JPS62232975A - Photoconducting detector - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] This invention relates to photoconductive detectors, and more particularly to high speed, high gain detectors used in optical communication systems.

[Background of the invention]

A typical intrinsic photoconductive detector consists of a body of photosensitive material with a pair of spaced apart bias contacts.

Incident optical radiation energy with an energy equal to or greater than the band gap of the photosensitized material is absorbed and generates electron-hole pairs, thereby increasing electrical conductivity.

It is known that the sensitivity of a detector is increased by photoconductive gain. Photoconductive gain is generally due to the time difference that electrons and holes exist within the photoconductive material.

For example, indium gallium arsenide at room temperature
The mobility of electrons in aAs) is about 40 times the mobility of holes in this material. When a photon is absorbed and an electron-hole pair is generated, many electrons are injected by the negatively biased contact and traverse the photoconductive material while the photogenerated hole is present in the material. . Therefore, the photoconductive gain is approximately equal to the number of electrons traversing the photoconductive material for each absorbed photomolecule.

Equally important to a detector's performance is its bandwidth, which is a practical measure of how fast the device responds, i.e. how many bits of information it can detect per second. . Bandwidth is determined by the time that the photogenerated hole exists in the photoconductive material. The longer the hole exists in the photoconductive material, the smaller the bandwidth, ie, the slower the response of the device.

In the field of high speed optical port systems with data rates in the gigabit/second range, lateral photoconductors are of particular interest due to their performance in bandwidth and photoconductive gain.

By lateral photoconductor we mean a device in which light is incident on a thin layer of photoconductive material in a direction perpendicular to the direction of current flow.

Lateral photoconductors are also ideal for use in monolithically integrated photoreceptors because their structure is compatible with that of field effect transistors.

Such lateral photoconductive devices usually use materials with high mobility ratios to obtain higher interest rates.
In order to obtain a wider bandwidth, a mechanism is required to reduce the residence time of holes in the active region. One method is to form a detector consisting of a substrate, typically of N+ conductivity type, and a typically undoped or lightly doped active region disposed on the substrate. An anode and a cathode are formed in an interdigital structure on the upper surface of the active region, and the anode-cathode spacing is sufficient for holes to be removed from the photosensitive region at least as fast as the data rate. It must be small. However, in doing so, approximately 30% to 50% of the incident light is weakened by the anode and cathode formed in the interdigitated structure on the detector surface, reducing the incoming signal. , when both functions are provided on one surface, a non-uniform electric field is created across the photosensitive area, which has the disadvantage of reducing bandwidth and producing spatially non-uniform benefits. Moreover, much of the incident light is absorbed near the anode and cathode, resulting in a loss of up to 50% of the achievable gain.

Another method that has been investigated to accommodate data rates in the gigabit/second range is a modification of the interdigital structure. For example, one method is to utilize a reverse biased pn junction behind the photoconductive layer to remove holes from the photoconductive material. However, such devices are characterized by reduced gain and increased generated-recombined noise.

Alternatively, vertical photoconductors have recently been studied.1 The prior art photoconductive device 10 shown in FIG. type active region 14 and a thin contact window 16 of N+ conductivity type.
It consists of The thickness of active region 14 is typically about 2
~31tm. The annular metal contact 20 delimits the light entrance opening 18, which has a large area compared to the lateral device. The N+ substrate 12 and the N+ contact window 16 are in ohmic contact with the active region 14 over its entire width. In theory, this should provide a uniform vertical electric field in the active region 14. Long wavelength (1,0~
1.611 m ) application, active region 14 is typically made of TnGaAs and contact window 16 is made of the same material for reasons related to noise and resistance levels. However, this ensures that the contact window 16 is relatively thin, i.e. about 0.5 to 1.5 mm, to prevent most of the light from being absorbed by the contact window 16. The thickness must be 1 m. With such a thickness, the presence of annular contact 20 causes a lateral electric field to be generated across contact window 16 and into active region 14 . Therefore, the electron and hole trajectories become transverse and the distance they travel is greater than the width of the active region 14, which increases the gain but significantly reduces the bandwidth. 3. Furthermore, since the substrate 12 is typically made of indium-phosphorus (N-P), holes that move toward the substrate 12 exist relative to minority carriers at the InGaAs/N-P interface 24. may be captured in a known charge trap. This would also increase gain but significantly reduce bandwidth.

