Thursday, January 21, 2016

Messengers from the Dark Age

Astrophysicists dream of putting radio
telescopes on the far side of the moon.
[Image Credits:]
An upcoming generation of radio telescopes will soon let us look back into the dark age of the universe. The new observations can test dark matter models, inflation, and maybe even string theory.

The universe might have started with a bang, but once the echoes faded it took quite some while until the symphony began. Between the creation of the cosmic microwave background (CMB) and the formation of the first stars, 100 million years passed in darkness. This “dark age” has so far been entirely hidden from observation, but this situation is soon to change.

The dark age may hold the answers to many pressing questions. During this period, most of the universe’s mass was in form of light atoms – primarily hydrogen – and dark matter. The atoms slowly clumped under the influence of gravitational forces, until they finally ignited the first stars. Before the first stars, astrophysical processes were few, and so the distribution of hydrogen during the dark age carries very clean information about structure formation. Details about both the behavior of dark matter and the size of structures are encoded in these hydrogen clouds. But how can we see into the darkness?

Luckily the dark age was not entirely dark, just very, very dim. Back then, the hydrogen atoms that filled the universe frequently bumped into each other, which can flip the electron’s spin. If a collision flips the spin, the electron’s energy changes by a tiny amount because the energy depends on whether the electron’s spin is aligned with the spin of the nucleus or whether it points in the opposite direction. This energy difference is known as “hyperfine splitting.” Flipping the hydrogen electron’s spin therefore leads to the emission of a very low energy photon with a wavelength of 21cm. If we can trace the emissions of these 21cm photons, we can trace the distribution of hydrogen.

But 21 cm is the wavelength of the photons at the time of emission, which was 13 billion years ago. Since then the universe has expanded significantly and stretched the photons’ wavelength with it. How much the wavelength has been stretched depends on whether it was emitted early or late during the dark ages. The early photons have meanwhile been stretched by a factor of about 1000, resulting in wavelengths of a few hundred meters. Photons emitted towards the end of the dark age have not been stretched quite as much – they today have wavelength of some meters.

This most exciting aspect of 21cm astronomy is that it does not only give us a snapshot at one particular moment – like the CMB – but allows us to map different times during the dark age. By measuring the red-shifted photons at different wavelengths we can scan through the whole period. This would give us many new insights about the history of our universe.

To begin with, it is not well understood how the dark age ends and the first stars are formed. The dark age fades away in a phase of reionization in which the hydrogen is stripped of its electrons again. This reionization is believed to be caused by the first star’s radiation, but exactly what happens we don’t know. Since the ionized hydrogen no longer emits the hyperfine line, 21cm astronomy could tell us how the ionized regions grow, teaching us much about the early stellar objects and the behavior of the intergalactic medium.

21 cm astronomy can also help solve the riddle of dark matter. If dark matter self-annihiliates, this affects the distribution of neutral hydrogen, which can be used to constrain or rule out dark matter models.

Inflation models too can be probed by this method: The distribution of structures that 21cm astronomy can map carries an imprint of the quantum fluctuations that caused them. These fluctuations in return depend on the type of inflation fields and the field’s potential. Thus, the correlations in the structures which were present already during the dark age let us narrow down what type of inflation has taken place.

Maybe most excitingly, the dark ages might give us a peek at cosmic strings, one-dimensional objects with a high density and high gravitational pull. In many models of string phenomenology, cosmic strings can be produced at the end of inflation, before the dark age begins. By distorting the hydrogen clouds, the cosmic strings would leave a characteristic signal in the 21cm emission spectrum.

CSL-1. A candidate signal for a cosmic
string, later identified as two galaxies.
Read more about cosmic strings here.
But measuring photons of this wavelength is not easy. The Milkyway too has sources that emit in this regime, which gives rise to an unavoidable galactic foreground. In addition, the Earth’s atmosphere distorts the signal and some radio broadcasts too can interfere with the measurement. Nevertheless, astronomers have risen up to the challenge and the first telescopes hunting for the 21cm signal are now in operation.

The Low-Frequency Array (LOFAR) went online in late 2012. Its main telescope is located in the Netherlands, but it combines data from 24 other telescopes in Europe. It reaches wavelengths up to 30m. The Mileura Widefield Array (MWA) in Australia, which is sensitive to wavelengths of a few meters, has started taking data in 2013. And in 2025, the Square Kilometer Array (SKA) is scheduled to be completed. This joint project between Australia and South Africa will be the yet largest radio telescope.

