Chapter 24 — Big Bang Nucleosynthesis: The First Three Minutes

"In the beginning there was an explosion. Not an explosion like those familiar on Earth, starting from a definite center and spreading out to engulf more and more of the circumambient air, but an explosion which occurred simultaneously everywhere, filling all space from the beginning." — Steven Weinberg, The First Three Minutes (1977)

Chapter Overview

Every proton in every water molecule you have ever drunk was forged in the first second after the Big Bang. Every helium atom in every party balloon traces its ancestry to nuclear reactions that occurred when the universe was roughly three minutes old. And the traces of deuterium, helium-3, and lithium-7 that survived from that epoch constitute one of the most powerful tools in all of cosmology — a set of cosmic fossils whose abundances were fixed by nuclear physics alone.

Big Bang nucleosynthesis (BBN) is the process by which the lightest elements were synthesized in the hot, dense early universe. It is a triumph of twentieth-century physics: the predictions of BBN, based on well-measured nuclear cross sections and the known laws of particle physics, agree spectacularly with the observed primordial abundances of deuterium and helium-4. That agreement — spanning nine orders of magnitude in abundance — provides one of the three observational pillars of the Big Bang model, alongside the cosmic microwave background (CMB) and the Hubble expansion.

But BBN also harbors an unsolved puzzle. The predicted abundance of lithium-7 is roughly three times higher than what is observed in the oldest, most metal-poor stars in our galaxy. This lithium problem has resisted explanation for over forty years and remains one of the most tantalizing open questions at the intersection of nuclear physics, stellar astrophysics, and cosmology.

In this chapter, we will:

• Trace the thermal history of the early universe from $T \sim 10^{11}\,\text{K}$ to $T \sim 10^{8}\,\text{K}$, identifying the critical nuclear physics events at each stage.
• Derive the neutron-to-proton freeze-out ratio from the competition between the weak interaction rate and the cosmic expansion rate.
• Explain the deuterium bottleneck — why nucleosynthesis is delayed until the universe is about three minutes old despite the fact that deuterium formation is exothermic.
• Follow the reaction network from deuterium through helium-4 to lithium-7 and explain why the chain stops there.
• Compare theoretical predictions to observations and confront the lithium problem.
• Show how BBN constrains fundamental cosmological parameters, especially the baryon density $\Omega_b$.

This chapter connects the nuclear reaction physics of Chapters 17–21 and the stellar nucleosynthesis of Chapter 22 to cosmology itself. The nuclear cross sections measured in laboratories around the world determine the chemical composition of the universe set in its first few minutes of existence — a stunning connection between the physics of femtometers and the physics of the cosmos.

🏃 Fast Track: If you are familiar with the thermal history of the early universe, skim Sections 24.1–24.2 and begin at Section 24.3 (the deuterium bottleneck). The essential quantitative material is in Sections 24.3–24.6.

🔬 Deep Dive: The full derivation of the freeze-out temperature (Section 24.2) and the Saha-equation analysis of the deuterium bottleneck (Section 24.3) reward careful study — they illustrate how microphysics (cross sections, binding energies) determines macrophysics (cosmic abundances).

📊 Spaced Review (Chapter 21): Recall the Gamow peak — the narrow energy window where the rising Coulomb barrier penetration probability and the falling Maxwell-Boltzmann tail overlap to determine the effective energy for thermonuclear reactions. This same physics governs every reaction in the BBN network.

📊 Spaced Review (Chapter 22): In stellar nucleosynthesis, hydrogen burns to helium over billions of years inside stars. In the Big Bang, the same net process ($4p \to {}^4\text{He} + 2e^+ + 2\nu_e$) occurred in roughly 200 seconds. The difference is the environment: BBN operates at temperatures and densities far beyond any stellar interior, but in a universe that is expanding and cooling rapidly.

24.1 The Hot, Dense Early Universe

24.1.1 The Expanding Fireball

In the standard hot Big Bang model, the very early universe was extraordinarily hot and dense. As the universe expanded, it cooled. The relationship between temperature $T$ and time $t$ in the radiation-dominated era is:

$$T \approx \left(\frac{3c^2}{32\pi G a_R}\right)^{1/4} \frac{1}{\sqrt{t}} \approx \frac{1.5 \times 10^{10}\,\text{K}}{\sqrt{t/\text{s}}} \cdot g_*^{-1/4}$$

where $a_R = 4\sigma_B/c$ is the radiation constant and $g_*$ is the effective number of relativistic degrees of freedom. More precisely, the Friedmann equation gives the Hubble parameter as:

$$H(T) = \sqrt{\frac{8\pi G}{3}\rho_{\text{rad}}} = \sqrt{\frac{8\pi^3 g_*}{90}} \frac{T^2}{M_{\text{Pl}}}$$

where $M_{\text{Pl}} = \sqrt{\hbar c / G} \approx 1.22 \times 10^{19}\,\text{GeV}/c^2$ is the Planck mass and the radiation energy density is:

$$\rho_{\text{rad}} = \frac{\pi^2}{30} g_* T^4 \quad (\text{in natural units with } \hbar = c = k_B = 1)$$

The factor $g_*$ counts the relativistic species weighted by their spin statistics:

$$g_* = \sum_{\text{bosons}} g_i + \frac{7}{8}\sum_{\text{fermions}} g_i$$

At $T \sim 10^{10}\,\text{K}$ ($kT \sim 1\,\text{MeV}$), the relevant species are photons ($g = 2$), electrons and positrons ($g = 4$, with the $7/8$ factor), and three flavors of neutrinos plus antineutrinos ($g = 6$, with the $7/8$ factor). This gives $g_* = 2 + \frac{7}{8}(4 + 6) = 2 + 8.75 = 10.75$.

💡 Key Insight: At $T \sim 10^{10}\,\text{K}$, the universe is a hot soup of photons, electrons, positrons, neutrinos, and a tiny admixture of baryons (protons and neutrons). The baryon-to-photon ratio is $\eta = n_b/n_\gamma \approx 6 \times 10^{-10}$ — for every baryon, there are roughly 1.6 billion photons. This enormous photon excess has profound consequences for nucleosynthesis.

24.1.2 Timeline of the First Three Minutes

The key events in BBN occur in a well-defined sequence, governed by the competition between nuclear reaction rates and the expansion rate:

💡 Key Insight: The window for BBN is remarkably narrow. At earlier times, the universe is too hot (photodissociation destroys nuclei). At later times, the universe is too cool and dilute (Coulomb barriers cannot be overcome). The entire epoch of primordial nucleosynthesis lasts roughly from 3 to 20 minutes after the Big Bang.

24.1.3 The Cosmic Inventory at 1 Second

To appreciate the environment in which BBN takes place, consider the contents of a cubic centimeter of the universe at $t = 1\,\text{s}$ ($T \sim 10^{10}\,\text{K}$, $kT \sim 1\,\text{MeV}$):

• Photons: $n_\gamma \approx 2 \times 10^{31}\,\text{cm}^{-3}$. The universe is a blinding fireball.
• Electron-positron pairs: $n_{e^+} \approx n_{e^-} \approx n_\gamma$ (still approximately in equilibrium since $kT \sim 2m_ec^2$).
• Neutrinos: $n_\nu \approx \frac{3}{11} n_\gamma$ per flavor (they have just decoupled or are about to).
• Baryons: $n_b \approx \eta \cdot n_\gamma \approx 10^{22}\,\text{cm}^{-3}$. This is comparable to the density of air — remarkably dilute for nuclear reactions.

The mass-energy density is dominated overwhelmingly by radiation (photons, $e^\pm$, neutrinos), not by baryons. The baryon mass density is $\rho_b \sim 10^{-2}\,\text{g/cm}^3$ — dense by interstellar standards but negligible compared to the radiation energy density $\rho_{\text{rad}} \sim 10^{12}\,\text{g/cm}^3$ (in mass-energy equivalent).

💡 Key Insight: BBN is a nuclear process occurring in a radiation-dominated universe. The baryons — the only particles that participate in nuclear reactions — are a tiny minority. This fact, encoded in the small value of $\eta$, determines virtually everything about BBN: the freeze-out ratio, the deuterium bottleneck, and the final abundances.

