SUBSTRATES

FIELD OF THE INVENTION

The invention relates to the implantation of semiconductor structures by means of one or more ionic or atomic species, in order to form a region of fragilization within such structures, with a view to the detachment of a thin layer bounded by the said fragilization region.

The invention may notably be applied to the fabrication of structures of the semiconductor-on-insulator type, also referred to by the acronym SOI, in which a layer of insulator is sandwiched between a thin layer of semiconductor material and a base substrate.

The term "insulator" is understood to refer to an electrical insulator. BACKGROUND TO THE INVENTION

A structure of the SOI type generally comprises at least one layer of insulator sandwiched between a thin upper layer and a base substrate, typically composed of silicon. "Thin layer" is taken to be a layer with a thickness typically in the range between 50 A and a few micrometres, for example around a hundred Angstroms, for example 120 A.

The layer of insulator can be a layer of oxide, such as Si0 ₂ , which is then referred to by the acronym "BOX", for "buried oxide", in other words oxide buried under the thin layer of silicon.

The fabrication of structures of the SOI type generally comprises the following steps:

- formation of a layer of oxide on the surface of a substrate made of a semiconductor material such as silicon,

implantation of atomic or ionic species into the substrate, through the layer of oxide, so as to form a fragilization region within the substrate, and to bound a thin layer of semiconductor material between the layer of oxide and the fragilization region,

- bonding of the substrate onto a second substrate, and

removal of the back part of the first substrate, by detachment along the fragilization region, so as to conserve only the structure of the SOI type. It is also possible to carry out the implantation of the atomic or ionic species into the substrate and to form the layer of insulator on the second substrate.

Nevertheless, implanting the species through the insulator offers certain advantages. This allows the implantation depth of the species to be made uniform, by notably reducing a phenomenon known as "channelling", during which the implanted species along a main direction of the crystal lattice formed by the thin layer propagate to a greater depth than the mean depth reached by the implanted species.

Furthermore, as indicated hereinabove, the fragilization region defined by the implanted species bounds the thin layer to be detached from the substrate. Making the implantation depth of the species uniform therefore allows the thickness of the thin layer to be made more uniform, this thickness being subject to very tight controls because it strongly influences the quality of the SOI structure obtained.

However, in the latest SOI structures developed, called UTBOX for Ultra- Thin Buried Oxide, in which the thickness of the layer of oxide is less than around 50 nm (such a layer will be referred to as "thin oxide film"), the inventors have observed that the thickness uniformity of implantation of the species is degraded when the implanter used is of the batch implanter type.

In this type of implanter, a plurality of substrates to be implanted are positioned on as many supports disposed around a perimeter of a wheel rotating about its axis of symmetry, and exposed in turn to a flux of species maintained in a constant direction.

Surprisingly, the use of a batch implanter for implanting a UTBOX structure does not allow satisfactory results in terms of implantation depth to be obtained. Furthermore, in the case of an SOI structure of the UTBOX type, the specifications on the thickness uniformity of the thin layer of semiconductor material to be detached from the substrate are even more severe than for conventional SOI structures.

It is therefore necessary to develop, notably for SOI structures of the UTBOX type, novel solutions to allow an improved uniformity of the implantation depth of the ionic or atomic species in the substrate. DESCRIPTION OF THE INVENTION

The aim of the invention is to overcome the aforementioned problem. In particular, one of the aims of the invention is to provide a novel method for implantation of atomic or ionic species into a substrate through a thin layer of insulator, and allowing a uniform implantation depth of the species to be obtained.

For this purpose, the invention provides a method for implantation of atomic or ionic species into a batch of substrates made of semiconductor material, in which:

- each substrate made of semiconductor material is positioned on a respective support of a batch implanter, each substrate comprising a thin layer of electrical insulator on its surface, and

- a dose of at least one ionic or atomic species is implanted over the whole surface of the said substrates, through their layer of insulator, so as to form a fragilization region within each substrate and to bound there a thin layer of semiconductor material between the thin layer of insulator and the fragilization region of the substrate,

the implantation method being characterized in that, during the said method, each support on which a substrate is positioned has at least two separate inclinations with respect to the plane orthogonal to the direction of implantation of the species, in order to make the implantation depth of the species in the substrate uniform.

