SiGe p-channel tri-gate transistor based on bulk silicon and fabrication method thereof

Abstract

A p-channel tri-gate transistor has a silicon fin that protrudes from a bulk silicon substrate, a thin silicon-germanium active layer is formed on three sidewalls of the silicon fin, and a hole well is formed between the gate insulating film and the silicon fin in the active layer surrounded by the tri-gate by a valence band offset electric potential against the silicon fin for moving holes collected in the hole well along the active layer with a high hole-mobility. Thus, it is possible to have the effects of not only an ultra-high speed, low power operation, but also a body biasing through an integral structure of the silicon fin-body. The p-channel tri-gate transistor can be fabricated together with an n-channel FinFET transistor in one substrate by the same CMOS process.

Claims

1. A method for fabricating a p-channel tri-gate transistor comprising: a first step of depositing an etching stopper thin film on a bulk silicon substrate for a planarization process; a second step of forming an etching mask on the etching stopper thin film and etching the etching stopper thin film and the bulk silicon substrate for forming an etching stopper pattern, source/drain supporters and a silicon fin; a third step of depositing an isolation insulating film material on the bulk silicon substrate and polishing by a CMP process until the etching stopper pattern is exposed; a fourth step of further etching the isolation insulating film material to a predetermined depth and removing the etching stopper pattern for protruding the source/drain supporters and the silicon fin from the isolation insulating film to a predetermined height; a fifth step of forming a silicon-germanium active layer having a predetermined thickness on the protruded source/drain supporters and the protruded silicon fin; a sixth step of forming a gate insulating film on the active layer; and a seventh step of depositing a gate material on the gate insulating film and forming a tri-gate to surround three sidewalls of the silicon fin by patterning and etching, wherein the thickness of the silicon-germanium active layer is reduced by forming the gate insulating film in the sixth step.

2. The method of claim 1, wherein the etching mask of the second step is formed by a sidewall spacer patterning process or EUVL (extreme ultraviolet lithography) for forming the silicon fin with a width smaller than that of the source/drain supporter, and wherein the active layer of the fifth step is directly grown as a silicon-germanium layer having a germanium content on the exposed surfaces of the source/drain supporters and the silicon fin.

3. The method of claim 2, wherein the gate insulating film of the sixth step is a silicon oxide film generated from a surface of the active layer, and the active layer has increased germanium content toward the source/drain supporters and the silicon fin as a result of a germanium condensation process of the active layer by a thermal oxidation process.

4. The method of claim 3, wherein the gate insulating film further comprises a high- film having permittivity higher than that of the silicon oxide film on the silicon oxide film after forming the silicon oxide film by the thermal oxidation process.

5. The method of claim 4, wherein the thickness of the active layer is 1 to 5 nm.

6. The method of claim 1, wherein the silicon substrate is an intrinsic substrate doped with no impurity or an n-type substrate doped with n-type impurity, and wherein, after the seventh step, the method further comprises steps of removing a gate insulating film by the tri-gate as an etching mask and processing an ion implantation with a p-type impurity.

7. The method of claim 6, wherein the second step forms a plurality of etching masks in a row with a predetermined interval by a fineness pattern and forms a plurality of silicon fins side by side between the source and the drain supporters by etching the bulk silicon substrate through the plurality of etching masks.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIGS. 1 to 11 are a process perspective view showing a fabrication method of a transistor according to an embodiment of the present invention.

(2) FIG. 12 is a cross-sectional view taken along line AA of FIG. 11.

(3) FIG. 13 is showing an energy-band diagram (FIG. 13a) and a hole-current density diagram (FIG. 13b) in a cross-sectional view taken along line BB of FIG. 12 when the transistor of FIG. 12 is turned on by applying a negative voltage to the gate.

(4) FIG. 14 is a perspective view showing a transistor structure fabricated by a fabrication method according to another embodiment of the present invention.

(5) FIG. 15 is a cross-sectional view taken along line CC of FIG. 14.

(6) FIG. 16 is an electrical characteristic diagram showing a comparison between switching operations when the fraction x of the Si.sub.1-xGe.sub.x active layer is x=0(Si), 0.5(Si.sub.0.5Ge.sub.0.5) or 1(Ge) in the structure of FIG. 11.

(7) FIG. 17 is an electrical characteristic diagram showing a comparison between leakage currents of switching operations at a turn-off state when the fraction x of the Si.sub.1-xGe.sub.x active layer is x=0.9(Si.sub.0.1Ge.sub.0.9) or 1(Ge) in the structure of FIG. 11.

