Method for manufacturing a silicon layer donor structure for the purpose of quantum applications

Abstract

Method (100) for manufacturing a donor structure (DonStruct) capable of providing a silicon layer (28SiSLay) consisting of at least 99.92% silicon isotope 28Si, the method comprising the steps of: forming (130SLay) the silicon layer (28SiSLay) on a first substrate (Si_WafSLay), thus forming a seed structure (SLay); forming (120Car) a silicon nitride layer (SiN) on a second substrate (Si_WafCar), thus forming a carrier structure (Car); forming (140SLay) a silicon oxide layer (28SiO2) on at least one of the silicon layer (28SiSLay) and the silicon nitride layer (SiN); assembling (160) the seed structure (SLay) and the carrier structure (Car) so that the second substrate (Si_WafCar), the silicon nitride layer (SiN), the oxide layer (28SiO2), the silicon layer (28SiSLay) and the first substrate (Si_WafSLay) are stacked in this order; and clearing (170, 180) a surface of the silicon layer (28SiSLay) located opposite the second substrate (Si_WafCar), thus forming the donor structure (DonStruct).

Description

METHOD FOR MANUFACTURING A SILICON LAYER DONOR STRUCTURE FOR THE PURPOSE OF QUANTUM APPLICATIONSTECHNICAL FIELD OF THE INVENTION

[0001] The invention relates to a method for forming a structure comprising a layer of28Si silicon to transfer part of this layer to form a base substrate for the manufacture of electronic devices whose operating principle is based on the manipulation of quantum objects, particularly qubits.TECHNOLOGICAL BACKGROUND

[0002] Silicon is recognized as a promising material for electronic devices based on the manipulation of the spin of electrons, which represents their magnetic moment. In particular, information can be stored and manipulated by altering the spin of electrons, rather than by using their electrical charge as in conventional electronics. The spin of electrons can also be used as a support for qubits, for the purpose of putting quantum information theory into practice, by making use of superposition of spin states with a well-defined phase. The documents EP 3 975 072 A1 and US 2023 / 0026518 may be consulted.

[0003] The period of maintenance of the phase, and therefore of the information, is defined by the coherence time. Long coherence times are necessary to ensure that operations on qubits can be performed, or that quantum information can be stored, before a read operation in which the qubit loses its coherence and the information is lost.

[0004] Silicon has three stable isotopes, namely the isotopes28Si,29Si and30Si, which constitute about 92.2%, 4.7% and 3.1% of natural silicon, respectively. One of the mechanisms that promotes the decoherence of spin states is the presence of29Si isotopes of silicon, whose nuclear magnetic moment is not zero. It is therefore preferable to eliminate, as far as possible, the proportion of29Si in silicon used as the basis of qubit manipulation devices. The30Si isotope should also be eliminated: although the30Si isotope has no spin, it causes variations in the lengths of bonds between atoms, and therefore in the local environment of qubits. The homogeneity of the local environment should be maximized, by minimizing the presence of the30Si isotope, so that the characteristics of the qubits are as similar as possible to each other.

[0005] One solution is to use silicon enriched with28Si isotopes. Schneider et al. (E. Schneider and J. England, “Isotopically Enriched Layers for Quantum Computers Formed by 28Si Implantation and Layer Exchange”, ACS Appl. Mater. Interfaces 2023, 15, 21609−21617) propose a technique for enriching a28Si silicon isotope layer based on the ion implantation of28Si ions into a layer of aluminum formed on a substrate of silicon without native oxide, followed by layer exchange crystallization: this produces a high-purity28Si layer on the silicon substrate. Schneider also explains that defects at the interfaces of the layers of the devices make it necessary to use high-quality dielectrics and / or techniques for keeping the qubits away from the interfaces and the noise that they create, depending on the qubit placement geometry and whether the qubits are close to or distant from the interfaces.

[0006] Devices for manipulating spin qubits can be produced by using silicon as a base, with standard CMOS (Complementary Metal Oxide Semiconductor) or FD-SOI (Fully Depleted Silicon on Insulator) manufacturing methods, as used in the semiconductor industry. In particular, the FD-SOI technique may be used to control the placement of the qubits, at a distance from the interfaces between the silicon and the gate dielectric or the buried oxide (BOX) layer of the substrate.

[0007] However, there are no commercially available substrates of the SOI (silicon on insulator) type that have a thin layer of the28Si isotope of silicon of quantum quality. Here, “quantum quality” is taken to mean28Si layers that are pure enough, with a28Si purity of more than 99.92%, for the spin coherence of the electrons to have a sufficient duration to permit operations on the qubits.

[0008] Thus, there is a need for silicon wafers that can be used as substrates for the manufacture of electron spin qubit manipulation devices that are capable of maintaining the isotopic purity of28Si layers and enabling qubits to be manipulated at chosen locations compatible with the type of device envisaged.

[0009] More specifically, there is a need for SOI substrates having a thin layer of28Si on a support, with an electrically insulating layer interposed between the28Si layer and the support. It is advantageous for the thin layer and the electrically insulating layer (the buried BOX layer mentioned above) to be thin enough for good coupling with back-gate which provides an additional knob for manipulating the qubit location and to form FD-SOI (Fully Depleted SOI) transistors with back-bias capability.

