SEMICONDUCTOR LIGHT-RECEIVING ELEMENT AND METHOD FOR MANUFACTURING SAME

[0001] This application is a continuation application of International Application PCT/JP2013/061625 filed on Apr. 19, 2013 and designated the U.S., the entire contents of which are incorporated herein by reference.

[0002] The present invention relates to a semiconductor light-receiving element and a method for manufacturing the same, and in particular, to a semiconductor light-receiving element of which the light absorbing layer is made of Ge provided on a Si substrate to be used for optical communication and data communication as well as a method for manufacturing the same.

[0003] As data transmission amounts between server and CPUs have increased in recent years, the handling of data through transmission of electrical signals using conventional Cu wires are approaching the limit. In order to resolve this bottleneck, optical interconnection, that is to say, data transmission through optical signals is required. Furthermore, from the points of view of reduction in power consumption and reduction in the area of devices, elements formed by integrating optical components such as optical transmitters, optical modulators and optical receivers that are required for optical transmission/reception on a Si substrate become necessary.

[0004] Meanwhile, optical components integrated on a Si substrate are connected through optical fibers and, therefore, it is preferable to use the wavelength of 1.55 μm, with which loss is low in optical fibers, as a transmission wavelength band. Therefore, it is preferable to use Ge, which has an absorption end in the vicinity of 1.55 μm, as an absorbing layer of a photodetector to be used for optical transmission with a band of the wavelength of 1.55 μm.

[0005] Meanwhile, wavelength division multiplexing (WDM) transmission is required as the data transmission amount increases. In order to implement WDM transmission, Ge photodetectors having a high response sensitivity in a broad region of which the wavelengths are longer than 1.55 μm are necessary.

[0006] In general, when the temperature is cooled to room temperature while Ge is growing on a Si substrate, the Ge epitaxial layer is subject to tensile strain in the directions within the plane of the substrate due to the difference between the coefficients of thermal expansion of Si and Ge. As a result, the absorption end of Ge on the Si substrate has a longer wavelength as compared to Ge layers in a bulk state, as has been reported (see Non-Patent Literature 1). This works advantageously from the point of view of expansion of the wavelength band of the photodetectors.

[0007] Meanwhile, the element capacitance of photodetectors is required to be lowered from the point of view of increase in the high-speed response properties and, thus, it is necessary to make the element area smaller (make the element width narrower). In addition, from the point of view of increase in the response sensitivity properties, it is required to prevent photocarriers generated in the depletion layer from recombining or becoming trapped by lattice defects while drifting. In order to do so, it is likewise necessary to make the element area smaller (make the element width narrower).

[0008] Thus, it has been reported that a photodetector having Ge on a Si substrate as an absorbing layer is processed as a mesa element having a width of approximately of several μm (see Non-Patent Literature 2). FIG. 20 is a schematic cross sectional diagram illustrating a conventional photodetector having Ge as an absorbing layer where the photodetector is formed using an SOI substrate. A Si layer 83 provided on top of a Si substrate 81 with a BOX layer 82 in between is processed to form a p type Si mesa portion 84 and p type Si slab portions 85 on the two sides thereof. At this time, a waveguide in stripe form is formed so as to be connected to the p type Si mesa portion 84 through a tapered portion, though this is not shown.

[0009] A non-doped Ge layer is formed on top of this p type Si mesa portion 84 through selective growth and n type impurities are implanted into the surface thereof so as to provide an n⁺⁺ type Ge contact layer 87, where the portion into which the impurities have not been introduced is an i type Ge light absorbing layer 86. Meanwhile, p type impurities are implanted into portions of the p type Si slab portions 85 so as to form p⁺⁺ type Si contact portions 88.

[0010] Next, an oxide film 89 that becomes the upper clad layer of the waveguide in stripe form is formed, plugs 90 and 91 are formed and, then, an n side electrode 92 and p side electrodes 93 are formed. Light that propagates through the waveguide in stripe form passes through the p type Si mesa portion so as to reach, and be absorbed by, the i type Ge light absorbing layer 86 through evanescent coupling.

