JPH10274728A - Photodetector module and manufacturing method thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a digital photodetector module for an analog receiver of an optical CATV with many channels and a digital photodetector module for a large optical input by reducing signal distortion. SOLUTION: A diagonal exit angle from a fiber is expressed by αand a distance between the center of a lens and a chip is expressed by L, and the center of the chip is shifted from the center of the lens by Ltan α in the direction of the bottom point S of the fiber end face. The center of the fiber is shifted in the opposite direction. The center O of the photodetector element, the center H of the lens, and the center Q of the fiber are moved in the direction perpendicular to the axis and are positioned so that diagonal light coming out of a diagonal cut fiber goes to the center of the chip passing through the center of the lens.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a photodiode module (PD module) in which a light receiving element used for optical communication and optical measurement is combined with an optical fiber, and a method of manufacturing the photodiode module. In particular, the present invention relates to a light receiving element module capable of receiving a large number of analog signals without distortion due to small distortion due to interference between signals and a method of manufacturing the same.

[0002]

2. Description of the Related Art Here, a light receiving element module is a combination of a light receiving element and an optical fiber. In the combination of fiber, lens and PD chip,
There have been several proposals to shift the position of the tip forward or backward along the axis from the position of the image of the fiber end by the lens. JP-A-64-79629 and JP-A-5-224101. Its purpose is to lower the level of reflected light or lower distortion to increase efficiency. The light receiving element is one in which a PD chip is housed in a package. The PD chip is a semiconductor chip having a pn junction, has electrodes on both sides, and the side into which light enters is an annular electrode. A conventional light receiving element chip will be described with reference to FIG.
In the long wavelength band of 1.3 μm or 1.55 μm, InG is mainly used.
A pin-PD using aAs as a light receiving layer is used. On an n-type InP substrate 81, an n-InP buffer layer 82, In
Epitaxial growth is performed in the order of the GaAs light receiving layer 83 and the InP window layer 84 to form an epitaxial wafer.

Further, the p-type region 8 is selectively diffused by Zn.
5 to form a pn junction. Passivation film 8
8, antireflection film 87, p-side electrode 86, n-side electrode 90
To form The p-side electrode 86 has a ring shape, and light 89 enters therein. n electrode is n-type InP
The shape is uniform on the bottom surface of the substrate and is not ring-shaped. Such a PD chip is mounted on a metal package having a glass window to serve as a light receiving element. Alternatively, it is integrated with an optical fiber or an optical connector so that the light intensity transmitted through the optical fiber can be detected. The one integrated with the optical fiber and the optical connector is called a light receiving element module (PD module). Since light emitted from a thin optical fiber is detected, alignment of the optical fiber and the chip becomes an important problem.

A conventional example of a light receiving element module combining an optical fiber and a light receiving element chip will be described with reference to FIG. The photodiode 1 is fixed on a package 12 via a submount 13. The light receiving part of the photodiode 1 is set at the center of the package, and the center of the optical fiber 14 is also on the center line of the package. That is, the alignment is performed so that all the optical members are aligned on the center line. The package has anode pin 15, cathode pin 16,
It has a case pin 17 and the like.

[0005] The package 12 is further fixed with a cylindrical cap 22 having a spherical lens 23. Package 12
A cylindrical sleeve 18 is attached to the outer periphery of the header. A cylindrical ferrule holder 19 is fixed to the upper surface of the sleeve 18. This holds the ferrule 20 into which the tip of the fiber 14 is inserted on the axis. The bend limiter 21 is made of a flexible member and protects the optical fiber 14 from being excessively bent at the base of the ferrule.

Since the n-electrode (cathode) of the PD (photodiode) chip 1 needs to be electrically floated from the case 12, a submount 13 of an insulator is interposed therebetween.
It is a rectangular plate made of AlN or alumina (Al ₂ O ₃ ). The front and back surfaces are covered with an appropriate metal film (called metallization). The PD chip 1 is fixed on the submount 13 by solder such as PbSn. The upper surface of the submount 13 is connected to the cathode pin 16 by a gold wire. The p electrode of the chip 1 is connected to the anode pin 15 by a gold wire.

[0007] The spherical lens 23 condenses the light emitted from the optical fiber so as to be incident on the light receiving surface of the photodiode (PD) 1 with high efficiency. The end face of the optical fiber is obliquely polished so that light reflected on the end face of the optical fiber does not return to the laser. This is cut so that it has an 8 ゜ slope.
The angle of inclination varies. The centers of the optical fiber, lens and PD are on the same straight line.

[0008]

Although it seems natural that the three members are on the same axis as described above, there is a drawback that mutual interference easily occurs between a large number of signals having different frequencies. Do you get it. Since this is not always easy to understand, it will be described in detail. Since the difficulty of the conventional example is evident in optical CATV, the operation in that case will be described. When a light receiving element module is used in a light receiving section of an optical CATV receiver, a signal is distorted if there is mutual interference between frequencies. Therefore, the number of channels is significantly limited. In an optical CATV, an analog electric signal is converted into an optical analog signal by a laser or a light emitting diode, and the converted signal is transmitted through an optical fiber. A plurality of signals having different frequencies propagate through one optical fiber and are photoelectrically converted by one light receiving element.

A unit having one frequency is called a channel. Through the signals from a number of channels, the receiver selects any one of them and receives and reproduces it. Since an analog signal is transmitted, it must be photoelectrically converted by a receiver having good linearity. Good linearity means that the energy of light is proportional to the converted electrical energy. If there is a non-linear part, such as proportional to the square of the light energy, two signals of different frequencies will interfere. That is, a signal component of a difference frequency (beat) or a sum frequency newly occurs and is included in the output of the light receiving element. This is called distortion.

[0010] Distortion is a non-linear phenomenon that occurs in any order of two or more. Among them, the second-order distortion is the largest and easy to measure. So second-order intermodulation distortion IM
Evaluating the light receiving element by D _2. Since it is second-order distortion, it is proportional to the strength of the input signal. If the input signal is weakened, the second-order distortion is reduced, but if the signal strength is low, the signal itself is weakened, so that the signal is buried in noise and a beautiful image cannot be reproduced. In a conventional light receiving element module, it has been difficult to reduce distortion without sacrificing signal strength.

