Overview of MAPS detectors Fergus Wilson Rutherford Appleton Laboratory (with lots of input and slides from Renato Turchetta and the RAL Sensor Design.

Overview of MAPS detectors

Fergus WilsonRutherford Appleton Laboratory

(with lots of input and slides from Renato Turchetta and the RAL Sensor Design Group)

Vertex 2015, Macha Lake, Czech Republic, 15-19 Sep 2014

Outline

Outline Introduction to Monolithic Active Pixel Sensors Some non-HEP and commercial uses (and why they matter). On-going and future HEP MAPS projects and detectors.

Overlapping presentations: PXL at STAR: M. Szelenicak/M. Simko (poster) ALICE ITS upgrade: F. Reidt ATLAS pixels: J. Grosse-Knetter HV-CMOS: D. Muenstermann

Workshop on CMOS Active Pixel Sensors for Particle Tracking (CPIX14), Bonn, 15-19 Sept 3 days, 37 talks I’ll do it all in 25 minutes…

16-Sep-2014 Fergus Wilson, RAL/STFC 2

CMOS Monolithic Active Pixel Sensors

• First invented in the 60’s but CCDs much better then.• Re-invented at the beginning of 90s: JPL, IMEC,

– Standard CMOS technology.– All-in-one detector-connection-readout – Monolithic.– Small size / greater integration.– Low power consumption.– Low noise.– Radiation resistance.– System-level cost.– Increased functionality.– Increased speed.– Increased readout speed (parallel processing).– Region of interest readout.– Etc…


Charged Particle Detection


Deep p-well: enhances charge collection, allows enhanced pixel structures

Thin epitaxial layer: shorter collection times, less multiple scattering, less chance of charge capture

Guard rings: improve resistance to radiation damage.

High-resistivity epitaxial layer: improved signal to noise.

High-resistivity epitaxial layer + low voltage bias (HR-CMOS): charge collection by drift, faster, radiation hardness

High voltage bias (HV-CMOS): charge collection by drift, faster, radiation hardness

Active pixels and In-Pixel electronics


Passive Active

Correlated Double Sampling (CDS), reduced noise

Move as much processing as you can on to the pixel

No need to stop at 4T…

A


Fabrication and Stitching.

C

BD

Reticle size is just over 2cm x 2cm ‘stitching’

Reticle is subdivided in blocks

A

B

AC A

BD B

C

D

C

D

56 mm

56 mm

Beyond Particle Physics

• MAPS have penetrated other science areas more quickly than particle physics.

• Commercially attractive (high yields, low cost).• Many overlaps with particle physics requirements:

– Radiation tolerance - Cost– Small and large pixels - Reliability– High Speed – Quantum efficiency– High dynamic range– Low power

• But particle physics detectors want them all !16-Sep-2014 Fergus Wilson, RAL/STFC 7


Transmission Electron Microscopy (TEM)

Slide taken from D. Contarato, LBNL, 2012


Detection of electrons in CMOS

Single electron detectionGood event

Bad event

Energy contained in one pixel

61x63 mm2 silicon area (4 dies per wafer)

0.35mm CMOS

16 million pixels, 4Kx4K array

14 µm pixels

32 analogue outputs, 10 Mpixs/sec

40 fps

Pixel binning 1X, 2X and 4X

ROI readout

83 e- rms noise

Full well 120ke-

Radiation hardness of >500 million of primary electrons/pixel (>20

Mrad)

20% QE for visible light

Achilles: a 16Mpixel sensor for TEM


www.fei.com

Novo virus

• Motivations– Extra-oral dental, mammography, chest imaging, security,…

• Requirements– High yield (commodity item).– Radiation hard:– Very large sensors:

• Wafer scale sensor.• One sensor per 8”/20cm wafer• 3-side buttable – 2 x N tilling

• Lassena characteristics– 6.7 Mpixels; 30 fps; 50µm pixels; Low noise: 68 e-– Large area: 3-side buttable to cover any length with 28 cm width– Binning x2, x4; Region-of Interest readout– High dynamic range, multiple programmable integration times


Wafer-scale sensor for X-ray medical imaging


Photon Science - Percival

Pixelated Energy Resolving CMOS

Imager,

Versatile and Large

Percival soft x-ray imager


Design goals

Back-thinned

4k x 4k pixels

120 fps (digital CDS)

