the pe-ald of ta based metals/nitrides: the growth ......stability • various phase: solid-solution...

41
Thomas J. Watson Research Center The PE The PE - - ALD of Ta Based Metals/Nitrides: The Growth, Materials ALD of Ta Based Metals/Nitrides: The Growth, Materials Properties, and Applications to Future Device Fabrications Properties, and Applications to Future Device Fabrications Hyungjun Kim and Steve M. Rossnagel Thomas J. Watson Research Center IBM Research Division Thomas J. Watson Research Center 12/12/02 TFUG/PEUG Northern California Chapter AVS

Upload: others

Post on 24-Oct-2020

1 views

Category:

Documents


0 download

TRANSCRIPT

  • Thomas J. Watson Research Center

    The PEThe PE--ALD of Ta Based Metals/Nitrides: The Growth, Materials ALD of Ta Based Metals/Nitrides: The Growth, Materials Properties, and Applications to Future Device FabricationsProperties, and Applications to Future Device Fabrications

    Hyungjun Kim and Steve M. Rossnagel

    Thomas J. Watson Research CenterIBM Research Division

    Thomas J. Watson Research Center

    12/12/02

    TFUG/PEUGNorthern California Chapter

    AVS

  • Thomas J. Watson Research Center

    Atomic Layer DepositionAtomic Layer Deposition

    ♦ ALD (Atomic Layer Deposition): Self-limiting film growth method by alternate exposure of chemical species in layer-by-layer manner.• Good conformality• Easily controlled thickness at the atomic scale• Atomic level composition control• Large area uniformity• Low impurity at low growth temperature

    ♦ For sub-100 nm node with Cu interconnect,• Ultra thin/highly conformal diffusion barrier.

    ALD is a promising technique.

  • Thomas J. Watson Research Center

    Tungsten plug for via hole

    Cu diffusion barrier/adhesion promoterCu seed layer

    Metal gate electrode Diffusion barrier

    Potential Applications of ALD Metals/Nitrides in Nanoscale DevicPotential Applications of ALD Metals/Nitrides in Nanoscale Deviceses

  • Thomas J. Watson Research Center

    Ta and TaN Thin FilmsTa and TaN Thin Films

    ♦ Ta• Good resistivity

    α phase (bcc): 15-60 µΩ cm(β phase (tetragonal): 170-210 µΩ cm)

    • High thermal stability (Tm = 2669 °C)• No intermetallic compound with Cu• Cu has very low lattice diffusivity in Ta.

    ♦ TaN• High chemical, mechanical, and thermal

    stability• Various phase: solid-solution α phase Ta,

    hexagonal Ta2N, hexagonal TaN, cubic TaN, hexagonal Ta5N6, tetragonal Ta4N5, orthorhombic Ta3N5

    Binary Alloy Phase diagrams, ASM International, 1990

  • Thomas J. Watson Research Center

    Ta and TaN layer for Cu interconnectTa and TaN layer for Cu interconnect

    ♦ TaN/Ta bilayer liner

    • Good adhesion: resisting delamination during process or thermal processing, high EM resistance

    • Liner/ILD and Cu/liner adhesion have conflicting dependencies on N content for TaNx

    – Ta adhesion on SiO2 (low k) is poor: TaN on SiO2 (low k) is require.– Cu adhesion on TaN is poor: Cu on Ta is required.

    TaN/Ta bilayer is the only viable option so far.– TiN/Ta causes corrosion for some CMP process– alpha Ta formation on TaN: low in-plane resistivity

    Ref) D. Edelstein et al, Proc. 2001 IITC, 9 (2001).

    Both Ta and TaN have good DB property. Why bilayer liner?

    Currently PVD is used, but if we want to implement ALD liner process…Need ALD process for both Ta and TaN liners!