24.1.4 The Baryon-to-Photon Ratio

The single most important cosmological parameter for BBN is the baryon-to-photon ratio:

$$\eta \equiv \frac{n_b}{n_\gamma}$$

where $n_b$ is the baryon number density and $n_\gamma$ is the photon number density. This ratio is conserved as the universe expands (both $n_b$ and $n_\gamma$ scale as $a^{-3}$ during the BBN epoch, where $a$ is the scale factor). Its value, determined independently by both BBN and the CMB, is:

$$\eta = (6.14 \pm 0.19) \times 10^{-10} \quad \text{(Planck 2018 + BBN)}$$

The baryon-to-photon ratio is related to the baryon density parameter by:

$$\Omega_b h^2 = 3.65 \times 10^7 \, \eta$$

where $h = H_0 / (100\,\text{km/s/Mpc})$ is the reduced Hubble constant. The Planck 2018 value $\Omega_b h^2 = 0.02242 \pm 0.00014$ corresponds to $\eta = (6.14 \pm 0.04) \times 10^{-10}$.

The physical significance of $\eta$ being so small is enormous: baryonic matter is a tiny contaminant in a universe dominated by radiation (at the BBN epoch) and later by dark matter and dark energy. But that tiny admixture of baryons is all the ordinary matter there is, and its nuclear processing in the first few minutes determines the raw material from which stars, planets, and people are eventually built.

24.2 Neutron-to-Proton Freeze-Out

24.2.1 Weak Equilibrium

At temperatures $T \gg 10^{10}\,\text{K}$ ($kT \gg 1\,\text{MeV}$), weak interactions maintain thermal equilibrium between neutrons and protons through the reactions:

$$n + e^+ \leftrightarrow p + \bar{\nu}_e \tag{24.1}$$

$$n + \nu_e \leftrightarrow p + e^- \tag{24.2}$$

$$n \leftrightarrow p + e^- + \bar{\nu}_e \tag{24.3}$$

In thermal equilibrium, the neutron-to-proton ratio is given by the Boltzmann factor:

$$\frac{n_n}{n_p} = \exp\left(-\frac{Q}{kT}\right) \tag{24.4}$$

where $Q = (m_n - m_p)c^2 = 1.293\,\text{MeV}$ is the neutron-proton mass difference. At high temperatures ($kT \gg Q$), the ratio approaches unity — roughly equal numbers of neutrons and protons. As the universe cools toward $kT \sim Q$, the equilibrium shifts toward protons because they are lighter.

⚠️ Important: The neutron-proton mass difference $Q = 1.293\,\text{MeV}$ is a critical number in BBN. If $Q$ were zero, the freeze-out ratio would be $n/p = 1$ and the universe would be almost entirely helium-4. If $Q$ were much larger, nearly all neutrons would have converted to protons before freeze-out, and there would be negligible helium. The actual value of $Q$ — determined by the up-down quark mass difference and electromagnetic contributions — is just right to produce a $\sim 25\%$ helium mass fraction.

24.2.2 The Freeze-Out Condition

Equilibrium is maintained as long as the weak interaction rate $\Gamma_w$ exceeds the Hubble expansion rate $H$. When the universe expands faster than the weak interactions can operate, the neutron-to-proton ratio freezes out — it ceases to track the equilibrium value and instead becomes locked in at its value at the freeze-out temperature.

The weak interaction rate for the process $n + \nu_e \to p + e^-$ scales as:

$$\Gamma_w \approx G_F^2 (kT)^5 / (\hbar c)^6 \tag{24.5}$$

where $G_F / (\hbar c)^3 = 1.166 \times 10^{-5}\,\text{GeV}^{-2}$ is the Fermi coupling constant. More precisely, summing over all three weak processes:

$$\Gamma_w \approx 1.63 \frac{G_F^2 (kT)^5}{\pi^3 \hbar} \left[1 + 3g_A^2\right]$$

where $g_A \approx 1.276$ is the axial-vector coupling constant of the nucleon. Numerically:

$$\Gamma_w \approx 0.40\,\text{s}^{-1} \left(\frac{T}{10^{10}\,\text{K}}\right)^5$$

The Hubble expansion rate in the radiation-dominated era scales as $T^2$:

$$H(T) = \sqrt{\frac{8\pi^3 g_*}{90}} \frac{(kT)^2}{M_{\text{Pl}} c^2} \approx 0.68\,\text{s}^{-1} \left(\frac{T}{10^{10}\,\text{K}}\right)^2 \tag{24.6}$$

The freeze-out condition $\Gamma_w = H$ gives:

$$T_f^3 \propto \frac{M_{\text{Pl}}}{G_F^2}$$

Solving:

$$kT_f \approx 0.8\,\text{MeV} \quad \Longleftrightarrow \quad T_f \approx 0.9 \times 10^{10}\,\text{K} \quad \Longleftrightarrow \quad t_f \approx 1\,\text{s}$$

🔄 Check Your Understanding: Verify the freeze-out temperature. Set $\Gamma_w = H$: $0.40 (T/10^{10})^5 = 0.68 (T/10^{10})^2$, giving $(T/10^{10})^3 = 1.7$, hence $T_f \approx 1.2 \times 10^{10}\,\text{K}$, or $kT_f \approx 1.0\,\text{MeV}$. (Exact calculations including all numerical factors give $kT_f \approx 0.7$–$0.8\,\text{MeV}$; the simple estimate captures the right order of magnitude.)

24.2.3 The Freeze-Out Ratio

At the freeze-out temperature $kT_f \approx 0.8\,\text{MeV}$, the equilibrium neutron-to-proton ratio is:

$$\frac{n}{p}\bigg|_{\text{freeze-out}} = \exp\left(-\frac{Q}{kT_f}\right) = \exp\left(-\frac{1.293}{0.8}\right) \approx e^{-1.62} \approx 0.20 \approx \frac{1}{5}$$

So at freeze-out, there is approximately one neutron for every five protons, giving $n/p \approx 1/5$.

💡 Key Insight: The freeze-out is not perfectly sharp — it occurs over roughly a factor of two in temperature. More detailed calculations, which integrate the rate equations through the freeze-out epoch rather than using an instantaneous approximation, give $n/p \approx 1/6$ at the end of the freeze-out process. This is the number we adopt.

24.2.4 Free Neutron Decay

After freeze-out, the neutron-to-proton ratio is no longer maintained by weak equilibrium, but it continues to decrease slowly due to free neutron beta decay:

$$n \to p + e^- + \bar{\nu}_e \qquad \tau_n = 879.4 \pm 0.6\,\text{s}$$

The ratio evolves as:

$$\frac{n}{p}(t) = \frac{(n/p)_f \, e^{-t/\tau_n}}{1 + (n/p)_f (1 - e^{-t/\tau_n})}$$

where $(n/p)_f \approx 1/6$ is the ratio at freeze-out and $\tau_n$ is the neutron mean life. By the time nucleosynthesis begins at $t \approx 180\,\text{s}$:

$$\frac{n}{p}(180\,\text{s}) \approx \frac{(1/6) \, e^{-180/879}}{1 + (1/6)(1 - e^{-180/879})} \approx \frac{(1/6)(0.815)}{1 + (1/6)(0.185)} \approx \frac{0.136}{1.031} \approx 0.132 \approx \frac{1}{7.6}$$

This is conventionally rounded to $n/p \approx 1/7$ at the onset of nucleosynthesis.

⚠️ Important: The neutron lifetime $\tau_n$ directly affects the primordial helium abundance. A longer-lived neutron means more neutrons survive to be incorporated into helium-4, producing a higher $Y_p$. The current $\sim 1\,\text{s}$ uncertainty in $\tau_n$ translates to an uncertainty of $\sim 0.1\%$ in $Y_p$ — small but not negligible at the current level of observational precision.

24.2.5 A Numerical Example: Tracing the Freeze-Out

Let us follow one neutron through the freeze-out process to build physical intuition.

