6th International Symposium on Andean Geodynamics (ISAG 2005, Barcelona), Extended Abstracts: 795-798

Evolution of , N. : Evidence from major and trace elements, Sr-, Nd-, Pb-isotopes, age dating and chemical zoning in sanidine megacrysts

W. Wegner, G. Worner, & A. Kronz

Abt. Geochemie, Universttat Gottingen, Goldschmidtstr. 1, 37077 Gôttingen, Germany; [email protected]

INTRODUCTION The Taapaca volcano in N. Chile (l8°06'S, 69 °30'W) forms a dacitic dome c1uster with rare flows (Fig. a-c). The is formed by folded Tertiary volcanic and volcaniclastic deposits and the 2,72 Ma . The voJcano is associated with an apron (W and N) and valiey-fiJling (SW) black-and-ash flows resulting From Frequent dome collapse. Ar -Ar ages for Taapaca rock s range From 1,27 Ma at the base to 32 ka for one of the youngest domes. Late Quaternary «20 ka) activity is indicated by isolated deposits block-and-ash flows in a distal transverse valley, suggesting descent From the summit over filled valleys.

' ",,) 0 r-----:,....--,....--,....-----,....-,....-- - -, - Faapaca anë esnes 10 ëacues ...... matie inclusions

~ 100.11 12 e- ~ '2 ~10 ~ H1. () o oï N c:J ~ ~ 8 o 1.0 o :;; N r ?: Z 6 •...... " ...... " t~ . .. . -.. . 0. 1 4 ......

tJ.OI 1 - -- . - RbPb l" ThE\a K r..a (' r!'\"bTaPr Sr~~dllr7rSmruOdTbO)'HoYbb \' TmTi Lu ~ ~ ~ M M W ~ W

Si0 2 wt% Fig. 1: Alkali-Silica - diagram showing the small Fig. 2: Trace element distribution diagram with a range for Taapaca compositions (main cone) com­ weak Nb-Ta trough and strong enrichments in Ba. pared to other stratovoJcanoes and El inclusions have less enriched patterns . N=733 Jargely unpublished data

Magma compositions are surpris ingly uniform throughout the volcano's history (Fig . l , 2), except one and abundant mafic inclusions. Thi s argues for a steady-state thermally buffered sy stem at 15-20km depth (AI-in-hbl barometry, see below). Sanidine megacrysts (up to 15 cm ) are Frequent in sorne domes, their included indicate shallower depths (5- 12km) of formation compared to the host.

PROCESS AND PRESSURE OF MAGMA EVOLUTION We used the Al-in- barometer (Schmidt (1992) and we are weil aware of the limitations and the large errors involved. Nevertheless, we found significant differences in composition between inclusions in the sanidine megacrysts and those of the host magma (Fig. 3) . Using 0-, Sr-, Nd-, and Pb­ isotope data, we modeled Recharge+Assimilation+Fractionation+ Tapping (Aitcheson & FOITest, 1994; J Pet 35:461-488) for a Taapaca dacite using a primitive island arc basaIt (Nye and Reid , 1986, JGR 91 BIO: 10271-

795 6th International Symposium on Andean Geodynamics (ISAG 2005, Barcelona), Extended Abstracts: 795-798

10287) and average nearby Belen basement (Wôrner et al., 2000, JSAES 13:717-737). The model is solved to satisfy all dacite isotope compositions for a RAFr process based on an input and assimilant co mpos itions. The solution then allows comparison between observed and calculated trace element patterns. The successful solution predicts c.18 % crustal assimilation ( =0 .18) at an assimilation/fractionation rate of r=O.15. The recharge rate is rather low compared to the assimilation rate ( =0,6). This is consistent with a restricted range of compositions and a long storage and evolution in a the rmally buffered magma system. Note, however, that the residence times of the crystals are relatively short (see below ).

core-. rim 12 34567 lIInI1

nI!•• ..l.. • rnnI.11. 1.1 ..l....l.. 1.

