Method for transferring surface layer to cavity

Abstract

The present invention relates to a method for transferring a surface layer (10) to a carrier substrate (20) comprising a cavity (23), the method comprising: providing a donor substrate, providing a carrier substrate (20) having a first surface and comprising cavities (23), each cavity opening at the first surface and having a bottom and a peripheral wall, creating at least one temporary column (30) in at least one of the cavities (23), the column (30) having an upper surface coplanar with the first surface of the carrier substrate (20), joining the donor substrate and the carrier substrate (20) at the first surface of the carrier substrate, thinning the donor substrate to form the surface layer (10), and removing the at least one temporary column (30).

Description

Technical Field

[0001] The present invention relates to the field of microelectronics and microsystems. In particular, the present invention relates to a method for transferring a surface layer onto a substrate comprising an array of cavities. Background Art

[0002] MEMS devices (MEMS is the acronym for Microelectromechanical systems) are widely used to manufacture various sensors for a variety of applications, examples of which may be mentioned are pressure sensors, microphones, radio frequency switches, electroacoustic and ultrasonic transducers such as piezoelectric micromachined ultrasonic transducers (pMUTs), etc.

[0003] Many of these MEMS devices are based on a flexible membrane suspended over a cavity. In operation, the deflection of the membrane, which is related to a physical parameter (e.g., acoustic wave propagation for the pMUT), is converted into an electrical signal (or vice versa, depending on whether the device is in receiver mode or transmitter mode).

[0004] To improve the performance of these cavity-on-MEMS devices, it may be useful to employ membranes with good crystal quality and uniform and well-controlled thickness. Silicon-on-insulator (SOI) substrates are particularly suitable for fabricating these devices because they provide a layer for forming a very high-quality surface for the membrane, as well as a buried oxide layer (and / or a carrier substrate) for housing the cavity below.

[0005] “Modeling, fabrication, and characterization of piezoelectric micromachined ultrasonic transducer arrays based on cavity SOI wafers.” (Journal of Microelectromechanical Systems 24.4 (2015) 1142-1149) published by Lu Yipeng and David A. Horsley presents an example of manufacturing pMUT devices from SOI substrates including buried cavities and the advantages brought about by this.

[0006] Depending on the type of device implemented, the geometry of the cavities (shape, lateral dimensions, depth), the geometry of the membranes (thickness) and their planar distribution (distance between cavities) will be different. Therefore, in certain geometries and distribution configurations, it may be complex to manufacture a substrate comprising a surface layer placed on a plurality of cavities and in particular to define a transfer method consistent with the transfer of a surface layer of small thickness onto a cavity of large dimensions.

[0007] Purpose of the Invention

[0008] The present invention aims to overcome all or some of the above-mentioned disadvantages. In particular, it relates to a method for transferring a surface layer to a substrate comprising a plurality of cavities.

[0009] Brief description of the invention

[0010] The present invention relates to a method for transferring a surface layer to a carrier substrate comprising a cavity, the method comprising:

[0011] Providing a donor substrate,

[0012] providing a carrier substrate having a first side and comprising cavities, each cavity opening on the first side and having a bottom wall and a peripheral wall,

[0013] generating in at least one of the cavities at least one temporary pillar having an upper surface coplanar with the first side of the carrier substrate,

[0014] joining the donor substrate and the carrier substrate via the first side of the carrier substrate,

[0015] Thinning the donor substrate to form the surface layer,

[0016] Remove at least one temporary column.

