Experimental Constraints on Silicate Perovskite Forming Reactions and Elastic Properties: Geophysical Implications for Chemical Heterogeneity in the Deep Mantle

Expe imen al Cons ain s on Silica e Pe o ski e Fo ming Reac ions and Elas ic P ope ies: Geophysical Implica ions o Chemical He e ogenei y in he Deep Man le Disse a ion zu E langung des G ades eines Dok o s de Na u wissenscha en -D . e . na .- de Fakul ä ü Biologie, Chemie und Geowissenscha en de Uni e si ä Bay eu h o geleg on Ashima Saikia 2007 Volls ändige Abd uck de on de Fakul ä ü Biologie, Chemie und Geowissenscha en de Uni e si ä Bay eu h genehmig en Disse a ion, zu E langung des akademischen G ades Dok o s de Na u wissenscha en (D . Re . Na .). P ü ungssausschuß: P o . Da id Rubie, Uni e si ä Bay eu h (1. Gu ach e ) P o . F. Langenho s , Uni e si ä Jena (2. Gu ach e ) P o . Jose B eu, Uni e si ä Bay eu h P o . L. Dub o insky, Uni e si ä Bay eu h P o . Klaus Bi ze , Uni e si ä Bay eu h P o . Ludwig Zölle , Uni e si ä Bay eu h P o . S. Pei e , Uni e si ä Bay eu h Da um de Ein eichung de Disse a ion: 5 h Oc obe , 2007 Da um des Wissenscha lichen Kolloquiums 6 h Feb ua y, 2008 2 Acknowledgemen s I eel immense pleasu e in a ailing his oppo uni y o exp essing my deepes sense o g a i ude and ega ds o my supe iso s, D . Dan F os and P o . Da e Rubie o hei dynamic guidance, cons uc i e sugges ions and un i ing e o s ex ended du ing my en i e doc o al esea ch wo k a he Baye isches Geoins i u , Uni e si ä Bay eu h, Ge many. I am acknowledging my obliga ions o he Eli ene zwe k Baye n, In e na ional G adua e School p og am o unding my esea ch p ojec s and he adminis a ion o he Baye isches Geoins i u , o p o iding me necessa y acili ies o pu sue my wo k. Mo e so I am indeb ed o subs an ial help and in aluable sugges ions o D . Tiziana Bo a Balla an wi h whom I ca ied ou my en i e diamond an il cell wo k, o P o . Masaki Akaogi, Uni e si y o Gakushuin, Tokyo (Japan) o allowing me o use his labo a o y acili y o calo ime ic measu emen s, o D . Ca he ine McCammon o helping me ou wi h he Mössbaue measu emen s and analysis, o D . Nobuyoshi Miyajima o assis ances in he TEM s udies and o D . Flo ian Heidelbach o he SEM s udies and o he Ge man ansla ion o my hesis summa y. I o e my special hanks o he en i e echnical and adminis a i e s u o he Baye isches Geoins i u o all he assis ances. Las bu no he leas I hank all who ha e di ec ly and indi ec ly sympa hized wi h me in comple ing my en i e disse a ion. I emain Ashima Saikia Bay eu h, Oc obe 2007 3 4 Table o Con en s Zusammen assung 9 Summa y 13 Chap e 1 In oduc ion 17 1.1 Man le mine alogy: in e ences om pe ological obse a ions, seismology and expe imen s 18 1.2 Mo i a ion 23 1.3 Aims o he s udy 27 Chap e 2 The calcium pe o ski e o ming eac ion in he ansi ion zone o he Ea h’s man le, implica ions o he mid- ansi ion zone seismic discon inui y a 520 km dep h. 31 2.1 In oduc ion 31 2.2 Expe imen al de ails 34 2.2.1 Syn hesis expe imen s 34 2.2.1.1 S a ing composi ion syn hesis 34 2.2.1.2 P essu e calib an syn hesis 35 2.2.2 Mul ian il expe imen al s udy 36 2.2.2.1 Mul ian il echnique 36 2.2.2.2 Mul ian il expe imen s 38 2.3 Cha ac e iza ion and analy ical echniques employed in his s udy 42 2.4 Resul s 49 2.4.1 The modynamic modeling 55 5 2.5 Discussion 58 2.6 Conclusions 65 Chap e 3 An equa ion o s a e s udy o (Fe,Al)-bea ing magnesium silica e pe o ski e single c ys als: implica ions o lowe man le p ope ies. 67 3.1 In oduc ion 67 3.2 Expe imen al de ails 71 3.2.1 S a ing ma e ials 71 3.2.2 Mul ian il syn hesis expe imen s 72 3.2.3 Quan i ica ion o pe o ski e Fe3+/∑Fe a ios 75 3.2.4 C ys al composi ions 78 3.3 Diamond an il cell comp ession expe imen s 80 3.3.1 Basics o diamond an il cell single c ys al comp ession echnique 80 3.3.2 Comp ession expe imen s 81 3.4 Equa ion o s a e esul s 84 3.4.1 Theo e ical backg ound 84 3.4.2 Uni cell la ice pa ame e a ia ion wi h p essu e 85 3.4.3 FE- E plo and EoS pa ame e s 87 3.4.4 Oc ahed al il ing 94 3.5 Discussion 96 3.5.1 The elas ici y o he Ea h’s lowe man le 96 3.5.2 The e ec o p essu e on pe o ski e subs i u ion 102 OAlFe VIVIII 3 33 ++ 3.6 Conclusions 105 Chap e 4 A calo ime ic s udy o he Mg3(Mg,Si)Si3O12(majo i e)- Mg3Al2Si3O12(py ope) ga ne solid solu ion 107 4.1 In oduc ion 107 4.2 S a ing ma e ial syn hesis 109 4.2.1 Pis on-cylinde syn hesis expe imen 110 6 4.2.2 Mul ian il syn hesis expe imen 111 4.3 Calo ime ic measu emen s 114 4.3.1 Basic p incipals 114 4.3.2 En halpy measu emen s 115 4.4 Resul s 119 4.5 Discussion 122 4.6 Conclusions 125 5 Conclusions 127 Appendix A 131 Appendix B 139 Re e ences 141 E klä ung 161 7 8 Zusammen assung Im Rahmen de o liegenden Dok o a bei wu den d ei expe imen elle Un e suchungen du chge üh , um zu e s ehen, wie die Zusammense zung des E dman els die S abili ä on Mine alen und ih e elas ische Eigenscha en bes imm , und wie diese wiede um die seismischen Eigenscha en des ie en E dman els beein lussen. Die Phasenbeziehungen on Calciumsilika -Pe owski wu den in Hochd uck-Hoch empe a u expe imen en un e such , um den E ek seine Bildung au die Cha ak e is ika de seismischen Diskon inui ä bei 520 km Tie e im E dman el zu bes immen. De E ek on a iable Zuammense zung au die Komp essibili ä on Magnesiumsilika -Pe owski wu de un e such , um die geophysikalischen Konsequenzen chemische He e ogeni ä im E dman el zu e s ehen. Kalo ime ische Messungen de G ana misch eihe Py op-Majo i wu den du chge üh , um g undsä zliche he modynamische Da en ü die Modellie ung de Bildungs eak ionen on Magnesium- und Calciumsilika -Pe owski be ei zus ellen. (i) Die Calciumsilika -Pe owski bildende Reak ion in de Übe gangszone des E dman els: Implika ionen ü seismische Diskon inui ä in de mi le en Übe gangszone bei 520 km Tie e. Globale seismische Beobach ungen zeigen, dass das Au e en de seismischen Diskon inui ä bei 520 km Tie e in de mi le en Man elübe gangszone ein komplexes E scheinungsbild ha . In einigen Regionen des E dman els e schein diese Diskon inui ä in zwei Diskon inui ä en in leich un e schiedlichen Tie en au gespal en zu sein. Es is dahe o geschlagen wo den, dass un e den Bedingungen de mi le en Übe gangszone ausse dem Phasenübe gang on 9 16 The da a also show ha he Mg3(Mg,Si)Si3O12(majo i e)-Mg3Al2Si3O12(py ope) solid solu ion has s ong non-ideal mixing p ope ies. Chap e 1 In oduc ion Knowledge o he physical and chemical s a e o he deep in e io o he Ea h is he key o unde s anding i s e olu ion and dynamics. Due o limi a ions in di ec sampling o he deep Ea h, mos in e ences abou i s in e io mus be based on indi ec in o ma ion ob ained h ough geophysical and geochemical obse a ions. Expe imen al da a on mine al p ope ies a deep man le condi ions become c ucial, pa icula ly in in e p e ing geophysical obse a ions and assessing hei implica ions o man le geochemis y and dynamics. The man le ep esen s he la ges geochemical ese oi o he silica e Ea h, comp ising abou 80% o olume o he o al Ea h. Ou main in o ma ion on he composi ion o he man le comes om xenoli hs, pieces o he man le b ough o he su ace o he Ea h by olcanic ac i i y, and pe ido i e massi s, sec ions o he man le emplaced in he c us by ec onic mo emen s. In addi ion, he geochemis y o olcanic ocks p oduced a mid-oceanic idges by pa ial mel ing o he man le aids ou unde s anding o man le composi ion and a compa ison wi h undi e en ia ed me eo i e samples helps o cons ain he likely composi ion o he Ea h as a whole. The cons ancy in composi ion o e up ed basal ic magmas wi h ime in e s ha he bulk composi ion o he man le may be ela i ely uni o m as expec ed o a con ec ing mixed ese oi . Aside om a ew inclusions in diamonds ha may ha e a deepe o igin, samples om he man le come om dep hs o no mo e han 200 km. So, in e ences abou deep man le s uc u e need o be based on geophysical obse a ions. Geophysical obse a ions o he man le ha e iden i ied a numbe o globally-obse ed adial discon inui ies ha e lec seismic wa es. These discon inui ies a e caused by sha p changes in man le elas ic p ope ies and densi y. Such 17 Chap e 1: Man le mine alogy obse a ions ha e led o he di ision o he man le in o an uppe man le, a ansi ion zone and a lowe man le, each sepa a ed by seismic discon inui ies a a pa icula dep h. The discon inui ies may esul om mine al phase ansi ions in an isochemical man le bu changes in he chemical composi ion o he man le wi h dep h would also be a possible explana ion. As di e en possibili ies exis , a ange o geophysical obse a ions need o be examined o de e mine he mos likely explana ion. Excep o s ongly incompa ible ace elemen s ha a e concen a ed in he c us and side ophile and chalcophile elemen s ha pa i ion in o he co e, he man le is he majo eposi o y o mos elemen s in he Ea h. An unde s anding o man le composi ion and i s composi ional s uc u e has, he e o e, g ea implica ions o he chemical composi ion o he en i e Ea h. In addi ion geophysical obse a ions con ain in o ma ion on he empe a u e o he man le ha can be ex ac ed h ough compa ison wi h expe imen ally de e mined measu emen s. Such in o ma ion is c ucial o unde s anding he dynamics and he mal e olu ion o he Ea h. 1.1. Man le mine alogy: in e ences om pe ological obse a ions, seismology and expe imen s Xenoli hs b ough o he su ace by kimbe li e and alkali basal magmas in addi ion o abyssal and massi pe ido i es emplaced in he c us , show ha he uppe pa o he man le is gene ally an ul ama ic ock e med pe ido i e, which is comp ised o he mine als oli ine, clinopy oxene, o hopy oxene and an Al-bea ing phase. A dep hs shallowe han 70 km, his Al bea ing phase is plagioclase, bu i ans o ms o spinel and hen ga ne wi h inc easing dep h. The deepes xenoli hs, howe e , gene ally become en ained om dep hs no deepe han ~200 km. Geophysical echniques, p ima ily seismology, hen become he only me hod o in es iga ing he deepe man le. Cons ain s also come om obse a ions o he momen o ine ia, g a i y, hea low o he Ea h and elec ical conduc i i y measu emen s. These geophyiscal obse a ions, howe e , equi e in e p e a ion using expe imen ally de e mined mine al p ope ies and phase equilib ia de e mina ions. The gene al app oach adop ed o deciphe ing he mine alogy o he man le is o assume ha he deep man le is o a simila composi ion o he uppe man le as in e ed 18 Chap e 1: Man le mine alogy om pe ido i e samples and o expe imen ally de e mine he mine alogy o his bulk composi ion as a unc ion o dep h along an adiaba ic he mal g adien . Based on such mine alogical models sound eloci ies a e hen calcula ed using a ailable he moelas ic pa ame e s and he esul s a e compa ed wi h seismic e e ence models. By examining how well seismic eloci ies in he man le ma ch he model de e mina ions, he assump ions made conce ning he chemis y and empe a u e o he man le as a unc ion o dep h can be assessed. E en hough he e a e limi a ions in he exis ing mine al physics da a ha inhibi a obus in e p e a ion o seismic da a, wi hin he unce ain ies o he cu en da a a easonable ma ch is ound be ween seismic obse a ions o S and P wa e eloci ies wi h dep h and models cons uc ed o a pe ido i ic man le composi ion (Camma ano e al., 2005). Howe e , seismic omog aphy has shown ha he e ogenei ies o he mal and chemical o igin, mos likely a ising om he p esence o subduc ed li hosphe ic slabs, exis in he man le. In addi ion obse a ions o seismic he e ogenei y in he lowe man le do no seem o be well explained by he mal a ia ions alone (T ampe e al., 2004). Some obse a ions ha e also been in e p e ed o esul om he p esence o pa ial mel ing in he deep man le (Williams and Ga ne o, 1996; Ka a o and Jung, 1998). A p essu es > 3 GPa such mel ing is likely o only a ise i he man le is sui ably en iched in ola ile elemen s like H and C a hese dep hs, which also aises ques ions as o he deg ee o chemical homogenei y ha can be expec ed in he deep man le. Chemical composi ion o he man le Va ious app oaches ha e been adop ed o in e he a e age chemical composi ion o he man le. The simples app oach is o assume ha he composi ion o he man le is he same as some p is ine pe ido i e samples e.g., KLB1 e c. eco e ed a he Ea h’s su ace. Ringwood, (1975) p oposed a chemical composi ion based on ecombining basal ic pa ial mel wi h he e ac o y man le esidue le behind. A mid-oceanic idges basal s a e gene a ed by app oxima ely 10% pa ial mel ing o pe ido i e man le. Ringwood a gued ha he composi ion o he p is ine man le, ha he e med py oli e, would be he same as a mix u e o pe ido i e mel esidue and p imi i e mid-oceanic idge basal (MORB) composi ions. Ra he han assuming a pa icula esidual and basal composi ion, as Ringwood did, an al e na i e app oach is o examine mel ex ac ion ends in pe ido i e 19 Chap e 1: Man le mine alogy xenoli hs and massi s and ex apola e hese ends back o an unmel ed p ecu so (Wal e , 2004). Models also exis based on chond i ic me eo i e composi ions (Alleg e e al., 1995; Ja oy, 1995). These models assume ha he bulk Ea h o med om chond i e me eo i e ma e ial bu ha side ophile elemen s we e ex ac ed om his composi ion o he co e lea ing mos majo elemen s in he man le in chond i ic p opo ions. Because all undi e en ia ed me eo i es ha e highe Si/Mg a ios han he uppe man le, such models equi e an addi ional ese oi ich in Si o o m. Usually i is p oposed ha Si was ex ac ed ei he o he co e o o he lowe man le. Such models a e o en ci ed as e idence ha he lowe man le may ha e a chemical composi ion di e en om ha o he uppe man le. Seismic obse a ions o he man le Vi al cons ain s can be placed on man le mine alogy om seismological e e ence models (T a el ime- ables o Je e ys and Bullen, 1940; 1066A, 1066B models o Gilbe and Dziewonski, 1975; PREM P elimina y Re e ence Ea h Model o Dziewonski and Ande son, 1981; IASP 91 model o Kenne and Engdahl, 1991; AK135, Kenne e al., 1995; Camma ano e al., 2005). Seismic e e ence models such as he P elimina y Re e ence Ea h Model (Dziewonski and Ande son, 1981) p o ide a adially symme ic eloci y dep h p o ile o he Ea h based on he in e sion o body wa e a el ime da a and ee oscilla ions o he Ea h. In hese models eloci ies in he Ea h’s in e io a e e ined o a se o polynomial unc ions ha ope a e o e a speci ied dep h in e al, wi h he assump ion ha man le discon inui ies occu a p ede e mined dep hs. These discon inui y dep hs ha e been de e mined by analyzing seismic wa es ha a e e ac ed by he discon inui y o om seismic wa es ha a e con e ed om S o P wa es, o isa e sa, a he discon inui y. In o ma ion on ampli ude o hese discon inui y jumps and dep h in e al o eloci y change ac oss he discon inui y can also be used o place much needed cons ain s on man le chemical s uc u e (S ix ude, 1997). Py oli e mine alogy as a unc ion o dep h The a ia ion in he p opo ion o mine als ha would c ys allize om a py oli ic bulk composi ion as a unc ion o dep h is show in Fig. 1.1 o he op 1000 km o he man le 20 Chap e 1: Man le mine alogy (Ringwood, 1991). Fo pu poses o discussion his diag am can be con enien ly di ided in o wo componen s, which occupy each side o he diag am. On he le side is oli ine and he highe p essu e oli ine polymo phs, while he igh side is composed o he non- oli ine Si and Al- ich phases o he man le. Phase ans o ma ions in oli ine occu wi h inc easing p essu e o e e y na ow dep h in e als, whe eas he phase ansi ions in non- oli ine phases a e g adual and occu o e b oad dep h in e als. A a dep h o 410 km, (Mg,Fe)2SiO4 oli ine ans o ms o he high-p essu e polymo ph o wadsleyi e β- (Mg,Fe)2SiO4 which is gene ally conside ed o cause he 410 seismic discon inui y ha is globally obse ed a his dep h. The 410 km seismic discon inui y ma ks he op o he ansi ion zone egion o he man le, so called because a ansi ion in seismic eloci y g adien occu s wi hin his egion. A a ound 17.5 GPa co esponding o a dep h o 520 km, wadsleyi e unde goes an iso-chemical phase ansi ion o ingwoodi e, which likely causes a weak seismic discon inui y obse ed egionally a app oxima ely his dep h. The bo om o he ansi ion zone and op o he lowe man le occu s a 660 km dep h, (app oxima ely 24 GPa), whe e ingwoodi e b eaks down o an assemblage o (Mg,Fe)(Si,Al)O3 wi h he pe o ski e s uc u e and (Mg,Fe)O magnesiowüs i e. This ans o ma ion causes a s ong globally obse ed seismic discon inui y a his dep h. Abo e 3 GPa he non-oli ine phases in a py oli e composi ion a e o hopy oxene and clinopy oxene and he Al- ich phase ga ne . Wi h inc easing p essu e, bo h o hopy oxene and clinopy oxene componen s s a o pa i ion in o ga ne . This esul s om he subs i u ion o Mg, Fe and Si on o he ga ne Al oc ahed al posi ion o c ea e a componen wi h py oxene s oichiome y called majo i e {i.e., (Mg,Fe)4Si4O12}. By mid- ansi ion zone condi ions py oxenes ha e comple ely dissol ed in o he ga ne s uc u e wi h ga ne ha ing he app oxima e s oichiome y (Mg,Fe,Ca)3(Mg,Al,Si)2Si3O12. A mid- ansi ion zone condi ions o app oxima ely 18 GPa, CaSiO3 s a s o exsol e o ming he 21 Chap e 1: Man le mine alogy Wadsle y i e (M g ,Fe) SiO 24 Uppe Man le Ma g nesium silica e Pe o ski e (M g ,Fe,Al)(Si,Al)O 3 Lowe Man le T ansi ion Zone Fe ope iclase (Mg,Fe)O Rin g woodi e (Mg,Fe) SiO 24 Ga ne (Mg,Fe,Ca) Al Si O 32312 Oli ine (Mg,Fe) SiO 24 O hopy oxene (Mg,Fe)SiO 3 Ma j o i ic g a ne (Mg,Fe,Ca) (Mg,Si,Al) Si O 32312 Clinopy oxene (Mg,Ca)SiO 3 00.2 0.4 0.6 0.8 1. 0 Calcium Silica e pe o ski e CaSiO 3 1000 800 600 400 200 D ep h ( k m ) Figu e 1.1 Py oli e man le mine alogy as a unc ion o mine al olume ac ion and dep h a ia ion. The small o ange and pink egion in uppe igh hand co ne ep esen s he s abili y ield o eldspa and spinel espec i ely. Mine als eco e ed om high p essu e-high empe a u e expe imen s a e shown in he inse s. Field o iew o inse is ~200 mic ons. (Wi h pe mission om D . D. J. F os ) sepa a e phase calcium silica e pe o ski e. A he op o he lowe man le he emaining ga ne s a s o dissol e in o magnesium silica e pe o ski e. By dep hs o app oxima ely 750 km in he lowe man le, a py oli e composi ion assemblage comp ises magnesium silica e pe o ski e, magnesiowüs i e and calcium silica e pe o ski e. This assemblage is belie ed o be s able h oughou he bulk o he lowe man le and only a p essu es co esponding o he D’’ laye on op o he co e man le bounda y does magnesium silica e pe o ski e ans o m o a pos pe o ski e polymo ph wi h he s uc u e o CaI O3 (Mu akami e al., 2004). 