The Minerva Masterclass

Introduction to Minerva

Minerva is a Masterclass tool allowing students to learn more about the physics that goes on in the ATLAS detector at the LHC. Currently, Minerva has two scenarios. One a tried and tested, well established exercise, where students must analyse events, using information from all parts of the detector to identify W and Z bosons. Another scenario which is currently under development is an exercise where students must find the decay products of neutral strange particles, which will eventually culminate in the calculation of the mass and lifetime of the strange particle.

W and Z scenario


The aim of the exercise is to identify particles in the ATLAS detector at the LHC. We ask the students to scan a mix of various physics events and classify them in the following categories

  • W->mu nu
  • W->e nu
  • Z->mu mu
  • Z->e e
  • QCD dijet events To do this we use the ATLAS event display called Atlantis. The setup was tailored to make it simpler and better looking for the students. Typically two students would work together on this exercise. The aim is to classify events in the different categories: W->enu, W->munu, Z->ee, Z->mumu and background events from jet production. To make it a bit more fun 1 Higgs->4l event is added to the sample. The group which finds this event will get a small prize e.g. an ATLAS 3D viewer. Each student will get 20 events to analyse. In case they are fast they are allowed to look as well at the other events to chase for the Higgs and obtain the prize. After the students have finished the results from each group are collected and we calculate together the ratio of W->enu/W->munu and the ratio of Z/W production and compare with the predictions.

This will teach the student some basics about how the detector works, e.g. what kind of 'traces' the different particles leave in the various detectors (e, mu, jet, missing transverse energy). They will need to learn how to select events via their signatures.

This exercise is based on the RAL Masterclass exercise given in 2008. This has been significantly extended and is now a joint venture between RAL and Birmingham.

Computer Set-up

MINERVA can be run either directly from the website (, or a standalone version can be downloaded below. MINERVA runs on Windows, Mac OSX and Linux, with a minimum requirement of Java version 1.6. In the MINERVA zip file is a version of Atlantis customised for this exercise with a special configuration file designed for students together with the events. If more groups of events are needed then they can be downloaded to each computer. To use these events, first extract the .zip file, then click File > Read Event Locally, then open the .zip file required for each group.

Real ATLAS data events:

  • Set of data events (~11MB)
  • Set of data events for up to 15 groups (~170MB)
  • Set of data events for up to 20 groups (~215MB) To run MINERVA, extract the file and double click on the 'atlantis.jar' file (Linux users consult the readme.txt file). MINERVA will then start, as long as you have a recent version of Java installed, version 1.6 or later. If you need Java installing please goto and download the software from the website. The default events are events which are shown in the introductory slides. To display the events of a given group, go to File (upper left corner of the right panel), then click on Read Events Locally, select the group you want to display and click Open.

Simulated events:


For the exercise, you will need some introductory slides, a results page per group, instructions for tutors etc. Here is what we need for the exercise
  • introductory ppt slides or in pdf format: This will teach the students how to identify the various categories of events in the detector and familisarise them with some Atlantis basics. In there is one small video on how the LHC works. You will have to copy it into the same directory as the powerpoint slides to play it within your presentation. If this fails you might have to include it yourself in a new slide (that happened already more than once to me)
  • result page per group: This page tells the students which set of events they have to analyse. Per event there is one box to tick per event category. The students have to tick (exactly) one box
  • overall result sheet: On this excel sheet the main tutor will collect the results from the various groups. Then he'll look together with the student at the ratio of W->enu/W->munu and the ratio of W/Z production which is approximately 10:1 (main effects: branching ratio, electro-weak couplings).
  • Atlantis instructions: summary of the main Atlantis command the student will need to do the exercise.
  • summary sheet of the various event signatures: in this sheet you find one event per category with some basic information how to identify them.
  • Overview sheet on how to identify electrons, muons and missing ET and which questions to ask yourself to identify the different type of signatures
  • cheat sheet for tutors: this sheet will tell the tutors which events correspond to each event type. To obtain this information, send a mail to Monika Wielers or Pete Watkins

Preparation for the exercise

At RAL there is a series of lectures to introduce the students to particle physics. This year more emphasis will be put on how a collider experiment work and on how to detect particles in the detector. Last years LHC lecture can be found here in pdf format. However, as the first set of students will do the exercise prior to this lecture, some basic details about collider physics and on how a detector work can be found as well in the introductory ppt slides, which will be shown before the students start the exercise. In there the basics on how to use Atlantis is covered as well.

