JP3620772B2 - Manufacturing method of semiconductor light receiving element - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a method for manufacturing a semiconductor light receiving element.
[0002]
[Prior art]
An outline of the refractive semiconductor light receiving element is shown in FIG. In FIG. 2, reference numeral 11 denotes a light incident end face, 12 denotes a 1 μm thick p-InP layer, 13 denotes a 1 μm thick InGaAs light receiving layer, 14 denotes a 1 μm thick n-InP layer, 15 denotes a semi-insulating InP substrate, and 16 denotes a p electrode. , 17 respectively indicate n electrodes.
As shown in FIG. 2, a chip is formed by cleavage or the like from the vicinity of the light incident inclined end face. When light (L) is incident on the light receiving element from a fiber or the like, the light emitted from the fiber or the like has a spread with respect to the distance, and therefore depends on how the emitted light spreads. It is necessary to set a distance of about several tens of μm or less from the light emitting end face of a fiber or the like. For this reason, as shown in FIG. 2, it is necessary to manufacture the cleaved end face 18 of the substrate 15 in the vicinity of the light incident end face 11 with high accuracy.
[0003]
Usually, in the cleavage, the substrate thickness is reduced to about 100 μm, and a wafer having a structure in which the refractive semiconductor light receiving element portion is arranged in parallel with the groove portion of the reverse mesa as shown in FIG. A cleaving flaw 20 is made by a needle 19 or the like, and a mechanical shock is applied thereto so that the flaw portion becomes a starting point of cleaving, so that the entire wafer is cleaved. Here, the needle 19 that forms the cleavage flaw 20 has a tip or a cone-shaped pyramid, and the tip is processed extremely finely. However, when a deep mesa exists like a refractive semiconductor light-receiving element, the mesa portion The groove width is sufficiently wide so that the needle 19 does not contact the mesa portion, and the cleavage flaw 20 is formed, but in order to prevent the needle from coming into contact with the upper end portion of the reverse mesa, it is several tens from the upper end of the reverse mesa. It was necessary to form cleaved scratches 20 separated by μm or more.
[0004]
In addition, when the spread angle of light emitted from a fiber or the like is large, it is important to bring the fiber close to the pole vicinity of the incident end face. For this reason, it is necessary to cleave the flaw 20 near the pole of the incident end face. In this case, the needle 19 is forcibly pressed near the entrance end face to form a cleavage flaw, but the cone or pyramid of the needle contacts the mesa portion of the entrance end face 11 near the position where the cleavage flaw 20 is applied. There was a problem that the position to be scratched was shifted or the position variation became large. For these reasons, there is a problem that the yield of chips cut out from the wafer is lowered.
[0005]
[Problems to be solved by the invention]
An object of the present invention is to provide a manufacturing method capable of cleaving in the vicinity of a light incident end with a high yield when cutting out an element from a wafer in manufacturing a refractive light receiving element.
[0006]
[Means for Solving the Problems]
According to a first aspect of the present invention, there is provided a method of manufacturing a semiconductor light-receiving element comprising: a light-receiving portion having a semiconductor multilayer structure including a light-receiving layer formed on a surface of a substrate; In the production of a refractive semiconductor light-receiving element in which an incident end face is provided so that incident light is refracted at the light incident end face so that the incident light passes through the light receiving layer obliquely with respect to the layer thickness direction. After the reverse mesa etching groove forming step for forming the end surface, the depth from the upper side of the light incident end surface to the bottom surface of the etching groove is H, and the depth of the cleavage flaw is Hc, and the needle for forming the cleavage flaw is formed. When the tip opening angle is 2δ, the groove width 2W at the upper end of the reverse mesa etching groove is 2W ≦ 2 (H + Hc) tan δ.
[0008]
DETAILED DESCRIPTION OF THE INVENTION
The manufacturing method of the semiconductor light receiving element of the present invention is such that, in the manufacture of the refractive semiconductor light receiving element, by setting the groove width at the upper end of the reverse mesa etching groove after the step of forming the reverse mesa etching groove for forming the light incident end face, After the end face formation etching, a groove having a shape as shown in FIG. 1 is formed.
