tesi_maria edera
TRANSCRIPT
1
UNIVERSITÁ DEGLI STUDI DI MILANO-‐BICOCCA
Scuola di Scienze
Corso di Laurea Magistrale in Fisica
Ion Irradiation of Nanocrystalline Graphene
on Quartz and Sapphire
Relatore: Prof. Alexander ZAITSEV
Correlatore: Prof. Alberto PALEARI
Tesi di Laurea di:
Maria EDERA
Matr. N. 769503
Anno Accademico 2014/2015
2
I want to thank
Professor Zaitsev for his guidance
For walking me down the path of science
Showing me its beauty
And making me discover a new way to look at it.
Professor Paleari
For his patience and paternal guidance
For believing in me
And sustaining me.
And The One who gives me everything
For His love beyond belief.
3
Index Chapter 1 Introduction ........................................................................................... 5
Chapter 2 Theory .................................................................................................... 9
2.1 Graphene ........................................................................................................ 9
2.1.1 Electronic properties ............................................................................ 11
2.1.2 Optical properties ................................................................................ 13
2.1.3 Other Properties ................................................................................... 14
2.1.4 Methods of Synthesis ........................................................................... 15
2.2 CVD .............................................................................................................. 16
2.2.1 Process of growth ................................................................................. 16
2.2.2 Types of CVD ....................................................................................... 17
2.2.3 Pros and Cons ....................................................................................... 18
2.3 Ion irradiation, implantation and sputtering ........................................... 18
2.3.1 Implantation ......................................................................................... 19
2.3.2 Implantation profile ............................................................................. 19
2.3.3 Damage .................................................................................................. 22
2.3.4 Annealing .............................................................................................. 23
2.3.5 Sputtering ............................................................................................. 24
2.3.6 Pros and Cons ....................................................................................... 29
Chapter 3 Equipment ........................................................................................... 31
3.1 Furnace ........................................................................................................ 31
3.2 Vacuum System ........................................................................................... 32
3.3 FIB ................................................................................................................ 34
3.4 Low Pressure Plasma System ..................................................................... 39
3.5 Conductance measuring system ................................................................. 41
3.6 Microscopes ................................................................................................. 42
Chapter 4 Growth of nanocrystalline graphene ................................................. 44
4.1 Experimental procedure for growth ......................................................... 44
4.2 Results .......................................................................................................... 46
4.2.1 Temperature ......................................................................................... 46
4.2.2 Pressure ................................................................................................. 48
4.2.3 Time ....................................................................................................... 50
4.2.4 Pulsed growth ....................................................................................... 52
4
Chapter 5 Plasma Treatments ............................................................................. 54
5.1 Procedure ..................................................................................................... 54
5.2 Results .......................................................................................................... 55
5.2.1 Argon treatments ................................................................................. 55
5.2.2 Oxygen treatments ............................................................................... 56
5.2.3 Nitrogen treatments ............................................................................. 57
5.2.4 Hydrogen treatments ........................................................................... 57
5.2.5 Krypton treatments .............................................................................. 58
Chapter 6 FIB Irradiation .................................................................................... 60
6.1 Experimental procedure ............................................................................. 60
6.1.1 Calculation of the dose ......................................................................... 60
6.1.2 Preparation and irradiation of the sample ........................................ 61
6.2 Results and discussion ................................................................................ 65
6.2.1 Conductance after irradiation and after annealing .......................... 65
6.2.2 Results in adhesion after irradiation and annealing ......................... 76
6.2.3 Results in Nucleation improvement of Graphene on Ion-Irradiated Substrates ....................................................................................................... 80
Chapter 7 Conclusion ........................................................................................... 84
Bibliography .......................................................................................................... 87
5
Chapter 1 Introduction
Graphene, a new prospective electronic nanomaterial, has attracted much
attention as an alternative to traditional semiconductors in a number of
applications [1-7]. Graphene has many advantageous physical properties
making it unique among other electronic materials. One of them is the
combination of a fair electrical conductance and very high optical
transmittance in a broad spectral range. This makes graphene a material of
choice for optoelectronics and photovoltaics where high optical
transmittance is required, but the moderate electrical conductance is not an
issue. Another one is the combination of the 2D atomic structure and a very
high mobility of charge carriers. This makes conductance of graphene very
responsive to the presence of adsorbates on its surface and hence makes
graphene an extremely sensitive electronic chemical sensor [8]. Another
advantage of graphene is that it is a true nanoelectronic material. Unlike
semiconductors with considerable energy bandgap between the valence and
conduction bands, graphene is a semimetal with zero bandgap energy.
Therefore, the unipolar conductivity – the key property of an electronic
material - can be induced in graphene without impurity doping but
controllably by a low external bias. In contrast, in conventional
semiconductors, the unipolar conductivity can be practically achieved only
via impurity doping. The impurity doping is one of the major limitations of
non-zero bandgap semiconductors, which limits the down-scaling of
electronic devices based on the principle of bipolar junction. For instance,
silicon field-effect transistor (FET) ceases to work as a practical device at a
size of 10 nm.
Graphene reveals its highest electronic parameters only when in a form of
isolated single crystal layer. However, direct growth of perfect single crystal
graphene on surface of standard electronic materials (silicon and silicon
oxide) remains an unmet challenge. By now, commercially viable
technologies are those based on two-step procedure of CVD growth of
graphene on copper followed by transfer of the grown graphene onto the
6
working substrate. The step of the transfer is one of the major limiting
factors of these technologies. The recently reported direct CVD growth of
graphene on germanium could be a promising alternative [9]. Yet it is not
applicable for direct deposition on silicon or SiO2.
A technologically simple and inexpensive alternative to the high quality
single crystal graphene is nanocrystalline graphene [10, 11]. Nanocrystalline
graphene can be grown by CVD deposition from gas precursor directly on
almost any material over large area [12-13]. Although the electronic
properties of nanocrystalline graphene are inferior to those of single crystal
graphene (low charge carrier mobility in the range 10 to 100 cm2/Vs and
rather high growth temperature over 800°C), nanocrystalline graphene
possesses reasonably high electrical conductance (between 10-4 and 10-5
siemens), high optical transmittance (about 95%) and high chemical
sensitivity (even non-structured nanocrystalline graphene can detect the
presence of NO2 molecules at a concentration of a few tens of ppb).
In the first part of this work, the CVD growth is studied, identifying the
relevant parameters, and finding the values which provide the best quality
nanocrystalline graphene layers.
The CVD growth is studied, in particular its dependence on temperature,
pressure, and time. The knowledge of the dependence on these parameters
allowed us to better understand the growth mechanism, and to interpret the
results of the whole investigation.
When adopting a new material in electronics, along with the technology for
its growth, it is equally important to develop a technology for its structuring
and patterning. Graphene, as a 2D material, is very suitable for planar
patterning, e.g. using lithography.
The second part of the work presents the analysis of the effects of plasma
pre-treatments of the substrate on graphene growth to investigate the
feasibility of patterned growth of nanocrystalline graphene. One part of the
sample was exposed to the plasma treatment, while the other was covered
with a photoresist. After the treatment, graphene growth was performed and
7
measurements of conductance were made in order to see the contrast
between the treated and pristine area, if any.
Unfortunately, after pre-treatments in Argon, Oxygen, Nitrogen, Hydrogen
and Krypton plasmas, no relevant effect of growth suppression or
enhancement has been observed.
A specific related feature of graphene, which can be used for its patterning,
is a very high sensitivity of conductance to radiation damage. It has been
shown in a number of publications [14-20], that practically any irradiation
(even with electrons of energy as low as 10 keV) damages graphene and
reduces its conductance. Upon achieving a critical concentration of defects
of 1%, the irradiated graphene becomes actually insulating [21-24].
Although the electron irradiation has been shown to change structural and
electronic properties of graphene [25-31], its efficiency is limited. The ion
irradiation is a more effective technique in many ways. First, the ion
irradiation is much more efficient in the energy transfer from the energetic
ions to the graphene atoms and, consequently, in defect production. Second,
due to the processes of the secondary irradiation by the recoil atoms
generated in graphene and in the underlying substrate, the spectrum of the
structural defects produced by ion irradiation is much broader than that
produced by electrons. Third, the ion irradiation offers unique opportunity
for impurity doping (ion implantation). This impurity doping can be
achieved both by incorporation of atoms from the primary ion beam and via
backscattered atoms from the substrate. Fourth, the ion irradiation is the
most precise technique of deterministic 2D and 3D modification of
materials at a level down to a few nanometers and the method of addressing
single elements in nanostructures. This is the object of the third part of this
work, in which ion irradiation were performed via FIB on nanocrystalline
graphene layers, grown on quartz and sapphire.
The first effect we observed and analyzed is the reduction of conductance
due to irradiation with Ga+ ions via FIB.
This effect offers an opportunity of achieving sharp contrast conductor-
insulator without physical removal of the graphene layer. If the ion
8
irradiation is performed with focused ion beam (FIB), the transition between
the conductive and non-conductive areas can be made as sharp as 1 nm.
This opportunity is attracting attention as a way towards the development of
a technology of maskless, resist-free patterning and as an approach to
fabricate novel graphene-based electronic nanodevices, like polarity-
reversible FETs [32] with on-off current ratio by far exceeding that of
conventional graphene FETs [33]. The development of a resist-free
technology of patterning could also resolve the issue of resist residues on
the surface of graphene, which considerably affect its transport properties
[34-36].
Another useful effect of ion irradiation we observed in this work is the
improvement of adhesion of thin films to the underlying substrates. The
effect of the irradiation-enhanced adhesion has been studied for decades and
is well known in material science, chemistry, biology, and medicine [37,
38]. Although the reliable adhesion of graphene to the working substrate is
an important technological property, the influence of irradiation and other
treatments with energetic particles (e.g. plasma) on adhesion of graphene
has not been studied systematically yet [15, 39].
Along with irradiation-induced suppression of conductivity and irradiation-
induced adhesion, the irradiation-induced enhancement of nucleation of
graphene on substrate is another effect we analyzed in the present work.
This effect may be used for patterning, even though its study is at the very
beginning. The only relevant publication we could find is [40], in which the
enhanced growth of graphene on SiC was demonstrated using FIB
irradiation with Si+ ions.
Below we show that all the effects of ion irradiation on graphene just
mentioned can be used for patterning nanocrystalline graphene as well.
Besides, we have found two new effects, specifically the restoration of
conductance after high temperature annealing and the evaporation at high
temperatures, which can also be used for patterning.
9
Chapter 2 Theory
2.1 Graphene
Nobel prize in physics 2010, devoted to graphene, has collected a lot of
attention in the world of scientific research since its discovery. The
possibility of research, discoveries and applications is becoming broader
every year.
The reason for this attention is due to its very particular, some say
miraculous, properties. It is the thinnest known material in the universe and
the strongest ever measured. Its charge carriers exhibit giant intrinsic
mobility, have zero effective mass, and can travel for micrometers without
scattering at room temperature. Graphene can sustain current densities six
orders of magnitude higher than that of copper, shows record thermal
conductivity and stiffness, is impermeable to gases, and reconciles such
conflicting qualities as brittleness and ductility [1].
It is also a very interesting material, since it is actually not so new. In fact,
graphene has been hidden behind the pencil trace since it was invented in
1656 in England. Ludwig Wittgenstein once commented that ‘the aspects of
things that are most important to us are hidden because of their simplicity
and familiarity’ [7].
Indeed, what we call graphene is simply a monoatomic layer of the much
more known material, graphite. It is the basic structural element of other
carbon allotropes, not only graphite but also carbon
nanotubes and fullerenes.
Picture 1 Carbon allotropes: graphite, fullerene and carbon nanotubes
10
It is a monolayer, honeycomb lattice of carbon atoms.
Each atom has four bonds, one σ bond with each of its three neighbors and
one π-bond that is oriented in the z-direction (out of the plane). The atoms
are about 1.42 Å apart.
One can visualize the π orbital as a pair of symmetric lobes oriented along
the z-axis and centered on the nucleus. Each atom has one of these π-bonds,
which are then hybridized together to form what are referred to as the π-
band and π*-bands. These bands are responsible for most of the peculiar
electronic properties of graphene.
The strong σ bounds instead are responsible for the amazing strength of this
material.
The hexagonal lattice of graphene can be regarded as two interleaving
triangular lattice (Picture 2). The atoms A and B in the picture, are actually
exactly the same, but the honeycomb structure is not a good basis to form a
Bravais lattice. With the identification of A and B atoms instead it is easy to
have a basis that can give a Bravais lattice.
Picture 3 Unit cell in the cristalline lattice and first Brillouin zone
Picture 2 Graphene’s Triangular sublattice and graphene’s bonds, sigma and pi
11
2.1.1 Electronic properties
Carbon is a chemical element of the group IV of the periodic table. It has
atomic number 6 and an electronic configuration 1s22s22p2. It is a non
metallic element and has four valence electrons, since the ones in orbital 1s
do not form bonds. Two are in the 2s subshell and two in the 2p subshell
when it is in the ground state.
Hybrid orbitals sp are formed when carbon is forming bonds with other
carbon atoms.
It also promotes one of its 2s electrons into its empty 2p orbital.
Three different types of sp orbitals can form, sp, sp2 and sp3 depending on
the number of s and p orbitals involved.
Carbon atoms with sp2 and sp3 hybrid orbitals are able to form three and
four bonds with neighboring carbon atoms, respectively, which form the
bases of graphene and diamond [6].
Looking at graphene band structure, the linear dispersion between energy
and momentum is evident (Picture 4). Because of this property, graphene
exhibits electronic properties for a two-dimensional (2D) gas of charged
particles described by the relativistic Dirac equation, rather than the non-
relativistic Schrodinger equation with an effective mass, and so the carriers
behave like particles with zero mass and an effective ‘speed of light’ of
around 106 m s–1 [5].
Picture 4 Graphene's band structure, taken from [7]
12
The electronic structure of single layer graphene (SLG) can be described
using a tight-binding Hamiltonian. Because the bonding and anti-bonding σ-
bands are well separated in energy (>10 eV at the Brillouin zone centre Г),
they can be neglected in semi-empirical calculations, retaining only the two
remaining π-bands. The electronic wavefunctions from different atoms on
the hexagonal lattice overlap. However, any such overlap between the pz(π)
and the s or px and py orbitals is strictly zero by symmetry. Consequently,
the pz electrons, which form the π-bonds, can be treated independently from
the other valence electrons. Within this π-band approximation it is easy to
describe the electronic spectrum of the total Hamiltonian and to obtain the
dispersion relations
E(kx, ky) restricted to first nearest-neighbour interactions only:
!±(!!,!!) = ±!! !+ ! !"# !!!!!
!"# !!!!+ ! !"# !!!
!
! (1)
where a = √3 acc (with acc = 1.42 A being the carbon–carbon distance) and
γ0 is the transfer integral between first-neighbour π-orbitals (typical values
for γ0 are 2.9–3.1 eV). The k = (kx, ky) vectors in the first Brillouin zone
constitute the ensemble of available electronic momenta.
With one pz electron per atom in the π–π* model (the three other s, px, py
electrons fill the low-lying σ-band), the (–) band (negative energy branch) in
equation (1) is fully occupied, whereas the (+) branch is totally empty.
These occupied and unoccupied bands touch at the K points. The Fermi
level EF is the zero-energy reference, and the Fermi surface is defined by K
and K′.
Expanding equation (1) at K(K′) yields the linear π- and π*-bands for Dirac
fermions:
!± к = ±ħ!! к (2)
13
where κ = k – K and νF is the electronic group velocity, which is given by νF
= √3 γ0a/(2ħ) ≈ 106 m s–1.
The linear dispersion given by equation (2) is the solution to the following
effective Hamiltonian at the K(K′) point H = �±ħνF (σ • κ), where κ = –i�
and σ are the pseudo-spin Pauli matrices operating in the space of the
electron amplitude on the A–B sublattices of graphene [5].
2.1.2 Optical properties
Graphene has shown also amazing optical properties, which combined with
its mechanical one make it a promising material for optoelectronic devices.
The linear dispersion of the Dirac electrons makes broadband applications
possible.
The optical image contrast can be used to identify graphene on top of a
Si/SiO2 substrate. This is possible because of interference between the
different layers, and with SiO2 acting as a spacer. The effect depends on the
number of layers.
Starting from the Fresnel equations in the thin-film limit is possible to
derive the transmittance of a freestanding SLG (single layer graphene) :
! = 1+ 0.5!" !! ≈ 1− !" ≈ 97.7% (3)
With α ≈ 1/137 the fine-structure constant. The high transmittance of
graphene is remarkable, as much as its low reflectivity, since it only reflects
<0.1% of the incident light in the visible region. Many layers stuck together
have a higher reflectivity, around 2% per layer. Because of this, we can take
the optical absorption of graphene layers to be proportional to the number of
layers.
We can take each layer absorbing A ≈ 1 – T ≈ πα ≈ 2.3% over the visible
spectrum.
14
While in a few-layer graphene (FLG) sample, each sheet can be seen as a
2D electron gas with little perturbation from the adjacent layers, making it
optically equivalent to a superposition of almost non-interacting SLG.
The absorption spectrum of SLG is quite flat from 300 to 2,500 nm with a
peak in the ultraviolet region (~270 nm), due to the exciton-shifted van
Hove singularity in the graphene density of states [5].
One of the biggest issues for graphene researchers is the fact that graphene
has no “band gap,” meaning that its conductive ability can’t be switched “on
and off” like that of silicon [3].
Chemical doping of graphene is one way to get around this problem,
together with producing bilayared graphene with a different symmetry,
changing the lattice orientation of the two layers.
One more option is to grow carbon nanotubes along different orientations of
the lattice, which brings as well to different bandgaps depending on the
chosen direction of growth.
2.1.3 Other Properties
Many works investigated other properties of graphene, beside the electronic
and optical ones, with equally surprising results, as very well reported in a
review by A.K. Geim [1].
The first ones to mention are graphene mechanical and thermal properties.
Graphene has a breaking strength of ~40 N/m, reaching the theoretical limit.
It has a room-temperature thermal conductivity of ~5000 Wm–1 K–1 and a
Young modulus of ~1.0 TPa.
Neither the melting temperature nor the order of the phase transition is
known.
The one-atom-thick foil of graphene showed also to be impermeable to
gases, including helium, opening the possibility to use it also for bio-
chemical applications as a filter. It also showed a high sensitivity to its
chemical environment, which suggests one more possible application as a
chemical sensor, for lab-on-chip technology.
15
Only a few of the other possible applications of graphene are: touch screens,
photovoltaic solar cells, graphene-ink for ink jet printing, gas diffusion
barrier, strain sensor, photo detectors, light emitting devices, ultrafast
tunable wavelength lasers, and many more keep being discovered and
studied by researchers.
With such good properties, one would expect an exorbitant price for such a
material.
Instead, one of graphene most exciting features is its cost.
Every few months, researchers develop new, cheaper methods of mass-
producing graphene and experts predict prices to eventually reach as low as
$7 per pound for the material . The thinnest, strongest material in the
universe may be closer to commercial applications than initially imagined
[3].
2.1.4 Methods of Synthesis
We mention here only some of the possible methods for the production of
graphene, and will go in depth only of the one that was used in this work.
The first method used to produce graphene, the one that the winners of
Nobel Prize in 2010 for its discover used, is mechanical exfoliation. It
consists basically in peeling off of a bulk of graphite the thinnest possible
layers, with adhesive tape. Other techniques are micromechanical cleavage,
carbon segregation from metal or SiC, deposition from Hydrocarbon gas on
metal, liquid phase exfoliation (graphite is put in a solution and exfoliation
happens applying ultrasounds to the solution) and chemical synthesis on
substrate , otherwise called CVD, chemical vapor deposition.
This is the technique that was used in this work.
16
2.2 CVD
2.2.1 Process of growth
Reactant gases are introduced in the chamber, and chemical reactions occur
on wafer surface leading to the deposition of a solid film.
The process starts introducing reactive gases into the chamber.
The gases are activated or decomposed by a source of energy which can
either be heat or plasma. The gas is adsorbed by the substrate surface. The
reaction takes place on the surface, leading to the deposition of the film on
top of the surface substrate.
The chemical reaction is exactly what makes the growth happen, exploiting
the energy provided to lead the molecules available to be deposited to
arrange on top of the sample surface forming a film.
The reaction can also produce byproducts, which are transported away from
the substrate together with the exhaust gases.
CVD is a very widely used technique, deposited films ranging from metals
to semiconductors to insulators. It is one of the premier techniques for
epitaxial growth of thin layer structures (semiconductors, oxides,
superconductors).
It also has a wide application for devices such as Lasers, LEDs, solar cells,
photo-detectors, HBTs, and FETs.
In each CVD procedure, a specific chemical reaction has to be studied, and
different parameters are involved. Not any chemical reaction on any type of
substrate will lead to an efficient deposition.
The reaction that was used for this deposition is thermal dissociation of
methane, and consequent deposition of nanocystalline graphene on quartz
and sapphire.
The details of the chemical reaction and the features of the resulting
graphene films were characterized in a previous work in the same
laboratory. The next section briefly summarizes the main results.
17
The data collected suggest that these films are nano-polycrystalline with the
crystallite size varying from 10 to 30 nm. The thinnest of these films have
thickness in the range of 1 nm, they are as transparent as 1–2 layer graphene
and their conductance is about 5x10-5 siemens. The charge carriers in these
films are holes and the hole mobility is a few tens of cm2/Vs [11].
2.2.2 Types of CVD
Chemical vapor deposition uses a precursor of the material to be deposited,
which can be from gas, volatile liquid or solid.
When CVD is performed from solid the target of material to deposit is
sputtered through a plasma.
The type of CVD that we used instead uses a gaseous precursor, which is
decomposed via heating.
The reaction can either happen in atmospheric pressure or at low pressure.
The atmospheric pressure one is cheaper, easier to realize and faster, but
produces less quality films with a worse coverage of the surface.
The low pressure CVD (gas pressure of a few torr) is slower, and requires a
good control of the temperature in the chamber, but films turn out of a much
better quality and a good coverage of the substrate is achieved.
In some types of CVD, the material can be dissociated via plasma or
photochemical CVD.
As mentioned above, the source of energy that allows the reaction to happen
in our case is heat. In most cases, as in ours, the heat was produced by a hot
filament, within or around the chamber.
It is possible to have the walls of the chamber heated up at the same
temperature of the substrate (hot wall reactors) or have the walls kept at a
lower temperature than the substrate (cold wall reactors).
In this work the CVD performed was in low pressure and with hot walls
reactor.
In our furnace a hot filament was running through the walls heating up all
the different areas uniformly.
18
The main pro of cold wall reactors is that deposition does not happen from
the walls of the chamber, which are not heated up, but only from the
precursor material.
This problem was avoided in our chamber since each part of it is graphite-
made, in this way any contamination from the walls of the chamber is
avoided. The pro of the hot wall reactor is that no gradient of temperature is
present, avoiding inconvenient convections in the chamber which can lead
to non uniform films.
2.2.3 Pros and Cons
Here we list briefly the pros and cons of this technique of growth.
Main advantages are the following ones: i) high growth rates are possible,
with a good reproducibility; ii) with this technique it is possible to deposit
materials which are normally hard to evaporate, through sputtering
performed with a plasma; iii) CVD allows to grow epitaxial films, and in
this case is also termed as “vapor phase epitaxy (VPE)”. For instance one
can grow also organic films with MOCVD (metal-organic CVD) which is
also called OMVPE (organo-metallic VPE); iv) it gives generally better film
quality, and a more conformal step coverage.
A few drawbacks have to be noticed: i) high process temperatures are
required; ii) it is a complex process, sometimes requiring toxic and
corrosive gasses; iii) exactly because of such gasses, which can be used as
catalysts, the films grown may not be pure (hydrogen incorporation…) [41].
2.3 Ion irradiation, implantation and sputtering
The basic process consists in the introduction of atoms in a solid phase,
through the ionization of a certain atomic species and its acceleration with
enough energy to cause the penetration in the solid.
Depending on the energy of the beam and dose of the atomic species the
damages produced in the solid are of very different entity.
19
This is because the ions will penetrate at different depths, depending on the
energy of the beam, and will cause collisions of different entity.
It is a very convenient method overall because it allows to control the
number of ions, the position and depth of the implantation.
2.3.1 Implantation
The implanter used in this work was a FIB, Focused Ion Beam, through
which the ion implantation was performed. FIB is described in detail in the
next chapter.
A very important parameter to characterize the implantation is the dose. It
expresses the number of incident ions in a unitary surface, and it indicates
the concentration of ions introduced.
Its unit can be expressed in atoms/µm2. Sometimes the level of doping can
also be expressed in number of impurities/volume (atoms/µm3) or in
Coulombs/µm2.
! = !"#$!!! ~
!"#$%!"! ~ !"
!!! ~!!"#$×!!×!
(4)
With the assumption that all ions within the beam have the same electrical
charge, the number of ions is proportional to the beam current integrated
over the time of exposure to the ion beam.
Typically the implanted ions are O2, N2, B, P, As, Sb at a typical density
between 1015-1017 atoms/cm2. These species though are very rarely used in
FIB systems, and more common for other implantation systems. The most
common specie used in FIB is Gallium.
2.3.2 Implantation profile
One very important and useful function to calculate is the implantation
profile.
Normally the crystal is considered to be isotropic and that no channeling is
happening. We will see later how this is not the case in our work, but this
20
function is still very important to estimate and still gives a good
approximation of the phenomenon studied.
The phenomenon of channeling is observed in ions moving in certain
directions in a crystalline material, where there are long range open spaces
through which the ions can travel without significant scattering. The
presence of these open spaces depends on the symmetry and crystalline
structure of the crystal.
During implantation, the high-energy-ion penetrating in the substrate
undergoes a series of stochastic collisions which deviate its trajectory, force
it to decelerate and in the end stop it.
Picture 5 Monte Carlo calculation of 128 ion trajectories for 50 keV boron implanted into Silicon,
taken from [41]
During the implantation the ion can produce two types of collisions: ion-
electron collision and ion-nucleus collision.
In the ion-nucleus collisions the interaction is dominated by the Coulombian
repulsion between the ion and the nucleus of the crystalline specimen. This
type of collisions causes abrupt deviations and decelerates the ions. These
collisions are the main reason for the displacement of the nuclei of the
21
crystal, thus the main cause of disorder and irradiation damage in the
crystal.
The collisions between ions and electrons only slow down the ions and
make them deflect slightly. The interaction is between the ions and the
electrons around the nuclei and the free conduction electrons on the
substrate. This type of interaction does not produce any crystal damage.
We introduce now some parameters to calculate the implantation profile.
SN and Se are the nuclear stopping power and the electronic stopping power.
The nuclear stopping power depends on the atomic mass and atomic number
of the incident ions and the atoms of the substrate.
!! ∝ !!!!
!!! !!!!
! !
!!!!!!!!!
(5)
Where Zi and Mi are the atomic number and the atomic mass of the incident
ion (i=1) and of the substrate atom (i=2).
