Thermal and Electrical Currents in Nanoscale Electronic Interferometers
Extracted source text ยท arXiv cond-mat/0403414
%\documentclass[preprint]{revtex4}
\documentclass[twocolumn,showpacs,preprintnumbers,amssymb]{revtex4}
%\documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
%\documentclass[twocolumn,aps]{revtex4}
%\documentstyle[aps,prl,twocolumn]{revtex}
%\documentstyle[aps,prl,twocolumn,showkeys]{revtex}
\begin{document}
%\draft
\title{Thermal and Electrical Currents in Nanoscale Electronic Interferometers}
\author{Sam Young Cho and Ross H. McKenzie}
\affiliation{Department of Physics, University of Queensland,
Brisbane 4072, Australia
}
\date{\today}
\begin{abstract}
We theoretically study thermal transport
in an electronic interferometer
comprising a parallel circuit of two quantum dots,
each of which has a tunable single electronic state which
are connected to two leads at different temperature.
As a result of
quantum interference,
the heat current through one of the dots
is in the opposite direction to the temperature gradient.
An excess heat current flows through the other dot.
Although locally, heat flows from cold to hot,
globally the second law of thermodynamics
is not violated because
the entropy production associated with
heat transfer through the whole device is still positive.
The temperature gradient also induces
a circulating electrical current, which makes the interferometer
magnetically polarized.
\end{abstract}
\pacs{85.35.Ds, 73.63.Kv, 73.40.Gk}
%\pacs{73.23.Hk, 73.63.Kv, 73.40.Gk, 85.35.Ds}
%
% 73.23.Hk Electronic transport in mesoscopic systems:
% Coulomb blockade; single-electron tunneling
% 73.63.Kv Electronic transport in mesoscopic or nanoscale materials and
% structures: Quantum dots
% 73.40.Gk Electronic transport in interface structures: Tunneling
%
% 85.35.Ds Quantum interference devices
%
%
%
\maketitle
{\it Introduction.}
Manipulation of quantum coherence and interference
in a controllable manner is of special interest
in nanoscale electron devices \cite{review}.
The coherence of
resonant electron tunneling through a quantum dot (QD)
has been demonstrated by using Aharonov-Bohm interference \cite{Yacobi95}.
Moreover, such interference effects have enabled
the realization of a phase sensitive probe of
the anomalous transmission phase \cite{Schuster97},
dephasing effects \cite{Buks98}, and
many-body correlation effects \cite{Wiel00}
in quantum coherent transport through a QD.
Very recently,
a quantum interferometer based on two QDs
has been fabricated and
control of coherent electron
transport by varying gate voltages of each dot \cite{Holleitner00}
has been demonstrated.
In such a double dot interferometer,
theoretical studies have focused on the subjects
of resonant tunneling~\cite{Kubala02}, cotunneling~\cite{Akera93},
many-body correlation effect~\cite{Izumida97},
magnetic polarization current~\cite{Cho03},
and two-electron entanglement in the context of quantum
communication \cite{Loss00}.
Also, there has been considerable interest
in thermal transport through nanoscale devices \cite{Sivan,Beenakker,Guttman,
Moskalets2,Appleyard,Schwab,Andreev,Matveev,Kim02,Humphrey02}
and possible ``violation'' of
the second law of thermodynamics for small colloidal
systems over short time scales
\cite{Wang}, small quantum systems \cite{Nie}, and nanoscale electric circuits
\cite{Nie2}.
In this paper,
we consider
the thermal transport induced by a temperature gradient
across a double dot interferometer
(see Fig. \ref{fig1}).
The thermal transport could be manipulated
in a controlled manner such as
varying gate potentials.
In the interferometer, electric current conservation does not require
that the total current through the interferometer should be
greater than the local current through each electron path.
The quantum interference of tunneling electrons
results in a circulating electric current which
can make the magnetic states of the device be up-, non-, or down-polarized.
It was recently shown that the magnetic polarization current
exists at a finite bias between the leads \cite{Cho03}.
In this study,
the temperature difference between the leads
can give rise
to a circulating electric current without an applied bias.
Furthermore, it is found that due to quantum interference
the heat current flows in the opposite direction
to the temperature drop through one dot
while the excess heat current flows through the other dot.
The behaviors of the local heat currents show
the existence of a circulating heat current.
We discuss the second law of thermodynamics associated with
the two unique thermal transport processes in the interferometer.
