91 Top-down synthesis of nanostructured materials: Mechanical and thermal processing methods

Corresponding author: C.C. Koch, e-mail: [email protected]

Rev.Adv.Mater.Sci. 5 (2003) 91-99
2003 Advanced Study Center Co. Ltd.
TOP-DOWN SYNTHESIS OF NANOSTRUCTURED
MATERIALS: MECHANICAL AND THERMAL PROCESSING
METHODS
C. C. Koch
Materials Science and Engineering Department, North Carolina State University, Raleigh, NC 27695, USA
Received: June 10, 2003
Abstract. This paper reviews the top-down methods for synthesis of nanostructured or
submicron grain size materials which utilize severe plastic deformation along with thermal
processes. Of the variety of methods discussed, it appears that only ball milling, high pressure
torsion, and accumulative roll bonding can regularly produce average grain sizes below 100 nm
that is, nanocrystalline materials. Recent examples of combining conventional deformation
processing methods with annealing show the possibility of forming nanocrystalline/submicron
grain sizes that can optimize mechanical properties in practical ways.
1. INTRODUCTION
Nanostructured materials, which can be defined as
materials with crystallite sizes less than 100 nm in
dimension, are synthesized by either bottom-up
or top-down processes [1]. The bottom-up ap-
proach starts with atoms, ions or molecules as
building blocks and assembles nanoscale clusters
or bulk material from them.
The top-down methods for processing of
nanostructured materials involve starting with a bulk
solid and obtaining a nanostructure by structural
decomposition. One such approach involves the
lithography/etching of bulk material analogous to
the processes used in the semiconductor industry
wherein devices are formed out of an electronic
substrate by pattern formation (such as electron
beam lithography) and pattern transfer processes
(such as reactive ion etching) to make structures
at the nanoscale. This broad area will not be covered
further in this paper. The main subject of this paper
will be the preparation of nanostructured materials
from the bulk by the use of mechanical/thermal
methods. The major methods to be discussed use
severe plastic deformation to introduce defects, i.e.
dislocations, into the material which can then self-
assemble into nanoscale grains, with or without the
aid of external thermal sources. In the remainder of
the paper the various methods for processing
nanostructured materials in these ways will be de-
scribed and discussed. This will include examples
from the author's laboratory as well as from the
literature. The nature of the microstructural evolu-
tion which provides the sought-for nanoscale struc-
tures will also be discussed with respect to pos-
sible mechanisms.
2. PROCESSING METHODS
2.1. BaII miIIing and consoIidation of powders
The ball milling of powders was first developed as a
powder metallurgy method to produce dispersion
strengthened alloys with fine, uniform dispersoid dis-
tributions [2]. Subsequently, it has been used as a
powerful nonequilibrium processing method that can
synthesize a variety of metastable structured mate-
rials. The ball milling of powders mechanical attri-
tion can be divided into two categories: 1. the
milling of elemental or compound powders me-
92 C.C. Koch
chanical milling and 2. the milling of dissimilar pow-
ders mechanical alloying in which material trans-
fer occurs. This subject has been reviewed by a
number of authors, e.g. [3-5]. Nanostructured ma-
terials are one class of metastable materials that
can be made by ball milling. Besides being dis-
cussed in the examples of general reviews of ball
milling, this specific topic has also been reviewed
by itself, e.g. [6-8]. The details of the mechanical
attrition processes, equipment used, etc. have been
covered in a number of reviews, e.g. [4], and will not
be repeated here.
Mechanical attrition has been found to refine the
grain size to the nanoscale of all solid elements
studied. The minimum grain size achieved is, how-
ever, dependent on a number of process variables
as well as properties of the element or alloy. The
minimum grain size obtainable by milling, d
min
, has
been attributed to a balance between the defect/
dislocation structure introduced by the plastic de-
formation of milling and its recovery by thermal pro-
cesses [9]. It has been found that the minimum grain
size induced by milling scales inversely with the
melting temperature of a group of fcc structure met-
als studied [9]. These data are plotted in Fig. 1 along
with data for other metals and carbon (graphite) [7].
