5-dimensional space with a
4-dimensional FRW spacetime as subspace
[1]We will wish to find five-d spaces with 4-d subspaces which are familiar to us. For example the metric for a five-dimensional space with a four-dimensional subspace which is homogeneous and isotropic:
ds2 = - dt2 + [Rt2/(1+kr2)2](dr2 + r2 dW2) + 5t2 5r2 dX52
The Einstein tensor components for spherically symmetric 4 and 5 d spaces have a simple relationship to each other (see appendix for derivation).
For the Grr tensor component: the relationship between the r parts of the 4 and 5-d components:
5rGrr =
4rGrr + 2(1/Rt2)(1/Rr2)(5'r/5r)(qr'/qr)
,
which for this metric with Rr = (1+kr2) and qr = r/(1+kr2) is:
5rGrr = 4rGrr + (2/Rt2)(1+kr2)2(5'r/5r)(1-kr2)/[r(1+kr2)]
= 4rGrr + (2/Rt2)(5'r/5r)(1-k2r4)/r
For the 5-d tensor component to be homogeneous, it must be independent of r, and since the 4-part is homogeneous and r-independent it must be that:
(5'r/5r)(1-k2r4)/r
is r-independent. That is:
(5'r/5r)(1-k2r4)/r = c ,
i.e:
5'r/5r = cr/(1-k2r4).
5'r/5r = c
[r/(1-k2r4)]
ln5r = c
[r/(1-k2r4)]
+ K
The integral can be easily performed using the substitution R = r2 , dR=2rdr.
The solution is:
5r
= K [(1-kr2)/(1+kr2)]-c/4k
so
that the 5-d metric for a homogeneous isotropic space is (squaring 5r
makes the exponent -c/2k)
ds2 = -dt2 +Rt2/(1+kr2)2 (dr2+r2dW2) + 5t2 K [(1-kr2)/(1+kr2)]-c/2k dX52
It
is amazing that this solution can be obtained so simply - for an appreciation
of how simple our method is, see
Chatterjee paper: he bravely solved this using the usual long and tedious
calculations.
The other results derived there are also easily found in this way.
[1] Edited excerpt from my file: “Part IV GR book”