Patent Application
20160056040
The present invention relates to a method for manufacturing a silicon carbide semiconductor device, and more particularly to a method for manufacturing a silicon carbide semiconductor device having an electrode capable of achieving a lowered contact resistance.
In order to allow a semiconductor device to be high in breakdown voltage and low in loss, silicon carbide has recently increasingly been adopted as a material forming a semiconductor device. Silicon carbide is a wide band gap semiconductor greater in band gap than silicon which has conventionally widely been used as a material forming a semiconductor device. Therefore, by adopting silicon carbide as a material forming a semiconductor device, a higher breakdown voltage and lowering in ON resistance of a semiconductor device can be achieved. A semiconductor device containing silicon carbide as a material is also advantageous in that lowering in characteristics when it is used in a high temperature environment is less than in a semiconductor device containing silicon as a material.
As a method for manufacturing an electrode in a semiconductor device containing silicon carbide, for example, Japanese Patent Laying-Open No. 2012-253291 (PTD 1) describes formation of an electrode through heat treatment in an argon gas after a metal film is formed on a silicon carbide substrate. Japanese Patent Laying-Open No. 2012-146838 (PTD 2) describes a contact electrode provided on a silicon carbide substrate, which contains Al atoms, Ti atoms, and Si atoms. Japanese Patent Laying-Open No. 2012-99599 (PTD 3) describes a method of forming an ohmic contact electrode by forming a Ti film on a silicon carbide substrate, forming an Al film on the Ti film, forming an Si film on the Al film, and thereafter carrying out annealing with laser beams.
Osamu Nakatsuka et al., “Low Resistance TiAl Ohmic Contacts with Multi-Layered Structure for p-Type 4H—SiC,” Materials Transactions, Vol. 43, No. 7, 2002, pp. 1684-1688 (NPD 1) describes influence by an Al concentration on electrical characteristics of a TiAl electrode establishing low-resistance ohmic contact with p-type silicon carbide. According to this document, electrical characteristics of the TiAl electrode are affected not by the number of TiAl layers but by a concentration of Al in the TiAl electrode.
Z. Q. Guan et al., “Phase formation during ball milling and subsequent thermal decomposition of Ti—Al—Si powder blends,” Journal of Alloys and Compounds, 252, 1997, pp. 245-251 (NPD 2) describes investigation about phase formation of a TiAlSi powdery mixture. According to this document, a decomposition process in two stages in a mechanically alloyed TiAlSi alloy has been found.
When an electrode in contact with a silicon carbide layer is formed with the methods described in Japanese Patent Laying-Open No. 2012-253291 (PTD 1), Japanese Patent Laying-Open No. 2012-146838 (PTD 2), and Japanese Patent Laying-Open No. 2012-99599 (PTD 3), a low contact resistance to an n-type region in the silicon carbide layer is obtained, however, a contact resistance to a p-type region in the silicon carbide layer has not sufficiently been low in some cases.
The present invention was made in view of the problems above, and an object thereof is to provide a method for manufacturing a silicon carbide semiconductor device having an electrode capable of realizing a low contact resistance to both of a p-type region and an n-type region in a silicon carbide layer.
The present inventor has conducted dedicated studies about a cause of a contact resistance between an electrode containing Ti (titanium), Al (aluminum), and Si (silicon) and a p-type region in a silicon carbide layer not being sufficiently low in a case that the electrode is formed on the silicon carbide layer, and obtained the finding below and found the present invention. As a result of detailed investigation of a silicon carbide semiconductor device high in contact resistance and a silicon carbide semiconductor device low in contact resistance, it has been found that composition (a concentration) of Al at an interface between an electrode of the silicon carbide semiconductor device high in contact resistance and a silicon carbide layer is lower than composition (a concentration) of Al at an interface between an electrode of the silicon carbide semiconductor device low in contact resistance and a silicon carbide layer.