The electronic stopping power depends on the square root of the ion energy:
!! ∝ ! (6)
The implantation profile N(x) is Gaussian, because of the randomness of the
collisions.
! !! = !!!∆!!
(7)
Rp is the average ionic penetration. The Gaussian has a standard deviation of
∆Rp.
Q is the quantity of ions introduced over a unitary area of the substrate.
The following equation is useful to calculate the average total distance that
the ions can travel in the bulk, before completely losing their energy and
stopping:
22
! = !"!! = !"
!! ! !!! !!!! (8)
An example of the Gaussian profile and its dependence on the beam energy
is shown in the graph:
Picture 6 Boron implanted atom distributions, comparing measured data points with Gaussian
fitted distributions, taken from [41]
The depth profiling of a Gaussian is centered at a certain distance from the
surface, typically fractions of micrometers inside the sample.
2.3.3 Damage
The ionic damage is provoked by ions over a certain dose and energy, as it
was also observed in this work.
The damage is caused by the collisions of ions that cause a displacement of
the atoms from their positions of equilibrium in the lattice.
High energy impacts produce a limited damage in the first superficial layers:
the thermalization of the incident ions happens mainly when the ion energy
is low and consequently the damage is localized close to the maximum of
the implantation N(x).
Regarding the recoil of the atoms, the number of substrate atoms pushed
back by the impacts grows with growing energy of the incident ions.
23
Usually, damage caused by ion implantation includes:
1) formation of crystal defects such as Frenkel defects, vacancies, di-
vacancies, higher-order vacancies, and interstitials;
2) creation of local amorphous regions included in the crystalline structure;
3) formation of continuous amorphous layers as the localized amorphous
regions grow and overlap.
Damage types 1 and 2 are categorized together as 'primary crystalline
damage'.
Normally the irradiation is followed by an annealing in order to eliminate
the damage and activate the dopant making it diffuse, if needed.
2.3.4 Annealing
Annealing is a process where the wafer is heated to repair the damage of the
lattice.
It has also the effect to activate the ions implanted as dopants, if they can
interact with the substrate and create new energy levels, and has also the
function of inducing dopant diffusion into the implanted target.
This second function of annealing was not applicable in our work, since
Ga+ ions were implanted on graphene, and these two materials do not
interact, Gallium is an inert material in graphene, electronically and
optically. This is why we focused more on the first function of annealing,
which is restoration of the broken bonds and of the pristine atomic structure
of graphene.
Primary crystalline damage annealing basically consists of:
1) recombination of vacancies and self-interstitials in the low temperature
range (up to 500 °C);
2) formation of dislocations at 500-600 °C which can capture impurity
atoms;
3) dissolution of these dislocations at 900-1000 °C.
24
Annealing of the continuous amorphous layers that extend to the surface has
been shown to occur by solid-phase epitaxy between 500-600 °C. Under this
phenomenon, the crystalline substrate beneath the amorphous layers initiates
the recrystallization of the amorphous layers, with the regrowth proceeding
towards the substrate surface. Factors affecting the recrystallization rate
include crystal orientation and the implanted impurities. Amorphous layers
that do not extend to the surface anneal differently, with the solid-phase
epitaxy occurring at both amorphous-crystal interfaces and the regrowth
interfaces meeting below the surface.
One more phenomenon that for sure cannot be restored through annealing is
the sputtering of the surface.
2.3.5 Sputtering
The picture below shows the main events that occur during ion irradiation of
a substrate (Picture 7).
The primary beam consists of Gallium ions, that are implanted in the target
generating defects and creating disorder in the crystalline structure.
When the energy of the incident ion is high enough it can happen that it hits
an atom of the specimen so hard to make it exit from the bulk.
When this phenomenon happens it leads to the corrosion of the bulk,
sputtering away considerable quantity of the material exposed to the beam.
25
Picture 7 Example of the events occurring during ion implantation, taken from [42]
Sputtering is a process whereby atoms are ejected from a solid target
material due to bombardment of the target by energetic particles. It only
happens when the kinetic energy of the incoming particles is much higher
than conventional thermal energies (� 1 eV). This process can lead, during
prolonged ion or plasma bombardment of a material, to significant
corrosion, and can thus be harmful. On the other hand, it is commonly
utilized for thin-film deposition, etching and analytical techniques.
Picture 8 Example of sputtering of the target by ion beam, taken from [42]
A good parameter in the study of sputtering is the etching rate.
26
Three different “Etching Rate” definitions are encountered in FIB literature:
!"#$%& !"# !"# ~ !!!
!"
!"#$%& !"# !"#$ !" !"#$ ~ !!!
!
!"#$ℎ !"# !"#$ !" !"#$ ~ !"!
The sputtering yield is given by ! = !"#$%"& !"#$%!"#$%&"' !"#$
= !"!!
(9).
It depends also on the angle of incidence.
Depth – Dose Relationship:
As a first-approximation depth of FIB etching is directly proportional to the
“Dose” of ion beam exposure. This approximation holds very well if:
- Actual depth is less then x3 of the narrowest dimension of the milled
area;
- Etching is done at 90° incident angle;
- Etching is done without any reactive gas
The phenomenon of channeling was mentioned in the previous paragraph. It
normally is omitted when calculating the implantation profile, but it actually
is relevant in our case.
The effects of channeling are well presented in Picture 9, which gives an
example of how changing simply the orientation of the same crystalline
structure the sputtering yield can change by far.
27
Picture 9 Dependece of Yield on angles of orientation, effect of channeling, taken from [42]
In our work in particular, channeling has a role since graphene on our
samples is of nanocrystalline type.
This means that the graphene layers consist mainly of crystalline grains.
Because of this, graphene layers do not have in general the same crystalline
orientation, and in this sense each and every grain has a different
orientation.
Because of this, it is impossible to evaluate in which direction the ion
channeling can happen, in order to avoid it, since in every crystal grain it
will be different.
On the other hand this fact can be convenient, since in order to have a full
channeling all the crystals should be oriented in the same way along the ion
trajectory, which is concretely almost impossible for the presence of the
grains.
Picture 10 Effect of channeling on different crystal grains taken from [42]
28
Another effect that was mentioned in the previous paragraph, as a
consequence of ion irradiation at high energy is complete amorphization of
parts of the sample.
The secondary electron yield and the atoms sputtering yield depend as well
on the orientation of the sample and on the energy of the beam, which
influences the implantation profile, and thus the area of the sample that is
interested in the ion irradiation.
This is well shown in Picture 12.
Picture 12 Example of dependence of secondary electron yield and sputtering yield on the orientation of the crystal, taken from [42]
One more effect that can happen during the irradiation is the sample
charging.
Picture 11 Example of amorphization of the sample with FIB irradiating the sample from the right, taken from [42]
29
The impact of ions produces charged particles. Some of them leave the
sample, while some of them do not have enough energy to exit and are
trapped in the sample, charging it and generating a current. This follows
Kirchoff’s Law, which applied for FIB is:
!!"#$ + !!"#$%&'() !"!#$%&'( − !!"#$%&'() !"#$ = !!"#$%& > 0 (10)
Due to impossibility of balancing current of the primary ion beam by
currents of secondary charged particles,, any un-grounded sample subjected
to FIB irradiation will accumulate positive charge. Under continuous ion
beam irradiation, the charge will keep accumulate until electric breakdown
creates a discharge path.
It is possible to solve this problem by charge neutralization, with an electron
gun dispensing electrons to compensate the positive charge of the
backscattered ions, or with biases placed around the chamber (negative bias
where positive ions will be sputtered back and bias applied to the sample).
Sometimes it is possible with certain instruments (with FIB for example) to
perform gas-assisted etching when the beam itself is not able to provide the
energy and depth necessary for the requested etching characterization of the
sample. This was not the case in our work.
2.3.6 Pros and Cons
Some of the advantages of this technique are the possibility to introduce any
type of impurity/ions in different substrates and to control accurately the
quantity of doping introduced together with control of the speed of the ions
and the depth of penetration.
It is possible to set the thickness of the irradiated layer and independently to
set also the implantation dose.
It gives a good uniformity and reproducibility of the operation.
30
This procedure allows us to set and control the implantation profile setting
the energy of the beam, and also to use photoresists in order to shield the
atoms in parts of the sample.
The operation can be performed at room temperature.
The disadvantages consist in the fact that it is hard to control the quality of
the implantation. The procedure causes damage of the crystalline structure
which has to be recovered with annealing, and it is not always possible to
fully recover.
When the irradiation is performed with masks it is possible that some ions
will pass through the mask and be implanted. What also needs to be
considered is the fact that the equipment is complicated and expensive.
“At present disorder at the atomic level produced by ion implantation in
Silicon is not understood in detail. Electron diffraction studies agree with
the interpretation above that ion implantation drives the silicon lattice
toward an amorphous-like structure, however the details of this transition
are difficult to ascertain.”
This is taken from the book Silicon VLSI Technology by Plummer, Deal
and Griffin, [41], which studies in detail the case of Silicon, but its
conclusions can be reasonably extended also to other materials, at least in
their general assertion.
31
Chapter 3 Equipment 3.1 Furnace
The main parts of the furnace are the heater, the sample holder, the
electrodes, the external vacuum chamber, the cooling system and the inlets
for gas and the pressure gauge.
The interior of the furnace is made of graphite (heater, sample holder and
electrodes), in order to avoid contaminations by other materials during CVD
growth. The external vacuum chamber is made of stainless steel.
The furnace is connected to a vacuum system, which keeps it in the pressure
range of a few millibars.
From a gas cylinder the precursor gases for the growth are injected in the
chamber, through a pipe connected to it, controlled by a valve.
Picture 13 Graphite and stainless steel furnace
32
For this type of growth the gas is methane, but also hydrogen or other gases
can be introduced in the chamber to enhance or suppress the growth, or
clean the chamber.
The heating system is connected and powered by the Sorensen DC power
supply DCR 20-125.
Through applying a certain potential difference (from 0 to 2,5 V with an
error of 0,25 V) to the heater (a graphite filament) it produces a current
(measured in Ampere and displayed on the same Power Supply) with an
error of about 2,5 A. It is possible to control the voltage with coarse and fine
adjusters.
The chamber is also connected to a pressure gauge that allows to control the
pressure in the chamber.
Two different pressure gauges were used during the experiment: Omega
DPG 3500B-2000MBARA [pressure range 0-1000 mbar] and Adixen ACC-
2009 [pressure range 5E-9 to 1000 mbar]. Temperature was measured with
a temperature controller and wire thermocouple.
The display with which readings of vacuum and temperature were made had
four digits for the temperature, and an error of 1°C and three digits for
pressure with an error of 10-2 mbars. The thermocouples and gauge errors
were both lower.
3.2 Vacuum System
The vacuum system consists of two pumps, that work in series.
The first one is a rotary vacuum pump, Two stage Edwards E2M2.
The second one which works at a higher vacuum regime is a
Turbomolecular pump, CFF 450 Turbo Alcatel Adixen ACC 2009.
Rotative pumps are mechanical pumps which work in idrodinamic regime,
able to produce 1.3x10-2 - 1.3x10-3 mbar vacuum, starting from atmospheric
pressure.
33
They basically consist of a rotor mounted on the same ax of the rotor of an
electronic motor.
The two rotors are at the center of a cylindrical chamber, in which the gas is
injected.
Two or four palettes are mounted on the rotor that firmly adhere to the
internal side of the chamber.
They have the function to “wipe” the volume of the chamber and compress
the air and push it toward the exhaust valve of the pump, that vents it out of
the chamber. All the parts are immersed in silicon oil that has the function
of lubricant and sealant between the rotor and the chamber stopping the
back diffusion of the pressurized gas toward the aspiration area.
The second pump used is a turbomolecular pump. It is a versatile type of
pump, since it can generate different degrees of vacuum, from intermediate
(~10-4 mbar) up to ultra-high vacuum (~10-10 mbar).
The physical principle it uses is furnishing moment to the gas molecules,
through rapidly spinning blades. The molecules in this way are pushed out
of the chamber, lowering the pressure of the gas.
What furnishes the momentum to the molecules is the repeated collisions
with the blades, that increases the speed of the molecules that are directed
from the inlet of the pump towards the exhaust in order to create or maintain
vacuum.
The gas molecules pass through different stages, all with the same working
principles, mounted in series. Once the gas passed from the first stage it is
directed toward the next one in which an higher vacuum is obtained by the
different size, shape and orientation of the blades. This process is continued,
finally leading the molecules of gas outwards through the exhaust.
Performance of a Turbo Molecular Pump (TMP) is strongly related to the
frequency of the rotor. As rpm increases, the rotor blades deflect more.
Turbomolecular pumps must operate at very high speeds, and the friction
heat buildup imposes design limitations. At atmospheric pressure, the mean
free path of air is about 70 nm. A turbomolecular pump can work only if
34
those molecules hit by the moving blades reach the stationary blades before
colliding with other molecules on their way. To achieve that, the gap
between moving blades and stationary blades must be close to or less than
the mean free path.
3.3 FIB
FIB stands for focused ion beam. The basic mechanism of this machine is
very simple.
It has an ion source, from which ions are extracted, focused through the
main column by lenses into a concentrated beam, to hit a target.
In the beam the energy of the ions, the sharpness of the beam and density of
ions can be calibrated, together with many more features.
The beam is directed toward the sample on which the ions are meant to
impact.
The area of the sample irradiated the beam can be changed by deflecting the
beam by octopole electrodes present in the column. Size of the area of
Picture 14 Elements of primary column in an ion beam, taken from [42]
35
irradiation can be set as well by changing amplitude of deflection.
Magnitude of ion beam current is defined by changing excitation of the
condenser lens and measured at the blanking aperture.
