Nuclear fusion: the true, the false and the uncertain

TRUE — Fusion can generate net positive energy within a limited scope.

The Lawrence Liv­er­more Nation­al Lab Nation­al Igni­tion Facil­i­ty (NIF) in Cal­i­for­nia proved this in 2022 with its laser igni­tion fusion set­up. The NIF set­up uses 192 of the world’s high­est-ener­gy lasers to pulse up to 4 mil­lion joules of ultra­vi­o­let ener­gy onto a deu­teri­um and tri­tium tar­get. The tar­get, approx­i­mate­ly as big as a pep­per­corn, is sus­pend­ed in a small x‑ray “oven” called a hohlraum, which can heat up to about 3 mil­lion degrees Cel­sius when hit by these pow­er­ful lasers. This caus­es the fuel to implode, cre­at­ing the con­di­tions for fusion. On 5^th Decem­ber 2022, NIF achieved tar­get gain, mean­ing that 2.05 mega­joules of laser ener­gy deliv­ered to the tar­get gen­er­at­ed 3.15 mega­joules of fusion ener­gy. More ener­gy came out of the tar­get than was put in. This point, called “igni­tion”, was a breakthrough.

FALSE — That reaction created energy, but nowhere near enough to power the facility.

Over­all, the sys­tem still expe­ri­enced a sub­stan­tial net ener­gy loss. To give a sense of scale, the target’s 1.1 mega­joules of net ener­gy is approx­i­mate­ly 0.306 kWh. By com­par­i­son, that would only keep a small 5W LED light bulb on for 20 hours after con­vert­ing the heat to elec­tric­i­ty. In addi­tion, NIF sci­en­tists esti­mate the NIF laser facil­i­ty typ­i­cal­ly requires about 100 times more ener­gy to run than the amount of ener­gy deliv­ered by the laser to the tar­get itself.

UNCERTAIN — It’s unclear when and how this crucial point, called the “engineering gain”, will be met.

At this point, engi­neers need to focus on find­ing addi­tion­al effi­cien­cies in the sys­tem to ensure it pro­duces more ener­gy than it uses. One way to do this could be to reduce the system’s ener­gy use rel­a­tive to out­put by mak­ing more ener­gy-effi­cient com­po­nents, such as lasers or super­con­duc­tors. Changes like tweak­ing the heat insu­la­tion or rolling out AI con­trols to oper­ate sys­tems faster than a human would also help. Oth­er gains can be made by improv­ing the machine’s mate­ri­als and com­po­nents to allow the sys­tem to oper­ate at high­er pow­er lev­els. To do so, engi­neers could look to include mate­ri­als that can with­stand extreme tem­per­a­tures and design even stronger mag­nets to bet­ter con­fine and con­trol the plas­ma used in fusion reac­tions. Anoth­er approach is to improve the process that cap­tures and con­verts the ener­gy from the fusion reac­tion to electricity.

TRUE — Fusion is generally seen as “clean” energy.

It pro­duces sub­stan­tial­ly less radioac­tive “waste” than fis­sion – though it is pos­si­ble that with emerg­ing tech­nolo­gies, waste from fusion and fis­sion could be reused. Still, like oth­er nuclear fis­sion, fusion will require appro­pri­ate and com­pre­hen­sive over­sight. One con­cern is that the reac­tion could be used to gen­er­ate fis­sile mate­ri­als usable in weapons. Fusion machines and relat­ed reac­tions do not direct­ly pro­duce mate­r­i­al use­ful for weapons. The reac­tion does, how­ev­er, cre­ate an enor­mous amount of neutrons.

