High Energy All Sky Transient Radiation Observatory HE-ASTRO

1
High Energy All Sky Transient Radiation
ObservatoryHE-ASTRO
W
A
E

By V. Vassiliev, S. Fegan, A. Weinstein
Cherenkov 2005   27-29 April 2005 - Ecole
Polytechnique, Palaiseau, France
2
Scientific Motivations in the realm of GLAST
epoch

• Studies of
• very high energy transient phenomena
• In the Universe

Studies of the highest energy radiation
Galactic sources
3
Cosmological studies of High Energy Transient
Phenomenato determine

• Population properties of AGN and GRBs
• Redshift evolution of these objects
• Redshift evolution of EBL (z0-6)
• Major contributors to EBL (stars, dust, AGN,
  Population III objects, relic particles, SFR,
  GFR, IMF, BH accretion histories, supernovae
  feedback, merger history)
• Cosmological magnetic fields and their evolution
• High energy properties of space-time

4
Cosmological Diffuse Background
nIn nW m-2 sr-1
l mm
5
Universes Opacity
Redshift Z
No CDB Evolution is assumed
Energy of g-ray TeV
6
Collecting Area Requirement

• To resolve a few min variability time scale in
  emission of Mrk 421-like AGN placed at Z1, the
  collecting area of the observatory must be 1
  km2.

E interval g-Rate GeV
min-1 25 - 50 1.3 50 -100
0.7 100 - 200 0.3
3C273 Energetic Quasar Jet by Chandra
7
Collecting Area Requirement

• Because the size of the HE-ASTRO, 1010 cm2, is
  much larger than the size of the Cherenkov light
  pool, 108 cm2, the number of telescopes required
  is 102.

A
Coupling distance d80m
Area per telescope A31/2/ 2 d2 5.542x107cm2
Number of telescopes N2173n(n-1) 1 n9
8
HE-ASTRO (1st specs)

• Target Energy Range 20 200 GeV
• Array of 217 telescopes (n9)
• Area 1.0km2
• Effective Area 1.6km2 (including boundary
  events)
• Detection Rates x 100Hz (flaring nearby AGN)
• x 1-10Hz (quiescent
  moderately distant AGN)
• x few per min
  (cosmologically distant AGN)
• Total Cost lt 200M (approximate cost of satellite
  mission)
• Cost per station lt 1M

9
Problem 1(energy vs area)
Although large aperture costly telescopes seem to
be favored for low energy range prerequisite,
they are incompatible with collecting area
constraint due to cost limitations
10
Solid Angle Requirement
Observing Mode
Sky Survey
RCR 300 kHz
Un-localized Source
RCR 30 kHz
RCR 3 kHz
Known Source
11
Problems 2 3

• A requirement of large solid angle coverage, p,
  by the system of a few large aperture telescopes
  possesses a very difficult engineering problem
  and costly large aperture secondary optics .

Sustaining high data rates in non-distributed
system, such as a few large aperture telescopes,
is another challenging electronics problem
12
HE-ASTRO (2nd specs)

• Energy Range 20200 GeV
• Array of 217 telescopes
• Area 1.0km2 (1.6km2)
• Field of View 15o
• FoV area 177 deg2
• Reflector Diameter 7m
• Reflector Area 40 m2
• Cost per station lt 1M

Energy Range
Cost per Telescope
13
GalacticHE Astrophysics
New Phase Space in E, T

• Max Acceleration Energy
• Crab Nebula
• 1 g/min gt 10 TeV 2 g/hour gt 100 TeV
• HE pulsars and transients

New Phase Space in F
If number of sources is a matter of sensitivity
in already probed energy domain ,100 GeV 10
TeV, then collecting area of 1km2 is a figure of
merit.
Horan Weekes
14
HE-ASTRO Challenge
Installation at A.E.R.E., Harwell 1962 -

• Can telescopes be relatively small, have
  moderately large FoV, be fairly inexpensive, and
  have differential peak photon detection rate at
  40 GeV?

