I ended up spending more time catching up with people this afternoon than going to talks after my session ended, but here are a couple of highlights:
- There was an invited session about the metal halide perovskites, and there were some interesting talks. My faculty colleague Aditya Mohite gave a nice presentation about the really surprising effects that light exposure has on the lattice structure of these materials. One specific example: under illumination, some of the 2D perovskite materials contract considerably, as has been seen by doing in situ x-ray diffraction on these structures. This contraction leads to a readily measured increase in electron mobility and solar cell performance. Moreover, the diffraction patterns show that some diffraction spots actually grow and get sharper under illumination. This kind of improved ordering shows that this is not just some sort of weird heating effect.
- In a session about imaging, I caught an excellent talk by Masaru Kuno, who described his spectroscopic infrared photothermal heterodyne imaging. The idea is elegant, if you have access to the right light source. Use a tunable mid-IR laser that can go across the "fingerprint region" of photon energies to illuminate the sample in a time-modulated way. If there is an absorptive mode (vibrational in a molecule, or plasmonic in a metal) there, the heating will cause a time-modulated change in the local index of refraction, which is then detected using a visible probe beam and a lock-in amplifier. It was an extremely clear, pedagogical talk.
- I spent much of my time in the strange metal session where I spoke. There were some very good (though rather technical) theory talks, trying to understand the origins of strange metallicity and key issues like the role of disorder.
I had wanted to attend the session about superconductivity measurements in materials at high pressures, because of the recent and ongoing controversies. However, the room was small and so packed that the fire marshal was turning people away all afternoon. I gather that it was quite an eventful session. If one of my readers was there and would like to summarize in the comments, I'd be grateful.
(BTW, it seems like this year there have been two real steps backwards in the meeting. The official app, I am told, is painful, and for the first time in several years, the aps wifi in the meeting venue is unreliable to the point of being unusable. Not great.)
It is said that the chair allowed no audience questions for the superconductivity session. Anyway, embargo lifts today. Looks like there will be some ‘interesting’ press perspectives on the whole saga.
ReplyDeleteI found a video. 1 GPa is small compared with the pressure I always feel as a leader of a big grant, so it is a normal pressure :-)
ReplyDeleteAnyway, JHS behaved like him and gave presentations in this way. They look so similar… But I cannot tell whether this is true or not…
https://youtu.be/XhhvOMuLF94
Nearly six minutes of introduction...Anyway...
ReplyDeleteLutetium hydride is known to have a residual resistivity of about 0.23 uOhm.cm. That is lower than ReO3. [J N Daou et al 1984 J. Phys. F: Met. Phys. 14 2983]. It is also known that these hydrides show metal-insulator transitions near RT, and under small stimuli (such as H2 partial pressure!). Note that their measurements seem to have H2/N2 as a pressure transmitting medium. See e.g. Huiberts, J. N., et al. "Yttrium and lanthanum hydride films with switchable optical properties." Nature 380.6571 (1996): 231-234.
In their paper, they say that " In some cases, small residual resistance from the instrument offsets was subtracted from the measured voltage." when doing resistivity measurements.
When they talk about resistivity under field (SI), which shows nice zero resistance at low temperature of course, they say "The temperature dependence of the resistance of a simple metal is written as: R(T) = Ro + aT2 + bT5. We fit the data below T < 220 K for each field, at which the resistance goes to the minimum value, to that function and subtracted it out."
The ac- susceptibility data are fit with a random background function to yield a negative result at low temperature. At some pressures, this has a positive slope, at others a negative slope... etc etc
Who do Nature get to review this stuff? There is a much bigger picture behind all this of course, but I personally blame the journal the most. To think that Bragg, Chadwick etc all published there. This is the ultimate end result of the Springer buy out IMO.
Oh yeah, and their 'superconducting' phase transition shows a ~10 K hysteresis on heating and cooling. Ffs...
ReplyDeleteKuno's work sounds interesting and similar to some recent work by Bumberg in Cambridge, UK.
ReplyDeleteThe talk is bizarre. The guy claims room temperature superconductivity, and in 10 min presentation to a specialized audience spends part of the time talking about superconducting trains and LBCO.
ReplyDeleteMaybe that was the investor version of the talk? More here, where they claim to have raised 20 M$:
ReplyDeletehttps://www.youtube.com/watch?v=vFXoFYx2YxI&t=2580s
This video was taken down. If anyone can please reupload the video, it would help give this discussion more context.
DeleteYou might want to put a hyphen in "strange metal session"
ReplyDeleteRe the app: Actually seems to function for me. It's not lightning quick, but it functions well enough to be usable (putting together a schedule, looking up other sessions to go to while you're thinking about leaving the one you're currently in, etc.). Wifi on my laptop however less so, as it takes several minutes to connect.
ReplyDeleteMy main complaint is that the screens in the rooms (as well as the venue in general) are too small, making them quite hard to read unless you are in the first few rows.
@Don: Convention seems to be that "strange metal" isn't hyphenated.
Thanks, everyone, for the perspectives on the session. Here is a video of the talk: https://www.youtube.com/watch?v=XhhvOMuLF94
ReplyDeleteI'd be very curious to see Hirsch's end-of-session talk, too.
@Don, are you thinking of this? https://xkcd.com/37/
For what it is worth, the embargo period is now over for the superconductor paper...
ReplyDeleteOh my goodness... I just went and plotted their data from supplementary figure 15, where they say that "The temperature dependence of the resistance of a simple metal is written as: R(T) = Ro + aT2 + bT5. We fit the data below T < 220 K for each field, at which the resistance goes to the minimum value, to that function and subtracted it out."
ReplyDeleteIt looks like rho(t) for e.g. BaFe2As2, i.e. with a broad drop, below which the data never even get close to zero resistance. They have subtracted off a GIGANTIC contribution to get 'zero resistance'.
I ask again, who reviews this stuff, and what are they smoking?
The New York Times just reported this as a breakthrough discovery. If and when the results fail to hold up under scrutiny, the credibility of scientists all around will be greatly harmed. Very frustrating.
ReplyDeleteI'm going to drown my sorrows with a beer (EU time!). What a time to be alive as a scientist...
ReplyDeleteThe resistance data from the SI...
ReplyDeletehttps://imgur.com/a/bm1gUnM
I wish I had learned about subtracting backgrounds from resistivity earlier, my PhD thesis would have been great!
Why would you not put your university as your address on a world changing result? Are UNLV not keen on superconductivity anymore?
ReplyDeleteUnearthly Materials Inc., Rochester, NY, USA
Keith V. Lawler & Ashkan Salamat
A week ago, I attended a Virtual Science Forum about reproducibility in condensed matter physics. Based on analysis from James Hamlin and others, I would not trust this group's results at all. It was way too reminiscent of analysis I did with for Hendrik Schoen's superconductivity data. Of course we will see.
ReplyDeleteBy the way, Kenneth Chang, who wrote the New York Times story also wrote about the Schoen case (he emphasized Batlogg) early on in those events. The current story looks like it was an in-progress story about the ongoing controversy that was supplemented and put out to reflect the breaking news. There's a lot of discussion about the skepticism (starting in the third paragraph, which isn't awful), but I agree that it comes across a little too credulous, at least for people that only read two paragraphs.
My name is Simon Kimber, and I am a co-author on a PRL with Dias and Salamat. We received information on the 27/10/22 from Dr. Hamlin, that our high pressure resistance data on MnS2 were remarkably similar to that for GeSe4 in Dias’ thesis.
