The New Physics We Awaited for 20 Years May Never Have Existed: The Muon g-2 Story

Forum topic · 小凯 · 2026-03-19

Summary

For two decades, physicists suspected the muon's anomalous magnetic moment deviated from the Standard Model's prediction, hinting at undiscovered particles or forces. This post recounts the full arc of the muon g-2 saga: Dirac's original g=2 prediction, the 2001 Brookhaven E821 result showing a 3.5-sigma tension, and Fermilab's E989 experiment, which in 2021 pushed the discrepancy to 4.2 sigma. The twist came from lattice QCD: on the very same day as Fermilab's 2021 announcement, the BMW Collaboration published a first-principles lattice calculation of hadronic vacuum polarization that disagreed with the traditional data-driven approach, shrinking the gap with experiment to about 1.5 sigma. In June 2025, Fermilab's final result — 116592040(25) × 10⁻¹¹ — matched the lattice QCD prediction (116592040(55) × 10⁻¹¹) exactly, while the data-driven prediction (116591810(43) × 10⁻¹¹) missed by 4.5 sigma. The conclusion: the hoped-for 'new physics' signal was likely a systematic error in the data-driven method, and the Standard Model remains correct to nine decimal places. The author frames this null result as a triumph of scientific self-correction and first-principles computation.

Have you ever imagined using a supercomputer to 'weigh' a ghost? Not as a metaphor — literally weighing, through pure mathematics and code, in a virtual universe made of 0s and 1s, the tiny influence one untouchable particle exerts on another.

That is the story that unfolded in particle physics in 2025. Its protagonist is a particle called the muon — the electron's 'fat cousin,' about 207 times heavier. At its heart is a mystery that has troubled physicists for 20 years: the muon's magnetism seemed 'wrong.'

And the latest answer to that mystery may surprise you — perhaps the 'new physics' we have been waiting for never existed.

But don't be disappointed yet. Behind this 'discovery of no discovery' hides one of the most fascinating stories about how humans come to understand the universe.

🧲 The Little Secret of Magnetism: Why Isn't g Exactly 2?

Start with the simplest concept. In 1928, British physicist Paul Dirac wrote down a famous equation predicting that the electron (and its sibling particles) should have a magnetic moment with a remarkably clean value: g = 2.

Here g is the 'g-factor,' describing how strong a particle's magnetism is. But in 1947, Isidor Rabi found that the electron's g-factor is not exactly 2 — it is slightly larger. This 'extra' part is called the anomalous magnetic moment.

Why isn't g exactly 2? The answer: the vacuum is not empty. In the quantum world, the vacuum is a boiling sea — particle-antiparticle pairs pop in and out of existence at every instant.

🎯 The Muon: The Electron's Heavy Cousin and Its Mystery

The muon was discovered in 1936. Its properties are nearly identical to the electron's — only heavier, about 207 times the electron's mass.

Precisely because the muon is heavier, physicists have long regarded its anomalous magnetic moment as the golden probe for new physics. If it disagrees with the Standard Model, that could signal new particles or forces.

🧪 Brookhaven's Surprising Finding

In 2001, the E821 experiment at Brookhaven National Laboratory published its result: a 3.5σ discrepancy between experiment and theory. Not enough to declare new physics, but enough to excite physicists.

🔬 Fermilab Takes the Relay

In 2013, Fermilab's E989 experiment began running, aiming to improve the measurement precision by a factor of 4. On April 7, 2021, the results were announced — in perfect agreement with Brookhaven, pushing the tension to 4.2σ.

🖥️ The Birth and Breakthrough of Lattice QCD

But in the theoretical calculation, the contribution of hadronic vacuum polarization (HVP) was especially tricky. QCD is non-perturbative at low energies, so traditional methods fail.

So physicists invented lattice QCD — discretizing four-dimensional spacetime into a grid and solving the QCD equations on the lattice. This is a first-principles computation, derived entirely from theory.

🎭 The 2021 Shock

On the very same day Fermilab announced its result, the BMW Collaboration published a paper in *Nature* — using lattice QCD to compute HVP, with a result differing from the data-driven approach by 2.2σ.

When the BMW result was plugged into the Standard Model, the predicted value increased and the gap with experiment shrank to about 1.5σ — no longer a significant deviation.

🔮 The Final Answer in 2025

In June 2025, Fermilab announced its final result:

Experimental value: 116592040(25) × 10⁻¹¹
Lattice QCD prediction: 116592040(55) × 10⁻¹¹ ✓ consistent
Data-driven prediction: 116591810(43) × 10⁻¹¹ ✗ off by 4.5σ

The verdict is clear: lattice QCD agrees with experiment; the data-driven approach does not.

🤔 But Does This Mean Failure?

Quite the opposite! This is a beautiful demonstration of the scientific method:

1. Lattice QCD achieved a major breakthrough — for the first time, we can compute such complex hadronic quantities from first principles. 2. It showcased the power of scientific self-correction — competing methods were tested, and the truth ultimately emerged. 3. The Standard Model remains correct at the most precise level — matching to nine decimal places.

🌟 Closing Thoughts

Science has never been about getting the answer you want. It is about asking the right questions and pursuing the answer with honesty and rigor — whatever that answer may be.

The 'new physics' signal awaited for 20 years was most likely caused by some systematic error in the data-driven method. But we gained something equally valuable: a deeper understanding of known physics, and a clearer expectation of the unknown.

---

📚 References: 1. Muon g-2 Collaboration, PRL 131, 161802 (2023) 2. BMW Collaboration, Nature 593, 51–55 (2021) 3. Muon g-2 Theory Initiative, arXiv:2505.21476 (2025) 4. Aoyama et al., Physics Reports 887, 1-166 (2020) 5. Physics World, Feb 2025