VCI 2004Gaia Lanfranchi-LNF/INFN1 Time resolution performances and aging properties of the MWPC and RPC for the LHCb muon system Gaia Lanfranchi – INFN/LNF

VCI 2004 Gaia Lanfranchi-LNF/INFN 1

Time resolution performances and aging Time resolution performances and aging properties of the MWPC and RPC for the LHCb properties of the MWPC and RPC for the LHCb

muon systemmuon system

Gaia Lanfranchi – INFN/LNFGaia Lanfranchi – INFN/LNF on behalf of the LHCb muon group on behalf of the LHCb muon group

• OverviewOverview• MWPC design;MWPC design;• Time resolution performances; Time resolution performances; • Aging properties;Aging properties;• RPC aging results.RPC aging results.• Conclusions.Conclusions.


5 muon stations, one in front (M1), 4 behind (M2-M5) the calorimeters. Muon detector must provide high-pt L0 muon trigger through a coincidence of hits in all five stations within the bunch crossing time of 25 ns. The muon momentum must be measured with δpt /pt ~ 20% accuracy.

The LHCb Muon DetectorThe LHCb Muon Detector

M1M2

M5M3 M4

R4

R3

R2

R1


Requirements:Requirements:

High (> 99%) efficiency per station in a 20 ns time window: good time resolution

Good rate capability: up to ~200 kHz/cm2 @ L = 5•1032 cm–2 s-1 in hottest region; fast

Aging resistant: up to integrated charge ~ 1 C/cm for safe operation in 10 years @ <L> = 2•1032 cm–2 s-1 .


Options:Options: M4 M5

M1

M2 M3

MWPC

GEM

RPC

~48% RPC ~ 52% MWPC ~0.1% GEM : M1R1, rate ~ 500 kHz/cm2 .

~ 99.9 % MWPC~435 m2, 1380 chambers


MWPC designMWPC design 4 wire layers ( for time resolution and redundancy); 5 mm gas gap; 2 mm wire pitch ( gold plated tungsten wires, 30 μm diameter ).

CathodePCB

StructuralPanel(poliuretanic foam)

2 layers are OR-ed in oneFE channel

2 FE channelsOR-ed onFE board

HV

HV

HV

HV

4 separateHV suppliesthe anode wires

Gold platedcathode pads

VCI 2004 6

MWPC operating parameters:MWPC operating parameters:

Gas mixture: Ar/C02/CF4 (40:40:20); Working point: Gain~ 5•104, total avalanche charge per mip ~ 0.4 pC Field on wires: 212 kV/cm, field on cathodes: ~ 5 kV/cm.

Drift velocity almost saturated at ~ 90-100 μm/ns

For 6 GeV muons we expect about:~ 45 clusters/cm, 53 e-/gap.

clusters/cm Drift velocity:

6 GeV

VCI 2004 Gaia Lanfranchi-LNF/INFN 7Different readouts and detector capacitances pose special requirements to the FE-chip.

MWPC readout schemes:MWPC readout schemes:

R4: wire pad readout R3: cathode pad readout

R1/R2-M2/M3: mixed readout

Wire pad, cathode pad and mixed (wires+cathode pads) readout schemes depending on the granularity requested from trigger and offline and on the particle rates.

Granularity goes from (1x2.5)cm2 to (25x30) cm2

MWPC dimensions from (20x48)cm2 to (149x31)cm2


A small MWPC prototype:A small MWPC prototype:

panel

Gold plated cathode pads

4 gaps

Lateral bars of FR4


PhotosM2R4 – wire readout

M3R3 – cathode readout

M2R1 prototype (mixed readout)

~ 15 prototypes built and tested with (almost) final front-end electronics at T11 test beam area at CERN.

The Front-End Electronics: the CARIOCA chipThe Front-End Electronics: the CARIOCA chip

Specifications important for time resolution and rate capability:

short peaking time: tp ~ 10 ns for Cdet = (40220) pF

low noise: ENC ~ 2000+40 e-/pF

high rate capability: pulse width ~ 50 ns, signal tail cancellation and baseline restoration circuits.

Carioca is the Amplifier-Shaper-Discriminator front-end chip developed for MWPC of LHCb.

