After a while that this subject hadn’t appeared in my blog, this week’s entry is entirely dedicated to the Hubble tension: a new observational update in light of Gaia DR3 parallaxes, an old model being killed, and prospects from future neutron star-black hole mergers. A heads-up that this will be the last entry for 2020, as I’m “going” on holidays. Stay tuned for next year’s entries, “see” you soon, Merry Christmas and Happy New Year!
#1 2012.08534: Cosmic Distances Calibrated to 1% Precision with Gaia EDR3 Parallaxes and Hubble Space Telescope Photometry of 75 Milky Way Cepheids Confirm Tension with LambdaCDM by Adam Riess et al.
This is the latest update by Riess’ group on the local distance ladder side of the H0 tension business, focusing on Cepheid-calibrated Type Ia Supernovae (SNeIa). In short, the local distance ladder consists of three rungs: 1) using geometrical distance measurements to calibrate the Cepheid period-luminosity relation; 2) identifying (now calibrated) Cepheids in distant galaxies containing SNeIa, to calibrate in turn the SNeIa magnitude-distance relation; 3) use this calibrated relation to move into the Hubble flow with SNeIa. One particularly crucial bit of the rung is the first one: any mistake in distances one gets there will propagate to the final parameters one infers, including the value of H0. Therefore, it is crucial to get geometrical distances to nearby Cepheids as accurately as possible.
The first rung of the local distance ladder usually makes use of parallaxes, or apparent shifts in the angular positions of objects as the Earth moves around the Sun. And the Gaia satellite, when it comes to parallaxes, is a goldmine of information. Unfortunately, the 1st and 2nd Gaia data releases (DR) suffered from large thermal oscillations which produced a variation in the angle between its two fields of view. In short, the telescope was sort of wobbling. As a result, Gaia’s parallaxes were affected by an additive error which made it hard to reliably estimate absolute parallaxes. To make matters worse, this “parallax offset”, as Riess and collaborators refer to it, is not a single value (which is why it is misleading to call it “parallax zero point”), but depended on color, magnitude, and location of the source which was being observed. Clearly not ideal if one wants to get reliable distances to Cepheids to calibrate the first rung of the distance ladder.
On December 3, Gaia released their DR3 data products, which had mostly removed this parallax offset, while at the same time modelling its dependence on color, magnitude, and location of the source in a much more precise way. What Riess and collaborators have done here has essentially been to calibrate the local distance ladder using the latest Gaia DR3 parallaxes, together with 75 Cepheids: 25 were new and 50 were already present in earlier analyses, but now have distances recalibrated using the Gaia DR3 parallaxes. The net result is a 1.8% determination of the Hubble constant, H0=73.2±1.3 km/s/Mpc. The previous go-to value was H0=74.03±1.42 km/s/Mpc from 1903.07603. While the central value has decreased a bit, the Hubble tension is still there, alive and well. And therefore it is up to us theorists to figure out what is going wrong in our model (if something is indeed wrong). Finally, in the next months we should probably expect the new Gaia DR3 parallaxes being applied to recalibrate the Tip of the Red Giant Branch-based local distance ladder, by Wendy Freedman’s group in Chicago. Definitely something more to look forward to!
#2 2012.07519: Updated constraints on massive neutrino self-interactions from cosmology in light of the H0 tension by Shouvik Roy Choudhury, Steen Hannestad, and Thomas Tram
An interesting proposal to address the Hubble tension was put forward in 1902.00534 by Christina Kreisch and collaborators. It makes use of non-standard neutrino interactions (i.e. interactions other than those dictated by the weak force) which delay the onset of neutrino free-streaming. Usually what happens in these models is that neutrinos interact with a new scalar via a Yukawa-like interaction. This new interaction does not affect neutrino decoupling, but makes neutrinos interact with each other even after they decouple from the rest of the plasma. This delays neutrino free-streaming, and results in the neutrino phase-shift signature being absent in the CMB, an effect which can be compensated by changing several other cosmological parameters, including H0. From a fit to Planck 2015 temperature and lensing measurements, BAO data, and a local H0 prior, this previous work found H0=72.3±1.4, clearly a very interesting result.
