PATTERNING IN A THIN LAYER USING A DOUBLE-IMPLANT HARD MASK

Abstract

Method for producing patterns in a semiconductor layer comprising: -Providing a hard mask (103) on the semiconductor layer (12). -Forming a first implantation mask (115) on the hard mask (103). -Carrying out a first implantation of the hard mask (103). -Removing the first implantation mask (115). -Forming a second implantation mask (125) on the hard mask layer (103). -Carrying out a second implantation.-Selective removal of given regions (103'') of the hard mask which were implanted during the first implantation and the second implantation with respect to portions (103N) of the hard mask which were not implanted or zones (103') implanted only during the first implantation or the second implantation, so as to form in the hard mask an openwork hard mask comprising one or more patterns, -Transfer of said patterns into the semiconductor layer (12) by etching the semiconductor layer covered with the openwork hard mask (figure 10B).

Description

Title of the invention: PRODUCTION OF PATTERNS IN A THIN LAYER BY MEANS OF A DOUBLE-IMPLANT HARD MASK TECHNICAL FIELD AND PRIOR ART

[0001] The present invention relates to the field of producing patterns in a thin semiconductor layer and applies in particular to the formation of particular patterns intended to accommodate quantum dots or quantum dots (in English "Quantum dot" or Qdot").

[0002] To form very small patterns with a high density in a layer, it is possible to use an approach called "multi-patterning" (i.e. "multiple patterning"), i.e. a shaping process by iteration of steps in which a sequence of several photo-lithography and etching steps is carried out, such as for example the process called "LELE" for "lithoetch litho etch") before transferring patterns into the thin layer of interest. Another example is the process commonly called SADP ("Self-aligned double patterning").

[0003] When it is desired to obtain patterns of a particular shape, for example square or comprising orthogonal portions joining together, one approach may consist of carrying out two iterations of photo-etching carried out with so-called "perpendicular" photo-lithographies, using several photo-lithography steps with successive maskings of the patterns orthogonal to each other from one masking to the next.

[0004] An example of a “multiple patterning” method is given in document US20220013366A1 for forming a semiconductor structure in a semiconductor substrate according to an arrangement of the type called “cross-bar”, i.e. formed of orthogonal and intersecting semiconductor bars. This method provides for using a molding structure to form a first mask and a second mask whose patterns are orthogonal to the first and, after removing the molding structure, transferring the patterns of the first mask and the second mask into the semiconductor substrate. Such a method comprises a significant number of steps, in particular photolithography, etching, deposition, cleaning and poses difficulties with regard to the removal of the molding structure and the masks without impacting the other layers. In particular, it induces selectivity constraints between the different materials.With such a process, these are spacer patterns from which the patterns are created. An additional hard mask level may in some cases be necessary.

[0005] Another “multiple patterning” method is described in the document US7901869B2. It provides in particular the use of a three-layer masking and that of a two-layer hard mask in order to be able to transfer final patterns into a semiconductor surface layer of an SOI substrate.

[0006] The use of the bi-layer hard mask induces additional steps and difficulties in order to be able to carry out the removal of the hard mask without impacting other underlying layers. In particular, if this method is carried out on an SOI substrate, the presence of a silicon oxide layer in a bi-layer hard mask can make it difficult to etch and remove it without impacting the oxide layer of the substrate.

[0007] The problem arises of finding a new method for producing patterns in a thin layer, in particular a semi-conductor layer, which is preferably improved with respect to at least one of the drawbacks stated above. Statement of the invention

[0008] It is therefore an aim of the present invention to provide a method for producing one or more patterns in a thin layer, the method comprising, in this order, the following steps:

[0009] -Provide a hard mask on the thin layer, then;

[0010] -Forming on the hard mask a first implantation mask, the first implantation mask being provided with one or more openings which extend mainly in a first direction and reveal the hard mask, then;

[0011] - Performing a first implantation using a given species of the hard mask through the first openings of the first implantation mask,

[0012] -Forming on the hard mask a second implantation masking, the second implantation masking being provided with second openings revealing the hard mask and which extend mainly in a second direction orthogonal or substantially orthogonal to the first direction, the second openings being arranged so as to reveal one or more given regions of the hard mask already revealed by the first openings during the first implantation then;