[0010] Furthermore, given that quantum applications require operations to be carried out at very low temperatures, of less than 3 K, in the order of 10 mK for example, while using control electronics that are physically as close as possible to the supports of the quantum magnitudes manipulated, the integration of FD-SOI transistors and devices for manipulating quantum magnitudes on the same substrate is advantageous: FD-SOI transistors can have a low threshold voltage and can operate while limiting the dissipation of energy into their support, and therefore while limiting the problems of heating of these quantum magnitude manipulation devices.

[0011] A method for manufacturing substrates adapted for the FD-SOI technology, with an isotopically pure surface layer of28Si silicon suitable for quantum applications, could be based on a layer transfer technique, which requires the formation of at least one layer of28Si on a donor substrate. This donor substrate may comprise a support in the form of a conventional solid silicon wafer, that is to say, a wafer formed from different silicon isotopes and containing a proportion of impurities. Now, the purity in general, and the28Si isotope purity in particular, of the28Si layer must be maintained (i) during the process of manufacturing the donor substrate, (ii) during the transfer to a receiver substrate and (iii) during the CMOS process for manufacturing transistors on the receiver substrate.

[0012] Because of the high price and low availability of precursors of quantum quality for28Si epitaxy, the manufacture of the donor substrate and the use of thicknesses of28Si produced require improvement.SUMMARY OF THE INVENTION

[0013] The aim of the applicant is to provide a method for cost-effective manufacturing a donor structure of a layer of pure28Si isotope of silicon for devices used in manipulating quantum objects, particularly qubits formed by electron spins.

[0014] For the purpose of achieving this aim, one aspect of the invention is a method for manufacturing a donor structure capable of providing a silicon layer consisting of at least 99.92% silicon isotope28Si, the method comprising the steps of: forming the silicon layer on a first substrate, thus forming a seed structure; forming a silicon nitride layer on a second substrate, thus forming a carrier structure; forming a silicon oxide layer on at least one of the silicon layer and the silicon nitride layer; assembling the seed structure and the carrier structure so that the second substrate, the silicon nitride layer, the oxide layer, the silicon layer and the first substrate are stacked in this order; and clearing a surface of the silicon layer located opposite the second substrate, thus forming the donor structure.

[0015] A first advantage of the method is that it can be implemented by known procedures that are in common use in the field of the semiconductor industry for the manufacture of quantum devices, thus accelerating the technological maturity of the manufacturing method undertaken and reducing the necessary investment of time and equipment.

[0016] A second advantage of this method is that it can be used to obtain a donor structure of an28Si isotope layer of the silicon supported by a conventional substrate, without contamination due to the diffusion of impurities and29Si and30Si isotopes present in this conventional substrate in the course of manufacture.

[0017] A third advantage is that this method can be used to obtain a donor structure capable of undergoing refresh operations, possibly including a thickening of the28Si layer, in advantageous conditions in terms of efficiency in the use of the manufacturing machines the precursor gases of the28Si, by comparison with the initial manufacture of such a layer. This is because, during its initial formation, the28Si layer is formed on layers that can transmit impurities and29Si and30Si silicon isotopes to it. The layer is therefore formed at a relatively low temperature to limit its contamination by thermally activated diffusion of the impurities and isotopes. When transferred to a substrate from which it is insulated by a layer of silicon nitride (SiN), it can be thickened by higher-temperature methods, allowing faster growth and yielding or cost savings (i) in the use of precursor gases with a high content of28Si silicon isotope, (ii) in the time required for using epitaxy frames for growing silicon layers essentially consisting of the28Si silicon isotope, and (iii) an option of grown thin thermal oxide, or chemical oxide or deposited chemical deposition oxide after deposition of SiN.

[0018] According to additional non-limiting characteristics of the support according to the invention, considered individually or in any technically feasible combination:the method can comprise a step of thickening the cleared silicon layer;the step of forming the silicon layer and the step of thickening the portion of the silicon layer can each be executed by epitaxial growth at a first temperature and at a second temperature, respectively, the second temperature being higher than the first temperature;the method can further comprise, before the step of forming the silicon layer on the first substrate, a step of forming a layer of a silicon-germanium alloy on the first substrate, so that the layer of silicon-germanium alloy is interposed between the first substrate and the silicon layer, and wherein the step of clearing the silicon layer can comprise the steps of: eliminating at least a part of the first substrate fixed to the layer of a silicon-germanium alloy, and eliminating the layer of a silicon-germanium alloy, by a selective chemical etch or physico-chemical etch, distinct from the step of eliminating the at least a part of the first substrate, so as to clear the silicon layer;the layer of a silicon-germanium alloy can have a proportion of silicon isotope28Si of at least 99.92%;the oxide layer can be formed by oxidation of the silicon layer;the silicon layer can have, before the step of forming the oxide layer, a thickness of between 20 and 110 nm;

[0019] - the method can further comprise the steps of: before the assembly step, forming a weakened plane in the first substrate by implantation of ions of a light species; and breaking the first substrate at the weakened plane, after the assembly step and before the step of eliminating at least a part of the first substrate that has remained fixed to the layer of a silicon-germanium alloy;

[0020] The invention extends to a method for manufacturing a structure comprising a carrier substrate supporting a silicon layer consisting of at least 99.92% of silicon isotope28Si, the method comprising the steps of: providing a donor structure having a structure as defined by the method for manufacturing a donor structure; providing a device structure comprising a third substrate; forming another oxide layer on at least one of the silicon layer and one face of the device structure; forming another weakened plane in the silicon layer by implantation of ions of a light species; assembling the donor structure and the device structure so that the third substrate, the silicon nitride layer, the other oxide layer, the silicon layer and the first substrate are stacked in this order; breaking the silicon layer at the other weakened plane, after the assembly step, so as to produce a support structure with the silicon layer on its surface.

