Absorption in 'CMBR' wavelengths - observations? processes?

[Nereid]

A read of the COBE and WMAP teams' papers on the methods they used to produce 'contamination-free' maps (or, to remove all 'foregrounds') is exhilarating.

However, unless I missed it, these methods largely address emission in the wavelengths of interest (i.e. those which the various instruments detect) - zodiacal light, free-free emission, warm dust, distant 'point sources', etc.

The only exception is the SZE (Sunyaev-Zel'dovich Effect - inverse comptonisation of the CMBR by hot electrons in the ICM of rich clusters), an effect which has now been unambiguously detected for several clusters, and which holds much promise as an independent probe of many cosmological parameters and processes.

OTOH, Copi et al. (and several others) have identified anomalies in the quadrupole and octopole CMBR, viz, a lack of power (cf inflation-based cosmological model predictions), and curious, 'non-fluke' coincidences (esp with the ecliptic).

So, two general questions:
1) irrespective of known processes, plausibility, etc, to what extent could absorption rob the l = 2 and 3 multipoles of power?
2) irrespective of observed examples, and sky distribution, what physical processes (other than the SZE) could give rise to absorption in the relevant wavebands?

 

[hellfire]

For me it is hard to imagine how any scattering effect on cosmological scales can affect the large scales of the lowest multipoles. For example, the Sunyaev-Zeldovich effect becomes relevant for scales smaller than 2 or 3 degrees (the angular size of the biggest clusters in heaven). As these lowest multipoles are due to gravitation on large scales (ISW effect), I would guess that the suppression of their power should be also due to some gravitational effect, if any.

 

[Garth]

Perhaps it is the other way round? That is the lowest l multipoles are not diminished by local contamination but do not exist in the first place?

The problem in removing all the foreground contamination is you have to know what it is before you can subtract it - how do you know you have assessed it correctly and accounted for it all? The fact that the quadrupole and octopoles are aligned with local geometry suggests that these have been added by local contamination, rather than diminished by it.

The WMAP spectrum would then be truly deficient at these large angles beyond a 3-sigma significance and a modification of the standard interpretation of the data would be required. e.g. the universe is finite after all so there was not enough room in the early universe for these large angle fluctuations to form in the first place.

So how can the data be consistent, with both a flat and finite universe? One solution would be that it is consistent instead with a conformally flat and finite universe! (Remember conformal means angle preserving and the WMAP data is angular in nature.) Such a conformally and finite universe might be Einstein’s original static cylindrical model, a conical model or a non-trivially connected model such as a torus

Garth

 

[SpaceTiger]

1) irrespective of known processes, plausibility, etc, to what extent could absorption rob the l = 2 and 3 multipoles of power?

In terms of robbing those multipoles of power, I would say that the answer is that there is no limit to the extent. In fact, absorbing gas could even add power, as Garth has suggested. Given the large number of possible types of contamination, I don't think one can prove it to be one or the other, but I'll agree that a statistically significant alignment with the ecliptic would initially suggest that power was added. I'm eager to see how this issue is dealt with in the next release. Perhaps this is just my prejudice, but I suspect that topology is not the reason for the low quadrupole.

2) irrespective of observed examples, and sky distribution, what physical processes (other than the SZE) could give rise to absorption in the relevant wavebands?

Unfortunately, there are a lot of possibilities. A large fraction of these concern the unknown properties of local, interstellar, and intergalactic dust. Not to diminish the many advances we've made in the subject of dust -- in fact, I worked with Bruce Draine, possibly the world's leading expert on the subject -- but the number of degrees of freedom in this problem is enormous and we can't reasonably expect to always be able to predict its distribution. The easiest way to deal with it is to see the effects it has on a known spectrum of light (as per the usual reddening correction), but it's exactly the known spectrum of light that's in question here.

 

[Nereid]

but I'll agree that a statistically significant alignment with the ecliptic would initially suggest that power was added.

Part of the why I asked was to see if there might be ways, in principle, to distinguish absorption effects from emission ones.

The ecliptic alignment is interesting - adding power in the ecliptic (I'm simplifying, 'in the ecliptic' is shorthand) results from an emission source whose distribution is close to l=2+3 peaks? or troughs?? OR an absorption 'source' whose distribution is close to these peaks?

In one case, you have a cosmological CMBR deficient in l=2,3 power (cf concordance models which include inflation); on the other you have a concordance CMBR, but l=2,3 power drained by 'local' absorption.