• [0011] Non-Patent Literature 1: Yasuhiko Ishikawa, Kazumi Wada, Jifeng Liu, Douglas D. Cannon, Hsin-Chiao Luan, Jurgen Michel, and Loinel C. Kimerling, Journal Of Applied Physics 98, 013501 (2005)
• [0012] Non-Patent Literature 2: Tao Yin, Rami Cohen, Mike M. Morse, Gadi Sarid, Yoel Chetrit, Doron Rubin, and Mario J. Paniccia, Optics Express, 15, 13966 (2007)
• [0013] Non-Patent Literature 3: L. Ding, T. Y. Liow, A. E. J. Lim, N. Duan, M. B. Yu, and G. Q. Lo, OFC/NFOEC Technical Digest, OW3G. 4 (2012)

[0014] It has been reported that, as the width of the element becomes narrower, lattice relaxation progresses within the plane and the tensile strain induced in the Ge layer is reduced (see Non-Patent Literature 3). As a result, the absorption end of Ge has shorter wavelengths, which lowers the response sensitivity on the long wavelength side when the width of the element is made narrower in order to lower the element capacitance or to reduce the loss of photocarriers.

[0015] Meanwhile, the element capacitance increases, which leads to the deterioration of the high-speed properties, and the loss of photocarriers increases, which leads to the deterioration of the response efficiency, when the width of the element is made wider in order to suppress the reduction in the tensile strain within the plane for the purpose of increase in the response efficiency with a long wavelength. That is to say, such a problem arises with the photodetector having the structure illustrated in FIG. 20 that expansion of the width of the wavelength band through the induction of tensile strain and increase in the high-speed response properties and the response efficiency through miniaturization cannot both be achieved at the same time.

[0016] According to one disclosed aspect, a semiconductor light-receiving element includes: a substrate of which a surface is a single crystal Si layer; a PIN type photodiode made by sequentially layering on top of the substrate a Si layer of a first conductivity type, a non-doped Ge layer and a Ge layer of a second conductivity type that is the opposite type of the first conductivity type; a Ge current blocking mechanism provided in at least a portion in the periphery of the layer structure made of the non-doped Ge layer and the Ge layer of the second conductivity type; a contact electrode for the second conductivity type provided on the Ge layer of the second conductivity type; and a contact electrode for the first conductivity type provided in the Si layer of the first conductivity type.

[0017] According to another disclosed aspect, a method for manufacturing a semiconductor light-receiving element includes: implanting impurity ions of a first conductivity type into at least a portion of a surface of a substrate of which the surface is a single crystal Si layer and annealing the impurity ions for activation; growing a non-doped Ge layer on at least a Si region of the first conductivity type into which impurity ions of the first conductivity type have been implanted; forming, on a surface of the non-doped Ge layer, a first dielectric mask that covers a region on which a photodiode is to be formed; implanting impurity ions of a second conductivity type that is the opposite conductivity type of the first conductivity type into an exposed portion of the non-doped Ge layer using the first dialectic mask as a mask; compensating the surface with impurities of the first conductivity type by implanting the impurities of the first conductivity type into the exposed portion of the non-doped Ge layer at a location that is shallower than the impurities of the second conductivity type using the first dielectric mask as a mask; forming a second dielectric mask from which only the region on which a photodiode is to be formed is exposed; implanting impurities of the second conductivity type into an exposed surface of the non-doped Ge layer using the second dielectric mask as a mask; and annealing the implanted impurities for activation.

[0018] The object and advantages of the invention will be realized and attained by means of the elements and combinations particularly pointed out in the claims.

[0019] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory and are not restrictive of the invention.

[0040] Here, the semiconductor light-receiving element according to an embodiment of the present invention is described in reference to FIGS. 1A through 2B. FIGS. 1A and 1B are a perspective diagram and a cross sectional diagram illustrating the semiconductor light-receiving element according to the present invention. FIGS. 2A and 2B are a top diagram and a cross sectional diagram illustrating the semiconductor light-receiving element according to the present invention. A PIN type photodiode 15 is formed on a substrate 11 of which the surface is a single crystal Si layer 14 by sequentially layering a Si layer 16 of the first conductivity type, a non-doped Ge layer 17 and a Ge layer of the second conductivity type 18.

