SILICON CARBIDE SEMICONDUCTOR DEVICE AND METHOD OF MANUFACTURING SILICON CARBIDE SEMICONDUCTOR DEVICE

Abstract

A silicon carbide semiconductor device, including a silicon carbide semiconductor structure, an insulated gate structure, an interlayer insulating film formed on the insulated gate structure, a poly-silicon film formed on the interlayer insulating film, and a main electrode formed on the poly-silicon film and in electrical connection with the silicon carbide semiconductor structure. The insulated gate structure includes a gate insulating film, which is a silicon dioxide film contacting the silicon carbide semiconductor structure, and a gate electrode formed on the gate insulating film.

Claims

1. A silicon carbide semiconductor device, comprising: a silicon carbide semiconductor structure; an insulated gate structure, including a gate insulating film, which is a silicon dioxide film contacting the silicon carbide semiconductor structure, and a gate electrode formed on the gate insulating film; an interlayer insulating film formed on the insulated gate structure; a poly-silicon film formed on the interlayer insulating film; and a main electrode formed on the poly-silicon film and in electrical connection with the silicon carbide semiconductor structure.

2. The silicon carbide semiconductor device according to claim 1, further comprising: a titanium film provided between the poly-silicon film and the main electrode.

3. The silicon carbide semiconductor device according to claim 2, further comprising: a titanium nitride film provided between the titanium film and the main electrode; another titanium film provided between the titanium nitride film and the main electrode; and an alloy layer formed between the another titanium film and the main electrode due to reaction between the another titanium film and the main electrode.

4. The silicon carbide semiconductor device according to claim 1, wherein the poly-silicon film has a thickness of 0.2 μm to 1.0 μm.

5. The silicon carbide semiconductor device according to claim 1, wherein the poly-silicon film has a thickness of 0.5 μm to 1.0 μm.

6. The silicon carbide semiconductor device according to claim 1, wherein the main electrode is a first main electrode; the silicon carbide semiconductor device further includes: a silicon carbide semiconductor substrate, and an n-type drift layer and a second main electrode, respectively provided on two opposite surfaces of the silicon carbide semiconductor substrate; the silicon carbide semiconductor structure includes: a p-type semiconductor region selectively formed in the n-type drift layer, and an n-type semiconductor region selectively provided in the p-type semiconductor region; the gate insulating film covers a surface of a portion of the p-type semiconductor region between the n-type drift layer and the n-type semiconductor region; and the first main electrode is electrically connected to the n-type semiconductor region.

7. The silicon carbide semiconductor device according to claim 6, wherein the silicon carbide semiconductor substrate is of an n-type and has an impurity concentration that is higher than that of the n-type drift layer.

8. A method of manufacturing a silicon carbide semiconductor device, comprising: providing a silicon carbide semiconductor base having a silicon carbide semiconductor structure formed thereon; forming an insulated gate structure, including thermally oxidizing the silicon carbide semiconductor structure to form a silicon dioxide film on a surface thereof, and forming a gate electrode on the silicon dioxide film; forming an interlayer insulating film to cover the insulated gate structure; forming a poly-silicon film on the interlayer insulating film; and forming a main electrode on the poly-silicon film and electrically connecting the main electrode to the silicon carbide semiconductor structure.

9. The method of claim 8, wherein the forming of the main electrode on the poly-silicon film includes forming a titanium film on the poly-silicon film; and forming the main electrode on the titanium film.

10. The method of claim 9, further comprising: after forming the main electrode, causing the titanium film and the main electrode to react, to thereby form an alloy layer between the titanium film and the main electrode.

11. The method of claim 9, wherein the forming of the main electrode on the titanium film further includes: forming a titanium nitride film on the titanium film, and forming another titanium film on the titanium nitride film; and after forming the main electrode, causing the another titanium film and the main electrode to react, to thereby form an alloy layer between the another titanium film and the main electrode.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0022] FIG. 1 is a cross-sectional view of a structure of a silicon carbide semiconductor device according to a first embodiment;

[0023] FIG. 2 is a cross-sectional view of a structure of the silicon carbide semiconductor device according to a second embodiment;

[0024] FIG. 3 is a cross-sectional view of a structure of the silicon carbide semiconductor device according to a third embodiment;

[0025] FIG. 4 is a cross-sectional view of a structure of a silicon carbide semiconductor device of a comparison example 1;

[0026] FIG. 5 is a cross-sectional view of a structure of a silicon carbide semiconductor device of a comparison example 2; and

[0027] FIG. 6 is a cross-sectional view of a structure of a conventional silicon carbide semiconductor device.

DETAILED DESCRIPTION OF THE INVENTION

[0028] Embodiments of a silicon carbide semiconductor device and method of manufacturing a silicon carbide semiconductor device according to the present invention will be described in detail with reference to the accompanying drawings. In the present description and accompanying drawings, layers and regions prefixed with n or p mean that majority carriers are electrons or holes. Additionally, + or − appended to n or p means that the impurity concentration is higher or lower, respectively, than layers and regions without + or −. In the description of the embodiments below and the accompanying drawings, main portions that are identical will be given the same reference numerals and will not be repeatedly described. Further, in the present description, when Miller indices are described, “-” means a bar added to an index immediately after the “-”, and a negative index is expressed by prefixing “-” to the index.

