Correction factor in nondiffusive Hall magnetometry

It is demonstrated how the correction factor α used in Hall magnetometry of localized magnetic field profiles depends on the sample geometry and the electron mean free path, in the quasiballistic and ballistic regimes. The frequently used approximation of a constant correction factor close to 1 is generally not justified, and may lead to large errors in the determination of the magnitude of the magnetic fields. α depends in a nontrivial way on the magnetic structure and the Hall cross and may vary considerably. The sensitivity peaks when the characteristic cyclotron diameter equals the size of the Hall cross.