GALLIUM DENITRIDE LAYER DONOR SUBSTRATE

The donor substrate with InGaN and AlGaN superlattices confines defects, addressing the inefficiencies of high hydrogen doses in the Smart Cut™ process, enabling cost-effective and high-quality GaN layer transfer for industrial applications.

FR3170822A1Pending Publication Date: 2026-06-26SOITEC SA +1

Patent Information

Authority / Receiving Office
FR · FR
Patent Type
Applications
Current Assignee / Owner
SOITEC SA
Filing Date
2024-12-20
Publication Date
2026-06-26

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

The present invention relates to a donor substrate for the transfer of a GaN layer, comprising a support substrate (10, 60) and a transfer stack (10) extending over the support substrate, said transfer stack (50) comprising, from its base to its surface: a first super-lattice for confining defects in the crystal lattice (20), a sacrificial layer (30) of GaN, a second super-lattice for confining defects in the crystal lattice (40), and a layer of GaN to be transferred (50), each super-lattice for confining defects (20, 40) comprising a plurality of alternations (25, 45) of a primary layer (21, 41) of InGaN and a secondary layer (22, 42) of AlGaN. Figure for the abstract: Fig. 1
Need to check novelty before this filing date? Find Prior Art

Description

Title of the invention: DONOR SUBSTRATE FOR THE TRANSFER OF A GALLIUM DENITRIDE LAYER FIELD OF INVENTION

[0001] The present invention relates to a donor substrate for the transfer of gallium nitride (GaN) layers. The invention also relates to a method for manufacturing such a donor substrate. STATE OF THE ART

[0002] Smart Cut™ technology enables the transfer of a layer from a donor substrate to a recipient substrate to form multilayer substrates, typically for the manufacture of electronic components. This process involves implanting atomic species into a donor substrate to create a zone of embrittlement at a predetermined depth, with the aim of detaching a thin layer delimited by said zone of embrittlement.

[0003] One objective of the transfer of the layer delimited by the embrittlement zone is to preserve the crystalline lattice of the transferred layer. The quality of this crystalline lattice is an important criterion for the manufacture of electronic components.

[0004] The atomic species create defects in the crystal lattice of the donor substrate, the set of defects forming the embrittlement zone. After the donor substrate is bonded to the recipient substrate, a heat treatment causes the donor substrate to fracture at the embrittlement zone, thus allowing the detachment of said thin layer bonded to the recipient substrate.

[0005] For the detachment of the thin film, it is important that the defects are distributed in the thickness of the substrate according to a regular distribution and that the area of ​​embrittlement extends substantially in a plane parallel to the main face of the donor substrate.

[0006] For the fabrication of certain components, particularly in optoelectronics, power electronics and radio frequency applications, gallium nitride (GaN) layers are used. The layer transfer and consequently the implantation of atomic species, typically hydrogen atoms, is preferably carried out in the (001) plane of the GaN crystal.

[0007] The defects created by the implantation of hydrogen atoms are generally vacancies propagating through the GaN substrate, particularly in the case of implantation in the (001) plane. A vacancy is an unoccupied atomic site in the crystal structure. Unlike the defects created by hydrogen implantation In other types of semiconductors, vacancies readily disperse towards the surface or into the interior of the GaN substrate before forming pyramidal defects in the crystal lattice. Vacancy mobility is a limiting factor for the preferential formation of pyramidal defects in a net embrittlement zone.

[0008] In order to create a weakened zone allowing for controlled fracture, it is currently necessary to implant high doses of hydrogen due to the dispersion of defects. However, this involves long irradiation times and significant energy consumption, which has a major impact on the cost and environmental sustainability of the process. This makes the transfer of GaN layers by the Smart Cut™ type process poorly suited to industrial-scale manufacturing.

[0009] Cherkashin et al. proposed confining the defects created by hydrogen atoms between two stacks of thin films of indium gallium nitride (InGaN) and GaN. The InGaN layers promote defect confinement due to the different lattice parameters of InGaN compared to GaN, resulting in compressive stress on the GaN crystal. This compressive stress has a detrimental effect on the crystalline quality of the GaN layers deposited on such a stack.

[0010] A compromise must therefore be found between preserving the crystal lattice of GaN for the manufacture of electronic components, and confining defects to allow localization of the embrittlement zone. Description of the invention

[0011] An object of the invention is to make available a donor substrate allowing the transfer of a layer of GaN delimited by a zone of embrittlement created by the implantation of a dose of hydrogen atoms lower than the doses used in known processes, while ensuring a good quality of the crystalline lattice of the GaN layer.

