Nickel isotope fractionation during metal-silicate differentiation of planetesimals: Experimental petrology and ab initio calculations

Texte intégral

(1)OATAO is an open access repository that collects the work of Toulouse researchers and makes it freely available over the web where possible. This is an author’s version published in: http://oatao.univ-toulouse.fr/25154. Official URL: https://doi.org/10.1016/j.gca.2019.10.028. To cite this version: Guignard, Jérémy and Quitté, Ghylaine and Méheut, Merlin and Toplis, Michael J. and Poitrasson, Franck and Connétable, Damien and Roskosz, Mathieu Nickel isotope fractionation during metal-silicate differentiation of planetesimals: Experimental petrology and ab initio calculations. (2020) Geochimica et Cosmochimica Acta, 269. 238-256. ISSN 0016-7037. Any correspondence concerning this service should be sent to the repository administrator: tech-oatao@listes-diff.inp-toulouse.fr.

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15) . ! % % ! #% % $! % % % %

(16) % % % " +'` 1J@[=MS@U]` ;`.LXDLYQ=` #+,` 1',` #,`.LXFZT=` M3J7=` 5.` 1J@[=MO@U]`9=`.LXDLYQ=` #+,` 1',` +` #,` 0XFXS=` N3J7=` +!`/` %*,`$(` #,`/`

(17) `3DD^=Ì@H` "LJOL` )

(18) ` ` 0XFXS=`8=<`

(19) ` N3J7=` "'"` #+,` 1"*` ` -M6LJJ=` 1J@[=MS@U_T` 2K@[=MS@U]` '@=MM=`=U` "3MA>`XMC` "XS]XI` &4U@LJ3D`9@OUL@M=`#3VXM=DH` '` ` ` MX=`X[@?` '3MBR` ` ` 3J7=è. . 1G9m¦ cncAG¦W9>czy9czy¦zP¦yc>lGn¦czzO¦a9¦;GGy¦GG cwGy9nn ¦9yc]GE¦9¦¦/¦ca¦z `Gy¦_`9>ccG¦9 cy`¦ Uzw¦

(20) ¦ z¦ !!¦9w¦9yE¦U|¦y¦E9czy¦Uzw¦¦z¦¦a¦'za¦lgyGc>¦9yE¦Gqc;j¦W9>czy9czy¦a9G¦;GGy¦E cGE¦%¦cG¦nzz¦G¦¦9¦GE¦cy¦ac>a¦aG¦wG9n¦GGzc¦c¦9¦G¦yc>lGo¦dG¦aznEcy`¦9¦cnc>9G¦wGn¦EzpG¦zP¦9yzacG¦ EczcEG¦GG>c>¦>zwzcczy¦)cy`¦aG¦>zG¦zP¦aG¦GGkwGy¦Ec\czy¦zP¦yc>lGn¦Uzw¦aG¦cG¦z¦aG¦cqc>9G¦zDGE¦ 8aG¦cxH>9nG¦z¦G9>a¦>aGwc>9n¦Gcnc; cw¦9¦P3 EGGyEGy¦9yE¦FLG9GE¦Uzw¦¦z¦¦az¦9¦>zyEcdzy¦;G>9wG¦wzG¦ GE>cy`¦ 8aG¦czzcI¦>zwcczy¦zP¦G9>a¦OGzc¦9¦EGGwjyGE¦; ¦1nc>znnG>z¦syE>dGn ¦(znGE¦5n9w9¦19¦6>zw¦ G ¦1(¦,(516¦ 9SG¦2c¦c]>9czy¦8aG¦czzcI¦>zwzcczy¦9¦U|yE¦z¦;G¦>zy9y¦cy¦aG¦wG9nnc>¦c¦ac>a¦aGG¦ U|G¦=a9GE¦9¦9y¦cy]yc¦ G¦GGzc¦3y¦aG¦>zy9 ¦zy`¦lcyGi?¦U9>czy9czy¦9¦z;JGE¦gy¦aG¦cncAG¦wGq¦{2c¦Ezy¦ z¦ " }~9w ¦ Go9cG¦z¦aG¦9yE9E¦-zzcI¦Gcnc;cw¦9¦ c>9qn¡¦K>aGE¦9PG¦¦az¦+z¦Gcoc;9GE¦9w eG¦9¦¦/¦ yz¦wG9n¦cnc>9G¦W9>czy9czy¦9¦z;GGE¦cacy¦y>G9cy ¦64¦ca¦¤2c 7fvfB $¦ ¦£¦ 9w ¥ 8aGzGc>9n¦ An>n9czy¦ zP¦ wG:n¦cnc>9G¦ czzG¦ U9>czy9czy¦ 9¦ Gcnc<cw¦ GG¦ 9nz¦ GU|wGE¦ zy¦EcXGGy¦ wG9o¦ cnc>9G¦ Gw¦8aOG¦>9n>n9czy¦>zy^w¦¦aG¦9;Gy>G¦zP¦U9@czy9czy¦9¦ acà¦Gw9G¦9yE¦ ¦9¦Kl¦GwG9G¦ EGGyEGy>G¦U|¦2c¦czzcI¦U9>czy9czy¦V|¦aG¦wG9n¦znccyG¦ 9yE¦wG9n¦ zGyG¦9c¦ca¦aG¦wG9q¦=cy`¦ rcàn ¦ncàG¦ gzzcAnn ¦ 3¦GG cwGy9n¦ E99¦ GG¦ ]y9nn ¦>zw9GE¦ ca¦ y99o¦ 9wnG¦ 6zG¦ wGzcEGcG¦ zy ¦czy¦ wGGz cG¦ az¦ 9¦ ¤2c 7hvhC >nzG¦z¦ GGcwGy9n¦ 9nG¦ 9¦Gcnc;cw¦ aGG9¦zaG¦Gac;c¦ zdcG¦wG9r¦cnc>9G¦U9>czy9czy¦ a9¦ >ztE¦GTuM¦ ljyGc>¦z>OG¦ (zyGGn ¦9nn9cG¦ Ecn9 ¦9¦zy`¦ yG`9cG¦ wG9n¦cngAG¦ Q>gzy9gzy¦8ac¦ wz¦ nclGn ¦ Gn¦Uzw¦lcyGc>¦z>OG¦ca¦2c¦EcYczy¦Uzw¦aG¦cnc>9G¦z¦aG¦wG9n¦a9G¦E¦z¦9¦>a9y`G¦zP¦2c¦9cczy¦>zG[>cGy¦ Ecy`¦>zzncy`¦.y¦ac¦GG>¦G¦yzG¦a9¦cy¦ aGG¦9nn9cG ¦czy¦czzG¦az¦wG9n¦cnc>9G¦R>dzy9dzy¦b9¦c¦zzcG¦ EcG>czy¦z¦2c¦zcy`¦aG¦cEG9¦zP¦lcyGc>¦czzG¦U9>czy9czy ¦9z>c9GE¦ca¦+G¦2c¦cyGEcZczy¦ ¢¦#¦ *nGdG¦0E¦&nn¦cà¦GNGE¦ \M:R` (C5H7JeCWNZNQW e*I2L7Z7WCK2IWe'7Z2e WCID52Z7e6C=8T7LZB2ZCNLe!`P7TCK7LZ2Ie P7ZTNIN?ae ".e 52I5\I2ZCNLe '<7NVCZ7We. NTT7WRL6BL?e2[Z@NTe2Ze $,*e0LC^7TWCZce67e.N\IN[X7e(,-e0+e(!-e/N\JN]7e "T2M7 I3@E`WGM=PP` F7T7Ka?\C?L2T6ET2PNKP7[e%e#[C?L2T6. @ZYS 6NCNT? G?52 .