In order to improve this, the present inventor has previously proposed a light receiving element module capable of reducing distortion by a condensing method that positively utilizes the aberration of a spherical lens. Japanese Patent Application No. 6-171873, "Analog PD Module and Manufacturing Method Thereof". FIG. 3 shows a conceptual diagram of the light receiving element module. The photodiode 1, the ball lens 3, and the obliquely polished optical fiber 5 are arranged on the same straight line. The number of members and the structure are not different from the conventional one. However, for the first time, the inventors have noticed that the distortion can be reduced without lowering the sensitivity by changing the distance Z between the lens and the end face of the optical fiber. We noticed that the distance dependence of sensitivity and distortion was different, and for the first time proposed an arrangement to reduce distortion without sacrificing sensitivity based on the difference in distance dependence.

FIG. 4 shows the sensitivity (AC sensitivity RAC (A / W)) of the light-receiving element module measured by the present inventor and IMD ₂
It is a graph of (dBc) dependence of the distance Z between a lens and a fiber end surface. The distortion is indicated by a broken line and the sensitivity is indicated by a solid line.
The distortion is maximized at Z = 1.2 mm. Conventionally, attention has been paid only to a region where Z is larger than this. Second order distortion is -7
It was required to be less than 5 dBc. Therefore, conventionally, when the optical fiber is moved far from the lens, the distortion becomes −75 d.
Z was set to be about 1.6 mm, which is Bc. However, this has lowered the IMD ₂ at the expense of the signal strength is outside the maximum sensitivity region. Since the signal is weak, it is still difficult to transmit and receive a large number of channels.

The inventor has repeatedly studied the possibility of lowering only the distortion without deteriorating the sensitivity. Then, he noticed that there was a portion where the distortion curve decreased faster than the sensitivity in a small range of Z. The point at which Z becomes -75 dBc in a small range is Z =
Located between 0.8 mm and 0.9 mm. Z = 1.
Unlike 6 mm, this is the maximum sensitivity. Therefore, the distance between the optical fiber and the lens is set to Z = 0.8 to 0.9 mm, and a light receiving element module that satisfies both requirements of sensitivity and distortion is provided.

The reason why the maximum sensitivity region extends from 0.8 to 1.3 mm is as follows. The diameter W of the light receiving surface surrounded by the p-electrode of the PD chip is considerably wide in the range of 100 μm to 200 μm. When the light emitted from the optical fiber is converged by a lens, the diameter becomes much smaller. Even if the distance between the PD and the lens does not change, the diameter of the light at the PD changes when the position of the optical fiber is moved in the Z-axis direction. Z giving the smallest diameter
When the fiber is fixed at the point, maximum sensitivity is obtained. However, not only that, even if the optical fiber is moved before and after this, as long as the light diameter U is smaller than the light receiving surface diameter (U ≦ W), all the light is absorbed by the PD, so that the maximum sensitivity is given. For this reason, the maximum sensitivity area is expanded to 0.8 to 1.3 mm.

Then, why does the distance dependence of the distortion and the sensitivity not match? The secondary distortion is the center Z _{0 of the} point of maximum sensitivity.
(1.1 mm) before the maximum (1.2 m
m). The above invention has noticed this asymmetry and has exploited it to subtly satisfy both sensitivity and distortion requirements. So why asymmetric? Assuming that the lens has no aberration, the spread of the light ring is symmetrical back and forth with respect to the imaging point. Then, the distortion should be symmetric even if the light of the optical fiber is shifted back and forth from the point where the light just forms.

However, a lens always has an aberration, and a spherical lens in particular has a large aberration. The far-axis light converges quickly and the paraxial light converges slowly. The position where the paraxial light crosses the axis is called a Gaussian image plane. Far-axis light traverses the axis earlier. The distance Z between the optical fiber and the lens and the distance Y from the lens to the image have a complementary relationship represented by the lens formula. When Z is large, Y is small. The inventor considers that the maximum distortion occurs when the Gaussian image plane coincides with the light receiving plane. At this time, light is most remarkably concentrated on one point. Since light is concentrated only on a part of the pn junction, 2
Next, a third-order effect appears, which causes mutual interference of different frequencies and amplifies distortion.

If the optical fiber is moved slightly outside, the light flux only diffuses, and the sensitivity starts to decrease. Conversely, if the optical fiber is moved toward the lens after the Gaussian image plane matches the light receiving surface, the convergence point of the far-axis light will match the light receiving surface, and the sensitivity will not decrease. Further, the light beam diameter U exceeds the light receiving surface diameter W only after the optical fiber is moved in parallel to the lens side. For that reason, sensitivity and distortion may be asymmetric. In this way, the present inventor has discovered for the first time an arrangement of an optical fiber lens capable of reducing the distortion to -75 dBc or less and maintaining the sensitivity. Again, the three members, lens, PD and optical fiber, are on the axis. That is, all three members are on the Z axis. An image of the light emitted from the optical fiber cannot be formed on the light receiving surface and is formed behind the light receiving surface. Since the light receiving element is placed at a position out of focus, this may be simply called a defocus method.

In an optical CATV, several channels (C
H) was sufficient, but 40CH is now standard. Also recently more 80CH ~ 110C
There is a demand for a system capable of transmitting H.
The demands on the number of channels are becoming increasingly demanding.
With the increase in the number of channels, the receiving frequency band of the receiver is 450 MHz, but 860 MHz is required. Bandwidth should be almost doubled. Since signals must be sent to more and more homes, laser (LD) light sources that emit a large amount of light have been used. Since a high-output laser is used as a light source, a large signal light of 1 mW or more may be incident on the PD. The greater the number of channels, the greater the effect of distortion, and the stronger the signal light, the greater the distortion. The need to improve distortion characteristics more than ever has been felt strongly by increasing the number of channels and increasing the output. It is necessary not only to improve the characteristics but also to provide a light receiving element module at a low price. Therefore, it is urgently necessary to manufacture a light receiving element module which is stable and has low distortion and high sensitivity.

It is a first object of the present invention to provide a light receiving element module with less distortion even for a large signal.
It is a second object of the present invention to provide a light-receiving element module with low distortion without sacrificing sensitivity. A third object is to provide a method for manufacturing a light-receiving element module with small distortion at low cost. It is a fourth object of the present invention to provide a method for manufacturing a low distortion light receiving element module with good yield.

