Sunday, 6 March 2016

Friday wrap-up: LIGO, chasing the 750 GeV excess...

Apologies for the long hiatus -- other aspects of life have been getting in the way. Here is a summary of the last month or so...

  • Of course the biggest news was the first observation of gravitational waves, a binary stellar-mass black hole system, and a binary black hole merger. Not bad for an 8 page paper! The signal is really quite striking; it's wonderful to see the agreement between the two detectors. I reproduce the observation plot below, just because one cannot admire it enough.

    Interest in the finding was phenomenal; the Physical Review Letters server even crashed (they were getting 10k hits per minute). One can find plenty of explanations at various levels online: e.g. for the layperson see Quanta, or Brian Greene on the Late Show; for the more scientifically minded there exists a digestible summary of each paper by Christopher Berry; or for a colloquium-level talk see Barry Barish at CERN. Lastly you can enjoy the xkcd.

    Here I just want to mention some interesting facts, taking as read the core ideas behind the phenomenon and the measurement. The event was actually observed before the first planned science run, during an engineering run. It was identified within 3 minutes and the decision was subsequently made to keep settings in place to take 16 more live days of data. This time period was chosen so that the data-driven background estimation could nail down the unlikeliness of the event to >5.1σ under the background-only hypothesis. Below is shown the event and background estimation; the detection is well in excess of 5.1σ, even including the event itself in the background estimation.

    The data is in fact completely open and you could analyse it yourself! In addition to the GW150914 event there are also two others that rise somewhat above the background ("GW151012" and "GW151226"). You can see them by eye in the above plot. They are clearly not statistically significant enough to announce a discovery alone, but still they are tantalising... with room for improvement to design sensitivity (by a factor of ~2 which increases the spatial reach by 2^3) and the construction of a third detector in India to triangulate the signal, the future of gravitational wave astronomy is exciting.
  • There's also that puzzling observation by Fermi of a gamma-ray burst 0.4s after the gravitational wave detection. There are good reasons for and against believing this was associated with the GW150914 event (see Quanta); the best way to tell is to just to wait and see if it happens again!
  • On the 750 GeV diphoton excess you can read Jester's "750 ways to leave your lover" on various explanations.

    As well, a comprehensive paper appeared on the arXiv reviewing some of the renormalisable and weakly-coupled explanations. In the authors' literature review they "found a wide range of mistakes or unjustified assumptions, which represent the main motivation that prompted this work." The suggestion is to utilise computational tools (e.g. SARAH) to automate the work which the phenomenologist should be doing anyway for a thorough analysis, and they provide 40 model files to match models already in the literature.

    Let us review these "mistakes or unjustified assumptions"; we will refer to the resonant 750 GeV state as $S$ throughout, and consider models where the effective coupling of $S$ to the diphoton/digluon vertices are induced by a loop of gauge-charged fermion(s) or scalar(s) [of course there are explanations which do not fit into this framework]...

    Next-to-leading-order (NLO) corrections to the $S$ decay widths matter. Compared to the LO result used in many papers, NLO corrections typically decrease $\Gamma(S\to\gamma\gamma)$ by O(10%), and N3LO corrections can increase $\Gamma(S\to gg)$ by a factor of almost 2. Overall this means $Br(S\to gg)/Br(S\to\gamma\gamma)$ is typically underestimated (for scalar $S$) by as much as a factor of 2 when using the LO estimate. This will change best fit regions and lead to stronger constraints from the dijet channel. Models which live on the edge of exclusion based on LO estimations may not survive.

    It is often assumed that $S$ does not mix with the SM Higgs. But mixing is necessarily generated at some loop level. If a fermion is in the loop it is a three-loop effect (but with large Yukawas and strong gauge couplings at the vertices). If a scalar is in the loop it arises at one-loop. This contribution can be turned off by tuning a quartic term to zero [this is not stable under the renormalisation group evolution], but there is always a pure-gauge two-loop contribution. This effect should be acknowledged and checked for consistency.

    Another common assumption is that the $S$ vev is zero. But since there is a $Sgg$ vertex the $S\to -S$ symmetry must be broken when expanded around the vacuum ($S=S_0 - \langle S_0 \rangle$). It is hard to imagine a non-finely-tuned potential with this property and a minimum at $\langle S_0 \rangle=0$. Another way to argue this is made in the paper: if the original state $S_0$ couples at a three-point vertex with a new fermion or scalar, then a tadpole term will induce a non-zero linear $S_0$ contribution which acts like a vev insertion.

    Decay channels have been missed in some works which can significantly change conclusions.

    In the proposed models it is necessary to have a rather large diphoton width. The authors identify three main methods for achieving this. There are worries with each of them which have not been addressed uniformly in the literature....

