3-D simulation of Si-ProtCharge distribution, signal development

D. Attié, P. Colas, E. Delagnes, S. Turnbull

- Introduction

- Orders of magnitude

- Simulation

- Results

- Data

April 17, 2008 RD51 workshop 2

Introduction• A single spark kills a TimePix chip

– High current, hot plasma -> destroys circuit– A sufficiently thick SiProt layer reduces and softens the sparks

(See M.Fransen’s talk)

• How protection works? – Allows charge to stay over the pad, lowering the local potential– Limits the current thru the amplifier– Protects mechanically– Avoids points, softens the surface?

• Any damage to the signal?– Amplitude loss?– Charge sharing with neighbouring pads?– Short-circuit of the amplifiers– Introduces dead time?

TO ANSWER, NEED A SIMULATION

April 17, 2008 RD51 workshop 3

Orders 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

April 17, 2008 RD51 workshop 4

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

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

April 17, 2008 RD51 workshop 6

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

April 17, 2008 RD51 workshop 7

Signal development Charge pulse: 1ns trapezoidal 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

April 17, 2008 RD51 workshop 8

ResultsLarge 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

0 1 2 3 4

5 6 7 8 9

10 11 12 13 14

15 16 17 18 19

20 21 22 23 24

Pads fully covered with Al

April 17, 2008 RD51 workshop 9

Results

0 1 2 3 4

5 6 7 8 9

10 11 12 13 14

15 16 17 18 19

20 21 22 23 24

"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.

April 17, 2008 RD51 workshop 10

thickness Signal frac. on central pad

(Full pads)

Signal frac. on central pad

(cross)

10 µ 94 % 76 %

15 µ 83 % 57 %

20 µ 70 % 44 %

Charge injected in the middle of the pad

April 17, 2008 RD51 workshop 11

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

April 17, 2008 RD51 workshop 12

Data

• With 20 µ SiProt, the spreading over several pads is clearly seen, but the grid misalignment also contributes.

April 17, 2008 RD51 workshop 13

Data

With 15 micron SiProt, pixel multiplicity of isolated clusters shows 91% of single pixels at a gain of 4000.

Difficult to relate precisely to the measurement, but no evidence for strong spreading.

InGrid-equipped TimePix, Nikhef

April 17, 2008 RD51 workshop 14

Conclusions• Charge spreading is limited both in space and

time for SiProt thickness less than 10-15 m.• Fully covered anodes spread less than partially

covered anodes

• SaclayDavid Attié

Max Chefdeville

Paul Colas

Xavier Coppolani

Eric DelagnesArnaud GiganonIoannis Giomataris

Marc Riallot

Stephen Turnbull

• CERNMichael CampbellXavier Llopart

• NikhefYevgen BilevychMarten BosmaMax ChefdevilleMartin Fransen Fred HartjesJan Timmermans Harry van der Graaf Jan Visschers

TimePix-Micromegas collaboration