At $t = 0.1\,\text{s}$ ($T = 4.7 \times 10^{10}\,\text{K}$, $kT = 4.1\,\text{MeV}$): The weak interaction rate is $\Gamma_w \sim 0.40 \times (47)^5 \approx 9.2 \times 10^7\,\text{s}^{-1}$. The Hubble rate is $H \sim 0.68 \times (47)^2 \approx 1.5 \times 10^3\,\text{s}^{-1}$. The ratio $\Gamma_w/H \sim 60{,}000$ — weak equilibrium is maintained easily. The equilibrium $n/p = \exp(-1.293/4.1) = 0.73$. Neutrons and protons interconvert freely.

At $t = 0.5\,\text{s}$ ($T = 1.9 \times 10^{10}\,\text{K}$, $kT = 1.6\,\text{MeV}$): $\Gamma_w/H \sim 0.40 \times (19)^5 / [0.68 \times (19)^2] \approx 97$. Still in equilibrium. The equilibrium $n/p = \exp(-1.293/1.6) = 0.45$. More protons than neutrons now.

At $t = 1\,\text{s}$ ($T = 1.3 \times 10^{10}\,\text{K}$, $kT = 1.1\,\text{MeV}$): $\Gamma_w/H \sim 7$. The reactions are beginning to fall behind. The actual $n/p$ is starting to deviate from the equilibrium value of $\exp(-1.293/1.1) = 0.31$.

At $t = 3\,\text{s}$ ($T = 7.7 \times 10^{9}\,\text{K}$, $kT = 0.66\,\text{MeV}$): $\Gamma_w/H \sim 0.3$. Freeze-out is essentially complete. The $n/p$ ratio is locked in at approximately $1/6$. The equilibrium value would be $\exp(-1.293/0.66) = 0.14 = 1/7.1$, but the ratio can no longer track this.

At $t = 180\,\text{s}$ ($T = 8 \times 10^{8}\,\text{K}$, $kT = 0.07\,\text{MeV}$): Free neutron decay has reduced $n/p$ from $\sim 1/6$ to $\sim 1/7$. The deuterium bottleneck breaks and nucleosynthesis begins. Our neutron — if it has survived — is about to be captured into a deuterium nucleus and then rapidly processed into helium-4.

This numerical trace makes the freeze-out process concrete. The transition from $\Gamma_w/H \gg 1$ (equilibrium) to $\Gamma_w/H \ll 1$ (frozen) occurs over about one decade in temperature — it is not instantaneous, but it is reasonably sharp.

24.2.6 Sensitivity to Fundamental Parameters

The freeze-out ratio, and hence the primordial helium abundance, depends sensitively on several fundamental quantities:

1. The neutron-proton mass difference $Q$. A larger $Q$ means a smaller $n/p$ at freeze-out (Eq. 24.4), hence less helium.

2. The Fermi constant $G_F$. A larger $G_F$ means faster weak reactions (Eq. 24.5), so freeze-out occurs later (at lower $T$), giving a smaller $n/p$ and less helium.

3. The number of neutrino species $N_\nu$. More neutrino species increase $g_*$, which increases the expansion rate (Eq. 24.6), causing earlier freeze-out (at higher $T$), a larger $n/p$, and more helium. This is why BBN constrains $N_\nu$.

4. The neutron lifetime $\tau_n$. A shorter $\tau_n$ means faster weak reactions (the same coupling governs both), delaying freeze-out; but it also means faster post-freeze-out decay, which reduces $n/p$ before nucleosynthesis begins. The net effect is complicated but quantifiable.

The sensitivity to $N_\nu$ is particularly important. Detailed calculations show:

$$\Delta Y_p \approx 0.013 \, \Delta N_\nu$$

BBN observations constrain $N_\nu = 2.94 \pm 0.38$ (at 95% confidence), beautifully consistent with three neutrino families and providing independent confirmation of the LEP measurement from $Z$-boson decays at CERN.

24.3 The Deuterium Bottleneck

24.3.1 The Paradox

Deuterium formation is exothermic:

$$p + n \to d + \gamma \qquad Q = B_d = 2.224\,\text{MeV} \tag{24.7}$$

At $kT \sim 1\,\text{MeV}$ (the time of weak freeze-out), the temperature is already below the deuterium binding energy. Naively, one might expect deuterium to form immediately and for nucleosynthesis to proceed at $t \sim 1\,\text{s}$.

But nucleosynthesis does not begin at $t \sim 1\,\text{s}$. It is delayed until $t \approx 180\,\text{s}$, when the temperature has dropped to $T \approx 8 \times 10^8\,\text{K}$ ($kT \approx 0.07\,\text{MeV}$). Why?

The answer lies in the enormous photon-to-baryon ratio. For every baryon, there are $\eta^{-1} \approx 1.6 \times 10^9$ photons. Even though the average photon energy ($\langle E_\gamma \rangle \approx 2.7 kT$) drops below the deuterium binding energy early on, the tail of the Planck distribution contains a sufficient number of high-energy photons to photodissociate deuterium:

$$\gamma + d \to p + n$$

The deuterium can only survive when the photodissociation rate drops below the formation rate — and given the enormous number of photons per baryon, this requires the temperature to drop far below $B_d/k$.

24.3.2 Quantitative Analysis: The Nuclear Saha Equation

The equilibrium abundance of deuterium in a bath of photons, protons, and neutrons is given by the nuclear Saha equation. For the reaction $p + n \leftrightarrow d + \gamma$, the equilibrium condition gives:

$$\frac{n_d}{n_p n_n} = \frac{3}{4} \left(\frac{2\pi \hbar^2}{m_N kT}\right)^{3/2} \exp\left(\frac{B_d}{kT}\right) \tag{24.8}$$

where the factor $3/4$ comes from the spin statistical weights ($g_d = 3$, $g_p = g_n = 2$, so $g_d/(g_p g_n) = 3/4$), $m_N$ is the nucleon mass, and $B_d = 2.224\,\text{MeV}$ is the deuterium binding energy.

We can rewrite this in terms of the baryon-to-photon ratio. Using $n_b \approx n_p + n_n$ (neglecting the tiny deuterium fraction initially) and $n_\gamma = 2\zeta(3)T^3/\pi^2$ (in natural units):

$$\frac{n_d}{n_b^2} = \frac{3}{4} \left(\frac{2\pi \hbar^2}{m_N kT}\right)^{3/2} \exp\left(\frac{B_d}{kT}\right) \tag{24.9}$$

Multiplying both sides by $n_b = \eta n_\gamma$:

$$\frac{n_d}{n_b} = \frac{3}{4} \eta \, n_\gamma \left(\frac{2\pi \hbar^2}{m_N kT}\right)^{3/2} \exp\left(\frac{B_d}{kT}\right) \tag{24.10}$$

Substituting numerical values:

$$\frac{n_d}{n_b} \approx \eta \cdot \frac{T^{3/2}}{T_{\text{nuc}}^{3/2}} \exp\left(\frac{B_d}{kT}\right)$$

where $T_{\text{nuc}} \sim m_N c^2 / k \sim 10^{13}\,\text{K}$.

The critical temperature for deuterium survival is defined as the temperature at which $n_d/n_b \sim 1$ — that is, when a significant fraction of baryons are in the form of deuterium. Setting $n_d/n_b = 1$ and solving for $T$:

$$\frac{B_d}{kT_D} \approx \ln\left(\frac{1}{\eta}\right) + \frac{3}{2}\ln\left(\frac{m_N c^2}{kT_D}\right) + \text{const.}$$

The dominant term is $\ln(1/\eta) \approx \ln(1.6 \times 10^9) \approx 21.2$. This gives:

$$kT_D \approx \frac{B_d}{21 + \frac{3}{2}\ln(m_N c^2/kT_D)} \approx \frac{2.224\,\text{MeV}}{21 + 13} \approx 0.065\,\text{MeV}$$

corresponding to $T_D \approx 7.5 \times 10^8\,\text{K}$ and $t_D \approx 200\,\text{s} \approx 3.3\,\text{min}$.