[0013] - Performing a second implantation in the hard mask through the second openings of the second implantation mask, the second openings being arranged so as to reveal one or more given regions of the hard mask already revealed by the first openings during the first implantation;

[0014] -Selectively removing said given regions of the hard mask which were implanted during the first implantation and the second implantation with respect to portions of the hard mask which were not implanted or areas implanted only during the first implantation or the second implantation, so as to form in the hard mask an openwork hard mask comprising one or more patterns,

[0015] -Transfer said patterns into the thin layer by etching the thin layer re- covered with the hard openwork mask.

[0016] With such a method, it is possible, with a reduced number of steps, to produce patterns of particular shape while obtaining a high pattern density and well-controlled dimensions of the patterns.

[0017] Such a method also makes it possible to reduce the constraints of etching selectivity between the different materials to enable the patterns to be produced.

[0018] By “thin” layer we mean a layer with a thickness of less than 1 μm and typically between 10 and 100 nm.

[0019] Preferably after the first implantation, the first masking is removed.

[0020] Also preferably, after the second implantation, the second masking is removed.

[0021] Typically, the first implantation is performed with a first dose and the second implantation is performed with a second dose, the first dose and the second dose being such that the sum of the first dose and the second dose is greater than or equal to a first threshold and the difference between the first dose and the second dose is less than a second threshold.

[0022] According to one possible implementation, the first implantation is carried out with a given species and the second implantation is carried out with said same given species.

[0023] Advantageously, the first implantation is carried out with a first dose and the second implantation is carried out with a second dose equal to the first dose. This makes it easier to implement the implantation method.

[0024] This also makes it possible to obtain, before transferring the patterns into the semiconductor layer, a hard mask of more uniform thickness.

[0025] Advantageously, the first implantation is carried out by plasma and according to a first duration of exposure to the plasma and in which the second implantation is carried out by plasma and according to a second duration of exposure to the plasma identical to the first duration.

[0026] According to one possible implementation, the first implantation and the second implantation are provided such that the areas implanted only during the first implantation and only during the second implantation are made selectively etchable with a selectivity of at least 10:1 with respect to non-implanted portions of the hard mask and that said given regions of the hard mask which are implanted both during the first implantation and during the second implantation are made selectively etchable with a selectivity of at least 2:1 with respect to the areas implanted only during the first implantation and only during the second implantation.

[0027] According to one possible implementation, the first implantation masking is formed from portions producing a first given design oriented in a given direction parallel to the main plane of the substrate, the second implantation masking being formed from portions producing a second design.

[0028] Advantageously, the second drawing may be identical to the first drawing and oriented orthogonally to said given direction.

[0029] Advantageously, the first implantation masking is formed by a first photolithography using a photolithography mask and wherein the second implantation masking is formed by a second photolithography using said photolithography mask oriented at 90° relative to its orientation during the first photolithography.

[0030] It is thus possible to carry out a manufacturing process for which the number of photolithography masks used is minimized.

[0031] Preferably, the first masking is removed after the first implantation and prior to carrying out the second masking.

[0032] Advantageously, the thin layer is a semiconductor layer, in particular made of silicon. In this case, the semiconductor layer may be formed on a semiconductor-on-insulator type substrate or may be the surface layer of a semiconductor-on-insulator substrate, the substrate being provided with a buried insulating layer called “BOX” (Burried Oxide) made of silicon oxide.

[0033] According to one embodiment, the given species is light ions, in particular hydrogen or helium. Preferably, the hard mask is formed from a monolayer and / or a single given material. Rather than using a hard mask formed from a stack of several layers, such a hard mask imposes fewer constraints in terms of etching selectivity than a hard mask formed from several stacked layers of different materials.

[0034] Advantageously, the hard mask can be formed from one of the following materials: SiN, SiCO, HfO2, TiN.

[0035] According to a particular implementation, the hard mask can be provided in SiN. One way to obtain patterns in the SiN hard mask is to carry out selective removal by wet etching and based on HF. The transfer of said patterns into the semiconductor layer can be carried out using plasma etching. This allows significant selectivity, i.e. at least 10:1 between the hard mask and the thin layer.

