Muon Interaction at ATLAS Hadron Calorimeter: Investigating Energy Dependence

[ChrisVer]

Can actually muons interact and give a signal while they go through the Hadron Calorimeter of the ATLAS? I think in general this isn't impossible, but it should depend on the muon energy...
I would like to see that energy dependence, even quantitatively.

 

[mfb]

They do interact, but probably not at a level the calorimeter is sensitive at. Their energy loss is basically given by the Bethe formula, at $\frac{2 MeV}{g cm^2}$. ATLAS has a sampling calorimeter and this energy loss is mainly to (relatively) low-energetic electrons - if they are produced in the iron absorbers, you can miss them.

I didn't check the total number of scintillating tiles in detail (design report here) - we have two meters of hadronic calorimeter, let's assume half of that distance has a material with 2g/cm^3, then we get a total energy loss of 400 MeV in the tiles. The value is a bit larger for particles crossing it at an angle, up to a factor of ~2. That is not enough to do anything useful with it and I'm not sure if it could be read out at all.
As comparison, typical calorimeter triggers are at tens of GeV.

The same should be true for CMS.

This is quite independent of the muon energy. At multi-TeV energies, Bremsstrahlung can get significant, similar to electrons, but those muons are extremely rare.

 

[ChrisVer]

to be more specific, my main reason is that working with ATLANTIS event display software, and checking muon events, I see several times the lines of muon in the inner and outer detector to allign with a trace/signal in the Hadron Calorimeter (check attachment).
So I thought this would be a reason...

There are only a few events which don't come with a signal in the HCAL. Unfortunately when I conducted the experiment I didn't notice this, so I couldn't check the deposited energy and see if it's actually measureable or 0...

 

[mfb]

Hmm... hard to tell based on event displays, and I have no idea about the scale ATLAS uses there. If they are frequent and of the order of 1 GeV, then it is probably the minimal ionization of the muon itself.
If it is significantly more for just a few events, then those muons could be produced as part of the hadronic shower in the calorimeter (pion decays).
If it is significantly more for most events, then something weird is going on. In this case: what is the reconstructed energy in the inner detector and the muon chambers? Are there multiple tracks in the inner detector?

Unfortunately when I conducted the experiment I didn't notice this, so I couldn't check the deposited energy and see if it's actually measureable or 0...

Did you delete your dataset?
Those event display images are very colorful, but in terms of physics they are not important.

 

[ChrisVer]

the energy within the first pic for example is 85.28 GeV in the inner detector and 53.92GeV in the outer (yes a difference of 31.36 GeV! lost inbetween the detectors- i think they are not well calibrated )
For the second it's less problematic 43.40 GeV inner detector... 43.83GeV outer (it seems to have gained energy but OK, it's up to the error in measuring it- so probably it lost almost nothing)...
The third is 48.89GeV in, 44.77GeV out (lost ~4GeV)

So I think it covers up all possibilities (I also got events where the muons out had gained 9GeV more energy). It was actually a mess a whole spectrum of nonsense results.
I don't know about the tracks of the inner detector, since in those cases I have only one neutrino it's coming out from W->mu nu channels...

I have no more access to the dataset...sorry...

 

[mfb]

the energy within the first pic for example is 85.28 GeV in the inner detector and 53.92GeV in the outer (yes a difference of 31.36 GeV! lost inbetween the detectors- i think they are not well calibrated )

That sounds too large for a calibration issue. I would expect a more physical reason here.

For the second it's less problematic 43.40 GeV inner detector... 43.83GeV outer (it seems to have gained energy but OK, it's up to the error in measuring it- so probably it lost almost nothing)...
The third is 48.89GeV in, 44.77GeV out (lost ~4GeV)

That looks good. 0.4 GeV is within the uncertainties and the energy loss is small in both cases.

(I also got events where the muons out had gained 9GeV more energy)

Very high-energetic muons? The uncertainty goes up fast for higher energies.

I don't know about the tracks of the inner detector, since in those cases I have only one neutrino it's coming out from W->mu nu channels...

You rarely just get a W decay without anything else.

 

[Vanadium 50]

First, if your first reaction to not understanding something is "the experiments did it wrong" (I am thinking about your statement "i think they are not well calibrated") you are not going to make many friends. This is really, really offensive.