(21) 1. INTRODUCTION Non traditional metal stable isotopes (e.g., Fe, Ni, Zn, Cu, Cr, Cd, Ga, Ge, Hg, Mo, U, W) have received growing attention over the past decade as they are powerful tools to trace (bio ) geochemical processes ranging from alteration at the Earth’s surface (e.g. McManus et al., 2002; Rouxel et al., 2003; Hohl et al., 2015; Noordmann et al., 2015; Wang et al., 2016; Baronas et al., 2018) to metal silicate dif ferentiation in planets and asteroids (e.g. Poitrasson et al., 2005, 2009; Georg et al., 2007; Schoenberg and von Blanckenburg, 2006; Luais, 2007; Hin et al., 2013; Bonnand et al., 2016; Kempl et al., 2016; Mahan et al., 2017; Liu et al., 2017; Bourdon et al., 2018). Nickel (Ni) is the second most abundant element in the metallic cores of planetesimals and planets. It is thus an ele ment of major interest in the study of metal silicate differen tiation in the early solar system. Until recently, nickel isotopes have mainly been used for nucleosynthetic anoma lies applications (Quitte´ et al., 2006a; Regelous et al., 2008; Steele et al., 2011; Render et al., 2018) and in radiochronol ogy (Shukolyukov and Lugmair, 1993; Quitte´ et al., 2011; Tang and Dauphas, 2014) to date early events in the solar system based on the decay of 60Fe to 60Ni with a half life of 2.62 Ma (Rugel et al., 2009). Coupled with 26Al, 60Ni has also been used to quantify the available heat sources required for planetary melting, and thus to unravel thermal histories of meteorite parent bodies that accreted early in the protoplanetary disk (e.g. Tachibana and Huss, 2003; Telus et al., 2018). Mass dependent fractionation of nickel stable isotopes in meteorites and planetary bodies has received much less attention than other elements such as iron and silicon, mostly due to technical and analytical issues. Nonetheless, for roughly ten years, analytical devel opments have allowed the study of stable nickel isotopes across a wide range of scientific fields, from biological (Cameron et al., 2009), to oceanic (Fuji et al., 2011; Gueguen et al., 2016), environmental (Ratie´ et al., 2015a; 2016), planetary differentiation and cosmochemical pro cesses. First measured in the metallic phases (kamacite and/or taenite) of different types of meteorites (Quitte´ et al., 2006a; Cook et al., 2007; Dauphas, 2007; Moynier et al., 2007), nickel isotopic signatures exhibit large varia tions, from 0.05 to 0.30‰.amu!1. Similarly, the Ni iso tope composition also varies within meteorite groups, e.g., among ordinary chondrites or iron meteorites types (Gall et al., 2017 and references therein). A fractionation is also observed as well in metal grains of CB/CH chondrites (Weyrauch et al., 2019). Last but not least, recent studies based on in situ measurements of nickel isotopes in silicate and metal phases in pallasites and mesosiderites show vari able metal silicate fractionation of nickel isotopes. In detail, pallasites exhibit negative metal silicate fractionation of nickel isotopes, from 0.106 ± 0.043 to 1.016 ± 0.080‰.amu!1, whereas mesosiderites have positive metal silicate fractionation of nickel isotopes, spanning a narrower range, from 0.045 ± 0.009 to 0.170 ± 0.040‰. amu!1. These differences in sign, range and absolute values possibly reveal various processes such as equilibrium, kinetic (diffusion and/or volatility) (Chernonozhkin et al.,. 2016, 2017). Thus, Ni isotope fractionation seems a power ful tool to unravel important questions in planetology. However, nickel is strongly siderophile, which means that the Ni content in silicates is usually low. Nickel can be con sidered as a trace element in the main silicate minerals found in meteorites. For instance, nickel contents in mete oritic olivine is typically between 20 and 40 ppm in palla sites, but it can reach up to 0.25 wt% NiO in the olivines of more oxidized meteorites such as Rumurutiites (Petry et al., 1996). It is to note that in planetesimals, the main sil icate reservoir of Ni is olivine. One of the key questions in planetology is whether the Ni isotope composition of the Bulk Silicate Earth (BSE) is chondritic or not. Based on various studies (Steele et al., 2011; Gueguen et al., 2013; Gall et al., 2017; Quitte´ et al., 2017; Klaver and Elliott, 2018), dNi values for the BSE span from 0.025‰.amu!1 to 0.115‰.amu!1 relative to chondrites. Hence it remains unclear whether the BSE and chondrites show the same isotopic composition within uncertainty or not, and nickel isotope data in planetary objects remain scarce. If this difference is confirmed, it could be explained by metal silicate differentiation. In par allel, the question can be tackled using an experimental approach. Contrary to iron, nickel is only present as Ni2+ in silicates. It can then help to decouple the various pro cesses responsible for isotope fractionation. According to Elardo and Shahar (2017), nickel is also an important ele ment to investigate, as it might be the ingredient controlling iron isotope fractionation during core formation. This remains a debated issue, however (Poitrasson et al., 2005, 2009; Chernonozhkin et al., 2016, 2017; Liu et al., 2017). Until now, a single experimental study exists and explores the isotope fractionation of Ni in a metal talc system (Lazar et al., 2012). Besides metal silicate differentiation, Ni isotopes can be fractionated by volatile depletion pro cesses in a way similar to Fe and Si, as suggested by Quitte´ et al. (2017) to explain the differences in isotope sig natures between the Earth, the Moon, Mars and Vesta. Indeed, nickel is a moderately volatile element that can eventually evaporate for long duration experiments at high temperature. In the present study, high temperature (1623 K) experi ments have been performed under reducing conditions (fO2 = 10!8.2 and 10!9.9 atm) to quantify kinetic and equi librium fractionations of nickel isotopes between a solid metal and a silicate melt under similar conditions. Accord ingly, most meteorites including pallasites, mesosiderites and primitive achondrites formed in a reducing environ ment (along the iron wu¨stite (IW) buffer to 5 log units below, i.e., from IW to IW 5). As stated above, Ni can only be present as Ni0 or Ni2+ and its metal silicate partition coefficient depends on the oxygen fugacity. Hence, two dif ferent fO2 conditions have been investigated in the frame of the present work. Additionally, first principles calculations were performed on different metal silicate mineral systems, namely Ni olivine and Ni diopside to determine theoretical equilibrium Ni isotope fractionation factors and their tem perature dependence. Both experimental and theoretical data are finally combined to interpret the variability of nickel isotopes signatures in meteorites..