From the observations, can you tell (even in principle)?

I'm eager to see how this issue is dealt with in the next release.

Take a ticket, get in line! :tongue2:

 

[SpaceTiger]

From the observations, can you tell (even in principle)?

I don't think it's an insurmountable problem, but it's certainly a difficult one. The best way to approach it would probably be to look for the signature of the contaminant in other bands. If you find it, you can constrain its distribution and subtract it from the CMBR. This may be a long-term theoretical and observational problem, however, since all-sky maps are few in number and usually don't go very deep.

 

[Nereid]

A large fraction of these concern the unknown properties of local, interstellar, and intergalactic dust.

Thus there seem - to me - to be two classes of approaches (not necessarily fully independent, of course):
- from observations (probably thousands of different kinds, probably the work of at least 50 years), constrain the properties of the various dust populations sufficiently to put limits on their contribution to 'the CMBR'
- from the CMBR, identify anomalies (wrt concordance cosmology) and use these as inputs to focussed research into these anomalies (at the top of the list, for such programs, would be 'dust, whose distribution 'looks like' that of the anomaly').

The easiest way to deal with it is to see the effects it has on a known spectrum of light (as per the usual reddening correction), but it's exactly the known spectrum of light that's in question here.

D'accord.

How about turning it around? For example, from other means, determine the 'known spectrum of light', do the (CMBR) observations, turn the handle, and voilà! Of course it may be humongously difficult to do, but observing 'at night'* would be key.

*how do you know the 'diffuse X-ray background' is? Well, you observe 'at night' - point your X-ray telescope to the dark side of the Moon. So, for 'non-cosmological' dust, you look for changes in signal as 'occulters' move across it - the Moon, Jupiter (we've been there, so we know 'all about' sources and sinks between here and there), the blackest of black Bok globule (not really; it'll likely be lit up like a flare, in the microwave region. However, you get the idea?), SZE 'shadows', ...

 

[Nereid]

For example, the Sunyaev-Zeldovich effect becomes relevant for scales smaller than 2 or 3 degrees (the angular size of the biggest clusters in heaven).

Hey, what if we live near the outer edge of just such a one? Or, some way in from the edge of a mini version of one? Wouldn't there be a large angle effect? (doesn't a nearby tree loom large, no matter how tiny the giant's twin appears, 50 km away?)

 

[SpaceTiger]

...but observing 'at night'* would be key.

*how do you know the 'diffuse X-ray background' is? Well, you observe 'at night' - point your X-ray telescope to the dark side of the Moon. So, for 'non-cosmological' dust, you look for changes in signal as 'occulters' move across it - the Moon, Jupiter

The contaminating signal could very easily be from the outer parts of the solar system (say, an extended dust disk/torus) and I can't think of reasonable occulting objects for this observation beyond the inner solar system. Even if we could find such an object, it wouldn't tell us about the distribution of the contamination (needed for subtraction) because presumably the object would only cover a very small area of sky (much much less than is needed to constrain the low l multipoles of the contamination). Also, your method only works for contamination from emitting sources, as best I can tell.

The most important thing to remember, however, is that it's not enough to identify the physical nature of the source of contamination, you also need to know how it's distributed. If it were, by some miracle, very simple in nature (like a slightly aspherical solar wind due to rotation of the sun), one might be able to construct a theoretical model of its distribution without observing the whole sky. I doubt this will be the case, however.

 

[SpaceTiger]

Hey, what if we live near the outer edge of just such a one? Or, some way in from the edge of a mini version of one? Wouldn't there be a large angle effect? (doesn't a nearby tree loom large, no matter how tiny the giant's twin appears, 50 km away?)

One of the nice things about the SZ effect is that you know exactly where to look for it. As far as I know, the only significant SZ contributors to the microwave background will be dense clusters because that's where the temperatures are highest. If the local X-ray gas were hot enough to create a significant amount of inverse compton scattering of the CMB, it would probably have been previously constrained by X-ray measurements.

It would be an interesting exercise to compute, however, so perhaps I'll run some numbers when I get some free time.

 

[Chronos]

Nereid's scenario is more plausible, in my mind, but ST raises good questions. I lean toward local effects being underestimated. The apparent distortion inline with the solar ecliptic is suspicious.

