According to Howard Caroline, a senior structural engineer with the
underground, there had always been rumours about the spiral escalator at
the station but it was not until late 1988, during a routine structural
inspection, that an engineer climbed through a duct next to the line and
discovered the machinery, covered in rubble.

Now the floor slabs around the remains have become unsafe so the engineers
have taken the opportunity of cutting the machinery into three pieces and
removing it. One of the pieces will be put on display at the London Transport
Museum, and another in the station itself which is in the process of being
renovated.

According to Paul Hadley, a researcher in transport history, the escalator
was built by Reno Electric Stairways and Conveyors, the British arm of an
American company. The Holloway Road escalator was planned to coincide with
the opening of the Great Northern, Piccadilly and Brompton Railway (now
the Piccadilly line) in December 1906. Hadley is unsure whether the escalator
was ever finished, but it was certainly never certified for use by the Railways
Inspectorate. The Reno company was wound up a few years later.

The escalator’s guide rails wind up in a gentle helix, level out at
the top and wind down again to complete a single unbroken loop. The moving
part is a form of chain in which the links are joined together with universal
joints. Wooden treads, fastened to the top of the links, have rollers at
each side that run along the guide rails.

The treads do not form steps, as they do on a modern escalator, but
a flat ramp – something like a inclined travolator. Hadley says it is not
clear from the remains but the chain was probably driven from beneath.

In recent years the Japanese company Mitsubishi has built a few spiral
escalators in Japan and the US. Otis, the world’s largest escalator maker,
has built a prototype but, because of the cost, has no plans to sell any
at the moment.

These gravitational waves are predicted by Einstein’s general theory
of relativity, but because they are very weak and hard to detect, no one
has yet managed to spot one. Detecting the waves would not only confirm
gravitational theories but could also be immensely valuable to astronomers
studying violent cosmic events, such as supernovae or two neutron stars
crashing together. ‘It would be a Nobel-prize-level discovery,’ says Richard
Marsden, deputy project scientist for Ulysses, one of the spacecraft involved
in the experiment.

Ulysses is a European probe on its way to study the north and south
poles of the Sun. Two NASA spacecraft are also involved: the Mars Observer,
now heading for Mars to study its climate and geology (see ‘Rendezvous
with the red planet’, this issue); and Galileo, en route for Jupiter. None
of these spacecraft were designed to detect gravity waves, but in a three-week
experiment running until 11 April they will form part of a detector.

Catastrophic events such as supernovae cause gravitational waves, rather
as a falling stone causes ripples on the surface of a pond. Since the 1960s,
physicists have been trying to detect the waves, but without success. They
are now waiting for a new generation of giant interferometers that measure
the time it takes a laser beam to travel through vacuum tubes several kilometres
long.

One such detector may begin construction in the US this year, and one
in Europe is awaiting final approval. Bernard Schutz, professor of astrophysics
at the University of Wales College of Cardiff, says he is confident that
these detectors will spot gravity waves before the end of the decade.

The experiment with the three spacecraft uses a similar approach. As
a gravitational wave passes through the Solar System, the spacecraft will
briefly bob about like corks on the surface of a pond. To detect this, researchers
are analysing the radio signals that are used to communicate with the satellites.
The signals are superimposed on a carrier wave that is transmitted at a
fixed frequency. The spacecraft picks this up, and sends it straight back
with its own messages superimposed on it.

But the frequency of the returning carrier wave will have changed. The
spacecraft is moving away at high speed, and this stretches out the wave,
reducing its frequency – an effect known as Doppler shift. If a gravitational
wave makes the spacecraft wobble, there will be a wobble in the Doppler
shift of the returning carrier wave.

The wave should produce three signals: the first as it passes the spacecraft,
and the second as it passes the Earth, because this too will affect the
relative speed between the craft and Earth. The third is caused by the fact
that as the wave passes Earth, a carrier wave will be sent out with a signal
on it. This will arrive back at the Earth some minutes later after a round
trip to the spacecraft. If the researchers detect all three of these signals
they can be confident that they have seen a gravitational wave.