(8) In these drawings, the following reference numbers are used throughout: reference number 10 indicates a bulk silicon substrate, 12 a source supporter, 14 a drain supporter, 16 and 18 a silicon fin, 20 an etching stopper thin film, 22 an etching stopper pattern, 34 an isolation insulating film, 40 and 46 an active layer, 41 and 45 an active layer sidewall, 42 a source contact, 43 an active layer top surface, 44 a drain contact, 50, 52 and 53 a gate insulating film and 60 a tri-gate.

DETAILED DESCRIPTION

(9) Detailed descriptions of preferred embodiments of the present invention are provided below with accompanying drawings.

(10) A transistor according to an embodiment of the present invention comprises, as shown in FIGS. 11 and 12, a silicon fin 16 integrally formed on and protruded from a bulk silicon substrate 10, the silicon fin 16 providing a body contact to the bulk silicon substrate 10; an isolation insulating film 34 filled from the bulk silicon substrate 10 to a predetermined height (h1) of the silicon fin 16; an active layer 40 or 46 having a predetermined thickness (t1, t2) formed of Si.sub.1-x (0.2x<1) on both opposite sidewalls and a top surface of the silicon fin 16 and on the isolation insulating film 34; a gate insulating film 50 formed to surround three sidewalls 41, 43 and 45 of the active layer 40; and a tri-gate 60 formed on the isolation insulating film 34 to surround the gate insulating film 50, wherein a hole well is formed in the active layer 40(41) between the gate insulating film 50 and and the silicon fin 16 by a valence band offset electric potential (Ev) against the silicon fin 16 as shown in FIG. 13 and is controlled by the tri-gate 60.

(11) FIG. 13a is an energy-band diagram in a cross-sectional view taken along line BB of FIG. 12 when the transistor according to an embodiment of FIG. 12 is turned on by applying a negative voltage to the tri-gate 60 and FIG. 13b is showing a current density of holes that flow the active layer 40 or 41 at the turn-on state.

(12) The active layer 40 is preferred to be formed by Si.sub.1-xGe.sub.x (0.2x<1). In this case, more preferably, the germanium Ge fraction x is 0.5x<0.9.

(13) It is preferred that the germanium Ge fraction x of the active layer 40 is closer to 1 because the valence band offset electric potential (Ev) against the silicon fin 16 and the hole-mobility can be raised. However, if the fraction x is above 0.9 to be mostly germanium Ge, there is a problem that a leakage current of switching operation at a turn-off state can be relatively large as shown in FIG. 17.

(14) And when the germanium Ge fraction x is less than 0.2, there are some problems that because the valence band offset electric potential (Ev) against the silicon fin 16 can not be increased, it is difficult to capture and store enough amount of holes and that because the hole-mobility can be reduced to be subthreshold swing (SS) smaller than that of Si.sub.0.5Ge.sub.0.5 as known from FIG. 16, it is difficult to gain an ideal switching operation characteristic.

(15) The gate insulating film 50 can be a silicon oxide film (SiO.sub.2). According to embodiments, the gate insulating film 50 can further comprise a high- film having permittivity higher than that of the silicon oxide film on the silicon oxide film. The gate insulating film 50 can be also just a high- film having permittivity higher than that of the silicon oxide film without the silicon oxide film. Thus, it is possible to collect more holes in the hole well than the other cases without the hole well at the same applied voltage and to operate with a low voltage by an increasing gate controllability in the channel.

(16) The silicon fin 16 can have various sizes, but preferably, comprise sidewalls that are protruded from the isolation insulating film 34 having a height (h2) larger than a width (w1) of a top surface. In this case, it is possible to preferably increase a current density without an area expansion by increasing a width (w1) of a channel formed under a tri-gate 60 through the large area of both sidewalls 41 and 45 of an active layer 46.

(17) And, as shown in FIGS. 6 and 11, preferably, at both ends, the silicon fin 16 can be integrally formed together with source/drain supporters 12 and 14 protruded from and surrounded by the isolation insulating film 34 having a height equal to the height (h2) of the sidewalls of the silicon fin 16 and a width wider than the width (w1) of the top surface of the silicon fin 16 by etching the bulk silicon substrate 10.

(18) In this case, the active layer 40, as shown in FIGS. 7 and 11, can further comprise source/drain contacts 42 and 44 formed on the source/drain supporters 12 and 14 protruded from the isolation insulating film 34 with a thickness equal to that on the silicon fin 16.

