New experimental data for a re-calibration of the Fe-Ti oxide thermo-oxybarometers [Elektronische Ressource] / vorgelegt von Ursula Sauerzapf

English
247 Pages
Read an excerpt
Gain access to the library to view online
Learn more

Description

NewExperimentalDataforaRe CalibrationoftheFe TiOxideThermo Oxybarometers.INAUGURAL DISSERTATIONzurErlangungderDoktorwürdederNaturwissenschaftlich MathematischenGesamtfakultätderRuprecht Karls UniversitätHeidelbergvorgelegtvonDiplom MineraloginUrsulaSauerzapfausHeidelbergNewExperimentalDataforaRe CalibrationoftheFe TiOxideThermo Oxybarometers.Gutachter:Prof. Dr. DominiqueLattardHochschuldozentinDr. AgnesKontnyTagdermündlichenPrüfung: 29.05.2006ContentsAbstract vKurzfassung ix1 Introduction 11.1 AshortpresentationofFe Tioxidemineralsandtheirsignificance . 11.2 Stateoftheart: crystallinephasesinthesystemFe Ti O . . . . . . 51.2.1 A general introduction to relevant Fe Ti oxide phases:chemistry,structureandstability . . . . . . . . . . . . . . . 51.2.2 Wüstite . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91.2.3 Titanomagnetite . . . . . . . . . . . . . . . . . . . . . . . 91.2.4 Titanomaghemite . . . . . . . . . . . . . . . . . . . . . . . 181.2.5 Ilmenite . . . . . . . . . . . . . . . . . . . . . . . . . . . 20ss1.2.6 Pseudobrookite . . . . . . . . . . . . . . . . . . . . . . . 21ss1.3 Thermo oxybarometry . . . . . . . . . . . . . . . . . . . . . . . . 231.4 Aimsofthepresentstudy . . . . . . . . . . . . . . . . . . . . . . . 342 Experimentalandanalyticalmethods 352.1 Generalapproach . . . . . . . . . . . . . . . . . . . . . . . . . . . 352.2 High temperatureexperiments . . . . . . . . . . . . . . . . . . . . 362.2.

Subjects

Informations

Published by
Published 01 January 2006
Reads 26
Language English
Document size 90 MB
Report a problem

