Process for preparing a microdevice of the light-emitting diode type emitting in the ultraviolet or visible spectral range

Abstract

The invention relates to a method for preparing a microdevice emitting in the ultraviolet or visible spectral range (or a microLED-type device), the microdevice obtained by implementing said method, and the uses of said microdevice, particularly for wireless optical communication applications, as a print head in maskless photolithography systems, or in the healthcare field. Figure to be published with the abbreviation: 5

Description

Title of the invention: Method for preparing a microdevice of the type light-emitting diode emitting in the ultraviolet or visible spectral range

[0001] The present invention relates to the field of light-emitting diode type devices (abbreviated as LEDs in French, or LEDs, from English: "Light-Emitting Diode").

[0002] More specifically, the invention relates to a method for preparing a microdevice emitting in the spectral range of the ultraviolet or visible (or microLED type device), the microdevice obtained by implementing said method, as well as the uses of said microdevice, in particular for wireless optical communication applications, as a print head in maskless photolithography systems, or in the field of health.

[0003] Devices emitting ultraviolet (UV) light are of increasing technological importance, with a growing range of applications in sectors such as healthcare, particularly for UV purification / sterilization applications, medical imaging, optoelectronics, and agriculture. UV light inactivates microorganisms (bacteria, spores, and viruses) by disrupting their DNA or RNA, thus ensuring water disinfection. This application is of great importance given that water quality is a global problem, with millions of people lacking access to safe drinking water and over a billion people without access to sanitation facilities.

[0004] Most current UV light technologies use UV discharge lamps (mercury, deuterium or xenon) and gas lasers (ArF, KrF, XeCl, etc.) which have a fixed and limited emission, and are generally bulky, energy-intensive, potentially toxic and short-lived.

[0005] On the contrary, the emission wavelength of gallium aluminum nitride (AlGaN)-based light-emitting devices can be adjusted across the entire UVA (380-315 nm), UVB (315-280 nm), and UVC (280-200 nm) spectral range by adding aluminum nitride (AIN) to the gallium nitride (GaN) alloy system. Furthermore, AlGaN-based devices are energy-efficient, robust, compact, environmentally friendly, and offer a long lifespan, making them ideal candidates for a wide range of UV-based applications. Thus, near-UVA (400-365 nm) and InGaN-based LEDs are already in use. for UV curing, whereas commercial UVC LEDs (270-280 nm) are already being used for water disinfection applications.

[0006] However, existing UV light technologies based on AlGaN devices still have insufficient efficiency and low luminous power, combined with a relatively high cost, which currently limits their mass adoption. The external quantum efficiency (EQE) of AlGaN-based LEDs (< 365 nm) is indeed limited to a few percent (< 10%), with a notable dip in the UVB range (310–350 nm) and a sharp drop below 250 nm, thus limiting their integration and mass adoption in practical devices and applications. These limitations stem in particular from the high defect density in AIN and AlGaN materials, the low efficiency of hole injection, and poor light extraction.

[0007] The advantages of microLEDs compared to conventional millimeter-wave LEDs are improved light extraction, current distribution, and integration compatibility. Furthermore, when these microLEDs are transferred and bonded to silicon substrates, it is possible to produce matrix-addressable micropixels with better heat dissipation, enabling high current density and high output power. Finally, as the diameter of the microLEDs decreases, the modulation bandwidth increases, which also justifies the interest of these micrometer-wave devices for wireless optical communication.

[0008] For about ten years, research and development on InGaN-based microLEDs emitting in the visible spectrum has been booming, primarily driven by various display applications ranging from smartwatches to virtual and augmented reality headsets, and even the automotive industry. AlGaN-based UV microLEDs are the solution for a wireless optical communication system, both indoors and outdoors. Indeed, by adjusting the wavelength of the UV device (below 320 nm), it is possible to eliminate noise from solar radiation in the communication system. Furthermore, light scattering increases at shorter wavelengths (oc 1 / X4), which is particularly advantageous for communication devices without direct line of sight, especially when an obstacle, rain, or particles are located between the transmitter and the receiver.Finally, its low absorption coefficient in water around 340 nm, an order of magnitude lower than the values ​​obtained in the visible wavelength range, would allow its application to be extended to underwater wireless optical communications. The use of UV microLEDs is not limited to optical communication applications and also represents a major interest for the market of maskless UV lithography and the processing of... skin and medical imaging. However, AlGaN-based UV microdevices are still far from reaching sufficient maturity to be used in these applications.

[0009] There are various approaches and strategies for fabricating microLEDs. The top-down approach is a subtractive process that relies first on using a lithography technique to define a mask with a specific configuration (diameter and spacing), and second, on removing material from the area defined by the mask. This removal is generally performed by dry etching, which allows for good control over the etching anisotropy according to the following equation: Anisotropy = 1 - ((lateral etching rate) / (vertical etching rate)), and this holds true for a wide range of semiconductor materials. However, the dry etching used to create microLEDs has the disadvantage of generating surface damage and non-radiative recombination centers on the sidewalls, the presence of which limits the overall efficiency of the microLEDs. Although sidewall passivation can be used to minimize these effects, the sidewall defects generated during the etching process cannot be completely avoided.

[0010] An alternative to the top-down approach is to grow these microstructures from one or more epitaxial layers. This is referred to as a bottom-up approach. This growth can be self-organized or organized depending on the strategy adopted. Numerous strategies have been developed in the visible spectrum for InGaN. However, these approaches are not necessarily transferable to AlGaN and therefore to the UV spectrum.

[0011] There is therefore a need for a process enabling access to microdevices without the disadvantages of prior art processes.

