WO2020074837A1

WO2020074837A1 - Device for detecting electromagnetic radiation, comprising a suspended three-dimensional structure

Info

Publication number: WO2020074837A1
Application number: PCT/FR2019/052415
Authority: WO
Inventors: Sébastien Becker; Jean-Jacques Yon
Original assignee: Commissariat à l'énergie atomique et aux énergies alternatives
Priority date: 2018-10-12
Filing date: 2019-10-10
Publication date: 2020-04-16
Also published as: FR3087262A1

Abstract

The invention relates to a device for detecting electromagnetic radiation, comprising a substrate and at least one thermal detector (20) that has a three-dimensional structure (22) in which a reflector (40) extends in a planar manner on an intermediate level that is separate from and located between a lower level comprising the thermal insulation arms (30) and an upper level comprising the absorbent membrane (60).

Description

RADIATION DETECTION DEVICE

ELECTROMAGNETIC HAVING A SUSPENDED THREE-DIMENSIONAL STRUCTURE

TECHNICAL AREA

The field of the invention is that of electromagnetic radiation detection devices comprising at least one thermal detector with an absorbent membrane thermally isolated from the reading substrate. The invention is particularly applicable to the field of infrared or terahertz imaging, thermography, or even gas detection.

PRIOR STATE OF THE ART

An electromagnetic radiation detection device may include sensitive pixels each formed by a thermal detector comprising an absorbent membrane thermally isolated from the reading substrate. The absorbent membrane comprises an absorber of the electromagnetic radiation to be detected associated with a thermometric transducer whose electrical property varies in intensity as a function of the heating of the transducer.

The temperature of the thermometric transducer being however largely dependent on its environment, the absorbent membrane is thermally isolated from the substrate and the reading circuit, the latter being disposed in the substrate. Thus, the absorbent membrane is generally suspended above the substrate by anchoring pillars, and is thermally insulated therefrom by thermal insulation arms. These anchoring pillars and thermal insulation arms also have an electrical function by ensuring the electrical connection of the absorbent membrane to the reading circuit.

However, there is a need to be able to produce such thermal detectors in the form of a matrix of sensitive pixels arranged periodically with a pitch of the array (pixel pitch) of the order of ten microns or even less. To avoid degrading the performance of such thermal detectors, one approach consists in placing the absorbent membrane in a plane different from that of the thermal insulation arms.

Patent application WO2018 / 055276 thus describes an example of a detection device in which the thermal detector comprises a three-dimensional structure suspended above the reading substrate by the anchoring pillars. This structure three-dimensional consists of a lower stage comprising the thermal insulation arms, which are suspended above the substrate by the anchoring pillars and connected to the reading circuit by the latter, and of an upper stage comprising the absorbent membrane detection, which is held above the thermal insulation arms by conductive pillars which bear on the thermal insulation arms. The absorbent membrane comprises an absorber of the electromagnetic radiation of interest associated with a thermometric transducer, here an MOS transistor. The absorbent membrane forms a quarter-wave interference cavity with a reflector which rests on the reading substrate.

Document US2002 / 0179837A1 describes another example of a detection device in which the thermal detector has a three-dimensional structure suspended above the reading substrate by the anchoring pillars. The three-dimensional structure comprises a lower stage containing the thermal insulation arms, an intermediate stage containing the reflector, and an upper stage containing an absorbent membrane. The upper floor is suspended above the lower floor by hollow conductive pillars. However, there is a need to improve the performance of the detection device and to simplify the manufacturing process.

STATEMENT OF THE INVENTION

The invention aims to remedy at least partially the drawbacks of the prior art, and more particularly to provide a device for detecting electromagnetic radiation which allows a reduction in the lateral dimensions of the sensitive pixel, while optimizing the performance of the thermal detector.

For this, the object of the invention is a device for detecting electromagnetic radiation, comprising:

o a substrate comprising a read circuit;

o at least one thermal detector, comprising:

• a three-dimensional structure adapted to detect electromagnetic radiation, suspended above the substrate and electrically connected to the reading circuit by at least one anchoring pillar, comprising:

^■ thermal insulation arms extending planar manner in a floor of said lower dimensional structure,

^■ an electromagnetic radiation absorbing membrane containing a thermometric transducer, extending planarly in a so-called upper stage of the three-dimensional structure, distinct and superimposed on the stage lower, suspended above the thermal insulation arms by conductive pillars which extend from the thermal insulation arms;

• a reflector of electromagnetic radiation, arranged between the substrate and the absorbent membrane so as to form a quarter-wave interference cavity for said electromagnetic radiation.

The reflector extends planarly in a so-called intermediate stage of the three-dimensional structure, distinct and located between the lower and upper stages. The conductive pillar is a hollow pillar having side walls formed of at least one conductive layer, the side walls defining an empty internal space.

According to the invention, the side walls of each conductive pillar are entirely formed, according to their thickness in a plane parallel to the substrate, by said conductive layer. The side walls are therefore entirely made of at least one electrically conductive material. In other words, the conductive layer defines the opposite faces of the side walls. Consequently, the side walls do not have an additional layer made of an insulating material, attached to the conductive layer, which makes it possible to reduce the thermal mass of the thermal detector, and improves the performance of the detection device. In addition, the manufacturing process for obtaining this detection device is simplified, since it is not necessary to make an opening in the additional insulating layer in order to then be able to contact a conductive layer of the thermal insulation arms. .

Some preferred but non-limiting aspects of this detection device are as follows.

Preferably, the thermometric transducer rests on and in contact with bias electrodes, which are in electrical contact with the conductive layer. In addition, the thermometric transducer is here made of a material different from that of the conductive layer. This structural configuration makes it possible to prevent the side walls from comprising an insulating layer, this insulating layer then being present between the thermometric transducer and the conductive layer.

The conductive pillar may include an upper opening leading to the internal space, said upper opening being through, that is to say that it is not closed and that it connects the internal space and the environment of the conductive pillar.

The conductive pillar may have an upper opening leading to the internal space and closed by the absorbent membrane, and may include a through lateral opening located at the side walls. The reflector may include at least one through opening, through which the conductive pillar extends.

The detection device may include a matrix of thermal detectors, the reflector extending continuously opposite each of the absorbent membranes.

The reflector can be suspended above the thermal insulation arms by at least one retaining stud which bears on an anchoring pillar.

The invention also relates to a method of manufacturing a detection device according to any one of the preceding characteristics, comprising the following steps:

i) production of the thermal insulation arms, of the reflector, then of the absorbent membrane resting on at least one conductive pillar, by means of various sacrificial layers deposited successively on the substrate, the conductive layer being deposited before deposition of the transducer material thermometric;

ii) removal of the various sacrificial layers, so as to obtain the suspension of the three-dimensional structure above the substrate.

