¿ìè¶ÌÊÓƵ

WHO wants to detect subatomic particles? Once it was just nuclear and
particle physicists who study the collisions of atomic nuclei and elementary
particles. They devised ingenious detectors – cloud and bubble chambers
and nuclear emulsions that no one else used. Now, the new detectors developed
to detect the exotic particles produced in accelerators at CERN, the European
Laboratory for Particle Physics in Geneva are among those finding uses in
many other areas of science.

These devices are enabling medical scientists to study how the muscles
in frogs’ legs contract, engineers to trace the flow of oil around the inside
of a jet engine, and doctors to observe how tumours grow. Detectors that
measure X-rays allow astronomers to observe collapsing stars, radiologists
to obtain moving images of patients, and materials scientists to understand
the structure of superconductors. The detectors not only measure the energy
and number of particles but also indicate where the particles come from.
They are called position-sensitive detectors.

Every baby comes into the world equipped with position-sensitive detectors
– its eyes – which respond to photons of visible light. Visible photons
are elementary particles with wavelengths of a few hundred nanometres, each
carrying a few electronvolts of energy. To detect more energetic particles,
such as neutrons, electrons, protons, and photons with high energies – ultraviolet
light, X-rays and gamma-rays – physicists and engineers have come up with
some ingenious concepts.

The first position sensitive detectors for particles were cloud chambers,
bubble chambers and nuclear emulsions. Physicists have used them ever since
the 1920 to study neutrinos, mesons, muons, strange particles and so on,
so establishing the particle physicist’s bestiary of fundamental objects.
In all three of these techniques, the tiny signals from a particle are directly
amplified into something visible, leaving a track in the material: a line
of droplets in a cloud chamber, bubbles in a bubble chamber or grains of
silver in a nuclear emulsion.

An energetic charged particle passing through matter behaves a bit like
a bullet going through a box of eggs. Some of the atoms and molecules are
completely shattered; electrons are knocked out, so that the atoms become
ionised, and sometimes even the nuclei of the atoms are broken up. Other
atoms are hit less hard; they absorb some of the energy and become ‘excited’.
Eventually, the ionised atoms recombine and the excited atoms return to
their lowest state of energy. In either case, the atoms lose energy, which
is dissipated as heat. (In a bubble chamber, for example, this heat causes
bubbles to form along the tracks of particles.)

Some crystals, such as barium fluoride or sodium iodide, convert some
of the energy of excitation into visible photons. These materials are called
scintillators and are among the standard detectors used in particle physics.

Another useful way of measuring the signal produced by subatomic particles
is to apply a strong electric field across the material of the detector.
This prevents the electrons from recombining with their ionised atoms. They
drift away and can then be collected. The electrons drift easily in gases
and in certain condensed materials such as silicon and liquid argon.

Yet all these different kinds of signals are too small to be detected
by the human senses. A charged particle going through a chamber filled with
gas releases just a few hundred electrons. Similarly, a gamma-ray passing
through a sodium iodide crystal releases a few tens of scintillation photons.
So the primary signals have to be amplified millions of times.

To detect and amplify the light from a scintillator, physicists use
an electronic valve called a photomultiplier tube. This consists of an evacuated
tube with a transparent window to let the light shine onto a photocathode,
a specially coated electrode. Photons cause electrons to be ejected from
the photocathode into the vacuum of the tube. An applied electric field
accelerates them so that they collide with a succession of other electrodes,
dynodes. Each collision splashes out more electrons from the metal of the
dynodes, producing a signal with a typical output of 10 million electrons
for every initial photoelectron. The combination of a scintillator with
a photomultiplier tube is called a scintillation counter. It is the standard
fast detector that particle physicists use, and it can detect particles
arriving at intervals as short as 100 picoseconds.

Amplifying electrons in a detector can be a ‘noisy’ process – think
of all the hiss and hum that you get from a cheap record player, or the
howl from a badly used public-address system. The electronic equipment involved
can introduce noise signals of its own which mask the signal that you want
to hear or measure. To overcome this problem, many detectors have amplification
built into the detection process, so that the signal reaching the electronic
amplifiers has already been boosted considerably before electronic noise
is added in.

