JP2021100023A - Photodetector and imaging device using the same - Google Patents

Abstract

To suppress surface leakage of a light absorption layer and reduce crosstalk between pixels by a photodetector sensitive to at least two wavelengths.SOLUTION: A photodetector with an array of a plurality of pixels includes a laminate including a first contact layer to which a common bias voltage is applied, a first light absorption layer that responds to light of a first wavelength, and a second light absorption layer that responds to light of a second wavelength longer than the first wavelength, and the first contact layer is located on the light incident side, the second light absorption layer is located on the element surface side, and the second light absorption layer is not physically divided and is continuous over the entire pixel region including the array of the plurality of pixels, the second contact layer separated by a pixel separation groove is arranged on the surface of the second light absorption layer, and a partition corresponding to each pixel is formed in the first light absorption layer by one or more grooves extending in the stacking direction.SELECTED DRAWING: Figure 2

Description

The present invention relates to a photodetector and an imaging device using the same.

Conventionally, an optical sensor that responds to two wavelengths with one element (see, for example, Patent Document 1) and an infrared detector (for example, see Patent Document 2) have been known.

When a Type-II Super-Lattice (T2SL) is used as the photoabsorbing layer of a photosensor or photodetector, the incident light is absorbed by the T2SL layer and electrons and holes, which are photoexcited carriers, are generated. The photocurrent signal is obtained. In addition to photoexcitation, carriers are also induced by thermal energy, and surplus current other than photoresponse flows. This surplus current is called "dark current". Since the dark current is a noise component, it is necessary to suppress the dark current in order to improve the signal-to-noise (S / N) ratio of the optical sensor. In particular, a T2SL device including an absorption layer in a long wave length (LW) band with a small bandgap is vulnerable to thermal excitation and tends to generate a dark current.

In the T2SL element, the dark current flowing on the exposed surface or the side wall such as the pixel separation groove is dominant rather than the dark current generated in the bulk. It is known that a photodetector having a single wavelength has a structure in which pixels are defined by metal contacts to reduce surface leakage without providing a pixel separation groove in the light absorption layer (see, for example, Patent Document 3).

Japanese Unexamined Patent Publication No. 11-15288 Japanese Unexamined Patent Publication No. 2015-155839 U.S. Pat. No. 9613999

It is difficult to directly apply the configuration of a single-wavelength photodetector whose pixels are defined by metal contacts to a photodetector having sensitivity to two wavelengths. In a photodetector in which two or more light absorbing layers having sensitivity to different wavelengths are laminated, even if the light absorbing layer located near the metal contact can suppress signal crosstalk, the light incident side away from the metal contact. This is because the excitation carriers generated in the light absorption layer of the above are likely to be mixed in the output of the adjacent pixel.

An object of the present invention is to provide a photodetector having sensitivity to at least two wavelengths, which suppresses surface leakage of a light absorption layer and reduces crosstalk between pixels.

In one aspect of the invention, in a photodetector having an array of multiple pixels,
A first contact layer to which a common bias voltage is applied, a first light absorption layer that responds to light of a first wavelength, and a second that responds to light of a second wavelength longer than the first wavelength. The first contact layer is located on the light incident side, and the second light absorbing layer is located on the surface side of the device.
The second light absorption layer is not physically divided and is continuous over the entire pixel region including the array of the plurality of pixels, and is separated on the surface of the second light absorption layer by a pixel separation groove. A second contact layer is placed
In the first light absorption layer, a partition corresponding to each pixel is formed by one or more grooves extending in the stacking direction.

A photodetector that is sensitive to at least two wavelengths can suppress surface leakage of the light absorption layer and reduce crosstalk between pixels.

