Aspects of Grid Pix technology Nikhef R D group

Aspects of Grid. Pix technology Nikhef R&D group Presented by Fred Hartjes 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes

Contents Introduction into Grid. Pix Practical directions Mechanical support Field cage HV connection DAQ Passivation Choice of the gas Peculiarities of Time. Pix and Grid. Pix Time walk Cross talk 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 2

Principle of Grid. Pix technology High granularity pixel chip Cell pitch 55 – 60 μm in X and Y Drift gap: from 1 mm (Gossip in ATLAS) on to ATLAS trigger layers (1 cm) and ILC TPC (dms) Gas avalanche from a grid (In. Grid) 14 mm Produced by wafer postprocessing (256 c olumn ) s) Gain 5000 - 10000 ws o r 56 (2 m 4 m 1 Fred Hartjes Scaled up 4 x for better visibility 9 th RD 51 collaboration meeting, CERN, February 22, 2012 3

Grid. Pix functioning Pixel chip with integrated Micromegas (In. Grid) Very small pitch => often detecting individual electrons 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 4

Parallel-plate amplification field under grid Moderate amplification field (810 k. V/mm) => detector might be less prone to ageing 100 V/mm Amplification gap 8 k. V/mm Pixel chip 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 5

Timepix as a pixel chip q q q Derived from Medi. Pix (X-ray detection) Matrix of 256 x 256 pixels 55 µm pitch => 14. 08 x 14. 08 mm 2 sensitive area Common clock (100 MHz) to measure drift time for each pixel Two modes of operation Charge arrival time to stop mode Time-Over-Threshold (TOT) mode to measure charge signal magnitude But both modes not simultaneously In. Grid Postprocessing to make gaseous detector 1. 8 µm resistive Si 3 N 4 (Si. RN) for spark protection 2. Amplification grid (In. Grid) on Time. Pix 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 6

Status of Grid. Pix production Transferring production from MESA+ (4” wafer processing) to IZM (8” wafer processing) Time. Pix chips produced on 8” wafers December 14, 2011: successful production Most processing done at IZM by Yevgen Bilevych Part done at MESA+ Deposition protective layer (Si. RN) Chemical activation of polyamide Yield ~ 80 chips 1 st chip tried was OK Excellent HV behaviour, good tracks seen Possibly somewhat bigger grid distance => higher grid voltage needed Production two additional wafers starting now Completed end April? ? 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 7

First one tried out by Harry Chips from 14 -12 -2011 Nice tracks from 90 Sr Good uniformity No conclusive results yet 3 other chips initially worked fine But broken because of unsuited HV power supply (at Fermilab) 14 mm run 55 Fe 0 50/5 2 O E/C M D 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 8

R&D on production technology Done at MESA+ Additional HV protection by resistive grid Replacing SU-8 (used for pillars and dykes) BCB KMPR Si. O 2 GEM grid Structures surviving cool down to 165 K of liquid Xe (XENON) 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 9

Infrastructure for Time. Pix Carrier board for MUROS RO (one chip) Relaxd RO possible Dedicated Re. Laxd boards almost finished (place for 4 chips) 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 10

In development: Octopuce 8 Time. Pix chips in a structure for ILC study Saclay (David Attié and Paul Colas) Nikhef (Jan Timmermans) Intermediate board up to 8 chips Bonn (Jochen Kaminski) 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 11

Example of a field cage (100 and 20 mm) Copper plate (guard electrode) surrounding the pixel matrix Field shaping By strips By wires 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 12

HV connection needs attention Example: Time. Pix with field cage Filtering HV noise Limiting discharge current from filtering capacitor Vfield Cathode plane In. Grid connection most precious Use resistor 10 – 100 MΩ Leakage current few n. A Guard potential should follow In. Grid potential Vguard Guard plane Short distance => avoid sparking HV regulation Vgrid is the master Vguard is the slave Automatic remote control system is preferred 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Ingrid Voltage divider circuit Fred Hartjes 13

Use for Vgrid HV supply with sensitive trip Most commercial supplies not suited Gentle ramping not easy No sensitive trip May deliver output currents in m. A region => broken Time. Pix Nikhef mini. HV dedicated for Grid. Pix application Sophisticated ramping and tripping until -1000 V Current measurement and trip setting in n. A region Versions for -2 k. V and higher in development 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 14

Systems basically developed for Medi. Pix DAQ MUROS (Nikhef) obsolete Re. Laxd (Nikhef) Developed together with industry (Panalytical) => closed FPGA code 10 Gb system in development at Nikhef (SPIDR) open source code Scalable Readout System (SRS) interface for Time. Pix (Universities of Mainz and Bonn) 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 15

