This week I cover a rebuttal against the claims of having detected phosphine (possibly connected to life) on Venus, a proposal for solving the core-cusp problem using late-time dark matter oscillations, and how self-interacting dark matter could provide the seeds for the observed supermassive black holes. This week I've found it particularly important to repeat a few important caveats: my covering particular papers does not necessarily mean I endorse the science or the writing style, just that I found the paper particularly interesting (it may even be directly outside my field of research). In other words, your mileage may vary, so don’t forget to take whatever you read here with a grain of salt. One thing I’m starting to do from this week on is to explicitly point out when a certain paper has been signed alphabetically (which is often the case within hep and gr-qc), in an attempt to hopefully give more visibility to early-career physicists whose last name initials, due to no fault of their own, reside in the second half of the alphabet.
#1 2010.14305: No phosphine in the atmosphere of Venus by Geronimo Villanueva et al.
I’ll start with an important disclaimer, namely that I am not an astrobiologist and hence no expert on the topic covered in this short paper (though I appreciate on a daily basis issues pertaining to miscalibration and introducing too many parameters in a statistical fit). So, as anything I write on this blog, take what I write with a grain of salt. I’m sure everyone has heard about the news of the discovery of phosphine (PH3) on Venus, possibly tied to the existence of life, which made media headlines recently. Now, the reasonining leading from PH3 to life is itself debated, as it is not 100% clear how PH3 is created on Earth, for example. But we’ll leave this aside, as it is irrelevant here. What is being questioned in this week’s paper by Villanueva and collaborators, submitted to Nature Astronomy “Matters Arising”, is the genuinity of the detection itself.
PH3 is detected through its first rotational transition at a frequency of ~267 GHz. What Villanueva and collaborators argue is that the PH3 signal is very easy to confuse with that of sulfur dioxide (SO2), which is naturally present on Venus, as the spectral resolution of both JCMT and ALM are comparable to the velocity separation of the two lines, so it is spectroscopically challenging to separate the two. More specifically, the two main issues pointed out here are a possible calibration issue and a possible overfitting issue. Regarding calibration, it appears like setting certain flags in the calibration script to True rather than False makes a huge difference in the residual spectra, and that is something being currently investigated by several teams. Regarding overfitting, the original team fitted their data with a 12th-order polynomial. Have they been overzealous in doing so, and unwillingly oversubtracted noise to the point of tricking themselves into creating artificial features which wouldn’t normally be there? This question is still up for grabs. As a note, I haven’t read the original PH3 detection paper so I don’t know what type of statistical procedure the original team followed (at the core, whether a frequentist or bayesian one). I’ve read in popular media accounts, in relation to this 12th-order polynomial issue, that one should “know when to stop”. Well, Bayesian analysis gives you precisely the tools to figure this out, and realize when you’re overfitting the data and introducing too many parameters, against the spirit of Occam’s razor.
I also noted that the first version of this week’s paper actually had a rather hard final sentence in its abstract (“We ultimately conclude that this detection of PH3 in the atmosphere of Venus is incorrect and invite the Greaves et al. team to revise their work and consider a correction or retraction of their original report“) which was mildened in the second version (“We ultimately conclude that this detection of PH3 in the atmosphere of Venus is not supported by our analysis of the data.“) - in my opinion the first version is definitely way too harsh given that this week’s paper is, after all, still a preprint which has yet to undergo the scrutiny of peer review. For the record, at least another recent preprint raised similar concerns, in this case focusing on the 12th-order polynomial fit and whether it leads to spurious results: 2010.09761. Regardless of what the outcome of this debate turns out to be, I think it is very healthy and is precisely how science works. And in any case making the leap PH3->life is clearly not as immediate as many media outlets would make it sound. I wouldn’t be surprised if settling this issue once and for all required actually going to Venus (which shouldn’t be impossibly hard, as we’ve done it at least as early as the 60s or 70s, depending on whether we’re talking about soil landing or not).