Detecting the same wave with more than one spacecraft gives them more
evidence. And by looking at the exact time it passed the three craft and
the Earth, the researchers can work out the direction of the source of the
wave.

Detectors based on the ground can only pick up gravitational waves with
frequencies of greater than 10 hertz. The space experiment will be looking
for waves with around a ten-thousandth of that frequency, which are believed
to be characteristic of events such as stars falling into giant black holes
at the centres of galaxies.

According to Marsden, it is hard to predict the intensity of such low-frequency
radiation, so the chances of positively identifying a wave are very small.
Schutz says that gravitation theories predict that the space detector would
need to be 100 times as sensitive as it is at the moment. ‘But it’s not
impossible that they will see something,’ he says.

The other two major tasks are replacing the wide-field/planetary camera,
one of the telescope’s main instruments, and installing a package of optics
called COSTAR to solve problems in three other instruments caused by the
infamous flaw in the main mirror. These two pieces of equipment, at around
350 kilograms each, have greater masses than the arrays, so are harder to
move around. But they are enclosed in boxes that simply slot into place,
so are less prone to damage.

The arrays are made by British Aerospace in Bristol. The five astronauts
involved in the repair have been practising in water-filled tanks in the
US. Their visit to Bristol was to familiarise themselves with the real equipment
they will next see in orbit.

The repair involves the most complicated series of spacewalks that NASA
has yet carried out. ‘It’s not cut and dried. This is a very complex mission,’
says Jeffrey Hoffman, one of the astronauts. Another, Kathryn Thornton,
likens working in space to repairing a car while hanging upside down and
wearing ski gloves.

Hubble, launched in April 1990, was designed to be repaired in orbit
about every three years. The two arrays, each 12 metres long and weighing
160 kilograms, have a total of 50 000 photovoltaic cells and were designed
to supply at least 4.5 kilowatts after two years in orbit. They were also
designed to last at least 5 years but are being replaced early because
they judder.

This happens mainly when the telescope passes from the Earth’s shadow
into sunlight and the temperature climbs by as much as 200 °C. The
booms that support the array expand from the increase in temperature, but
this happens much more quickly and jerkily than expected, in a rapid series
of stick and slip movements. The vibrations make it impossible to keep
the telescope pointing on target.

NASA has been able to reduce the effect by actively vibrating the telescope
in the opposite direction. But this manoeuvre uses so much of the telescope’s
on board memory that it is preventing it doing other things. NASA decided
to replace the arrays early to solve the problem.

According to Mike Newns, the project manager at British Aerospace, the
problem was tackled in two ways. First, on the new arrays his team fitted
thermal insulation to the support booms to slow the temperature changes.
BAe engineers took 1000 plastic discs coated with aluminium and welded them
together to form a concertina that fits over the support booms.

Second, the drum from which the flexible arrays are unrolled was fitted
with a drum brake to hold it steady while in space. The far end of each
wing of the array is fitted with nine coil springs to hold the array taught.
The engineers hope these changes will stop the stick-slip behaviour.

Although it is designed at present to represent a bit with around 100
electrons, they can be marshalled in and out of the cell one at a time.
The researchers say the cell can be easily modified to hold just one electron
– today’s chips use 500 000 electrons to represent a single bit.

The researchers predict that a chip the size of a 50 pence coin, made
using this technique, could hold a terabit of information – a million million
bits. Today, the best chips on sale hold 16 megabits, and a terabit of data
would take up the area of a tennis court.

The Japanese electronics company Hitachi funded the joint research by
the company’s laboratory in Cambridge and researchers from the University
of Cambridge’s Cavendish Laboratory, led by Haroon Ahmed, professor of microelectronics
(Electronics Letters, 18 February).

