Scribd Logo
Upload

Welcome to Scribd!

  • Seismo 1549

    Uploaded by

    Phochan Manomayidthikarn
    10 views18 pages

    Original Title

    Seismo_1549

    Copyright

    © Attribution Non-Commercial (BY-NC)

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    10 views18 pages

    Seismo 1549

    Uploaded by

    Phochan Manomayidthikarn

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    of 18
    JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 96, NO. B7, PAGES 11,891-11,848, JULY 19, 1991 The Equation of State of a Molten Komatiite 1. Shock Wave Compression to 36 GPa Gescony H. Muxea,'? Epwaxp M. Stour, axD Tuomas J. Aurens! Division of Geological and Planeary Sciences, California Intute of Technology, Pasedena "The equation of sate (BOS) ofan int oie (155I°C) annie (27 wt Se MgO) war desermined in the 5-36 GPa pressure range va shock wave Oompresion, Stock seve velony Ud parle velocity Uy (reeds) few te near ebsonhip UJ, = 3126008) + .STE003) Uy. Raed os cee density at 150°C, 0 bar of 2.745#0.005 plant, his U,-U, elatonship gives the isentop bulk modulus Ky = 270 £ 06 GPa, and its fr and seeond itenvopie pesaure dactaves, Kg = 49 & OL snd Ky = “01109 + 0,003 GPa? The ealeuats! heuidus compression cure agreer within et wil the sine ‘compression resis of Agee and Walker (1988) to 6 GPa tt i Tess dese thn tir extrpoitl valee at higher prevres, We deteraio that clivine (Foye) willbe noallybuoyaa: in konaitie melt of the cor potion tat we studied usar &2 GPs. Clinopyoxene would alo be newally buoyant ee this pressure Eiguidus gamet-majocte may be lets dense than thie Komalis>ligud in the 20-24 GPs interval, however, pytopie amet and perovsli phares ar denser th ths kimattie Laud ln their reapestive ligidus pros ‘ve fcevals 0 36 GPx Lighidas perovskite may be neatly bucyet nar 70 GPa, At 40 GPa, the de ty of shock-compresaed malin Kématite would be approxietsl equal to te tlasaed deni of an ‘quivers misture of dads void oxide components, This ebservaion suppor the del of Rgds (0989) for compresibilerof liquid ode componets. Using tel theoretical BOS foe liad forante and fayalite calcul te denier af « spectum of melts frm besa Trough pedottic that ae ese the experimentally stadied Koma guid by ain or rubraction of ebvine. At low presiur, olivine fexctination lowers the density of asi ragmas, bat above 13-14 GPa ts tread is evened. Al of these base io utabasi Ligeds ase predicted to have sina denies at 13-14 GPa, and dhs densy is aprox tnaely eqel to the density of te bulk fpeiminary reference Earth moe) mane Sn thie pressor rage, “Thir oggess hat mel derived from a pardsie mane way be ithibied from ascending eam das rete han 400 km, 1. Iwrropuction ‘The pressure-volume-temperature (PVT) equation of state (EOS) of silcaté melts at pressures of several (ens to hundseds of kilobars has zecenlly become an area of considerable interes ‘This interest stems in part from the recognition thar magmatic sotvity extending over much of the depth of the upper mantle and pechaps even the lower mantle may have played a significant sale it the early evolution and differentiation of the Barth. If so, knowledge of the equation of state woul be of fundumnental importance in setting constsins on the conse: ‘quences of such deep-sesied igneous activity. Tn partic, i thas boen suggesied [Stolper et al, 1981; Nisbet and Walker. 1982; Rigden et al, 1984; Obani, 1985; Takahashi, 1986; Agee ‘and Walker, 1987] that a sufficiently high pressures, silicate ‘oells could become so dense that ususl erystal-melt segregation (aystals. sinking), would be impeded. This could result in a ‘maximum depsh ffom which mels could rise and could have the ‘important consequence of burying beat and incompatible ele ‘ments that might otherwise have reached the Earth's surface. cexueme cases, silicate mets could be denser than surrounding sande rocks. resulting in a downward segregation of melts, Teal, and heat-producing elements, quite the opposite of most ‘oder shallow igneous activity T Riso-w fllen B. Linchuot Labortry of Experimental Geopiysies, ‘Caifornia Intute of Testnology, Pasion 7 Now at Deperunent of Gealogy and Geophys Universiy of Ca fowni, Berksiey ‘Copytigh 1991 by the American Geophysical Union. apes number 915801208 (148.0227 6158-01204805.00 Rigden et al. (1984, 1988) deermined the BOS of a model ‘aslt to 340 Kbar and verified that silicate mel can be of com. parable density