METHOD OF MAKING A TRANSISTOR HAVING A SOURCE AND A DRAIN OBTAINED BY RECRYSTALLIZATION OF SEMICONDUCTOR

Abstract

Method of making a transistor, comprising the following steps: make a gate and a first spacer on a first channel region of a first crystalline semiconducting layer; make first crystalline semiconductor portions on the second source and drain regions; make the second regions amorphous and dope them; recrystallise the second regions and activate the dopants present in the second regions; remove the first portions; make a second spacer thicker than the first spacer; make second doped crystalline semiconductor portions on the second regions, said second portions and the second regions of the first layer together form the source and drain of the transistor.

Claims

1. Method of making a transistor, comprising at least the following steps: make a gate and a first dielectric spacer in contact with the side walls of the gate, on a first region of a first crystalline semiconducting layer that will form the transistor channel; make first portions of crystalline semiconductor on second regions of the first layer that will form part of the transistor source and drain; make at least the second regions of the first layer amorphous and dope them; recrystallise at least the semiconductor in the second regions of the first layer and activate the dopants present at least in the semiconductor of the second regions of the first layer; remove the first portions; make a second dielectric spacer in contact with the lateral walls of the gate such that the thickness of the second dielectric spacer is more than the thickness of the first dielectric spacer; make second portions of doped crystalline semiconductor on the second regions of the first layer such that at least said second portions and the second regions of the first layer together form the source and drain of the transistor.

2. Method according to claim 1, in which the first layer is a surface layer of a semiconductor on insulator type substrate.

3. Method according to claim 2, in which the transistor is of the FDSOI type and the thickness of the first layer is between about 4 nm and 8 nm.

4. Method according to claim 1, in which the thickness of the first dielectric spacer is between about 1 nm and 3 nm.

5. Method according to claim 1, in which manufacturing of the second dielectric spacer includes a step to deposit dielectric material around the first dielectric spacer such that said dielectric material and the first dielectric spacer together form the second dielectric spacer, or in which manufacturing of the second dielectric includes removal of the first dielectric spacer followed by deposition of dielectric material forming the second dielectric spacer.

6. Method according to claim 1, also including a step between making the second dielectric spacer and making the second portions, to etch an upper part of the second regions of the first layer such that the second portions are made on remaining parts of the second regions of the first layer and partially facing the first region of the first layer.

7. Method according to claim 1, in which the first and second portions are made by epitaxy.

8. Method according to claim 1, in which the amorphisation and doping of the second regions of the first layer include implementation of a single ion implantation of doping species in the second regions of the first layer.

9. Method according to claim 1, in which the amorphisation and doping of the second regions of the first layer include a first ion implantation of non-doping species in the second regions of the first layer and a second ion implantation of doping species in the second regions of the first layer.

10. Method according to claim 1, in which the recrystallisation and activation of dopants include the use of at least one of the following steps: annealing under an inert atmosphere at a temperature between about 500° C. and 800° C., local heating of the semiconductor containing the dopants by laser.

11. Method according to claim 1, in which: amorphisation and doping are implemented such that they also result in amorphisation and doping of a lower part of the first portions and such that the semiconductor of an upper part of the first portions is kept in a crystalline state, and recrystallisation and activation of the dopants are done such that the semiconductor of the lower part of the first portions is also recrystallised and that dopants present in the semiconductor of the lower part of the first portions are also activated.

12. Method according to claim 1, also including a step between making the gate and the first dielectric spacer and making the first portions, to make an etching stop layer on the second regions of the first layer, in which the step to remove the first portions includes etching of the first portions until reaching the etching stop layer.

13. Method according to claim 12, also including a step between removal of the first portions and producing the second portions, to remove the etching stop layer.

14. Method according to claim 1, in which the first layer comprises silicon or SiGe comprising a proportion of germanium between about 20% and 60%.

