The TensorSubstittuition is not finding all of the eta[~mu,~mu]*l[nu](X)=l[~mu](X) substitutions.

The first term should change eta[~alpha,~beta]*l[alpha] -> l[~beta]. I made eta symmetric so the order should not matter. I notice that the command seems to work when the substitution is at the beginning of the term.  Do I need to use the new RepositionRepeatedIndicesAsIn command?

I am having trouble with the SubstitutionTensor command.  It kicks out an error message for operand with different free indices, but the operands seems to be fine.  I checked all of the terms and they all seem fine. Thank you for your help.

We first define the Christoffel symbol in terms of the metric,   .

Now we can substitute into the null condition for the Ricci tensor, .

  Do the first term

Do same thing with the first term but use the Simplify command

Simplify command does make the algebra easier, but the indices are not the same. Now, equation 16 should correspond to equation 14, but there is no combination of alphas that is consistent.  The variables alpha1 and alpha5 must be rho. One term is always wrong when I try to change the other indices.

alpha2 and alpha3 must be either mu or nu based on the first term.  But alpha6 should also be either mu or nu based on the last term, however that will make alpha2 and alpha3 either (nu and rho) or (mu and rho).  Neither combination makes all of the terms consistent with (14).  Very frustrating.