`
`
`United States Patent ti
`
`
`Gasner
`
`[11] Patent Number:
`
`
`
`
`
`
`[45] Date of Patent:
`
`
`4,599,789
`
`
`Jul. 15, 1986
`
`
`
`
`4,453,306
`
`
`6/1984 Lynchet al. ....ceeeeees 29/571
`
`
`
`
`
`
`OTHER PUBLICATIONS
`
`
`
`
`
`
`
`“Fabrication of High-Performance LDDFET’s with
`
`
`
`
`
`Oxide Sidewall-Spacer Technology”; Tsang et al.;
`
`
`
`
`
`
`
`
`
`IEEE; vol. ED-29, No. 4, Apr. 1982, pp. 590-596.
`
`
`
`
`
`
`“An Optimally Designed Process for Submicrometer
`
`
`
`
`
`
`
`
`
`MOSFET?’s”; Shibata et al.; IEEE, vol. ED-29, No. 4,
`
`
`
`
`Apr. 1982, pp. 531-535.
`
`
`
`
`
`“Twin-Tub CMOS—A Technology for VLSI Cir-
`
`
`
`
`
`
`
`cuits”; Parrilio et al.; IEDM, 1980, pp. 752-755.
`“Twin-Tub CMOS II—An Advanced VLSI Technol-
`
`
`
`
`
`
`
`
`
`
`
`
`
`ogy”; Parrillo et al.; IEDM 1982, pp. 706-709.
`
`
`
`“Quadruple-Wall CMOS—A VLSI Technology”;
`
`
`
`
`
`Chen; IEDM 1982, pp. 791-792.
`
`
`
`Primary Examiner—Upendra Roy
`
`
`
`
`
`Attorney, Agent, or Firm—Barnes & Thornburg
`
`
`ABSTRACT
`[57]
`
`
`
`
`
`
`
`
`CMOSdevices are formed in self-aligned wells in a
`
`
`
`
`
`
`
`substrate produced by a two mask, one photolitho-
`
`
`
`
`
`
`
`
`
`
`graphic step process wherein the first mask is used as a
`
`
`
`
`
`
`
`
`template to form the second inverse mask of substan-
`
`
`
`
`
`
`
`
`tially equal thickness. The gates are used as alignment
`
`
`
`
`
`
`
`
`mask for shallow source and drain regions and subse-
`
`
`
`
`
`
`
`
`quently formed lateral gate spacers are used as align-
`
`
`
`
`
`
`
`
`
`ment mask for deep source and drain regions. Exposed
`
`
`
`
`
`
`
`
`
`source and drain regions and silicon gates havesilicide
`
`
`
`
`
`formed thereon by a non-selective process.
`
`
`
`
`
`
`
`
`
`
`
`
`[54]
`
`
`[75]
`
`
`[73]
`
`[21]
`
`[22]
`
`[51]
`
`[52]
`
`
`[58]
`
`
`[56]
`
`PROCESS OF MAKING TWIN WELLVLSI
`
`
`
`
`
`
`CMOS
`
`
`Inventor:
`
`
`
`
`
`
`
`John T. Gasner, Melbourne, Fla.
`
`
`
`
`
`
`
`
`
`Assignee: Harris Corporation, Melbourne, Fila.
`
`
`
`Appl. No.: 620,835
`
`
`
`
`Filed:
`Jun. 15, 1984
`
`
`
`
`
`
`
`Int. Ch4 oo. eeceeeeeeee HO1L 21/265; BO1J 17/00
`
`
`
`
`
`
`U.S. CL. cecccccsccsssecreecsseeteeseseenes 29/571; 29/576 B;
`
`
`
`
`
`148/1.5; 148/187; 156/643; 357/42; 357/91
`
`
`
`
`
`
`
`Field of Search................. 148/1.5, 187; 29/571,
`
`
`
`
`
`29/576 B; 156/643; 357/91, 42
`References Cited
`
`
`U.S. PATENT DOCUMENTS
`
`
`
`9/1973 Wang .ecssccssercrserssseseseerenens 148/188
`3,759,763
`
`
`
`
`
`...........
`4,209,350 6/1980 Ho etal.
`- 148/188
`
`
`
`
`
`
`
`
`7/1981 Egawa et al. occ 29/571
`4,280,272
`
`
`
`
`
`
`
`3/1982 Lund et al. ws
`esesesenseceenens 29/571
`4,319,395
`
`
`
`
`
`
`
`
`4,333,099
`6/1982 Tanguay et al. ores 357/67
`
`
`
`
`
`
`
`
`4,343,082
`8/1982 Lepselter et al.
`. 29/576 B
`.
`
`
`
`
`
`
`
`
`4,356,040 10/1982 Fuetal. oe
`wee 148/1,5
`
`
`
`
`
`
`
`4,356,623 11/1982 Hunter..............
`vee 29/571
`
`
`
`
`
`
`4,366,613
`1/1983 Ogura et al. w.sessssesseeeaees 29/571
`
`
`
`
`
`
`
`
`4,385,947
`5/1983 Halfacreet al. ......
`.. 148/187
`
`
`
`
`
`
`
`4,411,058 10/1983 Chen... eee
`wee 29/571
`
`
`
`
`
`4,420,344 12/1983 Davies...
`vee 148/1.5
`
`
`
`
`
`
`3/1984 Schwabeet al
`. 29/576 B
`4,434,543
`
`
`
`
`
`
`
`4,435,896
`3/1984 Parrillo et al...
`eeeceseeee 29/571
`
`
`
`
`
`
`
`
`4/1984 Haken ........
