EP2656143A1

EP2656143A1 - Nanoimprint lithography method

Info

Publication number: EP2656143A1
Application number: EP11804568.1A
Authority: EP
Inventors: Sébastien PAULIAC
Original assignee: Commissariat a lEnergie Atomique CEA; Commissariat a lEnergie Atomique et aux Energies Alternatives CEA
Current assignee: Commissariat a lEnergie Atomique et aux Energies Alternatives CEA
Priority date: 2010-12-22
Filing date: 2011-12-21
Publication date: 2013-10-30
Also published as: WO2012085164A1; US20140004313A1; US9305783B2; JP5916753B2; FR2969772B1; FR2969772A1; JP2014515874A

Abstract

The invention relates to a nanoimprint lithography method, comprising: a step of pressing a mould (130, 50) in a photosensitive resin (120) in order to form at least one imprint pattern (127) defined by a stamped area (129) and an adjacent area (128), said adjacent area (128) being less stamped or not stamped at all, and being thicker than the stamped area (129); and a step of exposure to a certain amount of sunlight, characterised in that the respective thicknesses of said two areas are defined such that said two areas absorb a different amount of the sunlight and in that the amount of sunlight provided by the exposure step is predetermined so as to be great enough to activate the resin in whichever of the two areas has the greater absorption, and so as not to be great enough to activate the other of said two areas.

Description

Nanoprint lithography process »

TECHNICAL FIELD OF THE INVENTION

The present invention generally relates to lithography methods. It receives for privileged application lithography processes used by the microelectronics industry for the manufacture of semiconductor devices, including integrated circuits. It relates more particularly to an improved method of nanoscale lithography.

STATE OF THE ART

The industrial manufacturing of new generations of integrated circuits involves the ability to engrave ever smaller size patterns, which are now measured in nanometers (nm = 10 ^"9 meters) only, photolithography used from the outset, based on insolation. of photosensitive resins through optical masks reproducing the patterns to be engraved, however, faces physical barriers that require the use of increasingly sophisticated techniques to accompany the growth of the desired integration density. to limit the diffraction of light through masks shorter wavelengths (ultraviolet or even x-rays) and complex techniques (for example, immersion lithography) which require considerable investment in their development and industrial implementation are required.

Since the mid-1990s a very different technique, which allows to completely overcome the problems of diffraction mentioned above, was invented by Professor Stephen Y. CHOU in the laboratory of nanoscale structures of the university from Minnesota to the United States. The initial principle of this technique called "nanoscale lithography" has been unveiled by the latter in several publications, including the one entitled "Nanoimprint Lithography", published with his collaborators Peter R. Krauss, and Preston J. Renstrom in "Journal of Vacuum Science and Technology »reference B 14 (6), Nov./Dec. This technique immediately attracted a lot of interest and was at the origin of many research and development projects. Nanoscale lithography is now part of the international technology roadmap for semiconductors (ITRS), and more specifically for integrated circuit technologies under development or industrialization phase. the basic functional element, the node, has been defined by the waybill successively at 32 nm and at 22 nm.

Nanoscale lithography consists of two main variants. The first, originally proposed by Professor Chou, thermal nanoimprint lithography, usually referred to by its acronym T-NIL for thermal nanoimprint lithography, consists of printing, with an opaque mold, monomers or heated thermoplastic polymers. After cooling the mold can be removed, the printed patterns remain in place.

The second technique, nanoprinting with photosensitive resin, usually designated by its acronym P-NIL for "photo-curable nanoimprint" consists of printing a photosensitive resin with a transparent mold and to perform an optical insolation of the resin film at through this one. Insolation causes the hardening of the resin. As above, the mold can then be removed.

In both cases, however, there remains a residue at the bottom of the nano printed patterns that must be removed to allow their transfer to the substrate that is desired to burn. The implementation of lithography by nanoscale printing thus requires currently also having to perform a reactive ion etching, usually designated by the acronym RIE for "reactive ion etching", in the presence of oxygen to remove residues remaining at the bottom of the nano-printed trenches. Another way is to perform a post etch step in which a controlled thickness of material is removed chemically. This step is usually referred to by its English word "etch-back".

These known techniques for removing the residue present in the background of the nano printed patterns are relatively complicated, time consuming and expensive to implement.

The object of the invention is to provide an improved nanoscale lithography printing method which solves at least one of these problems.

SUMMARY OF THE INVENTION

The subject of the invention is thus a nanometric printing lithography method comprising a preparation step during which a photosensitive resin is placed on a substrate, a step of pressing a mold in the resin to form in the resin at I have a reason for impression. The printing pattern is defined at least in part by two areas including a stamped area and an area adjacent to said stamped area, said adjacent area being less or not stamped and having a thickness greater than that of the stamped area.

The method further comprises a step of exposing at least one of said two zones to an insolation dose. In other words, both areas receive the insolation dose during this exposure step.

Typically, the respective thicknesses of said two zones are defined so that said two zones have an absorption differential of the exposure dose and the exposure dose provided by the exposure step is determined so as to be large enough to activate the resin at that one of said two zones which has the highest absorption and so that it is not large enough to activate the resin at that of said two zones which has the lowest absorption. In other words, the thicknesses of said two zones are defined so that, in order to be activated, the resin at one of said two zones requires an insolation dose different from the exposure dose necessary to activate the resin at the level of the other of said two zones and the exposure dose provided by the exposure step is determined so as to be large enough to activate the resin at only one of said two zones.

Thus, the resin thicknesses and the insolation dose provided by the exposure step are determined so that the insolation dose provided is between the dose required to activate the zone with the highest absorption and the dose required to activate the zone with the lowest absorption.

Thus the invention takes advantage of the variation of the absorption of the resin film as a function of the thickness of this film. This absorption variation, usually considered as a serious disadvantage, is used in the context of the invention to selectively activate the resin at the level of the pattern or the area that surrounds it.

By using a positive resin, it is for example possible to activate the resin only at the level of the pattern to remove the residue after development of the resin. Similarly, by using a negative resin it is then possible to activate the resin only outside the pattern to remove the residue after development of the resin.

The invention thus makes it possible to eliminate the resin in the bottom of the patterns in a particularly precise and simple manner. One can indeed do without the usual steps of RIE or post engraving mentioned above.

In addition, the process for removing the residue according to the invention makes it possible to obtain a very good resolution of the patterns obtained by nanoimprinting. Indeed, the insolation and development stages of the resin make it possible to keep the rib of the nano printed patterns, unlike the steps usually used to remove the residue, which can alter the flanks of the nano-printed patterns. In addition, these techniques tend to degrade the resin.

In a particularly advantageous manner, the invention also makes it possible to obtain, after development of the resin, a final pattern that is the reverse of that obtained by pressing the mold in the resin. This final pattern corresponds to the relief of the mold.

In fact, with a positive resin, by choosing resin thicknesses such that the adjacent zone has an absorption greater than that of the strongly embossed zone constituting the background of a pattern, the exposure makes it possible to activate only the adjacent zone by making it soluble during development. After development, the adjacent area is removed and the resin of the bottom of the pattern, which she has not absorbed a sufficient dose remains in place. An inverse photo of the patterns obtained by printing is then obtained. Likewise, with a negative resin, by choosing resin thicknesses such that the adjacent zone has a lower absorption than that of the strongly embossed zone constituting the bottom of the pattern, the exposure makes it possible to crosslink the resin at the bottom of the pattern. only. During development, the adjacent area is removed and the bottom resin pattern, which has not absorbed a sufficient dose, remains in place. As will be detailed later, it is thus possible to easily obtain final patterns in projection corresponding to protrusions protruding from the mold. Advantageously, these protruding patterns may be narrow and may form, for example, lines.

