The invention relates to devices for detecting thermal radiation and methods of manufacturing thereof. These devices have detectors of thermal radiation with thermal insulation. One of many innovative aspects of this invention is new structures for thermal insulation of radiation detectors.

The structure for thermal radiation detection can be implemented using the super-sensitive thermometer for the thermal radiation detection and measurement. In particular, such structures can be used for detection and measurement of infrared (IR) radiation. Structures for detection of IR radiation are the temperature responsive detector thermally coupled with an IR radiation absorber; whereas the temperature responsive detector and absorber are heat-insulated from a substrate on which the electrical circuits for reading and processing of the values of the thermal radiation absorbed by the temperature responsive detector are made. Sensitivity of such IR detectors is determined, in addition to other factors, by quality of heat insulation of the temperature responsive detector. If the heat insulation of the temperature responsive detector is insufficient, the heating of the temperature responsive detector by thermal radiation is also insufficient that in turn leads to the failure of detection of property changes of the temperature responsive detector caused by thermal radiation, by means of the electronic instruments. As usual, the material layer with the high temperature coefficient of resistance (TCR) is used as the temperature responsive detector. As usual, this TCR is expressed as the percentage change of resistance per degree Kelvin. At present, the best materials, such as vanadium oxide or amorphous silicon, have the TCR change in a range of several percent per degree Kelvin, which is sufficient for detection and registration of IR radiation. The thermal radiation detectors can have more complex structure. For example, semiconductor p-n diodes can be used as IR radiation detectors. In this case, detection and measurement of IR radiation occurs on the basis of change in the current voltage characteristics of the diodes caused by heating.

The heat insulation of the temperature responsive detector (and the absorber) from a substrate, on which the detector is fabricated, as usual, is implemented by means of so-called legs. The legs are elongated structural elements of the detector which enable avoiding the direct physical contact of the temperature responsive detector (and the absorber) with a substrate. As usual, the leg in a combination with posts allow suspending the detector (and the absorber) over a substrate surface. The elongated form of the legs enables implementing a low heat conductivity of this element and increasing the heat insulation of the temperature responsive detector from the substrate. One end of each leg is connected to the temperature responsive detector or a microbridge structure on which the temperature responsive detector is placed, other end of each leg is connected to the respective post providing the suspending of the temperature responsive detector above a substrate surface. The patent EP 1212592 Bl describes such a structure, wherein the temperature responsive detector placed on a micro-membrane is suspended over a substrate using the legs and the posts. The legs in this patent have an elongated form and are made out of vanadium oxide and silicon nitride. In this patent, the heat insulation of the temperature responsive detector is achieved by means of the length of the legs.

In order to understand the invention disclosed herein, it is necessary to formulate the following definitions of physical quantities and laws used for the description of the invention.

Young's modulus designated as E, (a modulus of longitudinal elasticity) is the physical quantity characterizing the properties of material to resist extension/compression in case of elastic deformation.

The Hooke's law is the statement according to which the deformation in an elastic body (a spring, a rod, a console, a beam, etc.), is proportional to force applied to this body. The Hooke's law can be described by the following equation σ = ε*Ε, where σ is a normal stress in a cross- section area being equal to force F applied to the cross-section of an elastic body divided by the area of the cross-section S, ε is the strain being equal to the change of a length, ΔΙ, of the elastic body in the direction and under the influence of the applied force, divided by length of the elastic body, /, in the direction of the applied force.

The Poisson ratio designated as μ is the negative ratio of transverse to axial strain.

The biaxial modulus designated as M is the value calculated according to the equation Μ=Ε/(1-μ).

IR radiation is the electromagnetic radiation in the wavelength range of0.7 μηι - 1000 μι ^η . The absorption of IR radiation in a layer of material which absorption coefficient does not depend on thickness, is described by Lambert-Beer law: The I _d =I ₀ e ^"kd where I ₀ is a radiation flux received by a layer of material, the I _d is a radiation flux after traversing the layer of a thickness d, k - absorption coefficient ¹ . If the absorption properties of the material do not depend on thickness of the layer made of this material, the absorption coefficient depends only on wavelength and does not depend on layer thickness. A metal nitride, as understood here, can be a single-phase or multiphase compound having stoichiometric or non-stoichiometric composition.

A nitride, oxide, and oxynitride of silicon, as understood here, can be a hydrogenated or not hydrogenated silicon compound having amorphous structure, stoichiometric or non- stoichiometric composition.

The structure for detection of IR radiation comprises at least one temperature responsive detector on the microbridge structure positioned above a substrate surface. The microbridge structure is suspended above a substrate surface using legs and posts.

The legs are the elongated members having low heat conductivity providing a heat insulation of the microbridge structure and the temperature responsive detector placed thereon from a substrate. One end of each leg is connected to a microbridge structure; other end of the leg is connected to the respective post providing a gap between the substrate surface and the microbridge structure. The legs and the posts provide for an electrical connection with electronic components (circuits) on a substrate, wherein the electronic components are used for registering temperature of the temperature responsive detector. In order to improve the sensitivity of the IR radiation temperature detector, the absorber of IR radiation being thermally coupled to the temperature responsive detector of IR radiation can be used. The IR radiation can be absorbed both by the absorber and the temperature responsive detector and/or material of a microbridge structure. Besides, the substrate surface under the microbridge structure can be coated with the material reflecting IR radiation for increasing the absorption percentage of IR radiation modified by the temperature responsive detector. The microbridge structure, legs being integrally connected to the microbridge structure, temperature responsive detector allocated on the microbridge structure form a planar structure. This planar structure can also include the absorber.

Such structure can be made using thin-film technologies widespread in the semiconductor industry. As usual, in order to simplify the manufacturing method, one layer of material is structured in such a way that different parts of this layer are used in different parts of a product. For example, one dielectric layer can be used for manufacturing not only a microbridge structure but legs as well. Besides, the suitable material of the microbridge structure enables functioning of the microbridge structure as an absorber. Such materials as oxide, nitride, or oxy-nitride of silicon have the absorption peaks in the IR range. For example, the vibration modes of chemical bonds of Si-N, Si-H and N-H cause the appropriate absorption peaks at wavelengths of 11.76 μπι, 4.5 μπι, and 2.99 μηι respectively. However the use of one layer in different structure members imposes many requirements on layer properties. For example, if one dielectric layer is used for making microbridge structures and legs and in addition this layer has to function as an absorber, then an intrinsic stress in this layer has to be optimized in a way that it does not cause bending of the legs and a the microbridge structure; moreover, the absorption of the layer used for fabrication of a microbridge structure has to be high in the operational IR range; and last but not the least, the heat conductivity of the same layer used for fabrication of legs has to be low to minimize the thermal losses via legs.

Even when there is a solution for such a multi-objective layer property optimization problem, it is very difficult to find it. One of the aspects of this invention is the reduction of the heat conductivity of one or more dielectric layers used in legs by applying tensile stress to them. The reduction of heat conductivity of the leg comprising an electrically conductive layer and one or more dielectric layers being adjacent to the electrically conductive layer can be achieved when the conductive layer is made of a stiff material with compressive stress. In this case, the electrically conductive layer causes tensile stress in the one or more adjacent dielectric layers, whereas the tensile stress reduces heat conductivity of the one or more adjacent dielectric layers. In the semiconductor industry a layer having intrinsic stress causing mechanical stress in adjacent layer is called stressor. This effect enables simplification of the optimization problem, for example if the dielectric layer is used for fabrication of legs and a microbridge structure and it has to have in addition specified absorption properties in the IR range, the application of the aforementioned approach for reducing of the heat conductivity of the dielectric layers enables optimizing deposition parameters of the dielectric layers first of all for the optimum absorption in the IR range and avoiding bending of the microbridge structure. Moreover, with independent of the optimization problem the aforementioned method of applying of tensile stress to the dielectric layers enables improving parameters of the detectors (i.e. reduction in heat conductivity of legs) fabricated using already optimized processes. For this purpose, it is necessary to optimize the intrinsic stress in the electrically conductive layer in a way that it causes tensile stress in the one or more adjacent dielectric layers.

The problem of fabrication optimization of structures for detection of IR radiation can include the following stages. At first, a library of technological processes for the deposition of layers used for fabrication of the structures is generated. Not only are the technological parameters of deposition included in this library but also properties of layers which were deposited as a result of use of these technological processes. The properties of the layers in the library include properties which are additively taken into account in the analysis of characteristics of the structures for IR radiation detection. For example, a knowledge of refractive indexes and absorption coefficients of materials enables calculating an absorption spectrum of a microbridge structure and/or an absorber consisting of several layers of materials which properties are included in library together with the corresponding parameters of deposition. Besides, mechanical properties of the layers, such as values of biaxial moduli and intrinsic stress, can be stored in the library. Whereas, these parameters enable selecting optimal layer thickness values for achieving the aforementioned effect of reducing the heat conductivity of legs and at the same time avoiding inadmissible deformation of a microbridge structure.

After the generation of the library, selection of layers which, according to calculation / mathematical simulation, can provide specified characteristics of one or more structural elements of structures for IR radiation detection is performed. One or more structural elements with the aforementioned characteristics enable manufacturing structures for IR radiation detection with the specified characteristics. For example, a foreknowledge of absorption coefficients and refractive indexes of single layers enable simulating absorption of a multi-layer structure consisting of several layers, i.e. absorption of IR radiation in the structure for IR radiation detection. Selection of layers can be also performed in conjunction with maximization of the aforementioned effect of reducing the heat conductivity of legs. The knowledge of mechanical properties of single layers used for fabrication of legs enables to analytically solve one of the aspects of the problem of multiple objective optimization according to which the selected layers for fabrication of legs have to provide for an absence of an intolerable deformation of legs and at the same time for maximizing the aforementioned effect of reducing the heat conductivity of legs. This can be achieved by using the electrically conductive layer for formation of legs which causes tensile stress in one or more adjacent dielectric layers used for formation of the legs as well. Besides, the problem of multiple objective optimization can include a problem of obtaining the required absorption within the framework of the aforementioned absorption simulation if at least one layer is used for the formation of the legs and construction elements of the structure for IR radiation detection, wherein the construction elements absorb IR radiation and transfer the absorbed IR radiation to the temperature responsive detector (for example a microbridge structure and/or an absorber). Afterwards, fabrication of structures and evaluation of their characteristics are performed. At this stage, not only the measurement of the characteristics which were previously simulated on the basis of the properties of materials (for example an absorption spectrum), but also the measurement of characteristics which are determined by complex effects due to cross impact of processes in layers (for example the heat conductivity of the leg) are carried out. The heat conductivity of the legs consisting of several layers, in addition to other factors, is determined by the following ones: a mean free path of phonons transferring heat, which is comparable with the dimensions of a leg cross-section; a scattering of the phonons on boundaries of layers; and the aforementioned effect of reducing the heat conductivity in dielectric layers. Also, at this stage, the measurement of integral characteristics of structure for IR radiation detection such as the sensitivity in the operating range can be performed. If as a result of the measurements, it is found that characteristics of the structures do not correspond to the specified ones, the selection of a new combination of layers and/or optimization of thickness of already selected layers can be made. Further, new structures are fabricated taking into account the changes made.

