(12) United States Patent
`(10) Patent N0.:
`US 6,967,409 132
`
`Segawa et al.
`(45) Date of Patent:
`Nov. 22, 2005
`
`USOO6967409B2
`
`(54) SEMICONDUCTOR DEVICE AND METHOD
`OF MANUFACTURING THE SAME
`
`(56)
`
`(75)
`
`Inventors: Mizuki Segawa, Osaka (JP); IsaO
`Miyanaga, Osaka (JP); Toshiki Yabu,
`Osaka (JP); Takashi Nakabayashi,
`Osaka (JP); Takashi Uehara, Osaka
`(JP); Kiji Yamashita, Osaka (JP);
`Takaaki Ukeda, Osaka (JP); Masatoshi
`Arai, Osaka (JP); Takayuki Yamada,
`Osaka (JP); Michikazu Matsumoto,
`Osaka (JP)
`
`(73) Assigneei Matsushita Electric Industrial C0.,
`Ltd., Osaka (JP)
`
`( * ) Notice:
`
`Subject to any disclaimer, the term of this
`patent is extended or adjusted under 35
`U.S.C. 154(b) by 127 days.
`
`(21) Appl. No.: 10/454,682
`
`(22)
`
`Filed:
`
`Jun. 5, 2003
`
`(65)
`
`Prior Publication Data
`
`US 2003/0205820 A1 NOV. 6, 2003
`
`Related US. Application Data
`
`(62) Division of application No. 09/902,157, filed on Jul. 11,
`2001, now Pat. No. 6,709,950, which is a division of
`application No. 08/685,726, filed on Jul. 24, 1996, now Pat.
`No. 6,281,562.
`
`(30)
`
`Foreign Application Priority Data
`
`Jul. 27, 1995
`Dec. 19, 1995
`
`(JP)
`(JP)
`
`........................................... .. 7—192181
`........................................... .. 7—330112
`
`Int. Cl.7 .............................................. .. H01L 29/40
`(51)
`(52) US. Cl.
`..................................................... .. 257/774
`(58) Field Of Search ............................... .. 257/304, 510,
`257/774, 311
`
`References Cited
`US. PATENT DOCUMENTS
`
`4,578,128 A
`4,966,870 A
`5,177,028 A
`5,196,910 A
`
`3/1986 Mundt et al.
`10/1990 Barber et al.
`1/1993 Manning
`3/1993 Moriuchi et al.
`
`(Continued)
`FOREIGN PATENT DOCUMENTS
`
`EP
`EP
`
`0234988 A1
`0 243 988
`
`4/1987
`11/1987
`
`(Continued)
`
`Primary Examiner—Roy Potter
`(74) Attorney, Agent, or Firm—McDermott Will & Emery
`LLP
`
`(57)
`
`ABSTRACT
`
`An isolation which is higher in a stepwise manner than an
`active area of a silicon substrate is formed. On the active
`
`area, an FET including a gate oxide film, a gate electrode, a
`gate protection film, sidewalls and the like is formed. An
`insulating film is deposited on the entire top surface of the
`substrate, and a resist film for exposing an area stretching
`over the active area, a part of the isolation and the gate
`protection film is formed on the insulating film. There is no
`need to provide an alignment margin for avoiding interfer-
`ence with the isolation and the like to a region where a
`connection hole is formed. Since the isolation is higher in a
`stepwise manner than the active area, the isolation is pre-
`vented from being removed by over-etch in the formation of
`a connection hole to come in contact with a portion where
`an impurity concentration is low in the active area. In this
`manner, the integration of a semiconductor device can be
`improved and an area occupied by the semiconductor device
`can be decreased without causing degradation of junction
`voltage resistance and increase of a junction leakage current
`in the semiconductor device.
`
`81 Claims, 21 Drawing Sheets
`
`
`
`Exhibit 2076
`
`TSMC v. IP Bridge
`IPR2016-01246
`
`Page 1 of 40
`
`Page 1 of 40
`
`Exhibit 2076
`TSMC v. IP Bridge
`IPR2016-01246
`
`

`

`US 6,967,409 132
`
`Page 2
`
`US. PATENT DOCUMENTS
`
`FOREIGN PATENT DOCUMENTS
`
`5,286,674 A
`5,319,235 A
`5,384,281 A
`5,393,708 A
`5,397,910 A
`5,401,673 A
`5 413 961 A
`’
`’
`5,433,794 A
`5,497,016 A
`5,521,422 A
`5,561,311 A
`
`5,777,370 A
`5,804,862 A *
`6,022,781 A
`6,281,562 B1
`
`2/1994 Roth et 211.
