SCT Activities Nick Bedford, Mateusz Dyndal, Alexander Madsen, Edoardo Rossi, Christian Sander DESY ATLAS Weekly Meeting 03. Jun. 2016 1
Semi-Conductor Tracker Barrel 4 Layers 2112 identical modules Endcaps 2 end-caps, 9 disks each 1976 modules C. Sander DESY Weekly Meeting - 03 Jun 2016 2
Modules Two sides, back-to-back, 40 mrad 285 μm thick, high-resistivity p-strips on n-bulk [-5,-10] ºC 17 μm resolution in R-φ plane End-cap modules Barrel modules 3 rings of different modules 4 sensors/module (2 e/side) Outer and middle: 4 sensors/module Identical Inner: 2 sensors/module 768 strips/side 57-94 μm pitch 80 μm pitch CIS and Hamamatsu Hamamatsu C. Sander DESY Weekly Meeting - 03 Jun 2016 3
Activities - Overview DESY joined SCT operation (monitoring+calibration) effort in 2011 as institutional responsibility What are we doing? Keep 24h calibration loop working (Mateusz, Christian) Maintenance and development of monitoring code (Mateusz, Christian) Long-term monitoring of SCT noisy strips (Christian) Performance studies (Nick B., Edoardo) DCS on-call expert (Alex, starting this fall) Further activities (not covered here) SONAR (Cecile, Abby, Alex) Heater pads (Alex) MC Tuning (Akanksha, Mahsana, and Thorsten from Zeuthen) Regular meetings together with Zeuthen on Thursdays 1:30pm 4
Prompt Calibration Loop t = 0 Data taking (cosmics or collisions) Online partial reconstruction; special calibration streams 5-10 Hz; e.g. data from beam gaps to study noisy strips Prompt calibration loop; offline monitoring Dead strips Efficiency Noise occupancy Lorentz Angle Bytestream errors Data Quality Monitoring Noisy strips: <occupancy> > 1.5% masked Recent activities: Update cron job scheduling to avoid upload problems Fixed some bugs (e.g. web mail form, mail bombs ) Manual re-running of wrongly released jobs (hopefully not needed in the future) Uploaded to conditions data base (COOL) t = 24-48h Bulk reconstruction at TIER0 5
Monitoring of Noisy Strips Total number of noisy strips per run shows no alarming increase in 2016 Large fluctuations from run to run, resulting in warnings (so far, not critical) Number of noisy strips 4 10 3 10 2 10 10000 number of defects 9000 8000 7000 12/2009 12/2010 01/2012 12/2012 12/2013 12/2014 01/2016 12/2016 Date 14 12 10 6000 5000 4000 3000 2000 1000 0 0 10000 20000 30000 40000 50000 60000 70000 80000 duration (sec) 8 6 4 2 0 Hint of trend: longer runs more noisy strips For this purpose: added LB dependent HitMaps to prompt calibration loop 6
#NS/module vs. time (in LB) Run 299584 (so far longest run in 2016) Plot shows distribution of #noisy strips per module (x axis) for barrel (left) and endcap C (right) in bins of 20 luminosity blocks (y axis) Red: <number of noisy strips per module> 50 For end-cap C, large number of modules show strange time dependence 100 Number of noisy strips vs. Lumi block NNS2D_B Entries 70180 Mean x 2.198 Mean y 65.92 100 RMS x 11.66 RMS y 31.27 100 Number of noisy strips vs. Lumi block NNS2D_EC Entries 42340 Mean x 14.16 70 Mean y 66.94 RMS x 58.51 RMS y 60 29.6 80 80 80 50 60 60 60 40 40 40 40 30 20 20 20 20 10 100 200 300 400 500 600 700 800 900 1000 0 100 200 300 400 500 600 700 800 900 1000 0 7
Single Modules EC-C (Disk 2, 3 & 4) chips 0 1 2 3 4 5 Plots show occupancy for single modules as function of time (in units of 20 LB) Striking time dependent feature, not constrained to single chips Pattern shows some correlation of affected strips for each disk Patterns seem to be in sync for each disk Often, the module is ok for a significant fraction of time but will be marked as noisy for most strips 8
Properties of Suspicious Modules In end-cap C all affected modules ( 20) are from CIS are located in middle ring (i_eta = 1) are facing air gap (link-0) from disk 7 show some pattern in φ (only odd modules show feature, those closer to support structure) from other disks show no pattern in φ 9