For these data, only the lower melting point metals
show a clear inverse dependence of minimum grain
size on melting temperature. The minimum grain
size for elements with higher melting temperatures
(>T
M
for Ni), exhibit essentially constant values with
melting temperature for given crystal structure
classes. For these elements it appears that d
min
is
Fig. 1. Minimum nanocrystalline grain size for ball
milled elements vs. their melting temperature. Data
from references 7, 9, 10, 11.
in the order: fcc < bcc < hcp. However, these data
must be considered with some skepticism since
they were obtained by several different research
groups using different milling equipment, and the
measurements of grain size were mainly by the
analysis of x-ray diffraction line broadening, which
can be questionable in terms of absolute values. In
spite of these potential problems, however, the val-
ues for d
min
are in remarkable agreement between
the data of Fecht et al. [10] who used a high energy
shaker mill (Spex 8000) and that of Olesak and
Shingu [11] who used a conventional horizontal low
energy ball mill. These results suggest that total
strain, rather than milling energy or ball-powder-ball
collision frequency, is responsible for determining
the minimum nanocrystalline grain size. This is dif-
ferent from ball milling induced amorphization or dis-
ordering where it appears the energy and frequency
of ball-powder-ball collisions determine the final
structures formed in driven systems. It is, how-
ever, consistent with observations of nanocrystallites
formed at high strain values using other non-cyclic
deformation methods, as will be discussed later in
this paper. These results suggest that mill energy
per se is not critical to the final microstructure, but
of course, the kinetics of the process are depen-
dent on the energy, and times for attaining the same
microstructure can be several orders of magnitude
longer in the low energy mills than in high energy
mills.
Effect of Milling Temperature. Milling tempera-
ture has been observed to affect the rate at which
the nanocrystalline structure develops. The milling
time at which a given grain size was attained in a
TiNi intermetallic compound was a function of milling
temperature, with smaller grains formed at lower
milling temperatures [12]. Shen and Koch [13] also
observed smaller nanocrystalline grain sizes in both
Cu and Ni milled at -85

C compared to samples
milled at room temperature. For example, for Cu,
d = 26 3 nm for room temperature milling and
d = 17 2 nm for milling at -85 C. Evidence for
smaller nanocrystalline grain sizes formed by milling
at low temperatures have now been observed in a
number of materials including the intermetallic
compound CoZr [14] and elemental Zn [15]. Since
milling at temperatures lower than ambient can bias
the defect accumulation induced by plastic defor-
mation with respect to thermal recovery, higher dis-
location densities, and therefore, as observed, finer
grain sizes can be obtained. The first uses of milling
at cryogenic temperatures, however, have been for
the purpose of introducing fine nanoscale nitrides or
93 Top-down synthesis of nanostructured materials: Mechanical and thermal processing methods
oxynitrides into Al [16]. This method has subse-
quently been used to similarly dispersion harden
other materials including NiAl [17] and Inconel 625
[18]. In this case, liquid nitrogen is introduced into
the milling vial along with the powders and milling
balls. The chemical reaction of the nitrogen with the
metal matrix powders produces the fine nitride par-
ticles which can help stabilize the nanoscale grain
size during subsequent thermal/mechanical pow-
der consolidation, as well as serve as dispersion
hardening agents.
A recent study of the milling of Zn at cryogenic
temperatures, albeit with the Zn milled under an
argon atmosphere, has revealed a modulated cy-
clic variation in hardness/strength with milling time
which has implications on the formation mecha-
nisms for nanocrystalline grain sizes by plastic de-
formation [19]. The measured density and hardness
values for cryomilled Zn are shown in Fig. 2 as a
function of milling time. The hardness values for Zn
milled at room temperature are also shown. A relative
density greater than 98% of theoretical density was
obtained for all of the Zn samples after milling and
powder compaction. The magnitude of the hardness
peaks decreases with milling time, thus exhibiting
a modulated oscillatory hardening manifest as
damped oscillations of the hardness. At longer
milling times the hardness reaches a steady-state
value about twice that of unmilled Zn. Transmission
electron microscopy showed that large variations in
Fig. 2. Hardness vs. milling time for cryomilled Zn
(solid circles) and Zn milled at room temperature
(solid triangles). The relative density for the
cryomilled samples is shown by the open squares.
Reference 19.
the dislocation density and grain-size distribution
occurred during cryomilling. The observations sug-
gest that dynamic recrystallization takes place in
larger grains when the dislocation density due to
strain-hardening reaches a critical level. A reaction-
rate model was developed which accounts for the
dynamic recrystallization effect and the observed
oscillations in hardness.
Proposed Mechanisms for Formation of Nano-
structures by Ball Milling. The first description of
the formation of nanocrystalline materials by me-
chanical attrition was given by Fecht et al. [10]. The
observed phenomenology of nanocrystallization by
mechanical attrition was summarized as occurring
in three stages, namely:
Stage 1. Deformation localization in shear bands
containing a high dislocation density.
Stage 2. Dislocation annihilation/recombination/
rearrangement to form cell/subgrain structures
with nanoscale dimensions further milling ex-
tends this structure throughout the sample.
Stage 3. The orientation of the grains becomes
random, that is, low angle grain boundaries dis-
appear as high angle grain boundaries replace
them, by presumably grain boundary rotation/
sliding.
