Part 1: From Linacs to Lasers
From their humble beginnings as offshoots of the ordinary electric light bulb, particle accelerators have evolved in surprising directions. Among the most productive and promising developments have been light sources, first in the form of electron storage rings and increasingly as versatile and sophisticated free electron lasers.
Simulation of a laser wakefield accelerator: bunches of electrons (yellow and green) are injected and accelerated by surfing the plasma waves (blue surfaces) generated by a laser pulse. (Simulation by Cameron Geddes, LOASIS Program, at the National Energy Research Scientific Computing Center, NERSC)
Strange but true: the recently restarted Large Hadron Collider, the most powerful accelerator in the world, is the direct descendant of Thomas Edison’s light bulb. The light bulb was invented before electrons were discovered but nevertheless led to the first vacuum tubes, which for a long time were the principal means of accelerating and controlling charged particles in radios, medical x-rays, and other practical applications.
The late 1920s saw a burst of inventions intended for science. Cockroft and Walton’s voltage multiplier and Van de Graaff’s generator were powerful accelerators, but limited by the need to store an enormous electric charge. It was Rolf Widerøe’s design for a linear accelerator, or linac, enclosed in a segmented vacuum tube and designed to accelerate sodium and potassium ions (more massive and conveniently slower than electrons), that pointed Ernest Lawrence to a high-energy solution.
In effect, Lawrence wound Widerøe’s linear accelerator around and around itself to create the cyclotron, a machine that could boost a charged particle to very high energy by rhythmically switching the orientation of a relatively modest electric field. Almost overnight the word “accelerator” came to mean “atom smasher.”
Although vacuum-tube accelerators stayed around and continued to evolve for decades in radio, television, computer technology, and other fields, scientists who called themselves accelerator physicists pursued higher energies with mightier machines … until, one day, accelerators came to a fork in the road.
Enter the light source
When a charged particle is accelerated, jiggled in a light bulb’s hot filament or forced to turn a corner, for example, it loses energy in the form of light. This was a particular annoyance to designers of electron synchrotrons, who named the lost energy “synchrotron radiation.” (Protons and ions are so much more massive than electrons that the proportional loss of energy through synchrotron radiation, while significant, is less of an issue for the designers of these kinds of particle accelerators.)
It had already occurred to scientists that this light might be worth something for its own sake, but the first machine conceived and designed specifically for the production of synchrotron radiation didn’t come online until 1974. In a typical synchrotron light source, light comes from electrons moving through a simple bend magnet. The brightness is usually boosted by “insertion devices,” however — trains of dipole magnets called undulators or wigglers, whose poles alternate north and south to make the electrons slalom back and forth.
The resulting light – x-rays, ultraviolet, and a wide range of lower frequencies – continues along the straight path of the insertion devices and is taken off at a tangent to the storage ring through beamlines that end in experimental stations. Here the light can be used to resolve protein structures, investigate chemical reactions, and analyze the atomic-scale properties of materials, among many other investigations.
Light sources rapidly emerged as vital tools for scientific research. Within the light-source community, advanced accelerator methods have continually aimed at producing light that is brighter and comes in shorter pulses, and whose frequency, polarization, pulse length, repetition rate, and other variables are readily tunable.
Roger Falcone, director of Berkeley Lab’s Advanced Light Source (ALS) and the Lab’s associate director for photon sciences, remarks that “the ALS was built more than 15 years ago, but a few technical advances have brought our technology up to the level of the newest machines being built in Europe today.”
Progress has been so rapid, in fact, that synchrotron light sources are approaching the limits of what they can achieve in the way of short, bright pulses. For that reason, Falcone, John Corlett, and other creative scientists in the Lab’s ALS, Accelerator and Fusion Research (AFRD), and Engineering divisions are developing the technology for a new light source, of a kind known as a free electron x-ray laser.
Laser light from electrons
Like all lasers, a free electron laser (FEL) emits amplified coherent light, although the lasing medium is not a solid, liquid, or gas but rather the accelerated electrons themselves. In the technique called Self-Amplified Spontaneous Emission (SASE, pronounced “sassy”), emitted photons catch up and pass the electrons ahead of them in the beam, herding them into tight bunches. The undulator’s alternating magnets are spaced so that the photon wavelength and electron undulations are periodically matched in phase even though they are traveling at different speeds; thus each emission further amplifies the energy of the FEL light pulse, which can reach many orders of magnitude greater energy than that of a pulse from a synchrotron.
