3.155J6.152J Physical Vapor Deposition(PVD) SPUTTER DEPOSITION ◆. PECVD lasma enhanced surface diffusion without need for ◆… Dry etching Momentum transfer from plasma to remove surface species e We will see evaporation: Evaporate source material, Peg, vap P. s 10- Torr (another PVD) Poor step coverage, alloy fractionation: A P vapor Now sputter deposition. Noble or reactive gas P=10 mTorr What is a plasma? 6.1523.155 Nov.5.2003 What is a plasma? A gas of ionized particles, conducting at low freq Initiate ionization(breakdown) with spark Voc>>Vbkdn ions(e.g. Ar*+e) ions vnkdn P≈10-100 m Torr Ionizatio cathode What goes on inside a plasma?
3.155J/6.152J 6.152J/3.155J Nov. 5, 2003 1 Physical Vapor Deposition (PVD): SPUTTER DEPOSITION We will see evaporation: (another PVD) Evaporate source material, Peq.vap.Pg £10-6 Torr Poor step coverage, alloy fractionation: D Pvapor u u Now sputter deposition. Noble or reactive gas P § 10 mTorr We saw CVD Gas phase reactants: Pg § 1 mTorr to 1 atm. Good step coverage, T > > RT u …PECVD Plasma enhanced surface diffusion without need for elevated T u u …Dry etching Momentum transfer from plasma to remove surface species What is a plasma? 6.152J/3.155J Nov. 5, 2003 2 Initiate ionization (breakdown) with spark VDC > > Vbkdnfi ions (e.g. Ar+ + e- ) ions Vbkdn What goes on inside a plasma? P § 10-100 m Torr cathode anode V § 1 kV - Ex ve - vAr+ Ionization event What is a plasma? A gas of ionized particles, conducting at low freq
3.155J6.152J What is molecular density at 10 mT? Ideal gas: n=P/(hBT) 2.5x1025m3 LogIn(#/m)I I Atm= 0.1 MPa n=3.2x1030m I Torr :0 mi Spacing between molecules =l/5=0.15 microns. Is this=2? 6.1523.155 Nov.5.2003 Spacing between molecules rl/=0.15 microns. (Ar’ much less 0.1% to 1% of A, are ions: E field accelerates Ar e between collisions Eq WoLEy 2×10° 6.152J.155J Nov.5.2003
3.155J/6.152J 6.152J/3.155J Nov. 5, 2003 3 012345 Log[P (N/m2 )] 25 24 Log[n (#/m3)] 23 22 21 20 1 Atm= 0.1 MPa §14 lb/in2 2.5 x 1025 m-3 1 Torr 10 mT n = 3.2 x 1020 m-3 Ideal gas: n = P/(kBT) Spacing between molecules § n-1/3 = 0.15 microns. Is this = l ? What is molecular density at 10 mT? 6.152J/3.155J Nov. 5, 2003 4 l l l l l l l l l l l 0.1% to 1% of nAr are ions: E field accelerates Ar+, e- between collisions. v f 2 = v0 2 + 2ax ª 2 Eq m l lAr ª 3 cm ( l[Ar + ] much less) l = kBT 2pd2 P Spacing between molecules § n-1/3 = 0.15 microns. Is this = l ? No! ve- ª 2 ¥109 cms v , Ar ª107 cms J = nqvx cathode anode V § 1 kV - Ex ve - vAr+ Ionization event And ne- >> nAr+
3.155J6.152J Which species, e-or Ar+ is more likely to dislodge an atom at electrode P≈10-100 m Torr cathode Cathod source material Ar--- V≈lkv (1/2)MxVA2=916m2V A=/43 =(1/2)mv2=(1/2)mv2 PA=MV=1832mv/43 Momentum transfer? No surprise From ion implantation, most energy transfer when: AE=E, 4M, A i.e. incoming particle has mass close to that of target. 6.1523.155 Nov.5.2003 Sputtering process Ar* impact, momentum transfer at cathode e avalanche and released target atoms, ions Atomic billiards Elastic energy transfe 4M.M st for m E,4 For e- hitting anode, substrate, M,<<M But e" can give up all its Er in inelastic collision mv→△U Excitation of atom or ion 6.152J.155J
3.155J/6.152J 6.152J/3.155J Nov. 5, 2003 5 Which species, e- or Ar+ , is more likely to dislodge an atom at electrode ? P § 10-100 m Torr cathode anode V § 1 kV - Ex ve - vAr+ Ekin § 1 keV = = (1/2) MArVAr2 = 916 meVAr2 = (1/2) merve 2 = (1/2) meve 2 \ VAr = ve/43 Momentum transfer? PAr = MV = 1832mv/43 pe = mv No surprise. From ion implantation, most energy transfer when: i.e. incoming particle has mass close to that of target. DE = E1 4M1M2 ( ) M1 + M2 2 P § 10-100 m Torr cathode anode V § 1 kV - Ex ve - vAr+ Cathode is “target”, source material 6.152J/3.155J Nov. 5, 2003 6 Sputtering process Ar+ impact, momentum transfer at cathode fi e- avalanche and released target atoms, ions. For e- hitting anode, substrate, M1 < < M2 E2 E1 ª 4M1 M2 (small) But e- can give up all its EK in inelastic collision: 1 2 me ve 2 fi DU Excitation of atom or ion Elastic energy transfer E2 E1 µ 4M1M2 ( ) M1 + M2 2 cos2 q E2 greatest for M1 q @ M2 E1 E2 Atomic billiards
