performance of time projection chamber prototypes for the...
TRANSCRIPT
Performance of Time Projection Chamber Prototypes
for the Linear Collider Experiment
Makoto Kobayashi
on behalf of part of the ILC-TPC Collaboration
KEK, Saclay, Orsay, Carleton, Montreal, MPI, DESY, MSU、Tsukuba U, TUAT, Kogakuin U, Kinki U, Saga U, Hiroshima U.
Workshop on MPGDsJanuary 27, 2006 · Research Center for Nuclear Study, Osaka university
ILC-TPC R&D GroupsEuropeEurope
RWTH RWTH Aachen Aachen CERNCERNDESYDESY
U HamburgU HamburgU U FreiburgFreiburgU U KarlsruheKarlsruhe
UMM KrakowUMM KrakowLundLund
MPIMPI--MunichMunichNIKHEFNIKHEF
BINP BINP NovosibirskNovosibirskLAL LAL OrsayOrsayIPN IPN OrsayOrsayU U RostockRostockCEA CEA SaclaySaclay
PNPI PNPI StPetersburgStPetersburg
AmericaAmericaCarleton U
Cornell/PurdueIndiana U
LBNLMIT
U MontrealU Victoria
Asian ILC gaseousAsian ILC gaseous--tracking groupstracking groups
Chiba UChiba UHiroshima UHiroshima U
MinadamoMinadamo SUSU--IITIITKinki UKinki U
U OsakaU OsakaSaga USaga U
Tokyo UATTokyo UATU TokyoU Tokyo
NRICP TokyoNRICP TokyoKogakuinKogakuin U TokyoU Tokyo
KEK TsukubaKEK TsukubaU TsukubaU TsukubaTsinghuaTsinghua UU
Other Other MIT (LCRD)MIT (LCRD)
Temple/Wayne Temple/Wayne State (UCLC)State (UCLC)
YaleYale
Examples of Prototype TPCsCarleton, Aachen, Cornell/Purdue,Desy(n.s.) for B=0or1T studies
Saclay, Victoria, Desy(fit in 2-5T magnets)
Karlsruhe, MPI/Asia, Aachen built test TPCsfor magnets (not shown), other groups built small special-study chambers
Size: Effective volume = 4.1 m (diameter) ×4.6 m (full length) Spatial resolution r -φ: 100 ‒ 200 μm z: ≲ 1 mm Two-track resolving power
r -φ: ≲ 2 mm z: ≲ 10 mm Number of coordinate samples per track: ~ 250 Momentum resolution (TPC alone)
δPt / Pt ~ 10⁻⁴ Pt [GeV/c] for high momentum tracksParticle identification capability: dE/dx ~ 4%
ILC-TPCA Large, High precision, High 3-D granularity Time Projection Chamber
operated under a high magnetic field of 3 – 4 T
Unprecedented requirements to TPC especially for granularity
→ use of micro pattern gas detector (MPGD) as a readout device
→ performance tests using prototypes
Prototype TPC
Field cagemaximum drift length: 260 mmtypical cathode H.V. : - 6 kV
Readout planeeffective area: 100 mm×100 mm
MWPCGEMs (3-stage)
orMicroMEGAS
Experimental setup
• Beam: mostly 4 GeV/c pions (0.5-4 GeV/c hadrons &
electrons)
• Magnet: super conducting solenoidwithout return yoke (max field: 1.2 T)
TPC in JACCEE magnet
Beam
JECCEEinner diameter : 850 mmeffective length: 1 m
KEK π2 Beam Line
We have conducted a series of beam experiments at KEK PS.