15
Understanding Collecting Area
Cell Operation Mode
QEElevation or Dish Size
Concept of IACT-cell Aharonian at el.
Astroparticle Physics 6 (1997) 343-377
16
Important Angular Scales
a ?
a ?
Trigger Pixel Size Trigger Efficiency
Image Pixel Size Reconstruction Efficiency
17
Challenges

• Trigger
• Efficiency at low energies (peak of
• detection rate around 30 40 GeV)
• High Data rates
• Array trigger
• Low Cherenkov light level regime
• High angular resolution
• Background regection efficiency
• Engineering cost issues
• Light Sensors
• Wide Field of View Optics
• Costs

18
Trigger Pixel Optimization

• Simulations

19
Wavelength Response(assumption)
Double reflection from aluminized coated mirror
Relevant wavelength response window 200 400
nm Range of implied QEs 0.5 1.0 (50 100 )
Cherenkov Light Spectrum x (300 nm/l)2
Efficiency
Wavelength nm
20
Summary of parameters
(nth-Nnsb) / Q
Trigger Pixel Size degree
Reflector diameter D7 m NSB integration window
t20 ns NSB differential flux 0.4

• QE 0.5 1.0
• FoV 10 15 degrees
• Rnsb 0.1 1.0 kHz

21
Lowest trigger threshold(Effects of QE, Rnsb,
FoV)
Number of photons collected by telescope in
trigger pixel in 20 ns from a g-ray shower to
trigger pixel of a given size
QE0.5, FoV15o, Rnsb0.1 kHz
QE0.5, FoV10o, Rnsb0.1 kHz
(nth-Nnsb) / Q
QE0.5, FoV15o, Rnsb1.0 kHz
QE0.5, FoV10o, Rnsb1.0 kHz
QE1.0, FoV15o, Rnsb0.1 kHz
Effects
QE1.0, FoV10o, Rnsb0.1 kHz
1) QE
QE1.0, FoV15o, Rnsb1.0 kHz
2) Rnsb
QE0.5, FoV10o, Rnsb1.0 kHz
3) FoV
Trigger Pixel Size degree
22
Trigger Efficiency vs Pixel Size(central
telescope)
QE 1.0, D7m QE 0.5, D10m
QE 1.0, D7m QE 0.5, D10m
El 3.5 km
El 4.5 km
QE 0.5, D7m
QE 0.5, D7m
Trigger Efficiency
Trigger Pixel Size degree
Optimum Trigger Sensor pixel size
range 0.07o-0.25o
Parameters Eg42 GeV FoV15o Rnsb1kHz
Weakly Depends on QE, D, El
23
Efficiency versus Pixel Size (Array)
Array Trigger Three telescopes above
operational threshold
p0.05o
p0.08o
p0.10o
p0.13o
Array Parameters Elevation 3.5 km QE
0.5 Reflector 7 m FoV 15o
p0.16o
p0.20o
Efficiency
Efficiency gt 50 p0.05o 0.25o for E gt 20-30
GeV
Photon Energy GeV
24
Single Telescope Trigger Efficiency
Diff. spectral index 2.5
12 GeV
Diff. Rate
15 GeV
Trigger Efficiency
20 GeV
27 GeV
El4.5km, QE 1.0, D7m El4.5km, QE 0.5,
D10m El3.5km, QE 1.0, D7m El3.5km, QE 0.5,
D10m El3.5km, QE 1.0, D7m El3.5km, QE 0.5,
D10m
Photon Energy GeV
Photon Energy GeV
Effects 1) Cell operation mode 2) Optimum
trigger pixel size 3) QE, Reflector Size 4)
Elevation 5) Rnsb
Parameters Trigger pixel size 0.146o Obs. Mode
Un-localized Source
(FoV15o) Rnsb 1kHz
25
Single Telescope Trigger Efficiency
Diff. spectral index 2.5
14 GeV
Diff. Rate
17 GeV
Trigger Efficiency
22 GeV
Rnsb 100kHz Rnsb 10kHz Rnsb 1kHz
Photon Energy GeV
Photon Energy GeV
Effects 1) Cell operation mode 2) Optimum
trigger pixel size 3) QE, Reflector Size 4)
Elevation 5) Rnsb
Parameters Trigger pixel size 0.146o Obs. Mode
Known Source (FoV3.5o)
26
Array trigger
El4.5km, QE 1.0, D7m El4.5km, QE 0.5, D10m
El3.5km, QE 1.0, D7m El3.5km, QE 0.5, D10m
Average Number of Telescopes in Trigger

