eskesore
For some silicates there is a relation between site occupancies and lattice parameters. Entering relations in Topas would decrease the degrees of freedom for quantitative Rietveld of complex mixtures.
Does anyone know of a macro which retrieves the avarage atomic radii for a given site?
alancoelho
Hi eskesore
If you knew the sites then you could use the sites_geometry keyword which follows the z-matrix format; ie.
sites_geometry c4
load site_to_restrain { C4 C3 C2 C1 }
prm = Sites_Geometry_Distance(c4); : 0
prm = Sites_Geometry_Angle(c4); : 0
prm = Sites_Geometry_Dihedral_Angle(c4); : 0
The Sites_Geometry_Distance function will then return the distance between C4 and C3; Sites_Geometry_Angle will return the angle between C4, C3 and C2 and Sites_Geometry_Dihedral_Angle is self explanatory. Best to view the test_examples of SITES_GEOMETRY_1.INP and SITES_GEOMETRY_2.INP.
If you did not know the sites in question then you could use the atomic_interaction keyword:
atomic_interaction avg
= R / 6 ;
ai_radius 3 ' optional
ai_closest_N 6
ai_sites_1 Ti1
ai_sites_2 O*
The use of the parameter 'avg' will then return the average bond distance between the Ti1 site and any O* sites. Look up atomic_interaction in the technical_reference.pdf for details.
cheers
alan
eskesore
Thank you for the detailed answer, i am aiming a bit different
What i want to get is the radii as stated in the tables with atomic radii
In some old papers they found relations between lattice parameters and site occupancies for forexample olivine
Paper is:
Lumpkin and Ribbe (1983) American mineralogist vol 68, page 164-176
I am trying to apply this to the modelling as it would reduce the degrees of freedom.
Also i would like to output the chemical formula sum of the minerals to compare directly with chemical assays
mfisch
is it possible that you're talking about average bond length between e.g. Fe and O and Mg and O instead of radii? Decreasing polyhedral bond length may influence lattice parameters, however, a clear correlation between single bond length and lattice aprameters may be difficult to obtain. I
guess, it would be easier to correlate lattice parameters directly with site occupancies of e.g. the Fe/Mg sites, which is rather easy to do in Topas.
For outputting a chemical sum, wirte a macro. This will work for the one phase only.
Regards
Martin
eskesore
Thanks again, i guess youre right. What i want to do is correlate with cell parameters with site occupancies and off course write macros to output the chemistry of the minerals in batch scripts, ....since i am running phase quantifications on large sample batches ( will run 81 samples to make mineralogical map of mafic intrusion), and hence i want to model chemical variation in the mafic silicate minerals as well.
:0)
mfisch
give it a go:
name the lattice parameters, e.g. lp_a, lp_b and so on. from the formula in the paper, create a new parameter including the formula, which may be sthing like this:
prm occupancy_Fe = lp_a * 3.123(=some value from the paper);:0
Occupancy Mg = 1 - occupancy_Fe. Use the new occupancy_Fe as value in site occupancy of Fe.
Output for olivine is then (Mg*(1-occupancy_Fe),Fe*occupancy_Fe)2SiO4
Be aware that Fe and Mg in olivine will not have the same displacement parameters (and in a phase mixture, you probably do not have enough data quality to refine/obtain this value correctly).
Thus, your quantification will not be based on the same olivine structure throughout the batch (leading to the fact that even relative quantification may be inaccurate).
It may be better to quantify the mixture using only one olivine structure (e.g. close to an average Mg/Fe value (XRF?)) and determine Mg/Fe ratio from XRD in a second step.
eskesore
You are correct that the thermal displacement parameters cant be modelled modelled in a mixture. I usually keep them fixed, but i might have a second look at it if they influence quantification results. They are probably known for the end-members and some of the intermediate solid solutions....
The olivines are the most prossing mineral for this since they are simpler than e.g. Pyroxenes, which have significant tetrahedral substitutions as well. The idea is to have a single str. That covers that common olivine in mafic rocks and this would be usefull in mapping modal mineralogy because the mafic minerals typically varies in composition due to crystallisation processes in the magma chamber.
mfisch
Let us know if it works...