Newtonian gravity
Attrictive force F=GMm/r² between "test mass" m and a large mass M. Here, r is separation between M and m and F is in the direction r (or, F = GMm r/r3).

Integrating from r to infinity ¥ we obtain the gravitational potential energy V=-GMm/r.

Now we imagine a large enough spherical volume of radius R (~10Mpc) so that space can be considered homogeneous and the force acting on a particle m at R from the interior volume acts as if all the mass M=4/3prr³ within the volume is a point mass R distant and the rest of the universe exerts no net force on m (the rest of the universe cancels for r>R -- shells dr with r>R do not give a net force on a particle at R).

The Energy Equation
The total energy U of a test particle m at r will be the sum of the potential V and kinetic T=½mv² (v=dr/dt) vis U=T+V:
(Also, mass M = density times volume = rV = 4/3prr³ )

U = ½mv² - GMm/r = ½mv² - 4/3pGrmr²

Which gives the evolution of the separation of r from the origin or, invoking the cosmological principle, the separation r between any two particles...we move to a comoving coordinate system, a system carried along with the expansion of the universe. Since the expansion is uniform, the relationship between r and a comoving distance x can be written

r = a(t) x    or loosely as    r = a x.

The homogeniety ensures a(r,t)=a(t).

The coordinate grid expands with time as the universe expands, but local inertial reference frames remain at fixed locations in the comoving system. Here, a(t) is the scale factor that determines how physical separations change with time. If a(t) doubles, the separation of all local inertial frames doubles.

Now rewrite the total energy using r = a(t) x = a x    resulting in

U = ½m(da/dt)²x² - 4/3pGrma²x²

Then multiply by 2/ma²x² to get the Friedmann equation

[(da/dt)/a]² = (8pG/3)r - kc²/a²

where k = -2U/mc²x². This is the standard form of the Friedmann equation and is identical to that obtained with General Relativity. (But the GR derivation is much more satisfying!)

In this equation k must be independent of x like all the other terms (homogeneity) so while U is constant for a given particle m it changes with separation x, Uµx².

The k=-2U/mc²x² is often called the curvature and in the Friedmann universe retains its value throught its evolution.


Critical Density
Now consider a sphere of comoving radius R (expanding along with the universe) enclosing a mass M. The "Energy equation" U/m=v²/2-GM/R applied to a test mass m tells us for U=0 v=0 at R=infinity and the system is "critical" or just unbound.

If U<0 the system is "bound" and the velocity goes to zero at a radius Rmax=-GM(m/U)and the particle would fall back.

If U>0 the system is unbound. For any given velocity v there exists a corresponding critical mass that makes U=0:

Mcrit=v²R/2G
or, with M=rV=4/3prR³ we can write the mass density as M/V:
3Mcrit/4pR³=3v²/8pGR²
and in an expanding universe, v=HR on the comoving sphere so
3Mcrit/4pR³=3H²/8pG.
The quantity 3H²/8pG does not depend on R ... This quantity is called the critical density of the universe...
rcrit=3H²/8pG
If the average density of the Universe is less than or equal to critical it will expand forever, and the Universe is open. If greater, the Universe is closed. A Universe with critical density is flat, infinite, and the expansion rate approaches zero at t=infinity. If H=71 km/s/Mpc then rcrit=9.45E-27 kg/m³ (about E-26 kg/m³ or E-29 gm/cm³).
The Cosmological Constant L
Rewrite the energy equation as
v²=2GM/R+2U/m.
Suppose the Energy U is related to m via U~mc²... so U/m=-kc² or-kc²/2 (choice depends on whether you like k to be 0,±1 or 0,±½) ...then
v²=2GM/R-kc².
For a flat Universe k=0, k=+1 for a closed and -1 for an open Universe. (In GR other values of k are not allowed.) Note that for k=+1 if R>2GM/c² v² goes negative, not allowed... this is a closed universe.

Now we can put the cosmological constant Lambda into our energy equation:

v²=2GM/R-kc²+LR².

The effect of the new term is to make the expansion rate increase exponentially with time. The new term dominates when

L>2GM/R³ or L>8prG/3.
and R(t)=RoeHt/to
Useful for inflation in the early Universe or the new results from SNeIa and WMAP etc.
Einstein introduced the constant to keep the universe from expanding (biggest blunder?) by setting k=0 and L=-8prG/3 which results in v=0. Of course Einstein used real relativity! He added a term L to his field equation:
Gmn=Tmn+gmnL
Gamow in My World Line 1970 (Viking, New York) writes "when I was discussing cosmological problems with Einstein, he remarked that the introduction of the cosmological term was the biggest blunder he ever made in his life."

The latest results imply L=0.73 and the matter density 0.27 (total is unity) which gives an older Universe than if L were zero, since the expansion was slower in the past, relieving the "Universe is too young" problem.


Now use the first law of thermodynamics dE + p dV = T ds
applied to E = mc² = 4/3pr

Then a change of E in time dt is

dE/dt = 4prc²(da/dt) + 4/3pa³(dr/dt)c²

while the volume changes by

dV/dt = 4pa² da/dt

and if dS=0 (reversible) the first law gives the fluid equation

{dE/dt + dV/dt = 0} times 3/4pa³c²   or
dr/dt + 3(da/dt)/a (r+p/c²)=0.

The density evolution is due to volume increase (adot/a times r) and loss of energy due to work done as the volume of the universe is increased (adot/a times p/c²) = the energy converted into gravitational potential.