In the tutorial room Atlantis is already launched and the first event a given group has to analyse is displayed. This saves a bit of time. The summary sheet and the additional sheets for the students is distributed.


Each group of two students gets 20 events. The first group is asked to go through event 1-20, the 2nd group to use event 21-40 etc. Each group uses another set of events. Next to each computer the students will find their result page, some Atlantis instructions, and a summary of the various signatures they have to identify. Now they have 30 min for the exercise. Then we ask them to stop and sum up the events per category. If they didn't finish analysing the 20 events, this doesn't matter. If they manage to finish scanning the events ahead of time, they can look at the other events and hunt for the Higgs.

The tutors go around and help the students using ATLANTIS and help them classifying the events. In our trials the most complicated part is the distinction between electrons and jets..


After the exercise each group is asked how many events they found per category. The main tutor fills in the results in the overall result sheet. Note, the lower part should be covered for this part. Then once all the results are collected, scroll down a bit and show the measured results for the ratio of W->enu/W->munu, Z->ee/Z->mumu (not much stat, so make people aware of this) and the ratio of W/Z production which is approx. 10:1 (main factors, couplings, branching ratio). Then at last you ask, if someone found the Higgs. If yes, show the event to everyone and give the prize, e.g. an Atlas 3d viewer.

Neutral Decay scenario


The aim of this experiment (led by Stockholm University) is to allow students to understand the concepts of momentum conservation and invariant mass by investigating strange particle decays.

A test version of Minerva for this scenario is available here.

More events for this scenario can be found here.

This includes 25 'ideal' K0 events and Λ0 events, as well as 10 sets of 25 events for each K0 and Λ0.

Computer Setup

Preparation of the event display (Data tab, InDet tab and Cuts tab might already be set at start-up)

Data tab

Click Data tab and under Status, then InDet and tick Track Collection, Space Point and RecVertex.

InDet tab

Make the collision point (primary vertex) and the decay points (secondary vertices) well visible by clicking on InDet tab, then RecVertex and set the Symbol size to 7. Tick Force symb to see vertex also on top display.

Cuts tab

Set pT to > 0.5 GeV (pT is the momentum in the direction at right angle to the colliding beams). Tick the box, and highlight the number. Enter the new number and do Return. This removes quite a few tracks and makes it easier to explore the event. Set z0 < 20 cm (z0 is the z coordinate of the primary vertex and should be rather close to 0) and d0Loose to < 4 cm (d0 is the distance the track misses the primary vertex) The d0 cut can be used during the exploration of the event. Normally require d0> 0.5 mm. This focuses the interest on the particles from secondary vertices. Note that this cut can sometimes also remove a track from the secondary vertex. By removing the cut, all the tracks from the primary vertex are also seen.

Exploring the Scenario

Use both display projections to inspect primary and secondary vertices close to the collision point, typically within 40 cm from the collision point. Use the zoom facility to focus on that part of the detector. Use the information from both views as some tracks and vertices are easier to observe in one of the views. Sometimes the vertices are somewhat displaced from the real vertex. The best is to complement the display with your own pattern recognition capacity
  • During the exploration of the event, it is useful to switch from including and excluding the d0 cut to see tracks that come from the primary and the secondary vertex respectively.
  • One of the main tasks is the determination of the mass of the particle. The mass of the particle is the same in all frames of reference %u2013 it is invariant. The invariant mass, is a characteristic of the total energy and momentum of a system of particles.
  • Also make an approximate estimate of the particle μs lifetime.