The conductive pillar can be a hollow pillar having side walls formed entirely, according to their thickness in a plane parallel to the substrate, by said conductive layer, the side walls delimiting an internal space. Step i) can then include the following sub-steps:

- realization of the thermal insulation arms on a first sacrificial layer;

- Realization of the reflector on a second sacrificial layer resting on the first sacrificial layer;

- Deposition of at least a third sacrificial layer on the second sacrificial layer, and production of a vertical orifice through at least the second and third sacrificial layers so as to lead to a thermal insulation arm;

- conformal deposition of a conductive layer on the lateral flanks of the vertical opening so as to form the hollow conductive pillar having an upper opening;

- depositing an additional sacrificial layer so as to cover at least the third sacrificial layer and to fill the internal space of the conductive pillar through the upper opening;

- deposition of the material of the thermometric transducer on and in contact with polarization electrodes, which are in electrical contact with the conductive layer.

Step i) may include the following sub-steps:

- removal of the additional sacrificial layer, apart from a part located in the internal space of the conductive pillar and flush with the level of the upper opening; - Production of the absorbent membrane so as to leave the upper opening through;

and step ii) may include the removal of the additional sacrificial layer located in the internal space, this being evacuated through the upper through opening.

Step i) may include the following sub-steps:

- Production of an intermediate pad located between the reflector and the absorbent membrane, the intermediate pad forming a lug projecting into the vertical orifice, made of a material sensitive to an etchant used during step ii) of removal different sacrificial layers, so as to induce a break in the continuity of the conductive layer during conformal deposition, thus forming a lateral opening through the conductive pillar;

- removal of the additional sacrificial layer, except for a part located in the internal space of the conductive pillar and flush with the level of the upper opening;

- realization of the absorbent membrane so as to close the upper opening; and step ii) may include the removal of the additional sacrificial layer located in the internal space, this being evacuated through the through lateral opening.

The sacrificial layers can be made of an inorganic material, and can be removed by wet chemical etching with hydrofluoric acid, the stud forming the lug being made of titanium, tantalum oxide Ta ₂ 0 ₅ or a silicon nitride.

The conductive layer forming the conductive pillar can be made of WSi, TiN or TiW.

BRIEF DESCRIPTION OF THE DRAWINGS

Other aspects, aims, advantages and characteristics of the invention will appear better on reading the following detailed description of preferred embodiments thereof, given by way of non-limiting example, and made with reference to the accompanying drawings in which:

Figure ÎA and ÎB are schematic and partial views, respectively in cross section and in top view, of a detection device according to a first embodiment;

FIGS. 2A to 2G illustrate different stages of a method of manufacturing a detection device according to a second embodiment;

FIGS. 3A to 3G illustrate different steps of a method of manufacturing a detection device according to a third embodiment. DETAILED PRESENTATION OF PARTICULAR EMBODIMENTS

In the figures and in the following description, the same references represent the same or similar elements. In addition, the different elements are not shown to scale so as to favor the clarity of the figures. Furthermore, the different embodiments and variants are not mutually exclusive and can be combined with one another. Unless otherwise indicated, the terms "substantially", "approximately", "in the order of" mean to the nearest 10%, and preferably to the nearest 5%. Furthermore, the expression "comprising a" should be understood, unless otherwise indicated, as "comprising at least one" and not as "comprising a single".

The invention relates to a device for detecting electromagnetic radiation, for example infrared or terahertz radiation. The detection device can thus be particularly suitable for detecting infrared radiation from the LWIR (Long Wavelength Infrared) range, the wavelength of which is between approximately 8 pm and 14 pm. It has a quarter-wave interference cavity formed between an absorbent membrane and a reflector, thereby maximizing the absorption of infrared radiation to be detected by the absorbent membrane.

In the following description, the detection device comprises one or more thermal detectors each containing a thermometric transducer, located in an absorbent membrane thermally isolated from the substrate. A thermometric transducer is an element having an electrical property varying with its heating, and can be a thermistor material formed for example of vanadium or titanium oxide, or of amorphous silicon, a capacitance formed by a pyroelectric or ferroelectric material, a diode (pn or pin junction), or even a field effect transistor with a metal - oxide - semiconductor structure (MOSFET).

According to the invention, each thermal detector comprises a three-dimensional structure, suspended above the substrate, and comprising at least three distinct functional stages and superimposed on each other, that is to say arranged opposite and parallel one another. The three-dimensional structure thus comprises a lower stage of thermal insulation, an intermediate stage of optical reflection, and an upper stage of detection of electromagnetic radiation. The three stages are distant from each other along the Z axis, that is to say they are spaced from each other by a non-zero distance along the Z axis. Furthermore, the conductive pillars which maintain suspended from the upper stage vis-à-vis the lower stage are hollow conductive pillars whose side walls are entirely formed by a conductive layer.

Figures îA and îB are schematic and partial views, respectively in cross section and in top view, of a detection device i according to a first embodiment. Fig.iA is a view along the section plane A-A illustrated in fig.iB. In this example, the thermometric transducer is a thermistor material, for example amorphous silicon or a vanadium oxide. A single thermal detector 20 is shown here, but the detection device i advantageously comprises a matrix of identical thermal detectors 20 (sensitive pixels).

We define here and for the rest of the description a direct orthogonal three-dimensional reference (C, U, Z), where the plane (X, Y) is substantially parallel to the main plane of the reading substrate 10 of the detection device 1 , and where the axis Z is oriented in a direction substantially orthogonal to the main plane of the reading substrate 10 and oriented towards the absorbent membrane 60. In the following description, the terms “lower” and “upper” are understood to mean being relative to an increasing positioning when one moves away from the reading substrate 10 in the + Z direction.

The detection device 1 comprises a functionalized substrate, called the reading substrate 10, produced in this example based on silicon, comprising a reading circuit 12 allowing the control and reading of the thermal detectors 20.

