arxiv Find: The Local Density of Dark Matter

By Sean Carroll | July 2, 2009 11:44 am

One of the big hopes of particle- and astro-physicists over the next few years is to experimentally pin down the nature of dark matter. In a perfect world, we’ll make the dark matter particle at the LHC, observe gamma rays produced when dark matter annihilates in the galaxy, and detect it directly in experiments here on Earth. The world isn’t always perfect, but sometimes it’s even better, so everyone is sitting on the edges of their seats waiting to hear what the experiments tell us.

For the direct-detection strategy here on Earth, we build giant detectors and wait for ambient dark-matter particles to interact with something in the detector. If the dark matter is a weakly interacting massive particle (WIMP), that’s not so hard; the difficult part is distinguishing a purported signal from various backgrounds. To know what the signal should be, of course, we need to know how many dark matter particles are zipping through the laboratory. It should be a good number: roughly speaking, there would be about one weak-scale-sized dark matter particle per coffee-cup-volume in the universe, and in our galaxy these particles will typically be trucking along at around 300 kilometers per second.

Still, you’d like an accurate estimate of how much dark matter there is supposed to be in your detector. That’s what Riccardo Catena and Piero Ullio claim to have provided:

A novel determination of the local dark matter density
Authors: Riccardo Catena, Piero Ullio

Abstract: We present a novel study on the problem of constructing mass models for the Milky Way, concentrating on features regarding the dark matter halo component. We have considered a variegated sample of dynamical observables for the Galaxy, including several results which have appeared recently, and studied a 7- or 8-dimensional parameter space – defining the Galaxy model – by implementing a Bayesian approach to the parameter estimation based on a Markov Chain Monte Carlo method. The main result of this analysis is a novel determination of the local dark matter halo density which, assuming spherical symmetry and either an Einasto or an NFW density profile is found to be around 0.39 GeV cm$^{-3}$ with a 1-$sigma$ error bar of about 7%; more precisely we find a $rho_{DM}(R_0) = 0.385 pm 0.027 rm GeV cm^{-3}$ for the Einasto profile and $rho_{DM}(R_0) = 0.389 pm 0.025 rm GeV cm^{-3}$ for the NFW. This is in contrast to the standard assumption that $rho_{DM}(R_0)$ is about 0.3 GeV cm$^{-3}$ with an uncertainty of a factor of 2 to 3. A very precise determination of the local halo density is very important for interpreting direct dark matter detection experiments. Indeed the results we produced, together with the recent accurate determination of the local circular velocity, should be very useful to considerably narrow astrophysical uncertainties on direct dark matter detection.

So they’re claiming the density is about .39 GeV per cubic centimeter (where one GeV is about the mass of the proton), whereas the standard figure is something closer to .30 GeV per cubic centimeter. More importantly, they claim to trust their estimate to a precision of about 7%, while the usual number is supposed to be uncertain by a factor of 2 or 3.

I’m not expert enough to judge whether they are right, but it would certainly be very impressive to pin down the density to such high precision. They do assume spherical symmetry, however, which I suspect is not a very good assumption. There are ongoing arguments about how lumpy the distribution of galactic dark matter really is, and I can easily imagine that lumpiness can distort the local density by much more than 7%. But work like this is going to be very important in interpreting the results, if (when?) we do directly detect the dark matter.


    When will they bring the LHC back online? I’m getting bored!

  • weichi

    you’ve got a typo – should be .39 and .30 Gev/cm^3

  • Sean

    Oops, you’re right. Fixed.

  • Jason Dick

    For the direct-detection strategy here on Earth, we build giant detectors and wait for ambient dark-matter particles to interact with something in the detector.

    Well, not always. Some of the dark matter detectors are truly giant, but not all. See DAMA/LIBRA, for instance:

    250kg of sodium iodide isn’t small, but I wouldn’t exactly call it giant. It’s quite a neat idea, though: if I’m reading their material correctly, they are able to detect single elastic scattering events off of the nuclei of their sodium iodide targets. They also claim a positive result (they see yearly variation in their detectors). I still remain skeptical that this result will actually turn out to be dark matter, and would definitely like to see some independent confirmation. But cool nonetheless!

  • Alberto

    and 7% is at 1 sigma

  • Sili

    Stupid question, but will the DM be able to leave a signature noticable above the background in CMS and ATLAS? Or are they wholly unsuited for noticing?

  • Sean

    Pretty much unsuited. ATLAS and CMS can detect cosmic rays (which are useful for calibrating the detectors), but for dark matter the whole trick is to accurately distinguish from backgrounds, which the big particle-accelerator experiments aren’t designed to do.