Mechanical structure of a Time. Pix chip equipped with In. Grid Dyke (~ 250 µm wide) surrounding the pixel matrix for mechanical termination Dyke 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 16

Passivation of edge preferred May be OK for 400 V He/Ar mixtures Sparking at < 600 V DME/CO 2 etc 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 17

Passivation by Glob Top Using Dymax type 9001 -EV-3. 1 UV curing To be deposited in 3 – 4 layers Slightly conductive but still low leakage current Few n. A 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 18

Choice of the gas Ar/i. C 4 H 10 80/20 Regular drift chamber gases like Ar/He in C*H*, T 2 K, …… Easy in use Fair mobility (~ 50 µm/ns at 1 k. V/cm) Diffusion 150 – 200 µm/√cm Best for bigger drift distance (> few cm) 3 mm Avoiding large drift voltages and long charge collection times But paying with larger diffusion DME/CO 2 50/50 Extreme property gas (DME/CO 2) Affects many plastics (swelling) Very low mobility (~10 µm/ns at 2 k. V/cm) Very low Lorentz angle (~ 4°/T) Low diffusion ~ 70 µm/√cm Best for short drift distances (≤ 1 cm) 3 mm Small diffusion, beautiful tracks Better match with small (55 µm) pixel size Less scattering in space due to time walk 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 19

Peculiarities of Time. Pix and Grid. Pix 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 20

Present Time. Pix much hindered by time walk Same event from aside and from top Very small diffusion but big time walk Measurement in DME/CO 2 19. 3 mm drift distance From aside From top 3 mm 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 21

Indication for cross talk in the drift time spectrum Time walk less dominant at high gain But tail from small signals remains => possibly cross talk Caused by Si. RN layer (no cross talk at unprotected chip) Vg = -530 V -570 V /CO 2 Gossip (1 mm Grid. Pix) -600 V /50 0 -620 V 5 ME D 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 22

Conclusions Grid. Pix technology brings many benefits for accurate tracking But one has to learn how to profit from them Extreme gas mixtures is less easy but advantageous Develop the photolithographic production Newly produced In. Grids are probably fine, but no conclusive results yet Time walk will be greatly improved by dedicated front end chip (Time. Pix-3) Time. Pix-3 has Gossipo frontend Very low noise level (60 – 80 e- RMS) Peaking time 5 – 30 ns Simultaneously measuring arrival time to stop and TOT => time walk correction possible 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 23

(Potential) users list as given by Harry Institute/group Nikhef/Vertex-Pixel Nikhef/ILC Nikhef/LVL 1 Nikhef/TRT Nikhef/Xenon. Darwin Saclay Univ. Bonn ZEPLIN PSI NA 61 LHe. C CAST ECAP/Uni. Erlangen MIT/LNS HIP 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Experiment ATLAS ILC ATLAS XENON ILC ATLAS ILC TPC ATLAS Mueee NA 61 Pola. Pix Dark. Forces Helsinki RD 51 Fred Hartjes Representative Harry van der Graaf Jan Timmermans Nigel Hessey Anatoli Romaniouk Patrick Decowski, Matteo Alfonsi Paul Colas Norbert Wermes Klaus Desch Vitaly Chepel Malte Hildebrandt Marek Gazdzicki Alessandro Polini, Peter Kostka Ionnis Giomataris Thilo Mechil Peter Fisher Francisco Garcia Max Chefdeville 24

Spare 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 25

Rise of pixel efficiency curve after plateau Additional peak at charge signal spectrum (Time Over Threshold (TOT) is indication of the magnitude of the charge signal) Vg = -480 V Vg = -440 V C e/i H 9 th RD 51 collaboration meeting, CERN, February 22, 2012 H 10 4 0 0/2 8 Fred Hartjes 26

Single electron hits from Hit plot (from 55 Fe conversion) one 55 Fe conversion Plot focused on small signals Larger signals all dark red Small signal pixels (bright colour) always accompanied by a pixel having a high signal => may be caused by cross talk 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 27

Comparing DME/CO 2 to Ar/isobutane Tracks under 10° DME/CO 2 50/50 Ar/i. C 4 H 10 80/20 In Y direction 9 th RD 51 collaboration meeting, CERN, February 22, 2012 In Y direction Fred Hartjes 28

Chamber gas: DME/CO 2 50/50 o o Very slow and “cool” gas High drift field required Very low diffusion Suited for TPC Drift fields used in Gossips o 2 k. V/cm (lowest diffusion) o 6 k. V/cm (Vd = 50 µm/ns) o LHC tracking 9 th RD 51 collaboration meeting, CERN, February 22, 2012 Fred Hartjes 29