[0012] To this end, the invention proposes a donor substrate for the transfer of a GaN layer, comprising a support substrate and a transfer stack extending over the support substrate, said transfer stack comprising, from its base to its surface:

[0013] • a first super-network for confining defects in the crystal lattice,

[0014] • a sacrificial layer of GaN,

[0015] • a second super-network for confining defects in the crystal lattice, and

[0016] • a layer of GaN to be transferred,

[0017] each confinement supernetwork comprising a plurality of alternations of a primary InGaN layer and a secondary AlGaN layer.

[0018] Vacancies cannot easily propagate through InGaN and therefore remain confined within the small volume of available GaN. The InGaN layers of the confinement superlattice thus limit the volume available for vacancy propagation within the crystal lattice, thereby creating a localized embrittlement zone within the donor substrate. The AlGaN layers mitigate the impact of the stresses induced by the presence of InGaN on the crystalline quality of the GaN.

[0019] Preferably, the plurality of alternations comprises between 2 and 20 alternations, preferably between 5 and 15 alternations.

[0020] Advantageously, each primary layer has a thickness of less than 20 nm, preferably less than 10 nm, and more preferably equal to 3 nm.

[0021] Advantageously, each secondary layer has a thickness of less than 20 nm, preferably less than 10 nm, and more preferably equal to 3 mm.

[0022] Preferably, the donor substrate comprises a plurality of superimposed transfer stacks.

[0023] The invention also relates to a method for manufacturing a donor substrate as described above, comprising providing a first support substrate, a GaN seed layer and / or a GaN buffer layer extending over an upper face of said first support substrate, and the formation of a transfer stack comprising successively the following steps:

[0024] (i) the formation of the first super-confinement network on the germ layer or the buffer layer formed by alternating deposition of primary and secondary layers,

[0025] (ii) the formation of the sacrificial GaN layer on the first confinement layer,

[0026] (iii) the formation of the second confinement super-lattice on the sacrificial GaN layer by alternating deposition of primary and secondary layers,

[0027] (iv) the formation of a transferable layer of GaN on the second confinement super-lattice,

[0028] Advantageously, steps (i) to (iv) of formation of a transfer stack are carried out by epitaxial growth.

[0029] Advantageously, wherein the steps (i) to (iv) of transfer stack formation are carried out successively in a single epitaxial chamber.

[0030] In some embodiments, the process further includes, before the step of forming a transfer stack, a deposition of the GaN buffer layer or the transfer of the GaN seed layer onto the first support substrate.

[0031] Preferably, steps (i) to (iv) of forming the transfer stack (500) are repeated.

[0032] In certain embodiments, the process further comprises, after the formation of the transfer stack (500), the following steps:

[0033] (v) the bonding of a front face of the transfer stack (500) onto a second substrate support (60),

[0034] (vi) the removal of the first support substrate, so as to expose a layer of GaN to be transferred.

[0035] The manufacturing process may further include, during step (vi), the removal of the first support substrate is carried out so as to transfer a plurality of transfer stacks onto the second support substrate.

[0036] The invention also relates to a method for transferring a layer of GaN onto a first receiving substrate, comprising the following steps: the implantation of atomic species in the sacrificial GaN layer of a donor substrate as described above to form a zone of embrittlement, the defects created during said implantation being confined in the sacrificial GaN layer between the first confinement superlattice and the second confinement superlattice,

[0037] o the bonding of the transfer stack onto the first receiving substrate,

[0038] o the detachment of the donor substrate along the embrittlement zone.

[0039] Preferably, the transfer process further includes a step of selectively etching a remnant of the sacrificial layer and the second confinement super-lattice onto the first receiving substrate, so as to expose the transferred GaN layer.

[0040] The transfer process may further include, after selective etching, bonding the first receiving substrate to a second receiving substrate via the GaN layer and removing the first receiving substrate, so as to reverse the polarity of the transferred GaN layer.

[0041] The transfer process may further include a step of selectively etching a remnant of the sacrificial layer and the confinement super-lattice on the donor substrate so as to expose a second layer of GaN to be transferred. DESCRIPTION OF THE FIGURES

[0042] Other features and advantages of the invention will become apparent from the detailed description that follows, with reference to the accompanying drawings, in which:

[0043] Fig. 1 is a cross-sectional view of a donor substrate according to the invention.

