P. Lecoq CERN 1November 2010 Workshop on Timing Detectors - Cracow 2010

Time of Flight: the scintillator

perspective

Paul LecoqCERN, Geneva

P. Lecoq CERNNovember 2010 2Workshop on Timing Detectors - Cracow 2010

Where is the limit?Philips and Siemens TOF PET achieve

– 550 to 650ps timing resolution – About 9cm localization along the LOR

Can we approach the limit of 100ps (1.5cm)?

Can scintillators satisfy this goal?

P. Lecoq CERNNovember 2010 3Workshop on Timing Detectors - Cracow 2010

For the scintillator the important parameters are– Time structure of the pulse– Light yield– Light transport

affecting pulse shape, photon statistics and LY

€

Δt ∝ τN phe ENF

Timing parameters

decay time of the fast component

Photodetectorexcess noise factor

number of photoelectrons generated by the fast component

General assumption , based on Hyman theory

P. Lecoq CERNNovember 2010 4Workshop on Timing Detectors - Cracow 2010

Light output: LYSO example

Statistics on about 1000 LYSO pixels 2x2x20mm3 – produced by CPI – for the ClearPEM-Sonic project

(CERIMED)

Mean value = 18615 ph/MeV For 511 KeV and 25%QE:

2378 phe Assuming ENF= 1.1

Nphe/ENF ≈ 2200 phe

P. Lecoq CERN 5November 2010 Workshop on Timing Detectors - Cracow 2010

t = 40 ns

Nphe

t = 40 nsNphe

Nphe

Nphe

Statistical limit on timing resolution

W(Q,t) is the time interval distribution between photoelectrons = the probability density that the interval between event Q-1 and event Q is t

= time resolution when the signal is triggered on the Qth photoelectron

LSO

Nphe=2200€

WQ t( ) =

N pheQ × 1− e

− tτ

⎛

⎝ ⎜ ⎜

⎞

⎠ ⎟ ⎟

Q−1

exp −N phe 1− e− t

τ ⎛

⎝ ⎜ ⎜

⎞

⎠ ⎟ ⎟

⎡

⎣ ⎢ ⎢

⎤

⎦ ⎥ ⎥e

− tτ

τ Q −1( ) !

P. Lecoq CERNNovember 2010 6Workshop on Timing Detectors - Cracow 2010

Light generation

€

y(t) = Ae−

tτ

€

N phe = y(t)dt = Aτ0

∞

∫ Rare Earth4f

5d

P. Lecoq CERNNovember 2010 7Workshop on Timing Detectors - Cracow 2010

Rise time is as important as decay timeRise time

€

y(t) = Ae− t

τ d 1− e− t

τ r ⎛ ⎝ ⎜

⎞ ⎠ ⎟

P. Lecoq CERNNovember 2010 8Workshop on Timing Detectors - Cracow 2010

Photon counting approach

LYSO, 2200pe detected, td=40ns

tr=0ns tr=0.2nstr=0.5ns tr=1ns

P. Lecoq CERNNovember 2010 9Workshop on Timing Detectors - Cracow 2010

Cross-Luminescent crystals (very fast, low LY)– BaF2 (1400ph/MeV) but 600ps decay time produces more

photons in the first ns (1100) than LSO (670)! Direct bandgap semiconductors S. Derenzo, SCINT2001

– Sub-ns band-to-band recombination in ZnO, CuI,PbI2, HgI2

Nanocrystals– Bright and sub-ns emission due to quantum confinement

Faster than Ce3+?Intrinsic limit at 17ns

Pr3+

– Pr3+ 5d-4f transition is always 1.55eV higher than for Ce3+

€

t Pr

τ Ce≈ λ Pr

λ Ce

⎛ ⎝ ⎜

⎞ ⎠ ⎟3

≈ 3.5eV3.5eV +1.55eV ⎛ ⎝ ⎜

⎞ ⎠ ⎟3

= 13

P. Lecoq CERNNovember 2010 10Workshop on Timing Detectors - Cracow 2010

Material Density (g/cm3)

Radiation length X0

(cm)

Refraction index n

Critical angle

Fondamental absorption

(nm)

Cerenkov threshold

energy for e (KeV)

Recoil e range

above C threshold

(mm)

# C photons / 511KeV g ray

*

PbWO4 8.28 0.89 2.2 63° 370 63 513 21

LSO:Ce 7.4 1.14 1.82 57° 190 101 527 15

LuAG:Ce 6.73 1.41 1.84 57° 177 97 582 22

LuAP:Ce 8.34 1.1 1.95 59° 146 84 487 28

Ultimately fast using Cerenkov emission?

Even low enegy g ray produce Cerenkov emission in dense, high n materials

This emission is instantaneous with a 1/l2 spectrum

* Low wavelength cut-off set at 250nm for calculations on LSO, LuAG and LuAP Ce absorption bands subtracted from Cerenkov transparency window

P. Lecoq CERNNovember 2010 11Workshop on Timing Detectors - Cracow 2010

22Na

PMT left (2150V) PMT right (1500V)

LuAG 2013 (undoped -> shows no scintillation) LSO 1121

8cm 8cm

Crystals wrapped on5 sides with teflon.

