Multiphoton transitions in a macroscopic quantum two-state system
Extracted source text ยท arXiv cond-mat/0403425
\documentclass[aps,prl,twocolumn,superscriptaddress,showpacs]{revtex4}
%\documentclass[aps,prl,preprint,superscriptaddress,showpacs]{revtex4}
\usepackage{graphicx}
\begin{document}
\title{Multiphoton transitions in a macroscopic quantum two-state system}
\author{S. Saito}
\affiliation{NTT Basic Research Laboratories, NTT Corporation,
Kanagawa 243-0198, Japan}
\affiliation{CREST, Japan Science and Technology Agency, Saitama 332-0012,
Japan}
\author{M. Thorwart}
\affiliation{Institut f\"ur Theoretische Physik IV, Heinrich-Heine-Universit\"at D\"usseldorf,
40225 D\"usseldorf, Germany}
\author{H. Tanaka}
\affiliation{NTT Basic Research Laboratories, NTT Corporation,
Kanagawa 243-0198, Japan}
\affiliation{CREST, Japan Science and Technology Agency, Saitama 332-0012,
Japan}
\author{M. Ueda}
\affiliation{NTT Basic Research Laboratories, NTT Corporation,
Kanagawa 243-0198, Japan}
\affiliation{CREST, Japan Science and Technology Agency, Saitama 332-0012,
Japan}
\affiliation{Department of Physics, Tokyo Institute of Technology,
Tokyo 152-8551, Japan}
\author{H. Nakano}
\affiliation{NTT Basic Research Laboratories, NTT Corporation,
Kanagawa 243-0198, Japan}
\affiliation{CREST, Japan Science and Technology Agency, Saitama 332-0012,
Japan}
\author{K. Semba}
\affiliation{NTT Basic Research Laboratories, NTT Corporation,
Kanagawa 243-0198, Japan}
\affiliation{CREST, Japan Science and Technology Agency, Saitama 332-0012,
Japan}
\author{H. Takayanagi}
\affiliation{NTT Basic Research Laboratories, NTT Corporation,
Kanagawa 243-0198, Japan}
\affiliation{CREST, Japan Science and Technology Agency, Saitama 332-0012,
Japan}
\begin{abstract}
We have observed multiphoton transitions between two macroscopic
quantum-mechanical superposition states formed
by two opposite circulating currents
in a superconducting loop with three
Josephson junctions.
%The clockwise and counterclockwise circulating
%persistent currents correspond to the two macroscopically distinct states.
Resonant peaks and dips
of up to three-photon transitions were observed in spectroscopic measurements
when the system was irradiated with a strong RF-photon field.
The widths of the multiphoton absorption dips are shown to scale
with the Bessel functions in agreement with theoretical predictions derived from
the Bloch equation or from a spin-boson model.% with an Ohmic bath.
\end{abstract}
\pacs{74.50.+r, 03.67.Lx, 42.50.Hz, 85.25.Dq}
\maketitle
A macroscopic quantum two-state system (TSS) offers a unique testing ground
for exploring the foundations of quantum mechanics \cite{Leggett}.
This system can be in a quantum-mechanical superposition of macroscopically
distinct states, and its quantum nature can be revealed by measuring
the absorption of an integer number of photons from an externally applied
photon field \cite{Grifoni}. Since a macroscopic quantum system
cannot be completely decoupled from its environment, dissipative and
decoherence effects are unavoidable \cite{Leggett,Grifoni,Weiss}.
In addition to the investigation of fundamental physics,
the quantum TSS also serves as an elementary carrier of information
in a quantum information processor in the form of a quantum bit (qubit)
\cite{Bouwneester}.
Artificially designed superconducting circuits with mesoscopic Josephson
junctions constitute an important class of macroscopic quantum systems.
The charge degree of freedom of Cooper pairs is used to induce coherent
quantum oscillations between two charge states of a Cooper pair box
\cite{Nakamura and Vion}. A circuit with a single relatively large Josephson
junction, which is current-biased close to its critical current, forms a
so-called Josephson phase qubit \cite{Yu and Martinis}. Moreover,
three Josephson
junctions arranged in a superconducting loop threaded by an externally
applied magnetic flux constitute a flux qubit \cite{Mooij}.