Generally in the context of the present invention, the patterns in the resin are recessed or projecting. Preferably, they are obtained by nano printing. The reliefs of the mold may also be hollow or protruding.

Optionally, the method according to the invention further comprises at least any of the following characteristics:

the resin thicknesses are determined so that the difference between the dose required to activate the zone with the lowest absorption and the dose required to activate the zone with the highest absorption is at least 5mJ / cm ² , for example 10mJ / cm ² . Thus, if for a given resin the resin thickness of a zone requires a dose of 15 mJ / cm ² , the thickness of the adjacent zone should be such that for this thickness the minimum dose required for the activation of the resin about 20 mJ / cm ² . The exposure dose provided by the exposure step should therefore be greater than or equal to 15 mJ / cm ² and less than 20 mJ / cm ² . Preferably, contrast curves are defined to determine these thicknesses.

Preferably, the adjacent zones delimiting the same pattern formed during the pressing step of the mold receive the same dose of insolation.

Advantageously, the absorption of the resin as a function of its thickness defines a substantially sinusoidal curve and in which the thickness of the resin at one of said stamped zone or said adjacent zone corresponds to a maximum of said sinusoidal curve and the thickness of the resin at one of said stamped area or said adjacent area corresponds to a minimum of said sinusoidal curve.

According to a first embodiment, a final pattern corresponding to the relief of the mold is obtained. This final pattern is therefore the inverse of the pattern obtained by the printing step.

According to an alternative of this first embodiment, the resin is a positive photosensitive resin, the thicknesses of the resin at the level of the stamped zone and at the level of said adjacent zone are determined so that the resin at the level of the stamped zone presents a lower absorption than that of the resin at said adjacent zone and wherein the exposure dose provided by the exposure step is set so as to activate the resin at said area adjacent and not to activate the resin at the stamped area, so as to obtain a final pattern reverse the printing pattern. Preferably, the setting of the thickness of strongly pressed resin corresponds to a minimum on the light energy absorption curve, the adjustment of the thickness of the resin with little or no stamping corresponds to a maximum on the absorption curve of luminous energy.

According to another alternative of this first embodiment, the negative photosensitive resin, the thicknesses of the resin at the level of the stamped zone and at the level of said adjacent zone are determined so that the resin at the level of the stamped zone has a higher absorption. to that of the resin at said adjacent area and wherein the exposure dose provided by the exposing step is set so as to activate the resin at the stamped area and not to activate the resin at the of said adjacent area, so as to obtain an inverse final pattern of the printing pattern. This final pattern also corresponds to the relief of the mold. Preferably, the setting of the thickness of strongly pressed resin corresponds to the maximum on the light energy absorption curve, the adjustment of the resin thickness with little or no stamping corresponds to a minimum on the absorption curve of luminous energy.

Thus, thanks to the inversion, the invention makes it easy to obtain a final projecting pattern, such as a narrow line for example. In addition, the dimensions of this final projecting pattern can be very small and precisely controlled. However, with the known nano-printing processes, obtaining protruding patterns is particularly delicate. Indeed, their obtaining requires the presence of a recessed relief in the mold and it is very difficult to marry the resin form of a hollow relief of the mold. The presence of air in the hollow relief of the mold makes the obtaining of narrow projecting motifs even more delicate.

Optionally, after the development of the resin, an additional etching step is carried out to remove a residue of resin possibly remaining on the substrate at the level of the adjacent zone after the development step. Typically, these post-etch steps are RIE or etchback type steps.

According to a second embodiment, the resin residue present in the background of the pattern obtained by the printing step is removed.

According to an alternative of this second embodiment, the resin is a positive photosensitive resin, the thicknesses of the resin at the level of the stamped zone and at the level of said adjacent zone are determined so that the resin at the level of the stamped zone presents Absorption greater than that of resin at said adjacent area and wherein the exposure dose provided by the exposing step is set to activate the resin at the drawn area and not to activate the resin at said adjacent area so as to remove the resin residue at the stamped area, i.e., typically in the bottom of the printing pattern. Preferably, the setting of the strongly pressed resin thickness corresponds to a maximum on the light energy absorption curve, the adjustment of the resin thickness with little or no stamping corresponds to a minimum on the absorption curve of 'luminous energy.

According to another alternative of this second embodiment, the resin is a negative photosensitive resin, the thicknesses of the resin at the level of the stamped zone and at the level of said adjacent zone are determined so that the resin at the level of the stamped zone has a lower absorption than that of the resin at said adjacent area and wherein the exposure dose provided by the exposing step is set to activate the resin at said adjacent area and not to activate the resin at the stamped area, so as to remove the resin residue at the stamped area, i.e. typically in the bottom of the printing pattern. Preferably, the setting of the thickness of strongly pressed resin corresponds to a minimum on the light energy absorption curve, the adjustment of the thickness of the resin with little or no stamping corresponds to a maximum on the absorption curve of luminous energy.

Opposite or significantly different opening rates are obtained for two areas of the same plate.

In a first embodiment, after the pressing step, a plurality of printing patterns having different thicknesses are obtained, at least one of these thicknesses corresponds to an absorption maximum, and at least one other of these thicknesses corresponds to a minimum of absorption. More generally, these thicknesses correspond to different levels of absorption. Thus, by exposing all of the resin, both resin residues in the background of printing patterns can be removed and both an inverse image of other printing patterns can be obtained.

Advantageously, a full plate exposure is performed.

To obtain, after the pressing step, printing patterns having variable thicknesses, it is possible for the mold to have protruding reliefs of different heights.

The invention is not limited to a single pressing step to obtain resin zones of different thickness on the same substrate. Advantageously, the method also comprises a step of removing the mold after the pressing step. Preferably, the exposure is performed after removal of the mold. In an alternative embodiment of the invention, it can be performed before removal of the mold, the latter then being configured to pass, partly to the self, the insolation dose. In this variant, the mold is preferably substantially transparent.

In another embodiment, alternative or combined with the first embodiment, resin portions are insulated with different insolation doses. The exposure is thus carried out in a non-homogeneous manner over the entire plate. These differences in exposure can be obtained using a mask blocking part of the exposure.

Preferably, at least one first unit having a first dimension is insulted with a first dose of insolation. Said dimension is taken in a direction normal to the thickness of the resin is typically corresponds to the width of a trench or step formed in the resin. At least one second unit having a second dimension smaller than said first dimension is insulted with a second dose of insolation greater than said first dose of insolation. Specifically, performs the exposure step so that the first dose of insolation is sufficient to activate only one of the swaged area or the less or unsworn area of the first pattern and so the second dose of Insolation is insufficient to activate the second pattern but is sufficient to activate the area bordering the second pattern. The second pattern may be a trench, in which case the areas bordering the pattern are areas with a greater thickness of resin. The second pattern may also be a projection, in which case the areas bordering the pattern are areas having a lower resin thickness. According to an alternative to this embodiment, during the exposure step all the resin is exposed to the insolation dose. The invention thus allows a full plate exposure, particularly advantageous in terms of cost of speed. The insolation dose is provided by a coherent light source which makes it possible to generate interference phenomena in the resin film. This makes it possible to generate the absorption differential that the invention makes use of.

Preferably, the exposure step successively involves several light sources having different wavelengths so as to increase the absorption differential. Preferably, during the preparation step, a step is provided during which the photosensitive resin is deposited on a layer or a substrate making it possible to amplify the absorption variations of the resin as a function of its thickness. Typically said layer or said substrate is taken from the following materials SiC, Ge, Ag, W, AlSi. Alternatively, to achieve the same objective of amplification of absorption variations, a silicon substrate can be provided.

