C hallenges with decay vertex detection in CBM using an ultra-thin pixel detector system

M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN 1

Challenges with decay vertex detection in CBM using an ultra-thin pixel detector system

Outline:• The CBM experiment (a reminder)• Requirements for the CBM vertex detector• Technology choices and consequences• Running conditions• Summary and Conclusion

M. Deveaux, S. Amar-Youcef, C. Dritsa, J. Heuser, I. Fröhlich, C. Müntz, C. Schrader, F. Rami, S. Seddiki, J. Stroth, T. Tischler, C. Trageser and B. Wiedemann on behalf the CBM collaboration. :


RICHRICH

TRDTRD

TOFTOF

ECALECAL

MVD+STS

Beamand

Target

CBM@FAIR

CBM:Fixed target heavy ion experiment10-45 AGeV beam energyLocated at GSI/FAIR, Darmstadt, GermanyFirst collision ~2015


CBM

The Physics case of CBM

Explore the nuclear phase diagram in the region of high net baryon density


Observables of interest

Onset of chiral symmetry restoration at high B

• in-medium modifications of hadrons (,, e+e-(μ+μ-), D)

Deconfinement phase transition at high B • excitation function and flow of strangeness (K, , , , )• excitation function and flow of charm (J/ψ, ψ', D0, D, c)• charmonium suppression, sequential for J/ψ and ψ' ?

The equation-of-state at high B

• collective flow of identified hadrons• particle production at threshold energies (open charm)

QCD critical endpoint• excitation function of event-by-event fluctuations (K/π,...)

flow

charm

fluctuations

dileptons

Our topic for today


Global layout of the CBM experiment

MVD: Micro Vertex Detector, STS: Main Silicon Tracking SystemRICH + TRD: electron identification, TOF: Time of Flight system for hadron identification, MUCH Muon detector (replacing the RICH), ECAL

• electron ID: RICH & TRD suppression 104

• muon ID: absorber + detector layer sandwich move out absorbers for hadron runs

MVD + STS

aim: optimize setup to include both, electron and muon ID

→ would be the best crosscheck of results we can ever do!


[W. Cassing, E. Bratkovskaya, A. Sibirtsev, Nucl. Phys. A 691 (2001) 745]

SIS18 SIS100/ 300

The open charm challenge

D : c ~ 312 m

D+ K- + + (9.5%)

D0 : c ~ 123 m

D0 K- + (3.8%)

D0 K- + + - (7.5%)

Ds : c ~ 150 m

D+s K+ K- + (5.3%)

High interaction rate required => CBM is designed as fixed target experimentBeam intensity provided by SIS300: Up to 109 ions/s (slow extraction)Needs excellent secondary vertex resolution ~50 µm

Multiplicities / central collision

+c : c ~ 60 m

+c p K- + (5.0%)


Can’t tolerate material of the beam pipe. Vertex detector must be placed in (moderate) vacuum.

How to get the resolution

Sec. vtx. resolution demands for thin sensors like CMOS-Sensors (MAPS)

Hybrid pixels

CMOS-Sensors

Assuming first station 5cm behind the target.

Simulation of the sec. vtx resolution of the MVD

C. Dritsa


Minimum Ionizing Particle MOS Active Pixel Sensor

Features of the MIMOSA – detectors:

• Single point resolution 1.5µm - 2.5µm

• Pixel – pitch 10 - 40 µm

• Thinning achieved 50 - 120µm

• S/N for MIPs 20 – 40

• Radiation hardness: 1MRad ; 1 x 1013 neq/cm²

• Time resolution ~ 20 µs (massive parallel readout)MIMOSA IV

We follow the R&D for SoI-Detectors and 3D-VLSI-detectors

CMOS-sensors (MAPS)


The time resolution of MAPS vs. collision rate

Sensor array

Discriminators On - chip zero suppression

Bonding pads + output Amplis.

Pixel column

Readout busConsecutive readout of pixels required=> expected time resolution ~10 µs (~100 kFrame/s)

Expected collision rates:

Up to 109 heavy ions / sec (up to uranium @ 35 AGeV)1% - target => Up to 107 collisions / s

Expect pile up of nuclear collisions in the vertex detector


Running conditions (1)H

its/m

m²/

coll.

x 1

000

X (cm)

Y (cm)

~130 hits / evt

Typical non-centralcollision

Average hit density from nuclear collisions (@ 10cm)


Running conditions (1)

Mean hit density from nuclear collisions (@ 10cm)

Incl. delta electrons from the target

~700 hits / evt

Hot spots up to 1 hit / evt / mm2

Hits

/mm

²/co

ll. x

100

0

100 kHz => 0.7k (1/mm²) 1 MHz => 7k (10/mm²) 10 MHz => 70k (100/mm²)

Adapt beam to abilitiesof the detector!

