Performance and Aging of the BaBar Drift Chambersuperb04/talks/Kelsey.pdf · 2004-01-22 ·...

Performance and Aging of the BaBarDrift Chamber

Michael H. Kelsey

Stanford Linear Accelerator Center™

B

AB

AR

™&

©

L . de B r unhoff

Super B Factory Workshop in Hawaii

19–22 January 2004

DCH High Lumi

Outline

• Drift Chamber Design and Operation

• Tracking, dE/dx, and Physics

• Aging, Damage and Remediation

• Future Operations, Performance

• Summary and Outlook

Michael H. Kelsey SuperB 2

DCH High Lumi

BaBar Detector


DCH High Lumi

BaBar Drift Chamber

Main tracking detector in BaBar

Surrounds beam pipe, final focus magnets, and

silicon vertex tracker


DCH High Lumi

BaBar Drift Chamber

Small hex cells (1∼2 cm)

HEX2 - Wire 168 Isocrones every 50 ns HEX2 - Wire 185 Isocrones every 50 ns

10 superlayers of 4 layers each

Axial and stereo (∼ 4◦)

96–256 cells/superlayer

Sense wires 20 µm W-Rh (Au)

Field wires 120 µm Al (Au)

Guard wires 80 µm Al (Au)


DCH High Lumi

DCH Operation

Gas mixture 80% helium, 20% isobutane,

3500–4000 ppm water vapor, ∼ 80 ppm O2

Designed to operate at 1960V

Initially operated without water vapor

Discharges observed in small region of chamber

Reduced to 1900V (October 1999–July 2000)

Since January 2001 (85% of total data) at 1930V


DCH High Lumi

Premature Aging

May 1999: 80:20, O2 < 10 ppm, H2O <∼ 100 ppm

28 July 1999: Large spikes in HV current

HV problem in SL6, OUT

Day in July 1999

Cur

rent

(µA

)

Minutes from 7/28/99 7 AM

Cur

rent

(µA

)

0

5

10

15

20

25

30

28 29 30

0

5

10

15

20

25

30

0 10 20 30 40 50 60

Hit Map High Charge

-100

-80

-60

-40

-20

0

20

40

60

80

100

-100 -80 -60 -40 -20 0 20 40 60 80 100

2000/05/31 11.02

x(cm)y(

cm)

October 1999: Turned off affected region, added water

=⇒ No discharges observed in chamber since


DCH High Lumi

Accumulated Charge

Measure aging vs. accumulated charge per unit

wire length

276.4 cm × 7104 cells

= 19.6 km sense wire

400–600 µA w/beams

(0.25 nA/cm)

Recorded every second

DCH Dose Since Startup (May 1999 - Present)

0

2.5

5

7.5

10

0

2.5

5

7.5

10

20000

2.5

5

7.5

10

20010

2.5

5

7.5

10

20020

2.5

5

7.5

10

20030

2.5

5

7.5

10

2004

Acc

umul

ated

Cha

rge

(mC

/cm

)

Total charge 12.2 mC/cm in nearly five years


DCH High Lumi

Reconstruction Performance

Hit resolution 〈σ(resid)〉 ∼ 125 µm

Target: 140 µm in middle region of cell

Momentum σ(pT )/pT = 0.45% + 0.15% pT (GeV/c)

Target: 0.21% + 0.14% pT

Tracking > 95% matching with SVT tracks

dE/dx resolution ∼ 7.5%, ± (0.5-0.7)% bias

Early test results: 7.0%


DCH High Lumi

dE/dx, Particle ID

dE/dx vs momentum

10 3

10 4

10-1

1 10Track momentum (GeV/c)

80%

tru

ncat

ed m

ean

(arb

itra

ry u

nits

)

eµ

π

K

p dBABAR

p<0.

6 G

eV/c

Drift Chamber K/π Separation

0

50 π K

0.6<

p<0.

8

0

100π K

0.8<

p<1

0

100π K

DCH dE/dx - dE/dx(K)(arb. units)

p>1

GeV

/c

0

1000

-600 -400 -200 0 200

πK

Drift Chamber K/π Separation

BABAR

Good π/K separation up to ∼ 700 MeV/c.

Provides confirmation of DIRC, and coverage

outside of DIRC acceptance.


