JP5369737B2 - Optical communication system and manufacturing method thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a connection passage that is connected to an optical modulator having a small size with a high phase and high speed modulation, to provide an optical communication system that is composed of such optical modulator and the connection passage, and also to provide methods of manufacturing the connection passage and the optical communication system. <P>SOLUTION: The connection passage 30 connects an optical modulator and an optical waveguide installed outside the optical modulator. Inside the connecting passage 30, there are installed at least a portion of a semiconductor layer 8 that is dope-processed so as to exhibit a first conductive type and at least a portion of a semiconductor layer 9 that is dope-processed so as to exhibit a second conductive type, in a manner superimposed with a dielectric layer 11 in-between. In addition, the semiconductor layer 8 of the first conductive type, the semiconductor layer 9 of the second conductive type, and the dielectric layer 11 vary in the width and/or the height, so that they form a tapered shape or an inverted tapered shape from the optical modulator toward the optical waveguide. <P>COPYRIGHT: (C)2010,JPO&amp;INPIT

Description

  The present invention relates to a connection path connected to an optical modulator using a capacitor structure, an optical communication system, and manufacturing methods thereof.

  Optical fiber communication, which was mainly for business use, has become widespread for home use. Accordingly, a high-performance optical communication device is demanded. Optical communication devices for various optical communication systems, such as home optical fibers and local area networks (LAN), include silicon-based optical communication devices that function at 1330 nm and 1500 nm optical fiber communication wavelengths. This silicon-based optical communication device is a very promising device that enables the integration of optical functional elements and electronic circuits on a silicon platform by utilizing complementary metal oxide semiconductor (CMOS) technology. .

  As silicon-based optical communication devices, passive devices such as waveguides, optical couplers, and wavelength filters have been extensively studied. Further, active devices such as silicon-based optical modulators and optical switches are attracting much attention as important elements of the means for manipulating optical signals for the optical communication system described above. Optical modulators and optical switches that change the refractive index using the thermo-optic effect of silicon are slow, and can only be used in devices with an optical modulation frequency of 1 Mb / second or less. In order to use it in a device having a light modulation frequency larger than that, a light modulator using the electro-optic effect is required.

  Pure silicon shows no change in refractive index due to the Pockels effect, and the change in refractive index due to the Franz-Keldysh effect or the Kerr effect is very small. Therefore, many of the currently proposed optical modulators using the electro-optic effect change the real part and the imaginary part of the refractive index by changing the free carrier density in the silicon layer using the carrier plasma effect. It is a device that changes the phase and intensity of light.

The free carrier density in the electro-optic light modulator can be changed by free carrier injection, accumulation, removal, or inversion. Many of these devices studied to date have poor light modulation efficiency, require a length of 1 mm or more for optical phase modulation, and require an injection current density higher than 1 kA / cm ³ . In order to realize such a small, highly integrated and low power consumption device, it is necessary to have a device structure that can achieve high light modulation efficiency. It is possible to shorten the length. In addition, when the size of the optical communication device is large, it is easily affected by the temperature on the silicon substrate, and the electro-optic effect that should have been originally obtained by the change in the refractive index of the silicon layer due to the thermo-optic effect. It is also possible to cancel.

FIG. 12 is an example of a related technology of a silicon-based electro-optic light modulator using a rib waveguide formed on an SOI (Silicon on Insulator) substrate. On the substrate 3, a buried oxide layer 2 and an intrinsic semiconductor 1 including a rib-shaped portion are stacked in this order. A p + doped semiconductor 4 and an n + doped semiconductor 5 are formed on both sides of the rib-shaped portion of the intrinsic semiconductor 1 with an interval therebetween. The p + doped semiconductor 4 and the n + doped semiconductor 5 are formed by partially doping the intrinsic semiconductor 1 with a high concentration. The structure of the optical modulator shown in FIG. 12 is a PIN (P-intrinsic-N) diode. By applying a forward and reverse bias voltage, the free carrier density in the intrinsic semiconductor 1 is changed, and the carrier By using the plasma effect, the refractive index is changed. In this example, the electrode contact layer 6 is disposed on one side of the rib-shaped portion of the intrinsic semiconductor 1, and the above-described p + doped semiconductor 4 is formed at a position facing the electrode contact layer 6. Similarly, the electrode contact layer 6 is disposed also on the other side of the rib-shaped portion of the intrinsic semiconductor 1, and the n + doped semiconductor 5 is formed at a position facing the electrode contact layer 6. Further, the waveguide including the rib-shaped portion is covered with the oxide cladding 7. In the PIN diode structure described above, the semiconductors 4 and 5 can be highly doped so that the carrier density is about 10 ²⁰ / cm ³ .