[0021] According to additional non-limiting characteristics of the support according to the invention, considered individually or in any technically feasible combination:after the step of breaking the silicon layer, a portion of the silicon layer, remaining fixed to the donor structure, can be refreshed so that the donor structure can be re-used in the execution of the method for manufacturing a structure comprising a carrier substrate supporting a silicon layer consisting of at least 99.92% of silicon isotope28Si;refresh step can comprise a chemical mechanical polishing only or a chemical mechanical polishing followed by a step of reforming the silicon layer;

[0022] - the step of forming the silicon layer and the step of thickening the portion of the silicon layer can each be executed by epitaxial growth at a first temperature and at a second temperature, respectively, the second temperature being higher than the first temperature;

[0023] - before the step of forming the other oxide layer, the face of the device structure can be covered with a layer of silicon nitride.

[0024] During the step of forming the silicon layer on the first substrate, the silicon layer can be formed by addition of silicon onto the first substrate.

[0025] During the step of forming the silicon layer on the first substrate, the silicon layer can be formed by addition of silicon onto the first substrate, the silicon could have an isotope 28 purity of silicon28Si comprised between 99.920% and 99.995%.

[0026] A first advantage of this method is that it can be used for manufacturing a structure that can act as a support for a device for manipulating quantum objects having high stability.

[0027] A second advantage is that it allows the donor structure to be re-used and refreshed.BRIEF DESCRIPTION OF THE FIGURES

[0028] Other features and advantages of the invention will become apparent from the following detailed description of the invention, this description being given with reference to the appended figures, in which:

[0029] shows a manufacturing method according to a first embodiment of the invention;

[0030] shows the continuation of the method of;

[0031] is a diagram summarizing the manufacturing method of Figures 1 and 2;

[0032] shows a manufacturing method according to a second embodiment of the invention;

[0033] shows a continuation of the method of;

[0034] shows a continuation of the method of Figures 4 and 5;

[0035] is a diagram summarizing the manufacturing method of Figures 4 to 6.DETAILED DESCRIPTION OF THE INVENTION

[0036] First embodiment

[0037] A first embodiment of the present invention is described by means of Figures 1 to 3 and the associated description below.

[0038] Figures 1 and 2 show the steps of a method for manufacturing the structure shown at (I) in.is a diagram indicating the steps of the manufacturing method shown in Figures 1 and 2.

[0039] shows, at (I), a structure DonStruct formed by a barrier layer SiN of silicon nitride, a layer of 28SiO2 oxide, and a seed layer 28SiSLay, stacked in this order on a carrier substrate Si_WafCar, each of these elements preferably being in direct contact with the element or elements immediately adjacent to it.

[0040] The carrier substrate Si_WafCarhere consists of a conventional monocrystalline silicon wafer, also referred to as a “wafer”. The seed layer 28SiSLayis formed of this silicon. The 28SiO2 layer is formed by a silicon oxide layer. The layer 28SiSLay, and preferably the oxide layer 28SiO2, have a28Si silicon isotope purity of at least 99.92%, for example between 99.92% and 99.995%. Thus the layer 28SiSLaycomprises less than 800 ppm of29Si silicon isotope.

[0041] Such a structure can be used to form quantum devices: because of the28Si purity of the layer 28SiSLay, the spin coherence time of the electrons of this layer is sufficient for carrying out operations on qubits associated with these electrons. When integrated into a quantum device, the surface layer 28SiSLayis therefore a good qubit support.

[0042] Furthermore, the layer SiN and the oxide layer 28SiO2 prevent, or at least limit, the migration of impurities, particularly atoms of the29Si silicon isotope, from the carrier substrate Si_WafCarto the layer 28SiSLay. Without a barrier layer, the phenomenon of impurity diffusion would increase the concentrations of29Si,30Si or dopants to levels incompatible with the quality required for the correct operation of quantum devices, resulting in a reduction of the coherence times and causing these devices to operate in a similar manner to devices formed in conventional substrates for FD-SOI transistors.

[0043] In preliminary steps 110SLayand 110Carof the method 100 for manufacturing the structure DonStruct, a seed substrate Si_WafSLayand a carrier substrate Si_WafCarare provided. In this example, each substrate consists of a monocrystalline silicon wafer, but any other support conventionally used in the semiconductor industry could be employed.

[0044] shows at (A) the successive formation, in this order, of (1) an etch stop layer of silicon-germanium alloy SiGe, and (2) the seed layer 28SiSLayon the seed substrate Si_WafSLay, in two successive steps 120SLayand 130SLay. The resulting structure, called the “seed structure”, is denoted SLay in the figures. The layer of silicon-germanium alloy SiGe can be grown, by epitaxy for example, on one face of the seed substrate Si_WafSLay, and the seed layer 28SiSLaycan then be grown on the layer of silicon-germanium alloy SiGe thus formed. Each of these two layers is formed by adding material that constitutes the layer.

