Magnetic-field induced superconductor-metal-insulator transitions in bismuth metal-graphite
Extracted source text ยท arXiv cond-mat/0308312
%% ****** Start of file template.aps ****** %
%%
%%
%% This file is part of the APS files in the REVTeX 4 distribution.
%% Version 4.0 of REVTeX, August 2001
%%
%%
%% Copyright (c) 2001 The American Physical Society.
%%
%% See the REVTeX 4 README file for restrictions and more
%% information.
%%
%
% This is a template for producing manuscripts for use with REVTEX 4.0
% Copy this file to another name and then work on that file.
% That way, you always have this original template file to use.
%
% Group addresses by affiliation; use superscriptaddress for long
% author lists, or if there are many overlapping affiliations.
% For Phys. Rev. appearance, change preprint to twocolumn.
% Choose pra, prb, prc, prd, pre, prl, prstab, or rmp for journal
% Add 'draft' option to mark overfull boxes with black boxes
% Add 'showpacs' option to make PACS codes appear
% Add 'showkeys' option to make keywords appear
%\documentclass[aps,prl,preprint,groupedaddress]{revtex4}
%\documentclass[aps,prl,preprint,superscriptaddress]{revtex4}
%\documentclass[aps,prl,twocolumn,groupedaddress]{revtex4}
\documentclass[aps,prb,twocolumn,showpacs,groupedaddress]{revtex4}
\usepackage{graphicx}
% You should use BibTeX and apsrev.bst for references
% Choosing a journal automatically selects the correct APS
% BibTeX style file (bst file), so only uncomment the line
% below if necessary.
\bibliographystyle{apsrev}
\begin{document}
% Use the \preprint command to place your local institutional report
% number in the upper righthand corner of the title page in preprint
% mode.
% Multiple \preprint commands are allowed.
% Use the 'preprintnumbers' class option to override journal defaults
% to display numbers if necessary
%\preprint{}
%Title of paper
\title{ Magnetic-field induced superconductor-metal-insulator
transitions in bismuth metal-graphite }
% repeat the \author .. \affiliation etc. as needed
% \email, \thanks, \homepage, \altaffiliation all apply to the current
% author. Explanatory text should go in the []'s, actual e-mail
% address or url should go in the {}'s for \email and \homepage.
% Please use the appropriate macro foreach each type of information
% \affiliation command applies to all authors since the last
% \affiliation command. The \affiliation command should follow the
% other information
% \affiliation can be followed by \email, \homepage, \thanks as well.
\author{Masatsugu Suzuki }
\email[]{suzuki@binghamton.edu}
%\homepage[]{Your web page}
%\thanks{}
%\altaffiliation{}
\affiliation{ Department of Physics, State University of New York at
Binghamton,
Binghamton, New York 13902-6016}
\author{Itsuko S. Suzuki }
%\email[]{itsuko@binghamton.edu}
%\homepage[]{Your web page}
%\thanks{}
%\altaffiliation{}
\affiliation{ Department of Physics, State University of New York at
Binghamton,
Binghamton, New York 13902-6016}
\author{Robert Lee}
%\email[]{}
%\homepage[]{Your web page}
%\thanks{}
%\altaffiliation{}
\affiliation{ Department of Physics, State University of New York at
Binghamton,
Binghamton, New York 13902-6016}
\author{J\"{u}rgen Walter }
%\email[]{}
%\homepage[]{Your web page}
%\thanks{}
%\altaffiliation{}
\affiliation{Department of Materials Science and Processing,
Graduated School of Engineering, Osaka University, 2-1, Yamada-oka,
Suita, 565-0879, JAPAN}
%Collaboration name if desired (requires use of superscriptaddress
%option in \documentclass). \noaffiliation is required (may also be
%used with the \author command).
%\collaboration can be followed by \email, \homepage, \thanks as well.
%\collaboration{}
%\noaffiliation
\date{\today}
\begin{abstract}
% insert abstract here
Bismuth-metal graphite (MG) has a unique layered
structure where Bi nanoparticles are encapsulated between adjacent
sheets of nanographites. The superconductivity below $T_{c}$
(= 2.48 K) is due to Bi nanoparticles. The Curie-like susceptibility
below 30 K is due to conduction electrons localized near zigzag
edges of nanographites.
A magnetic-field induced transition from metallic to
semiconductor-like phase is observed in the in-plane resistivity
$\rho_{a}$ around $H_{c}$ ($\approx$ 25 kOe) for both $H$$\perp$$c$
and $H$$\parallel$$c$ ($c$: $c$ axis). A negative magnetoresistance
in $\rho_{a}$ for $H$$\perp$$c$ (0$<H\leq$3.5 kOe) and a logarithmic
divergence in $\rho_{a}$ with decreasing temperature for
$H$$\parallel$$c$ ($H$ $>$ 40 kOe) suggest the occurrence of
two-dimensional weak localization effect.
\end{abstract}
% insert suggested PACS numbers in braces on next line
\pacs{71.24.+q, 74.80.Dm, 72.15.Rn, 71.30.+h}
% insert suggested keywords - APS authors don't need to do this
%\keywords{}
%\maketitle must follow title, authors, abstract, \pacs, and \keywords
\maketitle
% body of paper here - Use proper section commands
% References should be done using the \cite, \ref, and \label commands
%\section{}
% Put \label in argument of \section for cross-referencing
%\section{\label{}}
%\subsection{}
%\subsubsection{}
% If in two-column mode, this environment will change to single-column
% format so that long equations can be displayed. Use
% sparingly.
%\begin{widetext}
% put long equation here
%\end{widetext}
\section{\label{intro}Introduction}
A weak localization theory predicts a logarithmic divergence
of the resistivity in the two-dimensional (2D) electron systems
as the temperature ($T$) is lowered.\cite{Abrahams1979} In high-mobility Si
metal oxide-semiconductor field-effect transistor (MOSFET),
the in-plane resistivity for a system with an electron density $n$
larger than a critical electron density $n_{c}$ decreases with
decreasing $T$, indicating a metallic behavior.
\cite{Kravchenko1995,Simonian1997,Kravchenko2000} This metallic
state is completely destroyed by the application of an external
magnetic field ($H$) applied in the basal plane when $H$
is higher than a threshold field $H_{c}$. Such coplanar fields
only polarize the spins of the electrons, indicating that the
spin state is significant to the high conductivity of the metallic
state. The scaling relation of the in-plane resistivity collapses
into two distinct branches above and below $H_{c}$. Such behaviors
are very similar to those observed in amorphous ultrathin metal
films of InO$_{x}$,\cite{Hebard1990} MoGe,\cite{Yazdani1995,Mason1999}
and Bi,\cite{Markovic1998} which undergo magnetic
field-induced transitions from superconducting phase to insulating
phase.
Bi-metal graphite (MG) constitutes a novel class of materials
having unique layered structures. This system can be prepared
from the reduction by Li-diphenylide from an acceptor-type BiCl$_{3}$
graphite intercalation compound (GIC) as a precursor material.