The device could be prepared in a quantum superposition of two states
carrying opposite macroscopic persistent currents \cite{Wal}. Coherent Rabi
oscillations have been reported, when the qubit and
the readout device are connected to obtain a large signal \cite{Chiorescu}.
Rabi oscillations have also been observed in a system, where the qubit and
the readout device are spatially separated \cite{Tanaka1}.
Since these solid-state devices are thought to be scalable up to a large
number of qubits, they are of particular interest in the context of
solid-state quantum information processing \cite{Makhlin}.
The energy scale of quantum circuits containing Josephson junctions is
in the microwave regime. This property was demonstrated
in the current-voltage characteristics of a Josephson junction under microwave
irradiation displaying the well-known Shapiro steps \cite{Shapiro}.
They appear at voltages corresponding to integer multiples of the applied
microwave energy. With this phenomenon, the superconductor phase difference at
the junction, which is a macroscopic degree of freedom, can be treated
as a classical degree of freedom moving in the Josephson potential.
%in phase space.
In contrast, the quantum-mechanical behavior of
the macroscopic phase difference was demonstrated using one-photon
absorption processes between quantized energy levels within a single Josephson
potential well \cite{Devoret}. Recently, Wallraff {\em et al.\/} presented
experimental evidence of multiphoton absorption between quantized energy levels
within the single potential well formed by a large current-biased Josephson
junction \cite{Wallraff}.
In this Letter, we report the first observation of muliphoton transitions
between superposition states of \textit{macroscopically distinct states}
\cite{Leggett02}, which are formed in the double-well potential system
of a superconducting flux qubit.
The presence of an energy gap at the degeneracy point of the macroscopically
distinct states will be shown to be a prerequisite if we are to
excite the qubit by applying a microwave even when the operating point is
far from the degeneracy point.
\begin{figure}
\includegraphics[width=1.0\linewidth]{fig1}%
\caption{(a) Scanning-electron-microscope picture of a superconducting flux
qubit system. (b) Signal-to-noise ratio $\Delta I_{\mathrm{qb}}/I_{\mathrm{std}}$
(solid curve) and average switching current $I_{\mathrm{sw}}$
(dashed curve) as a function of applied flux through the qubit loop. The two
arrows indicate the signal-to-noise ratios at
$\Phi_{\mathrm{qubit}}/\Phi_{0}$=0.5 and 1.5.
(c) Spectroscopic data of the qubit. Resonant frequency is plotted
as a function of the half distance between the resonant peak and dip
in $\Phi_{\mathrm{qubit}}/\Phi_{0}$.
The solid curve represents a numerical fit to the data.
The dashed line is the energy difference between localized states
$|\downarrow\rangle$ and $|\uparrow\rangle$. From this fit, we obtained
$E_{\mathrm{J}}/h$=380 [GHz] and $\Delta/2\pi$=0.56 [GHz].
\label{fig1}}
\end{figure}
Our device is fabricated by lithographic techniques that define
the structure of an inner aluminum loop forming the qubit and an outer
enclosing SQUID loop for the read-out (see Fig.~\ref{fig1}(a)).
They are spatially separated but magnetically coupled by the mutual inductance
$M\simeq$7 pH.
The inner loop contains three Josephson junctions, one of which has an area
$\beta$(=0.7) times smaller than the nominally identical areas
of the other two.
The larger junctions have a Josephson energy of
$E_{\mathrm{J}}=\hbar I_{\mathrm{c}}/2e$, where $I_{\mathrm{c}}$ is
the critical current of the junction and $e$ is the electron charge.
The outer loop contains two Josephson junctions.
By carefully designing the junction parameters \cite{Mooij, Wal},
the inner loop can be made to behave as an effective TSS \cite{Nakano}.