[0045] shows at (B) the formation of the layer 28SiO2 on the seed substrate Si_WafSLayin a step 140SLay. In step 140SLay, the seed layer 28SiSLayis surface oxidized to a certain thickness, to form the oxide layer 28SiO2.

[0046] shows at (C) the implantation Imp of light species such as hydrogen, helium, or a combination of such species in the volume of the seed substrate Si_WafSLaythrough the layers 28SiO2, 28SiSLayand SiGe, to form a weakened plane Frgl in the seed substrate Si_WafSLay, in a step 150SLay. This implantation may be an implantation of hydrogen and / or helium, that is to say an ion bombardment of the seed substrate Si_WafSLaywith hydrogen and / or helium through the layers 28SiO2, 28SiSLayand SiGe. In a known way, as shown at (C), the implanted ions form a weakened plane Frgl. The nature and dose of the implanted species and the implantation energy are chosen according to the nature and the thickness of the layers passed through, and the desired implantation depth.

[0047] In the preceding steps, the layers 28SiSLayand 28SiO2, and optionally the layers SiGe, are formed so that they have28Si silicon isotope 28 purities of between 99.920% and 99.995%. Thus the maximum concentrations of isotope 29 of silicon in these layers are, initially at least, between 800 and 50 ppm, these concentrations being low enough to allow the manufacture of quantum devices based on the spin of the electrons in the seed layer 28SiSLay. Such compositions of the layers 28SiSLay, SiGe and 28SiO2 can be obtained by using, in a conventional epitaxial growth method, silane gas, having a corresponding purity to that of the layers of28Si, as the precursor gas of the silicon of the layers.

[0048] Although the manufacture of a layer SiGe having a silicon isotope 28 purity of between 99.920% and 99.995% is only optional, such a purity makes it possible to avoid the contamination of the layer 28SiSLayby29Si and30Si isotopes in the epitaxy frame during its formation. It should be noted that, in this case, the layer 28SiSlayis relaxed, and the quantum confinement of a quantum device formed on this layer should be provided by a vertical electrostatic field. Preferably, the same epitaxy frame is used for the successive formation of the two layers 28SiDonand SiGe.

[0049] Since the oxide layer 28SiO2 is produced by the oxidation of the layer 28SiSLay, it has the same characteristics of28Si purity as the latter. This operation of oxidation can be carried out in a conventional manner by application of a heat treatment under an oxygen atmosphere. The layer 28SiO2 is preferably formed by thermal oxidation so as to provide a high interface quality between the structures SLay and Car (see below), after the assembly of the two structures (see step 150, shown at (E) in). Evidently, this is a silicon oxide formed of oxygen and the28Si isotope of the silicon of the layer 28SiSLay. The layer 28SiCaris oxidized only to a certain thickness, which is strictly less than its thickness and is dependent on the intended thickness of the layer 28SiO2.

[0050] Additionally, the layer 28SiO2 may also be formed by deposition of the HDP-CVD (High Density Plasma CVD) type, which is carried out at relatively low temperature by comparison with thermal oxidation, thus making it possible, advantageously, to limit the thermal budget imposed on the layer 28SiCar, and thereby to limit the diffusion of the other isotopes of Si in the layer intended to include the devices for manipulating quantum objects.

[0051] Before the oxidation step 130Car, the layer 28SiSLaymust have a thickness approximately equal to the sum of the intended thickness of the layer 28SiO2 and the intended final thickness of the layer 28SiCarof the structure SLay. For example, thicknesses of between 10 and 100 nm, or preferably between 20 and 40 nm could be accepted for the layer 28SiSLayimmediately after the formation of the layer 28SiO2, or between 5 and 20 nm for the oxide layer 28SiO2. Thus a thickness of between 15 and 120nm can be specified for the surface layer 28SiSLay, during its deposition, before it is oxidized to form the oxide layer 28SiO2. The etch stop layer SiGe may have a thickness of between 10 and 100 nm, or preferably between 20 and 50 nm, typically 30 nm.

[0052] shows, at (D), the formation of the barrier layer SiN of silicon nitride on the carrier substrate Si_WafCar, in a step 120Car. The resulting structure, called the “carrier structure”, is denoted Car in the figures. The barrier layer SiN can be deposited on one face of the carrier substrate Si_WafCarby conventional deposition methods, such as chemical vapour deposition (CVD), or possibly by plasma-enhanced chemical vapour deposition (PECVD)or atomic layer deposition (ALD). The layer of silicon nitride may have a thickness of between 10 and 50nm, or preferably between 20 and 30 nm.

[0053] As a variant for the carrier structure Car, a carrier structure Car’ can be formed with an oxide film Ox, such as a silicon oxide film or a silicon nitride oxide film, that can be grown or deposited 1 to 10 nm thick on the barrier layer SiN of silicon nitride, as illustrated at (D’) of. Then, the carrier structure Car’ can replace the carrier structure Car in the processes described below. A first advantage of this variant is to facilitate the bonding to another oxide layer in a process such as the one illustrated below at (E) of. A second advantage of this variant is a reduction of mechanical stress-induced damages due to the compressive stress generated by the layer of silicon nitride SiN in any structure incorporating this layer. This effect is obtained by at least a partial relaxation or absorption of the stress within the oxide film Ox.