 

[Garth]

Right on cue on today's arXiv: Local Pancake Defeats Axis of Evil First we note the unlikelihood of the low mode deficiency being a statistical fluke:

The original low quadrupole anomaly has long been dismissed as either the result of some residual systematic error or as a statistical fluke. However, the higher quality of data now available from WMAP strongly challenge the residual systematic explanation (Tegmark et al. 2003), and while a statistical fluke cannot be ruled out, the odds against are uncomfortably long. In one recent study, Copi et al. (2005) have used the multi-pole vector formalism to show that a purely accidental alignment is unlikely in excess of 99.9%. They have also shown that most of the ℓ = 2 and ℓ = 3 multi-pole vectors of known Galactic foregrounds are located far away from those observed in WMAP data, strongly suggesting that residual contamination by foregrounds which are currently included in the analysis is not a viable explanation. It is precisely this combination of a complete lack of any known systematic error, and long odds against random alignment that has earned the low-ℓ alignment anomaly the nickname “Axis of Evil”

The suggestion in this paper is that the "Axis of Evil" is explained by the gravitational lensing of the CMB dipole.

Weak lensing of the CMB has long been a topic of interest to cosmologists (e.g. Seljak 1996; Zaldarriaga & Seljak 1998; Hu 2000; Challinor & Lewis 2005). The effect is both simple and inescapable; all light which reaches us from the surface of last scattering (or any other source, for that matter) is deflected from its original path by the weak gravitational lensing
interaction with the matter distribution along the line of sight (see Bartelmann & Schneider 2001, for a comprehensive review), and no exceptions are made for photons from the CMB dipole. Although the dipole owes its existence to the motion of the observer with respect to the background, this makes no difference from the perspective of someone in the same frame of reference as the observer; one side of the universe is simply hotter than the other, and this anisotropy will be lensed. We note that since the dipole term measured by WMAP (given in Bennett et al. 2003b, as 3.346 cosθ mK in the direction (ℓ, b) = (263.◦85, 48.◦25) in Galactic coordinates) is more than two orders of magnitude larger than the quadrupole term (and is by far the largest anisotropy in the CMB), even sub percent level scatter will strongly effect the low ℓ moments. Also, because the dipole is coherent over the whole sky, it will couple best to lensing effects that are also coherent over much of the sky, so that local structures will be the dominant lenses.

This lensing may be modeled by a spherical mass (DM) 10¹⁷M_solar (~twice Great attractor) at a distance of 30 Mpc, or, as an alternative, by closer and smaller masses, even by that of our own Galaxy if the DM halo is more massive and planar than expected.

Therefore explanations of the "Axis of Evil" do exist, however to reiterate my original point, post #3 above, the paper continues:

This leads us directly to our next point; while it appears that we may have gone a long way toward eliminating the axis alignment problem, the effect of dipole lensing is to take power away from the dipole and add it to the higher moments, so that the low quadrupole anomaly is stronger than ever.

An over-estimation the low-l power would cause problems in the WMAP analysis of other cosmological parameters, so again perhaps talk the "age of precision cosmology" is a little premature!

Garth

 

[SpaceTiger]

An interesting idea. Thanks for the link, Garth. In fact, it seems to be the lensing equivalent of what Nereid was suggesting for the SZ effect.

 

[Garth]

An interesting idea. Thanks for the link, Garth. In fact, it seems to be the lensing equivalent of what Nereid was suggesting for the SZ effect.

Yes, as far as the SZ explanation is concerned that paper says

The fact that the AOE alignment includes sources of both cosmological and local origin suggests that the explanation might be found in some sort of interaction with local structure which is not already included in the foreground analysis. If this is the case, then the fact that the AOE points in the direction of the Virgo cluster (de Oliveira-Costa et al. 2004) is certainly intriguing, and has led some to suggest (Abramo & Sodre 2003) that the SZ (Sunyaev & Zeldovich 1972) imprint of the local supercluster might help explain at least some of the AOE. Although this idea is initially attractive, the SZ is probably 3 or 4 orders of magnitude too small to do the trick (Dolag et al. 2005).

Note as far as the M_solar¹⁷ "attractor" is concerned it would also have to be moving wrt the CMB isotropic frame comparable with our own local velocity. Given that much of our local velocity may be due to infall into the Great Attractor and other local potential wells, we may ask whether it is reasonable for the lensing "attractor" to give an interesting effect. If that is not the case then the lensing source may be closer to home.

