ShallowJunctions-_MS_contacts

ShallowJunctions-_MS_contacts - Shallow Junctions &...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Shallow Junctions & Metal Semiconductor Contacts Prof. Krishna Saraswat Department of Electrical Engineering Stanford University Stanford, CA 94305 saraswat@stanford.edu araswat! tanford University 1 EE311 / Shallow Junctions/MS Contacts Outline • Junction/contact scaling issues • Shallow junction technology • Ohmic contacts • Technology to form contacts araswat! tanford University 2 Page 1 EE311 / Shallow Junctions/MS Contacts MOS Device Scaling ! L xox Xj L N+ xox N+ Na N+ Xj N+ Na lo lo P P Why do we scale MOS transistors?! Constant E Field Scaling All device parameters are scaled by the same factor. •  gate oxide thickness xox ↓ •  channel length L ↓ •  source/drain junction depth Xj ↓ •  Channel doping ↑ •  Supply voltage VDD ↓ 1. Improve frequency response α 1/L " 3.  " "Increase device packing density" 2.  Improve current drive (transconductance gm) R. Dennard, et al., IEEE Journal of Solid State Circuits, Oct. 1974. " gm = " ID "VG VD = const # W Kox µ V L n to x D # W Kox µ (VG $ VT ) for VD > VDSAT , saturation region L n to x for VD < VD SAT , linear region Why do we need to scale junction depth? araswat! tanford University 3 EE311 / Shallow Junctions/MS Contacts Short Channel Effects on Threshold voltage Ddepletion width in a long channel device W= 2" ( 2# F + VBG ) qN A Gate Gate We can approximate, the bulk charge as ! # L + L' & QB " L = q " N A " W " % %2( ( $ ' By simple geometry, we can write: N+ source N+ source L Depletion Depletion region egion Lʼ !"#$%&'(")*"+"(',-' Q depleted QB depleted ,%+.'#/+"'/&('0%123"' by source by source P-Si P-Si N+ drain N+ drain rj !"#$%&'(")*"+"(',-' QB depleted B depleted ,%+.'#/+"'/&('(2/$&' by drain by drain $ 'r L + L' 2#W j = 1" & 1 + " 1) # & )L 2L rj % ( We can then approximate the threshold voltage as: VT = VFB ! 2 " # F ! ' rj QB * $ 2"W " ,1 ! & 1 + ! 1) " / Cox + % rj ( L. Threshold voltage is a function of junction depth, depletion width and channel length? araswat! tanford University L. Yau, Solid-State Electronics, vol. 17, pp. 1059, 1974 4 Page 2 EE311 / Shallow Junctions/MS Contacts Need for Shallow Source/Drain Junctions ! VDD poly gate poly gate n+ n+ n+ n+ p-well p-well VTH VTH V TH 0 VTH rj VD L L VD ' rj Q*$ 2"W VT = VFB ! 2 " # F ! B " ,1 ! & 1 + ! 1) " / Cox + % rj ( L. • Roll-off in threshold voltage as the channel length is reduced and drain voltage is increased • To minimimize VT roll-off • Reduce as junction depth and drain depletion width • Increase in Cox should increase gate control araswat! Sheet tanford University resistance increases as junction depth is reduced 5 EE311 / Shallow Junctions/MS Contacts Silicide metal Poly-Si Rc Silicide Xj source Rs Rs’ Rch Rd’ Rd drain Lch t ox ⇒  Scaled with Lg Rch " (V gs ! Vth ) (Lch ↓, tox↓) 1 Rsd ! Rsh ! N sd X j 70 60 50 40 2001 ITRS Physical Gate Length 50 Max. Ratio of Rsd to Ideal Rch 40 30 30 20 20 10 0 2000 60 10 SDE Junction Depth 2004 2008 2012 2016 Rsd/Rch-ideal [%] Gate Length or SDE Depth [nm] S/D Junction Scaling Trend 0 Year ⇒  Difficult to scale (Nsd const, Xj↓) ⇒  Rsd/Rch ↑ Ref: J. Woo (UCLA) •  As Lg scales down, Rsd becomes comparable to Rch •  Rsd becomes important factor for device current •  Parasitic portion of the device is now playing important role in device performance and CMOS scaling araswat! tanford University 6 Page 3 EE311 / Shallow Junctions/MS Contacts Sidewall Series Resistance (ohms) Impact of Parasitic Series Resistance: NMOS y=0 Gate Silicide Rcsd Rdp Rext Rov Nov(y) x 140 120 NMOS 100 80 60 Rdp Rcsd 20 0 30 nm 50 nm 70 nm 100 nm Physical Gate Length Relative Contribution [%] 70 60 50 tanford University Rcsd NMOS 40 Rext 30 Rov 20 10 0 Rdp 32 nm 53 nm 70 nm 100 nm Physical Gate Length Kim, Park & Woo, IEEE TED, MARCH 2002 araswat! Rov Rext 40 Next(x) Problem in junction scaling: •  Sheet resistance of a junction is a strong function of doping density •  Maximum doping density is limited by solid solubility and it does not scale ! •  Contact resistance Rcsd is one of the dominant components for future technology •  Silicidation can minimize the impact of junction sheet resistance Scaled by ITRS Roadmap 7 EE311 / Shallow Junctions/MS Contacts Relative Contribution [%] Series Resistance (ohms) Relative Contributions of Resistance Components: PMOS 200 PMOS Scaled by ITRS Roadmap 150 Rov 100 Rext 50 Rdp Rcsd 0 30 nm 50 nm 70 nm 100 nm Physical Gate Length 