An example is the multiwire proportional chamber invented by George
Charpak at CERN. Here, amplification results from collisions of electrons
with gas molecules to give an initial boost to the primary signal. A flat,
gas-filled box has a plane of fine parallel ‘sense-wires’ stretched through
it, which are about 2 millimetres apart (see Figure opposite). The walls
of the chamber are at zero volts but the wires have a positive electrical
potential of about 2 kilovolts.

When a particle releases negative electrons into the gas, they drift
rapidly towards the positively charged wires. The collisions with the gas
molecules control their speed. When the electrons come within a tenth of
a millimetre of the sense-wire, they are in such a high electric field –
due to the fineness of the wire – that they gain enough energy between collisions
to ionise gas molecules. This process produces more electrons, which are
accelerated in the field and make further ionising collisions, so producing
an avalanche of electrons. One electron drifting from a millimetre away
may be multiplied into 100 000 electrons reaching the wire, giving a pulse
that is much bigger than the amplifier noise.

The cleverness of Charpak’s invention was not in using gas multiplication
– the Geiger counter invented by Hans Geiger, 60 years ago, does the same
thing – but in avoiding its unwanted side effects. After a gas molecule
has been ionised, it also scintillates, sending out secondary photons of
visible and ultraviolet light as the excited atoms settle back into their
lowest state of energy. If these secondary photons are not stopped, they
knock out further electrons from the walls of the chamber. These drift to
their local sense-wire and generate secondary avalanches, until the whole
length of the available wire is covered in a glowing plasma. It gives a
very large pulse, but tells us only that a particle has hit the whole counter
somewhere on this sensitive area.

Charpak’s idea was to mix another gas, such as isobutane, ethane or
methane, with the argon in his multiwire chamber. This mixture absorbs the
secondary photons within a fraction of a millimetre of their origin. It
stops one avalanche from setting off another. The avalanche started by electrons
from a particle can be confined, not only to one wire, but also to one position
along its length. We can tell the position of the track in the direction
perpendicular to the wires (the x-direction on Figure 1) by seeing which
wire has the pulse. If we require two position coordinates at the same time,
there are several techniques that we can employ. We may, for instance, make
the sense-wires from a metal with a large electrical resistance, then measure
the relative sizes of the pulses at both ends of each wire to indicate how
far along the wire the pulse forms.

Multiwire chambers can detect many kinds of radiation. Energetic charged
particles will trigger avalanches in each chamber that they cross. Physicists
use a computer to build up pictures of the tracks of particles from the
hits in successive chambers. They are used like the tracks from bubble chambers
to reconstruct interactions between particles and their subsequent decays.

Numerous ‘daughters’ of the multiwire chamber are being used in particle
physics and astrophysics, many of them also invented by Charpak. For tracking
subatomic particles, the drift chamber has become the most popular. It economises
on the number of wires, which cuts the cost of the readout electronics by
measuring the time taken by the electrons to drift to a sense-wire. A scintillation
counter detects a particle once it has passed through the chamber, and starts
a timer that is stopped when the pulse comes from the avalanche on the sense-wire.
The drift distance may be as much as a metre, allowing the sense-wires to
be very widely spaced.

Multiwire chambers not only detect particles; they can also detect X-rays
very efficiently if another inert gas, xenon, is used in the chamber instead
of argon. David Anderson, now working at Fermilab near Chicago, added an
organic vapour called tetrakis (dimethyl amine)ethane, or TMAE, which made
the chambers directly sensitive to near-ultraviolet light. A number of research
groups are developing multiwire chambers loaded with TMAE to detect gamma-rays.
This could be particularly useful in medicine for a method of imaging called
positron emission tomography, or PET. This analytical technique can probe
the internal structure of objects, such as living bodies or small engineering
structures, producing a three-dimensional image of the interior of the material.

