Case Study 1 — From Discovery to Bomb to Reactor: The Physics of the Chain Reaction

The Problem

Between January 1939 and December 1942 — less than four years — nuclear fission went from a laboratory curiosity to a self-sustaining chain reaction in Enrico Fermi's Chicago Pile-1 (CP-1). Between December 1942 and August 1945 — less than three more years — the chain reaction was weaponized. This case study traces the physics reasoning at each critical juncture, focusing not on the politics or the ethics (which are addressed elsewhere) but on the quantitative arguments that determined feasibility. At every step, the question was: can we make $k_{\text{eff}} > 1$? The story of the chain reaction is ultimately a story about a neutron budget — and the margin was thinner than most people realize.

The Discovery: What Fission Means for a Chain Reaction

Within days of the Meitner-Frisch explanation in January 1939, physicists worldwide recognized the implications. If fission releases neutrons, and if those neutrons can cause further fissions, a chain reaction is possible. The first critical question was:

How many neutrons per fission?

In March 1939, experiments by Herbert Anderson, Enrico Fermi, and Leo Szilard at Columbia University (and independently by Frédéric Joliot, Hans von Halban, and Lew Kowarski in Paris) measured $\bar{\nu} \approx 2$–$3$ neutrons per thermal fission of $^{235}$U. The precise value matters enormously: if $\bar{\nu}$ were less than about 1.1, no chain reaction would be possible after accounting for neutron losses to parasitic absorption and leakage. The measured value of $\bar{\nu} \approx 2.4$ provided a comfortable margin — but only if neutron losses could be kept sufficiently small.

The Feasibility Calculation: Fermi's Analysis

Fermi's approach was systematic. He needed to estimate each factor in what would become the four-factor formula: $k_\infty = \eta f p \varepsilon$.

The reproduction factor $\eta$. For natural uranium (0.72% $^{235}$U), the thermal neutron absorption is dominated by $^{238}$U because of its much greater abundance. The effective $\eta$ for natural uranium is:

$$\eta_{\text{nat}} = \bar{\nu} \cdot \frac{N_{25}\sigma_f^{25}}{N_{25}\sigma_a^{25} + N_{28}\sigma_a^{28}}$$

Using modern values: $N_{25}/N_{28} = 0.0072/0.9928$, $\sigma_f^{25} = 584$ b, $\sigma_a^{25} = 683$ b, $\sigma_a^{28} = 2.68$ b:

$$\eta_{\text{nat}} = 2.43 \times \frac{0.00725 \times 584}{0.00725 \times 683 + 0.9928 \times 2.68} = 2.43 \times \frac{4.23}{4.95 + 2.66} = 2.43 \times 0.556 = 1.35$$

This is already a tight margin. Each thermal neutron absorbed in natural uranium fuel produces only 1.35 fast neutrons — far less than the 2.43 from pure $^{235}$U.

The thermal utilization factor $f$. The moderator absorbs some thermal neutrons. Fermi needed a moderator with the lowest possible absorption cross section. His choices were:

Light water absorbs too many neutrons — $k_\infty < 1$ with natural uranium in light water. This is a physics fact, not an engineering limitation. (The Canadian CANDU reactor achieves criticality with natural uranium only because heavy water has a neutron absorption cross section 600 times lower than light water.)

Fermi chose graphite because high-purity D$_2$O was unavailable in the U.S. in 1942, while graphite could be industrially purified. The penalty: graphite slows neutrons much less efficiently per collision ($\bar{\xi} = 0.158$ vs. 1.0 for hydrogen), requiring a much larger pile.

The resonance escape probability $p$. As neutrons slow down from fission energies ($\sim 2$ MeV) to thermal energies ($\sim 0.025$ eV), they pass through the enormous resonance absorption peaks of $^{238}$U, particularly the giant resonance at 6.67 eV ($\sigma_\gamma^{\text{peak}} \approx 24{,}000$ b). If uranium and graphite are homogeneously mixed, almost every neutron is captured in these resonances — $p \approx 0$ and no chain reaction is possible.

Fermi's critical insight was to use a heterogeneous lattice: lumps of natural uranium embedded in a matrix of graphite. The uranium lumps are large enough that the interior is shielded from resonance-energy neutrons (the resonance neutrons are absorbed in the outer layer of the lump, never reaching the interior). Meanwhile, neutrons slowing down in the graphite have a lower probability of encountering uranium at resonance energies. This "self-shielding" dramatically increases $p$ — from near zero (homogeneous) to $p \approx 0.80$–$0.90$ (heterogeneous lattice).

