08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.

FABRICATION, STRUCTURE, AND TRANSPORT

PROPERTIES OF NANOWIRES
Yu-Ming Lin,1 Mildred S. Dresselhaus,1,2 and Jackie Y. Ying3*
1
Department of Electrical Engineering and Computer Science, 2Department of
Physics, and 3Department of Chemical Engineering, Massachusetts Institute of
Technology, Cambridge Massachusetts 02139

I. Introduction 168
II. Fabrication and Structural Characteristics of Nanowires 168
A. Template-Assisted Synthesis 169
B. Laser-Assisted Synthesis 181
C. Other Synthesis Methods 184
III. Theoretical Modeling of Nanowire Band Structures 185
A. Band Structures of One-Dimensional Systems 185
B. The Semimetal–Semiconductor Transition in
Semimetallic Nanowires 188
IV. Transport Properties 191
A. Semiclassical Model 192
B. Temperature-Dependent Resistivity of Nanowires 193
V. Summary 198
References 199

Nanowire systems have attracted a great deal of attention recently

due to their technological potential. They are of fundamental interest
because they exhibit unique quantum confinement effects. In this ar-
ticle, advances in the fabrication of nanowires via template-assisted
and laser-assisted approaches are reviewed. The structure and charac-
teristics of different nanowire systems are discussed. To understand
and predict the unusual properties of nanowires, we have developed
a generalized theoretical model for the band structure of these one-
dimensional systems. A unique semimetal–semiconductor transition
that occurs in bismuth nanowires is described. Transport measure-
ments on bismuth and antimony nanowires illustrate that these novel
materials are very different from their bulk counterparts. A transport

* To whom correspondence should be addressed.

167
Copyright C 2001 by Academic Press

All rights of reproduction in any form reserved.

ADVANCES IN CHEMICAL ENGINEERING, VOL. 27 0065-2377/01 $35.00
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

168 YU-MING LIN ET AL.

model, based on the band-structure calculations, is presented to ex-

plain the experimental results and to gain insight into the transport
phenomena of nanowire systems.  2001 Academic Press.
C

I. Introduction

Nanostructured materials have received significant attention in recent

years because of their fundamental importance and potential applications in
areas ranging from chemistry, physics, biology, and materials science. In the
fields of electronics and optics, the drive to miniaturize devices and increase
storage density has fueled research in nanotechnology. Several techniques
have been developed to fabricate nanostructures, such as epitaxial growth,
electron beam lithography, chemical vapor deposition, and self-assembly ap-
proaches. Nanostructures represent a new class of materials with properties
different from molecular species and bulk solid-state structures. They exhibit
quantum confinement effects, giving rise to unique behavior that can be ex-
ploited in novel optical, magnetic, electronic, and thermoelectric devices.
Multiple quantum well structures are probably the most studied nanostruc-
tured systems; the carriers are confined to two dimensions in these systems.
In comparison, one-dimensional quantum wires and zero-dimensional quan-
tum dots are expected to show even stronger quantum confinement effects.
Quantum wires are perhaps the most amenable for the design of novel elec-
tronic devices because they exhibit more pronounced quantum effects than
the two-dimensional structures; unlike most zero-dimensional systems, they
maintain transport continuity along the wire axis. This chapter describes the
fabrication, structure, and transport properties of nanowire systems of in-
terest to electronic applications. It presents a generalized theoretical model
for the band structure of quantum wires, which is responsible for the novel
behavior of these systems. In addition, a semiclassical transport model is
also developed so that the unusual transport properties that are experimen-
tally observed in nanowire systems can be explained and compared to the
predicted behavior.

II. Fabrication and Structural Characteristics of Nanowires

The preparation of one-dimensional quantum systems represents one

of the greatest challenges in materials fabrication. The synthesis of highly
crystalline and continuous nanowires is essential for studying the interesting
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 169

quantum phenomena in these low-dimensional systems. Over the past de-

cade, significant progress has been made in deriving nanowires via tech-
niques, such as high-pressure injection (Huber et al., 1994, 1999; Zhang
et al., 1998a, 1999), vapor deposition (Heremans et al., 2000; Cheng et al.,
1999), electrochemical deposition (Foss et al., 1992; Hornyak et al., 1997;
Fasol, 1998; Piraux et al., 1994, 1999; Martin, 1994; Sun et al., 1999a; Whitney
et al., 1993; Blondel et al., 1994; Liu et al., 1998a; Routkevitch et al., 1996a,b;
Yi and Schwarzacher, 1999; Peng et al., 2000; Zeng et al., 2000), laser abla-
tion (Morales and Lieber, 1998; Yu et al., 1998; Zhang et al., 1998b, 2000a;
Duan et al., 2000), thermal evaporation (Wang et al., 1998a; Tang et al.,
1999), molecular beam epitaxy (Nötzel et al., 1992; Omi and Ogino, 1997),
and electron beam lithography (Chou et al., 1996). These different meth-
ods led to the generation of a broad range of one-dimensional nanostruc-
tured materials (Martin, 1994). Although some approaches are more suit-
able than others for specific applications, each approach may present certain
synthetic limitations with regard to nanowire diameter, crystallinity, fabri-
cation costs, and scalability. In this section, the nanowire-fabrication tech-
niques will be reviewed, and the structure of the nanowires produced will be
discussed.

A. TEMPLATE-ASSISTED SYNTHESIS

The template-assisted synthesis of nanowires is a conceptually elegant

way of fabricating nanostructures in an organized assembly (Ozin, 1992;
Tonucci et al., 1992; Ying, 1999). The template is typically a nonconducting
host matrix containing nanometer-sized elongated pores or voids, which are
filled by the material of choice that adopts the pore morphology. If the pore
diameters are sufficiently small, the pore-filling nanowires would exhibit
quantum confinement effects. In this synthesis approach, the important fac-
tors for consideration include the template characteristics, such as chemical
stability, mechanical properties, pore diameter, uniformity, and density. Tem-
plates that have been used for nanowire synthesis include anodic alumina,
nanochannel glass, ion track-etched polymers, and mica films.
Anodic alumina templates are produced by the anodic oxidation of alu-
minum films in acidic electrolyte solutions (Diggle et al., 1969; O’Sullivan
and Wood, 1970; Li et al., 1998a). The resulting aluminum oxide film pos-
sesses a regular hexagonal array of parallel and nearly cylindrical channels
(Keller et al., 1953; Masuda et al., 1997; Li et al., 1998a), as depicted in Fig. 1.
Depending on the anodization conditions, such as the voltage applied and
the nature and concentration of electrolyte used, the pores of anodic alumina
can be systematically varied from less than 10 nm to 200 nm in diameter,
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

170 YU-MING LIN ET AL.

FIG. 1. Schematic of a porous anodic alumina template.

with a packing density on the order of 109 to 1011/cm2 (Diggle et al., 1969;
O’Sullivan and Wood, 1970; AlMawlawi et al., 1991; Zhang et al., 1999).
The film thickness (or pore length) (< 1 µm to > 50 µm) can be controlled
by the duration of the anodization process. With their synthetic flexibility,
ordered pore structure, high pore density, and high mechanical strength, an-
odic alumina systems have gained great popularity as a template material for
nanowire fabrication. In addition, with recent improvements on the degree
of pore ordering, anodic alumina has stimulated much interest for other de-
vice applications (Govyadinov and Zakhvitcevich, 1998; Davydov et al., 1999;
Li et al., 1999; Masuda et al., 1999; Li et al., 2000; Kouklin et al., 2000) because
it provides a highly ordered two-dimensional nanopattern, readily and inex-
pensively, on a large scale compared to conventional lithographic approaches.
In the processing of anodic alumina films (Keller et al., 1953; Masuda
et al., 1997; Li et al., 1998a; Zhang et al., 1999), a thin aluminum sheet is first
mechanically and electrochemically polished to produce a smooth surface.
It is then anodized in an acidic solution at a constant voltage and tempera-
ture. The anodization voltage V determines the interpore distance D by the
empirical relation D (nm) = −1.7 + 2.81·V (volts) (Li et al., 1998a). Differ-
ent electrolytes are usually used for different anodization voltage ranges:
20 wt% sulfuric acid (H2SO4) for less than 20 V, 4 wt% oxalic acid (H2C2O4)
for 30–65 V, and 3.5 wt% phosphoric acid (H3PO4) for 70 V or more. The
as-prepared anodic alumina film has open pores on the top surface of the
substrate and is capped by a barrier layer on the other side (see Fig. 1).
Therefore, the sample has to be etched by an acidic solution to remove the
barrier layer for suitable applications (Li et al., 1998a; Jessensky et al., 1998).
Figures 2(a) and (b) show scanning electron microscopy (SEM) images of
the top surfaces of porous alumina templates anodized in 4 wt% oxalic acid
and 20 wt% sulfuric acid, respectively. The pore ordering and uniformity of
these materials have been optimized by a two-step anodization technique
(Li et al., 1998b; Lin et al., 2000).
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 171

FIG. 2. SEM images of the top surfaces of porous anodic alumina templates anodized in (a)
4 wt% H2C2O4 and (b) 20 wt% H2SO4. The average pore diameters in (a) and (b) are 44 nm
and 18 nm, respectively.

The self-organized pore structure in anodic alumina is derived from two

coupled processes, pore formation and pore ordering. Pore formation is
generally believed to be a result of several mechanisms, including oxide
formation and dissolution. During the anodization process, anions (O2– or
OH–) migrate through the oxide layer and form Al2O3 at the oxide–metal
interface, whereas some of the Al3+ ions produced at the oxide–metal in-
terface move through the oxide layer and become ejected into the elec-
trolyte. The nonuniform electric field and current density present in the
sample from surface topological variations are essential to the pore growth
mechanism. The surface variations may come from either the initial sample
polishing or the self-induced variation during steady-state pore growth.
The field-enhanced dissolution or the increased local temperature promotes
the dissolution rate in some areas at the oxide-electrolyte interface to
yield the observed pore morphology. The self-ordering of the pores may be
attributed to the volume expansion during the oxide formation, which
produces a repulsive force between the pores. In response to the mutual
repulsive forces between neighboring pores, self-arrangements of the cylin-
drical pores will occur to maximize their packing density in the regular hexag-
onal order at steady state.
Porous templates can also be fabricated by chemically etching particle
tracks originated from ion bombardment (Ferain and Legras, 1993; Sun et al.,
1999a). The pores produced by this track-etching method are randomly dis-
tributed with a packing density of 107 to 109/cm2, which is substantially lower
than that in anodic alumina. By controlling the duration of chemical etching,
the pore diameter can be varied from several hundred nanometers to about
5 nm. Track-etched polycarbonate membranes are commercially available
in a variety of pore sizes and have been widely used in the template-assisted
synthesis of nanowires (Martin, 1994; Blondel et al., 1994; Liu et al., 1998a).
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

172 YU-MING LIN ET AL.

FIG. 3. (a) SEM image of the particle track-etched polycarbonate membrane, with a pore
diameter of 1 µm (Martin, 1994). (b) SEM image of 2-µm pores in a single-crystal mica film
prepared by particle track-etching (Sun et al., 2000a).

