Self-Reversal of Remanent Magnetisation of Basalts - Origin, Mechanisms and Consequences

One of the main goals of current palaeomagnetic research is the
attempt to acquire high-resolution data on palaeodirections and
-intensities in order to obtain detailed information about the
Earth's magnetic field in the geological past. The material best
suited for such studies are basaltic rocks. For these high-quality
directional investigations and especially for palaeointensity
determinations, a profound knowledge about the stability,
magnetomineralogical character and the domain state of the carriers
of remanence is imperative. The emphasis of the present work was
placed on the investigation of basalts exhibiting partial or
complete self-reversal of natural remanent magnetisation (NRM).
This phenomenon is not an exotic rarity but a widespread
characteristic of many basaltic rocks. However it remains usually
unnoticed by routine palaeomagnetic measurements as it requires
special techniques for its detection. In this work samples from
Olby and Laschamp (France) and Vogelsberg (Germany) showing the
phenomenon were studied with rock magnetic, microscopic and
microanalytical techniques in order to identify the carriers of
self-reversed remanent magnetisation. Further aims were to
determine the exact mechanism of self-reversal acting in basalts
and to evaluate the consequences for the reliability of
palaeomagnetic data. On the basis of the experimental work a
numerical model was developed which shows that, from the physical
point of view, the observed magnetomineralogy is capable of causing
self-reversal. The present work provides the following new
insights: The phenomenon is caused by two magnetic phases with
different blocking temperatures which are magnetically coupled. The
lower blocking temperature corresponds to the primary
titanomagnetite (mother phase) crystallising from the basaltic
magma. The remanence with higher blocking temperature is carried by
titanomaghemite (daughter phase) evolving from the primary
titanomagnetite by partial low-temperature oxidation. The daughter
phase forms narrow bands along cracks in the otherwise unaffected
mother phase particles. This yields a close side-by-side assemblage
of titanomagnetite and titanomaghemite with markedly different
magnetic properties in one and the same grain. By applying the
various microscopic techniques on identical grains, it was possible
to directly correlate magnetomineralogy with magnetic domain
structure. Numerical calculations of remanence acquisition
demonstrate that two-phase particles with the experimentally
observed geometry and magnetic properties are able to acquire a
partially or completely self-reversed remanent magnetisation. The
calculations also prove that the two magnetic phases present in the
studied samples are coupled by magnetostatic interaction. The
experimental results indicate that the low-temperature oxidation
process responsible for the formation of the second magnetic phase
takes place at temperatures at or above the blocking temperature of
this daughter phase during primary cooling. This titanomaghemite
phase is thus carrying a stable remanence in direction of the
ambient magnetic field. Although the original titanomagnetite as
the mother phase is in a strict sense the primary magnetic mineral,
it does not carry the primary magnetic remanence but is at least in
part magnetostatically coupled to the titanomaghemite. Therefore,
its remanence is - at least in part - antiparallel to the external
field. MFM domain observations present evidence that the mother
phase is in the magnetic multidomain range. Hence, its magnetic
remanence is not stable and is replaced by a viscous overprint
acquired at ambient temperatures. In contrast, the daughter phase
has a higher coercivity due to oxidation induced stresses and an
increased domain width. Regarding the samples from Olby, these
magnetomineralogical investigations directly lead to new arguments
in favour of the existence of the Laschamp geomagnetic event: As
the high blocking temperature daughter phase carries a stable
remanence in direction of the external magnetic field, the local
geomagnetic field direction was indeed reversed at the time of
emplacement. Due to their complex magnetomineralogy and remanence
acquisition, samples exhibiting partial or complete self-reversal
are not suitable for palaeointensity determinations. In order to
identify and exclude such samples in the course of such
experiments, a modification of the existing Thellier-Thellier
method is proposed. Additionally, this new procedure is also able
to detect remanence carried by multidomain (MD) particles. The
method substantially improves the reliability and quality of
palaeointensity estimates as multidomain behaviour is among the
most common reasons for erroneous results in Thellier-type
palaeointensity determinations.