1:    % This document demonstrates how to use commands

2:    % described in the aguguide.tex user manual to create GRL

3:    % manuscripts for on-line submission.  Most

4:    % instructions and explanations are commented out;

5:    % thus they will not appear if you LaTeX this file

6:    % and view it as a sample document.

7: 

8:    % You may use this file as a LaTeX document template for

9:    % your GRL manuscript by replacing the sample text with

10:    % your own text and deleting any extraneous commands.

11: 

12: 

13: %%   PLEASE NOTE:

14: %%   1) Don't include any extra style files.

15: %%   2) Use LaTeX "thebibliography" environment for the reference

16: %%      list.  Use \cite to cite references in the main text.

17: %%

18: %% ------------------------------------------------------ %%

19: 

20:    % The first item in a LaTeX file should be a \documentstyle

21:    % command to declare the overall style of the paper.  The

22:    % \documentstyle line that is relevant for GRL manuscripts

23:    % is shown below.

24: 

25: %\documentstyle[12pt,pomegranate]{article}

26: %\documentstyle[12pt,agums,epsfig]{article}

27: %\documentstyle[grlga]{article}

28: %\documentstyle[jgrga,epsfig]{article}

29: \documentstyle[agupp,epsfig]{article}

30: 

31: %\documentstyle[12pt,agums]{article}

32:     % The agums (manuscript) style creates double-spaced wide-margined

33:     % manuscripts.  Use it to create copy to submit to

34:     % GRL editors for first-time review.

35: 

36: %\documentstyle[agupp]{article}

37:       % This substyle creates two-column preprints

38:    % (i.e., for distribution to colleagues; do not

39:         % submit two-column manuscripts to AGU).

40: 

41: %documentstyle[grlga]{article}

42:     % The grlga (galley) style can be used to create

43:     % camera-ready pages. You can use this to measure

44:     % the size of your manuscript.

45: 

46: %% ------------------------------------------------------ %%

47: %%

48: %%  PREAMBLE

49: %%

50: %%

51:    % Running heads, manuscript dates, manuscript

52:    % information, copyright information, author

53:    % addresses and slug comments are placed in the

54:    % preamble (before the \begin{document} command).

55:    %

56:    % AGU's style files cause the actual text to be

57:    % typeset after the reference section.  If you

58:    % have no reference section, you must include a

59:    % \forcesluginfo command or the preamble information

60:    % will not print.

61: 

62: 

63: %% ------------------------------------------------------ %%

64: %

65: %%  RUNNING HEADS

66: %

67: %% ------------------------------------------------------ %%

68: 

69: \lefthead{MARTON ET AL.}

70:     % Type only the last name(s) of the author(s) in all

71:     % capital letters between the curly brackets in the

72:     % \lefthead command.  If your manuscript has three

73:     % or more authors, type only the last name of the

74:     % first author followed by "ET AL."

75: 

76: %          1 <----------------------- SIXTY -----------------------> 60

77: \righthead{PEROVSKITE EOS AND LOWER MANTLE COMPOSITION}

78:     % Type a short version of the article title (or

79:     % include the entire title, if it will fit) in

80:     % all capital letters between the curly brackets

81:     % in the \righthead command.  You may type a total

82:     % of 60 characters (including letters, punctuation,

83:     % and spaces) between the curly brackets in the

84:     % \lefthead or \righthead commands (i.e., 10

85:     % characters in one command and 50 in the other

86:     % is acceptable).

87: 

88: %% ------------------------------------------------------ %%

89: %

90: %%  MANUSCRIPT DATES

91: %

92: %% ------------------------------------------------------ %%

93: 

94: \received{March 29, 2000}

95: \revised{September 28, 2000}

96: \accepted{December 5, 2000}

97:    % Editors supply the received, revised, and

98:    % accepted dates.  Type your manuscript dates

99:    % between the curly brackets in the \received,

100:    % \revised, and \accepted commands.  Use the

101:    % date format shown above.

102: 

103: %% ------------------------------------------------------ %%

104: %

105: %%  MANUSCRIPT INFORMATION  (GRL authors may leave these

106: %%                           commands blank)

107: %

108: %% ------------------------------------------------------ %%

109: 

110: \paperid{2000JB900457}

111:    % Leave this empty.

112: 

113: \cpright{AGU}{2001}

114:    % \cpright{PD}{1998}

115:    % \cpright{Crown}{1998}

116:    % If your paper is AGU copyright, choose the

117:    % "AGU" \cpright command.  If your paper is in

118:    % the public domain, choose the "PD" \cpright

119:    % command.  If your paper is Crown copyright,

120:    % choose the "Crown" \cpright command.  If you

121:    % are not sure which copyright to choose, please

122:    % contact your production coordinator.  You must

123:    % choose one of these copyright options.

