FR2620531A1 - THERMAL DETECTOR IN NETWORK - Google Patents

Abstract

The invention relates to thermal detectors. The thermal detector network, of the type comprising discrete detector elements 1, each of which is supported by a common support layer 7, and mounted in thermal contact with a corresponding collector 3 made of a radiation-absorbing material, the area of each collector 3 being larger that the surface of the corresponding detector element 1 is essentially characterized in that the contact between each collector 3 and the support layer 7 is limited to a zone disposed substantially inside the periphery of the edge of the detector 3 and away from the latter. Application to the production of infrared detector networks.

Description

The present invention relates to improvements to detectors

  and, more particularly, to infrared thermal detector networks. In a thermal detector, the energy of the incident radiation absorbed increases the temperature of each sensing element.

  tor. This increase in temperature leads to changes in

  in the sensor properties that depend on the temperature, which can be controlled to detect incident radiation. Thermal detectors

  especially interesting in the present. context

  take the pyroelectric detectors and / or

  dielectric bolometer capacitors in the form of a two-dimensional network interfaced to a semiconductor scanning device to provide a solid and uncooled compact storage array for

  the formation of thermal images and applications

  annexes. There is increasing interest in obtaining good performance of thermal detectors in two-dimensional, fine-pitched arrays (e.g., networks of x 100 elements and more, at steps of 100 μm or

  less) approaching the fundamental physical limits

  for a thermal detector, this performance being evaluated for example in terms of signal-to-noise ratio of the detector network. To obtain these possibilities, it is

  first of all, to make the signal

  detector (the answer) by optimizing the structure of the device and choosing

  the detector material (for example, by choosing

  a material with a high coefficient of response), making the electronic operating scheme optimal and by minimizing as much as possible all the sources of noise associated for example with the dielectric loss of the detector material as well as the sources of noise in

  the electronic reading or operating circuit.

  However, when these noise sources are sufficiently reduced, a physical limit noise is approached which is determined by the thermal conductance of the structure and the radiation collection or reception efficiency as well as the absorption in the unit cells. . This value of the limiting noise, or thermal fluctuation noise limit, is given by the expression: Thermal fluctuation noise 2 equivalent (NEP) = KT2GH no GH is the total thermal conductance of a network element, n is the radiation reception efficiency and the absorption in the surface of the elementary cell, K is the Boltzmann constant and T is the absolute temperature.

  In addition, it can be shown that the

  to the power of equivalent noise fluctuation

  thermal is proportional to YH / n o YH is the admit-

  thermal energy of a network element, a part of which is constituted by the term jWCH where w is the angular frequency of operation of the device and CH is the thermal capacitance of the element. This expression implies that small thicknesses of detector elements are necessary such that wCH is smaller or of the order of thermal conductance, the actual part of YH in the operating frequency range. In practice, total thermal conductance values below 0.2 w at 2 K-1 are required, with radiation reception and absorption efficiencies approaching unity, for steps equal to or less than pm and thicknesses. elements less than 15 μm (and preferably less than 8 μm) so that the thermal fluctuation noise limit still provides useful performance for operating frequencies of the order of 10 to 50 Hz. For other frequencies , the preferred thickness varies inversely with respect to

at the frequency.

  A pyroelectric detector, presenting a perfect NEP

  lioré, is described in the British patent application

  No 2 100 058 of 1982. In this application, a detection element

  thermal unit is mounted adjacent to an element

  absorbent surface and it is supported elastically

  on one or more flexible films. The electrical connection is provided by drivers who

  protrude laterally on the internal surfaces of

  cles. While this arrangement can be prolonged

  in the form of a one-dimensional network,

  the need for a lateral extension, as mentioned above.

  above, generally prohibits a two-dimensional

fine-pitched.

