Method for producing multispectral filter for electromagnetic radiation

Abstract

To provide a method for manufacturing a multispectram filter for electromagnetic radiation.SOLUTION: A filter includes: a first reflective layer 205; a second reflective layer 210; and a Fabry-Perot cavity dielectric layer 204A1, and includes the steps of: deposing a layer of a structuring material on a carrier substrate including at least two photoelectric transducers; three-dimensionally structuring the layer of the structuring material for obtaining at least two patterns of a different height, in which one of the patterns has a maximum reference height relative to the carrier substrate 200,; conformally deposing the first reflective layer 205; deposing a layer made of a dielectric material intended to form a dielectric pattern of a Fabry-Perot cavity; planarising the dielectric material, using a selective stop at a top part of the highest structuring material pattern; and deposing the same dielectric material for completing the formation of the Fabry-Perot cavity.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The technical field of the invention is that of imaging applications and of spectral filtering, in particular for making color filters for CMOS type image sensors, liquid crystal display devices or light emitting diodes. The invention can also be implemented in light emitting devices.

[0002] The present invention relates to a method for producing a multispectral filter for electromagnetic radiation. [Background technology]

[0003] Spectral or color filters are used to filter light by wavelength to provide information about the intensity of light at some wavelengths. Several color filters can be associated to form, for example, a Red-Green-Blue (RGB) filter, which gives information about the intensity of these three colors.

[0004] Metal / dielectric color filters made from Fabry-Perot cavities are especially known. These filters comprise one or more dielectric (or possibly semiconductor) cavities formed between two thin metal films with metal mirror function to form a Fabry-Perot cavity. Generally, the metal / dielectric stacks differ depending on their position on the optoelectronic component (for example an image sensor). The transmittance of the filter is set by adjusting the thickness of the cavity. So, in operation, a part of the incident light corresponding to the wavelength of the filter is transmitted through the filter in the form of a colored beam, while the remaining part of the incident light is reflected. Generally speaking, the thickness of the dielectric layer fixes the central wavelength transmitted, while the thickness of the metal layer makes it possible to set the transmission spectral width. Furthermore, the use of several Fabry-Perot cavities makes it possible to modify the spectral transmission profile of the filter. This type of filter is made using conventional semiconductor manufacturing techniques. So, to obtain a red-green-blue filter, it is advisable to form at least one dielectric cavity, the thickness of which should have three different values. This entails significant technological constraints, in particular at least one step of masking and then etching each dielectric cavity created.

[0005] It has also been shown that multi-spectral filter arrays may be used for imaging performed directly on the optical sensor array according to CMOS-compatible fabrication techniques (bottom-up approach).

[0006] More recently, document GB2574805 described the feasibility of creating an optical functional device following the creation of a Fabry-Perot cavity above a CMOS image sensor device using so-called grayscale lithography techniques. The method for manufacturing a multispectral filter in this document comprises the following steps: - depositing a first metal layer on a substrate containing a CMOS type photoelectric transducer; - structuring a layer of resist by grayscale lithography to create a number of resist patterns, each pattern forming a Fabry-Perot cavity; - depositing a second layer of metal over the pattern; Based on.

[0007] The latter solution also has some drawbacks. The first of these consists in carrying out greyscale lithography on a stack that already contains a metal layer: reflection phenomena at the metal-resist interface can then significantly affect the control over the resist thickness during lithography. Furthermore, greyscale lithography makes it possible to obtain resist pattern thicknesses of at least 40 or 50 nm: such values ​​are sufficient in the visible range, but it may be necessary to obtain lower thicknesses to filter out lower wavelengths in other spectral ranges. Such thicknesses below 40-50 nm are difficult to obtain using greyscale lithography as described in document GB2574805. [Prior art documents] [Patent documents]

[0008] [Patent Document 1] DE 2574805 A1 Summary of the Invention [Problem to be solved by the invention]

[0009] The present invention provides a solution to the problems discussed above by providing a method for manufacturing a multispectral filter for electromagnetic radiation that allows better control of the thickness of the Fabry-Perot cavity and allows obtaining Fabry-Perot cavity thicknesses of less than 40 nm. [Means for solving the problem]

[0010] Therefore, one object of the present invention is a method for manufacturing a multispectral filter for electromagnetic radiation, said filter comprising at least two color filters, each filter comprising a first reflective layer, a second reflective layer, a layer of Fabry-Perot cavity dielectric material between the first and second reflective layers, the thicknesses of the dielectric layers of the two color filters being different, each of the two filters facing a photoelectric converter, said method comprising the following steps: - depositing a layer of structured material on a carrier substrate comprising at least two photoelectric converters; - three-dimensionally structuring the layer of structured material to obtain at least two structured material patterns of different heights, at least one of the patterns having a maximum reference height relative to the carrier substrate, the height being measured perpendicular to the plane of the substrate, each of the patterns facing a photoelectric converter; - conformally depositing a first reflective layer on the at least two structured material patterns; - depositing a layer of dielectric material intended to form the dielectric pattern of a Fabry-Perot cavity, said layer of dielectric material covering all the patterns by having an upper surface, each point of which is located at a height higher than a maximum reference height; - planarizing the dielectric material by removal with selective stopping on top of the highest structured material pattern covered by the reflective layer; - depositing a layer of the same dielectric material to complete the formation of the Fabry-Perot cavity; - depositing a second reflective layer on the at least two Fabry-Perot cavities; Includes.

[0011] According to the invention, the Fabry-Perot cavity is made by a structuring step, for example grayscale lithography, carried out before the step of depositing the first reflective layer. This feature avoids the drawbacks linked to reflection phenomena at the metal-resist interface. Moreover, the thickness of the cavity with the lowest thickness is set directly by the thickness of the dielectric material made during the second deposition of the dielectric material directly above the first reflective layer. In fact, this deposition is preceded by a planarization step with a stop on top of the structuring material pattern covered by the reflective layer: it is therefore the thickness of the second layer of dielectric material that makes it possible to fix the lowest thickness of the Fabry-Perot cavity. By proceeding in this way, the method according to the invention does away with the thickness control limitations associated with grayscale lithography and makes it possible to achieve thicknesses much lower than known solutions, in particular below 40 nm, using conventional deposition techniques. Finally, it should be noted that the method according to the invention applies to any type of substrate, whether it is a substrate for integrating a CMOS photodetector or a transparent substrate, for example made of glass, on which the photodetector is arranged, for example, below the substrate. Thus, "carrier substrate" refers to both the substrate and the optoelectronic converter, whether the optoelectronic converter is integrated into the substrate or is above or below said substrate.