123 4 5 6 789

0, 25 ~,------:,...---,.------_ Fig. 4: Results of the RAFT model showing the "'E œ 0.20 " convergence of "'E ..- - ;c-'=';":-- ';:; parameters for a particular Vl 2 0. 15 u set of parameters. Q.~ According to this a Dacite TAP·DD2 o 0 .1 0 solution, the amount of '0 - - - "Sr/ " Sr 0,7065 ~ - E-Nd -5.7 crustal assimilation is Vl Vl - ,ooPb/,o'Pb 18.099 E O.os ,- ,o7Pb/'04Pb 15,61 9 about 18% and the - 208Pb/,o'Pb 38 399 relative rate of - ô18ü 7.4 %0 assimilation to crysta lli­ o.os o 10 r 0 .15 0 .20 0. 25 sation is 0.16. rate of (assrmitaticn/fractio na! cry st allizatio n) f3 rate of (recbarge/ asairrétatlcn )

--- observed - init ial arc basait - crustal contaminant lD l rromN( e& Reld ( 1986 ) -+-- calculated range of crustal compositions observed (womer " al.• 2001) 1000

Fig. 5: The RAFr model (Fig. 4) gave good solutions in the trace element modeling for the following fractionating assemblages: 8%, Opx 10%, Cpx 18 %, 30 %.

Ba Rb Th Ta Nb La Ce Sr Nd Sm Zr Eu Hf Tb YYb

796 The fractionating assemblage included amphibole and gamet but little to satisfy the strong LREE/HREE fractionation in the RAFT-process. The host dacite thus evolved during deep (>40 km) crustal assimilation (Fig. 5). Compositional and isotope data thus suggest 18% of crustal melt in the formation of the (Fig. 4). Subsequent evolution took place at shallower levels (15-20 km), based on AI in Hbl barometry of host dacite amphibole phenocrysts. The sanidine megacrysts formed at still shallower depths and were taken up as xenocrysts during ascent of the dacite.

ZONATION IN SANIDINE MEGACRYSTS 20 sanidines were mapped and profiled by microprobe analysis for zonations in Ba, Ca, K, and other trace elements. Both, small and large crystals show large Ba variations with Ba-poor cores, saw-tooth-type oscillations (0.2-3.1%BaO) and high-frequency alternating zonations in the outer rim (Fig. 6,7).

,"""'"'""'" • """"", ~~ lJ ld '«l

! vJ rr ~lt • Of lt-,iin ~ l ~ l''''

_ -'::" _ ...10' _ &'1 '0

Fig. 6: Largest megacryst mapped with Fig. 7: Correlations of growth zones using slight 87Sr/86Sr ratios in individual growth zones. manual stretching and compression to improve fit.

The sharpest compositional contrasts (Fig. 8) were selected from core to rim for high-resoJution chemical profiling using quantitative microprobe measurements and X-ray profiling. These compositional steps were then modeled for diffusional smoothing to derive minimum residence times for the crystals.

797 CD TAP-02-28-07 o TAP- 02-28-14 A) COMPO -v:; .. l , l . ~' fI'" ' .. , .: . .' J • -.J . ' 1

Fig . 8: Two examples of mea sured step-profiles. The top two images are "COMPO"_mode images of the zoned crystals. Grayscales represent Ba contents (Jight: high Ba, dark. Iow Ba). The second two samples are colored (see our poster !!) images of Ba zonation. The profiles below are based on fully quantitative electron microprobe measurements. These profiles show a "saw-tooth" zon ation pattern with steep increases of Ba at resorption interfaces . (.~ ,''1 ) J ~ . The lower profiles are X-ray - scans at high spatial (',QI' ~ 1o:fr4 R .... resoJution across steep Ba compositions steps, These steps Dl S1 &p.profil were modeled for diffusional smoothing to obtain crystal residence times.

", .. ..- [loo.·...... I,,1"It

.i

10000 Fig . 9: Residence times derived from several compositional steps in each crystal (calculations after ~ 1000 Morgan et al, 2004, EPSL 222: 933-946). Position is represented as calcuJated volume % from -----.. core. Note the generally negative --- correlation suggesting as expected - shorter residence times for yo unger growth zones. la L-- . -"

a la 20 30 40 50 60 70 60 90 100 volume Irom core in %

CONCLUSIONS 1. Isotopie and trace element compositions indi cate a RAFT process at high pressures (gamet involved, > 40 km) . AI-Hbl barometry indicates further storage and cry stallization at shallower leveJs (host: [5-20 km, sanid ines 5-12 km) . 2. element maps of sanidine megacryst allow correlation of zonation patterns. Small and large crystals have the same growth zones. indicating that large crystals grew faster than sm ail crystals. 3. Scans acro ss steep Ba contrasts give residence times between sorne thousand (core) and sorne few tens of years (rim) for a temperature of 845°C. Residence times are independent of crystal size. 4. Linear growth rate s derived from residence time and later al distance to the rim range between 10-9 and 10-13 cm/s for temperatures between 845°C and 915 °C.

798