Why Is the Universe Made of Matter Rather Than Antimatter?

Somewhere in the first fraction of a second after the Big Bang, the universe made a choice it did not have to make. For every billion pairs of matter and antimatter particles that flickered into existence in the hot plasma of creation, one extra particle of matter survived without a partner to annihilate. That lonely surplus — one part in a billion — is the entire physical content of today's universe. Every galaxy, every star, every atom in your body is made of leftovers, the tiny residue of a near-perfect cancellation that almost, but not quite, erased everything.

If the cancellation had been perfect, the universe would now be a cold, thin bath of photons and neutrinos, expanding forever into a darkness with no one to observe it. It was not perfect, and nobody yet knows why. This is one of the oldest unresolved debts in physics: an accounting error so small that we can barely detect its influence in laboratories today, yet so consequential that it decided whether existence would contain anything at all.

The strange part is that we can calculate, with the equations of the Standard Model in hand, how large that error should have been if only the known laws of physics were in charge. The calculation gives an answer many orders of magnitude too small. Something is missing from the ledger — some physics beyond what we have measured directly, acting in the first instants of cosmic history, that we have not yet caught in the act.

svg[The three Sakharov conditions, and where the Standard Model falls short of each]

Two independent lines of evidence tell us the baryon asymmetry has a definite, precisely measured value, and both agree strikingly well with each other. The abundances of light elements produced in Big Bang nucleosynthesis — deuterium, helium-4, lithium-7 — depend sensitively on the baryon-to-photon ratio during the first few minutes of cosmic history, and matching the observed abundances (mostly deuterium, measured in distant quasar absorption systems, is now the most precise) pins down η to good accuracy. Independently, the detailed pattern of acoustic peaks in the cosmic microwave background, mapped exquisitely by the Planck satellite, depends on the baryon density at the time of recombination, 380,000 years later. These two measurements, separated by hundreds of thousands of years of cosmic evolution and derived from completely different physics, agree: η ≈ 6.1 × 10⁻¹⁰.

Now run the Standard Model's own mechanism forward and see what it predicts. Both weak points identified in Sakharov's second and third conditions turn out to be fatal.

The CP violation problem is a matter of raw magnitude. The single CKM phase does generate CP-violating effects, and those effects have been measured precisely in kaon and B-meson systems, matching Standard Model predictions well. But translating that phase into an effective baryon-number-generating capacity during the electroweak epoch, physicists in the 1990s (notably Gavin Farrar and Michael Shaposhnikov, refining and then debating earlier estimates) found the CKM mechanism suppressed by a product of tiny quark mass-difference factors — a Jarlskog invariant divided by powers of the electroweak temperature — that drives the effective CP violation down to something like one part in 10¹⁰ to 10²⁰ relative to what's needed, depending on exactly how the estimate is done. Different calculations land in different places within that range, but they agree on the essential verdict: it is off by many, many orders of magnitude. There simply is not enough CP violation in the CKM matrix to do the job.

The equilibrium problem is, if anything, more damning, because it is not a matter of a weak effect but a qualitative failure of the mechanism's premise. Electroweak baryogenesis needs the Higgs transition to be strongly first-order — a genuine phase transition with bubble nucleation, latent heat, and a sharp discontinuity, the way water abruptly becomes steam. Whether the transition is first-order or not depends, in the Standard Model, almost entirely on the mass of the Higgs boson. Calculations through the 1990s showed that a first-order transition required a Higgs mass below roughly 70–80 GeV. When the Higgs boson was discovered at the LHC in 2012 with a mass of about 125 GeV, that possibility was closed. Lattice gauge theory calculations — nonperturbative numerical simulations of the full thermal field theory, carried out by groups including Kari Rummukainen, Mikko Laine, and collaborators through the late 1990s and 2000s — confirmed that for the measured Higgs mass, the electroweak "transition" is not a transition at all in the thermodynamic sense. It is a smooth crossover, with no bubbles, no latent heat, no departure from equilibrium of the kind Sakharov demanded.

Put the two failures together, and the verdict is unambiguous: the Standard Model, exactly as measured in every laboratory experiment to date, cannot generate the observed baryon asymmetry. It can generate an asymmetry — the mechanism is not fictional — but one many orders of magnitude too feeble, using a phase transition too gentle to lock anything in. This is one of the very few places in modern physics where the Standard Model, tested with such precision everywhere else, fails not marginally but catastrophically at a task it was in principle built to perform. Whatever completes the story must lie beyond it.

