Two species of dark matter?



At this point in time, evidence for the existence of dark matter has accumulated in many ways:

  • it affects galactic rotation curves
  • plays a major role in cosmology, and the evolution of structure in the universe
  • is predicted in copious amounts by gravitational lensing on a wide range of scales
  • influences the dynamics of galaxy clusters

to name a few.

There are many known candidates for dark matter particles: WIMPs, axions, WISPs, neutrinos, etc (in fact, even bricks, though some other considerations would exclude them).

The question then is: Why do we expect that only one type of dark matter particles is responsible for phenomenological dark matter?

For example, $\Lambda$CDM cosmology, the standard cosmological model, requires dark matter to be cold (slow, non-relativistic), which is used to constrain the possible properties of dark matter particles. However, this doesn't actually imply, that dark matter is cold for all the astrophysical systems. For example, galactic halos could be made of warm dark matter, and halos of dwarf galaxies could be made of cold dark matter.

One might of course say that one-species model is the simplest one. The counter-argument would be that in reality there well may be many species. This in turn might have profound implications for astrophysical models.

To summarize the question: Is there any good reason, preferably supported by observations, to think that only one species of dark matter is present in all the models currently used?

Alexey Bobrick

Posted 2013-10-25T15:45:17.243

Reputation: 1 641

We can't really expect one type or several types of dark matter. Dark matter could be a wide variety of particles, so long as they do not collide with our matter and other energies we can easily observe, yet still have energy. – trevorKirkby – 2013-12-08T06:10:23.533

@someone-or-other: True, wide, but as indicated by the answers, not too wide a variety, e.g. masses should be comparable. So it can be many kinds of particles rather than one, but with little phenomenological consequences. – Alexey Bobrick – 2013-12-08T17:11:53.297

Since there is no evidence of the nature of dark matter maybe there is an entire zoo of DM particles much like there is for "normal" matter and that there are "dark" forces that only effect dark matter and this creates and entire dark matter universe of existence that we are completely unaware of except that the we sense the gravity it generates. – HisDivineShadow – 2016-09-28T21:34:14.717

The idea is interesting, and yet we actually do know that compared to the ordinary matter the dark matter is much much less self-interacting if at all. – Alexey Bobrick – 2016-09-29T14:02:38.527

3Very nice question! – Dilaton – 2013-10-25T16:09:44.143

1Couple of things. What are WISPs, and what do you mean by 'bricks'? Also, the word 'cold' in cold dark matter means that at the time dark matter decoupled, it was non-relativistic (slow compared to the speed of light). LCDM predicts very well structures on large scales, it's a mismatching on smaller scales that motivates people to think of things like warm/hot dark matter since warm things have less structure on small scales. – astromax – 2013-10-25T17:49:37.927

3I do think this is a good question though. It is quite possible that the dark matter component to the universe is actually more than one type of massive particle, perhaps one that interacts weakly, and one not. Adding forces by which particles can interact through would add additional avenues for the transfer of energy into and out of these components. That dark matter is one 'species' of particle is simply the most natural thing to think of first. – astromax – 2013-10-25T17:53:27.490

1@astromax, thanks for your input! WISP, according to wiki, for example, stands for Weakly Interacting Sub-eV Particles such as axions. Bricks is more of a joke. However, if you do have objects, which weight as ordinary bricks do, appropriately spaced, they would dynamically behave more or less exactly as dark matter would (no forces except for gravity). Then, "cold (slow)" implies non-relativistic velocities, though I will add it for clarity. – Alexey Bobrick – 2013-10-25T18:43:44.743

@Alexey Would you also add a link to WISPs since I'm not familiar with them? Also, I think I've heard of some N-body simulations which contain more than a single dark component. If/when I find them I'll try to write up an answer to this question. – astromax – 2013-10-25T18:48:15.630

1@astromax, thanks also for pointing out the scales. I actully had in mind that dark matter, which is not cold, the large scale structure would look much more blurred, than it is, and therefore concluded that it is particularly important for small structures that dark matter is cold, and less so for larger structures. Can you comment on where does the discrepancy come from? Otherwise, your ideas look reasonable. I would be more than happy to see them in a bit more elaborated form as an answer. – Alexey Bobrick – 2013-10-25T18:48:21.283

@astromax: Thanks again! WISPs don't have their own page, but I will add a link to the page, where they are mentioned. – Alexey Bobrick – 2013-10-25T18:49:57.297



Hot dark matter would be made from very light, fast moving particles. Such particles could not possibly be gravitationally bound to any structure, but rather would be dispersed all across the universe.