The read circuit 12 is here in the form of a CMOS integrated circuit located in a support substrate 11. It has portions 14 of conductive lines, for example metal, separated from each other by a dielectric material 13 , for example a mineral material based on silicon such as a silicon oxide SiO _x , a silicon nitride SiN _x , or their alloys. It may also include active electronic elements (not shown), for example diodes, transistors, or passive electronic elements, for example capacitors, resistors, etc., connected by electrical interconnections to the thermal detector 20 on the one hand, and on the other hand to a connection pad (not shown), the latter being intended to connect the detection device 1 to an external electronic device. By way of illustration, the conductive portions 14 and the conductive vias 15 can be produced, for example, from copper, aluminum or tungsten. The copper or tungsten may optionally be located between sublayers made of titanium nitride, tantalum or the like. The reading substrate 10 here has an upper face 10a formed in particular by a surface of an inter-metal insulating layer 13 and a surface of conductive portions 14 of the last level of electrical interconnection. The upper face is advantageously coated with a protective layer 16, in particular when the three-dimensional structure 22 is produced by the use of mineral sacrificial layers, which are then eliminated by chemical attack in HF acid medium (hydrofluoric acid) . The protective layer 16 then has an etching stop function, and is therefore suitable for ensuring protection of the support substrate il and of the inter-metal dielectric layers 13 when they are made of an inorganic material with respect to HF chemical attack. This protective layer 16 thus forms a hermetic and chemically inert layer. It is electrically insulating to avoid any short circuit between the metal line portions. It can thus be made of alumina Al ₂ 0 ₃ , or even of aluminum nitride or fluoride, or even of intrinsic amorphous silicon. It may have a thickness of between a few tens and a few hundred nanometers, for example between πm and soonm, preferably between 2 µm and τm.

The detection device 1 comprises at least one thermal detector 20, and here a matrix of thermal detectors 20 identical to each other. The thermal detectors 20 thus form sensitive pixels of the matrix, arranged periodically, and can have a lateral dimension in the XY plane (known as the pixel pitch), of the order of a few tens of microns, for example equal to approximately πm or even less. .

The thermal detector 20 comprises a three-dimensional structure 22 for detecting electromagnetic radiation, suspended above the reading substrate 10 by anchoring pillars 21 and connected to the reading circuit 12 by the latter.

The anchoring pillars 21 are conductive pads made of at least one electrically conductive material, which extend along the axis Z from the reading substrate 10 to the three-dimensional structure 22. They are in contact portions 14 of conductive lines, and thus ensure the electrical connection of the three-dimensional structure 22 to the reading circuit 12. The anchoring pillars 21 can be made, for example, of copper, aluminum or tungsten, possibly encapsulated in at minus a protective underlay 16 of titanium nitride, or the like.

The three-dimensional structure 22 has at least three different distinct functional stages and superimposed on each other. The functional stages are thus arranged in separate planes which are parallel to each other, and are opposite one another.

The first functional stage of the three-dimensional structure 22 is a lower stage having in particular a thermal insulation function. It thus comprises the thermal insulation arms 30, which extend planarly in a first plane Pi parallel to the plane XY, from the anchoring pillars 21. They provide thermal insulation of the absorbent membrane 60 vis- with respect to the reading substrate 10, the electrical connection of the thermistor material 64, and participate in maintaining the absorbent membrane 60 suspended above the reading substrate 10. The thermal insulation arms 30 comprise a conductive layer 32 made of an electrically conductive material, here encapsulated between two dielectric layers 31, 33 participating in the stiffening of the thermal insulation arms 30. By way of example, the thermal insulation arms 30 can be formed of a thin conductive layer 32 of TiN encapsulated in two dielectric layers 31, 33 of amorphous silicon. They extend longitudinally preferably in a serpentine fashion, so as to maximize their length in the surface of the sensitive pixel. They extend here longitudinally between a first end 34 assembled and connected to an anchoring pillar 21 and a second opposite end 35 on which rests and is connected a conductive pillar 50 for holding the absorbent membrane 60.

The second functional stage of the three-dimensional structure 22 is an intermediate stage having a function of optical reflection of the electromagnetic radiation to be detected. It comprises a reflector 40 adapted to reflect towards the absorbent membrane 60 a proportion of the electromagnetic radiation to be detected that has not been absorbed by the absorbent membrane 60. The reflector 40 is a layer of a reflective material, for example metallic, which extends planarly in a second plane P2 parallel to the plane XY and distinct from the first plane Pi, from a holding stud 41 which rests on an anchoring pillar 21. The holding stud 41 extends thus along the Z axis between an upper end of the anchoring pillar 21 and the reflector 40. Also, the reflector 40 is arranged above and facing the thermal insulation arms 30 along the Z axis. support 41 can be made differently or identical to the anchor pillars 21. It can thus be made based on copper, aluminum or tungsten, optionally laterally encapsulated in a sub-co uche based on titanium nitride. The reflector 40 can also be made from copper, aluminum or tungsten. The reflectors 40 of the different sensitive pixels can be metallic layers distinct from each other, that is to say separated from each other in the plane P2. Alternatively, they can be formed from the same metallic layer which extends continuously at the level of each sensitive pixel. Alternatively also, the holding pad 41 may rest on and in contact with the reading substrate, and not rest on a thermal insulation arm.

The third functional stage of the three-dimensional structure 22 is an upper stage having a function of detecting the electromagnetic radiation of interest. For this, it comprises the absorbent membrane 60 containing an absorber of the electromagnetic radiation to be detected associated with the thermistor material. The absorbent membrane 60 extends planarly in a third plane P3 parallel to the plane XY and distinct from the planes Pi and P2. It is kept suspended above the reflector 40 by at least one conductive pillar 50 which rests on a thermal insulation arm 30. The absorbent membrane 60 is therefore suspended above the reading substrate 10 and is connected to the reading circuit 12, by the action of the pillars anchor 21, thermal insulation arms 30, and conductive pillars 50. It is also thermally isolated from the reading substrate 10 by the thermal insulation arms 30.

The absorbent membrane 60 is here conventionally formed of a stack of a lower insulating layer 61 made of a dielectric material, of two electrodes 62 electrically isolated from each other by lateral spacing, of a layer intermediate insulator 63 made of a dielectric material and covering the electrodes 62 and the lateral spacing, except in two openings opening onto the electrodes 62, of a thermistor material 64, for example amorphous silicon or a vanadium or titanium oxide. The thermistor material 64 is in contact with the two electrodes 62 via the openings. An upper protective layer 65 covers the thermistor material 64, in particular to protect the thermistor material during chemical attack with hydrofluoric acid implemented subsequently. The absorber is here formed by the electrodes 62, which are made of at least one metallic material, for example titanium nitride.

The conductive pillar 50 thus extends along the axis Z between the second longitudinal end 35 of a thermal insulation arm 30 and the absorbent membrane 60. The conductive pillar 50 can be produced differently or identical to anchoring pillars 21 and the support stud 41. It comprises an electrically conductive material, for example copper, aluminum or tungsten, optionally laterally encapsulated in a sublayer based on titanium nitride. The conductive pillar 50 is in contact with the conductive layer 32 of the thermal insulation arm 30 at a lower end, and in contact with a bias electrode 62 at an opposite upper end. It can be made in one piece with a bias electrode 62 or be separate and in contact with the latter. It can be a solid pillar, that is to say that the internal space delimited by its periphery, in the XY plane, is filled with at least one material (conductive or not), or be a hollow pillar, it that is to say that the conductive material extends at its periphery so as to form side walls which delimit an internal space not filled with a material.

