San Francisco
IN A padded room deep inside an old Parisian building, Mathias Fink is eating
his words. Fink stares at an array of microphones and loudspeakers that are
recording his voice and throwing it back at him in reverse. The effect is
curiously like running a film backwards. The reversed words sound as if they are
tumbling out of Fink鈥檚 mouth, while the speakers seem to be silent鈥攍ike a
bizarre high-tech ventriloquist act.
But this isn鈥檛 show business. Fink is the director of the waves and acoustics
laboratory at the Ecole Sup茅rieure de Physique et de Chemie in the Latin
Quarter of Paris and it is in his laboratory that this extraordinary
demonstration is taking place. Fink and his colleagues have pioneered a
technique that can record sound waves and project them back exactly to their
point of origin. The array of speakers and microphones is called a time-reversal
mirror and if the technology lives up to its promise, it could be used in a
dizzying range of areas from destroying tumours and gall bladder stones to
detecting defects in metals. It could even allow long-distance conversations
with the crews of submarines deep beneath the ocean surface.
Advertisement
The idea behind time-reversal mirrors is simple. Imagine a sound wave
propagating away from its source, like ripples created by a stone hitting the
surface of a pond. Now play this mental video in reverse, so that the waves
converge on the source instead of moving away. In theory, it is possible to
pinpoint the position of any acoustic source simply by time-reversing the
sounds it produces.
But time reversal can only occur under certain conditions. The process must
be adiabatic, for example. This means that the waves must move without losing
energy as heat since these losses could never be replaced. Another condition is
that the path the wave travels from the source to the receiver must be the same
as the path taken from the receiver to the source, a condition known as spatial
reciprocity. This may sound trivial, but waves in rotating fluids or materials
that strongly dissipate waves rarely satisfy this condition. However, acoustic
waves in air and water generally satisfy both conditions, says Fink.
Exploding cube
Even so, time reversal is a very difficult process. Fink demonstrates the
problems with a thought experiment. Imagine a small cube exploding in slow
motion with the debris particles racing away from the explosion at various
velocities. Now imagine a time-reversal mirror that could reconstitute the cube
exactly. How would such a device work?
In his thought experiment, Fink conjures up a device that reverses the
velocity of each debris particle as it crosses a closed surface surrounding the
cube. But simply reflecting the particles at this surface will not reconstitute
the cube. To rebuild the block, the slowest particles to arrive at the surface
must be sent back first. But these cross the surface last. So the return of
faster particles has to be delayed by an appropriate period of time.
Of course, constructing such a time-reversal mirror would be impossible.
Apart from the practical difficulties of building such a device, it would never
be possible to measure each particle鈥檚 motion with enough precision, says Fink.
And because the equations governing the return journey are chaotic, any error
renders time reversal impossible. But waves are a different kettle of fish.
An acoustic wave can be characterised entirely by measuring the way pressure
changes as the wave passes by. What鈥檚 more, any variations in the pressure occur
over distances of roughly the same order of magnitude as the wavelength. This is
extremely important since the measurements required to reconstruct the wave need
only be carried out on the scale of its wavelength. By comparison, there is
almost no limit to the accuracy with which the velocities of the debris from an
exploding cube must be measured if it is to be reconstructed.
Measuring the pressure created by a wave is simple because there is no need
to entirely surround the source. Instead, simply measure pressure changes over a
small angle using an array of sensors. As long as the gap between these sensors
is no more than half the wavelength of the wave to be measured, the wave can be
reconstructed entirely. The sensors must record the information digitally as the
entire wave passes by. Only when all the data from the wave have been recorded
can the process of time reversal take place.
Reverse replay
For the time-reversed sounds to emanate from exactly the spot where they were
recorded, the sensors must double up as speakers. Fink uses piezoelectric
transducers that convert pressure changes, or sound, into a voltage when the
wave passes and then produce pressure changes when a voltage is applied across
them. Creating the time-reversed sounds is simple, says Fink. 鈥淚t鈥檚 like playing
a tape recorder backwards.鈥�
Nevertheless, putting all these ideas together is a tricky business. Each
sensor has to pass its data to a converter that changes the analogue signal to a
digital one before storing it. The signal then has to be inverted when the time
reversal takes place. This task is straightforward with one signal, but when an
array consists of more than 100 sensors, each carrying out this process many
thousands of times a second, the amount of data involved is formidable.
Then there are the practical problems to overcome. In theory, each sensor
should record the wave at a single infinitely small point. But the sensors are a
finite size, so their measurements are an average of many points. Errors creep
in as a result, preventing the reversed sound from returning exactly to its
source.
To stop his happening, Fink keeps his sensors as small as possible. He also
squeezes the waves digitally, so that any small errors become even smaller when
the time reversal occurs. 鈥淭here are a lot of tricks to this,鈥� he says.
With many of the problems ironed out, Fink is eager to develop applications.
