Architecture of InGaAs / GaAsSb superlattices on InP substrates

Abstract

The present invention relates to an infrared detector device comprising at least one pixel including a first superlattice made up of a repeating basic group including a first layer having a first band gap and a first conduction band edge, at least one second layer having a second band gap and a second conduction band edge strictly lower than the first conduction band edge, and a third layer having a third band gap narrower than the first and second band gaps and a third conduction band edge strictly lower than the second conduction band edge, the basic groups being fabricated in a first stack configuration in the order of second layer, third layer, second layer and first layer, or in a second stack configuration in which the third layer is confined between the first and second layers.

Description

[Technical Field]

[0001] This invention relates to IR imaging (IR stands for infrared), and more particularly to radiation detectors or photodetectors manufactured from heterostructures based on III-V semiconductors. More specifically, this invention relates to an apparatus for short-wave infrared (SWIR) imaging at non-microtemperature temperatures. [Background technology]

[0002] Imagers operating in the infrared range are generally formed by assembling a matrix array containing multiple photodiode-type primary pixels that convert an incident photon flux into photoexcited charge carriers, and a readout integrated circuit (ROIC) for processing the electrical signals generated by the detector pixels.

[0003] The "quantum efficiency" of a semiconductor infrared detector refers to the ratio of the number of electron-hole pairs generated via the photoelectric effect to the number of photons passing through the pixel structure belonging to the infrared detector. Quantum efficiency depends on the absorption coefficient and the diffusion length within the pixel structure along the pixel axis. The absorption coefficient depends on the material used to create the pixel structure. The diffusion coefficient is inversely proportional to the effective mass of positive charge carriers (holes) and negative charge carriers (electrons) along the pixel axis.

[0004] Quantum efficiency is a fundamental technical characteristic that influences the electro-optical performance of infrared detectors.

[0005] This invention addresses a technical challenge in this field, namely, a method for designing a matrix array detector that improves quantum efficiency compared to prior art solutions while operating in the near-infrared range with a cutoff wavelength longer than 2 μm.

[0006] To further clarify the problems addressed by this invention, we will first describe the general structure of pixels belonging to semiconductor-type matrix array infrared detectors.

[0007] Figure 1a shows a perspective view of an example of a pixel belonging to a matrix array detector for detecting infrared radiation operating in the infrared frequency range. While this figure is limited to a single pixel Pxl for simplicity, it does not preclude the integration of pixels into a matrix array containing multiple juxtaposed pixels.

[0008] The pixels Pxl of an infrared detector are fabricated using a stack of semiconductor layers that form the pixel structure on a substrate SUB. The pixel axis Δ is an axis perpendicular to the horizontal plane (x,y) formed by the upper surface of the substrate SUB. The substrate SUB is fabricated, for example, from bulk III-V semiconductors. The selection of the substrate SUB material is important because it determines not only the technical steps in the manufacturing process of the device, but also the technical characteristics (optical characteristics, electrical characteristics, mechanical characteristics, etc.) of the matrix array detector. The pixel Pxl has the following structure, namely, starting from the substrate and in the direction of the pixel axis Δ, including a lower contact structure CONT_INF, a flat absorption structure SPA, and an upper contact structure CONT_SUP.

[0009] The lower contact structure CONT_INF is n + The lower contact structure is fabricated from a bulk semiconductor of type n, and its upper valence band is lower than the upper valence band of the flat absorption structure SPA. Preferably, the semiconductor on which the lower contact structure is fabricated is a group III-V semiconductor, which is, for example, gallium arsenide, indium arsenide, gallium nitride, gallium antimonide, boron phosphide, or ternary, quaternary, or pentary alloys thereof. The lower contact CONT_INF may also be fabricated using a heterostructure obtained by stacking multiple thin layers of a group III-V semiconductor, such as gallium arsenide, indium arsenide, gallium nitride, gallium tantimonide, boron phosphide, or ternary, quaternary, or pentary alloys thereof. The lower contact structure CONT_INF is, for example, n + It consists of a broad-bandgap superlattice doped into a mold.

[0010] The flat absorption structure SPA has a band gap that is narrower than or equal to that of the lower contact structure n -They are fabricated from doped bulk semiconductors (or superlattice heterostructures). The band diagram features (valence band, conduction band, band gap) of a flat absorption structure SPA are effective in that they are inherent in bulk materials and a combination of various thin-film features in the case of superlattices. A flat absorption structure SPA converts an incident photon flux of wavelength λ into negative electric carriers (electrons) in the (intrinsic or effective) conduction band and positive charge carriers (holes) in the (intrinsic or effective) valence band of the structure SPA. The semiconductors used to fabricate flat absorption structures SPAs (bulk or superlattice form) can be III-V semiconductors, such as gallium arsenide, indium arsenide, gallium nitride, gallium antimonide, boron phosphide, or their ternary, quaternary, or pentary alloys. The band structure of a flat absorption structure SPA is key to increasing the cutoff frequency of matrix array detectors, including pixel Pxl.

[0011] The upper contact structure CONT_SUP is P + They are fabricated from mold-doped broad-bandgap bulk semiconductors. These semiconductors are preferably III-V semiconductors, such as gallium arsenide, indium arsenide, gallium nitride, gallium antimonide, boron phosphide, or ternary, quaternary, or pentary alloys thereof. The lower contact CONT_SUP may also be fabricated using a heterostructure obtained by stacking multiple thin films of preferably III-V semiconductors, such as gallium arsenide, indium arsenide, gallium nitride, gallium antimonide, boron phosphide, or ternary, quaternary, or pentary alloys thereof. The upper contact structure CONT_SUP is, for example, n + It consists of a broad-bandgap superlattice doped into a mold.

[0012] Figure 1a shows a pixel with a structure obtained through continuous deposition and etching. The limits of the pixel in the plane (x,y) are therefore defined by its three-dimensional structure obtained through etching.

[0013] Alternatively, it is also possible to manufacture a pixel architecture that is the same along the pixel axis Δ but whose limits in the plane (x,y) are defined by the doped regions rather than the structure obtained as a result of etching. FIG. 1b shows a perspective view of a plurality of pixels belonging to a matrix array detector operating in the infrared frequency region, and the limits of the pixels are the p + -type doped regions. The characteristics of the various parts of each pixel along its axis Δ are the same as those described with respect to FIG. 1a. The main difference is the way in which the limits of the pixels of the matrix array are defined during the manufacturing process used for their manufacture (etching in 1a, doping in 1b).

[0014] FIG. 1c shows a cross-sectional view of an example of a prior art infrared detection pixel including a flat absorption structure SPA formed by a superlattice SR0.

[0015] The flat absorption structure SPA is manufactured in this figure using a periodic heterostructure that forms the superlattice SR0. In general, the superlattice SR0 is a periodic stack of a basic group G0 formed by a plurality of thin semiconductor layers. In this example, the basic group is formed by a first layer C0 with a thickness e0 and a second layer C'0 with a thickness e'0. Therefore, the period of the superlattice is equal to e0 + e'0. From a quantum point of view, if the layers C0 and C0' are thin enough (in the case of a single atomic monolayer, 0.3 nm to 10 nm), quantum confinement effects can be obtained at various junctions at the interfaces of the layers. This quantum confinement allows charge carriers (electrons and holes) to approach new energy minibands. This results in a band diagram that is different from that of the bulk semiconductor from which the superlattice SR0 was then fabricated. In the following description of this patent application, the band diagram obtained from the association of the thin layers C0=(C0,C'0) that form the superlattice SR0 is called the "effective band diagram".

[0016] The effective band diagram has - a lower end Ec eff and an effective conduction band BC eff , - an upper end Ev effThe effective valence band BV eff , - Lower end of conduction zone Ec eff and the upper edge of the valence band Ev eff The effective band gap Eg is equal to the difference between the two. eff It is defined by [the following].

