-
FABRICATION METHODS
Manufacturing takes place in very large facilities. lf you want
to build a computeichip, you need a giant semiconductor fabrication
facility. But nature can growcomplex molecular machines using
nothing more than a planto
RALPH MERKLE
-
178
time and time again later in the texto The physicsdivision of
the text-chapter 5 through chapter 8-engages the study of
nanomaterial properties andphenomena.
Please take note that the fabrication methodslisted in this
chapter are but a few of the multi-tude that actually exist. We
have tried to catego-rize in a generic sense the major forms and
triedto illustrate the processes with commonly prac-ticed
fabrication techniques. Much can be learnedabout nanomaterials by
understanding how theyare made.
Introduction to Nanoscience and Nanotechnology
Characterization methods have been presented,addressed, and
discussed, albeit without providingsignificam detail. The catalog
nature of chapter 3 isdeliberately extended into this chapter,
which isthe last chapter in the "Nanotools" division of thetext
..Because reference is made continually to vari-ous kinds of
fabrication techniques throughout thetext, it is prudent to place
introductory materialconcerning fabrication early in the book. In
this way,the student should be able to establish a levei ofcomfort
with, perspective on. and understandingof fabrication methods when
the subjects emerge
THREADS
4.0 FABRICATION OF NANOMATERIALSThere is nothing more
gratifying, arguably, than holding in one's hand the
physicalmanifestation of an idea, concept, or theory. The link
between the idea, concept, ortheory and its physical form is the
process of fabrication. The fabrication processbegins in a
laboratory with atomistic simulations, experiments, mock-ups,
andprototypes. Eventually, after a battery of testing, the physical
embodiment of theidea. concept, theory, simulation, mock-up, and
prototype makes it way into amanufacturing facility. We have
already listed several characterization methods. Itis now time to
discuss the fabrication of nanomaterials.
Nanomaterials are made by two generalized processes: top down
(e.g., sub-traction from bulk starting materiais] or bottom up
(e.g., addition of atomic ormolecular starting materiaIs). Each
scheme has a unique set of advantages anddisadvantages. We
recommend that you make a checklist of the advantages,
dis-advantages, limitations, and issues confronting each method as
we discuss themthrough the course of this chapter.
We also add a brief section on molecular modeling, which is a
fabricationtoo1. It is part of the design processo Molecular
modeling has become one of themost powerful tools in nanoscale
research, developrnent, and material designoThere exists a perfect
fit between simulation and nanomaterials since atoms andmolecules
in nan~scale materiaIs are finite and countable, and computer
capa-bility in this day and age is stilllimited with regard to
capacity. Depending onthe quality of input parameters, molecular
simulation is able to generate anaccurate rendition of nanoscale
material behavior. Low-energy states, structure,dynamical behavior,
chemical reactions, fluxes and flows, and more have beenmodeled
with some form of atomistic-molecular simulation.
4.0.1 BackgroundLike anything else that we present in this text,
boundaries are drawn for thesake of convenience and clarity,
although sharp ones are not always possible.
-
Pabrication Methods 179
Boundaries defining fabrication methods are no different;
however, weproceed unabated and present the first bifurcation in
the road. Like the greatbaseball player Yogi Berra said, "When you
come to a fork in the road, take it."
There are two generic strategies for nanomaterial fabrication:
top down andbottom up. Top-down fabrication methods begin with bulk
materials (top) thatare subsequently reduced into nanoparticles
(down) by way of physical, chemical,or mechanical processes.
Bottom-up methods, on the other hand, begin with
r atoms and molecules (bottom). These atoms or molecules react
under chemicalor physical circumstances to form nanomaterials (up
). Growth proceeds in zero,one, or two dimensions to form dots,
wires, or thin films, respectively. There aretwo generic types of
bottom-up procedures: In the first, nanomaterials retainsome level
of structural and functional independence; in the second,
nanoma-terials become identical components of a bulk material. An
example of theformer is an array of gold quantum dots in an
electronic device. Examples of thelatter case include bulk metals
formed from nanocrystallites and the structure ofbone tissue.
:i There is, of course, further blurring of boundaries. Two
general kinds ofoverlap, and possibly a third kind, occur between
the two types (bottom up andtop down) of fabrication strategies. In
one case, a technique may be designatedas top down but its
microscopic mechanism suggests otherwise. The bestexample of this
is the formation of carbon nanotubes by laser ablation. Thestarting
material is a target made of compacted graphite and catalyst
particles-certainly considered to be a bulk material in a compacted
formo However, carbonnanotubes form from atoms and molecules via a
catalytic process-definitelyfrom the bottom up. Bismuth metal,
obtained in bulk form, is melted andsubsequently evaporated into
atoms that deposit on the surface of a templatematerial.
Evaporation of a melted metal source to produce atoms (and
perhapsnanoclusters) is a top-down procedure, but the formation of
the thin layer ofbismuth from those evaporated atoms is certainly
from the bottom up.
In the second case, a manufacturing process may consist ofboth
top-down andbottom-up methods. During the course of the fabrication
of a computer chip,application of a photoresist material by a
process called spin-coating is top down(from a bulk liquid phase).
Photolithography is top down; chemical etching ofthe photoresist or
the silicon substrate to reveal features is top down, but
chemicalderivatization to form a monolayer comprising different
materials is bottom up.
Hybrid fabrication technology is a combination of distinct
top-down andbottom-up mechanisms that occur simultaneously. This
category of fabricationis exclusive to the nanoscale, where
top-down and bottom-up techniques con-verge at the 30-nm size scale
[1]. At the 3-nm scale, even hybrid technologieswill be challenged
by supramolecular and molecular technology that in tummay give way
to atomic and nuclear technologies at the subnanometer scale-the
realm of the single atom, single electron, single spin, and single
photon [1].These developments will require that we redefine
top-down, bottom-up, andhybrid fabrication technologies. In the
final analysis, it matters not whichdesignation is assigned to a
specific procedure, but for the sake of pedagogicalpurposes, we
will continue to explore many types of fabrication methods andlabel
them as one or the other or both.
Nanofabrication methods, just like characterization methods,
have a longhistory. Fabrication and synthesis processes are the
descendents of well-developed
-
180 lntroduction to Nanoscience and Nanotechnology
chemical and physical techniques developed over millennia.
Engineers tend tomanufacture components from the top down and then
assemble them to makea device. Chernists. on the other hand, have
always made materiais by reactingatoms and molecules to form
chemicals in bulk quantities-from the bottom-up.Chemical synthesis
is by definition a bottom-up processo With regard to thebiological
processes, ali structures are formed from the bottom up. Are you
ableto think of any exceptions to this mie?
The convergent nature of nanotechnology is well represented by
fabricationmethods. Engineers, physicists, chernists, and
biologists respectively bring top-down, top-down, bottorn-up, and
bottom-up methods to the same table. Thefuture of fabrication will
require more cooperation between and among thedisciplines, and the
design parameters of future nanofabs must include such for-ward
thinking in arder to accommodate diversity and to enhance
interactionamong ali the participants.
It is not practical to build an automobile engine from the
bottom up and,conversely, it is not practical to synthesize aspirin
from the top down, However,in nanotechnology, similar structures
can be built from either fabricationperspective [2]. Features on a
silicon wafer can be produced by a standard top-bottom procedure
called lithography (bulk wafer ~ application of a photoresistlayer
~ mask-UV exposure ~ etch) ar by a bottom-up procedure (bulk wafer
~polymer or seed crystals ~ self-assembly) [2]. Once again,
nanotechnology andnanoscience are changing the way we do things and
fabrication methods are noexception.
4.0.2 Types ofTop-Down Fabrication MethodsWe begin our catalog
of fabrication methods with top-down methods. Physicalfabrication
techniques are considered to be mostly from the top down.
Top-downmethods are extremely diverse. Nanomaterials are farmed
frorn the top downby mechanical-energy, high-energy, thermal,
chemical, lithographic, and naturalmethods.
Top-Down Mechanical-Energy Pabrication Methods. Cutting,
rolling, beating,machining, compaction, millng, and atomization
comprise a few examples ofmechanical methods used to produce
nanomaterials from the bulk. A mechanicalmethod employs a physical
process that does not involve chemical change-accarding to the
traditional definition of chemical change (a reaction).
Beatingmetais into a thin film is an ancient mechanical procedure
used by the Egyptiansand other pre-Hellenistic cultures to make
swords, spear tips, and ornamentalcoatings. Mechanical energy
methods such as ball milling operate on the principieof mechanical
attrition. Kinetic energy, translated by hard, high-speed pellets,
isimparted to samples by collision and friction. Samples are ground
into finepowders by this method. An overview of mechanical top-down
methods is shownin Table 4.1.
Top-Down Thermal Pabrication Methods. In the purest sense, a
thermal fabri-cation method employs a physical process (heating)
that does not initiatea chernical change in the sample-according to
the traditional definitionof chemical change (a reaction). Once
again, it has proven difficult to place
-
Pabrication Methods 181
TABLE 4.1 'Iop-Doum Mechanical-Energy Pabrication Methods
CommentsProduction of nanoparticles by mechanical attrition to
produce grain size
-
182
TABLE 4.2MethodAnnealing
Chill block meltspinningElectrohydrodynamicatomization
(EHOA)
Electrospinning
Liquid dynamicompaction (LJlC)Gasatam ization
Evaporation
Electrospinning
Extrusion
Template synthesis +evaporation
Sublimation
Thermolysis;pyrolysis
Combustion
Carbonization ofcopolymers
Introduction to Nanoscience and Nanotechnology
Top-Dtrum. Thetml Fabticaton Methods
CommentsThere are two applications of annealing: (1) anneal of
bulk materiais to form nanocrystallites,and (2) transformation of
nanomaterials into another physical phase [6].Microphase separation
to form nanoscopic structures occurs in copolymer bulk
materiaisupon application of thermal anneal above the glass
transition temperature.Metal is melted with RF coil and forced
through nozzle on rotating drum, where it solidifies;strips formed
with nanostructure [7].Production of monodisperse droplets; melt or
liquid materiais at nozzle with electric fieldbetween nozzle and
surface: cone ~ thin jet ~ droplets
EDHA + pyrolysis to produce 10-nm Pt nanoparticles [8]A high
voltage is applied to a polymer melt solution to induce charging.