124: 

125: \ccc{0148-0227/01/2000JB900457\$09.00}

126:    % Type your Copyright Clearance Center code between

127:    % the curly brackets in the \ccc command.  Crown

128:    % copyrights and manuscripts in the Public Domain

129:    % have no "\ccc{}" information.  Be sure to place

130:    % a backslash in front of the dollar sign or it

131:    % will be interpreted as a math command.

132: 

133: %% ------------------------------------------------------ %%

134: %

135: %%  AUTHOR ADDRESSES

136: %

137: %% ------------------------------------------------------ %%

138: 

139: \authoraddr{R.~E.\/ Cohen, and F.~C.\/ Marton,

140: Geophysical Laboratory and Center for High Pressure Research,

141: Carnegie Institution of Washington,

142: 5251 Broad Branch Road, NW, Washington, D. C. 20015-1305.

143: (cohen@gl.ciw.edu; marton@gl.ciw.edu)}

144: 

145: \authoraddr{J.\/ Ita, Shell International Exploration and

146: Production, P. O. Box 60, Rijswijk, 2280 AB, Netherlands.

147: (j.j.ita@siep.shell.com)}

148: 

149:    % For author postal addresses, group authors

150:    % by affiliation and list them in alphabetical

151:    % order.  Names include first initial only,

152:    % optional middle initial, and last name.  If

153:    % available, include e-mail addresses as shown.

154:    %

155:    % E-mail addresses may be broken over a period

156:    % or after the @ sign.  If your e-mail address

157:    % includes an underscore, do not forget

158:    % to place a backslash in front of it or it will

159:    % be interpreted as a command (see e-mail in

160:    % second \authoraddr for example).  You may add

161:    % additional \authoraddr commands if necessary.

162:    %

163:    % Note that author addresses include complete

164:    % postal address information, while author

165:    % affiliations include only department,

166:    % institution, town, and state (if state is

167:    % not included in the name of the institution)

168:    % or country (if not the United States).

169: 

170: %% ------------------------------------------------------ %%

171: %

172: %%  SLUG COMMENTS (optional)

173: %

174: %% ------------------------------------------------------ %%

175: 

176:   % Slug comments appear only in manuscript and

177:   % preprint styles, not in the camera-ready styles.

178: 

179: %% ------------------------------------------------------ %%

180: %

181: %%  CITE COMMANDS WITH OLD VERSION OF LATEX 2.09

182: %

183: %% ------------------------------------------------------ %%

184: 

185:   % \input{update.tex}

186: 

187:    % ONLY if you are using a VERY OLD version of LaTeX 2.09

188:    % and include \cite commands, LaTeX may have trouble

189:    % with the \reset@font command.  If so, LaTeX will pause

190:    % at every undefined citation the first time the input

191:    % file is LaTeXed.  Authors may either press [Enter]

192:    % each time this occurs in their first LaTeX run, or

193:    % they may type \input{update.tex} in front

194:    % of the \begin{document} command.  The "update.tex" file

195:    % is available on the kosmos.agu.org anonymous ftp site.

196: 

197: %% ------------------------------------------------------ %%

198: %

199: %%  NUMBERING YOUR SECTIONS

200: %

201: %% ------------------------------------------------------ %%

202: 

203: \setcounter{secnumdepth}{4}

204: 

205:    % If you want LaTeX to automatically number your

206:    % sections, type \setcounter{secnumdepth}{4} in

207:    % the preamble of your input file, as shown. If

208:    % you do not want section numbers, type

209:    % \setcounter{secnumdepth}{0}.

210: 

211: %% ------------------------------------------------------ %%

212: %%

213: %%  \BEGIN{DOCUMENT} and FRONT MATTER

214: %%

215: %%

216:    % The body of the paper starts with the \begin{document}

217:    % command, which is followed by the front matter (title,

218:    % author, affiliation, and abstract).

219: 

220: \begin{document}

221: 

222: 

223: %% ------------------------------------------------------ %%

224: %

225: %%  TITLE

226: %

227: %% ------------------------------------------------------ %%

228: 

229: \title{Pressure-volume-temperature equation of state \\

230: of MgSiO$_{3}$ perovskite from molecular dynamics \\

231: and constraints on lower mantle composition}

232: 

233:   % Type the title of your manuscript between the

234:   % curly brackets in a \title command.  Capitalize

235:   % only acronyms, first letter of the first word,

236:   % first letter of proper nouns, first letter of

237:   % the first word after colons, and first letter

238:   % of the first word of a subtitle.  If the title

239:   % exceeds one line, break it so that the first line

240:   % is longer than the second line; break the title

241:   % before articles, prepositions, and conjunctions.

242:   % To break the title, type a double backslash where

243:   % you want the break to occur, as shown above.