The process of such iterative optimization can be repeated several times until the measured characteristics of newly fabricated structures correspond to the specified values (for example until the measured characteristics are within the specified intervals) or until this iterative process reaches saturation, i.e. the changes in (improvements of) one or more characteristics resulting from the next iteration are insignificant (for example when the relative change of one or more characteristics is less than a predefined value). The electrically conductive layer being adjacent to one or more dielectric layers in legs can be made of any conductive stable stiff material capable of causing tensile stresses in the one or more adjacent dielectric layers which lead to a reduction in their heat conductivity. Nitrides of transition metals of the fourth, fifth or sixth groups of the Periodic Table, such as titanium, vanadium, chrome, zirconium, niobium, molybdenum, hafnium, tantalum or tungsten can be suitable materials. The conductive layer can be also made of solid solutions or alloys of these metals or their nitrides. Besides, the conductive layer can be a multi-layer, i.e. consist of several conductive layers. The conductive layer can be deposited by magnetron sputtering, plasma enhanced chemical vapor deposition (PECVD), or ion beam sputtering. Kauffman's source can be used as a source of the ion beam. Nitrides of metals can be deposited using targets of the respective metal nitrides or by reactive sputtering method. In case of the reactive sputtering using the ion beam, the nitrogen can be let directly into a vacuum chamber, in which the deposition is performed, and/or via an additional ion beam source for nitriding of a surface of the layer being deposited.

The dielectric layers can be fabricated, for example, of nitride, oxide, or oxy-nitride of silicon. Each of the dielectric layers can be a multi-layer, i.e. consist of several dielectric layers. The dielectric layers can be deposited by magnetron sputtering, PECVD or ion beam sputtering. In order to neutralize a charge of a sputtering target surface and/or a surface of a substrate on which the deposition is performed, an auxiliary source of electrons can be used. Nitrides, oxides and oxy-nitrides of silicon can be deposited using targets of the corresponding nitrides, oxides and oxy-nitrides of silicon or by reactive sputtering method. In a case of reactive sputtering using the ion beam, nitrogen and/or oxygen can be let directly into a vacuum chamber in which the deposition is performed, and/or inflow of oxygen and/or nitrogen can be implemented via an additional ion beam source for nitriding and/or oxidation of a surface of the layer being deposited.

In order to generate the tensile stresses in one or more dielectric layers adjacent to an electrically conductive layer in legs, the following process can be used: formation of a sacrificial layer on a substrate; formation of legs on a sacrificial layer; removal of a sacrificial layer. The formation of legs includes deposition of the following layers: deposition of the electrically conductive layer, deposition of the first one of the one or more dielectric layers before or after the deposition of the electrically conductive layer, and optional deposition of the second of the one or more dielectric layers before or after the deposition of the electrically conductive layer, wherein the second dielectric layer is deposited after the deposition of the electrically conductive layer if the latter is deposited after the deposition of the first dielectric layer, wherein the second dielectric layer is deposited before the deposition of the electrically conductive layer if the latter is deposited before the deposition of the first dielectric layer. The deposition parameters of the electrically conductive layer are selected such that the deposited electrically conductive layer has intrinsic compressive stress. The topology of the legs is selected such that the electrically conductive layer is adjacent to the first dielectric layer and to the second optional dielectric layer when the latter is deposited. In case of such fabrication of legs, the electrically conductive layer causes the tensile stresses in the adjacent first dielectric layer (and in the adjacent second dielectric layer when it is deposited) after removal of the sacrificial layer. Employment of deposition processes enabling deposition of the dielectric layers having no intrinsic stress before the removal of the sacrificial layer is preferable for deposition of the dielectric layers.

Mechanical properties of the layers used for fabrication of legs can satisfy the following rules enabling optimizing legs in a way that their bending is insignificant and/or planarity and parallelism of a microbridge structure with respect to a substrate surface are not distorted:

1. I M _\ *d _\ - M ₂ *d ₂ I / (M ₁ *d ₁ +M ₂ *d ₂ ) < 0.1, where "I I" is an operation of calculation of the modulus, i.e. absolute value, "*" is a multiplication operation, M and di are the biaxial modulus and thickness of the first dielectric layer adjacent to one surface of the conductive layer in legs, M ₂ H d ₂ are the biaxial modulus and thickness of the second dielectric layer adjacent to opposite surface of the conductive layer in legs. In some cases more rigorous criterion is preferable: I - M ₂ *d2 I / (M ₁ *d ₁ +M ₂ *d ₂ ) < 0.05. These criteria are applicable when the following is fulfilled: there is no gradient of intrinsic stress throughout a thickness of the electrically conductive layer in the fragments of the electrically conductive layer used for fabrication of legs and there is no intrinsic stress in the first and second dielectric layers immediately after their deposition and until when, in the course of fabrication of a pixel of the IR detector, the electrically conductive layer having intrinsic compressive stress causes tensile stresses in fragments of the first and second dielectric layers used for formation of legs.

2. M!*di > M ₂ *d ₂ , this criterion is applicable, when the intrinsic compressive stress in the electrically conductive layer changes substantially throughout the layer thickness, when an absolute value of the compressive stress in the electrically conductive layer at an interface with the first dielectric layer having the biaxial modulus Mi and the thickness di is bigger than an absolute value of the compressive stress in the electrically conductive layer at an interface with the second dielectric layer having the biaxial modulus M ₂ and the thickness d ₂ . The aforementioned criterion can be formulated for "a mirror case" when the absolute value of the compressive stress in the electrically conductive layer at the interface with the first dielectric layer having the biaxial modulus Mi and the thickness di is less than the absolute value of the compressive stress in the electrically conductive layer at an interface with the second dielectric layer having the biaxial modulus M ₂ and the thickness d ₂ .. In the latter case Mi*di < M ₂ *d ₂ . Besides an additional constraint has to be complied with for applicability of the criteria of point 2. It should be no intrinsic stress in the first and second dielectric layers immediately after their deposition and until when, in the course of fabrication of a pixel of the IR detector, the he electrically conductive layer having intrinsic compressive stress causes tensile stresses in fragments of the first and second dielectric layers used for formation of legs.

When only one dielectric layer is adjacent to the electrically conductive layer in the leg, then in order to avoid intolerable bending of the leg, the electrically conductive layer has to have a though-the -thickness gradient of compressive stress before the removal of the sacrificial layer, i.e. the absolute value of the compressive stress near one surface of the electrically conductive layer is higher than in the vicinity of the opposite surface of the electrically conductive layer before the removal of the sacrificial layer. The dielectric layer has to be adjacent to the surface of the electrically conductive layer near which the absolute value of the compressive stress in the electrically conductive layer is higher before the removal of the sacrificial layer.

Layers without gradients of intrinsic stress can be obtained as a result of the following optimization of a deposition process. Frist at least two layers of different thickness are deposited. The thickness of one layer is equal to a specified thickness, i.e. equal to the thickness of a layer which is used for fabrication of a structure for IR radiation detection. The other layers, further called test layers, have thickness which is less than the specified thickness. For example, it is possible to use at least one test layer having thickness in the range of 30% - 40% of the specified thickness. If values of the intrinsic stress evaluated according to n Stoney's equation ² in said layer and all test layers are not within a predefined range (e.g. predefined range of +1-5%), the optimization of the deposition process is performed, as a result of which one or more tuning technological deposition parameters are selected. They are changed in the course of deposition in such a way that values of intrinsic stress in said layer and all test layers are within the given range. For example, it is possible to use the plasma discharge frequency or a ratio of the periods of plasma excitation at different frequencies for chemical vapor deposition (CVD) processes of deposition as tuning technological parameters. Whereas, for physical vapor deposition (PVD) processes, it is possible to use the intensity of the ion bombardment of a surface of the layer being growth (ion assistance) as tuning parameter.

Additional objects, features, aspects and advantages of the present invention will be set forth, in part, in the description which follows and, in part, will be obvious from the description or may be learned by practice of the invention. In the description, reference is made to the accompanying drawings which form a part thereof and in which is shown by way of illustration specific embodiments for practicing the invention. These embodiments will be described in sufficient detail to enable those skilled in the art to practice the invention, and it is to be understood that other embodiments may be utilized and that structural changes may be made without departing from the scope of the invention. The following detailed description is, therefore, not to be taken in a limiting sense, and the scope of the present invention is best defined by the appended. The accompanying drawings, which are hereby incorporated into and constitute a part of this specification, illustrate preferred embodiments of the invention and, together with the description, serve to explain the principles of the invention.

