Hydrogen line

A hydrogen atom in its ground state has two slightly different energy levels, depending on whether the electron’s spin points the same way as the proton’s or the opposite way. The parallel arrangement sits a tiny bit higher. When an atom drops from the parallel to the antiparallel state it gives up that difference as a photon at 1420.405751768 MHz, a wavelength of about 21.1 cm. That is the hydrogen line, and it is one of the most precisely measured frequencies in physics.

For any single atom the transition is absurdly slow: the spontaneous lifetime of the upper state is on the order of ten million years. It works anyway because there is so much neutral hydrogen between the stars. Collisions keep re-populating the upper level, and along any line of sight through the Galactic plane there are enough atoms that the emission becomes one of the brightest features in the radio sky. The line was predicted in 1944 and first detected in 1951 by Ewen and Purcell with a horn antenna, which is still more or less the antenna we recommend people start with.

What it looks like in a receiver

You will not see a sharp spike at 1420.4058 MHz. The gas is moving, so every cloud along the line of sight is Doppler shifted by its own radial velocity, and the Earth is moving too. Toward the Galactic plane the result is a broad, lumpy bump a few hundred kilohertz wide, often with two or three distinct peaks, each one a body of gas at a different distance and speed. Rule of thumb: 1 km/s of velocity moves the line by about 4.74 kHz.

To read velocities off that bump you subtract the Earth’s motion first, which is what the local standard of rest correction does and what our Doppler calculator works out for your site and date. LAB-012 walks through a first spectrum end to end.

Why amateurs can detect it

The line is strong in brightness temperature terms, tens of kelvin and more in the plane, and it is extended, so it fills even a wide beam. That is why a horn on a tripod and a $37 SDR can see it, as we show in LAB-009. Point sources at the same frequency are a different story entirely.

The band around it, 1400 to 1427 MHz, is allocated to passive use only by the ITU, so there are no licensed transmitters in it. That keeps the line’s neighbourhood quieter than most of the spectrum, though not free of interference; see RFI. It is also the bottom edge of the water hole, which is why we were listening there in the first place. LAB-005 has that reasoning in full.