

Therefore, when the guide layer 3 is inserted in the polarization-independent SOA, the optimum distortion amount ε for obtaining the polarization-independent characteristic is changed in a decreasing direction. Further, the optimum strain amount ε is also decreased by changing the optical confinement ratio (Γ _TE / Γ _TM ) of the photoactive layer 5 between the TE and TM polarized waves by insertion of the guide layer 3. Also in this case, there is a relationship of Γ _TE > Γ _TM , and the insertion of the guide layer 3 makes the value of Γ _TM relatively large.

  FIG. 7 shows the result of comparing the optimum strain amount of the stretching strain that can realize polarization independence in each of the cases where the guide layer 3 is present and not present. According to FIG. 7, it can be seen that the optimum strain amount is reduced by about 10% by adopting the guide layer 3.

  From the above, in the SOA having the structure to which the present embodiment is applied, the required strain amount of the photoactive layer 5 can be reduced, so that the critical film is compared with the comparative example shown in FIGS. 4 (a) and 4 (b). The thickness can be increased. Therefore, it is possible to realize a highly efficient polarization-independent SOA by increasing the thickness of the photoactive layer 5 without causing deterioration of characteristics due to deterioration of crystal quality.

FIG. 8 shows that the magnitude of the effect of reducing the internal loss coefficient α of the polarization-independent SOA depends on the thickness of the n-GaInAsP guide layer 3. According to FIG. 8, the internal loss coefficient decreases as the guide layer 3 is thickened. When the thickness of the guide layer 3 is 500 nm, a sufficiently large internal loss reduction effect of about 5 cm ⁻¹ or more can be obtained.

Further, the internal loss coefficient α is greatly reduced by increasing the thickness of the guide layer 3 from 0 nm to 500 nm. However, when the thickness is greater than 500 nm, the reduction rate of the internal loss coefficient α is decreased and tends to be saturated. Become. This means that if the thickness of the guide layer 3 is 500 nm or more, even if a deviation occurs in the thickness of the guide layer 3, the fluctuation of the internal loss coefficient α is small and can be set to a stable value. Therefore, when the internal loss coefficient α due to the insertion of the guide layer 3 is reduced, the film thickness is set to 500 nm or more.

FIG. 9 shows that the internal loss coefficient difference between polarizations (α _TE −α _TM ) depends on the thickness of the guide layer 3. According to FIG. 9, in the structure of the comparative example shown in FIGS. 4A and 4B, the thickness of the guide layer 3 is 0 nm, the internal loss of the TM polarization mode is large, and (α _TE − α _TM ) is a negative value. On the other hand, when the thickness of the guide layer 3 is 500 nm or more, preferably 600 nm or more, the magnitude relationship between α _TE and α _TM is reversed and (α _TE −α _TM ) is changed to a positive value. In addition, the amount of distortion ε required for making the SOA polarization independent is reduced.

  According to the relationship between FIG. 8 and FIG. 9, by forming the n-GaInAsP guide layer 3 to a thickness of 500 nm or more, preferably 600 nm or more, more preferably 630 nm or more, a sufficiently large internal loss coefficient reduction effect can be obtained. Inversion occurs in the magnitude relationship between the TE and TM polarization modes of the internal loss factor. Therefore, it can be seen that by inserting the guide guide 3 having such a film thickness, the strain amount ε of the extension strain of the photoactive layer 5 necessary for making the polarization independence is reduced.

  Even when the thickness of the guide layer 3 exceeds 500 nm shown in FIG. 8, there is no particular problem, so there is no upper limit value for the thickness of the guide layer 3. Further, since the internal loss of the optical waveguide is reduced, the gain of the SOA is increased and the noise figure is reduced, so that the SOA with good characteristics is realized.

  In the polarization independent SOA according to the above embodiment, the optical confinement factor of the TM polarization mode is increased and the TE and TM polarization modes are compared with the polarization independent SOA not having the guide layer 3. The loss factor in the waveguide is reduced. Furthermore, according to the polarization independent SOA of this embodiment, the waveguide internal loss coefficient in the TE polarization mode is larger than the waveguide internal loss coefficient in the TM polarization mode. Therefore, according to such a structure, even if the strain amount of the extension strain MQW photoactive layer 5 is reduced and the extension strain MQW photoactive layer 5 is made thicker than that of the SOA without the guide layer 3, no polarization is generated. Dependence can be achieved. Thereby, a polarization-independent SOA with low power consumption can be realized.

