Amphibole-saturated liquid

[zhudan]

I have a question concerning the amphibole-saturated liquid(basaltic system).
My calculation is the crystallization process. At the late stage of crystallization, the amphibole becomes a saturated phase, if the starting material containing water(e.g. 2 wt%). what I want to know is how much water does the melt contain then the amphibole becomes a saturated phase(could you list some references that MELTs based on? ), and the accuracy of the prediction.

[Paula]

Hi Dan,

The short answer is that there are no reactions involving amphibole included in the MELTS calibration.  See Table 3 in Ghiorso & Sack (1995) for a list of the phases and reactions that were used.  (Table 1 in Ghiorso et al., 2002, is the equivalent for pMELTS.)  Some of the experimental data used in the calibration database are for assemblages that include amphibole.  However, it is only the compositions of the coexisting liquid and certain anhydrous minerals that become constraints on the liquid thermodynamic model.

You could try looking at some of the experimental data with amphibole and comparing simple batch calculations (e.g. option 3 in adiabat_1ph) with the experimental results to see how realistic the MELTS calculations are for coexisting melt and amphibole.  Bear in mind the systematic offset in temperature in MELTS - it is probably better to compare the assemblage at similar melt fractions instead.

Naturally there are a lot of references in Ghiorso & Sack (1995) and I haven't gone through them exhaustively to see which ones have amphibole but you should be able to narrow down the search by going to Mark Ghiorso's MELTS website:

http://melts.ofm-research.org/database.html

It you select 'Advanced Search' you can look for experiments included in the calibration that have, say, more than 5 wt. % water in the liquid (you can, but you don't have to, enter any other oxide compositions).  Only phases from Table 3 in Ghiorso & Sack are listed on the website so you'll have to consult the original papers to see if amphibole was present.

A good example of a reference included in the calibration with amphibole (hornblende) in some of the experimental run products is Sisson & Grove (1993a; note that in Ghiorso & Sack and the database the year is given incorrectly as 1992).  Sisson & Grove do a specific study of hornblende stability and show that Na₂O content is important as well as H₂O.  For basalts and basaltic andesite liquids with H₂O <= 6 wt.% they conclude that 'amphibole will only appear as a phenocryst if the melts are also soda rich and low temperature'.

Whilst the conclusions of Sisson & Grove (1993) are probably a good rule of thumb, I don't know what the composition of the liquid in your calculations is like by the time amphibole is stabilised or the P, T conditions.  You might be able to use the calculated composition, P and T in the database search above to find something more suitable for your bulk composition and / or pressure of crystallization.

Hope that answers your question...

Paula

P.S. I moved this to the board about MELTS as that seemed the closest.