Pointing the way

For decades, giant accelerators have used radio frequency (RF) waves to push elementary particles to near light speeds. The particles’ collisions, at increasingly stupendous energy levels, have created enough suspected but previously unidentified kinds of matter to fill all major blanks in the particle physics list known as the Standard Model.

But more is out there, theorists suspect, and it can be found – provided even more advanced accelerator technology can be built to find it. “Because particle physics is a statistical science, the more particles you can get, the greater the chances you have for making discoveries,” says James Amundson, who heads Fermi National Accelerator Laboratory’s scientific software infrastructure department in the Illinois lab’s Scientific Computing Division.

To increase the odds, the Department of Energy launched the Community Project for Accelerator Science and Simulation (ComPASS). The Fermilab-led multi-institutional program enlists advanced computer modeling to increase collider performance, allowing particle physicists to probe matter’s outer limits. ComPASS, a partnership of computational physicists, applied mathematicians and computer scientists, was formed in coordination with DOE’s Scientific Discovery through Advanced Computing program, or SciDAC.

DOE computers’ increasing capability is invigorating the program, Amundson says. “Computing technology has gotten so much better that we can do things now that 10 years ago were completely outside the realm of possibility for us.” Engineers still develop accelerator hardware prototypes, but “if we can do it virtually on a computer first, it makes sense, because it’s cheaper, faster and more flexible.”

Amundson is the primary architect of Synergia, a modeling tool created under ComPASS. Soon after its creation in 2002, researchers used Synergia to probe performance problems in the Booster, an intermediate part of Fermilab’s accelerator system that pushes protons to high beam intensities.

Many of those boosted particles fed what was then the world’s highest-energy collider, called the Tevatron because it delivered 1 trillion electron volts (TeV) to smash protons with antiprotons. Many other particles were converted into neutrinos for Fermilab’s MiniBooNE project, which studied how these practically weightless particles can morph, in particle physics vernacular, from one “flavor” to another.

The neutrinos would be beamed at out-of-state targets to study the particles’ physics.

Because Synergia models the interaction of particles with electromagnetic fields, it was tailor-made to simulate how proton accumulations affect each other and affect fields from external magnets and key device components called RF cavities.

The simulations showed how the positively charged particles could shove each other apart, leading to proton losses. That prompted Fermilab to install a collimator, a device to remove the lost particles before they become a problem, says Amundson.

The Tevatron discovered the top quark, a key Standard Model particle, but has since shut down. Attention has turned to the Swiss-French border facility CERN and its 13 TeV Large Hadron Collider (LHC) – now the world’s highest-energy collider – where physicists identified the coveted Higgs boson particle even before technicians completed a recent energy increase.

Synergia also will evaluate space charge at CERN as the LHC aims for higher luminosity: a proton injector upgrade will boost the number of particles generated for the same amount of energy. With more particles, for instance, “you can study the properties of the Higgs boson in greater detail,” says Amundson.

The simulation code also is helping Fermilab upgrade its accelerator complex into a new system that will generate unprecedented numbers of protons for new missions. Under the Proton Improvement Plan II (PIP-II), some of the proton swarm will create “the most intense neutrino beams anywhere,” Amundson says. In another thrust toward high luminosity, other PIP-II protons will be converted into an unprecedented few quintillion muons, short-lived subatomic particles.

The neutrinos would be beamed at out-of-state targets to study the particles’ physics. All those ghostly muons will go toward an attempt to document an unimaginably rare event: the direct conversion of just one of them into an electron.

ACE3P, another ComPASS simulation code, will evaluate superconducting radio frequency (SRF) cavities, devices that will accelerate the PIP-II protons. These prototypes resemble strings of hollow pearls and are made of carefully structured niobium, a superconducting metal. Some are bathed in ultracold liquid helium to quench heat buildup and improve efficiency. An electric field running through the cavities oscillates between positive and negative at 1.3 billion cycles per second, pushing the protons forward at the furious rate needed for experiments.

SRF cavities have fewer losses, making them more power efficient, Amundson says, but they’re “built with things we know how to do right now. They aren’t next generation.”

Another ComPASS thrust, based at other participating institutions, does focus on the next generation of smaller and cheaper accelerators. They would use new methods to move particles because conventional RF technology is approaching an outer limit where electrical breakdown begins. “I think it’s fair to say we operate close to the limit most of the time,” Amundson adds.

The problem is the acceleration gradient, or the energy charged particles transfer as they swoop through the devices. Gradients exceeding about 50 million to 100 million volts per meter create literal sparks in standard metal RF cavities, which operate in a vacuum. “You’ll get a spark and end up killing the system,” Amundson says.

One solution, driving particles through plasma with laser beams, could accommodate acceleration gradients exceeding 50 billion volts per meter, ComPASS researchers say. Another idea, using a laser to push particles through special materials known as dielectrics, such as photonic crystals, could allow gradients of 1 billion volts per meter.

Neither dielectric nor plasma accelerators would create sparks because they don’t operate in vacuums, Amundson says. They also could be more compact and cheaper than today’s giant machines.

ComPASS members Lawrence Berkeley National Laboratory and the University of California, Los Angeles are using the Warp, Osiris and Quick-PIC simulation codes to investigate the plasma approach. The SLAC National Accelerator Laboratory in California is using ACE3P to evaluate the dielectrics solution.

Other ComPASS members include Argonne National Laboratory, where Synergia has been battle tested on two Blue Gene supercomputers, the University of Texas, and Tech-X, a private corporation based in Boulder, Colorado.