On drawings:

Fig. la - 13a illustrate a cross-section of a post on a substrate at various stages of structure fabrication;

Fig. lb - 13b illustrate a cross-section of a microbridge structure, the temperature responsive detector, an absorber, legs and a substrate at various stages of structure fabrication;

Fig. 14 illustrates a top view of a pixel with a temperature responsive detector;

Fig. 15 illustrates a 3D image of a pixel with a temperature responsive detector. Fig. 16 illustrates intrinsic stress and values of a biaxial module in silicon nitride layers versus film thickness;

Fig. 17 illustrates intrinsic stress and values of a biaxial modulus in titanium nitride layers versus film thickness; Fig. 18 illustrates intrinsic stress and values of a biaxial modulus in titanium nitride layers versus film thickness;

Fig. 19a- 19b illustrate a test structure for detection of an intrinsic stress gradient in layers of different materials;

Fig. 20a-20c illustrate test structures after removing a sacrificial layer, for films with different gradients of intrinsic stress throughout thickness;

Fig. 21a-21c illustrate a test structure for measuring of biaxial moduli of materials;

Fig. 22 illustrates a flow chart diagram of optimization stages of pixel fabrication;

Fig. 23 illustrates a pixel matrix;

Fig. 24 illustrates a pixel matrix. According to one embodiment the present invention relates to a thermal infrared detector apparatus comprising: a pixel on a semiconductor substrate, said pixel comprising a first section and a second section, the first section being on a surface of the semiconductor substrate and comprising integrated circuit means, the second section being spaced from and immediately above the first section, the second section being planar and comprising leg portions, a microbridge structure, and a temperature responsive detector on the microbridge structure, the second section being supported by posts, one of the leg portions having one end integrally connected to the microbridge structure and another end integrally connected to one of the posts, another one of the leg portions having one end integrally connected to the microbridge structure and another end integrally connected to another one of the posts, the leg portions providing an electrical connection of the temperature responsive detector to the integrated circuit means via the respective posts and a thermal insulation of the microbridge structure and the temperature responsive detector from the semiconductor substrate. The one of the leg portions comprises: a first portion of a first dielectric layer, and a portion of an electrically conductive layer, the portion of the electrically conductive layer providing the electrical connection, wherein the first portion of the first dielectric layer is adjacent to a first surface of the portion of the electrically conductive layer, wherein the portion of the electrically conductive layer is a stressor causing a first tensile stress in the first portion of the first dielectric layer.

According to another embodiment the one of the leg portions comprises: a first portion of a second dielectric layer, wherein the first portion of the second dielectric layer is adjacent to a second surface of the portion of the electrically conductive layer, wherein the first and the second surface are opposite surfaces of the portion of the conductive layer, wherein the portion of the electrically conductive layer is a stressor causing a second tensile stress in the first portion of the second dielectric layer.

According to another embodiment an absolute value of compressive stress decreases through a thickness of the portion of the electrically conductive layer from the first surface to the second surface.

According to another embodiment the electrically conducting layer comprises nitride of one or more of the following metals: titanium, vanadium, chrome, zirconium, molybdenum, hafnium, tantalum, or tungsten. According to another embodiment a first product is a product of a layer thickness of the first portion of the first dielectric layer and a value of biaxial modulus of the first portion of the first dielectric layer, wherein a second product is a product of a layer thickness of the first portion of the second dielectric layer and a value of biaxial modulus of the first portion of the second dielectric layer, wherein an absolute value of a difference of the first product and the second product divided by a sum of the first product and the second product is less than 0.1.

According to another embodiment the first dielectric layer is a multilayer structure comprising several layers and/or the second dielectric layer is another multilayer structure comprising several layers, wherein each of the layers of both first and second dielectric layers has a respective product of its biaxial modulus and its thickness, wherein the first sum is equal to a sum of the products of all of the layers of the first dielectric layer, wherein the second sum is equal to a sum of the products of all of the layers of the second dielectric layer, wherein an absolute value of a difference of the first sum and the second sum divided by a sum of the first sum and the second sum is less than 0.1.

According to another embodiment a first product is a product of a layer thickness of the first portion of the first dielectric layer and a value of biaxial modulus of the first portion of the first dielectric layer, wherein a second product is a product of a layer thickness of the first portion of the second dielectric layer and a value of biaxial modulus of the first portion of the second dielectric layer, wherein the first product is bigger than the second product when an absolute value of compressive stress decreases through a thickness of the portion of the electrically conductive layer from the first surface to the second surface.

According to another embodiment the first dielectric layer is a multilayer structure comprising several layers and/or the second dielectric layer is another multilayer structure comprising several layers, wherein each of the layers of both first and second dielectric layers has a respective product of its biaxial modulus and its thickness, wherein the first sum is equal to a sum of the products of all of the layers of the first dielectric layer, wherein of the second sum is equal to a sum of the products of all of the layers of the second dielectric layer, wherein the first sum is bigger than the second sum product when an absolute value of compressive stress decreases through a thickness of the portion of the electrically conductive layer from the first surface to the second surface.

According to another embodiment the first dielectric layer comprises one or more of the following layers: a first layer of silicon nitride, a first layer of silicon oxide, a first layer of silicon oxynitrite.

According to another embodiment the second dielectric layer comprises one or more of the following layers: a second layer of silicon nitride, a second layer of silicon oxide, a second layer of silicon oxynitrite.

According to another embodiment the pixel further comprises infrared radiation absorber being thermally coupled to the temperature responsive detector, wherein the infrared radiation absorber comprises a second portion of the first dielectric layer and/or a second portion of the second dielectric layer, wherein the leg portions provide the thermal insulation of the infrared radiation absorber from the semiconductor wafer.

According to another embodiment the microbridge structure comprises a third portion of the first dielectric layer and/or a third portion of the second dielectric layer.

According to another embodiment the second portion of the first dielectric layer is comprised in the third portion of the first dielectric layer if the infrared radiation absorber comprises the second portion of the first dielectric layer and the microbridge structure comprises the third portion of the first dielectric layer, wherein the second portion of the second dielectric layer is comprised in the third portion of the second dielectric layer if the infrared radiation absorber comprises the second portion of the second dielectric layer and the microbridge structure comprises the third portion of the second dielectric layer. According to another embodiment the temperature responsive detector comprises a vanadium oxide, VO _x , layer, wherein 1.7<x<1.9, wherein electrical resistance of the vanadium oxide decreases with increase in temperature in a range of 1.7-2.6 % per degree Kelvin.

According to another embodiment said pixel comprises an infrared reflective layer covering at least a portion of the first section and facing the second section.

According to another embodiment the thermal infrared detector apparatus of further comprises an array of said pixels.

According to another embodiment the present invention relates to a method for fabrication of an infrared thermal detector apparatus. The infrared thermal detector apparatus comprises an array of pixels on a semiconductor substrate. Each of the pixels comprises a respective first section of the each of the pixels and a respective second section of the each of the pixels being spaced from and immediately above said respective first section and being supported by at least two respective posts, wherein each of the posts supports only one of the second sections or each of the posts of a first portion of the posts supports only two of the second sections and each of the posts of a second portion supports only one of the second sections. Each of the first sections is on a surface of the semiconductor substrate and comprises respective integrated circuit means of the each of the first sections. Each of the second sections is planar and comprises respective leg portions of the each of the second sections, a respective microbridge structure of the each of the second sections, and a respective temperature responsive detector of the each of the second sections on to said respective microbridge structure. The method comprising, for each of the pixels, the step of: forming the respective leg portions, wherein each of the formed leg portions has one end integrally connected to the microbridge structure of the pixel comprising the formed leg portions and another end integrally connected to the respective post supporting the second section comprising the formed leg portions, wherein at least two of the formed legs have their ends integrally connected to the different posts, wherein the formed leg portions provide an electrical connection of the temperature responsive detector of the pixel comprising the formed leg portions to the integrated circuit means of the pixel comprising the formed leg portions via the respective posts, wherein the formed leg portions further provide thermal insulation of the microbridge structure of the pixel comprising the formed leg portions and the temperature responsive detector of the pixel comprising the formed leg portions from the semiconductor substrate. The method further comprises the step of: forming a sacrificial layer on the semiconductor substrate. The forming of the respective leg portions comprises the following steps for forming of one of said respective leg portions: forming a first portion of a first dielectric layer; and forming a portion of an electrically conductive layer, wherein the portion of the electrically conductive layer provides the electrical connection, wherein the formed portion of the electrically conductive layer has compressive stress, wherein the portion of the electrically conductive layer is adjacent to the first portion of the first dielectric layer. The method further comprises the step of: removing the sacrificial layer, wherein after the removing of the sacrificial layer the portions of the electrically conductive layer function each as stressor causing a tensile stress in the respective first portion of the first dielectric layer.

According to another embodiment the forming of the respective leg portions comprises further the following steps for forming of one of said leg portions: forming a first portion of a second dielectric, wherein the portion of the electrically conductive layer is adjacent to the first portion of the second dielectric layer, wherein in a case when the first portion of the second dielectric layer is formed on the portion of the electrically conductive layer the first portion of the first dielectric layer is formed on the sacrificial layer and the portion of electrically conductive layer is formed on the first portion of the dielectric layer, wherein in a case when the first portion of the first dielectric layer is formed on the portion of the electrically conductive layer the first portion of the second dielectric layer is formed on the sacrificial layer and the portion of electrically conductive layer is formed on the second portion of the dielectric layer, wherein after the removing of the sacrificial layer the portions of the electrically conductive layer function each as stressor causing a tensile stress in the respective second portion of the first dielectric layer.

According to another embodiment each of the pixels comprises a respective infrared radiation absorber, wherein each of the infrared radiation absorbers is thermally coupled to the exactly one of the temperature responsive detectors of the pixel comprising the each of the infrared radiation absorbers, a respective second portion of the first dielectric layer and/or a respective second portion of the second dielectric layer, and is thermally insulated from the semiconductor wafer by the leg portions of the pixel comprising the each of the infrared radiation absorbers. The method further comprises, for each of the pixels, the step of: forming the respective infrared radiation absorber, wherein the forming of the respective infrared radiation absorber comprises the steps of: forming the respective second portion of the first dielectric layer on the sacrificial layer if the respective infrared radiation absorber to be formed in the step of the forming of the respective infrared radiation absorber comprises the respective second portion of the first dielectric layer; and forming the respective second portion of the second dielectric layer if the respective infrared radiation absorber to be formed in the step of the forming of the respective infrared radiation absorber comprises the respective second portion of the second dielectric layer. According to another embodiment, each of the microbridge structures comprises a respective third portion of the first dielectric layer and/or a respective third portion of the second dielectric layer, the method further comprising, for each of the pixels, the step of forming the respective microbridge structure, wherein the forming of the respective microbridge structure comprises the steps of: forming the respective third portion of the first dielectric layer on the sacrificial layer if the respective microbridge structure to be formed in the step of the forming of the respective microbridge structure comprises the respective third portion of the first dielectric layer; and forming the third portion of the second dielectric layer if the respective microbridge structure to be formed in the step of the forming of the respective microbridge structure comprises the respective third portion of the second dielectric layer.

According to another embodiment the second portions of the first dielectric layer are comprised each in the respective third portion of the first dielectric layer if the infrared radiation absorbers comprise each the respective second portion of the first dielectric layer and the microbridge structures comprise each the respective third portion of the first dielectric layer, wherein the second portions of the second dielectric layer are comprised each in the respective third portion of the second dielectric layer if the infrared radiation absorbers comprise each the respective second portion of the second dielectric layer and the microbridge structures comprise each the respective third portion of the second dielectric layer.

