Update Upd te o on the n the dual dual phase phase light ght simulati si ation i on in l n larsoft arsoft J. Soto - 1 / 12 - 12 - | 09 09/10 10/18
Some activities ongoing ● Towards the TDR: – Working on generating a sample of events to work with. ● Improving the simulation: – S1 photon library with timing. – S2 photon library. J. Soto - - 2 / 12 12 - | 09 09/10 10/18 18
Goal of event samples Including with charge and light: ● 1. Radiological Bkg ✔ ● 2. Mono energetic SN nu (energy scan) ✔ ● 3. Mono energetic e- (energy scan) ● 4. 2 +1 ● 5. 3 + 1 ● 6. NDK (channel p->nubarK+) ● 7. 6 + 1 J. Soto - 3 / 12 - 12 - | 09 09/10 10/18 18
Radiological Background ● Thanks to Jason Stock! ● We adapted the parameters needed to the dual phase geometry. ● Background is generated randomly in a window of 2 times the drift time. ● It includes all argon inside the cryostat (liquid and gas). ● Charge and light is included. J. Soto - - 4 / 12 12 - | 09 09/10 10/18 18
Main parameters for the samples ● Main parameters: – Charge: ● Purity / tau = 3ms (0.9 less electrons in 12m drift) ● Gain = 20 ● Birks Recomb (for 0,5kV/cm) ~0.7 for a MIP ● We = 23.6eV (4e7 electrons/GeV) - dE/dx ~ 2.12 MeV/cm (tipically for a MIP) ● No noise included!!!. – Light ● Light yield of 24k photons/MeV. ● For EM saturations: fNumScintPhotons = VisibleEnergyDeposition * scintYield – http://www.apc.univ-paris7.fr/~franco/g4doxy/html/classG4EmSaturation.htm l#4c7405e2bb881dc85dc9c958aec0ad67 ● Gain of 1.e7, 1ms readout window, 250MHz sampling, 50kHz DC, 0.12QE including TPB. J. Soto - 5 / 12 - 12 - | 09 09/10 10/18 18
Benchmark Memory CPU time Output file 1 radiological evt Peak (s) size (Mb) (Mb) gen 1300 5 G4 (no S1prop) 6300 1500 detsim 6700 1700 reco 7500 150 440Mb ana 1800 1 ~200Mb Memory CPU time Output file 1 SN evt Peak (s) size (Mb) gen 1000 0.7 G4 (no S1prop) 1200 6 detsim 2800 1200 reco 1100 2 0.5M ana 1900 3 30k J. Soto - 6 / 12 - 12 - | 09 09/10 10/18 18
S1 extended photon library ● Using cryostat boundaries (we want to generate light all over the volume), also outside the TPC active volume. ● 15x15x62 voxels (1m3 voxels). ● 1M photons per voxel. ● Including timing (extended PL). ● Photon library already in pardata. J. Soto - - 7 / 12 12 - | 09 09/10 10/18 18
S1 extended photon library ● Timing is exported using the SimPhotonCounter module. ● Then a root macro is used to do the Landau fit to every pair voxel PMT. ● I only do the fit if I get more than 50 photons. ● Thus we only get a TF1 for 102249 pairs voxel- PMT (~10%). ● Then, to use the library with the timing we move from 1.5kMb of RAM to 6.5kMb! High but still affordable. ● Issues: – No timing interpolation between voxels, and they are too big. – Timing range is hard coded (fixed in the next release). – If we are interpolating visibility between voxels, and timing is active, code crash… (working in a solution). ● Still some things to improve… So timing won’t be included for now. J. Soto - 8 / 12 - 12 - | 09 09/10 10/18 18
S2 photon library ● Visibility in the top of the detector is very very low. ● Usual code doesn’t work, since we need much more statistics of photons, to have a reasonable value. ● To solve this I run several libraries in parallel, and then merged them. ● This was relatively fast: We need a lot of memory to simulate 3M photons, but most of them got killed very fast (as they are generated very close to the lems, i.e. solid copper). ● In this way, I generated a 18M photons/voxel photon library. ● 1x12x60m voxels of 1m2. All voxels! All voxels! ● Se still have too low statistics to do the timing fit! J. Soto - 9 / 12 - 12 - | 09 09/10 10/18 18
S2 photon library ● A possible/fast solution would be to parametrize the timing for all voxels and all PMTs and provide a general TF1 for all the S2: Landau fit does not work... ● 31k photons out of 18Mph/voxel * 720voxels: – Global S2 visibility of 2e-6 (2 photons/Mph) in all PMTs. J. Soto - 10 / 12 12 - | 09 09/10 10/18 18
Next steps ● Test the hybrid library: – To do this we need a high statistics and high voxel density photon library to work with: ● Generating now a 1Mph/25cm3 voxel PL. ● But it is taking too much time in the grid… jobs don’t get to run… too long and heavy… Exploring other alternatives now. J. Soto - 11 / 12 12 - | 09 09/10 10/18 18
Backup - PL Benchmark 1evt, 6GeV electron # photons # voxels Voxel size Mb CPU Time (s) Lib 300k (only tpc active) 300k 16x16x81 0.75m 3 1400 18 Lib 1M (all cryostat) 1M 15x15x62 1m 3 1400 16 Lib 1M with timing (all 1M 15x15x62 1m 3 6400 34 cryostat) J. Soto - 12 / 12 12 - | 09 09/10 10/18 18
Recommend
More recommend
