Transcript TWEPP2011_IPHC - Indico

A Reticule Size CMOS Pixel Sensor (ULTIMATE) Dedicated to
the STAR HFT Upgrade
Thanh Hung PHAM
on behalf of the IPHC (Strasbourg) PICSEL group
Outline

STAR HFT Upgrade: PXL detector

CMOS Pixel Sensor requirements

Sensors optimization

Recent ULTIMATE test results


Lab test and Beam test
Summary + Perspectives
Heavy Flavor Tracker (HFT) at STAR

HFT is an upgrade of the inner tracking system of STAR
detector comprising :
 SSD – Silicon Strip Detector
 IST – Inner Silicon Tracker
 PXL – Pixel Detector (2 layers at 2.5 & 8 cm)

Physical Goals : Identification of mid rapidity Charm and
Beauty mesons and baryons through direct reconstruction
and measurement of the displaced vertex with excellent
pointing resolution
TPC
26-30/09/2011
~1 mm
TWEPP 2011
SSD
~300 µm
IST
~ 150 µm
~250 µm
IPHC [email protected]
PXL
<30 µm
vertex
STAR
2
PIXEL SENSORS FOR HFT
Sensors Requirements

Multiple scattering minimisation:
Sensors thinned to 50 µm, mounted on a flex
kapton/aluminum cable
 X/X0 = 0.37% per layer


Sufficient resolution to resolve the secondary
decay vertices from the primary vertex



~200-300 (600) hits / sensor (~4 cm2) in the
integration time window
Short integration time ~< 200 µs
Low mass in the sensitive area of the detector
 airflow based system cooling



carbon fiber sector tubes
(~ 200µm thick)
Luminosity = 8 x 1027 / cm² / s at RHIC_II


< 10 µm
Work at ambient (~ 35 °C ) temperature
Power consumption <~ 150 mW / cm²
Sensors positioned close (2.5 - 8 cm radii) to
the interaction region


~ 150 kRad / year
few 1012 Neq / cm² / year
RDO
buffers/
drivers
MAPS
4-layer kaptonconductor
cable with aluminium
Aluminum
Laddertraces
Flex Cable
Ladder with 10 MAPS sensors (~ 2×2 cm² each)
Insertion from one side
2 layers
10 sectors
4 ladders/sector
Leo Greiner @ St Odile CMOS Workshop, Sep 2011
26-30/09/2011
TWEPP 2011
IPHC [email protected]
STAR
3
Main characteristics of ULTIMATE (Mimosa-28) sensor









0.35 μm process with high-resistivity epitaxial layer
column // architecture with in-pixel CDS & amplification
end-of-column discrimination and binary charge encoding,
followed by zero suppression logic
active area: 960 columns of 928 pixels (19.9×19.2 mm²)
pitch: 20.7 μm ~0.9 million pixels
 charge sharing  >~ σsp 3.5 μm expected
tr.o. ≤ 200 μs ( ~ 5×103 frames/s)
 suited to >106 part./cm²/s
2 outputs at 160 MHz
≤ 150 mW/cm² power consumption
Radiation tolerant (~150 kRad/year & 3x1012 neq/cm²)
26-30/09/2011

Submitted by end-January 2011

Received early April 2011
TWEPP 2011
IPHC [email protected]
STAR
4
Main characteristics of ULTIMATE (Mimosa-28) sensor (suite)
Based on expertise's acquired from
M26 chip for EUDET, the design
of ULTIMATE has been optimized
for STAR environment
8 analog outputs
(Test purpose only)
Pixel Optimization:
Row sequencer

Minimize the delays of
signals over ~2cm
Pixels Ref Regulator &Analog
Power Supply regulator
(Optional)

Reduce I/O pads

Programmable (Ref
Regulator)
Radtol ~ 150 kRad/year &
3x1012Neq/cm²/year

Consumption < 150mW/cm2

Large reticule size
(~2cmx2cm)
End-column 960 discriminators:
Offset compensation

Encoding & Zero
suppression logic:

STAR conditions
Bias current & Ref DACs
High data size & Rate:
I/O Pads: Powers, LVDS &
Controls
JTAG Configuration
26-30/09/2011


