| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Innovation |
1 Department of Radiology and Radiological Sciences, Vanderbilt University
Medical Center, 1221 21st Ave. S, Nashville, TN 37235-2675.
2 Vanderbilt University W. M. Keck Free Electron Laser Center, PO Box 351816
Station B, Nashville, TN 37235-1816.
Received January 29, 2003;
accepted after revision April 30, 2003.
Address correspondence to F. E. Carroll.
Introduction
 Top Introduction Materials and MethodsResultsDiscussionReferences |
|---|
Historically, X rays for imaging and crystallography have generally been produced through the use of bremsstrahlung and line X rays from electrons impinging on a metallic anode. Such sources are inexpensive, simple, and robust but provide little control over the X rays produced. More recently, synchrotron sources have been used for both applications, with good results. Unfortunately, synchrotrons are large, expensive facilities with less than ideal beam geometry and are therefore not entirely practical for routine imaging applications.
The excellent results of experiments with monochromatic sources [3–5] show the desirability of improving on the current broadband X-ray imaging practice. No alternative has existed for experiments that need to operate at various X-ray energies. The availability of such a source may fundamentally change the practice of X-ray imaging and provide much wider availability of tuned X rays to crystallographers.
A compact source of pulsed tunable monochromatic X rays has been designed, built, and tested. This device can deliver "hard" X rays from 10- to 50-keV at narrow bandwidths (1–10%), with a flux of 1010 photons in each 8-psec pulse. These are produced in a cone-beam area geometry useful for human imaging, small animal imaging, protein crystallography, and nondestructive testing in industry. The machine integrates a laser with a linear accelerator (LINAC) and can be used in an unshielded environment.
The source described here is a tabletop-terawatt (T3) laser-based Compton backscattering system, which uses few-joule pulses from a 1,052-nm laser to collide with a 20- To 50-MeV electron beam to produce an intense pulse of narrowband X rays. The entire system footprint is 4 m wide by 10 m long, and it requires no shielding vault. It produces X rays in a small-angle cone-beam geometry in the 10- to 50-keV range, with up to 1010 photons in an 8-psec pulse, which is sufficient flux for medical and industrial imaging to be performed in a single shot. This source is certainly not the first Compton backscattering or laser-synchrotron X-ray source built. Experiments have been carried out at a number of the large accelerator facilities [6, 7] that have produced modest fluxes of photons over a wide range of interesting energies. Also, sources similar in concept to this one have been proposed [8, 9] and operated on a small scale. However, none of these sources has been designed and built in a small practical form and with a high enough flux to be deployed as a common laboratory-scale or clinical resource. Further, most of the current generation of sources produce high levels of background radiation from the linear accelerator and require the source to be embedded in a shielding vault.
The source at Vanderbilt University has its roots in a project that was built as an add-on to the free-electron laser at Vanderbilt that was proposed in 1987. It produced a flux of approximately 104 X rays per second in 1998 [10]. However, it proved difficult to tune the electron beam for the X-ray production and to operate the laser in an optimal manner at the same time. The machine required the linear accelerator to be in a vault, so servicing the experiments was slow and access to the X rays that were produced required the use of a complex beam-transport system that caused high flux losses. Shortly after the feasibility of the system was proven, the prototype was decommissioned to allow the design and construction of its successor, which is the source we describe.
|
TopIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The theoretic X-ray yield of this source can be computed directly from the
Thomson scattering cross-section and some basic beam-geometry and flux
considerations. This is a straightforward computation. The Thomson
cross-section (
) is
 |
 |
Vx), and Vx is the
fractional energy spread of the X-ray output that is useful to the
application. For this system, with 10 J of laser light, 1 nC of electron
charge, and an f /10 final focus of the laser, the output is
approximately 5 x 109 X rays into a 10% bandwidth.
Another common figure of merit for such systems is the brilliance in
 |
 |
|
|
The Machine
The system now in use at Vanderbilt consists of several basic
subcomponents. Because the unit produces an electron beam, an infrared light
beam, and an X-ray beam, the subcomponents can be grouped by their functional
tie to each of these beams.