  • Sili

    Thanks. I expected as much, but it’s good to know.

    Oh, and by the way. Thanks for the link to the Susskind lectures. Unfortunately the GR ones ran out just as they were getting juicy (did you do that in order make better sales?), but the Classical Mechanics is good too – I did take it yeeeeeaaaars ago, though, so I’m better primed for it.

  • macho

    7% uncertainty is a bit too optimistic (loosening the assumption of spherical symmetry will increase this uncertainty by a significant amount). Also, the uncertainties in some of the galactic parameters are larger than they’ve assumed here.

    There was another very recent paper that came at this with a different approach

    Reconstructing WIMP Properties in Direct Detection Experiments
    Including Galactic Dark Matter Distribution Uncertainties

    and found a local density of 0.32 Gev/cm$latex ^3$ with an uncertainty of 22% for
    their “baseline” model — with full acknowledgement that this is a model dependent

  • Lab Lemming

    Is GeV a mass equivalent energy unit? If so, why don’t they just use mAMU or (shock, horror) grams?

  • Jimbo

    That number is enuf to make a cosmologist go wild.
    The density of dark energy is 3.9 Gev/m^3; Thus they’re claiming a dark matter density which is exactly 5 orders of magnitude larger than that of dark energy !?
    The relative density of DE:DM is supposed to be about 0.72/0.24 ~ 3:1 with regular baryonic matter accounting for the remaining 4%.
    What gives ?

  • Olli

    Unlike DE, DM clusters due to gravity, i.e. it accumulates in objects like galaxies and clusters of galaxies. In the paper, they estimate the *local* density in the solar neighbourhood, which is significantly enhanced over the *average* density of DM and thus higher than the density of DE.

  • T.

    Hi Sean,

    Thanks for the always great posts.

    However, are you aware of this unnerving ad, which appears to the right of the page, saying:
    “The athesit riddle: so simple even a child can understand, so complex no atheist can solve” ?
    The link leads to a pseudo-science blog..

  • Sean

    Yeah, that is being taken care of as we speak.

  • Jimbo

    Crystal clear & thanx !

  • David Moles

    So why is dark matter so sparse? If the total mass of dark matter in the neighborhood of our galaxy is say 6x the mass of baryonic matter (Olli@12’s figure), why do we find the mass of the Earth, or the Sun, or the solar system, to be just the mass of the baryonic component? Does the hypothesized weak interaction repel WIMPs from one another sufficiently to keep them from clumping at densities greater than ~1 / coffee cup?

  • Peter Erwin

    David Moles @ 16:
    The basic answer is that the dark matter (DM) is pretty smoothly distributed, while the baryons tend to be clumped into very dense objects (e.g., stars and planets). And also that the volume of even the solar system is pretty small compared to the volume of interstellar space, so that you don’t expect to find a whole lot of DM within the solar system.

    Take the Earth, for example: if the local density of DM is ~ 6 x 10^-25 gm/cm^3 (i.e., 0.35 GeV/cm^3), then within the volume of the Earth (~ 10^27 cm^3), you expect a measly 600 gm of DM. Compared to the total mass of the Earth (6 x 10^27 gm), that’s kind of small….

    Similarly, we can ask how much DM there is inside the solar system. Using 50 Astronomical Units (slightly larger than Pluto’s orbit) as a working size, the expected DM is 10^21 gm — which you’ll notice is still much less than the mass of the Earth, let alone the Sun. So gravitational interactions in the solar system will still be dominated by the Sun and the planets.

  • confused in the early morning monday

    I think the other side of the answer to David Moles @16 is that CDM is collisionless at least at low pressures, so that even near (or within) normal stars diffuse DM gas tends to remain diffuse rather than clumping up due to degenerate phase DM self-interaction. Since it’s diffuse and cold, individual DM particles move very slowly in orbits about solar and planetary masses; their very low escape velocities (please forgive the use of this term by way of analogy) with respect to these masses keep the DM particles from falling far down the gravitational potential gradient, and their non-participation in electromagnetism means they are not kicked into interstellar space by solar radiation.

    I think one would expect that there is more DM in regions of lower gravitational potential and that in principle this can be observed when studying galactic cores, whereas it’s difficult to test this in the solar system because of the low densities as Peter Erwin says at @17.

    However, from the point of view of an observer of DM who is at rest while above and rotating with the galaxy, the peculiar motion of the total mass-energy of our solar system (including the DM gravitationally bound up in the solar system) probably appears to leave a clear kelvin wake in the otherwise fairly still (i.e., cold) and sparser galactic DM gas. In due course, some of that gas will be swept up into moving with the solar system, and some of the solar system’s DM will escape.