In fact, the read-out result of the qubit changes greatly with
the qubit design ranging from the purely classical to the quantum regime
\cite{Takayanagi}. It is described by the Hamiltonian
$\hat{H}_{\mathrm{qb}}=\frac{\hbar}{2}(\varepsilon_{0} \hat{\sigma}_{z}+\Delta
\hat{\sigma}_{x})$, where $\hat{\sigma}_{x,z}$ are the Pauli spin operators.
Two eigenstates of $\hat{\sigma}_{z}$ are localized states
$|\downarrow \rangle$ corresponding to the clockwise persistent current
of the qubit and $|\uparrow \rangle$ corresponding to the counterclockwise
current. Energy eigenstates $|0 \rangle$ and $|1 \rangle$ of
$\hat{H}_{\mathrm{qb}}$ show energy anticrossing with energy gap
$\hbar \Delta$. An externally applied static magnetic
flux $\Phi_{\mathrm{qubit}}$ generates the energy imbalance
$\hbar \varepsilon_{0}=I_{\mathrm{P}}\Phi_{\mathrm{0}}(\Phi_{\mathrm{qubit}}
/\Phi_{\mathrm{0}}-f_{\mathrm{op}})$
between the two potential wells, where $\Phi_{\mathrm{0}}=h/2e$ is the flux
quantum, $I_{\mathrm{P}}=I_{\mathrm{c}} \sqrt{1-(1/2\alpha)^{2}}$
is the persistent current of the qubit, and $f_{\mathrm{op}}$
is a qubit operating point which is a half integer.
The effective energy gap in the biased qubit is
$\hbar\Delta_{\mathrm{b}}=\hbar\sqrt{\varepsilon_{0}^{2}+\Delta^{2}}$.
The qubit dynamics is controlled by a time-dependent RF field
$s(t)=s \cos{\omega_{\mathrm{ex}}t}$ %with amplitude $s$ and frequency $\omega_{\mathrm{ex}}$
provided by an on-chip RF-line. This leads to an additional term
in the Hamiltonian $\hat{H}_{\mathrm{RF}}(t)=-\frac{\hbar}{2} s(t)\hat{\sigma}_{z}$.
If $\Delta$ is zero, $\hat{H}_{\mathrm{qb}}+\hat{H}_{\mathrm{RF}}(t)$ has only
diagonal terms. In this case, the RF field cannot excite the qubit
from the ground to the first excited state. A finite $\Delta$ is
therefore a prerequisite for a spectroscopic experiment. In other words,
resonant peaks and dips in the qubit signals are direct evidence
for the coherent superposition of the macroscopically distinct states and
for the multiphoton transitions to occur between them.
The qubit state was detected by measuring the switching currents
of the dc-SQUID. We defined the switching current as the current
when the $I$-$V$ characteristic exceeded 30 $\mu$V. The
probability distribution of the switching current was obtained by repeating
the measurements typically 500 times. The measurements were carried out
in a dilution refrigerator at a temperature of 30 mK. The bias voltage had
a triangular waveform and was fed through large bias resistors
of 1 M$\Omega$ to achieve a current-bias measurement. The sweep rate of
the current was set at 120 $\mu$A/s and the frequency of the wave was 310 Hz.
The RF line used to apply microwaves contained three attenuators in the input
line: 30 dB at room temperature, 10 dB at 4.2 K and 10 dB at the base
temperature. We fabricated an on-chip strip line to achieve strong
coupling between the qubit and the RF line.
To achieve a good read-out resolution, we chose the operating point
$f_{\mathrm{op}}$ at $\Phi_{\mathrm{qubit}}/\Phi_{0}$=1.5.
Figure~\ref{fig1}(b) shows
the signal-to-noise ratio (S/N)
$\Delta I_{\mathrm{qb}}/I_{\mathrm{std}}$ and the average switching current
$I_{\mathrm{sw}}$
as functions of the applied flux through the qubit loop, where
$I_{\mathrm{std}}$ denotes the standard deviation of the SQUID switching
currents over 150 events, which may be considered to be a noise level
in the qubit readout. The qubit signal amplitude $\Delta I_{\mathrm{qb}}$
should be proportional to the flux derivative of $I_{\mathrm{sw}}$,
and this means that we can estimate $\Delta I_{\mathrm{qb}}$ by using the signal
amplitude $\Delta I_{\mathrm{qb0}}$ at
$\Phi_{\mathrm{qubit}}/\Phi_{0}$=1.5 (see Fig.~\ref{fig2}(a)).
