Optoelectronic device and manufacturing process

Abstract

Optoelectronic device and manufacturing method. The invention relates to a manufacturing method for three-dimensional (3D) structures for optoelectronics, each 3D structure comprising, in stacking along (z): - a lower part (21) bearing on a substrate (1), - an active region (22) configured to emit light radiation, said active region (22) bearing on a vertex (211) of the lower part (21), and - an upper part (23) bearing on a vertex (221) of the active region (22), said method comprising: - A substrate (1) carrying a plurality of lower parts (21) of 3D structures, said lower parts (21) having distinct vertices (211) such that the vertices (211) of two adjacent lower parts (21) are separated from each other by a separation distance ds of less than 180 nm,- Formation by organometallic precursor vapor-phase epitaxy (MOVPE) of the active regions on the vertices (211) of the lower parts (21), - Formation of the upper parts (23) on the vertices (221) of the active regions (22). The invention also relates to an optoelectronic device based on a plurality of these 3D structures. Figure for the abstract: Fig. 3.

Description

Title of the invention: Optoelectronic device and manufacturing method technical field

[0001] The invention relates to the field of optoelectronics. It finds a particularly advantageous application in the field of Gallium nitride (GaN) based light-emitting diodes having a three-dimensional structure. STATE OF THE ART

[0002] Light-emitting diodes (LEDs) typically include an active region where radiative recombinations of electron-hole pairs occur, which make it possible to obtain light radiation having a main wavelength.

[0003] For display applications, LEDs can be configured to produce light radiation whose main wavelength is in the blue, or in the green, or in the red.

[0004] This principal wavelength depends in particular on the composition of the active region. To produce light in the green or red range, the active region typically includes InGaN-based quantum wells. The higher the indium [In] concentration, the higher the principal wavelength. It may therefore be necessary to incorporate an indium [In] concentration > 10 at.% to obtain a red-emitting LED.

[0005] GaN-based LEDs are generally manufactured using a so-called planar technology, which consists of forming on a base plane of a substrate, a stack of two-dimensional (2D) layers in a direction normal to the base plane.

[0006] As illustrated in [Fig.1], this stacking can typically include, from the substrate 10, a buffer region 101 in GaN, an N-doped region 102 in GaN, the active region 103 comprising the InGaN-based quantum wells, a P-doped region 104 in GaN.

[0007] A structuring of this stack a posteriori, for example by lithography / engraving steps, then makes it possible to form a plurality of LEDs or micro-LEDs each having a mesa structure typically comprising a summit face 1040 and lateral walls 100 ([Fig.l]).

[0008] One drawback of this type of mesa structure is related to the post-processing. The lateral walls 100 obtained by etching generally exhibit defects that promote the occurrence of non-radiative surface recombinations. The radiative efficiency of the LEDs is reduced.

[0009] One solution for reducing sidewall defects is to directly form a three-dimensional (3D) GaN-based structure. These 3D structures can be in the form of GaN-based microwires 2 or nanowires 2 extending mainly in a z-direction normal to the substrate 1, as illustrated in [Fig. 2]. They typically comprise a lower portion 21 bearing on the substrate 1, an active region 22 bearing on the lower portion 21, and an upper portion 23 bearing on the active region 22, stacked in the longitudinal z-direction.

[0010] These 3D structures 2 typically exhibit an axial architecture in which the InGaN-based quantum wells of the active region 22 extend transversely, parallel to the xy plane of the substrate 1. Such an axial architecture makes it possible in particular to incorporate a significant concentration of indium (In) in the quantum wells of the active region 22. This axial architecture can therefore be used to fabricate green or red 3D LEDs or micro-LEDs.

[0011] The fabrication of these so-called axial 3D structures is typically carried out by molecular beam epitaxy (MBE) from a GaN layer 11. This growth technique, unlike other conventionally used growth techniques, allows for the localized deposition of InGaN at the tips of the GaN-based wires. InGaN-based quantum wells can thus be formed parallel to the substrate. Axial 3D LEDs in the form of wires configured to emit in the green or red range, each comprising a lower GaN-based portion, an active InGaN-based region arranged transversely to the wire, and an upper GaN-based region, can thus be fabricated.

[0012] However, molecular beam epitaxy (MBE) is not a technique compatible with an industrial manufacturing process. Axial 3D LEDs exhibiting good emission performance in the green or red are therefore not viable on an industrial scale.

[0013] The present invention aims to overcome at least partially these drawbacks.

[0014] In particular, an object of the present invention is to propose an improved method for manufacturing a plurality of 3D structures for optoelectronics.

[0015] Another object of the present invention is to propose an optoelectronic device, in particular an axial 3D LED based on GaN, which can be manufactured in an optimized way.

[0016] The other objects, features and advantages of the present invention will become apparent from an examination of the following description and accompanying drawings. It is understood that other advantages may be incorporated. SUMMARY

[0017] To achieve the objectives mentioned above, the present invention provides according to a first aspect a manufacturing process for a plurality of three-dimensional (3D) structures for optoelectronics, each 3D structure comprising, stacked along a longitudinal direction z: - a lower part comprising a base resting on a substrate, - an active region configured to emit or receive light radiation, said active region resting on a vertex, opposite the base, of the lower part, and - an upper part supported by a vertex of the active region,

[0018] said method comprising: - A supply of a substrate carrying a plurality of lower parts of 3D structures, said lower parts having distinct vertices such that the vertices of two adjacent lower parts are separated from each other by a separation distance ds of less than 180 nm, - Formation by organometallic precursor vapor-phase epitaxy (MOVPE) of active regions on the apexes of the lower parts, - Formation of the upper parts on the apexes of the active regions.

[0019] The formation by MOVPE organometallic precursor vapor phase epitaxy of the active regions advantageously makes the manufacturing process of axial 3D LEDs industrially compatible.

[0020] According to a technical prejudice, it is generally accepted that MOVPE does not allow the formation of such axial 3D LEDs and that it is necessary to use MBE to manufacture such structures.

[0021] Thus, known solutions based on MBE aim in particular to size the equipment enabling the implementation of MBE, so that they are compatible with industrial production.

