Potential of
TMT polarimetry
Polarized
glass
Koji S. Kawabata
(Hiroshima Univ., Japan)
Index
1. Why Polarimetry?
2. Polarimetry Science Cases with TMT
•
SNe, AGN jet
3. Idea for Instrumentation for
polarimetry
•
WFOS/Mobie, IRIS
1. Why Polarimetry?
Because we explore…
 Geometry of Light-scattering media (free electron, dust
grain, etc) in distant point-like objects
(Morphology: bipolar/clumply mass-loss, jet, disk…)
 Physical properties of scattaring matter /
Extracting Light from Hidden source
 Magnetic Field (High-energy Electrons)
They are difficult to access by normal
imaging and spectroscopy.
2. Polarimetry Science Cases
W. Skidmore et al.
YSOs, Stars, ISM/IGM, SNe/GRBs, Exo-planets/Solar system, Extragalaxies, …
Currently, ~35 cases in various fields are listed.
Case 1: Supernovae
Image
© SNAP/LANL
Light curve
peak
Spectrum
Time Variation over ~100 days After Explosion
http://snap.lbl.gov/multimedia/animations/
5
SN Polarization: Continuum light
free electron
Scattered light
(polarized orthogonally to
scattering plane)
Polarization Pattern of Electron-scatteringdominated photosphere (Chandrasekhar 1960)
Spherical photosphere
Aspherical photosphere
Pol. Dir.
(mild 𝜏 case)
Polarization vector
Surface integration
of Stokes Q, U
Cancelled out and
Unpolarized (p~0%)
Not fully cannceled
and Polarized (up to a
few %; Hoeflich 1991)
SN Polarization: Absorption Line
Homogeneous ion distribution
Inhomogeneous ion distribution
Partially
absorbed
Homogeneously
absorbed
Polarization vector
P Cyg prof.
flux
%pol.
constant
𝜆0
𝜆
flux
%pol.
𝜆0
𝜆
In inhomogeneous case, polarization does not cancel out at
absorption line, and thus polarized (up to 𝑝 ~0.5 − 4% )
A few good samples.
Subaru/FOCAS spectropolarimetry
Tanaka+ (2012)
Polarization features at absorption lines
Probing 3D structure of SN photosphere
PA=90°
PA=0°
Tanaka+ 2012
See also Wang and Wheeler 2008;
KSK+ 2002; Leonard+ 2002; Wang+
2003
Tanaka+ 2012
Loop structure in QU diagram is
an evidence that the ion
distribution is NOT axisymmetric
(2D asymmetric), but
inhomogeneous (3D).
 Global mixing in expanding
atphosphere?
2D (axi-symmetry)
3D (inhomogeneous)
From 2-3D explosion models…
2D model
First principle 3D simulation
(Suwa+ 2010)
Entropy
0.0015ms
(Takiwaki+ 2012)
Density/Velocity
Blondin+ 04
0.0125ms
In 1D shock front stalls
Multi-D effects (SASI/rotation, mixing instability) accelerate
neutrino heating  Asymmetric engine is expected
From late-time spectroscopy…
Emission line profile may depend on viewing angle.
polar (single peak)
Torus-like O(Mg)-rich ejecta
w/ GRB
Fe(Ni)
Maeda+ (2002)
Mazzali, KSK+ 05; Maeda,
KSK+ 08. (cf. Modjaz+ 08;
Taubenberger+ 09)
w/o GRB
equator
double peaks
O-rich
 5 out of 18 (28%) ES CC SNe showed double-peak
profile in [O I] emission line. This high incidence is
consistent with the prediction in case that all of the
observed SNe are moderately asymmetric.
GRB jet – SN asphericity connection?
GRB-associated
Unpolarized hypernova
non-GRB
hypernova
Polarized
If the SN engine is bipolar associated with GRB jet, little or
no polarization is expected; while, non-GRB hypernovae may
show strong polarization
 New clue to the engine and mass—ejection process of
GRB-SN explosion with systematic observation with TMT
From X-ray SNR observation…
Cas A, a young SNR ~330 years old; Type IIb SN
(Fesen+ 2006; Krause+ 2008)
220”
Imaging + Doppler projection (X-ray, IR, ground-based Opt);
DeLaney+ 2010
(Seen) Ion distribution is inhomogeneous. How?
 Interaction with inhomogeneous CSM ?
 Already inhomogeneous at SN explosion ? – Clue to engine
Past line polarimetry for envelope-stripped
core-collapse SNe with 8-10m telescopes
SN ID
Ty
pe
Dist.
(Mpc)
Polarization
Loop or 3D ?
