Solar System limits on gravitational dipoles

The gravitational dipole theory of Hadjukovic (2010) is based on the hypothesis that antimatter has a negative gravitational mass and thus falls upwards on the Earth. Astrophysically, the model is similar to but more fundamental than Modified Newtonian Dynamics (MOND), with the Newtonian gravity $g_{_\mathrm{ N}}$ towards an isolated point mass boosted by the factor $\nu = 1 + \left(\alpha /X \right) \tanh \left(\sqrt{X}/\alpha \right)$ , where $X \equiv g_{_\mathrm{ N}}/a_{_0}$ and $a_{_0} = 1.2 \times 10^{-10}$ m s^-2 is the MOND acceleration constant. We show that α must lie in the range 0.4-1 to acceptably fit galaxy rotation curves. In the Solar System, this interpolating function implies an extra Sunwards acceleration of ${\alpha a_{_0}}$ . This would cause Saturn to deviate from Newtonian expectations by 7000(α/0.4) km over 15 yr, starting from known initial position and velocity on a near-circular orbit. We demonstrate that this prediction should not be significantly altered by the postulated dipole haloes of other planets due to the rather small region in which each planet's gravity dominates over that of the Sun. The orbit of Saturn should similarly be little affected by a possible ninth planet in the outer Solar System and by the Galactic gravity causing a non-spherical distribution of gravitational dipoles several kAU from the Sun. Radio tracking of the Cassini spacecraft orbiting Saturn yields a 5σ upper limit of 160 m on deviations from its conventionally calculated trajectory. These measurements imply a much more stringent upper limit on α than the minimum required for consistency with rotation curve data. Therefore, no value of α can simultaneously match all available constraints, falsifying the gravitational dipole theory in its current form at extremely high significance.