Readout scheme and Gas
ALEPH TPC Electronics：charge sensitive preamp. + shaper amp. (shaping time = 500 ns)
+ digitizer (time bucket = 80 nsec)
Readout Device Pads： 32 pads ×12 pad rows Width (pitch)×Length (pitch)
Gas (1 atm.) Drift field
MWPC SW spacing 2 mm SW - Pads 1 mm
2 (2.3) mm×6 (6.3) mm
TDR [Ar- 4CH (5%)- 2CO (2%)]
220 V/cm GEM (triple stage) CERN Standard transfer gaps 1.5 mm induction gap 1.0 mm
1.17 (1.27) mm×6 (6.3) mm
staggered
TDR 220 V/cm
Ar – methane (5%) 100 V/cm
MicroMEGAS 50-μm mesh 50-μm gap
2 (2.3) mm×6 (6.3) mm
Ar - isobutane (5%) 220 V/cm
Examples of Experimental Data
Z = 26 cm
Z = 0 cm
0 4-4
1
0.5
0
Pad responses for different drift distances
GEM Gas: P5, B = 1 T
[mm]
Xtrack – Pad center
Pad response width vs. z
TDR
P5
Drift distance (z) [cm]
σPR2[mm]
0 25
0.4
0.8
1.2
10 20
GEM B = 1 T
zD22PR0
2PR += σσ
PR0σ (μm) D (μm / √cm)
TDR
443 ± 5 207 ± 1
P5 511 ± 2 168 ± 1
Spatial resolution vs. z
TDR
P5
0 2510 20
σx[mm]
0.05
0.1
0.15
0.2
row6
0.25 GEM B = 1 T
zNDX eff22
X0 /+= σσ
Drift distance (z) [cm]
X0σ (μm) effN
TDR
81 ± 6 30 ± 2
P5 67 ± 6 27 ± 2
Track coordinate measurement
Amplification Gap
Readout Pads
Beam
Drift Volume Drift and Diffusion
Amplification andfurther Diffusion
Pad Response
Ionizations
Coordinatex
Liberation of Electrons
wPad j
ixxx ∆+= *
*xx =
ijq , ijq ,1+ijq ,1−
Pad response function
drift path ofelectron i
ijq ,2+
,/)()(
/)(
* ∑∑ ∑
∑ ∑ ∑∑∆+=
⎟⎟⎠
⎞⎜⎜⎝
⎛=
ii
i jiji
j i i jjiji
QxxFjwQ
qqjwX
ijiij
jjii
QqxxF
/)(
,
* ≡∆+
= ∑where
Charge Centroid X
i runs through 1 to N, where N is the total number of drift electrons(not a constant).
is the charge on pad j created by electron i .
ijq ,
is the displacement due to lateral diffusion and its variance is given by
, where is thediffusion constant and z is the driftdistance.
DzDx 22 =>∆<
ix∆
ｗ : pad pitch j : pad #
Formulation of Spatial Resolution
)()()( ** xxFPxdxxF jDj ∆+∆≡∆+ ∫∞+
∞−
[ ]2
*
,
**2 )()()(ii ⎟⎟⎠
⎞⎜⎜⎝
⎛∆+−∆+∆+≡ ∑∑
jj
kjkj xxFjwxxFxxFjkw
[ ] ∑∆
≡j
jwNQ
q 222
2
)(1)(III
( ) .2
exp21
2
2
⎟⎟⎠
⎞⎜⎜⎝
⎛ ∆−≡Dx
DPD π
where
1
2
2
eff1
−
⎪⎩
⎪⎨⎧
⎪⎭
⎪⎬⎫
≡Q
QN
N
for tracks perpendicular to the pad row
with
and
q∆( ： electric noise charge on a pad)
The definition is similar for < Fj ⋅Fk > .
[ ]2
** )()(i⎭⎬⎫
⎩⎨⎧
−∆+≡ ∑j
j xxxFjw
( ) [ ] [ ] [ ]IIIii1ieff
2 21
21
* +⎭⎬⎫
⎩⎨⎧
+= ∫+
− Nd w
xXσ
Remarks on the Formulation
ξξ dxfxFwjw
wjwj )()(2/
2/∫+
−−≡
with f (x) being the pad response function.
,11
2
2
eff RN
Q
QN
N ≡⎪⎩
⎪⎨⎧
⎪⎭
⎪⎬⎫
≡
−
where R is defined as ><><+≡ −1)1( NNKRwith ,/ 22 ><≡ QK Qσ
the relative variance of avalanche fluctuationfor a single drift electron.
><>< −1NN is greater than 1 because of asymmetric distribution of N.
Therefore the effective number of electrons is smaller than the averagenumber of drift electrons.
Therefore Fj can be evaluated once the pad response function is defined.
Remarks (cont’d)
Distribution of N (<N> = 71)
Distribution of 1/N(<1/N> = 0.028)
Distribution of Q(K = 0.67)
21
32.3028.071)67.01()1(
eff
1
≈=
≈××+≈><><+≡ −
RN
N
NNKR
An example of estimation of Nefffor 4 GeV/c pions and pad row pitch of 6.3 mm
in pure argon
Polya type with θ = 0.5<Q> = 1
[I] Finite pad pitch systematic biases due tocharge centroid method.Rapidly decreases with drift distance (z) because ofdiffusion. N independent.
This term can be reduced by calibration if the signal charge is shared by two or more pads.In reality, however, electronic noise significantly degrades the resolution when the charge sharing among the padsis not sufficient.