  An integrated two-dimensional detector array is described in U.S. Patent No. 4,162,420, 1979. In this device, however, the elements are not

  not crosslinked. The plane of the detector is distant from a cir-

  cooked with silicon but electrically

  nected to the latter, by means of a series of metal rods

  (square section of 10 pm with 25 pm length).

  The use of metallic materials in this form, as well as the lack of crosslinking, if translated by a device having a relatively thermal insulation

  mediocre and thermal intermodulation between

adjacent areas.

  So far there is a problem for the realization of a large integrated infrared detector network

  two-dimensional, fine-pitched, thermally insulated

  satisfactory performance and low intermodulation.

  The present invention therefore aims to provide structures

  detectors and corresponding processes

  for two-dimensional thermal detector networks

  important dimensions in which a high degree of iso-

  thermal equation (low GH) is combined with

  high reception and absorption of radiation.

  According to the invention, a thermal detector network comprises discrete thermal detector elements arranged in a two-dimensional array, each element being supported on one side of a layer of an elastic material with low thermal conductivity, which is also intended to space said elements with respect to each other, the opposite face of said layer

  carrying a common electrode structure which is in compliance

  tact with each element and form interconnections with each other by means of electrically conductive tracks and high thermal resistance, each element having an electrode element in contact with an inner surface of said element, each electrode element being connected by means of a connection to an input circuit of a semiconductor medium such that the connecting members act to support the layer with respect to the support,

  and each element being associated with a radiation absorber

is lying.

  The components of the housing may be of low conductivity

  thermal activity. They can be arranged in a posi-

  between the respective elements and the support.

  According to a variant, the connecting members can be

  adjacent to the sides of the remaining elements

  poles, extending directly between the support and the layer. The present invention also provides methods

  treatment for the organs above. Such structures

  as mentioned above, can be achieved by processing raw layers of detector material, or

  well they can be formed by implicit techniques

as thin films.

  These improved structures and methods are suitable for dielectric and pyroelectric bolometer detector arrays that employ monocrystalline materials or ferroelectric oxide ceramics, materials such as: barium-strontium titanate; lead-magnesium niobate; lead-scandium tantalate; niobate

  lead-iron; lead-iron tungstate; as well as zircon

lead nate / lead-iron niobate.

  The complete physical separation of the sensor elements from the thermal sensor material of the network ensures low thermal conductance and intermodulation

  weak thermal between the elements.

  The support layer is preferably of a polymeric material. The polymeric materials are most suitable for the support layer and of course for any other characteristic involving a support or a mechanical connection but must also have the low thermal conductance required. Polymeric materials generally consist of insulating materials having a complex molecular structure, frequently arranged to form an amorphous solid rather than

  crystal, show the values of thermal conductivity

  than the lowest at room temperature among the

  classes of solids as shown in the table below.

  Material class Thermal conductivity range W.M l. K-1 Metals 10400 Semiconductors 1-200 Mineral Crystalline Solids 2-200 Mineral Amorphous Solids 0.8-2.0 0.1-0.4 Organic Polymers A preferred class of polymeric materials for

  support layer is constituted by the family of poly-

  imides high temperature polymers. We can find

  examples of the substances that have been developed

  Currently, they are extensively used as dielectric passivation and interlayer coatings for silicon integrated circuits. A particular example of a polyimide material which has proved advantageous in this case is the so-called "PIQ" polyimide manufactured by the firm Hitachi. The use of a polyimide film backing material, which is strong and flexible with a high glass transition temperature and good adhesion to metal or oxide surfaces, is fully compatible with the ferroelectric materials mentioned above. as well as with the planar welding connection, a preferred technique for the purpose currently sought. The polyimide films are coated and cured, in particular between

  300 and 350 C. Coating thicknesses can be

  0.2 micron, with sub-micron coatings being preferred. The difference in thermal expansion coefficients of the polyimide film and the ferroelectric oxides forming the detector elements (.50 ppm.K-1 and less than 10 ppm.K -1) places the support film under slight positive tension after cooling.

  from the crosslinking stage. This property

  holds the film taut after crosslinking and results in low microphony. Other polymer materials

  possible to form a composite support film

  take polyamides, polyurethanes (including UV curing types) and polymers deposited

  from a vapor phase (for example parylene).

