PHILIP RUSSELL鈥檚 professional life is riddled with holes, but that鈥檚 the way
he likes it. Ask him the secret of his remarkable light-conducting fibres and
he鈥檒l tell you it is not the material from which they are made, but the lack of
it. Where there should be solid silica glass鈥攁s in a conventional optical
fibre鈥擱ussell and his team at the University of Bath have left dozens of
tiny holes. Strangely, the holes are there to keep the light in.
Russell鈥檚 fibres are an impressive piece of engineering鈥攖he holes are
just tens of nanometres wide and run perfectly for hundreds of metres. This
unique design lets them carry far more light than conventional fibres, so in
theory they could boost the capacity of fibre-optic networks and even help to
create powerful 鈥減ocket lasers鈥� for use by surgeons to cauterise tissue or by
astronauts to repair orbiting spacecraft. The fibres also let you play some
remarkable tricks, such as making light leap from one fibre to another, or
squirting atoms through a 鈥渉osepipe鈥� of light鈥攖ricks that are vital for
building high-speed optical computers or super-sensitive gravity detectors.
Conventional optical fibres carry light through a glass core surrounded by a
cladding layer鈥攁 bit like the insulation on an electrical cable. In most
optical fibres both the core and cladding are made from silica glass, but the
core has slightly different optical properties, thanks to a sprinkling of atoms
known as dopants, usually germanium or phosphorus. These atoms raise the
refractive index of the glass so that as light travels through the core, it is
reflected from the interface with the cladding by a process called total
internal reflection, and is trapped.
Advertisement
Russell鈥檚 fibres are very different. Core and cladding are still made of
silica glass, but there鈥檚 no need for dopants: instead the fibre has tiny holes
running through it. These make the perforated glass impermeable to light.
Why should holes prevent light from passing through an otherwise transparent
material? If the holes are arranged in a regular array with a size and
separation of roughly the same magnitude as the light鈥檚 wavelength, they act as
a picket fence that stops light from passing through. An orderly array of
particles or gaps that excludes light in this way is called a photonic crystal
(鈥淭ricks of the light鈥�, 快猫短视频, 26 August 1995, p 26).
Photonic crystals were predicted in 1987 by two physicists working
independently: Eli Yablonovitch of Bell Communications Research in Red Bank, New
Jersey, and Sajeev John of the University of Toronto in Canada. Light passing
through a field of obstacles, such as a suspension of particles in water, is
scattered if the size of the particles is similar to the wavelength of the
light. The researchers realised that if the particles were arranged in a regular
lattice, they would scatter incident light of the right wavelength back the way
it came, like a mirror. Such crystals have a photonic 鈥渂and gap鈥濃�攁 band of
forbidden wavelengths which they will not allow to pass.
Instead of filling empty space with a regular array of objects, you could
make a photonic crystal by peppering a solid block of material with a regular
array of holes.
Photonic crystals offer more than just fancy mirrors: they can confine and
guide light too. Disrupt the regularity of photonic materials and you open up a
鈥渃rack鈥� through which light can squeeze. Missing out a row of holes in a
perforated sheet, for example, provides a channel through which light can
travel.
This is the principle that Russell and his colleagues, Jonathan Knight and
Tim Birks, use to make their photonic crystal fibres. Their strands, each
thinner than a human hair, are laced with holes running their length.
But through the centre of the fibres runs a solid, light-conducting
channel confined by the surrounding photonic crystal. They have made fibres
hundreds of metres long, with tiny channels running their entire length. If the
smallest of these channels were scaled up to the width of the Channel Tunnel, it
would stretch from here to Jupiter.
Whisker thin
To create this perfect array, the researchers used a neat trick: they packed
together narrow glass tubes to make a bundle about two centimetres across. Like
logs stacked in a woodpile, the tubes arrange themselves in a hexagonal fashion.
To convert this hefty bundle of tubes into a perforated thread three hundredths
of a millimetre across, the researchers simply heated the bundle to 2000 掳C
to make the glass soft and viscous, and then stretched it until it was as thin
as a whisker. This preserved the holey cross-section, but shrunk it in scale by
a factor of about a thousand. At this temperature, the glass becomes soft enough
to ooze into and fill up the very narrow spaces between adjacent tubes whereas
their hollow interiors remain. The result: a flexible fibre riddled with
channels too small to see.