Polymer solutions atroom temperature are also used routinely. At an
acquired threshold, an electrospun fluid jetemerges from a needle
tip to form a Taylor cone. The substrate, held at a lower
potential, iscovered by the charged polymeric solutionMolten stream
of metal is atomized by high-velocity pulses of an inert gas and
thesemisolidified droplets are collected on a chilled, metallic
substrate [9].Molten metal is subjected to high-velocity inert gas
impact that forms metal droplets [7,10].Kinetic energy is
transferred to metal, resulting in small droplets upon
solidification to formpowders. Powders are then compacted to form
high-strength bulk materiais.Evaporation of solid metal or other
material samples to form thin films; usually performedunder high
vacuum (10-6 torr). Heat is produced by electrical resistance. If
nanoclusters areformed during the evaporation process, it is top
down. If atoms or molecules are formedduring the evaporation
process that recombine to form a thin layer without any
chemicalreaction, it is a crossover technique.
The process of electrospinning utilizes electricity to form thin
layers of filaments from bulkpolymer, composite, or ceramic
solutions; fibers with nanoscale diameter can be fabricated
[11].
Nanowires by extrusion of bismuth melt by pressure injection
into porous template materialsuch as alumina [12]. Parallel Bi
nanowires with diameter -l3 nmFormation of single-crystal Bi
nanowires by a vapor-phase technique into porous
aluminatemplate-7-nm Bi nanowires [13]; 400-500C with N2 trapo Only
phase changes areinvolved in this processo
The physical process of sublimation involves a phase change from
a solid into gaseous formwithout a liquid intermediate phase. If
nanoclusters are formed by this process, it isconsidered to be a
top-down processo If atoms or molecules are formed first and
thenagglomeration into nanoparticles occurs, it is considered to be
a crossover technique inwhich both top-down and bottom-up processes
occur nearly simultaneously. Sublimationdoes not involve a chemical
change of the material.Oecomposition of bulk solids at high
temperature (top-down). These terms are also appliedto the
decomposition of molecules-nanomaterials are formed after
decomposition in abottom-up way by agglomeration. Because of this
crossover, it is hard to place pyrolysis/thermolysis into one
category or the other. The most common sense of the terms
impliesthat molecules are simply converted into other molecules. In
this sense, pyrolysis andthermolysis are neither top-down-nor
bottom-up methods. In such reactions (Iikedecomposition), chemical
change does occur.
Chemical combustion is a top-down process in which there is
chemical coriverson of bulkorganic materiais + impurities into
molecules like CO2, H20, and nanomaterials such as ashwith micron
to submicron dimensions. The process of combustion involves
oxygen.Spun fibers from polymethylmethacrylate
(PMMA)--polyacrylonitrile (PAN) microspheres inPMMA matrix (top
down? or bottom up?). Ternperature treatment at 900C removes
PMMAand converts PAN into MWNTs [14]. Carbonization is another
example of the difficultyencountered in cataloging such
processes.
-
Pabrication Methods
result in nanomaterials. Evaporation, a thermal method based on
resistive heating,is considered to be a crossover technique in that
a bulk material is convertedinto small partides (molecules or
clusters)-a top-down process-that are thendeposited to form a
nanomaterial (thin film)-a bottorn-up processo
Top-Doum Chemical Pabrication Methods. If chemical
transformations occurduring a fabrication process, we shall
designate that process as a chemical fabri-cation method. AIthough
fabrication (a.k.a. synthesis) methods that employchemical
procedures rightfully reside within the domain of the bottom up.
thereare several that can be considered to be top down. Combustion
is an ambiguous
TABLE 4.3MethodArc discharge
Laser ablation
Solar energyvaporizationRF sputtering
lon milling
Electron beamevaporation
Reactive ionetching
Pyrolysis
CombustionHigh-energysonication
Top-Doum High-Energy and Particle Fabrication Methods
CommentsHigh-intensity electrical are discharge directed on a
graphite target (anode) + catalyst to producesingle-walled carbon
nanotubes that accumulate on the cathode
Temperature -4000 K [15,16]High-intensity laser beam directed on
a graphite target + catalyst to produce single-walled
carbonnanotubes; sample warmed to 1200-1500C by lurnace, laser
Sample is collected on water-cooled copper collector [17]. This
process can be considered to be athermal and a high-energy
method.Solar energy locused on graphite target + catalyst to
produce single-walled carbon nanotubesTemperature -3000+K [18]lon
bombardment 01 metal, oxide, or other material targets to lorm thin
film coatingsUsually performed under moderate vacuum (10-3 torr).
Atoms, molecules, and clusters are formedby this processo
Argon ion plasma is used to subtract material from a surlace.
The purpose is to clean surface orremove (thin) materiais for TEM.
No change in the chemical nature of the sample happens duringthis
processoThis is similar to evaporation in Table 4.2 but uses an
electron beam source to heat material,Evaporated material condenses
on target substrate. High vacuum is required.
Thin-Iayerantireflection, scratch-resistant coatings are formed by
this technique.Sensitive materiais are etched by reactive chemical
species in charged plasma. Chemical change01 the etched material
takes place during this processoThe etching process is guided
bymaskant materiais.Pyrolysis can also be considered a high-energy
method. Application 01 high-energy source likefire to bulk
hydrocarbon materiais (Iike a steak) in the absence of oxygen
creates polyaromatichydrocarbons (PAHs)-a top-down process (or if
considering intermediates-for example,carbon atoms-it can be
considered to be a bottom-up process).Pyrolysis 01 solid refractory
nanoscale materiais like Si-C-N substrate to torrn nanotubes
at1500-2200C is a crossover technique [19].Large-scale synthesis 01
multiwalled carbon nanotubes occurs in liame environments by
burningcarbon sources such as methane, ethylene, or
benzene.Combustion can be considered to be a high-energy, thermal,
or chemical labrication method.Ultrasonication uses high-energy
sound waves to make nanomaterials trem bulk materiais.The technique
is also used to disperse carbon nanotubes in a suitable solvent.
The dispersion01 bundles 01 nanotubes into individual tubes is top
down. Probe tips are made 01 titanium,vanadium, and other metais
and alloys. Micron- to nanosized residual tip metal is introduced
intosolutions during the sonication processo
183
-
chemical top-down method, depending on the starting material.
The chemicalstructure of solid constituents is completely altered
following a combustionprocesso Nanosized PAHs and fly ash are
by-products of a top-down pyrolysisprocess, e.g.. the burning of
coal.
Chemical etching of solid substrates like a silicon wafer
(masked or other-wise) is a top-down chemical method. Chemical
etching processes, on the otherhand, adhere to a slightly different
classification criterion-specifically, thatchemical alteration
occurs only in the layers exposed to, and subsequentlyremoved from,
the solid substrate. In other words, although nanofacets orporous
structures are formed on or within the solid substrate, the
chemicalstructure of the solid substrate remains intact. Only the
surface is altered (passiv-ated, oxidized). The process of chemical
alteration is only applicable to substratematerial removed during
the etching process, e.g.. transformation of the solidinto a
water-soluble oxide.
Anodizing is a chemical etching process that involves
electricity (e.g., electro-chemical etching). This process is a
crossover technique and consists of fourparts.
l. Metal is electrochemicalIy removed top down from the surface
andreleased into solution in ionic forrn. AP+, during the anodizing
ofaluminum metal. The cationic products of anodizing are not
nano-materiais; they are ions.
2. HexagonalIy distributed. monodisperse scalIoped structures
[nano-facets) are formed on the surface of the aluminum ano de
duringanodizing. The diameter and curvature of individual nanoscale
scalIopsare dependent on the applied anodic voltage. This is a true
example oftop-down fabrication. The other two parts of the ano dize
equationare bottorn-up procedures.
3. The reaction of metal cations with anions originating from
the cathodereaction or with solution anions leads to the formation
of nanoscalecolIoidal oxides that eventually form the porous layer
(from the bottomup). Anionic species include oxides, hydroxides,
and other negativelycharged species (phosphates, sulfates.
oxalates, or chromates).
4. The hexagonal porous anodic oxide layer is formed from the
bottomup by the electrochemical reaction of AP+ cations with
various oxideanions. The scalloped top-down metal surface
structures direct thesize. orientation, and distribution of the
bottorn-up pore channels.
Overall, if we had to choose we should probably consider
anodizing as a top-down fabrication processo Top-down chemical
fabrication methods are listed inTable 4.4.
184 Introduction to Nanoscience and Nanotechnology
Top-Doum Lithographic Pabrication Methods. Many powerful
top-down tech-niques involve some form of lithography. Lithographic
techniques are whatmade the integrated circuit industry what it is
today, and it continues to be the mostviable method to form
nanostructures that actually has widespread applications.The
history of lithography was presented briefly in chapter 1.
Traditionally,electromagnetic sources ranging from the visible
wavelengths are still the mostpopular-especially in MEMS and
circuit fabrication. Ultraviolet and x-raysources are increasingly
in demand as smaller features are required. Electron
-
186
TABLE 4.5MethodLIGA techniques
Photolithography
Immersionlithography
Oeep ultravioletlithography(OUV)
Extreme ultravioletlithography(EUVL)
X-ray lithography(XRL)
Electron beamlithography (EBL)
Electron beamwriting (EBW)
Electron beamprojectionlithography (EPL)
Focused ion beamlithography(FIBL)
Microcontactprinting methods
Nano-imprintlithography (NIL)
lntroduction to Nanoscience and Nanotechnology
Top-Down Lithographic Fabrication Methods
CommentsLIGA is a German acronym for "Lithographie
Galvanoformung Abformung," a microlithographicmethod developed in
the 1980s. It was one of the first major techniques to demonstrate
thefabrication of high-aspect ratio structures. Beam sources
include x-ray, ultraviolet, and reactiveion etching. MEMS devices
are fabricated using LIGA techniques.Light is used to transfer
patterns onto light-sensitive photoresist substrates.
Photolithographyis primarily used in the manufacture of integrated
circuits and MEMS devices. The wavelengthrange of
opticallithography techniques ranges from the visible to the near
ultraviolet-ca.300 nm. The resolution of photolithography
techniques is -100 nm [20].
