High-energy lasers could change particle-accelerator design by producing relativistic electrons over a much shorter distance than existing machines.
In experiments that test the most fundamental theories of physics, subatomic particles are accelerated to near light speeds in machines many kilometers in length. Work now underway using tabletop-terawatt (T3) laser systems promises to reduce the size (and perhaps the cost) of these particle accelerators by a hundredfold.
For years physicists have recognized the possibility of using the enormous electric-field gradients that can exist in a plasma as a less-expensive means of producing high-energy particles. At Lawrence Berkeley National Laboratory (LBNL; Berkeley, CA), experiments to test the practicality of this concept use T3 systems that cost less than some homes.
In the early 1960s, the scattering of laser light from relativistic electrons was used to produce low levels of x-rays and even gamma rays. The power of lasers at that time, together with the minute size of the Thomson cross section (<10-24 cm2) that governs the scattering of low-energy photons from free electrons, combined to limit the generation of high-energy photons. The development of T3 laser systems has rekindled an interest in colliding visible light with relativistic electrons for possible applications in material and life sciences, for creating particles directly from high-energy photons, and for revolutionizing the design of particle accelerators.
Oscillating regions of negative charge in a plasma-a plasma wave-can generate an intense longitudinal electric field. An electron injected into the plasma with a velocity that matches the phase velocity of this plasma wave will stay in phase with the wave and its electric field and absorb energy. In practice, the injected electrons move at essentially the speed of light, c, and the plasma is formed so that its plasma frequency also equals c. The acceleration that the injected electrons gain from the plasma wave goes into increasing their relativistic mass-their velocity is virtually unchanged. This allows the electrons to stay in phase with the plasma wave and to continue to absorb energy.
One method of producing the plasma wave is to propagate a short intense laser pulse through the plasma. The exchange of momentum between the pulse and the plasma leaves behind a wake much like the wake that follows a boat through water. Using a laser pulse to produce the accelerating field is referred to as laser wakefield acceleration, or LWFA. The optimum duration of the laser pulse to generate a wakefield for a typical plasma is on the order of 100 fs.
Just as the peak of an ocean wave breaks and falls into the neighboring trough when the wave exceeds a maximum amplitude, "wave breaking" limits the amplitude of the plasma oscillations and sets the maximum value of the longitudinal electric field. For a phase velocity of c, corresponding values of the electric field maximum vary from 10 GV/m to greater than 1 TV/m.
The Next Linear Collider (NLC), a proposed 30-km facility under study, is designed to operate at 50 MeV/m. To achieve this goal using 11-GHz accelerating fields, the NLC would use thousands of 1- to 2-m copper cavities with diamond-polished surfaces and would be aligned with micrometer-level precision over the length of the accelerator. By comparison, a LWFA machine some day may achieve the same particle energies in 100 m or less.
Workhorse at l`OASIS
Generating the wakefield is only part of the task of a T3 system. A synchronized multipicosecond-pulse output produces a channel in the plasma that acts as a waveguide to increase the acceleration distance. Almost as critical is the need for synchronized pulses for channel diagnosis and probing the wakefield. Finally, the laser is used to trigger the injection of the electron bunch to be accelerated. All of these pulses need to be generated at 10 Hz or more to achieve a respectable output.
The l`OASIS facility of Wim Leemens at LBNL is dedicated to LWFA and related studies. The laser oscillator is a modelocked Ti:sapphire system that produces 0.25 W of average power (see Fig. 1). The oscillator pulses are first stretched up to as much as 300 ps (the exact duration is controlled by an intracavity slit in the oscillator). The stretched pulses are fed into a regenerative amplifier that is pumped by a 1-kHz intracavity-doubled Nd:YLF laser. About 5% of the output is sent to the compressor and frequency-doubled to 400 nm for use as the diagnostic beam.
Most of the output of the regenerative amplifier, 1.2 to 1.5 mJ, is sent to a Ti:sapphire preamplifier pumped by a Q-switched frequency-doubled Nd:YAG at a 10-Hz repetition rate. The 50-mJ pulses are then boosted in a 2 + 2-pass Ti:sapphire main amplifier, pumped from both sides by 7- to 8-ns pulses from another frequency-doubled Nd:YAG. Output pulses of about 400 mJ have been obtained from 1.8-J pump beams.
About 60% of the pulse energy is compressed to 100- to 120-mJ pulses lasting 70 and 75 fs. These pulses are used to excite the wakefield. The remaining pulse energy is double-passed through the main amplifier again to produce 400- to 500-mJ pulses used to produce the plasma channel waveguide. Optical delay lines ensure the necessary synchronization between the terawatt pulse (wakefield generation), the long pulse (waveguide generation), and the blue pulse (diagnostics).