Stage 2 might be considered to be a form of
self-assembly since the dense dislocation arrays
form into subgrain boundaries in order to lower the
energy of the system. This mechanism proposed
by Fecht and co-workers appears to be a logical
description, and may in most cases be the process
of nanocrystallization. However, from the results on
cryomilled Zn discussed above, in some special
cases high angle nanocrystalline grains can also
be formed by a discontinuous dynamic recrystalli-
zation process in contrast to the continuous
rearrangement of dislocation structures during de-
formation which leads to subgrains and then grains
with high angle boundaries.
Nanocrystalline grains are observed during the
mechanical alloying of dissimilar component pow-
ders. Klassen et al [20] followed the phase forma-
tion and microstructural development during me-
chanical alloying of Ti and Al powder blends of over-
all composition Ti
25
Al
75
. TEM revealed nano-
crystalline grains of partially ordered L1
2
phase with
a crystallite size of 10-30 nm in the alloy layers at
the interface between the pure Ti and Al lamellae at
very early stages of the milling process. The alloy
phase which develops between the pure powder
components consists of nanocrystalline grains pre-
sumably because of the multiple nucleation events
94 C.C. Koch
and the slow growth which occur at the relatively
low temperatures (100 200 C above ambient)
during milling. Trudeau et al. [21] prepared nanocrys-
talline FeTi by both low milling energy mechanical
alloying of elemental Fe and Ti powders, and by
mechanical milling of FeTi compound powders.
Higher mill energies resulted in amorphization. The
grain size of the mechanically milled FeTi steadily
decreased with milling time while that for the me-
chanically alloyed Fe/Ti first increased and then
decreased to values essentially identical to those
for the milled samples. This effect is illustrated in
Fig. 3.
It might be expected that brittle components
would simply fracture during milling and be reduced
in size to the limit of comminution observed in the
grinding of brittle mineral powders. However, it has
been found that ball milling of some nominally brittle
materials can lead to alloying of the brittle
components, e.g. Si and Ge [22] and the introduc-
tion of significant plastic deformation and high dis-
location densities in brittle compounds, e.g. Nb
3
Sn
[23]. Si and Ge are completely brittle at room tem-
perature and yet complete solid solutions of Si-Ge
alloys were obtained across the binary phase dia-
gram. Thus, alloying on the atomic scale was ob-
served by mechanical alloying of brittle components.
Nb
3
Sn is an extremely brittle intermetallic compound
which fractures elastically until tested at tempera-
tures above about 1400 C. However, ball milling
Nb
3
Sn can produce large amounts of plastic defor-
Fig. 3. Crystallite size for FeTi intermetallic phase
for low energy mechanical alloying of Fe and Ti pow-
ders (solid circles), and mechanical milling of the
FeTi compound powders (solid triangles) [21].
mation as observed by TEM of the milled powder.
The dislocations so produced then presumably in-
duce a nanocrystalline grain structure similar to that
formed in milled ductile metals. It is not yet clear
how ball milling can produce large plastic deforma-
tion in materials that are very brittle under uniaxial
stress conditions. It is suggested that the high hy-
drostatic stress component that may exist in the
powders during milling can favor plastic deforma-
tion over fracture and allow a large dislocation den-
sity to be generated [23]. Mechanical attrition has
also been found to induce nanocrystalline micro-
structures in brittle ceramics such as ZrO
2
and ce-
ramic powder mixtures such as Fe
2
O
3
/Cr
2
O
3
and
ZrO
2
/Y
2
O
3
[8].
BaII MiIIing of PoIymeric MateriaIs. The appli-
cation of mechanical attrition to polymeric materi-
als was initiated by Shaw and co-workers [24]. In
order to fracture the polymer particulates, and on
the microscopic level the polymer chains, the milling
was conducted at temperatures below the glass tran-
sition temperature of the given polymer. Shaw's
group has studied a number of homopolymers such
as polyamide, polyethylene, acrylonitrile-butadiene-
styrene, polypropylene, and polystyrene. Refinement
of the microstructure typically occurred and milling-
induced structural and property changes were noted
that were very material specific. Subsequently, oth-
ers have studied milling-induced changes in the
structure of several semi-crystalline and amorphous
homopolymers [25]. Milling of poly(methyl methacry-
late) (PMMA) resulted in monotonic decreases in
molecular weight and glass transition temperature,
reflecting the milling-induced scission of the poly-
mer chains. Polyisoprene (PI) exhibited much dif-
ferent behavior in that the decrease in glass transi-
tion temperature, T
g
, given by T
g
=T
g,0
T
g
(t
m
),
where T
g,0
is the glass transition temperature of the
unmilled polymer and t
m
is the milling time, first in-
creased and then decreased as illustrated in Fig. 4.
