Why so Many Twists and Turns in the Ongoing IISc Superconductivity Saga?

It has been over five weeks since the supervisor-student duo, Anshu Pandey and Dev Kumar Thapa, from the Indian Institute of Science (IISc), Bengaluru, posted their experimental findings via a preprint paper, stirring up the condensed matter and materials research community of the world.

The paper had claimed that they had found superconductivity at room-temperature and pressure in a nanostructure composite of gold and silver. Both these noble metals, which are good conductors, are individually known not to become superconductors even at temperatures approaching absolute zero (-273.15º C).

The paper has apparently been submitted to Nature, the publication of which would in a sense authenticate the discovery. But nothing of significance has transpired since then. The unduly long period of peer review by the journal is also somewhat surprising. Let us review the status of the claim and the issues that have surrounded it since.

Given the discredited history of several such room temperature superconductivity (RTS) claims from around the world, the IISc duo’s astonishing claim naturally received mixed reactions from scientists. Extraordinary claims also immediately trigger scientific scrutiny by others, an integral part of the scientific process, to validate and confirm the claim or negate it. But superconductivity experimentalists around the world have been feeling frustrated that the paper does not provide enough details on the process used to make the material in which RTS was apparently detected, making it difficult for them to verify its claims.

The process used by Pandey and Thapa is stated in their paper as follows. Silver nanoparticles about 1 nm wide, prepared by “standard colloidal techniques”, were embedded in a gold matrix in a chemical sintering process. The nanostrucured gold-silver composite globules so obtained were about 10-20 nm in size. These were then turned into thin films and pellets for electrical resistivity and magnetic susceptibility measurements respectively. There are no further details beyond this sketchy description in the paper. It is, however, quite likely that what has been posted by the authors in their preprint paper is only a teaser, and that what has been submitted to Nature includes more details.

Biasing the narrative

“It was a remarkable claim, so there was lots of interest,” Pratap Raychaudhuri of the Tata Institute of Fundamental Research (TIFR), an experimentalist in condensed matter physics including superconductivity, had told Nature. “Several laboratories quickly leapt into action to try to replicate the results. But their efforts were frustrated because the preprint did not provide the details needed to manufacture the gold-silver material, and because Thapa and Pandey declined requests to share their samples. After two weeks or so, the community started to get more impatient.”

“In 1986 when [Georg] Bednorz and [Karl] Muller announced their discovery of high-temperature superconductivity [at -238º C], labs around the world reproduced their results within a week to 10 days. When a Japanese team announced a novel bismuth sulphide superconductor in July 2012, we synthesised it within a week, following their recipe,” V.P.S. Awana, a condensed matter experimentalist from the National Physical Laboratory, was quoted as saying by The Telegraph. “They should let others make similar measurements on their novel superconductor.”

“The Nature embargo is a lame excuse. According to the journal’s embargo policy, there is absolutely no embargo on any communication with scientific colleagues,” Raychaudhuri told The Wire. In a post on Facebook, he had written, “At least two dozen labs in the country can do this measurement for Thapa and Pandey. For the sake of healthy academic discourse, it is of paramount importance that Thapa and Pandey come out openly with their data and their samples. Their silence is harming them, and in the process harming their institution and the whole of Indian science.”

Indeed, much of the discussion around this process of scientific scrutiny has played out on the social media and on blogs – a common sight these days. But the reach and influence of exchanges in these fora go beyond that of the research community itself, including via the media. This has its merits and demerits.

Consider the case of the RTS claim itself. The biggest controversy that surrounds it now is rooted in a Twitter thread by Brian Skinner, a physicist at the Massachusetts Institute of Technology (MIT), whose keen eyes noticed something unusual in the data presented in the Pandey-Thapa paper. He found that background noise fluctuations – which are supposed to be random – in two independent magnetic susceptibility measurements (at two different applied magnetic field values) were identical. On the face of it, this neither seems right nor has any obvious explanation.

Skinner had also recalled in his post how identical noise patterns detected in an earlier case of research from the highly reputed Bell Labs led to the unraveling of one of the biggest instances of scientific fraud, which had gone on undiscovered for many years and across several reputed journals – well known today as the Schön scandal.