(documentation in: http://www.cern.ch/riegler)

VCI 2004 11

Time resolution:Time resolution:

rms = 3.5 ns

rms = 4.3 ns

rms = 7.4 ns

20 ns time window

ε (quadrigap station) > 99% in 20 ns time window; ε (double gap) 95% in 20 ns time window time rms for double gap < 5 ns

Double Gap Time Resolution: Double Gap Time Spectra:

Cathode pad readout

5 ns


Time resolution and efficiencyTime resolution and efficiency

ε(double gap) 95% in 20 ns rms (double gap) 5 ns: HV(95%) = 2.45 kV for wires readout; HV(95%) = 2.55 kV for cathodes/mixed readout;

Double gap efficiency for cathode readout

Working Point = 2.6 kV

VCI 2004 13

Time Resolution and Efficiency Uniformity:Time Resolution and Efficiency Uniformity:

Example of uniformity measurements on a M3R3 prototype with cathode pad readout:

> 99% pads are inside the specifications

Time resolution: Double Gap Efficiency in 20 ns:

5 ns95 %

Pad number Pad number

Time rms = (4.30.2) ns

ε (HV=2.6 kV) = (96.7 ±0.1) %


High Rate behaviorHigh Rate behavior

Time resolution stable (no space charge

effects)

Small Efficiency drop due to signal pile-up

Test @ GIF up to 500 kHz/FEE channel:

~ 2%


Aging Test of MWPCsAging Test of MWPCs

- Q < 10 mC/cm for 76% of area.

- 10 mC/cm <Q < 100 mC/cm for 19% of area.

- 100 mC/cm < Q < 1 C/cm for 5% of area.

- Particle rates are from LHCb-Muon TDR2001.- Safety factors of have been included to take into account the deposited charge of low energy background hits.

Integrated charges in 10 years @ 2•1032 cm-2 s-1


Aging tests @ ENEA Casaccia Calliope Facility:Aging tests @ ENEA Casaccia Calliope Facility:

Source 60Co (~ 1015 Bq), <Eγ> ~ 1.25 MeV.

3 MWPC irradiated with dose rates up to ~ 0.3 Gy/hr and different gas flows (vented and re- circulating gas system).

Standard mixture: Ar:C02:CF4=40:40:20 Gas gain ~ (11.5) 105 (~double wrt to the nominal one).

LNF,CERN1, CERN2

PC

Co60

HV

CONTROLROOM

4.5

m

P

T

win

dow

~0.3 Gy/hr

VCI 2004 17

Gas System: Chambers materials:Gas System: Chambers materials:

LNF chamber:• Adekit 140 for chamber closing.

CERN chambers:• Natural rubber O-rings.• Kapton foils to protect HV-traces and glued with epoxy.• Low temperature soldering, carbon film resistors and SMD capacitors.

CERN CHAMBERS

Common Materials:Gold plated 30 μm tungsten wires.Gold plated cathode pads.Adekit 145/450 epoxy for wire glueing.

One bottle of each gas (Ar, CO2, CF4).

Mass flowmeter with ORing in Viton.

Total gas flow: 40/40/20 cc/min = 6 l/hr to mixer.

Typical flows: Open Mode ~ 4.8 l/hr (~2 volumes/hr of LNF) Closed Loop: fresh gas ~1.35 l/hr, circulating gas:~ 5.2 l/hr

All pipes in Copper except 10 cm connections to chambers in Rilsan + brass connectors.

(with final components)


Aging tests @ ENEA Casaccia Calliope FacilityAging tests @ ENEA Casaccia Calliope Facility

32 days of tests. Integrated charge per cm of wire: Q ~ (440480) mC/cm (510) years depending on the rate evaluation. Typical current density: I ~ (1÷1.4) µA/cm2 (active areas = 500÷1200 cm2); In each chamber, one gap (out of 4) was switched on only for short periods to be used as reference.

Integrated Charge:

Reference gap

test gaps

VCI 2004 19

Aging tests @ ENEA Casaccia Calliope FacilityAging tests @ ENEA Casaccia Calliope Facility

We normalized the currents of tested gaps respect to the reference gap in order to remove T and P dependence and accidental gas mixture changes.

Current ratios are constant within ~10% for all chambers.

Current ratios – vented mode

CERN 1 I(A,C,D)/B(ref)

0,92

0,94

0,96

0,98

1

1,02

1,04

0 5 10 15 20 25

Days

A.U

.

Series1

Series2

Series3

Current ratios - closed loop

~ 4%

~ 8%

Malter currents: The self-sustaining rest currents were measured following the source off, using current monitors with a resolution of 1 nA. All gaps of all chambers draw currents of the order of few tens of nA or smaller with a decreasing trend.


