I recently participated in a spirited discussion with lawyers, academics, writers, and human rights workers on the topic of free speech. None of us questioned free speech’s importance. However, we didn’t all agree on what exactly it should mean, or how we should balance it in the context of other rights. When does free speech’s potentially harmful consequences outweigh its benefits? Should spending money to promote particular laws or candidates be limited in any ways? A lawyer explained how the U.S. Supreme Court relied on the right to free speech—along with the idea of spending money as a form of expression—to decide the Citizens United case, which allows unrestricted political contributions from corporations. But others in our discussion were concerned that unlimited spending by corporations can drown out individual citizens’ voices—arguing too that free speech was intended for people, not corporations. After all, neither money nor corporations can speak—freely or not—without a human voice.
But that Supreme Court decision being what it is, and with the resulting flood of money that now enters politics, let’s consider the different ways that individuals—and corporations—can spend their money to influence public opinion.
Financial contributions might be focused to target advertising in local regions such as cities and towns, where they can readily change people’s viewpoints and influence the outcome of a vote. Or donors can contribute more diffusely, spreading their wealth and blanketing their claims over a larger region—yielding some general shaping of opinion but generating less clearly delineated effects. The two strategies together hold stronger sway than either type of advertising alone. But the rate of change should be greater in the targeted regions, clearly reflecting the greater density of advertising in the smaller but more concentrated locales.
Similarly in physics, the gravitational influence of a thinner, denser disk would more sharply influence the motion of stars than a thicker, more diffuse one. As with the more pronounced influence of local advertising, the positions and velocities of stars moving in and out of the galactic plane would be more noticeably influenced by a thinner, denser disk.
Because the Milky Way would contain disks of both ordinary matter and dark matter, the stars’ motions in and out of the galactic plane would depend on both, generating a combined influence that varies sharply, then gradually, as you move away from the dense region at the galactic midplane—analogous to the consequences of local and global advertising together. With a thin dark disk embedded in a thicker ordinary matter one, dark matter’s concentrated pull would combine with the more diffuse tug of ordinary matter to yield a distinctive measurable influence on stars that would vary with the distance from the midplane of the Milky Way.
We live in a data-rich era and we certainly don’t want to overlook any possible search targets—especially when looking for something as astounding but elusive as a disk of dark matter. This chapter will explain how measuring the Milky Way disk’s gravitational influence using the motion of stars will establish or undermine a dark disk’s existence. But before getting to that, this chapter will first explore other general considerations about disk dark matter’s possibilities and the potential for its discovery through more conventional dark matter searches that are currently under way. After that, it will present some of the dark disk’s intriguing astronomical implications.
DIVERSE DARK MATTER
When first studying partially interacting dark matter, I was astonished to find that practically no one had considered the potential fallacy—and hubris—of assuming that only ordinary matter exhibits a diversity of particle types and interactions. Although a few physicists had tried to analyze models such as one known as mirror dark matter, which features dark matter that mimics everything about ordinary matter, exemplars such as this one were rather specific and exotic. Their implications were difficult to reconcile with everything we know.
A small community of physicists had studied more general models of interacting dark matter. But even they assumed that all the dark matter was the same and therefore experienced identical forces. No one had allowed for the very simple possibility that although most dark matter doesn’t interact, a small fraction of it might.
One potential reason might be apparent. Most people would expect a new type of dark matter to be irrelevant to most measurable phenomena if the extra component constitutes only a small fraction of the dark matter inventory. Having not even observed the dominant component of dark matter, concerning oneself with a smaller constituent might seem premature.
But when you remember that ordinary matter carries only about 20 percent of the energy of dark matter—yet it’s essentially all that most of us pay attention to—you can see where this logic could be flawed. Matter interacting via stronger nongravitational forces can be more interesting and more influential even than a larger amount of feebly interacting matter.
We’ve seen that this is true for ordinary matter. Ordinary matter is unduly influential given its meager abundance because it collapses into a dense matter disk where stars, planets, the Earth, and even life could form. A charged dark matter component—though not necessarily quite as bountiful—can collapse to form disks like the visible one in the Milky Way too. It might even fragment into starlike objects. This new disklike structure can in principle be observed, and might even prove to be more accessible than the conventional dominant cold dark matter component that is spread more diffusely in an enormous spherical halo.