In this case cryomilled PI does not exhibit a
monotonic increase in T
g
but instead shows a
sharp maximum at relatively short milling times (2h)
followed by a drop to almost zero before again in-
creasing slightly for longer milling times. This
unusual, but reproducible, behavior strongly sug-
gests that the PI chains undergo chemical
crosslinking during cryomilling. In this case we imag-
ine a dynamic competition between chains break-
ing (causing a decrease in molecular weight) and
crosslinking (promoting an increase in molecular
weight) under the nonequilibrium conditions of
milling. Sol-gel analysis and FTIR spectroscopy gave
95 Top-down synthesis of nanostructured materials: Mechanical and thermal processing methods
further evidence for the milling-induced crosslinking
in PI. Mechanical attrition was also performed at
cryogenic temperatures to incorporate PI into
PMMA [26]. TEM clearly showed that the solid-state
blending by mechanical attrition of polymeric mate-
rials can yield nanoscale dispersions of immiscible
polymers. The strong evidence for free radical for-
mation and crosslinking induced by milling suggests
many possibilities for the design of novel new poly-
meric materials with nanoscale microstructures.
Problems: Contamination and Powder Conso-
lidation. A serious problem with the milling of fine
powders is the potential for significant contamination
from the milling media (balls and vial) or atmosphere.
If steel balls and containers are used, iron
contamination can be a problem. It is most serious
for the highly energetic mills, for example, the Spex
shaker mill, and depends upon the mechanical be-
havior of the powder being milled as well as its
chemical affinity for the milling media. For example,
milling Ni to attain the minimum grain size in a Spex
mill resulted in Fe contamination of 13 at.%, while
the Fe contamination in nanocrystalline Cu simi-
larly milled was only <1 at.% [27]. Lower energy
mills result in much less, often negligible, Fe
contamination. Other milling media, such as tung-
sten carbide or ceramics can be used but
contamination from such media is also possible.
Interstitial element (oxygen, nitrogen) contamination
can be controlled by milling and subsequent pow-
der handling in a pure inert gas atmosphere, with
Fig. 4. Dependence of T
g
on milling time, t
m
, for
cryomilled polyisoprene. Reference 24.
care taken that the milling vial is leak-free during
processing.
Powder consolidation to theoretical density of
nanocrystalline materials prepared by mechanical
attrition without significant coarsening of the micro-
structure is necessary for many property
measurements, for example, mechanical behavior,
and for applications that require bulk materials. There
is not room in this paper to adequately review the
increasing effort in this important area. However,
research on particulate consolidation in this regard
by both conventional (e.g. hot pressing, hot isos-
tatic pressing, extrusion, etc.) and innovative
methods (e.g. microwave sintering, plasma activated
sintering, etc.) has been documented and is reviewed
by Groza [28]. For the special case of very ductile,
relatively low melting temperature metals such as
Zn, it has been found that in the absence of a pro-
cess control agent during milling, the cold-welding
can dominate the process such that the powders
can be consolidated during milling into spherical
shaped balls that can be up to 6 10 mm in diam-
eter [29]. Thus, in situ consolidation can be attained.
Disks can be formed from the spherical samples by
compression. Such disks are suitable for mechani-
cal tests such as miniaturized disk bend tests or
small tensile sample tests.
2.2. BuIk processing methods for severe
pIastic deformation
The possibility of producing very fine grain size struc-
tures by severe plastic deformation was suggested
by research using conventional deformation
methods taken to very high degrees of strain. It has
been known for many decades, going back to the
1950's, that the structure of deformed metals can
change with increasing plastic deformation such
that random dislocation arrays can lower the energy
of the system by self assembly into cells or
subgrains such that there is a high dislocation
density in the cell walls and a lower dislocation den-
sity within the cells. The cells are typically the result
of plastic deformation, the cell boundaries are some-
what diffuse, and strengthening due to cell struc-
tures gives an m = 1 dependence in the equation
=
0
+kGbd
-m
where is the shear flow stress, G
the shear modulus, b the Burgers vector, d the cell
size, and k and m are constants [30]. Subgrains,
like cells, show small misorientations with their
neighbors, but have sharper boundaries, and are
formed by plastic deformation and thermal recovery
processes. The strengthening gives an m value of
0.5, the same as for high angle grain boundaries in
96 C.C. Koch
the Hall-Petch equation. In most cases, the early
studies of microstructures produced by severe plas-
tic deformation gave cell or subgrain sizes in the
micron down to submicron size scale, but not into
the nanoscale. However, there have been some
reports of severe plastic deformation by multiple
pass rolling or drawing that produced nanoscale
microstructures [31]. For example, Pavlov has given
evidence for the decomposition of the microstruc-
ture of Ni, Pt, and several Pt-base alloys first into
fragments of about 15 nm diameter. The fragment
size decreased further with continued deformation
and the misorientation between crystallites in-
creased. Eventually, partial amorphization was
reported. These results have not been reproduced
to the author's knowledge. Rack and Cohen reported
[32] the cell structure developed in a series of Fe-Ti
alloys by wire drawing to large values of true strain
up to about 7. The size of the cells decreased with
increasing deformation and reached values of about
50 nm at the highest strains. However, these were
all cells with very low angle misorientations. In recent
years special methods of mechanical deformation
have been developed for producing submicron and
even nanoscale grains with high angle grain bound-
aries. These methods, the microstructures devel-
oped, and the properties of the materials with the
refined grains so produced have been reviewed by
Valiev et al. [33].