  During the optical modulation operation, a forward bias voltage is applied to the PIN diode from the power source connected to the electrode contact layer 6, thereby injecting free carriers into the waveguide. At this time, the refractive index of the intrinsic semiconductor 1 changes due to the increase of free carriers, and thereby phase modulation of light propagated through the waveguide is performed. However, the speed of this light modulation operation is limited by the free carrier lifetime inside the rib shape of the intrinsic semiconductor 1 and carrier diffusion when the forward bias voltage is removed. An electro-optic light modulator having such a related-art PIN diode structure usually has an operating speed in the range of 10 to 50 Mb / sec when a forward bias voltage is applied. On the other hand, in order to shorten the carrier lifetime, it is possible to increase the switching speed by introducing impurities into the intrinsic semiconductor 1, but the introduced impurities have the problem of reducing the light modulation efficiency. is there. The largest factor affecting the operating speed is due to the RC time constant, and the capacitance when the forward bias voltage is applied becomes very large due to the reduction of the carrier depletion layer at the PN junction. Theoretically, high-speed operation of the PN junction can be achieved by applying a reverse bias voltage, but it requires a relatively large drive voltage or a large element size.

  As another example of the related art, the buried oxide layer 2 and the first conductivity type main body region are sequentially stacked on the substrate 3, and the main body region and the second stack are stacked so as to partially overlap the main body region. Patent Document 1 discloses a silicon-based electro-optic light modulator having a capacitor structure including a conductive type gate region and having a thin dielectric layer 11 formed at the laminated interface. In the following, “thin” is intended to be in the submicron order (less than 1 μm).

  FIG. 13 shows a silicon-based electro-optic light modulator having a SIS (silicon-insulator-silicon) structure according to the related art. The electro-optic light modulator is formed on an SOI substrate including a substrate 3, a buried oxide layer 2, and a main body region, and the main body region is formed by doping a silicon layer of the SOI substrate, and a p-doped semiconductor 8 is formed. The p + doped semiconductor 4 is formed by doping at a high concentration and an electrode contact layer 6. The gate region is composed of an n-doped semiconductor 9 formed by doping a thin silicon layer stacked on an SOI substrate, an n + -doped semiconductor 5 formed by doping at a high concentration, and an electrode contact layer 6. Yes. The gap between the buried oxide layer 2 and the main body region and the gate region and the upper portion of the main body region and the gate region have an oxide cladding 7.

  In the doped region, the carrier density change is controlled by the external signal voltage. Further, when a voltage is applied to the electrode contact layer 6, free carriers are accumulated, removed, or inverted on both sides of the dielectric layer 11. As a result, optical phase modulation is performed. For this reason, it is desirable that the optical signal electric field and the region in which the carrier density is dynamically controlled externally are matched.

JP 2006-515082 A

  According to Patent Document 1, although optical phase modulation is possible, the thickness of the region where the carrier density dynamically changes is very thin, about several tens of nanometers. A modulation length is required, the size of the optical modulator is increased, and high-speed operation is difficult. Therefore, in a silicon-based optical modulator that can be integrated on a silicon substrate, low-cost, low current density, low power consumption, high modulation depth, low voltage drive, and high-speed modulation have a thickness of submicron order (1 μm) It is difficult to realize an optical modulator based on the carrier plasma effect that can be realized in the region of

  In addition, in an optical modulator coupled to an optical waveguide provided outside the optical modulator, the optical loss at the coupling portion is not only the increase in the insertion loss of light into the optical modulator, but also the light from the optical modulator. There is a possibility of reducing the modulation efficiency. For this reason, a connection path for reducing optical loss is provided between the optical waveguide and the optical modulator. For example, as shown in FIG. 29 and FIG. 30 of Patent Document 1, in the input / output portion of the optical modulator, a single taper (input increase) is provided for each of the input portion and the output portion of each of the silicon layers stacked in two stages. By arranging a taper, an output decrease taper, an input decrease taper, or an output increase taper), the optical loss at the coupling portion between the optical waveguide and the optical modulator is reduced. In order to perform high-efficiency optical modulation, not only the performance of the optical modulator itself is improved, but also an optical waveguide provided outside the optical modulator and the optical modulator using a connection path suitable for the optical modulator. It is better to connect. That is, even if a small-sized optical modulator having a high phase and a high degree of modulation is realized, if it is not connected to the optical waveguide using a connection path suitable for it, the optical loss is large and the optical modulation is deteriorated.

  The object of the present invention is to solve the above-mentioned problem that the optical modulator having a small, high-phase and high-speed modulation degree is directly connected to the optical waveguide, resulting in a large optical loss and a poor optical modulation degree. It is to provide a connection path connected to the optical modulator and the optical waveguide, an optical communication system constituted by such an optical modulator and the connection path, and a method for manufacturing them.

  The connection path of the present invention connects an optical waveguide provided outside the optical modulator and the optical modulator. In the connection path, at least a part of the semiconductor layer doped to exhibit the first conductivity type and at least a part of the semiconductor layer doped to exhibit the second conductivity type are dielectric layers. It is provided so as to overlap each other. Further, the width and / or height of the first conductive type semiconductor layer, the second conductive type semiconductor layer, and the dielectric layer from the optical modulator toward the optical waveguide have a tapered shape or an inversely tapered shape. Is changing.

  According to the present invention, an optical modulator having a small size, a high phase, and a high degree of modulation can be connected to an optical waveguide using a connection path suitable for the optical modulator, thereby suppressing optical loss and increasing optical modulation efficiency. Can be realized.