Advantageously, but optionally, the invention furthermore comprises at least one of the following features:

- the inclinations of the support have an angle in the range between 2° and 15°, preferably between 4° and 10° with respect to the plane orthogonal to the direction of implantation.

- during the implantation step, a predetermined dose of atomic species is implanted, the said implantation comprising the implantation of a first part of the dose while the support presents a first inclination, and of a second part of the dose while the support presents a second inclination.

- the first inclination and the second inclination are opposing with respect to a plane orthogonal to the direction of implantation. - the first part of the dose is in the range between 25% and 75% of the total dose, preferably equal to 50%, and the second part of the dose is its complement with respect to 100%.

- the thin layer of insulator has a thickness in the range between 1 nm and 50 nm, preferably between 15 nm and 25 nm.

- the implanted species is hydrogen.

- the implantation step comprises a co-implantation of hydrogen and helium. The invention furthermore provides a method for fabrication of a heterostructure of the SOI type comprising a thin layer of insulator sandwiched between a substrate referred to as 'receiver' and a thin layer coming from a substrate made of semiconductor material referred to as 'donor', the method comprising the steps consisting in:

- forming at least one thin layer of insulator on a donor substrate,

- forming a fragilization region in the donor substrate by implantation of the donor substrate by applying the implantation method according to the invention,

- bonding the donor substrate and the receiver substrate so that the layer of insulator is sandwiched between them, and

- separating the thin layer from the rest of the donor substrate on the fragilization region.

According to one embodiment, the thin layer is separated from the rest of the donor substrate by heat treatment.

According to one embodiment, the heat treatment comprises an increase in temperature to a temperature less than 500°C, and according to a ramp of less than 7°C per minute, preferably less than 5°C per minute.

The invention also relates to a heterostructure of the SOI type, comprising a thin layer of insulator sandwiched between a thin layer of semiconductor material and a base substrate, the structure being directly obtained by the implementation of the method of fabrication according to the invention, and the thin layer of semiconductor material having a variability in thickness of less than 5 A. The implementation of the method according to the invention allows the uniformity of the depth of implantation of the species into a substrate to be improved, and allows this even when the substrate is only covered by a thin layer of insulator.

DESCRIPTION OF THE FIGURES

Other features, aims and advantages of the present invention will become apparent upon reading the detailed description that follows, with reference to the appended figures presented by Way of non-limiting examples and in which:

- Figure 1 shows a fabrication process for an SOI structure during which the method according to the invention is implemented,

Figure 2 shows schematically the installation for the implementation of the method according to the invention,

- Figure 3a is a top view of a substrate during implantation, for illustrating notation conventions adopted for the description of the method,

Figure 3b is a graph presenting results of the application of the method according to the invention.

Figure 3c shows topographies of the thin layer surfaces obtained following the fracturing of the substrate for various implantation angles. - Figure 4 is a cross-sectional diagram of a substrate implanted by the method according to the invention,

DETAILED DESCRIPTION OF THE INVENTION

The invention is implemented during the fabrication of structures of the semiconductor-on-insulator or SOI type, in which a layer of insulator is sandwiched between a thin layer of semiconductor material and a substrate.

The thin layer of semiconductor material has a thickness typically in the range between 50 A and a few micrometres, for example of the order of a hundred Angstroms. This is notably the case when the thin layer is used in a product of FDSOI (Fully-depleted Semiconductor-on-lnsulator) type. A possible value of this thickness is 120 A.

The layer of insulator, also called BOX, may preferably, but in a non-limiting manner, be a thin layer, then called UTBOX, with a thickness in the range between 1 nm and 50 nm, preferably between 15 nm and 25nm. The invention may nevertheless be applied to thicker BOX layers, for example of the order of a few hundred nanometres.

