JP2009076669A - Active layer evaluation method and optical semiconductor device manufacturing method - Google Patents

Abstract

Polarization characteristics of a core layer including an active layer formed on a semiconductor substrate are measured with high sensitivity.
A core layer including an active layer formed on an upper surface of a semiconductor substrate is irradiated with excitation light, and an intensity ratio of the TE mode and TM mode of the emitted photoluminescence light is measured to measure the intensity of the active layer. In the active layer evaluation method for analyzing polarization characteristics, a semiconductor layer 8 having a refractive index different from that of the core layer 7 is formed on the core layer 7, and a slab type optical waveguide comprising the semiconductor substrate 1, the core layer 7, and the semiconductor layer 8. Configure. Since the photoluminescence light is confined in the core layer to form the optical waveguide, light amplification is performed in a long excitation region. As a result, the light intensity difference in each mode is expanded, and the light intensity difference is measured with high sensitivity.
[Selection] Figure 3

Description

  The present invention relates to a method for evaluating an active layer used in an optical semiconductor device having an optical waveguide including an active layer and a method for manufacturing the optical semiconductor device, and in particular, an active layer capable of precisely evaluating the polarization characteristics of the active layer. The present invention relates to a method for manufacturing an optical semiconductor device using the evaluation method for adjusting the formation conditions of an active layer.

  In an optical semiconductor device using a waveguide including an active layer, for example, a semiconductor optical amplifier (SOA: Semiconductor Optical Amplifier) that performs optical amplification using a striped waveguide, the polarization dependence (polarization characteristics) of the characteristics of the device, For example, it is desired that the polarization characteristic of the amplification characteristic is small. For example, a semiconductor optical amplifier used for optical communication is required to have a difference in gain between the TE mode and the TM mode of 1 dB or less. Therefore, there is a need for a method for producing and evaluating an active layer having a small polarization dependency.

  On the other hand, optical semiconductor devices using a quantum dot active layer containing quantum dots in the active layer, for example, semiconductor optical amplifiers, have excellent amplification characteristics such as a small pattern effect and a wide gain band. Development is underway. However, quantum dots may have a strong polarization dependency depending on the formation conditions, and the characteristics of an optical semiconductor device using such quantum dots have a large polarization dependency.

The polarization characteristics of quantum dots strongly depend on the shape, size, strain, material, and the like. However, these shapes and the like can be controlled to be close to a desired shape and the like by controlling the formation conditions of the quantum dots. For example, a semiconductor material having a lattice constant different from that of the lower layer is deposited thinly, a flat island-shaped semiconductor made of the semiconductor material is formed on the lower layer, and the quantum dots having a shape close to a cube are formed by stacking such island-shaped semiconductors. In this method, quantum dots having desired polarization characteristics can be manufactured by controlling formation conditions such as material, temperature, deposition amount, and number of stacked layers. (For example, refer nonpatent literature 1.)
As described above, the active layer of the optical semiconductor device is required to have a sufficiently small polarization dependency. For this reason, in the manufacture of an optical semiconductor device, a process of depositing an active layer on a semiconductor substrate, measuring and evaluating the polarization characteristics of the deposited active layer, and confirming that the prescribed polarization characteristics are satisfied is adopted. Yes. However, if the prescribed polarization characteristics are not satisfied, the active layer formation conditions (ie, deposition conditions) are changed based on the measured values so that the polarization characteristics approach the prescribed values, and the active layer formation (ie, deposition) and polarization are again performed. Repeat measurement and evaluation of characteristics. This repetition is continued until an active layer that satisfies the polarization characteristics is formed. As a result, an active layer that reliably satisfies the prescribed polarization characteristics is manufactured.

  A conventional active layer evaluation method for evaluating the polarization characteristics of the active layer deposited on the semiconductor substrate will be described below.

  FIG. 8 is a diagram for explaining the structure of a conventional evaluation sample, and shows a laminated structure of a sample on which an active layer is evaluated. 8A shows the cross-sectional structure of the core layer including the active layer formed on the upper surface of the semiconductor substrate, and FIG. 8B shows the vertical direction of FIG. 8A (the vertical direction of the upper surface of the semiconductor substrate). The refractive index distribution along is shown.

  Referring to FIG. 8A, in the evaluation sample, the core layer 7 having a structure in which the upper and lower sides of the active layer 5 are sandwiched between the light guide layers 2 and 6 is formed on the upper surface of the semiconductor substrate 1. The active layer 5 includes an intermediate layer 4 and quantum dots 3. The semiconductor substrate 1 is cleaved, and the end surface of the core layer 7 is exposed on the cleaved surface.

  Referring to FIG. 8B, the refractive index of the light guide layers 2 and 6 is larger than that of the semiconductor substrate 1 and slightly smaller than that of the active layer 5. Accordingly, the core layer 7 has a higher refractive index than the upper and lower semiconductor substrates 1 and air, and functions as a slab type optical waveguide layer.

  Such quantum dots 3 are manufactured as follows, for example. A semiconductor material having a lattice constant significantly different from that of the semiconductor substrate 1 and the lower light guide layer 2 is deposited in a flat island shape, and an intermediate layer 4 having a lattice constant close to that of the semiconductor substrate 1 is deposited thereon. Further, by stacking a plurality of island-shaped semiconductor materials and intermediate layers 4 alternately, columnar quantum dots 4 in which flat island-shaped semiconductor materials are stacked vertically are formed.