ReplyDeleteThe agreement was so striking that the next day (28/10/22) I contacted all co-authors and PRL recommending retraction. I heard nothing in response.
After this matter became public in Feb (see PubPeer), I aggressively contacted divisional editors etc. We now have a response, which in my opinion only reinforces the case for retraction.
This is a huge mess, and in my opinion, will only get worse. As you might imagine, the stress, and the silence from co-authors and APS, have been utterly devastating.
For anyone who has not seen it, James Hamlin’s talk last week. Extraordinary:
ReplyDeletehttps://m.youtube.com/watch?v=KiGgcNR2O5o
Possibly my favorite bit so far (from the Quanta piece):
ReplyDelete"In a 2021 talk organized by the Sri Lanka Association for the Advancement of Science and later posted to YouTube, Dias said “We recently raised $20 million just to, you know, focus on the science part of this. And these are the investors that we used for this kind of technology.” At that moment (43:25 in the video) a list of names appears on the screen. Under the category labeled “Investors (Series A),” the names include Sam Altman, CEO of OpenAI, and Daniel Ek, co-founder and CEO of Spotify. Following the publication of this article, a representative for Dias told Quanta that those were “aspirational statements,” that the company had not raised the money, and that the names listed were prospective investors."
The talk seems to have disappeared now, however some people might have predicted this, and downloaded it a few days ago ;-)
ReplyDelete@anonymous: how did you make your left (raw data) plot? I don't see the coefficients of the Ro + aT2 + bT5 fit in their paper
ReplyDeleteIt is just the raw data shown on the left. I have not tried it, but somebody else wrote on twitter that they also could not replicate the published data with a Ro + aT2 + bT5 fit. Download the data yourself and have a play!
ReplyDeleteSee also this plot of the raw data from Fig. 2 vs. the published result. One data set (10 kBar) seems to be just noise, however, it has the same file length as the 16 kBar data. What is it?
https://imgur.com/a/J3jfrCV
Screen shots from the (now removed) talk with details of the investors...
ReplyDeletehttps://imgur.com/a/n4HyRfi
See also:
https://imgur.com/a/cV5XnT5
He describes himself as "Discoverer of metallic hydrogen and room-temperature superconductivity, two of the top three "holy grails" of Physics".
And "mother of dragons", no wait.. not that yet!
ReplyDeleteGot it. I downloaded the raw data from the website (can't believe this is not shown in the Extended Data Figs).
ReplyDeleteThey declare they fit below 220 K, which works perfectly well, not over the whole T range. The lowest T in the available data is 150 K, so they presumably fit from 150 to 220 K, what a joke!
Pure nonsense. As you say, how can this have been accepted by referees?
I am wondering whether this is just cynism from Nature. The referee reports at all rounds should be made public. What is the credibility of Nature after this? Even less than before.
Here is a nice Twitter thread from a young scientist. They compare the atmosphere at the March meeting, to the days of cuprate discovery. It nicely illustrates how far we have all fallen:
ReplyDeletehttps://twitter.com/HongzeLi/status/1633735304787165184
Sighhhh, this whole thing is just disheartening as early career faculty. People like this get easy grant money for "groundbreaking work," despite the retractions, get to have the benefit of the doubt from editors looking for "impact," use these soon to be retracted works on grant applications, and the money spout doesn't stop (This guy was a recent NSF CAREER award winner even after multiple retractions for the same crap he applied to the CAREER award with). Then there's the rest of us not manipulating our data, doing things the right way, and get told by NSF that it isn't groundbreaking enough to get scraps (albeit I personally have been lucky).
ReplyDeleteSorry for venting in the blog comments.
What I don't understand is the fall in community standards. Ranga Dias's thesis contains large tracts of text from James Hamlin's thesis. These are publically available documents, take a look.
ReplyDeleteA few journalists have asked him about it, and Rochester are aware, but he just puts his hands up and says "gee, shucks, I made an error", and it gets let go.
What happened to a code of honor in science? What happened to some things (like a thesis) being absolutely sacrosant? Once people stop caring about these 'little' things, it is a slippery slope...
https://www.nature.com/articles/s41567-023-01970-3#article-info
ReplyDelete^Disagree on that, it is lunatics who are bad for science (and elsewhere). A bit of healthy competition never hurt anyone.
ReplyDeleteOh, and money is bad too. One thing this case highlights is that a simple "statement of conflicts" policy can never measure up against 20 M$ of funding. Not fit for purpose.
If people want to spin-out so badly, they should be encouraged to spin-out of the university and give it a real go. You shouldn't get to keep your cushy tax payer funded job too.
“Healthy” competition is the key word.
DeleteThis is one of the reasons I like doing research in a medical field - there, if your results are misleading and/or fraudulent, the consequences can be much more serious than 'merely' wasting other scientists' time and money. As a result, there's comparatively more incentive to take reproducibility and data integrity seriously. Obviously things are far from perfect, and there's a lot of shoddy work that falls through the cracks. But clinical trials and FDA approval at least provide a mechanism (however slow and inefficient it may be) for ensuring that unreliable science gets filtered out before it results in medical malpractice lawsuits.
ReplyDeleteIt is clear that Dias et al made an error when they uploaded supporting data for the resistivity curves in Fig. 2. The data set supplied for 10 kBar, is just noise.
ReplyDeleteHowever, it also has the exact same number of data points (65534) as the curve supplied for 16 kBar (which looks like the figure). Weird number too right?
Now, why should the same number of data points imply a link between data sets? Well, in another context, similarities between resistance data sets have been scoffed at by Salamat and Dias since:
"the x-component, which reflects the rate of cooling, is unique for each measurement. Measurements are done during warming cycles and the cryostat is left to warm up naturally whilst collecting e.g., resistance as a function of temperature. "
Any thoughts on what on earth this accidentally uploaded noise data set is? Also interesting that nobody seems to have plotted it during review or editing. I start to wonder if the reviewers had access to the supporting data. In my experience, Nature journals ask for it after acceptance.
The 10 kbar data also doesn't go above the temperature that they claim for the transition. And they have put it on a separate axis scale to the rest of the data, which makes it look a lot more comparable to the other pressures.
DeleteSo for me it is hard to say if it is the correct data truncated or simply incorrect data.
I think this has serious implications beyond the realms of science. First there is a foul play involved given their history of fabricated data. We should take Simon Kimber’s comment very seriously. And this paper was submitted in April 2022 with the 2020 paper still not retracted. Please note that Ranga Dias has already procured grants using the retracted papers as leverage and has claimed to have raised 20 million USD (or claimed to have been promised 20 million USD) for the startup.
ReplyDeleteTo make matters worse, they are not making their samples accessible quoting IP. If they claim to have discovered RT superconductivity, they must make sure that their results are reproducible. On the contrary, they visibly seem reluctant to leave room to that possibility. Rather they are busy fundraising. If that was the goal, why publish such fabricated, questionable data in Nature after all? And the Nature editor and the reviewers should be held accountable on this. Why publish papers from a group which has been caught red handed for fraudulent data manipulation and had their work subsequently retracted? And based on the fact that Ranga Dias won’t share his sample, hindering potential reproduction and slowing down the developments, what impact is there to be warranted for a Nature publication? I see no merit in this!
More importantly, if this secrecy is only to get the 20 million dollars using the questionable results, there should actually be a real investigation to look into the affairs. I think Dias is now behaving like Schön and it is not normal anymore.