[II] Diffusion, gas-gain fluctuation, finite pad pitchcombined.Gradually increases with drift distance,asymptotically .
[III] Random electronic noise. z independent.
Origin and characteristic of each term
( )zDwN 22eff 12//1 +
. when 0at 12/2 δ== fzw
Illustration of each contribution
z
2Xσ
[III]
[II][I]
12
2w
⎟⎠
⎞⎜⎝
⎛+≈ zDw
N2
2
eff 121
z
2PRσ
2PRF
22PR0
22PR0
2PR
12with σσ
σσ
+=
+=
w
zD
Pad Response
(when f = δ -function)
Charge spread with respect to the centroid
( )
( ) ( )
( )
( )∑ ∑∑
∑
∑∑
∑ ∑∑
∑
∑ ∑∑
∑∑
∑
∑ ∑∑
∑∑ ∑∑
∑
=
=
=
==
⎟⎟⎟
⎠
⎞
⎜⎜⎜
⎝
⎛−−−=
⎟⎟⎟
⎠
⎞
⎜⎜⎜
⎝
⎛−⋅−
⎪⎭
⎪⎬
⎫
⎪⎩
⎪⎨
⎧
⎟⎟⎟
⎠
⎞
⎜⎜⎜
⎝
⎛−+−=
⎪⎭
⎪⎬
⎫
⎪⎩
⎪⎨
⎧
⎟⎟⎟
⎠
⎞
⎜⎜⎜
⎝
⎛−−−
⎟⎟⎟
⎠
⎞
⎜⎜⎜
⎝
⎛−+−=
⎪⎭
⎪⎬
⎫
⎪⎩
⎪⎨
⎧
⎟⎟⎟
⎠
⎞
⎜⎜⎜
⎝
⎛−−−=
⎟⎟⎟
⎠
⎞
⎜⎜⎜
⎝
⎛−
N
ik
k
kkk
i
kk
i
kk
kkkN
ik
k
kkk
i
kk
i
N
ik
k
kkk
i
kk
kkk
i
kk
i
N
ik
k
kkk
i
kk
iN
ik
k
kkk
i
kk
i
xq
qxxx
xq
qxx
q
qxxx
xq
qxxxx
q
qxxx
xq
qxxx
q
qxx
1
2
*
#
2*#
2
*
#
1
2
*
#
2*#
1
*
#
*#
2
*
#
2*#
1
2
*
#
*#
1
2#
#
2
2
( ) ( ) ,2*#2*#
1
2#
# xXxxq
qxx
qqN
ik
k
kkk
i
kk
i −−−=⎟⎟⎟
⎠
⎞
⎜⎜⎜
⎝
⎛−∑ ∑
∑∑=
.
#
#
∑∑
≡
kk
kkk
q
qxXwith
:#ix
:*x original (true) track coordinatearrival position of electron i, quantized by the pad pitch (w)
Averaging over x , x , q and N,＊
i i
( ) 2d
22*#
12σ+=−
wxxNotice that and ( )2*# xX − is nothing but the spatial resolution ( ).2Xσ
Notations
Width of Pad Responsepad response function: δ - function
Ｚ (drift distance)
⎟⎠
⎞⎜⎝
⎛⋅+≈ zDw
N2
2
eff 121
zDw⋅+ 2
2
12
zN
DwN
⋅⎟⎟⎠
⎞⎜⎜⎝
⎛−+⎟⎟
⎠
⎞⎜⎜⎝
⎛−≈
eff
22
eff
1112
11
12
2w [II]
[I]
[III]
Schematic: not to be scaledPa
d re
spon
se w
idth
squa
red
[I] : with respect to thetrue track coordinate
[II] : spatial resolution[III]: with respect to the
charge centroid
[III] = [I] – [II]
Brief Summary of Formulation
• w: known
• D: measured: determined from z dependence of pad-response width (or given by Monte-Carlo simulation)
• Neff: measured: determined from D and z dependence of spatial resolution(or estimated with assumptions on ionization statistics and avalanche
fluctuation)
• f (pad response function): not known (Monte-Carlo simulation?) δ-function is assumed for f.
