Early Universe
The Primordial Lithium Problem: The Element Big Bang Theory Gets Wrong
Take the exquisitely tuned Big Bang, feed it the baryon density that the cosmic microwave background pins down to better than 1 percent, and it predicts that roughly one hydrogen atom in two billion should be lithium-7. Go measure that same ratio in the oldest, most pristine stars in the Milky Way's halo, and you find three times less lithium than the theory demands — a gap of four to five standard deviations that has refused to close for four decades.
The primordial lithium problem (also called the cosmological lithium problem) is this stubborn discrepancy between the abundance of 7Li predicted by standard Big Bang Nucleosynthesis (BBN) and the abundance actually observed in old, metal-poor halo stars. It is remarkable precisely because everything else about BBN works: deuterium and helium-4 match theory beautifully. Lithium is the one light element the standard picture appears to get wrong.
- TypeCosmological / nuclear astrophysics anomaly
- RegimeBig Bang Nucleosynthesis, t ~ 3-20 min after the Big Bang
- DiscoveredSpite plateau, Monique & François Spite (1982)
- DiscrepancyPredicted / observed Li-7 ~ 3x (4-5 sigma)
- Key ratioLi-7/H: BBN ~4.7e-10 vs observed ~1.6e-10
- Observed inMetal-poor halo dwarf stars ([Fe/H] < -1.5)
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
What the Problem Actually Is
Standard Big Bang Nucleosynthesis is a nearly parameter-free theory. Once you fix a single number — the baryon-to-photon ratio η = n_b/n_γ ≈ 6.1 × 10-10 — the physics of nuclear reactions in the expanding, cooling plasma during the first ~20 minutes determines the primordial yield of every light element. Crucially, η is no longer a free knob: the cosmic microwave background, measured by WMAP and then Planck, fixes the baryon density to Ω_b h² = 0.0224 ± 0.0001, which is equivalent to η.
With that pin in place, BBN predicts a lithium-7 abundance of Li/H ≈ (4.5–5.0) × 10-10. But when astronomers measure lithium in the atmospheres of the oldest halo stars — whose gas has been barely touched by later stellar processing — they find only Li/H ≈ (1.6 ± 0.3) × 10-10.
- The mismatch is a factor of roughly 2.4 to 4.3, a statistical tension of about 4–5σ.
- It is a one-sided problem: deuterium and helium-4 agree with BBN to within a few percent.
That selective failure is what makes lithium so provocative — the same theory that nails two nuclides overshoots the third.
The Mechanism: How BBN Makes Mass-7
To understand the problem you have to know that at the CMB-fixed baryon density, most primordial mass-7 is not made as lithium at all — it is made as beryllium-7. Two channels compete:
- Direct Li-7: 4He(3H,γ)7Li, followed by destruction via 7Li(p,α)4He. This dominates at low baryon density.
- Via Be-7: 3He(4He,γ)7Be. At the high baryon density the CMB actually favors, this channel wins.
Beryllium-7 is radioactive. Long after BBN freezes out, it captures an atomic electron with a half-life of 53.22 days, decaying to 7Li: 7Be + e- → 7Li + ν_e. So the lithium we should observe today is dominated by the radiogenic daughter of beryllium-7. The predicted abundance therefore hinges on the cross-section of 3He(4He,γ)7Be and the beryllium-7 destruction reactions — the very nuclear inputs that any 'nuclear physics' solution must revise.
Because primordial abundances depend only on η in standard BBN, the theory has essentially no wiggle room: raise η to fit lithium and you ruin deuterium.
The Numbers: Dex, Spite Plateau, and a Worked Comparison
Astronomers quote lithium as A(Li) = log₁₀(N_Li/N_H) + 12. The BBN+CMB prediction sits at A(Li) ≈ 2.7, while the observed Spite-plateau value is A(Li) ≈ 2.2.
- Prediction: A(Li) = 2.7 → N_Li/N_H = 10(2.7−12) = 10−9.3 ≈ 5.0 × 10−10.
- Observation: A(Li) = 2.2 → N_Li/N_H = 10−9.8 ≈ 1.6 × 10−10.
- Ratio: 10(2.7−2.2) = 100.5 ≈ 3.2 — the 'factor of three.'
The Spite plateau is the empirical anchor. François and Monique Spite showed in 1982 that lithium in warm, metal-poor halo dwarfs is remarkably flat: A(Li) ≈ 2.1–2.3 essentially independent of effective temperature (roughly 5700–6400 K) and of metallicity across [Fe/H] from about −1.5 down to −3. That flatness is the whole argument — it strongly suggests these stars display a fossil, near-primordial value rather than lithium made or destroyed by Galactic chemistry.
How It Is Observed
The measurement rests on a single spectral feature: the neutral lithium resonance doublet at 6707.8 Å (Li I 6708), seen in absorption in high-resolution stellar spectra from instruments like ESO's UVES on the VLT and HIRES on Keck. The line strength, combined with a model stellar atmosphere and an effective temperature, yields the lithium abundance.
The targets are carefully chosen metal-poor halo main-sequence and subgiant stars — old, low-mass stars formed from gas that had seen little chemical enrichment. Warm dwarfs are preferred because their thin surface convection zones should not have dragged lithium down into hot interiors where it burns (lithium is destroyed above ~2.5 × 106 K).
- Metallicity is read from iron lines as [Fe/H]; the most pristine stars reach [Fe/H] < −3.