  Such columnar quantum dots 4 can be formed in a cubic shape in which the horizontal width and the vertical height are substantially equal by controlling the formation conditions. Such cubic quantum dots have approximately the same characteristics for both horizontally polarized light and vertically polarized light. For this reason, the optical semiconductor device with small polarization dependence is realized by using the active layer 5 including the quantum dots 4 on a part of the core layer 7 of the optical waveguide constituting the optical semiconductor device.

  FIG. 9 is an explanatory view of a conventional evaluation method, showing excitation light for irradiating a sample and photoluminescence light emitted from the core layer.

  Referring to FIG. 9, in the conventional method for evaluating an active layer, referring to FIG. 9, excitation light 12 (laser light) is vertically applied to core layer 7 formed on semiconductor substrate 1 from above. ) Is excited to excite the active layer 5 in the core layer 7 (the active layer 5 in FIG. 8). Then, the polarization characteristic of the photoluminescence light 20 from the active layer 5 emitted from the cleavage plane is measured. (For example, refer nonpatent literature 2.).

  The excitation light 12 is expanded into a parallel beam through a collimator and then shaped into a light beam having an elongated rectangular cross section through a slit. The excitation light 12 is applied to the upper surface of the semiconductor substrate 1 from the vertical incident direction 12a so that the long side is perpendicular to the cleavage plane.

  In the active layer 5 excited by the excitation light 12, an elongated rectangular region is excited according to the shape of the excitation light 12. The core layer 7 including the excited rectangular region functions as a waveguide having a lateral confinement effect. For this reason, the photoluminescence light 20 propagates in the core layer 7 along the long side of the rectangular region, is amplified, and is emitted perpendicularly to the cleavage plane from the cleavage plane.

  The emitted photoluminescence light 20 measures the light intensity of the TM mode having a polarization direction 20M perpendicular to the core layer 7 (or the active layer 5) and the TE mode having a polarization direction 20E parallel to the core layer 7. Then, the polarization characteristic of the photoluminescence light 20 is evaluated.

  In this conventional evaluation method, the active layer 5 is excited to be elongated. As a result, the photoluminescence light 20 generated in the active layer 5 propagates through the excited core layer 7 and is amplified. When the length of the long side of the excited rectangular region is L and the optical gain is g, the intensity of the amplified photoluminescence light is proportional to (exp (gL) −1) / g. As a result, when there is a difference in optical gain between the TE mode and the TM mode, the difference in light intensity between the two modes expands exponentially with respect to the length L of the excited region. For this reason, the difference in the optical gain between the TE mode and the TM mode of the active layer 5 appears as a significant difference in the intensity of the photoluminescence light, compared with the case where the excitation light is simply a circular cross section or a short rectangle. Therefore, the polarization characteristics of the active layer 5 can be accurately evaluated.

As a result of the evaluation, when one of the light emission in the TM mode or the TE mode is too strong, the active layer formation condition is adjusted accordingly, the active layer is formed again, and the polarization characteristics of the active layer are evaluated again. As a result of the evaluation, if the polarization characteristics are within the standard range, an active layer for an optical semiconductor device is formed on the semiconductor substrate according to the conditions for forming the active layer. As a result, an optical semiconductor device that reliably satisfies the polarization dependency is manufactured.
T.Kita, N.Tamura, and O.Wada `` Artificial control of optical gain polarization by stacking quantum dot layers '' APPLIED PHYSICS lETTERS 88,211106 (2006) Takashi KITA, Osamu WADA, Hiroji EBE, Yoshiaki NAKATA and Mitsuru SUGAWARA “Polarization-Independent Photoluminescence from Columnar InAs / GaAs Self-Assembled QuantumDots” Jpn. J. Appl. Phys. Vol. 41 (2002) pp.L1143-L1145

  As described above, in a conventional method for evaluating the polarization characteristics of an active layer, a sample in which a core layer including an active layer is formed on a semiconductor substrate is used as it is as a sample for evaluation. When manufacturing an optical semiconductor device, which is an actual product, this core layer forms a core layer of an optical waveguide of an optical semiconductor device by forming a thick cladding layer on the upper portion, and this is sufficient to confine light. Does not have a sufficient thickness.

  For example, the thickness of the active layer is the thickness of a columnar quantum dot in which InAs islands are stacked in about 10 layers with an InGaAsP intermediate layer (side barrier layer) interposed therebetween, for example, about 10 nm, and light formed thereon The thickness of the guide layer (or the cap layer that protects the active layer) is, for example, about 200 nm. Accordingly, the thickness of the core layer composed of the active layer and the upper and lower light guide layers is at most about 1 μm or less.

  When the photoluminescence light is generated by exciting the active layer in the thin core layer in this way, most of the propagating light is in a radiation mode that is diffused into the semiconductor substrate, and the light confined in the core layer is small. For this reason, even if an elongated rectangular region is excited, sufficient light amplification over the entire length of the rectangular region is not performed. In this, the polarization dependence of the optical gain of the active layer (difference in optical gain for different polarizations) does not grow exponentially with the length of the excitation region, so that the intensity difference between photoluminescent light with different polarizations Is small. Therefore, there is a problem that it is difficult to accurately evaluate the polarization dependency of the active layer or the core layer from the observation of the polarization dependency of the photoluminescence light.

  Further, in the conventional method for evaluating the active layer, after the laser beam is widened, excitation light composed of a long and narrow rectangular beam is formed through the slit. However, since the laser light is a Gaussian beam, the light intensity distribution of the excitation light also has a Gaussian distribution. For this reason, the excitation light has a large intensity distribution along the long side of the rectangular cross section.