70 Rcsd 60 PMOS 50 40 30 Rov 20 Rext 10 0 Rdp 32 nm 53 nm 70 nm 100 nm Physical Gate Length •  Problem even more serious for PMOS •  Rcsd will be a dominant component for highly scaled nanometer transistor ( Rcsd/Rseries ↑ >> ~ 60 % for LG < 53 nm) Kim, Park & Woo, IEEE TED, MARCH 2002 araswat! tanford University 8 Page 4 EE311 / Shallow Junctions/MS Contacts Contact Resistance Problem: Light Emission Metal Rc Rdiff n+ Ge i Ge •  Direct band gap emission from Ge p-i-n heterojunction photodiode. •  Shift in wavelength with higher current density •  Power dissipation through Rc + Rdiff heats the diode araswat! tanford University E. Kasper et al, University of Stuttgart, Germany 9 EE311 / Shallow Junctions/MS Contacts Outline • Junction/contact scaling issues • Shallow junction technology • Ohmic contacts • Technology to form contacts araswat! tanford University 10 Page 5 EE311 / Shallow Junctions/MS Contacts Dopant Diffusion Implanted region •  Solutions to diffusion equations (Fick's laws) gives bulk diffusivity Di = D io " e _ EO k" T • In shallow junction technologies, numerous effects alter these values resulting in enhanced diffusion. • Transient enhanced diffusion D = Di + D o " e _t # • Diffusion affected by defects, e.g., oxidation induced point defects araswat! tanford University 11 EE311 / Shallow Junctions/MS Contacts Diffusion Affected by Oxidation Induced Point Defects TSUPREM IV simulations of oxidation enhanced diffusion of boron (OED) and oxidation retarded diffusion of antimony (ORD) during the growth of a thermal oxide on the surface of silicon." Sb B Oxidation increases interstitials (CI) and decreases vacancies (CV) from their equilibrium values. This in turn changes diffusivity." (Ref: Plummer, et al., Silicon VLSI Technology - Fundamentals, Practice and Models)! araswat! tanford University 12 Page 6 EE311 / Shallow Junctions/MS Contacts Transient Enhanced Diffusion (TED)! 40 keV, 10-14 cm-2 B 750ºC anneal τ At lower temperatures, the ion implantation damage can stay around longer and enhances the dopant diffusion, while at higher temperatures the damage annihilates faster. Thus the diffusivity is a function of time during the transient. % t( D = Di + Do " exp'# * & $) Where # E& Di = Dio exp%" 0 ( is intrinsic diffusity $ kT ' (Ref: Plummer, et al., Silicon VLSI Technology - Fundamentals, Practice and Models)! araswat! tanford University 13 ! EE311 / Shallow Junctions/MS Contacts ! Shallow Junction Formation Technologies: Low Energy Implantation 12 keV B implants" Concentration (cm-3)" Concentration (cm-3)! 40 keV As and B implants" Boron! Arsenic! Boron! BF2! Depth! As Concentration (cm-3) Depth! ) 3 - 10 10 22 m c 1018 ( s A 1016 araswat! tanford University as-implanted 20 5 keV 1 keV 0 20 40 60 Depth (nm) 14 Page 7 Ref. Kasnavi, PhD Thesis ! Stanford Univ. 2001! 80 EE311 / Shallow Junctions/MS Contacts Junction Depth Vs. Sheet Resistance Tradeoff 60 Junction Depth (nm) 5 keV limit 50 Roadmap Y=2000, L g=180nm 40 1 keV limit ) m 30 n ( j X 2002, 130nm 2005, 100nm 20 2008, 70nm 2011, 50nm 10 0 2014, 35nm 0 250 500 Rs ( ! / ) 1020C spike 750 1000 Ref. Kasnavi, PhD Thesis Stanford Univ. 2001 ! It will be difficult to meet the ITRS scaling requirments of junction depth and sheet resistance araswat! tanford University 15 EE311 / Shallow Junctions/MS Contacts Solid Source Diffusion In Si after silicide removal Depth (nm)! Depth (nm)! B Concentration (cm-3)" In COSi2   Boron profiles after diffusion at 950°C of 50 nm COSi2 implanted with 5 X 1015 cm-2 BF2 (a) and (b) in Si after silicide removal.   This technique is used to obtain shallow junctions with high surface concentration to reduce contact resistance araswat! tanford University 16 Page 8 EE311 / Shallow Junctions/MS Contacts Temperatue-time ranges of various conventional and advanced annealing techniques Ref: G. Thareja, PhD Thesis, Stanford Univ., 2011 araswat! tanford University 17 EE311 / Shallow Junctions/MS Contacts Laser Annealing of Ion Implants SIMS profiles for ion Implanted Sb Laser annealing scenario Sb - 10keV, 5 x 1015 cm-2 22 Incident Laser Energy (I) Temperature L 10 Concentration (cm-3) a/c boundary depth SIMS - as implanted SIMS - 0.4 J/cm2 SIMS - 0.2 J/cm2 SIMS - 0.7 J/cm2 SRP - 0.7 J/cm2 21 10 20 10 19 10 18 10 -1 Absorption Depth ~ ! 