It is a figure explaining the problem when the dark current suppression structure of a single wavelength optical sensor is applied to a two wavelength sensor. It is sectional drawing of the pixel array of the 2 wavelength sensor of an embodiment. It is a top view seen from the element surface side of a pixel array. It is a top view seen from the light incident side of a pixel array. FIG. 4 is a cross-sectional view taken along the line BB'in FIG. It is a manufacturing process diagram of the photodetector of an embodiment. It is a manufacturing process diagram of the photodetector of an embodiment. It is a manufacturing process diagram of the photodetector of an embodiment. It is a manufacturing process diagram of the photodetector of an embodiment. It is a manufacturing process diagram of the photodetector of an embodiment. It is a manufacturing process diagram of the photodetector of an embodiment. It is a manufacturing process diagram of the photodetector of an embodiment. It is a manufacturing process diagram of the photodetector of an embodiment. It is a manufacturing process diagram of the photodetector of an embodiment. It is a manufacturing process diagram of the photodetector of an embodiment. It is a manufacturing process diagram of the photodetector of an embodiment. It is a manufacturing process diagram of the photodetector of an embodiment. It is a manufacturing process diagram of the photodetector of an embodiment. It is a manufacturing process diagram of the photodetector of an embodiment. It is a schematic perspective view of the photodetector of an embodiment. It is a schematic diagram of the image pickup apparatus using the light detection apparatus of embodiment.

FIG. 1 is a diagram for explaining a problem that arises when the dark current suppression structure of a single-wavelength optical sensor is applied to a two-wavelength sensor as it is. In the pixel array of the optical sensor having sensitivity to two different wavelengths, the common contact layer 16, the MW (Mid Wavelength) band absorption layer 15, the block layer 14, and the LW band absorption layer 13 are laminated in this order. When the dark current suppression structure of the single-wavelength optical sensor is applied to the pixel array of the two-wavelength sensor as it is, the laminate from the common contact layer 16 to the LW band absorption layer 13 is not physically divided.

On the surface of the LW band absorption layer 13, a contact layer 12 that defines a pixel and a metal contact 36 that is electrically connected to the contact layer 12 are arranged. The pixel separation groove 21 is formed only in the vicinity of the surface of the device, and the separation groove is not formed in the LW band absorption layer 13 having a small band gap and easy flow of dark current. With this configuration, it is possible to reduce the exposure of the side surface of the LW band absorption layer 13 and reduce the dark current.

Although the LW band absorption layer 13 is not physically divided, the photoexcitation generated in the LW band absorption layer 13 close to the contact layer 12 due to the bias voltage applied between the common contact layer 16 and the individual contact layers 12 The carrier is pulled out for each pixel. In the LW band absorption layer 13 located directly below the contact layer 12, the leakage of photocurrent to the adjacent pixels when an electric field is applied is so small that it can be ignored.

In the case of T2SL, the MW band absorption layer 15 having a short response wavelength is arranged on the light incident side due to the relationship between the band gap of the light absorption layer and the wavelength response characteristic. Being close to the light incident surface means that it is separated from the contact layer 12 on the surface of the device. By switching the polarity of the bias voltage applied between the common contact layer 16 and the individual contact layers 12, the photocurrent PC generated in the MW band absorption layer 15 can be taken out, but since it is far from the contact layer 12 on the surface, The mileage of the photoexcited carrier in the undivided region becomes long. As a result, as shown in FIG. 1, a part of the photoexcited carriers flows out to the adjacent pixels. Although the dark current in the LW band absorption layer 13 can be reduced, another problem arises in which signal crosstalk is generated by the photoexcited carriers generated in the MW band absorption layer 15. In other words, the problems that were avoided with single-wavelength sensors become apparent with multi-wavelength sensors.

FIG. 2 is a schematic cross-sectional view of the pixel array 10 used in the photodetector of the embodiment. In the embodiment, in order to solve the problem of crosstalk due to photoexcited carriers generated in the MW band absorption layer 15, pixels are partitioned in the MW band absorption layer 15 without physically dividing the LW band absorption layer 13. One or more grooves 17 are provided.

The pixel array 10 is, for example, a two-dimensional array in which a large number of pixels 101 are arranged in the X direction and the Y direction. The common contact layer 16, the MW band absorption layer 15, the block layer 14, and the LW band absorption layer 13 are laminated in this order in the Z direction. On the surface of this laminated body, a contact layer 12 separated by a pixel separation groove 21 for each pixel 101 and a metal contact 36 electrically connected to the contact layer 12 are provided.

