RU2595306C1 - Heat radiation sensor and its manufacturing method - Google Patents

Abstract

FIELD: heat engineering.

SUBSTANCE: invention can be used for heat insulation of heat radiation detectors. Core of invention is that a device for thermal detection of infrared radiation includes a pixel on a semiconductor substrate, the pixel includes the first section and the second section, the first section is located onto the surface of the semiconductor substrate and includes electric circuits, the second section is separated from the first section and is located immediately above it, the second section is planar and includes legs, a micro-membrane and a temperature detector located on it, the second section is supported by columns, one of the legs has one end integrally connected with the micro-membrane and the other end integrally connected to one of the columns, the other leg has one end integrally connected with the micro-membrane and the other end integrally connected to the other of the columns, legs provide electrical connection of the temperature detector with electric circuits via appropriate columns and thermal insulation of the temperature detector and the micro-membrane from the semiconductor substrate, one of the legs includes the first part of the first dielectric layer, the first part of the second dielectric layer, part of the electroconductive layer, this part of the electroconductive layer provides the above said electrical connection, the first part of the first dielectric layer adjoins the first surface of the electroconductive layer and the first part of the second dielectric layer adjoins the second surface of the electroconductive layer, the first and the second surfaces of the electroconductive layer are opposed surfaces of the part of the electroconductive layer, part of the electroconductive layer is a source of mechanical stresses inducing tension stresses in the first part of the first dielectric layer and tension stresses in the first part of the second dielectric layer.

EFFECT: technical result is the possibility of reduction of heat conductivity of the dielectric layers.

20 cl, 24 dwg, 2 tbl

Description

The invention relates to samples for recording thermal radiation and methods for manufacturing such devices. These devices have heat radiation detectors with thermal insulation. One of the many innovative aspects of this invention are new structures for the thermal insulation of thermal radiation detectors.

The structure for detecting thermal radiation can be implemented using an ultra-sensitive thermometer for recording and measuring thermal radiation. Such structures can be used, in particular, for detecting and measuring infrared (IR) radiation. The structures for detecting IR radiation are a temperature detector thermally coupled to an IR absorber, in turn, the temperature detector and absorber are thermally insulated from the substrate on which electrical devices are formed for reading and processing the values of thermal radiation absorbed by the temperature detector. The sensitivity of such IR detectors is determined, among other factors, by the quality of thermal insulation of the temperature detector. If the thermal insulation of the temperature detector is insufficient, then heating the temperature detector with thermal radiation is also insufficient, which, in turn, makes it impossible to detect changes in the properties of the temperature detector by electronic devices caused by thermal radiation. Typically, a layer of material with a high coefficient of thermal resistance (CTC) is used as a temperature detector. This CCC is usually expressed as a percentage change in resistance per degree of temperature. At the moment, the best materials, such as vanadium oxide or amorphous silicon, have a change in CCC of a few percent per degree of temperature, which is sufficient for the detection and registration of infrared radiation. Thermal radiation detectors can have a more complex structure. For example, semiconductor pn diodes can be used as IR radiation detectors. In this case, the detection and measurement of infrared radiation occurs on the basis of changes in the current-voltage characteristics of the diodes caused by their heating.

Thermal insulation of the temperature detector (and absorber) from the substrate on which the detector is made, as a rule, is carried out using the so-called legs. The legs are elongated structural elements of the detector, which eliminates the direct physical contact of the temperature detector (and absorber) with the substrate. As a rule, legs in combination with columns allow the detector (and absorber) to be suspended above the substrate surface. The elongated shape of the legs allows for low thermal conductivity of this element and to increase the thermal insulation of the temperature detector from the substrate. One end of each leg is connected to a temperature detector or a membrane on which the temperature detector is located, the other end of each leg is connected to a corresponding column that suspends the temperature detector above the surface of the substrate. Patent EP 1212592 B1 describes a structure where a temperature detector located on a micro-membrane is suspended above the substrate by legs and columns. The legs in this patent are elongated and are made using vanadium oxide and silicon nitride. Thermal insulation of the temperature detector is achieved in this patent due to the length of the legs.

To understand this invention, it is necessary to introduce the following definitions of physical quantities from the laws used to describe the invention.

Young's modulus, denoted as E (modulus of longitudinal elasticity), is a physical quantity characterizing the properties of a material to resist tension / compression during elastic deformation.

Hooke's law is a statement according to which the deformation arising in an elastic body (spring, rod, cantilever, beam, etc.) is proportional to the force applied to this body. Hooke's law can be described by the following formula: σ = ε * E, where σ is the normal stress in the cross section equal to the force F applied to the cross section of the elastic body divided by the cross section S, ε is the relative elongation equal to the elongation of the elastic body Δl in the direction and under the action of the applied force divided by the length of the elastic body l in the direction of the applied force.

The Poisson's ratio, denoted by µ, is the ratio of the relative transverse compression to the relative longitudinal tension.

The biaxial module, denoted as M, is a value calculated by the formula M = E / (1-µ).

IR radiation is electromagnetic radiation in the wavelength range 0.7 µm - 1000 µm.

The absorption of IR radiation in a layer of material, the absorption coefficient of which is independent of thickness, is described by the Lambert – Behr law: I _d = I ₀ e ^-kd , where I ₀ is the intensity of the radiation entering the material layer, I _d is the radiation intensity after passage of a layer of thickness d, k is the absorption coefficient ¹ . If the absorption properties of the material do not depend on the thickness of the layer made of this material, then the absorption coefficient depends only on the wavelength and does not depend on the thickness of the layer.

Under the metal nitride in the description of this invention refers to both single-phase and multiphase compounds having both stoichiometric and non-stoichiometric composition.

Silicon nitride, oxide, oxy nitride in the description of this invention refers to both hydrogenated and non-hydrogenated silicon compounds having an amorphous structure and having both a stoichiometric and non-stoichiometric composition.

The structure for detecting infrared radiation includes at least one temperature detector on a micro-membrane located above the surface of the substrate. The micro-membrane is suspended above the surface of the legs through the legs and columns. The legs are an elongated element that has low thermal conductivity to provide thermal insulation of the micro-membrane and the temperature detector located on it from the substrate. One end of each leg is connected to the micro-membrane, the other end of the leg is connected to the corresponding column, providing a gap between the surface of the substrate and the micro-membrane. The legs together with the columns provide electrical connection with electronic components (circuits) on the substrate, designed to read data from a temperature detector. To improve the sensitivity of the temperature detector of IR radiation, an IR absorber thermally coupled to a temperature detector of IR radiation can be used. IR radiation can be absorbed by both the absorber and the temperature detector itself and / or the micro-membrane material. In addition, the surface of the substrate under the micro-membrane can be coated with a material that reflects IR radiation to increase the percentage of absorption of IR radiation converted by the temperature detector. The micro-membrane, the legs integrally connected to it, the temperature detector located on it form a planar structure. This planar structure may also include an absorber.

Such a structure can be fabricated using thin-film technologies widely used in the semiconductor industry. As a rule, to simplify the process, one layer of material is structured so that different parts of this layer are used in different parts of the product. For example, one dielectric layer can be used to form not only a micro-membrane but also legs. In addition, the properly selected material of the micro-membrane allows it to function as an absorber. Materials such as silicon oxide, nitride or hydroxy nitride have absorption peaks in the IR range. For example, the vibrational modes of Si-N, Si-H, and N-H chemical bonds cause the corresponding absorption peaks at wavelengths of 11.76 μm, 4.5 μm, and 2.99 μm, respectively. However, the use of one layer in various structural elements imposes many requirements on the properties of layers. For example, if one dielectric layer is used to produce both a micro-membrane and legs, and, in addition, this layer should function as an absorber, then the internal mechanical stresses in this layer should be optimized so as not to cause bending of the legs and micro-membrane in addition, the absorption of the layer used to make the micro-membrane in the working IR range should be high, in addition, the thermal conductivity of the same layer used for the manufacture of legs should be low to minimize heat ovyh loss through the legs.

The solution to such a multi-purpose problem of optimizing the properties of layers can be, if not impossible, then very difficult. One aspect of the present invention is to reduce the thermal conductivity of the dielectric layers used in the legs by applying tensile stresses to them. Reducing the thermal conductivity of a three-layer leg, where the electrically conductive layer is located between two layers of dielectrics, can be achieved when the electrically conductive layer is made of a rigid material having compression stresses. In this case, the electrically conductive layer causes tensile stresses in the dielectric layers, which, in turn, reduce the thermal conductivity of the dielectric layers. This effect makes it possible to simplify the optimization problem, for example, if the dielectric layer is used to make legs and a micro-membrane, in addition, it must also have certain absorption properties in the infrared range, then using the aforementioned technique to reduce the thermal conductivity of the dielectric layers allows optimizing the deposition parameters of the dielectric layers First of all, in order to obtain optimal absorption in the IR range and to prevent the bending of the micro-membrane. In addition, regardless of optimization problems, the aforementioned technique for applying tensile stresses to dielectric layers allows one to improve the parameters of detectors (i.e., reduce the thermal conductivity of legs) manufactured using already optimized processes. To do this, it is necessary to optimize the internal stresses in the electrically conductive layer so that it causes tensile stresses in adjacent dielectric layers.

In turn, the task of optimizing the manufacture of structures for detecting IR radiation may include the following steps. First, a library of technological processes is created for applying the layers used for the manufacture of structures. This library includes not only the technological parameters of application, but also the properties of the layers that were deposited as a result of using these technological processes. The properties of the layers in the library include properties that are additively taken into account when analyzing the characteristics of structures for detecting IR radiation. For example, knowledge of the refractive indices and absorption of materials allows us to calculate the absorption spectrum of a micro-membrane and / or absorber, consisting of several layers of materials whose properties are listed in the library along with the corresponding spraying parameters. In addition, the mechanical properties of the layers, such as the values of biaxial modules and internal stresses, can be stored in the library. These parameters, in turn, allow you to select the optimal values of the thicknesses of the layers to obtain the aforementioned effect of reducing the thermal conductivity of the legs, which also avoid unacceptable deformation of the micromembrane.

