A statistical analysis of the Stanford fractional charge experiment

We report arguments indicating that the residual fractional charges observed on niobium spheres in the experiments at Stanford are consistent with a statistical distribution of one kind of fractionally-charged object among the spheres, with a large average number of charges per sphere. Our statistical analysis enables us to conclude with 96% confidence that the spheres annealed on a tungsten substrate have a substantially higher concentration of fractional charges than those annealed on a niobium substrate. We draw attention to circumstantial evidence that such fractional changes were transferred from the tungsten substrate to the surfaces of the niobium spheres. We point out that surface and grain-boundary segregation effects, which previously have not been considered in assessing this experiment, could play a major role in enhancing the concentration of fractionally charged impurities on the surface of the tungsten substrate in contact with the niobium spheres. We conclude that the Stanford experiment would be consistent with a concentration of 1 fractional charge in 1013 to 1014 tungsten atoms, which would not be contradicted by other searches for fractional charges in tungsten. (orig.)