The surface on which the contact layer 12 and the metal contact 36 are formed is the element surface, and the common contact layer 16 is located on the light incident side. In the XX cross section of FIG. 2, the common contact layer 16 is drawn so as to be separated, but the common contact layer 16 is continuous in the XY plane, and a bias voltage common to all pixels is applied. Be supplied.

The MW band absorption layer 15 on the light incident side is, for example, a layer of T2SL having sensitivity in a wavelength range of 3 to 5 μm. The LW band absorption layer 13 far from the light incident side is a layer of T2SL having sensitivity in a wavelength range of, for example, 8 to 12 μm. T2SL is formed by repeatedly laminating dissimilar crystal materials having similar lattice constants in a short cycle. One of the materials to be laminated has a lower electron potential and a higher hole potential than the other material.

In T2SL, in addition to the optical transition between the band gaps of the semiconductor material, light is absorbed by the optical transition between the mini bands formed in the superlattice, so that the quantum efficiency is high. The cutoff wavelength can be changed by controlling the film thickness of the material forming the superlattice. In particular, from medium-wavelength infrared rays (3 to 5 μm) to long-wavelength infrared rays (8 to 12 μm), the wavelength to be detected can be relatively easily designed.

A block layer 14 is arranged between the MW band absorption layer 15 and the LW band absorption layer 13. The block layer 14 does not inhibit the movement of excited carriers (minority carriers) generated in the LW band absorption layer 13 or the MW band absorption layer 15, but allows the movement of each of the large number carriers of the LW band absorption layer 13 and the MW band absorption layer 15. Block.

In the pixel array 10, the LW band absorption layer 13 having a small band gap is not physically divided, but the MW band absorption layer 15 is formed with a groove 17. The LW band absorption layer 13 is directly below the contact layer 12 separated for each pixel 101, and unless the thickness is made excessively thick, the minority carriers generated by photoexcitation are metal contacts along the applied electric field. It is pulled out to 36, and leakage to adjacent pixels can be sufficiently suppressed.

On the other hand, the MW band absorption layer 15 has a section 151 corresponding to each pixel 101 defined by the groove 17. Since the MW band absorption layer 15 has a wider bandgap than the LW band absorption layer 13, even if the side surface of the MW band absorption layer 15 is exposed in the groove 17, there is almost no effect of the increase in dark current.

The MW band absorption layer 15 does not need to be completely separated into individual pixels by the groove 17, and may be partially connected in the XY plane as in the common contact layer 16. As long as the optical carriers generated in the compartment 151 partitioned by the groove 17 inside the MW band absorption layer 15 are pulled out along the corresponding metal contact 36 along the electric field in the stacking direction, even if the adjacent pixel regions are partially connected. , The effect on signal crosstalk is small.

The groove 17 does not need to penetrate the MW band absorption layer 15 in the stacking direction, and may have a depth that can suppress the outflow of photoexcited carriers to the adjacent pixels 101. As long as crosstalk between the pixels 101 can be prevented, the depth of the groove 17 may have some manufacturing variation. The groove 17 may partially partition the block layer 14, or may reach the LW band absorption layer 13 to the extent that the surface dark current does not affect the groove 17.

As described above, the groove 17 is formed so that the continuity of the common contact layer 16 is maintained and the groove 17 conducts over the entire pixel region when a common bias is applied. This point is different from the configuration in which the pixel separation groove 21 on the surface of the element completely separates the contact layer 12 for each pixel 101.

FIG. 3 is a diagram showing a planar arrangement of the pixel separation grooves 21 on the surface of the element, and FIG. 4 is a diagram showing an example of the planar arrangement of the grooves 17 seen from the light incident side of the back surface of the element. The vertical cross section along the AA'line of FIG. 4 is the cross-sectional shape of FIG.

In FIG. 3, the contact layer 12 is separated for each pixel by the pixel separation groove 21. Below the contact layer 12, a continuous LW band absorption layer 13 is arranged without being divided.