  The structure of the polarization independent SOA having the BH structure as described above is not limited to the above structure, and other means and structures may be adopted. For example, a p-type InP substrate may be used for the substrate. In this case, a p-InP cladding layer is formed from the photoactive layer 5 to the substrate side. The upper electrode side becomes a cathode electrode, and an n-GaInAsP guide layer and an n-InP clad layer are formed on the optical waveguide layer 5.

  However, in the structure using the p-type InP substrate, if the n-GaInAsP guide layer and the n-InP cladding layer with sufficient thickness are not formed, the optical waveguide mode distribution is likely to be affected by the upper electrode, so Thus, the structure using the n-type InP substrate has a higher degree of design freedom.

  In the above embodiment, the polarization-independent SOA having the extension strain MQW photoactive layer is formed. However, the structure of the photoactive layer is not limited to this, and the extension strain bulk photoactive layer or the quantum dot photoactive layer is not limited thereto. It is possible to apply a guide layer under the above conditions. Various structures can be applied to the presence or absence of the SCH layer.

Regarding the arrangement of the n-GaInAsP guide layer, in the above structure, the SCH layer and the n-GaInAsP guide layer are directly laminated, but the present invention is also applicable to a structure in which an n-InP clad layer is inserted between them. However, in this case, in the conduction band potential distribution of the n-InP cladding layer and the n-GaInAsP guide layer, there is a possibility that a barrier may be generated to hinder electrical conduction. Therefore, a structure in which the n-InP cladding layer is not inserted is preferable.

(Second Embodiment)
10A to 10D are cross-sectional views illustrating a process of forming a polarization-independent semiconductor optical amplifier (SOA) having a ridge waveguide structure according to the second embodiment.

  Next, steps required until a structure shown in FIG. 10A is formed will be described. Note that elements such as gallium, indium, phosphorus, and arsenic that represent the compound semiconductor layer are represented by element symbols Ga, In, P, As, and the like.

  First, an n-InP clad layer 22 having a thickness of about 200 nm, an n-GaInAsP guide layer 23 having a thickness of about 500 nm or more, for example, 1000 nm, and an i-GaInAs stretching strain having a thickness of about 110 nm are formed on the n-type InP substrate 21. Bulk photoactive layer 25 is formed. Further, a p-InP cladding layer 27 having a thickness of about 2000 nm and a p-GaInAs contact layer 28 having a thickness of about 300 nm are formed on the extension strain bulk photoactive layer 25. The n-InP cladding layer 22 to the contact layer 28 are epitaxially grown in order using, for example, MOVPE.

  The tensile strain bulk photoactive layer 25 is adjusted to a strain amount of, for example, 0.89% by adjusting the composition of Ga, In, and As, but is not limited thereto.

  The composition of Ga, In, As, and P of the n-GaInAsP guide layer 23 is adjusted so that the energy band gap wavelength is intermediate between the stretch-strained bulk photoactive layer 25 and the n-InP cladding layer 22. The energy band gap wavelength of the composition of the guide layer 23 is set to 1.05 μm, for example.

  After the contact layer 28 is formed, the n-InP substrate 21 is transferred to a CVD apparatus, and a silicon oxide film having a thickness of, for example, about 200 nm is formed as a dielectric film 29 on the upper surface of the contact layer 28 in the chamber. Further, after the n-InP substrate 21 is taken out from the CVD apparatus, a photoresist is applied on the dielectric film 29, and this is exposed and developed, whereby a waveguide-shaped resist pattern 30 having a stripe width of 2.0 μm. Form.

Next, steps required until a structure shown in FIG. 10B is formed will be described.
First, the dielectric film 29 is etched using the resist pattern 30 as a mask to form a hard mask 29 a made of the dielectric film 29. For the etching of the dielectric film 29, for example, dry etching such as reactive ion etching (RIE) method using a fluorine-based gas is adopted, but wet etching using hydrofluoric acid may be used.