In experiments, the qubit signal appears when $\Phi_{\mathrm{qubit}}/\Phi_{0}$
is a half-integer and the two arrows in Fig.~\ref{fig1}(b) indicate the S/N at
$\Phi_{\mathrm{qubit}}/\Phi_{0}$=0.5 and 1.5. Hence we chose the operating
point at $\Phi_{\mathrm{qubit}}/\Phi_{0}$=1.5 to achieve a higher readout
resolution.
\begin{figure}
\includegraphics{fig2}%
\caption{Applied magnetic flux dependence of the qubit signal
$\delta I_{\mathrm{sw}}$.
We subtracted a sinusoidal background signal from the averaged
switching current of the dc-SQUID $I_{\mathrm{sw}}$. (a) Without microwave
irradiation. The data were obtained after averaging 2000 measurements.
(b) With microwave irradiation. The data were obtained after averaging
500 measurements. The microwave power and frequency
were 5 dBm at the generator and 9.1 GHz, respectively. Resonant peaks
of up to two-photon processes and dips of up to three-photon processes
were observed. \label{fig2}}
\end{figure}
Figure~\ref{fig2}(a) shows the qubit signal $\delta I_{\mathrm{sw}}$
as a function of the external flux $\Phi_{\mathrm{qubit}}$ at
$\Phi_{\mathrm{qubit}}/\Phi_{0} \simeq 1.5$,
which is derived by subtracting the sinusoidal background from the SQUID
switching current $I_{\mathrm{sw}}$.
It shows a change in the thermal averaged persistent current of the qubit.
We obtained the sample temperature of 65 mK
from a numerical fit to the data by using the qubit
parameters $E_{\mathrm{J}}$ and $\Delta$, which are derived from the
spectroscopy measurements (see Fig.~\ref{fig1}(c)).
Figure~\ref{fig2}(b) shows the $\Phi_{\mathrm{qubit}}$ dependence of
the qubit signal under microwave irradiation.
The observed distinct resonant peaks and dips
are attributed to situations, in which the effective energy separation
of the two qubit states $\Delta_{\mathrm{b}}$ matches an integer
multiple of the RF photon energy $n \hbar \omega_{\mathrm{ex}}$.
We have detected up to three resonant peaks and dips for various
fixed RF frequencies. It should be noted that, at the resonant
peaks and dips, the qubit is
in a macroscopic quantum superposition of the two energy eigenstates.
The half width at half maximum (HWHM$_{n}$) and the normalized
amplitude $A_{n}$ of the dips are shown in Fig.~\ref{fig3}(a) and (b)
with various microwave amplitude $I_{\mathrm{RF}}$ for $n$=1, 2, 3,
which is defined by $I_{\mathrm{RF}}=\exp[P_{\mathrm{RF}}/20]$.
Here $P_{\mathrm{RF}}$ [dBm] is the microwave power at the signal generator.
We derived the HWHM$_{n}$ and $A_{n}$ from fitting the
$\Phi_{\mathrm{qubit}}/\Phi_{0}$ dependence of $\delta I_{\mathrm{sw}}$
at around the resonant dips (see Fig.~\ref{fig2}(b))
by using a Lorentzian with a linear background.
The $A_{n}$ values are normalized by the full amplitude
of the dips (3 [nA]).
The HWHM$_{n}$ carries important information on the dephasing and
relaxation processes caused by the interaction with the environment.
To calculate the line shape of the one-photon absorption dip,
we have used phenomenological Bloch equations in the first approach.
To describe the strong driving region and the
$n$-th dip, we replace the single-photon Rabi frequency
in the Bloch equations by the
corresponding multi-photon frequency
derived from theory of the dressed-atom approach \cite{Tannoudji}.