This aside from Skinner had the undesirable effect of most blogs and exchanges on the social media commenting on the Pandey-Thapa work from this perspective, with some commentators – particularly from the West – already branding the work as fraudulent. One author who had posted a comment on the work on the arXiv preprints repository, when asked to elaborate on it, said to The Wire, “Don’t waste your time. It is not a superconductor. Also, it seems that these guys have cooked the data. IISc should form a team of scientists and investigate their measurements and data.”

Controversies aren’t everything

To be fair to the authors, they have, through email exchanges with Skinner, thanked him for pointing out the repeated noise pattern and have told him that they had not noticed it, they don’t know of its origin, and that they will look into it and get back to him. They have also stated that they are getting the data independently validated by others. A fraudster is unlikely to respond in this manner. The Wire has also reliably learnt that Pandey and Thapa have indeed shared their samples with another IISc group from a different department for data validation.

As a scientist, The Wire found, Pandey is held in high regard by scientists across Indian research institutions for the quality of his past work. “I am saddened by the unnecessary controversy that has been created… Give them sufficient time to get their work peer-reviewed and published. Pandey has done excellent work in [the] past and his group is reputed in the field,” said Vivek Polshettiwar, a chemist specialising in nanocatalysts at the Tata Institute of Fundamental Research (TIFR), Mumbai.

“We know Pandey’s earlier work, I have followed his very interesting work on luminescence in nanomaterials,” remarked T. Pradeep, a chemistry professor at IIT Madras known for his research on noble-metal nanoparticle clusters and their applications. “He is known to be a thorough and meticulous experimentalist,” Arun Grover, an ex-TIFR superconductivity experimentalist and former vice-chancellor of Punjab University, said.

Given this, one would have liked to see discussions on the paper from a more academic perspective, setting aside the controversial ‘noise-repeat’ part of it. Unfortunately, such discussions and commentaries have been limited.

The first such discussion was from a theorist named Ganapathy Baskaran of the Institute of Mathematical Sciences (IMSc), Chennai, even before the controversy had surfaced. He postulated a theory to explain the observation in terms of the manifestation of latent – or hidden – superconductivity in systems involving monovalent metals (such as gold and silver) due to special quantum processes that can arise in such composite nanostructures. The Wire had taken note of this hypothesis in its earlier report itself. The theory also predicts the occurrence of superconductivity in other such composite monovalent metal nanostructures, which may be checked for.

David Pekker and Jeremy Levy, physicists from the University of Pittsburgh, in a comment posted on arXiv on August 17 (i.e. after the controversy), set aside the unexplained noise correlation problem and suggested an interpretation other than superconductivity for the observation. According to them, a physical process they called “temperature-driven percolation transition” could mimic the signals of a superconducting transition, particularly that of zero resistance, in such systems. Pointing out similar temperature dependent signals in earlier experiments, they also suggested a simple follow-up experiment that could rule out such a possibility.

Raychaudhuri, one of the most outspoken and harshest critics of the authors’ silence, and of their reluctance to give complete details of the material preparation process and for not sharing the samples with other researchers, had, however, offered a possible explanation for the identical noise patterns.

According to him, this “noise” might not be noise in the conventional sense of the term but a feature of the signal itself, arising from the motion of nanoparticles. This, he says, can occur in a not-too-well-compacted sample that allows the particles to move with respect to each other under the influence of the magnetic field, causing the sample to become deformed but only to the extent of the material being able to return to its original shape when the magnetic field is switched off. When a different low magnetic field is applied, the particles go through similar motion and the same “noise” gets repeated as the material retains its memory of deformation. At stronger magnetic fields, the deformation is large and the particles get detached from each other, and there is no memory of the original any longer.

However, this explanation has not found many takers who don’t find it convincing enough. The possibility for repetition of the noise pattern, according to some others, is that what is measured is magnetisation and not magnetic susceptibility (which is magnetisation divided by magnetic field). The magnetisation data itself may not have this noise repetition, and which could be arising from some peculiar behaviour of the magnetic field (perhaps in the field producing equipment itself) at low field values.