[0044] Figure [Fig. 2] is a cross-sectional view of a supporting substrate according to a first mode of realization.

[0045] Fig. 3 is a cross-sectional view of a support substrate according to a second embodiment.

[0046] Fig. 4A is a cross-sectional view of a first containment super-network.

[0047] Fig. 4B is a cross-sectional view of a second containment super-network.

[0048] Figure 5A illustrates a donor substrate according to the invention comprising several superimposed transfer stacks.

[0049] Figure 5B illustrates an alternative embodiment of a donor substrate according to the invention comprising several superimposed transfer stacks.

[0050] Fig. 6 illustrates a first step in transferring the stacks onto a pseudo-donor support substrate.

[0051] Fig. 7 illustrates a second step of transferring the stacks onto a pseudo-donor support substrate.

[0052] Figure 8A illustrates a first step of transferring a layer of GaN from a first embodiment of a donor substrate according to the invention to a receiving substrate.

[0053] Fig. 8B illustrates a first step in transferring a layer of GaN from a second embodiment of a donor substrate to a recipient substrate.

[0054] Fig. 9A illustrates a second step of transferring a layer of GaN from a first embodiment of a donor substrate to a recipient substrate.

[0055] Fig. 9B illustrates a second step of transferring a GaN layer from a second embodiment of a donor substrate to a recipient substrate.

[0056] Fig. 10 illustrates a third step of transferring a GaN layer from a donor substrate according to the invention to a recipient substrate.

[0057] Figure 11 illustrates the donor substrate during the third step of transferring a layer of GaN from a donor substrate according to the invention to a recipient substrate.

[0058] Figure 12 illustrates a fourth step in transferring a GaN layer from a donor substrate according to the invention to a recipient substrate.

[0059] Fig. 13A illustrates a first polarity reversal step of a GaN layer transferred onto a receiving substrate.

[0060] Fig. 13B illustrates a second polarity reversal step of a GaN layer transferred onto a receiving substrate.

[0061] Fig. 14 illustrates a donor substrate comprising several transfer stacks prepared for a second transfer of a layer of GaN.

[0062] For reasons of readability of the figures, the illustrated elements are not necessarily represented to scale. DETAILED DESCRIPTION OF THE INVENTION

[0063] Figure 1 illustrates a donor substrate according to the invention. The donor substrate comprises, from its base to its surface, a support substrate 10 and a transfer stack 500 from which a layer of gallium nitride (GaN) can be transferred onto a receiving substrate.

[0064] The support substrate 10 is typically made of silicon (Si), silicon carbide (SiC), aluminum oxide (Al2O3) or GaN.

[0065] With reference to [Fig. 2], the support substrate may consist of a base substrate 11 and a buffer layer 12 extending over a front face of the base substrate 11. The base substrate 11 supports the buffer layer and ensures the mechanical strength of the assembly. By way of illustration and without limitation, the base substrate 11 may, for example, be made of Si, sapphire, SiC, polycrystalline AIN, or be a composite substrate comprising a stack of several of these materials.

[0066] The buffer layer 12 is made of aluminium nitride (AIN) or aluminium gallium nitride (AlGaN) and allows for taking into account the network asymmetries and the thermal expansion constraints of the donor substrate in order to ensure good crystalline quality during the deposition of the layers of the transfer stack 500.

[0067] Alternatively, with reference to [Fig.3], the support substrate may consist of a base substrate 11 and a GaN seed layer 13 extending over a front face of the base substrate 11. The seed layer 13 may, for example, be transferred onto the base substrate 11 by a transfer process from another donor substrate.

[0068] In some cases, the substrate support 10 may consist of a central layer of polycrystalline AIN surrounded by several dielectric layers, and carry a layer such as a silicon layer (111) transferred onto its surface. Such substrates are, for example, manufactured using QST™ technology and marketed by Qromis.

[0069] In other embodiments (not shown), the support substrate is a bulk GaN substrate and does not include any buffer layer and no additional seed layer.

[0070] A GaN layer to be transferred 50 is arranged on the surface of the donor substrate. In order to transfer this GaN layer onto a receiving substrate, it is necessary to form a weakening zone between said GaN layer and the supporting substrate. For this purpose, a sacrificial GaN layer, in which the weakening zone will be created, is arranged between the GaN layer to be transferred and the supporting substrate. Said sacrificial layer is delimited by two confinement superlattices to locate the weakening zone within the sacrificial GaN layer. The GaN layer to be transferred and the sacrificial GaN layer delimited by the confinement superlattices together form a transfer stack 500.

