Bump hunting

Every now and then physics experiments surprise us with something unexpected. Today, CMS, a competing experiment, announced such a result [1], a “bump” in their distributions. Bumps are very exciting in particle physics, because they are usually easy to identify, and they appear in unexpected places. A bump occurs where we expect to see a smooth, flat spectrum, but instead we see a peak that nobody predicted.

Roughly speaking, the detector is shaped like a cylinder, or a soda can. We describe the detector using two numbers. The first number, Φ. is the angle around the cylinder, and since the cylinder is symmetric around this angle, we expect to see no major bumps in any Φ distribution. The other number is η, called the pseudorapidity. It’s similar to the angle from the axis of the cylinder. Using these two numbers we can describe a region in the detector. If we take two regions in the detector, we can then define the difference in these two numbers, ΔΦ and Δη. These are the variables where the bump was seen!

It helps to have a nice picture in your head when thinking about these kinds of distributions, so if you have a cylinder to hand, such as a soda can, bottle or coffee mug, grab it and take a look at the shape. If not, you can take a look at the photos of some plastic cups that I used to help visualize the detector. You should see that if we have a very large value of ΔΦ between two regions, then these regions must be on the opposite side of the cylinder. If we have a very large value of Δη between two regions, then one must be near the top of the cylinder and one near the bottom. What CMS saw was two regions with a very small ΔΦ and a very large Δη. That means that they saw two regions that were on the same side of the cylinder, but one was near the top of the cylinder, and one was near the bottom. This is a very odd arrangement of regions.



Left: Definition of the Φ and η coordinates.
Center: Two regions separated by large ΔΦ and large Δη, which is what we expect.
Right: Two regions separated by small ΔΦ and large Δη, where the bump was seen.

In order to understand things better, we need to know some more details about how CMS reconstructed the events. The physicists looked for charged particles, known as tracks, and looked that the values of ΔΦ and Δη between pairs of tracks. Most of the tracks at the LHC come from jets, which are huge cones of tracks. The apex of the cone is at the center of the detector and cone points away from the center of the detector. This is easy to understand, as each track in the cone moves with a slightly different value of Φ and η, so they move away from each other as they travel along. This means that most of the track pairs come from these jets, and have very small values of ΔΦ and Δη. This huge peak was removed from the analysis, giving the physicists a chance to see the bump.


The plot showing the location of the bump. [1]

Taking a look at the plots presented by the analysts, it is obvious that there is definitely some effect there. What could be causing this? So far I have not seen any explanation of this bump in terms of new physics, or new particles. This seems a little odd, as bumps like this usually lead to detailed discussions over coffee, and people rush to the library to look for papers. This time, the results are being presented as they appear, and the auditorium is full. With no theoretical explanation, and an odd angular distribution, things seem a little odd to me.

History has taught us to be cautious of bumps like this. Some famous bumps include the J/ψ particle [2], which showed that the charm quark exists. That was a bump discovery that ended well. Unfortunately, another bump, which was thought to be the Υ particle, turned out to be fake [3]. Perhaps the most infamous bump was the pentaquark, which was seen in the mid 2000s, and later “confirmed” on several experiments [4]. As these experiments gathered more data, the bumps disappeared, and it seemed that several experiments managed to see the same bias in their reconstruction.

Competition in particle physics is fierce, so it is not unusual for resuts to be scrutinized. In fact, the results seen today have been known by CMS since ICHEP, but a whole series of additional tests have been performed since then. It is essential that physicists eliminate as many explanations as possible. Our detectors are arranged in a particular shape, and event rejection depends on the geometry of the detector. It’s very easy to unwittingly shape distributions with event selection, or to accidentally change the shape of a spectrum. The physicsts have checked and rechecked a huge number of effects, using everything they can think of. This does not mean that all effects have been taken into account, but it means that if there is a mundane effect which has generated the bump, it is a subtle one. It has escaped simulation and it has escaped reasonable skepticism for now. The next stage is to see what other experiments see. If ATLAS sees this effect, then there could be a physical process that we do not yet understand which is responsible for this effect. It is worth noting that this bump only appears for “high multiplicity” events, that is events with very large numbers of particles in the final state. This hints that this effect may be associated with a state of energy and mass known as the quark-gluon plasma, and several references have been made to heavy ion collider experiments which study this state. The bump could be more prominent on other experiments.

The most concerning and thought provoking feature for me is the value of ΔΦ. Suppose one particle is moving from the center of the detector, and then it decays into two others, which fly away from each other. If the momentum of each particle is drawn as a straight line, then they fit nicely on a plane. There seems to be no reason why this plane would prefer to line up so that they have similar values of Φ, which would give a small value of ΔΦ. Whatever process is responsible for this bump must explain this effect. It could be a very odd new physical process, or could simply be the result of the geometry of the detector.

Finally, the bump seems to be dependent on the transverse momentum of the tracks. In particular, the bump is most prominent when both tracks have momentum larger than 1 GeV-1 and less than 3 GeVc-1. A GeV-1 is a unit of momentum that can be compared to mass, and it is roughly equal to the mass of a proton. This means that a proton traveling with a momentum of 1 GeV-1 would share its energy equally between its rest energy (using E=mc2) and its kinetic energy. For the LHC this momentum is very small. The momentum of the incoming proton is roughly 1000 times larger than this, and it seems odd that the effect only applies to very low transverse momentum tracks.

These results will be discussed in much more detail over the coming weeks. It may be that the effect goes away, or that it is a hardware effect which means that other experiments will not see it. Either way, the experimentalists and theorists from all over the world will be discussing what has been seen. These kinds of discoveries are very exciting and they make working with physics research worthwhile. Based on past experience with other geometry based bumps (such as the pentaquarks), this will probably just be an artifact of the hardware and will not be seen elsewhere. Even so, there is not much which will fill out the main auditorium in a lab. It’s new, it’s exciting, and if it can’t be explained away it’s exactly the kind of discovery we all search for.

This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>