22 Chap e 1: Mo i a ion Man le He e ogenei y Though he majo elemen composi ion o he uppe man le appea s o ha e emained uni o m o e eco ded geologic ime, ace elemen and iso opic s udies (Sun and McDonough, 1989) imply he p esence o signi ican he e ogenei ies in he basal sou ce egion. This may indica e he compa a i ely mo e mobile na u e o incompa ible ace elemen s in compa ison o majo elemen s o i may indica e he e ogenei ies being c ea ed by he p esence o subduc ed oceanic li hosphe e esiding in he man le (Ch is ensen and Ho maan, 1994). I is qui e likely ha a subduc ed slab would ake a signi ican leng h o ime o be ehomogenized by con ec i e s i ing in he man le (Holzap el e al., 2005). I is, he e o e, qui e possible ha la ge egions o he man le a e comp ised o mechanical mix u es o mel -deple ed pe ido i e and subduc ed oceanic c us on a a ie y o leng h scales. This will ha e signi ican implica ions o he s uc u e and composi ion o he man le in addi ion o i s po en ial e ec on geophysical obse a ions. 1.2 Mo i a ion Changes in man le mine alogy as a esul o phase ans o ma ions ha occu o e ela i ely sho dep h in e als ha e been ound o coincide wi h majo seismic discon inui ies in he Ea h’s man le as shown in Fig. 1.2. A signi ican numbe o expe imen al s udies ha e been de o ed o de e mining he phase ela ions and physical p ope ies o mine al phases associa ed wi h majo seismic discon inui ies such as hose a 410 km (Ka su a and I o, 1989; I i une and Isshiki, 1998; Kie e e al., 2001; F os , 2003; Li e al., 1998) and 660 km (I o and Takahashi, 1989; Shim e al., 2001; Li and Li, 2003). O key in e es is o unde s and how empe a u e and chemical a ia ions in he man le may a ec seismically obse able phenomena such as he dep h, dep h in e al and ampli ude o he discon inui ies in addi ion o he ambien sound eloci y and densi y o he man le. Fo his expe imen al da a a e equi ed on he in luence o a iable man le chemis y on phase s abili ies and elas ic p ope ies. 23 Chap e 1: Mo i a ion 12 14 16 18 20 22 24 26 1200 1300 1400 1500 1600 1700 1800 P essu e (GPa) Te m pe a u e (° C ) 450 500 550 600 650 700400 Dep h (km) Oli ine Wadsleyi e Ringwoodi e Pe o ski e + Fe ope iclase [1] [2] [4] [5] [3] Figu e 1.2. P essu e- empe a u e slopes o phase ans o ma ions in he Ea h’s man le compa ed wi h a e age seismic discon inui y dep hs (solid e ical lines) o he 410 km (in g een), 520 km (in blue) and 660 km (in ed) and global opog aphy ( e ical shaded egions). Double cu es ma k he beginning and end o di a ian egions o he oli ine o wadslyei e and wadsleyi e o ingwoodi e ans o ma ions, and he shaded a ea in he cu es show he empe a u e anges compa ible wi h globally-obse ed opog aphy o hese discon inui ies. Indi idual cu es o he Mg2SiO4 ingwoodi e o pe o ski e + e ope icalse eac ion a e shown om di e en s udies [1] I i une e al., (1998) [2] Ka su a e al., (2003) [3] Fei e al., 2004 [4] I o and Takahashi, (1989) [5] Shim e al., (2001). (Wi h pe mission om D . D. J. F os ) The oli ine o wadsleyi e ans o ma ion ha causes he 410 km discon inui y occu s in he MgO-FeO-SiO2 sys em and is he e o e insensi i e o chemical a ia ions apa om he Fe/Mg a io, al hough la ge concen a ions o H2O may also ha e an e ec (Wood 1995; Smy h and F os , 2002; F os and Doleĵs, 2007). The 660 km discon inui y, on he o he hand, is likely o be mo e a ec ed by a ia ions in chemis y because, in addi ion o he MgO-FeO-SiO2 sys em, pe o ski e will also be in luenced by he Al2O3 con en o he man le and he Fe2O3 con en . Recen expe imen al s udies ha e shown ha Fe2O3 has a s ong a ini y o magnesium silica e pe o ski e, which is s abilized by Al in he s uc u e as a esul o a coupled subs i u ion. Such a subs i u ion mechanism can also 24 Chap e 1: Mo i a ion a ec he elas ic p ope ies o pe o ski e. F om his analysis we can expec ha he o ma ion o magnesium silica e pe o ski e in he Ea h’s man le is likely o be a complex p ocess bo h physically and chemically. In addi ion, howe e , he e is e idence ha seismic he e ogenei y in he deep lowe man le may exis and ha he obse ed a ia ions a e no well co ela ed wi h changes in empe a u e (T ampe e al., 2004). To unde s and and in e p e hese obse a ions we equi e in o ma ion on how chemis y may in luence he elas ic p ope ies o majo lowe man le mine als. In addi ion o majo seismic discon inui ies, mino weak seismic discon inui ies a e also known o exis in he Ea h’s in e io such as he Hales discon inui y a abou 60-90 km dep h, possibly caused by MgAl2O4 spinel ans o ming o ga ne and he Lehmann discon inui y a 220 km dep h ha may ma k a change in man le aniso opy. Ano he weak discon inui y was epo ed by Shea e , (1990), om he mid- ansi ion zone a 520 km dep h. This has been conside ed o be caused by he wadsleyi e o ingwoodi e ansi ion. Howe e , ecen indings o Deuss and Woodhouse, (2001) show ha his discon inui y is spli a some loca ions in he man le in o wo dis inc discon inui ies one close o 500 km and he o he a app oxima ely 560 km. The mos likely explana ion o he occu ence o a spli in he 520 is ha i esul s om he o ma ion o calcium silica e pe o ski e (CaSiO3) om majo i ic ga ne , which also occu s in his dep h ange. P e iously he calcium silica e pe o ski e o ming eac ion had been poo ly s udied in p essu e- empe a u e space and i was ha d o cons ain he exac p essu e and empe a u e ange o e which i occu s. S udying his eac ion is u he complica ed by he ac ha majo i ic ga ne om which CaSiO3 pe o ski e exsol es is a mul i-componen solid solu ion and as such he exsolu ion eac ion is likely o be dependen on a signi ican numbe o a iables. A numbe o issues ha e emained unsol ed abou pe o ski e- o ming eac ions in he Ea h’s man le, o which labo a o y s udies can p o ide subs an ial in o ma ion. Fo ins ance, (i) The p essu e, empe a u e and composi ional e ec s on he o ma ion o calcium silica e pe o ski e om majo i e ga ne ha e emained unadd essed. This exsolu ion eac ion could possibly cause a seismic discon inui y bu he exis ing da a a e insu icien o e i y his. 25 Chap e 2: In oduc ion o ma ion o calcium silica e pe o ski e he e o e is clea ly dependen on he condi ions whe e ga ne becomes sa u a ed in CaO. Th oughou he lowe man le calcium silica e pe o ski e is he dominan CaO bea ing mine al. The o ma ion o calcium silica e pe o ski e om ga ne in na u al sys ems wi h pe ido i ic and basal ic composi ions occu s a app oxima ely he same p essu e inspi e o di e ences in bulk CaO con en , wi h basal ic composi ions con aining double he amoun o CaO han con ained in a pe ido i ic composi ion (I i une and Ringwood, 1993; Nisiha a and Takahashi, 2001). This clea ly implies ha he calcium solubili y in ga ne is dependen on some a iable also o he han p essu e. The mos likely explana ion could be ha he solubili y o CaO in ga ne is also dependen on he Al/Si a io o ga ne i.e., he p opo ion o he majo i e componen in ga ne . The o ma ion o he dense calcium silica e pe o ski e mine al in he Ea h’s in e io could cause discon inui ies in he speed o sound wa es as hey pass h ough he in e io , which is de ec able a he Ea h’s su ace. Recen seismic obse a ions o he ansi ion zone ha e iden i ied a discon inui y a he mid- ansi ion zone dep h o 520 km o en designa ed as 520d (520 km seismic discon inui y) (Shea e , 1990; Shea e , 1996; Shea e , 2000; Deuss and Woodhouse, 2001; Gilbe e al., 2003; an de Meijde e al., 2005; Deuss e al., 2006). This discon inui y is ound o be spli in o wo discon inui ies in some egions o he man le, one a an app oxima ely 500 km dep h and ano he a a deepe dep h o 560 km (Deuss and Woodhouse, 2001). The wadsleyi e (β) o ingwoodi e (γ) ansi ion is o en implica ed o be he cause o he 520d (Weidne and Wang, 2000). As he exsolu ion o calcium silica e pe o ski e om majo i ic ga ne occu s a a simila man le dep h, i could also cause a discon inui y and hus esul in a double o spli 520d (I a and S ix ude, 1992). The a iabili y in he dep h o hese discon inui ies could be a s ong unc ion o change in ei he empe a u e o composi ion be ween di e en egions o he man le. The mos likely majo ype o chemical a ia ion in he silica e man le a ises due o ac iona ion o oceanic c us and li hosphe e a he mid-oceanic idges h ough pa ial mel ing. The majo elemen s ha a e ac iona ed du ing pa ial mel ing a e Al, Ca, Si, and Na, which become concen a ed in he oceanic c us and co espondingly deple ed in he li hosphe e (Wal e , 2003). These componen s a e e en ually ecycled back o he man le by subduc ion o oceanic c us . Ongoing con ec i e s i ing in he man le may homogenize hese chemically dis inc domains, al hough his may ake a signi ican pe iod o ime o 32 Chap e 2: In oduc ion homogeniza ion o achie e local chemical equilib ium ( an Keken, 2002; Holzap el, 2005). Howe e , i hese chemically-dis inc domains end o accumula e due o densi y o heological con as s, i may led o o ma ion o long- e m he e ogenei ies in he man le ha a e esis an o homogeniza ion (Tackeley e al., 1993; Ch is ensen and Ho mann, 1994; Hel ich and Wood, 2001). The majo global seismic discon inui ies a 410 km and 660 km dep hs in he Ea h’s man le co espond o phase ans o ma ion in ol ing oli ine {α-(Mg,Fe)2SiO4} o i s high p essu e polymo ph o wadsleyi e {β-(Mg,Fe)2SiO4} and ingwoodi e {γ- (Mg,Fe)2SiO4} b eaking down o magnesiowüs i e and magnesium silica e pe o ski e espec i ely (Hell ich and Wood, 2001). As hese phases do no in ol e componen s like Ca and Al, which a e signi ican ly ac iona ed in he man le, discon inui ies a ising due o hese phase ans o ma ions ell us li le abou likely chemical a ia ions in he man le. Howe e , a seismic discon inui y like ha o 520d which may as well a ise due o exsolu ion o calcium pe o ski e om majo i e ga ne in ol ing componen s like Ca and Al, would be e y sensi i e o la ge scale man le chemical he e ogenei ies; such as hose ha migh esul om he p esence o signi ican p opo ions o emnan subduc ed oceanic c us in he man le as he Ca and Al componen a e signi ican ly ac iona ed a mid- oceanic idges. As majo i e ga ne is a mul i-componen solid solu ion, he exsolu ion eac ion o calcium pe o ski e om majo i e ga ne is pe cei ed o be qui e complex. Exis ing da a (I i une and Ringwood, 1993; Gaspa ik, 1996; Nishiha a and Takahashi, 2001; Li aso and Oh ani, 2005) on his eac ion a e no su icien ly consis en o modeling his eac ion o e he ange o p essu e, empe a u e and bulk composi ion ele an o he man le. Mo eo e , he p essu e in e al be ween expe imen s in exis ing s udies is no na ow enough o accu a ely desc ibe he shape o he eme ging calcium pe o ski e s abili y ield. The exsolu ion o calcium silica e pe o ski e o a ange o man le composi ions was s udied in he p esen wo k using high-p essu e and high- empe a u e mul ian il expe imen s. The solubili y o CaSiO3 in ga ne was measu ed as a unc ion o ga ne majo i e con en be ween 17 and 23 GPa in he empe a u e ange 1200-1600°C in o de o asce ain wi h high p ecision, how he dep h in e al o his eac ion compa es o he wadsleyi e o ingwoodi e ansi ion and whe he his eac ion could occu o e a na ow enough p essu e in e al o cause an obse able seismic discon inui y. 33 Chap e 2: Expe imen al de ails 2.2. Expe imen al De ails 2.2.1 Syn hesis expe imen s 2.2.1.1 S a ing composi ion syn hesis When we conside majo ypes o man le ocks such as a pe ido i e o basal , he majo ype o chemical a ia ion in ga ne s o hese ocks a ises due o he subs i u ion o he majo i e componen i.e., (Si, Mg) becoming inco po a ed in o he oc ahed al Al posi ion in he ga ne s uc u e. Fo his eason we choose ou Ca- ee ga ne composi ions on he (Mg,Fe)4SiO12(majo i e) - (Mg,Fe)3Al2Si3O12(py ope) join o s a ing ma e ials (Table 2.1). As can be seen in he Fig. 2.1, a no mal man le pe ido i ic ga ne will be mo e majo i ic han a ga ne om a subduc ed basal ic composi ion, which has been ecycled back o he man le. Glasses we e syn hesized ou o eagen g ade oxide mix u es o SiO2, Al2O3, Fe2O3 and MgO by using hem a 1600°C in a 1-a mosphe e u nace ollowed by apid quenching in wa e . Glasses we e analyzed o chemical composi ion by elec on mic op obe analysis using a poin beam ope a ing in wa eleng h dispe si e mode a 15nA and 15kV. Quench eco e ed glasses we e g ound o powde s and in o de o educe Fe3+ con en o he glasses o Fe2+, he glass powde s we e educed in a CO2/H2 gas mixing u nace, a a gas composi ion o 04/06 H2/CO2 (an app oxima e O2 o 2 log uni s abo e i on-wüs i e bu e ) a 650°C o a day. Mössbaue analysis was pe o med on he educed glass powde s ha con i med he absence o Fe3+ in he glass powde s. (De ails on cha ac e iza ion echniques a e discussed in nex sec ion). Table 2.1: Chemical composi ion o he s a ing oxide mixes o ga ne glass syn hesis based on 12 oxygens pe o mula uni . Composi ion SiO2 Al2O3 FeO MgO To al Si Al Fe Mg ∑Ca ions Pe ido i e 53.0 9.2 5.3 33.5 100 3.63 0.74 0.3 3.33 8 Basal 48.3 17.4 4.8 29.5 100 3.3 1.4 0.274 3.026 8 Py ope 43.9 24.9 4.4 26.8 100 3 2 0.249 2.751 8 Majo i e 54.8 6.2 5.5 33.5 100 3.75 0.5 0.311 3.439 8 34 Chap e 2: Expe imen al de ails 0.8 0.85 0.9 0.95 1 0 0.2 0.4 0.6 0.8 1 Mg/(Mg+Fe) Ex eme majo i e Pe ido i e ga ne Basal ic ga ne Py ope Mg4Si4O12 Mg3Al2Si3O12 Cubic Te agonal Al2O3 mol% Figu e 2.1: The glass s a ing composi ions used in he p esen s udy a e shown plo ed on he majo i e–py ope solid solu ion join as a unc ion o Mg numbe . In o de o measu e he solubili y o Ca in he ga ne a high p essu e hese glass powde s we e sa u a ed in he CaSiO3 componen by adding wollas oni e (CaSiO3). The wollas oni e phase was syn hesized om an oxide mix u e o calcium ca bona e (CaCO3) and silica (SiO2) powde . The CaCO3 was i s deca bona ed o 24 h s a 225°C in a 1- a mosphe e u nace. The oxide mix u e o deca bona ed CaCO3 and SiO2 was han placed in a 1-a mosphe e u nace o 16 ½ hou s a 1000°C. This u nace empe a u e was scheduled o hea slowly a a a e o 2°C pe minu e, which ensu ed slow elease o any emaining CO2 om he sample. These samples we e eg ound and again used in a 1- a mosphe e u nace a 1300°C in successi e s ages o e 48 hou s du a ion in o al o inally c ys allise wollas oni e. Powde X- ay di ac ion was employed o con i m he syn hesis o wollas oni e. 2.2.1.2 P essu e calib an syn hesis P ecise p essu e de e mina ion is o u mos impo ance in ou s udy. The e o e, in addi ion o using he no mal oil p essu e calib a ion o ou mul ian il expe imen s, we also used in-si u p essu e calib a ion by including an oli ine sample in all ou expe imen al uns. Oli ine composi ions we e syn hesized om s oichiome ic eagen g ade oxide mixes o SiO2, MgO and Fe2O3. A e ho ough g inding o oxide mix u e, pelle s we e made ou o 35 Chap e 2: Expe imen al de ails hese mix u es, ha we e educed in a CO2/CO gas-mixing u nace a an O2 o 14.1, log uni s a 1000-1200°C depending on he Fe concen a ion o he samples. A e h ee successi e s eps, each o 24 hou s o educ ion and ehomogeniza ion by c ushing and g inding, inally homogeneous oli ine c ys allized. Powde x- ay di ac ion on he p oduc phases mixed wi h Si powde as an in e nal s anda d was used o phase iden i ica ion. (De ails o he cha ac e iza ion echniques men ioned he ein a e discussed in sec ion 2.3.) 2.2.2 Mul ian il expe imen al s udy 2.2.2.1 Mul ian il echnique O e he pas 20 yea s use o mul ian il appa a us o high-p essu e, high- empe a u e expe imen s simula ing man le condi ions ha e inc eased mani old and a numbe o publica ions ha e e iewed his echnique and i s applica ions (Kawai and Endo, 1970; Walke e al., 1990; Rubie, 1993; Rubie e al., 1993; I i une, 2002; F os e al., 2004; Kepple and F os , 2005). A mul ian il p ess wo ks on he concep o educ ion o a ea (A) by applying a cons an o ce (F), he eby inc easing he p essu e (P) acco ding o he ela ion P = F/A. Essen ially, in a mul ian il appa a us, a hyd aulic p ess gene a es an uniaxial o ce which is exe ed on o a se o 6 s eel an il, which is e e ed o as he i s s age an ils. Two a ia ions o i s s age an il design a e known, a spli sphe e o a spil cylinde . This se o 6 an ils c ea es a cubic olume ha is illed wi h a se o eigh cubes (ei he o ungs en ca bide o sin e ed diamond) wi h unca ed co ne s, which unc ions as he second s age an ils. These unca ed an ils c ea e an oc ahed al p essu e chambe . In his p essu e chambe i s in he p essu e cell usually an oc ahed a o MgO con aining he sample, which is comp essed o he equi ed p essu e. A hole is d illed in he MgO oc ahed a o inse ion o a ubula esis ance hea e , usually made o g aphi e, me al oils (inconel, pla inum, henium) o LaC O3. S epped hea e s, whe e hickness o he hea e wall is inc eased in he cen al po ion a e also used some imes o educe he mal g adien s ac oss he la ge sample olumes. Slee es o insula ing ma e ial usually Z O2 a e placed a ound he hea e o p e en excessi e hea anspo o he ungs en ca bide an ils. Expe imen al sample is placed in he cen e o he p essu e assembly and sepa a ed om he hea e by an MgO slee e. MgO space s ill up he space abo e and below he sample capsule. The op 36 Chap e 2: Expe imen al de ails MgO space has a hole o inse ion o he mocouple wi h an alumina ube o empe a u e measu emen s. Py ophylli e gaske s a e used o suppo ing he unca ions and o p essu e sealing he high-p essu e chambe . The maximum p essu e achie able is dependen on he o ce applied by he hyd aulic p ess, he unca ion edge leng h o he second s age an ils, he edge leng h o he MgO p essu e cell and ul ima ely by he s eng h and ha dness o he ungs en ca bide. Fo ou expe imen al in es iga ions in he p esen wo k we ha e used di e en mul ian il appa a us wi h a ied designs loca ed a he Baye isches Geoins i u . Fo ou high-p essu e s udies we ha e employed di e en p essu e assemblies i.e., di e en unca ion edge leng hs o he cubes and di e en oc ahed on edge leng hs sui ing ou expe imen al p essu e equi emen s. We ha e used ungs en ca bide (WC) an ils as he second s age an ils om he comme cial supplie s Toshiba (Japan) and Widia (Ge many). WC cubes we e isola ed om he s eel an ils by epoxy shee s and coppe oils we e used o con ac be ween he p essu e assembly (de ail desc ip ion o p essu e assembly used in ou s udy is gi en in sec ion 2.2.2.2) and he i s s age s eel an ils. A he mocouple was inse ed in o he p essu e assembly o empe a u e measu emen s and a coppe coil p o ec ed he he mocouple in he gaske egion (Fig. 2.2). A eu o he m con olle con e ed he he mocouple e.m. . o empe a u e. Figu e. 2.2: De ails o a mul ian il appa a us and he expe imen al se up. On he le is a schema ic iew showing how he six inne an ils c ea e a cubic space whe e he 8 ungs en ca bide inne an ils i in con aining he p essu e cell and di ec ions o applica ion o o ce. On he le a pho og aph showing he ungs en ca bide an il se up o an expe imen . The MgO p essu e medium can be seen placed inside he oc ahed al ca i y o med by unca ed ungs en ca bide (WC) an ils. Ou o he se o eigh WC cubes, wo cubes a e no shown he e o show he MgO oc ahed a inside. 37 Chap e 2: Expe imen al de ails 2.2.2.2 Mul ian il expe imen s Mul ian il expe imen s we e pe o med in he p essu e and empe a u e ange o 17-23 GPa and 1200-1600°C, o cons ain he CaSiO3 pe o ski e o ming eac ion as a unc ion o composi ion in he mid- ansi ion zone o he Ea h’s man le. Ini ial expe imen al conside a ions Ini ially, o ou mul ian il expe imen s we mixed he Ca- ee ga ne glass powde and he CaSiO3 wollas oni e in a 1:1 a io on a ail basis. The i s expe imen was conduc ed a 1400°C o du a ion o 24 hou s; on analysis o he un p oduc s we saw clea zones in he ga ne s showing ha he expe imen had no eached equilib ium (Fig 2.3 A). This led us o inc ease he expe imen al du a ion o 48 hou s, which did no imp o e he equilib ium kine ics. We he e o e, inc eased he expe imen al empe a u e o 1600°C. The un p oduc showed a clea lack o CaSiO3 o eac ion wi h he ga ne (Fig 2.3 B). So we inally chose a s a ing mix u e o Ca- ee ga ne glass and CaSiO3 wollas oni e in a 1:2 a io (Fig 2.3 C), whe e we could clea ly obse e ga ne in equilib ium wi h calcium pe o ski e. Ga ne Ga ne Ca-P Ca-P Ga ne Ca-P 10 µm 10 µm (A) (B) (C) Figu e. 2.3: Seconda y and back-sca e ed elec on images o di e en expe imen al un samples (A) Zona ion in majo i e ga ne showing disequilib ium in a p elimina y expe imen . (B). Lack o CaSiO3 o u he di usion in o majo i e ga ne . (C). Majo i e ga ne and calcium pe o ski e (Ca- P ) in equilib ium. Du ing he expe imen CaSiO3 di uses in o he ga ne un il i becomes sa u a ed in his componen . Fo he capsule ma e ial we ini ially ied an Al2O3 ou -hole slee e in which we placed he s a ing composi ions in each hole. I eac ed away a expe imen al empe a u es and he e was huge con amina ion om he capsule ma e ial in o he s a ing ma e ials (Fig 2.4 A). Then we ied a molybdenum (Mo) od wi h spa k e oded sample chambe s o he 38 Chap e 2: Expe imen al de ails capsule; in his case he Fe o he s a ing composi ions di used in o he capsule ma e ial. Finally, henium od wi h spa k-e oded sample chambe s o he capsule ma e ial p o ed o be he igh choice because i was s able a high empe a u es and had no e ec on he sample composi ion. Al 2 O 3 Al O capsule 23 eac ed away Mo capsule Re capsule 100 µm 100 µm 100 µm (A) (B) (C) Figu e 2.4: Di e en capsule ma e ials ied ou o ou mul ian il expe imen s (A) Alumina ou hole slee e which eac ed away and only one sample could be eco e ed ha oo had con amina ion o Al2O3, (B) Mo (molybdenum) capsule made ou o 1 mm diame e Mo od wi h i e spa k e oded holes and (C) Shown he e is a adial sec ion h ough he high-p essu e assembly he ou e p essu e medium and u nace wi h an inne ou chambe Re ( henium) capsule (whi e). Th ee sample chambe s con ain ga ne plus Ca-pe o ski e assemblages (which appea ligh e ), while one sample chambe con ains a (Mg,Fe)2SiO4 p essu e calib an sample (da ke ). Mul ian il expe imen s we e ca ied ou using mul ian il p esses o 1000- on and 1200- on axial comp ession capaci ies o his s udy. Expe imen s we e ca ied ou using 10/5 and 10/4 p essu e assemblies i.e., using a C 2O3 doped MgO oc ahed a o 10 mm edge leng h in combina ion wi h ungs en ca bide cubes wi h 5 o 4 mm unca ion edge leng hs. A semi-conduc ing LaC O3 esis ance u nace p o ided he elec ical hea ing and he sample empe a u e was moni o ed using a W3%Re-W25%Re he mocouple, ha was inse ed axially in o he u nace (Fig 2.5). The sample capsule was made ou o 1mm long Re od, ha was spa k e oded o p oduce ou o i e sample chambe s each abou 250 mic ons in diame e (Fig 2.4.C). The mul i-chambe ed capsule allows us o un a leas h ee o he ga ne