UFSD: Ultra-Fast Silicon Detector

Size: px

Start display at page:

Download "UFSD: Ultra-Fast Silicon Detector"

Merry Maxwell
5 years ago
Views:

1 UFSD: Ultra-Fast Silicon Detector Basic goals of UFSD A parameterization of time resolution State of the art How to do better Overview of the sensor design First Results Nicolo Cartiglia with M. Baselga, M. Bruzzi, S. Ely2, V. Fadeyev, Z. Galloway, F. Marchetto, J. Ngo, M. Obertino, C. Parker, H. F.-W. Sadrozinski, M. Scaringella, D. Schumacher, A. Seiden, A. Vinattieri, A. Zatserklyaniy 1

2 UFSD goals Create a silicon detector with a factor of ~ 10 larger signal 1) ps time resolution: large signals allow for much better timing capabilities 2) Very high rate capability: excellent counters for charge particles 3) µm thin detector: ideal for low mass system 4) Very rad-hard: charge trapping has a moderate effect Here I examine 1) & 2) 3) & 4) are a direct consequence of larger signal ratio 2

3 Current Silicon sensor R&D Sensors This is what we want to do Radiation Resistance Integrated Low Mass Large signal - Vert. Electr., 3D - n-in-n - n-in-p - Silicon Substrate (FZ, CZ) - Diamond - CMOS 3T, 4T pixel - Semi-monolithic, DEPFET - CMOS monolithic, MAPS - SOI - HV CMOS, LePix - 3D IC - Thinned - Self Supporting - Active Edge - Vertical Vias - Micro-channel cooling - UltraThin - ps precision - GHz rate - Very rad-hard 3

4 UFSD: a time-tagging detector Pixel Pre-Amplifier Time measuring circuit Time is set when the signal crosses the comparator threshold The timing capabilities are determined by the characteristics of the signal at the output of the pre-amplifier and by the TDC binning: σ Total 2 = σ Jitter 2 + σ Time Walk 2 + σ TDC 2 4

5 Time walk and Time jitter Time walk: the voltage value Vo is reached at different time for signal of different amplitudes =! t V $ rise th t # & " S % σ TW RMS Jitter: the noise is summed to the signal, causing amplitude variations σ t J = N dv dt Due to the physics of signal formation (see backup slides for full calculation and reduction techniques) Mostly due to electronic noise (see backup slides for capacitance and noise values used) 5

6 Details of Collected Charge in Sensors To calculate time walk, we need a parameterization of the energy loss 11 JINST 6 P06013 S. Meroli, D. Passeri and L. Servoli 6 Note: In thin sensors, the yield of e-/h pairs is lower, and the signal amplitude spread is higher Hartmut Sadrozinski, Hiroshima Conference è Thin detectors have larger time walk

7 Time walk calculation Signals cross a given threshold with a delay that depends on their amplitude, on the rise time and on the value of the threshold: t delay = t rise V th V The rms of the time delay is a measure of the time walk. σ (t delay ) = ps t delay The CFD reduces this effect by at least a factor of 3. 7

[ff] Depends on the pitch and thickness N: Noise at preamp. Dominated by the voltage term S: Signal.

8 A parameterization of σ t σ 2 t = t 2! $ rise # & + t V 2!' * $! rise th ), " S / N # % "( S & + TDC $ bin # & + % " 12 % Jitter d: detector thickness [micron] l: pitch [micron] C: Detector capacitance [ff] Depends on the pitch and thickness N: Noise at preamp. Dominated by the voltage term S: Signal. Time Walk Proportional to the thickness t rise : Pre-Amp Shaping time V th : Comparator threshold Depends on the noise level TDC:Width of the TDC LSB [ps] l RMS d C Det = εε o l *l d N C Det t rise S = 75e *d V th =10* N LSB = 20 2 (1) TDC + 0.2*4l +50 8

9 What is the best shaping time (t rise )? % ' S & (' t rise Const N C Det t rise if " " t rise t col if " " t rise > t col V th N ) ' ' % ' ' * σ t & ' ' ' (' ' + C det t rise if " " t rise t col if C det * t rise " " t rise > t col To minimize time resolution: t rise = t col Note: This value also minimizes fake signals in neighboring pixels. Time Resolution [ps] Detector thickness: 100 micron Collection Time = 1250 ps Shaping time = Collection time Shaping Time [ps] 9

State of the Art Following eq. (1) (Using NA62 for normalization) Time Resolution [ps] 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 180.0 160.0 140.0 120.0 100.0 80.