0 500 1000 1500 2000 Depth (A) •  It is possible to get metastable electrically active doping density higher than the equilibrium doping density Ref: G. Thareja, PhD Thesis, Stanford Univ., 2011 araswat! tanford University 18 Page 9 EE311 / Shallow Junctions/MS Contacts Gas Immersion Laser Doping (GILD) •  Si wafer showing the adsorption of the dopant species onto the clean silicon surface. The dopant is incorporated into a very shallow region upon exposure to the excimer laser pulse. •  It is possible to get metastable electrically active doping density higher than the equilibrium doping density araswat! tanford University 19 EE311 / Shallow Junctions/MS Contacts Solutions to Shallow Junction Resistance Problem Poly-Si spacer Silicide source source Poly-Si drain drain extensions Extension implants Elevated source/ drain Silicide Poly-Si source Silicide Silicide drain Schottky Source/Drain Silicidation araswat! tanford University 20 Page 10 EE311 / Shallow Junctions/MS Contacts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`$"%U$&"1)&8@).&a=0$18/$@$17&4:&\@"#81"!&=403%"K&>38$9&:43&>""L& !"#$%")& (#<#$7&-) /&0) 079/&!() 8'5') /$!#?/!'0) @;) 5/9#0) !"'5,/<) -EF(-.& -!F5)6-& $'& ()*M*9JKI;&D$!12&:43&<=>&6(9;& =0/;$%349&]YZ=a.&\==>\<]&5HHH)& 7"# 8292:2*'2&# /&&'/<2) >'?#$') 957$'((#&6) 8/() $7,9<'!'0) 8#!") /) @/$%) !"#$%"& D$12& 5((9;& /) ABCD) #&!'5) <'?'<) 0#'<'$!5#$) /&0) E#FE#GF3<)RGS&J)&Q4U8`494&"1)&8@).&b=403%"K&>38$9&\9E$9""3$9E&:43&=0/O5((9;& '130%103"'.& %4;L83"!& 14& '&0) 74) <=>&B49& 89!& B4::& D8'& 8@'4& ?2"& 3"@81$49'2$L& ,'!/<2)3)0#/65/,)74)/)4#&#("'0)0'?#$')#()("78&)#&)4#6H5')]YZ=&c'$9E&="@"%1$#"&\L$18T$8@&_34D12&?"%29$d0"e.&B\>Y&*((()& /"1D""9&6(9;& 3":"3"9%"& !"#$%")& N4& 6559JKI;& D$!12& :43& 12"& R5S&B?<=&<48!;8L.&*((5)& $9#"'1$E81"!.& 8'& '24D9& $9& :$E03"& 55)&CJK) 9#$!H5') $9& /) $7,9<'!'0)R,S&_)&_0"E89&"1)&8@).&b_81"&89!&=403%"K&>38$9&\9E$9""3$9E&:43&6(9;& $;L34#";"91& 43&I2) 3) $57(() ('$!#7&/<) B4::.& ;"8'03"!& 81& fO]2899"@&YZ=[\?e.&\==>\<]&*((5)& 4:&\@"#81"!&=403%"K&>38$9&:43&>""L& !"E38!81$49& $9& B49& $'& '""9& 74) 12"'"& R*S&<)&_34`$"%U$&"1)&8@).&a=0$18/$@$17& -EF(-.& -!F5)6-& $'&KLCMJE) (!5H$!H5') #() ("78&) #&) 4#6H5') 1-) /&0) /) !#<!'0)R6S&])&]8$@@81&"1)&8@).&bP69;&f27'$%8@&_81"&X"9E12&NYZ=[\?'&D$12& ()*M*9JKI;&D$!12&:43&<=>&6(9;& !"#$%"'&:43&8&E$#"9&B4::)& CJK)9#$!H5')74)/)KLCMJE)(!5H$!H5'-)("78#&6)!"')/$!#?') =0/;$%349&]YZ=a.&\==>\<]&5HHH)& !"#$%"& D$12& 5((9;& <=>& '130%103"'.& %4;L83"!& 14& Q"8#7&B49&B;L@891"!&f4%U"1'&89!&Q$E2@7&<"@$8/@"&*9;O&?2$%U&_81"& /5'/)@7H&0'0)@;)!"')NLOLC)7:#0/!#7&)7:#0'-)#()("78&) RGS&J)&Q4U8`494&"1)&8@).&b=403%"K&>38$9&\9E$9""3$9E&:43&=0/O5((9;& ?2"& '0/O123"'24@!& '@4L"& E"') '9#!/:#/<) </;'5() /5') $'& 8@'4&YZ=&c'$9E&="@"%1$#"&\L$18T$8@&_34D12&?"%29$d0"e.&B\>Y&*((()& #&) 4#6H5') P2) A=C& 3":"3"9%"& !"#$%")& 45''-) 6559JKI;& D$!12& :43& 12"& 6(9;& 4:& ';8@@& !"#$%"'&4/$'!) N4& !"')ZT$!"&:43&5)6-&ZL"381$49e.&-X=B&=7;L4'$0;.&5HHH)& $;L34#"!& /7& 12"&957$'(() #() ('<'$!#?') !7) @7!") (#<#$7&) 0#7:#0') /&0) (#<#$7&) R,S&_)&_0"E89&"1)&8@).&b_81"&89!&=403%"K&>38$9&\9E$9""3$9E&:43&6(9;& 8!!$1$49& 4:& <=>&'130%103"')&?2"&6(9;& $;L34#";"91& 43& !"E38!81$49&!"'5') B49& $'& '""9& $9& 12"'"& &#!5#0'-) /&0) $9& #() &7) @5#06#&6) @'!8''&) !"') 5/#('0) fO]2899"@&YZ=[\?e.&\==>\<]&*((5)& !"#$(#' 3":"3"9%"& 949O"@"#81"!& '$@$%49& !"#$%"& 28'& 8& '0/O (7H5$')/&0)05/#&)/&0)!"')97<;(#<#$7&)6/!')'<'$!570'2) !"#$%"'&:43&8&E$#"9&B4::)& 4:& 5(*;-K!"%)& =& :43& 12"& 6(9;& R6S&])&]8$@@81&"1)&8@).&bP69;&f27'$%8@&_81"&X"9E12&NYZ=[\?'&D$12& 123"'24@!& '@4L"& A=C& Raised S/D Technology by Selective Epitaxial Growth (SEG) !&#&$-.(=&%1&8"( $'& 14& 3"!0%"& 12"& 8;4091& 4:& !4L891& !$::0'$9E& $914& 12"& !&8">+##( "T1"9'$49'& 4:& 12"& 38$'"!& '403%"K& !38$9& !"#$%"'& 83"& M#6H5') 12) O57(() ('$!#7&/<) CJK) 9#$!H5') 74) /) KLCMJE) (!5H$!H5'2) !"#$(#' !"#$%#' •  Plot of VG vs. ID for a 50nm raised source/ drain !"#$(## MOSFET with a 100nm !"#$%#& # #") ! !") epitaxy layer, measured at a /373!( !"#$%&'(#")$%*&+ !"#"$%&'"()*&%+,&+#( drain voltage of 1.5V (dashed !&#&$-.(/+0"1( & !"#$%#' line), and 0.1V (solid line). [$E03"& P)& f@41& 4:& E81"& #4@18E"& #')& !38$9& %033"91&•:43& 8&   Improvement in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f4%U"1'&89!&Q$E2@7&<"@$8/@"&*9;O&?2$%U&_81"& ) !"#$(## ZT$!"&:43&5)6-&ZL"381$49e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f@41& 4:& E81"& #4@18E"& #')& !38$9& %033"91& :43& 8& (7H5$')/&0)05/#&)5'6#7&(2) #"+) &#!5#0')(#0'8/<<)(9/$'5(2) "T1"9'$49'& !03$9E& L34%"''$9E)& 4:& 12"&;"89'& 1281& 12"& 5((9;& '1""L"3& '0/& 123"'24@!& '@4L"& ?2$'& <=>& !"#$%"'& $'& 6(9;& 38$'"!& '403%"K& !38$9& YZ=[\?& D$12& 8& 5((9;& !"! #$%&'()&*$+,%-.$'-+ #"+ • 4:& 5)6-& "TL@8$9"!& 12"& 38$'"!