Just like with immersion optical microscopy, resolution can be
enhanced by 30-40% withapplication of a liquid medium between the
aperture and the sample with higher refractiveindex. The medium
needs to conform to the following criteria: (1) refractive index n
1, (2) lowoptical absorption at 193 nm ., (3) immersion fluid
compatible with the photoresist and the lens,and (4) be
noncontaminating.Resolution with deep ultraviolet with . = 248-193
nm, resulting in features on the order of 50 nm
Short wavelength ultraviolet, . = 13.5 nm. EUVL resolution: -30
nm [20].The major problem with EUVL is that ali matter absorbs
EUVand damage to substrates is verylikely. High vacuum is required
and mask must be made of Mo-Si.
X-rays are produced by synchrotron sources. XRL is capable of
producing features down to10 nm. Problems include damage to
substrate materiais.
An electron beam source is used instead of light to generate
patterns.Although e-beams can be generated below a few nanometers,
the practical resolution isdetermined by the electron scattering of
the photoresist material. Just like in SEM, electroninteraction
volumes are generated during exposure.Line width
-
TABLE 4.5(CONTO.)MethodNanospherelithography(NSL)
Scanning AFMnanostencil
Scanning probenanolithographies
2-Photonpolymerization
MethodErosionEtchingHydrolysis
Volcanic activity
Forest andbrush fires
Solar activityPressure andtemperatureBiologicaldecomposition
Digestion
Fabrication Methods
Top-Down Lithographic Pabrication Methods
CommentsNSL is used to fabricate nanometer-scale patterns. It is
a straightforward economical process withhigh throughput and high
resolution. It is difficult to categorize this technique as top
down orbottom up. Micron-scale latex spheres are often used as the
template material. The intersticesare nanoscale in size.NSL
utilizes nanospherical materiais in close-packed configuration as a
mask to aid in thefabrication of periodic particle arrays (PPAs).
Polymer nanospheres (diameter
-
(4.1)
188 lntroduction to Nanoscience and Nanotechnology
Bottom-up fabrication techniques are divided into four general
categories: (1)gaseous phase methods, (2) liquid phase methods, (3)
solid phase methods,and (4) biological methods.
Iust as with top-down methods, it is difficult to pigeonhole a
technique intoa general category. Many bottom-up processes are
characterized by tandemapplications of liquid and gaseous
techniques onto solid substrates. There arethree generalized states
of matter: gaseous, liquid, and solid. The distance dbetween
molecules in a gas is proportional to
where V is the volume and N is the number of molecules. For an
ideal gas atstandard temperature and pressure (STP), V = 22.4 L and
N = 'lI[N Avogadro'snumber, 6.022 x 1023. The distance between
atoms or molecules, centerto center,in an ideal gas is equal to
3.34 nm.
A liquid is a state of matter that has volume but not shape.
Although theatoms and molecules in a liquid are compressed as
tightly as a solid, the mole-cules in a liquid are free to move
randomly and unfettered. The distance betweenmolecules or atoms in
a solid is like that of a liquid, but random movement isseverely
restricted due to structural factors. Solids, of course, constitute
the mostcondensed form of matter.
A technique is designated as gaseous, liquid, or solid if the
process takes placein that appropriate medium or if the active
constituent from which nanomaterialsare formed is a gas, liquid, or
solid. Once again, some difficulty in nomendatureis encountered
when more than one phase is present during synthesis, but from
apractical point of view, such dassification is relatively
straightforward.
Bottom-Up Gas-Phase Pabrication Methods. Gases represent a
highly dispersedphase of atoms and molecules. Some nanomaterials
formed in the gas phase, likeclusters, remain in the gas phase.
More commonly, gas-phase precursors interactwith a liquid- or a
solid-phase material. lf one of the precursors of a nanomate-ria]
originates from the gas phase or if the reaction takes place in the
gas phase,we shall call it a bottom-up gas-phase fabrication method
(Table 4.7).
Nonbiological Bottom-Up Liquid-Phase Pabrication Methods.
Bottorn-up liquidmethods are numerous and diverse (Table 4.8). The
choice of solvent is anextremely important parameter in any
liquid-based bottom-up fabricationmethod. The liquid medium can be
hydrophilic or hydrophobic, ionic oranionic, or heterogeneous
(e.g., for the purpose of phase transfer of productbetween two
immiscible liquids). The new field of supramolecular chemistry
isconducted entirely in liquid media. AlI bottom-up biological
fabrication pro-cesses occur in liquid media. The liquid phase is
also where most chemists feelat home, and it is also going to be
one of the prime drivers of nanotechnology.Scale-up of liquid-phase
fabrication methods is a relatively straightforward pro-cess and it
is at the scale-up stage where the chemists turn over the reins of
aprocess to the chemical engineers.
Bottom-Up Lithographic Pabrication Methods. We add a special
category forlithography once again, but this time featuring
bottorn-up lithographic methods.
-
TABLE 4.7
MethodChemical vapordeposition (CVO)
Atomic layerdeposition (ALO)
Combustion
Thermolysis;pyrolysis
Metal oxide (MOCVO)Organometallic vaporphase
epitaxy(OMVPE)Molecular beamepitaxy (MBE)
lon implantation
Gas phasecondensation;thermolysis
Solid templatesynthesis
Fabrication Methods
Bottom-Up Gas-Phase Pabrication Methods
CommentsCVO involves the formation of nanomaterials from the gas
phase, usually at elevatedtemperatures, onto a solid substrate or
catalyst. Carbon nanotubes are formed by catalyticdecomposition of
carbon feedstock gas in inert carrier gas at elevated
temperature.
Single-walled carbon nanotube production by CVD requires
nanoscale Fe, Co, or Ni catalystplus Mo activator on high surface
area support (alumina) at >650C. Methane gas serves asthe carbon
source [26]. ALO is an incredibly precise sequential surface
chemistry layer deposition method to formthin films on conductors,
insulators, and ceramics. The layer formed by ALO conforms
tosurface topography. Precursor materiais are kept separate until
required. Atomic scalecontrol pinhole-free layers are formed.
AI203 layers are generated from hydroxylated Si substrate +
AI(CH3h(g), then H20 vapor isapplied to remove methyl groups. The
process is repeated until a target thickness is attained ..Layer
thickness: 1-500 nmThe formation of Si nanoparticles from the
combustion of SiH4 (silane gas) and othersilicon-containing gases
like hexamethyldisiloxane under low-oxygen conditions producesSi
nanoparticles as small as 2 nm. AI203 and Ti02 can also be formed
by combustion.Solid Si nanoparticles can also be formed by the
thermal decomposition of silane gas in theabsence of oxygen. The
bottom-up decomposition of ferrocene to form Fe nanoparticles isone
of the best examples of a bottom-up gas-phase fabrication
method.Chemical characteristics of precursor materiais utilize
reactive gas-phase-organometalliccompounds that decompose to form
nanometer-scale thin films or nanoparticles.H2 carrier gas, group
III metal-organic compounds + group V hydrides 500-1500Cat 15- to
750-torr pressure are representative conditions under which MOCVO
is performed.MBE is a thin film growth process conducted under high
vacuum. A heated Knudsen cell oreffusion cell is used to introduce
reactants by molecular beams. MBE ls able to deposit oneatomic
layer per application.Examples include alternate layers of GaAs and
AIGaAs with each layer of 1.13 nm in thicknessand InGaAs quantum
dots [27]. The temperature used in MBE is commonly 750-1 050C inH2
carrier gas.
This is a tough method to categorize. Nanovoids, for example,
can be created by ionimplantation of Cu ions into silica and
subsequent annealing [28]. It is bottom-up actionperformed on a
bulk material. If the ions come from a bulk source, it has a bottom
upcomponent. Once the ions are formed, ion implantation is bottom
up.Formation of Fe nanoparticles by decomposition of ferrocene at
200C is an example ofgas-phase process to form nanoscale
Fe.Formation of lithium nanoclusters by decomposition of LiN3 is
another example [7].Temperature at decomposition depends on the
material.Provides a solid template substrate for gas-phase
deposition of materiais on the solidsubstrate. This is considered
to be a mixed bottom-up system.Final nanomaterial size, shape, and
orientation are predetermined by template parameters.
Bottorn-up lithography methods are limited to a few kinds, based
on templateprocesses or direct writing (Table 4.9).
Bottom-Up Biological and Inorganic Pabrication Methods.
Biological processesare overwhelmingly formed from the bottom up
(Table 4.10). More detail isallotted to this topic in chapter
14.
189
-
190
TABLE 4.8MethodMolecularself-assembly
Supramolecularchemistry
Nucleation andsol-gel processes
Reduction of metalsalts
Single-crystal growthElectrodepositionElectroplating
Electroless deposition
Anodizing
Electrolysis in moltensalt solutions
Solid templatesynthesis
Liquid templatesynthesis
Supercritical fluidexpansion
lntroduction to Nanoscience and Nanotechnology
Nonbiological Bottom-Up Liquid-Phase Pabrication Methods
CommentsThis generic process is supported in liquid media. From
some perspectives, supramolecularchemistry is a subset of molecular
self-assembly. Almost ali molecular self-assembly takesplace in
liquids. The liquid plays a major role in supporting intermolecular
interactions andintermediate metastable species.