[0017] Generally, the use of superlattices in the pixel's flat absorption structure makes it possible to achieve cutoff frequencies that cannot be obtained with absorption structures based on bulk semiconductors. Design choices made for flat absorption structures (SPAs) include, but are not limited to, the composition of the materials used (materials engineering), the use of bulk or superlattice structures, and the thickness of the layers used (structural design of the device). These design choices offer the potential to control the detector's absorbance, increase the cutoff frequency of matrix array detectors, improve the quantum efficiency of pixels, and limit crosstalk between adjacent pixels.

[0018] The technical solution according to the present invention relates to improving the performance of matrix array detectors in the field of infrared imaging (imaging of wavelengths from 1 μm to 70 μm, and therefore in the THz region), particularly SWIR spectrum imaging (SWIR stands for short-wave infrared, and is 1 to 2.5 μm). Specifically, prior art focal-plane matrix arrays, including a flat absorption structure made of InGaAs, have very good electro-optical performance, but are limited to a cutoff wavelength of 1.7 μm at 300 K. In certain applications, such as gas detection in aerospace engineering and detection of laser sources with emission wavelengths exceeding 1.6 μm, it is necessary to maintain a certain level of quantum efficiency and electro-optical performance while increasing the cutoff wavelength to 2 to 2.5 μm in order to ensure satisfactory detector operation.

[0019] Therefore, one technical challenge to be addressed in this field is how to design matrix array detectors that operate in the near-infrared region with cutoff wavelengths longer than 2 μm and improve the internal quantum efficiency of the detector compared to prior art solutions.

[0020] Prior art / constraints of prior art European Patent No. 3482421B1 describes a matrix array infrared image detector containing multiple pixels. The planar absorption structure of each pixel in the solution proposed in this patent is manufactured using bulk InGaAsSb on a substrate fabricated with GaSb of a specific vertical structure. Furthermore, European Patent No. 1642345B1 describes a matrix array infrared image detector containing multiple pixels. The planar absorption structure of each pixel in the solution proposed in this patent is manufactured using bulk InGaAsSb on a substrate fabricated with GaSb. The drawback of the solutions proposed in these two documents is that the process for manufacturing such structures on a GaSb substrate is complex to perform, and the immaturity of the process results in high manufacturing costs.

[0021] The paper "InGaAs / GaAsSb Type-II Superlattices for Shoft-Wavelength Infrared Detection" by J. Easley et al. exemplifies a second approach, based on a typical vertical detector architecture containing an InP substrate and InGaAs / GaAsSb, which yields a cutoff wavelength of 2.1 μm to 2.5 μm. However, existing InGaAs / GaAsSb superlattices cannot provide high-performance imagers at cutoff frequencies above 2.1 μm. Specifically, the periodic thickness of these superlattices is high, resulting in hole confinement in the GaAsSb layer, thus leading to low quantum efficiency and overall poor performance. [Prior art documents] [Patent Documents]

[0022] [Patent Document 1] European Patent No. 3482421B1 [Patent Document 2] European Patent No. 1642345B1 [Non-patent literature]

[0023] [Non-Patent Document 1] J. Easley et al., “InGaAs / GaAsSb Type-II Superlattices for Shoft-Wavelength Infrared Detection” [Overview of the project] [Problems that the invention aims to solve]

[0024] To partially overcome the limitations of existing solutions regarding increasing the SWIR region cutoff frequency to 2.5 μm at non-cryogenic temperatures while simultaneously having good quantum efficiency, the present invention provides multiple embodiments of pixel structures, including specific superlattice structures for flat absorption structures. More specifically, the present invention provides a solution compatible with InP technology, i.e., a manufacturing process in InP technology that has a higher level of technical maturity than other technologies such as GaSb technology. The present invention describes embodiments in detail with examples of the selection of inventions for superlattices, the range of compositions of semiconductor alloys forming the superlattice, and the range of layer thicknesses of the superlattice according to the present invention. A periodically repeating basic group in the superlattice according to the present invention consists of at least three thin layers, with additional low-bandgap layers inserted to obtain the following advantages compared to prior art solutions: A target cutoff wavelength longer than -2 μm can be achieved. -By increasing the potential relative to holes in the wells of a superlattice, the effective mass of positive electric carriers decreases, which induces an improvement in quantum efficiency. - It must be possible to use it in combination with mature manufacturing process technologies, such as technologies based on InP substrates. [Means for solving the problem]

[0025] One subject of the present invention is an infrared detection device including at least one pixel. The pixel includes a first superlattice comprising a stack along the stacking direction of a basic group of semiconductor layers. Each of the semiconductor layers of the basic group is arranged in a crystal lattice structure of a unit cell. The basic group is • The first band gap and 〇 Lower end of the first conduction band A first layer made of a first semiconductor having, • The second band gap, The second conduction band is strictly lower than the first conduction band. A second layer comprising at least one second semiconductor having, • A third band gap that is narrower than the first and second band gaps, The third conduction band lower edge is strictly lower than the second conduction band lower edge. A third layer consisting of a third semiconductor having, The basic group is manufactured in the following order: a first stack configuration of a second layer, a third layer, a second layer, and a first layer, or a second stack configuration in which the third layer is confined between the first and second layers.

[0026] According to one particular aspect of the present invention, the first semiconductor further has a first valence band upper edge, and the second semiconductor further has a second valence band upper edge that is strictly lower than the first valence band upper edge.

[0027] According to one particular aspect of the present invention, a first superlattice is fabricated by epitaxy on a substrate made of a fourth semiconductor arranged in a crystal lattice structure of a unit cell. The first superlattice is fabricated such that, for each semiconductor layer of the first superlattice, internal mechanical strain is generated in the lattice of the semiconductor layer to match the lattice of the crystal structure of the substrate.

[0028] According to one particular aspect of the present invention, the first, second, third, and fourth semiconductors are III-V semiconductors.

[0029] According to one particular aspect of the present invention, the fourth semiconductor is indium phosphide (InP).

[0030] According to one particular aspect of the present invention, the composition of the material used to manufacture the semiconductor layer of the basic group is selected such that the band diagrams of the conduction and valence bands in the stacking direction of the first superlattice have an effective band gap, an effective valence band upper limit, and an effective conduction band lower limit. The effective band gap is 400 meV to 750 meV.

[0031] According to one particular aspect of the present invention, the effective mass of positive charge carriers in the stacking direction of the superlattice is three times smaller than the mass of free electrons.

[0032] According to one particular aspect of the present invention, the third semiconductor is a binary composite material InAs.

[0033] According to one particular aspect of the present invention, the second semiconductor is a ternary alloy In x Ga 1-x As is alloy In x Ga 1-x This is the mole fraction of indium in As.

[0034] According to one particular aspect of the present invention, the mole fraction x of indium In in the second semiconductor is less than 0.55.

[0035] According to one particular aspect of the present invention, the first semiconductor is a ternary alloy GaAs y S 1-y And y is the alloy GaAs y S 1-y This is the mole fraction of arsenic inside.

[0036] According to one particular aspect of the present invention, the mole fraction y of arsenic As in the first semiconductor is less than 0.55.

[0037] According to one particular aspect of the present invention, the ratio of the sum of the thicknesses of the layers of the basic group, weighted by the amplitude of the strain generated in each layer, to the total thickness of the basic group is lower than or equal to a predetermined value.

[0038] According to one particular aspect of the present invention, Inx Ga 1-x The strain occurring in the lattice of the As layer is tensile strain, GaAs y S 1-y The strain that occurs in the lattice of the layer is compressive strain.

[0039] According to one particular aspect of the present invention, the amplitude of strain generated in the lattice of any one layer of the basic group is smaller than the critical strain at which dislocations occur.

[0040] According to one particular aspect of the present invention, the thickness of the base group layer is 0.3 nm to 10 nm.

[0041] According to one particular aspect of the present invention, the pixels are arranged along the stacking direction (Z), ·substrate • The upper end of the fifth valence band is strictly lower than the upper end of the effective valence band of the first superlattice, ○ A fifth band gap that is wider than or equal to the effective band gap of the first superlattice. n has + A lower contact layer consisting of a fifth semiconductor doped into the mold, • A flat absorption structure comprising at least one first superlattice, wherein the layer of the first superlattice is n-type doped. • The upper end of the sixth valence band is strictly lower than the upper end of the effective valence band of the first superlattice, The sixth conduction band lower end is strictly higher than the lower end of the effective conduction band of the first superlattice. Having p + Upper contact layer consisting of a sixth semiconductor doped into the mold It includes them in this order.