glass plus CaSiO3 s a ing ma e ials wi h di e en majo i e componen s (i.e., Al/Si a ios) in h ee chambe s in a single expe imen al un along wi h a (Mg,Fe)2SiO4 powde in ano he chambe o p essu e calib a ion. A Re disc o 0.025 mm hickness and an alumina disc o 0.2 mm hickness successi ely co e ed he uppe su ace o he sample capsule o a oid eac ion wi h he he mocouple. Expe imen s we e i s comp essed o he desi ed load and we e subsequen ly hea ed o he equi ed empe a u e o a leas 24 hou s. Expe imen s we e quenched by cu ing he powe supply o he 39 Chap e 2: Expe imen al de ails u nace. Decomp ession o he expe imen was usually pe o med o e a pe iod o 17 hou s. MgO + 5 w % C O oc ahed a 23 Z O slee e 2 LaC O u nace 3 MgO slee e Sample in a Re od capsule Alumina cemen and py ophylli e Mo disc The mocouple Coppe coil Al O disc 23 Re oil disc 4mm Figu e 2.5: A schema ic diag am o a 10/4 p essu e assembly used o mul ian il expe imen s o he p esen s udy, showing an axial c oss sec ion h ough he assembly. (10/4 assembly = 10 mm MgO oc ahed on edge leng h and 4 mm ungs en ca bide unca ion edge leng h). Du ing he expe imen s he s a ing glass plus CaSiO3 mix u e c ys allized as ga ne and calcium silica e pe o ski e and CaSiO3 dissol ed by di usion in o he ga ne . We measu ed his solubili y as a unc ion o p essu e, empe a u e and composi ion. When he majo i e ga ne became sa u a ed in he Ca componen , i s a ed o exsol e Ca- pe o ski e wi h inc easing p essu e. As he ga ne di usion was ound o be slow, we ied ou luxes o ou expe imen s. We did an expe imen whe e we chose he basal ic ga ne plus CaSiO3 composi ion and mixed i wi h abou 5-mol% o wo di e en luxes B2O3 and NaCl. The mul i-chambe ed Re capsule allowed us o pu hese wo luxed composi ions in o wo capsule chambe s along wi h a composi ion wi hou lux and a p essu e calib an in he o he wo chambe s in he same expe imen . Expe imen was ca ied ou a 19.5 GPa (oil ba p essu e) a 1400°C. Analysis o he expe imen al un p oduc showed ha he composi ion wi h NaCl lux showed no enhancemen in CaO solubili y in he ga ne s, 40 Chap e 2: Expe imen al de ails ins ead i c ys allized NaCl c ys als. Howe e , he one wi h he B2O3 lux showed signi ican ly enhanced eac ion in compa ison wi h he same composi ion wi hou any lux. This indica ed ha he B2O3 gene a ed mel a he expe imen al empe a u e ha p omo ed equilib ium. So, all he subsequen expe imen s we e luxed wi h 5-mol% o B2O3. This inc eased he a e o eac ion conside ably a 1600°C and 1400°C bu eac ion a 1200°C was s ill a oo slow o equilib ium o be achie ed on a easible ime scale. Re e sal expe imen s To ensu e ha we achie ed equilib ium in ou expe imen s, e e sal expe imen s we e pe o med using Ca-bea ing ga ne s as he s a ing ma e ial. Fo his we chose he maximum calcium-bea ing ga ne composi ions om ou o wa d uns. High pu i y oxide powde s o CaO, Fe2O3, MgO, Al2O3 we e mixed in he p ope s oichiome ic p opo ions and used a 1600°C in a u nace and apidly quenched in wa e o p oduce he desi ed glass phase. Glass powde s we e educed unde condi ions simila o hose used o he o he glass powde s (see sec ion 2.2.1) and we e cha ac e ized by Mössbaue analysis. Syn heses o he Ca-bea ing ga ne s we e ca ied ou using a 5000- on mul ian il p ess, whe e la ge sample olumes could be u ilized. Expe imen s we e pe o med using 15 mm edge leng h unca ed ungs en ca bide cubes, which comp essed a 25 mm edge leng h C doped MgO oc ahed a con aining he sample capsule jacke ed by an MgO space and a LaC O3 u nace. Sample capsules o 3.5 mm in leng h and 2 mm in diame e we e made ou o Re oil, which was in con ac wi h a Re75%W25%-W3%Re97% he mocouple o empe a u e measu emen s. Ga ne s we e c ys allized a 15 GPa and 1600°C wi hin an expe imen al du a ion o 1.5 hou s. Phase iden i ica ion was ca ied ou using powde x- ay di ac ion. Table 2.2: Re e sal Ca –bea ing ga ne glass composi ions in ca ion p opo ions based on 12 oxygen pe o mula uni as de e mined om elec on mic op obe analysis. Composi ion Si Al Fe Mg Ca ∑Ca ions Ca-Pe ido i e 3.62 0.65 0.31 2.66 0.80 8 Ca-Basal 3.37 1.06 0.29 2.32 1.04 8 Ca-Py ope 3.27 1.34 0.27 1.99 1.20 8 Ca-Majo i e 3.67 0.49 0.35 2.83 0.76 8 41 Chap e 2: Cha ac e iza ion echniques In he p esen s udy, he SEM echnique was used on a ou ine basis o imaging he high-p essu e, high- empe a u e mul ian il un p oduc s ha had been polished a e being embedded in epoxy esin and wi h ca bon coa ing o educe cha ging on he su ace (simila o ha o EPMA analysis) a an accele a ion ol age o 20 KeV and wi h di e en magni ica ions. Raman spec oscopy This spec oscopic me hod is based on he inelas ic sca e ing o ligh called he Raman e ec . When ligh is allowed o in e ac wi h a specimen, i exci es he cons i uen molecules, which subsequen ly sca e he ligh . Mos o his sca e ed ligh has a simila wa eleng h as he inciden ligh while some o i is sca e ed wi h a di e en wa eleng h. This inelas ically sca e ed ligh is he Raman sca e , which esul s om he changed molecula mo ions o he sample. The di e ence in ene gy be ween he inciden ligh and Raman sca e ed ligh is equal o he ene gy o he sca e ing molecule and will be cha ac e is ic o a molecule and i s en i onmen in a specimen. By plo ing he ene gy di e ence and he in ensi y o sca e ed ligh we ob ain a Raman spec um (Nasdala e al., 2004). In he p esen s udy we used a LABRAM Raman spec ome e wi h a He-Ne lase wi h he 632 nm ed line exci a ion o phase iden i ica ion o ou expe imen al un p oduc s. This was especially use ul in case o ou p essu e calib an s whe e we could easily dis inguish be ween he (Mg,Fe)2SiO4 polymo phs using his echnique. Raman spec a we e collec ed a ambien empe a u e wi h an ins umen al esolu ion o 2 cm-1 o he peak posi ions. By compa ing he ob ained spec a wi h he ele an s anda d Raman da a a ailable o he mine al phase om he li e a u e we ca ied ou he phase iden i ica ion (Fig. 2.8) 48 Chap e 2: Resul s 600 800 1000 9000 10000 11000 12000 13000 14000 15000 In ensi y Wa e numbe cm-1 600 800 1000 1200 800 1000 1200 1400 1600 1800 In ensi y Wa e numbe cm-1 Wadsleyi e Ringwoodi e Figu e 2.8: Raman spec a o wadsleyi e and ingwoodi e as c ys allized by he in-si u p essu e calib an s o ou expe imen s. Phase iden i ica ion was ca ied ou by compa ing hese spec a wi h he exis ing Raman spec a om he li e a u e o he conce ned phase. 2.4. Resul s Du ing he high p essu e expe imen s he ga ne glass and CaSiO3 composi ions c ys allized apidly o ga ne and pe o ski e espec i ely and CaSiO3 was dissol ed in o he ga ne du ing he hea ing pe iod. Expe imen al du a ions o a leas 24 hou s ensu ed equilib ium composi ions a 1600°C a lowe p essu es. Howe e , in some cases a highe p essu es, he inne co es o ga ne emained unequilb a ed bu equilib ium was achie ed in he ims wi h he aid o he B2O3 lux as men ioned in he expe imen al sec ion. As discussed ea lie , equilib ium was c osschecked by analyzing he e e sal expe imen s in which Ca-bea ing ga ne s exsol ed Ca pe o ski e a high p essu e. We had equilib ium in ou expe imen s because o a gi en p essu e he ex en o CaO solubili y in he ga ne phase a e CaSiO3 exsolu ion con e ged o he same alues as ob ained in he o wa d uns wi hin expe imen al e o . A 1400°C, high Ca-bea ing ga ne was o med up o p essu e o ~20 GPa, beyond which Ca con en s in he ga ne declined d as ically. Time s udies indica ed ha equilib ium was no achie ed a hese condi ions; a possible explana ion could be ha a hese condi ions he B2O3 lux c ys allized. In he case o expe imen s a 1200°C, e en he luxed expe imen s ailed o each equilib ium on a ime scale o 48 hou s. Longe du a ion equi emen s we e no easible. Du ing he expe imen s he (Mg,Fe)2SiO4 oli ine p essu e calib an c ys allized o o m co-exis ing high-p essu e phases. P essu e was de e mined om he Mg2SiO4( o s e i e)-Fe2SiO4( ayali e) phase diag am using he ield o coexis ence 49 Chap e 2: Resul s be ween (Mg,Fe)2SiO4 and (Mg,Fe)O magnesiowüs i e plus s isho i e (SiO2) (Fig. 2.9). This di a ian egion ex ends om he Fe2SiO4 ingwoodi e o magnesiowüs i e plus s isho i e ans o ma ion a app oxima ely 16 GPa o 23 GPa wi h he Fe/(Fe+Mg) a io o bo h ingwoodi e and magnesiowüs i e (when coexis ing wi h s isho i e) dec easing wi h inc easing p essu e. This eac ion has been well s udied (Ma suzaka e al., 2000; F os e al., 2001; F os , 2003a). The Fe/(Fe+Mg) a io o ingwoodi e o example dec eases by app oxima ely 10% pe GPa. As we can de e mine his a io wi h an accu acy o app oxima ely 1%, his ga e us a p ecision in p essu e de e mina ion o 0.1 GPa. A p e ious s udy had shown ha he bulk Fe concen a ion o magnesiowüs i e could be appa en ly high i e ic Fe is p esen . We calcula ed he p essu e using he ingwoodi e Fe concen a ion and used he phase ela ions p e iously de e mined a a simila oxygen ugaci y (F os e al., 2001; F os , 2003a). In his way we calib a ed accu a ely he p essu e o Ca-pe o ksi e o ma ion ela i e o he p essu e o phase ans o ma ions in he Mg2SiO4-Fe2SiO4 sys em (F os e al., 2001; F os , 2003b). The absolu e e o in p essu e depends on de e mina ions o end membe phase ansi ions used o cons uc he exis ing Mg2SiO4-Fe2SiO4 phase diag am which a e di icul o assess bu could be up o 1 GPa. Ou s udy elied mo e on he high p ecision o p essu es de e mined ela i e o he Mg2SiO4-Fe2SiO4 phase diag am, a he han on accu acy in absolu e p essu e. When he magnesiowüs i e g ains we e oo small in ou expe imen s o ge a eliable analysis, he p essu e was de e mined using he ingwoodi e Fe/(Fe+Mg) a io alone. In hose cases whe e a poo ly chosen Fe/(Fe+Mg) a io o he s a ing oli ine composi ion led o c ys alliza ion o a single phase ingwoodi e, p essu e was de e mined om he oil p essu e o he mul ian il expe imen . In expe imen s whe e he p essu e was measu ed, he de e mined p essu e showed excellen ela ionship wi h he mul ian il oil p essu e. The ela i e unce ain y using his calib a ion cu e in compa ison o p essu es de e mined om phase ela ions in Mg2SiO4-Fe2SiO4 sys em was app oxima ely ± 0.5 GPa. In some o ou expe imen s he p essu e calib an s e en c ys allized in he s abili y ield o pe o ski e plus magnesiowüs i e and s isho i e (Run No.S3478, S3784), see Appendix A, Table A.2 o he pe o ski e composi ions. Expe imen al esul s a e gi en in Table 2.4,whe e he ela i e p essu e e o s we e calcula ed using he misma ch be ween he analysed Fe con en s o ingwoodi e and magnesiowüs i e compa ed wi h he phase diag am o F os e al., (2001). 50 Chap e 2: Resul s 0 0.2 0.4 0.6 0.8 1 16 18 20 22 24 26 Fe/(Fe+Mg) P essu e (GPa) Mg2SiO4Fe 2 SiO4 1600oC Ringwoodi e Ring + Mw + S ish P + Mw P + Mw+ S ish Mw+S ish Figu e 2.9: Phase ela ions in he Mg2SiO4-Fe2SiO4 sys em be ween 16- 26 GPa a 1600 oC (F os e al., 2001) used o calcula ing he p essu es in ou expe imen s. Oli ine p essu e calib an s du ing expe imen s c ys allized in o i s co esponding high p essu es phases o ingwoodi e, magnesiowüs i e and s isho i e. F om p obe da a we de e mined he Fe/Fe+Mg a io o coexis ing ingwoodi e and magnesiowüs i e a e co ec ion o e ic i on con en (F os , 2003a). By plo ing he Fe/Fe+Mg alues as shown ma ked by s a s we could exac ly de e mine he p essu e in each o ou expe imen s. Table 2.4: Expe imen al esul s o ga ne composi ions eco e ed om mul ian il expe imen s as de e mined by elec on mic op obe analysis o di e en expe imen al uns. Resul s a e lis ed as pe ido i e, basal , py ope and majo i e based on he di e en ga ne s a ing composi ions as lis ed in able 2.1.Lis ed he e a e he expe imen al un numbe s, p essu es in GPa o each expe imen de e mined as desc ibed in ex and he ca ion p opo ions o he ga ne composi ions calcula ed based on 12 oxygen pe o mula uni . Da a in Ring ( ingwoodi e) and MW (magnesiowüs i e) columns a e he Fe/Fe+Mg a ios used o calcula ion o p essu e as desc ibed in ex . (See Appendix A, Table A.3 o he p obe da a o all expe imen al uns) 51 Run no. P essu e Si Al Fe Mg Ca To al Ring MW 1600°C (in GPa) Py ope S3550 19.2(3) 3.41(2) 1.08(4) 0.21(1) 2.25(6) 1.10(4) 8.050 0.626(3) 0.937641 S3549 19.9(2) 3.12(1) 1.52(2) 0.17(2) 2.25(5) 0.90(5) 8.038 0.542(3) 0.87(3) S3470 18.8(5) 3.32(5) 1.36(4) 0.05(2) 2.03(8) 1.22(1) 7.993 S3475 22.6(5) 3.11(6) 1.82(6) 0.207(6) 2.47(6) 0.37(3) 7.982 S3478 23(5) 3.09(2) 1.77(3) 0.28(5) 2.56(6) 0.32(3) 8.023 S3484 22.3(5) 3.08(5) 1.81(5) 0.24(7) 2.57(7) 0.30(7) 8.011 0.36(2) S3490 21.4(5) 3.10(5) 1.76(9) 0.23(1) 2.47(6) 0.44(1) 8.013 Basal S3551 20.4(3) 3.33(6) 1.11(7) 0.1(1) 2.94(8) 0.61(4) 8.096 0.496(2) 0.88(2) S3550 19.2(3) 3.29(2) 1.36(4) 0.178(9) 1.90(6) 1.27(7) 8.023 0.626(3) 0.937641 Chap e 2: Resul s S3549 19.9(2) 3.38(1) 1.14(3) 0.16(2) 2.70(3) 0.65(3) 8.044 0.542(3) 0.87(3) S3548 19.5(2) 3.41(2) 1.04(3) 0.30(2) 2.56(3) 0.75(3) 8.068 0.605(5) 0.863(4) S3547 19.5(5) 3.45(2) 1.04(2) 0.21(1) 2.58(5) 0.74(4) 8.020 0.592(2) 0.844(6) S3764 21.2(3) 3.34(3) 1.26(2) 0.21(2) 2.85(6) 0.35(1) 8.021 0.426(5) 0.73(1) S3757 20.7(5) 3.36(3) 1.29(4) 0.199(8) 2.65(3) 0.47(3) 7.990 0.469(2) 0.774(6) S3783 22.3(2) 3.34(2) 1.31(2) 0.262(7) 2.81(4) 0.29(1) 8.010 0.350(4) 0.71(5) S3657 19.9(3) 3.38(2) 1.01(5) 0.44(1) 2.62(5) 0.64(2) 8.108 0.545(3) 0.88(1) S3655 19.9(5) 3.36(5) 1.25(7) 0.39(3) 2.40(6) 0.62(4) 8.018 S3655 19.9(5) 3.42(1) 1.04(3) 0.24(3) 2.77(9) 0.58(2) 8.060 S3460 18.5(5) 3.502(7) 1.06(2) 0.049(4) 2.22(6) 1.12(3) 7.970 S3470 18.8(5) 3.44(2) 1.09(1) 0.03(7) 2.37(3) 1.05(2) 8.006 S3478 23(5) 3.30(2) 1.33(3) 0.25(1) 2.89(6) 0.24(2) 8.028 S3480 22.1(5) 3.47(5) 1.16(7) 0.27(2) 2.67(4) 0.37(3) 7.950 0.380(5) S3484 22.3(5) 3.33(3) 1.34(2) 0.20(6) 2.93(3) 0.188(7) 8.000 0.36(2) S3498 21.4(5) 3.33(5) 1.28(9) 0.276(8) 2.80(6) 0.34(3) 8.031 0.28(2) Pe ido i e S3551 20.4(3) 3.54(4) 0.87(8) 0.26(3) 2.9(1) 0.44(8) 8.022 0.496(2) 0.88(2) S3550 19.2(3) 3.62(2) 0.70(3) 0.22(8) 2.58(3) 0.92(3) 8.030 0.626(3) 0.937641 S3549 19.9(2) 3.58(1) 0.76(2) 0.11(4) 3.04(2) 0.54(2) 8.033 0.542(3) 0.87(3) S3548 19.5(2) 3.61(1) 0.68(3) 0.23(2) 2.93(4) 0.59(3)8 8.051 0.605(5) 0.863(4) S3547 19.5(5) 3.57(2) 0.77(3) 0.167(8) 2.88(5) 0.63(2) 8.037 0.592(2) 0.844(6) S3764 21.2(3) 3.60(3) 0.80(2) 0.28(3) 3.02(2) 0.283(1) 8.000 0.426(5) 0.73(1) S3757 20.7(5) 3.65(1) 0.69(1) 0.24(5) 3.082(4) 0.322(1) 7.998 0.469(2) 0.774(6) S3784 23.5(4) 3.60(2) 0.68(9) 0.31(4) 3.4(1) 0.054(1) 8.061 0.545(3) 0.88(1) S3460 18.5(5) 3.67(4) 0.63(3) 0.10(2) 2.79(7) 0.81(5) 8.010 S3470 18.8(5) 3.664(3) 0.683(5) 0.05(2) 2.75(3) 0.84(3) 7.994 S3475 22.6(5) 3.62(2) 0.791(2) 0.28(1) 3.21(5) 0.08(2) 7.984 Majo i e S3460 18.5(5) 3.78(4) 0.47(2) 0.098(9) 2.97(6) 0.65(4) 7.979 S3470 18.8(5) 3.77(1) 0.49(1) 0.034(5) 2.87(7) 0.81(3) 7.985 Re e sals S3784 23.5(4) 3.15(1) 1.68(3) 0.26(2) 2.59(6 0.30(4) 8.000 0.545(3) 0.88(1) S3538 17.9(5) 3.24(5) 1.36(5) 0.05(2) 2.82(7) 0.57(6) 8.067 S3543 18.4(5) 3.14(6) 1.57(6) 0.16(3) 2.63(4) 0.55(5) 8.069 S3515 19.8(5) 3.41(2) 1.08(2) 0.22(7) 2.42(3) 1.01(4) 8.042 S3521 20.7(5) 3.55(4) 0.87(4) 0.33(2) 2.77(6) 0.48(4) 8.008 S3523 20.8(5) 3.40(2) 1.18(3) 0.23(1) 2.53(2) 0.63(2) 8.015 0.464(3) 1400°C Py ope S3611 18.1(5) 3.17(3) 1.61(7) 0.144(7) 2.01(3) 1.08(5) 8.023 S3614 18.6(5) 3.15(3) 1.56(5) 0.197(6) 2.09(6) 1.055(5) 8.064 H2370 19.5(5) 3.10(6) 1.75(9) 0.236(5) 2.35(5) 0.58(4) 8.020 0.60(1) 0.93(1) H2375 17.9(5) 3.24(1) 1.46(4) 0.200(5) 2.00(3) 1.11(3) 8.025 H2241 19.6(5) 3.12(4) 1.76(6) 0.19(1) 2.41(5) 0.51(3) 7.996 Basal S3611 18.1(5) 3.301(8) 1.31(2) 0.199(4) 2.12(2) 1.11(2) 8.044 S3614 18.6(5) 3.33(2) 1.17(4) 0.223(7) 2.49(2) 0.86(2) 8.080 H2375 17.9(5) 3.40(3) 1.13(7) 0.223(8) 2.22(6) 1.04(3) 8.028 Pe ido i e S3614 18.6(5) 3.56(1) 0.69(2) 0.263(7) 2.91(4) 0.66(4) 8.090 S3611 18.1(5) 3.60(3) 0.71(5) 0.26(5) 2.68(4) 0.78(4) 8.042 52 Chap e 2: Resul s The Ca con en s o ga ne in equilib ium wi h Ca-pe o ski e a 1600°C and a 1400°C a e shown in Fig. 2.10 (A) and (B). Da a o all he e e sal expe imen s a 1600°C a e shown in Fig. 2.10(A). I was obse ed om he expe imen al da a a 1600°C ha he solubili y o he Ca in ga ne in equilib ium wi h calcium pe o ski e dec eases wi h p essu e. Howe e , he Ca solubili y inc eases wi h he ga ne majo i e con en (i.e., Al/Si a io) a a gi en p essu e. This implies ha he CaO sa u a ion o he ini ia ion o CaSiO3 exsolu ion will occu a a highe CaO con en wi h dec easing majo i e componen in majo i e ga ne a ansi ion zone p essu e- empe a u e condi ions. Low Al/Si a ios o abou 0.7 a e ep esen a i e o ga ne s o med in pe ido i ic bulk composi ions, while highe alues o abou 1.4 a e hose o ga ne s o med in a basal ic composi ion. Expe imen al da a a 1400°C also showed a simila Ca solubili y end in ga ne , i.e., CaO solubili y in he majo i e ga ne composi ion dec eases wi h inc easing majo i e componen Al/Si+Mg. Howe e , he e is a small bu dis inguishable dec ease in he Ca solubili y compa ed wi h esul s ob ained a 1600°C. As Fe is also an impo an componen in ansi ion zone phases we ca ied ou a ew expe imen s o obse e he e ec o a ying Fe concen a ion on he CaO solubili y in he ga ne s in equilib ium wi h calcium pe o ski e. Fo his we mixed ou luxed s a ing composi ions wi h a powde o 5-mol% Fo s e i e 80 composi ion, which is a ele an composi ion o he man le dep h a which his eac ion akes place. We placed one i on- bea ing sample along wi h a no mal s a ing composi ion in a single un using he mul i- chambe ed capsule. On analysis o he expe imen al un p oduc s no signi ican change in he CaO solubili y was obse ed compa ed wi h ha o he s a ing samples wi hou addi ional mixing o Fe. 53 Chap e 2: Resul s (A) P essu e(GPa) (B) 18.0 18.5 19.0 19.5 0.6 0.8 1.0 1.2 Py ope Al = 2 Basal Al = 1.4 Pe ido i e = 0.7 Ga ne Ca con en (p. .u.) P essu e in GPa CaSiO3 ga ne solubil y Figu e 2.10. Expe imen al esul s o he Ca-con en o ga ne in equilib ium wi h Ca- pe o ski e as a unc ion o p essu e (A) a 1600°C and (B) a 1400°C. Symbols e e o s a ing composi ions wi h di e en ga ne majo i e p opo ions named a e he ock ypes in which he ga ne s would occu . Al (p u) e e s o he Al con en pe ga ne o mula uni , i.e. Al (p u) = 2 e e s o he o mula (Mg,Fe)3Al2SiO12. These Al con en s only e lec he s a ing ma e ial because he Al con en is lowe ed du ing he expe imen s. Re e sals we e pe o med using p esyn hesised Ca-bea ing ga ne s as s a ing ma e ials. The cu e in (A) shows he calcula ed Ca con en o ga ne o a ock o pe ido i e composi ion based on pe ido i e composi ion om McDonough and Sun, (1995). Fo discussion on he pe ido i e model see sec ion 2.5. 54 Chap e 2: Resul s 2.4.1 The modynamic modeling The exsolu ion o calcium silica e pe o ski e om majo i e ga ne can be desc ibed by he ollowing eac ion Ca3Al2Si3O12+3/4(Mg,Fe)4Si4O12 = 3 CaSiO3 + (Mg,Fe)3Al2Si3O12 [1] Ga ne Ga ne Pe o ski e Ga ne whe e, he g ossula componen in he ga ne exsol es o o m calcium silica e pe o ski e, which lea es he ga ne inc easingly aluminium ich and combined wi h he majo i e componen , p oduces py ope. The expe imen al esul s we e i ed o a he modynamic model based on eac ion [1]. The impo an a iables in his i a e he olume change o he eac ion and he non- ideal mixing pa ame e s o majo i ic ga ne . Because he da a co e a ange o ga ne - majo i e composi ions, his model was used o calcula e he exsolu ion o CaSiO3 o e a ange o bulk composi ions ele an o he man le. This also allowed ou esul s o be ex apola ed o lowe empe a u es a which slow kine ics inhibi ed he achie emen o equilib ium in expe imen s. A equilib ium condi ions he s anda d s a e ee ene gy change o eac ion [1] is ela ed o he ac i i ies o he eac ing componen s by he equa ion, 0 ,TP G∆ [] [ ] [ ] 4 3 0 ,12441232312323 lnln aaa G OSiMg G OSiAlCa G OSiAlMg TP RTKRTG −=−=∆ [2] whe e R is he gas cons an , K is he equilib ium cons an and is he ac i i y o componen i in ga ne . In o de o es ima e ac i i ies i was assumed ha Ca mixes only on he dodecahed al si e in ga ne and ha Fe beha es iden ically o Mg. The ac i i y o CaSiO G i a 3 is uni y whe eas he ac i i ies o he ga ne componen s a e desc ibed by () ( ) 3 1244 1244 XX adodec Mg oc Maj G OSiMg G OSiMg γ = [3] () ( ) 3 1 12323 12323 XX adodec Ca oc Maj G OSiAlCa G OSiAlCa −= γ [4] () ( ) 3 1 12323 12323 dodec Mg oc Maj G OSiAlMg G OSiAlMg XX a−= γ [5] whe e,      − =2 2nAl Xoc Maj [6] 55 Chap e 2: Resul s      −− =3 3nFenCa Xdodec Mg [7]       =3 nCa Xdodec Ca [8] nAl, nCa and nFe a e elemen p opo ions in ga ne based on 12 oxygen o mula uni s and γ is he ac i i y coe icien . Si and Mg a e he e o e assumed o be locally o de ed on he oc ahed al si e. The ac i i y coe icien s (γ) we e i s de e mined using a ou -componen symme ic solu ion model ha include e ms o he non-ideali y o majo i e mixing. Howe e , all majo i e e ms, which we e e ined in he i ing, we e e y small and we e he e o e igno ed. A e na y symme ic solu ion model was hen used ha accoun ed o non-ideali y esul ing om Mg, Ca and Fe mixing only on he dodecahed al si e, i.e., ( ) WWW XX W X W XCaFeMgFeMgCa FeCa MgFeMgCa G OSiAlMg FeCa RT −+++= 22 12323 ln γ [9] ( ) WWW XX W X W XMgFeCaFeMgCa FeMg CaFeMgCa G OSiAlCa Fe Mg RT −+++= 2 2 12323 ln γ [10] whe e W a e Ma gules in e ac ion pa ame e s. The ollowing alues we e used based on published alues (O’Neill and Wood, 1979; F os , 2003b). 