Supporting Online Material for
To the origin of Icelandic rhyolites: insights from partially melted leucocratic xenoliths
Andrey A. Gurenkoa,b,c*, Ilya N. Bindemand, Ingvar A. Sigurdssone
a Centre de Recherches Pétrographiques et Géochimiques, UMR 7358, Université de Lorraine, 54501 Vandoeuvre-lès-Nancy, France
b Woods Hole Oceanographic Institution, Geology and Geophysics, Woods Hole, MA02543, USA
c Max-Planck-Institut für Chemie, Postfach 3060, 55020 Mainz, Germany
d Department of Geological Sciences, 1272 University of Oregon, Eugene, OR97403, USA
e South Iceland Nature Centre, Strandvegur 50, Vestmannaeyjar, IS 900, Iceland
This PDF file includes:
Text
Reference list
Tables A1 and B1 through B8
* Corresponding author and present address: Andrey A. Gurenko, Centre de Recherches Pétrographiques et Géochimiques, 15 rue Notre-Dame des Pauvres, BP 20, 54501 Vandoeuvre-lès-Nancy, France. Phone: +33 (0)3 83 59 48 75, Fax: +33 (0)3 83 51 17 98, E-mail:
1
Appendix A: Analytical methods
A1. Whole rock major and trace elements
Whole rock major and trace elements in the studied leucocratic xenoliths were analyzed by the Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP-AES) and flow injection Inductively Coupled Plasma Mass Spectrometry (ICP-MS) methods at the Analytical Service of Rocks and Minerals (SARM), Centre de Recherches Pétrographiques et Géochimiques (CRPG, Nancy, France); for more technique details see Carignan et al. (2001). The precision and accuracy of the method claimed by SARM are given in Table A1 but the accuracy wasadditionally monitored usingdifferent international reference materials (BHVO-2, Basalt, Hawaiian Volcanic Observatory; BCR-2, Basalt, Columbia River; AGV-2, Andesite; BIR-1a, Icelandic Basalt; GSP-2, Granodiorite, Silver Plume, Colorado; the United States Geological Survey), which were analyzed as unknown samples together with the samples of interest (Table A1). The analytical error declared by SARM is 15% relative for major elements, except for P2O5 (<10% relative), and 515% relative for trace elements, depending on their concentrations. This matches well the deviations from the certified values of the international reference materials analysedas unknowns (Table A1).
A2. Electron microprobe analysis
Major element compositionof minerals and glasses and S and Cl concentrations (glasses only) were determined using the JEOL Superprobe JXA-8200 electron microprobe at the Max Planck Institute for Chemistry (Mainz, Germany; hereafter referred as MPI-Mainz). We applied 15 kV accelerating voltage, 12 nA electron beam current and 1-2 µm size of the beam for analyses of clino- and orthopyroxenes, plagioclase, feldspar and Fe-Ti oxides but defocused it to 5-10 µm during analysis of interstitial glasses in order to minimize possible Na volatilization (Gurenko et al. 2005). The counted X-ray intensities of the elements were subject to a ZAF(“atomic numberabsorptionfluorescence”) matrix correction algorithm (Reed 2005 and references therein). Peak counting times on major elements were 60s and 30 s of background. Sulfur and chlorine were analyzed at the same analytical conditions as other major elements in the glass. A set of reference materials (natural and synthetic oxides, minerals and glasses; Micro-Analysis Consultants Ltd, Cambridgeshire, UK) and the Smithsonian Institution standard set for electron microprobe analysis (Jarosewich et al. 1980) were used for routine calibration and instrument stability monitoring. Typical analytical uncertainties (2RSD = 2 relative standard deviation) are 0.22.2% for SiO2, 0.84.6% for Al2O3, 1.812% for FeO, 1.03.2% for MgO, 0.83.2% for CaO, 1.07.8% for TiO2, 1.47.2% for Na2O, 2.020% for K2O, 1627% for MnO and 1140% for P2O5, depending strongly on their absolute concentrations, as inferred from replicate analyses of basaltic (USNM 111240/52 VG-2) and rhyolitic (USNM 72854 VG-568) reference glasses. As monitor samples to control precision and accuracy of S and Cl measurements, we used the VG-2 basaltic glass (0.1340.143 wt% S and 0.0290.032 wt% Cl; Dixon et al. 1991; Thordarson et al. 1996; Witter et al. 2005). The values obtained during this study are 0.144 ± 0.034 wt% S and 0.028 ± 0.016 wt% Cl (2SD, N = 53) and agree well with the reference values within the ±2SD uncertainty. Under the applied conditions, the detection limit of S and Cl was around 250350 g/g.