[0012] The present invention therefore has as its first object a method for manufacturing a microdevice emitting in the ultraviolet or visible spectral range comprising a plurality of micropallets of a semiconductor material based on an alloy of aluminum nitride and gallium nitride (AlGaN) or an alloy of aluminum nitride and indium nitride (AlInN), said method comprising, in this order, at least the following steps: a) the epitaxial growth of a sacrificial gallium nitride (GaN) layer A on the surface of a single-crystal substrate, said layer having a thickness EpA, b) the epitaxial growth of a layer B of an alloy of formula AlxXbxN, in which X is an element III chosen from gallium and indium, on the surface of said sacrificial layer A, said layer B having a thickness EpB, c) the creation of a mask C on the surface of said layer B, said mask C having a defined pattern, d) the transfer of the pattern of mask C by dry etching through layers B and A, to the surface of the single-crystal substrate to create a plurality of pillars, each of the pillars being composed of a gallium nitride base (corresponding to layer A), said base being surmounted by a layer of AlxXbxN (corresponding to layer B), e) removal of mask C, f) selective etching of the gallium nitride base of the pillars formed in step d) to obtain a plurality of gallium nitride micro-pillars, each supporting an individualized AlxXi_xN microelement of said layer B, (g) the epitaxial growth of a stack D of element III nitride layers on the surface of each individual microelement of said layer B, said stack D corresponding to an epistructure of a light-emitting diode and comprising in this order 1) a layer of an n-doped material, 2) an active region consisting of alternating barrier layers and quantum well layers, 3) an electron-blocking layer, and 4) a layer of a p-doped material, said process being characterized in that: - in the alloy of formula AlxXbxN of layer B, x is such that 0.05 <x<0,40, - the selective etching of the gallium nitride base of the pillars formed in step d) is carried out by thermal evaporation of layer A under a controlled atmosphere, at a temperature T equal to or greater than the evaporation temperature TA of layer A and strictly less than the evaporation temperature TB of layer B, and in that it further comprises: (h) a step of separation and transfer of the assembly formed by said stacks D and the individualized microelements of said layer B onto a transfer support, by mechanical fracture of the gallium nitride micropillars, i) the removal of individual microelements from said layer B and any gallium nitride residues from the micropillars, (j) the deposition of a stack of metallic layers on the surface of the n-type doped material layer of each individual microelement of the stack D in order to achieve metallic contact with said n-type layer, and k) the deposition of a stack of metallic layers on the surface of the layer of doped material p of each individual microelement of the stack D in order to to make metallic contact with said p-type layer, said step k) being carried out before or after step h) of separation and transfer of the assembly formed by said stacks D and layer B.

[0013] Thanks to the process according to the invention, it is now possible to obtain micro-LED devices that do not have the drawbacks of prior art processes and that exploit the good structural quality of the sacrificial gallium nitride layer A to allow the subsequent growth of an AlxGaixN-based stack D of different compositions without generating defects such as dislocations or cracks. Furthermore, the design used to fabricate the final micro-device ensures the removal of the GaN via transfer onto the host substrate and etching of the back side, which prevents any potential reabsorption of emitted photons in the active area while maximizing light extraction, a major problem in the fabrication of UV devices.Finally, the proposed approach achieves bottom-up growth of micrometer-sized LED structures without going through top-down etching of the entire LED structure, which is known to significantly degrade the efficiency of these microcomponents.

[0014] According to the invention, epitaxial growth of the surface layer A of gallium nitride can be carried out on any single-crystal substrate that allows for the growth of hexagonal-phase GaN along the c-axis. In a preferred embodiment, said single-crystal substrate is selected from c-planar sapphire (Al₂O₃), silicon (111), silicon carbide, bulk substrates of gallium nitride or aluminum nitride, and glass. Among such substrates, c-planar sapphire and silicon (111) are particularly preferred.

[0015] According to one embodiment, the thickness of layer A is approximately 1000 to 10000 nm, preferably approximately 4000 to 6000 nm.

[0016] The thickness of layer B can vary from 40 to 1000 nm. This thickness depends on the respective contents of aluminium and element X in the alloy of formula AlxXbxN and also on the nature of element X.

[0017] According to a first embodiment concerning layer B, and where X is gallium, the thickness of layer B can be from approximately 40 nm to 1000 nm. Typically, for an aluminum content of around 5%, the thickness of layer B will be approximately 800 to 1000 nm, for an aluminum content of around 15% the thickness of layer B will be approximately 100 to 150 nm, while for an aluminum content of 40% the thickness of layer B will be approximately 40 nm. Thus, when the aluminum content in the alloy of formula AlxXi_xN is between 5 and 40% inclusive, the growth of layer B occurs under optimal conditions, that is to say, under pseudo-optimal conditions morphic, without plastic relaxation and without introduction of defects into its structure, such as cracks or dislocations.

[0018] According to a second embodiment concerning layer B, and when element X is indium, the thickness of layer B can be from approximately 40 nm to 300 nm. Typically, for an indium content of around 17%, the AUnN alloy is in lattice match with GaN, so the thickness of layer B will be around 300 nm before the generation of defects such as holes. For indium contents below or above 17%, the AUnN is either in tension or compression, so the thickness of layer B will be around 40 nm for an indium content close to the limits of the range, i.e., around 14% or 20%.Thus, when the indium content in the alloy of formula AlxXbxN is between 14 and 20% inclusive, the growth of the B layer is carried out under optimal conditions, i.e. under pseudo-morphic conditions, without plastic relaxation and without introduction of defects into its structure, such as cracks, dislocations or holes.

[0019] The growth of layers A and B is preferably carried out by organometallic vapor phase epitaxy.

[0020] In step c), the nature of the mask C deposited on layer B is not critical. For example, the mask C may consist of a layer of resin, a layer of a dielectric material, or even a metallic layer.

[0021] The patterns of mask C can of course be of variable geometry, for example circular, square, rectangular, etc. and of variable dimensions as well, typically from dimensions that can vary from micron to hundreds of microns.

[0022] According to an embodiment of step c) of the process of the invention, the mask C consists of a resin layer having a thickness Epc. In this case, the thickness Epc of the resin mask C is preferably such that Epc > 2x(EpA+EpB). Typically, and depending on the thicknesses EpA and EpB of layers A and B, the thickness Epc of the resin mask is approximately 2000 to 22000 nm, preferably approximately 10000 to 14000 nm.

[0023] Still according to this embodiment of step c), the creation of the mask C on the surface of layer B can then be carried out by optical lithography according to techniques well known to those skilled in the art.

[0024] In step d), the geometry of the pillars is determined by the patterns of the mask C, their height being fixed by the respective thicknesses EpA and EpB of layers A and B. Typically, the height of the pillars is from 2000 to 10000 nm, and even more preferably from 4000 to 6000 nm. By way of example, and when said mask has circular patterns, the pillars then have a circular cross-section whose diameter is typically in the range of 1 to 100 pm approximately, preferably from 5 to 20 pm approximately.

[0025] The dry etching of step d) can be carried out in an enclosure equipped with an inductive plasma source in the presence of gaseous precursors, in particular chlorinated, which can in particular be chosen from chlorine (Cl2), boron trichloride (BC13) and argon (Ar).

[0026] Step e) of removing mask C can be carried out using techniques well known to those skilled in the art, depending on the nature of the layer constituting mask C. Thus, for example, when mask C consists of a resin layer, step e) can be carried out by treatment with an oxygen plasma, followed by one or more chemical treatment steps, in particular with a mixture of sulfuric acid and hydrogen peroxide and then with an etching solution such as a mixture of ammonium fluoride (NH4F) and hydrofluoric acid (HF). This step e) also makes it possible to clean the pillars obtained at the end of step d), which facilitates the subsequent execution of step f) of selective etching of the gallium nitride base of said pillars.