300= WMgFe J/mol [11] 2000= W CaFe J/mol [12] Al hough many s udies show he py ope-g ossula solid solu ion o ha e signi ican asymme ic excess en halpy, en opy and olume p ope ies, ex apola ion o he many published models o p essu es in excess o 20 GPa and 1873K yields wildly dispa a e ac i i y coe icien s. We he e o e ha e no op ion bu o ea WMgCa as an adjus able pa ame e bu he ange o Ca con en s co e ed by ou da a makes an asymme ic i uncons ained and unnecessa y o ou pu poses. The e o e, we e ine symme ic e ms o WMgCa bu cons ain hem o a y o e a ange p ede ined om li e a u e alues o ge , P WMgCa 3008000 += J/mol [13] whe e P is in GPa. 56 Chap e 2: Resul s Fig. 2.11, shows he RTlnK calcula ed om he expe imen al da a using he ac i i y model. The da a we e i ed in a leas squa es e inemen o an equa ion o ha ga e, 0 ,TP G∆ PTG TP 12560773.26140763 0 ,−+−=∆ J/mol [14] wi h T in K and P in GPa. As is he s anda d Gibbs ee ene gy change o he pu e componen s a p essu e and empe a u e, his can be simply desc ibed in e ms o en opy, en halpy and olume changes by, 0 ,TP G∆ VPSTHG TP ∆+∆−∆=∆ 0 , [15] We i he calcula ed alues o he -RTlnK e m o equa ion [15] .The olume change o equa ion [1] is de e mined as 12.56 cm3 mol-1 which is e y easonable gi en he olumes and equa ion o s a e p ope ies o he phases in ol ed 20000 40000 60000 80000 100000 120000 140000 17 18 19 20 21 22 23 24 Basal ic Pe ido i ic Py ope Re e sals 1400°C Fi 1600°C Fi 1400°C RTlnK (kJ/mol) P essu e (GPa) Figu e. 2.11: A plo o RTlnK as de ined in equa ion (2) as a unc ion o p essu e. Da a om a ious s a ing composi ions a 1600°C a e indica ed, while py ope, basal ic and pe ido i ic composi ions a e plo ed oge he a 1400°C. Fi ed cu es o desc ibed using equa ion (14) a e shown a bo h empe a u es. 0 ,TP G∆ 57 Chap e 2: Discussion 1200 1300 1400 1500 1600 440 480 520 560 600 Majo i e - Calcium pe o ski e ansi ion mid poin Wadsleyi e-Ringwoodi e ansi ion mid poin Tempe a u e °C Dep h in Km Figu e. 2.14: Compa ison o he ansi ion p essu e o wadsleyi e- ingwoodi e and majo i e-Ca-pe o ski e as unc ion o man le empe a u es o 1400°C and 1600°C showing posi i e Clapey on slopes as ep esen ed by he solid lines o bo h he eac ion. The closed ci cles a e he ansi ion mid poin s o wadsleyi e- ingwoodi e based on wo k o F os , (2003) and he open squa es a e he ansi ion mid poin s o he majo i e o Ca-pe o ski e om he p esen expe imen al s udy. A mo e consis en explana ion is ha in many egions whe e only a single 520 km discon inui y is obse ed he Ca-pe o ski e eac ion is in isible because he Ca con en o he man le is oo low i.e., he egion has unde gone pa ial mel ing a he su ace a some poin in i s his o y and is deple ed in a e ile componen . A s ong 560 e lec ion, on he o he hand, would indica e he p esence o e ile man le o man le ha con ains a signi ican componen o ecycled MORB c us . The impedance con as o he Ca- pe o ski e eac ion in MORB is 2.8 %. As no wadsleyi e o ingwoodi e ansi ion occu s in MORB, man le wi h an en iched ecycled MORB componen would ha e a weake 510 discon inui y bu a much s onge e lec ion a app oxima ely 560 km, in line wi h some o he obse a ions. The 520 km discon inui y becomes upli ed o app oxima ely 500 km as i spli s, which based on he Clapey on slope o he wadsleyi e o ingwoodi e ansi ion implies local man le empe a u es ha a e coole by a leas 100°. As hese egions also display a s ong Ca-pe o ski e ansi ion a good explana ion o he spli ing would be ha i occu s in a eas ha con ain signi ican amoun s o subduc ed oceanic c us . I subduc ed slabs descend a high angle di ec ly in o he lowe man le hese egions would no be expec ed o 64 Chap e 2: Conclusions be nea ly as globally ex ensi e as hey appea o be. Howe e , he widesp ead spli ing o 520 would be easily explained i subduc ed ma e ial is accumula ing a he base o he ansi ion zone, as a esul o ei he densi y o heological con as wi h he lowe man le (Ch is ensen, 1997; Ka ason and an de Hils , 2000), c ea ing ela i ely cool la - lying egions ich in oceanic c us . 2.6. Conclusions Ou expe imen al da a suppo he hypo hesis ha he wadsleyi e o ingwoodi e ansi ion and he Ca-pe o ski e o ming eac ions could p oduce wo dis inc discon inui ies be ween 500 and 560 km. The p esence and ampli ude o a discon inui y a app oxima ely 560 km will be sensi i e o he Ca-con en o he man le and will consequen ly be an indica o o he p esence o e ile man le o man le en iched in ecycled oceanic c us . This is, he e o e, he i s man le discon inui y o be iden i ied ha is sensi i e o he main ype o la ge-scale chemical he e ogenei y expec ed in he Ea h’s man le. The obse ed a iabili y in 520 km seismic discon inui y spli ing indica es ha he man le a his dep h may be he e ogeneous o e qui e ex ensi e egions. On conside ing he geog aphical dis ibu ion o he spli 520 km seismic discon inui y, i is obse ed ha Deuss and Woodhouse, (2001), in hei seismological s udy epo ed he p esence o a dis inc ly spli 520 km seismic discon inui y om aces ac oss No h Ame ica, he Eas Paci ic egion, Indonesia and in he a eas a ound he No h A ican Shield. Like wise, Gilbe e al., (2003) also epo ed a spli 520 km seismic discon inui y in a eas benea h he Colo ado Pla eau, he Basin and Range p o ince and he Rocky Moun ains in he No h Wes e n Uni ed S a es. These egions a e all undoub edly si es o mode n day o pas subduc ion. The e o e, a co ela ion can be d awn suppo ing ou idea o en ichmen o Ca in he man le by subduc ion o oceanic li hosphe e benea h con inen al pla es om he geog aphical dis ibu ion o he obse ed spli 520 km seismic discon inui y. Fu he s udies in o he a iabili y and spli ing shown by he 520 km seismic discon inui y will p o ide key insigh s in o he ci cula ion o subduc ed c us and he la e al dis ibu ion o chemical he e ogenei y in he ansi ion zone. 65 66 Chap e 3 The comp essibili y o (Fe,Al)-bea ing magnesium silica e pe o ski e de e mined by single-c ys al X- ay di ac ion: implica ions o lowe man le p ope ies. 3.1 In oduc ion The ans o ma ion o (Mg,Fe)2SiO4 ingwoodi e o (Mg,Fe)SiO3 pe o ski e and (Mg,Fe)O magnesiowüs i e ma ks he beginning o he lowe man le. This ansi ion is he likely cause o he 670 km seismic discon inui y (Ringwood and Majo , 1967). A dep hs below 670 km, (Mg,Fe)SiO3 pe o ski e becomes he dominan man le cons i uen , accoun ing o app oxima ely 80 olume pe cen o he Ea h’s lowe man le. Ou unde s anding o he s uc u e and empe a u e o he lowe man le is mainly based mainly on he compa ison o seismic wa e eloci ies wi h model calcula ions o wa e eloci ies o mine al assemblages a lowe man le condi ions. The eliabili y o such a compa ison depends on he accu acy o expe imen al and heo e ical es ima es o mine al elas ic p ope ies a he condi ions o he deep man le. A comple e knowledge o he elas ic p ope ies o (Mg,Fe)SiO3 pe o ski e o e he ange o p essu e, empe a u e and composi ion likely encompassed by he lowe man le is, he e o e, a p e equisi e o cons aining he physical and chemical p ope ies o he lowe man le. Because he lowe man le comp ises he bulk o he silica e Ea h, knowledge o i s composi ion and empe a u e is i al o ou unde s anding o he Ea h’s composi ion as a whole and man le dynamics in gene al. Pe o ski e- ype compounds ha e ABX3 s oichiome y and a e cha ac e ised by a ne wo k o BX6 co ne -sha ing oc ahed a. In he ideal pe o ski e s uc u e he oc ahed al amewo k o ms a cubic a ay and he A ca ions occupy he la ge 12- old coo dina ed si e. 67 Chap e 3: In oduc ion This s uc u e is cubic and has he space g oup mPm3 . The majo i y o pe o ski e s uc u es, howe e , including (Mg,Fe)SiO3 pe o ski e, a e dis o ed de i a i es o he cubic s uc u e. The mos common ype o dis o ion, as in he case o (Mg,Fe)SiO3 pe o ski e, a ises when he size o he A-ca ion is oo small o he 12- old si e. To accommoda e such ca ions he oc ahed a il abou he pseudocubic axes, in his way he A- X bond-leng hs a e no longe all equal wi h a consequen change o he A-si e coo dina ion and educ ion in symme y om he cubic a ys o ype. Oc ahed al il ing in pe o ski e has been discussed in de ails by se e al au ho s (see o example Glaze , 1972, 1975; Megaw, 1973; Woodwa d 1997; Howa d and S okes, 1998). (Mg,Fe)SiO3 pe o ski e has an o ho hombic symme y (space g oup Pbnm) due o he il ing o he BO6 amewo k which b ings some oxygens close o he A ca ions, esul ing in a lowe ing o coo dina ion o he A ca ions om [XII] o [VIII] old (Fig. 3.1). The A si e o (Mg,Fe)SiO3 pe o ski e is mainly occupied by di alen Mg and Fe ca ions and he B si e by Si. Di alen Fe subs i u es o Mg up o app oxima ely 25 % in o he A si e a condi ions compa ible wi h he op o he Ea h’s lowe man le, causing a sligh expansion o he s uc u e and a dec ease in dis o ion (Ross and Hazen, 1989; Ross and Hazen, 1990; Hemley and Cohen, 1992). B x 6 oc ahed a A ca ion a b c Figu e. 3.1: Dis o ed o ho hombic MgSiO3 pe o ski e s uc u e, showing co ne sha ed BX6 oc ahed a wi h dodecahed al A ca ion si e. 68 Chap e 3: In oduc ion Oc ahed al il angles can be es ima ed using uni cell dimensions. Fo o ho hombic Pbnm pe o ski es, wi h no dis o ion o he BX6 oc ahed a, h ee il angles can be de ined ( ollowing he no a ion o Zhao e al., 1993a,b). These angles, θ, φ, and Φ ep esen o a ions o he oc ahed a abou he pseudo-cubic axes [110], [001] and [111] espec i ely and can be calcula ed acco ding o he ollowing equa ions: ba/cos = θ (1) ca/2cos = φ (2) bca /2cos 2 =Φ (3) Al hough he uni cell dimension me hod no mally unde es ima es he il angles, as eal oc ahed a a e ypically sligh ly dis o ed, i is s ill e y use ul when he a omic coo dina es o o ho hombic pe o ski es a e no known. T i alen ca ions also subs i u e in o (Mg,Fe)SiO3 pe o ski e al hough he subs i u ion mechanism may a y depending on he ca ion and on p essu e and empe a u e. The e is e idence, o example, ha Al en e s he s uc u e h ough a coupled subs i u ion on o bo h A and B si es (S ebbins e al., 2001). O he wo ke s, howe e , ha e aised he possibili y ha Al may eplace Si on he oc ahed al B si e wi h cha ge balance p o ided by he c ea ion o an oxygen acancy (Na osky, 1999). Ce ainly a p essu es >30 GPa a coupled subs i u ion seems o occu wi h up o 25 % subs i u ion o an AlAlO3 componen being possible (Wal e e a., 2004). Al subs i u ion causes he oc ahed al dis o ion o inc ease (Hemley and Cohen, 1993). I was ecen ly es ablished expe imen ally ha a signi ican po ion o Fe in Al- bea ing (Mg,Fe)SiO3 pe o ski e is in he e ic (Fe3+) s a e (McCammon, 1997). Suppo ha his may also be he case in he lowe man le comes om he analyses o high e ic Fe con en s in mine al inclusions apped in diamonds ha appea o ha e o iginally had he pe o ski e s uc u e (Ha e e al., 1999). Expe imen al s udies ha e also shown ha he Fe3+/∑Fe a io in magnesium silica e pe o ski e is s ongly co ela ed wi h Al con en (Wood and Rubie, 1996; McCammon, 1997; F os and Langenho s , 2002; Lau e bach e al., 2000; McCammon e al., 2004). Such a s ong coupling would imply ha a coupled subs i u ion akes place. Richmond and B odhol , (1998) pe o med compu e simula ions o examine he ene gy associa ed wi h a ious i alen ca ion subs i u ion mechanisms i.e., 69 Chap e 3: In oduc ion Si4+[VI] + Mg2+[VIII]↔Al3+[VI] + Fe3+[VIII] (4) 2Si4+[VI] + O2-↔2Fe3+[VI] + V0 (5) 2Si4+[VI] + O2-↔2Al3+[VI] + V0 (6) These au ho s epo ed ha he coupled subs i u ion mechanism (4), whe e Al subs i u ions on o he Si si e and Fe3+ on o he Mg si e, is mo e ene ge ically a ou able han he oxygen acancy mechanisms (5) and (6). The inco po a ion o Al and he esul ing s abiliza ion o Fe3+ in he pe o ski e s uc u e can po en ially a ec i s elas ic p ope ies (Na o sky e al., 2003). A numbe o obse a ions o seismic anomalies in he lowe man le ha e been a ibu ed o chemical he e ogenei y (T ampe , 2004) and he e a e a numbe o possibili ies as o how Al and Fe concen a ions could a y in he lowe man le. In o de o unde s anding he o igin o seismic anomalies in he lowe man le a comple e unde s anding o how ca ion subs i u ions in luence he elas ic p ope ies o (Mg,Fe,)SiO3 pe o ski e is equi ed. Values o he bulk modulus o end membe MgSiO3 pe o ski e de e mined using s a ic comp ession and dynamic echniques such as B illouin sca e ing, co e a ange o 250-265 GPa wi h K' no mally ixed a 4 (And aul e al., 2001; Daniel e al., 2001; Wal e e al., 2004; Yagi e al., 2004; Wal e e al., 2006). S udies o (Mg,Fe,)SiO3 pe o ski e indica e alues in a simila ange (Kni le and Jeanloz, 1987; Mao e al., 1991).S udies on he e ec o Al subs i u ion on he elas ic p opoe ies o MgSiO3 pe o ski e, howe e , a e a mo e inconclusi e. A numbe o s udies ha e shown ha Al subs i u ion causes a dec ease in he bulk modulus o pe o ski e (Zhang and Weidne , 1999; Kubo e al., 2000; Daniel e al., 2001; Yagi e al., 2004). Some s udies indica e an inc ease in he bulk modulus (And aul e al. 2001; Ono e al., 2004), while se e al o he s indica e ha he e is no change in he bulk modulus wi h Al inco po a ion (Yagi e al., 2004; Jackson e al., 2004; Li e al., 2005). This disc epancy obse ed in elas ic p ope y beha iou migh be a ibu ed o di e en subs i u ion mechanisms ope a ing o e di e en p essu e o Al concen a ion anges. On he o he hand i may esul om pa ial amo phisa ion ollowing sample syn hesis o om he use o di e en measu emen echniques. Aside om he disc epancies in hese p e ious s udies i is likely ha in he Ea h he p esence o Fe3+ oge he wi h Al will a ec qui e di e en ly he beha iou o pe o ski e in compa ison 70 Chap e 3: Expe imen al de ails wi h Fe- ee samples. To da e he e a e no s udies ha ha e speci ically examined he e ec o a ying pe o ski e Fe3+ con en and bulk i on concen a ion on he elas ic p ope ies o Al-bea ing (Mg,Fe,)SiO3 pe o ski e. In his s udy single-c ys al X- ay di ac ion expe imen s ha e been pe o med a oom empe a u e in a diamond an il cell (DAC) o de e mine he equa ion o s a e pa ame e s o well-cha ac e ized samples o Fe and Al-bea ing pe o ski e, wi h di e en Fe3+ con en s. 3.2 Expe imen al de ails 3.2.1 S a ing ma e ials Th ee s a ing composi ions we e employed o he syn hesis o pe o ski e single c ys als in his s udy (Table 3.1). The alumina con en s o hese composi ions we e ixed o a alue o 5 w %, conside ed o be ypical o pe o ski e in he lowe man le, while he bulk Fe concen a ion was a ied. Table 3.1 The glass s a ing composi ions in w % oxide and ca ion p opo ion used o Fe and Al- bea ing pe o ski e syn hesis. Ca ion p opo ions a e calcula ed based on 3 oxygens pe o mula uni . No. SiO2 MgO FeO Al2O3 To al Si Mg Fe Al To al X Fe 1 54.58 36.64 3.63 5.15 100 0.929 0.935 0.052 0.103 2.019 0.05 2 52.96 35.13 6.76 5.17 100.02 0.916 0.911 0.097 0.105 2.031 0.1 3 52.8 32.07 10.09 5.1 100.06 0.925 0.843 0.147 0.1052 2.021 0.15 S a ing powde s o high-pu i y SiO2, MgO, Al2O3 and Fe2O3 we e g ound oge he in he p opo ions indica ed in Table 3.1, hen placed in a P c ucible and used in ai a 1600°C o 15 minu es. The c ucible was han apidly quenched in wa e o p oduce a silica e glass. The glass was g ound o a powde , hen pelle ised and educed in a CO2-H2 gas-mixing u nace main ained a an O2 o app oxima ely 2 log uni s abo e he i on- wüs i e bu e a 650°C o 24 hou s. The low empe a u es p e en ed he glass om c ys allizing. Composi ional cha ac e iza ion o he glasses was pe o med using an elec on mic op obe. The analysis condi ions we e he same as lis ed in Chap e 2 sec ion 71 Chap e 3: Expe imen al de ails 2.3. Mössbaue analysis on he glass powde s con i med he comple e educ ion o Fe2O3 o FeO. 3.2.2 Mul ian il syn hesis expe imen s The glasses we e ans o med o pe o ski e using a mul i-an il p ess a 25 GPa and 1800- 2000°C. Mul i-an il expe imen s we e pe o med using a p essu e medium o C 2O3 doped MgO oc ahed a o 8 mm edge leng h in combina ion wi h ungs en ca bide cubes wi h 3 mm unca ion edge leng hs. The p essu e assembly employed in his s udy is as shown in Fig. 3.2. The educed glass powde s we e packed in o Re o Au gold oil capsules o 2 mm leng h and 1 mm diame e . The hea ing du a ion o he expe imen s was a ied om an hou o en minu es and expe imen s we e quenched by cu ing he powe supply. No he mocouple was employed in he expe imen s bu ins ead a p e ious calib a ion o empe a u e e sus powe was used o p o ide an es ima e o he empe a u e. I has been ound ha in many expe imen s whe e a he mocouple was employed, i has led o leakage o pa ial mel om he capsule and ailu e o he expe imen due o damage caused by compac ion o he capsule by he ha d Al2O3 he mocouple slee e. The p esence o small deg ee pa ial mel s is key o he g ow h o la ge single c ys als. In expe imen H2369, mel ing occu ed due o he high empe a u es and high Fe con en , while in un H2438, H2O was added o lux mel ing (See Appendix B Table B.1 o un de ails). 72 Chap e 3: Expe imen al de ails MgO + 5 w % C O oc ahed a 23 Z O slee e 2 LaC O u nace 3 MgO slee e Sample in a me al oil capsule MgO od Figu e 3.2: The 8/3 mul ian il p essu e assembly used o mul ian il syn hesis expe imen s (no o scale) as desc ibed in he ex . Fo he c ys als o he composi ions wi h XFe = 0.10 (C ys al 1, Run no. H2369) and 0.15 (C ys al 2, Run no. H2438) we eco e ed single c ys als o size up o a maximum o 150 mic ons in leng h on decomp ession o he expe imen as shown in Fig. 3.3 (A) and (B). As H2O was used as lux in he syn hesis o C ys al 1, FTIR (Fou ie ans o m in a ed) spec oscopic measu emen s we e ca ied ou on hese pe o ski e c ys als o check o he p esence o wa e . No e idence o he inco po a ion o OH- in he pe o ski e sample was de ec ed, in acco dance wi h p e ious measu emen s (Bol an- Casano a e al., 2000). 73 Chap e 3: DAC expe imen s he e is li le e idence, would p oduce an oxygen excess composi ion due o Fe3+ o Al3+ subs i u ion o Mg2+ o Fe2+. As can be seen in Fig. 3.8, c ys als 2 and 3 plo along he cha ge coupled subs i u ion mechanism join. C ys als 1 and 4, howe e , which ha e he lowes i alen ca ion con en s, plo be ween he cha ge coupled and he oxygen acancy join. Oxygen acancies a e also indica ed by he o mulas o hese pe o ski e c ys als ha con ain less han 3 oxygens when no malized o 2 ca ions. 0.25 0.50 0.75 1.00 0.25 0.50 0.75 1.00 0.25 0.50 0.75 1.00 mol% SiO2 MgO+FeO Al2O3+Fe2O3 Oxygen excess join Cha ge coupled join Oxygen acancy join C ys al 1 C ys al 2 C ys al 3 C ys al4 mol% mol% Figu e. 3.8: A e na y concen a ion plo wi h he axes MgO + FeO, Al2O3+Fe2O3 and SiO2 showing he composi ions o he (Fe,Al) MgSiO3c ys als used in he p esen s udy in ela ion o di e en possible subs i u ion mechanisms. 