After creating the library, layers are selected that, according to calculation / mathematical modeling, can provide the specified characteristics of one or more structural elements of structures for detecting infrared radiation. One or more structural elements with the aforementioned characteristics make it possible to fabricate a structure for detecting infrared radiation with desired characteristics. For example, the previously measured absorption and refraction coefficients of individual layers make it possible to simulate the absorption of a multilayer structure consisting of several layers, i.e. absorption of IR radiation in the structure for detecting IR radiation. The selection of layers can also be carried out taking into account maximization of the above effect of reducing the thermal conductivity of the legs. Knowledge of the mechanical properties of the individual layers used for the manufacture of legs allows one to analytically solve the multi-purpose optimization problem, according to which the selected layers for the manufacture of legs should ensure that the legs are not unacceptably deformed and at the same time maximize the aforementioned effect of reducing the thermal conductivity of the legs (i.e. use an electrically conductive layer used to form legs, to create tensile stresses in the dielectric layers adjacent to it). In addition, the task of multi-purpose optimization may include the task of obtaining the required absorption in the framework of the above absorption modeling, if at least one layer is used to form both legs and structural elements of the structure for detecting IR radiation, which absorb IR radiation and transmit absorbed IR radiation temperature detector (e.g. micro-membrane and / or absorber). Subsequently, structures are made and their characteristics measured. At this stage, not only the characteristics that were previously modeled on the basis of the properties of the materials (e.g., absorption spectrum) are measured, but also the characteristics that are determined on the basis of complex effects due to the mutual influence of the processes occurring in the layers (e.g. thermal conductivity of the legs). The thermal conductivity of the legs consisting of several layers, among other factors, is determined by the mean free path of the phonons that transfer heat, which is comparable with the size of the cross section of the legs, the scattering of phonons at the boundaries of the layers, and the aforementioned effect of reducing thermal conductivity in dielectric layers. Also at this stage, the measurement of the integral characteristics of the structure for detecting IR radiation, such as, for example, sensitivity in the operating range, can be carried out.

If, as a result of measurements, it is established that the characteristics of the structures do not correspond to the specified ones, then a new combination of layers is selected and / or the thicknesses of the already selected layers are optimized. In the future, new structures are made taking into account the changes made.

The process of such iterative optimization can be repeated several times until the measured characteristics of the newly fabricated structures correspond to the specified values (for example, will not be in the given intervals) or until this iterative process reaches saturation, i.e. changes (improvements) in one or more characteristics resulting from the next iteration will be insignificant (for example, when the relative change in one or more characteristics is less than a certain value).

The electrically conductive layer located between the layers of dielectrics in the legs can be made of any electrically conductive hard, stable material that can cause tensile stresses in adjacent dielectric layers, which lead to a decrease in their thermal conductivity. Suitable materials may be nitrides of transition metals of the fourth, fifth or sixth groups of the Periodic system of chemical elements, such as titanium, vanadium, chromium, zirconium, niobium, molybdenum, hafnium, tantalum or tungsten. The electrically conductive layer may also be made of solid solutions or alloys of these metals or their nitrides. In addition, the electrically conductive layer may be multilayer, i.e. consist of several electrically conductive layers. The electrically conductive layer can be deposited by magnetron sputtering, by plasma-chemical gas-phase deposition, or by sputtering a target with an ion beam. As a source of an ion beam, a Kauffman source can be used. Metal nitrides can be sprayed using targets from the corresponding metal nitrides or by reactive spraying. In the case of reactive sputtering using an ion beam, nitrogen can be injected directly into the vacuum chamber where the sputtering is carried out, and / or an additional ion beam source can be used to nitrate the surface of the grown layer.

The dielectric layers can be made, for example, of silicon nitride, oxide or hydroxy nitride. Each of the dielectric layers can be multilayer, i.e. consist of several dielectric layers. The dielectric layers can be deposited by magnetron sputtering, by plasma-chemical vapor-phase deposition, or by sputtering a target with an ion beam. An auxiliary electron source may be used to neutralize the charge on the surface of the sputtered target and / or on the surface of the substrate to be sputtered. Silicon nitrides, oxides and oxy-nitrides can be deposited using targets from the corresponding silicon nitrides, oxides and oxy-nitrides or by reactive spraying. In the case of reactive sputtering using an ion beam, nitrogen and / or oxygen can be injected into the vacuum chamber where the sputtering is carried out, and / or an additional ion beam source can be used to nitrate and / or oxidize the surface of the grown layer.

To create tensile stresses in the layers of dielectrics adjacent to the electrically conductive layer in the legs, the following process can be used: formation of a sacrificial layer on a substrate; the formation of legs on the sacrificial layer; removal of the sacrificial layer. The formation of the legs includes: sequentially forming one of the dielectric layers on the sacrificial layer, forming an electrically conductive layer having a compressive stress on the formed dielectric layer, and forming another dielectric layer on the electrically conductive layer. With this manufacture of legs, the electrically conductive layer causes tensile stresses in adjacent dielectric layers after removal of the sacrificial layer. For deposition of dielectric layers, it is most preferable to use spraying processes to obtain unstressed layers of dielectrics on a substrate.

The mechanical properties of the layers used to make the legs can satisfy the following rules, which optimize the legs so that their bending is insignificant and / or the planarity and parallelism of the micro-membrane relative to the substrate surface are not violated:

1. | M ₁ * d ₁ -M ₂ * d ₂ | / (M ₁ * d ₁ + M ₂ * d ₂ ) <0.1, where "| |" module calculation operation, i.e. absolute value of a quantity; M ₁ and d ₁ are the biaxial module and the thickness of the first dielectric layer adjacent to one surface of the electrically conductive layer of the legs; M ₂ and d ₂ are a biaxial module and the thickness of the second dielectric layer adjacent to the opposite surface of the electrically conductive layer of the legs. In some cases, a more stringent criterion is preferable: 1 M ₁ * d ₁ -M ₂ * d ₂ 1 / (M ₁ * d ₁ + M ₂ * d ₂ ) <0.05. These criteria are applicable when there are no internal stresses in the first and second dielectric layers immediately after sputtering, and until the moment during which the internal compression stresses in the electrically conductive layer cause tensile stresses in the fragments of the first and second dielectric layers used to form the legs during pixel manufacturing . In addition, these criteria are applicable when there is no gradient of internal stresses across the thickness of the conductive layer in the fragments of the electrically conductive layer used to make the legs.

2. M ₁ * d ₁ > M ₂ * d ₂ , this criterion is applicable when the internal compressive stresses in the electrically conductive layer change substantially along the thickness of the layer, when the absolute value of the compressive stresses in the electrically conductive layer at the interface with the first dielectric layer having a biaxial module M ₁ and a thickness d ₁ greater than the absolute value of the compression stresses in the electrically conductive layer at the interface with the second dielectric layer having a biaxial module M ₂ and a thickness d ₂ . The aforementioned criterion can also be formulated for the “mirror case” when the absolute value of the compressive stresses in the electrically conductive layer at the boundary with the first dielectric layer is less than the absolute value of the compressive stresses in the electrically conductive layer at the boundary with the second dielectric layer. In this case, M ₁ * d ₁ <M ₂ * d ₂ . The criteria of paragraph 2 are applicable when there are no internal stresses in the first and second dielectric layers immediately after sputtering, and until the moment during which the internal compression stresses in the electrically conductive layer cause tensile stresses in the fragments of the first and second dielectric layers used to form legs.

Layers without gradients of internal stresses can be obtained as a result of the following process of optimization of the deposition process. First, at least two layers of different thicknesses are applied. The thickness of one layer is nominal, i.e. equal to the thickness of the layer that is used to make the structure for detecting IR radiation. The remaining layers, hereinafter referred to as test layers, have thicknesses that are less than the thickness of the nominal salt. For example, you can use at least one test layer having a thickness in the range of 30-40% of the nominal thickness. If the values of internal stresses measured using the Stoney formula ² in the layer and all test layers do not lie in a predetermined interval (for example, a preset value of +/- 5%), then the process of applying the layers is optimized, as a result of which one or more tuning technological parameters are selected deposition, which change during the deposition of layers in such a way that the values of internal stresses in the layer and all test layers lie in a given interval. For example, as the tuning technological parameters, you can use the frequency of the plasma discharge or the ratio of the periods of the excitation of the plasma at different frequencies for processes of plasma-chemical gas-phase deposition (CVD) deposition of layers ³ . In turn, for processes of gas-phase deposition of layers (PVD), the intensity of ion bombardment of the surface of the sprayed layer (ion assistance) can be used as a tuning parameter.

The implementation of this invention and the advantages that arise from it will become more clearly apparent as a result of consideration of illustrative examples, which follow below, are given for information only, but do not mean limitations, supported by the attached drawings.

In the drawings:

FIG. 1a - 13a illustrate a cross-section of a column on a substrate at various stages of fabrication of a structure;

FIG. 1b to 13b illustrate a cross section of a micro-membrane, a temperature detector, an absorber, legs, and a substrate at various stages of fabrication of the structure;

FIG. 14 illustrates a top view of a pixel equipped with a temperature detector;

FIG. 15 illustrates a three-dimensional image of a pixel equipped with a temperature detector;

FIG. 16 illustrates the internal stresses and values of a biaxial module in silicon nitride layers as a function of film thicknesses;

FIG. 17 illustrates the internal stresses and values of the biaxial module in titanium nitride layers as a function of film thicknesses;

FIG. 18 illustrates the internal stresses and values of the biaxial module in titanium nitride layers as a function of film thicknesses;

FIG. 19a and 19b illustrate a test structure for detecting a gradient of internal stresses in layers of various materials;

FIG. 20a and 20c illustrate test structures after removal of the sacrificial layer, for films with different thickness gradients of internal stresses;

FIG. 21a and 21b illustrate a test structure for measuring biaxial modules of materials;

FIG. 22 illustrates a flowchart for optimizing pixel manufacturing;

FIG. 23 illustrates a pixel array;

FIG. 24 illustrates a pixel matrix.