[0020]

When an axial ray propagating along the axis of the optical fiber exits the optical fiber end Q at an angle of α with the axis due to the angle of inclination θ of the tip of the optical fiber, an extension of the inclined ray α The lens center H and the light receiving element center O are arranged side by side in the light receiving element module of the present invention. The light receiving element is located closer to the lens than the Gaussian image plane at the end of the optical fiber by the lens. At least the oblique rays from the fiber and the HO are parallel. Since QH is determined by the alignment, it is not necessarily parallel to the inclined ray α, but ideally α and QH should also be parallel. If the center H of the light receiving element and the center H of the lens are set on the axis, the distance between the lens center H and the optical fiber is Lf, and the distance between the lens center and the center of the light receiving element is L. −L tanα). The center of the optical fiber is shifted from the center of the package in the opposite direction (+ X direction) by a value close to Lftanα. Further, the rotation angle of the optical fiber is determined so that the lowermost point S of the polished surface of the optical fiber is closest to the Z axis.

The above is the principle of the present invention. Actually, L and Lf are asymmetric. L is a constant determined in advance from the height of the cap and the like. Lf is determined for each element by a centering operation to minimize distortion and maximize sensitivity. When L and θ are known, distortion can be minimized by determining the position of the center of the light receiving element with respect to the lens center in the direction perpendicular to the axis as described above. In the present invention, both the optical fiber and the light receiving element are deviated from the axis. In contrast to the inventor's improvement described earlier, which is called defocusing, this is tentatively called off-axis. Therefore, the present invention is a defocus + off-axis having both of them. A particularly new claim is off-axis. That is, the center H of the lens and the center O of the light receiving element are arranged on the light beam coming out of the fiber end. Defocus has been described in the earlier application. So why do we shift the optical fiber and light-receiving element sideways (off-axis) to improve the distortion? Let me explain this.

The present inventor has manufactured a large number of defocused analog PD modules in which the optical fiber is brought closer to the lens than the focal position according to the above-mentioned prior application method. I noticed something strange during the manufacturing process. Perhaps due to variations in manufacturing conditions, a module with better distortion characteristics and a module with worse distortion characteristics than the standard can be produced. What comes from the variation in distortion characteristics? FIG. 4 shows the distance Z between the lens and the fiber.
The distortion is specified by, but are there any other parameters that affect the distortion?

Therefore, a module having a large distortion or a module having a small distortion was disassembled and investigated. Then I found an unexpected fact. As shown in FIG. 5, a module having a good distortion characteristic has an axis (Z axis) where the optical fiber center Q passes through the lens center H.
The center O of the light receiving element chip is also shifted from the axis to the opposite side. Moreover, the lowest point S of the optical fiber is closest to the axis. The diameter of the obliquely polished surface of the optical fiber in the oblong direction is ST, S is the lowest point, and T is the highest point. Points T, Q,
S, H, and O exist on the same plane including the Z axis, and points Q, H, and O
Are almost collinear. Since the alignment is performed by defocusing, as shown in FIG.
Should be on the shaft, but there may be deviations from the shaft due to manufacturing errors. Even in that case, there is a case where the distortion is small and the sensitivity is good, which is the case of FIG. Light rays along the central axis of the optical fiber emerge from point Q and refract to the side of S, and the Z axis and α
Advance through the lens center H and pass through the center H of the element 1
to go into. Ray QHO is substantially straight.

FIG. 9 shows the relationship between the oblique cut angle θ and the inclination angle α with respect to the axis of the light beam. The light ray RQ traveling along the fiber axis is refracted toward the side S on the inclined surface TS, and draws a locus RQM. The deviation from the right angle of the surface TS is θ.
The normal established on the surface TS is defined as Qn. According to Snell's law, n ₀ sin (α + θ) = n ₁ sin θ (1). n ₀ is the air refractive index, and n ₁ is the fiber core refractive index. From this, α becomes α = sin ⁻¹ ｛(n ₁ sin θ) / n ₀ ｝ −θ (2) When θ is determined, α is determined.

FIG. 6 shows an arrangement where the distortion is large. The center Q of the optical fiber is shifted from the Z axis, and the element center O is also shifted from the Z axis in the opposite direction. The lens center H is on the Z axis. The uppermost point T of the fiber cut surface is close to the Z axis.
The direction of the cut surface of the optical fiber is opposite to that of FIG. The optical fiber center Q, the lens center H, and the chip center O are substantially on the same straight line. However, since the light following the center line of the optical fiber is refracted to the S side of the oblique cut surface, the light travels at an angle of α outward from the Z axis and enters the outer shell of the lens. It is strongly refracted by the lens and enters the center O of the PD. Although the image of Q is the same at the center O of the light receiving element, the optical path of the beam is greatly different since the direction of the optical fiber is opposite. According to geometrical optics considerations, the light quantity is the same in any case, and it is unlikely that distortion will be affected.

The present inventor considered why such an asymmetric phenomenon occurs. If the distortion characteristic is not symmetric about the optical axis, it must be due to the components containing the asymmetric elements. Both the spherical lens and the light receiving element chip have symmetry. So the cause must be the optical fiber. The end of the optical fiber is polished at an angle of about 4 ° to 10 ° so that reflected return light does not return to the light source.
The direction of the polished surface seems to affect the distortion characteristics. If the direction of the polished surface is different, the trajectory of the light beam is greatly different as shown in comparison with FIGS. Geometrically, the image of Q can be at point O regardless of the orientation of the optical fiber. Light emitted from the optical fiber end Q passes through the lens and forms an image at the center O of the light receiving element chip. However, actually, there is not one ray. Light emerges from point Q in a conical shape. The vertex angle γ of the light cone depends on the core refractive index and the cladding refractive index of the fiber.

If the lens has no aberration, all the light emitted along the cone should converge on the center O of the chip. However, lenses actually have aberrations. Especially in the case of a spherical lens, the aberration is remarkable. Paraxial light converges slowly (Gaussian image plane) and far-axis light converges fast. Therefore, the convergence point on the axis of the light beam differs depending on where the light passes through the lens. 5 and 6
Is different. In the case of FIG. 5, the light beam enters the chip at a right angle. Therefore, the ray distribution becomes almost symmetric with respect to the point O. Light rays included in the cone enter the chip surface with a substantially uniform distribution.

However, in the case of FIG. 6, a light beam hits the chip from outside the lens. Therefore, the light enters the surface more obliquely. Since the light is incident obliquely, the light beam density on the light receiving surface greatly differs before and after the point O. Density fluctuations have little effect on sensitivity but strongly affect distortion. The light density greatly fluctuates due to the inclined incident line. In some places the light density is high and in others it is low. Significant non-uniformity. For that reason, in the arrangement shown in FIG. 6, mutual interference increases and distortion increases.