If the Standard Model can't supply enough CP violation or a strong enough departure from equilibrium, the natural move is to ask what new physics could supply either — or both. Several research programs have grown up around this question, each with real appeal and real difficulties, and none yet confirmed.

Extended electroweak baryogenesis. The most conservative extension keeps the electroweak-scale mechanism but adds new particles that couple to the Higgs sector, strengthening the phase transition and adding new sources of CP violation beyond the CKM phase. Supersymmetric extensions of the Standard Model were the most studied version of this idea through the 1990s and 2000s: extra Higgs bosons and light stops (superpartners of the top quark) could, in principle, make the transition strongly first-order while providing new CP-violating phases in the soft supersymmetry-breaking parameters. Its appeal is that it stays close to known physics and makes concrete, testable predictions — light stops and extra Higgs states that colliders could find. Its trouble is that LHC searches have pushed supersymmetric particle masses well above the range that the "light stop" scenario required, largely closing that particular window. More general two-Higgs-doublet models or singlet extensions of the Higgs sector remain viable and are actively studied, and they make a genuinely exciting promise: if the electroweak phase transition really was strongly first-order, it should have produced a stochastic background of gravitational waves, at frequencies that future space-based detectors could catch.

Leptogenesis. The most theoretically favored alternative shifts the action away from the electroweak scale entirely, up to energies vastly beyond anything a collider will reach. Leptogenesis, proposed by Masataka Fukugita and Tsutomu Yanagida in 1986, begins from a different puzzle that already independently motivates new physics: neutrino masses. Neutrinos, discovered to have tiny but nonzero masses through the observation of neutrino oscillations, most plausibly get those masses through a "seesaw" mechanism involving extremely heavy right-handed neutrinos — particles far beyond current collider reach, perhaps at energies of 10⁹ to 10¹⁵ GeV. If these heavy neutrinos exist, their out-of-equilibrium decays in the very early universe, occurring through CP-violating channels governed by their own complex Yukawa couplings (entirely independent of the CKM phase), could generate a net lepton number. The electroweak sphalerons — which, recall, violate B and L in a correlated way — would then partially convert that lepton asymmetry into the baryon asymmetry we observe today, in a well-defined ratio.

Leptogenesis is appealing precisely because it connects two mysteries — neutrino mass and the baryon asymmetry — with a single new sector of physics, and it can naturally produce an asymmetry of the right order of magnitude across a wide range of model parameters. Its trouble is that it is nearly untestable directly: the heavy right-handed neutrinos are typically far too massive to produce at any conceivable accelerator, so the case for leptogenesis rests on indirect consistency arguments — whether the CP violation measured in ordinary neutrino oscillation experiments (like those at T2K and the future Hyper-Kamiokande and DUNE) is compatible with the CP violation the seesaw mechanism would need at ultra-high energies, and whether the pattern of neutrino masses and mixings fits the models well. It is a compelling story that may never be pinned down beyond reasonable doubt in the way a direct detection would.

Affleck-Dine baryogenesis. A third family of ideas, proposed by Ian Affleck and Mark Dine in 1985, draws on supersymmetry in a different way: scalar fields carrying baryon or lepton number could have been driven to large values during cosmic inflation, then relaxed back toward zero along paths in field space that are not symmetric between particles and antiparticles, generating a large baryon asymmetry that later gets diluted down to the observed level. It is flexible and can be embedded naturally in supersymmetric models of inflation, but its very flexibility is a weakness: with enough free parameters, almost any answer can be arranged, making it hard to test or falsify.

GUT baryogenesis, an older idea from the late 1970s (Sakharov's original motivation, later developed by Motohiko Yoshimura and others), imagines baryon-number-violating decays of extremely heavy particles predicted by grand unified theories, occurring at energies near 10¹⁶ GeV, far above anything colliders will probe, and constrained by the requirement that any relic particles or topological defects from such a high-scale transition not overproduce unwanted relics like magnetic monopoles — a constraint that pushed the field, historically, toward inflationary cosmology in the first place.

None of these is a leading contender in the sense of having any confirmed evidence behind it. Each is a self-consistent, mathematically well-developed proposal that satisfies Sakharov's conditions in a different corner of energy scale — from the electroweak scale that colliders can reach, up through leptogenesis's inaccessible heavy neutrinos, to the truly speculative energies of grand unification. The honest state of the field is that we do not know which regime, if any of these, nature actually used.