Ideally, the staging structure of Bi-MG would be the same as
that of BiCl$_{3}$ GIC.\cite{Walter1998,Walter1999a,McRae2001}
In Bi-MG, Bi atoms would form intercalate
layers sandwiched between adjacent graphite layers. For the stage-$n$
(= 1, 2,$\cdots$) structure, the packages of $n$ graphite layers
and one Bi layer would be periodically stacked along the $c$
axis perpendicular to the graphite basal plane. In BiCl$_{3}$ GIC,
the ratio of $c$-axis resistivity $\rho_{c}$ to the in-plane
resistivity $\rho_{a}$ ($\Delta = \rho_{c}/\rho_{a}$)
at room temperature is (7.4 - 4.9) $\times$ 10$^{3}$,\cite{McRae2001}
suggesting that
virtually all $\pi$-electrons are directionally localized;
i.e., they can move freely along the graphite basal planes, but
are unable to diffuse across the stack of layers.\cite{Suzuki1996} Such a
large ratio $\Delta$ is due to the insulating BiCl$_{3}$ layer
sandwiched between the adjacent graphite layers. There is no
overlapping of the wave functions over nearest neighbor graphite
layers. The situation may not drastically change in Bi-MG in
spite of the fact that BiCl$_{3}$ layer is replaced by metallic
Bi layer. In fact the $c$-axis resistivity of Bi-MG (= 0.1 $\Omega$cm
at 298 K) is almost the same as that of BiCl$_{3}$ GIC at the same $T$,
suggesting that Bi-MG behaves like a quasi 2D conductor.
In this paper we have undertaken an extensive study on the transport
and magnetic properties of Bi-MG. We show that this compound
undergoes a superconducting transition at $T_{c}$ = 2.48 K. A
magnetic-field induced transition from metallic to semiconductor-like
phase is observed in $\rho_{a}$
around $H$ $\approx$ 25 kOe for $H$$\perp$$c$ and $H$$\parallel$$c$
($c$: $c$ axis). These results of Bi-MG are compared with
those of Si MOSFET, and amorphous ultrathin metal films of
InO$_{x}$,\cite{Hebard1990}
MoGe,\cite{Yazdani1995,Mason1999} and Bi.\cite{Markovic1998}
Structural studies of Bi-MG reveal that Bi layers are formed
of Bi nanoparticles which are encapsulated between adjacent sheets
of nanographites. The size of nanographites in Bi-MG is much
smaller than the in-plane coherence size of the graphite layers
of pristine graphite. The superconductivity is mainly due to
Bi nanoparticles, while the magnetism is mainly due to nanographites.
Nanographites are nanometer-sized graphite fragments which represent
a new class of mesoscopic system intermediate between aromatic
molecules and extended graphene sheets. Fujita and
co-workers\cite{Fujita1996,Wakabayashi1998,Wakabayashi1999}
have theoretically suggested that the electronic structures of
finite-size graphene sheets depend crucially on the shape of
their edges. The graphene edge of an arbitrary shape consists
of two-types of edges, zigzag type and armchair type. The former
has a trans-polyacetylene type structure and the latter has a
cis-polyacetylene one. Finite graphite systems having zigzag
edges exhibit a special edge state. The corresponding energy
bands are almost flat at the Fermi energy, thereby giving a sharp
peak in the density of states (DOS) at the Fermi energy. This
is in contrast to the case of 2D graphene sheet with infinite
size, where the DOS is zero at the Fermi energy. According to
Wakabayashi et al.,\cite{Wakabayashi1999} the magnetism of nanographites
is characterized
by Pauli paramagnetism and orbital diamagnetism from conduction
electrons. They have predicted that the Pauli paramagnetic susceptibility
for the nanographites with zigzag edges shows a Curie-like behavior
at low temperatures, which is in contrast to a $T$-independent
Pauli paramagnetism in normal metals. In this paper we show that
the susceptibility of Bi-MG has a Curie-like behavior at low
T. This behavior is discussed in the light of the prediction
by Wakabayashi et al.\cite{Wakabayashi1999}
\section{\label{exp}Experimental procedure}
BiCl$_{3}$ GIC samples as a precursor material, were prepared by
heating a mixture of highly oriented pyrolytic graphite (HOPG)
[grade ZYA from Advanced Ceramics, Ohio] and an excess amount
of BiCl$_{3}$ at 200 $^\circ$C in a ampoule filled with chlorine
gas at a pressure of 375 Torr.\cite{Walter1998,Walter1999a}
The reaction was continued
for three days. It was confirmed from (00$L$) x ray diffraction
(Rigaku RINT 2000 x-ray powder diffractometer) that the BiCl$_{3}$
GIC sample consists of stage-2 as majority phase and stage-3
and stage-4 as minority phase. The $c$ axis repeat distance is
13.17 $\pm$ 0.05 $\AA$ for stage-2, 15.85 $\pm$ 0.25 $\AA$
for stage-3, and 20.22 $\pm$ 0.25 $\AA$ for stage-4, respectively.
No Bragg reflection from the pristine graphite is observed.
The synthesis of Bi-MG was made by the reduction by Li-diphenylide
from BiCl$_{3}$ GIC. BiCl$_{3}$ GIC samples were kept for three
days in a solution of lithium diphenylide in tetrahydrofuran
(THF) at room temperature. Then the samples were filtered, rinsed
by THF, and dried in air. Finally the samples were annealed at
260 $^\circ$C in a hydrogen gas atmosphere for one day. The
structure of Bi-MG thus obtained was examined by (00$L$) x-ray
diffraction, and bright field images and selected-area electron
diffraction (SAED) (Hitachi H-800 transmission electron microscope)
operated at 200 kV. The same methods for the structural analysis
were used for Pd-MG.\cite{Walter1999b,Walter2000a,Walter2000b}
The (00$L$) x-ray diffraction pattern
of Bi-MG is much more complicated than that of BiCl$_{3}$ GIC, which
makes it difficult to calculate the average particle thickness from
the identity period in Bi-MG.
Note that graphite reflections appear in Bi-MG, suggesting that
a part of Bi atoms leaves from the graphite galleries occupied
by BiCl$_{3}$ intercalate layers in BiCl$_{3}$ GIC during the reduction
process. Such Bi atoms tend to form multilayered Bi nanoparticles
in the graphite galleries in Bi-MG. The number of Bi layers in
possible multilayered structures could not be exactly determined
at present.
SAED pattern of Bi-MG consists of polycrystalline diffraction rings,
suggesting that there are at least four Bi layers in thickness.
Reflections from Bi and graphite were observed. As listed in Table
\ref{table:one}, all
the Bi reflections were labeled and attributed to the formation of
rhombohedral Bi, according to the standard ICDD PDF (Card No. 05-0519).
This result indicates that Bi nanoparticles are crystallized as
rhombohedral Bi phase in Bi-MG. The observed spacings of Bi-MG are
1-2 \%
shorter than those of bulk Bi metal.
The size of Bi
nanoparticles distribute widely around the average size 110 $\AA$.