The cause of a low concentration of Al at the interface between the electrode and the silicon carbide layer is estimated as follows. After a metal layer containing Ti, Al, and Si is formed on a silicon carbide layer, the metal layer is annealed at a temperature around 1000° C. Then, an electrode establishing ohmic contact with the silicon carbide layer is formed. A eutectic point of AlSi is around 577° C. and a melting point of Al is around 660° C. Therefore, during a period of temperature increase of the metal layer from a room temperature to 1000° C., liquefaction of AlSi starts at a temperature around 577° C. and liquefaction of Al starts at a temperature around 660° C. Therefore, it is expected that, in a temperature range from around 577° C. to around 1000° C., Al evaporates from a surface of the electrode and some Al separates from the electrode, and consequently a concentration of Al at the interface between the electrode and the silicon carbide layer lowers.
As Al separates and a concentration of Al at the interface lowers, a concentration of Al which diffuses into the silicon carbide layer also lowers. Al functions as a p-type dopant for silicon carbide, and increase in concentration of a dopant in a p-type region brings about a low contact resistance. Therefore, as a concentration of Al at the interface between the electrode and the silicon carbide layer is lower, a contact resistance between the electrode and the p-type region in the silicon carbide layer is higher. As a result of dedicated studies conducted by the inventors, the inventors have found that, by forming a metal layer including a first region containing an Al element and an Si element and a second region containing a Ti element and formed on the first region and thereafter annealing the metal layer, an electrode achieving prevented loss of Al and having a low contact resistance to both of a p-type region and an n-type region can be formed. Namely, by arranging the second region containing the Ti element which is highest in melting point among Ti, Al, and Si (Ti having a melting point around 1670° C. and Si having a melting point around 1414° C.) on the first region containing the Al element, the second region functions as a Al separation prevention layer.
A method for manufacturing a silicon carbide semiconductor device according to the present invention has the following steps. A silicon carbide layer having a main surface and including a p-type region and an n-type region in contact with the p-type region is prepared. A metal layer in contact with the p-type region and the n-type region at the main surface is formed. After the step of forming a metal layer, the p-type region, the n-type region, and the metal layer are annealed. The step of forming a metal layer includes the steps of forming a first region in contact with the p-type region and the n-type region at the main surface and forming a second region arranged to be in contact with a surface of the first region opposite to a surface in contact with the main surface. The first region has an aluminum element and a silicon element. The second region has a titanium element.
As is clear from the description above, according to the present invention, a method for manufacturing a silicon carbide semiconductor device having an electrode capable of realizing a low contact resistance to both of a p-type region and an n-type region in a silicon carbide layer can be provided.
An embodiment of the present invention will be described hereinafter with reference to the drawings. It is noted that, in the drawings below, the same or corresponding elements have the same reference characters allotted and description thereof will not be repeated. In addition, regarding crystallographic denotation herein, an individual orientation, a group orientation, an individual plane, and a group plane are shown in [ ], < >, ( ), and { }, respectively. Moreover, a crystallographically negative index is normally expressed by a number with a bar “-” thereabove, however, a negative sign herein precedes a number. In expressing an angle, a system in which a total azimuth angle is defined as 360 degrees is employed.
Overview of an embodiment of the present invention will initially be described in (1) to (8) below.
(1) A method for manufacturing a silicon carbide semiconductor device 1 according to an embodiment has the following steps. A silicon carbide layer 10 having a main surface 10a and including a p-type region 18 and an n-type region 14 in contact with p-type region 18 is prepared. A metal layer 16 in contact with p-type region 18 and n-type region 14 at main surface 10a is formed. After the step of forming metal layer 16, p-type region 18, n-type region 14, and metal layer 16 are annealed. The step of forming metal layer 16 includes the steps of forming a first region 16a in contact with p-type region 18 and n-type region 14 at main surface 10a and forming a second region 16b arranged in contact with a surface 16a5 of first region 16a opposite to a surface 16a4 in contact with main surface 10a. First region 16a has an aluminum element and a silicon element. Second region 16b has a titanium element.