Photo 0SEM Analysis of Wires

Wires are clean


Photo 1

Boundary of FR4 etching

Etching of the FR4 frame:

The FR4 is etched also where there is no electric field.This effect is visible also in the reference gap : due to ionized gas fluorine etching.


The etching of the FR4 frame goes with the gas flow :

B1 gap, “gas in”side

B2 gap, “gas in”

B2 gap, “gas out” A2 gap, “gas in”

No etching….

…No etching….…No etching yet…

...Etching starts……FR4 etched..

..FR4 etched!!

A2 gap, “gas out”B1 gap, “gas out”

Photo 2Gas flow direction: B1 B2 A2 A1, B1=reference gap


Photo 3

“Bumps” under the guard strips of the cathode pads

(not found in the reference gap)


Conclusions on the MWPC aging testConclusions on the MWPC aging test

In 32 days we integrated up to Q~480 mC/cm corresponding to 510 years (depending on the rates foreseen) of LHCb @ 2 1032 cm-2 s-1.

Materials exposed to CF4 under irradiation show a surface etching BUT no drop in gas gain observed within 10%:

we decided do not change chamber design and materials.


RPC for LHCB Muon SystemRPC for LHCB Muon System

M1

M2

M3

M4

M5

R1

R2

R30.75

1.1

0.65

1.0

R40.25

0.4

0.23

0.3

RPC

Historical review

1998: RPC were proposed for LHCb Muon detector in regions with rates < 1 kHz/cm2.

1999: 2 prototypes built with identical characteristics:

- bakelite electrodes (ρ ~ 1010 Ω cm); - linseed oil; - graphite ( 100 kΩ/) ;� - 50 x 50 cm2 area. - 2 mm gas gap (C2H2F4 : iC4H10 : SF6 = 95:4:1) - avalanche mode (HV ~10.6 kV).

2000: rate capability was measured to be 3 kHz/cm2

(NIM A 456 (2000) 95.)

2001: an extensive test for study the aging properties started at GIF…

Rates (kHz/cm2) andintegrated charge (C/cm2) for L = 5•1032 cm-2 s-1


Aging test in 2001 @ GIF Aging test in 2001 @ GIF

RPC A irradiated at GIF during 7 months up to Q~ 0.4 C/cm2 . RPC B not irradiated, used as reference. I, V0 and T continuously monitored. Bakelite resistivity ρ extracted from (I,V0) curve using a model for RPC operating under high flux conditions ( NIMA 498 (2003) 135) and corrected for T dependence.

RPC A (high flux):QA~0.4 C/cm2

RPC B (low flux): QB~ 0.05 C/cm2

GIF test setup


Aging test in 2001 @ GIF: Aging test in 2001 @ GIF: RPC A resultsRPC A results

Inte

grat

ed c

har

ge (

mC

/cm

2 )

ρ (T

=20

°) (

10 1

0 Ω•c

m)

jan01

aug01

dec01

mar01

month

Q ~ 420 mC/cm2

ρ ~ 40 1010 Ω•cm

irradiation

ρ ~ 6 1010 Ω•cm

Observed a steady increase of ρ with and without irradiation.

no irradiation

ρ ~ 70 1010 Ω•cm


Aging test in 2002 @ GIFAging test in 2002 @ GIF Both detectors slowly irradiated ( Q~0.05 C/cm2 accumulated charge ). Resistivity continuously measured during ~ 350 days. Observed a steady increase of ρ with time for both deterctors probably due to drying up of bakelite. Addition of 1.2% of H2O vapor to the nominal gas mixture produced a decrease of resistivity. This effect in any case disappeared as soon as dry gas was flowed.

ρ ~ 200 •1010 Ωcm

RPCA RPCB

ρ ~ 220 •1010 Ωcm

ρ ~ 10 •1010 Ωcm

ρ ~ 70 •1010 Ωcm

H2O vaporH2O vapor

VCI 2004 29

RPC Rate capabilityRPC Rate capability

T=25.0 oCT=24.5 oC

HV

HV

August 2001: ρ = 40•1010 Ω•cm (T=20°) July 2002: ρ = 110•1010 Ω•cm (T=20°)

Rate capability ~ 640 Hz/cm2 @ 20 oC

Rate capability ~ 200 Hz/cm2@ 20 oC

These results brought the LHCb Collaboration to abandon the RPC technology for the Muon Detector.


ConclusionsConclusions

Three years of extensive tests showed that our design of MWPCs operating with CF4 based gas mixture satisfies all the requirements for the LHCb Muon System.

We built a fast detector, with good time resolution and aging resistant.

It will cover ~ 435 m2 area with ~ 1380 chambers and ~126000 readout channels.

The production is started and the detector should be ready for the 1st LHC beams.


Spares:Spares:

VCI 2004 32

Gas System:Gas System:

One bottle of each gas (Ar, CO2, CF4). Mass flowmeter with ORing in Viton. Total gas flow: 40/40/20 cc/min = 6 l/hr to mixer. Typical flows: Open Mode ~ 4.8 l/hr (~2 volumes/hr of LNF) Closed Loop: fresh gas ~1.35 l/hr, circulating gas:~ 5.2 l/hr All pipes in Copper except 10 cm connections to chambers in Rilsan + brass connectors.

Ar

GAS SYSTEM

CO2

CF4

REF(A,B) LNF CERN 2

REF(C,D)CERN 1

Fresh gas

Circulating gas

CLOSED LOOP

OPEN MODE

Exhaust

( with final components)


Mechanical Tolerances:Mechanical Tolerances:

The specifications for a single gap were defined such as the gas gain is within : - 0.8*G0 and 1.25*G0 in 95% of the chamber area; - G0/1.5 and 1.5*G0 in 5% of the chamber area.

What chamber imperfections are allowed in order to keep the gain within specifications?

SPECIFICATIONS:

• Gap: 95% in 90 µm 5% in 180 µm • Wire pitch: 95% in 50 µm 5% in 100 µm • Wire y-offset: 95% in 100 µm 5% in 200 µm • Wire plane y-offset: 95% in 100 µm 5% in 200 µm

The most critical parameter for our chambers is the gap dimension !!


Measurement of gain uniformity:Measurement of gain uniformity:

A: B:

C: D:

good~ good

very good ~ good

-We scan each gap using a radioactive source (137Cs, 40 mCi, 0.66 MeV photons).

-The current drawn by group of wires is measured by a nano-amperometer with 1 nA resolution.

137Cs source

Lead case

137Cs sourceLead case

MWPC


M1 M2 M3 M4 M5

R1

460 GEM

37.5 mixed r/o

10 mixed r/o

6.5 cathode r/o

4.4 cathode r/o

R2 186 GEM or MWPC

26.5 mixed r/o

3.3 mixed r/o

2.2 cathode r/o

1.8 cathode r/o

R3 80cathode r/o

6.5 cathode r/o

1.0 cathode r/o

0.75 cathode r/o

0.65 cathode r/o

R4 25 wires r/o

1.2 wires r/o

0.4 wires r/o

0.25 wires r/o

0.23 wires r/o

Rates and ReadoutRates and Readout


Average ratesAverage rates andand Integrated Charges: Integrated Charges:

- Rates (TDR2000) and integrated charges in 10 equivalent (~ 108 s) years for average luminosity <L> = 2 • 1032 cm-2 s-1 , - Safety factors (2 in M1 and 5 in M2-M5) are included.

M1 M2 M3 M4 M5

R1

184 GEM

15

132MWPC

4

35MWPC

2.6

46MWPC

1.76

31MWPC

R2 74.4

328 GEM or MWPC

10.6

185MWPC

1.32

25MWPC

0.88

16 MWPC

0.72

13 MWPC

R3 32

141 MWPC

2.6

46MWPC

0.4

7 MWPC

0.3

6 MWPC

0.26

5 MWPC

R4 10

44 MWPC

0.48

4.3

MWPC

0.17

1.5MWPC

0.1

1MWPC

0.09

0.8 MWPC

kHz/cm2

mC/cm


MWPC readout schemes: MWPC readout schemes: rates + granularityrates + granularity

M1 M2 M3 M4 M5

R1

460 1x2.5 GEM

37.5 0.6x3.1 mixed r/o

10 0.67x3.4 mixed r/o

6.5 2.9x3.6 cathode r/o


R2 186 2x2.5 GEM

26.5 1.3x6.3mixed r/o

3.3 1.35x3.4 mixed r/o



R3 80

4x10 cathode r/o



0.75 11.6x14.5 cathode r/o

0.65 12.4x15.5 cathode r/o

R4 25 8x20 wires r/o

1.2 5x25 wires r/o

0.4 5.4x27 wires r/o

0.25 13.1x29 wires r/o

0.23 24.8x30.9 wires r/o


Comparison with simulationsComparison with simulations

Comparison with simulations has been done by plotting double gap efficiency and time resolution as a function of threshold.