In FIG. 4, a plurality of grooves 17 for partitioning regions corresponding to individual pixels 101 are formed in the common contact layer 16. The common contact layer 16 is not completely divided and is continuous throughout the pixel array 10. In this example, the region corresponding to the pixel 101 is partitioned by the four grooves 17, and the common contact layer is connected to the adjacent region at the four corners of the partitioned region.

FIG. 5 is a cross-sectional view taken along the BB'line of FIG. In the vertical cross section along the BB'line, the common contact layer 16 and the MW band absorption layer 15 are continuous, and the end of the groove 17 is located at the position indicated by the broken line. The positions where the common contact layer 16 is continuous are not limited to the four corners of the region surrounded by the four grooves 17, and may be continuous at three corners or two corners of the rectangular region. The portion where the common contact layer 16 is continuous is not limited to the corner of the rectangular region, and may be continuous with the adjacent region in the vicinity of the corner or in the middle of the side of the rectangular region. As long as the compartment 151 in which the photoexcited carriers generated in the MW band absorption layer 15 can be held is formed and the common contact layer 16 is continuous in the entire pixel array 10, the arrangement configuration of the groove 17 can be appropriately designed.

During operation, for example, by applying a common bias to the common contact layer 16 and applying a positive voltage to the selected metal contact 36, electrons generated in the LW band absorption layer 13 directly below the selected contact layer 12 are generated. It is pulled out by the metal contact 36. Even if the LW band absorption layer 13 is not physically separated, the carriers generated in the LW band absorption layer 13 close to the contact layer 12 can travel a short mileage along the electric field applied to the selected region. Reach the metal contact 36.

By applying a common bias to the common contact layer 16 and applying a negative voltage to the selected metal contact 36, holes generated in the corresponding region of the MW band absorption layer 15 are pulled out to the metal contact 36. Since the partition 151 corresponding to the pixel 101 is formed in the MW band absorption layer 15 due to the presence of the groove 17, it is possible to suppress the outflow of the carrier to the adjacent pixel when the photoexcited carrier is pulled out. When the carriers generated in the MW band absorption layer 15 pass through the block layer 14 and enter the LW band absorption layer 15, they are pulled out to the metal contact 36 in a short mileage along the applied electric field, and the polarity is reversed. An optocurrent signal is obtained.

By switching the polarity of the bias voltage applied to the contact layer 12, the photocurrent signal that responds to the incident light in the MW band and the photocurrent signal that responds to the incident light in the LW band can be individually extracted from one pixel 101. A dual-wavelength infrared sensor can be realized.

By using the configurations of FIGS. 2 to 5, the side exposure in the LW band absorption layer 15 having a small bandgap is reduced to suppress the dark current, and the signal generated by the carrier in the MW band absorption layer having a large bandgap is suppressed. Crosstalk can be suppressed.

Since the pixel array 10 of the embodiment has a simple configuration in which each pixel 101 has one metal contact 36, miniaturization is easy. Based on the correlation between the two wavelength outputs obtained from one pixel, processing may be performed to improve the accuracy of the temperature absolute value measurement of the observation target. Further, a process of discriminating between the reflected light component of natural light (sunlight) and the temperature radiation component from the object itself may be performed from the infrared information received from the target object.

6A to 6N are manufacturing process diagrams of the photodetector of the embodiment. Of these, FIGS. 6A to 6L are pixel array manufacturing processes.

In FIG. 6A, the contact layer 12, the LW band absorption layer 13, the block layer 14, the MW band absorption layer 15, and the common contact layer 16 are crystal-grown on the surface of the substrate 11 by the molecular beam epitaxy method in this order. When a GaSb substrate is used as the substrate 11, for example, the contact layer 12, the LW band absorption layer 13, the block layer 14, and the MW band absorption layer 15 are formed by repeating different semiconductor materials having a lattice constant close to that of GaSb of the substrate 11. Will be done.

As an example, the contact layer 12 is a superlattice in which InAs having a thickness of 4 nm and GaSb having a thickness of 2 nm are repeated for 80 cycles, and p-type impurities are added at a concentration of ^{1 × 10 18} cm ^-3. A GaSb layer having a p-type impurity concentration of 1 × 10 ¹⁸ cm ^{-3 and} a thickness of about 500 nm may be inserted between the GaSb substrate 11 and the contact layer 12.

