Saturday, 13 December 2014

Planck: what's new

Slides from the recent Planck collaboration meeting are now available online. One can find there preliminary results that include an input from Planck's measurements of the polarization of the  Cosmic Microwave Background (some which were previously available via the legendary press release in French). I already wrote about the new  important limits on dark matter annihilation cross section. Here I picked up a few more things that may be of interest for a garden variety particle physicist.








  • ΛCDM. 
    Here is a summary of Planck's best fit parameters of the standard cosmological model with and without the polarization info:

    Note that the temperature-only numbers are slightly different than in the 2013 release, because of improved calibration and foreground cleaning.  Frustratingly, ΛCDM remains  solid. The polarization data do not change the overall picture, but they shrink some errors considerably. The Hubble parameter remains at a low value; the previous tension with Ia supernovae observations seems to be partly resolved and blamed on systematics on the supernovae side.  For the large scale structure fans, the parameter σ8 characterizing matter fluctuations today remains at a high value, in some tension with weak lensing and cluster counts. 
  • Neff.
    There are also better limits on deviations from ΛCDM. One interesting result is the new improved constraint on the effective number of neutrinos, Neff in short. The way this result is presented may be confusing.  We know perfectly well there are exactly 3 light active (interacting via weak force) neutrinos; this has been established in the 90s at the LEP collider, and Planck has little to add in this respect. Heavy neutrinos, whether active or sterile, would not show in this measurement at all.  For light sterile neutrinos, Neff implies an upper bound on the mixing angle with the active ones. The real importance of  Neff lies in that it counts any light particles (other than photons) contributing to the energy density of the universe at the time of CMB decoupling. Outside the standard model neutrinos, other theorized particles could contribute any real positive number to Neff, depending on their temperature and spin. A few years ago there have been consistent hints of Neff  much larger 3, which would imply physics beyond the standard model. Alas, Planck has shot down these claims. The latest number combining Planck and Baryon Acoustic Oscillations is Neff =3.04±0.18, spot on 3.046 expected from the standard model neutrinos.  This represents an important constraint on any new physics model with very light (less than eV) particles. 
  • Σmν.
    The limit on the sum of the neutrino masses keeps improving and gets into a really interesting regime. Recall that, from oscillation experiments, we can extract the neutrino mass differences: Δm32 ≈ 0.05 eV and Δm12≈0.009 eV up to a sign, but we don't know their absolute masses.  Planck and others have already excluded the possibility that all 3 neutrinos have approximately the same mass. Now they are not far from probing the so-called inverted hierarchy, where two neutrinos have approximately the same mass and the 3rd is much lighter, in which case Σmν ≈ 0.1 eV. Planck and Baryon Acoustic Oscillations set the limit Σmν < 0.16 eV at 95% CL, however this result is not strongly advertised because it is sensitive to the value of the Hubble parameter. Including non-Planck measurements leads to a weaker, more conservative limit Σmν < 0.23 eV, the same as quoted in the 2013 release. 
  • CνB.
    For dessert, something cool. So far we could observe the cosmic neutrino background only through its contribution to the  energy density of radiation in the early universe. This affects observables that can be inferred from the CMB acoustic peaks, such as the Hubble expansion rate or the time of matter-radiation equality. Planck, for the first time, probes the properties of the CνB. Namely, it measures the  effective sound speed ceff and viscosity cvis parameters, which affect the growth of perturbations in the CνB. Free-streaming particles like the neutrinos should have ceff^2 =  cvis^2 = 1/3, while Planck measures ceff^2 = 0.3256±0.0063 and  cvis^2 = 0.336±0.039. The result is unsurprising, but it may help constraining some more exotic models of neutrino interactions. 


To summarize, Planck continues to deliver disappointing results, and there's still more to follow ;)

14 comments:

Anonymous said...

What about BICEP2? Does the new Planck result say anything about it? Thanks!

Xezlec said...

Neff is neat, but I'm surprised you can't get more information about heavy particles from these cosmological analyses. I'd think that studying The Big Event would be the only way to get at that kind of stuff since you can't build a collider that big. :(

Yannick Copin said...

A new systematic has been identified for SNe Ia (Rigault+13, and its impact on H0 is indeed to reduce the tension between CMB and SNe Ia estimates.

Jester said...

Nothing about primordial B-modes in the BICEP window yet, but it should appear soon.

Cédric Bardot said...

What about the topology of the universe? From this quote :" The future release of Planck polarization data will allow us to improve considerably the constraints on models whose size L [of the fundamental domain] is smaller than DLSS [diameter of the last scattering surface] ...(http://arxiv.org/abs/1311.3509)" one could have expected new piece of information on this (less topical) issue.

Anonymous said...

Note that the standard value of N_eff = 3.046 for 3 active neutrinos relies on several assumptions:
*) 3 active neutrinos in the SM
*) No partly thermalised lights species
*) Reheating happened above ~4 MeV
*) No entropy production between 1 MeV and today
*) No cooling of photons between 1 MeV and today (e.g. through Dark Sector mixing)

Only the first assumption was verified by LEP, so there were plenty of room for N_eff to be different from 3.046.

Jester said...

I completely agree, maybe except that "Reheating happened above ~4 MeV" is independently confirmed by nucleosynthesis. I didn't mean that Neff is not useful, on the contrary. I meant that it is often presented as a measurement of the number of neutrinos, which may be misleading.

Stacy McGaugh said...

There are lots of direct distance measurements (not just SN) that favor H0 in the low to mid 70s. These are not so easy to reconcile with 67. I have feared for some time that confirmation bias would drive us to find systematics that reconcile traditional results with CMB fits. Now that we "know" the "right" answer, sure enough, we're finding such systematics.

coraifeartaigh said...

Don't entirely agree with the comment
"Planck continues to deliver disappointing results".
Disappointing would be if the instruments did not work as expected, or the resolution was less than that expected!
I wonder did you mean"some theorists may be disappointed by the empirical results offered by PLANCK so far" ?
Regards, Cormac

Jester said...

Right. But i suspect some experimentalists may feel this way too. You don't get a Nobel prize for shrinking error bars...

Anonymous said...

Neff =3.04±0.18,
does this imply that the possibility of eV scale sterile neutrinos is now gone in a model independent way?

Jester said...

eV sterile neutrinos that could address the short-baseline neutrino anomalies have always been disfavored, as they would give too much hot dark matter. After Planck, you can still have a light sterile neutrino as long as its mixing angle with the active neutrinos is small enough such that it doesn't thermalize. But then, I believe, the mixing angle is too small to probe this sterile neutrino in oscillation experiments...

Anonymous said...

That the Rigault (13) effect hasn't been confirmed by other studies (Betoule 14), both in terms of not seeing the effect (at modest significance) and it seeming to be rather unimportant for the cosmological parameters (although H_0 wasn't studied; however Rigault 13 claimed a much larger effect for the other variables as well). So further study is needed to see if that really helps.

Yannick Copin said...

Just a follow-up on the aforementioned impact of the SN Ia star formation bias on the Hubble constant measurement: Rigault+14.