[0036] After transferring the given hard mask pattern into the thin layer, the hard mask is typically removed, in particular by selective etching with respect to the thin layer, this removal being able to be carried out by plasma or by wet method.

[0037] According to one possibility, the first masking and the second implantation masking are formed by:

[0038] - production of a stack comprising at least one anti-reflective layer coated with at least one layer of photosensitive resin,

[0039] - formation of patterns in the photosensitive resin layer,

[0040] - reproduction of said patterns in the anti-reflective layer

[0041] - removal of the photosensitive resin layer.

[0042] Among said formed patterns transferred into the semiconductor layer there is at least one square-shaped pattern and / or at least one pattern formed of two bars joined and orthogonal to each other.

[0043] According to a particular aspect, the present invention relates to a method for manufacturing a quantum device comprising the implementation of a method as defined previously, the patterns being intended to accommodate one or more quantum boxes. Brief description of the drawings

[0044] The present invention will be better understood on the basis of the following description and the attached drawings in which:

[0045] [Fig. 1] illustrates an example of a possible starting structure for implementing an example of a method according to the present invention making it possible to produce particular patterns in a thin layer, in particular a semi-conductor layer.

[0046] [Fig.2] illustrates the production of a hard mask layer arranged on the thin layer.

[0047] [Fig.3A]

[0048] [Fig.3B]

[0049] [Fig.4] illustrate the production of a first implantation masking arranged on the hard mask layer.

[0050] [Fig.5] illustrates a first implantation of the hard mask layer through openings of the first implantation mask.

[0051] [Fig.6A]

[0052] [Fig.6B] illustrate one or more regions modified by implantation in the hard mask layer.

[0053] [Fig.7A]

[0054] [Fig.7B] illustrate the production of a second implantation masking arranged on the hard mask layer.

[0055] [Fig.8A]

[0056] [Fig.8B]

[0057] [Fig.8C] illustrate a second implantation through openings of the second masking, and the formation of areas of the hard mask layer implanted only during the second implantation, areas of the hard mask layer implanted only during the first implantation, regions of the hard mask layer implanted during both the first implantation and the second implantation, and non-implanted portions of this hard mask layer.

[0058] [Fig.9A]

[0059] [Fig.9B]

[0060] [Fig.9C] illustrate selective etching of regions of the hard mask layer implanted in both the first implantation and the second implantation, versus areas implanted only once or not modified.

[0061] [Fig.10A]

[0062] [Fig. 10B] illustrate a transfer of patterns from the hard mask into an underlying thin layer.

[0063] [Fig. 11] illustrates a method for determining the exposure time of a hard mask layer to an implantation plasma.

[0064] [Fig. 12] illustrates, by means of a curve obtained by simulation, a concentration gradient of susceptible species obtained following the implantation of a nitride layer by light ions.

[0065] Identical, similar or equivalent parts of the different figures bear the same numerical references so as to facilitate the transition from one figure to another.

[0066] The different parts represented in the figures are not necessarily on a uniform scale, in order to make the figures more readable.

[0067] Furthermore, in the following description, terms that depend on the orientation of the structure such as "upper", "lower", "above", "on", "underlying", apply considering that the structure is oriented as illustrated in the figures.

[0068] DETAILED DESCRIPTION OF PARTICULAR EMBODIMENTS

[0069] An example of a method for producing patterns in a thin layer, in particular a semi-conductor layer, will now be given in connection with FIGS. 1 to 10B.

[0070] We now refer to [Fig.l] which represents a substrate 1 coated with a semi-conductor layer 12 in which it is desired to produce one or more particular patterns intended for example each to accommodate a quantum box also called “quantum dot” (i.e. “Qdot according to English terminology”). The semi-conductor layer 12 can be provided with a thickness typically between 10 nm and 100 nm, for example 16 nm.