- Below [Fe/H] ≈ −3 the plateau appears to break downward and scatter — the 'meltdown' — complicating the primordial extrapolation.
- Corroborating checks come from lithium in the interstellar medium, in the Small Magellanic Cloud, and even from 7Be detected in classical novae ejecta.
Distinguishing It From Its Cousins
The primordial lithium problem is easy to confuse with adjacent puzzles; keeping them separate matters.
- The 'second' (or Li-6) lithium problem: some early spectra hinted at surprisingly high 6Li in halo stars — a purely fragile isotope BBN makes in negligible amounts. Later work showed most of that signal was a spectral-line asymmetry artifact, so the Li-6 problem has largely evaporated. The Li-7 problem has not.
- The solar / stellar lithium depletion problem: the Sun's surface lithium is ~140 times lower than its meteoritic birth value. That is a stellar-structure puzzle about mixing and burning, not a cosmological one.
- Galactic chemical evolution: at high metallicity, lithium rises (from cosmic rays, novae, AGB stars). The Spite plateau lives below that, in the regime where such production is negligible.
In short, the cosmological lithium problem is specifically about the primordial value inferred from the flattest, most metal-poor stars versus the BBN+CMB prediction — distinct from how individual stars later process their lithium.
Proposed Solutions and Open Questions
Four decades on, no single explanation commands consensus. The candidate fixes fall into three families:
- Astrophysical / stellar depletion: the leading contender. If halo stars have gently destroyed part of their surface lithium via atomic diffusion plus turbulent mixing below the convection zone, the observed plateau would sit below the true primordial value. Models tuned to reproduce the plateau's flatness can erase much of the factor of three, but require fine-tuning to keep the plateau so tight.
- Nuclear physics: revising the cross-sections that build or destroy mass-7 (especially 7Be(n,p)7Li and 7Be(d,p)2α). Repeated laboratory campaigns have not found a large enough missing resonance to solve it.
- New physics: decaying dark-matter particles, variations in fundamental constants, or a sterile-neutrino that suppresses 7Be during BBN. These can work but are unconstrained and often disturb deuterium.
The honest status: most cosmologists lean toward stellar depletion, but the flatness of the Spite plateau across ~2 dex of metallicity remains hard to reproduce naturally. The lithium problem endures as one of the few genuine cracks in the otherwise triumphant standard cosmology.
| Nuclide | BBN prediction (Planck baryon density) | Observed (primordial) | Verdict |
|---|---|---|---|
| D / H | (2.5 +/- 0.03) x 10^-5 | (2.53 +/- 0.04) x 10^-5 | Excellent agreement |
| He-4 (Y_p) | 0.2470 +/- 0.0002 | 0.245 +/- 0.003 | Agreement |
| He-3 / H | ~1.0 x 10^-5 | ~1.1 x 10^-5 (local ISM) | Consistent (weak constraint) |
| Li-7 / H | (4.5 - 5.0) x 10^-10 | (1.6 +/- 0.3) x 10^-10 | Factor ~3 too high: the problem |
Frequently asked questions
What is the primordial lithium problem in simple terms?
It is the mismatch between how much lithium-7 the Big Bang should have made and how much we actually see in the universe's oldest stars. Big Bang nucleosynthesis, using the baryon density measured from the cosmic microwave background, predicts about three times more lithium-7 than the observed value. Unlike deuterium and helium, which match theory well, lithium is the one light element that stubbornly disagrees.
How big is the discrepancy, exactly?
BBN predicts a lithium-7 to hydrogen ratio of roughly (4.5-5.0) x 10^-10, while old metal-poor halo stars show about (1.6 +/- 0.3) x 10^-10. That is a factor of about 2.4 to 4.3, corresponding to a statistical tension of roughly 4 to 5 standard deviations. In the A(Li) notation astronomers use, prediction is ~2.7 and observation is ~2.2.
What is the Spite plateau?
Discovered by François and Monique Spite in 1982, the Spite plateau is the observation that lithium abundance in warm, metal-poor halo stars is nearly constant, A(Li) ~ 2.2, regardless of the star's temperature or metallicity. That flatness suggests these stars preserve a near-primordial lithium value, which is why the plateau is used as the observational benchmark for the primordial lithium problem.
Why is beryllium-7 important here?
At the high baryon density the CMB implies, most primordial mass-7 is actually synthesized as beryllium-7 via helium-3 plus helium-4, not directly as lithium-7. Beryllium-7 is radioactive and later decays by electron capture (half-life 53.22 days) into lithium-7. So the predicted lithium abundance mainly reflects beryllium-7 production, which is why nuclear-physics solutions target the reactions that build and destroy beryllium-7.
Could old stars have simply destroyed their lithium?
This is the most popular explanation. Lithium burns at temperatures above about 2.5 million kelvin, so if atomic diffusion and turbulent mixing slowly carried surface lithium into a star's hot interior over billions of years, the observed plateau would lie below the true primordial value. Depletion models can account for much of the factor of three, but reproducing the plateau's remarkable tightness without fine-tuning remains difficult.
Does the lithium problem threaten the Big Bang theory?
Not seriously. Big Bang nucleosynthesis correctly predicts deuterium and helium-4 across ten orders of magnitude of abundance, and the framework is one of cosmology's great successes. The lithium problem is a genuine unresolved anomaly, but most researchers expect it to be resolved by stellar physics (surface depletion) or refined nuclear cross-sections rather than by overturning the Big Bang model itself.