15. Method according to claim 1, in which: when the transistor is of the NMOS type, the first and/or second portions comprise N type doped silicon and/or or silicon strained in tension, or when the transistor is of the PMOS type, the first portions comprise P type doped silicon and/or silicon strained in compression, and/or the second portions comprise P type doped SiGe for which the proportion of germanium is between 20% and 60%.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0048] This invention will be better understood after reading the description of example embodiments given purely for information and in no way limitative with reference to the appended drawings on which:

[0049] FIG. 1 diagrammatically shows an FDSOI transistor obtained when dopants are activated by classical SPER;

[0050] FIGS. 2 to 7 show steps in a first embodiment of a method of making a transistor;

[0051] FIGS. 8 to 12 show some of the steps in a second embodiment of a method of making a transistor;

[0052] FIGS. 13 to 18 show some of the steps in a third embodiment of a method of making a transistor;

[0053] Identical, similar or equivalent parts of the different figures described below have the same numeric references to facilitate the comparison between different figures.

[0054] The different parts shown on the figures are not necessarily all at the same scale, to make the figures more easily understandable.

[0055] The different possibilities (variants and embodiments) must be understood as not being mutually exclusive and can be combined with each other.

DETAILED PRESENTATION OF PARTICULAR EMBODIMENTS

[0056] Refer firstly to FIGS. 2 to 7 that show steps in a first embodiment of a method of making a transistor 100.

[0057] The transistor 100 in this case is of the NMOS type and is made from a semiconductor on insulator type substrate, for example of the SOI (silicon on insulator) type and comprising a thick semiconducting layer (not shown on FIGS. 2 to 7) on which there is a buried oxide (BOX) dielectric layer 102, for example comprising SiO.sub.2 and the thickness of which may for example be of the order of 20 nm, itself covered with a first layer 104 of crystalline semiconductor, in this case comprising silicon for example with a thickness of between about 4 nm and 8 nm. Silicon in the first layer 104 may or may not be strained in tension.

[0058] Deposition, photolithography and etching steps are then implemented to make a gate dielectric and one or more gate conductors that together form a gate 112 of the transistor 100, on a first region 106 of the first layer 104 that will form the channel of the transistor 100.

[0059] A first dielectric spacer 114 is then made on the first region 106 and around the gate 112, in other words adjacent to the lateral walls or flanks of the gate 112 (see FIG. 2). The first spacer 114 for example comprises SiN and is thin, in other words its thickness (direction parallel to the X axis shown on FIG. 2) is for example between about 1 nm and 3 nm.

[0060] As shown on FIG. 3, a first epitaxy is then implemented on access regions of the transistor 100, in other words on the second regions 116, 118 of the first layer 104 that will form part of the source and drain regions of the transistor 100, forming the first portions 120, 122 of the crystalline semiconductor located on the second regions 116, 118 of the first layer 104. On FIG. 3, the first region 106 and the second regions 116, 118 of the first layer 104 are symbolically delimited from each other by dashed lines. The first portions 120, 122 made are higher than the level of the first region 106 and the second regions 116, 118. Therefore these first portions 120, 122 are placed in contact with parts of external lateral walls of the first dielectric spacer 114 and are not arranged facing the first region 106. Moreover, each of these first portions 120, 122 has a thickness (dimension parallel to the Z axis shown on FIG. 3) that is for example between about 5 nm and 15 nm. This first epitaxy can be carried out such that the first portions 120, 122 comprise doped or undoped semiconductor, for example N type Si doped with phosphorus.

[0061] The semi-conductor of the first portions 120, 122 can also be elastically strained in tension. According to one variant embodiment, the semiconductor obtained by this first epitaxy can contain a low concentration of carbon, for example between about 0.5% and 2%, which confers a smaller mesh parameter on the crystalline semiconductor of the first portions 120, 122 than the mesh parameter of the semiconductor in the second regions 116, 118 that does not contain any carbon, thus conferring a tensile strain on the semiconductor of the first regions 120, 122.