The subject of the invention is also a multilayer assembly comprising a substrate coated with a layer of photosensitive resin, the resin having at least one pattern delimited at least in part by two zones including a stamped zone and an area adjacent to said stamped zone. . The thickness of each of the two zones corresponds to a maximum or a minimum of an absorption curve of said resin as a function of its thickness.

More generally, the thickness of each of the two zones corresponds to activation thresholds at least 5mJ / cm ² apart. Thus, the minimum dose to activate one of the zones is at least 5mJ / cm ² lower than the minimum dose to activate the other zone, for example less than 10mJ / cm ² .

BRIEF DESCRIPTION OF THE FIGURES

Other characteristics, details and advantages of the invention will emerge more clearly from the detailed description given below as an indication in relation to drawings in which:

FIG. 1 illustrates the steps of an exemplary nanoscale lithography process according to the invention.

FIGURE 2 illustrates the dependence of insolation parameters on the thickness of the deposited resin layer.

FIG. 3 illustrates with examples the behavior of the photosensitive resins as a function of parameters including the dose of insolation received and the size of the exposed patterns.

FIG. 4 describes four variants for implementing the invention, with a positive and negative resin, and by matching the two thicknesses of resins obtained after printing to different levels of absorption of the insolation energy, typically either at a maximum or at a minimum of absorption. FIGURE 5 illustrates the influence of the substrate or the material located under the resin layer for the implementation of the invention.

FIGURE 6 illustrates different zones of lithography, those where narrow trenches must be opened in the resin, and others where only narrow lines will remain.

FIG. 7 describes an example of application of the invention in which a variable topography mold is used, that is to say having reliefs of variable heights.

FIG. 8 describes another example of application of the invention in which the dose is varied as a function of the type of patterns of the zone to be irradiated.

FIG. 9 illustrates an example of a stack of layers modeled to determine the absorption curves of the resin as a function of its thickness.

FIGURE 10 illustrates an absorption curve of a resin layer as a function of its thickness.

Figures 11 to 11 e illustrate an example of a method according to the invention for performing a pattern reversal.

FIGURES 12a and 12b are examples of curves for determining the contrast of a negative and a positive resin respectively.

The accompanying drawings are given by way of example and are not limiting of the invention.

DETAILED DESCRIPTION OF THE INVENTION

Figure 1, which includes Figures 1a to 1e, illustrates the steps of the improved nanoscale lithography printing method of the invention.

On the substrate 1 10 where, on the surface, it is desired to reproduce and etch patterns which will contribute to the production of a device during manufacture, a layer of a photoresist 120, for example of the type used, is deposited. in a standard way by the microelectronics industry for optical lithography. The invention makes no assumption about the type of substrate from which the method of the invention is implemented. In particular, the substrate may for example already have on the surface numerous layers (not shown) in which patterns may already have been previously defined, with the method of the invention, or by other means, in particular with the aid of classical optical lithography or by electronic lithography.

As shown in FIG. 1a, the first step 101 therefore consists in depositing on the surface a layer of resin preferably controlled in thickness. The deposit can be done by any standard means implemented by the microelectronics industry. Most often, in this case, by centrifugation, method often designated by its English term "spin-coating". The thickness of the deposited layer is controlled by adjusting the speed of rotation as a function of the viscosity of the resin. After spreading, the resin generally undergoes a heat treatment to evacuate the solvent residues and mechanically stabilize the resin. This treatment may for example be of the type usually designated by its English word soft bake for soft cooking.

As shown in FIG. 1b, the following step 102 consists in pressing into the resin a mold 130 having reliefs 132. The pressing of the mold 130 causes the reliefs 132 to penetrate into the resin 120, which makes it possible to transfer these reliefs 132 to forming nano-printed patterns in the resin. Advantageously, the mold can be applied over the entire surface of the substrate and can therefore reproduce all the patterns of all devices produced simultaneously on a plate made for example of a semiconductor material. Typically, the plate is silicon. It can be of very large size, for example several tens of centimeters, with regard to patterns of nanometric sizes to reproduce. For simplicity, without in any way impeding the understanding of the method of the invention, only one of these protruding reliefs 132 is represented while a very large number, typically hundreds of thousands, may actually have been shaped on the bottom surface of the mold. The mold can be made of an opaque material, transparent or partially transparent.

As will be seen later the height 131 of the protrusions 132 projecting from the mold and / or the thickness 121 of the deposited resin layer are important parameters for controlling the implementation of the method of the invention.

In the present application, relief height or thickness ee _fj e ₀ , e _i; e ₂ , e _3j resin 120, dimensions taken in directions substantially perpendicular to the main plane of the substrate and / or substantially parallel to the direction of penetration of the mold 130 in the resin 120.

Preferably, while the mold 130 is pressed into the resin 120, and remains in place, the substrate 1 10, which rests on a support (not shown), is heated to facilitate printing by making the resin more malleable: for this purpose uses a temperature around the glass transition temperature of the resin. The heating temperature must be such that it does not affect the photosensitive qualities of the resin used. In particular, in the case of a so-called positive resin, the heating temperature must remain below the so-called deprotection temperature of the latter. In the case of a so-called negative resin, the heating temperature must remain below the crosslinking temperature. Depending on the case, positive and negative resins are indeed commonly used in lithography so that the parts Exposed to light become, respectively, soluble or insoluble after insolation, allowing to reproduce the patterns of the masks where their negatives.

In the following step 103, as shown in FIG. 1c, the mold 130 can then be removed. The printed patterns 127 remain in place in the resin layer 120.

The next step 104 is illustrated in Figure 1d. An insolation of the printed resin 120 is then carried out. Preferably, this insolation is carried out on the entire surface of the resin. This full plate insolation simplifies and speeds up the patterning process. According to one variant, only a portion of the resin is insoluble. This localized exposure can be obtained using a mask partially obscuring the insolation of the resin. The insolation provided to the resin, at least at some of the patterns is referenced 140 in Figure 1 d. It is clear from this figure that the adjacent zones 128, 129 delimiting a pattern 127 receive the same dose of insolation during the exposure step.

The invention is based on the observation that the behavior of the resin can be very different at the end of the insolation phase depending on its thickness. The behavior after insolation depends on the insolation dose absorbed. However, the absorbed dose depends on the absorption capacity of the resin which itself depends on the thickness of the resin. For a given reason, two thicknesses are to be considered. That of the resin which has been significantly stamped by the mold. This area 129 corresponds to the relief 132 projecting from the mold 130, that is to say, _r e 124; and the greater thickness of the resin, where it has only little or not been stamped by the reliefs 132 of the mold 130. This zone 128 corresponds to the hollows generated by the relief 132 of the mold 130. This zone 128 is designated subsequently adjacent area 128 to the pattern. Its thickness is referenced: e _f 122 in FIG.

Thus, if a mold has stepped reliefs, a first zone adjacent to a stamped zone may itself constitute a stamped zone delimited by a second adjacent zone that is less or not embossed than the first. This is the case of the reliefs 52 and 54 illustrated in FIG. 5 and described later.

In the present invention, it will be described as an em area, deformed, compressed or compressed zone and less or not stamped, deformed, compressed or compressed to characterize the (or) difference in thickness induced by the penetration of the mold into resin. This penetration of the mold in the resin generates at least two adjacent zones, one having a thickness greater than that of the other zone. Thus, the present invention covers both elastic deformations inelastic resin, that is to say, the deformations with or without significant compression of the resin.

