If dark matter where a lot of photons, then we could detect it. For example the experiments to detect the cosmic microwave background radiation https://en.wikipedia.org/wiki/Cosmic_microwave_background can detect a very small amount of light that is the leftover of the radiation produced soon after the big bang. We have very good estimations of how many photons are out there.
I don't understand, how would you detect photons that are going in a direction orthogonal to you in intergalactic space? They have to be coming toward you for you to detect them, right?
An implicit assumption is that the universe is isotropic and homogenous at that scale. Meaning, it looks the same from any point and any direction.
So you might not detect photons in one particular direction, but we should assume that they should be going in all directions. Otherwise that theory is just moving the goal from "where's this energy" to "why is it pointing towards that".
CMB is consistent with, every direction is the same, so you would have to say "photons are going everywhere in this range, but towards that in this segment of the spectrum".
But isn't it a hell of a lot less of a stretch to imagine there are more photons in other parts of the universe than here, than to imagine there is some sort of mysterious dark matter permeating all of the universe? If we're going to be unable to interact with whatever dark matter is to test may hypothesis, we might as well propose the hypothesis that requires the least radical changes to our fundamental theories of physics, no?
Assume have an equation like `Energy = 0`, where `Energy = Mass + Photons`. It works great for almost everything, we can get the curvature of space by `f(Energy)` where `f(Photons) = 0`.
However there's this one experiment that says the answer is `Energy = 42`. `f(x)` is still useful, everything else works.
We can just amend the equation to be: `Energy = Mass + Photons + DarkMatter` where `DarkMatter = -42` and `f(DarkMatter) = 0`.
Or we can keep the E=M+P but change the way f works to be: `f(Photons, X) = Y` where for almost any value `f(Photons, X') = 0` and in this one case `f(Photons, X") = 42`.
To me it's clear that the first approach is less complicated. The first approach raises one question, "why -42?" The second approach raises "why does X?" exist and "why 42?"
======
By the way, you are mistaken, we do interact with Dark Matter, that's how we detected it (Bullet Cluster). Gravity does interact with it.