Rate Hits (Max.)


The occupancy is dominated by delta electrons generated in the target.Handling them needs detector with very high granularity

Running conditionsH

its/e

vent

distance between station and target[cm]


Occupancy estimate:

Running conditions and consequences

IPHC-StrasbourgCEA-Saclay

3.5 mV

7.8 mV

A. Besson (modified)

Fired pixels / hit(from Mimosa16 beam test)

Pixel pitch: ~ 15 µm => 4.4k pixel /mm²

Assume collision rate of ~ 1MHz for thefirst years + 5 fired pixels / hit:

1.2 % occupancy @ 10cm 4.0 % occupancy @ 5cm

Data flow estimate:

Assume on-chip zero suppression:

40-80 bit/cluster (including addresses)1 MHz collision rate

~ 5-10 GB/s @ 10 cm ~ 8-16 GB/s @ 5 cm

Tracking?• Start in fast detectors of the STS,• Extrapolate tracks to the MVD.Validated excluding Delta-electronsTo be confirmed accounting for them


Mi-9

Mi-9

Mi-15

Mi-18

Assumption today:Detector may tolerate 1013 neq/cm²=> two weeks beam on target (1 MHz)~ 5 x 1011 coll. @ 5 cm~ 1 x 1012 coll. @ 10 cm=> Replace detector after one beam time

Radiation dose and tolerance

z=5 cm

z=10 cm

Radiation tolerance of MAPSFull collision rate => few 1015 neq/year


How it might look like

S. Belogurov et. al

Artistic view on the CBM-MVD

Power: ~ 1W /cm²

Cool with heat conduction: TPG - ~1400 W/(m K), 3-4 times better than Cu


Material budget?

Acceptablearea

Assuming 15 µm pitch

Prelim

inary

More speed needs more power => More material for cooling & cablesLarger diameter needs more pixels => Even more power & cablesLarger diameter needs longer ways for heat conduction => More material

Time resolution might need fine tuning accounting for material budget


Some simulation results (25 AGeV)

We expect ~ 5k (up to few 10k) reconstructed D+ within 2 weeks of measurement (1012 coll. at 106 coll./s, 25 AGeV). λC can be seen.

Expect better results for 35 AGeV (higher multiplicities)

But: Material budget of the detector must not exceed ~0.3% X0 per layer

M ~ 4.17 10-5

BR = 9.51%

ε D+ = 1% and S/B2σ

= 0.4

Multiplicity based

on HSD model


Rad. hardness studies:Frankfurt, IPHC, FRM II

Research lines for the MVD

MAPS R&D@ IPHC

Electronic system integration R&D@ Frankfurt

Design studies

Mechanical system integration R&D@ Frankfurt

mm

Do

se [

n eq /

cm2 / c

oll. ]

mm

BE

AM

Delta electrons produced in the target

Simulation studies @ GSI + IPHC

The CBMMicro Vertex Detector


Summary and conclusion

The CBM experiment aims for a first time to measure open charm being emitted from heavy ion collisions at beam energies between 10 and 45 AGeV

This requires a vertex detector with:• very low material budget (~0.3 % X0 /layer)• good radiation hardness (minimum 1013 neq/cm², optimal > 1015 neq/cm²)• good time resolution (minimum ~10µs, optimal < 100 ns)• good spatial resolution (~ 5 µm or better)• vacuum compatible design and cooling

Currently we consider CMOS sensors (by IPHC Strasbourg) as guideline technology Excessive R&D seems required to match the requirements.

R&D on radiation hardness of CMOS sensors, mechanical and electronical detector integration have started at GSI (Darmstadt) and Goethe University Frankfurt

We will presumably build our detector in two phases:• A CMOS sensor based vtx detector operating roughly 2015 (min. performance)• A novel technology based detector operating few years later

Challenges in vertex detector desig

n?

We have plenty of them

We are open to alternative technologies, advices and collaborators.


The CBM collaboration

Russia:IHEP ProtvinoINR TroitzkITEP MoscowKRI, St. Petersburg

China:CCNU WuhanUSTC Hefei

Croatia: RBI, Zagreb

Portugal: LIP Coimbra

Romania: NIPNE Bucharest

Poland:Krakow Univ.Warsaw Univ.Silesia Univ. KatowiceNucl. Phys. Inst. Krakow

LIT, JINR DubnaMEPHI MoscowObninsk State Univ.PNPI GatchinaSINP, Moscow State Univ. St. Petersburg Polytec. U.

Ukraine: Shevchenko Univ. , Kiev

Cyprus: Nikosia Univ.

