Abstract 1. Free Electron Lasers 2. X

PixFEL: Sviluppo di rivelatori a pixel per
applicazioni a macchine Free Electron Laser
G. Batignani, S. Bettarini, G. Casarosa, F. Forti, "
F. Morsani, A. Paladino, E. Paoloni, G. Rizzo, Università e INFN, Pisa
Abstract 3. PixFEL Long Term Goals Develop a four side bu/able module for a large area X-­‐ray camera for FELs using a mulAlayer device: acAve edge thick pixel sensor, two-­‐Aer CMOS readout chip (analog+digital/memory) with low/high density TSV, 65 nm technology to increase memory and funcAonality, in a small pixel pitch of 100 μm. Il proge/o PixFEL ha lo scopo di sviluppare un rivelatore per il piano focale per la rivelazione di raggi X alla prossima generazione di macchine Free Electron Laser, con la funzione di ricostruire le immagini di diffrazione prodo/e dai fasci coerenA ad alta brillanza. Le applicazioni sono molteplici, e spaziano dalla fisica dei materiali alla biologia. Nel proge/o PixFEL si vogliono migliorare le prestazioni dei rive-­‐latori esistenA uAlizzando tecnologie avanzate, come ele/ronica CMOS a 65nm, integrazione verAcale, e sensori di silicio edgeless, puntando a realizzare degli elemenA di rivelazione con cui costruire un mosaico di grandi dimensioni e zone morte ridoLssime, in grado di rivelare fotoni tra 1 e 10 keV con un range dinamico di 104 fotoni per pixel. Nel lungo periodo la collaborazione PixFEL vuole sviluppare una X-­‐ray camera versaAle che possa essere operata sia in modo impulsato che in modo conAnuo alle future macchine FEL come Eu-­‐XFEL o LCLS-­‐II Good efficiency up to 10 keV (450 um thick)
9 bit resolu-on (effec-ve), 5 MHz sampling rate 1. Free Electron Lasers wide dynamic H&1.,(" ^^VR" 3-6+*,(2" '%(" +(*A" ?,&1%'0(22" -5" 3-0D(0'&-0*/" /*2(,2" *04" '%(&," (<'(02&-02"
&0'-"
'%(" range (1 to 10000 photons), single photon sensi-vity ./',*D&-/('",(1&6(":&'%"2J03%,-',-0",*4&*'&-0"2-.,3(2"*04"UHTW2S""
Coherent X-­‐Ray beams with very high intensity and large energy range. Very short (fs-­‐as) and brilliant pulses: burst and con-nuous mode opera-on 1 kframe •  Burst mode: rapid sequence with long interval for readout. EU-XFEL
in
input
out
" Cf=Cmin, Gain=Gle
Based on the non linear feature of a
MOSFET operated in inversion mode
out|<<VTh
" Cf=Cmax, Gain=Ghe
High voltage needed to have good collecAon efficiency even with the plasma effect caused by the large number of photons. Vdd
out|<<VTh
Today: LCLS, …
•  |Δv
•  |Δv
AcAve edge, to allow Aling 450 μm thickness for efficiency OpAmized edge geometry: Suitable choice of W and L to configure the
gain in the low and high energy regime, under
the constraint set by the preamplifier
output range
Silicon pixel sensors D. Comotti, “Low noise readout channel with a novel dynamic
signal compression for future XFEL applications”, N18-2
•  ConAnuous mode: up to 1 MHz pulses. L. Ratti, “PixFEL: advanced X-ray pixel cameras for the next generation FELs”, 2014 IEEE NSS-MIC, Seattle, USA
4. PixFEL Technologies Dynamic compression with MOS capaci
W/L
Burst and conAnuous mode beam Ame C#N-($&**4IK"
structure Peak brilliance vs. Energy for sources compression amplifier )(*A" ?,&//&*03(" -," ?,&1%'0(22" D(,2.2" +%-'-0" (0(,1J" -5" 3-0D(0'&-0*/" /*2(,2" Dynamic *04" %&1%(,"
Q
in
%*,6-0&3" 1(0(,*'&-0" 2-.,3(2" cW*2(,28" ?/.(" (0D(/-+(d" 8" 2J03%,-',-0" 2-.,3(2" c>J03%,-',-028"
1,*J"
R oo
Non-­‐linear M
OS c
aps f
or h
igh d
ynamic range electrons
*04"e<V,*J"
5,(("Frame/Burst (/(3',-0" /*2(,2"rcUVHTW28"
,(4" (0D(/-+(d"
c5,-6"
Electron beam (0D(/-+(d"
Photon nergy epe--on Number X " 7S" 9//,&3%8" #S" O.4(0A-8" OS"
= 60 e-­‐RMS @ 50ns shaping #00S" O(DS" )%J2S" =%(6S"
PO8"
[]Y" cRX!RddS" N-'("
'%*'" :&'%" W=W>V^^" '%(" 2+(3',*/" ENC ,*01(" &2"
energy [GeV] L-2%*66(,8"
[keV]
rate [Hz]
pulses/burst
1,(*'/J"&03,(*2(4"'-"RYX"(Q"k"!_"A(Q"*04"'%("+(*A"?,&//&*03("(<3((42"'%*'"-5"-'%(,"2-.,3(2S""
NMOS
in
out
Start of opera-on
2009
14.5
SACLA@RINKEN
2011
8
European-­‐XFEL
2015
SwissFEL
LCLS II
2.4
0.03-­‐0.3
5
800@1us
0.3-­‐10
120
1
input
2016
0.01-­‐0.06
10
1
FELS in operaAon or under construcAon "
imaging
2. X-­‐Ray Protein
FEL Detectors Challenges Measurement principle: diffracAon Frame rate Using extremely
and
intense
X-ray
pulses
to
Single shot short
imaging (<100fs pulse, 200 ns readout) capture images of objects such as proteins before the XFrame storage of complete bunch train (2700 pulses) rays destroy
the sample.
