麻豆传媒

A BULLET train slips quietly out of Singapore at 500 kilometres an hour,
hovering just above the superconducting track. Its power is carried by
superconducting cables that never waste a single watt. In first class, a
businessman taps away at his superconducting laptop, which contains the fastest
processor ever seen. Until the coolant runs out, and it turns into a piece of
junk.

Superconductors, which carry electricity without resistance and repel
magnetic fields, could be the key to many technical marvels. But even the
鈥渉igh-temperature鈥� superconductors don鈥檛 work above about 130 kelvin (鈭�140
掳C), so without expensive and bulky refrigeration they are useless. And when
it comes to increasing this critical temperature, scientists have a big problem:
they know very little about how high-temperature superconductors work.

But one strange little phenomenon could hold the answer. Stripes. Nothing to
do with naff wallpaper, these stripes are microscopic rivers of charge that run
through the high-temperature superconductors. Some researchers are hailing them
as the key to one of the most puzzling mysteries in physics. Others claim the
stripes have nothing to do with superconductivity at all. And a few are using
stripes to try to build a superconductor that works at room temperature.

For more than a decade, physicists studying high-temperature superconductors
have seen hints of parallel strips just a few billionths of a metre wide that
have different magnetic properties from the rest of the material. They are
difficult to see, so their nature and even their existence was disputed.

But not any more. Experiments at the Massachusetts Institute of Technology,
reported in Physical Review Letters last month, have proved that
stripes do indeed exist, and that they swarm with electrical charges. And that
gives older studies new significance. Last year, for example, researchers from
Germany, Poland and the US used a beam of particles called muons to probe a
superconductor surface. The researchers found that 80 per cent of the area was
superconducting and 80 per cent was taken up by stripes鈥攅vidence that
high-temperature superconductivity and stripes occur in the same regions.
And another study last year, by Japanese and American researchers, showed that
decreasing the width of stripes can push up the temperature at which the
materials become superconducting. Hence the excitement: control this transition
temperature and the world is at your feet.

Antonio Bianconi of La Sapienza University in Rome believes that stripes are
like thin superconducting wires separated by bands of insulating material, and
that this configuration somehow increases the normal superconducting temperature
of a material. He has been granted a European and Japanese patent for a
鈥渟uperconducting lattice鈥� of wires that is supposed to mimic the behaviour of
stripes in high-temperature superconductors.

Bianconi believes his approach will get superconducting materials up to room
temperature long before anyone understands how the stripes work. His formula is
simple: carve slender wires from the right material and weave them into a
three-dimensional lattice, in which the wires are separated by exactly the right
distance. He reckons that such devices could give us superconducting computers
sooner than anyone thought possible.

鈥淲e have some indication already that in aluminium we can increase the
superconducting temperature by three or four times,鈥� he says. And he believes
that tenfold or even hundredfold improvements are possible. If so, then a
lattice made of niobium, which is normally superconducting around 10 K, would be
superconducting at room temperature.

The gap between experiments and Bianconi鈥檚 claims is rather large, to say the
least, and it could be years before he can prove he鈥檚 right. That hasn鈥檛
deterred him, however. To make further progress, he needs wires less than 10
nanometres thick鈥攖he thinnest in the world. A few researchers are working
on this, but progress is slow. 鈥淭his is at the forefront of nanotechnology,鈥�
Bianconi says. 鈥淣ormally people work to 500 nanometres with
蝉耻辫别谤肠辞苍诲耻肠迟辞谤蝉.鈥�

Other researchers in the field won鈥檛 be holding their breath. 鈥淚f Bianconi鈥檚
device ever turns out to be superconducting, it will be superconducting for the
wrong reasons,鈥� says Philip Anderson, a theorist at Princeton University, New
Jersey. He is one of those who think the stripes are a red herring, an
irrelevance in the quest to explain high-temperature superconductivity.

Nothing that mimics the stripes will make the slightest bit of difference to
the superconducting transition temperature, he says. Anderson believes that the
stripe fad is making people forget all the experiments where no one saw any
stripes. 鈥淭hey鈥檝e carefully focused in on the samples that show these peculiar
behaviours and not looked at the broad background of samples that don鈥檛, he
蝉补测蝉.鈥�

But there is a middle ground. Vic Emery of Brookhaven National Laboratory in
New York also thinks Bianconi is chasing a dream, but he believes stripes are
more than a curiosity. They are, he says, the key to high-temperature
superconductivity. Along with Steven Kivelson of the University of California,
Los Angeles, Emery has developed an ingenious theory that explains why stripes
form and why they make materials superconducting.

Hoping holes

To understand what鈥檚 going on, we have to look at the atomic structure of
superconducting materials. High-temperature superconductors are based on copper
oxides. Unlike metals, which conduct electricity well because they have plenty
of electrons that can move freely, copper oxides are poor conductors. All their
electrons are pinned to atoms. What鈥檚 more, each sheet of copper oxide has a
highly ordered 鈥渁ntiferromagnetic鈥� structure: every copper atom has a magnetic
field, defined by the direction of the spin of its outer electron, and adjacent
copper atoms align themselves in opposite directions.

Because such materials are nonconducting, to get them to behave as
superconductors you have to introduce charge carriers. Adding lanthanum and
strontium, for example, produces positively charged 鈥渉oles鈥濃�攇aps where
electrons are missing鈥攖hat can move through the lattice.