Univ. Mannheim Univ. MünsterFZ RossendorfGSI Darmstadt

Czech Republic:CAS, RezTechn. Univ. Prague

France: IPHC Strasbourg

Germany: Univ. Heidelberg, Phys. Inst.Univ. HD, Kirchhoff Inst. Univ. FrankfurtUniv. Kaiserslautern

Hungaria:KFKI BudapestEötvös Univ. Budapest

India:VECC KolkataSAHA KolkataIOP BhubaneswarUniv. ChandigarhUniv. VaranasiIlT Kharagpur

Korea:Korea Univ. SeoulPusan National Univ.

Norway:Univ. Bergen

Kurchatov Inst. MoscowLHE, JINR DubnaLPP, JINR Dubna

46 institutions

> 400 members

Strasbourg, September 2006


Backup


First vacuum compatibility tests

High-tech prototype ladder for vacuum tests (silicon glued on TPG)


Radiation hardness of MAPS


P+ P- P+

Outlook Radiation hardness

Non-ionizing radiation hardness:

Limitation: Slow charge collection in the sensor + reduced lifetime of signal electrons

Approach (1): Speed up charge collection:

Diode

P+ P- P+P+ => P-

Pixel with gradient doping.Attractive build in potential?

Diode


Outlook Radiation hardness

Approach (2): Restore lifetime of electrons

Recombination destroys the signal

Efficient approach for depleted N-doped detectors. BUT: MAPS are P-doped and undepleted.


The power dissipation of MAPS

Sensor array

Discriminators On - chip cluster-finding processor

Bonding pads + output Amplis.

Pixel column

Readout bus

Only one pixel / col. ac-tive, low power dissipation

Permanently biased,high consumption.1 cell : „Pcol“

Power scales withoccupency. Neglegted in the follwing

Basic assumption: Power scales with number of the end of column logics (discriminator + cluster finding processor)


The power dissipation of MAPS

pN

2pNS ftN Int

p

p

The surface of one collumn is:

With:

f pixels / s

²/1~²²

cmWpft

P

cm

P

Int

Col

Time resolution of MAPSReadout freq. (Pixels / time)Power of one logic cellmWP

MHzf

µst

Col

Int

5.0

105

10

pixels


The material buget of the MVD

pft

DLD

L

L

dInt

cable

cable

Sensorcable

²

/1~²

²

cmW

pft

PcmP

Int

Col

µm

pft

L

d

µm

LL

d

Int

Blind

Si

SensorBlind

Si 50

50

µm

TTL

d cool

500

100

)(

²5.0

21

otherschip

blindcable

ladderColladder

Intladder

dDLDLTTp

PL

pftd

)(

5.0121

2

Details: See talk of M.Deveaux, MVD-working group


Correlations of the thickness

322

1

1CLCLC

pftd ladderladder

Intladder

Big stations are thick stations

Small pixels make thick

Fast frame rates make thick stations

Fast pixels save material

othersSicablecoolladder ddddd

Time resolution of MAPSReadout freq. (Pixels / time)Pixel pitchµmp

MHzf

µstInt

30...10

105

10

(Simplified)


Correlation of thickness

Material budget (revised)

Material budget according to earlier studies

z = 10 cm

3000 D0/lifetime, S/B=0.3, Det. Eff. 0.7%@ Au+Au 25AGeV (preliminary)


Electronic system integration studies (Readout concept)

Demo-Aux-PCB

@ End of the ladder• Voltage regulation• Signal buffering• Single ended signal=> differential signal

Mi17 Mi17 Mi17 Mi17 Clock, bias,slow control

Sync. signals

Analogue output16 x 50 MHz <=> 1.2 GB/s

FPCUp to

280 pinsA

nalogoutput

Sync

Clock

JTA

G

ADC + FPGA

Bias (8V

DC

)

PC

Data transfer

Online monitoring

Reset

Data processing/reduction


How to study things more in detail => R&D in Frankfurt

4 x ADC12-bit, 50 MHz

Digital front-end

MAPS Analogsignal

MAPSslowcontrol

Virtex 4FPGA

512 MB

SD

RA

M

MAPS clock, fastcontrol

MAPS sync

Optical link2 Gbit/s

Sparsifieddata

Add-onboard

TRB-II

DAQ

12 Gbit/s

Bases on existing USB-FPGA board (IPHC, Strasbourg)

Allows for: • FPGA-based data sparsification (CDS, clustering, data compression)• Pipeline readout of MIMOSA-17 (~ 1 ms time resolution) without dead time


Running conditions

Needs still to be scaled with pile – up (up to factor 100)

Documents

C hallenges with decay vertex detection in CBM using an ultra-thin pixel detector system