Supramolecular chemistry, for reasons to be explained m chapter
11, is conducted in liquidmedia. Weak intermolecular forces are
supported in liquids that allow many kinds ofintermolecular
interactions to take place. Ali significant biological metabolic
processes occurin a liquid rnedium.Precursor chemicals in a
supersaturated state combine by self-assembly ar chemical
reactionto form seed particles. Thermodynamics drives a nucleation
process that forms nanoparticles.The nucleation process depends on
prevailing conditions of pH, temperature, ionic strength,and time
[5]. Due to van der Waals attractions, colloids are formed.Sol-gel
methods are irreversible chemical reactions of homogeneous
solutions that result in athree-dimensional polymer. Sol-gel
methods yield nanostructured materiais of high purityand uniform
nanostructures formed at low temperatures [5]. Negative replicas of
colloidalhierarchical structures, upon drying, yield aerogels ar
xerogels. Such gels can be back-filledto produce nanocomposites ar
hybrid materiais [5]. These are ali pure bottom-up processes.Noble
metal clusters and colloids are formed by the reduction of metal
salts like HAuCI4 andH2PtCls. Common reducing agents come in the
form of organic salts like sodium citrate-Na3CsH57'By means of
phase transfer reactions (consisting of an interface between
twoimmiscible liquids), metal clusters and colloids are stabilized
by the addition of organic ligands.For exarnple, phosphine ar
thiols are adsorbed onto gold-55 to produce a stable cluster
[29].Nucleation process to form single crystals in liquid
mediaElectrodeposition is direct deposition of metais from metal
salt solutions to form thin layersar films on a solid conducting
substrate. Electrodeposition is an electrolytic process thatforms
thin metal films on the cathode of the cell. The process conforms
to Faraday's law.Electroless deposition is the autocatalytic
deposition of metais without electrical assistance.It requires
metal cation + catalytic (activated) surface + reducing agents like
formaldehyde,alkali diboranes, alkali borohydrides, ar
hypophosphorous acid. Pt, Ni, Co, Au, andnumerous other metais can
be deposited on many kinds of substrates, including
plastics.Electroless deposition has been used to create negative ar
positive replicas of porousnanostructures [30].
We have already characterized anodizing as a top-down processo
We mentioned earlier thatanodizing method contains a top-down
component (formation of scalloped structure).Here, we focus on the
bottorn-up formation of the porous alumina.
Aluminum metal is made the anode in an electrolytic cell
consisting of a polyprotic acid(usually sulfuric, phosphoric, ar
oxalic).Pore diameter of 200 nm; with pore density: 20-80+% and
film thickness: 100 11m.Anodized titanium several nanometers thick
generates bright interference colors.Utilization of molten alkali
halide salts with graphite electrodes with 3- to 5-A current
[31]Erosion at the cathode to form tubesThe product is transferred
to toluene.Provides a solid template substrate for electrochemical,
chemical, polyrnerization, and otherliquid-phase reactions. Most
methods are accomplished in a liquid medium.
Final nanomaterial size, shape, and orientation predetermined by
template parameters.Liquid templates (micelles and reverse
micelles) are commonly used to make quantum dotsfrom the bottom
up.Solvent removal under hypercritical conditions forms aerogels
and xerogels that containnanometer-sized voids. Supercritical
conditions imply that the medium is in neither liquidnor solid
phase.
-
TABLE 4.9MethodNanolithography:Dip-pen methods(DPNL)
Nanosphere templatemethods
Nanopore templatemethods (shadowmask evaporation)
Block copolymerlithography (BCPL)
Local oxidationnanolithography
STM writing
Fabrication Methods
Bottom-Ilp Lithographic Fabrication Methods
CommentsNanoprobe lithography in the form of dip-pen
nanolithography was invented by Chad Mirkin'sgroup at Northwestern
University in Chicago [32]. DPNL is considered as an
AFM-basedsoft-lithography technique. The operation of this method
is quite simple. A water meniscusis formed between an AFM tip and a
substrate. The AFM tip, in conjunction with the watermeniscus
conduit, is able to transfer molecules to the surface. The method
has high spatialresolution 10 nm), has high registration capability
(probe can both read and write), and isable to deliver complex
molecules such as DNA to a surface [20]. The major
disadvantage,like that of STM writing, is low throughput.Nanosphere
lithography is a template method for fabrication of nanomaterials.
Latex spheresare arranged on a substrate surface in various
configurations: hexagonal close packed,or into a square array. The
interstitial spaces between latex spheres serve as sites
throughwhich deposition can occur-a very straightforward, sim pie
processo Although thedistribution and placement of the spheres can
be considered to be a top-down process,the deposition of material
through the interstices definitely occurs from the bottom up.Use of
porous alumina membrane templates as templates to form arrays of
nanoparticles.The size of the nanoparticles can be controlled from
5 nm to >200 nm. The space betweennanoparticles can also be
adjusted. Nanoparticle aspects are adjusted by theheight of the
mask, the pore size, and the direction of evaporation [33].
This technique is good for direct patterning without the need
for additional steps such asetching or lift-off. The combinations
of masks, materiais, and substrates are enormous,and the process
allows for straightforward upscale.
Arrays have been used in the secondary fabrication of memory
devices and carbonnanotubes.
BCPs applied by spin-coating (top down) self-assemble into an
ordered array of nanoscopicdomains on a surface. Selective removal
of one component yields an etch mask. Thesubstrate pattern is
formed by plasma etching. In a specific example: a 35-nm
thickpolystyrene-PMMA copolymer layer is applied to a Si3Ni4-coated
Si wafer. Removalof the PMMA leaves an ordered array of polystyrene
nanodots. Reactive ion etching (REI)with CHF3 transfers the pattern
to the Si3Ni4 layer. The Si3Ni4-formed pattern is etched againby
REI with HBr. The result is an ordered array of silicon pillars
(wires) [34]. Blockcopolymer lithography was able to produce
periodic arrays of 1011 holes per crrr- [35].One problem that faces
this procedure is long-range order.A scanning probe tip (a dynamic
AFM tip) is placed a few nanometers above a substratesurface. The
environment consists of saturated water vapor. A bias voltage is
appliedbetween the tip and the surface. Oxidation of the surface,
if silicon, produces lines ofsilicon oxide. The breadth of the
meniscus and the distribution of the electric field withindetermine
the size of the feature [36]. Features as small as 7 nm were
produced.One-nanometer projections were formed in the
z-direction.
The IBM logo pictured in chapter 1 was fabricated bya bottom-up
method. Startingwith xenon atoms, each atom was manipulated by the
scanning probe tip into its finalposition. Other examples of this
technique include the quantumcottet- circular arrayof Fe atoms
placed on a Cu surface [37]. Ali scanning probe fabrication methods
arehindered by low throughput.
4.0.4 Nebulous Bottom-Up Fabrication CategoriesFabrication of
nanoscaIe materiaIs (structures, domains) within solids is
diffi-cult to pinpoint. lt is difficult to track the history of an
atom or moIecuIethroughout the course of a soIid material. Solids
contain a number of diverse
191
-
MethodProtein synthesis
Bottom-Up Biological and Inorganic Pabrication Methods
CommentsFormation of proteins from precursor amino acids by
elaborate process of protein synthesisTransfer RNA transports amino
acids to ribosomal RNA and link with peptide bonds.Synthesis of
nucleic material (RNA, DNA) from sugars, phosphate, and nuclides
(adenosine,guanine, cytosine, and thymine) from the bottom up
The processes of mitosis and meiosis are template (replication)
methods.Bottom-up agglomeration of Iipids, phospholipids to form
organized membrane structuresthat make life possibleMother of pearl
(nacre)95% Inorganic aragonite (platelets 200-500 nm thick) +
organic biopolymerDeformable nanograins [38]Nucleation depends on
P,T, concentration, and composition.Flaws reduce surface energy by
nucleation.Direction of growth depends on nanostructure.
192 Introduction to Nanoscience and Nanotechnology
TABLE 4.10
Nucleic acid synthesis
Membrane synthesis
Inorganic biologicalstructures
Crystal formationmethods
defects that have nanoscale dimensions. Are these considered to
be "nanornate-rials" or nanofacets? Or are they merely nanodomains
ofthe bulk type material?Voids formed by ion implantation do
agglomerate to form nanovoids from thebottom up. We address this
nebulosity in more detaillater.
4.0.5 The Nano PerspectiveThere are many kinds of nanomaterials.
When discussing fabrication methods,it is essential that the nature
of the end product be understood. For example,some types of
nanomaterials retain their nanoscale dimensions (e.g.,
quantumdots). Others form into components of more complex
structures (e.g., one-dimensional, two-dimensional, or
three-dimensional arrays of quantum dots).In these instances, the
quantum dot retains its identity as a unique nanomaterial.In other
cases, nanomaterials form the structure of an integrated bulk
material.An example of a bulk material that is composed of
nanostructured componentsis a Cu-Fe alloy in which nanodomains of
one or the other metal exist within abulk material. Steel made of
nanosized grains has better mechanical propertiesthan steel made of
micron-sized grains.
Silk, collagen, elastin, and keratin tissue found in animals are
composed of ahierarchy of increasingly larger structures [39]. The
hierarchy begins with sub-nanometer materiais and ends with a
functional macroscopic material [39,40].The relationship of
nanostructure, muscle fibers, and connective tissue is shownin
Table 4.11. A similar table can be created for bone tissue and
other organsystems in animal bodies. From the purely structural
point of view, it is clearthat nature 'begins' from the bottom up
to build any kind of macroscopicfunctional material.
Fabrication of inorganic nanomaterials is bottom up. but some
well-knownmethods such as erosion certainly operate from the top
down. The constructionof a snowflake is a nucleation process that
emphasizes eccentricities in the unitcell of each snowflake, a
bottom-up processo With regard to nanoscienceand technology,
materiaIs are constructed from the top down, bottom up. or a
-
Pabrication Methods 193
TABLE 4.11 The Nanostructure of Tendons
Structural componentAmino acidsCollagen
Dimensions Description/functionThe building blocks of
proteinsPrimary structure polypeptide (the protein ofconnective
tissue)
Three polypeptide strands form a cooperative~quatern~ry
structure.
-ct-nrn1.5-nm Diameter
Triple-helix coil (tropocollagen)
MicrofibrilsSubfibrilFibril
1.5-nm Diameter; 300 nmlength
-
194 lntroduction to Nanoscience and Nanotechnology
engineering and physics. Top-down fabrication dominates
nanotechnology today,although significant ground has been gained by
bottom-up methods [44].
Although tried and true, there are many challenges that confront
top-downmethods as miniaturization continues unabated towards the
nanoscale.Contamination, machine cost and complexity, dean roam
cost and complexity,physicallimits (photolithography), material
damage, and heat dissipation are afew of the issues that confront
top-down methods. There seems to be a strong linkbetween the cost
of a procedure and the size of the intended product.
