Transcript WFXT_Pareschi

WFXT optics: design optimization and
development
Giovanni Pareschi
JKCS041:, z = 1.8, Andreon et al., 2009
1
Acknowledgement
The WFXT optics team at OAB
Paolo Conconi, Sergio Campana , Oberto Citterio, Marta Civitani, Vincenzo
Cotroneo , Giovanni Pareschi, Laura Proserpio, Gianpiero Tagliaferri , Giancarlo
Parodi, (BCVProgetti)
Thank you to the whole WFXT collaboration for supporting this work and
for many useful discussions!
• R. Giacconi, A. Ptak, C. Norman – JHU
• S. Murray, A. Vikhlinin – CfA
• M. Weisskopf, R. Elsner, S. O’Dell, B. Ramsey – NASA/MSFC
• S. Borgani, P. Rosati, P. Tozzi – INAF/OATrieste
• S. Molendi – INAF/IASF- Milano
ASI is supporting the pre-Phase A study in the context of the contract “High Energy
Astrophysics Studies” . INAF is also funding the activities with internal resources.
2
Outline
 Introduction the optical design of wide-field X-ray telescopes
 Optical design & Optimization of the WFXT mirrors
 A few remarks on manufacturing and implementation of the WFXT
optics
3
WFXT payload top level requirements
• Number of X-ray optics modules: 3
• Total payload (optics + detectors) mass: 1440 kg
4
Grasp
Grasp = Aeff x FOV
 measured at 1.5 keV in cm2 deg2
• Grasp measures the speed in which a survey can cover an area of the sky down to
a given flux limit.
• Better angular resolution results in better efficiency and source identification.
WFX eROSITA
T*
XMM
ROSAT
IXO
Chandra
Grasp
(cm2 deg2)
9000
1150
900
630
1500
50
HEW across
5/10
20-40
15-25
15-40
~5
1-5
the field
(arcsec)
5
Wolter’s solution to the X-ray imaging
H. Wolter, Ann. Der Phys., NY10,94
6
Wolter I optical system
7
Wolter I Point Spread Function (PSF)
ABRIXAS/e-Rosita
140
25’
20’
120
15’
10’
HEW [arcsec]
100
41’
5’
0’
80
60
40
20
ABRIXAS “cross scan”
Credits: MPE
calculation
0
0
5
10
15
20
Off-axis angle [arcmin]
25
R. Giacconi, “AN EDUCATION IN ASTRONOMY”, ARAA.
2005.43: 1- 30, 22
“A further extension of this line of thinking is that experiments could be designed by
modelling both the hardware and software as part of the initial design. I myself,
together with Richard Burg and Chris Burrows, used this approach in designing in the
1980s what I believe was one of the best experiments I ever conceived. The purpose
was to scan the sky and to detect distant clusters of galaxies through their X-ray
emission.
The idea was that it would be possible to equal or exceed the sensitivity of Chandra
with an X-ray telescope of one tenth the area (and cost). This could be achieved by
dedicating an entire mission of a small satellite to this purpose and by designing a
telescope that would have a >16-fold increase of the field of view with respect to
Chandra. ……..”
X-ray optics with polynomial profile: general remarks
• Mirrors are usually built in the Wolter I (paraboloid-hyperboloid) configuration
which provides, in principle, perfect on-axis images.
• This design exhibits no spherical aberration on-axis but suffers from field
curvature, coma and astigmatism,
 rapid degradation with increasing off-axis angles
• More general mirror designs than Wolter's exist
 the primary and secondary mirrors are expanded as a power series
• Optimization of polynomials
increase the angular resolution at large off-axis positions but degrading the
on-axis performances
See Burrows, Burgh and Giacconi (1992)
10
Why short mirror lengths for WFXT
•
The aspect ratio mirror shell length / focal length plays a very important role; in general
the height of the shells should be kept as short as possible
•
With short mirror shells the spherical aberration contribution to the PSF is reduced;
moreover a better control of the curvature of the field is achieved
spherical abberation
tan 2 
 k
tan 
coma
L
2