💡 Key Insight: The deuterium bottleneck temperature $kT_D \approx 0.07\,\text{MeV}$ is roughly 30 times smaller than the deuterium binding energy $B_d = 2.224\,\text{MeV}$. This huge factor arises from $\ln(1/\eta) \approx 21$ — the logarithm of the photon-to-baryon ratio. It is the enormous number of photons per baryon that keeps "photodissociating" deuterium long after the average photon energy has dropped well below $B_d$. Only the high-energy tail of the Planck distribution matters, and even a tiny fraction of $10^9$ photons is a lot of photons.

🔄 Check Your Understanding: Consider a hypothetical universe with $\eta = 10^{-5}$ (much more baryonic). Then $\ln(1/\eta) \approx 11.5$, and $kT_D \approx 2.224/(11.5 + 10) \approx 0.10\,\text{MeV}$. Deuterium would survive at a higher temperature, nucleosynthesis would begin earlier, and more neutrons would survive (less time for free neutron decay). This universe would produce more helium-4 and significantly more deuterium.

24.3.3 Dependence on $\eta$

The deuterium bottleneck temperature depends logarithmically on $\eta$:

$$kT_D \propto \frac{B_d}{\ln(1/\eta)}$$

• Higher $\eta$ (more baryons per photon): photodissociation is less effective, deuterium survives earlier, nucleosynthesis starts sooner, more neutrons are captured into helium-4, and less deuterium survives (because it is more efficiently burned to helium). This is why the final deuterium abundance is a steeply decreasing function of $\eta$.

• Lower $\eta$: the bottleneck lasts longer, more neutrons decay, and more deuterium survives (it is less efficiently burned). The deuterium abundance is a sensitive "baryometer."

24.4 The BBN Reaction Network

24.4.1 The Key Reactions

Once deuterium survives, the nuclear reaction network runs rapidly. The main reactions are:

Deuterium production: $$p + n \to d + \gamma \qquad Q = 2.224\,\text{MeV} \tag{R1}$$

Deuterium burning (two-body reactions that consume deuterium): $$d + d \to {}^3\text{He} + n \qquad Q = 3.269\,\text{MeV} \tag{R2}$$

$$d + d \to {}^3\text{H} + p \qquad Q = 4.033\,\text{MeV} \tag{R3}$$

$$d + p \to {}^3\text{He} + \gamma \qquad Q = 5.493\,\text{MeV} \tag{R4}$$

Helium-3 and tritium reactions (building helium-4): $${}^3\text{He} + d \to {}^4\text{He} + p \qquad Q = 18.353\,\text{MeV} \tag{R5}$$

$${}^3\text{H} + d \to {}^4\text{He} + n \qquad Q = 17.589\,\text{MeV} \tag{R6}$$

$${}^3\text{He} + n \to {}^3\text{H} + p \qquad Q = 0.764\,\text{MeV} \tag{R7}$$

$${}^3\text{H} + {}^4\text{He} \to {}^7\text{Li} + \gamma \qquad Q = 2.467\,\text{MeV} \tag{R8}$$

Lithium and beryllium production: $${}^3\text{He} + {}^4\text{He} \to {}^7\text{Be} + \gamma \qquad Q = 1.587\,\text{MeV} \tag{R9}$$

$${}^7\text{Be} + n \to {}^7\text{Li} + p \qquad Q = 1.644\,\text{MeV} \tag{R10}$$

$${}^7\text{Li} + p \to {}^4\text{He} + {}^4\text{He} \qquad Q = 17.347\,\text{MeV} \tag{R11}$$

$${}^7\text{Be} + e^- \to {}^7\text{Li} + \nu_e \qquad Q = 0.862\,\text{MeV} \tag{R12}$$

💡 Key Insight: The network has a clear hierarchical structure. Deuterium is the gateway — nothing heavier can be made until deuterium survives. Once it does, the $d + d$ reactions quickly produce ${}^3\text{He}$ and ${}^3\text{H}$, which then react with deuterium to produce ${}^4\text{He}$. The flow from light to heavy is rapid because the intermediate species ($d$, ${}^3\text{He}$, ${}^3\text{H}$) are all weakly bound and react quickly. But the flow stops at ${}^4\text{He}$ and ${}^7\text{Li}/{}^7\text{Be}$.

24.4.2 The Mass Gaps at A = 5 and A = 8

The most important feature of the BBN reaction network is what does not happen. There is no significant production of elements heavier than lithium-7 / beryllium-7. The reason is profoundly simple: there are no stable nuclei with $A = 5$ or $A = 8$.

• ${}^5\text{He}$ and ${}^5\text{Li}$ are unbound (they decay by nucleon emission in $\sim 10^{-22}\,\text{s}$).
• ${}^8\text{Be}$ is unbound by only 92 keV, decaying into two alpha particles in $\sim 10^{-16}\,\text{s}$.

This means: - ${}^4\text{He} + n \to {}^5\text{He}$ — unstable, immediately re-emits the neutron. - ${}^4\text{He} + p \to {}^5\text{Li}$ — unstable, immediately re-emits the proton. - ${}^4\text{He} + {}^4\text{He} \to {}^8\text{Be}$ — unstable, immediately breaks apart.

In stellar interiors, the $A = 8$ gap is bridged by the triple-alpha process ($3\alpha \to {}^{12}\text{C}$) via the Hoyle resonance in ${}^{12}\text{C}$ at 7.65 MeV (Chapter 22). But the triple-alpha process requires three particles to be in nearly the same place at nearly the same time. Its rate scales as $\rho^2$ (density squared), and while stellar cores are dense enough ($\rho \sim 10^5\,\text{g/cm}^3$), the BBN environment is not — the baryon density at $t \sim 3\,\text{min}$ is only $\rho_b \sim 10^{-5}\,\text{g/cm}^3$.

⚠️ Important: The absence of stable $A = 5$ and $A = 8$ nuclei is a contingent fact of nuclear physics — it follows from the specific properties of the nuclear force. But its cosmological consequences are immense. If there were a stable $A = 5$ nucleus, BBN would have burned through the mass gap and produced significant amounts of carbon and oxygen. The primordial universe would have been chemically rich rather than chemically simple, and the entire history of stellar nucleosynthesis — and the chemical evolution of the universe — would have been radically different.

24.4.3 The Timescale Hierarchy

A crucial feature of the BBN reaction network is the hierarchy of timescales. Once deuterium survives, the subsequent reactions are fast — much faster than the expansion rate. To see this quantitatively, consider the deuterium burning timescale at $T_9 = 0.8$ (just after the bottleneck breaks):

The density of baryons is $n_b \sim 10^{18}\,\text{cm}^{-3}$. The $d + d$ cross section at $kT \sim 0.07\,\text{MeV}$ gives $\langle\sigma v\rangle \sim 10^{-17}\,\text{cm}^3/\text{s}$. The deuterium burning timescale is:

$$\tau_{dd} \sim \frac{1}{n_b \langle\sigma v\rangle} \sim \frac{1}{10^{18} \times 10^{-17}} \sim 0.1\,\text{s}$$

Compare this to the expansion timescale:

$$\tau_H = \frac{1}{H} \sim 300\,\text{s}$$

So $\tau_{dd}/\tau_H \sim 3 \times 10^{-4}$ — the nuclear reactions are nearly $10^4$ times faster than the expansion. This means the nuclear network reaches a quasi-equilibrium very quickly once the bottleneck breaks. The reactions $d + d \to {}^3\text{He} + n$ and $d + d \to {}^3\text{H} + p$ consume deuterium almost as fast as it forms, and the subsequent reactions ${}^3\text{He} + d \to {}^4\text{He} + p$ and ${}^3\text{H} + d \to {}^4\text{He} + n$ convert the $A = 3$ species to helium-4 on a similar timescale.

The entire process — from deuterium survival to completion of helium-4 synthesis — takes only about 100–200 seconds. This is the nuclear "flash" that produces the light elements.

24.4.4 The Flow to Helium-4

Once the deuterium bottleneck breaks, the nuclear reactions proceed so rapidly that essentially all available neutrons end up in ${}^4\text{He}$. The logic is straightforward:

1. Every neutron that has survived to $t \approx 180\,\text{s}$ is captured into deuterium.
2. Deuterium is rapidly processed through the network to ${}^4\text{He}$.
3. The mass gaps at $A = 5$ and $A = 8$ prevent further processing.
4. The residual abundances of D, ${}^3\text{He}$, ${}^3\text{H}$, ${}^7\text{Li}$, and ${}^7\text{Be}$ are tiny — they represent the small fraction of material that did not make it all the way to ${}^4\text{He}$.

The resulting helium-4 mass fraction can be calculated from simple neutron counting.

24.5 Predicted Primordial Abundances

24.5.1 Helium-4: The Neutron Counter

If all neutrons end up in ${}^4\text{He}$, and each ${}^4\text{He}$ nucleus contains 2 neutrons and 2 protons, then:

Number of ${}^4\text{He}$ nuclei per baryon: $n_n / 2$ neutrons are paired with $n_n / 2$ protons to form $n_n / 2$ helium-4 nuclei. Each helium-4 nucleus has mass 4 (in nucleon mass units).

Helium-4 mass fraction:

$$Y_p \equiv \frac{4 \cdot (n_n/2)}{n_p + n_n} = \frac{2 n_n}{n_p + n_n} = \frac{2(n/p)}{1 + (n/p)} \tag{24.11}$$

With $n/p \approx 1/7$:

$$\boxed{Y_p = \frac{2 \times (1/7)}{1 + (1/7)} = \frac{2/7}{8/7} = \frac{2}{8} = 0.25}$$

Detailed numerical calculations, which track the full reaction network and include small corrections (not all neutrons end up in ${}^4\text{He}$; some remain in D, ${}^3\text{He}$, and ${}^7\text{Li}$), give:

$$\boxed{Y_p = 0.2470 \pm 0.0002 \quad \text{(standard BBN prediction, } \eta = 6.14 \times 10^{-10}\text{)}}$$

The uncertainty is dominated by the neutron lifetime. The simple estimate $Y_p = 0.25$ is remarkably close — this is because essentially all neutrons do end up in helium-4.

24.5.2 Dependence of $Y_p$ on Parameters

The helium mass fraction depends on $\eta$ only weakly (logarithmically, through the deuterium bottleneck temperature):

$$\frac{\partial Y_p}{\partial \ln \eta} \approx +0.012$$

Higher $\eta$ means the deuterium bottleneck breaks earlier, fewer neutrons have decayed, and $Y_p$ is slightly higher. But the dependence is weak — $Y_p$ changes by only about 1% as $\eta$ varies by a factor of 10.

$Y_p$ is much more sensitive to: - $N_\nu$: $\Delta Y_p / \Delta N_\nu \approx 0.013$ (more neutrino species $\to$ faster expansion $\to$ earlier freeze-out $\to$ more neutrons $\to$ more helium). - $\tau_n$: $\partial Y_p / \partial \tau_n \approx 2 \times 10^{-4}\,\text{s}^{-1}$ (longer neutron lifetime $\to$ more neutrons survive to nucleosynthesis $\to$ more helium).

24.5.3 Deuterium: The Baryometer

Unlike helium-4, the primordial deuterium abundance depends strongly on $\eta$:

$$\frac{\text{D}}{\text{H}} \propto \eta^{-1.6}$$

This steep dependence makes deuterium the most sensitive BBN "baryometer." Higher $\eta$ means higher baryon density, which means deuterium is burned more efficiently to helium (through reactions R2–R6), leaving less residual D.

The standard BBN prediction at $\eta = 6.14 \times 10^{-10}$ is:

$$\boxed{\frac{\text{D}}{\text{H}} = (2.52 \pm 0.03) \times 10^{-5}}$$

Deuterium is also a "one-way" tracer: it is destroyed but never produced in significant quantities in any astrophysical process after BBN (stars destroy D, never create it). So any measurement of D/H in a primitive environment provides a lower limit on the primordial value. In practice, measurements in high-redshift, low-metallicity gas clouds (damped Lyman-$\alpha$ systems) observed in quasar absorption spectra give:

$$\frac{\text{D}}{\text{H}}\bigg|_{\text{obs}} = (2.55 \pm 0.03) \times 10^{-5}$$

This spectacular agreement between prediction and observation is one of the great achievements of modern cosmology.

24.5.4 Helium-3

Helium-3 is produced alongside deuterium and has a predicted primordial abundance:

$$\frac{{}^3\text{He}}{\text{H}} \approx 1.0 \times 10^{-5}$$

However, ${}^3\text{He}$ is both produced and destroyed in stars, making it difficult to use as a cosmological probe. Its chemical evolution is complex, and the primordial value is poorly constrained by observations.

24.5.5 Lithium-7: The Problem Child

The BBN prediction for lithium-7 comes from two main production channels:

1. Direct production: ${}^3\text{H} + {}^4\text{He} \to {}^7\text{Li} + \gamma$ (reaction R8)
2. Production via beryllium-7: ${}^3\text{He} + {}^4\text{He} \to {}^7\text{Be} + \gamma$ (reaction R9), followed by ${}^7\text{Be} + e^- \to {}^7\text{Li} + \nu_e$ (reaction R12) after BBN, or ${}^7\text{Be} + n \to {}^7\text{Li} + p$ (reaction R10) during BBN.

At the Planck value of $\eta$, the beryllium-7 channel dominates. The standard prediction is:

$$\frac{{}^7\text{Li}}{\text{H}}\bigg|_{\text{BBN}} = (5.24 \pm 0.71) \times 10^{-10}$$

The observation, from the surfaces of old, metal-poor halo stars (the "Spite plateau," named after Monique and Francois Spite who first identified it in 1982), is:

$$\frac{{}^7\text{Li}}{\text{H}}\bigg|_{\text{obs}} \approx (1.6 \pm 0.3) \times 10^{-10}$$

The discrepancy is a factor of $\sim 3$–$4$:

$$\frac{({}^7\text{Li}/\text{H})_{\text{BBN}}}{({}^7\text{Li}/\text{H})_{\text{obs}}} \approx 3.3$$

This is the lithium problem — the most significant discrepancy in BBN, persisting for over four decades. We discuss it in detail in Case Study 1.

24.5.6 Summary of BBN Predictions vs. Observations

The agreement for D and ${}^4\text{He}$ spans the range from $Y_p \approx 0.25$ (a quarter of all baryonic mass!) down to D/H $\approx 2.5 \times 10^{-5}$ — nearly four orders of magnitude — using a single parameter $\eta$ determined independently by the CMB.

24.6 BBN as a Cosmological Probe

24.6.1 Constraining the Baryon Density

The primary cosmological output of BBN is a determination of the baryon-to-photon ratio $\eta$ (equivalently, the baryon density parameter $\Omega_b h^2$). Since each light-element abundance depends differently on $\eta$, combining them yields a precise determination.

The concordance is best seen graphically. In the standard BBN plot (known as the Schramm plot, after David Schramm who pioneered this analysis), the predicted abundances are plotted as curves versus $\eta$, and the observed values are shown as horizontal bands. The vertical band where all observations overlap gives the concordance value of $\eta$.

The key features of the Schramm plot are:

1. $Y_p$ (helium-4): Nearly flat — depends weakly on $\eta$ (logarithmically). The horizontal band is wide because the observational uncertainty on $Y_p$ is relatively large ($\pm 0.004$).

2. D/H (deuterium): Steeply declining ($\propto \eta^{-1.6}$). This is the most powerful baryometer because of its steep $\eta$-dependence and precise observational determination.

3. ${}^3\text{He}/\text{H}$: Mildly declining with $\eta$. Less useful due to complex chemical evolution.

4. ${}^7\text{Li}/\text{H}$: Has a characteristic "valley" — decreasing at low $\eta$ (where ${}^7\text{Li}$ is made directly), reaching a minimum near $\eta \sim 3 \times 10^{-10}$, then increasing at high $\eta$ (where the ${}^7\text{Be}$ channel dominates).

