Rit (Vísindafélag Íslendinga) - 01.06.1984, Blaðsíða 187

mineralchkmistry and relationships
183
reported, but primarily using iron-titanium oxides of' low minor element
content. The effects of higher minor element content on the results obtained
have been little studied. Most authors, including Buddington & Lindsley
(op.cit.), Carmichael (1967a & b) and Anderson (1968), combine the minor
elements on the basis of ilmenite and spinel structure and discard them. If
some minor elements are left after this procedure, Fe and Ti amounting to
the remainder are discarded before determination of temperature and
oxygen fugacity. If the minor element contents are low this has a very small
effect on the result. If the minor element contents are, on the other hand,
high, as is the case with the Jan Mayen oxides, the results differ consider-
ably from those obtained by incorporating the minor elements in the
calculations (see Table 30). Experimental investigations bearing on the
effects of minor elements on the oxide thermometer-baromcter of Budding-
ton & Lindsley (op.cit.) are few and give an incomplete picture of the total
effect of the combined minor elements. Mazzullo et al. (1975) report on the
effect of'Mn at 1 kbar, 700—850°C and fö2 of the NNO- and FMQ buffers.
Fhe effects are apparently rather small. Speidel (1970) studies the effects of
Mg on the iron-titanium oxides and says that Mg strongly prefers the spinel
structure over that of the ilmenite one at high f02, but at fÓ2= 10~7 to 10“9
this preference disappears and the distribution coeflicient approaches 1. He
further says that the f02 predicted from Fe—Ti contents of coexisting oxides
are an order of magnitude too low if Mg is present in the oxides, but the
temperature effect is difficult to evaluate. Pinckney & Lindsley (1967), on
the other hand, claim that Mg prefers the ilmenite structure and
increasingly so with falling temperature, and that the effect of Mg on the
oxide thermometer is similar to that of Mn. Because of the incomplete
knowledge of the minor element effect on the oxide thermometer, the author
has no solid grounds for choosing between methods of calculation and
follows therefore the tradition ofusing only the amounts of Fe and Ti which
amount to pure ilmenite-hematite and ulvospinel-magnetite solid solutions
in these minerals.
Titanomagnetite phenocrysts (rather small) without contemporaneous
dmenite are common in all rock types of the Jan Mayen rock suite, but
relatively early crystallized contemporaneous ilmenite and titanomagnetite
occur only as exceptions. This is in agreement with the statement of
Anderson (1968) that alkaline basalts with a relatively high TiO/(FeO-l-
Fe203) ratio and a low silica activity generally crystallize titanomagnetite at
a relatively early stage in contrast to subalkaline basalts, where both
ilmenite and titanomagnetite commonly crystallize at a relatively early
stage. Contemporaneous ilmenite and titanomagnetite crystals occur in the
groundmass of some of the basic lavas of Jan Mayen. From their habit and
size relationships, these grains are judged to have crystallized shortly before
the final quenching of the groundmass itself. These mineral pairs have been
analysed in one ankaramite sample (Jan 166), one ankaramitic basalt (Jan