An essential part of the g−2 experiment is to measure the 1.45T magnetic field across the muon storage ring to 70ppb precision. This is done using several hundred magnetometer probes, which determine the magnetic field strength from the NMR frequency of protons inside a water sample.

All these probes must be calibrated against a standard probe, to correct for local shifts due to magnetic properties of the probe material. This is done using a calibration probe with a spherical water reference sample, for which the magnetic field shift due to the electronic screening can be calculated, and that due to the bulk magnetic susceptibility is zero.

The systematic uncertainty on this probe was not so significant for the E821 Brookhaven experiment. But in the new Fermilab experiment, it is predicted to be the largest single uncertainty contributing to the magnetic field measurement.

We have started a programme in Oxford to develop a new standard calibration probe using a ³He gas sample, with a negligible diamagnetic susceptibility. This avoids the temperature and sample shape dependence which limits the accuracy of the water calibration probe to ~35ppb. But using ³He introduces new challenges as we need to use laser optical pumping to polarize the sample.