These are my personal notes from day four of the Topics in Astroparticle and Underground Physics Conference. The full conference materials are online: https://indico.cern.ch/event/1199289/timetable/#20230831
August 31, 2023
The fourth day kicked off with a focus on gravitational waves [Mavalvala]. The speaker began with a history of the technology. The LIGO goal is 
strain noise at 200 Hz. This is the upgrade target, allowing for detection of further, fainter objects. The expected range will increase to 240-325 Mpc.
Quantum engineering plays a role in the target of LIGO, as the uncertainty principle begins to demand serious consideration to achieve the goal. There is a fundamental uncertainty in the number of photons striking the mirrors in LIGO. This appears as a phase shift in the light. “Shot noise” is usually what this is referred to, and it’s fundamentally a Poisson process. Pumping more photons into the laser beam helps, but the 
means that there is then more variable photon pressure on the mirrors. This then degrades the measurement.
To avoid this limit, one has to embrace quantum mechanics via “squeezed states” of light. You can either squeeze the amplitude or the phase of the light. What you really want is frequency-dependent squeezing, achieving a phase shift in the light centred at 100Hz. To do this, they use 300m filter cavities, squeeze the light, and then inject it into LIGO.
Observing run 5 will go from 2026-2029, and for this more upgrades are planned (laser power, thermal noise, larger mirrors). After this, there is no choice left but to simply build overall larger interferometers.
We then moved on to gravitational wave observational results [Brady]. This provided a helpful “gravitational wave spectrum”. This runs from 
Hz, which can be probed by the CMB and tells us about events on the scale of the Hubble time; then to 
Hz, which can be probed using pulsar timing arrays and tells us about supermassive black holes; then to 
Hz, which can be probed using LISA and tells us about galactic binaries (with periods of hours); and finally to 
Hz, which can be covered by LIGO and other current-generation ground-based interferometers and provides information on compact mergers and, potentially, supernovas (with periods of seconds).
The speaker moved on to compact object mergers, about which we have the most information. The gravitational wave strain goes like 
. A gem of the recent years of observations has been the neutron star binary merger that was preceded by a gamma ray burst. This lent confidence to the idea that GRBs are connected to NS mergers. Observatories are now telling us about populations of compact objects. The mass sprectrum is so far inferred from observations. This currently peaks around 10 solar masses. There may be a secondary (tertiary) peak at 20 (30) solar masses. A power law + peak model works well with the data so far. For those other peaks, more work is needed to draw firm conclusions. NS-NS and NS-BH mergers are teaching us about the relative speeds of gravity and light, as well as the formation of heavy elements.
There are active searches for sub-solar-mass binaries, which could be used to put constraints on primordial black holes and the DM fraction represented by PBHs. (c.f. PRL 129, 061104) There are also efforts to observe gravitational waves from single pulsars, which could be induced by asymmetries on the NS surface (e.g. “mountains”). In addition, there are Hubble constant measurements using the luminosity distance determined from the GW waveform combined with the telescope observations.
In the Q+A, it was noted that GRBs are likely “beamed”, so the EM counterpart will only be detectable with GW when the beam is aimed at us. We got “lucky” in 2017, and from this the guess is that maybe 10% of GRBs are actually visible. It was noted that the 2017 NS-NS merger was a close-by object.
The next speaker then focused on next-generation GW observatories [Punturo]. The goal is to probe the cosmic “dark ages”, and this will require future-generation instruments. Examples are the Einstein Telescope (Europe) or Cosmic Explorer (US) concepts. These are moving through conceptual design stages, design reports, etc. The idea would be to network such instruments together between the EU and US.
These concepts expect to detect 100,000 BH-BH mergers per year and an equal number of NS-NS mergers in the same period. Current experience suggests that 
EM counterparts will be observed per year. Low-frequency sensitivity becomes crucial to understand high-mass objects. BH-BH coalesces are also crucial to test general relativity in strong field conditions and look for new phenomenon.
The Einstein Telescope effort is exploring two options: a triangle configuration for the interferometers, or a pair of “Ls”. There are a lot of open questions on the mirrors, squeezing the light, etc. The designs will require 30km of tunels, large caverns, underground operations, cryogenic systems, vacuum systems … in fact, the largest vacuum system ever constructed.
After the morning break, the next speaker focused on the LISA instrument [Danzmann]. The talk began recounting some history of the LISA concept, which dates back to 1985. With an expected launch in 2035, that will mark a 50-year journey. The basic concept is a series of space-based, free-floating test masses. The system has 17 degrees of freedom in their motion. LISA Pathfinder used 4 cm cubes of gold for the test masses, and this experiment represented “the stillest place in the universe”. With just a 38cm baseline, the test masses were held motion-free to better than the weight of a virus. LISA Pathfinder broke all expectations, beating even the full LISA mission requirements.