On the contact layer 12 of the InAs (4 nm) / GaSb (2 nm) superlattice, InAs having a thickness of 4 nm and GaSb having a thickness of 2 nm are repeated for 200 cycles to form an LW band absorption layer 13. The p-type impurity concentration of the LW band absorption layer 13 is, for example, 1 × 10 ¹⁶ cm ^-3 .

A block layer 14 is formed on the LW band absorption layer 13 of the InAs (4 nm) / GaSb (2 nm) superlattice with a superlattice in which InAs having a thickness of 5 nm and AlSb having a thickness of 1 nm are repeated for 20 cycles. The p-type impurity concentration of the block layer 14 is, for example, 1 × 10 ¹⁶ cm ^-3 . The block layer 14 blocks the movement of the majority carriers of the LW band absorption layer 13 and the MW band absorption layer 15, but causes the movement of the minority carriers (photoexcited carriers) generated in the LW band absorption layer 13 or the MW band absorption layer 15. Does not interfere.

On the block layer 14 of the InAs (5 nm) / AlSb (1 nm) superlattice, the MW band absorption layer 15 is formed by repeating 200 cycles of InAs having a thickness of 3 nm and GaSb having a thickness of 1 nm for 200 cycles. The p-type impurity concentration of the MW band absorption layer 15 is, for example, 1 × 10 ¹⁶ cm ^-3 .

A common contact layer 16 is formed on the MW band absorption layer 15 of the InAs (3 nm) / GaSb (1 nm) superlattice with a superlattice in which InAs having a thickness of 3 nm and GaSb having a thickness of 1 nm are repeated for 100 cycles. The concentration of p-type impurities in the common contact layer 16 is, for example, 1 × 10 ¹⁸ cm ^-3 . An undoped InAs cap layer having a thickness of about 30 nm may be provided on the common contact layer 16.

In FIG. 6B, a groove 17 for the MW band is formed. A SiON mask having a predetermined opening is formed on the outermost surface of the crystal-grown laminate (common contact layer 16 in the example of FIG. 6B) by lithography and dry etching, and a common contact layer is formed by dry etching using the SiON mask. A groove 17 is formed from the surface of 16 to reach, for example, the LW band absorption layer 13. As described above, the depth of the groove 27 does not necessarily have to reach the LW band absorption layer 13, and is a depth that can prevent the outflow of photoexcited carriers to adjacent pixels, for example, up to the middle of the block layer 14. It may be.

The grooves 17 are independent of each other, and the common contact layer 16 and the MW band absorption layer 15 are continuous in the in-plane direction. The planar arrangement of the grooves 17 may be the arrangement configuration shown in FIG. 4, or a partition 151 corresponding to a pixel may be defined by combining two L-shaped grooves. After forming the groove 17, the SION mask is removed.

In FIG. 6C, the insulating film 18 is formed on the entire surface including the inside of the groove 17. As an example, a SiO ₂ film is formed by plasma CVD. It is not always necessary that the entire inside of the groove 17 is embedded with the insulating film 18. Since the MW band absorption layer 15 may be provided with a partition from the adjacent pixel, the air layer may remain inside the groove 17. Even if a part of the MW band absorption layer 15 is exposed inside the groove 17, the band gap of the MW band absorption layer 15 is wider than that of the LW band absorption layer 13, so the influence of the dark current on the surface is ignored. Small enough to get.

In FIG. 6D, a support substrate 20 having an insulating film 22 formed on its surface is prepared. The insulating film 18 of the substrate 11 and the insulating film 22 on the support substrate 20 are bonded together by turning the substrate 11 having the laminate formed by the processes of FIGS. 6A to 6C upside down. As long as the insulating films are bonded to each other, the types of the insulating film 22 and the insulating film 18 may be the same or different. In this example, the insulating film 18 and the insulating film 22 are pressure-bonded by using the GaAs support substrate 20 on which the insulating film 22 of _{SiO 2 is formed.} Due to _{the intermolecular binding force of SiO 2} , the insulating film 18 and the insulating film 22 are integrated to form the insulating film 23.