[0071] With reference to [Fig. 1], the transfer stack 500 comprises a first super-lattice 20 for confining crystal lattice defects arranged on the support substrate 10 and a second super-lattice 40 for confining crystal lattice defects. A sacrificial layer 30 of GaN extends between the first confinement super-lattice 20 and the second confinement super-lattice 40. The sacrificial GaN layer 30 may have a thickness between 10 and 500 nm, preferably between 10 and 300 nm, for example 60 nm.

[0072] The GaN layer 50 intended to be transferred onto a receiving substrate is arranged on the front face of the second confinement super-lattice 40. The GaN layer to be transferred typically has a thickness between 100 nm and 2 pm and may have doping depending on the intended application of the layer 50 after the transfer.

[0073] With reference to [Fig.4A] and [Fig.4B], each confinement superlattice 20, 40 comprises a plurality of alternations 25, 45 of a primary layer 21, 41 of indium gallium nitride (InGaN) and a secondary layer 22, 42 of AlGaN.

[0074] Each primary layer 21, 41 is made of InxGal-xN with x between 0.05 and 0.60, preferably between 0.05 and 0.50. A primary layer may, for example, be made of Inl5Ga85N. InxGal-xN generally has slightly larger lattice parameters than pure GaN due to the incorporation of indium, which has a larger atomic radius than gallium. The higher the indium content x, the more the lattice parameters of the InxGal-xN ternary alloy deviate from those of GaN. The parameter x is chosen so as to allow compressive stresses to be exerted on the sacrificial GaN layer 30, while avoiding dislocations in the crystal lattice of the GaN layer to be transferred 50 and the sacrificial GaN layer 30 that can be caused by an excessively high indium content.The parameter x can depend on the thickness of the primary 21 and secondary 4L layers. The primary layer in such a material exerts a compressive stress on the crystal lattice of the GaN layer to be transferred 50 and the sacrificial GaN layer 30. This stress can reduce the propagation of defects in the crystal lattice, in particular vacancies caused by irradiation by hydrogen atoms.

[0075] The presence of the primary layers 21,41 in each confinement superlattice 20, 40 thus makes it possible to confine defects in the crystal lattice in the sacrificial GaN layer 30 arranged between the first and second confinement superlattice 20, 40.

[0076] Each secondary layer 22, 42 is made of AlxGal-xN with x between 0.2 and 1, preferably between 0.2 and 0.6. For example, a secondary layer may be made of A140Ga60N. AlxGal-xN has slightly smaller lattice parameters than GaN due to the incorporation of aluminum, which has a smaller atomic radius than gallium. The lattice parameters of AlxGal-xN are therefore also smaller than those of InxGal-xN. The higher the aluminum content x, the more the lattice parameters of the ternary alloy AlxGal-xN deviate from those of GaN and InxGal-xN.

[0077] The presence of the secondary layers 22, 42 mitigates the impact of the stresses exerted by the primary layers 21, 41 on the GaN crystal lattice. The average lattice parameter of the confinement superlattice is close to that of GaN, which allows GaN to be deposited on the superlattice while avoiding significant stresses on the crystal lattice. This optimizes the crystalline quality of the sacrificial layer 30 and the transfer layer 50, resulting in GaN layers of a quality suitable for use in electronic applications. The average lattice parameter is calculated by taking into account the lattice parameters of each layer and weighting them by the thickness of each layer. This yields an average value that represents the entire confinement superlattice.

[0078] The first and second layers of each confinement superlattice thus act synergistically. The InGaN layer limits vacancy propagation by exerting a compressive stress on the GaN crystal lattice. The AlGaN layer does not significantly affect vacancy propagation. This layer exerts an extensive stress that compensates for the compressive stress of the InGaN and thus readjusts the overall stress exerted on the crystal of each GaN layer. Thanks to this compensating effect, the stresses exerted by the InGaN and AlGaN layers do not, or only minimally, modify the crystal lattice of the GaN layers extending above and below the confinement superlattice. The confinement superlattice thus limits defect propagation and simultaneously preserves the crystalline quality of the GaN layers.