FIG. 1a - 13a and FIG. 1b to 13b illustrate cross-sections of a column, legs, micro-membrane, temperature detector, substrate, and absorber during the manufacturing process of a pixel equipped with a temperature detector. As will be shown below, pixels can be combined into a matrix of pixels, which allows to obtain an image in the infrared range. At the initial stage, electronic circuits and logic transistor elements 102 are formed on the substrate 100 to read the state of the temperature detector. In the case where a vanadium oxide layer with a high temperature coefficient of electrical resistance is used as a temperature detector, the electronic circuits and logic transistor elements are configured to measure the electrical resistance of the vanadium oxide layer. In FIG. 1a also shows an aluminum contact 101 of electrical circuits on which a micro-membrane supporting column will subsequently be formed. As will be further illustrated, the electrical contact of the temperature detector with electrical circuits is through an aluminum contact 101 and a column formed thereon.

FIG. 2a and FIG. 2b illustrate the next step in manufacturing a pixel: applying and structuring an IR reflector layer. Part of the layer 103b performs the function of an IR reflector, another part of the layer 103a performs the protective function of the aluminum contact 101, and it can also improve the adhesion of the materials from which the column is made. This layer can be made of various metals having good reflectivity in the IR range. For example, this layer may be made of titanium, nickel, chromium or a nickel-chromium alloy. Also, this layer may be multilayer and may consist of several metal layers. This layer can be deposited by magnetron sputtering, by sputtering a target with an ion beam, or by thermal evaporation.

FIG. 3a and FIG. 3b illustrates a further step in manufacturing a pixel. At this stage, a sacrificial layer 104 is formed and structured. It can be structured through a mask from the structured layer 105 by plasma-chemical etching in oxygen or oxygen-hydrogen plasma. As a result of the structuring of the sacrificial layer, a window is formed above the contact 101. The bottom of the formed window opens the contact surface 101, if the element 103a has not been formed, or the surface of the element 103a, if it was formed.

As the sacrificial layer, organic materials such as polyimides, photoresists, or inorganic materials based on low-temperature glasses obtained by plasma-chemical gas-phase deposition or centrifugation from solutions can be used. Layer 105 can be removed after structuring the sacrificial layer 104.

If the layer 105 is not removed after structuring the sacrificial layer, then parts of it can be used to make legs and / or micro-membranes.

As an alternative, a photosensitive sacrificial layer can be used. In this case, the process of structuring such a layer does not require the use of a mask 105. The structuring of the photosensitive layer is also carried out as the structuring of the photoresist, i.e. flash method with subsequent development and heat treatment.

Depending on the processing of the sacrificial layer and / or the process of structuring the sacrificial layer, the walls of the formed window may be vertical or, as shown in FIG. 3a, inclined. Inclined walls can be obtained by developing a photosensitive sacrificial layer, followed by heat treatment or by optimizing the etching process.

The thickness of the sacrificial layer is determined by the IR range in which the pixel should work. For example, if a pixel is designed to operate in the IR range of 8-12 μm, then the thickness of the sacrificial layer above the reflector 103b is selected in the range of 1.7-2 μm. The best option is the thickness of the sacrificial layer above the reflector, equal to a quarter of the wavelength, which is the arithmetic mean of the wavelengths, which are the boundary wavelengths of the working infrared range of the device.

After structuring the sacrificial layer 104, as shown in FIG. 3a and FIG. 3b, the dielectric layer 106 is sprayed as shown in FIG. 4a and FIG. 4b. As will be shown later, various parts of this layer after structuring will be used for the manufacture of legs and micro-membranes. Alternatively, this layer can be used to make only micro-membranes or legs. For this, layer 106 must be appropriately structured. If this layer is used only for the manufacture of a micro-membrane, then part of this layer should be removed in the region of the structure, the cross section of which is shown in FIG. 4a. If this layer is used only for the manufacture of legs, then part of this layer should be removed in the region of the structure, the cross section of which is shown in FIG. 4b.

An illustration of the next manufacturing step is shown in FIG. 5a and 5b. A layer of a temperature detector 108 and a layer of a protective dielectric 109 are applied to the mask from photoresist 107. The layer 109 is not an obligatory element of the structure. In certain embodiments of the structure, only layer 108 can be used. Layer 108 can be made of vanadium oxide VO _x , where 1.7 <x <1.9. Layers with such a chemical composition have a negative coefficient of change in electrical resistance in the range of 1.7-2.6% per degree Kelvin. The vanadium oxide layer can be sprayed by reactive ion-beam sputtering of a vanadium target, by ion sputtering of a vanadium target using an auxiliary ion beam of oxygen ions directed toward the surface of the grown film, or by reactive pulsed magnetron sputtering. Subsequently, the layers 108 and 109 are structured by explosive photolithography, as shown in FIG. 6a and 6b. As a result, a structure is formed from parts of layers 108 and 109 located one above the other, or only from part of layer 108, on layer 106 or 105 (depending on the options of the technological route described above). This structure is located above layer 103b. Alternatively, layers 109 and 108, or if layer 109 is not used, only layer 108 can be structured by etching through a photoresist mask. Etching can be carried out by the ion beam method. Layer 108 can be applied by reactive magnetron sputtering using a vanadium target in an argon and oxygen atmosphere or by reactive ion-beam spraying using a vanadium target and an ion beam of argon and oxygen ions to sputter a vanadium target. The thickness of the layer 108 used to form the temperature detector may be in the range of 50-150 nm.

In FIG. 7a and 7b illustrate the formation of electrical contacts to the layer of the thermal detector 109. The contacts can be made of any metal, for example vanadium, which provides ohmic contact to the layer 108. To effectively use the area of the structure of the thermal detector, a pair of contacts 119 is located near the opposite sides of the structure 108, as shown in FIG. 7b. If a layer 109 is used in the manufacturing process, then before forming the contacts 119, the windows for the contacts 119 are etched in the layer 109, as shown in FIG. 7b. The thickness of the layer used to form the contacts 119 may be in the range of 30-150 nm.

The next step of forming the structure is shown in FIG. 8a and 8b. As shown in FIG. 8a, a window is etched in the layer 106 above the contact 103a. This window subsequently provides the formation of an electrical contact between the contact 101 and the contact 119. The window opens a part of the surface of the element 103a, as shown in FIG. 8a. If the element 103a is not used, the window opens part of the contact surface 101. The formation of the window is performed if a layer 106 is used in the manufacturing process.

In FIG. 9a and 9b show the next step in the formation of the structure: deposition of an electrically conductive layer 112. This layer not only provides electrical contact between contacts 119 and 101, where contact 101 can be coated with a layer 103a. As will be shown later, part of this layer is used to form the legs, therefore, to realize the effect of reducing the thermal conductivity of the dielectric layers of the legs, this layer should be made of a rigid, stable material with a high internal compression stress, which, due to the properties of the material, does not decrease over time.

Suitable materials for the manufacture of layer 112 may include transition metal nitrides of the fourth, fifth or sixth groups of the Periodic System of Chemical Elements, such as titanium, vanadium, chromium, zirconium, niobium, molybdenum, hafnium, tantalum or tungsten. The electrically conductive layer may also be made of solid solutions or alloys of these metals or their nitrides. In addition, the electrically conductive layer may be multilayer, i.e. consist of several electrically conductive layers. These materials can be obtained by reactive magnetron sputtering in an atmosphere of nitrogen and argon using targets from the corresponding metals. The thickness of the layer 112 may be in the range of 10-100 nm.

In FIG. 10a and 10b show the next step in the formation of the structure: the formation of the supporting element 113 of the column 118. As shown in FIG. 10a, the material of this element 113 covers the bottom and walls of a window previously formed in the sacrificial layer. In addition, the material of the element 113 may form a cuff above the sacrificial layer around the perimeter of the window in the sacrificial layer. Element 113 may be made by magnetron sputtering of aluminum or other suitable metal, followed by etching through a photoresist mask. The thickness of the metal layer used to form the element 113 may be in the range of 0.3-1.5 μm.

In FIG. 11a and 11b show the next step in the formation of the structure: deposition of a dielectric layer 114. As shown in FIG. 11a and 11b, a dielectric covers the conductive layer 112 and the element of the supporting column 113.

In FIG. 12a and 12b show the next step in the formation of the structure: the formation of the structure of the legs 116. This step can be carried out by etching the layers 105, 106, 112, 114 through a mask of photoresist. As a result, slots 115 are formed separating the legs 116 from the micro-membrane 117.

In FIG. 13a and 13b show the next step in the formation of the structure: removal of the sacrificial layer 104. This step can be carried out by plasma-chemical etching in an oxygen atmosphere. As a result, a micro-membrane 117 and legs 116 are formed suspended above the surface of the substrate 100, or, depending on the pixel configuration, above the surface of the layer 102 including electronic circuits and logic transistor elements. The mechanical connection of the micromembrane with the substrate is carried out by legs 116 and columns 118, as shown in the top view of the pixel in FIG. 14. Elements 103a, 103b, as well as details of the surface relief of the micro-membrane, are not shown in FIG. 14 for greater clarity of image. In FIG. 14 shows the contours of structures 112a and 112b made of an electrically conductive layer 112, the contours of contact structures 119, and the contours of structured layers 108 and 109. FIG. 1a-13a illustrate the steps of forming pixel elements in section AA shown in FIG. FIG. 1b - 13b illustrate the steps of forming pixel elements in section BB shown in FIG. fourteen.

In FIG. 15 is a three-dimensional image of a pixel illustrating a micro-membrane 117 with a temperature detector. Thermal insulation of the micro-membranes is carried out using a pair of legs 116, which, in turn, are connected to a pair of columns 118 supporting all of these elements above the surface of the substrate 100 with electronic circuits, logical transistor elements 102 and a reflector 103b.

It is easy to see from the above description of the manufacturing process and FIG. 1a - 13a and FIG. 1b - 13b, various fragments of the dielectric layers 105, 106 and 114 can be used to form legs and / or micro-membranes. In addition, depending on the absorption properties of these layers in the working IR range of a pixel, fragments of these layers can function as an absorber. Also, fragments of the layers 109 and 108 can function as an absorber if they absorb radiation in the working infrared range of the pixel. For example, depending on the specific technological route, legs can be formed from fragments of the following groups of layers: 105, 106, 112, 114; or 106, 112, 114; or 105, 112, 114. In turn, the micro-membrane, depending on the specific technological route, can be formed from fragments of the following groups of dielectric layers: 105, 106, 114; or 106, 114; or 105, 114.