In FIG. 6E, the GaAs substrate 11 is removed. By mechanical grinding and wet etching, the substrate 11 is removed so that the InAs of T2SL of the contact layer 12 is exposed on the surface. The reason why the outermost layer is InAs is that GaSb is easily oxidized, and if metal Sb is generated on the surface, it may contribute to a dark current. When performing wet etching, an etchant having a high selection ratio of GaSb is used. As an example, a mixed solution of phosphoric acid, hydrogen peroxide solution, citric acid, and water is used. This mixed solution has a high etching rate for GaSb, but hardly etches InAs. As a result, a laminate of epitaxial crystal growth layers having InAs as the outermost layer is transferred onto the support substrate 20.

As described above, in the illustrated cross section, the groove 17 or the insulating film 18 appears in the common contact layer 16 of the laminate transferred to the support substrate 20, but the common contact layer 16 is continuous in the plane. ..

In FIG. 6F, a resist mask having a predetermined opening pattern is formed on the surface of the contact layer 12, and a pixel separation groove 21 reaching the LW band absorption layer 13 or its vicinity is formed by dry etching. Unlike the groove 17 formed in the MW band absorption layer 15, the pixel separation groove 21 completely partitions the pixels (see FIG. 3). After that, the resist mask is removed.

In FIG. 6G, a groove 24 for a common electrode is formed on the outside of the region where the pixels are two-dimensionally arranged on the support substrate 20. A SiON hard mask having an opening pattern corresponding to the shape of the groove 24 is formed on the outermost surface of the laminate, and the groove 24 reaching the common contact layer 16 is formed by dry etching. Then remove the hardmask.

In FIG. 6H, the protective film 25 is formed on the entire surface. The protective film 25 is, for example, a film of _{SiO 2 or SiN formed by plasma CVD.}

In FIG. 6I, an ohmic electrode 26 connected to the contact layer 12 defining each pixel and an ohmic electrode 26C connected to the common contact layer 16 on the bottom surface of the groove 24 for the common electrode are formed. Specifically, by forming a resist mask by photolithography and etching using the resist mask, an opening is formed at a required portion of the protective film 25, and the contact layer 12 and the common contact layer 16 are exposed in the opening. .. After that, for example, an AuGe film is formed by vacuum deposition. By removing the AuGe film on the resist mask together with the resist mask by lift-off, an ohmic electrode 26 that makes ohmic contact with the contact layer 12 of each pixel and an ohmic electrode 26C that makes ohmic contact with the common contact layer 16 are formed.

In FIG. 6J, the base electrodes 27a and 27b of the bump are formed. A resist mask having an opening pattern is formed at a portion where the bump electrode is provided and a portion where the wiring is drawn out from the groove 24 for the common electrode, and a Ti / Pt metal laminated film is formed on the entire surface by a sputtering method. Then, the Ti / Pt film on the resist mask is removed by lift-off. In each pixel region, a base electrode 27a connected to the ohmic electrode 26 is formed. In the peripheral region, a base electrode 27b is formed which is connected to the ohmic electrode 26 on the bottom surface of the groove 24 and is drawn out from the side wall of the groove 24 to the element surface.

In FIG. 6K, the entire surface is again _{covered with the protective film 28 of SiO 2} , and then dry etching is performed on the portion where the bump electrode is arranged to form the opening 29. In the pixel region, the base electrode 27a is exposed in the opening 29, and in the peripheral region, a part of the base electrode 27b drawn out to the surface of the element is exposed. A part of the protective film 28 formed in this step is integrated with the conventional protective film 25.

In FIG. 6L, the bump electrode 31 is formed on the exposed base electrode 27a, and the bump electrode 32 is formed on the base electrode 27b. The bump electrode 31 is an electrode for reading a signal from each pixel 101. The bump electrode 32 is an electrode for applying a common bias voltage to the common contact layer 16 in the peripheral region. The bump electrodes 31 and 32 are formed by forming a resist mask having a predetermined opening pattern, depositing In on the entire surface, and lifting off. As a result, the processing of the pixel array 10 having the pixels 101 arranged two-dimensionally and the dummy pixels 102 for applying the common bias is completed.