[0079] Each confinement superlattice typically comprises between 2 and 20 alternations 25, 45 of primary layers 21, 41 and secondary layers 22, 42, preferably between 5 and 15 alternations. The number of alternations will be chosen according to the thickness of the primary and secondary layers and the desired crystalline quality of the GaN in order to find a compromise between the number of steps during layer deposition and the efficiency of defect confinement and the preservation of the crystal lattice. The first and second confinement superlattices may have the same or different numbers and thicknesses of layers.

[0080] Typically, with reference to [Fig. 5A], a donor substrate according to the invention comprises a plurality of transfer stacks 500 superimposed on the same support substrate 10. This superposition makes it possible to transfer, from a single donor substrate, a plurality of GaN layers onto different recipient substrates intended to the manufacture of electronic components and thus increase the efficiency of the manufacture of substrates for electronic applications.

[0081] Each transfer stack 500 comprises a first confinement superlattice 20, a sacrificial layer 30 of GaN arranged on the first confinement superlattice, a second confinement superlattice 40 arranged on the sacrificial layer 30, and a layer of GaN to be transferred 50 arranged on the second confinement superlattice 40. The number of transfer stacks depends on the desired number of layers to be transferred and the maximum thickness of the donor substrate. The maximum thickness of the donor substrate will be chosen based on the type and size of the support substrate and its material, particularly in terms of the crystalline lattice and the coefficient of thermal expansion of the support substrate 10 and any buffer layer 12.For example, the total thickness of the transfer stacks can be limited to a value between 1.5 pm and a few hundred pm for a number of transfer stacks between 1 and 40, preferably between 1 and 30.

[0082] Fabrication of the donor substrate

[0083] We will now describe a method for manufacturing a donor substrate as described above.

[0084] The process begins by providing a first support substrate, which can be silicon (Si), silicon carbide (SiC), aluminum oxide (Al₂O₃), or GaN. This first support substrate may include a buffer layer or a seed layer to facilitate the growth of single-crystal GaN on its surface. Optionally, one or more layers can be deposited on the substrate, for example, a GaN layer. Referring to [Fig. 5A], the epitaxial growth of a first super-lattice for confining defects in the crystal lattice is then carried out. This step is performed by alternating epitaxial growth of primary layers 21 of InGaN and secondary layers 22 of AlGaN. In some embodiments, an AlGaN layer is deposited as the final layer of a confinement super-lattice to facilitate the resumption of growth of a GaN layer. A sacrificial layer 30 of GaN is then deposited by epitaxy.On the sacrificial layer 30, a second confinement super-network 40 is created by alternating epitaxial growth of primary layers 41 of InGaN and secondary layers 42 of AlGaN. .

[0085] On the second superlattice, a GaN layer to be transferred 50 is deposited by epitaxy. During this step, the layer to be transferred can be doped according to the intended applications of said layer. The confinement superlattices and the GaN layers are typically deposited successively on the substrate in a single epitaxial chamber to avoid transfer steps that would render The process is longer and more expensive. It also avoids introducing contaminants into the deposition chamber and / or onto the substrate to be manufactured.

[0086] The deposition steps can be repeated in order to deposit a plurality of transfer stacks 500 on the same substrate, each comprising a first confinement superlattice 20, a sacrificial layer 30 of GaN, a second confinement superlattice 40, and a layer of GaN to be transferred 50. This makes it possible to obtain a donor substrate suitable for the transfer of several layers of GaN onto several respective receiving substrates, making the process more economical.

[0087] Gallium nitride crystallizes in a crystalline structure exhibiting intrinsic polarity, which can be Ga-faced, terminated by gallium atoms, or N-faced, terminated by nitrogen atoms, depending on whether the growth direction is

[0001] (Ga polarity) or [000-1] (N polarity). The electrical and electrochemical properties of a GaN layer depend on its polarity on the substrate. Most GaN-based electronic components are manufactured with Ga polarity due to its advantages in terms of crystalline quality and electrical properties.

[0088] During epitaxial deposition, a Ga-face polarity is typically preferred due to the better crystalline quality that can be obtained in this polarity. During the transfer of a layer onto a receiving substrate, the GaN layer is inverted and therefore exhibits an N-face polarity on the receiving substrate. It is therefore necessary for most electronic components to perform a second transfer step to invert the layer's polarity a second time. Such a double transfer makes it possible to obtain a Ga-face polarity on the receiving substrate on which the electronic components will be fabricated.