Increased Noise for HPK Modules A noise anomaly has been detected in end-caps: All modules affected are Hamamatsu link-0 (chip 0-5) modules link-1at module boundaries (6 & 11) shown even/odd differences study by Taka Kondo and Mutsuto Hagihara 10
Facing the Airgap 11
Ion Wind? Hypothesis: radiation ionises air between disks, creating ion-wind increased noise Plan: use ion gun to monitor noise levels on test module (in future @DESY) First step: estimate ion flux on modules, see if gun can provide it For given module, Nick integrated number of expected charged particles over volume of air gap next to module, and calculate expected number of produced ions from average path length This assumes, that all ions do reach the modules! Is this true? How about dissipation? Idea: solve diffusion equation (1D, no Bx and By); no drift c: concentration of ions ions = A 2 f LHC dn± d q: source term (produced ions from charged particles passing air gap) α: radiative recombination coefficient dn ions L hli 10... 80 na @ c(~x, t) =D c(~x, t)+q(~x) c(~x, t)2 @t 12
Solution with MATHEMATICA without dissipation with dissipation c [cm -3 ] c [cm -3 ] x [cm] time [s] x [cm] time [s] Nick s preliminary results show that dissipation is important! Consequence: flux reaching module surface is much reduced w.r.t. naive estimate Current challenge: calculate the flux on modules Simple estimate (x, y, z = L) = 1 2 Z L 0 q(~x)dz Z L 0! c(~x) 2 dz does not work! (limited numerical accuracy) 13
Summary Several SCT activities ongoing, e.g. maintenance of prompt calibration loop New time-dependent HitMaps allow for detailed further studies Some modules, in particular in end-cap C, show suspicious time dependent noise patterns under investigation Increased noise for HPK modules in end-caps facing air gaps under investigation Ongoing: estimation of expected ion flux More studies are ongoing (heater pads, SONAR, MC tuning), or will start soon (beam induced noise ) 14
Backup 15
SCT Indexing Scheme ibec_ilayer_iphi_ieta_iside -2/0/+2 0..8 0 55-6 +6 0/1 16
Single Modules EC-C (Disk 7)?? 17
Another Run (00300415) Number of noisy strips vs. Lumi block This run shows same structures for same modules as before (for disk 3), modules of disk 2,4,7 seem not to be affected 30 25 20 15 NNS2D_EC 100 Entries 7425 Mean x 17.1 90 Mean y 20.92 RMS x 79.18 80 RMS y 7.376 70 60 50 40 10 5 30 20 10 100 200 300 400 500 600 700 800 900 1000 0 18
Diffusion Equation Determine ion flux on modules facing air gaps c: concentration of ions q: source term (production of ions from charged particles passing the air gap) α: radiative recombination coefficient 0 0 0 0 1 D: diffusion coefficient; in strong Bz fieldd = @ 0 0 0 A 0 0 D zz with T: temperature [K] M1 and M2: molar masses [g/mol] p: pressure [atm] σ12: collision diameter [Å] @ c(~x, t) =D c(~x, t)+q(~x) c(~x, t)2 @t D zz = 1.858 10 3 T 3/2p 1/M 1 +1/M 2 p 2 12 Ω: collision integral ( 1) https://en.wikipedia.org/wiki/mass_diffusivity 19
Dissipation Radiative recombination (e - + N2 + N2) can reduce the number of ions reaching surface of modules 0.39 For nitrogen: = 2.2 10 7 Te cm 3 300 K s 1 Journal of Geophysical Research, Vol. 109, 2004 If recombination is not a sizeable effect, all the ions which are produced are reaching the surface Nick s previous calculation will hold (maybe reduced by factor 1/2) If recombination is sizeable, the ion flux on modules is (in the static case) given by! (x, y, z = L) = 1 2 To understand if the system will reach a static behaviour we have to solve the diffusion equation, no? Z L If not static things will become more complicated!!! 0 q(~x)dz Z L 0 c(~x) 2 dz 20