More than 50 \% of Bi nanoparticles has sizes ranging between
10 and 50 $\AA$. The largest particle size is 750 $\AA$.
\begin{table}
\caption{\label{table:one} Experimentally observed spacings $d_{exp}$ for Bi-MG
and $d_{PDF}$ given by powder diffraction file (PDF) 5-0519 for rhombohedral Bi
[spacegroup: R-3m ($a_{0}$ = 4.546 $\AA$, $c_{0}$ = 11.860 $\AA$ for the notation of the
hexagonal close-packed structure)]. All reflections can be indexed as
($hkl$) reflections of Bi. Graphite reflections are not included.}
\begin{ruledtabular}
\begin{tabular}{lll}
$d_{exp}$($\AA$) & $d_{PDF}$($\AA$) & ($hkl$)\\
3.24 & 3.28 & 102\\
2.35 & 2.39 & 014\\
1.92 & 1.970 & 113\\
1.71 & 1.868 & 022\\
1.51 & 1.515 & 025\\
1.40 & 1.443 & 212\\
\end{tabular}
\end{ruledtabular}
\end{table}
The measurements of DC magnetization and resistivity were made
using a SQUID magnetometer (Quantum Design MPMS XL-5) with an
ultra low field capability and an external device control mode.
In the present work we used a Bi-MG sample based on HOPG which
is partially exfoliated. The stoichiometry of C and Bi was not
determined. The in-plane resistivity $\rho_{a}$ and the $c$
axis resistivity $\rho_{c}$ were measured by a conventional
four-probe method. The sample had a rectangular form with a base
6.0 $\times$ 1.6 mm$^{2}$ and a height 0.47 mm along the $c$ axis. For the
measurement of $\rho_{a}$, four thin gold wires (25-$\mu$m
diameter) that were used as as the current and voltage probes
were attached to one side of the $c$ surfaces by silver paste
(4922N, du Pont). For the measurement of $\rho_{c}$, two
thin gold wires as the current and voltage probes were attached
to each $c$-surface of the sample. The current ($I$ = 10
mA for $\rho_{a}$ and 3 mA for $\rho_{c}$) was supplied
through the current probes by a Keithley type 224, programmable
DC current source. The voltage $V$ generated across the voltage
probes was measured by a Keithley 182 nanovoltmeter. The linearity
of $I$-$V$ characteristics was confirmed for the measurements
of $\rho_{a}$ and $\rho_{c}$.
\section{\label{result}Result and discussion}
\subsection{\label{resultA}Meissner effect due to Bi nanoparticles}
\begin{figure}
\includegraphics[width=8.0cm]{BiMGfig1.eps}%
\caption{\label{fig:one} $T$ dependence of ZFC and FC
susceptibilities of Bi-MG based on HOPG. $H$ = 100 Oe.
$H$$\perp$$c$ ($c$: $c$ axis). The inset shows the magnetic
phase diagram of $H$ vs $T$ for
$H$$\perp$$c$, where the peak temperatures
of ZFC and FC susceptibilities are plotted as a function of
$H$. The solid line denotes the least-squares fit of the ZFC
data ({\Large $\bullet$}) to Eq.(\ref{eq:one}).}
\end{figure}
\begin{figure}
\includegraphics[width=8.0cm]{BiMGfig2.eps}%
\caption{\label{fig:two} $T$ dependence of ZFC and FC
susceptibilities of Bi-MG. $H$ = 100 Oe.
$H$$\parallel$$c$ ($c$: $c$ axis). The inset shows the magnetic phase diagram of
$H$ vs $T$ for $H$$\parallel$$c$. The solid line
denotes the least-squares fitting curve for the ZFC data ({\Large $\bullet$})
to Eq.(\ref{eq:one}).}
\end{figure}
We measured the DC magnetization of Bi-MG. After the sample was
cooled from 298 to 1.9 K at $H$ = 0, the measurement of zero
field cooled (ZFC) magnetization $M_{ZFC}$ was made with increasing $T$
from 1.9 to 15 K in the presence of $H$. The sample was kept
for 20 minutes at 50 K. Then the sample was cooled from 50
to 15 K. The measurement of field-cooled (FC) magnetization $M_{FC}$
was made with decreasing $T$ from 15 to 1.9 K in the presence
of the same $H$. For convenience, hereafter, the direction
of $H$ in the $c$ plane is denoted as $H$$\perp$$c$ ($c$: $c$ axis).
Figures \ref{fig:one} and \ref{fig:two} show typical examples of the $T$ dependence
of the susceptibility $\chi_{ZFC}$ (=$M_{ZFC}/H$) and $\chi_{FC}$
(=$M_{FC}/H$) at $H$ = 100 Oe for $H$$\perp$$c$ and $H$$\parallel$$c$,
respectively. Both $\chi_{FC}$ and $\chi_{ZFC}$ show
a sharp peak around 2.5 K at $H$ = 20 Oe, which results from
the competition between the diamagnetic susceptibility due to
the Meissner effect and the Curie-like susceptibility (which
will be described later). The peak shifts to the low $T$ side
with increasing $H$. No peak is observed above 1.9 K for $H\geq$ 250
Oe. For convenience, the peak temperature is defined as a critical
temperature $T_{c}(H)$. The peak temperature $T_{c}(H)$
for $\chi_{FC}$ is almost the same as that for $\chi_{ZFC}$.
In the insets of Figs. \ref{fig:one} and \ref{fig:two}, we show the $H$-$T$ magnetic
phase diagram for Bi-MG for $H$$\perp$$c$ and $H$$\parallel$$c$,
respectively. We find that the data of $H$ vs $T$ (ZFC susceptibility)
for $H$$\perp$$c$ and $H$$\parallel$$c$ are well described by
\begin{equation}
H=H_{0}(\frac{T_{c} -T}{T_{c} } )^{\beta},
\label{eq:one}
\end{equation}
where $H_{0}$ is the critical field at $T$ = 0 K ($H_{0}$ = $H_{c2}$ for
$H$$\perp$$c$ and $H_{0}$ = $H_{c3}$ for $H$$\parallel$$c$), $\beta$ is an
exponent, and $T_{c}$ is a critical temperature at $H$ = 0. The field
$H_{c3}$ is a field to nucleate a small superconducting region near the
sample surface. The least squares fits of the data of $H$ vs $T$ (ZFC
susceptibility) yield the values of $H_{c2}$ = 489.4 $\pm$ 0.5 Oe, $\beta$
= 0.48 $\pm$ 0.02, $T_{c}$ = 2.48 $\pm$ 0.06 K for $H$$\perp$$c$, and
$H_{c3}$ = 838 $\pm$ 1 Oe, $\beta$ = 0.69 $\pm$ 0.02, $T_{c}$ = 2.48 $\pm$
0.02 K for $H$$\parallel$$c$.
Here we note that the superconductivity is observed in ultrathin films of
amorphous Bi grown in top of layer of amorphous Ge (thickness, 6 - 10
$\AA$). Haviland et al.\cite{Haviland1989} have reported that the
superconductivity occurs for the thickness of Bi films, $d$ = 6.73 - 74.27
$\AA$. The critical temperature $T_{c}$ decreases with decreasing $d$:
$T_{c}$ = 5.6 K for $d$ = 74.27 $\AA$ and $T_{c}$ $\approx$ 0.8 K for $d$
$\approx$ 6.73 $\AA$. Markovic et al.\cite{Markovic1998} have shown that
$T_{c}$ decreases with decreasing $d$ for $d>d_{c}$ ($d_{c}$ = 12.2 $\AA$):
$T_{c}$ = 0.5 K for $d$ = 15 $\AA$. Weitzel and Micklitz\cite{Weitzel}
have reported surface superconductivity in granular films built from
well-defined rhombohedral Bi clusters (mean size = 38 $\AA$) embedded in
difference matrices (Kr, Xe, Ge) or with H$_{2}$ or O$_{2}$ gas adsorbed on
the cluster surface. The critical temperature $T_{c}$, which is dependent
on matrices and gases used, is between 2 and 6 K. This value of $T_{c}$ is
on the same order as that of Bi-MG.
