Transcript (Y.Sugimoto) - JLC

Study of FPCCD
Vertex Detector
12 Jul. 2005
@8th ACFA WS
Y. Sugimoto
KEK
FPCCD Vertex Detector







Accumulate hit signals for one train and read out between trains
(Other options read out 20 times per train)
Fine pixel of ~5mm (x20 more pixels than “standard” pixels) to
keep low pixel occupancy
Fully depleted epitaxial layer to minimize the number of hit pixels
due to charge spread by diffusion
Two layers in proximity make a doublet (super layer) to minimize
the wrong-tracking probability due to multiple scattering
Tracking capability with single layer using cluster shape can help
background rejection
Three doublets (6 CCD layers) make the detector
Operation at low temperature to keep dark current negligible (r.o.
cycle=200ms)
Tracking efficiency

Large number of background hits may cause
tracking inefficiency: mis-identification of signal hit
with background hit
For a normal incident track, the probability
of mis-identification of hit is given by;
pmis  2R0 2 , R0  d 0
 : Background hit density
0 : Multiple scattering angle
Angular and momentum dependence;
pmis  p 2 sin 4 
Tracking efficiency
pmis
Mis-identification Probability (p=1 GeV/c,tSi=50mm)
d=10mm, =40/mm2
d=10mm, =2/mm2
Expected hit density
40/mm2/train (TESLA)
20/mm2/train (Nominal)
15/mm2/train (LowQ)
at R=20 mm, B=3 T
d=2mm, =40/mm2
cos
Cluster shape
Standard pixel
FPCCD
(tracking capability with single layer!)

Effective background rejection will be possible by
requiring angles of hit clusters in two layers of a doublet
and the angle of the line connecting the two hits to be
consistent with the extrapolated track.
Baseline design
Advantages of FPCCD





Free from beam-induced RF noise
x~1.4mm even with digital readout
Simple structure : advantageous for large size
Active circuit on one edge : easy to control
temperature
Readout speed: 15MHz is enough
(128(V)x20000(H)/200ms=12.8MHz)
(CPCCD:>50MHz)
CCD
MAPS/FAPS/ISIS
Impact parameter resolution


Full simulation study by Jupiter
Parameters:
 Old standard option:




FPCCD option:




R=24, 36, 48, 60 mm
t = 330 mm/layer
 = 4 mm
R=20, 22, 39, 41, 58, 60 mm
t = 80 mm/layer
 = 2 mm
Be beam pipe (common):


R=18 mm
t = 500 mm
Impact parameter resolution
p=1, 3, 5, 10, 20, 50, 100 GeV/c
T. Nagamine
Lorentz angle

Lorentz angle in depleted-layer


tan=mnB
mn: electron mobility
Carrier velocity saturates at
high E field:




mn =0.07
@T=300K, E=1x104V/cm
mn =0.045 m2/Vs
@T=300K, E=2x104V/cm
m2/Vs
Small angle can be cancelled
by tilting the wafer
May not be a serious problem

Number of hit pixels does not
increase so much
B=3T
B=5T
E=1x104V/cm
=12deg
=19deg
E=2x104V/cm
=7.7deg
=13deg
Lorentz angle

Calculation of E-field in epi-layer


Tools
 FEMLAB (COMSOL in Japan) 3.1
 Solve Poisson equation by finite element analysis (FEA)
Parameters
 Material is assumed fully depleted (No free charge)
 n-layer: ND=1x1016/cm3=1.6x103 C/m3
 Epi-layer: NA=1x1013/cm3=-1.6 C/m3
 VG=4 V
 tSiO2=100 nm
 tn=1 mm
 tepi=15 mm
Lorentz angle

Result of E-field calculation
Lorentz angle

Result of E-field calculation – Potential
Lorentz angle

Result of E-field calculation – Summary



Almost constant E-field of ~104V/cm in epi-layer can be
achieved
E-field in epi-layer depends on gate voltage
 Higher (positive) gate voltage gives higher E-field
 Positive gate voltage should be applied during train
crossing in order to get saturated carrier velocity and less
Lorentz angle (Inverted (MPP) mode can be maintained for
~1ms)
The Lorentz angle of 12 degrees is expected at B=3T
Summary and outlook




We propose FPCCD option for the ILC Vertex Detector
 Fully depleted CCD with 5mm-square fine pixel size
 Accumulate 2820 BX and readout between trains
 Two layers make a doublet (super layer) to pick up signal
hits out of background hits
If the thickness of the wafers is less than 80 mm, the impact
parameter resolution better than b = 5  10/(pbsin3/2) mm can
be achieved with Rin=20 mm and B=3T
Electric field in the epitaxial layer has been calculated. From the
results, the Lorentz angle of ~12 degrees at B=3T is expected
Tracking efficiency under beam background is the most critical
issue for FPCCD. Simulation study is urgent.