[0071] The semiconductor layer 12, for example made of silicon, in particular Si28, is here a so-called “surface” layer of a substrate, in particular of the semiconductor on insulator type such as an SOI substrate (for “Silicon On Insulator”, i.e. “silicon on insulator”), arranged on an insulating layer 11, itself arranged on a support layer 10 made of semiconductor material, for example silicon.

[0072] The semiconductor layer 12 is first coated ([Fig.2]) with a layer 103 called a “hard mask” made of a given material chosen here to be able to resist subsequent etching of the semiconductor layer 12, in particular when it is desired to form the patterns in this layer. The hard mask layer 103 may be made of a dielectric such as, for example, HfO2 or a dielectric derivative of silicon such as, for example, SiO2, SiN, SiCO or a metallic material such as, for example, TiN.

[0073] In the case in particular where a substrate 1 of the semiconductor on insulator type is used, provided with an insulating layer 11 commonly called BOX (for “Burried Oxide”, i.e. buried oxide) and made of silicon oxide, a material capable of being etched selectively with respect to that of the insulating layer IL is preferably chosen.

[0074] Silicon nitride (SiN) is suitable for example for forming a hard mask layer 103 on a surface layer 12 of a SOL substrate. A hard mask layer 103 with a thickness typically between 5 and 100 nm and advantageously between 10 and 40 nm, for example of the order of 20 nm, can be provided.

[0075] A deposition at a temperature above 600°C, for example between 10 and 100 nm, can in particular be implemented to produce this layer. Alternatively, it is possible to provide for a deposition at a lower temperature but sufficient to not excessively reduce the resistance of the material to a subsequent etching process, using HF.

[0076] We then produce (figures 3A and 3B giving respectively a top view and a cross-sectional view of the stack) a masking 115 for implantation on the hard mask layer 103. The implantation masking 115 is typically formed or provided with a layer 114 of photosensitive resin structured by photolithography.

[0077] The implantation mask 115 is here formed with one or more oblong regions 115a, 115b, for example rectangular, here several regions 115a, 115b spaced apart and parallel and which extend mainly in a first direction (direction parallel to the y axis of an orthogonal reference frame [0; x; y; z]), parallel to a main plane of the substrate (the main plane of the substrate being defined here and throughout the description as a plane passing through the substrate and parallel to the plane [0; x; y] given in [Fig. 3A]). The implantation mask 115 here also comprises one or more oblong openings, for example rectangular.

[0078] In order to obtain a better resolution of patterns that one wishes to produce in the hard mask layer 103, a bottom anti-reflective layer 107 (BARC for “Bottom layer Anti-Reflective Coating”), such as an organic layer, can here advantageously be provided between the hard mask layer 103 and the masking layer 115. This layer 107 is then structured so as to extend the opening(s) 116a, 116b of the masking 115 into the anti-reflective layer 107 ([Fig.4]). This structuring is preferably carried out by selective etching with respect to the hard mask layer 103. For example, when the anti-reflective layer 107 is organic (i.e. formed from carbon and elements among the following elements: nitrogen, oxygen and hydrogen) and the hard mask layer 103 is made of silicon nitride, etching using oxidizing or reducing plasmas (derived from O2, N2, H2) can be provided such as H2 / N2 or HBr / O2.

[0079] A first ion implantation is then carried out ([Fig.5]) through the opening(s) 116a, 116b of the masking 115 and, where appropriate, of the layer 107 so as to modify the composition of the hard mask 103 in one or more zones 103A, 103B located opposite the openings of the masking 115.

[0080] Ion implantation can be a so-called “classical” implantation in which ions are transported in a beam, or else a plasma implantation, in particular using a plasma immersion technique.

[0081] The implanted species is intended to modify the hard mask material and is here chosen so that, when the dose of this implanted species is sufficient, selective removal relative to the non-implanted hard mask material or to the hard mask material implanted at a lower dose and in particular below a given threshold is permitted.

[0082] By "selective removal" is meant here that the removal speed, depending on the rate of implanted elements, is much higher for the modified material so that this modified material can be removed without altering or by removing with a selectivity greater than 10:1 compared to the unmodified material.