A3. Laser ablation ICP-MS
Interstitial glasses fromthe studied leucocratic xenoliths and their transporting hyaloclastites were analyzed for trace elements by laser ablation ICP-MS at Max Planck Institute for Chemistry (Mainz, Germany) using a New Wave UP-213 laser system (solid-state Nd:YAG laser with 213 nm wavelength operated at 10 Hz) combined with a single-collector sector-field ThermoFinnigan ICP mass spectrometer ELEMENT2 (for more detail see Jochum et al. 2011). Briefly, the ablation occurred in a He atmosphere (gas flow rate of ~0.8 l/min) that then was mixed with Ar (gas flow rate ~0.6 l/min) prior to the plasma torch. Spot analyses were done using a typical crater diameter of 65 µmat an energy density of about 15 J/cm2. Washout time between spots was 30 s, ablation time was 50 s and blank count rate was 16 s prior to ablation. The mass spectrometer was tuned to give maximum, stable signals at low oxide formation (ThO/Th <1%), no additional oxide correction was applied. Data reduction was done by calculating the blank-corrected count rates of the isotopes relative to the internal standard 43Ca. Instrument calibration was performed by ablating the NIST SRM 612 glass standard. The USGS and MPI-DING reference glasses (NIST SRM 612, KL2-G, ATHO-G; Jochum et al. 2006, 2011) were repeatedly analyzed throughout analytical sessionsand were used as reference materials to calculate relative sensitivity factors (RSF) of the target elements. External reproducibility of element concentrations measured within each of three reference glasses was always better than 10% relative. The obtained overall analytical error also was better than 10% relative for all elements except Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu and Hf, whose uncertainties were between 10 and 15% relative.
A4. Secondary Ion Mass Spectrometry
U-Pb zircon dating
Zircons were extracted from the crushed and sieved xenolith fragments by dissolving them in the HF. Approximately 3050 zircon grains from each sample were then hand-picked, mounted in epoxy resin, polished to ~7550% of their original size, investigated and imaged under petrography microscope using transmitted and reflected light, and then mapped using back-scattered electrons (BSE) and cathodo-luminiscence imaging (CL) with the JEOL Superprobe JXA-8200 electron microprobe at MPI-Mainz, Germany.
The U-Pb zircon dating (total 91zircons, 12 to 17 grains from each of 6 samples was performed using the NENIMF CAMECA IMS 1280 ion microprobe (Woods Hole Oceanographic Institution, USA) in October 2009. Then, in June 2012, the zircons were repeatedly dated using the CAMECA IMS 1270 instrument at the University of California Los Angeles (UCLA). The analytical conditions (see below) were similar for both laboratories. Also, a subset of the “youngest” zircons (9 grains, 23 from 4 of 6 samples) was additionally dated at UCLA in continuation of the June2012 session by U-Th geochronology method (the technique details are given in Schmitt et al. 2003, 2006).
During the U-Pb zircon analyses, ion intensities of94Zr2O (1 s), 204Pb (2 s), 206Pb (8 s), 207Pb (6 s), 208Pb (2 s), 232Th (1 s), 238U (3 s), and 238UO(1 s) peaks were measured in 10cycles, using a mass-filtered 16O-primary ion beam of ~20 nA focused to an oval 2530 um spot (counting times for the each mass are given in the brackets). To increase Pb+ yields (by about 2 times), oxygen flooding at O2 pressure of ~2e3 to 4e3 Pa of the sample surface was used during analyses. Secondary ions accelerated at 10 keV and passed through the energy slit centered and opened to 50 eV were analyzed in a peak jumping mode at mass resolution of ~4800 using an axial electron multiplier collector. During the analytical session (totally 230 individual spot analyses, including standards and unknown zircons), relative sensitivities of 238UO ions were calibrated by repeated measurementsof the concordant 91500 reference zircon (Wiedenbeck et al. 1995; 5 times at the beginning of the analytical session and then one time every 5 measurements of unknown grains). Also two, R33 and Plešovitse, reference zircons (Black et al. 2004; Sláma et al. 2008) were analyzed as unknown samples 3-5 times each at the beginning and the end of each sub-session (~5070 individual analyses) to additionally control the relative sensitivity factors obtained on the 91500 zircon standard.