The threshold is expressed as a fraction of the average signal in order to be independent from the gain.

Full simulation: - primary ionization (HEED); - drift, diffusion (MAGBOLTZ); - induced signals (GARFIELD).

Good agreement between data and simulations


Efficiency and crosstalk

Crosstalk: probability of firing the neighboring pad

Pt accuracy measurement requires:Crosstalk < 5%

Main source is the pad-to-pad capacitive coupling(crosstalk ~ s Rin Cpp)

~4%

2.6 kV

Cro

ssta

lk in

20

ns T

WC

ross

talk

in 2

0 ns

TW

Eff

icie

ncy

in 2

0 ns

Eff

icie

ncy

in 2

0 ns


The ATLAS Cathode Strip Chambers are intended for position resolution. Amplifier peaking time 80ns, bipolar shaping, ‘crosstalk intended’ on cathode strips for center of gravity.

The CMS Cathode Strip Chambers are intended for position resolution (cathodes strips) and timing (wires). Cathode amplifier peaking time 100ns, wire amplifier peaking time 30ns.

The LHCb MWPCs are intended for highly efficient timing within a certain spatial granularity at the LVL0 trigger. Amplifier peaking time of 10ns, pulse width<50ns, unipolar shaping, low crosstalk.

Since crosstalk is RinCpp and since (20 MHz) is high we have to minimize the pad-pad capacitance Cpp and amplifier input impedance Rin. Because we want unipolar shaping we need a baseline restorer in the front end.

Comparison with ATLAS/CMSComparison with ATLAS/CMS


Principle of RPC operation

qi

-----------+++++++++++

q=GqiV0

gas

Electrodesd,S,

particle

G~106-107; G1 for V0< VT

Due to space-charge effects

(especially with SF6):

qVgap-VT Vgap=V0 -IR

I = q(Vgap) ( = particle flux)

dR = 2ρ

S

If low I small VgapV0



qi

-----------+++++++++++

q=GqiV0

gas

Electrodesd,S,

particle

G~106-107; G1 for V0< VT





dR = 2ρ

S



0 TV - VXI =

1+X RX=k R

High flux conditions If we must have q 0, i.e. Vgap VT and Imax=(V0 –

VT)/R

Linear dependence on HV

Exponential temperature dependence of I via R:( )

20R R e

More generally:

Saturation with flux



qi

-----------+++++++++++

q=GqiV0

gas

Electrodesd,S,

particle

G~106-107; G1 for V0< VT





dR = 2ρ

S



0 TV - VXI =

1+X RX=k R

High flux conditions If we must have q 0, i.e. Vgap VT and Imax=(V0 –

VT)/R

Linear dependence on HV


20R R e

More generally:



Current saturation

X obtained fitting I at different

At GIF 1/Abs0.7

where Abs is the filter setting(Abs=1,2,5,.......)

0 TV - VXI =

1+X RX=k R

(1)



Electrodesd,S,

G~106-107; G1 for V0< VT





qi

-----------+++++++++++

q=GqiV0

gas

particle

dR = 2ρ

S



0 TV - VXI =

1+X RX=k R

High flux conditionsHigh flux conditions

If we must have q 0, i.e. Vgap VT and Imax=(V0 –

VT)/RLinear dependence on HV


20R R e

More generally:


I = q(Vgap) , = particle flux, q = G qi

qi

-----------+++++++++++

q=Gqi

V0

gas

particle

dR = 2ρ

S


Current saturationCurrent saturation

X obtained fitting I at different

At GIF 1/Abs0.7

where Abs is the filter setting(Abs=1,2,5,.......)

0 TV - VXI =

1+X RX=k R

(1)


Measurement of RMeasurement of R

R can be measured in a non-destructive way from the slope of I-V0 curve Its changes can be easily monitored Precise knowledge of X is unimportant if X is large (an estimate is sufficient)

We can obtain R by fittingI vs. V0 curve

Documents

VCI 2004Gaia Lanfranchi-LNF/INFN1 Time resolution performances and aging properties of the MWPC and RPC for the LHCb muon system Gaia Lanfranchi – INFN/LNF