¿ìè¶ÌÊÓƵ

Look down on Grenoble from the air and you might catch sight of what
looks like a giant Polo mint straddling a peninsula between the rivers Isere
and Drac. This huge circular building, 850 metres in circumference, houses
the bare bones of Europe’s latest grand collaborative high-technology project
– the European Synchrotron Radiation Source (ESRF). In two years’ time,
when the ESRF starts working, it will become the world’s brightest source
of X-rays. This will make it an invaluable tool for probing the structure
of matter at many levels, from the minutiae of magnetic interactions in
atomic orbitals, through the crystal steps that form a catalyst’s surface,
to the complex structure of viruses. The X-rays could even image blocked
arteries in patients with coronary problems.

The ESRF is being built next to the home of another radiation source,
the Laue-Langevin Institute, which has a nuclear reactor providing neutrons
for a wide range of experiments (‘Neutrons tackle ‘sludge science’ ‘, New
¿ìè¶ÌÊÓƵ, 25 January). The instruments at the ESRF will complement those
at the institute, providing an unprecedented choice of analytical techniques
for visiting European researchers from all disciplines. In the same complex
is one of the outstations of the European Laboratory for Molecular Biology
and several French national research institutions, including the CENG (Grenoble
Nuclear Energy Research Centre) and laboratories of the CNRS, the French
national agency for scientific research. The area has recently been renamed
the Polygone Scientifique Louis Neel – a sort of scientists’ EuroDisney.

Like a theme park, the ESRF is an engineering project on the grand scale.
It is being financed by 12 European countries, including Britain, and is
on target to keep costs within its construction budget of £350 million.
According to Andrew Miller, a British biochemist, who has just retired as
the ESRF’s director of research, the financial planning is well managed.
He adds wryly: ‘This is the first time that I have worked in an adequately
funded facility.’ And construction is six months ahead of schedule.

Studies to build the ESRF started as long ago as 1976 and in 1985 France,
Germany, Italy, Spain and Britain agreed to go ahead with the project. Switzerland,
Denmark, Finland, Norway and Sweden joined the collaboration in 1987, followed
by Belgium and more recently the Netherlands. Construction started in 1988.

The ESRF is the first of a new generation of sources of so-called synchrotron
radiation – a type of electromagnetic radiation (photons) that was first
observed in 1947 in a circular particle accelerator called a synchrotron,
belonging to the American company General Electric. Synchrotron radiation
is emitted when charged particles, particularly electrons and positrons,
are accelerated by a magnetic field perpendicular to their direction of
motion. It covers a large part of the electromagnetic spectrum, from the
far infrared to the hard X-ray region.

At first, researchers working in high-energy physics regarded synchrotron
radiation as a nuisance because it caused energy to leak away from the particles
they were trying to accelerate. But gradually some researchers realised
that the radiation might actually have some use.

Photons of all wavelengths provide one of the best ways of studying
the properties of matter. The shorter the wavelength the sharper the detail
that can be seen and the further it is possible to probe into the microscopic
realm of molecules and atoms. And the brighter the radiation, the more quickly
a clear image of an object can be obtained without damaging it. Synchrotron
light is ideal because as the charged particles approach the speed of light,
the number of photons they emit (the flux) is extremely high. The light
also shifts to higher energies and, therefore, towards the shorter ultraviolet
and X-ray wavelengths which are comparable in dimension with the spacing
between atoms in crystals. At these wavelengths, light interacts with electrons
in atoms.

But what is particularly significant is that the photons that make up
synchrotron light are emitted in a narrow, flattened cone, reaching perhaps
a few millimetres across and a fraction of a millimetre high, in the direction
that the particles are moving. If the particles are moving in a circle then
the photons are emitted tangentially to the circle. This provides a beam
of brilliant radiation from which different wavebands can be selected and
focused. The light is also partially polarised in the horizontal plane.