However, these fibres, first fabricated three years ago, don鈥檛 operate by
setting up a full photonic band gap in the cladding layer. Rather, Russell and
his colleagues believe that the picket fence of holes acts like a material with
a different refractive index from that of the solid glass core, so the light is
confined by total internal reflection.
But they鈥檝e done far more than create a conventional fibre using
unconventional means. They鈥檝e discovered that no matter how big they make their
photonic crystal fibre, it behaves like a conventional fibre with a core just a
few micrometres across. This gives their fibre some huge advantages.
Make the core of a conventional fibre too wide and light travelling along it
can follow several paths. This sets up patterns or 鈥渕odes鈥� in the
light鈥攍ike the patterns of vibration on a drumhead. A fibre with a core
more than about 20 micrometres across can support several of these modes and is
called a multimode fibre. Send a narrow signal pulse along a multimode fibre and
it smears out as it travels, blurring the information content and limiting the
amount of data that it can carry.
One solution is to use fibres with cores just a few micrometres across, as
light can only follow a single pathway through them. But these 鈥渟ingle mode鈥�
fibres are so narrow that you can鈥檛 fit much light into them.
On the other hand, you can make a photonic crystal fibre as wide as you like
and it will still behave like a single-mode fibre. The 鈥渇undamental鈥� mode
remains trapped, but all the light following other pathways leaks out of the
core by 鈥渟queezing鈥� between the holes, so light pulses don鈥檛 smear out as they
travel along the photonic crystal fibre.
These fibres are also made of pure silica, without the dopants which can
absorb light, particularly at blue and ultraviolet wavelengths, and reduce the
intensity of the signal bouncing through the fibre. In principle the photonic
crystal fibre should behave as a single mode fibre for light from the
ultraviolet to the infrared parts of the spectrum鈥攁 feat unmatched by any
other fibre.
And because the fibre core can be made wider than conventional single-mode
fibres, you can squeeze far more light down it鈥攗p to 20 times
more鈥攐pening up new possibilities for delivering high-power laser light to
cauterise tissue in microsurgery, for instance, or to slice up materials with
great precision for engineering.
Higher power also makes the fibres ideal for telecommunications, since it
means that fewer amplifiers will be needed along the cable to boost the signal
for long-distance transmission. Optical signals must be amplified at regular
intervals along the transmission line to compensate for losses through
absorption and scattering. This creates technical problems when cables are laid
on the seabed, for example. But greater power means the signals can survive for
longer between boosts, and because amplifiers are prone to breakdown, fewer of
them means fewer costly repairs of undersea cables.
All of this represents a radical departure from the way light is normally
guided around in optical fibres. 鈥淩ussell鈥檚 photonic crystal fibres have
breathed new life into a well-established field,鈥� says Douglas Allan of Corning,
a leading glass technology company based in New York. 鈥淭heir work is an
extremely valuable source of new ideas for controlling light in fibres,鈥� he
adds.
Totally new
Late last year, the Bath group went one step further. It reported a photonic
crystal fibre in which the core was not solid silica but an extra hole (
Science, vol 282, p 1476). The researchers believe that, whereas in their
solid-core fibres the light was still confined by total internal reflection, the
hollow-centred fibres trap the light because there is a true photonic band gap
in the cladding鈥攁 totally new principle of operation for an optical
fibre.
Nevertheless, in their new fibre the light is confined inside the glass that
surrounds the central hole, rather than in the hole itself. Russell would dearly
love to create a fibre in which the light passes down the hollow
core鈥攖hrough air instead of silica. There would be almost no absorption or
scattering and the power capacity would be vastly increased. This, says Russell,
is his team鈥檚 ultimate aim, the target that he originally envisaged back in
1992. 鈥淲e鈥檙e not there yet,鈥� he says, 鈥渂ut we鈥檙e close.鈥�
Ultimately, he sees such fibres being used as the amplifying cavity of a
laser. Fibre lasers are already in use, but if they pumped out more power they
would be far more useful. The power produced by a hollow-core fibre that carries
light through the hole could be at least ten times as great as that from a
silica-core fibre, upgrading 50-watt laser beams into the kilowatt region.