& E81"& %49134@& 49& 12"& %2899"@& "T1"9'$49'& 4:& /7& 12"& /"11"3&'403%"K& !38$9& !"#$%"'& 83"& "L$18T7& @87"3.& ;"8'03"!& 81& 8& !38$9& #4@18E"&  Plot of VT vs. channel !"#"$%&'"()*&%+,&+#( !"#"$%&'"(61->%C(-.( L41"91$8@& 8'& 12"& '28@@4D"3& V09%1$49'& 12"& '$@$%49& S0) ?(2) #"&) T6) $H5?'() ,'/(H5'0) 7&) /) P+&,) UC>) '28@@4D"3& 89!& ;43"& 3"'$'1$#"& 1289& 12"'"& 4:&3"!0%"& 12"& length for devices with !&#&$-.( <-#0;&#&$-.(6+%"( A!8'2"!&@$9"C.&89!&()5-&A'4@$!&@$9"C)&GKLC) 0'?#$')/!)"#6")/&0)<78)05/#&)@#/()/5')("78&)#&)4#6H5').2) L"9"1381$49&4:&12"&"@"%13$%&:$"@!&:34;&12"&'403%"&89!&!38$9& )#"$%1-8"( +(9;& 3":"3"9%"&!"#$%".&89!&1281&12$'&"::"%1&$'&;43"&;83U"!&8'& various SEG Si layer 3)9<7!)74)!"5'("7<0)?7<!/6')?(2)0'?#$')$"/&&'<)<'&6!")#() #"& $914& 12"& !"#$%"& %2899"@)& ?2"& 2$E2"3& 3"'$'1$#$17& 4:& 12"& ("78&) #&) 4#6H5') V2) E"') !"5'("7<0) ?7<!/6') 57<<) 744) W!"') 12"& 12$%U9"''& 4:& 12"& "L$18T$8@& '$@$%49& @87"3& $'& $9%3"8'"!)& thicknesses. ABB(?@*#+.%"8(&.%-( "T1"9'$49'&89!&3"!0%1$49&$9& '2431&%2899"@&123"'24@!&34@@& #")) 0#44'5'&$') @'!8''&) !"') !"5'("7<0) ?7<!/6') 74) P+&,) !)6(!&#&$-.( #", P(9;& <":"3"9%"& ?2"& '28@@4D"3& V09%1$49'& 4:& 12"& "@"#81"!& !"#$%"'& E$#"& 0'?#$'()/&0)<7&6)$"/&&'<)0'?#$'(X)#()("78&)#&)4#6H5')*2) 4::&8%%4091'&:43&12"&@4''&4:&%033"91&!3$#"&$9&12"&"@"#81"!& •  Improvement in Vt E"')#") '44'$!) 74) !"') ('<'$!#?') '9#!/:#/<) (#<#$7&) 957$'(() /"11"3&!"#$%"')&?2$'&%89&/"&%4;L"9'81"!&:43&/7&4L1$;$'81$49&4:& '2431& %2899"@& /"28#$403& 89!& >BWX& R*S)& ?2"& #"+) rolloff and DIBL 5'0H$#&6) !"') !"5'("7<0) ?7<!/6') 57<<) 744) $/&) $<'/5<;) @') 5((9;& '1""L"3& '0/& 123"'24@!& '@4L"& 4:& 12"& <=>& !"#$%"'& $'& #"*) 12"&Q>>&$;L@891&:43&12"&"@"#81"!&!"#$%"')& (''&2)E"')!"5'("7<0)?7<!/6')57<<)744)74)!"')P+&,)(#<#$7&) #"+ 5'4'5'&$')0'?#$'()W8#!")&7)5/#('0)(7H5$'F)05/#&)(!5H$!H5'X) "TL@8$9"!& /7& 12"& /"11"3& E81"& %49134@& 49& 12"& %2899"@& #"* #() QY*,T-) $7,9/5'0) !7) ZH(!) PI,T) 475) !"') 0'?#$'() 8#!") L41"91$8@&+"# 8'& ,)*'-(&%)*&# 12"& '28@@4D"3& V09%1$49'& 3"!0%"& 12"&) #"#! #"! ! !# #"&) !"')Q++&,)!"#$%)'9#)</;'5(2) M#6H5') I2) M#&/<) KLCMJE) 0'?#$') (!5H$!H5'2 L"9"1381$49&4:&12"&"@"%13$%&:$"@!&:34;&12"&'403%"&89!&!38$9& Waite, et al., essderc-esscirc-2003 +(9;& ,$3.4$%05"//$(%6$/)#5%*78+ <8$'"!&'403%"K&!38$9&YZ=[\?'.&D$12&%2899"@&@"9E12'& araswat! E"') #"& #,957?','&!) #&) >SAN) 9'5475,/&$') 74) !"') $914& 12"& !"#$%"& %2899"@)& ?2"& 2$E2"3& D$12& 8& :8%"1& :3""& & !4D9& 14& 6(9;.& 28#"& /""9& ;8!"& 3"'$'1$#$17& 4:& 12"& 0'?#$'() $/&) @') (''&) #&) 4#6H5') Y-) ("78#&6) !"') 0#44'5'&$') tanford University EE311 / Shallow Junctions/MS Contacts #&) !"5'("7<0) ?7<!/6')21 WT0#@<X) 74) !"') 0'?#$'() ,'/(H5'0) /!) "T1"9'$49'&89!&3"!0%1$49&$9& '2431&%2899"@&123"'24@!&34@@& #")) '"@"%1$#"&"L$18T$8@&'$@$%49&L34%"''&D$12&8&12"3;8@&/0!E"1& T0[+2QT) /&0) T0[Q2PT2) T0#@<) 475) !"') P+&,) %2899"@& P(9;& <":"3"9%"& [$E03"& M)& f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f@41& 4:& ?23"'24@!& #4@18E"& #')& %2899"@& @"9E12& :43& Raised S/D !"#$%"'& D$12& #83$40'& '"@"%1$#"& "L$18T7& '$@$%49& @87"3& Technology by SEG 12$%U9"''"')& From A. Hokazono et al (Toshiba), IEDM2000 Rcsd !L $ ! = c coth # contact & LT " LT % LT = " q" b !c ! exp $ $N # contact !c R sh ,dp % ' ' & •  Elevated S/D structure ⇒ Reduction of Rcsd by increasing Ncontact & reducing Rsh,dp underneath silicide •  Elevated S/D structure reduce E-field in the drain region and DIBL araswat! tanford University 22 Page 11 EE311 / Shallow Junctions/MS Contacts New Structures and Materials for Nanoscale MOSFETs 5 3 Top Gate G 4 2 S C Si D Source SiO2 1 High µ channel Si BULK Drain Bottom Gate High-K Double gate SOI 1. Electrostatics - Double Gate - Retain gate control over channel - Minimize OFF-state drain-source leakage 2. Transport - High Mobility Channel - High mobility/injection velocity - High drive current for low intrinsic delay 3. Parasitics - Schottky S/D - Reduced extrinsic resistance 4. Gate leakage - High-K dielectrics - Reduced power consumption 5. Gate depletion - Metal gate araswat! tanford University 23 EE311 / Shallow Junctions/MS Contacts Schottky Source/Drain MOSFET Schottky S/D MOSFET Junction S/D MOSFET Possible advantages •  Better utilization of the metal/semiconductor interface  Possible option to overcome the higher parasitic resistance •  Modulation of the source barrier by the gate  High Vg ⇒ barrier thin ⇒ tunneling current ⇑ ⇒ ION ⇑  Low Vg ⇒ barrier thick ⇒ tunneling current ⇓ ⇒ IOFF ⇓ •  Better immunity from short channel effects