The electron beam.—A master timing oscillator provides an 81.6-MHz laser-mode-locking signal and the 2,856-MHz master signal for the radiofrequency drive circuits. A timing-monitoring and feedback system watches for and corrects any timing drift caused by the transit time of signals through the laser and radiofrequency system, allowing few-picosecond timing consistency throughout the system. A neodymium: yttrium lithium fluoride (Nd:YLF) laser, pulse compressor, and frequency quadrupler provide up to 200 µJ of 263-nm light in 5-psec pulses that strike a copper photocathode. This radiofrequency drive system and bare copper photocathode electron gun–linear accelerator produce a high-brightness electron beam using characteristics of the Brookhaven accelerator test facility (ATF) electron gun system (Advanced Energy Systems, Medford, NY). A superconducting solenoid focusing magnet is used downstream from the linear accelerator to focus the electron beam to its point of interaction with the infrared light beam.
The infrared beam.—A 200 fsec Ti:Sapphire Tsunami seed laser (Spectra-Physics, Mountain View, CA) runs at 1,052 nm and drives a stretcher–regenerative Spitfire amplifier combination (Positive Light, Los Gatos, CA), which produces a 480-Hz train of 200-µJ pulses stretched to approximately 1 nsec. This laser seeds both the Nd:YLF laser used to generate the electron beam and a multistage Nd: glass amplifier. This Nd: glass laser (Positive Light) and pulse compressor provide up to 10 J of 1,052-nm light compressed to 8-psec infrared pulses for the collision with the electron beam. This infrared beam is focused on the interaction zone after being deflected by a beryllium mirror pointed upstream into the linear accelerator.
The X-ray beam.—An interaction zone and beam alignment system are used where the electron beam and photon beam collide head-on in a 50-µm spot. The X rays produced traverse the infrared beryllium mirror, and a thin beryllium exit window is used to allow the X rays out of the vacuum beam line into the imaging area.
Some components of this hardware deserve more detailed commentary because they have been specifically designed to make this system robust and stable for daily operation. Because this facility will become an open user facility, extra efforts have been taken in the design to reduce instability in the system that could result from environmental changes and parameter drift.
The timing, monitoring, and feedback system previously described in the Electron Beam section is particularly critical to the system. The transit time of an optical pulse from the seed laser to the interaction zone is close to 1 µsec and needs to be stable to approximately 2 psec. Air, with an index of refraction of roughly n– 1 = 3 x 10–4, contributes a 3-psec delay for each 1% change in atmospheric pressure. Also, small fluctuations in the pressure of the fill gas in the wave guides and small changes in the temperature of the electron gun can alter the phase of the electric field in the gun, resulting in energy shifts in the electron beam. To avoid having to develop heroic environmental controls for the system, we have opted instead to monitor these timing variations and correct for them dynamically with a set of phase comparators that monitor the various timing offsets relayed back to the master oscillator. This end-to-end–timing monitoring system allows the system to remain stable without adding complexity or requiring much operator effort.
The electron gun has also been modified slightly from the ATF design to
make it more robust, although it requires high photocathode-drive laser
energy. The modern trend in electron guns is often to pick
high–quantum-efficiency materials for the photocathode to minimize the
required laser drive energy. Unfortunately, these materials, such as
magnesium, form oxide layers that need to be cleaned via laser ablation to
maintain optimum performance. Instead, this machine uses bare polished copper
that requires no intentional laser cleaning. The operating vacuum in the gun
is reasonably good (
3 x 10–7 Pa), so the copper
may slowly clean itself under the normal laser irradiation, although it is
assumed that—effectively—the machine uses a copper oxide cathode
surface. Using a laser energy of approximately 100–200 µJ
gives an electron pulse with a charge of 1 nC, which the machine needs in
order to operate. Under these conditions, the cathode never needs any
reprocessing to maintain sufficient emission, which saves a good deal of time
and effort.
Imaging with the device.—Images have been made using various phantoms including a human hand phantom consisting of a preserved human hand skeleton embedded in plastic, multiple custom-made breast phantoms, contrast agent phantoms, cell button preparations, and whole animals. Each image is obtained using a single 8-nsec pulse of X rays.
Film-screen combinations currently in use in standard X-ray imaging are not a good match to this new beam because of the extremely high speed of the X-ray pulse. For these preliminary imaging experiments, several detectors have been used, including a fluorescent image plate coupled to an image intensifier, a tapered fiber optic bundle, and a charge-coupled device camera (Nanocrystal Technology, Briarcliff, NY); a MAR345 protein crystallography camera (which incorporates an Agfa computed radiography plate [Agfa, Morstel, Belgium]); and a detector (Hologic, Bedford, MA) (which consists of amorphous selenium on thin-film transistor array). Because the beam is not collimated for these images, the X-ray energy spectrum is slightly variable, with the highest energy photons in the center of the beam and varying over the bandwidth of the beam at the time of imaging. In this case, the bandwidth is set at 1%.
|
TopIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
The actual photon yield of the system is variable and depends on the considerations previously noted and the actual beam quality and alignment, which are adjustable. We have not yet accurately measured the absolute output flux of the system. This requires careful consideration because of the extremely highspeed pulsed nature of the beam and the measurement devices that are currently available. These measurements are all the more difficult because some high-energy photons from stray electrons in the electron beam mix with the Compton backscattered photons and alter the readings. We are, therefore, in the process of verifying our measurements. However, any single 8-psec pulse has sufficient photon flux to produce high-quality X-ray images, which is the first priority for this machine.