Still, the astronomers’ dream would be to get rid of the distortion caused by Earth’s atmosphere. Their most ambitious plan is to put an array of telescopes on the far side of the moon. But this idea is, unfortunately, still far-fetched – for not to mention underfunded.

Only a few decades ago, cosmology was a discipline so starved of data that it was closer to philosophy than to science. Today it is a research area based on high precision measurements. The progress in technology and in our understanding of the universe’s history has been nothing but stunning, but we have only just begun. The dark age is next.

[This post previously appeared on Starts With a Bang.]


Uncle Al said...

" If dark matter self-annihilates," then the Tully-Fisher relation for spiral galaxies must change over redshift. It doesn't, it doesn't - including (central) black hole capture. Milgrom acceleration is immune. Its universal source is bench top testable in existing apparatus. Look - for it also sources baryogenesis.
Moon, "dark" side"

Lunar LOAR construction requires massive importation of structural materials plus labor round trip. Moon pristine hard vacuum atmosphere yields to burning propellants (H_2O; CO_2, HCl).
Galactic noise
Water dielectric loss
Atmospheric opacity, anhydrous.
Atmospheric opacity, including water

JimV said...

Typo patrol: "might give us a peak at cosmic strings" - unless this is a clever pun which went over my head, it should be "peek".

Note: my typo density is much higher than what I see here.

Sabine Hossenfelder said...


Thanks, I've fixed that :)

Uncle Al said...

Science 351(6270) 30, 249 (2015)
Lunar atmosphere mean free path exceeds escape trajectory.
"the rocket exhaust expended from each Lunar Module temporarily doubled the total mass of the natural lunar atmosphere"

United States Space Science Program: Report to COSPAR
National Academies, 1967
QB500 U54 10th Mtg. 1967 c. 1 (Dewey Decimal System)
"Lunar Atmosphere," p. 168

nicolas poupart said...

Question for the pros: where is now the energy lost by the stretched photons ?

Sabine Hossenfelder said...

Answer from the pros: energy conservation is an expression of time-translation invariance (Noether's theorem). An expanding background is not invariant, consequently energy doesn't have to be conserved.

Lucy M said...

Dear Sabine - I'm hearing a lot at the moment about the 'information paradox'. I understand how the paradox itself is constructed. But I don't really understand why information cannot be lost forever beyond an event horizon and into the hole. It must be so obvious because no-one ever sees a need to explain why that violates the laws of physic. Or why the remainder of the Quantum Mechanical universe on our side should be left corrupted, since the information that is destroyed is entirely to do with the mass that is also destroyed. What is a word "destroyed", why can't we say it a mechanical one-way (positive feedback) divergence to threshold mutual causal isolation, and beyond.
The wordiness is just tongue in cheek, but I thought it would help if I stated what I don't understand as a 'why can't we do it like this' example, so you can easily see what my problem is, should you be minded to grant my request. Which isn't taken for granted but PLEASE PLEASE PLEASE you explain things clear that I understand everyone else just tries to mess me up cos I'm perwittier.

Sabine Hossenfelder said...


I'm somewhat short on time at the moment. Please try this older post first and let me know if it doesn't answer your question.

Don Foster said...


“The universe might have started with a bang, but once the echoes faded it took quite some while until the symphony began.”

It is interesting and engaging to see once again how enterprising physicists must be in finding obscure phenomenological processes that allow exploring very remote physical landscapes. One must apparently also be incredibly patient to await experimental results to satisfy curiosity.

And it is comforting to find this universe of thirteen billion years ago to be a fairly sophisticated place with hydrogen atoms doing their familiar dance and the laws of physics already presiding.

As to the “echoes,” was that simply a turn of phrase or was there actually something physically akin to echoes in the earlier phase?

And one more question regarding the emissions of these 21cm photons. Were they emitted in random a direction or was that direction somehow determined by the specifics of the atomic collision?

Thank you.

Lucy M said...

“The universe might have started with a bang, but once the echoes faded it took quite some while until the symphony began.”

Everyone give or take everyone else has been going around regurgitating that line verbatim (one recital each, per unique context in 'human resource', very likely). Unique as in not shared with everyone else give or take everyone else else, and not showing up to the same show twice either.

It's effortless for any one person, effortlessly synchronized in the participating linguistically short-trouser sharing/caring & phraseological bulimia support group across the corridor from physics in the upchuck-therapeutics laboratory. Try regurgitating that pygmy pet.

Anyway, it is effortless because it's all handled in social instinct because it's been going on for eons. Unfortunately it hasn't quite caught up with the Internet age (or Bronze age more accurately) and really does not work very well online