[0009]
FIG. 1 also schematically shows a needle for simultaneously forming a cleavage flaw. As shown in FIG. 1, when the depth from the upper side to the bottom surface of the light incident end face 11 is H, the depth of the cleavage flaw is Hc, and the tip opening angle of the needle for forming the cleavage flaw is 2δ, When the groove width at the upper end of the reverse mesa after manufacturing the light incident end face is set to 2W ≦ 2 (H + Hc) tan δ in the figure, and the cleavage needle 19 is lowered from above to the vicinity of the groove center, Works as a guide and can form a cleavage flaw in the center with good reproducibility. Even if the groove width is narrowed and the upper sides of both mesas are damaged by the contact of the inclined portion of the needle 19, the portion always acts as a guide, and therefore a cleavage flaw can be formed in the central portion with good reproducibility. For this reason, by applying a mechanical shock, cleavage occurs with good reproducibility starting from the cleaved portion. Further, if the groove width 2W is set to 10 μm or less, the cleavage can be performed in the very vicinity of 5 μm or less from the light incident end face.
[0010]
FIG. 4 shows the etching when the (001) surface InP substrate 15 is etched by bromine-based etching when the etching groove width is sufficiently wide to expose the (111) plane to form a reverse mesa shape (θ = 55 degrees). It is a schematic diagram of a cross section. If the depth of the cleavage flaw is Hc and the groove width 2W at the upper end of the reverse mesa after manufacturing the light incident end face is 2W ≦ 2 (H + Hc) tan δ, the reverse mesa portion is formed on both sides of the needle. Thus, when the cleavage needle is lowered from above to the vicinity of the center of the groove, the upper side of the opposite mesa serves as a guide, and a cleavage flaw can be formed in the center with good reproducibility.
[0011]
Here, as a typical example, when light from a silica-based single mode fiber (SMF) is received with high efficiency without a lens system by a refractive semiconductor light receiving element, a silica-based single mode fiber (SMF: spot size is 10 μm). ) Has a beam diameter of 15 μm and a light intensity of 99% or more, the depth D of the light incident end face 11 must be at least 15 μm. In order to set D to 15 μm, H was required to be 26 μm according to experiments. Usually, the opening angle of the tip of the needle used in the automatic cleaving flaw forming apparatus is often about 2δ = 60 degrees or more.
[0012]
Here, when 2δ = 60 degrees and the cleavage crack depth Hc is set to zero, the groove width of the upper end of the reverse mesa after manufacturing the light incident end face is 30 μm or less (2 × (26 × tan (30 degrees)) = When the end face forming mesa etching is performed so as to be 30) and the cleaving needle is lowered from above to the vicinity of the groove center, the upper side of the opposite mesa acts as a guide, and a cleavage flaw can be formed at a substantially central position with good reproducibility. Actually, since the needle is further lowered by a finite cleavage flaw depth Hc, the inclined portion of the needle is surely brought into contact with and guided by the reverse mesa portion. For this reason, by applying a mechanical shock, cleavage occurs with good reproducibility starting from the cleaved portion.
[0013]
This is different from the prior art in that in the formation of cleavage flaws, there is almost no positional displacement for scratching, and cleaving can be performed with desired accuracy in the vicinity of the incident end face.
[0014]
As described above, this manufacturing method has almost no variation in misalignment in the formation of cleavage flaws and can be cleaved in the vicinity of the incident end face. Therefore, it is possible to easily cut out the element while sufficiently satisfying the distance to the light incident end face required for optical coupling with a fiber or the like. In addition, since the device can be cut out using both sides of the etched groove as the light incident end face, the yield of the device from the wafer is greatly improved.
[0015]
【Example】
Hereinafter, although the suitable Example of this invention is described, this invention is not limited to this.
[0016]
[Example 1]
FIG. 5 is a diagram for explaining the first embodiment of the present invention. 11 is a light incident end face, 12 is a 1 μm thick p-InP layer, 13 is a 1 μm thick InGaAs light receiving layer, 14 is a 1 μm thick n-InP layer, 15 is a semi-insulating InP substrate, 16 is a p electrode, and 17 is an n electrode. It is. The light receiving layer area of the element is 14 μm × 20 μm. The light incident end face 11 was formed by wet etching using bromomethanol using a silicon nitride film mask. At this time, the light incident end face 11 was formed by utilizing the fact that the (111) A face was formed in a reverse mesa shape as shown in the figure by wet etching using bromomethanol on a (001) surface wafer. The groove width at the upper end of the reverse mesa etching groove after this step is 7 μm. Even if the groove width at the upper end of the reverse mesa etching groove is as narrow as 7 μm, the depth D of the light incident end face 11 is as deep as 20 μm and the depth H to the bottom is about 35 μm, so that a good incident end face can be formed. ing.
