JP2014165224A - Optical semiconductor device and optical semiconductor device manufacturing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a PD array having a plurality of light-receiving parts which can reduce cross talk to a further degree; and provide a method of manufacturing a PD array simpler and at low cost.SOLUTION: An optical semiconductor device of the present embodiment comprises: an optical semiconductor element in which a plurality of optical elements each composed of a semiconductor junction layer formed in a semiconductor composite layer laminated on a semiconductor substrate; and a base plate joined to the semiconductor substrate, in which a groove piercing the semiconductor composite layer and the semiconductor substrate to reach the base plate is formed every between neighboring optical elements formed in the optical semiconductor element.

Description

  BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical semiconductor device and a method for manufacturing the optical semiconductor device. More specifically, the present invention is used in an optical semiconductor device mainly used for optical fiber communication, and is a light receiving element, for example, a photodiode (PD), : Photodiode), an optical semiconductor device including the optical element, and a method for manufacturing the optical semiconductor device.

  In recent optical wavelength communication for supporting large-capacity optical communication, a communication device having more optical channels is required. On the other hand, there is a strong demand for downsizing and cost reduction of optical components constituting communication equipment.

  In particular, in order to reduce the size of an optical module that is an optical component, a configuration that can reduce the manufacturing cost of the optical module while having a large number of channels is preferable. By using an array type optical semiconductor element in which a plurality of light receiving portions or light emitting portions are integrated, further downsizing of the optical module has been promoted. Hereinafter, as an example of an array-type module in which a plurality of light receiving units are integrated, a multi-channel optical signal monitor that monitors each optical signal power in an optical system having a plurality of optical signal channels is taken as an example. The multi-channel optical signal monitor is configured around a PD array which is an optical semiconductor element. Device examples serving as the basis of a multi-channel optical signal monitor will be described with reference to FIGS.

  FIG. 1 shows a cross-sectional view of a first conventional PD array, and shows a cross-sectional configuration of a PD array 100 which is an optical semiconductor element having a plurality of light receiving portions. The PD array 100 is manufactured using a semiconductor substrate 101 made of, for example, InP. An example using an n-type semiconductor substrate 101 will be described with reference to FIG. In the PD array 100, a semiconductor composite layer in which a light absorption layer 102, an n-type buried layer 103, and a protective film 107 are sequentially stacked is stacked on an n-type semiconductor substrate 101. Further, the p-type diffusion region 104 is formed in a part of the n-type buried layer 103. A p-type electrode 105 is formed in the p-type diffusion region 104, while a first n-type electrode 106 is also formed on the back surface of the n-type semiconductor substrate 101.

  A region of the light absorption layer 102 sandwiched between the n-type semiconductor substrate 101 and the p-type diffusion region 104 (a portion surrounded by a dotted line in FIG. 1) is a first light absorption region 110. In the first light absorption region 110, a pin semiconductor junction layer is configured. In the first light absorption region 110 shown in FIG. 1, a p-i-n semiconductor junction layer is formed so as to be partially depleted, or reversely through the p-type electrode 105 and the first n-type electrode 106. A fully depleted layer is formed by applying a bias voltage. Therefore, the incident optical signal generates a carrier pair of electrons and holes and is taken out as a photocurrent, that is, an electric signal.

  In the conventional PD shown in FIG. 1, an optical signal is incident from the p-type diffusion region 104 side. A region where an optical signal is incident from the p-type diffusion region 104 side is referred to as a light receiving unit 111. Since the PD array 100 having a plurality of light receiving portions 111 is collectively formed by a semiconductor process, the light receiving characteristics of the respective channels are uniform. As typical dimensions, the interval between the light receiving portions is 250 μm, each light receiving diameter is 80 μm, and the thickness of the light absorbing layer is 3.6 μm. For a 24-channel PD array, a PD array having dimensions of 6000 μm in length, 350 μm in width, and 150 μm in thickness has been developed.

