As an experimental physicist, I go back to basics:
Here is a double slit experiment a single photon at a time:
The top panel shows dots from single photons coming in at the slits. The others the slow accumulation by which an interference pattern appears.
That is what the experiment shows, dots like one would expect from single particles, localized at a specific (x,y). From a single photon no wave nature can be guessed at. It is the statistical accumulation that shows the classical wave interference, similar to water waves and acoustic waves.
What we have experimentally is a probability distribution of where to find a photon thrown at these two specific slits.
One talks of the "wave nature" of photons and particles at the individual particle level referring to this probability distribution. The particle is not spread out all over the screen , it has a specific trajectory after the fact of passing through one or the other slit. Which slit it passes through depends on the boundary conditions of the solution of the quantum mechanical problem "photon and slits" which gives the probability distribution.
No matter how close to the slits in the z direction one registers the photons the trajectories will have this behavior, of a particle leaving the two slits. Except the behavior of the aggregate is not the classical behavior of little balls thrown at the slits.
So the quantum mechanical way of looking at it is that a particle is described by a wavefunction with a probability distribution for any interaction it will have . The particle interacting with the two slits has a probability distribution which displays its wave nature. The individual trajectory displays its particle nature.
There exist which way experiments that show how the boundary conditions affect the interference patterns after the slits, in this case with electrons:
Overall, the results suggest that the type of scattering an electron undergoes determines the mark it leaves on the back wall, and that a detector at one of the slits can change the type of scattering. The physicists concluded that, while elastically scattered electrons can cause an interference pattern, the inelastically scattered electrons do not contribute to the interference process.
So back to your question:
In the double-slit experiment, we emit a photon that is in a state of superposition (wave form) which travels through both slits to interfere with itself.
The photon ( or other particle) is described by a quantum mechanical wave function that gives the probability of finding it in a specific trajectory in space. Psi(x,y,z,t).
When we measure which slit it went through, it "collapses" to a particle at that point removing the interference from the other wave.
Collapse is a funny word. The wave function is not a balloon. A value is picked from the probability distribution that will give a particle trajectory for it.
From that point to the detector wall, does the photon now go back into a wave form to travel from the collapsed point to the detector wall?
Due to the boundary condition of two "slits" the wavefunction, and therefor the probability distribution, changes and carries phase information that the two slits imposed.
Is a photon always in a state of superposition while traveling through space?
The correct statement is that the photon/particle is described by a probability distribution in space and time.