Transcript Diapositive 1 - Institut Pluridisciplinaire Hubert Curien

irfu
saclay
Achievements & Perspectives of MIMOSA Sensors
(MAPS) for Vertexing Applications
Christine Hu-Guo (IPHC)
on behalf of IPHC (Strasbourg) & IRFU (Saclay) collaboration
Outline

Achieved MAPS (Monolithic Active Pixel Sensors) performances

R&D for improving MAPS performances

Increase readout speed  Fast readout architecture


Applications and perspectives

Improve radiation tolerance

Projection beyond present (2D sensors) & perspectives
Conclusion
Development of MAPS for Charged Particle Tracking

In 1999, the IPHC CMOS sensor group proposed the first CMOS pixel sensor (MAPS)
for future vertex detectors (ILC)



Numerous other applications of MAPS have emerged since then
~10-15 HEP groups in the USA & Europe are presently active in MAPS R&D
Original aspect: integrate sensitive volume (EPI layer) and front-end readout
electronics on the same substrate

Charge created in EPI, excess carries propagate
thermally, collected by NWELL/PEPI , with help of reflection
on boundaries with P-well and substrate (high doping)



Compact, flexible
EPI layer ~10–15 µm thick



thinning to ~30–40 µm permitted
Standard CMOS fabrication technology


Q = 80 e-h / µm  signal < 1000 e-
Cheap, fast multi-project run turn-around
Room temperature operation
R.T.
Attractive balance between granularity, material budget, radiation tolerance, read
out speed and power dissipation

BUT




14-18/09/2009
Very thin sensitive volume  impacts signal magnitude (mV!)
Sensitive volume almost un-depleted  impacts radiation tolerance & speed
Commercial fabrication (parameters)  impacts sensing performances & radiation tolerance
NWELL used for charge collection  restricts use of PMOS transistors
Vertex-2009
IRFU - IPHC [email protected]
2
Achieved Performances with Analogue Readout

MAPS provide excellent tracking performances

Detection efficiency ~100%


ENC ~10-15 eS/N > 20-30 (MPV) at room temperature
MIMOTEL

Single point resolution ~ µm, a function of pixel pitch

~ 1 µm (10 µm pitch), ~ 3 µm (40 µm pitch)
M18
MAPS: Final chips:

MIMOTEL (2006): ~66 mm², 65k pixels, 30 µm pitch
MIMOSTAR
EUDET Beam Telescope (BT) demonstrator

MIMOSA18 (2006): ~37 mm², 262k pixels, 10 µm pitch
Chip dimension: ~2 cm²
High resolution EUDET BT demonstrator

MIMOSTAR (2006): ~2 cm², 204k pixels, 30 µm pitch

LUSIPHER (2007): ~40 mm², 320k pixels, 10 µm pitch
Test sensor for STAR Vx detector upgrade
LUSIPHER
Electron-Bombarded CMOS for photon and radiation imaging detectors
14-18/09/2009
Vertex-2009
IRFU - IPHC [email protected]
3
Radiation tolerance (preliminary)

Ionising radiation tolerance:

O(1 M Rad) (MIMOSA15, test cond. 5 GeV e-, T = -20°C, tint~180 µs)
Integ. Dose
0
1 Mrad


Noise
9.0 ± 1.1
10.7 ± 0.9
S/N (MPV)
27.8 ± 0.5
19.5 ± 0.2
Detection Efficiency
100 %
99.96 % ± 0.04 %
tint << 1 ms, crucial at room temperature
Non ionising radiation tolerance: depends on pixel pitch:

20 µm pitch: 2x1012 neq /cm2 , (Mimosa15, tested on DESY e- beams, T = - 20°C, tint ~700 μs)
Fluence (1012neq/cm²)
S/N (MPV)
Det. Efficiency (%)


2.1
14.7 ± 0.3
99.3 ± 0.2
5.8 (5/2)
8.7 ± 2.
77. ± 2.
5.8 (4/2)
7.5 ± 2.
84. ± 2.
10 µm pitch: 1013 neq /cm2 , (MIMOSA18, tested at CERN-SPS , T = - 20°C, tint ~ 3 ms)

0
1026
28.5 ± 0.2
99.93 ± 0.03
6
680
20.4 ± 0.2
99.85 ± 0.05
10
560
14.7 ± 0.2
99.5 ± 0.1
parasitic 1–2 kGy gas  N ↑
Further studies needed :