The absorbent membrane 60 is spaced vertically and is arranged opposite the reflector 40 along the Z axis, so that the absorber (electrodes 62) and the reflector 40 together define a quarter-wave interference optical cavity, allowing thus maximizing the absorption of the electromagnetic radiation to be detected by the absorbent membrane 60. Thus, the detection device i comprises at least one sensitive pixel, and preferably a matrix of identical sensitive pixels, each having a thermal detector 20 containing a three-dimensional structure 22 suspended above the reading substrate îo by the pillars anchor 21. This three-dimensional structure 22 thus comprises at least three separate stages: a lower stage in which the thermal insulation arms 30 extend; an intermediate stage in which the reflector 40 extends, and an upper stage in which the absorbent detection membrane 60 extends.

This vertical arrangement of the different functional stages of thermal insulation, optical reflection, and absorption / detection of the electromagnetic radiation of interest makes it possible both to produce a sensitive pixel with small lateral dimensions in the XY plane, by example of the order of ten microns or less, while optimizing the performance of the thermal detector 20. In fact, it is then possible to improve the thermal insulation of the absorbent membrane 60 (and therefore to increase the thermal resistance of the thermal detector 20) by making thermal insulation arms 30 of considerable length in the plane Pi insofar as there is no constraint on the presence of the absorbent membrane 60. In addition, the thermal detector 20 has a large filling factor FF, this parameter FF (for Fill Factor) being defined as the ratio of the surface of the absorbent membrane 60 to the a total area of the sensitive pixel, in an XY plane parallel to the plane of the substrate. The filling factor FF is particularly important insofar as the absorbent membrane 60 is not limited in size by the presence of the thermal insulation arms 30 in the plane P3. Also, the optical efficiency of the thermal detector 20, produced by the parameter FF by the absorption e of the thermal detector 20 is preserved, the absorption e being defined as the proportion absorbed per unit area of the incident energy of the electromagnetic radiation to be detected. .

In addition, the fact that the thermal insulation arms 30 are located in the lower stage and not in the intermediate stage, and therefore are located outside the quarter-wave interference cavity, allows preserving the performance of the thermal detector 20. In fact, the presence of the thermal insulation arms 30 in the quarter-wave interference cavity can lead to a disturbance of the interference cavity and consequently to a degradation of the absorption e. The absorber of the absorbent membrane 60 is usually arranged at a distance from the reflector such that the reflected wave generates constructive interference with the incident wave at the absorbent membrane, that is to say at a substantially equal distance. at l _o / 4h of the reflector 40, l _o being a central wavelength of the spectral range of the electromagnetic radiation to be detected (for example isopm for the LWIR range), and n being the refractive index of the medium located in the quarter-wave interference cavity, usually vacuum. The presence of the arms of thermal insulation 30 within the quarter-wave interference cavity can cause a reduction in the absorption e of the thermal detector 20, due to a modification of the optical path making it possible to generate the constructive interference at the level of the membrane absorbent on the one hand, and due to a non-zero absorption of electromagnetic radiation by the thermal insulation arms 30 on the other hand. It appears that these drawbacks are eliminated in the context of the invention insofar as the thermal insulation arms 30 are arranged outside the quarter-wave interference cavity. The performance of the thermal detector 20 is then preserved.

FIGS. 2A to 2G illustrate different stages of a method of manufacturing the detection device 1 according to a second embodiment. The detection device 1 is here similar to that illustrated in FIG. 2A and is essentially distinguished from it in that the conductive pillar 50 for maintaining the upper floor is a hollow pillar. In addition, the reflector 40 is formed of a planar layer made of a reflective material, which rests on a planar lower layer 42 made in one piece and of the same material with the holding stud 41.

In this example, the thermal detector 20 is produced using mineral sacrificial layers intended to be subsequently removed by wet etching in an acid medium (HF vapor).

Referring to FIG. 2A, the reading substrate 10 is produced, formed of a support substrate 11 containing the reading circuit 12 suitable for controlling and reading the thermal detector 20. The reading circuit 12 thus includes conductive portions 14 which are flush with the upper face 10a of the reading substrate 10, which is substantially planar. The conductive portions 14 and the conductive vias 15 can be made of copper, aluminum and / or tungsten, among others, for example by means of a damascene process in which trenches made in the inter-metal insulating layer 13 are filled The leveling of the conductive portions 14 at the level of the upper face 10a can be obtained by a mechanical-chemical planarization technique (CMP).

We can then deposit a protective layer 16 so as to cover the inter-metal insulating layer 13. This etching stop layer is made of a material substantially inert to the etching agent used subsequently to remove the layers sacrificial minerals, for example in the HF medium in the vapor phase. It thus prevents the underlying mineral insulating layers 13 from being etched during this step of removing the sacrificial layers. It can be formed from an aluminum oxide or nitride, from aluminum trifluoride, or from intrinsic amorphous silicon (not intentionally doped). It can be filed for example by PVD (for Physical Vapor Déposition, in English) and can have a thickness of the order of ten nanometers to a few hundred nanometers.

Referring to fig.2B, we realize the anchoring pillars 21 and the lower stage of the three-dimensional structure 22, that is to say the thermal insulation arms 30. For this, we deposits a first sacrificial layer 71 on the reading substrate 10, for example made of an inorganic material such as a silicon oxide SiO _x deposited by chemical vapor deposition assisted by plasma (PECVD). This mineral material is capable of being removed by wet chemical etching, in particular by chemical attack in an acid medium, the etchant preferably being hydrofluoric acid (HF) in the vapor phase. This mineral sacrificial layer 71 is deposited so as to extend continuously over substantially the entire surface of the reading substrate 10 and thus cover the etching stop layer 16. The thickness of the sacrificial layer 71 along the Z axis allows to define the distance separating the thermal insulation arms 30 from the reading substrate 10. It can be of the order of a few hundred nanometers to a few microns.

Vertical holes are then made for the formation of the anchoring pillars 21. They are made by photolithography and etching, and pass through the first mineral sacrificial layer 71 and the protective layer 16, to lead to the conductive portions 14 of the read circuit 12. The vertical orifices may have a cross section in the plane (X, Y) of square, rectangular, or circular shape, with a surface area substantially equal, for example, to 25.25 pm ² . The anchoring pillars 21 are then produced in the vertical holes. They can be produced by filling the orifices with one or more electrically conductive materials. For example, they may each comprise a layer of TiN deposited by PVD or MOCVD (for Metal Organic Chemical Vapor Deposition, in English) on the vertical sides of the orifices, and a copper or tungsten core filling the defined space. transversely through the TiN layer. A CMP step then makes it possible to remove the excess filling materials and to planarize the upper face formed by the sacrificial layer 71 and the anchoring pillars 21.