  • Thomas J. Watson Research Center

    Ta/TaN ALDTa/TaN ALD : Precursors: Precursors

    ♦ Ta metal precursor• Halides: TaCl5, TaBr5, TaI5, all solid source• MO source: N is contained

    ♦ Ta ALD: Ta halides + reducing agents• Molecular H2: require above 700 °C

    ♦ TaN ALD• Halides + NH3 : high resistivity Ta3N5• MO sources: high C content

    TaN

    NN

    NN CH3

    CH3

    CH3

    CH3

    CH3CH3

    CH3

    CH3

    CH3

    CH3Ta N

    N

    N

    N

    C

    CH3

    CH3

    CH3

    CH3CH3

    CH3CH3

    CH3 CH3

    PDMAT TBTDET

  • Thomas J. Watson Research Center

    Strategy for Ta based metals/nitrides ALDStrategy for Ta based metals/nitrides ALD

    TaCl5

    + hydrogenplasma

    Ta

    + hydrogen/nitrogenplasma

    TaNx

    + hydrogen/nitrogenplasma

    +SiH4

    TaSiyNx

    + SiH4

    TaSi2

    OR

    + nitrogen plasma

  • Thomas J. Watson Research Center

    Experimental Setup for PEExperimental Setup for PE--ALDALD

    Quartz tubeRF coil

    200 mm wafer

    H and/or N 2 2

    Leak valve

    TaCl solid source5Ceramic heater

    Turbo pump

    Leak valve

  • Thomas J. Watson Research Center

    Growth ProcedureGrowth Procedure

    TaCl adsorption 5

    extraction

    tTaCl5

    TaCl on5 TaCl off5 H2 on

    TaCl pump out

    5

    plasma on

    tp

    plasma off

    pump out

    Gate valve off

    TaCl purgeby H

    5

    2

    1 Cycle

    TaCl5 (a) + 5H (g) → Ta(a) + 5HCl (g)↑

  • Thomas J. Watson Research Center

    PEPE--ALD Ta: SelfALD Ta: Self--limited Adsorptionlimited Adsorption

    ♦ Saturation condition: tH > 4 s, tTaCl5 > 1 s♦ Saturated growth rate: 5.3x1013 /cm2 (0.08 Å/cycle)

    0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.50

    1

    2

    3

    4

    5

    6

    7Ts = 250 °CtH= 5 s

    PE-ALD Ta

    Num

    ber o

    f Ta

    atom

    s/C

    ycle

    (101

    3 /cm

    2 )

    tTaCl5 (s)

  • Thomas J. Watson Research Center

    PEPE--ALD Ta: Plasma Exposure TimeALD Ta: Plasma Exposure Time

    0 2 4 6 80

    1

    2

    3

    4

    5

    6

    7

    8Ts = 250 °CtTaCl5

    = 2 sPE-ALD Ta

    Num

    ber o

    f Ta

    atom

    s/C

    ycle

    (101

    3 /cm

    2 )

    tH (s)

  • Thomas J. Watson Research Center

    PEPE--ALD Ta: MicrostructureALD Ta: Microstructure

    30 35 40 45 50

    x1000

    On TaN On SiO2

    (b) ALD Ta

    Inte

    nsity

    (a.u

    .)2θ (°)

    β Ta (002)

    α Ta (110)

    On TaN On SiO2

    (a) PVD Ta

    ALD Ta

    PVD Ta

    ♦ Resistivity: ρ = 150-180 µΩcm → close to β-Ta phase♦ RBS: Cl concentration = 0.5-2 at% at Ts < 300 °C

  • Thomas J. Watson Research Center

    PEPE--ALD TaNALD TaNxx : Mixture of N and H: Mixture of N and H

    10-3 10-2 10-1 1

    0.2

    0.4

    0.6

    0.8

    1.0

    1.2

    1.4

    1.6

    1.8

    tTaCl5= 2 s

    tp = 5 sTs = 300 °C

    Number of Ta atoms/Cycle

    PE-ALD TaN

    Num

    ber o

    f Ta

    atom

    s/C

    ycle

    (101

    4 /cm

    2 )

    PN2/PH2

    0.2

    0.4

    0.6

    0.8

    1.0

    1.2

    1.4

    1.6

    N/Ta

    N/T

    a ra

    tio

    ♦ Growth rate (for N/Ta = 1): 1.2x1014 /cm2 (0.24 Å/cycle)♦ Easy control of N content