According to another embodiment the forming of the first portion of the first dielectric layer on the sacrificial layer comprises deposition of the first dielectric layer on the sacrificial layer using a process enabling the deposition of the stress-free first dielectric layer, wherein the forming of the first portion of the second dielectric layer on the portion of the electrically conductive layer comprises deposition of the second dielectric layer on the portion of the electrically conductive layer using a process enabling the deposition of the stress-free second dielectric layer. According to another embodiment, each of the microbridge structures comprises a respective third portion of the first dielectric layer and/or a respective third portion of the second dielectric layer, wherein the second portions of the first dielectric layer are comprised each in the respective third portion of the first dielectric layer if the infrared radiation absorbers comprise each the respective second portion of the first dielectric layer and the microbridge structures comprise each the respective third portion of the first dielectric layer, wherein the second portions of the second dielectric layer are comprised each in the respective third portion of the second dielectric layer if the infrared radiation absorbers comprise each the respective second portion of the second dielectric layer and the microbridge structures comprise each the respective third portion of the second dielectric layer.

Figs, la- 13a and Figs, lb- 13b illustrate cross-sections of a post, legs, a microbridge structure, a temperature responsive detector, a substrate, and an absorber in the course of fabrication of the pixel with the temperature responsive detector. As it will be shown further, pixels can be integrated in a pixel matrix enabling obtaining of images in the IR range. At the initial stage, electronic circuits and logical transistor elements 102 for registering a state of the temperature responsive detector are manufactured on a substrate 100. In a case when a vanadium oxide layer with a high temperature coefficient of electrical resistance is used as the temperature responsive detector, the electronic circuits and the logical transistor elements are configured to measure an electrical resistance of the vanadium oxide layer. Fig. la shows the aluminum contact 101 of the electronic circuits on which afterwards the post supporting a microbridge structure will be fabricated. As it will be shown further, the electrical contact of the temperature responsive detector with electrical circuits is implemented via the aluminum contact 101 and the post on it.

Fig. 2a and fig. 2b illustrate the next stage of pixel manufacturing: deposition and structuring of a layer of an IR reflector. The part of the layer 103b functions as the IR reflector, another part of the layer 103a performs protective function of the aluminum contact 101, and also, it can improve an adhesion of materials of which the post is made. This layer can be made of different metals with good reflectivity in the IR range. For example, this layer can be made of titanium, nickel, chrome, or nickel - chrome alloy. Also, this layer can be a multi-layer and consist of several metal layers. This layer can be deposited by a magnetron sputtering, an ion beam sputtering, or thermal evaporation.

Fig. 3 a and fig. 3b illustrate a further stage of the pixel manufacturing. At this stage a sacrificial layer 104 is deposited and structured. It can be structured with a mask of the structured layer 105 by the plasma-chemical etching method in oxygen or oxygen - hydrogen plasma. As a result of structuring a sacrificial layer, the window over contact 101 is formed. The bottom of the formed window opens a surface of the contact 101 if the element 103a was not fabricated or a surface of the element 103a if it was fabricated.

The organic materials like polyimide, photo resist, or inorganic materials on a basis of the low- temperature glasses produced by PECVD or by spin coating of solutions can be used as the sacrificial layer. The layer 105 can be removed after structuring the sacrificial layer 104. If the layer 105 is not removed after the structuring of the sacrificial layer, its fragments can be used for fabrication of legs and/or a microbridge structure.

Alternatively, the photosensitive sacrificial layer can be employed. In this case the structuring process of such layer does not require utilization of the mask 105. The structuring of the photosensitive layer is carried out in the same way as photo resist structuring, i.e. by exposure with the subsequent development and heat treatment.

Depending on processing of the sacrificial layer and/or structuring process of the sacrificial layer, the window can have vertical walls or, as shown in Fig. 3a, the window can be tapered. The tapered window can be obtained as a result of development of the photosensitive sacrificial layer with the subsequent heat treatment or as a result of optimization of etching process.

A thickness of the sacrificial layer is defined by the IR range in which the pixel is ought to function. For example, if the pixel is designed for operation in the IR range of 8-12 μιη, the thickness of the sacrificial layer on the reflector 103b is selected in the range of 1.7 - 2.8 μιη. Preferable thickness of the sacrificial layer on the reflector is equal to a quarter of a wavelength which is an arithmetic mean value of maximum wavelength and minimum wavelength of the operational IR range of the device.

After the structuring of the sacrificial layer 104 as shown in Fig. 3a and Fig. 3b, the deposition of a dielectric layer 106 is made as shown in Fig. 4a and Fig.4b. As it will be shown further, after structuring different parts of this layer, will be used for fabrication of legs and a microbridge structure. As an alternative this layer can be used for fabrication of either a microbridge structure or legs only. For this purpose, the layer 106 has to be appropriately structured. If this layer is used only for fabrication of a microbridge structure, the part of this layer has to be deleted in the area of structure, the cross-section of which is illustrated in Fig. 4a. If this layer is used only for fabrication of legs, the part of this layer has to be deleted in the area of structure the cross-section of which is illustrated in Fig. 4b.

The illustration of the following manufacturing stage is given in Figs. 5a and 5b. The deposition of the temperature responsive detector 108 and protective dielectric 109 is made on a photoresist mask 107. The layer 109 is an optional element of the structure. Only the layer 108 can be used in certain variants of structure. The layer 108 can be made of vanadium oxide, VO _x , where 1.7 < x < 1.9. The layers with such chemical composition have the negative coefficient of change in electrical resistivity in the range of 1.7 - 2.6% per degree Kelvin. The layer of vanadium oxide can be deposited by a method of reactive ion beam sputtering of vanadium target, by a method of ion sputtering of vanadium target with a use of an auxiliary ion beam of oxygen ions directed towards a surface of a film being grown, or by a method of pulsed reactive magnetron sputtering. Afterwards, a structuring of layers 108 and 109 is made by lift-off lithography method as shown in fig. 6a and 6b. As a result the structure is formed of parts of layers 108 and 109 located one above another or only of a part of a layer 108 on a layer 106 or 105 (depending on the options of a process flow described above). This structure is located over a layer 103b. As an alternative, the layers 109 and 108 or if the layer 109 is not used, the layer 108 only can be structured by etching with a photoresist mask. Etching can be implemented by an ion beam etching. The layer 108 can be deposited by reactive magnetron sputtering method using vanadium target in the argon and oxygen atmosphere or by reactive ion beam deposition method using vanadium target and ion beam of argon and oxygen ions for vanadium target sputtering. Thickness of the layer 108 used for formation of the temperature responsive detector can be in the range of 50-150 nanometers.

In Figs. 7a and 7b, the formation of electrical contacts to a layer of the thermal detector 109 is illustrated. Contacts can be made of any metal for example of vanadium which provides the ohmic contact to the layer 108. For effective use of the structure area of the thermal detector, the pair of contacts 119 is located near the opposite sides of structure 108 as shown in fig. 7b. If the layer 109 is used in the course of manufacture, the etchback of windows is made in a layer 109 for contacts 119 before formation of contacts 119, as shown in fig. 7b. Thickness of the layer used for formation of contacts 119 can be in the range of 30-150 nanometers.

The following step of structure formation is given on figs. 8a and 8b. As shown on figs. 8a, the window is etched in a layer 106 over contact 103a. Afterwards this window provides formation of electrical contact between contact 101 and contact 119. The window opens a part of a surface of an element 103a as shown in fig. 8a. If the element 103a is not used, the window opens a part of a surface of contact 101. Formation of a window is made if the layer 106 is used in the course of manufacture.

On Figs. 9a and 9b the following step of structure formation is illustrated. In this step an electrically conductive layer 112 is deposited. This layer provides not only electrical contact between contacts 119 and 101 where contact 101 can be coated with a layer 103a. As it will be shown further, the part of this layer is used for formation of legs, therefore for implementing the effect of reducing the heat conductivity of dielectric layers of legs, this layer has to be made of stiff and stable material with high intrinsic compressive stress which does not decrease in time due to properties of the material. Nitrides of transition metals of the fourth, fifth or sixth groups of the Periodic Table, such as titanium, vanadium, chrome, zirconium, niobium, molybdenum, hafnium, tantalum or tungsten can be suitable materials for fabrication of the layer 112. The electrically conductive layer can be also made of solid solutions or alloys of these metals or their nitrides. Besides, the conductive layer can be multi-layer, i.e. consist of several conductive layers. These materials can be obtained by reactive magnetron sputtering method in a nitrogen and argon atmosphere with use of targets of the appropriate metals. Thickness of the layer 112 can be in the range of 10-100 nanometers.

On Figs. 10a and 10b the next step of structure formation is illustrated. In this step the bearing element 113 of a post 118 is formed. As shown in Fig. 10a the bottom and walls of the window, which have been earlier formed in the sacrificial layer, are coated by a material of this element 113. Besides, the material of the element 113 can form a seal over a sacrificial layer along a perimeter of the window in the sacrificial layer. The element 113 can be made by a method of magnetron sputtering of aluminum or other suitable metal with the subsequent structuring by etching with a photoresist mask. A thickness of the metal layer used for formation of the element 113 can be in the range of 0.3-1.5 micrometers.

In Figs. 11a and l ib, the next step of structure formation is illustrated. In this process step a layer of dielectric 114 is deposited. As shown in Figs. 11a and l ib, the dielectric covers a conductive layer 112 and an element of the bearing post 113. Deposition of the dielectric layer 114 is optional.

In Figs. 12a and 12b, the next step of structure formation illustrated, wherein legs 116 are formed. This step can be carried out by etching of layers 105, 106, 112, 114 with a photoresist mask. As a result gaps 115 separating the legs 116 from the microbridge structure 117 are made.

The nest step of structure formation is illustrated on Figs. 13a and 13b. In this step the sacrificial layer 104 is removed. This step can be performed by reactive ion etching (RIE) in an oxygen atmosphere. As a result, both the microbridge structure 117 and the legs 116 are suspended over a substrate surface 100 or, depending on a pixel configuration, over a surface of the layer 102 including electronic circuits and logical transistor elements. Mechanical connection of the microbridge structure with the substrate is provided by legs 116 and posts 118 as shown on the pixel top view in Fig. 14. Elements 103a, 103b and details of a surface relief of a microbridge structure are not shown in fig. 14 for the sake of the demonstrativeness of the picture. On Fig. 14 the contours of the structures 112a and 112b made of the electrically conductive layer 112, contours of the structures of contacts 119, and contours of the structured layers 108 and 109 are illustrated. Figs, la- 13a illustrate cross-section views A-A of elements of the pixel depicted on Fig. 14 throughout the manufacturing steps. Figs, la- 13a illustrate cross-section views B-B of elements of the pixel depicted on Fig. 14 throughout manufacturing steps the.