[0042] According to one particular aspect of the present invention, the pixels are arranged in the stacking direction. • Circuit board (SUB) • The upper end of the second effective valence band is strictly lower than the upper end of the first effective valence band of the first superlattice, ○ A second effective band gap that is wider than or equal to the effective band gap of the first superlattice. n has +The lower contact is manufactured using a second superlattice doped into the mold. • A flat absorption structure comprising at least one superlattice, wherein the layer of the superlattice is n-type doped. • The upper end of the third effective valence band is strictly lower than the upper end of the first effective valence band of the first superlattice, The lower end of the third effective conduction band is strictly higher than the upper end of the first superlattice's effective conduction band. Upper contact manufactured using a third superlattice having It includes them in this order.

[0043] According to one particular aspect of the present invention, the flat absorption structure is • On the other hand, the lower end of the effective conduction band of the first superlattice, On the other hand, the lower end of the sixth conduction band or the lower end of the third effective conduction band The lower end of the seventh conduction band between, • On the other hand, the upper end of the effective valence band of the first superlattice and On the other hand, the upper end of the sixth valence band or the upper end of the third effective valence band The upper end of the seventh valence band between It further includes a transition layer made of a seventh semiconductor doped to the n-type having, The transition layer is confined between the superlattice and the upper contact layer.

[0044] Other features and advantages of the present invention will become clearer from the following description with respect to the attached drawings. [Brief explanation of the drawing]

[0045] [Figure 1a] A perspective view of the first example of a pixel belonging to a matrix array detector in the infrared frequency range is shown. [Figure 1b] A perspective view of a second example of a pixel belonging to a matrix array detector in the infrared frequency range is shown. [Figure 1c] A cross-sectional view of an example of an infrared detection pixel using prior art, which includes a flat absorption structure formed by a superlattice, is shown. [Figure 1d]Figure 1c shows an electronic diagram of the superlattice along the pixel axis, and the superlattice is obtained by epitaxial growth that is lattice-matched with the substrate. [Figure 1e] Figure 1c shows the potential diagram of the superlattice along the pixel axis, and the superlattice is obtained by strain-compensated epitaxial growth on a substrate. [Figure 2a] This shows a cross-sectional view of an infrared detection pixel including a superlattice according to the first embodiment of the present invention. [Figure 2b] This shows the potential diagram along the pixel axis of a superlattice according to the first embodiment of the present invention. [Figure 3a] This shows a cross-sectional view of an infrared detection pixel including a superlattice according to a second embodiment of the present invention. [Figure 3b] The potential diagram along the pixel axis of a superlattice according to a second embodiment of the present invention is shown. [Figure 4] The curves of absorption simulation results at a cutoff wavelength equal to 2.3 μm are shown for a two-layer superlattice and a superlattice according to the second embodiment of the present invention. [Figure 5a] A cross-sectional view of a first example of a pixel containing a flat absorption structure according to any one embodiment of the present invention is shown. [Figure 5b] Figure 5a shows a band diagram along the pixel axis. [Figure 5c] A cross-sectional view of a second example of a pixel containing a flat absorption structure according to any one embodiment of the present invention is shown. [Figure 5d] Figure 5c shows the band diagram along the pixel axis. [Figure 6] This shows a cross-sectional view of a device for detecting infrared radiation containing multiple pixels according to the present invention. [Modes for carrying out the invention]

[0046] First, we will explain the concept of growing a thin-layer lattice that matches the substrate. Generally, solid-state bulk semiconductors are organized into a crystalline structure obtained through the spatially periodic repetition of a basic unit cell composed of atoms of the material. For any material, its mechanical, physical, and electrical properties are all determined, in particular, by the structure of the unit cell and the resulting crystal lattice. In the case of semiconductor alloys, the selection of mole fractions of the various materials constituting the alloy defines the crystalline structure and geometric parameters of the resulting alloy's crystal lattice.

[0047] In the case of infrared detectors based on semiconductor thin layers, the layers forming the superlattice are fabricated via epitaxial growth on the substrate. Molecular beam epitaxy or organometallic vapor phase epitaxy can be used. In semiconductor alloys, the parameters of the crystal lattice of the deposited layer can be controlled by selecting the mole fraction used in the epitaxial growth steps of each component of the alloy. Therefore, through a reasonable selection of mole fractions, a thin-layer lattice that matches the crystal structure of the substrate can be deposited by epitaxy. This is called homoepitaxy.

[0048] When the mole fraction used in the alloy deposited on a substrate differs from the value corresponding to lattice matching, it is possible to grow a layer of semiconductor alloy several nanometers thick on the substrate; this is called heteroepitaxy. Specifically, the crystal lattice of the deposited layer undergoes internal mechanical strain during epitaxial growth, matching the crystal lattice of the substrate. This internal mechanical strain can be tensile or compressive strain, depending on the mole fraction selected for the alloy. At the crystal lattice level, the internal mechanical strain occurring in the unit cell of the deposited layer results in asymmetry of the unit cell and modulation of interatomic electromagnetic forces. From an electrical standpoint, this alters the band structure of the deposited layer as described below.

[0049] The following explanation of Figures 1d and 1e will detail the limitations of using two-layer superlattices for applications requiring cutoff frequencies above 2.1 μm. Specifically, while near-infrared cutoff frequencies above 2.1 μm can be achieved by fabricating a flat absorption structure with a superlattice composed of two layers, the electro-optic performance is poor. The quantum efficiency of this type of structure is low, and the mobility of positive charge carriers along the pixel axis is reduced. This degradation is due to the large effective mass of the success. This increase in effective mass is explained by the very high potential barrier observed due to holes in the superlattice along the pixel axis. All of these observations and results are discussed in the explanation of Figures 1d and 1e.

[0050] Figure 1d shows the potential diagram of the pixel axis in the z direction of the superlattice SR0 in Figure 1c, and the superlattice is obtained by lattice-matched epitaxial growth with the substrate.

[0051] The potential diagram in Figure 1d shows the lattice of the superlattice matched with the InP substrate and GaAs y0 S 1-y0 The first layer C1 and In consist of x0 Ga 1-x0 Corresponding to the basic group G0, which consists of a second layer C2 made of As, x0 = 0.53 is consistent with the InP substrate. x0 Ga 1-x0 This is the mole fraction of indium in the As alloy lattice, where y=0.52 is consistent with GaAs in the InP substrate. y0 S 1-y0 This is the mole fraction of arsenic in the alloy lattice. To achieve a cutoff frequency of 2.5 μm, GaAs y0 S 1-y0 The thickness e1 of layer C1, which consists of In x0 Ga 1-x0 The thickness e2 of layer C2, which is made of As, is equal to 7 nm.

[0052] The band diagram shows the valence electron and conduction band structure of various layers along the z-direction of the superlattice axis Δ, and therefore through the periodic stacking of the superlattice layers. In the example shown in the figure, the first layer C1 of the basic group G0 is a ternary alloy GaAs having a first band gap Eg1.y0 S 1-y0 It is manufactured using a ternary alloy, which has a first conduction band lower edge Ec1 and a first valence band upper edge Ev1. The second layer C2 is a ternary alloy In x0 Ga 1-x0 It is fabricated using As, which has a second band gap Eg2, a second conduction band lower edge Ec2, and a second valence band upper edge Ev2. The material constituting the superlattice is selected such that Ec1 > Ec2 and Ev1 > Ev2, and a type II band alignment is obtained. From the perspective of positive charge carriers (holes), this is a matter of alternation between the potential barrier in the second layer C2 and the potential well in the first layer C1. From the perspective of negative charge carriers (electrons), this is a matter of alternation between the potential barrier in the first layer C1 and the potential well in the second layer C2.