In FIG. 6M, the pixel array 10 is flip-chip connected to the read circuit 5 by the bump electrodes 31 and 32. The bump electrode 31 is fused with a protrusion electrode or an In bump provided in the read circuit 5 to form a protrusion electrode 34 that electrically connects each pixel 101 (see FIG. 6L) of the pixel array 10 to the read circuit 5. Will be done. The bump electrode 32 in the peripheral region is fused with the protrusion electrode or the In bump provided in the read circuit 5 to form the protrusion electrode 34COM that supplies a common bias to the pixel array 10. An underfill may be filled between the pixel array 10 and the readout circuit 5 to cure the pixel array 10.

In FIG. 6N, the support substrate 20 that is no longer needed in the pixel array 10 is removed. For example, the GaAs support substrate 20 is back-ground to a thickness of several tens of μm, and then removed by wet etching. As a result, the laminated portion of the light receiving element whose surface is covered with the insulating film 23 remains in the pixel array 10. The photodetector 30 is formed by the read circuit 5 and the pixel array 10 connected to the read circuit 5.

FIG. 7 is a perspective view of the photodetector 30 in which the pixel array 10 is flip-chip connected to the read circuit 5. The MW band absorption layer (MW) and the LW band absorption layer (LW) are laminated in the direction perpendicular to the substrate with the common contact layer 16 as the light incident side. In this schematic diagram, the block layer 14 is omitted. Each pixel 101 (see FIG. 6L) of the pixel array 10 is connected to the reading circuit 5 by one protruding electrode 34.

A bias that gives a positive potential and a bias that gives a negative potential are alternately applied from the read circuit 5 to each pixel 101 of the pixel array 10 in a time-divided manner, and at each timing, the MW band absorption layer 15 and the LW band absorption layer 13 From one side, the photocurrent is input to the read circuit 5.

The read circuit 5 has, for example, a CMOS circuit for reading pixels, and charges of light currents corresponding to each wavelength are integrated in the CMOS circuit for each wavelength to generate a voltage output signal corresponding to the incident intensity of each wavelength. To do. The reading circuit 5 sequentially scans and reads the response output from each pixel 101, and outputs an infrared image signal as a time series signal. Specifically, the infrared image signal in the MW band and the infrared image signal in the LW band are output alternately.

FIG. 8 is a schematic view of an image pickup apparatus 1 using the photodetector 30. The optical system 2 is arranged on the light incident side of the photodetector 30, and an incident infrared light image is formed on the incident surface of the pixel array 10. Infrared intensity information detected at two wavelengths for each pixel is alternately input to the reading circuit 5. The infrared intensity information obtained from the pixel array 10 represents the infrared radiation from the observation target, that is, the temperature distribution.

The output of the read circuit 5 is connected to, for example, the image processing circuit 3. The image processing circuit 3 generates an image according to the temperature distribution of the observation target. In the pixel array 10 of the photodetector 30, the dark current in the LW band absorption layer 13 is suppressed, and the crosstalk of the photoexcited carriers generated in the MW band absorption layer 15 is suppressed, so that high-quality two-wavelength infrared rays are suppressed. Images can be acquired.

The entire image pickup apparatus 1 of FIG. 8 may be housed in the cooler. The photodetector 30 may be provided with a shutter, a temperature sensor, or the like. The image processing circuit 3 may include a correction processing circuit that performs correction processing on the output signal from the photodetector 30. The correction process includes correction of variations in sensitivity and non-linearity for each pixel 101. A display recording unit may be connected to the image processing circuit 3.

Although the photodetector 30 and the image pickup apparatus 1 have been described above based on the specific configuration example, the present invention is not limited to the specific example described above. If the manufacturing process is complicated, the MW band absorption layer 15 may be pixel-separated without providing the groove 17 in the common contact layer 16. For example, in FIG. 6A, after laminating the contact layer 12, the LW band absorbing layer 13, the block layer 14, and the MW band absorbing layer 15 on the substrate 11, a plurality of grooves 17 extending in the stacking direction are formed in the MW band absorbing layer 15. , The inside is embedded with an insulating film 18. The common contact layer 16 may be formed after the surface of the insulating film 18 is flattened to expose the MW band absorption layer 15. In this case, the MW band absorption layer 15 is divided by one groove 17, and the common contact layer 16 is a continuous layer on the entire surface without being physically divided.