  When excitation light having such an intensity distribution is used, an effective excitation region is limited to the vicinity of the center of the region, so that an effective amplification distance of photoluminescence light is shortened. For this reason, there is a problem that it is difficult to accurately evaluate the polarization dependency of the active layer.

  The present invention provides an active layer evaluation method capable of precisely evaluating the polarization dependency of the active layer, and a method of manufacturing an optical semiconductor device having a small polarization dependency characteristic using the evaluation method. Objective.

  The active layer evaluation method of the first configuration of the present invention for solving the above-described problem forms a semiconductor layer on a core layer including an active layer formed on an upper surface of a semiconductor substrate. The semiconductor substrate and the semiconductor layer have a refractive index smaller than that of the core layer, and form a slab type optical waveguide including the semiconductor substrate, the core layer, and the semiconductor layer. Then, the active layer is irradiated with excitation light, and the polarization characteristics of the photoluminescence light emitted from the end face of the active layer are analyzed.

  In the first configuration, a semiconductor layer is formed on the core layer, and a semiconductor substrate having a refractive index lower than that of the core layer, a slab type optical waveguide including a core layer and a semiconductor layer having a refractive index lower than that of the core layer are formed. . In this structure, light is confined by the refractive index difference above and below the core layer. Therefore, the photoluminescence light propagating through the core layer is confined in the core layer and propagates, and light is amplified in the meantime. For this reason, the polarization characteristic of the active layer is expanded by optical amplification of the photoluminescence light, and the polarization characteristic of the active layer is measured with high sensitivity. As a result, the polarization characteristics can be accurately evaluated.

  The semiconductor layer needs to have a thickness sufficient to confine the photoluminescence light in the core layer and a refractive index difference with respect to the core layer together with the semiconductor substrate. That is, the semiconductor substrate and the semiconductor layer sandwiching the core layer from above and below have a refractive index difference sufficient to form a slab waveguide having the core layer as an optical waveguide, and are necessary for light confinement in the core layer. Has a thickness. Note that the refractive indexes of the core layer, the semiconductor substrate, and the semiconductor layer may be of any magnitude as long as they form a waveguide. For example, the semiconductor substrate and the semiconductor layer may have a refractive index smaller than that of the core layer and constitute a waveguide.

  Using the semiconductor material of the clad layer provided on the core layer of the optical semiconductor device as this semiconductor layer, the refractive index of the sample to be evaluated for polarization characteristics and the semiconductor device have the same distribution. This is preferable from the viewpoint of accurately evaluating the polarization characteristics. Of course, in order to increase the polarization dependency of the active layer and perform a sharp measurement, the semiconductor layer can be formed using a semiconductor material having a greater light confinement effect.

  The core layer may be composed of only the active layer. Moreover, you may have a light guide layer in one or both of the upper and lower sides of an active layer. Furthermore, a thin film for protecting the active layer may be formed on the active layer.

  The active layer is not particularly limited in structure and material as long as it is excited by excitation light and amplifies the light. For example, in addition to an active layer including quantum dots, an active layer having a quantum well structure or a bulk active layer may be used. In particular, quantum dots, quantum well structures, or strained active layers tend to have polarization dependence, and precise evaluation of their polarization characteristics reduces the polarization dependence of the active layer by adjusting the formation conditions of the active layer. Therefore, it is suitable for applying the present invention.

  In the first configuration, the excitation light is preferably a light beam having a rectangular cross section having a long side perpendicular to the end face of the active layer. By setting the excitation light in such a cross-sectional shape, the excited region becomes a line segment, so that the photoluminescence light is confined in this line segment region and propagates in this region. Therefore, since the amplification of the photoluminescence light is made effective, the polarization dependency of the active layer is sufficiently expanded, and the polarization dependency of the active layer can be accurately evaluated. In addition, since excitation light excites a long line segment area | region uniformly, it is desirable that especially the light intensity distribution of a line segment direction is uniform.

  Such excitation light having a uniform light intensity distribution and an elongated rectangular cross section can be formed as follows.

  First, the diameter of the laser beam is expanded through a beam expander. Next, the laser light whose diameter has been expanded is incident on a beam shaper that converts incident light having a Gaussian intensity distribution into a uniform intensity distribution, and converted into a light beam having a uniform intensity distribution. Next, a light beam having a uniform intensity distribution is incident on the cylindrical lens to form excitation light composed of a light beam having a rectangular cross section.

  The beam shaper used here converts a light beam having a uniform light intensity distribution while transmitting the incident Gaussian beam of parallel rays. Therefore, the emitted light can be condensed into a line segment (elongated rectangular cross section) using a cylindrical lens.

  According to this method, the laser light having a circular cross section is limited by the slit to form a rectangular cross section, and compared with the conventional method of forming the excitation light, there is little or no loss of light due to the slit, and thus a high light intensity that is easily uniform. Luminous excitation light is formed.

  Another method of forming excitation light having a rectangular section with uniform intensity is a method using a diffusion plate that converts a light beam having a Gaussian intensity distribution into a light beam having a square section having a uniform light intensity distribution.

  That is, the laser beam is incident on the diffuser plate, the laser beam is converted into a light beam having a square cross section having a uniform light intensity distribution, and the converted light beam is incident on a cylindrical lens to generate a light beam having a rectangular cross section. Excitation light consisting of If the converted light beam is divergent light, it is converted into parallel rays and then incident on the cylindrical lens.

