Magic clusters on group IV surfaces 3

C
C
H
H
A
A
P
P
T
T
E
E
R
R


3
3


103


C
C
h
h
a
a
p
p

t
t
e
e
r
r


3
3
:
:


E
E
x
x
p
p
e
e
r
r
i
i
m
m
e
e

n
n
t
t
a
a
l
l


P
P
r
r
o
o
c
c
e
e
d
d
u
u
r
r
e
e
s
s

3.1 The UHV System


In the study of surface structures in the nano-scale regime, effects of surface and
interfacial energy become more pronounced as opposed to studies of bulk material
characteristics. This inevitably requires an atomically clean and flat substrate surface
without the undesirable influence of contaminants on surface/interfacial energies in order
to accurately ascertain surface phenomenon. As main sources of contaminants are
typically oxygen, water or carbon-related species etc, which are found abundantly in the
atmosphere, a UHV system is crucial towards providing a clean environment for in-situ
sample preparation and material deposition. It allows source materials to be maintained at
elevated temperature in UHV over long period of time which tends to eliminate moisture
and results in higher purity source material and higher quality monolayer film. This
environment also allows the sample surface to remain contaminant-free within the
experimental time frame so as to ensure accuracy and consistency in observations

The reduction of contaminants such as moisture is important, as these impurities
are known to affect the preparation of clean surfaces and prevent high quality layer
growth due to creation of crystal defects and carrier traps. This impedes surface migration
C
C
H
H
A
A

P
P
T
T
E
E
R
R


3
3


104
of adsorbates as the defects often act as nucleation centers for the formation of lattice
defects, such as stacking faults and dislocations, which modify the adsorption and growth
chemistry of materials.

As the UHV environment does not suppress the occurrence of contaminants
indefinitely, it is important to be able to appreciate the rate of contamination in order to
plan the experiments. By assuming that the rate of contamination is analogous to the rate
of arrival of gaseous molecules on a clean surface and from consideration of the kinetic
theory of gases, we can describe the rate of surface bombardment by molecules, Z, as
given by:

mkT
p
Z
π

2
= cm
-2
.s
-1
Eqn. (3.1)

where p is the ambient pressure (in N.cm
-2
), m is molecular mass (in Kg.molecule
-1
), T is
absolute temperature (in K), and k is Boltzmann constant (in J.K
-1
).

As the rate of surface contamination also depends on the sticking probability, S(θ),
we can, by assuming the worst case of S(θ) = 1, to estimate the coverage of CO (a typical
gaseous contaminant at 300K) at ambient pressures of 10
-6
Torr and 10
-10
Torr,
respectively (1 Torr = 1.333 x 10
-2
N.cm
-2
), using Eqn. (3.1). At 10
-6
Torr, Z is

determined to be 3.82 x 10
14
cm
-2
.s
-1
. Assuming an atomic density of 10
15
cm
-2
(typical of
most surfaces), this will imply a rate of contamination of 0.382 ML.s
-1
. Alternatively, the
time taken for a clean surface to be saturated with 1 ML of CO will be 2.6 seconds. If the
C
C
H
H
A
A
P
P
T
T
E
E
R
R

3
3


105
ambient pressure is now 10
-10
Torr, the time taken to saturate the clean surface will be
26178 seconds, which is about ~ 7 hours. Hence these calculations demonstrate the
importance of conducting experiments under UHV regimes where ambient pressures are
< 10
-9
Torr, which will allow sufficient time for sample preparation, film growth and
characterization before surface contamination.

For this work, the samples were introduced into an UHV environment via a fast
entry lock and the experiments which involve in-situ sample preparation, XPS scanning
and STM imaging were performed in an OMICRON UHV System. It comprises of 3
main chambers; a fast entry lock chamber, a preparation chamber and an analysis
chamber which are all interconnected to allow the experiments to be conducted in-situ.
Figures 3.1, 3.2 and 3.3 show the OMICRON UHV system layout.