  FIG. 2 shows an optical module using the PD array of the first conventional example, and shows an embodiment in which the PD array 100 is put into practical use. FIG. 3 shows a cross-sectional view of an optical module using the PD array of the first conventional example, and shows a cross-sectional structure of the PD array 100. 2 and 3, in order to ensure the reliability of the PD array 100, the PD array 100 is fixed in a ceramic package 201 and the glass lid 210 is hermetically sealed with a metal solder thin film 206. This is an optical module. The PD array 100 is fixed by using a metal solder thin film 203 on a common ground electrode 202 formed on the inner surface of the package in the first n-type electrode 106 formed on the back surface. On the other hand, each p-type electrode 105 is connected to each signal electrode 205 wired in the package 201 by a bonding wire 204. Each signal electrode 205 and common ground electrode 202 in the package 201 are taken out from the inside of the package (not shown).

JP 2007-266251 A

  The conventional PD array 100 shown in FIGS. 1 and 2 has a problem that crosstalk occurs. That is, when an optical signal is incident on the light receiving unit 111 of a certain predetermined channel for which an optical signal is desired to be input, the incident light is not completely absorbed by the first light absorption region 110 of the corresponding light receiving unit. Then, there is a phenomenon that part of the incident light reaches the back surface of the PD array as it is and repeats reflection and scattering again to return to the absorption layer. When an optical signal is incident on the light receiving unit 111, a part of the light that has repeatedly reflected and scattered enters the first light absorption region 110 of the light receiving unit 111 of other channels other than the predetermined channel, that is, other channels that are not originally desired to receive light. Will reach. Part of the light that has repeatedly reflected and scattered may be output as an unnecessary photocurrent due to light absorption in the first light absorption region 110. Therefore, the generation of a photocurrent which is completely unnecessary in other channels causes crosstalk, and there is a problem that the performance as an optical monitor is deteriorated.

  The problem of lowering the performance as an optical monitor is caused by the structure of the optical semiconductor element having a plurality of light receiving portions 111. Therefore, even if the module 200 is hermetically sealed as shown in FIG. 2, the problem cannot be avoided. In particular, the influence of crosstalk tends to become a problem when the light receiving portion pitch is 500 μm or less. However, if the light input / output device using an optical fiber is generally assumed, the maximum pitch of the light receiving portions may be 250 μm at the maximum.

Crosstalk is represented by the ratio (10 * Log10 (IX _{i + 1} / IC _i )) between the photocurrent IC _i and the photocurrent IX _{i + 1} . Photocurrent IC _i is incident after sealing the PD array 100 to the package 201, the light of the 1550nm light receiving portion 111 of the input channel _{C i} to desired Bring cleaved end face of the single mode fiber to the glass lid 210 on the surfaces of an optical current drawn from the input channels C _i desired when allowed to. The photocurrent IX _{i + 1} is a photocurrent that leaks out from the desired input channel C _i when the same light is incident on the light receiving surface 111 of the channel C _{i + 1} adjacent to the desired input cannel.

  Crosstalk also affects the distance between the light input end and the light receiving unit 111. The distance between the light input end and the light receiving unit 111 is practically used, and the light input position is the optical path length from the cleavage end surface of the single mode fiber on the glass lid surface to the light receiving surface of the PD array, and the optical path length is 350 μm. Yes. In addition, a reverse bias voltage of −5 V is applied to each PD via the p-type electrode 105 and the first n-type electrode 106.

  FIG. 4 shows a cross-sectional view of a second conventional PD array. As shown in FIG. 4, a second light absorption region 120 that absorbs unnecessary reflected and scattered light generated between channels is arranged between the light receiving portions 111, although it does not participate in light signal reception. A PD array 100 has been developed (see, for example, Patent Document 1). The second light absorption region 120 formed in the configuration shown in FIG. 4 can be handled only by changing the mask in the semiconductor process for manufacturing FIG.

  FIG. 5 shows a cross-sectional view of an optical module using the second conventional PD array. By changing the optical module configuration of FIG. 3 from the conventional embodiment to the optical module configuration of FIG. 5, the crosstalk between adjacent channels is reduced by about 5 to 10 dB when the input light wavelength is 1550 nm. The improvement effect is obtained.

