Ben Zastovnik's Work at CERN

CERN Globe and a "Superconducting Dipole Magnet" similar to those used in the LHC
(ATLAS Experiment is in the background on the right and the Jura Mountains are in the background on the left)

Updated 28 September 2009
During summer of 2008, I had the opportunity to work at the European Organization for Nuclear Research (French: Organisation Eurpéenne pour la Recherche Nucléaire), also known as CERN. My tasks were to attend the summer student lectures on various topics in high energy physics and to develop improvements to Virtual Point One, a event display software used by the ATLAS Experiment. Though the lectures were very interesting and informative, I feel they are pretty self-explanatory (you can even watch them here), so I am just going write about my contributions to the ATLAS Experiment.

A brief introduction to the setting: Located near Geneva, Switzerland on the Franco-Swiss border, CERN is the largest particle physics laboratory in the world, with physicists, engineers, and so many others from all around the world working there. It has made several important contributions to the area of particle physics since it was established in 1954, under the name of Conseil Européen pour la Recherche Nucléaire (CERN). We also have CERN thank for the World Wide Web which has become an important part of the culture today. During the time I was there, they were in the final stages of construction on the much talked about, Large Hadron Collider (LHC). The collider has a circumference of 27 km (17 miles) and lies 500 feet below the French and Swiss countryside. It accelerates two protons to 99.9999991% of the speed of light (~671 million MPH) and collides them together at four points along the ring. In these collisions, the kinetic energy is converted to mass (E=mc² remember?), and a shower of new, more massive particles are created. These new particles are highly unstable and quickly decay into other particles. Four massive detectors placed around the LHC are designed to detect this shower of new particles coming out of the collisions. The two biggest detectors, ATLAS and CMS are searching for the Higgs Boson, a particle theorized by Peter Higgs that would help explain the origin of mass in the universe. The other two detectors are ALICE and LHCb, which will study quark-gluon plasma and CP-violation (why there is more matter than anti-matter) respectively.

The huge detectors on the LHC need software to help scientists to make sense of all the data collected from particle collisions, which is where my job came in. Fresno State recently hired Dr. Yongsheng Gao who was able to sign Fresno State up as part of the ATLAS Collaboration and start a high energy physics group working on the ATLAS Experiment. I became part of this group in Fall of 2007 and attended his lectures on high energy physics. With most of the hardware completed on the LHC, most of the work that remained was developing software. Through Dr. Yongsheng Gao and his colleague, Dr. Charlie Young (affiliated with SLAC), I was introduced to Dr. Thomas Kittelmann (University of Pittsburgh) who I would be assisting with the development of Virtual Point One. Virtual Point One (VP1) is an event display for the ATLAS experiment that takes data recorded by the various sub-detectors of the ATLAS Detector and recreated the collision in a virtual environment. The name, VP1, come from the fact that ATLAS is at Point One on the LHC and the software is a virtual creation of the detector and events at Point One.

I had only one formal programming class before going to CERN, which was a Java class in high school. I had also picked up a little Perl and PHP over the years, so I had some programming experience, but not much. Also, the programming I would be doing would be in C++, which I had no experience with. However, Thomas said I should be alright since Java is object oriented and similar to C++. So I brushed up on some Java and took a look at C++ before I left. I also installed the ATLAS software package, Athena, which I would need. This downloaded an offline version of the software package used by the ATLAS collaboration, which including VP1. Most software at CERN is designed for use in Linux, including Athena. CERN uses a variation of Red Hat Linux called Scientific Linux CERN (SLC), however most of the software will run on other distributions of Linux. I ended up needing to reinstall this a few times after I got to CERN however. I first setup Athena in Ubuntu Linux but I found that it was easiest to setup a virtual machine in VMware running SLC-4. I found that certain libraries were required in Ubuntu in order to compile the VP1 code I would be working with. It took a couple days to get my computer all set up to compile and run VP1, but I was soon ready to start working on my first task.

The ATLAS detector: Before I start describing my project, allow me to explain the ATLAS Detector. The ATLAS detector is actually several layers of various sub-detectors that detect different particles and different attributes of the particles. The very first layer that a particle created in a collision comes to is the Inner Detector. The Inner Detector actually contains three sub-detectors, the Pixel Detector, Semi-Conductor Tracker (SCT), Transition radiation tracker (TRT), all of which tracks and determines the momentum of charged particles as the leave the center of the collision. The next layers are the Calorimeters, which absorb all the particles (except muons) and determine the energy of the particles. The outer layer is the Muon Spectrometer, which measures the momentum of the muons, which pass through the other layers of the detector.

The layer I worked on was the Pixel Detector. The pixel detector works similar to your digital camera, except in one "color" and in three-dimensions. There are 1744 modules are arranged on three layers of concentric cylinders and three layers on each end of the cylinder. Each module has a silicon wafers, with 46,080 pixels. When charged particle passes through a pixel it ionizes some of the silicon, which creates a signal that can be recorded. By mapping out the location of each pixel that a particle went through, we can trace the path the particle took to find the momentum and charge of the particle as it traveled through the magnetic field (see Lorentz Force and ATLAS Magnetic System).

When I started working on VP1, the virtual display of the Inner Detector was all the color blue. My task was to change the colors of the pixel detector to make the various components easier to see (Troels Jacobsen did the same with the SCT detector). To accomplish this task, I loaded the detector geometry (a 3D model of the detector) in VP1 and clicked each of the different types of components (i.e. chips, silicon, pigtail, etc.) to determine their name. With the list of names I was able to go into the code and set the attributes of each material, including the color. Using photos of the real Pixel Detector, I was able to match the colors of the Pixel Detector in VP1 to the colors of the real detector. This was a major improvement over the solid blue detetor.

Photo of half of the inner cylinder of real Pixel Detector

As I mentioned earlier, the new, more massive particles that are created from the collisions at the LHC are unstable and decay into other, less massive particles. The point at which these particles decay is called a vertex and the next task I was assigned was working with the vertices in VP1. When I started on this project, code had been written to display one large cross in the center where the collision took place. My job was to loop this through all the vertices in the event and create a cross for each vertex and be able to adjust the size of the cross according to various attributes. A lot of times scientists use simulated data to simulate what certain events will look like as well as other uses such as to debug software. These crosses show for simulated data, for real data I had to develop ellipses around vertices that are reconstructed from tracing particles through the detector. These ellipses show the spatial (three-dimensional) error of the reconstructed vertices, similar to error bars on graph. (This page is under construction, check back later for more)


See: CERN, Large Hadron Collider, ATLAS Experiment, and Virtual Point One

Home