DCH High Lumi

Physics Performance

Reconstruction of K0S → π+π− at large radii

)2Vtx+Y

2

Vtx(X√Decay Radius

2M

eV/c

0

2

4

6

8

10

12

0 10 20 30 40 50

RMS

hwhm

)2Vtx+Y

2

Vtx(X√Decay Radius

2M

eV/c

0

2

4

6

8

10

12

Bea

mP

ipe

Bea

mP

ipe

SV

TS

VT

Su

pp

ort

Su

pp

ort

DC

HD

CH

Fits in DCH comparable to

DCH+SVT

2) GeV/c

os)-M(K-π +πM(

)2Ev

ents/

(0.5 M

eV/c

0

50

100

150

200

250

300

350

25→Decay Radius 24. Entries = 4932oN

Sig. Events = 4460.50 MeV-0.045771

0.0458304mean = -0.156774RMS = 5.08925 MeV

/Ndof = 197.51/1922χhwhm = 2.38295 MeV

frac1 = 0.474382sigma1 = 3.8942 MeV

frac2 = 0.391389sigma2 = 1.5711 MeV

frac3 = 0.13423sigma3 = 11.4963 MeV

frac4 = 0sigma4 = 0.5 MeV

-0.05 0 0.05

“Jumps” due to material scattering uncertainty and fewer

hits per track fit


DCH High Lumi

Physics Performance (II)

B → D(∗)D(∗), D∗+ → π+D0, D0 → K−π+

(Monte Carlo generated events)

Momentum σ(p)/p 4.7 × 10−3

D0 → K−π+ σ(mass) 6.5 ± 0.2 MeV/c2

D∗ → π+D0 σ(mass) 0.80 ± 0.03 MeV/c2


DCH High Lumi

Gain vs. Time

G(Q) = {G0 + ∆G1|Q>Q1+ ∆G2|Q>Q2

+ . . .} exp(−AQ)

Accumulated Charge (mC/cm)

Cor

rect

ed G

ain

DCH Gain Since Startup (May 1999 - Present)

0.4

0.6

0.8

1

1.2

0 2.5 5 7.5 10

Aging rate: -0.549± 0.023 %/(mC/cm) over 12.17 mC/cm


DCH High Lumi

Long-term Damage

Besides aging, “sudden damage” always possible

Transient discharges, voltage trips

Buildup of deposits on wires

Self-sustaining discharge (Malter effect)

Long-term studies of aging and remediation

Lu Changguo (Princeton)

Pisa Frontier Detectors Meeting (2003)

Adam Boyarski (SLAC)

DESY Aging Workshop (2001)

IEEE NSS/MIC (2003)

Other groups (Colorado, Montreal, Novosibirsk, ...)

Accumulated several times BaBar lifetime charge


DCH High Lumi

Suppressing Damage

Running chamber without water vapor allows

polymer (dielectric) buildup, increases likelihood

of discharges

Presumed mechanism for damage seen in July 1999

Princeton test chamber run dry up to 130 mC/cm,

saw discharges, high singles rate, 10% drop in gain

Adding 1500–4000 ppm H2O eliminated discharges

Gain stabilized at 0.9 of initial value, up to 300 mC/cm

Poor performance returned when water removed


DCH High Lumi

Remediating Damage

Excess current can trigger self-sustaining discharges

Maximum safe current significantly reduced over time

0 20 40 60 80 100 120

Time (Min)

0

100

200

300

400

500

An

od

e C

urr

en

t (n

A)

Adam Boyarski, SLAC

“Training” with oxygen (500 ppm) gradually raises current

limit to original construction. Maintained after O2 removed.

Further study underway to confirm long-term performance


DCH High Lumi

Future Performance

Can we operate at ten times design luminosity?

DCH currents vs. Luminosity

0

1000

2000

3000

10 20 30 40Luminosity (1033 cm-2 s-1)

DC

H c

urre

nt, 1

930V

(µA

)

20002001

20022003

2004

2005

2006

20072008

250

500

750

1000

DC

H o

ccup

ancy

(Q

>50

)

High occupancy?

High currents?

High rates?

Aging problems?

DCH performance computed from beam currents and luminosity

By 2008, average occupancy, current, trigger rates

will be tripled from current (2003) conditions.