The major methods reviewed [33] and that have
received the most attention from researchers in the
field are severe plastic torsion straining under high
pressure (SPTS) and equal channel angular press-
ing (ECAP). In the case of SPTS a disk shaped
sample is compressed to pressures of a few GPa
and then one of the dies is moved with respect to
the other. With enough rotation significant shear
strains can be achieved. Even though it would be
predicted that the strain distribution might be radial
from the disk center, electron microscopy studies
have shown approximately uniform microstructures
across the disk samples. This is consistent with
more detailed calculations that take into account
both the compressive and torsional stress states.
While in most studies, submicron grain sizes have
been produced by SPTS, in some cases
nanostructured materials have been prepared. This
method has also been successfully used for the
consolidation of powders. The ECAP method which
allows for the deformation of bulk samples by pure
shear was first developed by Segal [34]. In this
method a billet is pressed through a die with two
channels at angles of intersection typically 90 to
120. The billet is subjected to severe deformations
without changing its dimensions. Multiple passes
through the die provide accumulative strain. The grain
sizes developed by this method are typically in the
submicron, 200 to 300 nm, range. A large body of
experimental research and modeling studies has
been developed for this technique as reviewed by
Valiev et al. [33] and in subsequent journal articles
and conference proceedings. There are examples
of submicron size grain structures induced by the
severe strain of ECAP in several metals that pro-
vide an excellent combination of both increased
strength along with good ductility [35].
Another bulk deformation method that has been
applied to billet deformation is multiple forging. This
was developed by Salischev and co-workers [36]. It
involves repeated free forging of a billet with changes
of the axis of the applied stress and has been done
at elevated temperatures in a variety of Ti, Ni, and
Mg- base alloys. The mechanism of grain refinement
has been attributed to dynamic recrystallization at
the elevated temperatures of processing [33]. For
example, Belyakov et al. [37] have deformed 304
stainless steel by multiple compressions with the
deformation axis changed after every pass to total
strains of over 6 at 873K. The formation of the sub-
micron grains observed was attributed to a
continuous dynamic recrystallization phenomenon.
Accumulative roll-bonding is a process which can
be applied to the grain refinement of sheet samples
and in principle scaled up for commercial applica-
tion to large production quantities of sheet material.
The process was first proposed by Saito et al. [38]
although similar methods had been previously used
to prepare nanostructured or amorphous materials
by cyclical rolling and stacking of sheets of dis-
similar metals [39, 40]. Accumulative roll bonding
involves the rolling of a sheet, cutting, cleaning the
surfaces, stacking, and rolling the stacked pair to
sufficient deformation (typically 50%) to attain good
bonding. The process is then repeated and the strain
so accumulated can reach large values only depen-
dent on the number of cycles used. While the earlier
studies using dissimilar metal sheets, foils, resulted
in nanocrystalline or even amorphous structures,
the more recent work applied to single phase sheets
usually results in submicron rather than nanoscale
grain sizes. However, recent work on dissimilar foils
also exhibits nanostructured grain formation [e.g.
41]. The number of rolling/stacking cycles in these
studies of dissimilar metal composites were large,
typically 60-75 times, while the reports of accumu-
lative roll bonding on single composition sheets were
97 Top-down synthesis of nanostructured materials: Mechanical and thermal processing methods
only 7-8 times. Since the total strain is proportional
to the number of rolling/stacking cycles, this may
explain why nanostructured grains have been ob-
tained in the dissimilar stacked foil experiments and
only submicron grains in the single composition
studies. Thus, under the appropriate conditions,
accumulative roll bonding can be a method to ob-
tain nanostructured microstructures.