14-18/09/2009
0.47
21.8 ± 0.5
99.9 ± 0.1
5.8·1012neq/cm² values derived with standard and with soft cuts
Fluence (1012neq/cm²)
Q cluster (e-)
S/N (MPV)
Det. Efficiency (%)

0
27.8 ± 0.5
100.
Tolerance vs diode size, Readout speed, Digital output, ... , Annealing ??
Vertex-2009
IRFU - IPHC [email protected]
4
System integration

Industrial thinning (via STAR collaboration at LBNL)

~50 µm, expected to ~30-40 µm


Ex. MIMOSA18 (5.5×5.5 mm² thinned to 50 μm)
Development of ladder equipped with MIMOSA chips (coll. with LBNL)

STAR ladder (~< 0.3 % X0 )  ILC (<0.2 % X0 )
LVDS drivers
PIXEL Ladder
10 MAPS Detectors
low mass / stiffness
cables
to motherboard
40 LVDS Sensor output pairs
clock, control, JTAG, power,
ground.
% radiation length
MIMOSA detector
0.0534
Adhesive
0.0143
Cable assembly
0.090
Adhesive
0.0143
CF / RVC carrier
0.11
0.282
Total

Edgeless dicing / stitching  alleviate material budget of flex cable
14-18/09/2009
18-21/05/2009
Vertex-2009
FEE09
IRFU - IPHC [email protected]
IRFU - IPHC [email protected]
5
MAPS performance Improvement
R&D organisation : 4 (5) simultaneous prototyping lines
MIMOSA22
Architecture of pixel array organised
Pixel array
136 x 576
pitch 18.4 µm



128
discriminators

in //
columns read out:
Pre-amp and CDS in each pixel
A/D: 1 discriminator / column (offset
compensation)
Power vs Speed
 Power  Readout in a rolling shutter
mode
 Speed  Pixels belonging to the same
row are read out simultaneously
MIMOSA8 (2004), MIMOSA16 (2006),
MIMOSA22 (2007/08)
Zero suppression circuit:
SUZE-01



Reduce the raw data flow of MAPS
Data compression factor ranging from 10
to 1000, depending on the hit density per
frame
SUZE-01 (2007)
Pixel Array
Analogue processing / pixel
A/D: 1 ADC ending each column
Zero suppression
Bias DC-DC Data transmission
4–5 bits ADCs (~103 ADC per sensor)

5-bit ADC


Serial link transmission with clock recovery

14-18/09/2009
Prototype (2008-2009)
Vertex-2009
Potentially replacing column-level
discriminators
5 bits: sp ~1.7–1.6 µm
4 bits: sp < 2 µm for 20 µm pitch
Next step: integrate column-level ADC
with pixel array
Voltage regulator & DC-DC converter
IRFU - IPHC [email protected]
6
MIMOSA22 + SUZE-01 Test Results

MIMOSA22: (15 µm EPI) 136 x 576 pixels + 128 column-level discriminators

Laboratory test:
 Temporal Noise: 0.64 mV  12 e FPN:
0.22 mV  4 e-
0.64 mV


Beam test at CERN SPS (120 GeV pions)
 Threshold ~ 4 mV  6 x σ noise
 Detection Efficiency > 99.5%
 Single point resolution < 4 µm
 Fake rate < 10-4
0.22 mV
SUZE-01:
Lab. test :
 Design performances reproduced up to 1.15 × design read-out frequency (115 MHz at room Temp ):



No pattern encoding error, can handle > 100 hits/frame at rate ~200 ns per pixel row
Still to do : improve radiation tolerance (SEU, SEL) of digital circuits (including memories)
14-18/09/2009
Vertex-2009
IRFU - IPHC [email protected]
7
MIMOSA26: 1st MAPS with Integrated Ø
CMOS 0.35 µm OPTO technology, Chip size : 13.7 x 21.5 mm2

Integration time: ~ 100 µs
 R.O. speed: 10 k frames/s
 Hit density: ~ 106 particles/cm²/s



Testability: several test points
implemented all along readout
path
 Pixels out (analogue)
 Discriminators
 Zero suppression
 Signal transmission




Pixel array: 576 x 1152, pitch: 18.4 µm
Active area: ~10.6 x 21.2 mm2
In each pixel:
 Amplification
 CDS (Correlated Double Sampling)
Row sequencer
Width: ~350 µm
1152 column-level discriminators
 offset compensated high
gain preamplifier followed
by latch

