Every so often the nuclear world produces a familiar ripple. Someone completes a cool nuclear experiment. A low-key (sometimes) but earnest announcement circulates. Observers without much familiarity with the field confuse what happened with something far larger. A few political figures speak as if a reactor is nearly ready for the grid. And in the background, every nuclear engineer sighs and wonders how many decades it will take for the basics to sink in.
But seriously, remember when LLNL announced its December 2022 laser experiment, and I found myself on CBS explaining (don’t laugh—this was my first time with John Dickerson, and I was terrified) what this actually meant (versus what the headlines claimed)? The media loved the ‘fusion breakthrough’ headline, but what followed was a masterclass in misunderstanding what nuclear engineering actually involves.
Ok, the lasers produced about two megajoules of light, the capsule produced a bit over three megajoules of fusion energy (yes! that’s the energy production we want to talk about)…but the facility consumed far more than that to power the lasers in the first place. It was a remarkable physics result but not a power plant. Nonetheless, it was treated by many as if electricity were about to pour from a fusion grid any moment.
The same confusion appears repeatedly in fission. A team conducts a subcritical experiment. Or a zero power experiment. Or a non-nuclear test experiment. A press release goes out. People unfamiliar with the field think a reactor is around the corner. The truth is that all of these physics and engineering tests are vital but they are not a precursor to power production in the way most people imagine. They belong to an entirely different category of activity.
If you want to understand why, it helps to start with something even simpler than zero power. Subcritical piles have been used as teaching tools of the fundamentals of neutron behavior for generations. At Georgia Tech, we maintain a highly streamlined graphite pile that I use in class to introduce the foundational ideas of reactor physics in a hands-on setting. It is little more than stacked graphite blocks with embedded channels where a small neutron source can be inserted. There is no heat to manage. There are no control systems in any industrial sense. There is no path to commercial operation. What the pile provides is transparency of neutron physics: students see diffusion in real space rather than in a Monte-Carlo-generated plot on their computer. And interestingly, I have to teach that there IS a negative neutron flux (don’t worry, it’s an artifact of mathematical expressions we use to describe the neutron flux to solve for things). They measure how a neutron reflector alters the neutron population. They observe buckling effects that usually exist only as integro-differential equations. They learn what moderation (or neutron slowing down critical for some reactor designs) feels like in practice. They watch neutron populations respond to changes in geometry/materials rather than in theory alone.
Subcritical piles exist because the field needs a clean way to observe neutrons without the complications of temperature, fluid flow or structural stress. Zero power assemblies exist for a related reason. They allow researchers to push a system to criticality without creating meaningful heat, and therefore without inviting the full suite of engineering questions that real reactors must confront. In a zero power experiment, the goal is to confirm that the physics calculations are not drifting into unreality. The multiplication factor is measured directly (if we have the proper setup). The behavior of control elements can be mapped with precision. Spectral shifts can be quantified. Delayed neutron behavior can be anchored to something observable rather than left to tables or codes. The experiment validates the reactor model at the level of the neutron.
This kind of work is essential. It is the reason our simulation tools perform as well as they do. It is the reason safety analyses rest on something other than faith in a integro-differential equations (yes, I’m a big fan of our seven-variable neutron diffusion equation). Yet zero power experiments do not tell you how to build a reactor system any more than a wind tunnel tells you how to build an airliner. They cleanly isolate one variable at the cost of ignoring almost everything else.
A commercial reactor, by contrast, is a machine that must survive heat and time. The physics of criticality is the beginning of the story and not the middle. The real questions are thermal, but also include the copious neutrons. Can heat be removed at full power without entering an instability? Will the coolant remain chemically stable at operating temperatures? How does the fuel deform when it absorbs fission products for years? Will the cladding resist corrosion, swelling and creep? How’s the neutron moderation affected by material creep? Do the control systems behave sensibly when the reactor is hot and under load rather than cold and quiet? Can the supply chain fabricate the components with the tolerances required? Will the regulators accept the safety case? Will the plant integrate into the grid that exists rather than the grid that is hoped for?
Zero power experiments do not illuminate these issues. They help the neutrons behave correctly on paper and in practice. Everything else comes later.
This distinction is the same one that motivated my earlier piece titled Critical Not Connected. A reactor can reach criticality and still sit idle for reasons that have nothing to do with physics. Grid integration, regulatory delays, supply chain bottlenecks, financing structures and the political conditions that surround siting decisions often matter more than neutron behavior. A reactor that is technically operational but economically or institutionally stranded is no more useful to a nation than a reactor that exists only in a code. In that essay I argued that nuclear success depends on the system around the reactor rather than the criticality event itself.
The same logic applies at the opposite end of the spectrum. Zero power experiments reach criticality in a narrow and carefully controlled sense, but they do not cross the threshold into engineering. If Critical Not Connected was about reactors that operate without connecting to the world, this piece is about experiments that connect to physics but not as much to the world of reactor systems. They are essential for science but less relevant to power production until many years of engineering turn neutrons into heat and heat into electricity.
None of this diminishes the value of the teams doing zero power work. They are the inheritors of a tradition that began with Chicago Pile 1 and matured through the ZPR series and ZPPR in Idaho. Those facilities produced the benchmark measurements that still anchor fast reactor physics today. Without them, our models would wander and our safety margins would rest on guesswork. The people who run today’s benchmark experiments (big thanks to LANL’s NCERC—we have worked with that facility for years) maintain skills that only a few nations still possess. They ensure that the nuclear field remains tied to measurement rather than drifting into simulation driven optimism.