Because the baryogenesis event happened once, in conditions no accelerator will ever fully recreate, the path to an answer runs mostly through indirect evidence — precision measurements sensitive enough to detect the fingerprints of whatever new physics did the work, even though we cannot rerun the event itself.

The most direct laboratory approach searches for new sources of CP violation, because almost every viable baryogenesis mechanism needs more CP violation than the CKM phase provides. One of the cleanest ways to look is the electric dipole moment of the neutron and of the electron — a tiny separation of positive and negative charge within these particles that would exist only if CP is violated in whatever interaction generates it. The Standard Model predicts electron and neutron EDMs so absurdly small (suppressed by many powers of small mixing angles, landing many orders of magnitude below current sensitivity) that any positive detection would be an unambiguous announcement of new physics. Experiments have been pushing the limits down for decades: the ACME collaboration's measurement of the electron EDM using thorium monoxide molecules, and neutron EDM experiments at the Paul Scherrer Institute and elsewhere, have already ruled out significant portions of the parameter space for supersymmetric and other electroweak-scale baryogenesis models. Each new order of magnitude of improved sensitivity — and several experiments aim for roughly tenfold gains over the coming decade — either finds new CP violation or closes off entire classes of theories.

A second avenue is neutrino physics. If leptogenesis is correct, the CP violation that seeded the baryon asymmetry, though occurring at inaccessible energies, is connected by the seesaw mechanism to CP violation observable today in neutrino oscillations — specifically a parameter called δ_CP, which governs whether neutrinos and antineutrinos oscillate between flavors at different rates. Current data from the T2K and NOvA experiments already show hints of large CP violation in the neutrino sector, though not yet at the level of a definitive discovery. The next generation of long-baseline experiments — Hyper-Kamiokande in Japan and DUNE in the United States, both under construction with results expected through the 2030s — aim to measure δ_CP with enough precision to establish neutrino CP violation unambiguously. That alone would not prove leptogenesis happened; the connection between low-energy neutrino CP violation and high-energy leptogenesis is model-dependent and not guaranteed to be simple. But a nonzero δ_CP would remove one obstacle and strengthen the plausibility case considerably, while a value inconsistent with viable leptogenesis models in a given framework would weaken it.

A third, newer, and genuinely exciting avenue is gravitational waves. If the electroweak (or some other) phase transition was strongly first-order, colliding and merging bubbles of true vacuum would have stirred the plasma violently enough to produce a stochastic gravitational wave background with a characteristic spectrum, redshifted by cosmic expansion down to frequencies in the millihertz range today. The planned space-based gravitational wave observatory LISA, expected to launch in the 2030s, is sensitive to exactly this frequency band. A detection of a stochastic background with the right spectral shape would be striking direct evidence that the early universe underwent a genuine first-order phase transition around the electroweak scale — the missing ingredient for electroweak baryogenesis to work, in its extended, beyond-Standard-Model forms.

Colliders remain relevant too. Future high-energy machines — a possible Future Circular Collider at CERN, or other proposed successors to the LHC — could directly produce the new Higgs-sector particles or other electroweak-scale states that extended electroweak baryogenesis models require, either confirming a viable mechanism or ruling out large swaths of remaining parameter space.

None of these routes offers a single decisive experiment with a clean yes-or-no answer on a fixed timeline, the way, say, the Higgs discovery had a clear threshold. The realistic expectation among researchers is a slow tightening of the net: EDM experiments closing off electroweak-scale models piece by piece, neutrino experiments constraining leptogenesis's viable parameter space, and gravitational wave observatories either finding or excluding a strong electroweak phase transition — each result narrowing, rather than instantly answering, the question of which mechanism (if any we have imagined) actually operated.

It would be easy to file this away as a technical accounting problem in cosmology, a mismatch of exponents to be patched by the right particle at the right energy scale. But sit with the plain fact underneath it: the theory that describes particle physics with extraordinary success, tested to parts per billion in some cases, predicts that you should not exist. Not "predicts something exotic about you" — predicts nothing at all, a universe of photons with no galaxies, no chemistry, no observers to be puzzled by their own absence.

That is a strange position for a scientific theory to be in: enormously successful and simultaneously unable to account for the existence of its own theorists. Resolving it is not a matter of tidying up a decimal point. It requires physics beyond the Standard Model — new particles, new interactions, new symmetry violations — operating in an era of the universe we have never directly probed and, with current technology, largely cannot. In that sense the baryon asymmetry functions as a lamppost, illuminating where new physics must be hiding even though we can't yet see the physics itself, in exactly the way beyond-the-standard-model physics is more broadly sought through anomalies, dark matter, and the hierarchy problem.