  Even with this method, since there is little or no loss due to the slit, it is possible to obtain excitation light with high light intensity (high brightness) having uniform light intensity.

  The second configuration of the present invention relates to a method of manufacturing an optical semiconductor device having excellent polarization characteristics with small polarization dependence, and in particular, a step of forming a core layer including an active layer on the upper surface of a semiconductor substrate, A step of evaluating the polarization characteristics, and a step of adjusting the formation conditions of the active layer based on the evaluation.

  In the second configuration, the method for evaluating the active layer having the first configuration described above is used in the step of evaluating the polarization characteristics of the active layer.

  As described above, the polarization characteristic of the active layer formed on the semiconductor substrate can be accurately evaluated by the active layer evaluation method having the first configuration. Therefore, it is easy to control the formation conditions of the active layer so that the polarization characteristics of the active layer fall within the required range. For this reason, an optical semiconductor device in which the polarization dependency is reliably within a required range is manufactured.

  According to the present invention, the polarization dependency of the active layer can be evaluated with high sensitivity. Therefore, since the formation conditions of the active layer having a small polarization dependency are easily determined, an optical semiconductor device having a small polarization characteristic can be reliably manufactured.

  The first embodiment of the present invention relates to a method for manufacturing an optical semiconductor device, and more particularly to a method for manufacturing a semiconductor optical amplifier (SOA).

  FIG. 1 is a manufacturing process diagram of an optical semiconductor device according to the first embodiment of the present invention, and shows a manufacturing process of a semiconductor optical amplifier. FIG. 2 is an explanatory view of the structure of an evaluation sample according to the first embodiment of the present invention, and shows the structure of a sample 30 on which an active layer to be evaluated is formed. 2A shows a cross-sectional structure of the sample 30, and FIG. 2 shows a refractive index distribution in a direction perpendicular to the upper surface of the sample 30.

  Referring to FIG. 1, the process of the first embodiment includes a condition determining process (S1 to S4) for determining the formation conditions of the active layer, and a product manufacturing process for manufacturing an actual semiconductor device ( S5 to S10).

  Referring to FIGS. 1 and 2, in the condition determination step (S1 to S4), first, test deposition of active layer 5 is performed on semiconductor substrate 1 as step S1. Next, in step S2, the InP clad layer 8 is deposited as the optical semiconductor layer 8 on the test-deposited active layer 5, and a sample 30 for evaluation is produced. The clad layer 8 is an example of the above-described semiconductor layer, and the same reference numerals as those of the semiconductor layer 8 are given to the clad layer 8 for the sake of clarity. Hereinafter, details of Step S1 and Step S2 will be described.

  Referring to FIG. 2A, in step S1, first, a light guide layer 2 made of InGaAsP having a thickness of, for example, 150 nm is deposited on a semiconductor substrate 1 made of n-type InP. Next, an active layer 5 was deposited, and a light guide layer 6 made of, for example, InGaAsP having a thickness of 150 nm was deposited thereon. Both the active layer 2 and the light guide layers 2 and 6 were deposited using MOCVD (metal organic chemical deposition).

  The active layer 5 is composed of quantum dots 3 and an intermediate layer 4 that fills between the quantum dots 3. The active layer 5 is formed by alternately depositing an extremely thin (for example, monoatomic layer or less) InAs layer and an InGaAsP layer alternately a predetermined number of times (for example, nine InAs layers).

  Since InAs has a lattice constant significantly different from that of the underlying light guide layer 2 or the intermediate layer InGaAsP, a thin InAs layer is deposited in an island shape. The InGaAsP deposited thereon covers the island-like InAs and is deposited almost flatly. The island-shaped InAs in the second and higher layers are spontaneously formed immediately above the lower island-shaped InAs. Accordingly, by alternately depositing InAs layers and InGaAsP layers a plurality of times, quantum dots 3 in which a plurality of island-like InAs are vertically stacked via thin InGaAsP layers are formed. On the other hand, the InGaAsP layer deposited between adjacent quantum dots 3 is deposited flat to form an intermediate layer 4 that separates the quantum dots 3. Thereby, an active layer 5 having a structure in which the quantum dots 3 are separated by the intermediate layer 4 is formed. Although the thickness of the active layer 5 varies depending on the formation conditions of the quantum dots 3, for example, the quantum dots 3 having a height of approximately 13 nm, that is, the active layer 5 having a thickness of 13 nm is formed by depositing nine InAs layers.

  Next, in step S2, as described above, the light guide layer 6 made of InGaAsP having a thickness of, for example, 150 nm to 350 nm is deposited on the active layer 5. By this step, the core layer 7 having a structure in which the active layer 5 is sandwiched from above and below by the light guide layers 2 and 6 is formed.

  Next, in step S2, a clad layer 8 made of InP as a semiconductor was deposited on the core layer 7 using MOCVD. The thickness of the clad layer 8 is set to a thickness sufficient for optical confinement in the core layer 7, for example, 3 μm. If the clad layer 8 is thin, the light confinement effect is small, and the polarization characteristics of the active layer 5 are not sufficiently reflected in the polarization characteristics of the photoluminescence light, but also the polarization characteristics of the photoluminescence light due to the variation in the thickness of the clad layer 8. This is not preferable because changes occur. Conversely, if it is too thick, there is a problem that the deposition time is long. Accordingly, the thickness of the clad layer 8 is preferably 1 μm or more from the viewpoint of the light confinement effect and measurement stability, and 5 μm or less from the viewpoint of the deposition time.