But clarity matters. Splitting atoms is routine for people trained in this field. A subcritical pile does it quite well actually. A zero power assembly does it with slightly greater formality. Neither activity resembles the challenge of running a power reactor. The milestone that counts is not the appearance of neutrons arranged in a stable pattern. The milestone that counts is the creation of a system that can take those neutrons and produce heat in a stable and predictable way. The path from one to the other is long, technical and filled with work (and a lot of paperwork!) that never fits into a headline.
Physics is where nuclear begins. Engineering is where it becomes useful. Confusing the two is how we end up believing that a fusion shot signals a power plant or that a zero power experiment heralds the arrival of a reactor. Nuclear progress demands a different kind of literacy, one that recognizes that the hardest parts are not just producing the neutrons but everything that must surround them.
Appendix: How Zero-Power Physics Has Actually Worked Throughout Nuclear History
The confusion between zero-power physics and power reactors is not new. It has followed the field since its earliest days, and history offers several clear cases where benchmark experiments played a crucial role without being mistaken for power-production milestones. Understanding this lineage helps clarify how today’s zero-power work fits into a much longer intellectual tradition.
1. Chicago Pile-1 (1942)
CP-1 was the world’s first controlled, sustained fission chain reaction, built under the west stands of Stagg Field. It went critical on December 2 and remained at essentially zero power. The entire structure was a hand-stacked graphite pile with natural uranium metal and oxide pieces. There were no cooling systems, no pressure containment structures, and no intention of generating heat.
CP-1 demonstrated one thing: that a reactor could be made self-sustaining with the known cross sections and neutron moderation physics of the era. It was a physics proof, not an engineering prototype. CP-1 didn’t teach anyone about heat removal or fuel lifetime. It told us almost nothing about the materials performance problems that would later hunt commercial reactors. Its purpose was to validate neutronics at the most fundamental level, not to produce meaningful power.
CP-1 is the spiritual ancestor of today’s zero-power assemblies. It answered the physics question, then quietly stepped aside for real engineering to begin. Image below shows students in my research lab loading counting foils into the Georgia Tech subcritical pile to map neutron flux to calculate materials buckling.
2. The ZPR Series at Argonne (1950s–1990s)
The Zero Power Reactor (ZPR) series were the backbone of U.S. fast-reactor physics for nearly four decades. These facilities allowed researchers to build full-scale, room-temperature mock-ups of fast reactor cores using precisely machined plates of uranium, plutonium, sodium, steel, and other structural materials. The goal was to measure (and I’m only including some examples):
• fission rates and spectra • control rod worths • sodium void effects • core leakage • reflector behavior • and the subtle feedbacks that underpin fast-reactor dynamics
ZPR assemblies were not reactors. They produced negligible heat and required no engineered cooling systems. Their physics was invaluable, but their engineering relevance stopped at the neutron level. You could run ZPR experiments all year and still not know how a fast reactor fuel pin deforms at 650°C, how sodium behaves under thermal transients, or how cladding responds to swelling and creep.
ZPR defined what zero-power physics is supposed to do: quantify the neutrons in a representative geometric and materials configuration and leave the rest to actual reactors.
3. ZPPR and the Era of Benchmark Global Cooperation
The Zero Power Physics Reactor (ZPPR) succeeded ZPR at Idaho and became a cornerstone of U.S.-international collaboration on fast reactor physics. Nations building sodium-cooled reactors sent teams to Idaho to participate in benchmark campaigns that refined data libraries and resolved cross-section discrepancies. Many modern codes still rely on ZPPR measurement datasets to reduce uncertainties for fast-spectrum modeling. One of my early experiences at national labs was participating on similar a benchmark (not ZPPR) during a summer internship at Argonne National Lab — using zero power physics to benchmark various codes around the world.
Again, the pattern holds. ZPPR validated core physics and compared how various codes treat reactor physics. It did not demonstrate thermal-hydraulic performance, fuel reliability, fabrication feasibility, or operating behavior. The reactor engineers who designed EBR-II or the later GE PRISM concepts depended on both types of information. From ANL’s description:
Argonne had a very large inventory of critical facility materials that could be assembled in various combinations to construct any reactor within just a few weeks. Because of the very low reactor power, the materials became only slightly radioactive, and could therefore be used again and again. This feature, combined with the short time required to assemble a core and the absence of a cooling system, meant that nuclear reactors could be built and tested in ZPPR for about 0.1% of the capital cost of construction of the whole power plant.
4. The Modern Echo: Zero-Power Today
Today’s zero-power experiments, where they still exist, remain crucial for validating codes, quantifying uncertainties, and building confidence in new reactor designs directly following the lineage of CP-1, ZPR, and ZPPR. The historical separation, however, is clear: CP-1 did not solve the engineering for Watts Bar, just as ZPR did not enable thirty years of EBR-II operation or qualify fast-reactor cladding and fabrication protocols.
Understanding this history doesn’t diminish the achievements of today’s experimenters. It gives their work the correct context. They are part of an eighty-year chain of experimentalists whose data define the neutron world the engineers must later tame.
The neutrons are the beginning of the story. The reactors come later.
Hear, hear. I grow tired of salivating news stories that fundamentally either, a) neglect the years or decades of engineering required to bring it to practical fruition, or b) turn a truly minor non-event into something worthy of full-blown public panic, so long as it denigrates nuclear.
Hear, hear. I grow tired of salivating news stories that fundamentally either, a) neglect the years or decades of engineering required to bring it to practical fruition, or b) turn a truly minor non-event into something worthy of full-blown public panic, so long as it denigrates nuclear.