Then performs the thermal insulation arms 30. For this, a lower dielectric layer 31 is deposited here on the mineral sacrificial layer 71, then a conductive layer 32 and an upper dielectric layer 33. The electrical contact between the layer conductive 32 and the anchor pillar is obtained through an opening made in the lower dielectric layer 31 and filled with the conductive layer 32. The conductive layer 32 is in contact with the upper end of the anchor pillars 21. It is made of an electrically conductive material, for example TiN with a thickness of a few nanometers to a few tens of nanometers, for example μm. The lower dielectric layers 31 and upper 33 can be made of silicon amorphous, silicon carbide, alumina Al ₂ 0 ₃ or aluminum nitride, among others. The lower dielectric layers 31 and upper 33 can have a thickness of a few tens of nanometers, for example 20 nm, and participate in ensuring the stiffening of the thermal insulation arms 30. A structuring of the dielectric layers 31, 33 and of the conductive layer 32 is then carried out by photolithography and localized etching, so as to define the topology of the thermal insulation arms 30 in the first plane Pi. Thus, the thermal insulation arms 30 extend longitudinally, for example in a serpentine, between a first end 34 of electrical connection between the anchoring pillar 21 and the conductive layer 32, and a second opposite end 35 intended to receive the conductive pillar 50 for holding the absorbent membrane 60.

Referring to FIG. 2C, the intermediate stage of the three-dimensional structure 22 is then produced. For this, a second mineral sacrificial layer 72 is deposited so as to cover the underlying sacrificial layer 71 and the arms d thermal insulation 30. The thickness of the sacrificial layer 72 makes it possible to define the distance separating the lower and intermediate stages of the three-dimensional structure 22, that is to say the reflector 40 with respect to the isolation arms thermal 30. It can be of the order of a few hundred nanometers to a few microns, for example comprised between soonm and spm, and preferably comprised between ispm and 2pm.

Then, vertical orifices intended for the formation of the holding pads 41 of the reflector 40 are produced. They are produced by photolithography and etching, and pass through the second mineral sacrificial layer 72, to lead to the thermal insulation arms 30, preferably on the first end 34 of the thermal insulation arms 30 resting on the anchoring pillars 21, so that the holding pads 41 are preferably located opposite the anchoring pillars 21. As indicated previously, as a variant , the holding pads 41 can rest on and in contact with the reading substrate 10 and not the thermal insulation arms.

Then, in this example, a continuous holding layer 42 is deposited, preferably made of a dielectric material such as amorphous silicon, so as to fill the vertical orifices and to cover the second sacrificial layer 72. The layer of holding 42 is then covered by a reflective layer 40, forming the reflector, made of a material reflecting the electromagnetic radiation of interest, for example a metallic layer, for example made of aluminum, copper, tungsten, or the like. The retaining layer 42 and the reflector 40 can extend continuously at the level of all the sensitive pixels, or be distinct from one sensitive pixel to the other.

Referring to fig.2D, the hollow conductive pillar 50 is produced. If necessary, a through opening 43 is made beforehand by localized etching of the layer. reflector 40 and the retaining layer 42 perpendicular to the second end 35 of the thermal insulation arm 30 intended to receive the conductive pillar 50.

Then depositing a third mineral sacrificial layer 73 so as to cover the underlying reflective layer 40. The thickness of this sacrificial layer 73 makes it possible to define the distance separating the intermediate and upper levels of the three-dimensional structure 22, that is to say the absorbent membrane 60 with respect to the reflector 40. It thus defines the size of the quarter-wave interference cavity of the thermal detector 20. It can be of the order of 1.5 pm to 2.5 pm in the case of the detection of LWIR radiation.

Then deposited on the surface of the third sacrificial layer 73 a dielectric layer 61, intended to form the lower dielectric layer of the absorbent membrane 60. The lower dielectric layer 61 can be made of at least one dielectric material. By way of example, it can be formed for example of a thin under-layer of protection against chemical etching carried out subsequently, for example in amorphous silicon, in Al ₂ 0 ₃ or in AIN d ' a thickness between Ionm and 50nm, coated with a passivation sublayer, for example in SiN with a thickness between Ionm and 30nm.

Then depositing a conductive layer 62 made of an electrically conductive material, and here advantageously absorbent vis-à-vis the electromagnetic radiation to be detected. This conductive layer 62 is intended to form the polarization electrodes of the thermistor material, which also performs the function of absorber. It is preferably made of Ti, TiN, TaN, WN, or the like, and has a thickness of between 3 nm and 20 nm approximately. Here it covers the lower dielectric layer 61.

Vertical holes 51 are then produced for the formation of the conductive pillars 50. They are produced by photolithography and etching, and pass, from top to bottom, the conductive layer 62, the lower dielectric layer 61, the third and second sacrificial layers 73, 72, and the upper dielectric layer 33 of the thermal insulation arms 30, to lead to the conductive layer 32 of the latter, here at their second ends 35. The vertical orifices 51 may have a cross section in the plane (X, Y) of square, rectangular, or circular shape, with an area substantially equal, for example, to 0.25pm ² . They can have a dimension in the XY plane of the order of approximately 0.5 μm, and a height of the order of approximately 3 μm, within the framework of a detection of the infrared range LWIR.

Referring to fig.2E, the conductive pillars 50 are produced, intended to hold the absorbent membranes 60 suspended, while connecting them to the reading circuit 12. For this, a conductive layer 52 is deposited so as to continuously cover the sides of the vertical orifice 51 and to come into contact on the one hand with the conductive layer 32 of the thermal insulation arms 30, and on the other hand with the conductive layer 62 intended to form the bias electrodes. The conductive layer 52 can be deposited by a conformal deposition technique, for example by chemical vapor deposition (Chemical Vapor Deposition, in English), by atomic thin layer deposition (Atomic Loyer Deposition, in English), or even by physical deposition in vapor phase (Physical Vapor Deposition, in English), for example by sputtering, or other.