[0027] During step f), the selective etching of the gallium nitride pillars (or layer A) is made possible by the judicious choice of temperature T, allowing the selective decomposition of layer A constituting the pillars while preserving the integrity of layer B of the AlxXbxN alloy as indicated above. Thus, at the end of step f), a plurality of micro-pillars are obtained, each supporting an individual element of layer B whose shape and dimensions correspond to the patterns of the mask C.

[0028] Step f) can be carried out in a chamber heated to temperature T in the presence of a carrier gas. The carrier gas is preferably chosen from dihydrogen, nitrogen and a mixture thereof.

[0029] According to one embodiment of the invention, step f) is carried out at a pressure of approximately 50 to 300 mbar, preferably approximately 75 to 150 mbar.

[0030] During step f), and according to a preferred embodiment of the invention, the temperature T is approximately 900 to 1100 °C.

[0031] At the end of step f), the micropillars of layer A are such that, in cross-section along an axis substantially perpendicular to their height, they exhibit a larger dimension, preferably from approximately 0.2 to 5 pm, and even more preferably from 1 to 3 pm. It has indeed appeared that beyond 5 pm, the mechanical fracture of the micropillars during step h) is more difficult to achieve.

[0032] Step g) allows the epitaxial growth of the stacking D of the different layers of element III nitrides necessary for the formation of a plurality of light-emitting diode epistructures on the surface of each microelement individualized from said layer B according to techniques known in the field. The fabrication of these different layers can, for example, be carried out with the same technique as that used to fabricate layers A and B, namely metal-organic vapor deposition.

[0033] According to one embodiment of the invention, the stack D can be based on an aluminum nitride and gallium nitride alloy, with a variation in composition (respective percentages of Al and Ga) allowing for the stacking of layers necessary to obtain an LED epistructure emitting in the UVA, UVB, or UVC range. In particular, if we consider an alloy with the formula AfCa, yN, the value of y can vary from 0 to 1 depending on the wavelength we wish to target. Furthermore, in general, the first n-doped AlGaN layer, as well as the AlGaN barriers, will have an Al composition y higher than that of the quantum wells. The electron-blocking layer will have the highest Al composition y, and that of the p-doped layer will generally be equal to or lower than that of the n-doped layer.

[0034] According to another embodiment, the stack D can be based on an alloy of indium nitride and gallium nitride with a variation in composition (respective % in In and Ga) allowing the stacking of layers necessary to obtain an LED epistructure emitting in the visible (blue, green, or even red).

[0035] The total thickness of the different layers constituting the stack D is preferably from 1 to about 10 pm, and even more preferably from 2 to about 4 pm.

[0036] Step h) allows the set of said stacks D to be separated from the original substrate by transferring them onto a transfer substrate in order to carry out the subsequent manufacture of said micro device emitting in the spectral range of ultraviolet or visible depending on the LED epistructure adopted in the previous step, either directly on the transfer support if it allows the realization of said device, or on another support after an additional transfer.

[0037] According to one embodiment, the transfer support used in step h) is preferably chosen from substrates having a flat surface suitable for receiving a metallic deposit and undergoing a pressure bonding step at a temperature of around 300 to 400 °C. By way of non-limiting example, said transfer support may be chosen from a single-crystal silicon substrate, a metallic substrate, a flexible substrate, and an application-specific integrated circuit (ASIC), in particular of the complementary metal-oxide-semiconductor (CMOS) type.

[0038] According to a preferred embodiment of the invention, the transfer support is a single-crystal silicon substrate, which allows current to be injected into the final micro-device.

[0039] The mechanical fracture of step h) can, for example, be induced by bringing the assembly formed by said stacks D and the individualized micro-elements of said layer B into contact with the surface of the transfer support followed by the application of pressure.

[0040] According to one embodiment, the transfer on said transfer support of step h) can be carried out by a metal-to-metal type bonding, for example of the gold-to-gold type when the stacking of the metal layers on the surface of the doped material p carried out in step 1) is a stacking of gold layers.

[0041] Step i) of the process according to the invention allows for the roughening of the back face of the D stacks while also removing the previous layers, namely the gallium nitride residues forming the micro-pillars (layer A) as well as the individual micro-elements of layer B, and texturing the surface of the back face of said D stacks in order to maximize light extraction. At the end of said step i), only the D stacks remain, their back faces having been roughened.

[0042] According to the invention, step i) is an etching step which can be carried out by chemical etching or by dry etching.

[0043] According to one embodiment of the invention, step i) is a chemical etching step which can, for example, be carried out using an alkaline solution, in particular a potassium hydroxide solution.

[0044] According to another embodiment of the invention, step i) is a dry etching step which can, for example, be carried out in an enclosure equipped with an inductive plasma source in the presence of gaseous precursors, in particular chlorinated, which can in particular be chosen from chlorine (Cl2), boron trichloride (BC13) and argon (Ar).

[0045] Steps j) and k) of deposition of a stack of metallic layers can be carried out by vacuum evaporation, in particular by electron beam evaporation, thermal evaporation, or radio frequency magnetron sputtering.

[0046] The invention also relates to a micro device emitting in the spectral range of the ultraviolet or the visible, said micro device directly obtained by the process defined according to the first object of the invention.

[0047] A third object of the invention is the use of a microdevice directly obtained by implementing the process as defined according to the first object of the invention for applications emitting in the UV range, for example as a print head in maskless photolithography systems, or in the health sector, particularly for sterilization and / or disinfection of surfaces, skin treatment by phototherapy, medical imaging, or for applications emitting in the visible spectrum, for example for display (television screens, smartwatches, augmented reality and virtual reality technologies, etc.) or finally for applications covering both wavelength ranges such as wireless optical communication.