Figure 3.1: Plan view of OMICRON UHV schematic
Fast Entry
Load Lock
Preparation
Chamber
Analysis
Chamber
STM
C
C
H
H
A
A
P
P
T
T

E
E
R
R


3
3


106










Figure 3.2: Side view of OMICRON UHV schematic

Figure 3.3: 3-D view of OMICRON UHV schematic.
Turbo Pump
Ion Pump
Pneumatic
Gate Valve
Electron Beam
Sources
Manipulator
XPS Analyzer
STM
Sample Transfer
Magnetic Probe
Arms
LEED
STM
Manipulator
C
C
H
H
A
A
P
P
T

T
E
E
R
R


3
3


107
The fast entry lock (FEL) chamber is pumped by a rotary pump and a turbo-
molecular pump which makes it possible for the chamber to be pumped down in stages
from atmospheric pressures until pressures equivalent to preparation chamber conditions.
As the FEL is separated from the preparation and analysis chamber by a gate valve, this
allows for the introduction of samples and tips without disrupting the vacuum conditions
in the preparation chamber. The main function of the preparation chamber is to clean
sample surfaces and deposit thin films in-situ. This chamber is pumped primarily by a
VARIAN turbo-molecular pump (useful for pumping light gases) with a pump speed of
500 litres/sec and a VARIAN VacIon Plus 150 ion pump with a pump speed of 150
litres/sec (useful for maintaining oil free UHV conditions). The analysis chamber is
linked to the preparation chamber via a gate valve and consists of the main analytical
tools such as STM and XPS. This chamber is pumped by a second VARIAN turbo-
molecular pump with a pump speed of 500 litres/sec and a VARIAN VacIon Plus 300 ion
pump with a pump speed of 300 litres/sec. Both preparation and analysis chambers are
also equipped with titanium sublimation pumps, which sublime titanium to getter
chemically active gases if required.
C
C

H
H
A
A
P
P
T
T
E
E
R
R


3
3


108
3.2 The sample holder and transfer system

The substrates are cut into 4mm by 9mm strips before ex-situ preparation and
mounted onto Direct Heating (DH) sample holders provided by OMICRON. The holders
are made mainly from molybdenum and tantalum due to the high temperature stability of
these materials as samples are expected to be heated to temperatures of up to 1200°C by
passing a direct current through the sample. Figure 3.4 shows the DH sample holder and
a sample mounting.

Figure 3.4 Schematic of the OMICRON DH sample holder [1].

C
C
H
H
A
A
P
P
T
T
E
E
R
R


3

3


109
The sample holders held by a rotary grip mechanism as shown in Figure 3.5, and
are transferred from one chamber to another via the transfer arms linearly using magnetic
probe sliding mechanisms. As both preparation and analysis chambers are equipped with
manipulators for high precision x, y, z positioning to allow fine control of sample
positions for film deposition, LEED and XPS analysis, the sample transfer between
transfer arm to manipulator or manipulator to STM carousal is facilitated by both the
transfer arms or wobble stick pincers as shown in Figure 3.6. The manipulators which are
shown in Figure 3.7, are equipped with 2 modes of sample heating. They are (1) direct
current heating (DH), whereby current is passed through the sample via contact brushed
and it is resistively-heated by the internal resistance of the sample itself; and (2) resistive
heating (RH), whereby current is passed through rows of pyrolytic boron nitrite (PBN)
wires located on the manipulator to heat up the back-end of the sample holder.










Figure 3.5: Schematic of the head of the magnetic probe transfer arm, showing the rotary
grip mechanism during gripping and releasing of sample holders. Rotation of the transfer
arm shaft opens and closes the gripping arm of the transfer head [1].
Gripping arm

C
C
H
H
A
A
P
P
T
T
E
E
R
R


3
3


110

Figure 3.6: Schematic diagram showing the sample transfer process between (A)
manipulator and transfer arm via transfer head and (B) manipulator and STM carousal via
wobble stick pincer [1].



















Figure 3.7: Schematic diagram showing the manipulator head in (A) 3-D view (B) Plane
view and (C) Side view pincer [1].