NewExperimentalDatafora
Re CalibrationoftheFe TiOxide
Thermo Oxybarometers.
INAUGURAL DISSERTATION
zur
ErlangungderDoktorwürde
der
Naturwissenschaftlich MathematischenGesamtfakultät
derRuprecht Karls Universität
Heidelberg
vorgelegtvon
Diplom MineraloginUrsulaSauerzapf
ausHeidelbergNewExperimentalDataforaRe Calibrationof
theFe TiOxideThermo Oxybarometers.
Gutachter:
Prof. Dr. DominiqueLattard
HochschuldozentinDr. AgnesKontny
TagdermündlichenPrüfung: 29.05.2006Contents
Abstract v
Kurzfassung ix
1 Introduction 1
1.1 AshortpresentationofFe Tioxidemineralsandtheirsignificance . 1
1.2 Stateoftheart: crystallinephasesinthesystemFe Ti O . . . . . . 5
1.2.1 A general introduction to relevant Fe Ti oxide phases:
chemistry,structureandstability . . . . . . . . . . . . . . . 5
1.2.2 Wüstite . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2.3 Titanomagnetite . . . . . . . . . . . . . . . . . . . . . . . 9
1.2.4 Titanomaghemite . . . . . . . . . . . . . . . . . . . . . . . 18
1.2.5 Ilmenite . . . . . . . . . . . . . . . . . . . . . . . . . . . 20ss
1.2.6 Pseudobrookite . . . . . . . . . . . . . . . . . . . . . . . 21ss
1.3 Thermo oxybarometry . . . . . . . . . . . . . . . . . . . . . . . . 23
1.4 Aimsofthepresentstudy . . . . . . . . . . . . . . . . . . . . . . . 34
2 Experimentalandanalyticalmethods 35
2.1 Generalapproach . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.2 High temperatureexperiments . . . . . . . . . . . . . . . . . . . . 36
2.2.1 Preparationofstartingmixtures . . . . . . . . . . . . . . . 36
2.2.2 High temperaturetechniques . . . . . . . . . . . . . . . . . 38
2.2.3 Synthesisexperiments . . . . . . . . . . . . . . . . . . . . 39
2.2.4 Re equilibrationexperiments . . . . . . . . . . . . . . . . . 41
2.2.5 Annealingexperiments . . . . . . . . . . . . . . . . . . . . 42
2.2.6 Quenchingprocedures . . . . . . . . . . . . . . . . . . . . 42
2.3 Characterisationandanalysisoftherunproducts . . . . . . . . . . 43
2.3.1 Opticalexamination(eye,binocular) . . . . . . . . . . . . . 43
2.3.2 X raypowderdiffraction . . . . . . . . . . . . . . . . . . . 43
2.3.3 Scanningelectronmicroscopy . . . . . . . . . . . . . . . . 44
2.3.4 Electronmicroprobe(EMP)analysis . . . . . . . . . . . . . 45
2.3.5energy lossspectroscopy . . . . . . . . . . . . . . 47
i3 Textures of synthetic sub solidus Fe Ti oxide assemblages in the sys
temFe Ti±Al±Mg O 49
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.2 Generalremarksonexperimentalconditionsandprocedures . . . . 50
3.3descriptionofcommontexturalfeatures . . . . . . . . . . . 51
3.3.1 Crystalshapes,sizesanddistribution . . . . . . . . . . . . 51
3.3.2 Porosityandcracks . . . . . . . . . . . . . . . . . . . . . . 51
3.4 Formationofconcentrictextures-correlationtothestartingmaterial 54
3.4.1 Generalremarks . . . . . . . . . . . . . . . . . . . . . . . 54
3.4.2 ExperimentswithFe O +TiO startingmixturesfrompre 2 3 2
annealedreagents . . . . . . . . . . . . . . . . . . . . . . . 55
3.4.3withFe O +TiO startingmixtures . . . . . . 562 3 2
◦3.4.4 ExperimentswithFe +TiO starting . . . . . . . 642
◦3.4.5withFe O +Fe +TiO startingmixtures . . 662 3 2
3.4.6 Time dependentexperiments . . . . . . . . . . . . . . . . . 67
3.4.7 Interpretationandconclusion . . . . . . . . . . . . . . . . . 68
3.5 Oxy exsolutiontextures . . . . . . . . . . . . . . . . . . . . . . . . 71
3.5.1 SEMobservationsandEMPanalysis . . . . . . . . . . . . 71
3.5.2 Interpretationandconclusion . . . . . . . . . . . . . . . . . 73
3.6 Chemicalinhomogeneitiesofphasesinproductsofsynthesisruns . 75
3.7 Summaryandconclusion . . . . . . . . . . . . . . . . . . . . . . . 77
4 PhaserelationsandphasecompositionsinthesystemFe Ti O 81
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.2 Generalexperimentalapproach . . . . . . . . . . . . . . . . . . . . 82
4.3 Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.3.1 Indicationsfortheattainmentofequilibrium . . . . . . . . 82
4.3.2 Re equilibrationexperiments . . . . . . . . . . . . . . . . . 84
4.3.3 Phaseboundariesandcompositions . . . . . . . . . . . . . 90
4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