Other Bulk Processing Methods that Produce
Submicron Grain Sizes. There have been several
attempts to design and use alternate techniques
for producing refined grain structures by severe plas-
tic deformation. Among these are repetitive
corrugation and straightening [42], friction stir weld-
ing [43], multipass coin-forging [44], constrained
groove pressing [45], and twist extrusion [46]. Re-
petitive corrugation and straightening involves
bending and straightening of sheet samples
repetitively to build up significant plastic strain. The
advantages cited for this method, which can be
adapted to rolling mill technology, are the creation
of bulk sheet material with fine grains and free from
contamination and porosity [42]. A range of fine
grains were observed in Cu subjected to repetitive
corrugation and straightening, but the average grain
size was submicron, not nanocrystalline. Friction
stir welding carried out with the sample cooled to
liquid nitrogen temperature, has resulted in submi-
cron grain sizes in aluminum alloys [43]. The mecha-
nism has been attributed to dynamic recrystalli-
zation. In multipass coin-forge processing a metal
sheet surface is coined between two sine wave
shaped dies, with successive rotation of the
workpiece, followed by flat forging or rolling. The
process is repeated until the deformed surface zones
meet. Submicron grain sizes have been obtained.
Constrained groove pressing [45] is very similar in
principle and results to multipass coin-forging. The
concept of twist extrusion extruding a prismatic
billet through a die with a twist channel has been
proposed [46] but no experimental results have been
presented.
2.3. Surface deformation methods
It has been known for some time that the micro-
structures of surfaces subjected to wear processes
can be similar to those observed by mechanical
attrition; that is nanostructured or submicron grain
sizes [8]. Recently K. Lu and co-workers have de-
veloped a method to produce controlled severe sur-
face deformation on metallic surfaces [47]. This
method is ultrasonic shot peening. Steel balls (3
mm diameter) are vibrated at ultrasonic frequencies
and impact the sample surface when resonated by
a large number of balls over a short period of time.
The surface layers of the sample are severely de-
formed and nanoscale microstructures are devel-
oped in the near surface regions. The refined grains
become coarser as the distance from the surface
increases. By using TEM the progress of the defor-
mation-induced grain refinement could be followed.
This surface mechanical attrition (SMA) technique
promises to provide a surface treatment to enhance
mechanical and possibly corrosion properties of
metals so treated.
Static recrystallization of severely deformed
metals. In order for practical production of struc-
tural nanostructured materials in industrial opera-
tions it would be desired to use conventional defor-
mation processing methods such as rolling, extru-
sion, or wire drawing followed by appropriate
annealing treatments to give microstructures for
optimized mechanical behavior. While it has been
demonstrated in the past [e.g. 32] that severe con-
ventional deformation processes can produce refined
microstructures, these microstructures on the
nanoscale have been cell or subgrain boundaries.
High angle grain boundaries are needed as a sub-
stantial percentage of the boundaries to take ad-
vantage of the nanocrystalline properties. There have
been some recent results that indicate such hoped
for processes may be possible.
Wang et al. [48] have reported on the prepara-
tion and tensile testing of nanostructured/submicron
grain sized Cu. They rolled the Cu to 93% at liquid
nitrogen temperatures and then annealed it at low
temperatures up to 200 C. The original heavily cold
worked Cu had a high dislocation density along with
some resolvable grains less than 200 nm in size.
Annealing resulted in development of well defined
grains with high angle boundaries. The annealing
treatment (3 minutes at 200 C) that optimized
strength and ductility produced a mixture of
nanoscale/ultra fine grains (80 nm to 200 nm) along
with about 25% volume fraction of coarser grains (1
to 3 m). The coarser grains were the result of sec-
ondary recrystallization. The nanoscale grains ap-
parently supplied the strengthening while the larger
grains allowed for significant strain hardening which
prevents localized deformation and premature frac-
ture. Rolling the Cu at room temperature did not
produce the above effects, suggesting that biasing
the accumulation of defects (dislocations) versus
recovery by thermal processes by rolling at liquid
nitrogen was needed to obtain the required deformed
structure as a precursor to the development of the
fine grained structures on subsequent annealing.
98 C.C. Koch
Another example of using conventional
thermomechanical processing to obtain a mixture
of nanostructured and submicron grain size micro-
structures was given by Ueji et al. [49]. The authors
used plain low carbon steels that were austenitized
and quenched to produce a martensitic structure.
The as-quenched sheets were rolled to a modest
strain of 0.8 and then annealed at temperatures of
673K to 973K. The steel annealed about 823K ex-
hibited a microstructure consisting of fine grained
ferrite with grain sizes of 50 to 300 nm along with
uniformly distributed carbide particles. This mate-
rial had both high strength (710 MPa yield strength
and 870 MPa tensile strength) and good ductility
(9% uniform elongation, 20% total elongation). The
as-quenched lath martensite in these low carbon
steels had a high dislocation density which on sub-
sequent modest plastic deformation by rolling ap-
parently allowed a sufficiently high dislocation den-
sity to produce the fine grained microstructure ob-
served. These examples suggest that it should be
possible to obtain nanostructured/submicron grain
sizes in materials using conventional deformation
processing methods combined with annealing in
carefully chosen procedures for given materials. An
initial high dislocation density appears to be a pre-
requisite for such behavior.
3. SUMMARY AND CONCLUSIONS
Of the variety of methods discussed in this paper to
prepare nanoscale microstructures by severe plas-
tic deformation, only ball milling, high pressure tor-
sion, and accumulative roll bonding appear to be
able to regularly produce average grain sizes below
100 nm that is nanocrystalline materials. Ball
milling results in powders which then must be
compacted without losing the refined microstruc-
ture but with theoretical density and complete par-
ticulate bonding. This remains a challenge to this
otherwise very versatile method. High pressure tor-
sion is an excellent way to prepare nanocrystalline
materials but is limited to small sample sizes. Ac-
cumulative roll bonding does not have the limitation
of size, but to reach the nanoscale microstructure
requires many repetitions which is very labor inten-
sive and may not be a practical industrial process.
However, a few recent examples of combining
conventional deformation processing with annealing
are encouraging. This suggests that with better un-
derstanding of the deformation and recrystallization
processes in given materials, thermomechanical
methods may be developed that can produce mix-
tures of nanoscale and submicron grain sizes that
can optimize mechanical properties.
ACKNOWLEDGEMENTS
The author's research in this area is supported by
the National Science Foundation, grant number
DMR-0201474 and the US Department of Energy,
grant number DE-FG02-02ER46003.
REFERENCES
[1] E. L.Hu and D. T. Shaw, In: Nanostructure
Science and Technology, ed. by R. W. Siegel,
E. L. Hu and M. C. Roco (Kluwer Academic
Publ: Dordrecht, 1999), p. 15.
[2] J. S. Benjamin // Sci. Amer.234 (1976) 40
[3] C. Suryanarayana // Prog. Mater. Sci. 46
(2001) 1.
[4] C. C. Koch, In: Materials Science and Tech-
nology A Comprehensive Treatment, ed. by
R. W. Cahn, P. Haasen and E. J. Kramer
(VCH: Weinheim,1991) vol. 15, p.193.
[5] B. S. Murty and S. Ranganathan // Inter.
Mater. Rev. 43 (1998) 101.
[6] C. C. Koch // NanoStructured Mater. 2 (1993)
109.
[7] C. C. Koch // NanoStructured Mater. 9 (1997)
13.
[8] H. J. Fecht, In: Nanostructured Materials:
Processing, Properties, and Applications, ed.
by C. C. Koch (William Andrew Publ.:Norwich,
2002) p. 73.
[9] J. Eckert, J. C. Holzer, C. E. Krill III and W. L.
Johnson // J. Mater. Res. 7 (1992) 1751.
[10] H. J. Fecht, E. Hellstern, Z. Fu and W. L.
Johnson // Met. Trans. A 21A (1990) 2333.
[11] D. Olesak and P. H. Shingu // J. Appl. Phys.
79 (1996) 2975.
[12] K. Yamada and C. C. Koch // J. Mater. Res.
8 (1993) 1317.
[13] T. D. Shen and C. C. Koch // Mater. Sci.
Forum 179-181 (1995) 17.
[14] D. K. Pathak, Ph. D. Dissertation , North
Carolina State University, 1995.
[15] X. Zhang, Ph. D. Dissertation, North Carolina
State University, 2001.
[16] M. J. Luton, C. S. Jananth, M. M. Disko,
S. Matras and J. Vallone // MRS Symp. Proc.
132 (1989) 79.
[17] J. D. Whittenberger, E. Arzt and M. J. Luton
// J. Mater. Res. 5 (1990) 271.
[18] J. He and E. J. Lavernia // J. Mater. Res. 16
(2001) 2724.
99 Top-down synthesis of nanostructured materials: Mechanical and thermal processing methods
[19] X. Zhang, H. Wang, R. O. Scattergood,
J. Narayan and C. C. Koch // Acta Mater. 50
(2002) 3995.
[20] T. Klassen, M. Oehring and R. Bormann //
J. Mater. Res. 9 (1994) 47.
[21] M. L. Trudeau, R. Schultz, L. Zaluski,
S. Hosatte, D. H. Ryan, C. B. Doner,
P. Tessier, J.O. Strom-Olsen and A. Van
Neste // Mater. Sci. Forum. 88-90 (1992)
537.
[22] R. M. Davis and C. C. Koch // Scripta Metall.
