11
the critical supersaturation of Si in the Au–Si liquid alloys has been
measured. For details, readers are encouraged to refer to Ref. 55 and
the associated supplementary material.
In summary, the phenomenon of nucleation at the nanoscale has
been observed in situ in a TEM. Simple kinetic models have been
developed to quantitatively explain the data. The experimental
and theoretical approaches implemented and/or developed are
general and applicable to other material systems. The development
of nanotechnology relies on precise placement and control over
the nanoscale structure and composition of each device. And, since
nucleation is essential to the synthesis of any material, there is a
need to understand the kinetic factors underlying nucleation at the
nanoscale. Hence, we expect that similar studies will be carried to
investigate the kinetics of nucleation in other materials.
1.4 Silicon Nanowire Growth Kinetics
In this section, we will review the recent literature on in situ
observations of the VLS growth of Si nanowires in a UHV TEM. All the
experiments are carried out on clean, Au-covered Si(111) substrates
heated to temperatures above the Au–Si eutectic temperature
(~ 363°C) to form liquid Au–Si alloy droplets. Upon the introduction
of disilane gas, Si is deposited preferentially under the droplet-
substrate interface and results in the growth of nanowires. As the
nanowires grow away from the substrate, they are observed in bright
In situ observations provide a wealth of information regarding
the nanowire growth orientation, structure, crystallinity, and shapes
of the wire and the droplet. This is best illustrated in Fig. 1.6, which
shows a series of TEM images acquired in situ during the growth of
Si nanowires. The characteristics of VLS growth are readily visible in
the image—1-D morphology with a smoothly curved feature at the
tip, characteristic of a liquid droplet, which in this case is the Au–
Si alloy. From the images, we notice that the liquid–solid interface
between the nanowires and the catalyst droplets is planar. Most of
the nanowires in this experiment, carried out in “clean” environment
grow along 〈111〉. (In the presence of small amounts of oxygen,
however, Si nanowires grow along 〈110〉.
57
) From the TEM images in
Silicon Nanowire Growth Kinetics
12
In situ Observations of Vapor–Liquid–Solid Growth of Silicon Nanowires
the Si nanowires grown in the UHV TEM are single crystals. A few of
the wires show structural defects, such as stacking faults and twins,
highlighted in Fig. 1.6A.
Figure 1.6 Collage of typical in situ
from Si(111) samples during the growth of Si nanowires via
VLS using disilane as the Si source and Au as the catalyst.
The images show single-crystalline, mostly 〈111〉-oriented, Si
nanowires with liquid Au–Si alloy droplets (darker contrast
smoothly curved features) at the tips. Solid circles indicate
the absence of Au –Si droplets on a few wires (see A and D).
Dotted circles highlight structural defects that are occasionally
of surface facets and droplet shapes. (B) Under appropriate
imaging conditions, surface structure of the wires can be
better visualized. The arrows in D show tapered nanowires
with increasing diameter.
In Au–Si system, all the Si nanowires oriented along 〈111〉 are
bound by sawtooth facetted surfaces.
58
These surface facets are
clearly visible in the TEM (see Fig. 1.6B,C). Moreover, using the
appropriate imaging conditions, one can resolve the apparent 3D
structure of the nanowires (see Fig. 1.6C). The shapes of the droplets
can be measured as a function of deposition parameters and wire
diameter. These data were found to be essential in understanding the
mechanisms controlling the Au-catalyzed growth of Ge nanowires.
59
Note that the images also show nanowires with a range of diameters.
annealing and the variation in droplet size via Ostwald ripening.
While this is undesirable for the large-scale fabrication of nanowire
13
devices, we take advantage of this variation in diameters to quantify
the growth kinetics.
Another interesting and probably surprising aspect that is ob-
served in the images is the variation in the nanowire diameters along
their lengths. As seen in Figs. 1.6A,D, a few of the wires are tapered
with wider bases and narrower tips, while others are narrower at
the bases and wider at the tips. Such morphologies are often attrib-
uted either to direct non-catalyzed deposition on the sidewalls,
28
or to increased accumulation of the material at the tips.