[0029] A structure of the silicon carbide semiconductor device according to a first embodiment will be described taking a planar gate structure SiC-vertical MOSFET as an example. FIG. 1 is a cross-sectional view of the structure of the silicon carbide semiconductor device according to the first embodiment. In FIG. 1, an active region that bears driving current (region through which current flows in an on-state) is depicted for one unit cell (semiconductor element functional unit), and other unit cells arranged adjacent to this unit cell and an edge termination structure portion surrounding the active region are omitted from the drawing (similarly, in FIGS. 2, 3). The edge termination structure portion is a region that mitigates the electric field of a base front surface side of an n.sup.−-type drift layer 2 and, for example, has an edger termination structure combined with a guard ring, a field plate, RESURF, and the like.

[0030] In the silicon carbide semiconductor device depicted in FIG. 1, on a front surface of an n.sup.+-type silicon carbide substrate 1 that becomes an n.sup.+-type drain region, an n.sup.−-type silicon carbide epitaxial layer that becomes the n.sup.−-type drift layer 2 is deposited. In a surface layer on an opposite side of the n.sup.−-type drift layer 2 from an n.sup.+-type silicon carbide substrate 1 side, a p-type semiconductor region 3 is selectively provided. On a surface on an opposite side of the n.sup.−-type drift layer 2 from the n.sup.+-type silicon carbide substrate 1 side, a p.sup.−-type silicon carbide epitaxial layer that becomes a p.sup.−-type well layer 4 is deposited so as to cover the p-type semiconductor region 3. The p-type semiconductor region 3 and the p.sup.−-type well layer 4 function as a base region (p-type semiconductor region).

[0031] The p-type semiconductor region 3 has an impurity concentration that, for example, may be higher than an impurity concentration of the p.sup.−-type well layer 4. As a result, when a high reverse bias is applied to a pn junction between the p-type semiconductor region 3 and the n.sup.−-type drift layer 2, punchthrough of the p.sup.−-type well layer 4 may be prevented. In the p.sup.−-type well layer 4, an n.sup.−-type region (JFET region) 7 that penetrates the p.sup.−-type well layer 4 in a depth direction and reaches the n.sup.−-type drift layer 2 is selectively formed. In other words, the JFET region 7 is provided on the surface of a portion of the n.sup.−-type drift layer 2 between adjacent p-type semiconductor regions 3 and has functions as a drift region with the n.sup.−-type drift layer 2. The JFET region 7 has an impurity concentration that may be, for example, higher than an impurity concentration of the n.sup.−-type drift layer 2 to reduce the JFET resistance.

[0032] In a portion of the p.sup.−-type well layer 4 facing the p-type semiconductor region 3 in the depth direction, a p.sup.+-type contact region 5 and an n.sup.+-type source region (n-type semiconductor region) 6 are selectively provided. The p.sup.+-type contact region 5 is provided so as to contact the n.sup.+-type source region 6, the p.sup.+-type contact region 5 contacting the n.sup.+-type source region 6 on an opposite side from a JFET region 7 side, for example. Further, the p.sup.+-type contact region 5 may be provided so as to penetrate the p.sup.−-type well layer 4 and reach the p-type semiconductor region 3. On the surface of a portion of the p.sup.−-type well layer 4 between the JFET region 7 and the n.sup.+-type source region 6, a gate electrode 9 is provided so as to extend onto the JFET region 7 and onto the n.sup.+-type source region 6, via a gate insulating film 8.

[0033] The gate insulating film 8 is a silicon dioxide (SiO.sub.2) film formed by thermally oxidizing a front surface (surface of a p.sup.−-type well layer 4 side) of a silicon carbide semiconductor base (stacked base constituted by the sequentially stacked n.sup.+-type silicon carbide substrate 1, n.sup.−-type drift layer 2, and p.sup.−-type well layer 4). In this manner, on the front surface side of the silicon carbide semiconductor base (semiconductor chip), a MOS gate structure constituted by silicon carbide semiconductor portions (i.e., a silicon carbide semiconductor structure, including semiconductor regions such as the p-type semiconductor regions 3, the p.sup.−-type well layer 4, the n.sup.+-type source region 6, and the like), the gate insulating film 8, and the gate electrode 9 is provided. An interlayer insulating film 10 is provided so as to cover the gate electrode 9. A contact hole that penetrates the interlayer insulating film 10 in the depth direction and reaches the p.sup.+-type contact region 5 and the n.sup.+-type source region 6 is provided. On a silicon carbide semiconductor portion exposed in the contact hole, a nickel silicide (NiSi) layer 11 that forms an ohmic contact with the silicon carbide semiconductor portion is provided.

[0034] On the surface of the interlayer insulating film 10 and the nickel silicide layer 11, a poly-silicon (poly-Si) film 16 is provided. The poly-silicon film 16 has a function of absorbing hydrogen (H) atoms/hydrogen ions generated from inside a source electrode 12 described hereinafter and having aluminum (Al) as a main constituent. Therefore, hydrogen atoms/hydrogen ions generated from inside the source electrode 12 are suppressed from moving into an interface (SiO.sub.2/SiC interface) of the gate insulating film 8 and the silicon carbide semiconductor portion, or into the gate insulating film 8. As a result, the generation of positive charge by hydrogen atoms/hydrogen ions maybe suppressed. Hydrogen atoms/hydrogen ions are particles having hydrogen atoms as the smallest constituent unit and in particular, are hydrogen atoms, hydrogen ions, and hydrogen molecules.