3.155J6.152J Sputtering process P≈10-100 m Torr cathode Cathode is" target” V≈1kV cathode e Momentum transfer of Ar+ on cathode erodes cathode atoms source atoms Target material(cathode s flux to anode substrate must be conductive or use RF sputtering (later) 6.1523.155 Nov.5.2003 Inside a plasma Cathode e Art impact on cathode electrons ② 1.5eV Faraday dark space ①Ek ②1.5≤Eks3eV③ e. induced no action optical excitation ofAr→ visible glow 6.152J.155J Nov.5.2003
3.155J/6.152J 6.152J/3.155J Nov. 5, 2003 7 P § 10-100 m Torr cathode anode V § 1 kV - Ex ve- vAr+ Cathode is “target”, source material cathode anode V § 1 kV - Mostly-neutral source atoms Target material (cathode) must be conductive …or use RF sputtering (later) Momentum transfer of Ar+ on cathode erodes cathode atoms fi flux to anode, substrate. Sputtering process Ar+ 6.152J/3.155J Nov. 5, 2003 8 ve - vAr+ Ex Cathode Anode x Ar+ impact on cathode fimostly electrons ¿ ¡ ¬ x 1.5eV EK ~ ve - 2 3 eV Ionization glow v f 2 = 2ax ¿ EK e - 3eV fi ionization, High conductivity plasma Faraday dark space
3.155J6.152J Inside a plasma 3ev (D.C. or cathode sputtering 1. 5eV ionization Cathode relative to anode Film species Cathode sheath Substrate ow ion density High conductivity How plasma results in deposition 5)Deposited at anode 1)Ar+ accelerated to cathode cathode ano 2)Neutral kicked off Ar =Ar++e 3) some neutral⊙ V≈1kV (e.g.O→0) 6)Some physical of al by Ar
3.155J/6.152J 6.152J/3.155J Nov. 5, 2003 9 ¿ ¡ ¬ x 1.5eV EK ~ ve - 2 3 eV Ionization glow Je- ,ve- >> JAr+ ,vAr + fi \ plasma § 10 V positive relative to anode cathode anode - Inside a plasma (D.C. or cathode sputtering) Cathode sheath: low ion density Plasma High conductivity Cathode Anode e- Ar+ + VD.C. Film species Target Substrate 6.152J/3.155J Nov. 5, 2003 10 cathode anode V § 1 kV - Al Ar Al + Ar+ + Ar+ 1) Ar+ accelerated to cathode Al Ar =Ar+ + ee- 2) Neutral target species (Al) kicked off; 3) some neutral Ar and e- also. 4) e- may ionize impurities (e.g. O => O- ) O- 5) Deposited at anode: Al, some Ar, some impurities 6) Some physical resputtering of Al by Ar Al How plasma results in deposition
3.155J6.152J Sputtering rate of source material in target is key. Typically 0.1-3 atoms/Ar Sputtering rates vary little from material to material. Vapor pressure or source NOT important (this differs greatly for different materials cathode anode kicked off CAr=Ar++e V≈1kV G4 6Some physica uttering of Al by Ar 6.152J3.155J Nov.5.2003 Sputtering vield atoms molecules from target #f incident ions E, En Oboo Oooo Td2 # area excited in Random walk each laver Oring, Fig. 3. 18. Table 3-4 Sputter rate depends on angle of incidence, relative masses, kinetic energy
3.155J/6.152J 6.152J/3.155J Nov. 5, 2003 11 cathode anode - V § 1 kV Al Ar Al + Ar+ + Ar+ Al Ar =Ar+ + e- 2) Neutral target species (Al) kicked off. 6) Some physical Al resputtering of Al by Ar Sputtering rate of source material in target is key. Typically 0.1 - 3 atoms/Ar Sputtering rates vary little from material to material. Vapor pressure or source NOT important (this differs greatly for different materials). cos q 6.152J/3.155J Nov. 5, 2003 12 S = Sputtering yield = # atoms, molecules from target # incident ions S =s 0nA E2 4 Ethresh ¥ ... ln Ev Eb Ê Ë Á ˆ ¯ ˜ È Î Í ˘ ˚ ˙ # excited in each layer p d 2 # / area Random walk to surface; Eb = binding energy Sputtering yield q S 90° Oring, Fig. 3.18, Table 3-4 q Sputter rate depends on angle of incidence, relative masses, kinetic energy