%
%%%%%%%%%%% Figure %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{figure}
\vspace*{5.00cm}
\special{psfile=fig1.eps vscale=58 hscale=58 hoffset=5 voffset=5}
\caption{(color online) A quantum interferometer based on two quantum dots.
Both dots are tunnel-coupled to the left and right leads.
The tunneling amplitudes between the dots and the leads
are denoted by $\Gamma_1$ and $\Gamma_2$.
Each lead is described by an equilibrium Fermi-Dirac
distribution with temperature $T_L$ and $T_R$ and
electrochemical potential $\mu_L$ and $\mu_R$.
The energy level position in each dot is measured
as $\varepsilon_1$ and $\varepsilon_2$ relative to the Fermi energy
in the leads.
}
\label{fig1}
\end{figure}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
{\it Model.}
We start with a general model Hamiltonian
\begin{equation}
H = \sum_{\alpha} H_\alpha \{ c^\dagger_{k\sigma}; c_{k\sigma} \}
+ \sum_{\alpha,j} H^T_{\alpha,j}
+ \sum_{j} H^D_{j} \{ d^\dagger_{j\sigma}; d_{j\sigma} \},
\label{Hamiltonian}
\end{equation}
where $\alpha(=L,R)$ and $j(=1,2)$ are the indices of leads and dots.
The Hamiltonians $H_\alpha$, $H^D_j$, and
$$H^{T}_{\alpha,j}=\sum_{k\sigma \in \alpha;j}
\left[
V^{\alpha}_{k,j} c^\dagger_{k\sigma} d_{j\sigma} + {\rm h.c.}
\right],$$
respectively
represent the leads, the interferometer, and tunneling between
the leads and the dots.
$c_{k\sigma}$ and $d_{j\sigma}$ are the annihilation operators
with spin $\sigma$
for electrons in the leads and the dots.
%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
The energy (number of electrons) flowing into the interferometer
is defined as the rate of change in the energy (number of electrons)
in the lead $\alpha$ \cite{Mahan}:
$I^E_\alpha =-{d \langle H_\alpha \rangle / dt}
= (i/ \hbar) \langle [ H_\alpha, H ] \rangle$ and
$I_\alpha =-{d \langle N_\alpha \rangle / dt}
= (i/ \hbar) \langle [ N_\alpha, H ] \rangle$,
where $H_\alpha=\sum_{k\sigma\in \alpha} \varepsilon_k
c^\dagger_{k\sigma} c_{k\sigma}$ and
$N_\alpha=\sum_{k\sigma\in \alpha}
c^\dagger_{k\sigma} c_{k\sigma}$.
The heat current flowing into the interferometer from the lead $\alpha$
is defined by
\begin{equation}
I^Q_\alpha = I^E_\alpha - \mu_\alpha I_\alpha,
\end{equation}
where $\mu_\alpha$ is the electrochemical potential in the lead $\alpha$.
Using the Keldysh Green function
$G^<_{k\sigma,j\sigma}(t-t') \equiv i\langle d^\dagger_{j\sigma}(t')
c_{k\sigma}(t) \rangle$
which involves electron
operators for the leads and for each dot,
one writes the heat currents as
\begin{eqnarray}
I^Q_\alpha
\!\! &=& \!\!
- \!\! \sum_{k\sigma\in \alpha}\!\! \int\!\! \frac{d\varepsilon }{2\pi\hbar}
(\varepsilon_k -\mu_\alpha)
\left[ V_{k,1} G^<_{k\sigma,1\sigma}(\varepsilon) + {\rm h.c.} \right]
\nonumber \\ &&
\!\! - \!\!
\sum_{k\sigma \in \alpha}\!\! \int \!\!\frac{d\varepsilon}{2\pi\hbar}
(\varepsilon_k -\mu_\alpha)
\left[ V_{k,2} G^<_{k\sigma,2\sigma}(\varepsilon) + {\rm h.c.} \right] \, .
\label{current}
\end{eqnarray}
The first (second) line of Eq.~(\ref{current}) describes heat
transfer from the left lead to QD 1 (QD 2) or vice versa.
Then each heat transfer
can be defined as a local heat current through each dot, $I^Q_{\alpha,j}$.
Thus the total heat current is the sum of the local heat currents
through each dot,
$I^Q_\alpha = I^Q_{\alpha,1} + I^Q_{\alpha,2}$.
The heat current is written in terms of
the electron Green functions, $G^<$.