[0012] By "photoelectric converter" is meant a device that can act as a collector of light from a filter or as a transmitter of light to a filter. When the photoelectric converter is a collector, the converter can be, for example, a CMOS type photodiode. When the photoelectric converter is a transmitter, the converter can be an LED diode, a QLED type diode, or a laser diode, in which case the transmitter has a broader emission spectrum than the emission spectrum of the corresponding Fabry-Perot cavity.

[0013] In addition to the features previously discussed in the paragraph above, the method may have one or more of the following further features, taken individually or according to any technically possible combination:

[0014] - according to a first embodiment, the structuring material is a resist, said structuring step being carried out by a so-called greyscale lithography step on a layer of structured material followed by a step of crosslinking the resist; advantageously, before a step of conformally depositing a first reflective layer on the at least two structured material patterns, the structuring step is followed by a step of conformally depositing an encapsulation layer on the at least two structured material patterns, said first reflective layer covering the encapsulation layer;

[0015] According to a second embodiment, the structuring material is an organic or inorganic isolating material and the structuring step comprises the following sub-steps: o depositing a layer of organic or inorganic isolation material followed by the sub-step of depositing a layer of resist; a sub-step of performing so-called greyscale lithography on the layer of resist in order to three-dimensionally structure the layer of resist and obtain at least two resist patterns of different heights; a substep of transferring two resist patterns into a layer of organic or inorganic material in order to obtain said at least two structured material patterns of different heights; Includes.

[0016] the deposition of a layer made of a dielectric material intended to form the dielectric pattern of a Fabry-Perot cavity and the deposition of a layer made of the same dielectric material to complete the formation of the Fabry-Perot cavity are conformal depositions;

[0017] - three-dimensionally structuring the layer of structured material is a step of obtaining at least two structured material patterns of different heights with a void space between said at least two structured patterns.

[0018] According to a first alternative, the method according to the invention comprises, after the step of conformally depositing a first reflective layer and before depositing a layer made of dielectric material intended to form the dielectric pattern of the Fabry-Perot cavity, the following steps: o depositing a layer called isolation layer, said isolation layer covering all the structured material pattern by having an upper surface, each point of the isolation layer upper surface being located at a height relative to the carrier substrate that is higher than a maximum reference height; depositing a layer of resist above said isolating layer; o removing the resist in order to retain the resist only above the void spaces between the structured material patterns and to leave a clear separation layer above the structured material patterns; o removing material of the separation layer from clear zones of the separation layer; a step of removing the remaining resist in order to preserve walls made of said material of the separating layer between the structuring material patterns; Includes.

[0019] According to a second alternative, the method according to the invention comprises, after the step of conformally depositing a first reflective layer and before depositing a layer made of dielectric material intended to form the dielectric pattern of the Fabry-Perot cavity, the following steps: o depositing a layer of resist having an upper surface thereby covering all the structured material pattern, each point of the upper surface being located at a height relative to the carrier substrate that is higher than a maximum reference height; o removing the resist between the structured material patterns in order to retain the resist only above the structured material patterns; depositing a layer, called separation layer, which covers all the structured material patterns covered by the resist and fills the spaces between said patterns by having an upper surface, each point of which is located at a height relative to the carrier substrate that is higher than a maximum reference height; removing the material of the separation layer above the structuring material patterns and the remaining resist in order to retain walls made of said material of the separation layer between the structuring material patterns; Includes.

[0020] - said material of the separation layer may be a sacrificial material; said method comprising the steps of: depositing a layer of resist on the second reflective layer; o removing the resist from the zones above the spaces between the structured material patterns; o removing the material of the second reflective layer, the dielectric material intended to form the dielectric patterns of the Fabry-Perot cavities, and the sacrificial material from zones above the spaces between the structured material patterns, in order to form void walls between the structured material patterns and between the overhanging Fabry-Perot cavities; Remove remaining resist Includes.

[0021] - said material of the separation layer may be a low or high optical index dielectric material; by high or low optical index material it is meant that the material has an optical index strictly lower or strictly higher than the index of the resist or the index of the material of the sublayer and of the dielectric material intended to form the dielectric pattern of the Fabry-Perot cavity; said method comprises the following steps: depositing a layer of resist on the second reflective layer; o removing the resist from the zones above the spaces between the structured material patterns; o removing the material of the second reflective layer and the dielectric material from the zones not protected by the remaining resist in order to reach the low or high optical index dielectric material; o depositing a layer of low or high optical index dielectric material over the remaining resist to fill the voids between the overhanging Fabry-Perot cavities, complementing the dielectric material already present between the structured material patterns; Removing the resist and the low or high optical index dielectric material above the resist zones while leaving the low or high optical index material between the structured material patterns and between the overhanging Fabry-Perot cavities to form a dielectric isolation wall. Includes.

[0022] The invention and its various applications will be better understood by reading the following description and examining the accompanying figures.

[0023] The figures are presented to illustrate the scope of the invention and in no way to limit it. [Brief description of the drawings]

[0024] [Figure 1] 1 illustrates the different steps of a method according to a first embodiment of the invention; [Diagram 2] 5A-5D illustrate different steps of a method according to a second embodiment of the invention. [Diagram 3] 2A-2D show different steps of two alternative methods for the method according to the invention for ensuring optical isolation between pixels as well as between Fabry-Perot cavities. [Figure 4] 1A-1D show different steps of two alternative methods for the method according to the invention for ensuring optical isolation between pixels as well as between Fabry-Perot cavities. [Diagram 5] 1A-1D show different steps of two alternative methods for the method according to the invention for ensuring optical isolation between pixels as well as between Fabry-Perot cavities. [Figure 6] 1A-1D show different steps of two alternative methods for the method according to the invention for ensuring optical isolation between pixels as well as between Fabry-Perot cavities. [Figure 7] 1A-1D show different steps of two alternative methods for the method according to the invention for ensuring optical isolation between pixels as well as between Fabry-Perot cavities. [Figure 8] 1A-1D show different steps of two alternative methods for the method according to the invention for ensuring optical isolation between pixels as well as between Fabry-Perot cavities. [Figure 9] 1A-1D show different steps of two alternative methods for the method according to the invention for ensuring optical isolation between pixels as well as between Fabry-Perot cavities. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0025] The figures are set forth to illustrate the scope of the invention and are not to be construed as limiting the scope of the invention in any way.