(22) 2. METHODS. results. In details, experiments at 24 hours were at least duplicated for both types of experiments.. 2.1. Experimental petrology 2.2. Nickel separation and analysis Experiments have been performed to approach metal silicate equilibrium in planetesimals. The metallic reservoir is a nickel wire 0.5 mm in diameter (Alfa Aesar, puratonic, 99.999%) and the silicate reservoir is a melt of anorthite diopside eutectic composition which corresponds to a good analogue of silicate composition in natural samples and that can cover a wide range of temperature under its liquid form. Pure Ni wire has been chosen instead of a Fe Ni alloy, more representative of a planetary metallic core, because Ni covers a wider range of fO2 in its metallic form; its solubility in the silicate glass is therefore easier to control under reducing conditions. Besides, using pure Ni avoids interactions between different elements such as Fe Ni inter diffusion processes or competition between both. A large batch of the glass ("50 g) was synthesized from reagent grade oxides (SiO2; Alfa Aesar, 99.8% and Al2O3; Acros Organics, 99+%, extra pure) or carbonates (CaCO3; Acros Organics, 99+% and 4(MgCO3)#Mg(OH)2#5(H2O); Acros Organics, containing 40.39 wt% MgO). The starting mixture was first fired to 1073 1273 K for several hours to ensure complete dehydration and decarbonation. The mix ture was then melted in air in a platinum crucible at 1723 K for 2 hours, quenched and finely crushed in an alumina mortar. These steps were repeated twice to obtain a well homogenized glass as demonstrated by the multiple ("100) electron microprobe analyses of its composition (Table 1). Time series experiments were performed using the wire loop set up. Glass droplets of approximately 30 mg were prepared and suspended on a Ni wire loop ($25 mg). In this setup, at high temperature, nickel diffuses from the wire to the silicate melt where its concentration rapidly reaches chemical equilibrium. For each experiment, 5 samples were loaded simultaneously in the furnace and quenched in air at different times, thereby allowing to quantify kinetic effects. All experiments were performed at a single temperature (1623 ± 1 K). Two different oxygen fugacities (10!8.2 and 10!9.9 atm; i.e., 2 and 3.7 log units below the Ni NiO buffer) were used, fixed by CO:CO2 gas mixture. These conditions of temperatures and fO2 have been chosen to mimic condi tions of metal silicate differentiation in natural objects. Run durations ranged from 30 minutes to 7 days (Table 2). Replicate experiments were carried out in order to confirm the reproducibility of the experimental conditions and. After experiments, the metallic wire and the silicate glass droplet were mechanically separated under a binocular microscope. Each part was then digested on a hotplate at 120 130 !C for at least 3 days in a mixture of HCl, HF and HNO3 with volume ratios 2:1:1 and 1:1:0.05, for metal and silicate, respectively, and in a total volume of 4 ml. After digestion, samples were dried and then taken up in 2 ml of 6 N HCl twice to convert fluorides into chlorides. A small aliquot of 5 % in volume was then saved for con centration analysis (using ICP OES and/or ICP MS) and the remaining solution was dried down again, ready for nickel separation. The latter consists of four steps and follows the method developed by Quitte´ and Oberli (2006b). The main differ ence in the chemical separation between metal and silicate is that the dimethylglyoxime (DMG) step (see below) is not used for Ni wires. Briefly, the samples were re dissolved in 2 ml 9 N HCl and loaded on a AG1 X8 anion exchange resin. As nickel elutes immediately, it was directly collected in the load fraction. Elution was completed with additional 4 ml of 9 N HCl. After evaporation, the sample was taken up in 2 ml 2 N HCl and loaded once more on the same resin. Nickel elution was again completed with 4 ml of HCl 2 N. These two steps allowed efficient separation of iron and zinc, two elements that have isobaric interferences with nickel at masses 58 and 64 respectively. The dry sam ples were then dissolved in 5 mL of 1 N HCl and 1 ml of 1 N ammonium citrate and the pH was adjusted to 8 9 with a few drops of concentrated ammonia. The samples were subsequently loaded on a Ni specific resin containing DMG that complexes Ni and lets most of all other elements flow through the resin. Only few elements form chelates with DMG: Ni, Co, Cu, Pd and Pt. Copper and Cu were separated during the first step using AG1 X8 resin, hence they are not present in the Ni cut anymore. Contrary to Ni, palladium only reacts with DMG in acid solution, and platinum and Pd are anyway removed later during the separation procedure (see below). Elution of this Ni DMG organic complex was achieved by adding 12 ml of 3 N HNO3. In order to break the Ni DMG complex, a few drops of HClO4 were added to the solution before it was evaporated to dryness. To destroy organic matter orig inating from the resin and residual DMG in the case of. Table 1 Starting composition of the silicate glass.. SiO2 CaO Al2O3 MgO a b. Expected starting composition (wt%). a. 50.33 23.49 15.39 10.79. 50.52(35)b 22.95(31)b 15.22(25)b 11.30(19)b. Measured starting composition (wt%). Average value of measurements made on 100 electron microprobe points. Number in parentheses indicate two standard deviation on the last two digits..

(23) Table 2 Experimental data. dNiM (a.m.u 1). b. b. dNiS (a.m.u 1). DNiM S. Replicatesc. (08) (11) (05) (04) (11) (04) (04) (06) (06) (06) (01). 0.52 0.53 0.72 0.16 0.18 0.10 0.13 0.00 0.07 0.03 0.01. (11) (11) (06) (10) (12) (07) (05) (07) (07) (08) (01). 7 6/4 5/5 3/4 5/3 4/5 3/4 3/4 4/5 3/4 3/4 3/3. (04) (03) (03) (07) (01). 0.50 0.36 0.15 0.03 0.03. (05) (03) (04 (07) (03). 3/3 3/5 3/4 3/4 3/4. Sample. Temperature (K). Log fO2 (atm). Time (hours). Concentration (103ppm)a. Niwire Niwire+(An:Di)melt Niwire+(An:Di)melt Niwire+(An:Di)melt Niwire+(An:Di)melt Niwire+(An:Di)melt Niwire+(An:Di)melt Niwire+(An:Di)melt Niwire+(An:Di)melt Niwire+(An:Di)melt Niwire+(An:Di)melt Niwire+(An:Di)melt. 298 1623 1623 1623 1623 1623 1623 1623 1623 1623 1623 1623. Air 8.2 8.2 8.2 8.2 8.2 8.2 8.2 8.2 8.2 8.2 8.2. 0 0.5 1 2 3 5 10.6 17.9 24 24 48 168. 103 2.64 3.74 4.88 5.68 6.01 6.97 8.81 9.13 8.95 9.20 8.90. (49)d (77)d (60)d (98)d (43)d (41)d (16)d (17)d (11)d (38)d (15)d. 0.25 0.25 0.27 0.26 0.26 0.23 0.28 0.27 0.21 0.26 0.27 0.13. (06) (08) (03) (03) (09) (05) (06) (02) (05) (02) (05) (01). 0.77 0.80 0.98 0.42 0.40 0.38 0.39 0.21 0.33 0.30 0.14. Niwire+(An:Di)melt Niwire+(An:Di)melt Niwire+(An:Di)melt Niwire+(An:Di)melt Niwire+(An:Di)melt. 1623 1623 1623 1623 1623. 9.9 9.9 9.9 9.9 9.9. 0.5 1 2.8 24 24. 1.01 1.14 1.31 1.35 1.34. (02)e (01)e (01)e (01)e (01)e. 0.25 0.26 0.24 0.24 0.24. (02) (02) (02) (03) (02). 0.74 0.62 0.39 0.27 0.27. Processed Ni std. 298. Air. 0.0003. 0.02 (02). 3. Glass + Ni std. 298. Air. 0.0003. 0.01 (02). 3. a b c d e. Numbers in parentheses represent 2SD on the last two digits, determined from at least three replicates. M = Metal, S = silicate melt. Metal/silicate. Measured by ICP OES. Measured by ICP MS.. Ni rich samples, the dried samples were then dissolved twice in a 30% H2O2 + 16 N HNO3 (1 ml + 1 ml) mixture and evaporated again to dryness. Finally, the samples were taken up in 0.2 ml H2O and loaded onto a AG50W X8 cationic resin for a final purification step. The main goal of this third step using small columns is to remove organic matter, as well as Pt and Pd if any. Matrix elements were washed out with 2.5 mL of 0.2 N HCl and nickel was eluted with 0.5 ml of 3 N HCl and evaporated to dryness. The yield of this procedure was 95 ± 3%. The small loss of Ni occurred when using dimethylglyoxime (DMG) on the sec ond column and induced no Ni isotope fractionation as wit nessed by the isotope composition of the processed standard, similar to that of the unprocessed isotope standard. Nickel isotope measurements were performed using a combination of the standard sample bracketing technique and Cu doping using a Neptune MC ICPMS (ThermoFis cher Scientific) in ‘‘medium resolution mode”, i.e., a resolv ing power M/DM " 7000 (5 95% peak edge definition) at the Observatoire Midi Pyreńeés, Toulouse. A double spike approach was not used for Ni isotope analyses in this study to avoid clean laboratory and our instrument contamina tion. Besides this study, we also conduct in Toulouse the measurement of mass independent variations of Ni isotopes in cosmochemical samples, which are particularly sensitive to small uncertainties in 61Ni, 62Ni and 64Ni, especially when looking at nucleosynthetic anomalies. The Cu doping method adopted in this study has a reproducibility compa rable to that attained using a double spike approach. All. nickel isotopes (58, 60, 61, 62, 64) were measured simulta neously as well as 57Fe and 66Zn in order to correct for iso baric interferences of 58Fe on 58Ni and 64Zn on 64Ni. These corrections are efficient if the Fe/Ni and Zn/Ni ratios are lower than 0.1 and 10!3 respectively (Quitte´ and Oberli, 2006b). This is always the case for Fe/Ni ratios, but more difficult to control for Zn/Ni due to external Zn contamina tion, hampering the systematic use of 64Ni. After nickel separation and purification, samples were dissolved in 0.1 N HCl and their Ni concentration was adjusted to match within 10 % the concentration of the bracketing standard (an Aldrich standard solution). More over, in order to correct for the instrumental mass bias, copper (Cu) was used as an internal standard and an expo nential law was considered for the calculation. Copper was added both to samples and standards, typically to half the concentration of nickel. Hence, 63Cu and 65Cu were mea sured in a second cycle during the run in dynamic mode (one cycle for Ni, Fe and Zn isotopes, the other for Cu). 60 Ni is a radiogenic isotope whose abundance may vary in meteorites due to in situ decay of 60Fe. We therefore pre fer to avoid 60Ni when reporting mass dependent isotope fractionation in meteorites or in experiments when the experimental data are dedicated to interpret results obtained in natural samples. Besides, once evidence is pro vided that the observed fractionation follows a mass depen dent law (see Fig. 2), the reported di/jNi ratio does not matter (i and j can be any isotope of Ni). Reporting the iso tope fractionation per atomic mass unit (a.m.u.) should allow a direct and easy comparison with literature data..