[0062] As shown on FIG. 4, one or several ion implantation steps are implemented such that dopants are implanted throughout the thickness of the second regions 116, 118 of the first layer 104. In the first embodiment described herein, dopants are also implanted in lower parts 124, 126 of the first portions 120, 122. This or these ion implantations destroy the crystalline structure of the lower parts 124, 126 of the first portions 120, 122 and the second regions 116, 118 and make this semiconductor amorphous. Therefore the lower parts 124, 126 of the first portions 120, 122 and the second regions 116, 118 together form amorphous semiconducting regions that are thicker than the first layer 104. Upper parts 128, 130 of the first portions 120, 122 are passed through by ion beams 132 used during this or these implantations without these ions destroying the crystalline structure of the semiconductor of these upper parts 128, 130. Therefore the upper parts 128, 130 of the first portions 120, 122 keep the crystalline semiconductor that will subsequently be used for recrystallisation of the amorphous semiconductor in the lower parts 124, 126 and the second regions 116, 118.

[0063] The amorphous semiconductor can be obtained by means of a first implantation of Si.sup.+ ions and then a second implantation of dopant ions, for example phosphorus or arsenic, in the second regions 116, 118 of the first layer 104 and in the lower parts 124, 126 of the first portions 120, 122. As a variant, the implantation of dopant species can be applied before the implantation of non-dopant species. It would also be possible for a single implantation of ions of dopant species, for example phosphorus ions, to be used. Ions may be implanted for example with an energy of between about 15 keV and 40 keV and with a dose par example equal to between about 1×10.sup.14 at/cm.sup.2 and 5×10.sup.14 at/cm.sup.2. Implanted ion species and the energy and doses with which these ions are implanted are chosen particularly as a function of the thickness and the nature of semiconductors of the second regions 116, 118 of the first layer 104 and the first portions 120, 122.

[0064] A recrystallisation to transform the amorphous semiconductor of the second regions 116, 118 of the first layer 104 and the lower parts 124, 126 of the first portions 120, 122 into crystalline semiconductor in which dopants are activated by SPER is then implemented. To achieve this, the entire structure formed above can be annealed, for example under an inert atmosphere, the temperature of this annealing for example being between about 500° C. and 800° C. and its duration may be between a few seconds, for example less than 10 seconds when the temperature is high (for example equal or approximately equal to 800° C.), and a few minutes (for example less than 30 minutes) when the temperature is lower (equal or approximately equal to 500° C.).

[0065] Recrystallisation of the amorphous semiconductor takes place due to the crystalline semiconductor of the upper parts 128, 130 of the first portions 120, 122 that forms a recrystallisation front for the amorphous semiconductor of the lower parts 124, 126 of the first portions 120, 122 and then for the amorphous semiconductor of the second regions 116, 118 of the first layer 104. Furthermore, this annealing and the SPER recrystallisation approach also activate dopants implanted in the semiconductor of the second regions 116, 118 of the first layer 104 and the lower parts 124, 126 of the first portions 120, 122. After this recrystallisation and this activation of dopants, the second regions 116, 118 of the first layer 104 and the lower parts 124, 126 of the first portions 120, 122 comprise doped crystalline semiconductor (FIG. 5).

[0066] As a variant, this recrystallisation and this activation of dopants can be implemented by a laser to locally heat the semiconductor of the second portions 116, 118 of the first layer 104 and the upper parts 124, 126 of the first portions 120, 122.

[0067] During recrystallisation, the mesh parameter of the crystalline semiconductor of the first portions 120, 122 (of the upper parts 128, 130 in this case) is at least partially transferred to the semiconductor located under it, in other words the semiconductor of the second regions 116, 118 of the first layer 104. Thus, if the semiconductor of the first portions 120, 122 is strained, this strain can be transferred into the semiconductor in the second regions 116, 118 of the first layer 104.