Recently, single-crystalline mica films have also been used for the fabrication
of track-etched pores (Sun et al., 1999a). Polymer membranes have the disad-
vantage that they are relatively soft, so the definition of the pore morphology
is not straightforward, and their internal pore surfaces can be quite rough.
Also, they are fairly limited in working temperatures for sample fabrication
and characterization compared to mica films, which are chemically stable
up to 770 K. Figures 3(a) and (b) show the SEM images of etched tracks in
polycarbonate membranes (Martin, 1994) and mica films (Sun et al., 2000),
respectively. The track-etched pores in the polymeric membranes are ap-
proximately cylindrical. Due to anisotropic etching rates, the pores in mica
films have a diamond-shaped cross section with a tapered pore wall along
the ion track (Sun et al., 1999a).
Nanochannel glass (NCG) has also been proposed for the template-
assisted synthesis of nanowires (Tonucci et al., 1992; Huber et al., 1994). It
contains a regular hexagonal array of capillaries similar to the pore structure
in anodic alumina. It is prepared by arranging two dissimilar glasses in a pre-
determined configuration. For example, Fig. 4(a) shows the use of cylindrical
rods of an acid-etchable glass as the cores in a hexagonally packed matrix
of inert glass tubes. The array of this core-tube assembly is then drawn at
high temperatures to reduce the cross-sectional area. By repeating the draw-
ing process, the channel diameters can be made as small as 33 nm, with a
packing density of 3 × 1010 pores/cm2 (Tonucci et al., 1992). Finally, the core
glass rods are etched away by an acid, yielding an array of cylindrical pores.
Figure 4(b) shows a SEM micrograph of an NCG with 33-nm channels ar-
ranged in a hexagonal close packing (Tonucci et al., 1992).
A number of other porous materials may be used as the host matrices
for nanowire fabrication. A mesoporous molecular sieve termed MCM-41
(Beck et al., 1992) possesses hexagonally packed pores with very small
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 173

FIG. 4. (a) Schematic diagram illustrating the fabrication process for nanochannel glass
array (Tonucci et al., 1992). (b) SEM image of a glass array with 33-nm channels after acid
etching (Tonucci et al., 1992).

channel diameters, which can be systematically varied between 2 nm and

10 nm using a surfactant-based supramolecular template (Ying et al., 1999).
The surfactant molecule consists of a hydrophilic head group that has a
high affinity for water and a long hydrophobic tail group. In the presence
of water, surfactant species will undergo self-aggregation to form micellar
structures, where the hydrophilic head groups are directed outward in con-
tact with water, whereas the hydrophobic tail groups form the micellar core
to minimize contact with water. Silicate precursors are then deposited on
the ordered arrays of micelles, which are self-assembled at a high enough
micellar concentration to form inorganic–organic mesostructures. The or-
ganic surfactants can be removed via heat treatment to yield silicates with
hexagonally packed cylindrical pores. Wu and Bein (1994) have fabricated
conducting organic filaments in the nanochannels of MCM-41. Han et al.
(2000) have also prepared metallic nanowires in SBA-15 silica with hexago-
nally packed mesopores templated with triblock copolymers. Recently, the
DNA molecule was also used as a template for growing nanometer-sized
wires (Braun et al., 1998).
The variety of porous solid materials that can be used as templates for
nanostructure synthesis has been reviewed by Ozin (1992). In the template-
assisted synthesis of nanowires, the pores or voids of the template are filled
with the chosen material using a number of approaches. Nanowires have
been derived via pressure injection, electrochemical deposition, and vapor
deposition, as described in the following sections.

1. Pressure Injection
In the pressure injection method, the nanowires are formed by inject-
ing the desired material in its liquid form into the pores of the template.
This technique has been used to fabricate a number of metallic and
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

174 YU-MING LIN ET AL.

FIG. 5. Schematic of the experimental setup for the pressure injection of materials into the
nanochannels of a porous template.

semiconducting nanowires (Huber et al., 1994; Zhang et al., 1998a; Lin et al.,
2000b). Before the injection process, the template is first cleaned to remove
any particles or substances that may prevent the pores from being filled.
The template is then placed with the solid material in a high-pressure cham-
ber. The chamber is evacuated at a temperature slightly below the melting
point of the solid material for a few hours to degas the template. It is then
brought to a temperature above the melting point of the solid, so that the
porous template is immersed in a liquid melt. Next, the chamber is discon-
nected from the vacuum system and filled with an inert gas such as argon.
The high pressure within the chamber forces the liquid melt into the pores
of the template. Following this pressure injection process, the chamber is
slowly cooled to solidify the impregnated material within the nanochan-
nels of the template before releasing the inert gas. By carefully controlling
the cooling rate for nanowire solidification, it is possible to fabricate essen-
tially single-crystalline wires (Zhang et al., 1999). Figure 5 shows a schematic
diagram of the experimental setup for the pressure injection process. The
detailed pressure injection procedures are described by Huber et al. (1994)
and Zhang et al. (1998a).
In the pressure injection process, a greater pressure is needed to overcome
the surface tension for the liquid melt to enter pores of a smaller diameter.
The relation between the pore diameter dw and the required pressure P is
given by the Washburn equation (Adamson, 1982),

dW = −4γ cos θ/P, (1)

08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 175

where γ is the surface tension of the liquid and θ is the contact angle between
the liquid and the template. For most liquid metals and semiconductors of
interest, the surface tension ranges from 100 to 600 dyne/cm. Assuming
the least favorable case of a nonwetting interface (θ = 180◦ ) and a medium
surface tension of 400 dyne/cm, filling channels 40 nm in diameter would
require a pressure of about 400 bar. To reduce the pressure required or to
maximize the filling factor, additives may be used to decrease the surface
tension or the contact angle. For example, it was found that the introduction
of a small quantity of copper atoms into the bismuth melt could facilitate the
injection of liquid bismuth into porous anodic alumina (Zhang et al., 1999).
For the pressure injection approach, the templates employed have to
be chemically stable and structurally robust under the high pressures and
temperatures involved. Anodic alumina and nanochannel glass can be used
as template materials for this nanowire synthesis approach. Metallic
(e.g. Bi, In, Sn and Al) and semiconducting (Se, Te, GaSb and Bi2Te3)
nanowires have been pressure injected into anodic alumina templates
(Huber et al., 1994; Zhang et al., 1998a; Lin et al., 2000b). The following
describes some of the important properties of bismuth nanowires fabricated
by the pressure injection method.
Figure 6(a) shows a SEM image of bismuth nanowires embedded in an
anodic alumina template after pressure injection. The pore diameter of the
template was about 42 nm, and the maximum pressure and temperature ap-
plied for the injection process were about 310 bar and 325◦ C, respectively.
A small amount of copper was introduced into liquid bismuth to enhance
its injection; the copper atoms should be segregated from bismuth during
solidification because they have zero solubility in solid bismuth. Upon solid-
ification, the copper flakes were brought to the top surface of Bi due to their

FIG. 6. (a) SEM image of the bottom surface of an anodic alumina template filled with
bismuth. The pore diameter is 42 nm. (b) TEM micrograph of the cross section of a 65-nm
bismuth nanowire array (Zhang et al., 1999).
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

176 YU-MING LIN ET AL.

FIG. 7. (a) A HRTEM image of a 40-nm freestanding bismuth nanowire, showing lattice
fringes. The amorphous surface layer is bismuth oxide formed upon air exposure of bismuth
nanowire. (b) SAED pattern of a single Bi nanowire (Zhang et al., 1999).

lower density, leaving pure bismuth nanowires within the pore channels.
Figure 6(b) shows a TEM image of the cross section of a 65-nm Bi nanowire
array, illustrating a high pore-filling factor (Zhang et al., 1999). Because
Bi has a higher electron density than anodic alumina, it is shown in dark
contrast in the TEM image. Figure 7(a) shows a high-resolution transmis-
sion electron microscopy (HRTEM) image of a 40-nm freestanding bismuth
nanowire, which was obtained by dissolving the anodic alumina template in
a special acid solution that did not attack the bismuth nanowires. The free-
standing wire was found to have a nearly uniform diameter (within 10%)
along its length. The lattice fringes in Fig. 7(a) indicated a highly crystalline
structure, as confirmed by the selected-area electron diffraction (SAED)
pattern shown in Fig. 7(b). It was found that the alumina matrix protected
the bismuth nanowires against oxidation, whereas freestanding wires were
gradually oxidized upon exposure to air. The amorphous surface layer in
Fig. 7(a) is the bismuth oxide developed when the nanowire was exposed to
air prior to imaging.
Figure 8 shows X-ray diffraction (XRD) patterns of bismuth nanowire
arrays (Lin et al., 2000b). It illustrates that the crystal structure of bismuth
nanowires is the same as that of bulk bismuth and that no copper phases
were present. The nanowires have a preferred wire orientation dependent
on their diameters. The major orientations of the 95-nm and 40-nm bis-
muth nanowire arrays were normal to the (202) and (012) lattice planes,
respectively, indicating that most (> 80%) of the nanowires were oriented
along the [101̄1] and [011̄2] directions for dw ≥ 60 nm and dw ≤ 50 nm, re-
spectively (Zhang et al., 1999; Lin et al., 2000b). The existence of more
than one dominant orientation in the 52-nm Bi nanowires (Fig. 8(b)) was
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 177

FIG. 8. XRD patterns of bismuth–anodic alumina nanocomposites with average bismuth

wire diameters of (a) 40 nm, (b) 52 nm, and (c) 95 nm (Lin et al., 2000b). The Miller indices
corresponding to the lattice planes of bulk Bi are indicated above the individual peaks.

attributed to the transitional behavior of nanowires of ‘intermediate’ di-

ameters as the preferential growth orientation was shifted from [101̄1] to
[011̄2] with decreasing dw. Huber et al. (2000) found a preferred crystal ori-
entation [0001] along the wire axes for bismuth nanowires fabricated at a
much higher pressure (about 1.5 kbar), indicating the possibility that the wire
growth orientation might also depend on the applied pressure and other pro-
cess parameters during the injection process. Additional XRD studies also
revealed the presence of a metastable high-stress phase in the as-prepared
bismuth nanowires (Zhang et al., 1999), which could be converted to the
normal phase by a thermal annealing treatment.
The smallest diameter attained for bismuth nanowires by the pressure
injection method was about 13 nm, using a pressure of approximately 0.3
kbar (Zhang et al., 1999). Finer nanowires might be fabricated by increasing
injection pressures (Huber et al., 2000), but it remains to be seen if the anodic
alumina templates would retain their structural integrity under those high
pressures.

