Chemistry Reference
In-Depth Information
It is more simple to assume that iron is mainly in the ferrous state and
to calculate the corresponding number of oxygen for the cations other than
K, Na, Ca (that is 21 oxygens in this part of the unit cell), regardless of OH
ions that will be latter independently calculated.
Example 4
: A hornblende A
0-1
X
2
Y
5
Z
8
O
22
(OH)
2
In amphiboles, the A site (occupied by Na and K) is not necessarily
complete, and on the other hand, Na may also occupy the X site (Ca). It is
better to ignore these elements and calculate the structural formula on the Y
and Z sites of the unite cell, that is on 13 cations. The ratio ferrous - ferric
iron is then recalculated from the corresponding number of oxygen (21) and
OH fixing OH
4.
If there is sodium in site X, the calculation may be erroneous. So consult
Leake (1978) for calculating the structural formulas of amphiboles in func-
tion of the type of amphibole.
When we the various parameters (H
2
O, FeO, Fe
2
O
3
) have been calcu-
lated, it is possible to recalculate the analysis in weight of the mineral, to
make the total and compare the recalculated analysis the raw results of the
microprobe.
The criteria of quality of analysis are:
+
F
+
Cl
=
1
Match with the structural formula.
2
To be compatible with other analysis of the same mineral in the same
sample or in related samples; indeed there are regular laws of variation
of the chemical composition of minerals that reflect the geological proc-
esses; the new analysis should integrate the set of analyses that describe
this phenomenon. If not, it is worth reflecting on why it does not fit.
3
The total of the analysis should be 100 wt% after recalculation; this is
perhaps the least important criterion; some water-rich mineral (such as
chlorites), or minerals with a light matrix (cordierite with Si, Al, Mg
and possibly zeolitic water) or minerals with a sharp contrast in weight
between the various elements (garnet with heavy elements like Fe, Mn,
and light ones, Si and Mg), give an apparently poor result; some toler-
ance of the total may be accepted, as long as this tolerance is of the same
order for all the minerals that we want to compare.
There is not single way to calculate a structural formula. We have to
adapt this calculation to the minerals - and the problems - that are studied.
However, it must be clearly indicated how the calculation was done.