[0022] According to the technical assumption mentioned above, MOVPE precursor vapor-phase epitaxial growth produces substantially conformal layers. Thus, according to this assumption, an MOVPE deposition of InGaN on the lower part of the 3D structure, for example a GaN-based wire, forms a continuous InGaN layer on the sides and top of the lower part or of the wire, and the resulting 3D structure therefore exhibits a so-called radial architecture.

[0023] On the contrary, in the context of the development of the present invention, it has surprisingly emerged that such a MOVPE deposit of InGaN on a set of lower parts of 3D structure based on GaN close to each other, makes it possible to obtain an axial 3D structure, where the InGaN-based material is located predominantly on the top of the lower parts.

[0024] According to one principle of the invention, the apexes of the lower parts are separated from each other by a separation distance ds of less than 180 nm. The deposited part Periodic MOVPE of InGaN quantum wells and AlGaN barriers advantageously allows the formation of active regions on the vertices of the lower GaN-based parts. These active regions can exhibit a truncated pyramid shape at the top of the lower parts, for example at the tips of the wires.

[0025] These high density conditions of the lower parts or wires are favorable for MOVPE to advantageously produce axial 3D structures. These high density conditions are typically obtained for separation distances ds, between adjacent vertices of the lower parts or wires, of less than or equal to 180 nm. This process is compatible with the industrial fabrication of axial 3D LEDs.

[0026] In addition, the active summit regions may have sufficient concentrations of indium, for example [In] > 10%at, to form LEDs configured to emit light radiation in the green or in the red.

[0027] According to one possibility, the lower parts are in the form of wires and are formed by MOVPE through openings in a masking layer. In this case, the openings in the masking layer are regularly distributed in a lattice with a pitch less than or equal to 700 nm. This pitch partly determines the separation distance ds between the wire tips. After the growth of the GaN-based wires, they are therefore relatively close to each other. This promotes axial growth of InGaN at the wire tips, to form 3D structures with an axial architecture.

[0028] A second aspect of the present invention relates to an optoelectronic device comprising a plurality of three-dimensional (3D) structures for optoelectronics, each 3D structure having a wire shape and comprising, according to a so-called axial architecture: - a lower part extending in a longitudinal direction and comprising a base resting on a substrate, - an active region, preferably comprising at least one quantum well extending along a plane normal to the longitudinal direction, configured to emit or receive light radiation, optionally from at least one quantum well, said active region resting on a vertex, opposite the base, of the lower part, and - an upper part supported by a vertex of the active region.

[0029] Advantageously, this axially structured 3D device further comprises a masking layer in contact with a surface layer of the substrate. This masking layer advantageously includes openings through which the lower, wire-like parts extend. The bases of the lower parts thus bear against the surface layer of the substrate.

[0030] The integration of such a masking layer advantageously allows the use of MOVPE to form, respectively, the lower part, active region, and upper part of the axial 3D structure. This masking layer can be retained after the fabrication of the 3D structures.

[0031] This masking layer also serves to mechanically reinforce the resulting axial 3D structure. It can be made of a dielectric material so as to electrically insulate the substrate from a metallic contact subsequently deposited on the top of the 3D structures, for example. It can also be transparent so as to allow light emitted or received by the 3D structures to pass through.

[0032] Advantageously, two vertices of two adjacent lower parts are separated from each other by a separation distance ds of less than 180 nm. This makes it possible in particular to obtain an axial 3D structure by MOVPE.

[0033] It is understood that the features and advantages of one aspect of the invention can be transposed, mutatis mutandis, to another aspect of the invention. BRIEF DESCRIPTION OF THE FIGURES

[0034] The aims, objects, features and advantages of the invention will become clearer from the detailed description of embodiments thereof, which are illustrated by the following accompanying drawings in which:

[0035] [Fig.1] Fig.1 illustrates a 3D LED having a mesas structure according to the prior art.

[0036] [Fig.2] Fig.2 illustrates a 3D LED in the form of wires, each exhibiting an axial architecture according to the prior art.

[0037] [Fig.3] Fig.3 illustrates a plurality of 3D structures in the form of wires, each exhibiting an axial architecture according to an embodiment of the present invention.

[0038] [Fig.4] The [Fig.4] is a scanning transmission electron microscopy (STEM) image of 3D structures according to an embodiment of the present invention.

[0039] [Fig.5A] The [Fig.5A] is an enlarged view of the thumbnail a materialized on the image of the [Fig.4] showing quantum wells of a 3D structure according to an embodiment of the present invention.

[0040] [Fig.5B] The [Fig.5B] is an enlarged view of thumbnail b materialized on the image of the [Fig.4] showing a side of a 3D structure according to an embodiment of the present invention.

[0041] [Fig.6] The [Fig.6] presents a curve showing different values ​​of the ratio of the thicknesses deposited by MOVPE radially and axially on the wires, as a function of different separation distances ds between wires, according to an embodiment of the present invention.

[0042] [Fig.7] The [Fig.7] is a scanning electron microscopy (SEM) image of an optoelectronic device comprising a plurality of 3D structures according to an embodiment of the present invention.

[0043] [Fig.8] The [Fig.8] is a scanning transmission electron microscopy (STEM) image of a 3D structure according to an embodiment of the present invention.

[0044] [Fig.9] Fig.9 is an EDX map of a portion of the image shown in Fig.8, showing the distribution of indium at the top of the lower part of a 3D structure according to an embodiment of the present invention.

[0045] [Fig. 10A] The [Fig. 10A] is an EDX profile established according to the profile A materialized on the image of [Fig.8] showing the presence of quantum wells at the top of a 3D structure according to an embodiment of the present invention.

[0046] [Fig.1OB] The [Fig.1OB] is an EDX profile established according to the profile B materialized on the image of [Fig.8] showing the absence of quantum wells on a flank of a 3D structure according to an embodiment of the present invention.

[0047] [Fig. 11] The [Fig. 11] is a scanning transmission electron microscopy (STEM) image of 3D structures according to an embodiment of the present invention.

[0048] [Fig.12A] Fig.12A shows a scanning transmission electron microscopy (STEM) image of a detail of a 3D structure according to an embodiment of the present invention, and an EDX profile established according to the profile materialized on the STEM image.