The process of fabrication of such a structure is shown in Figure 1. With reference to Figure 1 a, a substrate 10 made of semiconductor material, also called donor substrate, is provided from which a thin layer, or useful layer 12, of semiconductor material, typically of silicon, will be taken.

The substrate 10 is covered by a thin layer of insulator 1 1 , for example of silicon oxide Si0 ₂ .

The thin layer 12 taken from the donor substrate 10 will be transferred onto another substrate 20, also called receiver substrate.

Furthermore, the substrate 20 may also be covered with a layer of insulator 21. This layer 21 is nevertheless optional because the thin layer of insulator covering the substrate 10 can be sufficient for forming the thin layer of buried oxide.

With reference to Figure 1 b, a fragilization region 13 is created within the donor substrate 10, in such a manner as to bound, between the fragilization region 3 and the layer of insulator 11 , the thin layer of semiconductor material 12 to be transferred.

' This fragilization is carried out by implantation of ionic or atomic species, during a step which is described in more detail hereinafter.

With reference to Figure 1 c, once the fragilization region 13 has been created, the donor substrate 10 is bonded to the receiver substrate 20, so that the layer of insulator 1 1 on the donor substrate 10 and, where it exists, the layer of insulator 21 on the receiver substrate 20, are sandwiched between the thin layer 12 and the receiver substrate 20.

Finally, in Figure 1d, the back part of the donor substrate 10 is removed by fracturing, in such a manner as to leave only the thin layer of semiconductor material 12 on the layer of insulator 1 1 , 21. At this step, a structure 1 of the SOI type is obtained. The control of the implantation depth of the species during the implantation step for creating a fragilization region 13 of uniform depth will now be described.

During the implantation step, the species to be implanted are projected towards the substrate 10 with an energy that is predefined depending on the species, so that the latter penetrate inside the substrate to the desired depth, through the thin layer of insulator 11.

The invention is preferably implemented by implantation of hydrogen.

However, implanting other species may also be envisaged. For example, a co-implantation of hydrogen and helium, which allows the quality of the final thin layer 12 to be improved, notably by reducing its roughness, may also be used for this step.

In order to carry out the implantation with a controlled implantation depth of the species, a device known as an 'implanter' is used, which comprises the system for projection of the species to be implanted, and a support for fixing the substrate or substrates to be implanted, and holding them in the desired position.

The implantation is carried out in 'full sheet' mode, in other words, for each substrate, over the whole surface of the substrate.

The species are generally emitted by a source 32 then injected into a particle accelerator, selected by a mass analyzer, before being projected onto the wheel 30.

The implanter used in the method according to the invention is a batch implanter 3. As illustrated in Figure 2, the implanter 3 comprises a plurality of supports 31 disposed along an outside perimeter of a wheel 30.

Each support 31 can have a surface area and a shape substantially equivalent to the surface area and to the shape of the substrate that it is designed to carry.

Each substrate 10 is positioned on a support 31 , and the contact between a substrate 10 and a support 31 is established over essentially the entirety of the surface area of the said substrate 0. "Essentially the entirety of the surface area of the said substrate 10" is understood to mean a surface area greater than 90%, preferably greater than 95%, even more preferably the totality of the surface area of the substrate. The contact between a support 31 and a substrate 10 is designed to stabilize the temperature of the substrate uniformly. Stabilizing the temperature of the substrate is taken to mean cooling or heating the substrate.

The wheel 30 is driven in rotation about its axis of symmetry during the implantation step. When the wheel 30 is rotating, the substrates 10 are only held around their outside edge on the supports 31 by fixing means in the shape of a gutter. On this subject, those skilled in the art will find a technical description on the mechanism allowing the substrates 10 to be held on the supports 31 in the document EP 2320454 A1 (T.BESNARD), 1 1 May 201 1 (1 1.05.201 1 ), column 1 , paragraphs [0003] and [0004].

