Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Forward pixels – on tracking at higher η in CMS
Frank MeierUniversity of Nebraska-Lincoln
Januar 23, 2015
1 / 40
Introduction
2 / 40
Introduction
3 / 40
Overview
Introduction
Phase-I forward pixelsPhase-I: Motivation
Phase-II forward pixels
Very forward pixelsSimulationsSensor R&D:HardwareConclusions
Conclusions
4 / 40
Phase-I forward pixels
5 / 40
Phase-I forward pixels
The upgrade is foressen to be installed at YETS2016
I Baseline: L = 2 · 1034 cm−2s−1 @ 25 ns → 50 PU withnegligible efficiency loss
I Tolerate: L = 2 · 1034 cm−2s−1 @ 50 ns → 100 PU withreduced performance
I Survive integrated luminosity of 500 fb−1
I Evolutionary upgradeI Robustify tracking: 4 instead of 3 hits; from 2 to 3 disks
(can compensate point losses in strips)6 / 40
Phase-I: Motivation
Requirements (TDR)I Running at 50 or more pile-up, same or better than current
detector in low pile-up (PU)I Maintain or improve the high efficiencies and low fake ratesI Maintain or improve the track impact resolutions and vertex
resolutions
I Maximize 4-pixel-hit coverage over an η range of ±2.5
I No increase of material in the tracking volume; Minimizedegradation due to radiation damage
I Reuse patch panel and off-detector cables and fibers
I Have one FPix module type, simplifying production andmaintenance
I Be compatible with the new smaller diameter beam pipe
I Switch to 2-phase CO2 cooling, target −25◦C
I Install during a slightly extended year-end technical stop
7 / 40
Phase-I forward pixels
Plots show expected fluence in innermost disk:
Comparison based on x-rays.
8 / 40
Phase-I forward pixels
Comparing old and new for efficiency:
9 / 40
Phase-I forward pixels
Improvements in impact parameter resolution:
10 / 40
Phase-I forward pixels
New design reduced mass
Biggest impact: optimized cabling, end-flange of barrel, cooling
11 / 40
Phase-I forward pixels
B-tagging efficiency
12 / 40
Phase-I forward pixels
One physics example: H → ZZ → 4`Analysis does:
1. trigger on di-lepton
2. kinematic reconstruction of 2 Zs from isolated di-leptons
3. reconstruct invariant mass of H
Plot on next slide shows a cut flow-chart to emphasize on cutsthat make the biggest improvement
13 / 40
Phase-I forward pixels
14 / 40
Phase-I forward pixels
Where are we?
I FPix passed CD reviews (critical decision), production canstart
I Final pre-production of modules end of February
I Module production seems to work
I Mechanics starts now as well
Want to emphasize on some module manufacturing issues
15 / 40
Phase-I forward pixels
102Delivery of BBM
101Deliveryof HDI
103Glueing of
HDI to BBM
104Wirebonding
105Encapsulation
106Final inspectionand shipping
107Reworking
206Electrical
acceptancetest module
204Electrical
acceptancetest HDI
203IV test
202Visual
inspection201
Visualinspection
205Visual
inspection module
208Electrical
acceptancetest BBM
207Pull testing
of wirebonds
This is the workflow at UNL.The second production center,Purdue, differs slightly
We use a semi-automatedmanufacturing process using arobotic gantry, a Delvotecwirebonder and some testequipment
Bare modules (sensor + ROC)are made at a commercialvendor (RTI), as well as theHDI (Compunetics)
16 / 40
Phase-I: Motivation
17 / 40
Phase-I forward pixels
https://www.youtube.com/watch?v=ofdntTIwKY4
Beware: it is not an action movie. . .18 / 40
Phase-I forward pixels
19 / 40
Phase-II forward pixels
I Phase-II is beyond about 2023
I 10× more instantaneous luminosity, even more PU
I Major challenge
I Simply said: what layer 1 of the barrel is in Phase-I wouldbecome the outer layers and new technology will be neededfor innermost layer(s)
I will focus on some very forward studies and will briefly come backto this at the end
20 / 40
Very forward pixelsI VFPix effort studies optins in high pseudorapidity tracking
using optimised pixels (2.5 < |η| < 4.0)I Worked on three areas:
I Simulation: Explore physics potential and resolution impactby studying different strawmans and pixel cell geometries
I Sensor R&D: Explore potential of other pixel sizes thancurrently in use
I Hardware: Finish up the construction of a precision telescope(from a PIRE project)
I Will flash status on all three areas.I Baseline layout from TP:
z [mm]0 500 1000 1500 2000 2500
r [m
m]
0
200
400
600
800
1000
1200
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
1.8
2.0
2.2
2.5η
this proposal
21 / 40
Very forward pixels
Latest proposal (strawman 5):
Observe: Extended inner layer, two types of modules. Has insertionalready in mind.