[0054] shows, at (E), the assembly of the structures SLay and Car to produce an assembly Set, required for the formation of the structure DonStruct comprising a stack of the layers SiN, 28SiO2 and 28SiSLay. Following the formation of the weakened plane Frgl, the seed structure SLay is turned over and the layer 28SiO2 is brought into close contact with the upper layer of the carrier structure Car, in this case the layer SiN, and is assembled to it, by molecular bonding or any other assembly technique in which surfaces of elements to be assembled are placed directly in contact with each other, in a step 160.

[0055] shows the steps of the method for moving from the state shown at (E) to the structure DonStruct shown at (I).

[0056] shows, at (F), the breaking of the seed substrate Si_WafSLayat the weakened plane Frgl produced by ion implantation, so that a part Si_WafSlay_2of the seed substrate Si_WafSLayis detached from the carrier structure Car, while a part Si_WafSLay_1of the seed substrate Si_WafSLayand the layers SiGe and 28SiSLayand 28SiO2 remain fixed there, in a breaking and removal step 170. This break may be made by heat treatment, assisted if necessary by mechanical force to initiate breaking, according to conventional layer transfer techniques that are well known in the field of microelectronics, such as the method known as “Smart Cut”. Following the assembly and the breaking of the donor substrate, heat treatment can be applied in order to consolidate the assembly and repair the assembled layers.

[0057] An alternative to the formation of a weakened plane and the breaking of the substrate Si_WafSLayat this plane may consist in thinning the donor substrate Si_WafSLayafter its assembly, for example by etching, grinding and / or chemical mechanical polishing (CMP). This thinning may be applied for the purpose of removing the seed substrate either completely or only partially, for example to reach the situation shown at (F) in.

[0058] shows at (G) the elimination, in a step 180, of the residual layer Si_WafSLay-1of the seed substrate Si_WafSLaywhich has remained fixed to the support structure Car at the end of step 180. The purpose of this step is to clear layer SiGe from the residual layer Si_WafSLay-1. This step may be implemented by selective etching of the silicon Etch (Si_WafSLay-1), by a wet method using a chemical solution, for example, or by a dry method using a plasma, for example, or by chemical mechanical etching of the CMP type. Such etching can be used to eliminate the residual layer Si_WafSLay-1formed of silicon, using the layer SiGe as an etch stop layer, by a method of selectively eliminating the silicon of the layer Si_WafSLay-1relative to the silicon-germanium alloy forming the layer SiGe. For example, the patent documents US 8,389,416 B2, US 9,984,890 B2 or US 10,934,485 B2, which describe methods of selectively etching silicon relative to a silicon-germanium alloy, may be consulted. One may etch the silicon constituting the residual layer Si_WafSLay-1selectively with regard to SiGe by using a KOH:H2O solution.

[0059] shows, at (H), the elimination, in step 190, of the layer SiGe cleared as a result of step 180. This step may be implemented by selective wet or dry etching of the silicon-germanium alloy (SiGe), stopped by the 28Si film. The wet method can use known chemical solutions, for example a solution of HF / HNO3 / H2O with 1 / 100 / 100 relative proportions in volumes, to remove the SiGe film over the 28Si film with selectivity varying from 10:1 to 100:1, which depends on the process condition and optimization. Another SiGe etching techniques is a dry method based on a plasma silicon-germanium etching using either fluorine-based or chlorine-based chemistries to etch the SiGe film with selectivity controlled by adjusting plasma conditions and gas mixtures. Such etching can be used to eliminate the layer SiGe formed from a silicon-germanium alloy, using the layer 28SiSLayas an etch stop layer, and using a method of selectively eliminating the silicon-germanium alloy of the layer SiGe relative to the silicon forming the layer 28SiSLay. For example, the patent documents US 9,236,265 B2, US 11,875,997 B2 or WO 2006 / 027332 A1, which describe methods of selectively etching silicon-germanium alloys relative to silicon, may be consulted.

[0060] Regarding the selective etch Etch(Si_WafSLay-1), a selective etch of the silicon relative to the silicon-germanium alloy may be considered to be an etch having a silicon etch rate at least twice as high as the silicon-germanium alloy etch rate. Conversely, for the selective etch Etch(SiGe), a selective etch of the silicon-germanium alloy relative to the silicon may be considered to be an etch having a silicon-germanium alloy etch rate at least twice as high as the silicon-germanium alloy etch rate.

[0061] Step 190 is distinct form step 180: step 180 is to etch silicon selectively with regard to silicon-germanium while step 190 is to etch silicon-germanium selectively with regard to silicon.

[0062] Following step 190, a finishing process can be applied to the seed layer 28SiSLayobtained from the seed structure SLay. This finishing process may consist of one or more mechanical and / or chemical etches to thin the seed layer 28SiSLayto a desired thickness and / or to polish its accessible face so as to improve its smoothness and reduce its roughness. The finishing step may be similar to those used in the manufacture of substrates for the FD-SOI technology. This finishing step may be provided in such a way that the layer 28SiSLaycan act as a seed layer for the epitaxial growth of silicon.

[0063] shows, at (G), an optional step 195 of thickening the layer 28SiSLayand the formation of a donor structure DonStruct. Step 195 can be implemented by epitaxial growth of a layer of silicon on the seed layer 28SiSLayin the same way as that in which the seed layer 28SiSLaywas initially formed on the layer SiGe in step 130SLay, to form a layer of silicon isotope 28Si having a purity in isotope 28 of silicon28Si of between 99.920% and 99.995%. The layer 28SiSLaythen acts as a crystal growth seed.