The exponent $\beta$ for $H$$\perp$$c$
is very close to that ($\beta$ = 0.5) predicted for homogeneous
system of isolated superconducting grains.\cite{Deutscher1980}
The ratio $H_{c3}$/$H_{c2}$
is calculated as 1.71, which is very close to the predicted value
1.695.\cite{SaintJames1963} Similar behavior is observed in the critical fields
for $H$$\perp$$c$ and $H$$\parallel$$c$ in stage-1 K GIC
(KC$_{8}$).\cite{Tanuma1992}
The coherence length $\xi$ is estimated as $\xi$
= 820 $\AA$ from the value of $H_{c2}$ using the relation $H_{c2}$
= $\Phi_{0}$/(2$\pi \xi ^{2}$), where $\Phi_{0}$
(= 2.0678 $\times$ 10$^{-7}$ Gauss cm$^{2}$) is a fluxoid.
The coherence length $\xi$
is much larger than the size of islands (110 $\AA$ in average).
\begin{figure}
\includegraphics[width=8.0cm]{BiMGfig3.eps}%
\caption{\label{fig:three} Hysteresis loop of DC magnetization for
Bi-MG. $H$$\perp$$c$. $T$ = 1.9 K.
The measurement was made with increasing $H$ from 0 to 500 Oe
(denoted by {\Large $\bullet$}), with decreasing $H$ from
500 Oe to -500 Oe ({\Large $\circ$}), and with increasing $H$
from -500 Oe to 500 Oe ($\triangle$).}
\end{figure}
\begin{figure}
\includegraphics[width=8.0cm]{BiMGfig4.eps}%
\caption{\label{fig:four} Hysteresis loop of DC magnetization minus a
diamagnetic background (= $\chi$$_{d}$$H$)
for Bi-MG. $H$$\perp$$c$. (a) $T$ =
1.9 K. $\chi$$_{d}$ = -5.143 $\times$ 10$^{-7}$ emu/gOe.
(b) $T$ = 3.3 K. $\chi$$_{d}$ = -7.317 $\times$
10$^{-7}$ emu/gOe.}
\end{figure}
Figure \ref{fig:three} shows the hysteresis loop of the
magnetization $M_{a}(H)$
for $H$$\perp$$c$ at $T$ = 1.9 K. The sample was cooled
from 298 K to 1.9 K at $H$ = 0. The magnetization $M_{a}(H)$
was measured with increasing $H$ from 0 to 500 Oe, with decreasing $H$
from 500 Oe to -500 Oe, and with increasing $H$ from -500 Oe
to 500 Oe. The magnetization consists of superconductivity contribution
and diamagnetic background. In Fig. \ref{fig:four}(a) we show the hysteresis
loop of $\Delta M_{a}(H)$ that is defined by $M_{a}(H)$
at 1.9 K minus a diamagnetic background given by $\chi_{d}H$
with $\chi_{d}$ = -5.143 $\times$ 10$^{-7}$ emu/gOe. A hysteresis
loop characteristic to a type-II superconductor is observed with
a lower critical field $H_{c1}$ (= 80 Oe). In contrast, the magnetization
hysteresis loop$M_{a}(H)$ at $T$ = 3.3 K is very different
from that at 1.9 K. It seems that there is neither local minimum
nor local maximum. In Fig. \ref{fig:four}(b) we show the hysteresis
loop of $\Delta M_{a}(H)$
which is defined by at $M_{a}(H)$ at 3.3 K minus a diamagnetic
background given by $\chi_{d}H$ with $\chi_{d}$
= -7.317 $\times$ 10$^{-7}$ emu/gOe. As $H$ decreases, a trapped magnetic
flux, corresponding to a paramagnetic moment $M_{r}$ ($\approx$
2.7 $\times$ 10$^{-5}$ emu/g) remains in the sample. With further cycling
of $H$ from -500 to 500 Oe a characteristic hysteresis loop
is observed.
We also measured the hysteresis loop of the magnetization $M_{c}(H)$
for $H$$\parallel$$c$ at $T$ = 1.9 K. The difference $\Delta M_{c}$
is defined by $M_{c}(H)$ at 1.9 K minus a diamagnetic background
given by $\chi_{d}H$ with $\chi_{d}$ = -2.980 $\times$
10$^{-6}$ emu/gOe. The hysteresis loop of $\Delta M_{c}$
for $H$$\parallel$$c$ is very similar to the corresponding data $\Delta M_{a}$
for $H$$\perp$$c$. The value of $H_{c1}$ (= 75 Oe) for $H$$\parallel$$c$
is a little lower than that for $H$$\perp$$c$. The peak
value of $\Delta M_{c}$ is almost the same as that of $\Delta M_{a}$.
The hysteresis loop of $\Delta M_{c}$ at 1.9 K exhibits
a nearly reversible behavior. In bulk samples this could be explained
in terms of a lack of structural defects to provide pinning sites
in the vortex state ($H_{c1} < H < H_{c2}$). In Bi-MG,
however, the size of islands is much shorter than the coherence
length $\xi$. Thus the pinning effect is not relevant
for these islands. In contrast, the hysteresis loop of $\Delta M_{a}$
at 1.9 K exhibits some hysteretic behavior. Note that similar
magnetization curves are observed in a superconductor TaC nanoparticles
which are synthesized using a vapor-solid reaction path starting
with carbon nanotube precursor.\cite{Fukunaga1998}
\subsection{\label{resultB}Curie-like susceptibility due to nanographites}
\begin{figure}
\includegraphics[width=8.0cm]{BiMGfig5.eps}
\caption{\label{fig:five} $T$ dependence of FC
susceptibilities for Bi-MG. $H$ = 1 kOe.
$H$$\perp$$c$ and $H$$\parallel$$c$. The inset
shows the reciprocal susceptibility ($\chi-\chi_{0}$)$^{-1}$ as a function of $T$.
The solid lines are described by Eq.(\ref{eq:two}) with parameters
given in the text. }
\end{figure}
\begin{figure}
\includegraphics[width=8.0cm]{BiMG2fig6.eps}
\caption{\label{fig:six}$T$ depedence of $c$-axis resistivity $\rho_{c}$($T$,$H$)
for Bi-MG at various $H$. $H$$\parallel$$c$. $I$$\parallel$$H$.}
\end{figure}
Figure \ref{fig:five} shows typical data of $\chi_{FC}$ for $H$ = 1
kOe for $H$$\perp$$c$ and $H$$\parallel$$c$. The susceptibility
is negative at high $T$ and drastically increases with decreasing $T$.