[0017]
When the narrow needle 19 having a needle tip opening angle of 2δ = 20 degrees is lowered from above to near the center of the groove and the cleaving flaw depth Hc is about 10 μm, the conical portion contacts the upper side of the opposite mesa. This part served as a guide, and a cleavage flaw was formed in the center with good reproducibility.
This satisfies the condition 2W = 7 <2 × (35 + 10) × tan (10 degrees) = 15.9 μm of claim 1.
The cleavage flaw Hc was formed over a length of about 200 μm in a part of the wafer edge. Of course, since the strength of the semiconductor is weaker than that of the needle, the reverse mesa part where the needle comes into contact is finely damaged. I was able to form. Therefore, by providing a mechanical shock here, it was possible to perform good cleavage over several centimeters starting from the cleavage flaw, and both sides of this cleavage could be used as the light incident end face of the light receiving element.
[0018]
In this embodiment, the groove width at the upper end of the reverse mesa etching groove after the process of forming the reverse mesa etching groove for forming the light incident end face is 7 μm. However, even if the groove width is 15 μm, good cleavage is possible at the groove center. It was. The reverse mesa portion may be formed using another wet etching solution or a dry etching method, or may be formed using another crystal plane or controlling the angle using the adhesion of the etching mask. .
[0019]
When a non-reflective film is formed on the incident end face, and light with a wavelength of 1.55 μm is introduced using a tip-end fiber, a large value of light receiving sensitivity of 0.9 A / W or more can be obtained with an applied reverse bias of 3 V, and a 3 dB band A high-speed operation of 40 GHz was possible. At this time, the protrusion of the substrate portion immediately below the light incident end face portion is as small as several microns, and hardly interferes with the proximity of the fiber.
[0020]
In this embodiment, the groove of the etching mask uses the linear stripe-shaped etching mask 21 as shown in FIG. 7A, but the uneven etching mask as shown in FIGS. 7B and 7C. Various shapes such as 22 and tapered concavo-convex etching mask 23 may be used as long as the width of the groove where the cleavage flaw is formed matches the purpose of the present invention.
[0021]
In this embodiment, the p-InP layer on the surface side is formed by crystal growth. However, in the crystal growth, an undoped InP layer is used, and the main conductivity type of the semiconductor on the surface side is changed to Zn diffusion or ion implantation. And subsequent annealing.
[0022]
The semiconductor light-receiving element portion is on a semiconductor layer having the first conductivity type, and includes a light-receiving layer and a Schottky formed of an intrinsic or first conductivity type semiconductor layer, a superlattice semiconductor layer, or a multiple quantum well semiconductor layer. A multilayer structure in which a semiconductor layer having a high Schottky barrier height having a Schottky barrier higher than the Schottky barrier between the light receiving layer and the Schottky electrode is interposed between the electrode and an electrode. the or a semiconductor light receiving element formed by configured on a substrate, the Schottky barrier height with high semiconductor _{layer, in 1-x-y Ga} x Al y As (0 ≦ x ≦ 1,0 ≦ y ≦ 1) or in _{1-x-y Ga x Al} y as (0 ≦ x ≦ 1,0 ≦ y ≦ 1) thin on the Part _{In 1-u Ga u as 1} -v P v (0 ≦ u ≦ 1,0 ≦ v ≦ 1) It may be constituted by a semiconductor light receiving device that.
[0023]
Further, in this embodiment, a semi-insulating InP is used as a substrate and an n-InP layer is used on the substrate side. However, even if a p-InP layer is used, the above-described p and n are reversed in the same manner. It can be manufactured, and can also be manufactured using an n-InP or p-InP substrate.
