Docstoc
EXCLUSIVE OFFER FOR DOCSTOC USERS
Try the all-new QuickBooks Online for FREE.  No credit card required.

Back end of Line Hard Mask Cleaning Challenges Back end of Line

Document Sample
Back end of Line Hard Mask Cleaning Challenges Back end of Line Powered By Docstoc
					                Cleaning Challenges: Back-end-of-Line
                Hard Mask Integration
                Challenges in Device Scaling TechXPOT

                Semicon® West 2008




July 15, 2008
David J. Maloney
DuPont/EKC Technology
                                                           2




Outline

1. Overview of Hardmask Interconnect Process Flow

 •   Why a hardmask? And why a metal hardmask? Tradeoffs

 •   Example Process Flow

 •   Illustration of Residues Created during etching

2. Residues’ Chemical Composition

 •   Composition variation as a function of location

 •   Change in chemical nature of residues over time

3. Approaches to Cleaning Hardmask-generated Residues

4. Selected Results
                                                                                        3


Example Hard Mask Integration Scheme

PR

 Metal Hard Mask
 Dielectric mask




 dielectric


     Barrier


Trench Mask        Hard Mask Open    Trench Mask Strip     Via Mask     Via Mask Open




  Via Etch         Strip and Clean     Trench Etch       Barrier Open    Final Clean
                                                                                   4


    Why a hardmask? And why a metal hardmask?
• 90nm     65nm      45nm     32 nm:organic etch masks struggle with selectivity
  during etch

   • Hampers ability to progress to tighter pitch

   • loses edge definition    rounding @ edges

• Hardmasks preserve patterning of tighter-pitch features

• Cleaning considerations – any change in the materials & process change the
  chemical nature of post-etch residues

Comparison of metal vs organic (PR) hardmasks
                              Metal Hardmask    Organic (PR) Hardmask
etch selectivity                    better                worse
notching/dielectric damage            ?                     ?
extendability                       better                worse
ease of residue cleaning            worse                 better
queue time problems                 worse                 better
hardmask removal                    worse                  n/a
• Some research into dissoluble hardmasks, other approaches
                                                                               5

Illustration of Residues Created During MHM Etch;
Chemical Composition of Residues




 Sidewall polymer   Inorganic particle            Bottom via residues: CuOx,
                    contaminants (TiFxOy):        CuOxFy…..
                    size of the particles grows
                    with time
                                                                                    6


         Chemical Composition of Residues (2)
                                          Residue and issues by location:

                                          1. Σ (Ti, F, O); grow and get tougher
                                          over time



                                           2. Σ (Si, C, O, F); selectivity of
                                           removal of damaged material;
                                           notching; swelling



                                           3. Σ (Si, Cu, C, O, F); selectivity of
                                           removal of damaged material;
                                           notching; swelling

Lots of different residues – demanding selectivity in the presence of sensitive
materials
                                                                                                     7




           Approaches to Cleaning Hardmask Residues


                                                                    Methodology

                     Etch chemistry
                                                                            Electrical
                           +                                                test data
                     Device materials                                       Evidence of
                                                                             Cleaning
                      Residue composition =                                (microscopy)
                      Σ (process materials, device materials)
                                                                Compatibility; thin-film metrology



Challenges include:
• taking plasma damage of dielectric into account
• accounting for different levels of dielectric damage by topography
8
                                                                                                                                        9


              Compatibility of blanket plasma damaged ULK (k~2.5)– after low-k etch only
                        Blanket layer exposed to plasma trench low-k etch
                        Clean: EKC experimental-1 (coupons in beaker)
                                                                                                0.06
                                                                                                           plasmatreated
                                                                                                                            Si-O-Si
                         FTIR
                                                                                                           1minEKC1
                                                                                                0.05       2minEKC1

             0.06                                                                               0.04
                                plasmatreated                           Si-O-Si




                                                                                   Absorbance
                                1minEKC1                                                        0.03

             0.05               2minEKC1                                                        0.02


                                                                                                0.01
             0.04
                                                                                                0.00
Absorbance




                                                                                                -0.01
             0.03                                                                                   1500                   1000         500
                                                                                                                    wavenumber (cm-1)


             0.02                                                Si-CH3
                                                                                   Si-H                      No minor
             0.01
                                   Porogens-
                                   CH2-CH3            Si-H
                                                                                                             absorbance
                                                                                                             changes observed
             0.00                                                                                            in the Si-O-Si peak

             -0.01
                 4000       3500     3000      2500     2000     1500       1000                 500
                                            wavenumber (cm-1)
                                                                                                                                                                                           10