[0089] In this case, with reference to [Fig. 5B], one or more transfer stacks are deposited on the first support substrate 10 in the reverse order of that described above to form a temporary substrate 100. Optionally, a demounting layer (not shown) may be deposited before the transfer stacks are deposited. For example, a transfer layer 50 of GaN is deposited first, which may include doping depending on the electronic application for which the GaN layer is intended. The first and second confinement superlattices 20, 40 are deposited on the transfer layer 50, with a sacrificial GaN layer 30 arranged between the confinement superlattices 20, 40. The deposition of the transfer stacks 500 may be repeated one or more times.

[0090] Subsequently, with reference to [Fig. 6], a second support substrate 60 is bonded to the front face of the temporary substrate 100. The second support substrate is, for example, made of silicon and may have a layer of silicon oxide on its surface. In other embodiments, the second substrate may be made of another material such as polycrystalline aluminum nitride or polycrystalline silicon carbide. Typically, Without limiting the invention, the second support substrate is bonded to a confinement superlattice 20 extending over a sacrificial layer of GaN 30 on the front face of the first support substrate containing the transfer stacks. The second support substrate 60 can be made of Si, SiO2, A1N, Al2O3 or GaN, of Si with a surface layer of SiO2, or any other substrate suitable for receiving the transfer stacks 500.

[0091] The first support substrate 10 can then be removed with reference to [Fig.7]. The removal can be carried out by various techniques, for example by mechanical removal, grinding, by irradiation of the interface by a laser (LLO, acronym for the Anglo-Saxon term "Laser Lift-Off" for laser detachment), or by dismantling a demounting layer deposited on the first support substrate 10 before the deposition of the transfer stacks 500.

[0092] In any case, deposition can begin with any layer chosen from a GaN transfer layer, a sacrificial layer, a first or second confinement superlattice for the fabrication of a temporary substrate, and the unused layers on the donor substrate can be removed before or after transferring the stack to the second support substrate. In most cases, the first and second confinement superlattice are identical, which also eliminates the need to reverse the layer deposition order when fabricating a temporary substrate 100. In an alternative embodiment, a temporary substrate 100 is formed as described above by deposition of one or more transfer stacks onto a support substrate, typically such that the GaN layers on the temporary substrate 100 exhibit Ga polarity.The polarity reversal can then be achieved by transferring a single layer of GaN onto a first receiving substrate, and subsequently onto a second receiving substrate in order to recover the initial Ga polarity, as described below. Transfer of a layer onto a receiving substrate

[0093] The donor substrate is intended for the transfer of a GaN layer onto a receiving substrate, which is typically a polycrystalline silicon carbide, passivated polycrystalline aluminum nitride, or a substrate with a SiO2 surface layer, or silicon. Depending on the desired polarity of the GaN layer after transfer, a donor substrate fabricated by epitaxial growth of one or more transfer stacks on a single support substrate is chosen, or a donor substrate fabricated using a temporary substrate and a second support substrate as described above. In both cases, the donor substrate comprises a support substrate and one or more transfer stacks. A GaN layer to be transferred extends over the free front face of the donor substrate. Hereafter, reference number 10 will be used for the support substrate in the process of transferring a layer onto a receiving substrate. This substrate can also be a second support substrate bonded to the transfer stacks in a polarity reversal step as described above.

[0094] To carry out a transfer of a GaN layer, with reference to [Fig.8A] and [Fig.8B], an implantation of atomic species, typically hydrogen, helium or a mixture of hydrogen and helium, is implemented so as to form a weakening zone 51 in the sacrificial GaN layer 30 between the first confinement superlattice 20 and the second confinement superlattice 40. In the case of a plurality of transfer stacks on a donor substrate, as illustrated in [Fig.8B], the implantation is carried out in the sacrificial layer closest to the front face of the donor substrate.

[0095] Implantation is carried out with a maximum concentration c near the middle of the GaN sacrificial layer, at a depth p that is typically between 300 and 700 nm, for example 400 nm from the surface of the donor substrate. The graph on the left side of [Fig. 8A] and [Fig. 8B] illustrates the concentration of the implanted species as a function of depth from the surface of the donor substrate.

[0096] The implantation causes the formation of vacancies in the crystalline lattice of the sacrificial layer, but these remain localized between the first and second confinement superlattices. The confinement superlattices thus promote the agglomeration of these vacancies and their subsequent precipitation as pyramidal defects within the sacrificial layer 30. These pyramidal defects are thus localized within the sacrificial layer. The set of defects weakens a precise zone located within the sacrificial layer 30, such that thermal or mechanical treatment can initiate the fracture of the substrate along this zone of embrittlement.