2 Memories 2048x32-bits

2 Outputs at 160 MHz
PLL (Optional)
TWEPP 2011
IPHC [email protected]
STAR
5
Pixel optimization
Slct_Row
Radiation Tolerant and Power Consumption

Enclosed layout transistor M4
Tradeoff
between Power Consumption and Radiation
Tolerant -> Optimization of pixels size (20.7x20.7µm²)

~2x2cm2
16 pix

Slct_Row
Slct_Gr
Slct_Row
reticule size
Multiplex pixels output to reduce the capacitance of
output nodes
Slct_Row
Optimization of output buffer stage (transistors M7,
M8 & M9) in order to drive 2cm of metal line
Slct_Row

16 pix

Slct_Gr
R ~ 1.9 KΩ
C ~ 4 pF
Slct_Row
~ 2 cm long!!!
~50µA
RD
~3µA
CALIB
Out group
LATCH
Column-level
Discriminator
M9
Select_Gr
26-30/09/2011
TWEPP 2011
IPHC [email protected]
STAR
6
Digital conception challenges
200 ns

19295µm



250µm
2cm of row controls signal
 Row sequencer logic : Uniformly distributed with dispersion < 500ps
Optimization of zero suppression to cope with STAR environment
 Up to 9 states /row
 Segmented by 15 groups of 64 columns -> Symmetrical distributions of digital controls over ~2cm (at
50MHz)
High density & High speed readout :~0.9 Mpix & < 200µs frame readout

2 memories of 2048x32 bits

2 outputs of 160MHz
 Increase frequency up to 160MHz
Layout constraint : 2261µmx19872µm

The output of SRAM is serialized at 160MHz
Row sequencer
SDS
Memory management
2261 µ m
MUX
SRAM 2048 x 32
Seq
Serializer
26-30/09/2011
TWEPP 2011
1
SRAM 2048 x 32
2
19872 µm
IPHC [email protected]
STAR
7
On-chip Regulators

Internal Pixel References Voltages
RDO
buffers/
drivers
MAPS
4-layer kapton cable with aluminium traces
Vcl
Integrated VCL Regulator 
Ultimate
~2x2cm
²
Ladder of 10 Ultimate sensors using external Vcl

Reduce crosstalk between sensors

Reduce material budget: no extra
decoupling capacitors
 Crosstalk between sensors through Vclp
Serial RC
network
Error amplifier
Buffer
Output stage
Vdda
Mz
c1
cz
Vbias
Msf1
Vbg
Vfb
Mpw
Msf2
c2
Vclamp
RA
RB
26-30/09/2011
TWEPP 2011
IPHC [email protected]
Size
0.0389 mm²
Capacitive
Load
> 5nF
Consumption
< 1mW
Output range
1.9-2.3V
Noise
74.8 nV/√Hz @ 1 kHz
PSRR
52 dB @ 10 kHz
38 dB @ 1 MHz
STAR
8
LAB TEST RESULTS
Analogue output noise (Mode Test):


Conditions:

Pixel array scan at 40MHz

T = 20°C

Nominal JTAG load

Analog Power Supply (VddA) = 3.3V
ENC ~ < 15 e- (On-chip reference regulator)
Gain ~ 65µV/eGood Noise uniformity
The CCE is very little sensitive to Temperature and Analog
Power Supply variations

Charge Collection Efficiency (Mode Test):
CCE
Ultimate
Sensor
Temperature
Calib
peak
(UADC)
ENC
(e-)
~20 °C
395
~35 °C
~45 °C
26-30/09/2011
Calib
peak
(UADC)
Seed
pixel
2x2
pixels
3x3
pixels
5x5
pixels
Analog
Power
Supply
13.8
24%
62%
82%
94%
Vdd_a = 3.3V
395
385
16.4
24%
62%
83%
96%
Vdd_a = 3V
390
369
20.7
23%
63%
85%
99%
TWEPP 2011
IPHC [email protected]
CCE
ENC
(e-)
Seed
pixel
2x2
pixels
3x3
pixels
5x5
pixels
13.8
24%
63%
82%
95%
13.9
24%
62%
83%
95%
STAR
9
Scan of matrices pixel and discriminator
Sub-matrix A
Sub-matrix A
Sub-matrix A
(Column 1-240)
Sub-matrix B
(Column 241- 480)
Sub-matrix C
Column 481 – 720)
Sub-matrix D
Column 721 -960
Temporal Noise (mV)
1
0.95
0.92
0.9
Fixed Pattern Noise (mV)
0.57
0.49
0.48
0.47
26-30/09/2011
TWEPP 2011
IPHC [email protected]
STAR
10
Beam Test Results (July 2011)

