Multispectral imaging sensor provided with means for limiting crosstalk

Abstract

A hybrid multispectral imaging sensor, characterized in that it comprises a photosensitive backside-illumination detector (DET) that is made on a substrate (100) made of InP, and that is formed of a matrix of pixels (105, P1, P2, P3) that are themselves made in a structure based on InGaAs (103), and a filter module (MF) that is formed of a matrix of elementary filters (1, 2, 3) reproducing said matrix of pixels, and that is mounted into contact with said substrate (100), said substrate (100) made of InP having a thickness less than 50 m, and preferably less than 30 m.

Claims

1. An infrared multispectral imaging sensor comprising: a photosensitive backside-illumination detector (DET) that is made on a first substrate (100) made of InP, and having a backside face (101) and a frontside face (102), and that is formed of a matrix of pixels (105, P1, P2, P3) that are themselves made in a structure (103) based on InGaAs and deposited by epitaxy on the frontside face of the first substrate (100) made of InP; and a filter module (MF) that is formed of a matrix of elementary filters (1, 2, 3) reproducing said matrix of pixels, and that is constituted by a stack made up of a first mirror (MIR1) and of a second mirror (MIR2) that are separated by a spacer (SP), said filter module defining a plurality of filter cells (IF11, IF12, . . . , IF44), each of which comprises at least two filters (FP1, FP2, FP3); said infrared multispectral imaging sensor being characterized in that: said filter module (MF) is formed on a second substrate (SUB) on which the first mirror (MIR1), the spacer (SP), and the second mirror (MIR2) have been deposited, in that order; the second mirror (MIR2) is mounted into contact with said first substrate (100), thereby forming a hybrid multispectral imaging sensor that functions for wavelengths greater than 1000 nm and ranging up to 2200 nm; and said first substrate (100) made of InP has a thickness less than 50 m, and preferably less than 30 m.

2. A sensor according to claim 1, characterized in that said filter module (MF) is adhesively bonded to said detector (DET) around its periphery.

3. A sensor according to claim 1, characterized in that said filter module (MF) is provided with alignment patterns.

4. A sensor according to claim 3, characterized in that at least one of said filters (FP1, FP2, FP3) has a bandpass transfer function.

5. A sensor according to claim 4, characterized in that at least some of said filter cells (IF11, IF12, IF13, IF14) are in alignment in a first strip.

6. A sensor according to claim 5, characterized in that at least some of said filter cells (IF21 to IF24) are in alignment in a second strip that is parallel to and disjoint from the first strip.

7. A sensor according to any one of claim 4, characterized in that at least two of said filters (FP1, FP2, FP3) that are adjacent are separated by a crosstalk barrier.

8. A sensor according to any one of claim 4, characterized in that at least one of said filters (FP1, FP2, FP3) is panchromatic.

9. A sensor according to claim 1, characterized in that said detector (DET) is mounted onto a read circuit (110).

10. A sensor according to claim 9, characterized in that said read circuit (110) is integrated using CMOS technology.

11. A sensor according to claim 1, characterized in that said substrate (100) made of InP is made thinner by polishing.

12. A sensor according to any one of claim 1, characterized in that said substrate (100) made of InP is made thinner by etching.

Description

(1) The present invention appears in greater detail from the following description of embodiments given by way of illustration and with reference to the accompanying figures, in which:

(2) FIG. 1 is a diagram of a known detector made on an InP substrate;

(3) FIG. 2 is a theoretical diagram of a one-dimensional filter cell, and more specifically:

(4) FIG. 2a is a plan view of said cell; and

(5) FIG. 2b is a section view of said cell;

(6) FIGS. 3a to 3c show three steps in making an embodiment of a filter module;

(7) FIG. 4 is a theoretical diagram of a two-dimensional filter module;

(8) FIG. 5 is a diagram showing a filter module having 64 filters and having a screening grid;

(9) FIG. 6 is a diagram of a filter module in which each of its cells has nine filters;

(10) FIG. 7 is a section diagram of a device of the invention; and

(11) FIG. 8 shows a device designed to form a detector.

(12) Elements present in more than one of the figures are given the same reference in all of them.

(13) The description begins with a filter module that has a plurality of generally identical filter cells formed on a substrate SUB.

(14) With reference to FIGS. 2a and 2b, a filter cell comprises three interference filters of the Fabry-Prot type FP1, FP2, FP3 in successive alignment so as to form a strip.

(15) The cell is constituted by a stack on said substrate SUB of the filter module, the substrate being made of glass, or of silica, or of silicon, for example, and the stack being made up of a first mirror MIR1, of a spacer SP, and of a second mirror MIR2.

(16) The spacer SP, which defines the central wavelength of each filter, is thus constant for any given filter and varies from one filter to another. Its profile is staircase-shaped since each filter has a surface that is substantially rectangular.

(17) An advantageous method of making the filter module using thin layer technology is given by way of example.