[0049] [Fig.12B] [Fig.12B] shows a scanning transmission electron microscopy (STEM) image of a detail of a 3D structure according to an embodiment of the present invention, and an EDX profile established according to profile b materialized on the STEM image.

[0050] [Fig.12C] The [Fig.12C] shows a scanning transmission electron microscopy (STEM) image of a detail of a 3D structure according to an embodiment of the present invention, and an EDX profile established according to the profile c materialized on the STEM image.

[0051] The drawings are given by way of example and are not limiting of the invention. They constitute schematic representations of principle intended to facilitate understanding of the invention and are not necessarily to scale with practical applications. In particular, the dimensions of the various elements of the 3D structures are not necessarily representative of reality. DETAILED DESCRIPTION

[0052] Before beginning a detailed review of embodiments of the invention, it is recalled that the invention according to its first aspect includes in particular the following optional features which may be used in association or alternatively.

[0053] According to one example, each 3D structure has a wire shape and comprises, according to a so-called axial architecture: a lower part extending in a longitudinal direction and comprising a base resting on a substrate, an active region comprising at least one quantum well extending along a plane normal to the longitudinal direction and configured to emit or receive light radiation from at least one quantum well, said active region resting on a vertex, opposite the base, of the lower portion, and an upper part supported by a peak in the active region.

[0054] For example, the process includes: The supply of a substrate comprising at least one surface layer; the formation of a masking layer on the substrate, said masking layer comprising openings through which areas of the surface layer are exposed. A formation, starting from the exposed areas of the surface layer, of lower parts in the form of threads, the bases of said lower parts bearing on the surface layer of the substrate, through the openings, Organometallic precursor vapor phase epitaxial (MOVPE) formation of active regions on the apexes of the lower parts, Formation of upper parts on the apexes of the active regions.

[0055] According to an example, the MOVPE formation of active regions includes an axial deposition occurring along the longitudinal direction and a radial deposition occurring along a direction normal to the longitudinal direction, and a thickness of the radial deposit is less than or equal to 10% of a thickness of the axial deposit.

[0056] According to one example, the method is configured so that the lower part has sides without an active region.

[0057] According to one example, the formation of the lower parts is configured so that the vertices of the lower parts of the 3D structures are separated from each other by a separation distance ds of less than 180 nm, preferably less than 150 nm.

[0058] According to one example, the separation distance ds is less than or equal to 100 nm.

[0059] According to one example, the lower parts of the 3D structures are distributed within the plurality so as to present a surface density greater than or equal to 4 pm2.

[0060] According to one example, the formation of the masking layer is configured so that the openings of the masking layer are spaced at a distance d less than 700 nm.

[0061] According to one example, the formation of the masking layer is configured so that the openings of the masking layer are distributed in such a way as to present a surface density greater than or equal to 4 pm2.

[0062] According to one example, the method is configured so that the active region extends only from the top of the lower part.

[0063] According to one example, the lower and upper parts are chosen to be based on GaN and the active region is chosen to be based on InGaN.

[0064] According to one example, the active region includes at least one InGaN-based quantum well.

[0065] According to one example, the active region comprises an InGaN layer with a thickness greater than 5nm.

[0066] According to one example, the active region comprises an array of InGaN-based quantum dots.

[0067] According to one example, at least one InGaN-based quantum well is formed by MOVPE at a temperature greater than or equal to 700°C.

[0068] According to one example, the active region extends transversely to the longitudinal direction z.

[0069] According to one example, the surface layer has a thickness between 50 nm and 200 nm.

[0070] According to one example, the 3D structures are entirely formed by MOVPE organometallic precursor vapor phase epitaxy.

[0071] According to one example, the openings in the masking layer are regularly spaced by a step less than or equal to 700 nm.

[0072] According to one example, the formation of the active region is configured so that the InGaN-based quantum wells exhibit an indium [In] content > 10%at.

[0073] The invention according to its second aspect includes, in particular, the following optional features which may be used in combination or alternatively:

[0074] According to one example, the principal wavelength X of the light radiation is greater than or equal to 400 nm and / or is less than or equal to 700 nm.

[0075] According to one example, the main wavelength X of the light radiation is between 500 nm and 650 nm.

[0076] According to one example, the lower part of the wire has a height greater than or equal to 150 nm.

[0077] According to one example, the lower part of the wire has a diameter greater than or equal to 30 nm and / or less than or equal to 500 nm.

[0078] According to one example, the vertices of the lower parts of the 3D structures are separated from each other by a separation distance ds less than or equal to 100 nm.

[0079] According to one example, the bases have a diameter less than a diameter of the vertices of the lower parts, preferably at least 10% less.

[0080] According to one example, the lower parts have a section that increases sharply between an area within the openings and an area outside the openings.

[0081] According to one example, preferably for each 3D structure, the active region extends only from the top of the lower part.

[0082] According to one example, preferably for each 3D structure, the active region has a truncated pyramidal shape.

[0083] According to one example, preferably for each 3D structure, the active region in the form of a truncated pyramid comprises inclined faces at an angle of approximately 30° with respect to the longitudinal direction, these inclined faces corresponding substantially to semi-polar planes of type {10-11}.

[0084] According to one example, preferably for each 3D structure, the top of the lower part of the wire is surrounded by a collar.

[0085] According to one example, preferably for each 3D structure, the collar has continuity with the inclined faces of the truncated pyramid forming the active region.

[0086] According to one example, InGaN-based quantum wells have an indium [In] content > 10%.

[0087] Except in cases of incompatibility, it is understood that the manufacturing process and the optoelectronic device may include, mutatis mutandis, any of the above optional features.

[0088] In the present invention, the method for manufacturing axial 3D structures is in particular dedicated to the manufacture of 3D LEDs.

[0089] The invention can be implemented more widely for various 3D structure optoelectronic devices, and in particular those comprising active regions.

[0090] The active region of a 3D structure of an optoelectronic device is understood to be the region from which the majority of the light radiation supplied by this structure is emitted, or the region from which the majority of the light radiation received by this structure is captured.