There is a region of similar magnitude (but not identical) noise directly below the transition for the 16 kBar data (plot it on a log scale). Can these noisy features come from short circuits? That measurement does recover to yield a stable and finite resistance at the lowest temperatures, characteristic of a good metal.
ReplyDeleteIn addition, why are nearly all the data in the paper cut off below 100 K, and sometimes higher (~235 K for the 20 kBar data in Fig. 2)? A sceptic might suggest that there are other features, characteristic of a non-superconducting ground state down there...
I don't think the reviewers have dowload access to the data (although they could probably ask for it), but how can it be that they did not request the raw resistivity data and the fits to be actually shown in the Extended Data Figs ?
ReplyDeleteThe perfect hybrid of Schön and Theranos...
ReplyDeleteI'll uncloak (Simon K here), to add some of the things which were weird about the paper writing. I never had any direct access to Dias (never met the guy in fact), and he was regarded as the 'guy with the magic hands' by all. He was brought in by Salamat. He had helped out with a couple measurements on some (much) earlier work on MnS2 [I analysed all data on those].
ReplyDeleteAt some point, I received a photo of a data set on a monitor, sent to me by WhatsApp. This was claimed to show a superconducting transition for MnS2 at 25 K under pressure.
Now that would be incredibly weird right? A cubic, S=5/2 material showing such a high Tc? In addition, the normal state resistance still looked kind of quadratic. Nothing like the linear beahaviour you would expect above an unconventional superconducting transition.
I remember humming and hawing a bit, and sent the result to some theorist friends. They also were not terribly impressed. Soon enough, I received a message along the lines of 'it was a measurement fault, a short-circuit with a thermocouple'. Shortly afterwards, the data appeared which were ultimately published.
Now, it is really important that people openly share results during the preparation of a paper, and evident that mistakes will appear. Especially for something as complex as high pressure measurements. Probably it is just one of those coincidental funny stories that i'll end up spinning to impress some graduate students one day (if I ever get a job again...). But still...
The data that appeared in the paper was the same data that you got a screenshot of? The data with the measurement fault?
Delete^That comment was about our PRL obviously! SAJK.
ReplyDeleteI'd also add that, in general, it was fairly fruitful collaborating with Salamat. He did the laser heating for our first MnS2 paper, while we were post-docs at ESRF. His group later did high-pressure Raman and DFT on two other projects (MnS2 and CuMnO2). No complaints from me. It was good data from good experiments, although just finishing up old projects, and not very high impact stuff.
ReplyDeleteI don't suppose i'll be getting or sending a Christmas card this year, but those are the facts. SAJK.
I never saw the screenshot data again, which would support it being a measurement artifact spotted by whoever measured it. SAJK.
ReplyDeleteI think what worried me was more that it was considered as a reasonable result. As described above, MnS2 is about the last possible material which should show superconductivity. It is a classical 3D magnet! SAJK.
ReplyDeleteYeah implies some lack of common sense and skepticism
DeleteI played around a bit with A+BT^2 +CT^5 fits to the supplied data and subtracting (which is by itself an unjustified procedure to find a superconductor), but I could not reproduce the data in Fig. 15. In fact it seems unlikely that you could reproduce it because the slop of the raw data is greater below the transition than above. Yet you are subtracting a progressively steeper curve. But the data of Fig. 15 shows basically zero slope above and below the transition. So its not clear how their (unjustified) procedure can even give you the data in Fig. 15.
ReplyDeleteJames Analytis’s reply in Ana Akrap’s tweet about Supp Fig. 15: “While subtracting a background from a magnetization curve in a pressure cell is relatively common, I don’t know why/how you would do it for electrical resistance, which would always be shorted by the most conducting path - the superconductor.”
Deletehttps://twitter.com/profanalytis/status/1633874418354241536?s=46&t=vNVANugoxJ8gpyl7v3IvlQ
More pressing question is, why is Dias so desperate?
Bonus points to Quanta for the title on this one: https://www.quantamagazine.org/room-temperature-superconductor-discovery-meets-with-resistance-20230308/
ReplyDeleteSigh. I am old enough to recall being chided by a colleague (name withheld) for not being "open-minded" enough to accept the possibility that Pons and Fleischmann's cold fusion claims were credible, because "after all, hydrogen has unique properties that make this possible." Here we are again.
ReplyDeleteDon Monroe: >You might want to put a hyphen in "strange metal session"
ReplyDeleteYou might enjoy this video https://www.youtube.com/watch?v=SP9Ol8Tv4Tw
I understand that it is necessary to subtract a background from a magnetization measurement in a pressure cell, but you have to have some connection to reality right?
ReplyDeleteThat is to say, the cell is going to behave e.g. largely as a diamagnet, or a Pauli paramagnet, depending on construction. There might also be some Curie tail at the lowest temperature.
Dias et al have subtracted completely arbitrary polynomials, with no link to physical reality. These switch between positive and negative slopes between neighbouring measurements (in pressure). Completely random in fact.
It seems that they demonstrated ac-susceptibility for the NYT journalist, and showed him such a glitch. However, it proves nothing, except that there was a phase transition between two phases with slightly different susceptibilities. Or am I missing something?
Peter Armitage had an exchange of tweets with NYT reporter Kenneth Chang who claims that “he saw something in the lab”. Please take a look it’s hilarious. And on Peter’s part, he did an excellent job of bringing Ken Chang back to reality.
Deletehttps://twitter.com/npeterarmitage/status/1633514508873515008?s=46&t=vNVANugoxJ8gpyl7v3IvlQ
Plagiarism: "The fraudulent representation of another person's language, thoughts, ideas, or expressions as one's own original work. Although precise definitions vary, depending on the institution, such representations are generally considered to violate academic integrity and journalistic ethics as well as social norms of learning, teaching, research, fairness, respect, and responsibility in many cultures."
ReplyDeleteOnce upon a time, that would have mattered.
In Supp. Fig. 15 raw data, there is a crossover between the 0T (black) and 3T data (red) at around 225K temperature. If an arbitrary background is subtracted from it, still this trend will be visible in the published data. However, there is no intersection between the 0T and 3T plots in the published data. Also, 0T and 1T (blue) plots don't show any crossover although they show on in the raw data.
ReplyDeleteI think this minute detail itself can raise a lot of red flags about the data fabrication. I hope experts will look into this more carefully. Someone who can claim superconductivity in a S = 5/2 3D magnet can go too far.
Absolutely brutal take down in Physics magazine by Dan Garisto:
ReplyDeletehttps://physics.aps.org/articles/v16/40
Is that the sound of the APS cavalry arriving?
I had no ideas that text from Hamlin’s thesis was also plagiarised in one of Dias’s arXiv papers too!
ReplyDeleteDias is so desperate because he is still an assistant professor. Needs tenure.
ReplyDelete@Anonymous But this is exactly the opposite of what you should do to get tenure.
ReplyDeleteI think that a future in real estate, or second hand cars beckons for Mr. Dias....
Maybe he could become George Santos’ running mate in a presidential campaign!
Delete"In response to the allegations that he plagiarized Hamlin’s thesis, Dias says he has done nothing wrong: “I have appropriate citations.” Washington State University, which awarded Dias his PhD declined to comment on whether they have carried out a misconduct investigation. A statement from the University of Rochester says, “Dr. Dias has taken responsibility for these errors and is working with his thesis advisor…to amend the thesis.”"
ReplyDeleteThe denial is just amazing, and this also reflects incredibly badly on Rochester. I thought they were a reputable place?