Specifically,it becomes more expensive to make smaller materiais
and devices. According topundits, however, once the R&D phase
is accomplished and the manufacturingline is in place, the cost of
nanomaterial-enhanced products should go down.
A few selected top-down processes wil! be reviewed in the
following sections.There are many we leave out. For the purposes of
this course, a representativesample has been compiled that should
provide enough insight and informationinto top-down fabrication
methods.
4.1.1 Mechanical Methods (Mechanosynthesis)Any procedure that
involves the action of a bulk implernent. tool, ar machineon
samples made of bulk materiais is a top-down mechanical
method.Mechanical methods base their action on kinetic energy: a
hammer falling, acanister revolving, a roller thinning, adie
extruding, a compacter compressing.ete. Beating and rolling methods
to form thin metal films with nanometerdimensions and extrusion of
soft materiais in plastic phase to form wires arewidespread
industrial practices [5].
,.
Ball Milling. One of the most important mechanical top-down
methods isball mil!ing (and shaker milling), a technique that is
able to produce nanoscalemateriais by mechanical attrition. ln ball
milling, the kinetic energy of a grind-ing medium (e.g., stainless
steel or tungsten carbide ball bearings) is transferredto
coarse-grained metal, ceramic, ar polymeric sample materiais with
the directpurpose of size reduction [3]. Rotation or rapid
vibration of a drum ar canisterimparts kinetic energy to the
grinding medium (under controlled atmosphericconditions to prevent
oxidation) [5]. During the ball mil! process, severe
plasticdeformation of the sample material initiates the formation
of defects and dislo-cations. Any type of mechanical deformation
subjected to high sheer and strainconditions leads to the
forrnation of nanograined material [45]. Figure 4.1displays a
rendition of a generic ball mil!.
The result of the procedure, however, yields nanoparticulate
materiaispeppered with defects with a wide distribution of size. On
the upside, mechanicalattrition is one of the least sophisticated
technical processes and hence the leastcostly. Although the process
has roots in ceramic processing and powder metallurgyfor severa
decades, it is considered to be a rapidly evolving field [3]. Ball
milling,first accomplished by J. Benjamin in 1966, produces
mechanically alloyed mate-riais. Alloys, metastable phases,
quasi-crystalline phases, and amorphous alloysare formed by such
mechanical attrition techniques [3].
The principie of mechanical attrition is relatively
straightforward. A samplematerial is placed in a canister filled
with ball bearings. The canister is activatedand begins to rotate
at increasingly higher revolutions per minute. The ball
-
Pabrication Methods 195
1
Thete are two ens of the fabricaton specttum: At one enthere is
the high-priced lithogtaphic equipment that requires ahigh-vacuum
environment and expensioe energy sources. At theother en there is
the ball mill-a purely mechanical machinethat fabricates
nanomaterials by mechanical methods. Kineticenergy from a rotating
or vibrating canister is imparted to hardspherical materiais Uke
ball bearings. The ball bearings in turnreduce bulk precursor
materiais into nanoparticles.
bearings impart significant kinetic energy to the samples, a
much softer mate-rial. Several processes occur in the following
arder. The first event to happen iscompaction and then
rearrangement of particles. Secondly, elastic and
plasticdeformation and welding occur. Particle fracture and
fragmentation furtherreduce the particle size. Griffith theory
describes particle fracture in a mathe-matical sense:
(4.2)
where (JF is the stress at which crack propagation leads to
catastrophic failure,y is the surface energy of the particle
(joules per square meter), E is Young'smodulus, and c is the length
ofthe crack [3]. The tipping point is reached whenthe stress equals
the strength of cohesion between atoms of an isotropic solid.As
particles get smaller, due to enhanced surface energy,
agglomeration forces(antifracture) predominate. A balance is struck
among the stress, increasedresistance to fracture, increased
agglomeration, and maximum energy that isexpended in milling.
-
196 Introduction to Nanoscience and Nanotechnology
There are several types of mechanical attrition devices. Shaker
mills are themost popular form used by scientists and are able to
produce particles 1000 cycles-rnirr ', ball velocity>5 m S-I)
applied to a vial with milling balls ensures that samples
pulverizeproperly. Planetary ball mills are commonly used in
laboratories. In this form ofmechanical attrition, rotational
forces are the source of kinetic energy impartedto the grinding
media and the sample.
Compaaioti and ConsoZidation. Following a ball mlling process
(e.g., of a compo-sition that consists of copper and iron metal
constituents), materiaIs are compactedwith a tungsten-carbide dye
under high pressure for extended periods of time [7].After
compaction, heat is applied, also under pressure, to the alloy. The
result is ametal formulation that is characterized by an average
grain size of 40 nm within arange of 15-75 nm. Thewhole point
ofthis procedure is to produce a material withsmaller grain size
that demonstrates superior physical properties to that of a
materialwith larger grain size. Nanograined alloys demonstrated
fracture stress that was fivetimes betterthan pure iron with larger
grain size (50 nm-150 um) [7,46].
Compaction of ceramic and superconducting nanomaterials by
applicationof shock waves limits the grain growth [47]. Ceramic
superconductor materiaisformed by such advanced techniques
demonstrate higher current capacity, largermagnetic fields, and no
energy loss through resistance.
Chill Block MeZt Spinning and Solidification. This is a process
that initiallyapplies heat to bulk material with the intent of
melting that material and per-forming an extrusion processo Quick
solidification of the metal is induced tofreeze the metal into a
desired formo An RF (radio frequency) heating source isutilized to
create a metal melt. The liquid metal is then forced through a
nozzlein the form of a stream that is oriented over the surface of
a rotating drum [7].A bulk alloy material consisting of aluminum
nanoparticles, 10-30 nm in size,made by this method demonstrated
tensile strength in the gigapascal range.
4.1.2 Thermal MethodsA top-down method is considered to be
thermal if an external source of heat isapplied to the processo
Melting a bulk material and converting the liquid intonanomaterials
are considered to be a thermal top-down method. Many meth-ods
produce heat during operation, such as laser ablation and solar
flux, but areconsidered to be high energy rather than thermal
methods per se.
Gas Atomization. This is another top-down method that is suited
for the manu-facture of nanoparticulates. In this process, a
high-energy stream of some inert gasis directed at a molten metal
stream. Iust like in the ball milling, kinetic energy istransferred
to the sample=-this time from the high-energy inert-gas beam.
Theimpact initiates the formation of finely divided metal particles
that upon solidifi-cation form into a finely divided powder. The
nanopowder is then compacted toform a bulk metal with superior
mechanical properties.
EZectrohydrodynamic Atomization (EDHA). Electrohydrodynamic
atomizationis an offshoot of electrostatic spray technology and is
a subset of liquid disrup-tion processes. The formation of a Taylor
cone that terminates in a fine-stream jet
-
Fabrieation Methods 197
forms the basic mechanism of EHDA. An electrostatic atomizer
causes a netcharge to develop on the surface of a droplet that
causes dispersion due to cou-lombic repulsive forces. This process
prevents agglomeration of droplets andhence partieles are formed.
The EHDA process is capable of producing partielesas small as
quantum dots.
The products of EDHA procedures depend on the flow rate of the
liquid, thediameter of the needle orifice, the distance between the
needle tip and groundedsurface, and the strength of the applied AC
field [48]. One of the primary goalsof this procedure is to be able
to synthesize nanopartieles rapidly and over largeareas. The EHDA
technique was used to atomize a solution of chloroplatinicacid
[H2PtCI6 (H20)6] in ethanol. The purpose of the atomization
procedurewas to produce Pt metal partieles. Droplets are sprayed on
a Si-Si02 substrateand heated at 700C for a short period of time.
The dimensions of the Pt parti-eles were on the order of 10 nm
[8].
4.1.3 High-Energy Methods
Arc discharge, laser ablation. and solar vaporization are three
high-energy top-down methods that are able to generate
nanomaterials by the application ofhigh energy electric currents,
monochromatic radiation, or solar radiation to asolid substrate.
Each method is capable of forming carbon nanotubes fromgraphite
substrates that contain catalytic Fe, Mo, or Co partieles. We
considerany process that involves plasma to be a high-energy
processo High-energy meth-ods, with the possible exception of the
solar version, are not practical to upscaledue to the intense
investment in energy that is required.
Are Discharge Plasma Method. The first deliberate attempt to
produce carbonnanotubes with an arc discharge method was
accomplished with an arc plasmadischarge method developed by Y.
Ando in 1982 [15,16]. The formation ofcarbon nanotubes by arc
discharge (plasma arcing) process is dependent onthe pressure of
He, the process temperature, and the applied current.
Typicalconditions utilize an applied voltage of 20 V, current
ranging from 50 to 100 A,and He pressure of 50-760 torro Two
graphite rods are plaeed millimeters apart(Fig. 4.2). The
sacrificial anode consists of graphite that is doped with
metaleatalyst partieles. In this configuration, single-walled
earbon nanotubes arefabricated. Multiwal!ed earbon nanotubes are
formed if no metal eatalyst ispresent in the graphite. At 100 A,
earbon vaporizes in a hot plasma. Carboneations are formed at the
anode and the soot is eol!ected at the cathode. Theare method,
although relatively simple, produces an array of unwanted
by-produets. Samples originating from are diseharge methods often
require extensivepurifieation. Basing scientifie conelusions on
unpurified materiaIs is not arecommended practiee.
Laser Ablation of Solid Targets. In 1995, carbon nanotubes were
synthesized bypulsed laser method. Graphite rods containing Co and
Ni catalyst were heated to1200C and then exposed to laser pulses
[17]. Heat is, therefore, generated bytwo means in this process-the
furnaee and the laser. The vaporized earbon iscollected on a cooled
finger downstream of the earbon targets. Continuous waveCO2
(,..,2kW) infrared, ultraviolet, or Nd:YAG lasers are the most
common typesoflasers used in the ablation method. A generic scheme
is shown in Figure 4.3.
-
198 lntroduction to Nanoscience and Nanotechnology
Fullerenes and carbon nanotubes were first fabricated in
arc-discharge apparatus. Once again, a relatively simple
mechanismis capable of fabricating nanomaterials-in this case with
nomovingparts. A high-energy spark is created between two
closelyspace electrodes, the anode of which contains a graphitic
targetmaterial.
