Two weeks ago, I wrote that the recent OPERA/CERN findings that showed certain neutrinos to be moving slightly faster than light were almost certainly the result of an experimental error. I wrote that the timing and the clocks would be of specific interest to getting to the bottom of this.

I had quipped that it would be important to check General Relativistic effects on the experiment and that it would be charming if Einstien reached out from the grave and killed this himself with GR.

That looks to have happened.

The OPERA neutrino velocity result and the synchronisation of clocks

Alt link:

xxx.lanl.gov

Carlo R. Contaldi

Theoretical Physics, Blackett Laboratory, Imperial College, SW7 2BZ, London, United Kingdom

(Dated: September 30, 2011)

Abstract:

The CERN-OPERA experiment [1] claims to have measured a one-way speed of neutrinos that is apparently faster than the speed of light c. One-way speed measurements such as these inevitably require a convention for the synchronisation of clocks in non-inertial frames since the Earth is
rotating. We argue that the effect of the synchronisation convention is not properly taken into account in the analysis of [1] and may well invalidate their interpretation of superluminal neutrino velocity.

For a more lay person’s article:

Faster-than-light neutrinos face time trial

Did gravity mess with the clocks that measured particles breaking cosmic speed limit?

But one paper2, posted on 28 September by theorist Carlo Contaldi of Imperial College London, bears the distinction of being the first to challenge the experimental calculations.

The OPERA team timed the neutrinos using clocks at each location that were synchronized using GPS (Global Positioning System) signals from a single satellite. Contaldi’s paper says the group’s calculations do not take into account one aspect of Albert Einstein’s general theory of relativity: that slight differences in the force of gravity at the two sites would cause the clocks to tick at different rates.

Because of its location relative to the centre of Earth, the CERN site feels a slightly stronger gravitational pull than Gran Sasso. Consequently, a clock at the beginning of the neutrinos’ journey would actually run at a slower rate than a clock at the end. “It would reduce the significance of the result,” Contaldi says.

Dario Autiero of the Institute of Nuclear Physics in Lyons (IPNL), France, and physics coordinator for OPERA, counters that Contaldi’s challenge is a result of a misunderstanding of how the clocks were synchronized. He says the group will be revising its paper to try to make its method clearer. Autiero notes that OPERA has been careful to present its startling observations without concluding that the laws of physics have been upended. His e-mail discussion with Contaldi — being followed by dozens of other physicists — is ongoing.

Going deep

Because two clocks are needed to time the neutrinos’ journey — one at the beginning, and one at the end — they must be synchronized to within nanoseconds to get an accurate measurement, explains Toby Wiseman, a theoretical physicist also at Imperial College London. Measuring the speed of light on this journey would be much easier, because the beam can be reflected back to its origin, and the round trip timed with just one clock. “Whether they have or haven’t synchronized their clocks correctly is the crucial question,” says Wiseman.

Contaldi admits that his original analysis posted at arXiv wrongly assumed that OPERA’s timings relied on a clock being moved from one end of the beam to the other. But even synchronizing the clocks using GPS does not remove the difference in the time dilation effect, which Contaldi says could amount to tens of nanoseconds.

That effect would reduce the statistical significance — which OPERA claimed was six standard deviations — of the group’s result (five is enough to count as strong evidence in the field of particle physics). Contaldi says the additional error would reduce that number to two or three standard deviations, enough to make only a tentative claim of a faster-than-light effect.

For what it is worth, after following this, and reading the PDF linked above, I agree with Contaldi’s calculations. There is still room for debate and other potential sources of error to track down, but I strongly suspect that as this gets resolved, this will be a dominant issue.