The inventor has made such a hypothesis. The simulation by the ray tracing method was performed to confirm the hypothesis. In other words, it is assumed that a large number of light beams are emitted from the end Q of the optical fiber into the inside of the cone at the same density as the solid angle, and each of the light beams is traced geometrically. Each light ray is refracted by a lens, and the point at which this is incident on the chip is determined. It is called ray tracing because it traces rays one by one. Refraction at the lens can be easily calculated because it can be handled by geometrical optics. FIGS. 7 and 8 show the incident point on the chip by dots. These are calculated results, not experimental results. It is not actually observed that the light spot enters the element in this way.

Let X be the deviation of the chip center O from the Z axis,
Let W be the deviation of the optical fiber center Q from the Z axis. X and W
Is determined as follows. As shown in FIG. 6, when the lowermost point S of the fiber end face ST is farther from the Z axis, W is negative and X
Is positive. Since the three points Q, H, and O are on a straight line, W has the opposite sign to X. In FIG. 5, W is positive and X is negative because S is close to the Z axis. In FIG. 6, the X axis is defined to the left, and in FIG. 5, the X axis is defined to the right. The optical system will be described below with reference to FIGS. 5 to 8 by giving specific numerical values of the present invention. Here, all lenses are common and have a diameter of 1.5
mm, a spherical lens with a refractive index of 1.5. At this time, the focal length of the lens with respect to the paraxial ray is 1.12 mm. Therefore, in the following example, Lf = 1850 μm, 1650 μm
The Gaussian image plane at the time of is 3000 μm from each lens,
It comes to the position of 3500 μm. Assuming that the distance of the light receiving surface of the light receiving element from the lens is L = 2000 to 2100 μm, the light receiving element is at a position closer to the lens side by 1000 μm to 1400 μm from the Gaussian image plane. That is, the light receiving element is at the defocus position. According to the present invention, the light receiving element and the light emitting element are defocused and shifted in the direction perpendicular to the axis (off-axis). Therefore, the present invention is defocused and off-axis.

In the arrangement shown in FIG. 5, the fiber shift W = 120 μm
m, chip deviation X = −140 μm, distance Lf in the z direction between the lens center H and the fiber end Q = 1850 μm, distance L between the lens center and the chip L = 2100 μm, α = 3.7 °, θ =
8 ゜. FIG. 7 shows a distribution of incidence points of paraxial rays in the arrangement of FIG. The horizontal axis in FIG. 7 is the X axis and the vertical axis is the Y axis. Both are orthogonal to the Z axis. Circle C is the spread of the light receiving surface of the light receiving element. A point indicates an element incident point of one ray. Points indicating the light incident points are uniformly distributed on the light receiving surface. Although the horizontal axis in FIG. 7 is the X axis, the distribution of points is almost the same regardless of whether it is + X or -X. The same can be said for the + Y axis direction and the −Y axis direction. That is, light spots are distributed isotropically on the xy plane. In particular, there is no high-density area. Near uniform distribution.

This must not be taken for granted, since the light beam enters the element in a straightforward manner as shown in FIG. Quite the opposite. If there is no aberration in the lens and the light-receiving element is located at the image forming point of the fiber end point Q, rays emitted in any direction should converge to the point O. Every trace of thousands of rays should concentrate at point O. When the light beam is concentrated at one point, the interference is the largest and the distortion is large. Therefore, in a lens having no aberration, the distortion becomes maximum at the maximum sensitivity, and the lens becomes symmetrical in the front-back direction with respect to z. It is difficult for an aberration-free lens to increase sensitivity and reduce distortion.

As a result, the asymmetry of sensitivity and distortion in FIG. 4 is caused by large aberration. The Gaussian image plane of Q is formed behind the light receiving surface of the element because the lens has an aberration and the fiber is placed before the image forming point. The imaging point does not match the light receiving surface. For this reason, light rays emitted from the point Q are scattered as much as possible on the element surface. When light spots are uniformly distributed in this manner, interference is reduced and distortion is reduced. Moreover, since most of the light spots are inside the light receiving circle C, the sensitivity is good. Since almost all light is dispersed and enters the light receiving surface, sensitivity is good and distortion is small.

This trick is possible because the element is located out of focus using an aberrated lens. It is not the case that the imaging point should be moved behind the element surface using an aberrated lens. It's not that simple. Element mounting position,
This is because the position of the fiber may be off the axis, in which case the strain is increased and sometimes the strain is reduced.

FIG. 8 shows an incident point distribution created by far-axis light as shown in FIG. The fiber shifts away from the chip. The plane of displacement is defined as an xz plane. That is, it is assumed that both the fiber and the chip are shifted in the x direction including the sign. The parameters in FIG. 6 are fiber deviation W = −160 μm, chip deviation X = + 140 μ.
m, Lf = 1650 μm, L = 2100 μm, α = 3.
This is an example of 7 ° and θ = 8 °. Almost all light rays enter the light receiving surface circle C. However, the density is uneven. The crescent-shaped area GJEF near x = −c (J) is blank and no light beam reaches it. However, EFGK just inside the curve EFG
ON has high-density light spots. This high density light spot increases the distortion. The high-density distribution is deviated to the left of the center O, and the right side of the point O is a low-density area.

Light spots are not uniformly distributed, and there are dense portions,
This causes interference between different frequencies and increases distortion. Although almost all of the light spots are included in the light receiving circle C, the sensitivity is high, but the light spot distribution is not uniform, and there is a dense portion, which increases the distortion. The sensitivity is high as long as the light spot is included in the light receiving circle C. The distortion increases when there are congested parts. What can be done in overcrowded areas? That's not easy to understand.

As shown in FIG. 3, when neither the optical fiber nor the chip deviates from the Z axis, the light spot variation is intermediate between FIGS. Therefore, in the case of FIG. 7, the distortion is smaller than in the case of FIG. 3 on the axis, and in the case of FIG. 8, the distortion is larger than in the case of FIG. In the case of FIG. 7, since the light spot is diffused over the entire light receiving surface C by using the aberration of the spherical lens, the light power per unit area is reduced and the distortion is reduced.

