Collaborative research: 21 CM power spectrum analysis of hydrogen epoch of reionization array phase II observations

The National Science Foundation Division of Astronomical Sciences awarded the Board of Regents of Nevada System of Higher Education, doing business as University of Nevada, Las Vegas (UNLV), $210,198 in project grant funding on September 1, 2026, under the Mathematical and Physical Sciences program (CFDA 47.049) for collaborative research analyzing observations from the Hydrogen Epoch of Reionization Array (HERA).

UNLV and a research team from the University of California, Berkeley will analyze three years of HERA data—collected by an NSF-supported radio telescope array in South Africa—to develop artificial intelligence-based techniques for separating the faint hydrogen signal from cosmic dawn (the era of the first stars and galaxies) from bright radio contamination originating in nearby galaxies, the Milky Way, human-caused radio interference, and instrumental systematics. The central technical challenge is that galactic and extragalactic radio emission outshines the target cosmological signal by approximately 100,000 to 1. The work directly observes the impact of the universe's first luminous objects by detecting the distinct 21-centimeter radio signature of neutral hydrogen in the intergalactic medium during the epoch of reionization and earlier cosmic dawn.

The project will also broaden participation in cutting-edge research by providing paid summer research experiences in data analysis for undergraduates at both Berkeley and UNLV, including mentorship and training designed to build pathways to graduate school and careers in science and technology.

Performance occurs in Las Vegas, Nevada from September 1, 2026, through August 31, 2028.

Abstract

The cosmic dawn?the ERA when the first STARS and black holes formed and LIT up the universe?is one of the last unexplored frontiers in the history of our cosmos. As those first generations of STARS and galaxies formed, they heated and ultimately ionized the hydrogen gas that filled space between them, ending the cosmic dark ages. That hydrogen gas has a distinct radio signature that astronomers can detect to directly observe the impact of the universe's first luminous objects. The hydrogen epoch of reionization array (HERA), an NSF-supported radio telescope made of hundreds of dishes in south africa, was purpose-built to detect this signal. The central challenge is that radio emission from our own galaxy and others outshines the cosmic dawn signal by a factor of roughly 100,000 to 1. A team of scientists from the university of california, berkeley and the university of nevada, LAS vegas will analyze three years of already-collected HERA data?the most sensitive data set of its kind?and develop new AI-based techniques for separating the faint cosmological signal from this bright contamination, as well as from human-caused radio interference and systematics introduced by the telescope itself. The project will also broaden the participation in cutting-edge research by providing paid summer research experiences in data analysis for undergraduates at berkeley and unlv, complete with mentorship and training designed to build pathways to graduate school and careers in science and technology. HERA is an interferometer optimized to measure 21 CM emission from neutral hydrogen in the intergalactic medium (IGM) during the epoch of reionization (6 ? Z ? 12) and the earlier cosmic dawn, constraining the timing and morphology of reionization, the properties of the first galaxies, and the early sources of IGM heating. This project will analyze three years of HERA phase II observations already in hand: 579 nights with ~190 working antennas over 50?230 MHZ. This data set is roughly 30 times more sensitive to the 21 CM power spectrum than any published measurement and is the first with a realistic likelihood of detecting the cosmological signal at 6 ? Z ? 9. Building on the power spectrum estimation and foreground avoidance techniques the team pioneered and applied to produce the current world-leading limits with HERA phase I, the team will produce and validate the most sensitive 21 CM power spectra to date. To maximize sensitivity, the team will advance the state of the art in mitigating instrumental systematics, using phase II data as a testbed for improved foreground filtering and subtraction, bandpass calibration, machine-learning-based radio frequency interference detection, and DE-mixing of antenna-to-antenna mutual coupling. A detection would tightly constrain the escape fraction of ionizing photons from early galaxies and the electron scattering optical depth, a key CMB degeneracy; the techniques developed will guide the design of next-generation 21 CM arrays. The project's broader impacts recreate HERA's successful champ program locally, funding eight berkeley and unlv undergraduates per summer for cohort-based research experiences that culminate in presenting at the AAS meeting. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the foundation's intellectual merit and broader impacts review criteria.- subawards are not planned for this award.