← quantum · contents

§ Phase xv · Open Questions · Chapter 01

Dark matter

In 1933 a sharp-tongued Swiss astronomer noticed that the Coma cluster was spinning too fast for its visible stars to hold it together. Forty years later, a careful American astronomer measured how spiral galaxies turn, and found the same impossible result by the hundred. The universe is, by every method we have, mostly made of something that pulls on everything else and shines on nothing. We can weigh it, map it, and predict its effects across the sky. We still cannot say what it is.

The story begins, as so many stories in 20th century astronomy do, with a man who annoyed almost everyone he met. In 1933, Fritz Zwicky was a thirty-five-year-old Bulgarian-born Swiss physicist at Caltech, famous for his combative manners and an unusually creative eye for what telescopes ought to be doing. He had recently turned that eye on the Coma cluster, a swarm of more than a thousand galaxies some 320 million light-years away in the constellation Coma Berenices. He measured how fast the individual galaxies were moving relative to the cluster’s center. By the standard tool of the trade, the virial theorem, the average speed of the members tells you the mass that must be holding them together. If the members are moving fast and the gravity is weak, the cluster should have flown apart long ago.

Zwicky did the arithmetic and was startled. The visible galaxies, counted up by the light of their stars, came to a certain mass. The motion of those same galaxies required a mass about four hundred times larger. Something was holding the cluster together, and that something had to outweigh every star, every gas cloud, every patch of dust in the cluster by more than two orders of magnitude. In his 1933 paper, written in German for the Helvetica Physica Acta, Zwicky gave the unseen something a name that has stuck for nearly a century: dunkle Materie, dark matter.

The community’s response was to file the paper and move on. Zwicky was easy to ignore. He called his colleagues “spherical bastards” (so named, he explained, because they were bastards when viewed from any angle), and his methods were sometimes ahead of his data. The Coma observations rested on a small number of galaxy redshifts measured with the technology of the early 1930s. Perhaps the virial theorem did not apply cleanly to a cluster that was not in equilibrium. Perhaps the cluster’s distance was wrong. Perhaps the missing mass was just dim stars and cool gas that nobody had counted yet. Plausible-sounding objections piled up. For four decades, dark matter sat in a corner of the literature and gathered dust.

The corner stayed quiet until the early 1970s, when a careful young astronomer at the Carnegie Institution in Washington decided to look at something simpler than galaxy clusters. Her name was

Rubin and her instrument-builder colleague Kent Ford had a new tool, the Carnegie image-tube spectrograph, that let them measure the Doppler shifts of emission lines in faint outer regions of galaxies more precisely than anyone before. They started with Andromeda, the nearest big spiral, the galaxy you can see with the naked eye on a dark autumn night. They expected, if anything, a confirmation of the Keplerian fall-off at large radii. They got something else.

The rotation curve of Andromeda did not fall off. It went out and out, past the visible disk, past the last pinpoint of starlight Rubin could measure, and the orbital velocity stayed almost dead flat. The stars at the edge were moving as fast as the stars further in. By the simple Newtonian arithmetic, the mass interior to the orbit had to keep growing in lockstep with r, even out where there was no visible matter to grow. Either Newton’s law of gravity was wrong on the scale of a galaxy, or there was a great deal of mass out there that emitted no light at all.

Andromeda was not a fluke. Through the 1970s, Rubin, Ford, and their collaborators worked through one spiral after another. Sa, Sb, Sc; large, small; gas-rich, gas-poor. By 1980 they had a sample of more than sixty galaxies, every one of which showed the same impossible flat rotation curve. The result was not a single anomaly. It was a universal feature of how spiral galaxies turn. Whatever was holding their outer stars in orbit was not the visible disk. It was a roughly spherical halo of unseen matter, several times more massive than the stars and gas combined, extending well past the optical edge of the galaxy. The corner Zwicky had inhabited alone for forty years suddenly had a great deal of company.

By the late 1970s the evidence had passed from one careful astronomer’s notebooks into a community problem. Sandra Faber and Jay Gallagher wrote an influential 1979 review titled “Masses and Mass-to-Light Ratios of Galaxies” that surveyed the field and concluded, with reluctant honesty, that the case for unseen mass was now strong. A new generation of theorists began to ask what dark matter might actually be. The question split immediately into two: what does it weigh in total, and what is each individual lump of it?

The total has been answered, with stunning precision, by cosmology. In the 1990s and 2000s, satellites mapped the cosmic microwave background, the cooled afterglow of the Big Bang, in exquisite detail. The faint pattern of hot and cold spots in that radiation encodes the densities of every component of the universe at the moment, 380,000 years after the bang, when atoms first formed and let light escape. Ordinary matter (protons, neutrons, the stuff we are made of) leaves a fingerprint that depends sharply on its abundance. Dark matter leaves a different fingerprint. Dark energy, the mysterious tension stretching the universe apart, leaves a third. The COBE, WMAP, and Planck satellites measured the fingerprints. The numbers that come out are the same regardless of which method you use, and the numbers do not flatter our self-importance.