  • Thomas J. Watson Research Center

    PEPE--ALD TaN: Growth TemperatureALD TaN: Growth Temperature

    100 200 300 4000

    5

    10

    15

    20

    25

    30

    tTaCl5= 2 s

    tp = 5 sPN2

    /PH2 = 0.025

    Cl H

    PE-ALD TaN

    Cl,

    H c

    onte

    nt (a

    t %)

    Ts (°C)0

    200

    400

    600

    ρ

    ρ (µ

    Ωcm

    )

  • Thomas J. Watson Research Center

    PEPE--ALD TaN: XRDALD TaN: XRD

    25 30 35 40 45 50 55 60 65

    cubic TaN

    cubic TaNTa3N5

    x2

    (e) PN2/PH2

    = 0.65

    Inte

    nsity

    (a.u

    .)

    2θ (°)

    (d) PN2/PH2

    = 0.25

    (c) PN2/PH2

    = 0.025

    (b) PN2/PH2

    = 0.004

    ALD TaN/SiO2Ts = 300 °C

    Ta2N

    (a) PN2/PH2

    = 0.001

    25 30 35 40 45 50 55 60 65

    (c) 100 °C

    Inte

    nsity

    (a.u

    .)

    2θ (°)

    (b) 200 °C

    PN2/PH2

    = 0.025ALD TaN/SiO2(a) 300 °C

    111 200 220

  • Thomas J. Watson Research Center

    PEPE--ALD TaN: TEMALD TaN: TEM

    SiO2

    TaN

  • Thomas J. Watson Research Center

    PEPE--ALD TaN: ResistivityALD TaN: Resistivity

    ♦ ρ 300 − 400 µΩ cm for N/Ta = 1

    0.2 0.4 0.6 0.8 1.0 1.2 1.4

    103

    104

    tTaCl5= 2 s

    tp = 5 sTs = 300 °CPE-ALD TaN

    ρ (µ

    Ωcm

    )

    N/Ta ratio

  • Thomas J. Watson Research Center

    PEPE--ALD TaN: MEIS (Medium Energy Ion Spectroscopy)ALD TaN: MEIS (Medium Energy Ion Spectroscopy)

    80 85 90 95 100

    102

    103

    104

    Ta

    SiON

    30 A TaNPE-ALD TaN/Si

    Ion

    yiel

    d (C

    ount

    s)

    Energy (KeV)

    TaNOx

  • Thomas J. Watson Research Center

    Comparison with Previously Reported ALD TaN Comparison with Previously Reported ALD TaN

    1) M. Ritala, P. Kalsi, D. Riihelä, K. Kukli, M. Leskelä, and J. Jokinen, Chem. Mater. 11, 1712 (1999).2) M. Juppo, M. Ritala, and M. Leskela, J. Electrochem. Soc. 147, 337 (2000).3) P, Alén, M. Juppo, M. Ritala, M. Leskelä, T. Sajavaara, and J. Keinonen, J. Mater. Res. 17, 107 (2002).4) P, Alén, M. Juppo, M. Ritala, T. Sajavaara, J. Keinonen, and M. Leskelä, J. Electrochem. Soc. 148, G566 (2001).5) J.-S. Park, H.-S. Park, and S.-W. Kang, J. Electrochem. Soc. 149, C28 (2002).

    3)0.1 – 0.35-Cl > 2C: 1 - 20> 1300TaNTaCl5a)+tBuNH2

    (allylNH2)+(NH3)

    2)0.13 – 0.31-Cl < 0.5Too highAmorphousTaCl5+DMHy

    5)0.87.9C: 15-35400 - 2000TaNTBTDET+HPlasma

    This study0.249.8Cl < 0.5350TaNTaCl5 +H+N

    1.1

    0.66 – 0.94

    0.15 – 0.2

    0.24

    Growth rate(A/cycle)

    5)3.6C: 15> 105AmorphousTBTDET+NH3

    4)-Al: 30-32Br: 4> 6600Ta(Al)N(C)TaCl5 +NH3+TMA

    1)-Cl < 4> 900TaNTaCl5a) + NH3+Zn

    1)-Cl < 1> 105Ta3N5TaCl5a) + NH3

    Thermal

    Ref.Density (g/cm3)Impurityb)