It is the schematic of an example of the optical modulator used in this invention. It is AA 'sectional drawing of the optical modulator in FIG. FIG. 2 is a cross-sectional perspective view of a main part of an embodiment of a connection path that connects the optical modulator in FIG. 1 and an optical waveguide provided outside the optical modulator. It is the schematic of another example of the optical modulator used in this invention. FIG. 5 is a cross-sectional perspective view of a main part of an embodiment of a connection path that connects the optical modulator in FIG. 4 and an optical waveguide provided outside the optical modulator. It is the schematic of another example of the optical modulator used in this invention, (a) is the figure seen from the propagation direction of light, (b) is BB 'sectional drawing in (a), (c) is ( It is CC 'sectional drawing in a). It is a principal part cross-section perspective view of one Embodiment of the connection path which connects the optical modulator in FIG. 6, and the optical waveguide provided in the exterior. It is the figure which showed the manufacturing process of the connection path in FIG. FIG. 9 is a diagram illustrating a manufacturing process of the connection path continued from FIG. 8. It is the figure which showed the relationship between the optical modulation length and optical phase shift amount in the optical modulator used in this invention, and the optical modulator of a prior art. It is the figure which showed the relationship between the carrier density and frequency band in the optical modulator used in this invention, and the optical modulator of a prior art. It is a schematic block diagram of an example of the optical modulator of a prior art. It is a schematic block diagram of another example of the optical modulator of a prior art.

  Embodiments of the present invention will be described below with reference to the accompanying drawings. In addition, the same number is attached | subjected to the structure which has the same function in an accompanying drawing, and the description may be abbreviate | omitted.

  Before describing an exemplary structure of an optical modulator used in the present invention, a mechanism for modulating a carrier density in a silicon layer on which the operation is based will be described. The silicon-based electro-optic light modulator used in the present invention utilizes the carrier plasma effect described below.

As described above, pure silicon shows no change in refractive index due to the Pockels effect, and the change in refractive index due to the Franz-Keldysh effect or the Kerr effect is very small. Therefore, only the carrier plasma effect and the thermo-optic effect can be used for the light modulation operation. However, an optical modulator that changes the refractive index using the thermo-optic effect is slow. Therefore, only carrier diffusion by the carrier plasma effect is effective for high-speed operation (Gb / sec or more). The change in the refractive index due to the carrier plasma effect is explained by a first-order approximation of the following relational expression derived from the Kramers-Kronig relational expression and the Drude expression.

Where Δn and Δk represent the real and imaginary parts of the refractive index change of the silicon layer, e is the charge, λ is the light wavelength, ε ₀ is the dielectric constant in vacuum, and n is intrinsic the refractive index of silicon, m _e is the effective mass, m _h of the electron carrier is the effective mass of the hole carriers, mu _e is the mobility of electron carriers, mu _h is the mobility of hole carriers, .DELTA.N _e is the concentration of the electron carrier The change, ΔN _h, is the change in hole carrier concentration. Experimental evaluation of the carrier plasma effect in silicon has been carried out, and the refractive index change with respect to the carrier density at 1330 nm and 1500 nm wavelengths used in the optical communication system agrees well with the case obtained by the above formula. I understood. Further, in the electro-optic light modulator using the carrier plasma effect, the phase change amount is defined by the following equation.

  Where L is the length of the active layer along the light propagation direction of the electro-optic light modulator.

  In the present invention, the amount of phase change is a greater effect than light absorption, and the electro-optic light modulator described below can basically exhibit the characteristics of a phase modulator.

  An optical modulator using a free carrier plasma effect having a silicon-dielectric layer-silicon capacitor structure on an SOI (Silicon on Insulator) substrate, and a connection path connecting the optical modulator and the optical waveguide will be described below. To do.

  FIG. 1 is a schematic sectional view of an example of an optical modulator used in the present invention. The basic structure of this optical modulator will be described. A buried oxide layer 2 is formed on a substrate 3, and a first conductivity type semiconductor 8 having a rib structure thereon, a dielectric layer 11, and a second conductivity type. The semiconductor layers 9 are sequentially stacked. The substrate 3, the buried oxide layer 2 and the first conductivity type semiconductor 8 constitute an SOI substrate. In the figure, the arrow indicates the light propagation direction. The surface of the first conductivity type semiconductor (hereinafter referred to as “p-doped semiconductor”) 8 constituting the rib waveguide formed on the SOI substrate is recessed in the direction perpendicular to the light propagation direction (the length of the recess). The direction is parallel to the light propagation direction), and a concave-convex shape is formed. All the portions where the uneven shape is formed are covered with a thin dielectric layer 11 (hereinafter, “thin” indicates a submicron order (less than 1 μm)). On the thin dielectric layer 11, a second conductivity type semiconductor (hereinafter referred to as “n-doped semiconductor”) 9 is further deposited to form a rib shape. In the slab regions on both sides of the rib shape, a highly doped region (hereinafter referred to as “p + doped semiconductor”) 4 is formed, and the second conductivity type semiconductor 9 is also doped at a high concentration. A treated doped region (hereinafter all referred to as “n + doped semiconductor”) 5 is formed. Further, electrode contact layers 6 are provided on the p + doped semiconductor 4 and the n + doped semiconductor 5, respectively. The entire waveguide is covered with an oxide cladding 7.