The heat transfer from the lead to one of the dots
is then accompanied by electron dynamics including
a complex trajectory through the entire interferometer,
as well as a direct tunneling to the dot.
The Green functions, $G^<$, can be expressed in terms of
the dot Green functions defined by
$G_{jj',\sigma}^{\rm r}(t-t')
=-i\theta (t-t')\langle \{ d_{j'\sigma}(t),d_{j\sigma}^\dagger(t')\}\rangle$
and
$G_{jj',\sigma}^{<}(t-t')
=i\langle d_{j'\sigma}(t)d_{j\sigma}^\dagger(t')\rangle$ \cite{Meir92}.
As a consequence, a general expression of the heat current
through a nanoscale electron interferometer is given by
\begin{eqnarray}
I^Q_\alpha
&=&
i \sum_{jj'\sigma} \int
\frac{d\varepsilon}{2\pi\hbar}
(\varepsilon-\mu_\alpha)
\Gamma_{jj',\alpha}(\varepsilon)
\left[
G^<_{jj',\sigma}(\varepsilon)
\right. \nonumber \\ && \hspace*{1.4cm} \left.
+ f_\alpha(\varepsilon)
\left( G^r_{jj',\sigma}(\varepsilon)
- G^a_{jj',\sigma}(\varepsilon) \right)
\right],
\end{eqnarray}
where the tunnel couplings between the leads
and the dots are denoted by
$\Gamma_{jj',\alpha}=2\pi {\cal N} V^\alpha_{j} V^{\alpha *}_{j'}$
with the density of states of the lead ${\cal N}$.
The Fermi-Dirac distribution functions of the leads
are
$f_\alpha (\varepsilon)=f(\varepsilon-\mu_\alpha)$,
where
the chemical potentials are
$\mu_L=-\mu_R=eV/2$ with applied bias $V$ between the two leads.
For $i \neq j$, the terms of the current describe
the interference between the electron waves through two dots.
In the absence of one dot, i.e.,
$V_{k,1}=0$ or $V_{k,2}=0$, only the current
through the other dot exists and
any interference resulting from the existence of the one dot
disappears.
Therefore, the expression of the heat current is reduced to
the heat current formula in a single dot electronic device.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Our interferometer has the two electron pathways which
are allowed for electron transport from one lead to the
other via two dots not being coupled to each other directly.
Electron tunneling through the dots are
manipulated by varying the gate voltages.
The dots makes it possible to
control a coherent electron passing through
the two electron pathways in the interferometer where
it is required to satisfy current conservation
at the leads. The current conservation
gives rise to a circulating current on
a closed path through the dots and the leads \cite{Cho03}.
To clarify the origin of the circulating electric/heat currents
in the interferometer
the intradot electron-electron interaction
is not taken into account in this study.
Then the level spacing in each dot is larger than
the applied bias and temperature
because electron transport through a single level in the dots.
Although intradot Coulomb interactions are considered
in the Coulomb blockade regime,
the resonant transport could be well explained
in the Hatree-Fock mean-field level where
the energy level of the dots can be described by
a simple shift of the interaction parameter.
In fact, we focus on studying the interference effects
that are present for near resonant transport
and employ the resonant level model to describe the dots;
$
H^D_j = \sum_{j\sigma} \varepsilon_j d^\dagger_{j\sigma} d_{j \sigma},
$
where $\varepsilon_1$ and $\varepsilon_2$ are the level energy in each dot,
measured, relative to the Fermi energy of the leads.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
With the Keldysh technique
for nonlinear current through the system,
the local heat currents through each dot at the lead $\alpha$
are given by
\cite{Meir92,Cho99},
\begin{equation}
I^Q_{\alpha,j} = \sum_\sigma \int \frac{d \varepsilon}{2\pi\hbar} \,
(\varepsilon -\mu_\alpha)
(f_\alpha(\varepsilon)-f_{\alpha'}(\varepsilon)) {\cal T}_j(\varepsilon),
\label{heat}
\end{equation}
and similarly the local electric currents are obtained as
\begin{equation}
I_{\alpha,j} =
e \sum_\sigma \int \frac{d \varepsilon}{2\pi\hbar} \,
(f_\alpha(\varepsilon)-f_{\alpha'}(\varepsilon)) {\cal T}_j(\varepsilon),
\label{electric}
\end{equation}
where the local transmission spectral functions are defined by
${\cal T}_j(\varepsilon)=
\left\{ \mbox{\boldmath $\Gamma$}^L
{\bf G}^{\rm r}_\sigma (\varepsilon)
\mbox{\boldmath $\Gamma$}^R
{\bf G}^{\rm a}_\sigma (\varepsilon) \right\}_{jj} $
which is the $j$-th diagonal component of the matrix transmission
spectral function.