Th-U zircon dating
The Th-U disequilibrium dating was performed, using basically the same instrument setup but using higher primary beam current (4080 nA). The ion intensities of 90Zr2O4 (0.5 s), 244.038 (3 s), 90Zr92ZrO4 (0.5 s), 246.028 (10 s), 246.3 (3 s), 232ThO (1 s) and 238UO (1 s) masses were measured in 30 cycles. Measured 238UO+/232ThO+ and radiogenic 206Pb/208Pb on the concordant reference zircon 91500 (Wiedenbeck et al. 1995) were used to determine Th/U relative sensitivity factors. The intensity of 230ThO+ was a subject for background correction by subtracting the averaged intensities measured on two mass stations at 244.038 and 246.3 amu. Mass 244.028 (232ThC+) was monitored as a proxy for 232Th2CO2+ isobaric interference resulting from possible beam overlap with epoxy resin, but none analysis was filtered, if exceeding the intensity of >5 cps. Uranium concentrations in unknown zircons were calculated using measured 238UO+/90Zr2O4+ ratios, by comparison with the reference 91500 zircon ([U] = 81.2 g/g; Wiedenbeck et al. 1995).
Oxygen isotope analyses
Oxygen isotope composition of zircon (112 individual measurements, including 42 replicate measurements, were done in 59grains, in which core, mantle and rim zones were analyzed where possible) was studied in CRPG (Nancy, France) during two analytical sessions, in March 2011 using the CAMECA IMS 1270 ion microprobe and in June 2012 using the CAMECA IMS 1280 HR. Multiple grains of the KIM-5 zircon standard (Valley 2003) mounted together with the unknown samples were sputtered with a 10 kV Cs+ primary beam of 810 nA current focused to 2025 μm spots. Pre-sputteringof samples during 60180 s was applied before each measurement. A liquid-N2 cold trap was used to ensure a vacuum pressure of <108 Torr in the sample chamber. The normal-incidence electron flood gun was used to compensate for sample charge. Secondary 16O and 18O ions were accelerated at 10 kV and analyzed at a mass resolving power of 2500 using a circular focusing mode and a transfer optic of 150 μm. A 400 μm contrast aperture and a 25003000 μm field aperture were used, giving a field of view of approximately 40 μm. The energy slit was centered and opened to 50 V. Automatic routine of secondary beam centering in the field aperture was used at the beginning of each isotopic measurement. The 18O/16O isotopic ratios were analyzed in multi-collection mode using two off-axis L’2 and H1 Faraday Cup (FC) detectors for counting simultaneously the 16O and 18O ion intensities, respectively. The gain of the Faraday cups was calibrated daily at the beginning of each analytical session using the CAMECA built-in amplifier calibration software, and the signal was then corrected for the FC backgrounds measured during pre-sputtering. The obtained ion intensities of ~3.5e+9 and ~0.71.1e+7 cps obtained on the 16O- and 18O-peaks, respectively, yield an internal 1 SE (standard error) uncertainty of better than ±0.1‰ that was reached after ca. 150 s (30 cycles of 5 sec of analysis time each).