  In the structure shown in FIG. 1, by providing an uneven shape at the junction interface of the capacitor structure, the overlap between the light field and the carrier density modulation region is large, and light modulation is sufficiently possible even if the light modulation length is short. Therefore, the size of the optical modulator can be reduced. Further, by further increasing the doping density of the region doped to exhibit the first conductivity type adjacent to the junction interface of the capacitor structure and the region doped to exhibit the second conductivity type, the series resistance component is reduced. It is also possible to reduce the RC time constant.

  In order to reduce the optical absorption loss due to the overlap between the region where the doping density is increased and the optical field, a rib waveguide as shown in FIG. 1 is used, and a structure in which the doping density in the slab region is increased is adopted. It is also possible to obtain an optical modulator that operates at high speed with small optical loss and small RC time constant.

In the region near the junction interface of the capacitor structure, assuming that the thickness of the portion where carrier modulation occurs is W, the maximum depletion layer thickness (thickness causing carrier modulation) W is given by the following formula in the thermal equilibrium state.

Here, ε _s is the dielectric constant of the semiconductor layer, k is the Boltzmann constant, N _c is the carrier density, _ni is the intrinsic carrier concentration, and e is the charge amount. For example, when N _c is 10 ¹⁷ / cm ³ , the maximum depletion layer thickness is about 0.1 μm, and as the carrier density increases, the depletion layer thickness, that is, the thickness of the region where the carrier density modulation occurs becomes thinner.

  FIG. 2 is a cross-sectional view taken along the line AA ′ of FIG. Assuming that the interval between the concave portion and the convex portion provided on the p-doped semiconductor 8 is X, X is preferably 2 W or less. When the interval between the concave and convex portions is 2 W or less, the carrier modulation regions between the adjacent concave and convex portions overlap, so that a higher light modulation effect can be obtained. However, it is possible to obtain an effect of improving the light modulation efficiency even when the interval between the adjacent concave and convex portions having a concave and convex shape is 2 W or more.

When the effective refractive index felt by the optical signal electric field is n _eff and the optical signal wavelength is λ, the light field size is λ / n _eff . Therefore, in the optical modulator shown in FIG. 1, the height from the concave portion to the convex portion provided on the surface of the p-doped semiconductor 8 is preferably λ / n _eff or less. By doing so, the overlap between the optical field and the region where carrier density modulation is performed is maximized, and efficient optical phase modulation can be realized.

  Next, the connection path will be described.

  FIG. 3 is a principal cross-sectional perspective view showing a connection path 30 that is a main characteristic part of the present invention and connects the optical modulator and the optical waveguide. A cross section 4 in FIG. 3 is a portion connected to the optical modulator shown in FIG. 1, and a cross section 1 in FIG. 3 is a portion connected to an optical waveguide provided outside the optical modulator. The cross section 1 and the cross section 4 are connected by the connection path 30 of the present invention. The optical waveguide is a silicon fine wire optical waveguide having a rectangular cross section similar to the cross section 1, and the optical modulator is a rib waveguide similar to the cross section 4 (see FIG. 1). When these two waveguides are connected without passing through the connection path 30, an optical connection loss occurs because the distribution of the optical field confined in each waveguide is different. Therefore, the connection path 30 has a cross-sectional shape 1 that matches the cross-sectional shape of the optical waveguide at the end portion of the optical waveguide, and a cross-sectional shape that matches the cross-sectional shape of the optical modulator at the end portion of the optical modulator. Further, the shape gradually changes from the shape of the cross section 1 to the shape of the cross section 4 through the shape of the cross section 2 and the shape of the cross section 3. By using the connection path 30 whose cross-sectional shape changes in this way, it is possible to suppress optical loss both between the connection path 30 and the optical modulator and between the connection path 30 and the optical waveguide. As a result, high optical modulation efficiency can be realized without impairing the characteristics of a small-sized optical modulator having a high phase and a high degree of modulation.

  Hereinafter, the structure of the connection path 30 will be described in detail.

  In the cross section 1, the optical waveguide 30 is composed of a first conductivity type semiconductor (hereinafter referred to as “p-doped semiconductor”) 8 having a buried oxide layer 2 formed on a substrate 3 and a rectangular cross section thereon. ing.