In recent years, experimenters at LBNL, the Naval Research Laboratory, UCLA, the University of Michigan, the University of Maryland, and elsewhere have produced accelerating fields using LWFA of between 10 and 100 GV/m over a few millimeters. Plasma channels have been used to propagate moderately intense pulses over 70 Rayleigh lengths (more than 2 cm). While clearly still in its infancy, the field has shown enough promise so that planning for a new generation of experiments has begun.
The T3 systems would be involved more directly in the investigation of the universe in the proposed gamma-gamma collider. Physicists hope to search for clues to the existence of the Higgs boson in the products of colliding gamma photons, which would shed light on the origin of mass. It is also possible that the photon itself might have structure that could be revealed by colliding gamma rays, and there is even speculation that the mystery of our perception of the arrow of time might be related to a field arising from these photon components.
The gamma-gamma collider would use Compton scattering of light produced by T3 systems off the beams in a high-energy accelerator to produce gamma photons (see Fig. 2). To compare favorably with the production of exotic particles by accelerators, a gamma-gamma collider will need to provide a gamma photon from each electron in the accelerator. The physics of Compton scattering therefore dictates that a pulse of about 1019 photons should be incident on a typical electron bunch in the accelerator. There also exists a constraint on the photon wavelength to avoid unwanted types of particle production that would reduce the gamma-gamma interaction. This optimum wavelength is about 1 ?m, so taken together with other considerations, the resulting requirement is a 1-J pulse at 1-?m wavelength that spans a few picoseconds.
Further requirements for the laser system are more daunting. Total particle-production requirements demand that the average power of the µm source be between 10 and 20 kW, far beyond existing T3 systems. The laser polarization must be variable. In addition, sensitive hardware and varieties of high-energy radiation would be crowded together in the target area. There are proposals to chain together banks of T3 systems or recirculate the laser pulses and reduce the need for high average power.
There may be more immediate and more practical applications for this technology. The 100-fs time scale has particular importance in materials studies and in the life sciences. For example, the time-resolved study of the phase transformation between the solid and the molten states of a silicon surface has special interest for the semiconductor industry. T3 lasers have been used to study such phenomena indirectly, but visible light cannot measure the fundamental motion of the atomic core.
In the life sciences, the dynamics of protein folding have great importance. A biological protein molecule, when fully uncoiled, may reach a meter or more in length. The protein goes from its stretched condition to its characteristic coiled state in about 1 µs, but along the way assumes a number of intermediate configurations that affect the interaction of the protein with membranes and other microscopic structures. Thousands of different types of proteins evidently take on just a few distinct intermediate coiled states, but little is currently known about these transformations.
An ultrashort x-ray pulse could match the temporal and spatial scale of these phenomena. T3 lasers can produce x-rays by scattering from solids and by harmonic generation in gases, and advanced synchrotron facilities produce high-brightness, tunable x-rays. However the pulsed output from these sources is either too long in duration or in wavelength, or the average power is insufficient for the applications mentioned above. Researchers at the Beam Test Facility (BTF) at LBNL, using a novel 90° Thomson-scattering geometry for producing T3 pulses off a low-energy but still relativistic electron beam, have generated the shortest pulsed hard x-ray pulses to date.
This scattering geometry allows production of femtosecond x-ray pulses (see Fig. 3). The convolution of the transit time of the light through the electron bunch with the laser-pulse duration determines the pulse length of the scattered x-rays. This transit time can be made to match the femtosecond laser pulse by focusing the electron beam to a narrow waist in the interactive region.
Using 100-fs, 0.5-TW Ti:sapphire pulses scattered at 90° from a 50-MeV electron beam, the BTF has generated
x-rays as short as 0.4 Å in pulses less than 300 fs long. While the conversion efficiency currently is low, the flux is already sufficient to provide diffraction peaks from a silicon surface, and prospects are excellent for increasing the conversion efficiency.
Neither the LWFA nor the gamma-gamma collider will be producing new results in high-energy physics for decades. However the dramatic promise they hold and the convenience of performing prototype feasibility studies using T3 laser systems insure that this will remain an area of active research into the next century.
Says Swapan Chattopadhyay, head of the Center for Beam Physics at LBNL, "All accelerators for relativistic particles today demand an unrealistically large footprint and enormous finances. It is time that we start exploring the emerging technology of light waves interacting with electron beams, which holds considerable promise both for compact electron accelerators and for radiation sources of unusual characteristics for fundamental scientific investigation."