[0097] With reference to [Fig. 9A] and [Fig. 9B], the donor substrate is then bonded to a recipient substrate 80. Bonding can be achieved by direct contact, possibly preceded by activation of the surface of the donor substrate and / or the recipient substrate, for example by irradiation. Bonding can also be achieved by hydrophilic or hydrophobic bonding and / or by means of one or more bonding layers that can be applied to the donor substrate and / or the recipient substrate, for example of SiO2, Si, silicon nitride (SiN), tungsten, or titanium.

[0098] With reference to [Fig. 10], the donor substrate is detached along the embrittlement zone 51, for example by heat treatment or by applying mechanical stress. Heat treatment is typically carried out at a temperature between 300°C and 400°C. Detachment of the substrate donor leads to the transfer of the GaN layer 50 with the second confinement superlattice 40 and a portion 31 of the sacrificial layer 30 of GaN onto the receiving substrate 80. With reference to [Fig. 11], the first confinement superlattice 20 and a second portion of the sacrificial layer 32 remain on the donor substrate.

[0099] With reference to [Fig. 12], the second confinement superlattice 40 and the portion 31 of the sacrificial GaN layer are subsequently removed from the receiving substrate by a chemical and / or mechanical method. Preferably, the removal is carried out selectively, for example by wet etching, plasma etching or chemical-mechanical polishing (CMP), in order to avoid degradation of the transferred GaN layer 50 which will be used for the fabrication of one or more electronic components.

[0100] In some embodiments, an inversion of the polarity of the transferred layer 50 is desired, in particular when the polarity of the transfer stacks has not been reversed during the fabrication of the donor substrate as described above and illustrated in Figures 6 and 7.

[0101] For example, when the polarity of the layer to be transferred onto the donor substrate is a Ga polarity, the GaN layer 50 exhibits an N polarity after the transfer onto the recipient substrate 80.

[0102] In this case, the receiving substrate 80 is a first receiving substrate and is used only temporarily. Optionally, the first receiving substrate 80 may include a demounting layer (not shown) on its surface. During the transfer steps described above and illustrated in Figures 9A to 12, the GaN 50 layer is transferred onto the first receiving substrate 80. When the first receiving substrate 80 includes a demounting layer, said demounting layer is arranged at the interface between the first receiving substrate 80 and the transferred GaN 50 layer.

[0103] In a subsequent step, the GaN 50 layer arranged on the first receiving substrate 80 is bonded to a second receiving substrate 81 as illustrated in [Fig. 13A],

[0104] With reference to [Fig. 13B], the first receiving substrate 80 is then removed, for example by mechanical removal, grinding, an LLO process, or by dismantling the removal layer arranged at the interface between the GaN layer 50 and the first receiving substrate 80.

[0105] After these steps, the GaN layer 50 is arranged on the second receiving substrate 81 in a reversed position with respect to its arrangement on the first receiving substrate 80. Thus, the polarity of the transferred GaN layer 50 is also reversed, typically to a Ga polarity.

[0106] After transferring the GaN layer onto the second receiving substrate 81, any residues on its surface can be removed. Any residues from the transfer stacks of the first receiving substrate 80 can also be removed so that the first substrate can be reused. receiving substrate 80 for a new transfer of a layer of GaN onto one of other receiving substrates intended for the manufacture of electronic components.

[0107] After the transfer steps, a finishing treatment of the transferred layer can be implemented, so as to cure the defects related to the implantation and to smooth the free surface of said layer.

[0108] When the donor substrate includes other transfer stacks, the first confinement superlattice 20 and the second portion 32 of the sacrificial layer are also removed, as shown in [Fig. 14], to expose a new transferable GaN layer 50A. The removal is carried out, for example, by wet etching, plasma etching, or chemical-mechanical polishing (CMP). A surface treatment can be applied to prepare the donor substrate for the transfer of a new GaN layer 50A onto another receiving substrate. The transfer steps can then be repeated to transfer each transferable GaN layer contained in the donor substrate onto a respective receiving substrate. REFERENCES

[0109] N. Cherkashin et al., “Confinement of vacancies during annealing of H implanted GaN sandwiched between two {InGaN / GaN] superlattices”, Appl. Phys. Lett 101, 023105 (2012).

