Sulfur's Effect on Cu111 Surface Morphology

1
Sulfur's Effect on Cu(111) Surface Morphology
J. de la Figuera K. Pohl M.C. Bartelt N.C.
Bartelt P.J. Feibelman and Robert Q. Hwang
Sandia National Laboratories
Funding DOE-AC04-94AL85000 MEC/Spain PB96-0652
2
Sulfur on Cu(111)

• Learn about the evolution of nanoscale morphology
  of surfaces under gas exposure
• Cu(111)
• benchmark system for morphology evolution. Well
  studied in the clean case.
• Interest in smaller and smaller features for the
  Cu-interconnect technologies
• Sulfur is a common contaminant!

3
Cu/Ru(0001) Cu(111) Mesas
2 mm
500nm

• Up to 100ML of Cu on Ru(0001)
• Annealed to 650 C
• Cu/Ru(0001) grows layer-by-layer only for lt3ML
• Add additional 1-5ML to form mounds

4
Morphology changes rapidly on Cu mesas upon S
exposure
20ML
1 mm x 1 mm
19min

• Islands adopt a triangular shape
• Mounds flatten FAST!
• Coarsening between same level terraces!

43min
Critical S exposure 0.1L, rate 10-10 torr
5
Hexagonal to Triangular Shape Change
Compact steps in fcc(111) surfaces Step type A
(100) microfacets Step type B
(111) microfacets
6
A
A
B
B
A-steps recede faster upon S exposure
7
Which side disappears faster?
We need to find the orientation of the Cu(111)
mesa
The hcp substrate prevents determination of the
orientation by any means other than a local probe
like STM...
Upper Cu(111) terrace
B step
Lower Cu terrace
8
Aspect ratio for clean Cu(111)
The observed ratio between both types of steps in
clean Cu(111) islands is close to 1

• Ab-initio GGA calculations give
• EA 0.27eV/atom
• EB 0.28eV/atom
• From that ratio, the equilibrium shape (following
  Wulff construction) is
• LB/LA1.14

9
Sulfur on Cu(111) steps

• From ab-initio calculations
• S prefers to bind to A-type steps

Same S-Cu tetramer unit found on
and Cu(111)
On Cu(100)
1st layer S
2nd layer S
Cu
M. Foss et al, Surf. Sci. 388,5 (1997)
M.L. Colaianni et al, PRB 50,8798 (1994)
10
Sulfur atoms appear in (2x1) geometry at steps

• GGA calculations using the (2x1) S position at
  steps,
• EA 0.11eV/atom (preferred, fourfold hollow-like
  sites)
• EB 0.15eV/atom
• From that ratio, the equilibrium shape is
• LA/LB2.4

B
A
11
Analysis of STM image sequences showing island
decay

• Decay rate 103 atoms/sec
• Area decays linearly
• Attachment-detachment limited?
• The sulfur effects begin upon exposures of 0.1L

12
A decrease in the Schwoebel barrier?
On clean Cu(111), EES.22eV (M. Giesen and H.
Ibach, Surf. Sci. 431 (1999) 109)

• Although the Schwoebel barrier may be modified,
  it does not explain the fast coarsening on the
  same terrace
• In experiments where the Schwoebel barrier is
  assumed to disappear, the island decay is still
  gt102 slower than with S

Decrease in the Schwoebel barrier is NOT enough!
M. Giesen et al., Phys. Rev. Lett. 80 (1998) 552.
13
Is Cu atom detachment from islands irreversible?
t15min
06_25_99/m110
06_25_99/m124
t21min
t19min
06_25_99/m127
06_25_99/m129

• On the same level, some islands ripen
• The detachment is NOT irreversible
• But it may be modified by the presence of sulfur
  at the island steps and in the adatom sea

14
Modified reattachment to step edges?

• S may inhibit reattachment

A
B
15
Problems in Cu/Ru(0001) vs Cu(111)

• Higher density of dislocations than in bulk Cu
  crystals due to the network of misfit
  dislocations at the Cu-Ru interface

Dissociated edge threading dislocation emerging
at the mesa surface

• The reactivity towards sulfur of the areas
  between mesas (mainly 2ML Cu) is different than
  Cu(111)

Decay behavior of mesas depends on a large
neighbor-hood
10ML
t135min
750nm
16
Summary

• We have observed the evolution of Cu(111)
  structures grown on Ru(0001)
• Sulfur induces
• Fast decay of mounds!
• Fast ripening!
• Our observations of coarsening on the same
  terrace ruled out that the observations could be
  explained
• By a lowering of the Schwoebel barrier
• By irreversible detachment of Cu from islands
• We expect these features to reflect the
  modification of the reattachment at island step
  edges