`w-- 29/571
`4,442,591
`
`
`
`
`
`
`4,444,605
`4/1984 Slawinski ose 148/187
`
`
`
`
`
`
`14 Claims, 20 Drawing Figures
`
`
`
`
`
`
`
`
`
`
`
`
`
`Page 1 of 11
`
`TSMC Exhibit 1037
`TSMC vy.IP Bridge
`IPR2016-01246
`
`Page 1 of 11
`
`TSMC Exhibit 1037
`TSMC v. IP Bridge
`IPR2016-01246
`
`
`
`
`
`
`
`
`
`
`
`
`
`f6
`
`
`
`
`
`
`
`
`
`
`
`
`
`4,599,789
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Page 2 of 11
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`eet20f6 4,599,789=
`
`
`
`
`
`
`
`
`
`
`
`
`Page 3 of 11
`
`
`
`
`
`- U.S. Patent
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Sheet30f6
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Jul. 15, 1986
`
`
`
`4,599,789
`
`
`
`
`
`
`
`Page 4 of 11
`
`Page 4 of 11
`
`
`
`
`
`U.S. Patent
`
`Jul. 15, 1986
`
`
`
`
`
`Sheet4of6
`
`
`
`
`
`4,599,789
`
`
`P+
`
`
`
`
`
`
`
`
`
`
`2EMPT(Pe= SF
`
`
`TTL
`
`
`
`
`
`
`
`
`
`
`
`
`
`Page 5 of 11
`
`Page 5 of 11
`
`
`
`U.S. Patent
`
`
`
`
`
`Jul15, 1986
`
`
`
`
`
`Sheet 5of6
`
`
`
`
`
`4,599,789
`
`
`
`
`
` *
`
`58
`
`we
`oe ~LTE fe
`
`.
`wre ete
`
`
`
`Page6 of 11
`
`Page 6 of 11
`
`
`
`
`U.S. Patent
`
`
`
`Jul. 15, 1986
`
`
`
`
`
`Sheet60f6
`
`
`
`
`4,599,789
`
`
`\(e a
`
`
`
`
`
`
`
`
`
`
`
`p
`
`. wt
`
`Page7 of 11
`
`Page 7 of 11
`
`
`
`1
`
`
`
`4,599,789
`
`
`
`PROCESS OF MAKING TWIN WELL VLSI CMOS
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`10
`
`
`
`
`
`
`
`25
`
`
`30
`
`
`
`
`40
`
`
`45
`
`
`BACKGROUND AND SUMMARYOF THE
`
`
`
`
`INVENTION
`
`
`
`
`
`
`
`The present invention relates generally to integrated
`
`
`
`
`
`
`
`circuits and method of manufacturing and morespecifi-
`
`
`
`
`
`
`cally to a method of manufacturingself-aligned junction
`
`
`
`
`
`
`isolated complementary insulated gate field effect tran-
`sistors.
`
`
`
`
`
`
`
`
`The industry is constantly working towardincreasing
`
`
`
`
`
`
`
`
`
`
`
`the device density on a wafer or chip. The number of
`
`
`
`
`
`
`
`
`
`
`
`
`devices that can be placed on a chip is limited by the
`size of the device and the electrical interaction. These
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`place restraints on the spacing between devices. Further
`
`
`
`
`
`
`
`
`limitation are in the processing steps dealing with pho-
`
`
`
`
`
`
`
`
`
`tolithography and the ability to form doped regions of
`
`
`
`
`
`
`
`the controllable size and impurity concentration. For
`
`
`
`
`
`
`
`junction isolated insulated gate field effect transistors,
`
`
`
`
`
`
`
`
`the design must minimize device latch-up resulting from
`
`
`
`
`
`
`
`
`parasitic bipolar devices. It is also desirable to reduce
`
`
`
`
`
`
`
`
`
`
`
`the capacitance of the device as well as the contact and
`
`
`
`
`
`
`
`interconnect system. Hot electron and the substrate
`
`
`
`
`
`
`
`
`current injection is another problem which must be
`
`
`
`
`
`
`
`
`
`
`faced for insulated gate field effect transistors as well as
`
`
`
`
`
`
`
`
`reducing the resistance of the contacts and junctions.
`
`
`
`
`
`
`
`
`Althoughall of these problems have been addressed
`
`
`
`
`
`
`
`
`
`singly by the prior art,
`the industry is continuously
`
`
`
`
`
`
`
`
`searching for complete process flows which address as
`
`
`
`
`
`
`
`
`
`
`
`many of these problemsas possible. Thus,it is an object
`
`
`
`
`
`
`
`
`
`
`of the present invention to address as many of the de-
`
`
`
`
`
`
`
`
`sign restraints as possible while increasing the density of
`
`
`
`
`
`the devices on a chip.
`
`
`
`
`
`
`
`
`Another object of the present invention is to provide
`
`
`
`
`
`
`
`
`a process flow with a minimum number of masking
`
`steps.
`
`
`
`
`
`
`
`
`Still another object of the present invention is to
`
`
`
`
`
`
`
`
`provide a unique process for forming a reverse image
`mask.
`
`
`
`
`
`
`
`
`
`
`Yet another object of the present invention is to mini-
`
`
`mize latch-up.
`
`
`
`
`
`
`
`
`
`A further object of the present inventionis to provide
`
`
`
`
`
`
`
`
`
`‘ a process which allowstailoring of the threshold of the
`devices.
`
`
`
`
`
`
`
`
`
`
`An even further object of the present invention is to
`
`
`
`
`
`
`
`inhibit hot electron and substrate current injection.