[0064] There is an important distinction between steps 130SLayand 195: in step 130SLay, the layer of 28SiSLayis formed on the layer SiGe of silicon-germanium alloy which separates it from the seed substrate Si_WafSLayformed from conventional silicon, relatively rich in impurities and silicon isotopes29Si and30Si. In this configuration, an excessively rapid temperature increase would cause the impurities and the29Si and30Si isotopes of silicon to be diffused into the 28SiSLaythrough the layer SiGe. On the other hand, in step 195, the seed layer 28SiSLaythat is grown by an additional deposition of silicon isotope28Si is separated from its carrier substrate Si_WafCarby the silicon oxide layer 28SiO2, and especially by the layer of silicon nitride SiN, which is an effective barrier to the diffusion of the impurities and the29Si and30Si isotopes contained in this carrier substrate Si_WafCar.

[0065] Because of this distinction, it is possible to use higher epitaxial growth temperatures for step 195 than for step 130SLay, thus achieving greater efficiency, with advantages in terms of the associated costs such as the amounts of precursor gases to be used and the time required for growth and consequently the period of use of the epitaxial growth frames.

[0066] The epitaxial growth of silicon of step 130SLaycan be carried out at a first temperature of between 350 and 500°C, and the epitaxial growth of silicon of step 195 can be carried out at a second temperature of between 600 and 1000°C, higher than the first temperature.

[0067] The thicknesses of the layers 28SiO2 and SiN can be chosen to provide an adequate barrier effect to prevent the contamination of the seed layer 28SiSLay, particularly by the29Si and30Si from the support Si_WafCar.

[0068] Another advantage of the donor structure DonStruct, due to the quality of the barrier formed by the layer SiGe, is that the donor structure can be used several times for transferring some of the thickness of its layer 28SiSLayto another substrate, while maintaining an acceptable purity of the layer 28SiSLayfor quantum applications. The layer in question may be the layer 28SiSlayas initially formed on the layer SiGe, or the layer 28SiSlayas thickened following step 195.

[0069] The second embodiment, described below, illustrates a method for using the donor structure DonStruct, produced by the method 100 for example, to form an adequate support for an electronic device suitable for quantum applications.

[0070] Second embodiment

[0071] The second embodiment of the present invention is described in Figures 4 to 7 and the associated description below.

[0072] Figures 4 and 5 show the steps of method 200 for manufacturing the structure shown in.shows a refresh of the donor structure for its re-use in the steps of the method shown in Figures 4 and 5.is a diagram indicating the steps of the manufacturing method shown in Figures 4 to 6.

[0073] shows a structure DevStruct that forms a support for the formation of an electronic device suitable for quantum applications. This structure comprises a device support Si_WafDevon which are stacked, in this order, a silicon nitride layer SiN, a silicon oxide layer 28SiO2, and a silicon layer 28SiSLay-1, each of these elements preferably being in direct contact with the element or elements immediately adjacent to it. Here, the device support Si_WafDevconsists of a conventional monocrystalline silicon wafer. The layer 28SiSLay-1, and preferably the layer 28SiO2, each have a28Si silicon isotope purity of at least 99.92%, for example between 99.92% and 99.995%. This structure is produced as shown in Figures 4 and 5.

[0074] A structure DonStruct, as defined in the first embodiment, and the device support Si_WafDevare provided, respectively, in preliminary steps 210Donand 210Devof method 200 for manufacturing the structure DevStruct.

[0075] shows, at (B), the application of steps 220Donand 230Donto the structure DonStruct shown at (A).

[0076] Step 220Don, which may be similar to step 140SLay, consists in forming a layer of silicon oxide 28SiO2 on the layer 28SiSLay. As for step 140SLay, the oxidation of the layer 28SiSLayprovides a layer of silicon oxide having a silicon isotope28Si purity at the same level as that of the layer 28SiSLay.

[0077] Step 230Donconsists of forming a weakened layer Frgl in layer 28SiSLayby the implantation Imp of light species such as hydrogen, helium, or a combination of such species, into the volume of layer 28SiSLay, through layer 28SiO2. This implantation may be an implantation of hydrogen and / or helium, that is to say, an ion bombardment of the layer 28SiSLay with hydrogen and / or helium. In a known way, as shown in (B), the implanted ions form a weakened plane Frgl. The nature and dose of the implanted species and the implantation energy are chosen according to the nature and the thickness of the layers passed through, and the desired implantation depth.

[0078] shows, at (C), the formation of a silicon nitride layer SiN on the device support Si_WafDev, in step 220Dev. This layer may be formed in the same way and according to the same parameters as the silicon nitride barrier layer formed in step 120Car.

[0079] shows, at (D), the assembly of the structures DonStruct and DevStruct to produce an assembly Ass, required for the formation of the finished structure DonStruct, the assembly Ass comprising a stack of the layers SiN, 28SiO2 and 28SiSLay.

[0080] More specifically, following the formation of the weakened plane Frgl in step 230Don, the structure DonStruct is turned over and the layer 28SiO2 formed on the layer 28SiSLayis brought into close contact with the upper layer of the structure DevStruct, in this case, the layer SiN, and is assembled to it, by molecular bonding or any other assembly technique in which surfaces of elements to be assembled are placed directly in contact with each other, in step 240.