  However, in the configuration in which the second light absorption region 120 is inserted, the insertion space of the second light absorption region 120 is inevitably reduced when the interval between the light receiving portions is to be reduced in order to further reduce the size. Therefore, there is a problem that it is impossible to expect the amount of reflected / scattered light absorption, that is, the reduction of crosstalk. In addition, since the first light absorption region 110 and the second light absorption region 120 need to be formed separately from each other, there has been a problem that formation on a semiconductor process becomes difficult.

  FIG. 6 shows a cross-sectional view of a third conventional PD array. As shown in FIG. 6, a PD array in which both the p-type electrode 105 and the second n-type electrode 116 are formed on one side of the PD array has been developed as a PD electrode configuration. That is, the electrode on the n-type semiconductor substrate 101 side is taken out from the front surface side of the PD array, and the metal thin film 117 used for fixing the solder is disposed on the back surface side of the n-type semiconductor substrate 101.

  FIG. 7 shows a cross-sectional view of an optical module using a third conventional PD array. As shown in FIG. 7, the PD array 100 is fixed in the same manner as the conventional form shown in FIG. 3, but the bonding wire 204b connected to the second n-type electrode 116 taken out on the surface is common in the package. It is connected to a ground electrode (not shown). Even in the PD array form shown in FIG. 7, the problem of crosstalk which remains a problem as in the conventional form shown in FIG. 5 remains.

  FIG. 8 is a perspective view showing an optical module using individual PDs. In order to suppress the reflected and scattered light in the PD array made of a single semiconductor substrate, as shown in FIG. 8, a configuration in which individual PD300 physically separated is mounted one by one at a desired channel interval. The lowest crosstalk. The amount of crosstalk that is ideally required as a multi-channel optical signal monitor is the level of an array mounting of PDs that are physically separated as shown in FIG.

  However, the method of mounting each individual PD 300 with high pitch accuracy has a limit in the number of mounted PDs, that is, the number of channels. Also, metal solder is used for optical semiconductors from the viewpoint of fixed reliability. However, if solder reflow is performed each time an individual PD is mounted, problems such as solder oxidation deterioration due to remelting occur, so the point of solder oxidation deterioration However, there is a problem that there is a limit to the number of mounted channels, that is, the number of channels.

  The present invention has been made in view of the above-described problems that have occurred in the prior art, and an object of the present invention is to provide a PD array having a plurality of light receiving portions that can further reduce crosstalk. A further object is to provide a method for producing a PD array more easily and at a lower cost.

  In order to achieve the above object, according to the present invention, a plurality of optical elements each including a semiconductor junction layer formed in a semiconductor composite layer laminated on a semiconductor substrate are formed. An optical semiconductor element, and a substrate bonded to the semiconductor substrate, penetrating the semiconductor composite layer and the semiconductor substrate between the adjacent optical elements formed in the optical semiconductor element, A reaching groove is formed.

  A second aspect of the present invention is the optical semiconductor device according to the first aspect, characterized in that vias for connecting the metal thin films formed on the front surface and the back surface of the substrate are formed.

  A third aspect of the present invention is the optical semiconductor device according to the first or second aspect, wherein the optical elements are two-dimensionally arranged.

  The invention according to claim 4 is a method of manufacturing an optical semiconductor device, wherein the optical semiconductor element includes a plurality of optical elements each including a semiconductor junction layer formed in a semiconductor composite layer stacked on a semiconductor substrate. A step of bonding the semiconductor substrate and the substrate, and a groove penetrating the semiconductor composite layer and the semiconductor substrate and reaching the substrate between the adjacent optical elements formed in the optical semiconductor element And a step of performing.