`
`
`
`
`
`
`
`
`
`
`A still even further object of the present invention is
`
`
`
`
`
`
`
`
`to provide a reduced resistance of junction in contacts.
`
`
`
`
`
`
`
`
`
`Another object of the present invention is to lower
`
`
`
`
`
`
`
`the overall capacitance of the integrated circuit.
`
`
`
`
`
`
`
`
`These and other objects of the invention are attained
`
`
`
`
`
`
`
`
`
`by forming a first mask on a substrate ofa first conduc-
`
`
`
`
`
`
`
`tivity type and introduction of second conductivity
`
`
`
`
`
`
`
`
`type impurities to form a second conductivity type well.
`
`
`
`
`
`
`
`
`
`
`This is followed by a second mask whichis the reversed
`
`
`
`
`
`
`
`
`
`image of the first mask and introduction ofa first con-
`
`
`
`
`
`
`
`
`ductivity type impurities to form a first conductivity
`
`
`
`
`
`
`
`
`
`
`type well. The second mask may be formed by other
`
`
`
`
`
`
`
`
`
`processes.It is preferred that the second mask edges be
`
`
`
`
`
`
`
`
`
`
`self-aligned to the first mask. A gate insulative layer and
`
`
`
`
`
`
`
`
`
`
`
`gate materials are formed on the two well areas. This is
`
`
`
`
`
`
`
`
`followed by forming shallow source and drain regions
`
`
`
`
`
`
`
`
`
`
`
`
`using the gate as a mask in each of the wells. Next,
`
`
`
`
`
`
`
`insulative spacers are formed extending laterally from
`
`
`
`
`
`
`
`
`
`
`the first and second gates over the source and drain
`
`
`
`
`
`
`
`
`regions and impurities are introduced to form deeper
`
`
`
`
`
`
`
`
`
`
`source and drain regions using the gate and the spacers
`
`
`
`
`
`
`
`
`
`
`
`as a mask. This is followed by the forming of the
`
`
`
`50
`
`
`
`
`60
`
`
`65
`
`
`Page 8 of 11
`
`
`2
`
`
`
`
`
`
`
`
`
`contacts to the respective layers. The gate material is
`
`
`
`
`
`
`polycrystalline silicon and the impurity introduction
`
`
`
`
`
`
`
`
`
`steps are carried out by ion implantation. A metal sili-
`
`
`
`
`
`
`
`
`
`cide step is performed to reduce the contact resistance
`
`
`
`
`
`
`
`
`
`by forming metal silicide over the source and drain
`
`
`
`
`
`
`
`
`
`
`regions as well as the gate. Insulative inserts are formed
`
`
`
`
`
`
`
`
`
`between the device regions by etching and filling by
`
`
`
`
`
`
`
`
`deposition to form planar oxide inserts separating the
`
`
`
`device regionslaterally.
`
`
`
`
`
`
`
`A method offormingself-aligned well regions includ-
`
`
`
`
`
`
`
`
`
`ing the inverse image masking step includes forming the
`
`
`
`
`
`
`
`
`first mask followed by introducing impurities to form
`
`
`
`
`
`
`
`
`
`
`the first well region. The inverse mask is formed by
`
`
`
`
`
`
`
`
`applying a second masklayer overfilling the openingsin
`
`
`
`
`
`
`
`
`
`
`the first mask and covering the first mask. The second
`
`
`
`
`
`
`
`
`
`
`mask layer is removed sufficiently to expose at least a
`
`
`
`
`
`
`
`
`
`
`portion of the first mask layer. The exposed first mask
`
`
`
`
`
`
`
`
`
`layer and any first mask layer superimposed thereonis
`
`
`
`
`
`
`
`
`
`selectively removed to iorm the second mask having the
`
`
`
`
`
`
`
`
`
`
`reversed image of the first mask. This is followed by
`
`
`
`
`
`
`
`introducing the impurities of the opposite conductivity
`
`
`
`
`
`
`type to form a second well.
`
`
`
`
`
`
`
`Other objects, advantages and novel features of the
`
`
`
`
`
`
`
`present invention will become apparent from the fol-
`
`
`
`
`
`
`
`lowing detailed description of the invention when con-
`
`
`
`
`
`
`sidered in conjunction with the accompanying draw-
`ings.
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`
`
`
`
`
`
`
`
`
`
`
`FIGS. 1-3 illustrate a portion of the process incorpo-
`
`
`
`
`
`
`
`rating the principles of the present invention.
`
`
`
`
`
`
`
`
`FIGS. 4-6 illustrate a first method of forming an
`
`
`
`
`
`
`
`inverse mask incorporating the principles of the present
`invention.
`
`
`
`
`
`
`
`
`
`
`FIGS.7 and8 illustrate a second method of forming
`
`
`
`
`
`
`
`
`
`an inverse mask according to the principles of the pres-
`ent invention.
`
`
`
`
`
`
`
`
`
`FIGS. 9-16 illustrate the remainder of the process
`
`
`
`
`
`
`
`incorporating the principles of the present invention.
`
`
`
`
`
`
`
`FIGS. 17-20 illustrate of a modification of the pro-
`cess of FIGS. 10-13.
`
`
`
`
`DETAILED DESCRIPTION OF TH
`
`
`
`DRAWINGS
`.