[0026] Generally speaking, and as is well known to those skilled in the art, it should be borne in mind that a metal / dielectric type color filter made from a Fabry-Perot cavity is obtained by dimensioning the thickness of a dielectric layer formed between two metal layers. If several colors are to be filtered on the same component, it is necessary to be able to obtain a variable thickness of dielectric on the same component.

[0027] This dimensioning is performed using, for example, electromagnetic calculation programs such as the Abeles matrix transfer formalism, or diffraction calculations on pixels whose size is close to the wavelength, such as the Fourier Expansion Modal Method or Rigorous Coupled Wave Analysis (RCWA) formalisms.

[0028] These calculation programs allow the optimum parameters for the dielectric metal stack to be determined for each pixel. The calculations take into account, among other things, the thicknesses of the metal and dielectric layers, as well as their indices, the spectral and angular distribution of the incident light. For example, in the case of a Fabry-Perot filter, the central wavelength of the filter is determined by the following formula:

number

[0029] Once the cavity order has been chosen, the angle of attack is known, and the index and phase shift are known, all that remains is to determine the approximate thickness h so that the cavity is centered at a particular wavelength. Once the filter function has been calculated for each filter and each wavelength, the dielectric thickness h is then adjusted according to the required performance (searching for good signal-to-noise ratio, maximum transmission, etc.).

[0030] An alternative, more empirical method is to calculate the response of the stack for several thicknesses h and choose h such that the filter resonance peak (λres) is positioned according to specification.

[0031] FIG. 1 shows the different steps of a method 100 according to a first embodiment of the invention in which no isolation is provided between the Fabry-Perot cavities nor between the optoelectronic converters.

[0032] The first step 101 of the method 100 consists in depositing a layer 201 of resist on a substrate 200, called the carrier substrate, and then three-dimensionally structuring said layer 201 of resist. The substrate 200 may for example be a Si substrate, a Silicon On Isolator (SOI) substrate or even a glass or sapphire substrate.

[0033] The carrier substrate 200 includes a plurality (at least two) photoelectric converters 202Ai (here three photoelectric converters 202A1, 202A2 and 202A3 are shown). The converters 202Ai are shown here diagrammatically below the substrate 200, it being understood that the converters 202Ai may be formed inside or above the latter. According to the invention, the photoelectric converters 202Ai are considered as integral parts of the carrier substrate 200. The photoelectric converters can operate as collectors of light from the filter or as transmitters of light to the filter. When the photoelectric converter is a collector, the converter can be, for example, a CMOS type photodiode. When the photoelectric converter is a transmitter, the transmitter can be an LED diode, a QLED type diode or a laser diode, in which case the transmitter has a broader emission spectrum than the emission spectrum of the corresponding Fabry-Perot cavity.

[0034] The 3D structuring of the layer of resist 201 is carried out by a lithography step. This lithography can preferably be grayscale lithography according to electronic or optical terminology. Other lithography techniques such as two-photon lithography or nanoimprint lithography can be used as well to make the resist structure 201. The structure 201 comprises a number of 3D patterns 203Ai (here three patterns 203A1, 203A2 and 203A3). It is possible to freely fix the dimensions of each pattern in three directions of the space Oxyz (Oxy is the plane of the drawing, the axis Oy is along a direction perpendicular to the plane of the substrate 200 and the axis Oz is along a direction perpendicular to the plane of the drawing).

[0035] According to the invention, it is advisable to have at least two resist patterns 203Ai of different heights (here, three patterns each having a different height). Among all these patterns 203Ai, one of them, in this case the pattern 203A1, has a maximum height Hmax, called the reference height, the height being measured perpendicular to the plane of the substrate 200 along the axis Ox. More generally, the height of the pattern 203Ai is determined by the following equation: resist-i Therefore, in Fig. 1, Hmax is expressed as h resist-i Each pattern 203Ai faces a corresponding transducer 202Ai along the axis Oy.

[0036] The method 100 then includes a step 102 of crosslinking the structured layer of resist 201 to densify the layer of resist 201 by stabilizing the grayscale resist using thermal or ultraviolet methods.

[0037] Advantageously, the method 100 according to the invention comprises a third step 103 consisting in conformal deposition on the resist pattern 203Ai (above and on the flanks of the pattern 203Ai) of a thin (a few tens of nm) encapsulation layer 204. The material of the encapsulation layer 204 may be, for example, an oxide or nitride of the Al2O3 or SiO2 type, the deposition being preferably carried out by a low-temperature deposition technique such as Atomic Layer Deposition (ALD).

[0038] The method 100 according to the invention then comprises a step 104 of depositing a reflective layer 205, for example with a thickness between 20 nm and 100 nm, covering the encapsulation layer 204 above the resist patterns 203Ai. This reflective layer 205, typically a metal layer, forms the first reflective layer of the resulting Fabry-Perot cavity type color filter. The reflective layer 205 is conformally deposited by Chemical Vapour Deposition (CVD or Physical Vapour Deposition (PVD). The layer 205 is deposited successively on all the resist patterns 203Ai covered by the encapsulation layer, including on the flanks of the resist patterns 203Ai. The conformal deposition allows the thickness of the reflective layer 205 to be constant, at least on the top of the patterns, but preferably also on the flanks of the patterns 203Ai.