(24) This notation has already been used by different authors for, e.g., Cu, Zn, Ca, Ni. All nickel isotopic ratios are reported relative to the Ni Aldrich standard corrected for mass bias using the internal standard, and expressed according to the following equation: ! ! " ð i Ni= j NiÞsample!Cucorr 1000 1 ( dNi ‰:amu!1 ¼ ð i Ni= j NiÞstd!Cucorr ði jÞ. where (iNi/jNi)sample-Cu corr is the measured iNi/jNi ratio in the sample corrected for the mass bias using copper and (i Ni/jNi)std-Cu corr is the average of the iNi/jNi ratio of the standards that bracket the sample in the analytical sequence, corrected from mass bias using copper. Measure ments of the Aldrich standard solution relative to the certi fied SRM986 isotope standard confirmed that both have the same isotopic composition (dNi = 0.05 ± 0.06‰. amu!1). A standard processed through the whole separa tion procedure shows an isotopic composition of 0.02 ± 0.02‰.amu!1 demonstrating that our nickel purification procedure does not fractionate its isotopes. The external reproducibility (2 SD) achieved on standard replicates over a year, i.e., "300 measurements, is 0.04‰.amu!1 and 0.07‰.amu!1 for d(62Ni/58Ni) and d(64Ni/58Ni), respec tively. Similar reproducibility is found on the nickel wire (dNistarting metal = 0.25 ± 0.06‰.amu!1, n = 7 replicates). To check accuracy of the isotope measurements, a glass droplet comparable to those used for the experiments, but doped with our Aldrich Ni standard, was also processed: it yields a dNi of 0.01 ± 0.02‰.amu!1, undistinguishable from the matrix free processed and unprocessed standards. Hence, isotope analyses are not hampered by matrix effects. We also note that there is no evidence of correlation between the reproducibility and odd / even isotope pairs, indicating no significant mass independent fractionation effect, if any. The external reproducibility on the samples is calculated on at least three replicate measurements (Table 2). Finally, the nickel isotope fractionation between metal and silicate melt (DNiM-S in amu!1) is expressed as: DNiM!S ¼ dNiM. ð2Þ. dNiS ;. with dNiM the nickel isotope composition in the metal (in amu!1) and dNiS the isotope composition of the silicate melt (in amu!1). 2.3. ab-initio calculation A theoretical approach was also used to estimate the DNiM!S at equilibrium for different metal silicate systems with the aim to compare them with the experimental data, to refine the interpretation of the latter and to extrapolate to lower temperature conditions. Equilibrium fractionation properties of condensed phases are expressed theoretically by their b factors, which are related to the change in free energy associated with iso topic substitution in a particular phase. The connection between the D values and the b factors is expressed via the isotopic fractionation between phases A and B: DNiA!B ¼ 1000 ln bNiA. 1000 ln bNiB ;. ð3Þ. where ln bΝiA(B) are the so called b factors of Ni isotopes in phase A (respectively B). b factors of metal and silicate phases have then been computed from their phonon frequencies based on the gen eral theory (see e.g. Meheut et al., 2007). For the metal (pure Ni, fcc structure), the phonon fre quencies were computed from first principles using density functional theory (DFT) (Hohenberg and Kohn, 1964; Kohn and Sham, 1965). The calculation was based on the exchange correlation functional of Perdew, Burk and Ernz erhof (PBE) (Perdew et al., 1996), a plane wave basis set, and atomic pseudopotentials as implemented in the Quan tum Espresso package (Giannozzi et al., 2009). Pseudopo tentials used for Si, O, H are described in Me´heut et al. (2007). The pseudopotential used for Ni comes from the GBRV (Garrity, Benneth, Rabe, Vanderbilt) pseudopoten tial library version 1.4 (Garrity et al., 2014), verified within the SSSP (Standard Solid State Pseudopotentials) effort (Lejaeghere et al., 2016). For the silicate phases, the calculation may be more dif ficult. Indeed, insulators containing transition elements appear sometimes challenging to model, due to the failure of classical DFT methods to model strongly correlated elec tronic systems (Cococcioni et al., 2005). In this case, a + U correction might be necessary for the calculation to cor rectly reproduce the insulating behavior of these systems (Cococcioni et al., 2005). However, we note that the insulat ing materials considered in this study (Ni olivine and Ni diopside) were found to be insulators even without a +U correction, suggesting that this latter correction is not nec essary in the calculations (Zbiri et al., 2008). Therefore, in the present study calculations were performed without the +U correction. Finally, these different phases show com plex magnetic orders (Hagemann et al., 2000; Durand et al., 1996). In order to quantify the effect of this order, dif ferent tests have been performed on Ni olivine structure (see discussion in the Appendix). Such tests show that the type of magnetic order has a negligible effect on the calcu lated fractionation properties (Fig. 1 of the Appendix). Following these considerations, in the rest of this study, the magnetic structure considered for Ni olivine is the one referred to as configuration 1 by Cococcioni et al. (2005) (who studied Fe olivine), as it is slightly better in reproduc ing experimental structure (see Appendix). For Ni metal and Ni diopside, a ferromagnetic structure was considered. The structural parameters of Ni metal were found to be vir tually indistinguishable from experimental ones, whereas those of silicates were found to be 1 2% larger than the experimental ones, which is typical of this type of method (Table 1 of Appendix). Electronic wave functions were expanded in plane waves up to an energy cutoff ecut = 60Ry for the metal and ecut = 80Ry for the silicates, and the charge density cut off set to 8ecut. The electronic structure integration was performed by sampling the first Brillouin zone with a 20 ( 20 ( 20 k points grid for the Ni metal, a 2 ( 2 ( 2 k point grid for Ni2SiO4 and a 2 ( 3 ( 2 k point grid for CaNiSi2O6. The dynamical matrices were computed using the linear response of Baroni et al. (2001), with the PWSCF package.

(25) C}P¶|»^©» P » -))6 »^²PV©´»±}ªz}» @Bk»» \}p^^ª v¡}\¤» j» »¯c©¤» 4» ³»4»³» 4» n» G}» ^©P» -»³» -»³»-» n G}K}· » -» ³»/»³» -» r»?PG}K}H Lz^» z» s^¬^» V}^¤» ±^^» ªz^» S©P}^\» }» P´» }©» j»^VVP» ¤PV^ ¬¤}v»©z^»¤©P\P\»B¬}^»}©^¡Pª}»©cz}¬^»B»¤}" }VP©^»Pª^P¤» ª{^»¯~TPª~P» Pª~ª» tV©~»±P¤» V» ¯^v^\»±}ªz»P»2»2»2»}©^ Pª}» v\