Claims

Demands

1. Donor substrate for the transfer of a layer of GaN, comprising a support substrate (10, 60) and a transfer stack (500) extending over the support substrate, said transfer stack (500) comprising, from its base to its surface: • a first super-lattice for confining defects of the crystal lattice (20), • a sacrificial layer (30) of GaN, • a second super-lattice for confining defects (40) of the crystal lattice, and • a layer of GaN to be transferred (50), each super-lattice for confining (20, 40) comprising a plurality of alternations (25, 45) of a primary layer (21, 41) of InGaN and a secondary layer (22, 42) of AlGaN.

2. Donor substrate according to claim 1, wherein the plurality of alternations comprises between 2 and 20 alternations, preferably between 5 and 15 alternations.

3. Donor substrate according to claim 1 or claim 2, wherein each primary layer (21,41) has a thickness of less than 20 nm, preferably less than 10 nm, and more preferably equal to 3 nm.

4. Donor substrate according to any one of the preceding claims, wherein each secondary layer (22, 42) has a thickness of less than 20 nm, preferably less than 10 nm, and more preferably equal to 3 mm.

5. Donor substrate according to any one of the preceding claims, comprising a plurality of superimposed transfer stacks (500).

6. A method for manufacturing a donor substrate according to any one of claims 1 to 5, comprising providing a first support substrate (10), a seed layer (13) of GaN and / or a buffer layer (12) of GaN extending over an upper face of said first support substrate (10), and the formation of a transfer stack (500) comprising successively the following steps: (i) the formation of the first confinement super-lattice (20) on the seed layer (13) or the buffer layer (12) by alternating deposition of primary layers (21) and secondary layers (22), (ii) the formation of the sacrificial GaN layer (30) on the first confinement layer (20), (iii) the formation of the second confinement super-lattice (40) on the sacrificial GaN layer (30) by alternating deposition of primary layers (41) and secondary layers (42), (iv) the formation of a transfer layer of GaN (50) on the second confinement super-lattice (40),

7. A method for manufacturing a donor substrate according to claim 6, wherein the steps (i) to (iv) of forming a transfer stack (500) are carried out by epitaxial growth.

8. A method for manufacturing a donor substrate according to claim 6 or claim 7, wherein the steps (i) to (iv) of forming the transfer stack (500) are carried out successively in a single epitaxial chamber.

9. A method for manufacturing a donor substrate according to any one of claims 6 to 8, further comprising, prior to the step of forming a transfer stack (500), a deposition of the buffer layer (12) of GaN or the transfer of the seed layer (13) of GaN onto the first support substrate (10).

10. A method for manufacturing a donor substrate according to any one of claims 6 to 9, wherein the steps (i) to (iv) of forming the transfer stack (500) are repeated.

11. A method for manufacturing a donor substrate according to any one of claims 6 to 10, further comprising, after the formation of the transfer stack (500), the following steps: (v) bonding a front face of the transfer stack (500) to a second support substrate (60), (vi) removing the first support substrate (10), so as to expose a layer of GaN to be transferred (50).

12. A method for manufacturing a donor substrate according to claim 11 in combination with claim 10, wherein, in step (vi), the removal of the first support substrate (10) is carried out so as to transfer a plurality of transfer stacks (500) onto the second support substrate (60).

13. A method for transferring a layer of GaN (50) onto a first receiving substrate (80), comprising the following steps: • the implantation of atomic species into the sacrificial GaN layer (30) of a donor substrate according to any one of claims 1 to 5 to form a weakening zone (51), the defects created during said implantation being confined in the sacrificial GaN layer (30) between the first confinement superlattice (20) and the second confinement superlattice (40), • the bonding of the transfer stack (500) onto the first receiving substrate (80), • the detachment of the donor substrate along the weakening zone (51).

14. Transfer method according to claim 13, further comprising a step of selectively etching a remnant (31) of the sacrificial layer and of the second confinement super-lattice (40) on the first receiving substrate (80), so as to expose the transferred GaN layer (50).

15. Transfer method according to claim 14, further comprising, after selective etching, the bonding of the first receiving substrate (80) to a second receiving substrate (81) via the GaN layer (50) and the removal of the first receiving substrate (80), so as to reverse the polarity of the transferred GaN layer (50).

16. A transfer method according to claim 14 or claim 15 in combination with claim 5, further comprising a step of selectively etching a remnant (32) of the sacrificial layer and the confinement super-lattice (20) on the donor substrate so as to expose a second layer of GaN (50A) to be transferred.