244: 

245: 

246: %% ------------------------------------------------------ %%

247: %

248: %%  AUTHOR NAMES, AFFILIATIONS, and ALTERNATE AFFILIATIONS

249: %

250: %% ------------------------------------------------------ %%

251: 

252: \author{Frederic C.\/ Marton, Joel Ita,\altaffilmark{1} and Ronald

253: E. Cohen}

254: %

255: \affil{Geophysical Laboratory and Center for High Pressure Research,

256: \\Carnegie Institution of Washington, Washington, D. C.}

257: 

258: \altaffiltext{1}{Now at Shell International Exploration and

259: Production, Rijswijk, Netherlands.}

260: 

261: %%%% Type author names between the curly brackets in

262:    % the \author command(s).  Author names consist of

263:    % first name or initial, optional middle name or

264:    % initial, and last name.  If it is necessary to

265:    % break lines, make sure you break them between

266:    % two author's names (using \linebreak commands).

267: 

268: %%%% Type affiliation information between the curly

269:    % brackets in an \affil command that immediately

270:    % follows each \author command.

271: 

272: %%%% If your authors have alternate affiliations,

273:    % type the footnote number between the curly

274:    % brackets in an \altaffilmark command and type

275:    % the alternate affiliation information in

276:    % \altaffiltext commands below the \author and

277:    % \affil commands, as shown.  There is a separate

278:    % \altaffiltext for each alternate affiliation

279:    % indicated above.  This information prints

280:    % as footnotes after the reference section in

281:    % your camera-ready copy, and will be placed

282:    % on the bottom left-hand corner of the first

283:    % page of your published manuscript.

284: 

285: %% ------------------------------------------------------ %%

286: %

287: %%  FOOTNOTED AFFILIATIONS    (For more than three

288: %%                             author/affiliation groups.)

289: %

290: %% ------------------------------------------------------ %%

291: 

292:    % Important:  If your manuscript contains more than three

293:    % author and affiliation groupings, all of the affiliate

294:    % information should be footnoted using \altaffilmark{}

295:    % commands as shown.  Please include commas inside the

296:    % footnote numbers.

297: 

298:    % \author{   S. Kurtz,\altaffilmark{1}

299:    %            Evan R. King,\altaffilmark{1,2}

300:    %            A. E. Simon,\altaffilmark{1,3}

301:    %            R. J. Hall,\altaffilmark{4}

302:    %            C. D. Qian,\altaffilmark{5}

303:    %            Jason M. Albert,\altaffilmark{5}

304:    %            G. G. Kaye,\altaffilmark{6}

305:    %            and Antony Ignatov\altaffilmark{7}     }

306:    %

307:    % \altaffiltext{1}{Astronomy Department,

308:    %                  University of California, Berkeley.}

309:    % \altaffiltext{2}{Also at Physics Department,

310:    %                  University of California, Berkeley.}

311:    % \altaffiltext{3}{Now at National Center for Atmospheric

312:    %                  Research, Boulder, Colorado.}

313:    % \altaffiltext{4}{Space Telescope Science Institute,

314:    %                  Baltimore, Maryland}

315:    % \altaffiltext{5}{Woods Hole Oceanographic Institution,

316:    %                  Woods Hole, Massachusetts}

317:    % \altaffiltext{6}{Space Telescope Science Institute,

318:    %                  Baltimore, Maryland}

319:    % \altaffiltext{7}{Department of Astronomy, California

320:    %                  Institute of Technology, Pasadena.}

321: 

322: 

323: %% ------------------------------------------------------ %%

324: %

325: %%  ABSTRACT

326: %

327: %% ------------------------------------------------------ %%

328: 

329: \begin{abstract}

330: The composition of the lower mantle can be investigated

331: by examining densities and seismic velocities of

332: compositional models as functions of depth.  

333: In order to do this it is necessary to know the 

334: volumes and thermoelastic properties of the compositional

335: constituents under lower mantle conditions.

336: We determined the thermal equation of state (EOS)

337: of MgSiO$_{3}$ perovskite using the nonempirical 

338: variational induced breathing (VIB) interatomic potential

339: with molecular dynamics simulations

340: at pressures and temperatures of the lower mantle.

341: We fit our pressure-volume-temperature results to a thermal EOS

342: of the form

343: $P(V,T)$ = $P_{0}(V,T_{0})$ + $\Delta P_{\rm th}(T)$, where

344: $T_{0}$ = 300 K and $P_{0}$ is the isothermal Universal EOS.

345: The thermal pressure $\Delta P_{\rm th}$ can be represented

346: by a linear relationship $\Delta P_{\rm th} = a + b T$.

347: We find $V_{0}$ = 165.40 \AA$^{3}$, $K_{T_{0}}$

348: = 273 GPa, $K^{\prime}_{T_{0}}$ = 3.86,

349: $a$ = -1.99 GPa, and $b$ = 0.00664 GPa K$^{-1}$

350: for pressures of 0-140 GPa and temperatures of 300-3000 K.