Once you start thinking along these lines, the possibilities quickly multiply. After all, electromagnetism is only one of several nongravitational forces experienced by Standard Model particles. In addition to the force that binds electrons to nuclei, the Standard Model particles of our world interact via the weak and strong nuclear forces. Still more forces might be present in the world of ordinary matter, but they would have to be extremely weak at accessible energies since so far, no one has observed any sign of them. But even the presence of three nongravitational forces suggests that the interacting dark sector too might experience nongravitational forces other than just dark electromagnetism.
Perhaps nuclear-type forces act on dark particles in addition to the electromagnetic-type one. In this even richer scenario, dark stars could form that undergo nuclear burning to create structures that behave even more similarly to ordinary matter than the dark matter I have so far described. In that case, the dark disk could be populated by dark stars surrounded by dark planets made up of dark atoms. Double-disk dark matter might then have all of the same complexity of ordinary matter.
Partially interacting dark matter certainly makes for fertile ground for speculation and encourages us to consider possibilities we otherwise might not have. Writers and moviegoers especially would find a scenario with such additional forces and consequences in the dark sector very enticing. They would probably even suggest dark life coexisting with our own. In this scenario, rather than the usual animated creatures fighting other animated creatures or on rare occasions cooperating with them, armies of dark matter creatures could march across the screen and monopolize all the action.
But this wouldn’t be too interesting to watch. The problem is that cinematographers would have trouble filming this dark life, which is of course invisible to us—and to them. Even if the dark creatures were there (and maybe they have been) we wouldn’t know. You have no idea how cute dark matter life could be—and you almost certainly never will.
Though it’s entertaining to speculate about the possibility of dark life, it’s a lot harder to figure out a way to observe it—or even detect its existence in more indirect ways. It’s challenging enough to find life made up of the same stuff we are, though extrasolar planet searches are under way and trying hard. But the evidence for dark life, should it exist, would be far more elusive even than the evidence for ordinary life in distant realms.
We have yet to directly see gravity waves emitted by a single object. Even black holes and neutron stars, which astronomers have detected in other fashions, have so far evaded gravity-wave detection. We stand little to no chance of detecting the gravitational effect of a dark creature, or even an army of dark creatures—no matter how close all of them might be.
Ideally, we would want somehow to communicate with this new sector—or have it correspond with us in some distinctive manner. But if this new life doesn’t experience the same forces that we do, that’s not going to happen. Even though we share gravity, the force exerted by a small object or life-form would almost certainly be too weak to detect. Only very big dark objects, like a disk extending throughout the Milky Way plane, could have visible consequences—like those discussed below.
Dark objects or dark life could be very close—but if the dark stuff’s net mass isn’t very big, we wouldn’t have any way to know. Even with the most current technology, or any technology that we can currently imagine, only some very specialized possibilities might be testable. “Shadow life,” exciting as that would be, won’t necessarily have any visible consequences that we would notice, making it a tantalizing possibility but one immune to observations.
In fairness, dark life is a tall order. Science fiction writers may have no problem creating it, but the Universe has a lot more obstacles to overcome. Out of all possible chemistries, it’s very unclear how many could sustain life, and even among those that could, we don’t know the type of environments that would be necessary. Enticing as it is, dark life is not only hard to test for. It is hard for the Universe to create. I will therefore leave this possibility aside—at least for now—and focus on the targets and the searches for a big dense disk that I expect to be more promising.
SIGNS OF A DARK DISK
To be systematic and to start with the most minimal assumptions, JiJi Fan, Andrey Katz, Matt Reece, and I first investigated the simplest DDDM model we could think of. In addition to the usual feebly interacting dark matter, our model contained dark-charged particles and a dark force analogous to electromagnetism, through which the charged dark matter particles interact. The model included a heavy particle that is positively charged like a proton and another type of negatively charged particle akin to the electron.