[0091] The invention can therefore also be implemented in the context of laser or photovoltaic devices.

[0092] Unless explicitly stated otherwise, it is specified that, within the framework of the present invention, the relative disposition of a third layer intercalated between a first layer and a second layer does not necessarily mean that the layers are directly in contact with each other, but means that the third layer is either directly in contact with the first and second layers, or separated from them by at least one other layer or at least one other element.

[0093] The formation steps of the different elements are understood in a broad sense: they can be carried out in several sub-steps which are not necessarily strictly successive.

[0094] The diameter is understood to be the largest transverse dimension at the point considered in the 3D structure (for example, the diameter of the base of the lower part, the diameter of the apex of the lower part, the diameter of the collar, the diameter of the apex of the truncated pyramid). In the present invention, the 3D structures do not necessarily have a circular cross-section. In particular, in the case of GaN-based 3D structures, this cross-section may be hexagonal. The diameter then corresponds to the distance between two opposite vertices of the hexagonal section. Alternatively, the diameter may correspond to an average diameter calculated from the diameter of a circle inscribed in the polygon of the cross-section and the diameter of a circumcircle of this polygon.

[0095] The 3D structures may have a hexagonal or polygonal cross-section. The active regions in the form of truncated pyramids rest on one end of the lower parts of the 3D structures. The truncated pyramids preferably have the same hexagonal or polygonal base as these lower parts.

[0096] The term "wire" refers to a 3D structure elongated along its longitudinal direction. The longitudinal dimension of the wire, along the z-axis in the figures, is greater, and preferably much greater, than the transverse dimensions of the wire in the xy-plane in the figures. For example, the longitudinal dimension is at least five times, and preferably at least ten times, greater than the transverse dimensions.

[0097] The surface density of 3D structures depends on the separation distance ds between the vertices and lower parts of two adjacent 3D structures. In particular, it can be inversely proportional to this distance ds according to k / ds², where k is a proportionality factor. This density is expressed in pm², i.e., the number of 3D structures per square micrometer.

[0098] In this patent application, the terms "concentration", "rate" and "content" are synonymous.

[0099] More particularly, a concentration can be expressed in relative units such as mole or atomic fractions (%at), or in absolute units such as the number of atoms per cubic centimeter (at.cm3).

[0100] In the following, concentrations are atomic fractions expressed in %at, unless otherwise stated.

[0101] In this patent application, the terms "light-emitting diode", "LED" or simply "diode" are used synonymously. An "LED" may also be understood to mean a "micro-LED".

[0102] In the following, the following abbreviations relating to a material M are possibly- actually used:

[0103] Mi refers to intrinsic or unintentionally doped material M, according to the terminology usually used in the field of microelectronics for the suffix -i.

[0104] Mn refers to the N, N+ or N++ doped material M, according to the terminology usually used in the field of microelectronics for the

[0105] suffix -n.

[0106] Mp refers to the P-doped, P+ or P++ material M, according to the terminology usually used in the field of microelectronics for the

[0107] suffix -p.

[0108] A substrate, layer, or device "based on" a material M is understood to mean a substrate, layer, or device comprising only that material M or that material M and optionally other materials, for example, alloying elements, impurities, or dopant elements. Thus, a 3D structure based on gallium nitride (GaN) may, for example, comprise gallium nitride (GaN or GaN-i) or doped gallium nitride (GaN-p, GaN-n). An active region based on gallium-indium nitride (InGaN) may, for example, comprise gallium-aluminum nitride (AlGaN) or gallium nitride with varying aluminum and indium content (GalnAIN). In the context of the present invention, the material M is generally crystalline.

[0109] A coordinate system, preferably orthonormal, comprising the x, y, z axes is shown in the attached figures.

[0110] In this patent application, the terms thickness for a layer and height for a structure or device will be preferred. The thickness is taken along a direction normal to the principal extension plane of the layer, and the height is taken perpendicular to the basal xy plane of the substrate. Thus, a surface layer typically has a thickness along z, and a wire has a height along z. The thickness of an axial deposit is taken along z, and the thickness of a radial deposit is taken in the xy plane.

[0111] Dimensional values ​​are understood to be within manufacturing and measurement tolerances. Thus, two separation distances ds that are identical in theory or two wire diameters that are identical in practice may exhibit slight dimensional variation.

[0112] The terms "approximately", "about", and "on the order of" mean, when referring to a value, "within 10%" of that value, or, when referring to an angular orientation, "within 10°" of that orientation. Thus, a direction substantially normal to a plane means a direction having an angle of 90+10° with respect to the plane.

[0113] To determine the geometry of the 3D structures and the compositions of the different elements (wire, active region, collar for example) of these 3D structures, one can carry out analyses of Scanning Electron Microscopy (SEM) or Transmission Electron Microscopy (TEM or TEM for the English acronym for "Transmission Electron Microscopy") or even Scanning Transmission Electron Microscopy (STEM) (English acronym for "Scanning Transmission Electron Microscopy").

[0114] TEM or STEM are particularly well suited to the observation and identification of quantum wells – whose thickness is generally on the order of a few nanometers – in the active region. Various techniques, listed below in a non-exhaustive manner, can be implemented: dark-field and bright-field imaging, weak beam imaging, and high-angle annular dark field (HAADF) diffraction.

[0115] The chemical compositions of the different elements can be determined using the well-known EDX or X-EDS method, an acronym for "energy dispersive x-ray spectroscopy" which means "energy dispersive analysis of X-ray photons".

[0116] This method is well suited for analyzing the composition of small optoelectronic devices such as 3D LEDs. It can be implemented on metallurgical sections within a Scanning Electron Microscope (SEM) or on thin sections within a Transmission Electron Microscope (TEM).

[0117] The optical properties of the different elements, and in particular the principal emission wavelengths of the axial 3D LEDs based on GaN and / or the active regions based on InGaN, can be determined by spectroscopy.

[0118] Cathodoluminescence (CL) and photoluminescence (PL) spectroscopies are well suited to optically characterize the 3D structures described in the present invention.