2. Electrochemical Deposition
Electrochemical deposition has attracted increasing attention as a tech-
nique for nanowire fabrication. Traditionally, electrochemistry has been
used to grow thin films on conducting surfaces. Because electrochemical
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

178 YU-MING LIN ET AL.

deposition is usually controllable in the direction normal to the substrate

surface, this method can be readily extended to fabricate one-dimensional
or zero-dimensional nanostructures if the deposition is confined within the
pores of an appropriate template. The electrochemical technique involves
first coating a thin conducting metal film onto one side of the porous mem-
brane to serve as the cathode for electroplating. The length of the nanowires
deposited can be controlled by the duration of the electroplating process.
Electrochemical deposition has been used to synthesize nanowires and su-
perlattices (Piraux et al., 1994; Blondel et al., 1994) of a variety of materials,
including metals (e.g. Bi (Piraux et al., 1999; Liu et al., 1998a), Co (Ferre
et al., 1997; Zeng et al., 2000), Fe (AlMawlawi et al., 1991; Peng et al., 2000),
Ni (Ferre et al., 1997; Sun et al., 1999a), Cu (Piraux et al., 1994; Blondel
et al., 1994), Ag (Bhattacharrya et al., 2000), and Au (Hornyak et al., 1997)),
semiconductors (e.g., CdS (Routkevitch et al., 1996a; Kouklin et al., 2000)),
superconductors (e.g., Pb (Yi and Schwarzacher, 1999)), and conducting
polymers (Martin, 1994; Piraux et al., 1999).
For electrochemical deposition, it is critical to choose a template system
that is chemically stable in the electrolyte used and during the electrolysis
process. The presence of cracks and defects in the templates will affect the
electrochemical process, because deposition will occur predominantly in the
more accessible cracks, leaving most of the nanochannels unfilled. Particle
track-etched mica films or polymer membranes have been employed in sim-
ple dc electrolysis. To use anodic alumina in dc electrochemical deposition,
the insulating barrier layer has to be removed first, and a metal film has to be
evaporated onto one side of the template. However, the rectifying properties
of the oxide barrier layer make it possible to use the ac deposition method
directly. Although the applied voltage is sinusoidal and symmetrical, the cur-
rent is greater during the cathodic half-cycles, so that deposition is dominant
over the stripping that occurs in the subsequent anodic half-cycles. By re-
taining the barrier layer and using the ac electrolysis approach, the problem
with cracks may be avoided, because there is no rectification at the defects,
and the deposition process is reversed during each anodic half-cycle. By this
approach, metals (e.g., Co (Zeng et al., 2000) and Fe (AlMawlawi et al., 1991;
Peng et al., 2000)) and semiconductors (e.g., CdS (Routkevitch et al., 1996a;
Kouklin et al., 2000)) have been deposited within the pores of anodic alumina
templates without removing the barrier layer.
In contrast to nanowires synthesized by the pressure injection method,
nanowires derived electrochemically are typically polycrystalline with no
preferred crystal orientation. However, some exceptions occur. For example,
polycrystalline CdS nanowires fabricated by ac electrodeposition in anodic
alumina were shown to have a preferred wire growth orientation along the
c axis (Routkevitch et al., 1996). Recently, Xu et al. (2000a, 2000b) prepared
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 179

FIG. 9. (a) TEM image of a single Co(10 nm)/Cu(10 nm) multilayered nanowire. (b) A
selected region of the sample at high magnification. (Piraux et al., 1994.)

a number of single-crystalline nanowires of II–VI semiconductors (e.g., CdS,

CdSe, and CdTe) in anodic alumina by dc electrodeposition in a nonaque-
ous electrolyte. Yi and Schwarzacher (1999) found that single-crystalline
Pb nanowires can be formed by pulse electrodeposition under an overpo-
tential, but no specific crystal orientation was noted along the wire axis.
One advantage of the electrochemical deposition technique is the possi-
bility of fabricating multilayered structures within the nanowire systems. By
varying the cathodic potentials in the electrolyte that contains two different
kinds of ions, different metal layers are deposited controllably. Co–Cu mul-
tilayered nanowires have been synthesized to study the giant magnetoresis-
tance (GMR) effect (Piraux et al., 1994; Blondel et al., 1994). Figure 9 shows
TEM images of a single Co/Cu nanowire that is approximately 40 nm in
diameter prepared by Piraux et al. (1994). The light bands represent Co-rich
regions and the dark bands represent Cu-rich layers. This electrodeposition
method provides a low-cost, effective approach for preparing multilayered
one-dimensional nanostructures.

3. Vapor Deposition
Vapor deposition methods include physical vapor deposition (PVD)
(Heremans et al., 2000) and chemical vapor deposition (CVD) (Cheng et al.,
2000) or metallorganic chemical vapor deposition (MOCVD) (Berry et al.,
1996). Vapor deposition and electrochemical deposition are capable of
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

180 YU-MING LIN ET AL.

deriving nanowires of smaller diameters (≤ 20 nm) more readily than pres-

sure injection methods because they do not involve the use of high pressures
to deposit materials inside the porous channels.
The experimental setup and procedures for the physical vapor deposition
of bismuth nanowires are described by Heremans et al. (2000). The target
material is placed in a crucible that is covered with the porous template to
be filled. The template used for vapor deposition has open pores on both
sides. The crucible, surrounded by heating wires, is inserted into a vacuum
chamber. It is heated to melt the target material and to provide a high vapor
pressure. Because the pressure outside is lower than that within the crucible,
the vapor passes through the pores of the template. A slow cooling process is
then initiated to produce a temperature gradient across the porous template.
Because the temperature of the outer surface of the template is lower than
that of the inner surface, the vapor begins to condense within the pores from
the outer surface, and nanowires are grown inward. The process is completed
once the temperature on the inner surface of the template falls below the
melting point of the target material. Heremans et al. (2000) have synthesized
nearly single-crystalline bismuth nanowires within anodic alumina by this
approach; these nanowires possess a preferred crystal growth orientation
along the wire axis, similar to that prepared by the pressure injection method
(Zhang et al., 1999; Heremans et al., 2000).
Compound materials can also be prepared by the vapor deposition tech-
nique with two reacting gases (Cheng et al., 2000). The reacting species with
the lower melting point is placed in the crucible. The crucible, covered by the
porous template, is inserted into a tube that contains the other reacting gas.
The tube is heated up to vaporize the reactants in the crucible. The two gases
then react to produce compound wires within the nanochannels of the tem-
plate. Single-crystalline GaN nanowires have been synthesized in an anodic
alumina template through reaction of Ga2O vapor with a flowing ammonia
stream (Cheng et al., 1999, 2000). A different liquid/gas-phase approach has
been used by Berry et al. (1996) to prepare polycrystalline GaAs and InAs
nanowires in a nanochannel glass array. In this case, the nanochannels are
filled with one liquid precursor (e.g., trimethyl gallium or triethyl indium)
via a capillary effect, and the nanowires are formed by reaction between the
liquid precursor and the gas reactant (e.g., AsH3).
Recently, carbon nanotubes, an important class of one-dimensional nano-
structures, have been fabricated within the pores of anodic alumina via CVD
(Davydov et al., 1999; Li et al., 1999; Iwasaki et al., 1999; Suh et al., 1999).
A small amount of metal (e.g., Co) is first electrochemically deposited on
the bottom of the pores as a catalyst for the carbon nanotube growth, and
the template is heated to 700 to 800◦ C in a flowing gas mixture of N2 and
acetylene or ethylene. The hydrocarbon molecules are then pyrolyzed to
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 181

form carbon nanotubes within the porous template. Using anodic alumina
as the matrix, a highly ordered two-dimensional array of carbon nanotubes
has been achieved. Such well-aligned nanotube array has stimulated much
interest with its great potential for applications such as cold-cathode flat-
panel displays.

B. LASER-ASSISTED SYNTHESIS

In addition to the template-assisted method, laser-assisted synthesis has

been developed for generating nanowires. It is particularly useful for produc-
ing large quantities of crystalline semiconducting nanowires with ultrafine
diameters (≤ 10 nm). Nanowire synthesis by the laser-assisted method is
based on the vapor-liquid-solid (VLS) growth of single-crystalline silicon
whiskers (Wagner and Ellis, 1964), which was discovered in the early 1960s.
In VLS growth, a liquid metal droplet or catalyst cluster forms an energeti-
cally favored site for the adsorption of gas-phase reactants and the nucleation
site for crystallization when supersaturated. The reactant crystallizes at the
surface of the liquid cluster, and a preferentially one-dimensional structure
is developed. Figure 10 shows a schematic of the silicon nanowire growth
by the VLS mechanism. The one-dimensional structure obtained by such
growth has a diameter larger than 0.1 µm, which is limited by the minimum
size of the liquid droplet.
Recently, a laser ablation–condensation technique was used to produce
nanometer-sized catalyst clusters to grow nanowires by the VLS method.
A schematic of the laser ablation apparatus used by Morales and Lieber
(1998) to produce silicon nanowires is shown in Fig. 11. The target consists
of silicon and the catalyst material (e.g., Si1-xFex), and a pulsed laser is used
to produce nanometer-sized catalyst clusters within a reaction chamber at
1200◦ C. The ablated materials are carried by an argon gas flow, and the

FIG. 10. Schematic illustrating the growth of silicon nanowires by the VLS mechanism.
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

182 YU-MING LIN ET AL.

FIG. 11. Schematic of the laser ablation apparatus for the production of nanowires (Morales
and Lieber, 1998).