This substitution has been verified by numerical simulations
\cite{Goorden1}. In a second approach, we have used a microscopic driven
spin-boson model with a standard Ohmic spectral density
\cite{Weiss,Hartmann00} to calculate the line-shape of the resonances.
Both methods yield similar results in excellent agreement with the
experimental ones as seen in Fig.~\ref{fig3}(a).
\begin{figure}
\includegraphics{fig3}
%\includegraphics{model}%
%\vspace*{5cm}%
\caption{Half width at half maximum (a) and normalized amplitude (b)
of the resonant dips as functions of the microwave amplitude
$I_{\mathrm{RF}}$. The microwave frequency is $\omega_x / 2 \pi =$ 3.8 [GHz].
Solid (dashed) curves represent theoretical fits obtained by
the Bloch equations combined with the dressed-atom approach
(by the real-time path-integral method). We obtained
$T_{\mathrm{r1}}$=31(27) [ns], $T_{\mathrm{r2}}$=84(66) [ns],
$T_{\mathrm{r3}}$=430(240) [ns], $T_{\phi}$=4.0(4.2) [ns], $c$=0.55(0.55).
We have obtained similar results for $\omega_x / 2 \pi = 9.1$ [GHz] and
$11.4$ [GHz] (not shown).
\label{fig3}}
\end{figure}
The Bloch equations describe the dynamics of a spin $\frac{1}{2}$
in a constant field in the $z$-direction and a time-dependent field perpendicular
to it. To keep the analogy between the spin $\frac{1}{2}$ and the flux qubit,
we write the Hamiltonian in terms of
the energy eigenstates \cite{Goorden} and obtain
%
\begin{equation}
\hat{H}=\frac{\hbar}{2} \left( \left\{ \Delta_{\mathrm{b}}-\frac{\varepsilon_{0}}
{\Delta_{\mathrm{b}}}s(t) \right\} \hat{\sigma}_{z}+\frac{\Delta}{\Delta_{\mathrm{b}}}
s(t)\hat{\sigma}_{x} \right).
\label{H}
\end{equation}
%
The second term of the Hamiltonian leads to a non-adiabatic periodic variation
in the Larmor frequency.
When $\frac{\varepsilon_{0}}{\Delta_{\mathrm{b}}}s \simeq
\Delta_{\mathrm{b}}$, the Rabi frequency is decreased. However, we can
disregard this term under the usual experimental condition, $\frac{\varepsilon_{0}}
{\Delta_{\mathrm{b}}}s < \Delta_{\mathrm{b}}$. We have confirmed this by
numerical simulations. If we adopt the rotating-wave
approximation, the motion of the qubit spin $\langle \hat{\sigma}(t) \rangle$
in the laboratory frame can be described by the Bloch equations
%
\begin{eqnarray}
\frac{d\langle \hat{\sigma}_{x/y}(t)\rangle}{dt} &=& [\gamma \langle \hat{\sigma}(t)
\rangle \times \vec{B}(t)]_{x/y}-\frac{\langle \hat{\sigma}_{x/y} (t)\rangle}{T_{\phi}}, \nonumber \\
%\frac{d\langle \hat{\sigma}_{y}(t)\rangle}{dt} &=& [\gamma \langle \hat{\sigma}(t)
%\rangle \times \vec{B}(t)]_{y}-\frac{\langle \hat{\sigma}_{y} (t)\rangle}{T_{\phi}}, \nonumber \\
\frac{d\langle \hat{\sigma}_{z}(t)\rangle}{dt} &=& [\gamma \langle \hat{\sigma}(t)
\rangle \times \vec{B}(t)]_{z}-\frac{\langle \hat{\sigma}_{z} (t)\rangle-\sigma_{0}}{T_{\mathrm{r}}}.
