Fusion Energy: Why the Hardest Engineering Problem Is Getting Closer to Being Solved

# The Tritium Problem: Fusion's Fuel Supply Challenge Nobody Discusses

When fusion's remaining challenges are discussed, the conversation typically focuses on plasma physics (ignition, confinement stability), engineering challenges (neutron damage to materials, reactor efficiency), or economics (cost versus renewables). All of these are real and important. But there's a challenge that consistently gets skipped in popular fusion coverage that is, in my view, the most underappreciated problem in the entire field: tritium supply.

The D-T fusion reaction we're building all these reactors to use requires tritium. And tritium barely exists.

## Tritium's Unusual Properties

Tritium is hydrogen-3: one proton, two neutrons. It's radioactive, with a half-life of 12.3 years, decaying to helium-3 by beta emission. This short half-life means that tritium doesn't accumulate naturally — any tritium produced must be used relatively quickly before it decays.

Tritium doesn't occur in meaningful quantities in nature. The entire natural inventory from cosmic ray interactions with the atmosphere is about 3.5 kilograms globally at any given time. The total tritium inventory on Earth — including what's held in nuclear programs — is approximately 20-25 kilograms. That's roughly the weight of a medium-sized dog. The entire global supply of fusion fuel.

## Where Tritium Comes From

Tritium is produced in fission reactors by neutron bombardment of lithium-6:

Li-6 + n → T + He-4 + 4.8 MeV

The CANDU reactors operated by Ontario Power Generation in Canada are the world's primary commercial tritium producers. They produce tritium as a byproduct of their heavy water moderator becoming contaminated through neutron bombardment. Canada extracts and sells this tritium to medical, industrial, and defense customers. The annual global tritium production is on the order of 1-2 kilograms per year.

## The Scaling Problem

A commercial D-T fusion power plant burning 1 gigawatt of thermal power would consume approximately 55 kilograms of tritium per year. A fleet of 100 commercial fusion reactors would need 5,500 kilograms per year — many hundreds of times the current global production capacity.

This isn't just an economic problem. It's a physical impossibility under current conditions. You cannot build 100 D-T fusion plants without a massive expansion of tritium production. And the only way to massively expand tritium production is either to build many more fission reactors (which creates its own political and economic complications) or to use fusion's own neutrons to breed tritium.

## The Breeding Blanket Concept

The solution to the tritium supply problem, in principle, is the breeding blanket. Fusion reactions produce 14.1 MeV neutrons in large quantities. If you surround the fusion plasma with a blanket containing lithium, those neutrons can react with lithium to produce new tritium:

Li-6 + n → T + He-4 (plus energy)
Li-7 + n → T + He-4 + n' (slightly slower neutrons, net tritium gain possible with multipliers)

A fusion reactor with a tritium breeding blanket would, in theory, produce more tritium than it burns — tritium breeding ratio (TBR) > 1 — allowing a fleet of reactors to be tritium self-sufficient after an initial startup investment.

## The Bootstrapping Problem

Here's the circular logic that makes this hard. You need tritium to run fusion reactors. You need fusion reactors (with breeding blankets) to produce tritium at scale. How do you start?

The current plan involves using the limited existing tritium inventory to start the first few fusion reactors, which then breed enough tritium to fuel additional reactors. This requires:
- Breeding blankets that actually work as designed (not yet demonstrated at reactor scale)
- Tritium breeding ratios above 1.05 or so to accumulate enough surplus for reactor fleet growth
- Minimal tritium losses (tritium permeates through metals, which is an engineering headache)

ITER will not breed tritium — its design includes a test blanket module program but not a full breeding blanket. The question of whether tritium breeding blankets can actually achieve TBR > 1 in a real reactor is not yet answered experimentally.

This isn't a reason to dismiss fusion, but it's a reason to be skeptical of timelines that assume rapid commercial scale-up. Before a first commercial D-T fusion plant can start a breeding program, the breeding blanket technology needs to be demonstrated. Before a fleet can grow rapidly, the breeding ratio needs to comfortably exceed 1. None of this is physically impossible, but none of it is solved.

The tritium problem doesn't appear in popular fusion coverage because it doesn't fit the narrative. It's not a dramatic physics barrier or a corporate funding story. It's a supply chain and engineering problem that will take decades to solve even after fusion plasma performance is proven. That's exactly why it deserves more attention than it gets.

The Tritium Problem: Fusion's Fuel Supply Challenge Nobody Discusses

// COMMENTS

ON THIS PAGE