First, more fun re Ranga Dias. Fresh off renewed controversy about claims of room temperature superconductivity in Lu-N-H at high pressures (claims of reproduction of the effect seem unreliable to me), it’s come out that this paper, already under an “expression of concern”, is being retracted. This has been widely reported - see here (hurray for student journalism) and here and here for example. It is abundantly clear that data fabrication/copying has taken place here. Between this, the other retraction, and the clear evidence of thesis content plagiarism, it’s hard to see any signs of credibility remaining.
Then there is the claim via preprints (here, here) of room temperature superconductivity at ambient pressure in a lead oxide compound from investigators in Korea. Cutting to the chase: it is very unlikely, in my view, that this pans out, for multiple reasons. Extraordinary claims hardly ever hold up. There are multiple weird issues with the data in the figures (e.g., magnetic susceptibility data that shows up in both papers with the same units but axes that differ in magnitude by a factor of 7000 - which numbers are reliable, if either? Resistivity that seem bizarrely large (0.01 Ohm-cm is bigger than the Mott-Ioffe-Regel limit - again, are the units right?). A specific heat that doesn’t reach 3R at high temperatures. Not clear of the resistance is really zero in the nominally superconducting part of the V-I curves.). That said, if the video and the not-crazy-scale susceptibility data are trustworthy, this stuff is very diamagnetic, more so than graphite, which is quite unusual. At least the authors do provide a comparatively straightforward synthesis recipe, so replication attempts should clear this up in a week or two.
None of this is thaaaaaaat unusual, by the way. There are claims of weird superconductivity at some rate. It’s easy to screw up measurements, especially in inhomogeneous materials. Unfortunately, social media (esp the site formerly known as twitter) drastically amplifies this stuff. I assume Michio Kaku is going to be on tv any second now talking about how this will change the world. Hopefully responsible journalists will be effective at pointing out that a non-reviewed preprint on the arxiv is not conclusive.
I’m traveling, so updates will be sparse, but I will try to keep an eye on the comments.
52 comments:
You were too modest to mention this (or maybe you didn't want to draw attention to it...), but you have been directly quoted in some of the journalism reporting this. See, e.g., https://www.scientificamerican.com/article/viral-new-superconductivity-claims-leave-many-scientists-skeptical/
You might add La3Ni2O7 to the list of superconducting discoveries (looks much more convincing too). T_c up to 80 K under pressure, in a very well known material.
Perovskites rock ;-P
https://www.nature.com/articles/s41586-023-06408-7
There is a paper on LK-99 published past March in a Korean scientific journal (in Korean): Sukbae Lee, Jihoon Kim, …, Keun Ho Auh, «Consideration for the development of room-temperature ambient-pressure superconductor (LK-99),» Journal of the Korean Crystal Growth and Crystal Technology 33: 61-70 (30 Mar 2023), doi: https://doi.org/10.6111/JKCGCT.2023.33.2.061 [Google Translate PDF https://www.docdroid.net/UiUrs8c/kci-fi002955269-1-pdf].
I suspect the authors of the LK-99 paper just didn't do a thorough enough analysis and critique of their own work. I would conjecture that there are major flaws in their measurement setups and analysis; they probably just have some high-conductivity diamagnetic material as is the case for pyrolytic graphite. For this reason, I think the community (i.e., scientists) should remain calm and carry on if the results are not reproducible. We should have a positive attitude towards these researchers and thank them for sharing the preprint. This all assumes they were not dishonest or manipulated data. I would have also preferred a less grandiose title, maybe something like "Intriguing new finding in the field of superconductive materials".
In contrast, R.P. Dias is turning out to be a fraudulent and dishonest researcher on multiple occasions as @Doug points out. I wonder how his institution will handle this, would he still get tenure? I also think arXiv and others should make a public apology to J. Hirsch. He has taken the time to investigate the findings of Salmat and Dias and communicated his concerns but institutions have treated him like he is just trolling Dias.
Completely agree with the above comment. ArXiv appear to have implemented a blanket ban on Hirsch. His recent preprint on percolation effects (which likely explains Dias’ results) never appeared, and came out elsewhere.
Whatever happened in the past, something stinks today.
I agree; the Korean paper may (likely) have wrong conclusions, and the authors may be less expert in the field than we are used to, but they'll learn from this. No indication (so far) of any nefarious aspects to that paper, even if I am fairly convinced its conclusions won't be confirmed.
I also note they did not seem to have targeted a lot of publicity, as Dias did.
Love the Kaku reference. He really just rides a hype train wherever it will go.
