Image: National Institute of Standards and Technology, public domainWhy moving clocks run slow, with one triangle
A muon lives for 2.2 microseconds on average, and in that time light travels 659 metres. Muons are made by cosmic rays about 15 kilometres up in the atmosphere, and most of them reach the ground. Something about moving fast makes their internal clock run slow, and the amount by which it runs slow can be worked out with one right-angled triangle.
A clock made of light
Build a clock out of two parallel mirrors a distance \(h\) apart, with a pulse of light bouncing between them. Every time the pulse hits the lower mirror, the clock ticks. The time between ticks, measured by someone sitting next to the clock, is
\[ \Delta\tau = \frac{2h}{c}. \]Now let the clock move past you at speed \(v\), with the mirrors at right angles to the motion. You see the pulse go up to the top mirror and down again, but while it does so the mirrors move sideways. The path you see is two slanted lines, each the hypotenuse of a right-angled triangle whose vertical side is \(h\) and whose horizontal side is half the distance the clock moves during one tick.
Call the time between ticks, as you measure it, \(\Delta t\). During that time the clock moves \(v\,\Delta t\), so the horizontal side of each triangle is \(v\,\Delta t/2\). The light moves at \(c\) in your frame too, because that is what Einstein's second postulate says, so each hypotenuse has length \(c\,\Delta t/2\). Pythagoras gives
\[ \left(\frac{c\,\Delta t}{2}\right)^2 = \left(\frac{v\,\Delta t}{2}\right)^2 + h^2. \]Multiply out, collect the terms in \(\Delta t\), and use \(h = c\,\Delta\tau/2\):
\[ \Delta t^2\,(c^2 - v^2) = c^2\,\Delta\tau^2, \qquad \Delta t = \frac{\Delta\tau}{\sqrt{1 - v^2/c^2}} = \gamma\,\Delta\tau. \]That is the whole derivation. The factor \(\gamma = 1/\sqrt{1 - v^2/c^2}\) is called the Lorentz factor, and \(\Delta\tau\), the time between ticks measured by the clock itself, is called the proper time. The moving clock's ticks are spaced further apart in your time than in its own. At everyday speeds the factor is so close to one that nothing is noticed: at 300 metres per second, the speed of an airliner, \(\gamma - 1\) is \(5 \times 10^{-13}\). At \(0.6c\) it is 1.25, at \(0.8c\) it is 1.667, and at \(0.995c\) it is 10.
Why it is not a trick of the mirrors
The first objection every student raises is that this is a property of a light clock, and a wristwatch or a heartbeat might not care. The answer is the first postulate. Put a light clock and a wristwatch side by side on a train and let them agree. If the wristwatch ran at a different rate from the light clock when seen from the platform, then the two clocks would disagree on the train too, since the number of ticks of one between two ticks of the other is a plain count that everyone can agree on. The passenger could then tell that the train was moving by watching two clocks, and the first postulate forbids exactly that. Whatever runs slow for the light clock runs slow for every clock, including the chemistry of a living cell and the decay of a particle.
The second objection is that if the clock on the train runs slow seen from the platform, the clock on the platform should run slow seen from the train, and that both cannot be true. Both are true. Each observer compares one moving clock with two of their own synchronized clocks, placed where the moving clock passes, and the two observers disagree about whether those two clocks are synchronized. Unit 3 of the masterclass spends most of its length on this point, because once it is understood, the rest of special relativity is bookkeeping.
The muons
Cosmic rays, mostly protons, hit nuclei at the top of the atmosphere and produce pions, which decay into muons within a few metres. The muons are left moving downward at speeds very close to \(c\). A muon at rest decays with a mean lifetime of 2.197 microseconds, measured in the laboratory with muons brought to a stop in a block of material. The value is in the tables of the Particle Data Group to seven significant figures.
Suppose the muons moved at \(0.995c\) and did not experience time dilation. In 2.197 microseconds they would travel 656 metres, and after 15 kilometres, which is 23 mean lifetimes, the fraction left would be \(e^{-23}\), about one in ten billion. Hardly any would reach the ground. With time dilation at \(\gamma = 10\), the lifetime in the ground frame is 22 microseconds and the mean path is 6.6 kilometres, so after 15 kilometres about \(e^{-2.3}\), one in ten, survive. The ground is hit by about one muon per square centimetre per minute, and the particle detectors in every physics teaching laboratory count them.
The cleanest version of the measurement was made by David Frisch and James Smith in 1963, with a scintillation counter on the summit of Mount Washington in New Hampshire, about 1,900 metres above sea level, and then at sea level. They selected muons in a narrow range of speed around \(0.995c\) and counted about 560 per hour at the top. Without time dilation, the 6.4 microseconds of flight to sea level is 2.9 lifetimes, and they would have counted about 30 per hour at the bottom. They counted about 410. The muons' own clocks had advanced by only about 0.7 microseconds, and the factor between the two numbers, 8.8, was the Lorentz factor of their muons within the uncertainties.
The muon's own point of view
From the muon's rest frame the story is the same and the roles are swapped. The muon's clock is fine. What it sees is the atmosphere rushing up at \(0.995c\), and a layer of air that is 15 kilometres thick in the Earth's frame is contracted to 1.5 kilometres in the muon's frame, by the same factor \(\gamma\). A 1.5 kilometre trip at nearly \(c\) takes 5 microseconds of the muon's time, about two lifetimes, and one in ten survive. The two descriptions agree on the one thing that can be checked, the number of muons that hit a detector on the ground, and that agreement is the content of the theory. Length contraction and time dilation are two faces of the same transformation, and Unit 4 writes down the transformation that contains them both.
If you want to see the effect for yourself, the triangle is all you need, plus a calculator for \(\gamma\). At \(0.9c\) the factor is 2.29; at \(0.99c\) it is 7.09; at \(0.9999c\) it is 70.7. The protons in the Large Hadron Collider move at a speed for which \(\gamma\) is about 7,250, and if they were unstable they would live 7,250 times longer than at rest. They are stable, which is lucky, because the muons in the same ring, used for calibration, decay before they get round once.