4
tan

tan

 
F
 = on-axis incidence angle
 = angular off-set
L = mirror height
F = focal length
Van Speybroeck & Chase, Appl. Opt., 1971
•
A typical aspect ration between focal length and mirror shell of 14 - 15 must has to be
taken (this was for the old WFXT design), but it should change in the interval 10 – 30,
depending on the f-number.
N.B.: short mirror shells  increase of the manufacturing problems!
11
Optimization of the Single Mirror Shell
Best merit function for optimization of
surveying telescopes
f = Focal Length / Shell Entrance Radius
l = (100 x Total Mirror Length / Focal Length)
Simplified formula for the HEW of a polynomial optics (BBG, 1992) weighted
over the FOV coming from the optimizations (Conconi et al., Applied Optics ,
2009, submitted)
12
Image quality for different f x l products
1) f = 5 ; l = 10
2) f = 7.1 ; l = 7.1
3) f = 10 ; l = 5
(Conconi et al., Applied Optics , 2009, submitted)
13
Polynomial shell (f = 5, l = 7) versus Wolter I and W-S
(Conconi et al., Applied Optics , 2009, under submitted)
14
Aberration Analysis
(f = 5, l = 7)
(Conconi et al., Applied Optics , 2009, submitted)
15
Butterfly-like configuration
The curvature of the field is dependent on f1.8:
In order to maintain the same focal plane curvature, and the same f x l product the
length L should change along the series of nested shells.
average focal profile
internal shell
external shell
Butterfly-like assembly must be used,
with mirror shells shorter at the
center.
See Conconi and Campana, 2001 – Conconi et al., 2004 - Conconi et al., 2009
16
Intersection planes at different focal distances
• Even if with different height along the series, images produced by different
mirror shells do not superimpose exactly, having different plate scales and then
they have not the same best focus positions (plate scale problem)
• The problem is attenuated by using mirror shells with intersection planes at
different positions (i.e. they have to be moved relatively to each-other)
See Conconi and Campana, 2001 – Conconi et al., 2004 - Conconi et al., 2009
17
Shift among the intersection planes
The plate scale of the shells is different along the
series of nested shells. If not corrected the effect
that the focal spots of different shells do not
coincide.
20 arcmin off axis
FL = 1 m
Outermost and innermost
mirror shells
Correction for 1 m focal length
In order to correct the effect a shift
among the intersection planes has to be
introduced.
18
Angular resolution optimization strategy
• Use of 3-order polynomia (x 2) for optimizing the “parabola” and
“hyperbola”
•
Figure of merit:
Effective Area per resolution element
Ang _ Re s( ) 