Notwithstanding these, it is a testimony of Pandey’s fellow scientists that, given his reputation, they are standing by Pandey-Thapa and giving the duo the benefit of doubt. The authors may have good reasons – beyond the pre-publication embargo placed by Nature – for not revealing how exactly their materials were prepared. Raychaudhuri’s point about sharing samples with others is a valid complaint but while the authors have not shared their samples with many laboratories, they have said that samples have been shared with one other research group for validating their results. Once done, they have also agreed to share the results of that exercise, which seems reasonable enough.

Attempts to replicate

The ‘noise-repeat’ issue also appears to have dissuaded some groups from trying to reproduce the results. Pushan Ayyub of TIFR, who had told The Wire in its earlier report that his group would be attempting to repeat the experiment with material prepared through a different process, has now been quoted in Nature saying that his lab had not completely stopped trying to replicate the results, but the pace was no longer frenetic.

Speaking to The Wire, Ayyub said that, using a different process, his group had prepared some material samples, gold-silver composites as well as gold matrix with a substitute for silver that would behave just like silver, but they had not found anything of interest. Mingda Li of MIT, according to Nature, is another researcher who had planned to reproduce the results. But following Skinner’s observation, Li called a meeting of his group and it was collectively decided to abandon the replication attempts.

On September 3, a research team from the Indian Institute of Science Education and Research (IISER), Pune, and the UGC-DAE Consortium for Scientific Research, Indore, posted a preprint paper on the arXiv server describing the results of their attempt to detect superconducting transition in gold-silver nanostructured thin films made by the pulsed laser deposition (PLD) technique, which is quite different from the colloidal technique used by Pandey-Thapa. The group found no signature of superconductivity, either magnetic or resistivity transition, over a wide temperature range: -268º C to 27º C. “The sample was fully metallic down to the lowest temperature,” they state in their paper. Not only did the resistivity not reach zero, it did not even show any drop throughout the temperature range.

Using PLD and different metallic targets, the researchers deposited alternate layers of gold and silver of specific thickness, with silver’s thickness much lower than gold’s, on silicon and quartz substrates and repeated the process a hundred times to produce films of a total thickness of about 80-100 nm. The characterisation of the material so produced, using X-ray diffraction and field emission scanning electron microscope techniques, confirmed the presence of a nanostructured configuration in the film samples.

As the authors emphasise, it is important to note that the sample state here is completely different from what Thapa and Pandey have reported. “There cannot be any comparison between the two,” the IISER Pune team wrote. Therefore, the upshot of this paper – as the authors state very clearly – is that there is no signature of superconductivity in the limited case of gold-silver modulated nanostructured thin films made by the PLD technique. So this does not immediately imply that the Pandey-Thapa RTS result is fake or a false alarm.

It would however seem a bit curious, at least to a non-specialist, that the various attempts to reproduce the results have not employed the same material fabrication method – colloidal processing – especially when their paper states, “Sample preparation was done using standard colloidal techniques” and when the authors also claim to have found superconducting transition for a range of different proportions of gold and silver. From a lay perspective, one would imagine that researchers engaged in the field should be able to prepare nanostructured gold-silver composite samples with random proportions of the constituent metals and look for superconductivity signatures.

Things are not as simple even though this technique for preparing nanoparticles is at least as old as nanotechnology itself. One particular  statement in the IISER Pune paper reflects this problem: “We realised that it may be non-trivial to realise the specific state of the sample discussed in their work without clear and detailed information about the specific steps in the synthesis.”

“The colloidal route to preparing such structures with 1-2 nm silver particles in 10 nm gold particle is not standard,” Raychaudhuri pointed out. “In the absence of a detailed recipe, several groups have tried to replicate using different routes, for example sputtering. None of those attempts so far showed any superconductivity.

“However, since the detailed nanostructures of those samples are like to be different, they do not per se negate the Anshu-Pandey results, but merely cast doubt. But they have kept absolutely quiet, effectively reinforcing the doubt, and reducing the incentive for people to try to reproduce the result,” he added.

“We routinely make gold and silver nanoparticles of different shapes and sizes. We have earlier looked for superconductivity in gold-silver composites but did not find any,” Pradeep explained. “So when this paper came out, I was very excited and thought everything [process details] will be revealed soon, but nothing has happened.”