The goal is to detect the “galactic background” in the region of 
Hz. All binary black holes cross through the LISA sensitivity band at some point in their lifecycle. The lead on LISA is ESA and the partner is NASA, with a dozen states in the LISA consortium. The full LISA mission will total to EUR 2 billion.
The mission will be sensitive enough to resolve strains at the femtometre scale, not 
as for the ground-based instruments.
The next talk moved us along to MeV-scale gamma-ray astronomy [Ajello]. This focused on COSI, AMEGO, and FERMI-LAT. There is an MeV-scale gap in space-based instruments. Lower and higher than this region have excellent sensitivity, but not in this region. Coverage is needed here.
There are mysteries to resolve in this region. There is the 511 keV line, nuclear emission lines, polarization effects, and multimessenger detection in this region. We want to learn something about all of this. For example, consider the need to understand the formation of heavy elements in recent cosmic time, via stars, mergers, etc. Nuclear emission lines are critical as a tool here, meaning gamma and beta production. A key example of this is 
, which emits a positive beta ray that subsequently annihilates with an electron and can lead to 511 keV lines.
The size of the observed galactic centre 511 keV gamma ray line is still a mystery. It’s compatible with the structure of the bulge and the disk in the Milky Way, but inconsistent with star formation. Could it be from SMBH injection, pulsars, dark matter, or something else? We need an MeV map of the galactic centre. For example, we might discover young supernova due to the delayed decay of key isotopes with lifetimes on the scale of decades.
Regarding polarization, there are many sources of emission that themselves can have large polarizations. Good examples are GRBs and blazars, where 50% and 30% polarization are expected from models, respectively. Emission via synchrotron and Compton effects can in turn alter the polarization.
In the multimessenger space, we know now of high-energy neutrino emission from AGNs. We also have GW coincidences with short GRBs that peak in the keV band optically. Cosmic rays and galactic acceleration are also key targets, looking at bremsstrahlung from electrons that leads to MeV/keV gamma rays.
The field needs instrumentation. COSI (the Compton Spectrometer and Imager) is planned by NASA and will launch in 2027. It will provide sensitivity from 0.2-5 MeV. It is host to a cryogenically cooled calorimeter using germanium detectors at 5K. It will provide a 25% instantaneous field of view of the sky. Type-II supernova at distances out to the Large Magellenic Cloud will be visible to COSI. The instrument is optimized for sensitivity to key lines such as the 511 keV line and those from Co-56, Ti-44, Fe-60, and Al-26.
There are also concept missions: the AMEGO concept includes AMEGO-X from the US and e-ASTROGRAM from the EU. These would be able to see 200 short GRBs each year vs. the expectation of 10/year from COSI.
The Q+A here focused on the complexity of operating cryogenics in space and the skepticism from space agencies about using liquid nobles in that environment. If such concepts succeeded they would provide excellent resolution.
A key target for the 511 keV line program is the detection of shifts. This will tell us something about the source(s).
The final talk of the plenaries focused on diversity and inclusion in physics [Muino]. She began with definitions. “Diversity” refers to the recognition that people differ in many possible ways and that we can learn from the richness of experience. “Inclusion” refers to valuing each person so that they feel they belong and can contribute to the fullest of their abilities.
The speaker motivated why we should worry about these two things. First, there is the general issue of fairness and social justice. But there is also the fact that there is plenty of evidence that diversity boosts innovation and productivity, and referenced studies of companies that demonstrated these effects. The speaker reviewed the statistics for women in physics. She noted the GENERA project (an EU study) that looked at gender discrimination in physics, the causes and the effects. GENERA resulted in a number of recommendations, such as training in the various issues, awareness of unconscious bias, efforts at family friendliness, and issues of mobility, among many others.
The speaker then reviewed the state of racial minorities in physics. She drew from a pre-pandemic American Institute of Physics report, TEAM-UP. There have been huge gains by African Americans in geosciences, but declines in physics. This makes the point that not all STEM fields are the same or suffer from the same issues and challenges. The report identified 5 factors that determined persistence in STEM: a sense of belonging, a sense of identity and recognition of one’s identity in others, academic support (e.e. effective teaching via emphasis on what is going right and how to learn from it, rather than focusing on failures exclusively), and financial support. The report proposed a set of measures.
The speaker then reviewed the state of LGBTQ+ members of the field, beginning with a review of statistics. The point was clearly made that not being able to be who you are creates a large additional stress on individuals in the field and distracts from their career. One avenue to alleviate this issue on the road to inclusivity is the establishment of support networks, especially role models and mentors. Another activity on the road to inclusion is the simple statement of preferred pronouns (e.g. on nametags). If everybody does this, then nobody feels apart and everyone knows who to address others. “If I do it, you can feel safe to do it, too,” to paraphrase the speaker.