[0024]
Here, a bulk of uniform composition is used as the light receiving layer. However, a separate-absorption-graded-multiplication (SAGM) structure, a separate abstraction and multiplexing superstructure (SAM-SL) structure, and the like used for an avalanche photodiode are used. Needless to say, a semiconductor layer having a superlattice structure may be used. Further, it goes without saying that a material system such as InGaAlAs / InGaAsP or AlGaAs / GaAs system other than the InGaAsP / InP system or a material system with inherent strain may be used.
[0025]
[Example 2]
FIG. 6 is a diagram for explaining a second embodiment of the present invention. 11 is a light incident end face, 12 is a 1 μm thick p-InP layer, 13 is a 1 μm thick InGaAs light receiving layer, 14 is a 1 μm thick n-InP layer, 15 is a semi-insulating InP substrate, 16 is a p electrode, and 17 is an n electrode. It is. The light receiving layer area of the element is 30 μm × 50 μm. The light incident end face was formed by wet etching using bromomethanol using a silicon nitride film mask. At this time, the light incident end face was formed by utilizing the fact that the (111) A face was formed in a reverse mesa shape as shown in the figure when wet etching using bromomethanol was performed on a (001) surface wafer. The groove width at the upper end of the reverse mesa etching groove after this step is 30 μm. Even if the groove width at the upper end of the reverse mesa etching groove is 30 μm, the depth D of the light incident end face is increased to 20 μm to form a good incident end face, and the depth H to the bottom is about 35 μm. When an automatic cleaving flaw forming apparatus having a pyramidal needle is used and the needle 19 having a needle tip opening angle of about 2δ = 60 degrees in a plane perpendicular to the direction in which the cleaving flaw is formed is lowered from above to the vicinity of the groove center The pyramid part was in contact with the upper side of the reversed mesa, and this part served as a guide, and a cleavage flaw was formed in the center with good reproducibility.
[0026]
Here, the needle 19 is symmetrical, but if the groove width is set so that the upper ends of both mesas serve as guides even if the left and right are asymmetrical, the cleavage flaw position is intended from the center of the groove. Can be formed with good reproducibility. Cleaved scratches were formed on a part of the wafer edge to about 200 μm. Of course, the strength of the semiconductor is weaker than that of the needle, so the reverse mesa part where the needle comes into contact is finely damaged, but the mesa part still acts as a guide, and the crevice scratches are reproducible in the central part of the groove. I was able to form.
[0027]
Therefore, by providing a mechanical shock here, it was possible to perform good cleavage over several centimeters starting from the cleavage flaw, and both sides of this cleavage could be used as the light incident end face of the light receiving element. The reverse mesa portion may be formed using another wet etching solution or a dry etching method, or may be formed using another crystal plane or controlling the angle using the adhesion of the etching mask. . When a non-reflective film was formed on the incident end face, and light having a wavelength of 1.55 μm was introduced using a single mode fiber (SMF), a large value of light receiving sensitivity of 0.9 A / W or more was obtained with an applied reverse bias of 3 V. . At this time, the protrusion of the substrate portion immediately below the light incident end face portion is 10 μm or more, but the misalignment tolerance in the optical axis direction at which the light receiving sensitivity to the SMF of the present light receiving element is reduced by 1 dB is 80 μm, which is practically a problem. Don't be.
[0028]
In this embodiment, the groove of the etching mask has a linear stripe shape as shown in FIG. 7A, but the uneven shape or tapered uneven shape as shown in FIGS. 7B and 7C. In addition to the shape and the like, various shapes may be used as long as the width of the groove where the cleavage flaw is formed matches the purpose of the present invention. In this embodiment, the p-InP layer on the surface side is formed by crystal growth. However, in the crystal growth, an undoped InP layer is used, and the main conductivity type of the semiconductor on the surface side is changed to Zn diffusion or ion implantation. And subsequent annealing.
The semiconductor light-receiving element portion is on a semiconductor layer having the first conductivity type, and includes a light-receiving layer and a Schottky formed of an intrinsic or first conductivity type semiconductor layer, a superlattice semiconductor layer, or a multiple quantum well semiconductor layer. A multilayer structure in which a semiconductor layer having a high Schottky barrier height having a Schottky barrier higher than the Schottky barrier between the light receiving layer and the Schottky electrode is interposed between the electrode and an electrode. the or a semiconductor light receiving element formed by configured on a substrate, the Schottky barrier height with high semiconductor _{layer, in 1-x-y Ga} x Al y As (0 ≦ x ≦ 1,0 ≦ y ≦ 1) or in _{1-x-y Ga x Al} y as (0 ≦ x ≦ 1,0 ≦ y ≦ 1) thin on the Part _{In 1-u Ga u as 1} -v P v (0 ≦ u ≦ 1,0 ≦ v ≦ 1) It may be constituted by a semiconductor light receiving device that.
Further, in this embodiment, a semi-insulating InP is used as a substrate and an n-InP layer is used on the substrate side. However, even if a p-InP layer is used, the above-described p and n are reversed in the same manner. It can be manufactured, and can also be manufactured using an n-InP or p-InP substrate. Here, a bulk of a uniform composition is used as the light receiving layer. However, a separate-absorption-graded-multiplication (SAGM) structure, a separate abstraction and multiplexing superstructure (SAM-SL) structure, and the like used in an avalanche photodiode are used. Needless to say, a semiconductor layer having a superlattice structure may be used. Further, it goes without saying that a material system such as InGaAlAs / InGaAsP or AlGaAs / GaAs system other than the InGaAsP / InP system or a material system with inherent strain may be used.
[0029]
【The invention's effect】
As described above, in the method of manufacturing a semiconductor light receiving element according to the present invention, the groove width of the upper end of the reverse mesa etching groove after the step of forming the reverse mesa etching groove for forming the light incident end face is determined. By setting it narrow, when the needle for cleaving the groove portion is lowered from above after the end surface formation etching, the cone or pyramid of the needle contacts the mesa portion of the incident end surface, and this portion Works as a guide, and can cleave flaws with high accuracy in the center. Therefore, by applying a mechanical shock here, it is possible to cleave with good reproducibility starting from this cleavage flaw portion. As described above, this manufacturing method can accurately form a cleavage flaw at the center portion of the groove, can be cleaved with good reproducibility starting from this cleavage flaw portion, and the protruding portion of the substrate from the light incident end face portion is short. it can. For this reason, it becomes possible to cut out the element easily and with good reproducibility while sufficiently satisfying the distance to the light incident end face required for optical coupling with a fiber or the like. In addition, the device can be cut out using both sides of the etched groove as the light incident end faces, and the area of the groove portion that is wasted on the wafer can be reduced, so that the yield of the device from the wafer is greatly improved.
[Brief description of the drawings]
FIG. 1 schematically shows a shape after end face formation etching according to the present invention.
FIG. 2 is a diagram illustrating a refractive semiconductor light receiving element.
FIG. 3 is a schematic view of a wafer structure in which a refractive semiconductor light receiving element portion is arranged in parallel with a groove portion of an inverted mesa.
FIG. 4 schematically shows an inverted mesa shape in which InP on a (001) surface is etched by bromo etching and a (111) surface is exposed.
FIG. 5 is a diagram for explaining a first embodiment of the present invention.
FIG. 6 is a diagram for explaining a second embodiment of the present invention.
FIG. 7 is an example of a groove shape of an etching mask.
[Explanation of symbols]
11 light incident end face 12 1 μm thick p-InP layer 13 1 μm thick InGaAs light receiving layer 14 1 μm thick n-InP layer 15 semi-insulating InP substrate 16 p electrode 17 n electrode 18 cleavage end face 21 etching mask 22 etching mask 23 etching mask

Claims (1)

Receiving light at the light incident end surface by providing a light receiving portion having a semiconductor multilayer structure including a light receiving layer formed on the surface of the substrate and a light incident end surface inclined inward as the distance from the surface side increases. In the manufacture of a refractive semiconductor light receiving element in which incident light passes through the light receiving layer obliquely with respect to the layer thickness direction,
After the reverse mesa etching groove forming step for forming the light incident end face, the depth from the upper side of the light incident end face to the bottom face of the etching groove is H, and the depth of the cleavage flaw is Hc. 2. A method of manufacturing a semiconductor light receiving element, wherein a groove width 2W at an upper end of a reverse mesa etching groove is 2W ≦ 2 (H + Hc) tan δ, when an opening angle of a tip of a needle is 2δ.

Images