In this example, the conductive layer 52 intended to form the conductive pillars 50 is distinct from that intended to form the bias electrodes 62, but as a variant, it could be a single conductive layer. In this case where the conductive layer 52 forms a single layer with the electrodes 62, for example made of TiN, it is advantageous to structure the conductive layer 52 at the absorbent membrane to adapt it to the impedance of the vacuum , for example by making a network of through orifices The conductive layer 52 here has a thickness less than the lateral dimension of the vertical orifice 51 in the XY plane, so that this vertical orifice 51 is not entirely filled by the conductive material. By way of example, the conductive layer 52 may have a substantially constant thickness of a few tens of nanometers to a few hundred nanometers, for example isonm approximately in an XY plane, and for example sonm in a plane substantially orthogonal to the XY plane, while the vertical orifice 51 may have a lateral dimension equal, for example, to approximately 0.5 μm. The conductive layer 52 is made of at least one electrically conductive material, for example based on tungsten or titanium, such as WSi, TiN, TiW, among others. Also, the conductive pillar 50 is formed by a conductive layer 52 extending vertically along the axis Z and laterally delimiting a hollow non-conductive internal space which opens onto an upper opening 53. The conductive pillars 50 therefore have side walls 50a which extend over the lateral flanks of the vertical orifice 51. The lateral walls 50a are made in one piece. They can be rectilinear side walls joined to each other, or correspond to different parts of the same circular or oval peripheral wall. They therefore extend substantially orthogonal to the XY plane of the reading substrate 10. The conducting pillars 50 are in electrical contact with the reading circuit 12 via the anchoring pillars 21 and the thermal insulation arms 30 , and make it possible to connect the bias electrodes 62 of the absorbent membranes 60.

Finally, a fourth mineral sacrificial layer 74 is deposited so as to cover the conductive layer 52 forming the conductive pillars 50 and fill the space internal hollow of these. The mineral sacrificial layer 74 is preferably made of a dielectric material identical to that of the underlying sacrificial layers.

Referring to fig.2F, we realize the upper stage of the three-dimensional structure 22, that is to say we finish the realization of the absorbent membrane 60. For this, we delete part of the fourth sacrificial layer 74 so as to expose the upper face of the conductive layer 52. Part of the fourth sacrificial layer 74, located in the internal space of the conductive pillars 50, is thus preserved. It is flush with the part parallel to the XY plane of the conductive layer forming the conductive pillars 50. This gives an essentially flat surface which facilitates the continuation of technological operations (without which, it would be necessary to remove the layers which could fall into the empty internal space of the supporting pillars)

Then, by photolithography and etching, the conductive layer 62 is structured in the XY plane so as to define the polarization electrodes. The part parallel to the XY plane of the conductive layer 52 forming the conductive pillars 50 is also etched, so as to keep only an upper portion in contact with the bias electrodes 62. Preferably, the material of the electrodes 62 is different. from that of the conductive layer 52, so as to have an etching selectivity between these two materials. Thus, the conductive layer is preferably made of WSi and the electrodes are preferably made of TiN. The bias electrodes 62 are substantially coplanar and electrically isolated from each other. They are made so as to have an electrical resistance of the order of 377 Îl / square. The electrodes 62 each contact a different support pillar and are separated from each other in the XY plane by a distance preferably less than about X _c / 5 or even less than 1 pm so as not to disturb the absorption electromagnetic radiation to be detected with a central wavelength l _o .

Finally, by a series of deposition, photolithography and etching steps, the absorbent membrane 60 is obtained comprising an intermediate insulating layer 63 made of a dielectric material, for example of Al 0 ₃ , or AIN, and covering the electrodes 62 and the lateral spacing between them, except at the level of openings opening onto the electrodes 62. A portion of thermistor material 64, for example made of amorphous silicon or of a vanadium or titanium oxide, is deposited in electrical contact electrodes 62 via the openings. It may have a thickness for example between 20nm and 200nm. Finally, an upper protective layer 65, for example made of amorphous silicon, AI O or AIN with a thickness of between Ion and Sonm, is deposited so as to cover the thermistor material 64. Finally, there is provided a through hole opening on the upper opening 53 of the conductive pillars 50, by localized etching of the intermediate dielectric layers 63 and upper 65, in order to allow subsequent removal of the part of the fourth sacrificial layer 74 located in the internal space of the conductive pillars 50. The through orifice has lateral dimensions smaller than those of the upper opening 53, this to preserve the integrity of the conductive layer 52. The subsequent removal of the fourth sacrificial layer 74 located in the conductive pillar 50 is particularly advantageous insofar as it makes it possible to reduce the calorific mass of the conductive pillars 50, and thus to reduce the thermal time constant of the thermal detector 20.

Referring to fig.2G, we delete the various sacrificial layers 71, 72, 73, 74, so as to suspend the three-dimensional structure 22, and therefore its different stages, above the reading substrate 10. The suspension can be carried out after having encapsulated the thermal detector 20 in a housing (not shown) defining a vacuum cavity intended to be hermetic. The suspension can be obtained by chemical etching of the various mineral sacrificial layers, here by wet chemical etching by attack with hydrofluoric acid in the vapor phase. The part of the fourth sacrificial layer 74 located in the internal space of the conductive pillars 50 is evacuated at the same time through the upper opening 53 of the conductive pillars 50.

A detection device 1 is thus obtained here comprising a matrix of sensitive pixels which may have a particularly small pixel pitch, each sensitive pixel having a reduced time constant. As detailed above, the quarter-wave interference cavity is not disturbed by the presence of the thermal insulation arms 30, unlike the example of the prior art mentioned above. In addition, the conductive pillars 50 are here hollow pillars, the internal space of which is not filled with an electrically conductive material or with a dielectric material of a sacrificial layer. The absence of filling material in this internal space makes it possible to reduce the thermal capacity associated with the conductive pillar 50, thereby reducing the thermal time constant of the thermal detector 20.

In addition, the fact that the side walls 50a of each conductive pillar 50 are formed only of the conductive layer 52, and not also of an additional insulating layer as in document US2002 / 0179837, makes it possible to reduce the mass thermal detector, and therefore improve the performance of the detection device. In addition, the manufacturing process is simplified, since it is not necessary to produce, as in document US2002 / 0179837, an opening in the additional insulating layer, so as to lead to a layer conductive of the thermal insulation arms, this step being particularly delicate due to the local topology. It is also possible to reduce the width in the XY plane of the conductive pillars 50.

The thermometric transducer 64 is here made of a material different from that of the bias electrodes 62 and of the conductive layer 52. It is deposited after the step of making the hollow pillars 50, which is made possible by the fact that the hollow pillars 50 are momentarily filled with the sacrificial layer 74. Also, the thermometric transducer 64 rests on and in contact with polarization electrodes 52, which are in electrical contact with the conductive layer 52. This makes it possible to obtain a membrane absorbent with high performance. Indeed, making the vertical hole 51 and the conductive pillars 50 after the deposition of the thermometric transducer 64 can lead to a degradation of the performance of the absorbent membrane, as well as to a potential mechanical fragility of the latter, in particular if other vertical orifices must be made through the thermometric material 64 to come into contact with the electrodes 62.

