Skip to content

The Gauge Origin Problem

Where the Gauge Origin Enters

A uniform magnetic field B is represented in quantum mechanics not directly, but through a vector potential A, defined so that:

B = ∇ × A

This definition doesn't pin down A uniquely — for a uniform field, a common choice is:

A(r) = ½ B × (r − R₀)

where R₀ is an arbitrary reference point, the gauge origin. Physically observable quantities must not depend on where R₀ is placed; different choices are supposed to describe exactly the same physics.

Why It Matters Computationally

With an exact, complete basis set, the choice of R₀ genuinely has no effect on the computed shielding — the gauge invariance holds exactly. In any real calculation, however, the basis set is finite, and this invariance is only approximate. Different choices of gauge origin can give numerically different shielding values for the same molecule at the same level of theory — an artifact of basis-set incompleteness, not a real physical effect.

This is a practical problem: a shielding constant that depends on an arbitrary coordinate choice isn't usable for comparing to experiment or between different molecules.

Two Ways to Fix It

There are two general strategies:

  • Use an extremely large, flexible basis set, so the calculation approaches the complete-basis-set limit where gauge invariance is restored — computationally expensive, and still only approximate.
  • Change the basis functions themselves so that gauge invariance holds exactly, even with a finite basis — this is the approach Gaussian uses, via Gauge-Including Atomic Orbitals (GIAO).