But dark matter is always "found" (or "inferred") either gravitationally bound to some visible structure (e.g. weak lensing detection of dark matter associated with colliding galaxy clusters / flat rotation curves of spiral galaxies / abnormal velocity dispersion in galaxy clusters) or not associated to anything visible but nevertheless forming clumps (weak lensing detection of galaxy clusters previously unseen). That is why dark matter is thought to be cold.

Additionally, there is a clear distinction between both types: there is not such thing as dark matter that is "not too cold but not too hot either" (see footnote as well). Dark matter is either made of particles with less than ~10 eV (hot dark matter, made of light particles, mostly dispersed everywhere) or particles with more than ~2 GeV (heavier, slower particles gravitationally bound to some structure). Both limits are found when imposing the maximum amount in which the candidate particles (neutrinos or something more exotic) can possibly contribute to the actual value of the density parameter due to matter in our expaning Universe.

Thus, either DM appears gravitationally bound (cold DM) or dispersed (hot DM), and both types are clearly distinct (10 ev vs 2 Gev). Observations favour the first case. However, Cold Dark Matter is not the ultimate solution, and still faces some problems.

Regarding the possibility of mixed solutions, many of them have been already ruled out. Microlensing has ruled out the possibility of unseen compact objects (brown dwarfs, stars, stellar black holes) in galactic halos, in our galactic neighbourhood as well as in the extragalactic domain. Ordinary matter (stones, bricks, dust) cannot possibly be, otherwise they would become hot and re-radiate. Any exotic mix of known particles doesn't work.

All we think we know is that DM must be made of some heavy particles yet to be discovered. In order to introduce a more complex model (e.g. different types of particles depending on the structure they appear attached to) one needs a justification (i.e. some predictions that better agree with reality) and nobody has been able to do that yet.

Remark Note that Dark Matter particles, either from the hot or the cold type, cannot possibly "slow down" and clump too much (e.g. forming planets) because they don't interact electromagnetically like ordinary matter, that is why DM is said to be collisionless. Wherever infalling ordinary matter forms any structure (e.g. protostars or accretion disks), a very important part of the process is thermalisation, i.e. the redistribution of energy of the infalling particles by means of numerous collisions. This cannot happen with Dark Matter.

Eduardo Guerras Valera

Posted 2013-10-25T15:45:17.243

Reputation: 352

There are some other issues aside from the "missing satellite problem", and they occur at the inner regions of simulated clusters and galaxies. Simulated galaxies and galaxy clusters appear to have steeper density profiles than observed clusters do - even in the Bolshoi Simulaton. People think that there may need to be additional physics simulated at the centers of these objects - where baryonic content may come into play. These considerations can be thought of as legitimate physical corrections to the dominant cdm paradigm. As @EduardoGuerrasValera has already mentioned, warm/hot dark matter – astromax – 2013-11-21T04:07:54.067

(cont'd) wipes out small scale structures, and it is clear that they exist. – astromax – 2013-11-21T04:09:05.773

Warm dark matter; This answer is out of date and far to didactic.