[0119] The techniques mentioned above make it possible, in particular, to determine whether an optoelectronic device with an axial 3D wire structure includes InGaN-based quantum wells formed at the top of a GaN-based wire, and a masking layer indicative of the implementation of a MOVPE-type deposition, as described in the present invention. The possible presence of a collar is also easily observable using these techniques.

[0120] The method according to the invention can be implemented using two main approaches. A first approach consists of forming the lower parts of the 3D structures, the active regions, and preferably the upper parts, directly by successive growths, preferably by successive MOVPE growths. This first approach is called "bottom-up" according to the terminology generally used, since the 3D structures are formed from the bottom up.

[0121] A second approach consists of forming the lower parts of the 3D structures from a pre-existing 2D layer by etching this 2D layer. The active regions, and preferably the upper parts, are then formed by successive growths, preferably by successive MOVPE growths. This second approach is called "top-down" according to the terminology generally used, since the lower parts of the 3D structures are formed from top to bottom.

[0122] The following examples fall within the framework of the "bottom-up" approach.

[0123] A first example of the realization of 3D structures according to the invention will now be described with reference to Figures 3, 4 and 5A, 5B.

[0124] Fig. 3 illustrates a plurality of adjacent 3D structures 2 distributed on the same substrate 1. The following description, relating to one of these 3D structures, naturally extends to the other 3D structures of this plurality, which are deemed to be substantially identical to each other.

[0125] The 3D structure 2 comprises at least a lower part 21 in the form of a wire and an active region 22 at the top of this lower part 21. It is preferably formed directly from the substrate 1. This substrate 1 can be in the form of a stack comprising, along the z direction, a support 10, a surface layer 11 called the nucleation layer, and a masking layer 12. The substrate 1 is substantially planar and parallel to the xy plane.

[0126] The support 10 can be made of sapphire to limit lattice parameter mismatch with GaN, or of silicon to reduce costs and address technological compatibility issues. In the latter case, it can be in the form of a wafer with a diameter of 200 mm or 300 mm. It serves, in particular, as a support for 3D structures.

[0127] The nucleation layer 11 is preferably based on AIN. It can alternatively be based on other metal nitrides, for example GaN or AlGaN. It can be formed on the silicon support 10 by epitaxy, preferably by metal-organic vapor phase epitaxy (MOVPE). As an example, the nucleation layer 11 has a thickness between 1 nm and 10 pm. Preferably, it has a thickness on the order of a few hundred nanometers, for example about 100 nm or 200 nm, to a few microns, for example on the order of 2 pm. It can also have a thickness less than 100 nm. This limits the mechanical stresses induced by this layer 11 on the support 10. This prevents detrimental curvature of the support 10. Such a thickness also helps to limit the appearance of structural defects in the nucleation layer 11.In particular, the growth of this nucleation layer 11 can be pseudomorphic, that is to say that the epitaxial constraints (related in particular to the difference in lattice parameters between Si and AlN, . GaN or AlGaN) can be elastically released during growth. The crystalline quality of this nucleation layer 11 can thus be optimized.

[0128] The masking layer 12 is preferably made of a dielectric material, for example silicon nitride Si3N4. It can be deposited by chemical vapor deposition (CVD) onto the nucleation layer 11. It partially masks the nucleation layer 11 and includes openings 120, preferably circular, exposing areas of the nucleation layer 11. These openings 120 typically have a dimension, for example a diameter <e>or an average diameter, between 30 nm and 500 nm. The apertures 120 can be distributed regularly within the masking layer 12, for example in the form of an ordered array. The pitch d, i.e., the distance separating the centers of two adjacent apertures 120, is preferably less than or equal to 700 nm. It can be between 50 nm and 650 nm. The apertures 120 advantageously have a surface density greater than 4 pm². This ultimately allows for densely distributed 3D structures on the substrate 1. These apertures 120 can be produced, for example, by UV or DUV lithography (Deep UV), by electron beam lithography, or by NIL (Nanoinprint lithography). Such a masking layer 12 allows localized growth of a 3D structure at the level of each opening 120.In particular, during a preliminary growth stage called germination, a GaN-based seed forms at the opening 120 and then fills said opening 120. Subsequent growth of the lower part of the 3D structure then occurs from this seed in a localized manner. This seed thus forms the base 210 of the lower part 21 of the 3D structure. The lower part 21 is supported by the nucleation layer 11 of the substrate 1 via its base 210.

[0129] The lower part 21 of the 3D structure is preferably based on GaN, in particular based on GaN-n. It is preferably oriented parallel to z along a crystallographic direction

[0001] corresponding to the c-axis of a hexagonal crystallographic structure.

[0130] The formation of this lower GaN-based part 21 can be carried out by epitaxy, preferably by metal-organic vapor phase epitaxy (MOVPE), in particular as defined in publication WO2012136665. The gallium source in the form of an organometallic precursor (precursor III) can typically be trimethylgallium (TMGa) or triethylgallium (TEGa). The nitrogen source can typically be ammonia (NH3) or nitrogen (N2) (precursor V). The growth temperature is preferably above 700°C, for example on the order of 1000°C. The gas pressure within the growth reactor is, for example, on the order of 425 Torr. The Growth preferably occurs under a neutral and / or reducing atmosphere, typically by adding nitrogen N2 and / or dihydrogen H2. The fluxes of the different gases can be adapted in a way known to those skilled in the art, depending in particular on the volume of the reactor.

[0131] The formation of the lower part 21 can alternatively be carried out by HVPE (Hydride Vapor Phase Epitaxy) gaseous precursor vapor epitaxy, by CVD and MOCVD (MetalOrganic Chemical Vapor Deposition).

[0132] Optionally, conventional surface preparation steps for the germ 210 (chemical cleaning, heat treatment) can be carried out prior to the epitaxial growth of the lower part 21.

[0133] The lower portion 21 of the 3D structure 2 may include an N-doped GaN-based region (GaN-n). This N-doped region can be obtained by growth, implantation, and / or activation annealing. N-doping can, in particular, be achieved directly during growth from a silicon or germanium source, for example, by adding silane, disilane, or germanium vapor. The growth conditions required for the formation of such a lower portion 21 are widely known.