@Maxwell's Dog
Just read Dias' Wikipedia page:
https://en.wikipedia.org/wiki/Ranga_P._Dias
It is not clear how this guy can even teach, never mind research, with any credibility, When are Rochester going to act? Their reputation is trashed...
@Anon 2:30 AM
It's likely that the University of Rochester is content because he has some significant chunk of funding coming from NSF/DOE/DOD. Hopefully, his newly developed reputation prevents Dias from getting any of these precious dollars in the future. Most universities are administrative heavy and just care about treating thier institution like a business.
I wish the NSF/DOE chokes off *all* funding to the Univ of Rochester for failing to act against this persistent and deliberate fraud. You'll be surprised to see how quickly the University suddenly develops a set of scruples and principles.
I sympathize with your anger, but I don’t think it’s fair to punish all the other PIs at U Rochester like that, just because they happen to be at the same institution as this fraudster. Some punitive consequences are definitely needed, but they should be directing at the corrupt administrative upper levels, not at any innocent bystander faculty.
I.e. overhead on PI grants will increase to pay off the upper level punitive...
Is there any way NSF or DOE can place limits of caps on how much overhead URochester’s allowed to collect?
I am guessing most departments would have internal red teams to cross-check at least big discoveries if not regular claims. Isn't it a standard practice for PIs or other lab members to independently check/verify the set-up/device-measurements and confirm them before its shipped of to a big-time journal?
Also, wouldn't UoR have assembled a faculty committee to check these claims?
@PPP, indirect cost rates are set by complicated negotiations btw the university and the feds, and they are not renegotiated very often or easy to modify. Remember, the fed agencies have their own offices of research integrity or equivalent, and there are procedures that govern how universities conduct misconduct investigations and how they report findings. All of those procedures have ample opportunities for the investigated parties to respond to allegations and appeal findings. Nothing is fast.
@anon@12:35, in general, no, there is no internal red teaming at universities, certainly not on a formal level. PIs are free agents. I don’t have to run my papers by anyone before submission. Now, a lot of people do have colleagues read draft manuscripts of big results to get feedback, because that likely does result in better-written, stronger manuscripts, but it’s all on the PI to decide how to proceed. I assume Rochester has and/or had investigative panels to look at these issues. However, those panels are not intended to actually double-check the measurements- they’re to see whether misconduct has taken place. (BTW, the federal definition of misconduct is rather narrow - fabrication, falsification, or plagiarism in the presentation of research results. There is all kinds of poor practice and behavior that doesn’t fit neatly in there, which is why a lot of universities have broader policies about research integrity.)
Doug, thanks for the information, very enlightening.
A follow-up, possibly naive, question: if indirect cost rates are so set and difficult to change, then wouldn’t Rochester itself not be able to increase overhead on PI grants as a way of “paying off the upper level punitive”, as Anon@4:17PM suggested would happen?
In particular, the professor from a reputed university who is the moderator in the superconductivity section should apologize to Hirsch. Everyone knows how he promoted Dias’s work in his Linkedin. Also trashed all the works in quantum information in his most recent FB.
You only have to watch the ‘redmatter’ YouTube video to know that something is seriously wrong at Rochester.
It features the Dean (in charge of academic misconduct) and head of department waxing lyrical about how the discovery will change the university and make it a national centre. Not to mention the PI of the major DoE grant that Dias used (‘Rip’ Collins). That video was made, and released, after they were aware of problems with e.g. the CSH paper.
The sad part is you don't even need to understand Korean to get the fist of things.
I apologize for the newbie question, but I couldn't find scales on which exact figures differ by a factor of 7000?
In one paper they specified e-4 on axis, while in the other mentioned that their units are *e-4. I am talking about fig 1 and fig 4 respectively.
What am I missing?
@anon@7:20, they have posted a revised version of one of the papers, and they’ve changed the units. Here is the original version: https://arxiv.org/pdf/2307.12037v1.pdf
Do you know what they mean by this comment, referring to their I-V measurements in Fig. 6? It's being used to explain their non-zero resistivity at very low currents, and as far as I'm aware, Joule heating shouldn't occur in the superconducting state, and what breaks the superconductive state at a critical current density is the magnetic fields created by that current?
"In regions I, J, and K, we observe an increase in resistance, indicating
the breaking of the superconducting energy gap due to Joule heating when the current exceeds a certain threshold."