[0048] Other features and advantages of the present invention will become apparent from the following examples and the accompanying figures in which: [Fig.l] is a schematic cross-sectional view of a device being developed at the end of step c), said device comprising a single-crystal substrate 1 surmounted by a sacrificial layer A of gallium nitride 2, itself surmounted by a layer B of an alloy of formula AlxXbxN 3, layer 3 having on its surface a mask C 4,4'; [Fig.2] is a schematic cross-sectional view of a device being developed at the end of step d), said device comprising a single-crystal substrate 1 on which rest gallium nitride pillars 5,5', said pillars each being themselves composed of a sacrificial layer A of gallium nitride 2,2', itself surmounted by a layer B of an alloy of formula AlxXi_ XN 3,3', the layer 3,3' having on its surface a residual layer from the mask C 4,4'; [Fig.3] is a schematic cross-sectional view of a device being developed at the end of step f), said device comprising a single-crystal substrate 1 on which rest gallium nitride micropillars 6,6' whose diameter is necessarily smaller than that of the pillars 5,5' shown in [Fig.2], said micropillars 6,6' supporting individualized microelements 7,7' of layer B; [Fig.4] is a schematic cross-sectional view of a device being developed at the end of step g), said device comprising a single-crystal substrate 1 on which rest micro pillars 6,6' of gallium nitride supporting individual microelements 7,7' of layer B on which a stack D of layers corresponding to a light-emitting diode epistructure 8,8' has been grown epitaxially; [Fig. 5] is a schematic cross-sectional view of a device being developed during step h), said device comprising a single-crystal substrate 1 on which rest gallium nitride micropillars 6,6' supporting individualized microelements 7,7' of layer B on which rest a stack D of layers each corresponding to a light-emitting diode epistructure 8,8' was grown epitaxially. Each of the stacks 8,8' has on its surface a layer 9,9' comprising a metallic contact and a layer of a material allowing the junction with a transfer support 10. The elements 11,11' represent the mechanical fracture action carried out during step h); [Fig. 6] is a schematic cross-sectional view of a device at the end of step j), said device comprising said transfer support 10 on which rest a plurality of stacks D of layers each corresponding to a light-emitting diode epistructure 8,8', said stacks 8,8' being in contact with said transfer support 10 via said layer 9,9', said epistructures 8,8' having a rough surface 12,12' surmounted by a metallic contact 13,13' [Fig.7] gives scanning electron microscopy (SEM) images of the surface of a device according to the invention at different stages of its manufacture according to example 1 (tilted view): (a) pillar of Alo.15Gao.85N / GaN alloy, (b) and (c) microdisc of Alo.15Gao.x5N supported by a pillar of GaN, micropallets of Alo,5Gao,5N after (d) one hour of growth and (e) and (f) after 6 hours of growth; [Fig.8] gives scanning electron microscopy (SEM) images of the surface of a device according to the invention according to example 2.1 after the step of transferring the micropallets onto a flexible adhesive carbon film; [Fig.9] is a schematic cross-sectional representation of the transfer process on a silicon carrier support 10 with (a) the deposition of gold 14 on said carrier support 10, as well as 14' on an assembly of AlGaN micropallets 8,8' supported by GaN micropillars 6,6' resting on a sapphire substrate 1 according to example 2.2, (b) the gold-gold bonding 14,14' between the assembly of AlGaN micropallets 8,8' and the silicon carrier support 10, (c) the assembly of micropallets 8,8' is bonded to the carrier support 10 by means of the gold-gold bonding 14,14', after the fracturing of the GaN micropillars 6,6' and the removal of the sapphire substrate 1; [Fig. 10] gives scanning electron microscopy (SEM) images of the surface of a device according to the invention according to example 2.2 after the gold-to-gold bonding step and the transfer of the micropallets onto a Si support (Fig. 10a magnification x90, Fig. 10b magnification x500 and Fig. 10c magnification x50000); [Fig. 11] gives atomic force microscopy images of the surface of samples of the device prepared in example 1 after the different stages of the manufacturing process: (a) micro pillar of Alo.isGao.ssN / GaN, (b) micro disk of Alo.15Gao.x5N, and micro pallets of Alo,5Gao,5N after (c) 1 hour and (d) 6 hours of growth; [Fig. 12] shows reciprocal space maps made for the asymmetric (10-15) reflection of the device prepared in Example 1 after the different fabrication steps: (a) the 80 nm planar layer of AlGaN on 5 pm of GaN, and (b) the AlGaN / GaN micropillars. The horizontal dashed arrows indicate which layer the peak is associated with. The AlGaN layer is constrained by the GaN layer when aligned along the solid vertical line; [Fig. 13] shows reciprocal space maps made for the asymmetric (10-15) reflection of the device prepared in Example 1 after the different fabrication steps: (a) the AlGaN microdisks, and (b) the 6h AlGaN regrowth. The horizontal dashed arrows indicate which layer the peak is associated with. The AlGaN layer is constrained by the GaN layer when aligned along the solid vertical line. The AlGaN layer is no longer constrained by the GaN layer when aligned along the dashed vertical line; [Fig. 14] gives cathodoluminescence spectra of quantum well structures emitting at different wavelengths: 229 nm for the solid line spectrum, 251 nm for the dashed line spectrum, 285 nm for the dotted and dashed line spectrum, 309 nm for the dotted line spectrum. Each spectrum has been normalized by the maximum intensity relative to the emission peak of the quantum wells; [Fig. 15] shows cathodoluminescence spectra of LED structures emitting at different wavelengths: 285 nm for the solid line spectrum and 312 nm for the dashed line spectrum. Each spectrum has been normalized by the maximum intensity relative to the emission peak of the quantum wells.

[0049] EXAMPLE 1: Preparation of a micropallet network

[0050] In this example, the feasibility of the steps of the process according to the invention, which enables the fabrication of micrometer-sized semiconductor micropallets via a combination of top-down and bottom-up approaches, is demonstrated. One of the essential characteristics of the process according to the present invention, namely the production of compliant AlGaN microdisks by selective thermal etching, will also be clearly established.

[0051] The growth of element III nitrides was carried out in a metal-organic vapor epitaxy (MOVEP) reactor. The system is equipped with the organometallics necessary for obtaining AlGaN or InGaN-based DEL epistructures, namely trimethylgallium (TMGa), triethylgallium (TEGa), trimethylaluminum (TMA1), and trimethylindium (TMIn) for element III, and bis(cyclopentadienyl)magnesium (Cp2Mg) for p-type doping. Ammonia (NH3) provides element V, and silane (SiH4) provides n-type doping. All these gases were transported to the substrate surface by carrier gases, which are hydrogen (H2) or nitrogen (N2) or a combination of both. The temperature at the sample surface was estimated by optical measurement and adjusted according to the epitaxial layers. a) Growth of layer A