On the bright side, these neu­trons could help gen­er­ate more fuel for the fusion reac­tion — many designs plan to incor­po­rate a “breed­ing blan­ket,” a lay­er of mate­ri­als that acts as heat insu­la­tion, but is also lined with mate­ri­als that can cap­ture the neu­trons to cre­ate more tri­tium. Ura­ni­um or tho­ri­um could also be placed in some breed­ing blan­ket designs. The con­cern is that these mate­ri­als, once irra­di­at­ed, could gen­er­ate ura­ni­um-235 that can be used in nuclear weapons. Phys­i­cal ways to deter this process exist, such as requir­ing the use of lithium‑6 in the blan­ket mod­ules. The IAEA will be impor­tant in ensur­ing non-pro­lif­er­a­tion safe­guards and oversight.

FALSE — Fusion energy may be near limitless, but that doesn’t necessarily translate to unending energy.

Some fusion ener­gy tech­nolo­gies under devel­op­ment could the­o­ret­i­cal­ly pro­duce more fuel than they con­sume and, there­fore, be essen­tial­ly lim­it­less. But that does­n’t mean that this would pro­vide the ener­gy soci­ety needs. Most researchers expect build­ing and oper­at­ing ear­ly fusion plants to be very expen­sive. Whether soci­ety will be will­ing to pay to run cost­ly fusion reac­tors will depend on how fusion fits in with oth­er clean ener­gy systems. 

UNCERTAIN — When fusion will start powering the world is still unclear.

I am opti­mistic that fusion will even­tu­al­ly pro­vide clean ener­gy to at least parts of the world. How­ev­er, the tech­nol­o­gy is unlike­ly to be ready to entire­ly sup­port the tran­si­tion away from car­bon fuels in its own. This lag could put it at a dis­ad­van­tage com­pared to oth­er prod­ucts adopt­ed ear­ly on that could be deployed at a much larg­er scale. Still, we could see a cas­cade of break­throughs that quick­ly accel­er­ates progress for fusion. Or we may have to wait a long time before the next break­through arrives.

TRUE — Startups are saying they are ready to build commercial pilot plants.

Sev­er­al star­tups have announced very ambi­tious time­lines. Helion, for instance, has promised to start pro­duc­ing ener­gy from a fusion plant by 2028¹. In their 2023 report, the Fusion Indus­try Asso­ci­a­tion found that many oth­ers believe a fusion plant will deliv­er elec­tric­i­ty to the grid before 2035². Fusion firms are indeed mak­ing iter­a­tive progress and march­ing steadi­ly toward suc­cess. The NIF, which proved fusion igni­tion, has pro­vid­ed cru­cial data that will help guide research programs—particularly for laser igni­tion type of fusion. Some firms have also start­ed rolling out arti­fi­cial intel­li­gence to opti­mise their approach to fusion, gen­er­at­ing some inter­est­ing results. Still, mak­ing progress is dif­fer­ent from hav­ing a prod­uct available.

FALSE — No startup has shown evidence that they reached the stage of development needed to roll out fusion to the market in the near term.

Star­tups still need to demon­strate their sys­tems’ net ener­gy gain and over­all engi­neer­ing gain. They also have to learn how to run their fusion reac­tion on scales that can gen­er­ate a prof­it. As much as fusion com­pa­nies want to put a time­line on achiev­ing these tech­no­log­i­cal mile­stones, these are break­throughs you can’t schedule.

UNCERTAIN — There’s also a tricky commercial case to overcome.

Fusion is still a very high-risk invest­ment because the tech­nol­o­gy has yet to be proven. It’s not like invest­ing in solar cells 15 years ago when they were already at 20% effi­cien­cy. It’s more like invest­ing in solar cells 40 years ago when they had 1% effi­cien­cy and a very small roll­out. Star­tups will need to build pilot plants to prove their con­cept. Rais­ing enough cap­i­tal to make those plants will like­ly be tricky. First-gen­er­a­tion sta­tions are like­ly to be cost­ly and unreliable—that is just a part of inno­va­tion. Still, assum­ing we’re going to move away from fos­sil fuels, that pop­u­la­tion growth dra­mat­i­cal­ly increas­es, and devel­op­ing nations’ demand for ener­gy con­tin­ues to grow, there’s a vast poten­tial mar­ket for all clean ener­gy sources mov­ing forward.

Marianne Guenot