Working on a novel idea that hasn’t yet entered the physics canon is almost invariably a bit of an uphill battle. For some physicists and astronomers, double-disk dark matter is a stretch. Even for particle physicists, despite the rather daring nature of their research that aims to uncover the fundamental building blocks of matter, many colleagues—and scientists in general—are on the whole a conservative lot. This is not entirely unwarranted: if there is a conventional explanation for an observation, it is almost always the right one. Radical departures should be accepted only when they explain phenomena that older ideas fail to accommodate. In only very rare instances are new ideas truly necessary to explain observations.
Even when the scientific community agrees that something new is called for, deviating beyond the few “accepted” proposals for which the preponderance of work has lent weight can meet with resistance. Supersymmetry and WIMPs, for example, are often viewed by particle physicists as almost established, even though experimental evidence for them is as yet nonexistent. Only in the face of increasingly constraining data do many members of the community begin to acknowledge doubt and start to consider new possibilities for what lies beyond the established research canon.
Once a newer concept takes hold, everyone works it to death, figuring out and testing every corner of the space of parameters—even for hypotheses that are not yet proven to be true. But before an idea reaches that level, a lot of (often justifiable) criticism reigns. A few particle physicists—myself and my collaborators among them—simply try to keep an open mind in the face of uncertainty. We might favor some theories that we find more elegant or economical, but we don’t decide what is correct—or what to work on—until data’s arbitrating influence opens or closes a door.
My collaborators and I soon realized that interacting dark matter, which behaves very differently from non-interacting dark matter, should have distinctive observational implications. But given the initial motivation behind the DDDM proposal, I’ll also briefly consider its implications for more conventional search methods, such as the indirect detection signal that first stimulated our research—as well as for one place where DDDM might resolve an issue facing conventional dark matter scenarios. I’ll start by considering indirect signals, such as the Fermi photon signal that led to our research.
A thin dark disk is dense, meaning the concentration of dark matter particles is high. Within the dense disk, more dark matter encounters and hence more annihilations should occur than for dark matter distributed in a conventional cold dark matter halo. This doesn’t mean that DDDM models will all be observable in this way. For DDDM to generate an indirect photon signal, which was the initial stimulus for our idea, an additional ingredient over and above the charged dark matter I just described would be required. Because a Fermi-like signal requires dark matter to turn into photons, which are a form of ordinary matter, an observable interaction will arise only if there is a particle that is charged under both the usual electromagnetic force and the dark one—the analogue of the person who both watches Fox News and listens to NPR or is signed in to both Facebook and Google+. If a particle charged under both types of electromagnetism exists, then dark matter could annihilate into photons by producing this intermediary particle that connects to both the dark and visible sector. This makes the Fermi signal a possible prediction, but not a generic one, of DDDM.
The dense disk does, however, mean that if observable interactions exist, they will occur at a faster-than-expected rate. The even better news is that if DDDM generates any indirect detection signal, be it photons or positrons or antiprotons, the result will be distinguishable from that of any other type of dark matter model. With the dark matter indirect detection signal from the usual type of dark matter, the prediction for the rate is highest near the center of the galaxy, where the dark matter density is greatest. The signal from DDDM would also be stronger toward the galactic center, but any signal that comes from the galactic center should exist throughout the entire plane too—since dark matter is dense throughout the region. Such visible annihilations throughout the galactic plane would be a smoking gun for DDDM.
Of interest too are the potential implications of DDDM for direct detection experiments, which are, after all, the holy grail of many dark matter seekers. Recall that direct detection relies on a small interaction between dark matter and ordinary matter, which allows for the deposition of tiny recoil energy that a detector can potentially record. As with indirect detection, any direct detection signal of DDDM models would also rely on the optimistic (and nongeneric) assumption that dark matter has some interaction with ordinary matter—feeble enough to be consistent with everything we know but strong enough to lead to a detectable signal.
The direct detection signal also depends on the local dark matter density, since, after all, the more dark matter the better. Disk dark matter may or may not exist in the vicinity of ordinary matter—it depends on the thickness of the dark disk plane—but if it does, it should have much greater density than dark matter in the halo.