[0052] The sacrificial GaN layer (layer A), the surface of which is a plane (0001), was epitaxially grown on a two-inch (50 mm diameter) sapphire substrate with a thickness of approximately 5 pm. The growth steps of the GaN layer A were as follows: 1) Nitriding of the sapphire substrate with 3 standard liters per min (Lsm) of NH3 at 1000°C, at 200 mbars under carrier gas H2 and for 300 sec; 2) SiN treatment of the sapphire surface with 3 Lsm NH3, 200 standard cm3 per minute (cmsm) of SiH4 (100 ppm in H2) at 1000°C, at 200 mbars under carrier gas H2 and for 250 sec. 3) The temperature was reduced to 550°C in order to carry out a 3D deposition of GaN with 3 Lsm of NH3, 34.5 cmsm of TMGa, at 200 mbars under carrier gas H2 and for 140 sec; 4) The temperature was then increased to 1040°C in order to coalesce the 3D deposit of GaN with 12 Lsm of NH3, 130 cmsm of TMGa, at 300 mbars under carrier gas H2 and for 3000 sec; 5) Finally, the temperature is increased to 1060°C to smooth the GaN deposition with 10.3 Lsm NH3, 89 cmsm TMGa, 150 mbar under H2 carrier gas for 3000 sec. b) Epitaxial growth of layer B

[0053] Layer B is an alloy layer of formula AlxGabxN with a thickness of 80 nm and an aluminum composition of 15% (x=0.15), composition determined by X-ray diffraction. The growth of this layer was carried out following layer A of GaN under the following conditions: the temperature was fixed at 1040°C with 9 Lsm of NH3, 30 cmm of TMGa, 60 cmm of TMA1 at 150 mbar under carrier gas H2 and for 300 sec. c) Deposition of a resin mask C

[0054] The stack of layers A and B was transported to a cleanroom to create the resin mask C. After spreading a resin (AZ® 12XT-20PL sold by Merck KGaA, Darmstadt, Germany) onto the surface of layer B (conditions: 2000 rpm, 30 sec, and annealed at 110°C on a hot plate for 180 sec), disc-shaped patterns were defined by optical lithography (annealed at 90°C on a hot plate for 120 sec and developed in a tetramethylammonium hydroxide solution (AZ726 MIF Developer, Merck) for 80 sec) with a diameter and height of 20 and 12 pm, respectively. d) Transfer of the pattern from mask C

[0055] The disc-shaped resin mask pattern C was then transferred to layers B and C by dry etching in a chamber equipped with an inductively coupled plasma (ICP) source containing Cl2 BC13 type chlorinated precursors (15 cm³ of Cl2, 2 cm³ of BC13, 2 cm³ of Ar, 2 cm³ of N2, at a pressure of 0.66 Pa (5 mtorr), radio frequency (RF) magnetic field power 40 W and inductive coupling (ICP) power 800 W for 15 min). An optical detection system was used to detect when the sapphire substrate layer had been reached and, consequently, when layers B and A had been completely etched.

[0056] After cleaning and removal of the remaining mask (O2 plasma, 3:1 v / v sulfuric acid and hydrogen peroxide mixture (Piranha solution), and BOE etching solution (buffered oxide etch) composed of a 7:1 v / v mixture of ammonium fluoride (NH4F) and hydrofluoric acid (HF)), a network of Alo.15Gao.85N / GaN pillars was obtained. Figure 7a, a scanning electron microscopy (SEM) image, shows the morphology of an Alo.15Gao.85N / GaN pillar after dry etching and cleaning. e) Selective etching of GaN pillars

[0057] The pillar array is subsequently reintroduced into the EPVOM reactor to perform annealing under a hydrogen and ammonia environment. The estimated surface temperature is approximately 970°C, the pressure is set at 100 mbar, the ammonia flow rate at 1 Lsm, and the hydrogen carrier gas flow rate at 8 Lsm. Rapid alternation between annealing under hydrogen alone for 2 seconds and annealing under hydrogen plus ammonia for 3 seconds was performed a number of times to ensure controlled lateral etching of the sacrificial GaN layer (layer A). Figures 7b and 7c show SEM images of the annealing after 350 loops alternating between hydrogen alone and hydrogen plus ammonia. Microdisks of Al0.15Ga0.85N (layer B) supported by a GaN micropillar (layer A) can be observed. It is thus demonstrated that the higher thermal stability of the Alo.i5Gao.85N layer allows for selective etching of GaN without attacking the Alo.isGao 85N.The diameter of the Alo.15Gao.85N microdisk was 20 pm, while the diameter of the GaN pillar varied down to 2 pm at the interface with the Alo.isGao.85N. With the experimental parameters used, it is therefore possible to achieve controlled and uniform thermal etching between the microdisks.

[0058] f) Development of an AlGaN micropallet network

[0059] Following the formation of a network of AlGaN microdisks, the resumption of growth of layers of an AlGaN alloy (layer D) with aluminum compositions ranging from 0 to 100% and varying thicknesses was achieved, again in the same EPVOM reactor. A network of micropallets of an AlGaN alloy was thus obtained. Figure 7d shows a SEM image after 1h of growth of an AlGaN alloy layer with an aluminum composition of 50% (composition determined by X-ray diffraction), while Figures 7e and 7f show SEM images after 6h of growth for an AlGaN alloy with the same 50% aluminum composition. The surface of the micropallet did not exhibit any holes or cracks.

[0060] The growth of this layer D was carried out under the following conditions: the temperature was fixed at 1030°C with 0.5 Lsm of NH3, 121 cmm of TEGa, 33 cmm of TMA1 at a pressure of 100 mbars under carrier gas H2 and for a time that depends on the thickness to be obtained. Under these conditions the growth rate was approximately 750 nm / h.

[0061] EXAMPLE 2: Experiments transferring a network of micropallets onto a transfer support

[0062] 2.1) Transfer of micropallets onto an adhesive support

[0063] A flexible adhesive carbon film was manually glued to the surface of the micropallet network obtained above in Example 1, then pressure was applied with a flat surface object and the adhesive film was finally removed.

[0064] The result is illustrated by the attached [Fig. 8], which shows SEM images of this transfer test on the adhesive carbon film (Fig. 8a, 70x magnification, and Fig. 8b, 500x magnification). This test was not performed under industrial bonding conditions, but it demonstrates the feasibility of the approach with more than satisfactory transfer without any optimization. At 500x magnification, the indentation of the GaN pillar can be observed on the back face, and its diameter is estimated to be around 1 to 2 µm. This suggests that a diameter of 2 µm of the GaN pillar at the interface with the AlGaN micro-pallet appears sufficient to initiate fracture and thus the separation of the micro-pallets.