  Inside the connection path 30, at least part of the layer of the p-doped semiconductor 8 and at least part of the layer of the second conductivity type semiconductor (hereinafter all referred to as “n-doped semiconductor”) 9 sandwich the dielectric layer 11. Are overlapping. Specifically, an unevenness is formed in the layer of the p-doped semiconductor 8 so that the shape of the cross-section 1 is changed to the shape of the cross-section 2 next, and the dielectric is formed so as to cover the concave portion and the upper portion of the p-doped semiconductor 8 layer. Layer 11 is formed. The lower part of the p-doped semiconductor 8 layer also extends in the horizontal direction. Further, the gap between the concave portion and the convex portion of the layer of the p-doped semiconductor 8 is increased so that the shape of the cross section 3 is formed, and the layer of the n-doped semiconductor 9 is sandwiched between the dielectric layers 11 between the concave portion and the convex portion. In addition, a layer of n-doped semiconductor 9 is also formed on the dielectric layer 11. The lower portion of the p-doped semiconductor 8 layer further extends in the horizontal direction. And it becomes the shape of the cross section 4, and this corresponds with the cross-sectional shape of an optical modulator mentioned above. The inside of the connection path 30 has a taper shape or an inverse taper shape from the shape of the p-doped semiconductor 8 of the cross section 1 to the layers of the p-doped semiconductor 8 and the n-doped semiconductor 9 and the dielectric layer 11 sandwiched between these layers. So that the width and / or height thereof gradually change to change to the layers of the p-doped semiconductor 8 and the n-doped semiconductor 9 in the cross section 4 and the dielectric layer 11 sandwiched between these layers. ing. At this time, the width of the interval between the concave and convex portions of the concavo-convex shape of the layer of the p-doped semiconductor 8 is from the optical waveguide side to the optical modulator side so as to gradually change from the shape of the cross section 2 to the shape of the cross section 4. Gradually grows. Further, the optical modulator, the connection path 30 and the optical modulator are all covered with the oxide cladding 7. By connecting such a connection path 30 to the optical modulator described above, the optical communication system of the present invention is configured.

  FIG. 4 is a schematic view showing another example of an optical modulator used in the present invention. In this example, the surface of the SOI layer has a concavo-convex shape in a direction orthogonal to the light propagation direction. This optical modulator has a slab waveguide shape, but has a rib structure in the opposite direction to the structure of FIG. A thin dielectric layer 11 is deposited on the p-doped semiconductor 8 having an uneven shape, and an n-doped semiconductor 9 is further deposited. In order to reduce the size of the optical field, the thickness of the p and n doped semiconductors 8 and 9 spreading left and right for electrode extraction is set to 100 nm or less. This makes it possible to dispose the heavily doped p + doped semiconductor 4 and n + doped semiconductor 5 regions adjacent to the light modulation region, reduce the series resistance component, and accelerate the accumulation and removal of carriers. As a result, the size of the optical modulator is reduced, and higher speed and lower power are realized. An electrode contact layer 6 is provided on the p + doped semiconductor 4 and the n + doped semiconductor 5, and portions other than the above are covered with an oxide cladding 7.

  Next, the connection path 30 will be described.

  FIG. 5 is a cross-sectional perspective view of the main part showing the connection path 30 of the present invention for connecting the optical modulator and the optical waveguide. 5 is a portion connected to the optical modulator shown in FIG. 4, and cross section 1 in FIG. 5 is a portion connected to an optical waveguide provided outside the optical modulator. The cross section 1 and the cross section 4 are connected by a connection path 30. Also in this case, if the connection by the appropriate connection path 30 is not performed, an optical connection loss is caused for the reason described above.

  Therefore, the connection path 30 has a cross-sectional shape 1 that matches the cross-sectional shape of the optical waveguide at the end portion of the optical waveguide, and a cross-sectional shape that matches the cross-sectional shape of the optical modulator at the end portion of the optical modulator. Further, the shape gradually changes from the shape of the cross section 1 to the shape of the cross section 4 through the shape of the cross section 2 and the shape of the cross section 3.

  Hereinafter, the structure of the connection path 30 will be described in detail. The structure of the optical waveguide is as described above.

  In the connection path 30, at least a part of the layer of the p-doped semiconductor 8 and at least a part of the layer of the n-doped semiconductor 9 are overlapped with the dielectric layer 11 interposed therebetween. Specifically, an unevenness is formed in the layer of the p-doped semiconductor 8 so that the shape of the cross-section 1 is changed to the shape of the cross-section 2 next, and the concave portion and a part of the upper portion of the p-doped semiconductor 8 are covered. A dielectric layer 11 is formed. Further, an n-doped semiconductor 9 is formed on the dielectric layer 11. Further, the n-doped semiconductor 9 is formed so that the gap between the concave portion and the convex portion of the layer of the p-doped semiconductor 8 is increased so that the shape of the cross-section 3 is formed, and the dielectric layer 11 is sandwiched between the concave portion and the convex portion. The Further, on the side opposite to the rib-structured substrate 3, the p-doped semiconductor 8 and the n-doped semiconductor 9 layers spread in the horizontal direction in opposite directions. And it becomes the shape of the cross section 4, and this corresponds with the cross-sectional shape of an optical modulator mentioned above. The inside of the connection path 30 has a taper shape or an inverse taper shape from the shape of the p-doped semiconductor 8 of the cross section 1 to the layers of the p-doped semiconductor 8 and the n-doped semiconductor 9 and the dielectric layer 11 sandwiched between these layers. So that the width and / or height thereof gradually change to change to the layers of the p-doped semiconductor 8 and the n-doped semiconductor 9 in the cross section 4 and the dielectric layer 11 sandwiched between these layers. ing. At this time, the width of the interval between the concave and convex portions of the concavo-convex shape of the layer of the p-doped semiconductor 8 is from the optical waveguide side to the optical modulator side so as to gradually change from the shape of the cross section 2 to the shape of the cross section 4. Gradually grows. Further, the optical modulator, the connection path 30 and the optical modulator are all covered with the oxide cladding 7.