Last year, the researchers made a switch that could control the flow
of single electrons. On a semiconductor chip they fabricated strips of gallium
arsenide just a few atoms thick, doped with silicon to provide more electrons.
In places, the strips narrow to just a fraction of a micrometre wide. These
constrictions form the switches: three are shown in the picture above.

Other strips laid down close to the constrictions control the flow of
electrons: applying a negative voltage to a control electrode forces electrons
out of the constriction. Those that remain cluster into a string of little
islets, each a few tens of nanometres across.

Mostly, the gaps between the islets act as barriers to electrons, but
quantum mechanics allows a few to cheat and ‘tunnel’ through the barriers.
By carefully controlling the applied voltage, it is possible to let electrons
through one by one. The researchers call the switches multiple tunnel junctions.

The memory cell, which is about 20 micrometres across, can be seen in
the centre of the picture as a collection of strips isolated by two junctions.
The cell can hold single electrons because of a phenomenon called Coulomb
blockade. If the cell is sufficiently small, transferring one electron into
it changes its voltage so much that other electrons are prevented from entering
and the one inside cannot escape. So the cell has two stable states: either
empty or containing one electron, and these can be used to represent the
digital states 0 or 1. The junction in the bottom right of the picture acts
as a detector to determine which state the cell is in.

The cell is switched between the two states by the strip coming down
from the top left of the picture. This changes the voltage of the cell:
making it more negative will force the trapped electron out again, while
a positive voltage will lower the Coulomb blockade and allow another electron
in.

The main problem is that the cell only works at the extremely low temperature
of 0.03 K. To work at room temperature, the cell would have to shrink to
less than a thousandth of its present size. The researchers believe it will
be at least ten years before devices that small can be made.

An organic molecule, nicknamed George, has provided a key step towards
making electronic circuits from organic molecules rather than inorganic
semiconductors. British researchers have demonstrated that a layer of the
material only a few molecules thick can behave as a rectifier, in other
words it will allow electric current to flow in one direction but blocks
current flowing the other way.

The most obvious application for molecular circuits would be in sensors
for organic compounds and chemicals, says Roy Sambles of the physics department
at the University of Exeter, leader of the team that developed the rectifier.
Reactions between the compounds and sensing circuit generate an electrical
current that passes to a conventional circuit to display the result.

To make circuits, Sambles says, you need an insulating material – fatty
acids in the case of organic circuits – and switches so that you can turn
currents on and off or store charge. Rectifiers are the key to switches:
in semiconductor electronics, basic transistors are made by fabricating
two rectifiers back-to-back.

The ideal of molecular electronics is to produce single chain-like molecules
that behave as circuits, with different structures along the chain acting
as electronic components. These could be made in a laboratory using chemical
synthesis without the complex and expensive wafer fabrication plants needed
for semiconductors.

Layers of organic molecules have been shown before to behave like rectifiers.
But researchers were never sure that it was the molecule causing the results
or simply an interaction between the active groups of the molecule and the
metal contacts. Sambles along with Scott Martin, also from Exeter, and Geoff
Ashwell from the Cranfield Institute of Technology have shown conclusively
that it is the molecule itself (Physical Review Letters, 11 January, p 218).

Ashwell, a synthetic chemist, prepared a molecule with the formula C₁₆H₃₃-&ggr;Q3CNQ,
dubbed George for convenience. It is a type of double ion, known as a zwitterion,
that has both a positive and a negative ion group attached. Also, one end
of the molecule is water soluble and the other end, the paraffin group,
is not. So when the molecules are spread on the surface of water they form
a layer one molecule thick, known as a Langmuir-Blodgett film, with all
the molecules the same way up.

The film is lifted off the water surface with a silver-coated slide,
the silver later acting as one electrode of the rectifier. This process
is repeated seven times to produce a layer seven molecules thick which is
then dried for two weeks. A number of small magnesium electrodes are then
fabricated on top of the layers of molecules.

