How astronomers measured primordial helium abundance to 0.5%

Key takeaways

• The LBT Y_p Project, a team led by astronomers at the University of Minnesota and Ohio State University, reports a primordial helium-4 mass fraction of Y_p = 0.2458 ± 0.0013. That is a precision of about 0.5%, which the team calls the most precise determination to date.
• The value comes from spectra of metal-poor star-forming regions observed with the Large Binocular Telescope (LBT) in Arizona. The final average uses 15 low-metallicity targets.
• It agrees with the standard Big Bang prediction, Y_p = 0.2467 ± 0.0002, which is calculated from the baryon density measured by the Planck satellite.
• Combined with deuterium and cosmic microwave background (CMB) data, the result constrains the number of light neutrino-like species in the early universe to N_ν = 2.925 ± 0.082. That is consistent with the three neutrino types of the Standard Model.
• The papers appeared as a five-paper series in The Astrophysical Journal in September 2026, and science news sites have recently highlighted the result again.

Helium as a fossil from the first minutes

The primordial helium abundance, the helium-4 forged in the universe's first minutes, has been measured at Yp = 0.2458 ± 0.0013, a precision of about 0.5%. The LBT Yp Project used the Large Binocular Telescope, and the value agrees with Planck-based predictions. On Earth, helium is now a policy issue: the US has named helium a critical material, citing chipmaking.

Helium-4 is the main product. In the standard picture, reviewed by Richard Cyburt, Brian Fields, Keith Olive and Tsung-Han Yeh in Reviews of Modern Physics, roughly a quarter of the mass of ordinary matter ended up as helium-4. Astronomers call this mass fraction Y_p, where "p" stands for primordial. According to an Ohio State University summary of the new work, about 90% of the helium in today's universe was made in the Big Bang. The other 10% has been added by stars over roughly 13.5 billion years.

The helium fraction is set mainly by how many neutrons were around when nuclei could finally form. Early on, neutrons and protons constantly converted into each other through weak nuclear reactions. Within about the first second, the universe cooled enough that these reactions could no longer keep up with the expansion. The neutron-to-proton ratio then "froze out" at roughly one neutron for every six protons. Free neutrons decay, so the ratio kept falling slowly until nuclei began to form a few minutes later. Almost all the surviving neutrons were locked into helium-4. That is why Y_p works like a stopwatch for the first second of cosmic history.

Why helium is sensitive to new particles

The timing of that freeze-out depends on how fast the universe was expanding. The expansion rate in turn depends on the total energy in light, fast-moving particles: photons, the three known neutrino types, and any unknown "dark radiation." Add an extra light species and the universe expands faster. Freeze-out then happens earlier, when more neutrons remain, and more helium is made. Physicists summarise this with a parameter for the effective number of neutrino-like species, written N_ν or N_eff. The Standard Model of particle physics has three. A precise helium measurement therefore tests whether anything else was present one second after the Big Bang.

How the LBT team measured it

Primordial helium cannot be observed directly, because stars have been adding helium ever since. Astronomers instead look for gas that has been processed as little as possible: ionised clouds of hydrogen (H II regions) in small, faint galaxies that contain very few heavy elements. Astronomers use oxygen as the tracer for these "metals." Gas with very little oxygen has seen little star formation, so its helium content should be close to the primordial value.

In the fourth paper of the series, Erik Aver, Evan Skillman, Richard Pogge and colleagues describe how they obtained uniform spectra of 54 metal-poor H II regions with the LBT on Mount Graham, Arizona. The telescope carries two 8.4-metre mirrors on a single mount. The team used two of its instruments:

• MODS (Multi-Object Double Spectrographs) recorded optical and near-infrared emission lines of hydrogen and helium.
• LUCI, a near-infrared spectrograph, measured the helium line at 10,830 ångströms. That line is especially sensitive to the density of the gas, which helps pin down the conditions in each cloud.

According to a University of Minnesota release, the project used 130 hours of observing time on the LBT.

Turning line brightnesses into a helium abundance is the hard part. The team fitted nine physical parameters for every target at once, including electron temperature, electron density, dust reddening, the optical depth of helium lines, and the stellar absorption underneath the emission lines. They then removed targets whose fits were poor or whose spectra showed warning signs. These signs included unusual stellar features, very high optical depth, or temperature measurements that did not agree. Of the 54 regions, 47 passed a statistical goodness-of-fit cut and 41 remained after screening.

Earlier studies usually plotted helium against oxygen and extrapolated the trend to zero oxygen. For this analysis, the authors argue that their 15 lowest-metallicity, highest signal-to-noise targets are close enough to pristine that a weighted average of those 15 is "well-justified and produces a robust result." The result is Y_p = 0.2458 ± 0.0013.

The numbers side by side

The table lists published research values. They do not change over time, so no "data as of" timestamp applies.

Why it agrees with the Planck-based prediction

In standard BBN with three neutrino types, the light-element abundances depend on one main number: how many baryons (protons and neutrons) there are per photon. The Planck satellite measured this independently from the cosmic microwave background, the light released about 400,000 years after the Big Bang. Feeding Planck's baryon density into BBN calculations predicts Y_p = 0.2467 ± 0.0002.