`
`
`
`
`
`
`
`
`
`The process of fabrication begins as illustrated in
`
`
`
`
`
`
`
`
`
`FIG. 1 with a substrate 20 which for purposes of exam-
`
`
`
`
`
`
`
`
`
`ple will be an N-conductivity type silicon having an
`
`
`
`
`
`
`
`
`impurity concentration in the range of 1015 atoms per
`
`
`
`
`
`
`
`
`
`
`cubic centimeter. The substrate 20 may be an original
`
`
`
`
`
`
`
`
`
`
`substrate or may be an epitaxial layer formed on another
`
`
`
`
`
`
`
`
`substrate layer. The substrate is processed to form lat-
`
`
`
`
`
`
`
`
`eral dielectric isolation regions. This results in mesa
`areas in which thefield effect transistors will be formed.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`An insulative layer, for example silicon dioxide 22, is
`
`
`
`
`
`
`
`
`
`formed on the silicon substrate 20 and a photoresist
`
`
`
`
`
`
`
`
`
`
`
`layer 24 is formed on the oxide layer 22. The photoresist
`
`
`
`
`
`
`
`
`
`layer is then patterned by the well-known technique to
`
`
`
`
`
`
`
`
`
`
`form openings and act as a mask. The photoresist layer
`
`
`
`
`
`
`
`
`
`
`
`24, the oxide layer 22 and the substrate 20 are then
`
`
`
`
`
`
`
`
`
`
`etched to form trenches 26. The etchant may be a wet
`
`
`
`
`
`
`
`
`
`
`
`etch or a dry etch. Whenusing the wet etch, the angle
`
`
`
`
`
`
`
`
`
`the walls form with respect
`to the horizontal
`that
`
`
`
`
`
`
`
`
`
`
`should be between 60° and 90°. The steeper walls are
`
`
`
`
`
`
`
`
`preferred since it creates devices with effectively wider
`channels.
`
`
`
`
`
`
`
`
`
`
`The photoresist layer 24 and oxide layer 22 are re-
`
`
`
`
`
`
`
`moved immediately after etching the trenches. Then the
`
`
`
`
`
`
`
`
`
`
`
`
`Page 8 of 11
`
`
`
`
`3
`
`
`
`
`
`
`
`
`
`
`entire surface is thermally oxidized to form a thin oxide
`
`
`
`
`
`
`
`
`
`region 28. An insulative material, for example oxide,is
`
`
`
`
`
`
`
`
`then chemically vapor deposited to back fill
`the
`
`
`
`
`
`
`
`
`trenches 26. The excess chemical vapor deposited oxide
`
`
`
`
`
`
`
`
`
`are then removedin a planarization process to provide
`
`
`
`
`
`
`
`
`
`
`
`inserts 30 planar with the top of the substrate 20. The
`
`
`
`
`
`
`
`
`inserts 30 have, for example, a depth of 5000 Angstroms
`
`from the surface. The substrate 20 is then oxidized to
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`provide a thin oxide coating 32. The resulting structure
`
`
`
`
`
`
`
`
`
`
`
`
`is illustrated in FIG. 2. The inserts 30 may be formed by
`
`
`
`
`
`
`
`
`
`other processes and it
`is preferred that an insulator
`
`
`
`
`
`
`
`
`insert be used to provide lateral dielectric isolation.
`
`
`
`
`
`
`
`
`The next sequence of steps produceself-aligned twin
`wells in which the field effect transistor devices are to
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`be formed. This process involves using a first mask and
`
`
`
`
`
`
`
`
`
`
`then forming a second mask whichis the inverse image
`
`
`
`
`
`
`
`
`
`
`of a first mask. A first masking material 34, for example
`
`
`
`
`
`
`
`
`
`
`a photoresist or a polyimide, is formed on the surface of
`
`
`
`
`
`
`
`
`
`
`
`
`the substrate 20. An opening 36 is formed in the mask 34
`
`
`
`
`
`
`
`
`
`
`
`to expose the areas in which the P-wells are to be
`
`
`
`
`
`formed. A P-type impurity, for example boron, is ion
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`implantedat sufficient energy level and dose to form the
`
`
`
`
`
`
`
`
`
`P— well 38 illustrated in FIG. 3 having an impurity
`
`
`
`
`
`concentration greater or equal to 10!6 atoms per cubic
`
`
`
`
`
`
`
`
`
`
`
`centimeter and a depth of, for example, 15,000 Ang-
`
`
`
`
`
`
`
`
`
`
`stroms in mesa areas and 5000 Angstromsin oxidefilled
`
`
`
`
`
`
`
`
`
`areas. The photoresist layer 34 generally has a thickness
`
`
`
`
`
`
`
`
`in the range of 10,000 to 20,000 Angstroms. Additional
`
`
`
`
`
`
`
`
`
`
`
`boron implants can optionally be done atthis time, using
`
`
`
`
`
`
`
`
`different doses and energies, to create a tailored impu-
`
`
`
`
`
`
`
`rity doping concentration for the P— for device re-
`
`quirements.
