Back in the day, pre and post Harold as a student, we followed a routine as he describes.  However, in general, I did not assume stoichiometric Al,  I did assume nothing else in Si site.  So if the analysis had a bit of Fe+3 for Al site, that was fine.  However, total Fe was generally off.  I preferred an almandine garnet for Fe rather than fayalite or hedenbergite.  My suspicion was that the the high CN for Fe in garnet (viii) affected the Fe based on Fe (vi) in olivine etc.

A running rule was that a good garnet analysis totaled slightly over 101 (or was it less than 99?).  Sorry , senior moment here...

Basically, black magic.

Lincoln


On Mar 28, 2018, at 7:18 PM, Stowell, Harold <[log in to unmask]> wrote:

Hi all

My experience is also perplexing.
A few ideas and observations:

1 Most garnet in pelites, diorite gneiss, and eclogite is stoichiometric and has likely has minimal amounts of Fe3+
Therefore, criteria based on 8 cations, with 3 Si and 2 Al are generally OK. I strive for +/-0.02 cations for each of these numbers.

2 Analyses can be difficult if the stds are significantly different compositions than the unknowns. Remember B-A corrections etc….

3 Si, Al, Fe are generally where the difficulties are most obvious.

4 For pelitic amphibolite facies garnet with lots of Fe, I use spessartine garnet [fairly pure] for Si and Al std. I generally use fayalite for Fe because it is hi Fe - similar to many grt.

5 In general, #4 has been adequate for the Ca and Mg-rich grt in Fiordland eclogite and mafic granulite.

cheers
H
On Mar 28, 2018, at 5:49 PM, Wintsch, Robert P. <[log in to unmask]> wrote:

HI Howard,

wishing I could help, but I cant.
I used very well characterized olivines and kyanite and wollastonite as standards,
and that seemed to help.

but really I wanted to say that I am also FINALLY retiring from IU, and moving east to Conn.
Maybe we can cross paths in either Conn or Maine some time?

bob

Robert P. Wintsch
Dept. of Geological Sciences
Indiana University
1005 E. 10th Street
Bloomington, IN 47405
Tele. off: 812-855-4018; Tele. dept: 812-855-5582
Fax: 812-855-7899
[log in to unmask]
________________________________________
From: Metamorphic Studies Group <[log in to unmask]> on behalf of day <[log in to unmask]>
Sent: Wednesday, March 28, 2018 6:23 PM
To: [log in to unmask]
Subject: [geo-metamorphism] What is a good garnet analysis?

Colleagues,

Some thirty years ago, when we acquired a new microprobe, I spent a frustrating year attempting to produce good analyses of garnet.  Although we had some successes, we were unable to produce good analyses in a systematic way.  Every suite seemed to be a new problem.  Then, I became chair of department and that effort came to a halt.  Now, I am  back to the problem again in the context of eclogite petrology, and it occurs to me first to ask “What is a good garnet analysis?”

Because the quality of analyses in the literature is quite variable, I ask for your thoughts on the criteria by which we should judge the quality of routine garnet analyses as sufficient to be published. Obviously, the criteria may differ depending on the proposed application of the data, but is there a consensus on what constitutes a good routine analysis?

Thanks

Howard


Howard W. Day, Professor Emeritus
The Dept. Formerly Known as Geology
University of California Davis
One Shields Ave.
Davis CA 95616

Harold Stowell
Geological Sciences, University of Alabama
201 7th Ave Tuscaloosa
AL 35487-0338