[0134] The lower part 21 of the 3D structure 2 has a diameter $ greater than or equal to 30 nm and / or less than or equal to 500 nm. The cross-section of the lower part 21, in the xy plane, can typically have a more or less regular hexagonal shape. The diameter $ can in this case be an average diameter. The diameter $ can be greater than the diameter ¢0 of the opening 120 and the base 210 that gave rise to the lower part 21. The cross-section can thus exhibit an abrupt change in diameter between its base 210 contained within the opening 120 and its main part outside the opening 120, called the main lower part 21.

[0135] The base 210 of the lower part 21 of the 3D structure 2 is enclosed by the masking layer 12. The main lower part 21 can also bear on the masking layer 12. This thus makes it possible to mechanically reinforce the 3D structure 2. This mechanical reinforcement is all the more important as the aspect ratio of the 3D structure increases.

[0136] The main lower part 21 has, in particular, a height h greater than or equal to 100 nm, preferably greater than or equal to 200 nm. The main lower part 21 preferably has an aspect ratio h / <e>greater than 1, and preferably greater than 5. This improves the crystalline quality at the apex 211 of this lower part 21. It also allows the apex 211 to be moved further away from the underlying plane substrate 1. The local environment at the apex 211 is thus not disturbed by the underlying plane substrate 1. The apex 211 of the part lower 21 is preferably substantially flat and parallel to the xy plane, so as to accommodate the active region 22.

[0137] In the case of an LED, the active region 22 can typically comprise a plurality of quantum wells configured to emit light radiation along a principal wavelength X. These quantum wells are, for example, based on InGaN. They can conventionally be separated from each other by barriers based on AlGaN.

[0138] The active region 22 is preferably oriented along the same crystallographic direction as the lower portion 21. In the example illustrated by Figures 4, 5A, and 5B, it includes, in particular, the InGaN-based quantum wells 220 visible in the STEM HAADF image of [Fig. 4], and more specifically in the STEM HAADF image of [Fig. 5A], which is an enlargement of the area labeled a in [Fig. 4]. The active region 22 may include at least one quantum well 220. The number of quantum wells 220 in the active region 22 may be between 1 and 20. The quantum wells 220 preferably extend along xy planes and are typically separated along z by quantum barriers 222 ([Fig. 5A]). Alternatively, the active region 22 can be in the form of an InGaN-based layer with a thickness greater than or equal to 5 nm, for example 30 nm. Alternatively, the active region 22 can comprise an array of InGaN-based quantum dots.

[0139] The formation of this active region 22 is preferably achieved by metal-organic vapor phase epitaxy (MOVPE). This includes, in particular, CVD epitaxy techniques, such as high-velocity vapor phase epitaxy (HVPE). The growth conditions required for the formation of the active region 22 differ from those required for the formation of the lower portion 21. In particular, an indium source in the form of an organometallic precursor, for example trimethylindium (TMIn) or triethylindium (TEIn), is added to the gallium (TEGa) and / or trimethylgallium (TMGa) and nitrogen (NH3) sources to grow the InGaN-based quantum wells 220. The ratio of the indium precursor element (TMIn, TEIn) to all the precursors III (TEGa, TMGa and TMIn...) can be on the order of 0.3. The growth temperature can be on the order of 800°C.The gas pressure within the growth reactor is, for example, on the order of 100 Torr. The V / III or In / III ratio, the pressure, and the growth temperature can be adjusted according to the design of the epitaxial reactor and the target emission wavelength. The formation of the lower part 21 and the formation of the active region 22 can advantageously take place in a single reactor or growth frame.

[0140] The immediate environment at the apex 211 of the lower part 21 can influence the growth morphology. In particular, the proximity of other apexes 211 Adjacent regions can locally modify the growth conditions of the active region 22, and in particular of the InGaN-based quantum wells 220. During the development of the present invention, it became apparent that a high density of lower parts 21 in the form of wires, particularly greater than 4 pm², promotes the growth of the quantum wells 220 along xy planes. Thus, the active region 22 formed by MOVPE is confined to the apex 211 of the lower part 21 of the 3D structure. The flanks 212 of the lower part 21 can therefore be at least partially devoid of the active region 22, in particular of the InGaN-based quantum wells, as shown in the STEM HAADF image in [Fig. 5B]. An axial 3D structure is thus obtained by MOVPE.

[0141] The distance ds between the vertices 211 of two adjacent lower parts 21 is a parameter influencing the morphology of the MOVPE deposit allowing the formation of the active regions 22.The distance ds typically corresponds to the minimum separation between the edges of the vertices 211 of the two adjacent lower parts 21 concerned. In particular, this MOVPE deposit may comprise a proportion of axial deposition, i.e., along the z-axis, and a proportion of radial deposition, i.e., along a direction normal to z. These proportions vary as a function of the distance ds, as illustrated in [Fig. 6]. This [Fig. 6] shows that the experimental curve 6, which plots the ratio of radial to axial deposit thicknesses, decreases sharply and unexpectedly for distances ds less than or equal to 200 nm. In particular, for a distance ds less than 100 nm, on the order of 90 nm, the experimental point 61 corresponds to a radial to axial deposit thickness ratio of less than 10%.Axial deposition refers to an increase in thickness along the z-axis; radial deposition refers to an increase in thickness along a direction perpendicular to the previous one, particularly at the flanks of the elements protruding from the substrate. The deposition is thus predominantly axial for this distance ds between vertices 211, on the order of 90 nm. This behavior deviates from the expected linear behavior represented by curve 60 extrapolated from experimental points 62, 63, 64. The proximity of the vertices 211 to the lower parts 21 of the 3D structures clearly and surprisingly favors axial growth of the active regions 22.

[0142] Figures 7 and 8 are further electron microscopy images of these 2-axial 3D structures obtained by MOVPE.

[0143] Fig. 7 is a SEM image showing dense 3D structures, regularly arranged on a substrate (not visible) via a masking layer (not visible). In this example, the 3D structures have a diameter of approximately 200 nm (for an aperture diameter of 50 nm) and a surface density on the order of 10 pm².

