For decades, the prime candidates for dark matter were WIMPs (weakly interaction massive particles), that interacted only via the weak nuclear force and gravity, and that had masses in the GeV to thousand GeV range. Evidence from multiple experiments essentially rules out the existence of such particles as an important constituent of dark matter.
Direct searches for WIMPS have excluded the entire parameter space of WIMP dark matter candidates, and all hints of WIMP dark matter in any given experiment have been contradicted by many other independent experiments. Astronomy data compel the conclusion that if dark matter exists, that its particles must look like 2 keV sterile neutrinos, rather than GeV scale WIMPS. Weakly interacting and SUSY particles have been ruled out in the appropriate mass ranges.
Note also that the direct dark matter searches, by pushing down the maximum possible cross-section of interaction of any heavy WIMPS to about 10^-44 v. 10^-36 or so for neutrinos, rule out Cold Dark Matter models in which the CDM particles deviate meaningfully from being sterile and collisionless. This makes astronomy simulation exclusions ruling out pure collisionless CDM conclusive for pretty much all kinds of CDM that couldn't be detected by Xenon-100.
Direct Searches For WIMPs have come up empty
The Xenon-100 experiment is the most sensitive of all of the direct dark matter detection experiments that has reported results. Xenon-100 contradicts all positive results for direct dark matter detection.
Its results from 2012
ruled out the existence of dark matter particles with the properties purportedly seen in five other experiments, only a couple of which are consistent with each other (as well as those associated with Fermi line (ca. 130 GeV) and AMS-02 experiment (of at least 350 GeV) that are looking for evidence of dark matter-antimatter annihilation).
For comparison purposes, the cross-sections of WIMP-Nucleon interaction probed are many orders of magnitude weaker than neutrino-nucleon cross-sections of interaction over almost all of the region excluded.
As Lubos Motl notes
You see various claims of this kind – not really compatible with each other – made by DAMA/I, DAMA/Na, CoGeNT, CRESST-II. An extra shape could perhaps be added to reflect the data from PAMELA, Fermi, and the newest three events from CDMS II (which wouldn't be too far from the CoGeNT potato).
On the other hand, the long lines depict the statements by the "negative experiments" that claim that all the points above their curve are excluded: SIMPLE, COUPP, ZEPLIN-III, EDELWEISS, CDMS (2010/2011: later betrayed the axis), and – most famously – XENON100. I say "most famously" because XENON100 is by far the most powerful experiment of this kind, at least among the negative ones.
The newest exclusion curve makes this priority even more obvious. Note the blue line inside the green-and-yellow (Brazil) band at the bottom (the exclusion is about a sigma stronger than expected). It is safely below all the "positive" potato ellipses and it is also well below the other exclusion curves. The contradiction between the latest XENON100 results and the "positive" experiments couldn't be stronger. Well, it could but it's already strong enough! ;-) One may say that pretty much all the preferred regions are disfavored by XENON100 at 5 sigma or more.
Note that the liquid xenon is a relatively diverse mixture of many isotopes (or do they filter which ones they use?). So the absence of signals is probably not due to some special properties of a xenon nucleus. On the other hand, the absence could be explained if the signals ultimately involved the interaction of a particle with the electrons – because xenon (unlike germanium, silicon, and all the other elements used in the experiments) is an inert gas with full electron shells and L=S=J=0 , when it comes to atomic physics. The events don't look like interactions with the electrons but there could be some subtleties. The tension between XENON100 and others seems so strong that the inert character of xenon seems "almost necessary" for me to understand the apparent xenophobia of the dark matter particle – but the de Broglie wavelength of the new particle would have to be of atomic size or longer for the vanishing atomic angular momentum to matter at all (which seems like an insanely low momentum, too). Also note that the collisions with the electrons are supposed to be "background" and distinguished from the dark-matter-like collisions with the nuclei but there could be a reason why some particle's collisions with the electrons look nucleus-like.
