Why in the News
The upgraded Giant Metrewave Radio Telescope (uGMRT) at Khodad near Pune has measured the average hydrogen mass of 8,122 star-forming galaxies in the "COSMOS" field of the sky by combining (stacking) their 21-cm radio signals from atomic hydrogen - these galaxies appear to us as they were about 8-9 billion years ago, because their signal took that much time to reach the Earth. The research was published online in The Astrophysical Journal Letters (ApJL) on 25 September 2026 (open access); the release of NCRA-TIFR and the main news coverage came between 13 and 18 October 2026 - keep both dates separate. The first author is Balpreet Kaur (post-doctoral fellow at IUCAA, Pune; earlier a PhD at NCRA-TIFR); co-authors are Nissim Kanekar, Aditya Chowdhury and Jayaram N. Chengalur (NCRA-TIFR/TIFR). Summary of the finding: the stock of hydrogen in early galaxies was far more than in today's galaxies, but the rate of star formation was also many times faster - that is, the fuel was used up quickly, and when no new gas was received from outside, star formation slowed down.
Background: How Stars Form from Gas
In the universe, stars are formed from gas: atomic hydrogen first becomes molecular hydrogen, and stars are born from it. Astronomers have known for a long time that the star-formation activity of the universe was at its peak about 8-10 billion years ago and has been falling steadily since then - but why? The direct answer rests on measuring how much atomic hydrogen (the basic fuel of star formation) was present in the galaxies of that time. The problem is that the 21-cm radio signal of the hydrogen atom is extremely weak; catching it from a single distant galaxy is almost impossible even with today's technology. The basic scientific goal of building the GMRT about 40 years ago was this very thing - to catch the 21-cm signal from galaxies of the early universe. An upgrade about a decade ago (new receivers, new correlator, new electronics) gave the telescope new life. This measurement is a new step in the same direction: it is only the second measurement of the average hydrogen amount at z≈1 (an era about 8 billion years old) - the first was done in 2020 by a team of the same tradition on galaxies of the DEEP2 field - and the first in the COSMOS field. That is, there was also an independent confirmation of the first result.
Method: 216 Hours of Observation and the Art of "Stacking"
According to the research paper, the team observed the COSMOS field for 216 hours with uGMRT's Band-4 (550-850 megahertz) - a total area of 2.4 square degrees in 7 pointings, 67 hours in the central pointing and 89 hours in the other six. The signal of a single distant galaxy is so faint that it cannot be measured directly, so the "HI 21 cm stacking" technique was adopted: signals of galaxies with precise redshifts already known from optical telescopes were aligned at their expected frequency and averaged. In the institutional release Kaur spoke of aligning the signals of about 10,000 galaxies; in the abstract of the paper the final analysis is of 8,122 galaxies - both figures are of different stages, consider the analysis figure of 8,122 as the standard. The result: an average signal at the 4.2-sigma level at a spatial resolution of 90 kiloparsecs, and an average hydrogen mass of (15.5 ± 3.7) × 10⁹ solar masses. Also, by stacking the continuum images at 680 megahertz, a weighted-median star-formation rate of 8.90 ± 0.92 solar masses per year was derived from the 1.4 gigahertz radio continuum.
What Was Found: More Fuel, Even Faster Consumption
There are three main numerical findings. First, an average hydrogen mass of about 15.5 billion solar masses - a far larger stock than in today's ordinary galaxies. Second, the ratio of hydrogen to stellar mass is 1.15 ± 0.27 - that is, in the galaxies of that era the share of gas was considerably more than that of stars. Third, and most important, the "depletion timescale" - the time for the fuel to run out at the present rate - is merely 1.74 ± 0.45 billion years (at an average redshift of 1.081). That is, the galaxies of that era would turn their gas into stars in just a billion or two years; if the flow of new gas (accretion) from the intergalactic medium was not sufficient, star formation stopping was natural. According to the researchers, this gives a "natural explanation" of the fall in the star-formation rate of the universe after z<1 - this conclusion matches and strengthens the 2020 results of the DEEP2 survey. According to Kanekar this detection is an important step towards the bigger goal that no one has yet achieved - measuring the 21-cm signal from single galaxies of this same era; for this the team is doing a still deeper GMRT observation of the COSMOS field.