  FIG. 6 is a schematic view showing still another example of the optical modulator used in the present invention. 6A is a diagram viewed from the propagation direction of light, FIG. 6B is a diagram showing a BB ′ cross section of FIG. 6A, and FIG. It is the figure which showed CC 'cross section of (a). The direction of the arrow is the direction of light propagation (in FIG. 6A, from the front to the back).

  In this example, the surface of the p-doped semiconductor 8 in the rib waveguide formed on the SOI substrate is recessed parallel to the light propagation direction (the longitudinal direction of the recess is perpendicular to the light propagation direction). The concave and convex shape is formed, and all portions on the concave and convex shape are covered with the thin dielectric layer 11. An n-doped semiconductor 9 is deposited on the thin dielectric layer 11. Further, on the n-doped semiconductor 9, an n + doped semiconductor 5 doped at a high concentration is deposited. Highly doped p + doped semiconductors 4 are formed in the slab regions on both sides of the rib-shaped region. The p + doped semiconductor 4 and the n + doped semiconductor 5 are provided with an electrode contact layer 6, and the entire waveguide is covered with an oxide cladding 7.

Y is preferably 2 W or less for the reasons described above, where Y is the distance between the concave and convex portions of the concavo-convex shape formed on the p-doped semiconductor 8 and W is the thickness of the region where the carrier density is modulated. In addition, the period of the concavo-convex shape is formed so as to reduce the group velocity of the optical signal, or in order to suppress the reflection of the optical signal, the effective refractive index that the optical signal electric field feels aperiodically is n _eff , where the optical signal wavelength is λ, the interval may be less than λ / n _eff .

  Next, the connection path 30 will be described.

  FIG. 7 is a cross-sectional perspective view of the main part showing the connection path 30 of the present invention for connecting the optical modulator and the optical waveguide. A cross section 4 in FIG. 7 is a portion connected to the optical modulator shown in FIG. 6, and a cross section 1 in FIG. 7 is a portion connected to an optical waveguide provided outside the optical modulator. The cross section 1 and the cross section 4 are connected by the connection path 30 of the present invention. Also in this case, if the connection by the appropriate connection path 30 is not performed, an optical connection loss is caused for the reason described above.

  Therefore, the connection path 30 has a cross-sectional shape 1 that matches the cross-sectional shape of the optical waveguide at the end portion of the optical waveguide, and a cross-sectional shape that matches the cross-sectional shape of the optical modulator at the end portion of the optical modulator. Further, the shape gradually changes from the shape of the cross section 1 to the shape of the cross section 4 through the shape of the cross section 2 and the shape of the cross section 3.

  Hereinafter, the structure of the connection path 30 will be described in detail. The structure of the optical waveguide is as described above.

  In the connection path 30, at least a part of the layer of the p-doped semiconductor 8 and at least a part of the layer of the n-doped semiconductor 9 are overlapped with the dielectric layer 11 interposed therebetween. Specifically, the dielectric layer 11 is formed on the layer of the p-doped semiconductor 8 so that the shape of the cross section 1 is changed to the shape of the cross section 2 next, and further in the central portion of the upper layer of the dielectric layer 11. A layer of n-doped semiconductor 9 is formed. Further, the lower part of the p-doped semiconductor 9 layer extends in the horizontal direction. Next, the layer of the n-doped semiconductor 9 spreads in the horizontal direction from the center so as to have the shape of the cross section 3. Furthermore, the entire layer of the p-doped semiconductor 8 extends in the horizontal direction. And it becomes the shape of the cross section 4, and this corresponds with the cross-sectional shape of an optical modulator mentioned above. The inside of the connection path 30 has a taper shape or an inverse taper shape from the shape of the p-doped semiconductor 8 of the cross section 1 to the layers of the p-doped semiconductor 8 and the n-doped semiconductor 9 and the dielectric layer 11 sandwiched between these layers. So that the width and / or height thereof gradually change to change to the layers of the p-doped semiconductor 8 and the n-doped semiconductor 9 in the cross section 4 and the dielectric layer 11 sandwiched between these layers. ing. Further, the optical modulator, the connection path 30 and the optical modulator are all covered with the oxide cladding 7.

  8, 9 and 10 show an example of a method for forming the connection path 30 of the present invention so as to have a shape suitable for the above-described optical modulator having a carrier modulation region having an uneven shape.

  FIG. 8A is a cross-sectional view of an SOI substrate used for forming the connection path 30 of the present invention. This SOI substrate has a structure in which a buried oxide layer 2 is laminated on a substrate 3 and a silicon layer 8 of about 100 to 1000 nm (1 μm) is further laminated thereon. In order to reduce optical loss, the thickness of the buried oxide layer 2 was set to 1000 nm (1 μm) or more. The silicon layer 8 on the buried oxide layer 2 is formed by using a substrate that has been previously doped so as to exhibit the first conductivity type, or by doping phosphorus or boron into the silicon surface layer by ion implantation or the like. Heat treatment may be performed. In FIG. 8A, it is assumed that boron is doped, and the silicon layer 8 is a p-doped semiconductor.