FIGS. 3A to 3G illustrate different steps of a method of manufacturing the detection device 1 according to a third embodiment. The detection device 1 is similar to that illustrated in FIG. 2G and is essentially distinguished in that the conductive pillars 50 for maintaining the upper level are hollow pillars closed in their upper openings 53 by the absorbent membrane 60, and each comprising a transverse lateral opening 54 (ie not closed) allowing the evacuation from the side of the sacrificial layer located in the internal space of the conductive pillar 50. This variant of the manufacturing process is advantageous, in particular when the dimensioning of the absorbent membrane 60 does not release the upper opening 53 of the conductive pillar 50, as described above with reference to Fig.2A to 2G.

With reference to FIG. 3A, the lower and intermediate stages of the three-dimensional structure 22 are produced. The structure obtained is then identical or similar to that described with reference to FIG. 2C. The steps for obtaining this structure are not detailed again.

Referring to FIG. 3B, intermediate pads 2 are produced which are intended to form lugs 2a projecting within each vertical orifice 51 of the conductive pillars 50. Each lug 2a is here defined as being a portion of the pad 2 projecting from the vertical opening 51 of the conductive pillar 50, that is to say projecting from the lateral flanks of the vertical opening 51. Beforehand, an opening 43 is produced if necessary by localized etching of the reflective layer 40 and the retaining layer 42 perpendicular to the part of the thermal insulation arms 30 intended to receive the conductive pillar 50, here opposite the second end 35 of a thermal insulation arm 30.

Then depositing a third mineral sacrificial layer 73.1 so as to cover the reflective layer 40 and the underlying sacrificial layer 72. The thickness of this sacrificial layer 73.1 participates in defining the distance separating the intermediate and upper stages of the three-dimensional structure 22, that is to say the absorbent membrane 60 with respect to the reflector 40. It thus participates in defining the size of the quarter-wave interference cavity of the thermal detector 20.

Then produced, by deposition, photolithography and etching, a pad 2 made of a sensitive material (that is to say capable of being etched) to the etching agent used subsequently to remove the mineral sacrificial layers. This material can be chosen from titanium, tantalum oxide Ta ₂ 0 ₅ , or a silicon nitride deposited preferably by PECVD at low temperature, for example at 300 ° C., among others. It has a thickness which depends on the nature and the thickness of the other layers to be etched, and may be of the order of a few tens of nanometers to a few hundred nanometers, for example between about ½ m and about 300 nm. It is positioned in the XY plane so that part of the stud 2 protrudes into the vertical hole 51 intended to make the conductive pillar 50.

Referring to FIGS-3C and 3D, the vertical orifices 51 are then produced to form the hollow conductive pillars 50. For this, a first mineral sacrificial layer 73.2 is first deposited so as to cover the underlying sacrificial layer 73.1 as well as the stud 2 intended to form the lug. The thickness of the third and fourth sacrificial layers 73.1, 73.2 makes it possible to define the size of the quarter-wave interference cavity along the axis Z between the intermediate and upper stages of the three-dimensional structure 22.

Then deposited on the surface of the sacrificial layer 73.2 the lower dielectric layer 61 of the absorbent membrane 60, as well as the conductive layer 62 intended to form the polarization electrodes, then the vertical orifices 51 intended for formation are produced. conductive pillars 50 hollow. They are produced by photolithography and etching, and pass, from top to bottom, the conductive layer 62, the lower dielectric layer 61, the fourth, third and second sacrificial layers 73.2, 73.1, 72, and the upper dielectric layer 33, to open up. on the conductive layer 32 of the thermal insulation arms 30, here at the second end 35 of the thermal insulation arms 30. To obtain the projection of the lug 2a with respect to the lateral flanks of the orifice vertical 51, and more precisely the projection of the lug 2a with respect to the third sacrificial layer 73.1 (ie under the lug along the axis Z), the etching of the sacrificial layers, and in particular of the third layer sacrificial 73.1, has a slight isotropy.

Due to the prior positioning of the pad 2, a portion 2a of the pad 2 is then projecting vis-à-vis the lateral flanks of the vertical orifice 51. This portion therefore forms a lug 2a. Preferably, it projects from the lateral flanks over a distance in the XY plane at least greater than the thickness that the conductive layer 52 is intended to have at the level of the lateral flanks. Thus, if the vertical walls of the conductive pillar 50 are intended to have a thickness of approximately sonm, the lug 2a is then projecting over a distance advantageously at least equal to sonm. Preferably, to preserve the mechanical strength of the conductive pillar 50, the lug 2a can extend over at most half of the local circumference of the vertical orifice 51.

Referring to FIG. 3E, the conductive pillars 50 are produced which are intended to hold the absorbent membranes suspended, and to connect them to the reading circuit 12. For this, a conductive layer 52 is deposited so as to cover the sides of the vertical orifice 51 and coming into contact on the one hand with the conductive layer 32 of the thermal insulation arms 30, and on the other hand with the conductive layer 62 intended to form the polarization electrodes. The conductive layer 52 is deposited by a conformal deposition technique, for example by chemical vapor deposition (Chemical Vapor Deposition, in English), or by physical vapor deposition (Physical Vapor Deposition, in English), for example by spraying cathodic (Sputtering, in English), or other. The conductive layer 52 is made from a material of interest chosen from a tungsten silicide WSi, TiN or TiW, among others.

The inventors have thus found that when a conductive layer 52 of such a material of interest (WSi, TiN, etc.) is deposited conformally, preferably by CVD, in the vertical orifice 51 in which protrudes the lug 2a, there is a break in the continuity of the conductive layer 52, in particular under the lug 2a. A lateral opening 54 is therefore formed which will be used later to remove the sacrificial layer 74 which will be present in the internal space of the conductive pillar 50.

Finally, a fifth mineral sacrificial layer 74 is deposited so as to cover the conductive layer 52 forming the conductive pillars and to fill the hollow internal space of the latter. The mineral sacrificial layer 74 is preferably made of a dielectric material identical to that of the underlying sacrificial layers.

Referring to fig.3F, we finalize the production of the absorbent membrane 60. The steps are similar or identical to those described above, and differ essentially in that the absorbent membrane 60, here the dielectric layer intermediate 63, covers the upper opening of the conductive pillar 50. As before, a fifth mineral sacrificial layer 74 is then located in the internal space of the conductive pillar 50 to obtain an essentially flat surface which facilitates the continuation of the technological operations. The steps for producing the absorbent membrane 60 are not detailed again.

With reference to FIG. 3G, the sacrificial layers 71, 72, 73.1, 73.2, 74 are removed, so as to suspend the three-dimensional structure 22, and therefore its different stages, above the reading substrate 10. The suspension is obtained here by chemical etching of the various mineral sacrificial layers, here by wet chemical etching by attack with hydrofluoric acid in the vapor phase. The part of the fifth sacrificial layer 74 located in the internal space of the conductive pillars 50 is removed at the same time, through the lateral opening 54 of the conductive pillars 50, and the studs 2 forming the pins 2a are removed in the since they are made of a material sensitive to the etchant used, which makes it possible to prevent them from degrading the performance of the detection device 1 by falling on the reflectors 40 or by remaining fixed to the conductive pillars (and thus by disturbing the electromagnetic radiation in the optical cavity).