It is also well known that the dark matter detection rate depends on the dark matter particle’s mass, which helps determine whether the recoil energy is sufficiently big to be recorded and, if so, the amount of energy that would register. The detectability of the signal has a similar dependence on a more overlooked characteristic of dark matter, its velocity, which is also critical to the kinetic energy and hence the amount of recoil energy. Faster dark matter is easier to detect than slower dark matter since the energy deposited would be greater.
DDDM has much lower velocity in and out of the galactic plane than ordinary dark matter because it has cooled down. Moreover, the dark matter orbits around the galaxy in the same fashion as the Solar System, so its velocity relative to us is very small too. The low velocity of the new dark matter component relative to us means that DDDM would impart so little energy in a direct detection experiment—even if it did interact—that it would almost certainly be below the energy detection threshold and would therefore not be seen. Without more sensitive detectors or some additional ingredient in the model, conventional DDDM interactions would go unrecorded in the usual direct detectors.
However, experiments with lower thresholds are in the works and variants on the model might allow a signal even before their completion. What’s interesting here too is that should a signal be seen, it would be distinctive enough to identify a DDDM origin. The low velocity of the dark matter would lead to a signal that was much more concentrated in energy than any other dark matter candidate that has previously been proposed.
One further interesting test of our model—or any dark matter model containing charged dark matter that combines into atoms—comes from detailed studies of the microwave background radiation. Several astronomers and physicists used the CMB and galaxy distribution data to look for evidence of dark atoms and DDDM in an interesting new way.
Remember, radiation in normal matter can wipe out density variations in charged matter—much as wind on a beach smoothes out the evidence of tides—while dark matter simply attracts further growth in structure. The distinctive influences that imprint themselves on the cosmic microwave background radiation can be used to distinguish dark and ordinary matter. Ordinary matter can also leave an imprint when charged matter combines into neutral matter, much as you see a distinctive rise in the sand at the maximum extent of the water creeping up onto a beach.
If dark matter—or at least some of it—also interacts with dark radiation, effects that are similar to those of ordinary matter will imprint themselves on the background radiation. Since our model contains both a heavy dark matter particle and a light one with opposite charge—very much akin to a proton and an electron—these particles would combine into dark atoms that would register in ways very similar to ordinary matter.
Detailed studies of the cosmic microwave background radiation showed that the fraction of dark matter that can have interactions of the sort we suggested is constrained. If the temperatures of the two sectors are reasonably similar, as would be the case if the dark and ordinary sectors interacted enough early on, the amount of interacting dark matter might have to be as low as five percent of the amount of total dark matter—about one-quarter the amount of visible matter. Fortunately this value is still interesting and it should also be observable using the method presented below. It is also well within the value needed to explain the periodic meteoroid strikes that I’ll get to in the following chapter.
MEASURING THE SHAPE OF THE GALAXY
The research just described was interesting in that it demonstrated not only the power of the cosmic microwave background radiation, but also the significance of large data sets in the modern cosmological era—the processing of which astronomers are well set up to do. With the input of a model-building perspective and the technological and numerical advances currently under way, we have a much better chance of finding influences of unconventional dark matter—even when they are only subtle effects on the observed distribution of structure. My collaborators and I recognized that most interesting and robust signals are probably not the ones targeted by the usual dark matter searches that I’ve already described. More promising observational consequences of a dark disk follow from the gravitational pull of the disk itself. In today’s era of “big data,” the best places to look for distinctive dark matter properties could well be seemingly ordinary astronomical data sets.
The most obvious and decisive generic implication of the DDDM proposal is the existence of a thin dark disk in the central plane of the galaxy. If the dark matter particle is heavier than the proton, the disk will be narrower than the one containing stars and gas, making the gravitational potential exerted by the Milky Way galaxy—and all the others—different from what would be expected without the new form of dark matter. Like targeted advertising, the dark disk will add extra heft to the more diffuse ordinary matter component—and furthermore change matter’s distribution—influencing the gravitational potential most dramatically near the galactic midplane, where the dark matter disk is concentrated. Because the gravitational influence of this matter distribution would influence the motion of stars, when enough positions and velocities of stars are measured with adequate precision, the distribution will confirm or rule out a dark disk (at least one with big enough density to make a difference).
