At 10:03 AM 01/10/2001 -0400, Frank Spear wrote:
>Dugald and others,
>I can't resist weighing in on this discussion.
>Dugald started off with an elegant equilibrium discussion of how to 
>view buffering, but ended with a kinetic argument...

...that rests on the same strong foundation as an elegant technique of
determining a P-T path from a garnet zoning profile (Spear & Selverstone
1983 CMP 83,348-357) - the postulation that diffusion within garnet is
negligible. :-)
>A useful quantity to calculate for your s-buffer assemblage in the 
>KFMASH-O2 system for the assemblage  Grt+Bt+Mu+Mt+Qtz+aqueous vapor 
>phase is the partial derivative (¸Xalm/¸f(O2))P,T. This will provide 
>an equilibrium view of the sensitivity of the garnet Fe/Mg to changes 
>in f(O2). Another useful quantity is (¸Mgrt/¸f(O2))T,P (where M is 
>the amount of the phase), which will provide a measure of the amount 
>of garnet (or any other phase) that will be consumed/produced as a 
>function of f(O2). It is also useful to quantify the amount of oxygen 
>required, in which case a useful derivative  is (¸Mgrt/¸M(O2))T,P.

How about the partial derivative (dM(O2)/df(O2))T,P),normalized to 100
oxygens in the less oxidized assemblage? This could be used to compare the
limited equilibrium mass-transfer capacity of various s-buffer reactions to
the unlimited mass transfer of any r-buffer reaction.  But none of the
above will ease my suspicion that many assemblages with F>2 are former
divariant assemblages that ran out of r-buffering capacity due to the
relatively fast rate of equilibration of divariant assemblages.
Mass-transfer of oxygen would not be necessary; loss or gain of _any
component_ would lead to complete consumption of one mineral and hence loss
of r-buffering for all components of the assemblage (including oxygen).  

>Since most minerals are dominantly oxygen, I suspect one would find 
>that most metamorphic assemblages are pretty effective s-buffers. 
>That is, the value of f(O2) is pretty nearly totally controlled by 
>the internal equilibrium.

Most minerals being dominantly oxygen is not sufficient; f(O2) buffers must
be able to produce or consume O2 and this depends on altervalent cations
(principally Fe). But for sure the equilibrium f(O2)-buffering capacity of
crustal rocks is huge. In a 1948 paper in J Geol, Tom Barth showed that
lateritic weathering of average continental crust to a depth of ~240m would
consume virtually all the oxygen in the atmosphere. 

>I know of no examples where major changes 
>in f(O2) have been documented in regionally metamorphosed rocks as a 
>result of infiltration.

Me neither. My intersecting isograds in southern Ontario (1970 J Petrol)
were cited by Rice and Ferry (1982 Reviews in Mineralogy 10) as evidence of
km-scale "infiltration and the external control of fluid composition", yet
sillimanite-zone metapelites containing either graphite or abundant
magnetite occur within a few meters of each other, reflecting a steep
gradient of f(O2) that evidently was not much flattened during the

It looks as if Rice and Ferry fuzzified "buffering" quite deliberately:
"Following the treatment of buffers in the chemical literature, we adopt
the more general concept of buffering phenomena. Buffers are considered as
those equilibria that _resist changes_ in certain intensive variables that
are imposed on the system under consideration. The conventional petrologic
definition of buffers -- equilibria that _fix_ intensive variables -- is
then simply a special case of this more general definition of buffers." (op
cit p.268). 

Cheers, Dugald

Dugald M Carmichael                    Phone/V-mail: 613-533-6182
Dept of Geological Sciences and Geological Engineering
Queen's University                              FAX: 613-533-6592
Kingston  ON  K7L3N6             E-mail: [log in to unmask]