High dynamic range (4 gains per pixel)

2*105 photons @ 250 eV

~120dB or full well >10 Me-

12+1bit ADC

15 bits per pixel (2 gain bits + 13 bits)

Digital I/O (LVDS)

60 Gbit/sec continuous data rate

Pixel array4kx4k

@25µm pitch)

28,000 ADCs(7 ADCs per column)

Serialiser and LVDS I/OM

ulti-

leve

l row

con

trol

SPI a

nd b

ias

gene

rato

r210x160 25µm pixel prototype under front illumination at DESY

Time-Of-Flight Mass Spectroscopy

Courtesy of A. Nomerotski et al., Oxford University


• Separate chemical species by (mass/charge) ratio and identify where they are in the specimen

• Requirements:• Timing information• Spatial Information

16-Sep-2014

)/

Fergus Wilson, RAL/STFC 15

PImMS family

PImMS1: 72 x 72 pixelsPImMS2: 324 x 324 pixelsPImMS camera

• 70 um x 70 um pixels• 25 ns time resolution (12.5ns has been demonstrated).• Continuous 40 Mfps for 100µs. • 4 events can be stored in each pixel.• 12-bit time-code resolution.• Each pixel can be trimmed.• Analogue readout of intensity information.• Equivalent pixel rate for a standard full frame camera 2 x 1012 pixels/sec

Looks a bit like Linear Collider specs…

Ultra-high speed uCMOS - Kirana

• High resolution: 924 x 768 30µm pixels• Die size 32.5 x 25.5 mm.• In-pixel storage and Correlated Double Sampling

(CDS).• Burst mode: 180 frames at 5 MHz.• Continuous mode: 1180 fps.• Noise: <10e-; full well: 11,700 e-• Commercialised (Specialised Imaging)


Looks a bit like Linear Collider specs…


Performance summary

Parameter Unit ValuePixel pitch (X) um 30Pixel pitch (Y) um 30

Pixel format (X) 924Pixel format (Y) 768Number of pixels 709,632

Frame rate (burst mode) fps 5,000,000Frame rate (continuous mode) fps 1,180

Pixel rate (burst mode) Pixel/sec 1.42 TPixel rate (continuous mode) Pixel/sec 0.84 G

Noise e- rms <10 e- rmsFull well capacity e- 11,700

Camera gain µV/e- 80Dynamic range >1,170

dB 61.4 bit 10.2

Fill Factor 11%

Quantum efficiencyWithout

microlens2.3% (red)2.2% (blue)

MAPS HEP progression

Where is MAPS being proposed?


0.16 m2

1.9 m2

10 m 2

~100? m2

STAR PXL(now)

mu3e(2015)

ALICE ITS(2018)

ATLAS Tracker Phase II?

(2023)

Linear Collider(20??)

Vertexer ?

Tracker ?

Digital Calorimetry ?

STAR PXL at RHIC


Design: LBNL, UT at Austin; PICSEL group, IPHC, StrasbourgSee M Szelezniak talk and M Simko poster.

DCA Pointing resolution (12* 24 GeV/pc) m

Layers Layer 1 at 2.8 cm radiusLayer 2 at 8 cm radius

Pixel size 20.7 m X 20.7 m

Hit resolution 3.7 m* (6 m geometric)

Position stability 6 m rms (20 m envelope)

Radiation length first layer X/X0 = 0.39% (Al conductor cable)

Number of pixels 356 M

Integration time (affects pileup) 185.6 s

Radiation environment 20 to 90 kRad / year2*1011 to 1012 1MeV n eq/cm2

Rapid detector replacement ~ 1 day

PRELIMINARY PRELIMINARY

μ3e at PSI


• µ→eee lepton flavour violation• 109 muon decays/s. Low Pt

tracks, resolution dominated by multiple scattering.