Images produced with this machine are presented in Figures 4A, 4B, 5A, 5B, 6, 7. Because the energy of the electron beam can be adjusted from 20 to 50 MeV, the energy of the X rays produced is tunable. Figures 4A and 4B clearly show the difference between two energy levels. At 19 keV (Fig. 4A), the finger of the phantom shows good soft-tissue detail, but the bone is radiopaque. At 29 keV (Fig. 4B), the bone becomes X-ray transparent, revealing the cortex and medullary cavity inside, but the soft tissues become less distinct. Images of a breast phantom composed of breast-equivalent materials of differing glandular percentages, calcifications of different sizes, and paraffin are shown in Figures 5A and 5B. The monochromatic image (Fig. 5A) clearly depicts "lesions" not easily seen on the polychromatic image (Fig. 5B). Figure 6 shows use of the monochromatic beam in a mouse at various energies, and Figure 7 shows the acquisition of an image of the entire mouse in a single 8-psec X-ray shot using the conebeam-area geometry of the beam.
|
|
|
|
|
|
To retune the machine from lower to higher energy settings currently requires approximately 10 min of operator time, but this could be reduced to less than a minute with computer-based saving and recalling of the few parameters requiring adjustment.
|
TopIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Picosecond pulses become valuable for studying physical, chemical, or
mechanical processes that occur on the X-ray beam's picosecond time scale.
Because a machine of this type is easily scaled to higher photon energy by
simply lengthening the LINAC, it should be simple to adapt this technology for
other applications. Because the X-ray photon energy increases as the square of
the electron beam energy, doubling the LINAC length quadruples the X-ray
energy, and shortening it can make it useful for other applications in
medicine and biologic research. Development of these machines for protein
crystallography [14] and
high-resolution small-animal imaging is already under way. Because it can be
modified to deliver high average power as well as high peak power, this device
can be used to perform protein crystallography at home without restriction to
the Cu k
of the standard X-ray tube home crystallography
machine. It is likewise suited to perform multiple anomalous dispersion and
Laue crystallography. This use is important for institutions with extensive
proteomics laboratories or drug-development programs.
A more compact device has been designed and will soon be built that delivers X rays from 8- to 50-keV at 20 Hz yet is easily scalable to higher energies above the diagnostic range for other uses. The small focal spot of this method of X-ray production makes this type of device an ideal source for performance of phase-contrast imaging [15–18].
Monochromatic or—more properly—narrowband X-ray imaging has shown great promise for improving the quality of images used in medical diagnostics while potentially reducing the dose to patients [19–21]. Also, time-of-flight techniques have the potential to reduce scatter and improve image contrast. Previously, few sources have been capable of producing the requisite X-ray beam parameters to realize these goals. Such sources were well developed and useful in the laboratory but were not suitable for widespread clinical applications. The source we describe will provide access to pulsed tunable X rays to a wider range of users and, ultimately, accrue benefits for clinical patients, X-ray crystallographers, industrial radiographers, and others who routinely use X rays. It has high brightness, short pulses, and a variable bandwidth, and it is small enough to fit in a modest laboratory or clinical facility. We believe that the time has finally come for practical applications of monochromatic X rays.
Acknowledgments
We thank Virginia Adcox, Vera Merriweather, Scott Degenhardt, Travis Henry,
Adam Winchell, Chris Baughman, Matt Mellon, Gary Shearer, and the personnel of
the Vanderbilt University W. M. Keck Free Electron Laser Center.
|
TopIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Keyrilainen, M. Fernandez, M.-L. Karjalainen-Lindsberg, P. Virkkunen, M. Leidenius, K. von Smitten, P. Sipila, S. Fiedler, H. Suhonen, P. Suortti, et al. Toward High-Contrast Breast CT at Low Radiation Dose Radiology, October 1, 2008; 249(1): 321 - 327. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
