Dimensional regularization

Dimensional regularization replaces the spacetime dimension d with a complex parameter, typically writing d = 4 − 2ε, where ε is a small complex number. Loop integrals that diverge in four dimensions become finite for generic ε, and the divergence reappears as poles in 1/ε as ε → 0. These poles are then absorbed into renormalization constants, leaving finite physical predictions.¹ The method is particularly elegant because it respects Lorentz invariance and gauge symmetry, unlike cutoff regularization which breaks these symmetries explicitly.

The technique was developed independently by Gerard 't Hooft and Martinus Veltman, and by Carlos Bollini and Juan Giambiagi, both in 1972. 't Hooft and Veltman used it to prove the renormalizability of non-Abelian gauge theories, work that earned them the 1999 Nobel Prize in Physics.² Bollini and Giambiagi, working in Argentina, published their version in Nuovo Cimento the same year, but their contribution remained less widely recognized for decades.³ The method quickly superseded older schemes like Pauli–Villars regularization in most perturbative calculations.

Dimensional regularization is the default method in perturbative quantum chromodynamics (QCD) and electroweak calculations. It underlies the modified minimal subtraction (MS-bar) scheme, which is the standard renormalization scheme in particle physics.⁴ It is also essential in calculations of anomalous dimensions, beta functions, and critical exponents in statistical field theory. A subtlety arises with γ₅ in dimensional regularization, because the fully antisymmetric tensor ε_μνρσ only exists in integer dimensions; various prescriptions (e.g., 't Hooft–Veltman, naive, and dimensional reduction) handle this differently, leading to scheme-dependent finite terms.⁵

Dimensional regularization has surprising connections to other fields. It is closely related to the Riemann zeta function and analytic continuation, as the divergent integrals often evaluate to zeta values at negative integers.⁶ In mathematics, it inspired the method of dimensional renormalization in the study of Feynman integrals. A niche application is in the calculation of Casimir forces, where it can be used to regularize sums over modes. Also, the technique works for any complex dimension, and it has been extended to fractional dimensions in the context of fractal spacetimes, though such extensions remain speculative. The original papers by Bollini and Giambiagi were initially rejected by some journals because the idea seemed too formal, but they were eventually published and recognized.³

Ultraviolet divergence

A divergence in a Feynman integral arising from high-momentum (short-distance) behavior.

Renormalization

The procedure of absorbing divergences into redefinitions of physical parameters.

γ5

The chirality matrix in Dirac algebra, which requires special treatment in dimensional regularization.

MS-bar scheme

A renormalization scheme that subtracts the poles and a specific constant (ln 4π − γ) at each order.