  The use of a discrete element structure, mounted on a polymer support layer, which is under slight voltage, provides a very low microphonic response. A certain degree of piezoelectric activity is inherent both in the pyroelectric and dielectric materials of bolometers under the influence of the applied bias field and the study of the detector must be

  conduct to minimize this unspoken

visable.

  Absorbent radiation structure mentioned

  above can be formed from a metal film

  a thin, resilient carrier having an impedance

  that of free space (377 10) which can also be used

  vir as a common electrode. Such a film absorbs

  % of broadband incident radiation. Coats

  Anti-reflection elements can be added to increase the absorption efficiency over a wavelength band

  more limited. To further increase the absorption efficiency

-7 2620531

  In the band of interest, it is possible to use a reflective metal element active electrode layer.

  on the opposite side of the detector element while

  the infrared optical properties of the materials of the structure and by controlling the thickness of the detector element. Yields of 0.6 to 0.8 in the infrared band of 8 to 14

  such multilayer structures. However, a structure

  The preferred absorption layer comprises a metallic black such as platinum black which has a

  absorption efficiency close to unity in a structure

  very low thermal mass.

  While the absorbent elements may be of a surface equal to that of the respective sensor elements,

  Absorbent and detection surfaces in structures

  Detectors in cross-linked networks may vary independently. Indeed, the independent nature of these two structures can be exploited to obtain

  a high efficiency of reception and absorption of

  combined with a high degree of thermal insulation

detector elements.

  The preferred electrode materials, particularly for the interconnecting tracks between the elements and to the connections on the elements, include metals, metal alloys, cermets and amorphous metal alloys having a high thermal resistivity, exhibit good adhesion to other materials of the structure and can be drawn to form narrow tracks. By way of example, mention may be made of anodized chromium as well as metal films of nickel-chromium alloy, of chromium-silica cermets and of amorphous nickel-indium. Such films, required at a small thickness (for example from 10 to 10 μm) and at a small width (for example from 2 to 10 μm), can be deposited by means of a number of thermal deposition techniques including evaporation

  on filaments and electron beam, or the spraying

  tion. Spraying processes, and in particular

  magnetron sputtering methods are preferably

  rable. An electrode comprising one of the above metals, with a coating of an inert metal with high electrical conductivity such as gold, is generally required for

  the deposition of metallic blacks such as platinum black.

  Inert and high conductivity metal can be removed from the interconnecting tracks of the elements after deposition

  black to eliminate the associated thermal shunt.

  Open openings or through holes in the polymeric support film above the elements are desirable to provide good thermal contact between the absorbent and the sensor element. If an impedance balanced thin film absorption structure is used, the polymer support film can also be

  serve as an anti-reflection coating for the absorbent.

  Other features and advantages of the present

  the invention will come out better from the description that will

  FIG. 1 is a sectional view of an integrated thermal detector array according to one embodiment of the present invention; Figure 2 shows a detailed sectional view of a unit cell of the network of Figure 1, showing in detail the structure of its connecting member; FIG. 3 represents a sectional view of a network

  integrated thermal detector according to a variant of

of the present invention;

  Figures 4 and 5 show respective views of

  in section and plan of a part of a form of

  preferred embodiment of the network shown in the preceding figure.

toothed;

  Figures 6 and 7 show schematic views of

  in section of the networks illustrated in the preceding figures.

  dentes, showing construction and assembly details

93 2620531

from the raw material; and

  Figures 8 and 9 show schematic views of

  in section of a preform and a corresponding network

thin film treatment.