`6/1994 Kihara et 211.
`1/1995 Kenney et 211.
`2/1995 Hsia et 211.
`Ishimaru
`3/1995 Urayama
`5/1995 Kim
`7/1995 Fazan et 211.
`3/1996 Koh
`5/1996 Mandelman et 211.
`10/1996 Hamamoto et a1.
`.
`7/1998 Omld—Zohoor et 211.
`9/1998 Matumoto ................ .. 257/396
`2/2000 Noble, Jr.
`8/2001 Segawa et a1.
`
`EP
`EP
`JP
`JP
`JP
`JP
`JP
`JP
`JP
`JP
`JP
`JP
`
`0 513 639
`0 706 206 A2
`59181062 A
`62_85461
`03079033 A
`448647
`4—68564
`4605922
`6—45432
`6-163843
`7-273330
`09162392 A
`
`11/1992
`4/1996
`10/1984
`4/1987
`4/1991
`2/1992
`3 1992
`“51992
`2/1994
`6/1994
`10/1995
`6 1997
`/
`
`* cited by examiner
`
`Page 2 0f 40
`
`Page 2 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 1 0f 21
`
`US 6,967,409 B2
`
`F1G.1(a)
`
`4 V[9%m
`
`50a
`
`FIG 1 b n q 2"
`
`FIG. 1(0)
`
`FIG. M)
`
`Page 3 0f 40
`
`Page 3 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 2 0f 21
`
`US 6,967,409 B2
`
`
`
`Page 4 0f 40
`
`Page 4 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 3 0f 21
`
`US 6,967,409 B2
`
`,
`
`FIG. 3<d> ‘
`
`
`
`
`
`V‘
`
`
`
`FIG. 3(6)
`
`..
`
`FIG. 3(f)
`
`’
`
`Page 5 0f 40
`
`Page 5 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 4 0f 21
`
`US 6,967,409 B2
`
`
`25b 5‘“‘
`
`
`
`
`Page 6 0f 40
`
`Page 6 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 5 0f 21
`
`US 6,967,409 B2
`
`FIG. 5(a) ‘
`
`FIG.5(C)
`
`FIG. 5(b)
`
`Page 7 0f 40
`
`Page 7 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 6 0f 21
`
`US 6,967,409 B2
`
`“nim-
`“n‘ 7 """'£‘
`
`
`
`
`133
`
`
`
`"\V
`
`
`12
`
`at“, \
`,.““"IVIJI. “
`
`
`
`Page 8 0f 40
`
`Page 8 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 7 0f 21
`
`US 6,967,409 B2
`
`FIG.7(a)
`
`FIG.7(C)
`
`FIG. 7(b)
`
`Page 9 0f 40
`
`Page 9 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 8 0f 21
`
`US 6,967,409 B2
`
`FIG.8(a)
`
`FIG.8(b)
`
`F1G.8(c)
`
`
`
`
`
`
`
`
`23W
`a
`u
`
`
`
`
`
`Page 10 0f 40
`
`Page 10 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 9 0f 21
`
`US 6,967,409 B2
`
`FIG.9(a)
`
`
`
`23
`
`‘N 2”“.
`i
`ugh“;- ,.
`
`
`
`
`
`-'«gin
`
`
`
`Page 11 0f 40
`
`Page 11 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 10 0f 21
`
`US 6,967,409 B2
`
`
`
`
`
`Vl/I/I/I/I/IA
`
`.’ ‘—“"i‘...'