Scope

Coincidence:Th_left=-4mV, th_right=-500mV

CFD

LuAG Cerenkov/LYSO Scintillation coincidence measurement

FWHM=374psLuAG=259ps

FWHM=650psLuAG=587ps

P. Lecoq CERNNovember 2010 12Workshop on Timing Detectors - Cracow 2010

Light Transport

– -49° < θ < 49° Fast forward detection 17.2%– 131° < θ < 229° Delayed back detection 17.2%– 57° < θ < 123° Fast escape on the sides 54.5% – 49° < θ < 57° and 123° < θ < 131° infinite bouncing 11.1%

For a 2x2x20 mm3 LSO crystalMaximum time spread related to

difference in travel path is424 ps peak to peak

≈162 ps FWHM

P. Lecoq CERNNovember 2010 13Workshop on Timing Detectors - Cracow 2010

Photonic crystals to improve light extraction

Periodic medium allowing to couple light propagation modes inside and outside the crystal

M. Kronberger, E. Auffray, P. Lecoq, Probing the concept of Photonics Crystals on Scintillating Materials TNS on Nucl. Sc. Vol.55, Nb3, June 2008, p. 1102-1106

24% 34%

P. Lecoq CERNNovember 2010 14Workshop on Timing Detectors - Cracow 2010

LuAP Light gain2.1

LYSO Light gain2.08

BGO Light gain2.11

LuAG:Ce Light gain1.92

Expected Light Output Gain for different crystals

Litrani + CAMFR simulation

P. Lecoq CERNNovember 2010 15Workshop on Timing Detectors - Cracow 2010

How does the PhC work?

Section of the plane crystal- air interface: (EM – fieldplot)

Crystal- air interface with PhC grating:

θ>θc

Total Reflection at the interface Extracted Modeθ>θc

Diffracted modes interfere constructively in the PhC- grating and are therefore able to escape the Crystal

P. Lecoq CERNNovember 2010 16Workshop on Timing Detectors - Cracow 2010

PhC fabrication

Nano Lithography

PhC is produced in cooperation with the INL (Institut des Nanotechnologies de Lyon)

Three step approach:1. Sputter deposition of an auxiliary layer2. Electron beam lithography (EBL)3. Reactive ion etching (RIE)

RAITH® lithography kit:

P. Lecoq CERNNovember 2010 17Workshop on Timing Detectors - Cracow 2010

PhC fabrication

Reactive Ion Etching (RIE)

1. Chemically reactive plasma removes Si3N4 not covered by the resist

2. Change the composition of the reactive plasma to remove the resist (PMMA) without etching the Si3N4

x

z

y

Scintillator

ITOSi3N4

a

Hole depth: 300nm

hole diameter: 200nm

x

z

yScintillator

ITOSi3N4

Ion Bombardment

PMMA Resist

P. Lecoq CERNNovember 2010 18Workshop on Timing Detectors - Cracow 2010

PhC fabrication

Results

Scanning Electron Images:a = 340nm

D = 200nm

P. Lecoq CERNNovember 2010 19Workshop on Timing Detectors - Cracow 2010

Use larger LYSO crystal: 10x10mm2 to avoid edge effects

6 different patches (2.6mm x 1.2mm) and 1 (1.2mm x 0.3mm) of different PhC patterns

PhC first results

0° 45°

Preliminary

P. Lecoq CERNNovember 2010 20Workshop on Timing Detectors - Cracow 2010

PhC improves light extraction eficiency

But also collimation of the extracted light

P. Lecoq CERNNovember 2010 21Workshop on Timing Detectors - Cracow 2010

Conclusions Timing resolution improves with lower threshold Ultimate resolution implies single photon counting High light yield is mandatory

– 100’000ph/MeV achievable with scintillators Short decay time

– 15-20ns is the limit for bright scintillators (LaBr3)– 1ns achievable but with poor LY

Crossluminescent materials Severely quenched self-activated scintillators

SHORT RISE TIME– Difficult to break the barrier of 100ps

P. Lecoq CERNNovember 2010 22Workshop on Timing Detectors - Cracow 2010

New approaches?

Conclusions

Crystals with a highly populated donor band (ZnO)

Metamaterials loaded with quantum dots

Make use of Cerenkov light

Improve light collection with photonic crystals

P. Lecoq CERNNovember 2010 23Workshop on Timing Detectors - Cracow 2010

Our TeamCERN

– Etiennette Auffray– Stefan Gundacker– Hartmut Hillemanns– Pierre Jarron– Arno Knapitsch– Paul Lecoq– Tom Meyer– Kristof Pauwels– François Powolny

Nanotechnology Institute, Lyon– Jean-Louis Leclercq– Xavier Letartre– Christian Seassal