In the case where the adjacent zone is not stamped by the mold, its thickness substantially corresponds to the resin thickness deposited during the first step 101. If the total surface of the reliefs is large, there can be a significant reflux of the resin in the poorly stamped areas and therefore an increase in the initially deposited resin thickness. The thicknesses should be chosen accordingly, depending on the density and size of the patterns. Preliminary tests will advantageously be carried out to determine the effective thicknesses after pressing which are those which are important for the choice of doses.

As will be seen in detail in the description and the figures which follow, the insolation dose provided during the exposure phase may be such, by adjusting the thicknesses e _f and e _r , that the thicker parts remain or become effectively soluble during development phase whereas, respectively, the compressed parts become or remain insoluble depending on the type of resin used, that is to say negative or positive.

This makes it possible to obtain with the method of the invention, for example, the result illustrated in FIG. 1e or 1e 'at the end of step 105 of developing the resin after exposure. In the case illustrated in FIG. 1e, the resin residue located in the bottom of a matrix 27 absorbs an insolation dose which causes it to shrink after development while the resin adjacent to pattern 127 remains in place.

In the case illustrated in FIG. 1 e ', a transfer is obtained in the resin of the patterns 126 which correspond to the reliefs 132 of the mold, whereas it is the opposite result which is obtained with the standard method where the said RIE etching step in the chapter on the state of the art instead makes the resin that has been stamped, where it is the least thick 124.

Thus, by performing an inversion, it is easy to obtain a final projecting pattern. In addition, the dimensions of this final projecting pattern can be very small and precisely controlled. However, with the known nano-printing processes, obtaining protruding patterns is particularly delicate.

Figure 2, which is comprised of Figures 2a and 2b, illustrates the above-mentioned dependence of the insolation parameters on the thickness of the deposited resin layer.

The deposited resin layer 120, together with the underlying substrate 110, constitutes a semi-transparent and semi-reflective optical system of the Fabry-Perot interferometer type. The behavior of the layer for the insolation operation is then depending on its thickness. Indeed, the interference phenomena that appear in the resin film cause a variation in the energy absorbed by it. As a result, the optimal insolation dose, which makes it possible to transform the chemical structure of the resin so that it becomes soluble or insoluble for the subsequent development phase, varies according to its thickness. Diagram 210 is an example of experimentally determined characteristic data that shows this dependency. It is in this example a negative resin whose commercial reference is indicated 212. On the ordinate, we find the exposure dose necessary for the chemical transformation of the exposed resin. In the case of a negative resin, this energy dose, expressed here in milli joules per square centimeter, causes its crosslinking so that it becomes insoluble. The optimum dose for obtaining this result is most often designated by the English word "dose-to-size" 214, that is to say optimal dose that allows to obtain after development the nominal size of the exposed patterns. Curve 218 shows the dependence of the optimal dose as a function of the thickness 216 of the resin. This curve which is cyclic, typically sinusoidal, has a series of minima and maxima whose repetition period depends on the wavelength of the coherent light source used, 248 nm in this case. The insolated patterns are 9 mm squares.

This phenomenon of variation of the absorption of a resin film can also be calculated using the Fabry-Perot interferometer model already mentioned above. Diagram 220 shows the result of a simulation of the normalized 0-1 absorption 222 of a resin film based on its thickness 224 from the resin optical data provided by the manufacturer. This simulation was carried out under conditions similar to those of the diagram 210 which makes it possible to compare the experimental curve 218 and the calculated curve 226 and to note, for example for a thickness of 200 nm, that the minimum of absorption of the curve 226 corresponds well on the curve 218 to a maximum of crosslinking dose to bring to the resin to obtain its activation. Indeed, the lower the absorption, the higher the insolation dose to obtain the same result. It is therefore expected that a minimum of absorption corresponds to a maximum of the "dose-to-size" to be applied.

This significant variation in the optimum dose to be applied depending on the thickness of resin deposited is unanimously considered as a serious drawback by the skilled person. To overcome this problem, it is often resorted to the deposition of additional layers (such as those called BARC of the English "bottom anti reflective coating" for anti-reflective coating located under the resin) for prevent or minimize any reflection of the substrate by depositing thereon, prior to the resin layer; this layer will not reflect the incident light and attenuate the amplitude of the sinusoids 218. Many other techniques such as the deposition of an anti-reflective surface coating usually referred to as "top anti reflective coating" have been developed to reduce the undesirable consequences of the absorption variation.

On the contrary, the invention takes advantage of this phenomenon to propose the method described in FIG. 1, a method that can be implemented in four different ways as explained in FIG. 4 below.

Prior to this description, FIGS. 3a and 3b give additional details on the behavior of the photosensitive resins as a function of parameters such as the thickness of the resin deposited, the dose of insolation received and the size of the exposed patterns and which are useful. to the understanding of the process of the invention.

Diagram 230 shows an example of an experimental determination of a dose window 232 which produces inverse results after exposure according to the resin thickness considered. It can be seen for example on its curves, called contrast curves, that a dose of 15 mJ / cm ² , situated in the middle of the window 232, will be suitable for selectively activating the negative resin in question (NEB22A2), if it has a thickness of 172 nm or 235 nm 234, thicknesses for which the absorption is strong. On the other hand, it will not activate thicknesses of 208 nm or 270 nm 236, thicknesses for which absorption is low. The entire range of doses included in the window 232 is likely to be suitable. In this example the curves are established for square patterns of 9 mm side.

Another very important parameter that determines the choice of doses to be applied is the size of the patterns. Diagram 240 represents, on the ordinate, the evolution of the dose required for the activation of the resin, usually designated by its English term "dose to size", expressed in milli joules per cm ² depending on the size of the exposed patterns expressed in microns, that is to say 10 ^"6 meter. the two curves correspond to two resin layers, one where absorption is high 244, the other in which the absorption is low 242. of course, the dose -to-size to apply is more important for resin thicknesses where absorption is lower.

FIG. 4, which comprises FIGS. 4a to 4e, describes four variants for implementing the invention, with a positive and negative resin, and by matching the two thicknesses of resins obtained after printing to either a maximum or at least a minimum of the sinusoidal curve of absorption of the insolation light energy by the resin layer.

In order to facilitate the disclosure of the invention, in all the examples which follow, the resin thicknesses correspond to either a maximum or a minimum of absorption. The invention is however not limited to resin thicknesses corresponding to extremums. It encompasses any processes involving resin thicknesses having sufficient absorption differences to selectively activate the resin at the compressed zone or at the adjacent area less or not compressed.

FIG. 4a shows the resin layer printed at the end of step 103 of the method as described in FIG. 1. At this stage, four alternative embodiments are possible, which are described below in the figures. 4b to 4e.

FIG. 4b illustrates a first variant in which the resin used is positive and where an inversion of the nano-printed patterns 127 is obtained, that is to say a transfer in the resin of the protruding reliefs 132 of the mold 130 as described in FIG. 1st. To obtain this result, that is to say to obtain the patterns 126, it is necessary that the thickness of the pressed resin e _r is adjusted to a minimum of absorption 420 of the sinusoidal curve described in FIG. 2b. At the same time, it is necessary that the thickness of the resin with little or no stamping e _f by the reliefs of the mold is set to an absorption maximum 410 of the sinusoidal curve. More generally, it is necessary that the resin spacer in the bottom of the pattern corresponds to a significantly lower absorption than that in the area adjacent to the pattern. Thus, by setting the optimal dose of insolation or "dose-to-size" on this absorption maximum 410 is not provided to the most stamped resin parts a dose sufficient to transform them chemically. In the case of a positive resin, the dose is then insufficient to render it soluble for development and the patterns 126 remain in place for the subsequent substrate etching operation. As already noted in FIG. 1, this first way of proceeding makes it possible to obtain a transfer of the protrusions 132 projecting from the mold 130 contrary to a standard operation of nanometric printing lithography where it is the compressed parts of the resin, those which in this case cases are generally referred to as residues, which are removed by an engraving operation RI E that follows. This first implementation of the invention instead advantageously uses these parts most stamped or residues to perform a pattern inversion.

