The Speed of Neutrinos: Part 1

Map showing the path of the neutrino beam as it passes from CERN to the Gran Sasso Laboratory. Map courtesy of Google

Map showing the path of the neutrino beam as it passes 730 km through the Earth from CERN to the Gran Sasso Laboratory. The width of the beam is exaggerated - it is only around 6 km wide when it reaches the detector. Map courtesy of Google

You may have read recently about an experiment at CERN which found that neutrinos appeared to travel faster than light. I was going to write about the significance of this result (should it be confirmed) – but when I looked at what the experimenters actually did I was filled with admiration for their work. And I felt that it might be nice to just describe what they did. And in doing this I hope you will see that although it is quite possible that they are right, there is also a distinct possibility that they have made a mistake. You can read the researchers own account of their work here.

The neutrinos are generated in CERN using a proton accelerator. Every 6 seconds a ‘kicker magnet’ sends protons out of a ring in two pulses each 10 microseconds long, and separated by 50 milliseconds. Through a complicated series of interactions the protons generate neutrinos of a type known as muon neutrinos which travel in roughly the same direction as the protons. Neutrinos barely interact with matter at all, and can travel through the Earth and barely notice it. Of the 1020 neutrinos generated in this experiment over a period of three years, only around 16,000 were detected in the Gran Sasso Laboratory, deep underground in Italy – 730 km away. In other words only roughly 1 neutrino was detected for every 10,000,000,000,000,000 that were emitted from CERN. How could the transit time of this one neutrino in 10 million billion possibly be timed?!

Well first the researchers synchronised clocks at the two laboratories using a clock on a GPS satellite as a common reference source. They took great care over this and think their clocks agree within 1 nanosecond. This may sound incredible, but your computer is probably processing instructions at a rate of at least 1 per nanosecond (if you have a 1 Gigahertz processor) so you can think of this synchronising the clocks of two computers within one processor cycle – difficult but imaginable.

Then the researchers measured the distance as best they could using GPS satellites and a ground survey. They found the distance to be 730534.61 metres with an uncertainty of only 20 centimetres – and they could easily detect a 7 centimetre shift in this distance after the L‘Aquila Earthquake.

So how could the transit time of this one neutrino in 10 million billion possibly be timed?! It can’t. Instead the researchers first measured the current of protons headed towards the target where the neutrinos were created. They did this by wrapping a transformer around the beam which gave a voltage when the proton ‘current’ pulsed through the transformers coils. They reasoned that the rate of production of neutrinos ought to be proportional to the rate at which protons hit the target. When they had measured the shape of the proton pulse averaged of the millions of pulses in three year, they tried looking for neutrinos arriving at the detector 2.436801 milliseconds later – the transit time for light. By restricting their view in this way they were able to ignore the possibility that a neutrino from the rest of the universe would arrive in this tiny time window. They expected that as the neutrinos were detected, they would see a pulse shape slightly delayed with respect to time that light would have taken (had it been able to travel through 730 km of solid rock!). To their great surprise they found that although the 16111 neutrinos they detected over three years did have the same shape as the average proton pulse shape, the neutrino pulse occurred slightly earlier than they would have expected.

The graph below shows the average proton pulse shape from each of the two 10 microsecond pulses which occur every six seconds

The average shape of the two pulses of protons that generate the neutrinos averaged over all the pulses used in three years.

The average shape of the two pulses of protons that generate the neutrinos averaged over all the pulses used in three years.

The graph below shows the detected neutrino arrival times alongside the expected arrival time (in red) based on the average proton pulse shape assuming the neutrinos travelled at the speed of light. The neutrino distribution is in advance of its expected arrival time by around 1 microsecond. The researchers can understand most of that in terms of signal delays in equipment. But there is a stubborn 0.06 microseconds that they are unable to explain – and that is why they have publicised their work – basically to ask for help.

The averaged arrival times of neutrinos

The dots show the measured arrival times of neutrinos at the detector compared with the expected timing (shown in red) if the neutrinos travelled at the speed of light. Notice that the rise in neutrino count occurs a little earlier than expected.

Now there are lots of places that they could have made mistakes, and that is what I expect they will find eventually. But I just want to express my admiration for their work.

  • First of all for the sheer chutzpah of trying to do such an audacious experiment! Detecting neutrinos is hard in itself, but this is just amazing!
  • Secondly, it is an example of how precision measurement can reveal the physical details of a phenomenon in the same a way a microscope shows visual details: if they hadn’t tried to time the pulses so accurately, they would never have been able to meaningfully ask the question about whether neutrinos travelled faster than light.
  • And finally, although I think they will find an error, I really hope they don’t, and that their result stands: it would be great to have an experiment that nobody on Earth understood – how exciting would that be!

Tags: , ,

4 Responses to “The Speed of Neutrinos: Part 1”

  1. Emma Says:

    Thank you very much Michael! (I’ll be honest … I’ve been waiting for your summary before answering “Heh, you’re a physicist, do you think they’ve broken the speed of light?”)

  2. The Speed of Neutrinos: Part 2 « Protons for Breakfast Blog Says:

    […] Protons for Breakfast Blog Making sense of science « The Speed of Neutrinos: Part 1 […]

  3. Jeff Says:

    Confused says: there is already a strict limit based on the neutrinos that accompanied supernova 1987a arriving in step with the photons is this somehow different?

    • protonsforbreakfast Says:


      Good Point. I was not aware of that work which is based on the simultaneous arrival time of (a) neutrinos and (b) light from a supernova explosion which occurred roughly 160,000 years ago. You are so clever and well read. As I mentioned (and everyone else too) the CERN-Gran Sasso experiment is probably a mistake. You and I both know how complex these experiments are and how subtle oversights can manifest themselves in strange ways. However playing Devil’s Advocate for the moment I could point out that in the recent terrestrial experiment, every aspect of the process was under human control. And although the distance was tiny compared to the SN1987A observations, the timing and the distance measurements should both have been good enough.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: