Misfit Dislocation Networks: a possible system for the growth of nanostructures

1
Misfit Dislocation Networks a possible system
for the growth of nanostructures
Juan de la Figuera Oliver Schaff Andreas K.
Schmid Norm C. Bartelt Robert Q. Hwang
Sandia National Laboratories California, USA
Funding DOE-AC04-94AL85000
2
Self-Organized Growth

• Goals
• To grow ordered arrays of nanometer-sized objects
• With the ability to tailor the size, separation
  and lattice type to different applications
• Make devices grow themselves

3
Misfit Dislocations Networks
Au(111)
50nm
3ML Cu/Ru(0001)
on (111) surfaces
4
Classic Example Fe,Co,Ni on Au(111)
Grown at RT
Au(111),O. Schaff, SNL
Co/Au(111), 343nm x 373nm J.V. Lauritsen et al.,
J. of Cat. 197 (2001) 1
D. Chambliss et al, JVSTB 9, 933 (1991)
5
Building Blocks of the Dislocation Patterns
31nm x 29nm
R.Q. Hwang and C.B. Carter, PRB 51 (1995) 4730
6
Misfit Shockley Partial Dislocations
1ML Au/Ru(0001)
6.6 misfit
1ML Cu/Ru(0001)
2ML Cu/Ru(0001)
HCP
FCC
HCP
-5.5 misfit
7
Threading Dislocations Meeting point of Shockley
Partials
8
Structure of the Au(111) Dislocation Network
4.4 misfit between last layer and bulk
in-plane lattice parameter
250nm
9
Threading Dislocations move by exchanging adatoms
in Au(111)
STM at 80 C
Inserting one atom here moves dislocation core
laterally by one lattice constant
Theory
10
Quantitative Analysis of the STM Data
Analyse STM data line by line and determine
position of cores
Map out trajectories of single cores with time
Measure the distances of cores with time
11
What is the barrier for exchange?
Measure mean residence time of the displacement
between two dislocations
At 80 C, distributions of times is exponential
with frequency of 0.4s
Number consistent with dilute Au lattice gas
randomly exchanging with cores -- sum of
formation energy and diffusion barrier is 1.2eV.
1
2
(sec)
12
Quantitative Model for Distortions
What is the energy cost of displacing cores
with respect to each other? It is a nearly
harmonic potential with spring constant k
As obtained from calculations based on a
Frenkel-Kontorova model of the film
x
k 3.0 meV/Å2
Can we measure this spring constant
experimentally?
13
Determination of the spring constant
Experimental result
Probability distribution of core distances gives
dislocation spring constant
k 2.53 0.21 meV/Å2
Compare with theoretical result
k 3.0 meV/Å2
14
How nucleation of islands takes place on surfaces
1
Homogenous nucleation
2
3
1
Exchange induced inhomogeneous nucleation
3
4
2
15
How to grow islands into ordered patterns
Confine the adatoms
Provide inhomogeneous (ordered) nucleation points
16
Possible Effects of Shockley Partials on Adatoms

• In-plane atomic arrangement over partial
  dislocations is nearly perfectly hexagonal
  (exchange is difficult).
• The atoms close to the dislocations are strained.
• Different diffusion barrier!
• If the temperature is low enough, they are able
  to confine diffusing adatoms.

H. Brune at al, PRB 52 (1995) R14380
B. Fischer et al, PRL 82 (1999) 1732
17
Possible Effects of Threading Dislocations on
Adatoms

• Acting directly as nucleation points (unproven)
• Enhanced Exchange for diffusing adatoms which
  once embebbed serve as nucleation spots
• Proposed mechanism for Ni/Au(111)
• Enhanced exchange directly observed in 1ML
  Cu/Ru(0001) and Au(111)

J. A.Meyer et al, Surf. Sci. 365 (1996) L647
A.K. Schmid et al. Phys. Rev. Lett. 78, 3507
(1997), O. Schaff et al, in preparation
18
Variable Temperature STM
LN2 cooling with free-hanging supporting mass
110K to 450K (so far)
19
1.X ML Ag/Ru(0001) Misfit Dislocation Network
20
0.X ML Ag on Ru(0001) Misfit Dislocation Network
Short period herringbone
R.Q. Hwang et al, PRL 75 (1995) 4242
21
Growth on thick Ag on Ru(0001)
110K
Pseudo Diffusion Limited Growth, with random
nucleation
Long mean free path for adatoms (compact surface)
150nm x 150nm
22
Ag on 1ML Ag/Ru(0001)
T110K
200nm x 200nm
23
Ag on 0.X ML Ag/Ru(0001)
Disordered, bimodal distribution
110nm x 120nm
24
Co on Ag/Ru(0001)
170nm x 140nm
T110K
25
Cu on Ag/Ru(0001)
T110K
600nm x 430nm
26
Why double islands?Threading Dislocations can
dissociate
J. de la Figuera et al, submitted to PRL
27
Annealed S2ML Cu/Ru(0001)
200nm x 140nm
60nm x 30nm
J. de la Figuera, Surf. Sci. 433-435 (1999) 93 J.
Hrbek et al, J. of Phys. Chem. B 103 (1999) 10557
28
Cu on S2ML Cu/Ru(0001)
Cu 140nm x 120nm
T110K
29
Co,Ag on SCu/Ru(0001)
Ag 120nm x 120nm
Co 110nm x 70nm
30
Summary Au(111)

• 80C temperature measurements (with a conventional
  STM!)
• Direct observation of dislocation climb through
  exchange with adatoms
• Measurement of the energy barrier involved.
• Measurement of the interaction between threading.
  dislocations and comparison with theoretical
  estimates.

31
Summary growth on dislocation networks

• Results so far
• Several metals (Co,Cu,Ag) grow on the misfit
  dislocation networks employed by nucleating
  preferentially on top of the preexisting
  threading dislocations.
• Ordered patterns where found with Co,Cu on
  Ag/Ru(0001) where all the threading dislocations
  have islands on top.
• On CuS/Ru(0001) at 110K not all threading
  dislocations are decorated by islands of either
  Co,Cu or Ag.

32
Future

• Determine whether exchange is needed for
  threading dislocations to act as nucleation
  centers.
• If so, study the relative energies for diffusion
  compared to exchange at threading dislocations.
• Do quantitative studies of barriers involved (by
  measuring nucleation density versus temperature)

33
Ag/Ag/Pt(111) no threading dislocations?
K. Bromann et al., Euro. Phys. J. D9 (1999) 25 H.
Brune at al., Nature 394 (1998) 451
2 ML Ag/Pt(111) (annealed to 800k)
Deposition of aditional Ag at 110K