Another means of amplification is by “seeding” – shooting a separate laser beam through the electron beam to bunch the electrons, which then emit radiation at the seed laser frequency. Seeding produces light at a precise energy peak. If the bunches are sent through a second undulator, the emission can come from a “harmonic,” its frequency now two or three or more integer times the original seed laser frequency. Undulator pairs can be arranged in series so that each emits outgoing radiation at a harmonic of the incoming radiation, in a “seeded harmonic cascade.”
Dozens of beamlines issue from the electron storage ring of the Advanced Light Source. For over fifteen years the ALS has been in the forefront of synchrotron light sources providing soft x-rays for a uniquely wide range of scientific endeavors. |
A number of FEL-based light sources are under construction or already in operation around the world, such as the Linac Coherent Light Source (LCLS) at the SLAC National Accelerator Center, the world’s first x-ray laser, and the soft x-ray Free-Electron Laser in Hamburg (FLASH) at Germany’s German Electron Synchrotron (DESY); FLASH is a pilot program for the hard-x-ray European XFEL scheduled to begin operation in 2014.
Meanwhile scientists in Berkeley Lab’s AFRD, ALS, and Engineering divisions are designing a unique new kind of light source that would outstrip the performance of any other soft x-ray FEL machine now in the works. At the heart of the Next Generation Light Source (NGLS) would be a superconducting linac operating in continuous-wave mode, allowing the repetition rate of electron bunches to be up to a million times greater, and beam power a thousand times greater, than existing linac-based light sources.
“The advantage as opposed to a synchrotron light source is that when electrons radiate coherently in a laser beam, you get about four orders of magnitude more electromagnetic power per electron,” Falcone says. “The downside is that, in a linear device, you usually get just one beam. So in thinking about a new generation machine, we plan to multiplex it.”
Synchrotrons, being roughly circular, can accommodate a great many beamlines, but a linac’s one-way electron beam can be switched between several undulators, each a FEL tailored for different experimental needs. The NGLS would feed up to 10 FELs, each of which could supply x-rays to two or more beamlines.
A SASE FEL might be optimized for brilliance, intensity, and the ability to provide a very large flux of photons with a high repetition rate – a hundred thousand times a second or more – but with relatively long x-ray pulses, each a leisurely few femtoseconds to 100 femtoseconds (100 quadrillionths of a second), for example. (This is still a very fine slice of time; there are as many femtoseconds in one second as there are seconds in 32 billion years.)
Another FEL might have less intensity but be capable of pulses a thousand times shorter, measured in mere attoseconds (quintillionths of a second) – with the ability to provide a pair of these ultrashort x-ray pulses at different wavelengths (“colors”) and controlled separation between the pulses, an important capability for x-ray pump-probe experiments.
Timing the pulses
The NGLS contemplates producing millions of electron bunches per second in a continuous and uniform stream. (Today’s best x-ray FELs are limited to 10 to 100-plus bunches per second.) The capacity for very high repetition rates ultimately resides in the injector and electron gun at the front end of the linac.
“You can’t improve the fundamental beam quality after the beam has been generated,” says John Corlett, who heads the Center for Beam Physics in AFRD. “So the performance of the whole machine depends on the quality of the beam generated by the injector and source – the electron gun.” Here the NGLS would outperform any existing or planned machine.
In the gun, a small, circular photoemission cathode gives off shaped bunches of electrons when hit by a pulse from a dedicated laser; the laser pulse rate determines the pulse rate of the electron bunches. The electrons are then accelerated by a strong, low-frequency electric field before entering the injector and the superconducting linac.
A prototype of the electron gun is being built for testing at a dedicated facility at the ALS. One challenge is to choose the right material for the photocathode, one that produces a high ratio of electrons to photons in a tight beam and holds up under continuous use at very high repetition rates, in a high accelerating field.
Timing and synchronizing the many components of an NGLS, including its photocathode laser, is a problem facing not just light-source designers but other accelerator designers as well. Widely separated components – up to a kilometer or more apart – must be synchronized to within femtoseconds, an interval in which light travels only a third of a millionth of a meter.
Berkeley Lab scientists lead in designing systems that use light frequency, not speed, to provide a master clock. Using a reference laser beam and an interferometer, the researchers count the number of light waves in the master-clock beam, which travels the length of the machine through an optical fiber. Any change in the length of the fiber changes the number of interference fringes per unit length.
“So we monitor the number of fringes and add or subtract length to maintain the right number,” says AFRD’s John Staples, who led the development of the synchronization system, which is already in place at Stanford’s LCLS.
Not only machine components but experiments maintain synchronization by means of the master clock. Pump-probe experiments can send in a laser pulse to excite a sample and follow with a probe pulse to see what happens after some precisely timed interval, which may be as little as a few hundred attoseconds.
