Low-frequency radio cutoff
A Socratic walk-through of the low-frequency radio cutoff — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why would astronomers put a radio telescope on the far side of the Moon to hear the longest wavelengths?
Radio astronomy is the branch that got to stay at home. Optical astronomers fled to mountaintops and then to orbit; radio astronomers built dishes in Cheshire and New Mexico and did fine, because radio waves stroll through cloud and weather without noticing them. So why propose the most awkward observatory site in the solar system — the far side of the Moon, permanently out of contact with Earth — purely to listen below about ten megahertz?
The reflex answer is "to escape interference". That is half of it, and the weaker half. Interference is a nuisance you can legislate against. What sits below ten megahertz is not a nuisance. It is a wall.
Reasoning it through
REASONING #Ask what the upper atmosphere actually is. Above roughly eighty kilometres, solar ultraviolet strips electrons from air molecules faster than they can recombine, leaving a permanent partly-ionised layer: free electrons and heavy ions, neutral overall. Now put a radio wave into that. The wave's electric field pushes the electrons — the ions are thousands of times heavier and barely stir — and moving charges radiate. So the medium answers back. The question is how fast it can answer.
Try a thought experiment. Displace all the electrons in a slab of that plasma sideways by a small distance x, leaving the ions behind. You have made a charged sheet at each face, of surface density n e x, where n is the electron number density. Between them the field is E = n e x / ε₀, pulling the electrons back with force -n e² x / ε₀ — proportional to displacement, opposite to it. That is a harmonic oscillator, and its angular frequency follows directly:
ω_p² = n e² / (ε₀ m_e)
So a plasma has a natural ringing rate of its own, set only by how many electrons are in it. Now the payoff. If a wave arrives slower than that — lower frequency — the electrons have ample time to rearrange each cycle, cancel the wave's field, and hand the energy back. Nothing propagates; the wave is reflected. Arrive faster than the plasma can respond and the electrons lag hopelessly behind, the medium is effectively transparent, and the wave sails through.
Put in numbers. Using the recalled constants e = 1.602 x 10^-19 C, ε₀ = 8.854 x 10^-12 F/m and m_e = 9.109 x 10^-31 kg, take a daytime peak electron density of about 10^12 per cubic metre. Then n e² ≈ 2.57 x 10^-26 and ε₀ m_e ≈ 8.07 x 10^-42, giving ω_p² ≈ 3.2 x 10^15, so ω_p ≈ 5.6 x 10^7 rad/s and f_p = ω_p/2π ≈ 9 MHz. At night, ionisation recombines and the density falls by roughly a factor of ten; the frequency goes as the square root, so the wall drops to about 2.8 MHz.
That is the first half of the answer. From the ground you cannot observe below a few megahertz at any hour, nor below about nine for most of the day, because the sky's own light at those wavelengths is mirrored back into space before it reaches you. Above the wall the ionosphere is not innocent either — its patchiness scintillates the signal and wrecks the phase calibration an array depends on — but that is a difficulty, whereas below the wall there is nothing to detect.
Why care about frequencies that low? Neutral hydrogen's 21-centimetre line sits at 1420 MHz at rest, and cosmic expansion divides that by (1+z). At redshift 30 it arrives at 46 MHz; at 100, at 14 MHz; at 150, at 9.4 MHz. The era before the first stars — the one epoch nobody has yet observed — radiates precisely into the band the ionosphere hides.
The refuting observation: ionosondes measure the local critical frequency continuously. If a ground array ever recorded a celestial source at one megahertz while the ionosonde overhead read eight, the plasma-cutoff account would be dead. What is seen instead is a cutoff tracking the measured density hour by hour, sinking at night and rising through the solar cycle.
The analogy
THE ANALOGY #A plasma is like a pool of water you are trying to push a wave through with your hand. Wave slowly and the water simply slides around it — the surface stays flat, nothing travels, the effort comes back at you. Move faster than the water can get out of the way and a wave leaves your hand and crosses the pool. There is a rate the medium can keep up with, and below it your push accomplishes nothing but sloshing.
water's response time depends on depth and geometry, whereas a plasma's depends on one number only — free electrons per cubic metre — and not at all on the size or shape of the region.
Clarifying the model
THE MODEL #The Moon does two separate jobs, and they are worth keeping apart.
The far side blocks Earth. Some 3,500 kilometres of rock stands between the site and every broadcast transmitter, over-the-horizon radar and switching power supply on the planet. That is the interference half, and it matters because terrestrial leakage also propagates best at long wavelengths.
But the Moon's real gift is the absence of an ionosphere. It has only a wisp of exosphere, and on the sunlit side a thin photoelectron sheath, so the cutoff essentially vanishes and the window opens downward. How far down? Here the mechanism proves its own generality: the solar wind is itself a plasma, about five electrons per cubic centimetre near Earth, or 5 x 10^6 per cubic metre. Frequency goes as the square root of density, so from 10^12 per cubic metre giving 9 MHz, a density lower by 2 x 10^5 gives a frequency about 450 times lower — roughly 20 kHz. Below that the interplanetary medium itself is opaque, and no site in the solar system will help.
One correction worth making explicitly: the ionosphere is not absorbing this radiation, and calling it a "blocking layer" invites that misreading. Below the critical frequency the energy is turned around essentially intact — which is exactly why the layer that hides the early universe is the layer that lets a shortwave broadcast reach another continent.
A picture of it
THE PICTURE #How to readThis is a packet-layout diagram repurposed as a frequency axis — read the numbered ranges as megahertz, not as bits. The leftmost block lies below even the night-time plasma frequency of about 2.8 MHz, so it never reaches the ground at any hour. The middle block sits between the night and day values, opening after dark and closing at sunrise. Only the right-hand block, above the daytime 9 MHz derived earlier, is reliably available — and the redshifted dark-ages hydrogen signal falls in the two blocks on the left.
What became clearer
WHAT CLEARED #A plasma has one natural frequency, fixed by its electron density alone, and it is transparent above that and mirrored below it. Earth wears such a plasma as a permanent shell, so the long-wavelength sky is not faint from the ground — it is absent. The far side of the Moon is not merely a quieter site; it is a site where the wall has been removed, leaving only the far thinner plasma of the solar wind as a floor.
Where to go next
ONWARD #- Why the same reflection that hides the cosmic signal is what makes shortwave broadcasting work.
- How an array of thousands of simple dipoles, with no rigid dish at all, becomes the practical design at 30-metre wavelengths.
Key terms
TERMS #| Term | What it means |
|---|---|
| Plasma frequency | the natural oscillation rate of free electrons about their ions, sqrt(n e²/ε₀ m_e)/2π, above which a plasma transmits and below which it reflects. |
| Critical frequency | the plasma frequency of the densest ionospheric layer overhead, measured routinely by ionosondes. |
| Redshifted 21-centimetre line | neutral hydrogen's 1420 MHz emission, stretched by cosmic expansion into the low-frequency band. |
Every term the collection defines is gathered in the glossary.