This is a comparison between three typical cases as shown in FIGS. 5, 6 and 3, and the relationship between the axis shift and the distortion was examined in more detail. The parameter is the deviation X of the chip center O from the Z axis. A constraint condition is set such that the three points Q, H, and O are substantially on a straight line, and the sensitivity is 1.3 μm for all.
Lf and L are adjusted so as to be 0.9 A / W with respect to light. In each case, the Gaussian image plane is not on the chip but farther from the chip plane. This is the same as the concept (defocus) of the inventor's earlier application. Therefore, not only X is changed, but L, Lf, and W are also changed. The optical fiber is a single mode fiber used for 1.3 μm to 1.55 μm, and the polishing angle θ is 8 °, n ₁ =
Assuming 1.5, the emission angle α is 3.7 °.

FIG. 10 shows the result of measuring the distortion IMD ₂ while changing the chip deviation X. However, when the lowermost point S of the end face of the optical fiber is separated from the Z axis and the uppermost point T is closer to the Z axis, X is defined as positive. X in FIG. 6 is positive, and X in FIG. 5 is negative. That is, the direction of the orthogonal projection of the vector ST onto the xy plane is the + x direction.

X = −140 μm, minimum distortion value −86 dB
c, which corresponds to the arrangement of FIG. 5 (W = 120 μm, Lf =
1850 μm, L = 2100 μm). X is -15
It is -85 dBc or less from 0 μm to −100 μm.

X = 0 μm is a case where there is no axis deviation as shown in FIG. IMD ₂ is -79 dBc. X = -20
In the range of 0 μm to +50 μm, the distortion is −75 dBc or less.
If the required value of the distortion is -75 dBc as before, this range is sufficient. When X becomes positive, the distortion increases sharply. When the position of X = + 140 μm is as shown in FIG. 6 (W = −1)
60 μm, Lf = 1650 μm, L = 2100 μm). The distortion is -59 dBc.

[0042]

DESCRIPTION OF THE PREFERRED EMBODIMENTS In order to carry out the present invention, care must be taken when placing a chip on a package. The simplest way is to move the chip from the package center P to -Ltanα.
That is, the center H of the shifted lens is matched with the center P of the package. The next possible method is to match the chip center O with the package center P and set the lens center H to Ltanα.
It is a method of shifting only. Further, the chip is placed at an arbitrary position, and the lens is placed at a position shifted by Ltanα therefrom.

That is, when these members are represented by deviations from the package center line (T in ST is defined as positive), chip center O = -Ltanα, lens center H = 0 chip center O = 0, lens center H = + Ltanα Chip center O = β Lens center H = β + Ltanα Three cases are possible. Since the direction of the fiber is known since the direction ST of the end face is known, the fiber is moved in parallel in the xy plane, and the fiber is aligned while checking the distortion.

The optimum position of the fiber end Q is: Q = Lftanα Q = Lftanα + Ltanα Q = Lftanα + Ltanα + β However, since Lf is determined by adjustment in the z direction, it cannot be determined in advance.

Then, why does the light ray distribution as shown in FIGS. 7 and 8 differ depending on the rotation direction of the fiber?
This is the problem. Since the results calculated by the ray tracing method are shown in FIGS. 7 and 8, it is difficult to understand intuitively. The reason why the non-uniform distribution shown in FIG. FIG. 13 shows a ray trajectory of the light emitted from the fiber end Q passing through the spherical lens H and reaching the chip. The lenses are exaggerated to highlight the differences. The light emerging from the fiber end exits in the direction of the oblique angle α, which has a conical spread. The spread angle γ is given by sin γ = (n ₁ ² −n ₂ ² ) ^1/2 . n
₁ core index, n ₂ is the refractive index of the cladding. Even if it comes out diagonally, it spreads at almost the same cone angle.

The light beam emitted from one point Q is refracted by the spherical lens, but has a different convergence point on the optical axis QbHmFg. Light rays passing outside converge faster. A, b, c, d,
Consider points e, f, and g and light rays passing therethrough.
Light passing through the center passes through QbHm and converges on a Gaussian image plane Fb. The convergence point is the farthest. Light beam Q that bends slightly
cl and Qan converge at Fc. A ray passing through d intersects the optical axis at Fd. A ray passing through e crosses the optical axis at Fe. Light rays that pass through the outer shell g further advance to ghFg and converge at Fg. In other words, light rays passing farther from the center H are refracted more strongly, so that the convergence point approaches the lens. This is the aberration of the spherical lens. As shown in FIG. 5, paraxial light passing through H is incident at ab on the lens surface and exits at mn in FIG. This is F
The optical axis is cut at b to Fc.

The principle discovered by the present inventors (Japanese Patent Application No. 6-17 / 1990)
1873), the distortion is smaller when the chip is placed in front of the convergence point (Gaussian image plane). Assume that the light receiving element chip is placed at a position pq where paraxial light passing through the center H of the lens is cut ahead of the Gaussian image plane as shown in FIG. In that case, the light spread from the light source point Q to the minute surface ab is uniformly distributed, and even mn is emitted with a uniform distribution. This results in a uniform light distribution on the light receiving element chip surface pq. This is the case in FIG. Paraxial light enters the light receiving element.

In contrast, in the case of FIG. 6, the direction of the slope of the fiber is opposite, so that the fiber passes outside the lens. That is, it is far-axis light. The far-axis light is largely refracted and enters the center of the chip. In FIG. 13, light beams Qf to Qg entering the oblique surface fg of the lens correspond to this. It passes from gh to fi and exits from i to h. To receive this, the chip is placed at a position sr for receiving the far-axis light on the lens side from the Gaussian image plane. This corresponds to the arrangement in FIG. Then, the light beams is and hs intersect at the chip surface s. Then the lens surface f
G through the lenses fi to gh and the angle h
The light passes through the inside of si and converges intensively at one point s on the chip surface. This corresponds to the dense light spot in the EFGKON portion of FIG.

In the case of such a spherical lens, there is always a portion where the light energy becomes overcrowded because the light beam passing on the outside and the light beam passing on the inside intersect in front of the Gaussian image plane. The far-axis light passing outside crosses fast. Where they cross, the energy becomes overcrowded. This causes dense light energy as shown in FIG. 8 and increases the secondary distortion. Since paraxial light does not intersect with the Gaussian image plane, the energy does not become overcrowded, and a uniform light energy distribution is realized by placing the chip before the focal point. This is simply the case. In FIG. 13, the point at which the light beam enters the lens is bcdefg.
This corresponds to changing the displacement x from −140 μm in the positive direction in FIG.

