Papex

Open problem · Physics · Open

Neutron lifetime puzzle

Why do the two main ways of measuring the free neutron’s lifetime give different results?

Summary

Experiments that count neutrons surviving in a trap (“bottle”) and those that count protons from decays in a beam disagree on the free neutron’s lifetime by more than their stated uncertainties; the cause of the gap is unknown.

Why it is open

In the Standard Model, a free neutron decays into a proton, an electron and an antineutrino, and its lifetime is measured in two ways [2]. In bottle experiments, ultracold neutrons, slow enough to be kept in a container, are stored for a time comparable to their lifetime and the survivors are counted [2]. In beam experiments, a cold neutron beam passes through a trap that catches the decay protons, and the ratio of the absolute numbers of neutrons and protons gives the lifetime [13]. In 2018 Bartosz Fornal and Benjamín Grinstein put the bottle average at 879.6 ± 0.6 s and the beam average at 888.0 ± 2.0 s, a discrepancy of 4.0 standard deviations (σ) [2]. Writing in 2024, Fred Wietfeldt gave the difference as 9.8 ± 2.0 s [10].

The lifetime matters beyond neutron physics: it affects the abundance of light elements formed in the early universe [10], and a 2023 white paper from the US neutron-physics community sets a lifetime precision of about 0.3 s as the target for determining V_ud, a basic parameter of the weak interaction, from neutron decay as precisely as from nuclear decays [9].

Fornal and Grinstein stated the alternatives: "either one of the measurement methods suffers from an uncontrolled systematic error, or there is a theoretical reason why the two methods give different results" [2]. Wietfeldt judges that the discrepancy "probably has an experimental origin", with new physics "a more exciting prospect" [10].

No body sets a bar for a resolution. The puzzle would end if a cause were found that accounts for the gap, or if independent methods converged on one value; the planned routes include a new proton-counting beam experiment aiming at the 0.3 s level [15] and a direct measurement of the fraction of neutron decays that are ordinary beta decays [9].

What is known

Beam measurements (2005–2013). BL1, the proton-counting beam experiment at the National Institute of Standards and Technology (NIST) described in 2013 as the most precise beam measurement, gave 886.3 ± 1.2 (statistical) ± 3.2 (systematic) s in 2005, with a dominant uncertainty of 2.7 s from the absolute count of neutrons in the beam [1, 9]. In 2013 the group measured its neutron monitor's efficiency five times more precisely, found it consistent with the earlier value and updated the lifetime to 887.7 ± 1.2 ± 1.9 s [1].

Bottle measurements (2021–2025). The UCNτ experiment at the Los Alamos Neutron Science Center stores ultracold neutrons in a bowl-shaped magnetic trap from which they lack the energy to escape against gravity, and counts the survivors with a detector lowered into the trap [7, 9]. From about 38 million surviving neutrons counted in 2017 and 2018, the collaboration reported 877.75 ± 0.28 (statistical) +0.22/−0.16 (systematic) s in 2021 [7]. Adding data from 2020 to 2022, taken with four detector geometries and a smaller correction for neutrons scattered by residual gas, it reported an average of 877.83 ± 0.22 +0.20/−0.17 s in 2025 [12]. The Particle Data Group's 2022 world average, 878.4 ± 0.5 s, carries a scale factor of 1.8 on its uncertainty, which the 2023 white paper reads as "suggesting underestimated systematic uncertainties even within the UCN experiments" [9].

A link to the axial coupling (2018). Andrzej Czarnecki, William Marciano and Alberto Sirlin related the lifetime to the axial coupling g_A, a constant measured from asymmetries in neutron decay; with the average of g_A measurements published since 2002, the relation favours the bottle value [3].

Testing a dark decay (2018–2022). A beam experiment counts protons, so it measures the lifetime divided by the fraction of decays that produce a proton [2]. Fornal and Grinstein proposed that about 1% of neutrons decay into dark-sector particles, one of which could be dark matter [2]. Requiring beryllium-9 to stay stable confines the dark final state to a narrow mass window just below the neutron mass; a decay into a dark particle and a photon would emit a photon of 0.782–1.664 MeV [2]. A search for such photons from neutrons stored in a coated stainless-steel bottle excluded a branch large enough to explain the discrepancy with 97% confidence [5]. Data from the UCNA (Ultracold Neutron Asymmetry) experiment excluded a dark particle plus an electron–positron pair as the dominant channel at well above 5σ for pair energies of 100–644 keV [4]. Gordon Baym and colleagues showed that if neutrons could decay into a stable, feebly interacting dark fermion, no neutron star could exceed 0.7 solar masses, yet all well-measured ones exceed one [6]. Czarnecki and colleagues derived a general limit of 0.27% on exotic branching ratios [3]. A separate proposal, that neutrons convert into "mirror neutrons" in the 4.6 T magnetic field of the NIST beam experiment, was excluded in 2022 by a search in a 6.6 T field at the Spallation Neutron Source [8].