                           Compatibility of blanket plasma damaged ULK (k~2.5) – after low-k etch only
Blanket layer exposed to plasma of trench low-k etch                                                                                        Clean: EKC experimental-1 (coupons in beaker)

                                            Thickness                                                                                                    K-value
                                Thickness pristine                Refractive Index pristine                                                         Aurora ELK HM
                                Thickness plasma treated          Refractive Index plasma treated                                 2.75
                                                                                                                                                                 pristine
                                Aurora ELK HM                                                                                                                    Plasma treated
                         -1.5                                                        0.016
                                                                                                                                   2.7

                                                                                     0.014
                          -2




                                                                                              Refractive Index change
 Thickness change (nm)




                                                                                                                                  2.65




                                                                                                                        k-value
                                                                                     0.012
                         -2.5
                                                                                                                                   2.6
                                                                                     0.01

                          -3
                                                                                     0.008                                                                                          Removal of
                                                                                                                                  2.55
                                                                                                                                                                                    damaged
                         -3.5                                                                                                                                                       low-k
                                                                                     0.006
                                                                                                                                   2.5
                                                                                                                                     -0.5     0    0.5     1      1.5    2        2.5
                          -4                                                         0.004
                            0.5            1             1.5          2           2.5                                                                dip time (min)
                                                     dip time (min)
                                                                                                                                            K-value changes < 0.1 are not significant
After trench plasma etch and 2 min 45°C EKC clean:
                                                                                                                                             Study on blanket layers show
• ULK loss of around 2 nm, no change in RI
                                                                                                                                             EKC chemistry to be compatible
• k-value change ~ 0.16                                                                                                                      with ULK (k~2.5)
                                                                                                           11

  Electrical evaluation of EKC experimental-1 as post-etch clean in dual damascene
  structures
Via etch/ash/Trench low-k etch                                               DD 90nm ½ pitch ULK (k = 2.5)
Clean: 1 and 2 min 45°C EKC experimental-1 on single                   FIB-TEM 100 nm via chains
wafer spin tool
          100nm Kelvin structures                                                              Cu
single via contact which is sensitive to organic contam.
removal and low-k etching (via diameter)                                                              Low-k
the lower the contact resistance, the better the clean
                                                                      Cu Low-k                      SiCN/SiCO



                                                              No clean                 After strip/etch +
                                  POR
                                                                                       2 min EKC @45°C
                                                   (1)          Full Via etch &
                                                                      ash

                                           EKC
                                           (1&2min)
      (2)                                                                         Damaged region
                                                                                  after via etch/ash

                                                               Good yield on isolated via at cost of CD
                                                           loss
(1) Probably from more uniform clean;
                                                               Profile distortion of via;
     POR clean performed on prototype tool                 Ash & etch damage + etching of damaged LK
(2) contact clean with POR better at wafer center          by clean (polymer removal by undercutting)
                                                              Clean is too aggressive
                                                                                    12


 Compatibility of patterned plasma damaged ULK (k~2.5) after trench etch
Trench low-k etch
Clean: 2min 45°C EKC experimental-1 on single     DD 90nm ½ pitch ULK (k = 2.5)
wafer spin tool                                 FIB-TEM Kelvin structure
       Damaged low-k by trench etch
Metal 2
                                                               trench
     trench                       100 nm
                       Low-k                             via   125 nm      105 nm
                                  125 nm
           via
Metal 1



    Thickness loss blanket plasma
    treated layer : ~ 2 nm


• TEM images reveal low-k loss of ~ 20nm (trench level) for plasma
damaged patterned layer     possible MHM impact on compatibility?
• Plasma damaged blanket layer not representative for patterned layer?
   check influence of ash plasma in previous step
                                                                                                                                                           13

                   Compatibility of blanket plasma damaged ULK– after ash and low-k etch
         FTIR              Blanket layer exposed to plasma of ash and low-k etch
                           Clean: EKC experimental-1 (coupons in beaker)                                                 Si-O-Si
                                                                                                           0.06                               plasma
                                                                                                                                              1 min 30C EKC-1
                                                                                                           0.05                               2 min 30C EKC-1
                                                                                                                                              1 min 45C EKC-1
                                                                                                           0.04
                                                                                                                                              2 min 45C EKC-1
                                    plasma




                                                                                        intensity (a.u.)
                                                                           Si-O-Si
                   0.06             1 min 30C EKC-1                                                        0.03


                                    2 min 30C EKC-1                                                        0.02


                   0.05             1 min 45C EKC-1                                                        0.01


                                    2 min 45C EKC-1                                                        0.00


                   0.04                                                                                    -0.01
intensity (a.u.)