The concordance value from BBN (primarily from deuterium) is:

$$\eta_{\text{BBN}} = (6.14 \pm 0.19) \times 10^{-10} \quad \Longleftrightarrow \quad \Omega_b h^2 = 0.0224 \pm 0.0007$$

24.6.2 Agreement with the CMB

The Planck satellite measured the baryon density from the acoustic oscillation pattern in the CMB power spectrum:

$$\Omega_b h^2\big|_{\text{CMB}} = 0.02242 \pm 0.00014$$

The agreement between the BBN and CMB determinations of $\Omega_b h^2$ is remarkable:

$$\frac{\Omega_b h^2|_{\text{BBN}} - \Omega_b h^2|_{\text{CMB}}}{\Omega_b h^2|_{\text{CMB}}} < 1\%$$

These two measurements probe completely different physics at completely different epochs: - BBN measures nuclear reactions at $t \sim 3\,\text{min}$ ($T \sim 10^9\,\text{K}$, $z \sim 10^9$). - The CMB measures photon-baryon acoustic oscillations at $t \sim 380{,}000\,\text{yr}$ ($T \sim 3000\,\text{K}$, $z \sim 1100$).

Their agreement to $\sim 1\%$ is powerful evidence that the standard cosmological model — the hot Big Bang with known particle physics — is fundamentally correct over a factor of $\sim 10^5$ in time and $\sim 10^6$ in temperature.

24.6.3 Constraining New Physics

The precision of BBN makes it a sensitive probe of physics beyond the Standard Model. Proposed new particles or interactions that would alter the expansion rate or the weak interaction rates during the BBN epoch are tightly constrained. Specific examples include:

Additional neutrino species. The effective number of neutrino species is conventionally parameterized as $N_{\text{eff}}$. In the Standard Model, $N_{\text{eff}} = 3.044$ (slightly above 3 due to non-instantaneous neutrino decoupling and $e^+e^-$ heating). BBN constrains:

$$N_{\text{eff}} = 2.94 \pm 0.38 \quad \text{(BBN alone, 95\% CL)}$$

$$N_{\text{eff}} = 2.99 \pm 0.17 \quad \text{(BBN + CMB, 95\% CL)}$$

Any new light particle that was in thermal equilibrium at $T \sim 1\,\text{MeV}$ would contribute to $N_{\text{eff}}$ and is constrained by these limits.

Variation of fundamental constants. If the fine-structure constant $\alpha$, the strong coupling constant, or the electron-to-proton mass ratio were different in the early universe, BBN abundances would be altered. Current BBN constraints limit the variation of $\alpha$ to $|\Delta\alpha/\alpha| < 0.02$ at the BBN epoch.

Dark matter annihilation. Late-decaying or annihilating dark matter particles could inject energetic particles into the BBN plasma, altering the light-element abundances. BBN constrains the electromagnetic energy injection rate to $< 10^{-10}\,\text{GeV/s}$ per baryon during the nucleosynthesis epoch.

24.7 The BBN Reaction Network: Coupled ODEs

24.7.1 Structure of the Rate Equations

The evolution of nuclear abundances during BBN is governed by a system of coupled ordinary differential equations. For a species $i$ with mass fraction $Y_i = n_i / n_b$ (where $n_b$ is the total baryon number density), the rate equation is:

$$\frac{dY_i}{dt} = \sum_j \Gamma_j^{(1)} Y_j + \sum_{j,k} n_b \langle\sigma v\rangle_{jk} Y_j Y_k + \sum_{j,k,l} n_b^2 \langle\sigma v\rangle_{jkl} Y_j Y_k Y_l + \ldots \tag{24.12}$$

where: - $\Gamma_j^{(1)}$ represents single-particle decay rates (e.g., neutron beta decay, tritium beta decay). - $\langle\sigma v\rangle_{jk}$ is the thermally-averaged cross section times relative velocity for two-body reactions $j + k \to$ products (Chapter 21). - $\langle\sigma v\rangle_{jkl}$ is the three-body reaction rate (relevant only for a few reactions, such as the triple-alpha process, which is negligible at BBN densities).

The sum includes terms with appropriate stoichiometric coefficients and signs: positive for production of species $i$, negative for destruction.

24.7.2 The Minimal Network

A minimal but physically complete BBN network tracks eight species: $n$, $p$, $d$, ${}^3\text{H}$, ${}^3\text{He}$, ${}^4\text{He}$, ${}^7\text{Li}$, and ${}^7\text{Be}$. The 12 reactions listed in Section 24.4.1, plus neutron decay (R0: $n \to p + e^- + \bar{\nu}_e$), form the minimal set.

The system of 8 coupled ODEs has the form:

$$\frac{dY_n}{dt} = -\frac{Y_n}{\tau_n} - n_b\langle\sigma v\rangle_{np \to d\gamma} Y_n Y_p + n_b\langle\sigma v\rangle_{dd \to {}^3\text{He}\,n} Y_d^2 + \ldots$$

$$\frac{dY_p}{dt} = +\frac{Y_n}{\tau_n} - n_b\langle\sigma v\rangle_{np \to d\gamma} Y_n Y_p + n_b\langle\sigma v\rangle_{dd \to {}^3\text{H}\,p} Y_d^2 + \ldots$$

$$\frac{dY_d}{dt} = +n_b\langle\sigma v\rangle_{np \to d\gamma} Y_n Y_p - 2n_b\langle\sigma v\rangle_{dd} Y_d^2 - n_b\langle\sigma v\rangle_{dp} Y_d Y_p - \ldots$$

and so on for each species. The full system is stiff (reaction rates span many orders of magnitude) and requires an implicit ODE solver (e.g., BDF or Rosenbrock methods).

24.7.3 Thermonuclear Reaction Rates

The thermally-averaged reaction rates $\langle\sigma v\rangle$ for BBN reactions are computed from measured cross sections, typically parameterized in terms of the astrophysical $S$-factor (Chapter 21):

$$\sigma(E) = \frac{S(E)}{E} \exp\left(-2\pi\eta_s\right)$$

where $\eta_s = Z_1 Z_2 e^2 / (\hbar v)$ is the Sommerfeld parameter (not to be confused with the baryon-to-photon ratio $\eta$). The thermal average is:

$$\langle\sigma v\rangle = \left(\frac{8}{\pi \mu}\right)^{1/2} \frac{1}{(kT)^{3/2}} \int_0^\infty S(E) \exp\left(-\frac{E}{kT} - \frac{E_G^{1/2}}{E^{1/2}}\right) dE \tag{24.13}$$

where $E_G = 2\mu (Z_1 Z_2 e^2 \pi / \hbar)^2$ is the Gamow energy.

For BBN calculations, the reaction rates are typically taken from compilations such as: - NACRE II (Nuclear Astrophysics Compilation of REaction rates) - Smith, Kawano, & Malaney (1993) — the classic BBN rate compilation - Coc et al. (2015) — updated rates with modern nuclear data - Particle Data Group reviews

Each rate is parameterized as a function of temperature, often in the REACLIB format:

$$N_A\langle\sigma v\rangle = \exp\left(a_0 + \sum_{i=1}^{5} a_i T_9^{(2i-5)/3} + a_6 \ln T_9\right)$$

where $T_9 = T / 10^9\,\text{K}$ and $N_A$ is Avogadro's number.

24.7.4 Time-Temperature Relation

The integration variable in the BBN code can be either time $t$ or temperature $T$. They are related by:

$$\frac{dT}{dt} = -H T \cdot \left(1 + \frac{1}{3}\frac{d \ln g_{*s}}{d \ln T}\right)^{-1} \tag{24.14}$$

where $g_{*s}$ is the effective number of entropic degrees of freedom. During the BBN epoch ($T \lesssim 10^{10}\,\text{K}$), $e^+e^-$ annihilation is occurring, which changes $g_*$ and heats the photon bath relative to the neutrinos. This must be tracked carefully.

