Introduction

X-ray tubes have been used ever since Roentgen’s discovery of the new kind of radiation in 1895, and they have been developed into remarkably reliable and useful devices. Surprisingly, the brightness of these X-ray sources have not shown any radical increase in the course of time, the main innovations being the hot filament tube, invented by William Coolidge in 1913, and the rotating anode tube, devised by R. E. Clark in 1934. A major break-through occurred in 1947, when F.R. Elder, A.M. Gurewitsch, R.V. Langmuir and H.C. Pollock first observed synchrotron radiation as visible light from the General Electric 70-MeV synchrotron.

Synchrotron X-radiation is emitted by electrons or positrons orbiting at high energies in synchrotrons or storage rings. The first sources were true synchrotrons utilized in a parasitic manner, the machines being built and run by the high-energy physicists. Second generation facilities were purpose built electron storage rings, dedicated as light sources. Present third generation sources are based mainly on insertion devices, called wigglers and undulators. Examples of such rings are the Advanced Light Source (ALS) in Berkeley and the Advanced Photon Source (APS) in Argonne, both in USA, the European Synchrotron Radiation Facility (ESRF) in Grenoble, France, and the SPring-8 facility in Japan.

During the last three decades, synchrotron radiation has proved its usefulness in many different fields of science, ranging from physics, chemistry and biology to materials science and medical diagnostics. Advances in the creation, compression, transport and monitoring of bright electron beams have made it possible to base the next generation of synchrotron radiation sources on linear accelerators rather than on storage rings. This development has been spurred by the availability of advanced low-emittance electron guns, linear accelerators in the GeV energy range, and precise undulators based on new materials for permanent magnets, such as SmCo5. Therefore, the next big leap forward with high-power light sources will most probably be the introduction of free-electron lasers producing VUV and X-radiation with unprecedented brightness and coherence.

Undulators

Synchrotron radiation is produced by electrons or positrons travelling along a curved trajectory with a velocity close to that of light. The radiation is emitted in the forward direction, tangentially to the orbit and confined within a narrow cone, having an opening angle y given by

where the relativistic factor is being the electron mass and rest mass, respectively, E the electron energy and c the velocity of light. Thus, for electron energies in the GeV range, the angular divergence y of the emitted radiation is less than 1 mrad.

The traditional device for producing synchrotron radiation in the first and second generation machines is the bending magnet. In the third generation facilities, the most important sources are devices inserted in the alternating polarity, causing the electrons to oscillate straight sections of the electron storage ring. The insertion devices are arrays of dipole magnets with around their linear trajectory (Fig. 1). In the following it is assumed that the symmetry axis of the insertion device is the z-axis, and that the magnetic field is directed along the x-axis. The magnetic flux density Bx is assumed to vary sinusoidally with the distance z along the undulator:

where B0 is the amplitude of the magnetic field, and l0 is the magnetic period. From the equation of motion of an electron in a periodic magnetic field, it can be shown that the maximum deflection angle a (Fig. 2) is given by

where e is the electron charge. The so called undulator parameter K is defined as

For an undulator, K is of the order of one or less, i.e. the electron beam is kept within the angular emission of the synchrotron radiation from each wiggle. Thus, the radiation from different points within the undulator can interfere constructively, giving rise to a series of monochromatic spectral lines. When K >> 1 (typically > 20), there are no coherence effects, and the insertion device is called a wiggler.

The interaction between the speeding electrons and the electromagnetic field of the emitted radiation has to be calculated using advanced quantum mechanics [1]. Fortunately, the resonance condition can be understood from a simple model. While the radiation emitted by the electron in the forward direction propagates along the z-axis with the speed of light, the electron moves at an average velocity determined by its energy and its path length per magnetic period. After one period, the electron lags behind the radiation emitted at the corresponding point in the preceding period by a distance corresponding to one wavelength of the radiation. This condition can be formulated in the following way:

where l is the wavelength of the emitted radiation, and vz is the velocity of the electron along the z-axis. The total velocity of the electron is

where vy is the y-axis velocity component caused by the periodic magnetic field. A relativistic calculation using equations (5) and (6) shows that the resonance wavelength l is given by

The wavelength is determined by the electron energy through g in equation (7). Tuning is then possible by varying the undulator parameter K. This is most easily done by varying the gap of the magnets, thereby changing the magnetic flux density B0. Higher harmonics are obtained by substituting nl (n = 1, 3, 5, …) for l in equation (7). Only odd harmonics appear on axis. The free-electron laser

A free-electron laser (FEL) can be realized by having an undulator situated in an optical cavity defined by mirrors
(Fig. 3). The periodic magnetic field causes the electrons to undergo transverse oscillations. Thus the electrons will have a velocity component in the direction of the electric field vector of the light in the cavity, and there will be an interaction between the electron beam and the light. The interaction induces a density modulation of the electrons (microbunching) with a periodicity equal to the wavelength of the light. When this occurs, the electrons will emit synchrotron radiation coherently, i.e. the intensity of the radiation is proportional to the square of the number of electrons participating, rather than proportional to the number of electrons as is the case with normal synchrotron radiation. In addition, the intensity of the light interacting with the electron beam stimulates the emission of synchrotron radiation, leading to a further increase of light intensity and a lasing action of the device.