Limits and Cautions
Keep several things in measure. First, this is an average measurement - the stacking technique gives the combined signal of thousands of galaxies, not of any one galaxy; single-galaxy measurement is still the remaining goal. Second, the detection is at 4.2 sigma - a statistically solid measurement, but below the conventional 5-sigma "discovery" threshold of astrophysics; read it as accumulated evidence, not as a single decisive proof. Third, "8-9 billion years ago" is light-travel time (lookback time); the paper's average redshift is 1.081 - it tells the era, not the distance. Fourth, the explanation of the fall by a lack of gas accretion is an interpretive conclusion - the flow of new gas was not measured directly; it is the best explanation based on data consistent with DEEP2. Fifth, two figures for the number of galaxies are in circulation - "about 10,000" (aligned signals) in the release/news and 8,122 (final analysis) in the paper; we have adopted the one in the paper. Sixth, the star-formation rate is a median derived from radio-continuum stacking, not a separate rate for each galaxy. Seventh, keep the dates apart: arXiv preprint 11 September, paper online 25 September 2026, print issue 1 October (ApJL volume 1009, L38), while the NCRA release is of 13 October and the news coverage of 13-18 October - this is the public explanation of a month-old paper, not fresh news of a new laboratory discovery.
What to Watch Next?
Three things: (1) the team's ongoing deep GMRT observation - whether the 21-cm signal will reach single galaxies; (2) studies linking gas properties with the environment of galaxies from multi-wavelength data of the COSMOS field, as Chengalur hinted; and (3) the global context - South Africa's MeerKAT has also recently (a separate study, an ApJL paper of July 2026) detected gas 4-5 billion years old through hydrogen intensity mapping, and the under-construction SKA will take these same techniques to a large scale. From the policy point of view this work is an example of Indian infrastructure (uGMRT), the funding of the Department of Atomic Energy and the Anusandhan National Research Foundation (ANRF), and the human-capital cooperation of premier Indian institutions (NCRA-TIFR, IUCAA) - the same kind of open-access discussion like SERC/ONOS that is seen in other fields too.
Reference Points for the Reader
Remember four things. First, the 21-cm line: the radio signal emitted when the electron-proton spin of the hydrogen atom flips (rest frequency about 1420 megahertz); through redshift it shifts to low frequencies like Band-4. Second, redshift z≈1 means the era of the universe when its age was less than half - lookback time about 8 billion years. Third, stacking: the average of many weak signals - it tells the collective, not the individual, truth. Fourth, the GMRT: an array of 30 forty-five-metre dishes at Khodad near Pune, operated by NCRA (a constituent of TIFR, Department of Atomic Energy) - in Prelims the institution, the place and the purpose have all been asked.
Mains Practice Question
'The mystery of the falling rate of star formation in the universe has faced space science for decades.' In support of this statement, evaluate the contribution of the recent uGMRT HI 21 cm stacking measurement, and comment on the strategic role of Indian radio-astronomy infrastructure. (General Studies-III: Science and Technology; 250 words, 15 marks)
Sources and Limits
- Primary research paper (The Astrophysical Journal Letters, open access): Kaur, Kanekar, Chowdhury, Chengalur, "The GMRT CAT z1-COSMOS Survey: HI 21 cm Emission from Star-forming Galaxies at z≈1 in the COSMOS Field", volume 1009, L38; online 25 September 2026; DOI 10.3847/2041-8213/aea3e8. Abstract, author affiliations, dates and funding read.
- arXiv preprint 2609.13382 (submitted 11 September 2026): full abstract - 216 hours, 8,122 galaxies, 4.2σ, 15.5±3.7×10⁹ M☉, SFR 8.90±0.92, ratio 1.15±0.27, depletion 1.74±0.45 Gyr.
- Crossref record: confirmation of received 25 June, accepted 4 September, online 25 September, print 1 October 2026; DAE/DST grants.
- NCRA-TIFR release (13 October 2026): quotes of Kaur/Chowdhury/Chengalur/Kanekar, the claim of "second measurement", DAE+ANRF funding.
- Times of India (13-14 October 2026): detailed coverage, 200+ hours, quotes.
- Deccan Herald (16-17 October 2026) and The Intelligent Hindi (18 October 2026): summary of the finding and importance (news).
- Limits: average (stack) measurement, not a single-galaxy detection; 4.2σ (below 5σ); the gas-accretion explanation is inferred; number of galaxies "about 10,000" (release) versus 8,122 (paper) - paper is the standard; star-formation rate is a median; date difference between paper (25 September) and release (13 October); the MeerKAT mention is only an independent-study reference.
Download This Article: PDF