This week’s post covers the resurgence of cosmic triangles in the context of the Hubble tension, how to use redshift drift to probe dark energy in a model-independent way, and how to self-calibrate polarization efficiency in CMB experiments. Enjoy the read and have a nice weekend!
#1 2102.05066: The trouble beyond H0 and the new cosmic triangles by José Luis Bernal et al.
I was too young to have been around when the term cosmic triangles was in vogue, but apparently it was first introduced by Bahcall, Ostriker, Perlmutter, and Steinhardt in astro-ph/9906463. Despite their looking like a Mason symbol, their purpose was effectively a way of representing the possible past and future histories of our Universe by placing the matter, cosmological constant, and curvature density parameters on the sides of a triangle and exploiting the fact that the three are related to one another because they should sum to unity (on the subject of Masonic symbolism, I recently finished re-reading The Lost Symbol by Dan Brown, highly recommended thrilling read!). In this week’s paper, Bernal and collaborators revive the concept of cosmic triangles within the context of the Hubble tension. The community is used to thinking about the Hubble tension in the r_d-H0 plane of sound horizon versus Hubble constant, which seems to be pointing towards pre-recombination solutions to the tension in order to maintain consistency with BAO and Hubble flow SNeIa measurements. However, there are other important pieces in the puzzle which go neglected if one only focuses on r_d and H0. For example, the age of the Universe t_U is one such piece. As recent works by some of the authors in this paper has shown (see e.g. 1902.07081 and 2007.06594), one can obtain a model-independent estimate of the age of the Universe from the ages of its oldest inhabitants, such as globular clusters. This represents an important piece in the Hubble tension puzzle, as it may support or disfavor late-time solutions to the Hubble tension, given that t_U is essentially only affected by the expansion at the Universe at z<10, and is instead completely insensitive to anything that happened prior to recombination.
Bernal and collaborators therefore propose the use of at least three different cosmic triangles (see figure below) to shed light on possible solutions to the Hubble tension. These triangles involve two cosmological parameters on two of their sides, with the third side being the product of the two. This is useful because different probes are sensitive to different combinations of these parameters, with certain probes sensitive to their product. The couples of parameters in question are: H0 and t_U (and H0*t_U constrained by BAO+SNeIa data), H0 and r_d (and H0*r_d also constrained by BAO+SNeIa data - actually even BAO alone), and h^2 and Ωm (and of course Ωm*h^2 constrained by the CMB). These triangles are a powerful consistency diagnostic, since models yielding agreement across all three triangles should be favored by data. Mathematically speaking, this is the statement that if we take all precision cosmological measurements at face value, we are in the presence of an over-constrained algebraic system (within statistical uncertainties), whose constraints are visualized by these triangles: incidentally, this over-constrained status is what makes it really hard to solve the Hubble tension. I think of it as a whac-a-mole game, where you solve a problem (kill a mole) somewhere but because of the over-constrained status of the system, you just end up creating a problem (having a mole pop up) somewhere else. A related issue was also brought up in the very interesting 2010.04158, which I discussed in my 2020 Week 41 post. I found the revival of cosmic triangles in the context of the Hubble tension very interesting, and I wonder what other triangles might be useful in this context (e.g. H0-T0-H0*T0, with T0 the CMB temperature?).
#2 2102.03774: Redshift drift as a model independent probe of dark energy by Asta Heinesen
Due to the expansion of the Universe, the redshifts of all cosmological sources are expected to change, given a sufficient amount of time. In fact, the approximation of redshifts being constant is, strictly speaking, only valid instantaneously. This is known as redshift drift, or the Sandage-Loeb test (from the names of the two scientists who first proposed its use as a cosmological probe, Allan Sandage and Avi Loeb), and is typically expected to be of order a few m/s per century. It will be extremely hard to measure, but succeeding in doing so would provide a clean and direct probe of the expansion of the Universe, in real time. For this reason, experiments such as CODEX and SKA are trying to detect this minuscule signature, and are expected to succeed given a few decades of observation time. Of course, if we manage to detect redshift drift, it will be equally important to have a solid theoretical understanding of what this redshift drift should look like in specific theoretical models of the Universe. Or, better still, to be in the position of analyzing this detection in such a way as to draw model-independent conclusions about the Universe.
Typically, a positive redshift drift signature is associated to a signature of dark energy or, more precisely, a violation of the strong energy condition (SEC), which basically stipulates that the equation of state of any given component should satisfy w>=-1/3 (but we know that w<-1/3 is needed in order to have acceleration, as the second Friedmann equation teaches us). However, this association between positive redshift drift and SEC violation is strictly speaking valid only on the homogeneous and isotropic FLRW metric. On sufficiently small scales, the FLRW metric is clearly an approximation to the truth, and it is therefore important to understand the expected redshift drift in a manner which is as independent of assumptions about the underlying metric tensor as possible. This is what Heinesen does in this week’s paper, finding that within General Relativity and without making any assumptions about the metric tensor, a measured positive value of redshift drift indicates a violation of the SEC. The only way this conclusion can be bypassed is basically if observers have a very special 4-acceleration profile, something which looks unlikely. This is a very important conclusion, because if we were to measure a positive redshift drift, we can be fairly confident that dark energy or something that looks very much like it exists, and that cosmic acceleration is not due to more complex effects such as backreaction (which is another area of research where Heinesen and the PI of her group in Lyon, Thomas Buchert, are experts) - though whether backreaction can successfully mimic cosmic acceleration as inferred is still an open question in the field.
#3 2102.03661: Breaking the degeneracy between polarization efficiency and cosmological parameters in CMB experiments by Silvia Galli et al.
Recent results from Cosmic Microwave Background (CMB) surveys have demonstrated that the polarization signal is rapidly coming to dominate over the temperature one in terms of signal-to-noise. This will definitely the case in upcoming Stage-3 and Stage-4 CMB surveys, so it is important to have good control over possible systematic errors in polarization which could propagate to the cosmological parameters we want to infer. One important parameter in this sense is the polarization efficiency, directly related to how much power is absorbed by a polarized detector per unit time, and related to the overall calibration of the polarization maps per frequency, P_cal. Typically, polarization calibration parameters are included as (one of the many) nuisance parameters one marginalizes over in a standard cosmological analysis, with priors informed from external calibration. However, it would be desirable if in principle one could do away with this external calibration, i.e. allowing the data to self-calibrate these polarization calibrations/efficiencies.
In this week’s paper, Galli and collaborators propose a way to self-calibrate the polarization overall calibration, P_cal, by exploiting the different way this parameter impacts the TE and EE spectra. In other words, given a particular cosmological model, P_cal is treated on an equal footing with all other cosmological parameters, and internally determined. This provides a very valuable cross-check to band-averaged polarization efficiency measurements determined using other approaches. Within the minimal ΛCDM model, the authors show that allowing P_cal to float has little impact on the values and uncertainties of the inferred cosmological parameters, with the exception of the amplitude of the primordial scalar power spectrum A_s, whose uncertainty increases by up to 50%. Their results are stable to extensions which free up the lensing amplitude A_L, the effective number of neutrinos N_eff, and the sum of the neutrino masses M_nu, although the upper limits on M_nu degrade slightly. Moreover, they show that the planned CMB-S4 experiment should be able to self-calibrate its polarization efficiency to the 0.2% level if only using TE and EE information, and to the 0.06% level if also including TT. On the other hand, Planck calibration comes with a 0.5% uncertainty, and therefore an external Planck-based calibration would be insufficient for CMB-S4.