In most present free-electron lasers the light from several passes of the electron beam through the undulator is stored in the optical cavity. Free-electron lasers using optical cavities are successfully operating at wavelengths l ³ 200 nm. Extending these devices to shorter wavelengths poses problems, however, because of the lack of high-reflecting mirrors for VUV and X-rays. Recently, it has become possible to consider another mode of operation, based on a single pass of a high-brightness electron bunch through a long undulator. No optical cavity is needed, and lasing action is achieved by a self-organizing process called Self-Amplified Spontaneous Emission (SASE) [3].

In the SASE mode, the laser has to start up from noise. In the first part of the undulator the electrons radiate incoherently, the power growing linearly with distance. Later, the interaction between the radiation and the electron beam leads to an amplification of the spontaneous emission, giving rise to coherent radiation. The power will then grow exponentially with distance until saturation is reached (Fig. 4).

The basic concept of SASE has been demonstrated experimentally in the microwave region. An experimental verification in the short-wavelength range has been slow due to the tight requirements on the electron beam characteristics. With the development of new low-emittance electron guns, linear electron accelerators have become an attractive option for short-wavelength SASE. Serious considerations on the realization of VUV and/or X-ray free-electron lasers are now carried out at the Stanford Linear Accelerator Center (SLAC) in the USA, and the electron synchrotron DESY in Hamburg, Germany. In both places a number of workshops on this topic has been held [4-7].

The Stanford machine would utilize part of the existing 50 GeV linear accelerator. In the near future, the first 2/3 of the 3 km linac will be used to inject electrons and positrons in the soon to be completed PEB-IIB-Factory. The last 1/3 of the linac will then be available for the production of an up to 15 GeV electron beam that could be used to create a hard X-ray free-electron laser. A collaboration of American and Japanese scientists is investigating the technical issues and the potential scientific applications of the radiation produced. Single-pass free-electron laser projects are also planned at UCLA and BNL, utilizing 20 and 230 MeV linacs to reach wavelengths of 10 mm and 10 nm, respectively.

At DESY it is proposed to make use of the unique electron beam properties of the TESLA Test Facility (TTF) to construct a VUV free-electron laser based on the SASE principle. The project will proceed in two phases. In the first phase, a 390 MeV superconducting linear accelerator will be used to drive a free-electron laser operating at a wavelength of 42 nm. This facility is intended to test various components and, last but not least, to prove the SASE principle for the first time at short wavelengths. The phase I free-electron laser is under construction. The first cryomodule has been successfully tested, and first tests with the electron gun will take place in 1998. In the same year, two additional cryomodules will be mounted. After the installation of the undulator, the first experiments are expected to be carried out in the spring of 1999. In the second phase, the electron energy will be raised to 1 GeV and the undulator will be extended, producing radiation of 6.4 nm wavelength. An experimental hall completes the construction, and the facility will be available to external users as well. In a more distant future the DESY team is planning a hard X-ray free-electron laser for wavelengths in the 0.1 nm (1 Å) range. The latter facility would be integrated into a 500 GeV e+e- linear collider.

The scientific case

VUV and X-ray free-electron lasers would produce coherent radiation orders of magnitudes greater in power and brightness than present 3rd generation synchrotron radiation sources. The large increase in brightness, coupled with spatial coherence and the very short pulse duration, will certainly open new and interesting research possibilities in physics, chemistry, biology and allied sciences.

The high brightness of the free-electron laser source would facilitate the study of diluted systems and other weak scattering systems, such as atom clusters, magnetic scattering, surface atoms and biological materials. The spatial coherence would be beneficial to all kinds of imaging techniques, such as X-ray microscopy, holography and interferometry. The high brightness per pulse is certainly the most outstanding feature of the free-electron laser light. It will permit scattering and spectroscopic measurements based on a single pulse, having a duration of less than 1 ps. Thus, time resolved experiments would be possible in the sub-picosecond range. While the coherent radiation is monochromatic, the free-electron laser also produces a broadband pulse of radiation, incoherent but quite intense. This white radiation could be used for ultra-fast Laue crystallography, e.g. for the study of proteins.