The impurity concentration of the contact layer 12 and the common contact layer 16 of each pixel 101 can be set to an appropriate concentration for ensuring conductivity, for example, 1 × 10 ^{18 to} 3 × 10 ¹⁸ cm ^-3 . Impurities may be added to the LW band absorption layer 13, the block layer 14, and the MW band absorption layer 15 at ^{an appropriate concentration of 5 × 10 17} cm ^{-3 or less.}

The first wavelength and the second wavelength to be detected can be designed to be desired wavelengths by adjusting the composition, period, number of repetitions, stepwise change mode of composition, etc. of the T2SL type light absorption layer. .. A compound semiconductor material other than GaSb / InAs may be used for each layer forming the epitaxial laminate. For example, even if T2SL composed of GaAs, InAs, AlAs, GaSb, InSb, AlSb, GaP, InP, AlP, and a mixed crystal material thereof is formed, such as a superlattice in which InAs and InAsSb are repeated in a short cycle. Good.

1 Image pickup device 2 Optical system 3 Image processing circuit (processing circuit)
5 Read circuit 10 Pixel array 11 Substrate 12 Contact layer (second contact layer)
13 LW band absorption layer (second light absorption layer)
14 Block layer 15 MW band absorption layer (first light absorption layer)
151 Section 16 Common contact layer (first contact layer)
17 Groove 18 Insulation film 20 Support substrate 21 Pixel separation groove 30 Photodetector 34 Projection electrode 101 pixels

Claims (9)

In a photodetector having an array of multiple pixels
A first contact layer to which a common bias voltage is applied, a first light absorption layer that responds to light of a first wavelength, and a second that responds to light of a second wavelength longer than the first wavelength. The first contact layer is located on the light incident side, and the second light absorbing layer is located on the surface side of the device.
The second light absorption layer is not physically divided and is continuous over the entire pixel region including the array of the plurality of pixels, and is separated on the surface of the second light absorption layer by a pixel separation groove. A second contact layer is placed
In the first light absorption layer, a partition corresponding to each pixel is formed by one or more grooves extending in the stacking direction.
Photodetector. The one or more grooves penetrate the first contact layer in the stacking direction and extend into the inside of the first light absorption layer, and the first contact layer is formed in the pixel region in the in-plane direction. The photodetector according to claim 1, wherein the photodetector is arranged so as to be continuous over the entire surface. The photodetector according to claim 2, wherein the compartments corresponding to the individual pixels inside the first light absorption layer are partially connected to adjacent pixels in the in-plane direction. .. The third aspect of claim 3, wherein the compartments inside the first light absorption layer corresponding to the individual pixels are connected to the adjacent pixels at the corners of the rectangular region in the in-plane direction. Light detector. A block layer arranged between the first light absorption layer and the second light absorption layer,
Have,
The block layer does not inhibit the movement of a small number of carriers generated in the first light absorption layer or the second light absorption layer by light absorption, and the first light absorption layer and the second light absorption layer. The photodetector according to any one of claims 1 to 4, wherein the movement of a large number of carriers existing in the layer is blocked. The photodetector according to claim 5, wherein at least a part of the one or more grooves penetrates the first light absorption layer in the stacking direction and extends to the block layer. The photodetector according to any one of claims 1 to 6, wherein at least a part of the one or more grooves is embedded with an insulating film. A protruding electrode electrically connected to each of the second contact layers,
A read circuit that is electrically connected to the individual pixels by the protruding electrodes.
The photodetector according to any one of claims 1 to 7, wherein the polarity of the bias voltage supplied from the reading circuit to the individual pixels is switched in a time division manner. The photodetector according to any one of claims 1 to 8.
An optical system arranged on the light incident side of the photodetector,
A processing circuit that processes the output signal of the photodetector, and
An imaging device having.

Images