The Gravitational Universe

1
This is a slide from Dick Bond that packs a lot
of information. A lot of information is on this
slide, so it must be from Dick. Where is Dick
anyway? This is a slide from Dick Bond that packs
a lot of information. A lot of information is on
this slide, so it must be from Dick. Where is
Dick anyway? This is a slide from Dick Bond that
packs a lot of information. A lot of information
is on this slide, so it must be from Dick. Where
is Dick anyway? This is a slide from Dick Bond
that packs a lot of information. A lot of
information is on this slide, so it must be from
Dick. Where is Dick anyway? This is a slide
from Dick Bond that packs a lot of information.
A lot of information is on this slide, so it must
be from Dick. Where is Dick anyway? This is a
slide from Dick Bond that packs a lot of
information. A lot of information is on this
slide, so it
2
On Predicting the Polarization of
Low-frequency Emission by Diffuse
Interstellar Dust
IAS 12 September 2005
3
ID
? Peter Martin
? CITA
4
Motivation CMB Polarization
As we have heard, and will hear, several of the
recent and next-generation cosmic microwave
background (CMB) experiments have polarimetric
capability, promising to add to the finesse of
precision cosmology. Among these are Archeops,
Boomerang (B2K in 2003), and the Planck Surveyor.
5
Archeops and Planck HFI
__
__
__
?
Archeops 10 to 20 _at_ 545 353 217
143 GHz
6
BOOMERanG
? ? ?
7
Contaminating components
Dust dominates above 100 GHz
Higher latitude
Figure from http//www.planck.fr/heading136.html
Giard and Lagache
8
Motivation Cirrus
One of the diffuse foregrounds contaminating the
CMB signal near a few 100 GHz (mm to
submillimetre range) is cirrus thermal
emission by diffuse interstellar dust.
9
Cirrus
IRAS 100 micron Faint diffuse emission
everywhereeven at high latitude
10
Cirrus Mitigation
Not the topic of this talk. Plan A mask out
regions of bright cirrus. But wide sky coverage
is needed for precision cosmology. Only 20 of
the sky has H I column density below 1020 /
cm2. Even that produces a non-negligible
foreground ( 1 MJy/sr at 100 microns). Plan
B measure properties of cirrus at high frequency
where CMB is not important, and extrapolate to
lower frequencies where one does have to address
component separation.
11
Motivation Dust Polarization
Since optical polarization is commonly seen, from
differential extinction by aligned aspherical
dust particles, it is expected that thermal
emission from these grains will be
polarized. Note Galaxy is optically thin in
submm. Therefore, we see the whole galaxy, or
right out of it. Unlike star probes which rely
on differential extinction along path. But at
high latitude, not dissimilar if stars are
sufficiently distant.
12
Polarization Optical and FIR
Both depend on aligned grains. Orientation of
E-vector of optical polarization is orthogonal to
that of the emitted radiation.
Figure from http//www.planck.fr/article263.html
Pontieu and Lagache
13
Alignment Theory
Alignment for Dummies coming soon to a
discerning supermarket checkout counter near you.
14
Polarization of Diffuse Emission
Indeed, in the Galactic plane and in dark
(molecular) clouds, dust emission in the infrared
and submillimetre has been measured to be
polarized. (next talk) It seems likely that the
faint diffuse cirrus emission, of more relevance
to CMB experiments, will be polarized too.
15
Now that were motivated
16
What has been accomplished? (1)
We discuss how well the degree of polarization of
the diffuse cirrus component can be predicted.
To do this we draw on what is known about
alignment from optical (and infrared and
ultraviolet) interstellar polarization. We
emphasize the importance of the polarized
intensity and its spectral dependence (needed
also for extrapolation to CMB frequencies).
17
What has been accomplished? (2)
We comment on polarization (alignment) of small
grains, possibly relevant to the anomalous
emission. We do not assess the power spectrum,
which depends on the spatial variation of the
alignment. (other talks)
18
Polarization of Emission
Polarized intensity P and intensity I are summed
over all grains species. The ratio is gives the
degree of polarization of the submillimetre
emission, p_emission.
Non-aligned grains dilute the net
polarization. Because of different weighting,
the spectral dependence of the polarized
intensity can be different than that of total
emission.
19
Calculations Submillimetre
In the submillimetre range of interest, the size
of the grains is much smaller than the wavelength
? simple analytical formulae can be used for
absorption ( emission) cross section per unit
volume e.g., for spheroids
20
Basic Model
For a single grain composition (silicate) and
axial ratio, independent of size,
There is a slight wavelength dependence across
the submillimetre range of interest, due to
changes in m, but the large nubeta dependence
cancels out. Depends on composition too (but
grains of other materials not aligned?).
21
p_emission for Single Grains
P/I for astronomical silicate (and amorphous
carbon)
22
Challenges
Wide range of grain sizes. Different grain
compositions. Grain shape how
flattened/elongated? Which grains are aligned?
How well?
23
Grain sizes (and compositions)
Grains come in many sizes (perhaps a function of
composition). Which grains produce the
submillimetre emission? Which grains produce the
extinction in the optical and ultraviolet? Which
grains polarize in the optical and
ultraviolet? Does this result in significant
submm polarization?
24
Lessons from Extinction Curve
Continued rise in extinction into ultraviolet
requires smaller and smaller grains. Bump at
2200 A. Separate grain components.
Fig. from Cardelli, Clayton, Mathis 1989
25
Extinction into IR
Follows a power law of index about 2 (1.84 here).
Silicate absorption at 10 microns (requires most
of Si to be depleted in amorphous silicate
grains).
Fig. from Martin and Whittet 1990
26
Lessons from IRAS and ISO
Spectrum components Fig from Desert,
Boulanger, Puget (1990)
grains of size 0.1 microns ?
? 1 mm
27
Origin of the Emission