According to the method for manufacturing silicon carbide semiconductor device 1 according to the present embodiment, after metal layer 16 in which second region 16b containing the titanium element is arranged on first region 16a having the aluminum element and the silicon element is formed, metal layer 16 is annealed. Therefore, since metal layer 16 is annealed while aluminum is covered with titanium, evaporation of aluminum and separation thereof from metal layer 16 can be prevented. Consequently, a silicon carbide semiconductor device having an electrode capable of realizing a low contact resistance to both of the p-type region and the n-type region in the silicon carbide layer can be manufactured.
(2) In the method for manufacturing silicon carbide semiconductor device 1 according to the embodiment, preferably, first region 16a further contains the titanium element. Thus, a silicon carbide semiconductor device having an electrode capable of realizing a lower contact resistance to both of the p-type region and the n-type region in the silicon carbide layer can be manufactured.
(3) In the method for manufacturing silicon carbide semiconductor device 1 according to the embodiment, preferably, the step of forming first region 16a includes the following steps. A first layer 16a1 being in contact with p-type region 18 and n-type region 14 and containing the titanium element is formed. A second layer 16a2 being in contact with first layer 16a1 and containing the aluminum element is formed. A third layer 16a3 being in contact with second layer 16a2 and containing the silicon element is formed. Since third layer 16a3 containing the silicon element is formed on first layer 16a1 containing the aluminum element, evaporation of aluminum can efficiently be suppressed. Consequently, a contact resistance between p-type region 18 and the electrode can be lowered.
(4) In the method for manufacturing silicon carbide semiconductor device 1 according to the embodiment, preferably, first layer 16a1 has a thickness not smaller than 140 angstroms and not greater than 340 angstroms. Thus, a contact resistance between the electrode and the n-type region and a contact resistance between the electrode and the p-type region can effectively be lowered.
(5) In the method for manufacturing silicon carbide semiconductor device 1 according to the embodiment, preferably, second layer 16a2 has a thickness not smaller than 190 angstroms and not greater than 390 angstroms. Thus, a contact resistance between the electrode and the p-type region can effectively be lowered.
(6) In the method for manufacturing silicon carbide semiconductor device 1 according to the embodiment, preferably, third layer 16a3 has a thickness not smaller than 230 angstroms and not greater than 430 angstroms. Thus, a contact resistance between the electrode and the n-type region can effectively be lowered.
(7) In the method for manufacturing silicon carbide semiconductor device 1 according to the embodiment, preferably, second region 16b further contains the silicon element. Thus, oxidation of titanium contained in second region 16b can be suppressed. As a result of oxidation of silicon, silicon dioxide is produced, however, silicon dioxide can readily be removed with hydrofluoric acid.
(8) In the method for manufacturing silicon carbide semiconductor device 1 according to the embodiment, preferably, the step of forming second region 16b includes the following steps. A fourth layer 16b1 being in contact with first region 16a and containing the titanium element is formed. A fifth layer 16b2 being in contact with fourth layer 16b1 and containing the silicon element is formed. Thus, oxidation of titanium contained in second region 16b can effectively be suppressed.
(9) In the method for manufacturing silicon carbide semiconductor device 1 according to the embodiment, preferably, the step of forming second region 16b includes the step of forming a layer containing a titanium silicide alloy. Thus, oxidation of titanium contained in second region 16b can effectively be suppressed.
(10) In the method for manufacturing silicon carbide semiconductor device 1 according to the embodiment, preferably, the step of forming second region 16b includes the step of forming a layer containing a titanium carbon alloy. Thus, oxidation of titanium contained in second region 16b can effectively be suppressed.
(11) In the method for manufacturing silicon carbide semiconductor device 1 according to the embodiment, preferably, second region 16b has a thickness not smaller than 200 angstroms and not greater than 300 angstroms. When second region 16b has a thickness Tb in a range not smaller than 200 angstroms and not greater than 300 angstroms, evaporation of aluminum contained in first region 16a can effectively be suppressed and productivity of silicon carbide semiconductor device 1 can be improved.