In the simplified treatment (constant $g_*$), the relation reduces to:

$$t \approx \frac{132\,\text{s}}{T_9^2} \cdot \left(\frac{10.75}{g_*}\right)^{1/2}$$

24.8 Numerical BBN: State of the Art

24.8.1 Historical Development

The first quantitative BBN calculations were performed by Ralph Alpher, Hans Bethe, and George Gamow in their famous 1948 "alpha-beta-gamma" paper (the inclusion of Bethe was a pun — Gamow could not resist having the author list read $\alpha$-$\beta$-$\gamma$). Alpher's doctoral thesis, supervised by Gamow, contained the first serious attempt to compute primordial element abundances.

The modern era of precision BBN began with the work of Wagoner, Fowler, and Hoyle (1967), who developed the first comprehensive numerical codes with realistic nuclear reaction networks. Robert Wagoner's code (and its descendants) remains the standard. The key subsequent developments include:

• Kawano (1992): The NUC123 code — a publicly available Fortran code that became the community standard. It tracks 26 nuclides and 88 reactions.
• Smith, Kawano, & Malaney (1993): Updated nuclear rates.
• Cyburt, Fields, & Olive (2001–2015): Systematic updates with Monte Carlo uncertainty propagation.
• Pisanti et al. (2008), Consiglio et al. (2018): The PArthENoPE code — a modern public BBN code.
• Pitrou et al. (2018): The PRIMAT code — includes QED corrections, neutrino oscillations, and incomplete neutrino decoupling.
• Fields (2011), Coc & Vangioni (2017): Reviews of the current state of BBN.

24.8.2 Modern Precision

State-of-the-art BBN calculations include:

1. Full nuclear network: 12 nuclides ($n$, $p$, $d$, ${}^3\text{H}$, ${}^3\text{He}$, ${}^4\text{He}$, ${}^6\text{Li}$, ${}^7\text{Li}$, ${}^7\text{Be}$, ${}^8\text{Li}$, ${}^{11}\text{B}$, ${}^{11}\text{C}$) and $\sim 100$ reactions. The minimal 8-species, 12-reaction network captures $> 99.9\%$ of the physics.

2. QED radiative corrections: Finite-temperature QED corrections to the electron mass and the equation of state at the 0.1% level.

3. Neutrino decoupling: Neutrinos do not decouple instantaneously at $T \sim 1\,\text{MeV}$. Residual $\nu$-$e$ interactions during $e^+e^-$ annihilation slightly heat the neutrinos, giving $N_{\text{eff}} = 3.044$ rather than 3.000.

4. Neutron lifetime: The input $\tau_n = 879.4 \pm 0.6\,\text{s}$ (PDG 2023) is the dominant source of theoretical uncertainty in $Y_p$.

5. Nuclear rate uncertainties: Monte Carlo sampling over experimental rate uncertainties, propagated through the full network.

The resulting theoretical uncertainties are impressively small: - $\delta Y_p \approx 0.0002$ ($\sim 0.1\%$) - $\delta(\text{D/H}) \approx 3\%$ (dominated by nuclear rate uncertainties, especially the $d(p,\gamma){}^3\text{He}$ rate) - $\delta({}^7\text{Li}/\text{H}) \approx 12\%$ (dominated by the ${}^3\text{He}(\alpha,\gamma){}^7\text{Be}$ rate)

24.8.3 The Role of Nuclear Cross Section Measurements

The precision of BBN predictions is ultimately limited by the precision of the input nuclear cross sections. Ongoing experimental programs at underground laboratories (to reduce cosmic-ray backgrounds) and with radioactive beams are continually improving these measurements:

• LUNA (Laboratory for Underground Nuclear Astrophysics, Gran Sasso): Has measured several key BBN reactions, including $d(p,\gamma){}^3\text{He}$ and ${}^3\text{He}({}^4\text{He},\gamma){}^7\text{Be}$, at energies directly relevant to BBN.

• $d(p,\gamma){}^3\text{He}$: This reaction is the dominant destruction channel for deuterium. The LUNA measurement (Mossa et al. 2020) reduced the uncertainty in D/H from $\sim 4\%$ to $\sim 1\%$, making the BBN prediction for deuterium as precise as the best observational determination.

• ${}^3\text{He}(\alpha,\gamma){}^7\text{Be}$: This reaction controls ${}^7\text{Be}$ production and hence the predicted ${}^7\text{Li}$ abundance. Improved measurements have not resolved the lithium problem — the predicted ${}^7\text{Li}$ remains too high.

🔗 Connection to Chapter 21: The Gamow peak formalism developed in Chapter 21 is exactly how these reaction rates are computed. The BBN temperatures ($T_9 \sim 0.1$–$1$) correspond to center-of-mass energies of $\sim 10$–$300\,\text{keV}$ — squarely in the Gamow window for the light-element reactions.

24.9 Observations of Primordial Abundances

24.9.1 Primordial Deuterium

Primordial deuterium is measured in high-redshift ($z \sim 2$–$4$), low-metallicity gas clouds seen in absorption against background quasars. These damped Lyman-alpha (DLA) systems have undergone minimal chemical processing and preserve a near-primordial D/H ratio.

The measurement technique exploits the isotope shift between the hydrogen and deuterium Lyman-series absorption lines. At the Lyman-$\alpha$ transition (1216 A), the deuterium line is shifted by $-81.6\,\text{km/s}$ (corresponding to the reduced-mass difference between H and D). In a sufficiently simple absorption system, the deuterium absorption feature is cleanly separated from the hydrogen line.

The current best determination, from a compilation of seven high-quality DLA systems (Cooke et al. 2018), is:

$$\frac{\text{D}}{\text{H}}\bigg|_{\text{obs}} = (2.547 \pm 0.025) \times 10^{-5}$$

This $\sim 1\%$ measurement is in beautiful agreement with the BBN prediction of $(2.52 \pm 0.03) \times 10^{-5}$.

24.9.2 Primordial Helium-4

Measuring $Y_p$ is challenging because helium is produced in stars, and even a small amount of stellar processing increases $Y_p$ above its primordial value. The strategy is to measure the helium abundance in metal-poor extragalactic H II regions (ionized gas clouds in dwarf galaxies) and extrapolate to zero metallicity.

The helium mass fraction in an H II region is determined from the relative intensities of helium and hydrogen recombination lines (He I 5876 A, 4471 A, 6678 A, etc.). Systematic effects — including radiative transfer corrections, temperature fluctuations, underlying stellar absorption, and collisional excitation — make $Y_p$ determinations notoriously difficult.

The current best estimates are:

$$Y_p = 0.2449 \pm 0.0040 \quad \text{(Aver et al. 2015)}$$ $$Y_p = 0.2446 \pm 0.0029 \quad \text{(Peimbert et al. 2016)}$$

These are consistent with the BBN prediction $Y_p = 0.2470 \pm 0.0002$, though the observational uncertainties are $\sim 20$ times larger than the theoretical ones. Improving the $Y_p$ determination is an active area of observational cosmology.

24.9.3 The Neutron Lifetime Puzzle

Before discussing primordial lithium, we should note a related puzzle in the nuclear physics input to BBN: the neutron lifetime problem.

The neutron lifetime $\tau_n$ is measured by two independent methods:

1. Bottle experiments: Ultracold neutrons are stored in a material or magnetic trap, and the number surviving after a measured time is counted. The most precise bottle measurements give $\tau_n = 877.75 \pm 0.28\,\text{s}$ (Serebrov et al. 2018, UCN$\tau$ 2021).

2. Beam experiments: A beam of cold neutrons passes through a detector that counts the protons (or electrons) from beta decay. The most precise beam measurement gives $\tau_n = 887.7 \pm 1.2\,\text{s}$ (Yue et al. 2013).

The discrepancy between bottle and beam results is $\sim 10\,\text{s}$ — roughly $4\sigma$. This "neutron lifetime puzzle" has persisted for over a decade. One speculative explanation (Fornal & Grinstein, 2018) is that neutrons have a "dark decay" channel ($n \to \text{dark matter} + \gamma$) that would be missed by beam experiments (which detect only the standard $\beta$-decay products) but counted as a loss in bottle experiments. If such a dark decay exists, it would affect BBN by altering both the effective neutron lifetime and the $n/p$ freeze-out.