    (at %)Resistivity

    (µΩcm)PhasePrecursors

    a) Similar results for TaBr5b) Cl content at 400 °C

  • Thomas J. Watson Research Center

    PEPE--ALD TaNALD TaNxx : With nitrogen plasma: With nitrogen plasma

    ♦ N/Ta = 1.3, ρ = 105 µΩ cm: high resistivity Ta3N5 phase ♦ Low Cl (< 1%)

    30 35 40 45 50

    Orthorhombic

    PE-ALD Ta3N5

    132

    113

    023

    Inte

    nsity

    (a. u

    .)

    2θ (°)

  • Thomas J. Watson Research Center

    PEPE--ALD ALD TaSiNTaSiN

    ♦ Use SiH4 as Si source

    ♦ a: SiH4 →TaCl5 → H+N plasma• High resistivity, polycrystalline• Ta (15%) N(45%) Si(40 %)

    ♦ b: TaCl5 → SiH4 → H+N plasma• Low resistivity, amorphous film• Ta (20%) N(25%) Si(55 %)

    30 40 50

    TaCl5 → SiH4 → H+N plasma

    TaN (111)

    Inte

    nsity

    (a. u

    .)

    2θ (°)

    TaN (200)

    SiH4 → TaCl5 → H+N plasma ALD TaSiN

  • Thomas J. Watson Research Center

    Thermal ALD TaSiThermal ALD TaSi22

    ♦ Growth without plasma :TaCl5 + SiH4♦ RBS: Ta/Si = 0.5, Cl < 1%, no H

    10 20 30 40 50 60 70 80 90 1000

    Si223

    312311

    220

    203

    211112

    200

    Si

    003

    111

    110102

    101

    100

    HexagonalALD TaSi2

    Inte

    nsity

    (a.u

    .)

    2θ (°)

  • Thomas J. Watson Research Center

    ConformalityConformality

    Ta TaN

  • Thomas J. Watson Research Center

    200 cycles

    Ta TaN

    RMS: 0.206 nm

    MorphologyMorphology

    1600 cycles

    RMS: 0.123 nm

    RMS: 0.106 nm

    RMS: 0.153 nm

  • Thomas J. Watson Research Center

    Thermal Stability: PEThermal Stability: PE--ALD TaALD Ta

    200 300 400 500 600 700 800 900 10000

    10

    20

    30

    40

    50

    60 SiALD Ta/Si

    Si c

    onte

    nt (a

    t.%)

    Annealing Temp (°C)

    0

    2

    4

    6

    8

    10

    12

    14 H Cl

    Cl,

    H c

    onte

    nt (a

    t. %

    )

    ♦ 30 minutes annealing in He.♦ PVD Ta reacts with Si at 700 °C.(Ref)

    Ref) K. Holloway, P.M. Fryer, C. Cabral, Jr. J.M.E. Harper, P.J. Bailey, KH. Kellerher, JAP 71, 5433 (1992)

  • Thomas J. Watson Research Center

    Thermal Stability: PEThermal Stability: PE--ALD TaNALD TaN

    ♦ Surface O reduced by annealing. (3Å thick after 900 °C)

    80 85 90 95 100

    102

    103

    104

    Ta

    SiON

    As-dep 700 °C, 60 secs 800 °C, 60 secs 900 °C, 60 secs 1000 °C, 30 secs

    30 A TaNPE-ALD TaN/Si

    Ion

    Yie

    ld (C

    ount

    s)

    Energy (KeV)

  • Thomas J. Watson Research Center

    Cu Diffusion Barrier Property Test StructuresCu Diffusion Barrier Property Test Structures

    PVD Cu (200 nm)

    Poly Si

    Si

    ALD Ta or TaN

    RMS: 23.51 nm

    RMS: 2.126 nm

    Before annealing

    After annealing

    Cu/Ta/Poly Si

    ♦ Roughness, resistivity, phase change

  • Thomas J. Watson Research Center

    In Situ Characterization During Rapid Thermal AnnealingIn Situ Characterization During Rapid Thermal Annealing