FIG. 4c describes a second variant of implementation which makes it possible to obtain, with a positive resin, the opposite result. In this case, as with a standard operation of nanoscale lithography, it is the unpressed parts 128 of the resin, those which correspond to the reliefs 132 forming a hollow in the mold, which remain in place. This result is obtained by adjusting the thickness of the pressed resin e _r on an absorption maximum 410 of the sinusoidal curve. At the same time, the thickness of the unswashed resin e _f by the reliefs of the mold must be set to a minimum of absorption 420. More generally, it is necessary that the resin thickness in the bottom of the pattern corresponds to an absorption significantly greater than that in the area adjacent to the pattern. Thus, as above, by setting the optimum dose of insolation or "dose-to-size" on this maximum of absorption 410 it is to the non-stamped resin parts that this time a dose is not added. sufficient to transform them chemically. The resin being positive, it is initially insoluble, and adjacent areas 128 less absorbent will therefore remain in place during development.

It will be noted that this second embodiment makes it possible to eliminate the stamped parts or residues without using an RI E etching as is necessary in a standard operation of nanometric printing lithography. Particularly advantageously, the invention makes it possible to preserve the rib of the patterns and thus offers an improved resolution compared with existing methods involving a subsequent etching step during which the sides of the patterns 127 can be significantly degraded during etching.

Figures 4d and 4e are the dual figures of the two previous figures. They respectively describe the third and fourth implementation variants of the invention using this time a negative resin. What has been said for Figures 4b and 4c applies. Only the result obtained is reversed because of the use of a negative resin, which is therefore initially soluble, and some parts of which are made insoluble by exposing them to an optimal dose of light determined by a maximum of 410 of the sinusoidal curve. absorption 226.

Thus, with a negative resin for FIG. 4d, the resin located in the bottom of the patterns 127, that is to say here the resin of the embossed zone 129, has a height such that its absorption is lower than the absorption of the adjacent zone 128 in the pattern 127. The exposure is thus carried out so that:

the resin situated on said adjacent zone 128 absorbs a sufficient dose when it is activated. It remains in place after development.

- The resin located in the bottom of the pattern 127 absorbs a dose not sufficient to its activation. It will be removed during development. The invention thus makes it possible with a negative resin to remove the residues at the bottom of the patterns without resorting to the existing stages of RIE or post etchback.

On the other hand, with a negative resin for FIG. 4e, the resin located in the bottom of the patterns 127 has an absorption such that its absorption is greater than the absorption of the zone 128 adjacent to the pattern 127. The exposure is therefore performed so that:

the resin located in the bottom of the units 127 absorbs a sufficient dose at its activation. It remains in place after development.

the resin located in said adjacent zone 128 in the pattern 127 absorbs a dose which is not sufficient for its activation. It is removed during development.

The invention thus makes it possible, with a negative resin, to easily reverse the patterns 127 obtained by nano-printing. It then makes it possible to obtain patterns similar to the reliefs 132 of the mold 130.

Concerning the general implementation of the invention the following remarks apply:

The optical properties of the resin, the substrate and more particularly those of the resin / substrate interface will advantageously be adapted to adjust the production method to a particular application and / or to widen its application window. The conditions of the optical insolation, especially the wavelength of the optical source but also to a lesser extent, the optical aperture, the illumination, the depth of field, the angle of incidence are to be considered.

The substrate, or the material placed under the resin, has a very strong influence on the absorption of the resin film as a function of its thickness. In view of the simulation results shown in Figure 5, we see that some materials are more favorable than others, for example: SiC, Si, Ge, Ag, AISi and W, offer the possibility to have a strong difference in absorption between two thicknesses of resin. As in the diagrams of the previous figure, it is the normalized absorption which appears in ordinate according to the thickness of resin expressed in nm.

In order to be placed under particularly advantageous conditions of application, it is preferable to adjust the thickness of the resin parts which are strongly pressed and those which are less so or not. For this purpose, as shown in Figures 1a and 1b, it is possible to play on the one hand on the thickness 121 of the resin layer initially deposited, and on the other hand on the height 131 of the protruding projections of the mold . This so that the thickness of the parts of resins pressed and those which are less or not correspond to the more exactly possible to the chosen minima and maxima of the sinusoidal absorption curve 226.

The resins used are photosensitive resins, for example resins with chemical amplification conventionally used in microelectronics, for example the resin usually referenced CAP 1 12 and marketed by the Japanese company TOK, which must also be able to preserve the mold cavity without deformation. and without the heating undergone during this operation does not alter their photosensitive properties.

- If a chemical amplification resin is used, attention must be paid to the temperatures and pressures applied during the pressing process. The pressing temperature must remain below the thermal crosslinking temperature of the resin which is dependent on the pressure applied to the resin film.

In order to generate the interference phenomena in the resin film, it is possible to use a coherent light source, that is to say having a given wavelength, such as a laser or a UV lamp provided with a suitable filter.

It is also possible to use polychromatic sources which are filtered or have a restricted spectrum width, typically less than 200 nm. It is also possible to use sources with several distinct wavelengths, or to use several light sources successively to effect the insolation of the resin if these different wavelengths make it possible to increase the absorption differential.

All sources usually used for optical lithography may be suitable. One can for example use a mercury lamp, usually referred to by its Anglo-Saxon Mercury Arc Lamp, filtered to obtain a peak intensity for a specific wavelength. Typically, a mercury lamp configured to have an intensity peak at a wavelength of 436 nm or 405 nm or 365 nm can be used. G-line lithography for 436nm wavelength, H-line lithography for 405nm wavelength and l-line lithography for 365nm wavelength will be discussed respectively. It is also possible to use an excimer or exciplex laser (KrF, ArF, F _2, etc.). The source and its wavelength must be chosen according to the sensitivity of the resin used.

- If, as we have seen, the optimal dose of insolation or "dose-to-size" which makes it possible to obtain a pattern of nominal size varies according to the thickness of the resin film, this optimal dose of The exposure must also be adapted according to the dimensions and / or the configuration of the patterns to be produced. In general, the dose Optimal increases when the dimensions of patterns, lines or spaces, decrease. Therefore, it is easier to perform a reverse lithography of that obtained by performing a standard nano-imprint lithography operation, employing a positive resin, as shown in FIG. 4b. In the same way, it is easier to remove the strongly drawn parts of the resin, the residues, using a negative resin as illustrated in Figure 4d.

Indeed the bottom of the pattern has a very small size which increases the optimal dose to bring to this pattern to be activated. The optimum dose difference between the pattern background and the area adjacent to the pattern is therefore important. This facilitates the activation of the adjacent area without activating the background of the pattern.

In the case of a positive resin, the bottom of the pattern, whose resin is not activated, then remains in place. We then obtain a pattern inversion which forms for example a line as illustrated in Figure 4b.

In the case of a negative resin, the pattern background is not activated and disappears during development. The residue is removed which forms a trench as shown in Figure 4d.

- Depending on the conditions used (resin, substrate, patterns to be made, etc.), it is possible for a resin residue to be present on the inverted image of the nanoimprint lithography (in which the thicknesses of resins are reversed) . In this case, it is sufficient to remove the residue using the techniques usually employed for a nanoprint lithography and indicated previously.