[0083] By controlling the implantation depth and considering that the concentration of implanted species follows a decreasing gradient with a higher concentration at the surface, it is possible to provide for creating a significant gradient of implanted species in the zones 103A, 103B and to keep at least a lower portion 1031 arranged on and in contact with the underlying semiconductor layer 12 which has not been modified by the implantation, or with too low a concentration of implanted species to allow it to be selectively etched with respect to the unmodified material. A concentration ratio well above 100 can be obtained between, on the one hand, the zones 103A, 103B and, on the other hand, the lower portion 1031.

[0084] An example of simulation of implantation of a layer of Si3N4 obtained with the SRIM tool (“Stopping and Range of Ions in Matter”) and in energy conditions of 250 eV for a Si3N4 material with a density of 3.44 g.cm-3 is given in [Fig. 12] and shows a distribution tail of very reduced width and consequently the possibility of obtaining a very marked interface zone between implanted zones and non-implanted portion.

[0085] An implantation of light ions can be advantageously provided. Here, the term "light ions" means ions originating from atom(s) or molecule(s) whose number atomic number in the periodic table of elements is low and typically less than 10. Light ions such as Helium (He) or Hydrogen (H) ions may be particularly suitable, especially when the hard mask material is silicon nitride.

[0086] Alternatively, other inert species such as argon (Ar) or more reactive species such as nitrogen (N2) may be used, for example.

[0087] The first implantation is carried out here at a first dose chosen here preferably lower than a given threshold dose.

[0088] The implantation can be carried out essentially vertically, that is to say perpendicular to the plane of the layers and to a principal plane of the substrate.

[0089] The possible use of a plasma to carry out the implantation can make it possible to locate the implanted species at depths lower than the minimum depths likely to be obtained by conventional implantation.

[0090] According to a particular embodiment, for a hard mask 103 in particular made of silicon nitride and with a thickness of the order of 20 nm, an H2 plasma can be used according to conditions, in particular power, pressure, exposure time such as described in the doctoral thesis document “Development and characterization of etching processes for Si3N4 and SiCO spacers for 14nm FDSOI technology”, by Maxime Garcia Barros, Micro and nanotechnologies / Microelectronics. Université Grenoble Alpes, 2018.

[0091] The implantation mask 115 and the background anti-reflective layer 107 are then removed (figures 6A and 6B), for example using an oxidizing (02) or reducing or (H2 / N2) plasma.

[0092] Then, a second implantation mask 125 is formed (FIGS. 7A and 7B) on the hard mask layer 103. The implantation mask 125 is typically formed or provided with a layer 124 of photosensitive resin structured by photolithography.

[0093] The second implantation masking 125 is here produced with one or more oblong regions 125a, 125b in particular rectangular, here several regions 125a spaced and parallel and which extend mainly in a second direction parallel to a main plane of the substrate and orthogonal to the first direction in which the regions of the masking 115 extended. The implantation masking 125 here also comprises one or more oblong openings 126a, 126b typically rectangular, which extend in the second direction, orthogonally to the openings of the first masking 115 used previously during the first implantation. Typically, the second masking with implantation 125 is identical, in particular in terms of design, to the first masking 115 but with a 90° orientation of its patterns relative to those of the first masking 115. Advantageously, to obtain identical designs but oriented at 90° to each other between the first masking 115 and the second masking 125, it is planned to carry out the first and second masking by carrying out a first and a second photolithography each time using the same photolithography mask but oriented differently at 90° from one photolithography to the next.

[0094] Here again, a background anti-reflective layer 117 can advantageously be provided between the hard mask layer 103 and the masking layer 125. This anti-reflective layer 117 is then structured so as to extend the opening(s) 126a, 126b of the masking 125 into the anti-reflective layer 117.

[0095] A second ion implantation is then carried out (figures 8A giving a sectional view Y1'Y1, 8B giving a sectional view Y2'Y2, 8C giving a top view) through the opening(s) 126a, 126b of the second masking 125 and where appropriate of the layer 117, so as to modify the composition of the hard mask 103, in one or more zones 103C, 103D located opposite the openings of the masking 125.

[0096] The second implantation, “classical” or by plasma, can advantageously be carried out with the same species as that used during the first implantation. An implantation of light ions, for example using Helium (He) or Hydrogen (H) ions can thus be provided.

[0097] The second implantation is here implemented with a second dose chosen here preferably such that the sum between the first dose used for the first implantation and the second dose is greater than a first threshold which, when exceeded, allows subsequent selective removal of the material modified by implantation. Preferably, the difference between the first dose used for the first implantation and the second dose used for the second implantation is provided to be sufficiently small, and in particular less than a second threshold. This can make it possible to subsequently ensure that selective removal of material having undergone both implantations can be carried out with respect to the material implanted only once. A selectivity of at least 2:1 between a material modified twice by implantation compared to a material modified only once is preferably aimed for.

[0098] A selectivity of at least 10:1 for an implantation-modified material over a non-implanted material may be expected.

[0099] These selectivity constraints depend on the thickness of the implanted hard mask so that the thinner this hard mask is chosen to be, the higher selectivities can be expected. Preferably, a minimum thickness of 5 nm is maintained after implantation and selective removal for transfer into the underlying layers.

[0100] Advantageously, it is possible to provide a dose during the second implantation equal to the dose implanted during the first implantation. In this case, it is therefore possible to advantageously use the same conditions (time, energy, pressure, etc.) for the first and the second implantation step and apply each time half a dose compared to the dose required to make the hard mask material selectively etchable.

[0101] The first and second implantations can thus be implemented with equal durations of exposure to a plasma.

[0102] Alternatively, durations of the two implantation steps may be considered which are not equal as long as the doubly exposed regions have the equivalent of a total dose sufficient to allow them to be etched selectively.

[0103] [Fig.8C] serves to illustrate (according to a top view in which the second masking is not shown) different unmodified portions 103N of the hard mask 103, protected by the masking 115 during the first implantation and by the masking 125 during the second implantation and which were thus not implanted during the aforementioned implantations. Implanted areas 103' once, either during the first or the second implantation, and regions 103” implanted twice, both during the first and the second implantation are also shown.

[0104] The zones 103C, 103D implanted during the second implantation thus extend orthogonally to the zones 103A, 103B, implanted during the first implantation and the common regions 103” between these zones 103A, 103B, 103C, 103D result from the two successive implantations.

[0105] The implantation masking 125 and the background reflective layer 117 are then removed, for example using a plasma and according to a method which may be similar to that implemented previously to remove the first masking 115.

[0106] The doubly implanted 103” regions of the hard mask 103 are then selectively removed (Figures 9A giving a cross-sectional view X' 1X1, 9B giving a cross-sectional view X'2X2, 9C giving a top view). Etching is provided such that the thickness of hard mask 103 consumed depends on the number of implantations undergone and their respective doses. The doubly implanted 103” regions are consumed more quickly than single-implanted or non-implanted areas 103'. The single-implanted areas 103' can also be etched more quickly than portions 103N of non-implanted material. Thus, at the end of an etching carried out at the time and after the lapse of a determined duration, there are portions 103N of the hard mask 103 where the material thickness is greater than that of the zones 103' of the hard mask 103 implanted once and parts where the hard mask 103 is entirely removed.

[0107] The zones 103' exposed to a single implantation have, in this particular example and advantageously, substantially all the same thickness. However, in the case where implantations have been carried out with different exposure times from one implantation to another, there is a risk of having a difference in thickness between im planted by the first iteration and those implanted during the second iteration of the process.

[0108] As can be seen in Figures 9A and 9B, the regions 103” removed from the hard mask 103 leave space for holes 133 revealing the semiconductor layer 12.

[0109] The selective removal of the regions 103” can be carried out by wet etching. This etching can for example be carried out using HF, in particular when the hard mask 103 is made of silicon nitride. HF diluted to 1% can then be used in particular.

[0110] The graph in [Fig.l 1] gives a thickness of silicon nitride hard mask removed by HF after implantation using an H2 plasma, and this as a function of a plasma exposure time. To define the conditions of the first implantation and the second implantation in terms of dose, one method here consists of defining the time necessary for a total removal of the zones 103” of the hard mask layer 103 and to place oneself at half of this duration to define the exposure duration for each implantation.