Graphiterod
,,:f C-soot~-----~
8
FIG.4.3
Laser ablation is a cleaner means of fabricating
nanotubesthanarc-discharge. Graphite targets are placed insie a
quartztube.The tube is heate to ca. 1200C. Vaporized carbon
ptoducts arecollected on a cold finger downstream of the ablation
processo
High-Flux Solar Furnace. Solar power has also been used to
fabricate carbonnanotubes by a top-down procedure [18]. Since
scale-up of the arc dischargeand laser ablation methods is
problernatic, the goal is to increase the powerofthe solar furnace
to a levei of 500 kW [49]. At the National Renewable
EnergyLaboratory (NREL) in Golden, Colorado. researchers were able
to producefullerenes from a lO-mm diameter graphite pellet with a
lO-kW high-flux solar
Ar, 0.5 bar
--Ar, 0.5 bar (;>. Nd: YAG-Iaser,U Q-switched, 1064 nm
-
Pabrication Methods 199
fumace (HFSF) [18 J. Temperatures in the range of 3000-4000
Kwere attained.NREL's high-flux fumace has 25 hexagonal mirrors to
concentrate solar radia-tion that provide flux at 2500 suns ar,
with adjustments, 20,000 suns-quiteimpressive.
Plasma Methods. We place ion milling, RF sputtering, plasma
cleaning, andreactive ion etching into the category of high-energy
methods. Plasma (fromthe Greek plasma ar plassein, "to mold, to
spread") is an tonized gas that is con-sidered to be a distinct
phase of matter. Plasmas contain ions and electrons andexist best
in a vacuum environment for obvious reasons. Plasmas are
electri-cally conductive and are strongly influenced by electric
and magnetic fields.A simple reactive ion etching system is shown
in Figure 4.4.
Reactive ion etching (RIE) is an effective means of subtmcting
material from a substrate-hence, a top-oum method. Molecules
(usually oxygen, sulfur hexaJluoride, Jluorine, or otherreactive
species) are ionized to form chemically reactive plasmas bY the
action of an appliedelectromagnetic field (parallel plate
configuration) under low-pressure conitions. The appa-ratus
consists of a cylindrical chamber kept under a fow millitorr vacuum
conditions.Inductively coupled plasma (ICP) produced by RF magnetic
fields is another mode of creatingRIE plasmas. Combinations of
parallel plate and ICP alSo existo Since the trajectory of
onsproduced in RlE is mostly normal to the plane of a substrate,
the process is capable of aniso-tropic etching~as opposed to
chemical etching, which tends to act in an isotrCJf1icashion.
Upper electrode
Chamber
Wafer
PI"m: :o:~"o:":')O o O O OO O O
O O O OO o O O O
Chamber :-vacuumpump
Exhaust outlet
RF generator ~
-
200 Introduction to Nanoscience and Nanotechnology
RF sputtering is a physical (as opposed to chemical) vapor
deposition (PVD)method. Atoms from a solid target source (hence the
top-down designation) areejected via the process of momentum
exchange into the plasma by the action ofhigh-energy ions, usually
originating from argon. The ejecta are then depositedon a surface
of a sample material to provide a coating. A radio frequency
altemat-ing current is commonly used to generate the plasma and a
bias voltage isapplied to the target to promote acceleration of
ions.
Ion milling, another PVD process, is similar to RF sputtering
except that nocoating is formed. In actuality, the opposite is
true. Material is removed to promotethinning or shaping of a sample
material (e.g., formation of nanofacets). Thinfilms with dimensions
on the order of a few nanometers for the purposes ofTEM preparation
are formed after exposure to ion mill plasmas. Reactive ionetching
is a chemical process in which a reactive chemical species is added
to theplasma mixture. Oxygen, fiuorine derivatives, or etchant
species that are knownto react with targeted substrate materiais
are commonly used in reactive ionetching (RIE) procedures.
4.1.4 Chemical Fabrication MethodsCombustion of Bulk Materials.
Combustion is a top-down chemical methodthat is capable of
producing nanomaterials. However, impurities in bulk
carbonmateriais such as coal and oil contain contaminants that
contribute to the for-mation of fly ash and acid aerosols.
Polyaromatic hydrocarbon clusters (PAH)can be produced under
incomplete combustion conditions. Pure hydrocarbonsproduce CO2 and
H20 under efficient combustion conditions. Combustionalso is a
bottom-up method that is capable of producing na no materiais.
Following the combustion ofbulk Mg to MgO, a cluster-based
nanoparticlebonding mechanism was the cause of agglomeration. This
is apparently acommon phenomenon that applies with equal validity
to titania and aluminaparticles. For alumina, it was found that the
primary Al203 aggregate was on theorder of 1 um in size, but that
it was composed of clusters 10 nm in size [50].
Chemical Etching of Silicon. Chemical etching is important in
numerousindustrial production procedures, lithography in
particular. The anisotropicetching of silicon with KOH is a major
industrial procedure. The reaction yieldssilicates [51]:
(4.3)
The Si(110) surface undergoes the fastest etch rate of all the
primary low-index planes surfaces. For example, the etch rates of
Si in a 30% w-w solutionof KOH at 70C for the (110), (100), and
(lll) surfaces are equal to l.5, 0.79,and 0.005 um min-1'.A common
isotropic etching solution used for silicon isHNA (HF + HN03 +
CH3C02H). Isotropic etchants operate independently ofcrystal
direction. The trench profile following isotropic etching looks
like aninverted "C" by uoss-section; from anisotropic etching, the
trench looks like a"V" with a flat bottom [52]. Etching with
hydrofiuoric acid is driven by thestability of the [SiF6P-
complex:
(4.4)
-
Pabrication Methods 201
As a result of lithographic procedures and subsequent top-down
chemicaletching, nano- to mieron-sized features can be formed on
the surface of siliconwafers.
Chemical-Mechanical Polishing. This method is a combination of a
chemicaletching and a mechanical attrition method. The process of
polishing jade witha corundum-based abrasive has been traced back
to Neolithic farmers in ancientChina 6000 years ago [53]. Grinding
is the planar rernovl of material frorn atarget surface by a tixed
abrasive. In polishing, the abrasive is allowed to roll. Thesurface
roughness, determined by profilometers or AFM, is shown to be a
fewnanometers.
Chemical-mechanical polishing combines the mechanical grinding
charac-teristics of abrasives with the chemical action of an
etchant. Pressure is appliedon the abrasive and hence on the
surface through a conformal pad. This allowsfor free movement of
the abrasive under the pad. The method is important tothe
lithography industry, where depth of focus (DoF) is ever shrinking
withsmallerwavelength sources and larger numerical apertures
(N.A.). The smootherthe surface of a Si wafer becomes, the better
is the accommodation of shrinkingDoF.
Anodizing and Electropo1ishing. These two techniques are
integrally relatedand differ only with regard to purpose and
conditions. Anodizing is a processthat creates an insulating porous
oxide layer on a conductive metal anode, usu-ally aluminurn, in an
electrolytic solution, usually a dilute polyprotic acid.
Byproviding hexagonally packed pore channels that are sim pie to
fabricate and theability to manipulate pore diameter and length
during and after anodizing, theporous anodic film offers a perfect
template for nanoscale material synthesis.Anodizing conditions
consist of an electrolytic bath made of a polyprotic acid(H2S04,
H3P04' H2C204, ar H2Cr04) at OCwith applied voltage of 2-100 V
de.The formation of nanoscale pores with diameters ranging fiam a
few nanometersto several hundred nanometers is the major product of
anodizing. The chemicalreactions in anodizing are
Anodic reaction (4.5)
Oxide-electrolyte interface (4.6)
Cathodic reaction (4.7)
Overall oxide formation reaction: 2Alo(s)+ 3H20 ~ Al203(S) +
3H2(g) (4.8)
Anodizing, however, is a mixed fabrication method. Technically,
it containscomponents that can be classified as top down or bottom
up. The top-downcomponent is the electrochemically assisted
dissolution of bulk aluminum toform AP+ cations. During this
process. nanostructured scallops are formed inthe surface of the
aluminum metal. Pore diameter is directly proportional to
theapplied anodic de potential (dpore oc 1.4 V) and is controlled
by the diameter ofthe scallops on the metal surface. A schematic
illustration of an anodic film isshown in Figure 4.5.
-
202 Introduction to Nanoscience and Nanotechnology
Electropolishing involves the removal of metal to form a smooth
surfacewithout forming a thick oxide layer. The conditions for
electropolishing arerather severe 'compared to anodizing: elevated
temperature (70-90C), elevatedlevei of current (10-20 A), and
concentrated acid or base solutions. Electro-polishing often
precedes anodizing to prepare a smooth surface.
Hydrolysis Reactions. These reactions can affect inorganic,
organic, and biologicalmateriaIs. Hydrolysis occurs by the action
of water to disrupt a bond. The bondcan be as strong as a covalent
bond, ionic bond, or any kind of intermolecular
-
Pabrication Methods 203
attraction. For example, dissolution of proteins from the top
down by acid-catalyzed hydrolytic mechanisms is a common means to
regenerate the con-stituent amino acids. The degree of hydrolysis
determines the size of the finalproduct.
4.1.5 Lithographic Methods
A brief history oflithography was presented in chapter 1.
Lithographic methodsare the most widely utilized industrial process
in the high-technology sector.The computer industry. for example,
depends heavily on lithography. Integratedcircuits,
microelectromechanical machines (MEMS), and numerous
otherapplications require lithography during some phase of their
manufacture.However, challenges facing lithography today are
numerous as well. Fabricationof increasingly smaller features
requires sources with smaller wavelength. Withincreasingly srnaller
wavelength (e.g., electron beams and x-rays), the resolvingpower of
the procedure is enhanced but the substrate sustains more
damage.Fabrication of increasingly smaller features also requires
increasingly moreexpensive equipment. Wavelength-based lithographic
techniques, although wellestablished. are rather costly to
operate.