• 4 layers 80x80m2 pixel size, 275 MP

• Thin <50µm. 180nm HV-CMOS.• Fast charge collection by drift.• Power consumption 7.5 µW/pixel

MuPix design: Heidelberg, PSI, Zürich, Genf

3mm

μ3e at PSI: recent DESY test-beam results


• Recent DESY test beam results (MuPix4):

• Timing resolution 18ns• Track residuals: 28µm• Hit efficiency > 99%

ALICE Inner Tracker System Upgrade


See Felix Reidt talk

Many competing/collaborating architectures: MISTRAL/ASTRAL (IPHC), Cherwell (RAL), ALPIDE (CCNU/CERN/INFN/Yonsei)

Also being considered for forward tracker

ATLAS Phase II Tracker

• Challenges– 200 bunches in pile-up, increased particle

densities. (1-2 GHz/cm2)– Increased radiation damage (2 x 1016

neq/cm2)– Increased power requirements.– Reduced material required.

• Pixel+microstrip still the baseline but have ~2-3 years to show that CMOS could be viable technology.– Strips -> elongated pixels.– MAPS with HV-CMOS or HR-CMOS for

radiation hardness and speed.– MAPS not the only candidate: thin planar

silicon, diamond, 3-D detectors…16-Sep-2014 Fergus Wilson, RAL/STFC 23

See Daniel Muenstermann talk

A hybrid MAPS ?

Vertexing and Tracking for Linear Collider

• Pixels are a baseline technology for CLIC/LC vertexing; could become baseline technology for tracking.

• CLIC detector development has been progressing; LC development has been on hold for ~6 years.

• But CLIC and ILC have very different bunch structures.– ILC: 5Hz, 2625 bunches in 1ms followed by 199ms gap.– CLIC: 50Hz, 312 bunches, 0.5ns between bunches, 20ms gap.

• MAPS (Mimosa, Chronopixels, LBL, INFN ...), clixpix, CCD, ISIS, DEPFET, SoI, 3D,…


See S.Redford CLIC, A.Besson ILC

Example of MAPS performance• Cherwell sensor.• 99.7% hit efficiency.• 3.7μm hit resolution.• Power pulsing.

Digital Calorimetry for Linear Collider


An alternative to silicon wafers or scintillators.Results from TPAC chip in CERN test beam.Shows correct behaviour as function of energy.Demonstrates DECAL/MAPS concept validity

1

0

( )

( )

a btdE bt eE b

dt a

TPAC sensor:• 168 x 168 pixels• 50x50μm• Digital readout• Sample every 400ns

T.Price, Birmingham, 2013

• MAPS are already commercially available.• MAPS have already penetrated non-HEP areas

– Medical, photon science, space, X-rays, neutron, lasers,…• In HEP

– Capabilities proven at STAR.– Soon to be used in μ3e vertex detector.– Expect to see used in a tracker in ALICE ITS, Forward Tracker.– Already seeing radiation hardness and speeds (not to mention

power consumption, material thickness, cost, …) that are suitable for LHC phase II upgrades

– MAPS an excellent candidate for LC/ILC vertex detectors and trackers.


Conclusions.

Backup



Kirana pixel. 1

Photodiode

Memory bank- A vertical entry (VEN)

bank with 10 cells- Ten rows of lateral (LAT)

banks, each with 16 cells- A vertical exit (VEX) bank

with 10 cells

- Total of 180 memory cells


Kirana pixel. 2

Highly scalable architecture:- Number of memory cells- Number of pixels


Burst mode

Vertical transfers x10 @ full speed


Burst mode

Charge moved into lateral memory bank


Burst mode

Ten more vertical transfers


Burst mode

Lateral transfer x1 @ full speed / 10


Burst mode

… and so on, seamless


Burst mode



Burst mode



Burst mode



Burst mode



Burst mode



Burst mode



Burst mode


Burst mode

Charge in the vertical exit registers is dumped in the reset node …

… until receipt of the trigger. The status of the memory bank is then frozen and the sensor read out.


Continuous mode

Memory bank acting simply like a delay line


Continuous mode



Continuous mode



Continuous mode



Continuous mode



Continuous mode



Continuous mode



Continuous mode



Continuous mode


Continuous mode


Continuous mode


Continuous mode


Continuous mode


Continuous mode


Continuous mode

Overview of MAPS detectors Fergus Wilson Rutherford Appleton Laboratory (with lots of input and slides from Renato Turchetta and the RAL Sensor Design.

Documents

cmos active pixel sensors

radiation resistance

charge collection

pixel electronics

low noise

radiation damage

readout monolithic

particle physicsmaps