Quite right! Orbitals are not physical observables (expectation values of linear Hermitian operators) and hence are not observable via experiment.
This is a difference map; the subtraction of a density distribution constructed from the fit of a basis set (the presumed orbitals) to the observable structure factors (the observed scattering) - however obtained - from a model distribution constructed from the superposition of atoms in chosen (and hopefully appropriate) states. The later distribution is entirely arbitrary.
This is a difference map; the subtraction of a density distribution constructed from the fit of a basis set (the presumed orbitals) to the observable structure factors (the observed scattering) - however obtained - from a model distribution constructed from the superposition of atoms in chosen (and hopefully appropriate) states. The later distribution is entirely arbitrary.