[0041] A Ge current blocking mechanism 19 is provided in at least a portion in the periphery of the multi-layered structure made of the non-doped Ge layer 17 and the Ge layer 18 of the second conductivity type, and a contact electrode 22 for the second conductivity type is provided on top of the Ge layer 18 of the second conductivity type. Meanwhile, a contact electrode 23 for the first conductivity type is provided so as to be connected to the Si layer 16 of the first conductivity type.

[0042] This Ge current blocking mechanism 19 typically has a structure where a Ge layer 20 of the second conductivity type and a Ge layer 21 of the first conductivity type are layered on top of each other starting from the substrate 11 side. As a result, a thyristor structure made of the Ge layer 18 of the second conductivity type, the Ge layer 21 of the first conductivity type, the Ge layer 20 of the second conductivity type and the Si layer 16 of the first conductivity type is formed between the contact electrode 22 for the second conductivity type and the contact electrode 23 for the first conductivity type.

[0043] When the thyristor structure is formed, a current does not flow through the thyristor structure in the backward direction when backward biased to the breakdown voltage or lower in the backward direction. Accordingly, the thyristor structure functions as a carrier blocking area which prevents photocarriers generated in the non-doped Ge layer 17 from dispersing into the periphery. As a result, the carrier path is narrowed and the element capacitance is lowered. At the same time, it is possible to reduce the number of photocarriers that recombine or are trapped by lattice defects within the depletion layer due to drifting.

[0044] The PIN type photodiode 15 typically has a rectangular parallelepiped form of which three sides out of the four sides may make contact with the Ge current blocking mechanism 19. In this case, on the one side that does not make contact with the Ge current blocking mechanism 19, a single crystal Si core layer 24 in stripe form may be connected to the Si layer 16 of the first conductivity type via a tapered waveguide 25.

[0045] Thus, the PIN type photodiode 15 is surrounded by the Ge current blocking mechanism 19 made of Ge so that the strain relaxation induced in the periphery of the non-doped Ge layer 17 does not transfer to the non-doped Ge layer 17. Therefore, wavelengths are not shortened at the absorbing end of the non-doped Ge layer 17 due to the reduction in the tensile strain.

[0046] Alternatively, all of the sides of the PIN type photodiode 15 may be surrounded by the Ge current blocking mechanism 19 and, in this case, the top surface of the PIN type photodiode 15 may be a surface through which light enters. In addition, the substrate 11 may be a single crystal Si substrate, but typically an SOI substrate where a single crystal Si layer 14 is provided on a crystal Si substrate 12 with a SiO₂ film 13 in between is used.

[0047] In the case where such a semiconductor light-receiving element is formed, impurity ions of the first conductivity type are implanted into at least a portion of the surface of a substrate 11 of which the surface is a single crystal Si layer 14 and are activated through annealing, and a non-doped Ge layer 17 is grown on the region of the Si layer 16 of the first conductivity type into which impurity ions of the first conductivity type have been implanted. Next, a first dielectric mask that covers the region on which a photodiode is to be formed is formed on the surface of the non-doped Ge layer 17. The first dielectric mask is used as a mask to implant impurity ions of the second conductivity type into the exposed portion of the non-doped Ge layer 17 so that a Ge layer 20 of the second conductivity type is formed.

[0048] Next, impurities of the first conductivity type are implanted into the exposed portion of the non-doped Ge layer 17 into which impurities of the second conductivity type have been implanted using the first dielectric mask as a mask at a location shallower than the implanted impurities of the second conductivity type so that the surface is compensated with impurities of the first conductivity type and, thus, a Ge layer 21 of the first conductivity type is formed. Next, a second dielectric mask is formed for exposing only the region on which a photodiode is to be formed. This second dielectric mask may be used as a mask so that impurities of the second conductivity type can be implanted into the surface of the exposed non-doped Ge layer 17 so as to form a Ge layer 18 of the second conductivity type.

