Title: Presentazione di PowerPoint
1(No Transcript)
2Motivations
- In adatoms as a tracer of vacancy selfdiffusion
- (R. van Gastel et al. Nature (London) 408
(2000)) - (R. van Gastel et al. Phys. Rev. Lett. 86
(2001)) - In adatoms foster the layers by layer selfgrowth
on Cu(100) - (H.A. van der Vegt et al., Phys. Rev. B 51,
14806 (1995))
No ab initio calculations available
3Simulation details
- Ab initio - density functional theory (DFT)
- Electron-ions interaction US-PP
- Code VASP using PAW, supercells
- NEB for diffusion barriers
- Brillouin zone summation
- Monkhorst Pack 8x8x1
- Methfessel-Paxton, ?0.2 eV
- Slabs 3x3x545 atoms, 5x5x5125 atoms (100)
- 5x4x3 for 110-step and 6x6x3 for 100-step
- all layers (exc. bottom) relaxed, vacuum region
10 Ã…
4In adatom adsorbtion on Cu(100)
- Indium adsorbtion Hollow vs Substitutional
5In segregation
As any good surfactant, In floats (dE1.1 eV)...
Also, low Insub-Cuad exchange activation of 0.46
eV (220 K)
6Vacancy formation energy and diffusion
- Eform of a vacancy beside an embedded In atom is
0.37 eV, - i.e 0.12 eV less than in a clean terrace !!
- gtIndium-Vacancy complex
- is formed
- Extremely small barrier for
- In Vac exchange just 0.07 eV (0.24 eV via EAM,
Van Gastel et al. 2001)
- Cu Vac. barriers in proximity of an embedded
In are lower, between 0.28 eV and 0.41 eV
7In-Vac correlated motion
vac
Ediff 0.41 eV
Both formation and diffusion activation energies
are consistent with experiment (notice
improvement due to entropy)
8Vacancy generation at surface kinks
- Barrier for generation via direct expulsion of
surface Cu 1.72 eV ! ... but ....
Generation at 110-kink step
Generation at 100-step kink
Eatt0.69 eV
Eatt0.7 eV
9Indium incorporation at steps
(STM picture by R. van Gastel et al. Nature
(London) 408 (2000))
No barrier against In incorporation at kink
vacancy ! according to our ab initio calculations
10Insubst-Cuadat Interactions
Indium fosters layer by layer Cu(100) selfgrowth
(surfactant).
Why?
First step gtcheck Insubst Cuadat interactions
Indium-Adatoms interaction is slightly repulsive
0.1eV
Qualitatively similar to the interaction of Sb
(surfactant) with Ag adatoms in Sb/Ag(111)
Is the surfactance understandable similarly in
the two systems ?
11Summary
- Surface-substituted In favored over In at surface
hollow or bulk typical behaviour of a
surfactant - Cu-Insost exchange barrier is not high
(0.46 eV) gt In - segregation relatively inexpensive
-
- first-principle calculations support the
- a) vacancy-assisted In diffusion in the form of
In-V complex - b) vacancy-assisted In incorporation at steps
- c) Cu nucleation at Inads
-
- Insub Cuad interactions remind Sbsub-Agad on
Sb/Ag(111) - but smaller and shorter-range.
12Acknowledgements
Project MIUR-FIRB 01 Microscopic bases of thin
film morphology SLACS-HPC Cluster at CASPUR
Roma INFM Supercomputing at CINECA Bologna
13Indium as a surfactant
(R. van Gastel PhD thesis Leiden Univerisity
2001)
14Indium as a vacancy tracer ?
(R. van Gastel et al. Nature (London) 408 (2000))
15Cu adatom diffusion on Cu(100) by NEB
2 Diffusion processes are considered
Jump
Ediff 0.55 eV
Cu-Cu exchange
Ediff 0.95 eV via NEB
Experimental Barrier 0.40 eV !!!
16Incorporation of In atoms
- Nucleation
- Eleg-0.27 eV gt
lt Eleg-0.60 eV
- Exchange
- Eatt 0.86 eV
- Tatt gt400 K
17Insubst-Insubst interaction (eV)
-0.05
-0.025
-0.02
0.13
-0.03
18Mass transport due to vacancy diffusion in Cu(100)