[0035] Further, the poly-silicon film 16 is electrically insulated from the gate electrode 9 by the interlayer insulating film 10 and functions as source wiring. The poly-silicon film 16 may have a thickness about, for example, 0.2 μm or more and 1.0 μm or less, and more particularly may have a thickness of 0.5 μm or more. The reason is as follows. By making the thickness of the poly-silicon film 16 0.2 μm or more, the hydrogen (H) atoms/hydrogen ions absorbing function is sufficiently obtained.

[0036] On the surface of the poly-silicon film 16, a source electrode (first main electrode) 12 having aluminum as a main constituent is provided so as to be embedded in the contact hole. The source electrode 12 is electrically connected to the p.sup.+-type contact region 5 and the n.sup.+-type source region 6, via the poly-silicon film 16 and the nickel silicide layer 11. The source electrode 12 functions as source wiring. On the source electrode 12, a passivation protective film 14 that protects the chip front surface is provided. On a rear surface of silicon carbide semiconductor base (surface on the n.sup.+-type silicon carbide substrate 1 side, i.e., a rear surface of the n.sup.+-type silicon carbide substrate 1), a rear electrode 13 that becomes a drain electrode (second main electrode) is provided.

[0037] The method of manufacturing a silicon carbide semiconductor device according to the first embodiment will be described. First, the n.sup.+-type silicon carbide substrate (semiconductor wafer) 1 of four layer periodic hexagonal silicon carbide (4H-SiC) and becoming the n.sup.+-type drain region, for example, is prepared. The front surface of the n.sup.+-type silicon carbide substrate 1 may be, for example, a (000-1) face (so-called, C-face). Next, on the front surface of the n.sup.+-type silicon carbide substrate 1, the n.sup.−-type drift layer 2 doped with, for example, 5×10.sup.15/cm.sup.3 of an n-type impurity such as nitrogen (N) is deposited (formed) by epitaxial growth to have a thickness of, for example, about 10 μm.

[0038] In the surface layer of the n.sup.−-type drift layer 2, the p-type semiconductor region 3 is selectively formed by ion implantation of a p-type impurity. On the n.sup.−-type drift layer 2, the p.sup.−-type well layer 4 doped with, for example, a p-type impurity such as 5×10.sup.15/cm.sup.3 of aluminum (Al) is deposited by epitaxial growth so as to cover the p-type semiconductor region 3 and have a thickness of, for example, about 0.5 μm. By the processes up to here, the silicon carbide semiconductor base (epitaxial wafer) constituted by the sequentially stacked n.sup.+-type silicon carbide substrate 1, n.sup.−-type drift layer 2, and p.sup.−-type well layer 4 is formed.

[0039] Next, for example, in the p.sup.−-type well layer 4, the JFET region 7 is selectively formed by ion implantation of an n-type impurity such as nitrogen, the JFET region 7 penetrating the p.sup.−-type well layer 4 in the depth direction (base depth direction) and reaching the n.sup.−-type drift layer 2. Next, for example, in the p.sup.−-type well layer 4, the n.sup.+-type source region 6 is selectively formed away from the JFET region 7 by ion implantation of an n-type impurity such as phosphorus (P). Further, for example, in the p.sup.−-type well layer 4, for example, the p.sup.+-type contact region 5 is selectively formed by ion implantation of a p-type impurity such as aluminum, the p.sup.+-type contact region 5 being formed so as to contact the n.sup.+-type source region 6. Next, for example, activation annealing (heat treatment) at a temperature of 1600 degrees C. in an atmosphere of argon (Ar) is performed.

[0040] Next, for example, on the surface of a portion of the p.sup.−-type well layer 4 between the n.sup.+-type source region 6 and the JFET region 7, the gate insulating film 8 is formed to have a thickness of, for example, about 70 nm, the gate insulating film 8 being formed by thermal oxidation in a nitrous oxide (N.sub.2O) atmosphere. Next, on the gate insulating film 8, a poly-silicon layer that becomes the gate electrode 9 is formed. Next, on the entire front surface (surface on the p.sup.−-type well layer 4 side) of the silicon carbide semiconductor base, the interlayer insulating film 10 is formed so as to cover the gate electrode 9. Next, a contact hole is formed by photolithography and etching to penetrate the interlayer insulating film 10 in the depth direction, exposing the p.sup.+-type contact region 5 and the n.sup.+-type source region 6 in the contact hole. Next, on the silicon carbide semiconductor portion exposed in the contact hole, a nickel (Ni) film is formed, silicon carbide semiconductor portion and the nickel film are caused to react by sintering (heat treatment), and the nickel silicide layer 11 is formed to have a thickness of, for example, 1.0 μm. Before the sintering, on the entire rear surface of the silicon carbide semiconductor base (rear surface of the n.sup.+-type silicon carbide substrate 1), the rear electrode 13 is formed.