Recent results (2024–2026). In December 2024 a team at the Japan Proton Accelerator Research Complex (J-PARC) posted a beam measurement that counts decay electrons, not protons, relative to neutron captures on helium-3 in a gas detector, so its systematic uncertainties differ from earlier beam experiments' [11]. Its value, 877.2 ± 1.7 +4.0/−3.6 s, is consistent with the bottle method and in 2.3σ tension with the proton-counting beam average; the team's first result, 898 ± 10 +15/−18 s, had been consistent with both [11]. The preprint had no journal reference on 15 October 2026 [11]. Its four data subsets combine with χ²/DOF = 15.8/3, and "the underlying cause of this deviation remains undetermined", the authors write [11]. In a single-author 2026 paper, Ashish Desai argues that a pressure-dependent detector response is "a plausible explanation" for such variations [14]. A 2025 NIST study showed that trapped protons can exchange charge with molecular hydrogen, but concluded that the NIST beam result "is unlikely to have been significantly affected" [13].

Main approaches

An error in the beam measurements. Proton-counting experiments could lose, gain or misidentify protons, or miscount the neutrons in the beam [1, 13]. Strongest evidence: the electron-counting J-PARC measurement agrees with the bottles [11], and the relation to g_A favours the bottle value [3]. Main difficulty: two specific effects tested so far, the neutron monitor's efficiency and charge exchange with hydrogen, did not change the NIST result [1, 13], and the J-PARC result is a preprint whose subsets disagree with one another [11, 14].

An error in the bottle measurements. Storage experiments could underestimate their systematic uncertainties [9]. Strongest evidence: the scale factor on the world average, which the white paper takes to suggest underestimated uncertainties among the storage experiments [9]. Main difficulty: the UCNτ magnetic trap combines "extremely low UCN loss" with detection inside the trap [9], its 2020–2022 runs, with four detector geometries, gave 877.94 ± 0.37 s against the 877.75 s it published in 2021 [7, 12], and both the electron-counting beam result and the g_A relation agree with the bottle value [3, 11].

New decay or conversion channels. About 1% of neutron decays produce no proton, which only the beam method would notice [2]. Strongest evidence: Fornal and Grinstein built models that fit the size of the gap and, with the dark particles' masses in a narrow window, evade proton-decay and nuclear-stability limits [2]. Main difficulty: the photon and electron–positron channels are excluded as the explanation [4, 5], neutron-star masses require a multiparticle dark sector "with highly constrained interactions" [6], the general limit of 0.27% on exotic branches is below the 1% needed [2, 3], and mirror-neutron conversion is excluded [8]. A decay into dark particles only, which leaves no photon or electron to detect, "relies on higher order radiative processes" [2].

Decisive new measurements. Methods with different systematics could show which side is wrong. Strongest evidence: the J-PARC result shows that a beam method with different systematics can land on the bottle value [11]; BL3 at NIST is designed to raise the proton trapping rate a hundredfold and to test beam systematics at the 10⁻⁴ level [15], and UCNProβe at Los Alamos to measure the beta-decay branching ratio directly [9]. Main difficulty: none has reported yet, and schedules have moved. In 2023 BL3's first run was planned for 2026 [9]; a June 2025 abstract expects it to run after the NIST Center for Neutron Research returns to operation, expected in 2027 [15].

Open sub-questions

  1. Pressure dependence at J-PARC. Does a pressure-dependent detector response explain the spread among J-PARC's four data subsets, and does correcting for it move the combined value towards the bottle or the beam average [11, 14]?
  2. Proton-trap systematics. Beyond charge exchange with molecular hydrogen, which process in a proton trap could shift a proton-counting result by about 1%, and could BL3 test it at its 0.3 s goal [13, 15]?
  3. Direct branching ratio. How far below the 0.27% bound derived from g_A could a direct measurement of the beta-decay branching ratio, such as UCNProβe's, push exotic decays [3, 9]?
  4. Fully dark decays. For a neutron decaying only into dark particles, which dark-sector interactions keep neutron stars above two solar masses while allowing a 1% branch [2, 6]?
  5. Trap systematics at 0.1 s. Does the UCNτ lifetime depend on detector geometry or the stored neutrons' energy spectrum at the 0.1 s level that the τSPECT and PENeLOPE traps target [9, 12]?

Work on this in Papex

If you have a testable answer to one of the sub-questions above, you can draft it as a private hypothesis in Papex, or prepare the draft from Claude or ChatGPT with the Papex connector. You review the exact version, sign it and decide whether to publish it; a public record that names this problem is listed on this page. A Papex record shows who signed which version and when. It is not peer review and does not mean the hypothesis is correct.