[0017] Other advantageous and non-limiting features according to the invention (alone or in any technically feasible combination):

[0018] Providing a donor substrate comprises: implanting a light substance in the donor substrate so as to form a buried fragile zone between a first portion of the donor substrate and a second portion of the donor substrate, the first portion being intended to form a surface layer and the second portion being intended to form a remainder of the donor substrate;

[0019] Thinning of the donor substrate includes: separating the surface layer from the rest of the donor substrate via a buried fragile region;

[0020] the first portion of the donor substrate has a thickness comprised between 0.2 micrometers and 2 micrometers;

[0021] At least one column is separated from the peripheral wall of the cavity;

[0022] The upper surface of the column has a circular, square, rectangular or cross-shaped outline;

[0023] At least one column engages at least one peripheral wall of the cavity;

[0024] The upper surface of at least one of the pillars forms a grid that engages the peripheral wall of the cavity;

[0025] a plurality of pillars forming an array of parallel walls that join the peripheral wall of the cavity at the ends of the parallel walls;

[0026] The bonding comprises directly bonding the first side of the donor substrate on the one hand and the upper surface of the at least one pillar of the carrier substrate on the other hand;

[0027] Removing the pillars includes: locally etching the surface layer to form through holes in the surface layer; and chemically etching the pillars through the holes;

[0028] The pillars are chemically etched by wet or dry etching;

[0029] The holes are formed perpendicular to the pillars;

[0030] The hole has a cross-sectional area that is less than or equal to the area of ​​the upper surface of the pillar;

[0031] The holes have a cross-sectional area that is greater than the area of ​​the upper surface of the pillars;

[0032] • A hole is formed in the surface layer outside a region overhanging the at least one cavity.

[0033] Removing the pillars includes locally etching the second side of the carrier substrate until the cavity to form a hole communicating with the cavity; and chemically etching the pillars through the hole.

[0034] The pillars include at least one material selected from silicon oxide, silicon nitride, single crystal silicon, polycrystalline silicon, amorphous silicon, and porous silicon.

[0035] • The donor substrate comprises at least one semiconductor or piezoelectric material. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Other features and advantages of the present invention will become apparent from the following detailed description of the invention, which is given with reference to the accompanying drawings, in which:

[0037] [ Figure 1 ] Figure 1 shows a structure comprising a surface layer placed on a buried cavity obtained using the transfer method according to the invention;

[0038] [ Figure 2a ]

[0039] [ Figure 2b ]

[0040] [ Figure 2c ]

[0041] [ Figure 2d ]

[0042] [ Figure 2e ]

[0043] [ Figure 2f ]

[0044] [ Figure 2g ]

[0045] [ Figure 2h ]

[0046] [ Figure 2i ]

[0047] [ Figure 2j ] Figures 2a to 2j The steps of the transfer method according to the present invention are shown;

[0048] [ Figure 3a ]

[0049] [ Figure 3b ]

[0050] [ Figure 3c ] Figure 3a to Figure 3c Other steps of the transfer method according to the present invention are shown;

[0051] [ Figure 4a ]

[0052] [ Figure 4b ]

[0053] [ Figure 4c ] Figures 4a to 4c shows a variant of the step of removing the temporary column, which step is included in the transfer method according to the invention;

[0054] [ Figure 5a ]

[0055] [ Figure 5b ]

[0056] [ Figure 5c ]

[0057] [ Figure 5d ]

[0058] [ Figure 5e ]

[0059] [ Figure 5f ]

[0060] [ Figure 5g ]

[0061] [ Figure 5h ]

[0062] [ Figure 5i ] Figures 5a to 5i An example of implementation of the transfer method according to the present invention is shown. DETAILED DESCRIPTION

[0063] In the description, the same reference numerals may be used for the same type of elements in the drawings. The drawings are schematic representations and are not drawn to scale for ease of reading. In particular, the thickness of the layers along the z-axis is not proportional to their lateral dimensions along the x-axis and y-axis; and the relative thickness of the layers relative to each other is not necessarily considered in the drawings.

[0064] The present invention relates to transferring the surface layer 10 to the containing cavity 23 ( Figure 1 ) of a carrier substrate 20, the transfer method resulting in the manufacture of a structure 100 comprising a buried cavity 23.

[0065] The method according to the invention comprises the step of providing a donor substrate 1 having a front side 11 intended to be joined to a carrier substrate 20 and a rear side 12 ( Figure 2a ). By way of example and not limitation, the donor substrate 1 may include at least one semiconductor (e.g., silicon, silicon carbide, gallium nitride, etc.), or a piezoelectric material (e.g., lithium tantalate, lithium niobate, aluminum nitride, zinc oxide, PZT, etc.).

