Science is said to be moving faster than ever before. In fact, the pace is usually limited by the flow of information—I cant respond to your results until I know about them, and journals are notoriously slow. The arXiv, which hosts manuscripts that may be submitted for peer review, is not. Put your preprint in today, watch it go public tomorrow… and see it get slapped down three weeks later.
This has all played out before our eyes with an extraordinary claim for room-temperature superconductivity, which is marred by some suspicious data.
Why is superconductivity important?
Superconductivity is a hidden workhorse of todays medical and scientific establishment. The way to get high magnetic fields is to use a very large current loop. In ordinary conductors, the imperfect conduction of electrons heats the metal, which increases the resistance of the metal. The whole feedback loop terminates when your finely designed metallic coil turns into a small glowing metallic puddle—often considered a bad thing.
This can be avoided because some metals become superconductors at low temperatures and have zero resistance. But superconductivity has limits: resistance returns if the current is too high or if the magnetic field is too high.
Although there are various high-temperature superconductors (where high means somewhere between -196 to -90°C), they have proven to be difficult to use. All the magnets used in big apparatus use superconductors that operate at liquid helium temperatures (-270°C). Liquid helium is really expensive, and it is a limited resource that can be the subject of some very odd market forces.
This combination of applications and circumstances makes any claim of room-temperature superconductivity very exciting—and makes everyone who researches superconductivity very suspicious.
Sobering up electrons
Superconductors work at low temperatures because they rely on some pretty delicate physics. Resistance is due to an electron not really knowing what is going on. It finds itself surrounded by a seemingly regular array of metal atoms. Based on that information it sets off, like a confident drunk driver. But atoms vibrate, and crystals are not perfect, so our drunk electron bounces off atom after atom. Each time, it changes direction and loses energy.
Below a certain temperature, electrons can pair up to form something that behaves like a single extended particle. In a kind of physics irony, the vibrations that helped create resistance before now help keep the pairs of electrons together. The pairs spread out in a giant wave that holds information about the atomic positions. The electrons are guided by this wave so that they can flow without resistance. The electrons will even pass through thin non-conducting barriers and still not lose energy.
That pairing process is very delicate. If the temperature is too high, the magnetic field too large, or the number of electrons too high, the pairs get broken up, and superconductivity vanishes.
Pairing up while high on temperature
The researchers, a pair of scientists from the Indian Institute of Science, were aiming for something different. They wondered if the structure of a material could cause electrons to scatter in a way that causes them to pair up—using a material's properties to promote superconductivity.
To test this, the researchers used silver nanoparticles embedded in gold. The choice of silver and gold makes sense because electrons in gold and silver dont much notice vibrations in either metal. This means that the pairing mechanism from standard superconductivity is inefficient for their materials.
Unfortunately, I dont see how randomly structuring a material can cause electrons to pair up and drop into a superconducting state. I cant even really discuss how they made their choices for nanoparticle sizes or ratios of gold to silver because there isnt a lot of information to go on.
According to their data, though, they observe a transition to superconductivity at temperatures between -123°C and 77°C—yes, that is +77°C—depending on the ratio of gold to silver. Actually, for the higher transition temperature, they never actually observed the transition: they claim the critical temperature, if it exists, is higher than 77°C and could not be reached in their equipment.
The data is also remarkable for another reason: the critical magnetic field seems to be extremely high. The researchers still observe superconductivity at -123°C for applied fields of three to five Tesla. This is big: I think niobium is the best pure metallic superconductor with a critical field of about 0.2T (some alloys are better).
Thats not noise
The biggest problem, though, seems to be the data itself. A sharp-eyed researcher noticed that two of the traces had identical noise, just offset from each other slightly. I cannot think of a single reasonable explanation for this. Noise is random and should never repeat. The graph could be the result of a series mistakes: the same dataset processed twice with a mistake made in the process in one run. Then one dataset gets mislabeled and used in the paper. But I seriously doubt that, which raises the possibility of data misrepresentation.
I also thought it strange that no one else seems to have investigated the electrical properties of silver nanoparticles in gold. After a quick search it seems that I am right and wrong. The optical properties—the bit that I am usually interested in—are more than done to death. There was nothing obvious about the electrical properties, although that does not mean there are no papers on the conductivity of silver nanoparticles in gold.
It is possible that the researchers got lucky and what they observe is mostly reality (with the exception of the dodgy noise). But, I think I would have to see this done in another lab before accepting the results as presented. Luckily, I think that can be done quite quickly.
arXiv.org, 2018, ID: 1807.08572, 1808.02929, (About the arXiv).
[contf] [contfnew] 
Ars Technica
[contfnewc] [contfnewc]