The present invention employs defocus and off-axis. The figure shows the relationship more intuitively. FIG. 14 shows the relationship between the light receiving element, the lens, and the fiber of the defocus + off-axis of the present invention. This corresponds to the cases of FIGS. The lower point S of the fiber end is near (inside) the axis. The central ray emerges at an angle α to the axis. The center ray passes through the lens center H. Other paraxial rays also enter the light receiving element PD after passing through the lens. The Gaussian image plane is behind it. Since paraxial rays cut through the PD, the ray distribution is uniform.

FIG. 15 shows a case where the lowermost point S of the fiber end is outside. 6 and 8. FIG. The central ray emerges at an angle α to the axis but does not pass through the lens. A paraxial ray is strongly refracted through the very outside of the lens and enters the PD. The far-axis ray that is far from the axis as viewed from the lens is the central ray from the fiber (at an angle of α)
Close to QM. This is Qxyz. It is the least refracted. Qxyz cuts straight line QH near the Gaussian image plane.
Although several rays have been written, the opposite confinement is the ray Quvw, which refracts at a critical angle near the periphery of the lens. This is the strongest refraction. Let F be the point at which this critical ray Quvw enters the PD chip. There is no light ray incident on the PD on the + X side any more. This gives the boundaries of the EFG of FIG. Beyond the critical refraction angle of the lens, P
Light rays do not enter deep in the D plane. For this reason, in FIG. The blank in FIG. 8 results from the critical refraction of such a lens. Such an intuitive description will make the advantage of the PD of the present invention of defocusing + off-axis and positioning the lens center on the central ray exiting the fiber more apparent.

An intuitive explanation has been given. When a light beam passing through the peripheral portion of the lens is applied to the chip, the light beam overlaps at the outer peripheral portion, so that the light becomes too dense and the distortion becomes excessive. The distortion is small because the paraxial light does not intersect and no dense area occurs. If all the light enters the light receiving surface c, the sensitivity is maximized, but the distortion varies depending on the light distribution. The present invention has been successful in considering distortion in relation to lens aberrations.

[0053]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS By using the assembling method of the present invention, it is possible to obtain a light receiving element module having extremely excellent distortion characteristics without using a special element member method. The optimum can be provided in an analog PD module in which distortion has a serious effect. When the method of the present invention is used not only for an analog receiver but also for a digital receiver having a high light input level (distortion is likely to be large), the light incident density is made uniform and a more accurate signal waveform is reproduced. be able to. In other words, it is possible to provide an optimum receiver with little distortion in both analog and digital. The light receiving element module according to the present invention includes the light receiving element chip itself and the module. Therefore, the chip itself and the module will be described.

[Embodiment of PD Chip] A conventional chip can be used as it is. This is because the present invention is not an invention of the chip itself. FIG. 1 is a sectional view of a light receiving element chip. On the n-InP substrate 81, n-InP
P buffer layer 82, n-InGaAs light receiving layer 83, n-
An InP window layer 84 is epitaxially grown. A passivation film is formed on this epitaxial wafer, and a hole is formed in a film portion corresponding to the central portion of the device.
Select to spread. This forms a pn junction. On the light receiving surface on which light is incident, an antireflection film made of a dielectric such as SiON or SiNx is formed on the surface of the epitaxial layer in order to reduce reflection loss. This is 1.3 μm or 1.
A multilayer film having a combination of a film thickness and a refractive index that does not allow light of 55 μm to pass through and reflect the light is used.

In this example, the light receiving layer is made of InGaAs. However, the light receiving layer can be made of InGaAsP. For example, if the composition is such that the absorption edge wavelength is λg = 1.4 μm, a photodiode having wavelength selectivity of receiving 1.3 μm light and not receiving 1.5 to 1.6 μm light can be realized. Can be. A p-side electrode 86 is formed on the ring or partially on the Zn-doped p region. On the back surface of the n-type InP substrate 81, an n-side electrode 90 is formed.
The p and n electrodes are reverse biased. The incident light 89 enters the light receiving surface from the p-side electrode through the antireflection film 87. This is absorbed by the InGaAs light receiving layer 83 to generate an electron-hole pair. This becomes a current, and a photocurrent proportional to the optical signal is generated.

In the case of an optical CATV receiver, a high-speed response of at least 1 GHz is required. Therefore, the light receiving diameter (2c) of the light receiving element is often set to 70 μm to 100 μm. The light receiving element of FIG. 1 has such a structure.

FIG. 11 shows a light receiving element chip obtained by further devising. This is proposed by the present inventor in Japanese Patent Application No. 2-230206. At the same time as the diffusion of the light receiving region at the center of the device, the Zn diffusion region 9
5 are formed. This portion (Zn diffusion region for diffusion shielding) 95 is completely insulated from the central light receiving portion 85 by being separated by an n layer and having a double pn junction. Light incident on the element peripheral portion 95 generates an electron-hole pair, but disappears soon because holes cannot go to the p-electrode 86 and electrons cannot go to the n-electrode 90. That is, since the p region 95 absorbs excess carriers, there is no delay in response. It is most suitable as an optical CATV receiver when the number of channels is further increased.

[Embodiment of PD Module] FIG. 12 is a sectional view showing a basic structure of a light receiving element module according to the present invention. The present invention provides a fiber end so that light emitted from the obliquely cut fiber center Q passes through the lens center H and reaches the chip center O.
The light receiving element chip is shifted from the axis in the opposite direction. The orientation of the fiber is important. This shows a type very similar to the conventional example shown in FIG. 2, but the module in FIG. 12 is not exactly the same.

A PD chip 31 (light receiving element) is die-bonded on a header 32 via a submount 33. The header (package) 32 has pins 41, 42, 43
Etc. The submount 33 is an insulator, but the front and back surfaces are metallized. The p-electrode of chip 1 is connected to pin 41 by wire 36. The n-electrode is bonded to the upper surface of the submount 33. The upper surface is wire 37
Is connected to the pin 43. The chip center is not at the center of the header. It is shifted by a certain amount from the center in a certain direction. The amount of displacement x is given by the following equation: the distance L between the lens center H and the chip depends on the inclination angles α and L of the light emitted from the optical fiber.
represented by anα. The chip is die-bonded in advance to such a deviation position.