The dielectric layers 105, 106, 109 and 114 can be made of various dielectrics, such as silicon oxide, silicon nitride or hydroxy nitride silicon. These layers can be deposited, for example, by magnetron sputtering or by plasma-chemical gas-phase deposition. The thickness of these layers can be in the range of 50-200 nm. Each of the layers, in turn, may consist of one or more layers of dielectrics. The choice of layers, their thicknesses, deposition parameters can be selected from the considerations of optimal absorption in the wavelength range in which the IR radiation detector should work.

For example, the process of plasma-chemical gas-phase deposition of silicon oxide can be optimized by selecting the discharge power and the values of the gas flows SiH ₄ and N ₂ O. Depending on the different values of these technological parameters, various absorption values in the IR range can be obtained due to the molecular bonds Si-O, Si- H and Si-OH, having maxima at wavelengths of 9.6 μm, 4.4 μm and 3 μm, respectively. Similarly, absorption in the IR range for silicon nitride and silicon oxy nitride obtained by plasma-chemical gas-phase deposition can be optimized.

Internal stresses in the layers of the aforementioned dielectrics can also be optimized by optimizing the technological parameters of layer growth. For example, when applying silicon nitride in an atmosphere of SiH ₄ and NH _3, an increase in SiH ₄ flux or an increase in discharge power causes an increase in compression stresses, while an increase in gas pressure during the deposition process can reduce compression stresses and even obtain layers having tensile stresses. Another dynamics is shown by the process of deposition of silicon oxide by the method of plasma-chemical gas-phase deposition in the atmosphere of N ₂ O and SiH ₄ . An increase in SiH ₄ flux causes a shift in the internal stresses of the silicon oxide layers toward tensile stresses, while an increase in power causes the opposite effect. In the case of using magnetron sputtering, the internal stresses in the layers are also optimized by the selection of technological parameters. As a rule, layers deposited at low gas pressures have compressive stresses, and layers deposited at high pressures have tensile stresses.

Layers can also have gradients of internal stresses in thickness. Optimization of process parameters can minimize internal stress gradients. In the case of using plasma-chemical processes of deposition from the gas phase, the use of installations using, in addition to high-frequency (HF) power at a frequency of 13.56 MHz or other multiple frequencies, also low-frequency (LF) power at frequencies of 50-500 kHz, to minimize the gradients of internal stresses in the layers . More preferred for generating low-frequency power is the frequency range of 100-300 kHz. The use of LF power to excite the plasma allows us to obtain layers with compression stresses, while the use of HF power allows us to obtain layers with tensile stresses ³ . The ratio of the length of time periods when the plasma is alternately excited by RF or LF power allows you to control the internal stresses in the deposited layers. An increase in the percentage of time periods when the plasma is excited by LF power can reduce the values of internal stresses, i.e. reduce the values of tensile stresses, or increase the absolute values of the compressive stress, or go from tensile stresses to compressive stresses, etc. Changing the ratio of the time intervals of plasma excitation using LF power and HF power can be carried out in the process of layer growth, and, therefore, can be used to optimize the layer growth parameters to obtain layers without gradients of internal stresses in thickness. For example, if a layer grown in the process of plasma-chemical gas-phase deposition, during which all technological parameters remained unchanged, has a gradient of internal stresses, when the absolute value of the compression stresses in the part of the layer adjacent to the substrate is greater than the absolute value of the compression stresses in the surface part of the layer, then for To minimize the gradient of internal stresses over the thickness, it is necessary to increase the duration of the time periods when the plasma is excited by low-frequency power during the layer growth enshat duration time periods, when the plasma is excited by RF power. In another example, if a thin film grown in the process of plasma-chemical gas-phase deposition, during which all technological parameters remained unchanged, has a gradient of internal stresses, when the absolute value of the tensile stress in the part of the layer adjacent to the substrate is greater than the absolute value of the tensile stress in the surface parts of the layer, then to minimize the gradient of internal stresses over the thickness of the layer, it is necessary to increase the duration of the time periods when the film grows PCA is excited by RF power and decrease the duration of periods of time when the plasma is excited by the LF output.

A similar approach is applicable to minimize the gradients of internal stresses over the thickness in layers (thin films) obtained by magnetron sputtering. As a tuning parameter to minimize the gradients of internal stresses across the thickness in the layers (thin films), you can use the value of the RF power applied to the substrate holder. For example, if a thin film grown during magnetron sputtering, during which all technological parameters remained unchanged, has a gradient of internal stresses, when the absolute value of the compression stresses in the part of the film adjacent to the substrate is greater than the absolute value of the compression stresses in the surface part of the film, then, to minimize the gradient of internal stresses over the thickness, it is necessary to increase the RF power applied to the substrate holder during film growth. In another example, if a thin film grown during magnetron sputtering, during which all technological parameters remained unchanged, has a gradient of internal stresses when the absolute value of the compressive stresses in the part of the film adjacent to the substrate is less than the absolute value of the compressive stresses in the surface part film, in order to minimize the gradient of internal stresses over the thickness, it is necessary to reduce the RF power applied to the substrate holder during film growth.

Also, the aforementioned approach is applicable to minimize the gradients of internal stresses over the thickness in the layers (thin films) obtained by sputtering the target with an ion beam. In this case, another source of inert gas ions is used, which irradiates the surface of the sprayed layer. As a tuning parameter to minimize the gradients of internal stresses over the thickness in the layers (thin films), the intensity of the ion beam irradiating the surface of the sprayed layer can be used. For example, if a thin film grown during the deposition process, during which all technological parameters remained unchanged, has a gradient of internal stresses, when the absolute value of the compression stresses in the part of the film adjacent to the substrate is greater than the absolute value of the compression stresses in the surface part of the film, then, to minimize the gradient of internal stresses over the thickness, it is necessary to increase the intensity of the ion beam irradiating the surface of the sprayed layer during film growth. In another example, if a thin film grown during the deposition process, during which all technological parameters remained unchanged, has a gradient of internal stresses, when the absolute value of the compression stresses in the part adjacent to the substrate is less than the absolute value of the compression stresses in the surface part of the film , in order to minimize the gradient of internal stresses over the thickness, it is necessary to reduce the intensity of the ion beam irradiating the surface of the sprayed layer during film growth.

The application of these technologies to minimize the gradients of internal stresses across the thickness by adjusting the above parameters during film growth is illustrated in FIG. 16-18. Each point shown in these graphs corresponds to a film, where the thickness is plotted on the corresponding horizontal axis, and internal stresses are on the corresponding vertical axis. The values of internal stresses in the films were calculated using the Stoney formula ² based on measurements of the curvature of the substrate caused by internal stresses in the film. Thin specially prepared silicon wafers with a thickness of 100-200 μm and having a crystalline orientation of (100) were used as substrates.

Each point in FIG. 16 corresponds to a silicon nitride film grown by plasma-chemical gas-phase deposition in a reactor with parallel electrodes and capacitive plasma matching. The graph indicated by circles in FIG. 16 represents the internal stresses in silicon nitride films grown under the same technological parameters: discharge power of 100 watts, SiH ₄ stream 75 sccm (standard cubic centimeter per minute), NH ₃ 500 sccm stream, gas pressure during deposition 77 Pa , substrate temperature 160 ° C, the ratio of the time period when the plasma is excited by LF power (T _LF ) to the total period of time when the plasma is excited by LF (T _HF ) or LF (T _LF ) power is 0.5 (T _LF / (T _LF + T _HF ) = 0.5). As the graph clearly shows, the voltage indicated by the circles in thin films strongly depends on the thickness, i.e. in the films there is a gradient of internal stresses along the thickness. To eliminate the internal stress gradients, a change in the parameter T _LF / (T _LF + T _HF ) was used. As a result of optimization, a regime was selected in which the parameter T _LF / (T _LF + T _HF ) gradually decreased from 0.5 to 0.1 during the deposition of a film 200 nm thick. When thinner films were deposited, the parameter T _LF / (T _LF + T _HF ) gradually decreased from 0.5 to the corresponding value in the range 0.1-0.5. For example, during the deposition of a film 50 nm thick, the parameter T _LF / (T _LF + T _HF ) gradually decreased from 0.5 to 0.36. All films grown in this regime with tuning of the parameter T _LF / (T _LF + T _HF ) during the growth of the films have the same value of internal stresses close to zero, as the curve indicated by the triangles in FIG. 16.

In FIG. Figure 17 shows the values of internal stresses for titanium nitride films of various thicknesses grown by reactive magnetron sputtering of a titanium target in a glow discharge of a direct current plasma. The graph indicated by circles in FIG. 17 represents the internal stresses in titanium nitride films grown under the same technological parameters: DC discharge power of 100 watts, Ar 100 sccm stream, N ₂ 7 sccm stream, substrate holder temperature 20 ° C, pressure in the spraying chamber 0.7 Pa . As this graph clearly shows, the absolute value of compression stresses decreases with increasing film thickness. To eliminate the internal stress gradients, a change in the parameter of the RF power value applied to the substrate holder was used. As a result of optimization, a regime was selected in which the value of the RF power applied to the substrate holder gradually increased from 0 to 20 watts during the growth of a film 200 nm thick. When thinner films were sputtered, the value of the RF power applied to the substrate holder gradually increased from 0 to the corresponding value in the range of 0–20 watts. For example, during the deposition of a film with a thickness of 50 nm, the value of the RF power applied to the substrate holder gradually increased from 0 to 3 watts. All films grown in this mode with fine tuning of the RF power applied to the substrate holder during the film growth have the same value of internal stresses, as the curve indicated by the triangles in FIG. 17.