[0066] The method further includes providing a carrier substrate 20 ( Figure 2b ) step; the carrier substrate 20 has a first side 21 intended to be bonded to the donor substrate 1 and a second side 22. As a non-limiting example, the carrier substrate 20 may include silicon, glass, sapphire, etc. The carrier substrate 20 includes a plurality of cavities 23 opened on the first side 21 of the carrier substrate 20. Each cavity 23 has a bottom 23a and a peripheral wall 23b.

[0067] The geometry of each cavity 23 depends on the target MEMS device and is defined as:

[0068] the shape of the cavity 23 in the plane of the first side 21 of the carrier substrate 20 (referred to as the main plane (x, y) in the rest of the description): it will probably be circular, square, rectangular or polygonal;

[0069] The lateral dimensions of the cavity 23 in the main plane (x, y): they will probably vary from a few micrometers to a few millimeters;

[0070] • The depth of the cavity 23 along the z-axis perpendicular to the main plane (x, y): it will probably vary from a few tens of nanometers to a few tens of micrometers or even a few hundred micrometers.

[0071] The planar distribution of the cavities 23 (ie their distribution in the main plane (x, y)) also depends on the target device and will define the cavity spacing 24 ( Figure 2c): It will likely vary from a few microns to hundreds of microns or even a few millimeters. The cavity spacing 24 will likely be uniform and the same over the entire surface of the carrier substrate 20 or vary between areas on the surface of said carrier substrate 20 .

[0072] It should be noted that carrier substrate 20 will likely contain cavities 23 having different shapes, lateral dimensions, depths and / or planar distributions, particularly where it is intended that various types of devices be co-integrated into structure 100 including buried cavities.

[0073] Various layers will likely be deposited on the bottom 23 a and / or walls 23 b of the cavity 23 (eg, silicon nitride, silicon oxide, etc.), depending on the type of MEMS device that is intended to be produced using the structure 100 including the buried cavity 23 .

[0074] Furthermore, the transfer method according to the invention also provides for a step of generating at least one temporary column 30 in at least one of the cavities 23 (and preferably in each cavity 23) Figure 2d ). The number and position of the pillars 30 will likely be selected depending on the lateral dimensions of each cavity 23.

[0075] The pillar 30 has an upper surface 31 in the principal plane (x, y) that is coplanar with the first side 21 of the carrier substrate 20. The lower surface of the pillar 30 is securely fastened to the bottom 23a of the cavity 23.

[0076] By way of example and not limitation, pillars 30 include at least one material selected from silicon oxide, silicon nitride, single crystal silicon, polycrystalline silicon, amorphous silicon, and porous silicon.

[0077] According to a first variant, the column (or more than one) 30 is separate from the peripheral wall 23b of the cavity 23. Figure 2e As shown, the pillars 23 are not in contact with the walls 23b of the cavities 23. Preferably, they are evenly distributed on the bottom 23a of each cavity 23.

[0078] The upper surface 31 of the pillar 30 may have various types of profiles. Figure 2f Several examples of the profile of the upper surface 31 are illustrated in : a circular, square, rectangular or cross-shaped profile.

[0079] According to this first variant, the pillars 30 will likely have dimensions ranging from a few microns to about 15 microns, for example 5 microns, 7 microns or even 10 microns, for diameter of a circular profile or side length of a square or rectangular profile.

[0080] According to a second variant, the pillar 30 or some of the pillars 30 are engaged with at least one peripheral wall 23b of the cavity 23. Figure 2g2, a plurality of examples of columns 30 are illustrated, each of which forms a cloison. According to one example, the upper surface 31 of the column 30 (or a plurality of columns 30) forms a grid that engages the peripheral wall 23b of the cavity 23. According to another example, some of the columns 30 in the cavity engage the peripheral wall 23b while other columns 30 are separated from the peripheral wall 23b. According to yet another example, the columns 30 form an array of parallel cloisons that engage the peripheral wall 23b of the cavity 23 at their ends.

