At what point did stars, galaxies, and galaxy clusters make their first appearance in the universe? There’s now good evidence that galaxy clusters were starting to form about 650 million years after the big bang. So galaxies must have begun forming earlier than that, and the first stars even earlier.
The evidence consists of the discovery that a known bright galaxy at z≈8 has 4 dimmer companions within a radius of about 10 million light-years. The known galaxy was originally found using the Wide Field Camera 3 (WFC3) of the Hubble Space Telescope, by means of the Lyman-break technique. The was done as part of a search for z≈8 galaxies, known as the Brightest of Reionizing Galaxies (BoRG) survey. The region in which the bright galaxy was found was designated BoRG58.
The Lyman-break technique is based on the fact that photons having wavelengths less than the Lyman-α length of 121.6 nm (medium ultraviolet) are easily absorbed by neutral hydrogen gas. Consequently, most of the light from hot young stars is extinguished by neutral hydrogen, since such stars are brightest at wavelengths shorter than Lyman-α. The wavelength of Lyman-α photons emitted at z≈8 is stretched by a factor of z+1 to 1094 nm, in the infrared range. So in order to identify bright galaxies at z≈8, images are made through a series of filters that pass photons on either side of 1094 nm. Objects that seem to disappear in images made through filters that pass only light with wavelengths below 1094 nm are probably at the desired redshift.
Four bright objects presumed to be at z≈8 were identified early in the survey process. Computer simulations have indicated that at that redshift the most massive galaxies (which are presumably the brightest) should have a number of less massive, less luminous companions. So the search was extended around each of the four candidates, using longer exposure times to detect fainter objects. 17 less luminous objects were detected with substantial confidence. Field BoRG58, which had the best original candidate, turned out to have 4 less luminous companions of the candidate at the best level of confidence.
It hasn’t yet been possible to obtain direct spectroscopic evidence for the actual redshift of the detected objects. This is probably because so much of the available light from those objects has been extinguished by neutral hydrogen. However, the detection of fainter z≈8 dropout objects near the original candidate provides additional evidence that the objects are not spurious dropouts (which can occur with low probability).
The computer simulations suggest that the halo mass (which includes dark matter) around the brightest object should be in the range 400 to 700 billion M⊙. (The Milky Way’s halo is estimated to be more than 1 trillion M⊙.) There should also be additional smaller halos nearby about 100 billion M⊙, hosting the fainter dropouts. By the present time the system should have grown into a cluster with halo mass 200 trillion M⊙, a fairly typical cluster size.