351: By fixing $V_{0}$ to the experimentally determined value

352: of 162.49 \AA$^{3}$ and

353: calculating density and bulk sound velocity profiles

354: along a lower mantle geotherm

355: we find that the lower mantle cannot consist solely of

356: (Mg,Fe)SiO$_{3}$ perovskite with X$_{\rm Mg}$ ranging from

357: 0.9-1.0.  Using pyrolitic compositions of 67 vol \%

358: perovskite (X$_{\rm Mg}$ = 0.93-0.96) and 33 vol \% magnesiow\"ustite

359: (X$_{\rm Mg}$ = 0.82-0.86), however, we obtained density and

360: velocity profiles that are in excellent agreement

361: with seismological models for a reasonable geotherm.

362: \end{abstract}

363: 

364:    % The abstract consists of a one-paragraph summary

365:    % of your paper, 250 words or fewer.  Do not cite

366:    % references unless absolutely necessary.  If you

367:    % must cite, place the citation in square brackets

368:    % using italic type (this abstract citation format

369:    % is new).  Do not include displayed material.

370: 

371: %% ------------------------------------------------------ %%

372: %

373: %%  SECTION HEADS

374: %

375: %% ------------------------------------------------------ %%

376: 

377: %%%% LEVEL 1 HEADS

378:    %

379:    % Use \section{} commands to identify Level 1 heads;

380:    % type the appropriate Level 1 head between the curly

381:    % brackets, as shown.

382:    %

383:    % Capitalize the first letter of each word (except for

384:    % prepositions, conjunctions, and articles that are 3

385:    % letters or shorter).

386:    %

387:    % Do not hyphenate Level 1 heads.  To break lines,

388:    % type \protect\\ where you want a break to occur.

389:    %

390:    % If you wish to number your sections, see the

391:    % "Numbering Sections" information in the preamble,

392:    % above.

393: 

394: % \section{Introduction}

395: 

396: %%%% LEVEL 2 HEADS

397:    %

398:    % Use \subsection{} command to identify Level 2

399:    % heads; type the appropriate Level 2 head between

400:    % the curly brackets, as shown.

401:    %

402:    % Capitalize the first letter of each word (except for

403:    % prepositions, conjunctions, and articles that are 3

404:    % letters or shorter).

405:    %

406:    % Do not hyphenate Level 2 heads.  To break lines,

407:    % type \protect\\ where you want a break to occur.

408: 

409: % \subsection{This is a Level Two Head}

410: 

411: 

412: %%%% LEVEL 3 HEADS

413:    %

414:    % Use the command \subsubsection{} to identify

415:    % Level 3 heads; type the appropriate head between

416:    % the curly brackets, as shown.

417:    %

418:    % Capitalize only acronyms, the first letter of the

419:    % first word, first letter of proper nouns, and first

420:    % letter of the first word after colons.

421:    %

422:    % Hyphenation is permitted in Level 3 heads, if needed.

423:    %

424:    % You must include at least two Level 3 heads per

425:    % Level 2 head.

426: 

427: % \subsubsection{This is a level three head}

428: 

429: 

430: %%%% LEVEL 4 HEADS

431:    %

432:    % Use the command \subsubsubsection{} to identify

433:    % Level 4 heads; type the appropriate head between

434:    % the curly brackets, as shown.

435:    %

436:    % Capitalize only acronyms, the first letter of the

437:    % first word, first letter of proper nouns, and first

438:    % letter of the first word after colons.

439:    %

440:    % Hyphenation is permitted in Level 4 heads, if needed.

441:    %

442:    % A level 4 head cannot directly follow a Level

443:    % 3 head; there must be a least one sentence between

444:    % the two heads.

445: 

446: % \subsubsubsection{This is a level four head}

447: 

448: 

449: %% ------------------------------------------------------ %%

450: %%

451: %%  TEXT

452: %%

453: %%

454:    % The body of the paper must start with a

455:    % \begin{article} command, and you must include

456:    % an \end{article} command after the bibliography

457:    % section (otherwise your text will not print at

458:    % the proper column width).

459: 

460:    % The first \section{} command is optional.  You may

461:    % delete the command or you may replace the word

462:    % "Introduction" with text appropriate to your article.

463: 

464: \begin{article}

465: \section{Introduction}

466: By comparing density and seismic velocity profiles of

467: compositional models with seismological models it is possible

468: to investigate the composition of the Earth's lower mantle.

469: Therefore an understanding of the high pressure, high temperature

470: properties, and equation of state (EOS) of

471: (Mg,Fe)SiO$_{3}$ perovskite is vital to our understanding

472: of the lower mantle, as this mineral accounts for

473: perhaps two thirds of the mineralogy of this region \cite{bina98}

474: and one third of the volume of the entire

475: planet.  Experiments have been performed

476: up to pressures approaching those found at the base of the mantle,

477: but direct coverage of the lower mantle geotherm has been limited

478: to perhaps the uppermost one third (\callout{Figure~\ref{fig1}}).