DCH High Lumi

Tracking at High Luminosity

Mix Monte Carlo B → D(∗)D(∗) with real random-trigger data

Multiple triggers overlaid to match luminosity extrapolations

Evaluate physics quality relative to current performance

Compared to design 2 × 1034 4 × 1034 4 × 1034

3 × 1033 cm−2 s−1 (3× bkg) (5× bkg) (10× bkg)

Tracking eff. (%) 98.6 ± 0.1 ± 0.7 97.4 ± 0.1 ± 1.0

Momentum

σ(p)/p = 4.7×10−3+4.2 × 10−5 +5.5 × 10−5

D0→ K+π− (%) 96.0 ± 0.5 95.5 ± 0.5 80 ± 3

D0 Mass, σ

6.5 ± 0.2 MeV/c26.5 ± 0.2 6.4 ± 0.2 7.0 ± 0.3

D∗ → D0π (%) 84.4 ± 1.1 75.0 ± 1.3 25 ± 2

D∗ Mass, σ

0.80±0.03 MeV/c20.97 ± 0.04 1.50 ± 0.08 3.2 ± 0.8


DCH High Lumi

High Rate Data Acquisition

Modular, parallel electronics

4-buffer pipeline per channel

16 elements per quadrant

via 1 GHz fiber to processor

FEA−3 180 ch

FEA−2 144 ch

FEA−1 96 ch

24 ch

16x FEE

to DIOM

Readout time set by single element’s occupancy

TDAQ = 8

{

8 +

N∑

(32mi + 4)

}

× 16.7 + 33 ns

N non-empty chips, mi = 1–8 hits each

=⇒ 45 hits requires 200 µs (5 kHz)


DCH High Lumi


Readout time scales with HV current, luminosity

(uniform occupancy)

IDCH (µA)

Rea

do

ut

tim

e (µ

s)

0

50

100

150

200

0 100 200 300 400 500 600 700 800

M. Cristinziani

)-1 s-2 cm33

Luminosity (x 100 5 10 15 20 25 30 35 40

Dea

dti

me

(%)

0

20

40

60

80

100

Deadtime Projection

Existing DAQ

With upgrade

2003

2004

2005

2006

2007-2010

C. Jessop

Deadtime when readout ∼ trigger rate

Expect 30–40% d.t. by 2006, 50+% by 2007


DCH High Lumi


Redesign front-end with extreme parallelism

Three chips per logical “element”

≤ 24 channel serial readout

< 105–140 µs readout time, 7 kHz rate

Essentially no deadtime for normal running,

through lifetime of Babar.

Requires 20 fiber readout processors

Targeting installation in mid-2005


DCH High Lumi

Effect of Aging on Gain

Integrate current to estimate accumulated charge

Include expected running year by year

Linear lumi improvement within years

DCH Projected Dose (1999 - 2009)

0

20

40

60

80

100

2000 2002 2004 2006 2008

Acc

umul

ated

Cha

rge

(mC

/cm

)

DCH Projected Gain (1999 - 2009)

0

0.25

0.5

0.75

1

0 20 40Accumulated Charge (mC/cm)

Est

imat

ed G

ain

(196

0V=

1)

Use fitted aging function to extrapolate gain

Assume 1930V operation, no gas changes


DCH High Lumi

Compensating for Aging

By 2007, gain at 1930V could be >∼ 25% below

current performance

Significant impact on trigger, track finding

Non-uniform acceptance, efficiency

Compensate for gain loss by increasing voltage

Lower gain, same HV current as now

+10V ⇒ ∆G ∼ +0.09

Raise when G <∼ 0.65

E.g. 2006, 2008

DCH Projected Gain (1999 - 2009)

0

0.25

0.5

0.75

1

0 20 40Accumulated Charge (mC/cm)

Est

imat

ed G

ain

(196

0V=

1)


DCH High Lumi

Summary and Outlook

BaBar Drift Chamber has been operating smoothly

since October 1999

Tracking performance up to design

dE/dx performance good, improvable

Excellent operational efficiency

Reasonable aging rate observed so far

Accumulated 12.2 mC/cm after five years

Gain fits well for simple aging 0.55%/(mC/cm)

Overall reduction 15% since startup

Comparable to other large experiments


DCH High Lumi

Summary and Outlook

Expect reliable performance over 10-year lifetime

50–80 mC/cm total charge, 25% loss of gain

⇒ Well below limits found in test chambers

Minimal degradation of physics performance

Increasing data rate will require significant elec-

tronics upgrade

Damage and aging may be remediated with

additives

3500 ppm H2O should prevent discharges

Aging may be compensated by adjusting voltage


DCH High Lumi

Supplemental Information


DCH High Lumi


DCH High Lumi

The PEP-II B Factory

Asymmetric e+e− collider: E(e+) = 3.1 GeV, E(e−) = 9 GeV

⇒ 10.58 GeV CMS: Υ(4S) → B+B−, B0B0

B decay length increased from ∼ 30 µm (CMS) to ∼ 250 µm

(lab), allowing precision time-dependent measurements

PEP-II delivers luminosity in ex-

cess of 6× 1033 cm−2 s−1

Over 165 fb−1 delivered to

BaBar (∼ 9% off resonance)