`
`
`
`
`
`
`
`
`The formation of the second mask includes applying
`
`
`
`
`
`
`
`
`
`
`
`a metal layer 40 whichfills the previous opening 36 in
`
`
`
`
`
`
`
`
`
`
`
`the first mask layer 34. The metal layer 40 is then plana-
`
`
`
`
`
`
`
`
`rized by applying a planarization layer 42 which may
`
`
`
`
`
`
`
`
`be, for example a photoresist or polyimide layer. By
`
`
`
`
`
`
`
`
`
`
`wayof example, the metal layer 40 which may be alumi-
`
`
`
`
`
`
`
`
`
`num, is formed to have a thickness of 10,000 to 20,000
`
`
`
`
`
`
`
`
`
`
`
`
`Angstroms on the surface of the first masking layer 34
`
`
`
`
`
`
`
`
`
`and the planarization layer 42 has a thickness above the
`
`
`
`
`
`
`
`
`
`
`
`planar part of the metal layer in the range of 10,000 to
`
`
`
`
`
`
`
`
`20,000 Angstroms. The resulting structureis illustrated
`in FIG. 4. The planarization process includes etching
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`the planarization layer 42 and the metal layer 40 at
`
`
`
`
`
`
`
`
`
`appropriate rates such that the resulting metal layer 40
`
`
`
`
`
`
`
`
`
`is planar with the top ofthe first mask layer 34 as illus-
`
`
`
`
`
`
`
`
`
`
`
`trated in FIG. 5. This may be accomplished by reactive
`
`
`
`
`
`
`
`
`
`ion etching using the appropriate gas and energylevel
`
`
`
`
`to accomplish planarization. Planarization techniques
`
`
`
`
`
`
`
`
`
`
`
`are well knownin the prior art and thus will not be
`
`
`
`
`
`
`
`
`
`
`described herein in detail. Thefirst mask layer 34 is then
`
`
`
`
`
`
`
`
`selectively removed by an appropriate wet or dry etch-
`
`
`
`
`
`
`
`
`
`
`
`
`ant. This results in a mask layer 40 of metal whichis the
`
`
`
`
`
`
`
`
`
`inverse imageofthe original mask layer 34 asillustrated
`in FIG.6.
`
`
`
`
`
`
`
`
`
`
`An alternative method of forming the second mask
`
`
`
`
`
`
`
`
`
`
`layer which is the inverse of the first mask layer in-
`
`
`
`
`
`
`
`
`
`
`cludes forming the second mask layer or metal layer 40
`
`
`
`
`
`
`
`
`
`
`
`
`to a sufficient depth to fill the opening 36 in the first
`
`
`
`
`
`
`
`
`
`
`
`mask layer 34 as illustrated in FIG. 7. For a first mask
`
`
`
`
`
`
`
`layer 34 having a thickness of 10,000 to 20,000 Ang-
`
`
`
`
`
`
`
`
`
`
`
`stroms,
`the second masking layer 40 would have a
`
`
`
`
`
`
`
`
`
`thickness on the first mask layer 34 of 10,000 to 20,000
`
`
`
`
`
`
`
`
`
`
`
`Angstroms. The metal layer 40 is then etched using
`
`
`
`
`
`
`
`
`either wet or dry etch until a portion of the first mask
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`layer 34 is exposed. Since only a single layer is being
`etched, the etch rate is uniform and the corners of the
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`20
`
`
`
`25
`
`
`
`
`
`
`40
`
`
`45
`
`
`50
`
`
`55
`
`
`60
`
`
`65
`
`
`Page 9 of 11
`
`4,599,789
`
`
`
`
`
`
`
`
`
`
`
`
`
`4
`
`
`
`
`
`
`
`
`opening 36 will become exposed because ofthe topol-
`
`
`
`
`
`
`
`
`
`
`
`
`ogy of the metal layer 40. Asillustrated in FIG. 8, edge
`
`
`
`
`
`
`
`
`
`
`44 of mask layer 34 is exposed and partially etched. A
`
`
`
`
`
`
`
`
`
`
`
`first mask layer 34 is then selectively removed using a
`
`
`
`
`
`
`
`
`
`
`wetetchant which not only removesthefirst mask layer
`
`
`
`
`
`
`
`
`
`
`
`34, but also removes the portions of the metal layer 40
`
`
`
`
`
`
`
`which are superimposed thereon. Thislift-off technique
`
`
`
`
`
`
`
`
`
`
`results in the same mask layer 40 of FIG. 6 having in-
`
`
`
`
`
`
`
`
`
`
`
`verse image of the mask layer 34 of FIG.3.
`
`
`
`
`
`
`Since the alternative inverse mask forming process
`
`
`
`
`
`
`
`
`
`does not attempt to form a planar surface before etch-
`
`
`
`
`
`
`
`ings, the deposited second layer may be substantially
`
`
`
`
`
`
`
`
`
`thinner with a pronounced indenture. This reduces the
`
`
`
`
`
`
`
`
`
`
`time and cost of applying and etching the second mask
`
`layer.
`
`
`
`
`Although two techniques are described for forming
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`an inverse mask, other methods may be used to form the
`
`
`
`
`
`
`
`
`
`
`second mask. As will be described below, the edges of
`
`
`
`
`
`
`
`
`
`the mask openings should be aligned so as to form self-
`
`
`
`
`
`
`
`
`
`
`aligned wells with little if any overlap. Any dual photo-
`
`
`
`
`
`lithic techniques requires lateral spacing because of the
`
`
`
`
`
`
`
`
`
`
`
`tolerances of the lithographic processes to assure no
`
`
`
`
`
`
`
`
`overlap of the formed regions. Similarly, prior single
`
`
`
`
`
`
`lithographic techniques could not truly form inverse
`
`
`
`
`
`
`
`
`
`
`
`mask, and, thus, lateral spacing or guard rings were also
`
`required.