to manile minerals and rocks at high pressures. ‘The requisite pressures (26 GPs) ate, however, much greater than those at which basaltic magmas are generally believed to have formed (Green and Ringwood, Y967], Melts generated at such high pressures aze believed 19 be more olivine normative than common basalic magmas. Komatitic liquids are examples of magmas that may originate at the high pressures at which silicate magmas may be comparable in density to olivine-tich residual materials. Recent high-pressure melting experiments [Tatahashi and Scarfe, 1985; Scarfe and Takahashi, 1986] have demonstrated that Tiquids similar to komatitic magmas ean be generated by partial melting of gamet Theeoite ot pressures of 5-7 GPa. At higher pressures, parial melting of garnet Yherzo- Tite produces even more olivine rich magmas, approsching mande-derived lserzolites in composition at =14 GPa [Scarfe ‘and Tekahashi, 1986). Large-scale melting of the mantle may have occurred during accretion [Kaula, 1979; Abe and Matsui, 1986; Mazsul and Abe, 1986; Zale et a 1988; Ahrens, 1990} (or impact formation of the Miaon [Benz ef al, 1986, 1987; Cam ‘ron and Benz, 1989; Stevenson, 1989), pechaps generating & komaliie or peidottic magma ocean extending to great depth. ‘The equations of state of ultrabasic liquids are cvtical for ‘evaluating hypotheses for the generation of Komatiic liquids snd for quantitative evaluation of early magmatic evolution that imay have extended to great depths in the Earth, Extapoimton of the high-pressure EOS of « model basalt [Rigden et of, 1984, 1988) mad simple mineral melts [Rigien et al, 1989] 10 more olivine-normative compositicns involves con- siderable uncertainty. In particular, liquid olivine components are believed to have a relatively large isentropic bulk modulus at 118% (Ks) to 100 kisst but their detailed behavior is unknown (c high Kes and low AKs/@P ls (K's) versus 1ow Ks and high ’qg} and can mtake u big difference in extrapolation of density hove a few tens of kilobars (Rivers, 1985; Herzberg, 1987] Furthermore, aihough J-bar bulk moduli of Basie and ultrabasic liquids can be modeled with reasonable accuracy Lange and Carmichael, 1987; Rivers and Carmichael, 1987] tht extrapo- lation to 100.kbar pressures is complicated by the fact shat the {orm of the EOS eannot be anticipated [Rigden et al, 1988}. Direct high-pressure experimentation on ultrabasic liquids is required, e.g, the static compression experiments of Agee and Watter {1988} Another direct experimental method for the ‘devermination of high-pressure, high-temperature liquid equa- tions of sate is shock wave compression. Shock wave ‘compression fas previously been used to measure the EOS of silicate Liquids in the system anorihite-dippside to pressures of 400 Kar [Rigden et al, 1984, 1988, 1989]. These results are both consistert with avsilable low-pressure compressibility data and extend these data wo lower manile pressures through diect ‘measurement, In this paper, we report de results of a study of, the equation of state of molten Komasite to 360 kbar using shock wave techniques, ‘Ths is the first report of the EOS of a ‘complex, naturally occurring melt composition to these pres ‘The results of the high-pressare komatite EOS are usod as a basis for evaluating models of the pettogentsis of komatitic ‘magmas and the early evolution of the manile in a compaaion paper (Miller et al, this issue), 2, EXPERIMENTAL METHOD Although the molten silicate shock wave experiment was described ¥y Rigden et al. [1988], we have meds some ‘modifications (0 both the experimental procedure and method of data analysis. We therefore will descate the entire experimen: tal and analytical procedure, placing emphasis on those aspects ‘that differ from the earlier work. ‘A synthetic Komatite (glass plus approximately 10% spinifex olivine) was prepared at Corning Technical Cente, Corning, Now York, by G. Fine. The nominal composition (Table 1) was based on ¢ natural komative from the Pike Hill area of Munro ‘Township, Ontario (Walker et al, 1977). Batches of 200g spectoscopic grade oxides were mixed, then melisd at 1650°C for 4 hours in a Fe-saturated Pt crucible in ai. The Liquid was then poured onto a 6 by 6 inch cold steel sab to make « patty TABLE 1, Experimental Konutite Composition Mesaured Aer Heating in Mo Oxide Nominal Measures” Contin 50, 461 4 465 Mad 2 2656 m7 Ro 95 86 8s ceo 14 18 1” ALO, 16 32 80 20 09. "1 a9, TO; 04 04 04 C0, 04 oa 03 Ko 02 03 02 Mao 04 NA NA MoO, = 00. 