BaBar runs > 95% efficiency:

nearly 120 million BB pairs

Nov 1

Jan 1

1999

Mar 1

May 1

Jul 1 Sep 1 N

ov 1 Jan 1

2000

Mar 1

May 1

Jul 1 Sep 1 N

ov 1 Jan 1

2001

Mar 1

May 1

Jul 1 Sep 1 N

ov 1 Jan 1

2002

Mar 1

May 1

Jul 1 Sep 1 N

ov 1 Jan 1

2003 2004

0

10

20

30

40

50

60

70

80

90

100

110

120

130

140

150

160

Inte

grat

ed L

umin

osity

(fb

-1)

BA BA R

PEP-II Delivered 165.13/fbBABAR Recorded 157.78/fb

BABAR off-peak 14.65/fb


DCH High Lumi

BaBar Drift Chamber

IP1618

469236

324 681015 1749

551 973

17.19ÿ20235

Dimensions in mm

24 cm inner radius

81 cm outer radius

2.8 m length

IP 37 cm behind center

Be & Al inner cylinder

Carbon fiber/Nomex outer

2.5 (1.2) cm Al end plates

All electronics on rear


DCH High Lumi

Track Fitting

Residuals of hits vs. distance from sense wire

0

0.005

0.01

0.015

0.02

0.025

-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1

Mean 125 µm

Signed distance from wire (cm)

Res

olut

ion

(cm

)

〈σ(resid)〉 ∼ 125 µm

Design target: 140 µm in middle region of cell

d(t): pair of 7th order Chebyshev functions, each

side of cell. Adjusted for angle through cell.


DCH High Lumi

Track Fitting

Momentum resolution from cosmic rays

0

1.0

2.0

0 4 8Transverse Momentum (GeV/c)

σ(p

t)/p

t (%

)

1-2001 8583A23

σ(pT )/pT = 0.45% + 0.15% pT (GeV/c)

Design target: 0.21% + 0.14% pT


DCH High Lumi

Tracking Efficiency

Polar Angle (radians)

Transverse Momentum (GeV/c)

Eff

icie

ncy

0.6

0.8

1.0

0.6

0.8

1.0

Eff

icie

ncy

0 1 2

3-2001 8583A40

1960 V

1900 V

a)

1960 V

1900 V

b)

0.0 0.5 1.51.0 2.0 2.5

“Pseudo-efficiency”

Count DCH+SVT tracks

vs. total SVT tracks

Measure acceptance in

momentum or pT

polar angle

azimuth (not shown)

DCH and SVT not independent


DCH High Lumi

Physics Performance (III)

D0 → K−π+ reconstruction efficiency vs. φ

ask David Williams for plot


DCH High Lumi

Quantifying Gain

Normalize gain to absorb environmental effects

Pressure and temperature ⇒ density

Detailed gas mixture (He:i-C4H10, H2O, O2)

Precise operating voltage

Scale dE/dx from Bhabha-scattering events

[e+e− → e+e−(γ)]; normalization done ∼ hourly


DCH High Lumi

Normalizing Gain

Most significant correlation with gas density

absgain(time)

DCH Gain vs. Density (1930V, Aug-Dec 2001)

0

500

0.68 0.735 0.79

dchdens()

0

500

3.34 3.41 3.48

P/T-3.43

Gai

n

0

1

-0.04 -0.02 0 0.02

14.24 / 48

A0 0.7307 0.1112E-02

A1 -0.9868 0.3940E-01

Density (P/T)

Gco

rr

0

1

3.35 3.4 3.45 3.5

nR/V = Pabs/Tabs

Several short time intervals

Jan–Jul 2000 (1900V)Jul–Dec 2000 (1960V)Mar–Jul 2001 (1930V)Aug–Dec 2001Feb–Jun 2002Feb–Jun 2003

Correct gain using slope of fit

Applying offset could obscure aging effects


DCH High Lumi

Gain vs. Time

Continuous decrease G = G0 exp(−AQ) due to aging


Cor

rect

ed G

ain


0.4

0.6

0.8

1

1.2

0 2.5 5 7.5 10

A B C D

1 2 3 4 5

Steps due to operating changes (e.g., voltage)