[0053] As mentioned above, the quantum coupling between the various layers of the superlattice allows for the effective lower edge Ec eff Effective conduction band and effective upper end Ev eff An effective valence band with the following properties can be created. Superlattice Eg eff =Ec eff -Ev eff The effective band gap is equal to 0.488 eV. Therefore, by the Planck-Einstein relationship, the cutoff frequency λc = hc / Eg eff This is approximately 2.5 μm, where h is Planck's constant and c is the speed of light in free space.

[0054] Therefore, the basic group G0 = GaAs y0 S 1-y0 / In x0 Ga 1-x0 A superlattice with a thickness of e1=e2=7nm, matched to an InP substrate containing As, can be used to obtain a cutoff wavelength of 2.5 μm. In this case, bandgap overlap refers to the difference between the upper edge Ev1 of the first valence band and the lower edge Ec2 of the second conduction band. The selection of a thickness of 7 nm is necessary to obtain a bandgap overlap of 0.380 eV.

[0055] However, the holes in the potential wells within the first layer C1 are large in combinations of layer thickness (7nm / 7nm) and Eveff As a result of the large potential difference between Ev2 and Ev2 (estimated to be 0.350 eV), the holes receive a high potential. The effect of the potential on the holes themselves manifests as an increase in the effective mass of the heavy holes in the effective valence band, whose mass is calculated to be 71 times the mass of the free electrons represented by m0 in this combination. This corresponds to a decrease of more than 10 in the coefficient of optical efficiency for a flat absorption structure with acceptable optical performance. In general, infrared detectors provide acceptable electro-optical performance when the effective mass of the holes is three times smaller than the mass of the free electrons m0.

[0056] Table 1 shows G0=GaAs that matches the InP substrate to achieve cutoff wavelengths of 2.1 μm, 2.3 μm, and 2.5 μm. y0 S 1-y0 / In x0 Ga 1-x0 This document summarizes the results obtained for various combinations of superlattice SR0 having an As lattice.

[0057] [Table 1]

[0058] From the results in the table, G0=GaAs is consistent with InP substrates. y0 S 1-y0 / In x0 Ga 1-x0 While superlattices SR0 with As lattices can achieve cutoff wavelengths exceeding 2.1 μm, their electro-optical performance is unsatisfactory, and it has been proven that they have low quantum efficiency (very large effective hole mass along the pixel axis). Therefore, a two-layer lattice-matched superlattice SR0 with a cutoff frequency of 2.1 μm can obtain an effective hole mass equal to 9.8 times the mass of free electrons m0. A two-layer lattice-matched superlattice SR0 with a cutoff frequency of 2.3 μm can obtain an effective hole mass equal to 52 times the mass of free electrons m0. A two-layer lattice-matched superlattice SR0 with a cutoff frequency of 2.5 μm can obtain an effective hole mass equal to 71.8 times the mass of free electrons m0.

[0059] Figure 1e shows the dislocation diagram of the pixel axis in the z direction of the superlattice SR0 in Figure 1c, and the superlattice is obtained by strain-compensated epitaxial growth on a substrate.

[0060] Figure 1e shows a superlattice structure SR'0 that is the same as Figure 1c, but with a different mole fraction of the alloy constituting the superlattice layers. This change in composition is achieved by controlling the balance of the alloy components during the epitaxial growth phase, thereby depositing layers that are not lattice-matched with the InP substrate. As already explained, this lattice mismatch generates internal mechanical strains in the crystal lattices of layers C1 and C2, which align with the crystal lattice of the substrate SUB, resulting in the new band diagram shown in Figure 1d. In this case, the superlattice SR'0 is called a "strain-compensated" superlattice.

[0061] The band diagram in Figure 1e shows GaAs on an InP substrate. y’0 S 1-y’0 The first layer C1 and In x’0 Ga 1-x’0 This corresponds to a strain-compensated superlattice having a basic group G'0 composed of a second layer C2 made of As, where x'0 = 0.25 is the strain-compensated In on the InP substrate. x’0 Ga 1-x’0 This is the mole fraction of indium in the As alloy, and Y'0 = 0.2 is the strain compensation GaAs on the InP substrate. y’0 S 1-y’0 This is the mole fraction of arsenic within. To achieve a cutoff frequency of 2.5 μm, GaAs y’0 S 1-y’0 The thickness e1 of layer C1 consisting of can be reduced to 2.9 nm, In x’0 Ga 1-x’0 The thickness e2 of the As layer C2 can be reduced to 2.9 nm.

[0062] GaAs y’0 S 1-y’0 Compressive strain occurs in the lattice of layer C1, which consists of In x’0 Ga 1-x’0Tensile strain occurs in the lattice of layer C2, which is made of As. The internal strain applied to the lattice of the superlattice layer must not exceed the dislocation limit. The total strain occurring in the basic group G'0 is zero.

[0063] The band diagram in Figure 1e shows the valence and conduction band structure along the same axis Δ as in Figure 1d. As previously mentioned, this is a problem with the band diagram of a type II band alignment with different upper and lower bands. Specifically, a quantum effect known as "degeneracy lift" occurs as a result of strain induced by internal strain in the lattice. Degeneracy lift is the separation of energy levels occupied by heavy holes from those occupied by light holes. Therefore, in the first layer C1 and the second layer C2, a difference is observed between the upper valence band occupied by heavy holes, shown as Ev1-HH in the first layer and Ev2-HH in the second layer, and the upper valence band occupied by light holes, shown as Ev1-LH in the first layer and Ev2-LH in the second layer. In addition, the positions of the conduction and valence bands change under the influence of strain. When it is a tensile strain problem, the band gap decreases (the lower conduction band decreases and the upper valence band increases). When this is a compression strain problem, the band gap increases (the lower end of the conduction band increases, and the upper end of the valence band increases). This makes it possible to reduce the band gap overlap to 240 meV, and as a result, the thickness of the first layer C1 and the second layer C2 can be reduced to 2.9 nm.

[0064] The quantum coupling between various layers of a superlattice is at the lower end of the effective conduction band Ec eff The effective conduction band and the upper end of the effective valence band Ev eff This leads to the creation of an effective valence band with the effective superlattice band gap Eg. eff =Ec eff -Ev eff This is equal to 0.492 eV. Therefore, by the Planck-Einstein relationship, the cutoff frequency λc = hc / Eg eff This is approximately 2.5 μm, where h is Planck's constant and c is the speed of light in free space.

[0065] Therefore, on the InP substrate, the basic group G0 = GaAsy’0 Sb 1-y’0 / In x’0 Ga 1-x’0 A strain-compensated superlattice with e1 = e2 = 2.9 nm having As can obtain a cut-off wavelength of 2.5 μm. However, holes in the potential wells within C1 receive a high potential as a result of the thick layer thickness combination (2.9 nm / 2.9 nm) and the high potential difference (evaluated to be 0.520 eV) between Ev eff and Ev2. The effect of the potential received by the holes manifests itself as an increase in the effective mass of the heavy holes in the valence band, and the mass is calculated to be 24 times the mass of free electrons for this combination.

[0066] Table 2 summarizes the results obtained for various combinations of the strain-compensated superlattice SR’0 having a G’0 = GaAs y’0 Sb 1-y’0 / In x’0 Ga 1-x’0 As lattice on an InP substrate to achieve cut-off wavelengths of 2.1 μm, 2.3 μm, and 2.5 μm at an operating temperature of 300 K.

[0067]

Table 2

[0068] From the results in the table, G’0 = GaAs y’0 Sb 1-y’0 / In x’0 Ga 1-x’0Distortion-compensated superlattice SR'0 with an As lattice can achieve cutoff wavelengths greater than 2.1 μm, but its electro-optical performance is unsatisfactory, and it has been proven that its quantum efficiency is low (the effective hole mass along the pixel axis is very large). Therefore, a two-layer distortion-compensated superlattice SR'0 with a cutoff frequency of 2.1 μm can obtain an effective hole mass equal to 1.2 times the free electron mass m0. A two-layer distortion-compensated superlattice SR'0 with a cutoff frequency of 2.3 μm can obtain an effective hole mass equal to 6 times the free electron mass m0. A two-layer distortion-compensated superlattice SR'0 with a cutoff frequency of 2.5 μm can obtain an effective hole mass equal to 24 times the free electron mass m0. Improvements are observed compared to lattice-matched superlattice SR0, but these improvements are still insufficient and are most pronounced for cutoff wavelengths of 2.3 μm and 2.5 μm. The above description demonstrated the limitations of using a two-layer superlattice to fabricate a flat absorption structure (SPA) of pixel Pxl belonging to an infrared detector operating at the upper cutoff frequency of SWIR at non-microtemperatures. The present invention provides a novel superlattice structure that can overcome the limitations of two-layer superlattice structures with respect to the 2.1 μm to 2.5 μm cutoff frequency with high quantum efficiency achieved by reducing the effective hole mass.