An embodiment of the present invention will now be described in further detail.
Initially, a construction of a metal oxide semiconductor field effect transistor (MOSFET) representing the silicon carbide semiconductor device according to one embodiment of the present invention will be described.
Referring to
Single-crystal substrate 11 is composed, for example, of hexagonal silicon carbide having a poly type of 4H. Single-crystal substrate 11 contains such an impurity as N (nitrogen) at a high concentration and has an n-type. Epitaxial layer 12 is a layer resulting from epitaxial growth of silicon carbide. Epitaxial layer 12 is formed on one main surface of single-crystal substrate 11, for example, to a thickness approximately from 10 to 15 μm.
Drift layer 9 included in epitaxial layer 12 contains such an impurity as N (nitrogen) and has an n conductivity type. Such an impurity as nitrogen contained in drift layer 9 is higher in concentration than an n-type impurity contained in single-crystal substrate 11. A concentration of such an impurity as nitrogen contained in drift layer 9 is, for example, around 5×1015 cm−3. First main surface 10a may be a surface angled off by approximately at most 10°, for example, relative to a {0001} plane.
P body 13 is a region having a p conductivity type. P body 13 is formed in epitaxial layer 12 to be in contact with first main surface 10a. P body 13 contains such an impurity as Al (aluminum) or B (boron). A concentration of an impurity contained in p body 13 is, for example, around 1×1017 cm−3.
N+ source region 14 is a region having the n conductivity type. N+ source region 14 is formed in p body 13 to be in contact with first main surface 10a and to be surrounded by p body 13. N+ source region 14 contains such an impurity as P (phosphorus) at a concentration higher than that of the n-type impurity contained in drift layer 9, for example, at a concentration of 1×1020 cm−3.
P+ region 18 is a region having the p conductivity type. P+ region 18 is in contact with n+ source region 14, first main surface 10a, and p body 13. P+ region 18 is formed to pass through n+ source region 14 from first main surface 10a to p body 13. P+ region 18 contains such an impurity as Al or B at a concentration higher than that of the impurity contained in p body 13, for example, at a concentration of 1×1020 cm−3.
Gate insulating film 15 is formed on first main surface 10a of epitaxial layer 12 so as to be in contact with first main surface 10a and extend from an upper surface of one n+ source region 14 to an upper surface of the other n+ source region 14. Gate insulating film 15 is composed, for example, of silicon dioxide.
Gate electrode 17 is arranged to be in contact with gate insulating film 15, so as to extend over from one n+ source region 14 to the other n+ source region 14. Gate electrode 17 is composed of polysilicon doped with an impurity or such a conductor as Al.
Source contact electrode 16 has first surface 16a4 in contact with first main surface 10a and second surface 16b5 opposite to first surface 16a4. Source contact electrode 16 is in contact with gate insulating film 15, p+ region 18, and n+ source region 14. Source contact electrode 16 contains, for example, a titanium (Ti) element, an aluminum (Al) element, and a silicon (Si) element. Preferably, when a concentration profile of Ti in source contact electrode 16 is measured along a direction of normal to first main surface 10a, a region where a Ti concentration is highest is located close to second surface 16b5, relative to a position intermediate between first surface 16a4 and second surface 16b5.
Source interconnection 19 is formed to be in contact with source contact electrode 16, and formed, for example, of such a conductor as Ti/Al. Source interconnection 19 is electrically connected to n+ source region 14 through source contact electrode 16. Source interconnection 19 is formed to cover interlayer insulating film 21 which will be described later.
Drain electrode 20 is formed to be in contact with second main surface 10b of silicon carbide layer 10. Drain electrode 20 may be constructed similarly, for example, to source contact electrode 16, or may be formed of another material such as Ni having the n-type, which can establish ohmic contact with single-crystal substrate 11. Drain electrode 20 is electrically connected to single-crystal substrate 11. A backside pad electrode 23 composed, for example, of Ni/Au is formed to be in contact with drain electrode 20.