${\bf G}^{\rm r}_\sigma (\varepsilon)$
is the matrix dot Green function defined in time space as
$G_{jj',\sigma}^{\rm r}(t-t')
=-i\theta (t-t')\langle \{ d_{j'\sigma}(t),d_{j\sigma}^\dagger(t')\}\rangle$.
The matrix coupling to the leads is described by
${\bf \Gamma}^L={\bf \Gamma}^R=
\left(\begin{array}{cc} \Gamma_1 & \sqrt{\Gamma_1 \Gamma_2 }
\\ \sqrt{\Gamma_1 \Gamma_2 } & \Gamma_2 \end{array} \right)$.
The symmetric tunnel-coupling between the dots and the leads
will be assumed to be independent of energy.
The matrix Green function of the dots is
\begin{equation}
\label{Green's function}
{\bf G}_\sigma^{\rm r}(\varepsilon) =
\left( \begin{array}{cc}
\varepsilon-\varepsilon_1 + i \Gamma_1
& i \sqrt{\Gamma_1 \Gamma_2} \\
i \sqrt{\Gamma_1 \Gamma_2}
& \varepsilon -\varepsilon_2+i \Gamma_2
\end{array} \right)^{-1}.
\end{equation}
From the relation,
${\bf G}_\sigma^{\rm a}(\varepsilon)
=[{\bf G}_\sigma^{\rm r}(\varepsilon)]^\dagger$,
the advanced Green function can be obtained.
%
Accordingly, the local transmission spectral functions
in terms of the total transmission function are given by
%\begin{subequations}
\begin{eqnarray}
{\cal T}_1(\varepsilon) &=&
{\Gamma_1 (\varepsilon -\varepsilon_2)
\over
\Gamma_2(\varepsilon \! - \! \varepsilon_1)
+\Gamma_1(\varepsilon \! -\! \varepsilon_2)} {\cal T}(\varepsilon) , \\
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
{\cal T}_2(\varepsilon) &=&
{\Gamma_2 (\varepsilon -\varepsilon_1)
\over
\Gamma_2(\varepsilon \! - \! \varepsilon_1)
+\Gamma_1(\varepsilon \! -\! \varepsilon_2)} {\cal T}(\varepsilon) .
\end{eqnarray}
%\end{subequations}
%
The total current is the sum of current through each dot,
$I^Q=I^Q_1+I^Q_2$, which is just the current conservation.
This leads to the total transmission spectral function as
${\cal T}(\varepsilon)={\cal T}_1(\varepsilon)+{\cal T}_2(\varepsilon)$,
\begin{equation}
{\cal T}(\varepsilon) \! = \!\!
{
\left[\Gamma_2(\varepsilon\! -\! \varepsilon_1)
\! +\!\Gamma_1\!(\varepsilon\! -\!\varepsilon_2)\right]^2
\over
(\varepsilon \!-\! \varepsilon_1)^2
(\varepsilon\!-\!\varepsilon_2)^2
\! + \!
\left[\Gamma_2(\varepsilon \! - \! \varepsilon_1)
+\Gamma_1(\varepsilon \! -\! \varepsilon_2)\right]^2 } .
\end{equation}
%
In the linear response regime, the transport electric/heat currents
are expanded
up to the linear terms of $\Delta T = T_L-T_R$ and $\Delta V = V_L - V_R$.
The electric currents and the heat currents are related
to the chemical potential difference, $\Delta V$,
and the temperature difference, $\Delta T$,
by the thermo-electric coefficients $L_{mm'}$ :
\begin{equation}
\left( \begin{array}{c} I \\ I^Q \end{array} \right)
=
\left( \begin{array}{cc} L_{11} & L_{12} \\ L_{21} & L_{22}
\end{array} \right)
\left( \begin{array}{c} \Delta V \\ \Delta T \end{array} \right).
