Papex

Open problem · Cosmology · Open

Cosmological lithium problem

Why do very old stars hold less lithium than Big Bang nucleosynthesis predicts?

Summary

The amount of lithium-7 predicted to come from Big Bang nucleosynthesis differs from the amount observed in very old stars, and the discrepancy is unexplained.

Why it is open

Big Bang nucleosynthesis (BBN) describes how the lightest nuclei, deuterium, helium-3, helium-4 and lithium-7, were made between about one second and three minutes after the Big Bang [5]. Once measurements of the cosmic microwave background (CMB) fix the density of ordinary matter, standard BBN becomes a "zero-parameter theory" [13]. Deuterium and helium-4 then agree with observations; lithium-7 does not [5, 13].

The size of the gap is stated in different ways. Brian Fields wrote in 2011 that with the first-year data of the WMAP satellite the discrepancy was a factor of 2–3, and that it had since grown to a factor of 3–4, or 4–5σ [5]. The Particle Data Group (PDG) review revised in August 2025 compares a predicted primordial ratio Li/H = (4.72 ± 0.7) × 10⁻¹⁰ with (1.45 ± 0.25) × 10⁻¹⁰ inferred from halo stars: a factor of 3.3, "corresponding to a 4.4σ discrepancy" [13].

The proposed solutions fall into three classes: the stellar measurements do not give the primordial abundance, the nuclear reaction rates are wrong, or new physics acted during or after BBN [5]. The PDG review notes that the stellar value "should be regarded as a lower bound rather than a direct measurement", because stars may have destroyed some of their lithium [13].

No body sets a bar for a resolution. The sources point to what would settle it: a stellar depletion mechanism that reproduces both the thin plateau of lithium in most old stars and the scatter at the lowest metallicities [4, 13], lithium measurements that stellar processing cannot touch [6, 13], or new physics that lowers lithium without spoiling the precisely measured deuterium [13]. In its 2025 revision the PDG review calls the problem "an unresolved issue in BBN" [13].

What is known

The Spite plateau (1982–2010). In 1982 M. Spite and F. Spite reported that extreme population II dwarfs, very old and very metal-poor stars, have significantly less lithium than younger stars, and noted that, if attributed to the Big Bang, this abundance suggests a rather low density of ordinary matter [1]. In the hottest halo stars, which have thin convection zones, lithium hardly varies with iron abundance; this flat trend, the "Spite plateau", is read as a sign that the lithium is primordial [5]. The measurement rests on one lithium line at 670.8 nm, which sees only neutral lithium, while most lithium in these stars is ionised; the correction depends steeply on the star's temperature [5]. Fields reported that raising the temperature scale by about 500–600 K would remove the problem, but that later studies agreed with the standard scale [5].

Measuring metallicity as the logarithm of a star's iron-to-hydrogen ratio relative to the Sun's, L. Sbordone and colleagues observed 28 halo dwarfs with metallicities between −2.5 and −3.5 and established a "meltdown" of the plateau below about −3: the scatter doubles towards lower metallicity, while the plateau above −2.8 is very thin [4]. No star below −3 lies above the plateau, which they read as stars born at plateau level and depleted later [4]. Stars accreted into the Milky Way from other galaxies show the same lithium as native stars [13].

Stellar depletion (2006–2022). A. J. Korn and colleagues observed 18 stars at different evolutionary stages in the metal-poor globular cluster NGC 6397 [2]. Element-specific abundance trends matched stellar models with atomic diffusion, which carries lithium below the visible surface, and turbulent mixing; the model gives an original lithium abundance 78% higher than measured in stars near the main-sequence turn-off and on the subgiant branch, which agrees with the CMB plus BBN value within the 1σ errors [2]. Fields cautions that systematic differences between lithium in globular clusters and in field stars raise concern about using clusters to constrain primordial lithium [5]. Xiaoting Fu and colleagues modelled lithium destroyed before the main sequence and partly restored by late accretion of gas, and reproduced the plateau for stars of 0.62–0.80 solar masses [7]. A key objection to depletion was the reported detection of the more fragile lithium-6 in some plateau stars [5]. In 2022 Fields and Keith Olive noted that newer searches give only firm upper limits on lithium-6, implying its depletion by factors up to 50, and that an equal depletion of lithium-7 "is more than sufficient to resolve" the problem [11].

Interstellar gas (2012). J. Christopher Howk and colleagues measured interstellar lithium-7 in the Small Magellanic Cloud (SMC), a nearby galaxy with a quarter of the Sun's metallicity, where stellar depletion plays no part [6]. They found it "nearly equal to the BBN predictions", which standard BBN can match only with "an extremely fine-tuned depletion of stellar Li with metallicity"; the result is also consistent with non-standard BBN [6]. The 2025 PDG review summarises the SMC measurements as "consistent with Galactic halo stars" [13].