Manipulator
Transfer Arm inserts sample
into manipulator groove

Rotary Grip Mechanism – Head
rotates to release grip
Transfer Arm retracts leaving
sample in groove
Sample
holder
(B) Transfer Arm (A) Wobble Stick Pincer
Sample holder
(B) Plane view
(a) 3-D view (C) Side view
Contact brush
for DH heating
PBN wires for
RH heating
C
C
H
H
A
A
P
P
T
T
E
E
R
R

3
3


111

3.3 Scanning Tunneling Microscope

STM images are acquired by scanning an atomically sharp metal tip across a conducting
surface at a distance of approximately 5 Å to 10 Å, such that the wave functions of the tip
and sample overlap. Consequently, quantum tunneling of electrons across these two
materials can occur. This situation is schematically illustrated in Figure 3.8.


Figure 3.8: Overlapping wave functions of sample and tip.

There is a finite probability that electrons can cross the gap from one end to the
other. Fowler et al [2] derived an expression, which is shown in Eqn. (3.2), for the
tunneling current based upon the tip – sample separation and the work functions of the
materials involved.


)exp(
φ
kd
d
U
I −∝ Eqn. (3.2)

Sample

Tip
Vacuum
Separation
~ 5 Å to 10 Å
C
C
H
H
A
A
P
P
T
T
E
E
R
R


3
3


112
where I is the tunneling current, U represents the applied bias across the gap, d is the
separation of the gap, Φ is the average work function (U << Φ), and k is a constant.

The tunneling current depends exponentially on the separation of the tip and
sample surface. The constant k has a value of 1.025 Å

-1
(eV)
-1/2
for a vacuum gap.
Therefore, there is roughly an order of magnitude change in the tunneling current for a
change in separation of only 1 Å. This is the reason for the extremely high vertical
resolution of STM.

The basic setup of STM is given in Figure 3.9a, which shows the various key
components of STM. The feedback loop is essential for maintaining a constant tunneling
current in the constant-current mode, which will be discussed later. The current amplifier
is employed for amplifying the weak tunneling current. There are 2 basic modes of STM
operations, namely “constant-current” and “constant-height” mode, which are shown in
Figure 3.9b. In the constant-current mode, the vertical position of the tip is controlled by
the feedback loop. The value of the tunneling current measured at the tip position is
compared with a value preset by the operator via the feedback loop. If the measured value
exceeds the present value (e.g. smaller tip – sample separation), a voltage will be applied
to the z – piezotube, which then retracts the tip so as to maintain the current at the preset
value. Similarly, if the current is too low (e.g. smaller tip – sample separation), the
feedback loop will drive the tip towards the surface. The tip is raster scanned by means of
the x – and y – piezotubes across an area of up to several thousands of angstroms and the
voltages applied to the z – piezotube will change accordingly in order to maintain a
C
C
H
H
A
A
P
P

T
T
E
E
R
R


3
3


113
constant tunneling current. The voltages applied are then plotted as a function of position,
which will then yield topographic or electronic image of the surface.

In constant-height mode, the feedback loop is deactivated. The tip, moving at
constant height, experiences a tunneling current, which varies with the atomic
corrugation of the surface (see Figure 3.9c). This mode has the advantage that the tip can
be scanned rapidly across the surface. However, it is restricted to samples that are very
flat, since sudden large changes in surface topography can result in tip crashes.



(a)
C
C
H
H
A

A
P
P
T
T
E
E
R
R


3
3


114

Figure 3.9 Schematic illustration of (a) the STM apparatus, (b) “constant-current” mode
and (c) “constant-height” mode of operations [3].

Eqn. (3.2) shows that the tunneling current includes contributions from both
topographic and electronic features (i.e. d and Φ). Figure 3.10 illustrates the principle of
sampling different electronic states in the region of the bandgap using STM. By choosing
a particular bias for scanning, different electronic states can be probed. Applying a
negative sample bias allows occupied-states of the sample to be imaged (Figure 3.10a),
while positive sample bias yields empty-state information (Figure 3.10b). All states
within the energy window defined by the bias voltage may contribute to tunneling, but
major contribution comes from those close to the Fermi energy as they have the largest
decay lengths.