A detector using the secondary electrons produced by ion beam interaction
with the sample allows to obtain images of the area where the beam is
rastering on the sample.
In the images are shown the main components of the ion column (Picture
14), and the block diagram of the whole system (Picture 15).
The sample, together with the sample holder and the whole system is kept in
a high vacuum (10-7 torr).
Different vacuum pumps and controllers are involved. In particular the
specimen holder can be pulled out in order to introduce the sample, and
inserted back in. Sample exchange is the main reason for different chambers
to exist, at a different vacuum levels. Multiple-chamber configuration
allows to keep the beam and the rest of the system under high vacuum while
the sample is introduced into the chamber, from atmospheric pressure.
Picture 15 Block diagram of Micrion 2500 FIB, taken from manual
36
The system of vacuum pumps and gauges, the ion beam column, the
detector and the imaging system are controlled by a computer, from which
is possible to set all the different parameters and
control the irradiation.
The source of Gallium Ions is a liquid metal.
The metal is placed on top of a tungsten needle.
The needle is heated by a tungsten wire coil that
runs around the needle. In order to heat up the
coil a current runs through it, and this current can
be easily regulated.
The heat melts the metal drop on the needle, and
the ions are attracted from the source by high-
strength electric field of the extractor electrode. Since the metal has to be
melted normally a material with a low melting point is chosen.
In order to extract the ions a potential difference is applied between the
electrodes, one of them is the needle itself and the other is a metal plate
toward which the ions are attracted. This second electrode is called the
extractor electrode. Potential applied between the needle and extractor
generates en electric field that influences the shape of the drop of metal on
the needle. In the picture is depicted the difference of the shape of the liquid
Ga surface at the tip of the needle depending on the applied field (picture
17).
Picture 16 Gallium ion source, taken from [42]
37
Picture 17 Dependence of shape of liquid Ga on the applied field, taken form [42]
Through all the column of the FIB there are electrostatic lenses and
blanking plates with the function of focusing and directing the beam of ions.
The parameters that can be set for the beam are many, in particular to
calibrate the number of ions one can change both the final aperture and
potential of the condenser lens to define the ion beam current. The aperture
changes the size of the beam and potential of the condenser lens changes the
current the density of ions in the beam. Two beams with same aperture can
have very different number of ions, depending on the potential applied to
the condenser lens. Both the beam current and the final aperture can be set
from the computer that controls the FIB system.
The octopole has the function of correcting astigmatism by shaping the ion
beam. Shape correction is achieved by applying same polarity potentials to
the opposing plates of the octopole, thus either compressing with positive
potential or expanding it with negative potential. Such shape correction
removes ellipticity from the ion beam, thus correcting astigmatism.
The deflector plates have the role to change the direction of the beam, and
use as well an electric field of different polarity applied to opposite plates of
38
octopole, thus pushing ion beam away from the positive plate and attracting
it to the negative.
The sample can also be tilted, so that implantation of ions can happen also at
an angle different from the normal.
As mentioned, the whole system is kept in vacuum. The specimen is
introduced in a secondary chamber, placed on its holder and the pushed
forward, under the beam. From here, once the sample gets on the stage, it
can be moved in all directions and tilted.
The computer controls the position of the stage, so that it’s possible to
calibrate exactly its movements. In this way orientation on the sample is
possible also without using the function of imaging, available on Micrion
2500 FIB only through primary ions. The impact of ions during the
observation of the sample can damage the surface, so normally when
possible it was avoided during our experiment.
In most FIBs though a second column of electrons is present, specifically
with the function of imaging (Picture 18). The system of imaging is based
on secondary electrons (it is so even when the imaging happens through
bombardment of the sample by ions, but only the electrons are detected
from the detector and used to depict the image). The principle is the same
one used in SEM microscopes.
Picture 18 SEM layout and function, taken from [42]
39
Also a system to neutralize current in order to avoid damaging of the sample
when it’s overcharged is present in the FIB system. There are two main
ways to neutralize the current that runs through the sample. One of them is
through an electron gun that injects electrons directly on the sample (since
as explained in chapter 2 the current on the sample is always positive, as a
consequence of Kirchoff’s law), or through applying a bias to the sample
[42].
3.4 Low Pressure Plasma System
The low pressure plasma system that was used in order to treat with plasma
the samples, is produced by Diener electronic. It is the PICO plasma-
surface-technology model, with vacuum pump trivac D2,5E.
This machine can be used for different purposes, like cleaning of surfaces,
activation of surfaces, etching of surfaces and deposition of surfaces-
plasmapolymerization. [Manual of the instrument]
As will be explained more in depth in chapter 5, in this work the plasma
treatments were performed to see if the growth of graphene could be
enhanced or suppressed through such treatments, in order to characterize the
surfaces, isolating the not treated areas covering them with a photoresist.
A plasma can be created by heating a gas or subjecting it to a strong
electromagnetic field applied with a laser or microwave generator. This
decreases or increases the number of electrons, creating ions, and is
accompanied by the dissociation of molecular bonds, if present.
There are several means for plasma
generation, however, one principle is
common to all of them: there must be
energy input to produce and sustain
it. For this case, plasma is generated
when an electrical current is Picture 19 Artificial plasma produced in air by a Jacob's Ladder
40
applied across a dielectric gas or fluid (an electrically non-
conducting material) as can be seen in the image to the left, which shows
a discharge tube as a simple example (DC used for simplicity).
The potential difference and subsequent electric field pull the bound
electrons (negative) toward the anode (positive electrode) while
the cathode (negative electrode) pulls the nucleus. As the voltage increases,
the current stresses the material (by electric polarization) beyond
its dielectric limit (termed strength) into a stage of electrical breakdown,
marked by an electric spark, where the material transforms from being
an insulator into a conductor (as it becomes increasingly ionized). The
underlying process is the Townsend avalanche, where collisions between
electrons and neutral gas atoms create more ions and electrons. The first
impact of an electron on an atom results in one ion and two electrons.
Therefore, the number of charged particles increases rapidly (in the
millions) only "after about 20 successive sets of collisions", mainly due to a
small mean free path (average distance travelled between collisions).
This is also the way in which the plasma is produced by the generator used
in this work.
The chamber contains two metal plates, about 4 inches wide and 16 inches
long.
They are kept in a medium vacuum (between 0,1 to 10 mbars) and the
sample is placed on the underneath one, on top of an holder.
It is possible to connect the machine to two different gas cylinder.
Once the sample is placed inside the vacuum rotary pump is turned on and
once a vacuum between 0,2 and 0,4 mbar is reached the gas is let flow
inside, and the power is turned on, in order to create an electric field
between the plates that generates the plasma.
41
3.5 Conductance measuring system
The conductance measuring system consists of a working table (Went
Worth Labs), with a stage in the center where the samples are placed in
order to perform electrical measurements on them (Picture 20)
Supports for the two tungsten needles are installed on the table.
The needles can be moved in all directions, be tilted, raised and lowered.
The stage as well can be set at different heights. This system gives a full
mobility and allows to take measurements in a wide area of the stage and
with different needles distances.
Picture 20 Working table, with needles and needle holders and microscope
On top of the stage is placed an optical Bausch&Lomb Microscope, which
allows a better view of the sample and of the contacts with the needles.
42
The electrical conductance was measured using Keithley 2300
Semiconductor characterization system analyzer placing the tungsten
needles on the sample surface with identical force.
The distance between the needles tips touching graphene was about 20
microns. Most of the measurements were performed at a voltage of 1 V.
Through the computer is possible to set the voltage, start the measurements
and read on the monitor the level of conductance on the sample.
3.6 Microscopes
Microscope Olympus SZ30 was used for three dimensional observation of
the sample in order to verify the cleanness of the surface before deposition
and for quick and three dimensional observation of the samples in general. It
is a stereo microscope.
The stereo or stereoscopic or dissecting microscope is an optical
microscope variant designed for low magnification observation of a sample,
typically using light reflected from the surface of an object rather than
transmitted through it. The instrument uses two separated optical paths with
two objectives and eyepieces to provide slightly different viewing angles to
the left and right eyes. This arrangement produces a three-
dimensional visualization of the sample being examined.
The model Olympus SZ30 is developed primarily for mounting onto
bonders and probers, offers excellent cost performance. The combination of
a long working distance (110 mm) and large zoom range of 0.9X to 4X
means the user can expect superb optical performance and maneuverability.
It is provided with Auxiliary Objectives 0.3x, 0 4x, 0 5x, 0.75x. 1.5x, 2x.
[Manual of the instrument]
Optical microscope Nikon Eclipse LV100- Marrell is an Advanced
polarized light microscope that works under both diascopic and episcopic
illumination. This more precise instrument was used to make more detailed
43
observation of the samples, in particular to identify the irradiated areas of
graphene. The lenses are 5X/0,15A 10X/0,30A 50X/0,80A 100X/0,90A,
all installed together on the microscope.
Nikon's Eclipse polarizing microscopes are renowned for their ability to
produce brighter, clearer, and higher contrast images. The LV100 POL,
available in diascopic and episcopic microscope illumination types,
continues this tradition and offers a completely reengineered base for even
easier operation. It also features an exclusive high-intensity halogen light
source, which delivers brighter images, lower power consumption and less
heat generation, thereby reducing the chance of heat-induced focus drift.
Only some of the key features of this instrument are: the diascopic
(transmitted) light source, lead and arsenic free objective lenses, reverse
type nosepiece, Bertrand Lens incorporated into design of intermediate tube,
reduced focus drift and many more [43].
44
Chapter 4 Growth of nanocrystalline graphene 4.1 Experimental procedure for growth
The method used to grow graphene layers on quartz substrate is Chemical
Vapor deposition, CVD.
The substrates are single crystal quartz plates of size 4 mm×4 mm×0.5 mm
with the large surfaces cut perpendicular to the z-axis and polished to a
roughness of Ra < 1 nm (commercial product of MTI company).
The quartz is cleaned in acetone in ultrasonic bath for five minutes and,
when needed, wiped with cotton soaked in acetone to mechanically remove
dust.
After cleaning it, the sample is placed on the sample holder. As mentioned
before, both the sample holder and the furnace are all-graphite, in order to
avoid contaminations during the growth.
The furnace is opened and the holder is placed inside of it. The furnace is
closed and the pump (Rotary Two stage Edwards E2M2) is turned on, and
keeps working until the furnace reaches a vacuum of at least 10-2 mbar.
Pressure was measured with two gauges: Omega DPG 3500B-2000MBARA
and Adixen ACC-2009.
Methane is then inflated in the chamber and let flow through it for about 30
seconds, to make sure that no other gas except methane is present in the
chamber during the growth.
Then, the methane in the chamber is pumped out, down to the pressure
chosen for the growth (normally between 1 and 5 mbar). Through the
heating system (Sorensen DC power supply DCR 20-125) the temperature is
raised, up to the value chosen for the growth (normally between 900°C and
1200°C). In these conditions of pressure and temperature, methane
decomposition occurs and nanocrystalline graphene layers start to grow on
top of the substrate.
The temperature is kept constant during the growth. To stop the growth the
heating system is quickly turned off and the chamber cools down rapidly.
45
When the temperature is low enough (around 200°C), air is inflated in the
chamber and the sample can be taken out and measured. To prevent the
damage of rubber parts of the vacuum system, these are constantly cooled
down by a water cooling system.
Electrical conductance of graphene films was measured at room temperature
using a Keithley 2300 analyzer and tungsten needles placed on the sample
surface with identical force. Since we had to perform hundreds of
measurements in different places of many samples, we practically could not
use a four probe method with deposition of metal electrodes. Instead, a
simple two probe method was used. The distance between the needle tips
touching the graphene surface was about 20 microns.
The apparatus is connected to a computer in which the KYTE software
allows us to set a bias between the needles and control the current passing
through them when they touch a conductive layer. Most of the
measurements were performed at a voltage of 1 V. This voltage was low
enough to avoid non-linear effects in the current flow. The level of
sensitivity of the measurements was in the range 0.01 to 0.1 pA. The quality
of the electrical contacts between the tungsten tips and the graphene layer
was not always good and that could cause considerable fluctuations of the
current. After having performed many measurements on different samples at
different voltages and at different conditions of placing probes on sample
surface, we estimated the experimental error of the measurements as an
order of magnitude. However, despite this seemingly high experimental
error, the discussion of the electrical measurements is not compromised
since the changes in conductance occurred over many orders of magnitude.
46
4.2 Results
In order to better understand the growth process and which parameters play
a role in it, many growths were performed in different conditions.
The parameters that we focused on in particular are temperature, pressure
and time of growth. They most likely influence the growth of the films in
their quality, thickness and conductance. Other factors that can influence the
growth are the substrate surface roughness, substrate chemical composition
and surface termination. In each growth was possible to choose the
temperature of growth, the pressure of the methane in the chamber and the
time of growth, which determine the methane decomposition and the
deposition of graphene. Our study was performed to see if and how these
parameters influence the growth of graphene.
Growths were performed keeping two of the three parameters fixed and
changing the third one in order to observe the influence of one parameter at
a time.
In this way was also possible to find the conditions in which the growth is
more effective. In fact, judging by the transparency of the layers and the
conductance, it was possible to deduce when a higher quality graphene was
grown, and so record which were the best conditions of temperature,
pressure and time to effect the growth.
4.2.1 Temperature
The first parameter to be studied was the temperature.
In this case the growths were all performed with pressure in the chamber of
5mbar for 15 minutes.
The lowest temperature at which the growth was performed is 900°C. We
considered that a lower temperature did not make sense because the
standard temperature for methane decomposition is around 1200°C on
average. The highest temperature of growth was 1300°C. We did not
47
performed growths at higher temperatures since the threshold was already
found and 1300°C is the average temperature of quartz degradation.
