Monday, February 15, 2021

Major Embarrassment for Microsoft: No Majorana Particle After All


In 1937, the Italian physicist Ettore Majorana published a paper predicting the existence of something that came to be known as the Majorana particle.  In the society of subatomic particles, the Majorana is rather standoffish:  without a positive or negative charge, without an antiparticle (technically, it's its own antiparticle) and without even a magnetic or electric dipole moment.  Even the famously neutral neutron has a magnetic dipole moment.  A few months after writing the paper, Majorana sent an enigmatic note to a colleague saying he was sorry for what he was about to do, got on a boat bound from Palermo to Naples, and was never seen again. 


Of course, the physics community started looking for Majorana particles right away, and the search intensified after people began trying to make quantum computers.  Theoretically, a quantum computer can perform certain kinds of calculations many orders of magnitude faster than ordinary bit-based computers, because each "qubit" can hold a combination of states and thus process more information in a given amount of time.  (That explanation probably gives physicists a headache, but it's the closest I can get in the space I have.)


Anyway, it turns out that if engineers could make Majorana particles, their standoffish nature would become a virtue, because the quantum computers people have devised up to now all suffer from a common problem:  insufficient isolation from the environment.  The quantum states needed to do quantum calculations are very delicate, and any little disturbance from magnetic or electric fields, or just the passage of time, busts up the party so much that extensive error correction and multiple processing of the same problem are necessary.  Theorists say that a quantum computer using Majorana particles would be much less prone to such errors because the particles are so inert, relatively speaking. 


So the quantum-computing world was quite impressed back in 2018 when researchers funded by Microsoft announced that they'd finally made a Majorana particle.  The alleged particle wasn't "fundamental" in the sense that it was a single entity.  Rather, they said it was a kind of collective phenomenon created by electron interactions in a cold semiconductor. 


There's nothing fishy about that.  Even my EE undergrads learn about positively charged "particles" called holes, which turn out to be a collective effect of electrons in a semiconductor.  But in January of 2021, the same research group published a new paper saying basically, "Oops, we screwed up."  Some critical data tending to falsify the result was omitted from the 2018 paper, which they are withdrawing. 


The new paper came about when Sergey Frolov, another physicist, questioned the results of the 2018 paper and obtained their raw data, which included points that were not shown in the 2018 paper. 


Leo Kouwenhoven, the leader of the Microsoft research team, released the new paper before peer review along with a note that retracted their earlier paper.  He refused to comment further on the new paper because of peer review, but it's fairly clear what has happened, as described by a recent report in Wired. 


Under pressure to deliver results, the Microsoft team omitted a part of their data, allegedly for "esthetic" reasons, and published the 2018 claim to have discovered a Majorana particle.  In retrospect, omitting the esthetically displeasing data was not a good idea.  But they did the right thing in providing Frolov with unpublished as well as published data, and in issuing a new paper showing that they were basically incorrect in their 2018 interpretation of the same data.


Physics is hard enough even when the only motivation is intellectual curiosity.  When the auxiliary pressures of continued funding, fame, fortune, or tenure get into the mix, it's tempting to make claims that later can't withstand intense scrutiny.


In my own peculiar little field of ball lightning research, I see this quite often.  Ball lightning is an atmospheric phenomenon which thousands of people have seen over the centuries.  There is a consistent set of characteristics which leaves little doubt that there is a real thing there which occurs rarely, but not so rarely that people never see it.  However, there is as yet no generally accepted scientific explanation for ball lightning, and no one has ever been able to produce anything in a lab that shows the most common characteristics of ball lightning.  Worse yet, there are no photographs or videos that are generally accepted as showing an actual ball lightning object.


Of course, taking a photo or video or obtaining other kinds of objective data on ball lightning would be a major accomplishment, and many people have claimed to do that over the years.  But subsequent investigations, either by the original researchers or someone else, usually shows that there is a simpler explanation than ball lighting for what was photographed, or just leaves the question unresolved.


I don't attribute base motives to people who publish exciting-looking data that later turns out to be not so exciting.  There's always the chance it will prove to be the real thing, and one important reason for publishing scientific data and interpretations is to get it out in the open so others can look at it and criticize it if necessary, just as Forlov did with the Microsoft data.  And while it's embarrassing and can lead to adverse career consequences, admitting that you made a mistake is a part of being an adult, and Kouwenhoven and his group have done the right thing by publishing the later paper and retracting the 2018 one.   


Some people would look at this situation and say there's something wrong with the way physics works, but I disagree.  As my wife says in a different context, "More communication is better than less communication."  Let anybody who even thinks they have something worth publishing go ahead and publish it, and let the reviewers and critics have at it as hard as they like, without being mean, of course.  That's the way progress happens.  


Allegedly, a person matching Majorana's description was seen in the late 1950s in Valencia, Venezuela.  For several years leading up to his disappearance, Majorana had become increasingly isolated, almost like his eponymous particle.  And while he may have lived through World War II and after in professional silence in Venezuela, he may have ended his life in the waters off the Italian coast on March 25, 1938.  Some things we just can't know for sure yet, and that goes for physics too.


Sources:  The report by Tom Simonite "Microsoft's Big Win in Quantum Computing Was an 'Error' After All," appeared on Feb. 12, 2021 at  I also referred to Wikipedia articles on Ettore Majorana and the Majorana particle. 

No comments:

Post a Comment