Possible Disadvantage •  Tradeoff between short channel effect vs. ION reduction due to the Schottky barrier araswat! tanford University 24 Page 12 EE311 / Shallow Junctions/MS Contacts Metal Source/Drain MOS Performance Issue   For NMOS need ΦBn= 0 to increase Ion and reduces drain leakage   Schottky S/D NMOS Similarly need ΦBp= 0 for PMOS For Ge ΦBp= 0.1 eV, excellent for PMOS, ΦBn= 0.56 eV, really bad for NMOS   On state   For Si ΦBn≈ 0.67 EG and ΦBp= 0.33 EG bad for both NMOS and PMOS   For InAs ΦBn= -0.2 eV, excellent for NMOS, ΦBp= 0.38 eV, really bad for PMOS Ideal Schottky Contact for NMOS Off state ΦBN = 0 Higher on current Lower ambipolar leakage ΦBP = Eg araswat! tanford University 25 EE311 / Shallow Junctions/MS Contacts Doped vs. Schottky S/D DG Device Comparison Simulations ION vs. IOFF CV/I DelayBody thickness SB NM OS Conventional NMOS Source: King/Bokor,U.C. Berkeley Ref: R. Shenoy, PhD Thesis, Stanford 2004 Low barrier height metal contact required to achieve high ION and low CV/I delay   Extensive research needed to develop a low barrier technology araswat!   tanford University 26 Page 13 EE311 / Shallow Junctions/MS Contacts Outline • Junction/contact scaling issues • Shallow junction technology • Ohmic contacts  Need to understand the physics of contacts resistance and develop technology to minmize it • Technology to form contacts araswat! tanford University 27 EE311 / Shallow Junctions/MS Contacts Conduction Mechanisms for Metal/ Semiconductor Contacts ! I Low doping φB Ef V Schottky (a) Thermionic emission Medium doping (b) Thermionic-field emission Heavy doping Ohmic (c) Field emission.! Contact resistance strongly depends on barrier height (φB) and doping density araswat! tanford University 28 Page 14 EE311 / Shallow Junctions/MS Contacts Specific Contact Resistivity (ρc)! V = Vbulk + 2Vcontact = I (Rbulk + 2Rcontact) ! n+ dVbulk !l = dI A Rbulk = !V For a uniform current density " !V Rcontact = dVcontact !c = dI A • Specific contact resistivity and not contact resistance is the fundamental parameter characterizing a contact" araswat! tanford University 29 EE311 / Shallow Junctions/MS Contacts Tunneling - Ohmic Contacts ! Fm ϕB Jsm Xd = Fs 2 K !o "i q Nd When Xd ≤ 2.5 – 5 nm, electrons can “tunnel” through the barrier. Required doping for Si is:" N d min " Net semiconductor to metal current is a function of available electrons and tunneling probability " 2 K #o $i " 6.2 %10 1 9 cm &3 2 q Xd J sm = A*T F P( E )(1 " Fm )dE k!s # 2! B P(E) is the tunneling probability given by" P( E ) ~ exp% - ! $ Current can be shown to be" for X d = 2.5 nm Fs and Fm are Fermi-Dirac distribution functions in metal and semiconductor " sm * & ( N' [ * J s m " exp #2 xd 2 m (q$ B # qV ) / ! % *( '2 # B $ s m * Specific contact resistivity is of the form" "c = " co exp' ! N* & ) 2 ohm + cm 2 ρc primarily depends upon " •  the metal-semiconductor work function, φΒ, " •  doping density, N, in the semiconductor and " •  the effective mass of the carrier, m*. " araswat! tanford University ! 30 Page 15 EE311 / Shallow Junctions/MS Contacts Specific Contact Resistivity to P-type Si! P-type Si! $ 2" !c = ! co exp & B & q! % #sm* ' ) N) ( ohm * cm 2 Specific contact resistivity, ρc ↓! • As doping density N↑! • Barrier height φB ↓! Specific contact resistivity (Ωcm2)" Specific contact resistivity! (S. Swirhun, PhD Thesis, Stanford Univ. 1987)! NA (cm-3)" araswat! tanford University 31 EE311 / Shallow Junctions/MS Contacts Specific Contact Resistivity to N-type Dopants ! Specific contact resistivity (Ωcm2)" • Similar trends for N-type Si" • For a given doping density specific contact resistivity is higher for n-type Si than p-type." • This can be attributed to the barrier height" •  φBn > φBp " • However, lowest value is obtained for N-type Si because of higher solid solubility of n-type dopants" (S. Swirhun, PhD Thesis, Stanford Univ. 1987)! araswat! tanford University ND (cm-3)" 32 Page 16 EE311 / Shallow Junctions/MS Contacts Solid Solubility of Dopants in Silicon • Maximum concentration of dopants is limited by solid solubility (SS) • Problem is worse for p-type dopants (B), solid solubility is lower • Behavior in Ge is opposite: Higher SS for B and lower for P, As PROBLEM: Solid solubility of dopants does not scale !! araswat! tanford University 33 EE311 / Shallow Junctions/MS Contacts Barrier Height of Metals and Silicides to Si Ideal Schottky model Experimental barrier height to n- and p-type Si (φBN hollow symbols and φBP solid symbols) φBn Φm < χ! Φm > χ! Practical barrier exhibit Fermi level pinning φBN ⇒ 2Eg/3! araswat! φBp S. Swirhun, PhD Thesis, Stanford Univ. 1985 φBP ⇒ Eg/3 ! φBN + φBP = Eg ! tanford University 34 Page 17 EE311 / Shallow Junctions/MS Contacts Fermi Level Pinning in Ge Ni contacts to Ge Φbn vs. Φm Ohmic Schottky Pethe and Saraswat, IEEE DRC, 2007 Dimoulas, APL 89, 252110 (2006) Various metals show Schottky behavior with n-Ge and ohmic with p-Ge indicating that the Fermi level is pinned close to EV araswat! tanford University 35 EE311 / Shallow Junctions/MS Contacts Charge Neutrality Level ECNL Yeo, King, and Hu, " J. Appl. Phys., 15 Dec. 2002"   In actual practice, surface states often dominate the observed characteristics of a metal semiconductor contact.   A semiconductor or dielectric surface has gap states due to a variety of reasons. These are spread across the energy gap.   Some of these states are acceptor like and may be neutral or negative   Other energy states are donor like and may be neutral or positive   Energy level at which the dominant character of the interface states changes from donor like to acceptor like is called the charge neutrality level ECNL araswat! tanford University 36 Page 18 EE311 / Shallow Junctions/MS Contacts Metal-induced Gap States (MIGS)   A semiconductor or dielectric surface has gap states due to the broken surface bonds. These are spread across the energy gap. However, this model can not explain pinning.   The wave functions of electrons in the metal tail or decay into the semiconductor in the energy range where the conduction band of the metal overlaps the semiconductor band gap. These resulting states in the forbidden gap are known as metal-induced gap states (MIGS) or simply intrinsic states. Higher ΦBeff ECNL EFM Free state penetration  Pinning -q   The gap states consist of acceptor-like and donor-like states in the bandgap where the gap-state charges are balanced. Assuming, Fermi level is initially above ECNL. Then some portion of acceptor level is occupied and negatively charged. These negative charges pull positive image charges at metal side and creates interface dipole. This dipole pushes the Fermi level down close to ECNL and thus Fermi level is pinned close to ECNL. araswat! tanford University 37 EFS +q EE311 / Shallow Junctions/MS Contacts Fermi Level Pinning: Dipole Model Efm,vac qφBn Efm,eff In the bond polarization theory, the interaction of metal and semiconductor wavefunctions at a Schottky junction are accounted for the formation of an interface specific region, where the electronic states are not fully metal or semiconductor-like, but rather a mixture of the two. Minimization of the total interface energy causes the metal and semiconductor charge density to relax, creating an interface dipole that pins the Fermi level. R. Tung, Phys. Rev. B., 64, 205310-205324 (2001). J. Hu, K. Saraswat & H.-S. P. Wong, J. Appl. Phys. 107(7) 063712, March 2010. araswat! tanford University 38 Page 19 EE311 / Shallow Junctions/MS Contacts Phenomenological Model of Pinning Schottky barrier height φn depends on charge transfer at interface •    CNL aligns to metal EF   φn = S(φm - φCNL) + (φCNL- χ) S is a dimensionless pinning factor given by •  S= d" n = d" m 1 Ne 2# 1+ $ ε is optical (electronic) portion of the dielectric constant •  !   No charge transfer, S=1 e.g. SiO2   Charge transfer, strong pinning, S=0 CNL = charge neutrality level, N surface states/m3/eV δ = state decay length = dipole layer width araswat! tanford University Robertson, JVST B18 1785 (2000) Yeo, King, and Hu, J. Appl. Phys., 15 Dec. 2002 39 EE311 / Shallow Junctions/MS Contacts Barrier Heights with Fermi Level Pinning Si  The metal work function is pinned near the charge neutrality level. Ge  Charge neutrality level is situated at ~ one-third of the band gap in silicon φbn ≈ 2Eg/3 and φbp ≈ Eg/3 Nishimura et. al, APL, 2007  Pinning in other materials depends on the location of Ecnl and its alignment with metal Fermi level 0.6 araswat! tanford University 40 Page 20 EE311 / Shallow Junctions/MS Contacts araswat! 