`‘
`
`
`Page 12 0f 40
`
`Page 12 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 11 0f 21
`
`US 6,967,409 B2
`
`
`
`
`
`
`WWI/IA
`
`‘
`
`‘
`
`‘
`
`
`'1 “Ln-fa V
`
`'
`
`w
`
`Page 13 0f 40
`
`Page 13 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 12 0f 21
`
`US 6,967,409 B2
`
`23
`
`90
`
`L‘
`
`
`
`Vflfi'ffl‘
`
`‘s'.v.v.n.u
`
`‘2
`'Cr'nnbll'l W0)...rl
`
`
`
`
`
`Page 14 0f 40
`
`Page 14 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 13 0f 21
`
`US 6,967,409 B2
`
`FIG.13(b)
`
`Ill!!!!|llllllillllllll
`
`
`m \\\\\\‘
`
`
`
`\\w
`
`
`
`
`
`
`FIG.13(C)
`
`N “\v
`
`
`
`
`
`50b
`
`
`
`
`
`m
`
`
`
` h
`
`‘
`WIIIII/fi
`'A'IIIIII
`
`
`Page 15 0f 40
`
`Page 15 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 14 0f 21
`
`US 6,967,409 B2
`
`'
`
`.u ‘-
`
`
`
`- _""""""I
`
`N \\\\\\“
`
`
`2b
`
`m P
`
`age 16 0f 40
`
`Page 16 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 15 0f 21
`
`US 6,967,409 B2
`
`
`
`Page 17 0f 40
`
`Page 17 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 16 0f 21
`
`US 6,967,409 B2
`
`
`
`\““‘“
`
`Page 18 0f 40
`
`Page 18 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 17 0f 21
`
`US 6,967,409 B2
`
`FIG. 17
`PRIOR ART
`
`
`3:4!‘s
`“$42!le 12
`
`
`Page 19 0f 40
`
`Page 19 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 18 0f 21
`
`US 6,967,409 B2
`
`
`ffifiN
`fl?‘\
`
`
`m m _'
`
`
`
`‘2
`
`’l
`
`DIAMETER
`
`OF HOLE
`
`END OF ISOLASION
`
`FIG. 18 (b)
`PRIOR ART
`
`PRIOR ART
`
`FIG.18(C)
`
`Page 20 0f 40
`
`Page 20 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 19 0f 21
`
`US 6,967,409 B2
`
`FIG. 19
`PRIOR ART
`
`Q.‘I."‘
`
`2
`
`Page 21 0f 40
`
`Page 21 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 20 0f 21
`
`US 6,967,409 B2
`
`PRIOR ART
`
`FIG.20(a)
`
`
`-ACRCYO'Q'A"
`"‘VAV-VAVA"..C;.L.L..’...'.'.Q57".
`FIG. 20(b) m s\\\\‘
`\
`
`PRIOR ART
`
`
`
`FIG. 20(0):“.
`
`PRIOR ART
`
`a
`
`
`
`F I G_
`
`(d)
`
`PRIOR ART
`
`\““{Illllfl w’llll‘l.
`
`
`
`
`FIG.20(e)1
`
`
`kV‘VAVA-VAV.‘
`D.'..I/l
`
`
`PRIOR ART
`
`Page 22 0f 40
`
`Page 22 of 40
`
`

`

`US. Patent
`
`Nov. 22, 2005
`
`Sheet 21 0f 21
`
`US 6,967,409 B2
`
`FIG. 21(3)
`PRIOR ART
`
`IMPURITY ION
`
`
`
`
`
`10701080 / / /
`VIA
`
`QY/l/l/l/l/l/ll”I
`
`
`1050
`
`115
`
`101
`
`FIG. 21(b)
`PRIOR ART
`
`113
`
`
`
`Page 23 0f 40
`
`Page 23 of 40
`
`

`

`US 6,967,409 B2
`
`1
`SEMICONDUCTOR DEVICE AND METHOD
`OF MANUFACTURING THE SAME
`
`The Application is a Divisional Application of applica-
`tion Ser. No. 09/902,157 filed on Jul. 11, 2001 now US. Pat.
`No. 6,709,950, which is a Divisional Application of appli-
`cation Ser. No. 08/685,726 filed on Jul. 24, 1996, which is
`now US. Pat. No. 6,281,562.
`
`BACKGROUND OF THE INVENTION
`
`The present invention relates to a semiconductor device
`including transistors and connection between the transistors
`for constituting an LSI with high integration and a decreased
`area.
`
`With the recent development of a semiconductor device
`with high integration and high performance,
`there are
`increasing demands for more refinement of the semiconduc-
`tor device. The improvement of the conventional techniques
`cannot follow these demands, and novel
`techniques are
`unavoidably introduced in some technical
`fields. For
`example, as a method of forming an isolation, the LOCOS
`isolation method is conventionally adopted in view of its
`simpleness and low cost. Recently, however, it is considered
`that a trench buried type isolation (hereinafter referred to as
`the trench isolation) is more advantageous for manufactur-
`ing a refined semiconductor device.
`Specifically, in the LOCOS isolation method, since selec-
`tive oxidation is conducted, the so-called bird’s beak occurs
`in the boundary with a mask for preventing the oxidation. As
`a result, the dimension of a transistor is changed because an
`insulating film of the isolation invades a transistor region
`against the actually designed mask dimension. This dimen-
`sional change is unallowable in the refinement of a semi-
`conductor device after the 0.5 pm generation. Therefore,
`even in the mass-production techniques, the isolation form-
`ing method has started to be changed to the trench isolation
`method in which the dimensional change is very small. For
`example, IBM corporation has introduced the trench isola-
`tion structure as a 0.5 pm CMOS process for the mass-
`production of an MPU (IBM Journal of Research and
`Development, VOL. 39, No. 1/2, 1995, pp. 33—42).
`Furthermore, in a semiconductor device mounting ele-
`ments such as a MOSFET in an active area surrounded with
`
`an isolation, an insulating film is deposited on the active
`area, the isolation and a gate electrode, and a contact hole is
`formed by partly exposing the insulating film for connection
`between the active area and an interconnection member on
`
`a layer above the insulating film. This structure is known as
`a very common structure for the semiconductor device.
`FIG. 17 is a sectional view for showing the structure of a
`conventional semiconductor device. In FIG. 17, a reference
`numeral 1 denotes a silicon substrate, a reference numeral 2b
`denotes an isolation with a trench isolation structure which
`
`is made of a silicon oxide film and whose top surface is
`flattened so as to be at the same level as the top surface of
`the silicon substrate 1, a reference numeral 3 denotes a gate
`oxide film made of a silicon oxide film, a reference numeral
`4a denotes a polysilicon electrode working as a gate
`electrode, a reference numeral 4b denotes a polysilicon
`interconnection formed simultaneously with the polysilicon
`electrode 4a, a reference numeral 6 denotes a low-
`concentration source/drain region formed by doping the
`silicon substrate with an n-type impurity at a low
`concentration, a reference numeral 7a denotes an electrode
`sidewall, a reference numeral 7b denotes an interconnection
`sidewall, a reference numeral 8 denotes a high-concentration
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`2
`source/drain region formed by doping the silicon substrate
`with an n-type impurity at a high concentration, a reference
`numeral 12 denotes an insulating film made of a silicon
`oxide film, and a reference numeral 13 denotes a local
`interconnection made of a polysilicon film formed on the
`insulating film 12.
`The local interconnection 13 is also filled within a con-
`
`nection hole 14 formed in a part of the insulating film 12, so
`as to be contacted with the source/drain region in the active
`area through the connection hole 14.
`In this case,
`the
`connection hole 14 is formed apart from the isolation 2b by
`a predetermined distance. In other words, in the conven-
`tional layout rule for such a semiconductor device, there is
`a rule that the edge of a connection hole is previously located
`away from the boundary between the active area and the
`isolation region so as to prevent a part of the connection hole
`14 from stretching over the isolation 2b even when a mask
`alignment shift is caused in photolithography (this distance
`between the connection hole and the isolation is designated
`as an alignment margin).
`However, in the structure of the semiconductor device as
`shown in FIG. 17, there arise problems in the attempts to
`further improve the integration for the following reason:
`A distance La between the polysilicon electrode 4a and
`the isolation 2b is estimated as an index of the integration.
`In order to prevent the connection hole 14 from interfering
`the isolation 2b as described above,
`the distance La is
`required to be 1.2 pm, namely, the sum of the diameter of the
`connection hole 14, that is, 0.5 pm, the width of the electrode
`sidewall 7a, that is, 0.1 pm, the alignment margin from the
`polysilicon electrode 4a, that is, 0.3 um, and the alignment
`margin from the isolation 2b, that is, 0.3 pm. A connection
`hole has attained a more and more refined diameter with the
`
`development of processing techniques, and also a gate
`length has been decreased as small as 0.3 pm or less. Still,
`the alignment margin in consideration of the mask alignment
`shift in the photolithography is required to be approximately
`0.3 pm. Accordingly, as the gate length and the connection
`hole diameter are more refined, the proportion of the align-
`ment margin is increased. This alignment margin has
`become an obstacle to the high integration.
