CANDELS and the Extragalactic Background Light

Is the night sky dark?

Depending on your current location and what instrument you’re using, the answer can vary wildly.

The zodiacal light on the left and the Milky Way on the right. 

(Image credit: Daniel Lopez)

In the first place, any non-city dweller can instantly report that a large area of the night sky is not completely dark: the band of the Milky Way proves this. Our own Milky Way galaxy, seen edge on, is the most well-known “diffuse” structure in the night sky – diffuse, in this context, means any astronomical source of light that is spread out from our perspective and not limited to a single object. Thus individual stars, galaxies, planets, the Sun, and the Moon are all referred to as “sources,” but everything else is referred to as diffuse emission.

Another diffuse structure in the night sky is the Zodiacal light, which is mainly visible for observers with exceptionally good nighttime seeing. The zodiacal light, like the Milky Way, looks like a faint, glowing band in the sky. Unlike the Milky Way, the zodiacal light is located in the ecliptic plane along with the planets, the Sun, and the Moon. Its brightness is concentrated near the sun, and for this reason can be seen only shortly after sunset or shortly before sunrise. For this reason, it has sometimes been called “the false dawn”. Frequently Venus will reside within the zodiacal light from our perspective. (For this reason, native Australian peoples referred to the zodiacal light as the rope that held Venus to the earth.)

The diffuse zodiacal sky brightness is produced by the solar system’s dust cloud, which reflects sunlight back to us from the ecliptic plane as a hazy glow. Similar dust disks also reside around other stars. In fact, if an alien civilization were to image our solar system from their own solar system, the zodiacal dust cloud would probably be the most prominent feature in the image, apart from the Sun itself!

Both the Milky Way and the zodiacal light appear to the naked eye as “bands” on the sky. But take the Hubble Space Telescope and point it at any apparently dark area of the sky. The first surprise you would see, as you might guess if you’ve read some of the previous posts on this blog, is that you’ll see thousands of stars and galaxies (each with hundreds of millions of stars in them). Even then, between the stars and galaxies in a deep Hubble image, there’s a faint glow present. Primarily, though not exclusively, the faint glow in the image would be zodiacal and Galactic (Milky Way) diffuse emission. Thus, while the “bands” we see are the brightest concentrations of the zodiacal and Galactic diffuse emission, the entire sky, even when looking outside of the ecliptic and the Galactic planes, is aglow with faint emission. In other words, not a single part of the night sky is truly dark if your eyes are powerful enough! 

But the story doesn’t end here. One can imagine measuring the total brightness from the Galactic and zodiacal diffuse emission and subtracting them, seeing if then, finally, the night sky would truly be dark. The details are how to do this are complicated, and lead to some uncertainty. Nevertheless, when astronomers have subtracted out the sky intensity from the Galaxy and the solar system, there is a glow remaining. This glow appears to be coming from every direction in the sky with about the same brightness. Because this residual brightness therefore probably comes from extreme distances far beyond the Milky Way or other nearby galaxies, we call this remaining sky brightness the Extragalactic Background Light, or the EBL.

Because of the uncertainties, not everyone agrees on the EBL intensity, or indeed whether it has actually been detected. If it is there, what could be producing it? Certainly, to some degree, it consists of unresolved faint galaxies that we haven’t observed individually yet, but which are so numerous that they become smeared together as a “glow” from our perspective. For example, most of the galaxies visible in the deep fields of CANDELS would have looked like an “extragalactic background light” a generation ago before they were individually resolvable by Hubble. As part of my PhD thesis, I took the deepest galaxy counts from Hubble and extrapolated them down to fainter magnitudes. Reasonable extrapolations fall short of explaining the EBL.

The edge-on galaxy NGC 5907 with its faint tidal streams. 
(Image credit: D.Martínez-Delgado et al.)

The EBL can also consist of fainter components of galaxies we have already resolved. Take, for example, some images of local galaxies which have extremely faint components. The image to the left shows a long tidal trail from a smaller galaxy being ripped apart. The interesting thing about this image is that the faint parts weren’t known about until better cameras imaged this bright edge-on galaxy (NGC 5907, which had already been well-known). So if one imagines that all galaxies have some faint structures like these, the currently unseen parts of known faint galaxies such as those in CANDELS might sum together to appear as a significant part of the EBL. Also as part of my own recent PhD work (which used some of the CANDELS data) I showed that a significant fraction of the EBL at some wavelengths can be explained by the summation of faint wings of galaxies in deep Hubble images.

Thus, even after we extrapolate the faint galaxy counts and account for the missing wings of galaxies, we still haven’t accounted for the measured EBL. So we have a puzzle. There are two solutions: (1) the subtraction of the foreground light is incorrect, or (2) there are some unexpected very faint sources producing this light. We need to look for clues in the EBL itself.

The all-sky infrared background light, from the DIRBE instrument on board the COBE satellite. (Image Credit: Michael Hauser, the COBE/DIRBE Science Team, and NASA.)

One possibility is that the EBL is coming from very distant galaxies forming their first stars. These galaxies are individually very faint, but if they formed enough stars, collectively they might have produced enough light to dominate the EBL. Using CANDELS data, we are looking to test whether the EBL comes from very distant galaxies in two ways. First, if the galaxies are very distant, they will have high enough redshifts that their light would only contribute to the long-wavelength CANDELS images. Second, the theory of structure formation makes a pretty solid prediction that the background won’t be entirely uniform, but will fluctuate. The largest fluctuations should appear on scales of about ten to twenty arcminutes. The telltale signature that the EBL is coming from very early galaxies would be to find the predicted fluctuations in the long-wavelength CANDELS images, but not find it in the shorter-wavelength images. 

This is hard. We need to be certain to remove all of the scattered light from Earthshine and nearby stars. We have to make sure that we have accurately calibrated the sensitivity of the WFC3 camera across its full field of view. Fortunately, for some of our fields, we are taking many images at different times with different shifts and rotations. So we have a lot of cross checks. But we also have a lot of work to do to get this right.

The measurement is difficult enough that observations are continuing on a number of other fronts (independent of CANDELS). One example of another project is CIBER (Cosmic Infrared Background ExpeRiment, see image to the right), which is a rocket mission that has launched from Wallops Island in Virgina as well as the White Sands missile range in New Mexico. It contains a camera specialized for diffuse, faint emission as well as a near-infrared spectrometer. Another possibility currently being explored is to mount an EBL-specialized camera on a spacecraft traveling to the outer planets, to Jupiter or beyond, where the zodiacal light is drastically reduced when one looks out of the solar system.

The Black Brant IX sounding rocket, which carries CIBER.

(Image credit: NASA)

Returning to our initial question: the night sky is not dark! With the eyes of the Hubble Space Telescope, the sky-spanning glows in the zodiacal light and the Milky Way’s emission are only the tip of the iceberg – a faint glow, with contributions from the solar system, the Galaxy, distant galaxies, and possibly the first galaxies in cosmic history, lurks everywhere you look. There is a famous thought experiment called Olbers' Paradox (really going back to the time of Kepler in the sixteenth century) which asked the question: if the universe were infinitely old and infinitely vast, why is the night sky then dark at all? Shouldn’t every dark area be a window to infinitely more distant stars which, while faint, would be so numerous that they must add up to a diffuse glow as bright as the Sun itself? There are many reasons why an absurdly high brightness is the wrong prediction, but the correct prediction is still not zero. And so the most important part of the answer is the night sky only appears dark to us because our eyes aren’t powerful enough see the trillions of galaxies within every pinhole in the blackness. Fortunately, we have the Hubble Space Telescope to make up for it.