Speed limits - check your distance

It has been an exciting 24 hours since it became clear that the OPERA experiment at Gran Sasso in Italy may have detected neutrinos travelling faster than light. If this measurement is true, it would appear to break a fundamental part of Special Relativity that states that matter can't go faster than light. Physicists (including the ones making the announcement) are being cautious as dramatic claims require dramatic evidence.

What exactly is claimed? Well, one of the experiments at CERN in Switzerland smashes protons into a graphite target. Those collisions produce mesons that then decay into muons and neutrinos. The muons get detected at CERN. The neutrinos hardly interact with matter so stream merrily* away from the experiment through the Earth's crust. Nearly all of those neutrinos will pass through the crust and stream out into space just like the billions of neutrinos from the Sun that pass through our planet every second. A team of particle physicists then placed a detector in central Italy that sits in the path of those neutrinos to catch some of them.

Over the past few years the OPERA team have precisely measured the distance between the source of neutrinos in CERN and their detector in Italy. They have also measured the time it takes between the production of the neutrinos and their arrival at their detectors. They have made some amazingly precise measurements and there are some really neat aspects to this work such as creating modulated bunches of protons and seeing identical profiles in the detected neutrinos at the other end.

Dividing the distance by the time taken gives velocity. When they put the numbers in it seems to be faster than light. They think the neutrinos arrived 60 ns (+/- 6.9 ns from statistical and +/- 7.4 ns from systematic uncertainties) quicker than photons would have. That means they travelled about 1.000025 times faster than the speed of light. Wow!

As an astronomer I know there are some problems with this claim because a previous test did not show the same result. Back in 1987, neutrinos from the famous supernova SN1987a were detected after astronomers saw it (Update 24 Sep 10:41 BST: I shouldn't have written all this so late at night. In fact, the neutrinos were detected 3 hours before the light. Peter Coles says: "The few hours delay between neutrinos and photons is explained by the fact that neutrino emission occurs when the core of the progenitor star collapses, whereas visible light is released only when a shock wave reaches the surface of the imploding object"). Those neutrinos had not travelled faster than light. As Peter Coles says, if they had been travelling faster than light they would have arrived a few years earlier given the massive distance involved. The SN1987a result put a more precise limit on the speed of neutrinos which doesn't agree with this OPERA result. The OPERA paper does acknowledge the SN1987a measurements and notes that they were at a much lower energy - 10 MeV compared to 17 GeV. To make them both agree, the faster-than-light ability would have to be energy dependent. That is possible.

It is also possible (and as an experimentalist, I'd say more likely) that there are some subtle effects that they have not accounted for in their experiment. Ultimately there could be an error in their timing or in their distance. From what I caught of the web-cast (I didn't see all of it as I had work to do) their timing is very good so my hunch is on the distance measurement. It only has to be wrong by 18 m for the result to disappear.

They have used GPS receivers to get very precise positions and I don't doubt that. It is much more accurate than 18 m and their claim of 20 cm sounds plausible. However, the GPS receivers need to see satellites so the very precise positions are for locations on the surface. The source and detector are deep underground though. Using other means they've had to work out where their experiments are relative to those points on the surface. Unfortunately, they don't go into much detail about how those final distances (and directions) were determined.

This afternoon Chris North, Bob Watson, and I had a discussion about things that could affect their measurement of separation. They may have determined their distance vertically and assumed that "vertical" meant "towards the centre of the Earth" (or towards the centre of the ETRF2000 reference frame). However, that is not always true. Local concentrations of mass near the surface of the Earth will slightly alter this because there will give a component of gravity sideways. A large mountain to one side of you could alter "up" such that going 1.4 km "down" could actually move you off a geometric line to the centre of the Earth by a few metres (Update 24 Sep 11:41 BST: probably less than a metre over 1.4 km). If there was a mass concentration to the north, it would deflect "down" towards the north and you'd be a little closer to CERN than you thought at the bottom of your tunnel. The OPERA experiment does happen to sit to the south of the mountain Gran Sasso.

The physicists at OPERA may have taken the vertical deflection into account. If they haven't it will be a source of systematic uncertainty.

* Please forgive the anthropomorphism.
† An under-statement
‡ I am aware that ETRF includes "reference frame"

Posted in astro blog by Stuart on Saturday 24th Sep 2011 (01:26 BST) | Permalink
[an error occurred while processing this directive]
[an error occurred while processing this directive]