Thus, in summary, the application of the method according to the invention corresponding to the first and the fourth variant, as illustrated respectively by FIGS. 4b and 4e, makes it possible to invert the image produced by a printing mold. nanometric, that is to say, can directly transfer into the resin protruding reliefs of the mold.

Moreover, the application of the method corresponding to the second and third variants, as illustrated respectively by FIGS. 4c and 4d, makes it possible, on the contrary, to eliminate the resin parts located in the bottom of the patterns 127 obtained by nano printing. that is to say, the resin parts which have been strongly stamped by the protrusions 132 projecting from the mold. These two variants thus provide an alternative to a standard nano-print lithography operation where the stampings, most commonly referred to as residues, are removed in a subsequent etching process. The method of the invention thus offers the advantage of very well keeping the dimensions of the resin patterns. Finally, it should be noted that the process of the invention makes it possible simultaneously to produce lithographs with opposite or significantly different opening rates on the same resin layer. By opening rate of a given zone of a plate is meant the ratio between the surface of the resin left in place in this zone and respectively the resin surface in which recessed patterns are made during printing. in this same area. As shown in FIG. 6, lithographs comprising both areas where narrow trenches are to be opened in resin 610, and others where only narrow lines 620 of resin should remain. As previously indicated, obtaining narrow, projecting end patterns, such as lines, is particularly delicate with known nano printing methods. Generally, in lithography, narrow holes and trenches are made with a positive resin. This is for example the case of vias or vertical interconnections between different levels of metallization. As for the lines and networks of lines, which include, for example, the active areas and the grids of the transistors, they are made with a negative resin. This involves two different resins and therefore two successive series of steps of spreading the resin, insolation and development. In addition, this implies that different masks must then be used, considerably increasing the cost. This is not the case with the method of the invention which makes it possible to treat the two types of zones simultaneously as in the two examples of application of the invention described hereinafter.

This possibility offered by the invention of being able to obtain opposite or significantly different opening rates for two zones of the same plate is particularly advantageous in applications such as the manufacture of micro or nanoelectromechanical systems called NEMS or of devices. optics.

FIGS. 7 and 8 illustrate exemplary embodiments of the invention making it possible to obtain on the same plate opposite opening rates, that is to say final patterns projecting narrow in certain places and final patterns in narrow hollows in other places.

FIG. 7, which comprises FIGS. 7a to 7c, describes an example of application of the invention in which a mold 50 with variable topography is used, that is to say comprising reliefs 51, 52, 53, 54, 55 protruding from different heights.

The pressing of the mold 50 in the resin 120 transfers the impression of the reliefs 51, 52, 53, 54, 55 to form the patterns 61, 62, 63, 64, 65. The patterns 61, 62, 63, 64, 65 having respectively the thicknesses in _, er ₂ , er ₃ , er ₂ and e _, as illustrated in Figure 7b.

The area adjacent to these patterns, that is to say where the resin has been the least stamped or has not been stamped has a height er ₀ .

The resin 120 is then exposed. This figure illustrates that the adjacent zones delimiting a pattern receive the insolation dose.

The result after development of the resin is illustrated in FIG. 7c.

The final result reveals trenches 71, 72 at the bottom of which the resin residue has been removed during development. These trenches correspond to the reliefs 51, 55 of the mold 50.

This same final result shows final patterns 73 reversed with respect to the patterns obtained by nano printing. The pattern 73 thus forms a line conforming to the relief 53 of the mold 50.

With the same mold, one thus obtains at the same time, in some places, an inversion of patterns obtained by printing and in other places a disappearance of the residues in the pattern background. This result can be achieved with a single exposure step. The invention thus greatly simplifies the known methods of integrated circuits.

This final result can be obtained with a positive resin. In this case, the thicknesses er ₀ , en, er ₂ , er ₃ will be chosen so that er ₀ and er ₃ correspond to an absorption minimum and and er ₂ correspond to an absorption maximum.

More generally, it is necessary that the absorptions corresponding to the thicknesses er ₀ and er ₃ are significantly lower than those of the thicknesses in and er ₂ . An absorption difference of 5mJ / cm ² is sufficient. This difference indeed offers a sufficiently large process window. A larger difference, greater than 10mJ / cm ² will significantly increase this window.

This final result can be obtained with a negative resin. The thicknesses er ₀ , en, er ₂ , er ₃ will then be chosen so that en and er ₂ correspond to an absorption minimum and er ₀ and er ₃ correspond to an absorption maximum.

More generally, it is necessary that the absorptions corresponding to the thicknesses in and er ₂ are significantly lower than those of the thicknesses er ₀ and er ₃ .

FIG. 8, which comprises FIGS. 8a to 8d, describes another example of application of the invention in which the dose is varied as a function of the type of patterns of the zone to be irradiated. In this case, as illustrated in FIG. 8a, the height 131 of the reliefs 132 projecting may be identical over the entire surface of the mold 130. The impression of the resin is as described above. In a nonlimiting manner, the resin is negative in this example. The result of the printing is shown in Figure 8b. As in Figure 1, two resin layers are to be considered: the thickness e _r 124 resin zones pressed by the relief 132 projecting from the mold, and the thickness e _f 122 resin zones less or not pressed. In this example of application of the invention the thickness e _r is adjusted to have a high absorption, and for example correspond to a maximum 410 of the sinusoidal absorption curve 226. The thickness of the non-stamped parts e _f is as for it adjusted to have a weak absorption and for example to correspond to a minimum 420 of this curve.

This application of the invention is characterized in that one will proceed to two successive insolations. The first exposure 142, corresponding to a dose D 1, is limited to the zones containing relatively wide open patterns, for example 123. As seen above, the thicker zone of insolated resin corresponds to a weak zone. absorption and that of the compressed parts at high absorption. D1 dose is adjusted to allow crosslinking of the insolated areas compressed but is not sufficient to cause crosslinking of thick insolated areas where the energy absorption is lower. As will be seen in Figure 8d it is the stamped parts 126 which will remain in place after development of the resin. Negative resin in this example, which is initially soluble, and remains so where an insufficient dose is applied.

As a practical example, referring again to diagram 240 of FIG. 3b, the thickness e _f 122 can be chosen to be equal to 208 nm and correspond to the low absorption curve 242. The stamped parts are then having a thickness e equal to 124 _r 172 nm. They correspond to the strong absorption curve 244. For the parts where the patterns to be etched are wide, for example of the order or greater than 500 microns, as the pattern 123 of FIG. 8b, it is possible to see on the diagram 240 of FIG. FIG. 3b for this pattern size 248, that a dose D1 of 20 mJ / cm 2 is sufficient to activate the resin of the deep-drawn zone 123 but is not sufficient to activate the resin of the less-drawn zones. This is suitable for obtaining, in this first insolation zone, the result shown in FIG. 8d.

The invention does not make any hypothesis on the way in which the zones containing such or such type of patterns are selected nor on the means used to selectively irradiate them. For example, a mask obscuring the exposure may be used in some places. As already noted above, and as can be seen precisely in the diagram 240 of FIG. 3b, the optimum dose to be applied increases when the dimensions of the patterns to be produced, lines or spaces, decrease. The dose 144 that will be applied to the narrow pattern areas 125, that is to say D2, is higher than D1. This will allow this time the crosslinking of the thick parts 128 of resin. It will, however, be insufficient to cross-link the bottom of the narrow trenches 125 despite the fact that the thickness of the stamped parts is set for maximum absorption.

