E.G. Adelberger, M.G. Harris and B.R. Heckel

We have built¹ a new electron spin polarized torsion pendulum to test for a spin-dependent interaction with the monopole-dipole form² , where m=/c is the mass of a low-mass mediating boson. The torsion pendulum has been installed in the 'Eotwash' rotating apparatus.³ Such an interaction would produce a field whose magnitude and direction depend on the range , and upon variations in the local topography. For our pendulum, a signal of 10 nrad would correspond to a measurement of 1.2 × 10^-20eV electron spin.

To check for systematic errors, we have applied large gravity gradients (by reversing the compensators that normally cancel the gradients),⁴ indicating an effect of at most 5 nrad of pendulum deflection under normal conditions. We have also investigated effects arising from the coupling of the remnant magnetic dipole moment of the pendulum to the lab-fixed magnetic field (reduced by a 3-axis Helmholtz coil, and 3 layers of mu-metal shielding). We installed a 'dummy' pendulum with a large, known magnetic dipole moment (about 1000 times larger than that of the spin-polarized pendulum). By observing the deflections of this pendulum both in the normal operating magnetic field, and in a large external field, we were able to establish a maximum magnetic systematic error also of about 5 nrad.

We observe spin pendulum signals of the order of hundreds of nrad, and (worse) a lack of reproducibility in measurements taken more than a week apart. We have recently isolated instabilities in the rotation rate of the apparatus, and in the electronics, which presumably caused the signals. The reproducibility problem has now been fixed. The remaining work which needs to be done involves making further and more careful measurements of the systematic errors mentioned above, as well as other minor ones (such as temperature effects), and a long-term collection of spin-potential data.