\label{Bloch}
\end{eqnarray}
%
Here, $\gamma \vec{B}(t) = - \frac{\hbar}{2}(\frac{\Delta}{2\Delta_{\mathrm{b}}}s \cos
\omega_{\mathrm{ex}}t, \frac{\Delta}{2\Delta_{\mathrm{b}}}s \sin \omega_{\mathrm{ex}}t, \Delta_{\mathrm{b}})$,
$\sigma_{0}$ is the thermal equilibrium value of
$\langle \hat{\sigma}_{z}(t)\rangle$, and $T_{\mathrm{r}}$ and $T_{\phi}$ are
the relaxation and dephasing times. The steady-state solutions
of Eqs.~(\ref{Bloch}) can be obtained in the rotating frame by setting
$\frac{d\langle \hat{\sigma}_{i}(t)\rangle}{dt}=0$.
A resonant dip with a Lorentzian line shape appears at around
$\omega_{\mathrm{ex}} \simeq \Delta_{\mathrm{b}}$.
The HWHM$_{1}$ and the amplitude $A_{1}$ of the resonant dip follow as
%
\begin{eqnarray}
\mathrm{HWHM}_{1}&=&\sqrt{\left( \frac{1}{T_{\phi}} \right)^{2}+\omega_{1}^{2} \left(
\frac{T_{\mathrm{r}}}{T_{\phi}} \right)^{2}} \label{HWHM}, \\
A_{1} &=& \frac{\omega_{1}^{2} T_{\mathrm{r}} T_{\phi}}{1+\omega_{1}^{2} T_{\mathrm{r}}
T_{\phi}}\sigma_{0}, \label{Amplitude}
\end{eqnarray}
%
where $\omega_{1}=\frac{s \Delta}{2 \Delta_{\mathrm{b}}}$
is the Rabi frequency of the one photon absorption process.
To describe the regime of strong driving and $n$-photon absorption processes,
we apply the dressed-atom aproach to the flux qubit. The Hamiltonian is
given by $\hat{H} = \hat{H}_{\rm qb} +
\hbar \omega_{\mathrm{ex}} a^{\dagger}a + g \hat{\sigma}_{z}(a+a^{\dagger})$, where $a$ and $a^{\dagger}$ are the
annihilation and creation operators
for a field mode with angular frequency $\omega_{\mathrm{ex}}$ and $g$ is
a coupling constant between the qubit and the field.
%The first two terms represent the qubit Hamiltonian in the basis of the
%localized state $|\downarrow\rangle$ and $|\uparrow\rangle$.
The Hamiltonian can be explicitly diagonalized when $\Delta$=0. The eigenstates are given by
$|\uparrow(\downarrow);N \rangle_{\mathrm{dressed}}=\exp[-(+)g(a^{\dagger}-a)/\hbar \omega_{\mathrm{ex}}]
|\uparrow (\downarrow) \rangle |N \rangle$ and
their eigenenergies are $E_{N}^{\uparrow (\downarrow)} = N \hbar \omega_{\mathrm{ex}}-g^{2}/\hbar
\omega_{\mathrm{ex}}+(-)\frac{1}{2}\hbar \varepsilon_{0}$, where $|N \rangle$
are eigenstates of $\hbar \omega_{\mathrm{ex}} a^{\dagger}a$. For $\Delta \ll \varepsilon_{0}$,
first-order perturbation theory
in the qubit Hamiltonian yields the term linear in $\Delta$ and when
$n \hbar \omega_{\mathrm{ex}} \simeq \hbar \Delta_{\mathrm{b}}$,
the two dressed states show anticrossing due to
off-diagonal coupling $\frac{\hbar \Delta}{2} J_{n}(\alpha)$. Here,
$J_{n}(\alpha)$ is the $n$-th order Bessel function of first kind and
$\alpha \equiv 4g \sqrt{\langle N \rangle}/\hbar \omega_{\mathrm{ex}}$ is the scaled
amplitude of the driving field. If we prepare a localized initial state,
Rabi oscillations occur with frequency
$\omega_{n}=|\Delta| J_{n}(\alpha)$ \cite{Nakamura01}.
We also applied a second approach starting from a microscopic
driven spin-boson model with weak coupling to
an Ohmic bath \cite{Weiss,Hartmann00}.
