Detector of electromagnetic radiation and in particular infrared radiation, and process for producing said detector

Abstract

An infrared radiation detector includes an array of elementary imaging bolometric detectors, each of the elementary bolometric detectors being formed of a bolometric membrane including a film made of vanadium oxide VOx, having a resistivity in the range from 6 ohm.Math.cm to 50 ohm.Math.cm, said membrane being suspended above a substrate integrating a signal for reading out the signal generated by said elementary detectors and for sequentially addressing the elementary detectors. The detector includes at least one getter intended to ensure the trapping of residual gas during and after the forming of the detector, and includes a hermetically-sealed cavity having said array and said at least one getter housed therein, having an upper cap including a window transparent to infrared radiation, said cap being sealed by means of a seal on a chip supporting the array of elementary detectors or on a package at the bottom of which the chip supporting the array of elementary detectors has been attached, said cavity being under vacuum or a low pressure.

Claims

1. An infrared radiation detector comprising: an array of elementary imaging bolometric detectors, at least one getter intended to ensure the trapping of residual gas during and after the forming of the detector; a hermetically-sealed cavity having said array and said at least one getter housed therein, which cavity has an upper cap comprising a window transparent to infrared radiation; in which: each of the elementary bolometric detectors is formed of a bolometric membrane comprising a film made of amorphous vanadium oxide VO.sub.x, x being comprised between 1.8 and 2.3, said membrane being suspended above a substrate integrating a signal for reading out the signal generated by said elementary detectors and for sequentially addressing the elementary detectors, said cap is sealed by means of a seal on a chip supporting the array of elementary detectors or on a package at the bottom of which the chip supporting the array of elementary detectors has been attached, said cavity being under vacuum or a low pressure.

2. The infrared radiation detector of claim 1, wherein the seal is made of a AuSn metal alloy.

3. A method of forming an infrared radiation detector, comprising the steps of: installing an array of elementary imaging bolometric detectors and at least one getter intended to ensure the trapping of residual gas during and after the forming of the detector in said cavity, each of the elementary bolometric detectors being formed of a bolometric membrane comprising a film made of amorphous vanadium oxide VO.sub.x, x being comprised between 1.8 and 2.3, said membrane being suspended above a substrate integrating a signal for reading out the signal generated by said elementary detectors and for sequentially addressing the elementary detectors, hermetically sealing said cavity with a upper cap comprising a window transparent to infrared radiation by means of a seal on a chip supporting the array of elementary detectors or on a package at the bottom of which the chip supporting the array of elementary detectors has been attached, said cavity being under vacuum or a low pressure; in which method: the sealing of the upper cap of the cavity thereon, after the elementary imaging bolometric detectors and the getter have been installed therein, is performed at a temperature in the range from 280° C. to 320° C. for a duration in the range from 10 to 90 minutes; and the activation of said getter is performed concurrently to the sealing of the upper cap of the cavity.

4. The infrared radiation detector forming method of claim 3, wherein the film made of vanadium oxide VO.sub.x used to form the bolometric membrane has a resistivity measured at 30° C. in the range from 6 ohm.Math.cm to 24 ohm.Math.cm.

5. The infrared radiation detector forming method of claim 4, wherein the film made of vanadium oxide VO.sub.x used to form the bolometric membrane has a resistivity measured at 30° C. in the range from 6 ohm.Math.cm to 9 ohm.Math.cm, and wherein the sealing of the upper cap of the cavity thereon, after the elementary detectors and the getter have been installed therein, is performed at a temperature in the range from 280° C. and 300° for a duration in the range from 10 to 90 minutes.

6. The infrared radiation detector forming method of claim 4, wherein the film made of vanadium oxide VO.sub.x used to form the bolometric membrane has a resistivity measured at 30° C. in the range from 9 ohm.Math.cm to 24 ohm.Math.cm, and wherein the sealing of the upper cap of the cavity thereon, after the elementary detectors and the getter have been installed therein, is performed at a temperature in the range from 280° C. and 320° C. for a duration in the range from 10 to 90 minutes.