[0144] Fig. 8 is a STEM HAADF image of one axial 3D structure among the plurality of 3D structures illustrated in Fig. 7, viewed in cross-section. The base 210 is embedded in a opening of the masking layer 12, the main lower part 21 and the active region 22 comprising 3 quantum wells 220 are clearly visible.

[0145] The active region 22 rests on the vertex 211 of the lower part 21 and has a vertex 221 suitable for receiving an upper part 23 (not shown in the 3D structure of [Fig. 8]). The active region 22 typically has the shape of a truncated pyramid ([Fig. 8]). This truncated pyramid typically has a more or less regular hexagonal cross-section in the xy plane. The active region 22 in the form of a truncated pyramid thus comprises inclined flanks or faces 224 extending from the vertex 211 of the lower part 21 to the vertex 221. In particular, there may be six of these faces 224. These faces 224 may be inclined at an angle of approximately 60° to the xy plane. Such faces 224 may, in particular, correspond to planes of the type {10-11}. According to another possibility, the faces 224 can be inclined at an angle of approximately 80° with respect to the xy plane.Such an inclination of the faces 224 coincides approximately with semi-polar planes of type {20-21}.

[0146] The active region 22, in the form of a truncated pyramid, can extend below the apex 211 of the lower portion 21, for example, in the form of a collar 223 ([Fig. 8]). This collar 223 may include facets extending from the faces 224 of the active region 22. This collar 223 typically forms, together with the truncated pyramid portion of the active region 22, a cap covering the apex 211 of the lower portion 21. The collar 223 allows, for example, improved mechanical cohesion between the lower portion 21 and the active region 22. The collar 223 may have a significant height, on the order of one-third or one-half of the height of the truncated pyramid, for example. It can also extend towards the base 210 of the lower part 21, in the form of a thin layer of a few nanometers, for example on the order of 1 to 5 nm, covering the flanks 212 of the lower part 21.The collar 223 is not necessarily continuous with the truncated pyramid of the active region 22. It may be independent of it.

[0147] The distribution of indium in the active region 22 is localized at the quantum wells 220, as shown in the EDX mapping of the indium element presented in [Fig. 9]. The incorporation of indium at the collar 223 and the flanks 212 of the lower part 21 is low or even zero ([Fig. 9]).

[0148] Figures 10A and 10B confirm this distribution of indium within the raw axial 3D structure by MOVPE.

[0149] Figure 1OA shows an EDX profile acquired at the active region 22, along profile A shown in Figure 8. This profile allows the atomic fractions (%at) of the different chemical elements (O, Si, In, Ga) present in the profile acquisition zone to be obtained. In Figure 1OA, the first part, "SiO2", of the profile This corresponds to the silicon dioxide protective layer deposited on vertex 221 of the 3D structure during the preparation of a sample observable by STEM. The second part, "InGaN QW," of the profile corresponds to the active region 22 of the 3D structure. Three indium peaks, corresponding to three InGaN quantum wells in this active region 22, are clearly detected. The third part, "GaN," of the profile corresponds to the lower part 21 of the 3D structure, below vertex 211. Indium is not detected. Therefore, indium has not diffused into the lower part 21.

[0150] Figure 1OB shows an EDX profile acquired at a flank 212 of the lower part 21, along profile B shown in Figure 8. This profile allows the atomic fractions (%at) of the different chemical elements present in the profile acquisition area to be obtained. In Figure 1OB, the first part, "SiO2," of the profile corresponds to the silicon dioxide protective layer deposited on the flanks 212 of the 3D structure during the preparation of a sample observable by STEM. The second part, "In free GaN," of the profile corresponds to the lower part 21 of the 3D structure, at a flank 212. Indium is not detected. Indium has therefore not diffused onto the flanks 212 of the lower part 21. The 3D structure obtained by MOVPE thus clearly exhibits an axial architecture.

[0151] In the case of an optoelectronic device, for example an LED, the active region 22 is typically surmounted by an upper portion 23 based on GaN, in particular based on P-doped GaN. This upper portion 23 typically covers the active region 22 and allows carriers to be injected into the active region 22. The growth of the upper portion 23 on the active region 22 is preferably carried out by MOVPE. The thickness of the deposit used to form this upper portion 23 is preferably limited to a few tens of nanometers, for example less than 100 nm, or even less than 50 nm, so as to limit the reabsorption of the light emitted by the active region 22.

[0152] Fig. 11 illustrates raw axial 3D structures by MOVPE each comprising a lower part 21, an active region 22, and an upper part 23, 23b.

[0153] This upper part 23 can extend only over the summit 221 of the active region 22. Alternatively, the upper part 23, 23b extends partly over the summit 221 of the active region 22, and partly over the flanks of the active region 22 and over the flanks 212 of the lower part 21 below the summit 211, as illustrated in [Fig. 11].

[0154] Figures 12A, 12B and 12C present different EDX profiles acquired on the axial 3D structures illustrated in [Fig. 11].

[0155] Figure 12A presents a STEM image of a 3D structure and an EDX profile acquired along the profile materialized on said STEM image, through the upper part 23 and the active region 22 of the 3D structure. This profile allows obtaining the fractions atomic percentages (%at) of the different chemical elements present in the profile acquisition zone. In [Fig. 12A], the first part, "protective coating," of the profile corresponds to the protective layer deposited around the 3D structure during the preparation of a sample observable by STEM. The second part, "AlGaN," of the profile corresponds to the upper part 23 of the 3D structure. The third part, "InGaN," of the profile corresponds to the active region 22 of the 3D structure. The fourth part, "GaN," of the profile corresponds to the lower part 21 of the 3D structure, below the apex 211.

[0156] Figure 12B shows a STEM image of a 3D structure and an EDX profile acquired along profile b, materialized on said STEM image, through the upper part 23b of the 3D structure. This profile allows the atomic fractions (%at) of the different chemical elements present in the profile acquisition area to be obtained. In Figure 12B, the "protective coating" portion of the profile corresponds to the protective layer deposited around the 3D structure during the preparation of a sample observable by STEM. The "GaN" and "AlGaN" portions of the profile correspond to the upper part 23b of the 3D structure. The "InGaN" portion of the profile corresponds to the detection of traces of indium between the upper part 23b and the lower part 21 of the 3D structure. The "GaN" part of the profile corresponds to the lower part 21 of the 3D structure, below the vertex 211.