`Therefore, attempts have been made to form the connec-
`tion hole 14 without considering the alignment margin in
`view of the alignment shift in the photolithography. Manu-
`facturing procedures adopted in such a case will now be
`described by exemplifying an n-channel MOSFET referring
`to FIGS. 18(a) through 18(C).
`First, as is shown in FIG. 18(a), after forming an isolation
`2b having the trench structure in a silicon substrate 1 doped
`with a p-type impurity (or p-type well), etch back or the like
`is conducted for flattening so as to place the surfaces of the
`isolation 2b and the silicon substrate 1 at the same level. In
`
`an active area surrounded with the isolation 2b, a gate oxide
`film 3, a polysilicon electrode 4a serving as a gate electrode,
`an electrode sidewall 7a, a low-concentration source/drain
`region 6 and a high-concentration source/drain region 8 are
`formed. On the isolation 2b are disposed a polysilicon
`interconnection 4b formed simultaneously with the polysili-
`con electrode 4a and an interconnection sidewall 7b. At this
`
`point, the top surface of the high-concentration source/drain
`region 8 in the active area is placed at the same level as the
`top surface of the isolation 2b. Then, an insulating film 12
`of a silicon oxide film is formed on the entire top surface of
`the substrate.
`
`Next, as is shown in FIG. 18(b), a resist film 25a used as
`a mask for forming a connection hole is formed on the
`
`Page 24 of 40
`
`Page 24 of 40
`
`

`

`US 6,967,409 B2
`
`3
`insulating film 12, and the connection hole 14 is formed by,
`for example, dry etching.
`Then, as is shown in FIG. 18(c), the resist film 25a is
`removed, and a polysilicon film is deposited on the insulat-
`ing film 12 and within the connection hole 14. The poly-
`silicon film is then made into a desired pattern, thereby
`forming a local interconnection 13.
`At this point, in the case where the alignment margin in
`view of the mask alignment shift in the formation of the
`connection hole 14 is not considered in estimating the
`distance La between the polysilicon electrode 4a and the
`isolation 2b, a part of the isolation 2b is included in the
`connection hole 14 when the exposing area of the resist film
`25a is shifted toward the isolation 2b due to the mask
`
`alignment shift in the photolithography. Through over-etch
`in conducting the dry etching of the insulating film 12,
`although the high-concentration source/drain region 8 made
`of the silicon substrate is not largely etched because of its
`small etching rate, the part of the isolation 2b included in the
`connection hole 14 is selectively removed, resulting in
`forming a recess 40 in part of the connection hole 14. When
`the recess 40 in the connection hole 14 has a depth exceed-
`ing a given proportion to the depth of the high-concentration
`source/drain region 8, junction voltage resistance can be
`decreased and a junction leakage current can be increased
`because the concentration of the impurity in the high-
`concentration source/drain region 8 is low at that depth.
`In order to prevent these phenomena, it is necessary to
`provide a predetermined alignment margin as is shown in the
`structure of FIG. 17 so as to prevent the connection hole 14
`from interfering the isolation 2b even when the alignment
`shift is caused in the lithography. In this manner, in the
`conventional layout rule for a semiconductor device, an
`alignment margin in view of the mask alignment shift in the
`photolithography is unavoidably provided.
`Furthermore, a distance between the polysilicon electrode
`4a and the connection hole 14 is also required to be provided
`with an alignment margin. Otherwise, the connection hole
`14 can interfere the polysilicon electrode 4a due to the
`fluctuation caused in the manufacturing procedures, result-
`ing in causing electric short-circuit between an upper layer
`interconnection buried in the connection hole and the gate
`electrode.
`
`As described above, it is necessary to provide the con-
`nection hole 14 with margins for preventing the interference
`with other elements around the connection hole, which has
`become a large obstacle to the high integration of an LSI.
`Also in the case where a semiconductor device having the
`so-called salicide structure is manufactured, the following
`problems are caused due to a recess formed in the isolation:
`FIG. 19 is a sectional view for showing an example of a
`semiconductor device including the conventional
`trench
`isolation and a MOSFET having the salicide structure. As is
`shown in FIG. 19, a trench isolation 105a is formed in a
`silicon substrate 101. In an active area surrounded with the
`
`4
`gate interconnection 107b is provided with interconnection
`sidewalls 108b on its both side surfaces. On the gate
`electrode 107a, the gate interconnection 107b and the high-
`concentration source/drain region 106b, an upper gate elec-
`trode 109a, an upper gate interconnection 109b and a
`source/drain electrode 109C each made of silicide are respec-
`tively formed. Furthermore,
`this semiconductor device
`includes an interlayer insulating film 111 made of a silicon
`oxide film, a metallic interconnection 112 formed on the
`interlayer insulating film 111, and a contact member 113
`(buried conductive layer) filled in a connection hole formed
`in the interlayer insulating film 111 for connecting the
`metallic interconnection 112 with the source/drain electrode
`109C.