• Components/Mechanisms
• gt 100 microns thermal emission by larger
  grains (size 0.1 microns)
• 60 and 25 microns non-equilibrium emission by
  smaller grains, 0.007 micron 70 A 7 nm
• 12 microns non-equilibrium emission by tiny
  grains/PAHs, 1 nm
• All of these components of course radiate at
  longer wavelengths too.
• Tiny grains also spin rapidly and emit microwave
  radiation which could be another foreground
  contaminant of the CMB (anomalous emission).

28
PAHs(simple ones)
Coronene C24H12
Naphthalene
Phenanthrene
Chrysene
29
Submillimetre Spectrum
In the submillimetre the thermal emission is
characterized by T and often a single beta, the
spectral index of the dust emissivity
Total intensity is volume weighted, since C/V is
size independent. In ISM, large grains carry
most of the volume. Is beta constant ( 2) with
frequency? Is T constant with size? Is epsilon
constant with T?
30
Spectral Index Variations

• Evidence for excess emission at 217 GHz (1.5 mm)
  (Archeops experiment Bernard et al. talk)
• Comments
• was attributed to cold dust at 5 7 K. But
  diffuse dust being that cold seems unphysical
• effect is seen everywhere (so a property of dust,
  not environment)
• Conclusion
• beta is not constant with wavelength over the
  range of interest
• 1.8 for lambda lt 600 microns (gt 500 GHz)
• 0 at 1 mm
• 2.2 at lambda gt 2 mm (lt 150 GHz)
• due to intrinsic processes in amorphous grains

31
Optical (and FIR) Polarization
Both depend on aligned grains. E-vector of
optical polarization is perpendicular to the
projected direction of magnetic field.
Figure from http//www.planck.fr/article263.html
Pontieu and Lagache
32
Interstellar Polarization Basics

• extinction scattering absorption
• grains are aspherical
• aligned, so that in plane of sky the ensemble
  average grain profile is elongated
• long axis of profile is oriented perpendicular
  to the magnetic field B
• differential extinction according to orientation
  of electric vector with respect to this profile
• ? net polarization of transmitted light
• greater extinction for E parallel to long axis
• ? E parallel to short axis, hence parallel to B

33
Wavelength Dependence of Polarization and
Extinction
Polarization reaches a peak while extinction is
still rising.
Fig. from Rogers and Martin 1979
Wavenumber ?
34
Polarization Curve
C
low polarization in the UV, whereas extinction
keeps rising
power law rise in IR (not unlike extinction)
Fig. from Martin, Clayton and Wolff 1999
35
Implication of Low UV Polarization
Despite a wide range of grain sizes for
extinction, only the larger grains are aspherical
and aligned. Figs. from Kim and
Martin 1994
Aligned grain mass distr.
low polarization in the UV
? small grains not aligned
36
Polarization of IR Features
It will be interesting to see if the 3.1 micron
ice band is polarized as it would be if the
aligned silicate grains were ice-coated. (Martin
1975) it is!
This is a line of sight to an embedded source,
the Becklin-Neugebauer object in OMC 1. Still,
the silicate to ice mass ratio is 15 45 thin
frost.
15
Fig. from Martin and Whittet 1990
37
Lessons from IR Extinction Features