One of the most incredible developments that JiJi, Andrey, Matt, and I learned about when we first started thinking about the dark matter disk in the summer of 2013 was that precisely this measurement in the Milky Way galaxy was set to occur. A satellite that was scheduled to launch that fall (or in the spring for those at the French Guiana launch site, as pointed out by my bemused Australian colleague) should measure this distinctive gravitational influence.
The GAIA satellite will in effect measure the shape of the galaxy. Within five years we will know the result. Preparations for the satellite were already well under way when we worked on our first paper, but it will conduct precisely the dark disk measurement we might have requested had we been asked during its preparation. In fact, although they didn’t have our precise model or methodology in mind, astronomers had argued for the GAIA mission in large part based on its ability to determine the mass distribution in the galaxy—no matter what type of matter or where in the galaxy it is located. Though the takeoff was delayed a couple of months from the initially scheduled date, the launch in December of that year—only several months after we had finished our paper—certainly seemed remarkably fortuitous.
Particle physicists don’t encounter this type of surprise very often. We know which experiments are possible and try to find if they can be tweaked or interpreted in a way that will test new ideas. Experimenters at the Large Hadron Collider (LHC) at CERN investigate some of the proposals that Raman Sundrum and I and others devised to explain the mass of the Higgs boson, for example. Though the LHC experiments were initially designed with other models in mind, Raman and I were fully aware of them and their potential when conducting our research on a warped extra dimension of space.
On the other hand, sometimes an idea is sufficiently compelling and testable that experimenters will respond and design a relatively small-scale experiment to rule out or verify the proposal, as when physicists designed experiments to measure the gravitational force more precisely in response to large extra-dimension ideas.
But rarely does it happen that an experiment just happens to be starting that is poised to test an idea that was studied independently for completely different purposes. Yet this is what materialized. The GAIA satellite houses a space observatory that will measure the positions and velocities of a billion stars of the Milky Way, with the goal of creating an extremely precise and extensive three-dimensional galactic survey. Its measurements will map onto a particular galactic potential and thereby tell us the galaxy’s density distribution. If this distribution demonstrates the presence of a dark disk, the disk’s thickness and density will tell us in turn about the mass of the new type of dark matter particle and how much of the interacting dark matter exists.
The method is based on an idea suggested by Jan Oort—the astronomer who also established the existence of the Oort cloud. Oort realized that the velocities of stars as they go in and out of the galactic plane depend on the shape and the density distribution of the disk, since their motion responds to the disk’s gravitational pull. Measuring the velocities and the positions of stars oscillating in and out of the plane therefore pins down the density and spatial distribution of matter in the disk.
This is precisely what we would like to know to test or confirm our dark disk proposal. The gravitational attraction of a dark disk affects the motion of stars since they respond to the galaxy’s gravitational pull. Knowing positions and velocities precisely for so many stars will reveal the galaxy’s gravitational potential and establish whether or not a dark disk exists. With detailed information about the disk potential and the spatial distribution of matter within it, we can hope to determine more about the properties of the disk and about the interacting dark matter that created it.
But we don’t have to wait for GAIA data to test the method and get a preliminary result. We already have useful data from the Hipparcos satellite, which the European Space Agency launched in 1989 and which continued to operate until 1993. Hipparcos was the first to do the detailed position and velocity measurements, but they did so with less accuracy and with fewer stars than GAIA will survey. Yet its results, though not as complete as GAIA’s will be, already constrain the form that a dark disk might take.
This insight, though new to us as particle physicists, was well known among some astrophysicists. In fact, using this method, a few researchers had even gone so far as to conclude that existing data already rules out a dark disk. This cavalier disk denial confused many people, including one of our paper’s referees. A moment’s reflection, however, tells you that this result (at least as stated) is not possible. No matter how accurate the measurement, the density can always be sufficiently low to evade any existing bound. What astrophysicists were really saying was that there was no need for a dark disk. Given the uncertainties in densities in all the known gas and star components, the measured potential could be accounted for by known matter alone.
