Summary of Blayer Replacement Workshop

1
Summary of B-layer Replacement Workshop

• October ATLAS Week
• CERN, 10 October 2007
• G. Darbo - INFN / Genova
• Workshop page
• http//indico.cern.ch/conferenceDisplay.py?confId
  20119

2
B-Layer Workshop
http//indico.cern.ch/conferenceDisplay.py?confId
21107

• B-layer Replacement WS
• Sept. 28-29, 2007
• Workshop numbers
• 1 1/2 days
• 6 Sessions round table
• 56 registered participants

3
Why Pixel Replacement

• B-layer was designed to survive 3 years of LHC
  nominal luminosity (1034cm-2s-1)
• NIEL fluence 1015 neq/cm2 (mainly from pions)
• Ionizing Dose 500 kGy (50 Mrad)
• Sensors
• Not fully depleted at 600 V ? lower charge
  collected, lower efficiency
• Leakage current (noise)
• Charge trapping ? lower charge collected, lower
  efficiency
• Electronics
• Transistor VT shift due to charge trapping in
  the gate oxide (Ionizing Radiation). FE and MCC
  tested to 500 kGy)
• New LHC machine scenario could move replacement
  date into the future, but large unknowns on
  operation and cooling runaway may make it
  sooner.
• Target date for the Workshop studyWinter
  shutdown 2012

B-layer bias voltage T 0ºC operation, T
-20ºC no operation, T 20ºC for short
accesses 50 headroom uncertainty
Max DETbias
4
Replacement Constraints

• Aim of the replacement is to improve, or at least
  restore, initial performance of the Pixel (ID)
  detector.
• This scenario involves improvements to
  sensor/electronics design, to module and
  mechanical geometry.
• Hope for beam pipe radius reduction, such that a
  3237 mm radius pixel detector can be inserted
  inside the existing detector.
• The new b-layer could be inserted together with
  the beam pipe in this scenario.
• Two scenarios have been studied and analysed in
  the workshop
• Add a small R b-layer leave existing one
  (having reduced efficiency)
• Add a small R b-layer remove the existing one
  (remove material)
• Both A B requires to move the pixel package
  out of the pit. Can we do an "in situ" insertion
  without removing the pixel package, or is this
  completely excluded?
• The additional constraint we used in the study
  was the replacement date 2012 and the shutdown
  time 6 months

Replacement
Insertable
A
B
5
Layout Simulation Method and Software

• DC1 model of ID upgraded pixel.
• Latest xKalmanOO (tracking).
• Private version of b-tagging software with
  secondary vertices and recalibration for
  different trackerprocess configurations
• B-tagging software was not specifically tuned for
  new setup (all track selection cuts were the same
  as for current pixel).

Ref. V. Kostyukhin, P. Nevski, A. Rozanov
Results presented for upgraded pixel with added
layer (4L) and with modified single b-layer
together with results obtained with current
pixel detector (3L).
Modified layout with single b-layer
Layout with 4-layers L4
Rb1 37.0 mmRb2 absentR1 88.5 mmR2
122.5 mm
Rb1 37.0mm, 1.2 X0 Rb2 50.5mm , 2.2 X0 R1
88.5mmR2 122.5mm
6
Material Studies Performance
B-tagging vs. ?

• The b-layer with z-pitch reduced from 400 to
  250µm improves substantially Z-vertex resolution
  (top left)
• The addition of material is critical to b-tagging
  at large ??? (figure top right is b-tagging for
  present pixel detector)

Ref. A. Rozanov
7
B-tagging Results
Ref. V. Kostyukhin
Preliminary!!!
WH(120)?uu(bb), no pileup, ATLFAST jets,
reconstructed primary vertex.

• Tracking performance seems ok for 4 layers case
  but b-tagging is worse comparing with current
  design. Track part of b-tagging is mainly
  responsible for worsening.
• Layout with new single b-layer at R37mm and
  removed current b-layer gives significantly
  better performance with existing tracking
  software.
• Single b-layer at R37mm provides some increase
  of b-tagging rejection at high ? region weak
  point of existing pixel detector.

8
Disconnect / Extract the Pixel
Ref. D. Giugni
TASKS in this ENVIROMENT Set up and align the
DST, Extract the detector, Rotate the DST, Hook
up the DST to the crane. DURATION, MANPOWER,
INTEGRAL DOSE 84h, 3 people, 3.1mSv
lt50 µSv/h ? Simple controlled radiation area
/ Low occupancy areas lt 2 mSv/h ? Limited stay
area / Low occupancy areas

• The area of the ATLAS beam line will be
  classified as Controlled Radiation Area.
• Assuming that the Full Body Dose exceeds 50µSv/h
  the area should also be classified as Limited
  Stay Area(lt 2mSv/h).
• This means that workers have a controlled access,
  must wear a personal dosimeter an operational
  dosimeter and be classified as Radiation
  Workers.