A photoconductive detector has been developed that is suitable for high speed operation, has high gain, reduced noise and trapping, and is easily amenable to monolithic integration.

[Invention (already required)]

The photoconductive detector of the present invention is formed of a semiconductor material of a first conductivity type that is substantially transparent to light at the wavelength to be detected and is in ohmic contact with an overlying photoconductive active region. The substrate layer is doped to a sufficient extent to The active region comprises a body of undoped semiconductor material that absorbs light at the wavelength to be detected, and the three substrate layers having first and second major surfaces cover the entire first major surface of the active region. A metal or metal alloy acts as a second ohmic contact overlying the entire second major surface of the active region.

[Detailed explanation of recommended examples]

In FIG. 2, the anti-conductive vertical photoconductive detector 25 includes a substrate layer 28 and a heterojunction interface 32 on the substrate layer 28.
0 active region 3 consisting of a photoconductive layer 30 located through
4, it consists of a portion of the photoconductive layer 30 to which a bias is applied. Substrate layer 28 acts as a first optically transparent ohmic contact, and a second ohmic contact 36 is located over the entire active area 34 . The area of the 20th ohmic contact 36 contacting the active area 34 is defined by a dielectric area 38 . An annular electrical contact means 40 defines an open area 42 for light to enter the substrate layer 28 . A typical diameter of the aperture region 42 is on the order of 20 to 301 Lm, thereby allowing substantially all of the light to enter the substrate layer 28. As an anti-reflection coating, for example, S
i3N4, S-, o, 5102, etc. are used.

Substrate layer 28 is of a first conductivity type and is sufficiently doped to make ohmic contact with active region 34 .

As a representative example, the substrate layer 28 may be a heavy substrate layer 28.
must be substantially transparent to light at the wavelength to be detected, i.e. must have a transmittance of at least 90%, so that the substrate layer 28 has a thickness of approximately 150 p,m.
Shouldn't get any thicker.

The substrate 7g28 should be at least 57 zm thick to avoid lateral electric fields from the annular contact means 40. Substrate layer 28 is typically made of gallium arsenide (GaA
s) or of a semiconductor of a higher conductivity type, such as indium phosphorus (nP), and can serve, for example, as the substrate of the detector or be a layer on top of another substrate.

The active region 34 may be defined by a mesa structure or may be a portion of the light guide layer 30 across which a bias is applied. In the latter case, the amount of photoconductive layer 30 in contact with second ohmic contact 36 substantially determines the width of active region 340. The effective width of the active region 34 is typically 2 to 31 Lm thick in the opening region 42.
The amount of spread of the electric field across is small (1p, m or 211
m), so in practice it may be slightly larger than the width of the second ohmic contact 36. Active region 34 may be intrinsic or of the same conductivity type as substrate layer 28, but is typically undoped or naturally lightly doped. The active area 34 is about 1o15/cy! It should have a very low concentration of mobile charge carriers, preferably in the range of 10''/l near 1f3 or even lower.

Alternatively, active region 34 can be doped to increase its resistance to reduce noise. However, this reduces the mobility of charge carriers in the active region 34. Active region 34 consists of a photoconductive semiconductor material that absorbs the light to be detected. The material used in the active region 34 has a high mobility ratio, that is, a ratio of majority carrier to minority carrier mobility (electrons/holes in an N-type material) between about 10:1 and 20=1; Preferably a ratio of 40:1 or higher is advantageous. Ga
More N-conductivity type materials such as As, GaAs, etc. are well suited for this purpose.

The 20th ohmic contact 36 is typically a metal alloy that can continuously supply majority carriers to the active region 34 without voltage drop. For example, for a detector 26 having an N+ type nP substrate layer 28 and an N- type nGaAS active region 34, gold-tin, gold-germanium, etc. are suitable materials for the second ohmic contact 36. This second ohmic contact 36
may be formed to a desired width for the active region 34, or the width of this contact may be formed by forming a second ohmic contact 36 on a region 38 of dielectric material, such as arsenic dioxide.
The successive layers may be specified as appropriate.

The electrical contact means 4°, which is shown as annular in FIG. .