But sometimes the right question is what else might be consistent, and therefore be a viable alternative interpretation of the data. The only way to find out whether something is allowed or even preferred is to evaluate the consequences of new assumptions and determine their experimental implications. My collaborators and I asked a different question from the astronomers. We didn’t ask for proof that a dark disk exists. The real question is how substantial a disk can be while maintaining consistency with all observations. And whether perhaps the introduction of a dark disk component might even match the data better.
This different way of thinking reflects in large part the difference between the sociology of particle physicists—particularly model builders—and many astrophysicists. To give credit where it’s due, the astrophysicists taught us quite a lot. We learned how they approach the problem and what data currently exist. Their methods are extremely useful. But approaching a problem from a different angle often leads to new insights and opens up new possibilities. Whether or not a dark disk exists, we will only know by making the assumption that it does and figuring out what is allowed. Everyone wins in the end.
We wanted to know if a dark disk is allowed, or even possibly favored, by the data—not just whether or not one can fit the measured stars’ properties without one. Each of the ordinary matter components that is added to compute the Milky Way disk’s gravitational potential is known only so well. Allowing for the uncertainties in the measurements certainly creates room for something new. This is the task I set out for a student, Eric Kramer, who studied the Hipparcos data as well as gas density measurements in the galactic plane. Together we identified many assumptions that went into the astrophysicists’ analysis that needed to be revisited. Although a cursory examination of the Hipparcos results could prematurely lead to the conclusion that a dark disk was disfavored, a more careful and up-to-date analysis demonstrated that the data didn’t suffice to make such a claim.
The Hipparcos data itself provide some of the uncertainties. But the relatively poor measurement of some of the visible matter in the Milky Way is a major source of uncertainty too. The more wiggle room there is, the more room there is for a dark disk. On top of that, because all components of matter experience the gravity exerted on them by the other components, only by including all matter—including the dark disk—from the beginning can one extract the true constraints. This is one of the merits of having a model. It gives a well-defined target and a fixed computational strategy when evaluating the results of a search.
With a careful analysis, we found that there is room for a dark disk. Indications are promising, but before more conclusive data becomes available we won’t know if DDDM models will prove to be correct or whether simpler, more standard scenarios suffice to account for the matter in our Universe.
This brings me to the question: what is the dark disk density we hoped to target in the first place? That is, how strong a constraint would be interesting? From many perspectives, any value is worth pursuing. Finding a dark disk, no matter how low its density, would be a fundamental change in our view of the Universe. But we will soon see that another target comes from the periodic-meteoroid-strike-inducing aspect of the dark disk. I will simply say for now that the value we found that was necessary to trigger meteoroid attacks is consistent with current data.
Furthermore, although it wasn’t our original intention, partially interacting dark matter might also help solve some outstanding mysteries of more conventional cold dark matter scenarios. The astronomer Matthew Walker, now a professor at Carnegie Mellon University, suggested that DDDM might help address the problem with dwarf satellite galaxies in Andromeda alluded to in Chapter 18. A world with ordinary matter or even conventional cold dark matter offers no explanation should these results hold up. A Harvard postdoctoral fellow, Jakub Scholtz, and I showed that self-interactions for a component of the dark matter might be the unique solution to the problem of how dark-matter-dominated dwarf galaxies that are aligned in a plane are formed. Jakub, Matthew Reece, and I are also investigating the potential implication of DDDM for primordial black holes, which are bigger than they should be in standard scenarios.
The Fermi gamma ray signal that motivated our project now appears to have been a red herring since the signal has faded over time. But the dark disk scenario that grew out of trying to understand it has wide-ranging implications that should make DDDM observable in other ways. The scenario might even have more interesting implications for galaxy formation and dynamics that we can now begin to explore.
So, after our expansive exploration of the cosmos and the Solar System, let’s culminate our journey by bringing together a lot of these ideas. We’ll now consider how dark matter might affect us close to home—affecting the motion of stars and potentially the stability of objects at the outskirts of the Solar System.