3.3 Diamond an il cell comp ession expe imen s 3.3.1 Basics o diamond an il cell single c ys al comp ession echnique In his s udy, a BGI-design diamond an il cell (Allan e al., 1996) has been employed o pe o m s a ic comp ession expe imen s. Fo he high-p essu e expe imen s a single c ys al is placed in a hole d illed h ough a ha dened me al oil (gaske ), which is inden ed 80 Chap e 3: DAC expe imen s be ween wo opposed diamond an ils wi h la pa allel aces. This p oduces a p essu e chambe as shown in Fig.3.9. A p essu e calib an is placed along side he sample o p essu e de e mina ion wi hin he chambe and he ee olume in he p essu e chambe is illed wi h a p essu e- ansmi ing luid o main aining hyd os a ic condi ions. P essu e is applied o he sample by mechanically p essing he diamonds oge he and single c ys al di ac ion lines a e eco ded using a 4-ci cle di ac ome e . De ails o he echnique can be ound in Mile ich e al., (2000) and Mile ich, (2005). Figu e. 3.9: A schema ic sec ional iew o diamond an il cell as is used o ou comp ession s udy. 3.3.2 Comp ession expe imen s: Single c ys als wi h well-de ined habi s we e selec ed by ca e ul obse a ion unde an op ical mic oscope. Mos c ys als we e oo opaque o base his selec ion on ex inc ion and anspa ency. The inal selec ion o he c ys als was pe o med by checking he e lec ion in ensi y and peak p o iles using a Hube single c ys al di ac ome e . Single c ys als wi h dimensions o app oxima ely 120 µm x 80µm x 20 µm (C ys al 1), 130 µm x 100 µm x 50 µm (C ys al 2), and 120 µm x 90µm x 30 µm 81 Chap e 3: DAC expe imen s (C ys al 3) we e loaded in o a diamond an il cell sample chambe d illed in o a s eel (T301) pla e gaske p e-inden ed down o 90 µm dep h. Diamonds wi h 600 µm cule s we e employed and he sample chambe was 300 µm in diame e . Qua z was used as an in e nal p essu e calib an and a 4:1 mix u e o me hanol:e hanol was used as he p essu e ansmi ing medium o ou s udy. The uni cell pa ame e s we e de e mined a ambien empe a u e a a ious p essu es up o 9.16 GPa o C ys al 1, 7.10 GPa o C ys al 2 and 6.97 GPa o C ys al 3 on a Hube ou -ci cle di ac ome e (non-monoch oma ised MoKα adia ion) using he 8-posi ion cen ing p ocedu e acco ding o King and Finge (1979) and Angel e al., (2000). The maximum p essu e eached du ing each expe imen was de e mined by ailu e o he gaske and consequen loss o he pe o ski e single-c ys als, he e o e i was no possible o collec da a du ing decomp ession. The cen ing p ocedu e and ec o leas squa e e inemen o he uni cell cons an s we e pe o med by SINGLE04 so wa e acco ding o he p o ocol o Ralph and Finge , (1982) and Angel e al., (2000). The uni cell da a a di e en p essu es a e abula ed in Table 3.3. Table 3.3. Uni -cell la ice pa ame e s o (Fe, Al)- MgSiO3 single c ys als. S anda d de ia ions a e in pa en heses in e ms o leas uni s ci ed. P (GPa) a(Å) b(Å) c(Å) V((Å) C ys al 1 0.00010(1) 4.78638(19) 4.94261(16) 6.9188(4) 163.680(12) 0.295(6) 4.7844(2) 4.94052(19) 6.9161(5) 163.479(15) 1.063(6) 4.77899(18) 4.93630(15) 6.9085(4) 162.976(12) 1.982(7) 4.7732(2) 4.9310(2) 6.8995(6) 162.392(17) 2.818(9) 4.7674(2) 4.9265(3) 6.8916(6) 161.860(18) 3.991(11) 4.7595(2) 4.9198(3) 6.8800(8) 161.10(2) 4.611(11) 4.7558(2) 4.9165(2) 6.8748(6) 160.747(19) 6.501(10) 4.7441(2) 4.9066(2) 6.8574(7) 159.62(2) 7.944(10) 4.73509(18) 4.89956(16) 6.8453(5) 158.810(14) 8.785(18) 4.7306(4) 4.8953(4) 6.8376(10) 158.34(3) 9.16(2) 4.7284(4) 4.8934(3) 6.8342(9) 158.13(3) C ys al 2 0.00010(1) 4.79239(14) 4.95152(11) 6.9343(4) 164.548(11) 0.350(5) 4.79005(15) 4.94953(11) 6.9311(5) 164.327(13) 82 Chap e 3: DAC expe imen s 0.681(5) 4.78789(16) 4.9480(2) 6.9278(6) 164.123(16) 1.204(4) 4.78423(18) 4.94443(15) 6.9222(6) 163.747(15) 1.685(4) 4.78099(14) 4.94197(15) 6.9179(6) 163.455(14) 2.455(5) 4.77571(15) 4.93737(15) 6.9100(5) 162.933(13) 3.491(6) 4.76913(11) 4.93217(14) 6.9005(5) 162.315(12) 4.728(7) 4.76121(16) 4.92566(17) 6.8893(6) 161.568(15) 5.420(6) 4.75672(9) 4.92167(9) 6.8825(4) 161.127(9) 6.429(9) 4.75072(12) 4.91680(10) 6.8734(4) 160.552(10) 6.910(8) 4.74773(9) 4.91416(9) 6.8695(3) 160.272(9) 7.101(9) 4.74647(13) 4.91332(12) 6.8682(4) 160.173(12) C ys al 3 In ai 4.79977(10) 4.97947(9) 6.98054(11) 166.837(5) 0.00010(1) 4.79978(10) 4.97961(11) 6.9799(3) 166.826(10) 0.396(4) 4.79701(11) 4.97753(13) 6.9756(4) 166.558(11) 0.513(5) 4.79639(12) 4.97699(14) 6.9745(4) 166.492(11) 0.745(5) 4.79465(14) 4.97563(16) 6.9718(5) 166.321(14) 1.176(5) 4.79181(13) 4.97311(16) 6.9671(4) 166.029(13) 1.510(5) 4.78929(17) 4.97096(18) 6.9634(6) 165.781(17) 1.943(5) 4.78665(14) 4.96852(16) 6.9591(5) 165.504(14) 2.334(5) 4.78420(13) 4.96607(14) 6.9545(4) 165.231(12) 2.756(7) 4.78122(13) 4.96370(16) 6.9502(4) 164.950(12) 3.302(7) 4.77772(13) 4.96071(13) 6.9447(4) 164.596(11) 4.285(7) 4.77151(11) 4.95556(12) 6.9346(4) 163.973(10) 4.540(8) 4.76976(12) 4.95431(12) 6.9323(4) 163.817(10) 4.930(7) 4.76740(12) 4.95202(13) 6.9283(4) 163.566(11) 5.463(7) 4.76404(11) 4.94915(13) 6.9231(4) 163.232(11) 5.969(9) 4.76110(18) 4.9465(2) 6.9188(6) 162.942(17) 6.423(8) 4.75831(12) 4.94434(13) 6.9145(4) 162.674(11) 6.729(8) 4.75623(13) 4.94286(14) 6.9115(4) 162.486(12) 6.968(8) 4.7550(2) 4.9420(3) 6.9089(8) 162.36(2) 83 Chap e 3: Equa ion o s a e esul s 3.4 Equa ion o s a e esul s 3.4.1. Theo e ical backg ound The incomp essibili y o a ma e ial is exp essed in e ms o i s bulk modulus K, which is de ined as K = -V(δP/δV). The bulk modulus is also a unc ion o p essu e and can be desc ibed by he p essu e de i a i e K' = δK/δP and po en ially also K'' = δ2K/δP2. S a ic comp ession measu emen s a e pe o med a cons an empe a u e and a e he e o e desc ibed using he iso he mal bulk modulus K0. Comp ession in he Ea h, howe e , is no iso he mal bu adiaba ic and is he e o e desc ibed using he adiaba ic bulk modulus (KS). The ela ionship be ween he iso he mal and adiaba ic bulk moduli is KS = K0 (1+αγT) (7) whe e γ is he G uneisen pa ame e and α is he he mal expansion coe icien . The ela ionship be ween olume and p essu e is desc ibed using an equa ion o s a e (EoS). A numbe o EoS o mula ions exis (Angel, 2000; Du y and Wang, 2000) bu he mos commonly used, pa icula ly o desc ibe expe imen al da a on he comp ession o mine als, is he Bi ch–Mu naghan equa ion o s a e (Bi ch, 1947). This is based on he assump ion ha he s ain ene gy o a solid unde going comp ession can be exp essed as a Taylo expansion o he Eule ian s ain E = [(V0/V)2/3-1]/2. Expansion o he ou h-o de in E gi es he Bi ch-Mu naghan EoS: P = 3K0 E (1+2 E) 5/2{1+3/2(K' –4) E +3/2{K0K'' + (K'- 4)(K' –3)+35/9} E 2} (8) Howe e , o mos ype o ma e ials his EoS is unca ed a he second o de in s ain, which equi es K' o be ixed o 4 o a he hi d-o de in s ain which equi es he coe icien o E2 o be ze o wi h an implied alue o K'' = -1/K0(3- K')(4- K')+35/9} (Ande son, 1995). 84 Chap e 3: Equa ion o s a e esul s 3.4.2 Uni cell la ice pa ame e a ia ions wi h p essu e The changes in uni cell olume as a unc ion o p essu e o all h ee c ys als a e shown in Fig.3.10, whe e he ela i e olume (V/V0) is he olume no malized o he ambien p essu e olume. The maximum p essu e o each expe imen is cons ained by he poin whe e he gaske s a s o ail and di ac ion lines a e b oadened. This occu ed a sligh ly di e en p essu es in each o he h ee expe imen s wi h c ys al 1 achie ing he highes p essu e o jus abo e 9 GPa. Fo compa ison, da a o pu e MgSiO3 pe o ski e (Vanpe eghem e al., 2006) and MgSiO3 pe o ski e wi h 5-mol% Al2O3 (Zhang and Weidne , 1999) a e also plo ed in Fig. 3.10. 0246810 0.960 0.965 0.970 0.975 0.980 0.985 0.990 0.995 1.000 1.005 C ys al 1 XFe = 0.07 C ys al 2 XFe = 0.13 C ys al 3 XFe = 0.24 MgSiO3 MgSiO3- 5mol% Al203 Rela i e olume V/V0 P essu e (GPa) Figu e. 3.10:Uni cell olume a ia ion o (Fe,Al)-MgSiO3 single c ys als as a unc ion o p essu e o he 3 c ys als o p esen s udy. The solid cu e ep esen s he olume a ia ion o pu e MgSiO3 o Vanpe eghem e al., (2006) and he do ed cu e has been calcula ed using he EoS pa ame e s epo ed o a MgSiO3 pe o ski e wi h 5-mol % Al2O3 by Zhang and Weidne , (1999). Al hough he di e ences be ween he P-V da a collec ed in his s udy and hose o MgSiO3 pe o ski e (Vanpe eghem e al., 2006) a e e y small, he e is a sligh inc ease in comp essibili y wi h inc easing Fe con en . In addi ion, he Fe and Al-bea ing pe o ski es 85 Chap e 3: Equa ion o s a e esul s syn hesised in his s udy a e less comp essible han he Fe- ee Al-bea ing pe o ski e s udied by Zhang and Weidne , (1999). The axial comp essibili ies a/a0, b/b0, c/c0 o all he h ee c ys als as a unc ion o p essu e a e shown in Fig. 3.11. The axial comp ession is aniso opic wi h he b-axis being he leas comp essible. Fo C ys al 1 (XFe = 0.07) and C ys al 2 (XFe = 0.13) he axial comp essibili ies o he a and c axes a e i ually iden ical and a e e y close o hose o pu e MgSiO3 pe o ski e (Vanpe eghem e al., 2006). Fo C ys al 3 (XFe = 0.24) he di e ence in comp essibili y be ween he a and c axes is la ge , wi h a-axis being less comp essible han c-axis. As a consequence, he axial a io c/a emains cons an o C ys al 1 a 1.445 and o C ys al 2 a 1.446, whe eas i dec eases wi h p essu e o C ys al 3. The b/a axial a io ins ead inc eases wi h p essu e o all h ee c ys als. The axial comp essibili ies epo ed o he Fe- ee Al bea ing pe o ski es by Zhang and Weidne (1999) show he a-axis o be sligh ly mo e comp essible han he c-axis. This sugges s ha he e is a di e en mechanism by which Al is subs i u ed in o he pe o ski e s uc u e in he absence o in he p esence o Fe. 86 Chap e 3: Equa ion o s a e esul s 01234567891011 0.984 0.986 0.988 0.990 0.992 0.994 0.996 0.998 1.000 1.002 C ys al 3 a / a 0 b / b 0 c / c 0 Rela i eaxis P essu e in GPa 0246810 0.986 0.988 0.990 0.992 0.994 0.996 0.998 1.000 1.002 C ys al 1 a / a 0 b / b 0 c / c 0 Rela i e axis P essu e in GPa 01234567891011 0.986 0.988 0.990 0.992 0.994 0.996 0.998 1.000 1.002 C ys al 2 a / a 0 b / b 0 c / c 0 Rela i e axis P essu e in GPa Figu e 3.11. Axial comp essibili ies o he (Fe,Al)-MgSiO3 pe o ski e single c ys als. The solid cu es a e he 3 d-o de Bi ch Mu naghan EoS i s o he da a: (a) C ys al 1, XFe = 0.07 (b) C ys al 2, XFe = 0.13 and (c) C ys al 3, XFe = 0.24. 3.4.3 FE- E plo s and EoS pa ame e s A no malised s ess, de ined as FE = P/3 E (1+2 E)5/2 has been plo ed e sus he Eule ian s ain E o each c ys al (Fig. 3.12) in o de o ha e a isual assessmen o he o de o he Bi ch-Mu naghan EoS equi ed o i he comp ession da a (Angel, 2000). 87 Chap e 3: Equa ion o s a e esul s 0.000 0.002 0.004 0.006 0.008 0.010 0.012 230 240 250 260 No malised p essu e FE GPa Eule ian s ain E C ys al 1 BM III i BM II i 0.000 0.002 0.004 0.006 0.008 0.010 240 255 270 285 C ys al 2 BM III i BM II i No malised p essu e FE GPa Eule ian s ain E 0.000 0.002 0.004 0.006 0.008 0.010 220 230 240 250 260 C ys al 3 in DAC BM III i BM II i No malised p essu e FE GPa Eule ian s ain Figu e. 3.12 FE- E plo s based on he Bi ch-Mu naghan EoS o (Fe,Al)-MgSiO3 pe o ski e single c ys als. The alues o V0 used o calcula e he ini e s ain a e hose measu ed a 1 ba . The open ci cles shown in he FE- E plo o C ys al 3, a e he ini e s ain alues calcula ed wi h olume o he c ys al de e mined in ai (Equa ions o he weigh ed i s a e men ioned in he ex ). 88 Chap e 3: Equa ion o s a e esul s Weigh ed linea i s o da a poin s o he h ee c ys als indica e a posi i e slope, which implies ha K' is g ea e han 4. Fo compa ison, weigh ed ho izon al linea i s (implying K' = 4) also a e shown in Fig. 3.12, al hough in his case we obse e an inc ease o he chi-squa e. In he case o C ys al 3 he FE- E da a do no de ine a s aigh line, sugges ing ha a ou h-o de unca ion o he Bi ch-Mu naghan EoS migh be conside ed. Howe e , i has been shown (Angel, 2000) ha such cu a u e o small s ains, i.e., o da a collec ed a low p essu es, is o en due o a w ong alue o V0 being used in he calcula ion o he Eule ian s ain. I he alue o he olume measu ed wi h c ys al 3 glued on a glass ib e (i.e., in ai , Table 3.3), ins ead o ha measu ed wi h he same c ys al in he DAC is used o calcula e he F- plo , linea beha iou is ob ained (Fig. 3.12, C ys al 3 open ci cles). Mo eo e , he in e cep on he F axis is equal o he bulk modulus, K0, alue and can hence be used o assess he goodness o he EoS i p ocedu e. F- plo s ha e been calcula ed also o he indi idual c ys al axes and a e shown in Fig. 3.13 indica ing posi i e slope. Since he FE- E plo s (Fig. 3.12) sugges ha K' is la ge han 4 o all h ee c ys als, he P-V da a ha e been i ed wi h a 3 d o de Bi ch–Mu naghan EoS using he Eos i 52 p og am (Angel, 2000). Howe e , o compa ison wi h da a p esen in he li e a u e, i ed p ac ically uniquely wi h K' ixed o he alue o 4, a 2nd o de Bi ch–Mu naghan EoS has also been used. A linea ised Bi ch-Mu naghan EoS (implemen ed in he Eos i 52 so wa e) in which he cube o a uni -cell axis is used ins ead o he olume has been used o i he uni cell a, b and c pa ame e s o he h ee c ys als. The bulk moduli so ob ained a e 00 31 β = K, whe e β 0 is he axial comp essibili y. The esul ing equa ion o s a e pa ame e s a e epo ed in Table 3.4. The alues ob ained a e in good ag eemen wi h hose ob ained om he linea i s o he FE- E plo s. 89 Chap e 3: Discussion 01234567891011 18.0 18.5 19.0 19.5 20.0 20.5 21.0 C ys al 1 C ys al 2 C ys al 3 MgSiO3 Dis o ion angle Φ(°) P essu e in GPa Figu e 3.17: Dis o ion angle Φ o (Fe,Al)-MgSiO3 pe o ski es as a unc ion o composi ion and p essu e. The o ho hombic dis o ion is ound o inc ease wi h inc easing p essu e o all composi ions. Pu e MgSiO3 da a a e also plo ed o a compa ison om Vanpe eghem e al., (2006). 3.5.Discussion 3.5.1 The Elas ici y o he Ea h’s lowe man le Assuming a py oli e bulk man le composi ion, he Ea h’s lowe man le should be composed o app oxima ely 80- olume % magnesium silica e pe o ski e coexis ing wi h magnesiowüs i e and calcium silica e pe o ski e. Compa ison o seismically-in e ed p ope ies o he lowe man le wi h model p ope ies based on measu ed bulk and shea moduli o such a man le assemblage p o ides he bes es o he composi ion and empe a u e o he lowe man le. The elas ici y da a de e mined om single c ys al comp ession expe imen s ha e been combined wi h exis ing he moelas ic da a in o de o calcula e he a ia ion in densi y and bulk modulus o e he dep h ange o 670–2571 km, which co e s he en i e lowe man le excluding he D’’ low eloci y zone. A numbe o s udies ha e epo ed he moelas ic p ope ies o silica e pe o ski e om expe imen s conduc ed a simul aneous high p essu e and empe a u e (Mao e al., 1991; Yeganeh–Hae i, 1991; 96 Chap e 3: Discussion Wang e al., 1994; U sumi e .al, 1995; Funamo i e al., 1996; Fique e al., 1998; Fique e al., 2000). Fo hese calcula ions we ha e used he high empe a u e p ope ies om he da a se o Funamo i e al., (1996) [ K0,T = 261 GPa, K'0,T = 4, (dK/dT)P =-0.028 Gpa K-1 , α0 = 1.98*10-5 K -1, α1 = 0.82*10-8 K-2 and α2 = -0.47(K)] which we e de e mined o MgSiO3 pe o ski e. The oom p essu e olume (V0) a high empe a u e was calcula ed using he mal expansion coe icien s and he equa ion ∫ = T TTT dTVV 298 0,0,,0 exp α (9) In iew o he lack o high- empe a u e elas ic p ope y da a o Al-bea ing pe o ski es and he limi ed da a o Fe-bea ing pe o ski es, we ha e no op ion bu o assume ha hey a e he same as o he MgSiO3 end membe . Mao e al., (1991) p oposed ha he mal expansi i i es o pu e and Fe-bea ing pe o ski e we e simila , bu ha he empe a u e dependence o he bulk modulus o Fe-bea ing pe o ski e was g ea e , i.e., (dK/dT)P = -6.3*10-2 Gpa K-1. In o de o examine his p oposal we ha e also made calcula ions assuming he alue o (dK/dT)P epo ed by Mao e al., (1991). Almos all p e ious s udies on elas ic p ope ies o MgSiO3 pe o ski e ha e i ed he expe imen al da a assuming a second o de Bi ch-Mu naghan equa ion o s a e, i.e., wi h K' ixed a 4. Ou da a, howe e , p o ide s ong e idence ha , a highe Al and Fe concen a ions a hi d-o de Bi ch-Mu naghan equa ion o s a e is equi ed wi h a K' highe han 4. To ob ain an o e iew o how ixing K' o 4 and K' > 4 can e ec calcula ed lowe man le p ope ies; calcula ions ha e been pe o med using bo h 2nd and 3 d o de i s o ou da a se . All he calcula ions we e pe o med along he adiaba ic geo he m o B own and Shankland, (1981) ixed a 1873 K o a dep h o 670 km. Densi ies calcula ed along an adiaba ic empe a u e g adien o he h ee c ys als a e shown as a unc ion o p essu e in he man le in Fig. 3.18. When K' is ixed a 4 and (dk/dT)P = -0.028 GPa K-1 (Funamo i e al., 1996), he e is a clea inc ease in densi y wi h inc easing Fe and Al concen a ion a lowe man le condi ions. Densi ies o C ys als 1 and 2 a e only sligh ly la ge han hose calcula ed o MgSiO3 pe o ski e h oughou mos o he lowe man le. Howe e , he sligh ly lowe bulk moduli o C ys als 1 and 2 means ha hei densi ies become smalle han MgSiO3 pe o ski e close o he co e man le bounda y. The di e ence in densi ies be ween C ys als 1 and 2 also dec eases h oughou he lowe man le due o he mino di e ences in hei bulk moduli. The densi y o C ys al 3, which 97 Chap e 3: Discussion has he highes Fe con en emains signi ican ly la ge han he o he wo c ys als h oughou he lowe man le e en hough i has a lowe bulk modulus. When (dK/dT)P = - 0.063 GPa/K, as p oposed by Mao e al., (1991), is employed o his calcula ion, he e ec o Fe and Al on densi ies in he lowe man le becomes much s onge as a esul o he bulk moduli being lowe . The P elimina y Re e ence Ea h Model (PREM), (Dziewonski and Ande son, 1981), densi y cu e is qui e consis en wi h a monomine alic pe o ski e lowe man le, bu a pe ec ma ch would only be ob ained i he lowe man le inc eased in Fe con en owa ds he base. Based on simila compa isons, p e ious s udies ha e a gued ha he lowe man le may ha e a highe Si/Mg a io han he uppe man le and he e o e be composed mos ly o pe o ski e (Ande son e al., 1995). La ge payo s exis , howe e , be ween he alue o (dK/dT)P used and he p opo ion o coexis ing magnesiowüs i e in he modelled assemblage (Fique e al., 1998). As discussed p e iously, he comp ession da a epo ed in his chap e a e ac ually consis en wi h alues o K' >4 o Fe and Al bea ing pe o ski e, and K' was also ound o inc ease wi h inc easing Fe con en . In he lowe panel o Fig. 3.18 densi y calcula ions ha e been pe o med using he esul s o he 3 d o de Bi ch-Mu naghan equa ion o s a e i s. The esul is a much smalle change in densi y in he lowe man le, compa ed o oom p essu e, as a esul o Fe and Al subs i u ion, due o he highe alues o K' which make Fe- ich pe o ski es less comp essible a lowe man le condi ions. Ve y la ge changes in Fe concen a ion in he lowe po ion o he lowe man le would be equi ed o ma ch he changing slope o he PREM densi y cu e when K' > 4. 98 Chap e 3: Discussion 20 40 60 80 100 120 4.2 4.4 4.6 4.8 5.0 5.2 5.4 5.6 5.8 6.0 6.2 Wi hK'>4 Wi h Funamo i e al. 1996 da a C ys al 1 C ys al 2 C ys al 3 MgSiO3 PREM Wi h Mao e al. 1991 da a C ys al 1 C ys al 2 C ys al 3 Densi y (gm/cm 3 ) P essu e (GPa) 20 40 60 80 100 120 4.2 4.4 4.6 4.8 5.0 5.2 5.4 5.6 5.8 6.0 6.2 Wi h K' = 4 Wi h Funamo i e al.1996 da a C ys al 1 C ys al 2 C ys al 3 MgSiO 3 PREM Wi h Mao e al. 1991 da a C ys al 1 C ys al 2 C ys al 3 Densi y (gm/cm 3 ) P essu e (GPa) Figu e 3.18: Densi y p o iles calcula ed o he h ee c ys als wi h K' = 4 and > 4 as de e mined om he comp ession da a o his s udy. Two (dk/dT)P alues we e used o compa ison as s a ed in he ex . PREM da a o Dziewonski and Ande son, (1981) and pu e MgSiO3 densi y calcula ed based on comp ession da a o Vanpe eghem e al., (2006), a e also plo ed o compa ison. We ha e also calcula ed KT, he bulk modulus a p essu e and empe a u e, o he h ee c ys als o his s udy, using he equa ion: K T,0= K 298,0 + (dKT,0/dT)P(T-298) (10) 99 Chap e 3: Discussion whe e, KT,0 is he bulk modulus a empe a u e, K298,0 is he bulk modulus a ambien condi ions and (dKT,0/dT)P is he empe a u e de i a i e o he bulk modulus, aken om ei he Funamo i e al., (1996) o Mao e al., (1991), and he equa ion K T= K T,0 + (dKT,0/dP)TP (11) whe e (dKT,0/dP)T is he p essu e de i a i e o he bulk modulus. An es ima e o he alue o KT o he lowe man le can be ob ained om PREM KS ( he adiaba ic bulk modulus) es ima es using he equa ion: KT=KS/(1+αγT) (12) whe e, α is he he mal expansi i y and γ is he g uneisen pa ame e , alues o which a e abula ed o he lowe man le in B own and Shankland, (1981). Rega dless o he alue o (dKT,0/dT)P employed, KT cu es calcula ed o pe o ski e samples assuming K' > 4 display a much g ea e di e gence as a unc ion o composi ion han when K' is assumed o be 4 (Fig. 3.19). This is he opposi e beha iou om ha obse ed o densi ies, whe e he g ea es e ec o Fe con en occu s o K' = 4. Ve y la ge changes in Fe con en would be equi ed, on he o he hand, o in luence KT in he lowe man le i K' = 4. In addi ion, he slope o he PREM alue o KT is shallowe han hose calcula ed o he h ee c ys als o his s udy i K' > 4, which means ha Fe con en s would ha e o dec ease in he lowe po ion o he lowe man le o be consis en wi h he PREM slope. Again his is he opposi e conclusion o ha ob ained based on densi y. KT de e mina ions wi h K' > 4 a e much highe a he base o he man le compa ed wi h PREM and p o ide a poo i o he obse a ions compa ed o when K' = 4. The addi ion o magnesiowüs i e o he model assemblage would educe KT bu no by mo e han 10 % (Fique e al., 1998). Unde hese ci cums ances, alues o K' = 4 and (dk/dT)P = -0.028 GPa K-1 gi e he bes ag eemen wi h he PREM alue o KT o lowe man le. 100 Chap e 3: Discussion 40 60 80 100 120 200 300 400 500 600 700 800 900 1000 (dK/dT) P = -0.028GPaK -1 Wi h K' = 4 C ys al 1 C ys al 2 C ys al 3 Wi h K' > 4 C ys al 1 C ys al 2 C ys al 3 MgSiO 3 (K' = 4) PREM K T (P) P essu e in GPa 40 60 80 100 120 200 300 400 500 600 700 800 900 1000 (dK/dT) P = -0.063GPaK -1 Wi h K'= 4 C ys al 1 C ys al 2 C ys al 3 Wi h K' > 4 C ys al 1 C ys al 2 C ys al 3 MgSiO 3 K' = 4 PREM K T (P) P essu e in GPa Figu e 3.19: Compa ison o bulk modulus KT(P) p o iles along wi h PREM lowe man le model (Dziewonski and Ande son, 1981) along he empe a u e p o ile o B own and Shankland, (1981). Bulk modulus o MgSiO3 is also plo ed o a compa ison based on da a o Vanpe eghem e al., (2006). KT p o ile o he h ee c ys als calcula ed using (dk/dT)P o –0.028 GPa K-1 and -0.063 GPa K-1 alues as epo ed by Funamo i e al., (1996) o a MgSiO3 composi ion and Mao e al., (1991) o a Fe-bea ing MgSiO3 composi ion bo h o a K' o 4 and K' > 4 as ob ained in his s udy by 3 d o de Bi ch-Mu naghan i . 