In particular, we simplify the high-frequency approximation
by numerically solving the pole
equation (8) of Ref.~\cite{Hartmann00}. We find that the solutions
for the $n-$photon transition are given by
$\theta=\theta_0$ and $\theta=\theta_n$. Inserting this in the expression for
$P_\infty$, we obtain an expression for the HWHM$_n$ similar to that in
Eq.~(\ref{HWHM}), where the phenomenological dephasing and relaxation times
$T_{\phi/r}$ follow as the inverse of the weak-coupling rates
of the spin-boson model \cite{Weiss}. This result differs slightly from
that of the Bloch equations, which can be recovered by setting the
field-dressed level spacing
$\Delta_{0} \equiv \Delta J_{0}(\alpha) \approx \Delta$ in the second approach.
To analyze HWHM$_{n}$ and $A_{n}$ of the resonant dips, we substituted
$\omega_{n}$ for Eqs.~(\ref{HWHM}) and (\ref{Amplitude}). We took
$T_{\mathrm{r1}}$, $T_{\mathrm{r2}}$, $T_{\mathrm{r3}}$,
$T_{\phi n}=T_{\phi}$, and $c$ as fitting parameters. Here $T_{\mathrm{r}n}$
and $T_{\phi n}$ are the relaxation and dephasing times related to the $n$-th
photon absorption process. We defined a coupling constant $c$ as
$c I_{\mathrm{RF}}=\alpha$. We found excellent agreement with the experimental
data for both HWHM$_{n}$ and $A_{n}$ (see Fig.~\ref{fig3}).
It should be noted that the relaxation and dephasing times $T_{\mathrm{r}n}$
and $T_{\phi}$ do not depend on the frequency of the corresponding
multi-photon transition for a pure structure-less
Ohmic environment (see also Eqs.~(\ref{Bloch})). This is no longer the case
for a more complicated structured harmonic environment \cite{ChemPhys} as it
is present in our device. Here, the plasmon frequency of the dc-SQUID
provides an additional energy scale of the environment.
Nevertheless, the global physics is captured by our simplified model.
In conclusion, we have reported on measurements of macroscopic
superconducting circuits that reveal their quantum-mechanical behavior
at low temperature. We have identified multiphoton transition processes
in the qubit and found that the width of the $n$-photon resonance scales
with the $n$-th Bessel function with its argument given as the ratio of
the driving-field strength to the frequency of the photons.
Our results add another
example to the few cases of systems that exhibit clear
quantum-mechanical behavior at a macroscopic scale.
\begin{acknowledgments}
We thank J. E. Mooij, C. J. P. M. Harmans, M. Grifoni, I. Chiorescu,
Y. Nakamura, and D. Vion for useful discussions; T. Kutsuzawa
for experimental help.
This work has been suported by the CREST projiect
of Japan Science and Technology Agency (JST).
\end{acknowledgments}
\begin{thebibliography}{}
\bibitem{Leggett}A. J. Leggett, in {\em Chance and Matter\/}, Lecture Notes
of the Summer School of Theoretical Physics in Les Houches,
Session XLVI, 1986, eds. J. Souletie, J. Vannimenus and R. Stora
(North Holland, Amsterdam, 1987).
\bibitem{Grifoni}M. Grifoni and P. H\"anggi, Phys. Rep. {\bf304}, 229 (1998).
\bibitem{Weiss}U. Weiss, \textit{Quantum Dissipative Systems}
(World Scientific, Singapore, 2nd ed., 1999).
\bibitem{Bouwneester}D. Bouwmeester, A. Ekert, and A. Zeilinger,
\textit{The Physics of Quantum Information} (Springer, Berlin, 2000).
\bibitem{Nakamura and Vion}Y. Nakamura, and Y. A. Pashkin, J. S. Tsai,
Nature {\bf398}, 786 (1999); D. Vion,
{\em et al.,\/}
% A. Aassime, A. Cottet, P. Joyez,
% H. Pothier, C. Urbina, D. Esteve, and M. H. Devoret,
Science {\bf296}, 886 (2002).