During an implantation, each substrate 10 is exposed to a flux of atomic or ionic species 40 in a constant direction. The flux of implanted species tends to heat up the wafer. In a known manner, for implanting into a batch of substrates 10, the temperature of each substrate 10 is controlled by the support on which it is sitting in order to reduce its temperature. However, the applicant has observed that the temperature is not uniform over the whole surface of the substrates 10. This non- uniformity of the temperature of the substrates 10 is probably due to the at least partial detachment of the substrates 10 from their support 31 during the rotation of the wheel 30. The temperature stabilization of the substrates 10 is then no longer uniform. Consequently, the non-uniformity of the implantation depth of the implanted species is probably due to the non-uniform control of the temperature of the substrates 10.

This effect is not observed in the case of single-wafer implantors. The reason for this is that, in the case of a single-wafer implanter, only one substrate is positioned in the implanter on a support. An electrostatic contact is established between the wafer and the support. The contact between the substrate 10 and its support is therefore relatively unaffected. In other words, as there is no detachment between the support and the substrate, the control of the temperature of the substrate is uniform. During the implantation of the species in a batch implanter 3, each substrate

10 is inclined by tilting each support 31 for a substrate 10 with respect to the direction of implantation of the species.

The inclination is illustrated in Figure 3a, which shows schematically a support 31 exposed to the flux of species, the support being seen from above. An axis x is defined, parallel to the direction of the flux 40 of implantation of the species, and an axis y, defining with x a direct orthogonal reference frame. Thus the axis y is included in a plane orthogonal to the direction of implantation.

The support, and hence the substrate which is positioned on it, is tilted with respect to the direction of the flux 40 by an angle a with respect to the axis y, in other words to the plane orthogonal to the direction of implantation, such that the flux 40 of implantation of the species is tilted by the angle a with respect to the normal N to the wafer 10.

Advantageously, the tilting of the support 31 may be modified during the implantation step, in such a manner that the substrate presents at least two separate inclinations with respect to the direction of implantation of the species during the implantation step.

For example, the implantation step may be implemented in several sub- steps, for example two sub-steps, between which the tilt, in other words the value of the angle a, is modified.

For example, the tilt can be modified between the sub-steps such that a part of the total dose of species is implanted with a first inclination a, and a complementary part of the total dose of species is implanted with an inclination a' different from a.

For example, the first part of the dose can be in the range between 25% and 75% of the total dose, preferably 50%, and the second part of the dose is its complement with respect to 100%.

The implantation step may also be divided into three or four sub-steps during which the inclinations are respectively different. The wheel is stopped when the inclination of the substrate supports is changed.

In the case of two sub-steps for which two separate inclinations are applied, the inventors have observed that, surprisingly, the results are better when the inclinations are respectively opposing (α' =- a) with respect to the axis y, in other words with respect to the plane orthogonal to the direction of implantation of the species. However, a decrease in the thickness variability is also observed even if the inclination a' of the second part of the dose to be implanted is not the opposite of the inclination a of the first part. The absolute values of the inclinations can be different. The results of one preferred embodiment of the invention are illustrated in

Figure 3b, where the thickness variability of the thin layer 12 is shown as a function of the value of the tilt a applied during the implantation. This thickness variability is represented as a percentage of the variability obtained when no tilt is applied to the substrate. Thus, for positive percentages, the thickness variability is increased, and the result is therefore degraded. On the other hand, the negative percentages indicate a decrease in the thickness variability and hence an improvement in the quality of the surface of the thin layer.

The thickness variability is defined with reference to Figure 4 by the difference in thickness between the point on the layer 12 where the thickness is the greatest, and that where the layer is thinnest.

Again with reference to Figure 3b, the experimental measurements are indicated by the dots. The mean values M of these measurements are illustrated by the horizontal segments, and the extrema E by the apices of the triangles whose bases are formed by the horizontal segments.

Four series of measurements are illustrated in this figure, respectively from left to right: . implantation without inclination, implantation with constant inclination equal to 7°, implantation with constant inclination equal to -7°, and implantation whose first half is carried out with an inclination equal to 7°, and the second half with inclination equal to -7°. The other implantation parameters, including in particular the energy of the implanted species, are the same for these series of measurements.