Now if we only knew how P varies, the equation of state ... P=P(r,T??)


If we differentiate (d/dt) the Friedmann equation we get

d/dt[(da/dt)/a]²=2[(da/dt)/a] [a(d²a/dt²)-(da/dt)²]/a² = 8/3pG(dr/dt) + 2kc²(da/dt)/a²

and substituting for rhodot from the field equation we have

(d²a/dt²)/a - [(da/dt)/a]² = -4pG(r+p/c²) + kc²/a²

so from the Friedmann equation again we see

(d²a/dt²)/a = -4/3pG(r+3p/c²)

and yes, that is +3p/c², any pressure actually decreases (decelerates) the expansion!
(The curvature k cancelled out.)


Most GR equations use "natural units" where c=1 so the Friedmann equation becomes

[(da/dt)/a]² = (8pG/3)r - k/a²

then k has units t-2 (and time is measured in metres).


The Hubble Law
Hubble observed v=Hr where since r=a(t)x
v=[(da/dt)/a]dr/|r| so in the spirit of the Friedmann equation
H=(da/dt)/a.

Since H=H(t) we denote the value of H we observe today Ho and, since Ho>0 we know the universe is expanding. (This is in terms of our Friedmann universe, but the Hubble expansion can be explained in simpler ways and usually is.) But using H² for (adot on a) squared we have

H²=(8pG/3)r - k/a²    (c²=1)

So H=H(t) and H should decrease with time as the expansion of the universe is slowed by gravitational attraction. If the Friedmann model is right and if there is no extra "dark physics".

The redshift used to justify the big bang is a result of the time dependent scale factor adot on a... the time between emission and absorption of a photon is dt=dr/c,

dl/lo=[(da/dt)/a][dr/c]=[(da/dt)/a]dt=da/a

where dl=l-lo, lo is the "lab" wavelength, l is the (redshifted) wavelength, and integrating:

ln(l/lo)=ln(a) i.e. lµa.

We usually talk in terms of a "redshift parameter" z=Dl/lo (if you don't see a subscript it is there!) then "redshift" z is related to the scale of the universe by

1+z=l/lo=ao/a.

So if the wavelength has doubled, z=1 and the universe is twice as large as when the light was emitted. (The universe was half the present age when the light was emitted.)


The Equation of State for -- Dust
Dust? Well, what we mean here is any material that exerts negligible pressure so we can set p=0. This is obviously the simplest case but it is not trivial since it is nearly exact for stars and galaxies which really only interact gravitationally (no star collisions even when galaxies collide!) and not too bad for a universe that has cooled to the point where atoms are neutral and collisions infrequent. Good for dust too, but there is not too much of that in the universe. We use the term "dust" for nonrelativistic matter like the stellar interior people use "z" for "metals", anything not H or He!

The Equation of State for radiation
Relativistic particles, photons included, exert a pressure p=rc²/3. For photons this is called radiation pressure. Neutrinos may also be important? (Neutrinos have to be included if they have rest masses of order of an eV.) Now we can solve a couple of interesting cases....
Friedmann Universe for k=0 and p=0
The fluid equation

dr/dt + 3(da/dt)/a (r+p/c²)=0

for dust becomes dr/dt + 3r(da/dt)/a=0

i.e. d(ra³)/dt=0   (zero times a³ if you wonder where that factor went!)

Integrating, ra³=constant. Hardly surprising, the density goes as one over the volume! But this is useful since as we are only interested in a ratio of adot to a we can rescale a(t) and it is usual to choose a(present)=1. Then the physical and comoving coordinate frames coincide at the present time and we can write

r=ro/a³

and if we substitute this into the Friedman equation with k=0 we get

[(da/dt)/a)]² = [8pG/3] ro/a³

[da/dt]² = [8pGro/3] [1/a]

which obviously :) has a solution a µ tq... hmmmm.... the LHS would be adot squared goes as t to the minus 2q-2 and the RHS 1/a goes as t to the minus q so setting the powers equal we have 2q-2=-q or q=2/3. (Wow!) So a(t) goes as t2/3 and since (we set now=to) a(t)=[t/to]2/3 and

r(t) = ro/a³ = ro(to/t)².

The universe expands forever but the expansion rate decreases with time: H goes as 2 over 3t

[(da/dt)/a] = H = 2/3t.


Friedmann Universe for Radiation
If the equation of state is p=rc²/3 the fluid equation

dr/dt + 3[(da/dt)/a] (r+p/c²)=0   becomes

dr/dt+4[(da/dt)/a]r = 0.

Solving this as before we find now r=ro/a4   instead of /a³. Applying this to the Friedmann equation and guessing a solution :) we get a=a(t)=(t/to½) so

r(t) = roa(t)-4 = ro(to/t)2.

The expansion is slower if radiation dominated (extra deceleration due to pressure) but as before density falls off as t². The density falls as the fourth power of the scale factor because of increasing volume (three of the four) and redshift of the radiation (the other one.) The drop can also be thought of as the work done p dV by the pressure as the univeerse expands.


Mixtures of Dust and Radiation

The total density is rtot = rdust + rrad   where rdustµa-3   and rradµa-4.

Usually one or the other will dominate, but if radiation dominates note that the rapid fall with scale will eventually cause the "dust" to become important and after a long enough time dominate. At present dust dominates and so will continue to do so for the rest of time.


You want real cosmology? Try the NCSA (want a job in supercomputing?) Cosmos in a Computer site. ²³