Three to 5 measurements were run on the KIM-5 zircon standard at the beginning and at the end of each block of data acquisition that includes 2 to 5 unknown zircon grains, each of which was probed 25 times (10 to 25 point analyses in total), employing so-called “contiguous bracketing” technique. To correct raw data for instrumental mass fractionation (IMF), we used an average value of IMF derived on the standard at the beginning and the end of a given data block. If a systematic shift of IMF values during one or several data blocks was observed (usually it is 0.05 to 0.1‰ per hour), the unknown data were corrected for IMF calculated as a function of time. The external reproducibility obtained on the standard during multiple replicate measurements bracketing data blocks was nearly identical, suggesting that (i) all mounted standard grains are equally homogeneous and (ii) instrument stability was maintained. The uncertainty of an individual 18O measurement u(IM) (‰) was defined as:
u2(IM) = (signal)2/n + u2(IMF)/m+
+ u2(RM)(A1)
where signalis the relative standard deviation (RSD, usually ±0.20.5‰) of the18O/16O ratio over n cycles (n = 30), u(IMF) is the uncertainty of instrumental mass fractionationdefined by multiple (m), concurrent runs of the KIM-5 reference zircon, as stated above (1SD, usually ±0.10.5‰) and u(RM) is the uncertainty of the “true”18O (‰) value of the KIM-5 reference zircon used for calibration (±0.06‰, 1SD; Valley 2003). The resulting cumulative error for individual zircon measurement was always better than ±0.4‰ (2 SE), it is provided for each individual zircon in Table B8(Apendix B). The 42 zircon analyses (either representing core, rim or mantle parts of the crystals) were replicated within the uncertainty of 0.010.28‰, demonstrating rather limited O-isotope heterogeneity within the respective crystal zones. The 18O values are given in ‰ and defined relative to the Standard Mean Ocean Water (SMOW, 18O/16O = 0.0020052 0.00000043; Baertschi 1976) standard:
18O = ([18O/16O]sample – [18O/16O]SMOW)/
/ [18O/16O]SMOW 1000(A2)
Ti-in-zircon thermometry
An additional set of zircons was analyzed using the CRPG-Nancy CAMECA IMS 1280HR instrument (1 to 3 but mostly 2 spots in each of29 grains; none of them were characterized for O isotopes or dated). We applied a ~8 nA 16O beam focused to ~2025 µm spot. Titanium was analyzed at high mass resolving power (MRP of ~7000) to resolve 48TiH+ hydride interference on the 49Ti+ peak. The only remaining interference, 98Mo2+, can be quantitatively determined by measuring 98Mo+ in zircon, and was found to be negligible for all natural zircons examined so far.The Ti/Zr relative sensitivity factor was calibrated by 49Ti+/90Zr+ measurements of the 91500 reference zircon (5.3 ± 0.6 µg/g Ti, 1SD; Schmitt and Vazquez 2006). The 57Fe+/90Zr+ ratio also was monitored to control possible beam overlap on melt (glass), apatite and Fe-Ti oxide inclusionsthat can be detected by coherently elevated or loweredFe and Ti signals, as compared to those on pure zircons. We formally rejected those measurements, whose 57Fe+/90Zr+ ratios were exceeding 2SD deviation from the average 57Fe+/90Zr+ value calculated from all unknown zircons (57Fe+/90Zr+ = 1.32e4 ± 1.59e5, 2SD, N = 76).This value correspondswithin 2-sigma uncertainty to 57Fe+/90Zr+ obtained for the 91500 reference zircon (1.22e4 ± 1.68e5, 2SD, N = 19). By this rationale, we also excluded several high-57Fe+/90Zr+ analyses, where Ti concentrations could have been considered as “normal” and had no evidence for beam overlap with inclusions.
A5. Single-grain laser fluorination
Oxygen isotopic compositions of nodule-forming quartz and feldspar grains and variously colored interstitial glasses were obtained by laser fluorination in the Stable Isotope Laboratory (University of Oregon, USA; Bindeman 2008) during four separate sessions, in April and December 2009 and in April and May 2012.Single grains of minerals and glass chunks were analyzed using a home-built laser fluorination line equipped with a 35W New Wave CO2 IR laser and combined with aFinnigan MAT 253 large radius gas source mass spectrometer. The gas generated in the laser chamber was purified through a series of cryogenic traps held at the temperature of liquid N2 and then traces of fluorine excess were removed by a mercury diffusion pump. Oxygen was converted to CO2 gas, the yield was measured, and then O isotopic composition of CO2 gas was analyzed on the gas spectrometer. Based on the concurrent multiple runs of the primary Gore Mt. Garnet standard (UWG-2, 18O = 5.8‰; Valley et al. 1995) and onesecondary Gore Mt. reference garnet (UOG, 18O = 6.52‰), the precision of the method was maintained to be at ±0.050.13‰, 1SD, depending on the session and is given for each individual measurement in Table 7 of the printed article. Totally, 57 individual glass, feldspar and quartz grains were analyzed, and21 of 57 grains were replicated within the uncertainty of 0.0010.17‰, 1SD, demonstrating a limited inter-grain O-isotope heterogeneity.