Chicago Pile-1: The Numbers

On December 2, 1942, Fermi's team assembled CP-1 under the stands of Stagg Field at the University of Chicago. The pile contained approximately:

• 385 tonnes of graphite (purity was critical — boron impurities of a few ppm would kill the chain reaction)
• 36 tonnes of uranium metal (in cylindrical slugs)
• 9 tonnes of uranium oxide (in pressed blocks)

The final configuration achieved:

The pile went critical with $k_{\text{eff}}$ barely above 1 — there was no margin to spare. The excess reactivity was about 0.06% ($\rho \approx 0.0006$), controlled by cadmium strips (cadmium has a thermal neutron absorption cross section of about 2,520 b).

The pile operated at a maximum power of about 200 watts — deliberately limited because there was no radiation shielding beyond the graphite itself. The purpose was proof of concept, not power generation.

From Pile to Bomb: The Critical Mass Problem

CP-1 demonstrated that a self-sustaining chain reaction was possible with natural uranium. But a bomb requires a fast, supercritical chain reaction — no moderator, and the multiplication factor must be much greater than 1 during the brief assembly time.

The physics challenges:

1. Enrichment or plutonium production. Natural uranium cannot sustain a fast chain reaction ($k_\infty < 1$ without a moderator). Options: (a) enrich uranium to >80% $^{235}$U, or (b) produce $^{239}$Pu in a reactor and separate it chemically.

2. Critical mass. A bare sphere of $^{235}$U metal has a critical mass of approximately 52 kg (radius $\approx 8.5$ cm). With a natural uranium reflector (which returns escaping neutrons to the core), the critical mass drops to $\sim 15$ kg.

3. Assembly speed. The weapon must be assembled from subcritical to supercritical faster than the chain reaction can disassemble it. For $^{235}$U, the prompt generation time is $\sim 10^{-8}$ s (no moderator). Each generation multiplies the neutron population by $k_{\text{eff}} \approx 2$. In 80 generations ($\sim 0.8$ $\mu$s), the neutron population increases by $2^{80} \approx 10^{24}$, releasing roughly one megaton of energy — if the assembly holds together that long. In practice, the energy release is sufficient to blow the assembly apart after $\sim 50$–$60$ generations.

The Two Designs

Little Boy (uranium, gun-type). A projectile of $^{235}$U (approximately 25 kg) was fired down a gun barrel into a target ring of $^{235}$U (approximately 39 kg) at a velocity of about 300 m/s. Assembly time: $\sim 1$ ms. This was sufficient for $^{235}$U because its spontaneous fission rate is low enough ($\sim 1$ fission per 60 $\mu$s per critical mass) that premature initiation during the 1 ms assembly is unlikely. The yield was approximately 15 kilotons of TNT equivalent, and the efficiency (fraction of $^{235}$U that actually fissioned) was only about 1.4%. Most of the fuel was blown apart before it could fission.

Fat Man (plutonium, implosion). Reactor-produced plutonium contains $\sim 6$% $^{240}$Pu, which has a spontaneous fission rate $\sim 40{,}000$ times higher per gram than $^{235}$U. During a 1 ms gun assembly, there would be a near-certainty of spontaneous fission neutrons initiating a premature chain reaction (a "fizzle") — the chain reaction would start before full assembly, and the energy released would blow the weapon apart before an efficient explosion developed. The implosion design compresses a subcritical sphere of about 6.2 kg of plutonium to supercritical density by converging detonation waves from a shell of conventional explosives. The compression increases the density by a factor of approximately 2–3, which reduces the critical mass (since $M_{\text{crit}} \propto \rho^{-2}$) and allows assembly to supercriticality in $\sim 1$ $\mu$s — 1,000 times faster than the gun method — reducing the predetonation probability to an acceptable level. The yield of Fat Man was approximately 21 kilotons.

The implosion design was far more technically challenging than the gun design. It required the development of explosive lenses (shaped charges that convert a diverging detonation wave into a converging one) and precision machining to tolerances of $\sim 0.01$ inches. The theoretical and experimental work on implosion hydrodynamics, led by Seth Neddermeyer and George Kistiakowsky at Los Alamos, was one of the most demanding physics and engineering programs of the war.

