Episode:
63
SuperPhénix
Country:
France
Years of Operation:
1985-1998
Category:
Commercial & Power
Reactor Type:
SFR
Coolant:
Sodium
Fuel Type:
Mixed Oxide (MOX)
Moderator:
Thermal Power (MWth):
3000
Electrical Power (MWe):
3000
Status:
Commercial & Power
timeline
First Criticality Year
1985
Commercial Op Year
1986
Shutdown Year
1998
Lessons Learned
• Coolant choice determines maintenance reality.
Thermal performance means little if inspection, repair, and leak isolation are slow, complex, and
invasive.
• Capacity factor is the ultimate truth metric.
A reactor that works occasionally is economically indistinguishable from one that doesn’t work
at all.
• Peripheral systems dominate reliability.
Valves, seals, piping, sensors, and access paths matter as much as fuel and neutronics.
Promising technology. Unforgiving reality. The burden of proof is still very much on
sodium
#ForgottenReactor #NuclearEngineering #NuclearHistory #FastReactors #AdvancedNuclear
#LiquidSodium
SuperPhenix died because the long outages chained together destroyed its economic reputation:
1.1987-1989: Sodium Fuel Drum Leak (20+ months) -> Solved, but expensive.
2.1990-1992: Argon Air Ingress + Turbine Building Roof Collapse (20+ months) -> Made
the plant look cursed.
3.1994-1995: IHX Argon Leak (12 months) -> Regulatory nightmare.
4.1996-1998: Scheduled Maintenance -> Political opportunity to kill it.
It is accurate to say SuperPhenix died from a "death by a thousand cuts." The liquid sodium
made ALL nuclear repairs slow, but the Turbine Building roof collapse and the administrative
freezing of the final maintenance outage made the downtime percentage so catastrophically high.
sources
ARTICLE
A 1,200-MW reactor that worked… and still barely ran.
Superphénix is what happens when beautiful nuclear physics collides head-on with ugly industrial reality.
Construction began in 1977 near Creys-Malville in eastern France. The goal was bold: build a
1,200+ MWe liquid-sodium-cooled fast breeder reactor that could generate power while
producing more fuel than it consumed. In theory, this was the long-term future of nuclear
energy—close the fuel cycle, stretch uranium resources for centuries, and move beyond
conventional light-water reactors.
On paper, Superphénix was magnificent.
After roughly a decade of construction, the plant entered operation in the late 1980s. And here’s the part that often gets rewritten by history: the reactor itself worked. The core behaved. The neutronics were fine. When the plant was online, it produced massive amounts of power exactly
as designed.
So what went wrong?
Liquid sodium.
Over its operational life, Superphénix achieved an average capacity factor of about 7.9%. For a gigawatt-scale plant that costs billions, that number is catastrophic. Not because the reactor was unstable—but because sodium leaks – though small in number – were catastrophic for uptime.
They were very difficult to locate, slow to repair, and punishingly expensive to qualify and restart.
Each small leak triggered long outages. Draining systems. Inspecting kilometers of piping.
Repairing exotic materials. Cleaning up chemically reactive metal. Re-certifying systems. Restarting. Again. And again.
This wasn’t a failure of nuclear physics. It was a failure of maintainability at an industrial scale.
Superphénix was permanently shut down in 1998, far earlier than its technical lifetime. The plant was rendered vulnerable by its downtime and politically killed in 1998 by a new French administration.
Modern fast reactor designers argue they’ve solved these historical problems. I genuinely hope they have.
But hope is not data.
Until someone operates a sodium-cooled reactor for decades at a high capacity factor with
reasonable maintenance burden, Superphénix remains one of the clearest reminders in nuclear engineering that reliability is built not just in the core…
but in every pipe, valve, seal, sensor, access hatch, and maintenance procedure surrounding it.
SLIDE DECK