479: Experiments include resistively heated multianvil apparati 

480: \cite{wang94, kato95, utsumi95, funamori96}, 

481: diamond anvil cells (DACs) at ambient temperature

482: \cite{kudoh87, ross90},

483: resistively heated DACs \cite{mao91, saxena99},

484: laser heated DACs \cite{knittle87,fiquet00}, and

485: shock wave experiments \cite{duffy93, akins99}.

486: 

487: Theoretical methods, such as those used here, permit the 

488: investigation of thermodynamic and thermoelastic properties of

489: minerals at high pressures and temperatures.

490: Previous theoretical work on orthorhombic

491: MgSiO$_{3}$ perovskite (Mg-pv) 

492: has been done using molecular dynamics (MD) using the 

493: semi\-empirical potential of {\it Matsui} [1988]

494: at 0-1000 K and 0-10 GPa (at 300 K) \cite{matsui88} and 300-5500 K

495: and 0-400 GPa \cite{belonoshko94, belonoshko96} and combined 

496: with lattice dynamics at 500-3000 K and 0-100 GPa \cite{patel96}.

497: {\it Wolf and Bukowinski} [1987] used a rigid ion

498: modified electron gas (MEG) model at 0-800 K and 0-150 GPa,

499: while {\it Hemley et al.} [1987] used a MEG model combined

500: with shell stabilized ion charge densities at 0-2500 K and

501: 0-200 GPa.  {\it Cohen} [1987] did quasi-harmonic

502: lattice dynamics calculations from 0-3000 K

503: and 0-150 GPa using the potential induced breathing (PIB) 

504: model.  First principles static lattice calculations ($T$ = 0 K)

505: have also been done up to pressures of 150 GPa, using plane wave

506: pseudopotential (PWPP) MD \cite{wentzcovitch95} and the

507: linearized augmented plane wave (LAPW) method \cite {stixrude93}.

508: {\it Karki et al.} [1997] also used the PWPP method to

509: examine the athermal elastic moduli of Mg-pv.

510: 

511: In addition to the equation of state the thermodynamic

512: stability of perovskite in the lower mantle is an open

513: question and has been studied with experimental

514: and theoretical methods.  Work by {\it Meade et al.} [1995],

515: {\it Saxena et al.} [1996, 1998], and

516: {\it Dubrovinsky et al.} [1998]

517: indicate that perovskite will break down to oxides under

518: lower mantle conditions.  

519: The thermodynamic analysis of {\it Stixrude and

520: Bukowinski} [1992] and the experimental 

521: work of {\it Serghiou et al.} [1998],

522: however, suggest that it will not.

523: 

524: We used MD simulations using the nonempirical variational

525: induced breathing (VIB) model, similar to the model of

526: {\it Wolf and Bukowinski} [1988],

527: to investigate the properties and EOS of Mg-pv

528: at pressures (0-140 GPa) and temperatures (300-3000 K)

529: that cover the bulk of the conditions found in the lower mantle.

530: Newton's equations of motion are solved as functions of time,

531: and equilibrium properties are obtained as time averages

532: over a sufficiently long interval.  

533: This accounts for all orders of anharmonicity but not for

534: quantum lattice dynamics effects.  Thus our results are more

535: appropriate at high temperatures above the Debye temperature

536: (1076 K \cite{anderson98}) and are entirely suitable for the Earth's mantle.

537: We then used our results, combining them with data for other

538: components and phases where appropriate, to calculate density and

539: seismic velocity profiles along a geothermal gradient for different

540: compositional models.  These profiles were then compared with 

541: profiles from a reference Earth model in order to test the 

542: compositional models' fitness for the lower mantle.

543: 

544: \section{Method}

545: MD simulations were performed on

546: supercells of 540-2500 atoms of orthorhombic ({\it Pbnm})

547: Mg-pv using the nonempirical VIB potential.

548: VIB is a Gordon-Kim type model \cite{gordon72},

549: where the potential is obtained by overlapping 

550: ionic charge densities which are computed using the local

551: density approximation (LDA) \cite{hedin71}.  The total energy is a sum of

552: (1) the self-energy of each atom, (2) the long-range 

553: electrostatic energy computed using the Ewald method, and

554: (3) the short-range interaction energy, i.e.,

555: the sum of the kinetic, short-range electrostatic,

556: and exchange-correlation overlap energies from the LDA.

557: Free oxygen ions are not stable and are stabilized 

558: in VIB by surrounding

559: them with Watson spheres with charges equal in magnitude but opposite

560: in sign to the ions, e.g., 2+ spheres for O$^{2-}$.  

561: Interactions are obtained for overlapping ion pairs at different

562: distances with different Watson sphere radii for each oxygen.