In FIG. Figure 18 shows the values of internal stresses for films of vanadium nitride of various thicknesses grown by reactive sputtering by an ion beam using two sources of ion flux. The first source was used to sputter a vanadium target with argon ions. The second source was used to nitrate the surface of the grown layer. Argon was poured into the first source, a mixture of argon and nitrogen was poured into the second source. The graph indicated by circles in FIG. 18 represents the internal stresses in the films of vanadium nitride grown under the same technological parameters: the ion current of the first source was 82 mA, the ion energy of the first source was 1400 eV, the argon inlet to the first source was 13 sccm, the ion current of the second source was 42 mA, the ion energy of the second source was 110 eV, the argon inlet to the second source was 2 sccm, the nitrogen inlet to the second source was 2 sccm. As this graph clearly shows, the absolute value of compression stresses decreases with increasing film thickness. To eliminate the internal stress gradients, we used a change in the parameter of argon inlet into the second source. As a result of optimization, a regime was selected in which the argon discharge gradually increased from 2 to 3.3 sccm during the growth of a film 200 nm thick. The ion current of the second source, respectively, increased from 42 mA to 51 mA. During deposition of thinner films, the argon inlet gradually increased from 2 to the corresponding value in the range of 2-3.3 sccm. For example, during the deposition of a film with a thickness of 50 nm, the argon admission value increased from 2 sccm to 2.3 sccm. All films grown in this mode with adjustment of the argon inlet into the second source during the film growth have the same value of internal stresses, as the curve indicated by the triangles in FIG. eighteen.

The presence of stress gradients in thin films can also be checked using a cantilever test structure. depicted in FIG. 20a, 20b, 20c. FIG. 19a illustrates a top view of the test structure before removing the sacrificial layer 203. FIG. 19b illustrates a side view of the test structure before removing the sacrificial layer 203. At position 200, two structures of the same thickness are formed having an adjacent side. One of the adjacent structures 203 is formed from a material of the sacrificial layer, the same as that used in the above pixel formation process. Another of adjacent structures 201 is formed from a material whose etching rate is significantly lower compared to the etching rate of the sacrificial layer during the removal of the latter. Pattern 201 may be formed, for example, from silicon nitride or silicon oxide. Also on the surface of structures 201 and 203, a strip 202 is formed from a material in which it is necessary to study the uniformity of internal stresses over the thickness. FIG. 20a-20c illustrate the state of the cantilever structure formed from the strip of test material 202. In the absence of an internal stress gradient in the cantilever structure 202a, it does not bend after removal of the sacrificial layer, as shown in FIG. 20a. In the presence of an internal stress gradient in the cantilever structure 202b or 202c, it bends upward, as shown in FIG. 20b, or bends downward, as shown in FIG. 20c, after removing the sacrificial layer. The situation shown in FIG. 20b, is realized, for example, when internal stresses increase in the direction from the lower part of the film 202b adjacent to the structure 201 to the upper (surface) part of the film. The situation shown in FIG. 20c is realized, for example, when internal stresses decrease in the direction from the lower part of the film 202b adjacent to the structure 201 to the upper (surface) part of the film.

The optimal dimensions of the cantilever structures are a length of 150 + - / 50 microns and a width of 20 +/- 10 microns. The bending of the cantilever structures can be measured using an interference microscope or using a scanning electron microscope.

A technique based on the fabrication of the test cantilever structures shown in FIG. 19a-19b and 20a-20c, confirms how the presence of gradients in the thickness of internal stresses in the films indicated by circles in FIG. 16-18, as well as the absence of gradients in the thickness of internal stresses in the films indicated by triangles in FIG. 16-18.

Comparison of measurements of internal stresses in films obtained on the basis of measurements of the bending of the positon using the Stoney formula and on the basis of measurements of the bending of cantilever structures made it possible to develop an easily applicable numerical criterion for the absence of a gradient of internal stresses in the layers of various materials by thickness, the use of which does not lead to unacceptable deformation of the legs and / or membranes. Further, the formulated criteria are conservative, i.e. valid for all materials from which the electrically conductive and dielectric layers can be made in this invention. The gradient of internal stresses over the thickness in the layer is insignificant (or negligible in the framework of this technology), i.e. not causing unacceptable deformations of the pixel structural elements, if the internal stress in the layer differs by less than 10% from the internal stress of the test layer, which is grown under the same conditions as the layer itself, and has a thickness equal to 20-30% of the layer thickness where the internal stresses in the layers are calculated using the Stoney formula based on the bending of the substrate. The criterion threshold of 10% is applicable for layers having a thickness equal to or greater than 90 nm. In the case where the layer thickness is less than 90 nm, a more stringent formulation is preferred when the values of internal stresses in the layer and test layer differ from each other by less than 5%. Hereinafter, layers having negligibly small gradients of internal stresses in thickness are simply called layers without internal gradients in thickness for ease of presentation. In addition, if the layer and the test layer have absolute values of internal stresses not exceeding 15 MPa, then such a layer is hereinafter referred to as unstressed.

The uniformity of the mechanical properties of the layers in thickness is also an important factor necessary for optimizing the manufacturing process of pixels. The values of the biaxial modules for the materials from which the dielectric layers 105, 106, 116 and the electrically conductive layer 112 are made can be measured using the test structures shown in FIG. 21a-21v. In FIG. 21a is a plan view of a test structure in one of the manufacturing steps, and FIG. 21b shows section AA of the same structure. A structure 301 is formed on the substrate 300, having a window filled with sacrificial material 303. The material from which the structure 301 is formed is not critical, it can be any material whose etching rate, in the subsequent manufacturing steps, is significantly lower than the etching rate of the sacrificial layer. For example, structure 301 may be made of silicon oxide or nitride. The window in the structure 301 and a portion of the surface of the structure 301, which is parallel to the surface of the substrate and adjacent to the perimeter of the window, is covered by a layer 302, the biaxial module of which is to be measured. Such a topology ensures the tightness of the window 304 covered by the layer 302 after removing a part of the substrate located under the structure from the sacrificial material 303, as well as the structure 303 itself, as shown in FIG. 20th century Subsequently, the test structure is mounted in a measuring system that measures the bending of the membrane formed from the layer 302 covering the window 304, depending on the difference in gas pressures P ₁ and P ₂ on opposite sides of the membrane. The membrane bending data, depending on the pressure difference measured for different test structures with different window sizes and the same layers 302, allows us to calculate the biaxial module of the material from which the layer 302 ^{4 is made} .

This technique also allows you to check the dependence of the biaxial module on the layer thickness. To do this, test structures 305 with various membrane thicknesses 302 are manufactured. The above-mentioned methods for growing layers without gradients of internal stresses in thickness also make it possible to obtain dielectric layers 105, 106, 116 and an electrically conductive layer 112 with a constant value of the biaxial module throughout the layer, i.e. layers of different thicknesses grown using the same technological process have almost the same biaxial modules, differing from each other in most cases by less than 3%. Further down the text, for simplicity of presentation, such layers are called layers with uniform mechanical properties. In FIG. 16-18, the squares denote the values of biaxial modules for layers of silicon nitride and titanium nitride of various thicknesses grown in the above mode with a change in the adjusted parameter, ensuring the absence of a gradient of internal stresses over the thickness. As these graphs clearly show, the values of biaxial modules for films of various thicknesses lie in the three percent range.

Most unstressed dielectric layers grown using the above methods also have optical properties that are uniform in thickness, i.e. their absorption and refraction coefficients in the IR range of 7–13 μm depend only on the wavelength. The absorption of these layers in a given wavelength range is described by the aforementioned Lambert-Behr formula ¹ with an error not exceeding 2%, where the absorption coefficients are independent of the thicknesses of the corresponding layers. Hereinafter, for the sake of simplicity, such dielectric layers are called layers with uniform optical properties.

In addition to the aforementioned criteria, which impose restrictions on internal stresses in individual layers, it is possible to use a number of criteria / formulas that relate the mechanical properties of the layers, during which the bending of the legs is insignificant, and / or the planarity and parallelism of the micro-membrane relative to the surface of the substrate is not violated:

, где "| |" - операция вычисления модуля, т.е. абсолютного значения величины, М₁ и d₁ являются биаксиальным модулем и толщиной первого диэлектрического слоя, примыкающего к одной поверхности электропроводящего слоя ножек, М₂ и d₂ являются биаксиальным модулем и толщиной второго диэлектрического слоя, примыкающего к противоположной поверхности электропроводящего слоя ножек. Этот критерий применим, когда толщина слоев равна или превышает 90 нм. В отдельных случаях, когда используются слои тоньше 90 нм, предпочтителен более жесткий критерий:

. Для выполнения этих критериев необходимо отсутствие внутренних напряжений в первом и втором диэлектрических слоях перед удалением жертвенного слоя 104 под ножками 116 и микро-мембраной 117. Для пояснения смысла последнего утверждения следует упомянуть, что после удаления жертвенного слоя электропроводящий слой вызывает напряжения растяжения в первом и втором диэлектрических слоях, что, в свою очередь, снижает их теплопроводность. Также в электропроводящем слое должны быть однородные по толщине напряжения сжатия, т.е. должен отсутствовать градиент внутренних напряжений. Кроме того, все слои должны обладать однородными механическими свойствами. Сформулированные в этом пункте критерии являются консервативными, т.е. определенными на основе данных, полученных с использованием слоев изготовленных из наиболее жестких материалов (т.е. имеющих наибольшие биаксиальные модули), которые используются для изготовления ножек или/и микро-мембраны в технологическом процессе, описанном выше. Кроме того, вышеперечисленные в этом пункте критерии, накладывающие ограничения на биаксиальные модули и толщины первого и второго диэлектрического слоев, применимы только тогда, когда первый и второй диэлектрический слои не являются многослойными, т.е. когда каждый из этих слоев состоит из одного слоя. Применение этих критериев этого пункта позволяет упростить процесс оптимизации технологии изготовления. Зная биаксиальные модули материалов, можно сразу выбирать толщины диэлектрических слоев, неприводящих к недопустимым вышеупомянутым искажениям структуры микро-мембраны и/или ножек, т.е. можно не проводить дополнительную оптимизацию технологического процесса, основанную на изготовлении тестовых консольных структур 202а-в с использованием тех же but)

where "| |" - operation of calculating the module, i.e. the absolute value, M ₁ and d ₁ are the biaxial module and the thickness of the first dielectric layer adjacent to one surface of the conductive layer of the legs, M ₂ and d ₂ are the biaxial module and the thickness of the second dielectric layer adjacent to the opposite surface of the conductive layer of legs. This criterion is applicable when the layer thickness is equal to or greater than 90 nm. In some cases, when layers thinner than 90 nm are used, a more stringent criterion is preferable:

. To fulfill these criteria, the absence of internal stresses in the first and second dielectric layers is necessary before removing the sacrificial layer 104 under the legs 116 and the micro-membrane 117. To clarify the meaning of the last statement, it should be mentioned that after removing the sacrificial layer, the electrically conductive layer causes tensile stresses in the first and second dielectric layers, which, in turn, reduces their thermal conductivity. Also in the electrically conductive layer should be uniform in thickness compression stresses, i.e. there should be no gradient of internal stresses. In addition, all layers must have uniform mechanical properties. The criteria formulated in this clause are conservative, i.e. determined based on data obtained using layers made of the most rigid materials (i.e., having the largest biaxial modules), which are used to make the legs and / or micro-membranes in the process described above. In addition, the criteria listed in this clause, which impose restrictions on the biaxial modules and thicknesses of the first and second dielectric layers, are applicable only when the first and second dielectric layers are not multilayer, i.e. when each of these layers consists of one layer. The application of these criteria of this paragraph allows to simplify the process of optimization of manufacturing technology. Knowing the biaxial modules of the materials, one can immediately choose the thickness of the dielectric layers that do not lead to unacceptable above-mentioned distortions in the structure of the micro-membrane and / or legs, i.e. You may not need to carry out additional process optimization based on the manufacture of test console structures 202a-b using the same

processes and materials that are used to make the legs (i.e., in this case, the cantilever structure 202a-b is multilayer and consists of layers: the first dielectric layer / electrically conductive layer / second dielectric layer);

b) M ₁ * d ₁ > M ₂ * d ₂ , this criterion is applicable when the internal compressive stresses in the electrically conductive layer have a gradient, when the absolute value of the compressive stresses in the electrically conductive layer at the boundary with the first dielectric layer having a biaxial module M ₁ and a thickness d _{1 is} greater than the absolute value of the compression stresses in the electrically conductive layer at the boundary with the second dielectric layer having a biaxial module M ₂ and a thickness d ₂ . The criterion formulated above in this paragraph can also be formulated for the “mirror case” when the absolute value of the compressive stresses in the electrically conductive layer at the interface with the first dielectric layer is less than the absolute value of the compressive stresses in the electrically conductive layer at the interface with the second dielectric layer. In this case, M ₁ * d ₁ <M ₂ * d ₂ . The formulation of the criteria of this paragraph is valid for the case when the electrically conductive layer is grown on the first dielectric layer, and the second dielectric layer, in turn, is grown on the electrically conductive layer. As in paragraph a), the first and second dielectric layers must be unstressed until the sacrificial layer is removed and not be multilayer, in addition, the first and second dielectric layers must have mechanical properties uniform in thickness. The criteria formulated in this clause are conservative, i.e. determined based on data obtained using layers made of the most rigid materials (i.e., having the largest biaxial modules), which are used to make the legs and / or micro-membranes in the process described above.

In this case, it is impossible to completely abandon the manufacture of test console structures 202a-c mentioned in paragraph a) to optimize the manufacturing technology, however, the criterion formulated as inequality in this paragraph can significantly reduce the parameter space in which optimization should be carried out. For example, if the thickness d ₂ is chosen randomly or based on considerations unrelated to preventing unacceptable distortions in the shape of the legs and / or membrane, then, based on the criterion M ₁ * d ₁ > M ₂ * d ₂ , the thickness d ₁ has the limiting criterion d ₁ > M ₂ * d ₂ / M ₁ , which allows to significantly limit the range of parameters in which it is necessary to search for the optimal thickness d ₁ . On the other hand, within the framework of the criteria formulated in paragraph b), it is possible to manufacture legs with substantially different first and second dielectric. For example, if the first and second dielectric are made of the same material, then according to the criteria of paragraph a) their thicknesses should be the same, while according to the criteria of paragraph b) their thicknesses should differ.

In the case where the first dielectric layer consists of several dielectric layers, in points a) and b) the corresponding product M ₁ * d ₁ is replaced by the sum of the products of the biaxial modules and the thicknesses of the dielectric layers making up the first dielectric layer. For example, if the first dielectric layer consists of two dielectric layers A and B having respectively biaxial modules M _A and M _B and thicknesses d _A and d _B , then the product M ₁ * d ₁ in paragraphs a) and b) is replaced by the sum of the products M _A * d _A and M _B * d _B. A similar replacement procedure is carried out for the second dielectric layer in rules a) and b) if it, in turn, consists of several dielectric layers. Thus, the criteria formulas in paragraphs a) and b) can be rewritten in a generalized mathematical form:

[one]

where N is the number of layers in the first dielectric layer, M _i and d _i are the biaxial modules and layer thicknesses in the first dielectric layer, M is the number of layers in the second dielectric layer, M _k and d _k are biaxial modules and layer thicknesses in the second dielectric layer where the thicknesses of the first dielectric layer, the second dielectric layer and the electrically conductive layer of the films are equal to or greater than 90 nm, and the layers of the first and second dielectric layers are unstressed until the sacrificial layer is removed and have uniform mechanical properties in thickness e, in addition, the electrically conductive layer also has uniform mechanical properties in thickness and uniform compression stresses in thickness, i.e. there is no gradient of internal compression stresses;

где N - количество слоев в первом диэлектрическом слое, M_i и d_i - биаксиальные модули и толщины слоев в первом диэлектрическом слое, М - количество слоев во втором диэлектрическом слое, M_k и d_k - биаксиальные модули и толщины слоев во втором диэлектрическом слое, где толщины первого диэлектрического слоя, второго диэлектрического слоя и электропроводящего слоя менее 90 нм, а остальные критерии, определяющие свойства слоев, такие же, как для формулы [1];[2]

where N is the number of layers in the first dielectric layer, M _i and d _i are the biaxial modules and layer thicknesses in the first dielectric layer, M is the number of layers in the second dielectric layer, M _k and d _k are biaxial modules and layer thicknesses in the second dielectric layer where the thicknesses of the first dielectric layer, the second dielectric layer and the electrically conductive layer are less than 90 nm, and the remaining criteria determining the properties of the layers are the same as for the formula [1];

[3]

where N is the number of layers in the first dielectric layer, M _i and d _i are biaxial modules and layer thicknesses in the first

dielectric layer, M the number of layers in the second dielectric layer, M _k and d _k biaxial modules and layer thicknesses in the second dielectric layer, where there is a gradient of internal stresses in the electrically conductive layer, the internal compression stresses in the part of the electrically conductive layer adjacent to the first dielectric layer are greater in absolute value than the compressive stresses in the part of the electrically conductive layer adjacent to the second dielectric layer, in addition, the layers of the first and second dielectric layers are unstressed and before removing the sacrificial layer and have uniform mechanical properties in thickness;

[four]

where N is the number of layers in the first dielectric layer, M _i and d _i are the biaxial modules and layer thicknesses in the first dielectric layer, M is the number of layers in the second dielectric layer, M _k and d _k are biaxial modules and layer thicknesses in the second dielectric layer where there is a gradient of internal stresses in the electrically conductive layer, the internal compression stresses in the part of the electrically conductive layer adjacent to the first dielectric layer are smaller in absolute value than the compressive stresses in the part of the electrically conductive layer adjacent to the second dielectric layer, in addition, the layers of the first and second dielectric layers are unstressed until the sacrificial layer is removed and have uniform mechanical properties in thickness;

Criteria similar to criteria a) and b) can be developed and applied to select the topology of the layer fragments that are used to make the micro-membrane 117, the temperature detector 108, its contacts 119, the absorber 109 and the electrical connection between the contacts of the temperature detector and the electrically conductive fragments of the layer 112a used to make the legs 116, since fragments of the electrically conductive layer causing tensile stresses in the dielectric layers of the legs 116 can also be used to create This electrical connection 112b and, hence, cause stresses in the fragments of layers which are used for the manufacture of micro-membrane 117, the temperature detector 108, its contacts 119 and absorber 109 (Fig. 13b and 14). However, in this case, in contrast to the legs 116, where the only topological difference between the fragments of the layers used for their manufacture is their thicknesses, it is necessary to take into account not only the thicknesses of the fragments of the layers used to manufacture the micro-membrane 117, the temperature detector 108, its contacts 119 , absorber 109, electrical connections 112b, providing electrical connection of contacts 119 with fragments of the conductive layer 112a, but their geometry in a plane parallel to the substrate 100. To exclude direct influence I internal stresses in the conductive layer 112 on the planarity of the micro-membrane 117 itself, the electrical connections 112b connecting the contacts 119 of the temperature detector 108 with the fragments of the conductive layer 112a used to make the legs 116, can be made of fragments of the same layer from which fragments are made detector contacts 119.

The above methods and criteria make it possible to apply the pixel manufacturing optimization technique shown in FIG. 22. This technique allows for stage-by-stage iterative optimization of manufacturing, which greatly simplifies the optimization process, since at each stage a small number of parameters are optimized.

At the initial stage of optimization 400, a library of spraying processes of dielectric and electrically conductive layers is created. Each process for spraying dielectric layers allows the growth of dielectric layers with uniform optical and mechanical properties. In addition, the process library includes the deposition of electrically conductive layers having internal compression stresses. The spraying processes in the library allow spraying of electrically conductive layers both with a gradient of internal stresses in thickness and without a gradient of internal stresses in thickness.

During the next optimization step 401, one or more pixel manufacturing technology paths are selected. The technological route of manufacturing a pixel must satisfy the following requirements: minimizing the number of layers and the number of layer structuring processes used in the manufacturing process of a pixel, in addition, the layers used must provide the required sensitivity of the pixel in the operating range of the device, and also not cause unacceptable bending of the legs and / or micro membranes. Simplification of the technological route leads to the need to use fragments of the same layers in the maximum number of pixel functional elements. For example, the first dielectric layer, which may consist of either one or both of the layers 105 and 106, is used in the above-described pixel manufacturing process for the manufacture of both legs and micro-membranes. In addition, the first dielectric layer functions as an absorber, since the above materials used for its manufacture absorb infrared radiation. Such multipurpose use of the first dielectric layer leads to conflicting requirements. On the one hand, increasing the thickness of this layer allows you to increase the absorption of infrared radiation and thereby increase the sensitivity of the pixel, on the other hand, increasing the thickness of this layer leads to an increase in the thickness of the legs and, accordingly, their thermal conductivity, which, in turn, reduces the thermal insulation of the temperature detector and, as a result, pixel sensitivity. Exactly the same dilemma underlies the determination of the thickness of the second dielectric layer 114, which is used to manufacture the same pixel functional elements as the first dielectric layer. The optimal resolution of these conflicting requirements for the thicknesses of the first and second dielectric layers can be found by using additional layers in the pixel, which are used only for the manufacture of one of the functional elements. For example, layer 109 only functions as an absorber. This layer may consist of several layers providing the necessary absorption of IR radiation in the operating range of the device. This, as it should be noted, is only one of the options for optimizing the technological process and topology of the layers. For example, layer 105 can be structured so that it only functions as an absorber.