  Referring to FIG. 2B, the sample 30 manufactured by the above-described steps S1 to S2 has a higher refractive index of the core layer 7 than the semiconductor substrate 1 made of InP and the cladding layer 8 made of InP, Photoluminescence light is effectively confined in the core layer 7. The light guide layers 2 and 6 have substantially the same refractive index as that of the active layer 5 so that light is confined in the core layer and propagates.

  Next, referring to FIG. 1, in step S3, the polarization characteristics of the active layer are evaluated.

  FIG. 3 is a perspective view for explaining the evaluation method according to the first embodiment of the present invention, and shows excitation light incident on the sample and photoluminescence light emitted from the sample.

  Referring to FIG. 3, excitation light 12 is incident on the upper surface of cladding layer 8 (semiconductor layer 8) formed on the upper surface of sample 30 from an incident direction 12 a perpendicular to the upper surface of sample 30. The excitation light 12 was produced by the same method as the excitation light in the above-described conventional active layer evaluation method. That is, after the laser beam is widened and made into a parallel beam through a collimator, it is passed through a slit and shaped into a light beam having an elongated rectangular section, for example, a rectangular section having a width of about 10 μm and a length of 400 μm. did. The excitation light 12 has a Gaussian light intensity distribution in the long side direction of the rectangle.

  The excitation light 12 is incident so that the long side of the light beam having a rectangular cross section is perpendicular to the cleavage plane 1a. Accordingly, the core layer 7 (more precisely, the active layer 5 included in the core layer 7) under the cladding layer 8 is excited only in the linear region extending perpendicular to the cleavage plane 1a. An optical waveguide through which light propagates along the linear region is formed. The left and right sides of the linear region are confined by the refractive index difference between the excited region and the non-excited region, and the upper and lower regions are confined by the semiconductor substrate 1 and the cladding layer 8 due to the refractive index difference.

  Photoluminescence light is generated in the linear region of the excited active layer 5. This photoluminescence light is amplified while propagating through the linear region, and is emitted perpendicularly from the cleavage plane 1a as photoluminescence light 20 emitted from the end face of the active layer 5 exposed on the cleavage plane 1a.

  The photoluminescence light 20 radiated from the cleavage plane 1a includes a TE mode light having a polarization direction 20E parallel to the plane of the active layer 5 and a TM mode having a polarization direction 20M perpendicular to the plane of the active layer 5. Including light. In step S3, the light intensity ratio between the TE mode and the TM mode is measured using, for example, a polarizing plate and a spectroscope. The light intensity of these TE mode and TM mode depends on the optical gain for the TE mode and TM mode polarization of the active layer, respectively. Therefore, the light intensity ratio between the TE mode light and the TM mode light can be measured, and the polarization characteristics of the active layer 5 can be evaluated from the light intensity ratio.

  Next, in step S4, it is evaluated whether or not the light intensity ratio measured in step S3 is within a specified value. If it is within the specified value, a product manufacturing process (S5 to S10) described later is started. When the light intensity ratio exceeds the specified value range, the condition determination process from step S1 to step S4 is repeated again. The repetition of steps S1 to S4 is repeated until the light intensity ratio measured in step S4 is within a specified value.

  In the manufacturing determination process (S1 to S4) repeated again, the formation conditions of the active layer 5 and the intermediate layer (for example, the deposition temperature, the deposition layer thickness of the InAs layer), the number of InAs layers, and the formation conditions of the intermediate layer 4 in the process S1. (For example, the deposition temperature, the composition ratio of the intermediate layer, and the layer thickness) are adjusted so that the polarization characteristic of the active layer 5 approaches the specified value.

  For example, when a light intensity greater than that of the TM mode photoluminescence light 20 is measured in the TE mode photoluminescence light 20, the deposition temperature is lowered, the thickness of the InAs layer is increased, or the number of InAs layers is increased. To do. By adjusting the formation conditions of these active layers 5, the ratio of the height to the diameter of the quantum dots is increased to increase the optical gain for the TM mode.

  Next, when the light intensity ratio measured in step S4 is within a specified value, for example, within ± 0.5 dB, a product manufacturing process (S5 to S10) for manufacturing an optical semiconductor device as a product is started. This product manufacturing process (S5 to S10) is the same as the manufacturing process of a normal optical semiconductor device (in this embodiment, a semiconductor optical amplifier having an optical waveguide structure).

  First, in step S5, referring to FIG. 2, the light guide layer 2, the quantum dots 3, and the active layer 5 including the intermediate layer 4, the light guide layer 6, and the cladding layer 8 are formed in this order on the semiconductor substrate 1. The core layer 7 composed of the light guide layer 2 to the light guide layer 6 is formed under the formation conditions in which the light intensity ratio is within a specified value, that is, the formation conditions of the test deposition step S1 of the immediately preceding active layer. The clad layer 8 may have the same formation conditions as in the immediately preceding step S2. If necessary, the deposition conditions of the cladding layer 8, for example, the thickness of the cladding layer 8 can be changed. Such a change in the formation conditions of the cladding layer 8 does not significantly affect the polarization characteristics of the active layer 5.

