Time Reversal of electromagnetic waves
G. Lerosey, J. de Rosny, A. Tourin, A. Derode, G. Montaldo, M. Fink
Laboratoire Ondes et Acoustique, ESPCI, Université Paris VII, UMR 7587,
10 rue Vauquelin, 75005 Paris, France
We report the first experimental demonstration of time-reversal focusing with electromagnetic waves. An antenna
transmits a 1-µs electromagnetic pulse at a central frequency of 2.45 GHz in a high-Q cavity. An other antenna records
the strongly reverberated signal. The time-reversed wave is built and transmitted back by the same antenna acting now
as a time-reversal mirror. The wave is found to converge to its initial source and is compressed in time. The quality of
focusing is determined by the frequency bandwidth and the spectral correlations of the field within the cavity.
In acoustics, time reversal experiments can be carried
out with broadband wave forms [1]. In such
experiments, a source sends a short pulse that
propagates through a more or less complex (but ideally
non dissipative) medium and is captured by a
transducer array, termed a Time Reversal Mirror
(TRM). The recorded signals are digitized, stored in
electronic memories, time-reversed, re-analogized and
finally transmitted back by the TRM. The time-
reversed wave is found to converge back to its source
all the more accurately than the medium is complex
and the frequency bandwidth is larger [2]. This is very
appealing for applications in telecommunications.
Indeed, it was recently shown with ultrasonic waves
that it is possible to take advantage of the complexity
of a medium to convey more information through it by
means of a TRM [3]. From a practical point of view, a
TRM was used to focus random series of bits
simultaneously to different receivers which were only
a few wavelengths apart. In the language of
communication, it corresponds to a MIMO-MU
configuration (Multiple Input, Multiple Output –
Multiple Users). While the transmission was free of
error when strong multiple scattering occured in the
propagation medium, the error rate was huge in the
homogeneous medium (free space) due to crosstalk
between receivers. Indeed, the spatial resolution of a
TRM can be much thinner in a multiple scattering
medium than in free space. It is the well-known
“super-resolution” that has been experimentally
highlighted [2,4] and theoretically discussed [2,5] in
the past.
Is it possible to transpose this idea to the
electromagnetic case? It is a challenging question
because in many real environments (buildings or
cities), microwaves with wavelengths between 10 and
30 cm are scattered by objects such as walls, desks,
cars, etc which produces a multitude of paths from the
transmitter to the receiver. In such situations, a time-
reversal antenna should be able not only to compensate
for these multipaths, but also increase the information
transfer rate thanks to the many
reflections/reverberations, as it was already shown
with ultrasound [3]. However, the first step is to prove
the feasability of a time reversal experiment with
electromagnetic waves in the GHz range. It is the aim
of this letter. To that goal, we present the first one-
channel electromagnetic time reversal mirror (TRM)
working around 2.45 GHz.
Fig. 1 Sketch of a transmit/receive experiment.
The baseband signal m
I
(t) (resp. m
Q
(t) ) is fed into a
mixer that multiplies it by cos(2πf
0
t) (resp. sin(2πf
0
t)),
with f
0
=2.45 GHz. After amplification, the output rf
signal is sent by antenna A , which is located in the
reverberating room. The reverberated field is recorded
by antenna B. Its spectrum is downshifted to the
baseband by a multiplier and a low-pass filter. The
resulting signals m’
I
(t) and m’
Q
(t) are sampled and
stored in the computer memory. Then they can be time-
reversed, re-analogized and, following the same principle,
sent back by antenna B while antenna A acts now as a
receiver.
From a practical point of view, the main
difficulty to transpose the time reversal technique
developed for ultrasound directly to the
electromagnetic case lies in the much higher sampling
frequencies that are needed to digitize radiofrequency
(RF) signals. One way to overcome this limitation is to
work only with quasimonochromatic signals and to do
a phase conjugation using the so-called three-wave or
four-wave mixing in a non linear material in order to
naturally produce the analogic phase-conjugated wave
[6]. Here, we want to perform a truly broadband time
reversal for an electromagnetic pulse m(t)cos(2
π
f
0
t+
ϕ
),
with f
0
the carrier frequency and m(t) a lower
frequency signal (“baseband signal”). The time-
reversal operation can be decomposed in two steps:
phase conjugation of the wave carrier f
0
, plus time-
reversal of the baseband signal m(t). The advantage is
that the sampling of the baseband signal requires a
sampling frequency much lower than 2 f
0
.