In conclusion, a wide range of new and exciting experimental possibilites would be opened up by VUV/X-ray free-electron lasers. Such devices should lead to the same sort of revolutionary developments in VUV/X-ray studies of matter that was produced in optical studies by the introduction of lasers for visible light.

References

1.   J.M. Madey, Stimulated Emission of Bremsstrahlung in a Periodic Magnetid Field. J. Appl. Phys. 42, 1906          (1971).
2.   J.M. Ortega, The CLIO Infrared FEL Facility. Synchrotron Radiation News 9(2), 20 (1996).
3.   K.-J. Kim and M. Xie, Self-amplified spontaneous emission for short wavelength coherent radiation. Nucl. Instrum.          Methods A 331, 359 (1993).
4.   J. Arthur, G. Materlik, R. Tatchyn and H. Winick. The SLAC Linac Coherent Light Source. Synchrotron Radiation         News 7(5), 22 (1994).
5.   M. Cornacchia, The Linac Coherent Light Source. Synchrotron Radiation News 11(1), 28 (1998).
6.   J. Feldhaus and B. Sonntag, The Vacuum Ultraviolet Free-Electron Laser at DESY. Synchrotron Radiation News          11(1), 14 (1998).
7.   A Superbrilliant X-ray Laser Facility. http://www.desy.de/~wroblewt/scifel.html

An obstetrician friend bumps into you at a party and starts recounting a recent challenging case. One of her patients reported suffering from lethargy and abdominal pain. This in itself was not immediately alarming, however, given the patient’s history of two early mid-trimester miscarriages and several probable 1st-trimester miscarriages, the cause of this abdominal pain was investigated. The patient was found to be suffering from lead poisoning, as indicated by numerous factors, the most telling of which being elevated blood-lead levels. Given the possible health risks to the foetus and subsequently a nursing child, the source of this lead exposure was sought. However, therein lay the mystery. Thorough analysis of the family environment revealed no source of lead contamination, and her husband and two other children had normal blood-lead levels. It was revealed that the patient had suffered acute lead encephalopathy when she was 18 months old but has since had no known lead exposure.

Your friend speculated that with calcium turnover greatly increased during pregnancy, the patient was suffering from lead poisoning due to the mobilisation of lead stores from her bone. The obvious choice of treatment was chelation therapy, but without concrete evidence of high bone-lead levels it was a difficult decision to treat due to the potential harm to the foetus from chelation. In this case, a means for assessing bone-lead levels non-invasively, in vivo, would have provided an invaluable tool for your friend in helping the patient as soon as the problem was identified, rather than waiting until after childbirth during which time the child may have been lost. This is precisely the type of motivation for ongoing research at McMaster University and elsewhere in the field of in vivo trace element analysis.

For the past 20 years, the application of x-ray fluorescence (XRF), a standard elemental analysis technique, has been investigated for the purposes of measuring many toxic elements in vivo. In some cases, the source of exposure may be environmental. However, in the majority of cases, the development of measurement systems is driven by the need for better monitoring and health care for individuals exposed to toxic elements in occupational settings, e.g. Pb, Cd, U, and Hg. In addition, the need for measuring the in vivo concentrations of such elements as gold and platinum has arisen due to exposure to these toxic metals through medical procedures; gold salts are used in treating rheumatoid arthritis and platinum-based drugs are a common choice in chemotherapy. Research has focussed on developing systems to measure these elements generally in one of two sites, the bone matrix or the kidneys. The bone matrix represents the site of long term retention for elements such as lead and uranium. Therefore, measurements in this site give insight into the history of exposure of a given individual to these toxic metals. Other elements, such as Pt, Cd, Hg and Au, are better measured in the kidneys, as this organ can be a major retention site as well as often the organ to suffer damage due to metal accumulation.

XRF makes use of the unique electron energy levels to identify the elemental composition of an unknown sample. The sample is irradiated with x-rays or g-rays that interact through scatter and photoelectric absorption. When the incident radiation has energy greater than the K edge of an element, there is a significant probability that the element will absorb the photon and eject a K shell electron. With the element in an excited state, it can return to the ground state through emitting an Auger electron or an x-ray, with an energy corresponding to the difference between the K shell binding energy and that of an outer electron shell. Since the energy levels are unique to each element, the x-ray series emitted by an element has a characteristic energy signature that enables its identification. Furthermore, through careful calibration, the quantity of these x-rays detected can be used to determine the absolute amount of an element present in the sample.