22 / 40
Simulations
Results from TkLayout studies:
Strawman 0: TP; Strawman 5: Previous slide.Pixel size is 75 × 100µm2. Estimated resolutions were (assumingdigitized analog pulse-height and charge sharing based onmounting angle/B-field):
FPix/VFPix: 6.8 × 28µm2
BPix: 6.8 × 14.8µm2
Details: https://twiki.cern.ch/twiki/bin/view/CMS/VFPix
23 / 40
Simulations
This already allows to justify the importance of a well-chosenenvelope.
I Conical geometry performs better than cylindrical
I Keep in mind: there is a strip tracker around this volume
I For a good pT measurement, three points need to bemeasured to good precision:
I Why do we need good pT ? Think of calorimetry and particleflow
24 / 40
Simulations
Results from TkLayout studies:
Strawman 4, similar to 5Observe: To establish charge of a track, a good pT resolutionmatters, even one is tempted to optimize on z resolution
NB: pT resolution comes from lever arm and hit resolution, zresolution dominated by distance to first hit
25 / 40
SimulationsHow to recover z resolution?
layer closer to BS
layer further away
Overlap helps. Some we get in the disks “for free”, somedownstream.
The resolution studies already have some assumptions on zresolution beyond just the local pixel resolution.
26 / 40
SimulationsHot off the press: Resolution studies using PixelAV (courtesyMorris Swartz)
27 / 40
Simulations
Next steps here:
I Use this information in TkLayout simulation
I Have to deal with clusters of size 30. . . – but we already havesome ideas how to handle this
I ROC: psi46 architecture can be translated into more advancedtechnologies. I.e. 110 nm offers reduction of pixel cell by factor5, good for pixels of size 60× 90µm2. Would allow forevolutionary design.
28 / 40
Simulations
Results from Delphes using resolution data from TkLayout:
Details: See TWiki, table of presentations.
29 / 40
Simulations
Results from Delphes using resolution data from TkLayout:
Details: See TWiki, table of presentations.
30 / 40
Sensor R&D:
30
0 µ
m
50 µm
60
0 µ
m
25 µm
I All our pixel geometriesused same area
I Want to study resolutionusing some clever metallayer routing
I Made a sensor designgood for current Psi46pixel, available spring2015Disadvantage: increasedcapacitance, of course
I Will have 300 × 50µm2
and 600 × 25µm2
geometries
31 / 40
Hardware
I We work on external triggering and soft TBM for the DTB
I Build a telescope based on PIRE project
I Uses thinned-down strip sensors with 25µm pitch
I APC128 chip with analog readout
I Telescope had some issues with signal quality. Improving thatto get decent S/N
I Should be ready in spring 2015
I Target resolution: ≈ 1µm
32 / 40
Conclusions
I Established simulation chain
I Resolution studies show very nice results
I One physics case studies with clear benefit from VFPix
I More physics channels in the come
I Hardware studies will soon start as well
I Telescope should be available for beam tests
I Will keep you updated
33 / 40
Conclusions
I Phase-I forward pixel ongoing and on track
I For Phase-II, I’ve shown some studies we did for very forwardtracking
I Efforts ongoing
34 / 40
35 / 40
36 / 40
37 / 40
38 / 40
39 / 40
40 / 40