Zero suppression logic

Reference Voltages
Buffering for 1152
discriminators

I/O Pads
Power supply Pads
Circuit control Pads
LVDS Tx & Rx


Current Ref.
Bias DACs


14-18/09/2009
Vertex-2009
Readout
controller
JTAG controller


Memory
management
Memory IP blocks
IRFU - IPHC [email protected]

PLL, 8b/10b
optional
8
Test MIMOSA26 (Lab. + beam test)

Measured temporal noise = 0.6-0.7 mV and FPN = 0.3-0.4 mV for pixel array with its
associated discriminators.


These values are equivalent to those obtained with Mimosa22.
It shows a good uniformity of the whole 576 x 1152 pixels with the 1152 discriminators
Entries
576 x 288
Entries
576 x 288
Mean
0.64 mV
Mean
-0.93 mV
RMS
0.07 mV
RMS
0.29 mV
Noise distribution [mV]
Threshold distribution [mV]

~ 30 MIMOSA26 chips are tested (only 1 "dead")

The characterization of Mimosa26 is complemented by the beam tests (Sept. 2009)



6 MIMOSA26 chips are running simultaneously at nominal speed
Tracking successful
data analysis is underway, preliminary results show similar performances as MIMOSA22
14-18/09/2009
Vertex-2009
IRFU - IPHC [email protected]
9
MIMOSA26 = Final Sensor for EUDET Beam Telescope

EUDET supported by the European Union in the 6th Framework Programme

Provide to the scientific community an infrastructure aiming to support the detector R&D
for the ILC
 JRA1 (Joint Research Activity): High resolution pixel beam telescope (BT)


Two arms each equipped with three layers of pixel sensors (MIMOSA)
DUT is located between these arms and moveable via X-Y table
MIMOSA26
13.7 mm
Pixel Sensor
y
z
x

21.5 mm
(DUT)
Being Mounted on EUDET beam telescope
EUDET beam telescope specifications:




14-18/09/2009
High extrapolated resolution < 2 µm
Large sensor area ~ 2 cm2
High read-out speed ~ 10 k frame/s
Hit density: up to 106 hits/s/cm2
Vertex-2009
Preliminary test results from the EUDET
collaboration:


3 MIMOSA26 chips mounted as DUT in BT
demonstrator in July
BT tracks reconstructed in the 3 planes
 residues compatible with σsp ~ 3.5-4 μm
IRFU - IPHC [email protected]
10
Extension of MIMOSA-26 to STAR

Final HFT (Heavy Flavour Tracker) - PIXEL sensor :

MIMOSA-26 with active surface × ~1.7

1088 col. of 1024 pixels  1.1 million pixels
Pitch : 18.4 μm  (~20.0 x 18.8 mm²)
 Integration time 200 μs
 Design from now  fab. Feb. 2010
 1st physics data expected in 2011/12

PIXEL at 2.5 and 8 cm

Critical points:
Reduction of power consumption
 Radiation tolerance improvement
Also:
 Integrated voltage reference
 High speed transmission

Pixel Vx Detector
~22.4 mm
IST at 14 cm
Inner Layer:
10 ladders
SSD at 23 cm
STAR Detector Upgrade
14-18/09/2009
Vertex-2009
Outer Layer: 30 ladders
10 sensors / ladder
IRFU - IPHC [email protected]
~20.4 mm
11
Extension of MIMOSA-26 to CBM/FAIR & ILC

Micro Vertex Detector (MVD) of the CBM H.I. fixed target expt :

2 double-sided stations equipped with MIMOSA sensors
 MIMOSA-26 with double-sided read-out  readout speed !
 Active surface : 2 x 1152 columns of 256 pixels


tint. ~ 40 μs



21.2 x 9.4 mm2
< 25 μs in < 0.18 μm techno.
Prototyping until 2012  start of physics in 2013/14 (?)
Vertex detector of the ILC:
ILD design: 2 options
tint. ~ 25 μs (innermost layer)  double-sided readout
 tint. ~ 100 μs (outer layer)  Single-sided readout
 2 μm (4-bit ADC, 20 µm) < sp < 3 μm (discri. 14 µm pitch)
 Pdiss < 0.1–1 W/cm² × 1/50 duty cycle


Critical points:


Power pulsing
Design for the innermost layer:


14-18/09/2009
Small pixel pitch
Faster readout speed
Vertex-2009
5 single layers
IRFU - IPHC [email protected]
3 double layers
12
Radiation Tolerance Improvement

Ionising radiation tolerance
1. Special layout

Diode level (already realized): Remove thick oxide surrounding N-well diode by
replacing with thin-oxide without design rule violation!!
Vdd!

