Detector R&D at NIKHEFSAC meeting, 20 April 2007
Jan Timmermans
• Pixel readout of gas detectors: Gridpix, Timepix, …
• Medipix, RELAXd• 100 m RASNIK alignment
Micro Patterned Gaseous Detectors
GEM
Micromegas
• High field created by Gas Gain Grids
• Most popular: GEM & Micromegas
Use ‘naked’ CMOS pixel readout chip as anode
Results pixel readout gas detectors
Observation of min. ionising cosmic muons: high spatial resolution + individual cluster counting !
δ ray
NIKHEF-Saclay-CERN-Twente
NIM A540 (2005) 295 (physics/0409048)
55x55 μm2 pixels
TimePix(EUDET: Bonn,Freiburg, Saclay, CERN, NIKHEF)
•Distribute clock to full 256x256 pixel matrix (50-100-160MHz)
•Enable counting by first hit after ‘shutter’ opens, until ‘shutter’ closes (common stop); also time-over-threshold possible
•Dynamic range 214 x 10 ns = 160 μs
•(for the time being) no zero-suppress
to remain fully compatible with Medipix2
•Shaping time ~200 ns
•Keep same chip-size, pixel-size, readout protocol
•1st full reticle submit done July 2006; IT WORKS!
11 22 44
55
66
55μm
55μm
First tracks with TimePix+SiProt+Micromegas:
cosmics + α particle
V_grid= -400 V; gain ~5000
He/Isobutane 80/20
Acquisition (‘shutter’) time ~ 1sec
Time_over_threshold mode pulseheight
And triggered cosmics (sum of several triggers)
Shutter time = 15 μs
TimePix chip stayed alive for 40 days with He/isobutane 80/20
And ‘died’ on 26/1/07 after 1.5 day with Ar/isobutane 80/20
Blue = late
Red = early
First INGRID tests
• Post-processing of a 4” wafer successful!– SiProt covered– InGrid– Diced to individual chips
• Handling critical• SU8 photo-resist dike
around the pixel matrix
• Perfect alignment pads/holes
• Pillars in between pads
• 2 chips tested in He 20% iC4H10
• 55Fe signals at Vgrid = -280 V i.e. G ~ 200• Almost full homogeneous response• Still short lifetime (sparks even at low gains)
Micromegas covered Timepix
InGrid covered Medipix
Moire effects with Micromegas
TwinGridAlternative:
Fred Hartjes 10
NIKHEF detector R&D meeting. January 23, 2007
MIP response for Gossip 21
charge signal (ke-)
0 10 20 30 40 50
# of
eve
nts
0
100
200
300
400
500
Measured histogramFitted convoluted PoissonAssumed pedestal peak
Vmesh = - 600 V
Vcath = -1000 V
Drift gap 1.2 mm=> Edrift = 3.3 kV/cm
pad pattern on glassBrass Micromegasgas: Kr/DME 74/26Exitation by mips from 90Sr source2.55% pedestal events14-2-2007
GOSSIP: 1.2 mm thin gas detector + Micromegas Gas: Kr/DME 74/26
Measured histogram with fitted convoluted Poisson
Cluster density calculated from two independent methods
1. Magnitude pedestal peak 2.55 +/- 0.07 % pedestal events
=> cluster density 39.1+/-0.8 cm-1
2. Fitting convoluted Poisson to experimental data
Using pdf =1 the best fit occurs at 38.2 +/- 2.6 cm-1 ¥
¥Note that this number depends on the pdf value
LiteratureCluster density Kr/DME 74/26: 40.2 cm-1
GOSSIPO-2 chip:
• CMOS 130nm technology
• Multi-project wafer run (design ready Nov. 2006)
•16 x 16 pixels
•Each pixel: preamp, shaper, discriminator + 700 MHz clock (active for short interpolating period)
Waiting for the run to take place
Medipix and RELAXd
18 April 2007 - Jan Visschers
Medipix Collaboration
- Univ + INFN CagliariUniv + INFN Cagliari
- CEA-LIST SaclayCEA-LIST Saclay
- CERN Genève- CERN Genève- Univ d'Auvergne Univ d'Auvergne
- Univ Erlangen- Univ Erlangen
- ESRF Grenoble- ESRF Grenoble
- Univ Freiburg- Univ Freiburg
- Univ Glasgow- Univ Glasgow
- IFAE Barcelona- IFAE Barcelona
- Mitthoegskolan - Mitthoegskolan
Sundsvall Sundsvall
- MRC-LMB Cambridge - MRC-LMB Cambridge
- Univ + INFN Napoli- Univ + INFN Napoli
- - NIKHEF AmsterdamNIKHEF Amsterdam
- Univ + INFN Pisa- Univ + INFN Pisa
- FZU CAS Prague - FZU CAS Prague - IEAP CTU Prague IEAP CTU Prague - SSL Berkeley SSL Berkeley
Spokespersons: Spokespersons: Michael CAMPBELL CERNMichael CAMPBELL CERNJan VISSCHERS NIKHEFJan VISSCHERS NIKHEF
Quad Medipix
1sensor on 4 Medipix chips6-layer ultra-fine line designLaser-drilled micro-via’sKapton Buildup technology0.25 Megapixels
RELAXd: Tile-able Microsystems
1. Edgeless Planar Silicon Sensors
2. Through-CMOS contacting
3. High-bandwidth readout
Medipix 2.1 ROIC
1. Canberra
2. IMEC
3. NIKHEF4. PANalytical (application)
Eureka Project
Raw image of 90Sr source with current Quad carrier.
Medipix ASIC
Through-CMOS connections
Multi-layered chip carrier board
Medipix ASIC
Quad sensor
• Through-silicon interconnect• Medipix wafer thinning: 100 um• Ball-Grid Array on the back side of the Medipix chip• Balls 500 um diameter, 800 um pitch
Frame and Cooling
Large area (9x9 cm2)< 100 Watts total powerCeramics FrameCooling pipes through framePower through frame
Roadmap
Through-via carrierZdenek Vykydal
Multi-Gbps carrierBas van der Heijden
RELAXD module(end 2007)John Hug
Current Quad
Frame, Power,Cooling (2009)
100m RASNIK alignment
Harry van der Graaf
Coded Mask
Light Source LensCCD
8000 systems installed in ATLAS
Principle of CCD-RASNIK(as used for muon chamber alignment in L3)
100 m RASNIK for CLICCLIC
Accelerator length: 2 x 15 kmNumber of elements: 15.000Number of ‘4 m’ RASNIKs: 15.000/30.000Number of ‘100 m’ RASNIKs: 600
Laser zone lenshole dia. 50 mm
RasCam
100 m (vacuum tube!)
Image position on sensor. Response of earthquake in Mid-Atlantic,5 Richter Scale, on March 1, 2007
Tidal movement: crust deformation due to angular positionof Sun and Moon, measured with RasClic, CERN, Geneva (filtered raw data)
The ‘Earth Hum’, measured after the ‘tsunami’ earthquake, Xmas 2004[Park et al]
(Direct) Fourier coverted raw X-data of RasClic, Feb 2007New wide-band seismometer?
‘Hum’ modes