563: The interactions are fit with a 23-parameter analytical

564: expression as functions of the interatomic distance $r$ and

565: $U_{i} = z_{i} / R_{i}$, where $U_{i}$, $z_{i}$, and $R_{i}$ 

566: are the Watson sphere potential, charge, and radius for atom $i$.

567: For each oxygen $i$, $R_{i}$ is adjusted to minimize the total

568: energy at each time step.

569: This gives an effective many-body potential.

570: The oxygen ions respond to changes in their environment.

571: For example, they are compressed at high pressures relative

572: to low pressures.

573: Previous work on Mg-pv using 

574: the related PIB model

575: gave an equation of state in excellent agreement with 

576: experiment \cite{cohen87}.  Also, work using VIB

577: on MgO has shown that this model accurately predicts

578: EOS and thermal properties \cite{inbar95}.

579: 

580: Nominally charged ions give semiquantitative, but

581: less accurate, results than desired for Mg-pv.

582: This is due to a small degree of covalent bonding relative

583: to ionic bonding as revealed by electronic structure

584: calculations of cubic ($Pm$3$m$) Mg-pv 

585: performed using the first principles LAPW method \cite{cohen89}.

586: These calculations show that while the Mg is nearly a perfectly

587: spherical Mg$^{2+}$ ion, there is some charge transfer from

588: O to Si and a small covalent bond charge 

589: (involving $\ll$ 0.1 electrons) between Si and O.

590: By varying the ionic charges on Si (3.1+ to 3.55+) and

591: O (1.70- to 1.85-) to account for the covalency,

592: but otherwise using the same methods as before, 

593: we compared the resulting zero temperature pressure-volume data to

594: the LAPW results of 

595: {\it Stixrude and Cohen} [1993].  The best agreement

596: was found with Si$^{3.4+}$ and O$^{1.8-}$.

597: Using these charges, the resulting potential gives

598: excellent agreement with the octahedral rotation energetics

599: obtained using the LAPW method \cite{stixrude93, hemley96}

600: (\callout{Figure~\ref{fig2}}).

601: 

602: MD runs were performed for 20 ps with a 1 fs time step

603: using a sixth-order Gear predictor-corrector scheme

604: \cite{gear71} in the constant pressure-tempera\-ture ensemble

605: using the thermostat and barostat of {\it Martyna et al.} [1994].

606: Initial atomic positions for the genesis run were the same 

607: as those for the unrotated octahedra of

608: {\it Stixrude and Cohen} [1993].  

609: Subsequent runs at higher pressures or temperatures

610: used the positions generated by a previous run.

611: Statistical ensembles were obtained in $\sim$2000 iterations

612: for the genesis run and in $<$ 1000 iterations for subsequent runs.

613: 

614: We fitted pressure-volume-temperature (P-V-T) 

615: data obtained from the MD simulations

616: to a thermal EOS of the form

617: \begin{equation}

618:    P(V,T) = P_{0}(V,T_{0}) + \Delta P_{\rm th}(T)

619: \end{equation}

620: with a reference temperature $T_{0}$ = 300 K

621: and the thermal pressure $\Delta P_{\rm th}$ relative to the 300 K

622: isotherm.  We analyzed our results using the third-order

623: Birch-Murnaghan isothermal EOS \cite{birch52}

624: \begin{equation}

625:    P_{0} = 3 K_{T_{0}} f (1 + 2f)^{5/2} (1 - \xi f),

626: \end{equation}

627: with the Eulerian strain variable $f$ and the coefficient

628: $\xi$ given by

629: \begin{equation}

630:    f \equiv \frac{1}{2} \left[ \left( \frac {V_{0}}{V} \right)

631:    ^{2/3} - 1 \right] , ~~~~

632:    \xi \equiv - \frac{3}{2} ( K^{\prime}_{T_{0}} - 4 ),

633: \end{equation}

634: and the Universal EOS \cite{vinet87}

635: \begin{equation}

636:    P_{0} = 3 K_{T_{0}} (1 - x) \, x^{-2} \exp \left[ 

637:    \frac {3}{2} \, (K^{\prime}_{T_{0}} - 1) (1 - x) \right],

638: \end{equation}

639: where $x = ( V / V_{0} )^{1/3}$,

640: $K_{T}$ is the isothermal bulk modulus, and $K^{\prime}_{T}$

641: is its pressure derivative.

642: 

643: $\Delta P_{\rm th}$ is given by

644: \begin{equation}

645:    \Delta P_{\rm th} = \int^{T}_{T_{0}} \left( \frac {\partial {P}}

646:    {\partial T} \right)_{V} d{\hat{T}} =

647:    \int^{T}_{T_{0}} ( \alpha K_{T} ) d{\hat{T}},

648: \end{equation}

649: where $\alpha$ is the volume coefficient of thermal expansion.