Each experiment that claims to have seen a dark matter particle with a particular cross-section of interaction and mass has about ten other experiments that have seen nothing in the same parameter space. The case that all of the claimed direct dark matter detections reported to date are erroneous is quite powerful. This is particularly so because the kinds of signals purportedly seen are more or less identical to the kinds of signals that would be seen if an unaccounted for source of background noise was omitted from the analysis
Yet one more sobering fact (NYU’s Neal Weiner emphasized this in his talk last week in Princeton) is that in all of these underground experiments, a failure to account for a small background will typically show up as a few extra low-energy collision candidates, which will then closely resemble what you’d expect for a low-mass dark matter particle. In other words, lightweight dark matter is what an oops! will look like.
Efforts are currently underway to build a Xenon based sensor that is more sensitive by the same factor as the leap from Xenon-10 to Xenon-100 illustrated above.
Astronomy data rule out Cold Dark Matter models
Astronomy simulations that show galactic scale structures and halo shapes different from those that cold dark matter would produce also rule out most cold dark matter in the 1 GeV and up range, and favor warm dark matter models instead with particles closer to a single kind of 2 keV mass particle that behaves like a sterile neutrino, although the exclusions of heavier and lighter dark matter particles is more definitive than the positive evidence that a 2 keV dark matter particle (give or take) could fit the data.
A good summary of the reasons that cold dark matter is experimentally excluded can be found at H.J. de Vega and N.G. Sanchez, “Warm dark matter in the galaxies:theoretical and observational progresses. Highlights and conclusions of the chalonge meudon workshop 2011″ (14 Sept 2011) http://arxiv.org/abs/1109.3187
Here are some key quotes from the abstract and body text:
Warm Dark Matter (WDM) . . . essentially works, naturally reproducing the astronomical observations over all scales: small (galactic) and large (cosmological) scales (LambdaWDM). Evidence that Cold Dark Matter (LambdaCDM) and its proposed tailored cures do not work at small scales is staggering. . . .
The most troubling signs of the failure of the CDM paradigm have to do with the tight coupling between baryonic matter and the dynamical signatures of DM in galaxies, e.g. the Tully-Fisher relation, the stellar disc-halo conspiracy, the maximaum disc phenomenon, the MOdified Newtonian Dynamics (MOND) phenomenon, the baryonic Tully-Fisher relation, the baryonic mass discrepancy-acceleration relation, the 1-parameter dimensionality of galaxies, and the presence of both a DM and a baryonic mean surface density. . . .
It should be recalled that the connection between small scale structure features and the mass of the DM particle follows mainly from the value of the free-streaming length lfs. Structures smaller than lfs are erased by free-streaming. WDM particles with mass in the keV scale produce lfs ∼ 100 kpc while 100 GeV CDM particles produce an extremely small lfs ∼ 0.1 pc. While the keV WDM lfs ∼ 100 kpc is in nice agreement with the astronomical observations, the GeV CDM lfs is a million times smaller and produces the existence of too many small scale structures till distances of the size of the Oort’s cloud in the solar system. No structures of such type have ever been observed. Also, the name CDM precisely refers to simulations with heavy DM particles in the GeV scale. . . . The mass of the DM particle with the free-streaming length naturally enters in the initial power spectrum used in the N-body simulations and in the initial velocity. The power spectrum for large scales beyond 100 kpc is identical for WDM and CDM particles, while the WDM spectrum is naturally cut off at scales below 100 kpc, corresponding to the keV particle mass free-streaming length. In contrast, the CDM spectrum smoothly continues for smaller and smaller scales till ∼ 0.1 pc, which gives rise to the overabundance of predicted CDM structures at such scales. . . .
Overall, seen in perspective today, the reasons why CDM does not work are simple: the heavy wimps are excessively non-relativistic (too heavy, too cold, too slow), and thus frozen, which preclude them to erase the structures below the kpc scale, while the eV particles (HDM) are excessively relativistic, too light and fast, (its free streaming length is too large), which erase all structures below the Mpc scale; in between, WDM keV particles produce the right answer.