101 Chap e 3: Discussion 3.5.2 The e ec o p essu e on pe o ski e subs i u ion. OAlFe VIVIII 3 33 ++ Expe imen al obse a ions indica e (F os e al., 2004) ha e en a he lowes plausible oxygen ugaci y o he lowe man le, (Fe,Mg)(Al,Si)O3 pe o ski e con ains a signi ican Fe3+ con en (Fe3+/∑Fe > 0.5). As he uppe man le has a e y low Fe3+/∑Fe a io (<0.03) wo possible scena ios o he edox s a e o he lowe man le can be p oposed: ei he he lowe man le is mo e oxidized han he uppe man le o a lowe man le condi ions Fe2+ is oxidised o Fe3+ by some agen also p esen in he bulk man le. E idence o whole man le con ec ion con adic s he i s possibili y, as i would ha e been impossible o main ain a low Fe3+/∑Fe a io in he uppe man le i i we e mixed h ough geologic ime wi h a highly oxidised lowe man le ese oi . Al hough he second possibili y is he e o e mo e likely, none o he oxidising agen s ac i e in he uppe man le (e.g., CO2, SO2) a e abundan enough o p oduce he equi ed Fe3+ con en s. In he absence o such an agen , disp opo iona ion o FeO i.e., 3FeO = Fe+Fe2O3. (13) becomes he only mechanism capable o p oducing Fe3+ in su icien abundance. F os e al., (2004) showed expe imen ally ha abou 1 w % o Fe me al would be equi ed o balance he Fe3+ equi emen o pe o ski e a 25 GPa wi hin a py oli ic lowe man le bulk composi ion. The equilib ium be ween (Fe,Mg)(Al,Si)O3 pe o ski e and Fe me al can be desc ibed by he equa ion: AlAlO3 + 3FeO = Fe + 2AlFeO3 (14) Pe o ski e Mw Me al Pe o ski e whe e, Mw is magnesiowüs i e. F os e al., (2004) measu ed high concen a ions o he AlFeO3 componen in pe o ski e in equilib ium wi h me allic Fe and concluded ha his equilib ium mus be shi ed s ongly o he igh a condi ions compa ible wi h he op o 102 Chap e 3: Discussion he lowe man le. By assessing he e ec o ( ), ( Fe ) and ( Fe ) subs i u ions on he olume o (Fe,Mg)(Al,Si)O O AlAl VIVIII 3 33 ++ Fe Si Mg AlAl VI VIII VIVIII + + ++ +↔+ 4 2 33 OSiVIVIII 3 42 ++ MgVIII VIII + +↔2 2 OAlFe VIVIII 3 33 ++ Si Mg Al VI VIII VIVIII + + ++ +↔+ 4 2 33 3 pe o ski e, i should be possible o de e mine whe he disp opo iona ion o FeO will be a ou ed a he highe -p essu e condi ions o he deep lowe man le. The olume and comp ession esul s p esen ed in his chap e p o ide da a ha help o unde s and he likely olume e ec . 24.4 24.6 24.8 25 25.2 25.4 25.6 0 5 10 15 20 25 30 AlAlO3 (Wal e e al., 2004, 2006; Anda ul e al., 2001;Kubo e al., 2000) Nishio-Hamane e al., (2005) P esen s udy FeSiO3 (Yagi e al., 1979; Fei e al., 1996; Anda ul e al., 2001) Vanpe eghem e al., (2006) mola olume (cm3mol-1) mol% AlAlO3, FeSiO3, FeAlO3 MgSiO3 AlAlO3 Fe3+AlO3 (4) (4) (4) FeSiO3 0 (8) (3) Fe3+AlO3 Figu e 3.20. E ec o , , subs i u ions on he mola olume o magnesium silica e pe o ski e. OSiFe VIVIII 3 42 ++ O AlAl VIVIII 3 33 ++ OAlFe VIVIII 3 33 ++ Fig. 3.20 shows he change in mola olume o MgSiO3 pe o ski e as a unc ion o possible subs i u ion mechanisms de e mined using oom p essu e olume da a om his and p e ious s udies. The addi ion o and componen s ha e simila e ec s on he mola olume o pe o ski e wi h he o me ha ing a sligh ly g ea e e ec . O AlAl VIVIII 3 33 ++ OSiFe VIVIII 3 42 ++ 103 Chap e 3: Discussion Al hough he pe o ski e samples syn hesised in his s udy ha e high Fe3+/ΣFe a ios, all samples con ain Fe2+ and he e o e do no lie on a composi ional join be ween MgSiO3 and . All single c ys als syn hesised in his s udy also con ain 4-mol % o , which is indica ed by he igu e in b acke s nex o each da a poin . These da a show ha in compa ison o and subs i u ions he (Fe,Mg)(Al,Si)O OAlFe VIVIII 3 33 ++ OSiVI 3 4+ FeVIII 2+ OAlVI 3 3+ OSiVI 3 4+ O AlAl VIVIII 3 33 ++ 3 OSiFe VIVIII 3 42 ++ SiFeVIII 2+ 3 pe o ski es ha e la ge mola olumes. The only sample wi h a composi ion on he MgSiO3- join is ha syn hesised by Nishio-Hamane e al., (2005) a app oxima ely 50 GPa in a diamond an il cell. Al hough he Fe OAlFe VIVIII 33 ++ OSiVIVIII 3 42 ++ 3+ con en was no independen ly con i med, i was assumed o be he same as ha o he s a ing composi ion. The olume o his sample is smalle han he sample syn hesised in ou s udy a a simila con en and a end s a s o eme ge pa icula ly when he olumes and con en s o (Fe,Mg)(Al,Si)O FeVIII 3+ FeVIII 2+3 pe o ski e samples syn hesised by Vanpe eghem e al., (2006) a e also conside ed. Inc easing he con en o (Fe,Mg)(Al,Si)O O VI 3 4+ 3 pe o ski e samples inc eases hei mola olumes d ama ically. In o he wo ds, he subs i u ion o in o Al- ee pe o ski e has a smalle e ec on mola olume han when i subs i u es in o Al-bea ing pe o sk ie. As pe o ski e mus con ain Al in he lowe man le he la ge e ec on he mola olume should make Fe Fe 2+ subs i u ion in pe o ski e un a ou able wi h inc easing p essu e. Using da a in Fig. 3.20 and comp ession da a om his s udy and he li e a u e (Fei, 1996; Wal e e al., 2004) he mola olume change o he equilib ium (eq. 14), a p essu es o he lowe man le is calcula ed o be app oxima ely –2 cm3/mol. This implies ha disp opo iona ion o FeO should be a ou ed wi h inc easing p essu e and he e o e is likely o ake place h oughou he pe o ski e-bea ing egion o he lowe man le. In addi ion, howe e , he la ge e ec on he pe o ski e mola olume o subs i u ion in o (Fe,Mg)(Al,Si)O OSiFe VIVIII 3 42 ++ 3 pe o ski e should d i e Fe2+ ou o pe o ski e wi h inc easing p essu e and in o magnesiowüs i e. This should cause pe o ski e o dec ease in o al Fe con en wi h p essu e bu o become mo e Fe3+ ich. Some suppo o his can be ound in he ac ha while in his s udy i was no possible o syn hesise pu e Fe3+ bea ing pe o ski e a 25 GPa, as a emp ed in he syn hesis o c ys al 104 Chap e 3: Conclusions 3 which con ained only Fe3+ in he s a ing composi ion, his was possible in he s udy o Nishio-Hamane e al., (2005) pe o med a app oxima ely 50 GPa. 3.6 Conclusions The subs i u ion Al and Fe3+ in o MgSiO3 pe o ski e occu s by a coupled subs i u ion mechanism wi h he possibili y ha small concen a ions o oxygen acancies a e p esen only a low i alen ca ion concen a ions. S a ic comp ession measu emen s show ha he subs i u ion o Al and Fe in o MgSiO3 pe o ski e inc eases he comp essibili y. The axial comp ession is aniso opic wi h he b axis being he leas comp essible, and a and c axes ha ing i ually iden ical comp essibili ies. Axial comp essibili ies o Fe- ee Al-bea ing pe o ski es (Zhang and Weidne , 1999) show he a axis o be sligh ly mo e comp essible han he c axis implying ha he e is a di e en mechanism by which Al is subs i u ed in o he pe o ski e s uc u e in he absence o in he p esence o Fe. Fi ing he da a o a 2nd o de Bi ch-Mu naghan EoS (K' = 4) gi es alues o he bulk modulus ha dec ease om 247 o 240 GPa as he Fe con en inc eases om 0.12 o 0.22 o mula uni s. These alues a e lowe han he alue o 253 GPa epo ed o MgSiO3 pe o ski e (Vanpe eghem e al., 2006) based on single c ys al de e mina ions. Howe e , - F plo analyses indica e ha a 3 d o de Bi ch- Mu naghan i o he da a is s a is ically jus i ied. These i s indica e a la ge d op in bulk modulus wi h Fe and Al subs i u ion om 243 o 234 GPa and an inc ease in K' om 5 o 6.5. The dec ease in bulk modulus likely a ises om an inc ease in polyhed al comp essibili y. Combining hese da a wi h he moelas ic da a om he li e a u e indica es ha pe o ski e densi ies calcula ed along a lowe man le geo he m a e insensi i e o Fe-Al subs i u ion i he 3 d o de Bi ch- Mu naghan i o he da a is employed. The esul ing alues o K' om his i which a e >4 p o ide a poo i o he es ima ed bulk modulus o he lowe man le. I on he o he hand he 2nd o de Bi ch- Mu naghan i is employed he calcula ed densi ies when compa ed o PREM a e consis en wi h an inc ease in Fe-Al subs i u ion in he lowe po ions o he lowe man le. 105 Chap e 4: S a ing ma e ial syn hesis MgO ce amic is inse ed om one side o he assembly h ough he molybdenum elec ode, so ha he junc ion is loca ed jus abo e he sample. Syn hesis expe imen s we e pe o med be ween 17-20 GPa and 1000-1600°C (Table 4.1). The p essu e calib a ion cu es o F os e al., (2004) we e adop ed o de e mina ion o p essu e in ou expe imen s. The hea ing du a ion a ied om one o h ee hou s and samples we e quenched by cu ing he powe supply o he u nace. C doped MgO p essu e medium Z O slee e 2 Re capsule LaC O u nace 3 MgO slee e LaC O disc 3 MgO disc MgO od MgO slee e Molybdenum slee e The mocouple Al O 23 Coppe coil Figu e. 4.2: A diag amma ic ep esen a ion o he box u nace assembly is shown he e. The sample was con ained in a Re capsule as shown su ounded by a MgO slee e and a LaC 2O3 u nace. The molybdenum ubes se e as con ac s o hea ing. This se up ensu es low he mal g adien in he la ge sample olume. Table 4.1: Syn hesis condi ions o calo ime ic samples used in his s udy. Pis on cylinde and a mul ian il p ess we e used o syn hesise hem. The uni cell pa ame e s de e mined by powde X- ay di ac ion (as discussed la e in his sec ion), o ga ne s a e also lis ed he e. Ga ne s wi h cubic symme y ha e a=b=c and o e agonal symme y ha e a = b ≠ c whe e a, b, c a e c ys al axes so, as a and b axes a e equal, only, a and c a e lis ed he e. Run no. Composi ion P essu e P essu e Tempe a u e Uni cell pa ame e s (in Å uni s) assembly GPa °C a-axis c-axis Z493 Py 40 18/8 19 1000 11.476 Z494 Py 20 18/8 19 1600 11.483 Z496 Py 10 18/8 18 1600 11.497 11.457 Z501 Py 30 18/8 17 1600 11.484 Z507 Majo i e 18/8 19.5 1750 11.517 11.433 Z508 Py 80 18/11 17.5 1200 11.463 Z525 Py 15 18/8 18 1400 11.482 PC Py ope 1/2inch 3 1200 11.460 112 Chap e 4: S a ing ma e ial syn hesis Small chips o he eco e ed samples we e embedded in epoxy esin o quan i a i e chemical analysis using a JEOL JXA-8200 WD/ED elec on mic op obe ope a ing in wa eleng h dispe si e mode wi h a poin beam a 15 nA cu en and 15 kV. S anda ds employed we e and adi e o Si, spinel o Al and ens a i e o Mg ( o de ails on ope a ing condi ions see Table 2.3 o Chap e 2). The de e mined composi ions a e lis ed in Table 4.2. Table 4.2: The ca ion composi ion o he ga ne s a ing ma e ials as analyzed by elec on mic op obe calcula ed based on 12 oxygen pe o mula uni . Abb e ia ion Maj. (majo i e) and Py . (py ope). Composi ion Si Mg Al To al Maj. con en Py . con en Py ope 3.007 2.974 2.007 7.989 0.007 0.993 Py ope 80 3.224 3.133 1.611 7.970 0.224 0.776 Py ope 40 3.606 3.548 0.827 7.980 0.605 0.394 Py ope 30 3.677 3.689 0.638 8.004 0.677 0.323 Py ope 20 3.755 3.828 0.442 8.024 0.755 0.245 Py ope 15 3.867 3.783 0.321 7.972 0.867 0.133 Py ope 10 3.897 3.846 0.240 7.983 0.897 0.103 Majo i e 4.025 3.95 0 7.974 1.025 0 The emaining po ion o each sample was g ound o a powde and a po ion o his powde , which is p esumed o ha e been ep esen a i e o he en i e sample, was cha ac e ized by X- ay powde di ac ion. The samples we e mixed wi h a small amoun o Si as an in e nal s anda d (NBS s anda d numbe 640b). X- ay di ac ion was pe o med using a Philips X’Pe P o X– ay di ac ome e ope a ing in e lec ion mode; using Co Kα adia ion wi h an wa eleng h o 1.78892 Å selec ed using a ocused monoch oma o . The di ac ion condi ions we e se o a s ep size o 0.2°, a s ep ime o 1000s wi h scan speed o 0.0002°/sec wi h a o a ing pla o m o a ing a 1 o a ion pe second. The da a we e collec ed o e a 2θ ange o 20-120°. Cell pa ame e e inemen s we e pe o med using he GSAS so wa e package (see Table 4.1). Ob ained Cell pa ame e s a e plo ed in Fig. 4.3. No phases o he han cubic o e agonal ga ne we e 113 Chap e 4: Calo ime ic s udies de ec ed. The cubic o e agonal ga ne ansi ion was obse ed a sligh ly lowe py ope con en s (< 13 %) han in he s udy o Heinemann e al., (1997). Figu e 4.3: Re ined cell pa ame e s o he ga ne composi ions in he majo i e–py ope join a e syn hesis in mul i-an il and pis on cylinde ( illed squa es). The cell pa ame e s as epo ed by Heinemann e al., 1997 along he majo i e–py ope join a e plo ed (solid iangles) o compa ison. They epo ed he s abili y ield o e agonal ga ne s o commence om he Maj 80 composi ion; howe e ou Maj 0.867 s ill appea s o be a cubic ga ne . This disc epancy may be a ibu ed o di e ences in he syn hesis condi ions. 4.3 Calo ime ic measu emen s 4.3.1 Basic p incipals A calo ime e measu es he change in hea associa ed wi h he change o s a e o a sample. A numbe o calo ime ic me hods ha e been employed in Ea h sciences o p o ide basic he modynamic da a. The mophysical measu emen s such as low empe a u e adiaba ic calo ime y o di e en ial scanning calo ime y (Akaogi, 1990; Na o sky, 1997; Na o sky, 2004) a e used o de e mine hea capaci ies as a unc ion o empe a u e, which a e also equi ed o de e mine he s anda d en opy o a phase. The mochemical o eac ion calo ime y, on he o he hand, p o ides a measu emen o he en halpy o a eac ion, which can be combined wi h o he eac ion en halpies o gi e he hea o o ma ion o a 114 Chap e 4: Calo im e ic s udies compound. A widely used ype o eac ion calo ime y is o measu e he hea o dissolu ion as a sample dissol es in a sol en , ei he an acid o an oxide mel , o in ini e dilu ion. The e a e wo me hods by which his can be accomplished, which de ia e wi h espec o he ini ial empe a u e o he sample. In solu ion calo ime y he ini ial empe a u e o he sample is iden ical o ha o he sol en be o e dissolu ion. In d op solu ion calo ime y he sample is ini ially a oom empe a u e and is d opped in o he high- empe a u e sol en . In his s udy d op solu ion calo ime y using an oxide mel was employed. This was deemed mo e sui able because he eco e ed me as able high-p essu e phases could po en ially b eakdown du ing he high empe a u e equilib a ion s age (a ~700°C) o solu ion calo ime y. I should be no ed, howe e , ha in p e ious s udies solu ion calo imen y has been ex ensi ely employed o measu e solu ion en halpies o high-p essu e phases (Akaogi e al., 1987; Yusa e al., 1993; Akaogi and I o, 1999). 4.3.2 En halpy measu emen s A win cal e ype mic ocalo ime e , based on he design desc ibed by Kleppa, (1976), ins alled a Gakushuin Uni e si y, Tokyo, was used o ou calo ime ic measu emen s. I consis ed o wo sample chambe s each o which was su ounded by a he mopile o P - P 10Rh he mocouples. The he mopiles we e connec ed in opposi ion inside a massi e kan hal block. The block was main ained a 978 K by hea e s si ua ed ou side o he block (Fig. 4.4). The e.m. . om he he mopile a e ampli ica ion was eco ded using an elec onic in eg e a , which ga e a g aphical ep esen a ion o he hea low o he expe imen (see Fig. 4.6) and a nume ical alue o he a ea unde calo ime ic peak a e p ocessing. 115 Chap e 4: Calo ime ic s udies Insula ion Sample chambe Fu nace The mopile Calo ime ic block Figu e. 4.4 A Twin Cal e ype mic ocaloi me e o he ype desc ibed by Kleppa, (1976) as used in his s udy. 2PbO.B2O3 was used as he sol en in ou expe imen s. A single ba ch o sol en was made by mixing PbO and H3BO3 in a 1:1 mola p opo ion. The mix u e was ini ially hea ed in a 1-a mosphe e u nace a 1073 K o 30 minu es. The quenched glass was ehomogenised by g inding and emel ed a 1323 K o one and hal hou s o ensu e dehyd a ion (Cha lu e al., 1975). Fo each expe imen 5 g ams o 2PbO.B2O3 glass we e measu ed in o each o wo P ubes, one o each sample chambe o he calo ime e , and mel ed in a u nace a 700°C o an hou . Each ube was placed in o one o he calo ime e sample chambe s housed in a second P ube jacke ed by a silica glass ube inside an inconel ube. The inconel ube was inse ed in o he calo ime e p io o he beginning o each expe imen and was allowed o each he mal equilib ium as indica ed by a s able base line om he he mopiles. A gas was bubbled h ough he sol en a a low a e o 5 cm3/minu e in o de o has en dissolu ion by s i ing as shown in Fig. 4.5. The sample pelle s o abou 3 mg weigh we e made using a minia u e pelle p ess and we e d opped om oom empe a u e in o he 2PbO.B2O3 sol en a 978 K. D ops we e al e na ed be ween he wo samples chambe s. Hea was ei he abso bed o libe a ed du ing he eac ion and he he mopiles de ec ed he change o empe a u e be ween he sample chambe and he kan hal block. The libe a ed 116 Chap e 4: Calo im e ic s udies hea gi es ise o a calo ime ic peak, which dies ou exponen ially o he o iginal baseline wi h ime as hea is ans e ed om he sample chambe o he block (Fig. 4.6). The eco ded hea change is equal o he hea con en o he sample plus he en halpy o solu ion a he empe a u e o he calo ime e . The py ope- ich samples we e dissol ed wi hin an hou o d opping in o he sol en ; howe e , he majo i e ich composi ions equi ed app oxima ely 1.5 hou s o comple e dissolu ion. When he calo ime ic peak dies ou and ge s back o he o iginal base line he sample is conside ed o be o ally dissol ed. P Gas ube Inconel ube SiO glass ube 2 P es ube co e P es ube 2 PbO.B O lux 23 Sample palle A gas bubbling A gas bubbling Figu e 4.5: A schema ic diag am showing he se o he sample chambe s used o d op solu ion calo ime ic echnique. A gas was bubbled h ough he sample chambe s o has ening dissolu ion. 117 Chap e 4: Calo ime ic s udies 0 5000 10000 15000 20000 -15 -10 -5 0 5 10 15 Hea low /mW Time / sec Base line ''Endo he mic'' peak ''Exo he mic'' peak Figu e 4.6. A d op solu ion calo ime ic cu e o he py ope composi ion showing he e ol ing hea o dissolu ion as a unc ion o ime. Each peak ep esen s a d op solu ion un and a e ma ked exo he mic and endo he mic peaks o dis inc ion o hea low alues ob ained om le and igh sample chambe s o he calo ime e (see Fig. 4.4). By in eg a ing he a ea unde he peaks, hea o dissolu ion associa ed wi h each un is de e mined. The hea e ec is p opo ional o he a ea unde he calo ime ic peak bu he p opo ionali y ac o mus be calib a ed. This calib a ion ac o was de e mined o each sample chambe using Al2O3 as a known s anda d. Pelle s o Al2O3 o app oxima ely 3 mg we e d opped in o 5 gm o lead bo a e sol en . The in eg a ed hea con en o Al2O3 dissolu ion was compa ed wi h he s anda d ∆Hd-sol o Al2O3 o 106.73 kJ/mol o gi e calib a ion ac o s o 0.875 + 0.01 J/mol o he le -side and 0.896 + 0.01 J/mol o he igh -side sample chambe s (Table 4.3). 118 Chap e 4: Discussion Table 4.3: Calib a ion ac o calcula ion using Al2O3 pelle s. (Abb e ia ions: obse ed H is he obse ed hea , ∆H is he en halpy o solu ion and S is he calib a ion ac o ). Le Side Run No. Al2O3 mass / g Al2O3 /mol Obse ed H / J Obse ed H/ KJ ∆H o Al2O3 / KJ.mol-1 S 1 0.003253333 3.1908E-05 3.0124 0.0030124 94.41003578 0.884568873 2 0.003303333 3.2398E-05 3.0242 0.0030242 93.34524406 0.874592374 3 0.00337 3.3052E-05 3.0815 0.0030815 93.23229125 0.87353407 4 0.003186667 3.1254E-05 2.9089 0.0029089 93.07354171 0.872046676 5 0.00346 3.3935E-05 3.1445 0.0031445 92.66368916 0.868206588 6 0.003286667 3.2235E-05 3.0253 0.0030253 93.8527221 0.879347157 7 0.003333333 3.2692E-05 3.1171 0.0031171 95.34678993 0.893345732 8 0.0034 3.3346E-05 3.061 0.003061 91.79488853 0.860066416 A e age calib a ion ac o 0.875713486 Righ side Run No. Al2O3 mass / g Al2O3 /mol Obse ed H / J Obse ed H/ KJ ∆H o Al2O3 / KJ.mol-1 S 1 0.00325 3.1875E-05 3.0643 0.0030643 96.13510532 0.900731803 2 0.003526667 3.4588E-05 3.3203 0.0033203 95.99464319 0.899415752 3 0.003216667 3.1548E-05 3.0454 0.0030454 96.53223712 0.904452704 4 0.003353333 3.2888E-05 3.1942 0.0031942 97.12241338 0.909982323 5 0.00328 3.2169E-05 3.0046 0.0030046 93.40000628 0.875105465 6 0.00344 3.3738E-05 3.2499 0.0032499 96.32646916 0.902524774 7 0.00339 3.3248E-05 3.2332 0.0032332 97.24492779 0.911130214 8 0.003306667 3.2431E-05 3.0598 0.0030598 94.34887131 0.883995796 9 0.00325 3.1875E-05 3.0179 0.0030179 94.67941597 0.887092813 10 0.00316 3.0992E-05 2.9567 0.0029567 95.40129389 0.893856403 A e age calib a ion ac o 0.896828805 4.4 Resul s The measu ed d op solu ion en halpies (∆H d-sol) o majo i e-py ope ga ne s a e epo ed in Table 4.4. Due o he size o he syn hesized samples a numbe o d ops o app oxima ely 3 mg each could be pe o med o each sample and he inal ∆H d-sol is he e o e he a e age o be ween 3-6 d ops. 119 Chap e 4: Discussion Table 4.4: D op solu ion en halpies o Mg3(Mg,Si)Si3O12 (majo i e)-Mg3Al2Si3O12 (py ope) solid solu ion join in 2PbO.B2O3 a 978K, whe e Pyx is he py ope composi ion de e mined om elec omic p obe, ∆H d-sol is he en halpy o dissolu ion. Mass (g) ∆H d-sol (kJ /mol) Mass (g) ∆H d-sol (kJ /mol) Py 0.993 Py 0.776 0.00270 406.6921 0.00259 344.5220 0.00251 420.5045 0.00278 344.6312 0.00296 396.3194 0.00282 358.7211 0.00253 388.5248 0.00313 400.6793 0.00281 405.9892 A . 