A T&M View of the Higgs Boson Discovery: Pt 1, Herding cats on the Franco-Swiss border
The test conditions are provided by the Large Hadron Collider (LHC) at CERN (the European Center for Particle Physics Research – the acronym stands for Conseil Européen pour la Recherche Nucléaire but the lab's purpose evolved from nuclear to particle physics in the last few decades and it boasts perhaps the finest cafeteria dining on earth, due to the prevalence of French chefs, along with a nice choice of wines and a few beers).
The LHC is a 27 km long pipe laid out in a circle beneath the Jura Alps (see map below). Beams of protons circulate in opposite directions through the beam pipe. Protons, of course, are positively charged hydrogen nuclei. The protons travel in bunches that are accelerated to 4 TeV so that when the beams collide at the center of the two detectors the total collision energy of any pair of interacting protons is 8 TeV; 8 TeV is 1.28 micro-Joules or the amount of energy acquired by an electron in traversing eight trillion volts. The detectors are cylindrical objects oriented so that the beam pipe enters the center of one end cap and exits the other.
Here is a short list of the biggest measurement challenges:
- Protons are complicated objects
Protons are 1 fm diameter bags of point-like objects called quarks and gluons most of which carry electric charge. Of course the sum of the charges within the bag must add up to 1 positive electric charge, but the occupants of the bag constantly varies.
- The signal-to-noise ratio of these experiments is ridiculous
Bunches of protons cross the center of the detectors at 40 MHz. When they cross, chances are that twenty protons from each beam actually notice each other. Now picture two bags of marbles colliding. Since the bags are mostly empty space, chances are that just one marble from each bag actually hit one another. When this happens, it's called an event. But no one knows exactly which marbles are colliding.
- The Standard Model of Particle Physics (a.k.a., the theory) predicts that just a few out of every 1013 collisions produces the particular Higgs signal they're after.
It has been said that this sort of experiment is like throwing fine mechanical watches at each other and then trying to figure out how they work by the debris remaining after the collision. But it's more like throwing a 800 million watches every second.
- Event rate + data size = major data acquisition headache
The data acquisition systems include thousands of miles of cable and millions of electrical channels – all of which require enough analysis at that 40 MHz to reduce the amount of data for storage on disk to a reasonable level. Three levels of " triggers," including a large parallel system of ASIC and FPGA-based processors along with over 2000 PCs are used to reject more common/less interesting events and reduce the data to be storage rate to a reasonable 2.5 Gb/s.
- Each of the experiments has over 2000 people working on them and there's no "boss."
The vast majority of these people would like to spend their time analyzing data. But the vast majority of the work is in designing, building, and maintaining the experimental hardware, software, and supporting computing.
Now, it might sound like the experiments, called ATLAS and CMS, would be right-sized companies to work for, and they're great for sole contributors. But the leaders of the experiments only directly employ a dozen or so of the people who work on them. In other words, think of it as directing 2000 employees you can neither hire, fire, nor promote/demote. Leadership is by consensus not compulsion and the result is that many of the "Spokespeople" (c.f., CEOs) are gifted leaders.
Continue to Part 2, A T&M View of the Higgs Boson Discovery Pt. 2, What they detect.
(Ransom, the author, worked on one of the experiments from 1994 to 1999 and would like his appreciation for the subsidized gourmet food and beer duly noted)