(18) With reference to FIG. 3a, the method starts by depositing the first mirror MIR1 on the substrate SUB, followed by a dielectric layer or a set of dielectric layers TF for defining the spacer SP. The mirror is either metallic or dielectric.

(19) With reference to FIG. 3b, the dielectric TF is etched: initially at the second and third filters FP2 and FP3, in order to define the thickness of the spacer SP at the second filter FP2; and subsequently at the third filter FP3, in order to define the thickness of said spacer at the third filter.

(20) At the first filter FP1, the spacer SP has the deposition thickness.

(21) With reference to FIG. 3c, the second mirror MIR2 is deposited on the spacer SP in order to finalize the three filters.

(22) The spacer SP may be obtained by depositing a dielectric TF followed by successive etching as described above, but it may also be obtained by depositing a plurality of thin layers in succession.

(23) By way of example, it is possible to scan the range of wavelengths from 900 nm to 2000 nm by modifying the optical thickness of the spacer.

(24) It should be observed at this point that the thickness of the spacer needs to be small enough to obtain only one transmission band in the range to be probed. Specifically, the greater the thickness, the greater the number of wavelengths that satisfy the condition [ne=k /2].

(25) The invention thus enables a set of aligned filters to be made, which filters can thus be referenced in one-dimensional space.

(26) With reference to FIG. 4, the invention also enables the filter cells to be organized in two-dimensional space. Such an organization is often referred to as a matrix organization.

(27) Each one of four identical horizontal strips comprises four cells. The first strip, i.e. the strip that is shown at the top of the figure, corresponds to the first row of a matrix and comprises cells IF11 to IF14. The second, third, and fourth strips respectively comprise cells IF21 to IF24, filters IF31 to IF34, and cells IF41 to IF44.

(28) The organization is said to be a matrix organization because cell IFjk belongs to the j.sup.th horizontal strip and also to a k.sup.th vertical strip that comprises cells IF1k, IF2k, . . . , IF4k.

(29) With reference to FIG. 5, it is desirable for the various filters of the filter module to be well separated in order to avoid partial overlap of one filter on a filter adjacent to it, and in order to minimize any problem of crosstalk. To achieve this, it is possible to add a grid over the filter module (the grid being shown in black in the figure) so as to form a crosstalk barrier for delimiting all of the filters. The grid should be absorbent. By way of example, an absorbent grid may be made by depositing and etching a black chromium (chromium+chromium oxide) while a reflective grid may be made by depositing and etching chromium.

(30) With reference to FIG. 6, each filter cell now has 9 filters. Each of these cells is in the form of a square within which a corresponding filter lies that is tuned to a distinct wavelength 1, 2, 3, 4, . . . , 9.

(31) In this figure, for reasons of clarity, the spacing between the cells has been voluntarily increased compared with the spacing between two filters. Naturally, in reality, these spacings are identical.

(32) The filter module is thus associated with a detector capable of measuring the light fluxes produced by the various filters.

(33) With reference to FIG. 7, the filter module MF that is shown in FIG. 6 is reproduced.

(34) The detector DET is made using InGaAs technology on an InP substrate as described in the introduction of the present application.

(35) The filter module MF comes to bear against the detector DET in contact with the InP substrate so that the filters 1, 2, 3 are facing the pixels P1, P2, P3 and are interposed between the InP substrate of the detector DET and the substrate SUB of the filter module.

(36) In this way, the distance separating the pixels from the filters is minimized, and can be reduced to the thickness of the InP substrate, and crosstalk is also minimized.

(37) Positioning this module MF is performed by means of alignment patterns, which is a technique known to the person skilled in the art of photolithography and is therefore not described in any further detail below.

(38) The filter module MF is fastened to the detector DET by means of a margin of adhesive ST.

(39) In particular, the filter module MF is made on a first type of substrate, e.g. made of glass, silica, or silicon, and the detector DET is made on a second type of substrate, which is made of InP in this example, these two elements as assembled together forming the hybrid structure of the imaging sensor.

(40) For reasons of clarification, it is specified that the pixels commonly have a size of about 15 micrometers.

(41) Furthermore, it is naturally well understood that the InP substrate 100 needs to be made thinner. To achieve this, two solutions are proposed.

(42) The first solution consists in polishing the substrate mechanically down to a thickness of approximately in the range 20 m to 30 m.

(43) In the second solution, with reference to FIG. 8, a stop layer 108 is grown on the InP substrate, and then a thin layer 109 of InP is grown on the stop layer by epitaxy.

(44) The active layer 103 of InGaAs is then grown on said thin layer 109.

(45) It is thus necessary to etch the substrate 100 to the stop layer 108 by selective etching, and then to etch said stop layer also by selective etching. This results in obtaining the required thickness for the InP backing 109.

(46) The above-described embodiments of the invention have been chosen because of their concrete nature. However, it is not possible to list exhaustively all possible embodiments covered by the invention. In particular, it is naturally possible to replace any of the means described by equivalent means without going beyond the ambit of the present invention.