From Bomb to Reactor: The Return to Controlled Fission

After the war, the physics of the chain reaction was redirected toward power generation. The key difference is profound: a power reactor operates at delayed critical ($k_{\text{eff}} = 1$, with reactivity controlled by the delayed neutron margin $\beta$), while a weapon operates at prompt supercritical ($\rho \gg \beta$). The entire field of civilian nuclear power rests on staying safely below $\rho = \beta$ at all times.

The first nuclear power plant to deliver electricity to a grid was the AM-1 reactor in Obninsk, Russia (1954, 5 MWe). The first commercial nuclear power plant in the U.S. was Shippingport (1957, 60 MWe), designed by Admiral Hyman Rickover's team, which had already developed nuclear submarine propulsion. The Shippingport reactor was a pressurized water reactor (PWR) — the design that came to dominate the world fleet — using enriched uranium fuel and ordinary water as both moderator and coolant.

The transition from weapons to power required solving several new physics problems:

1. Long-term reactivity management. A power reactor must operate continuously for 12–24 months between refueling. The initial fuel loading must contain enough excess reactivity to compensate for fuel depletion and fission product buildup (especially $^{135}$Xe and $^{149}$Sm poisoning), while this excess reactivity must be suppressed during early operation by burnable absorbers or soluble boron.

2. Fuel burnup. In a power reactor, the fuel remains in the core for years, achieving burnups of 40,000–60,000 MWd/tHM (megawatt-days per tonne of heavy metal). At these burnups, the isotopic composition of the fuel changes significantly: $^{235}$U is consumed, $^{239}$Pu is bred from $^{238}$U (and itself fissions, contributing 30–40% of the energy by end of life), and fission products accumulate. The neutronics must be tracked throughout the fuel cycle.

3. Safety and feedback. A power reactor must be inherently stable — any increase in power must be met by a decrease in $k_{\text{eff}}$ through negative reactivity feedback (Doppler broadening, moderator temperature/density effects). This is not required in a weapon, where the goal is maximum energy release before disassembly.

Discussion Questions

1. Fermi's CP-1 achieved $k_{\text{eff}} \approx 1.0006$. If the graphite had contained just 1 ppm more boron impurity, the pile might not have gone critical. Calculate the increase in macroscopic absorption cross section $\Delta\Sigma_a$ from 1 ppm boron in graphite ($\sigma_a^{\text{B}} = 767$ b for natural boron, graphite density 1.6 g/cm$^3$, $A_C = 12$) and estimate the effect on $k_{\text{eff}}$.

2. Why is it physically impossible to make a nuclear weapon from natural uranium (without a moderator)? Frame your answer in terms of $\eta$ for fast neutrons in natural uranium.

3. The Hiroshima bomb used about 64 kg of $^{235}$U enriched to about 80%. Using the approximate critical mass formula, explain why more than one critical mass was needed.

4. Modern nuclear reactors are designed so that $\rho$ can never exceed $\beta$ during any credible accident scenario. Discuss the physical design features that enforce this constraint (negative temperature coefficient of reactivity, Doppler broadening of $^{238}$U resonances).

Key Physics Takeaways

• The chain reaction depends on a tight neutron budget: $\eta \approx 1.35$ for natural uranium leaves almost no margin for neutron losses. Every factor in the four-factor formula matters. CP-1 achieved $k_{\text{eff}} = 1.0006$ — one of the smallest margins in the history of technology.
• The heterogeneous lattice (lumping the fuel rather than mixing it homogeneously with the moderator) was the critical innovation that made a natural-uranium chain reaction possible. The self-shielding of uranium lumps dramatically improved the resonance escape probability $p$.
• The distinction between delayed critical (reactors) and prompt supercritical (weapons) is the fundamental physics boundary between controlled energy and uncontrolled energy release. The delayed neutron fraction $\beta \approx 0.0065$ is the physical parameter that separates these regimes.
• Purity of materials (graphite, uranium) was as important as the physics design. Fermi's team spent as much effort on graphite purification as on neutron physics. Boron impurities of even a few parts per million in the graphite would have prevented the chain reaction entirely.
• The spontaneous fission rate of $^{240}$Pu forced the development of the implosion design — a pivotal example of how a nuclear physics measurement (the $^{240}$Pu spontaneous fission half-life) drove a major engineering decision.
• The entire history of nuclear technology illustrates how a ~1 MeV pairing energy difference between odd-$N$ and even-$N$ compound nuclei shaped the trajectory of the 20th century. If $^{235}$U were not fissile with thermal neutrons, the path to both reactors and weapons would have been fundamentally different.