In vivo XRF-based systems have been successful in measuring toxic elements of interest with high atomic numbers, e.g. Pb, Au, Hg, U, Pt, and to some extent Cd. This is because three critical parameters increase with increasing Z; fluorescence yield, K edge energy and characteristic x-ray energy. The higher fluorescence yield, as the name suggests, means that there is a greater probability of characteristic x-ray emission with every shell vacancy created instead of de-excitation through Auger electron emission. Therefore, there is an increased signal produced per unit irradiating flux. Higher K edge energies imply that a higher energy photon can be used to irradiate the subject. This results in the use of more penetrating radiation. Similarly, higher characteristic x-ray energies with higher Z elements result in greater penetrating capability of the outgoing signal. This is important in non-invasive measurements as the site of metal retention, bone, kidney, or liver, is shielded by a significant thickness of attenuating material.

Given that the element of interest has sufficiently high Z for XRF to be feasible, typically greater than 50, the optimisation of this in vivo probe must be undertaken for each element. There are a number of parameters that must be selected with a firm understanding of the underlying radiation physics principles. The first choice is whether the subject will be irradiated by g-rays from a radionuclide source or polarised x-rays produced by an x-ray tube. Polarised x-rays have been used in the development of systems to measure Cd, Hg, Pt, and Au primarily. The major difficulty in all XRF systems is that the element of interest is present in trace quantities within a low Z matrix. Therefore, the vast majority of detected photons are scatter events that only give rise to spectral background and increase the count rate in the detector. Polarised photons can be used to minimise the number of scattered events detected, as there is a much-reduced probability of scatter along the direction of polarisation. The isotropic characteristic x-rays are detected with a suppressed background when the detector is positioned along this axis, thereby enhancing the signal to noise ratio.

The pros and cons of these two classes of XRF systems can be looked at in parallel. A successful radionuclide-based system uses a g-ray source with an energy that is just above the K edge of the element under investigation, as this corresponds to the greatest photoelectric cross section. Once a source has been chosen with an appropriate g-ray energy, a sufficient half-life such that the source does not require frequent replacement, and minimal additional radiation emissions that merely give rise to subject dose, the irradiation geometry must be optimised. The source-sample-detector geometry is set such that the angle of scatter from source to detector gives rise to the Compton scatter distribution in the spectrum as far removed in energy from the characteristic x-rays as possible in order to reduce the background under the signal. These are essentially all the variables that can be optimised in a source-based system and therefore the sensitivity of the system is largely dependent on the availability of an appropriate radionuclide. This is the main reason for the outstanding performance of the 109Cd lead measurement system – with a g-ray that is only 30 eV above the K edge of lead, a reasonably long half-life of 462 days, and a lower energy photon emission that is readily shielded to reduce subject dose, 109Cd is the ideal radionuclide for measuring lead. Furthermore, the energy difference between the 88 keV g-ray scattered through ~ 160o and the lead x-rays is such that the signal is located in a reasonably low background portion of the detected spectrum, see fig. 1.

With polarised systems, the energy of the incident radiation is also a variable to optimise. However, a 90o source-sample-detector geometry must be chosen as this corresponds to the direction of initial polarisation and therefore minimal scatter. This limits the flexibility with which a fluorescing source can be designed since the energy dependence of the background spectrum now can only be altered by changing the energy dependence of the incident spectrum. Therefore, appropriate tube voltage, polarising material, and beam filters must be selected such that the incident energy distribution corresponds to high photoelectric absorption in the target element and results in the appropriate positioning of the characteristic x-rays in the background spectrum detected at the 90o scatter angle. The coupling of these factors in the polarised systems is the reason for the poor performance in polarised uranium measurements.

XRF measurements of lead concentrations in bone have been taking place clinically for many years now. This tool has become extremely useful in monitoring occupational exposure in workers at risk as it provides a measure of their long-term lead exposure and a means to ensure that their job is not putting their health at risk. In vivo XRF has allowed researchers to demonstrate that bone-lead levels are a direct measure of cumulative lead exposure and repeat studies of the same populations over many years are beginning to shed new light on physiological parameters such as the biological half-life of lead in bone. Furthermore, it has been demonstrated through these measurements that bone lead can be a source of endogenous lead exposure, just as your obstetrics friend speculated in her challenging case. It is hoped that with continued research this tool may become readily available to aid in occupational monitoring, optimising therapeutic procedures making use of gold (chrysotherapy for rheumatoid arthritis) or platinum (chemotherapy with cisplatin), as well as potentially monitoring environmental exposure to a wide range of toxic heavy metals.

Previous

Main

Next