But in moving through the material, such a hole should encounter massive
resistance. For a hole to hop from one atom to another, an electron has to hop
in the other direction, changing the spin on each atom and so upsetting the neat
antiferromagnetic pattern. 鈥淚nstead of having adjacent spins pointing opposite
to each other you have spins pointing in the same direction. That costs energy,鈥�
Emery says.

And that could be why stripes form. The holes are uncomfortable in an
environment where they disturb the antiferromagnetic structure. Instead, they
congregate in stripes, leaving hole-free anti-ferromagnetic bands in
between.

Once formed, stripes may encourage holes to pair up. Single holes cannot
travel through the antiferromagnetic parts of the material鈥攖hey are
imprisoned in their stripe. And according to quantum mechanics, a tightly
confined particle bounces around violently, so a trapped hole would have a lot
of energy. But holes can escape the stripe, and therefore lower their energy, by
cooperating: two holes can pair up to form a single entity whose magnetisation
is a mixture of up and down and so is barely affected by the strips of
antiferromagnetic material. Free to skip in all directions, the total energy of
the pair is much lower. They can relax.

This complicated line of reasoning solves an old problem. Researchers have
known for a long time that the pairing of charge carriers is at the heart of
superconductivity, because it allows charge carriers to move through a material
without being scattered by the electric fields of atoms. But the two holes in a
pair are both positively charged, so the enormous repulsive electrostatic force
between them ought to tear them apart. No one had been able to explain what
might overcome this repulsion鈥攂ut what could do the trick is the reduction
in energy that pairs gain from their freedom of movement.

So stripes may be essential for these substances to be superconducting.
Unfortunately, even if this theory is right, Emery doesn鈥檛 see how the
temperature could be much improved. 鈥淭he outlook is pessimistic if you鈥檙e
looking for room-temperature superconductors,鈥� he says.

Anderson is on Emery and Kivelson鈥檚 side, to an extent. 鈥淚 think they鈥檙e
absolutely right, but they鈥檙e not original,鈥� he says. Their suggested pairing
mechanism, Anderson claims, is simply the development of a theory鈥攐ne that
has nothing to do with charge stripes鈥攈e put forward ten years ago. He is
not accusing them of plagiarism, but says that the whole field is just going
round in circles, devoid of new ideas. The cause of all the fuss about stripes
is frustration among scientists, not new findings, he says. 鈥淚t鈥檚 just that we
theorists are a little bit stuck. There have been waves of fashion of this sort
in the subject over and over again.鈥�

However, almost everywhere stripes appear, says Anderson, the superconducting
transition temperature falls sharply. Indeed, the existence of static stripes
should destroy superconductivity because it allows large oscillations of the
charge to build up. These 鈥渃harge density waves鈥� should split pairs apart.

When John Tranquada鈥檚 research group at Brookhaven National Laboratory
managed to pin down the stripes by adding neodymium atoms, the superconductivity
disappeared. But that鈥檚 not a problem, Emery says. It鈥檚 a good thing. 鈥淭he fact
that you are able to kill superconductivity by making stripes more parallel is
an important piece of the puzzle,鈥� he argues. In general, the stripes are not
static, but meandering, he believes, which stops charge density waves becoming
too powerful. This could be why stripes have sometimes been so hard to see. And
it may also mean, Emery and Kivelson believe, that Bianconi鈥檚 rigid stripe
lattice is doomed to fail.

Not everyone is convinced. But Emery and Kivelson can take heart from the
latest breakthrough, published in Physical Review Letters last month
(vol 82, p 4300). A team from MIT used radio waves to excite the nuclei of
copper atoms in their sample and measured exactly which frequencies were
absorbed. Within stripes, they found a clear evidence of nearby charges.
Previous studies using different methods had detected only the antiferromagnetic
areas between the stripes.

Most importantly, they found strong evidence that stripes are detectable when
materials are superconducting. It is not yet clear whether the stripes are fixed
in place, or moving very slowly, but they are definitely there.

However, as Takashi Imai of the MIT group admits, this does not prove that
stripes are essential. Superconductivity is most suppressed when the stripes are
steadiest, so sceptics could argue that it is disrupted by the stripes. Or, if
you prefer, the results show that when the stripes are allowed to fluctuate
slowly, superconductivity is much stronger. 鈥淢aybe stripes and superconductivity
have some sort of love-hate relationship: they need each other, but if they are
too close they end up fighting,鈥� Imai suggests.

Anderson would prefer it if experimentalists didn鈥檛 put too much spin on
their data. 鈥淚f you look at the data and close your ears to what the speaker is
saying, you find that different people get the same stuff,鈥� he says.

In the short term, things look bleak. Scientists are struggling to find
superconductors that work at higher temperatures. Between 1986 and 1995, it was
trial and error approach, simply mixing up new compounds and testing them, that
raised the highest superconducting transition temperature from 鈥�238 掳C
to 鈥�139 掳C. And it hasn鈥檛 moved since.

Meanwhile, theorists are in noisy disagreement about stripes. If Anderson is
right, and stripes are an unimportant fad, the noise will fade away in a few
years鈥� time. Emery believes the experimental evidence for their importance will
grow and grow. Stripes are here to stay, he says. Imai thinks it could be twenty
years before anyone knows the truth. By that time, Bianconi hopes he will be
holding a room-temperature superconductor, built from tiny wires, in his
hand.