To continue the preceding practical example, while still referring to the diagram 240 of FIG. 3b, the doses that must be applied for 5 micron patterns, such as for example the pattern 125 of FIG. 8c, are significantly higher as it can be seen 246. In this example, a D2 dose of 40 mJ / cm ² is however sufficient to insolate activate the wide patterns of this second zone without, however, sufficiently insolvent the stamped areas at the bottom of the patterns such as 125. last, which will not remain after development of this negative resin since the dose applied has been insufficient. According to curve 244, for this dimension 246, it would indeed have been necessary to apply a minimum dose of approximately 70 mJ / cm ² . The dose of 40 mJ / cm ^2, however, is sufficient to activate the broad patterns, in this example the patterns having a width greater than 500 mirons, having a thickness of 208 nm which corresponds to a low absorption and the curve 242 more low absorption.

The final result after development is that shown in Figure 8d where it was possible to transfer in the resin, during the same operation, both wide patterns 126 and narrow trenches 125.

To carry out the present invention, a person skilled in the art would easily establish absorption curves of the resin used as a function of the thickness of this resin. By way of example, a method for determining the absorption curve of a resin layer as a function of the thickness of this resin layer is given below. This method can be applied to determine the curves shown in Figures 2b, 4 and 5.

The multilayer assembly or stack of layers comprising the photoresist to be printed is illustrated in FIG. 9 and is comparable to a Fabri-Pérot interferometer.

In this model we set E0 the amplitude of the electric field of the incident plane electromagnetic wave and Er the resultant amplitude of the waves reflected by the resin / substrate stack. The reflection coefficients rij and transmission coefficients corresponding to the complex wave readings (Fresnel coefficients) are:

with:

H _j : reflection coefficient at the interface between the media i and j

ti _j : transmission coefficient at the interface between media i and j

ni: complex index of the resin (n = n-ik)

ni: refractive index of the medium i. nor is the real part of the complex index n. ki: extinction coefficient of the medium i. ki is the imaginary part of the complex index n.

Let φ be the phase difference of a wave passing through the resin film: ή ₂ : complex index of the resin (n = n-ik)

d: thickness of the resin film

δ: optical path traveled by the wave in the resin

Θ: refraction angle

In our case, we are in normal incidence so: θ = 0 and φ

Referring to FIG. 9, it can be seen that the phase difference between two consecutive waves reflected or transmitted is equal to 2φ. So the resulting amplitude Er of the waves reflected by the resin film is equal to the sum:

F = 'r12 ^ F0 + ^~ t ^t ₁₂ ' r ₂ 3 t ^t ₂ l β ^{t; ~ 2ίφ} ^ F0 - t '12' r2 ² 3 'r12 t' 21 e ^{e ~ Μφ} ^ F0 + ~ t ¹ 12 'r 23' r 1 ² 2t '21 β ^{c ~ 6ίφ} ^ F0 - t' 12 'r2 ⁴ 3 'r12 t' 21 β ^{e ~ & ίφ} ^ F

The reflection amplitude "r" is then equal to:

F rtr ---- r + tn ^{- r} n ⁺

1 + rr ₂₃ -2ίφ

1-r, 12

or ^ 2 211 r = -2ί

3 e φ

l12 1 + r ₁₂ r ₂

By proceeding in the same way, the transmission amplitude "t" is obtained:

E _f - ^ 2 ^ ~ ^ 23 ^ 12 ^ 23 ^ 23 ^ 12 ^ ^ 0 ^{"^} ^ 12 ^ 23 ^ 12 ^ 23 ^ ^ 0 ^_ ^ 12 ^ 23 ^ 12 ^ 23 ^ ^ 0 t-

0 l + r _n r ₂₃ e

1 + ^ ₁₂ ^ 23 ^

The reflection and transmission coefficients corresponding to the intensities of the waves, named reflectivity R and transmission T, are equal to the squares of the respective modules of the amplitude coefficients: R = \ _r \ ² = n * and T = \ t \ ² = tt *

From the reflectivity and the transmission it is possible to determine the absorption of the resin film thanks to the following relation:

R + T + A = 1

With:

R: reflectivity

T: the transmission

A: absorption

An exemplary embodiment of a pattern inversion, non-limiting example, will now be described with reference to Figures 10, 11a to 11e.

Figure 10 illustrates the absorption of the resin used as a function of its thickness. This resin is a positive resin of CAP112 type.

In this example, the resin layer 120 initially has a thickness of 375 nm (e _f ). The resin 120 is disposed on a silicon substrate 110.

The mold 50 employed has projecting patterns of 100 nm thickness.

So, we find ourselves in the configuration where the initial resin thickness (e _f ) is close to an absorption peak 410 and where the residual resin thickness (e _r ) after nano-printing (ie about 275 nm) is close to an absorption minimum 420. The thicknesses e _f and e _r corresponding to the adjacent areas delimiting each pattern are shown in Figure 10.

The reliefs of the mold have dense lines which make it possible, by the printing step as illustrated in FIG. 11b, to print in the resin 120 patterns forming trenches of approximately 250 nm in width separated by spacers. 250nm also. Thus, parallel lines of about 250 nm in width are obtained, a pattern 127 forming a hollow line (thickness of _r ) being adjacent to two patterns each forming a projecting line of thickness e _f .

The patterns obtained are illustrated in Figure 11a with two different scales.

Following the printing step, an exposure step is performed, for example at a wavelength λ = 248 nm. During this step, only half a plate is exposed. The lower part 1 1 1, located below the dotted line in Figure 1 1 c is not exposed. The upper part 1 12, located above the dotted line in Figure 1 1 c is exposed. In this upper part 112, the zones delimiting a pattern receive the same dose of insolation. Thus, resin areas having a resin thickness e _f and resin areas having a resin thickness e _r receive the same dose of insolation in that portion 112 of the plate.

This insolation dose is chosen so as to be sufficient to activate the resin in areas of high absorption (areas having a thickness e _f in this example) and so as not to be sufficient to activate the resin in low areas. absorption (areas with a thickness e _r ). The resin being positive, areas u épaisseu re _f are activated and disappear during development. Areas with a thickness e _r are not activated and do not disappear during development. Thus, only the resin parts stamped are preserved. For the part 1 12 of the plate subjected to exposure, the patterns schematized in FIG.

FIG. 11d also illustrates the schematic patterns observed for the portion 11 1 of plate not subjected to the exposure step.

FIGS. 11c and 11d thus clearly show that, following the same step of pressing a mold 50 in the resin 120, it is possible to obtain, by means of the process according to the invention, patterns which are the reverse of those which the we obtain without the insolation step. In the case of the insolated plate, the salient patterns following the pressing step have disappeared. Following the exhibition and development stage, motifs 126 protrusions were formed from the resin residue located in the bottom of the patterns

127 recessed from the pressing step. Particularly advantageously, the salient patterns 126 obtained do not have residues and are therefore directly exploitable.

Fig. 11e is a photograph showing the difference in patterns at the junction between exposed and non-exposed portions 11-11 of the resin. This figure clearly shows that instead of the patterns 127 formed by hollow pressing of the mold 50, the resin 120 which has been exposed has patterns 126 projecting. In the context of the present invention, it is particularly advantageous to use so-called "threshold" resins. We speak of threshold resin, when the chemical structure of the resin is modified from a relatively precise dose of insolation. In the case of a negative resin, this modification of the chemical structure of the resin is comparable to a crosslinking. In the case of a positive resin, this modification of the chemical structure of the resin is comparable to a deprotection. Threshold resins are often characterized by high contrast. This contrast is preferably greater than 1.

