It is now generally believed that most of the ordinary matter in the
universe resides in the form of a connected network of galaxies and gas, the so-called cosmic web.
Galaxies are embedded in a gaseous matrix (the intergalactic medium or IGM) consisting mostly of a hydrogen-helium plasma; they grow partly by accreting gas from the IGM and partly by merging with other galaxies.
The traditional picture of galaxies as "Island Universes", separated by vast
oceans of nothingness, has been superseded by a scenario where galaxies
are more like raisins in a cake, with most of the matter in the universe
being in the "dough", i.e., the intergalactic medium, as opposed to having fallen into the galaxies themselves.
Until recently, very
little was known about how the cosmic web at high redshift actually moves.
Does it strictly follow the general expansion of the universe, or can we
see the pull of gravity directing streams of gas into future galaxies and
galaxy clusters ?
My collaborators and I have been attempting to measure the kinematics of the gaseous cosmic web at redshifts (2-3.5) to answer these questions.
We observed hundreds of absorption systems that simultaneously appeared in the lines of sight to a number
of close pairs of QSOs and measured the velocity differences between the
lines of sight for each absorber, as a function of perpendicular distance across the sky.
The spatial separation between the points where the absorption was measured are large enough that these absorbing "clouds" generally must
be large scale filaments or sheets of gas, rather than galaxies.
The velocity differences (stricly speaking we only observe the one-dimensional velocities
projected along the lines of sight to the QSOs) allow us to estimate the gradients in
the velocity field between two points in space. On large enough scales
most of these gradients
are expected to be due to the stretching of the cosmic web, following the general expansion of the universe, but this cannot
be the whole story: without
local departures from a uniform Hubble flow (i.e., regions in the universe breaking away from
the expansion) there could never be any galaxies, for example.
The left panel shows schematically a flattened cosmic gas cloud, intersected by two lines-of-sight to background QSO images, where the gas will cause an absorption line in each spectrum. These absorption lines would be shifted in velocity with respect to each other because of the expansion of the universe and the gravitational contraction of the cloud.
The histogram in the panel on the right shows the observed distribution of such velocity shifts for a sample of clouds seen against the gravitationally lensed images of a high redshift background QSO. The solid line gives the predicted shape of the distribution from a toy model, where the clouds are randomly oriented gaseous pancakes expanding with the Hubble flow. The dotted line shows predictions from a more sophisticated cosmological simulation of the cosmic web.