[0065] 2,2) Transfer of micropallets onto a silicon substrate by gold-to-gold bonding

[0066] Another transfer test was carried out, this time by performing a gold-on-gold bond. To do this, and with reference to the attached [Fig. 9], a deposit of gold 14,14' was made by electron beam evaporation on an array of AlGaN micropallets 8,8' as obtained according to the process of Example 1 above, as well as on a silicon transfer substrate 10, as schematically represented in [Fig. 9].a. At the evaporator outlet, the sample comprising the array of micropallets 8,8' was brought into contact with the gold layer 14 of the silicon support 10, via the a layer of gold 14' and pressure was applied manually using a flat surface object, as shown schematically in [Fig.9].b. Finally, the sapphire substrate 1 was removed, leaving a set of micropallets 8,8' transferred onto the silicon transfer substrate 10, as shown schematically in [Fig.9].c.

[0067] The final result is illustrated in [Fig. 10], which shows SEM images of this transfer test on a silicon substrate 10. (Fig. 10a, 90x magnification; Fig. 10b, 500x magnification; and Fig. 10c, 50,000x magnification). Although a conventional wafer bonding technique was not used for this gold-on-gold bonding, this test demonstrates the strong potential of the approach for transferring micropallets or microdevices based on element III nitrides. It is highly probable that the yield of transferred micropallets could approach 100% with the use of a conventional wafer bonding technique in which the bonding pressure and temperature can be finely adjusted.

[0068] EXAMPLE 3: Characterization of the different surface samples prepared in example 1 3.1) Study by atomic force microscopy

[0069] Atomic force microscopy (AFM) was used to evaluate more precisely the quality of the planar surface (0001) along the manufacturing process of the device prepared above in Example 1.

[0070] The results are presented in the attached [Fig. 11], which provides AFM images of the surface of samples after the different stages of the manufacturing process: (a) A10.15Ga0.85N micropillar, (b) A10.15Ga0.85N microdisk, and A10.5Ga0.5N micropallets after (c) 1 hour and (d) 6 hours of growth. It is observed that the surface of the micropillar (Figure 11a), the A10.15Ga0.85N microdisk (Figure 11b), and the A10.5Ga0.5N micropallets having an aluminum composition of 50% (Figure 11a and d for 1h and 6h of growth, respectively) has a low surface roughness, or RMS, as shown in the images of [Fig. 11]. These AFM images also allow the extraction of other useful information such as the density of through-dislocations (DDT). These dislocations emerge at the surface of the (0001) plane and form a nanometer-sized crater that can be seen as a black spot on the AFM images.We observe that the DDT remains of the same order of magnitude throughout the fabrication process and reaches values ​​in the low 10⁸ cm⁻² for 5 pm of A10,5Ga0,5N growth, which represents values ​​consistent with the state of the art in the field. Indeed, through-hole dislocations are defects that directly impact the internal quantum efficiency of the LED structure. It is predicted that values ​​in the low 10⁸ cm⁻² are necessary to maximize efficiency. internal quantum and obtain values ​​greater than 60-70% (Ban, K. et al., “Internai quantum efficiency of whole-composition-range AlGaN multiquantum wells.”, Appl. Phys. Express, 2021, 4, 052101). 3.2) Mapping of Reciprocal Space

[0071] Figures 12 and 13 show reciprocal space maps performed for asymmetric (10-15) reflection for (Fig. 12a) the 80 nm planar layer of Alo,i5Gao,85N on 5 pm of GaN, (Fig. 12b) the micropillars of Alo.isGao.ssN / GaN, (Fig. 13a) the microdisks of Alo^Gao^N, and (Fig. 13b) the 6h Alo^Gao^N regrowth. On Fig. Figure 12a shows that the peak corresponding to the 80 nm Alo15Gao85N is aligned with that of GaN (along the vertical solid line), which means that the 80 nm Alo15Gao85N layer is indeed under stress on the GaN layer, as expected in the case of pseudomorphic growth. When the micropillars are formed, the Alo15Gao85N layer partially relaxes but still remains under stress on the GaN (along the vertical solid line). After annealing and formation of the Alo15Gao85N microdisks, the Alo15Gao85N peak is clearly no longer aligned with that of GaN as shown in Figure 12a.Figure 13a (dashed line, not solid line) suggests that the Alo.i5Gao.85N microdisk is even more relaxed. Finally, after 6 hours of Alo.5Gao.5N growth, Figure 13b shows that the Alo.5Gao.5N peak aligns with that of the Alo.isGao.ssN microdisk (along the dashed line) while being completely uncorrelated with GaN (solid line). Thus, the Alo.i5Gao.85N microdisk acts as a compliant material, allowing the accumulation of stress imposed by the growth of Alo.5Gao.5N.

[0072] EXAMPLE 4: Preparation and characterization of quantum wells emitting at different wavelengths according to the method of the invention

[0073] Steps a) to e) of the process of Example 1 were reproduced identically and then the resumption of growth of an AlGaN layer and of a quantum well type structure (composed of six barriers and five quantum wells) were then carried out in the same EPVOM reactor.

[0074] Four quantum well structures emitting at different wavelengths were fabricated under the following conditions:

[0075] 1 — quantum well structure emitting at 309 nm: The AlGaN layer was fabricated under the following conditions: the temperature was set at 1025°C with 0.5 Lm of NH3, 160 cmm of TEGa, 30 cmm of TMA1, a pressure of 100 mbar under a hydrogen carrier gas (H2), and for 2 hours. The barriers were fabricated under the following conditions: the temperature was set at 1025°C with 3 Lm of NH3, 100 cmm of TEGa, 15 cmm of TMA1, a pressure of 100 mbar under a hydrogen carrier gas (H2), and for 60 seconds. The quantum wells were fabricated under the following conditions: the temperature was set at 1025°C with 3 Lsm of NH3, 150 cmm of TEGa, 10 cmm of TMA1, at a pressure of 100 mbars under carrier gas H2 and for 7 seconds.

[0076] 2 — quantum well structure emitting at 285 nm: The AlGaN layer was produced under the following conditions: the temperature was set at 1025°C with 0.5 Lm of NH3, 121 cmm of TEGa, 33 cmm of TMA1, a pressure of 100 mbar under carrier gas H2, and for 3 hours. The barriers were produced under the following conditions: the temperature was set at 1025°C with 3 Lm of NH3, 65 cmm of TEGa, 16 cmm of TMA1, a pressure of 100 mbar under carrier gas H2, and for 60 seconds. Quantum wells were performed under the following conditions: the temperature was set at 1025°C with 3 Lsm of NH3, 110 cmm of TEGa, 10 cmm of TMA1, at a pressure of 100 mbar under carrier gas H2 and for 7 seconds.