21 (1987) 305.
[23] Y. S. Cho and C. C. Koch // Mater. Sci. and
Engr. A A141 (1991) 139.
[24] W. J. D. Shaw // Mater. Sci. Forum 269-272
(1998) 19.
[25] C. C. Koch, A. P. Smith, C. Bai, R. J.
Spontak and C. M. Balik // Mater. Sci. Forum
343-346 (2000) 49.
[26] A. P. Smith, H. Ade, C. M. Balik, C. C. Koch,
S. D. Smith and R. J. Spontak // Macromol-
ecules 33 (2000) 2595.
[27] T. D. Shen and C. C. Koch // Acta Mater. 44
(1996) 753.
[28] J. R. Groza, In: Nanostructured Materials:
Processing, Properties, and Applications, ed.
by C. C. Koch (William Andrew Publ.
Norwich, 2002) p. 115.
[29] X. Zhang, H. Wang, M. Kassem, J. Narayan
and C. C. Koch // Scripta Mater. 46 (2002)
661.
[30] A. W. Thompson // Metall. Trans. A 8A
(1977) 833.
[31] V. A. Pavlov // Phys. Met. Metall. 59 (1985) 1.
[32] H. J. Rack and M. Cohen // Mater. Sci. and
Engr. 6 (1970) 320.
[33] R. Z. Valiev, R. K. Islamgaliev and I. V.
Alexandrov // Prog. Mater. Sci. 45 (2000)
103.
[34] V. M. Segal, V. I. Reznikov, A. E.
Drobyshevkij and V. I. Kopylov // Metally 1
(1981) 115.
[35] R. Z. Valiev, I. V. Alexandov, Y.T. Zhu and
T. C. Lowe // J. Mater. Res. 17 (2002) 5.
[36] O. R. Valiahmetov, R. M. Galeyev and G. A.
Salishchev // Russian Metally 10 (1990) 204.
[37] A. Belyakov, T. Sakai, H. Miura and
R. Kaibyshev // Phil. Mag. Lett. 80 (2000)
711.
[38] Y. Saito, H. Utsunomiya, N. Tsuji and
T. Sakai // Acta Mater. 47 (1999) 579.
[39] M. Atzmon, J. D. Verhoeven, E. D. Gibson
and W. L. Johnson // Appl. Phys. Lett. 45
(1984) 1052.
[40] P. H. Shingu, K. N. Ishihara, K. Uenishi,
J. Kuyama, B. Huang and S. Nasu, In: Solid
State Powder Processing, ed. by A. H.
Clauer and J. J. deBarbadillo (TMS:
Warrendale, PA) p. 21.
[41] L. Battezzati, P. Pappalepore, F. Durbiano
and I. Gallino // Acta Mater. 47 (1999) 1901.
[42] J. Y. Huang, Y. T. Zhu, H. Jiang and T. C.
Lowe //Acta Mater. 49 (2001) 1497.
[43] S. Benavides, Y. Li and L. E. Murr, In:
Ultrafine Grained Materials, ed. by R. S.
Misra, S. L. Semiatin, C. Suryanarayana,
N. N. Thadhani and T. C. Lowe (TMS,
Warrendale, PA,2000) p.155.
[44] A. K. Ghosh and W. Huang, In : Ultrafine
Grained Materials, ed. by R. S. Misra, S. L.
Semiatin, C. Suryanarayana, N. N. Thadhani
and T. C. Lowe (TMS, Warrendale, PA,2000)
p. 173.
[45] J-J Park and D. H. Shin, In: Ultrafine Grained
Materials II, ed. by Y. T. Zhu, T. G. Langdon,
R. S. Misra, S. L. Semiatin, M. J. Saran and
T. C. Lowe (TMS, Warrendale, PA, 2002)
p. 253.
[46] Y. Beygelzimer, D. Orlov and V. Varyukhin,
In: Ultrafine Grained Materials II, ed. By Y.T.
Zhu, T. G. Langdon, R. S. Misra, S. L.
Semiatin, M. J. Saran and T. C. Lowe (TMS
Warrendale, PA, 2002) p. 297.
[47] N. R. Tao, M. L. Sui, J. Lu and K. Lu //
NanoStructured Mater. 11 (1999) 433.
[48] Y. Wang, M. Chen, F. Zhou and E. Ma //
Nature 419 (2002) 912.
[49] R. Ueji, N. Tsuji, Y. Minamino, Y. Koizumi
and Y. Saito, In: Ultrafine Grained Materials II
ed. by Y. T. Zhu, T. G. Langdon, R. S. Misra,
S. L. Semiatin, M. J. Saran and T. C. Lowe
(TMS, Warrendale, PA, 2002) p. 399.