Pixel circuit level: ELT for the transistors
connected to the detection diode
PWR_On
M4
M6
V_clp
M3
Clp
M9
M5
M10
M7
2. Minimise integration time
 Increase readout speed:
Vertex-2009
Slct_Row
M11
Slct_Gr
M12
M2
Nwell/Pepi
M1
Gr_node
Pix_Out
Minimise leakage current
14-18/09/2009
M8
Pdiff/Nwell
Gnd!
IRFU - IPHC [email protected]
13
Radiation Tolerance Improvement

Non ionising radiation tolerance
High resistivity sensitive volume  faster charge collection

Exploration of a VDSM technology with depleted (partially ~30 µm) substrate:


Project "LePix" driven by CERN for SLHC trackers (attractive for CBM, ILC and CLIC Vx Det.)
Exploration of a technology with high resistivity thin epitaxial layer

XFAB 0.6 µm techno: ~15 µm EPI ( ~ O(103).cm), Vdd = 5 V (MIMOSA25)
 Benefit from the need of industry for improvement of the photo-sensing
elements embedded into CMOS chip
TCAD Simulation
15 µm high resistivity EPI compared to 15 µm standard EPI
For comparison: standard CMOS
technology, low resistivity P-epi
high resistivity P-epi: size of depletion
zone size is comparable to the P-epi
thickness!
14-18/09/2009
Vertex-2009
IRFU - IPHC [email protected]
14
MIMOSA25 in a high resistivity epitaxial layer
Landau MP (in electrons) versus cluster size
0 neq/cm²
0.3 x 1013 neq/cm²
1.3 x 1013 neq/cm²
3
MIMOSA25
saturation -> >90 % of charge is collected
is 3 pixels -> very low charge spread for
depleted substrate
16x96
Pitch 20µm




x 1013 neq/cm²
To compare: «standard» non-depleted EPI
substrate: MIMOSA15 Pitch=20µm, before and
after 5.8x1012 neq/cm2
20 μm pitch, + 20°C, self-bias diode @ 4.5 V, 160 μs read-out time
Fluence ~ (0.3 / 1.3 / 3·)1013 neq/cm2
Tolerance improved by > 1 order of mag.
Need to confirm det (uniformity !) with beam tests
14-18/09/2009
Vertex-2009
IRFU - IPHC [email protected]
15
Using 3DIT to improve MAPS performances
3DIT are expected to be particularly beneficial for MAPS :



Split signal collection and processing functionalities, use best suited technology for
each Tier :





Combine different fabrication processes
Resorb most limitations specific to 2D MAPS
Tier-1: charge collection system  Epitaxy (depleted or not), deep N-well ?  ultra thin layer X0 
Tier-2: analogue signal processing analogue, low Ileak, process (number of metal layers)
Tier-3: mixed and digital signal processing
Tier-4: data formatting (electro-optical conversion ?)

130 nm, 2-Tier run with "high"-res substrate
(allows m.i.p. detection)






3D - MAPS
Pixel Controller,
A/D conversion
Digital
Tier A to tier B bond Cu-Cu bond
3 D consortium: coordinated by FermiLab

digital process (number of metal layers)
feature size fast laser driver, etc.
2D - MAPS
Run in Chartered - Tezzaron technology
FermiLab
INFN
IN2P3-IRFU
Univ. of Bonn
CMP
Pixel Controller, CDS

Diode
Diode
Analog
~ 50 µm
Analog Readout Analog Readout
Circuit
Circuit
Sensor
Diode
Diode
Analog Readout Analog Readout
Circuit
Circuit
TSV: Through Silicon Vias
~ 20 µm
14-18/09/2009
Vertex-2009
IRFU - IPHC [email protected]
16
IPHC & IRFU 3D MAPS

Delayed R.O. Architecture for the ILC Vertex Detector

Try 3D architecture based on small pixel pitch, motivated by :