The least squares fit of the data ($\chi_{FC}$ vs $T$)
for 1.9 $\leq T \leq$ 30 K to the Curie-Weiss law
\begin{equation}
\chi =\chi _{0} +\frac{C}{T-\Theta },
\label{eq:two}
\end{equation}
yields $\Theta$ = -0.86 $\pm$ 0.05 K, $C$ = (1.999
$\pm$ 0.029) $\times$ 10$^{-6}$ emu K/g, $\chi_{0}$ = (-1.307 $\pm$
0.001) $\times$ 10$^{-6}$ emu/g for $H$$\perp$$c$, and $\Theta$
= -1.13 $\pm$ 0.02 K, $C$ = (1.460 $\pm$ 0.008) $\times$ 10$^{-6}$
emu K/g, $\chi_{0}$ = (-1.146 $\pm$ 0.001) $\times$ 10$^{-6}$ emu/g
for $H$$\parallel$$c$. The value of $\Theta$ is very close to
zero, showing a Curie-like law. In the inset of Fig. \ref{fig:five} we show
the reciprocal susceptibility ($\chi -\chi_{0})^{-1}$
as a function of $T$. The negative value $\chi_{0}$ is
from the orbital diamagnetic susceptibility. There is a crossover
from the high-temperature diamagnetic susceptibility to the low-temperature
Curie-like susceptibility around 50 K. We assume that the susceptibility
of Bi-MG at 100 K corresponds to the diamagnetic susceptibility
since the Pauli susceptibility is positive and is nearly equal
to zero at 100 K. From Fig. \ref{fig:five} the diamagnetic susceptibility for
$H$$\parallel$$c$
and $H$$\perp$$c$ is estimated as $\chi_{c} \approx$
-1.1 $\times$ 10$^{-6}$ emu/g and $\chi_{a} \approx$ - 1.3 $\times$ 10$^{-6}$
emu/g, which are almost isotropic. These values are in contrast
with the susceptibility of HOPG, which is very anisotropic: $\chi_{c}$
= -25.86 $\times$ 10$^{-6}$ emu/g and $\chi_{a}$ = -1.06 $\times$ 10$^{-7}$ emu/g
at $H$ = 1 kOe. The absolute value of $\chi_{c}$ in Bi-MG
is much smaller than that in HOPG, while the values of $\chi_{a}$
are on the same order for both systems.
Here it is interesting to compare our data of susceptibility
with that of nanographites prepared by heat treating diamond
nanoparticles\cite{Anderssen1998} and a disorder network of nanographites in
activated carbon fiber (ACF).\cite{Shibayama2000} The $T$ dependence of the
susceptibility for these compounds is similar to that of Bi-MG.
In particular, the values of $\Theta$, $C$, and $\chi_{0}$
for the ACF are on the same order as those for Bi-MG, where $\Theta$
$\approx$ -2 K, $C$ = 1.21 $\times$ 10$^{-6}$ emu K/g, and $\chi_{0}$
= -1.36 $\times$ 10$^{-6}$ emu/g for ACF prepared at the heat treatment temperature
HTT = 1500 $^\circ$C.\cite{Shibayama2000} Here we assume that the
spin $S$
is 1/2 and the Land\'{e} $g$-factor is 2 since the spins are
associated with carbon materials. The number of spin density
per 1 g of Bi-MG is estimated as (4.7 - 6.4) $\times$ 10$^{18}$/g, which
is comparable with 3.9 $\times$ 10$^{18}$/g for the ACF with HTT = 1500
$^\circ$C.\cite{Shibayama2000}
It has been theoretically predicted by
Wakabayashi et al.\cite{Wakabayashi1999}
that the Pauli susceptibility exhibits a Curie-like behavior
in the nanographites with zigzag edges, because of the DOS which
is sharply peaked at the Fermi energy. However, qualitatively
we think that the Curie-like behavior at low T in Bi-MG is due
to the conduction electrons localized around the zigzag edges
of nanographites, which have local magnetic moments (spin $S$
= 1/2 and a Land\'{e} $g$-factor $g$ = 2). The origin of spin
polarization in nanographites with zigzag edges has been discussed
by Fujita et al.\cite{Fujita1996} using the Hamiltonian that consists of the
on-site Coulomb repulsive interaction ($U$) when the site is
occupied by two electrons, and the electron transfer integral
between the nearest sites ($t$). They have shown that a ferrimagnetic
spin polarization emerges on the edge carbons even for weak $U/t$
$\approx$ 0.1.
The orbital diamagnetic susceptibility is very sensitive to the
size and edge shapes of nanographites. The orbital diamagnetic
susceptibility is almost isotropic in Bi-MG, while it is very
anisotropic in pristine graphite with infinite size. In Bi-MG
the orbital cyclotron motion of electrons in the presence of $H$
($H$$\parallel$$c$) is greatly suppressed. This result is in good agreement
with the prediction by Wakabayashi et al.\cite{Wakabayashi1999} In ribbon-shaped
nanographites with zigzag edges, the magnitude of the diamagnetic
susceptibility decreases as the ribbon width decreases. The flow
of the orbital diamagnetic ring current significantly depends
on the lattice topology near the graphite edge.
\subsection{\label{resultC}$c$-axis and in-plane resistivities}
\begin{figure}
\includegraphics[width=8.0cm]{BiMGfig6.eps}%
\caption{\label{fig:seven} $T$ dependence of in-plane
resistivity $\rho$$_{a}$($T$,$H$)
for Bi-MG at $H$ = 0, 100, 200, and 400 Oe.
$H$$\perp$$c$. $I$$\perp$$H$. The inset shows the $T$ dependence of
$\rho$$_{a}$($T$,$H$ = 0) ({\Large $\bullet$}) and d$\rho$$_{a}$($T$, $H$ =
0)/d$T$ ({\Large $\circ$}) at low
$T$.}
\end{figure}
The $c$-axis resistivity $\rho_{c}$ of Bi-MG was measured
using the four-probe method. After the sample was cooled from
298 K to 1.9 K at $H$ = 0, the $c$-axis resistivity $\rho_{c}$
was measured with increasing $T$ from 1.9 K to 50 K without
and with $H$ [$H$$\parallel$$c$ ($c$: $c$ axis)], where the current direction is
parallel to the field direction. This resistivity is denoted
as the longitudinal magnetoresistance. Figure \ref{fig:six} shows
the $T$
dependence of $\rho_{c}$ for Bi-MG at $H$ = 0 - 10 kOe.
The resistivity $\rho_{c}$ for Bi-MG shows a very broad
peak around 14 K at $H$ = 0, in contrast to $\rho_{c}$
for HOPG exhibiting a peak around 40 K.\cite{Suzuki1996} This peak shifts
to the low $T$ side with increasing $H$. For $H \geq$ 20
kOe, $\rho_{c}$ shows a semiconductor-like behavior: it
decreases with increasing $T$. Note that $\rho_{c}$ shows
a positive magnetoresistance: it increases with increasing $H$
at any fixed $T$ between 1.9 and 50 K. The value of $\rho_{c}$
at $T$ = 298 K and $H$ = 0 is 0.11 $\Omega$cm, which is
on the same order as that of HOPG
($\rho_{c}$ = 0.096 $\Omega$cm)\cite{Suzuki1996}
and BiCl$_{3}$ GIC ($\rho_{c}$ = 0.2 $\Omega$cm),\cite{McRae2001} leading
to a mean free path less than 1 $\AA$ according to the Drude formula.