[0041] Next, for example, by low-pressure chemical vapor deposition (CVD), on the interlayer insulating film 10 and the nickel silicide layer 11, for example, the poly-silicon film 16 doped with an n-type impurity such as phosphorus is deposited (formed) to have a thickness of, for example, about 0.2 μm. Next, for example, the poly-silicon film 16 is subject to annealing (heat treatment) at a temperature of about 800 degrees C. As a result, sheet resistance of the poly-silicon film 16, for example, becomes about 50Ω/□ (Ω/square) or less. By depositing the poly-silicon film 16 by low-pressure CVD, the poly-silicon film 16 may be formed to be stable, resistant to cracking, and to have favorable step coverage.

[0042] Next, for example, on the poly-silicon film 16, metal layer (hereinafter, aluminum layer) having aluminum as a main constituent and becoming the source electrode 12 is deposited by a sputtering method to have, for example, a thickness of 5.0 μm. Next, the source wiring of a predetermined pattern is formed by pattering the source electrode 12 by photolithography and etching. Next, on the source electrode 12, a polyimide layer that becomes the passivation protective film 14 is formed and, for example, the passivation protective film 14 is hardened (cured) by heat treatment at a temperature of about 380 degrees C. Thereafter, silicon carbide semiconductor base is cut (diced) into individual chips whereby the SiC-vertical MOSFET depicted in FIG. 1 is completed.

[0043] In the silicon carbide semiconductor device according to the first embodiment, although hydrogen atoms/hydrogen ions are generated from inside the source electrode 12 under high-temperature operation, these hydrogen atoms/hydrogen ions are absorbed by the poly-silicon film 16 beneath the source electrode 12. Therefore, the hydrogen atoms/hydrogen ions generated from inside the source electrode 12 do not diffuse near the gate insulating film 8 or inside the gate insulating film 8. SiC-vertical MOSFETs were produced according the method of manufacturing a silicon carbide semiconductor device according to the first embodiment described above and according to various conditions given as examples, and variation of threshold voltage was measured with respect to the SiC-vertical MOSFETs. The results confirmed that after negative voltage of −3 MV/cm is applied to the gate electrode 9 for 1000 hours under high-temperature operation in which the operating temperature is 200 degrees C., the variation range of the threshold voltage could be suppressed to 0.1V or less.

[0044] As described, according to the first embodiment, by provided the poly-silicon film beneath the source electrode, hydrogen atoms/hydrogen ions generated from inside the source electrode under high-temperature operation are absorbed by the poly-silicon film beneath the source electrode. Therefore, the hydrogen atoms/hydrogen ions generated from inside the source electrode under high-temperature operation may be suppressed from moving to the gate insulating film side and diffusing into the gate insulating film. As a result, the generation of positive charge near the interface of the gate insulating film and the silicon carbide semiconductor portion or inside the gate insulating film may be suppressed, enabling the variation range of the threshold voltage when negative voltage is applied to the gate electrode to be reduced. In other words, a gate insulating film for which the threshold voltage is stable may be formed. Therefore, even in cases where either positive or negative voltage is applied to the gate electrode, variation of the threshold voltage may be suppressed, enabling a highly reliable silicon carbide semiconductor device to be provided that has stable electrical characteristics. Further, according to the first embodiment, by forming the poly-silicon film beneath the source electrode by CVD, a stable poly-silicon film that is resistant to cracking and has favorable step coverage may be formed thereby, enabling variation of the threshold voltage to be suppressed.

[0045] A structure of the silicon carbide semiconductor device according to a second embodiment will be described. FIG. 2 is a cross-sectional view of the structure of the silicon carbide semiconductor device according to the second embodiment. The silicon carbide semiconductor device according to the second embodiment differs from the silicon carbide semiconductor device according to the first embodiment in that between the poly-silicon film 16 and the source electrode 12, a metal film (hereinafter, titanium film) 15 having titanium (Ti) as a main constituent is provided. The titanium film 15 has a function of blocking hydrogen atoms/hydrogen ions generated from inside the source electrode 12. In other words, a diffusion coefficient of the hydrogen atoms/hydrogen ions inside the titanium film 15 is small to an extent that hydrogen atoms/hydrogen ions moving in the titanium film 15 substantially do not reach the poly-silicon film 16 below.

[0046] The method of manufacturing a silicon carbide semiconductor device according to the second embodiment, for example, is a method where in the method of manufacturing a silicon carbide semiconductor device according to the first embodiment, after the poly-silicon film 16 is formed and before the source electrode 12 is formed, the titanium film 15 is deposited (formed) on the poly-silicon film 16. In other words, the poly-silicon film 16, the titanium film 15, and the source electrode 12 are sequentially deposited as source wiring. Respective thicknesses of the poly-silicon film 16, the titanium film 15, and the source electrode 12 at the time of deposition, for example, may be 0.3 μm, 0.1 μm, and 5.0 μm, respectively. An alloy layer 19 is formed between the titanium film 15 and the source electrode 12 by causing the titanium film 15 and the source electrode 12 to react.

[0047] As described, according to the second embodiment, effects identical to those of the first embodiment may be obtained. Further, according to the second embodiment, hydrogen atoms/hydrogen ions generated from inside the source electrode under high-temperature operation are blocked by the titanium film beneath the source electrode. Therefore, movement of the hydrogen atoms/hydrogen ions from inside the source electrode to the gate insulating film side may be suppressed, enabling variation of the threshold voltage to be further reduced.