The LBT value, 0.2458 ± 0.0013, differs from that prediction by 0.0009. That is less than one standard deviation of the measurement. In the fifth paper, Tsung-Han Yeh, Keith Olive, Brian Fields and co-authors combine the helium value with the latest primordial deuterium measurement and with CMB data. They find that the baryon density from the early-universe elements, Ω_Bh² = 0.02236 ± 0.00014, matches Planck's 0.0224 ± 0.0001. The authors write that their results "demonstrate consistency with the Standard Model of particle physics" and with the standard picture linking BBN, about one second after the Big Bang, to the CMB some 400,000 years later.

That agreement matters because the two probes are independent. Helium is measured in nearby dwarf galaxies with spectrographs. The baryon density is measured from temperature patterns in the oldest light in the sky. Different physics, instruments and teams arrive at the same picture of the universe's first minutes.

What it says about extra neutrino-like particles

Many proposed extensions of the Standard Model include new light particles, such as sterile neutrinos, axion-like particles or a "dark sector" with its own radiation. If such particles were abundant during BBN, they would have sped up the expansion and raised the helium fraction.

When Paper V lets the number of light species vary, the data give N_ν = 2.925 ± 0.082. That is consistent with three. If at least the three known neutrinos are assumed, any extra contribution is limited to ΔN_ν ≤ 0.125 at 95% confidence. For comparison, Planck 2018 combined with baryon acoustic oscillation data gives N_eff = 2.99 ± 0.17. The two numbers describe slightly different epochs, but the helium-based constraint is tighter.

A follow-up preprint posted in September 2026 by Marilena Loverde, Murali Saravanan and Zachary Weiner, which has not yet been peer reviewed, explores the consequences. It argues that the LBT value lets cosmologists analyse CMB data without assuming standard BBN, at almost no cost in precision. It also reports that the measurement rules out most, but not all, of the non-standard models the authors tested, including some proposed to ease the "Hubble tension." It notes an indirect link to the neutron-lifetime puzzle, in which two kinds of laboratory experiments disagree. The preprint finds the helium result consistent with "bottle" experiments and 2.4 standard deviations below "beam" experiments.

Caveats and open questions

• Systematic effects dominate. The authors identify stellar absorption underneath the helium lines, radiative transfer in dense clouds, and mismatched temperature estimates as the main risks. Targets affected by these were flagged or excluded.
• Sample selection. The final value rests on 15 targets. Public descriptions differ in how they count the sample. The Ohio State release refers to 48 high-quality samples, while the paper reports 54 observed, 47 after the fit cut and 41 after screening. Readers should rely on the paper's own numbers.
• Atomic physics. The analysis uses updated helium and hydrogen emissivities and new radiative-transfer calculations. Future revisions to atomic data could shift the value slightly.
• Lithium remains a puzzle. The BBN review by Cyburt and colleagues notes that predicted lithium-7 still disagrees with observations. The helium result does not resolve that tension.

The researchers framed the result in strong terms. "We promised a half-percent uncertainty in our proposal, and we got there," Evan Skillman of the University of Minnesota said in the university's release. Richard Pogge of Ohio State said the team had "learned something fundamental about the universe."

Closer to home, planetary scientists are chasing a different origin story. Japan's MMX mission is heading for Mars to collect soil from the moon Phobos and return it in 2031, which could help settle where the Martian moons came from; see our explainer on JAXA's Phobos sample mission.

What it means

For cosmology, the LBT measurement turns helium from a loose consistency check into a precision tool. A 0.5% value makes it about three times more precise than the best previous emission-line estimate. It is also roughly ten times more precise than what current CMB data can achieve on their own. The result supports the standard model of the early universe: three neutrino types, a baryon density that matches Planck, and no clear sign of extra radiation at the one-second mark.

For particle physics, the tighter limit on extra light species narrows the room for new particles that would have been abundant during BBN. Theories that predict such particles must now fit within a smaller margin. The Loverde preprint goes further. It estimates that, in the cases the authors studied, future CMB surveys would improve on the LBT helium constraint only modestly, by no more than about a factor of two.

What remains unknown is whether still-unrecognised systematic errors in nebular spectroscopy could move the value. Independent checks will matter, from other telescopes and from future high-resolution CMB surveys that can measure helium directly. For now, independent measurements made by very different methods point to the same answer.

Related coverage

• Orionid Meteor Shower 2026 Peak: When and How to Watch Oct 21-22
• NASA-led model doubles Thwaites Glacier sea-level projection
• How Common Is MS Misdiagnosis? What the Research Shows

Sources

• Aver et al., "The LBT Y_p Project IV: A New Value of the Primordial Helium Abundance," arXiv:2601.22238
• Yeh et al., "The LBT Y_p Project V: Cosmological Implications of a New Determination of Primordial ⁴He," arXiv:2601.22239
• Ohio State News, "Using helium, astronomers blow certainty into early Big Bang conditions" (9 September 2026)
• University of Minnesota via Newswise, "Researchers pinpoint key early universe measurement with record precision" (9 September 2026)
• Planck Collaboration, "Planck 2018 results. VI. Cosmological parameters," arXiv:1807.06209
• Loverde, Saravanan and Weiner, "Big Bang For Your Helium Buck," arXiv:2609.13140 (preprint)
• Cyburt, Fields, Olive and Yeh, "Big Bang Nucleosynthesis: 2015," Reviews of Modern Physics 88, 015004