7
In situ
observations, discussed below, reveal that tapering can also be due
to Ostwald ripening of the catalyst droplets during growth.
60
Real-time observations provide insights into the morphological
evolution of nanowires, as shown in Fig. 1.7.
Figure 1.7 In situ TEM images acquired during the growth of Si nanowires
at T = 655 °C using disilane pressure of 1 × 10
–6
Torr. The time
t
in this measurement sequence was acquired. The plot shows
time-dependent change in volumes of droplets on two wires
labeled 1 and 2. Adapted from Refs. 60 and 61.
The images show morphologies of two Si nanowires along with
their droplets obtained at various times during growth. The associated
plot shows droplet volume V as a function of deposition time t. Note
that volume of the droplet on the wire labeled 1 decreases, while
that of wire 2 increases. This behavior is characteristic of Ostwald
ripening, a curvature-driven phenomenon. In the Au–Si system, the
coarsening/decay of Au–Si alloy droplets occurs as a result of Au
60,61
In addition to direct visualization of nanowire growth, in
situ TEM allows quantitative description of the kinetics from the
measurements of the rates of growth of nanowires as a function
dL/dt, can be controllably varied over three orders of magnitude.
Silicon Nanowire Growth Kinetics
14
In situ Observations of Vapor–Liquid–Solid Growth of Silicon Nanowires
From the TEM images such as Fig. 1.6A, acquired as a function of
deposition time, dL/dt data were extracted for several wires with
d. Fig. 1.8 shows plots of dL/dt as a function of
d, P, and T. For the Si nanowires grown using disilane with Au as the
catalyst, under our growth conditions, dL/dt is practically invariant
with d, increases linearly with P, and increase exponentially with T.
The measured activation energy of 0.53 ± 0.02 eV for the VLS growth
of Si wires is nearly one-fourth of the activation energy associated
with the same process on a non-catalyzed Si surface, indicative of the
62
The experimental data shown in Fig. 1.8 can be understood
adsorption followed by dissociation of the gas precursor. This
can occur at the catalyst surface, on the sidewalls, and/or on the
substrate. In order to form a 1-D wire, the deposited adatoms
material is incorporated at the interface. For Au–Si, the rate-limiting
step for Si wire growth is the dissociative chemisorption of disilane
at the Audi alloy droplet.
28
In conclusion, in situ TEM revealed new and surprising aspects
of the VLS process of silicon nanowire growth. The images provided
direct evidence for the presence of a liquid catalyst, characteristic of
the VLS process. From the wire morphologies and time-dependent
observations, important insights into the growth dynamics and
the droplet stability were obtained. These studies are useful and
probably essential for developing large-scale fabrication of precisely
controlled nanostructures.
1.5 Summary and Outlook
In this book chapter, an overview of in situ electron microscopy
studies of silicon nanowire nucleation and growth via vapor–liquid–
solid process was presented. The combination of real-space and real-
time observation of phenomena such as the nucleation and growth
kinetics enables accurate determination of the underlying reaction
One of the unique advantages of an UHV TEM is the ability to
controllably vary the composition of the environment during in situ
15
Summary and Outlook
50 75 100 125 150
0
2x10
-2
4x10
-2
6x10
-2
8x10
-2
1x10
-
1
dL/dt (nm/s)
d (nm)
T = 655
o
C
: 1x10
-6
Torr
62
HSi
P
Si
2
H
6
(Torr)
10
-8
10
-7
10
-6
10
-5
T = 575
o
C
13.0 13.5 14.0 14.5 15.0
10
-2
10
-1
T (
o
C)
1/kT
(
1/eV
)
640 600 560 520
E = 0.56±0.05 eV
: 1x10
-6
Torr
62
HSi
P
Figure 1.8 (a) Plot of axial growth rate dL/dt as a function of Si nanowire diameter d. The data is obtained from Si nanowires grown
at T = 655 °C using disilane pressure of 1 × 10
–6
Torr. (b) Log–log plot of dL/dt versus pressure at 500 °C. (c) Arrhenius
plot of dL/dt versus T for Si nanowires grown using 1 × 10
–6
Torr disilane. The dotted and solid lines in all the plots are
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.