Single point resolution < 3 μm with binary output
Probability of > 1 hit per train << 10 %
12 μm pitch :
•
sp ~ 2.5 μm
•
Probability of > 1 hit/train < 5 %
Acquisition
Readout
~1 ms
~200 ms
~1 ms
Split signal collection and processing functionalities :


Tier-1: A: sensing diode & amplifier, B: shaper & discriminator
Tier-2: time stamp (5 bits) + overflow bit & delayed readout
 Architecture prepares for 3-Tier perspectives : 12 µm



Tier-1: CMOS process adapted to charge collection
Tier-2: CMOS process adapted to analogue & mixed signal processing
Tier-3: digital process (<< 100 nm ????)
12 µm
A
Tier 1
Tier 2
B
12 µm
TS & R.O.
Detection diode
or Q injection
Amplifier
Amp.+Shaper
Discriminator
Hit identification
+
5 bits (7?) Time Stamp
2nd hit flag
ReadOut
24 µm
14-18/09/2009
Vertex-2009
IRFU - IPHC [email protected]
ASD
Detection diode
& Amp
17
IPHC 3D MAPS: Self Triggering Pixel Strip-like Tracker (STriPSeT)

Combine Tezzaron/Chartered 2-tiers process with XFAB high resistivity EPI process
Tier-1
Tier-2
Cf~10fF
Tier-3
off <10 mV
G~1
Digital RD
Cc=100fF
Vth
Cd~10fF
Ziptronix
(Direct Bond Interconnect, DBI®*)

Tezzaron
(metal-metal (Cu) thermocompression)
DBI® – Direct Bond Interconnect, low temperature CMOS compatible direct oxide bonding with
scalable interconnect for highest density 3D interconnections (< 1 µm Pitch, > 10 8/cm /cm² Possible)

Tier-1: XFAB, 15 µm depleted epitaxy  ultra thin sensor!!!





Tier-2: Shaperless front-end: (Pavia + Bergamo)



Single stage, high gain, folded cascode based charge amplifier, with a current source in the feedback loop
 Shaping time of ~200 ns very convenient: good time resolution
Low offset, continuous discriminator
Tier-3: Digital: Data driven (self-triggering), sparsified binary readout, X and Y projection of
hit pixels pattern

14-18/09/2009
Fully depleted  Fast charge collection (~5ns)  should be radiation tolerant
For small pitch, charge contained in less than two pixels
Sufficient (rather good) S/N ratio defined by the first stage
“charge amplification” ( >x10) by capacitive coupling to the second stage
Matrix 256x256  2 µs readout time
Vertex-2009
IRFU - IPHC [email protected]
18
IRFU & IPHC 3D MAPS: RSBPix

FAST R.O. architecture aiming to minimise power consumption
Subdivide sensitive area in ”small” matrices
running INDIVIDUALLY in rolling shutter mode
 Adapt the number of raws to required frame r.o. time
 few µs r.o. time may be reached (???)

Vrst
Vclp1+Vth
CS
Av
~4
Vclp2
Vclp2
CS
Clamp0
Clamp
Vclp2
CS
Clamp1
Clamp4
Digital Memory
Latch
MOSCAP
(100fF)
PWRON_A
MOSCAP
MOSCAP
(20fF)
PWRON_D
PWRON_D
Digital Readout
Vclp2
Track Latch
Discriminator
Tier-1

14-18/09/2009
and
LATCH_D
DREAD
Tier-2
Planned also to connect this 2 tier circuit to XFAB detector tier
Vertex-2009
IRFU - IPHC [email protected]
19
Conclusion

2D MAPS have reached necessary prototyping maturity for real scale
applications :
Beam telescopes allowing for sp ~ few μm & 106 particles/cm²/s
 Vertex detectors requiring high resolution & very low material budget


The emergence of fabrication processes with depleted epitaxy / substrate
opens the door to :

Substantial improvements in read-out speed and non-ionising radiation tolerance
 "Large pitch" applications  trackers (e.g. Super LHC )

Translation to 3D integration technology :

Resorb most limitations specific to 2D MAPS

T type & density, peripheral insensitive zone, combination of different CMOS processes
Offer an improved read-out speed : O(μs) !
 Many difficulties to overcome (ex. heat, power)
 R&D in progress  2009/10 important step for validation of this promising
technology

14-18/09/2009
Vertex-2009
IRFU - IPHC [email protected]
20