A6. Oxygen isotope fractionation
The fractionation of oxygen isotopes among two different phases (X and Y; in our case between Fsp, Qz, Zrn and interstitial melt) can be calculated from the following basic equation (e.g. Faure 1986):
18OX18OY=XY ≈ 1000 ln(aXY) =
= AXY 106/T2(A3)
where A is an empirical constant usually defined experimentally (e.g.Chiba et al. 1989; Chacko et al. 2001), T is the absolute temperature in Kelvin andaXY is a fractionation factor between X and Y phases defined as
aXY = (18O/16O)X/(18O/16O)Y(A4)
or from aconsistent equationinferred from the increment method of cation-oxygen bond strength calculations, primarily developed by Schütze (1980) and further elaborated by Zheng (1991, 1993) and Zhao and Zheng (2003 and references therein):
1000 ln(aXY) = 1000 ln(X)
1000 ln(Y) =(AXAY) 106/T2 +
+ (BXBY)103/T + (CX
CY)(A5)
whereXand Yare thermodynamic oxygen isotope factors of X and Y phases, andAX, AY, BX, BY, CX and CY are constants and T is in Kelvin.
Appendix B: Chemical and O-isotope compositions, U-Pb zircon dating results
The whole rock chemical compositions (major and trace elements) of the studied leucocratic crustal xenoliths, their interstitial glasses and host, transporting hyaloclastites are listed in Tables B1 and B2. The compositions of xenolith-forming plagioclase and K-feldspar, clinopyroxene, orthopyroxene, magnetite and ilmenite are given in Tables B3 through B6. The results of U-Pb zircon dating and oxygen isotope results are listed in TablesB7and B8.
References
Baertschi P (1976) Absolute 18O content of Standard Mean Ocean Water. Earth Planet Sci Lett 31:341344
Black LP, Kamo SL, Allen CM, Davis DW, Aleinikoff JN, Valley JW, Mundil R, Campbell IH, Korsch RJ, Williams IS, Foudoulis C (2004) Improved 206Pb/238U microprobe geochronology by the monitoring of a trace-element-related matrix effect; SHRIMP, ID-TIMS, ELA-ICP-MS and oxygen isotope documentation for a series of zircon standards. Chem Geol 205:115140
Bindeman I (2008) Oxygen isotopes in mantle and crustal magmas as revealed by single crystal analysis. In: Putirka KD, Tepley III FJ (eds) Minerals, Inclusions and Volcanic Processes. Rev Mineral Geochem 69, Mineral Soc Am, WashingtonDC, pp 445478
Carignan J, HildP, Mevelle G, Morel J, YeghicheyanDE (2001) Routine analyses of trace element in geological samples using flow injection and low pressure on-line liquid chromatography coupled to ICP-MS: a study of geochemical reference materials BR, DR-N, UB-N, AN-G and GH. Geostandard Newslett 25;187198
Chiba H, Chacko T, Clayton RN, Goldsmith JR (1989) Oxygen isotope fractionations involving diopside, forsterite, magnetite, and calcite: application to geothermometry. Geochim Cosmochim Acta 53:29852995
Chacko T, Cole DR, Horita J (2001) Equilibrium oxygen, hydrogen and carbon isotope fractionation factors applicable to geologic systems. In: Valley JW, Cole DR (eds) Stable Isotope Geochemistry. Rev Mineral Geochem 43, Mineral Soc Am, WashingtonDC, pp 181
Dixon JE, Clague DA, Stolper EM (1991) Degassing history of water, sulfur and carbon in submarine lavas from Kilauea volcano, Hawaii. J Geol 99:371394
Faure G (1986) Principles of Isotope Geology. 2ndEdition, John Wiley and Sons, New York
Gurenko AA, Trumbull RB, Thomas R, Lindsay JM (2005) A melt inclusion record of volatiles, trace elements and Li-B isotope variations in a single magma system from the Plat Pays Volcanic Complex, Dominica, Lesser Antilles. J Petrol 46:24952526
Jarosewich EJ, Nelen JA, Norberg JA (1980) Reference samples for electron microprobe analyses. Geostand Newslett 4:4347
Jochum KP, Stoll B, Herwig K et al (2006) MPI-DING reference glasses for in situ microanalysis: New reference values for element concentrations and isotope ratios. Geochem Geophys Geosyst 7:Q02008, doi: 10.1029/2005GC001060