DCH High Lumi

Gain vs. Time Fit Details

G(Q) = {G0 + ∆G1|Q>Q1+ ∆G2|Q>Q2

+ . . .} exp(−AQ)

χ2/dof = 668.33 / 214

−A(%) = −0.549 ± 0.023

G0 = 0.975 ± 0.009

∆G1 = −0.429 ± 0.009 (Q1 = 0.3)

∆G2 = 0.432 ± 0.001 (Q2 = 1.2)

∆G3 = −0.276 ± 0.001 (Q3 = 2.8)

∆G4 = 0.053 ± 0.001 (Q4 = 4.4)

∆G5 = −0.035 ± 0.001 (Q5 = 8.7)


DCH High Lumi

Systematic Checks


Cor

rect

ed G

ain

DCH Gain (showing error on means)

0.4

0.6

0.8

1

1.2

0 2.5 5 7.5 10



Cor

rect

ed G

ain

DCH Gain (No data excluded)

0.4

0.6

0.8

1

1.2

0 2.5 5 7.5 10



Gai

n


0.4

0.6

0.8

1

1.2

0 2.5 5 7.5 10


No density correctionAccumulated Charge (mC/cm)

Cor

rect

ed G

ain


0.4

0.6

0.8

1

1.2

0 2.5 5 7.5 10


Coarser binning (50 vs. 200 bins)


DCH High Lumi

Comparative Aging

Aging measurements from other experiments

[mC/cm] [%/(mC/cm)]Experiment Gas Mix Charge ∆G/G Aging

CDF Ar:Eth:Alc 130 <1 ∼ 20(Run 2) 50:50:1

ZEUS Ar:Eth:CO2 100 <∼ 0.183:5:12

H1 Ar:Eth:H2O < 10 >∼ 150:50:0.1

HERA-B Ar:CF4:CO2 2300 ∼ none(test) 65:30:5

BaBar He:i-C4H10:H2O 12.2 0.680:20:0.4

From 2001 DESY Workshop presentations


DCH High Lumi

Suppressing Damage

Lu Changguo, Princeton


DCH High Lumi

Remediating Damage

Additives (alcohol, water, oxygen) can suppress

discharges, allow running at higher currents

Before With Additive AfterAdditive (%) Imax Time (hr) Imax Imax Cured?

Methylal 4 0.3 ≈0 >8 No2 3 0.4 No

2-Propanol 1.0 ≈0.5 ≈0 >12 No0.5 >10 No0.25 >13 0.2 No

H2O 0.35 0.4 ≈0 >27 No0.18 >9 0.5 No

O2 0.10 0.5 1.5 >32 >40 Yes0.05 0.4 2 >29 >16 Yes0.02 0.9 10 >35 >14 Yes

CO2 5 0.4 35 >40 >27 Yes

O2+H2O 0.4 40 10 3 Partly(0.05+0.35)

Adam Boyarski, SLAC

Temporary 500–1000 ppm O2 running appears to

“cure” discharge problem. Current limit restored

to initial level.


DCH High Lumi

PEP-II Luminosity Upgrades

[1033cm−2 s−1] [mA] [mA] [µA] [raw/rec]

Year L I(e−) I(e+) DCH HV Hits/ev

2003 6.3 1140 1450 532 456/ 332

2004 12.1 1600 2700 973 686/ 483

2005 18.2 1800 3600 1380 898/ 621

2006 23.0 2000 3600 1665 1046/ 719

2007 33.0 2200 4500 2284 1368/ 930

2008 33.0 2200 4500 2284 1368/ 930J. Seeman/PEP-II, SLAC


DCH High Lumi

Tracking at 4 × 1034cm−2 s−1

Momentum resolution

σ(p)/p

D∗ → D0π Mass

D* mass

0

25

50

75

100

2.002 2.004 2.006 2.008 2.01 2.012 2.014 2.016 2.018 2.02

568.8 / 66P1 0.4578 0.8226E-02P2 2.010 0.2885E-04P3 0.3429E-03 0.4997E-05P4 0.2012E-02 0.7171E-04P5 758.1 8.335


Performance and Aging of the BaBar Drift Chambersuperb04/talks/Kelsey.pdf · 2004-01-22 ·...

Documents

Transcript of Performance and Aging of the BaBar Drift Chambersuperb04/talks/Kelsey.pdf · 2004-01-22 ·...