[from Physics magazine]
Surely this would reflect worse on Washington State. Forgiving large chunks of plagiarized text in a phd thesis? Wow.
DeleteActually Rochester scores really bad on this matter from the day 1. When Jorge Hirsch raised concerns on the 2020 Nature paper, instead of doing a thorough investigation, they blatantly supported Ranga and took to the threat and intimidation tactics by threatening to sue Jorge. This questions their credibility and should have been a warning sign for them as to check whether any future claim by Dias is accurate or not. Looks like they haven’t learned from their mistakes.
DeleteMy understanding is that WSU have not commented, not that they have forgiven anything.
ReplyDeleteIt will be crucial to recover the raw data for all of the high-pressure transport experiments in that thesis. Universities (or indeed federal funding authorities) need to seize computers in such cases IMO. They never do of course.
"The University of Rochester has conducted two internal inquiries into the CSH data-manipulation allegations, led by Steven Manly and Gilbert Collins. According to a spokesperson, both inquiries “determined that there was no evidence that supported the concerns.” But the university has not made the remit of the investigations public and has not provided any rationale for the investigations or details on how they reached their conclusions."
ReplyDeleteTake a look how U of Rochester benefits from this, how for them a lot of money is a stake, when they want to become a fenter for everything that has to do with superconductivity, including manufacturing and applications.
ReplyDeleteJust take a look at this video:
https://www.youtube.com/watch?v=ryJxMYX7YEU
and please store it before it gets taken down (like the other video by Ranga Dias)
to later show the hybris of administrators and deans etc. who like to ride the wave
and stand in the sunshine.
That's not going to age well! Explains a lot.
ReplyDeleteOne of the people in that video is responsible for handling academic misconduct at Rochester...
ReplyDeleteMorning all, here is a plot of the raw resistance data supplied by Nature for Fig. 2 at 16 kbar.
ReplyDeletehttps://imgur.com/a/2FqZWm8
I have zoomed in on the low temperature area, to show that there are two regions. Directly below the transition (175 < T < 265 K), the resistance is glitchy and looks like a measurement artifact.
At still lower temperatures (100 < T < 173 K), the measurement recovers, and shows standard metallic behaviour, which can be fit with our favourite a+b*x^2+c*x^3 expression.
In my opinion, they have measured a metal-insulator transition, probably of the sort reported years ago for these materials (see earlier comments). They have (to my knowledge) honestly reported their data, but misinterpreted it. Occam's razor etc.
SAJK
The image link is not working
Deletehttps://imgur.io/a/2FqZWm8 Works
DeleteLast comment from me (probably...), but I want to know where the low temperature (T < 100 K) data is? The video above shows students happily loading cells into a brand new PPMS, so why would you not measure to the lowest temperatures? What does the T-dependence of the residual finite resistace look like? I think that all data in the paper is truncated at T<100 K or thereabouts. SAJK
ReplyDeleteOK. OK. Last one...
ReplyDeleteJust to highlight what I mentioned previously. The authors seem to have made an honest mistake with the data uploaded for the 10 kbar plot in Fig. 2. This is 'just noise', or some sort of null measurement.
If you plot it together with the 16 kbar data, then it overlaps the noisy area (in magnitude only, they are not identical). Assuming that the 10 kbar data are some sort of null measurement, this supports my interpretation that there was a measurement fault at 16 kbar between 175 and 265 K. It also supports my interpretation that the metallic conductivity at lower temperatures reflects the actual intrinsic groundstate of the sample, which is not a superconductor.
https://imgur.com/a/Zih9Tyb
But hey what do I know, i'm just some unemployed chemistry bum ;-) SAJK
SAJK: thanks for your analysis. I have this comment. You write:
ReplyDelete> They have (to my knowledge) honestly reported their data, but misinterpreted it. Occam's razor etc.
But then:
> What does the T-dependence of the residual finite resistace look like? I think that all data in the paper is truncated at T<100 K or thereabouts.
My comment: if cropped data point at non-superconducting behavior, this is not a case of "Occam's razor etc.". If the full data contradict the selected data, the paper would be just as invalid as if the data were made up and never measured.
In the case of Delft and Copenhagen Majorana papers, this is what we see. Data not shown in direct contradiction with what is written in the papers.
Sergey, good points, but I'm going to plead the Rumsfeld on that one:
ReplyDelete"There are known knowns, things we know that we know; and there are known unknowns, things that we know we don't know. But there are also unknown unknowns, things we do not know we don't know.”
Time will tell. I suppose somebody else will measure it too. I'm already bored of this. SAJK.
I don't want to start my daily cond-mat reading habit again, but there is this paper on:
ReplyDelete"Superconductivity above 70 K experimentally discovered in lutetium polyhydride"
https://arxiv.org/abs/2303.05117
This is actually exactly what I would expect, it is very high pressure (hundred of GPa), moderate T_c (70 K) and a very hydrogen rich composition (Lu4H23). Thus totally consistent with other reports on hydrides from reputable groups. Very different in both chemistry and physics terms from the claim discussed here for LuH3-x at 1 GPa.
SAJK
Congratulations on getting an investigation started at PRL, Simon, I think your post here above was the tipping point for them.
ReplyDeleteYet, they have nearly infinite opportunities to screw up the investigation. Also likely to focus on the plagiarism, because that is really important in physics papers.
I don't know Sergey, James Hamlin did all the hard work. I just sent a few emails questioning what was going on.
ReplyDeleteIn the end i'd say a few things:
1- Most importantly, we should all remind ourselves that it is just physics. Nobody outside of a very small circle of people cares. Even the argument of 'wasting tax payers funds' is a little flawed (you could argue that physics keeps a lot of people off the streets who would otherwise by unemployable!). I just hope that everyone keeps perspective, keeps safe, and things eventually calm down.
2- Personally, the whole circus of deans, department heads, big money, politics, spin-outs etc, is why I (usually) don't want to be an academic anyway. I also didn't apply for any faculty posts this fall because of this MnS2 story (how do you explain that in a cover letter?!). Probably that was the right choice, people like me (who like playing with measurements and doing cool, but uncorrelated projects) are simply not that welcome any more in universities IME.
So unless that mythical 'Bell labs style' place still exists, then I am out anyway (if it does, tell me!). I'm not sure that it does, since this whole story is symptomatic of much larger changes in academia. Finally, APS etc are ultimately responsible for keeping their houses in order, not me, and I quite like that :-)
SAJK
Note that the Dan Garisto who wrote the critical story on Physics is not the same as Robert Garisto, the managing editor of PRL (Dan's father).
ReplyDeleteDan is not the same as Robert? OK, got it :-P
ReplyDeleteOh yes, that Bell labs place, where you could just do science without any pressure
ReplyDeleteExactly, Bell labs at times had even more toxic culture than people would ever imagine these days.
DeleteDias might not even have the material he claims to be measuring. The EDX data looks fishy. There is a peak at 1.24 eV incorrectly labeled as Lu. A sum peak for C, O, and N should be 1.19 eV, so that cannot explain the 1.24 eV peak. The caption for the figure says that there are background peaks of C and Al in the spectrum, but there is no evidence of Al which has a peak at 1.48 eV. So he might have Mg or something in there. The counts in the raw data file are extremely low, unless he means cps, but I’m not sure why they would report that instead of taking a longer collection. Also the electron image that accompanies the EDX data is trash. You can’t make anything out in the image, and it looks like he took the EDX spectrum from inside a void? Or they really just do not know how to adjust contrast settings on the scope and don’t know how to focus. I know this is not the main point of the paper, but if this is any indication of the quality of work they are doing, y i k e s.