It is sectional drawing of PD array of the 1st prior art example. It is a figure which shows the optical module using PD array of the 1st prior art example. It is sectional drawing of the optical module using PD array of the 1st prior art example. It is sectional drawing of the PD array of the 2nd prior art example. It is sectional drawing of the optical module using the PD array of the 2nd prior art example. It is sectional drawing of PD array of a 3rd prior art example. It is sectional drawing of the optical module using PD array of the 3rd prior art example. It is a perspective view showing the optical module using individual PD. It is a figure showing the manufacturing process in the 1st Embodiment of this invention. It is a figure showing the manufacturing process in the 1st Embodiment of this invention. It is a figure showing the manufacturing process in the 1st Embodiment of this invention. 1 is a cross-sectional view of an optical semiconductor device according to a first embodiment of the present invention. It is a perspective view of the optical semiconductor device concerning the 2nd Embodiment of this invention. It is a perspective view of the optical semiconductor device concerning the 3rd Embodiment of this invention. It is sectional drawing of the optical semiconductor device concerning the 4th Embodiment of this invention. It is a figure showing the manufacturing process in the 5th Embodiment of this invention.

  Embodiments of the present invention will be described below with reference to the drawings. In addition, since it is the same structure unless it refuses especially about the part which is the same component as a prior art example, description is abbreviate | omitted. In addition, in order to take up a multi-channel configuration, repetitive component notations in the figure are omitted unless otherwise specified.

(First embodiment)
First, a method for realizing the first embodiment will be described with reference to FIGS. 9A, 9B, 10 and 11, step by step.

  FIG. 9A shows a sectional view of the PD array (optical semiconductor element) 100 according to the first embodiment of the present invention, and shows a sectional configuration of the PD array 100 having a plurality of light receiving portions. In the first embodiment, an example using a PD array 100 in which both the p-type electrode 105 and the second n-type electrode 116 are formed on the surface side shown in FIG. The PD array 100 is manufactured using a semiconductor substrate 101 made of, for example, InP. In the PD array 100, a semiconductor composite layer in which a light absorption layer 102, an n-type buried layer 103, and a protective film 107 are sequentially stacked is stacked on an n-type semiconductor substrate 101. Further, the p-type diffusion region 104 is formed in a part of the n-type buried layer 103. A p-type electrode 105 is formed in the p-type diffusion region 104, while a second n-type electrode 116 is formed on the surface of the PD array by taking out the electrode from the n-type semiconductor substrate 101 side. A metal thin film 117 used for soldering is disposed on the back side of the n-type semiconductor substrate 101.

  A region of the light absorption layer 102 sandwiched between the n-type semiconductor substrate 101 and the p-type diffusion region 104 (a portion surrounded by a dotted line in FIG. 9A) is a first light absorption region 110. In the first light absorption region 110, a pin semiconductor junction layer is configured. A region where an optical signal is incident from the p-type diffusion region 104 side is a light receiving unit 111.

  The point to be noted in the first embodiment of the present invention is that, as shown in FIG. 9A, unlike the third conventional example shown in FIG. The second light absorption region 120 is not required. The second light absorption region 120 shown in FIG. 6 absorbs unnecessary reflected and scattered light generated between the channels, but in the first embodiment of the present invention, the PD array 100 is separated and the light receiving parts are also separated. This is because unnecessary reflected and scattered light generated between the channels does not occur.

  In the first embodiment, similarly to the conventional example, the PD array 100 has a light receiving portion interval of 250 μm, each light receiving diameter of 80 μm, and a thickness of the light absorbing layer 3 of 3.6 μm. As the dimensions of the PD array of the channel, a PD array having a length of 6000 μm, a width of 350 μm, and a thickness of 150 μm was used.

  FIG. 9B shows a cross-sectional view of the carrier 400 in the first embodiment of the present invention. As shown in FIG. 9B, a carrier 400 on which the PD array 100 is mounted is prepared. In this embodiment, a Si substrate is used as the substrate 401 of the carrier 400, and the dimensions of the carrier 400 are approximately the same size as the PD array 100, 6400 μm long, 400 μm wide, and 300 μm thick. A thin film metal 402a is deposited or plated on the surface of the carrier 400 on which the PD array 100 is fixed. A thin film metal 402b is deposited or plated on the back surface opposite to the surface on which the PD array 100 is fixed. In the first embodiment, Au having a surface thickness of 2 μm is used for the thin film metals 402a and 402b. Further, an AuSn solder thin film 403 is pre-coated with a thickness of 7 μm on the thin film metal side of the surface.

  Next, as shown in FIG. 10, the PD array was fixed to the carrier substrate by reflowing the solder thin film 403.