[0081] shows step 250 of breaking the layer 28SiSLayat the weakened plane Frgl produced by ion implantation, so that the donor structure DonStruct is detached from the device structure DevStruct, leaving a part 28SiSLay-1of the layer 28SiSLayfixed.

[0082] This break may be made by heat treatment, assisted if necessary by mechanical force to initiate breaking, according to conventional layer transfer techniques that are well known in the field of microelectronics, such as the method known as “Smart Cut”. Following the assembly and the breaking of the donor substrate, heat treatment can be applied to consolidate the assembly and repair the assembled layers.

[0083] Following step 250 of breaking and removal, the two distinct and separate structures DonStruct and DevStruct are present.

[0084] Following step 250, the device structure DevStruct described above is provided with a layer of silicon 28SiSLay-1, obtained from the layer 28SiSLayand having substantially the same proportion of silicon isotope28Si and the same quantum properties as the latter. Immediately after the breaking step, the free surface of the layer 28SiSLay-1is irregular and rough, and a finishing step 260Devneeds to be executed on the layer 28SiSLay-1. This finishing process may consist of one or more mechanical and / or chemical etches to thin the seed layer 28SiSLayto a desired thickness and / or to polish its free face to improve its smoothness and reduce its roughness. The finishing step may be similar to those used in the manufacture of substrates for FD-SOI technology and / or to the finishing step following step 190. This finishing step may be provided in such a way that layer 28SiSLay-1can act as a substrate for the formation of quantum devices such as those incorporating FD-SOI transistors. Following step 260Dev, layer 28SiSLay-1may have a thickness of between 4nm and 50 nm, preferably between 5 and 20 nm. The finishing step may be performed by chemical mechanical polishing.

[0085] Following step 250, the donor structure DonStruct is provided with a silicon layer 28SiSLay-2, corresponding to the layer 28SiSLayfrom which the silicon layer 28SiSLay-1was removed by breaking at the weakened plane Frgl. For reasons similar to those mentioned for the layer 28SiSLay-1, a finishing treatment can be applied to layer 28SiSLay-1, in step 260Don, to obtain a sufficiently flat and smooth surface to allow the structure DonStruct to be directly re-used in step 210Donof the method 200 (alternative Alt1) or to allow the layer 28SiSLay-1to be used again as a crystallization seed for the epitaxial growth of silicon (alternative Alt2).

[0086] In the first alternative Alt1, the respective thicknesses of the layers 28SiSLay, 28SiSLay-1,and 28SiSLay-2can be such that it would be feasible to re-use the donor structure DonStruct, flattened and smoothed, in step 210Donof the method 200. This is the case when the residual thickness of the layer 28SiSLay-2, shown at (B) in, is sufficient to envisage the formation of the oxide layer and the weakened plane of steps 220Don and 230Don, followed by the breaking of step 250, to produce a new layer 28SiSLay-1that is sufficiently thick to form a quantum device, allowing for the finishing step 260Dev.

[0087] In a second alternative Alt2, if it is estimated that the residual thickness of layer 28SiSLay-2is insufficient to execute method 200 as described above, it is possible to thicken layer 28SiSLay-2to re-form a layer 28SiSLayin a step 270Don, and then to re-use the donor structure DonStruct in step 210Donof the method 200. In step 270Don, the layer 28SiSLay-2acts as a crystal growth seed for the epitaxial growth of layer 28SiSLay, in a method similar to that of step 195 of the first embodiment. Notably, the epitaxial growth for re-forming layer 28SiSLaycan take place at a higher temperature than the initial forming of this layer in step 130SLay.

[0088] A finishing step 280Donsimilar to step 260Doncan be applied to layer 28SiSLayproduced by thickening layer 28SiSLay-2, before the execution of steps 220Donand 230Don.

[0089] This method 200 offers high efficiency with respect of the resources needed to produce structures for accommodating quantum devices, in terms of precursor gases and the time required for using deposition or thin layer growing equipment.

[0090] In the methods described above, the layers referred to as “SiN” were presented as being formed of silicon nitride. An alternative would be for one or other, or both, of these layers to comprise a layer of silicon oxynitride in addition to the layer of silicon nitride. In these examples, the silicon nitride layer is preferably interposed between the layer of silicon oxynitride and the substrate in question, that is to say, the carrier substrate Si_WafCaror the device support Si_WafDev. The silicon oxynitride layer may contain nitrogen in a nitrogen / oxygen concentration ratio that is greater than, or equal to, 0.01, 0.05, or 0.1.

[0091] The invention is not limited to the embodiments described above, and variants may be employed without departing from the scope of the invention as defined by the claims.