\label{linearI}
\end{equation}
Similarly, with the local thermo-electric coefficients $L^{(j)}_{mm'}$
the local electric/heat currents in the linear
response regime can be written as
\begin{equation}
\left( \begin{array}{c} I_{j} \\ I^Q_{j} \end{array} \right)
=
\left( \begin{array}{cc} L^{(j)}_{11} & L^{(j)}_{12} \\
L^{(j)}_{21} & L^{(j)}_{22}
\end{array} \right)
\left( \begin{array}{c} \Delta V \\ \Delta T \end{array} \right).
\end{equation}
The thermo-electric coefficients associated with the local current
through each dot
are expressed as
$L^{(j)}_{11} = e^2 {\cal L}^{(j)}_0$, $L^{(j)}_{21}=TL^{(j)}_{12}=-e{\cal L}^{(j)}_1$,
and $L^{(j)}_{22}={\cal L}^{(j)}_2/T$, where
the integrals are defined as
$
{\cal L}^{(j)}_n (T)
= \frac{2}{h} \int d\varepsilon
\left(-\frac{\partial f}{\partial \varepsilon} \right)
\varepsilon^n {\cal T}_j(\varepsilon)$.
According to the current conservations, the thermo-electric
coefficients have the relations:
$L_{mm'}=L^{(1)}_{mm'}+L^{(2)}_{mm'}$.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5555555
\begin{figure}
\vspace*{6cm}
\special{psfile=fig2.eps vscale=70 hscale=70 voffset=-43 hoffset=-40}
\caption{Thermopower of the interferometer as a function of temperature
for different level positions of the dots,
$(\varepsilon_1/\Gamma,\varepsilon_2/\Gamma)$.
The tunneling amplitudes are taken as $\Gamma=\Gamma_1=\Gamma_2$.
For both energy levels of dots being above (below) the
Fermi energy, the sign of the thermopower is negative (positive).
The charge and heat are carried mainly through
the electron (hole) channels.
When the energy level of one dot is lying below the Fermi
energy and that of the other is lying above the Fermi energy,
the sign of the thermopower is changed as
the temperature varies. Therefore,
the main propagating channels for
charge and heat determine the sign of the thermopower.
The magnitude of the thermopower is of the order of
$k_B/e = 86.17 \mu {\rm V}/K$.
}
\label{fig2}
\end{figure}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
{\it Thermopower.}
The thermopower of the interferometer
can be found by measuring the induced voltage drop
across the interferometer when the temperature difference
between two leads is applied.
For zero electric transport current, $I=0$,
the thermopower is defined by the relation
\begin{equation}
S = - \lim_{\Delta T \rightarrow 0}
\frac{\Delta V}{\Delta T} \bigg|_{I=0}
= \frac{L_{12}}{L_{11}}.
\end{equation}
In terms of the defined integrals,
one can rewrite the thermopower,
$ S= -(k_B/e)({\cal L}_0/k_BT {\cal L}_1) $
with the constant $k_B/e\simeq 86.17 \mu V/K$.
In Fig. \ref{fig2},
the characteristics of the thermopower are shown to be
dependent on the energy level positions of the dots.
The sign of the thermopower can indicate
the main channel in transporting charge and heat.
When more transmission spectral weight lies in the electron channel
then in the hole channel,
the charge and heat are carried by mainly electron channels.
In this case the sign of the thermopower is negative.
In the opposite case,
since charge and heat transport through the hole channels is
predominant,
the thermopower is positive.
If the same amount of electron and heat are carried by each
electron and hole channel,
the sign of the electric/heat current is the same/opposite
for electron and hole channel.
This results in the thermopower being zero.
As shown in Fig. \ref{fig2},
when the energy level of one dot is lying below the Fermi energy
and that of the other dot is lying above the Fermi energy,
the sign of the thermopower is changed as temperature increases.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5555555
\begin{figure}
\vspace*{6cm}
\special{psfile=fig3a.eps vscale=70 hscale=70 voffset=-43 hoffset=-40}
\vspace*{6cm}
\special{psfile=fig3b.eps vscale=70 hscale=70 voffset=-43 hoffset=-40}
\caption{(a) Local electric currents induced by a temperature gradient
as a function of temperature.
% Temperature dependence of the local electric currents
% circulating on the closed path between the dots and the leads.
The dot energy levels are taken as
$(\varepsilon_1/\Gamma,\varepsilon_2/\Gamma)=(-0.9,0.3)$
for the tunneling amplitudes, $\Gamma=\Gamma_1=\Gamma_2$.
In contrast to the thermopower (compare Fig. \ref{fig2}),
the level positions taken in
this plot do not affect the physics of the local currents
but only change its amplitude and sign.