The speaker then turned to the issue of implicit bias. For example, statements like “You don’t look like a physicist” serve to reflect the implicit assumptions of one person about what a group should be. A simple example of how implicit biases work their way into actions is when people act fast without using all of the information that is actually needed to make decisions. Resources from CERN on this subject were provided.
The speaker then turned to the matter of attracting more people from underrepresented groups in high-energy particle astrophysics. In this field, 80% of members are male and 20% are female. One avenue to address this is to enlarge the pool of people who are arriving to physics as a career path. This requires many things. Outreach remains crucial. For example, using the International Day of Women and Girls in Science to theme activities and messaging and call attention to the value of women in the field. Teacher programs generally reach a ot more people.
In large collaborations, there is more effort to establish DI offices to try to manage issues, collect statistics, etc. There were examples of such collected information. Diversity charters for large organizations (e.g. ECFA, NuPECC, APPEC) can help to promote diversity and monitor it at the same time, especially through the establishment of intentional monitoring programs to gather that information. This should be run in parallel effort for a field-wide collection of large, unbiased statistics so that we as a field can draw correct conclusions.
The Q+A noted that the evidence shown was gathered prior to the pandemic, and so doesn’t include its effects (which were substantial). The speaker noted that efforts are underway to update this. The question was raised about what TAUP itself was doing to address this issue. For instance, are participant statistics being collected? How were talks distributed? A key member of the organizing committee responded by noting that they tried hard to take gender diversity into account. For example, of the plenary speakers 38% were female (excluding the two “flashplenaries” that were each only 15 minutes). They also attempted to manage national diversity in the plenary talks. They also noted that TAUP was actively running child care.
After the lunch break, I ran around the parallel sessions again. I took in some of the talks on directional detection of dark matter [Higashino, Bignell] and proposed efforts to develop scalable detectors (NEWAGE, CYGNUS). I was also curious about the Super Kamiokande results in searching for a diffuse supernova neutrino background [Harada]. Results from the 0.01% loading (2020) with gadolinium were shown, but the detector is still unable to distinguish the nu and signal hypotheses in the data. Nevertheless, they achieved a sensitivity in this phase equivalent to 10 years of pure-water data. They have now (as of 2022) loaded to 0.03% Gd and results are expected in the future.
I also wanted to see the results of the liquid helium dark matter experiments, like HeRALD, SPICE, and TESSARACT [Hertel]. One key thing that the speaker emphasized was the general concern that cesium has been used in HeRALD to keep liquid helium from “wetting” the sensor above the liquid volume. While human-made cesium (e.g. nuclear testing) is radioactive, natural abundant cesium (which they used in the construction) is stable. There was also an excellent talk on qubit-based dark matter detector [Linehan], explaining the detection method for such a device. Right now, such a device can reach sensitivity to energy deposits at the order of 10eV. It’s not great, but it’s a start.
There was a report from the low energy excess workshop, which I encourage everyone to read [Wagner, Baxter].
Finally, I saw the updates from COSINE-100 and ANAIS on physics results [Lee, Coarasa]. COSINE has 6.4 years of data but only 3 years are public. 3 more years worth of data will soon be made additionally public. There has been o evidence of a DM signal in the existing 3-year release. However, this has been a model-dependent search; the model-independent annual modulation search has limited sensitivity so far, and in the 1-5 keV recoild energy region of interest is not yet sensitive to the DAMA/LIBRA claims. They noted that when they apply a background-averaging technique similar to that reported by DAMA/LIBRA< they can create an artificial modulation but with the wrong phase.
Ongoin work in COSINE is to lower the detector threshold and perform a low-energy calibration for electron and nuclear recoils. They will be ready for a WIMP search with a 0.7 keV threshold. They are moving from Yangyang Lab to Yemilab, and COSINE-100U will restart by the end of ’23. They also discussed crystal purity efforts for COSINE-200. This has suffered setbacks, and it will take another 1-2 years for batch production at the target purity.
ANAIS has completed 643.48 kg-years of data exposure as of August of 2023, with nearly 100% livetime. Their public results are determined from only 3 years of data. They observe no modulation at the level of 
and a 
exclusion of the DAMA/LIBRA result. They have employed machine learning approaches to improve the signal-to-noise discrimination in their data. The speaker focused on a result obtained using a boosted decision tree (AdaBoost, specifically). The signal is Cf-252 calibration data, while the noise is obtained from a “blank module” run. The AdaBoost approach combines a series of individually weak discriminators into a more potent single discriminant. They define an energy-dependent BDT cut, and applying this to their existing data they expect to be able to exclude DAMA/LIBRA at 
with their current data (643.48 kg-years) and at better than 
in late 2025. For the model-independent analysis (annual modulation) they expect to be able to quote results in both the 1-6 keV and 2-6 keV regions. They also intend to take into account the potential difference in quenching factors between DAMA/LIBRA and their own experiment (20% for ANAIS vs. 30% for D/L).