`
`Now, the manufacturing procedures for the semiconduc-
`tor device including the conventional trench isolation and
`the MOSFET with the salicide structure shown in FIG. 19
`
`will be described referring to FIGS. 20(a) through 20(e).
`First, as is shown in FIG. 20(a), a silicon oxide film 116
`and a silicon nitride film 117 are successively deposited on
`a silicon substrate 101, and a resist film 120 for exposing an
`isolation region and masking a transistor region is formed on
`the silicon nitride film 117. Then, by using the resist film 120
`as a mask, etching is conducted, so as to selectively remove
`the silicon nitride film 116 and the silicon oxide film 117,
`and further etch the silicon substrate 101, thereby forming a
`trench 104. Then, impurity ions are injected into the bottom
`of the trench 104, thereby forming a channel stop region 115.
`Then, as is shown in FIG. 20(b), a silicon oxide film (not
`shown) is deposited, and the entire top surface is flattened
`until the surface of the silicon nitride film 117 is exposed.
`Through this procedure, a trench isolation 105a made of the
`silicon oxide film filled in the trench 104 is formed in the
`
`isolation region Reiso.
`Next, as is shown in FIG. 20(c), after the silicon nitride
`film 117 and the silicon oxide film 116 are removed, a gate
`oxide film 103 is formed on the silicon substrate 101, and a
`polysilicon film 107 is deposited thereon. Then, a photore-
`sist film 121 for exposing areas excluding a region for
`forming a gate is formed on the polysilicon film 107.
`Then, as is shown in FIG. 20(a), by using the photoresist
`film 121 as a mask, dry etching is conducted,
`thereby
`selectively removing the polysilicon film 107 and the gate
`oxide film 103. Thus, a gate electrode 107a of the MOSFET
`in the transistor region Refet and a gate interconnection
`107b stretching over the isolation 105a and the silicon
`substrate 101 are formed. After removing the photoresist
`film 121, impurity ions are injected into the silicon substrate
`101 by using the gate electrode 107a as a mask, thereby
`forming a low-concentration source/drain region 106a.
`Then, a silicon oxide film 108 is deposited on the entire top
`surface of the substrate.
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`Next, as is shown in FIG. 20(6), the silicon oxide film 108
`is anisotropically dry-etched,
`thereby forming electrode
`sidewalls 108a and interconnection sidewalls 108b on both
`
`isolation 105a, a gate insulating film 103a, a gate electrode
`107a, and electrode sidewalls 108a on both side surfaces of
`the gate electrode 107a are formed. Also in the active area,
`a low-concentration source/drain region 106a and a high-
`concentration source/drain region 106b are formed on both
`sides of the gate electrode 107a. A channel stop region 115
`is formed below the isolation 105a. Furthermore, in areas of
`the silicon substrate 101 excluding the isolation 105a and
`the active area, a gate interconnection 107b made of the
`same polysilicon film as that for the gate electrode 107a is
`formed with a gate insulating film 103b sandwiched, and the
`
`60
`
`65
`
`side surfaces of the gate electrode 107a and the gate inter-
`connection 107b, respectively. At this point, the gate oxide
`film 103 below the silicon oxide film 108 is simultaneously
`removed, and the gate oxide film 103 below the gate
`electrode 107a alone remains. Then,
`impurity ions are
`diagonally injected by using the gate electrode 107a and the
`electrode sidewalls 108a as masks, thereby forming a high-
`concentration source/drain region 106b. Then, after a Ti film
`is deposited on the entire top surface, high temperature
`annealing is conducted, thereby causing a reaction between
`the Ti film and the components made of silicon directly in
`
`Page 25 of 40
`
`Page 25 of 40
`
`

`

`US 6,967,409 B2
`
`5
`contact with the Ti film. Thus, an upper gate electrode 109a,
`an upper gate interconnection 109b and a source/drain
`electrode 109C made of silicide are formed.