As mentioned before the way to detect a layer of graphene is by measuring
its conductance. The lowest values found are around 10-14S, which is
comparable with the sensitivity of the instrument. When the conductance
reaches a value of 10-6 S or over, it is the clear sign of the presence of
graphene layers.
The conductance was measured on both the top and the bottom side, since
from previous studies we had evidence that graphene was growing also on
the bottom side of the sample.
Between 900°C and 1300°C, a sharp step is observed. For the top side, the
conductance is in the range of 10-14 S in the growths at 900°C, 1000°C and
1050°C.
When the growth was performed at 1075°C, a conductance of almost
6x10-6 S was observed. From this temperature on, the conductance was
between 6x10-6 S up to 7,9x10-5 S.
On the bottom side of the sample, the conductance was around 10-14 S up to
1175°C, at which the conductance resulted to be 2,3x10-5 S.
These results show how the dependence of graphene’s growth on
temperature follows the trend of a sharp threshold. Under a certain
temperature no conductance is present. This could either be the sign of the
complete absence of graphene on the substrate, or islands of graphene could
be present, which are not connected yet and do not result in a high
conductance of the sample.
Over a certain temperature, conductance suddenly increases, revealing the
formation of the first complete layer of graphene. From that temperature on,
the conductance remains stable, slowly growing linearly with the increasing
of the temperature.
The same trend is observed also in the case of the bottom side of the sample,
but with a higher temperature threshold. This is probably because this side
of the sample is less exposed to the methane flow. In fact it is in contact
48
with another piece of quartz or sapphire, to keep it clean and prevent the
direct contact with the graphite support.
In both cases the temperature dependence of the growth has the same trend.
There is a certain temperature at which the growth of graphene starts. This
is probably linked to the quantity of methane decomposed, and consequently
the number of C atoms available for the formation of graphene.
Picture 21 Temperature dependence of growth
4.2.2 Pressure
In order to study how pressure can influence the growth of graphene on
quartz, a series of growths were performed at constant temperature of
1200°C for 15 minutes, at different pressure values, as measured by the
gauge into the chamber, after it was filled with methane, and then pumped
out to the chosen pressure value. The pressure therefore is most likely the
value which indicates the quantity of methane present in the chamber. In
fact, pressure is not given exclusively by the presence of methane, because
little leaks are always possible as well as presence of other gases influencing
the pressure. However, it is a good parameter to indicate quite accurately the
relative quantity of methane from growth to growth.
49
Picture 22 registers, also in the case of pressure dependence, a sharp step
behavior. In this case, the top and bottom side of the sample show the same
trend of conductance.
The growths performed at 0,5 mbar and 0,75 mbar did not show any
presence of graphene deposited. From 2 mbar up, instead, the conductance
was found to be in the range of 10-5-10-4 S. The sudden appearance of
conductance is associated with the presence of a layer of graphene. It is
possible to notice how the increase of the pressure, and so the quantity of
graphene, does not determine a major change in the conductance. We can
affirm that it stays in the same range from 2 to 10 mbar.
This is probably because once there is enough Methane, and consequently
Carbon in the chamber to have a continuous layer of graphene, an increase
in the quantity of Carbon does not increase the conductance of the layer,
since most likely simply more layers are formed, one on top of the other, all
more or less of the same quality and having the same conductance.
Picture 22 Pressure dependence of growth
50
4.2.3 Time
The threshold as a function of the time was studied at three different
pressures: 3 mbar, 4 mbar and 5 mbar. In the case of 5 mbar, the threshold is
between 2 and 3 minutes for the top side and between 3 and 5 for the bottom
side. When the growth was performed with 4 mbar pressure, the observed
time threshold is between 2 and 3 minutes for both the top and the bottom
side. For the growth at 3 mbar the threshold in time is between 3 and 5
minutes for both sides. The temperature was kept stable at 1200°C in all
three cases.
In the time dependence measurements, quartz degradation is observed after
prolonged treatment at high temperatures. The degradation occurs in
particular on the bottom of the sample, which is likely at higher temperature
because of the contact with the graphite support. This phenomenon is
accompanied by the lack of detectable conductance in the samples,
specifically in the growths at 5 mbar for 30 minutes, at 4 mbar for 60
minutes and at 3 mbar after fifteen minutes.
5mbar:
Picture 23 Time dependence of growht, pressure 5 mbar
51
4mbar:
Picture 24 Time dependence of growth, 4 mbar
3mbar:
Picture 25 Time dependence of growth, 3 mbar
52
4.2.4 Pulsed growth
Pulsed growths were also performed, in order to find out at which
temperature graphene starts to grow instantaneously.
The method consisted of a very quick growth, with growth time of about
one second.
Once the sample is placed inside the furnace, starting from room
temperature, the temperature is raised very quickly, increasing the voltage
applied to the heating wire by DC Sorensen Power Supply .
In about ten seconds, the temperature reached the chosen value (between
1200°C and 1600°C) and then voltage was promptly turned down. The
estimated time of reaction to turn off the voltage regulator and reaction of
the machine is estimated to be around one second, time during which the
growth occurs.
This procedure was used to study at which temperature graphene can grow
instantaneously, in order to better control the growth of nanocrystalline
graphene in our furnace.
Different measurements were performed in this way, from 1200°C to
1600°C, finding the growth threshold at 1225°C. Growths were performed
at pressure of 5 mbar.
Picture 26 Pulsed growth
53
We observed in all cases that there are thresholds of temperature, pressure
and time at which continuous conductive film starts to form. The threshold
value of each parameter depends on the other parameters as well as on the
surface roughness, the surface termination and its atomic structure (presence
of defects). Therefore, the growth temperature, the methane pressure and the
growth time could be chosen so that a continuous conductive film formed
only on the surfaces of particular physical and chemical parameters.
The carbon films grown for this research were homogeneous, transparent
and possessed conductance at a level around 10-4 siemens. As shown in
[11], these films can be identified as a few layer nanocrystalline graphene of
a thickness of 1 nm.
54
Chapter 5 Plasma Treatments We carried out a thorough analysis of the effects of plasma treatments on
the substrate to study the possibility of changing the growth of graphene
selectively, either promoting or suppressing it.
If any effect like this was to be found, the specific gas would be used in the
following FIB treatments.
In the FIB Micrion 2500 it is possible to inject a gas together with the Ion
Beam, during the process of ion sputtering. The presence of a gas that could
promote or suppress the growth could be helpful to nanostructure the
graphene film.
Treatments were performed with Argon, Oxygen, Nitrogen, Hydrogen and
Krypton plasmas.
5.1 Procedure
The quartz sample was cleaned in ultrasonic bath for five minutes.
A photoresist was used in order to cover half of the sample. The drop of
photoresist was dried putting the sample on a heated plate for 15-20 min.
When the drop was dry the sample was placed into the Diener electronic
Plasma system PICO.
The chamber was pumped out until reaching a pressure of about 0,2-0,3
mbar. Then the gas was inflated into the chamber up to the chosen pressure
(between 0,15 and 1,75 mbar).
The plasma is ignited turning on the generator, at the power of 85 Watt. The
sample is kept into the chamber for about five minutes.
After the treatment air is let flow into the chamber.
The sample is extracted and the photoresist is removed with an ultrasonic
bath in acetone for five minutes.
Then the growth is performed on the sample as described above. The
temperature was always 1200°C, pressure 5 mbar for 15 minutes.
55
After the growth the sample is extracted and the conductance is measured in
the two different areas.
The presence of the photoresist covered half of the sample, thus it is
possible to compare the growth in the part which was treated with plasma
and the part that was covered, and is intact.
5.2 Results
5.2.1 Argon treatments
1. Argon plasma 0,3 mbar 5 min:
Conductance S Error Top treated with plasma 4,5E-05 ±1E-05 Top covered with photo resist 9,4E-05 ±1E-05 Bottom 3,0E-05 ±1E-05
2. Argon plasma 0,35 mbar 5 min:
Conductance S Error Top treated with plasma ∼E-13 ±1E-13 Top covered with photo resist 9,6E-5 ±1E-05 Bottom ∼E-13 ±1E-13 Note: Area covered with photo resist visible with optical microscope
3. Argon plasma 0,35 mbar 5 min:
Conductance S Error Top treated with plasma 9,9E-05 ±1E-05 Top covered with photo resist 9,5E-05 ±1E-05 Bottom 5E-06 ±1E-06
4. Argon plasma 0,35 mbar 5 min:
Conductance S Error Top treated with plasma 2,6E-5 ±1E-05 Top covered with photo resist 1,1E-4 ±1E-04 Bottom 2,4E-5 ±1E-05
56
5. Argon plasma 0,35 mbar 10 min:
Conductance S Error Top treated with plasma 1,4E-4 ±1E-04 Top covered with photo resist 2,0E-4 ±1E-04 Bottom ∼E-13 ±1E-13
In the treatment with Argon only once, during the second treatment, was
seen an evident difference in conductance between the area treated with
plasma and the pristine one.
The area treated with plasma is completely non conductive, and the one
covered with photoresist shows a conductance of about 1E-4 S.
In the rest of the treatments no important difference in the conductance
between the treated and not treated area is apparent. Because of this the
episode of the only difference of conductance was not judged as relevant
enough to proceed with plasma treatments with Argon during FIB
irradiations.
5.2.2 Oxygen treatments
1. Oxygen plasma 0,35 mbar 5 min
Conductance S Error Top treated with plasma 1,2E-4 ±1E-04 Top covered with photo resist 9,8E-5 ±1E-05 Bottom 1,5E-5 ±1E-05
2. Oxygen plasma 0,35 mbar 5 min:
Conductance S Error Top treated with plasma 8,4E-5 ±1E-05 Top covered with photo resist 8,1E-5 ±1E-05 Bottom 1,7E-5 ±1E-05
3. Oxygen plasma 0,35 mbar 5 min:
Conductance S Error Top treated with plasma 7,1E-5 ±1E-05 Top covered with photo resist 7,3E-5 ±1E-05 Bottom 8E-6 ±1E-06
57
In the case of oxygen plasma no relevant difference is evident between the
treated and not treated area. This result was expected since in the surface
(SiO2) a big quantity of oxygen is already present, and the oxygen plasma
does not change significantly the condition of the surface and the growth on
it.
5.2.3 Nitrogen treatments
1. Nitrogen plasma 0,35 mbar 5 min
Conductance S Error Top treated with plasma 1,0E-4 ±1E-04 Top covered with photo resist 1,0E-4 ±1E-04 Bottom 5,0E-5 ±1E-05
2. Nitrogen plasma 0,35 mbar 5 min:
Conductance S Error Top treated with plasma 2,0E-4 ±1E-04 Top covered with photo resist 1,6E-4 ±1E-04 Bottom 7,0E-5 ±1E-05
3. Nitrogen plasma 0,35 mbar 5 min:
Conductance S Error Top treated with plasma 1,2E-4 ±1E-04 Top covered with photo resist 9,3E-5 ±1E-05 Bottom 3,4E-5 ±1E-05
5.2.4 Hydrogen treatments 1. Hydrogen plasma 0,35 mbar 5 min:
Conductance S Error Top treated with plasma 1,2E-4 ±1E-04 Top covered with photo resist 1,2E-4 ±1E-04 Bottom 8,4E-5 ±1E-05
2. Hydrogen plasma 0,35 mbar 5 min:
Conductance S Error Top treated with plasma 9,1E-5 ±1E-05 Top covered with photo resist 1,2E-4 ±1E-04 Bottom ∼E-13 ±1E-13
58
3. Hydrogen plasma 0,2 mbar 5 min:
Conductance S Error Top treated with plasma 2,7E-4 ±1E-04 Top covered with photo resist 3,7E-4 ±1E-04 Bottom 2,0E-5 ±1E-05
4. Hydrogen plasma 1,75 mbar 5 min:
Conductance S Error Top treated with plasma 1,1E-4 ±1E-04 Top covered with photo resist 1,1E-4 ±1E-04 Bottom 3,1E-5 ±1E-05
5. Hydrogen plasma 0,2 mbar 5 min:
Conductance S Error Top treated with plasma 7,2E-5 ±1E-05 Top covered with photo resist 2,3E-4 ±1E-04 Bottom 4,1E-5 ±1E-05
5.2.5 Krypton treatments
1. Krypton plasma 0,35 mbar 5 min:
Conductance S Error Top treated with plasma 2,1E-4 ±1E-04 Top covered with photo resist 3,1E-4 ±1E-04 Bottom 9,2E-5 ±1E-05
2. Krypton plasma 0,35 mbar 10 min:
Conductance S Error Top treated with plasma 2,0E-4 ±1E-04 Top covered with photo resist 2,5E-4 ±1E-04 Bottom 1,1E-4 ±1E-04
3. Krypton plasma 0,2 mbar 5 min:
Conductance S Error Top treated with plasma 4,3E-4 ±1E-04 Top covered with photo resist 4,5E-4 ±1E-04 Bottom 1,2E-4 ±1E-04
59
4. Krypton plasma 1,5 mbar 5 min:
Conductance S Error Top treated with plasma 9,3E-5 ±1E-05 Top covered with photo resist 1,3E-4 ±1E-04 Bottom 5,4E-5 ±1E-05
In the cases of Nitrogen and Hydrogen no difference is apparent in the
conductance of the areas treated with plasma and the ones not treated with
plasmas.
This is true also in the case of Krypton, in which different attempts were
made with different pressures in the treatments (0,2 0,35 and 1,5 mbar).
Plasma treatments were performed in order to see if the interaction of the
substrate with the plasma could promote or suppress the growth of graphene
selectively.