It should be noted that a high contrast of the resin makes it easier to practice the present invention. The present invention can nevertheless be performed with resins having a low contrast.

It should also be noted that the contrast of a resin is dependent on many parameters. Among the most important we find: the type of substrate, the process used, and in particular the conditions of development of the resin. Among these resin development conditions are the following parameters: temperature and annealing time after insolation; nature, developer concentration and temperature; method and time of development.

The resin thickness after insolation and development varies depending on the patterns and the insolation dose. In order to approximate the value of the contrast, a curve representing the residual resin thickness can be plotted as a function of the insolation dose. Figures 12a and 12b illustrate such curves for negative and positive resins respectively.

We can then determine the contrast γ by the following equation:

For example, these curves can be obtained by irradiating identical patterns, on the same plate, with a dose of increasing insolation. It is then necessary to measure the residual resin thickness after development for each dose of insolation.

In the illustrated example, squares of 9mm side have been insolated to neglect the lateral diffusion phenomena photogenerated acid, because in these examples resins chemical amplification type NEB22 and CAP112 were used.

In conclusion, it will be noted that the method of the invention takes advantage of two phenomena: one is the absorption differential of the resin as a function of its thickness and the other is related to the size of the patterns and to the higher doses. high that must be applied to insolate smaller patterns. Depending on whether a positive resin or a negative resin is used, it is possible to take advantage of the two phenomena or the only phenomenon related to absorption according to the table below:

The embodiments of Figures 7 and 8 may be combined. In particular, for the same plate it is possible to use a mold with variable topography and different exposure doses.

The invention is not limited to the previously described embodiments but extends to any embodiment within its spirit.

Claims

1. A method of nanoscale lithography comprising a preparation step during which a photosensitive resin (120) is placed on a substrate (110), at least one step of pressing a mold (130, 50) in the resin (120) to form in the resin at least one printing pattern (127) delimited at least in part by two adjacent areas (128, 129), one (129) of said two areas (128, 129) having a thickness (e) less than the thickness (ef) of the other (128) of said two zones (128, 129), characterized in that it comprises a step of exposing at least said two zones (128, 129) to during which said two zones (128, 129) receive the insolation dose, and characterized in that the thicknesses (er, ef) of said two zones (128, 129) are defined so that, to be activated, the resin at one of said two zones (128, 129) requires an insolation dose different from the exposure dose nil to activate the resin at the other of said two zones (128, 129) and in that the exposure dose provided by the exposure step is determined so as to be large enough to activate the resin at the of only one of said two areas (128, 129) and so as not to be large enough to activate the other of said two adjacent areas (128, 129).

2. Method according to the preceding claim wherein the absorption of the exposure dose by the resin (120) according to its thickness defines a substantially sinusoidal curve (218, 226) and wherein the thickness of the resin (120 ) at one of said two areas (128, 129) substantially corresponds to a maximum (410) of said sinusoidal curve and the thickness of the resin at the other of said two areas (128, 129) corresponds substantially to a minimum (420) of said sinusoidal curve (218, 226).

3. A method according to any one of the preceding claims wherein the resin thicknesses (er, ef) of said two zones (128, 129) are determined so that the difference between the dose required to activate one of said two zones (128, 129) and the dose required to activate the other of said two zones (128, 129) is at least 5mJ / cm ² .

A method according to any one of the preceding claims wherein the resin is a positive photoresist, said thicknesses (er, ef) are determined so that the resin at the region (129) having the smallest thickness (e ) has a lower absorption than that of the resin at the zone (128) having the greatest thickness (ef) and wherein the exposure dose provided by the exposure step is set so as to activate the resin of the area

(128) having the greatest thickness (ef) and not activating the resin of the zone

(129) having the smallest thickness (er), so as to obtain a final pattern (126) inverse of the printing pattern (127).

A method according to any one of claims 1 to 3 wherein the resin is a negative photoresist, said thicknesses (er, ef) are determined so that the resin at the region (129) having the smallest thickness has a higher absorption than that of the resin at the zone (128) having the greatest thickness (ef) and wherein the exposure dose provided by the exposure step is set so as to activate the resin of the zone (129) having the smallest thickness (er) and not activating the resin of the zone (128) having the greatest thickness (ef), so as to obtain a final pattern (126) inverse to the pattern of printing (127).

6. Process according to any one of the two preceding claims, in which, after development of the resin, an additional etching step is carried out to remove a residue of resin remaining on the substrate (110) at the zone (128) exhibiting the highest thickness (ef) after the development stage.

A method according to any one of claims 1 to 3 wherein the resin is a positive photoresist, said thicknesses (er, ef) are determined so that the resin at the region (129) having the smallest thickness (e) has an absorption greater than that of the resin at the zone (128) having the greatest thickness (ef) and wherein the exposure dose provided by the exposure step is set so as to activate the resin at the zone (129) having the smallest thickness (er) and not activating the resin at the zone (128) having the greatest thickness (ef), so as to remove a resin residue at the area (129) having the smallest thickness (er).

A method according to any one of claims 1 to 3 wherein the resin is a negative photoresist, said thicknesses (er, ef) are determined so that the resin at the region (129) having the smallest thickness (e) has a lower absorption than that of the resin at the zone (128) having the greatest thickness (ef) and wherein the exposure dose provided by the exposure step is set so as to activate the resin of the area

(129) having the smallest thickness (er), so as to remove a resin residue at the zone (129) having the smallest thickness (er).

9. A method according to any preceding claim wherein after the pressing step a plurality of printing patterns (61, 62, 63, 64, 65) having different thicknesses (erO, er1, er2, er3), at least one of these thicknesses corresponds to an absorption maximum, and at least one of these thicknesses corresponds to a minimum of absorption.

10. Method according to the preceding claim wherein the mold (50) has protruding reliefs (51, 52, 53, 54, 55) of different heights.

A process according to any one of the preceding claims wherein during the exposure step all the resin (120) is exposed to the insolation dose.

12. A process according to any one of the preceding claims in which resin portions (120) are insulated with different insolation doses (142, 144).

13. Method according to the preceding claim wherein during the exposure step is insole at least a first pattern (123) having a first dimension, said dimension being taken in a direction substantially normal to the thickness of the resin, with a first dose of insolation (142) and at least one second pattern (125) having a second dimension smaller than said first dimension is insulted with a second dose (144) of insolation greater than said first dose of insolation (142). ).

14. Method according to the preceding claim wherein the first dose of insolation (142) is sufficient to activate only one of said two zones of the first pattern (123) and wherein the second dose of insolation (144) is insufficient to activate the second pattern (125) but is sufficient to activate one or more areas (128, 128) bordering the second pattern (125).

15. A method according to any one of the preceding claims wherein the insolation dose is provided by a coherent light source.

16. A method according to any one of claims 1 to 14 wherein the exposure step successively involves several light sources having different wavelengths.

17. A method according to any one of the preceding claims comprising, during the preparation step, a step in which the photoresist (120) is deposited on a layer or a substrate (1 10) taken from the following materials SiC, Ge, Ag, W, AlSi.

18. A method according to any one of claims 1 to 16 comprising, during the preparation step, a step during which the photoresist

(120) is deposited on a silicon layer or substrate (1 10).

19. Multilayer assembly comprising a substrate (1 10) coated with a layer of photoresist (120), the resin having at least one pattern (127) for printing, delimited at least in part by two zones (128, 129) having different thicknesses (er, ef), characterized in that the thicknesses of said two zones (128, 129) respectively correspond to a maximum and a minimum of an absorption curve (218, 226) of said resin (120 ) according to its thickness.