GOSSIP

1
GOSSIP a new vertex detector for ATLAS
Harry van der Graaf NIKHEF, Amsterdam Univ. of
Bonn, Nov 23, 2006
2
New gaseous detectors the application of pixel
sensors as direct anode
Harry van der Graaf, NIKHEF, Amsterdam University
of Bonn Oct 14, 2004
NIKHEF Auke-Pieter Colijn Alessandro
Fornaini Harry van der Graaf Peter
Kluit Jan Timmermans Jan Visschers Maximilie
n Chefdeville Saclay CEA DAPNIA Paul
Colas Yannis Giomataris Arnaud
Giganon Univ. Twente/Mesa Jurriaan
Schmitz CERN/Medipix Constm Eric Heijne Xavie
Llopart Michael Campbell
Thanks to Wim Gotink Joop Rovenkamp Arnaud
Giganon
3
MIP
MIP
InGrid
Cathode foil
CMOS pixel array
CMOS pixel chip slimmed to 30 µm
Drift gap 1 mm Max. drift time 16 ns
GOSSIP Gas On Slimmed SIlicon Pixels
4
Si (vertex) track detector GOSSIP
Gas 1 mm as detection medium 99 chance to have
at least 1 e- Gas amplification 1000 Single
electron sensitive All signals arrive within 16
ns

• Si strip detectors
• Si pixel detectors
• MAPs

5
April 2004 Micromegas MediPix
2 NIKHEF/Saclay/Univ. Twente
55Fe, 1s
No source, 1s
55Fe
Cathode (drift) plane
Drift space 15 mm
Micromegas
Baseplate
MediPix2 pixel sensor Brass spacer block Printed
circuit board Aluminum base plate
14 mm
d-ray!
He/Isobutane 80/20 Modified MediPix
MIPs
6
Integrate Micromegas and pixel sensor InGrid
wafer post processing by Univ. of Twente,
MESA there is plenty of room at the top
7
Prototypes
hidden pillars!
8
Energy resolution in Argon IsoC4H10 80/20

• Observation of two lines
• Ka _at_ 5.9 keV
• Kß _at_ 6.4 keV
• FWHM of the Ka distribution
• 16.7
• Gain fluctuations
• lt 5

Very good energy resolution Very precise
dimensions d lt 0.1 µm
May 2005
9
A-Si not adequate? Then TwinGrid
10

• Gas instead of Si Pro
• Low mass detector!
• low power 2 µW/pixel (analog), 0.6 µW/pixel
  (digital)
• Total 2.6 µW/pixel ? 0.1 W/cm2 due to
• extremely low source capacitance (10 fF)
• fast arbitrary large charge signals
• no bias current
• ? little material required
• for power cooling
• Thinning (Slimming) of CMOS pixel chip
• to 50 µm
• detection layer only 0.06 radiation length

11

• Gas instead of Si Pro
• a GOSSIP detector measures a track segment
  instead of a point-in-space
• single-electron sensitive (eff gt 95 )
• 10 e- along 1 mm track length
• spatial resolution 50 µm / v10 15 µm

14 mm3

• - radiation hardness
• gas is flushed
• CMOS chip (130, 90, 45 nm technology)
  sufficiently radiation hard for SLHC
• very low sensitivity for neutrons, X-rays and
  gammas
• simple CMOS chip
• modest (small area) analog input circuits
• no bias current simpeler input circuit
• pixel area available for data storage
  communication
• Cheap

12

• Gas instead of Si Con
• GOSSIP has 1 mm detection layer
• parallax error elimination requires drift time
  measurement
• ? with single-electron measurement track
  segment data per layer!
• Gas-filled detector
• chamber ageing (deposit on electrodes)
• discharges (sparks, too large signals) ruin CMOS
  chip
• ? showstoppers !

13
CMOS Chip protection against -
discharges - sparks - HV
breakdowns - too large signals
Silicon Protection SiProt
Amorph Si (segmented)
Emperical method Try RPC principle
14
plasma
A-Si
-
-
-
-

• RPC principle reduction of local E-field
• Avalanche charge electrostatic induction
  towards input pad
• Specific resistance - high enough to block
  avalanche charge
• - low enough to flow signal current
• - layer thickness 4 µm, Rvol gt 0.2 100 GO/cm

Technology A-Si deposit standard wafer post
processing, but wafers may get too hot ? Univ.
of Neuchatel/IMT/P. Jarron (CERN) uses this for
integrated X-ray sensor/convertor on MediPix 2
Test put Thorium in gas Radon a-decays -
large (proportional) signals - Discharges like
short circuits
15
The SiProt chamber

• Low temperature deposition (lt 250C) of a
  4 µm thick a-Si layer of 1011 O.cm resistivity
• Experimental setup
• 1 bare anode and 1 a-Si covered anode with
  Micromegas on top
• Gain curve with an Iron 55 source
• Induce discharges by means of 5 MeV alphas from
  Th source in gas
• Record grid signals