(26) » ±z}^»r» ^ªP P»-)»³»-)»³» -)»y}\»±P¤»¬¤e». . i . i. . i. . ?V^ª¡Pª}¤»j»G}»zP¯^» ´»S^^»^P¤¬^\»}»S¬ ¤} }Y©^» vP¤¦¤» Pjª^» ¤P^» \}v^¤©} » ¬¤}v» D?J»HAK» Pª DJCJ»JP¡}¤ »»D?J»FK» Pª»ªz^»HFJ »L¬¬¤^ »\^^\}v » }¤©¬^ª»P¯P}PS}}ª´»LPS^» -»»B}v» *»G}V^" V» V^ªP©» }» ªz^» ±}^» zP¤» ª» S^^» b¤¬^\» ¤}V^» }ª» }¤» P\^» j» ¬ªP¬^» G}» B» ^²^}_ª¤» ^r¡^\» Pª jH +)¥".» P©»B}v» P »ªz^»G}»VV^©Pª}» }V^P¤^¤ ±}©z»©}^ » s» -2))»» Pjª^»/)» }¬ªi» j»z^Pª» ª^Pª» ^ª»Pª»*2-/»©»¸6)))»»Pj©^»*36»z¬¤»B»v^ \¬Pª}¤ »VV^ªPª}»^P}¤»V¤©Pª»Pª»Eb¤ª»¬»ª»04 z¬¤»¤¬vv^¤ª}v»©zPª»Vz^}VP»^¬}}S}¬»}¤»^PVz^\»B ^²^^©¤»ô^\»¬\^»^»e¬V}v»V\}ª}¤ »}^ jH +) P©»B}v» S »©z^»G}»VV^©Pª}»}V^P¤^¤ s» +)))»» Pjª^» )1» z¬¤» j» z^Pª» ª^P©^ª» ª +/))»»Pjª^»-4»z¬¤:»B» v^»©}^¤»¬»ª»-0»z¬¤ ªz}¤»VZ©Pª}»^P}¤» ¤ªPS^»Lz¬¤

(27) » Vz^WP" ^¬}}S» }¬»}¤»^PVz^\»¸»2»ª}_¤»nR¤©^»¬\^»ªz^¤^»^»^\¬V}v V\}ª}¤»Lzi^» i¬ª¤» P^»}» ¯^´» v\»Pv^^^©»±}ªz ^¯}¬¤» ¤ª¬\}i» » ©z^»¤ ¬S}©´» j» G}» }»P» ¤}}Yª^» ^ª j»¤}}P»V¤}ª}»P©»¯P¡¬¤»©^^Pª¬^¤»P\»²´v^ nvPV}ª´»V\}ª}¤»¤^^»A»/»}» @}v±^»^©»P » +660. . i i i i i i . . . . . . i i . . . / Gi,i. i. . . . º . i. . .

(28) . @!iPi f0i2YKi i. i. . i. i. i.

(29)

(30). . h"i,i. .i i

(31) ii2[i i. i. +.i. i. i. i. i.

(32)

(33) 'EB i i &aOH\YEOLiO<i-Ei 7OM9LYV1YEOLiELiRSiELiYC9iWEHE71Y9iK9HYi bEZCi ZEK9i1L8iAViYDi YbOiWY_F98iOcdB9Li?^B17EYdi7OL8EYFOLWi 1i i gi 1ZKi 1L8i 4i i gi 1YK i$6KE71Hi 9U\EHE4VE]Ki EWi V917598 1<Y9Vi iCO\VWi ?QViYC9iKOWYiOcE8EeFLBi7OL8FYEOLWi1i1L8i1>Y9ViOLHdi i CO\Vi ELi YC9i OYC9Vi 71Xi 4 i )Li 4OYCi 71XWi 7ONLZV1YEOLWi K91W\V98i 1Yi 9U]EIE4VEì 1V9i ELi a9Vdi BOO8i 1BV:K9LYi bEYCi YC9i WOH]4EHEYdiO=i-EiELiWEIE71Z9iK9IYibEYCiWEKEH3Vi7OKTWEYEOLi71H7\I1Y98i =VOKiYC9iWOH\4FHEYdiH1bi1<Z9Vi%ELBb9HHi9Zi 1H ii i. .

(34)

(35) . B» bVz» ^²^¡`© » ªz^»}¤ª}a»V¤}ª}¤» j»©z^ ^ªP" ±}^»P\» j»ªz^» ¤}}VPª^» ^ª»zP¯^» Sf» \^ª^}e P\» P^» ^¤^ª^\» }» MPS^» - » B}v¤» -» P\» /»Lz^»¤ªPª}v ^ªP" G~!;» zP¤» P» V¤}©}» j» G}!" )-1» ))2P¬, »^P¤¬e»»3»^VP©^¤» D» P" ¤P^¤ G}» }» S©z» ^ªP" P\» ¤}}VP©^» }¤» }¤©}VP´» }vz©^» V£» P^\»©»ªz^»SPV^ª}v»¤ªP\P\¤<»ªz^}»G}»¯Q¬^¤»Pv}v s» )-4»©» ) */P¬ " j» v^¤©»^²^}`ª »¤^^ ^² PPª}» S^±» r» ªz^» ^ªP» P\» o » )64» ª )*0P¬ " r»ªz^» ¤Y©^»^©»LPS^» -» P» ªz^^» }¤©^»\}PvP»5 G}»¯¤» G}

(36) »G»\P©P»ª»»P»P¤¤ \^^\^ª»sPVª}Pª}» ^ »±}©z»P»¤^»j»)623»))21 V¤}\^v» P\»^»SP¤»»bVz»}©»?¤}\^v \}lm^^ª»\P©P¤^©¤»r»^PVz»zP¤^»\¬}v»}^©}V»»^¬}}S» }¬»^²^©¤» ¤z±»ªzP©» ¤^¤»P\» ^» SP¤» ©zPª» P^ }»ªz^»¤P^»Px»j» ªz^»±z^»\PªP¤^ª»B}v»-»Lzg^ª}VP ^¬}}U¡}¬» ¹» P\» }^©}V» ¹» ^²^ª¤» zP¯^» P¤ S^^» VPX¬P©^\ » P\»^ª^\»}»P ¬, »PV[\}v»ª»©z^ r±}v» ^¬Pª}¤» ]^\» o » N¬v» ^ª» P !» -))- B}v»-7» &> '8# 9 » =» %$'O. » ³»

(37) . . . . . . . . 0. =» ³» . . 1. ±z^^» }¤» ©z^» P¤¤» j» P»v¯^» ¤©^»¹" P\» ¹ " P^ r¬\»©»S^»¯^´»V¤^ »^¤^V«¯^ ´»)66-»P\» )640:»ªz^^» n^

(38) »»¤¬Vz»P±»Pv^» j» }¨ª}a»V¤}ª}

(39) »ªz^´ P^» SP^ ´» \}¤©}v¬}¤zPS ^»B}v» -»B}Pµ » SP¤^\» » ªz}¤ /»}¤ªi»©»»P»P±»Pv^»j»[¤}© »©z^»V¬ ^©»PP´©VP" ^V¤}»\^¤»ª»^}ª»©»\}¤V¬¤¤»©z^»}» v}» }^©X» P\(» h¬}}S}®» j» ªz^» S¤^°^\» }¤©^ ¯P}P©}¤»B}v»-

(40) »\}lm^^ª»¤^¤»r»ªz^»\}p^^©»¤¬S¤^©¤ Lz^» ^¯ ¬©}» j» }¤ª}a»V¤}ª}¤» G}» j» ^PVz ^¤^¯}»¢¤^V©¯^ ´»G

(41) " P\»5G}¤»r»^ªP»P\»§}VPª^ ±}©z»ª}^»P\»r»ª±»\}lm^^©»²´v^»nvPV}©´»V\}ª}¤ }¤»¤z±» }» B}v» /P»S»B»©z^»^ªP" i^¯}

(42) » s»I»ª 04»z¬¤»j z^Pª»ª^Pª_ª»P\»±zPª^¯^»©z^»²´v^»uvPV» }©´»V\}«¤ » G

(43) " ^P}¤»V¤ªP©»±}ªz}»¬V^ªP}ª}^¤ G}

(44) " )-1»))1P¬, » P\» }\¤ª}v¬}¤zPS^ s»©z^»¤ªP©v» ^©P" G w» " )-1» P¬,

(45) » »3»^XPª^¤»LPS^» - »B}v» /P»S»B» ªz^»¤}}» VPª^»o PV©}¤ »G}¤»}¤»¤©v´»©}^»\^^\^ª»P\»Sd^¤ z^P¯}^»±}©z» ª}_»\¬v»ªz^»q¤ª» -0» z¬¤ »¤©Pª}v»s ¯P¬^¤»Pª» )33»))4P¬ " Pjª^»/)»}¬©i »\^V^P¤ }v»ª» )64»))1P¬ " Pjª^»-»z¬¤» P©» jH +) .