  Subsequently, dry etching is performed from the contact layer 28 in the region exposed from the hard mask 29a to the middle of the p-InP cladding layer 27 to form two recesses 21a, whereby a width of 2.0 μm sandwiched between the recesses 21a, A ridge waveguide structure 26 having a depth of 2.0 μm is formed. As the dry etching in this case, for example, ICP / RIE method is used. Further, a chlorine-based gas is used as an etching gas. Here, when the ridge structure is formed, the photoactive layer 25 is not etched, but the depth to which the p-InP cladding layer 27 is etched is not particularly limited, and other structures can be implemented. .

  10C, after removing the hard mask 29a, a dielectric layer 31 of a polymer material such as benzocyclobutene (BCB), a polyimide organic compound, an epoxy organic compound, or an acrylic organic compound. Is filled into the recesses 21a on both sides of the ridge waveguide to flatten the upper surface.

  Further, after polishing the back surface of the n-InP substrate 21, the cathode electrode 42 is formed on the back surface of the n-InP substrate 21, and the anode electrode 43 is formed on the p-GaInAs contact layer 28. This completes the wafer process.

  Thereafter, an n-InP wafer as the n-InP substrate 1 is cleaved in an array, and if necessary, an input / output end face of the element is coated with an AR coating with a dielectric multilayer film, and chips are separated to obtain a polarization-independent SOA. Complete.

  The polarization-independent SOA manufactured by the above procedure has the TE polarization mode and TM polarization mode distribution shown in FIGS. 11A and 11B, and FIG. Compared to the polarization-independent SOA having a structure without the guide layer 23 shown in FIG. 12B, the waveguide mode of each polarization is closer to the n-InP cladding layer 22 side.

  Further, the polarization independent SOA according to the present embodiment has an anisotropic internal loss reduction effect with respect to the TE polarization mode and the TM polarization mode shown in FIG. 13 as in the first embodiment. As a result, as shown in FIG. 14, the gain independent of the polarization can be realized by the stretch strain bulk photoactive layer 25 having a strain amount of about 4% smaller than that of the structure without the guide layer 23. The critical film thickness can be increased. In addition, since the internal loss is reduced, the effect of increasing the gain and reducing the noise figure is obtained, and an SOA element having excellent characteristics is realized.

  Note that the polarization-independent SOA having a ridge waveguide structure is not limited to the above structure, and other means and structures may be employed. For example, a p-type InP substrate may be used for the substrate. In this case, a p-InP clad layer is formed on the p-type InP substrate side from the stretched strain bulk photoactive layer 25, a cathode electrode and an n-InP clad layer are formed on the opposite side of the substrate, and the n-GaInAsP guide layer is stretched. Formed above the strained bulk photoactive layer.

  In the present embodiment, the SOA having an extension strain bulk photoactive layer has been described. However, the photoactive layer structure is not limited to the above, and for example, a quantum dot photoactive layer may be adopted. In the present embodiment, a structure in which an SCH layer is formed above and below the photoactive layer 25 may be employed as in the first embodiment.

  In the present embodiment, since the optical waveguide centering on the photoactive layer 5 needs to have only a single waveguide mode for each of TE and TM polarization modes, the photoactive layer 5 and the guide layer 3 are used as cores. A structure in which a plurality of guided modes exist and optical power transition occurs between the modes is not adopted.

  All examples and conditional expressions given here are intended to help the reader understand the inventions and concepts that have contributed to the promotion of technology, such examples and It is interpreted without being limited to the conditions, and the organization of such examples in the specification is not related to showing the superiority or inferiority of the present invention. While embodiments of the present invention have been described in detail, it will be understood that various changes, substitutions and variations can be made thereto without departing from the spirit and scope of the invention.