7. The infrared radiation detector forming method of claim 3, wherein the seal sealing the upper cap to the cavity is made of AuSn alloy.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The foregoing features and advantages of the presently disclosed embodiments will now be discussed in the following non-limiting description of a specific embodiment, in relation with the accompanying drawings.

(2) FIG. 1 is a simplified representation of a bolometric pixel or elementary bolometric detector.

(3) FIG. 2 is a simplified representation illustrating the principle of the forming of a vanadium oxide film by IBS (acronym for “Ion Beam Sputtering”), which is the preferred technique in the state of the art.

(4) FIG. 3 shows two histograms showing the dispersion of the output signal of an array infrared radiation detector, respectively: for the case of a material which has not evolved during its integration into the detector (upper portion), and for the prior art case of a VOx material (lower portion) deteriorated by the thermal treatment.

(5) FIGS. 4 and 5 are graphs representative of the variation of the resistivity of films of vanadium oxide VOx, for different intrinsic resistivities measured at 30° C., according to the anneal temperature, respectively for anneal durations of 90 and of 10 minutes.

DETAILED DESCRIPTION

(6) An elementary bolometric detector has been shown in FIG. 1. Such a detector is basically formed of a membrane (1) suspended via thermal insulation “arms” (2) and pillars (3), ensuring the electric connection with the substrate (4).

(7) The membrane (1) comprises a thin film of thermistor material (5) on the most part of its surface, oriented opposite a window transparent to infrared radiation (and typically made of silicon or of germanium).

(8) Advantageously, and to optimize the performance of the elementary detector, a metallic reflector film (6) is affixed under the suspended membrane and at an adequate distance therefrom, to form a resonating cavity and thus optimize the absorption of the infrared radiation.

(9) The thermistor material (5) is made of a thin film of vanadium oxide VOx, having a typically thickness in the range from 20 to 200 nanometers and having a resistivity in the range from 6 ohm.Math.cm to 50 ohm.Math.cm. These resistivity values typically correspond to a value of x in the range from 1.8 to 2.3, such as measured by the RBS (acronym for “Rutherford Backscattering Spectroscopy”) technique.

(10) The thin film of vanadium oxide VOx is formed by IBS (acronym for “Ion Beam Sputtering”) deposition on a substrate at ambient temperature in a reactor in the presence of oxygen, with a partial pressure in the range from 3×10.sup.−5 Torr to 1×10.sup.−4 Torr.

(11) Such a reactor is illustrated in FIG. 2. This reactor (10), provided with a pumping system (11), comprises a support (12) receiving a target (13) of pure or almost pure vanadium. The target is bombarded by a beam of ionized krypton emitted by a gun (14) known per se. The use of krypton is not limiting, any other rare gas may be used.

(12) The vanadium atoms ejected from the target as a result of this ion bombarding are more or less oxidized by the oxygen (15) introduced into the reactor. The partial pressure of oxygen present in the reactor chamber is controlled by means of a regulation loop to set the final quantity x to the desired value. The sputtered vanadium atoms oxidize to form on a substrate (16) a layer of VOx having a resistivity depending on proportion x of atomic oxygen.

(13) The substrate (16) is kept at a temperature close to the room temperature during the deposition by a cooling system (not shown) using a heat-carrying fluid.

(14) To compare the thermal stability of VOx having an initial resistivity measured at 30° C. in the range from 0.5 to 24 ohm.Math.cm, the selected substrate is formed of a film of silicon nitride SiNx (or even of silicon oxide SiOx) deposited on a 200-millimeter wafer of single-crystal silicon. This provides an excellent electric insulation between the VOx film and the silicon substrate.

(15) A series of square patterns (“Van der Pauw”-type patterns) is then defined by photo-lithography and then etching of the VOx material. The material is then contacted at the four corners of this square by the deposition and then the definition of metal electrodes.