[0081] According to this second variant, the pillar 30 will likely have a width ranging from a few microns to about 15 microns, for example 5 microns, 7 microns or even 10 microns. It will likely have a length ranging from a few microns up to a size allowing engagement with the peripheral wall 23b of the cavity 23, and therefore of the order of magnitude of the size of said cavity 23.

[0082] The transfer method according to the present invention further comprises a step of bonding the donor substrate 1 and the carrier substrate 20 via the first face 21 of the carrier substrate 20 ( Figure 2h ).

[0083] Advantageously, this step comprises directly bonding the front side 11 of the donor substrate 1 on the one hand and the first side 21 of the carrier substrate 20 and the upper surface 31 of the (at least one) pillar 30 on the other hand by molecular adhesion. The principle of molecular adhesion is well known in the prior art and will not be described in detail here. It will be noted that the substrates must have a very good surface finish (cleanliness, low roughness, etc.) to obtain a high-quality joint.

[0084] In particular, special attention must be paid to obtaining good coplanarity between the first side 21 of the carrier substrate 20 and the upper surface 31 of each pillar 30 to ensure effective bonding of the first side 21 and the upper surface 31 to the front side 11 of the donor substrate 1 .

[0085] Advantageously, in order to ensure a good quality joint, the bonding step comprises cleaning the surfaces to be bonded of the donor substrate 1 and the carrier substrate 20 before the surfaces are brought into contact. For example, conventional procedures used in microelectronics, in particular for silicon-based substrates, include ozone cleaning, SC1 cleaning (SC1 is an acronym for Standard Clean 1) and SC2 cleaning (SC2 is an acronym for Standard Clean 2), with rinsing in between. The surfaces to be bonded will also possibly be activated (e.g. using plasma) before contacting to promote high bonding energies between the surfaces.

[0086] Optionally, the donor substrate 1 and / or the carrier substrate 20 will likely include a bonding layer on the front side 11 and / or the first side 21 , respectively, to improve the bonding quality and bonding energy of their interface.

[0087] Then, the transfer method comprises the step of thinning the donor substrate 1 to form the surface layer 10 .

[0088] According to a first variant, the step of thinning the donor substrate 1 is performed by mechanical grinding, chemical mechanical polishing and / or chemical etching of the back side 12 of the donor substrate 1. At the end of the thinning step, a surface layer 10 ( Figure 2i ).

[0089] According to a second advantageous variant, the thinning is performed using SmartCut TM A process is performed which is based on the implantation of light ions and the separation via the implanted areas.

[0090] Thus, according to this second variant, the aforementioned step of providing a donor substrate 1 comprises: injecting a light substance into said donor substrate 1 so as to form a buried fragile zone 2 between a first portion 3 of the donor substrate 1, intended to form the surface layer 10, and a second portion 4 of the donor substrate, intended to form the remainder of the donor substrate 1 ( Figure 3a ).

[0091] The thickness of the first portion 3 and therefore of the future surface layer 10 depends on the implantation energy of the light species (eg hydrogen or helium). Advantageously, the implantation energy is chosen such that the first portion 3 of the donor substrate 1 has a thickness of about 0.2 to 2 microns.

[0092] Then, in the bonding step of the method, the donor substrate 1 is bonded to the carrier substrate 20 ( Figure 3b ).

[0093] Still according to this second advantageous variant, the step of thinning the donor substrate 1 comprises separating the surface layer 10 (formed by the separated first portion 3) from the remaining portion 4 of the donor substrate 1 via the buried fragile zone 2 ( Figure 3c ). This separation preferably takes place during the heat treatment, at a temperature comprised between a few hundred degrees and 700° C. Alternatively, it can be mechanically assisted or achieved after the heat treatment by mechanical stress.