1.5 AlSb AlS b 1.0 0.5 Si Si G Gee IInP nP G aSb G aS b InSbb InS 0.0 U n s tra in n e d B a n d  ­E d g e A lig n m e n t [e V ] Fermi Stabilization Energy GaAs G aA s  ­0.5 GaP GaP  ­1.0 EFS InAs In As AlA s AlAs A lP 5 .4 5.6 5.8 6.0 6.2 6.4 L attic e C o ns tan t [Α] The dashed grey line represents the energy at which Fermi level is pinned tanford University 41 EE311 / Shallow Junctions/MS Contacts Variation of the Schottky barrier S factor with electronic dielectric constant S •  Materials with higher ε have lower s factor and thus higher pinning. •  These materials also show lower bandgap Robertson, JVST B18 1785 (2000) araswat! tanford University 42 Page 21 EE311 / Shallow Junctions/MS Contacts Potential Solutions for S/D Engineering Sidewall •  Rdp & Rcsd Scaling (ρc ↓) Silicide Rcsd Rdp Rext Rov Rov & Rext Scaling ⇒  Dopant Profile Control: ultra-shallow highly-doped boxshaped SDE profile (e.g., laser annealing) araswat! tanford University 43 EE311 / Shallow Junctions/MS Contacts Contact Resistivity – NiSiGe Bandgap Engineering 1E-05 Contact resistivity (ohm-cm2) •  ⇒  Maximize surface doping Nif (Rsh,dp ↓): - Laser processing - Elevated S/D ⇒  Minimize ΦB: - Dual low-barrier silicide (ErSi (PtSi2) for N(P)MOS) Gate p+Si • 1E-06 – Doping Density – Ge concentration p+SiGe 1E-07 Contact Resistivity decreases with • Si0.58Ge0.42 An order of magnitude improvement is possible by using SiGe Si0.49Ge0.51 c = 10-8 ohm-cm2 is possible 1E-08 1E-5 1E-4 1E-3 1E-2 B2H6 Partial Pressure (Torr) Source: M. C. Ozturk. (NC State Univ) •  Si1-xGex S/D & germanosilicide contact -  Assuming metal Fermi level is pinned near midgap -  Similar barrier heights on n- or p-type material -  Boron solid solubility is higher in Si1-xGex than in Si -  Smaller bandgap for Si1-xGex and hence lower barrier height -  Reduction of Rcsd with single contact metal araswat! tanford University 2005 International Conference on Characterization and Metrology for ULSI Technology 44 Page 22 EE311 / Shallow Junctions/MS Contacts Fermi-level De-pinning •  Ultrathin interfacial layer can reduce or even eliminate the Fermi level pinning •  Barrier height can be changed by this method. Wave function penetration model w/o interfacial insulator Metal w/ interfacial insulator Ge Metal Ge Lower Higher Φ Beff Bond polarization model Φ Beff Ge ECNL EFS EFM Free state penetration Pinning EFS EFM Ge ECNL SiN ECNL Heine, Phys. Rev. 138, A1689 (1965). Tersoff, Phys. Rev. Lett. 52, 465 (1984). Connely et al., IEEE Trans. Nanotech. March 2004. araswat! tanford University 45 Tung, Phys. Rev. B., 64, 205310-205324 (2001). EE311 / Shallow Junctions/MS Contacts Fermi-level De-pinning in Ge and III-V III-V Ge (pinning) 0.38 eV 0.41 eV 0.51 eV 0.40 eV Y , Er Al W Ti Hu, Saraswat and Wong, J. Appl Phys., 107, 063712 (2010) Kobayashi, Kinoshita, Saraswat, Wong and Nishi., J. Appl Phys., 105, 023702 (2009) Effective barrier height versus metal work function for Schottky diodes and MIS contacts. araswat! tanford University 46 Page 23 EE311 / Shallow Junctions/MS Contacts !"#$%&'()*+',-$ Pinning vs. Tunneling 21234$()*+$ 567$8$=04-<$ #2 J "e tunnel 2m(Vo#V ) d ! 01234$()*+$ 567$8$9:;4-<$ insulator V -V o Al/Al2O3/Ge !.#$%&'%&+*/',-$ d # q$ qV " e kT e kT B J Schottky ! B Optimum barrier thickness minimizes R Depinning MIGS Tunneling ! R d This technique may be useful to obtain low contact resistance for materials with high barrier height and low dopant solid solubility araswat! tanford University (c) (b) (a) 47 EE311 / Shallow Junctions/MS Contacts !"#$%& "*$%& '"($%& Depinning in N-Ge by TiO2 I)nterfaciaLayer Metal/Ge M Al/Al O 2O (d) etal/Al/Ge3/Ge 23 !"#$%&'()*+',-$ Tunnel resistance !"#$%& EFM *"*+$%&,&*"(+$%& Lower !BN Lower !BN 21234$()*+$ 567$8$=04-<$ Metal/TiO /Ge (e) Al/TiO2/Ge 2 Lower tunnel resistance EFS *"++$%& EFM '"#$%& 01234$()*+$ 567$8$9:;4-<$ EFS *"++$%& (")$%& ! !.#$%&'%&+*/',-$ •  Ultrathin interfacial layer can reduce or even eliminate the Fermi level pinning Tunneling resistance remains low for TiO2 •  TiO2 has ~0 conduction band offset, minimizing dielectric tunneling resistance. Lower !BN, tunneling resistance small •  Both Al/TiO2 (low ΦM) and Pt/TiO2 (high ΦM) show current increase by 1000x indicating Fermi level depinning and low dielectric tunneling resistance. Tunneling resistance dominates for Al2O3 ! Lin, Roy, Nainani, Sun & Saraswat, Appl. Phys. Lett. (Accepted) araswat! tanford University ! 