FOV
0
A
eff
( )
Ang _ Re s( )
2
 
HEW  D80%
2
• Number of modules: 3
• Diameter: large enough for compliance with requirements effective area
maximization
19
Sag of the first polynomial mirror wrt a Wolter I
20
Optimization of the acceptance angle
17 arcim is chosen as acceptance angle for the whole set of shells
21
Interface with spoke wheels: effects of the axial displacements
In order to reduce FEM dimensions only 9
MS have been included in the model.
MSs # 1-16-31-46-62 have been modelled
with their real characteristics (thickness
and material).
They have been used for the evaluation of
the optical degradation by ray-tracing.
In between there are four dummy MSs
having mass and stiffness respectively
equivalent to the MS groups:
from #2 to #15
from #17 to #30
from #32 to #45
from #47 to #61
22
Mirror modules parameters
N.B: the weight was calculated for the full set of shells concerning the 3 mirror modules;
at least 30 % more should be accounted for the mechanical structure
23
Effective area for SiC
Coating: Pt + C overcoating
24
Effective area for Glass
Coating: Pt + C overcoating
25
Why Carbon overcoating?
3 mm wall thickness
N.B.: spider vignetting not included
Measured reflectivity of Pt and Pt+C
Focal spots at different angular off-sets (1 keV)
0 arcmin
15 arcmin
5 arcmin
20 arcmin
10 arcmin
25 arcmin
28
HEW for the mirror module (theoretical design)
29
Main technical aspects
The realization of mirror shells with a small aspect ratio ( length/diameter more
than 3-4 times lower than XMM and Chandra)
 increased difficulty in reaching very good angular resolution:
• mechanical behavior closer to a “belt-like” configuration rather than a
“tube-like”
• border effect errors with a much higher weight in determining the PSF
• angular resolution more strongly affected by the slope errors caused by
out-of-phase azimuthal errors
30
Past and future X-ray telescopes: HEW vs. the
Mass/Collecting-Area ratio
IXO
WFXT goal
31
FEM analysis: short vs. long mirror shells
Example: LOADINGS GIVING SMALL SPATIAL SCALE DEFORMED SHAPE
(i.e. the deformed shape affects small portion of MS surface (at least in long MS) with
strong and local displacement gradients):
Loading 5:
twelve tangential
moments (10Nmm each)
applied at front section in
12 point 30° spaced.
Loading 6:
twelve outward radial
forces (0.1N each)
applied at front section in
12 point 30° spaced.
32
Why SiC and Glass for the WFXT mirrors
Higher rigidity mirror shells, based on materials with:
• low density (to increase the wall thickness)
• good mechanical parameters such as SiC and Glass
can be the solution for the above mentioned problems
Bending
merit figure
(2)
MS Material
Density
Ρ [t/m3]
Young
E [GPa]
Poisson
Ratio ν
CTE (1)
[°K-1]
Ther. cond.
[Wm-1K-1]
(1)
Electrof. Nickel
8.8
180
0.3
12.7×10-6
60
1.0
Glass 3)
2.51
72.9
0.208
7.2×10-6
0.93
16.6
Fused Silica
(HSQ300)
2.2
72.5
0.17
0.55×10-6
1.38
24.2
CVC SiC
3.18
456
0.21
2.33×10-6
140
51.1
33
WFXT heritage (SiC by epoxy replication)
see O. Citterio, et al., ”, SPIE Proc., 3766, 198 (1999) Ghigo et al., SPIE Proc., 3766, 209
(1999)
Tests @ Panter-MPE &
Marshall XRF
WFXT (epoxy replication on SiC)
Ø = 60 cm
Ni replication with same mandrel
Ø = 60 cm
Height = 20 cm
Height = 20 cm
F. L. = 300 cm
F. L. = 300 cm
HEW = 10 arcsec @ 0.1 keV
HEW = 35 arcsec @ 0.28 keV
34
Direct polishing approach for WFXT mirror shells
Angular resolution
Chandra, Rosat
(0.5” – 3”)
Direct polishing
Direct polishing
WFXT
goal 5”
Electroformed Ni
Effective area
WFXT goal
9000cm2
XMM Newton,
Jet X, Swift (15”)
Electroformed Ni
LEGENDA:
Technology proved
achievement
Technology under development
Difficult
35
Materials for carriers
Two materials under investigation:
• SiC (CVD for allowing the polishing)
• Glass (Fused Silica)
SiC CVD Pros
•
•
•
•
outstanding T/M parameters
already used in space applications
low density
polishable up to 2 Angstrom
SiC CVD Cons
• very hard (long time for polishing)
• cost
Fused Silica Pros
• well known material already used
in space applications
• low density
• polishable up to 2 Angstrom
(very easy)
Fused Silica Cons
• available just on thick tubes (to be
grinded!)
• T/M parameters lower than SiC
36
Processes envisaged for the mirrors production
37
Mirror shell on astatic support (1)
•During metrology and polishing (just for glass and
CVC SiC) operations the MS in vertical position
(axial gravity) rests on astatic supports, which
contrast the gravity by controlled axial forces.
•The astatic support number has to be computed in
a way that the gravity deflections are sufficiently
small.
38
Direct polishing & metrology
Credits: Zeeko, UK
Advanced technologies but thy have to be tested on thin shells asap
39
Glass tube of Heraeus HSQ 300 during grinding
Typical working time: 1 week/shell with 1 machine
(but it can improve)
60 cm
5 mirror shells already ordered (60 cm diam , 1.5
mm thick; + 4 48 cm diam, 1.5 mm thick).
They will be used for direct polishing testing.
Credits: Heraeus, Germany
40
Stress- free CVD SiC material available
Credit: TREX, USA
1 mirror shell already ordered ( 30
cm diam; 1.5 mm thick). A secvond
(60 cm cm; 1 mm thick) is available
from past projects.
They will be used for direct polishing
testing.
Typical deposition time: 100 mm / hour (a couple of days for a shell production)
41
Rough SiC shell produced at TREX
42
Jig for metrology
and machining
43
44
Thank you!
45