Next, features of the embodiment of the present invention will be described.
(Appendix 1) An n-type InP cladding layer formed above the InP substrate, a p-type InP cladding layer formed above the InP substrate, and between the n-type InP cladding layer and the p-type InP cladding layer And a thickness of 500 nm or more between the photoactive layer and the n-type InP clad layer, and a size between InP and the photoactive layer. And an n-type guide layer formed of a compound semiconductor having an energy bandgap wavelength.
(Supplementary Note 2) The compound semiconductor of the n-type guide _layer, Appendix 1, characterized in that the _{Ga x In 1-x As y} P 1-y layer (0 <x <1,0 <y <1) 2. A polarization-independent semiconductor optical amplifying device according to 1.
(Supplementary Note 3) An optical / carrier separation confinement layer made of a material having an energy band gap wavelength between the photoactive layer and the n-type guide layer between the photoactive layer and the n-type guide layer The polarization-independent type semiconductor optical amplifier according to appendix 1 or appendix 2, characterized in that is formed.
(Supplementary note 4) Any one of the supplementary notes 1 to 3, wherein the photoactive layer and the light / carrier separation confinement layer are formed in contact with the n-type guide layer. Polarization-independent semiconductor optical amplifier.
(Supplementary note 5) Any one of Supplementary notes 1 to 4, wherein the optical waveguide centering on the photoactive layer in the previous period has a single waveguide mode in each of the TE polarization direction and the TM polarization direction. The polarization-independent semiconductor optical amplifier according to one.
(Supplementary note 6) Any one of Supplementary notes 1 to 5, wherein the InP substrate is n-type, and the n-type guide layer is formed closer to the InP substrate with respect to the photoactive layer. 2. A polarization-independent semiconductor optical amplifying device according to 1.
(Supplementary Note 7) The photoactive layer, at least a part of the p-type InP cladding layer, and at least a part of the n-type InP cladding layer have a stripe shape and are sandwiched by buried heterojunction current blocking layers from both sides. The polarization-independent semiconductor optical amplifier according to any one of appendices 1 to 6, characterized in that
(Supplementary note 8) The polarization-independent semiconductor optical amplifier according to supplementary note 7, wherein the photoactive layer has a width larger than a thickness.
(Supplementary Note 9) The stripe shape is formed in the photoactive layer, the n-type InP clad layer, the p-type InP clad layer, and the n-type guide layer, and has a width taper structure in which the waveguide width changes at the end. And a waveguide mesa structure sandwiched between the buried heterojunction current blocking layers on the photoactive layer, the n-type InP cladding layer, the p-type InP cladding layer, and the entire n-type guide layer. The polarization-independent semiconductor optical amplifier according to appendix 7, which is formed.
(Supplementary Note 10) Of the n-type InP clad layer and the p-type InP clad layer, a layer higher than the InP substrate has a convex ridge structure in a region along the optical waveguide. The polarization-independent semiconductor optical amplifier according to any one of appendices 1 to 6.

1 n-type InP substrate 2 n-type InP clad layer 3 n-type GaInAsP guide layer 4 SCH layer 5 extension strain MQW photoactive layer 6 SCH layer 7 p-type InP clad layer 8 contact layer 11 current blocking layer 12 cathode electrode 13 anode electrode 21 n-type InP substrate 22 n-type InP cladding layer 23 n-type GaInAsP guide layer 25 extension strain bulk photoactive layer 27 p-type InP cladding layer 28 contact layer 31 current blocking layer 32 cathode electrode 33 anode electrode

Claims (5)

An n-type InP cladding layer formed above the InP substrate;
A p-type InP cladding layer formed above the InP substrate;
A photoactive layer formed between the n-type InP clad layer and the p-type InP clad layer and applied with a tensile strain;
An n-type formed from a compound semiconductor formed between the photoactive layer and the n-type InP cladding layer to a thickness of 500 nm or more and having an energy bandgap wavelength between InP and the photoactive layer. A guide layer,
A polarization-independent semiconductor optical amplifying device.   Between the photoactive layer and the n-type guide layer, there is formed a light / carrier separation confinement layer made of a material having an energy band gap wavelength having a size between the photoactive layer and the n-type guide layer. The polarization-independent semiconductor optical amplifier according to claim 1, wherein The InP substrate is n-type,
3. The polarization-independent semiconductor optical amplifier according to claim 1, wherein the n-type guide layer is formed closer to the InP substrate than the photoactive layer.   The photoactive layer, at least a part of the p-type InP clad layer, and at least a part of the n-type InP clad layer have a stripe shape and are sandwiched by buried heterojunction current blocking layers from both sides. The polarization-independent semiconductor optical amplifier according to any one of claims 1 to 3.   The upper layer of the n-type InP clad layer and the p-type InP clad layer with respect to the InP substrate has a convex ridge structure in a region along the optical waveguide. The polarization-independent semiconductor optical amplifier according to claim 3.