This result suggests that there is no overlapping over nearest-neighbor
layers along the $c$ axis. The \textit{c}-axis conduction can occur
via the hopping of carriers between layers.
The in-plane resistivity $\rho_{a}$ of Bi-MG was also measured
using the four-probe method. After the sample was cooled from
298 K to 1.9 K at $H$ = 0, the in-plane resistivity $\rho_{a}$
was measured with increasing $T$ from 1.9 K to 50 K without
and with $H$ ($H$$\perp$$c$), where the field direction
is perpendicular to the current direction in the $c$ plane.
This resistivity is denoted as the transverse magnetoresistance.
The resistivity ratio $\Delta$ (= $\rho_{c}$/$\rho_{a}$)
is estimated as 30 at 298 K using the measured $\rho_{a}$.
However, the actual value of $\Delta$ is considered to
be much larger than 30 because of possible contribution of $\rho_{c}$
to the measured $\rho_{a}$. Figure \ref{fig:seven} shows the $T$ dependence
of $\rho_{a}$ of Bi-MG at $H$ = 0 - 400 Oe. The zero-field
($H$ = 0) in-plane resistivity increases with increasing $T$
at low $T$, showing a metallic behavior. It has a maximum around
15 K, and it decreases with further increasing $T$, showing
a semiconductor-like behavior. Note that the value of $\rho_{a}$
for Bi-MG (= 4.2 m$\Omega$cm) is much larger than that of BiCl$_{3}$
GIC (= 27 $\mu \Omega$cm at 298 K). We find that the $T$
dependence of $\rho_{a}$ for Bi-MG is very similar to that
of grafoil: they even have similar magnitudes. The grafoil is
a pyrolytic graphite with a polycrystalline structures formed
of many domains. According to Koike et al.,\cite{Koike1985} $\rho_{a}$ of
grafoil (grade GTA, Union carbide) shows a semiconductor-like
behavior, while $\rho_{a}$ of HOPG (Union Carbide) shows
a metallic behavior. This result suggests that the semiconductor-like
behavior in Bi-MG may arise from nanographites where the degree
of disorder is greatly enhanced. Such a semiconductor-like behavior
is observed at least below 25 kOe.
In the inset of Fig. \ref{fig:seven} we show the detail of $\rho_{a}$
at $H$
= 0 near $T_{c}$ = 2.48 K. No drastic decrease in $\rho_{a}$
below $T_{c}$ is observed with decreasing $T$, while the $T$-derivative
$d \rho_{a}/dT$ gradually decreases with increasing $T$
around $T_{c}$. The causes for the finite value of $\rho_{a}$
below $T_{c}$ in spite of the superconducting phase are as follows.
(i) The sample used in the present work is an exfoliated Bi-MG
based on HOPG. The sample surface is not flat partly because
of cracks generated in the basal plane. Since the current path
is not always located on the same layer, the contribution of
large $\rho_{c}$ to the observed $\rho_{a}$ is not
negligibly small. Another possibility is the local superconductivity
in isolated islands. In such a system there is competition between
the Josephson coupling ($E_{J}$) and the charging energy ($E_{U}$)
between superconducting islands.\cite{Zwerger1991} For $E_{J} \gg E_{U}$,
the Cooper pairs are delocalized leading to a superconducting
state with vanishing resistivity. For $E_{J} \ll E_{U}$,
the pairs will be localized and the transport is possible only
by thermal activation leading to insulating behavior at $T$
= 0 K. The resistivity varies with $T$ as $\rho \approx$
exp($T_{a}/T$), where $T_{a}$ is related to the activation
energy. In this model, the resistivity should increase with decreasing $T$,
which contradicts with our result. Liu et al.\cite{Liu1992} have reported
an unusual $T$ dependence of resistivity at low $H$ in ultrathin
superconducting films of Pb, Al, and Bi. The resistivity varies
with $T$ as $\rho \approx$ exp($T/T_{0}$) at
low $T$, where $T_{0}$ is a characteristic temperature. In
Bi-MG, $d\rho/dT$ decreases with increasing $T$
around $T_{c}$. In superconducting thin films, $d\rho/dT$
[$\approx$ exp($T/T_{0}$)/$T_{0}$] increases with increasing $T$.
Therefore, the $T$ dependence of $\rho_{a}$ in Bi-MG is
not the case of $\rho \approx$ exp($T/T_{0}$).
\subsection{\label{resultD}2D Weak localization effect}
\begin{figure}
\includegraphics[width=8.0cm]{BiMGfig7.eps}%
\caption{\label{fig:eight} $H$ dependence of
$\rho$$_{a}$($T$,$H$) for Bi-MG at
fixed $T$ (2 $\leq$ $T$ $\leq$ 50 K).
$H$$\perp$$c$.}
\end{figure}
As shown in Fig. \ref{fig:seven}, the in-plane resistivity $\rho_{a}$ slightly
decreases with increasing $H$ at the same $T$ (at least for
1.9 $\leq T \leq$ 50 K), indicating a negative magentoresistance
(NMR). Figures \ref{fig:eight}(a) and (b) show the $H$ dependence of $\rho_{a}$
for Bi-MG with 2 $\leq T \leq$ 50 K, where $H$$\perp$$c$
and the field direction is perpendicular to the current direction.
For each $T$, $\rho_{a}$ decreases with increasing $H$
at low $H$ ($H \geq$ 100 Oe), exhibits a local minimum
around $H$ = 2.5 kOe, and increases with further increasing $H$.
The sign of the difference $\Delta \rho_{a}$
[= $\rho_{a}(T,H)-\rho_{a}(T,H=0)$]
is negative for 0 $\leq H \leq$ 3.5 kOe due to the possible
2D weak localization effect (WLE).\cite{Hikami1980} It changes from negative
to positive at $H$ = 3.5 kOe. The resistivity $\rho_{a}$
drastically increases with further increasing $H$, as is observed
in compensated metals such as bulk Bi.\cite{Suzuki1978} The $H$ dependence
of $\rho_{a}$ for 3 kOe $\leq H \leq$ 45 kOe can
be well described by $\rho_{a}$ = $\rho_{0}$ + $\rho_{1}H$
+ $\rho_{2}H^{2}$, where $\rho_{0}$, $\rho_{1}$,
and $\rho_{2}$ are constants: $\rho_{0}$ = 4.290 m$\Omega$cm, $\rho_{1}$
= (9.85 $\pm$ 0.11) $\times$ 10$^{-3}$ $\mu \Omega$cm/Oe, and $\rho_{2}$
= (4.56 $\pm$ 0.23) $\times$ 10$^{-8}$ $\mu \Omega$cm/Oe$^{2}$ at 1.9
K. The linear term ($\rho_{1}H$) is dominant compared
to the squared-power term ($\rho_{1}H^{2}$) for $H \ll H_{1}$,
where $H_{1}$ = $\rho_{1}$/$\rho_{2}$ = 216 kOe at
1.9 K. For comparison, we also measured the $H$ dependence of $\rho_{a}$
in Bi-MG for 1.9 $\leq T \leq$ 30 K, where $H$$\parallel$$c$.