[0049] Here, the single crystal Si layer 14 may be etched so that a tapered waveguide 25 to be connected to the photodiode region and a single crystal Si core layer 24 in stripe form to be connected to the tapered waveguide 25 can be formed.

[0050] In addition, in the step of growing the non-doped Ge layer 17, it is desirable to use a two-stage growth step including a first growth step for growing a layer at a relatively low temperature in order to form a flat layer through secondary growth and a second growth step for growing a layer at a relatively high temperature in order to make possible the growth of a high quality crystal. Here, typical examples of the method for growing the non-doped Ge layer 17 are low pressure chemical vapor deposition methods and molecular beam epitaxial growth methods.

[0051] In the embodiment of the present invention a Ge layer is provided around the non-doped Ge layer 17 that becomes the light absorbing layer and, therefore, strain relaxation is suppressed and the wavelength can be prevented from becoming shorter at the absorbing end of the non-doped Ge layer 17. At the same time, it is possible to reduce the element capacitance and to reduce the number of photocarriers that are trapped by lattice defects. As a result, it is possible to realize operation with a broad wavelength band, and a rapid high-speed response and a high response efficiency at the same time.

[0052] Next, a process for manufacturing the semiconductor light-receiving element according to Example 1 of the present invention is described in reference to FIGS. 3A through 10B. Here, FIGS. 3A, 4A, 5A, 6A, 7A, 8A and 9A are schematic perspective diagrams and FIGS. 3B, 4B, 5B, 6B, 7B, 8B and 9B are cross-sectional diagrams along the parallelogram depicted by the single-dotted chain line. First, as illustrated in FIGS. 3A and 3B, an SOI substrate is prepared where a single crystal i type Si layer 33 having a thickness of 0.3 μm is provided on a Si substrate 31 of which the main surface is (001) with a BOX layer 32 having a thickness of 3.0 μm in between.

[0053] Next, a resist is applied onto the i type Si layer 33 and patterned into the form of a Si passive waveguide through exposure to light using electron beam lithography, which is followed by development through wet etching. Next, a Si rib type waveguide 34 having a core layer 35 and a slab portion 36 is formed through inductively coupled plasma (ICP) dry etching. Here, a tapered portion 37 is provided on the side that makes contact with the i type Si layer 33.

[0054] Next, as illustrated in FIGS. 4A and 4B a SiO₂ film is deposited on the entire surface in accordance with a CVD method and is patterned to form a mask 38 for ion implantation, which has openings only for regions of a light-receiving portion and a thyristor portion. Next, B ions are implanted using the mask 38 for ion implantation as a mask. In this case, the implantation energy is 40 keV and the dosage amount is 5.0×10¹⁴ cm⁻². Next, an annealing process for activating the implanted B ions is carried out for one second at 1050° C. so as to form a p type Si layer 39. The concentration of impurities in the p type Si layer 39 is 1.0×10¹⁸ cm⁻³ to 3.0×10¹⁹ cm⁻³.

[0055] Next, as illustrated in FIGS. 5A and 5B, a mask 40 for selective growth made of a SiO₂ film is provided and an i type Ge layer 41, having a thickness of 0.8 μm is selectively grown in accordance with a low pressure CVD method. In this case, GeH₄ is used as a material gas and H₂ is used as a carrier gas. The layer is first grown by 0.1 μm at a low temperature of 400° C. and then the layer is grown by 0.7 μm at a high temperature of 700° C.

[0056] Next, as illustrated in FIGS. 6A and 6B, a mask 42 for ion implantation having an opening only for a portion in which a thyristor structure is to be formed is made of a SiO₂ film. Next, this mask 42 for ion implantation is used as a mask to implant P ions so that the entirety of the i type Ge layer 41 is doped. In this case, the implantation energy is 600 keV and the dosage amount is 1.0×10¹⁴ cm⁻². As a result, the concentration of impurities in the n type Ge layer 43 is 1.0×10¹⁸ cm⁻³ to 2.0×10¹⁸ cm⁻³.