650: If $\alpha K_{T}$ is independent of T, then the thermal

651: pressure can be represented by a linear relationship

652: \cite{anderson80,anderson83}

653: \begin{equation}

654:    \Delta P_{\rm th} = a + b T ,

655: \end{equation}

656: which applies 

657: to a wide range of solids at high $T$, including minerals,

658: alkali metals, and noble gas solids \cite{anderson84}.

659: The linearity of $\Delta P_{\rm th}$

660: in $T$ starts at 300 K \cite{masuda95} in minerals.

661: There should be a small anharmonic correction at high $T$, which

662: results in an additional $c\,T^{2}$ term in (6),

663: which is often sufficiently small so that it can be neglected

664: \cite{anderson84}.  Indeed, including the $c\,T^{2}$ term

665: did not statistically improve our fits.  Likewise,

666: our fits were not improved by including volumetric compression terms 

667: [e.g., {\it Jackson and Rigden}, 1996].

668: 

669: Once the P-V-T data were fit, other parameters, such as

670: $\alpha$,

671: \begin{equation}

672:    \alpha = { \left( \frac {\partial {\:\ln V}}

673:                 {\partial T} \right) }_{P},

674: \end{equation}

675: its volume dependence, the Anderson-Gr\"uneisen parameter,

676: \begin{equation}

677:    \delta_{T} = { \left( \frac {\partial {\:\ln \alpha}}

678:                 {\partial {\:\ln V}} \right) }_{T},

679: \end{equation}

680: the Gr\"uneisen parameter,

681: \begin{equation}

682:    \gamma = \frac {\alpha K_{T} V}{C_{V}},

683: \end{equation}

684: and its volume dependence,

685: \begin{equation}

686:    q = { \left( \frac {\partial {\:\ln \gamma}}

687:        {\partial {\:\ln V}} \right) }_{T},

688: \end{equation}

689: were obtained by numerical differentiation.  $C_{P}$ was

690: found by calculating enthalpy ({\it U \rm + PV\,}) 

691: at each MD run point and

692: differentiating with respect to $T$.

693: Using the relation $C_{P} = C_{V} (1 + \gamma \alpha T)$,

694: equation (9) can be rewritten as

695: \begin{equation}

696:    \gamma = \left( \frac {C_{P}}{\alpha K_{T} V} -

697:             T \alpha \right)^{-1}.

698: \end{equation}

699: 

700: \section{Results}

701: MD calculations were performed at 46 pressure-temp\-er\-ature

702: (P-T) points, ranging from 0-140 GPa and 300-3000 K, thus covering the

703: P-T conditions of the lower mantle (\callout{Table~\ref{tbl1}}).  

704: Axial ratios $b/a$ and $c/a$ follow smooth trends at lower mantle

705: pressures (\callout{Figures~\ref{fig3}}a-3b), with deviations

706: away from the {\it Pbnm} structure seen at $P$ $\le$ 12 GPa

707: and $T$ $\ge$ 2400 K.

708: At 3000 K and 12 GPa the structure can be observed

709: to move towards cubic lattice parameters 

710: ($b/a$ $\rightarrow$ 1, $c/a$ $\rightarrow$

711: $\sqrt{2}\,$).  This point,

712: in particular, is close to the melting curve of Mg-pv

713: \cite{heinz87,knittle89,poirier89}, and such temperature-induced

714: phase transitions have been predicted for Mg-pv by MD

715: \cite{wolf87} and lattice dynamics \cite{matsui91}.

716: In addition, it has been found for RbCaF$_{3}$

717: and KMnF$_{3}$ perovskites by MD simulations \cite{lewis89,nose89}

718: and observed experimentally in the fluoride perovskite

719: neighborite (NaMgF$_{3}$) \cite{chao61}.

720: Regardless of these changes, the volumes are well-behaved

721: as functions of pressure and temperature (Figure 3c),

722: and no other structural anomalies or melting was

723: encountered during runs at other P-T points.

724: 

725: The resulting P-V-T EOS

726: fits and their reduced $\chi^{2}$ values 

727: are given in \callout{Table~\ref{tbl2}}.  Both the 

728: Birch-Murnaghan and Universal EOSs are in excellent agreement

729: with experimental results.  As for the accuracy of our thermal

730: pressure expression, RMS differences between volumes found using

731: equation (1) for either EOS and those found using isothermal EOSs are of 

732: the order of 10$^{-3}$ \AA$^{3}$.  As for the differences between

733: the two equation of state forms, {\it Jeanloz} [1988] 

734: compared them over moderate

735: compressions and found that they are quite similar.  More recent

736: work by {\it Cohen et al.} [2000] supports this, but they found

737: that for large ($>$30\%) strains and for determining parameters such

738: as $V_{0},~ K_{T_{0}}, ~{\rm and } ~ K^{\prime}_{T_{0}}$, the

739: Universal EOS works best.  So, for the rest of our analyses

740: we use that here.  However, given that under the pressure range

741: studied, compressions will be no greater than 25\%, so we can

742: confidently compare our results with those of others that were

743: found with the Birch-Murnaghan EOS.