nanowires generated are collected at the cold finger (Morales and Lieber,
1998; Lieber 1998). The growth of nanowires is affected by the reaction tem-
perature and the catalyst material used (Morales and Lieber, 1998; Lieber,
1998; Duan and Lieber, 2000). It is believed that the nanowire growth oc-
curs only when the catalyst cluster remains a liquid (Morales and Lieber,
1998). The wire growth ceases as the catalyst cluster is solidified when it
is carried away from the hot zone of the furnace by the gas flow. All the
nanowires are found to terminate at one end with catalyst clusters of diam-
eters that are 1.5 to 2 times that of the wires. The nanowires range from
3 nm to tens of nanometers in diameter, with lengths up to tens of microm-
eters (Morales and Lieber, 1998; Duan and Lieber, 2000). Using the laser
ablation technique, Lieber and coworkers have prepared a wide range of
semiconducting nanowires, including group IV elements (Si and Ge), III–V
compounds (GaAs, GaP, InAs, and InP), II–VI compounds (ZnS, ZnSe, CdS,
and CdSe), and binary SiGe alloys (Morales and Lieber, 1998; Duan et al.,
2000; Duan and Lieber, 2000). Lee and coworkers also fabricated Si and
Ge nanowires using a similar laser ablation apparatus (Zhang et al., 1998b,
2000a).
Figure 12 shows TEM images of silicon nanowires fabricated by Morales
and Lieber (1998). The nanowires are produced by laser ablating a Si0.9Fe0.1
target. In Fig. 12(a), the darker spheres with a diameter larger than the
nanowires are the solidified FeSi2 catalyst clusters that terminate at one
end of the nanowires. Electron diffraction pattern indicates that the sil-
icon nanowires are single-crystalline and grow along the [111] direction
( Fig. 12(b)) (Morales and Lieber, 1998). The TEM image in Fig. 12(b) shows
the silicon nanowire to be uniform in diameter, with a crystalline core that
is surrounded by an oxide layer, which may have resulted from the presence
of residual oxygen (Morales and Lieber, 1998; Duan et al., 2000). Wang et al.
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 183

FIG. 12. (a) TEM images of Si nanowires produced after laser ablating a Si0.9Fe0.1 target.
The dark spheres with a slightly larger diameter than the wires are solidified catalyst clusters
(Morales and Lieber, 1998). (b) Diffraction contrast TEM image of a Si nanowire. The crystalline
Si core appears darker than the amorphous oxide surface layer. The inset shows the convergent
beam electron diffraction pattern recorded perpendicular to the wire axis (Morales and Lieber,
1998).

(1998a,b) found that these oxides played a more important role than metals
in assisting the nanowire growth. SiOx was discovered to be an effective cat-
alyst that significantly increased the yield of silicon nanowires (Zhang et al.,
1998b; Wang et al., 1998a, b). Similar yield enhancement was also found in the
synthesis of germanium nanowires by laser ablating a germanium powder
mixed with GeO2 (Zhang et al., 2000a). It was observed that targets com-
posed of equal molar ratios of Ge:GeO2 and Si:SiO2 in the absence of metal
catalysts gave the maximum yields for germanium and silicon nanowires,
respectively (Zhang et al., 2000a; Wang et al., 1998a,b). Based on these ob-
servations and other TEM studies (Zhang et al., 2000a; Wang et al., 1998a;
Lee et al., 1999), an oxide-enhanced nanowire growth mechanism different
from the classical VLS mechanism was proposed (Wang et al., 1998a). It was
postulated that the nanowire growth was catalyzed by the GemO or SimO
layer (m > 1) on the nanowire tips, which might be in or near their molten
states (Lee et al., 1999). Because the oxide-assisted laser ablation method
did not require any metal catalysts, the resulting nanowires could achieve
a high purity. The germanium and silicon nanowires produced from these
catalyst-free targets were noted to grow generally along the [112] crystal
direction (Lee et al., 1999).
Although laser ablation methods have been used to fabricate large quan-
tities of single-crystalline semiconducting nanowires with high aspect ratios,
the wires obtained are randomly oriented with a variety of morphologies
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

184 YU-MING LIN ET AL.

(Tang et al., 1999) and correspond to broad distributions of wire diameters

and lengths. Therefore, it would be important to devote future research
efforts toward controlling the wire diameter, length, morphology, and as-
sembly in such synthesis in order to utilize the nanowires derived effectively
for potential electronic applications.

C. OTHER SYNTHESIS METHODS

Recently, the VLS growth method has been extended beyond the gas-
phase reaction to synthesis of Si nanowires in Si-containing solvent (Holmes
et al., 2000). In this case 2.5-nm Au nanocrystals were dispersed in supercrit-
ical hexane with a silicon precursor (e.g., diphenylsilane) under a pressure of
200–270 bar at 500◦ C, at which temperature the diphenylsilane decomposes
to Si atoms. The Au nanocrystals serve as seeds for the Si nanowire growth,
because they form an alloy with Si, which is in equilibrium with pure Si. It
is suggested that the Si atoms would dissolve in the Au crystals until the sat-
uration point is reached; then they are expelled from the particle to form a
nanowire with a diameter similar to the catalyst particle. This method has an
advantage over the laser-ablated Si nanowire in that the nanowire diameter
can be well controlled by the Au particle size, whereas liquid metal droplets
produced by the laser ablation process tend to exhibit a much broader size
distribution. With this approach, highly crystalline Si nanowires with diame-
ters ranging from 4 nm to 5 nm have been produced by Holmes et al. (2000).
The crystal orientation of these Si nanowires can be controlled by the reac-
tion pressure.
In addition to the fabrication methods discussed here, special techniques
have been developed to prepare a variety of one-dimensional systems for
investigating interesting quantum phenomena. For example, quantized con-
ductance (in units of G 0 = 2e2/ h, where h is the Planck constant) has been
observed through thin metallic nanowires with scanning tunneling micro-
scopy (Pascual et al., 1993, 1995; Brandbyge et al., 1995) and mechanically
controlled break junctions (Muller et al., 1992, 1996). These techniques use
the same basic principle of pressing two metals together and pulling them out
to form a nanometer-sized wire between the two contacts. One-dimensional
silicon nanostructures have also been fabricated on a silicon-on-insulator
(SOI) substrate by a combination of lithography and orientation-dependent
etching (Namatsu et al., 1997). Other approaches include forming nanowires
by vapor deposition on cleaved superlattice planes (Arakawa et al., 1996) or
by shadow deposition on stepped lattice planes (Sugawara et al., 1997). These
fabrication techniques usually require unique experimental configurations
or complicated processing and produce nanowires with a very low yield or
at a very high cost. Therefore, although the resulting nanostructures may be
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 185

helpful toward understanding low-dimensional phenomena, their utilization

in commercial applications and devices is currently limited.

III. Theoretical Modeling of Nanowire Band Structures

Nanostructured materials exhibit behavior distinct from their bulk coun-

terparts due to quantum confinement effects. As their length scale shrinks
to a size comparable to the de Broglie wavelength of electrons, the energy of
electrons in the confined direction becomes quantized and forms a discrete
energy spectrum. If the separation between the quantized energy levels is
much larger than the thermal excitation energy kBT, virtually all the elec-
trons will occupy the lowest possible states. Because it is not likely to promote
electrons to a higher energy state in the absence of an external excitation, the
degree of freedom for electrons is quenched in the confined direction, pro-
ducing a system with reduced dimensionality. The systems confined in one,
two, and three dimensions are treated as a two-dimensional (quantum well),
a one-dimensional (quantum wire), and a zero-dimensional (quantum dot)
electron gas, respectively. The energy quantization changes the band struc-
ture of nanostructured materials and alters their optical, magnetic, and elec-
tronic properties dramatically. Thus, the effects of quantum confinement
on low-dimensional systems are of both fundamental and technological
importance. In the following sections, a theoretical framework is presented
for the modeling of one-dimensional nanowire systems. The Schrödinger
equation has been solved to obtain the band structure of nanowires. Based
on the band structure, a semiclassical model is developed to predict various
transport properties of nanowire systems.

A. BAND STRUCTURES OF ONE-DIMENSIONAL SYSTEMS

The electronic states of nanowire systems exhibit a very different spec-

trum from that of bulk materials. In order to understand their unique elec-
tronic properties, we have modeled the band structure of these
one-dimensional systems.
Without loss of generality, we assume a bulk material where the major
carriers are electrons with an effective mass me. In general, the electron
masses are anisotropic, and the effective mass is expressed as a symmetric
second-rank tensor. The dispersion relation of the electrons is written as
h̄ 2
E(k) = k · α · k, (2)
2
where k is the wave vector and α is the inverse tensor of me. From the
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

186 YU-MING LIN ET AL.

effective mass theorem, the envelope wavefunction of electrons, ψ(r), is

described by the Schrödinger equation:
h̄ 2
− ∇ · α · ∇ψ(r) = Eψ(r). (3)
2
For nanowires embedded in an insulating matrix with a large band gap
(e.g., alumina or mica), electrons are well confined within the wires. Thus, to
a good approximation, the electron wavefunction, ψ(r), can be assumed to
vanish at the wire boundary.
For an infinitely long wire with a circular cross section of diameter dw, we
take the z axis to be parallel to the wire axis, with the x and y axes lying on
the cross-sectional plane. The cylindrical symmetry of the wire is then used
to simplify Eq. (3) by making αx y = α yx = 0, which can be achieved by a
proper rotation about the z axis. The wave function ψ(r) then has the form
ψ(r) = u(x, y) exp(iξ · x) exp(iη · y) exp(ik z · z), (4)
where ξ and η are constants to be determined and kz is the wave number
of the traveling wave in the z direction. By letting ξ = −(αx z /αx x )k z and
η = −(α yz /α yy )k z , Eq. (3) is reduced to a concise second-order differential
equation in x and y only:
   
h̄ 2 ∂2 ∂2 h̄ 2 k z2
− αx x 2 + α yy 2 u(x, y) = E − u(x, y), (5)
2 ∂x ∂y 2m zz
where m zz = ẑ · me · ẑ is the transport effective mass along the wire axis.
Equation (5) is reminiscent of a two-dimensional Schrödinger equation with
in-plane effective mass components
m x ≡ αx−1x
m y ≡ α −1
yy (6)
in the x and y directions, respectively. Because u(x, y) must satisfy the bound-
ary condition: u(x, y) = 0 when x 2 + y 2 = (dw /2)2 , the eigenvalues of u(x, y)
in Eq. (5) are quantized, and the energy of the electrons is written as
h̄ 2 k z2
E nm (k z ) = εnm + , (7)
2m zz
where εnm is the eigenvalue of Eq. (5) corresponding to the subband edge
eigenstate at k z = 0, labeled by the quantum numbers (n, m).
In a nanowire system, the quantized subband energy ε nm and the transport
effective mass mzz along the wire axis are the two most important parameters
and determine almost all the electronic properties. Due to the anisotropic
carriers and the special geometric configuration (circular wire cross section
and high aspect ratio of length to diameter), several approximations were
used in earlier calculations to derive ε nm and mzz in bismuth nanowires. In the
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 187

first calculation carried out by Zhang et al. (1998c, 2000b), the quantized en-
ergy levels were evaluated by using a cyclotron effective mass approximation
for the in-plane effective mass. An improved model was subsequently devel-
oped by Sun et al. (1999b) based on a square wire cross-section approxima-
tion, which allowed an analytical expression to be obtained for the wave func-
tions. However, these two approximations were valid only for less anisotropic
carriers, producing significant discrepancies for highly anisotropic systems
(αx x /α yy 1 or αx x /α yy  1). Thus, a numerical approach was developed
recently to more accurately determine ε nm and mzz (Lin et al., 2000c).
For a simple case where αx x = α yy , the wave function of Eq. (5) has the
analytical solution
u nm (r) ∼ Jn (χnm r )ein θ , (8)
where Jn is the nth Bessel function and χnm is determined by the mth root of
Jn (x · dw /2) = 0. The subband energy εnm corresponding to the wavefunction
u nm (r) is given by
h̄ 2
εnm = αx x χnm
2
. (9)
2
For the general case where αx x = α yy , there are no analytical solutions, and
the only possible approach to determine the quantized subband energy εnm
from Eq. (5) is through numerical methods (Lin et al., 2000c). In this instance,
a mesh consisting of M concentric circles and N sectors is created within the
wire cross section, as shown in Fig. 13. The differential equation of Eq. (5) is
then transformed to a set of difference equations based on the grid points on