[0048] A structure of the silicon carbide semiconductor device according to a third embodiment will be described. FIG. 3 is a cross-sectional view of the structure of the silicon carbide semiconductor device according to the third embodiment. The silicon carbide semiconductor device according to the third embodiment differs from the silicon carbide semiconductor device according to the second embodiment in that a metal film (hereinafter, titanium nitride film) 17 having titanium nitride (TiN) as a main constituent, and a second titanium film 18 are provided between the titanium film (hereinafter, first titanium film) 15 and the source electrode 12. The titanium nitride film 17 and the second titanium film 18 have a function of absorbing hydrogen atoms/hydrogen ions generated from inside the source electrode 12.

[0049] The method of manufacturing a silicon carbide semiconductor device according to the third embodiment, for example, is a method where in the method of manufacturing a silicon carbide semiconductor device according to the first embodiment, after the first titanium film 15 is formed and before the source electrode 12 is formed, the titanium nitride film 17 and the second titanium film 18 are deposited (formed) on the first titanium film 15. In other words, the poly-silicon film 16, the first titanium film 15, the titanium nitride film 17, the second titanium film 18, and the source electrode 12 are sequentially deposited as source wiring. Although not depicted, the structure may be one in which the second titanium film 18 is not provided. Respective thicknesses of the poly-silicon film 16, the first titanium film 15, the titanium nitride film 17, the second titanium film 18, and the source electrode 12 at the time of deposition, for example, may be 0.3 μm, 0.1 μm, 0.1 μm, 0.1 μm, and 5.0 μm, respectively. The alloy layer 19 is formed between the second titanium film 18 and the source electrode 12 by causing the second titanium film 18 and the source electrode 12 to react.

[0050] As described, according to the third embodiment, effects identical to those of the first and the second embodiment may be obtained. Further, according to the third embodiment, movement of hydrogen atoms/hydrogen ions from inside the source electrode to the gate insulating film side may be suppressed by the titanium nitride film and second titanium film between the source electrode and the first titanium film, enabling variation of the threshold voltage to be further reduced.

[0051] An example will be described. The cause of variation of the threshold voltage in a conventional SiC-MOSFET (hereinafter, conventional example; refer to FIG. 6) will be described. In the conventional example, the interface state density of the SiO.sub.2/SiC interface being high is a problem particular to the SiO.sub.2/SiC interface and at present, it is not clear if the interface state density is high from differences in the band structure, the amount of distortion, and the amount of defects of the SiO.sub.2/SiC interface. Thus, arrangement of the aluminum layer formed as electrode layers was variously changed and the cause of threshold voltage variation in the conventional example was verified. FIG. 4 is a cross-sectional view of the structure of a silicon carbide semiconductor device of a comparison example 1. FIG. 5 is a cross-sectional view of the structure of a silicon carbide semiconductor device of a comparison example 2. As depicted in FIG. 4, a SiC-horizontal MOSFET (hereinafter, the comparison example 1) of a planar gate structure was produced in which the electrode layer and the interlayer insulating film 30 do not contact each other in the contact hole, and no electrode layer (aluminum layer) is arranged on an interlayer insulating film 30.

[0052] The comparison example 1 is a horizontal-type MOS gate structure of the conventional example. Thicknesses and impurity concentrations of regions of the comparison example 1 are respectively the same as the thicknesses and the impurity concentrations of the corresponding regions of the conventional example. Further, in the comparison example 1, an electrode layer (a source electrode 32a and a drain electrode 32b) and the interlayer insulating film 30 are arranged so as to not contact each other. In particular, first, on a front surface of an n.sup.+-type silicon carbide substrate 21, a silicon carbide epitaxial layer that becomes an n.sup.−-type drift layer 22 is deposited. Next, in the surface layer of the n.sup.−-type drift layer 22, a p-type semiconductor region 23 is formed by ion implantation. On the p-type semiconductor region 23, a silicon carbide epitaxial layer that becomes a p.sup.−-type well layer 24 is deposited.

[0053] In the p.sup.−-type well layer 24, an n.sup.+-type source region 26a and an n.sup.+-type drain region 26b are each selectively formed by ion implantation of phosphorus. Further, in the p.sup.−-type well layer 24, a p.sup.+-type contact region 25a is selectively formed by ion implantation of aluminum. The p.sup.+-type contact region 25a is arranged at a position that is farther away from a gate electrode 29 described hereinafter than the n.sup.+-type source region 26a is, the p.sup.+-type contact region 25a being arranged so as to contact the n.sup.+-type source region 26a. Next, activation annealing is performed at a temperature of 1600 degrees C. in an argon atmosphere.

[0054] Next, on a surface of a portion of the p.sup.−-type well layer 24 between the n.sup.+-type source region 26a and the n.sup.+-type drain region 26b, a gate insulating film 28 is formed by thermal oxidation in a nitrous oxide atmosphere. On the gate insulating film 28, a poly-silicon layer that becomes the gate electrode 29 is formed. The interlayer insulating film 30 is formed so as to cover the gate electrode 29. Next, first and second contact holes are formed by photolithography and etching to penetrate the interlayer insulating film 30 in a depth direction, exposing the p.sup.+-type contact region 25a and the n.sup.+-type source region 26a in the first contact hole and exposing the n.sup.+-type drain region 26b in the second contact hole.

