1 Simulations of MICE March 2005 BENE Week Rikard Sandström Geneva University

1 Outline What we do & how G4MICE – Features – Geant4 in MICE Example of G4 simulations – Detector response, reconstruction, analysis Example of Digitization Other tools – G4BeamLine – TURTLE – ICOOL – others Present status of G4MICE

1 What we do, how we do it Using software to simulate – Beamline – Particle physics – Detectors – Reconstruction of events – Analysis of (simulated) reconstructed events – Problems, such as RF background Misalignment Main path: Geant4 based simulations – G4MICE – 1.0 released a few weeks ago – Geant p02 Will move to Geant p01 soon

1 40 mm Detailed simulations of the full MICE cooling channel – RF cavities – absorbers – all detectors (TOF’s, CKOV’s, SciFi tracker, TPG tracker, EM-calorimeter) – iron shields, etc – beam contamination – RF background User specified number of cooling cells -> NuFact! All parameters can be given as input – Tweakability & user friendly – Turn on/off detectors, processes, RF field etc Digitization, reconstruction & analysis of most detectors Built in unit test for most subsystems Features of G4MICE

1 Geant4 in MICE Great care taken to confirm results with – Theory – PDG data – Independent simulations (ICOOL, EGSnrc)  Fair agreement, gets better with every G4 release Red = ICOOL Blue = G4MICE Energy loss in absorber (Chris) Energy deposition in EMCal fibers (Rikard)

1 Example: Energy deposition in EMCal Edep fiber/lead is constant for muons, and different for e and µ. Fibers sees electrons, photons convert in lead, fiber/lead dep decreases with z.

1 Example: Dark current simulations Scenario: – Electrons are ripped off RF cavities at high E field. – They are accelerated through the cavities, losing energy in Be windows. Can turn around if E-field flips. – Hit vacuum and absorber windows, and 35 cm liquid H2. Conversion to brems-strahlung photons. – Photons convert to electrons in trackers.

1 How I did it Measurements at Fermilab gave dark current per area. – Agrees with F ~ j x B – What area should we consider? – Assumed emission rate: Given the 40 kHz/cm2 of emitted electrons off a cavity at 8 MV/m, MHz frequency, and a conservative assumption of area to be considered, we have 7.92 e- hitting the absorber per energy peak (and period). 8 energy peaks * 2 linacs * 7.92 = e- per period in total. Tungsten coated windows lowers emission. Studies using 201 MHz cavity and atom probe tomography right now. Calculated propagation in Matlab, using energy loss from STAR. – More on next slide. Matlab results are started at cavity exits in G4MICE (Geant4). Windows are modelled as torispherical. – Starting time and energy as calculated in Matlab. Effect of background is digitized & reconstructed together with real muon, and analyzed. – SciFi trackers are OK with this rate of background.

1 RF phase calculations in Matlab The phases for the background electrons assumes phases optimized for a mu+ at 2 00 MeV/c on axis. – This causes a symmetry breaking in z! – Assumes phase difference between neighboring cavities is constant. – Phase diff = rad = ns. – This gives the muon an energy gain of 10.8 MeV per set of four RF-cavities, including energy loss in Be windows. – Energy loss is calculated using STAR data for ionization and bremsstrahlung.

1 Acce lerating e- in the RF, intro The electrons are assumed to have zero kinetic energy when emitted from beryllium windows. The electrons emitted at the peak values of the E- field only (+ and - respectively). They are accelerated using the same Matlab model as used for the muon. This results in 1. energy when leaving the RF-system 2. travel time for leaving the RF-system

1 Accelerating e- in the RF, 4 cavs Downstream direction Upstream direction

1 Detector response, recon & analysis Simulated hits in detectors are transformed to ADC counts etc (Digitization). – As close to real life as possible. – ex. use light yield and dead channel in tracker from test data – Reconstruction should see no formal difference between simulated signal and experimental signal. Reconstruction package reconstructs the events, given digits from detector response. The Analysis package calculates emittance along z (among other things). – User can add his own analysis. – Tools exist to convert all output files to ROOT files. – ”Analysis does everything for you, but perhaps not coffee” - Malcolm Ellis

1 Example: Digitization of TPG Objective: Find out whether 100 cm He or 18 cm Ne tracker is better. Simulation stores hundreds of points along a track. Digitization: – Poisson distributes clusters along the track. – Garfield determines how many drift electrons for a given cluster. – Driftelectrons drift to GEMs. – GEMs amplify signal, and gives further spread. – Readout strips picks up charge, converts into time dependent amplitude. – At given intervals, amplitudes are collected. – If there is more than one digit on a strip at a given sampling their amplitudes are summed up. – Noise is added. – The resulting digits are written to file, ready for Reconstruction. – Digits know about all their ancestors.

1 Other tools than G4MICE Acceptance, field matching, emittance is often calculated outside G4MICE – ICOOL (Ulisse Bravar) Less detailed than G4MICE, but well used, well known. Allows iterative optical matching. – TURTLE (Kevin Tilley) Used for optimizing beam line (rates & acceptance) – G4BeamLine (Tom Roberts) Useful tool here used in a similar manner as TURTLE. – Lahet, Mars, standalone Geant4 (Kenny Walaron) Target studies – FLUKA (Lara Howlett) Target studies – Matlab (Rikard Sandström) RF background transportation in time dependent fields – Garfield (Edda Gschwendtner) Low energy ionization in gas TOF0TOF1 Ckov1 Iron Shield TOF2 Ckov2 Cal ISIS Beam Diffuser Proton Absorber Iron Shield G4BeamLine Output from ICOOL, TURTLE and G4BeamLine can be used as input beams for G4MICE! – G4MICE can also reuse its own output as input beam.

1 Present status of G4MICE Beam line, cooling cell and detectors described. Detector response Some detectors have functional reconstruction. – Work in progress. Started PID – Expect results later (autumn?). Many questions of simulation physics solved. – bug fixes in Geant-4.6. – RF background will be measured this spring at FNAL. G4MICE is already used for production. Analysis tools -> listen to Chris.