FIG. 13. Schematic of the grid points used to transform the differential equation into a
difference equation. The mesh in the circular wire cross section consists of M concentric circles
and N sectors. In this figure, M = 5 and N = 12 (Lin et al., 2000c).
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

188 YU-MING LIN ET AL.

FIG. 14. Calculated subband energies in units of ε0 = 2αx xh̄ 2 /dw2 as a function of the
in-plane mass anisotropy α yy /αx x . The subband energies of nanowires of various diameters
can then be derived from this figure.

the mesh, which can be solved readily with the aid of computers. By refining
the mesh, εnm can be obtained with great accuracy (within 0.1%). Figure 14
shows the calculated subband energies in units of 2αx xh̄ 2 /dw2 as a function of
the mass anisotropy α yy /αx x .
The density of states (DOS) of electrons in nanowires is derived from
Eq. (7) as
√
2m zz 
g(E) = (E − εnm )−1/2 . (10)
πh̄ n,m

Figure 15 shows the calculated DOS for electrons in a 40-nm bismuth nano-
wire compared to that of bulk bismuth. The DOS in nanowires is a super-
position of one-dimensional transport channels, each located at a quantized
subband energy εnm . We note that the DOS in nanowires has sharp peaks at
the subband edges, whereas that in a bulk material is a smooth monotonic
function of energy. The enhanced DOS at the subband edges of nanowires
has important implications for many applications, such as in optics (Black
et al., 2000) and thermoelectrics (Hicks and Dresselhaus, 1993).

B. THE SEMIMETAL–SEMICONDUCTOR TRANSITION IN

SEMIMETALLIC NANOWIRES

For semimetals in bulk form, such as bismuth, the conduction band over-
laps in energy with the valence band, and the electronic properties are
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 189

FIG. 15. Calculated effective densities of states for 40-nm bismuth nanowires (solid curve)
and bulk bismuth (dashed curve). The zero energy refers to the band edge of bulk bismuth.
The nonparabolic effects of the electron carriers are considered in these calculations.

governed by both the electrons and holes. In nanowires, the quantum con-
finement effects cause the band edge of the electrons to move up in energy
(see Fig. 16), whereas the valence band edge decreases in energy. The band-
edge energies of electrons and holes shift in opposite directions in bismuth
nanowires, decreasing the energy overlap between the conduction band and
the valence band (Fig. 16). As the wire diameter continues to decrease, the

FIG. 16. Schematic illustrating the semimetal–semiconductor transition in nanowires made

of semimetals: (a) bulk semimetals with a band overlap between the electrons and the holes,
(b) nanowires with the critical wire diameter dc where the band overlap vanishes, and (c)
nanowires with diameters smaller than dc , exhibiting a bandgap between the conduction and
valence bands.
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

190 YU-MING LIN ET AL.

energy overlap eventually vanishes (Fig. 16(b)), producing a band gap be-
tween the lowest conduction subband and the highest valence subband. This
semimetal-to-semiconductor transition will occur at a critical wire diame-
ter dc, which depends on the band overlap energy, the electron and hole
effective masses, and the crystal orientation along the wire axis for the ma-
terial of interest. The critical wire diameters for two group V semimetals,
bismuth and antimony, are predicted to be about 50 nm and 10 nm (Lin
et al., 2000c; Heremans et al., 2001), respectively. This quantum confinement–
induced semimetal–semiconductor transition is one of the unique properties
of nanowires made of semimetallic materials. Such a transition dramatically
alters the electronic properties of nanowires, providing us with new possi-
bilities for manipulating the band structures of materials.
Due to the semimetal–semiconductor transition, the carrier concentra-
tion N(T ) of nanowires made of semimetallic materials is highly dependent
on the wire diameter and temperature and must satisfy the condition dw ≤ dc
to exhibit semiconducting behavior. As an example, Fig. 17 shows the cal-
culated total carrier densities for various bismuth nanowires oriented along
the [011̄2] growth direction as a function of temperature (Lin et al., 2000c).
Three different types of temperature dependence for the carrier density
are predicted for bismuth nanowires, depending on the wire diameters. For
10-nm bismuth nanowires, which are in the semiconducting regime, the car-
rier density increases exponentially with temperature. For 80-nm bismuth
nanowires, which remain in the semimetallic regime for all temperatures,
the carrier density is similar in temperature dependence to bulk bismuth.

FIG. 17. Calculated total carrier density (electrons and holes) as a function of temperature
for bulk 3D bismuth and bismuth nanowires of different diameters oriented along the [011̄2]
direction.
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 191

The lower carrier density of the 80-nm nanowires compared to bulk bismuth
is due to the smaller band overlap in the former. For the 40-nm bismuth
nanowires, the carrier density has a temperature dependence similar to bulk
bismuth at high temperatures, but it drops rapidly with decreasing tempera-
ture at low temperatures. Because the carrier density is highly dependent on
wire diameter, the transport properties of bismuth nanowires are expected
to be highly sensitive to wire diameter, as will be shown experimentally in
the section “temperature-dependent resistivity of nanowires.”

IV. Transport Properties

The transport phenomena in low-dimensional systems can be roughly di-

vided into two categories: ballistic transport and diffusive transport. Ballistic
transport occurs when the electrons travel across a nanowire without any
scattering. In this case, the conduction is determined mainly by the contacts
between the nanowire and the external circuit, and the conductance is quan-
tized into a universal conductance unit, G 0 = 2e2/ h (Wharam et al., 1988;
van Wees et al., 1988). Ballistic transport is usually observed in very short
quantum wires, such as those produced by mechanically controlled break
junctions (Muller et al., 1992, 1996) or by scanning tunneling microscopy
(Pascual et al., 1993, 1995; Brandbyge et al., 1995), whereby the wire length
is much shorter than the electron mean free path and the conduction is a pure
quantum phenomenon. An additional requirement for ballistic transport is
that k B T  ε j − ε j−1 , where ε j − ε j−1 is the subband separation between the
j and j − 1 subband energy levels. On the other hand, for nanowires with
lengths much longer than the carrier mean free path, the electrons or holes
undergo numerous scattering events when they travel along the wire. In this
case, electron transport is in the diffusive regime, and conduction is domi-
nated by scattering due to phonons (lattice vibrations), boundary or lattice
defects, and impurity atoms.
The ballistic transport of one-dimensional systems has been extensively
studied since the discovery of quantized conductance (Wharam et al., 1988;
van Wees et al., 1988). In contrast, due to the difficulty in fabricating long
nanowires (> 1 µm) and the inadequacy of conventional experimental tech-
niques, there have been fewer experimental and theoretical studies on trans-
port phenomena in the diffusive regime. Recently, a semiclassical transport
model based on the band structure of nanowires was developed for systems
in which scattering events are not negligible (Lin et al., 2000b,c). The diffu-
sive transport model has been applied to bismuth and antimony nanowire
systems (Lin et al., 2000b; Heremans et al., 2001) and agrees well with the
experimental findings.
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

192 YU-MING LIN ET AL.

A. SEMICLASSICAL MODEL

The diffusive transport phenomena in nanowires can be described by a

semiclassical model based on the Boltzmann transport equation. For car-
riers in a one-dimensional subband, important transport coefficients, such
as the electrical conductivity, σ , the Seebeck coefficient, S, and the thermal
conductivity, κ e, are derived as (Sun et al., 1999b; Ashcroft and Mermin,
1976a)
σ = L (0) (11)
1 L (1)
S=− (12)
eT L (0)
 
1 (L (1) )2
κe = 2 L −
(2)
, (13)
e T L (0)
where T is temperature and
  
4 dk df
L (α) = e2 − τ (k)ν(k)ν(k)(E(k) − E f )α , (14)
π 2 dw2 dE
where α = 0, 1, 2, k is the wave vector along the transport direction, E(k)
denotes the carrier dispersion relation, ν(k) is the group velocity, τ (k) is
the relaxation time, Ef is the Fermi energy, and f(E) is the Fermi–Dirac dis-
tribution function. Assuming a parabolic dispersion relation, the transport
elements L(α) in Eqs. (11)–(13) are derived as
 
1
L (0) = D F −1/2 (15)
2


(k B T )D 32 F1/2 − 12 ς ∗ F−1/2 (for electrons)

L (1) = 3 1 ∗
 (16)
−(k B T )D 2 F1/2 − 2 ς F−1/2 (for holes)
 
5 ∗ 1 ∗2
L (2) = (k B T ) D
2
F3/2 − 3ς F1/2 + ς F−1/2 , (17)
2 2

where D is given by
 1/2
8e 2m ∗ k B T
D= 2 2 = µ, (18)
π dw h̄ 2
in which µ is the carrier mobility along the nanowire and
 ∞
x j dx
Fj = (19)
0 exp(x − ς ∗ ) + 1
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 193

denotes the Fermi–Dirac related functions, with fractional indices. j = − 12 ,

, . The reduced chemical potential, ς ∗ , is defined as
1 3
2 2
 (0) 
∗
E f − εe /k B T (for electrons)
ς =  (0)  (20)
εh − E f /k B T (for holes),
(0) (0)
where εe and εh denote the band edges for electrons and holes, respectively.
In considering the transport properties of real one-dimensional nanowires,
contributions from all the subbands near the Fermi energy should be in-
cluded, and the L (α) s in Eqs. (11)–(13) should be replaced by the sum of
(α)  (α)
contributions from each subband i, L total = i L i , to obtain the various
transport coefficients.
It should be noted that the carrier mobility in nanowires is lower than
that in bulk single-crystalline material due to possible scattering at wire and
grain boundaries, uncontrolled impurities, and lattice defects. The overall
effect of this additional scattering is taken into account by Matthiessen’s
rule (Ashcroft and Mermin, 1976b),
1 1 1 1
= + + , (21)
µtot (T ) µbulk (T ) µbound µimp (T )
where µbulk is the carrier mobility in bulk crystalline material and the terms
µ−1 −1
bound and µimp account for boundary scattering and charged impurity scat-
tering, respectively. Also, µ−1bound is usually assumed to be independent of
temperature, whereas µ−1 imp has a temperature dependence of T
1.5
for most
charged impurity scattering processes.

