Chemistry Reference
In-Depth Information
Chapter 3
Wetting Layer Super-Diffusive Motion and
QSE Growth in Pb/Si
M.C. Tringides, M. Hupalo, K.L. Man, M.M.T. Loy, and M.S. Altman
Abstract
The unusual growth mode of uniform height islands discovered in Pb/Si
was related to the electronic energy modulation with island height due to quantum
size effects (QSEs). In addition to these energetic reasons provided by QSE, there
is also the question of kinetics, i.e., how atoms move at relatively low temperatures
(as low as 150 K) to build the islands in the short time of minutes. Controlled exper-
iments with different techniques have shown the intriguing role of the dense wet-
ting layer in transporting mass. STM experiments monitoring how unstable islands
transform into stable islands have shown that the wetting layer between the islands
moves selectively to the unstable islands, climbs over their sides, forms quickly
rings of constant width
20 nm, and finally it completes the island top, but at
a slower rate than the ring completion. This growth is independent of the starting
interface, whether it is the amorphous wetting layer on the Si(111) (7
∼
×
7) or the well-
α(
√
3
×
√
3
ordered Si(111)-Pb
)
surface (except Pb diffusion on the latter interface
∼
is faster by a factor of
5). Real-time low-energy electron microscopy (LEEM)
observations of mass transport phenomena have confirmed the fast mobility of the
wetting layer in Pb/Si and in addition have revealed some unusual features that are
unexpected from classical diffusion behavior. The experiment monitors the refilling
of a circular vacant area generated by a laser pulse. The concentration profile does
not disperse as in normal diffusion, the refilling speed
x
/
t
is constant (instead of
/
√
x
t
=
constant), and the equilibration time diverges below a critical coverage,
/τ
∼
(θ
c
−
θ)
−
κ
. The absolute value of the refilling speed 0.05 nm/s at 190 K
is orders of magnitude higher than what is expected from Pb diffusion on Pb crystals
at higher temperatures. These results are compared with predictions of three candi-
date models: (i) a conventional diffusion model with a step-like coverage-dependent
diffusion coefficient
D
c
(θ)
θ
c
,as1
, (ii) a model with mass transport due to adatoms on top
of the wetting layer with coverage-dependent adatom vacancy formation energy,
and (iii) the carpet unrolling mechanism proposed for other systems. None of these