Modern opticallithographic techniques utilize radiation sources
with wave-length from a few to 300 or 400 nm. Nano-imprint and
nanosphere lithographyoffer cost-effective facilitative
alternatives to the high-vacuum, high-energy,high-maintenance
processes. Once a few fundamental technical issues in
thesenanotechniques become better resolved, expect wavelength-based
lithographyfabrication to start giving way to these
nanorevolutionary procedures. With theadvent of nanosphere and
nano-imprint lithography, both extremely simplemethods capable
ofhigh resolution, the trend in operation costs may be reversedin
the near future.
In general, the underlying operation oflithographic techniques
has not changedmuch since the time of the inventor of the
technique, Bavarian author AloisSenefelder, in 1796.
Photolithography fo11ows the general procedure of patterntransfer
established by Senefelder but employs radiation or particle
projectiononto a resist material instead of writing on a limestone
substrate (Fig. 4.6):
Deposition of thin layer on substrate ~ deposition of
photoresist material ~exposure via mask (the master) by energy
source ~ development by etch (positiveor negative replica) of
excess material ~ stripping of a11 resist ~ chemicalmodification
(additive or subtractive)
There are numerous energy sources employed in lithographic
processes-visible to ultraviolet radiation and x-rays for
photolithography. Electron and ionbeams have also been applied in
lithographic procedures. Top-dowr. nanolitho-graphic sources
consist of photons (UV, DUV, EUV, and x-rays), particle
beams(electrons and ons}, physical contact printing (nano-imprint
methods), andedge-based techniques (shadow evaporation). Bottom-up
nanolithographicprocedures like dip-pen lithography and
self-assembly (surfactant systems andblock copolymers) will be
discussed in a later section.
There are three primary considerations for any lithographic
process: resolution,registration, and throughput. Resolution, first
discussed in chapter 3, is defined asthe best attainable physical
scale of a feature: the smaller the better. Registration
-
204
Directionof
process
lntroduction to Nanoscience and Nanotechnology
Lthography is the workhorse of the compuser chp industry. It is
the most common top-doummanufacturing process and it is one
nanomanufacturing technique that is undesprea. Atarget material is
first appled to the surface of a silcon substrate. Polymeric tesist
layeris then applied b1'spin coating. An energy beam, usual'y in
the visible to ultraviolet wave-length range, is shined through a
mask that contains a pteetetmine pattetn. Regionseepose to the EM
radiation are sensitize (positive resist) or protecte (negative) to
the sub-scquent eteh step. Following etching, the resist is
removed, transferring the fJattern inscribeby the mask to the
target material. Lithography is a rather expensioe process that
requiresclean room condions, high-vacuum conditions, and othetunse
expensiue equipment.
Application of photosensitive resist material (polymer)on target
material on substrate surface
_ Resist polymer material---Target materialSubstrate----L ~_
Sensitization/activationphotoexposure
_--- EM radiation
Etching process to form features on substrate surface
Removal (stripping) of resist material to leave surface
features- .~,----. - -,--.----is the process of aligning one layer
to another to form an integrated structure.Throughput is a gauge of
the balance between cost effectiveness and the rate ofproduction.
"
Optical Photolithography. Opticallithography employs visible and
ultravioletradiation to transfer a pattem onto a receptive
substrate. Ultraviolet radiation(deep ultraviolet lithography, DlN)
is the most common kind in use today.Three general methods are used
to expose wafers:
l. Contact prnting. in which the mask lies on top of the resist
(e.g.,there is no wafer-mask gap), requires no magnification but
resolution islimited (N500 nm). The mask degrades in this
configuration resultingin loss of planarity.
2. Proximity printing places the resist in dose proximity to the
mask.There is no magnification with this configuration and
resolution iseven lower (N1 um). Diffraction effects limit the
accuracy ofthe patterntransfer process [54J.
-
Pabrication Methods 205
3. Projection printing is a widely adopted technique. An image
isprojected through a mask and reduced by a factor of four to ten
timeson the resisto Resolution is much better (".,70 nrn}, but
equipment iscostly and accuracy is limited by diffraction [54].
A computer-generated pattem on a mask (from optical or electron
beamgenerators) is transferred to a chromium surface (".,100nm
thick) on fused silica[55]. The mask is then positioned over a
substrate, usually silicon. silicon oxide,or a semiconductor
material. The substrate is prepped beforehand with a thinlayer of
oxide, nitride, or other functional material, and then a
photoresist mate-rial is applied by spin-coating-a photoresist
material that is sensitive to the typeof radiation used in the
lithographic procedure. In opticallithography, the photo-resist is
illuminated through a mask and is rendered soluble (positive
resist) orinsoluble (negative resist) during the subsequent
developer step. The exposedresist (positive) or the unexposed
resist (negative) is removed byetching. Farexarnple, in a negative
scheme, the exposed resist polymer becomes cross-linkedafter
exposure to the radiation. Cross-linking implies that the resist
material ismore difficult to dissolve than areas that were
unexposed. Following develop-ment, an additive process deposits
material onto or into the etched areas. Insubtractive processes,
material may be removed by ion milling through thedeveloped areas.
Following these steps, the remnants of the resist are removed.
Resolution in projection lithography is diffraction limited but
has improvedover the years since the days of the first integrated
circuits. Line widths of the late1960s were on the arder of 5 um
[56]. In 1997, this was reduced to 350 nm.Today, sub-100-nm line
widths are commonly achieved. Some of the equationspresent below
willlook familiar.
Far contact style printing, radiation interacts with the sample
as a squarewave with limited or no diffraction. The near field (or
Fresnel diffraction limt),appropriate for proximity printing, and
resolution are given by
W=k..jId; (4.9)
where dg is the mask-to-wafer distance (gap), is the wavelength
of the imping-ing radiation, and k is a constant that is dose in
value to 1 and depends on resistmaterial and other technological
parameters associated with the processo Fresneldiffraction occurs
when
(4.10)
and the minimal resolvable feature is
(4.11)
For the projection style of lithography (the most commonly
applied form),the optical condition is called far field and the
mask is called a mask in the farfield. The optical description of
far-field lithography is similar to other types ofprojection
methods, whether optical or electronic. The minimal resolvable
fea-ture in a projection lithographic system is
-
206 lntroduction to Nanoscience and Nanotechnology
(4.12)
N.A. = n sin a (4.13)
where is the wavelength of the radiation used for exposure and
N.A. is thenumerical aperture of the optical lithographic
instrument (usually equal to",0.5). The facto r kj is a constant
for a specific lithographic procedure thatdepends on the index of
refraction and thickness of the photoresist material(0.4-0.8, a
quality descriptor). ln general, the line width is approximately
equalto the wavelength of the incident light.
The resolution limit for optical lithography is given by the
followingequations:
where N.A. is the numerical aperture, n is the refractive index
of the medium (ifvacuum, n = 1), and ais the half-angle of the cone
of light that can enter or exitthe lens. Does this look familiar?
The numerical aperture is a function of thedistance between the
lens and the sample and is an indication of the resolvingpower of
the system. The larger the numerical aperture is, the higher is
theresolution capability of the instrumento The following equation
should lookfamiliar as well; it also applies to lithography:
R = 1.22f = l.22fd n(2f sina)
0.6U 0.6U-----n sin N.A. (4.14)
DoF= k2---2
(N.A.)(4.15)
Depth of focus (DoF) (like depth of field) becomes a concern as
resolution isincreased in shorter wavelength tools. DoF is the
distance from the objectivelens that yields a focused image and it
gets worse (smaller) as N.A. becomeslarger:
where k2 is a constant associated with the photolithographic
system and is tradi-tionally equal to 1.
Contact, proximity, or projection modes are commonly used
photolithogra-phy techniques. Contact type of photolithography (or
shadow mode) is the casewhere the mask is right on top of the
resists. Resolution is calculated from
(4.16)
where 2b is the grating' period of a mask with equally spaced
lines and d is thethickness of the resist material. ln the
proximity method, a gap exists betweenthe mask and the photoresist
and its resolution is found from
(4.17)
where dg is the distance between the mask and the resist.
-
Pabrication Methods 207
Particle Beam Lithography (IPL). Because panicle beams do not
undergo dif-fraction and scattering is minimal, higher resolution
can be achieved with IPLthan optical, x-ray, or electron beam
methods. Resist materiaIs demonstrategreater sensitivity to ions
than to electrons. Ion lithography is mostly used torepair masks in
optical and x-ray lithographic procedures.
Extreme Ultraviolet Lithography (EUVL). This technique applies
radiation thatis as short as 11-14 nm, significantly lower than
those used in DUVL [57,58].Features smaller than 50 nm have been
achieved, but theoretically much smallerfeatures are possible
(e.g..
-
208 Introduction to Nanoscience and Nanotechnology
Nano-mprint lithography (NlL) was invented in the mi 1990s by
Stephe Chou of PrincetonUnversity. NIL is a method that is capable
ofgeneratingfeatures with nanoscale resolutionwith high throughput.
A stamp (top), mae by an electron beam lithographic method,
ispresseilinto a soft polymeric material at its glass transition
temperature. The film is hardenedbefore the stamp is remove.
Anisotropic etching by RIE process removes excess polymer.Metallic
pillars are fabricated n the nanofeatures formed by NIL.
Nanoimprint lithography
TOPOgraPhic{masterstamp
- Silicon substrate- Silicon oxide mold
Polymer heatedabove glasstransition
temperature
Poylmethylmethacrylate--- impressionable material-Silicon
substrate
1Polymer curedbelow glasstransition
temperatureafter imprinting
Stamp removed
Reactive ionetching used toremove excesscompressedresist
material
Reactive ion etchingand pattern transler
Gold pillars deposited into template
Remainder 01 PMMA removed with organic solvent
25-nm features spaced 70 nm apart. It is easy to understand why
NIL is preferredover wavelength-dependent lithographic techniques:
(1) NIL is able to achievesmaller features, (2) NIL takes less
time, and (3) NIL is an inexpensive processthat does not require
ultrahigh vacuum and expensive radiation or electronbeam equipment.
The biggest problem with NIL is defectivity; although recentmethods
have driven the defect density to
-
Pabrication Methods 209
resolution. Compared to other methods, NSL is quite fast and
economical.The NSL process is able to create ordered arrays of
differing configurations. Inone case, materiais are deposited
through the open spaces between spheres toform an array (Fig. 4.8).