[0069] Figure 2a shows a cross-sectional view of a pixel in an infrared detector including a superlattice according to the first embodiment of the present invention.

[0070] Figure 2b shows the potential diagram of the pixel axis in the z-direction within a superlattice according to the first embodiment of the present invention, where the superlattice is obtained by strain-compensated epitaxial growth on a substrate SUB.

[0071] Pixel Pxl includes a superlattice SR1 containing a stack of semiconductor layers of basic group G1 along the pixel axis Δ. The semiconductor layer of basic group G1 is strain-compensated. The periodically repeating basic group G1 is -A first layer C1 consisting of a first semiconductor SC1, -The second layer C2 consists of the second semiconductor SC2, -The third layer C3 consists of the third semiconductor SC3. - Layer C'2 has the same composition as layer C2, but with a different thickness e'2. This includes in this order. The first embodiment therefore corresponds to the superlattice SR1 based on the following basic group, namely G1=(C1 / C2 / C3 / C'2).

[0072] The first semiconductor SC1 of layer C1 has a first band gap Eg1, a first valence band upper limit Ev1, and a first conduction band lower limit Ec1.

[0073] The second semiconductor SC2 of layer C2 has a second band gap Eg2, a second valence band upper limit Ev2 that is lower than the first valence band upper limit Ev1, and a second conduction band lower limit Ec2 that is lower than the first conduction band lower limit Ec1.

[0074] The third semiconductor layer C3, SC3, has a third bandgap Eg3 that is strictly narrower than the first and second bandgaps Eg1 and Eg2, and a third conduction band lower edge Ec3 that is lower than the second conduction band lower edge Ec2. Generally, by inserting a third layer C3 of the narrow-bandgap semiconductor SC3 into the superlattice, the overlap REG of the bandgap can be reduced. Reducing the overlap REG of the bandgap allows for the selection of a thinner overall thickness of the fundamental group G1 in the region of the target cutoff wavelength. By thinning the thickness of the fundamental group G1, the potential barrier experienced by holes confined in the potential wells is reduced, thus reducing their effective mass and improving quantum efficiency and the mobility of positive charge carriers.

[0075] Compressive strain occurs in the first layer C1, tensile strain in the second layer C2, and compressive strain in the third layer C3. The total strain experienced by the basic group G1 = (C1 / C2 / C3 / C'2) is zero.

[0076] For example, without loss of generality, it is possible to manufacture a superlattice SR1 according to the first embodiment via a stack of multiple semiconductor layers, preferably III-V semiconductor layers such as gallium arsenide, indium arsenide, gallium nitride, barium antimonide, boron phosphide, or ternary, quaternary, or pentary alloys thereof. To illustrate this embodiment, Figure 2b specifically shows a band diagram obtained from a superlattice SR1 having the following dimensions, properties, and plasticity.

[0077] In the example shown in Figure 2b, the target cutoff wavelength is 2.5 μm. To obtain this cutoff wavelength, the materials constituting the superlattice according to the first embodiment are as follows: The first layer C1 is a ternary alloy GaAs with a mole fraction of arsenic y1 = 0.35. y1 S 1-y1 Using this method, it was manufactured to a thickness e1 = 1.2 nm, and the second layer C2 is a ternary alloy In with a mole fraction of indium x1 = 0.25. x1 Ga 1-x1 Using As, the first instance of the basic group (C2) is fabricated with a thickness e2 = 0.6 nm, and the second instance of the basic group G1 (C2') is fabricated with a thickness e'2 = 0.9 nm. The third layer C3, inserted between the two layers (C2, C'2), is fabricated using the narrow-bandgap binary composite III-V semiconductor, indium arsenide (InAs). The thickness of the third layer C3 is denoted by e3 and is equal to 0.7 nm.

[0078] The effect of combining the insertion of layer C3, made of InAs, with strain-compensated growth is to separate the heavy-hole and light-hole levels, which can be seen in the band diagram in Figure 2b. The resulting effect is a reduction in the overlapping band gap REG and therefore the ability to use thicknesses e1, e2, and e3 on the order of 1 nm for each layer. In this case, the overlapping band gap is effective and corresponds to the combination of overlapping band gaps of the various component layers of the superlattice. Therefore, 0.493 eV (λc = hc / Eg eff Effective band gap Eg (approximately 2.5 μm) eff =Ec eff -Ev effThe pixels have a total thickness of 2.2 nm, and Ev eff This is obtained in the basic group G1, where the potential difference between Ev2 and Ev2 is low (evaluated as 0.360 eV). The reduction in these two properties allows the potential barrier of the wells (layer C1) in the superlattice SR1 to be lowered with respect to holes. The effective mass of holes is reduced to 2.8 times the mass of free electrons with respect to the 2.5 μm cutoff wavelength.

[0079] Table 3 shows the G1=GaAs ratio on an InP substrate for achieving cutoff wavelengths of 2.1 μm, 2.3 μm, and 2.5 μm at an operating temperature of 300 K. y1 S 1-y1 / In x1 Ga 1-x1 As / InAs / In x1 Ga 1-x1 This report summarizes the results obtained for various combinations of the strain-compensated superlattice SR1 according to the first embodiment having As.

[0080] [Table 3]

[0081] From the results in the table, G1 = GaAs on the InP substrate y1 S 1-y1 / In x1 Ga 1-x1 As / InAs / In x1 Ga 1-x1Distortion-compensated superlattice SR1 with As can achieve cutoff wavelengths exceeding 2.1 μm, demonstrating improved electro-optic performance and higher quantum efficiency (lower effective hole mass along the pixel axis compared to results obtained with flat absorption structures formed using bilayer superlattices). Therefore, a superlattice SR1 according to the first embodiment with a cutoff frequency of 2.1 μm yields an effective hole mass equal to 0.89 times the free electron mass m0 (compared to 2.1 × m0 for SR'0 and 9.8 × m0 for SR0). A superlattice SR1 according to the first embodiment with a cutoff frequency of 2.3 μm yields an effective hole mass equal to 1.2 times the free electron mass m0 (compared to 6 × m0 for SR'0 and 52 × m0 for SR0). In the first embodiment of the superlattice SR1 with a cutoff frequency of 2.5 μm, an effective hole mass equal to 2.8 times the free electron mass m0 is obtained (compared to 24 × m0 for SR'0 and 71.8 × m0 for SR0). In general, the effective mass of holes in the potential well can be reduced by inserting a thin layer made of a narrow bandgap semiconductor into the basic group of the superlattice between two layers C2. The obtained value is quite close to that of the bulk material, thereby enabling good internal quantum efficiency for wavelengths of 2.1 μm to 2.5 μm at non-microtemperatures. The basic group G1=(C1 / C2 / C3 / C2) of the superlattice SR1 in the first embodiment is GaAs y S 1-y It consists of, and the mole fraction y of arsenic is y max The first layer C1 is lower than =0.55, and In x Ga 1-x It consists of As, and the mole fraction x of indium is x max It can be manufactured using a second layer C2 with a value lower than =0.55.

[0082] Maximum mole fraction limit y max =0.55 corresponds to lattice matching on the InP substrate. y S 1-y The mole fraction y of arsenic is y match This is because the strain reverses from compression to tension at a value exceeding =0.52. This reversal of strain from compression to tension increases the overlap of the band gap, and therefore the cutoff wavelength decreases. This limit is ymax It is acceptable for the value to exceed 0.55 because the inversion effect is small and operation at a cutoff frequency of 2.1 μm to 2.5 μm at non-microtemperature temperatures is acceptable. max This is because it is possible even if the value is less than 0.55.