At first researchers opportunistically siphoned off synchrotron radiation
from electron accelerators designed for high-energy physics. The first machine
to produce synchrotron radiation was the Tantalus storage ring, built in
1968 at the University of Wisconsin. Originally designed to store circulating
protons, the ring operated at an energy of 0.24 gigaelectronvolts. Over
the past 20 years many countries have converted electron accelerators into
synchrotron sources or have built dedicated ones such as the Brookhaven
National Light Source in the US and the Synchrotron Radiation Source at
Daresbury in Britain, which operates at 2 GeV.

Dedicated synchrotron facilities are up to a million times as powerful
as conventional sources of radiation such as gas discharge lamps which produce
ultraviolet light (with photons having energies between 10 and 1000 electronvolts)
and X-ray tubes which emit in the X-ray region (photons at energies from
1 to 100 kiloelectronvolts). A synchrotron facility supports a massive arsenal
of analytical techniques based on the reflection or absorption of light.
Synchrotron radiation is probably the most powerful research tool available.
Indeed much of the research in surface science and molecular biology would
not be possible without it.

The new synchrotron sources such as the ESRF will generate X-rays that
are much brighter than those from present sources. The ESRF will operate
at an electron energy of 6 GeV. Close on its heels is the 7-GeV Advanced
Photon Source (APS) being built at Argonne National Laboratory in Illinois,
to be ready by 1996. Two years later Japan’s Spring-8 source will be completed,
overtaking its Western counterparts with a ring energy of 8 GeV.

A modern synchrotron source is a complex arrangement. A linear accelerator,
or linac, accelerates electrons from an electron gun. The particles are
then injected into a ‘booster’ synchrotron where magnets bend and focus
the electron beam around the ring, accelerating them close to the speed
of light. The electron beam is then channelled into a similar but larger
storage ring which keeps the electrons circulating at constant energy for
several hours. Radio-frequency cavities around the ring top up the energy
lost through radiation and split the electron beam into ‘bunches’ a few
centimetres long. The electrons are replenished every few hours as they
are lost through interactions with residual gas molecules in the ring.

A typical storage ring is not completely circular but consists of alternating
curved and straight sections. Magnets installed on the curved sections bend
the electron path around the ring, producing synchrotron radiation with
a flux between a hundred and a million times as great as that of light from
conventional sources (see Figure 1).

Until a few years ago, ‘bending’ magnets provided the main source of
synchrotron radiation. But the new generation of synchrotron sources, including
the ESRF, will largely exploit magnetic devices called ‘wigglers’ and ‘undulators’
situated along the straight sections of the ring. These so-called insertion
devices make the electron beam twist and turn in particular ways so as produce
much more brilliant radiation at higher energies and in narrower beams
(see Figure 2). What is more, their geometrical configuration can be tailored
for specific experiments. Insertion devices consist of two parallel horizontal
columns of magnets with alternating north and south poles. This periodic
magnetic field causes the electron beam to wobble from side to side without
actually changing direction, but the intensity of the radiation increases
by a factor that depends on the number of magnets and their length.

The difference between a wiggler and an undulator is subtle but important.
Wigglers have fewer, longer magnets, and periodically deflect the electron
beam from its circular path at an angle that is larger than the angle at
which the light is emitted. Undulators are a newer development in which
the beam deflections are smaller than the angle at which light is emitted.
In this case something interesting happens: the cones of light waves from
each deflection overlap so that they interfere. This generates interference
patterns in which the intensity at certain wavelengths is significantly
increased. The intensity depends on the square of the number of magnets
making up the undulator. The radiation also has a laser-like quality, in
that it is emitted as a thin, parallel beam and is partially coherent. Undulators
are suitable for experiments that require extremely bright radiation of
a single wavelength. The wavelength can be varied by altering the gap between
the top and bottom rows of magnets.

