Cosmology Primer: The Early Universe
Given our understanding of the current state of the universe, and our knowledge of the appropriate laws of physics, we can extrapolate backwards in time to say what the early universe must have been like. Fortunately, we can then use current observations to test whether such an extrapolation is valid; the answer is that it is remarkably accurate.
The cosmic microwave background (CMB) is the leftover radiation from the Big Bang. When the universe was much smaller it was much hotter and denser. The ordinary matter that today resides in the form of stars, gas, and dust was packed together at incredible densities and temperatures, so much so that electrons moved freely rather than being attached to individual atomic nuclei (much like conditions at the center of a star). This hot plasma gave off copious amounts of radiation, just like any other hot object, and we can detect that radiation today. The plasma was also opaque; any photon would only travel a short distance before bumping into a free electron. At 370,000 years after the Big Bang, the temperature had dipped below about 3,500 Kelvin, cool enough for electrons to recombine with nuclei to make atoms, and the universe suddenly became transparent. As the universe expanded and photons redshifted to longer wavelengths, the radiation subsequently cooled to about 2.7 Kelvin, which is what we see today. The CMB was first discovered by Arno Penzias and Robert Wilson in 1965.
The CMB provides a snapshot of what the universe looked like at the moment of recombination, when electrons became attached to nuclei. What it looked like was something extremely smooth; fluctuations in density from place to place were only about one part in 100,000. But we can detect these fluctuations. The image reproduced here is the famous map from the Wilkinson Microwave Anisotropy Probe satellite, compared with the first detection of CMB fluctuations from the earlier COBE satellite. Blue regions are slightly colder than average, red regions are slightly hotter. These changes in temperature from place to place are referred to as "anisotropies," from the technical term meaning "changing as we look in different directions." Of course the relic radiation from the Big Bang fills all points of space, no matter where we are in the universe; but since we can only observe it from our location, we perceive the the changes from place to place as projected onto the sphere of the sky.
The smooth, slightly perturbed early universe visible in the CMB anisotropies grew into the lumpier universe of stars and galaxies we see today. This should come as no surprise. The hot and cold spots of the CMB correspond to regions of slightly higher or lower density than average. In the regions that were overdense, the pull of gravity brought matter closer together, further emptying out the regions that were less dense. The evolution of the universe under the influence of gravity thus acts to increase the contrast of the matter distribution, from a nearly featureless plasma to an intricate collection of galaxies. This process takes longer over larger distances, which is why the universe remains approximately smooth on very large scales.
There is a treasure trove of information contained in the CMB fluctuations. In particular, statistical properties of the fluctuations depend on two things: the original primordial perturbations from which they arose, and the recipe of ingredients in our universe that controls the subsequent evolution of the perturbations between early times and now. Remarkably, an extremely simple specification of primordial perturbations works very well -- simply imagining that the perturbations are (on average) of equal strength at all distance scales. From this guess, and the observed fluctuations in the CMB sky, we can derive very tight constraints on interesting cosmological parameters. In particular, the CMB provides independent support for the ideas that there is more matter in the universe than can be accounted for by ordinary atoms (thus implying the need for dark matter), and that there is more total energy than be accounted for by matter along (thus implying the need for dark energy).
The CMB provides a valuable picture of what the universe was like when it was 370,000 years old, but it also forms a barrier past which we can't see -- because the plasma of the early universe was opaque, we can never obtain a direct image of any event at earlier times. Nevertheless, we can obtain indirect information about the universe back to when it was a few seconds old, by observing the primordial abundances of light elements.
Just as the universe at times earlier than 370,000 years was too hot for electrons and nuclei to stick together to form atoms, at times earlier than a couple of minutes after the Big Bang it was so hot that protons and neutrons could not stick together to form atomic nuclei. As it expanded and cooled, the early universe was a nuclear reactor -- protons and neutrons combined to make light elements such as deuterium ("heavy hydrogen," with one proton and one neutron per nucleus), helium (two protons and two neutrons, or occasionally two protons with only one neutron), and lithium (three protons and four neutrons). The nuclei themselves would have liked to continue this process of fusion, combining all the way to form iron (the most stable nucleus), but the rapid expansion of the universe soon made the nuclear plasma too thin to sustain further reactions.
This process of primordial nucleosynthesis took place while the universe was between a few seconds and a few minutes old, and only one-billionth its current size. Since that time, further nuclear evolution has taken place inside stars and supernovae, providing us with the rich variety of elements in the universe today. However, by looking to regions that have been relatively untouched by exploding stars, we can infer the primordial abundances of deuterium, helium, and lithium. These abundances depend sensitively on two things -- the amount of ordinary matter in the form of protons and neutrons, and the expansion rate of the universe when it was just a few seconds old. We obtain perfect agreement with observations if two things are true -- the amount of ordinary matter is much less than the total amount of matter we deduce in the current universe (thus providing evidence for dark matter), and the expansion rate is exactly as predicted by Einstein's theory of general relativity (thus assuring us that our best theories can be safely extrapolated back to early times).
The agreement between the predictions of the Big-Bang cosmological model and our observations of primordial light-element abundances was by no means guaranteed -- it is certainly conceivable that the observed amounts of deuterium, helium and lithium could not be explained by conventional general relativity for any value of the matter density. The fact that we do get good agreement, indicating that we understand the behavior of the universe when it was only a few seconds old, is one of the most profound achievements of modern science.