[0083] Maximum mole fraction limit x max =0.55 corresponds to lattice matching on the InP substrate. x Ga 1-x x is the mole fraction of indium in As. match This is because the strain reverses from tension to compression at a value exceeding =0.53. This reversal of strain from tension to compression increases the overlap of the band gap, and therefore the cutoff wavelength decreases. This limit is x max It is acceptable for the value to exceed 0.55 if the inversion effect is small and the operation at a cutoff frequency of 2.1 μm to 2.5 μm at non-microtemperature temperatures is x max This is because it is possible even if the value is less than 0.55.

[0084] By selecting these mole fractions of the alloys used to fabricate the superlattice, the lattice of the crystalline structure of the superlattice layers can be strained such that the strain remains below the limiting strain that would cause dislocations, which would weaken the mechanical structure of the apparatus, while reducing the potential barrier as seen from holes. The strains and thicknesses of each layer C1, C2, and C3 of the basic group G1 must conform to the following inequality in order to avoid dislocation effects:

number

[0085] In the formula, e i This is the i-th layer C of the basic group G1. i thickness, c i This is the i-th layer C of the basic group G1. i This is the distortion that occurs in the grid.

[0086] Figure 3a shows a cross-sectional view of an infrared detection pixel including a superlattice according to a second embodiment of the present invention.

[0087] Figure 3b shows the potential diagram of the pixel axis in the z direction of a superlattice according to a second embodiment of the present invention, the superlattice being obtained on a substrate SUB by strain-compensated epitaxial growth.

[0088] Pixel Pxl includes a superlattice SR2 containing a stack of semiconductor layers along the pixel axis Δ of a basic group G2. The semiconductor layer of the basic group G2 is strain-compensated. The periodically repeating basic group G2 is -A second layer C2 consisting of a second semiconductor SC2 having the same characteristics (in terms of its unique band distribution) as the second semiconductor used in the first embodiment described above, -A third layer C3 consisting of a third semiconductor SC3 having the same characteristics (in terms of its intrinsic band distribution) as the third semiconductor used in the first embodiment described above, - A first layer C1 made of a first semiconductor SC1 having the same characteristics (in terms of its intrinsic band distribution) as the first semiconductor used in the first embodiment described above. The second embodiment therefore corresponds to the following basic group, namely, the superlattice SR2 based on G2=(C2 / C3 / C1).

[0089] Compressive strain occurs in the first layer C1, tensile strain in the second layer C2, and compressive strain in the third layer C3. The total strain in the basic group G2 = (C2 / C3 / C1) is zero.

[0090] For example, without loss of generality, the superlattice SR2 according to the second embodiment can be manufactured via a stack of multiple semiconductor thin layers, preferably layers of III-V semiconductors such as gallium arsenide, indium arsenide, gallium nitride, gallium antimonide, boron phosphide, or ternary, quaternary, or pentary alloys thereof. To illustrate this embodiment, Figure 3b specifically shows the band diagram obtained from a superlattice SR2 having the following dimensions, properties, and plasticity.

[0091] In the example in Figure 3b, the target cutoff wavelength is 2.5 μm. The first layer C1 is a ternary alloy GaAs with an arsenic mole fraction y1 = 0.35.y1 S 1-y1 It is manufactured using a material with a thickness e1 = 1.2 nm. The second layer C2 is a ternary alloy In with a mole fraction x1 = 0.25 of indium. x1 Ga 1-x1 The layers are fabricated using As to a thickness of e2 = 1.5 nm. The third layer C3, inserted between layers C2 and C1, is fabricated using indium arsenide (InAs), a narrow-bandgap binary composite material III-V semiconductor. The thickness of the third layer C3 is denoted by e3 and is equal to 0.7 nm.

[0092] As mentioned above, the effect of the combination of inserting layer C3 made of InAs and strain-compensated growth that separates heavy-hole and light-hole levels is seen in the band diagram of Figure 3b. The resulting effect is a reduction in band gap overlap and therefore the ability to use thicknesses e1, e2, and e3 on the order of 1 nm for each layer. In this case, band gap overlap is effective, and this corresponds to the combination of band gap overlaps of the various component layers of the superlattice. Therefore, the effective band gap Eg eff =Ec eff -Ev eff This is equal to 0.49 eV (λc = hc / Eg eff A pixel of approximately 2.5 μm in thickness has a total thickness of 2.2 nm, and Ev eff This is obtained in the basic group G2, where the potential difference between Ev2 and Ev2 is small (evaluated as 0.360 eV). The reduction of these two properties allows the potential barrier of the wells (layer C1) of the superlattice SR2 to be lowered from the perspective of holes. The effective mass of holes is reduced to 2.8 times the mass of free electrons when the cutoff wavelength is 2.5 μm.

[0093] According to the second embodiment of the present invention, G2=In on an InP substrate x1 Ga 1-x1 As / InAs / GaAs y1 S 1-y1 The distortion-compensated superlattice SR2, which has this feature, can achieve a cutoff wavelength greater than 2.1 μm, improves electro-optic performance, and has higher quantum efficiency (the effective hole mass along the pixel axis is lower than that obtained with a flat absorption structure formed using a two-layer superlattice).

[0094] More generally, the effective mass of holes in the potential wells can be reduced by inserting thin layers of narrow-bandgap semiconductors into the basic group of the superlattice. The resulting values ​​are quite close to those of bulk materials, thereby enabling good internal quantum efficiency at non-microtemperatures for wavelengths of 2.1 μm to 2.5 μm. In the second embodiment, the basic group G2=(C2 / C3 / C1) of the superlattice SR2 is GaAs y S 1-y It consists of a first layer C1 in which the mole fraction y of arsenic is lower than 0.55, and In x Ga 1-x It consists of As, and the mole fraction x of indium is x max It can be manufactured using a second layer C2 with a value lower than =0.55. The previous explanation regarding the selection made with respect to mole fractions also applies to this architecture. By selecting these mole fractions for the alloy used to manufacture the superlattice, it is possible to apply strains to the crystalline lattice of the layers of the superlattice such that they remain below the critical strain at which dislocations occur, thereby weakening the mechanical structure of the apparatus, while reducing the potential barrier as seen from holes. The strains and thicknesses of each layer C1, C2, and C3 of the basic group G2 must satisfy the following inequality in order to avoid dislocation effects:

number

[0095] In the formula, e i This is the i-th layer C of the basic group G2. i thickness, c i This is the i-th layer C of the basic group G2. i This is the distortion that occurs in the grid.

[0096] Generally, during the epitaxial growth phase, the composition of the material used to fabricate the semiconductor layer of the basic group G1 (or G2) is such that the effective band gap Eg of the superlattice SR1 (or SR2) is such that the conduction and valence bands are in the stacking direction. effHowever, to obtain a detection device with a cutoff frequency (λc) of 1.6 μm to 3.1 μm, the voltage is selected to be between 400 meV and 750 meV.

[0097] Figure 4 shows the absorption simulation results curves at a cutoff wavelength of 2.3 μm for the superlattice of the prior art and the superlattice according to the second embodiment of the present invention.

[0098] An absorption energy of 0.54 eV corresponds to a cutoff wavelength of 2.3 μm. Here, the advantages of the flat absorption structure SPA based on the superlattice according to the present invention over the solution using a superlattice with two basic groups are demonstrated. For example, at an incident wavelength of 2 μm (equivalent to 0.6 eV), the absorption of the incident wave shown by the solution according to the present invention is three times greater than that of a superlattice with two basic groups.

[0099] Figure 5a shows a cross-sectional view of a first example of a pixel Pxl containing a flat absorption structure according to the present invention. Figure 5b shows a band diagram along the axis of the pixel Pxl in Figure 5a.

[0100] Pixel Pxl is located in the stacking direction z (or along the pixel axis) on a substrate SUB, n having the characteristics described in detail below. + The present invention includes, in this order, a lower contact layer CONT_INF made of a fifth semiconductor SC5 doped in the n-type, a flat absorption structure SPA having n-type doped layers C1, C2, and C3, and an upper contact layer CONT_SUP made of a sixth semiconductor SC6 doped in the p-type.