• electronic noise ( ): measurable and different from pad to padconsidered here as a constant termindependent of z
NOISEσ
If w, D, f and Neff are known can readily be calculated for given z.Xσ
Examples of Calculated Resolution
Comparison with the real data
Data
zNDXX eff22
02 /+= σσ
m53129cmm/7.94.38/
0
eff
µσµ
±=
±=
X
ND
pad-pitch dominant region asymptotic regiondiffusion dominant
12/ mm3.2
Calculation
Data: MicroMEGAS
B = 1 TGas: Ar-isobutane (5%)
w = 2.3 mm
D = 469, 285 and 193 respectively for
B = 0, 0.5, and1.0 T(gas: Ar-isobutane (5%))
Neff = 27.5
f : δ function
Electronic noise: absent
cmm/µ
2PRF
22PR0
22PR0
2PR 12/with σσσσ +=+≈ wzD
Asymptotic behavior at long drift distances Expectation vs. Data
Pad Response (B = 1 T)
Spatial Resolution (B = 1 T)
2NOISE
2
eff
2X0
2
eff
2X0
2X 12
1with1 σσσσ +=+≈w
NzD
N
MWPC GEM MicroMEGAS Gas TDR TDR P5 Ar-isobutane (5%)
)m(PR0 µσ 1390 443±5 511±2 781±79
( )m12/ µw 663
(w = 2.3 mm) 367
(w = 1.27 mm) 663
(w = 2.3 mm)
Ｄ ( )cmm/µ 220 207±1 168±1 198±15
Ｄ [MAGBOLTZ] 200 200 166 193
MWPC GEM MicroMEGAS Gas TDR TDR P5 Ar-isobutane (5%)
)m(X0 µσ 220 81±6 67±6 129±53
( )m12
1
eff
µwN
135 67 71 128
( )cmm/eff
µND 45 38±1 32±1 38±10
effN 24 30±2 27±2 27±14
PRF dominant
small S/N ratio
Calculated Resolution for LC TPC
B = 3 Tcm/m82µ=D
5.27eff =N
It is important to reduce pad-pitchdominant region in Linear Collider TPC.
Reduced pad pitch with a larger number ofreadout channels
or
Use of resistive anode technique(see the next slide.)
PRF: δ - function
Charge dispersion in a MPGD with a resistive anode
•Modified GEM anode with a high resistivity film bonded to a readout plane with an insulating spacer.•2-dimensional continuous RC network defined by material properties & geometry.•Point charge at r = 0 & t = 0disperses with time.•Time dependent anode charge density sampled by readout pads.Equation for surface charge density function on the 2-dim. continuous RC network:
∂ρ∂t
=1
RC
∂2ρ∂r2 +
1r
∂ρ∂r
⎡
⎣ ⎢
⎤
⎦ ⎥
⇒ ρ(r, t) =RC
2t
−r 2RC4 te
ρ(r,t) integral over pads
ρ(r) Q
r / mmmm ns
Concept & first results:M.S.Dixit et.al., Nucl. Instrum. Methods A518 (2004) 721
TPC resolution with charge dispersionMicromegas-TPC with a resistive anode readout
ILC-TPC collaboration, KEK beam test (October 2005)
2 x 6 mm2 pads (B = 1 T) Ar/C4H10 95/5 Cd = 125 µ m⁄ √cm (Edrift= 80V/cm)
Drift distance
Tran
sver
se re
solu
tion
(mm
)
121cmm/ 125 m)547(
eff
0
±==
±=
ND
X
µ
µσ
Preliminary
Summary
TPC prototypes with MPGD readout have been operated successfully.Spatial resolution is understood in terms of pad pitch, diffusion, PRF, and the effective number of electrons (Neff).The expected spatial resolution can be estimated by a simple numerical calculation (NOT a Monte-Carlo) for given geometry, gas mixture and PRF if the relevant parameters are known.It is essential to make pad pitch small, physically or effectively, in order to minimize the pad-pitch dominant volume in the TPC.Resistive anode technique reduces the effective pad pitch, thereby eliminating the pad-pitch dominant region.
Future ProspectsPerformance tests with a larger prototype at DESY using atest beam and the super conducting magnet, PCMAG.Search for better gas mixtures.
high ionization statistics (electron yield)small diffusionlarge (ω )τsmall avalanche fluctuationhigh electron mobilitysmall fraction of hydrocarbons
Ar – CF4 (- isobutane) ?
Future Prospects (Cont’d)
Measurement of the pad response function and avalanche fluctuation using single electrons for better understanding of the pad response and spatial resolution.Study of positive ion backflow.Detailed Monte-Carlo simulation of the electron transportincluding avalanche process for the development and geometry optimization of detection devices.Inclusion of UV-photon emissions in avalanches?Positive ion buildup in the case of GEM?Development of (surface mount) readout electronics.Digital TPC?