3.155J6.152J Sputtering Yield Data for Metals (atoms/ion) Sputtering Gas He Ne Ar Kr XeArThreshold Energy(kev) 0.5 0.50.5 0.5 0.5 1.0 Voltage (ev) 0201.773 273.323.8 0160.731.050.960.8210 0071082.4013.0630136 0240420.510.48035 007-0.120.130.17 0.130.901.22108108 25 0241.802.352.352052 0.150.881.101.071001 080681.11.12104 25=E 4M, M2 M,+M, 0030480.800.8708711 061.101451.3012222 0030631401.82193 01304805005004206 26 007043051048043 6.152J3.155J Nov.5.2003
3.155J/6.152J 6.152J/3.155J Nov. 5, 2003 13 Aargon = 40 amu ATi = 48 amu DE = E1 4M1M2 ( ) M1 + M2 2
3.155J6.152J Sputtering miscellany cos 8= normal component cos"0: more-narrowly directed at surface rge target small substrate = good step coverage more uniform thick Moving substrates Higher pressure shorterλ 万x2P p 2(cm) better step cove 10mT2 1mT20 6.152J3.155J Nov.5.2003 Target composition vs film composition Sputtering removes outer layer of target =>more uniform (problem with multicomponent system only initially) Initial target composition A, B(e. g. Si, W) If sputter yield A>B omposition of target B Co-sputtering = composition control A2B AB 6.152J3.155J Nov.5.200
3.155J/6.152J 6.152J/3.155J Nov. 5, 2003 15 Sputtering miscellany cos q = normal component of flux Large target, small substrate => good step coverage, more uniform thickness Moving substrates cosn q: more-narrowly directed at surface Higher pressure => shorter l, better step coverage l = kBT 2pd2 P p l (cm) 10 mT 2 1 mT 20 6.152J/3.155J Nov. 5, 2003 16 Sputtering removes outer layer of target => more uniform composition (problem with multicomponent system only initially) Target composition vs. film composition A3B A2.5B A2B If sputter yield A > B time Surface composition of target A B Initial target composition A2B (e.g.Si2W) Co-sputtering => composition control, sample library A2B AB AB2 A B
3.155J6.152J Varieties of sputtering exper D.C. sputter deposition: Only for conducting materials if DC sputtering were used for insulator. e. g. carbon, charge would accumulate at each electode and quench plasma within 1-10 mico-sec I micro sec Ar =Ar++" e Therefore, use RF plasma 6.152J3.155J Nov.5.2003 ◆ RF plasma sputtering Plasma conducts at lowf 0-10 Target Plasma potential still >0 Substrate due to high e" velocity. Potential now symmetric Mobility of target species still sufficient to = deposition in 1/2 cycle 7×104m M Smaller target=> higher field Target A F.C.C. reserves: 13.56 MHz for sputtering
3.155J/6.152J 6.152J/3.155J Nov. 5, 2003 17 D.C. sputter deposition: Only for conducting materials. if DC sputtering were used for insulator. e.g. carbon, charge would accumulate at each electode and quench plasma within 1 - 10 mico-sec. Varieties of sputtering experience + Ar+ cathode anode V § 1 kV - C Ar t 0 due to high e- velocity. Potential now symmetric. Mobility of target species still sufficient to => deposition in 1/2 cycle: v £ 2eV M ª 7 ¥104 m /s Target Substrate Vplasma Smaller target => higher field + Ar+ Ar+ + Some re-sputtering of wafer V1 V2 = A2 A1 Ê Ë Á ˆ ¯ ˜ m m § 1,2 Plasma conducts at low f Thus, V = 0 in plasma w p ~107 s -1 F.C.C. reserves: 13.56 MHz for sputtering
3.155J6.152J Varieties of sputtering exper T Negative wafer bias enhances re-sputtering of film -I kV of Ta films vs esults hold for for process Used in SiO,: denser TARGET ewer Bias affects stress dielectric constant SUBSTRATE BIAS ( VOLTS) Nov.5.2003 Film stress in rF 匚Inc. pressure sputter deposition 68.8 0008 hode 等搭藏信北 energetic Pressure Better step coverage
3.155J/6.152J 6.152J/3.155J Nov. 5, 2003 19 u Bias sputtering Negative wafer bias enhances re-sputtering of film Varieties of sputtering experience Target Substrate Vplasma + Ar+ Vbias § -100 V Ar+ + Some re-sputtering of wafer Bias is one more handle for process control. Used in SiO2: denser, fewer asperities. Bias affects stress, resistivity, density, dielectric constant… Resitivity of Ta films vs. substrate bias voltage. Similar results hold for W, Ni. Au, Cr. 300 nm 160 nm V § -1 kV 6.152J/3.155J Nov. 5, 2003 20 Cathode -Vbias Film stress in RF sputter deposition Film Species Better step coverage Ar+