(46) >u z '>u >SH # az 1z lzw_H buHbwaH>w`#z#)az{6bGzHwzHzlH_ _Hbw_a>bzcIHbob@bubH>BaHG;aH>uHHwG b ]wG]H buHw > lzHT8 >u] # az;aHHWH >w Hblb@b bzz W>Bbzw>bzw ZB z zU-7d zU $>u AHQw uH>m >wG blbB>H a> @HHw bwWOHG ^zu Gb[HHw #a >wG #)a HHbuNw HHbuHw >'"4>wG T8 >u X #a >wG #)a HWvNG bBH =babw xDH>bw aHHbawzuH>lbtbB>HuHlW>Bbzw>bzwzT6bcz zH >Hblb@bwGHaHHBzwGbbzw 2zHH]aHlzw_HH buHw>lwcH') az@zauH>l >wG blbB>H uHl bzzHBzuzbbzw HzlH z>G lb_al aH>bH Bzubbzw {6b " >u >wG {6b #n>u HHBbHl;abbwGbB>Ha>lb_abzzH>H>b>ll lzTzuaHzH>lluH>oblbB>HHu abBau> bwHHHG >aHHlzT > u>ll >uzw zS 6b3z @ zl>blb bw aH zHw HHbuHw>l Hu > >oH>G z@HHG bw aH B>H zS bzw bzz bw bubl> Hu 9bH H>l )5zHzH >BBzGbw_ z =>w_ H>l **#>wG;Bac>u>>wG1icuzz**%aHH>z> bH3zzTbzwHWOHwb>llzBBbwuH>la>bwaHbr bB>H 2zHH 6b Hu>bw aH >uH babw wBH>bwbH>GHHubwHGW #a z#)aHHbuHw bH 6b >u __Hbw_ a> aH. buHElHVbzzbJHHblb@>bzwzTaH>ulHbZH a>wH>z>cH3zzSwbBjHl

(47) yeCkKp h|}K ;aH Hzlczw ba Hu >H W <,# zT HGBHG>bbzwYwBbzw>bz>ozB>llHGZBzzT zTuH>l>wGblbB>Hbbll>HGbw1b_#1z>ll a>H W?BzGHBH>lcwH>lbacwBH>bw_HuH> HbHGPH>bw_ n;!5H>6c~z1H~ Hab@blzH ZBzFu>HGzblbEHubwH>faHH6bbHHw >6b ( BzwbHwlbaaHsLzTau@Hl>bw_W>B bzw>bzw >wG HGz>H:Ba>@qH #HH >lz GbB bzw bw aH .HwGb P 1z aH q>H

(48) 6bbBa zlbbwH 6b:b8 zazazu@bB BH >wG 6bbBa zHwH0>6b:b uzwzBobwbBBH>H@>HlGb bw_ba>@lH>>wHuH>H__Hbw_a>wgBjHlbz zH a>H bubl> zHcH H_>GlH zS aH BH >wGBzuzbbzwzTaHH>waGzblbB>R 1zu aH B>lBl>bzw zT aHH V?Bz Hblb@bu uH>nblbB>HV>Bbzw>bzw6ba>H@HHwGHMubwHG > > WwBbzw zS Hu >H >wG HzHG bw >u 1b_ & /H \z aHElBl>bzw W uH>lzlbbwH >wG uH>nzHwHHu>HGHBb@HG@aHH>czw+. Q 0 0 12

(49) Q P . J 4B -GC6:HH:7QHM3@73F7HQ *;@/;=<53K8Q?8>K#QI=CD8Q%Q!!QNQQ ! (E/<>;53L8Q?8=L"QI>CD8Q&QNQ Q +<A,8M3#QH=CD8Q'Q QNQ Q OQ)E,9K3$QH=CD6Q%QQu| % I>CD8Q#Q. .

(50)

(51) . (. *. - Q. !%|

(52) v w x y z ! - !% !! ( % !| *.

(53) 3TP| #| <RhH|Tj`k`cD|CT>Ph>\|2>k>|NaZ[`r| kQD|khD]C|DscFkDC|M`h| \>jj|CDcD]CD]l|Lh>BlT`]>kT`]|TD|>|ZT_| rTkQ|>|jY`d|Dgm>Z|k`| "| ?Z>BV| ZT]D|7`]P|>]C|jQ`hk|C>jSC|ZT]Dj|hDchDjD]k|WT]DlTB|{X^|,+&|>]C|DgmTZT?iUo|{ ,,#|jZèj|B>ZBm[>kG|rTkQ|lRD|Nai\mZ>j|`L|=`m]P| Dl| >Z| # #|>]C| >hD|?>hD[t|CTjlT]PmTjQ>?ZD|`]| k@Tj|h>]PD|`M|p>ZmDj| 2>k>jDk|Mbh|D>B@| cQ>jD|@>j|?H]|jcZTk|?DkrH]|WT]DkTB|>]C|DgmT[T?hTm\| DscDhT\D]kj|;Z`cDj|>]C|Dhh`hj|#;2| Q>qD|?H]|CDkDh\T]DC|M`h| D>B@| jm?|C>k>jDl|B`]jUCJT]P|Dhh`h| Ahj|`]|?`kQ|>]C| 5]|>]t|B>jD|Tk| Tj| T\fjjT?ZD| k`|BZD>hZt| CTjkT]PmTjQ|?DlrI|WT]DkUB|>]C|DgmTZT?hTm\| DOEkj| :Y`lkDC| Dhh`hj| >hD|#;2| 2>iW|PhH]| BThBZDj|WT_lTB| jTZTB>kD|\DZk.| 8TPQk|PhI|BThBYDj|DgmT[T?iTm\|jTZTB>kD|\D[l/| hDC| BThBYDj| WT]DlTB|9T| rThD|`h>]PD|BThBZDj|DgmT[T?hTm\|9T| rThD/| C>hW|?ZnD|BThBZK|?h>BW| DkT]P|jk>]C>hCj|ZTPQk|?[mD|BThBYD|jl>]C>hCj|ch`BDjjDC|`]|B`Zm\]j|.

(54) |. J -.% -.& -.5.

(55) . J;. . . -). h#.%. UõøzÓ×ÑÂ|å Ù ãø . %. .. &ø. 5ø. ;. . #%. #&. #5. #;. KKøø ø j{{ø® Kø

(56)

(57)

(58)

(59)

(60)

(61)

(62)

(63)

(64)

(65) ¯Køø. -% ºø òø ???ø ó???ë ÷øUñ&. Ìø -i5ø.

(66) . öø -i.;. . -#ø h#%. . %. #W. #. . §. ªí£ø . sîsø .

(67) . 5. W. . . . & n % # ì#ø. jøôUzÒÔÁ»äÆ Úø´ø %W n. . ø. %. . 5. ;. ). m%. m&. )5. );.

(68)

(69) ²+køMkø O\ Dø Ýø 7 ø8¦O3øø7 øÖø7 ø!ø Y ø2ø øZ_4 ø8Søu4ø N$Møyø2ø Lø 4ø2ø81dø4ø LN$Møyø2ø ¤øLøø tïtøø± ø1ø ø $¸·ø ø 7 ø8Ëø¹1 ø$SøO ø>4ø 7Çø !3D1!øø 2ø¿`øL¨øÛø$& > !fø>!ø ø >ø~ 3 ø ! 31!JøO\ ø ø 7 ø Yø øZ_ ø ø ø FÍø ø!++ +øøÏ34 F 3ø1 >F Tø` }ø\ Dø 4 ø 7 ø YøD ø ø ø4 + aøF31 ø¶ ! 2ø ß! 1!øZ_ øø3øF a + 2ø+ ø`>ø~ Dø! ø8fø+ aø dJø .