`
`The-procedures to be conducted thereafter are omitted,
`but the semiconductor device including the MOSFET hav-
`ing the structure as shown in FIG. 19 can be ultimately
`manufactured. In FIG. 19, the metallic interconnection 112
`is formed on the interlayer insulating film 111, and the
`metallic interconnection 112 is connected with the source/
`drain electrode 109C through the contact member 113
`including a W plug and the like filled in the contact hole.
`When the aforementioned trench isolation structure is
`
`adopted, the dimensional change of the source/drain region
`can be suppressed because the bird’s beak, that is, the oxide
`film invasion of an active area, which is caused in the
`LOCOS method where a thick silicon oxide film is formed
`
`by thermal oxidation, can be avoided. Furthermore, in the
`procedure shown in FIG. 20(C), the surfaces of the isolation
`105a and the silicon substrate 101 in the transistor region
`Refet are placed at the same level.
`In such a semiconductor device having the trench type
`isolation, however, there arise the following problems:
`When the procedures proceed from the state shown in
`FIG. 20(a) to the state shown in FIG. 20(6), the silicon oxide
`film 108 is anisotropically etched so as to form the sidewalls
`108a and 101%. At this point, over-etch is required. Through
`this over-etch, the surface of the isolation 105a is removed
`by some depth.
`FIGS. 21(a) and 21(b) are enlarged sectional views
`around the boundary between the high-concentration source/
`drain region 106b and the isolation 105a after this over-etch.
`As is shown in FIG. 21(a), between the procedures shown
`in FIGS. 20(a) and 20(6), the impurity ions are diagonally
`injected so as to form the high-concentration source/drain
`region 106b. Through this ion injection,
`the high-
`concentration source/drain region 106b is formed also below
`the edge of the isolation 105a because the isolation 105a is
`previously etched by some depth. Accordingly, the high-
`concentration source/drain region 106b is brought closer to
`the channel stop region 115, resulting in causing the prob-
`lems of degradation of the junction voltage resistance and
`increase of the junction leakage current.
`In addition, as is shown in FIG. 21(b), in the case where
`the Ti film or the like is deposited on the high-concentration
`source/drain region 106b so as to obtain the silicide layer
`through the reaction with the silicon below, the thus formed
`silicide layer can invade the interface between the silicon
`substrate 101 and the isolation 105a with ease. As a result,
`a short-circuit current can be caused between the source/
`drain electrode 109C made of silicide and the channel stop
`region 115.
`
`SUMMARY OF THE INVENTION
`
`invention is improving the
`The object of the present
`structure of an isolation, so as to prevent the problems
`caused because the edge of the isolation is trenched in
`etching for the formation of a connection hole or sidewalls.
`In order to achieve the object, the invention proposes first
`and second semiconductor devices and first through third
`methods of manufacturing a semiconductor device as
`described below.
`The first semiconductor device of this invention in which
`
`a semiconductor element is disposed in each of plural active
`areas in a semiconductor substrate comprises an isolation for
`surrounding and isolating each active area,
`the isolation
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`6
`having a top surface at a higher level than a surface of the
`active area and having a step portion in a boundary with the
`active area; an insulating film formed so as to stretch over
`each active area and the isolation; plural holes each formed
`by removing a portion of the insulating film disposed at least
`on the active area; plural buried conductive layers filled in
`the respective holes; and plural interconnection members
`formed on the insulating film so as to be connected with the
`respective active areas through the respective buried con-
`ductive layers.
`Owing to this structure, in the case where a part of or all
`the holes are formed so as to stretch over the active areas and
`the isolation due to mask alignment shift
`in
`photolithography, a part of the isolation is removed by
`over-etch for ensuring the formation of the holes. In such a
`case, even when the top surface of the isolation is trenched
`to be lower than the surface of the active area, the depth of
`the holes formed in the isolation is small in the boundary
`with the active area because of the level difference between
`the top surface of the isolation and the surface of the active
`area. Accordingly, degradation of the junction voltage resis-
`tance and increase of the junction leakage current can be
`suppressed. Therefore, there is no need to provide a portion
`of the active area where each hole is formed with an
`
`alignment margin for avoiding the interference with the
`isolation caused by the mask alignment shift in the lithog-
`raphy. Thus, the area of the active area can be decreased,
`resulting in improving the integration of the semiconductor
`device.
