[foven]

Agreed, this looks bogus. Some suspicious points:

In the first paper, they claim to measure zero resistance (on a scale of microvolts), but are very careful not to show full RvT curves - in the second paper, we can still see significant changes below Tc where they include more complete curves. How can the resistance change significantly in the superconducting (zero resistance) state? We can actually see significant noise in paper 1 fig. 1c in the ohmic state and it even appears to behave as an insulator at 0 field (increasing resistance with decreasing temperature), but a metal with applied field. There's something wrong with the measurement.

400 K is an odd choice for your superconducting temperature, and just so happens to be the top end of what an MPMS system can measure so is not completely random. Surely it makes sense to measure significantly above this with one of the oven attachments, verify these results with collaborators at other labs even.

10 Gauss is an extremely small field to use for a ZFC-FC measurement and again if their superconducting Tc is at or above 400K they need higher temperature data to show anything about the phase transition.

The claim that they have measured the density of states is completely unjustified - not even a citation. I don't know how you can believe that to be the case.

And in general the presentation both of the data and the paper itself is poor - if you just made a groundbreaking discovery like this, wouldn't you care?

[zarzavat]

> if you just made a groundbreaking discovery like this, wouldn't you care?

Hell no! If I had made a discovery of similar magnitude I would have done exactly what they’ve done: push out a rough preprint ASAP to reserve my Nobel prize, then take a deep breath, relax and take my time dotting ‘i’s and crossing ‘t’s for the real paper in Nature.

That doesn’t mean they’re correct, but there’s nothing inherently suspicious about the way this has unfolded.

[stu255]

Exactly, they have made a very big claim and made it very easy to replicate / falsify. It takes a few days to produce this stuff.

Their paper is weak on data / results.

This is exactly what you would do if your team genuinely believed you had discovered something monumental.

In poker terms they are "all in" and they want to get called.

That's why it is so interesting. If they had posted lots of extreme results but it needed $10m to replicate then I would be thinking "fraud". It would look like a bluff.

[foven]

As I mentioned in my above post, they have really dodgy data. Ideally, with something like this, you would have collaborators to verify alongside you as joint co-authors. I think something people underestimate if how easy to replicate samples are - crystal growth is difficult, and impurities are important. It is unlikely anyone will produce exactly the same sample only something close based on the process they've given.

In realistic terms it seems they're grabbing for the prestige without the foundation of crossing their ts. Bad science like this shouldn't be encouraged. It's likely there's not very many groups growing the same material system so they have the time to spare. A paper like this wouldn't be on the arxiv at all if they were 100% sure because they would go straight for the nature publication and take the time to do more follow-up papers while they can.

Edit: to be clear as well, a lot of people are underestimating the time it takes to reproduce a growth even with a manuscript telling you how to do it. People always leave out steps and oversimplify. There is a lot of extra characterization that takes time to double check you have the right material that lines up with what they have here. Only the direct competitors actually already growing this material can do it in a few days.

[peteradio]

I think the politicking of the situation could explain why there was minimal involvement of outside collaborators.

[meroes]

The motivations for fakery are too high.

[nazgulnarsil]

excellent point that the noise should raise probability of measurement/set-up/equipment issues