B. TEMPERATURE-DEPENDENT RESISTIVITY OF NANOWIRES

A number of transport measurements have been performed on nanowires

prepared by the template-assisted approach. The nanowire-embedded tem-
plate provides a convenient package for making electrical contacts on both
ends of the wires, so that two-point transport measurements can be per-
formed. Transport properties of various nanowire arrays have been mea-
sured by Zhang et al. (1998c, 2000b), Heremans et al. (1998, 2000), Liu et al.
(1998b), Huber et al. (1999), Hong et al. (1999), Lin et al. (2000b), and Sun
et al. (2000). Although a two-point resistance measurement has the advan-
tage of simplicity, an absolute resistivity value for the nanowires cannot be
determined by this approach, because the number of wires in the template
contributing to the conduction measurement is not known. Progress was
recently made to characterize the absolute resistivity of a single nanowire
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

194 YU-MING LIN ET AL.

FIG. 18. SEM image of a 70-nm bismuth nanowire with four electrodes attached to the
nanowire. The circle on the large left electrode is a reference point used to find the nanowire
and to attach electrodes to it by a lithographic process (Cronin et al., 1999).

via a four-point setup (Cronin et al., 1999, 2000), which could provide more
physical information on the properties of nanowires than the normalized
resistance of a nanowire array. Figure 18 shows an SEM image of a four-
point electrode patterned on a 70-nm Bi nanowire (Cronin et al., 1999). The
circular dot in Fig. 18 represents one of the prepatterned grid points used
to locate the nanowires on the substrate. In this case, nanoelectrodes were
patterned by electron-beam lithography on top of a single nanowire on a
substrate coated with a thin insulating layer (Cronin et al., 1999, 2000). The
electrodes consisted of a gold layer (∼ 1000 Å thick) and a thin adhesive
layer, and their processing followed a standard lift-off method. We note that
most nanowires would undergo surface oxidation upon removal from the
template. The surface oxide layer on the nanowires imposes a serious prob-
lem in making nanoelectrical contacts; efforts are currently being devoted
to tackling this challenge.
Figure 19(a) shows the temperature dependence of resistance R(T) for
bismuth nanowire arrays (dw = 7 − 200 nm) synthesized by vapor deposition
and measured by Heremans et al. (2000). Hong et al. (1999) reported simi-
lar resistance measurements on bismuth wires of larger diameters (200 nm to
2 µm) prepared by electrochemical deposition (Fig. 19(b)). These two studies
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 195

FIG. 19. (a) Measured temperature dependence of resistance for bismuth nanowire arrays
of various wire diameters dw (Heremans et al., 2000). (b) R(T )/R(290 K ) for bismuth wires of
larger dw measured by Hong et al. (1999). (c) Calculated R(T )/R(300 K ) of 36-nm and 70-nm
bismuth nanowires (Lin et al., 2000b). The dashed curve refers to a 70-nm polycrystalline wire
with increased boundary scattering.

showed that R(T ) of nanowires is highly sensitive to dw and is very different

from that of bulk bismuth. In Fig. 19(a), the temperature dependence of resis-
tance for bismuth nanowires exhibits dissimilar trends for dw > 50 nm and dw
≤ 50 nm. Based on the semiclassical transport model and the band-structure
model of bismuth nanowires, R(T )/R(300 K) was calculated for 36-nm and
70-nm wires. The two wire diameters were chosen to represent semicon-
ducting and semimetallic bismuth nanowires, respectively. The solid curves
in Fig. 19(c) (Lin et al., 2000b) illustrate that the calculated R(T )/R(300 K)
trends are consistent with those obtained experimentally (Fig. 19(a)). We
note that the nonmonotonic R(T ) behavior for semimetallic nanowires ob-
served by Heremans et al., 2000 in Fig. 19(a) is not found in the wires of
Fig. 19(b). This inconsistency may be due to the crystal quality differences in
the bismuth wires prepared by the two different approaches, which can be
accounted for by µbound in the transport model. Instead of the nonmonotonic
behavior exhibited by single-crystalline semimetallic nanowires in Fig. 19(a),
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

196 YU-MING LIN ET AL.

FIG. 20. (a) Measured R(T )/R(270 K ) for 40-nm bismuth nanowires prepared with alloys
of different Te doping levels. (b) The calculated temperature dependence of µ−1
avg for 40-nm
undoped and Te-doped bismuth nanowires of different Nd . The dashed and solid lines are
fitting curves corresponding to undoped and Te-doped Bi nanowires, respectively.

R(T) is predicted to display a monotonic temperature dependence at a high

defect level. This is illustrated by the dashed curve in Fig. 19(c) for polycrys-
talline 70-nm bismuth wires. Because the nanowires prepared by electro-
chemical deposition were found to be polycrystalline, their carriers would
experience more boundary scattering, resulting in the monotonic R(T ) be-
havior noted experimentally in Fig. 19(b).
The same transport model has also been extended to describe the proper-
ties of Te-doped bismuth nanowires and antimony nanowires. Te, a group VI
element, is an electron donor in bismuth. Figure 20(a) shows the measured
R(T)/R(270 K) for 40-nm Bi nanowires with the Te concentrations used
to form the Bi–Te alloys for nanowire synthesis. The actual Te concentra-
tions in the bismuth nanowires would be smaller than these nominal values
because some Te atoms would segregate to the wire boundary during alloy
solidification (Zhang et al., 2000b). For simplicity, we assume that about 10%
of the Te dopants in the alloy melt are present in the final nanowire prod-
uct and that each Te atom donates one electron to the conduction band of
bismuth. Therefore, 0.025, 0.075, and 0.15 at% Te-doped Bi alloys give rise
to donor concentrations Nd of 6.67 × 1017 , 2.0 × 1018 , and 4.0 × 1018 cm−3 in
the respective nanowires. Based on the measured R(T ) in Fig. 20(a) and the
calculated temperature-dependent carrier density, the average carrier mo-
bility, µavg(T ), of the Te-doped bismuth nanowires can be obtained. Anal-
ogous to Eq. (21), µ−1 avg for doped bismuth nanowires can be related to the
various scattering processes by µ−1 −1 −1 −1
doped (T ) = µundoped (T ) + µimp (T ) + µdefect ,
where µundoped is the average mobility of the undoped bismuth nanowires
with the same diameter and µ−1 −1
imp and µdefect are associated with the in-
creased ionized impurity scattering and the higher defect level in Te-doped
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 197

FIG. 21. (a) Temperature dependence of the resistance measured for various antimony
nanowires, normalized to the resistance at 300 K. (b) Calculated R(T )/R(300 K ) for 10-nm
and 48-nm antimony nanowires.

bismuth nanowires, respectively. The µavg −1 (T ) values calculated are shown

in Fig. 20(b) as solid curves, which are qualitatively consistent with the ex-
perimental results (denoted by symbols).
Like bismuth, antimony is semimetallic in bulk form. The band overlap
between electrons and holes in antimony is about 180 meV at 4 K, which
is about five times larger than that in bismuth (∼ 38 meV). Therefore, the
semimetal–semiconductor transition would occur at a smaller wire diam-
eter in antimony nanowires (∼ 10 nm) compared to bismuth nanowires
(∼ 50 nm). Figure 21(a) shows the temperature dependence of the resis-
tance for antimony nanowires prepared by vapor deposition (Heremans
et al., 2001). One of the 10-nm nanowire arrays has less resistance varia-
tion with temperature than the other, probably due to differences in impu-
rity content or wire diameter distribution within the anodic alumina tem-
plates. Figure 21(b) illustrates the modeled R(T ) curves for 10-nm and
48-nm antimony nanowire arrays, which display trends that are qualitatively
consistent with the experimental results shown in Fig. 21(a) (Heremans
et al., 2001). The fact that the measured R(T ) of antimony and bismuth
nanowires can be explained by the same transport model suggests that the
different temperature dependences of the resistance between nanowires
and bulk materials arise from both quantum finite-size effects and classi-
cal finite-size effects. The classical finite-size effect decreases the carrier
mobility by limiting the carrier mean free path, whereas the quantum con-
finement effect alters the band structure (especially for semimetals) and
significantly changes the carrier density. Together, these two factors de-
termine the temperature dependence of resistance in the nanowire sys-
tems. Finite-size effects in nanowires have also been observed in the
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

198 YU-MING LIN ET AL.

magnetoresistance measurements of bismuth and nickel nanowire arrays

(Fasol, 1998; Sun et al., 2000).

V. Summary

Recent developments in the fabrication of nanowires have led to signif-

icant advances in research on these low-dimensional systems. The laser-
assisted synthesis has successfully produced a variety of semiconducting
nanowires with ultrafine diameters. In contrast, most metallic nanowires
have been prepared in a regular array with a well-controlled wire-packing
geometry via the template-assisted approach. Nanoporous templates pro-
vide a useful matrix for handling nanowires and for device integration. The
semimetallic nanowires of bismuth and antimony have been shown to un-
dergo a transition to semiconducting behavior with decreasing wire diam-
eters. A generalized theoretical framework has been developed to predict
the band structure and transport properties of nanowire systems. Theoretical
calculations indicate that nanowires exhibit a very different band structure
from that of their bulk counterparts, resulting in unusual optical, electronic
and thermoelectric properties. Transport studies suggest that quantum con-
finement effects significantly perturb the carrier density, whereas the carrier
mean free path is limited mainly by the wire diameter. Consequently, trans-
port phenomena in nanowire systems can be manipulated by tuning the
diameter of these one-dimensional structures.
Nanowire systems present a great challenge for the understanding and
utilization of low-dimensional materials. Although device fabrication based
on nanowire systems is still in its infancy, these nanostructures have demon-
strated a significant potential for technological breakthroughs with their
unique band structure and transport properties. They offer tremendous re-
search opportunities at the frontiers of nanotechnology. Future advances
in this field would entail further interdisciplinary efforts toward develop-
ing novel synthesis method and control of self-assembled nanostructures,
detailed understanding and theoretical modeling of quantum confinement
effects, and sophisticated characterization of nanostructures with high aspect
ratios.