[0057] Next, as illustrated in FIGS. 7A and 7B, the same mask 42 for ion implantation is used as a mask to shallowly implant B ions so that the upper layer, having a thickness of 0.5 μm, is compensated and becomes a p type Ge layer 44 while the bottom layer, having a thickness of 0.3 μm, is not compensated and remains as an n type Ge layer 43. In this case, the implantation energy is 100 keV and the dosage amount is 3.0×10¹⁴ cm⁻². As a result, the concentration of impurities in the p type Ge layer 44 is 1.0×10¹⁸ cm⁻³ to 1.0×10¹⁹ cm⁻³.

[0058] Next, as illustrated in FIGS. 8A and 8B, the mask 42 for ion implantation is removed with hydrofluoric acid, and after that a new mask 45 for ion implantation having an opening only for the region in which a photodiode is to be formed is made of a SiO₂ film. The mask 45 for ion implantation is used as a mask to implant P ions so that the upper layer, having a thickness of 0.3 μm, is converted to an n type Ge layer 46. In this case, the implantation energy is 40 keV and the dosage amount is 1.0×10¹⁴ cm⁻². As a result, the concentration of impurities in the n type Ge layer 46 is 4.0×10¹⁸ cm⁻³ to 1.0×10¹⁹ cm⁻³, and the concentration of impurities located at a depth of 0.3 μm from the surface is 1.0×10¹⁶ cm⁻³.

[0059] Next, as illustrated in FIGS. 9A and 9B, the mask 45 for ion implantation is removed with hydrofluoric acid, and after that annealing for activation is carried out for one second at 1050° C. so as to activate the implanted impurities. Next, a SiO₂ film, which becomes an upper clad layer 47 and also becomes a passivation film in the element portion, is deposited in accordance with a low pressure CVD method, and after that the surface of the SiO₂ film and the Ge layer is flattened in accordance with a chemical mechanical polishing (CMP) method.

[0060] Next, while not shown in the figures, a resist is applied onto the flattened surface and is patterned to the form of contact holes through exposure to light using electron beam lithography, which is followed by development through wet etching. Next, contact holes are created through ICP dry etching and wet etching with HF.

[0061] Next, a TiN film, which becomes a barrier metal, is formed in accordance with a spattering method, and after that the contact holes are filled in with Al. Next, the resist used for the creation of contact holes is removed and at the same time the Al layer that has been deposited on the resist is lifted off so as to form an n side electrode 48 and a p side electrode 49. Thus, the basic structure of the semiconductor light-receiving element according to Example 1 of the present invention is complete.

[0062] FIGS. 10A and 10B are diagrams illustrating the structure of the semiconductor light-receiving element according to Example 1 of the present invention. FIG. 10A is a top diagram and FIG. 10B is a cross-sectional diagram along the single-dotted chain line A-A′ in FIG. 10A. As illustrated in the figures, the three sides of the PIN type photodiode made of the n type Ge layer 46, the i type Ge layer 41 and the p type Si layer 39 are surrounded by the p type Ge layer 44 and the n type Ge layer 43, and therefore strain relaxation on the three end surfaces can be prevented from affecting the PIN type photodiode. In addition, a thyristor structure made of the n type Ge layer 46, the p type Ge layer 44, the n type Ge layer 43 and the p type Si layer 39 is formed on the two sides of the PIN type photodiode so as to function as a current blocking layer, which makes it possible to reduce a leak current. Here, a pseudo thyristor-like current block mechanism is formed of the n type Ge layer 46, the p type Ge layer 44, the n type Ge layer 43, the i type Si layer 33 and the p type Si layer 39 on the side opposite to the surface through which light enters.

[0063] Signal light that has entered into the core layer 35 propagates through the tapered portion 37 and reaches the p type Si layer 39 so as to be transmitted to the i type Ge layer 41 through evanescent coupling, and then is extracted from the n side electrode 48 as an electrical signal. Here, the p side electrode 49 is usually set to the ground potential.