ReplyDeleteIs the raw data supplied for Fig. 2 actually 'raw'? Both the 10 and 16 kbar data sets have 65534 points, which is close to 2^16 (unlikely coincidence for independent measurements, the T-scales are different too).
ReplyDeleteBoth also show fine structure when plotted, suggestive of over-fine binning.
https://imgur.com/a/iw43RjW
16 bit resolution of the instrument?
DeleteNo idea? Weird way to collect data as a function of temperature to decide to have ~2^16 steps between an arbitrary starting temperature of 107.2 K and a finish temperature of 289.07 K.
ReplyDeleteAnd the same for an independent measure at a different pressure from 99.638 K to 288.5 K.
ReplyDeleteNah man, you’re reaching at this point
ReplyDeleteMaybe, but it is not the raw (or source) data.
ReplyDeleteWhy not?
DeleteOK, correction. If it IS the raw data, then it is a seriously weird way to do an experiment.
ReplyDeleteIt sounds a bit like they have some home-made setup for this, so maybe that is just how it is done. No idea, just curious! Anyone know what it's about?
Pretty likely that the instrument resolution is 16 bit and the temperature “discrepancy” you mention is just working with the PPMS.
ReplyDeleteI can see why some people would be hesitant to share their data. Seems easy for things to go into some grand conspiracy direction.
Well, one of the data sets is anyway just noise, and bears no relation to the published figure. I personally think that a(?) referee, or editor should be tasked to make sure no mistakes like that go through to publication.
ReplyDeleteSomebody needs to check raw data files before publication, especially when there are so many.
Instrument resolution would show up on the y axis. Here we are talking about 2^16-2 data points on the x axis. This has NOTHING to do with "instrument resolution".
ReplyDeleteOnce this case is over, we need to discuss the role of Nature. How can it be that it publishes such a paper when the author has just had one of his papers retraced by NATURE itself? How can it be that he gets published and gets to cite his retracted papers, among which there is another one in PRL?
How can the editor of NATURE declare in the News & Views section to the article that "doubts remain", but still the paper gets published?
How can it be that people who point this out (in the end, it is all about scientific discussion in our quest for the truths of nature) be cancelled and silenced, yet the authors slip by nearly unharmed?
When will U of R finally take action and lay off Ranga Dias, and when will he finally be ousted from the scientific community?
Which consequences will this have to people such as APS past president Frances Hellmann, who state in this case:
The superconductor controversy "may stem in part from the ethos of physics, which has historically encouraged combativeness”, “The culture of physics is one that is more aggressive and not very welcoming”, and “would like to see that change”.
Physics and science in general are about truth, and not about a welcoming culture. Frauds such as Ranga Dias should feel the full condemnation of the physics community, also to deter potential follow ups. Jan Hendrik Schön and others serve as warning examples.
I get what you’re saying, and I agree that we need to make it clear that what this guy has been doing is completely unacceptable. However, I don’t think that is mutually exclusive with nurturing a more welcoming and less aggressive culture in physics. I doubt that doing so would have made a difference to fraudsters like Dias, but that doesn’t mean it shouldn’t still be a priority, for various other reasons.
Delete65536 is the maximum number of rows for the Excel 97-2003 file format - .xls - file format at which they provided data
Deletethey either didn't know it or they did it intentionally to hide data. the second is more likely since it is necessary to press extra buttons to save data in this format
In the future, NATURE must require from all authors to disclose their raw data, no further questions asked and no legal bs applied. And if they do not want to do this, and if they are not willing to publish in the methods section how their samples are being prepared, their paper will not get published.
ReplyDeleteBecause good science requires others to confirm the results by replicating the experiment. NATURE allows the fraud Dias to not disclose his methods of sample preparation by vague hints towards patenting.
BTW: one patents first, then publishes. Case closed.
Agree almost entirely. Only exception would be if the raw data involves human subjects and includes confidential protected information.
Deletethere are people out there with deep insight into how the ‘sausage was made’ in Rochester, and indeed Harvard. For the good of science, I hope that more of them speak up in public.
ReplyDeleteMy theory is that it's Lawrencium and not Lutetium.
ReplyDeleteIn the Wall Street Journal he breezily plans to swap Lu out for Aluminium to save money!
ReplyDeleteSure, why not? It’s all good, they are practically chemically identical!
Raw data appear to be available here: https://zenodo.org/record/7374510#.ZAvDWBjfsVF (link provided in the article).
ReplyDeleteFrom a discussion elsewhere on the temperature hysteresis shown for the 'superconducting' phase transition in S13 [It was pointed out that this can occur from too fast T changes, lac of exchange gas etc, normal stuff].
ReplyDeleteWhat interests me is the thought process:
1- It is known that a superconducting phase transition is continuous (no hysteresis).
2- It is known that showing evidence for hysteresis will cause others to claim that it is just e.g. some other kind of phase transition. And this is your Nobel prize winning paper right?
3- It is known that it is possible to eliminate experimental problems by sufficient care. For example, at those temperatures/pressures, you can do precise temperature sweeps with a 15 kg Paris-Edinburgh cell. See also careful measurements on e.g. 10 g samples with neutrons.
So why show the data, if it implies that either: a) Your explanation for the observed effect is wrong; or b) You weren't careful enough?; or c) That you hadn't thought about this at all?
And finally, why is this the only figure where the online source data is not supplied? This is the only other place where ~10 kbar (maximum T_c) resistance data is found in the paper. As discussed above, the data set provided for Fig. 2 for this pressure is wrong...
^That was SAJK by the way.
ReplyDeleteThe transition in the 16 kbar resistivity data (Fig. 2) is about 0.5 K wide. Does it sound credible to you?
ReplyDeleteThe 16 kbar data actually shows a little step before dropping to its minimum value.
ReplyDeleteThis roughly corresponds to the extrapolated value from the fit of A+B.x^2+C.x^5 to the low temperature finite region of resistance (see figure):
https://imgur.com/a/lSRkfki
As to the width, the transition width looks perfectly fine to me in the context of a metal-insulator transition ;-)
SAJK
Dias is basically randomly smashing elements based on the latest computational trends, hoping for more simulations to eventually back him up (and publishig whatever phase transition he gets in the meantime). To be future proof, he masks the exact measured stoichiometries and structures behind "patents pending".
ReplyDeleteA very advanced "fake it till you make it", which is the exact opposite of what an indipendent experimental verification of simulations should be.
I gather from tone of some comments that there is great skepticism about the metallic hydrogen result w Silvera from back in 2017. Similar concerns?
ReplyDeleteDoug- I would refer you to the famous (and dubious) Nature paper on 'Snappy the cloned dog'. Evidence provided in Fig. 1 is... a picture of a dog.
ReplyDeletehttps://www.nature.com/articles/436641a
I think that the Dias + Silvera metallic hydrogen provides a similar level of analysis!
Snuppy, sorry ROFL!
ReplyDeleteOK, some news. The online repository for the data has different files in it to the ones directly linked on the Nature web site. In particular, the 10 kbar resistivity plot now has more points (beyond the excel limit?) and shows a transition at higher temperature...
ReplyDeleteSome good comments directly on the article by the way.
https://zenodo.org/record/7374510#.ZAxM4y6ZM72
https://disqus.com/by/disqus_eyxKWVHH2V/?
ReplyDeleteSome comments
OK, a Nature editor has already suggested the writing of a comment to me. I have two problems with this:
ReplyDelete1- There are so many contributions, anonymous and open, that it is very hard to give credit (and a plagiarism case would be ironic right!)
2- Personally, I don't really want to engage with them (hence anonymous). Why give them the platform, when they are largely responsible for this situation?
Does anyone have any thoughts on how to proceed? I guess the comments will be flying out on arXiv next week anyway. If you plan to submit your own comment, good luck, and know that at least some editors will support it.
Which video?
ReplyDeleteThe video of Dias’ March meeting talk is linked in the tenth comment in this thread, and it still works.
ReplyDeleteOnly one video has disappeared from the net to my knowledge. I speculate that it was because of this 'aspirational' segment:
ReplyDeletehttps://streamable.com/774by8
This is amazing. In his mind, he's taking the audience to Mars. This new superconductor is going to solve everything that over 100 years of SC research couldn't. The whole talk is gloriously terrible, and you know this is all one big lie. But, TBH, I've heard similar-toned talks from some respected physicists. Maybe as a community we should rethink the fantasy parts of our talks.
DeleteRochester has said that the people named here, and the sum of money involved, were 'aspirational'. I wonder if they were investors at the time? This is quite an old video, and dates back to before the first Nature paper was retracted.
ReplyDeleteUnearthly Materials, their company, still lists six employees on LinkedIn. I wonder who is paying them? Rochester? Does that explain why senior people there still seem so invested in all this?
The width of the transition in Fig. 13 is ~1/5000 of Tc, in addition to being hysteretic. I asked Ranga Dias to supply the missing raw data for that Fig. He said he would upload them when he gets back to his office.
ReplyDeleteInternational press (real press, not science stuff) is appearing now.
ReplyDeleteIn The Hindu (paper of record, and second largest Indian paper in English):
https://www.thehindu.com/sci-tech/science/scientist-who-reported-room-temperature-superconductivity-faces-more-controversy/article66606747.ece
Il manifesto in Italy will run a story tomorrrow:
“the story of Ranga Dias, author of the scientific discovery of the century or the scam of the year.”
Le Monde also ran a very balanced piece in France already.
How long will it take US, UK media etc to catch up? There are quite a few outlets who haven’t run anything yet…
They obviously like his hyperbole at Rochester:
ReplyDeletehttps://www.youtube.com/watch?v=ryJxMYX7YEU
Note the appearance of the Dean (Wendy Heinzelman, responsible for academic misconduct), and 'Rip' Collins (who led the first investigation at Rochester, according to Physics Magazine).
This video appeared after e.g. Heinzelman was made aware of possible data replication in the MnS2 PRL, and the plagiarism in Dias's thesis.
They've gone all in. It's a bold move, let's see it if it pays off.
^Apologies, I wait years to use a dodgeball quote, then end up making a typo...
ReplyDelete@Anonymous 9:36 AM
ReplyDeleteThe whole talk is here:
https://vimeo.com/manage/videos/807275349
Note the interesting discussion by Julia Deitz on issues with the EDX data here:
ReplyDeletehttps://www.nature.com/articles/s41586-023-05742-0#article-comments
Its gone... all gone... I saw that comment earlier too, but it vanished.
ReplyDeleteJulia Deitz here.
ReplyDeleteI also saw that comment was deleted - not by me - and my Disqus account shows it was “detected as spam.” I just resubmitted the comment and it should be there again!
Looks like it's gone again!
ReplyDeleteSuggest localizing findings to https://pubpeer.com/publications/5B50A0D3400CDD252EC67D75F0841A now that a page has been started.
ReplyDeleteHere is a slide from another talk of Dias's, claiming room temperature superconductivity in yet another class of materials (the THIRD!). It is labelled as LaxSy and CexSy, which would go against every theoretical and chemical insight into the hydride superconductors.
ReplyDeletehttps://imgur.com/a/b0x8pcJ
So it begins!
ReplyDeletehttps://arxiv.org/abs/2303.05987
This comment totally missed the mark. Four point resistivity and VSM-type magnetization measurements are well-established ways to demonstrate superconductivity proven time and time again. There's a lot of justified ways to criticize the Duas group result, but this comment is not one of them.
DeleteFirst casualty of the war...
ReplyDelete@FraudHydride has vanished from Twitter. It's better to burn out, than to fade away, whoever you were!
@fraudhydride here, I think Ranga and his bots mass reported me or something, I can’t log in anymore. From my actual twitter handle, I tried to search the account, no results. Looks like they muzzled me. I do hope people took screenshots. Anyway, happy to move some of the commentary I have compiled to PubPeer.
DeleteIn the meantime, let’s also try in some other channel. I am not sure who was the reviewer of this paper but I know Tobias Roedel well. He was the editor for couple of my papers and I have done a few referee jobs commissioned by him. I can just try to contact him to keep an eye.
There is another concern. Ranga Dias has another CSH paper under review in Nature https://arxiv.org/abs/2302.08622. There is no way CSH is a RT SC. Data looks shoddy too. We have to faster before it’s too late.
So we must connect with other editors and tell them our concerns. Someone please warn Ian Osborne from science (he deals with condensed matter submissions) and any editor in Nature physics whose purview this RT SC fall under.
We have no time to lose.
Just because someone posts a counterpoint on Twitter or arxiv doesn’t make that person right. Feel like we just need to reestablish that
ReplyDeleteCexSy sounds like "sexy", is this the ultimate form of investor pitch?
ReplyDeleteJulia, I read your comment with interest yesterday. Since there is a risk of editing critical comments on Nature, could you please repost it here and on PubPeer? Perhaps you would also allow it to be reposted on Twitter? Or I could retweet from you.
ReplyDeleteHi All, I would also endorse moving all these conversations (and the one on the Nature site itself) to the PupPeer link given above. It is likely to be a better format for discussion and compilation. Thanks very much @Doug Natelson for having us over. Lots of discussion happening on Twitter as well. Perhaps substantial analyses from Twitter could get reposted there as well.
ReplyDeleteHi Peter, thanks for your hard work on this. Would you like to drop couple of twitter discussion links here just for the record? Yes, might be good to get them post their analyses in the PubPeer as well.
DeleteOne problem is that PubPeer is very highly moderated...Also comments can take an age to appear.
ReplyDeleteModerated by who?
ReplyDeleteI don't know, but some comments don't appear, and some are deleted after a short while. It is not very transparent. No reason not to mirror (or even preferentially put) comments there, just a heads up.
ReplyDeleteIt looks like according to PubPeer's policy https://pubpeer.com/static/faq#12b anonymous and "greylisted" authors can take more time to appear. I would suggest that people who feel comfortable not post anonymous and crosspost for sometime.
ReplyDeleteMy experience with pubpeer was that a comment I posted took about a day to appear. My guess is that if the comment is purely factual with no unscientific stuff it will not be moderated away. They probably have some set of very specific criteria to reject so that they handle the huge number of comments that they get. That said, no reason not to post in multiple places.
ReplyDeleteCorrect. Those running PubPeer are quite afraid of legal action and have experienced those (defamation!). So simply question the data, not state fraud.
DeleteDepending upon how good your Japanese is, there are some interesting tweets about noise here:
ReplyDeletehttps://twitter.com/iratagnubiz/status/1634879222861922304
This person also appears to have commented on the article itself. Summary- there is something weird about some of the resistance data above and below 'T_c'.