[0077] 3 — quantum well structure emitting at 251 nm: The AlGaN layer was produced under the following conditions: the temperature was set at 1025°C with 0.5 Lm of NH3, 110 cmm of TEGa, 33 cmm of TMA1, a pressure of 100 mbar under carrier gas H2, and for 3 hours. The barriers were produced under the following conditions: the temperature was set at 1025°C with 3 Lm of NH3, 75 cmm of TEGa, 35 cmm of TMA1, a pressure of 100 mbar under carrier gas H2, and for 40 seconds. Quantum wells were performed under the following conditions: the temperature was set at 1025°C with 3 Lsm of NH3, 60 cmm of TEGa, 15 cmm of TMA1, at a pressure of 100 mbar under carrier gas H2 and for 10 seconds.

[0078] 4 — quantum well structure emitting at 229 nm: The AlGaN layer was produced under the following conditions: the temperature was set at 1025°C with 0.5 Lm of NH3, 50 cmm of TEGa, 50 cmm of TMA1, a pressure of 100 mbar under carrier gas H2, and for 3 hours. The barriers were produced under the following conditions: the temperature was set at 1025°C with 3 Lm of NH3, 25 cmm of TEGa, 25 cmm of TMA1, a pressure of 100 mbar under carrier gas H2, and for 60 seconds. Quantum wells were performed under the following conditions: the temperature was set at 1025°C with 3 Lsm of NH3, 25 cmm of TEGa, 15 cmm of TMA1, at a pressure of 100 mbar under carrier gas H2 and for 10 seconds.

[0079] Cathodoluminescence spectra were acquired on the surface of a micropalette for each of the different quantum well structures and are shown in the attached [Fig. 14]. Four distinct emissions at 309 nm (dashed curves), 285 nm (curve with dashes and dashes), 251 nm (curves with dashes), and 229 nm (solid line curve) are observed. Each spectrum has been normalized by the maximum intensity associated with the quantum well peak emission. Note that the emissions from the structures at 285 nm and 309 nm have very low intensity from deep defects (located at longer wavelengths, between 350 and 550 nm). This example demonstrates that the process according to the present invention makes it possible to grow active regions with good optical characteristics and covering a wide emission range in the UV range, thus allowing their use for a wide range of applications.

[0080] EXAMPLE 5: Preparation of a micro LED type device according to the process of the invention

[0081] Steps a) to e) of the process of Example 1 were reproduced identically, then the resumption of growth of an n-doped Aloj5Gaoj5N layer (layer D) was carried out in the same EPVOM reactor and under the following conditions: the temperature was set at 1100°C with 0.5 Lsm of NH3, 180 cmsm of TEGa, 30 cmsm of TMA1, with an outlet silane flow of 0.69 (source 25 sccm, dilution 1000 sccm and injection 25 sccm), at a pressure of 100 mbar under carrier gas H2 and for 3 hours.

[0082] A quantum well structure composed of six barriers and five quantum wells was then fabricated. The barriers were fabricated under the following conditions: the temperature was set at 1025°C with 3 Lsm of NH3, 100 cmm of TEGa, and 15 cmm of TMA1, at a pressure of 100 mbar under carrier gas H2 for 60 seconds. The quantum wells were fabricated under the following conditions: the temperature was set at 1025°C with 3 Lsm of NH3, 150 cmm of TEGa, and 10 cmm of TMA1, at a pressure of 100 mbar under carrier gas H2 for 12 seconds.

[0083] The electron blocking layer was produced under the following conditions: the temperature was set at 1000°C with 3 Lsm of NH3, 90 cmsm of TEGa, 30 cmsm of TMA1, 200 cmsm of Cp2Mg, at a pressure of 100 mbar under carrier gas H2 and for 60 seconds.

[0084] The p-type layer is a stack of layers carried out according to the following conditions:

[0085] pAlGaNl: the temperature was set at 1000°C with 3 Lsm of NH3, 100 cmsm of TEGa, 20 cmsm of TMA1, 100 cmsm of Cp2Mg, at a pressure of 100 mbars under carrier gas H2 and for 600 seconds.

[0086] pAlGaN2: the temperature was set at 1000°C with 3 Lsm of NH3, 140 cmsm of TEGa, 10 cmsm of TMA1, 100 cmsm of Cp2Mg, at a pressure of 100 mbars under carrier gas H2 and for 300 seconds.

[0087] pGaNl: the temperature was set at 1000°C with 4.1 Lsm of NH3, 17 cmsm of TMGa, 150 cmsm of Cp2Mg, at a pressure of 100 mbars under carrier gas H2 and for 12 seconds.

[0088] pGaN2: the temperature was set at 1000°C with 4.1 Lsm of NH3, 17 cmsm of TMGa, 150 cmsm of Cp2Mg, at a pressure of 100 mbars under carrier gas H2 and for 30 seconds.

[0089] Finally, an activation annealing of the p-type layer is carried out under the following conditions: the temperature was set at 810°C, at a pressure of 100 mbars under carrier gas N2 and for 1200 seconds.

[0090] These conditions result in an LED structure emitting at 312 nm.

[0091] It was also possible to obtain an LED structure emitting at 285 nm in modifying the parameters as follows: The regrowth of an n-doped Alo^Gao^N layer (D layer) was carried out in the same EPVOM reactor under the following conditions: the temperature was set at 1100°C with 0.5 Lsm of NH3, 160 cmsm of TEGa, 30 cmsm of TMA1, with an outlet silane flux of 0.69 (source 25 sccm, dilution 1000 sccm and injection 25 sccm), at a pressure of 100 mbar under carrier gas H2 and for 3 hours.

[0092] A quantum well structure composed of six barriers and five quantum wells was then fabricated. The barriers were fabricated under the following conditions: the temperature was set at 1025°C with 3 Lsm of NH3, 65 cmm of TEGa, and 16 cmm of TMA1, at a pressure of 100 mbar under carrier gas H2 for 80 seconds. The quantum wells were fabricated under the following conditions: the temperature was set at 1025°C with 3 Lsm of NH3, 125 cmm of TEGa, and 10 cmm of TMA1, at a pressure of 100 mbar under carrier gas H2 for 10 seconds.

[0093] The electron blocking layer was produced under the following conditions: the temperature was set at 1000°C with 3 Lsm of NH3, 75 cmsm of TEGa, 37 cmsm of TMA1, 200 cmsm of Cp2Mg, at a pressure of 100 mbar under carrier gas H2 and for 60 seconds.

[0094] The p-type layer is a stack of layers carried out according to the following conditions: pAlGaNl: the temperature was set at 1000°C with 3 Lsm of NH3, 75 cmsm of TEGa, 22 cmsm of TMA1, 100 cmsm of Cp2Mg, at a pressure of 100 mbars under carrier gas H2 and for 700 seconds.