48 Page 24 EE311 / Shallow Junctions/MS Contacts Contact Resistance: 3D Model! Contact! I Metal! Majority carrier continuity equation outside the contact is! !" J = I I Current density in the semiconductor is ! Silicon! Silicide! #J x #J y #J z + + =0 #x #y #z J = !"E = "#v Current I Combining these two equations we obtain! ! " #!V = 0 Total current over the contact area is! •  Current flow in a contact is highly non-uniform! •  Contact resistance does not scale with area! I tot = " $ J # dA Solution of the above equations gives information about contact resistance. However, calculations are very involved.! ! W. Loh, et al., IEEE Trans. Electron Devices, August, 1986 araswat! tanford University 49 EE311 / Shallow Junctions/MS Contacts Transmission Line Contact Model ! A simplified 1D solution of the contacts is # & x I ( x ) = I1 exp % ! ( "c Rs ' $ = I1 exp( ! x lt ) lt = !c Rs lt is the characteristic length of the transmission line - the distance at which 63% of the current has transferred into the metal. " araswat! tanford University 50 Page 25 EE311 / Shallow Junctions/MS Contacts Measurement of Contact Resistance and Specific Contact Resistivity (ρc) ! Rf = Vf /I = Rs" c coth( d / lt ) w For a very large value of lt or for d << lt" Rf ! ! Re = Ve / I = "c wd Rs" c w sinh( d / lt ) •  Rf gives reasonable assessment of the source/drain contact resistance including the resistance of the semiconductor under the contact! •  Specific contact resistivity, ρc, can be calculated by measuring I, Vf or Ve! ! •  Measurement of Rf or Re is not straightforward and needs specialized test structures! araswat! tanford University 51 EE311 / Shallow Junctions/MS Contacts Test Structure to Measure Contact Resistance: Transmission Line Tap Resistor metal d semiconductor contact V24 = V f + IRSi + V f Rf! V24 Rt = = 2 R f + Rs ls w I Rs ! c R f = V f / I1 = coth (d / lt ) is a very small number w araswat! tanford University 52 Page 26 EE311 / Shallow Junctions/MS Contacts Test Structure to Measure Contact Resistance: Cross-bridge Kelvin Structure Cross-bridge Kelvin structure used to measure an average contact resistance, called RK in the figure. It gives a correct value only for dimensions of the contact << lt! araswat! tanford University 53 EE311 / Shallow Junctions/MS Contacts Error in Specific Contact Resistivity due to 1-D Modeling 1-D model Specific contact resistivity (ρc) 2-D model Contact resistance W. Loh, et al., IEEE Trans. Electron Devices, August, 1986 •  Specific contact resistivity (ρc) is a fundamental property of the interface and should be independent of contact area •  1-D models overestimate the contact resistance (Rc) as it includes parasitics •  2-D models give more accurate results and should be used araswat! tanford University 54 Page 27 EE311 / Shallow Junctions/MS Contacts Outline • Junction/contact scaling issues • Shallow junction technology • Ohmic contacts • Technology to form contacts araswat! tanford University 55 EE311 / Shallow Junctions/MS Contacts Aluminum Contacts to Si Aluminum Oxide N+ Oxide Silicon •  Silicon has high solubility in Al ~ 0.5% at 450ºC •  Silicon has high diffusivity in Al •  Si diffuses into Al. Voids form in Si which fill with Al: “Spiking” occurs. araswat! tanford University 56 Page 28 EE311 / Shallow Junctions/MS Contacts Al/Si Alloy Contacts to Si Al-Si phase diagram By adding 1-2% Si in Al to satisfy solubility requirement junction spiking is minimmized But Si precipitation can occur when cool down to room temperature ⇒ bad contacts to N+ Si araswat! tanford University 57 EE311 / Shallow Junctions/MS Contacts Silicide Contacts Barrier TiW TiN Aluminum Oxide N+ Oxide TiSi2 Contact PtSi Silicon • Silicides like PtSi, TiSi2 make excellent contacts to Si • However, they react with Al at < 400°C • A barrier like TiN or TiW prevents this reaction araswat! tanford University 58 Page 29 EE311 / Shallow Junctions/MS Contacts Silicide Contacts •  Similar methods are used for other silicides: TiSi2, NiSi, CoSi2 •  More on this in another chapter araswat! tanford University 59 EE311 / Shallow Junctions/MS Contacts Interfacial reactions ! Integrity of ohmic contacts due to a physical barrier between Al and silicide Schottky barrier reduction due to Al reaction with PtSi ΦB (eV) T (°C) araswat! tanford University 60 Page 30 EE311 / Shallow Junctions/MS Contacts Barriers Structure Al/PtSi/Si Failure Temperature (˚C) 350 Al/TiSi2/Si 400 Al/NiSi/Si 400 Al/CoSi2/Si 400 Al/Ti/PtSi/Si 450 Al/Ti30W70/PtSi/Si 500 Al/TiN/TiSi2/Si Failure Mechanism (Reaction products) 550 Compound formation (Al2Pt, Si) Diffusion (Al5Ti7Si12, Si at 550˚C) Compound formation (Al3Ni, Si) Compound formation Al9Co2, Si) Compound formation (Al3Ti) Diffusion (Al2Pt, Al12W at 500˚C) Compound formation (AlN, Al3Ti) • Silicides react with Al at T < 400°C • A barrier like TiW, TiN, TaN prevents this reaction upto T > 500°C araswat! tanford University 61 Page 31 EE311 / Shallow Junctions/MS Contacts ...
View Full Document

Ask a homework question - tutors are online