Even when the total electric transport current is zero $(I=0)$,
the non-vanishing local electric currents indicates
the existence of the circulating electric current
which makes the interferometer magnetically polarized.
The magnetic polarization current is then defined
as the local current: $I_M \equiv I_1 = -I_2$ for $I=0$.
(b)
The ratio of the magnetic polarization currents to the temperature
gradient, $\Delta T$, as a function of temperature.
The magnitude of the magnetic polarization currents
is of the order of ${\rm nA}$ for a temperature gradient of order
$1 {\rm K}$.
}
\label{fig3}
\end{figure}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
{\it Local electric currents and magnetic polarization currents.}
The requirement, that the electric transport current is zero,
for the thermopower
implies that the local electric currents are required
to ensure $I_1 = -I_2$.
If these local currents exist for $I=0$,
the local electric currents should circulate on the closed path
through the leads and the dots.
Then the interferometer can
be magnetized by the circulating electric currents.
One can define the circulating current as
a magnetic polarization current \cite{Cho03}, $I_M \equiv I_1=-I_2$ for $I=0$.
From Eq. (\ref{linearI}), the magnetic polarization current
is then expressed as
\begin{equation}
I_M = {\cal K} \, \Delta T,
\end{equation}
where
${\cal K}= - S L^{(1)}_{11} + L^{(1)}_{12}
= S L^{(2)}_{11} - L^{(2)}_{12}$.
This shows that the magnetic polarization current exists
even when the electric transport current is zero
because this magnetic polarization current is induced
by the temperature gradient
between the leads due to the quantum interference.
Figure \ref{fig3}(a) shows that the total sum of the local currents
is always zero because the local electric currents
are flowing along the opposite direction to each other, which
implies the existence of the magnetic polarization current.
It should be noted that the direction of the magnetic polarization current
is reversed as the temperature increases.
If one can define the local thermopower as $S_j = L^{(j)}_{12} /L^{(j)}_{11}$,
the magnetic polarization current
vanishes at a specific temperature, $T_0$, satisfying $S(T_0) =-S_j(T_0)$.
It is also shown in Fig. \ref{fig4} (b) that,
by manipulating the gate voltages of each dot,
the magnetic polarization current can be controlled.
This implies that changing the energy level positions of each dot,
one can magnetize the interferometer by the magnetic polarization
current induced by the temperature gradient as
up-, non-, and down-polarized.
In contrast to the measurement of the thermopower by means of
electron transport,
to observe the magnetic polarization current,
one can measure a magnetic field produced by the magnetic polarization
current by using a superconducting quantum interference device (SQUID).
Recent measurements of a persistent current by using a SQUID \cite{Rabaud01}
show that
magnitudes of the magnetic polarization current
of the order of ${\rm nA}$
should be experimentally observable
for a temperature gradient of order $0.1 \sim 1 {\rm K}$.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{figure}
\vspace*{5.5cm}
\special{psfile=fig4.eps vscale=75 hscale=75 hoffset=-45 voffset=-65}
\caption{(color online) The ratio of the magnetic polarization current
$I_M \equiv I_1=-I_2$
to the temperature gradient $\Delta T$
for $I=0$ as a function of the energy level
positions of each dot $(\varepsilon_1/\Gamma, \varepsilon_2/\Gamma)$
for $k_B T = 5.0 \times 10^{-2} \Gamma$.
The tunneling amplitudes are taken as $\Gamma=\Gamma_1=\Gamma_2$.
Note that by varying the energy level of each dot the interferometer
has a different magnetic state according to the direction of the
magnetic polarization current.}
\label{fig4}
\end{figure}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5555555
{\it Heat currents.}
The condition of an open circuit ($I=0$) to find the thermopower
can apply for a circulating heat current in the interferometer.
Under the condition, $I=0$, the local heat currents are rewritten as
\begin{equation}
I^Q_j = \kappa_j \Delta T,
\end{equation}
where $\kappa_j = -S L^{(j)}_{21} + L^{(j)}_{22}$.
Due to the quantum interference, the local heat currents
can be greater than the total heat current through the interferometer
at a given energy level positions of the dots.
If $I^Q_1 > I^Q$, there exists an excess heat current
through QD 1.
One can define the excess current as $I^Q_{exe} = I^Q_1 - I^Q$.
From the relation $I^Q=I^Q_1+I^Q_2$, the local heat current through
the QD 2 should be $I^Q_2 =-I^Q_{exe}$.