    1. If fermion in the loop, then a large Yukawa coupling $yS\psi\bar{\psi}$. Typically these need to be O(1). Naturally, one should make sure perturbativity is under control when calculating the one-loop effective coupling. There exist papers which don't. As well, even if it remains pertubatively controlled at the 750 GeV scale, the renormalisation group evolution can evolve that Yukawa to the non-perturbative regime at some higher energy. This should be checked and at least acknowledged. [It is an interesting fact that this does not happen in the standard model for the top Yukawa; it is ~1 at the electroweak scale and shrinks with energy scale due to Higgs/top quark gauge contributions].

    2. If scalar in the loop, then a large cubic term $\kappa S XX$. The authors point out that this generally leads to problems with stability of the electroweak vacuum. I will also add that if the cubic term is large (>TeV) compared to the desired sub-TeV particles, then it is likely that the vacuum potential must be somewhat tuned, and this will not be stable under radiative corrections.

    3. Instead of relying on a large Yukawa or cubic, increase the charge coupling the loop fermion/scalar to gluons/photons or have more particles in the loop. Papers exist with Q≥5 and N=9000. However, such changes induce a large correction to the gauge coupling renormalisation group running above threshold, and can lead to high-energy electroweak behaviour which is ruled out, or worse to a Landau Pole at energies below 1 TeV.
  • My conclusion from all this: even if one does not object to the phenomena of ambulance chasing [I personally do not object to the principle], one should object to the lack of quality that seems to go along with it. It is a problem that assumptions are made and effects are overlooked which change conclusions considerably. It is a problem that lower quality papers are (at least for a good while) cited on par with better considered ones. It is a problem that we are mostly seeing the same idea embedded into different models with no qualitatively new observations. In addition, it is frustrating that (non-)participation in trending topics has implications for whether you can continue to make your way in the field, especially for early career researchers.
  • On a lighter note, see the arXiv preprint "A Theory of Ambulance Chasing" by Mihailo Backović (a 750 GeV ambulance chaser himself!) for a bit of fun, where he attempts to model the total number of papers on a trending topic as a function of time. For the diphoton excess: "It follows that if the interest scales as an inverse power law in time, the cumulative number of papers on a topic is well described by a di-gamma function, with a distinct logarithmic behavior at large times." "Di-gamma" is just brilliant. A (testable) prediction of this model is that "the total number of papers will not exceed 310 by June 1, 2016". If you feel like you have a better model, then throw your hat into the ring! 
  • The HEP Postdoc Project has appeared. To quote the website: "The HEP Postdoc Project intends to be a tool for Postdocs, or even PhD students, in the area of High Energy Physics... When an applicant accepts an offer, she/he is lacking, however, important information about the senior researchers in the corresponding institution... The goal of the HEP Postdoc Project is to fill this gap. Please, send us your opinions on senior high energy physicists you have interacted with in the past..."
  • CERN is doing an "In Theory" series of articles on the CERN Theory department. The first-two installments are "Welcome to the Theory corridor" and "why bother with theoretical physics?"
  • Conferences/workshops:
    • CoEPP Annual Workshop 2016 (indico)
    • LHC Performance Workshop (indico)
    • UCLA Dark Matter 2016 (agenda)
    • CERN Winter School on Supergravity, Strings, and Gauge Theory 2016 (indico)
  • Links without thinks:
    • Stories of Australian Science: Looking for dark matter in a gold mine.
    • Sabine Hossenfelder via aeon: The superfluid Universe.
    • SciAm: Physicist Sabine Hossenfelder Fears Theorists, Lacking Data, May Succumb to "Wishful Thinking".
    • Smashpipe: Who's winning the string wars and why should you care? [Part 1 and Part 2]
    • Quanta: From Einstein’s Theory to Gravity’s Chirp.
    • Lawrence Krauss via New Yorker: Do the New, Big-Money Science Prizes Work?
    • symmetry magazine: The ABCs of particle physics.
  • In audio/video media:
    • Recordings of talks from "Why Trust a Theory? Reconsidering Scientific Methodology in Light of Modern Physics" [In my opinion, more conferences/workshops should record and make public their talks like this].
    • CMS Experiment: An introduction to the CMS Experiment at CERN. [7:25]
    • Katherine Freese at Perimeter Institute: The Dark Side of the Universe. [1:03:16]
    • Gianfranco Bertone: The Quest for Dark Matter. [1:00:23]
    • Camilo Garcia-Cely: Phenomenology of Left-Right Symmetric Dark Matter. [1:06:30]
    • Gero von Gersdoff: Light by light scattering and the 750 GeV diphoton excess. [58:00]
    • The Good Stuff: What the Heck is Dark Matter? [12:02]
    • Stephen Sekula SMU Godbey Lecture: "The Tail of the Lion: 100 Years of General Relativity, the Scientific Theory of Space and Time" [1:10:05]
    • La physique autrement: Physics and caffeine. [9:12]

1 comment:

  1. W against T, reference required. #conspiracyofscience