In Figure 3c, the same experimental results are shown in the same order, this time in the form of topographies of the surface of the thin layer obtained after fracturing.

Conventionally in topography, each contour line shows a particular elevation; the wider a given line, the less the surface exhibits variations in elevations.

Surprisingly, it is observed that the inclination at an angle a of 7° (second illustration starting from the left in Figures 3b and 3c) does not allow any beneficial effect to be obtained. This thickness variability is degraded when the substrate is tilted at an angle of 7° by 20% with respect to an inclination of 0°.

In Figure 3c, an increase in the elevation contour lines is noted, which indicates an increase in the thickness variability of the thin layer 12.

On the other hand, it is observed that the inclination by an angle of -7° allows the thickness variability of the thin layer to be reduced by 15% with respect to the inclination of 0°, as indicated in the third illustration of Figure 3b. A decrease in the variability in elevation can be seen in Figure 3c by wider and less numerous elevation contour lines.

Even better results can be seen in Figures 3b and 3c, with the successive use of opposing inclinations of absolute value equal to 7°, which allows a decrease in the thickness variability of the thin layer 12 of 40% with respect to the inclination of 0°. The thickness variability obtained is then less than 10 A and, in this implementation, even less than 5 A.

The values of the inclination successively at + and - 7° are preferred but non- limiting. Indeed, a decrease in the thickness variability has also been observed for absolute values of angles in the range between 2 and 15°, and preferably between 4 and 10°.

Nor is the invention limited to the successive use of a positive then negative inclination; the first part of the dose can be implanted with a negative inclination (for example a=-7°), and the second part of the dose with a positive inclination (for example a'=+7°). The implantation step such as implemented in the present invention, by directly applying the method for fabrication of an SOI structure such as previously described, allows a heterostructure to be obtained having a thin layer 12 over a buried oxide layer 1 1 , in which the thin layer exhibits a thickness variability that is reduced with respect to the conventional methods of fabrication.

Indeed, the depth of implantation of the species determines the fragilization region of the substrate 10, and hence the free surface of the thin layer after its detachment. The improvement in the thickness variability of the thin layer therefore results directly from the improvement in the variability of the depth of implantation of the species. The control of the fracturing step will now be described.

The fracturing step consisting in removing the back part of the donor substrate 10, in such a manner as to leave only the thin layer of semiconductor material 12 on the layer of insulator 11 , 21 , is executed by a heat treatment step.

The heat treatment step comprises an increase in temperature to a temperature less than 500°C.

In a particularly advantageous manner, the increase in temperature is applied according to a ramp of less than 7°C per minute, preferably less than 5°C per minute.

The term 'ramp' here means a rate of rise of temperature.

The applicant has observed the following facts:

- The variation in thickness of the thin layer after fracturing is substantially equivalent irrespective of the ramp used during the fracturing step, and according to a method involving an implantation according to only one inclination. For example, under conditions of implantation according to only one inclination, the variation in thickness of the thin layer after fracturing is substantially equivalent in the case of a ramp at 5°C/min or 16°C/min.

- In contrast, for a method involving an implantation according to two separate inclinations, the variation in thickness of the thin layer after fracturing is 4.5 A for a ramp of 5°C/min, whereas it is 8 A for a ramp of 16°.C/min.

The effect of the ramp on the variations in thickness of the thin layer after fracturing is very negligible when an implantation step involves only one inclination. On the other hand, the association of an implantation step according to two separate inclinations, and of a ramp of less than 7 °C, allows the variations in thickness of the thin layer to be significantly reduced, which is not the case when the ramp is greater than 7°C/min.

There is evidently a synergy between the mode of implantation and the ramp of the fracturing heat treatment.

The consideration of an implantation according to two separate inclinations combined with a heat treatment comprising an increase in temperature according to a ramp of less than 7°C per minute results in a greater effect than the sum of the effects taken separately.

In particular, the application of the method according to the invention allows a thickness variability of the thin layer of less than 10 A, or even less than 5 A, to be obtained.