PET works by exploiting positrons, the antiparticles of electrons, which
are produced when some short-lived nuclear isotopes decay. The positron
emitting isotope can be used as a ‘tracer’ when injected into a person or
object. A positron emitted in flesh, or metal, or engine oil, will travel
a fraction of a millimetre before meeting an electron from the material
and annihilating into a pair of gamma-rays with a unique energy of 511 kiloelectronvolts.
The two gamma-rays go off in almost exactly opposite directions. A scanner
detects the gamma-rays using pairs of detectors on opposite sides of the
patient or sample. The detectors measure the positions of these gamma-rays
as they arrive, defining a line between each pair of conversion points.
A computer combines the lines from many pairs to reconstruct the distribution
in space of the starting points of the gamma-rays; in other words, the distribution
inside the body of the radioactive tracer emitting the positrons.

Doctors already use PET to study the circulation of the blood, especially
to live tumours or localised organs. By contrasting how the tracer isotopes
are taken up and distributed, they can obtain precise and immediate details
about the organism’s metabolism. Radioactive isotopes of oxygen are particularly
important, because they will concentrate wherever reactions are taking place.
PET can also help to measure how drugs behave in the body by revealing how
a drug concentrates in target organs, and where else it goes to cause side
effects. Eighty per cent of drugs can be labelled with positron-emitting
tracers, so doctors are predicting that PET could revolutionise pharmacology.

At the moment, the detectors place severe limitations on PET scanners.
The scanners use pairs of large scintillating crystals placed opposite one
another in a ring around the patient or sample. The crystal converts the
gamma-rays into visible photons at the point where they impinge. Each crystal
is viewed by a tightly packed array of photomultiplier tubes, connected
to some fast electronics. This system is called an Anger camera. The camera
recognises two gamma-rays from an annihilation because they arrive at their
individual detectors at the same time. The position of each gamma-ray within
its crystal is calculated from the average position of the pulses in the
array of photomultiplier tubes. The advantage of this method is its efficiency.
Heavy scintillating crystals, such as sodium iodide, detect a large fraction
of the gamma-rays that hit them, so doctors can quickly build up a tomographic
(‘slice by slice’) picture from a small amount of tracer. This keeps down
the dose of radiation received by the patient. But this technique has not,
so far, achieved an accuracy of position much better than 5 millimetres,
because the photomultiplier tubes are too big. This is not good enough to
reconstruct the fine details of organs.

Charpak and his colleagues think that multiwire chambers could improve
resolution. Alan Jeavons, now in the nuclear physics laboratory at the University
of Oxford has done some impressive high-resolution PET using chambers loaded
with sheets of lead to convert the gamma-rays into electron tracks. But
the efficiency of conversion is small, so large doses of radioactive tracer
have to be given to build up a precise picture of organs or to record rapid
variations in metabolism. Now, Charpak has suggested a better method. Barium
fluoride is a dense scintillating material, which gives out plenty of ultraviolet
photons when it absorbs gamma-rays at 511 kiloelectronvolts. A number of
research groups in the Netherlands, Belgium and Britain, including Eddie
Bateman at the Rutherford Appleton Laboratory (RAL) and Derek Imrie at Brunel
University, are developing multiwire chambers loaded with TMAE to detect
these photons. They hope to combine the efficiency of an Anger camera with
the positional accuracy of a multiwire chamber.

Machines are not as sensitive to large doses of radiation as people
are. Multiwire chambers doped with lead, built by Bateman and his team of
former particle physicists at the RAL, have been very useful in bringing
PET techniques into nondestructive testing. Mike Hawkesworth and his colleagues
from the University of Birmingham have used these techniques with Rolls
Royce to study how oil flows inside a jet engine. They have also studied
the dynamics of what is called a fluidised bed. This is a trough full of
sand which has a strong blast of air coming up through it from fine holes
in the perforated bottom plate. The chemicals and energy industries employ
fluidised beds in a number of processes, including burning coal in power
stations. The research team used a PET scanner to follow a small solid particle
carrying a radioactive tracer as it moved around inside the bed. In both
the jet engine and in the fluidised bed, PET is the only technique we have
that can reconstruct processes deep inside the equipment as they actually
happen.

Researchers are using multiwire chambers to study structural details
obtained from rapidly changing X-ray diffraction patterns. If you look at
a distant street lamp through the fabric of a nylon umbrella, you will notice
a pattern of dots around the position of the lamp. This diffraction pattern
tells you something about the structure of the fabric. There is a wider
spacing between the dots when the strands of the nylon are closer together.