  Next, in step S6, the light guide layer 2 to the clad layer 8 are etched from the upper surface of the clad layer 8 using photoetching to form mesa stripes on which the semiconductor substrate 1 is exposed. The core layer 7 included in the mesa stripe becomes an optical waveguide of the semiconductor optical amplifier.

  Next, in step S7, an insulating buried layer for embedding the mesa stripe is formed on the semiconductor substrate 1 exposed on both sides of the mesa stripe.

  Next, in step S8, a flat layer including a contact layer is formed on the mesa stripe and the buried layer.

  Next, in step S9, electrodes are formed on the contact layer and on the back surface (lower surface) of the semiconductor substrate 1 to manufacture a semiconductor optical amplifier.

  Next, in step S10, characteristic evaluation including polarization characteristics of the manufactured semiconductor optical amplifier is performed, and the semiconductor optical amplifier is shipped excluding defective products. Through the above steps, the optical semiconductor device (semiconductor optical amplifier) according to the first embodiment is manufactured.

  In the first embodiment, the clad layer 8 made of the same InP as the clad layer 8 of the optical semiconductor device is formed in the step S2 of the condition determining step (S1 to S4). However, the clad layer 8 made of a different semiconductor material from the clad layer 8 of the optical semiconductor device to be the product may be formed, and the clad layer 8 having a thickness different from that of the clad layer 8 of the optical semiconductor device to be the product is formed. May be.

  In this embodiment, the case where the core layer 7 has a higher refractive index than the semiconductor substrate 1 and the clad layer 8 has been described, but conversely, the core layer 7 has a lower refractive index than the semiconductor substrate 1 and the clad layer 8. A waveguide structure may be formed.

  In the second embodiment of the present invention, instead of the excitation light used in the first embodiment, an optical semiconductor device using a light beam having a rectangular cross section (linear cross section) having a uniform light intensity distribution as excitation light is used. Regarding the method. The second embodiment is the same as the first embodiment except that the method for producing excitation light is different from that of the first embodiment and the light intensity distribution of the excitation light is uniform. Therefore, only steps S3 and S5 will be described below for the sake of simplicity.

  FIG. 4 is a block diagram of an excitation light production apparatus according to a second embodiment of the present invention, and shows the main configuration of the excitation light production method and production apparatus. 4A to 4D represent the cross-sectional shapes of the light beams at positions A to D shown in FIG. 4, respectively.

  Referring to FIG. 4, the laser light 13 a emitted from the Ar laser 13 is incident on the beam expander 14. With reference to FIG. 4A, the beam expander 14 widens the laser beam 13a made of an incident thin Gaussian beam to form a large-diameter parallel beam having a Gaussian light intensity distribution. For example, a beam expander (product symbol ZB-128X) manufactured by Newport Corporation having a known optical system can be used as the beam expander 14.

  Next, the laser beam 13 a that has been widened and made into parallel light is incident on the beam sieve 15. The beam sieve 15 converts the incident parallel Gaussian beam into a parallel light beam 15a having a circular cross section having a uniform light intensity distribution with reference to FIG. As the beam sieve 15 having such a function, for example, a beam sieve manufactured by Newport (product symbol GBS-NIR-H) can be used.

  Next, the parallel light beam 15a having a circular cross section passes through the cylindrical lens 16, and is focused in a direction perpendicular to the paper surface of FIG. 4 with reference to FIG. 4C, and linearly extending in the paper surface. The light beam 16a is converted into a narrow rectangular cross section. The shape of the light beam 16a in FIG. 4C represents a cross-sectional shape near the focal point of the cylindrical lens 16 for easy understanding.

    Next, referring to FIG. 4D, the light beam 16 is reduced to the excitation light 12 having an elongated rectangular cross section having a width of about 10 μm and a length of about 400 μm using the convex lens 17, and the excitation light 12 Is projected onto the surface of the sample 30.

  The sample 30 is the same as the sample 30 of the first embodiment, and is arranged so that the excitation light 12 is irradiated on the surface of the cladding layer 8.

  The photoluminescence light 20 propagated through the linear region of the core layer 7 irradiated with the excitation light 12 and amplified is emitted from the end face (cleavage surface 1a) of the core layer 7. The photoluminescence light 20 is collimated by the convex lens 25 and enters the polarizing plate 26. The polarizing plate 26 alternately transmits TE mode or TM mode photoluminescence light. The photoluminescence light 20 transmitted through the polarizing plate 26 is condensed at the tip of the optical fiber 28 by the convex lens 27 and introduced into the optical fiber. Then, it is introduced into the spectroscope 29 and the spectral intensity is measured. Finally, it is determined from the light intensity ratio between the TE mode and the TM mode whether the polarization characteristics of the active layer 5 (core layer 7) satisfy the standard.

  The third embodiment of the present invention relates to a method of manufacturing an optical semiconductor device using a diffuser plate in place of the beam expander 14 and the beam shaper 15 used for forming the excitation light of the second embodiment.

  FIG. 5 is a diagram showing the configuration of a pumping light producing apparatus according to a third embodiment of the present invention, and shows the main configuration of the pumping light producing method and the producing apparatus. In addition, the figure (E)-(G) in the circle | round | yen in FIG. 5 represents the cross-sectional shape of the light beam in the position EG shown in FIG.

  With reference to FIGS. 5 and 5A, the laser beam 13 a emitted from the Ar laser 14 is incident on the diffusion plate 18. The diffusion plate 18 converts the laser beam 13a formed of a thin Gaussian beam into a light beam having a substantially square cross section having a uniform light intensity distribution. The converted light beam spreads at an angle of about 10 degrees. As the diffusion plate having such a function, for example, a diffusion plate (product name Engineered Diffusers: product symbol ED1-S20) manufactured by Thorlabor can be used.