Next, as shown in FIG. 8B, a thermal oxide layer 12 of about 10 to 30 nm is formed on the p-doped semiconductor 8 by heat treatment, and a film forming method such as a low pressure CVD (Chemical Vapor Deposition) method is further formed thereon. Thus, the SiN _x layer 13 is formed.

Next, as shown in FIG. 8C, the SiN _x layer 13 is patterned so as to have an interval corresponding to the interval between the concave and convex portions formed on the p-doped semiconductor 8.

Next, as shown in FIG. 8 (d), thermal oxidation is performed using the SiN _x layer patterned in FIG. 8 (c) as a mask, and the thermal oxide layer 14 is also applied to the undoped masked p-doped semiconductor 8 layer. To form. These processes are called LOCOS (Local Oxidation of Semiconductor) processes, which are common processes in CMOS processing processes. However, in the case of microfabrication of 100 nm or less, shape control may not be sufficient. Therefore, instead of the LOCOS process, it is also effective to form an uneven shape on a desired surface by a method such as reactive ion etching using a photoresist as a mask.

Next, after immersing the SOI substrate in a phosphoric acid solution to remove the SiN _x layer 13 and the thermal oxide layers 12 and 14, heat treatment is performed as shown in FIG. 9A to form a dielectric layer on the surface layer of the p-doped semiconductor 8. A silicon oxide layer which is the body layer 11 is formed. The dielectric layer 11 may be at least one layer made of a silicon oxide layer, a silicon nitride layer, or another high-k insulating layer.

  Next, as shown in FIG. 9B, a polycrystalline silicon film 9 is formed by CVD or sputtering so that the surface irregularities on the dielectric layer 11 are sufficiently covered. At this time, due to the uneven shape of the dielectric layer 11, the same uneven shape is also formed on the polycrystalline silicon 9. Such an uneven shape on the polycrystalline silicon 9 causes light scattering loss when an optical signal propagates, and therefore it is desirable to smooth it by CMP (chemical-mechanical polishing process). Further, the polycrystalline silicon 9 is doped during film formation so as to exhibit the second conductivity type, or after the film formation, is doped with boron or phosphorus by an ion implantation method or the like (first conductivity type semiconductor). Dope with the opposite of the layer). In FIG. 9B, it is assumed that phosphorus is doped, and polycrystalline silicon 9 is an n-doped semiconductor.

  Next, as shown in FIG. 9C, processing is performed by a reactive plasma etching method or the like so as to form a rib-shaped waveguide. Further, as shown in FIG. 9D, a p + doped semiconductor 4 and an n + doped semiconductor 5 that are highly doped are formed in regions adjacent to the p-doped semiconductor 8 and the n-doped semiconductor 9.

  If the above method is appropriately used, the connection path from the optical waveguide side end of the connection path to the optical modulator side end can be formed.

  Further, since the optical modulator is formed integrally with the connection path 30, the method of forming the optical modulator is the same as that shown in FIGS.

  The dependence of the phase shift amount on the optical signal propagation direction in the optical modulator used in the present invention in the optical signal propagation direction was examined in the case where the surface of the first conductivity type semiconductor layer had an uneven shape. The interval between the concave and convex portions having an uneven shape was set to 160 nm or less. An example of the test result is shown in FIG.

  It can be seen that the light modulation efficiency is improved because the phase shift is increased by forming an uneven shape of about 160 nm or less, which is the same thickness as the carrier-modulated thickness between the concave and convex portions. Although the test results are not shown, the light modulation efficiency was improved by increasing the height from the concave portion to the convex portion.

In addition, the relationship between the carrier density and the optical modulation operating frequency band of the optical modulator was examined for the case where the surface of the first silicon semiconductor layer of the optical modulator used in the present invention had an uneven shape. The operating frequency band of optical modulation has a trade-off between the effect of reducing the size by improving the modulation efficiency and the effect of increasing the electric capacity by providing the uneven shape. The distance between the concave and convex portions of the concave-convex shape is 160 nm or less, the effective refractive index felt by the optical signal electric field is n _eff , and the optical signal wavelength is λ, the height from the concave portion to the convex portion is λ / n _When it is less than _eff , the operating frequency band of light modulation is widened.

As can be seen from the example of the test results shown in FIG. 11, by setting the carrier density to about 10 ¹⁸ / cm ³ , the operating frequency band of optical modulation becomes 10 GHz or more, and high-speed operation is possible.

  In addition to the above, in order to improve the frequency band, carrier mobility and lifetime are very important. In particular, the carrier mobility in the polycrystalline silicon layer is a problem in high-speed operation. Therefore, recrystallization by annealing treatment increases the particle size and improves carrier mobility, or improves the crystal quality of the second conductivity type semiconductor layer by using an epitaxial lateral growth (ELO) method or the like. It is effective.