Key references

  1. Yue, A. T., Dewey, M. S., Gilliam, D. M., Greene, G. L., Laptev, A. B., Nico, J. S. et al. (2013). Improved determination of the neutron lifetime. Physical Review Letters 111(22), 222501. 10.1103/PhysRevLett.111.222501 · Checked 15 Oct 2026
  2. Fornal, B. & Grinstein, B. (2018). Dark matter interpretation of the neutron decay anomaly. Physical Review Letters 120(19), 191801; erratum in 124(21), 219901 (2020), DOI 10.1103/PhysRevLett.124.219901, not read. 10.1103/PhysRevLett.120.191801 · Checked 15 Oct 2026
  3. Czarnecki, A., Marciano, W. J. & Sirlin, A. (2018). Neutron lifetime and axial coupling connection. Physical Review Letters 120(20), 202002. 10.1103/PhysRevLett.120.202002 · Checked 15 Oct 2026
  4. Sun, X., Adamek, E., Allgeier, B., Blatnik, M., Bowles, T. J., Broussard, L. J. et al. (2018). Search for dark matter decay of the free neutron from the UCNA experiment: n → χ + e⁺e⁻. Physical Review C 97(5), 052501. 10.1103/PhysRevC.97.052501 · Checked 15 Oct 2026
  5. Tang, Z., Blatnik, M., Broussard, L. J., Choi, J. H., Clayton, S. M., Cude-Woods, C. et al. (2018). Search for the neutron decay n → X + γ, where X is a dark matter particle. Physical Review Letters 121(2), 022505. 10.1103/PhysRevLett.121.022505 · Checked 15 Oct 2026
  6. Baym, G., Beck, D. H., Geltenbort, P. & Shelton, J. (2018). Testing dark decays of baryons in neutron stars. Physical Review Letters 121(6), 061801. 10.1103/PhysRevLett.121.061801 · Checked 15 Oct 2026
  7. Gonzalez, F. M., Fries, E. M., Cude-Woods, C., Bailey, T., Blatnik, M., Broussard, L. J. et al. (2021). Improved neutron lifetime measurement with UCNτ. Physical Review Letters 127(16), 162501. 10.1103/PhysRevLett.127.162501 · Checked 15 Oct 2026
  8. Broussard, L. J., Barrow, J. L., DeBeer-Schmitt, L., Dennis, T., Fitzsimmons, M. R., Frost, M. J. et al. (2022). Experimental search for neutron to mirror neutron oscillations as an explanation of the neutron lifetime anomaly. Physical Review Letters 128(21), 212503. 10.1103/PhysRevLett.128.212503 · Checked 15 Oct 2026
  9. Alarcon, R., Aleksandrova, A., Baeßler, S., Beck, D. H., Bhattacharya, T., Blatnik, M. et al. (2023). Fundamental neutron physics: a white paper on progress and prospects in the US. Preprint, arXiv:2308.09059 (version 1, 17 August 2023), not peer reviewed. Source · Checked 15 Oct 2026
  10. Wietfeldt, F. E. (2024). The neutron lifetime discrepancy and its implications for cosmology and dark matter. Symmetry 16(8), 956. 10.3390/sym16080956 · Checked 15 Oct 2026
  11. Fuwa, Y., Hasegawa, T., Hirota, K., Hoshino, T., Hosokawa, R., Ichikawa, G. et al. (2024). Improved measurements of neutron lifetime with cold neutron beam at J-PARC. Preprint, arXiv:2412.19519 (version 1, 27 December 2024), not peer reviewed. Source · Checked 15 Oct 2026
  12. Musedinovic, R., Blokland, L. S., Cude-Woods, C. B., Singh, M., Blatnik, M. A., Callahan, N. et al. (2025). Measurement of the free neutron lifetime in a magneto-gravitational trap with in situ detection. Physical Review C 111(4), 045501. 10.1103/PhysRevC.111.045501 · Checked 15 Oct 2026
  13. Caylor, J., Biswas, R., Crawford, B., Dewey, M. S., Fomin, N., Greene, G. L. et al. (2025). Detection of molecular hydrogen in a neutron beam lifetime experiment. Physical Review C 112(6), 065501. 10.1103/nr3b-3dtl · Checked 15 Oct 2026
  14. Desai, A. M. (2026). Pressure-dependent detector effects in beam-based neutron lifetime measurements. The European Physical Journal A 62(7), 154. 10.1140/epja/s10050-026-01929-x · Checked 15 Oct 2026
  15. Wietfeldt, F. (2025). The BL3 beam neutron lifetime experiment. Talk abstract, 15th Conference on the Intersections of Particle and Nuclear Physics (CIPANP 2025), Madison, 11 June 2025. Source · Checked 15 Oct 2026

Page reviewed 15 Oct 2026

Source

Open, as listed by Wikipedia · Unsolved problems in physics. Checked 08 Oct 2026.

The status is the source’s. Papex does not decide whether a problem is solved.

Records on this problem

No public record addresses this problem yet.

More in Physics