`
`
`
`
`N-type impurities, for example phosphorous, are then
`
`
`
`
`
`
`
`
`
`
`
`
`introduced for example, by ion implantation to form the
`
`
`
`
`
`
`
`
`
`N— well 46. The N— well which is aligned with the
`
`
`
`
`
`
`
`previously formed P-well 38 has an impurity concentra-
`
`
`
`
`
`
`
`tion of equal to or greater than 1016 atoms per cubic
`
`
`
`
`
`
`
`
`
`
`centimeter and a depth of 15,000 Angstroms. The sec-
`
`
`
`
`
`
`
`
`
`
`
`ond masklayer 40 is removed and the substrate is an-
`
`
`
`
`
`
`
`
`nealed. The resulting structureis illustrated in FIG.9.
`
`
`
`
`
`
`
`
`
`
`
`Theself-aligned dual well 38 and 46 remove the need
`
`
`
`
`
`
`
`for separate guard rings. Thus, removing additional
`
`
`
`
`
`
`
`
`
`
`
`
`steps to form the guard rings. It should be noted that by
`
`
`
`
`
`
`
`
`
`
`using ion implantation to form the wells, very little side
`
`
`
`
`
`
`
`
`diffusion occurs and therefore the spacing between the
`
`
`
`
`
`
`
`
`devices may be reduced. Similarly, very high peak dop-
`
`
`
`
`
`
`
`
`ing results which reduces the latch-up problem by form-
`
`
`
`
`
`
`
`ing high threshold parasitic field effect transistors. The
`
`
`
`
`
`
`
`resulting well’s surface doping also results in relatively
`
`
`
`low junction capacitance.
`
`
`
`
`
`
`Complementary insulated gate field effect transistors
`
`
`
`
`
`
`
`
`
`
`are formed by forming a gate oxide layer 48 on the
`
`
`
`
`
`
`
`
`
`
`surface of the substrate by, for example, exposing the
`
`
`
`
`
`
`
`
`substrate to an oxidizing atmosphere followed by the
`
`
`
`
`
`
`
`
`application of a gate material and delineation to form
`
`
`
`
`
`
`
`
`
`gate regions 50 and 52. In the preferred embodiment,
`
`
`
`
`
`
`
`
`
`the gates 50 and 52 are made from polycrystalline sili-
`
`
`
`
`
`
`
`
`
`
`
`con. A first type P impurity, for example boron,is ion
`
`
`
`
`
`
`
`
`
`implanted non-selectively using the gates 50 and 52 and
`
`
`
`
`
`
`
`
`
`
`the oxide inserts 30 as a mask to produce shallow re-
`
`
`
`
`
`
`
`
`
`
`
`
`
`gions 56 in well 38 and shallow regions 54 in well 46 as
`illustrated in FIG. 10.
`
`
`
`
`
`
`
`
`
`
`
`
`
`A mask is formed by applying photoresist layer 58
`
`
`
`
`
`
`
`
`
`and delineating to mask the N— well region 46 and
`
`
`
`
`
`
`
`
`expose the P~— well region 38. N-type impurities, for
`
`
`
`
`
`
`
`
`
`example arsenic, are implanted using the mask 58 and
`
`
`
`
`
`
`
`
`
`
`
`the gate 50 in combination with oxide inserts 30 to form
`
`
`
`
`
`
`
`shallow N+ source and drain regions 60. The ion im-
`
`
`
`
`
`
`
`
`
`
`plantation is carried out at a sufficient level to overcome
`
`
`
`
`
`
`
`
`
`the original P+ regions 56 in well 38. Theinitial intro-
`
`
`
`
`
`
`
`
`
`
`duction of P+ impurities into the P— well 38 allowsfor
`
`
`
`
`
`
`
`
`
`
`greater range for the formation of the N+ source and
`
`
`
`
`
`
`
`
`
`drain regions 60. The resulting structureis illustrated in
`
`
`
`
`
`
`
`
`
`
`
`FIG. 11. The mask layer 58 is removed. Thestructure at
`
`Page 9 of 11
`
`
`
`
`5
`
`
`
`
`
`
`
`
`
`this point has shallow source and drain regions formed
`
`
`
`
`
`
`
`
`in each of the complementary field effect transistor
`
`
`
`
`
`
`
`
`
`
`
`wells having a depth in the range of 1000 to 2000 Ang-
`
`
`
`
`
`
`
`
`stroms and an impurity concentration in the range of
`
`
`
`
`
`10!8 to 10!9 atoms per cubic centimeter.
