Dark Matter: Why Look for It Instead of Modifying Laws?

[exmarine]

Does anyone know why we are looking for dark mater - instead of say, modifying our laws of gravity to conform with observation?

 

[phinds]

Because modifying the laws of gravity, as a replacement for dark matter, has been tried and failed. Google "MOND"

 

[Vanadium 50]

People do look at modifying the laws of gravity. I think phinds overstates the case, but there are certainly regions (in particular, clusters of galaxies) where models of dark matter do much better than models of modified gravity.

 

[Bandersnatch]

Here's a quick overview of MOND:
http://www.talkorigins.org/faqs/astronomy/bigbang.html#mond

 

[exmarine]

OK thanks. Follow on: How do we know that the outer stars in spiral galaxies travel at the same speed as the inner ones? Because the spiral arms are still intact maybe?

 

[Vanadium 50]

In general, they don't.

 

[ohwilleke]

It is worth observing that any phenomenological gravity modification theory (such as MOND) that explains lots of dark matter phenomena in a parsimonious manner is valuable even if the true mechanism is dark matter, rather than a gravity modification.

This is because any such theory succinctly describes a target for dark matter models to replicate with the right dark matter halo distribution, and because the underlying mechanism of any dark matter model that can be replicated over a broad range of system size/mass with very few parameters must itself not be very elaborate either. It is easier to tweak a dark matter model to reproduce a simple equation than to try to individually fit huge numbers of data points that have that kind of subtle hidden structure.

And, if you can describe dark matter phenomena at all galactic or smaller scales with one empirically measured degree of freedom with moderate accuracy, then your dark matter model shouldn't require seven parameters related to seven independent degrees of freedom to achieve the same result (and dark matter models range from very simple to Byzantine in their complexity).

Similarly, since modified gravity theories, by definition, derive phenomena consistent with an alternate description in terms of inferred dark matter distributions entirely from the distribution of ordinary luminous matter in a system, it follows that dark matter distributions and ordinary matter distributions must influence each other in a very particular and very regularized way, through gravity and any other forces that moderate interactions between the ordinary matter sector and the dark matter sector.

Indeed, even posing the question as one or more new "dark matter" particles v. gravity or inertia modification is really somewhat misleading these days, as many of the "dark matter" models that best fit the astronomy observations involve both one or more new beyond the Standard Model fermions that are "dark matter" and also one or more new force carrying bosons that mediate self-interactions between dark matter particles in the dark sector that functionally, albeit indirectly, modify the consequences of gravitation arising from ordinary baryonic matter (i.e. ordinary matter via ordinary gravity influences the movement of dark matter fermions ==> dark matter fermions influence the movement of other dark matter fermions via "dark photons" ==> dark matter whose trajectories have been influenced by dark photons then influence the movement of ordinary matter via ordinary gravity).

The parameter space of dark matter models that have dark matter fermions without dark sector forces of their own has been particularly well combed through and mostly found wanting. Yet, I have seen numerous academic papers that demonstrate that dark matter models that do not either have one kind of dark matter fermion, or one dominant kind of dark matter fermion almost always under perform relative to simpler models in fitting the dark matter halos inferred from astronomy data. This is a problem because a model builder's first instinct when a simple model fails to fit what is observed is to move on to something a bit more complex. There are clearly limits to how far the data allows us to go down that road.

Finding simple dark matter models consistent with the quite generic set of properties that they must have to work with the highly successful at describing what is observed by astronomers λCDM model of cosmology, while simultaneously generating the right shaped halos and the right amount of local group of galaxies scale structure turns out to be remarkably hard, in part, because the most naive expected dark matter halo distribution in simple collisionless, singlet cold dark matter models (the NFW halo) is not a good match to what is observed in real galaxies.

The other problem potentially looming on the horizon for scientists is that the parameter space of dark matter candidates in the kind of relatively simple dark matter models that the limited success of theories like MOND (and its relativistic generalization TeVeS) is troublingly on the verge of being over constrained. It is all good and well to falsify a few of the leading gravity modification theories. But, what do you do when one set of observations, say the Lyman-Alpha data establish that generically any dark matter candidate must have a mass of at least 3.5 keV, and some other observational data set not yet measured (call it the "ultra-low frequency radio wave background" or what have you if you want to put it in a novel and make it sound real) is inconsistent with any dark matter candidate with a mass of more than 100 eV. Houston, we have a problem!

The example above is made up, but not far fetched. The mass parameter space of warm dark matter theories fixed by astronomy observations is only a few keV wide and then only by accepting that some observations are in mild tension with that mass range. The force carrying boson mass range in the self-interacting dark matter models that are the best fit to observations are only a couple of hundred MeV wide in light of current observations. Constraints on any kind of dark matter that interacts with any of the three Standard Model forces are also quite strict.

Improving the accuracy of cosmology simulations that dark matter models are tested into see if they reproduce something like the universe we live in not infrequently make old models that looked good appear untenable upon closer examination, while at other times they breath viability into dark matter models that had previously been ruled out when less sophisticated models were crudely calibrated to available data. So, old studies are not always reliable, even though they shape our prejudices about what to explore next.

For example, many dark matter simulations in use even today ignore almost all post-Newtonian gravitational effects even though all serious scientists agree that GR is a far more accurate description of reality than Newtonian gravity when the two are distinguishable, and most physicists in the field agree that this is a decent first order approximation, GR does add something in some important circumstances and there is not a real consensus regarding precisely how important it is to the accuracy of a simulation to include which post-Newtonian effects at a quantitative level. It might be the case, for example, that GR-specific effects matter in galactic cores, but not at galactic fringes, or visa versa.

As computing power improves, the data base of high quality astronomy observations improves, and HEP and direct dark matter detection experiments refine the parameter space of potential dark matter candidates, we are on track to have a very well defined potential dark matter candidate, or have over constrained models that require some new approach, within a decade or two, if not sooner.

Null results in direct detection experiments and HEP experiments to date, for example, have pretty much ruled out almost all of the weakly interacting massive particle (WIMP), thermal relic, cold dark matter parameter space at all masses from about 5 GeV to 1000 GeV, which investigators in the 1980s, 1990s and and the first decade of this century had been sure would contain a SUSY derived dark matter candidate. New non-SM particles of less than 5 GeV that interact via any of the SM forces are likewise largely ruled out by HEP experiments. A variety of astronomy data, meanwhile generically disfavor CDM particle masses of 100 GeV or more due to large scale structure and halo shape considerations. So, we're left looking for a particle that flies through everything on and in planet Earth almost without a trace, that has no strong force, weak force and electromagnetic interactions, that is lighter than a typical helium atom, has a well defined density per unit of space, and forms just the right shaped dark matter halos entirely as a result of gravitational interactions and its self-interactions in the dark sector that we can only infer indirectly.

But, if it is a thermal relic with a mass of tens of eV or less, it is hot dark matter that won't generate enough large scale structure in the universe and will screw up observations on the number of observed neutrino species is comic background radiation. Dark matter that isn't a thermal relic isn't quite so tightly constrained in its properties, but no longer fits the definition of dark matter for purposes of the highly successful λCDM model.

Basically, if we can't make either a keV scale warm dark matter model, or a considerably less constrained in dark matter particle mass simple self-interacting cold dark matter model with almost no non-gravitational interactions between the ordinary matter sector and the dark matter sector and a 100 MeV-ish scale force carrying boson that interacts with dark matter particles, we are over constrained and there are no sets of parameters for which any dark matter paradigm works to reproduce the data - which makes keeping any plausible option, be it a dark matter model, or a gravity modification model, open a good idea.