In another application, the size of the spheres is
Nanosphete lithography (NSL), lme NIL, is another ingenious
low-cost, high-throughputmethod to form nanomaterials and
nanopartide arrays. One simple method utilizes latexspheres that
are dose pached in a two-dimensional array. Deposion af metal
betweenspheres, the interstitial spaces, forms star-shaped pattems
of tetrahedrally [orme nano-structure materials. RIE etching,
depending on the type of active molecule, is able to reducethe size
of the spheres (thereby creating wider gaps among the spherical
matrix elements,or etch) in an anisotro ic manner, the substrate
under the s aces to orm pore channels.
y y y y y y y y y y y y y y
y y y y y y y y y y y y y y y y y y
-
210 Introduction to Nanoscience and Nanotechnology
reduced by standard RIE with oxygen [63]. Col umnar arrays
several micronsin height are then fabricated by application of a
deep-RIE (Bosch) top-downprocesso
The NSL is a mixed bago It is a bottorn-up template method
during the depo-sition of material at the base of the interstitial
regions of the spheres, but it is atop-down method during the
deep-RlE processo
Applications of NSL include its use in manufacturing
size-tunable noblemetal substrates in the range of 20-1000 nm. The
optical response of NSL-formed Ag nanoparticles to their local
environment was probed by localizedsurface plasmon resonance
spectroscopy (LSPR) [62]. Results showed that zep-tornole-Ievel
detection of adsorbed analytes was possible by LSPR
spectroscopy[62]. Large-scale fabrication of protein nanoarrays
based on NSL was demon-strated by Y.Cai et al. in 2005 [64]. Based
on nanospheres with 300-nm diame-ter, protein islands were formed
with ring shapes of 50-nm width and 1I8-nmdiameter [64].
4.2 BOTIOM-UP FABRICATIONBottorn-up fabrication approaches
selectively combine atoms ar molecules toform nanomaterials.
Bottom-up fabrication methods, therefore, are consideredto be
additive. Bottom-up fabrication methods reside within the realm of
chem-istry and biology. Nature, of course, has perfected bottorn-up
fabrication ofnanomaterials.
Advantages of bottorn-up methods are numerous. Self-assernbly
processes,for exarnple, occur under thermodynamic control
conditions. Because such pro-cesses exploit much weaker
intermolecular interactions, as opposed to strongcovalent bonds,
nanomaterials are fabricated under milder conditions of
tem-perature, pressure, and pH. The upscale potential ofbottom-up
methods is enor-mous. As with any other chemical process, it is
"relatively" straightforward toscale up a process that takes place
in a beaker on a lab bench (e.g., the domainof the chemist) to a
batch production process in a manufacturing line (e.g., thedomain
of the chemical engineer). However, there exist significant
challengesfacing bottorn-up methods. OveralI robustness, long-range
order (related tocomplicated patterns), and directed growth all
leave something to be desired.ln order for bottorn-up fabrication
of nanomaterials to become the dominantfabrication mo de of
industry, alI of these concerns need to be overcome.
We divide bottorn-up methods according to the phase within which
theprocess occurs. We also add a special section discussing the
solid state.
4.2.1 Gaseous-Phase Methods
Vapor phase reactions can be homogeneous (all reactants,
prcducts, and cata-lysts exist as a vapor) ar heterogeneous
(vapor-Iiquid or vapor-solid phasesexist within the same sphere of
reaction). lf there exists a vapor (or any highlydispersed phase,
e.g., a particle beam) in a process, we shall consider that
pro-cess to be a gaseous-phase fabrication method.
Chemical Vapor Deposition. Chemical vapor deposition (CVD) is
one of themost effective procedures used to produce advanced
materiaIs. CVD is the best
-
Pabticatum Methods 211
way to form carbon nanotubes because it is less energy intensive
and more con-trol is exerted over products. Si02, SiC, Si3N4, W,
and other materiais are rou-tinely deposited on surfaces via CVD
methods. In semiconductor industrypractice, wafers are exposed to
volatile precursor materiais that react or decom-pose on the
surface to form thin films. There are many kinds of CVD.
Chemical CVD (CCVD) is used in the fabrication of carbon
nanotubes-single-walled and multi-walled operating temperatures
range from the low400C to produce carbon fibers and multiwalled
carbon ~anotubes to tempera-tures >1000C. The decomposition of
methane, ethane, ethylene, propane,propylene, ar acetylene or the
disproportionation of carbon monoxide-allover catalysts-is an
example of some of the carbon source materiais (usually ingas form)
used in CVD techniques. The decomposition of methane in the
pres-ence of catalysts (usually Fe, Ni, or Co) at temperatures of
700C at atmosphericpressure yields SWNTs.
CH4(g)-7SWNT+H2(g) (4.18)
Polysilicon thin films are formed by the decomposition of
silanes in a low-pressure CVD (or liquid-phase CVD, LPCVD) chamber
at ca. 650C. If othergases, such as phosphine or arsine, are
present in the strearn, the silicon can bedoped in situo Silicon
dioxide layers are formed by the gas-phase decompositionof
tetraethylorthosilicate (TEOS). Since TEOS boils at ca. 168C, the
CVD pro-cess is conducted between the boiling point ofTEOS and
750C. TEOS breaksdown into solid silica and gaseous
diethylether:
Si(OC2Hs)4(g) -7 Si02(s) + 0(C2Hs)2(g) (4.19)
Plasma-enhanced CVD (PECVD) is another bottom-up CVD
fabricationmethod to produce thin films. The plasma is created by
radio frequency ordirect current discharge between electrodes [65].
Silicon dioxide, fromsilanes + O2 or TEOS + O2, can be formed with
the PECVD technique atreasonably low pressure (N 100 mtorr).
Silicon nitride thin films are alsodeposited with plasma
assistance. An example of a CVD apparatus is shownin Figure
4.9.
Metal oxide CVD (MOCVD) utilizes H2 as a carrier gas, Group-III
metal-organic compounds, and Group-V hydrides to make nanometer
scale thin filmsor nanopartides. Temperatures rangingfrorn 500 to
1500C at 15-750 torrpressure are representative conditions under
which MOCVD is performed.
Atomic Layer Deposition. Atomic layer deposition (ALD; a.k.a.
atomic layerepitaxy, ALE) was introduced in 1974 by Tuomo Suntola
of Finland with theintent of improving the quality of ZnS films
used in electroluminescence dis-plays. After a decade of
developrnent, high-quality phosphor layers anddielectric layers
were produced, and the process has since acquired majorimportance
to industrial manufacturing. ALD is the process of
fabricatinguniform conformal films through the cyclic deposition of
self-terminatingsurface half-reactions that allows for thickness
control at the levei of theatomic layer [66]. ALD is a derivative
of chemical vapor deposition (CVD),but one that differs from CVD in
severa I notable ways [67]. The comparisonis shown in Table
4.12.
ALD is a straightforward synthesis method that exploits specific
chemical react-ions with the intent of adding one molecular
monolayer at a time. The process is
-
212 lntroduction to Nanoscience and Nanotechnology
Chemical vapor deposton, especially in the case of casbon
nanotubes, is yet anothet low-eost, "low-tech" metho to form
nanomaterals. A carbon soutce gas (usuaUy methane,CO, acetylene,
propylene, or ethylene) is introduced into a chamber (the quartz
tube pic-tured) under reducng conditons. Upon contact with Co, Pe,
or Ni catalyst pattides, thegases decompose imo C and H atoms.
Nanotubes nudeate on the catalyst partide andgrows out from the
partide by eithet the tp-growth or base-growth mechanism.
1fpicalCVD conditions use 10% methane, 5% hydrogen, 85% argon
carrier gas at 700C, andatmospheric pressute.
Quartz tube
1000C Furnace, 1C Resolution
Roughing pump
characterized bya binary reaction that is split into two
half-reactions appliedsequentially. ALD is characterized by the
systematic use of self-terminating gas-solid reactions [68 J. A
self-terminating reaction depends on saturation of avail-able
surface sites and that precursors do not react with each other. The
ALDprocess offers a powerful arsenal of properties that are
specifically tailored fornanofabrication of thin films. First of
aIl, conformal coatings can be applied toparticulates or fiat and
curved surfaces ofbulk materiaIs. Secondly, atomic-scalecontrol of
thickness is possible by adding Iayers with stoichiometric
scalingbased on a chemisorption-saturation processo The process is
broken down intothe following general steps:
Surface activation ~ injection of A ~ purge ~ injection of B ~
purge ~injection of A ~ purge ~ injection of B ~ purge ~ ~
termination
Ultimatelya film composed of a structure ABABABA is formed. The
thicknessof the film can be estimated instantaneously by counting
the steps in theprocesso
The process is depicted in Figure 4.10. In essence, the ADL
sequential processalternates between chemisorption and saturation
steps. Purging of the processfollows each saturation step in the
cycle. The ALD film growth process is referredto as self-limiting
in that a stoichiometric process essentially terminates thereaction
upon saturation. Excess reactants and products are purged from
thechamber following each step.
-
Pabrication Methods 213
TABLE 4.12
Para meterPrecursor reactivity
Potential materiais
SelectivitySurfaces
Decomposition atreaction temperature
Time of processUniformity
Thickness
Conditions
Upscale potential
ALD-CVD Comparison
Chemical vapor depositionLess reactiveCan be autocatalyticMetal
oxides, semiconductors, and carboncompoundsLow selectivityLayers
conform according to surfacetopography of substrate.
Reactants can decompose.
VariableUniformity control by process parameters(partial
pressure of reactants, flow, pressure,temperature)-more difficult
to executeThickness contrai by process parameters-more difficult to
execute
Requires inert atmosphere and highertemperatures (>600C)P, T,
concentration, and gas flow distributionhave significant effect on
the processo
Good
The formation of alumina layers on a silicon surface will serve
as an exampleof the ALD processo The first step in the process is
the activation of a hydrogen-terminated silicon surface by exposure
to water vapor:
Atomic layer depositionHighly reactiveSelf-limiting at
saturationMetais, semiconductors, insulatorsWide range of
materiaisHighly selectiveLayers conform according to
surfacetopography of substrateSurfaces capable of activat