  Further, as shown in FIG. 11, a cut groove 500 is formed by dicing between the light receiving portions 111 serving as the respective signal channels to completely separate the light receiving portions. That is, the PD array portion that was originally the same is completely separated, and the carrier 400 is further cut.

  The configuration of the first embodiment of the present invention will be described with reference to FIGS. 9A, 9B, 10 and 11. The optical semiconductor device according to the first embodiment is formed on the semiconductor substrate 101. FIG. It consists of a pin semiconductor junction layer (light absorption region 110) formed on a laminated semiconductor composite layer (a composite layer in which the light absorption layer 102, the n-type buried layer 103, and the protective film 107 are sequentially stacked). A PD (optical element) includes a plurality of PD arrays (optical semiconductor elements) 100 and a substrate 401 in a carrier 400 bonded to a semiconductor substrate 101. A cut groove 500 that penetrates the semiconductor composite layer and the semiconductor substrate 101 and reaches the substrate 401 is formed between adjacent PDs (optical elements) formed in the PD array (optical semiconductor element) 100.

  A feature of the present invention is that a plurality of PDs are held in an array without changing the light receiving unit pitch by cutting the carrier 400 to a part rather than completely cutting it. is there. The kerf width of the dicing blade used in the first embodiment is 30 μm. In the cut groove 500, the cut amount entering from the surface of the carrier 400 is about 20 μm, and any cut amount may be used as long as the light receiving portions 111 of the original PD array 100 are completely separated. That is, the cut amount entered from the surface of the carrier 400 is not particularly specified as long as the plurality of light receiving portions are continuously fixed on the same substrate with the initial pitch after the PD array cut. In other words, the carrier does not necessarily have to be cut, and the PD array 100 is necessarily separated.

  The PD array produced as described above was assembled into an optical module as shown in FIG. FIG. 12 shows a cross-sectional view of the optical semiconductor device according to the first embodiment of the present invention. First, on the bottom surface of the package 201, the PD array 100 provided with the carrier 400 prepared earlier was fixed on the thin film metal 202 using Au as the surface metal with the metal thin film solder 203 having the same size as the carrier. After fixing, the wire bonding 204a connects the p-type electrode 105 on the PD array 100 and the electrode of the package 201, and the wire bonding 204b connects the second n-type electrode 116 on the PD array 100 and the electrode of the package 201. (Not shown). Finally, the glass lid 210 is used to hermetically seal the package.

  The crosstalk characteristic of the optical module according to the first embodiment manufactured as shown in FIG. 12 was compared with that of the conventional module FIG. In the conventional configuration, the crosstalk was about −40 dB, but according to the first embodiment, it was found that the crosstalk was reduced to −50 dB or less. The numerical value of −50 dB or less corresponds to the crosstalk level of an optical module manufactured by fixing each individual PD 300 shown in FIG.

  According to the first embodiment, there is a feature that it is only necessary to prepare a PD array that does not require the second light absorption region 120 as shown in FIG. Further, according to the first embodiment, after being fixed to another carrier 400 and then divided into individual PDs, the process of dividing into individual PDs can be realized by cutting the PD array. There is a feature that the assembly process of the PD array module can be simplified.

(Second Embodiment)
FIG. 13 is a perspective view of an optical semiconductor device according to the second embodiment of the present invention. The second embodiment is an embodiment where the n-type electrode of the PD array 100 is arranged on the back surface. In the second embodiment, before the cut groove 500 is formed, the PD array 100 is placed in the ceramic package 201 in order to ensure the reliability of the PD array 100, for example, as in the conventional embodiment shown in FIG. And an optical module in which the glass lid 210 is hermetically sealed with a metal solder thin film 206. In the PD array 100, the solder thin film 403 was reflowed and fixed to the carrier substrate in the first n-type electrode 106 formed on the back surface. On the other hand, each p-type electrode 105 is connected to each signal electrode 205 wired in the package 201 by a bonding wire 204a. The common ground electrode 202 is connected to the thin film metal 402a on the carrier 400 via a bonding wire 204c. Each signal electrode 205 and common ground electrode 202 in the package 201 are taken out from the inside of the package (not shown).