[0068] As shown on FIG. 6, the entire semiconductor of the first portions 120, 122, in other words both the upper parts 128, 130 and the lower parts 124, 126, are then removed, in this case by etching. The thickness of the remaining portions of doped crystalline semiconductor is very similar to the thickness of the first region 106 that will form the channel of the transistor 100 and correspond to the second regions 116, 118 of the first layer 104.

[0069] A second dielectric spacer 138 is then made by depositing a dielectric material, for example SiN, around the first dielectric spacer 114. The thickness (along the direction parallel to the X axis shown on FIG. 7) of the second dielectric spacer 138 (formed from the removed dielectric material and the first dielectric spacer 114) is equal for example to between about 3 nm and 10 nm. This second dielectric spacer 138 is formed on portions of the second regions 116, 118 that are juxtaposed on the first region 106.

[0070] As a variant, the first dielectric spacer 114 can be removed and the second dielectric spacer 138 can then be made such that it occupies the volume previously occupied by the first dielectric spacer 114.

[0071] A second epitaxy is then implemented from the doped crystalline semiconductor of the second regions 116, 118 that is not covered by dielectric spacers 114, 138, forming second portions 140, 142 of doped crystalline semiconductor (in this case N type by phosphorus ions). These second portions 140, 142 and the second regions 116, 118 form the source and drain 144, 146 of the transistor 100 (see FIG. 7).

[0072] Like the first semiconducting portions 120, 122, the semiconductor of the second portions 140, 142 can be strained in tension. To achieve this, the semiconductor obtained by this second epitaxy can contain a low concentration of carbon, for example between about 0.5% and 2%, which confers a smaller mesh parameter on the crystalline semiconductor of the second portions 140, 142 than the mesh parameter of the semiconductor in the second regions 116, 118 that does not contain any carbon, thus conferring a tensile strain on the semiconductor of the second portions 140, 142, this tensile strain inducing a tensile strain in the channel of the transistor 100.

[0073] According to one advantageous variant of this embodiment, a step could be performed between production of the second dielectric spacer 138 and the second epitaxy, to etch an upper part of the second regions 116, 118 of the first layer 104. Thus, the second portions 140, 142 are then made on the remaining parts of the second regions 116, 118 and therefore partially facing the first region 106, in other words the channel of the transistor 100. This variant is particularly advantageous when the second portions 140, 142 comprise strained semiconductor (in tension in the case of an NMOS transistor) because this strain is then better reproduced in the channel of the transistor.

[0074] According to another variant of this first embodiment, the ion implantation(s) are done only in the semiconductor in the second regions 116, 118. In this case, after the recrystallisation annealing and activation of dopants, only the second regions 116, 118 comprise doped crystalline semiconductor, and the first portions 120, 122 are entirely composed of only slightly doped crystalline semiconductor.

[0075] FIGS. 8 to 12 show some of the steps in a second embodiment of a method of making the transistor 100.

[0076] The first step is to make a structure similar to that described with reference to FIG. 2. The next step is that an etching stop layer 202 resistant to the etching agent used to etch the semiconductor of the first portions 120, 122 that will be formed later (in other works with an etching rate by this etching agent lower than the etching rate at which the first portions 120, 122 are etched by this etching agent) is then deposited on the parts of the first layer 104 not covered by the gate 112 and the first dielectric spacer 114, in other words on the second regions 116, 118 of the first layer 104 (FIG. 8). When the first portions 120, 122 that will be made by epitaxy on the etching stop layer 202 comprise silicon, the etching stop layer 202 advantageously comprises SiGe for which the proportion of germanium is between about 10% and 50%, for example. For example, the thickness of the etching stop layer 202 is between 1 nm and 5 nm. The semiconductor of the etching stop layer 202 may or may not be doped (in situ doping during epitaxy). The etching stop layer 202 may comprise SiGe for an NMOS or PMOS type transistor 100, or it may comprise SiGe:B for a PMOS type transistor 100.

