Superluminal Neutrinos, Take Two 98
Coisiche writes "To address the many responses to their original findings, the OPERA team who reported the detection of faster-than-light neutrinos is starting a new and improved version of their experiment. 'The neutrinos that emerge at Gran Sasso start off as a beam of proton particles at CERN. Through a series of complex interactions, neutrino particles are generated from this beam and stream through the Earth's crust to Italy. Originally, CERN fired the protons in a long pulse lasting 10 microseconds (10 millionths of a second). ... [In the new experiment], protons are sent in a series of short bursts — lasting just one or two nanoseconds, thousands of times shorter — with a large gap (roughly 500 nanoseconds) in between each burst. This system, says Dr Bertolucci, is more efficient: "For every neutrino event at Gran Sasso, you can connect it unambiguously with the batch of protons at CERN," he explained.'"
Helpful but not that helpful (Score:5, Informative)
This is helpful but not that helpful. There are at this point a variety of potential explanations for what went wrong in the OPERA experiment. These include mismeasuring the tunnel length, issues with the clock calibration, and issues with the statistical analysis among other issues. It is important to note that while the OPERA group is double checking most of these issues, this experiment only really helps deal with a single problem, the statistical analysis of the neutrinos. If they are associated to individual bursts, the statistical test will be much simpler. So even if this still gets the same result, this won't be that strong evidence that there's something real going on here.
A better replication attempt is that which is being done by MINOS http://en.wikipedia.org/wiki/MINOS [wikipedia.org], the equivalent experiment at Fermilab in the US. One reason that OPERA was paying careful attention to the arrival times (when their main interest was actually in measuring neutrino oscillation) was that MINOS had earlier reported data that tentatively suggested that some neutrinos might be going too fast. Now that OPERA has done their work, MINOS is working on doing a more detailed analysis that should be out by around February.
Overall, I still think that there's a mistake here, but it is interesting to see how long this is taking to find where the mistake was. The apparent initial sprint by physicists to find the error is turning into a marathon. The data though still needs to be somehow reconciled with the fact that neutrinos from SN 1987a (a supernova that occurred close to Earth and whose light and neutrinos reached Earth in 1987 ahref=http://en.wikipedia.org/wiki/SN_1987Arel=url2html-7691 [slashdot.org]http://en.wikipedia.org/wiki/SN_1987A> had the neutrinos arrive when conventional theory predicted them, that is a few hours before the light. This isn't due to neutrinos traveling faster than the speed of light, but due to the fact that neutrinos are produced at the way beginning of a supernova in the core and then fly out with a headstart because they can easily avoid most of the matter in the star but the light takes time to get through the star. But, if the neutrinos traveled faster than light to the extent OPERA data suggests then SN 1987A neutrinos should have arrived years earlier.
There are some other possibilities that would reconcile the two claims. For example, it is possible that neutrinos actually travel faster in a denser medium. This would be really weird. It is also possible that the reactions we think produce neutrinos actually produce a very short lived tachyon which itself decays into a neutrino. This starts running afoul of Occam's razor, but would explain why one would see too much velocity in the OPERA setting but not from the supernova. This hypothesis is actually also pretty easily testable: one needs to use a shorter distance for one's neutrino detectors and see if the apparent velocity goes up.
Overall, I still suspect that this is a fluke or error of some kind. But I really hope it isn't. This could be the Michelson–Morley experiment of our error, the first anomaly which leads to a glimpse of some fantastically deeper understanding of the universe. But I really wouldn't bet on it.
Re:Helpful but not that helpful (Score:5, Informative)
This is helpful but not that helpful. There are at this point a variety of potential explanations for what went wrong in the OPERA experiment. These include mismeasuring the tunnel length, issues with the clock calibration, and issues with the statistical analysis among other issues.
The distance measurement and clock calibration were initially proposed by people outside the calibration as simple explanations, but at this point it's clear that they are simply not credible explanations. Contaldi http://arxiv.org/abs/1109.6160 [arxiv.org] suggested early on that the clocks could have been put out of synchronization by transport, but the OPERA team clarified that they were calibrated after transport, through GPS. Van Elburg, who is apparently completely ignorant of how GPS works, proposed that it could be a special-relativistic time dilation effect due to the orbital motion of GPS satellites relative to the lab frame. The distance measurement would have to be off by 20 meters in order to explain the 60 ns shift, and that's completely implausible.
All of the really obvious, stupid explanations have been ruled out -- which is not a big surprise, since 170 PhD's in the OPERA collaboration had their reputations on the line, so they were highly motivated to detect any really dumb blunders. So the remaining sources of error really are things in the general category you're referring to as statistical analysis. Some serious suggestions have been made that seem viable: (1) There could be a correlation between the direction of emission of the neutrinos and the time at which they were emitted during the 10 us beam pulse. (2) There could be a correlation between the distribution of energies in the neutrino beam and the time of emission. (3) There could be spillover from previous beam pulses. (4) There could be subtle effects in the electronics such as dead-time. Every single one of these possible errors is eliminated in the design that they're currently running, with 1- or 2-ns pulses instead of 10 us ones.
A better replication attempt is that which is being done by MINOS http://en.wikipedia.org/wiki/MINOS [wikipedia.org] [wikipedia.org], the equivalent experiment at Fermilab in the US. One reason that OPERA was paying careful attention to the arrival times (when their main interest was actually in measuring neutrino oscillation) was that MINOS had earlier reported data that tentatively suggested that some neutrinos might be going too fast. Now that OPERA has done their work, MINOS is working on doing a more detailed analysis that should be out by around February.
The trouble with MINOS is that (1) they have poorer statistics, (2) the energy is lower than the one used in CNGS (and the FTL effect, if real, is energy-dependent), and (3) OPERA's design was closely based on MINOS's, so subtle sources of error that are present in OPERA are likely to be present in MINOS as well. A better candidate for totally independent checking of the OPERA result is Tokai to Kamioka (T2K).
But, if the neutrinos traveled faster than light to the extent OPERA data suggests then SN 1987A neutrinos should have arrived years earlier.
If you believe both the OPERA result and other results at lower energy, then there is an energy-dependence in the speed that is different than that predicted by special relativity. (If neutrinos were tachyons, which is consistent with SR, then OPERA neutrinos would have been slower than SN1987A neutrinos, because tachyons go slower when you put more energy in them. This is the opposite of what is actually claimed observationally.)
There is essentially no hope for reconciling this observation with theory, unless we are in the middle of a major scientific revolution where everything is so weird that we just can't make sense of it yet -- which I don't find plausible. If neutrinos really went faster than light, then they would emit