9
On Surface in the SR1
TASKS in this ENVIROMENT Remove SQPs, Access
the beam pipe, Remove the beam pipe, Remove and
insert the B-layer, Connectivity Test DURATION,
MANPOWER, INTEGRAL DOSE 600h (depending upon
the scenario), 4 people present, 28.9mSv
Supervised Radiation AREA lt 15 µSv/h
ITT
Simple Controlled Radiation AREA lt50µSv/h

• Great attention has to be paid to avoid
  radioactive contamination. Grinding and cutting
  activated material would cause a contamination in
  the airborne and on the surfaces.

Ref. D. Giugni
10
Services - Additions
Intreferes with T0 Cables
OSP
Detector
ISP
Requires Removal Of all SQP on this side
Modify
BPSS

• (Some) critical issues
• Service Panels (SQPs) are highly integrated both
  electrically and mechanically
• Electrical modularity of services locked in
  Mechanically
• Tight envelopes make re-arrangement of services
  difficult
• Penetration at PP1 uses almost all real estate at
  package ends
• Fibers case A requires new opto-fibers to be
  re-routed (difficult) and remapping with new
  electronics, case B fiber reuse requires same
  nominal location of New Optical drivers

Ref. E. Anderssen
11
What to Open

• Case AA-Side (to insert B-layer leaving existing
  one)
• Need to install new cooling circuits, cant be
  C-Side, not quite enough spare Holes
• Requires complete disassembly of at least 2 SQPs
  to get at least 6 cooling circuits in
• Gives access to BPSS to install new PP0s
• A lot of work, all after removal, likely on
  critical path
• Case BC-Side (to remove existing B-layer and add
  a new one)
• Need clear aperture to pass BP and new BL
  incannot pass PP1 Cruciform
• Only required removal of T0 cables and enough
  SQPs to be comfortable removing BPSS
• Remove SQPs but not dissassemble to modify for
  cable routing of new PP0s
• Access to BPSS to install new PP0s
• On Critical Path, but less work than A
• Both require removal of BPSS and at least 2 SQPs
  on one side.
• Schedule
• Case A and B differ in the needed operations but
  the total time will not differ greatly. Case B
  requires, in first analysis 9.5 months. Case A
  could be slightly faster.
• More detailed studies are needed.

Ref. E. Anderssen
Ref. D. Giugni
12
Envelope Studies

• New beam pipe radius
• Nominally reduced to R25, theoretically to R17
  (instead of R29 now) but
• R17 will require feedback from future Atlas run
  see Rays talk at
• Current B-layer envelope R45.5 to R74 28.5 mm.
• Case A IR35 OR41
• Case B IR35 OR71.5
• Studies which might give more space
• Insulation thickness between BP and B-Layer
• Radial Adjustment of Beam-pipe
• Assembly sequence/tooling which requires less
  clearance

Ref. A. Catinaccio
13
Layout Concepts Single Layer
Monolitic OD 88 mm ID 75 mm

• Developing directions
• Made stave prototype (for SLHC upgrade test) with
  carbon foams with good thermal conductivity but
  relatively high density (?0.55 g/cc).
• Thermal and FEA analysis.
• Look to low density (?0.15 g/cc) foam or
  nanotubes.

Ref. M. Gilchieriese, M. Garcia Sciveres
Stave - 2 OD 88 mm ID 70 mm
Stave - 1 OD 93 mm ID 70 mm
14
Layout Concept Double b-layer
Ref. N. Wermes
incl. angle B2 6o incl. angle B1 10o
nom. radius B2 49 mm nom. radius B1 37 mm
envelope ok
staves B2 20 staves B1 16
Wolfgan Dietsche, Walter Ockenfels
15
Local Support Structure - Stave
Ref. K.W. Glitza

• Wuppertal stave studies
• Use homogeneous structure all carbon based.
• Reduce material carbon foam for the cover (
  density 0.50.9 g/cm³) thermal conductivity
  (in plane 140 W/mK out of plane 70 W/mK)
• Material and stability woven carbon pipes

sensor
cover
base
base
pipe
Prototype
16
Sensors

• Basically all technologies covered by
  presentations
• N-on-n is the Pixel proven technology
• P-on-n gives the same radiation tolerance than
  n-on-n, but single-side process cheaper and with
  higher yield
• Thin sensors lower material, lower leakage
  current, lower charge
• Diamond works at room temperature, pCVD (spatial
  resolution?) and scCVD (large enough sensors?).
  Work at room temperature, low noise, high cost.
• 3D sensor high level of radiation hardness
  (highest collected charge after irradiation),
  active edge, but cost.
• Gossip gaseous detector, many interesting
  features, but still on RD stage.