• 10 micron polarization
• ? silicate component is aligned
• details of p/tau across the feature constrain
  the band strength and the shape and axial ratio
• ? Hildebrand and Dragovan 1995 find oblate with
  axial ratio 1.5

38
Fluffy silicate agglomerate IDP
Individual sub-grains the size of interstellar
silicates (0.1 micron)
39
GEMS
Glass with embedded metals and sulfides. Mg rich
silicate. Mid-IR spectrum like comets. Fe and
FeS inclusions. Lack of S depletion in gas a
problem if GEMS interstellar?
40
Lessons from Interstellar Polarization

• only the larger grains are aspherical and
  aligned
• 10 micron polarization ? silicate component is
  aligned
• axial ratio not extreme oblate
• certainly adequate to model with silicates
  alone

41
Summary so far
Large grains dominate submillimetre
emission. Only large (silicate) grains are
aligned. But shape and alignment?
42
Shape and Alignment
Both influence the degree of polarization. The
degree of interstellar polarization is also made
larger by larger column densities, but this is
just as for extinction. Thus the column
density can be normalized out by taking the ratio
of polarization to extinction, p/tau. The
observed envelope in p vs. tau constrains the
shape and the best achieved alignment.
43
Polarization/Extinction at V
Observed amount of optical polarization per unit
extinction provides the required measure of the
asphericity and degree of alignment.
Fig. from Serkowski, Mathewson, and Ford 1975

44
Polarization/Extinction
This ratio varies systematically over the range
infrared optical ultraviolet.
45
Bootstrapping

• Hildebrand and Dragovan 1995 find the effect of
  disalignment by comparing
• p_e for their model at 2.2 microns
• and
• p/tau observed at 2.2 microns.
• Problems
• former assumed pure absorption, whereas the
  latter involves grains of size comparable to
  wavelength, so scattering as well. Model p_e does
  not really apply at 2.2 microns.
• (ii) p_e at 2.2 microns for silicates is very
  sensitive to how dirty they are, which has
  little effect on submm p_e. Hard to scale.

46
p_emission for a Mixture
P/I for astronomical silicate and graphite (both
aligned)
47
Polarization/Extinction
p/tau 6 at 2.2 microns.
48
Calculating p/tau
Need to carry out calculations of extinction
(scattering absorption) by particles comparable
in size to wavelength (as in Mie theory for
spheres, harder for spheroids).
49
Detailed Models Recipe

• For a given axial ratio, and perfect alignment,
  find the aligned grain size distribution by
  fitting the wavelength dependence of interstellar
  polarization.
• Compare to model of interstellar extinction,
  keeping track of mass of all components
  (unaligned grains contribute to tau and not p,
  and so cause dilution). Use models of Kim and
  Martin 1995.
• Calculate p/tau.
• Compare this to observed p/tau and deduce a
  reduction factor R (lt1) due to imperfect
  alignment.

50
Results for Disalignment R
Repeat axial ratios, oblate and prolate
shapes. For example, for perfectly aligned
oblate silicate particles (R 1 in this case),
the axial ratio needs to be no higher than 1.4 to
produce the maximum p/tau observed. For larger
axial ratios, grains must be somewhat disaligned
by a quantifiable amount (reduction factor R lt 1)
to produce the same p/tau.
51
Apply to FIR and Submillimetre
For the same model (shape, axial ratio, grain
components), self-consistently calculate the
polarization of the low-frequency thermal
emission, p_e. Apply R from interstellar
polarization model for that axial ratio
(approximately correct). Repeat for different
axial ratios. ? predict maximum p_emission
expected.
52
Polarization of Emission of Aligned Grains

• predict maximum p_emission expected from the
  aligned grains. Large.
• Note how there is little dependence on shape and
  axial ratio.