The beams emerging from either the bending magnets or the insertion
devices are channelled down pipes called ‘beamlines’, which run tangentially
from apertures in the storage ring to nearby experimental areas. Each experimental
station is like a small laboratory. Its beamline will have a particular
configuration of optics to tailor the radiation for a specific experiment.
Slits and mirrors first guide and focus the light, and then wavelengths
are selected by a series of mirrors and monochromators. The kind of monochromator
used depends on the region of the spectrum. Traditional diffraction gratings
are used to select ultraviolet and soft X-rays (with wavelengths down to
100 nanometres). Harder X-rays (with wavelengths down to 10 nanometres)
require monchromators consisting of pairs of perfect crystals whose planes
of atoms reflect radiation of a particular wavelength. This phenomenon is
called Bragg scattering if the radiation bounces off atomic electrons, or
nuclear Bragg scattering if it is reflected from the nuclei, as happens
in the case of very hard X-rays.

The number of beamlines and experimental stations depends on the number
and arrangement of the magnets around the storage ring. At the ESRF, the
storage ring has 32 straight sections available for insertion devices, and
64 curves. Initially the ESRF will have 30 public beamlines, which any research
groups from the contributing countries can apply to use. Of these, 26 will
be piped from insertion devices on the straight sections; the other four
will be on the bending magnets.

The ESRF is also considering proposals from groups of users, known as
collaborating research groups, who will pay for further beamlines and instruments
to be built to their own specifications on the remaining bending magnets.
The groups may be international consortia or from industry. The only stipulation
is that one-third of the time must be available for public use. Four applications
have been agreed on so far. The American APS will fund all of its 68 beamlines
in this way.

The ESRF council expects at least seven beamlines to be operational
by the middle of 1994, with a further 11 to be commissioned within the following
12 months. Three new beamlines will then be put into operation each year
to reach a total of 30 by 1998. Gottfried Mulhaupt, joint director of the
project, expects that these targets should be met without problem. The 40-metre-long
linac, which will run at 200 MeV, and the 300-metre booster synchrotron
were tested last year. Last month the storage ring was injected with electrons
to test the beam’s behaviour. To reach the target X-ray energies of up to
100 keV will require an electron current of 100 milliamps concentrated into
bunches circulating at 350 000 revolutions per second.

One major problem the engineers have had to overcome is ensuring that
the electron beam remains stable within the narrow confines of the bending
and focusing elements. The beam must not be allowed to wander more than
a few micrometres, particularly in the insertion devices where the gap width
is crucial and the tolerances are incredibly small. This meant, for instance,
that the engineers had to develop a novel way of compensating for building
settlement, which in the Grenoble area can vary over a distance of 10 metres
by about 0.3 millimetres per year. Vibrations due to the surrounding rivers
and town are also a problem. Their solution was to support the storage ring
on girders equipped with remote controlled jacks monitored by a levelling
system which acts on the same principle as a spirit level.

Another crucial question is how long the electrons can remain in the
ring. The projected storage time is about 10 hours. Although the tube is
evacuated to a pressure of a million-millionth of an atmosphere, there remain
dust particles and gas molecules which form positive ions. They attract
electrons which are then lost from the beam. One way to reduce the problem
is to use positrons instead of electrons, as they repel positive ions. This
doubles the lifetime of the beam. The ESRF council will decide later this
year whether it will be necessary to install a positron converter and a
400-MeV positron linac.

The light from the wigglers and undulators is a hundred times more intense
than that obtained in current synchrotron sources. Developing the necessary
optics to deal with such intense radiation will not be easy. Silicon crystals
are ideal monochromators for X-rays but the beams will deliver so much energy
(as much as 100 keV) that they cause the surface of the crystal to heat
up, and therefore to expand and deform. A way around this is to cool the
crystals with liquid nitrogen to -150 °C; at this low temperature the
thermal expansion vanishes and the thermal conductivity is three times as
great as at room temperature. Engineers designing the APS have chosen liquid
gallium as a coolant to achieve the same trick. Another way of decreasing
thermal deformation is to use very thin crystals. The thickness required
to reflect a beam is much smaller than the absorption length so these very
thin crystals transmit much of the incident power whilst reflecting all
the useful X-rays.

Advanced X-ray optics will figure in one of the first beamlines to be
ready, the microfocus beamline. It will exploit a wiggler producing tunable
radiation with up to 40 keV of energy focused to a tiny spot only 10 micrometres
across. Such a fine beam will be ideal for diffraction studies on minute
crystals and a boon to researchers who study proteins that are difficult
to crystallise. Another high priority will be a ‘white’ beamline from a
wiggler for carrying out Laue crystallography on proteins. This technique
will allow researchers to follow the course of enzyme reactions by taking
snapshots every few microseconds – the time taken for a bunch of electrons
to circulate once around the storage ring. This time resolution is a hundred
times as fast as that of similar experiments now carried out at Daresbury
(‘The first enzyme picture show’, ¿ìè¶ÌÊÓƵ, 14 December 1991).

Larger biological crystals such as viruses, which do not readily scatter
X-rays, will be analysed by a beamline from an undulator optimised to give
the highest possible brilliance. A high-energy beamline generating X-rays
of more than 80 keV will be used to study defects in large crystals, and
the electronic properties of solids.

A further beamline will exploit an ‘exotic’ undulator that generates
a magnetic field with both vertical and horizontal components, forcing electrons
to follow a helical path. The resulting radiation will be circularly polarised,
and will be used to study ‘chiral’ molecules, whose geometrical arrangement
of atoms endows them with a handedness. It will be also used to look at
surface magnetism.

A newly developed analytical approach called magnetic resonance scattering
uses intense beams of X-rays to investigate exotic magnetic materials such
as alloys of uranium, antimony and thorium. X-rays can distinguish between
two types of magnetism – that from an electron’s spin and that from its
orbital motion. The neutrons traditionally used to study magnetism cannot
do this.

Dozens more experiments are planned at the ESRF but how much will British
researchers benefit? The Synchrotron Radiation Source at Daresbury has undoubtedly
been very successful, researchers there having resolved the structure of
the foot and mouth virus, for example. For many experiments the Daresbury
source is perfectly adequate, but before researchers can get time on many
of its beamlines they have to wait in a long queue. The ESRF will provide
much-needed extra capacity. More importantly, because the ESRF is a hundred
times as powerful, it will make possible experiments that could not be done
at Daresbury.

Michael Hart at the University of Manchester, who is vice-chairman of
the ESRF’s science advisory committee, says he is pleased that the experiments
chosen for the first eight stations to be built at the ESRF will complement
those at Daresbury. John Helliwell, also from Manchester and a member of
the committee, sees a clear need for both facilities. ‘The ESRF will take
over the time-resolved studies,’ he points out.

Much will depend on how much money Britain’s Science and Engineering
Research Council allocates for experiments at the ESRF. As yet no mechanism
has been set up for submitting research proposals, although it is most likely
they will be processed through the SERC’s Rutherford Appleton Laboratory
as is the case for applications to work at the Laue-Langevin Institute next
door. If the first beamlines start operating in March 1994, then the call
for proposals will probably come in the preceding autumn. According to one
SERC administrator, British researchers who want to use the ESRF will have
to show that their experiments can only be done with the more powerful source.

An alternative access route could be through the collaborating research
groups. William Stirling of the University of Keele and Malcolm Cooper of
the University of Warwick have proposed such a group to do magnetic scattering
experiments, which require the circularly polarised light avail-able only
at the ESRF. The SERC is assessing whether to fund them. It admits that
the group has a very compelling case but cautions that funds are limited.
This means that the Research Council is forced to choose ‘between compet-ing
priorities’. Nevertheless, Britain has already provided 14 per cent of the
cost of the new source, and as Ruprecht Haensel, the director-general of
the ESRF, points out: ‘The facility will bring about cross-fertilisation
between many small research groups from different disciplines and diff-erent
countries.’ So it would be a pity if British researchers could not take
full advantage of one of the world’s most remarkable scientific instruments.