ACKNOWLEDGMENTS

The authors acknowledge helpful discussions with Dr. Z. Zhang, Dr. G. Dressel-
haus, Dr. J. Heremans, O. Rabin, M. R. Black, and S. B. Cronin. They are grateful to
the NSF, the U.S. Navy, the MURI program, and DARPA for financial support.
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 199

REFERENCES

Adamson, A. W., “Physical Chemistry of Surfaces.” Wiley, New York, 1982, p. 338.
AlMawlawi, D., Coombs, N., and Moskovits, M., Magnetic properties of Fe deposited into
anodic aluminum-oxide pores as a function of particle size. J. Appl. Phys. 70, 4421
(1991).
Arakawa, T., Watabe, H., Nagamune, Y., and Arakawa, Y., Fabrication and microscopic pho-
toluminescence imaging of ridge-type InGaAs quantum wires grown on a (110) cleaved
plane of AlGaAs/GaAs superlattice. Appl. Phys. Lett. 69, 1294 (1996).
Ashcroft, N. W., and Mermin, N. D., “Solid State Physics.” Holt, Rinehart and Winston, New
York, 1976a, Chap. 13.
Ashcroft, N. W., and Mermin, N. D., “Solid State Physics.” Holt, Rinehart and Winston, New
York, 1976b, Chap. 16.
Beck, J. S., Vartuli, J. C., Roth, W. J., Leonowicz, M. E., Kresge, C. T., Schmitt, K. D., Chu,
C. T-W., Olson, D. H., Sheppard, E. W., McCullen, S. B., Higgins, J. B., and Schlenker, J.
L., A new family of mesoporous molecular sieves prepared with liquid crystal templates.
J. Am. Chem. Soc. 114, 10834 (1992).
Berry, A. D., Tonucci, R. J., and Fatemi, M., Fabrication of GaAs and InAs wires in nanochannel
glass. Appl. Phys. Lett. 69, 2846 (1996).
Bhattacharrya, S., and Saha, S. K., Nanowire formation in a polymeric film. Appl. Phys. Lett.
76, 3896 (2000).
Black, M. R., Lin, Y.-M., Dresselhaus, M. S., Tachibama, M., Fang, S., Rabin, O., Ragot, F.,
Eklund, P. C., and Dunn, B., Measuring the dielectric properties of nanostructures using
optical reflection and transmission: bismuth nanowires in porous alumina. MRS Symp.
Proc. 581, 623 (2000).
Blondel, A., Meier, J. P., Doudin, B., and Ansermet, J.-Ph., Giant magnetoresistance of
nanowires of multilayers. Appl. Phys. Lett. 65, 3019 (1994).
Brandbyge, M., Schiøtz, J., Sörensen, M. R., Stoltze, P., Jacobsen, K. W., Nørskov, J. K., Olesen,
L., Laegsgaard, E., Stensgaard, I., and Besenbacher, F., Quantized conductance in atom-
sized wires between two metals. Phys. Rev. B 52, 8499 (1995).
Braun, E., Eichen, Y., Sivan, U., and Ben-Yoseph, G., DNA-templated assembly and electrode
attachment of a conducting silver wire. Nature 391, 775 (1998).
Cheng, G. S., Zhang, L. D., Zhu, Y., Fei, G. T., Li, L., Mo, C. M., and Mao, Y. Q., Ordered
nanostructure of single-crystalline GaN nanowires in a honeycomb structure of anodic
alumina. Appl. Phys. Lett. 75, 2455 (1999).
Cheng, G. S., Zhang, L. D., Chen, S. H., Li, Y., Li, L., Zhu, X. G., Zhu, Y., Fei, G. T., and
Mao, Y. Q., Ordered nanostructure of single-crystalline GaN nanowires in a honeycomb
structure of anodic alumina. J. Mater. Res. 15, 347 (2000).
Chou, S. Y., Krauss, P. R., and Kong, L. S., Nanolithograpically defined magnetic structures and
quantum magnetic disk. J. Appl. Phys. 79, 6101 (1996).
Cronin, S. B., Lin, Y.-M., Koga, T., Sun, X., Ying, J. Y., and Dresselhaus, M. S., Thermoelectric
investigation of bismuth nanowires, in “The 18th International Conference on Thermo-
electrics: ICT Symposium Proceedings” (G. Chen, Ed.), p. 554. IEEE, Piscataway, NJ,
1999.
Cronin, S. B., Lin, Y.-M., Koga, T., Ying, J. Y., and Dresselhaus, M. S., Transport measurements
of individual bismuth nanowires. MRS Symp. Proc. 582, 10.4 (2000).
Davydov, D. N., Sattari, P. A., AlMawlawi, D., Osika, A., and Haslett, T. L., Field emitters based
on porous aluminum oxide templates. J. Appl. Phys. 86, 3983 (1999).
Diggle, J. W., Downie, T. C., and Goulding, C. W., Anodic oxide films on aluminum. Chem. Rev.
69, 365 (1969).
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

200 YU-MING LIN ET AL.

Duan, X., and Lieber, C. M., General synthesis of compound semiconductor nanowires. Adv.
Mater. 12, 298 (2000).
Duan, X., Wang, J., and Lieber, C. M., Synthesis and optical properties of gallium arsenide
nanowires. Appl. Phys. Lett. 76, 1116 (2000).
Fasol, G., Nanowires: Small is beautiful. Science 280, 545 (1998).
Ferain, E., and Legras, R., Track-etched membrane-dynamics of pore formation. Nucl. Instrum.
Methods B 84, 539 (1993).
Ferre, R., Ounadjela, K., George, J. M., Piraux, L., and Dubois, S., Magnetization processes in
nickel and cobalt electrodeposited nanowires. Phys. Rev. B 56, 14066 (1997).
Foss, C. A., Jr., Tierney, M. J., and Martin, C. R., Template synthesis of infrared-transparent
metal microcylinders—comparison of optical properties with the predictions of effective
medium theory. J. Phys. Chem. 96, 9001 (1992).
Govyadinov, A. N., and Zakhvitcevich, S. A., Field emitter arrays based on natural self-
organized porous anodic alumina. J. Vac. Sci. Technol. B 16, 1222 (1998).
Han, Y.-J., Kim, J. M., and Stucky, G. D., Preparation of noble metal nanowires using hexagonal
mesoporous silica SBA-15. Chem. Mater. 12, 2068 (2000).
Heremans, J., Thrush, C. M., Zhang, Z., Sun, X., Dresselhaus, M. S., Ying, J. Y., and Morelli,
D. T., Magnetoresistance of bismuth nanowire arrays: a possible transition from 1D to 3D
localization. Phys. Rev. B 58, R10091 (1998).
Heremans, J., Thrush, C. M., Lin, Y.-M., Cronin, S. B., Zhang, Z., Dresselhaus, M. S., and
Mansfield, J. F., Bismuth nanowire arrays: Synthesis and galvanomagnetic properties. Phys.
Rev. B 61, 2921 (2000).
Heremans, J., Thrush, C. M., Lin, Y.-M., Cronin, S., and Dresselhaus, M. S., Transport properties
of antimony nanowires. Phys. Rev. B. 63, 5406 (2001).
Hicks, L. D., and Dresselhaus, M. S., Thermoelectric figure of merit of a one-dimensional
conductor. Phys. Rev. B 47, 16631 (1993).
Holmes, J. D., Johnston, K. P., Doty, R. C., and Korgel, B. A., Control of thickness and orient-
ation of solution-grown silicon nanowires. Science 287, 1471 (2000).
Hong, K., Yang, F. Y., Liu, K., Reich, D. H., Searson, P. C., and Chien, C. L., Giant positive
magnetoresistance of Bi nano wire arrays in high magnetic fields. J. Appl. Phys. 85, 6184
(1999).
Hornyak, G. L., Patrissi, C. J., and Martin, C. R., Fabrication, characterization, and optical
properties of gold nanoparticle/porous alumina composites: the nonscattering Maxwell-
Garnett. J. Phys. Chem. 101, 1548 (1997).
Huber, C. A., Huber, T. E., Sadoqi, M., Lubin, J. A., Manalis, S., and C. B., Prater, Nanowire
array composites. Science 263, 800 (1994).
Huber, T. E., Graf, M. J., and Foss, C. A., Jr., Low contact resistance 30 nm and 200 nm diameter
Bi wire array composites, in “The 18th International Conference on Thermoelectrics: ICT
Symposium Proceedings” (G. Chen, Ed.), p. 558. IEEE, Piscataway, NJ, 1999.
Huber, T. E., Graf, M. J., Foss, C. A., Jr., and Constant, P., Processing and characterization of
high-conductance bismuth wire array composites. J. Mater. Res. 15, 1816 (2000).
Iwasaki, T., Motoi, T., and Den, T., Mutiwalled carbon nanotubes growth in anodic alumina
nanoholes. Appl. Phys. Lett. 75, 2044 (1999).
Jessensky, O., Müller, F., and Gösele, U., Self-organized formation of hexagonal pore arrays in
anodic alumina. Appl. Phys. Lett. 72, 1173 (1998).
Keller, F., Hunter, M. S., and Robinson, D. L., Structural features of oxide coating on aluminum.
J. Electrochem. Soc. 100, 411 (1953).
Kouklin, N., Bandyopadhyay, S., Tereshin, S., Varfolomeev, A., and Zaretsky, D., Electronic
bistability in electrochemically self-assembled quantum dots: A potential nonvolatile ran-
dom access memory. Appl. Phys. Lett. 76, 460 (2000).
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 201