I am very disappointed to see that my comment was yet again deleted. I am not sure why this would be or who makes these decisions - Disqus again just marked it as spam over an hour after it was posted.
ReplyDeleteI am working with some colleagues on EDX simulations, and I will likely write something more formal once I have those results. In the meantime, I'm posting the original comment here for anyone who is interested.
There are serious issues with the EDX data. Given the EDX results are in large part used as evidence that the authors are measuring an N-doped Lutetium hydride, it is of high importance that these are addressed. Issues are:
• The authors state that the EDX data was collected at 15 kV. This is a poor choice of experimental parameters for this measurement. Given the high atomic number of Lu, no electron microscopist would try to show a N signal using a 15 kV accelerating voltage as the detectability of N is not optimized by performing EDX at 15 kV. One would have to do this analysis at a few kV to get sensitivity for Nitrogen. A large N concentration would be required to observe the presented result.
• There are two peaks labeled as Lu: one at 1.24 keV and one at 1.58 keV. The peak at 1.24 keV is outright labeled incorrectly. There is no x-ray emission for Lu at 1.24 keV. This also cannot be a sum peak of the C, O, and N, as that would be 1.19 keV. Potentially this could be x-ray emission associated with Mg. In any case, this means there is another element in this material (possibly Mg) that is not being reported. A full table of x-ray emissions for different elements can be found here:https://www.edax.com/resour...
• As for the Lu peak at 1.58 keV – there are concerns. Lu has an M line emission at 1.58 keV, and the peak shown in this data looks too symmetrical to be an M line emission. Additionally, the raw data shows a miniscule Lα peak for Lu at 7.65 keV. Given how small the Lα peak is, I would not expect the peak at 1.58 keV M line peak to be as high as it is. It is possible (depending on the resolution of their detection) that much of this peak is Al Kα which has an emission at 1.48 keV. The authors do state in the caption that there is background Al signal from the stub they used, but it would be surprising if this was the case given the peak center is almost exactly 1.58 keV. If the authors go back and lower the accelerating voltage used to collect the EDX data, that should result in less Al signal being collected with the smaller interaction volume.
• The electron image accompanying the EDX spectrum raises questions. The red circle on the electron image indicates that they collected data from inside something like a void, although I can't say precisely what it is. They don’t mention particles, but maybe this a particle embedded in carbon paint on an Al stub? It does appear to have some very flat crystallographic edges, but it is not obvious what it is. Certainly, a backscatter electron image would have shown the particle bright against the C background. I cannot think of an explanation for why an image like this would be used.
• The overall EDX counts in the raw data are low. The x-ray spectra could have been acquired for a much longer time to increase the counts to some acceptable number. Also, it would be important to take a number of analysis points with statistically significant numbers of counts to show the qualitative homogeneous composition of the sample.
As it stands, the EDX data should be removed from the paper, and I'm not sure the paper would still be publishable without it. This data should be recollected using a lower accelerating voltage and possibly with a different stub type to better understand what the material is.
I second the comment of Peter Armitage above. I’m happy to have discussion taking place here, but PubPeer is designed for exactly this kind of analysis and post-publication hashing out of issues.
ReplyDelete@Julia, thanks for posting; I’d read your comment before it evaporated over at Nature. If they somehow length-limit comments, then they need to spell that out. I would surely hope that specific scientific discussion of the paper at hand would be considered appropriate for the comments section - otherwise, why have the comments?
Why are the journals and conferences destroying the credibility of the community? Is the peer review system so broken that in spite of the chronic issues with the authors, elementary concerns with their latest are remarkable claims are now clarified in blogspot / twitter conversations? How is anyone supposed to keep up to date? Even if there was outright data manipulation there's no real consequence for folks like those involved here, nanowire Majorana, etc and that's the real problem. They all just keep grifting, getting taxpayer money, and wasting the rest of our time. Like the Delft case, it you assemble a committee to review the provable data manipulation / violation of scientific integrity, they'd probably come back after 3 years with a verdict of 'poor judgement' or 'negligence'. It's a tragedy really. The conference organizers and journal editors should be ashamed of themselves. On a related note, why were no questions allowed in the APS talk?! It was a joke. Let's ignore this paper and move on. The authors can share their 'patented' material or present at conferences where you can actually ask them questions if they want to be taken seriously.
ReplyDeleteWe can ignore the paper, but there are other serious things in academia. Things like writing a PhD as a sole author. Things like collaborating honestly with people. Things like sharing data when asked.
ReplyDeleteThese are the foundation, nay the honor code, of the scientific system. I don't care what crap some glossy journal publishes, but we ignore the above at our peril. Certain universities and funding bodies need to step up to the plate. Openness, integrity, and honesty are needed now.
@fraudhydride
ReplyDeleteI wouldn't worry about science too much, they rejected the CSH paper at the editorial stage. They did not like the conditions the authors tried to impose on them, namely signing non-disclosure agreements, and choosing from an author-provided list of referees. Allegedly.
1) Academia cannot rely on an 'honor code' for credibility. That's naive and leads only to where we find ourselves now. 2) Unfortunately, the only thing that matters is what 'some glassy journal' publishes. That is the only currency in academia now. 3) There is no peril in ignoring chronic offenders - they're not obligated to our attention. 4) Let's clean up by first holding the past offenders accountable (there's plenty of those) and instituting rules for raw data, sample sharing, etc. Yes, universities and funding bodies should lead the charge not Nature and Science. 5) Isnt it astonishing and completely backwards that we are now afraid that Nature will publish the follow-up paper before it is torn apart online. Thats when you know that the peer review system is completely broken.
ReplyDeleteThe point I was trying to make is that if you e.g plagiarise your thesis, you should be out.
ReplyDeleteNo second chances, no “working with your supervisor to correct errors” etc. Just out.
I saw the post of 'Rafa' in the comment section of the LuH2 paper in Nature. This highlights how subtracting a a+bx^2+cx^5 background from the resistivity data yields strange results.
ReplyDeleteI thought I had seen something similar before, and remembered where. It is in this paper, which claimed to reproduce resistance data for C-S-H (it's the same gang):
https://pubs.rsc.org/en/content/articlelanding/2022/cc/d2cc03170a
The resistance data shown has strange peaks above T_c. These look very similar to the result of subtracting off a polynomial background from the low temperature region. So what happened? Interestingly, the re-submitted C-S-H 'Nature' paper (arXiv:2302.08622) has lost its resistance data.
You can find the comparison between Rafa's plot and the C-S-H data here:
https://imgur.com/a/GccwZCC
Yuck, there was some alpha channel thing, and the references were not visible. Better plot here:
ReplyDeletehttps://imgur.com/a/XFKwfEv
As a further aside, the peak in the (wrongly) background subtracted data will be at the onset of the phase transition (since the fitted background curve crosses the data at this pount).
ReplyDeleteSmith et al report much lower Tc than in their original work (only 190 vs 280 K. However, the curves in Smith et al actually have two features. It might be that it is possible to 'tune' between these features with a background subtraction, to make one or the other look like the onset of the superconductivity.
However, to do that, you would need something more complicated than the x+x^2+x^5 curve discussed above. Maybe something like a cubic spline?
What do you refer to? The arXiv paper here is the first mention of VSM measurements that I saw?