• Trigger Pixel size 0.146o

Photon Energy GeV
27
Single telescope CR rates
28
HE-ASTRO (3rd specs)

• Array of 217 telescopes
• Elevation 3.5km
• Telescopes coupling distance 80m
• Area 1.0km2 (1.6km2)
• Single Telescope Field of View 15o
• FoV area 177 deg2
• Reflector Diameter 7m
• Reflector Area 40 m2
• QE 50 (200-400 nm)
• Trigger sensor pixel size 0.146o
• Trigger Sensor Size 31.2cm
• NSB rate per Trigger pixel 3.2 pe per 20 ns
• Single Telescope NSB Trigger Rate 1KHz
• Energy Range 20200 GeV
• Differential Detection Rate Peak 30 GeV
• Single Telescope CR trigger rate 30 kHz

29
Event Reconstruction

• Angular scales

30
An event
g at 42 GeV
31
Voronoi Diagrams
Event 1 (42 GeV)
Event 2 (42 GeV)
32
g-g Separation ScalesImage NSB
g 21 GeV NSB 150 g/deg2
g 42 GeV NSB 150 g/deg2
Diff. density Arbitrary
g-g separation deg
g 100 GeV NSB 150 g/deg2
QE 0.5 Reflector Diameter 7m Elevation 3.5
km Trigger pixel size 0.146o
Voronoi Diagram g-g separation scales in
Image 0.01o-0.04o
33
Reconstruction Method
di , pi
uncertainty
34
Event Reconstruction(arrival direction)
ltd(qx,qy)gt ?di2/ ?1i
35
Cleaning and arrival direction reconstruction
Optimum cut 4 photons within circle of 0.02o
radius
36
An Event Cleaning
37
Ln(ng) vs. ltpgt
q lt 0.2o
No q cut
21 GeV 42 GeV 100 GeV CR
Ln(Ng) 1
Mean arrival distance m
q lt 0.05o

• The height of shower development
• is an effective discriminating factor
• which can be utilized within
• paradigm of IACT arrays
• e discrimination ?
• Roadmap to energy estimate!

38
ltpgt cut
21 GeV 42 GeV 100 GeV CR
21 GeV 42 GeV 100 GeV
Integrated flux p/deg2/min
Integrated flux p/deg2/min
Mean arrival distance m
g containment fraction
21 GeV 42 GeV 100 GeV
21 GeV 42 GeV 100 GeV
S/NS/SRT(B) arbitrary
S/NS/SRT(B) arbitrary
Mean arrival distance m
g containment fraction
39
Event Cleaning CRs
Ep127 GeV
40
Event Cleaning Photons
Eg42 GeV
Eg100 GeV
41
Ng vs. ltdgt
Regime A few hundred photons per event collected
by array
21 GeV 42 GeV 100 GeV CR
Ng 1
Mean cascade radius ltdgt m
42
Problem 4(number of pixels)

• Both, identification of primary particle and
  reconstruction of its arrival direction can be
  accomplished in 40-50 GeV energy domain.
• However, the imaging resolution scales required
  are in the range 0.01o-0.02o
• Optimal pixel size for triggering and imaging
  differ by a factor of 10.