                                                                                                               1500             1000                            500
                                                                                                                          wavenumber (cm-1)
                   0.03


                   0.02
                                                                    Si-CH3
                                                                                     Si-H                             Shift observed in
                   0.01                Porogens-
                                       CH2-CH3            Si-H
                                                                                                                      the Si-O-Si peak
                   0.00


                   -0.01

                       4000     3500     3000      2500      2000   1500      1000            500
                                                wavenumber (cm-1)
                                                                                                                                                                                                                                    14

                                       Compatibility of blanket plasma damaged ULK – after ash and low-k etch
                                                                Blanket layer exposed to plasma of ash and low-k etch
Ellipsometry
                                                                Clean: EKC experimental-1 (coupons in beaker)

                                           Thickness change                                                                                                     RI change
                                                                                                                                                       0.001                             30C_pristine   0.0035
                                      10                                           25
                                                                                                                                                                                         45C_pristine
                                       8                                                                                                              0.0005                             30C_plasma
                                                                                                                                                                                                        0.003
 thickness chan ge (nm ) _pristin e




                                                                                                                                                                                         45C-plasma




                                                                                        thickn ess chang e (nm )_p lasm a
                                       6                                           20                                                                      0
                                           30C_pristine                                                                                                                                                 0.0025




                                                                                                                                RI change _pristine
                                       4   45C_pristine                                                                                               -0.0005




                                                                                                                                                                                                                 RI change_plasma
                                           30C_plasma                              15
                                       2                                                                                                               -0.001                                           0.002
                                           45C-plasma
                                       0                                                                                                              -0.0015                                           0.0015
                                                                                   10
                                      -2                                                                                                               -0.002
                                                                                                                                                                                                        0.001
                                      -4                                                                                                              -0.0025
                                                                                   5
                                                                                                                                                                                                        0.0005
                                      -6                                                                                                               -0.003

                                      -8                                           0                                                                  -0.0035                                           0
                                                    1                      2                                                                                    1                    2
                                                                                                                                                                    dip time (min)
                                                          dip time (min)

                                                                                                                            PERR clean after trench etch
2 min EKC @ 45C clean after ash and
                                                                                                                                                                         Damage at trench level
trench etch  20 nm ULK loss
                                                                                                                                                                         higher than expected
Idem patterned DD wafers                                                                                                                                                 because ash plasma
                                                                                                                                                                         after via etch also has
                                                                                                                                                                         impact
                                                                                      15


                       Future work
                Chemistry screening on blanket and patterned wafers to understand
                etch/ash/clean interactions wrt ULK damaging and etching


                       • Blanket ULK layer (k~2.5)
                             - EKC chemistry blends
Assess compatibility




                             - look at impact ash plasma and ash/etch plasma
                             - look at process time (coupons in beaker setup)
                             - FTIR, ellipsometry and Hg-probe


                       • Checkerboard patterned layers: ULK layer (k~2.5) + TiN MHM
                             — EKC chemistry blends
                             - look at impact ash plasma and ash/etch plasma
                             - does presence of MHM influence compatibility?
                             - look at process time (coupons in beaker setup)
                             - FTIR, ellipsometry and Hg-probe
                                                                                                                   16


     Future work
       Chemistry screening on patterned wafers
                 •   Polymers not visible directly after etch/ash but are present (only visible after dHF clean)

                        etch/dry strip                          etch/dry strip + HF




                                                                                                  Polymer veils
Assess removal




                 • Create more polymers for faster chemistry screening
   efficiency




                       • Test vehicle: SD 90 nm ½ pitch ULK (k~2.5)
                       • EKC chemistry blends
                       • Detuned etch recipe : create more polymers to check removal more
                         efficiently
                       • look at clean process time (coupons in beaker setup)
                       • SEM inspection
17
                                                                                         18




Conclusions

• It is very likely that the use of metal hardmasks will be required
below 45nm
• The associated cleaning challenges are significant, but not
insurmountable
• For the work illustrated here, some notable points:
  • Dry strip and wet clean combination drives ‘patterning performance’
  • Use of EKC experimental-1 for PERR in dual damascene partial trench first with ULK
    (k~2.5) results in good yield on isolated via but at cost of CD loss
  • Cleaning products appear to remove damaged ULK at a very high rate – but not attack
    pristine ULK  high selectivity of removal towards damaged dielectric


Acknowledgments

Hua Cui (EKC)
Guy Vereecke, Martine Klaes (IMEC)

				
DOCUMENT INFO
Shared By:
Categories:
Stats:
views:28
posted:6/30/2010
language:English
pages:19
Description: Back end of Line Hard Mask Cleaning Challenges Back end of Line