How can an experiment feel the presence of particles it is far too weak to create? Quantum field theory gives a precise answer. Every process is shaped by all possible intermediate states, including those too heavy to produce directly. These virtual contributions are not particles in any literal sense. They are terms in a quantum amplitude, and they still leave fingerprints on the rates at which ordinary particles transform.

Those fingerprints are clearest in rare decays. The Standard Model forbids some transitions at the simplest level and permits them only through quantum loops. Their rates are suppressed by factors of a million or more. That suppression is useful: a small contribution from unknown physics does not have to compete with a loud background of ordinary processes.

This is the logic of indirect discovery, and it has worked before. Both the charm quark and the top quark were anticipated from rare processes years before anyone produced them. Today, decays of B mesons carry the same promise. For a decade they also offered tantalising hints that something beyond the Standard Model might be stirring.

In the Standard Model, the photon, the Z boson, the gluon and the Higgs all couple to quarks without changing their flavour. This is no accident. When the weak-interaction states of quarks are rotated into their mass states, the unitarity of the CKM matrix keeps every neutral current diagonal. A bottom quark cannot emit a Z and emerge as a strange quark. At the simplest level, the process does not exist.

It can still happen, but only indirectly. The bottom quark emits a W boson, becomes an up-type quark, and reabsorbs the W as a strange quark. These loop processes, called penguin and box diagrams, each pay a price of roughly g²/16π². There is a further cost. Contributions from the up, charm and top quarks partially cancel, a feature known as the GIM mechanism. If the three quarks had equal masses, the cancellation would be exact. Only their mass differences survive.

The result is extraordinary rarity. The decay of a B_s meson into two muons, which is also helicity-suppressed, occurs about three or four times in every billion decays. Such tiny rates are exactly what makes these processes sensitive. A new heavy particle could run inside the loop, or it could mediate the transition directly. Because the Standard Model amplitude is itself loop-suppressed, new physics competes on unusually even ground.

History rewards this reasoning. In 1970 Glashow, Iliopoulos and Maiani proposed a fourth quark, charm, to explain why certain kaon decays were so strongly suppressed. Gaillard and Lee then estimated its mass from kaon mixing before the J/ψ was discovered. In 1987 the ARGUS experiment observed unexpectedly large B meson mixing, which pointed to a very heavy top quark. Fermilab found the top quark in 1995.

Takeaway

A suppressed process is a quiet room: the less the Standard Model has to say, the more clearly any new voice can be heard.

B mesons are well suited to this kind of listening. The bottom quark is heavy, at about 4.2 GeV, so it has room to decay into a rich variety of final states. Its transitions to strange quarks are dominated by loops containing the top quark, the heaviest particle we know. Studying b→s processes is therefore a way of interrogating the most massive corner of the Standard Model, and whatever may lie beyond it.

The natural language for this is effective field theory. Heavy particles are mathematically integrated out, and their influence is compressed into a few numbers called Wilson coefficients. Three of them, C₇, C₉ and C₁₀, describe most of the physics of b→sμμ. The idea has a distinguished ancestor. Fermi's four-fermion theory of beta decay encoded the W boson long before it was discovered. Its coupling strength even pointed toward the right energy scale.

Experiments such as LHCb, Belle II and CMS measure several complementary observables. Branching fractions count how often a decay occurs. The angular distribution of B→K*μμ, with its four-body final state, separates different underlying interactions. Ratios such as R_K compare decays to muons with decays to electrons. The Standard Model predicts these ratios lie very close to one, and in the ratio most hadronic uncertainties cancel cleanly.

The reward is reach. Suppose new physics couples with a strength comparable to the Standard Model's own. Then precise flavour measurements probe energy scales of tens of TeV, well beyond what the LHC can produce directly. The reach depends on assumptions, though. A new sector with weak or carefully aligned couplings could hide. The measured B_s→μμ rate, which agrees well with prediction, already rules out many otherwise plausible ideas.

Takeaway

Effective field theory turns ignorance into a parameter: you need not know what lives at high energies to measure how strongly it whispers at low ones.

In 2013, LHCb reported that an angular observable in B→K*μμ, called P₅′, deviated from prediction by roughly three standard deviations in certain kinematic regions. Measured branching fractions for decays such as B→Kμμ and B_s→φμμ also fell systematically below expectation. Global fits found a coherent explanation: a negative shift in C₉, as though some new interaction were subtly interfering with the muons.

Lepton universality ratios seemed to sharpen the story. From 2014 onward, LHCb measured R_K below one, reaching a 3.1σ tension in 2021. Theorists proposed leptoquarks and new Z′ bosons that treated muons and electrons differently. Then, in December 2022, a refined LHCb analysis found R_K and R_K* consistent with the Standard Model. The analysis had improved its handling of misidentified electrons, which are notoriously hard to reconstruct, and the lepton-universality anomaly dissolved.

The muon-channel tensions remain, but their meaning is contested. Loops of charm quarks, sometimes called charming penguins, can mimic a C₉ shift because they couple to the lepton pair through a photon, identically for every lepton flavour. Whether the residual discrepancy is new physics or underestimated strong-interaction dynamics is still an open debate. Meanwhile, Belle II reported evidence in 2023 for B⁺→K⁺νν̄ at a rate roughly 2.7σ above prediction. This is another hint that more data must test.

The episode teaches patience rather than cynicism. A rare decay measures a sum: Standard Model loops, hadronic complications, detector effects, and perhaps something genuinely new. Separating these requires lattice QCD calculations, independent channels and growing datasets from upgraded LHCb and Belle II. A real signal will survive that scrutiny. The scrutiny itself is how the field earns its eventual claims.

Takeaway

An anomaly that vanishes is not a failure of the method but a demonstration of it; the discipline that dissolves false signals is the same discipline that will confirm a true one.

Dirac predicted antimatter by taking an equation more seriously than intuition allowed. Rare decays ask for a similar faith in the formalism. They assume that the quantum vacuum carries the imprint of every particle nature permits, whether or not we have ever produced it.

That faith has been rewarded twice already, with charm and with top. Whether B mesons will reward it a third time remains genuinely uncertain. The recent retreat of some anomalies is a reminder that nature does not owe us a discovery on any schedule.

Still, the principle endures. Matter as we know it is quietly shaped by possibilities that never fully materialise, and a few transitions in a billion may be enough to hear them.