10 State of the Art Following eq. (1) (Using NA62 for normalization) Time Resolution [ps] Pixel = 300 micron Gain = 1 Below d = 100 micron: signal too small Shaping Time [ns] Time walk / Jitter 0.0 Time resolution Detector Thickness [micron] NA62 Best resolution achievable: ~ 100 ps (assuming Time Walk reduction of ~ 3)

11 State of the Art Following eq. (1) (Using NA62 for normalization) Shaping Time [ns] Time Resolution [ps] Time walk / Jitter 0.0 Time resolution Detector Thickness [micron] Best resolution achievable: ~ 50 ps (assuming Time Walk reduction of ~ 3) Pixel = 100 micron Gain =

12 σ t 2 =! # " How can we do better? t rise S / N $ & % 2!' + t V * rise th #), "( S + RMS $ & % 2! + TDC bin # " 12 $ & % 2 Boost the signal: introduce gain S => G * S Impact ionization model: if the electric field is high enough, the carriers are multiplied according to: In Silicon, at 270kV/cm: α e 0.7 pair/µm α h 0.1 pair/µm N ( ) = N 0 *e (α* ) = G * N 0 (see backup slides for details) 12

Add protections: to avoid electrical breakdown Thin the detector: to reach very

13 UFSD Basic concept Add a p+ implant underneath the n+ electrode: a large area with high electric field where multiplication can be achieved. Add protections: to avoid electrical breakdown Thin the detector: to reach very high rates Gain ~ 10 (A low gain device) The design of the sensor is the main goal of the UFSD project 13

14 The effect of gain Time Resolution [ps] Time resolution Pixel = 300 um Gain 10 Gain = Gain Detector Thickness [micron] Effect of gain: thinner detectors and smaller time resolution Best resolution achievable ~ 15 ps (assuming Time Walk reduction of ~ 3)

15 Rate [MHz] Resolution for 100 and 300 µm pixel Time Resolution [ps] Shaping Time [ns] Higher rate Best Resolution Gain = 10 Pixel 300 um Pixel 100 um Detector Thickness [micron] High rate capability requires thinner detectors Small time resolution requires thicker detectors Time resolution Rate capability

16 No Time Walk Correction Rate [MHz] Time Resolution [ps] Shaping Time [ns] Detector Thickness [micron] Pixel = 100 micron Gain = 10 Time walk Jitter Time resolution Excellent timing resolution even without Time Walk correction è Much simpler electronics

17 Pixel size [µm] UFSD - Summary 300 σ t ~ 50 ps Rate ~ 1 GHz Rate ~ 100 MHz 100 No need for Time Walk correction 50 σ t ~ 10 ps Sensor Thickness [µm] 17

UFSD: The PPS project at LHC Can we measure the exclusive production

Standard event: pp à many particles PPS Current design for Low

Cherenkov radiators Central Exclusive Production: pp à X p p PPS

18 UFSD: The PPS project at LHC Can we measure the exclusive production of new particles at LHC? Standard event: pp à many particles PPS Current design for Low Luminosity: Position: Pixel silicon detector Timing: Quartz Cherenkov radiators Central Exclusive Production: pp à X p p PPS Upgrade for High Luminosity: Timing and Position: UFSD A much lighter detector allows collecting more data near the beam line 18

19 UFSD Gran plan We propose to realize a low-gain ultra-fast sensor that will concurrently measure: Time ~ ps: It depends on gain, noise levels and pixel size Space ~ µm: Edge effects in the p+ implant might be important, forcing larger pixels. Rate ~ GHz: The rate is determined by the detector thickness collection time = shaping time. 19

20 Backup slides 20

21 Charge Multiplication 21 Charge multiplication in path length l : N( ) α e, h = N 0 *exp( α * ) = g * N ( ) = ( e, h E α ) e, h *exp E At the breakdown field in Si of 270kV/ cm: α e 0.7 pair/µm α h 0.1 pair/µm gain g = 33 possible in l = 5 µm. In the linear mode (gain ~10), consider electrons only b 0 Need to raise E-field as close to breakdown field as possible for high gain but not too much to prevent breakdown! A. Macchiolo,16th RD50 Workshop Barcelona, Spain, May 2010

22 Noise vs Shaping Time Pixel size Noise [e-] 100 um 200 um 300 um NA Shaping Time [ps ] Noise values used in the parameterizations (NA62: Shaping time ~ 5500 ps, noise ~ 300 e-) 22

23 Capacitance vs Detector Thickness Pixel size 100 um Capacitance [ff] 200 um 300 um Detector Thickness [micron ] Capacitance: backplane ff/µm (perimeter) + 20 ff (fix term) 23

24 Are TDC fast enough? 24

UFSD: Ultra-Fast Silicon Detector

UFSD: Ultra-Fast Silicon Detector Basic goals of UFSD (aka Low-Gain Avalanche Diode) A parameterization of time resolution State of the art How to do better Overview of the sensor design Example of application