[0064] Next, a process for manufacturing the semiconductor light-receiving element according to Example 2 of the present invention is described in reference to FIGS. 11A through 18B. The basic process is the same as the above description for Example 1 except that the conductivity types are opposite. Here, FIGS. 11A, 12A, 13A, 14A, 15A, 16A and 17A are schematic perspective diagrams and FIGS. 11B, 12B, 13B, 14B, 15B, 16B and 17B are cross-sectional diagrams along the parallelogram depicted by the single-dotted chain line. First, as illustrated in FIGS. 11A and 11B, an SOI substrate is prepared where a single crystal i type Si layer 53 having a thickness of 0.3 μm is provided on a Si substrate 51 of which the main surface is (001) with a BOX layer 52 having a thickness of 3.0 μm in between.

[0065] Next, a resist is applied onto the i type Si layer 53 and patterned into the form of a Si passive waveguide through exposure to light using electron beam lithography, which is followed by development through wet etching. Next, a Si rib type waveguide 54 having a core layer 55 and a slab portion 56 is formed through ICP dry etching. Here, a tapered portion 57 is provided on the side that makes contact with the i type Si layer 53.

[0066] Next, as illustrated in FIGS. 12A and 12B a SiO₂ film is deposited on the entire surface in accordance with a CVD method and is patterned to form a mask 58 for ion implantation, which has openings only for regions of a light-receiving portion and a thyristor portion. Next, P ions are implanted using the mask 58 for ion implantation as a mask. In this case, the implantation energy is 150 keV and the dosage amount is 5.0×10¹⁴ cm⁻². Next, an annealing process for activating the implanted B ions is carried out for one second at 1050° C. so as to form an n type Si layer 59. The concentration of impurities in the n type Si layer 59 is 1.0×10¹⁸ cm⁻³ to 3.0×10¹⁹ cm⁻³.

[0067] Next, as illustrated in FIGS. 13A and 13B, a mask 60 for selective growth made of a SiO₂ film is provided and an i type Ge layer 61, having a thickness of 0.8 μm is selectively grown in accordance with a low pressure CVD method. In this case, GeH₄ is used as a material gas and H₂ is used as a carrier gas. The layer is first grown by 0.1 μm at a low temperature of 600° C. and then the layer is grown by 0.7 μm at a high temperature of 700° C.

[0068] Next, as illustrated in FIGS. 14A and 14B, a mask 62 for ion implantation having an opening only for a portion in which a thyristor structure is to be formed is made of a SiO₂ film. Next, this mask 62 for ion implantation is used as a mask to implant B ions so that the entirety of the i type Ge layer 61 is doped. In this case, the implantation energy is 200 keV and the dosage amount is 3.0×10¹⁴ cm⁻². As a result, the concentration of impurities in the p type Ge layer 63 is 2.0×10¹⁸ cm⁻³ to 6.0×10¹⁸ cm₋₃.

[0069] Next, as illustrated in FIGS. 15A and 15B, the same mask 62 for ion implantation is used as a mask to shallowly implant P ions so that the upper layer, having a thickness of 0.5 μm, is compensated and becomes an n type Ge layer 64 while the bottom layer, having a thickness of 0.3 μm, is not compensated and remains as a p type Ge layer 63. In this case, the implantation energy is 350 keV and the dosage amount is 8.0×10¹⁴ cm⁻². As a result, the concentration of impurities in the n type Ge layer 64 is 1.0×10¹⁸ cm⁻³ to 3.0×10¹⁹ cm⁻³.

[0070] Next, as illustrated in FIGS. 16A and 16B, the mask 62 for ion implantation is removed with hydrofluoric acid, and after that a new mask 65 for ion implantation having an opening only for the region in which a photodiode is to be formed is made of a SiO₂ film. The mask 65 for ion implantation is used as a mask to implant B ions so that the upper layer, having a thickness of 0.3 μm, is converted to a p type Ge layer 66. In this case, the implantation energy is 20 keV and the dosage amount is 2.0×10¹⁴ cm⁻². As a result, the concentration of impurities in the p type Ge layer 66 is 5.0×10¹⁸ cm⁻³ to 2.0×10¹⁹ cm⁻³, and the concentration of impurities located at a depth of 0.3 μm from the surface is 1.0×10¹⁶ cm⁻³.