[0039] The following step 105 consists in depositing a layer 206 made of a dielectric material intended to form the dielectric pattern of the Fabry-Perot cavity. The layer 206 of dielectric material covers all the resist patterns 203Ai covered by the sealing layer 204 and the reflective layer 205. The layer 206 has an upper surface 207 (not necessarily planar) and each point of the upper surface 207 is located at a height relative to the carrier substrate 200 higher than a maximum reference height Hmax. The material of the layer 206 is preferably, but not exclusively, a material that is transparent in the visible range, such as an organic material of polymeric or inorganic type (oxides, silicon nitride, alumina, etc.). The deposition is preferably a conformal deposition, for example carried out by Physical Vapor Deposition (PVD) or Chemical Vapour Deposition (CVD) or Low Pressure Chemical Vapour Deposition (LPCVD) or Plasma Enhanced Chemical Vapour Deposition (PECVD) techniques. It should be noted that the invention is not limited to the visible range and that other materials transparent to other wavelengths, for example in the infrared, can be used (eg using silicon).

[0040] The method 100 then comprises a step 106 of planarizing the layer 206 made of a dielectric material intended to form the dielectric pattern of the Fabry-Perot cavity, to form a planarized layer 208 at its surface by removing the dielectric material from the layer 206, the planarizing step may be carried out by an etch-back type etching step and / or a Chemical Mechanical Polishing (CMP) step. The planarization is performed with a stop on the reflective layer 205 located at its highest level (i.e. at the reference height of the pattern 203A1).

[0041] The method 100 according to the invention then comprises a step 107 of conformally depositing a layer 209 of the same dielectric material intended to form the dielectric pattern of the Fabry-Perot cavities as in step 105. This step is carried out, for example, by physical vapor deposition (PVD) or chemical vapor deposition (CVD) or low pressure chemical vapor deposition (LPCVD) or plasma enhanced chemical vapor deposition (PECVD) techniques. The height of the layer 209 is indicated as hc, the height being measured perpendicular to the plane of the substrate. After this step, a number of Fabry-Perot cavity dielectric patterns 204Ai (here three patterns 204A1, 204A2, 204A3) are obtained.

[0042] For each resist pattern 204Ai, the height of the pattern is denoted as hdiel_i, the height being measured perpendicular to the plane of the substrate. The height hdiel_i of the pattern 204i can be determined (based on the specifications for the desired filtering) by the methods described above. According to the method 100 according to the present invention, the height hdiel_i is calculated according to the following formula: hdiel_i=Hmax-h reist-i Set by +hc.

[0043] So, for pattern 204A1, hdiel_1 now directly equals the height hc of layer 209.

[0044] More generally, for each pattern 204Ai, the height is technically determined by the height difference between the height Hmax of the highest resist pattern 203A1 and the height of the resist pattern 203Ai to which the height of the second dielectric layer 209 is added. It can therefore be observed that the height of the Fabry-Perot cavity dielectric pattern 204Ai is not related to the thickness of the layer 201 of resist (unlike current state-of-the-art solutions that make it difficult to obtain a thickness of about 10 nm), but to the height difference between the patterns. Such a configuration makes it possible to achieve much smaller heights. This is especially true for the Fabry-Perot cavity dielectric pattern 204A1 formed directly above the highest resist pattern 203A1, which depends only on the height of the second dielectric layer 209, which is very easy to control and can reach values ​​of the order of 10 nm.

[0045] Step 108 of the method 100 then consists in depositing a second reflective layer 210, for example with a thickness between 20 nm and 50 nm, covering the second dielectric layer 209 above the Fabry-Perot cavity pattern 204Ai. This planar reflective layer 210 forms the second reflective layer of the Fabry-Perot cavity type color filter. The reflective layer 210 is conformally deposited by chemical vapor deposition (CVD) or physical vapor deposition (PVD). A plurality (here, three) of Fabry-Perot cavity type color filters Fi (i.e., F1, F2, and F3) are thus obtained, each formed by the first reflective layer 205, the pattern 204Ai, and the second reflective layer 210.

[0046] Optionally, a layer 211 may be deposited on the second reflective layer 210 according to CMOS technology.

[0047] FIG. 2 shows different steps of a method 300, an alternative to method 100 of FIG. 1, in which no isolation is provided between the Fabry-Perot cavities or between the optoelectronic converters.

[0048] The first step 301 of the method 300 consists in depositing a layer of material, called sublayer 401, such as polymeric organic material or inorganic type (oxide, silicon nitride, alumina, etc.), on a substrate 400, called carrier substrate. The substrate 400 is of the same type as the substrate 200 of FIG. 1 and comprises a plurality (at least two) photoelectric converters 402Ai (here three photoelectric converters 402A1, 402A2 and 402A3 are shown). A layer of resist 412 is then deposited on the sublayer 401. The layer of resist 412 is then three-dimensionally structured. The 3D structuring of the layer of resist 412 is carried out by a lithography step. This lithography can preferably be grayscale lithography in electronic or optical terms. Other lithography techniques, such as two-photon lithography or nanoimprint lithography, may also be used to make the resist structures 412. The structure 411 includes a number of 3D patterns 412Ai (here three patterns 412A1, 412A2 and 412A3), each of which can be freely fixed in its dimensions in three directions in the space Oxyz (Oxy is the plane of the drawing, the axis Oy is along a direction perpendicular to the plane of the substrate 400, and the axis Oz is along a direction perpendicular to the plane of the drawing).

[0049] According to the invention, it is advisable to have at least two resist patterns 412Ai of different heights (here, three patterns each have a different height). Among all these patterns 412Ai, one of them, in this case the pattern 412A1, has a maximum height Hmax, called the reference height, the height being measured perpendicular to the plane of the substrate 400 along the axis Ox. More generally, the height of the pattern 412Ai is determined by the following equation: resist-i Therefore, in FIG. 2, Hmax is expressed as h resist-1 Each pattern 412Ai faces a corresponding transducer 202Ai along the axis Oy.

[0050] The method 300 then comprises a step 302 of transferring the patterns 412Ai from the layer of structured resist 412 into the sublayer 401: in other words, at the end of this step 302, the layer of structured resist 412 is removed and completely transferred into the sublayer 401, so that the sublayer 401 is itself three-dimensionally structured with a number of patterns 403Ai (here the three patterns 403A1, 403A2 and 403A3) of dimensions similar to those of the patterns 412Ai. This transfer step can be carried out by techniques known to the person skilled in the art, such as dry or wet etching.