[0094] At the end of the thinning step, a surface layer 10 ( Figure 3c ). To review, SmartCut TM The processing allows to obtain thin layers with excellent thickness uniformity. This criterion can be very advantageous for certain MEMS devices requiring flexible membranes of controlled thickness.

[0095] Using SmartCut TMIn some cases where the thickness of the surface layer 10 transferred by the process is insufficient, it can be increased again during the finishing process mentioned below by depositing an additional layer on the free surface 12 ′ of the surface layer 10 , for example by epitaxial growth or other known deposition methods.

[0096] In both variants described, after the surface layer 10 has been transferred to the carrier substrate 20, the thinning step may comprise a finishing treatment aimed at improving the crystal quality (removing defects from the layer), the surface quality (removing residual roughness from the free surface 12') and / or modifying the thickness of the surface layer 10. Such treatments will likely include one or more thermal treatments, chemical mechanical polishing, chemical etching, epitaxial growth and / or deposition of additional layers.

[0097] The function of the (at least one) temporary pillar 30 located in the cavity 23 is to mechanically support the surface layer 10 during the thinning step.

[0098] During chemical mechanical thinning according to the aforementioned first variant, the surface layer 10 overhanging the cavity 23 is susceptible to deformation.

[0099] Furthermore, according to the second variant described, if the reinforcement effect on the front side 11 of the donor substrate 1 is insufficient during the weakening of the buried fragile zone 2 and until the first part 3 and the second part 4 of the donor substrate 1 are separated, there is a risk that the surface layer 10 is not transferred facing the cavity 23. The (at least one) temporary pillar 30 located in the cavity 23 ensures this reinforcement effect on the front side 11 and therefore allows the surface layer 10 to be completely transferred to the entire carrier substrate 20 (especially above the cavity 23).

[0100] Advantageously, for a surface layer 10 of about 1 to 1.5 microns thickness, the spacing between the pillars 30 themselves and between the peripheral wall 23b of the cavity 23 and each pillar 30 is chosen to be between 10 and 50 microns, and preferably about 20 microns.

[0101] The transfer method according to the invention finally comprises a step of removing the (at least one) temporary column 30 .

[0102] Removing the pillar 30 may include locally etching the surface layer 10 to form at least one through hole 13 a , 13 b , 13 c in the surface layer 10 .

[0103] This local etching can be performed by photolithography and dry chemical etching or wet chemical etching. In particular, the mask deposited on the free surface 12' of the surface layer 10 allows the area to be etched to form the hole to be defined and the remaining part of the free surface 12' to be protected. It will be noted that during the formation of the cavity 23 and the column 30 on the carrier substrate 20, the alignment marks defined in the periphery of the carrier substrate 20 and / or in the area provided for the cutting road on the first side 21 of the carrier substrate 20 and / or the second side 22 of the carrier substrate 20 allow accurate positioning relative to the buried cavity 23 and the column 30 during the step of removing one or more columns. These marks will also be possible for subsequent steps that need to be aligned relative to the cavity 23 in the structure 100 including the buried cavity.

[0104] Figure 4a , Figure 4b and Figure 4c An enlarged view from above the surface layer 10 is shown, the outline of the underlying cavity 23 and the upper surface 31 of the pillar 30 having been drawn in dotted lines. The holes 13a, 13b, 13c may in particular be produced in any of the configurations shown in these figures.

[0105] like Figure 4a As shown, the hole 13a may be generated perpendicular to each pillar 30 and have a cross-sectional area smaller than the area of ​​the upper surface 31 of the pillar 30. Then, dry chemical etching or wet chemical etching suitable for etching the material of the pillar 30 is performed through the hole 13a so as to remove the pillar 30 and release the surface layer 10 over the entire range of the cavity 23.

[0106] Alternatively, the hole 13b may be generated perpendicular to each pillar 30 and have a cross-sectional area ( 100 Å) larger than the area of ​​the upper surface 31 of the pillar 30. Figure 4b ). Dry chemical etching or wet chemical etching is performed through the holes 13 a so as to remove the pillars 30 and release the surface layer 10 over the entire range of the cavity 23 .