Nuclear experiments (2006–2018). Most primordial lithium-7 is made as beryllium-7 [15]; its production is dominated by ³He(α,γ)⁷Be, and its destruction runs through ⁷Be(n,p)⁷Li followed by ⁷Li(p,α)⁴He [5]. The LUNA collaboration measured the production reaction underground at Gran Sasso, at energies directly relevant to BBN, with a total uncertainty of 4% [3]; Fields notes that solving the problem this way would need its rate to be 3–4 times lower, which solar neutrino data exclude [5]. At CERN the n_TOF collaboration measured ⁷Be(n,α)⁴He, finding "a minor role of this reaction in BBN" [8], and ⁷Be(n,p)⁷Li, finding "only a relatively minor improvement" [10]. F. Hammache and colleagues searched carbon-10 and carbon-11 for resonances that could destroy beryllium-7 through reactions with helium-3 or helium-4, and their results did not support the first channel and excluded the second [9].

Recent claims. In September 2025 Oswaldo Miranda, in Astronomy & Astrophysics, coupled a chemical evolution model, with yields from Population III and Population II stars, to hierarchical structure formation and concluded that "there is no so-called cosmological lithium problem"; the model rests on an ansatz, a primordial infall term, and gives a plateau near Li/H ∼ 1.81 × 10⁻¹⁰ [12]. The PDG review, revised in August 2025, calls the problem unresolved [13]. In October 2026 Simone Taioli and colleagues, also in Astronomy & Astrophysics, computed weak and radiative destruction channels of beryllium-7 from first principles and found only percent-level corrections that "cannot resolve the cosmological lithium problem" [15].

Main approaches

Stellar depletion. The stars were born with the BBN abundance and have since lost lithium through diffusion, rotational mixing or burning before the main sequence [13]. Strongest evidence: diffusion models fitted in NGC 6397 [2] and the lithium-6 limits [11], which the PDG review says strengthen "the case for stellar depletion as the leading resolution" [13]. Main difficulty: reaching the required level "generally demands some degree of fine-tuned stellar parameters or ad hoc mechanisms", and diffusive models need extra turbulence to keep the plateau uniform [13]. Sbordone and colleagues find it "difficult to envision a single depletion phenomenon" for both the thin plateau and the scattered stars below it [4], and the SMC gas needs fine-tuning in this picture [6].

Measurement systematics. The stellar abundances could be biased, for example by the temperature scale [5]. Strongest evidence: the ionisation correction is "exponentially sensitive" to stellar temperature [5]. Main difficulty: detailed studies of the temperature scale agree with the standard one [5].

Nuclear physics. A missed or poorly measured resonance could enhance the destruction of beryllium-7 [5]. Strongest evidence: in 2011 Fields described a proposal by Cyburt and Pospelov: a known resonance in ⁷Be + d was poorly constrained experimentally and could solve the problem "in an elegant manner" [5]. Main difficulty: measurements at CERN and searches for resonances found no fix [8, 9, 10], and the PDG review says experimental searches have now closed off these possibilities, "making a nuclear fix increasingly unlikely" [13]; the 2026 first-principles rates point the same way [15].

New physics during BBN. Decaying particles, for example from dark matter, or changing fundamental constants alter the nuclear reactions [13]. Strongest evidence: Andreas Goudelis, Maxim Pospelov and Josef Pradler showed that light neutral particles with masses of 1.6–20 MeV and lifetimes from a few hundred to 10⁴ seconds can solve the problem without changing deuterium or helium, and that such particles could be probed at intensity-frontier experiments [14]. Main difficulty: reducing lithium tends to raise deuterium, and precise deuterium data now force particle-injection models to perturb lithium "far short of the level needed", with the Goudelis model an exception [13].

Open sub-questions

  1. Lithium-6 limits. Do upper limits on lithium-6 in further plateau stars imply depletion as large as Fields and Olive infer, and do stellar models deplete lithium-7 as much as lithium-6 [11, 13]?
  2. The meltdown. What produces the scatter below a metallicity of about −3, and does the same process act on plateau stars at higher metallicity [4, 13]?
  3. Diffusion beyond one cluster. Does the turbulent-mixing efficiency fitted in NGC 6397 reproduce lithium along the evolutionary sequence in other metal-poor clusters and in field stars [2, 5]?
  4. Interstellar lithium. Do measurements of interstellar lithium in other metal-poor galaxies match the SMC value, and do they favour depletion or non-standard BBN [6, 13]?
  5. Light particles. Which intensity-frontier experiments could test the couplings that Goudelis and colleagues require, within the window allowed by current deuterium data [13, 14]?
  6. The full network. Does a first-principles reassessment of the whole BBN reaction network change beryllium-7 destruction by more than a few per cent [15]?