[0051] Step 302 is followed by a step 303 of depositing a reflective layer 405, for example with a thickness between 20 nm and 50 nm, which covers the patterns 403Ai of the structured sublayer 401. This reflective layer 505 forms the first reflective layer of the resulting Fabry-Perot cavity type color filter. The reflective layer 505 is conformally deposited by chemical vapor deposition (CVD) or physical vapor deposition (PVD). The layer 405 is deposited successively on all the patterns 403Ai, including on the flanks of the patterns 403Ai. The conformal deposition allows the thickness of the reflective layer 405 to be constant, at least on the top of the patterns, but also preferably on the flanks of the patterns 403Ai. It should be noted that, unlike the first embodiment of FIG. 1, it is here quite possible to dispense with the deposition of an encapsulation layer. As in the first exemplary embodiment of FIG. 1, a structuring step, for example using grayscale lithography, is carried out before the step of depositing the first reflective metal layer.

[0052] The advantage of having the pattern made of another organic or inorganic material rather than resist is that resist is often a polymer resist, which can have characteristics that degrade over time, resulting in a change in color or loss of light transmission. The use of an organic material offers better stability over time, and the use of an organic material prevents the color from changing over time.

[0053] Steps 304 to 307 of the method 300 according to the invention are identical to steps 105 to 108 of the method 100 according to the invention.

[0054] Therefore, the next step 304 consists in depositing a layer 406 made of a dielectric material intended to form the dielectric pattern of the Fabry-Perot cavity. The layer 406 of dielectric material covers all the patterns 403Ai covered by the reflective layer 405. The layer 406 has an upper surface 407 (not necessarily planar) and each point of the upper surface 407 is located at a height relative to the carrier substrate 400 higher than a maximum reference height Hmax. The material of the layer 406 is preferably, but not exclusively, a material transparent in the visible range, such as an organic material of polymeric or inorganic type (oxides, silicon nitride, alumina, etc.). The deposition is preferably a conformal deposition, for example carried out by physical vapor deposition (PVD) or chemical vapor deposition (CVD) or low pressure chemical vapor deposition (LPCVD) or plasma enhanced chemical vapor deposition (PECVD) techniques.

[0055] The method 400 then comprises a step 305 of planarizing the layer 406 made of dielectric material intended to form the dielectric pattern of the Fabry-Perot cavity, in order to form a layer 408 planarized at its surface by removal of the dielectric material from the layer 406. The planarization is performed with a stop on the reflective layer 405 located at its highest level (i.e. at the reference height of the pattern 403A1).

[0056] The method 300 according to the invention then comprises a step 306 of conformally depositing a layer 409 of the same dielectric material intended to form the dielectric pattern of the Fabry-Perot cavities as in step 304. This step is carried out, for example, by PVD, CVD, LPCVD or PECVD deposition techniques. The height of the layer 409 is indicated as hc, the height being measured perpendicular to the plane of the substrate. After this step, a number of Fabry-Perot cavity dielectric patterns 404Ai (here three patterns 404A1, 404A2, 404A3) are obtained.

[0057] For each resist pattern 404Ai, its height hdiel_i is determined the same as in the method 100 of FIG.

[0058] Step 307 of the method 300 then consists in depositing a second reflective layer 410, for example with a thickness between 20 nm and 50 nm, covering the second dielectric layer 409 above the Fabry-Perot cavity pattern 404Ai. This planar reflective layer 410 forms the second reflective layer of the Fabry-Perot cavity type color filter. The reflective layer 410 is conformally deposited by CVD or PVD deposition. A number of (here three) Fabry-Perot cavity type color filters Fi (i.e. F1, F2, and F3) are thus obtained, each formed by the first reflective layer 405, the pattern 404Ai, and the second reflective layer 410.

[0059] Optionally, a CMOS layer 411 may be deposited on the second reflective layer 410 .

[0060] The embodiment shown in figures 1 and 2 illustrates a simple and efficient fabrication of Fabry-Perot resonant cavities. However, in the case of discretely present photoelectric converters, i.e. pixelated converters, where each pixel is formed by a photoelectric converter (e.g. photodetector) and a corresponding Fabry-Perot cavity, it may be particularly interesting to perform optical isolation between each pixel. In fact, when the dimensions of the pixels, and thus the color filters, are reduced, typically to the order of the wavelength of light, the spatial overlap of the light contribution from one Fabry-Perot cavity to another can reduce the efficiency of the overall device. This "crosstalk" phenomenon has already been observed in photoelectric converters (i.e. CMOS sensors). In recent years, pixel dimensions have been continuously reduced in order to provide high resolution sensors, especially for smartphones. This reduction in pixel dimensions has two very different consequences: - The total amount of incident light hitting each pixel is reduced, which means that fewer and fewer photons reach the light-sensitive component. Collecting each photon within the component and directing it towards the sensor is all the more important to ensure the efficiency of the sensor. - As the pixel size approaches the length of the light being collected, diffraction comes into play, increasing the optical and spatial losses in the device.

[0061] In addition to the problems associated with reducing pixel dimensions, the angle of incidence at which light penetrates the Fabry-Perot cavity is also a factor that should not be overlooked. In fact, when an oblique ray of light reaches the surface of the cavity, part of the light passes through the physical boundary of the filter and ends up in the neighboring cavity. This undesirable parasitic contribution distorts the light collection at the scale of the device and is another major source of "crosstalk" from one pixel to another.

[0062] 3 shows a carrier substrate 500 in a top view and in a cross section across the photoelectric converters along a plane P perpendicular to the plane of the substrate 500. The carrier substrate 500 comprises a plurality of pixelated photoelectric converters 502Ai (three of which 502A1, 502A2 and 502A3 are shown in cross section) arranged in the form of a matrix (here a 3×3 matrix) in the substrate 500. The substrate 500 is covered by a barrier layer 501.

[0063] Different alternatives to the method according to the invention will now be described, which allow colour filters to be made, each associated with a converter located above or in the substrate 500, with optical isolation that allows crosstalk to be reduced not only between pixelated photoelectric converters, but also between Fabry-Perot cavities.

[0064] The first step 601, shown in Fig. 4 and common to these alternatives, consists in making resist patterns 503Ai (here three pillars 503A1, 503A2 and 503A3), hereafter called pillars, of different heights above the barrier layer 501, it being understood that these pillars may be made directly above the substrate 500. To that end, a layer of resist is first deposited and then three-dimensionally structured by grayscale lithography using masks M1, M2 and M3 of different densities. The exposure of the resist is therefore different and after development three pillars 503Ai of different heights are obtained and the resist between the pillars is completely removed so that a void space is present between each pillar 503Ai.