\bibitem{Yu and Martinis}Y. Yu,
{\em et al.,\/}
%S. Han, X. Chu, S.-I. Chu, Z. Wang,
Science {\bf296}, 889 (2002); J. M. Martinis,
{\em et al.,\/}
%S. Nam, J. Aumentado, C. Urbina,
Phys. Rev. Lett. {\bf89}, 117901 (2002).
\bibitem{Mooij}J. E. Mooij,
{\em et al.,\/}
% T. P. Orlando, L. Levitov, L. Tian,
% C. H. van der Wal, and S. Lloyd,
Science {\bf285}, 1036 (1999).
\bibitem{Wal}C. H. van der Wal,
{\em et al.,\/}
%, A. C. J. ter Haar, F. K. Wilhelm,
% R. N. Schouten, C. J. P. M. Harmans, T. P. Orlando, S. Lloyd,
% and J. E. Mooij,
Science {\bf290}, 773 (2000).
\bibitem{Chiorescu}I. Chiorescu,
{\em et al.,\/}
%Y. Nakamura, C. J. P. M. Harmans, and J. E. Mooij,
Science {\bf299}, 1869 (2003).
\bibitem{Tanaka1}H. Tanaka, S. Saito, T. Kutsuzawa, M. Ueda, H. Nakano,
K. Semba, and H. Takayanagi, in preparation.
\bibitem{Makhlin}Y. Makhlin,
{\em et al.,\/}
%G. Sch\"on, and A. Shnirman,
Rev. Mod. Phys. {\bf73}, 357 (2001).
\bibitem{Shapiro}S. Shapiro, Phys. Rev. Lett. {\bf11}, 80 (1963);
W. C. Danchi,
{\em et al.,\/}
%F. Habbal, and M. Tinkham,
Appl. Phys. Lett. {\bf41}, 883 (1982).
\bibitem{Devoret}M. H. Devoret,
{\em et al.,\/}
%J. M. Martinis, D. Esteve, and J. Clarke,
Phys. Rev. Lett. {\bf53}, 1260 (1984).
\bibitem{Wallraff}A. Wallraff,
{\em et al.,\/}
%T. Duty, A. Lukashenko, and A. V. Ustinov,
Phys. Rev. Lett. {\bf90}, 037003 (2003).
\bibitem{Leggett02}A. J. Leggett, J. Phys. {\bf14}, R415 (2002).
\bibitem{Nakano}H. Nakano, H. Tanaka, S. Saito, K. Semba, H. Takayanagi,
and M. Ueda, in preparation.
\bibitem{Takayanagi}H. Takayangi, H. Tanaka, S. Saito, and H. Nakano,
Phys. Scr. \textbf{T102}, 95 (2002); S. Saito, H. Tanaka, H. Nakano,
M. Ueda, and H. Takayanagi, in \textit{Quantum Computing and
Quantum Bits in Mesoscopic Systems}, edited by A. J. Leggett,
B. Ruggiero, and P. Silvestrini (Kluwer, New York, 2004), p. 161.
\bibitem{Tannoudji}C. Cohen-Tannoudji, J. Dupont-Roc, and G. Grynberg
\textit{Atom-Photon Interaction}
(Wiley, New York, 1992).
\bibitem{Goorden1}M. C. Goorden and F. K. Wilhelm, Phys. Rev. B {\bf 68},
012508 (2003).
\bibitem{Hartmann00}L. Hartmann,
{\em et al.,\/}
%I. Goychuk, M. Grifoni, and P. H\"anggi,
Phys. Rev. E {\bf 61}, R4687 (2000).
\bibitem{Goorden}M. C. Goorden, Master's Thesis,
(TU Delft, Delft, 2002).
\bibitem{Nakamura01}Y. Nakamura, Y. A. Pashkin, and J. S. Tsai,
Phys. Rev. Lett. {\bf87}, 246601 (2001).
\bibitem{ChemPhys}M. Thorwart, E. Paladino, and M. Grifoni,
Chem. Phys. {\bf 296}, 333 (2004).
% (Special Issue on The Spin-Boson Problem:
% From Electron Transfer to Quantum Computing,
% ed. by H. Grabert and A. Nitzan).
\end{thebibliography}
\end{document}