  The integrated thermal detector network represented

  in FIG. 1 comprises a large number of detecting elements

  1, each of which is discrete and spaced in

  network formation by means of an elastic support layer

  only 3 in a material with low thermal conductivity. This

  layer provides sufficient support and mechanical integrity.

  sufficient for the whole. On an integrated surface

  layer 3 is arranged a structure 5 of elec-

  common trode which forms interconnection between the elements by means of electrically conductive links 7 and high thermal resistance. The opposite face of each element 1 carries an individual electrode 9. The network structure with supported element is disposed above a semiconductor support 11 and connecting members 13, which also form a support and electrical connection, are included between each element 1 and the support 11. These connecting members provide electrical continuity between the electrodes 9 of the elements and the plates 15

  corresponding input circuits of the operating interfaces

  on the semiconductor medium 11. At each

  detector element 1 corresponds to a plate 17 of material

  heat-absorbing material on the opposite surface of the layer 3. The thermal contact between each element 1 and the absorbent 17 is thus achieved, the effective thermal conduction being provided through the support (the layer)

  or via metallized openings of the support.

  As shown in FIG. 2, each organ of

  link 13 is in a composite structure and disposed

  below each element 1, which combines the contradictory requirements of electrical interconnection and thermal insulation. This composite structure can be made for example by forming a Messe structure 19 in a material with low thermal conductivity such that a

2620531

  polymer, immediately below the detector 1, and

  having a metal track 21 thin and high resistance

  temperature at the top of the mesa 19. The mesa

  19 in metallized polymer is then connected and electrically connected.

  to the input plate 15 of the support 11 for operation by means of, for example, a solder connection

  23 and a surface metallization 25, 27. This structure

  It can also be used in an inverted form, that is with the solder connection 23 adjacent to the thermal detector element 1 and the mesa

  of polymer 19 adjacent to the operating support 11.

  However, this structure is less favorable because the solder connection 23 then acts in thermal load on

  the detector element 1 and reduces the response accordingly.

  A range of polymeric materials can be selected to form the polymeric mesa structure 19. The polyimides are particularly suitable for the inverted form of the mesa structure where the mesa 19 is located on the support 11. The polyimides are less suitable for the original mesa structure, the fact that the cross-linking cycle that is involved can lead to some difficulties in

  the sequence of the process attached to the device. The utili-

  The addition of a "Novolac" photoresist, which can be drawn and then cast and cross-linked to form a mesa 19 with a rounded edge profile to facilitate subsequent metallization, is preferable for the structure according to FIG. 2. In this case, is desirable to

  implement a radiation stabilization treatment

  (X-ray UV or electron beam) or chemical (eg immersion in an acidified formaldehyde solution) in order to cross-link the Novolac resin and stabilize the dimensions of the mesa structure before

  metallization and the final binding operation by

tough.

  The adoption of a point structure where links

  sounds 13 are arranged below each element 1

  implies that the element itself can cover a proportion

11 2Z620531

  significant increase in the available area of the elementary cell. It can be shown that an optimal compromise is achieved between the reception factor (i.e. the reception efficiency) and thermal intermodulation between adjacent elements 1 for equal areas of the absorbent 17 and Element 1, on an element having a ratio interval / width of 4/1 equivalent

  at a radiation reception efficiency of 0.64. The UTI

  The use of an enlarged platinum black collector (see FIG. 4) and / or XAC-type large surface collector structures can still provide a significant additional gain in receiving efficiency without compromising thermal insulation. A reception and absorption efficiency approaching 0.9 can be obtained in this way, in particular for structures having

a step of 100 "or less.

