Massachusetts Institute of Technology
For a person devoted to physics and computers, it seemed a once-in-a-lifetime opportunity: doing research with the Large Hadron Collider (LHC), the world’s biggest particle-smashing machine.
“These big colliders, they open every 10 to 20 years,” says Dragos Velicanu, a Department of Energy Computational Science Graduate Fellowship recipient. He was seeking a graduate school just as the European nuclear science agency CERN was ramping up the LHC. Gunther Roland of the Massachusetts Institute of Technology was working with the giant experiment. Velicanu joined his group.
“I thought that this is a unique opportunity for me to study this stuff,” adds Velicanu, whose family moved from Romania to the United States in 1995. The bonus is spending several weeks each year gathering data at the LHC, located in a 17-mile circular tunnel deep beneath the Swiss-French border.
On his last trip, late in 2015, Velicanu went underground for his first look at the Compact Muon Solenoid (CMS), the detector that gathers data for his research. “It was much bigger than I expected it to be, even though you always hear that it’s as big as a cathedral.”
With its size, the CMS gets detailed pictures of the fallout when the LHC smashes lead ions into each other at nearly the speed of light. The collisions heat neutrons and protons in the ion’s nucleus to temperatures a million times as hot as the sun’s core. At that intensity, “you actually melt the protons and neutrons into a sort of liquid form of quarks and gluons,” the fundamental particles that comprise them, Velicanu says. “We melt matter to its most basic constituents and we study that form of matter.”
The resulting short-lived (about 10-23 seconds) quark-gluon plasma (QGP) exists only in these high-energy collisions and recreates conditions believed to have existed for a similarly brief (10-6 seconds) period after the Big Bang. Studying the QGP helps us understand the universe’s early stages. It also provides information about the strong force, which cements quarks and gluons into protons and neutrons and is one of the least understood of physics’ four fundamental forces.
Velicanu studies the particles that fly from the collisions and hit the CMS detectors, appearing as high-energy jets or electromagnetic deposits. He and his colleagues look for pairs of particles (quarks, gluons or others) that shoot in opposite directions and compare the energy each carries. That lets them calculate how much the QGP has slowed one of the pair and provides insights into the plasma’s properties.
That’s the physics part of the problem, Velicanu says. Computation first enters the picture as he and his fellow researchers sift the data billions of collisions generate – about a terabyte (one trillion bytes) per second.
That’s too much information to capture in computer memory as it’s gathered, so the first computational step applies algorithmic screens to record data from only the most interesting collisions. A hardware and algorithmic trigger chooses, in billionths of a second, only about 1 percent of the collisions based on thresholds the researchers set. From there, the data goes to thousands of computer servers, where higher-level algorithms examine them in depth, eventually discarding about 99 percent of the remaining collisions.
Even with these screens, Velicanu and his colleagues finish an LHC run with a petabyte (a quadrillion bytes) or more of data to analyze. Much of his research is devoted to developing algorithms to examine the information.
“Our current mathematical knowledge is at its limits” in systems like the QGP, Velicanu says. They’re strongly coupled, meaning “if you wiggle one of the pieces of the system it affects every other piece.”
The methods he and his group research could improve approaches to address this fundamental science problem.
Velicanu and Roland are part of the CMS Collaboration, an international group of more than 3,000 researchers. Most of the time Velicanu talks with them via video conferences, but when they’re collecting data at the LHC, “everybody from all over the world gathers in the same room and works together. That’s the part I really enjoy.”
The trips also give Velicanu a chance to explore some of the world’s most picturesque mountains, where he hikes and skis. On his last visit, Velicanu traveled to near the top of the storied Mont Blanc on the French-Italian border, but the snow was too thin for skiing.
Velicanu expects to graduate in summer 2016. After that, he hopes to find a postdoctoral research position or a job with a national laboratory or company.
Image caption: This shows a lead-on-lead LHC collision at a center of mass energy of 5 TeV (tera, or trillion, electron volts) per nucleon pair. This collision produced a large number of charged particles whose position and momenta are precisely tracked in the detector (green lines). Meanwhile, calorimeters measured the particles’ energy as shown by the blue (for neutral) and red (for charged) blocks. Two collimated high-energy deposits, called jets, are visible at the top and bottom of the detector and are the result of back-to-back high-energy quarks or gluons that shower into a spray of particles and energy. The fact that these energy deposits are not equal despite the initial particles starting with equal energy provides information about how the quark-gluon plasma slows down color particles via the strong nuclear force.