Case Study 2 — ITER: Building a Star on Earth

Background

In the fields of southern Provence, near the medieval town of Cadarache, France, an international consortium is building the most complex scientific instrument ever constructed: ITER — the International Thermonuclear Experimental Reactor (the name is also Latin for "the way"). When complete, ITER will be the first fusion device designed to produce significantly more power from fusion than is supplied to heat the plasma. Its target: $Q = 10$, producing 500 MW of fusion power from 50 MW of external heating, in pulses lasting 400–600 seconds.

This case study examines the physics basis of ITER, its engineering challenges, its troubled construction history, and what it will — and will not — demonstrate.

The Physics Case for ITER

Why is ITER so large?

ITER's major radius is $R = 6.2\,\text{m}$ and its plasma volume is $840\,\text{m}^3$ — roughly 10 times the volume of JET, the largest existing tokamak. This size is dictated by the empirical confinement scaling:

$$\tau_E \propto I_p^{0.93} R^{1.97} \kappa^{0.78} \cdots P^{-0.69}$$

The energy confinement time $\tau_E$ increases strongly with size ($R^{1.97}$) and plasma current ($I_p^{0.93}$), but decreases with heating power ($P^{-0.69}$, the "power degradation"). To reach the triple product $n\tau_E T$ required for $Q = 10$ with the known scaling, the device must be large enough that confinement improves faster than losses increase.

JET achieved $Q \approx 0.67$ with $R = 3.0\,\text{m}$. Scaling to $Q = 10$ required roughly doubling the linear dimension and operating at a higher magnetic field.

Key Design Parameters

The burning plasma regime

ITER's most important scientific goal is not merely producing fusion power — it is achieving a burning plasma, in which the dominant heating source is the fusion-produced alpha particles rather than the external heating systems. In a $Q = 10$ plasma, the alpha heating power is:

$$P_\alpha = \frac{E_\alpha}{E_{\text{fus}}} \times P_{\text{fus}} = \frac{3.5}{17.6} \times 500 = 100\,\text{MW}$$

This is twice the external heating power (50 MW). The plasma is predominantly self-heated — a qualitatively new regime that has never been explored experimentally. The physics of a burning plasma (alpha transport, alpha-driven instabilities, self-organized profiles) cannot be fully studied in non-burning experiments.

Engineering Challenges

Superconducting magnets

ITER will be the first large tokamak with all superconducting magnets — no copper coils. The toroidal field is produced by 18 D-shaped coils, each 17 m tall and weighing $\sim 360\,\text{tonnes}$. The conductor is Nb$_3$Sn (niobium-tin), cooled to 4.5 K by supercritical helium. The maximum field on the conductor is $\sim 11.8\,\text{T}$.

The central solenoid — the largest superconducting magnet ever built — stores $6.4\,\text{GJ}$ of magnetic energy and drives the 15 MA plasma current inductively. It was manufactured by General Atomics in the United States as six stacked modules.

The magnetic field energy stored in all ITER coils is approximately $41\,\text{GJ}$ — equivalent to about 10 tonnes of TNT. Managing this stored energy in the event of a quench (sudden loss of superconductivity) is a major engineering challenge.

The divertor

Exhaust particles and heat from the plasma are channeled along open magnetic field lines to the divertor — a set of tungsten-clad plasma-facing components at the bottom of the vacuum vessel. The divertor must handle:

• Heat flux: Up to $10\,\text{MW/m}^2$ steady state (comparable to the heat flux on the surface of the Sun)
• Particle flux: $\sim 10^{24}$ ions per $\text{m}^2$ per second
• Neutron damage: $\sim 5\,\text{dpa}$ over the ITER operational lifetime

ITER's divertor target plates are made of tungsten monoblocks bonded to copper alloy heat sinks, actively cooled by pressurized water. The design was validated in high-heat-flux test facilities, but performance under simultaneous neutron damage and plasma bombardment has never been tested.

Tritium systems

ITER will use D-T fuel, consuming approximately 0.5 kg of tritium per full-power year. Since the world's tritium supply is limited ($\sim 25\,\text{kg}$, primarily from CANDU heavy-water reactors), tritium management is critical.

ITER will include Test Blanket Modules (TBMs) — six ports equipped with prototype breeding blankets using lithium compounds to test tritium breeding and extraction technology. However, ITER itself is not designed for tritium self-sufficiency; it will rely on externally supplied tritium. Demonstrating a tritium breeding ratio (TBR) $> 1$ is deferred to the next-step device (DEMO).

Disruptions

A disruption is a sudden, uncontrolled loss of plasma confinement. In ITER, a disruption could deposit the entire plasma thermal energy ($\sim 350\,\text{MJ}$) and magnetic energy ($\sim 380\,\text{MJ}$) onto the first wall and divertor in $1\text{–}10\,\text{ms}$, along with mechanical forces (halo currents) of up to $\sim 70\,\text{MN}$ on in-vessel components.

ITER's disruption mitigation system will use shattered pellet injection (SPI) — frozen neon/deuterium pellets shattered and injected into the plasma to radiate the energy isotropically before it reaches the wall. Achieving reliable disruption mitigation with $> 95\%$ success rate is mandatory for ITER operation.

The Construction Story

ITER was proposed in 1985 at the Reagan-Gorbachev Geneva summit and formally established in 2006 with seven parties: the European Union (host), United States, Russia, Japan, China, South Korea, and India. Construction began at Cadarache in 2010.

Timeline and cost

The cost and schedule overruns have multiple causes: design changes, regulatory requirements (ITER is licensed as a nuclear facility under French law), manufacturing challenges (the superconducting magnets pushed the limits of metallurgy and cryogenics), supply chain disruptions, and the inherent complexity of coordinating in-kind contributions from seven parties on three continents.