For BBN calculations, the PDG average $\tau_n = 879.4 \pm 0.6\,\text{s}$ is typically used. The $\sim 10\,\text{s}$ difference between bottle and beam values translates to a shift $\Delta Y_p \approx 0.002$ — comparable to the current observational uncertainty. Resolving the neutron lifetime puzzle is therefore important for precision BBN.

24.9.4 Primordial Lithium-7

Lithium-7 is measured from the Li I 6708 A resonance absorption line in the atmospheres of old, metal-poor halo stars. The key observation, first reported by Monique and Francois Spite in 1982, is the Spite plateau: metal-poor halo stars with $[\text{Fe/H}] < -1.5$ all show approximately the same lithium abundance, independent of metallicity and temperature (over a range of effective temperatures where lithium is not expected to be depleted by convection):

$$A(\text{Li}) \equiv \log_{10}\left(\frac{n(\text{Li})}{n(\text{H})}\right) + 12 \approx 2.2 \pm 0.1$$

corresponding to ${}^7\text{Li}/\text{H} \approx (1.6 \pm 0.3) \times 10^{-10}$.

The constancy of the Spite plateau across stars with different temperatures and metallicities is striking — it suggests a universal, primordial origin rather than a stellar one. But the plateau value is $\sim 3\times$ lower than the BBN prediction. This is the lithium problem, discussed in detail in Case Study 1.

24.10 The Physics Behind the Numbers

24.10.0 The Anthropic Significance of the Mass Gaps

It is worth pausing to reflect on what the mass gaps at $A = 5$ and $A = 8$ mean for the universe we inhabit. Without these gaps:

• BBN would have synthesized significant quantities of carbon, nitrogen, and oxygen in the first few minutes. The primordial universe would have been chemically complex.
• Stars would have formed from gas that was already enriched in heavy elements. The first generation of stars (Population III) would not have been metal-free.
• The chemical enrichment of the universe — the process by which successive generations of stars build up the heavy elements — would be quantitatively very different. The metallicity-dependent processes that drive galaxy evolution would be altered.
• The "delayed" production of carbon inside stars, which takes hundreds of millions of years after the Big Bang, is what sets the clock for the emergence of carbon-based chemistry and eventually life. If carbon had been produced in the first three minutes, the timeline for cosmic chemical evolution would be dramatically compressed.

The mass gaps are not fundamental symmetries — they are contingent properties of nuclear physics that follow from the specific character of the nuclear force (Chapter 3). The absence of a bound $A = 5$ state, for instance, results from the spin-orbit structure of the nucleon-nucleon interaction. A slightly different nuclear force could have produced a bound ${}^5\text{He}$, with profound consequences for cosmology.

24.10.1 Why $Y_p \approx 0.25$ and Not Some Other Number

The helium mass fraction $Y_p \approx 0.25$ is not a coincidence — it is determined by fundamental physics. Let us trace the chain of reasoning:

1. $Q = 1.293\,\text{MeV}$ (neutron-proton mass difference) $\to$ set by QCD and QED.
2. $T_f \sim 0.8\,\text{MeV}$ (freeze-out temperature) $\to$ set by $G_F$, $M_{\text{Pl}}$, and $g_*$.
3. $n/p|_f \sim 1/6$ $\to$ from $\exp(-Q/kT_f)$ evaluated at $T_f$.
4. Delay of $\sim 3\,\text{min}$ (deuterium bottleneck) $\to$ set by $B_d = 2.224\,\text{MeV}$ and $\eta \sim 6 \times 10^{-10}$.
5. $n/p \sim 1/7$ at nucleosynthesis $\to$ from free neutron decay ($\tau_n = 879\,\text{s}$) during the delay.
6. All neutrons into ${}^4\text{He}$ $\to$ from the efficiency of the reaction network and the mass gaps at $A = 5, 8$.
7. $Y_p = 2(n/p)/(1 + n/p) = 2/8 = 0.25$ $\to$ simple neutron counting.

Each step depends on measured physical constants. The remarkable thing is that all seven numbers have been independently measured, and the prediction $Y_p \approx 0.247$ agrees with the observation $Y_p \approx 0.245 \pm 0.004$.

24.10.2 Why D/H $\sim 10^{-5}$

Deuterium is the "leftover" — the material that did not make it all the way to ${}^4\text{He}$. Its abundance is determined by the competition between the deuterium production rate and the rates of the deuterium-destroying reactions (R2–R6). At the CMB value of $\eta$, this competition leaves a residual $\text{D/H} \sim 10^{-5}$.

The steep $\eta$-dependence ($\text{D/H} \propto \eta^{-1.6}$) arises because higher baryon density means: - More targets for deuterium-burning reactions. - The deuterium bottleneck breaks earlier (at higher density). - Both effects accelerate deuterium destruction.

24.10.3 Why ${}^7\text{Li}/\text{H} \sim 10^{-10}$

Lithium-7 is produced through a minor side channel. At $\eta \sim 6 \times 10^{-10}$, the dominant production pathway is:

$${}^3\text{He} + {}^4\text{He} \to {}^7\text{Be} + \gamma \quad \to \quad {}^7\text{Be} + e^- \to {}^7\text{Li} + \nu_e$$

The tiny ${}^7\text{Li}$ abundance reflects the fact that this is a small branch off the main flow from D to ${}^4\text{He}$. The ${}^7\text{Li}$ abundance vs. $\eta$ curve has a distinctive minimum near $\eta \sim 3 \times 10^{-10}$, where the direct ${}^7\text{Li}$ production (dominant at low $\eta$) is balanced by the ${}^7\text{Be}$ channel (dominant at high $\eta$). At the Planck value of $\eta$, we are on the rising (high-$\eta$) side of this minimum.

24.11 The Progressive Project: BBN Network Solver

🖥️ Code Project — bbn_network.py

In this chapter's project, you will implement a simplified BBN reaction network as a system of coupled ODEs, solve it numerically, and explore how the primordial abundances depend on the baryon-to-photon ratio $\eta$.

What you will build: - A system of 8 coupled ODEs tracking $Y_n$, $Y_p$, $Y_d$, $Y_{{}^3\text{H}}$, $Y_{{}^3\text{He}}$, $Y_{{}^4\text{He}}$, $Y_{{}^7\text{Li}}$, and $Y_{{}^7\text{Be}}$. - Parameterized thermonuclear reaction rates as functions of temperature. - The time-temperature relation for the radiation-dominated era. - A stiff ODE solver (using scipy.integrate.solve_ivp with the Radau or BDF method). - Plots of abundance evolution vs. time and temperature. - A parameter study showing the dependence of final abundances on $\eta$.

Key learning outcomes: - Experience with stiff ODE systems (a common feature of nuclear reaction networks). - Understanding of how microphysics (cross sections) determines macrophysics (cosmic abundances). - Practice with the Gamow peak formalism from Chapter 21 applied to real astrophysical reactions.

See code/bbn_network.py for the complete implementation and code/project-checkpoint.md for the detailed project checkpoint.

What's Next

In Chapter 25, we turn to the densest objects in the universe that are not black holes: neutron stars. Where BBN probes nuclear physics at low density ($\rho \sim 10^{-5}\,\text{g/cm}^3$) and high temperature ($T \sim 10^9\,\text{K}$), neutron stars probe nuclear physics at extreme density ($\rho \sim 10^{14}$–$10^{15}\,\text{g/cm}^3$) and relatively low temperature. The equation of state of dense nuclear matter — the same nuclear force we studied in Chapter 3, pushed to its limits — determines the structure, mass, and radius of neutron stars. Gravitational wave observations from neutron star mergers (GW170817, Chapter 23) now constrain this equation of state, connecting the physics of femtometers to the physics of kilometers.

Chapter 24 draws on material from Chapters 21 (thermonuclear reaction rates, Gamow peak) and 22 (stellar nucleosynthesis, reaction networks). The nuclear cross sections measured in laboratories at LUNA, TUNL, and other facilities worldwide are the essential input. The lithium problem remains one of the most intriguing unsolved puzzles at the interface of nuclear physics and cosmology.