According to him, with the limited information provided, nobody will be able to reproduce the claimed RTS results. For example, he noted that saying ‘10 nm gold nanoparticle’ did not provide much useful information about the nanoparticle’s various attributes, in effect making it hard to characterise it. A nanoparticle cuboid is different from, say, a sphere or a rod or octahedron.” Moreover, the technique used to prepare the nanoparticles may have been standard “but one needs the exact synthesis protocol used. The chemical reduction process affects particle surfaces. Particles have surface charges and colloidal repulsion prevents particles from being stable and the ‘monolayer’ [of the chemical species used for reduction] decides the final stability.”

This level of detail has to be provided because, Pradeep continued,“there is an almost infinite variety of nanostructures that one can end up with depending on the protocol used. The structural characterisation details they have given are basic, rudimentary information.”

The thing about nanoparticles…

“In the synthesis of nanoparticles, almost all reaction parameters turn out to be critical,” D.D. Sarma, a colleague of Pandey and Thapa from the same department at the IISc, said – referring to how all those parameters need to be fixed to the right values. “So it is perfectly possible for the same nanomaterial with essentially the same size, shape, crystal facets and morphology to have completely different properties based on the surface treatment.” He provided the example of the photoluminescence quantum efficiency of a nanomaterial, which “may vary from nearly zero to nearly…100% just based on what is or is not there on the surface of the nanomaterial.”

“Although there are a large number of methods to make gold-silver nanoparticles, there is always a probability that every protocol produces nanoparticles with some different properties, in terms of coating, functionality, sizes, shape, crystal structure, etc.,” Polshettiwar elaborated. “Therefore, it will be important to use  exactly the same protocol to make these materials to check the reproducibility of their results, and hence we need to wait until they publish their work.”

He added that a close inspection of the duo’s electron microscopy and X-ray data indicated that “these are not typical bimetallic or core-shell gold-silver materials but look like several small silver nanoparticles embedded in big gold nanoparticles,” as a result of which synthesising them may not be a “trivial” process.

To get a perspective of these remarks: remember that an atom is about a tenth of a nanometre in size. So at 1 nm, one is dealing with an assembly of about tens of atoms. So when one says ‘1 nm silver particles embedded in a gold matrix in the form of 10 nm-20 nm size globules’, one is talking of a near-spherical assemblage of hundreds of atoms of gold and silver, which are almost of the same size, and the fabricated thin film or pellet is an aggregate of such globules.

Although both have similar crystal structures, it is important how the atoms are packed and how the silver is embedded with respect to the gold. This influences the quantum dynamics between the two species, such as  electron transfer between the two, which is ultimately responsible for any collective behaviour of the electrons, including superconductivity.

Polshettiwar pointed that “The colloidal route generally allows better control over size, shape, crystalline nature, defects and functional groups of nanoparticles, and all these properties play a significant role in their quantum properties and electron transport behaviour.”

Indeed, in a 2015 paper published in the journal Pramana, Pandey and colleagues outlined how counterintuitive quantum behaviour, particularly electron transport, can manifest in nanostructured materials prepared by the colloidal route. Since their objective, as Thapa and Pandey state in their paper, was to look for non-phononic electron pairing mechanisms, the colloidal technique seems to have been a natural choice for them over other methods for preparing the nanostructured materials.

But as to the continued silence of the authors about revealing details of the process, it is possible – as reports on the grapevine on Facebook say – that the authors may have filed for a patent for the unique fabrication process that produces superconducting gold-silver composites, and hence don’t want to share the process details at this time. The other scenario, which would be rather disappointing, is what Raychaudhuri has stated – again, hearsay – in his Facebook post: “They are unable to reproduce their samples and the original ones are not showing what they claimed in the paper; so they think it has degraded.”

So if the RTS claim was one more addition to the wayside bin of failed discoveries, so be it. False alarms and failures too are part and parcel of the scientific process of discovery. But let us wait and watch how the story pans out in the days to come. It is not yet time for its epitaph.

R. Ramachandran is a science writer.