– Rob Jeffries – 2016-03-07T10:06:23.900

@RobJeffries You are most welcome to update/improve it (click on Edit). Or, much better, I suggest you write a full new answer. – Eduardo Guerras Valera – 2016-03-08T11:11:12.230

@RobJeffries ...specially if it is based on a superpartner (so far completely speculative theory with zero evidence) of the (already fully hypothetical) graviton. In a forum like this one and a short informal answer like this one, I think it is enough (or at least didactic) to stick to what is more or less mainstream. But as said, you are most welcome to write your answer. – Eduardo Guerras Valera – 2016-03-08T11:41:24.990

What you judge to be "mainstream" is not what others do. Warm dark matter is very much an area of contemporary research; thermal relics and sterile neutrinos are certainly not some theoretical backwater. Two-component models involving both CDM and WDM are being proposed to solve a number of problems that CDM-only models encounter. I write plenty of answers on this forum already. I am not minded to answer one which already has an accepted answer with many upvotes, even if, in my opinion, it contains a couple of misleading paragraphs (3 and 4). – Rob Jeffries – 2016-03-08T15:57:41.897


Very nice answer :-). Maybe things concerning what mass the dark matter particles potentially have will be clarified soon ... Though personally I rather expect another null result and that the huge media brimborium made about this LUX dark matter results will rather be misused as an excuse for the US decision makers to cancel some experimental dark matter studies, as a commenter on TRF said ... :-/

– Dilaton – 2013-10-27T11:53:07.573

Great answer, thank you! I have just two questions and one secondary argument. Question 1: As I understand it, cold DM cosmology puts limits on cold DM properties using observational data, whereas hot DM cosmology puts similar limits on hot DM properties. Do you know of any work, where Hot plus Cold DM cosmological models have been tested against observations? Question 2: I guess you are implying, that if there were two CDM or two HDM species, it would not have significant physical consequences, compared to single CDM and signle HDM component cases, correspondingly. Is it right? – Alexey Bobrick – 2013-10-27T13:35:28.990

Now, a small argument: We do not expect from our current knowledge of particle physics, that there may be warm dark matter particles. However, both cold and hot dark matter particle candidates are also not in the standard model of particle physics. Therefore, generally, perhaps, when looking at phenomenological dark matter, one should still consider warm case as a possibility. – Alexey Bobrick – 2013-10-27T13:39:16.833

@AlexeyBobrick If there is hot DM out there, it should affect the density parameter $\Omega_{M}$ and would leave less room for all the bound matter that is inferred from the observations. As long as I know, there is no need for hot DM to explain the "detection" of DM bound to colliding galaxy clusters, galactic halos affecting rotation curves, etc. On the other hand, the computer simulations (such as the Bolshoi) reproduce the observed universe departing from cold dark matter, otherwise they don't manage to reproduce structures that resemble the...

– Eduardo Guerras Valera – 2013-10-27T19:52:30.990

(...structures that resemble the) Universe as we see it today. Hot dark matter was the first thing people thought of, because neutrinos were a possible candidate at that time (it was not know whether they had any mass at all, in fact the first constraints on the neutrino masses came from cosmological considerations in the 70s, not from a particle accelerator, I believe) but the simulations reproduce the big scale structure departing from cold DM. The problem is that they predict a lot of dwarf galaxies that are not observerd, not even in the high redshift Universe. – Eduardo Guerras Valera – 2013-10-27T19:57:59.260

1@Dilaton Thanks for the link, it is great. – Eduardo Guerras Valera – 2013-10-27T20:00:58.430

@AlexeyBobrick Regarding "do you know of any work where Hot plus Cold DM cosmological models have been tested against observations", I don't know (sorry) but consider this: "fine" dark matter "detection" (I mean, merely more than the problem of the rotation curves and cluster virial masses) is done by weak lensing observations, i.e. lots of very faint background galaxies whose shapes are statistically processed with a computer. This requires modern technology (both CCDs and computers), thus it has been done more or less recently, under the cold dark matter paradigm (which is not yet perfect... – Eduardo Guerras Valera – 2013-10-27T20:07:53.723

(which is not yet perfect but uses only cold dark matter). There must be still people trying to make hot DM fit into the models, as well as there is still people trying to find a modified gravity that does not need DM at all, and also people trying to dismiss all our cosmology distance estimates, by arguing that the Universe is not homogeneous enough to allow using the standard model of Cosmology we use now. There is still some room for all those alternatives, because cold DM still does not work perfectly, and particle physicists have not yet given us the particle that solves the mistery. – Eduardo Guerras Valera – 2013-10-27T20:11:59.883


@AlexeyBobrick I encourage you to dig deep into ArXiV and try to find such studies mixing hot and cold DM. It is a fantastic, free resource, where most of the cutting-edge research in astrophysics and particle physics is published. It is somewhat hard and you will have to filter out a lot of unknown or unrelated stuff, but eventually it is worth the effort.