Unfortunately no relevant effect was observed with any of the gases, so we
decided to perform the FIB irradiations with no plasma injected.
60
Chapter 6 FIB Irradiation 6.1 Experimental procedure The source in the FIB is liquid Gallium ion source. The graphene films were
irradiated with Gallium ions that collided on the graphene layer, in some
cases only partially damaging it, in other cases completely sputtering the
whole film.
The irradiations were performed at various ion energies and doses.
In order to better understand and study the events happening during the ion
irradiations, TRIM software was used. It is a software the produces
simulations of irradiations, giving predictions about distributions of atoms,
vacancies, number of collisions and sputtered atoms, useful to better
interpret our results.
6.1.1 Calculation of the dose
The dose indicates the concentration of ions introduced, it is defined as the
number of incident ions over a unitary surface.
There are more than one parameter that can be changed in order to have
different doses.
The parameter that more obviously regulates the number of incident ions is
the ion beam current.
Primary ion beam current is the current of ions after main aperture, which is
defined by combination of the main aperture diameter and voltage applied to
the condenser lens. This primary ion beam current is the current of ions
which irradiate the sample. When beam is blanked this primary ion beam
current is measured at the beam blanker. However measurements taken at
the beam blanker are not very accurate because of noise. If a more accurate
measurement is required then the beam is un-blanked and directed into the
ion Faraday cup for accurate measurement. This is the procedure that was
followed in our work.
61
With a fixed potential, the area of exposure is inversely proportional to the
dose, in fact with the same number of ions if they are spread through a
wider area the resulting concentration of ions per unit area will be smaller.
We kept the area always the same size, 150×150 µm2, and instead changed
the other parameters.
Because of problems focusing precisely the beam and for unpredictable
shakings and movements of the machine the area is never precisely a square,
but it approximates pretty well the area exposed to the beam.
With the constant flow of beams hitting the sample, the third parameter on
which clearly the dose depends is the time of exposure.
The dose of ions irradiating the target is directly proportional to the time of
exposure to the beam.
In the same treatment, energy of the beam, current, and area were kept
constant, and a line of subsequent squares was exposed to the beam for
various amounts of time.
The formula that was used in order to calculate the dose was:
!"#$ = !∗ !!∗!
(11)
Where I is the Faraday current, t is the time of exposure, Q is the ion
charge 1,6*10-19 C and a is the area of exposure.
Ion irradiation was performed with Ga+ ions using focused ion beam (FIB)
system Micrion 2500. The ion energy varied from 5 to 50 keV at an ion
beam current in the range from 0.6 to 1.3 nA. The doses were in a range
from 1013 to 1016 cm-2 by scanning ion beam.
6.1.2 Preparation and irradiation of the sample The quartz and sapphire samples were cleaned in acetone in ultrasonic bath
for five minutes.
Then the growth of graphene was performed as described in Chapter 4.
62
Only in the case of samples for the study of FIB-induced nucleation
enhancement, the growth of graphene was performed after the treatment
with FIB.
The samples were placed into the FIB and loaded on the stage.
Through the use of the console and the camera placed into the chamber was
possible to move the stage and place the sample right under the beam.
If necessary, before placing the sample under the beam, further
measurements and checking were made, to maximize the focusing of the
beam and set its energy.
To set correctly the beam current, the stage was moved so that the Faraday
cup is placed right under the beam. In order to ensure homogeneity of
irradiation, the ion beam was defocused to a spot of 2 µm. Vacuum in the
sample chamber during ion irradiation was higher than 2×10-7 mbar. Thus
we believe that no considerable contamination of the sample surface
occurred during irradiation.
Imaging is possible detecting secondary electrons that come from the impact
of primary ions with the surface, even though caution has to be paid not to
damage the sample. In fact, such a function was only used to identify the
sample and the exact zone where to perform the irradiations. Imaging
parameters (focus and contrast) can be set, by moving the beam on specific
copper grids and adjusting the settings.
After positioning the sample under the beam for the first irradiation, the
beam is turned on and the treatment starts. The time of exposure was
calculated from equation (11) consistently with the dose of ions to be
deposited on the sample. After turning off the beam, the sample is moved to
the next square and the beam is turned on again to treat the following area
with the chosen dose.
After FIB irradiation, the sample is extracted and the conductance of the
treated areas is measured with Keithley 2300 analyzer, as explained in detail
in Chapter 4.
63
In some cases, after the measurements, the sample undergoes a treatment of
annealing, to see if and to which extent the original conductance can be
restored.
Annealing temperature ranged from 400°C to 1500°C.
Annealing treatments were performed in the same all-graphite furnace in
which the samples of nanocrystalline graphene were grown.
The furnace was purged with Ar of ultra-high purity grade. Through this
expedient, and because of the fact that each and every part of the furnace is
high purity graphite made, we can be sure that no noticeable contamination
was present in the chamber during the annealing treatment.
A high vacuum is needed for the annealing process, and it was provided by
a turbomolecular pump, that works in series with a diffusive pump. Once a
vacuum of 10-4 mbar is reached the heater is turned on to increase the
temperature up to a first stay at 600 °C. The temperature is kept constant
around 600°C, in order to allow gases and impurities to be released, until the
chamber pressure reaches 10-4 mbar again. Then, the temperature is
increased to the set temperature of annealing and kept constant for the
chosen time.
After the annealing the system is let cool down and the sample is extracted
to be measured again.
Optical images of the irradiated structures were taken with Nikon Eclipse
LV100 microscope in the regimes of transmitted and reflected light.
Visualization of the irradiated areas was an issue and this should be
addressed separately. It was not problematic to find the irradiated squares
when the optical transmittance contrast between the irradiated and non-
irradiated areas was considerable, like for high dose irradiation.
In particular was found a high optical sensitivity of quartz to irradiation,
which made the irradiated areas easy to recognize.
Ion irradiation, even with relatively low doses, makes quartz permanently
gray and this color change is seen in microscope. In contrast, sapphire is not
so sensitive to ion irradiation. An irradiation test of pristine quartz and
sapphire substrates was performed in the regimes used for the irradiation of
64
the samples with graphene. It was found that the quartz substrates changed
their color in almost all irradiated areas, whereas the sapphire substrates did
not show any detectable changes in color (Picture 27). Thus, the color
changes observed on the graphene-on-quartz samples could be caused both
by graphene and quartz, while the color changes observed on the graphene-
on-sapphire samples were due to graphene only.
Picture 27 Optical images in reflected light of ion-‐irradiated areas on quartz (a, b) and sapphire (c, d) substrates after irradiation (a, c) and after subsequent annealing at 1400°C and complete removal of graphene (b, d). The irradiation dose for different squares varies from 1014 to 1016 cm-‐2. All the irradiated squares on quartz are clearly seen immediately after irradiation and after high temperature annealing and removal of graphene. On sapphire, only squares irradiated at high
65
doses (left squares on (c)) are clearly seen. The low dose irradiations shown on picture (c) are hardly seen if at all. After high temperature annealing and complete removal of graphene, all irradiated squares on sapphire are not visible. (e, f) Condensation of water on sapphire substrate covered with graphene after ion irradiation. The pattern of the condensed water droplets allows recognition of all irradiated areas.
The problem with the visualization of the irradiated areas on sapphire could
be overcome by gentle blowing warm air saturated with water vapor onto
the samples. Tiny water droplets condensate immediately on the sample
surface and make the irradiated squares clearly seen (Pictures 27e, 27f).
This effect of the selective water condensation was observed both on
sapphire and quartz and was used for the precise positioning of the
measuring probes over the irradiated areas.
6.2 Results and discussion 6.2.1 Conductance after irradiation and after annealing Dose dependences of the conductance of irradiated samples are shown in
Picture 28.
28 a
66
28 b
28 c
67
28 d
Picture 28. Dose dependence of conductance of graphene on quartz after irradiation with Ga+ ions of energies 40 keV (a), 20 keV (b), 5 keV (c) and on sapphire after irradiation with 50 keV Ga+ ions (d). Annealing time for all temperatures but 1500°C was 5 to 10 min. The annealing at temperature 1500°C was performed for 1 min.
The ion irradiation with doses below 3x1014 cm-2 has not affected the
conductance of nanocrystalline graphene both on quartz and sapphire
substrates. However, a dose of 3x1014 cm-2 of Ga+ ions is sufficient to
considerably suppress the conductance of single crystal graphene. An
obvious explanation of this difference is the inherent disorder of
nanocrystalline graphene [11]. This critical dose 3x1014 cm-2 does not
depend much on the ion energy (at least in the range of a few tens of keV)
and can be taken as a phenomenological parameter of the atomic disorder in
nanocrystalline graphene. Based only on this dose we cannot give a
meaningful estimation of the concentration of defects. The reason for that is
that the ion irradiation damage of graphene on substrate is a very complex
process involving the primary damage by fast ions, the damage by recoil
atoms from the substrate, the implantation of atoms from the substrate, the
68
sputtering as well as the processes of the secondary defect transformations.
We can just say that Ga+ ion irradiation with an energy of a few tens of keV
at doses up to 3x1014 cm-2 does not produce more atomic disorder than it is
present in nanocrystalline graphene CVD-grown on quartz and sapphire.
This conclusion is supported by the similarity of the Raman spectra of
nanocrystalline graphene [10] and single crystal graphene irradiated with
Ga+ ions [19, 44]. Both spectra are very much alike and typical of low
quality graphene (Picture 29). The main features are D-band (at about 1350
cm-1), G-band (at about 1580 cm-1) and 2D band (at about 2700 cm-1). All
these bands are broadened suggesting a considerable disorder of crystal
structure.
Picture 29. Raman spectrum of nanocrystalline graphene on quartz (re-‐plotted from [11]) is compared with that of single crystal graphene irradiated with Ga+ ions to a dose causing an order of magnitude reduction of its conductance (re-‐plotted from [19]).
At doses over 3x1014 cm-2, the conductance of irradiated graphene decreases
rapidly with the dose and becomes undetectable over the sensitivity at doses
exceeding 2x1015 cm-2. This irradiation-induced reduction in conductance is
an expected behavior for crystalline semiconductors and semimetals, the
conductance of which is strongly affected by defects working as scattering
centers and charge carriers traps. Another reason for the irradiation-induced
69
decrease in conductance of graphene is its gradual ion sputtering and
actually its physical removal. Indeed, the dose 2x1015 cm-2 is comparable
with the atomic density of graphene (3.9x1015 cm-2). Thus at this dose a
considerable fraction of graphene atoms experiences direct collisions with
the incident Ga ions and gets knocked out of the graphene layer.
Ion irradiation at doses of complete suppression of conductance can be used
as a method of maskless fabrication of electronic structures on graphene. In
this case, areas with as-deposited graphene are conductive, whereas
irradiated ones are insulating.
In Picture 30, the conductance contrast between an irradiated square and the
surrounding non-irradiated area is shown. Although a high conductance
contrast can be achieved this way, it may disappear after annealing, when
the graphene remaining in the irradiated area restores its conductance. That
is why in this case a sharp difference of doses also should be used, possibly
directly sputtering the region that is desired to be non conductive.
Picture 30. Change in conductance of nanocrystalline graphene on quartz measured with probes passing across a square as-‐irradiated with a dose of 5x1015 cm-‐2. The irradiated area is insulating, while the surrounding non-‐irradiated graphene exhibits its original conductance of 10-‐5 S. Zero mark on horizontal axis is set at one of the edges of the irradiated square.
70
The observed reduction in conductance of graphene after irradiation is not
unique among carbon materials. This effect is also known for bulk graphite.
In order to find out whether the ion dose range causing the reduction of
conductance is specific of graphene only, we performed comparative
measurements on high purity, high density polycrystalline graphite
irradiation in the same regimes. It was found that the current-voltage
characteristics of the irradiated areas on graphite revealed an electric
breakdown behavior (Picture 31). The current-voltage characteristics of this
type are typical for conductors covered with thin insulating layer. Thus this
result suggests formation of an insulating layer on the surface of graphite
after ion irradiation.
Picture 31. Current-‐voltage characteristics measured on the surface of a polished plate of
polycrystalline graphite before irradiation (dotted curve) and after irradiation with 50 keV Ga+
ions at doses 1.2x1014 (solid curve) and 1016 cm-‐2 (dashed curve). The characteristics of the
irradiated graphite exhibit current jump typical for electrical breakdown.
At low voltages, current is proportional to voltage and the conductance
remains unusually low till a certain voltage threshold. At this threshold,
current jumps up to a high value (compliance limit of the instrument) and,
71
with sweeping voltage down, current follows the dependence of the non-
irradiated graphite. This current jump is a typical electrical breakdown
through a thin insulating layer formed on the surface (which is what
determined the first trend of the curve, of low conductance for the voltage
applied). The breakdown starts as an avalanche sustained by injection of
high current and then ends up as a thermal breakdown, when the insulating
layer gets destroyed. The curve (1) in Picture 31 is an example of the
current-voltage characteristics when the thermal breakdown has not
occurred yet and the insulating layer survives. In this case, current returns
back to low values when the voltage is reduced. The curve (2) shows
current-voltage characteristic of full breakdown with the complete
destruction of the insulating layer. This is attested by the new high value of
conductance/voltage.
In this case, the conductance becomes high at any voltage and reaches the
value of non-irradiated graphite.
Conductance after Annealing
After the as-irradiated samples had been measured, they were annealed at
different temperatures up to 1500°C. The data show that the annealing at
400°C did not cause noticeable changes in conductance. After annealing at
temperatures over 600°C, a considerable recovery of conductance was
observed, and, at temperatures from 1000°C to 1300°C, the recovery
reached its maximum. The onset of the restoration of conductance at
temperatures 600°C to 800°C is expected. It is known that vacancies are the
most abundant defects in as-irradiated materials. This was confirmed also by
a TRIM simulation at energy 50keV, which is shown in Picture (32).