Unfortunately, despite the huge amount of attention it received, the interacting neutrino model is not all it cracked up to be, for several reasons:
-the analysis of 1902.00534 failed to include some key datasets, including Planck polarization data (at the time the 2018 polarization likelihood had not yet been released, and up to then we had been warned that the Planck 2015 polarization data might have been contaminated by systematics) and the Pantheon SNeIa compilation, both of which are key to assess whether a given model really has the potential to solve the H0 tension: in particular, Planck polarization data is expected to ultimately provide the most stringent constraints on this model;
-similarly, only the “lite” versions of the Planck likelihoods were used;
-the use of the local H0 prior is questionable at best: it is an euphemism to say that I have written this several times in my blog (see e.g. the social injustice metaphor I cooked up in these regards in my Week 26 entry);
-realistic particle physics constructions of these non-standard neutrino interactions are facing enormous difficulties when confronted against laboratory constraints, leading the community to suspect that a particle physics model underlying these interactions has to be baroque at best.
This week’s paper by Roy Choudhury and collaborators addresses in essence the first three of the above points. They revise constraints on (massive) neutrino Yukawa self-interactions in light of the Planck 2018 likelihood (not making use of the “lite” version thereof), including Planck polarization and Pantheon SNeIa data. The results one should focus on are given by the “CMB+EXT” column in Table 2: the “best” result which is obtained in terms of H0 is H0=67.4±2.2, clearly not at all an improvement over ΛCDM (recall that for ΛCDM and using a very similar dataset combination one gets H0=67.4±0.5). Even from the statistical point of view (e.g. Bayesian evidence) the interacting neutrino model does not appear to be preferred over ΛCDM. The only “positive” result here is that the neutrino mass constraint appears to be robust against these possible self-interactions. I found this paper a very interesting one because I was always hoping someone would have shown, clearly and without misunderstandings, that the minimal self-interacting neutrino model is by no means a contender to solve the H0 tension. This is exactly what this week’s paper does, in a crisp and clear writing style, which is why I really enjoyed reading it.
#3 2012.06593: Prospects for Measuring the Hubble Constant with Neutron-Star-Black-Hole Mergers by Stephen Feeney et al.
It is well known that H0 can be measured using gravitational wave (GW) standard sirens (SS), a result which to the best of my knowledge was first discussed in Bernard Schutz’s 1986 Nature paper. This essentially works by determining the luminosity distance to the source (whose redshift one should also find a way to estimate, either through an electromagnetic counterpart or using more sophisticated statistical methods). One can then use GWSSs to construct a Hubble diagram - calibrated, unlike the uncalibrated Pantheon one - from which one can estimate H0. One way to do this is through binary neutron stars (BNS). In fact, a similar group of authors as in this week’s paper showed in 1802.03404 that a future sample of ~50 BNS mergers might be able to arbitrate the H0 tension, depending on what the true value of H0 is (the ~50 estimate is if the local value is correct, if Planck is correct then ~3000 BNS would be needed).
However, much less attention has been devoted to another class of mergers, partly because we do not yet have convincing evidence for having observed such mergers. I am talking about neutron star-black hole (NSBH) mergers, of which there are hints that GW190425 might be one. In this week’s paper, Feeney and collaborators perform the first end-to-end realistic analysis of simulated NSBH samples with electromagnetic counterpart, fully accounting for specified parent populations, combined GW and EM selection, and a complete noise treatment. The result is a state-of-the-art forecast showing that NSBH mergers will be able to achieve an unbiased 1.5-2.4% measurement of H0 by 2030, thanks to observations from LIGO A+, Virgo AdV+, KAGRA, and LIGO India. Of course, it is very possible that by then we will have independent measurements of H0 (e.g. from the local distance ladder) which will be an order of magnitude more precise - after all, we currently are looking at 1.8% measurements of H0 from the local distance ladder. And of course, the hope is that by 2030 we might finally have wrapped our heads around the H0 tension (or, most likely, not)!
Honorable mention goes to 2012.09125: Emmy's letter to Santa Claus (and a reply): Vaidya geometries and scalar fields with null gradients by Valerio Faraoni, Andrea Giusti, and Bardia Fahim. This is a short paper showing the impossibility of constructing Vaidya solutions sourced by a massless scalar field with null gradient, i.e. \nabla_\mu \phi \nabla^\mu \phi = 0 (Vaidya solutions are among the simplest non-static generalizations of the Schwarzschild metric, and describe non-rotating objects emitting or absorbing null dusts, i.e. basically some form of radiation). This paper is written in a very entertaining style, in the form of a letter from Emmy (presumably Noether, since she is quoted as being a child prodigy who likes to play with symmetries) to Santa Claus, where Emmy asks for Vaidya solutions sourced by a massless scalar field, and Santa Claus replies that he is very sorry but this is not possible, for the technical reasons explained in the paper. On the other hand, dear reader, I hope that your Christmas wishes will have better luck than Emmy’s, and that your 2021 will be better than your 2020 (let’s face it, it can hardly be worse). Merry Christmas!