`
`
`
`
`
`
`
`
`
`
`The next sequence of operation form the deep source
`
`
`
`
`
`
`
`
`and drain region portions. This process begins with the
`
`
`
`
`
`
`
`formation of spacers extending laterally from the gate
`
`
`
`
`
`
`
`
`
`regions 50 and 52. These spacers are formed by deposit-
`
`
`
`
`
`
`
`
`ing an insulative layer, for example chemical vapor
`
`
`
`
`
`
`
`
`
`deposition of silicon dioxide over the surface of the
`
`
`
`
`
`
`
`
`
`
`substrate. The oxide layer is then reactive ion etched to
`
`
`
`
`
`
`
`
`
`
`form the spacer regions 62 extending from gate 50 and
`
`
`
`
`
`
`
`
`
`spacers 64 extending from gate 52. These spacersresult
`
`
`
`
`
`
`
`
`
`from the uneven topology of the deposited silicon oxide
`
`
`
`
`
`
`
`
`
`
`
`layer. This process is well known in the prior art and,
`thus, is not described in detail.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`After formation of the spacers, a masking layer 66 is
`
`
`
`
`
`
`
`
`applied over the substrate and delineated to mask N—
`
`
`
`
`
`
`
`
`
`well region 46 and exposed P— well region 38. N+
`
`
`
`
`
`
`
`impurities, for example phosphorous, are ion implanted
`
`
`
`
`
`
`
`
`
`
`
`
`using the mask layers 66, the gate 50 and the spacers 62
`
`
`
`
`
`
`
`
`
`
`
`
`
`as well as inserts 30 as a mask. This results in deep N+
`
`
`
`
`
`
`
`
`
`source and drain regions 68 extending laterally from the
`
`
`
`
`
`
`
`
`
`
`shallow N+ source and drain regions60 as illustrated in
`
`
`
`
`
`
`
`
`
`
`
`FIG. 12. The masking layer 66 is removed and a new
`
`
`
`
`
`
`
`
`
`masking layer 70 is applied and delineated to expose
`
`
`
`
`
`
`
`
`
`well 46 and mask well 38. P-type impurities, for exam-
`
`
`
`
`
`
`
`
`
`
`
`ple boron, are ion implanted using the mask layer 70, the
`
`
`
`
`
`
`
`
`
`
`
`gate 52, spacers 64 and the oxide inserts 30 as alignment
`
`
`
`
`
`
`
`
`
`
`mask. This results in the deep P+ source and drain
`
`
`
`
`
`
`
`
`regions 72 extending laterally from the shallow source
`
`
`
`
`
`
`
`
`
`
`
`and drain regions 54 as illustrated in FIG. 13 having a
`
`
`
`
`
`
`
`
`
`
`
`depth in the range of 3000 to 4000 Angstroms and an
`
`
`
`
`
`
`
`impurity concentration in the range of 10!8 to 1020
`
`
`
`
`
`
`
`
`
`
`
`atoms per cubic centimeter. The photoresist layer 70 is
`
`
`
`
`
`
`
`removed. This completes the device formation steps.
`
`
`
`
`
`
`
`
`It is evident from FIG. 13 that the polycrystalline
`
`
`
`
`
`
`
`
`
`
`gates 50 and 52 are of a lowerresistance and are doped
`
`
`
`
`
`
`
`
`
`as N+ and P+ respectively having an impurity con-
`
`
`
`
`
`
`
`
`centration in the range of 1018 to 1020 atoms per cubic
`
`
`centimeter.
`
`The formation of the field oxide and the contacts is
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`the final processing sequence. In order to reduce the
`
`
`
`
`
`
`
`junction resistance and provide barrier metal for shal-
`
`
`
`
`
`
`
`
`
`low junctions, a silicide forming metal layer 74 is ap-
`
`
`
`
`
`
`
`
`
`
`plied to the substrate as illustrated in FIG. 14. This layer
`
`
`
`
`
`
`
`
`
`may be, for example platinum. The wafer is then heated
`
`
`
`
`
`
`
`
`
`to cause platinumsilicide to form in those regions where
`
`
`
`
`
`
`
`the platinum contacts the polycrystalline silicon gate or
`
`
`
`
`
`
`
`
`
`the silicon substrate. Asillustrated in FIG. 15, this in-
`
`
`
`
`
`
`
`
`cludes the platinum silicide regions 76 on deep source
`
`
`
`
`
`
`
`
`
`
`
`and drain regions 68, region 78 on gate 50, regions 80 on
`
`
`
`
`
`
`
`
`
`
`deep source and drain regions 72 and region 82 on gate
`
`
`
`
`
`
`
`
`
`52. No platinum silicide is formed over the oxideinserts
`
`
`
`
`
`
`
`
`
`
`
`30 or the spacers 62 and 64. The portions of the plati-
`
`
`
`
`
`
`
`
`num which should not form platinum silicide is easily
`
`
`
`
`
`
`selectively removed using, for example, hot concen-
`
`
`
`trated Aqua Regia.
`
`
`
`
`
`
`
`
`
`
`The use of the oxide spacers inhibit hot electron and
`
`
`
`
`
`
`
`
`substrate current injection as well as providing better
`
`
`
`
`
`
`
`
`gate oxide reliability. Used in combination with the
`
`
`
`
`
`
`
`metal forming silicides, a self-alignment technique for
`
`
`
`
`
`
`
`the silicide formation without a mask results. The spac-
`
`
`
`
`
`
`
`
`
`
`ers also smooth out the sharp edges of the gate material
`
`
`
`
`
`for better contact metal coverage.
`
`
`
`
`
`
`
`
`
`
`A field oxide layer 84 is then formed over the wafer
`
`
`
`
`
`
`
`
`by chemical vapor deposition for example andvias are
`
`
`
`
`
`
`
`5
`
`
`
`— 0
`
`
`
`
`
`20
`
`
`25
`
`
`30
`
`
`
`
`40
`
`
`45
`
`
`
`
`
`
`
`
`65
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`4,599,789
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`6
`
`
`
`
`
`
`
`
`etched to provide contacts to the platinumsilicide por-
`
`
`
`
`
`
`
`
`
`
`tion of the source and drain regions and the gate re-
`
`
`
`
`
`
`
`
`
`
`gions. A layer of contact metalis then applied and delin-
`
`
`
`
`
`
`
`
`
`eated to form contacts 86. The resulting structure is
`
`
`
`
`
`
`
`
`
`illustrated in FIG. 16. An insulative layer may be pro-
`
`
`
`
`
`
`
`
`
`
`vided overthe first metal layer 86 and a second layer of
`
`
`metal provided.