ReplyDeletehttps://arxiv.org/abs/2303.05987
> I wouldn't worry about science too much, they rejected the CSH paper at the
ReplyDelete> editorial stage. They did not like the conditions the authors tried to impose on
> them, namely signing non-disclosure agreements, and choosing from an author-provided
> list of referees.
Is that true? How do you know that?
Do you mean Nature likely accepted to restrict their referee choice to an author-provided list? I can't imagine this.
If this were true, it is not peer-review system that is broken, it is Nature.
This comment has been removed by the author.
ReplyDeleteThe problem here is nearly 100 % with Nature. They are supposed to make sure that peer review is done properly. At some stage, all of us have had things knocked back, which upon reflection, were a bit over done. How can they get it so wrong?
ReplyDeleteFrom a comment by 'Cosmic Magnet' on Reddit:
ReplyDeletehttps://www.reddit.com/r/Physics/comments/11qpv7p/raw_data_vs_published_data_for_room_temperature/
"...my own PhD advisor was one of the original referees contacted by Nature to review Dias's original paper on room-temperature superconductivity in the high-pressure hydrides, the one that was retracted. My advisor refused to referee it, and instead wrote a letter to Nature requesting they not publish. Nature ignored and asked a different referee, so it seems they already had their mind made up and just needed the referee reports to rubber stamp. The point is that the blame for bad articles like this one squarely rests on the shoulders of the Nature editorial team, because without more details it's possible that it was published over the strong protests of the referees. These details will never come out in order to protect the confidentiality of the reviewers."
This is ancient history, but Nature also overrode serious objections during review of one of Hendrik Schön's "superconductivity" papers. They also declined to publish a letter from IBM stridently criticizing the "SAMFET" work. In both cases, the critics were right, and Nature was dead wrong. I would have thought they had learned something.
ReplyDeleteHow much longer does Tobias Roeder wants to cover this up? It is crystal clear that Dias once more substracted the residual resistance from his data.
ReplyDeleteIs this the editorial standard that is upheld by NATURE?
Everybody and their grandmother is now an expert on high Tc superconductivity measurements
ReplyDeleteYes unfortunately the Internet is not so good of a judge about this. Proper and justified scientific critique is mixed up with ignorant comments.
DeleteClear example of this is the comment by bioscience professor Srinivas Sridhar (https://arxiv.org/abs/2303.05987) who unintentionally rejects 90% of measurements of superconductors based on a misguided definition of a superconductor and ignorance of experimental studies in the field.
Although I think Sridhar’s argument has gaps, still let’s not better to tag him as “bioscience” professor. He has actually done a good amount of research in superconductivity and has a reasonable understanding of the phenomenon. It’s totally fine to have disagreements about interpretation. That helps nurture the science culture. But reducing someone’s comment mainly because he holds the professorship of a “bioscience department” won’t really help. Please understand.
DeleteWell, all these background subtractions certainly put a new meaning on the phrase “unconventional superconductivity”!
ReplyDeleteIf you want to reduce 'ignorant comments' on the internet, email Nature editors to do their job or cancel their subscription please. Armchair experts are on the rise because the academics in charge have recklessly turned scientific publishing into a racket.
ReplyDelete"Clear example of this is the comment by bioscience professor Srinivas Sridhar"
ReplyDeleteWhile I agree that the rejection of the measurements does is odd, reducing Srinivas Sridhar as a "bioscience professor" is a bit strange too. He has performed a lot of significant research in the area of superconductivity, i.e. see https://srinivassridhar.com/superconductivity/
And hey, some of use just wish we were any kind of 'Professor' :-P
ReplyDeleteMore questions about Dias publications:
ReplyDeletehttps://pubpeer.com/publications/69EDBAECD50F31B051ECECCD1DF346
I have written this down a couple of years ago already... Never shared anyhwhere - first time here! What do you think:
ReplyDeleteWays in which Nature failed on ‘Quantized Majorana Conductance’
Accepted in a very short time (of order one month)
Only two referees
Only one round of review, i.e. two substantive referee reports
Nature paper said data were available online but they were not
Took 1 year to retract, from the date the authors expressed intent to retract
Published a misleading expression of concern (‘The authors have alerted the editors’)
Which for months was the only piece of information that something was wrong
And did not explain what was wrong and the magnitude of the problem
Declined to publish or send for review a replication study with a negative result
Did not use editorial retraction
Did not make problematic paper open access right away
Inaccurate and confusing retraction notice which the original authors were allowed to write
Which does not acknowledge biased data selection
Nature News report does not discuss Nature’s own role in promulgating a false scientific claim
No separate statement from the editorial office as to what happened
Nature has no oversight from research community, such as an editorial board
Nature does not investigate complaints about their papers
Nature’s policies are poorly written, ambiguous and vague, hard to find
Editors responsible for each paper are not known to the public
It's a great job, being a Nature editor. No real responsibility, no real rules, and no negative citations. Their wasting of the scientific community's time goes completely unpunished.
ReplyDeleteWith great power comes great responsibility... Or not.
Sergey, your Majorana conclusions overlap a lot with the Dias stories. I can't imagine that Nature isn't doing this on purpose.
Does anyone else have the impression that the profile of a Nature editor has changed? They seem to be younger and less experienced now on average. In the lesser Nature journals, some are essentially in their first jobs post PhD.
ReplyDeleteI suppose this reflects the casualization of the workforce everywhere, and reducing salaries and working conditions. The end result of buyout, overexpansion, and managerialism. Probably the only way to get a pay rise/promotion is to oversee the lauching of some new product, and hence move further away from the research.
Effects which are also seen in e.g. National labs and universities nowadays...
Interesting that the pure LuH2 paper showed a color change with pressure but the recent submission of the N doped sample does not see it as Dias’ did.
ReplyDeletehttps://pubs.acs.org/doi/full/10.1021/acsomega.3c00207
ReplyDeletereally a missed opportunity here to claim this thing as a topological superconductor Majorana platform while we're at it
Well, from the abstract: "have van Hove singularities, which confirms their superconductivity."
DeleteThat disqualifies the rest of the paper, imo, already...
When will the career of Dias finally be shut down? JH Schön, after his faking was revealed, was finished and never got a foot on the ground again. This Dias character is a notorious lyer and faker, supported by his peers at Rochester and supported by a shady editor at NATURE who supported his fake news spreading TWICE. When will his career be ended, by COMPLETE embargo against him. No more talks, no more editors that decide to send his papers out, no more job at Rochester.
ReplyDeleteNever thought I’d see physicists acting like qanon conspiracists.
ReplyDeleteWhat is ACS Omega? Is it for when you are rejected from ACS alpha, ACS beta, ACS gamma etc?
ReplyDeletePRL have published an expression of concern on the MnS2 paper:
ReplyDeleteOn 30 June 2021, Physical Review Letters published the article “Colossal Density-Driven Resistance Response in the Negative Charge Transfer Insulator MnS2” by D. Durkee et al. Questions have arisen regarding the integrity of the data. At this juncture, we are investigating this allegation with the cooperation of the authors.
https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.130.129901
Anonymous@1:42pm
ReplyDeleteACS Omega appears to be ACS' version of Scientific Reports, i.e. it is an open access journal publishing "scientific articles that describe new findings in chemistry and interfacing areas of science, without any perceived evaluation of immediate impact."
Reading your blog was a journey through a landscape of profound ideas. Your ability to weave facts with narrative is exceptional. It's not just a read; it's an experience to savor.
ReplyDeleteelectronic components shop near me
electronic parts
electrical components
electronics component suppliers
electronic component suppliers