[0055] Next, on each silicon carbide semiconductor portion exposed in the first and the second contact holes, a nickel film is formed and nickel silicide layers 31a, 31b are formed by performing sintering and causing the silicon carbide semiconductor portions and the nickel films to react. Before the sintering, on a rear surface of the n.sup.+-type silicon carbide substrate 21, a rear electrode 33 is formed. On the interlayer insulating film 30 and the nickel silicide layers 31, an aluminum layer is deposited and patterned to leave only in the first and second contact holes, the aluminum layer forming the source electrode 32a and the drain electrode 32b. At this time, the source electrode 32a and the drain electrode 32b are formed to be away from the interlayer insulating film 30 so as to not contact the interlayer insulating film 30. Thereafter, the silicon carbide semiconductor base is cut into individual chips whereby the SiC-horizontal MOSFET of the comparison example 1 depicted in FIG. 4 is completed.

[0056] Further, as depicted in FIG. 5, a SiC-horizontal MOSFET of a planar gate structure in which an electrode layer 32c of the source potential is arranged on the surface of a portion of the interlayer insulating film 30 on the gate electrode 29 (hereinafter, comparison example 2) was produced. Configuration of the comparison example 2 excluding the electrode layer 32c is the same as that of the comparison example 1. A method of manufacturing the comparison example 2 may be a method of further leaving the aluminum layer on the interlayer insulating film 30 to form the electrode layer 32c when the aluminum layer deposited on the interlayer insulating film 30 and the nickel silicide layers 31 is patterned and the source electrode 32a and the drain electrode 32b are formed. The electrode layer 32c faces the gate electrode 29 in the depth direction, via the interlayer insulating film 30.

[0057] Further, the comparison example 2 is structured so that a distance X from an end of the electrode layer 32c to an end of the gate electrode 29 is 100 μm or more, and so that hydrogen atoms/hydrogen ions generated from the interface of the interlayer insulating film 30 and the electrode layer 32c or from inside the electrode layer 32c are suppressed from diffusing from near the end of the gate electrode 29 into the gate insulating film 28. The distance X from the end of the electrode layer 32c to the end of the gate electrode 29 is a distance from the end of the source electrode 32a side of the electrode layer 32c to the end on the source electrode 32a side of the gate electrode 29, and a distance from the end of the drain electrode 32b side of the electrode layer 32c to the end of the drain electrode 32b side of the gate electrode 29.

[0058] With respect to the comparison examples 1, 2, variation of the threshold voltage was measured after applying a negative voltage of −3 MV/cm to the gate electrode 29 for 10 minutes under high-temperature operation in which the operating temperature was 200 degrees C. The result for both of the comparison examples 1, 2 was a threshold voltage variation range of ±0.1V. In the comparison example 1 of a configuration in which the electrode layer (the source electrode 32a and the drain electrode 32b) and the interlayer insulating film 30 do not contact each other in this manner, it was confirmed that the threshold voltage does not vary. Further, it was confirmed that even in a configuration in which the electrode layer and the interlayer insulating film contact each other, the threshold voltage does not vary consequent to the arrangement of the electrode layer 32c on the surface of the portion of the interlayer insulating film 30 on the gate electrode 29 as in the comparison example 2.

[0059] Further, with respect to the comparison example 2 in which the interlayer insulating film 30 and the electrode layer 32c contact each other, elemental analysis of the interface of the interlayer insulating film 30 and the electrode layer 32c, and of the inside of the electrode layer 32c was performed by thermal desorption spectroscopy (TDS). The results reveal that when the chip temperature was raised to 200 degrees C. or higher, hydrogen molecules at an impurity concentration of 3×10.sup.14/cm.sup.2 or higher were detected. The generation of hydrogen atoms/hydrogen ions from the interface of the interlayer insulating film 30 and the electrode layer 32c and from the electrode layer 32c is thought to occur consequent to a reaction of aluminum, which is a constituent material of the electrode layer 32c and water (H.sub.2O) included in the vapor atmosphere during thermal oxidation.

[0060] Further, poly-silicon, which is a constituent material of the gate electrode 29 has a grain boundary and at this grain boundary, a dangling bond is present. A dangling bond of the grain boundary of poly-silicon is known to be terminated by hydrogen atoms when hydrogen is introduced. Therefore, the threshold voltage of the comparison example 2 not varying indicates that hydrogen atoms/hydrogen ions generated from the interface of the interlayer insulating film 30 and the electrode layer 32c or from inside the electrode layer 32c terminate dangling bonds of the grain boundary of poly-silicon, which is a constituent material of the gate electrode 29, whereby the hydrogen atoms/hydrogen ions are absorbed by the poly-silicon and substantially, do not diffuse into the gate insulating film 28.