72
Picture 32 TRIM simulation with Ga ions irradiation with 50 keV energy, distribution of vacancies
is displayed
In graphite, vacancies start to move at a temperature of 700°C [45]. Since
the atomic structure of a few-layer nanocrystalline graphene is closer to that
of bulk polycrystalline graphite than to single crystal single layer graphene,
we can also expect similar temperatures of the activation of defect mobility
in graphite and nanocrystalline graphene.
At temperatures over 1300°C, a reverse process of reduction of conductance
was observed. The most drastic changes occurred at doses in the range from
1x1015 to 4x1015 cm-2. Picture 33 shows the change in conductance of
nanocrystalline graphene on quartz after irradiation at a dose of 1015cm-2
with Ga+ ions of different energy. It is seen that after annealing at 1000°C
and 1200°C the conductance increases over 7 orders of magnitude and
actually restores its original value of 10-5 S. After annealing at temperatures
over 1300°C, the conductance reduces again and completely disappears
after annealing at a temperature of 1500°C.
73
Picture 33. Conductance of nanocrystalline graphene grown on quartz after irradiation with 20 and 40 keV Ga+ ions at doses 1.3x1015 cm-‐2 and 1.4x1015 cm-‐2 respectively versus annealing temperature. Although the ion energies differ by a factor of two, there is no essential difference between the annealing dependences.
Complex behavior of conductance of graphene subjected to ion irradiation
and annealing suggests involvement of several processes, the obvious ones
being the annealing of irradiation-induced defects and the degradation of
graphene at high temperature. We exclude surface contamination as a
possible cause of the observed conductance change for two reasons: first,
we took every precaution to minimize the contamination during the
irradiation and annealing. Second, any surface termination possibly
occurring in the vacuum chamber at residual pressure below 10-5 mbar
cannot change the conductance of graphene for many orders of magnitude.
A cause of graphene degradation at high temperature could be a strong
chemical interaction with the substrate and the sublimation of graphene. In
order to find out which of these mechanisms prevails, a few samples of
graphene on quartz and sapphire were subjected to isothermal annealing at
different temperatures in vacuum. Some of these samples were annealed
74
with their graphene-carrying surfaces fully open to vacuum. The graphene-
carrying surfaces of the other samples were covered with clean substrates of
the same type and same size (stack of two substrates with graphene film in
between) and so annealed. The result of this annealing test is shown in
Picture 34.
Picture 34 (a) Change in conductance of non-‐irradiated graphene film on quartz as a result of isothermal annealing at 1300°C (3, 4) and 1500°C (1, 2); graphene film open to vacuum (1, 3); graphene film between two quartz plates (2, 4). (b) Change in conductance of graphene film on sapphire with annealing: comparison of sublimation of as-‐grown graphene and graphene irradiated with doses of partial sputtering.
75
At 1300°C, a slow reduction of conductance of non-covered graphene
occurs for 1 hour of annealing and then conductance disappears abruptly. At
temperature 1500°C, the conductance vanishes in less than 2 minutes. In
contrast, the conductance of the covered graphene reveals no degradation
after annealing at 1300°C and only a 10 times reduction after annealing at
1500°C for 7 minutes. This behavior well supports the idea of sublimation
of graphene. Assuming that the thickness of highly transparent
nanocrystalline graphene film is about 1 nm, the sublimation rate at
temperature 1500°C can be estimated as 1 nm per minute. This is
unexpectedly high value, which is at least three orders of magnitude greater
than the sublimation rate of bulk graphite in vacuum [46-48] and free-
standing multilayer graphene. We do not have any solid explanation of this
discrepancy yet, however the chemical interaction with the substrate and
first of all with its volatile component (oxygen atoms) could be the reason of
this highly stimulated sublimation.
Combination of the two effects, the healing of the radiation damage and the
sublimation, can well explain the changes in conductance during annealing.
At temperatures below 1200°C, sublimation is negligible and the restoration
of conductance occurs due to healing of the damaged graphene lattice.
Full restoration occurs only if the irradiation dose has not exceeded 2x1015
cm-2. For higher doses, from 2x1015 to 6x1015 cm-2, conductance restores
only partially. For doses over 6x1015 cm-2, the irradiated areas do not show
any conductance neither after irradiation, nor after subsequent annealing at
any temperature. This dose 6x1015 cm-2 of the total irreversible destruction
of conductance is most probably the dose of the complete ion-sputtering of
the graphene layer. The dose range from 2x1015 to 6x1015 cm-2 corresponds
to a partial sputtering of the graphene film. TRIM simulation [49] of the
radiation damage of 1 nm thick carbon film on quartz and sapphire predicts
the sputtering yield of carbon atoms by Ga+ ions of energy from 10 to 50
keV to be from 1.6 to 2. Thus the area density of the carbon atoms
sputtered by a dose of 6x1015 cm-2 is expected to be about 1.2x1016 cm-2.
This number is very close to 2D atomic density of a three-layer, 1 nm thick
76
graphene (1.15x1016 cm-2). Experimental studies of the ion beam sputtering
of bulk graphite with 5 keV Ar+ ions give a value of sputtering yield of 1.5
[50]. This number is in good agreement with our data. Thus we may
conclude that the total destruction of conductance of graphene after high
dose ion irradiation is the result of the physical sputtering of the graphene
film and that the ion sputtering is one of the major effects causing the
degradation of graphene during ion irradiation.
It is seen in Fig. 28 that the conductance of the ion-irradiated graphene
increases with annealing temperature up to 1200°C and then disappears after
annealing at higher temperatures due to sublimation. The conductance
vanishes first in the areas irradiated with high doses and then all irradiated
areas become non-conductive. This behavior is quite expected as the
sublimation efficiency of materials strongly depends on their structural
quality. The higher sublimation rate of the irradiated graphene is an effect,
which can be used for patterning. Regimes of irradiation and annealing can
be found so as to completely remove the graphene, whereas the non-
irradiated graphene still retains continuity and a considerable conductance.
6.2.2 Results in adhesion after irradiation and annealing
It has been found that even low dose ion irradiation improves adhesion of
graphene grown on quartz and sapphire. The enhancement of adhesion is
especially pronounced for quartz substrates. Our tentative explanation of the
different efficiency of quartz and sapphire is the different chemical
composition and, first of all, the presence of silicon atoms. TRIM simulation
of the Ga ion irradiation of graphene on quartz shows that a considerable
intermixing of C, Si and O atoms occurs at the graphene-quartz interface.
Thus the formation of stable Si-C bonds between graphene and quartz
during ion irradiation, and especially after subsequent annealing, is quite
possible.
The results of the simulations and the distributions of C, O and Si are shown
in the following graphs (Picture 35):
77
Picture 35 TRIM simulation with Ga ions with 50 keV energy, distribution of C, O and Si
78
The irradiation enhanced adhesion was found when the samples with the
irradiated graphene were gently rubbed with cotton soaked in acetone
(Picture 36). Usually rubbing with cotton could easily remove as-deposited
graphene from all substrates. In contrast, the ion-irradiated graphene could
not be easily removed this way and, if on quartz, it could not be removed
completely even after vigorous rubbing.
Picture 36. (a) As-‐grown graphene of quartz substrate. Considerable removal of graphene film occurs after single gentle stroke with cotton soaked in acetone. (b) Graphene on sapphire substrate after irradiation and annealing at 1200°C.
Graphene film remains on irradiated square areas after multiple strokes with
cotton soaked in acetone, while it is completely removed from the non-
irradiated areas. Graphene uniformly covers the areas irradiated with doses
over 1014 cm-2 (upper row of squares), whereas it has been partially removed
79
from the areas irradiated with doses below 1014 cm-2 (bottom row of
squares). For 50 keV Ga+ ions, the irradiation-stimulated adhesion of
graphene occurs at doses over 2x1014 cm-2 and its strength comes to
maximum at doses over 7x1014 cm-2 (Picture 37a). The dose 7x1014 cm-2 is
low enough not to affect much the conductance of nanocrystalline graphene
(Picture 28).
Picture 37. Conductance of graphene on sapphire after 50 keV Ga+ ion irradiation, annealing at a temperature of 1300°C and single stroke with cotton over the sample surface. (a) Dose dependence of the irradiation-‐enhanced adhesion. At doses below 2x1014 cm-‐2, the adhesion
80
remains negligible and graphene is wiped away completely. In the dose range from 2x1014 cm-‐2 to 7x1014 cm-‐2, the adhesion improves and graphene remains partially on the irradiated areas. At doses over 7x1014 cm-‐2, complete retention of the graphene layer takes place. (b) Conductance contrast between the irradiated and non-‐irradiated areas. Conductance is measured with the probes scanning over two areas irradiated with doses 2x1014 cm-‐2 and 2x1015 cm-‐2. Graphene is completely removed from the non-‐irradiated surface leaving it nonconductive. The irradiated areas retain conductance due to remaining graphene film. The zero mark on horizontal axis is set at one of the edges of the irradiated area.
The enhancement of adhesion takes place immediately after the irradiation.
Subsequent annealing results in a further improvement of the adhesion.
After the irradiation and annealing, the adhesion of graphene on quartz may
become so strong that the conductive graphene film cannot be removed
from the substrate even by an intense rubbing.
The effect of the irradiation-enhanced adhesion of graphene to the substrates
can be used for the patterning and development of imprint lithography for
graphene. We have found that the application of a sticky tape to the samples
with the irradiated graphene and then its peeling off is a way to remove the
non-irradiated graphene and to leave the irradiated graphene areas in place.
When scanning the measuring probes over the irradiated areas of these
samples, a high conductance contrast between the irradiated areas with
graphene and the surrounding non-irradiated areas has been revealed. This
contrast was similar to that shown in Picture 37b.
6.2.3 Results in Nucleation improvement of Graphene on Ion-Irradiated
Substrates
Ion irradiation can be also used for selective growth of nanocrystalline
graphene on quartz and sapphire. It has been found that the formation of
continuous well conductive graphene on ion-irradiated areas can be
achieved in the temperature-pressure-time regimes at which no conductive
film grows on non-irradiated surface (Picture 38).
81
Picture 38 Deposition of graphene on sapphire substrate at temperature 1200°C, methane pressure 1 mbar, growth time 1 minute. At these parameters, visible and conductive graphene film grows only on the ion-‐irradiated areas.
Our experiments show that the higher the pressure of methane the broader
the dose range of the stimulated growth. For instance, at a pressure of 0.1
mbar and temperature 1280°C, the growth of graphene is observed on the
areas irradiated with doses from 7x1014 cm-2 to 2x1015 cm-2, whereas at a
pressure of 2 mbar and the same temperature all the areas irradiated with
doses from 2x1014 cm-2 to 6x1015 cm-2 reveal deposition of highly
conductive film (Picture 39). In both cases, the non-irradiated surface
remains insulating.
82
Picture 39 Deposition of nanocrystalline graphene on quartz ion-‐irradiated with different doses. (a) Conductance of the irradiated areas after: irradiation and deposition at methane pressure of 0,1 mbar (light blue symbols); cleaning in oxygen plasma (green symbols); plasma cleaning and deposition at methane pressure of 0.1 mbar (red symbols); plasma cleaning and deposition at
methane pressure of 2 mbar (blue symbols). (b) Measuring conductance with probes moved over the squares after irradiation with doses 1.1x1015 and 1.8x1015 and graphene deposition at 0.1 mbar pressure ( pink and purple symbols); after irradiation with doses 2x1014 and 5.7x1015 cm-‐2
and graphene deposition at 2 mbar pressure (orange and green symbols).
83
The enhancement of nucleation of graphene on ion-irradiated quartz and
sapphire is a very stable effect. It persists after multiple cleanings in
solvents, high temperature annealing and even after cleaning in oxygen
plasma. This stability suggests that the enhanced nucleation is the result of
permanent ion damage of the substrate and, as such, remains until the ion-
damaged layer is removed to its whole depth.
As to the preferential growth on the areas irradiated with certain doses after
plasma cleaning (oxygen plasma termination), as seen in Picture 39 (a), it is
hard for us to give a solid explanation, since we had no chance to study this
effect in detail.
We suppose it occurs as a manifestation of three concurring processes:
(i) stimulation of growth by low dose ion irradiation (below 2e15 cm-2),
(ii) suppression of growth by high dose irradiation (above 3e15 cm-2) and
(iii) suppression of growth by O-plasma treatment. The latter is a weaker
effect compared with ion irradiation and hence it is observed only for low
doses of irradiation (below 7e14 cm-2). At higher doses, the irradiation
stimulation of growth is stronger than the plasma suppression and thus the
growth happens at doses over 7e15 cm-2. However at high doses (over 2e15
cm-2), the damage of the substrate surface becomes too high and this
suppresses the growth.
84
Chapter 7 Conclusion Graphene, a very promising true 2D nanoelectronic material, is studied in
many different aspects.
Its zero bandgap and consequent unipolar transport makes it a very
interesting material for micro and nano-technology.
One aspect that we believe has not been explored with appropriate attention
yet is the characterization of its surface.
The aim of this work was to find new ways to characterize nanocrystalline
graphene.
As a first step the growth of this material on top of single crystal quartz and
sapphire was studied in detail, understating the dependence of the growth on
three main parameters, temperature pressure and time, and getting to know
under which conditions the best quality nanocrystalline graphene can be
grown.
The first attempt to characterize graphene, creating a sharp contrast between
a conductive and a non conductive area, was made with plasma pre-
treatments of the substrates.
