A large number of partial hysteresis curves called first order reversal curves (FORCs) are commonly measured to determine the FORC distribution of a sample. These distributions, as represented in FORC diagrams, provide information about micro coercivities and magnetic particle interactions in samples, which ultimately aid in deciphering the composition and grain sizes of the magnetic minerals. Because a variable number of FORCs are used in making FORC diagrams, with each curve potentially representing hundreds of data points, the time it takes to produce a single FORC diagram varies from only a few minutes when only a few curves are measured to several hours when over a hundred curves are measured. Using more curves in building a FORC diagram generally gives a more detailed look at the coercivity distribution and interactions between magnetic particles, but it also increases the measurement time. As might be expected, there is a point at which there is no significant gain in resolution no matter how many additional curves are measured. Our goal is to provide some "rules of thumb" for producing FORC diagrams for natural materials that fully resolve the FORC distributions with as few curves as possible. Such information is extremely valuable for studies involving large numbers of samples or for studies where measurement time is limited, as might be the case for researchers visiting other laboratories. Even though the answer depends somewhat on the magnetic properties of the samples, experiments with a range of rock and mineral types provide useful measurement strategies. For example, experiments with oceanic basalt samples indicate that, when using fewer than 40 curves, the FORC distribution is overly smoothed. In contrast, FORC diagrams produced with more than 100 curves reveal no new features relative to FORC diagrams produced with around 80 curves. At some point, variations between densely spaced data are more indicative of instrument noise than magnetic properties. Besides the number of curves measured, factors such as the size of the space examined in the FORC diagram and the amount of smoothing done in processing the raw data are important. These factors mean that we cannot make a single set of definitive rules governing how FORC diagrams should be produced, but we can provide some general rules of thumb for the efficient production of FORC diagrams that sufficiently resolve the underlying FORC distributions.

[ return to abstract list ]