[0061] Further, in general, when a SiC-MOSFET is manufactured, a majority of the hydrogen ions are taken in by the SiO.sub.2/SiC interface consequent to an annealing process at a high temperature of 800 degrees C. or higher, or a thermal oxidation process for oxide film formation at a high temperature of 800 degrees C. or higher. The hydrogen ions taken in by the SiO.sub.2/SiC interface consequent to the high-temperature heat treatment of 800 degrees C. or higher, bond with dangling bonds of the SiO.sub.2/SiC interface, forming silicon-hydrogen (Si—H) bonds and carbon-hydrogen (C—H) and becoming fixed. In this manner, hydrogen atoms of silicon-hydrogen bonds and carbon-hydrogen bonds formed at the SiO.sub.2/SiC interface by high-temperature heat treatment do not change (dissociate) under low-temperature heat treatment of 400 degrees C. or less.

[0062] On the other hand, the electrode layer (aluminum layer for wiring) is deposited on the interlayer insulating film by low-heat treatment of 400 degrees C. or less. When the electrode layer is deposited by low-temperature heat treatment, hydrogen atoms/hydrogen ions generated from the interface of the interlayer insulating film and electrode layer or from inside the electrode layer are not fixed and move to the SiO.sub.2/SiC interface when negative voltage is applied to the gate electrode of the SiC-MOSFET under high-temperature operation. It is surmised that consequent to these hydrogen atoms/hydrogen ions, fixed hydrogen atoms dissociate from silicon-hydrogen bonds and carbon-hydrogen bonds of the SiO.sub.2/SiC interface, form dangling bonds (Si.sup.+, C.sup.+) of silicon atoms and carbon atoms, and generate positive charge near the SiO.sub.2/SiC interface or inside the gate insulating film.

[0063] For example, the diffusion coefficient of the hydrogen atoms/hydrogen ions in the oxide film (SiO.sub.2 film) at 200 degrees C. is 1.0×10.sup.−8 cm.sup.2/second and the diffusion length thereof is 24.5 μm for 10 minutes. Therefore, as in the conventional example, when the interlayer insulating film 110 and the source electrode 112 contact each other in the contact hole, the hydrogen atoms/hydrogen ions generated from the interface of the interlayer insulating film 110 and the source electrode 112 or from inside the source electrode 112 under high-temperature operation easily move into the interlayer insulating film 110 and reach the gate insulating film 108 from the end side of the gate electrode 109, causing the threshold voltage to vary.

[0064] Such movement of hydrogen atoms/hydrogen ions from inside the electrode layer to the interlayer insulating film, for example, may be blocked by forming a titanium film or a titanium nitride film between the interlayer insulating film and the electrode layer, however, the titanium film or the titanium nitride film, for example, is formed by sputtering and thus, step coverage becomes poor. Further, although a SiC-vertical MOSFET having a structure in which the interlayer insulating film and the electrode layer do not contact each other in the contact hole is possible, the size of the unit cell (unit region in which one MOS gate structure is formed) increases consequent to a gap created between the electrode layer and the contact hole. Consequently, practical use is difficult.

[0065] As described, in the present invention, between the interlayer insulating film 10 and the source electrode 12, at least the poly-silicon film 16 is formed that has a function of absorbing hydrogen atoms/hydrogen ions generated from inside the source electrode 12. As a result, for example, when the poly-silicon film 16 has a thickness of 0.3 μm, variation of the threshold voltage after a negative voltage of −3 MV/cm is applied to the gate electrode 9 for 1000 hours under high-temperature operation where the operating temperature is 200 degrees C. may be suppressed to ±0.1V. Further, in the present invention, formation of the poly-silicon film 16 between the interlayer insulating film 10 and the source electrode 12 enables formation of the source electrode 12 without creating a gap between the electrode layer (the source electrode 12) and the side wall of the contact hole whereby increases in the unit cell size may be avoided.

[0066] Although the present invention, for example, is particularly effective for elements forming a channel (inversion layer) in a (000-1)-face of a 4-layer periodic hexagonal crystal silicon carbide (4H-SiC) semiconductor (i.e., elements having a C-face as the chip front surface), the same effects are achieved in elements forming a channel in other plane orientations (e.g., (0001)-face (a so-called Si-face), (11-20)-face, (03-38)-face). Further, in the embodiments, although a SiC-vertical MOSFET has been described as an example, the invention is applicable to other MOS-type silicon carbide semiconductor devices such as SiC-horizontal MOSFETs, SiC-IGBTs, etc., and the same effects are obtained. In place of the planar gate structure, even in the case of a trench gate structure, the same effects are obtained. Further, the structure may be such that without providing the p.sup.−-type well layer, the p.sup.+-type contact region and the n.sup.+-type source region are selectively formed in the p.sup.−type semiconductor region that functions as a base region. Further, in the embodiments described above, for example, dimensions, impurity concentrations, and the like of constituent parts may be variously set according to required specifications.

[0067] However, as a result of extensive research by the inventors, it was found that when the oxide film that becomes the gate insulating film of a SiC-MOSFET is formed by thermal oxidation in an atmosphere that includes nitrous oxide and negative voltage is applied to the gate electrode, the threshold voltage (Vth) varies greatly. In putting a silicon carbide semiconductor device into practical use, achieving high reliability enabling stable operation even during stress application (voltage, temperature) is problematic. For example, in a SiC-power MOSFET, at the time of driving, positive voltage and negative voltage that are also high voltages are applied to the gate electrode. Further, a SiC-power MOSFET has to guarantee operation under high-temperature environments where the junction (joining) temperature is 200 degrees C. or higher.