[0065] 1 and 2, it is sensible to have at least two resist pillars 503Ai of different heights (here, each of the three pillars is of a different height). Among all these pillars 503Ai, one of them, in this case pillar 503A1, has a maximum so-called reference height Hmax, the height being measured perpendicular to the plane of the substrate 500 along the axis Ox. More generally, the height of the pillar 503Ai will be denoted as hresist-i. So, in the figures, Hmax is expressed as h resist-1 Each pillar 503Ai faces a corresponding transducer 502Ai along an axis Oy. As in FIG. 1, the resist pillars 503i are cross-linked to densify them.

[0066] It should be noted that although pillars 503Ai in FIG. 4 are made of resist, they could also be made of organic or inorganic materials according to the sub-layers described with reference to FIG. 2 by transferring a resist pattern into the sub-layer.

[0067] The following step 602 shown in FIG. 5 is likewise common to these alternatives to the method according to the invention and corresponds to steps 103-104 and 303 in FIGS. 1 and 2. In practice, the aim is to deposit an encapsulation layer and then a reflective layer on the pillars 503Ai when they are made of resist, or to simply deposit a reflective layer on the pillars 503Ai when they are made of organic or inorganic material. For simplicity, the encapsulation layer and the reflective layer (in the case of resist pillars) or the reflective layer (in the case of sub-layer pillars) are represented by a single layer 505 in FIG. 5, referred to hereafter by the general term "reflective layer". The technique for making this layer is similar to the one already described with reference to FIGS. 1 and 2. The thickness of the reflective layer varies between about 10 nanometers and several hundred nanometers, depending on the metal chosen and the design choices made.

[0068] FIG. 6 illustrates steps 603 to 607 of a first alternative embodiment for achieving optical isolation between adjacent pixels (ie, between photoelectric converters 502Ai and between pillars 503Ai of resist or undercoating material).

[0069] Step 603 consists in depositing a layer called isolation layer 506. This substantially planar isolation layer covers all the pillars 503Ai covered by the reflective layer 505 and has an upper surface 507, each point of which is located at a height relative to the carrier substrate higher than a maximum reference height Hmax. As will be seen later, the isolation layer 506 may be made of a sacrificial material (i.e. intended to disappear) or, conversely, of a high or low optical index material intended to provide isolation between the photoelectric converters 502Ai. By high or low optical index material is meant a material that has an optical index strictly lower or strictly higher than the index of the resist or of the material of the sublayer and than the index of the dielectric material intended to form the dielectric pattern of the Fabry-Perot cavity.

[0070] Materials such as SiN may be used for the high optical index material and polymers containing fluorine groups may be used for the low optical index material. It will be seen later that when the material used is a sacrificial material, the purpose is to create an air gap to achieve isolation.

[0071] Step 604 consists in depositing a layer 508 of a photosensitive or electron sensitive resist above the isolation layer 506 .

[0072] Step 605 then consists in making gaps 509 aligned with the boundaries of the pillars 503Ai. For that purpose, a layer 508 of photosensitive or electron sensitive resist is exposed using a lithography tool in order to define a regular gap grid on its surface, after which the resist is developed. The lithography used here is conventional lithography. The gaps 509 are created after removal of the resist, leaving lines 510 of resist above the layer of isolating material 506 in the zones located between each pillar 503Ai.

[0073] According to step 606, in the open spaces in the resist formed by the gaps 509, the isolation material is etched with a stop on the reflective layer 505. At the end of this step 606, boundaries or walls 511 of isolation material remain on each edge of each pillar 503Ai, with lines 510 of resist extending from each boundary 511.

[0074] The next step 607 then consists in removing (by etching or stripping type techniques) the lines of resist 510 to leave only boundaries or walls 511 of isolating material on either side of the pillars 503Ai that fill the void spaces between the pillars 503Ai. The walls 511 can act as optical separators between adjacent pixels. The walls 511 can also act as containment walls for the pillars of resist or sub-layer material. The walls will likewise prevent deformation of the pillars of resist or sub-layer material during the following planarization step.

[0075] FIG. 7 illustrates steps 603' to 606' of a second alternative embodiment of optical isolation between adjacent pixels (ie, between photoelectric converters 502Ai and between patterns of resist or sub-layer material 503Ai).

[0076] Step 603' consists in depositing a layer 506' of photosensitive or electron sensitive resist, which layer 506' covers all the pillars 503Ai covered by the reflective layer 505 and has an upper surface 507', each point of which is located at a height relative to the carrier substrate higher than a maximum reference height Hmax.

[0077] Step 604' then consists in making gaps 509' creating void spaces between each pillar 503Ai. For that purpose a layer 507' of photosensitive or electron sensitive resist is exposed using a lithography tool to define a regular gap grid on its surface, after which the resist is developed. The lithography used here is conventional lithography. The gaps 509' are created after removal of the resist, leaving a resist 510' above each pillar 503Ai.

[0078] According to the next step 605', a layer called isolation layer 511' is deposited. This isolation layer 511' covers all the resist 510' and fills the gaps 509' to a height strictly greater than Hmax. As in the case of figure 6, the isolation layer 511' may be made of a sacrificial material (i.e. intended to disappear) or, conversely, of a high or low optical index material intended to ensure isolation between the photoelectric converters 502Ai.

[0079] Step 606' then consists in removing, for example by etching or stripping type techniques, the resist 510' as well as the material of the isolation layer 511' located above the resist 510', in order to leave, on either side of the pillars 503Ai, only walls 511 of isolation material filling the void spaces between said pillars 503Ai. It may be noted that the structure obtained after steps 603' to 606' of Figure 7 is identical to the structure obtained after steps 603 to 607 of Figure 6.

[0080] Based on the structure obtained at the end of step 607 of Fig. 6 or step 606' of Fig. 7, Fig. 8 shows steps 608 to 612 that enable a Fabry-Perot cavity to be made. It should be noted that these steps 609 to 612 are substantially the same as steps 105 to 108 and 304 to 307 of Figs. 1 and 2.

