Synchrotron Light and the Lyncean Compact Light Source
Conventional synchrotron light sources store multi-GeV electron beams in large rings of magnets to generate intense, bright 1 Å wavelength radiation. In a third generation synchrotron, the radiation is produced as the electron beam passes through a special magnet called an undulator magnet. An undulator magnet with a 2 cm-period requires an electron beam with an energy of about 5 GeV and a storage ring that is about 300 m in diameter.
The Compact Light Source uses a “laser undulator” to accomplish the same result. A laser beam colliding with an opposing electron beam has the same effect as an electron beam passing through an undulator magnet. The electric and magnetic fields of the laser beam cause the electron to wiggle and induce a radiation spectrum similar to that from a long undulator magnet. This radiation is often referred to as Inverse Compton Scattering. If a laser beam with a wavelength of one micron is used, the electron beam energy necessary for 1 Å radiation is only about 25 MeV. This reduces the scale of the device by a factor of about 200 and results in a storage ring with a footprint comparable to that of a large office desk.
A model of the Lyncean CLS is shown in Figure 1. An electron pulse is generated using an RF photocathode. This electron pulse gains energy by passing through accelerator sections, similar in size to the accelerators used in medical linacs. Once a chosen energy is reached, the pulse is injected into the storage ring. The storage ring is designed to allow a bunch to circulate for about one million turns. Periodically, a new pulse is injected to refresh the stored beam in order to maintain a steady, high quality X-ray output beam.
On one side of the storage ring, the electron beam is transversely focused to a small spot. This side of the ring also serves as one path of an optical cavity for a laser pulse. A laser resonantly drives the optical cavity to build-up a high-power laser pulse. By using an enhancement cavity, for example, a 10W laser can build up about 100kW of circulating power. The cavity length is adjusted so that round-trip travel time of the photon pulse in the optical cavity is equal to the revolution time of the electron bunch in the storage ring. The electron bunch and the laser pulse collide in the interaction region producing a burst of X-rays which are directed in a narrow cone in the same direction as the electron beam.
The native X-ray spectrum of the laser pulse/electron bunch collision is equivalent to that of a 20,000 period undulator magnet. The X-rays can be focused using conventional X-ray optics down to a size of about 45 microns. Fine-tuning of the X-ray energy (for scans near absorption edges) is achieved with a monochromator adjustment just as with a synchrotron beamline. A large change in X-ray energy, necessary for targeting different elements, is achieved by tuning the electron beam energy. The configuration described above operates with a similar photon flux up to X-ray energies of many tens of kilovolts. The naturally narrow bandwidth of the X-ray spectrum means the total produced X-ray power is already within the energy bandwidth useful for most X-ray applications. And since this power is only a fraction of a Watt, the relatively low power minimizes unwanted background scatter and simplifies X-ray optic design.
The Lyncean CLS X-ray beam has properties that are well suited to many conventional X-ray applications as well as some new ones. Although the initial incentive for this technology development was macromolecular crystallography, the Lyncean CLS provides a tunable, monochromatic, spatially coherent beam that may either be focused for general diffraction experiments such as powder diffraction and small angle scattering, or unfocused for imaging large samples. The large area beam can take direct advantage of the high quality native Lyncean CLS beam to provide high resolution tomographic imaging of biological specimens. The spatial coherence of the source enables new methods of contrast beyond conventional radiography, opening new avenues of X-ray science for medical applications.
Perhaps the most exciting new applications for the Lyncean CLS are in health care. New biological imaging techniques that provide exquisitely detailed images of soft tissue are being developed at the synchrotron facilities. The Lyncean CLS matches key aspects of the X-ray quality of these beamlines, but at a cost and scale that makes clinical applications of these powerful techniques practical. We believe that the Lyncean CLS will ultimately improve our nation’s health and impact millions of individual lives through better understanding of disease, more effective drug development, and by enabling clinical applications of emerging new techniques for biological imaging.
The Lyncean CLS has screened and collected data sets on several proteins at X-ray energies ranging from 12 keV to 18 keV. Details of the CLS’s first diffraction dataset on a novel protein are in the published paper with the ATCG3D collaboration. The protein structure was solved and published to the RCSB Protein Data Bank (3IFT).
Lysozyme example (March 2012)
As a benchmark test, a data set collected on Lysozyme was taken and solved using standard software. Diffraction data were measured on a Rayonix SX165 detector from a single frozen tetragonal lysozyme crystal, in frames of 1° rotation. Each frame was exposed for two 5 sec images and averaged. The energy was 15.2 keV (0.8157 Å), with a crystal-to-detector distance of 127.5 mm. A total of 120° were recorded. The data were integrated with both XDS and MOSFLM, and scaled with SCALA in the CCP4 package. The data reported below were integrated with XDS, but the statistics from the MOSFLM-integrated data were not significantly different. Also, several other data sets were recorded from other lysozyme crystals, with comparable results.
Table 1. Lysozyme data reduction statistics
|Dmin (Å)||Rsym||Rmeas||Rpim||Mean I/σ(I)||Nmeas||Nref||Multi-plicity||% complete|
These data were used for molecular replacement phasing using lysozyme structure PDB ID 3A8Z. Water, glycerol, Na+ and Cl- were removed from the model, and PHASER was run using data in the range 20–2.5 Å. The Z scores for the rotation and translation functions were 15.3 and 34.1. Data between 20 and 1.7 Å were used for initial rigid-body refinement, which gave an R value of 33.7%. A portion of the 2Fo-Fc (blue, 1.2 s) and Fo-Fc (green, +3.0 s) electron density maps made from the rigid-body refined model is shown below.
Phase Contrast Imaging
The requirements for differential phase contrast imaging (DPCI) are extremely well matched to the characteristics of the X-ray beam from the CLS because DPCI can use the full bandwidth (few percent), native CLS beam without any X-ray optics. The phase contrast mechanism leverages the high degree of spatial coherence from the small source spot, which translates to high contrast (i.e. deep fringes) and is comparable to the best synchrotron beamlines.
In 2007 the CLS performed its first 2D DPCI experiments which led to a publication in the January 2009 issue of The Journal of Synchrotron Radiation. Images from the experiment (below) were featured on the journal’s cover.
In 2010 the CLS performed an imaging experiment to create a 3-D tomographic image of the head of a mouse (below). Additional experiments have studied an array of soft tissue samples for medical research, ranging from characterizing eardrum structure to tumor detection.