See also in accord: S. Tulin, et al. “Beyond Collisionless Dark Matter: Particle Physics Dynamics for Dark Matter Halo Structure” (15 Feb 2013) http://arxiv.org/abs/1302.3898
As is well known, the collisionless cold DM (CCDM) paradigm has been highly successful in accounting for large scale structure of the Universe. . . . Precision observations of dwarf galaxies show DM distributions with cores, in contrast to cusps predicted by CCDM simulations. It has also been shown that the most massive subhalos in CCDM simulations of Miky Way (MW) size halos are too dense to host the observed brightest satellites of the MW. Lastly, chemo-dynamic measurements in at least two MW dwarf galaxies show that the slopes of the DM density profiles are shallower than predicted by CCDM simulations.
A number of more recent papers have highly constrained the mass range for warm dark matter and have disfavored models with multiple kinds of dark matter.
deVega and Sanchez, for example, offer up, "Dark matter in galaxies: the dark matter particle mass is about 2 keV" (Submitted on 2 Apr 2013) http://arxiv.org/abs/1304.0759
Warm dark matter (WDM) means DM particles with mass m in the keV scale. For large scales, for structures beyond 100 kpc, WDM and CDM yield identical results which agree with observations. For intermediate scales, WDM gives the correct abundance of substructures. Inside galaxy cores, below 100 pc, N-body classical physics simulations are incorrect for WDM because at such scales quantum effects are important for WDM. Quantum calculations (Thomas-Fermi approach) provide galaxy cores, galaxy masses, velocity dispersions and density profiles in agreement with the observations. All evidences point to a dark matter particle mass around 2 keV. Baryons, which represent 16% of DM, are expected to give a correction to pure WDM results. The detection of the DM particle depends upon the particle physics model. . . . So far, not a single valid objection arose against WDM.
, for example, C. Watso, et al. “Constraining Sterile Neutrino Warm Dark Matter with Chandra Observations of the Andromeda Galaxy” http://arxiv.org/abs/1111.4217
(10 Jan 2012) (WDM mass capped at 2.2 keV); R. de Souza, A. Mesinger, A. Ferrara, Z. Haiman, R. Perna, N. Yoshida, “Constraints on Warm Dark Matter models from high-redshift long gamma-ray bursts” (17 Apr 2013) http://arxiv.org/abs/1303.5060
(WMD mass at least 1.6 keV); D. Anderhaldena, et al. “Hints on the Nature of Dark Matter from the Properties of Milky Way Satellites” (12 Dec 2012) http://arxiv.org/pdf/1212.2967v1.pdf
(mixed CDM/WDM models disfavored); J. Viñas, et al. “Typical density profile for warm dark matter haloes” (9 Jul 2012) http://arxiv.org/abs/1202.2860
(models with more than one WDM species disfavored); Xi Kang, Andrea V. Maccio, aaron A. dutton, "The effect of Warm Dark Matter on galaxy properties: constraints from the stellar mass function and the Tully-Fisher relation" (8 April 2013) http://arxiv.org/abs/1208.0008
(WDM mass of more than 0.75 keV and consistent with 2 keV).
Weakly interacting particles light enough to be dark matter are ruled out experimentally
Particles that interact via the Standard Model weak force with masses of less than 45 GeV have been excluded for many years by precision electro-weak measurements at LEP (and this will soon rise to 62.5 GeV as Higgs boson decays are analyzed).
This means that any dark matter particles must interact with ordinary matter, if it interacts at all other than via gravity, via a force other than the three Standard Model forces (although it could conceivably couple to the Higgs boson).
SUSY can't supply particles that could be dark matter
Searches for SUSY particles at colliders like the LHC have likewise established that there are no sparticles with masses below the hundreds of GeV
in fairly simple MSSM and NMSSM SUSY models.
None of the SUSY models propose particles consistent with experimental data describing dark matter phenomena. Any viable explanation of dark matter effects needs to come from a source outside SUSY and SUGRA theories.
Experiments do not rule out the possibility that ephemeral and unstable heavy SUSY particles exist, but even if they do, they cannot the source of dark matter. SUSY theories, generically, have no features which could help solve this most glaring of remaining unsolved problems in fundamental physics.