[0081] Step 608 therefore consists in using the structure obtained at the end of step 607 in FIG. 6 or step 606' in FIG.

[0082] The next step 609 consists in depositing a layer 706 made of a dielectric material intended to form the dielectric pattern of the Fabry-Perot cavity. The layer 706 of dielectric material covers all the pillars 503Ai covered by the reflective layer 505. The layer 706 has an upper surface 707 (not necessarily planar) and each point of the upper surface 707 is located at a height relative to the carrier substrate 500 higher than a maximum reference height Hmax. The material of the layer 706 is preferably, but not exclusively, a material that is transparent in the visible range, such as an organic material of polymeric or inorganic type (oxides, silicon nitride, alumina, etc.). The deposition is preferably a conformal deposition, for example carried out by PVD, CVD, LPCVD or PECVD deposition techniques.

[0083] The method then comprises a step 610 of planarizing the layer 706 made of dielectric material intended to form the dielectric pattern of the Fabry-Perot cavity, to form a layer 708 planarized at its surface by removal of the dielectric material from the layer 706. The planarization is performed with a stop on the reflective layer 505 located at its highest level (i.e. at the reference height of the pillars 503A1).

[0084] The method according to the invention then comprises a step 611 of conformally depositing a layer 709 of the same dielectric material intended to form the dielectric pattern of the Fabry-Perot cavities as in step 609. The height of the layer 709 is indicated as hc, the height being measured perpendicular to the plane of the substrate. After this step, a number of Fabry-Perot cavity dielectric patterns 704Ai (here three patterns 704A1, 704A2, 704A3) are obtained.

[0085] For each dielectric pattern 704Ai, the height of the pattern is denoted as hdiel_i, the height being measured perpendicular to the plane of the substrate. The height hdiel_i of the pattern 704i can be determined according to the method described above. According to the method according to the invention, the height hdiel_i is calculated according to the following formula: hdiel_i=Hmax-h resist-i Set by +hc.

[0086] So, for pattern 704A1, hdiel_1 now directly equals the height hc of layer 709.

[0087] More generally, for each pattern 704Ai, the height is technically determined by the height difference between the height Hmax of the highest resist pattern 503A1 and the height of the resist pattern 503Ai to which the height of the second dielectric layer 709 is added.

[0088] Step 612 of the method according to the invention then consists in depositing a second reflective layer 710, for example with a thickness between 20 nm and 50 nm, covering the second dielectric layer 709 above the Fabry-Perot cavity patterns 704Ai. This planar reflective layer 710 forms the second reflective layer of the Fabry-Perot cavity type color filter. The reflective layer 710 is conformally deposited. A plurality (here three) of Fabry-Perot cavity type color filters Fi (i.e. F1, F2 and F3) are thus obtained, each formed by the first reflective layer 505, the patterns 704Ai and the second reflective layer 710. It should be noted that at this stage the photoelectric converters 502Ai as well as the pillars 503Ai are sufficiently separated by the walls 511, but the dielectric patterns 704Ai and therefore the Fabry-Perot cavities are not completely separated.

[0089] FIG. 9 shows two alternative embodiments of isolation aimed at reducing crosstalk not only between the photoelectric converters 502Ai and the pillars 503Ai, but also between the dielectric patterns 704Ai and thus the Fabry-Perot cavities.

[0090] Based on the structure obtained at the end of step 612 of FIG. 8, step 613 consists in depositing a layer of resist 711 on the second reflective layer 711 .

[0091] Step 614 then consists in making aligned gaps 712 in the walls 511. For that purpose, a layer 711 of photosensitive or electron sensitive resist is exposed using a lithography tool in order to define a regular gap grid on its surface after the resist is developed. The lithography used here is conventional lithography. The gaps 712 are created after removal of the resist, leaving lines 713 of resist above the second reflective layer 710 in the zones located above each pillar 503Ai.

[0092] According to step 615 , in the open spaces in the resist formed by the gaps 712 , the metal of the second reflective layer 710 as well as the dielectric material used for the Fabry-Perot cavity are etched to reach the walls 511 .

[0093] The method can then take two different routes depending on whether the material of the wall 511 is a sacrificial material or a high or low optical index material intended to ensure optical isolation.

[0094] According to a first alternative, in which the material of the walls 511 is a sacrificial material, the method includes a step 616 of removing the walls 511 to form an air gap 714 that completely separates the photoelectric converters 502Ai, the pillars 503Ai and the dielectric pattern 704Ai from each other and thus the Fabry-Perot cavity.

[0095] Finally, step 617 consists in removing the remaining resist 713 by stripping or etching.

[0096] According to a second alternative, in which the material of the wall 511 is a high or low optical index material, the method includes a step 616' of depositing a layer 714' of said high or low optical reflectivity material to form a wall made of this material that completely separates the photoelectric converter 502Ai, the pillars 503Ai and the dielectric pattern 704Ai and thus the Fabry-Perot cavity.

[0097] Finally, step 617' consists in removing, by stripping or etching, the remaining resist 713 as well as the low or high optical index material located above the resist.

[0098] The air gap 714 and the separation wall 714' together ensure complete isolation of the entire Fabry-Perot cavity as well as its associated opto-electrical transducer. Furthermore, each pillar of resist (or made of the material of the sub-layer) on which the first reflective layer is deposited acts as a structure to guide the light waves transmitted from the filter to the transducer or vice versa: such a configuration further limits the risk of crosstalk between two adjacent filters and between two adjacent transducers and increases the opto-electrical performance of the filter. [Explanation of symbols]

[0099] 200, 400, 500 Carrier Board 201, 412 Resist layer 202A1, 202A2, 202A3, 402A1, 402A2, 402A3, 502A1, 502A2, 502A3 Photoelectric converter 203A1, 203A2, 203A3, 403A1, 403A2, 403A3, 412A1, 412A2, 412A3 3D (resist) pattern 204A1, 204A2, 204A3, 404A1, 404A2, 404A3, 704A1, 704A2, 704A3 Farley-Perot Dielectric Pattern 204 Sealing layer 205, 405, 505 First reflective layer 206, 406, 706 Dielectric layer 207, 407, 507, 507', 707 Upper surface 208, 408, 708 Flattened layers 209, 409, 709 Second dielectric layer 210, 410, 710 Second reflective layer 211, 411 CMOS layer 401 Sub-Tier 501 Barrier Layer 503A1, 503A2, 503A3 Pillar M1, M2, M3 Masks 506, 511' separation layer 508, 506', 507', 711 Layer of photosensitive or electron sensitive resist 509, 509', 712 Gap 510, 713 Resist lines 510' Resist 511 Separation material boundaries or walls F1, F2, F3, Fi Fabry-Perot cavity type color filters 713 remaining resist 714 Air Gap