The result, repeated to four decimal places by every method now in use, is this. Of the total energy content of the present-day universe, about 68 percent is dark energy, the smooth tension driving the cosmic expansion ever faster. About 26 percent is dark matter, the cold quiet stuff that bunches around galaxies and holds them together. About 5 percent is ordinary matter, the protons and neutrons and electrons that make stars, planets, dust, oceans, vegetables, and people. A small fraction of a percent is in neutrinos and a thinner trace in photons. The everyday stuff of human experience is a five-percent rounding correction to the cosmic budget. The dominant components, taken together more than nine-tenths of the universe, do not radiate, do not absorb, and have never been seen in a laboratory.

So we know how much there is. We do not know what it is. The candidates are several, and after fifty years of hunting they remain frustratingly equally plausible. The most studied class is the WIMP, the weakly-interacting massive particle. The acronym was coined in the 1980s, when supersymmetric extensions of the Standard Model predicted a stable, massive neutral partner of one of the gauge bosons, exactly heavy enough and exactly aloof enough to explain dark matter without contradicting any laboratory test. WIMPs are appealing because they would be made in the hot early universe in roughly the right abundance to give the observed 26 percent (the so-called “WIMP miracle”). They would also, very rarely, scatter off an atomic nucleus and leave a tiny recoil signature, which is in principle measurable.

A second strong candidate is the axion, a hypothetical light boson originally invented in the late 1970s for an entirely different reason (to explain why the strong nuclear force does not violate the symmetry between matter and antimatter as strongly as it could). Roberto Peccei and Helen Quinn introduced the symmetry that bears their names; Frank Wilczek and Steven Weinberg later showed it implied a new particle. If axions exist with mass somewhere between a millionth and a thousandth of an electron-volt, they would behave as cold dark matter and account for the cosmic abundance. A third candidate is the sterile neutrino, a heavier cousin of the three neutrinos already in the Standard Model, distinguished by not feeling the weak force at all. A fourth, lately revived, is a population of primordial black holes formed in the first second of the Big Bang. The list is not closed.

v(r)² / r = G M(<r) / r²

M(<r) = v(r)² · r / G

How do you find a particle that interacts only through gravity and, at most, the weak force? Three strategies have been pursued in parallel for thirty years.

The first is direct detection. Bury an extremely clean and quiet detector deep underground (in a former gold mine, a converted nickel mine, a salt cavern beneath a mountain), shield it from cosmic rays and natural radioactivity, and wait for a dark-matter particle to bump into one of its nuclei. The bump deposits a few keV of energy, which the detector reads out as a tiny scintillation flash, an electron, and a small heat pulse. The current state of the art is the LZ experiment in the former Homestake mine in South Dakota and the XENONnT experiment in the Gran Sasso laboratory under the Italian Alps. Both use multi-ton tanks of liquid xenon. Both have run for years. Both have seen nothing yet beyond the expected backgrounds. The WIMP parameter space that fit so neatly with supersymmetry in 1985 has been almost entirely emptied by these null results, an extraordinary feat of experimental craftsmanship that has slowly retired its target.

The second is indirect detection. If two dark-matter particles meet and annihilate, the leftover energy might come out as gamma rays, positrons, or neutrinos, particles we can already see. The Fermi gamma-ray satellite, the IceCube neutrino observatory beneath the Antarctic ice, and various ground-based Cherenkov telescopes scan the sky for these signatures. There have been suggestive bumps in the data over the years, one near the galactic center, another in cosmic-ray positrons. None have survived rigorous follow-up. The third strategy is collider production. If dark matter can be created in particle collisions, the LHC at CERN might be quietly minting it; a dark-matter event would show up as missing transverse momentum, an invisible imbalance in the energy ledger of the collision. The ATLAS and CMS experiments have set upper limits across an enormous range of possible masses. Limits, not discoveries. Across all three strategies, after fifty years and many billions of dollars of effort, dark matter remains undetected in any non-gravitational channel.

Vera Rubin, who started all this, did not live to see the answer. She died on Christmas Day in 2016, at the age of 88, never having received a Nobel Prize that many of her colleagues thought she had earned several times over. She had spent her life arguing with the dominant assumptions of her field, often quietly, occasionally pointedly, and almost always correctly. In 2019 the United States Congress voted to rename the under-construction Large Synoptic Survey Telescope, then nearing completion on Cerro Pachón in Chile, after her. The Vera C. Rubin Observatory saw first light in 2025. Over the next ten years it will image the entire visible southern sky every few nights, building a deep, time-resolved map of more than thirty billion galaxies. Among its principal scientific goals: pin down the distribution of dark matter on every scale where it leaves a gravitational imprint.

That is, in some sense, the present state of the question. We have measured the dark matter’s mass, its density profile around galaxies and clusters, its role in the formation of cosmic structure, its imprint on the cosmic microwave background. We have built instruments far more sensitive than the technology of Rubin’s first observing runs would have allowed her to dream of, and we have searched in every plausible direction for the particle that constitutes it. The hunt continues. The universe, meanwhile, holds itself together with five times more invisible material than visible, and the simplest honest summary of our position, ninety years after Zwicky and fifty after Rubin, is that it works and we do not know how.

Bertone and Hooper, in a recent review of the field, opened with a line that distills the situation: dark matter is the only experimentally confirmed extension of the Standard Model that has no candidate experimentally confirmed. We have established that something is there. We have not established what. That is a stranger and more honest place to stand than physics is usually willing to occupy. It is also, after this long parade of triumphs from Planck and Bohr through Rubin herself, the chapter where the long arc of “and then we figured it out” finally pauses, and we wait.