[0095] pAlGaN2: the temperature was set at 1000°C with 3 Lsm of NH3, 100 cmsm of TEGa, 10 cmsm of TMA1, 100 cmsm of Cp2Mg, at a pressure of 100 mbars under carrier gas H2 and for 400 seconds.

[0096] pGaNl: the temperature was set at 1000°C with 4.1 Lsm of NH3, 17 cmsm of TMGa, 150 cmsm of Cp2Mg, at a pressure of 100 mbars under carrier gas H2 and for 12 seconds.

[0097] pGaN2: the temperature was set at 1000°C with 4.1 Lsm of NH3, 17 cmsm of TMGa, 150 cmsm of Cp2Mg, at a pressure of 100 mbars under carrier gas H2 and for 20 seconds.

[0098] Finally, an activation annealing of the p-type layer is carried out under the following conditions: The temperature was set at 810°C, at a pressure of 100 mbar under carrier gas N2 and for 1200 seconds.

[0099] Cathodoluminescence spectra were acquired on the surface of a micropalette for each of these two quantum well structures and are shown in the attached [Fig. 15]: 285 nm for the solid line spectrum and 312 nm for the dashed line spectrum. Each spectrum was normalized by the maximum intensity corresponding to the emission peak of the quantum wells.

Claims

1. Demands A method for manufacturing a microdevice emitting in the ultraviolet or visible spectral range comprising a plurality of micropallets of a semiconductor material based on an aluminum gallium nitride (AlGaN) alloy or an aluminum indium nitride (AlInN) alloy, said method comprising, in this order, at least the following steps: a) the epitaxial growth of a sacrificial gallium nitride (GaN) layer A on the surface of a single-crystal substrate, said layer having a thickness EpA, b) the epitaxial growth of a layer B of an alloy of formula AlxXbxN, in which X is an element III selected from gallium and indium, on the surface of said sacrificial layer A, said layer B having a thickness EpB, c) the creation of a mask C on the surface of said layer B, said mask C having a defined pattern, d) the transfer of the pattern of mask C by dry etching through layers B and A,up to the surface of the single-crystal substrate to create a plurality of pillars, each pillar consisting of a gallium nitride base (corresponding to layer A), said base being surmounted by an AlxXbxN layer (corresponding to layer B), e) removal of mask C, f) selective etching of the gallium nitride base of the pillars formed in step d) to obtain a plurality of gallium nitride micropillars, each supporting an individualized AlxXi_xN microelement of said layer B, g) epitaxial growth of a stack D of element III nitride layers on the surface of each individualized microelement of said layer B, said stack D corresponding to an epistructure of a light-emitting diode and comprising, in this order, 1) a layer of n-doped material, 2) an active region consisting of alternating barrier layers and quantum well layers, 3) a blocking layer electrons,and 4) a layer of a p-doped material, said process being characterized in that: - in the alloy of formula AlxXbxN of layer B, x is such that 0.05 <x<0,40, - la gravure sélective de la base en nitrure de gallium des piliers formés à l’étape d) est réalisée par évaporation thermique de la couche A sous atmosphère contrôlée, à une température T égale ou supérieure à la température TA d’évaporation de la couche A et strictement inférieure à la température TB d’évaporation de la couche B, et en ce qu’il comprend en outre : h) une étape de séparation et de transfert de l’ensemble formé par lesdits empilements D et les microéléments individualisés de ladite couche B sur un support de report, par fracture mécanique des micro piliers en nitrure de gallium, i) l’élimination des microéléments individualisés de ladite couche B et des résidus éventuels de nitrure de gallium des micro piliers,j) the deposition of a stack of metallic layers on the surface of the n-type doped material layer of each individual microelement of the stack D in order to achieve metallic contact with said n-type layer, and k) the deposition of a stack of metallic layers on the surface of the p-type doped material layer of each individual microelement of the stack D in order to achieve metallic contact with said p-type layer, said step k) being carried out before or after step h) of separation and transfer of the assembly formed by said stacks D and layer B.,

2. The method according to claim 1, characterized in that said single-crystal substrate is selected from c-plane sapphire (Al2O3), silicon (Al11), silicon carbide, gallium nitride or aluminum nitride bulk substrates and glass.

3. Method according to claim 1 or 2, characterized in that the thickness of layer A is from 1000 to 10000 nm.

4. A method according to any one of the preceding claims, characterized in that the thickness of layer B is 40 nm at 1000

5. nm. A method according to any one of the preceding claims, characterized in that the mask C is constituted by a layer of resin, a layer of dielectric material or a metallic layer.

6. Method according to claim 5, characterized in that the mask C is made of a layer of resin having a thickness Epc and in that Epc > 2x(EpA+EpB).

7. A method according to any one of the preceding claims, characterized in that at step d), the height of the pillars is from 2000 to 10000 nm.

8. A method according to any one of the preceding claims, characterized in that said mask has circular patterns and in that the pillars then have a circular cross-section whose diameter is from 1 to 100 pm.

9. A method according to any one of the preceding claims, characterized in that the dry etching of step d) is carried out in an enclosure equipped with an inductive plasma source in the presence of chlorinated gaseous precursors.

10. A method according to any one of the preceding claims, characterized in that step f) is carried out at a pressure of 50 to 300 mbar, preferably 75 to 150 mbar.

11. A method according to any one of the preceding claims, characterized in that during step f), the temperature T is from 900 to 1100°C.

12. A method according to any one of the preceding claims, characterized in that at the end of step f), the micro pillars have, in cross-section along an axis perpendicular to their height, a larger dimension which is from 0.2 to 5 pm.

13. A method according to any one of the preceding claims, characterized in that the stack D is based on an alloy of aluminium nitride and gallium nitride or on an alloy of indium nitride and gallium nitride.

14. A method according to any one of the preceding claims, characterized in that the total thickness of the different layers constituting the stack D is from 1 to 10 pm.

15. A method according to any one of the preceding claims, characterized in that the transfer support used in step h) is selected from a single-crystal silicon substrate, a metallic support, a flexible support, and an application-specific integrated circuit.

16. A method according to any one of the preceding claims, characterized in that the etching in step i) is a chemical etching step carried out using an alkaline solution.

17. Micro device emitting in the spectral range of ultraviolet or visible, said micro device being directly obtained by the process defined in any one of claims 1 to 16.

18. Use of a micro device directly obtained by implementing the process as defined in any one of claims 1 to 16 as a print head in maskless photolithography systems, or in the field of health, in particular for sterilization and / or disinfection of surfaces, skin treatment by phototherapy, medical imaging, for display, or for wireless optical communication applications.

Images