    A similar view will be had for all other solar-system-like clumps of mass-energy in the galaxy, but other than this “stirring”, the vast majority of individual galactic dark matter move in slow and fairly stable orbits about the galactic centre, with negligible chance of being knocked (and in most cases little chance of being dragged) out of the galaxy by radiation, and with very little infalling because there is no more than minuscule amounts of dark chemistry working to clump together dark mass-energy into larger structures that have larger galactic escape velocities. The galactic DM will be denser towards the centre of mass of the galaxy, and there will be some dynamical migration of DM in and out of the galaxy.

    This probably continues analogously at still larger scales.

    Or, of course, I could be completely out to lunch when I should be having breakfast…

  • David Moles

    Peter Erwin @17: Thanks, but I guess what I was really asking was “why is it so smoothly distributed” — when baryonic matter isn’t. Perhaps confused @18 has the answer.

  • Peter Erwin

    Peter Erwin @17: Thanks, but I guess what I was really asking was “why is it so smoothly distributed” — when baryonic matter isn’t.

    Ah, OK. The short answer is that baryons easily can lose energy via interactions with each other and radiation, and DM cannot.

    You can think of DM collapsing into a DM condensation (a DM halo) as a process whereby the DM particles trade potential energy for kinetic energy: they end up losing potential energy by moving closer to the center of the halo, but gain kinetic energy. The result is a DM halo that’s in quasi-equilibrium, with all the DM particles on orbits within the gravitational potential of the halo. This is something denser than the initial smooth background, but not super dense. In order for it to become denser, the DM particles would have to lose energy[*] somehow — but they can’t, at least not on reasonable timescales.

    The situation for baryons[**] would be similar, except for the crucial fact that they can interact with each other and give off radiation. So the baryons can lose energy via radiation, and sink closer to the center, and thus the baryon condensation gets denser. Which leads to more frequent baryon-baryon interactions, and more radiation (i.e., the collapsing cloud of gas heats up and glows), which just reinforces the process.

    Once it gets going, baryonic collapse tends to stop only when the center is dense enough to resist further collapse, either because it’s hot enough to produce nuclear reactions (which add kinetic energy to the particles) or because pressure support is stable (as in the case of planets, white dwarfs, and neutron stars).

    I’m skipping over a lot of details (e.g., baryonic collapse can slow down if the condensation becomes dense enough to be opaque, so that radiation only leaks out at the surface — this is the situation for protostars that haven’t started nuclear fusion yet), but I think that’s the gist of it.

    (I realize this doesn’t answer what might be a further question — why are baryons concentrated into lots of different clumps instead of, e.g., one big clump at the center of the galaxy? — but: a) that’s getting into detailed physics of gas-cloud fragmentation and so forth; and b) this comment is long enough as it is…)

    [*] Total energy, in the sense of potential + kinetic energy.

    [**] This includes electrons, since they’re usually considered part of “baryonic matter” even though they’re really leptons.

  • confused

    In order for it to become denser, the DM particles would have to lose energy[*] somehow — but they can’t, at least not on reasonable timescales.

    The timescales seem to be more reasonable when forming a DM halo (rather than a pure DM condensation). DM in a mixed cloud of DM and baryonic matter can transfer kinetic energy to the embedded baryonic matter gravitationally and through weak scale interactions. The denser the normal matter, the greater its opportunity for experiencing an acceleration attributable to DM which in turn cools that DM enabling it to follow a closer orbit. DM in regions very sparsely populated by baryonic matter will have far fewer opportunities to do so.

  • Michael Williams

    My understanding is that a spherically symmetric halo is a fairly well-justified assumption in the context of their method and goal. The CDM+baryon simulations (e.g. and references therein) and the particular observational evidence for the Milky Way (e.g. suggests that the axial ratio is perhaps around 0.8. I would like to see it checked, but it’s difficult to see how changing their halo model to better match these simulations and observations would result in a 25% change in their measurement (and thus make their number consistent with previous determinations).

  • David Moles

    Thanks, Peter & confused. That gives me a much clearer picture.

  • Pingback: Galactic Interactions()

  • Pingback: Länkar på parad « Stjärnstoft och kugghjul()


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Cosmic Variance

Random samplings from a universe of ideas.

About Sean Carroll

Sean Carroll is a Senior Research Associate in the Department of Physics at the California Institute of Technology. His research interests include theoretical aspects of cosmology, field theory, and gravitation. His most recent book is The Particle at the End of the Universe, about the Large Hadron Collider and the search for the Higgs boson. Here are some of his favorite blog posts, home page, and email: carroll [at] .


See More

Collapse bottom bar