  An alternative network structure is shown in FIG. 3. In this example, the thermal insulation

  of element 1 relative to the support of the operating circuit

  tation (the dominant heat source in the two structures) and between the adjacent elements 1, is obtained simultaneously by means of an envelope 3 in a polymer support strip around each element 1. The individual elements i of this structure are more small ones

  of the point-link structure (Figure 1) and the enve-

  loppe polymer support 3 is proportionately more

  great. The links 13 which form the electrical connections

  from the electrodes 9 of elements with the cir-

  The burners are placed over the envelope 3 and it is no longer essential to incorporate additional thermal barrier structures (mesas) in the links 13 as was the case for the dot structure. In this structure, thermal conductance

  total is compromised with respect to the

  radiation (for the case where the surfaces of

  element 1 and absorbent 17 are equal) and a dia-

  The optimal element meter is about 0.6 relative to the elementary cell width, giving an elementary cell reception factor (or a radiation reception efficiency) of about 0.35. total and lateral thermal conductance for this polymer film structure are substantially less than that of the dot-bonded structure

  (Figure 1) at a given step and for structural geometries

  feasible, which then provides a higher resultant performance of the device provided that the

  radiation reception efficiency can be increased

  you. A preferred embodiment of a two-dimensional film thermal detector array structure comprises a platinum black and enhanced surface collector 17 '(PBXAC) as shown in FIG.

  4. It is particularly noteworthy that this structure

  The preferred embodiment comprises truncating the corners giving octagonal rather than square elements, which increases the length of the thin film metallization tracks 29 which connect each element electrode 9 to each connection connection 13 on the operating circuit 15. giving a further increase in thermal insulation with minimal loss of absorbent surface. This track 29 of thin film is disposed between the electrode 9 of element on a polymeric film 31 isolated

  to the point where the links 13 are placed. The utilisa-

  Four common electrode connections 7 must also be noted. While two connections per element 1, plus additional links at the end of each row of elements, are sufficient for a full common connection, some redundancy is

  useful with fine-track geometries that are desirable

  tables in these structures with fine pitch. The use of three to four common connections is therefore desirable. Again, the planar solder connection is adopted as a preferred method of bonding

  13 between the network and the electrical exploitation circuit

13 2620531

  tronics. This technique is particularly suitable for the film-mounted structure, since the process

  welding link means that no mechanical pressure

  This is necessary in any solid phase bonding process such as bonding to

  indium), which could be incompatible with the location

  bonding to a flexible polymeric film.

  In contrast, the solder bonding method relies on the natural wetting action of the molten solder on the wettable metallization 25, 27 performed on the two components to form the bonds 13. The ratio of the width to the length of the solder connection, with a bond height comparable to the diameter of the wettable seal pad, allows each element

  thermal sensor 1 to be physically spaced from the

  the operating circuit, provided that the thickness of the

  the element is less than the height of the link

  welding sound. This imperative of connection geometry cor-

  responds to the need for very small thicknesses of elements

as indicated above.

  Examples of methods of manufacturing thermal detector arrays will now be given with the structures described above. These methods, while containing some novel features, are not necessarily the only processes or the only operating sequences by which such networks can be made. Processes suitable for detector arrays prepared from

  Ferro-electric raw materials and ferro-film

  thin electric wires deposited. As particular characteristics

  Examples of such methods include etching or selective thinning and cross-linking, new structures for forming mechanical supports for networks and processes.

  alignment of hybrid devices.

  The structures of the two-dimensional thermal sensing arrays of the dot-bond type and mounted with polymeric film, such as made from thinned raw materials, are illustrated in Figures 6 and 7. These figures show the edges of the structure. network and part of the network itself. The edges of the structure comprise a rim 33 forming a relatively thick and non-reticular base frame around the entire detector network. This rim forms a sufficiently rigid mechanical support frame for the polymer support or the film 3. It also forms a region of the structure that can be brought into contact during the

  and a location for a row or more

  solder linkages 35 of substantially

  greater than the solder connections 23 of the network itself.

  even. These solder connections 35 are intended to automatically align the hybrid structure during assembly,

  in particular to align the network of the entire

  its 23 small diameter on the support 11.