The team applied a voltage across the film which swept smoothly from
around -1 volts up to +1 volts and back again repeatedly. The resulting
current flowing through the film, from a single sweep, is shown in the diagram.
At positive voltages the current rises sharply, but negative voltages produce
virtually no current.

To show that the molecule was responsible for this effect, the researchers
made identical films using molecules that had been bleached to inactivate
the molecule’s active ion groups. This film simply acted as an insulator,
blocking current in both directions.

The team also made a layer of normal unbleached molecules sandwiched
between layers of w-tricosenoic acid, a material related to soap, to isolate
the molecule’s active groups from the metal contacts. This film sandwich
also behaved like a rectifier.

The team are now making films on transparent conducting electrodes so
that they can shine light at them. The conductivity of the brightly coloured
films is influenced by light and the researchers hope this will give clues
about how they work. The next stage, says Sambles, is to make a double,
back-to-back film – a molecular transistor.

One of the great frustrations of picnics, having to drink beer and soft
drinks that are lukewarm, may soon be at an end. An inventor in Florida
has developed a disposable self-cooling can which, counterintuitively, uses
boiling water to cool the drink it contains.

Water does not always boil at 100 °C. If the pressure of air above
the water is reduced, its boiling point also decreases. If it is put in
a vacuum, its boiling point is reduced to near 0 °C.

Israel Siegel of Miami Beach has built a prototype self-cooling can
which exploits this principle. The can has two chambers, of which the upper
one is almost filled with the beverage to be cooled. All the air is removed
from the can to create a vacuum but the space above the beverage is quickly
filled with water vapour, creating enough pressure to stop the beverage
boiling.

A pipe connects the space above the drink to the lower chamber, but
until cooling is needed, the pipe is covered with a breakable seal. The
bottom chamber contains a desiccant which absorbs water vapour, such as
silica gel or calcium sulphate.

When picnickers wants a cool drink, all they have to do is press a button
on the top of the can which breaks the seal on the tube. The desiccant then
absorbs all the water vapour in the can and the beverage finds itself in
a vacuum again.

As it will be at room temperature, it will also be above its boiling
point, so the beverage will begin to boil furiously until it cools to its
low boiling point near 0 °C. The desiccant becomes hot as it absorbs
the vapour, so the two chambers are insulated from each other with an air
gap. The nicely chilled drink can then be drunk through a ring pull at the
top of the can.

Siegel has also designed a version that can be built into the base of
a jug. A sealed chamber of water in the base is connected to the desiccant
by a tube with a valve. Opening the valve makes the water boil, so cooling
the beverage above it. The cooler can be turned on and off with the valve.
When the desicant becomes saturated it can be dried out with electric elements
built into the bottom chamber.

Siegel has patented his designs and is looking for a company to manufacture
the cans. He estimates the disposable cooler would add about 20 to 30p to
the cost of a can of drink.

Novaya Zemlya has always been secondary to the Soviet Union’s main testing
site at Semipalatinsk in Kazakhstan, although some of the largest bombs
were detonated at Novaya Zemlya. There have been about 40 underground tests
on the island.

Russia is observing a moratorium on testing until July, but the Russian
company Chetek, supported by the Ministry of Atomic Energy, is planning
to destroy chemical weapons by incinerating them in underground chambers
with redundant nuclear weapons.

John Hassard, an expert in nuclear proliferation at Imperial College,
London, says that it is well established that 30 per cent of Soviet tests
have leaked some radioactive material. ‘Such a cocktail of nuclear and chemical
material leaking would be disastrous,’ he says.

Johnny Skorve and John Kristen Skogan of the Norwegian Institute of
International Affairs have been studying recent satellite images of Novaya
Zemlya taken by the American Landsat satellites and the French SPOT satellite.
They compared these with aerial reconnaissance photographs taken by the
Luftwaffe in 1942.