Conditions:



CERN-SPS 120 GeV π− beam
BT made of 6 Ultimate sensors(20µm thick epi)

T = 20°C & 30°C

ionizing radiation dose: 0&150 kRad

Analog power supply : 3V & 3.3V
Results:

Efficiency > 99.5 % with a fake hit rate << 10-4

Spatial resolution < 4 µm
26-30/09/2011
TWEPP 2011
IPHC [email protected]
STAR
11
Conclusion & perspectives


ULTIMATE sensors for the PXL detector of STAR HFT experiment has
been designed and tested in 2011
The lab test and beam test showed :
 Robust regarding temperature variations
 Operational with analog power supply down to 3V
 High detection efficiency ( > 99%) with very low fake event of beam test
 High yield : > 90%
 12 sensors fully functional
 4 with 1% of death pixels

26-30/09/2011
The Ultimate sensor fulfils all STAR HFT specifications

Engineering run of ULTIMATE sensor (12 wafers) is being submitted in
September for equipping the engineering prototype detector
 Start of run at RHIC in FY 2012

The ULTIMATE sensor development allows to accumulate expertises for
future sensor designs (ALICE, AIDA, CBM, EIC, …)
TWEPP 2011
IPHC [email protected]
STAR
12
BACKUP SLIDES
26-30/09/2011
TWEPP 2011
IPHC [email protected]
STAR
13
MIMOSA26 with high resistivity EPI layer (1)

Charge collection efficiency for the seed pixel, and for 2x2 and 3x3 pixel clusters
Standard (~10 .cm) 14 µm
EPI layer
Seed
CCE
EPI
layer
source)
2x2
High resistivity (~400 .cm)
3x3
Standard (~10 .cm) 14 µm
~21%
~ 2x2
54 %
~ 3x3
71 %
Seed
(55Fe
EPI
thickness
seed

15EPI
µm
20 µm
10 µm
~ 22 %
~ 36 %
~ 57 %
~ 85 %
~ 76 %
~ 95 %
(a)
~21%
~ 54 %
~ 71 %
15 µm
~ 31 %
~ 78 %
~ 91 %
Standard
.cm) pixel
14 µm before20irradiation
High
resistivity
(~400
µm
~ 22
% after~exposure
57 % .cm)~ 76 %
noise ratio
for (~10
the seed
and
Signal to
EPI layer
to
a fluence of
6 xirradiation
1012 neq / After
cm² 6x1012 n(a)
Before
eq/cm²
S/N at seed
EPI
layer
pixel
(106Ru source)
S/N at seed
pixel
(106Ru source)
Standard (~10 .cm) 14 µm
~ 20
10.7
-/11.6 e-)
Before
After 6x1012 neq/cm²
(230 eirradiation
~ 20
(230 e-/11.6 e-)
Christine HU@TWEPP 2010
26-30/09/2011
3x3
~ 36resistivity
%
~ (~400
85 % .cm)~ 95 %
High
~seed
31 %
~ 2x2
78 %
~ 3x3
91 %
10 µm
thickness
CCE
55
( Fe source)
2x2
TWEPP 2011
(b)
10.7
EPI
thick
Before irradiation
10 µm
15EPI
µm
thick
After 6x1012 neq/cm²
35
22
High~ resistivity
(~400 .cm)
~ 41
2812
Before irradiation
After 6x10
neq/cm²
20 µm
10 µm
~ 36
~ 35
-------22
15 µm
~ 41
28
20 µm
~ 36
--------
(b)
IPHC [email protected]
STAR
14