(b) (c)
C
C
H
H
A
A
P
P
T
T
E
E
R
R


3
3


115

Figure 3.10: (a) Occupied-state imaging, and (b) empty-state imaging.

STM studies were performed using an OMICRON STM. A carousel, which
resides within the analysis chamber, provides storage space for up to 6 samples or tip
holders. The holders are transferred to and from the STM by means of a wobble stick.
Figure 3.11 illustrates the STM scanning stage.

Figure 3.11: Schematic diagram of the STM scanning stage [3].

E
F
eV

Sample
(-ve)
Tip
(+ve)
E
F
eV

Sample
(+ve)
Tip
(-ve)
E
F
(a)

(b)

C
C
H
H
A

A
P
P
T
T
E
E
R
R


3
3


116

The sample holder is inserted into the STM scanning stage and it will rest on the
3-contact bearing, which is shown in Figure 3.12. All STM images were taken with
tunneling currents from 0.1 to 1.2 nA and bias from –2 V to +2 V applied to the sample.
The OMICRON VT STM uses a single tube scanner with a maximum scan range of
about 10 x 10 µm with a z – travel of about 1.5 µm. A z – resolution of better than 0.1 Å
can be achieved. The STM has been configured such that voltage biases are measured
with respect to the substrate, this would imply that a negative bias voltage produces a
filled state image, where the electrons would be tunneling from the substrate surface to
the tip. All the images were recorded using the constant current mode. Dark features on
the images correspond to depressions and bright to elevations of the surface. Height
analysis, length and the surface reconstruction periods of the STM were carried out with
the use of a single scan line evaluation.

C
C
H
H
A
A
P
P
T
T
E
E
R
R


3
3


117

Figure 3.12: Schematic of loading of sample plate into STM stage [3]

For eddy current damping during STM scanning, a ring of Cu plates is mounted
on the STM stage. A ring of magnets is fixed at the columns of the spring suspension as a
counterpart successfully damping excursions in all directions. The resonance frequency
of the system is about 2Hz. The STM stage can be locked by using the push-pull motion
drive, in order to enable sample and tip exchange by means of a wobble stick. The eddy
current and vibration damping mechanisms are shown in Figure 3.13.

C
C
H
H
A
A
P
P
T
T
E
E
R
R


3
3


118

Figure 3.13: Side-view schematic of the VT STM [3]

C
C
H
H
A

A
P
P
T
T
E
E
R
R


3
3


119
3.4 X-Ray Photoelectron Spectroscopy (XPS)

XPS was developed in the mid 1960s by K. Siegbahn and his co-workers who
were later awarded the Nobel Prize for Physics in 1981. The phenomenon is based on the
photoelectric effect, where the electrons were ejected from a surface when bombarded
with photons. The photoemission principle states that if an atom absorbs a photon of
known energy, an electron from the core shell of the atom, where binding energy, E
B
, of
that electron is lower than the excitation photon, hυ, will be ejected out from the atom.
This photoelectron will possess a kinetic energy of (hυ – E
B
). However, this
photoelectron has to leave the surface and into vacuum, it must overcome the work

function, φ, of the specimen. Hence, the resultant kinetic energy, E
K
, possesses by the
photoelectron will be given in Eqn. (3.3). Figure 3.14 shows the photoemission process.

ϕυ
−−=
BK
EhE
Eqn. (3.3)

Figure 3.14: A typical photoemission process.

hυ
1s
2p
2s
2p
2s
1s
valence
valence
φ
φ
E
K
vacuum
vacuum
Core Level Electron
Photoelectron

C
C
H
H
A
A
P
P
T
T
E
E
R
R


3
3


120
The XPS technique is highly surface specific due to the short inelastic mean-free-
path of photoelectrons that are excited from the solid. Typical excitation X-ray sources
for XPS are Al K
α
(1486.6 eV) and Mg K
α
(1253.6 eV). Other X-ray lines can also be
chosen such as Ti K
α

(2040 eV). The energy of the photoelectrons leaving the sample is
determined using a concentric hemispherical analyzer (CHA) and this gives a spectrum
with a series of photoelectron peaks. Figure 3.15 shows the schematic diagram of the
CHA.