#2 2010.12583: Late-Time Dark Matter Oscillations and the Core-Cusp Problem by Jim Cline, Guillermo Gambini, Sam McDermott, and Matteo Puel (alphabetical)
The core-cusp problem (CCP - unfortunate acronym I know) is a long-standing problem of the collisionless cold dark matter (CDM) paradigm. The CCP amounts to the observation that CDM predicts DM halo profiles which are too “cuspy” (growing too fast as r->0) compared to observations, particularly of dwarf spheroidals, which instead appear to suggest “cored” profiles (constant density as r->0). There are many ways out. The usual “get out of jail free” card is “baryonic physics/feedback/gastrophysics”. For sure feedback can help redistributing material from the inner more dense regions to the outer less dense ones, changing the profile to something more cored. The problem is that the systems where the CCP is most pronounced - dwarf spheroidals - are mostly devoid of baryons, so it is unclear really how much this works. Other means, proposed at least as early as astro-ph/9909386, invoke DM elastic scattering with a cross-section per unit mass of the order σ/m~0.1-1 cm^2/g, to the limit of what is allowed by the Bullet Cluster, which also helps redistributing the matter. The problem here is that σv (with v the velocity) grows and hence makes the effect more important in larger systems, contrarily to what is observed. The alternative is to go for exothermic processes, such as fusion as studied by McDermott (one of the authors of this week’s paper) in 1711.00857, as it naturally predicts σv~const at low velocities, in better agreement with observations. Along a similar spirit, this week’s paper by Cline and collaborators also considers an exothermic process, albeit a slightly different one: DM annihilation. This idea had been proposed over 20 years ago in astro-ph/0005210, but is now being revisited much more quantitatively.
More precisely, Cline and collaborators consider an asymmetric DM (ADM) model, where there is an asymmetry between DM particles and antiparticles (usually within these models the relic density is determined by this asymmetry, much as the relic density of baryons in the visible sector is determined by the baryon asymmetry rather than by some freeze-out process). More precisely, the scenario being considered here envisages DM annihilation being effective at early times, shutting off later (I suppose once the symmetric component has been significantly depleted), and being reactivated at late times due to DM particle-antiparticle oscillations. Considering a Dirac fermion DM, these oscillations are activated by a DM-number violating mass term, essentially a Majorana term. Cline and collaborators consider two classes of models, where the DM couples either to a light vector boson or to a complex scalar. As a note of caution, at least the vector boson model cannot be an UV complete model because at the very least it should specify how the gauge symmetries of the model are broken by both the Majorana and vector boson mass terms (a Higgs- or dark Higgs-like mechanism probably would do), but in an EFT spirit it is totally fine. The rest of the paper is devoted to showing that these two toy models do indeed lead to cored profiles, with similar effects as those of DM elastic scattering, backing up the claim with both numerical studies and N-body simulations. The model proposed does not, in its simplest form, address the issue of the diversity of halo profiles. However, there are hints for the fact that this might work if both the vector boson and complex scalar mediator models are combined. Overall, definitely a very interesting read!
#3 2010.15132: Seeding Supermassive Black Holes with Self-Interacting Dark Matter by Wei-Xiang Feng, Hai-Bo Yu, and Yi-Ming Zhong (alphabetical)
We have observed quasars at very high redshift (z~7 or higher), which means that supermassive black holes (SMBHs) with masses of at least ~10^9M☉ existed less than a billion years after the Big Bang. The question of how these SMBHs managed to grow so heavy so quickly is an open puzzle. Direct collapse of Pop III stars is impossible as these stars would have to be way too massive. Another way is direct collapse of pristine baryonic gas, although observations pose challenges to this scenario too. Dark stars, stellar objects powered by DM annihilation rather than fusion which I discussed in my Week 39 entry, could grow extremely massive and in principle collapse to these SMBHs, although again the extent to which this would work is unclear.
In this week’s paper, Feng and collaborators study a qualitatively different scenario, in the form of self-interacting DM (SIDM). They show that SIDM can naturally trigger gravothermal instability, which is reached when the 3D velocity dispersion of the SIDM particles reaches ~0.57c. This mechanism is further accelerated by the presence of baryons, which shorten the collapse timescale. Self interactions-induced viscosity also helps the system dissipate its angular momentum. The mechanism is shown to work for both Eddington and sub-Eddington accretion rates (quantifying accretion efficiency). Even helped by this mechanism, the initial halo has to come from the high tails of the primordial density fluctuations, making these SMBHs extremely rare, which is qualitatively consistent with observations. It is interesting that the range of cross-sections per unit mass which explain this SIDM-induced gravitational instability, of order σ/m~1 cm^2/g, to the limit of what is allowed by the Bullet Cluster, are also the same ones required for SIDM to address the problem of the diversity of DM distributions across galaxies.