The authors also state in response to PubPeer comments something similar, i.e. "We say on the basis of experimental result. Q2: a high current flows to the sample, Joule heat occurs, then Superconductivity breaks, Then Gap(Superconductor) to nogap(Metal) transistion occurs." (see https://pubpeer.com/publications/82AF24DF343C16341B42CF36B22FA4#16).
There's a recent preprint from a group in China analyzing electronic structure of LK-99 structure with Au and Ag: https://arxiv.org/pdf/2307.16040.pdf. Anyone familiar with these kind of calculations for understanding superconductivity?
There is a preprint from India, stating that they have failed to see any superconductivity, but they'll keep trying. https://arxiv.org/abs/2307.16402
There are also some interesting DFT calculations. Basically, the author states that to get SC one needs to be careful with Pb substitution since it is not the lowest energy site that we want.
https://arxiv.org/abs/2307.16892
What makes me confused is that I thought that to run DFT calcs you need to know the lattice structure quite well. While here it seems that the original sample was quite messy (an impression that I got from reading chemists' comments). Does it mean that we might get some useful insights if someone manages to synthesize a pure sample with correct substitutions? Or does it show that we don't know anything because it is not quite related to the original sample?
I think they do use a lattice structure. They appear to know the lattice configuration of the parent (lead apatite) material, and replace some of the lead atoms with Cu based on a previous crystal structure. They found that the Cu doping converts the parent insulator to a metal, and that including electron-electron repulsion (the Hubbard "U") does not alter the basic computed electronic structure of LK-99.
I dont know how this pertains to the presence of a superconducting transition at room temperature though. Maybe not having such a structure implies that a superconducting transition? To my, very out of date knowledge, the electronic structure calculations are otherwise pretty standard (PBE density functionals with VASP's augmented plane wave pseudopotentials).
Sorry the above line *not having such a structure...* is incomplete and should be read as "maybe not having such a structure implies that a superconducting transition is unlikely".
The Wall Street Journal is in on the act, even including a quote from our host on this blog:
“ “We’re absolutely still waiting for confirmation that superconductivity at room temperature exists,” Douglas Natelson, a condensed-matter physicist at Rice University in Houston, said of the claims”
https://www.wsj.com/articles/room-temperature-superconductivity-plagiarism-df591741
Well here is another ab-initio calculation preprint from a researcher from LBNL indicating LK-99 indicating he thinks this might be it; I'm still doubtful. Also these are ground-state calculations so how do you account for room temperature and pressure? Here is the authors concluding statement:
"the calculations presented here suggest that
Cu substitution on the appropriate (Pb(1)) site displays
many key characteristics for high-TC superconductivity,
namely a particularly flat isolated d-manifold, and the
potential presence of fluctuating magnetism, charge and
phonons."
The following sentence after that one does indicate some potential challenges with synthesis because of lower formation energy of Pb(2).
Wondering if the author is trying to cement their role in the theory prediction side of things, i.e., 3rd Nobel prize winner, lol.
https://arxiv.org/pdf/2307.16892.pdf
From a solid state chemistry point of view, it is nuts to imagine Cu2+ substitution for Pb2+. The coordination requirements of these cations are incredibly unlike, due to their size, a Jahn-Teller distortion for Cu2+, a lone-pair for Pb2+ etc etc…
@Xirtam@3:28 That paper was also mentioned in the comment by Daniil Kudrin above.
A colleague also pointed out an experimental preprint from China: https://arxiv.org/abs/2307.16802 They synthesize the material and claim to find semiconducting transport rather than superconductivity. But I'm not sure what to make of it, as they also don't find the same strong diamagnetism as in the original preprints.
@Anon @4:20PM Thanks for this insight.
It also isn't that surprising to see flat bands with the proposed doped structure. The coppers are many angstrom apart, so localized states are expected.
@Anon @4:20PM & @Anonymous @8:49PM. Thanks! I find the DFT calculations to be misleading in regards to the room temperature Tc claims, since I dont understand how the band structure they described has much to do with specific value of the critical temperatures.
In my opinion, visualizing vortices using Bitter decoration (e.g. https://doi.org/10.1103/PhysRevB.67.092512) would be a much more convincing magnetic demonstration of (type-II) superconductivity than any "levitation" experiment.
Oh God. Must. Not. Respond. To. Berkeley. Tweet:
https://twitter.com/BerkeleyLab/status/1686517668290174976
Chemically and physically naive results, unreviewed, and unpublished. Yet they go with the 'mic drop' meme...
I replied. They hid my comment (I just said that it was improper for a national facility to be communicating science this way).