(19) The tri-gate 60, as shown in FIG. 11, can have a width (w2) smaller than a height (h3) of a surrounding region of the silicon fin 16 and can be integrally formed with a gate contact 60 having a width wider than the width (w2) of the surrounding region.

(20) As another embodiment, as shown in FIGS. 14 and 15, a plurality of silicon fins 16 and 18 having a structure which is equal to that of the above mentioned silicon fin 16 can be arranged in a row with a predetermined interval as a shape of fingers between the source supporter 12 and the drain supporter 14 of the above mentioned embodiment. An active layer can be further formed with the same thickness on a plurality of silicon fins 16 and 18 that protrude over an isolation insulating film 34.

(21) In this case, gate insulating films 52 and 54 can be formed to surround three sidewalls of each of the active layers 46 and 48 that are formed on a plurality of silicon fins 16 and 18, respectively. And the above mentioned tri-gate 60 can be formed as one body on the gate insulating films 52 and 54 to commonly surround three sidewalls of each of the active layers 46 and 48 that are formed on the plurality of silicon fins 16 and 18, respectively.

(22) Thus, because a plurality of active layers 46 and 48 are arranged in a row between the source/drain contacts 42 and 44, it takes an advantage that an operating current can be raised to be able to drive a low power operation.

(23) In each embodiment mentioned above, the source/drain contacts 42 and 44 can be the same conduction type as the active layer 46 surrounding the silicon fin 16 to form a junctionless transistor. A channel region surrounded by the tri-gate 60 in the active layer 46 can be also doped with no impurity or n-type impurity to have concentration of less than 10.sup.18/cm.sup.3. And the active layer 46 except the channel region and the source/drain contacts 42 and 44 can be doped with p-type impurity to have concentration of 10.sup.16 to 10.sup.20/cm.sup.3. Thus, it is possible to form a channel by enough accumulated or inversed holes in the mentioned hole well.

(24) In the latter case, preferably, the substrate 10, the silicon fin 16 and the source/drain supporters 12 and 14 can be doped with no impurity or n-type impurity for having the same conduction type as the channel region in order that it is possible to provide a body contact to the channel region.

(25) In each embodiment described above, the thickness of the active layer 40 or 46 can be preferably 1 to 5 nm. If it is less than 1 nm, it is profitable to increase the controllability of the tri-gate 60, but there is a problem that the amount of current itself can be decreased. However, if it is more than 5 nm, there are some problems that the active layer 46 can be uselessly thicker and that it can lead to reduce the possibility of device reduction and increase the process time and cost.

(26) According to some related simulation results, when gate insulating film 52 is a silicon oxide film, the concentration of holes can be maximized at about 1 nm distance from a surface of Si.sub.1-xGe.sub.x active layer 40 or 46. Thus, more preferably, the thickness of the active layer 40 or 46 is 1 to 2 nm.

(27) Descriptions of embodiments related to methods for fabricating the mentioned transistors are provided below with accompanying FIGS. 1 to 11.

(28) First, as shown in FIG. 1, an etching stopper thin film 20 is deposited on a prepared bulk silicon substrate 10 for a planarization process (a first step). Here, the bulk silicon substrate 10 can be an intrinsic substrate doped with no impurity, but preferably, it is an n-type substrate doped weakly with n-type impurity to have concentration of less than 10.sup.18/cm.sup.3.

(29) Next, an etching mask (not shown) is formed on the etching stopper thin film 20 and the etching stopper thin film 20 and the silicon substrate 10 are etched for forming an etching stopper pattern 22, source/drain supporters and a silicon fin 16 (a second step).

(30) Here, the etching mask can be formed by a sidewall spacer patterning process or EUVL (extreme ultraviolet lithography) for forming the silicon fin 16 with a width smaller than that of the source/drain supporters.

(31) And, as shown in FIGS. 14 and 15, for fabricating a transistor having a plurality of silicon fins 16 and 18 between the source supporter 12 and the drain supporter 14, the plurality of etching masks can be formed to arrange in a row with a predetermined interval by a fineness pattern and then the silicon substrate 10 can be etched to form a plurality of silicon fins 16 and 18.

(32) Next, as shown in FIG. 3, an isolation insulating film material 30 is deposited on the silicon substrate 10 and, as shown in FIG. 4, planarized by a CMP process until the etching stopper pattern 22 is exposed (a third step).

