A collimator with a magnetic personality?

We report an unusual technical problem with a gamma camera fitted with a pinhole collimator. An unexpected decrease of approximately 20 keV in the energy of the 99mTc photopeak is seen if the (circular) collimator is rotated relative to the camera head. The centreline of the photopeak varied over the range 118-141 keV depending on the orientation of the collimator. The count rate and photopeak width changed concurrently. Image uniformity was severely degraded for lower peak energies. The reason for this behaviour is that the pinhole collimator had become magnetized. The magnetic field was measured and had maximum values of 1.0 mT at the edge and 0.1 mT near the camera face. The magnetic field strongly influenced the response of the photomultiplier tubes. We conclude that acceptance testing of gamma cameras should include a check for collimator magnetization.