Post on 03-Apr-2015
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 11
RESISTRESIST
Etat des lieux, bulk Etat des lieux, bulk résistif, techniques résistif, techniques
disponiblesdisponibles
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 22
Introduction : butsIntroduction : buts Améliorer la résolution par un Améliorer la résolution par un
meilleur partage de chargemeilleur partage de charge Protéger le détecteur et Protéger le détecteur et
l’électronique contre les étincelles l’électronique contre les étincelles
Ceci est nécessaire pour obtenir la résolution voulue pour la TPC de l’ILC (100 microns) sans multiplier le nombre de canaux, et ceci peut être utile également pour T2K, si des problèmes de stabilité apparaissaient ou si une résolution meilleure que 500 microns s’avérait utile.
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 33
Théorie analytique de la Théorie analytique de la résolutionrésolution
Avec nos collègues japonais, nous avons mis au point une théorie analytique de la résolution:
Effets de fluctuation d’ionisation, de fluctuation du gain, et de taille finie des pads, limitent la résolution.
En raison du manque de partage de la charge entre pads voisins la résolution est, au mieux, de 10% de la largeur des pads
1/√12
1/√(12.Neff)
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 44
Extrapolation à la TPC de l’ILCExtrapolation à la TPC de l’ILC
Conclusion: même avec un pas de 1mm, Micromegas avec des pads standard ne permettrait pas d’atteindre le but de 130 m de résolution.
Cependant avec une feuille résisitive et des pads de 2.3 mm, ce but est largement atteint (courbe rouge).
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 55
Feuille Feuille résistiverésistiveUne feuille résistive collée sur le plan Une feuille résistive collée sur le plan
d’anode en fait un réseau de d’anode en fait un réseau de résistances et capacités qui étale la résistances et capacités qui étale la charge charge Proc. of LCWS 2005
On obtient une résolution de On obtient une résolution de 50 50 micronsmicrons à petite dérive à petite dérive
(et à toute dérive pour B assez (et à toute dérive pour B assez grand)grand)
Faisceau test à KEK B=1T
Micromegas avec feuille résistive
Q
ns
Cosmiques en champ magnétique
Dapnia 07-36, soumis à NIM A
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 66
Bulk résisitifBulk résisitif La technologie bulk La technologie bulk NIM A 560 (2006) 405NIM A 560 (2006) 405
permet une grande robustesse et permet une grande robustesse et l’absence de zones mortes (pas de cadre)l’absence de zones mortes (pas de cadre)
Idée: la combiner à la technologie du Idée: la combiner à la technologie du résistif résistif
RESIST : RESIST : 12 personnes+1 postdoc+ST12 personnes+1 postdoc+ST
Tests en labo: bulk résistif, mesures de Tests en labo: bulk résistif, mesures de résistivité, uniformité, etc… Nouveau labo résistivité, uniformité, etc… Nouveau labo opérationnel mi-novembreopérationnel mi-novembre
Tests en faisceau ou cosmique au CERN Tests en faisceau ou cosmique au CERN et/ou TRIUMF ou Fermilab et à DESY.et/ou TRIUMF ou Fermilab et à DESY.
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 77
TECHNIQUESTECHNIQUES
Film Sheldhal- AlSi cermet + colle Film Sheldhal- AlSi cermet + colle flobond + procédé drytac (Carleton)flobond + procédé drytac (Carleton)
Pâte résistive (CERN, Rui de Oliveira)Pâte résistive (CERN, Rui de Oliveira) Couche déposée (ASiH, AsiP) (MIT Couche déposée (ASiH, AsiP) (MIT
unine, Nicolas Wyrsch)unine, Nicolas Wyrsch)
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 88
Several techniques are being tested for the Several techniques are being tested for the resistive coatingresistive coating
1) Carbon-loaded kapton. An old technique first tested in Carleton, using a Dupont film 3 MOhm/sq. Improvements on application of the resistive foil and switch to bulk.
First results promising. One panel recently produced.