– Eduardo Guerras Valera – 2013-10-27T21:01:30.350

2@Dilaton Hey Dilaton. TRF is a great blog. I had not discovered it until now. I like Lubos's writing style a lot. I see you and Dimension10 are there too. – Eduardo Guerras Valera – 2013-10-28T04:26:00.957

2@Dilaton I hadn't discovered Lubos until now. His blog is incredible, wow! I am really having a good time reading his posts. He is ironic and corrosive (I am laughing like hell), and seems very accurate (at least conservative, because most of the stuff is new to me) in his scientific claims. He has a different style from Ron, but it is another "must". I hadn't paid any attention to TRF until you posted that link. – Eduardo Guerras Valera – 2013-10-28T12:19:12.593

@EduardoGuerrasValera: Thank you for so many answers! One last little question: You say hot DM tends to make too many small halos and dwarf galaxies in early universe. It seems counterintuitive, since hot DM is fast and hence is supposed to be more blurry and form structures on large, rather than small scales. Why is it then that hot DM likes to form small halos? – Alexey Bobrick – 2013-10-28T12:24:16.677

2@EduardoGuerrasValera yes Lumo rocks :-D! When reading TRF I often almost spit my coffee at my screen, because of his at times immensely funny style of writing, he makes me LOL regularely :-D. And of course learning cool cutting edge physics from him is very precious and valuable to me too! – Dilaton – 2013-10-28T12:24:29.043

2@Dilaton, I have the impression that he is a bit more careful and conservative when stating scientific facts than Ron, and that results in more accuracy. Ron very often dares entering in unknown territory, confident about his knowledge and intelligence, and then he ends doing statements that are his conclusions, fresh and usually amazing, but without much filtering. – Eduardo Guerras Valera – 2013-10-28T12:35:55.527

1@AlexeyBobrick No, you misinterpreted my comment (or I didn't explained it well enough). Hot DM does NOT form small halos. It is COLD DM who forms too many subhalos in the simulations. Structure Formation is a very complex field of research. The simulations that better reproduce how the density perturbations in the early universe grew until forming something close to our universe are today based upon cold dm only. I cannot assure 100% sure that there isn't anybody working in mixed models. But the most accepted paradigm today is cold dark matter alone, although it is still not perfect. – Eduardo Guerras Valera – 2013-10-28T12:49:14.177


@EduardoGuerrasValera I like both of them :-), and if Ron could still be among us elsewhere, certain things would probably be less hopeless. Sometimes I wish I had a time machine to undo certain key events ...

– Dilaton – 2013-10-28T18:18:18.993


@Dilaton Please don't miss this. You'll probably like it. It goes for you, CHEERS!

– Eduardo Guerras Valera – 2013-10-28T21:07:46.537


@EduardoGuerrasValera great answer, it summarizes what I was always saying since ... you know when :-). I'll have to login at Christmas exclusively to +1 it, if it (togethe with the question it is attached to) is still there then! If you have not yet seen it, you might be interested in this too ;-)

– Dilaton – 2013-10-28T21:34:54.367

1@EduardoGuerrasValera ha, I have just seen the last paragraph you added to your answer: exactly to the point! – Dilaton – 2013-10-28T23:16:47.213


Essentially, the answer is Occam's razor: look for the simplest solution and avoid complicated and contrived solutions, unless (observational) evidence requires them. Yes, it is possible that two or more types of dark-matter particles exist. But any solutions where not one species dominates require fine-tuning and hence are unfavourable. So, unless there is a theory that would naturally come with a mix of dark-matter particles (with different properties regarding their astrophysical implications, i.e. hot and cold etc, when Occam's razor does not apply), we should expect only one species to dominate.