Lee, S. T., Zhang, Y. F., Wang, N., Tang, Y. H., Bello, I., Lee, C. S., and Chung, Y. W., Semicon-
ductor nanowires from oxides. J. Mater. Res. 14, 4503 (1999).
Li, A. P., Müller, F., Birner, A., Neilsch, K., and Gösele, U., Hexagonal pore arrays with a
50–420 nm interpore distance formed by self-organization in anodic alumina. J. Appl.
Phys. 84, 6023 (1998a).
Li, F., Zhang, L., and Metzger, R. M., On the growth of highly ordered pores in anodized
aluminum oxide. Chem. Mater. 10, 2470 (1998b).
Li, J., Papadopoulos, C., Xu, J. M., and Moskovits, M., Highly ordered carbon nanotube arrays
for electronic applications. Appl. Phys. Lett. 75, 367 (1999).
Li, Y., Holland, E. R., and Wilshaw, P. R., Synthesis of high density arrays of nanoscaled gridded
field emitters based on anodic alumina. J. Vac. Sci. Technol. B 18, 994 (2000).
Lieber, C. M., One-dimensional nanostructures: chemistry, physics and applications. Solid State
Commun. 107, 607 (1998).
Lin, Y.-M., Sun, X., Cronin, S. B., Zhang, Z., Ying, J. Y., and Dresselhaus, M. S., Fabrication and
transport properties of Te-doped bismuth nanowire arrays. MRS Symp. Proc. 582, 10.3
(2000a).
Lin, Y.-M., Cronin, S. B., Ying, J. Y., Dresselhaus, M. S., and Heremans, J. P., Transport properties
of Bi nanowire arrays. Appl. Phys. Lett. 76, 3944 (2000b).
Lin, Y.-M., Sun, X., and Dresselhaus, M. S., Investigation of thermoelectric transport properties
of cylindrical Bi nanowires. Phys. Rev. B 62, 4610 (2000c).
Liu, K., Chien, C. L., Searson, P. C., and Kui, Y. Z., Structural and magneto-transport properties
of electrodeposited bismuth nanowires. Appl. Phys. Lett. 73, 1436 (1998a).
Liu, K., Chien, C. L., and Searson, P. C., Finite-size effects in bismuth nanowires. Phys. Rev. B
58, 14681 (1998b).
Martin, C. R., Nanomaterials: A membrane-based synthetic approach. Science 266, 1961 (1994).
Masuda, H., Yamada, H., Satoh, M., Asoh, H., Nakao, M., and Tamamura, T., Highly
ordered nanochannel-array architecture in anodic alumina. Appl. Phys. Lett. 71, 2772
(1997).
Masuda, H., Ohya, M., Asoh, H., and Nakao, M., Photonic crystal using anodic porous alumina.
Jpn. J. Appl. Phys. 38, 1403 (1999).
Morales, A. M., and Lieber, C. M., A laser ablation method for the synthesis of crystalline
semiconductor nanowires. Science 279, 208 (1998).
Muller, C. J., van Ruitenbeek, J. M., and de Jongh, L. J., Conductance and supercurrent discon-
tinuities in atomic-scale metallic constrictions of variable width. Phys. Rev. Lett. 69, 140
(1992).
Muller, C. J., Krans, J. M., Todorov, T. N., and Reed, M. A., Quantization effects in the conduc-
tance of metallic contacts at room temperature. Phys. Rev. B 53, 1022 (1996).
Namatsu, H., Horiguchi, S., Nagase, M., and Kurihara, K., Fabrication of one-dimensional
nanowire structures utilizing crystallographic orientation in silicon and their conductance
characteristics. J. Vac. Sci. Technol. B 15, 1688 (1997).
Nötzel, R., Ledentsov, N. N., Daweeritz, L., and Ploog, K., Semiconductor quantum-wire struc-
tures directly grown on high-index surfaces. Phys. Rev. B 45, 3507 (1992).
Omi, H., and Ogino, T., Self-assembled Ge nanowires grown on Si(113). Appl. Phys. Lett. 71,
2163 (1997).
O’Sullivan, J. P., and Wood, G. C., The morphology and mechanism of formation of porous
anodic films on aluminum. Proc. Royal Soc. London A 317, 511 (1970).
Ozin, G. A., Nanochemistry: synthesis in diminishing dimensions. Adv. Mater. 4, 612 (1992).
Pascual, J. I., Mendez, J., Gómez-Herrero, J., Baró, A. M., Garcia, N., and Binh, V. T., Quantum
contact in gold nanostructures by scanning tunneling microscopy. Phys. Rev. Lett. 71, 1852
(1993).
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

202 YU-MING LIN ET AL.

Pascual, J. I., Mendez, J., Gómez-Herrero, J., Baró, A. M., Garcia, N., Landman, U., Luedtke,
W. D., Bogachek, E. N., and Cheng, H. P., Properties of metallic nanowires—from con-
ductance quantization to localization. Science 267, 1793 (1995).
Peng, Y., Zhang, H.-L., Pan, S.-L, and Li, H.-L., Magnetic properties and magnetization reversal
of α-Fe nanowires. J. Appl. Phys. 87, 7405 (2000).
Piraux, L., George, I. M., Despres, J. F., Leroy, C., Ferain, E., Legras, R., Ounadjela, K., and
Fert, A., Giant magnetoresistance in magnetic multilayered nanowires. Appl. Phys. Lett.
65, 2484 (1994).
Piraux, L., Dubois, S., Duvail, J. L., and Radulescu, A., Fabrication and properties of organic
and metal nanocylinders in nanoporous membranes. J. Mater. Res. 14, 3042 (1999).
Routkevitch, D., Bigioni, T., Moskovits, M., and Xu, J. M., Electrochemical fabrication of CdS
nanowire arrays in porous anodic aluminum oxide templates. J. Phys. Chem. 100, 14307
(1996a).
Routkevitch, D., Tager, A. A., Haruyama, J., AlMawlawi, D., Moskovits, M., and Xu, J. M.,
Nonlithographic nanowire arrays: Fabrication, physics, and device applications. IEEE
Trans. Electron. Dev. 43, 1646 (1996b).
Sugawara, A., Coyle, T., Hembree, G. G., and Scheinfein, M. R., Self-organized Fe nanowire
arrays prepared by shadow deposition on NaCl(110) template. Appl. Phys. Lett. 70, 1043
(1997).
Suh, J. S., and Lee, J. S., Highly ordered two-dimensional carbon nanotube arrays. Appl. Phys.
Lett. 75, 2047 (1999).
Sun, L., Searson, P. C., and Chien, C. L., Electrochemical deposition of nickel nanowire arrays
in single-crystal mica films. Appl. Phys. Lett. 74, 2803 (1999a).
Sun, X., Zhang, Z., and Dresselhaus, M. S., Theoretical modeling of thermoelectricity in bismuth
nanowires. Appl. Phys. Lett. 74, 4005 (1999b).
Sun, L., Searson, P. C., and Chien, C. L., Finite-size effects in nickel nanowire arrays. Phys. Rev.
B 61, R6463 (2000).
Tang, Y. H., Wang, N., Zhang, Y. F., Lee, C. S., Bello, I., and Lee, S. T., Synthesis and character-
ization of amorphous carbon nanowires. Appl. Phys. Lett. 75, 2921 (1999).
Tonucci, R. J., Justus, B. J., Campillo, A. J., and Ford, C. E., Nanochannel array glass. Science
258, 783 (1992).
van Wees, B. J., van Houten, H., Beenakker, C. W. J., Williamson, J. G., Kouwenhoven, L. P.,
van der Marel, D., and Foxon, C. T., Quantized conductance of point contacts in a two-
dimensional electron gas. Phys. Rev. Lett. 60, 848 (1988).
Wagner, R. S., and Ellis, W. C., Nanowire formation in a polymeric film. Appl. Phys. Lett. 4, 89
(1964).
Wang, N., Tang, Y. H., Zhang, Y. F., Lee, C. S., and Lee, S. T., Nucleation and growth of Si
nanowires from silicon oxide. Phys. Rev. B 58, R16024 (1998a).
Wang, N., Zhang, Y. F., Tang, Y. H., Lee, C. S., and Lee, S. T., SiO2-enhanced synthesis of Si
nanowires by laser ablation. Appl. Phys. Lett. 73, 3902 (1998b).
Wharam, D. A., Thornton, T. J., Newbury, R., Pepper, M., Ahmed, H., Frost, J. E. F., Hasko,
D. G., Peacock, D. C., Ritchie, D. A., and Jones, G. A. C., One-dimensional transport and
the quantization of the ballistic resistance. J. Phys. C 21, L209 (1988).
Whitney, T. M., Jiang, J. S., Searson, P. C., and Chien, C. L., Fabrication and magnetic-properties
of arrays of metallic nanowires. Science 261, 1316 (1993).
Wu, C.-G., and Bein, T., Conducting polyaniline filaments in a mesoporous channel host. Science
264, 1757 (1994).
Xu, D., Chen, D., Xu, Y., Shi, X., Guo, G., Gui, L., and Tang, Y., Preparation of II–VI group
semiconductor nanowire arrays by dc electrochemical deposition in porous aluminum
oxide templates. Pure Appl. Chem. 72, 127 (2000a).
08/18/2001 02:18 PM Chemical Engineering-v27 PS068-05.tex PS068-05.xml APserialsv2(2000/12/19) Textures 2.0

FABRICATION, STRUCTURE, AND TRANSPORT PROPERTIES 203

Xu, D., Xu, Y., Chen, D., Guo, G., Gui, L., and Tang, Y., Preparation of CdS single-crystal
nanowires by electrochemically induced deposition. Adv. Mater. 12, 520 (2000b).
Yi, G., and Schwarzacher, W., Single crystal superconductor nanowires by electrodeposition.
Appl. Phys. Lett. 74, 1746 (1999).
Ying, J. Y., Nanoporous systems and templates: The unique self-assembly and synthesis of
nanostructures. Science Spectra 18, 56 (1999).
Ying, J. Y., Mehnert, C. P., and Wong, M. S., Synthesis and applications of supramolecular-
templated mesoporous materials. Angew. Chem. Int. Ed. 38, 56 (1999).
Yu, D. P., Lee, C. S., Bello, I., Zhou, G. W., and Bai, Z. G., Synthesis of nano-scale silicon wires
by excimer laser ablation at high temperature. Solid State Commun. 105, 403 (1998).
Zeng, H., Zheng, M., Skomski, R., Sellmyer, D. J., Liu, Y., Menon, L., and Bandyopadhyay,
S., Magnetic properties of self-assembled Co nanowires of varying length and diameter.
J. Appl. Phys. 87, 4718 (2000).
Zhang, Z., Ying, J. Y., and Dresselhaus, M. S., Bismuth quantum-wire arrays fabricated by a
vacuum melting and pressure injection process. J. Mater. Res. 13, 1745 (1998a).
Zhang, Y. F., Tang, Y. H., Wang, N., Yu, D. P., Lee, C. S., Bello, I., and Lee, S. T., Silicon nanowires
prepared by laser ablation at high temperature. Appl. Phys. Lett. 72, 1835 (1998b).
Zhang, Z., Sun, X., Dresselhaus, M. S., Ying, J. Y., and Heremans, J., Magnetotransport inves-
tigation of ultrafine single-crystalline bismuth nanowire arrays. Appl. Phys. Lett. 73, 1589
(1998c).
Zhang, Z., Gekhtman, D., Dresselhaus, M. S., and Ying, J. Y., Processing and characterization
of single-crystalline ultrafine bismuth nanowires. Chem. Mater. 11, 1659 (1999).
Zhang, Y. F., Tang, Y. H., Wang, N., Lee, C. S., Bello, I., and Lee, S. T., Synthesis and character-
ization of amorphous carbon nanowires. Phys. Rev. B 61, 4518 (2000a).
Zhang, Z., Sun, X., Dresselhaus, M. S., Ying, J. Y., and Heremans, J., Electronic transport
properties of single crystal bismuth nanowire arrays. Phys. Rev. B 61, 4850 (2000b).