The cap 38 has a window and the ball lens 34 is fixed. The center H of the lens 34 matches the center of the header 32. Of course, there is an error in mounting the cap 38, but the error is ignored at this time, and the discussion proceeds. Header 32
Is welded with a cylindrical sleeve 44. Fiber 3
The ferrule 45 is fixed to the tip of the fifth and the tip is polished obliquely. The oblique angle is θ. This is to prevent the end face reflected light from returning to the laser, but may be any number of times such as 5 ° to 10 °. Here, an 8 ° oblique angle is used.

At this time, α = 3.9 °. The distance L between the lens center H and the chip is determined by the cap, but here L = 2000 μm. Therefore, the chip is fixed while being shifted x = −136 μm from the center of the header (offset mounting). Of course, in this offset mounting, the lens may be shifted. In that case, the chip can be fixed at the center of the header. Therefore, X = L with respect to the lens center.
It is common to shift by tanα. However, considering the convenience of alignment, it is more convenient to shift the chip in one direction with the lens on the axis.

The ferrule 45 is a ferrule holder 46
Plugged in. The ferrule holder 46 is fixed on the end face G of the sleeve 44, but the center of the holder 46 is not on the center line of the header 32. The holder 46 is shifted to the side opposite to the chip shift. Moreover, the lowermost point S of the obliquely polished surface of the fiber exists closer to the axis. The direction of the lowermost point S and the uppermost point T of the feedback end face can be recognized in advance because the ferrule is marked. If the mark 51 is always present on a plane including the chip center O and the lens center H and perpendicular to the header surface, the direction as shown in FIG. 12 can be maintained.

The fiber alignment is performed by axial alignment, vertical axis movement, and rotation around the axis.
Here, it is assumed that the axial alignment has been completed, and only the in-plane movement and rotation will be described. The alignment in a plane is
This is usually done by horizontal translation of the holder on the G plane in the xy plane and rotation at a point. That is, in the conventional method, alignment = translation + rotation. However, in the present invention, rotation is unnecessary or almost unnecessary.

First, the holder is moved on the xy plane to
The maximum amount of light incident on the lens (maximum sensitivity) and the minimum distortion (IM
D ₂ minimum) made a point to look for. The same method as in the related art is required for the parallel movement alignment in the xy plane. Next, the ferrule holder (fiber) is rotated to perform centering to lower the point of the minimum distortion sensitivity. The present invention can significantly reduce the time required for alignment in rotation. In the case of the present invention, the rotation direction of the fiber is almost determined in advance. This is because the lowest point S is set to a position close to the axis. Therefore, the alignment of the fiber in the rotation direction may be performed at most within a range of 90 °.
The centering in the rotation direction may not be performed at all. In that case, it is only necessary to perform horizontal alignment in the xy plane. Therefore, in the case of the present invention, alignment = parallel movement + rotation within a limited range, or alignment = parallel movement. Since there is no alignment work by full rotation, time can be saved.

Further, in alignment, it is ideal to monitor the two parameters of distortion and sensitivity. However, the sensitivity takes a maximum value within a certain range (as shown in FIG. However, it is also possible to monitor only the distortion. Ideally, the light emitted at an angle of α goes straight and passes through the center of the lens and the center of the chip, but since there is an error, the light only passes near the center of the lens. I can't say. Minimum distortion,
Since the position of the fiber is determined based on the maximum light quantity (sensitivity) and there is a measurement error and a component error, the QHO is not always always linear as a result. However, they generally line up in a straight line.

[0066]

According to the method of the present invention, IMD ₂ = −8.
A light-receiving element module of 0 to -85 dBc was manufactured with good reproducibility. Since the distortion is extremely small, it is possible to stably manufacture a high-performance receiving module for an optical CATV having a higher power and a larger number of channels such as 100 channels. It can also be used as a digital receiving module with a large optical input.

Since the PD is mounted off the center of the package from the beginning, a clear asymmetry can be obtained. Since the inclination direction ST of the tip of the fiber is known, if the direction of ST is adjusted to an asymmetric direction, there is no need for alignment in the rotation direction. Alignment can be performed only by translation in the xy plane.
In this case, even without rotation, the deviation from the optimal azimuth is within 90 degrees and can satisfy the current requirement of -75 dBc or less. Since the rotational alignment is a time-consuming operation, it is very effective to omit this operation. The range of rotation can be limited to about 90 degrees without completely omitting it. In any case, the alignment time can be saved.

[Brief description of the drawings]

FIG. 1 is a cross-sectional view of a known light-receiving element chip that can be used in the present invention.

FIG. 2 is a cross-sectional view of a conventional photodiode module.

FIG. 3 is a Japanese Patent Application No. 6-17187 filed by the inventor of the present invention.
FIG. 3 is a diagram for explaining a configuration of a low-distortion light receiving element module proposed in No. 3; The fiber, lens and chip are collinear.

FIG. 4 shows the distance Z (mm) between the lens center and the fiber end, the sensitivity RAC (A / W) and the secondary distortion I when the fiber is moved along the axis in Japanese Patent Application No. 6-171873.
Graph showing the measured change in MD _2.

FIG. 5 is a schematic configuration diagram showing an arrangement of a fiber, a lens, and a PD according to the present invention in which light emitted from an end face of a fiber travels straight near the center of a lens and enters a chip. θ = 8 °,
α = 3.7 °, W = 120 μm, X = −140 μm, L
f = 1850 μm, L = 2100 μm.

FIG. 6 is a schematic configuration diagram showing an arrangement of a fiber, a lens, and a PD according to the present invention in which light emitted from an end surface of a fiber passes through a peripheral portion of a lens and is refracted largely into a chip. θ =
8 °, α = 3.7 °, W = 160 μm, X = + 140 μ
m, Lf = 1650 μm, L = 2100 μm.

FIG. 7 shows that, in the arrangement shown in FIG. 5, a plurality of light rays having the same density at the exit cone from the fiber end come to a limited solid angle, and the light rays are traced one by one geometrically optically to any point on the light receiving element surface. FIG. 4 is a ray incident point distribution diagram calculated by calculating whether or not incident light is incident and adding dots to the incident points.

FIG. 8 shows that, in the arrangement shown in FIG. 6, a plurality of light rays having the same density at the exit cone from the fiber end come out at a limited solid angle, and the light rays are traced one by one geometrically optically to any point on the light receiving element surface. FIG. 4 is a ray incident point distribution diagram calculated by calculating whether or not incident light is incident and adding dots to the incident points.

FIG. 9 is a diagram for explaining the relationship between the oblique polishing angles θ and α when light emitted from the oblique polishing fiber travels in a direction forming α with the axis.