[0107] Alternatively again, the hole 13c (or holes) may be created in a region of the surface layer 10 that overhangs the cavity 23 ( Figure 4c ). A wet chemical etch is performed via the holes 13 a in order to remove the pillars 30 and release the surface layer 10 over the entire extent of the cavity 23 .

[0108] It will be noted that for Figure 4a , Figure 4b and Figure 4c In the various configurations presented in FIG. 1 , the holes 13 a , 13 b , 13 c can be blocked, for example, by depositing polysilicon under vacuum or a controlled atmosphere.

[0109] According to a variant (not shown), the removal of the pillars 13 may comprise a local etching of the surface layer 10 so as to form at least one through hole 13 in said surface layer 10 in a region not positioned perpendicularly to the cavity 23. In this case, the hole 13 leads to a transverse channel generated in the carrier substrate 20 before the bonding step of the method; this transverse channel communicates with one or more surrounding cavities 23. Dry chemical etching or wet chemical etching may then be performed via the hole 13 and the transverse channel so as to remove the (at least one) pillar 30 and release the surface layer 10 over the entire extent of the cavity 23.

[0110] It will be noted that this variant allows the membrane (the portion of the surface layer 10 situated perpendicularly to the cavity 23 ) to remain intact by avoiding the passage of the holes 13 therethrough.

[0111] According to another variant (not shown), the removal of the pillars 30 may comprise forming at least one hole 13 by locally etching the second side 22 of the carrier substrate 20 up to the cavity 23. Advantageously, this etching of the second side 22 is carried out at the end of the manufacture of the MEMS device when the carrier substrate 20 is thinned to, for example, 400 microns, 200 microns, 100 microns, 50 microns or less. This allows the production of small-sized holes 13 while remaining within the ratio of etched thickness / size of the holes achievable using known chemical etching techniques.

[0112] At the end of the step of removing one or more temporary pillars 30, the structure 100 including the buried cavity (suitable for manufacturing MEMS devices due to the geometry of the cavity 23), the thickness of the surface layer 10 (flexible membrane) and the planar distribution of the cavity / membrane meet the specifications of the MEMS device. The transfer method according to the invention allows, during the thinning step of forming the surface layer 10, to transfer a high-quality surface layer 10 (in particular a layer 10 with a small thickness (less than a few micrometers)) to a cavity of any geometry (in particular a cavity with large dimensions (greater than tens of micrometers)) by using the temporary pillars 30 present in the cavity 23.

[0113] Example of Implementation

[0114] In this example, an attempt was made to form a structure 100 including a buried cavity 23 including a surface layer made of silicon with a thickness of 1.5 micrometers and cavities with a side length of 250 micrometers, a depth of 0.5 micrometers and a spacing of 100 micrometers.

[0115] The donor substrate 1 is a substrate made of silicon ( Figure 5a ). Before the light species are implanted, an oxide layer 5 of, for example, about 50 nm thick is formed on the front side 11 of the donor substrate 1, for example, by thermal oxidation. The implantation energy is set to 210 keV, and the dose of hydrogen species is about 7 E16 / cm2. Thus, a buried fragile region 2 is formed between the first portion 3 and the second portion 4 of the substrate 1.

[0116] Prior to the step of bonding to the carrier substrate 20 , the oxide layer 5 may be retained or removed.

[0117] The carrier substrate 20 is a substrate made of silicon. A thermal oxide layer 24 having a thickness of 0.5 micrometers is formed on the first side 21 and the second side 22 of the substrate 20. The thermal oxide layer present on the second side 22 may be partially or completely retained or deleted as appropriate. Alternatively, the oxide layer may be deposited (using known deposition techniques) only on the first side 21 of the carrier substrate 20.

[0118] Then, by photolithography, a mask 25 is defined on the first side 21 of the carrier substrate 20, said mask 25 comprising unmasked areas where the thermal oxide layer 24 will be etched and masked areas ( Figure 5b ). It will be noted that alignment marks are also defined at the periphery of the carrier substrate 20 and / or in the region of the scribe lines for subsequent lithographic steps to be able to determine the coordinates of the cavity 23 when it is required that the cavity 23 be buried beneath the surface layer 10 .