[0071] Next, as illustrated in FIGS. 17A and 17B, the mask 65 for ion implantation is removed with hydrofluoric acid, and after that annealing for activation is carried out for one second at 1050° C. so as to activate the implanted impurities. Next, a SiO₂ film, which becomes an upper clad layer 67 and also becomes a passivation film in the element portion, is deposited in accordance with a low pressure CVD method, and after that the surface of the SiO₂ film and the Ge layer is flattened in accordance with a chemical mechanical polishing (CMP) method.

[0072] Next, while not shown in the figures, a resist is applied onto the flattened surface and is patterned to the form of contact holes through exposure to light using electron beam lithography, which is followed by development through wet etching. Next, contact holes are created through ICP dry etching and wet etching with HF.

[0073] Next, a TiN film, which becomes a barrier metal, is formed in accordance with a spattering method, and after that the contact holes are filled in with Al. Next, the resist used for the creation of contact holes is removed and at the same time the Al layer that has been deposited on the resist is lifted off so as to form a p side electrode 68 and an n side electrode 69. Thus, the basic structure of the semiconductor light-receiving element according to Example 2 of the present invention is complete.

[0074] FIGS. 18A and 18B are diagrams illustrating the structure of the semiconductor light-receiving element according to Example 2 of the present invention. FIG. 18A is a top diagram and FIG. 18B is a cross-sectional diagram along the single-dotted chain line A-A′ in FIG. 18A. As illustrated in the figures, the three sides of the PIN type photodiode made of the p type Ge layer 66, the i type Ge layer 61 and the n type Si layer 59 are surrounded by the n type Ge layer 64 and the p type Ge layer 63, and therefore strain relaxation on the three end surfaces can be prevented from affecting the PIN type photodiode. In addition, a thyristor structure made of the p type Ge layer 66, the n type Ge layer 64, the p type Ge layer 63 and the n type Si layer 59 is formed on the two sides of the PIN type photodiode so as to function as a current blocking layer, which makes it possible to reduce a leak current. Here, a pseudo thyristor-like current block mechanism is formed of the p type Ge layer 66, the n type Ge layer 64, the p type Ge layer 63, the i type Si layer 53 and the n type Si layer 59 on the side opposite to the surface through which light enters.

[0075] Signal light that has entered into the core layer 55 propagates through the tapered portion 57 and reaches the n type Si layer 59 so as to be transmitted to the i type Ge layer 61 through evanescent coupling, and then is extracted from the p side electrode 68 as an electrical signal. Here, the n side electrode 69 is usually set to the ground potential.

[0076] Next, the semiconductor light-receiving element according to Example 3 of the present invention is described in reference to FIGS. 19A and 19B. The basic manufacturing process is the same as in the above description for Example 1 except that the semiconductor light-receiving element is a simple light-receiving element without the provision of a Si rib type waveguide, and therefore only the final structure is illustrated. Here, FIG. 19A is a top diagram and FIG. 19B is a cross-sectional diagram along the single-dotted chain line A-A′ in FIG. 19A.

[0077] An SOI substrate is prepared where a single crystal i type Si layer having a thickness of 0.3 μm is provided on a Si substrate 71 of which the main surface is (001) with a BOX layer 72 having a thickness of 3.0 μm in between. Next, a mask for ion implantation having openings only for the regions in which a light-receiving portion and a thyristor portion are to be formed is made on the i type Si layer, and this mask for ion implantation is used as a mask to implant B ions. In this case, the implantation energy is 40 keV and the dosage amount is 5.0×10¹⁴ cm⁻². Next, an annealing process for activating the implanted B ions is carried out for one second at 1050° C. so as to form a p type Si layer 73. The concentration of impurities in the p type Si layer 73 is 1.0×10¹⁸ cm⁻³ to 3.0×10¹⁹ cm⁻³.

[0078] Next, a mask for selective growth made of a SiO₂ film is provided, and an i type Ge layer 74 having a thickness of 0.8 μm is selectively grown on the p type Si layer 73 in accordance with a low pressure CVD method. In this case, GeH₄ is used as a material gas and H₂ is used as a carrier gas. First, the layer is grown by 0.1 μm at a low temperature of 400° C., and then the layer is grown by 0.7 μm at a high temperature of 700° C.