(70) . . . . ) A )./q/øQø:/rø Qø c. . . . . A ) A¥&øQøø:B/øQø c . q. X

(71) ø

(72) ø *Ræ ø ø Gø ø ÅÊ . (ø Hø ø '

(73) ø Pø

(74) ø øH]ø*^ø

(75) Eø çø ø ø(. ø

(76) ø ð0[ ø G'

(77) ø *Rø ¬I@¡ø

(78) &ø

(79) ø l/Il

(80) ø AB:oø ,øø

(81) CøøÞ 9ø *^ø

(82) Eø Ggøø rI@

(83) ø BI@ *^ø

(84) ø(èø vø 0 ø "E. = ø <ø ,0"=<9ø ø ø Àø

(85) ]ø

(86) ø Ã

(87) øø

(88) ø(øCø0 ,(. éø ø

(89) eø ,

(90) (. (ø ø â 9ø Cø

(91) ø 'ø ]Eø ,"Õø

(92) ø ø:))øø"ø/9øPÐÈø0³=<ø½(ø

(93) á ø ø . ø [

(94) Gø ø Øø*à

(95) gø 'ø *R :b@ ø&ø øBb@

(96) ø AB:oø v ø ø0"=<ø

(97) ø ø

(98) ø øø*R

(99) ø

(100) ø ø

(101) ø

(102) Eø

(103) øÎ'66ø

(104) ø ( 'ø G ø ,"ø /eTø . êø ø H'ø

(105) ¾ ø 0"=<ø ø ¢VPø Éø H' ø ,X wø "'¼669ø ø ø *Ü6ø Cø VP ,"ø /9 ". (Tø Ä '[ø

(106) øXw"ø ø ø ø ø0"=< ( ø

(107) ø ø 6øHø '6ø

(108) ø

(109) øC.

(110) jI¡C;sCu;l|{¡lj¡|¡ILlyI{;s¡G;;¡;¡l|lJ¡Il¡ sl>ly¡4l7¡¡¡!;y . 8jII]~I

(111) ¡ljl{¡ {CI;l{lI

(112) ¡ >|j¡ |¡ IIlyI{;s¡ ;{G¡ jSIlC;s¡ ;|;CjU¡;iII¡lj¡I;Cj¡|jI¡0li¡ C¡;{G¡# ¡ .

(113) . .IlyI{¡ ;{G¡ C;sCu;l|{¡g|y¡ jl¡ G¡ j|¡j;¡{GI¡C|{Gll|{¡ItI;{¡|¡yI;s¡lslC;I¡GlZ\I¡ I{l;l|{¡l{¡s;{IIpy;s

(114) ¡lI¡jlij¡IyI;I¡ & ¡ ;{G¡{Iislil>sI¡II¡IaIC

(115) ¡{lCqIt¡l||¡W=Cl|{;l|{¡ ;¡ Ilsp>ly¡ l¡ {Iislil?sI

(116) ¡ lj¡ 4l7¡ ;sI¡ |X¡ ¡ ¡!;y / l{¡ |¡ IlyI{¡ ;{G¡ $};y / l{`P¡ h|y¡ C=sCs;l|{¡ 8;>sI¡

(117) ¡ 0li¡ C¡;{G¡ # ¡,IlGI

(118) ¡ l{¡ IIpyI{

(119) ¡;;l{zK{¡ |Y¡l||lJ¡Ilsl>¡ ly¡l¡{|¡X5 7 GII{GI{¡0li¡ ¡jII;¡CjV{lC;s¡Il¡ sl>l¡ l¡ 0li¡ ¡ 9jl¡ X5 8 GLI{GI{CI¡ |Y¡ CjIylC;s¡ Ilsl@ly¡l¡GI¡|¡ jI¡ ;l;l|{¡|Y¡yI;slslC;I¡;l¡ l|{l{i¡CdClI{¡jlCj¡l¡jlijI¡;¡y|I¡IGCl{i¡C|{Gl¡ l|{

(120) ¡ sNGl{i¡|¡ _I¡ ;l|{¡ |X¡ lulC;I¡;¡ u|I¡ X5 7 5{¡ jI¡ |jI¡ j;{G ¡ ;;l{yI{¡|X¡ l||lJ¡Ilsl@ly¡ l¡. {|¡ GII{GI{¡ |{¡ jI¡;ll|{¡ EIdClI{+¡ jIIfI

(121) ¡l|¡ |I¡ Glbl|{¡ C|{l{U¡ ;YI¡ CjIynC;s¡ Iltl>py¡ l¡ I;CjIG¡ ;¡ s|{i¡ ;¡ IyI;I

(122) ¡ lI

(123) ¡ GlZ[ l|{¡ C|IelI{

(124) ¡ l¡jlij¡I{|ij¡ 9|¡jI¡>I¡|X¡|¡r{|tPiI

(125) ¡>IfI¡jI¡CI{¡|q

(126) ¡;¡ m{isI¡ILmyI{;s¡G¡j;¡AII{¡AsmkIG¡|{¡4m¡m||I¡ g;Cl|{;l|{¡BII{¡yI;s¡;{G¡lspC;I¡3;; I¡;s

(127) ¡ ¡ 2{¡j;¡G

(128) ¡p||I¡h;Cl|{;l|{¡>III{¡;¡{jIlC¡4l¡ ;sC¡ l||lC;ss¡ lqIG ¡ ;{G¡ 4l¡yI;t¡ {;;s¡ l||lJ¡ ;>{G;{D ¡;¡Ipy;IG¡p{i¡jI¡jQ¡l||I¡yIj|G

(129) ¡ l{¡ ;¡ l|{¡ Fsl{GI¡ ;;;¡ ;¡ GlaI{¡ IyI;I¡ ]|y¡ (( ¡ |¡ ( ¡ Is;lIs¡ s|¡ I¡ ¡ ¡ 16; ¡ ;{G¡ |iI{¡ î;Cl¡ >cMIG¡ ;|{G¡ jI¡4l4l5¡ Tlsl>¡ ly¡:{GI¡jII¡C|{Gll|{¡H¡'4l<x¡lI

(130) ¡H4l% 4l ¡ ;¡ f{G¡ |¡ >I¡ R;lC;ss¡ llI¡ ;{G¡ |¡ GICI;I¡ lj¡ l{COl{i¡ Iy;I¡ h|y¡ !"¡¡ |¡ )¡¡ ¡H9¡!¡-|{lGIl{i¡ jI¡ yI;s¡;sC¡ Iy

(131) ¡3=; I¡ ;t¡ ¡GIy|{;IG¡j;¡ jI¡g;Cl|{¡ ;l|{¡ C|vG¡ B¡ C;wCs;P¡ ;¡

(132) $ ¡¡ /;|s;l|{¡l{i¡jl¡Il|{¡j|¡ j;¡jI¡W;Cl|{;l|{¡_C|¡;¡C|I¡y;{sI¡>|{G;¡C|{¡ Gl l|{¡ l¡ {Iislio@sI*¡ 4l¡ yI;s¡;sC¡ l¡ ;y . ;¡ $¡

(133) ;{G¡ ;lI¡ >III{¡ $¡ ;{G¡ !$;y/ f¡. \llclR :&&&. 4!F. 0?'%. 0%&&. I&%. G(&. 4">. . 4"9. . . . . %H. . . Pm. . %F. %>. OfeK;))/ Onpi];-;

(134) X}b||oSr2$CPMEzX N<*1@ ZhoTrPo 2A[w,Q =)* J ^s~td#d{aN;+3B Uxqjg`vyk ,6||oV|[zY k ZhoLWxy,_uyk . %:. &. %. %D. 4. . 5"D. . *NL v v 6v1@>p>vNj`k`dNAvd6fkNkN`]vEm^kN`]vf6kN`juvG7=k`fjv`Fv N]v>NHI@g@]kv\O_f6PvjeN@jvfC@q6]kvJ`fvdT6]@kDN\6Ujvk6U=v6jv6 FXYZ[=lN`]vÈvl@\d@f6lnf@v =6U=nU6l@>v9tvlM@v )*3v\@kM`?v *`fv=`\d6fNj`]vuvE8=l`fjvÈv*@v>@k@f\N]@>v9tv/1,52v/oU@6fv1@j`]6]l ,]@U6jkQ=v5vf6tv2=6kk@fN]Lv6f@vjM`r]vKbfv jN\NT6fvjdBN@jv %&Ud @kv6U v "v %v4jXYZ[c6v@kv6U v #v%v.6=Sj`]v@kv6W v !$v %v)6mdM6jv@kv6U v v uvG7=k`fjv N]=f@6j@vrNkMv`sN>6kN`]vjl6l@vÈv/NvGi`\v/N l`v/N 6jvr@UUv6jvGbfv*@vuvG7=k`fj '@jN>@jv6vjO\NU6fv9@M6qN`fv=6]v<v`9j@fq@>vE`fvÙNqN]@v6]>vdtf`s@]@v=`]k6NvaN]LvOf`]v`hv]N=S@Vv/O2N06]>v(6/N2O 06f@vn]>OjkN]LnOjM6:U@ v.