4.4.1 TheTmt Ilm assemblage: phasecompositionsandbound ss
aries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
4.4.2 TheTmt Ilm Comparisontothermodynamicss
models . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
4.4.3 TheIlm Psb assemblage . . . . . . . . . . . . . . . . . 100ss ss
4.4.4 Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . 1025 High Tnon stoichiometryinmagnetite ulvöspinelsolidsolution 103
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
5.2 Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5.3 Theannealingmethod. . . . . . . . . . . . . . . . . . . . . . . . . 105
5.3.1 Principlesoftheannealingmethod . . . . . . . . . . . . . . 105
5.3.2 Uncertaintiesofthemethods . . . . . . . . . . . . . . . . . 112
5.3.3 Actualandallegedpitfalls . . . . . . . . . . . . . . . . . . 114
5.3.4 Presentationanddiscussionoftheresults . . . . . . . . . . 120
5.4 Non stoichiometryfromcelldimensions . . . . . . . . . . . . . . . 126
5.4.1 Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5.4.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5.4.3 Summaryandinterpretation . . . . . . . . . . . . . . . . . 129
5.5 Non stoichiometryfromEELS . . . . . . . . . . . . . . . . . . . . 132
5.5.1 Principlesandmeasurement . . . . . . . . . . . . . . . . . 132
5.5.2 Sampleselectionandpreparation . . . . . . . . . . . . . . 133
5.5.3 Pitfalls . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
5.5.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
5.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
5.6.1 Summary of annealing method results, lattice parameter
dataandEELSdatafromthepresentstudy . . . . . . . . . 138
5.6.2 Comparisonwithliterature . . . . . . . . . . . . . . . . . . 139
5.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
6 InfluenceofminorelementcontentsofMgandAl 143
6.1 Introduction: minorelementcontentsinFe Tioxides . . . . . . . . 143
6.2 Experimentalandanalyticalapproach . . . . . . . . . . . . . . . . 145
6.3 Phaserelations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
6.4 PhasecompositionsinthesystemFe Ti Mg Al O . . . . . . . . . . 148
6.4.1 Mg andAl contentsoftheFe Tioxidephases . . . . . . . 148
6.4.2 Projection of Mg Al bearing compositions into the Ilm
HemandMag Uspbinaries . . . . . . . . . . . . . . . . . 153
6.4.3 Comparisonoftheprojectionschemes . . . . . . . . . . . . 156
6.4.4 of X’ and X’ derived with equations 6.5Usp Ilm
and 6.6, with the data from the simple system Fe Ti O
andwiththeexistingthermodynamicthermo oxybarometer
models . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.5 Non stoichiometry in Mg Al bearing Tmt - results from annealing
experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
6.6 Discussionandconclusion . . . . . . . . . . . . . . . . . . . . . . 1667 Outlook 171
Bibliography 175
A Appendix 187
A.1 Tableofabbreviations . . . . . . . . . . . . . . . . . . . . . . . . . 187
A.2 Startingmixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
A.3 Tablesofexperimentalconditionsandanalysisresults . . . . . . . . 192
A.4 Conversionofwt%andvol%intomol% . . . . . . . . . . . . . . . 222
A.5 Tmtprojection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
B Author’spublicationsrelatedtothisthesis 227
B.1 Paper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
B.2 Abstractsandconferenceparticipations . . . . . . . . . . . . . . . 227
Acknowledgements 229
Erklärung 231Abstract
Despite their modest modal abundances, the Fe Ti oxide minerals are important
accessory minerals in igneous rocks. They are major carriers of rock magnetism
andpetrologicindicatorsoftemperatureandredoxconditionsduringthemagmatic
stage(Buddington&Lindsley,1964).
The titanomagnetite ilmenite (Tmt Ilm ) thermo oxybarometer has beenss ss
widely used to retrieve information on redox states in the Earth’s mantle, pro
cesses in magma chambers, the crystallisation conditions of lunar or martian
basalts,redoxconditionsduringmetamorphicevolution,orfluid rockinteractions.
However, the current thermodynamic formulations of the Tmt Ilm thermo ss
oxybarometer (Andersen & Lindsley, 1988; Ghiorso & Sack, 1991) fail to repro
