The future is now

Mansi M. Kasliwal, the Principal Investigator of the Global Relay of Observatories Watching Transients Happen (GROWTH) collaboration, shares her enthusiasm about the future of multi-messenger astrophysics. Mansi M. Kasliwal discusses the Global Relay of Observatories Watching Transients Happen (GROWTH) collaboration and shares her enthusiasm about the future of multi-messenger astrophysics.


Gold rush
We are living the twenty-first century gold rush 1 : astronomers and physicists have been working together to identify the sites of heavy element production in the Universe (gold included). The theory was clear decades ago: the rapid capture of free neutrons (r-process nucleosynthesis) synthesizes heavy elements (atomic mass range between 70 and 200 amu) and neutron star mergers are a promising site for this process to take place. However, the direct detection of gravitational waves (GWs) was needed to catch neutron stars in the act of merging, and telescopes were needed to pinpoint the location and characterize the electromagnetic (EM) emission. This is not an easy feat. As an example, I describe some lessons learned with a team comprised of 16 institutions around the world: the Global Relay of Observatories Watching Transients Happen (GROWTH) collaboration (Fig. 1).

Seeing is believing
On 17 August 2017, the majestic merger of two neutron stars (GW170817) sounded a loud ~100 s GW signal and lit up the entire EM spectrum 2,3 . Detections were reported in the γ-rays, optical, infrared, ultra-violet, X-rays and radio. The ground-based near-infrared data showed clear evidence of the nucleosynthesis of elements in the first abundance peak of the r-process around atomic mass ~80. Space-based mid-infrared data suggested that even the heaviest of the heavy elements in the second peak (around ~130) and third peak (around ~190) of the r-process were synthesized 4 . The GROWTH team presented a concordant picture as a panchromatic breakout of a wide-angle, mildly relativistic cocoon formed by jet-ejecta interaction [5][6][7] . Continued long-term monitoring has revealed evidence that the jet survived.

Patience
Despite GW170817's promise of a bright and loud future, no kilonovae were detected in the third LIGO/ Virgo observing run (O3; April 2019 to March 2020). This may disappoint, but it should not surprise anyone. During O3, GW alerts were sent out in real time for a total of six binary neutron star (BNS) and nine neutron star black hole (NSBH) mergers. The worldwide effort by several teams to discover kilonovae and characterize each candidate was exemplary (for example, papers and GCN (gamma-ray coordination network) circulars by the ENGRAVE, GRANDMA, GOTO, ASAS-SN, PS1, MASTER-Net, SAGUARO (searches after gravitational waves using Arizona observatories), DES-GW, BOOTES (burst observer and optical transient exploring system), VINROUGE, J-GEM (Japanese collaboration for gravitational-wave electro-magnetic follow-up) teams).
The GROWTH team leveraged three discovery engines: Zwicky Transient Facility (ZTF), Dark Energy Camera (DECam) and Palomar Gattini-IR (PGIR). Among these three, ZTF had the widest field and fastest response to most triggers. DECam had the deepest sensitivity and constrained the best localized event, GW190814 8 . PGIR had the reddest sensitivity and a wide-field, but a relatively shallow response 9 . After candidates were identified, the GROWTH team used a worldwide network of telescopes to characterize the properties of each candidate. Early career researchers, especially graduate students and postdocs, were the critical powerhouse of the GROWTH effort. Each GW alert triggered an automated phonecall to the team from 'GROBOT' . If the trigger was deemed a 'go' , hours of non-stop hard work to vet candidates followed.
ZTF promptly followed up 13 of the 15 BNS/NSBH triggers 10 . Within minutes, multi-step machine learning algorithms automatically reduced a list of 2.1 million candidates that were spatially and temporally coincident to 2,199. Within a few hours, the GROWTH team's real-time analysis further reduced this list to 127 candidates that were announced publicly via GCN circulars. Next, follow-up data were collected by various teams worldwide to characterize these 127 candidate counterparts. The GROWTH collaboration posted 82 GCNs with follow-up data during O3. Another 151 GCNs refer to follow-up of ZTF objects by other teams. After extensive follow-up analysis, we concluded that none of the candidates we found were kilonovae.
One may say that nothing was detected because the localizations were too coarse, the median being 4,400 square degrees, two orders of magnitude more than GW170817. I believe that multiple wide-field surveys -such as ZTF, DECam and ATLAS (Asteroid Nature reviewS | PHysics Terrestrial-impact Last Alert System) -rose to the challenge and successfully mapped thousands of square degrees. One may say that nothing was detected because the distances were much too great, the median being 214 Mpc, over five times further than GW170817. I believe that many searches were sensitive enough to detect GW170817. The key is that the joint probability of zero detections for ZTF, assuming kilonovae were at least as bright as GW170817, was only 4% (ReF. 10 ). The bottom line, for me, is that the zero detections illuminate the underlying physics -every kilonova is just not as intrinsically luminous as the first one.

Looking ahead
Preparations are underway for an even more sensitive fourth run (O4; planned for 2022). I think that the single most important thing for a successful O4 is clear, transparent and prompt communication. The GW data analysis is complex, and key numbers used by astronomers to make decisions change dramatically between online and offline analyses. Whereas the online analysis takes a few minutes, the final offline analysis could take more than a year. The difference can be as stark as a GW event being real or not, the EM counterpart being spatially coincident with the GW skymap, or not, the GW merger having at least one neutron star, or not. Surely, there must be a middle ground. I urge multiple, more frequent updates from offline analysis results. This will facilitate faster progress towards answering multi-messenger GW + EM questions [8][9][10] .
Scientifically, now that we have 'heard' neutron stars merge with black holes, I look forward to seeing it. To overcome the opacity due to the electrons having millions of possible line transitions in heavy elements, two infrared surveyors are under construction: WINTER at the Palomar Observatory (USA) and DREAMS at the Siding Springs Observatory (Australia). To maximize the infrared mapping speed, a surveyor in the Antarctic is planned. Sometimes black holes may swallow neutron stars whole; combining GW information with EM upper limits will be the best way to characterize whether and/or when that happens. I await the moment when a black hole rips apart the neutron star expelling so much lanthanide-rich material that we see brilliant, infrared fireworks; I want to catch this act red-handed.