43
Hardware implementation

• Approach
• Technologies
• Data Rates Array Trigger

44
1963
Japan Suga Italy ? Russia ?
45
The Story of Ground Based g-ray Astronomy (by
Jelley Porter)
46
Energy range gt 1 TeV Readout Event Rate lt 1kHz
All sky covered with 80 mega pixels in the CMOS
sensor arrays Optimized Baker-Nunn optical system
with three corrector normal lenses made of
acrylic resin and 1 m spherical reflector (spot
size less than 1 arcmin, 0.016o, for parallel
light rays incident at angles less than
25o). Focal sphere image intensifier, FIIT, of
60 cm aperture
47
Focal Plane Instrumentation
Fast random access CMOS sensor Image pixel size
0.0146o Readout image 128 x 128 pixels Readout
Image size 1.875o x 1.875o Readout rate 30-40
kHz
Optical or II-based delay
Two-mirror reflecting or one-mirror catadioptric o
ptical system
X-Y
Optical Splitter
Gate Shutter
X ?xi Y ?yi
OR
Gated Image Intensifier (MCP) (25-40 mm) Gate
20ns Rep. rate 40kHz P-43/P-24 , 2msec
Star tracker VETO
Large Aperture Image Intensifier (Electrostatic
or MCP) Photon detection efficiency 50 Fast
decay scintillator output screen 25 ns
Trigger Sensor 8200 pixels with 0.146o
Primary 7m Fov 15o
Array of rate compensated discriminators
Slow Control
48
CMOS Image Sensor
HE-ASTRO Image Sensor is not commercially
available yet. However, industry is very close
to meet specification. High-speed readout is
achieved with pipeline and parallel technologies.

Parallel processing macro-cell of 32x32 pixels
(1024) can be readout with gt 500 kHz, and 128 x
128 pixel image (16 macro-cells) with gt 30 kHz
Micron-MT9M413 1.3-Megapixel CMOS Active Pixel
Digital Image Sensor
Image pixel size 0.0146o Readout image 128 x
128 pixels Readout Image size 1.875o x
1.875o NSB per pixel 0.032 (20 nsec gate) ADC
8 bit (S/N improved, 10 gt8) Pixel dimension 12mm
x 12mm Sensor area 12.3 mm x 12.3 mm Shutter
exposure a few msec
49
Ultrafast Imaging
DRS technologies Inc. Variety of Ultrafast
Cameras for Military
applications CCD based 500fps to 100,000,000fps
e.g. 350 KHz at 250 x 250 Pixel exposures
from 5 nsec limited number of frames
120mm tank gun projectile
Photron CMOS based high speed cameras Ultima
APX-RS is the world's fastest video camera with
3,000 mega pixel frames per second (fps) or
250,000 fps at reduced resolution FASTCAM-X
1024 PCI is the first system to bring mega-pixel
CMOS to your personal computer at usable speeds
capable of operating as fast as 1,000 fps at
full 1,024 by 1,024 pixel resolution, or 109,500
fps through windowing. Ultima APX-i2 uses a
25mm MCP Gen II image intensifier, directly
bonded onto the APX's mega pixel sensor to
provide unmatched image quality with 20ns
gating.
airgun pellet impacting a matchstick
50
Gated Image Intensifiers