Tel./Instr.
Ref.
2002ap
Ic
9.7
Yes
Subaru/FOCAS,
Keck/LRISp,
VLT/FORS1
KSK+ 02,
Wang+02,
Leonard+ 03
2005bf
Ib
84
Yes
VLT/FORS1,
Subaru/FOCAS
Maund+ 07,
Tanaka+ 09
2007gr
Ic
9.3
No
Subaru/FOCAS
Tanaka+ 08
2008D
Ib
32
Yes
VLT/FORS1
Maund+ 09
2009jf
Ib
34
Yes
Subaru/FOCAS
Tanaka+ 12
2009mi
Ic
30
Yes
Subaru/FOCAS
Tanaka+ 12
5 our of 6 type Ib/c SNe show loop (i.e. 3D) signs.
Observation is restricted for nearest SNe (<~30 Mpc)
TMT reaches ~3 times distant SNe (~10 times more
SNe) and we can study time-variation of polarization
spectrum, probing structure of atmosphere.
Envelope-stripped (ES) CC SNe
H
He
C+O
H
He
C+O
II-P
II-L
H layer ~10M8
~1M8
Envelope-stripped CC SNe
Progenitor’s envelope structure just before SN explosion
H
He
C+O
He
C+O
C+O
IIb
Ib
Ic
0M8
0M8
~0.1M8
Mass loss (WR/LBV) or Binary effect
How the mass loss occur in a single star?
How the binary / rotation effects work?
Smith+ (2011)
15
Case 2: AGN jets
High spatial resolution polarimetry with HST
Structure of
magnetic field
1”
Pictor A: Thomson et al. (1995)
1”
M87 core and HST-1: Adams et al. (2012)
Wide-range emission (from radio to gamma-ray)
Time-variable
Requiring high-resolution NIR polarimetry with AO
→ Simultaneous polarimetry with ALMA (resolved)
and next generation X/gamma-ray satellites to
study jet physics (formation, structure, emission
mechanism)
Age of Next-Gen., Multi-wavelength Polarimetry
up to high-energy photons is coming..
1.9 yr
2.3yr
mm/sub-mm: ALMA
Hard X-ray: Astro-H
(2015-) ©ISAS/JAXA
©ESO/NAOJ
Optical 3-band polarization
(Δp~0.1%) of millions of stars
(<14mag) will be produced by
early ’2020 (?)
cm: SKA (’2020?-)
©Univ. of Manchester
Optical survey project
with 2m telescope
(Hiroshima Univ.+collabo.)
>100 times more samples than
current catalog (e.g. Heiles
2000)
©NASA
TMT
instruments
first
decade
Instrument
capabilities planned for theplanned
first decade of TMT for
operations.
The first
three are early
light instruments.
Instrument
Field of view / slit length
Spectral resolution
 (µm)
Comments
InfraRed Imager and
Spectrometer
(IRIS)
< 3 IFU >15”imaging
> 3500
5-100 (imaging)
0.8 – 2.5
0.6 –5(goal)
NFIRAOS
Wide-field Optical
spectrometer and
imager
(WFOS)
>40 arcmin2
>100 arcmin2 (goal)
Slit length>500”
1000-5000
>7500 @0.75” (goal)
0.31-1.0
0.3-1.5(goal)
Seeing-Limited
(SL)
InfraRed Multislit
Spectrometer (IRMS)
2 arcmin field 46
deployable slits
R=4660 @ 0.16 arcsec
slit
0.95-2.45
NFIRAOS
Multi-IFU imaging
spectrometer (IRMOS)
3 IFUs over >5’
diameter field
2000-10000
0.8-2.5
MOAO
Mid-IR AO-fed Echelle
Spectrometer (MIRES)
3 slit length
10 imaging
5000-100000
8-18
4.5-28(goal)
MIRAO
Planet Formation
Instrument
(PFI)
1 outer working
angle, 0.05 inner
working angle
R≤100
1-2.5
1-5 (goal)
108 contrast
109 goal
Near-IR AO-fed Echelle
Spectrometer (NIRES)
2 slit length
20000-100000
1-5
NFIRAOS
High-Resolution Optical
Spectrometer (HROS)
5 slit length
50000
0.31-1.1
0.31-1.3(goal)
SL
“Wide”-field AO imager
(WIRC)
30 imaging field
5-100
0.8-5.0
0.6-5.0(goal)
NFIRAOS
Courtesy TMT Observatory Corporation
Why NOT Polarimetry with TMT?
Essentially, polarimetry needs a number of photon in most
scientific cases.
(Requiring ~10 times larger S/N ratio than normal phot/spec)
Potentially TMT can be a great polarimeter.