Despite the delays, significant hardware has been completed and delivered: all 18 toroidal field coils have been manufactured, the vacuum vessel sectors are being assembled, and the central solenoid modules have been delivered. The tokamak pit (the 30-meter-deep concrete structure housing the machine) is complete, and machine assembly is underway.

Management reform

In 2024, the ITER Council commissioned an independent review that led to a "New Baseline" — a revised schedule and cost estimate with more realistic assumptions. The new plan phases construction to achieve first plasma in hydrogen/helium before beginning the more demanding D-T operations, reducing technical risk in the near term.

What ITER Will Demonstrate

If ITER succeeds in achieving $Q = 10$:

1. Burning plasma physics. The first extended study of a plasma heated predominantly by its own fusion products. This includes alpha-particle confinement, alpha-driven instabilities (Alfven eigenmodes), and self-consistent temperature and density profiles.

2. Confinement scaling extrapolation. Verification that empirical confinement scalings derived from existing devices hold when extrapolated to a reactor-scale plasma. This is essential for designing the next step.

3. Integrated technology. Simultaneous operation of superconducting magnets, fueling systems, heating systems, exhaust handling, and tritium processing at reactor-relevant parameters. No existing device tests all of these simultaneously.

4. Test blanket module testing. First integrated test of tritium breeding blanket concepts in a fusion neutron environment.

What ITER Will NOT Demonstrate

1. Tritium self-sufficiency. ITER relies on external tritium and will not achieve TBR $> 1$.

2. Electricity generation. ITER has no turbine or power conversion system. The fusion power is absorbed in the cooling water and rejected as waste heat.

3. Steady-state operation. ITER's inductively driven plasma is inherently pulsed ($\sim 400\,\text{s}$ burn), though advanced scenarios with non-inductive current drive may extend to $\sim 3000\,\text{s}$.

4. Reactor economics. ITER is a science experiment, not a prototype power plant. It cannot address questions about capital cost, maintenance schemes, or component lifetime.

The HTS Magnet Revolution

A transformative development in the fusion landscape is the emergence of high-temperature superconducting (HTS) magnets, specifically those using rare-earth barium copper oxide (REBCO) tape. Unlike the low-temperature Nb$_3$Sn conductors in ITER (which require cooling to 4.5 K and have a maximum practical field of $\sim 12\,\text{T}$), REBCO tape can operate at fields exceeding 20 T and at higher temperatures (20-30 K), simplifying the cryogenic systems.

The significance for fusion is captured by a simple scaling argument. The fusion power density in a tokamak scales as:

$$P_{\text{fus}} \propto \beta^2 B^4$$

where $\beta = nkT/(B^2/2\mu_0)$ is the ratio of plasma pressure to magnetic field pressure. If $\beta$ is limited by MHD stability (the Troyon limit), doubling the magnetic field quadruples the achievable plasma pressure and increases the fusion power density by a factor of 16. A compact, high-field tokamak could potentially achieve the same total fusion power as ITER in a device with roughly $1/10$ the plasma volume.

In September 2021, Commonwealth Fusion Systems (a spin-off from MIT) demonstrated a 20 T large-bore HTS magnet — the strongest magnet of its type ever built for fusion. This milestone validated the key technology underlying their SPARC tokamak and the subsequent ARC power plant concept.

SPARC design parameters (for comparison with ITER):

SPARC aims to demonstrate burning plasma physics in a device 28 times smaller by volume than ITER, at a fraction of the cost. If successful, it would validate the compact-high-field pathway and potentially compress the timeline to commercial fusion.

Beyond ITER: The Pathway to Fusion Power

ITER is intended to be followed by DEMO — a demonstration fusion power plant that would generate electricity and demonstrate tritium self-sufficiency. Several DEMO concepts are being developed:

• EU-DEMO: Conceptual design by EUROfusion, targeting $\sim 2\,\text{GW}$ fusion power and $300\text{–}500\,\text{MW}$ net electric. Target: late 2040s–2050s.
• CFETR (China): China Fusion Engineering Test Reactor, a two-phase approach targeting $Q > 10$ and then electricity production. Preliminary design complete.
• STEP (UK): Spherical Tokamak for Energy Production, a compact high-field spherical tokamak targeting net electricity by the 2040s.

Meanwhile, private companies are pursuing parallel paths, hoping to bypass the ITER-DEMO timeline with smaller, cheaper devices enabled by high-temperature superconducting magnets. Commonwealth Fusion Systems' SPARC device (under construction) aims to demonstrate $Q > 2$ in a tokamak with $R = 1.85\,\text{m}$ using 20 T HTS magnets — physics performance comparable to ITER in a device smaller than JET.

Discussion Questions

1. ITER's cost has increased from $\sim$\$5 billion to $\sim$\$22 billion. Is this justified? What would the consequences be if ITER were cancelled?

2. ITER will achieve $Q = 10$ in a pulsed mode ($\sim 400\,\text{s}$). A power plant needs steady-state operation. What additional physics and technology must be developed to bridge this gap?

3. The JET-to-ITER extrapolation relies on empirical scaling laws (IPB98(y,2)). What are the risks of this approach? Under what conditions might the scaling laws fail?

4. Private companies like Commonwealth Fusion Systems argue that HTS magnets can achieve ITER-level performance in a much smaller device. What physics and engineering advantages does a higher magnetic field provide? What are the risks of the compact approach?

5. China and South Korea are pursuing ambitious national fusion programs (CFETR, K-DEMO) in parallel with ITER. How does geopolitical competition affect the pace and direction of fusion research? Is competition or collaboration more productive?