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 99
2) Prepreg+ screen printing
This has been tried at CERN. 2 prototypes of 10x10 cm (2 and 8 MOhm/sq) have been tried at Saclay. There is not clear evidence that they are spark protected. Even one of the detector has been damaged by the HV during the test. Still such a layer will be applied to a CERN panel.
comparison screen printing with insulator
1000
10000
100000
390 400 410 420 430 440 450 460 470
V mesh
gain
gain R3i
gain R4i
tin
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 1010
Plasma deposition of thin layers (N. Wyrsch, Neuchatel, used for SiProt)
Preliminary tests going on. Next step: make a small bulk 12x14 cm2
with 2 layers of different resistivity, and then cover a PCB with pads and make a bulk out of it
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 1111
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 1212
In-situ TEST at DESYIn-situ TEST at DESY
Gas-box
Trigger counters
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 1313
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 1414
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 1515
New detector with resistive New detector with resistive foil in the gas boxfoil in the gas box
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 1616
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 1717
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 1818
La TPC du Large Prototype est presque prête pour prendre des données
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 1919
Took data with P5 gas with standard pads Took data with P5 gas with standard pads on October 2 and 3 at 50 MHz sampling on October 2 and 3 at 50 MHz sampling rate, and 200ns peaking time.rate, and 200ns peaking time.
Tested the whole chain again on October Tested the whole chain again on October 27, and switched to the new module (with 27, and switched to the new module (with C-loaded kapton resistive foil) in the test C-loaded kapton resistive foil) in the test box (took about 2 hours)box (took about 2 hours)
Took data since October 27 with P5 Took data since October 27 with P5 50 MHz sampling rate, 200 ns peaking time50 MHz sampling rate, 200 ns peaking time 100 MHz sampling rate, 200 ns peaking time100 MHz sampling rate, 200 ns peaking time 50 MHz sampling rate, 400 ns peaking time50 MHz sampling rate, 400 ns peaking time
Switched then to T2K gas Switched then to T2K gas (Ar/CF4/isobutane:95:3:2)(Ar/CF4/isobutane:95:3:2)
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 2020
plansplansTake beam data in the magnet in the period of weeks 46-47-48 (+1 or 2?)
- Install and commission beam trigger (today)
- Mount standard panel + 6 dummies on the endplate
- Finalize and connect cathode
- Cool down magnet
- See how to set HV (1st ring to 360-370 V)
- Order gas : 25 bar per bottle, 1 bottle for 18 hours at 60 l/h, 1 bottle in 55 hours at 20 l/h.
- Add bubbler
- Excite magnet on November 10
- Then start beam data taking and analysis
- switch to resistive anode week 47
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 2121
Simulating and understanding Simulating and understanding SiProtSiProt
A single spark kills a TimePix chipA single spark kills a TimePix chip High current, hot plasma -> destroys circuitHigh current, hot plasma -> destroys circuit A sufficiently thick SiProt layer reduces and softens the A sufficiently thick SiProt layer reduces and softens the
sparkssparks
How protection works?How protection works? Allows charge to stay over the pad, lowering the local Allows charge to stay over the pad, lowering the local
potentialpotential Limits the current thru the amplifierLimits the current thru the amplifier Protects mechanicallyProtects mechanically Avoids points, softens the surface?Avoids points, softens the surface?
Any damage to the signal?Any damage to the signal? Amplitude loss?Amplitude loss? Charge sharing with neighbouring pads?Charge sharing with neighbouring pads? Short-circuit of the amplifiersShort-circuit of the amplifiers Introduces dead time?Introduces dead time?
TO ANSWER, NEED A SIMULATION
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 2222
Orders of magnitudeOrders of magnitude
R neighbour
R thru
e
lL
Rneighb = L/el = /e for square pads
For 10 µ aSi, =1011 .cm, R=1014 /square
Rthru = e/Ll ~ 0.04 1014
Rneighb/Rthru = L2/e2
e<<L to avoid spreading the charge by side conductivity. The charge preferentially escapes thru the pad, but with an extremely high resistance.
Cneighb = r Le/l =r /e for square pads
Cthru = r Ll/e (note r ~11 for Si)
The influence acts preferentially thru the pad if L>>e
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 2323
R neighbour
R thru
e
lL
Also Rthru >> 1/C to leave the induced signal
(OK with 1011 cm within 3 orders of magnitude for 10 ns signals)
Equivalently, RC time constant >> 10ns
The signal is fully capacitive
1/Cthru
1/Cneighbour
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 2424
5 pixels
5 pixels
gas
Siprot
25 Pads
25 TimePixPreamps
gas: 10x10x10elements/pad
SiProt: 10x10x10elements/pad
« central » gas element.C network
(special ones fo edges)
« central » SiProt elementRC network
(special ones fo edges)Pads:Metallisation of the last SiProt elements
« Full » shape100% of pixel
aluminized
« Cross » shape12% of pixel Al,
rest SiN2
Charge Injection at the gas / SiProt interface in the central pixel Size = 4 elts 3 positions.
Top= HT (ac grounded)
1
23
SiProt Simulation with Analog ArtistE. Delagnes, S.Turnbull
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 2525
Note that SiProt is very different from the charge spreading for improving the resolution (M. Dixit et al.).
In that case, there is an insulating layer (50-100 µ) below a low-resistivity coating: 1 M/sq, 100 nm (i.e. 10 .cm). In that case, the charge obeys the telegraph equation (see Stephen Turnbull’s talk). In SiProt, the signal is capacitive and the charge slowly evacuates in thru the coating and the amplifier.
This Analog Artist simulation can also be used for double layers with 2 different resistivities.
Run time for SiProt: 10 h per set of parameters.
25 AFTERPreamps
gas
resistive
insulator
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 2626
Signal Signal developmentdevelopment
Charge pulse: 1ns Charge pulse: 1ns trapezoidal current pulsetrapezoidal current pulse
Small Cross vs Large Pads; Timepix Chip Simulation, 15um resistive layer.
-9,00E-04
4,91E-02
9,91E-02
1,49E-01
1,99E-01
2,49E-01
2,99E-01
3,49E-01
0,00E+00 1,00E-06 2,00E-06 3,00E-06 4,00E-06 5,00E-06 6,00E-06
Time (s)
Hig
h (
un
its
)
12 13 14 18 19 20 21 24 12 full 13 full
Injection in the middle
Injection in the corner
Injection in between
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 2727
ResultsResultsLarge Pad : injection in the middle of pad
0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1
0 5 10 15 20
Pad
Norm
aliz
ed s
ignal
FULL 10µm
FULL 15µm
FULL 20µm
00 11 22 33 44
55 66 77 88 99
1100
1111
1122
1133
1144
1155
1166
1177
1188
1199
2200
2211
2222
2233
2244
Pads fully covered with Al
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 2828
ResultsResults
00 11 22 33 44
55 66 77 88 99
1100
1111
1122
1133
1144
1155
1166
1177
1188
1199
2200
2211
2222
2233
2244
"Cross" Pad : injection in the middle of pad
0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1
0 5 10 15 20
Pad
Norm
aliz
ed s
ignal
Cross 10µm
Cross 15µm
Cross 20µm
Pads 12% covered.
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 2929
thicknesthicknesss
Signal frac. Signal frac. on central on central
padpad
(Full pads)(Full pads)
Signal frac. on Signal frac. on central padcentral pad
(cross)(cross)
10 µ10 µ 94 %94 % 76 %76 %
15 µ15 µ 83 %83 % 57 % 57 %
20 µ20 µ 70 %70 % 44 %44 %
Charge injected in the middle of the pad
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 3030
PRFs
0
0,1
0,2
0,3
0,4
0,5
0,6
0 0,5 1 1,5 2 2,5 3 3,5 4
distance of injection from pad center (in fraction of pad pitch)
Mea
sure
d C
har
ge Full 10µm
Cross 10µm
Full 15µm
Cross 15µm
Full 20µm
Cross 20µm
Pad Response Function
RESIST - 03 Nov 2008RESIST - 03 Nov 2008 P. ColasP. Colas 3131
The first resistive bulk works to perfection.
The first cosmics seen show a charge spreading as expected