Type No. Effective Ares Gate TIme Repetition Rate Photocathode MCP Number
C9546-01 18 mm 3 ns, 30 kHz GaAsP 1
C9546-02 18 mm 3 ns, 30 kHz GaAsP 2
C9546-03 18 mm 3 ns, 30 kHz Multialkali 1
C9546-04 18 mm 3 ns, 30 kHz Multialkali 2
C9547-01 25 mm 5 ns, 30 kHz GaAsP 1
C9547-02 25 mm 5 ns, 30 kHz GaAsP 2
C9547-03 25 mm 10 ns, 30 kHz Multialkali 1
C9547-04 25 mm 10 ns, 30 kHz Multialkali 2
Available 25 mm 20 ns, 2 kHz GaAsP or Multialkali 1 or 2
Available 25 mm 100 us, 100 Hz GaAsP or Multialkali 1 or 2
Available 40 mm 20 ns, 2 kHz Multialkali 1 or 2
Available 40 mm 100 us, 100 Hz Multialkali 1 or 2
Left C9016-2x Series Controller Center
C9546 Series Right C9547 Series
Hamamatsu products
Commercial products which almost satisfy
requirements of resolution, repetition rate, and
fast gating exist.
51
Trigger Sensor
Hamamatsu H9500 Flat Panel 52mm square Bialkali
Photocathode 16 x 16 Multianode 12 stage FoV
15o Trigger pixel size 0.146o Number of MAPMTs
32 Effective Area Ratio 89 Size 312 mm
52
Optical Systemone-mirror catadioptric
Due to a large number of telescopes in HE-ASTRO
array, 217, the monitoring of all sky can be
distributed alleviating very large FoV problem
for a single station. With a single telescope
FoV15o each event is viewed by 7 telescopes in
the sky 30o above horizon 19 telescopes in the
sky 45o above horizon
f/0.7 system to reduce size of secondary
optics Lens to 1m,
Optimized Schmidt-like Baker-Nunn optical system
with 2-4 active surfaces will satisfy optical
requirements. However, the design for D7m
FoV15o optical system must be cost effective,
e.g. Schmidt camera with large image angle is
probably prohibitively expensive due to large
costs of making 7m Schmidt corrector plate.
Another deficiency of Schmidt camera is
strongly curved focal surface and its position
inside the instrument.
Plate Scale Optical Coupler
Focal surface coupled II
7.5 deg
Low Loss Refractive Corrective Optics Aperture-1m
7m Primary, possibly discontinuous
Large aperture instruments with wide FoV are
usually prohibitively expensive
53
Optical SystemRitchey-Chrétien configuration
Design targets 1) Short tube length 2) Small
secondary mirror 3) Flat and accessible
focal surface
Cassegrain-like two-mirror design which is free
of spherical aberration and coma. Dominant
remaining aberrations are due to astigmatism
and field curvature
Two Hyperbolic mirrors system
Generalized Schwarzschild theorem For any
geometry with reasonable separations between the
optical elements, it is possible to correct n
primary aberrations with n powered elements.
(1905)
Traditional for Cherenkov telescopes Davies-Cotton
reflector compensates spherical aberrations by
discontinuous mirror. Discontinuous primary and
possibly secondary need to be explored for
astigmatism and field curvature compensation
Field curvature coupled II Field curvature is
inward curving concave toward the sky. Difficult
but may be not impossible to compensate it with
II design.
Primary 7m Fov 15o
Optical System RD is required
54
Focal Plane Image Intensifier
Electrostatic Image Intensifiers
MCP Image Intensifiers
Photek manufactures a range of 18, 25, 40, 75
and 150 mm active diameter image intensifiers.
SIEMENS image intensifiers. Large aperture units
(gt40cm) are developed for X-ray imaging.
Phosphor Scintillator P-47 - 80 ns lt10gt decay
time Lanthanum Bromide Scintillator, LaBr3 /
LaCl3 - 25 ns lt10gt decay time
High QE photocathode in 200-400 nm, 50,
continues to be an issue
55
Telescope data pipeline
80Mb/s
TD Veto Stars
Trigger Sensor
Disk
Retrieve Image to disk
L1
Position encoding X ?xi, Y ?yi Timing T
Array trigger L2 Broadcast
10 ms Memory Ring Buffer of Images Indexed by
local trigger timestamp
T
II (int. 20ns)
OR L1i
X,Y
delay
Timing Event Identification
L2 Telescope Trigger 40 kHz
Gated Image Intensifier P-43 2msec
Gate 20 ns
2.3kb per image
Shatter 2msec
Zero suppression
20kb per image
1024x1024 pixels
Readout 128x128 Pixels 1.9o x 1.9o ADC 10
bits/pixel, 16384 pixels
600 non-zero pixels (mostly NSB) Bitmask 20
bits ADC 10 bits
1.0 Mpixel CMOS Image Sensor (1000-500 fps
full frame)
15o x 15o
Data rate 80 Mb/s x 3600 s/h x 217 telescopes
62.5 Tb / array / hour
56
Array Trigger

• Distributed ?Every node acts as its own array
  trigger

• Telescope Trigger decision (30 kHz)
• Local trigger ? convert to GPS timestamp (good to
  100ns)
• Buffer timestamp locally
• Broadcast trigger packet of timestamp and node
  identifier (5 bytes) to all nearest and
  next-nearest neighbors (max. 2Mbps outflow rate)

• Local trigger together with any trigger of two
  telescopes from all neighbors and next
  nearest-neighbors is recognized as array trigger
• Local processing at node
• Receive trigger timestamps
• Buffer trigger timestamps (10-20 µs)
• Search for a coincidence (compensate for relative
  delays due to pointing)
• Coincidence ? retrieve pixel data, write to disk
  (90 Mb/s)

Data rate to center node 24 Mbps _at_ 30 kHz
57
HE-ASTRO (specs)

• Array of 217 telescopes
• Elevation 3.5km
• Telescopes coupling distance 80m
• Area 1.0km2 (1.6km2)
• Single Telescope Field of View 15o
• FoV area 177 deg2
• Reflector Diameter 7m
• Reflector Area 40 m2
• QE 50 (200-400 nm)
• Trigger sensor pixel size 0.146o
• Trigger Sensor Size 31.2cm
• NSB rate per Trigger pixel 3.2 pe
• per 20 ns
• Single Telescope NSB Trigger Rate 1KHz
• Energy Range 20200 GeV
• Differential Detection Rate Peak
• 30 GeV
• Single Telescope CR trigger rate
• 30 kHz

• Image pixel size 0.0146o
• Readout image 128 x 128 pixels
• Readout Image size
• 1.875o x 1.875o
• NSB per pixel 0.032 (20 nsec gate)
• ADC 8 bit (S/N improved,
• 10 gt8)
• Pixel dimension 12mm x 12mm
• Sensor area 12.3 mm x 12.3 mm
• Shutter exposure a few msec
• Image integration time - 20 ns
• Optical system TBD
• Array trigger protocol TBD
• Data Rates 80 Mb/secper node
• Online data processing TBD

58
Conclusions I

• Perhaps, sufficiently reach scientific goal,
  which can justify spending on the scale 100M,
  is studying high energy transient phenomena at
  cosmological distances and galactic phenomena
  with sensitivity improved by a factor of ten
  comparing to current IACTAs.
• Large array of moderate size telescopes may
  provide a viable cost effective solution to the
  problem of required large collecting area, large
  field of view, and low energy threshold at the
  same time, by combining new and reviving old
  ideas of using image intensifiers but based on
  the contemporary technology.
• Triggering scheme of individual moderate size
  telescopes with FoV15o and trigger pixel size
  of 0.15o is compatible with operation at 20-30
  GeV energy of the peak detection rate if array of
  telescopes is located at gt3.5 km elevation and
  QE50 D7m or QE25 D10m.
• Predicted CR background rates of 30 kHz can be
  already maintained with the use of commercially
  available technology.
• By using high resolution high speed CMOS imaging
  an arrival direction of photons can be
  reconstructed with high accuracy (21 GeV, 16.2
  arcmin 42 GeV 10.2 arcmin 100 GeV 3.6 arcmin
  within 50 g-containment radius).
• The potential of high resolution imaging for CR
  rejection in low energy light poor regime by a
  factor of 10-100 is very promising as well as
  accurate integrated g-ray event energy estimation
  
• To finalize feasibility of such approach several
  more detailed design studies are necessary
  optics, readout electronics, array trigger,
  background rejection efficiency, geomagnetic
  effect, and practicability of real time image
  processing in hardware or software to reduce data
  rates?

59
Conclusions II

• Small array of very large dish size telescopes
  sure guarantees low energy
• One can see development
• of the field when several such projects might be
  pursued in the future
• in Germany and France
• in US, UK, and Ireland
• in Japan
• In
• with price tag 3-5 times that of VERITAS or HESS
  each, 50M
• Is this the only option for the money?

Sensitivity of IACTAs

• There are alternatives which have not been
  explored yet those which could require broader
  international collaboration to accomplish a
  project with the price tag of 200M and
  utilizing technology that has been viewed as one
  with the greatest promise of potential advances
  in the field since its foundation.

60
Conclusions III

• Wide field of view, very high image resolution,
  and super-fast parallel data processing may
  provide technological basis for the next
  breakthrough ground-based g-ray astronomy.