Garth

 

[turbo]

Nereid's scenario is more plausible, in my mind, but ST raises good questions. I lean toward local effects being underestimated. The apparent distortion inline with the solar ecliptic is suspicious.

Indeed. There have been estimates of the temperature of the vacuum made since the 1800's, and many of these estimates ranged around 3-5 deg K - far more accurate than Gamow's prediction of 50 deg K, yet somehow when the CMB was confirmed, the Big Bang crowd claimed all the glory, and the Steady State folks like Eddington who made OOM more accurate predictions got no notice.

I firmly believe that WMAP's CMB anisotropies are all functions of the movements of the WMAP probe relative to the vacuum, and that the CMB is the temperature of the vacuum, and not a cosmological relic. If the small-scale (degree or two) anisotropies in the data of WMAP1 and WMAP2 were consistent, the WMAP2 would have been released promptly with a huge fanfare. If these tiny anisotropies do not accurately overlay the data from WMAP1, this is proof that the CMB is local, not cosmological, since projected back to the surface of last scattering, these tiny angles encompass huge regions of space that cannot have changed together in the space of a year. I hate to be negative and suspicious about this, but again, if WMAP2 agreed with WMAP1 on small scales, the data would already be in our hands.

 

[SpaceTiger]

I firmly believe that WMAP's CMB anisotropies are all functions of the movements of the WMAP probe relative to the vacuum

Movements that produce a nearly perfect Gaussian random field with a series of acoustic peaks in the power spectrum...

 

[turbo]

Movements that produce a nearly perfect Gaussian random field with a series of acoustic peaks in the power spectrum...

In previous posts you have quoted private communications with members of the WMAP team (not inconsistent with their public communications, BTW). Do you have any insight into the conditions ( I will not characterize them as problems) that have prompted the delay of WMAP2?

 

[SpaceTiger]

Do you have any insight into the conditions ( I will not characterize them as problems) that have prompted the delay of WMAP2?

What I said before is all I know. I spoke to one of the team members yesterday and his guess is that the release will be in November sometime. I wouldn't be surprised if this "Axis of Evil" had something to do with the delay, but that's just my speculation.

 

[Nereid]

Unfortunately, there are a lot of possibilities. A large fraction of these concern the unknown properties of local, interstellar, and intergalactic dust. Not to diminish the many advances we've made in the subject of dust -- in fact, I worked with Bruce Draine, possibly the world's leading expert on the subject -- but the number of degrees of freedom in this problem is enormous and we can't reasonably expect to always be able to predict its distribution.

Nereid's (observational) curiosity is piqued.

The observed CMBR is a 2.73K blackbody, all over the sky, within tiny error bars (this chart on Ned Wright's page is as good as any to summarise this).

Strip out this, and you find a dipole, down in the millikelvins.

Strip this out, and you find the holy grail, the angular power spectrum, in the microkelvins.

Real dust has a SED (spectral energy distribution) that deviates from a BB ... at the % and tens of % level!

Light echos (several have featured in APOD, over the last few/dozen years) make good probes of one sort of local dust, and assiduous monitoring of micrometeorite plasma trails tells us something useful about what comes in from the ISM.

Are there any particularly good (Bruce) papers you could recommend, ST?

 

[SpaceTiger]

Are there any particularly good (Bruce) papers you could recommend, ST?

Well, Draine & Lee 1984 is his most cited. Of course, you could also try Draine & Bond 2004.

 

[Nereid]

Well, Draine & Lee 1984 is his most cited. Of course, you could also try Draine & Bond 2004.

Clearly this Draine guy has done a lot of good work; but does anyone know anything about this young upstart, Bond?

Good work, SpaceTiger!

But back to the CMBR ... both the COBE and WMAP teams spent a lot of time (and computer power) addressing the 'foreground dust' contamination. What would you say are Bruce's 'top three' aspects of 'local' (out to the Great Attractor) dust, in terms of really great difficulty to tease apart from the more distant CMBR?

 

[SpaceTiger]

What would you say are Bruce's 'top three' aspects of 'local' (out to the Great Attractor) dust, in terms of really great difficulty to tease apart from the more distant CMBR?

I can't really speak for Bruce, but I suspect the main question mark would be the ubiquity of spinning and magnetic dust emission. There is sometimes a deviation in the thermal spectrum of dust emission in the microwave regime because of the fact that the grains are spinning, but it's not entirely clear how common this is in the galaxy and where we should be looking for it.