Fix definition of elemental mole fraction

[speth]

@thomasfiala, I was looking at the definition that was introduced for the elemental mole fraction in 4bc4fda, and I don't think it's correct. I thought it looked okay at the time, but thinking about it again, I think it should be invariant under composition changes which conserve atoms, e.g. a mixture with composition CH4:1, O2:2 should have the same elemental mole fractions as CO2:1, H2O:2. Also, like the relationship between species mass and mole fractions, the elemental mass and mole fractions should satisfy Z_mass,m = Z_mole,m * W_m / sum(Z_mole,j * W_j), which the existing definition does not. The new definition implemented in this PR satisfies both of these constraints.

[thomasfiala]

Ray, thanks a lot for changing this! I can't recall what I originally thought when I tried to derive an elemental mole fraction, but that clearly wasn't right. I looked through your implementation and found it to be correct. Also, the example and relation you mention perfectly makes sense.

I might have had a misunderstanding about the term 'elemental mole fraction', as I would intuitively use mole in combination with molecules but not with atoms. Since the term is not used very often (Google gives just 21 hits, some of them pointing to Cantera): Do you think that it might be helpful to be more descriptive in the documentation, like elemental mole fraction (which is the number of atoms of a certain element divided by the total number of atoms)?