(135) . q. q. %& ( ( q. q. q. . q. . . . . . . q. . . *%,"q*K *Qq. . 10@1Y@0U:MJdG4U0AC;Z:K4d +(d 10@1Y@0U:MJdG4U0A O^QN[4K4d+&d. `b%X?9d5Sd 0D d -%.!d

(136)

(137) G6V0@M@<Z9L6d+&-%."d abG7T0@ O_R\7L8d+'-'.#d cd $]O4P>H4KU0 Bd2/U0ddd)d *0Wd4Td 0A dG4T0@ U0E1 d4[Od.

(138) . . q. d. q. $LF

(139) q q #k[SgfM[Yq[>qeG8qX8e0SqcLSL50e8qMc[e[]8q@b05eL[Y0eL[Yq @25e[aq0eq8`gNSL3aLiqp+LqLYq 0 Xgq q0cq0q>gY5eM[Yq[>qf8X]8a0ega8

(140) q.G8qX8f0U [VLjLY8 ]oa[n8Y8qBb05eL[Y0eL[Yq)0ccq 68]8Y68Yeq@b05eL[Y0eL[Yq)!%q0a8q@\gY6qe[q38qY8F0fLk8q0Y6q 8nea0][S0eL[Yq mNfGq e8X]8a0ega8q cH[mq0q m80Rq 68]8Y7Y5o

(141) q #nâLX8Ye0Sq 60e0q [Yq X8f0Wq e0V5q (0aq 8fq 0S

(142) q q 6[q Y[eq X0f5Gq [gaq eJ[a8eL50Sq 50S5gS0eL[Yc

(143) q [Yl9adSoq [gaq 8n]8aLX8Ye1Sq 60e0q _LYecq[Yq X8e0SqcLSL50f8q X8Seq 0eq 4LFHq e8X]8a0ega8q0Fa;cqmLeHqeH8q 50S5gS0eL[Ycq mMfILYq 8aa[aq 30ac q [YfaL3geM[Yq [>qYg5S81aq E8S6qcHM>fq8C:ecq+$-#q40cq 3;Yq 51S5gV0e86q$hOq8fq0S qq0Y6qMcqY[fqcLFZLAP50Yeq a8S0eNk8qe[qfJq f[e0Sq>a05fL[Y0eL[Y q /L6fIq[>qeHLcq 5gaj8q cH[mcqeH8qk0aL03LSLeoq[?qeHLcq ]a[5<cq@\aq eH8q6LD8a8YeqMc[e[]8cqa0eM[c q. y?gk?p5y{o?u Nqig >i~h yi , -5p ?y 5a *?h:?5yy?gk?p5y{p?uiIg?y5auYaY:5y?>YKP?p?hyY5 yYih Y? 58i}? yX? g?ayYhW kiYhy iI )?/Y 5aai p?a?}5hy iIka5h?y?uYg5au hig?y5auYaY:5y?Nq5:yYih5 yYihiI/YYuiyik?u~i{a>i::{o~YyXYh5yz7Yh58a?5h5ayY:5b kp?:Ywih *i~?}?p ihaz5a:X5u8Eh Yh}?uyYW5y?> ~X\a? yXYugYh?p5ZYuhiyp?kp?vhz7yY}?iIuYaY;y?uV{h>Yhg?y? ipYy?u(ih}?pu?c yX?ktu?hyuy{>V:{vuihiaY}Yh?5h> kpi?h? y~iiIyX?g5_ipuYaY:5y?kX5uGiIka5h?yGYg5au &y

(144) ( i{o ?k?pYg@hyu 5h> :5a:{a5yYihu uXi~ p?uk?:yY}?a /Y1%5g{ 5h> 5g{ }5a{?u yX5y 5o? Yh Wii> 5Wp??g?hy ~YyX yXiv iI -55o 3,=I4U0B U0Fd %5g{ 5yyX? u5g? y?gk?p5y{o? +hy?p?uyYhWa g?y5az5a: hY:`?a Jr5:y[ih5y[ih >?:p?5u?u ~YyX Yh:o?5uYhW y?glp5y{o? ~XYa? yX? g?y5a iaY}YhBkpi?h?[h:pCu?u)YW +yYu5auigir?y?gk?p 5y{o? >?k?h>?hy [h zX? ;? iI yX? g?y5ay5a: uuy?g )YW3XYu>YR?p?h<iI8?X5}Yip9y~??hiaY}Yh?kpi ?h?5h>y5a:Yuhiyyp[}Y5ayiHka5Yh8?:5{v,=d uYy?u5p?uYg Ya5pYhyX?u?kX5u?u yX?iha>YR?th:?8?YhWyX?kp?u?h<iI X>pibWpi{ku ([hy5a:0h??ka5h5yYihYuyX5yyX\uYu p?a5y?>yiyX?YhX?o?hyaYgYyuiIyX?yXt?Yuiyik?ug?yXj[h yX?>?y?pg[h5y[ih iI ?n{YaY8pY|g Yuiyik?Nq5:yYih5yYih Vp. g?y5biY>? io u\aY;y? uuy?gu &u o?}Y?~D r?:?hya (5i5h>'5i

(145) 'i{p>ih?y5a yXYu?k?rYg?h y5a 5kkpi5:Xg5aC>yiYh:ioo?:y F{YaY8pY{gNq5:yYih5 yYih N6:yipu >{? yi u{T:Y?hya 5:={hy?> Vp `Yh?yY: ?KO?:yuYh yX?yXp??[xyik?>Y5Wp5g4 .ip?i}?p yX?{u?iI 5g?y5aaY::5m{a?Wia>YhyXYu:5u?>Yp?:yaYh:ihy5:y~YyX 5 g?y5afY: u5gka? [hypi>{:?u 5 yXYp> o?u?o}iYo yX5y g5 5KQ?:y yX? Yuiyik? Nq5:yYih5yYih iI Yhy?p?uy 3XYu Uh>YhW ~5u Yh>??> 5 g5_ip pG{ay iI yX? uy{> iI -55p ?y 5a yX5y XYWXaYWXy?> Vp dihW o{h >{o5y[ihu `Yh?yY: ?R?:yu :5{u?> 8 u{:X Yhy?o5:yYihu ){pyX?o ?k?o[g?hy5e 5h> yX?ip?y[:5a Yh}?uy[W5yYihu 5o? pF{Yp?> Vp h]:`?a [ui yik?u 5y ai~ y?glp5y{o? y?glp5y{p? p5hW? iI yi:a5pYIyX?p?5uihuVpyX[u>Yu:o?k5h: ( $(#$( "$$( ('( !##(. .5uu>?k?h>?hy`Yh?y^:Yuiyik?Nq5:yYih5yYihYh5g?y5a uYaY:5y?uuy?g:5h8?>{?yi>YKO?sAhykoi<uu?u!?}5ki p5yYih" :X?gY:5b >\S{uYih Yh 5 uYaY:5y? g?ay" Wo5Yh 8i{h>5p >YKL{u\ihYh uiaY> uYaY:5y?u# }ia{g? >YKL{u\ih Yh g?y5au$ yX?pg5a 2ip?y >YKL{uYih Yh uYf^:5y? g?ayu /iy?~ioyX yX?pg5a>YKM{uYihYhg?fyuYug5h ip>?puiI.