`
`In the first semiconductor device, at least a part of the
`plural holes can be formed so as to stretch over the active
`area and the isolation due to fluctuation in manufacturing
`procedures.
`In other words, even when no margin for the mask
`alignment
`in the lithography is provided,
`the problems
`caused in the formation of the holes can be avoided.
`
`Furthermore, the angle between a side surface of the step
`portion and the surface of the active area is preferably 70
`degrees or more.
`As a result, when the hole interferes the isolation, the part
`of the isolation included in the hole is definitely prevented
`from being etched through over-etch in the formation of the
`holes down to a depth where the impurity concentration is
`low in the active area.
`
`The isolation is preferably a trench isolation made of an
`insulating material filled in a trench formed by trenching the
`semiconductor substrate by a predetermined depth.
`This is because no bird’s beak is caused in the trench
`
`isolation differently from a LOCOS film as described above,
`and hence, the trench isolation is suitable particularly for the
`high integration and refinement of the semiconductor
`device.
`
`In the first semiconductor device, when the semiconduc-
`tor element is a MISFET including a gate insulating film and
`a gate electrode formed on the active area; and source/drain
`regions formed in the active area on both sides of the gate
`electrode,
`the following preferred embodiments can be
`adopted:
`The semiconductor device can further comprise a gate
`interconnection made of the same material as that for the
`
`gate electrode and formed on the isolation, each of the holes
`can be formed on an area including the source/drain region,
`the isolation and the gate interconnection, and the plural
`interconnection members can be connected with the gate
`interconnection on the isolation.
`
`Owing to this configuration, in the case where the inter-
`connection members work as local
`interconnections for
`
`Page 26 of 40
`
`Page 26 of 40
`
`

`

`US 6,967,409 B2
`
`7
`connecting a gate interconnection on the isolation with the
`active area, there is no need to separately form holes in the
`insulating film on the gate interconnection and the insulating
`film on the active area. In addition, there is no need to
`provide the separate holes with alignment margins from the
`boundary between the active area and the isolation.
`Accordingly, the area of the isolation can also be decreased,
`resulting in largely improving the integration of the semi-
`conductor device.
`
`The semiconductor device can further comprise electrode
`sidewalls made of an insulating material and formed on both
`side surfaces of the gate electrode; and a step sidewall made
`of the same material as the insulating material for the
`electrode sidewalls and formed on the side surface of the
`step portion. In this semiconductor device, at least a part of
`the holes can be formed by also removing a portion of the
`insulating film disposed on the step sidewall.
`Owing to this structure,
`the abrupt
`level difference
`between the surfaces of the isolation and the active area can
`
`be released by the step sidewall. Therefore, a residue is
`scarcely generated in patterning the interconnection
`members, and an upper interconnection is prevented from
`being disconnected and increasing in its resistance.
`The semiconductor device can further comprise a gate
`protection film formed on the gate electrode, and at least a
`part of the holes can be formed so as to stretch over the
`source/drain region and at least a part of the gate protection
`film.
`
`Owing to this structure, a part of the gate protection film
`included in the hole is removed by the over-etch in the
`formation of the holes. However,
`the gate electrode is
`protected by the gate protection film, and hence, electrical
`short circuit between the gate electrode and the intercon-
`nection member can be prevented. Accordingly, there is no
`need to provide an alignment margin from the gate electrode
`in the area where each hole is formed, resulting in further
`improving the integration.
`The interconnection members can be first layer metallic
`interconnections, and the insulating film can be an interlayer
`insulating film disposed between the semiconductor
`substrate, and the first layer metallic interconnections. In this
`case,
`the semiconductor device preferably further
`comprises, between the interlayer insulating film and the
`semiconductor substrate an underlying film made of an
`insulating material having high etching selectivity against
`the interlayer insulating film.
`The second semiconductor device of this invention in
`
`which a semiconductor element is disposed in each of plural
`active areas in a semiconductor substrate comprises a trench
`isolation for isolating and surrounding each active area, the
`trench isolation having a top surface at a higher level than a
`surface of the active area and having a step portion in a
`boundary with the active area; and a step sidewall formed on
`the side surface of the step portion of the trench isolation.
`Owing to this structure, in the impurity ion injection for
`the formation of an impurity diffused layer of the semicon-
`ductor device, the step sidewall disposed at the edg