[0079] Next, a mask for ion implantation having an opening only for a portion in which a thyristor structure is to be formed is made of a SiO₂ film, and this mask for ion implantation is used as a mask to implant P ions so that the entirety of the i type Ge layer 74 is doped. In this case, the implantation energy is 600 keV and the dosage amount is 1.0×10¹⁴ cm⁻². As a result, the concentration of impurities in the n type Ge layer 75 is 1.0×10¹⁸ cm⁻³ to 2.0×10¹⁸ cm⁻³.

[0080] Next, the same mask for ion implantation is used as a mask to shallowly implant B ions so that the upper layer, having a thickness of 0.5 μm, is compensated and becomes a p type Ge layer 76 while the bottom layer, having a thickness of 0.3 μm, is not compensated and remains as an n type Ge layer 75. In this case, the implantation energy is 100 keV and the dosage amount is 3.0×10₁₄ cm⁻². As a result, the concentration of impurities in the p type Ge layer 76 is 1.0×10¹⁸ cm⁻³ to 1.0×10¹⁹ cm⁻³.

[0081] Next, the mask for ion implantation is removed with hydrofluoric acid, and after that a new mask for ion implantation having an opening only for the region in which a photodiode is to be formed is made of a SiO₂ film. This mask for ion implantation is used as a mask to implant P ions so that the upper layer, having a thickness of 0.3 μm, is converted to an n type Ge layer 77. In this case, the implantation energy is 40 keV and the dosage amount is 1.0×10¹⁴ cm⁻². As a result, the concentration of impurities in the n type Ge layer 77 is 4.0×10¹⁸ cm⁻³ to 1.0×10¹⁹ cm⁻³, and the concentration of impurities located at a depth of 0.3 μm from the surface is 1.0×10¹⁶ cm⁻³.

[0082] Next, the mask for ion implantation is removed with hydrofluoric acid, and after that annealing for activation is carried out for one second at 1050° C. so as to activate the implanted impurities. Next, a SiO₂ film, which becomes a passivation film 78 is deposited, and after that the surface of the SiO₂ film and the Ge layer is flattened in accordance with a CMP method.

[0083] Next, a resist is applied onto the flattened surface and is patterned to the form of contact holes through exposure to light using electron beam lithography, which is followed by development through wet etching. Next, contact holes are created through ICP dry etching and wet etching with HF.

[0084] Next, a TiN film, which becomes a barrier metal, is formed in accordance with a spattering method, and after that the contact holes are filled in with Al. Next, the resist used for the creation of contact holes is removed and at the same time the Al layer that has been deposited on the resist is lifted off so as to form an annular n side electrode 79 and a p side electrode 80. Thus, the basic structure of the semiconductor light-receiving element according to Example 3 of the present invention is complete.

[0085] As described above, in Example 3 of the present invention, all sides of the PIN type photodiode are surrounded by the p type Ge layer 76 and the n type Ge layer 75, and therefore strain relaxation in the end portion can be completely prevented from affecting the PIN type photodiode. In addition, a thyristor structure made of the n type Ge layer 77, the p type Ge layer 76, the n type Ge layer 75 and the p type Si layer 73 is formed so as to completely surround the PIN type photodiode and functions as a current blocking layer, which makes it possible to reduce the leak current.

[0086] In the semiconductor light-receiving element in Example 3 as illustrated in the figure, light enters from above, transmits through the n type Ge layer 77, is absorbed by the i type Ge layer 74 and is extracted from the n side electrode 79 as an electrical signal. In this case, the p side electrode 80 is set to the ground potential.

[0087] All examples and conditional language provided herein are intended for the pedagogical purposes of aiding the reader in understanding the invention and the concepts contributed by the inventor to further the art, and are not to be construed as limitations to such specifically recited examples and conditions, nor does the organization of such examples in the specification relate to a showing of the superiority and inferiority of the invention. Although one or more embodiments of the present invention have been described in detail, it should be understood that the various changes, substitutions, and alterations could be made hereto without departing from the spirit and scope of the invention.