  A possible sequence of manufacture for these arrangements

  As shown in FIGS. 6 and 7, this includes the orientation, cutting, coating and polishing of the ferroelectric raw material, either in the form of a single crystal or in the form of a ceramic.

  polycrystalline, to form a pellet with parallel faces

  the desired thickness for the uncrosslinked rim 33 of the device. The area from which the network

  crosslinked must be achieved is then thinned selectively-

  the final thickness of the device. Suitable selective thinning techniques may include various wet and dry etch methods, for example, optically enhanced etching

  ion beam attack, depending on the

  riau and its crystalline form. Ion beam attack has the advantage of being a relatively universal etching method and is therefore preferred in many cases. Areas that should not be thinned are suitably masked from the etchants, for example using a foil mask or a product layer

AS4 2620531

  photoresist, according to the etching process.

  The different layers necessary for what will become the face 37 of incident radiation of the device

  are then added to selectively thinned wells.

  These wells generally comprise an etch stop layer for the subsequent crosslinking operation, the support polymer film or film layer 3, the common electrode layers and the infrared absorbent structure 17. The pellet is then returned and the elementary electrodes 9 and the layers are added.

  crosslinking masking. The crosslinking is preferably

  by an anisotropic etching process such as

* an attack by ion beam or reactive ions.

  Masking materials and etch layers can be suitably chosen depending on the material

  to attack, the attack process and the geometry

  cated. A set of wells 39 can be chemically attacked

  below the support rim 33 at the same time as

  crosslinking of the network element, to accept later

  the solder linkages to the dia.

  meter bigger. Wells attacked 39 are particularly suitable

  especially for the film-mounted structure, as shown in FIG.

  23, 35 are defined on half of the silicon pellet.

  operation of the final hybrid device, the said

  wells 39 may also contribute to the mechanical alignment

  components before melting the weld to complete the hybrid device bond. When the crosslinking is complete, the attack stop layer

  acts to protect the substrate or film

  polymer core. The attack stop layer, having

  served for this purpose can then be eliminated if desired.

  When the crosslinking is complete, we can add

  the different layers necessary for the side of con-

  nexion of the element of the structure. These may comprise a polymer insulation layer 31, the elementary electrode 9 and the polymer layers 19, the tracks

A, 4 2620531

  21 element / circuit interconnection, as well as

  Wettable metallization pads 25, 27 for accepting solder connections 23. A corresponding set of wettable metallization pads 27 are provided on the input pads 15 on the silicon operating circuit. The solder bosses 23 may be defined on either component of the hybrid device. The ferroelectric materials which form the sensitive elements of the thermal detector network can also be appropriately deposited directly in the form of a thin film at the desired thickness of the active element, which can avoid potentially costly processes. crystalline growth, ceramic production, orientation, cutting, coating and polishing inherent in the use of raw materials for such devices. The processes of

  deposition of thin films of such ferro-

  electric ones include spraying from.

  metal-organic vapor phase deposition (MCCVD) as well as the in situ decomposition of deposited metal-organic precursor films (eg films produced by the sol-gel process or by

  in situ hydrolysis of complex solutions agitated metal-

  organic product). Such films can be polycrystalline

  tallins with different degrees of crystallo-

  favorite graphics, or monocrystalline in nature

  deposition and / or decomposition conditions. Particularly important parameters in this regard include the deposition or decomposition temperature and the nature of the support on which these films are deposited. These processes, for relatively complex materials such as ferroelectric oxides mentioned above, generally involve deposition or decomposition temperatures above 400 C, particularly if highly crystalline films with preferred orientation or monocrystalline films are desired. For 41 /% d2620531

  this reason, the direct deposit on an isola-

  In this case, the appropriate thermal setting already in place on the operating device is generally difficult, the film deposition is therefore preferably carried out on a suitable temporary support. A support 41 with a deposited ferroelectric film 43 and a barrier layer 45 is illustrated in FIG. 8. The choice of support material

  appropriate refractory may remove the time constraint

  of the above-mentioned deposition process and may also allow verifications for the control