Figure 3.15: Schematic diagram of a concentric hemispherical analyzer (CHA).

Concentric Hemispheres
Entrance
Slits
Electron Path
Electrostatic Lens
Electron Detector
Sample
Exit
Slits
C
C
H
H
A
A
P
P
T
T
E
E
R

R


3
3


121
A CHA consists of two metal hemispheres. One hemisphere is concave in shape,
while the other is convex. They are arranged such that their centers of curvature are
coincident. Different voltages are applied on each hemisphere such that there is an
electric field between the two hemispheres. Photoelectrons are injected into the gap
between the hemispheres. If the electrons are travelling very fast, they will impinge on
the outer hemisphere. If they are travelling very slow, they will be attracted to the inner
hemisphere. Hence only electrons in a narrow energy region, called the pass energy,
succeed in getting all the way round the hemispheres to the detector. A series of lenses
are placed before the CHA. The lenses enable two operating modes - Constant Retard
Ratio (CRR), or Constant Analysis Energy (CAE). With CRR mode, the electrons are
slowed down in order to maintain a constant “actual:detected” energy ratio of the electron
to be analyzed. That is if the retard ratio is 10, and 1000eV electrons are to be detected,
then the electrons will be slowed down to 100eV, and the pass energy will be set to 100
eV. In the CAE mode, the pass energy is fixed. Hence if the pass energy is 50 eV, then
electrons of 1000eV will have to be slowed down by 950 eV in order to be detected. The
CRR mode gives constant resolving power and the CAE mode gives constant energy
resolution.

The binding energy of the peaks is characteristic of each element. The peak areas
can be used (with appropriate sensitivity factors) to determine the composition of the
materials surface. The shape of each peak and the binding energy will be slightly altered
by the chemical state of the emitting atom. Hence, XPS can provide chemical bonding

C
C
H
H
A
A
P
P
T
T
E
E
R
R


3
3


122
information as well. XPS is not sensitive to hydrogen or helium, but can detect all other
elements.

The XPS system comprises of a XR3E2 twin-anode X-ray source, an EA 125
hemispherical analyzer and associated electronic racks together with an output link to a
PC-based DAT 125 software system to capture and analyze XPS spectra. The schematic
diagram of the twin-anode X-ray source is given in Figure 3.16.

Figure 3.16: Schematic diagram of the XR3E2 twin-anode X-ray source [4]

The XR3E2 X-ray source is equipped with a dual anode, magnesium coated on
one side and aluminum on the other, which provides an X-ray line of 1253.6 or 1486.6
eV, dependent on the anode selected. The anode is seated on an annealed copper gasket
mounted through the source-mounting flange to project through the inner filament shield.
Three ceramic bushes centralize the position of the anode within the filament inner
shield. The filament is mounted concentrically with the inner shield enclosed by an outer
shield, in which a thin aluminum foil window is fitted. The electrical connections to the
Filament
Filament
Outer Shield
Window
Filament
Inner Shield

CU gasket
Anode
Assembly
Mounting
Flange
Insulated Tube

Anode

Cooling
Coupling Flange
Source cover
Cutaway Side
C
C
H
H
A
A
P
P
T
T
E
E
R
R


3
3


123
double filament are made through three filament supports mounted on the three-way feed
through connectors in the source-mounting flange. The high voltage connection to the
anode is made through the conduit supplying cooling water to the anode region of the
source.

The X-ray source operated at a potential of up to 15 KV and is controlled by the
8025 X-ray power supply. Electrons emitted by the coated filament, which is near earth
potential, are rapidly accelerated toward the high voltage potential of the anode. The
rapidly moving electrons impinge on the anode causing the emission of the X-rays. A
thin aluminum window mounted in the filament outer shield isolates the anode field and
prevents stray electrons from affecting experiments. The window also acts as a partial
filter for unwanted x-ray lines. When used in a hostile analysis chamber environment, the
window enables the filament section of the source to be differentially pumped.

