Dark Matter Still Hasn't Been Directly Observed

"Dark matter phenomena" are real, based upon the consensus interpretation of astronomers and astrophysicists looking at astronomy evidence.

Overwhelming evidence shows the existence of "dark matter phenomena" evidenced primarily by the dynamics of galaxies and galaxy clusters, by the gap between all discernible sources of ordinary matter and the amount of matter inferred using a Newtonian approximation of gravity, and by the gap between all discernible sources of ordinary matter and the amount of matter inferred from gravitational lensing of light using a weak field approximation of General Relativity. Cosmic background radiation patterns also support the conclusion that dark matter phenomena are real.

But Stacy McGaugh in his latest blog post at Triton Station, reminds us that there have been no credible and reliable detections of dark matter itself in the Milky Way, even though many hypotheses have been searched for with a variety of means.

Direct detection experiments have come up empty and have probed dark matter particle mass ranges from a bit below 1 GeV to about 1000 GeV. 

Macroscopic dark matter candidates like primordial black holes and MACHOs have been ruled out. 

No missing momentum signals in collider experiments up to the 13 TeV energies of the Large Hadron Collider have revealed any anomalies that are good dark matter candidates produced in these reactions.

Astronomy searches for dark matter annihilation signatures have come up empty and where there have been anomalies have other explanations that don't require new physics or dark matter particles.

The dynamics of dark matter also, generally speaking, rule out heavy dark matter candidates with particle masses in excess of 1 TeV, or for that matter, in excess of a KeV mass.

The lack of direct detections or detections of decay products doesn't in and of itself rule out the dark matter hypothesis. It just tightens the parameter space for dark matter to something that doesn't interact via Standard Model forces and is stable on a time frame of many billions of years of mean lifetime or more.

But, the problem you get when you impose those conditions is that you can't explain why dark matter isn't observed to have the NFW halo distribution that dark matter like that should have. A self-interaction of dark matter with dark matter only (SIDM) could partially remedy that problem, although efforts to model that and fit parameters for that self-interaction have largely been unsuccessful.

Some dark matter models are ruled out by evidence from Big Bang Nucelosynthesis.

Most importantly, you can't explain why dark matter phenomena can be accurately predicted from the distribution of ordinary matter in a system in a very tight correlation if it has no non-gravitational interactions with ordinary matter.

This is why I strongly favor gravity or fifth force explanations for dark matter phenomena. An extremely light bosonic dark matter particle candidate, however, starts to blur the line between a fifth force and a dark matter particle explanation. 

In all other aspects of physics, forces are carried by bosons (i.e. particles with integer intrinsic angular momentums like 0, 1, 2, etc.) that we sometimes simplify to think of as force fields, while the stuff that we think of as matter in a plain English sense of the word, is made up of fermions (i.e. particles with intrinsic angular momentum of 1/2, 3/2, etc.).