09 Tou 000 98 to. In weight perce ‘Avorige of approximately 100 andyser obtained at SO4um intervals slog hornocal and vrical anaes ae he sample TR e ealeuated as FO. ‘Mum Bt aL: KomariTe BQuantow of Save of appropriate thickness (© 0.5 em) for the shock wave sample assembly. The liquid could not be quenched rapidly enough 10 prevent growth of 5 ines, but by pouring slabe of similar thickness to our sample assembly we were able to ensore that the bulk composition used in our experiments was similar te the nominal composition Hsted in Table 1 despite the chemi- cal stratification of the slab. Cores of 1-em diameter were cut from the slabs with a water-cooled diamond coring bit. The ‘saagle size was doiermined by a ealeulation of the mass of melt realed to fM 85-90% of the Mo container (ee below) at 1550°C. The measured vokume of the Mo sample container, corrected to 1550°C [Touloukian etal, 1870, and the estntated smelt density at 1550°C (Lange ard Carmichael, 1981 deter: mined the mass of sample needed to fll the conitiner, The 10- 15% underfling was considered nacessary to prevent rupruring or distortion of the container. The cores were ground wo the desired size with 60.pm alumina. ‘Two cores, prepared as described above, were used to verify that the core preparation procedure did not bias the sample com position, Thin sections were cut along the axis of che eylindt eal cores and analyzed with the Calech IEOL, electron microprobe, Vertical and horizontal transects across the se tions were probed at approximately 50-m intervals, Averages ae presented in Table 1, colurm 2. Molybdenum sample containers were machined of high purity sintered Mo stock (grade ABL-2 from Climax Speciety Metals, Coldwater, Michigan) sccocding to the specifications of Rigden 4 al. 988], The inner surfaces ofthe container were polished 0 03 jum with alumina to minimize the possibiliy of bubble adhesion and tp provide « smooth and nominally flat mictaVliquid interface. Some curvature of the Mo. was into: duced by the polishing, however, resulting in a stighily vancave surface; the thickness of dhe Mo driver plate (Figure 1) was found 10 vary by 10-50 yn between the perimeter of the sample well and the center. Other dimensions of the Mo container (measured with #1-pm precision) were the thickness of the driver plate outside of the sample well and the sample wall depth and ameter. The density of the Mo container was doer ‘mined 10 40.02% by the Archimedean method. The measured density agreed with the published density [Straumanis and Shodhan, 1968} within the reported errors, confirming neslgible porosity and impurities, A Mo cover was machined to fit the sample container. The inner surface of the cover was lapped to 03 jum, and it thickness was measured to £1 jum, ‘The lapping produced no measurable curvature, The komatite ores were em Teemeoue ye Oy] al Prjetie Mo cover ee Molten; plate, teats Me iver ron — eo pate Vimp ig. 1. Schomati of shock wave sample assembly Maui ex aL: Kwari Bquaion oF Sram 11,833 placed in the Mo container, tho Mo cover was seated in place, and the cover was electron beam welded at 400°C in vacuo. ‘The heating under vacuum facilitates the eleetron beam welding ‘nd has the aided likely benefit of driving off volatiles adsorbod fon the glass. After welding, the sample asserably was measured to ensure that the cap was seated properly. Canting of the cap ‘was checked by measurement of the thickness of the assembly scross the cap fo #1 pm. The caps were found to be casted by Jess than 5 jum across their 1-cm diameter. ‘The wekied sample assembly was hung vertically in vertical tube furnace at 1500°C for 15.20 min wo verify weld integrity ‘The Tiguidus temperanre of the sample is appyoximataly 1465°, so the sample was completely melted. A continuous flow of dry Nz ges was maintsined in the furnace to minimize oxidation of the Mo. The dimensions of the sample container ‘were measured after heating to verify that the cap was not dis tended from overfiling of the sample container. The loss of thickness owing to Mo loss on oxidation was Iss than 5 pm. ‘One sample was cut open and examined to verify tha the sam- pile had been compleely melied. Before heating, the sample consnted of spnifex olivines in a glass matrix, After heating at 1500°C, all of the spinifex olivines were replaced by uniformly

    You might also like