So far, he has concentrated on the problem of how to remove kidney and gall
bladder stones effectively. The current method relies on using one ultrasound
beam to locate a stone and another to destroy it. The rapid changes in pressure
created by the ultrasound beam break up the stone. To increase the potency of
this mechanism, doctors use several beams so that the pressure changes are
highest where they cross.
But this process is somewhat hit and miss. Just by breathing, the patient can
move the stone by as much as 2 centimetres. Fink estimates that the ultrasound
is off-target in up to 70 per cent of cases, leading to internal bleeding and
discomfort for the patient.
Time-reversal technology could change all that, says Fink. The process would
occur in three stages. First, part of the array generates a brief ultrasonic
pulse to illuminate the target. A stone will be more dense than its
surroundings, so it will reflect the pulse more strongly, sending a wavefront
back towards the array.
Direct hit
The array then time-reverses this signal and the re-emitted wave focuses on
the stone and destroys it. This process is then repeated thousands of times a
second. The advantage is that the focused beam almost never misses its target
and so damage to the surrounding area is minimal. 鈥淭he waves have no choice but
to return to the stone,鈥� says Fink.
Fink has already given his apparatus a trial run in two hospitals in Paris
and Lyon. The results were a great success but the French company contracted to
build the machines went bankrupt and the technology has yet to go into
production. 鈥淲e have a set of patents, but now we have to find another company,
which has delayed its commercialisation,鈥� laments Fink.
A similar technique could also be used to destroy brain tumours. Today, small
brain tumours must be removed by surgery since other methods of destruction are
too coarse. However, Fink believes it will be possible to destroy tumours using
time-reversal mirrors. The trick here is to focus the beam so tightly that it
heats the target tissue to 60 掳C or more, destroying it. For this method to
work, the spot size must be tiny less than 1.5 millimetres across.
The task is complicated by the fact that sound travels at different
velocities through the skull and brain. 鈥淭his creates a mirage effect,鈥� says
Fink. He and his colleagues have had to develop special focusing techniques to
get round this. He hopes to begin animal trials within six months.
The medical industry isn鈥檛 the only one that could benefit from Fink鈥檚 work.
His time-reversal mirrors have also caught the imagination of American engineers
and physicists working on the problem of communicating with submarines beneath
the sea. Conventional radio waves do not pass through water, so they cannot be
used to communicate with a submarine. Sound waves can鈥檛 do the job either
because they disperse too quickly. As any child in a swimming pool soon figures
out, talking underwater is a muffled affair.
Multiple distortions
In the ocean, things get even more complicated. David Dowling at the
University of Michigan and Darrell Jackson at the University of Washington in
Seattle are studying the way sound propagates in the ocean. They say that
currents, changes in temperature and salinity and reflections off the surface
and seafloor create multiple distortions and reflections. These effects are not
even constant. 鈥淥ver a period of minutes or hours the internal structure of the
ocean changes radically because of the tide and current,鈥� says Dowling.
But time-reversal mirrors could overcome these problems. 鈥淭he spark of the
idea was the discovery that time-reversed acoustic waves could focus,鈥� says
Dowling, a mechanical engineer. This might allow acoustic communication between
a submarine and the surface. The inability to communicate with submerged
submarines has been a thorny problem since the First World War. 鈥淓ssentially,
you could say they鈥檙e not communicating,鈥� says Jackson. 鈥淭he US Navy is really
蝉迟耻尘辫别诲.鈥�
Recently, researchers from NATO鈥檚 SACA laboratory, in Lastezia, Italy, and
the Scripps Institution of Oceanography in La Jolla, California, tried using the
technique in the Mediterranean off the island of Elba. Led by William Kuperman
and William Hodgkiss from Scripps, they demonstrated that a time-reversal mirror
could compensate for the problems of currents and internal reflections and focus
sound waves at distances of up to 30 kilometres. 鈥淭he military applications are
really interesting,鈥� says Fink.
Time-reversal mirrors could be also used to locate very small defects in
metal. This is especially important for aircraft safety, since small defects can
cause catastrophic failure. Currently, defects are detected using ultrasound
beams, which have a poor resolution, or X-rays, which are difficult to handle
safely. What鈥檚 more, these methods can damage the very material being
tested.
Crack detector
Time-reversal mirrors will be able to find a defect in the metal, says Fink,
because the defect reflects pressure waves in the same way as a kidney stone.
Fink has already begun trials for the French organisation the National Society
for the Study and Construction of Aircraft Engines, which is based in Paris. His
team has produced a 128-element time-reversal mirror array designed to
detect extremely low-contrast defects that are only fractions of a millimetre
across鈥攖oo small for any conventional technique to pick up. Similar
machines could scan the steel walls of nuclear power plants looking for defects
or cracks.
Back in the time-reversal room in Fink鈥檚 lab, acoustic researchers keep busy
eating their words. Even if these applications never come to pass, Fink reckons
the technology has all the makings of a great show-biz voice throwing act. Even
so, it鈥檚 unlikely Fink and his team will need to give up their day jobs anytime
soon.