403.118 ± 8.085 A . 349.291 ± 8.166 Py 0.394 Py 0.323 0.00296 307.8711 0.00262 290.4055 0.00287 308.38 0.00255 286.598 0.00273 308.292 0.00281 292.6601 0.00261 292.6617 0.00270 303.6652 0.00273 291.9788 0.00289 294.4718 0.00273 301.6067 A . 301.798 ± 5.796 A . 293.472 ± 5.083 Mass (g) ∆H d-sol (kJ /mol) Mass (g) ∆H d-sol (kJ /mol) Py 0.245 Py 0.133 0.00261 214.3883 0.00299 252.4929 0.00256 230.8846 0.00268 241.8 0.00256 222.8271 0.00260 237.4546 0.00264 232.6677 0.00258 247.6614 0.00253 221.0675 A . 224.3688 ± 5.989 A . 245.8522 ± 7.131 Py 0.103 Maj 0.00257 281.2224 0.00257 244.5810 0.00261 265.012 0.00261 245.6691 0.00276 260.9722 0.00276 234.0641 0.00289 260.8478 0.00279 254.2783 0.00283 266.1986 0.00283 212.1408 A . 266.251 ± 6.881 A . 238.146 ± 12.968 The en halpies o dissolu ion a e plo ed in Fig 4.7. The da a show a nega i e de ia ion om a s aigh line joining he wo end membe s. This implies ha he mixing p ope ies o he solid solu ion de ia e posi i ely om ideali y. F om he py ope end membe he 120 Chap e 4: Discussion en halpies dec ease non linea ly wi h dec easing py ope con en , howe e , be ween a py ope con en o 0.32 and 0.24 a s ong b eak in slope occu s and en halpies s a o inc ease be o e a inal dec ease occu s a he majo i e end membe . 020406080100 220 240 260 280 300 320 340 360 380 400 420 ∆H0 d-s KJ/mol mol % Al2O3 Mg4Si4O12 Mg3Al2Si3O12 Figu e. 4.7: En halpies o dissolu ion o Mg3(Mg,Si)Si3O12-Mg3Al2Si3O12 ga ne s in 2PbO.B2O3 sol en a 978K. The d op dissolu ion en halpy (∆H d-sol) can be w i en as, ∫ + ∆ = ∆Η −978 298 .d Cp Hsolsold (1) whe e, ∆H sol is he en halpy o solu ion and he in eg al e m accoun s o he change in hea equi ed o aise he empe a u e o he sample om oom empe a u e o ha o he sol en . Values o he in eg al e m a e 298 kJ/mol o py ope (Robie e al., 1978) and 297 kJ/mol o majo i e (Yusa e al., 1993) and i is assumed ha his alue a ies linea ly ac oss he solid solu ion. The unce ain y in his co ec ion is o he o de o 1 kJ/mol. Values o ∆H sol a e mo e use ul o compa ing wi h p e ious measu emen s and o calcula ing en halpy changes o mine al eac ions. 121 Chap e 5: Conclusions his exsolu ion eac ion a mid- ansi ion zone condi ions, using mine al physics da a, show ha he impedance con as o he ini ial pe o ski e exsolu ion is o a sui able magni ude o cause a discon inui y. The e o e, coupled wi h he wadsleyi e- ingwoodi e ansi ion, wo discon inui ies a e, possible and a e consis en wi h he obse ed spli in he 520 km seismic discon inui y. Howe e , why he discon inui y should appea spli in some egions o he Ea h and no in o he s is shown no o be a unc ion o empe a u e. Al hough empe a u e would end o me ge he wo discon inui ies oge he i would do so only a dep hs much deepe han whe e a single discon inui y is ac ually obse ed. Wha is mo e likely is ha a single discon inui y is obse ed in egions whe e he Ca con en o he man le is low bu in egions ich in Ca, such as hose con aining signi ican ecycled oceanic c us , wo discon inui ies would be obse ed. Ou indings allow egional seismic obse a ions o his spli ing in he 520 km discon inui y o be used as a p obe o his majo ype o man le he e ogenei y in he man le. This s udy is, he e o e, he i s epo o a deep man le seismic discon inui y, which can be used as a sensi i e indica o o man le chemical he e ogenei y. Changes in he elas ic p ope ies o magnesium silica e pe o ski e ha e been examined as a unc ion o he inco po a ion o e ic Fe and Al in he s uc u e, using diamond an il cell comp ession and single-c ys al X- ay di ac ion echniques. Comp ession expe imen s on Al-and Fe-bea ing magnesium silica e pe o ski es show ha Fe and Al inco po a ion makes he magnesium silica e pe o ski e s uc u e mo e comp essible. The Al3+ and Fe3+ subs i u ion in magnesium pe o ski e will mos ly occu ia a coupled subs i u ion mechanism a man le condi ions; howe e a low i alen ca ion concen a ions a small p opo ion o oxygen acancies may o m. A hi d o de Bi ch– Mu naghan i o he comp ession da a yields a bulk modulus which dec eases om 243 o 234 GPa wi h inc easing Fe and Al con en wi h a ise in K' om 5 o 6.5, which can be a ibu ed o an inc ease in polyhed al comp essibili y. On conside a ion o he e ec s o possible subs i u ion mechanisms on mola olumes o Al-and Fe-bea ing magnesium silica e pe o ski e, i has been obse ed ha in magnesium silica e pe o ski es exis ing a lowe man le condi ions he subs i u ion o he componen will be ene ge ically less a ou ed. Ou esul s show ha disp opo iona ion o Fe is ene ge ically a o able in he lowe man le condi ions which would esul in he en ichmen o Fe OSiFe VIVIII 3 42 ++ 3+ a lowe man le condi ions. 128 Chap e 5: Conclusions Slow kine ics o silica e eac ions inhibi s he a ainmen o equilib ium unde easible expe imen al condi ions a lowe empe a u es. Consequen ly, he expe imen al esul s a e gene ally ex apola ed o lowe empe a u es using sui able he modynamic models. Unce ain ies in such models can be educed conside ably, i he modynamic pa ame e s used o such i s a e independen ly de e mined as has been done in ou s udy o he py ope-majo i e solid solu ion using he d op solu ion calo ime ic echnique. En halpies o solu ion along he majo i e–py ope join ob ained om d op solu ion calo ime y show a signi ican de ia ion om ideali y. The en halpy o dissolu ion dec eases non-linea ly o composi ions below 30mol% py ope due o a symme y change om cubic o e agonal. The alues o d op solu ion en halpies on majo i e–py ope join a e signi ican ly lowe han p e ious es ima es. An es ima ion o he excess p ope ies using a symme ic egula solu ion model gi es a alue o WH {in e ac ion pa ame e o mixing on one ca ion si e in ga ne i.e., (Mg,Si)↔Al} o 38 kJ/mol and a solu ion en halpy o he ic i e cubic majo i e ga ne o –37 kJ/mol. This would mean a la ge posi i e en halpy o mixing along his solid solu ion join. This hesis wo k emphasizes how a combina ion o complimen a y expe imen al echniques can lead o a obus assessmen o he deep in e io o he Ea h and con ibu es o a be e unde s anding o he e olu ion o he Ea h. Fu he wo k Du ing he cou se o his hesis a numbe o poin s ha e become ob ious whe e u he expe imen al s udies would bene i ou unde s anding o he silica e pe o ski e o ming eac ions. A co ela ion was obse ed be ween he seismically de e mined geog aphical dis ibu ion o he spli 520 km discon inui y wi h mode n and pas subduc ion zones as men ioned in chap e 2. This suppo s ou idea o Ca-en ichmen in he man le ia subduc ion o oceanic c us , which would lead o he o ma ion o Ca-pe o ski e causing he second discon inui y in he 520 km egion. So, u he geophysical s udies in o he spli ing and a iabili y o he 520 km seismic discon inui y combined wi h ou in e p e a ion will allow us o scale la e al dis ibu ion o man le he e ogenei y on a a ie y o scales and will p o ide signi ican insigh s in o he ci cula ion o subduc ed oceanic li hosphe e in he man le. Fu he mo e as he oceanic basal bulk composi ions a e ich in 129 Chap e 5: Conclusions he SiO2 componen , addi ional phase equilib ia s udies in ol ing he o ma ion o Ca- pe o ski e om Ca2SiO4 and CaSi2O5 will lead o be e cons ain s on he he modynamic modeling o he Ca-pe o ski e o ming eac ion. As e y li le he modynamic da a exis o calcium silica e pe o ski e, ano he impo an aspec will be o collec high quali y calo ime ic da a on Ca2SiO4 and CaSi2O5 o aid in he calcula ion o calcium pe o ski e he modynamic p ope ies om i s o ma ion eac ion. Ou s udy o he equa ion o s a e o Al- and Fe-bea ing magnesium silica e pe o ski e single c ys als was pe o med a ambien empe a u e up o p essu es o 10 GPa. I was obse ed in ou s udy ha e en hough a K' alue g ea e han 4 was ob ained om ou esul s, a calcula ion o KT o lowe man le condi ions wi h a K' = 4, p o ides a be e i o he p edic ed lowe man le bulk modulus. Gi en his disc epancy, i is necessa y o ex end he equa ion o s a e s udy o p essu e- empe a u e ange ele an o man le condi ions in o de o be e quan i y he changes in elas ic p ope y a hose condi ions due o Al and Fe inco po a ion in he magnesium silica e s uc u e. This will be possible, using a gas loading o he diamond an il cell o comp ession expe imen s, which would p o ide hyd os a ic condi ions o highe p essu es. In addi ion, a powde x- ay di ac ion equa ion o s a e s udy could be ca ied ou using a high in ensi y synch o on X- ay sou ce, in case o ailu e in p oducing sui able single c ys als. Ano he impo an objec i e would be o ca y ou c ys al s uc u al e inemen s udies using single c ys al X- ay di ac ion, p o ided we can syn hesize single c ys als o sui able size possibly win- ee, o a be e unde s anding o he subs i u ion mechanism o ca ions in di e en c ys al s uc u al si es. Fo he calo ime ic measu emen s on he py ope–majo i e join, i will be impo an o ep oduce he hea o dissolu ion da a o he majo i e- ich composi ions, whe e a s ong sca e in da a has been obse ed by ou s udy, in o de o con i m whe he he e is an e ec o o de ing o Mg and Si on he oc ahed al si e ela ed o syn hesis condi ions. A ho ough assessmen o he o de ing o Mg and Si o he majo i e- ich composi ions can be ca ied ou using a Raman spec oscopic s udy o a 27Al-NMR s udy on hese majo i e- ich composi ions. In addi ion, in o de o employ ou he modynamic da a a man le empe a u es, in o ma ion on he excess en opy e m will be equi ed. This can be de e mined by hea capaci y measu emen s along he majo i e–py ope join by employing di e en ial scanning calo ime y o he ecen ly-de ised PPMS (Physical P ope ies Measu emen Sys em) calo ime ic echnique. 130 Appendix A Table A.1: De ails o un numbe s and phases obse ed. Abb e ia ions: Pd (pe ido i ic composi ion), Basl (basal ic composi ion), Py (py ope composi ion), Fo ( o s e i e composi ion), Pd g (pe do i ic ga ne ), Basl g . (basal ic ga ne ), Py g (py ope ga ne ), Maj g (majo i ic ga ne ), Ca-P (calcium pe o ski e), Al–phase (a non-s oichome ic unknown phase wi h Al), Ring ( ingwoodi e), Mw (magnesiowüs i e), S ish (s isho i e), Oli (oli ine), Py x.(py oxene), Pe i (pe iclase),Wad (wadsleyi e), Pd (pe ido i e), P (pe o ski e),Ca-Pd ,Ca- Basl , (all Ca bea ing e e sal composi ions), Ca-Pd g , Ca-basl G .( e e sal phases wi h Ca) Run no. P essu e S a ing composi ions Phases obse ed (GPa) 1400°C H2375 17.9 Pd, Basl , Py , Fo20 Pd g + Ca-P , Basl g +Ca-P , Py g + Ca-P , Ring + Mw + S ish. S3611 18.1 Pd, Basl , Py , Fo20 Pd g + Ca-P , Basl g +Ca-P , Py g + Ca-P , Ring + Mw + S ish S3614 18.6 Pd, Basl , Py , Fo20 Pd g + Ca-P , Basl g +Ca-P , Py g + Ca-P , Ring + Mw + S ish H2370 19.5 Pd, Basl , Py , Fo20 Pd g + Ca-P , Basl g +Ca-P , Py g + Ca-P , Ring + Mw + S ish H2241 19.6 Pd, Basl , Py , Fo20 Pd g + Ca-P , Basl g +Ca-P , Py g + Ca-P , Ring + Mw + S ish 1600°C S3460 18. Pd, Basl , Maj, Fo90 Pd g + Ca-P , Basl g , Maj g + Ca-P , Wad S3470 18.8 Pd, Basl , Py , Fo30 Pd g + Ca-P , Basl g + Ca-P Pyx + Wad S3550 19.2 Pd, Basl , Py , Fo30 Pd g + Ca-P , Basl g + Ca-P Py G + Ca-P , Ring + Mw + S ish S3548 19.5 Pd, Basl , Fo30 Pd g + Ca-P , Basl g + Ca-P , Ring + Mw + S ish S3547 19.5 Pd, Basl , Fo30 Pd g + Ca-P , Basl g + Ca- P , Ring + Mw + S ish S3549 19.9 Pd, Basl , Py , Fo30 Pd g + Ca-P , Basl g + Ca-P Py G , Ring + Mw + S ish S3657 19.9 Basl , F090 + Basl , Fo30 Basl g + Ca-P , Basl g + Ca-P + Oli , Mw + Py x + S ish + Wad S3655 19.9 Basl , F090 + Basl Basl g + Ca-P , Basl g + Ca-P + Oli . S3551 20.4 Pd, Basl , Py , Fo30 Pd g + Ca-P , Basl g + Ca-P Py G + Ca-P , Ring + Mw + S ish S3757 20.7 Pd, Basl , Fo30; Fo50 Pd g + Ca-P , Basl g + Ca-P , 131 Ring + Mw + S ish, Ring + Mw + s ish S3480 22.1 Pd, Basl , Py , Fo30 Pd g + Ca-P +Al –Phase, Basl g , + Ca-P + Al phase, Ring +S ish S3764 21.2 Pd, Basl , Fo30 Pd g + Ca-P , Basl g + Ca-P , Ring + Mw + S ish S3498 21.4 Pd, Basl , Py , Fo30 Pd g + Ca-P , Basl g + Ca-P Py g + Al phase+ Ca-P , Ring + Mw + S ish S3783 22.3 Pd, Basl , Maj, Fo70 Pd g +Ca-P , Basl g + Ca-P , Ring + Mw + S ish + Py x. S3484 22.3 Pd, Ca-Basl , Py , Fo30 Pd g +Ca-P + Pe i, Ca-Basl g + Ca-P , Py g + Ca-P , Ring + Mw S3475 22.6 Pd, Basl , Py , Maj Pd g + Ca-P , Basl g + Ca-P , Py G + Ca-P , Maj g + Ca-P S3478 23 Pd, Basl , Fo30 Pd g + Ca-P , Basl g + Ca- P , + G + Ca-P , P + Mw + S ish S3784 23.5 Pd, Ca-Basl , Fo70, Pd Pd g +Ca-P , Ca-Basl g + Ca-P Mw + P , S ish + Mw +P + Co Re e sals (1600°C) S3538 17.95 Basl , Ca Basl , Fo30 Basl g + Pyx, Ca-Pd g + Ca-P , Pyx + Wad S3543 18.4 Basl , Ca Basl , Fo30 Basl g + Pyx, Ca-Pd g + Ca-P Pyx +Wad +Mw S3515 19.8 Basl , Ca Basl , Fo30, Fo98 Basl g + Ca-P , Ca-Basl g + Ca-P , Ring +Wad + Mw, Wad S3521 20.7 Pd, Ca Pd, Fo30 Pd g +Ca-P , Ca-Pd g + Ca-P Ring + Mw S3523 20.8 Pd, Ca-Pd, Ca-Basl , Fo30 Pd g + Ca-P + Al-phase +Ca-Pd g + Ca-P , Ring +S ish S3784 23.5 Pd, Ca-Basl , Fo70, Pd Pd g +Ca-P , Ca-Basl g + Ca-P , Mw + P , S ish + Mw +P + Co Table A.2: Elec on mic op obe da a o he pe o ski es c ys allized om he p essu e calib an s in mul i an il expe imen s calcula ed based on 3 oxygens pe o mula uni . Fo30 (Fo s e i e 30), Fo70 (Fo s e i e 70) Run no. P essu e calib an P essu e Si Mg Fe To al S3478 Fo30 23 GPa 0.961 0.763 0.310 2.034 S3784 Fo70 23.5 GPa 0.962 0.899 0.171 2.031 132 Table A.3: Elec on mic op obe da a o all he expe imen al uns be o e p ocessing o da a. All he concen a ions a e in w % oxide. (A) o expe imen s a 1600°C, (B) o expe imen s a 1400°C and (C) o e e sal expe imen s. Resul s a e lis ed as pe ido i e, basal , py ope and majo i e based on he di e en ga ne s a ing composi ions as lis ed in able 2.1. (Abb e ia ion Pd g –pe ido i ic ga ne , Ca-P -calcium pe o ski e, Basl g -basal ic ga ne , Maj g –majo i ic ga ne , Py g - py ope ga ne , Fo90- o s e i e 90 composi ion, Fo50- o s e i e 50 composi ion, Fo30- o s e i e 30 composi ion, (Mg,Fe)O-magnesiowüs i e, SiO2-s isho i e, (A) Runs a 1600°C. (Run no. Composi ion Phases SiO2 MgO FeO CaO Al2O3 To al obse ed S3460 Pe ido i e Pd g 52.92 24.96 1.71 11.10 7.82 98.49 Ca-P 59.54 10.71 0.77 16.88 3.12 91.03 Basal Basl g 49.31 21.02 0.83 14.81 12.70 98.66 Majo i e Maj g 54.51 28.92 1.68 8.36 5.79 99.26 Ca -P 47.89 0.09 0.00 44.03 0.01 92.02 Fo 90 Oli ine 42.18 56.04 2.32 0.02 0.10 100.66 S3470 Pe ido i e Pd g 53.26 25.98 1.12 11.21 8.26 99.82 Ca-P 49.25 6.34 0.58 38.21 1.74 96.13 Basal Basl g 49.17 22.74 0.55 14.09 13.23 99.77 Py ope Py g 46.97 19.24 0.82 16.16 16.30 99.48 Ca-P 49.68 0.75 0.03 43.66 0.58 94.71 Fo 90 Oli ine 42.08 56.53 1.71 0.00 0.10 100.43 S3550 Pe ido i e Pd g 50.26 23.90 3.67 11.91 8.26 98.01 Ca-P 50.08 0.17 0.10 47.33 0.08 97.77 Basal Basl g 45.55 17.47 2.95 16.70 16.06 98.73 Ca-P 47.57 2.16 0.42 43.91 1.84 95.89 Py ope Py g 47.39 20.84 3.49 14.30 12.83 98.84 Ca-P 49.34 0.78 0.23 46.39 0.55 97.29 Fo30 oli ine 32.87 17.17 51.65 0.01 0.05 101.76 (Mg,Fe)O 1.22 3.34 89.81 0.07 0.30 94.74 SiO2 98.20 0.27 2.06 0.01 0.21 100.75 S3548 Pe ido i e Pd g 51.49 27.95 3.90 7.95 8.33 99.62 Ca-P 49.20 0.20 0.10 46.98 0.07 96.56 Basal Basl g 47.90 24.00 5.07 9.83 12.45 99.25 133 Fo30 oli ine 33.86 17.88 49.40 0.02 0.05 101.21 (Mg,Fe)O 3.68 8.78 85.02 0.02 0.04 97.54 SiO2 98.83 0.63 2.40 0.01 0.08 101.95 S3547 Pe ido i e Pd g 51.11 27.53 2.87 8.44 9.38 99.32 Ca-P 50.24 0.13 0.06 47.51 0.07 98.00 Basal Basl g 48.58 24.21 3.56 9.72 12.46 98.55 Ca-P 46.68 0.62 0.24 46.01 0.57 94.12 Fo30 oli ine 33.66 18.91 49.39 0.00 0.05 102.02 (Mg,Fe)O 2.93 9.40 87.59 0.00 0.07 99.98 SiO2 96.86 0.18 1.62 0.01 0.19 98.87 S3549 Pe ido i e Pd g 50.47 28.51 1.81 7.08 9.17 97.04 Ca-P 45.78 1.70 0.22 44.84 0.84 93.38 Basal Basl g 47.46 25.27 2.70 8.55 13.60 97.58 Ca-P 48.52 0.36 0.10 46.44 0.10 95.52 Py ope Py g 44.08 20.65 2.82 11.52 17.84 96.91 Fo30 oli ine 33.65 21.30 45.37 0.00 0.06 100.39 (Mg,Fe)O 1.08 7.40 87.90 0.02 0.17 96.57 SiO2 95.05 0.39 3.64 0.01 0.03 99.11 S3657 Basal Basl ic maj 43.46 12.43 34.93 2.89 4.52 100.36 Ca-P 48.27 0.35 0.30 46.38 0.33 95.62 Fe+ basal Basl g 46.85 24.20 7.38 8.28 11.92 98.63 Ca-P 48.04 0.19 0.12 46.26 0.32 94.93 Fo30 oli ine 33.67 21.25 45.76 0.00 0.06 100.75 (Mg,Fe)O 2.84 4.62 86.96 0.06 0.38 94.87 SiO2 72.70 5.19 17.46 0.01 0.08 95.44 S3655 Fe+ basall Basl g 44.01 21.00 6.10 7.59 13.97 92.67 Basal Basl g 48.48 26.19 4.02 7.73 12.57 98.99 Ca-P 47.90 1.78 0.61 45.29 0.11 97.02 Fo30 oli ine 37.05 32.64 30.89 0.00 0.03 100.61 (Mg,Fe)O SiO2 99.07 0.08 0.72 0.00 0.02 99.89 S3551 Pe ido i e Pd g 50.14 27.43 4.46 5.82 10.48 98.33 Ca-P 48.22 0.23 0.16 46.70 0.57 95.88 Basal Basl g 46.41 27.00 1.73 7.95 14.18 97.27 134 Ca-P 39.42 0.92 0.24 43.37 3.69 87.64 Py ope Py g 45.03 21.99 3.80 9.54 18.46 98.82 Ca-P 50.29 0.50 0.16 47.27 0.67 98.89 Fo30 oli ine 34.64 24.08 42.60 0.01 0.06 101.38 (Mg,Fe)O 5.32 5.44 82.80 0.04 0.37 93.96 SiO2 94.77 1.42 4.41 0.02 0.06 100.67 S3480 Pe ido i e Pd g 53.91 28.18 4.23 4.92 8.64 99.88 Ca-p 51.35 0.07 0.03 50.05 0.03 101.53 Basal Basl g 50.82 25.90 4.62 5.35 13.73 100.42 Ca-P 50.85 0.09 0.09 49.40 0.05 100.48 Py ope Py g 46.51 19.56 8.61 8.77 16.70 100.15 Ca-P 50.21 0.18 0.34 49.42 0.43 100.58 Fo 50 Oli ine 37.98 29.59 32.44 0.00 0.07 100.07 SiO2 101.54 0.05 0.54 0.02 0.03 102.17 S3498 Pe ido i e Pd g 53.22 32.12 5.44 0.71 9.53 101.02 Ca-p 51.29 0.70 0.18 45.56 0.26 97.98 Basal Basl g 48.00 27.10 4.76 4.66 15.64 100.16 Ca-P 50.54 0.31 0.11 45.32 0.22 96.50 Py ope Py g 45.16 24.71 3.95 5.38 21.49 100.68 Ca-P 50.52 0.20 0.09 45.82 0.10 96.73 Fo 50 Oli ine 38.72 37.48 26.17 0.00 0.08 102.46 (Mg,Fe)O 6.21 12.41 70.20 0.14 0.73 89.69 S3484 Pe ido i e Pd g 52.31 33.49 3.00 0.89 9.24 98.93 Ca-p 53.01 1.76 0.12 42.10 0.49 97.48 Basal Basl g 47.63 28.78 3.44 2.38 16.66 98.88 Ca-P 51.29 0.23 0.09 46.93 0.09 98.63 Py ope Py g 44.03 25.42 3.84 2.04 23.32 98.64 Ca-P 49.31 1.60 0.33 45.24 1.43 97.91 Fo 50 Oli ine 37.01 30.79 31.97 0.00 0.09 99.87 (Mg;Fe)O 0.71 21.41 75.89 0.00 0.12 98.13 S3475 Pe ido i e Pd g 53.23 31.49 5.09 0.86 9.61 100.27 Ca-P 48.06 6.17 1.13 35.97 2.09 93.42 Basal Basl g 49.03 27.26 3.28 3.03 17.03 99.62 Ca-P 46.97 0.43 0.15 45.41 0.72 93.68 135 Majo i e Maj g 49.32 16.34 2.41 21.32 7.36 96.75 Ca -P 49.06 0.96 0.20 44.84 0.58 95.64 Py ope Py g 44.96 24.00 3.58 5.04 22.35 99.94 Ca-P 42.84 0.52 0.26 44.77 2.38 90.77 S3478 Pe ido i e Pd g 51.90 33.15 5.20 0.20 9.09 99.53 Ca-p 48.86 0.15 0.11 46.80 0.08 96.00 Basal Basl g 47.40 28.31 4.42 2.96 16.32 99.41 Ca-P 42.92 0.95 0.41 45.06 2.56 91.89 Py ope Py g 44.22 24.56 4.81 4.28 21.55 99.41 Ca-P 45.29 1.30 0.58 44.21 2.20 93.58 Fo 50 P 51.19 27.09 19.74 0.24 1.92 100.19 (Mg,Fe)0 0.38 20.28 78.75 0.00 0.10 99.51 SiO2 97.62 0.37 2.18 0.00 0.02 100.19 S3757 Pe ido i e Pd g 52.37 29.20 4.14 4.48 8.42 98.61 Ca-p 47.99 0.39 0.25 45.93 0.73 95.29 Basal Basl g 47.64 24.98 3.36 6.26 15.54 97.78 Ca-P 44.64 1.13 0.49 43.13 1.16 90.55 Fo50 Oli ine 34.19 26.74 39.20 0.02 0.10 100.25 (Mg,Fe)0 0.22 14.33 82.58 0.02 0.09 97.23 SiO2 79.56 7.10 12.08 0.02 0.17 98.93 Fo 30 Oli ine 35.04 25.43 40.32 0.01 0.07 100.86 (Mg,Fe)0 0.94 13.63 84.07 0.01 0.08 98.73 SiO2 94.94 1.56 3.37 0.02 0.21 100.10 S3764 Pe ido i e Pd g 52.70 28.41 4.82 3.67 9.60 99.19 Ca-p 48.12 1.14 0.19 43.70 0.96 94.11 Basal Basl g 48.06 27.32 3.58 4.68 15.44 99.09 Ca-P 49.43 0.04 0.04 46.96 0.03 96.50 Fo30 Oli ine 35.64 27.84 37.14 0.00 0.09 100.70 (Mg,Fe)0 1.98 17.16 80.91 0.01 0.10 100.16 SiO2 90.05 4.11 7.14 0.00 0.25 101.56 S3784 Pe ido i e Pd g 51.46 32.16 4.87 0.50 10.17 99.15 Ca-p 49.83 0.29 0.13 46.69 0.26 97.19 Fo70 (Mg,Fe)0 6.90 29.88 64.18 0.04 0.01 101.01 P 55.18 34.70 11.18 0.16 0.08 101.30 136 S3783 Pe ido i e Pd g 51.34 30.92 5.77 1.45 9.39 98.87 Ca-P 49.22 0.58 0.19 45.07 0.36 95.43 Basal Basl g 50.28 15.75 2.98 23.26 4.87 97.15 Ca-P 48.47 0.24 0.13 44.60 0.29 93.72 Majo i e Maj g 49.83 11.87 2.27 28.60 3.73 96.29 Ca-P 49.04 0.16 0.08 45.60 0.11 94.99 Fo70 Oli ine 37.18 32.30 31.24 0.00 0.07 100.79 (Mg,Fe)0 8.31 12.62 74.94 0.11 0.60 96.58 SiO2 91.16 1.26 3.35 0.02 0.15 95.93 (B) Runs a 1400°C Run no. Composi ion Phases