Disadvantage
• Requiring extremely large optical
components that have never been
realized
• A few % instrumental polarization due
to reflection at tertiary mirror
• Less image quality/sensitivity due to
additional optical components
There would be two types of
astronomers…
1. People who like polarization so much.
2. People who do no like polarization so much.
Bipolarity --- 2. is likely to be major...
We (1.) may explore further possibility…
Required Field of View for TMT
polarimetry cases
From the list by Skidmore et al.
Required FoV
Number of polarimetric
scientific cases
Single point-like
source (<~10”)
28
10 - 60”
3
>60”
4
Most cases (31, 88%) requires only <60” Fieldof-View
→ Optical components in less size can
realize polarimetry with TMT instruments
3. Idea for Instrumentation
Target : Two of three First light instruments
Example: Polarimeter at 8-10m Tel.
Subaru/HiCIAO
VLT/FORS
Subaru/FOCAS
ISAAC/VLT
Keck/LRIS
Naco/VLT
(Gemini/Michelle)
Add-on-type polarimetry module 1
Aux/Filter
Retarder
Polarizing prism
 8-10m class optical: FOCAS/FORS/(LRIS) type
Rotatable
Half-wave plate
(~120mm diam.)
Beam splitter
(~120mm cube)
Filter 2
Filter 1
Grism
Polarization Module
Focal mask
(~175mm diam.)
FOCAS/Subaru
ordinary ray +
extraortdinary ray
We (need to) insert
two additional optical
components in the
beam.
Add-on-type polarimetry module 2
Circular spectro-polarimetry with fiber-fed high-disp.
spectrograph at 2-9m telescopes
Pol. unit inserted in front of optical fiber
 HARPSpol ESO3.6m/HARPS
(Mayor+ 2003)
Pol. unit for HARPSpol
(Piskunov+ 2011)
Pol. unit in SEMPOL
Focal reducer
1998 ）
beam spliter
λ/4wave plate
Focal mirror/slit
 SOFIN-pol NOT2.6m/SOFIN (Pettersson+
Optical fiber (ord/ext)
 SEMPOL AAT3.9m/UCLES (Semel+ 1993)
(Semel+ 1993)
WFOS/Mobie case
Mobie/WFOS
Focal plane (mask)
1260mm
550mm
~300mm𝜑 pupil
Large Optical Components vs. Clearance
Mobie/WFOS
Focal plane (mask)
550mm
1260mm
Wave plate before mirror
>550mm×1260mm
Beam splitter
@pupil position
>350mmφ
~300mm𝜑 pupil
Too difficult to realize full-spec polarimetry
→ Requiring some `trade-off’
Possible solution for Mobie polarimetry
Accepting Less Field of View (e.g., ~1’𝝋),
• Feasible size of wave-plate (140mmφ×
10mm-thick) is inserted just below the
focal plane, which shold be rotatable
• Challenging size of beam splitter
(mosaic’ed~350mmφ ×~40mm-t) inserted
just above/below the grisms.
Monolithic WP of
~150φ available
(A corp. in Japan)
BBO crystal having 2° apex angle
produce a beam separation of 0.47°
for ordinary and extraordinary ray.
Mosaic’ed
crystal
o
e
Accepting Less Accuracy/Efficiency,
• Feasible size of wave-plate (140mmφ)
• Feasible size of polarizer (~350mmφ
Wire-grid
wire-grid plate) at two position angles
(0, 90°) Reducing cost ↔ 50% light loss
Add-on polarimetry with a
restricted FoV (<~1’) could be
possible.
Sufficient clearance in the
optical train / holder / support
of Mobie?
Feasibility study needed.
Possible solution for polarimetry: IRIS
• Feasible size of wave-plate (depending on
FoV) at entrance of NFIRAOS
Fabrication with highest quality (<~200nm RMS wave-front error)
Somewhere before reflecting optics (first OAP).
• Feasible size of beam splitter
Not large thanks to restricted FoV
Somewhere in collimated beam
section
NFIRAOS: Herriot et al. (2011)
wave plate
beam splitter
IRIS : ©IRIS web page
Summary
 TMT could be a strongest polarimeter
 Most TMT polarimetry scientific cases
requires only small FoV
→ (A proposal) Add-on type, smaller sizes of
optical components (than originally-expected)
could provide polarimetric capability.
Polarimetric accuracy is still an issue,
depending on astronomy and target field.
(Cf. Study of instrumental polarization by Warren Skidmore’s team)
Thank you !
Enjoy (and be careful for) Japanese Fall !
Colored leaves (Nikko)
Typhoon