The negative sign of the local heat current through the QD 2 implies
that the heat current conservation requires a heat current flowing
through the QD 2 against the temperature gradient.
According to the second law of the thermodynamics,
the entropy production defined by $I^S \equiv I^Q/T > 0$
should be greater than zero during thermoelectric process \cite{Guttman}.
If one can define a local entropy production as $I^S_j \equiv I^Q_j/T$,
the local heat current flowing against the temperature gradient
means that $I^S_2 < 0$.
However, for heat transfer through the entire interferometer,
the heat current conservation should make
the second law of the thermodynamics preserved
in heat transport through the entire interferometer.
As a result, the excess heat transport through the QD 1
is compensated with the local heat current flowing against
the temperature gradient through the QD 2.
Like the magnetic polarization current,
those thermoelectric processes imply
that there appears a circulating heat current
on the closed path between the dots and the leads
in order to satisfy the second law of the thermodynamics.
Therefore, for $I^Q_1 > I^Q$,
the excess current can be defined as a circulating heat current,
$I^Q_M \equiv I^Q_{exe}=-I^Q_2$.
Similarly, for $I^Q < I^Q_2$, the circulating heat current is determined.
In Fig. \ref{fig5}, we display
the circulating heat current as a function of
the energy level positions of each dot from the numerical calculation.
It is shown that the interference between the electron and hole
channels produces the circulating heat current.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5555555
\begin{figure}
\vspace*{5.5cm}
\special{psfile=fig5.eps vscale=75 hscale=75 hoffset=-45 voffset=-65}
\caption{(color online) The ratio of the circulating heat current
to the temperature gradient $\Delta T$
as a function of the energy level
positions of each dot $(\varepsilon_1/\Gamma, \varepsilon_2/\Gamma)$
for $I=0$.
The parameters are taken as $k_B T = 5.0 \times 10^{-2} \Gamma$
and $\Gamma=\Gamma_1=\Gamma_2$.
The existence of the circulating heat current indicates
that
while the local heat current through one dot
can be greater than the total transport
heat current through the interferometer
due to the quantum interference,
the local heat current through the other dot
flows against the temperature gradient.
The current conservation requires
that the local excess heat current circulate
on the closed path in the interferometer.
}
\label{fig5}
\end{figure}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
{\it Summary.}
We have investigated thermal transport
in nanoscale interferometers.
The expression of the heat current for the interferometer
has been derived based on
the nonequilibrium Green's function technique.
Controllable electronic states in two dots
make it possible to manipulate the quantum interference
which causes a heat current in the opposite direction
to the temperature drop through one dot
and an excess heat current through the other dot.
The circulating electric current induced by
a temperature gradient across the entire interferometer
is sufficiently large that
it should be experimentally observable.
{\it Note added.}
After completion of this work, we become aware of some related
work by Moskalets concerning a temperature-induced circulating electric current
in a one-dimensional ballistic ring \cite{Moskalets1}.
{\it Acknowledgments.}
This work was supported by the University of Queensland and
the Australian Research Council.
%\begin{references}
\begin{thebibliography}{}
\bibitem{review}
See, for a review, Y. Imry, {\it Introduction to Mesoscopic Physics},
2nd ed.
(Oxford University Press, Oxford, New York 2002);
S. Datta, {\it Electronic Transport in Mesoscopic Systems}
(Cambridge University Press, New York 1995).
\bibitem{Yacobi95}
A. Yacoby, M. Heiblum, D. Mahalu, and H. Shtrikman,
Phys. Rev. Lett. {\bf 74}, 4047 (1995).
%
\bibitem{Schuster97}
R. Schuster, E. Buks, M. Heiblum, D. Mahalu, V. Umansky, H. Shtrikman,
Nature {\bf 385}, 417 (1997).
%
\bibitem{Buks98}
E. Buks, R. Schuster, M. Heiblum, D. Mahalu, and V. Umansky,
Nature {\bf 391}, 871 (1998).
%
\bibitem{Wiel00}
W.G. van der Wiel, S. De Franceschi, T. Fujisawa, J.M. Elzerman, S. Tarucha,
and L.P. Kouwenhoven,
Science {\bf 289}, 2105 (2000);
%
%\bibitem{Ji00}
Y. Ji, M. Heiblum, D. Sprinzak, D. Mahalu, and H. Shtrikman,
Science {\bf 290}, 779 (2000).