Aluminum
a-Si
Micromegas grids prot. anode un-prot. anode
To digital scope or pre-amplifier
16
SiProt chamber
Th container
cathode
Digital scope
17
Discharge Signals study

• No preamplifier
• Ar 20 iC4H10
• Signals from 5 MeV alphas
• Fast digital scope

18
UN
PROT
Discharge signals short-circuit between grid and
anode due to plasma
UN
PROT

• no hot plasma on
• pixel input pads
• reduced charge
• current
• Looks like it works!
• Next try on Medipix
• chips

UN
PROT
19
Gain measurement
55Fesource Look at the pulses from a (calibrated)
preamplifier (low grid voltage) Look at the
current flowing through the power supply (high
grid voltage)
from current
from pulse height
No sparks up to 570 V on the grid ! Next step
SiProt (and InGrid) on Medipix, TimePix
20
Ageing

• Ageing of a GOSSIP detector versus wire chambers
• - Ratio of anode surface/chamber volume
• thin wire surface versus anode plane (20x)
• - Low gas gain (1 k) due to fast signal and low
  source capacity (20x)
• total factor 400 x
• So application as GOSSIP vertex detector in
  Super LHC
• 1016 MIP/cm2
• seems feasible

First try
21
Ageing remember the MSGCs.
22
Irradiation with 8 keV X-raysNo rate effects
up to anode current density of 0.2 µA / mm2 ?
very fast track counting possible!
After 0.3 Coulomb/mm2 ? (eq. 3.7 x 1016
MIPs/cm2 !!) deposit of carbon polymer on anode
is clearly visible. Micromegas is clean
(?!) Little deposit on cathode, and Chamber
still worked!
23
Ongoing projects
24
GOSSIPO-1 test of preamp-shaper-discriminator
for GOSSIP MultiProjectWafer in 0.13 µm
technology
Cpar 10fF50fF
Very low (parasitic) capacitance at the input
(Cpar ? 10fF) .
Coaxial-like layout of the input interconnection.
Parasitic metal-to-metal fringe capacitances.
25
Triple well layout isolation of
digital and analog sections
26

• - match extreme small source capacity 15 fF
• peaking time 40 ns
• noise (expected 60 e- input eq.)
• power 2 µW/pixel (!)

GOSSIPO chip Submitted December 2005.

• Input noise eq. reached
• No effect of digital switching
• within pixel

MultiProject Wafer Vladimir Gromov/NIKHEF CERN
Micro-electronics group
27
GOSSIPO-2

• test of
• preamp-shaper-discriminator
• and
• 700 MHz TDC per pixel
• 0.13 µm technology
• containing 16 x 16 pixels
• Submission Nov 29, 2006
• Can be used for GOSSIP demo!

28

• New mechanics cooling concepts for Gossip
• As little as possible material
• detector consists of foil!
• less power required (? less cooling) w.r.t. Si

laundry line
string power, chip support, cooling
29
Virtual goal ATLAS pixel upgrade

• - Ladder strings fixed to end cones
• Integration of beam pipe, end cones pixel
  vertex detector
• 5/10 layers (0.06 rad. length each!) seems
  feasible

30
data lines (Cu/kapton)
ladder cross section
casted aluminium
Gossip chip InGrid drift gap cathode foil
Stainless steel tube - string -
power - CO2 cooling
ladder side view
ladder top view
31

• First practical GOSSIP
• with
• CMS Vertex Pixel FE chip PSI 46 ( ATLAS FE
  pixel chip?)
• apply A-Si protection layer
• apply InGrid
• mount Gossips on pcb beam telescope
• Testbeam 1st half 2007

PSI, Univ. Nijmegen, NIKHEF,
32

• Gossip projects at NIKHEF
• Univ. Twente/
• Saclay
• CERN
• PSI
• EUDET
• Discharge protection
• InGrid/TwinGrid/TripleGrid
• Construction of detector MediPix2 SiProt
  InGrid
• Construction of detector TimePix SiProt
  InGrid
• Beam Telescope with CMS PSI 46 pixel chip
• Ageing studies
• - CO2 cooling

33
NIKHEF Harry van der Graaf Jan
Timmermans Jan Visschers Maximilien
Chefdeville Vladimir Gromov Ruud
Kluit Fred Hartjes Els Koffeman Martin
Fransen Saclay CEA DAPNIA Paul Colas Yannis
Giomataris Dan Burke Univ. Twente/Mesa Jurri
aan Schmitz Cora Salm Sander
Smits Victor Blanco Carballo CERN Erik
Heine Medipix Consortium
Thanks to Wim Gotink Joop Rovenkamp
Gossip the electronic bubble chamber
34
Sr-90 ß-source
1.2 mm