53
Frequency Dependence of Polarized Intensity
Because of different weighting, the spectral
dependence of the polarized intensity P can be
different than that of total emission, I ?
Non-aligned grains dilute the net polarization.
If the frequency dependence of emission for the
diluting component is different, then this
introduces a frequency dependence for p P/I. ?
Measuring P (rather than p) offers a direct way
of examining the spectral behaviour for the
aligned grains.
54
Complication Dilution
Non-aligned grains contribute to the thermal
emission and dilute the net polarization. In
the MRN/Draine and Lee/Kim and Martin model with
silicate and (large) graphite, this causes a
further reduction, by d 1/3. Note that the
dilution of p/tau in the optical by extinction by
unaligned graphite has resulted in a larger R,
and so this is payback.
55
Dilution Silicate Graphite Model
Silicate is only part of the submillimetre
emission. Fig. from Draine and Anderson 1985
56
Final Prediction of p_emission
Kim Martin 1995 interstellar polarization
models maximum interstellar (p/tau)_V 0.0267
dilution in submillimetre d 1/3
Uncertainty arising from using models, and R.
57
Further Dilution

• Self-consistent model predicts a maximum
  p_emission 7 /- 2 .
• When averaged over large regions with
• non-uniform alignment (beam dilution)
• or less than perfect alignment (perhaps the
  direction of the magnetic field),
• or with alignment which changes along the line
  of sight,
• typically half of this might be expected (recall
  SMF figure of p/E_B-V).

58
Observations
Archeops observes 5 , even averaged over large
regions near the plane. Note E-vector
orientation is as expected too.
59
Discussion alternative models
Desert et al. model uses one type of big grain.
Therefore, less dilution in submm (only by VSGs
d 0.9), but less in optical too, lowering R and
so compensating in product dR. Also no
refractive indices (or shape or axial ratio)
specified, so no detailed calculation possible.
60
Alignment of Small Grains
p/tau very low in the UV, where extinction comes
from small grains (VSGs, PAHs). What
polarization there is is consistent with coming
from big grains. Small grains, the ones that
spin most rapidly and might produce anomalous
emission, are not well aligned.
61
Polarization UV extinction bump
This star is rho Oph AB. Fig. from Wolff et al.
1997
62
Polarization UV bump (1)
Polarization feature is at the same position as
the UV extinction bump, has a positive excursion,
and shows no change in position angle. But only
seen in 2 of the 28 lines of sight in the Galaxy
observed by WUPPE and HST (FOS), despite
sufficient S/N (Martin, Clayton, and Wolff
1999). Not a common phenomenon.
63
Polarization UV bump (2)
The ratio delta_p/delta_tau is a measure of the
polarization efficiency of the carriers of the
extinction bump and polarization feature (if
present). This ratio is small compared to the
corresponding ratio for the continuum (often by
an order of magnitude). It is at least 2 orders
of magnitude smaller than the theoretical maximum
for perfectly aligned graphite carriers (0.8
Martin et al. 1995). Thus, either the alignment
is quite incomplete or only a small fraction of
the grains is aligned. .
64
(No Transcript)
65
(No Transcript)
66
BN pol
C tau 3, p 15 at 10 microns. p/t .2 (.22)
see martin 1975 Would this, compared to 2
microns, make sense
67
Bare silicates
Hildebrand and Dragovan 1995 uses bare silicates
argues ice (OConnell) cant be too thick. Does
not mention organic refractories.
68
Martin 75
It will be interesting to see if the 3.1 micron
ice band is polarized as it would be if the
aligned silicate grains were ice-coated.
(silicate to ice mass ratio 15 45).
69
Coated
Desert BG are dominated in mass by the
silicates, with coating of organic CG 89 volume
dominated (90) by organic refractories so
optical properties are those of mantle (but if
all Si is depleted, and only 26 C in mantles,
then Si mass is significant (density is
higher). Both need to absorb starlight to warm
up. (circumstellar silicates are warm without
organic refractory mantles) Both simulate optical
albedo, but dont have reliable IR (2 micron)
albedo.
70
Desert
Sizes PAH 1 nm 10 A 0.001 micron VSG 7 nm
70 A 0.007 micron BG 15 to 110 nm 150 to
1100 A 0.015 to 0.11 micron Most mass in
largest particles
71
(No Transcript)
72
template
C
73
template
C
74
17O anomaly
A grain enriched in 17O. Supernova condensate
or massive star?
75
IR spectrum
For