  The back electrode is originally formed as a common electrode. According to the second embodiment of the present invention, the PD array 100 is fixed to the carrier 400 and then separated by dicing. During the separation, the first n-type electrode 106 on the back surface, which was originally common, is divided. Therefore, the divided first n-type electrode 106 may be connected to the common ground electrode 202 in the package by the bonding wire 204 through the thin film metal 402a and the solder thin film 403. The increase in the bonding wires 204 can be handled automatically in the PD manufacturing process, and therefore, the process tact is not particularly increased.

  According to the second embodiment, unlike the first embodiment, in the process of manufacturing the PD array, the number of semiconductor process steps is increased as much as the process of bringing the n-type electrode on the surface becomes unnecessary. There is a feature that the effect that can be reduced is great.

(Third embodiment)
FIG. 14 is a perspective view of an optical semiconductor device according to the third embodiment of the present invention. Unlike the second embodiment, the third embodiment is an optical semiconductor device manufactured by stopping cutting into the carrier substrate halfway. According to the third embodiment, the divided back surface electrodes of the PD array 100 are connected by the uncut surface metal on the carrier substrate, so that the number of bonding wires can be reduced. .

(Fourth embodiment)
FIG. 15 shows a sectional view of a PD array (optical semiconductor element) 100 according to the fourth embodiment of the present invention, and shows a sectional configuration of the PD array 100 having a plurality of light receiving portions. The PD array 100 is manufactured using a semiconductor substrate 101 made of, for example, InP. In the PD array 100, a semiconductor composite layer in which a light absorption layer 102, an n-type buried layer 103, and a protective film 107 are sequentially stacked is stacked on an n-type semiconductor substrate 101. Further, the p-type diffusion region 104 is formed in a part of the n-type buried layer 103. A p-type electrode 105 is formed in the p-type diffusion region 104, while a first n-type electrode 106 is also formed on the back surface of the n-type semiconductor substrate 101.

  A region of the light absorption layer 102 sandwiched between the n-type semiconductor substrate 101 and the p-type diffusion region 104 (a portion surrounded by a dotted line in FIG. 15) is a first light absorption region 110. In the first light absorption region 110, a pin semiconductor junction layer is configured. A region where an optical signal is incident from the p-type diffusion region 104 side is a light receiving unit 111.

  Unlike the second or third embodiment, the fourth embodiment shows a mode in which the metal thin film 410 on the front surface of the carrier 400 and the metal thin film 411 on the back surface are electrically connected in advance via vias 600. ing. For the substrate 401 of the carrier 400 used in the fourth embodiment, alumina having a thickness of 100 μm was used. The metal type disposed in the via 600 may be a hollow type or a filling type. The arrangement of the via 600 is 250 μm, which is the same as the pitch of the light receiving portions 111 of the PD array 100 to be fixed, and the diameter of the via 600 is 100 μm. Before dicing cutting, the PD array 100 is fixed by the solder thin film 403 on the metal thin film 410 on the surface of the carrier so that one via 600 is arranged for one light receiving portion 111 of the PD array 100.

  According to the fourth embodiment, unlike the second and third embodiments, a bonding wire to the back surface electrode is not required, and is fixed in the package via thin film solder as in the first embodiment. Conducting is possible just by doing. Further, unlike the first embodiment, in the process of manufacturing the PD array, there is a great effect that the number of semiconductor process steps can be reduced to the extent that the process of bringing the n-type electrode on the surface becomes unnecessary. There is.

(Fifth embodiment)
FIGS. 16A, 16B, and 16C show manufacturing steps in the fifth embodiment of the present invention. Unlike the fourth embodiment, the fifth embodiment gives an example in which the used PD array is two-dimensionally developed.

  As shown in FIG. 16A, in the two-dimensional array PD array 700 used in the fifth embodiment, PD light receiving units 111 are arranged at a pitch of 250 μm in the vertical direction and the horizontal direction. Along with the arrangement of the PD light receivers 111, vias 600 formed in the carrier 400 are also two-dimensionally arranged at the same pitch.