Crystallographers have used such patterns to study the structure of
materials for many years. Traditionally, they have used X-ray tubes (like
those used by dentists). But now brighter and better sources of X-rays are
available from synchrotrons. These are circular accelerators that accelerate
electrons to produce very bright, so-called synchrotron radiation – a smooth
spectrum of electromagnetic radiation, ranging from hard X-rays, through
visible light to the far infrared. With such beams, researchers can produce
diffraction patterns of highly structured materials whose physical and chemical
nature changes rapidly with time. For example, they can study moving muscle.

A research team from the Laboratory of Molecular Biology in Cambridge
has been studying what happens when muscle fibres twitch. The researchers
use intense X-ray beams from the Synchrotron Radiation Source at the Daresbury
Labora tory in Cheshire to make hundreds of electronic ‘snapshots’ per second
of two simultaneous diffraction patterns. A fast multiwire chamber detects
the first, an X-ray pattern generated by diffraction from the fine network
of cross-bridges linking the muscle fibres. A charge-coupled device detects
the second pattern created as the fibres themselves diffract visible light.
This tells the researchers about the state of relaxation or contraction
of the muscle on a larger scale. This gives a moving picture of the progress
of the twitch, with simultaneous data on the cross-bridge links causing
it.

Finally, you can detect diffraction patterns not only due to X-rays
but also those from neutrons. ¿ìè¶ÌÊÓƵs working with the ISIS spallation
neutron source at the RAL and with the neutron source at the Institut Laue-Langevin
in Grenoble have made linear arrays of fine scintillators to pick up rows
of dots from neutron diffraction. Such methods are used, for example, to
study the structure of high-temperature superconductors. Researchers can
also detect neutrons by including certain nuclear isotopes in the gas of
the multiwire chamber or in its walls. Slow neutrons cause isotopes to break
up to form charged fragments which then drift towards the sense-wires.

As all these examples show, detectors originally developed for fundamental
particle physics have already become useful in more everyday applications.
It is worth remembering that three or four years ago, the government seriously
thought about ending Britain’s involvement in CERN and reducing our contribution
to some kinds of astronomy. The argument was that this would allow resources
to be concentrated on ‘exploitable and applicable research. . . leading
to . . . benefits . . . related to social, environmental or to government
policy (in 5, 10 or 20 years)’. These are quotations from a government discussion
document, called A Strategy for the Science Base, published in 1987.

The history of the development and the use of detectors illustrates
how dangerous it would be to follow this line of reasoning. Because ‘big
science’ (particle physics and astrophysics in particular) is fundamental,
it cannot help but be pervasive, not only in the long-term significance
of its results, but also in its short-term technical and educational spin-offs.
PET scanners and X-ray diffraction detectors are today’s timely and applicable
technologies. We have them because the nuclear, particle and astrophysicists
developed the techniques for purely ‘curiosity-oriented’ research. And almost
all of the people who built them have PhDs in particle physics or astrophysics.

* * *

Detecting particles in the solid

MODERN silicon technology can put the most intricate of fine patterns
onto a small chip. One use of this is in charge-coupled devices, (CCD)s,
which are used both for television cameras and to detect particles (see
an article in next week’s issue by Christine Sutton).

Another detector, with faster readout than the CCD, is the silicon microstrip.
A charged particle reveals its track by releasing electrons in the silicon.
Parallel strip electrodes, running a few tens of micrometres apart across
the surface of the silicon, collect the electrons. Hundreds of tiny amplifiers
have to be connected by fine wires to the pattern of strips.

Physicists have employed these detectors to measure production of short-lived
elementary particles containing charmed quarks. But the complicated array
of electronics around each tiny detector makes them very difficult to use.

Recently, researchers at Imperial College London, in Munich, and elsewhere
have succeeded in making silicon drift chambers that may overcome this complication.
As with the gaseous drift chamber, electrons drift over ‘long’ distances
(a few millimetres in this case, rather than tens of micrometres in a microstrip
detector) to relatively few electrodes, and the drift time gives the precise
position of the track of the original particle.

David Miller is in the department of physics and astronomy, University
College, London. He is a member of the OPAL collaboration, which is producing
many of the new results coming from the electron-positron collider, LEP
at CERN.