At each $T$, $\rho_{a}$ increases with increasing $H$.
Thus the sign of $\Delta \rho_{a}$ is always positive
for 0 $<H\leq$ 47.5 kOe and 1.9 $\leq T \leq$ 50
K, indicating an anti-localization effect which is similar to
that observed in Bi thin films\cite{Komori1983} with strong spin-orbit interactions.
The $T$ dependence of $\rho_{a}$ at $H$ = 45 kOe for $H$$\perp$$c$
is described by
\begin{equation}
\rho _{a} =a_{0} -a_{1} \ln (T),
\label{eq:three}
\end{equation}
for 1.9 $\leq T\leq$ 4.3 K, where $a_{0}$ = (4.8261 $\pm$
0.0002) m$\Omega$cm and $a_{1}$ = (3.79 $\pm$ 0.19) $\times$ 10$^{-3}$
m$\Omega$cm. As shown in previous papers,\cite{Suzuki1996,Suzuki2000} the theory
of the 2D WLE predicts that the following relation is valid for
the ratio $a_{2}$/$a_{1}$,
\begin{equation}
\frac{a_{1} }{a_{0} } =\frac{e^{2} }{2\pi ^{2} \hbar }
\frac{A}{\sigma_{2D}^{0} },
\label{eq:four}
\end{equation}
with $A=\alpha p+\gamma$, where $e^{2}/(2\pi^{2}\hbar$)
= 1.23314 $\times$ 10$^{-5}$ $\Omega^{-1}$ and $\sigma_{2D}^{0}$ is the
in-plane conductivity defined by $I_{c}$/$\rho_{a}^{0}$ ($\rho_{a}^{0}$
is the in-plane resistivity and $I_{c}$ is the $c$ axis repeat
distance). In the parameter $A$, $\alpha$ is nearly equal
to unity, $p$ is the exponent of the inelastic life time $\tau_{\epsilon}$
of the conduction electrons ($\tau_{\epsilon} \approx T^{-p}$),
and $\gamma$ is a numerical factor giving a measure of the
screening by other carriers. For convenience, here we use the
value of $\sigma_{2D}^{0}$ (= 1.89 $\times$ 10$^{-2}$ $\Omega^{-1}$) for
kish graphite ($I_{c}$ = 3.35 $\AA$ and $\rho_{a}^{0}$ = 1.77
$\mu \Omega$cm at $T$ = 4.2 K) obtained by Koike et al.,\cite{Koike1985}
instead of the corresponding data for Bi-MG.
A kish graphite is a high-quality single crystal of graphite which is
deposited on walls of blast furnaces for steel production.
Using $a_{1}/a_{0}$
= 7.853 $\times$ 10$^{-4}$ the parameter $A$ is estimated as $A$ = 1.20,
which is comparable with 1.14 for stage-4 MoCl$_{5}$ GIC.\cite{Suzuki2000} These
results suggest that the logarithmic behavior of $\rho_{a}$
can be explained in terms of the 2D WLE.
\subsection{\label{resultE}Field induced metal-semiconductor transition}
\begin{figure}
\includegraphics*[bb=50 50 425 750,clip,width=8.0cm]{BiMGfig8.eps}
\caption{\label{fig:nine} $T$ dependence of the difference $\mu$$_{a}$
between $\rho _{a} (T,H)$ and $\rho _{a} (T=20K,H)$ [$\mu _{a} =\rho
_{a} (T,H)-\rho _{a} (T=20K,H)$] for Bi-MG with various $H$
(1 $\leq$ $H$ $\leq$ 45 kOe).
$I$$\perp$$H$. (a)
$H$$\perp$$c$. (b) $H$$\parallel$$c$.
The inset of (b) shows the $T$ dependence of $\rho _{a} (T,H)$
at $H$ = 47.5 kOe. The solid line denotes the fitting curve of
the data to Eq.(\ref{eq:three}).}
\end{figure}
In Fig. \ref{fig:nine}(a) we show the $T$ dependence of the
difference ($\mu_{a}$)
between $\rho_{a}$ at $T$ and that at 20 K for various $H$
($H$$\perp$$c$): $\mu_{a} = \rho_{a}(T,H)
- \rho_{a}(T=20 K,H)$. The $T$ dependence of $\mu_{a}$
below 10 K is dependent on $H$. The difference $\mu_{a}$
shows a metallic behavior ($d\mu_{a}/dT>$ 0) below
15 - 20 kOe, while it shows a semiconductor-like behavior
($d\mu_{a}/dT<$ 0)
above $H$ = 25 kOe. In Fig. \ref{fig:nine}(b) we show the $T$ dependence of $\mu_{a}$
for $H$$\parallel$$c$. The $T$ dependence of $\mu_{a}$ below
7 K is dependent on $H$. The difference $\mu_{a}$ shows
a metallic behavior ($d\mu_{a}/dT>$ 0) below 10
kOe, while it shows a semiconductor-like behavior above 20 kOe.
Note that $\mu_{a}$ at $H$ = 47.5 kOe drastically increases
with decreasing $T$, leading to the insulating state. The magnitude
of $\mu_{a}$ for $H$$\parallel$$c$ is on the same order as that
for $H$$\perp$$c$. The $T$ dependence of $\rho_{a}$
at $H$ = 47.5 kOe for $H$$\parallel$$c$ is well described by
Eq.(\ref{eq:three})
for 1.9 $\leq T\leq$ 4.3 K, where $a_{0}$ = (5.8385
$\pm$ 0.0002) m$\Omega$cm and $a_{1}$ = (5.33 $\pm$
0.23) $\times$ 10$^{-3}$ m$\Omega$cm. The corresponding fitting curve
is denoted by the solid line in the inset of Fig. \ref{fig:nine}(b). The
ratio $a_{1}/a_{0}$
= 9.13 $\times$ 10$^{-4}$ is almost the same as that (= 7.85 $\times$ 10$^{-4}$) for $\rho_{a}$
at $H$ = 45 kOe for $H$$\perp$$c$ (see Sec. \ref{resultC}), indicating
the occurrence of 2D WLE.
Bi-MG undergoes a transition from the metallic phase to the semiconductor-like
phase at a critical field $H_{c}$. Similar crossover behavior
is observed in Si MOSFET,\cite{Simonian1997} and amorphous metal films
of InO$_{x}$,\cite{Hebard1990}
MoGe,\cite{Mason1999} and Bi.\cite{Markovic1998} The value of $H_{c}$ ($\approx$ 25 kOe)
for Bi-MG is almost the same as that for Si MOSFET\cite{Simonian1997} and amorphous
MoGe film.\cite{Mason1999} Note that the Zeemann energy $gS\mu_{B}H$
at 25 kOe corresponds to a thermal energy $k_{B}T_{H}$ with $T_{H}$
= 1.7 K, where $g$ = 2 and $S$ = 1/2. The temperature $T_{H}$
is slightly lower than $T_{c}$ (= 2.48 K) for the superconductivity.