duce experimental results, in particular at the high temperatures that are relevant
for basaltic assemblages. This is shown in chapter 1 (section 1.3) by applying the
formulationstoexperimentalliteraturedata. Theshortcomingsoftheexistingmod
els are partly due to the limited T-fO range of experiments used for the original2
calibrations. Further, theeffectofsmallcontentsofMg, Al, etc. typicalfornatural
Tmt and Ilm was only partly evaluated, and the non stoichiometry of the phasesss
(inparticularTmt)wasnotconsideredatall.
To support a revised version of the Tmt Ilm thermo oxybarometer we havess
synthesised Tmt+Ilm assemblages in the systems Fe Ti O, Fe Ti Mg O, Fe Ti ss
◦Al O and Fe Ti Mg Al O at 1 bar, 1000 −1300 C and under a large range of fO2
(ΔNNO 5 to +5, fO controlled by gas mixtures or solid state buffers, see chapter2
2 for experimental details). A large data set has also been provided for Ilm co ss
existing with pseudobrookite (Psb ). This forms the basis for an experimentalss ss
calibrationofathermodynamicIlm Psb (thermo )oxybarometermodel. Thelat ss ss
ter can be used to estimate oxygen fugacity in oxidised terrestrial igneous rocks,
moderately reduced assemblages in mantle xenoliths, or strongly reduced lunar as
semblages.
Thetexturesofthesyntheticsampleshavebeenusedtoevaluatewhetherequilib
riumcouldbeattainedoratleastcloselyapproachedintherunproducts. Indeed,the
synthetic run products essentially consist of polycrystalline, equi granular aggre
vgates, with idiomorphic, homogenous and evenly distributed grains (10−50μm).
Thissuggeststhattherunproductsreflectorareclosetoanequilibriumstate.
However,inmanysampleswehaveobservedconcentricvariationsinTi/(Ti+Fe).
Theirformationisstronglyrelatedtothestartingmaterialandisprobablydrivenby
the redox processes that occur during the experiment (chapter 3). Synthesis prod
ucts prepared from Fe O +TiO mixtures usually show an increase of Ti/(Ti+Fe)2 3 2
towardsthesamplepelletsurface. Asaresult,single phasesamplesdevelopacon
centric phase inhomogeneity, while binary assemblages show variations of phase
abundances, typically the formation of a Ti rich monomineralic rim. In contrast,
◦synthesisproductspreparedfromFe +TiO mixtureshaveFe richrimsatthesam 2
plepelletsurface. Nevertheless,localequilibriumisachievedineachofthesample
regions.
At the outer max. 100μm of the sample pellets, samples that have been exposed
to oxidising conditions during the quench process display oxy exsolution features
inthesenseofBuddington&Lindsley(1964),whichmayformwithin<5seconds.
Oxy exsolutiondoesnotaffectthephasecompositionsinthecentralsampleregion
butinvolvesanincreaseofthesamplebulkoxygencontent.
For the re calibration of the Fe Ti thermo oxybarometer, the data base on Tmt
coexisting with Ilm used to calibrate the models of Andersen & Lindsley (1988)ss
and Ghiorso & Sack (1991) has now been extended towards higher temperatures
◦(i.e. 1000−1300 C) and to a larger range of oxygen fugacities. The compositions
of coexisting Fe Ti oxide phases in the system Fe Ti O have been determined by
EMPasafunctionofTandfO (chapter4).2
Theexperimentaldataderivedwithinthisstudyhasbeencomparedtotheexisting
thermo oxybarometermodelsbyAndersen&Lindsley(1988,QUILF)andGhiorso
& Sack (1991). The transition within the ilmenite hematite solid solution series
¯ ¯(fromthelong rangeorderedR 3tothelong rangedisorderedR 3cstructure)causes
an abrupt change of the slope of the experimentally determined isotherms, which
gets increasingly pronounced with decreasing temperature. Andersen & Lindsley
(1988) did not account for this transition, and consequently their model isotherms
donotshowtherelatedsharpbend. Ghiorso&Sack(1991)incontrastdidaccount
fortheIlm Hemtransition,howevertheirmodelheavilyreliesonoldexperimental
data involving sample compositions that were not controlled by microanalytical
methods.
SyntheticTmtcompositionsaregenerallyTi richerthanpredictedfromthemod
els. At high fO , this is probably related to errors in the standard state properties2
usedfortheTmtendmembers. Atlow fO andhighT,thestrongdiscrepanciesbe 2