鈥淭here are many applications waiting in the wings for the world鈥檚 first
1-kilowatt single-mode fibre laser,鈥� suggests Russell. They might be used, for
example, in cutting and welding tools for repairs to orbiting spacecraft鈥攁
kind of pocket light sabre.
Holey fibres can also be used to process light. Simply fill the hollow
channels with a gas or liquid, and laser light would interact with it
continuously as the light bounces down the fibre. A potential product of
鈥渇illed鈥� fibres is an all-optical switch. Such switches exist already, but the
new fibres could overcome several of their limitations. All-optical switching
allows light signals to be rerouted without the need for electronic control,
which increases speed and reduces signal loss. It exploits the 鈥渙ptical Kerr
effect鈥濃�攚hen an intense burst of light triggers an increase in a
material鈥檚 refractive index. This can cause a pulse of light to jump tracks:
above some threshold intensity, it leaps from one glass fibre to another when
the two run side by side. In other words, a burst of light can be used to flick
this switch in an optoelectronic circuit.
But the optical Kerr effect is very weak in glass鈥攊ts 鈥淜err
coefficient鈥� is small鈥攕o the light pulse has to travel over a long
distance through the fibre before the cumulative effect is big enough to induce
switching. In the holey fibres the effect could be made much stronger by filling
the channels with a liquid such as methanol that has a higher Kerr coefficient.
Then, says Russell, it might be possible to achieve switching over one metre of
fibre rather than, as is typical at present, over half a kilometre of coiled
glass fibre.
Perhaps most intriguing of all is the possibility of using the channels in a
photonic crystal fibre as a pipe of light along which cold atoms could be sent.
Researchers can use the momentum of photons in a laser beam to pick up small
objects such as cells and move them about
(鈥淪pin doctors鈥�, 快猫短视频, 14 February 1998, p 34).
Atoms can be moved about in a similar way.
Tuning the energy of the light to one side of an element鈥檚 atomic
absorption line forces these atoms to move into the bright areas of the beam;
tuning it to the other side pushes them out of the light beam.
So Russell鈥檚 team is collaborating with Kishan Dholakia and coworkers at the
University of St Andrews in Scotland to use laser light and the hollow channels
in the fibre to confine and guide atoms鈥攁 sort of atom hosepipe. The idea
is simple: create a vapour of cold atoms and use light to guide the atoms one by
one into the channels. Then as the laser beam bounces along the core, it pushes
the atoms down the fibre. The atoms are attracted towards the charged glass
walls of the channel, but the laser beam nudges them back to the centre of the
hole. 鈥淭hey sort of rattle along the channels,鈥� says Dholakia.
Potential applications range from the separation of elements or
isotopes鈥攖uning the laser light to the right frequency will selectively
transport specific atoms or isotopes down the pipe鈥攖o ultra-sensitive
gravity meters, which would spot changes in gravity using interference. If you
cool atoms to a whisker above absolute zero, they display their wavelike nature.
When these 鈥渁tom waves鈥� meet, they generate interference fringes in much the
same way that light beams can, but since atoms have more mass than photons, the
spacing of these atom fringes is acutely sensitive to the gravitational field in
which the atoms sit. Use atom hosepipes to bring beams of cold atoms
together鈥攁s is done with light in an optical interferometer鈥攁nd you
can make measurements of minute changes in gravity. This could be useful for
investigating the structure of the deep Earth, suggests Dholakia, or for
detecting tiny changes in sea level.
Although all sorts of photonic materials are under development, most remain a
long way from the market. Russell鈥檚 fibres, on the other hand, could be put to
use much sooner. And they should be easy to manufacture in bulk, says physicist
Shawn Lin at Sandia National Laboratories in Albuquerque, New Mexico: 鈥淭his
means they should be cheap to produce.鈥� Several multinational companies involved
in laser design, communications and laser machining have already beaten a path
to Russell鈥檚 door. 鈥淕iven the level of interest,鈥� says Russell, 鈥渢here is a good
chance that the photonic crystal fibre will be commercialised within the next
few years.鈥� Light sabres, optical computers and gravity detectors may follow in
due course. Not bad for a material that鈥檚 full of holes . . .