The NGLS and the technology now being developed to build it – from electron guns to accelerator design to synchronization systems to multiplexed FELs – will not only open new frontiers of science but are scientific advances in themselves.
Read the original Part 1 at Berkeley Lab's website
Part 2: Accelerating with Light
Accelerators are far from achieving the highest energies their builders aspire to, but size and cost may limit the kinds of facilities funding agencies can support. In the future, new kinds of machines will be needed to make further progress. Perhaps the most promising is the laser plasma accelerator.
Sophisticated as it is, a superconducting linac is a conventional particle accelerator that, in a machine like the Next Generation Light Source (NGLS) now under study, can be used to produce superbright laser beams. The inverse is also true: powerful lasers can be used to accelerate charged particles – but in ways that are anything but conventional.
An accelerator’s beam energy depends on the strength of its accelerating field (the gradient) and on its length. With gradients of 10 to 20 million volts per meter, the NGLS’s superconducting linac will continuously accelerate electrons to 2.5 billion electron volts (2.5 GeV) in a facility two-thirds of a kilometer long. That’s a short distance for a conventional accelerator, but when plasma accelerators come online sometime in the future, it may seem long indeed.
Laser plasma acceleration
Plasma accelerators were first proposed 30 years ago by the late John Dawson of the University of California at Los Angeles and Toshiki Tajima, now at the Ludwig Maximilian University of Munich. A plasma resembles an ordinary gas, one so hot that the atoms have fallen apart into their constituent particles – in the case of heated hydrogen gas, protons and electrons. To create a strong electric field it’s necessary to separate these positive and negative charges.
“Laser wakefield acceleration works on the principle that by sending a laser pulse through a gas to create a plasma — separating negatively charged electrons from positively charged ions in the gas — some of the free electrons will be carried along in the wake of the plasma wave created by the laser,” explains Wim Leemans, head of the LOASIS Program in Berkeley Lab’s Accelerator and Fusion Research Division. LOASIS stands for Laser Optics and Accelerator Systems Integrated Studies. “Imagine that the plasma is a lake and the laser pulse is a speedboat. The electrons are surfers riding the wave created by the boat’s wake.”
Theoretically, a laser wakefield accelerator could create an electric field of 100 billion volts per meter, 10,000 times as strong as conventional accelerators. Before the mid-1980s, however, lasers couldn’t make pulses short enough or powerful enough to achieve laser wakefield acceleration at all, because very high-powered lasers tended to blow themselves to pieces.
Then came a new laser technology called chirped-pulse amplification, invented by Gérard Mourou and Donna Strickland of the University of Rochester. The technique begins with a very short pulse from a very low-power oscillator. The short pulse is then “stretched” a thousand times or more with a pair of optical gratings that spread out the wavelengths, creating a “chirp” with low frequencies in the front of the pulse and high frequencies in back. The stretched-out chirp can be amplified a million times or more without damage to the laser. Once amplified, the pulse is unstretched with gratings or prisms and compacted back to its original short length and wide bandwidth – but now with ultrahigh power.
With chirped-pulse amplification laser wakefield acceleration advanced rapidly, although poor beam quality was a problem. Some electrons in a bunch achieved high energies but were tailed by many more electrons of low energy. “The energy spreads were so wide that they weren’t beams so much as sprays of electrons,” Leemans remarks.
But by 2004 three groups (one based in England, one in France, and the LOASIS team at Berkeley Lab) were closing in on the problem. They published their results simultaneously in Nature.
With a custom-made laser nicknamed T-REX, having a peak power of 40 trillion watts, the LOASIS team accelerated a high-quality beam of electrons to an energy of a billion electron volts in just 3.3 centimeters — a little over an inch. (Photo Roy Kaltschmidt) |
Two of the groups essentially used brute force, but the Berkeley Lab technique was unique, subtle, and promised a pathway to the future. To maintain beam focus, LOASIS created plasma channels that acted like optical fibers to concentrate and guide the laser pulse. An initial laser pulse drilled such a channel through a plume of hydrogen gas, while a second pulse heated the gas to plasma. The combination prepared the way for a short, 10-trillion-watt (10 TW) driver pulse from a separate, chirped-pulse laser. In this way LOASIS produced a high-quality, 80-million-electron-volt (80 MeV) electron beam within a distance of a few millimeters.
If in 2004 the entire world was consuming a dozen terawatts of power every second, how could a single laser in a small laboratory put out that much power or more? By packing an otherwise modest amount of energy into a much shorter slice of time: a mere half a joule of energy concentrated in 40 or 50 femtoseconds – the time it takes light to travel a third the width of a human hair – delivers 10 to 15 TW of peak power. The LOASIS researchers built the drive laser themselves and named it Godzilla. (They also built the drill and heater laser, named Chihuahua.)
