Testing the independence of partial thermoremanent magnetizations of single‐domain and multidomain grains: Implications for paleointensity determination

The Thellier method of paleointensity determination requires that partial thermoremanent magnetizations (pTRMs) be additive, mutually independent, and reciprocal in their thermal blocking and unblocking. We tested the independence law by thermally demagnetizing a sum of orthogonal pTRMs produced by replacing part of an original TRM( T C , T 0 , H ) by pTRM( T , T 0 , H ), with H rotated 90° between the TRM and the pTRM ( T C , Curie temperature; T 0 , room temperature). The composite remanence simulates thermal remagnetization in nature. For single‐domain grains, thermal demagnetization resolved the orthogonal pTRMs cleanly. Two remagnetization temperatures were tried, T = 400°C and 550°C; the latter gave a better test because the two pTRMs were almost equal. Directions were recovered to within ∼5° for three of the four pTRMs, trajectories on vector diagrams intersected at exactly 400°C or 550°C, and the Arai plots were linear and yielded the correct field intensity for both pTRMs. However, multidomain magnetite violated the independence law. A remagnetization temperature T = 400°C produced almost equal pTRMs, but in thermal cleaning, because of overlap of their unblocking spectra they no longer appeared orthogonal. Each was displaced 13–15° toward the direction of the other as judged by quasi‐linear segments on a vector diagram. The angle between the fields that produced the pTRMs thus appeared to be ∼60°, not 90°. Arai plots for both pTRMs were strongly curved and yielded no usable paleointensity estimate. Phenomenological modeling correctly predicted the main results of the independence experiments.