[0119] The definition of the unmasked area depends on the size and the desired planar distribution of the cavities 23 of the structure 100 on the one hand and on the arrangement of the temporary pillars 30 on the other hand.

[0120] Typically, in this case, each cavity 23 measures 250 microns per side, and the temporary pillars 30 are placed 25 microns from the peripheral wall 23b of the cavity 23 and spaced 25 microns apart from each other. The upper surface 31 of each pillar 30 is a square with a side length of 7 microns; alternatively, the upper surface 31 may be a circle with a diameter of 7 microns or a cross with the maximum dimension of the cross set to 7 microns.

[0121] In the unmasked areas, a dry chemical etch or a wet chemical etch is performed directly through the thickness of the thermal oxide layer 24 (ie, 0.5 microns) on the thermal oxide layer 24. Figure 5c ). Then, the mask 25 is removed.

[0122] Thus, a carrier substrate 20 is obtained, which includes a plurality of cavities opening on the first side 21 of the carrier substrate 20 and in which temporary pillars 30 are placed, the upper surface 31 of the temporary pillars 30 being coplanar with the first side 21 of the substrate 20 ( Figure 5d and Figure 5e ).

[0123] After the cleaning and activation process, the front side 11 of the donor substrate 1 and the first side 21 of the carrier substrate 20 are in contact and directly bonded ( Figure 5f ). It will be noted that direct bonding will likely be performed in ambient atmosphere or in a controlled atmosphere (pressure and nature of the gas) or in vacuum. Annealing for consolidating the bonding interface may be applied to the bonded structure at a temperature of about 350°C.

[0124] The separation via the buried fragile zone 2 is carried out during the separation heat treatment at a temperature of about 500° C.

[0125] Then, the surface layer 10 ( Figure 5g ).

[0126] Preferably, finishing treatment operations are performed, such as thermal oxidation methods and chemical mechanical polishing, in order to ensure that the transferred surface layer 10 has good surface and structural quality and obtains a thickness of 1.5 micrometers.

[0127] For the step of removing the temporary pillars 30, a mask 14, for example made of silicon nitride, is defined by photolithography using alignment marks provided on the carrier substrate 20, so as to define unmasked areas in the surface layer where the through holes 13a will be formed, the remaining free surface 12 of the surface layer 10 being masked and thus protected. The surface layer 10 made of silicon is locally etched dry or wet so as to form the holes 13a, where the cross-sectional area of ​​each hole 13a is selected to be smaller than the area of ​​the upper surface 31 of each pillar 30 ( Figure 5h ).

[0128] In the presence of the holes 13 a , chemical etching (for example dry chemical etching based on hydrofluoric acid (HF) vapor) is performed to remove the thermal oxide making the pillars 30 and thus free the surface layer 10 over the entire extent of the cavity 23 .

[0129] The mask 14 may be removed prior to the chemical etching of the pillars 30 or at the end of the step of removing the pillars 30 .

[0130] Then, if necessary, the hole 13a can be blocked.

[0131] A structure 100 is obtained, wherein the structure 100 includes a buried cavity 23 ( Figure 5i ) and is suitable for manufacturing MEMS devices because the geometry of the cavity 23, the thickness of the surface layer 10 (flexible membrane) and the planar distribution of the cavity / membrane meet the above specifications. The transfer method according to the present invention allows, during the thinning step of forming the surface layer 10, by using a temporary pillar 30 placed in the cavity 23, a high-quality surface layer 10 (especially a layer 10 with a small thickness (about 1 micron in this example)) to be transferred to a cavity of any geometry (especially a cavity with large dimensions (250 microns×250 microns in this example)).

[0132] Of course, the invention is not limited to the embodiments and examples described and variant embodiments may be introduced therein without departing from the scope of the invention as defined by the claims.