I replied with some questions to the author of the DFT paper:
1. Handling e-e corr. effects: HTSC exhibit unconventional pairing mech. Does DFT+U work? Don't you need GW or DMFT?
2. DFT is GS, e-e corr. effects are dynamical, don't you need TD-DFT?
3. DFT handles SC gap and Eliashberg theory?
No reply. Can anyone explain how a single-particle theory (DFT) with localization treated by a Hubbard like model(+U) can be used to understand superconductivity? I get that DFT is exact if one has the exact exchange-correlation functional, but my reading seems all used are approximation and very deficient for true quantum many-body systems. Also this is a ground-state theory and what about phonons? Seems so much is missing? Maybe I'm missing something.
@Xirtam Esrevni,
I don’t think you can really, and agree with your point about the need for GW or DMFT methods. DFT is a ground state theory, and does not include electron-phonon couplings in its native form. I also don’t think it would be valid near a superconducting phase transition. The paper linked above is a straightforward gs calculation so, like you, I am mystified by the connections being made to room temperature superconductivity, and maybe I am missing something too.
People here are missing the point: not one of the now order 10 DFT papers claims anything about superconductivity. Not at all. No mechanism is proposed but some speculations are made. For example maybe something exotic occurs due to flat bands and a high DOS at Ef. These papers are pretty up front about this: they find some unusual properties at the band structure level which itself is intrinsically interesting but alone implies nothing much wrt superconducting behavior. To me band this narrow and localized with a double degeneracy at Ef should be deep in the Mott insulating phase and with JT distortion and super exchange should by AF insulators. This is at odds with what the papers claim…but none of them claim superconductivity as far as I can see…
@Anon 12:18 PM. How useful are speculations about 'something exotic caused by flat bands and a high DOS at Ef'? Surely these are so generic, as to be obvious?
And has the argument about the bands being flat because they are impurity bands (Anon @8:49 PM) been addressed?
I would say having a small number such flatbands isolated from the remaining bands is not very common nor is having d9 copper with 2 degenerate orbitals at Ef that somehow don’t distort (this I don’t understand) is uncommon in materials of this type
But I wouldn’t say more and I would expect this material to be insulating…
News from the other high-Tc game.... Ashkan Salamat of UNLV no longer appears as a co-funder of Unearthly Materials with Ranga Dias... The company founded to exploit their 'discoveries' in high pressure research.
https://www.unearthlymaterials.com/
What can it all mean?!
@Anon 1:38
Surely it means he got a job at Q-centre, Inc. working on LK-99
In my view studies of the (ground state) electronic structure.are certainly useful in a system where strange claims are being made. Same as when you have an arpes system; you'd be studying what you can if you had a sample.
Do these studies reveal the answer to all questions? No. But are they a prerequisite for understanding some answers? Yes. (Leaving the question of the stoichiometry and struck/symmetry alone...)
It depends on what he's trying to superconduct, hot or cold electrons? Present day categorizations of superconductivity per location per time have not been classified as being per location per time. The state that the election is in, is critical... Who knows life may in fact be on an electron!
I'm not familiar with Elsevier's "Journal of Materials Science & Technology", but it appears to have very fast referees:
First-principles study on the electronic structure of Pb10−xCux(PO4)6O(x=0,1)
Received date: 1 August 2023
Accepted date: 1 August 2023
https://doi.org/10.1016/j.jmst.2023.08.001
Preprint version was first submitted July 29: https://arxiv.org/abs/2307.16040
@Anonymus 12:48pm
It is doubtful that any of these DFT papers will be useful at all, because, as they write above, these are all ground state calculations at the mean-field e-e level, no phonons etc. and provide very little explanation.
They just do a simple calculation fast and throw some speculation to justify why you would do such paper to jump on the bandwagon of the hype, if it turns to be SC then they will be highly cited because they were the first to do it; if it is not, then it is just another bandstructure paper (nothing bad about it per se) that will be go to the pile of forgotten papers rather quickly.
An interesting accepted paper in PRL:
https://journals.aps.org/prl/accepted/5d07cYcbU43R9e1b00057383259ac34505b5a3f54
Particularly notable for its author list, which does not include Dias as accepting the retraction…
The plot thickens… the above link for the accepted retraction no longer works, and Salamat reappeared at Unearthly Materials…
I say we take off and nuke the entire site from orbit. It’s the only way to be sure...
The PRL retraction has been posted! https://doi.org/10.1103/PhysRevLett.131.079902
It does not provide many details, but Ranga Dias is not an author of it as "Of the authors on the original paper, R. Dias stands by the data in Fig. 1(b) and does not agree to retract the Letter."
Post a Comment