  Next, the widened converted light beam is converted into a parallel beam 17 a using the convex lens 17. Referring to FIG. 5F, the parallel beam 17a has a uniform light intensity with a substantially square cross section.

  Next, using the cylindrical lens 16, the parallel beam 17 a is condensed in a direction perpendicular to the paper surface of FIG. 5, converted to excitation light 12 having an elongated rectangular cross section with reference to FIG. Irradiate. If necessary, similarly to the first embodiment, the excitation light 12 can be further condensed by a convex lens to form excitation light having a high light intensity.

  Subsequent measurement of the light intensity ratio of the TE mode and the TM mode of the photoluminescence light 20 is performed in the same manner as in the second embodiment.

  FIG. 6 is an explanatory diagram of the effect of the present invention, and represents the TE mode and TM mode spectra of the photoluminescence light observed in the first and second embodiments of the present invention. 6A shows the spectrum of the second embodiment, and FIG. 6B shows the spectrum of the first embodiment. For comparison, FIG. 6C shows the spectrum measured by the conventional active layer evaluation method. Spectrum was shown.

  Referring to FIG. 6B, the sample 30 in which the semiconductor layer 8 (cladding layer 8) is formed on the core layer 7 is excited by the excitation light 12 having a Gaussian light intensity distribution. In the embodiment, the light intensity ratio of the TE mode light intensity (curve B in FIG. 6) and the TM mode light intensity (curve A in FIG. 6) in the used wavelength region of the optical semiconductor device, 1500 nm to 1560 μm is approximately It was measured as 3 dB.

  On the other hand, referring to FIG. 6C, in the conventional active layer evaluation method using a sample in which the semiconductor layer 8 (cladding layer 8) is not formed, the light intensity ratio between the TE mode and the TM mode in the same wavelength range. Was 0.2 dB or less. The sample has the same layer structure as in the first embodiment except for the presence or absence of the semiconductor layer 8 (cladding layer 8). Further, the same excitation light as that in the first embodiment was used.

  As described above, in spite of using a sample having the same active layer 5 to core layer 7, the conventional active layer evaluation method evaluates the polarization dependence to be 0.2 dB or less. In the first embodiment, A large polarization dependence of 3 dB was measured. As described above, according to the first embodiment, the polarization characteristics of the active layer can be more sensitively measured using the same excitation light as compared with the conventional active layer evaluation method.

  Furthermore, according to the first embodiment of the present invention, the light intensity of the measured photoluminescence light is about 20 dB higher than that of the conventional method for evaluating an active layer. Therefore, the S / N ratio is high and the light intensity can be measured accurately.

  The difference in the photoluminescence light seen in the first embodiment of the present invention and the conventional active layer evaluation method is caused by the difference in the propagation mode of the photoluminescence light, as will be described below. Is considering.

  FIG. 7 is an explanatory diagram of the light propagation mode in the evaluation sample of the present invention, and shows the light intensity distribution of the photoluminescence light propagating through the sample obtained by calculation. FIG. 7A shows the light intensity distribution of the sample used in the first embodiment of the present invention, and FIG. 7B shows the light intensity distribution of the sample used in the conventional active layer evaluation method. The samples of FIGS. 7A and 7B are the same as the samples of FIGS. 6B and 6C, respectively, that is, only the presence or absence of the semiconductor layer 8 (cladding layer 8) is different. . 7 represents the distance from the sample surface to the semiconductor substrate 1 direction.

  Referring to FIG. 7A, in the sample of the first embodiment of the present invention, a core layer 7 having a large refractive index is formed on a semiconductor substrate 1 having a small refractive index, and the refractive index is formed on the upper surface of the core layer 7. A small semiconductor layer 8 is formed.

  In this structure, the photoluminescence light is mostly confined in the core layer 7 having a large refractive index, as indicated by the curve c in FIG. Does not penetrate much. This indicates that the photoluminescence light travels in a propagation mode in which the inside of the core layer 7 is an optical waveguide.

  In this way, since the photoluminescence light confined in the active layer (that is, in the core layer) propagates along the long and narrow rectangular excitation region, it is amplified while propagating through the long excitation region that functions as an optical waveguide. The Since the optical amplification also increases the polarization dependency of the active layer, the light intensity between the TE mode and the TM mode of the photoluminescence light is also expanded with reference to FIG. Moreover, the light intensity of the emitted photoluminescence light is high.

  On the other hand, in the sample used for the conventional evaluation method of the active layer, the semiconductor layer 8 is not provided on the core layer 7 and is directly exposed to air. In addition, in order to protect the core layer 7, a thin (for example, 200 to 300 nm) protective layer may be provided.

  In such a sample, with reference to FIG. 7 (b), the photoluminescence light (the light intensity is shown as a curve d in the figure) has a light intensity peak in the core layer 7, but the light intensity. The distribution is large and penetrates into the semiconductor substrate 1. This indicates that the photoluminescence light propagates in the core layer 7 in a radiation mode in which it is diffused and propagated in the semiconductor substrate 1.

  In this radiation mode, the photoluminescence light propagating through the core layer 7 is greatly attenuated, and light amplification is not performed. For this reason, with reference to FIG.6 (c), the polarization characteristic of an active layer does not appear as a light intensity difference between TE mode and TM mode of photo-luminescence light. Even if the same excitation light is used, the intensity of the emitted photoluminescence light is small.