They found several circular depressions only a few kilometres from the
main testing base which they believe to be relics of underground tests.
They have also identified several rockslips that could have been triggered
by underground explosions.

Nuclear experts in both Russia and the US say that underground tests
are safe because the heat of the explosion melts the surrounding rock. This
cools into a glassy sphere around the cavern created by the bomb, sealing
in the radioactive debris. But the surrounding rock can crack as it cools
and the caverns collapse leaving the telltale craters on the surface. Some
tests on Novaya Zemlya were very close to existing caverns, so may have
damaged them.

The Norwegian researchers are particularly concerned that radioactive
elements may leak from caverns into the ground water and from there into
the Matochkin Shar Strait. Some tests were less than 2 kilometres from the
strait. The ground is also frozen into permafrost as deep as 600 metres
in places. The heat of explosions can melt the permafrost above the cavern,
weakening the ground and releasing water that could seep through the cavern,
flushing radioactive elements into the ground water.

Photonic crystals, a new type of material first demonstrated last year
in which photons behave similarly to electrons in a semiconductor, have
found their first application. The crystals have a range of frequencies
at which photons are ‘forbidden’, so act as perfect reflectors of photons
at those frequencies. American researchers have now used a photonic crystal
as a reflector for a miniature microwave antenna which could one day be
fabricated on a single chip along with control electronics.

When atoms are combined to form a semiconductor crystal, the energy
levels that electrons can reside in merge to form energy bands. The low-energy
valence band holds electrons which bind the crystal together. It is separated
by a band gap from the conduction band, where free-moving electrons conduct
electricity.

Photonic crystals are analogous to semiconductor crystals. Photons can
only pass through if their frequency is in certain frequency bands; if
their frequency lies between bands they are forbidden entry and reflected.
The crystal acts like a filter to cut out those frequencies.

A team led by Eli Yablonovitch of Bellcore, the research arm of the
American regional telephone companies based in Redbank, New Jersey, made
the first photonic crystal last year by drilling holes in material that
is transparent to microwaves. The holes were drilled in a regular pattern
at certain angles to produce a three-dimensional array of voids. These
act like the atoms of the semiconductor crystal: photons which pass through
can either reside in the low-energy voids or the high-energy regions of
material between them. A band gap of frequencies is created which depends
on the spacing of the voids.

This year, Yablonovitch collaborated with Elliot Brown and Chris Parker
of the Massachusetts Institute of Technology’s Lincoln Laboratory in Lexington,
Massachusetts, to reflect signals from a microwave antenna with a photonic
crystal. Previously, if an antenna was made on a semiconductor substrate,
most of the microwaves went into the substrate rather than into the air.
A simple dipole antenna fabricated on semi-insulating gallium arsenide will
only radiate 2 per cent of its power into the air.

Yablonovitch made a photonic crystal from an epoxy-based material with
the voids spaced 7.8 millimetres apart. This produces a frequency band gap
from 13 gigahertz to 16 gigahertz. Brown and Parker then fabricated a flat
antenna in the shape of a bow tie on the surface of the crystal and fed
a microwave signal into it at a frequency of 13.2 gigahertz. They measured
the signal radiated from the antenna at all angles and found that almost
all the signal was transmitted into the air.

As a control, the researchers made an identically shaped antenna on
a block of the same material but without the voids. As expected, the signal
was mostly radiated into the material.

‘It’s a wonderful situation,’ says Brown, because you can change the
frequency of the band gap simply by changing the spacing of the voids and
change the width of the gap by choosing a different material.

Brown and Parker are now working on making the photonic crystal out
of silicon, gallium arsenide or indium phosphide, which would allow them
to integrate the antenna and electronics on the same chip. Such chips would
be ideal for such uses as phased-array radars, Brown says.

Making a photonic crystal with a band gap that excludes visible light
needs voids much closer together – only a few hundred nanometres. Although
researchers are trying to drill such crystals with beams of ions, none have
succeeded yet.