FIG. 10 shows the secondary distortion when the fiber is displaced in the direction perpendicular to the axis in the direction of the end face ST of the fiber and the chip is displaced in the opposite direction so that the sensitivity is kept constant (0.9 A / W). Is a graph showing how the measured value changes with the displacement X.

FIG. 11 is a cross-sectional view of a light-receiving element chip proposed in Japanese Patent Application No. 2-230206 of the present invention in which a p-region is formed also by doping Zn around the chip to form a p-region and a response delay due to leaked light is eliminated. .

FIG. 12 is a sectional view of a light receiving element module according to an embodiment of the present invention.

FIG. 13 shows a case where light emitted from one point is refracted by a spherical lens and converges on the optical axis. However, the convergence point differs depending on where the lens passes, and since the far-axis rays intersect in front of the Gaussian image plane, the power is low. FIG. 4 is a diagram for explaining that an overcrowded region occurs.

FIG. 14 is a diagram showing a relationship between a light receiving element, a lens, and a fiber of defocus + off-axis according to the present invention. Since the S point is inside, it can be seen that the light distribution is uniform.

FIG. 15 is a diagram showing a relationship between a light receiving element, a lens, and a fiber, which is defocus + off-axis but is denied by the present invention. It can be seen that the light distribution is non-uniform because the point S is outside.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 photodiode chip 2 package 3 ball lens 4 oblique cut ferrule 12 package 13 submount 14 single mode fiber 15 anode pin 16 cathode pin 17 package pin 18 sleeve 19 ferrule holder 20 oblique cut ferrule 21 bend limiter 22 cap 23 ball lens 31 light receiving Element chip 32 Package 33 Submount 34 Lens 35 Optical fiber 36 Wire 37 Wire 38 Cap 40 Polished end face of optical fiber 41 Anode pin 42 Case pin 43 Cathode pin 44 Sleeve 45 Ferrule 46 Ferrule holder 51 Mark 52 Lowermost point of ferrule inclined surface 54 Tip of ferrule 81 n-InP substrate 82 n-InP buffer layer 83 n-InGaAs light receiving Layer 84 n-InP window layer 85 Zn diffusion region 86 p electrode 87 antireflection film 88 passivation film 89 incident light 90 n electrode 95 diffusion Zn diffusion region for diffusion shielding

Claims (13)

[Claims] 1. An outgoing angle α determined with respect to an inclination angle θ of the tip of an optical fiber is calculated, and a shift amount X = Ltan α is obtained, where L is a distance between a lens center H and a photodiode.
The PD chip is fixed to the package at a point shifted by X in the fixed direction from the package center P, the lens is mounted on the package so that the package center and the lens center H coincide, and the optical fiber is the lowest point of the inclined end surface ST. S is held in a direction such that S coincides with the direction of the displacement of the PD, and is translated in the xy plane orthogonal to the axis and along the Z direction parallel to the axis to provide the desired sensitivity and the lowest distortion rate. And fixing an optical fiber at that point. 2. An outgoing angle α determined with respect to the inclination angle θ of the tip of the optical fiber is calculated, and a shift amount X = Ltan α is obtained, where L is the distance between the lens center H and the photodiode.
The PD chip is fixed to the package at a point shifted by β (including β = 0) in a certain direction from the package center P, and the lens is β + X in the same direction from the package center P.
The optical fiber is held in a direction such that the lowermost point S of the inclined end surface ST coincides with the direction of the displacement of the PD with respect to the lens, and the optical fiber X is perpendicular to the axis.
Manufacturing a light-receiving element module characterized by searching for a point that has a desired sensitivity and gives the lowest distortion rate by translating in the Y plane and along the Z direction parallel to the axis, and fixing an optical fiber to that point. Method. 3. The light receiving device according to claim 1, wherein a light receiving surface of the photodiode is set at a position closer to the converging lens than a Gaussian image surface of a fiber exit end face by the converging lens. A method for manufacturing an element module. 4. When aligning an obliquely cut fiber at an optimal position deviated from a lens axis and an axis perpendicular to a PD light receiving surface,
The method for manufacturing a light-receiving element module according to any one of claims 1 to 3, wherein rotation alignment around the optical fiber is not performed. 5. An optical fiber having an obliquely cut end face, a lens for condensing light emitted from the optical fiber,
A light receiving element chip for receiving the light condensed by the lens, the lowermost point of the inclined end surface of the fiber is S, the uppermost point is T, and the oblique emission angle from the fiber end surface is α,
Assuming that the distance between the light receiving element chip and the lens center is L, the chip center O is shifted from the lens center H in the direction of the lowest point S of the fiber end face by Ltanα in the direction perpendicular to the axis, and the fiber center Q is more than the lens center. A light receiving element module, which is shifted from a direction perpendicular to an axis in a direction opposite to a lowermost point S of an end face. 6. The light-receiving element module according to claim 5, wherein the light-receiving surface of the photodiode is set at a position closer to the converging lens than the Gaussian image plane of the fiber exit end face by the converging lens. . 7. The light receiving element module according to claim 5, wherein the oblique cutting angle of the tip of the optical fiber is 4 ° to 10 °. 8. The light receiving element module according to claim 5, wherein the lens is a spherical lens. 9. A single mode fiber having a core diameter of about 1
0 μm, the oblique cutting angle of the optical fiber is about 8 °, the distance between the fiber and the lens center H is about 1850 μm, the lens is a spherical lens with a refractive index of 1.5 and a diameter of 1.5 mm, and the center of the lens is The distance L between H and the light receiving surface of the light receiving element is about 2100 µm, and the displacement X of the chip with respect to the center of the lens is 50 to 180 µm.
9. The light-receiving element module according to any one of items 1 to 8. 10. The light receiving element module according to claim 5, wherein the light receiving element chip has an InGaAs light receiving layer. 11. The light-receiving element module according to claim 5, wherein the light-receiving element chip has an InGaAsP light-receiving layer. 12. A light-receiving element chip comprising: an InP substrate;
Buffer layer, InGaAs or InGaAsP light-receiving layer,
11. An InP window layer, comprising:
2. The light receiving element module according to 1. 13. A light-receiving element chip is also provided with a Z around the light-receiving layer.
13. The light-receiving element module according to claim 5, wherein the light-receiving element module has a pn junction by n-diffusion, and removes unnecessary photocarriers due to leakage light around the light-receiving surface.