Joern,
The alteration products of cordierite are a messy mixture of clays, some
of which are amorphous and others intergrown on a near-atomic level. They
may dehydrate as well as undergo alkali migration. Usually the Al and Si
increase during this process. However, the carbon coating can also become
damaged during dehydration, and there may be sample charging, which gives
successively lower totals. it is difficult to obtain an adequate polish on
clay minerals, leading to low totals anyway (i.e., lower than expected from
the anticipated OH and H2O contents). These materials are best studied by
TEM although a first step is certainly to undertake SEM, EMP and XRD work.
There was an iconoclast, Prof. Stanton, who proposed some 20 years ago
in a series of articles mostly in obscure Australian journals, that such
clays seen in cordierite and sillimanite were the vestiges of prograde
metamorphism, and that pelitic isograds represent direct reaction from
sedimentary minerals. He concluded that metamorphic isograds and
metamorphic facies are controlled by sedimentary facies environments rather
than by regular changes in P-T! Few workers have noticed this work or have
taken it seriously (fortunately).
Eric
>Dear all,
>
>i am working on natural retrograde alteration products of cordierite.
>During the EM analyses i made an unusual observation (at least to me
>unusual):
>In extremely fine grained homogenous areas i increased the EM beam diameter
>up to 10-15 µm to hinder a loss of alkalis. Astonishingly, alkalis were
>more or less indifferent to the the size of the beam diameter, but Al was
>not. The sum of al2o3 went down up to 8 wt% (absolutely) when i reduced the
>beam diameter to 1 µm. Has anybody made similar observations?? Does anbody
>know what this could mean??
>
>cheers
>
>Joern Ogiermann
>
>***************************************
> Joern Ogiermann
>
> Mineralogisches Institut
> Universitaet Heidelberg
> Im Neuenheimer Feld 236
> 69120 Heidelberg
> GERMANY
>
> Tel.:++49(0)6221/54 60 15
>
> [log in to unmask]
>***************************************
|