Oxygen, Argon, Nitrogen, Hydrogen and Krypton plasmas did not produce
relevant results of selective enhancement or suppression of growth.
The main focus of the work was the study of the characterization of
nanocrystalline graphene with Gallium ions via FIB. The irradiations were
performed in a wide range of doses, with energy from 5 to 50 keV.
The results show that using ion irradiation it is possible to control the
electrical conductance of nanocrystalline graphene on quartz and sapphire
over many orders of magnitude. It is shown that nanocrystalline graphene
stands fairly high doses of ion irradiation (up to 3x1014 cm-2 for 5 to 50 keV
Ga+ ions) without degradation in conductance.
At higher doses, nanocrystalline graphene rapidly loses its conductance and
at doses over 2x1015 cm-2 becomes actually insulating.
Annealing in vacuum at temperatures over 600°C restores the conductance
of the ion-irradiated nanocrystalline graphene and, if the irradiation does not
85
exceed a dose of 3x1015 cm-2, this restoration can be almost complete. This
reverse increase in conductance after annealing occurs via the healing of the
radiation damage.
Ion irradiation at very high doses - approaching 1016 cm-2 - results in the
complete sputtering of a few graphene layers. This is confirmed by the
lowering of conductance, and the impossibility to recover it even after high
temperature annealing. A quick calculation, based on graphene density and
TRIM simulations of irradiation, confirms that at such high doses the
nanocrystalline graphene layers are sputtered by the colliding ions.
This can be a first method of graphene patterning, by irradiating areas of the
sample which give rise to non conductive areas with a sharp contrast with
respect to untreated areas in which highly conductive graphene is preserved.
It has also been found that along with radiation damage and ion beam
sputtering, the high temperature sublimation is an important effect involved
in the reduction of conductance of the ion-irradiated and annealed
nanocrystalline graphene.
It is also shown that ion irradiation improves adhesion of graphene to quartz
and sapphire substrates. As a result, graphene in the non treated areas can be
easily wiped away with cotton and acetone, while in the areas treated with
gallium ions the adhesion of graphene is increased. When the treated areas
undergo the same acetone cleaning, graphene is not wiped away, and
conductance measurements confirm its presence, with a sharp contrast
between pristine and treated areas. This can be an additional method of
patterning graphene.
For 50 keV Ga+ ions, irradiation-enhanced adhesion is observed at doses
over 2x1014 cm-2.
The increased efficiency of the graphene nucleation on these materials
during CVD growth is another effect studied. After observing the
enhancement of adhesion of nanocrystalline graphene as a result of
irradiation, we investigated whether the same effect could be induced by
treatments performed before the growth of graphene.
86
The results of irradiations in a broad dose range (from 1x1014 up to
5x1015cm-2) demonstrate the occurrence of selective growth on the treated
areas. A deeper study of the effect is desired to better understand the
responsible mechanisms and to optimize the range of irradiation to have the
most effective growth contrast.
The promoted graphene nucleation is observed in a broad dose range.
The above effects can be used for developing methods of patterning of
graphene deposited on insulating substrates and methods of imprint
lithography of graphene. As the modern FIB systems permits ion irradiation
with a sub 10 nm resolution [18, 51, 52], the patterning of graphene at a few
nanometer scale is feasible. It is important that the patterning done in this
way is maskless and resist-free, and, as such, would not require steps of
cleaning of graphene surface from the resist residues.
87
Bibliography 1. A. K. Geim, Science 324, 1530 (2009). 2. A. C. Ferrari, “Graphene Technology: Roadmap to Applications”, file:///C:/Users/You%20Mad/Documents/Geaphene/ferrari-21032011_en.pdf. 3. S. Gladstone, “Graphene and Its Applications: The Miracle Material of the 21st Century”, Dartmouth Undergraduate Journal of Science, Spring 2012, p. 6-8. 4. J. Liua, Y. Xuea, M. Zhanga, and L. Dai, MRS Bulletin 37, 1265 (2012). 5. F. Bonaccorso, Z. Sun, T. Hasan, and A. C. Ferrari Nature Photonics 4, 611-622 (2010). 6. Y. H. Wu, T. Yu, and Z. X. Shen, Journal of Applied Physics 108, 071301 (2010). 7. A. H. Castro Neto and K. Novoselov 2011 Rep. Prog. Phys. 74 102001. 8. G. Chen, T. M. Paronyan, and A. R. Harutyunyan, Applied Physics Letters 101, 053119 (2012). 9. G. Wang, M. Zhang,Y. Zhu, G. Ding, D. Jiang, Q. Guo, S. Liu, X. Xie, P. K. Chu, Z. Di, and X. Wang, Nature: Scientific Reports, 3, 2465 (2013). 10. S. V. Samsonau, E. Dzedzits, S.D.Shvarkov, F. Meinerzhagen, A. D.Wieck, and A. M. Zaitsev, Sensors and Actuators B: Chemical 186, 610 (2013). 11. S. V. Samsonau, S. D. Shvarkov, F. Meinerzhagen, A. D. Wieck, and A. M. Zaitsev, Sensors and Actuators B 182, 66 (2013). 12. J. Hwang, V. B. Shields, C. I. Thomas S. Shivaraman, D. Hao, M. Kim, A. R. Woll, G. S. Tompa, and M. G. Spencer, Journal of Crystal Growth 312, 3219 (2010). 13. U. Wurstbauer, T. Schiros, C. Jaye, A. S. Plaut, R. He, A. Rigosi, C. Gutierrez, D. Fischer, L. N. Pfeiffer, A. N. Pasupathy, A. Pinczuk, and J. M. Garcia, Carbon 50, 4822 (2012). 14. L. G. Canc-ado, A. Jorio, E. H. Martins Ferrreira, F. Stavale, C. A. Achete, R. B. Capaz, M. V. O. Moutinho, A. Lombardo, T. S. Kulmala, A. C. Ferrari, NanoLetters 11, 3190 (2011).
88
15. G. Compagninia, F. Giannazzo, S. Sonde, V. Raineri, and E. Rimini, Carbon 47, 3201 (2009). 16. L. Tapasztó, G. Dobrik, P. Nemes-Incze, G. Vertesy, Ph. Lambin, and L. P. Biró, Physical Review B 78, 233407 (2008). 17. J. –H. Chen, W. G. Cullen, C. Jang, M. S. Fuhrer, and E. D. Williams, Physical Review Letters 102, 236805 (2009). 18. Y. Zhang, C. Hui, R. Sun, K. Li, K. He, X. Ma, and F. Liu, Nanotechnology 25, 135301 (2014). 19. Q. Wang, W. Mao, D. Ge, Y. Zhang, Y. Shao, and N. Ren, Applied Physics Letters 103, 073501 (2013). 20. C. Deng, W. Lin, S. Eimer, D. Ravelosona, C. Chappert, and W. Zhao, “Raman Spectroscopy Study of Low Energy He+ Ion Irradiation Effect in Graphene Transferred onto SiO2”, Proc. of the 13th IEEE International Conference on Nanotechnology, Beijing, China, August 5-8, (2013). 21. A. Lherbier, S. M.-M. Dubois, X. Declerck, Y. –M. Niquet, S. Roche, and J. –C. Charlier, Phys. Rev. B 86, 075402(2012). 22. A. Lherbier, S. M.-M. Dubois, X. Declerck, S. Roche, Y. –M. Niquet, and J. –Ch. Charlier, Physical Review Letters 106, 046803 (2011). 23. J. Kotakoski, A. V. Krasheninnikov, U. Kaiser, and J. C. Meyer, Physical Review Letters 106, 105505 (2011). 24. S. Nakaharai, T. Iijima, S. Ogawa, S. Suzuki, S. –L. Li, K. Tsukagoshi, S. Sato, and N. Yokoyama, ACS Nano 7, 5694 (2013). 25. K. J. Kim, J. Choi, H. Lee, H. –K. Lee, T. –H. Kang, Y. –H. Han, B. –C. Lee, S. Kim, and B. Kim, J. Phys. Chem. C 112, 13062 (2008). 26. J. –H. Kim, J. H. Hwang, J. Suh, S. Tongay, S. Kwon, C. C. Hwang, J. Wu, and J. Y. Park, Applied Physics Letters 103, 171604 (2013). 27. D. Teweldebrhan and A. A. Balandina, Applied Physics Letters 94, 013101 (2009). 28. I. Childres, L. A. Jauregui, M. Foxe, J. Tian, R. Jalilian, I. Jovanovic, and Y. P. Chen, “Effect of electron-beam irradiation on graphene field effect devices”, http://arxiv.org/ftp/arxiv/papers/1008/1008.4561.pdf 29. I. Childres, R. Jalilian, M. Foxe, A. Chernyshov, L. Rohkinson, I. Jovanovic, and Y. P. Chen, “Effect of Energetic Electron Irradiation on
89
Graphene”,https://www.physics.purdue.edu/quantum/files/childres_ebeam_crete09_highres.pdf 30. Li Tao, C. Qiu, F. Yu, H. Yang, M. Chen, G. Wang, and L. Sun, J. Phys. Chem. C 117,10079 (2013). 31. Y. H. He, L. Wang, X. L. Chen, Z. F. Wu, W. Li, Y. Cai, and N. Wang, Applied Physics Letters 99, 033109 (2011). 32. S. Nakaharai, T. Iijima, S. Ogawa, K. Tsukagoshi, S. Sato, and N. Yokoyama, “Electrostatically-Reversible Polarity of Dual-Gated Graphene Transistors”, Functionality-Enhanced Devices Workshop FED’13, EPFL, March 25th, (2013). 33. A. V. Klekachev, A. Nourbakhsh, I. Asselberghs, A. L. Stesmans, M. M. Heyns, and S. De Gendt, The Electrochemical Society Interface, Spring 2013, 63 (2013). 34. M. C. Prado, D. Jariwala, T. J. Marks, and M. C. Hersam, Appl. Phys. Lett. 102, 193111 (2013) 35. S. Kumar, N. Peltekis, K. Lee, H.-Y. Kim, and G. S. Duesberg, Nanoscale Res. Lett. 6, 390 (2011). 36. N. Peltekis, S. Kumar, N. McEvoy, K. Lee, A. Weidlich and G. S. Duesberg, Carbon 50, 395 (2012). 37. V. N. Popok, “Compositional and structural alterations of polymers under low-to-mediumenergy ion implantation”, in “Surface Science Research”, Ed. Charles P. Norris, Nova Science Publishers Inc., 2005, p. 147-193. 38. J. E. Pawel and C. J. McHargue, “Analysis of pull tests for determining effects of ion implantation on adhesion of iron films to sapphire substrates” in “Adhesion Measurement of Films and Coatings”, Ed. K. L. Mittal, VSP BV Publications, p. 323-330 (1995). 39. A. V. Krasheninnikov, K. Nordlund, and J. Keinonen, Physical Review B 65, 165423 (2002). 40. S. Tongay, M. Lemaitre, J. Fridmann, A. F. Hebard, B. P. Gila, and B. R. Appleton, Applied Physics Letters 100, 073501 (2012). 41. J. D. Plummer, M. D. Deal, ad G n P. B. Griffin, “Silicon VLSI Technology Fundamentals, Practice and Modeling”, Prentice Hall (2000).
90
42. V. Ray, University of Maryland, College Park, “FIB short course”, [email protected] 43. www.nikoninstruments.com 44. L. G. Cancado, A. Jorio, E. H. Martins Ferreira, F. Stavale, C. A. Achete, R. B. Capaz, M. V. O. Moutinho, A. Lombardo, T. S. Kulmala, and A. C. Ferrari, Nano Lett. 11, 3190 (2011). 45. M. Li and N. V. Mokhov, “Experience with Moving from Dpa to Changes in Materials Properties”, 46th ICFA Advanced Beam Dynamics Workshop, Sept. 27 – Oct. 1, 2010, Morschach, Switzerland, (2010). 46. L. B. Tecchio, M. Tonezzer, J. Bermudez, M. Loriggiola, and M. Lollo, “Measurement of Graphite Evaporation Rate”, LNL Annual Report, 2008, p. 251-252. http://www.lnl.infn.it/~annrep/read_ar/2008/contributions/pdf_singles/251.pdf 47. J. R. Haines and C. C. Tsai, “Graphite Sublimation Tests for the Muon Collider/Neutrino Factory Target Development Program”, (2002), https://www.google.com/webhp?sourceid=chrome-instant&ion=1&espv=2&ie=UTF-8#q=Graphite%20Sublimation%20Tests. 48. P. Thieberger, “Upper Limits for Sublimation Losses from Hot Carbon Targets in Vacuum and in Gasses”, Brookhaven National Laboratory, MUC-0186, (2000). 49. J. F. Ziegler, “Particle Interactions with Matter”, http://www.srim.org/ 50. R. S. Williams, E. J. Snyder, D. A. Baugh, R.S. Williams, E. A. Eklund, and M. S. Anderson, "Evidence for the dependence of the sputter yield of graphite on ion bombardment fluence", Naval Research Technical Report, No. 19 (1992), http://www.dtic.mil/dtic/tr/fulltext/u2/a252500.pdf 51. A. N. Abbas, G. Liu, B. Liu, L. Zhang, H. Liu, D. Ohlberg, W. Wu, and C. Zhou, ACS Nano 8, 538 (2014). 52. A. Hemamouche, A. Morin, E. Bourhis, B. Toury, E. Tarnaud, J. Mathé, P. Guégan, A. Madouri, X. Lafosse, C. Ulysse, S. Guilet, G. Patriarche, L. Auvray, F. Montel, Q. Wilmart, B. Plaçais, J. Yates, and J. Gierak, Microelectronic Engineering 121, 87 (2014)