DESCRIPTION OF SYMBOLS 1 Intrinsic semiconductor 2 Embedded oxide layer 3 Substrate 4 P + doped semiconductor 5 N + doped semiconductor 6 Electrode contact layer 7 Oxide cladding 8 P doped semiconductor 9 N doped semiconductor 11 Dielectric layer 12 Thermal oxide layer 13 SiN _x layer 14 Oxide layer 30 Connection path

Claims (10)

It formed in the semiconductor layer which is doped to exhibit a first conductivity type, to exhibit an optical waveguide provided outside the optical modulator, a connecting passage for connecting the optical modulator, the first conductivity type An optical communication system having an optical modulator in which at least a part of a semiconductor layer doped with a semiconductor layer and at least a part of a semiconductor layer doped to exhibit a second conductivity type overlap each other with a dielectric layer interposed therebetween There,
At least a part of the first conductivity type semiconductor layer and at least a part of the semiconductor layer doped so as to exhibit the second conductivity type overlap each other with a dielectric layer interposed therebetween in the connection path. The first conductive type semiconductor layer, the second conductive type semiconductor layer, and the dielectric layer have a tapered shape or a reverse tapered shape from the optical modulator toward the optical waveguide. width and / or height is changing,
In the optical modulator, the surface of the semiconductor layer of the first conductivity type is uneven at a portion where the semiconductor layer of the first conductivity type and the semiconductor layer of the second conductivity type overlap with each other with the dielectric layer interposed therebetween. The dielectric layer is formed on the first conductive type semiconductor layer having the concavo-convex shape, and the second conductive type semiconductor layer is further formed on the dielectric layer. An optical communication system in which the uneven shape of the surface of the first conductivity type semiconductor layer is formed in a direction parallel to the propagation direction of the optical signal . 2. The optical communication according to claim 1, wherein the connection path has the same cross-sectional shape as that of the optical waveguide at an end portion of the optical waveguide, and has the same cross-sectional shape as that of the optical modulator at an end portion of the optical modulator. System . In the optical modulator, the surface of the first conductive type semiconductor layer has an interval between the concave and convex portions having the concavo-convex shape. The first conductive type semiconductor layer and the second conductive type semiconductor layer each internally accumulated on both sides of the free carriers said dielectric layer, removal, or the thickness W of the inversion region is less than 2W, optical communication system according to claim 1 or 2. The optical modulator has an effective refractive index that the optical signal electric field in the optical modulator senses the height from the concave-convex concave portion to the convex portion of the surface of the first conductivity type semiconductor layer. The optical communication system according to any one of claims 1 to 3 , wherein n _{eff is} equal to or less than λ / n _eff , where n _{eff is} an optical signal wavelength. The optical modulator has a rib waveguide structure in which an optical signal including a portion where the first conductive type semiconductor layer and the second conductive type semiconductor layer overlap each other with the dielectric layer interposed therebetween has a rib waveguide structure. The optical communication system according to any one of claims 1 to 4 . Region where light signal of the first conductivity type semiconductor layer and the semiconductor layer of the second conductivity type comprises overlapped portions across the dielectric layer propagates has a slab waveguide structure, according to claim 1 5. The optical communication system according to any one of items 1 to 4 . An optical waveguide provided outside the optical modulator, formed of a semiconductor layer doped so as to exhibit the first conductivity type, a connection path connecting the optical modulator, and the first conductivity type An optical communication system having an optical modulator in which at least a part of a semiconductor layer doped with a semiconductor layer and at least a part of a semiconductor layer doped so as to exhibit the second conductivity type overlap each other with a dielectric layer interposed therebetween Manufacturing method,
Inside the connection path, at least part of the semiconductor layer of the first conductivity type and at least part of the semiconductor layer doped so as to exhibit the second conductivity type are overlapped with a dielectric layer interposed therebetween, Forming a semiconductor layer of the first conductivity type, the semiconductor layer of the second conductivity type, and the dielectric layer so as to have a tapered shape or an inversely tapered shape ;
In a part of the optical modulator, an uneven shape is provided on the surface of the first conductivity type semiconductor layer in a direction parallel to the propagation direction of the optical signal, and the uneven shape of the first conductivity type semiconductor layer is provided. The dielectric layer is formed on the dielectric layer, and the second conductive type semiconductor layer is further formed on the dielectric layer, whereby the first conductive type semiconductor layer and the second conductive type semiconductor layer are formed on the dielectric layer. A method of manufacturing an optical communication system , comprising: overlapping with a dielectric layer interposed therebetween . The interval between the concave and convex portions having the concavo-convex shape on the surface of the first conductive type semiconductor layer can be freely set inside each of the first conductive type semiconductor layer and the second conductive type semiconductor layer. 8. The method of manufacturing an optical communication system according to claim 7 , wherein carriers are set to 2 W or less with respect to a thickness W of a region where carriers accumulate, remove, or invert on both sides of the dielectric layer. The height from the concave-convex concave portion to the convex portion of the surface of the semiconductor layer of the first conductivity type is defined as an effective refractive index n _eff , an optical signal electric field in the optical modulator, and an optical signal. The method of manufacturing an optical communication system according to claim 7 or 8 , wherein the wavelength is λ / n _eff or less when the wavelength is λ. The optical modulator and the connection passage is formed by integral molding, a manufacturing method of an optical communication system according to any one of claims 7 to 9.

Images