Claims

Method (100) for manufacturing a donor structure (DonStruct) capable of providing a silicon layer (28SiSLay) consisting of at least 99.92% silicon isotope28Si, the method comprising the steps of:- forming (130SLay) the silicon layer (28SiSLay) on a first substrate (Si_WafSLay), thus forming a seed structure (SLay);- forming (120Car) a silicon nitride layer (SiN) on a second substrate (Si_WafCar), thus forming a carrier structure (Car);- forming (140SLay) a silicon oxide layer (28SiO2) on at least one of the silicon layer (28SiSLay) and the silicon nitride layer (SiN);- assembling (160) the seed structure (SLay) and the carrier structure (Car) so that the second substrate (Si_WafCar), the silicon nitride layer (SiN), the oxide layer (28SiO2), the silicon layer (28SiSLay) and the first substrate (Si_WafSLay) are stacked in this order; and- clearing (170, 180) a surface of the silicon layer (28SiSLay) located opposite the second substrate (Si_WafCar), thus forming the donor structure (DonStruct).Method according to Claim 1, comprising a step (195, 270Don) of thickening (200) the cleared silicon layer (28SiSLay).Method according to Claim 2, wherein the step (130SLay) of forming the silicon layer (28SiSLay) and the step (270Don) of thickening the portion (28SiSLay-2) of the silicon layer are each executed by epitaxial growth at a first temperature and at a second temperature, respectively, the second temperature being higher than the first temperature.Method according to any of Claims 1 to 3, further comprising, before the step (130SLay) of forming the silicon layer (28SiSLay) on the first substrate (Si_WafSLay), a step (120SLay) of forming a layer (SiGe) of a silicon-germanium alloy on the first substrate (Si_WafSLay), so that the layer (SiGe) of silicon-germanium alloy is interposed between the first substrate (Si_WafSLay) and the silicon layer (28SiSLay), andwherein the step (170, 180) of clearing the silicon layer (28SiSLay) comprises the steps of:- eliminating (180) at least a part (Si_WafSLay_1) of the first substrate fixed to the layer (SiGe) of a silicon-germanium alloy, and- eliminating (190) the layer (SiGe) of a silicon-germanium alloy, by a selective chemical etch or physico-chemical etch (Etch(SiGe)), distinct from the step (180) of eliminating the at least a part (Si_WafSLay_1) of the first substrate, so as to clear the silicon layer (28SiSLay).Method according to Claim 4, wherein the layer (SiGe) of a silicon-germanium alloy has a proportion of silicon isotope28Si of at least 99.92%.Method according to any of Claims 1 to 5, wherein the oxide layer (28SiO2) is formed by oxidation of the silicon layer (28SiSLay).Method according to Claim 6, wherein the silicon layer (28SiDon) has, before the step (140SLay) of forming the oxide layer (28SiO2), a thickness of between 20 and 110 nm.Method according to any of Claims 1 to 7 in their dependency on claim 4, further comprising the steps of:- before the assembly step (150), forming (140Don) a weakened plane (Fgrl) in the first substrate (Si_WafDon) by implantation of ions of a light species; and- breaking (160) the first substrate (Si_WafDon) at the weakened plane (Frgl), after the assembly step (150) and before the step (170) of eliminating at least a part (Si_WafDon_1) of the first substrate that has remained fixed to the layer (SiGe) of a silicon-germanium alloy.Method for manufacturing a structure (DevStruct) comprising a carrier substrate (Si_WafCar) supporting a silicon layer (28SiDon) consisting of at least 99.92% of silicon isotope28Si, the method comprising the steps of:- providing (210Don) a donor structure (DonStruct) having a structure as defined by the method of Claim 1;- providing (210Car, 220Dev) a device structure (DevStruct) comprising a third substrate (Si_WafDev);- forming (220Don) another oxide layer on at least one of the silicon layer (28SiSLay) and one face of the device structure (DevStruct);- forming (230Don) another weakened plane (Fgrl) in the silicon layer (28SiSLay) by implantation of ions of a light species;- assembling (240) the donor structure (DonStruct) and the device structure (DevStruct) so that the third substrate (Si_WafDev), the silicon nitride layer (SiN), the other oxide layer (28SiO2), the silicon layer (28SiSLay) and the first substrate (Si_WafSLay) are stacked in this order;- breaking (250) the silicon layer (28SiSLay) at the other weakened plane (Frgl), after the assembly step (240), so as to produce a support structure with the silicon layer (28SiSLay) on its surface.Method according to Claim 9, wherein, after the step (250) of breaking the silicon layer (28SiSLay), a portion (28SiSLay-2) of the silicon layer (28SiSLay), remaining fixed to the donor structure (DonStruct), is refreshed (260Don, 270Don, 280Don) so that the donor structure (DonStruct) is re-used in the execution of the method according to Claim 9.Method according to Claim 10, the refresh step (260Don, 270Don, 280Don) comprising a chemical mechanical polishing only or a chemical mechanical polishing followed by a step of reforming the silicon layer (28SiSLay).Method according to Claim 11, the step (130SLay) of forming the silicon layer (28SiSLay) and the step (270Don) of thickening the portion (28SiSLay-2) of the silicon layer each being executed by epitaxial growth at a first temperature and at a second temperature, respectively, the second temperature being higher than the first temperature.Method according to any of Claims 11 to 14, wherein, before the step (220Don) of forming the other oxide layer, the face of the device structure (DevStruct) is covered with a layer (SiN) of silicon nitride.The method according to any one of claims 1 to 13, wherein, during the step of forming (130SLay) the silicon layer (28SiSLay) on the first substrate (Si_WafSLay), the silicon layer (28SiSLay) is formed by addition of silicon onto the first substrate (Si_WafSLay).The method according to any one of claims 1 to 13, wherein, during the step of forming (130SLay) the silicon layer (28SiSLay) on the first substrate (Si_WafSLay), the silicon layer (28SiSLay) is formed by addition of silicon onto the first substrate (Si_WafSLay), the silicon having an isotope 28 purity of silicon28Si comprised between 99.920% and 99.995%.

Images