Table 1 shows the specifications of the dielectric layers (mechanical properties and spraying processes) used to make the legs in various pixels / products. The selection of processes, materials and layer thicknesses was carried out by the method of selecting layers with suitable mechanical and optical properties in the library created at step 400, where one of the formulas [1] - [4] was used when choosing the thicknesses.

The problem of optimizing the pixel manufacturing process must also be analyzed from the point of view of the physical processes that determine the quality of the pixel, namely: the mechanical properties of the films used to make the pixel, the absorption of infrared radiation by a pixel absorber, the conversion of absorbed infrared radiation into a measured signal of a temperature detector, and heat transfer from the absorber to temperature detector and thermal conductivity of the legs. The mechanical properties of the layers and the absorption of infrared radiation in the layers are well-studied physical processes, in addition, these properties are easy to measure. The integral properties of a pixel associated with IR absorption and the mechanical properties of individual layers are easily predictable based on the properties of individual layers. That is why all technological processes in the process library are optimized in such a way as to obtain unstressed films of dielectrics with homogeneous mechanical and optical properties. As a result, you can choose a combination of layers in the process library that provides not only the desired absorption spectrum in the operating range of the device, but also taking into account formulas [1] - [4] which does not cause unacceptable deformation of the legs and / or micro-membranes. It should be noted that in the case of optimization of mechanical properties according to formulas [3] or [4], experimental optimization of the thickness of the first or second dielectric layer is still necessary. In contrast to the optical absorption and mechanical properties of the layers, parameters such as the thermal conductivity of the legs and the process of heat transfer from the absorber to the temperature detector are difficult to describe with a simple model based on a simple additive account of certain parameters of individual layers. The heat transfer processes in a pixel depend on a large number of parameters, which often depend on the specific topology of the pixel. For example, the phonon mean free path is comparable to the characteristic size of the layers in a pixel, in addition, it is possible to reduce heat transfer between different layers due to Kapitsa resistances at the boundaries between the layers. In other words, the heat transfer processes in a pixel depend not only on the processes occurring in each layer, but also on the interaction of the processes occurring in these layers. Thus, the library, which includes the processes of spraying layers and information about the properties of the layers obtained in these processes, simplifies the design process as much as possible, allowing us to narrow the search space for optimal solutions at the stage of planning the optimization process.

In the next stage of optimization of manufacturing process 402, if necessary, layer thicknesses are adjusted using test cantilever structures consisting of several layers, where the test structures correspond to the combinations of layers from which pixel elements should be made. As a result of this adjustment of the thicknesses, the absence of unacceptable deformations in such pixel elements as, for example, a micro-membrane and legs is ensured. The main objective of step 402 is to produce pixels according to one or more technological appropriate topologies taking into account possible adjustment of layer thicknesses.

In the next stage of the optimization of the manufacturing process 404, measurements are made of the sensitivity of the pixel in the operating range. These measurements may include measurements of the spectral and temporal response of a pixel or multiple pixels. In addition to measuring pixel characteristics, the characteristics of individual pixel nodes (e.g., thermal conductivity of the legs) and / or the characteristics of test structures can be measured. As a test structure that allows you to control the reflection coefficient of the reflector 1036, you can use a freestanding reflector made on the same technological route as the reflector 1036. In the future, the results are compared with the given characteristics. Correspondence criteria of the obtained characteristic to a given one can be formulated in the form of an interval in which the measured value of the characteristic should be located, or in the form of a threshold value that determines the maximum (minimum) allowable measured value of the characteristic. In case of discrepancy between the measured characteristics and the given ones, the technological manufacturing routes of the pixel are changed to achieve the specified characteristics, i.e. in fact, a complete or partial repetition of step 401 followed by repeating steps 402 and 404. For example, if as a result of measurements it was established that the heat sink to the substrate from the temperature detector to the substrate is unacceptably large, then during the repeated execution of optimization step 401, a decision can be made about reducing the thicknesses of the layers simultaneously used to make the legs and the absorber, and using additional layers to make the pixel, which are exclusively used to make the absorber. In some cases, the repetition of step 401 may be limited by changing the thickness of one or more layers.

As a very simple and effective method of measuring the parameters of a pixel can be a measurement of the current-voltage characteristics of the pixel in a quasi-static mode. If such measurements are made in vacuum with a residual pressure of less than 1 Pa, then the Joule heat generated in the temperature detector will be removed mainly through the legs, i.e. neglect of other heat removal mechanisms, such as radiation, does not significantly affect the measurement accuracy. In this case, the heat balance equation is formulated as follows: ΔT * G = I ² * R (ΔT), where the equation R (ΔT) = R ₀ (1 + α * ΔT) describes the temperature dependence of the electrical resistance of the temperature detector R (ΔT) where R ₀ is the base resistance value at the base temperature, ΔT is the temperature detector temperature exceeding the base temperature, α is the temperature coefficient of resistance, G is the thermal conductivity of the legs, I is the current flowing through the temperature detector and generating Joule heat. The right side of the heat balance equation describes the generation of Joule heat, and the right side of this equation describes the sink into the substrate through the legs of the generated heat. This equation can be written in another form: 1 / R (ΔT) = 1 / R ₀ -αI ² / G. The construction of a linear graph, where the square of the current is plotted on the horizontal axis and the inverse of the resistance is plotted on the vertical axis, allows you to determine the value of α / G from the slope of the straight graph. Thus, by manufacturing different pixels on the same plate with the same temperature detectors (for example, the same layer of vanadium oxide), we can compare the thermal conductivity of the pixel legs, which were made according to different technological routes.

Table 2 shows a selection of test pixel measurements illustrating the effect of reducing the thermal conductivity of the legs using electrically conductive layers that cause tensile stresses in the dielectric layers of the legs. Each line in the table corresponds to one technological process in which two types of pixels were made. For the manufacture of one type of pixels, the deposition of dielectrics and an electrically conductive layer from the library was used, the other type of pixels was produced in the same way, except for one difference, which consisted in the fact that unstressed layers were used as the electrically conductive layers in the legs. Since temperature detectors were manufactured for both types of pixels in a single process, the measurements described above make it possible to compare the thermal conductivity of the legs. The last column shows the percentage reduction in the thermal conductivity of the legs for pixels, which used electrically conductive layers with internal compression stresses, relative to reference pixels, where stressed electrically conductive layers were used. This decrease in thermal conductivity is not related to the possible influence of internal stresses on the thermal conductivity of the electrically conductive layers. Comparative measurements of thermal conductivity by the 3ω method in stressed and non-stressed electrically conductive layers made of the same materials did not reveal differences within 1.5% ⁵ . Thus, the decrease in the thermal conductivity of the legs, shown in the last column of the table, is caused precisely by the use of strained conductive layers that cause tensile stresses in the dielectric layers. Results similar to those presented in table 2 were obtained for other transition metal nitrides in combination with various dielectric layers.

The preferred range of internal stresses in the electrically conductive layers causing tensile stresses in the dielectric layers of the legs is the interval −3 GPa to −1 GPa. The most preferred interval of internal stresses in the electrically conductive layers, causing tensile stresses in the dielectric layers of the legs, is the interval -2 GPa - -1.5 GPa.

If the measured pixel characteristics match the specified optimization process ends at step 405. The transition to step 405 can also be carried out if the optimization process is saturated and further repetition of steps 401, 402 and 404 does not seem justified, for example, when a relative change in one or more characteristics as a result of the last iteration are less than the corresponding threshold values (for example, 5% or 10%). At step 405, documentation is prepared for the mass production of devices with pixels according to one or more technological manufacturing routes worked out during the above optimization steps.

In FIG. 23 is an example of a matrix of 500 pixels 502, for obtaining images in the infrared range. The pixels of this matrix can be made using the same technology as the single pixel manufacturing technology described above. This matrix includes electronic circuits that allow pixel-by-pixel polling with subsequent registration of the IR radiation power received by each pixel in the polled row. In this design, two columns correspond to each pixel. In FIG. Figure 24 shows a more compact design of a matrix of 510 pixels 512. In this case, one column 513 is used for adjacent pixels. The matrix 510 also includes electronic circuits that allow pixel-by-pixel polling with subsequent registration of the IR radiation power received by each pixel in the polled row. By analogy with the pixel image in FIG. 14, for clarity, the images in FIG. 23 and 24, elements 103a, 103b and details of the surface relief of the micro-membranes are not shown. In FIG. 23 and 24 show the contours of structures made of an electrically conductive layer 112, the contours of contact structures 119, and the contours of structured layers 108 and 109.

Information sources

1. "Optical thin films and coatings," ed. A. Piegari and F. Flory, Woodhead Publishing, p. 294, ISBN 9780857095947.

2. "Fundamentals of Microfabrication: The Science of Miniaturization, Second Edition" M.J. Madou, CRC Pres, p. 312, ISBN-10: 0849308267.

3. "Thin-film deposition: principles and practice," ed. D.L. Smith, McGraw-Hill, Inc., p. 495, ISBN 0-07-113913-3.