All wide scans were carried out in step sizes of 1 eV; narrow scans were carried
out in step sizes of 0.05 eV. Binding energy calibrations were carried out using Si 2p
3/2
as
the reference peak.
C
C
H
H
A
A
P
P
T
T
E
E
R
R

3
3


124

3.5 UHV Electron-Beam Evaporator

In this dissertation work, all materials deposition is done using a FOCUS EFM3T
UHV electron-beam evaporator. This is a 3-in-1 evaporator whereby 3 different materials
can be evaporated onto a substrate simultaneously. It consists of 3 separated evaporation
cells to avoid cross-contaminations. The 3 cells will be water-cooled during the
deposition process. Figure 3.17 below shows the exterior-view of the FOCUS EFM3T.









Figure 3.17: Exterior-view of the FOCUS EFM3T [5].

Evaporant materials, in this case, Si and Co, are in the form of rods. The
bombarding electron beam, from a heated filament, induces a temperature rise of the
evaporant, causing evaporation. The evaporator is equipped with a flux monitor. Once
calibrated, the flux monitor can replace the necessity of a quartz thickness monitor by
continuously monitoring the evaporation rate. Flux is measured directly, which allows a
much more precise rate adjustment and much faster rate control than an indirect method,

100 mm

Sample
Flux Monitor Feedthrough
HV Feedthrough
HV Feedthrough
Cooling water
Shutter Actuater
C
C
H
H
A
A
P
P
T
T
E
E
R
R


3
3


125
e.g. temperature measurement by a thermocouple. Figure 3.18 shows the various interior

components of the evaporator.














Figure 3.18: Interior-view of the evaporator [5].

The flux monitor is actually an ion collector, which is situated in the evaporant
exit column. At a given electron emission current, I
E
, and e-beam energy, the ion flux
measured there is directly proportional to the flux of evaporated atoms.

To begin evaporation, the manipulator has to be positioned such that the sample is
about 100 mm from the evaporator and is facing it normally. Filament current, I
f
, will be
increased to about 1.9 A to supply the bombardment electrons. A positive bias ranging
from 850 V to 950 V is applied to the evaporant rod. This is to attract and focus the
electrons towards the end of the rod. At this stage, I

E
would have increased and the end of
the evaporant rod should be heated up. The parameters mentioned can be adjusted until a
suitable and stable flux is achieved, where deposition can then commence by opening the
Flux
C
C
H
H
A
A
P
P
T
T
E
E
R
R


3
3


126
cell shutter. To terminate the deposition process, the cell shutter can be shut and the
control parameters decreased to cool the evaporant rods.
C
C

H
H
A
A
P
P
T
T
E
E
R
R


3
3


127
3.6 Pre-experimental Preparations and Procedures

3.6.1 Achieving and Maintaining UHV Conditions

As UHV conditions are necessary in both chambers to allow controlled
experimentation, there are various steps in achieving pressures of 10
-10
mbar in the
system. The UHV flanges are firstly sealed and tightened before the system is pumped
down from atmosphere to 10
-4

mbar using the rotary pumps and turbo molecular pumps.
Upon which, the ion pumps are turned on and the system is further pumped down to
pressures of 10
-7
mbar. At this point, leakage of atmospheric gases into the system due to
poorly sealed flanges or other faulty parts would be apparent, as pumping would be slow
or unable to reach the desired vacuum pressures. If a leak is suspected, the RGA (Inficon
mass spectrometer) is used as a leak detector, while He gas is passed around the entire
system.

Assuming there are no leaks, all non-bakeable items are removed from the system
and the bakeout covers put into place. The gate valves between the chambers are kept
open, as the seals maybe damaged if closed at elevated temperatures (~ 250°C). Once the
bakeout covers are in place, the temperature of the system is ramped up to 150 °C and
monitored by a thermocouple placed on the outside of the system. The system is kept at
this temperature for more than 72 hours. The length of the bakeout depends upon the
amount of time for which the system has been exposed to atmospheric pressure and the
introduction of any new components to the system. When the heaters are turned off, the