One reaction, two products, 17.6 MeV
The most practical fusion reaction for near-term power generation fuses two hydrogen isotopes, deuterium and tritium, into helium plus a neutron:
²H + ³H → ⁴He (3.5 MeV) + n (14.1 MeV) Total energy released: 17.6 MeV per reaction
The mass deficit between reactants and products converts directly to energy via ΔE = Δm·c² — about 10 million times more energy per unit mass than burning coal. The helium-4 nucleus, being charged, stays trapped by the magnetic field and reheats the plasma, which is what eventually lets a reactor sustain itself. The neutron, having no charge, passes straight through the magnetic confinement and deposits its much larger 14.1 MeV share of the energy as heat in a surrounding blanket — the heat that actually drives a turbine. Deuterium is abundant (about 30 g per cubic metre of seawater); tritium is radioactive and scarce, so power plants are designed to breed their own from lithium-6 in that same neutron blanket.
Tunnelling through the Coulomb barrier
Both nuclei carry positive charge, so they repel each other electrostatically. At nuclear range (~1 fm) that Coulomb barrier is about 1.4 MeV — but a plasma at 100 million °C only gives ions around 10 keV of thermal energy, two orders of magnitude short. Fusion still happens because of quantum tunnelling: a nucleus's wavefunction extends slightly through the barrier, giving a small but nonzero chance of spontaneous penetration on every close approach. D-T has the highest reaction cross-section of any practical fuel pair at the lowest energy, peaking around 65 millibarns near 100 keV, which is exactly why it's the first-generation fuel of choice.
The Lawson criterion: three ways to hit one number
For a plasma to produce more energy than it costs to sustain, the product of density n, temperature T, and energy confinement time τE has to clear a threshold:
n · T · τE ≥ 3×10²¹ m⁻³ · keV · s (D-T, 15–20 keV)
Because it's a product of three factors, there are three viable strategies: push density extremely high and hold it for nanoseconds (inertial confinement), hold a moderate density for seconds (magnetic confinement), or push temperature higher to raise the reaction cross-section — real reactors combine all three to some degree.
Tokamaks: magnetically suspended plasma
A tokamak confines plasma in a toroidal (donut-shaped) vessel using powerful superconducting magnets, so the 100–200 million °C plasma never touches a physical wall. Two field components — toroidal (running along the donut) and poloidal (wrapping around its cross-section) — combine into helical field lines that charged particles spiral along, unable to easily escape sideways. ITER, under construction in France since 2010, is the largest tokamak ever built, with an 847-tonne central solenoid producing a 13-tesla field; its design target is Q = 10, with D-T experiments expected around 2035. Japan's JT-60SA achieved first plasma in December 2023 and is currently the largest operating superconducting tokamak.
Q factor, ignition, and where things stand
The Q factor is simply fusion power out divided by heating power in. Q = 1 is scientific breakeven; Q = ∞, called ignition, is the point where the plasma's own fusion-generated heat (via the trapped helium-4) sustains the reaction with no external heating input at all — the ultimate goal of the whole field. In December 2022 the National Ignition Facility's 192 lasers, delivering 2.15 MJ to compress a millimetre-scale D-T capsule to 100 billion atmospheres in under 10 nanoseconds, produced 3.15 MJ of fusion energy from 2.05 MJ of laser light — the first time any experiment exceeded Q = 1 on that basis, though the roughly 300 MJ drawn from the grid to power the lasers means the plant as a whole is nowhere near net electricity yet. On the magnetic-confinement side, JET (UK) set a record of 59 MJ over five seconds of D-T plasma in 2022 (Q ≈ 0.67), and ITER's Q = 10 target, plus a wave of private ventures (Commonwealth Fusion, TAE, Helion) targeting net electricity in the 2030–2035 window, define the next decade of progress.
Frequently asked questions
What exactly does the D-T fusion reaction produce?
Deuterium and tritium fuse into helium-4 plus a free neutron: ²H + ³H → ⁴He (3.5 MeV) + n (14.1 MeV), releasing 17.6 MeV total. The helium nucleus (an alpha particle) stays trapped in the magnetic field and reheats the plasma, sustaining the reaction, while the uncharged neutron escapes the confinement entirely and deposits its larger share of the energy as heat in a surrounding blanket, which is what actually drives the steam turbines in a fusion power plant design.
What is the Lawson criterion and why does it define feasibility?
The Lawson criterion says that net energy gain requires the product of plasma density (n), temperature (T), and energy confinement time (τE) to exceed roughly 3×10²¹ m⁻³·keV·s for D-T fuel. Because it's a product of three quantities, there are three routes to satisfying it: inertial confinement pushes density extremely high for a few nanoseconds, magnetic confinement holds a much lower density for seconds, and both approaches also push temperature into the hundreds of millions of degrees to raise the reaction cross-section. Falling short on any one factor means falling short overall.
What does the Q factor mean, and has any experiment reached Q = 1?
Q is the ratio of fusion power produced to the power used to heat the plasma; Q = 1 is scientific breakeven, and Q = ∞ (ignition) means the plasma sustains itself on its own fusion-produced heat with no external heating at all. In December 2022 the National Ignition Facility became the first experiment in history to exceed Q = 1 on the laser-to-fusion-energy basis (3.15 MJ out from 2.05 MJ of laser light in), though total wall-plug power draw remained far higher. ITER's design target is Q = 10, with D-T operation expected around 2035.
Try it live
Every reaction, field and Lawson threshold above runs live in Nuclear Fusion — Tokamak & Lawson Criterion. Tune density, temperature and confinement time and watch how close the plasma gets to ignition.
▶ Open Nuclear Fusion simulation