By 2006 LOASIS was a full-fledged program in AFRD. The plasma channel had evolved to a physical channel in a block of sapphire, a technology pioneered by Simon Hooker and his group at Oxford University; an electrical discharge heated the gas inside to plasma. Using another “handmade” laser, this one with 40 TW peak power and named T-REX, the LOASIS crew achieved a high-quality beam of electrons with an energy of a billion electron volts (1 GeV) in a distance of just 3.3 centimeters, a little over an inch. A typical linear accelerator would require over two-thirds the length of a football field – 2,000 times as far – to reach the same energy.
Brighter beams and higher energies
Chirped-pulse amplification has so revolutionized the optics industry that today 10-TW lasers can be bought commercially for $1 million; 100-TW lasers are only about twice as much, an astonishing bargain by the standards of just a few years ago. But LOASIS is aiming higher.
Working with the private sector on a petawatt (quadrillion watt) laser system, one capable of delivering 40 joules of energy once every second in 40-femtosecond pulses, LOASIS is building the Berkeley Lab Laser Accelerator (BELLA) to accelerate a high-quality beam of electrons to 10 billion electron volts (10 GeV) in less than a meter. The U.S. Department of Energy has signaled its commitment to this quest for the accelerator of the future by granting BELLA a spate of simultaneous approvals.
Although it sounds like alphabet soup, DOE’s Critical Decision scheme is straightforward: once it has been agreed that there’s actually a need for a new thing, which DOE calls Critical Decision 0, CD-0, next comes the conceptual stage, CD-1, when a specific approach and cost range are determined. Then CD-2 looks at technical specifications to judge whether the project can be completed on time and on budget, CD-3 approves the start of construction, and finally CD-4 flips the “on” switch.
A high-priority project may move ahead on several fronts at once, developing detailed plans for some aspects while alternatives are still being evaluated for others. Meanwhile, construction can start on some essential components. That’s the case with BELLA, which received simultaneous CD-1, CD-2A, and CD-3A approvals in September of 2009.
“CD-2A/3A gives the approval to start the long process of procuring and constructing the laser systems, whose cost is well established,” says Sergio Zimmermann of the Engineering Division, BELLA’s project manager, “while CD-1 gives the BELLA team the go-ahead to design the remaining systems beyond the conceptual stage.”
BELLA’s chief aim is to accelerate electrons to 10 GeV in a one-meter module, but at the same time LOASIS is continuing research and development on equally important questions with its current lasers. There’s a limit to the distance over which a laser beam can retain critical focus, so a laser plasma accelerator, like a powerful conventional accelerator, will need many accelerating modules, each boosting the energy of the beam as it passes from one to the next.
LOASIS scientists are using the existing 3.3-centimeter modules to investigate injection and staging schemes for carrying well-formed bunches of electrons from one module to the next, while introducing the laser pulses needed to further focus and accelerate the bunches in close synchronization.
Achieving still higher energies with laser plasma accelerators will require staging and synchronization systems that can pass the beam from one accelerating module to the next, as in this scheme for an electron-positron laser plasma collider.
Sometimes called “tabletop” accelerators, a moniker that conveniently ignores the extensive laser systems needed to drive the electrons inside the accelerating module, laser plasma accelerators will nevertheless revolutionize high-energy physics. Once the challenges of staging and injection are mastered, hundreds of accelerating modules can be linked to produce electron colliders reaching center of mass energies in the trillions of electron volts (TeVs) – not in the scores of kilometers needed for conventional machines but in half a kilometer or less.
Other challenges remain. At one pulse per second, BELLA’s petawatt laser will have a high repetition rate for a big laser, but that’s a million times slower than what the NGLS electron injector aims for. Nevertheless, BELLA will be capable of unique science in its own right. The BELLA laser accelerator may form the core of a user facility unlike any in the world for research in physics, chemistry, biology, and materials.
For example, by colliding a 10-GeV electron beam with a separate petawatt laser pulse, extraordinarily high electric fields could be created in which accelerated electrons could gain energy equivalent to their own rest mass – enough to create electron-positron pairs out of the vacuum. It’s a phenomenon known as “snapping” or “boiling” the vacuum, a regime more familiar to astrophysicists than laser scientists.
Intense electron beams up to 10 GeV, as short as a femtosecond in duration, plus petawatt-peak-power laser beams; it’s a combination that will drive accelerator science in whole new directions – inside the laboratory.
Read the original Part 2 at Berkeley Lab's website