The invariant mass

The invariant mass is the mass of the decaying particle E2 = p2c2 + m2c4 where m is the invariant mass or just mass, E is the energy, p is the momentum and c is the speed of light. Energy and momentum must be conserved when the K0 particle decays into a π+ and a π- . Then : E = Eπ+ + Eπ- and p = pπ+ + pπ- remembering that the momentum p is a vector quantity! Then mK can be calculated: m2 = ( E2 - p2c2 )/ c4 Once you have identified a decaying particle, which very likely would be a K0 then: Click on 'pick' at the top of the GUI box of the software, then click on the two pion tracks one after the other The three components of the momentum will be displayed. Calculate the invariant mass of the original K0 particle in each case: mK = [ (Eπ+ + Eπ-)2 - (px π+ + px π-)2 - (py π+ + py π-)2 - (pz π+ + pz π-)2 ]1/2 An excel spread sheet could be designed to do this Explore the K0 events, and determine the mass from the momenta of the two pions. Repeat it for each K0 particle Make a histogram of the measured values, and determine the average mass of the K0.

The invariant mass of the Λ0 particle

The analysis of a potential lambda particle decay is a bit more challenging. For a decaying lambda, one of the charged particles tends to go in the same direction as the lambda particle. This particle will point back to the primary vertex, and will have a very small impact parameter at the primary vertex. In addition we do not know a priori whether the particle is a lambda or an antilambda particle. We have to test different hypotheses. The%uF020lambda particle will decay into a proton and a negative pion The antilambda particle will decay into an antiproton and a positive pion Determination of the invariant mass Click on %u201Cpick%u201D at the top of the GUI box of the software, then click on the two tracks one after the other The three components of the momentum will be displayed. Calculate the invariant mass of the original Λ0 particle in each case: mΛ = [ (Ep + Eπ-)2 - (px p + px π-)2 - (py p+ py π-)2 - (pz p+ pz π-)2 ]1/2 An excel spread sheet could be designed to do this. However, we also have to check the hypothesis that the positive particle is a pion and the negative particle is an antiproton as the decaying particle could be an antilambda.

The lifetime

Most particles are unstable. How long they live depends both on their lifetime and on their speed relative to the observer, that is us. The lifetime we observe is the particle lifetime at rest multiplied with the gamma factor (also called the Lorentz factor) The gamma factor, Γ = 1/(1-v2/c2)1/2 The gamma (or Lorentz) factor shows up frequently in special relativity The measured lifetime of a particle follows an exponential curve given by: N(t) = N0 e-t/τ Which describes the number of particles found at times t for a lifetime of τ. t is the measured lifetime divided with gamma, the Lorentz factor. Plotting the number of measured particles with a lifetime t on a logarithmic plot makes it rather easy to determine the particle lifetime. However, a complication is that the detection efficiency varies as a function of time, and that has to be corrected for. Without corrections, an average value of the measured lifetimes give an approximate value (order of magnitude) of the lifetime.

Recent Developments

Based on extensive feedback, the Masterclass has been improved to incorporate the feedback we received so far (thanks to everyone who provided feedback).
  • changes to the event display
    • to get an idea of the scale of the missing transverse energy vector the line thickness is now proportional to the magnitude.
    • muons can now be spotted more easily as they are now marked as a line in a similar way to the Inner Detector tracks
    • Atlantis can be downloaded directly via a web interface
    • several versions are provided depending on the number of groups needed (15, 20, 25)
  • changes in the W and Z exercise
    • each group now starts with a W or Z event, the events become more difficult towards the end of each batch of events. This should make it easier for the students to get started.


If you have any feedback concerning the exercise, please post it here. If you need more information, please send a mail to Monika Wielers or Pete Watkins or Tom McLaughlan.

Major updates:
-- ThomasMclaughlan - 04 Nov 2010
Topic revision: r11 - 19 Apr 2013 - 13:49:07 - ThomasMclaughlan
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback