Measuring Short Times

We have said that the new pulsed lasers produce pulses as short as 2 femtoseconds. How do we know that? Suppose we gave you the job of measuring the length of the laser pulse, and the best oscilloscope you had could measure times no shorter than a nanosecond. This is a million times too slow to see a femtosecond pulse. What do you do?
If you cannot measure the time directly, you can be sneaky and use the uncertainty principle. Send the laser pulse through a diffraction grating, and record the spread in wavelengths, i.e., the spread in energies of the photons in the pulse. If the line is very sharp, if they are all red photons of a single wavelength and energy, then you know that there is no measurable uncertainty ΔE in the photon energies, and the pulse must last a time Δt that is considerably longer than 2 femtoseconds. If, on the other hand, the line is spread out from the near infra red to violet, if the spread in energies is from 1 eV to 3 eV, and the spread is not caused by some other phenomena (like the Doppler effect), then from the uncertainty principle you know that the pulse is only about a femtosecond long. (You know, for example, it cannot be as long as 10 femtoseconds, or as short as a tenth of a femtosecond.)
Thus, with the uncertainty principle, you can use a diffraction grating rather than a clock or oscilloscope to measure very short times. Instead of being an annoying restriction on our ability to make experimental measurements, the uncertainty principle can be turned into an important scientific tool for measuring short times and, as we shall see, short distances.