[0101] The fifth semiconductor SC5, as shown in Figure 5b, is the upper edge Ev of the effective valence band of the superlattice SR1 (or SR2) according to the present invention. eff It has a more strictly lower fifth valence band upper limit Ev5.

[0102] The fifth semiconductor SC5, as shown in Figure 5b, has an effective bandgap Eg of the superlattice SR1 (or SR2) according to the present invention. eff It also has a wider or equal band gap Eg5.

[0103] As a non-limiting example, the following materials can be used to manufacture the lower contact: n + Type-doped InGaAs or n + Type-doped InAlAs or n + InP doped into the mold.

[0104] Alternatively, set the lower contact CONT_INF to N + It can be manufactured using a superlattice SR_INF having a stack of doped layers in the mold. The superlattice used for the lower contact CONT_INF is the upper end of the effective valence band Ev of the superlattice SR1 (or SR2) according to the present invention. eff A more strictly defined lower effective valence band upper limit Ev eff_inf The superlattice used for the lower contact CONT_INF has the effective band gap Eg of the superlattice SR1 (or SR2) according to the present invention, as shown in Figure 5b. eff A lower or equal effective band gap Eg eff_inf It also possesses.

[0105] The sixth semiconductor SC6 used to manufacture the upper contact CONT_SUP is, as shown in Figure 5b, the upper edge Ev of the effective valence band of the superlattice SR1 (or SR2) according to the present invention. eff It has a more strictly lower upper limit of the sixth valence band, Ev6.

[0106] The sixth semiconductor SC6, as shown in Figure 5b, is the lower end of the effective conduction band Ec of the superlattice SR1 (or SR2) according to the present invention. eff It also has a more strictly defined sixth conduction band lower limit Ec6.

[0107] As a non-limiting example, the following materials can be used to manufacture the upper contact: p + InGaAs or p doped into the type + InAlAs or p doped into the type + InP doped into the mold.

[0108] Alternatively, the upper contact CONT_SUP can be fabricated using a superlattice SR_SUP having a stack of p-type doped layers. The superlattice used for the upper contact CONT_SUP is the upper end of the effective valence band Ev of the superlattice SR1 (or SR2) according to the present invention. eff A more strictly defined lower effective valence band upper limit Ev eff_sup The superlattice used for the upper contact CONT_SUP is, as shown in Figure 5b, the effective conduction band lower end Ec of the superlattice SR1 (or SR2) according to the present invention. eff More precisely, the lower end of the effective conduction band Ec eff_sup It also possesses.

[0109] If the pixel limits are defined by a doped region as shown in Figure 1b, the p-type dopant used to fabricate the upper contact CONT_SUP may partially diffuse into the volume of the absorption region at the interface with the sixth semiconductor SC6 (or superlattice SR_SUP). This region is therefore considered to form part of the upper contact CONT_SUP.

[0110] Figure 5c shows a cross-sectional view of a second example of a pixel Pxl containing a flat absorption structure according to the present invention. Figure 5d shows a band diagram along the axis of the pixel Pxl in Figure 5c.

[0111] The pixel Pxl in Figure 5c has the same properties as the pixel shown in Figure 5a, as described above. The stack of pixel Pxl in Figure 5c differs in the insertion of an additional structure, indicated as the transition structure C_trans, which is confined between the flat absorption structure and the upper contact CONT_SUP, as shown in Figure 5d.

[0112] The transition structure C_trans is shown in Figure 5d, on the other hand, the effective conduction band lower end Ec of the first superlattice SR1 (or SR2) according to the present invention. eff On the other hand, it is fabricated from an n-type doped seventh semiconductor SC7 having a seventh conduction band lower end Ec7 located between it and the sixth conduction band lower end Ec6.

[0113] Alternatively, the upper contact CONT_SUP is p+ Considering the case where a mold-doped layer superlattice is used for manufacturing, as shown in Figure 5d, the lower end of the seventh conduction band Ec7 is, on the other hand, the lower end of the effective conduction band Ec of the first superlattice SR1 (or SR2) according to the present invention. eff On the other hand, the lower end of the effective conduction band Ec of the superlattice of the upper contact CONT_SUP eff_sup It is included between these two.

[0114] The seventh semiconductor SC7, doped to the p-type, is, as shown in Figure 5d, on the other hand, the upper edge Ev of the effective valence band of the first superlattice SR1 (or SR2) according to the present invention. eff On the other hand, it further has a seventh valence band upper edge Ev7 which is contained between it and the sixth valence band upper edge Ev6.

[0115] Alternatively, the upper contact CONT_SUP is p + Considering the case where it is manufactured using a superlattice with doped layers, the upper end of the seventh valence band Ev7 is, on the other hand, the upper end of the effective valence band Ev of the first superlattice SR1 (or SR2) according to the present invention. eff On the other hand, the upper end of the effective valence band Ev of the superlattice of the upper contact CONT_SUP eff_sup It is included between these two.

[0116] From a dimensional standpoint, the transition structure C_TRAN is deposited on a superlattice according to the present invention with an overall thickness of 1 μm to 3 μm, forming a ternary alloy In 0.53 Ga 0.47 n - It can be manufactured using layers produced with a superlattice of doped layers in a mold.

[0117] Figure 6 shows a cross-sectional view of a device D1 for detecting infrared radiation, which includes multiple pixels Pxl according to the present invention.

[0118] The device D1 for detecting infrared radiation is mounted on a substrate SUB. This is a problem of a hybrid optoelectronic system that includes an optical part OPT based on a matrix array formed by multiple pixels arranged in rows and columns, and an electronic part consisting of a readout integrated circuit ROIC on a semiconductor substrate that allows the signals of each pixel of the optical part OPT to be read out individually. A pixel belonging to the optoelectronic system may include one photoelectric element and multiple interconnected photodetectors.

[0119] Alternatively, it is conceivable to manufacture an optical part (OPT) with a single pixel Pxl for a specific application.

[0120] The readout integrated circuit (ROIC) is manufactured using CMOS technology (CMOS stands for complementary metal-oxide-semiconductor) with multiple transistors and thin layers of conductors, semiconductors, or dielectrics on a silicon substrate. Each pixel Pxl is associated with embedded electrodes for reading out signals generated by charge carriers photoexcited by the photodetector structure of the pixel Pxl.

[0121] In conclusion, the present invention provides a novel superlattice structure for a pixel's flat absorption structure for SWIR region detection. The superlattice structure according to the present invention can reduce the effective mass of positive electric carriers (holes) while achieving a target cutoff wavelength greater than 2.1 μm, thereby improving the quantum efficiency of the detector compared to prior art solutions on InP. In addition, the manufacturing process and selection of superlattice materials according to the present invention are compatible with mature manufacturing processes, such as technologies based on InP substrates.

Claims

1. An infrared detector (D1) comprising at least one pixel (Pxl), wherein the pixel comprises a flat absorption structure (SPA), and the flat absorption structure (SPA) comprises The semiconductor layer comprises a first superlattice (SR1, SR2) including a stack along the stacking direction (Z) of a basic group (G1, G2) of semiconductor layers, wherein each of the semiconductor layers of the basic group (G1, G2) is arranged in a crystal lattice structure of a unit cell, and the basic group is • The first band gap (Eg1), ○ Lower end of the first conduction band (Ec1) A first layer (C1) consisting of a first semiconductor (SC1) having, • The second band gap (Eg2), ○ A second conduction band lower end (Ec2) that is strictly lower than the first conduction band lower end (Ec1) A second layer (C2) consisting of a second semiconductor (SC2) having, • A third band gap (Eg3) that is narrower than the first and second band gaps (Eg1, Eg2) mentioned above, ○ A third lower end of the conduction band (Ec3) that is strictly lower than the second lower end of the conduction band (Ec2) A third layer (C3) consisting of a third semiconductor (SC3) having, Includes, The aforementioned basic groups (G1, G2) are: - The following order, namely, the first stack configuration of the second layer (C2), the third layer (C3), the second layer (C2), and the first layer (C1), or - A second stack configuration in which the third layer (C3) is confined between the first and second layers (C1, C2). Manufactured in An infrared detection device (D1) wherein the first semiconductor (SC1) further has a first valence band upper edge (Ev1), and the second semiconductor (SC2) further has a second valence band upper edge (Ev2) that is strictly lower than the first valence band upper edge (Ev1).

2. The infrared detection device (D1) according to claim 1, wherein the first superlattices (SR1, SR2) are manufactured by epitaxy on a substrate (Sub) made of a fourth semiconductor (SC4) arranged in a crystal lattice structure of a unit cell, and the first superlattices (SR1, SR2) are manufactured such that internal mechanical strain is generated in the lattice of each semiconductor layer (C1, C2, C3) of the first superlattices (SR1, SR2) to match the lattice of the crystal lattice structure of the substrate (Sub).

3. The infrared detection device (D1) according to claim 2, wherein the first, second, third, and fourth semiconductors (SC1, SC2, SC3, SC4) are III-V semiconductors.

4. The infrared detection device (D1) according to claim 3, wherein the fourth semiconductor (SC4) is indium phosphide (InP).

5. The composition of the material used to manufacture the semiconductor layer of the basic group is such that the band diagram of the conduction and valence bands in the stacking direction (Z) of the first superlattice has an effective band gap (Eg ｅｆｆ ), upper end of the effective valence band (Ev ｅｆｆ ), and the lower end of the effective conduction band (Ec ｅｆｆ ) is selected to have the effective band gap (Eg ｅｆｆ The infrared detection device (D1) according to any one of claims 3 and 4, wherein the range is 400 meV to 750 meV.

6. The infrared detection device (D1) according to claim 5, wherein the effective mass of positive charge carriers in the stacking direction (Z) of the superlattice (SR1, SR2) is three times smaller than the mass of free electrons.

7. The infrared detection device (D1) according to any one of claims 3 to 6, wherein the third semiconductor (SC3) is a binary composite material InAs.

8. The second semiconductor (SC2) is a ternary alloy In ｘ Ga １－ｘ As is, and x is the alloy In ｘ Ga １－ｘ An infrared detection device (D1) according to any one of claims 3 to 7, wherein the mole fraction of indium in As is.

9. The infrared detection device (D1) according to claim 8, wherein the mole fraction x of indium In in the second semiconductor (SC2) is less than 0.

55.

10. The first semiconductor (SC1) is a ternary alloy GaAs ｙ Sb １－ｙ where y is the molar fraction of arsenic in the alloy GaAs ｙ Sb １－ｙ The infrared detection device (D1) according to any one of claims 8 to 9, where y is the molar fraction of arsenic in the alloy GaAs

11. The infrared detection device (D1) according to claim 10, wherein the mole fraction y of arsenic As in the first semiconductor (SC1) is less than 0.

55.

12. On the one hand, the ratio of the sum of the thicknesses of the layers (C1, C2, C3) of the basic group (G1, G2) weighted by the amplitude of the strain generated in each layer (C1, C2, C3) to the total thickness of the basic group (G1, G2) is lower than or equal to a predetermined value, as described in any one of claims 8 to 11, for the infrared detection device (D1).

13. The aforementioned In ｘ Ga １－ｘ The strain occurring in the lattice of the As layer is tensile strain, and the GaAs ｙ Sb １－ｙ The infrared detection device (D1) according to any one of claims 8 to 12, wherein the strain occurring in the lattice of the layer is compressive strain.

14. The infrared detection device (D1) according to any one of claims 8 to 13, wherein the amplitude of the strain generated in the grid of any one layer (C1, C2, C3) of the basic group (G1) is smaller than the limiting strain at which dislocations occur.

15. The infrared detection device (D1) according to claim 12, wherein the thickness of the layers (C1, C2, C3) of the basic group (G1) is 0.3 nm to 10 nm.

16. The pixels are arranged along the stacking direction (Z), • Circuit board (SUB), - The upper end (Ev) of the effective valence band of the first superlattice (SR1, SR2) ｅｆｆ The upper end of the fifth valence band (Ev5) is strictly lower than ) and ○ The effective band gap (Eg) of the first superlattice (SR1, SR2) ｅｆｆ A fifth band gap (Eg5) that is wider than or equal to ) n has ＋ The lower contact layer (CONT_INF) consists of a fifth semiconductor (SC5) doped into the mold. - The flat absorption structure (SPA) in which the semiconductor layers (C1, C2, C3) of the first superlattice (SR1, SR2) are doped in the n-type, - The upper end (Ev) of the effective valence band of the first superlattice (SR1, SR2) ｅｆｆ The upper end of the sixth valence band (Ev6), which is strictly lower than ) ○ The lower end (Ec) of the effective conduction band of the first superlattice (SR1, SR2) ｅｆｆ ) Strictly higher than the lower end of the sixth conduction band (Ec5) p has ＋ Upper contact layer (CON_SUP) consisting of a mold-doped sixth semiconductor (SC6) An infrared detection device (D1) according to any one of claims 5 to 15, comprising the elements in this order.

17. The pixels are arranged along the stacking direction (Z), - The aforementioned substrate (SUB) ·each ○ The upper end (Ev) of the effective valence band of the first superlattice (SR1, SR2) ｅｆｆ The upper limit of the second effective valence band (Ev) is strictly lower than ) ｅｆｆ＿ｉｎｆ )and, ○ The effective band gap (Eg) of the first superlattice (SR1, SR2) ｅｆｆ A second effective band gap (Eg) that is wider than or equal to it. ｅｆｆ＿ｉｎｆ ) n has ＋ The lower contact (CONT_INF) is manufactured using a second superlattice (SR_inf) doped into the mold. - The flat absorption structure (SPA) in which the semiconductor layers (C1, C2, C3) of the first superlattice (SR1, SR2) are doped in the n-type, ·each ○ The upper end (Ev) of the effective valence band of the first superlattice (SR1, SR2) ｅｆｆ The upper edge of the third effective valence band (Ev) is strictly lower than ) ｅｆｆ＿ｓｕｐ )and, ○ The first superlattice (SR1, SR2) The lower end of the effective conduction band (Ec ｅｆｆ The third effective conduction band lower end (Ec) is strictly higher than ) ｅｆｆ＿ｓｕｐ ) Upper contact (CONT_SUP) manufactured using a third superlattice (SR_SUP) having the above An infrared detection device (D1) according to any one of claims 5 to 15, comprising the elements in this order.

18. The aforementioned flat absorption structure (SPA) is On the other hand, the lower end of the effective conduction band (Ec) of the first superlattice (SR1, SR2) ｅｆｆ )and, On the other hand, the lower end of the sixth conduction band (Ec6) The lower end of the seventh conduction band (Ec7) between these two points. On the other hand, the upper end (Ev) of the effective valence band of the first superlattice (SR1, SR2) ｅｆｆ )and, On the other hand, the upper end of the sixth valence band (Ev6) The upper end of the seventh valence band (Ev7) between It further includes a transition layer (C_trans) made of an n-type doped seventh semiconductor (SC7) having, The infrared detection device (D1) according to claim 16, wherein the transition layer (C_trans) is confined between the superlattice (SR1, SR2) and the upper contact layer (CONT_SUP).

19. The aforementioned flat absorption structure (SPA) is On the other hand, the lower end of the effective conduction band (Ec) of the first superlattice (SR1, SR2) ｅｆｆ )and, On the other hand, the lower end of the third effective conduction band (Ec ｅｆｆ＿ｓｕｐ ) The lower end of the seventh conduction band (Ec7) between these two points. On the other hand, the upper end (Ev) of the effective valence band of the first superlattice (SR1, SR2) ｅｆｆ )and, On the other hand, the upper end of the third valence band (Ev ｅｆｆ＿ｓｕｐ ) The upper end of the seventh valence band (Ev7) between It further includes a transition layer (C_trans) made of a fifth n-type doped semiconductor (SC7) having, The infrared detection device (D1) according to claim 17, wherein the transition layer (C_trans) is confined between the superlattice (SR1, SR2) and the upper contact layer (CONT_SUP).

Images