for Eu-­‐XFEL or fast readout for conAnuous machine (up to 1MHz) Single-molecule diffractive imaging with an X-ray freeDynamic electron
laser.Range Single photon counAng Individual
biological molecules will be made to fall
Up to 104 photons/pixel/pulse X-ray pulses
Diffraction
pattern
holes
in
PMOS
input
out
simulated performance << Vth
Cf = Cmin
Readout architecture • High energy photons
Vout
Burst mode operaAon Focal plane
oo
In-­‐pixel 10 bit SAR ADC @ 5MHz • Low energy photons Preamp schemaAc and UHTW2"*,("4&2'&03'"5,-6"3-0D(0'&-0*/"/*2(,2"*04"2J03%,-',-0",*4&*'&-0"2-.,3(2"&0"'%(".04(,/J&01"
4.5-­‐15
60
1
sampling +%J2&3*/" +,&03&+/(2" *04" '%(" +,-+(,'&(2" -5" '%(" /&1%'S=-0D(0'&-0*/" /*2(,2" *,(" ?*2(4" -0" /&1%'V
"
(6&''&01" (/(3',-0&3" ',*02&'&-02" ?(':((0" 4&2'&03'" g.*0'.6" 2'*'(2" &0" '%(" 3%-2(0" /*2(," 6(4&.6S"
17.5
0.4-­‐20
10
2700@220 ns
#6+/&5&3*'&-0"-5"'%("',*02&'&-0"&0'(02&'J"&2"*3%&(D(4"?J"?,&01&01"6*0J"(/(3',-02"&0'-"'%("2*6("
5.8 (<3&'(4"2'*'("c+-+./*'&-0"&0D(,2&-0d"*04"'%(0"2'&6./*'&01"6*0J"(/(3',-02"'-".04(,1-"'%("2*6("
1/12
100
2@50 ns
/*2(,2"
4-­‐14.5 ,*4&*'&D(" ',*02&'&-0S"
0.2-­‐25 H,((" (/(3',-0" 120 -­‐ 1*,("
06 ?*2(4" -0" ,*4&*'&-0"
1 (6&''(4" ?J" ,(/*'&D&2'&3" 5,(("
>2020
Particle injection
Power = 350 μW/pixel R
600
400
200
0
Eq. Capacitance 0(pF) E
InterconnecAons 10
1
5. PerspecAves 4 of 25
•  The PixFEL 3-­‐year demonstrator project has been approved by INFN –  explore applicaAon of innovaAve technologies to an X-­‐ray FEL imaging camera: acAve edge sensors, 65 nm, Through Silicon Vias, 3D integraAon •  Great potenAality and synergy in cross-­‐field ferAlizaAon Hardness TheRadiaAon pulse will ultimately
destroy each molecule, but not
10 M
Gy-­‐1 G
Gy odiffracted
ver 3 years from
operaAon before the
pulse
has
the undamaged
structure.
Pixel size: •  Longer term ambiAous plan to develop full instrument –  group mostly coming from HEP, learning its ways in X-­‐ray imaging –  need to interact with user community and exisAng detector development groups 700-­‐20 are
μm, combined
depending oto
n dform
istance angular The patterns
an&atomic-resolution
resoluAon required image of the molecule.
2000
Vout >> Vth
Cf = Cmax
Bandwidth needs for XFEL: 4.5 MHz frame Bump-­‐bonding for sensor to 0
!!"
10
rate, 1% duty cycle, 1k frames stored over 3k Charge Sensitive Amplifier Gain
preamp connecAon frames à 0.6 Gb/s/chip & 20 Gb/s/ladder dVThrough-­‐silicon out
vias for -1chip to G
=
ConAnuous operaAon 10
dph
With low repeAAon rate, OK. At 1MHz, very chip connecAon Equivalent feedback capacitance
Both high and low density TSV challenging. -2
⇥ 1
10
dVout
1
10
Cef = 280q
dph
through the X-ray beam, one at a time, and their
SensiAve Energy recorded
Range in the form of a
structural
information
0.25-­‐25 keV ideally with the same system diffraction
pattern.
Ambiguity in reconstructed images of a virus Large area coverage: coming from missing data due to dead area or The speed record of 25 femtoseconds for flash imaging
mulAple Ales with no dead area saturaAon of pixel. was achieved.
Cmax
LCLS@SLAC
2010
1.25
Vout
Ghe
2005
Q in
9
FLASH@DESY
Fermi@ELETTRA
Vout
800
Gain [mV/ph]
Project
gnd
PMOS 20
VB=VDD
Gle
"
65nm CMOS readout Gain (1000
mv/ph) Cmin
""""""""""""""""""""""""""""""
W/L
CSA Output Voltage [mV]
Tomorrow: LCLS II [email protected]
Ene