  of the structure of the film. Appropriate support materials

  These may include amorphous materials (eg, fused quartz) and crystalline oxides and other inorganic compounds. Monocrystalline substrates for example spinel, sapphire, magnesium oxide, cubic zircon, sodium chloride and lithium fluoride) are particularly interesting because they allow

  epitaxial growth of ferro-film or film

  monocrystalline electrical, if the parameters of the mesh

  or the spacings of the mesh are suitably equi-

  free. Monocrystalline ferroelectric films

  correctly oriented can provide a beautiful material

  more suitable shot (eg due to a lower dielectric loss than the polycrystalline film

equivalent).

  The barrier layer 45 shown in FIG. 8

  may be necessary in order to avoid a chemical interaction

  that between the film being grown and the substrate and / or to contribute to the nucleation of the film. The barrier layer 45 may itself be deposited on the support 41 so as to provide the preferred orientation or to form a monocrystalline epitaxial layer. The preferred orientation or monocrystalline nature of this barrier layer can be translated on the ferroelectric film that develops there. Barrier layer 45 can also be used to obtain a gradual transition between Af29 2620531

  the mesh parameter of the support 41 and that of the ferro-film

  43, in the case where occurs epitaxial growth. An example of a barrier layer 45 is platinum metal, which can be deposited with a preferred orientation on fused quartz or deposited in the form of an epitaxial film on a magnesium oxide.

monocrystalline spray.

  After depositing the thin ferroelectric film 43 of required thickness on the support 41, this structure can then be treated in a manner quite similar to the structure prepared from the raw material, as shown in FIG. support layer 41 is chemically etched from the active area of the array, suitably to the selected material of the support, but is retained around the perimeter of the structure to form the thicker and uncrosslinked support frame 33 for the film or the polymer support film 3 which carries the detector network. The barrier layer 45 can be

  included in the network structure or eliminated if one.

  the desire. The treatment also follows a complete pathway

  similar to the one described above and translates

  by a similar structure physically.

  Other structural variants can be derived from

  those described above and are also within the scope of the present invention. By way of a particular example, mention may be made of the implementation of other arrangements for the concentration and / or the reception of the incident infrared radiation on the detector elements 1. Thus, refractive or reflective optical structures comprising immersion optical lenses , individual and miniaturized as well as face reflection concentration structures

  planes or curves, can be included in the assembly.

  Such structures can be made for example

  by photochemical etching or by chemical etching

  of single-crystal silica with subsequent coating

  suitable antireflection or metallization.

  Such concentration structures can also be assembled with the hybrid device and substantially aligned with it using the planar weld bond.

SC0 2620531

Claims (5)

  A thermal detector array, comprising discrete thermal detector elements arranged in a two-dimensional array, characterized in that each element (1) is supported on a face of a layer   (3) an elastic material with low thermal conductivity   also intended to space said elements between   they, an opposite face of said layer carrying a structure   forming a common electrode (5) which is in contact with each element and forms interconnections with each other by means of electrically conductive and resistance tracks (7).   high temperature, each element (1) having an electrical   elementary trode (9) in contact with an integral surface   the element, each elementary electrode (9) being connected by means of a connecting member (13) to an input circuit (15) of a semiconductor carrier (11),   such that said connecting members (13) consis-   provide support for the layer (3) relative to the support (11) and each element (1) being associated with an absorber (17) radiation.   2. Network according to claim 1, characterized in that the connecting members (13) are made of   a material with low thermal conductivity.   3. Network according to one of claims 1 or 2,   characterized in that the connecting members (13) are arranged between the respective elements (1) and the support (11).   4. Network according to one of claims 1 or 2,   characterized in that the connecting members (13) are adjacent to the faces of the respective elements (1), extending directly between the support (11) and the layer (3).   5. Method of manufacturing a network according to one   any of claims 1 to 4, including in particular   thin film deposition, metallization, masking and the solder connection.