[0157] Figure 12C shows a STEM image of a 3D structure and an EDX profile acquired along the profile c materialized on said STEM image, through a flank 212 of the lower part 21 of the 3D structure. This profile allows the atomic fractions (%at) of the different chemical elements present in the profile acquisition area to be obtained. In Figure 12C, the "protective coating" portion of the profile corresponds to the protective layer deposited around the 3D structure during the preparation of a sample observable by STEM. The "GaN (In free) (Al free)" portion of the profile corresponds to the lower part 21 of the 3D structure. The presence of indium and aluminum is not detected. Indium and aluminum therefore did not diffuse along the flanks 212 of the lower part 21. This 3D structure obtained by MOVPE according to this second example also presents an axial architecture.

[0158] The following examples are based on the "top-down" approach.

[0159] According to this "top-down" approach, the lower parts 21 can be obtained from a previously formed two-dimensional (2D) layer, for example, a GaN-based layer. In a known manner, dense patterns substantially defining, in z-projection, the lower parts 21 are formed by lithography on the 2D layer. Etching this 2D layer along z then allows the formation of the plurality of lower parts 21. The lower parts 21 obtained by etching can be in the form of z-oriented wires, with an aspect ratio h / <t>>l, or in the form of mesas, with an aspect ratio h / <e>< 1. The etching is configured here so that the vertices 211 of two adjacent lower parts 21 are separated by a separation distance ds of less than 180 nm. This allows the active region 22 to be formed on the vertices of the lower parts 21 by MOVPE, according to an axial architecture. Thus, according to the invention, the "top-down" approach can be implemented to obtain the lower parts 21. Then, the formation of the active regions 22, and subsequently the upper parts 23, is carried out by MOVPE as in the "bottom-up" approach described above, maintaining a separation distance ds of less than 180 nm between the vertices 211 of two adjacent lower parts 21. This makes it possible to obtain an axial architecture by MOVPE, without creating defects on the flanks or faces 224 of the active regions.Thus, by combining the principle of the invention with a top-down formation of the lower parts, the diameters of the lower parts 21 can be significantly increased. The aspect ratio h / . <e>The lower parts 21 may be less than 1, or even much less than 1. The masking layer can be removed.

[0160] The invention is not limited to the embodiments previously described and extends to all embodiments covered by the claims.< / e> < / e> < / t> < / e> < / e>

Claims

1. Claims Method for manufacturing a plurality of three-dimensional (3D) structures (2) for optoelectronics, each 3D structure comprising, stacked in a longitudinal direction (z): • a lower part (21) comprising a base (210) resting on a substrate (1), said lower part (21) extending in the form of a wire in the longitudinal direction (z) so that each 3D structure has a wire shape, • an active region (22) configured to emit or receive light radiation, said active region (22) resting on a vertex (211), opposite the base (210), of the lower part (21), and • an upper part (23) resting on a vertex (221) of the active region (22), said method comprising: • A supply of a substrate (1) comprising at least one surface layer (11), • A formation of a masking layer (12) on the substrate (1), said masking layer (12) comprising openings (120) through which areas of the surface layer (11) are exposed, • A formation, from the exposed areas of the surface layer (11), of the lower parts in the form of wires (21), the bases (210) of said lower parts bearing on the surface layer (11) of the substrate (1), through the openings (120), said lower parts (21) having distinct vertices (211) such that the vertices (211) of two adjacent lower parts (21) are separated from each other by a separation distance ds less than or equal to 100 nm, • A formation by organic metal vapor phase epitaxy (MOVPE) of the active regions on the tops (211) of the lower parts (21), said formation by MOVPE of the active regions (22) comprising a so-called axial deposition taking place in the longitudinal direction (z) and a so-called radial deposition taking place in a normal direction (x, y) to the longitudinal direction (z), so that a thickness of the radial deposit is less than or equal to 10% of a thickness of the axial deposit, • A formation of the upper parts (23) on the peaks (221) of the active regions (22).

2. Method according to the preceding claim configured so that the lower part (21) has sides (212) devoid of active region (22).

3. A method according to any preceding claim wherein the lower portions (21) are distributed within the plurality so as to have a surface density greater than or equal to 4 pm2.

4. A method according to any preceding claim, configured so that the active region (22) extends only from the top (211) of the lower portion (21).

5. Method according to any one of the preceding claims in which the lower parts (21) and the upper parts (23) are chosen to be based on GaN and the active region (22) is chosen to be based on InGaN.

6. Method according to any one of the preceding claims in which the active region (22) comprises at least one quantum well (220) based on InGaN, or an InGaN layer with a thickness greater than 5nm, or a set of quantum dots based on InGaN.

7. Method according to the preceding claim in which the formation of the at least one quantum well (220) based on InGaN is carried out at a temperature greater than or equal to 700°C.

8. A method according to any preceding claim wherein the active region (22) extends transversely to the longitudinal direction (z).

9. Optoelectronic device comprising a plurality of three-dimensional (3D) structures (2) having a wire shape and each comprising: • a lower part (21) in the form of a wire extending in a longitudinal direction (z) and comprising a base (210) bearing on a substrate (1), • an active region (22) configured to emit or receive light radiation, said active region (22) bearing on a summit (211), opposite the base (210), of the lower part (21), and • an upper part (23) resting on a vertex (221) of the active region (22), • a masking layer (12) in contact with a surface layer (11) of the substrate (1), said masking layer (12) comprising openings (120) through which the lower parts (21) in the form of wire extend, the bases (210) of said lower parts (21) bearing on the surface layer (11) of the substrate (1), in which two vertices (211) of two adjacent lower parts (21) are separated from each other by a separation distance ds less than or equal to 100 nm, the device being characterized in that the lower parts (21) have a section increasing abruptly between an area included in the openings (120) and an area located outside the openings (120).

10. Device according to the preceding claim in which the active region (22) extends only from the top (211) of the lower part (21).

11. Device according to any one of claims 9 to 10 in which the active region (22) has a truncated pyramidal shape.

Images