[0068] In particular, electric field strength of about ±2 MV/cm to ±4 MV/cm is applied to the gate insulating film and a guaranteed operating temperature of about 200 degrees C. is required. In this case, under certain conditions, a phenomenon of the threshold voltage greatly varying has been observed. Hereinafter, the results of verification of electrical characteristics of the SiC-power MOSFET obtained by a reliability test will be described. First, according to the conventional method of manufacturing a silicon carbide semiconductor device described above, a SiC-MOSFET was produced (manufactured) under the various conditions described above (hereinafter, conventional example). Concerning this conventional example, under high-temperature operation in which the operating temperature (junction temperature) becomes 200 degrees C., 3 MV/cm (positive voltage) and −3 MV/cm (negative voltage) were each applied for 10 minutes to the gate electrode 109 and threshold voltage variation was observed.

[0069] As a result, when positive voltage was applied to the gate electrode 109, variation of the threshold voltage was small and the variation range (amount of variation) was confirmed to be ±0.1V or less. The variation range of the threshold voltage is the difference from the threshold voltage (reference value) at the time of product shipping, based on design conditions. On the other hand, when negative voltage was applied to the gate electrode 109, the threshold voltage varied greatly on the negative side (i.e., the threshold voltage decreased). This phenomenon of the threshold voltage varying on the negative side indicates that near the junction interface (SiO.sub.2/SiC interface) of the gate insulating film 108 and the silicon carbide semiconductor portion or in the gate insulating film 108 (SiO.sub.2 film), positive charge (holes) are captured and charge, generating a positive fixed charge consequent to the application of negative voltage to the gate electrode 109 under high-temperature operation.

[0070] There are few reports of the phenomenon of positive fixed charge being generated in the gate insulating film or at a junction interface (hereinafter, SiO.sub.2/Si interface) of the gate insulating film and the silicon semiconductor portion when negative voltage is applied to the gate electrode in a Si-MOSFET or a Si-insulated gate bipolar transistor (IGBT) using a silicon semiconductor. For example, although there are some reports of the phenomenon (slow trap phenomenon) of the gate threshold voltage varying in a Si-p channel-type MOSFET when negative voltage is applied to the gate electrode, the variation range of the threshold voltage is 0.1V even when the operating temperature is 150 degrees C. and a negative voltage of −3 MV/cm is applied to the gate electrode for 1000 hours.

[0071] The variation range of the threshold voltage of the SiC-MOSFET under the same conditions (operating temperature of 150 degrees C., gate voltage of −3 MV/cm) is −7V or more and therefore, the variation range of the threshold voltage for the Si-MOSFET and that for the SiC-MOSFET greatly differ. In particular, the interface state density of the SiO.sub.2/Si interface in the Si-MOSFET is 1.0×10.sup.11 cm.sup.−2 eV.sup.−1 or less. On the other hand, the interface state density of the SiO.sub.2/SiC interface in the SiC-MOSFET is 1.0×10.sup.12 cm.sup.−2 eV.sup.−1 or more. Although a majority of research is on reducing the interface state density of the SiO.sub.2/SiC interface, a technique of reducing the interface state density of the SiO.sub.2/SiC interface to about the same extent as the interface state density of the SiO.sub.2/Si interface has not been reported.

[0072] According to the present invention, under high-temperature operation, hydrogen atoms/hydrogen ions generated from inside the first main electrode are absorbed by the poly-silicon film beneath the first main electrode and therefore, movement of the hydrogen atoms/hydrogen ions to the gate insulating film and diffusion into to the gate insulating film may be suppressed. As a result, the generation of positive charge at the interface of the gate insulating film and the silicon carbide semiconductor portion or inside the gate insulating film may be suppressed, enabling the variation range of the threshold voltage when negative voltage is applied to the gate electrode to be reduced. Further, according to the present invention, since hydrogen atoms/hydrogen ions generated from inside the first main electrode are blocked by the first titanium film between the poly-silicon film and the first main electrode, movement of the hydrogen atoms/hydrogen ions to the gate insulating film side may be further suppressed. As a result, the variation range of the threshold voltage when negative voltage is applied to the gate electrode may be further reduced. Further, according to the present invention, since hydrogen atoms/hydrogen ions generated from inside the first main electrode are absorbed by the second titanium film and the titanium nitride film between the first titanium film and the first main electrode, movement of the hydrogen atoms/hydrogen ions to the gate insulating film side may be further suppressed. As a result, the variation range of the threshold voltage when negative voltage is applied to the gate electrode may be further reduced.

[0073] The silicon carbide semiconductor device and the method of manufacturing a silicon carbide semiconductor device according to the present invention achieve an effect in that predetermined electrical characteristics may be stably obtained and reliability may be improved.

[0074] As described, the silicon carbide semiconductor device and the method of manufacturing a silicon carbide semiconductor device according to the present invention is useful for power semiconductor devices used in control circuits of automobiles, industrial machines, and the like.

[0075] Although the invention has been described with respect to a specific embodiment for a complete and clear disclosure, the appended claims are not to be thus limited but are to be construed as embodying all modifications and alternative constructions that may occur to one skilled in the art which fairly fall within the basic teaching herein set forth.