[0077] As shown on FIG. 9, epitaxy is then used in a manner similar to that described above with reference to FIG. 3, forming the first portions 120, 122 in this case containing silicon.

[0078] As described above, the semiconductor of portions 120, 122 can be strained in tension and/or N type doped for an NMOS type transistor 100 or P type doped for a PMOS type transistor 100.

[0079] Then, as described above with reference to FIG. 4, one or several ion implantation steps are implemented such that dopants are implanted throughout the thickness of the second regions 116, 118 of the first layer 104 and in the lower portions 124, 126 of the first portions 120, 122. These dopants are also implanted in the semiconductor in the etching stop layer 202 that is located between the second regions 116, 118 of the first layer 104 and the first portions 120, 122. Preferably, in making a P type transistor, a first implantation of non-doping species such as silicon and/or germanium (for amorphisation) is made, and doping species such as boron are then implanted during a second implantation. As a variant, the second implantation can be made before the first implantation. P and/or As ions can be used to make an N type transistor.

[0080] The crystalline silicon and SiGe in which this ion implantation is made become amorphous semiconductors, the total thickness of the semiconductor thus formed being more than the thickness of the first layer 140 (FIG. 10). This ion implantation can be implemented with parameters similar to those described above for the first embodiment.

[0081] As shown on FIG. 11, a recrystallisation of the amorphous semiconductor of the second regions 116, 118 of the first layer 104 and the lower parts 124, 126 of the first portions 120, 122 and an activation of dopants implanted in this amorphous semiconductor by SPER are then done through the use of annealing of the entire structure and/or local recrystallisation by laser, as described above with reference to FIG. 5. Apart from the semiconductor of the second regions 116, 118 of the first layer 104 and the semiconductor of the lower parts 124, 126 of the first portions 120, 122 that has become crystalline and then doped subsequent to this recrystallisation and this activation by SPER, the semiconductor of the etching stop layer 202 is also recrystallised and becomes the doped semiconductor after these steps have been performed.

[0082] As shown on FIG. 12, the first portions 120, 122 are then removed by etching. Due to the presence of the etching stop layer 202, this etching is well controlled and the second regions 116, 118 of the first layer 104 are not deteriorated and therefore the thickness of semiconductor remaining adjacent to the first region 106 can be well controlled.

[0083] A second etching is then done to eliminate the etching stop layer 202. The structure obtained at this stage of the method corresponds to the structure previously described with reference to FIG. 6, however with the second regions 116, 118 of the first layer 102 having very regular upper surfaces not damaged by the etching used to remove the first portions 120, 122.

[0084] As a variant, the etching stop layer 202 may be kept.

[0085] The transistor 100 is then completed in a manner similar to that described above for the first embodiment with reference to FIG. 7, in other words forming the second dielectric spacer 138 and performing epitaxy forming the second portions 140, 142 (on the second regions 116, 118 of the first layer 102 if the etching stop layer 202 is removed or on the etching stop layer 202 if it is kept), thus completing production of the source and drain 144, 146 of the transistor 100.

[0086] The different variant embodiments described above for the first embodiment can also be applied to this second embodiment.

[0087] We will now describe the steps in a third embodiment of making the transistor 100 with reference to FIGS. 13 to 18.

[0088] The transistor 100 in this case is of the PMOS type and is made from a semiconductor on insulator type substrate, for example of the SGOI (silicon-germanium on insulator) type and comprising a thick semiconducting layer (not shown on FIGS. 13 to 18) on which a buried oxide dielectric layer 102 is located, for example comprising SiO.sub.2 and the thickness of which may for example be of the order of 25 nm, itself covered with a first layer of semiconductor 104, in this case comprising SiGe for example with a thickness of between about 5 nm and 8 nm. The SiGe in the first layer 104 may or may not be strained in compression, and the proportion of germanium can be between about 20% and 60% or between about 20% and 40%.

