Published on 2026-07-25
What is a preprint, and why does 'not yet reviewed' matter?
A superconductor that would have changed the world made headlines in 2023. It was a preprint, not a finished study, and the difference turned out to matter enormously.
In July 2023, two Korean researchers, Sukbae Lee and Ji-Hoon Kim, posted a paper online claiming they had built a material called LK-99 that conducted electricity with zero resistance, at room temperature, at normal pressure. If true, it would have been one of the biggest breakthroughs in physics in a century: cheaper power grids, lossless batteries, maybe even different trains. Within days, the claim was everywhere. Videos of a small grey chip half-levitating above a magnet spread across social feeds. Physicists who normally argue about the small print were suddenly all talking about the same object.
Here's the detail that got lost in the excitement: that paper had not been peer reviewed. It was a preprint, posted on a server called arXiv where researchers can share results instantly, before any outside expert has checked them. Nobody had confirmed the recipe worked. Nobody outside the original team had even tried.
What 'peer reviewed' actually buys you
A preprint is exactly what it sounds like: a manuscript shared publicly before it goes through peer review, according to a research guide from the University of Minnesota Duluth library. Peer review is the stage where independent experts in the same field read a study, poke at its methods, ask for missing details, and sometimes reject it outright. It's slow, occasionally frustrating for the researchers involved, and far from perfect. But it's also the point where a claim gets tested by people with no stake in it being true.
A preprint skips that gate entirely, by design. That's not automatically dishonest: preprints let researchers share findings fast, get early feedback from the whole field at once, and stake a claim to an idea before a slow publishing process finishes. Whole fields, especially physics and computer science, run partly on preprints, and plenty of preprints go on to be confirmed, refined and published essentially unchanged. The catch is that 'shared fast' and 'checked thoroughly' pull in opposite directions, and LK-99 landed exactly at that tension point: a claim big enough to reshape entire industries, sitting on a foundation nobody outside the original team had yet tested.
What happened once other labs tried
Within weeks, labs on multiple continents tried to reproduce the recipe. Teams in India and at Beihang University in Beijing made samples, but got ordinary semiconductor behaviour instead of superconductivity, according to Physics World's reporting on the replication attempts. A team at Princeton found the synthesis actually produced a mix of several different materials, not the single clean substance the original paper described. The instructions themselves were missing basic details, like exact equipment dimensions and the fact that the mixture turns molten partway through.
Then came the explanation nobody wanted but everybody needed. Researchers Wei Wu and Jianlin Luo, working at Beijing's National Laboratory for Condensed Matter Physics, showed that a common impurity, copper sulfide, undergoes its own structural shift at almost exactly the temperature where the original team had seen their dramatic resistance drop. One physicist called it the final nail in the coffin: a mundane impurity explained the spectacular result far better than a once-in-a-century discovery did.
The chip levitated. The claim did not survive contact with anyone else's lab.
The system worked exactly as intended
It's tempting to read the LK-99 story as science failing. Read it the other way and it's closer to a success story, just a slower and messier one than a single dramatic headline suggests. A striking claim appeared, spread fast because it was posted openly, got tested by dozens of independent groups within weeks, and was set aside once the evidence stopped supporting it. Nobody had to trust the original authors' word for it. The correction came from the same openness that let the claim spread in the first place.
That's the habit worth carrying into anything described as a preprint, a preliminary result, or 'not yet published'. It doesn't mean the finding is false. It means the finding hasn't been through the stage where other experts get to poke holes in it. Waiting for that stage isn't cynicism. It's just reading the label on the tin before deciding what's inside it.
None of this asks you to become suspicious of every early result you come across. It only asks for one small habit: when a headline mentions a study, glance for the word. 'Preprint' or 'not yet peer reviewed' means treat this as a lead, worth watching, not yet a settled fact. 'Published in' a named journal means other specialists have already had a chance to argue with it. The words are usually right there in the article, and they change what the rest of the sentence is allowed to mean.
Sources
- Room temperature superconductor LK-99 fails replication tests (Physics World) (physicsworld.com)
- Preprint vs postprint definitions (University of Minnesota Duluth Library Guides) (libguides.d.umn.edu)