(33) Next, as shown in FIG. 5, the planarized isolation insulating film material 32 is further etched to a predetermined depth for forming an isolating insulating film 34. And then, by removing the etching stopper pattern 22, as shown in FIG. 6, source/drain supporters 12 and 14 and a silicon fin 16 are protruded from the isolation insulating film 34 to a predetermined height (h2) (a fourth step).

(34) Next, as shown in FIG. 7, through a CVD, epitaxial process and the like, a silicon-germanium active layers 40, 42, 44 and 46 having a predetermined thickness is formed on the exposed source/drain supporters 12 and 14 and silicon fin 16 (a fifth step).

(35) Here, preferably, the silicon-germanium active layer can be directly grown with Si.sub.1-xGe.sub.x (0.2x<1) to have a thickness of 1 to 5 nm from the exposed source/drain supporters 12 and 14 and silicon fin 16.

(36) And a thickness (t1) of each of sidewalls and a thickness (t2) of a top surface of the silicon-germanium active layers 40, 42, 44 and 46 can be different from each other according to processes, but when it is grown by an epitaxial growth, it is possible to get an equal thickness. And as shown in FIG. 8, it is preferable to select a fabrication method that forms a thickness more than two times greater than that of the embodiment of FIG. 7, reduces the thickness by a germanium condensation process and is close a germanium Ge fraction x of Si.sub.1-xGe.sub.x to 1.

(37) At this time, a germanium condensation process can be performed by a thermal oxidation process. In this case, as shown in FIG. 9, a silicon oxide film is formed automatically as a gate insulating film 50 from the surface of an active layer 40. And the active layer 40 used to drive a device operation can have an increased germanium content close to 100% toward the source/drain supporters 12 and 14 and the silicon fin 16.

(38) As mentioned above, a germanium condensation can be used as a fabrication process of a gate insulating film 50. If it is not used, as shown in FIG. 9, a gate insulating film 50 is additionally formed on the active layer 40 (a sixth step).

(39) Here, the fabrication methods of the gate insulating film 50 can include a method for obtaining a thermal oxide film (a silicon oxide film) in a condensation process by a thermal oxidation mentioned above and a method for forming a high- film having permittivity higher than that of a silicon oxide film after surface passivation on the active layer 40 based on a plasma process. Of cause, after the silicon oxide film is formed by the condensation, a high- film having permittivity higher than that of a silicon oxide film can be further formed. In the process for directly forming a high- film on the active layer 40, the methods for a surface passivation or an interface passivation between an active layer and a gate insulating film can include a heat treatment with a gas and a treatment of plasma. It is possible to choice a suitable time to perform the surface and/or interface passivation before or after the gate insulting film formation.

(40) Next, a gate material is deposited on the gate insulating film 50, patterned and etched, as shown in FIG. 10, to form a tri-gate 60 surrounding three sidewalls of the silicon fin 16 (a seventh step). At this time, the gate material can be formed of a metal with a proper work function and not only silicon-based material such as poly-silicon, but also poly-silicon-germanium and poly-germanium.

(41) When a plurality of silicon fins 16 and 18 are formed between the source supporter 12 and the drain supporter 14 as shown in FIGS. 14 and 15, a tri-gate 60 can be integrally formed as one body on an isolation insulating film 34 to interlay each of gate insulating films 52 and 54 and to transverse a plurality of active layers 46 and 48 that surround a plurality of silicon fins 16 and 18, respectively.

(42) After the seventh step, as usual, after forming an interlayer insulating film (not shown), contact holes can be formed on source/drain contacts 42 and 44 and a gate contact to perform a wiring metal process. As shown in FIG. 11, the above step can be performed after removing the exposed gate insulating film 50 by using the tri-gate 60 as an etching mask, especially, it is preferable when the gate insulating film 50 is formed of a material such as a high- film that is different from the interlayer insulating film. And before and after removing the gate insulating film 50, it is possible to raise the conductivity in the tri-gate 60, the active layer 40 except the channel region and the source/drain contacts 42 and 44 by a further step of ion implantation with p-type impurity.

(43) As above mentioned, descriptions of preferred embodiments are provided with accompanying drawings, but it is not limited thereto. The embodiments can be variously modified. Because the modifications will be performed by the person of ordinary skill in the art with reference to the embodiments described above, the descriptions thereof are omitted.

(44) This work was supported by Business for Cooperative R&D between Industry, Academy, and Research Instituted funded by Korean Small and Medium Business Administration in 2015 (No. C0300518).