  Next, referring to FIG. 6A, the light intensity difference between the TE mode and the TM mode of the photoluminescence light observed according to the second embodiment reaches approximately 6 dB. This light intensity difference is about twice as large as the light intensity difference in the first embodiment shown in FIG. Therefore, in the second embodiment, the polarization dependency of the active layer can be measured more precisely than in the first embodiment.

  The difference between the configurations of the second embodiment and the first embodiment is only the method for producing the excitation light, and the others are the same including the sample. That is, in the second embodiment, the light intensity distribution is uniform over almost the entire surface of the elongated rectangular section of the excitation light. For this reason, since it excites uniformly over the full length of an excitation area | region, a long excitation area | region functions as an optical amplifier. As described above, the light intensity difference between the TE mode and the TM mode expands exponentially with respect to the length of the pumping region that functions as an optical amplifier, so that the light intensity difference between both modes increases.

  On the other hand, in the first embodiment, since the excitation light has a Gaussian light intensity distribution, a part of the long excitation region is strongly excited, and the long excitation region cannot be excited with uniform intensity. For this reason, since the area | region which functions effectively as an optical amplifier is short, the optical intensity difference with respect to TE mode and TM mode is smaller than 2nd Embodiment. In addition, the polarization characteristics of the optical semiconductor device manufactured in the product manufacturing process matched best with the evaluations according to the second and third embodiments having the thick cladding layer 8 and being uniformly excited.

  As described above, in the second embodiment using the excitation light having a uniform light intensity distribution, the polarization dependency is evaluated with higher sensitivity than in the first embodiment using the excitation light having the Gaussian light intensity distribution. In the third embodiment using excitation light having a uniform light intensity distribution as in the second embodiment, the same light intensity difference as in the second embodiment shown in FIG. 6A was measured.

  By applying the present invention to an optical semiconductor device, for example, a method for manufacturing a semiconductor optical amplifier, an optical semiconductor device with small polarization dependence can be reliably manufactured.

First Embodiment Optical Semiconductor Device Manufacturing Process Diagram of the First Embodiment First embodiment evaluation sample structure explanatory diagram of the present invention 1st Embodiment evaluation method explanatory drawing of this invention Second embodiment of the present invention Excitation light fabrication apparatus configuration diagram Third embodiment of the present invention Excitation light fabrication apparatus configuration diagram Effect explanatory drawing of this invention Explanatory drawing of the light propagation mode in the evaluation sample of the present invention Conventional evaluation sample structure Conventional evaluation method explanatory diagram

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Semiconductor substrate 1a Cleaved surface 2, 6 Light guide layer 3 Quantum dot 4 Intermediate layer 5 Active layer 7 Core layer 8 Clad layer (semiconductor layer)
DESCRIPTION OF SYMBOLS 12 Excitation light 12a Incident direction 13 Laser 13a Laser light 14 Beam expander 15 Beam shaper 15a, 16a, 17a Light beam 16 Cylindrical lens 17 Convex lens 18 Diffusion plate 20 Photoluminescence light 20E, 20M Polarization direction 21 Air 25, 27 Lens 26 Polarizing plate 28 Optical fiber 29 Spectrometer 30 Sample

Claims (6)

A step of forming a core layer including an active layer on a semiconductor substrate; and a polarization characteristic of photoluminescence light emitted from an end surface of the active layer by irradiating the upper surface of the semiconductor substrate with excitation light for exciting the active layer And an active layer evaluation method comprising:
The semiconductor substrate has a refractive index smaller than that of the core layer;
An activity characterized in that a semiconductor layer having a refractive index smaller than that of the core layer is formed on the core layer to form a slab type optical waveguide comprising the semiconductor substrate, the core layer, and the semiconductor layer. Layer evaluation method. The method for evaluating an active layer according to claim 1, wherein a refractive index of the semiconductor substrate and the cladding layer is smaller than a refractive index of the core layer. 3. The active layer evaluation method according to claim 1, wherein the excitation light is a light beam having a rectangular cross section having a long side perpendicular to an end face of the active layer. Expanding the laser beam through a beam expander; and
Next, the incident laser light having a Gaussian intensity distribution is incident on a beam shaver that converts the incident light into a uniform intensity distribution, and the laser light is converted into a light beam having a uniform intensity distribution. ,
4. The active layer according to claim 3, further comprising a step of entering the light beam having the uniform intensity distribution into a cylindrical lens to form the excitation light composed of a light beam having a rectangular cross section. Evaluation methods. A laser beam is incident on a diffusion plate that converts a light beam having a Gaussian intensity distribution into a light beam having a uniform cross-section having a uniform light intensity distribution.
Converting the laser beam into a square-shaped light beam having a uniform light intensity distribution;
4. The activity according to claim 3, further comprising: injecting a light beam having a square cross section having a uniform light intensity distribution into a cylindrical lens to form the excitation light including the light beam having a rectangular cross section. Layer evaluation method. Manufacturing an optical semiconductor device comprising: a step of forming a core layer including an active layer on an upper surface of a semiconductor substrate; a step of evaluating polarization characteristics of the active layer; and a step of adjusting formation conditions of the active layer based on the evaluation In the method
An optical semiconductor device manufacturing method using the active layer evaluation method according to claim 1 in the step of evaluating the polarization characteristics of the active layer.

Images