A joint Fermilab/SLAC publication

ATLAS' countdown to collisions

First 900 GeV candidate collision events in the ATLAS detector, November 23, 2009.

First 900 GeV candidate collision events in the ATLAS detector, November 23, 2009.

ATLAS, at 148 feet long, 82 feet high, and 82 feet wide, has the largest volume of any particle detector ever built and is designed to study the products of high-energy proton-proton collisions. Over the past year, ATLAS has been upgrading, repairing, and preparing for LHC collisions, which began on Monday, November 23, 2009. In the four months preceding the first collisions, the ATLAS collaboration’s focus shifted from repair and installation work to getting all pieces of the complex ATLAS detector operating as one. Here’s a look at how ATLAS prepared for collisions.

First, a checkup—4 months until collisions
When the ATLAS detector is closed up and ready for beam, the innermost portions are inaccessible. Therefore, much attention is paid to infrastructure maintenance when the commissioning process begins. ATLAS’ Luca Fiorini explained, “The detector is built like a Russian doll. There are things that we cannot repair once the cavern is closed and the magnets are on, though there are still outer levels that we can work on.”

Several weeks over the summer were devoted to maintenance work on ATLAS’ 12 sub-detector units and their internal components. Infrastructure maintenance occurred on all three levels of ATLAS: the inner detector, calorimeters, and muon chambers, to be sure that no wires were crossed and all detector channels were calibrated to work together.

Turn on the magnets—2 months until collisions
The ATLAS detector uses superconducting magnets to bend the paths of particles as they spray out from collisions in the very heart of the detector. The central solenoid magnet has a peak magnetic field of 2.6 Tesla and the toroid magnets have a peak magnetic field of about 4 Tesla. Any material that could be affected by this magnetic field must be removed from the cavern where the detector is kept prior to turning on the magnets. This prevents stray material from damaging the detector. Once the cavern had been cleared, the magnets were turned on and the cosmic data run was underway.

Test drive: The cosmic run—8 weeks until collisions
Before collisions, particle detectors like ATLAS record data from cosmic muons, particles produced by proton collisions in the atmosphere. The 1,000 cosmic muons passing through ATLAS each second provide important information for calibrating the detector. Cosmic muon data allows experimenters to recognize the signature, timing, and location of cosmic events. Once beam is circulating, recognizing cosmic muons ensures that the detector is functioning consistently, enables further calibration, and allows experimenters to distinguish cosmic muons from the muons that are generated by particle collisions in the center of ATLAS.

“The cosmic muon data lets us understand the detector,” said Sotiris Vlachos, who works with ATLAS’ muon chambers. “This way we don’t lose beam time doing what we have to do, especially with alignment studies where we need to know where things are in respect to each other within a precision of micrometers.”

Monitoring ATLAS 24/7—6 weeks until collisions
At this point, ATLAS crews monitored the detector as though there were collisions in the LHC. Shifts of fifteen to twenty people were present in the control room for 24 hours a day, seven days a week, simulating the experience of beam in the detector and recording data from cosmic muons. Through these shifts, the ATLAS experimenters prepared themselves for the task of reading out the detectors and taking data during collisions.

Run coordinator Christophe Clement observed, “What we were simulating is tougher than when beam is actually in the machine. We were trying to push ourselves.”

Circulating beam in the LHC—4 days until collisions
By the time real beams were circulating in the LHC, detector programs had been recalibrated from detecting cosmic muons to colliding particles. ATLAS’ first non-cosmic events were beam splashes, recorded on November 20. These splashes occur when the LHC closes a collimator upstream of ATLAS to stop the low-energy, low-intensity beam and produce a spray of particles that the detector reads.

Trigger timing is everything—1 day until collisions
On the Sunday before the first collisions of protons at 900 GeV, a meeting was held to study the timing of the detector’s trigger system based on the weekend’s collected splash events. The trigger allows the detector to focus on the exact moment of collisions, much as a photographer programs a camera’s timer.

When the LHC is running at its design energy, up to 40 million collisions events will take place at the center of ATLAS every second. Due to space restraints, however, only about 200 per second will be recorded. The trigger system is designed to block 99.9995% of events so that only the most interesting collisions events are kept. To do this, the trigger’s timing must be exactly calibrated so that detectors are tuned to the nanosecond with respect to each other and to the time of the collisions. The process of configuring and confirming correct trigger timing is crucial to detector data-taking, and continues throughout the experiment to ensure that measurements are as accurate and precise as possible.

First collisions and moving forward
On November 23, ATLAS saw its first collision events in the LHC. These events were read by calorimeters, muon systems, and two of the three inner detectors: the transition radiation tracker, outermost of the inner detector layers and active for months to record cosmic muon data; and the semiconductor tracker, the inner detector’s middle layer, which was set to a standby voltage. At this point, the process of starting ATLAS was nearly complete.

Before the inner detector can be fully activated, the beam needs to be stable to ensure the safety of the innermost detectors which are closest to the beam line. Once the LHC operations team has completed all the measurements necessary to ensure a stable beam, the semiconductor tracker can be set to its most sensitive, high voltage. The innermost detector, the pixel detector, will not be switched on until beam has been successfully circulated, monitored, and aborted.

“The philosophy is the following: we must compromise between low quality data that we can get as soon as possible and high quality data that does not risk the detector,” explained pixel detector project leader Beniamino di Girolamo.

Once the three layers of the inner detector are on, the entirety of ATLAS will be up and running for high energy collisions with stable beams.

by Daisy Yuhas

Latest news articles

The LHCb collaboration announced the discovery of a new system of five particles all in a single analysis


Dianna Cowern—a.k.a. Physics Girl—has one of those invent-it-yourself jobs that exist only in the age of the internet.


The PICO bubble chambers use temperature and sound to tune into dark matter particles.


An animated take on dark matter, voiced by Janna Levin.