Irradiation / testbeam studies are underway (or
about to start) with most of the technologies
using the present pixel electronics Some of the
technologies have RD proposal to SLHC
17
3D Sensors - New Results
Efficiency

• Efficiency studies of 3D sensors with 2006 test
  beam for perpendicular and inclined sensors
• Time walk (20ns plateau) measurements with
  injected charge for 3D structures and FE settings.

• 0º
• ? 95.90.1

• 10º
• ? 99.60.1

BEAM
Threshold 3200 e Structure 3E
Column shadow
Ref. M.Cristinziani E.Bolle, O.Rohne 3D ATLAS
RD Meeting 27/10/2007
Time walk
Charge overdrive (above threshold) for ?T 20 ns

• 20 gt power

• 40 gt power

?
?
18
Electronics - System Architecture

• Electronics designed for higher occupancy,
  radiation tolerance, SEU
• Larger FE chip increases module live fraction
  (70 ? 90)
• Feasible single chip (active edge) modules or
  standard multi-chip
• Remove MCC from module and have a SCC at stave
  end
• Opto-links use same SIMM/GRIN fibers but at
  higher speed.

19
Front-end Chip - FE-I4

• Laboratories involved Bonn, CPPM, Genova,
  Nikhef, LBNL.
• Very spread collaboration, difficult organize
  common effort.
• Submission planned 12/2008.

20
Block Design for FE-I4

• LBNL Prototype chip Submitted Feb07 in 0.13µm
  CMOS (IBM)
• Array of 40-cell columns.
• There are 21 total columns in a die which is
  3.6mm x 2.8mm with 110 I/O pads.
• First results very promising (bottom left noise
  vs. input C for TP design). Slope 76 e-/fF.
• designer is Abder Mekkaoui

• Bonn designing blocks for FE-I4
• LVDS TX/RX submitted to 0.13µm UMC.
• 8/10-bit DAC with monotonic behavior
• Designer is Michael Karagounis
• Nikhef submitted band-gap reference in 0.13 µm
  IBM
• Designer is V.Gromov et al.,
• Collaboration MPC will be submitted through
  MOSIS Jan08

LVDS RX
LVDS TX
21
Workshop Outcome

• The Workshop was quite successful, attracted more
  than 70 people and was the first serious attempt
  to look into the b-layer replacement.
• It was the first time we did substantial study
  in
• Simulating the b-tagging performance of the
  replacement scenarios.
• Analyzing the impact of activation in the
  operations to be made to replace the detector.
• Evaluating the impact on the schedule of
  different scenarios.
• We have been somewhat surprised
• Reduction of material is probably more important
  than reduction of the radius
• Scope is bigger than foreseen for short time
  scale. Services are on the critical path.
  Reconnect services cannot be done in short time.
  Replace services could be faster than
  disconnecting and connecting again, but cost
  would exceed that of the funded project (34
  MSF).
• Threshold to change is higher sensors and
  electronics could survive longer and risk of the
  intervention (in complexity and time scale) are
  higher than foreseen.

22
What to do next

• Develop a more precise and comprehensive analysis
  of the lifetime issues for the present B-layer
• Effect of the reduced collected charge from the
  sensors has to be studied throughout electronics
  signal generation and finally to b-tagging
  efficiency.
• Define some "thresholds" that would imply risks
  we are willing to take, or timescales we are
  willing to consider for replacements.
• B-layer is the unique layer, in the whole ID,
  which its reduced performance would have such a
  major effect on Physic achievements (basically,
  destroying or seriously degrading b-tagging).
• Develop a more complete plan, in collaboration
  with TC/PO for the complete replacement
  operation, to understand the timescales and
  projections for allowed shutdowns.
• Push RD to reduce material budget. This must
  happen mainly in the fields of sensors,
  electronics and integrated support structures
  (staves).
• Study ways of extending lifetime of present
  detector cooling, operating sensors for longer
  life (limit bias current to keep them cooler).

23
Last but not Least (Conclusions)

• The B-layer replacement, at a modest scope (no
  more than about 34 MSF CORE equivalent) is a
  funded project, though not approved in detail.
• it is not only a necessary intermediate step
  before going to a 1035 upgrade but should extend
  the secondary vertex related physics reach for
  ATLAS (or maintain it for longer).
• The timescale must fit with the SLHC, although
  that is not an approved project with a
  schedule.
• We will gain critical experience (pixel hookup
  and closing ATLAS) in next months so more natural
  time to reach a conclusion on baseline
  replacement option is again summer 2008.
• More engineering manpower will be released from
  the present priority task of Pixel detector
  completion and will be directed to the study of
  the new detector.