Claims

1. A method (100, 300) for manufacturing a multispectral filter for electromagnetic radiation, said filter comprising at least two color filters, each filter comprising a first reflective layer, a second reflective layer, a layer of Fabry-Perot cavity dielectric material between the first reflective layer and the second reflective layer, the thicknesses of the dielectric layers of the two color filters being different, each of the two filters facing a photoelectric converter, said method comprising the following steps: Depositing a layer of structured material on a carrier substrate comprising at least two photoelectric converters; Three-dimensional structuring (101, 102, 301, 302) of a layer of structured material to obtain at least two structured material patterns of different heights, at least one of the patterns having a maximum reference height relative to a carrier substrate, the height being measured perpendicular to the plane of the substrate, each of the patterns facing a photoelectric converter, Conformally depositing a first reflective layer on the at least two structured material patterns (104, 303); A step (105, 304) of depositing a layer of dielectric material intended to form the dielectric pattern of a Fabry-Perot cavity, said layer of dielectric material covering all the patterns by having an upper surface, each point of which is located at a height higher than a maximum reference height, Planarizing the dielectric material by removal with selective stopping on top of the highest structured material pattern covered by the reflective layer (106, 305); depositing a layer (107, 306) of the same dielectric material to complete the formation of the Fabry-Perot cavity; Depositing a second reflective layer on the at least two Fabry-Perot cavities (108, 307). A method (100, 300) comprising:

2. 2. The method (100) according to claim 1, characterized in that the structuring material is a resist and that the structuring steps (101, 102) are carried out by a so-called greyscale lithography step on a layer of structuring material followed by a step of crosslinking the resist.

3. 3. The method of claim 2, characterized in that the structuring step is followed by a step (103) of conformally depositing a sealing layer on the at least two structured material patterns before a step (104) of conformally depositing a first reflective layer on the at least two structured material patterns, the first reflective layer covering the sealing layer.

4. The structuring material is an organic or inorganic separating material, and the structuring step comprises the following sub-steps: depositing a layer of organic or inorganic isolation material followed by a sub-step (301) of depositing a layer of resist; a sub-step (301) of subjecting the layer of resist to so-called greyscale lithography in order to three-dimensionally structure said layer of resist to at least two resist patterns of different heights, a substep (302) of transferring two resist patterns into a layer of organic or inorganic material to obtain said at least two structured material patterns of different heights; 2. The method (300) of claim 1, comprising:

5. 5. Method (100, 300) according to any one of claims 1 to 4, characterized in that the deposition of a layer made of a dielectric material intended to form the dielectric pattern of a Fabry-Perot cavity and the deposition of a layer made of the same dielectric material to complete the formation of the Fabry-Perot cavity are conformal depositions.

6. 6. The method according to claim 1, characterized in that the step of three-dimensionally structuring (601) the layer of structured material is a step of obtaining at least two structured material patterns of different heights with a void space between said at least two structured patterns.

7. After the step of conformally depositing a first reflective layer and before depositing a layer made of dielectric material intended to form the dielectric pattern of the Fabry-Perot cavity, the following steps are carried out: Depositing (603) a layer called separation layer, said separation layer covering all the structured material pattern by having an upper surface, each point of said upper surface being located at a height relative to the carrier substrate higher than a maximum reference height, depositing a layer of resist over the isolation layer (604); Removing the resist (605) to retain the resist only above the void spaces between the structured material patterns and to leave a clear separation layer above the structured material patterns; removing (606) material of the separation layer from the apparent zone of the separation layer; Removing the remaining resist (607) to preserve walls of said material of the separation layer between the structuring material patterns. The method of claim 6, comprising:

8. After the step of conformally depositing a first reflective layer and before depositing a layer made of dielectric material intended to form the dielectric pattern of the Fabry-Perot cavity, the following steps are carried out: Depositing (603') a layer of resist having an upper surface thereby covering all the structured material pattern, each point of the upper surface being located at a height relative to the carrier substrate that is higher than a maximum reference height, Removing the resist between the structured material patterns to retain the resist only above the structured material patterns (604'); depositing a layer called an isolation layer, said isolation layer covering all the structured material patterns covered by the resist and filling the spaces between said patterns by having an upper surface (605'), each point of said upper surface being located at a height relative to the carrier substrate that is higher than a maximum reference height, Removing the material of the separation layer and the remaining resist located above the structured material patterns in order to retain walls made of said material of the separation layer between the structured material patterns (606'). The method of claim 6, comprising:

9. The material of the separation layer is a sacrificial material, and the method comprises the following steps: depositing (613) a layer of resist on the second reflective layer; Removing resist from zones above the spaces between the structured material patterns (614); removing (615, 616) the material of the second reflective layer, the dielectric material intended to form the dielectric patterns of the Fabry-Perot cavities, and the sacrificial material in the zones above the spaces between the structured material patterns in order to form void walls between the structured material patterns and between the overhanging Fabry-Perot cavities, Step 617: Remove remaining resist.

9. The method according to claim 7 or 8, characterized in that it comprises:

10. The material of the separation layer is a low optical index dielectric material, and the method comprises the following steps: depositing (613) a layer of resist on the second reflective layer; Removing resist from zones above the spaces between the structured material patterns (614); Removing (615) the material of the second reflective layer and the dielectric material from the zones not protected by the remaining resist to reach the low optical index dielectric material; depositing (616') a layer of low optical index dielectric material over the remaining resist to fill the voids between the overhanging Fabry-Perot cavities and complement the dielectric material already present between the structured material patterns; removing (617') the resist and the low optical index dielectric material above the resist zones while leaving the low optical index material between the structured material patterns and between the overhanging Fabry-Perot cavities to form dielectric isolation walls; 9. The method according to claim 7 or 8, characterized in that it comprises:

Images