`
`
`
`
`
`
`
`
`An alternative method for forming the shallow and
`
`
`
`
`
`
`
`
`
`deep source and drain regions includes, asillustrated in
`
`
`
`
`
`
`
`
`FIG.17, forming the masking layer 58 blocking the N—
`
`
`
`
`
`
`
`
`
`
`well region 46 and exposing the P— well region 38. The
`
`
`
`
`
`
`
`
`N-+ arsenic implant
`is performed to form shallow
`
`
`
`
`
`
`
`
`
`
`source and drain regions 60. The photoresist layer 58 is
`
`
`
`
`
`
`
`
`removed and a photoresist layer 88 is applied and delin-.
`
`
`
`
`
`
`
`
`
`eated to form a mask layer blocking the P— well region
`
`
`
`
`
`
`
`
`
`38 and exposing the N— well region 46. P-++ type impu-
`
`
`
`
`
`
`
`
`
`
`tities are implanted to form the P+ source and drain
`
`
`
`
`
`
`
`
`
`
`regions 54. The masking layer 88 is removed and a
`
`
`
`
`
`
`non-selective P-type boron implant is performed form-
`
`
`
`
`
`
`
`
`
`
`ing deep P+ source and drain regions 72 and P+
`
`
`
`
`
`
`
`
`
`
`source and drain regions 90 in wells 46 and 38 respec-
`
`
`
`
`
`
`
`
`
`tively. A masking layer 66 is applied and delineated to
`
`
`
`
`
`
`
`
`expose the P— well region 38 and block the N— well
`
`
`
`
`
`
`region 46. N-type impurities, for example phosphorous,
`
`
`
`
`
`
`
`
`
`are implanted to form the deep N+ source and drain
`
`
`
`
`
`
`
`
`regions 68. This implantation must be carried out at
`
`
`
`
`
`
`
`
`sufficient power levels to overcome the P+ regions 90
`
`
`
`
`
`
`
`
`originally formed in the prior step. The resulting struc-
`
`
`
`
`
`
`
`
`
`
`
`ture is illustrated in FIG. 20. The mask layer 66 is re-
`
`
`
`
`
`
`
`
`moved and the process is continued as previously de-
`scribed for FIGS. 14-16.
`
`
`
`
`
`
`
`
`
`
`
`The process of FIGS. 17-20 are substantially similar
`
`
`
`
`
`
`
`
`
`
`to the process steps of FIGS. 10-13 in that three mask-
`
`
`
`
`
`
`
`
`
`
`ing steps are used for four implantations to form the two
`
`
`
`
`
`
`
`part source and drain regions for complementary insu-
`
`
`
`
`
`
`
`
`lated gate field effect transistors. This results since one
`
`
`
`
`
`
`
`
`of the implantation or doping step is performed non-
`
`
`
`
`
`
`
`
`selectively. The difference is, in FIGS. 10-13, the non-
`
`
`
`
`
`
`
`
`selective implantation is at the beginning of the process,
`
`
`
`
`
`
`
`
`whereas in the process of FIGS. 17-20, the non-selec-
`
`
`
`
`
`
`
`
`
`
`
`tive implantation is in the middle of the process. As can
`
`
`
`
`
`
`
`
`
`
`
`be noted, the implantation of the P and N source and
`
`
`
`
`
`
`
`
`
`drain regions have been reversed in the two process
`
`
`
`
`
`
`
`
`
`flows. Since all the impurity introduction is by low
`
`
`
`
`
`
`
`
`
`
`temperature ion implantation,it is not critical that the N
`
`
`
`
`
`
`implantation precede the P implantation. The impor-
`
`
`
`
`
`
`
`
`
`
`tance of the sequenceofstepsis that the shallow source
`
`
`
`
`
`
`
`
`
`
`and drain regions are formed using the gate as the align-
`
`
`
`
`
`
`
`
`
`ment mask and that the deeper regions are formed using
`
`
`
`
`
`
`the spacer as the alignment mask.
`
`
`
`
`
`
`
`From the preceding description of the preferred em-
`
`
`
`
`
`
`
`
`
`bodiments,it is evident that the objects of the invention
`
`
`
`
`
`
`
`
`are attained, and although the invention has been de-
`
`
`
`
`
`
`
`
`
`
`scribed andillustrated in detail, it is to be clearly under-
`
`
`
`
`
`
`
`
`
`
`stood that the same is by way ofillustration and exam-
`
`
`
`
`
`
`
`
`
`
`
`ple only andis not to be taken by wayoflimitation. The
`
`
`
`
`
`
`
`
`
`
`spirit and scope of the invention are to be limited only
`
`
`
`
`
`
`
`by the terms of the appended claims.
`Whatis claimedis:
`
`
`
`
`
`
`
`
`
`
`1. A method for forming complementary insulated
`
`
`
`
`
`gate field effect transistors comprising:
`
`
`
`
`
`
`
`
`
`
`forming at least one insulative insert in a substrate of
`
`
`
`
`
`
`
`
`
`
`a first conductivity type to form atleast a first and
`
`
`
`
`
`
`
`
`second mesasin said surface separated laterally by
`
`
`said insert;
`
`
`
`
`
`
`
`
`
`forming a first mask layer having a first opening on
`
`
`said substrate;
`
`Page 10 of 11
`
`Page 10 of 11
`
`
`
`4,599,789
`
`
`7
`
`
`
`
`
`introducing second conductivity type impurities op-
`
`
`
`
`
`
`
`posite said first conductivity type into said sub-
`
`
`
`
`
`
`
`strate through said first opening to form simulta-
`
`
`
`
`
`
`
`
`neously a second conductivity type well in said
`
`
`
`
`
`
`second mesa having a higher impurity concentra-
`
`
`
`
`
`
`
`
`
`
`tion at the bottom of said well and guard ring under
`
`
`
`
`