A detection device 1 is thus obtained here comprising a matrix of sensitive pixels which may have a particularly reduced pixel pitch, each sensitive pixel having a reduced time constant. As detailed above, the quarter-wave interference cavity is not disturbed by the presence of the thermal insulation arms 30, unlike the example of the prior art mentioned above. In addition, the conductive pillars 50 are here hollow pillars, the upper opening of which is closed by the absorbent membrane 60. The sacrificial layer present in the internal space is then effectively removed during the removal of the various sacrificial layers, so that the thermal capacity associated with the conductive pillars 50 is reduced, thereby optimizing the thermal time constant of the thermal detector 20. The advantages indicated above in connection with the conductive pillars 50 whose side walls are formed entirely by the conductive layer 52 are also valid here . As well as those relating to the fact that the thermometric transducer 64 rests on and in contact with polarization electrodes 52, which are in electrical contact with the conductive layer 52.

Particular embodiments have just been described. Different variants and modifications will appear to those skilled in the art.

Claims

1. Device for detecting (1) electromagnetic radiation, comprising:

o a substrate (10) comprising a read circuit (12);

o at least one thermal detector (20), comprising:

• a three-dimensional structure (22) adapted to detect electromagnetic radiation, suspended above the substrate (10) and electrically connected to the reading circuit (12) by at least one anchoring pillar (21), comprising:

^■ thermal insulation arms (30) extending planarly in a so-called lower stage of the three-dimensional structure (22),

^■ an absorbing membrane (60) of electromagnetic radiation containing a thermometric transducer (64), extending planarly in a so-called upper stage of the three-dimensional structure (22), distinct and superimposed on the lower stage, suspended above thermal insulation arms (30) by conductive pillars (50) which extend from the thermal insulation arms (30), each conductive pillar (50) being a hollow pillar having formed side walls (50a) at least one conductive layer (52), the side walls (50a) defining an empty internal space;

^■ a reflector (40) of electromagnetic radiation, disposed between the substrate and the absorbent membrane (60) so as to form a quarter-wave interference cavity for said electromagnetic radiation, and extending planarly in a so-called intermediate stage of the three-dimensional structure (22), distinct and located between the lower and upper floors; characterized in that:

^■ the side walls (50a) of each conductive pillar (50) being entirely formed, according to their thickness in a plane parallel to the substrate, by said conductive layer (52).

2. Detection device (1) according to claim 1, in which the thermometric transducer (64) rests on and in contact with polarization electrodes (52), which are in electrical contact with the conductive layer (52), the transducer thermometer (64) being made of a material different from that of the conductive layer (52).

3. Detection device (1) according to claim 2, wherein the conductive pillar (50) has an upper opening (53) opening onto the internal space, said upper opening (53) being through.

4. Detection device (1) according to claim 2, wherein the conductive pillar (50) has an upper opening (53) opening onto the internal space and closed by the absorbent membrane (60), and has a lateral opening ( 54) through located at the side walls (50a).

5. Detection device (1) according to any one of claims 1 to 4, wherein the reflector (40) has at least one through opening (43), through which extends the conductive pillar (50).

6. Detection device (1) according to claim 5, comprising a matrix of thermal detectors, in which the reflector (40) extends continuously opposite each of the absorbent membranes (60).

7. Detection device (1) according to any one of claims 1 to 6, in which the reflector (40) is suspended above the thermal insulation arms (30) by at least one holding stud (41) which rests on an anchoring pillar (21).

8. Method for manufacturing a detection device (1) according to any one of the preceding claims, comprising the following steps:

i) production of the thermal insulation arms (30), of the reflector (40), then of the absorbent membrane (60) resting on at least one conductive pillar (50), by means of different sacrificial layers deposited successively on the substrate ( 10), the conductive layer (52) being deposited before deposition of the material of the thermometric transducer (64);

ii) removal of the various sacrificial layers, so as to obtain the suspension of the three-dimensional structure (22) above the substrate (10).

9. The manufacturing method according to claim 8, in which the conductive pillar (50) is a hollow pillar comprising side walls (50a) formed entirely, according to their thickness in a plane parallel to the substrate, by said conductive layer (52), the side walls (50a) delimiting an internal space, step i) comprising the following substeps:

realization of the thermal insulation arms (30) on a first sacrificial layer (71);

- Realization of the reflector (40) on a second sacrificial layer (72) resting on the first sacrificial layer (71);

depositing at least a third sacrificial layer (73) on the second sacrificial layer (72), and making a vertical orifice (51) through at least the second and third sacrificial layers (72, 73) so as to lead to a thermal insulation arm (30);

conformal deposition of a conductive layer (52) on the lateral flanks of the vertical orifice (51) so as to form the hollow conductive pillar (50) having an upper opening (53);

depositing an additional sacrificial layer (74) so as to cover at least the third sacrificial layer (73) and fill the internal space of the conductive pillar (50) through the upper opening (53);

depositing the material of the temperature transducer (64) on and in contact with bias electrodes (52), which are in electrical contact with the conductive layer (52).

10. The manufacturing method according to claim 9, in which step i) comprises the following substeps:

removal of the additional sacrificial layer (74), apart from a part located in the internal space of the conductive pillar (50) and flush with the upper opening (53);

realization of the absorbent membrane (60) so as to leave the upper opening (53) through;

and in which step ii) involves the removal of the additional sacrificial layer (74) located in the internal space, this being evacuated through the upper through opening (53).

11. The manufacturing method according to claim 9, in which step i) comprises the following substeps:

production of an intermediate stud (2) located between the reflector (40) and the absorbent membrane (60), the intermediate stud (2) forming a lug (2a) projecting into the vertical orifice (51), produced in a material sensitive to an etchant used during step ii) for removing the various sacrificial layers, so as to induce a break in continuity of the conductive layer (52) during conformal deposition, thus forming a lateral opening (54) through the conductive pillar (50);

making the absorbent membrane (60) so as to close the upper opening (53); and in which step ii) includes the removal of the additional sacrificial layer (74) located in the internal space, this being evacuated through the lateral opening (54) passing through.

12. The manufacturing method according to claim 11, wherein the sacrificial layers are made of an inorganic material, and are removed by wet chemical etching with hydrofluoric acid, the stud forming the lug being made of titanium, of oxide of tantalum Ta ₂ 0 ₅ or a silicon nitride.

13. The manufacturing method according to claim 11 or 12, wherein the conductive layer (52) forming the conductive pillar (50) is made of WSi, TiN or TiW.