CP violation and the matter-antimatter problem

Particle physics

10 min read

Why is there any matter at all?

In the early Universe, the temperature was high enough that particle–antiparticle pairs were constantly being created and destroyed. If the laws of physics treated matter and antimatter perfectly symmetrically, you would expect those processes to leave behind essentially nothing: as the Universe cooled, matter and antimatter would annihilate away into photons.

Instead, we see a small leftover excess of matter. The key point is that it only had to be a tiny imbalance: roughly “one extra baryon per billion particle–antiparticle pairs”. That small surplus is enough to make all the atoms, stars, and galaxies we observe.

So the puzzle is not “why no antimatter?”, but “how did a very small matter excess appear from nearly symmetric physics?”

Sakharov conditions: the checklist for making a baryon asymmetry

In 1967 Andrei Sakharov wrote down three ingredients that any successful explanation (baryogenesis) must include:

Baryon number violation
If baryon number B is exactly conserved, you can never generate a net excess of baryons over antibaryons. (In the Standard Model, at very high temperatures there are non-perturbative electroweak processes that can violate B + L, where L is lepton number.)
C and CP violation
C (charge conjugation) swaps particles with antiparticles. P (parity) flips spatial coordinates. CP does both. If CP were an exact symmetry, then for every reaction that produces a little more baryon number, the CP-mirrored reaction would produce an equal amount of antibaryon number. Averaged over the early Universe, the net would cancel. CP violation is what biases reactions so “matter-making” and “antimatter-making” are not exactly matched.
Departure from thermal equilibrium (or an effective alternative)
In thermal equilibrium, detailed balance tends to erase asymmetries. You need conditions where forward and reverse processes do not cancel perfectly. A classic way is a strong, first-order phase transition with rapidly expanding bubble walls; another is out-of-equilibrium decays of heavy particles.

These conditions explain why CP violation matters cosmologically: it is one of the few ways to build in a directional preference for matter over antimatter.

Why Standard Model CP violation seems insufficient

The Standard Model does contain CP violation. In the quark sector it appears through a complex phase in the CKM mixing matrix (the matrix that relates weak-interaction quark states to mass eigenstates). This is real, experimentally established CP violation.

The problem is “amount and circumstances”:

The CP-violating effects from the CKM phase are typically small in processes relevant for generating a cosmic asymmetry. In practice they come with suppressions: small mixing angles, hierarchy of quark masses, and the need for interference between at least two amplitudes that are both sizeable. When you combine these, the net CP-odd bias available for baryogenesis looks far too feeble.
The out-of-equilibrium ingredient is also weak in the Standard Model. For the observed Higgs mass, the electroweak phase transition is not strongly first order; it is closer to a smooth crossover. That makes it difficult to “freeze in” any produced asymmetry before washout reactions erase it.
QCD could, in principle, violate CP via a term controlled by an angle θ (theta). But experimental bounds (especially from electric dipole moment searches) imply θ is extremely tiny. So the Standard Model does not secretly gain a large extra source of CP violation there.

This is why many baryogenesis ideas invoke new physics: extra Higgs fields, new particles with new complex phases, or leptogenesis in which CP-violating lepton-number generation is partially converted into baryon number by electroweak processes.

What experiments actually measure when they say “CP asymmetry”

When experiments such as LHCb, BaBar, or Belle II report CP violation, they are not measuring “CP violation” directly as a standalone quantity. They measure differences between a process and its CP-conjugate.

A standard observable is the direct CP asymmetry:

A_CP = [ Γ(P → f) − Γ(P̄ → f̄) ] / [ Γ(P → f) + Γ(P̄ → f̄) ]

Here:

P is a particle (often a neutral meson like B^0, D^0, or K^0)
P̄ is its antiparticle
f is some final state, and f̄ is the CP-conjugate final state
Γ means the decay rate (probability per unit time)

Experimentally, this starts as a difference in event counts, but it cannot be taken at face value. Collaborations correct for “fake asymmetries”: different detection efficiencies for positive and negative tracks, production asymmetries in collisions, backgrounds, and particle mis-identification. The goal is to isolate a genuine physics asymmetry in the underlying rates.

Why does a non-zero A_CP happen? Usually because of interference. You need at least two different decay pathways leading to the same final state. They must contain:

different weak phases (from complex couplings such as CKM elements), and
different strong phases (from QCD rescattering, which supplies a relative phase that does not change sign under CP).

In neutral meson systems there is also mixing-induced CP violation: a meson can oscillate into its antiparticle before decaying. Then experiments use time-dependent rate asymmetries; the oscillation pattern contains the CP-violating phase information.

So “CP asymmetry” is a carefully defined comparison between conjugate processes, designed to pick out tiny CP-odd parts of quantum amplitudes. It is exactly the kind of bias Sakharov required — but, in the Standard Model, it seems too small to explain why the Universe ended up made of matter.