Fi
Fig. 2. (a) and (a’): baseband representation (resp. m
I
(t)
and m
Q
(t)) of the signal transmitted by antenna A. (b)
and (b’): baseband representation (i.e. m’
I
(t) and m’
Q
(t))
of the signal reverberated inside the cavity and received
by antenna B.
The experiment takes place in a strongly
reverberant cavity with dimensions 3.08 m × 1.84 m ×
2.44 m. Using an arbitrary waveform generator, we
deliver a short pulse m
I
(t) (central frequency 3 MHz, –
6dB bandwith 2 MHz) to the I analog input (cf. Fig.
2.(a)) of the transmit board. No signal is delivered to
the Q analog input (Fig. 2.(a’)). A mixer upconverts
this signal to the GHz band and delivers
e(t)=m
I
(t)cos(2
π
f
0
t). Then the waveform e(t) is
transmitted by antenna A. After propagation, the signal
s(t)=m’
I
(t)cos(2
π
f
0
t)+ m’
Q
(t)sin(2
π
f
0
t) is recorded by
antenna B and downconverted to produce the I and Q
components of the output signal m’
I
(t) and m’
Q
(t) that
can be observed at the oscilloscope ( Fig. 2. (b) and
(b’) ). The received signals last more than 8 µs, i.e.
eight times longer than the initial baseband pulse, or ~
3200 periods of the RF wave. The RF wave has
travelled more than two kms in the 14-m
3
cavity, i.e.
undergone ~ 6000 reflections.
Next, our goal is to time-reverse the received
radio signal
. To that end, the baseband I and Q
signals m’
)(ts
I
and m’
Q
are digitized by the oscilloscope at
a 40-MHz sampling rate, sent to a computer and time-
reversed. The wave carrier has to be conjugated too.
The following step consists in re-analogizing the time-
reversed I and Q signals, and encoding them on the
phase-conjugated wave carrier: the resulting RF signal
writes m’
I
(-t)cos(2
π
f
0
t)-m’
Q
(-t)sin(2
π
f
0
t) = s(-t). It is
then transmitted back by antenna B. After propagation,
the RF signal received on antenna A is downconverted
to baseband. As can be seen in Fig. 3.(a), the received
signal on channel I is compressed in time and recovers
its initial duration.
Fig. 3. (a) and (a’) : baseband representation of the signal
received by antenna A after time-reversal. (b) and (b’):
baseband representation of the signal received several
wavelengths away from antenna A after time-reversal.
Actually, since the reverberated wavefield has
been captured by a single antenna, the time-reversal
operation is not perfect. The waveform that is recreated
is not the exact replica of the initial pulse: there are
sidelobes around the peak on channel I, and a signal is
measured on channel Q (Fig. 3. (a’) ) although nothing
was sent on that channel. Similar effects have been
observed for time-reversal of ultrasonic waves: it can
be shown that the peak-to-noise ratio in a one-channel
time-reversal experiments varies as
ff
δ
/∆ where
∆
f
is the available bandwidth and
δ
f defines the
correlation frequency of the reverberated field [2].
δ
f is
the characteristic width of the field-field correlation
function
() ( )
dffff
∫
+
∗
δψψ
, with
ψ
the scattered
electromagnetic field. Therefore one could expect an
even stronger pulse compression if the bandwidth was
larger, or the correlation frequency smaller. Given the
dimensions of the cavity, the Heisenberg time (i.e. the
We have presented the first time reversal
experiment for electromagnetic waves in the GHz
domain. The experiments were carried out in a
reverberating cavity. To avoid digitizing the radio
signals at GHz frequencies, we have time reversed the
baseband signals and phase-conjugated the wave
carrier. Thanks to reverberation, the resulting signal at
the source is compressed in time and in space with a
signal-to-noise ratio depending on the ratio of the
bandwidth to the frequency correlation of the medium.