4. The MEMS handbook. Ed. M. Gad-el-Hak, CRC Press, pp. 16-110, ISBN 0-8493-0077-0.

5. Thermoelectrics handbook macro to nano. Ed. D.M. Rowe, Taylor & Francis group, pp. 23-11, ISBN 0-8493-2264-2.

Claims (20)

1. A device for thermal detection of infrared radiation, including:
a pixel on a semiconductor substrate, a pixel includes a first section and a second section, the first section is on the surface of the semiconductor wafer and includes electrical circuits, the second section is separated from the first section and located directly above it, the second section is planar and includes legs , the micro-membrane and the temperature detector located on it, the second section is supported by columns, one of the legs has one end, integrally connected to the micro-membrane, and the other end, integrally connected connected to one of the columns, the other of the legs has one end, integrally connected to the micro-membrane, and the other end, integrally connected to the other of the columns, the legs provide electrical connection of the temperature detector with electric circuits through the corresponding columns and thermal insulation of the temperature detector and micro- membranes from a semiconductor substrate, one of the legs includes:
the first part of the first dielectric layer,
the first part of the second dielectric layer,
part of the conductive layer, this part of the conductive layer provides the aforementioned electrical connection, the first part of the first dielectric layer borders the first surface of the conductive layer and the first part of the second dielectric layer borders the second surface of the conductive layer, the first and second surfaces of the conductive layer are opposite surfaces of the part of the conductive layer, part of the conductive layer is a source of mechanical stress, causing apryazheniya stretching in the first part of the first dielectric layer and the tensile stress in the first portion of the second dielectric layer. 2. The device for thermal detection of infrared radiation according to claim 1, where the electrically conductive layer includes a nitride of one or more of the following metals: titanium, vanadium, chromium, zirconium, niobium, molybdenum, hafnium, tantalum or tungsten. 3. The device for thermal detection of infrared radiation according to claim 1, where the thickness of the first part of the first dielectric layer (d1), the value of the biaxial module of the first part of the first dielectric layer (M ₁ ), the thickness of the first part of the second dielectric layer (d ₂ ), the value of biaxial the modules of the second part of the first dielectric layer (M ₂ ) satisfy the inequality | M ₁ * d ₁ -M ₂ * d ₂ | / (M ₁ * d ₁ + M ₂ * d ₂ ) <0.1. 4. The device for thermal detection of infrared radiation according to claim 1, where the first dielectric layer is a multilayer structure consisting of N layers, and / or the second dielectric layer is a multilayer structure consisting of M layers, where the layer thicknesses (d1 ₁ ... d1 _N respectively) of the first dielectric layer and their corresponding biaxial modules (M1 ₁ ... M1 _N, respectively) and layer thicknesses (d2 ₁ ... d2 _N, respectively) of the second dielectric layer and their corresponding biaxial modules (M2 ₁ ... M2 _N, respectively) satisfy the inequality to the

. 5. The device for thermal detection of infrared radiation according to claim 1, where the product of the thickness of the first part of the first dielectric layer by the value of the biaxial module of the first part of the first dielectric layer is greater than the product of the thickness of the first part of the second dielectric layer by the value of the biaxial module of the second part of the first dielectric layer, if the voltage compressions are reduced in part of the conductive throughout the thickness of the conductive layer from the first surface to the second surface. 6. The device for thermal detection of infrared radiation according to claim 1, where the first dielectric layer is a multilayer structure consisting of several layers, and / or the second dielectric layer is a multilayer structure consisting of several layers, where each layer of both the first and second layer corresponds to the product of the biaxial module of this layer by its thickness, where the sum of all products corresponding to all layers of the first dielectric layer is greater than the sum of all products corresponding to all layers of the second of the dielectric layer, if the compression stresses decrease in part of the conductive layer over the thickness of the conductive layer from the first surface to the second surface. 7. The device for thermal detection of infrared radiation according to claim 1, where the first dielectric layer includes one or more of the following layers: the first layer of silicon nitride, the first layer of silicon oxide, the first layer of oxy-silicon nitride. 8. The device for thermal detection of infrared radiation according to claim 1, where the second dielectric layer includes one or more of the following layers: a second layer of silicon nitride, a second layer of silicon oxide, a second layer of oxy-silicon nitride. 9. The device for thermal detection of infrared radiation according to claim 1, where the pixel includes an infrared absorber that is thermally connected to a temperature detector, the absorber includes a second part of the first dielectric layer and / or a second part of the second dielectric layer, the legs provide thermal insulation of the IR absorber radiation from the substrate. 10. The device for thermal detection of infrared radiation according to claim 9, where the micro-membrane structure includes a third part of the first dielectric layer and / or a third part of the second dielectric layer. 11. The device for thermal detection of infrared radiation according to claim 10, where the third part of the first dielectric layer includes the second part of the first dielectric layer, if the micro-membrane structure includes the third part of the first dielectric layer and the absorber includes the second part of the first dielectric layer, where the third part of the second dielectric layer includes the second part of the second dielectric layer, if the micro-membrane structure includes the third part of the second dielectric layer absorber includes a second portion of the second dielectric layer. 12. The device for thermal detection of infrared radiation according to claim 1, where the temperature detector includes a vanadium oxide layer VO _x , where 1.7 <x <1.9, where the electrical resistance of the vanadium oxide layer decreases with increasing temperature in the range 1.7-2.6% per degree Kelvin. 13. The device for thermal detection of infrared radiation according to claim 1, where the pixel includes a reflecting infrared layer covering at least a portion of the first section and facing the second section. 14. The device for thermal detection of infrared radiation according to claim 1, which also includes an array of the above pixels. 15. A method of manufacturing a device for thermal detection of infrared radiation, a device for thermal detection of infrared radiation includes a matrix of pixels on a semiconductor substrate, each of the pixels includes a corresponding first section of each of the pixels and a corresponding second section of each of the pixels separated from and located above said corresponding first section and supported by at least two corresponding columns, each of the columns supports only one second second July or each of the columns from the first group of columns supports only two of the second sections and each of the columns of the second group of columns supports only one of the second sections, each first section is on the surface of the semiconductor substrate and includes the corresponding electrical circuits of each first section, every second the section is planar and includes the corresponding legs of each second section, the corresponding micro-membrane of each second section, and the corresponding temperature detector of each second section and said corresponding micro-membrane, the method includes the following step for each of the pixels:
• the formation of the aforementioned corresponding legs, where each of the formed legs has one end integrally connected to a micro-membrane of a pixel including formed legs, and the other end integrally connected to a corresponding column supporting a second section including formed legs, where at least two formed the legs have their ends integrally connected to two different columns, where the formed legs provide electrical contact between the peak temperature detector mudflow, which includes formed legs with pixel electric circuits, including formed legs through corresponding columns, where the formed legs provide thermal insulation of the micro-membrane of the pixel, including formed legs, and a pixel temperature detector, which includes formed legs from the semiconductor substrates
The method also includes the following step:
• formation of a sacrificial layer on a semiconductor substrate,
the formation of the corresponding legs includes the following steps for forming one of the aforementioned corresponding legs:

the formation of the first part of the first dielectric layer on the sacrificial layer,

forming part of the conductive layer on the first part of the first dielectric layer, where part of the conductive layer provides the aforementioned electrical contact, where the formed part of the conductive layer has compression stresses, and

the formation of the first part of the second dielectric layer on the part of the conductive layer,
The method also includes the following step:
• removal of the sacrificial layer, where each part of the electrically conductive layer causes tensile stresses in the corresponding adjacent first part of the first dielectric layer and tensile stresses in the corresponding adjacent first part of the second dielectric layer. 16. The method according to p. 15, where each of the pixels includes a corresponding infrared absorber, where each of the infrared absorbers is thermally coupled to a single pixel thermal detector including each of the absorbers, includes a corresponding second part of the first dielectric layer and / or the corresponding second part of the second dielectric layer and is thermally isolated from the semiconductor substrate by means of pixel legs, including each of the absorbers mentioned, a method in Luciano a following step for each of the pixels:
• the formation of the corresponding infrared absorber, where the formation of the corresponding infrared absorber includes the following steps:

the formation of the corresponding second part of the first dielectric layer on the sacrificial layer, if the corresponding IR absorber, which must be formed in the process of forming the corresponding absorber of infrared radiation, includes the corresponding second part of the first dielectric layer, and

the formation of the corresponding second part of the second dielectric layer, if the corresponding IR absorber, which must be formed in the process of forming the corresponding absorber of infrared radiation, includes the corresponding second part of the second dielectric layer. 17. The method according to p. 16, where each of the micro-membranes includes a corresponding third part of the first dielectric layer and / or the corresponding third part of the second dielectric layer, the method includes the following step for each of the pixels:
• the formation of the corresponding micro-membrane, where the formation of the corresponding micro-membrane includes the following steps:

the formation of the corresponding third part of the first dielectric layer on the sacrificial layer, if the corresponding micro-membrane, which must be formed in the process of forming the corresponding micro-membrane, includes the corresponding third part of the first dielectric layer, and

the formation of the corresponding third part of the first dielectric layer on the sacrificial layer, if the corresponding micro-membrane, which must be formed in the process of forming the corresponding micro-membrane, includes the corresponding third part of the first dielectric layer. 18. The method according to p. 17, where every third part of the first dielectric layer includes a corresponding second part of the first dielectric layer, if each IR absorber includes a corresponding second part of the first dielectric layer and each micro-membrane includes a corresponding third part of the first dielectric layer, every third part of the second dielectric layer includes a corresponding second part of the second dielectric layer, if each IR absorber includes a corresponding etstvuyuschuyu second part of the second dielectric layer and each micro-membrane comprises a third portion corresponding to the second dielectric layer. 19. The method according to p. 15, where the formation of the first part of the first dielectric layer on the sacrificial layer includes spraying the first dielectric layer on the sacrificial layer using a process that allows you to get unstressed first dielectric layer, where the formation of the first part of the second dielectric layer on the part of the conductive layer includes spraying a second dielectric layer onto an electrically conductive layer using a process allowing the second dielectric layer to be unstressed. 20. The method according to p. 16, where each of the micro-membranes includes a corresponding third part of the first dielectric layer and / or a corresponding third part of the second dielectric layer, every third part of the first dielectric layer includes a corresponding second part of the first dielectric layer, if each infrared absorber includes a corresponding second part of the first dielectric layer and each micro-membrane includes a corresponding third part of the first dielectric layer, every third hour st second dielectric layer includes the corresponding second portion of the second dielectric layer if each infrared radiation absorber includes a corresponding second part of the second dielectric layer and each micro-membrane comprises a third portion corresponding to the second dielectric layer.

Images