Influence of oxygen fugacity on the viscosity of CO2-, H2O-, and halogen-bearing basaltic melts
by Wiebke Pischel
Date of Examination:2018-12-10
Date of issue:2019-03-04
Advisor:Prof. Dr. Sharon Webb
Referee:Prof. Dr. Sharon Webb
Referee:Prof. Dr. Harald Behrens
Files in this item
Name:Doktorarbeit_Wiebke Pischel.pdf
Size:13.1Mb
Format:PDF
Name:Abstract_Wiebke Pischel.pdf
Size:247.Kb
Format:PDF
Name:Melt composition_Wiebke Pischel.pdf
Size:288.Kb
Format:PDF
Abstract
English
Abstract The aim of this study is to demonstrate the effect of chlorine (Cl-), fluorine (F-), water (H2O) and carbon dioxide (CO2) on the viscosity of basaltic melts and a comparison to literature models. The basaltic composition is similar to a melt of flat-topped volcano from Mid-Atlantic Ridge (MAR). The iron content was decreased, to avoid crystallisation of the glass. The study is divided into three series with different oxygen fugacity adjusted by diverse synthesis conditions. Series (I), (II) and (III) represents basaltic melts doped with Cl-, F- and (Cl- + F-), which were synthesised in a 1 atm furnace at 1473 K for 4 to 9 h in air. Afterwards, all series (II) glasses are redox equilibrated in a vertical gas mixing furnace at 1473 K for up to 20 h under a carbon dioxide and hydrogen gas atmosphere. Therefore, different oxygen fugacity can be generated and results in diverse Fe2+/Fetotal. The hydrous glasses with 0.5, 1.5 and 3.5 wt% H2O and CO2-bearing glasses with 1000 to 3500 ppm CO2 of Series (III) were synthesised in the internally heated pressure vessel (IHPV). The experiments were performed at 1550 K and 300 MPa for 24 h with a rapid quench device. In addition to H2O and CO2-bearing glasses, dry and CO2-free glasses are synthesised in the IHPV. The high viscosity is measured by the micropenetration technique in the range of 108.5 to 1013.5 Pa s. The Raman spectroscopy and colorimetric micro-determination of Fe2+ content can be used for structural analysis. The volatiles (F-, H2O and CO2) decrease the viscosity of basaltic melts, whereas the addition of F- shows the smallest effect on the viscosity and CO2 results in the strongest decrease in viscosity. The addition of 12.56 mol% H2O in halogen-free basalt results in a decrease in viscosity by 5.2 log units, whereas the presence of 0.34 mol% CO2 to a basaltic melt results in a strong decrease in viscosity by 0.7 log units. The addition of 7.73 mol% F- to basaltic glass results in a decrease in viscosity by 2.5 log units. Amounts with up to 2.53 mol% Cl- in basaltic melts result in an increase in viscosity, whereas further addition of chlorine results in a decrease in viscosity. The comparison of the experimental viscosity data of basaltic melts reflects that the current viscosity models do not consider all volatiles (Cl-, F- and CO2) and the models are calibrated for diverse amounts of volatiles (Giordano et al. 2008; Duan 2014; Sehlke and Whittington 2016). Especially, the variable effect of chlorine on the viscosity of peralkaline and peraluminous silicate melts should be noted in future models. Furthermore, the addition of volatiles to basaltic melts changes the Fe2+/Fetotal. In peralkaline melts, the addition of fluorine results in a oxidation of Fe2+ to Fe3+, which reflects an increase in polymerisation due to the increase in network formers. The addition of water results in a depolymerisation of the structure due to increasing Fe2+/Fetotal. The addition of OH groups in halogen-free and –bearing basaltic glasses shows a linear decrease in Fe2+/Fetotal, whereas the total water content shows an exponential trend. The OH groups describe the water species, which is incorporated into the melt structure. A linear relation between the addition of chlorine and iron speciation cannot be confirmed. Also, no dependence between CO2 content and Fe2+/Fetotal is observed. Kress and Carmichael (1991) developed a model for the determination of Fe2+/Fetotal by means of melting temperature, oxygen fugacity and melt composition. The model considers the amounts of Al2O3, FeOtotal, CaO, Na2O and K2O. The results of the iron determination show that the volatiles (F- and H2O) have a strong effect on the Fe2+/Fetotal. Thus, the volatiles have to be included by future models. The falling sphere technique was used to extend the measurable viscosity range of CO2-bearing basaltic melts to low viscosity ranges. This experiment was performed at 1223 K and 200 MPa for 5 h in a water cooled RQ-CSPV (rapid quench - cold seal pressure vessels). In this temperature range, the CO2-bearing basaltic glass crystallises and the actual experiments cannot be realised with the present basaltic composition.
Keywords: viscosity; oxygen fugacity; CO2; H2O; halogen; iron