[0111] Under conditions defined in the aforementioned document “Development and characterization of etching processes for Si3N4 and SiCO spacers for 14nm FDSOI technology”, by Maxime Garcia Barros, it is noted that 90s of H2 implantation plasma allows a removal of 20 nm of the implanted material with 1% HF. Thus, by defining an implantation time of 45s for each implantation step, the entire thickness of a 20 nm SiN hard mask 103 can be removed from the doubly exposed 103” areas.

[0112] The method of removing the modified hard mask 103” used is selective with respect to the material of the semiconductor layer 12. Thus, an HF treatment is for example selective to silicon and therefore does not induce consumption of the semiconductor layer 12.

[0113] A method such as presented above with a SiN hard mask can also be adapted by a person skilled in the art for other types of hard masks, for example made of dielectric materials of the HfO2 type or metallic TiN. We then use charts giving the correlation between implantation time and removal speed to define the removal time necessary to obtain a desired operating point.

[0114] The hard mask 103 is then used to transfer by etching unprotected parts of the underlying semiconductor layer 12 patterns into this semiconductor layer 12.

[0115] A selective etching of the semiconductor layer 12 is thus carried out through the hole(s) 133 of the hard mask 103. Very selective silicon etching methods, in particular those greater than 20:1 with respect to a SiN hard mask, are known. For example, such etching can be carried out using HBr / O2 chemistry to selectively remove silicon with respect to silicon nitride and selectively with respect to silicon nitride. insulating layer 11 (BOX). Thus, with such a method, even if regions 103' implanted only once have a small thickness at this stage, this has only a negligible impact on the semiconductor layer 12.

[0116] Then, the hard mask 103 can be removed (figures 10A and 10B). For this, selective etching is carried out with respect to the semiconductor layer 12 and also preferably with respect to the insulating layer 11 of the substrate 1.

[0117] A treatment using H3PO4 selective to silicon and silicon oxide of the insulating layer 11 of BOX can be carried out for this purpose, for example, in particular when the hard mask 103 is made of silicon nitride. This removal can also be carried out with plasma etching solutions making it possible, from halogenated gases, to obtain very high silicon selectivities and in particular greater than 100:1.

[0118] Removal of the hard mask 103 by BC13 / C12 plasma can be used, for example, when the hard mask 103 is made of HfO2. Wet removal using a SCI (“Standard Clean 1”) type solution can be used, for example, when the hard mask 103 is made of TiN.

[0119] In Figures 10A, 10B, the etched semiconductor layer 12 is shown, once the hard mask 103 has been removed. A “cross-bar” type arrangement is here formed by orthogonal and intersecting semiconductor bars 121, 122. Square-shaped cavities 125 arranged around the semiconductor bars are also produced here.

[0120] Such a semiconductor structure can, for example, make it possible to form a matrix of quantum dots, and in particular a matrix of quantum dots as envisaged, for example, in document US 10,607,993B2 from the applicant.

[0121] A method according to the invention applies in particular to the production of specific patterns in a thin semiconductor layer but can also be implemented to carry out a transfer of patterns in other types of materials, for example metallic or insulating for interconnection applications or any application which cannot be carried out with a single level of photolithography for resolution and / or density constraints.

Claims

Claims

1. Method for producing one or more patterns in a thin layer, the method comprising, in this order, the following steps: -Providing a hard mask (103) on the thin layer (12), then; -Forming on the hard mask (103) a first implantation mask (115), the first implantation mask being provided with one or more first openings (116a, 116b) which extend mainly in a first direction and reveal the hard mask (103), then; - Carry out a first implantation of a given species in the hard mask (103), -Remove the first masking (115), -Forming on the hard mask (103) a second implantation mask (125), the second implantation mask being provided with second openings (126a, 126b) revealing the hard mask (103) and which extend mainly in a second direction orthogonal or substantially orthogonal to the first direction, the second openings being arranged so as to reveal one or more given regions (133”) of the hard mask (103) already revealed by the first openings during the first implantation then; - Performing a second implantation in the hard mask (103) through the second openings of the second implantation mask, the second openings (126a, 126b) being arranged so as to reveal one or more given regions (103”) of the hard mask already revealed by the first openings during the first implantation; - Removing the second mask, -Selectively removing said given regions (103”) of the hard mask (103) which have been implanted both during the first implantation and during the second implantation, with respect to portions (103N) of the hard mask (103) which are not implanted and / or areas (103') implanted only during the first implantation and / or only during the second implantation, so as to form an openwork hard mask comprising one or more patterns, -Transfer said patterns into the thin layer (12), by etching the thin layer (12) covered with the openwork hard mask.