  As shown in FIG. 16B, the two-dimensional array PD array 700 is joined to a carrier 400 having vias 600. Of course, before dicing cutting, the two-dimensional array PD array 700 is fixed with a solder thin film so that one via 600 is arranged for one light receiving portion 111.

  As shown in FIG. 16C, a cut groove 500 is formed in the two-dimensional array PD array 700 bonded to the carrier 400 having the via 600 by dicing between the light receiving portions 111 to be each signal channel, Separate the light receiving parts.

  Unlike the one-dimensional array PD array as seen in the first to fourth embodiments, the two-dimensional array PD array receives crosstalk from the vertical direction and the horizontal direction. Therefore, as in the fifth embodiment of the present invention, in the completely separated PD mounting mode, the effect of minimizing the influence of crosstalk from the vertical direction and the horizontal direction is great.

  As mentioned above, in each embodiment, although dicing was mentioned as a means to cut | disconnect PD array, this invention is not limited to a dicing means. In addition to dicing, for example, laser dicing, sand blasting, ion milling, reactive ion etching (RIE), or the like may be used. Depending on the type of means for cutting, the thickness of the PD array that can be cut is determined. However, in adjusting the thickness of the PD array, the substrate thickness of the PD array is set to a predetermined thickness within a range that can be actually processed. It is possible to cope with this by polishing to a certain extent.

  The present invention is not limited to the form described in each embodiment or the form shown in the drawings. For example, even if the illustrated PD array is four arrays, it is only a simplified illustration of what is intended by the invention, and the number of channels is not limited. The same applies to the number of arrays of the two-dimensional array PD array.

  The same applies to points relating to the size of the PD array and the substrate. For example, the pitch between PDs, the light receiving diameter size, the substrate thickness, the substrate via pitch, and the substrate via diameter are not limited to the sizes described in this specification.

  Furthermore, it goes without saying that the material configuration used for the PD array, the substrate, the package, and the like is not limited to the material configuration described in this specification. For example, p-type and n-type may be interchanged in the semiconductor configuration of the PD array, and even if a new layer configuration is inserted to improve the performance of the PD, as long as the implementation of the present invention is not hindered. It is not limited at all.

  Further, although the array type module in which the light receiving parts are integrated has been described, it may be configured as an array type module in which not only the light receiving parts but also the light emitting parts are integrated.

DESCRIPTION OF SYMBOLS 100 PD array 101 Semiconductor substrate 102 Light absorption layer 103 N type buried layer 104, 104a, 104b P type diffusion region 105 P type electrode 106 First n type electrode 107 Protective film 110 First light absorption region 111 Light receiving portion 116 2 n-type electrode 117 Metal thin film 120 Second light absorption region 200 Optical module 201 Package 210 Glass lid 202 Common ground electrode 203, 403 Solder thin film 204, 204a, 204b, 204c Bonding wire 205 Signal electrode 206 For airtight sealing Solder thin film 300 Individual PD
400 carrier 401 substrate 402a, 402b thin film metal 410 metal thin film 411 metal thin film 411 back metal thin film 500 cut groove 600 via 700 two-dimensional array PD array

Claims (4)

A plurality of optical semiconductor elements formed of a semiconductor junction layer formed in a semiconductor composite layer laminated on a semiconductor substrate; and
A substrate bonded to the semiconductor substrate;
An optical semiconductor device, wherein a groove that penetrates through the semiconductor composite layer and the semiconductor substrate and reaches the substrate is formed between the adjacent optical elements formed in the optical semiconductor element. 2. The optical semiconductor device according to claim 1, wherein vias for connecting metal thin films formed on the front surface and the back surface of the substrate are formed. The optical semiconductor device according to claim 1, wherein the optical elements are arranged two-dimensionally. Bonding the semiconductor substrate and the substrate in the optical semiconductor element in which a plurality of optical elements composed of semiconductor junction layers formed in a semiconductor composite layer laminated on the semiconductor substrate are formed; and
An optical semiconductor device comprising: a step of passing through the semiconductor composite layer and the semiconductor substrate and forming a groove reaching the substrate between the adjacent optical devices formed in the optical semiconductor device Manufacturing method.

Images