On Fig. 15 the three dimensional image of a pixel illustrating micro-membrane 117 with the temperature responsive detector is depicted. The heat insulation of a microbridge structure is provided by a pair of legs 116, whereas the legs are connected to a pair of posts 118 supporting all these elements over a surface of the substrate 100 with electronic circuits, logical transistor elements 102 and a reflector 103b.

As it is easy to see from the aforementioned description of the manufacturing process and Figs. la- 13 and Figs, lb - 13b, different fragments of dielectric layers 105, 106 and 114 can be used for formation of legs and/or a microbridge structure. Besides, depending on the absorption properties of these layers in the operating IR range of pixel the fragments of these layers can function as an absorber. Also fragments of layers 109 and 108 can function as an absorber if they absorb radiation in the operating IR range of pixel. For example, depending on a specific process flow, legs can be made of fragments of the following groups of layers: 105, 106, 112, 114; 106, 112, 114; 105, 112, 114; 106, 105, 112; 105, 112; 106, 112; or 114, 112. Whereas, depending on a specific process flow, the microbridge structure can be made of fragments of the following groups of dielectric layers: 105, 106, 114; 106, 114; or 105, 114.

Dielectric layers 105, 106, 109 and 114 can be made of different dielectrics, such as silicon oxide, silicon nitride, or silicon oxy-nitride. These layers can be made, for example, by magnetron sputtering or PECVD. Thickness of these layers can be in the range of 50-200 nanometers. Whereas, each of layers can consist of one or more dielectric layers. The choice of layers, their thickness, and parameters of deposition can be selected on the basis of optimum absorption in the wavelength range in which the IR radiation detector has to operate. For example, the PECVD process of silicon oxide deposition can be optimized by selection of a discharge power and values of SiH ₄ and N ₂ 0 gas flows. Depending on different values of these technological parameters, it is possible to obtain different values of absorption in the IR range caused by molecular bonds of Si-O, Si-H and Si-O-H having peaks at wavelengths of 9.6 μιη, 4.4 μπι, and 3 μηι respectively. Similarly, the absorption in the IR range for the silicon nitride and oxy-nitride of silicon obtained by PECVD can be optimized.

The intrinsic stress in layers of the aforementioned dielectrics can be also optimized by method of optimization of technological parameters of layer growth. For example, in a case of silicon nitride deposition in an atmosphere of Si¾ and NH ₃ , an increase in a SiH ₄ flow or increase in power of discharge causes increase in compressive stress, while an increase in a process gas pressure enables reducing an absolute value of compressive stress and even obtaining layers with tensile stress. The process of PECVD deposition of silicon oxide in an atmosphere of N ₂ 0 and S1H4 shows another behavior. An increase in a SiFL; flow causes a shift of intrinsic stress of silicon oxide layers towards the tensile stress, while an increase in power causes an opposite effect. In case of magnetron sputtering an intrinsic stress in layers is also optimized by selecting technological parameters. As usual, the layers deposited at low gas pressures have compressive stress, and the layers deposited at high pressures have tensile stress. Layers can also have gradients of intrinsic stress through their thicknesses. Optimization of technological parameters can minimize gradients of intrinsic stress. In case of usage of PECVD processes, employment of systems using a low frequency (LF) power at frequencies of 50-500 kHz in addition to a high-frequency (HF) power at a frequency of 13.56 MHz or other multiple frequencies for plasma generation enables minimization of gradients of intrinsic stress in layers. The frequency interval of 100-300 kHz is more preferable for generation of LF power. The use of LF power for excitation of plasma enables creating layers with the compressive stress while the use of HF power enables creating layers with tensile stress . The ratio of duration of time intervals when plasma is alternately excited by HF or LF power, enables controlling intrinsic stress in the deposited layers. The increase in the percentage ratio of time intervals when plasma is excited by LF power, enables reducing values of intrinsic stress, i.e. reducing the values of tensile stress or increasing the absolute values of compression stress, or transiting from tensile stress to compression stress, etc. Change in a ratio of time intervals of plasma excitation by means of LF power and HF power can be implemented in a course of growth of a layer and therefore it can be used for optimization of growth parameters of a layer for creating layers without gradients of intrinsic stress on thickness. For example, if the layer grown in the PECVD process in the course of which all technological parameters were fixed has a gradient of intrinsic stress, wherein the absolute value of compressive stress in a portion of the layer adjacent to a substrate is bigger than the absolute value of compressive stress in near-surface portion of the layer, in order to minimize a through-the-thickness gradient of intrinsic stress, it is necessary to increase duration of time intervals when plasma is excited by LF power in a course of growth of the layer, and to decrease duration of time intervals when plasma is excited by HF power. In other example, if a thin film grown in PECVD process in the course of which all technological parameters are fixed has a gradient of intrinsic stress, wherein the tensile stress in a portion of the layer adjacent to a substrate is higher than the absolute value of tensile stress in near-surface portion of a layer, in order to minimize a through-the-thickness gradient of intrinsic stress of a layer, it is necessary to increase duration of time intervals when plasma is excited by HF power in the course of growth of a film, and to reduce duration of time intervals when plasma is excited by LF power. Similar approach is applicable for minimization the gradients of intrinsic stress through- - thicknesses of layers (thin films) made by magnetron sputtering. As a tuning parameter enabling minimization of gradients of intrinsic stress through the-thickness of layers (thin films), it is possible to use the value of HF power applied to the substrate holder. For example, if a thin film which is grown in the course of magnetron sputtering, in the course of which all technological parameters remained unchanged, has a gradient of intrinsic stress wherein an absolute value of compression stress in a portion of the film adjacent to a substrate is higher than an absolute value of compression stress in a near-surface portion of the film, it is necessary to increase the HF power applied to the substrate holder in a course of growth of a film, in order to minimize of a gradient of intrinsic stress through its thickness. In other example, wherein a thin film deposited using magnetron sputtering process in the course of which all technological parameters are fixed has a gradient of intrinsic stress, wherein an absolute value of compressive stress in a portion of the film adjacent to a substrate is less than an absolute value of compressive stress in a near- surface portion of the film, it is necessary to reduce the HF power applied to the substrate holder in a course of growth of a film in order to minimize a gradient of intrinsic stress through its thickness.

The aforementioned approach is also applicable for minimization of the gradients of intrinsic stress through -thicknesses of layers (thin films) made by ion beam sputtering. In this case one more source of ions of inert gases irradiating a surface of a layer being deposited is used. As the tuning parameter enabling minimization of gradients of intrinsic stress through-the-thickness of layers (thin films), it is possible to use the intensity of the ion beam irradiating a surface of a layer being deposited. For example, if a thin film which is grown in a deposition process throughout which all technological parameters remained fixed, has a gradient of intrinsic stress, wherein an absolute value of compressive stress in a portion of the film adjacent to a substrate is higher than an absolute value of compressive stress in a near-surface portion of the film, it is necessary to increase the intensity of an ion beam irradiating a surface of a layer being deposited in the course of the growth of the layer in order to minimize a through-the-thickness gradient of intrinsic stress. In other example if the thin film which is grown in the course of a deposition process during which all technological parameters were fixed, has a gradient of intrinsic stress, wherein an absolute value of compressive stress in a portion of the film adjacent to a substrate is less than an absolute value of compressive stress in near-surface portion of the film, it is necessary to reduce the intensity of an ion beam irradiating a surface of a layer being deposited in the course of the growth of the layer in order to minimize a through-the-thickness gradient of intrinsic stress. Application of these technologies for minimization of through-the-thickness gradients of intrinsic stress using a method of tuning the aforementioned parameters in a course of growth of a film is illustrated in Figs. 16 - 18. Each point depicted on these diagrams corresponds to a film the thickness plotted on the respective horizontal axis and intrinsic stress plotted on the respective vertical axis. Values of intrinsic stress in films were calculated according to the Stoney's equation on the basis of measurements of the substrate bending caused by intrinsic stress in a film. The specially prepared thin silicon wafers 100-200 micrometers thick having crystalline orientation (100) were used as substrates.

Each point on diagram depicted on Fig. 16 corresponds to a silicon nitride film which is grown by PECVD in a reactor with parallel electrodes and a capacitive coupling of plasma. The curve denoted by circles in Fig. 16 shows the intrinsic stress in the silicon nitride films which are grown using the same technological parameters: discharge power 100 Watts, S1H4 flow of 75 seem (standard cubic centimeter per minute), N¾ flow of 500 seem, pressure of gases in the course of deposition 77 Pa, substrate temperature 160° C; the ratio of the time interval when plasma is excited by LF power (T _LF ), to a total time interval when plasma is excited by HF (T _HF ) or LF (TLF) power is equal to 0.5 ( T _LF /(T _LF + T _HF ) =0.5 ). As clearly demonstrated by the curve denoted by circles, intrinsic stress in thin films strongly depends on thickness, i.e. there is a gradient of intrinsic stress through film thickness. In order to eliminate the gradients of intrinsic stress, the change of the TLF ( _LF + THF) parameter is used. As a result of optimization, the regime, wherein the T _LF /(T _LF + a T _HF ) parameter is gradually decreased from 0.5 to 0.1 in the course of deposition of 200 nm thick film, was selected. In case of deposition of thinner films, the T _LF /(T _LF + T _HF ) parameter gradually decreased from 0.5 to the corresponding value in the range of 0.1-0.5. For example, in a course of deposition of 50 nanometers thick film, the TLF/(TLF + T _HF ) parameter is gradually decreased from 0.5 to 0.36. All films which are grown in this regime with tuning of the T _LF /(T _LF + a T _HF ) parameter in the course of growth of films have the same value of intrinsic stress being close to zero as clearly shown by the curve denoted by triangles on Fig. 16.

On Fig. 17 the values of intrinsic stress of titanium nitride films of different thickness which are grown by method of reactive magnetron sputtering of titanium target in glow discharge of direct current plasma, are given. The curve denoted by circles on Fig. 17 shows the intrinsic stress in the titanium nitride films grown using the same technological parameters: power of direct current discharge 100 Watts, Ar flow 100 seem, N ₂ flow 7 seem, substrate holder temperature 20 ^° C, deposition chamber pressure 0.7 Pa. As this curve shows, the absolute value of compressive stress decreases with increase in film thickness. In order to eliminate gradients of intrinsic stress, the change in HF power applied to the substrate holder is used. As a result of optimization, the mode wherein the value of HF power applied to substrate holder is gradually increased from 0 to 20 Watts in the course of growth of 200 nanometers thick film, was selected. In case of deposition of thinner films, the value of HF power applied to the substrate holder is gradually increased from 0 to the corresponding value in the range of 0 - 20 Watts. For example, in a course of deposition of 50 nanometer thick film, the value of HF power applied to the substrate holder is gradually increased from 0 to 3 Watts. All films which are grown in this regime with tuning of HF power applied to the substrate holder in the course of growth of films, have the same value of intrinsic stress as shown by the curve denoted by triangles on Fig. 17. On Fig. 18 the values of intrinsic stress in vanadium nitride films of different thickness grown by reactive ion beam sputtering using two ion beam sources are depicted. The first source was used for sputtering of a vanadium target by argon ions. The second source was used for nitriding of a surface of a layer being grown. Argon was injected in the first source. A mixture of argon and nitrogen was injected in the second source. The curve depicted by circles on Fig. 18 represents the intrinsic stress in the vanadium nitride films grown using the same technological parameters: ion current of the first source was 82 mA, ion energy of the first source was 1400 eV, argon injection in the first source was 13 seem, ion current of the second source was 42 mA, ion energy of the second source was 110 eV, argon injection in the second source was 2 seem, nitrogen injection in the second source was 2 seem. As the curve depicted on Fig. 18 clearly shows, the absolute value of compressive stress decreases with increase in film thickness. In order to eliminate gradients of intrinsic stress the change in argon flow parameter in the second source is used. As a result of optimization, the regime, wherein the argon injection in the second source is gradually increased from 2 to 3.3 seem in a course of growth of 200 nanometer thick film, was selected. Ion current of the second source was respectively increased from 42 mA to 51 mA. In case of deposition of thinner films, the argon injection was smoothly increased from 2 to a corresponding value in the range of 2 - 3.3 seem. For example, in a course of deposition of 50 nanometer thick film the value of argon injection increased from 2 seem to 2.3 seem. All films grown in this regime with adjustment of argon injection into the second source in a process of film growth have the same value of intrinsic stress as shown on the curve depicted by triangles on Fig. 18.

Presence of intrinsic stress gradients in thin films can be also checked using a test console structure shown on Figs. 20a, 20b, 20c. Fig.19a shows the top view of test structure before removing a sacrificial layer 203. Fig.19b illustrates a side view of test structure before removing a sacrificial layer 203. Two structures of the same thickness having the adjacent side are formed on a substrate 200. One of adjacent structures 203 is formed using sacrificial layer material being the same which is used in the above mentioned process of pixel formation. Another adjacent structure 201 is formed using material which etching rate is significantly lower in comparison with etching rate of a sacrificial layer in the course of the sacrificial layer removing. The structure 201 can be formed for example using silicon nitride or silicon oxide. In addition, a strip 202 of material in which it is necessary to study the homogeneity of intrinsic stress through thickness, is created on a surface of structures 201 and 203. Figs. 20a - 20c show a state of console structure formed using a strip of the material 202 being tested. In case of absence of intrinsic stress gradient in console structure 202a, it does not bend after removing a sacrificial layer as shown in Fig. 20a. In case of presence of intrinsic stress gradient in console structure 202b or 202c, it bends up, as shown in fig. 20b, or it bends down, as shown in Fig. 20c, after removing a sacrificial layer. The situation shown in Fig. 20b occurs for example when the intrinsic stress increases in the direction from the lower portion of film 202b being adjacent to structure 201 to the upper (near-surface) portion of film 202b. The situation shown in Fig. 20c occurs for example when the intrinsic stress decreases in the direction from the lower portion of film 202c being adjacent to structure 201 to the upper (near-surface) portion of film 202c.

The optimum sizes of console structures are the length being equal to 150 +-/50 micrometers and width being equal 20 +/-10 micrometers. The bending of console structures can be measured by means of an interference microscope or by means of a scanning electron microscope.

The technique based on the fabrication of the test console structures shown on Figs. 19a- 19b and 20a - 20c, confirms both the presence of through-the-thickness gradients of intrinsic stress in the films denoted by circles on Figs. 16 - 18 and absence of through-the-thickness gradients of intrinsic stress in the films depicted by triangles on Figs. 16 - 18. The comparison of intrinsic stress measurements in films obtained on the basis of measurements of a substrate bending and calculated using Stoney's equation with intrinsic stress measurements in films obtained on the basis of measurements of a console structure bending enabled to work out readily applicable numerical criterion of absence of the through-the-thickness gradient of intrinsic stress in layers of different materials, wherein utilization of said layers for manufacturing of legs and/or microbridge structure does not cause unacceptable deformations of the legs and/or or the microbridge structure. Further formulated criteria are conservative, i.e. applicable for all materials out of which the electrically conductive and dielectric layers can be made in this invention. The gradient of intrinsic stress through thickness in a layer is insignificant (or negligible within a framework of this technology), i.e. not causing unacceptable deformations of structural elements of pixel if a value of intrinsic stress in said layer differs less than 10% from a value of intrinsic stress in a test layer which is grown in the same conditions as said layer and has a thickness equal to 20-30% of thickness of said layer, wherein intrinsic stress in said layers is calculated using Stoney's equation on the basis of a substrate bending. The criterion threshold of 10% is applicable for the layers with thickness equal or exceeding 90 nanometers. In case when layer thickness is less than 90 nanometers, more rigorous formulation is preferable, wherein the values of intrinsic stress in a layer and a test layer differ from each other less than for 5%. Hereinafter the layers with negligible gradients of intrinsic stress through thickness are called as layers without intrinsic stress gradient through thickness, for the sake of simple presentation of the subject matter. Besides, if the layer and the test layer have absolute values of intrinsic stress not exceeding 15 MPa such layer is called stress-free further on in the text.

The homogeneity of mechanical properties of layers through thickness is also an important factor required for optimization of the pixel fabrication process. Values of biaxial modulus of materials, out of which the dielectric layers 105, 106, 116 and the electrically conductive layer 112 are made, can be measured using test structures shown on Figs. 21a - 21c. A top view of a test structure at one of manufacturing stages is shown on Fig. 21a, while a cross-section A- A of the same structure is shown on Fig. 21b. The structure 301 with the window filled with sacrificial material 303 is formed on a substrate 300. Material of which the structure 301 is formed is not critical; it can be any material, which etching rate in the subsequent manufacturing steps is significantly lower than etching rates of the sacrificial layer. For example, the structure 301 can be made of silicon oxide or nitride. A window in structure 301 and the part of a surface of structure 301, which is parallel to substrate surface and adjacent to a window perimeter, is covered by a layer 302 which biaxial modulus has to be measured. Such a topology provides air- tightness of the window 304 covered with the layer 302 after removing a part of the substrate located under the structure of sacrificial material 303 and the structure 303 as well as shown on Fig. 20c. Afterwards, the test structure is mounted in measuring system which measures a bend of the membrane made of the layer 302 and covering the window 304 versus, a difference of gas pressures of P] and P ₂ on the different sides from a membrane. Data of membrane bending versus the difference of the pressures measured for different test structures with different sizes of the window and the same layers 302 enables evaluation of the biaxial modulus of material of which the layer 302 is made ⁴ . This technique enables checking a dependence of the biaxial modulus on layer thickness as well. Test structures 305 with different thickness of the membrane 302 are made for this purpose. The aforementioned techniques of layer growth without through-the-thickness gradients of intrinsic stress enable obtaining of dielectric layers 105, 106, 1 16 and an electrically conductive layer 1 12 with values of the biaxial modulus being independent on layer thickness, i.e. the layers of different thickness which are grown using the same technological process have almost identical biaxial moduli differing from each other in most of the cases less than 3%. Hereinafter for the sake of simplicity of description, such layers are called as layers with uniform mechanical properties. On Figs. 16 - 18 small squares denote the values of biaxial moduli for the layers of silicon nitride and titanium nitride of different thickness which are grown in the aforementioned regime with the change of a tuning parameter providing the absence of the through-the-thickness gradient of intrinsic stress. As these diagrams show, the values of biaxial moduli for films of different thickness lie in a 3% interval.

The majority of stress-free dielectric layers grown by means of the aforementioned techniques also have the optical properties which are uniform through thickness, i.e. their absorption and refractive indexes in the IR range of 7-13 micrometers depend only on wavelength. Absorption of these layers in this interval of wavelengths is described by the aforementioned Lambert - Beer ¹ equation with an error margin below 2%, wherein absorption coefficients do not depend on thickness of the respective layers. Hereinafter, for the sake of simplicity of the description, such dielectric layers are called as layers with uniform optical properties. In addition to the aforementioned criteria imposing restrictions on intrinsic stress in single layers, the application of a number of criteria formulated as inequations linking mechanical properties of layers can be formulated. Compliance with these criteria can provide for avoiding intolerable bending of legs and as a result thereof planarity and parallelism of a microbridge structure with respect to a substrate surface: a) I Mi*di - M ₂ *d ₂ I / (M ₁ *d ₁ +M ₂ *d ₂ ) < 0.1 , where "I I" is a modulus operator, i.e. calculation of the absolute value, Mi and d _\ are the biaxial modulus and thickness of the first dielectric layer adjacent to one surface of the electrically conductive layer of legs, M ₂ and d ₂ are the biaxial modulus and thickness of the second dielectric layer adjacent to opposite surface of the conductive layer of legs. This criterion is applicable when thickness of layers is equal or exceeds 90 nanometers. In some cases, when at least one layer thinner than 90 nanometers is used, more rigorous criterion is preferable: I M^di - M ₂ *d ₂ I / (M ₁ *di+M2*d ₂ ) < 0.05. These criteria are applicable when the following additional constraints are complied with. The first and the second dielectric layer have to be stress free before the removing of the sacrificial layer 104 under the legs 1 16 and the microbridge structure 1 17. In order to explain the meaning of the this constraint, it is necessary to mention that after removing of the sacrificial layer, the conductive layer causes tensile stress in the first and the second dielectric layer, which in its own turn reduces their heat conductivity. In addition, a compressive stress has to be uniform through thickness in a conductive layer, i.e. a through-the-thickness gradient of intrinsic stress has to be absent. All layers must have uniform mechanical properties. The first and the second dielectric layer, each, have to be a single layer, i.e. not a multi-layer. The criteria formulated in this section are conservative, i.e. defined on the basis of data obtained with use of the layers made of the most rigid materials (i.e. having the highest biaxial moduli), wherein said layers are used for fabrication of legs or/and a microbridge structure in the technological process described above. Application of these criteria of this section enables simplifying optimization process of manufacturing technology. Knowing biaxial moduli of materials, it is possible to directly select the thicknesses of the dielectric layers which do not cause the aforementioned unacceptable structure deformations of the microbridge structure and/or the legs, i.e. it is possible to avoid an additional optimization of technological process based on manufacturing of the test console structures 202a-c with use of the same processes and materials that are used for fabrication of the legs (i.e. in this case, the console structure 202a-c is a multi-layer consisting of fragments of the following layers: the first dielectric layer / the electrically conductive layer / the second dielectric layer),

Mi*d _\ > M ₂ *d ₂ , this criterion is applicable when intrinsic compressive stress in the electrically conductive layer has a gradient, when an absolute value of compressive stress in the conductive layer at a boundary with the first dielectric layer having the biaxial modulus Mi and thickness d[ is bigger than an absolute value of compressive stress in the conductive layer at a boundary with the second dielectric layer having the biaxial modulus M ₂ and thickness d ₂ . In this section, the criterion formulated above can be formulated for "a mirror case" when an absolute value of compressive stress in the conductive layer at a boundary with the first dielectric layer, is less than an absolute value of compressive stress in the conductive layer at a boundary with the second dielectric layer. In this case M^d} < M ₂ *d ₂ . The criteria of this section are applicable when the following constraints are fulfilled. The layers used for the manufacturing of legs have to be deposited according to the following sequence. The first dielectric layer has to be deposited on the sacrificial layer. The electrically conductive layer has to be deposited on the first dielectric layer. The second dielectric layer has to be deposited on the electrically conductive layer. As well as in section a) the first and the second dielectric layer have to be stress-free before removal of the sacrificial layer. In addition the first and the second dielectric layer, each, have to be a single layer i.e. not a multilayer. Besides, the first and the second dielectric layer must have mechanical properties being uniform through thickness. The criteria formulated in this section are conservative, i.e. defined on the basis of data obtained with use of layers made of the most rigid materials (i.e. with the highest biaxial moduli) which are used for fabrication of the legs or/and the microbridge structure in the technological process described above. In this case it is not possible to completely abandon the manufacturing of mentioned in section a) test console structures 202a-c for optimization of the manufacturing technology, however the criterion formulated in the form of an inequality in this section enables to reduce significantly a parameter space wherein the optimization has to be performed. For example, if the thickness d ₂ is selected in a random way or its selection is based on reasons not related to the prevention of the unacceptable deformations of the legs and/or the microbridge structure, according to the criterion M]*d] > M ₂ *d ₂ thickness d] has a limiting criterion di > M ₂ *d ₂ /M ₁ enabling significant restriction of the range of parameters in which it is necessary to search for an optimal value of thickness di. On the other hand, within the framework of the criteria formulated in the section b), it is possible to make legs having the first and the second dielectric being significantly different. For example, if the first and second dielectric layers are made of the same material, according to criteria of section a) their thicknesses have to be the same, while according to criteria of section b) their thicknesses have to be different.

In a case when the first dielectric layer consists of several dielectric layers, in sections a) and b) the respective product is replaced with the sum of products of biaxial moduli and thicknesses of dielectric layers composing the first dielectric layer. For example, if the first dielectric layer consists of two dielectric layers A and B with respective biaxial moduli M _A and M _B and thicknesses d _A and ds the product M _\ *di in sections a) and b) is replaced with the sum of products M _A * IA and M _B * de- Similar procedure of replacement is made for the second dielectric layer in sections a) and b) if in turn it consists of several dielectric layers. Thus, the inequalities of criteria in sections a) and b) can be rewritten in the generalized mathematical form:

M

(=1 k=\

[1] < 0.1 , wherein N is a number of layers in the first dielectric layer, Mj ∑ , * 4 +∑ , * ^

/ = 1 k=\

and dj are biaxial moduli and thicknesses of layers in the first dielectric layer, M is a number of layers in the second dielectric layer, M _k and d _k are biaxial moduli and thicknesses of layers in the second dielectric layer, wherein thicknesses of the first and the second dielectric layer are not less than 90 nanometers, wherein each of the layers the first and the second dielectric layers is stress-free before the removal the sacrificial layer and has uniform mechanical properties through thickness, wherein the electrically conductive layer has also uniform mechanical properties through thickness and a uniform compressive stress through thickness, i.e. there is no gradient of intrinsic compressive stress in it;

N M

[2] < 0.05 , wherein N is a number of layers in the first dielectric layer,

∑J , * 4 +∑Ji/ ₄ * «/ _t

;=1 k=\

Mj and dj are biaxial moduli and thicknesses of layers in the first dielectric layer, M is a number of layers in the second dielectric layer, M _k and d _k are biaxial moduli and thicknesses of layers in the second dielectric layer, wherein a thickness of at least one of the first and the second dielectric layers is less than 90 nm, wherein each of the layers the first and the second dielectric layers is stress-free before the removal the sacrificial layer and has uniform mechanical properties through thickness, wherein the electrically conductive layer has also uniform mechanical properties through thickness and a uniform compressive stress through thickness, i.e. there is no gradient of intrinsic compressive stress in it;

M

[3] M _i * d _i > M _k * d _k , wherein N is a number of layers in the first dielectric layer, Mj and

;=1 k=\

dj are biaxial moduli and thicknesses of layers in the first dielectric layer, M is a number of layers in the second dielectric layer, M _k and d _k are biaxial moduli and thicknesses of layers in the second dielectric layer, wherein the electrically conductive layer has a through-the-thickness gradient of intrinsic stress, wherein an absolute value of an intrinsic compressive stress in a portion of the electrically conductive layer adjacent to the first dielectric layer is bigger than an absolute value of the compressive stress in a portion of the electrically conductive layer adjacent to the second dielectric layer, wherein each of the layers the first and the second dielectric layers is stress-free before the removal the sacrificial layer and has uniform mechanical properties through thickness;

N M

[4] * d _l < M _k * d _k , where N a number of layers in the first dielectric layer, Mj and d; are

(=1 k=\

biaxial moduli and thicknesses of layers in the first dielectric layer, M is a number of layers in the second dielectric layer, M _k and d _k are biaxial moduli and thicknesses of layers in the second dielectric layer, where there is a gradient of intrinsic stress in the electrically conductive layer, an absolute value of the intrinsic compressive stress in a portion of the electrically conductive layer adjacent to the first dielectric layer is less than an absolute value of the compressive stress in a portion of the conductive layer adjacent to the second dielectric layer, wherein each of the layers the first and the second dielectric layers is stress-free before the removal the sacrificial layer and has uniform mechanical properties through thickness.

The criteria similar to criteria of a) and b) can be developed and applied for the selection of topology of layer fragments which are used for the fabrication of the microbridge structure 117, the temperature responsive detector 108, its contacts 119, the absorber 109, and the electrical connection 112b between the contacts of the temperature responsive detector and fragments of the electrically conductive layer 112a used for the manufacturing of legs 116, since the fragments of the electrically conductive layer causing the tensile stresses in the dielectric layers of legs 116 can be also used for manufacturing of said electrical connection 112b and therefore they can cause mechanical stress in fragments of layers which are used for the fabrication of the microbridge structure 117, the temperature responsive detector 108, its contacts 119, and the absorber 109 (Figs. 13b and 14). However in this case, in difference from legs 116 where the single topological difference of fragments of the layers used for their fabrication is their thicknesses, it is necessary to consider not only thickness of fragments of the layers used for the fabrication of a microbridge structure 117, the temperature responsive detector 108, its contacts 119, the absorber 109, the electrical connections 112b providing electrical connection of contacts 119 with fragments of an electrically conductive layer 112a, but their geometry in the plane parallel to the substrate 100. In order to eliminate the direct influence of intrinsic stress in the electrically conductive layer 112 on planarity of the microbridge structure 117 itself, electrical connections 112b connecting the contacts 119 of temperature responsive detector 108 with fragments of the electrically conductive layer 112a used for fabrication of the legs 116 can be manufactured of fragments of the same layer of which fragments contacts of the detector 119 are made. In case when only one dielectric layer is adjacent to an electrically conductive layer in a leg (e.g. when the leg 116 consists only of the fragments of the layers 112 and 114) in order to avoid intolerable bending of the leg, the electrically conductive layer has to have a though 0 the 0 thickness gradient of compressive stress before the removal of the sacrificial layer, i.e. the absolute value of the compressive stress near one surface of the electrically conductive layer has to be less than in the vicinity of an opposite surface of the electrically conductive layer before the removal of the sacrificial layer. In addition the dielectric layer has to be adjacent to said opposite surface of the electrically conductive layer near which the absolute value of the compressive stress in the electrically conductive layer is higher before the removal of the sacrificial layer. The aforementioned techniques and criteria enable application of the optimization method of the pixel fabrication depicted on Fig. 22. This method enables execution of step-by-step iterative optimization of manufacturing that significantly simplifies the optimization process, since the optimization of a small number of parameters is made at each stage.

At the initial stage of optimization 400, a development of a library of deposition processes of dielectric and electrically conductive layers is performed. Each process for the deposition of dielectric layers enables growing the layers of dielectrics with uniform optical and mechanical properties. Besides, the library of processes includes processes of the deposition of the electrically conductive layers with the internal compressive stress. The deposition processes in the library enable deposition of the electrically conductive layers both with a gradient of intrinsic stress through thickness and without a gradient of intrinsic stress through thickness.