(16) The assembly is then encapsulated by a layer of silicon nitride SiNx deposited by PECVD (acronym for “Plasma-Enhanced Chemical Vapor Deposition”) at low temperature, that is, 280° C., to preserve the characteristics of the VOx material. The encapsulation is carried out so as to entirely cover the VOx patterns, to insulate the VOx material from any chemical interaction with the ambient atmosphere during anneal tests.

(17) The patterns are then biased via the metal electrodes. The resistance per square of the VOx film is then determined by the Van der Pauw method. The thickness of the VOx films is determined by ellipsometry on dedicated neighboring patterns. The thickness and the resistance per square of such films define their resistivity. This method has been applied before and after each anneal. The results appear on the graph of FIG. 4 for 90-minute anneals. It has been previously verified that the silicon nitride encapsulation layer has not modified the properties of the VOx film.

(18) Five wafers each integrating a VOx film, of respective resistivities of 0.5 ohm.Math.cm, 6.3 ohm.Math.cm, 9.3 ohm.Math.cm, 20 ohm.Math.cm, and 24 ohm.Math.cm measured at 30° C. have thus been formed. Such resistivities are correlated to the respective value of x (to within + or −0.1) of 1.6; 1.8; 1.9; 2, and 2.1.

(19) Samples from these wafers have then been submitted to anneals performed under a nitrogen flow and at different temperatures, staged between 240° C. and 330° C., to assess their thermal stability, or in other words their robustness in the presence of thermal stress. To deliberately be in the configuration which is the most constraining for the material, but the safest in terms of industrial reliability, particularly in terms of stability of the vacuum of the sealed components, the anneal time has been set to 90 minutes.

(20) The following observations can then be drawn from the curves of FIG. 4.

(21) The film having a 0.5-ohm.Math.cm initial resistivity sees its resistivity significantly drop as soon as the temperature reaches 250° C.

(22) The film having a 6.3-ohm.Math.cm initial resistivity is stable up to 300° C. and then sees its resistivity abruptly drop.

(23) However, the films having an initial resistivity of 9.3, 20, and 24 ohm.Math.cm have but a very small variation in terms of resistivity up to 310° C., said resistivity only dropping from 320° C.

(24) The stability threshold of the material thus appears to depend on its composition and on its intrinsic resistivity; the materials having the highest resistivities being the most stable.

(25) Such measurements have also been carried out on identical samples, but for anneals of short duration, typically 10 minutes, corresponding to a realistic minimum time for a vacuum sealing process.

(26) Thus, other devices integrating the same films of respective initial resistivities measured at 30° C. of 0.5 ohm.Math.cm, 6.3 ohm.Math.cm, and 20 ohm.Math.cm have been submitted to 10-minute anneals under a nitrogen flow, at different temperatures covering the range of interest, that is, 280° C., 300° C., 310° C., 320° C., and 330° C.

(27) The resistivity value measured after the anneal for each sample, plotted on the graph of FIG. 5, results in the following observations:

(28) The film having an initial resistivity of 0.5 ohm.Math.cm has not resisted the 280° C. anneal despite an anneal duration decreased to 10 minutes. This observation implies the incompatibility of such a material representative of the state of the art with the vacuum sealing process considered herein.

(29) The film having a 6.3-ohm.Math.cm initial resistivity keeps stable characteristics up to 320° C. at least for this short-time anneal (10 minutes).

(30) The film having an initial resistivity of 20 ohm.Math.cm withstands a 10-minute anneal up to 330° C.

(31) Thereby, by using elementary detectors integrating, as a thermistor material, vanadium oxide VOx, x being in the range from 1.8 to 2.3, characterized by a resistivity measured as previously indicated, that is, in the range from 6 ohm.Math.cm to 50 ohm.Math.cm at 30° C., it becomes possible to form infrared array detectors having high and lasting performances, despite the step of sealing the upper cap to the package and of getter activation, which as already indicated, is capable of occurring at at least 300° C. as soon as an appropriate metal seal (for example, of AuSn (80/20) type) and a getter material (for example, of the type commercialized by SAES under reference “Pagewafer”) are selected, at a relatively low activation temperature.