Interlayer insulating film 21 is formed to be in contact with gate insulating film 15 and gate electrode 17 and to cover gate electrode 17. Interlayer insulating film 21 is composed, for example, of silicon dioxide, and electrically isolates gate electrode 17 from the outside. A passivation film (not shown) may be formed on interlayer insulating film 21.
An operation of MOSFET 1 will now be described. While a voltage not higher than a threshold value is applied to gate electrode 17, that is, in an off state, a portion between p body 13 and epitaxial layer 12 located directly under gate insulating film 15 is reverse biased and rendered non-conducting. When a positive voltage is applied to gate electrode 17, an inversion layer is formed in a channel region around a portion of contact of p body 13 with gate insulating film 15. Consequently, n+ source region 14 and epitaxial layer 12 are electrically connected to each other and a current flows between a source electrode 22 and drain electrode 20.
A method for manufacturing MOSFET 1 in the present embodiment will now be described.
Initially, a silicon carbide substrate preparing step (S10:
Then, an oxide film composed of silicon dioxide is formed, for example, with chemical vapor deposition (CVD) on first main surface 10a of silicon carbide layer 10. After a resist is applied onto the oxide film, the resist is exposed and developed to form a resist film (not shown) having an opening in a region corresponding to a desired shape of p body 13. With the resist film serving as a mask, the oxide film is partially removed, for example, through reactive ion etching (RIE). Thus, a mask layer formed from the oxide film having an opening pattern is formed on epitaxial layer 12. Thereafter, after the resist film is removed, with this mask layer serving as a mask, such a p-type impurity as Al is introduced through ion implantation into first main surface 10a of silicon carbide layer 10. Thus, p body 13 is formed in epitaxial layer 12. Then, after the oxide film which has been used as the mask is removed, a mask layer having an opening in a region corresponding to a desired shape of n+ source region 14 is formed. Thereafter, with the mask layer serving as a mask, such an impurity as P (phosphorus) is introduced through ion implantation into epitaxial layer 12, to thereby form n+ source region 14. Then, a mask layer having an opening in a region corresponding to a desired shape of p+ region 18 is formed. With the mask layer serving as a mask, such an impurity as Al or B is introduced through ion implantation into epitaxial layer 12, to thereby form p+ region 18. P+ region 18 is formed to be in contact with n+ source region 14, first main surface 10a, and p body 13.
Then, heat treatment for activating an impurity introduced through ion implantation is carried out. Specifically, epitaxial layer 12 into which ions have been implanted is heated to a temperature around 1700° C., for example, in an Ar (argon) atmosphere and held for approximately 30 minutes. As above, silicon carbide layer 10 having first main surface 10a and including p+ region 18 and n+ source region 14 in contact with the p+ region is prepared.
Referring to
Then, a gate electrode forming step (S40:
Then, referring to
Then, a metal layer forming step (S80:
Referring to
Referring to
Referring to
Referring to
Referring to
Referring to
Then, an annealing step (S100:
Referring to
Then, referring again to
A function and effect of the method for manufacturing MOSFET 1 according to the present embodiment will now be described.
According to the method for manufacturing MOSFET 1 according to the present embodiment, after metal layer 16 in which second region 16b containing the titanium element is arranged on first region 16a having the aluminum element and the silicon element is formed, metal layer 16 is annealed. Therefore, since metal layer 16 is annealed while aluminum is covered with titanium, evaporation of aluminum and separation thereof from metal layer 16 can be prevented. Consequently, MOSFET 1 having source contact electrode 16 capable of realizing a low contact resistance to both of p+ region 18 and n+ source region 14 in silicon carbide layer 10 can be manufactured.
According to the method for manufacturing MOSFET 1 according to the present embodiment, first region 16a further contains the titanium element. Thus, MOSFET 1 having source contact electrode 16 capable of realizing a lower contact resistance to both of p+ region 18 and n+ source region 14 in silicon carbide layer 10 can be manufactured.