It will be appreciated that the regions and layers described above may be of opposite conductivity type as long as the relationship of relative conductivity types is maintained.

The active region 34 on the substrate layer 28 is preferably formed by liquid phase epitaxy or molecular beam epitaxy, as described, for example, in "Ga Techniques I" by T. P. Pearsal et al.
AllOySemiCON(lAl10ySe (Jo
hn w11ey & 5oz.

1Vapour-phas published in 1982)
It can be formed by vapor phase epitaxy as described in the paper by G.

In operation, the detector 26 has a highly uniform electric field due to the first ohmic contact, ie, the substrate layer 28, and the twentieth ohmic contact 36 completely covering the first and second major surfaces of the active region 34, respectively. It has significant advantages over interdigital channels. In the case of an InGaA6/InP nP detector, the InGaA substrate layer 28 is completely transparent to the wavelength of light to be detected, with an InGaA active region 34 having a thickness of 1507 zm or less. ing.

This substrate, 928, can therefore be made sufficiently thick to reduce any lateral field effects due to the annular electrical contact means 40, which is a typical feature of prior art vertical photoconductive devices. Although the detector will work when biased in either direction, it is preferable to bias it so that the substrate layer 28 is the anode and the second ohmic contact 36 is the cathode. By this method, holes generated near the nP/nGaAS interface migrate toward the second ohmic contact/active region interface, which is substantially free of charge traps compared to prior art detectors.

In this way, without creating a lateral electric field effect,
Additionally, the first and second ohmic contacts can easily provide a uniform electric field across the active region while minimizing light absorption by the substrate layer, resulting in significantly uniform gain. Additionally, bandwidths in the gigabit/second range are obtained due to the reduced charge trapping of holes at the active region/metal ohmic contact interface.

Another detector 44 of the present invention is shown in FIG. 3, having the same basic elements as detector 26 of FIG. In this detector 44, after forming the photoconductive layer 46 on the substrate layer 48, a mesa structure is formed by known methods, and the width of the active region 50 is defined by the width of the second ohmic contact 52. The contour of this second ohmic contact 52 is the dielectric region 5
Formed by 4. Due to the mesa structure, the electrical contact means 56, for example an annular structure, can be provided on the same side of the detector 44 on which the second ohmic contact 52 is provided, which provides clear design advantages. Also, the width of the active region can be specified by the mesa structure, but in this case, the second ohmic contact needs to cover the entire surface of the mesa structure.

In addition to the high gain, uniform electric field, and gigabits/second bandwidth mentioned above, the inverted photoconductive detector disclosed herein also has
During its manufacturing stage, it has the advantage that only one epitaxial layer needs to be grown on the substrate before forming the second ohmic contact using known photolithographic and metallization techniques. The detector can also be manufactured as a mesa structure or more conveniently as a flat structure,
It is well suited for monolithic integration and can be applied to optical receiver devices.

[Brief explanation of drawings]

1 is a sectional view of a prior art vertical photoconductive detector, FIG. 2 is a sectional view of a first embodiment of a vertical photoconductive detector according to the invention, and FIG. 3 is a sectional view of a vertical photoconductive detector according to the invention. FIG. 6 is a cross-sectional view of a second embodiment of child detection; 26... Photoconductive detector, 28... Substrate layer, 3o, 34
. . . active region, 32 . . . first ohmic contact, first main surface of the substrate layer and active region, 36 . . . second ohmic contact, 40 . . . electrical contact means. Patent applicant: RCSNY Corporation
Professor Tetsu Shimizu and 2 others 2nd mouth 2nd mouth 3rd mouth

Claims (1)

[Claims] (1) having an active region, a substrate layer acting as a first ohmic contact to the active region, a second ohmic contact to the active region, and an electrical contact to the substrate layer, characterized in that: The substrate layer is substantially transparent to light at the wavelength to be detected and to a sufficient extent that a first major surface of the substrate layer acts as the first ohmic contact to the active region. comprising a semiconductor material of a first conductivity type doped to the substrate layer, the active region comprising an undoped semiconductor material overlying the substrate layer and absorbing light at the wavelength to be detected; a first major surface in contact with the layer and a second major surface, the second ohmic contact being comprised of a metal or metal alloy and positioned over the entire second major surface of the active region. A photoconductive detector configured to.