(32) The thermal stability of such thermistor materials in terms of resistivity results in a limited dispersion of the signal obtained at the output of the readout circuit, and enables to keep at the output of the analog amplifier a narrow distribution histogram contained at the center of the electric dynamic range, as illustrated in FIG. 3 (upper portion).

(33) A film of intermediate initial resistivity (measured at 30° C.), typically smaller than 6 ohm.Math.cm, is unstable in the considered temperature range, which results in a very widened histogram which now occupies the most part of the electric dynamic range (FIG. 3—lower portion). Such a material would require, if the high reliability level of the previously-indicated sealing method is desired to be kept, complex electronic correction systems to take each output signal of the pixels to the center of the readout circuit dynamic range.

(34) The film having the lowest electric resistivity, 0.5 ohm.Math.cm, undergoes such an evolution in the applied temperature range that it should be considered incompatible with this sealing technology. Its TCR is zero or too low, and accordingly of no interest. Such a material would require using low-temperature sealing alloys, which do not enable to use efficient getters, which jeopardizes the durability of the vacuum and thus the reliability of the components formed by this method.

(35) The Applicants have also observed that a VOx film having a resistivity close to 1 ohm.Math.cm has faster degradation kinetics under an ordinary atmosphere than VOx films described in the present disclosure.

(36) Indeed, it has been described (see for example N. J. Podraza et al. —“Electrical and optical properties of sputtered amorphous vanadium oxide thin films”—Journal of Applied physics 111 n.sup.o 7—Apr. 1, 2012, or also M. A. Motyka et al. —“Microstructural evolution of thin film vanadium oxide prepared by pulsed-direct current magnetron sputtering”—Journal of Applied Physics 112, no 9 (2012)) that a layer of lower density (often called “roughness layer”) forms at the surface of VOx films exposed to ambient air. Measurements by ellipsometry interpreted by multilayer models, that is, models which consider, as a minimum, the film as being formed of two layers, one being integral, that is, made at 100% of VOx, and the other being made of 50% of VOx and 50% of air (according to the “Bruggeman effective medium approximation”) enable to determine the thickness of this “roughness layer”. The follow-up of the thickness of this layer along time, for an exposure to air and in the same temperature and humidity conditions, shows growth kinetics approximately twice faster in the case of a 1-ohm.Math.cm VOx film than for a 10-ohm.Math.cm film.

(37) As an example, 40 days after the deposition of the VOx films, which have a total thickness of 80 nanometers, the thickness of the roughness layer remains smaller than 5 nanometers for a VOx film having a 10-ohm.Math.cm resistivity but however exceeds 10 nanometers for a VOx film having a 1-ohm.Math.cm resistivity.

(38) This growth occurs to the detriment of the integral VOx layer. It has been shown (for example, in the two previously-mentioned publications) that this surface layer has a higher resistivity than the integral film, which necessarily modifies the resistance per square of films along time.

(39) Further, the presence of such a surface layer, of low density and of variable thickness, may cause problems of variability of the photolithography and etch methods, necessary to define the VOx patterns. These two difficulties are likely to deteriorate the final detector performances, either by increasing the electric resistance of the pixels beyond the desired value, or by increasing the dispersion of these values. It is however, possible to do away with an in-situ encapsulation after deposition, that is, with no contact with air, of the VOx films, such as advocated in literature (see U.S. Pat. No. 6,313,463), as soon as the thermistor films are used.

(40) The disclosed embodiments result in the possibility of making compatible the performance requirements of uncooled array-type infrared detectors and the constraints of the industrialization of the manufacturing of such detectors.

(41) Further, such detectors are easier to handle, during their manufacturing process, due to their greater capacity of resisting oxidation in ambient air.