next up previous contents
Next: Non-linear Perturbation Theories Up: Gravitational Instability Previous: Gravitational Instability

Linear Perturbation Theory

In the following discussion of the evolution of the density contrast, we assume that Newtonian mechanics can be applied, an approximation which simplifies the study of the development of structures significantly compared to Einstein's theory of gravity. This assumption is well justified if we stay within the non-relativistic domain; i.e. tex2html_wrap_inline5233 and tex2html_wrap_inline5261 (Hubble radius), where tex2html_wrap_inline5263 is the size of the particle horizon and H is the Hubble parameter at time t.

The matter will be treated as an ideal fluid, which will enable us to make use of the following three equations: the conservation of mass'


the Euler equation of motion,


and Poisson's equation for the gravitational potential tex2html_wrap_inline5269 ,


The velocity field tex2html_wrap_inline5271 and the mass density tex2html_wrap_inline5273 are both functions of the physical coordinates tex2html_wrap_inline5275 and time t. The subscript tex2html_wrap_inline5275 in the equations indicate that the term in question is to be calculated with respect to the proper coordinates. But, as in equation (2.1), we would like to use comoving coordinates. The relation between these two types of coordinates is as follows:


where a(t) is the scale factor describing the scale of the Universe as it expands. The velocity field can then be written as


The first term on the right hand side is recognized as the Hubble flow, tex2html_wrap_inline5283 , and the last term is the peculiar velocity relative to the general expansion, tex2html_wrap_inline5285 . If we assume that tex2html_wrap_inline5287 and tex2html_wrap_inline5289 have small values; i.e. much less then unity, we can linearize equations (2.2)-(2.4). This assumption is well justified in the early epochs of the Universe and on large-scales, where the mass distribution is approximately homogeneous. If we combine the linearized version of equations (2.2)-(2.4), we get the time evolution equation of the mass density contrast in linear perturbation theory:


This equation is local; i.e. it is only dependent on the conditions at one point in space. It was this equation Lif deduced.

Let us now take a look at some solutions of equation (2.7), assuming tex2html_wrap_inline5291 , where p is pressure and tex2html_wrap_inline5295 is the cosmological constant. The easiest solution is for the Einstein-de Sitter model, where there is no space curvature (k=0). The Friedmann-equations then gives us the following time evolution of the scale parameter a:


If we insert this relation into equation (2.7), we get


a partial differential equation which has the solution


where A and B are constants. Based on this solution, Lif jumped to the conclusion that the gravitational instability theory could not be used to explain the formation of structures in the Universe; the growth would be too slow compared with the exponential growth in a non-expanding Universe, which was the leading model at the time.

Another solution when tex2html_wrap_inline5291 can be found by rewriting the Friedmann-equations as


where tex2html_wrap_inline5307 is the density parameter and z is the cosmological redshift give by tex2html_wrap_inline5311 (subscript ``0'' denotes the present value). At large redshifts; i.e. tex2html_wrap_inline5315 , we get the Einstein-deSitter model, and equation (2.10) is valid for all models. If tex2html_wrap_inline5317 we have an open model, and equation (2.11) will reduce to the Milne-model, where tex2html_wrap_inline5319 and tex2html_wrap_inline5321 . If we insert this into equation (2.7) we get


which has the solution


These two solutions in equation (2.10) and (2.13) tell us that the perturbations will, assuming a tex2html_wrap_inline5291 model, initially grow tex2html_wrap_inline5229 , but later, after tex2html_wrap_inline5327 , grow slower and finally end up with a constant amplitude.

So far we have ignored any pressure in the fluid, which at early epochs of the Universe is a bad assumption. If we introduce pressure, the time evolution equation or the mass density contrast will be coordinate dependent, which was not the case for p=0. If we simplify the picture by assuming an ideal gas, we can, with help of equation (2.1), deduce tex2html_wrap_inline5331 , where tex2html_wrap_inline5333 is the speed of sound. This brings the density perturbation equation (2.7) to


To see the effect of the pressure term, we can write the density contrast as a Fourier series,


where the proper wavelength is tex2html_wrap_inline5335 . The coefficients in equation (2.15) is independent of tex2html_wrap_inline5251 , which gives us the following equation for each number vector:


The source terms on the right-hand side vanishes at wavelength


which is often referred to as the Jeans length. This was the scale Jeans-02 deduced as being the maximum size a gas cloud could have without collapsing. What does equation (2.16) tell us? Let's take a look at two special cases: If tex2html_wrap_inline5339 ; i.e. the wavelengths are long compared to the Jeans length, the pressure term in equation (2.14) becomes negligible, and we can apply the zero pressure solution. At wavelengths shorter then tex2html_wrap_inline5341 , the density contrast oscillates as a sound wave, which means that a density perturbations can not grow if it's scale is less then the Jeans length. If we have perturbations with amplitude decreasing with scale, the ones with tex2html_wrap_inline5343 will be the first to become non-linear.

next up previous contents
Next: Non-linear Perturbation Theories Up: Gravitational Instability Previous: Gravitational Instability

Trond Hjorteland
Mon Jul 5 02:59:28 MET DST 1999