Ferromagnetic resonance (FMR) is used to measure the magnetic anisotropy fi
eld H-k and its dispersion of poly-crystalline recording media. This method
does not use any single crystal substrate. A high-intensity field is appli
ed vertically to a sample surface to completely saturate the magnetization
and a high-frequency (over 50 GHz) magnetic field is applied in the surface
. An FMR simulator based on the Landau-Lifshitz-Gilbert equation confirms t
hat the magnetic (exchange and/or magnetostatic! interactions have very lit
tle effect on the resonant frequency in the configuration used. The linewid
th of the FMR spectrum is affected by the H-k dispersion and damping consta
nt. Methods using a torque or magnetization curve include magnetic interact
ions, resulting in a systematic error to the value of H-k. This FMR method
is applied to CoCrPt-alloy thin-film media with a Cr-alloy underlayer grown
on a glass substrate. A conventional torque method is also used for compar
ison. The axis of hexagonal-close-packed Co magnetic grain is distributed a
t random in the surface. In an FMR measurement, external magnetic field is
applied vertically to the surface up to 55 kOe for two frequencies (71 and
117 GHz). The obtained H-k is 7.0+/-0.2 kOe, which is a little higher than
the H-k of 6.4 kOe estimated with the torque loss method. Important physica
l parameters that affect the recording performance are also estimated using
the linewidth of FMR spectra for different frequencies with a micromagneti
c simulation: H-k dispersion=1.2+/-0.5kOe, Gilbert's damping constant alpha
=0.03+/-0.01, and a g value=2.19+/-0.02. (C) 1999 American Institute of Phy
sics. [S0021-8979(99)52108-7].