[0089] The gate 112 and the first spacer 114 are then made on the first layer 104, as described earlier for the first embodiment.

[0090] As in the second embodiment, the etching stop layer 202 is then made on the first layer 104, on the second regions 116, 118 (FIG. 14). The etching stop layer 202 may included doped or undoped SiGe. In the case in which the etching stop layer 202 is made of doped SiGe, the doping in this case is of the P type (since in this case the transistor 100 is of the PMOS type) and is obtained for example by implantation of boron ions. The thickness of the etching stop layer 202 may be between about 1 nm and 5 nm, and the proportion of germanium in SiGe can be between about 10% and 50%. When the first layer 104 comprises SiGe, as is the case here, the etching stop layer 202 can comprise SiGe with a proportion of germanium equal to at least 5% more than the proportion of germanium in the SiGe in the first layer 104.

[0091] As shown on FIG. 15, a first epitaxy is performed as described above in the previous two embodiments, forming the first portions 120, 122, in this case composed of silicon. The semiconductor of the first portions 120, 122 may be doped as P type or may be undoped, and may be strained in compression.

[0092] Then as described above with reference to FIGS. 4 and 10, one or several ion implantation steps are implemented such that dopants are implanted throughout the thickness of the second regions 116, 118 of the first layer 104 (FIG. 16). In this case, the dopants are also implanted in the etching stop layer 202. A first implantation of Si.sup.+ or Ge.sup.+ ions followed by a second implantation of B.sup.+ or BF.sub.2.sup.+ type doping ions, for example, are performed in this third embodiment. The semiconductor of the second regions 116, 118 of the first layer 104 in which this ion implantation is performed becomes an amorphous semiconductor. This ion implantation can be implemented with parameters (energy, dose, etc.) similar to those described above for the first and second embodiments, and adapted as a function of the thicknesses and the materials present. The implantation energies and doses used to form the amorphous semiconductor can be adapted such that the SiGe in the second regions 116, 118 of the first layer 104 and the etching stop layer 202 are amorphised selectively relative to the silicon in the first portions 120, 122. Examples of the implementation of such selected amorphisation are described for example in the document written by T. W. Simpson et al., “Amorphization threshold in Si-implanted strained SiGe alloy layers ”, Microstructure of Irradiated Materials, 1994, Meeting of the Materials Research Society.

[0093] As shown on FIG. 17, a recrystallisation of the amorphous semiconductor of the second regions 116, 118 of the first layer 104 (and also of the etching stop layer 202) and an activation of dopants implanted in this amorphous semiconductor are then done through the use of annealing of the entire structure and/or local recrystallisation by laser, as described above with reference to FIGS. 5 and 11. If the crystalline semiconductor of the first portions 120, 122 is silicon and the crystalline semiconductor of the first layer 104 is SiGe, the recrystallised SiGe in the second regions 116, 118 may become at least partially strained in compression.

[0094] As shown on FIG. 18, the first portions 120, 122 are then removed by etching. Due to the presence of the etching stop layer 202, this etching is well controlled and the second regions 116, 118 of the first layer 104 are not deteriorated and therefore the thickness of semiconductor remaining adjacent to the first region 106 is well controlled.

[0095] A second etching is then done to eliminate the etching stop layer 202.

[0096] The transistor 100 is then completed in a manner similar to that described above for the first embodiment with reference to FIG. 7, in other words forming the second dielectric spacer 138 and performing a second epitaxy forming the second portions 140, 142, thus completing production of the source and drain 144, 146 of the transistor 100. In this third embodiment, the semiconductor of the second portions 140, 142 is advantageously P type doped SiGe (for example with boron ions) in which the proportion of germanium is between about 20% and 60% such that the source and drain 144, 146 are strained in compression and that there is a compression strain in the channel of the transistor 100.

[0097] The different variant embodiments described above for the first and second embodiments can also be applied to this second embodiment. Moreover, this third embodiment can be used without making the etching stop layer 202.