SiO2 MgO FeO CaO Al2O3 To al obse ed S3611 Py ope Py g 44.92 18.91 2.48 14.50 19.34 100.14 Ca-P 48.94 0.44 0.12 45.83 0.45 95.78 Basal Basal g 46.66 19.99 3.38 14.66 15.74 100.42 Ca-P 49.61 0.93 0.18 46.22 0.50 97.43 Pe ido i e Pd g 51.22 26.26 4.36 9.21 8.56 99.61 Ca- P 49.55 0.19 0.12 47.35 0.05 97.26 S3614 Py ope Py g 44.15 19.66 3.29 13.64 18.79 99.53 Ca-P 48.32 0.22 0.12 47.13 0.38 96.18 Basal Basal g 47.00 23.50 3.77 11.35 14.07 99.69 Ca-P 48.49 0.28 0.12 47.16 0.36 96.41 Pe ido i e Pd g 50.74 27.91 4.54 8.60 8.24 100.03 Ca- P 49.20 0.11 0.11 47.41 0.13 96.96 H2375 Py ope Py g 45.21 18.65 3.23 14.01 17.93 99.02 Ca-P 50.21 0.10 0.06 46.94 0.08 97.39 Basal Basal g 47.46 20.57 3.73 13.36 13.67 98.79 Ca-P 48.97 0.18 0.13 46.71 0.17 96.15 H2370 Py ope Py 44.03 23.22 4.08 6.33 22.52 100.18 Ca-P 50.22 0.07 0.09 47.55 0.06 97.99 Fo 30 oli ine 33.45 20.65 51.13 0.01 0.04 105.28 (Mg,Fe)O 0.75 3.24 93.06 0.10 0.31 97.46 SiO2 94.96 0.53 2.74 0.15 0.16 98.54 H2241 Py ope Py g 44.04 22.89 3.25 6.31 21.20 97.68 137 Deuss, A. and Woodhouse, J. (2001) Seismic obse a ions o spli ing o he mid- ansi ion zone discon inui y in ea h’s man le. Science, 294: 354-357. Deuss, A., Red e n, S., Chambe s, K. and Woodhouse, J. H. (2006) The na u e o he 660-kilome e discon inui y in Ea h's man le om global seismic obse a ions o PP p ecu so s. Science, 311: 198- 201. Dziewonski, A. M. and Ande son, D. L. (1981) P elimina y Re e ence Ea h Model. Physics o Ea h and Plane a y In e io s, 25: 297–358. Fab ichnaya, O. B., Saxena, S. K., Riche , P. and Wes um E. F. (2004) The modynamic da a, models and phase diag ams in mul icomponen oxide sys ems. Sp inge , Be lin-Heidelbe g-New Yo k, p 198. Fei, Y., Mao, H. K. B. and Mysen, O. B. (1991) Expe imen al de e mina ion o elemen pa i ioning and calcula ion o phase ela ions in he MgO -FeO -SiO2 sys em a high p essu e and high empe a u e. Jou nal o Geophysical Resea ch, 96: 2157-2169. Fei, Y., Mao, H. K., Shu, J. and Hu, J. (1992) P-V-T equa ion o s a e o magneiowüs i e. Physics and Chemis y o Mine als, 18: 416-422. Fei, Y., Wang, Y. and Finge , L. W. (1996) Maximum solubili y o FeO in (Mg, Fe)SiO3-pe o ski e as a unc ion o empe a u e a 26Gpa: implica ion o FeO con en in he lowe man le. Jou nal o Geophysical Resea ch, 101(B5): 11525-11530. Fei, Y. and Be ka, C. M. (1999) Phase ansi ions in he Ea h’s man le and man le mine alogy. In Fei, Y., Be ka, C. M. and Mysen, B. O., edi o s, Man le pe ology: Field obse a ions and high p essu e expe imen a ion: A ibu e o F ancis R. Boyd, olume 6 o The Geochemical Socie y special Publica ion, p 189-207. The Geochemical Socie y, Hous on. Fei, Y., an O man, J., Li, J., an Wes enen, W., Sanloup, C., Mina ik, W., Hi ose, K., Komabayashi, T., Wal e , M. and Funakoshi, K. (2004) Expe imen ally de e mined pos -spinel ans o ma ion bounda y in Mg2SiO4 using MgO as an in e nal p essu e s anda d and i s geophysical implica ions. Jou nal o Geophysical Resea ch, 109(B02305), doi: 10.1029/2003JB002562. Fique , G., And aul , D., Dewaele, A., Cha pin, T., Kunz, M. and Haüse mann, D. (1998) P-V-T equa ion o s a e o MgSiO3 pe o ski e. Physics o Ea h and Plane a y In e io s, 105: 21-31. 144 Fique , G., Dewaele, A., And aul , D., Kunz, M. and Le Bihan, T. (2000) The moelas ic p ope ies and c ys al s uc u e o MgSiO3 pe o ski e a lowe man le p essu e and empe a u e condi ions. Geophysical Rease ch Le e s, 27(1): 21-24. Flanagan, M. P. and Shea e , P. M. (1998) Global mapping o opog aphy on ansi ion zone eloci y discon inui ies by s acking SS p ecu so s. Jou nal o Geophysical Resea ch, 103: 2673–2692. F os , D. J., Langenho s , F., an Aken, P. (2001) Fe-Mg pa i ioning be ween ingwoodi e and magnesiowüs i e and he e ec o p essu e, empe a u e and oxygen ugaci y. Physics and Chemis y o Mine als, 28: 455-470. F os , D. J. and Langenho s , F. (2002) The e ec o Al2O3 on Fe-Mg pa i ioning be ween magnesiowüs i e and magnesium silica e pe o ski e. Ea h and Plane a y Science Le e s, 199: 227- 241. F os , D. J. (2003a) The s uc u e and sha pness o (Mg,Fe)2SiO4 phase ans o ma ions in he ansi ion zone. Ea h and Plane a y Science Le e s, 216: 313-318. F os , D. J. (2003b) Fe2+-Mg pa i ioning be ween ga ne , magnesiowüs i e, and (Mg,Fe)2SiO4 phases o he ansi ion zone. Ame ican Mine alogis , 88: 387-397 F os , D. J., Liebske, C., Langenho s , F., McCammon, C., T nnes, R. G. and Rubie, D. C. (2004) Expe imen al e idence o he exis ence o i on- ich me al in he Ea h’s lowe man le. Na u e, 428: 409-412. F os , D. J., Poe, B. T., T Ønnes, R. G., Liebske, C., Duba, A., and Rubie, D. C. (2004) A new la ge olume mul ian il sys em. Physics o Ea h and Plane a y In e io s, 143-144: 507-514. F os , D. J. (2006) The s abili y o hyd ous man le phases. In Kepple , H. and Smy h, J. R., edi o s, Wa e in nominally anhyd ous mine als, olume 62 o Re iews in Mine alogy and Geochemis y, p. 243-271, Mine alogical Socie y o Ame ica, Chan illy, Vi ginia. F os , D. J. and Doleĵs D. (2007) Expe iemn al de e mina ion o he e ec o H2O on he 410 km discon inui y. Ea h and Plane a y Science Le e s, 256(1-2): 182-195. F os , D. J. (2007) The uppe man le and ansi ion zone. Elemen s (in p ess). 145 Funamo i, N., Yagi, T., U sumi, W., Kondo, T., Uchida, T. and Funamo i, M. (1996) The moelas ic p ope ies o MgSiO3 pe o ski e de e mined by in si u X- ay obse a ions up o 30 GPa and 2000K. Jou anl o Geophysical Resea ch, 101(B4): 8257-8269. Gaspa ik, T. (1990) Phase ela ions in he T ansi ion Zone. Jou nal o Geophysical Resea ch, 95 (B10): 15751-15769. Gaspa ik, T. (1996) Mel ing expe imen s on he ens a i e-diopside join a 70-224 Kba , including he mel ing o diopside. Con ibu ions o Mine alogy and Pe ology, 124: 139-153. Gaspa ik, T. (2003) Phase diag ams o geoscien is s. An a las o he Ea h’s in e io . Sp inge , Be lin Heidelbe g New Yo k, p 462. Gilbe , F. and Dziewonski, A. M. (1975) An applica ion o no mal mode heo y o he e ie al o s uc u al pa ame e s and sou ce mechanisms om seismic spec a. Philosophical T ansac ions Royal Socie y London, A 278: 187–269. Gilbe , H. J., Sheehan, A. F., Dueke , K. G. and Molna , P. (2003) Recei e unc ions in he wes e n Uni ed S a es, wi h implica ions o uppe man le s uc u e and dynamics. Jou nal o Geophysical Resea ch, 108, 10.1029/2001JB001194. Galze , A. M. (1972) The classi ica ion o il ed oc ahed a in pe o ski es. Ac a C ys allog aphica, B28: 3384-3392. Glaze , A. M. (1975) Simple ways o de e mining pe o ski e s uc u es. Ac a C ys allog aphica, A31: 756-762. Gwanmesia, G. D., Liu, J., Chen, G., Kesson, S., Rigden, S. M. and Libe mann, R. C. (2000) Elas ici y o he py ope (Mg3Al2Si3O12)-majo i e (MgSiO3) ga ne s solid solu ion. Physics and Chemis y o Mine als, 27: 445-452. Ha e, B., Ha is, Hu chinson, M. T., G. R., Wa , M. C. and Wilding, M. C. (1999) Lowe man le mine al associa ions in diamonds om São Luiz, B azil. In Fei, Y., Be ka, C. M. and Mysen, B. O., edi o s, Man le pe ology: Field obse a ions and high p essu e expe imen a ion: A ibu e o F ancis R. Boyd, olume 6 o The Geochemical Socie y Special Publica ion, p 125-153. The Geochemical Socie y, Hous on. 146 Hasel on, H. T. and New on, R. C. (1980) The modynamics o py ope-g ossula ga ne s and hei s abili ies a high empe a u es and high p essu es. Jou nal o Geophysical Resea ch, 85: 6973-6982. Ha ch, D. M. and Ghose, S. (1989) Symme y analysis o he phase ansi ion and winning in MgSiO3 ga ne : Implica ions o man le mine alogy. Ame ican Mine alogis , 74: 1221-1224. Heinemann, S., Sha p, T. G., Sei e , F. and Rubie, D. C. (1997) The cubic o e agonal phase ansi ion in he sys em majo i e (Mg4Si4O12)-py ope (Mg3Al2Si3O12), and ga ne symme y in he Ea h’s ansi ion zone. Physics and Chemis y o Mine als, 24: 206-221. Hein ich, K. F. G. and Newbu y, D. E. (1991) Elec on p obe Quan i a ion. Sp inge , p. 412. Hemley, R. J. and Cohen, R. E. (1992) Silica e pe o ski e. Annual Re iews in Ea h and Plane a y Science, 20: 553-600. Hel ich, G. R and Wood, B.J. (1996) 410 km discon inui y sha pness and he o m o he oli ine -β phase diag am: esolu ion o appa en seismic con adic ions. Geophysical Jou nal In e na ional, 126: F7-F12. Hel ich, G. R. and Wood, B. J. (2001) The Ea h’s man le. Na u e, 412: 501-507. Hi ose, K and Fei, Y. (2002) Subsolidus and mel ing phase ela ions o basal ic composi ion in he uppe mos lowe man le. Geocheimca Cosmochimica Ac a, 66: 2099-2108. Holzap el, C., Rubie, D. C., F os , D. J., Langenho s , F. (2005) Fe-Mg in e di usion in (Mg, Fe)SiO3 pe o ski e and lowe man le eequilib a ion. Science, 309: 1707-1710. Howa d, C. J. and S okes, H. T. (1998) G oup heo e ical analysis o oc ahed al il ing in pe o ski es, Ac a C ys allog aphica, B54: 782-789. I i une, T. (1987) An expe imen al in es iga ion o he Py oxene- Ga ne ans o ma ion in a py oli e composi ion and i s bea ing on he cons i u ion o he man le. Physics o he Ea h and Plane a y In e io s, 45: 324-336. 147 I i une, T. and Ringwood, A. E. (1993) Phase ans o ma ions in subduc ed oceanic c us and buoyancy ela ionships a dep hs o 600-800 km in he man le. Ea h and Plane a y Science Le e s, 117: 101-110. I i une, T. (1994) Absence o an aluminous phase in he uppe pa o he Ea h’s lowe man le. Na u e, 370: 131-133. I i une, T. and Isshiki, M. (1998) I on pa i ioning in a py oli e man le and he na u e o he 410-km seismic discon inui y. Na u e, 349: 409-411. I i une, T., Nishiyama, N., Ku oda, K., Inoue, T., Isshiki, M., U sumi, W., Funakoshi, K-I., U akawa, S., Uchida, T., Ka su a, T. and Oh aka, O. (1998) The pos -spinel phase bounda y in Mg2SiO4 de e mined by in si u x- ay di ac ion. Science, 279: 1698-1701. I a, J. and S ix ude, L. (1992) Pe ology, elas ici y, and composi ion o he man le ansi ion zone. Jou nal o Geophysical Resea ch, 97: 6849–6866. I o, E. and Takahashi, E. (1989) Pos spinel ans o ma ions in he sys em Mg2SiO4-Fe2SiO4 and some geophysical implica ions. Jou nal o Geophysical Resea ch, 94: 10637-10646. Jackson, I. and Rigden, S. M. (1996) Analysis o P-V-T da a: Cons ain s on he he moelas ic p ope ies o high-p essu e mine als. Physics o Ea h and Plane a y In e io s, 96: 85-112. Jackson, J. M., Zhang, J., Shu, J., Sinogeikin, S. V. and Bass, J. D. (2005) High-p essu e sound eloci y and elas ici y o aluminous MgSiO3 pe o ski e o 45 GPa: implica ions o la e al he e ogenei y in Ea h’s lowe man le. Geophysical Resea ch Le e s, 32L21305, doi: 10.1029/2005GL023522. Ja oy, M. (1995) The in eg al ens a i e chond i e model o he Ea h. Geophysical Resea ch Le e s, 22(16): 2219-2222. Je eys, H. and Bullen, K: E. (1940) Seismological Tables. B i ish Associa ion Seismological Commi ee, London. 148 Jiang, F., Speziale, S. and Du y, T. S. (2004) Single-c ys al elas ici y o g ossula -and almandine- ich ga ne s o 11 GPa by B illouin sca e ing. Jou nal o Geophysical Resea ch, 109, 10.1029/2004JB003081. Ka ason H. and an de Hils , R. D. (2000) Cons ain s on men al con ec ion om seismic omog aphy, In Richa ds, M. A., Go don R. G. and an de Hils , R. D., edi o s, His o y and dynamics o global pla e mo ion, olume 121 o Geophysical Monog aph Se ies, p. 277-288, Ame ican Geophysical Union, Washing on. Ka a o, S-I and Jung, H. (1998) Wa e , pa ial mel ing and he o igin o he low eloci y and high a enua ion zone in he uppe man le. Ea h and Plane a y Science Le e s, 157(3-4): 193-207. Ka su a, T. and I o, E. (1989) The sys em Mg2SiO4-Fe2SiO4 a high p essu es and empe a u es: P ecise de e mina ion o s abili ies o oli ine, modi ied spinel, and spinel. Jou nal o Geophysical Resea ch, 94(B11): 15663-15670. Ka su a, T., Yamada, H., Shinmei, T., Kubo, A., Ono, S., Kanzaki, M., Yoneda, A., Wal e , M. J., I o, E., U akawa, S., Funakoshi, K. and U sumi, W. (2003) Pos -spinel ansi ion in Mg2SiO4 de e mined by high P-T in si u X- ay di ac ome y. Physics o he Ea h and Plane a y In e io s, 136: 11-24. Kawai, N. and Endo, S. (1970) The gene a ion o ul ahigh p essu es by a spli sphe e appa a us. Re iew o scien i ic Ins umen s, 41: 1178-1181. Kenne , B. L. N. and Engdahl, E. R. (1991) T a el imes o global ea hquake loca ion and phase iden i ica ion. Geophysical Jou nal In e na ional, 105: 429–465. Kenne , L. N., Engdahl, E. R. and Buland, R. (1995) Cons ain s on he eloci y s uc u e in he Ea h om a el imes. Geophysical Jou nal In e na ional, 122: 108–124. Kepple , H. and F os , D. C. (2005) In oduc ion o mine als unde ex eme condi ions. In Mile ich, R., (edi o ), Mine al beha iou a ex eme condi ions, olume 7 o Eu opean Mine alogical Union No es in Mine alogy, p. 1- 7, Eö ös Uni e si y p ess, Budapes . Kie e , B., S ix ude, L., Ha ne , J. and K esse, G. (2001) S uc u e and elas ici y o wadsleyi e a high p essu es. Ame ican Mine alogis , 86: 1387-1395. 149 King, H. E. and Finge , L. W. (1979) Di ac ed beam c ys al cen e ing and i s applica ion o high- p essu e c ys allog aphy. Jou nal o Applied C ys allog aphy, 12: 374-378. Kleppa, O. J. (1976) Mine alogical applica ions o high– empe a u e eac ion calo ime y. In S ens, R. G. J., edi o , The physics and chemis y o Rocks and Mine als, p. 369-387, New Yo k, Wiley Kni le, E. and Jealoz, R. (1987) Syn hesis and equa ion o s a e o (Mg,Fe)SiO3 pe o ski e o o e 100 GPa. Science, 235: 688-670. Kubo, A., Yagi, T., Ono, S. and Akaogi, M. (2000) Comp essibil y o Mg0.9Al0.2Si0.9O3 pe o ski e. P oceedings o Japan Academy, 76B: 103-107. Kubo, T., Eiji Oh ani, E., Kondo, T., Ka o, T., Toma, M., Hosoya, T., Sano, A., Takumi Kikegawa ,T. and Nagase, T. (2002) Me as able ga ne in oceanic c us a he op o he lowe man le. Na u e, 420: 803-806. Lau e bach, S., McCammon, C. A., an Aken, P., Langenho s , F. and Sei e , F. (2000) Mössbaue and ELNES spec oscopy o (Mg,Fe)(Si,Al)O3 pe o ski e: a highly oxidised componen o he lowe man le. Con ibu ions o Mine alogy and Pe ology, 138: 17-26. Libe mann, R. C. and Wang, Y. (1992) Cha ac e isa ion o sample en i onmen in a uniaxial spli sphe e appa a us. In Syono, Y. and Manghnani, M. H., edi o s, High P essu e Resea ch: Applica ion o Ea h and Plane a y Sciences, p. 19–31,Te a Scien i ic Publishing Co., Tokyo/Ame ican Geophysical Union, Washing on, DC. Li, B., Liebe mann, R. C. and Weidne , D. J. (1998) Elas ic Moduli o Wadsleyi e (β-Mg2SiO4) o 7 Gigapascals and 873 Kel in. Science, 281(5377): 675-677. Li, B. and Li, B. (2003) Comp essional and shea wa e eloci ies o ingwoodi e {gamma}-Mg2SiO4 o 12 GPa. Ame ican Mine alogis , 88: 1312-1317. Li, L., B odhol , J. P., S ackhouse, S., Weidne , D. J., Al edsson, M. and P ice, G. D. (2005) Elas ici y o (Mg,Fe)(Si,Al)O3- pe o ski e a high p essu e. Ea h and Plane a y Science Le e s, 240: 529-536. 150 Li, L., Weidne , D.J., B odhol , J., Al è, D., P ice, G. D, Ca acas, R. and Wen zco i ch, R. (2006) Elas ici y o CaSiO3 pe o ski e a high p essu e and high empe a u e. Physics o Ea h and Plane a y In e io s, 155: 249–259. Li aso , K. D. and Oh ani, E. (2005) Phase ela ions in hyd ous MORB a 18-26 GPa: Implica ions o he e ogenei y o he lowe man le. Physics o Ea h and Plane a y In e io s, 150: 239-263. Mao, H. K., Hemley, R. J., Fei, Y., Shu, J. F., Chen, L. C., Jephcoa , A. P. and Wu, Y. (1991) E ec o p essu e, empe a u e and composi ion on la ice pa ame e s and densi y o (Fe,Mg)SiO3 peo ski es o 30 Gpa. Jou nal o Geophysical Resea ch, 96(B5): 8069-8079. Ma suzaka, K., Akaogi, M., Suzuki, T. and Suda, T. (2000). Mg-Fe pa i ioning be ween silica e spinel and magnesiowüs i e a high p essu e: expe imen al de e mina ion and calcula ion o phase ela ions in he sys em Mg2SiO4-Fe2SiO4. Physics and Chemis y o Mine als, 27: 310-319. McCammon, C. (1997) Pe o ski e as a possible sink o e ic i on in he lowe man le. Na u e, 387: 694-696. McCammon, C. (1998) The c ys al chemis y o e ic i on in Fe0.05Mg0.95SiO3 pe o ski e as de e mined by Mössbaue spec oscopy in he empe a u e ange 80-293K. Physics and Chemis y o Mine als, 25: 292-300. McCammon, C. (2004) Mössbaue spec oscopy: applica ions, in Spec oscopic me hods. In Be an, A. and Libowi zky, E., edi o s, Spec oscopic Me hods in Mine alogy, olume 6 o Eu opean Mine alogical Union No es in Mine alogy, p. 369-398, Eö ös Uni e si y P ess, Budapes . McCammon, C. A., Lau e bach, S., Sei e , F., Langenho s , F., and an Aken, P. A. (2004) I on oxida ion s a e in lowe man le assemblages I Empi ical ela ions de i ed om high–p essu e expe imen s. Ea h and Plane a y Science Le e s, 222: 435-449. McDonough, W. F. and Sun, S. S. (1995) The composi ion o he Ea h. Chemical Geology, 20: 223- 253. Megaw, H. D. (1973) C ys al s uc u e: a wo king app oach, Philadelpheia, Saunde s. 151 Meng, Y., Fei, Y., Weidne , D. J., Gwanmesia, G. D. and Hu, J. (1994) Hyd os a ic comp ession o γ- Mg2SiO4 o man le p essu es and 700 K: The mal equa ion o s a e and ela ed he moelas ic p ope ies. Physics and Chemis y o Mine als, 21: 407-412. Mile ich, R., Allan, D. R. and Kuhs, W. F. (2000) High-p essu e single-c ys al echniques. In Hazen, R. M. and Downs, R. T., edi o s, High-Tempe a u e and High–P essu e C ys al Chemis y, olume 41 o Re iews in Mine alogy and Geochemis y, p. 445-519, Mine alogical Socie y o Ame ica, Chan illy, Vi ginia. Mile ich, R., (2005) Di ac ion echniques: Shedding ligh on s uc u al changes a ex eme condi ions. In Mile ich, R., edi o , Mine al Beha iou a Ex eme condi ions, in olume 7 o Eu opean Mine alogical Union No es in Mine alogy, p. 281-331, Eö os Uni e si y P ess, Budapes . Mu akami, M., Hi ose, K., Kawamu a, K., Sa a, N., Ohisi, Y. (2004) Pos pe o ski ephase ansi ion in MgSiO3, Science, 304(5672): 855-858. Nasdala, L., Smi h, D. C., Kaindal, R. and Zieman, M. A. (2004) Raman Spec oscopy: Analy ical Pe spec i es in Mine alogical Resea ch. In Be an, A. and Libowi zky, E., edi o s, Spec oscopic Me hods in Mine alogy, olume 6 o Eu opean Mine alogical Union No es in Mine alogy, p. 281-243, Eö ös Uni e si y P ess, Budapes . Na o sky, A. (1997) P og ess and New di ec ions in High empe a u e calo ime y e isi ed. Physics and Chemis y o Mine als, 24: 22-241. Na o sky, A. (1999) A lesson om ce amics, Science, 284: 1788-1789. Na o sky, A., Schoeni z, M., Koji ani, H., Xu, H., Zhang, J., Weidne , D. J., Jeanloz and R. (2003) Aluminium in magnesium silica e pe o ski e: o ma ion, s uc u e, and ene ge ics o magnesium– ich de ec solid solu ions. Jou nal o Geophysical Resea ch, 108(B7), 2330, doi: 10.1029/2002JB002055. Na o sky, A. (2004) The mochemis y, ene ge ic modeling and sys ema ics. In G amaccioli, C, edi o , Ene ge ic modeling in mine als, olume 4 o Eu opean Mine alogical Union No es in Mine alogy, p. 5-31, Eö ös Uni e si y P ess, Budapes . 152 New on, R. C., Cha lu, T. V. and Kleppa, O. J. (1977) The mochemis y o high p essu e ga ne s and clinopy oxenes in he sys em CaO-MgO-Al2O3-SiO2. Geochimica e Cosmochimica Ac a, 41(3): 369- 377. Nishiha a, Y. and Takahashi, E. (2001) Phase ela ion and physical p ope ies o an Al-deple ed koma ii e o 23 GPa. Ea h and Plane a y Science Le e s, 190: 65-77. Nishio-Hammane, D., Nagai, T., Fujino, K., Se o, Y., Taka uji, N. (2005) Fe3+ and Al solubili ies in MgSiO3 pe o ski e: implica ions o he Fe3+AlO3 subs i u ion in MgSiO3 pe o ski e a he lowe man le condi ions. Geophysical Resea ch Le e s, 32:L16306, doi: 10.1029/2005GL023529. Ono, S., Kikegawa, T. and Iizuka, T. (2004) The equa ion o s a e o o ho hombic pe o ski e in a py oli ic man le up o 80 GPa: Implica ions o chemical composi ion o he lowe man le, Physics o he Ea h and Plane a y In e io s, 145(1-4): 9-17. O’Neill, H. S . C. and Wood, B. J. (1979) An expe imen al s udy o Fe-Mg pa i ioning be ween ga ne and oli ine and i s calib a ion as a geo he mome e . Con ibu ions o Mine alogy and Pe ology, 70: 59–70. Ralph, R. L. and Finge , L. W. (1982) A compu e p og am o e inemen o c ys al o ien a ion on ma ix and la ice cons an s om di ac ome e da a wi h la ice symme y cons ain s. Jou nal o Applied C ys allog aphy, 15: 537-539. Reed, S. (1996) Elec on Mic op obe Analysis and Scanning Elec on Mic oscopy in Geology. Camb idge Uni e si y P ess, Camb idge, p. 187. Richmond, N. C. and B odhol , J. P. (1998) Calcula ed ole o aluminium in he inco po a ion o e ic i on in o magnesium silica e pe o ski e. Ame ican Mine alogis , 83(9-10): 947-951. Ringwood, A. E. (1967) The py oxene-ga ne ans o ma ion in he Ea h’s man le. Ea h and Plane a y Science Le e s, 2: 255-263. Ringwood, A. E. and Majo , A. (1967) Some high-p essu e ans o ma ions o geophysical signi icance. Ea h and Plane a y Science Le e s, 2, 106-110. Ringwood, A. E. (1975) Composi ion and pe ology o he Ea h’s man le. McG aw-Hill, New Yo k, p. 618. 153