The suppression of the metallic phase by $H$ is independent
of the directions of $H$ ($H$$\parallel$$c$ and $H$$\perp$$c$)
for Bi-MG. These results suggest that the spin related effect
is significant compared to the orbital effect. This is consistent
with the result derived from the susceptibility measurement that
the Curie-like susceptibility is dominant at low $T$.
\section{CONCLUSION}
Bi-MG shows superconductivity at $T_{c}$ = 2.48 K, where the
coherence length is much larger than the size of nanoparticles.
The spin related effect characterized by a Curie-like susceptibility
is enhanced, while the orbital effect is greatly suppressed.
The transition from metallic phase to semiconductor-like phase
is induced by the application of $H$ above 25 kOe. A negative
magnetoresistance in $\rho_{a}$ and a logarithmic divergence
of $\rho_{a}$ with decreasing $T$ are indicative of the
2D WLE.
\begin{acknowledgments}
We would like to thank Kikuo Harigaya for valuable comments on
the magnetism of nanographites with zigzag edges. The work at
SUNY-Binghamton was supported by SUNY-Research Foundation (Grant
No. 240-9522A). The work at Osaka University was supported by
the Ministry of Cultural Affairs, Education and Sport, Japan
under the grant for young scientists (No. 70314375) and by Kansai
Invention Center, Kyoto, Japan. We were
grateful to Advanced Ceramics Corporation, Ohio for providing
us with HOPG (grade ZYA).
\end{acknowledgments}
\begin{references}
\bibitem{Abrahams1979}E. Abrahams, P.W. Anderson, D.C. Licciardello, and T.V. Ramakrishnan,
Phys. Rev. Lett. \textbf{42}, 673 (1979).
\bibitem{Kravchenko1995}S.V. Kravchenko, W.E. Mason, G.E. Bowker, J.E. Furneaux, V.M.
Pudalov and M. D'Iorio, Phys. Rev. B \textbf{51}, 7038 (1995).
\bibitem{Simonian1997}D. Simonian, S.V. Kravchenko, M.P. Sarachik, and V.M. Pudalov,
Phys. Rev. Lett. \textbf{79}, 2304 (1997).
\bibitem{Kravchenko2000}S.V. Kravchenko and T.M. Klapwijk, Phys. Rev. Lett. \textbf{84},
2909 (2000).
\bibitem{Hebard1990}A.F. Hebard and M.A. Paalanen, Phys. Rev. Lett. \textbf{65}, 927
(1990).
\bibitem{Yazdani1995}Ali Yazdani and A. Kapitulnik, Phys. Rev. Lett. \textbf{74}, 3037
(1995).
\bibitem{Mason1999}N. Mason and A. Kapitulnik, Phys. Rev. Lett. \textbf{82}, 5341 (1999).
\bibitem{Markovic1998}N. Markovic, C. Christiansen, and A.M. Goldman, Phys. Rev.
Lett. \textbf{81}, 5217 (1998).
\bibitem{Walter1998}J. Walter, H. Shioyama, and Y. Sawada, Carbon \textbf{36}, 1811
(1998).
\bibitem{Walter1999a}J. Walter and H. Shioyama, Carbon \textbf{37}, 1151 (1999).
\bibitem{McRae2001}E. McRae and J. Walter, Carbon \textbf{39}, 717 (2001).
\bibitem{Suzuki1996}M. Suzuki, C.Lee, I.S. Suzuki, K. Matsubara, and K. Sugihara,
Phys. Rev. B \textbf{54}, 17128 (1996).
\bibitem{Fujita1996}M. Fujita, K. Wakabayashi, K. Nakada, and K. Kusakabe, J.
Phys. Soc. Jpn. \textbf{65}, 1920 (1996).
\bibitem{Wakabayashi1998}K. Wakabayashi, M. Sigrist, and M. Fujita, J. Phys. Soc. Jpn. \textbf{67},
2089 (1998).
\bibitem{Wakabayashi1999}K. Wakabayashi, M. Fujita, H. Ajiki, and M. Sigrist, Phys.
Rev. B \textbf{59}, 8271 (1999).
\bibitem{Walter1999b}J. Walter and H. Shioyama, Phys. Lett. A \textbf{254}, 65 (1999).
\bibitem{Walter2000a}J. Walter, Adv. Mater. \textbf{12}, 31 (2000).
\bibitem{Walter2000b}J. Walter, Phil. Mag. Lett. \textbf{80}, 257 (2000).
\bibitem{Haviland1989}D.B. Haviland, Y. Liu, and A.M. Goldman, Phys. Rev. Lett. \textbf{18},
2180 (1989).
\bibitem{Weitzel}B. Weitzel and H. Micklitz, Phys. Rev. Lett.
\textbf{66}, 385 (1991).
\bibitem{Deutscher1980}G. Deutscher, O. Entin-Wohlman, and Y. Shapira, Phys. Rev.
B \textbf{22}, 4264 (1980).
\bibitem{SaintJames1963}D. Saint-James and P.G. de Gennes, Phys. Lett. \textbf{7}, 306
(1963).
\bibitem{Tanuma1992}S. Tanuma, in \textit{Graphite Intercalation Compounds II}, edited
by H. Zabel and S.A. Solin (Springer-Verlag, Berlin, 1992) p.163.
\bibitem{Fukunaga1998}A. Fukunaga, S. Chu, and M.E. McHenry, J. Mater. Res. \textbf{13},
2465 (1998).
\bibitem{Anderssen1998}O.E. Andersson, B.L.V. Prasad, H. Sato, T. Enoki, Y. Hishiyama,
Y. Kaburagi, M. Yoshikawa, and S. Bandow, Phys. Rev. B \textbf{58},
16387 (1998).
\bibitem{Shibayama2000}Y. Shibayama, H. Sato, T. Enoki, and M. Endo, Phys. Rev. Lett. \textbf{84},
1744 (2000).
\bibitem{Koike1985}Y. Koike, S. Morita, T. Nakamoto, and T. Fukase, J. Phys.
Soc. Jpn. \textbf{54}, 713 (1985).
\bibitem{Zwerger1991}W. Zwerger, in \textit{Quantum Coherence in Mesoscopic Systems},
edited by B. Kramer (Plenum Press, New York, 1991) p.491.
\bibitem{Liu1992}Y. Liu, D.B. Haviland, L.I. Glazman, and A.M. Goldman, Phys.
Rev. Lett. \textbf{68}. 2224 (1992).
\bibitem{Hikami1980}S. Hikami, A.I. Larkin, and Y. Nagaoka, Prog. Theor. Phys. \textbf{63},
707 (1980).
\bibitem{Suzuki1978}M. Suzuki and S. Tanuma, J. Phys. Soc. Jpn. \textbf{45}, 1645 (1978).
\bibitem{Komori1983}F. Komori, S. Kobayashi, and W. Sasaki, J. Phys. Soc. Jpn. \textbf{52},
368 (1983).
\bibitem{Suzuki2000}M. Suzuki, I.S. Suzuki, K. Matsubara, and K. Sugihara, Phys.
Rev. B \textbf{61}, 5013 (2000).
\end{references}
\end{document}
%
% ****** End of file template.aps ******