In the future, a larger bandwidth has to be used to
improve both spatial and temporal focusing. This
seems to be possible with the emergence of ultra wide
band electromagnetics components. Thus time-reversal
focusing could have promising applications in the field
of wide-band wireless communications in complex
reverberant environments.
inverse of the mean distance between eigenmodes) is
t
H
~ 80 µs. But the characteristic absorption time is t
a
=
3.6µs , therefore the modes are not resolved and the
correlation frequency
δ
f is determined by t
a
rather than
t
H
. Taking andMHzf 2=∆ kHztf
a
280/1 ≈=
δ
leads
to a predicted peak-to-noise ratio roughly equal to 3,
comparable to our experimental results.
The experiment shows that time reversal is
able to compensate for multiple reverberation and
recreate a short electromagnetic pulse at the source.
But there is more to it: we have also verified that the
amplitude of the re-created signal is stronger at
antenna A than anywhere else in the cavity ( Fig. 3. (b)
and (b’) ), i.e. the time reversed wave is spatially
focused. This is consistent with past experiments using
ultrasound in reverberant or strongly scattering media
that proved that even with a one-channel time-reversal
device, the pulse is sharply focused at the source and
only at the source, the noise level surrounding the
source being, once again, controlled by
ff
δ
/∆ [2].
Here, the role of reverberation as well as the large
frequency band must be emphasized. Since we only
have a one-channel time-reversal, no focusing would
occur in free space no matter how large the bandwidth;
in a reverberating medium, no focusing would occur
either if the bandwitdh was too narrow, as is the case
in classical phase-conjugation techniques. Indeed,
imagine a single time-reversal antenna trying to focus
a pure monochromatic wave on some receiver; the
phase-conjugated wave has no reason at all to be
focused on this point since the antenna only sends back
a sinusoidal spherical wave through the medium. But if
the frequency bandwidth ∆f is much larger than the
correlation frequency
δ
f, then the spectral components
of the reverberated field
ψ
at two frequencies apart by
more than
δ
f are decorrelated: there are roughly ∆f/δf
decorrelated components in the spectrum of the
reverberated signals. When we time-reverse (i.e.
phase-conjugate coherently all along the bandwidth,
and not just at one frequency) all these components,
they add up in phase at t=0 at the receiver position,
because all the phases have been set back to 0 all along
the bandwidth. Thus, the amplitude at this position
and at this time increases as ∆f/δf whereas outside the
receiver position (and outside t=0), the various
frequency components add up incoherently and their
sum rises as
ff
δ
/∆ . On the whole, the peak-to-
noise ratio increases as
ff
δ
/∆ as the bandwidth is
enlarged. It explains why a time-reversal experiment is
possible even with one single TR channel [2,5,7], as
long as the frequency bandwidth is sufficiently larger
than the correlation frequency of the propagation
medium.
ACKNOWLEDGMENTS
The authors wish to acknowledge the
“Département de Recherche en Electromagnétisme,
Laboratoire Signaux et Systèmes, Supélec”, Gif-sur-
Yvette, France (www.lss.supelec.fr) and particularly A.
Azoulay and V. Monebhurrun for having let us use
their reverberant cavity, as well as the “Conseil
Régional d’Ile de France” and the “Conseil
Départemental de l’Essonne” who financed the cavity.
This work is a part of the research projects developed
within the “Groupement de Recherche” ImCoDe
(GDR 2253, CNRS,
http://lpm2c.grenoble.cnrs.fr/IMCODE/IMCODE.html
).
REFERENCES
1. M. Fink, Physics Today 50 , 34 (1997).
2. A. Derode, A. Tourin and M. Fink, Phys. Rev. E
64, 036606 (2001).
3. A. Derode, A. Tourin, J. de Rosny, M. Tanter, S.
Yon, M. Fink, Phys. Rev. Lett. 90, 014301-1 (2003).
4. A. Derode, P. Roux, M. Fink, Phys. Rev. Lett. 75,
4206 (1995).
5. P. Blomgren, G. Papanicolaou, H. Zhao, J. Acoust.
Soc. Am., 111, 203 (2002).
6.
D.M. Pepper, Opt. Eng., 21, 156 (1982). Yian
Chang, H. R. Fetterman, I.L. Newberg, and S.K.
Panaretos
, Appl. Phys. Lett. 72, 745 (1998).
7. C. Draeger, M. Fink, Phys. Rev. Lett. 79, 407
(1997).