There are also deep connections to sibling puzzles. The strength and pattern of CP violation is itself mysterious in a second, independent way: quantum chromodynamics permits a CP-violating term (the theta term) that experiments show is suppressed to below one part in ten billion for no known reason — the strong CP problem, whose leading proposed solution invokes a new particle called the axion, itself now also a leading dark matter candidate. Whether nature's CP violation is stingy across the board, or generous in sectors we haven't yet probed, is a thread connecting the baryon asymmetry to the axion search and to the broader hunt for the missing 95 percent of the universe's mass-energy budget in the form of dark matter and dark energy. None of these puzzles is guaranteed to share a common solution, but they share a common lesson: the Standard Model, however successful, is a chapter, not the whole book.

There's also something almost vertiginous in what this problem says about contingency. If the mechanism turns out to be leptogenesis, then the precise number of protons in the observable universe was set, in some literal sense, by the tiny CP-violating phases in the couplings of particles that decayed and vanished before the universe was a trillionth of a second old, particles no experiment may ever directly detect. The universe's entire material content is a shadow cast by physics that has otherwise left the stage.

history

• 1964 — Cronin and Fitch discover CP violation in neutral kaon decays.
• 1967 — Sakharov proposes the three conditions for dynamical baryogenesis.
• 1986 — Fukugita and Yanagida propose leptogenesis via heavy neutrino decay.
• 2012 — LHC discovers the 125 GeV Higgs boson, closing the minimal electroweak baryogenesis window.

Stand back from the details and the shape of the problem is almost unbearably clean: we know the destination (η ≈ 6 × 10⁻¹⁰), we know the rules any journey there must obey (Sakharov's three conditions), and we know that the one route mapped out by confirmed physics — the Standard Model's own CP violation and electroweak transition — falls short by many orders of magnitude. Beyond that, the map runs out. Several roads have been sketched in pencil — electroweak baryogenesis with new Higgs physics, leptogenesis from unreachable heavy neutrinos, Affleck-Dine dynamics threading through the early universe's scalar fields, decays of grand-unified relics — each internally consistent, each capable in principle of reaching the destination, and none yet marked with a signpost confirming it as the one nature actually took.

What makes this frontier unusually humbling, compared to many open questions in physics, is the near-certainty that whatever answer is right, we may only ever confirm it indirectly. The event itself is inaccessible; only its consequences — a photon-to-baryon ratio, an absence of antimatter, perhaps a gravitational wave hum, perhaps a nonzero electron EDM, perhaps a measured neutrino CP phase — survive as evidence, and even a full set of consistent indirect signals will require care to interpret as proof rather than mere compatibility.

And yet the search keeps tightening. Every improved EDM limit, every year of neutrino oscillation data, every advance toward LISA's launch, narrows the space of viable stories. It is entirely possible that within a couple of decades, several of these avenues converge — a nonzero electron EDM at a scale matching some extended electroweak model, a measured δ_CP consistent with leptogenesis, a stochastic gravitational wave background with the telltale first-order-transition spectrum — and physicists will feel they finally understand, in outline if not in full, why there is something rather than a bath of photons.

Until then, the ledger remains open, and the imbalance in it is the reason there is anyone here to read a ledger at all.

Some of what remains genuinely unresolved:

• Which energy scale actually hosted the baryon-number-violating, CP-violating, out-of-equilibrium physics — electroweak, GUT, or something between?
• Is there a deep connection between the CP violation that shaped the early universe and the CP violation (or its suppression) at play in the strong-CP problem and the axion?
• Will next-generation neutrino experiments find CP violation of the right magnitude and character to make leptogenesis quantitatively convincing, or rule it out?
• Would a positive electron or neutron EDM detection point specifically toward electroweak-scale baryogenesis, or toward some other unrelated new physics?
• Could gravitational wave observatories like LISA directly detect the signature of the phase transition that set the baryon asymmetry, turning a cosmological relic into an observational fact?

Ask the tutor: If leptogenesis is right and the asymmetry was set by particles too heavy for any collider to ever produce, what kind of evidence — short of a time machine — could ever make us truly confident, rather than merely unconstrained, that this is the mechanism nature used?