At the following stage of optimization 401, the selection of one or more manufacturing routes of a pixel is carried out. The manufacturing route of the pixel fabrication has to meet the following requirements: minimization of a number of layers and a number of structuring processes of the layers used in a course of pixel manufacturing, besides, the used layers shall provide the required sensitivity of pixel in the operating range of the device and also not to cause intolerable bends of legs and/or a microbridge structure. The simplification of a manufacturing route results in the need of using fragments of the same layers in the maximum number of functional elements of the pixel. For example, the first dielectric layer which can consist of one of or as well as of both layers 105 and 106, is used in the above described process of pixel manufacturing for the fabrication of the legs and the microbridge structure. Besides, the first dielectric layer functions as the absorber, since the aforementioned materials used for its manufacturing absorb IR radiation. Such multiple objective use of the first dielectric layer leads to contradictory requirements. On the one hand, the increase in thickness of this layer enables an increase in absorption of IR radiation and thus increase in pixel sensitivity, on the other hand, the increase in thickness of this layer leads to the increase in thickness of legs and their heat conductivity respectively that in turn reduces the heat insulation of the temperature responsive detector and, as a result, reduces the pixel sensitivity. Just the same dilemma forms a basis for a calculation of a thickness of the second dielectric layer 114 which is used for manufacturing of the same functional elements of the pixel as the first dielectric layer. The optimum solution addressing these contradictory requirements on the thicknesses of the first and the second dielectric layers can be found by means of using additional layers in the pixel which are used only for manufacturing of one of the functional elements. For example the layer 109 functions only as an absorber. This layer can consist of several layers providing the necessary absorption of IR radiation in the operating range of a device. It should be noted that it is only one of optimization variants of technological process and topology of layers. For example, the layer 105 can be structured in such a way that it functions only as an absorber.

Specifications of dielectric layers (mechanical properties and deposition processes) used for manufacturing of legs in different pixels / products are listed in the Table 1. Selection of processes, materials, and thickness of layers was carried out by a method of selection of layers with suitable mechanical and optical properties in the library created at stage 400, wherein one of inequalities [1] - [4] was used for the selection of thickness.

Table 1

The problem of process optimization of pixel manufacturing needs to be analyzed also from the point of view of the physical processes defining the quality of pixel namely: mechanical properties of the films used for manufacturing of pixel, absorption of IR radiation by an absorber of a pixel, conversion of the absorbed IR radiation into a measured signal of the temperature responsive detector, heat transfer from the absorber to the temperature responsive detector, and heat conductivity of the legs. Mechanical properties of layers and absorption of IR radiation in layers are well studied physical processes; moreover, it is easy to measure these properties. The integral properties of pixel related to IR absorption and mechanical properties of separate layers are easily predictable on the basis of properties of separate layers. For this reason, all the technological processes in the library of the processes are optimized in a way to produce stress- free films of dielectrics with uniform mechanical and optical properties. As a result, it is possible to select the combination of layers in the library of the processes providing not only the required absorption spectrum in the operating range of the device, but also, taking into account inequalities [1] - [4], not causing the intolerable deformation of the legs and/or the microbridge structure. It should be noted that in case of optimization of mechanical properties according to inequalities [3] or [4], the experimental optimization of thickness of the first or the second dielectric layer is still necessary. In contrast to the optical absorption and the mechanical properties of the layers, such parameters as a heat conductivity of legs and a heat transfer process from an absorber to the temperature responsive detector are too complex to be described using a simple model based on taking into account of one or other parameters of separate layers in a simple additive way. Processes of a heat transfer in pixel depend on a large number of parameters which often depend on a specific topology of a pixel. For example, the mean free path of phonons is comparable with typical dimensions of layers in the pixel; besides, a reduction in the heat transfer between different layers caused by Kapitsa resistance at layer interfaces is possible. In other words, the processes of the heat transfer in the pixel depend not only on processes in each layer but also on interaction of processes in these layers. Thus, the library comprising the processes of layer deposition and information on properties of the layers obtained in these processes simplifies the design process as much as possible by enabling to narrow down a search space of optimum solutions at a stage of planning of an optimization process.

During the following stage of optimization of technological manufacturing process 402, if necessary, the adjustment of thickness of layers with use of the test console structures consisting of several layers is carried out, wherein the test structures correspond to combinations of layers of which pixel elements shall be made. As a result of this adjustment of thickness, the absence of intolerable deformations in such pixel elements as a microbridge structure and legs is provided. The main objective of stage 402 is the manufacturing of pixels according to one or more manufacturing routes with due account for possible adjustment of thickness of layers.

During the following optimization stage of technological manufacturing process 404 measurements of a pixel sensitivity in the operating range are carried out. These measurements can include measurements of a spectral and a temporal response of the pixel or several pixels. In addition to the measurement of pixel characteristics, characteristics of separate components of the pixel (for example heat conductivity of legs) and/or characteristics of test structures can be measured. It is possible to use a separate reflector made according to the same process flow as the reflector 103b as a test structure enabling controlling a reflection coefficient of the reflector 103b. Further, a comparison of the obtained results with the given characteristics is carried out. Criteria of compliance of the obtained characteristic with the specified one can be formulated in the form of an interval in which the measured value of the characteristic has to be, or in the form of the threshold value defining a maximum (minimum) permitted measured value of the characteristic. In case of mismatch of the measured characteristics with the specified ones, change in process routes of pixel manufacturing is carried out for achieving the specified characteristics, i.e. as a matter of fact, it is a full or partial repetition of stage 401 with the subsequent repetition of stages 402 and 404. For example, if as a result of measurements it is found that, that the heat sink to a substrate from the temperature responsive detector is unacceptably high, a decision on the reduction of layer thickness used for manufacturing of both the legs and the absorber and on the use of additional layers for manufacturing of the pixel which are only used for fabrication of an absorber can be taken during repeated execution of the stage of optimization 401. In some cases, the repetition of the stage 401 can be restricted to the change of thickness of one or more layers. The measurement of current-voltage characteristic of pixel in the quasi-static mode can be used as very simple and effective method of measurement of pixel parameters. If such measurements are performed in vacuum with residual pressure less than 1 Pa, the Joule heat generated in the temperature responsive detector, will be mostly sunk through legs, thus the neglect of another heat sink mechanisms such as radiation has no essential impact on the measurement accuracy. In that case an equation of a heat balance is formulated as follows AT*G = / *R(AT), where the equation R(AT)= Ro (1+ *ΔΤ) describes dependence of electrical resistance of the temperature responsive detector R(AT) on temperature where Ro - a reference value of resistance at a reference temperature, AT - a temperature rise of the temperature responsive detector above the reference temperature, a - a temperature coefficient of resistance, G - a heat conductivity of legs, I - a current which flows through the temperature responsive detector and generates the Joule heat.

The right member of the equation of the heat balance describes the generation of Joule heat, and the right part of this equation describes a sink of the generated heat to the substrate through the legs. This equation can be written in another form: 1/R(AT)= 1/Ro -a I ² / G. Plotting of a linear graph where the squared current is plotted on a horizontal axis, and a reciprocal value of resistance is plotted on a vertical axis, enables determining a / G value on a basis of an inclination of a straight line of the linear graph. Thus, by making different pixels with identical temperature responsive detectors on one substrate (for example, an identical layer of vanadium oxide) it is possible to compare heat conductivity of pixels legs which were fabricated according to different manufacturing routes.

The sample of measurements of test pixels illustrating the effect of reduction in the heat conductivity of legs with use of the electrically conductive layers causing the tensile stresses in dielectric layers of legs is given in the Table 2. One technological process wherein pixels of two types were made corresponds to each row in the table. In order to manufacture one type of pixels, the processes of deposition of dielectrics and a conductive layer from library were used; another type of pixels was made in the same way except for one difference which was that the stress-free layers were used as conductive layers in legs. Since temperature responsive detectors were manufactured for both types of pixels in one process, the measurements described above enable comparing the heat conductivities of legs. The percentage decrease in the heat conductivity of legs of pixels when the electrically conductive layer having compressive stress is used instead of the stress free electrically conductive layer for the manufacturing of the legs, is shown in the last column of the Table 2. This reduction in the heat conductivity is not related to potential influence of intrinsic stress on heat conductivity of electrically conductive layers. Comparative measurements of heat conductivity of stress free conductive layers and conductive layers having intrinsic stress made of the same materials using "3co" method did not reveal differences in the heat conductivity exceeding 1,5 % ⁵ . Thus, the reduction in the heat conductivity of legs shown in the last column of the table is due to the utilization of the conductive layers with compressive intrinsic stress causing the tensile stresses in dielectric layers. Results similar to the results presented in the Table 2 were obtained for other nitrides of transition metals in combinations with different dielectric layers. Table 2

The preferable interval of intrinsic compressive stress in the electrically conductive layers causing the tensile stresses in dielectric layers of legs is the interval of -3 GPa - -1 GPA. More preferable interval of intrinsic compressive stress in the conductive layers causing the tensile stresses in dielectric layers of legs is the interval -2 GPa - -1.5 GPA.

In case of compliance of the measured pixel characteristics with the specified ones the process of optimization ends with execution of stage 405. The transition to stage 405 can take place as well if the optimization process has reached saturation and further repetition of stages 401, 402 and 404 does not appear to be justified, for example, when the relative change of one or more characteristics as a result of execution of the last iteration is less than corresponding threshold values (for example 5% or 10%). At stage 405 the preparation of documentation for mass production of devices with pixels according to one or more manufacturing routes developed in a process of the execution of aforementioned optimization stages is made. In fig. 23, the example of an array 500 for acquisition of an image in the IR range is depicted. Pixels of this array can be made using the same technology, as the manufacturing technology of single pixels as described above. This array includes electronic circuits enabling row-wise polling of pixels with the subsequent registration of the IR radiation power received by each pixel in the polled line. In this design, two posts correspond to each pixel. In fig. 24, more compact design of an array 510 is given. In this case, one post 513 is used for adjacent pixels. This array includes the electronic circuits enabling row-wise polling of pixels with the subsequent registration of the IR radiation power received by each pixel in the polled line. By analogy with the image of pixel on Fig. 14, for illustrative purposes of images on Figs. 23 and 24, the elements 103a, 103b and details of a surface relief pattern of microbridge structures are not illustrated. On Figs. 23 and 24, the contours of the structures made of a conductive layer 112, contours of contact structures 119, and contours of the structured layers 108 and 109 are depicted.

References

"Optical thin films and coatings," edited by A. Piegari H F. Flory, Woodhead Publishing, page 294, ISBN 9780857095947.

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

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

The MEMS handbook. Edited by M. Gad-el-Hak, CRC Press, pages 16-110, ISBN 0- 8493-0077-0.

Thermoelectrics handbook: Macro to Nano. Edited by D.M. Rowe, Taylor & Francis group, pages 23-11, ISBN 0-8493-2264-2.