Claims

1. A transfer method for transferring a surface layer (10) to a carrier substrate (20) comprising a cavity (23), the transfer method comprising: - providing a donor substrate (1), - providing the carrier substrate (20), the carrier substrate (20) having a first side (21) and comprising cavities (23), each cavity (23) opening on the first side (21) and having a bottom (23a) and a peripheral wall (23b), - generating at least one temporary column (30) in at least one of the cavities (23), the column (30) having an upper surface (31) coplanar with the first side (21) of the carrier substrate (20), wherein the at least one temporary column (30) engages at least one peripheral wall (23b) of the cavity (23), - joining the donor substrate (1) and the carrier substrate (20) via the first side (21) of the carrier substrate (20), - thinning the donor substrate (1) to form the surface layer (10), and - Removing said at least one temporary column (30).

2. The transfer method according to claim 1, wherein: - providing the donor substrate (1) comprises: injecting a light substance into the donor substrate (1) to form a buried fragile zone (2) between a first portion (3) of the donor substrate (1) and a second portion (4) of the donor substrate (1), the first portion being intended to form the surface layer (10) and the second portion being intended to form the remaining portion (4) of the donor substrate (1), - thinning the donor substrate (1) comprises separating the surface layer (10) from the remaining portion (4) of the donor substrate via the buried fragile zone (2).

3. The transfer method according to claim 2, wherein: The first portion (3) of the donor substrate (1) has a thickness comprised between 0.2 micrometers and 2 micrometers.

4. The transfer method according to claim 1, wherein: The thinning of the donor substrate (1) comprises at least one mechanical grinding operation and / or at least one chemical mechanical polishing operation and / or at least one chemical etching operation performed on the back side (12) of the donor substrate (1).

5. The transfer method according to any one of claims 1 to 4, wherein: The column (30) is separated from the peripheral wall (23b) of the cavity (23).

6. The transfer method according to claim 5, wherein: The upper surface (31) of the column (30) has a circular, square, rectangular or cross-shaped outline.

7. The transfer method according to claim 1, wherein: The upper surface (31) of the at least one temporary column (30) forms a grid engaging the peripheral wall (23b) of the cavity (23).

8. The transfer method according to claim 1, wherein: A plurality of pillars (30) form an array of parallel walls that join the peripheral wall (23b) of the cavity (23) at the ends of the parallel walls.

9. The transfer method according to any one of claims 1 to 4, wherein: The joining comprises directly bonding the donor substrate (1) on the one hand and the first side (21) of the carrier substrate (20) and the upper surface (31) of the at least one temporary pillar (30) on the other hand.

10. The transfer method according to any one of claims 1 to 4, wherein: Removing the pillars (30) comprises: locally etching the surface layer (10) to form through holes (13, 13a, 13b, 13c) in the surface layer (10); and chemically etching the pillars (30) through the holes.

11. The transfer method according to claim 10, wherein: The holes (13, 13a, 13b) are formed perpendicular to the pillars (30).

12. The transfer method according to claim 10, wherein: The hole (13a) has a cross-sectional area smaller than the area of ​​the upper surface (31) of the column (30).

13. The transfer method according to claim 10, wherein: The hole (13b) has a cross-sectional area larger than the area of ​​the upper surface (31) of the column (30).

14. The transfer method according to claim 10, wherein: The holes (13) are formed in the surface layer (10) outside the area overhanging the at least one cavity.

15. The transfer method according to any one of claims 1 to 4, wherein: Removing the pillar (30) includes: locally etching the second side (22) of the carrier substrate (20) until the cavity (23) to form a hole communicating with the cavity; and chemically etching the pillar (30) through the hole.

16. The transfer method according to any one of claims 1 to 4, wherein: The pillars (30) include at least one material selected from silicon oxide, silicon nitride, single crystal silicon, polycrystalline silicon, amorphous silicon and porous silicon.

17. The transfer method according to any one of claims 1 to 4, wherein: The donor substrate (1) comprises at least one semiconductor or piezoelectric material.

Images