    ♦ At National Synchrotron Light Source, Brookhaven National Laboratory.♦ 3 °C/s in He, from 100 to 1000 °C.♦ Similar system without XRD for routine measurements

    NSLS

    Beamline X20C

    Lock-in #1

    Lock-in #2HeNeLaser

    X-Ray Diffraction - Phase FormationOptical Scattering - Surface RoughnessResistance - Electrical Characterization

    Measurement and Control Electronics

    LinearX-ray Detector

    AnnealingChamber

    FocusingMirror

    MultilayerMonochromator

    Sample

    BeamlineControl

    AnalysisWorkstation

    ExperimentalControl

  • Thomas J. Watson Research Center

    200 300 400 500 6000

    0.2

    0.4

    0.6

    0.8

    1.0

    Nor

    mal

    ized

    Res

    ista

    nce

    Temperature (°C)

    0

    0.5

    1.0

    Nor

    mal

    ized

    Opt

    ical

    Inte

    nsity

    52

    50

    48

    46

    2θ (d

    eg)

    PEPE--ALD Ta: Diffusion Barrier Failure TemperatureALD Ta: Diffusion Barrier Failure Temperature

    660 680 700 720 740 760 780 8000

    0.2

    0.4

    0.6

    0.8

    1.0

    Nor

    mal

    ized

    Res

    ista

    nce

    Temperature (°C)

    0

    0.5

    1.0 Light Scattering 0.5 µm Light Scattering 5 µm

    Nor

    mal

    ized

    Opt

    ical

    Inte

    nsity

    56545250484644

    2θ (d

    eg)

    Cu/ALD Ta (7 nm)/Poly-Si

    Cu(111)Cu Silicide

    Cu(200)

    Cu silicide formation: 773 °CCu silicide formation: 260 °C

    Cu/Poly-Si

    Cu(111) Cu Silicide

  • Thomas J. Watson Research Center

    ALD Ta (2.8 nm)

    ALD Ta (14 nm)

    PVD Ta (12 nm)

    Cu(111)Cu Silicide

    Cu(200)

    Increase in failure temperature

    PEPE--ALD Ta: Comparison with PVD TaALD Ta: Comparison with PVD Ta

    Thomas J. Watson Research Center

  • Thomas J. Watson Research Center

    PEPE--ALD TaN: ALD TaN: Diffusion Barrier Failure TemperatureDiffusion Barrier Failure Temperature

    550 600 650 700 750 8000

    0.2

    0.4

    0.6

    0.8

    1.0

    Nor

    mal

    ized

    Res

    ista

    nce

    Temperature (°C)

    0

    0.5

    1.0

    Light Scattering 0.5 µm Light Scattering 5 µm

    N

    orm

    aliz

    ed O

    ptic

    al In

    tens

    ity56545250484644

    2θ (d

    eg)

    Cu 111

    Cu 200

    ALD TaN (2.5 nm, N/Ta = 1)

    Cu silicide formation: 620 °C

    Cu Silicide

  • Thomas J. Watson Research Center

    PEPE--ALD Ta and TaN: ALD Ta and TaN: Diffusion Barrier Failure TemperatureDiffusion Barrier Failure Temperature

    0 5 10 15 20 25550

    600

    650

    700

    750

    800

    850

    ALD Ta PVD Ta ALD TaN PVD TaN

    PVD Cu(200 nm)/Ta or TaN/poly-Si

    Failu

    re T

    empe

    ratu

    re (°

    C)

    Film thickness (nm)

  • Thomas J. Watson Research Center

    PEPE--ALD Ta/TaN Bilayer Cu Diffusion BarrierALD Ta/TaN Bilayer Cu Diffusion Barrier

    500 600 700 800 900

    Ligh

    t sca

    tter.

    (5 µ

    m)

    Temperature (°C)

    Ligh

    t sca

    tter.