If such a theory fails to explain the evidence, only then does it make sense to go to a more complicated model with more than one type of dark-matter particle. Currently, we are not at that stage.


Posted 2013-10-25T15:45:17.243

Reputation: 4 362

Would the downvoters please indicate what I could improve with this answer and/or why it's not useful? – Walter – 2015-10-14T18:41:24.340

@AlexeyBobrick There is zero evidence for supersymmetry. AFAIK, supersymmetry has not made a single falsifiable prediction that was later verified. Supersymmetry is a classiacal WOMBAT (= Waste Of (tax payer's) Money, Brain, And Time). – Walter – 2015-10-15T07:20:33.540

yes - certainly so. It hasn't been confirmed, there is only some indirect non-robust evidence for it from observations/experiment (inflation, DM, neutrino masses ..). Theoretically it has been used a lot though, e.g. in ads-cft. – Alexey Bobrick – 2015-10-15T10:56:36.807

First, I agree with your Occam's razor argument. Second - I thought there were emerging problems for CDM in explaining the lack of "cuspiness" in deduced DM profiles of dwarf galaxies. Is there any mileage in a 2-phase DM model in solving this? Third, be careful - the term WOMBAT might equally be levelled at the whole dark matter industry until some is detected (though it is at least a model that makes falsifiable predictions)! – Rob Jeffries – 2015-10-15T11:34:03.977

@RobJeffries thanks for your comment. I'm fully aware that dark matter is only still a hypothesis and hinges on the correctness of GR, which in turn is not experimentally verified on the relevant field strengths (or shall I say 'weakness'). The cuspiness issue is not a fundamental problem, as its prediction ignores baryonic effects, which must play a role on small scales. Moreover, on galaxy-cluster scales, the cusps are there (the mass profile in the outer parts of central-cluster galaxies is what we expect for the inner parts of the cluster halo). – Walter – 2015-10-16T07:34:43.597

@RobJeffries I'm particularly concerned about the waste of brain, of which comparatively little is spend in the dark-matter 'industry'. – Walter – 2015-10-16T07:41:48.647

2Well, I think most natural theory actually would predict more than one species. And then, concerning Occam's razor, it also doesn't apply here. Imagine theories "A", "B", "A+B" giving three different predictions and are all viable. Then it is absolutely not justified to exclude "A+B" out of consideration. However, it is a correct point, the more parameters - the more uncertainties and fine-tuning. – Alexey Bobrick – 2013-12-29T15:54:28.703

@AlexeyBobrick Occam's razor says we should not start playing with more than one different DM particle type, unless there is convincing independent evidence or theory to the opposite. Here, a theory is not just a simple model (messing around), but a prediction for the relation between two DM species which naturally emerges from some deeper insight. So, if your "A+B" is a theory in this sense, then Occam's razor doesn't apply. However, AFAIK, no such DM theories with more than one species are currently seriously considered. – Walter – 2014-01-06T17:29:47.687

1Yes, @Walter, "A+B" is a theory in this very sense: as expected as the other two. For why it is exepcted, check for possible extensions of standard model. For why it is not seriously used, check the other given answer. – Alexey Bobrick – 2014-01-06T20:21:04.400

@AlexeyBobrick So, which theory naturally contains two different (one hot, one cold) species of DM particles in roughly equal proportions (so that no either dominates)? The other answer does not explain why such theories are not seriously considered. AFAIK, a mix of particles hot and cold particles cannot currently be ruled out, but Occam's razor is used. – Walter – 2014-01-07T12:54:37.160

Supersymmetry, for example. The key point, though, is that possible extensions do not contradict each other. As per the other answer: two main microscopically motivated models are hot and cold DM. Observations of large scale structure favour cold DM, cosmology gives limits on both, hence there are no significant amounts of hot component. Plus hot DM does not play much role on small scales. What do you think would be worthy to look at further here? – Alexey Bobrick – 2014-01-07T14:55:27.570