According to the method for manufacturing MOSFET 1 according to the present embodiment, in the step of forming first region 16a, first layer 16a1 being in contact with p-type region 18 and n-type region 14 and containing the titanium element is formed. Second layer 16a2 being in contact with first layer 16a1 and containing the aluminum element is formed. Third layer 16a3 being in contact with second layer 16a2 and containing the silicon element is formed. Since third layer 16a3 containing the silicon element is formed on first layer 16a1 containing the aluminum element, evaporation of aluminum can efficiently be suppressed. Consequently, a contact resistance between p+ region 18 and source contact electrode 16 can be lowered.
According to the method for manufacturing MOSFET 1 according to the present embodiment, first layer 16a1 has a thickness not smaller than 140 angstroms and not greater than 340 angstroms. Thus, a contact resistance between source contact electrode 16 and n+ source region 14 and a contact resistance between source contact electrode 16 and p+ region 18 can effectively be lowered.
According to the method for manufacturing MOSFET 1 according to the present embodiment, second layer 16a2 has a thickness not smaller than 190 angstroms and not greater than 390 angstroms. Thus, a contact resistance between source contact electrode 16 and p+ region 18 can effectively be lowered.
According to the method for manufacturing MOSFET 1 according to the present embodiment, third layer 16a3 has a thickness not smaller than 230 angstroms and not greater than 430 angstroms. Thus, a contact resistance between source contact electrode 16 and n+ source region 14 can effectively be lowered.
According to the method for manufacturing MOSFET 1 according to the present embodiment, second region 16b further contains the silicon element. Thus, oxidation of titanium contained in second region 16b can be suppressed. As a result of oxidation of silicon, silicon dioxide is produced, however, silicon dioxide can readily be removed with hydrofluoric acid.
According to the method for manufacturing MOSFET 1 according to the present embodiment, the step of forming second region 16b includes the following steps. Fourth layer 16b1 being in contact with first region 16a and containing the titanium element is formed. Fifth layer 16b2 being in contact with fourth layer 16b1 and containing the silicon element is formed. Thus, oxidation of titanium contained in second region 16b can effectively be suppressed.
According to the method for manufacturing MOSFET 1 according to the present embodiment, the step of forming second region 16b includes the step of forming a layer containing a titanium silicide alloy. Thus, oxidation of titanium contained in second region 16b can effectively be suppressed.
According to the method for manufacturing MOSFET 1 according to the present embodiment, the step of forming second region 16b includes the step of forming a layer containing a titanium carbon alloy. Thus, oxidation of titanium contained in second region 16b can effectively be suppressed.
According to the method for manufacturing MOSFET 1 according to the present embodiment, second region 16b has a thickness not smaller than 200 angstroms and not greater than 300 angstroms. When second region 16b has thickness Tb in a range not smaller than 200 angstroms and not greater than 300 angstroms, evaporation of aluminum contained in first region 16a can effectively be suppressed and productivity of MOSFET 1 can be improved.
It should be understood that the embodiment disclosed herein is illustrative and non-restrictive in every respect. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.
1 silicon carbide semiconductor device (MOSFET); 9 drift layer; 10 silicon carbide layer; 10a first main surface (main surface); 10b second main surface; 11 single-crystal substrate; 12 epitaxial layer; 13 p body; 14 n-type region (n+ source region); 15 gate insulating film (thermal oxidation film); 16 metal layer (source contact electrode); 16a first region; 16a1 first layer; 16a2 second layer; 16a3 third layer; 16a4 first surface; 16a5 surface; 16b second region; 16b1 fourth layer; 16b5 second surface; 16b3 sixth layer; 16b4 seventh layer; 16b2 fifth layer; 17 gate electrode; 18 p-type region (p+ region); 19 source interconnection; 20 drain electrode; 21 interlayer insulating film; 22 source electrode; and 23 backside pad electrode.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2013-093687 | Apr 2013 | JP | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/JP2014/055389 | 3/4/2014 | WO | 00 |