    (0.5

    µm

    )

    PE-ALD Ta, 10 Å PE-ALD TaN, 25 Å PE-ALD Ta 10 Å/TaN 25 Å

    She

    et re

    sist

    ance

    ♦ Bilayer deposition by switching off nitrogen flow

  • Thomas J. Watson Research Center

    Metal Gate for CMOSMetal Gate for CMOS

    ♦ Problems with poly-Si gate• Gate depletion at inversion: increase in teq• High gate resistance• Reaction with high k • Boron Penetration

    ♦ Requirements • Workfunction : 0.2eV from Ec(n-FET) or Ev(p-FET)

    – nFET : Ta, Ti, TaN, TaSiN…– pFET: Ru, Pt…

    • Thermal & chemical stability with gate oxide• Low resistivity

    Metal gate

  • Thomas J. Watson Research Center

    Deposition Technique of Metal GateDeposition Technique of Metal Gate

    ♦ Little damage to gate oxide♦ Low temperature deposition with low impurity to avoid reaction with gate

    oxide♦ Conformality at nanoscale structure

    a. Conventional Device Fabrication and oxide CMP

    b. Remove Sacrificial Gate Stackc. Deposit New Gate Stack and CMP patter gate

    SpacerPoly

    Silicide

    Oxideoxidespacer

    PolySilicide

    Spacer

    Silicide

    Oxide

    spacer

    Silicideoxide

    Spacer

    Silicide

    Gate dielectric

    OxideMetalMetal

    Gate Dielectric

    spacer

    Silicideoxide

    Ref) A. Chaterjee et al, IEDM Tech. Digest, 821 (1997)

    ALD of metal gate

    replacement gate processRef)

  • Thomas J. Watson Research Center

    PEPE--ALD TaN as Metal GateALD TaN as Metal Gate

    ♦ Capacitor structure: PE-ALD TaN/100 Å SiO2♦ Work function: 4.4 ± 0.1 eV♦ Little change in tox up to 1000 °C RTA

    -2 -1 0 1 20

    200

    400

    Quasi static High frequency

    PE-ALD TaN/100 Å SiO2C

    apac

    itanc

    e (n

    F/cm

    2 )

    Voltage (V)

  • Thomas J. Watson Research Center

    Nanoscale Device FabricationNanoscale Device Fabrication

    ♦ Self-assembling diblock copolymer process• Alternative to photoresists for nanoscale dimensions

    1.surface pretreatment

    2.spin diblock copolymer and bake

    3. liquid develop

    PS PMMA

    remove PMMAsilicon

    polystyrene

    K.W. Guarini, C. T. Black, Y. Zhang, H. Kim, and I. V. Babich, 2002 3-Beams Conference, Anaheim, CA

  • Thomas J. Watson Research Center

    PEPE--ALD TaN ElectrodeALD TaN Electrode

    1159%10:1

    570%5:1

    98%1:1

    9%0.5:1

    capacitance increase

    aspect ratio

    20 nm pores, 42 nm spacing 4 nm oxide thickness

    PE-ALD TaN

    ♦Pattern transfer to Si substrate: enhanced surface area ♦Fabrication of capacitor with high storage capacity

  • Thomas J. Watson Research Center

    ConclusionsConclusions

    ♦ Successful growth of Ta based metals/nitrides at low Ts by ALD using chloride precursor and plasma.

    ♦ Ta and TaN PE-ALD processes as a function of key growth parameters

    ♦ Analysis of PE-ALD Ta and TaN thin film properties

    ♦ Applications of PE-ALD Ta based metals/nitrides in modern semiconductor devices processing

  • Thomas J. Watson Research Center

    CollaboratorsCollaborators

    ♦ Andrew Kellock♦ Cyril Cabral, Jr. ♦ Christian Lavoie♦ Christophe Detavernier♦ Dae-Gyu Park♦ James Harper♦ Kathryn Guarini♦ Matt Copel♦ Michael Lane♦ Mike Gribelyuk♦ Roy Carruthers♦ Vijay Narayanan

  • Thomas J. Watson Research Center

    Evaluation for Cu linersEvaluation for Cu liners(ref)(ref)

    Ref) D. Edelstein et al, Proc. 2001 IITC, 9 (2001).