Second, there's no way one can tell what is going on from just those two numbers. It could be almost anything - possibly as simple as comparing p with pT.

 

[ChrisVer]

It's not about my understanding, it's about the fact that the muon lost ~1/3 of its energy. However checking the background knowledge:
1) to ECAL, the muon don't lose a lot of energy, except for if they are very energetic (so Bremhstralung can occur) and the critical energy is then E~1000GeV... my muon had an energy of just 85GeV, not really close to 1000GeV.

2) to HCAL, the muons can lose energy because of ionization, Brehmstralung and collisions with the atomic electrons. The first 2 are true for high energetic muon, whereas the last is the dominant one. Can it really cause such a large energy loss? Checking the attachment I give, I'm in an extreme point. And I think this graph is representative of what I did because (after doing the calculations for 20 different muon events) I got a similar average energy loss ($5 \pm 2$GeV).

That's why I think it's an experiment thing, and I didn't really want to be insulting(?). However I also tend to believe that with just the $E_{in},p_{T}^{in},E_{out}, p_{T}^{out}$ you can't possibly explain what is going on.

 

[mfb]

It's not about my understanding, it's about the fact that the muon lost ~1/3 of its energy.

We don't know that.
Two different measurements of two different quantities gave two different values. To make sense of that, we first have to know what exactly has been measured.

The main difference between ECAL and HCAL is the size - both are sampling calorimeters with steel and some other materials as absorbers, and the forward HCAL even uses the same active material as most of the ECAL (liquid argon).

Which muon energy was used in the slide you attached?

That's why I think it's an experiment thing, and I didn't really want to be insulting(?). However I also tend to believe that with just the $E_{in},p_{T}^{in},E_{out}, p_{T}^{out}$ you can't possibly explain what is going on.

Probably. At least the result from the calorimeter(s) would be interesting.

 

[ChrisVer]

I don't understand what you mean by "what has been measured"... suppose you have a muon moving from the inside (track system) to the outside (muon chambers), you measure at those points its energies... if the muon didn't lose any energy then the one you measure inside will have to be equal to what you measure outside. If they are different, their difference will have to be the energy that was lost in your system (in ATLAS this is mainly the calorimeter material), no?
I think the energy used in that slide is the energy loss (since the slide refers to that) of the muons.

 

[mfb]

Reconstructing an event is not as easy as reading off values from a voltmeter.

The detector itself gives a collection of digital and analog signals, corresponding to hits and (for some subdetectors) energy deposition.
Then you have to look for tracks / showers in those signals. This process is not perfect - sometimes uncorrelated hits look like a track without an actual particle there, sometimes you miss so many hits of a particle that you do not reconstruct it as track, sometimes a real particle gets assigned wrong hits as part of the track, sometimes the hits are so dense that tracking does not give proper results at all. All those options will give wrong results.
You get a similar issue in calorimeters: the overlap of multiple showers is a very common effect.

Once you assigned those hits to tracks, you can fit for particle properties like the energy, the transverse momentum, the origin of the particle in the interaction region and so on. From those tracks, you can find the primary vertices in the event - points where multiple tracks go through.
Depending on the analysis, you might want to re-adjust the track fits now to let multiple tracks come from a common primary vertex or to let individual tracks come from a specific vertex. This is especially important for photons and neutral kaons, as you just see them in the calorimeters, without tracking in the inner part of the detector.

Tracking also has to take alignment into account, for example: the detector elements are never exactly at the design positions, because construction of a meter-sized tracker is not precise down to the micrometer-level. The deviations have been measured and they constantly get updated, together with other improvements - reconstructing tracks again later ("reprocessing", often done a year after the data has been taken) will change all those values a bit and make them more precise.

Calorimeters are very complex devices and need a lot of calibration to get realistic results. The ratio of energy deposition in the absorber to energy deposition in the active material both for hadronic and electromagnetic showers, the fraction of energy escaping as neutrinos or muons, the responses of individual active elements to scintillation light and corrections for dead elements (or infrastructure), just to name a few things.

This is just a very short part of a long list of effects that are considered in the reconstruction. Sometimes there is no "best" way to handle things and different analyses use different methods.I am sure you did not have to consider all those things. But someone else did. And your numbers and their correct interpretation depend on the way this was done.