2. Method according to one of the preceding claims, in which the first implantation is carried out with a first dose and in which the second implantation is carried out with a second dose, - the sum of the first dose and the second dose being greater than or equal to a first predetermined threshold, - the difference between the first dose and the second dose being less than or equal to a second predetermined threshold.

3. Method according to one of the preceding claims, in which the first implantation is carried out with a first dose and in the second implantation is carried out with a second dose equal to the first dose.

4. The method of claim 3, wherein the first implantation is performed by plasma and according to a first duration of exposure to the plasma and wherein the second implantation is performed by plasma and according to a second duration of exposure to the plasma identical to the first duration.

5. Method according to one of the preceding claims, wherein the second implantation is carried out using said given species used for the first implantation.

6. A method according to one of the preceding claims, wherein the first implantation and the second implantation are provided such that the areas (103') implanted only during the first implantation and only during the second implantation are made selectively etchable with a selectivity of at least 10:1 with respect to non-implanted portions (103N) of the hard mask (103) and that said given regions (103”) of the hard mask (103) which have been implanted both during the first implantation and during the second implantation are made selectively etchable with a selectivity of at least 2:1 with respect to the areas (103') implanted only during the first implantation and only during the second implantation.

7. Method according to one of claims 1 to 6, in which the first masking (115) to be implanted is formed of portions producing a first pattern oriented in a given direction parallel to the main plane of the substrate, the second masking (125) to be implanted being formed of portions producing a second pattern, identical to the first pattern and oriented orthogonally to said given direction.

8. The method of one of claims 1 to 7, wherein the first implantation mask (115) is formed by a first photolithography using a photolithography mask and wherein the second implantation mask (125) is formed by a second photolithography using said photolithography mask oriented at 90° relative to its orientation during the first photolithography.

9. Method according to one of claims 1 to 8, in which the thin layer is a semiconductor layer (12), in particular made of silicon.

10. Method according to claim 9, in which the semiconductor layer (12) is formed on a semiconductor-on-insulator substrate or is the surface layer of a semiconductor-on-insulator substrate (1), the substrate being provided with an insulating layer (11) called a "BOX" made of silicon oxide.

11. Method according to one of claims 1 to 10, the given species being light ions, in particular hydrogen or helium.

12. Method according to one of claims 1 to 11, in which the hard mask (103) is formed from a monolayer of a single given material.

13. Method according to one of claims 1 to 12, wherein the hard mask layer (103) is formed from one of the following materials: SiN, SiCO, HfO2, TiN.

14. Method according to claim 13, in which the hard mask (103) is made of SiN, the transfer of said patterns into the semiconductor layer (12) being carried out by means of etching, in particular by plasma.

15. Method according to one of claims 1 to 14, further comprising, after transfer of said patterns from the hard mask (103) into the thin layer (12): removing the hard mask (103) by selective etching with respect to the thin layer (12).

16. Method according to one of claims 1 to 15, in which the first masking (115) and the second implantation masking (125) are formed by: - ​​producing a stack comprising at least one anti-reflective layer (107, 117) coated with at least one layer (114, 124) of photosensitive resin - structuring elements in the layer of photosensitive resin, - reproducing said elements in the anti-reflective layer.

17. Method according to one of claims 1 to 16, in which among said patterns transferred into the thin layer (12) there is at least one square-shaped pattern and / or at least one pattern formed of two bars joined and orthogonal to each other.

18. Method for manufacturing a quantum device comprising implementing a method according to one of the preceding claims, the transferred patterns being intended to accommodate one or more boxes quantum.

Images