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. Spite, M. & Spite, F. (1982). Lithium abundance at the formation of the Galaxy. Nature 297(5866), 483–485. 10.1038/297483a0 · Checked 16 Oct 2026
  2. Korn, A. J., Grundahl, F., Richard, O., Barklem, P. S., Mashonkina, L., Collet, R., Piskunov, N. & Gustafsson, B. (2006). A probable stellar solution to the cosmological lithium discrepancy. Nature 442(7103), 657–659. 10.1038/nature05011 · Checked 16 Oct 2026
  3. Bemmerer, D., Confortola, F., Costantini, H., Formicola, A., Gyürky, G., Bonetti, R. et al. (LUNA Collaboration) (2006). Activation measurement of the ³He(α,γ)⁷Be cross section at low energy. Physical Review Letters 97(12), 122502. 10.1103/PhysRevLett.97.122502 · Checked 16 Oct 2026
  4. Sbordone, L., Bonifacio, P., Caffau, E., Ludwig, H.-G., Behara, N. T., González Hernández, J. I. et al. (2010). The metal-poor end of the Spite plateau. I. Stellar parameters, metallicities, and lithium abundances. Astronomy & Astrophysics 522, A26. 10.1051/0004-6361/200913282 · Checked 16 Oct 2026
  5. Fields, B. D. (2011). The primordial lithium problem. Annual Review of Nuclear and Particle Science 61, 47–68. 10.1146/annurev-nucl-102010-130445 · Checked 16 Oct 2026
  6. Howk, J. C., Lehner, N., Fields, B. D. & Mathews, G. J. (2012). Observation of interstellar lithium in the low-metallicity Small Magellanic Cloud. Nature 489(7414), 121–123. 10.1038/nature11407 · Checked 16 Oct 2026
  7. Fu, X., Bressan, A., Molaro, P. & Marigo, P. (2015). Lithium evolution in metal-poor stars: from pre-main sequence to the Spite plateau. Monthly Notices of the Royal Astronomical Society 452(3), 3256–3265. 10.1093/mnras/stv1384 · Checked 16 Oct 2026
  8. Barbagallo, M., Musumarra, A., Cosentino, L., Maugeri, E., Heinitz, S., Mengoni, A. et al. (n_TOF Collaboration) (2016). ⁷Be(n,α)⁴He reaction and the cosmological lithium problem: measurement of the cross section in a wide energy range at n_TOF at CERN. Physical Review Letters 117(15), 152701. 10.1103/PhysRevLett.117.152701 · Checked 16 Oct 2026
  9. Hammache, F., Coc, A., de Séréville, N., Stefan, I., Roussel, P., Ancelin, S. et al. (2013). Search for new resonant states in ¹⁰C and ¹¹C and their impact on the cosmological lithium problem. Physical Review C 88(6), 062802. 10.1103/PhysRevC.88.062802 · Checked 16 Oct 2026
  10. Damone, L., Barbagallo, M., Mastromarco, M., Mengoni, A., Cosentino, L., Maugeri, E. et al. (n_TOF Collaboration) (2018). ⁷Be(n,p)⁷Li reaction and the cosmological lithium problem: measurement of the cross section in a wide energy range at n_TOF at CERN. Physical Review Letters 121(4), 042701. 10.1103/PhysRevLett.121.042701 · Checked 16 Oct 2026
  11. Fields, B. D. & Olive, K. A. (2022). Implications of the non-observation of ⁶Li in halo stars for the primordial ⁷Li problem. Journal of Cosmology and Astroparticle Physics 2022(10), 078. 10.1088/1475-7516/2022/10/078 · Checked 16 Oct 2026
  12. Miranda, O. D. (2025). The cosmological lithium problem. Astronomy & Astrophysics 701, A164 (published online 9 September 2025). 10.1051/0004-6361/202554482 · Checked 16 Oct 2026
  13. Fields, B. D., Molaro, P. & Sarkar, S. (2025). 24. Big Bang nucleosynthesis, revised August 2025. In S. Navas et al. (Particle Data Group), Review of Particle Physics, Phys. Rev. D 110, 030001 (2024) and 2025 update. Source · Checked 16 Oct 2026
  14. Goudelis, A., Pospelov, M. & Pradler, J. (2016). Light particle solution to the cosmic lithium problem. Physical Review Letters 116(21), 211303. 10.1103/PhysRevLett.116.211303 · Checked 16 Oct 2026
  15. Taioli, S., Triggiani, F. & Simonucci, S. (2026). Revisiting ⁷Be weak and radiative transition rates in Big Bang nucleosynthesis: implications for the primordial lithium problem. Astronomy & Astrophysics 714, A152 (published online 9 October 2026). 10.1051/0004-6361/202659016 · Checked 16 Oct 2026

Page reviewed 16 Oct 2026

Source

Open, as listed by Wikipedia · Unsolved problems in physics. Checked 29 Sep 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 Cosmology & astronomy