%
%
\bibitem{Holleitner00}
A.W.Holleitner, C. R. Decker, H. Qin, K. Eberl, and R.H. Blick,
Phys. Rev. Lett. {\bf 87}, 256802 (2001).
%
\bibitem{Kubala02}
B. Kubala and J. K\"onig,
Phys. Rev. B {\bf 65}, 245301 (2002).
%
\bibitem{Akera93}
H. Akera, Phys. Rev. B {\bf 47}, 6835 (1993).
%
\bibitem{Izumida97} W. Izumida, O. Sakai, and Y. Shimizu, J. Phys. Soc.
Japan {\bf 66}, 717 (1997).
%
\bibitem{Cho03}
S. Y. Cho, R. H. McKenzie, K. Kang, and C. K. Kim,
J. Phys.: Condensd Matter {\bf 15}, 1147 (2003).
%
\bibitem{Loss00}
D. Loss and E.V. Sukhorukov,
Phys. Rev. Lett. {\bf 84}, 1035 (2000).
%
\bibitem{Sivan}
U. Sivan and Y. Imry,
Phys. Rev. B {\bf 33}, 551 (1986).
\bibitem{Beenakker}
C. W. J. Beenakker and A. A. M. Staring,
Phys. Rev. B {\bf 46}, 9667 (1992).
\bibitem{Guttman}
G.D. Guttman, E. Ben-Jacob, and D.J. Bergman,
Phys. Rev. B {\bf 51}, 17758 (1995).
\bibitem{Moskalets2}
M. V. Moskalets,
JETP {\bf 87}, 991 (1998)
[Soviet Zh. Eksp. Teor. Fiz. {\bf 114}, 1827 (1998)].
\bibitem{Appleyard}
N. J. Appleyard, J. T. Nicholls, M. Y. Simmons, W. R. Tribe,
and M. Pepper, Phys. Rev. Lett. 81, 3491 (1998).
\bibitem{Schwab}
K. Schwab, E. A. Henriken, J. M. Worlock, and M. L. Roukes,
Nature {\bf 404}, 974 (2000).
\bibitem{Andreev}
A. V. Andreev and K. A. Matveev,
Phys. Rev. Lett. {\bf 86}, 280 (2001).
\bibitem{Kim02}
T.-S. Kim and S. Hershfield,
Phys. Rev. Lett. {\bf 88}, 136601 (2002).
\bibitem{Matveev}
K. A. Matveev and A. V. Andreev,
Phys. Rev. B {\bf 66}, 045301 (2002).
\bibitem{Humphrey02}
T. E. Humphrey, R. Newbury, R. P. Taylor, and H. Linke,
Phys. Rev. Lett. {\bf 89}, 116801 (2002).
\bibitem{Wang}
G. M. Wang, E. M. Sevick, Emil Mittag, Debra J. Searles,
and Denis J. Evans,
Phys. Rev. Lett. {\bf 89}, 050601 (2002);
S. Tasaki, I. Terasaki, and T. Monnai, cond-mat/0208154 (unpublished);
Th. M. Nieuwenhuizen and A. E. Allahverdyan, cond-mat/0207587 (unpublished).
\bibitem{Nie}
Th. M. Nieuwenhuizen and A. E. Allahverdyan,
Phys. Rev. E {\bf 66}, 036102 (2002).
\bibitem{Nie2}
A. E. Allahverdyan and Th. M. Nieuwenhuizen,
Phys. Rev. B {\bf 66}, 115309 (2002).
\bibitem{Mahan}
G. D. Mahan, {\it Many-Particle Physics}, 2nd ed. (Plenum, New York 1990),
p. 227.
%
\bibitem{Meir92}
Y. Meir, and N. S. Wingreen,
Phys. Rev. Lett. {\bf 68}, 2512 (1992).
%
\bibitem{Cho99}
S. Y. Cho, K. Kang, and C.-M. Ryu,
Phys. Rev. B {\bf 60}, 16874 (1999).
%
%
\bibitem{Rabaud01}
W. Rabaud, %{\it et al.},
L. Saminadayar, D. Mailly, K. Hasselbach, A. Benoit, and B. Etienne,
Phys. Rev. Lett. {\bf 86}, 3124 (2001).
\bibitem{Moskalets1}
M. V. Moskalets,
Europhys. Lett. {\bf 41}, 189 (1998).
%\end{references}
\end{thebibliography}
%%
\end{document}