744: 

745: Once we determined the EOS parameters, we determined volumes

746: for Mg-pv over a wide P-T range

747: \linebreak

748: (-10-150 GPa, 0-3300 K).  Using this extended data set, we

749: determined the isothermal bulk modulus 

750: $[ K_{T} =  

751: %\linebreak 

752: -V ( {\partial P}/{\partial V} )_{T} ]$

753: at those points and

754: $\alpha$, $\delta_{T}$, $\gamma$, and $q$

755: using equations (7), (8), (11), and (10).

756: We found $({\partial K}/{\partial T} )_{P=0}$

757: = -0.0251 GPa K$^{-1}$, close to those found experimentally,

758: -0.023 GPa K$^{-1}$ \cite{wang94} and -0.027 GPa K$^{-1}$ \cite{fiquet98},

759: as well as one found by the inversion of multiple experimental

760: data sets, -0.021 GPa K$^{-1}$ \cite{jackson96}.

761: 

762: \callout{Figure~\ref{fig4}} shows $\alpha$ at pressures

763: from 0 to 140 GPa.  Experimental results shown have higher

764: $T$ slopes, as well as higher extrapolated values for most temperatures,

765: though an average value found by {\it Jackson and Rigden} [1996]

766: at zero GPa of $2.6 \times 10^{-5} \: {\rm K}^{-1}$ over 300-1660 K

767: matches our value over the same $T$ range exactly.

768: {\it Kato et al.} [1995] found an average value of 

769: $2.0 \pm 0.4 \times 10^{-5} \: {\rm K}^{-1}$ over 298-1473 K and 25 GPa, 

770: close to our average value of $1.89 \times 10^{-5} \: {\rm K}^{-1}$

771: at the same conditions.  The lack of quantum effects in our MD

772: results can be seen at low temperatures;

773: instead of tending toward zero at zero temperature,

774: $\alpha$ remains large.

775: The $\delta_{T}$ increases with $T$, but decreases

776: as $P$ increases (\callout{Figure~\ref{fig5}}).  The dependency

777: of $\delta_{T}$ on $T$ decreases as a function of $P$, with

778: convergence to a value of 2.87 at $\sim$130 GPa, similar to the 

779: high-pressure behavior in MgO found by the same MD method \cite{inbar95}.

780: Our value of $\delta_{T}$ = 3.79 at 

781: ambient conditions is in excellent agreement with the 

782: experimentally derived value of {\it Wang et al.} [1994] of 3.73,

783: though it is lower than the value of 4.5, found by {\it Masuda and

784: Anderson} [1995] from the experimental data of {\it Utsumi et al.}

785: [1995].

786: Other theoretical determinations of $\delta_{T}$ are much higher,

787: however.  {\it Hama and Suito} [1998a] found a value of

788: 5.21, on the basis of calculations using the LAPW results of 

789: {\it Stixrude and Cohen} [1993] and

790: the Debye approximation for lattice vibration.

791: Using MD and semiempirical potentials \cite{matsui88}, 

792: {\it Belonoshko and Dubrovinsky} [1996] found $\delta_{T}$ = 5.80, and

793: {\it Patel et al.} [1996] found $\delta_{T}$ = 7.0 using

794: the same potentials, combining MD with lattice dynamics.

795: {\it Anderson} [1998], using Debye theory, calculated zero pressure

796: values ranging from 4.98 at 400 K

797: to $\sim$4 at 1000 K $\le$ $T$ $\le$ 1800 K, coming close to

798: our values at those temperatures.

799: 

800: The Gr\"uneisen parameter 

801: also increases as a function of $T$ and decreases as a

802: function of $P$ and approaches a value of 1 at

803: pressures of 130-140 GPa (\callout{Figure~\ref{fig6}}).

804: Our value of 1.33 at zero $P$ and 300 K matches 

805: well with {\it Wang et al.{\rm 's}} [1994] value of 1.3 and 

806: {\it Utsumi et al.{\rm 's}} [1995] and {\it Masuda and Anderson{\rm 's}}

807: [1995] value of 1.45.

808: {\it Stixrude et al.} [1992], using the experimental data of

809: {\it Mao et al.} [1991], found a higher value of

810: $\gamma$ = 1.96.  Values determined by

811: the inversion of multiple experimentally determined P-V-T data

812: sets match well also: 1.5 \cite{bina95}, 1.33 \cite{jackson96},

813: and 1.42 \cite{shim00}.

814: Two shock wave studies find a value of 1.5 \cite{duffy93, akins99}

815: on the basis of limited data sets: four for the former (with $q$ fixed equal

816: to 1) and two for the latter.

817: Values of $\gamma$ from theoretical studies range from very close

818: to ours, 1.279 \cite{hama98} and 1.44 \cite{hemley87}, 

819: to 1.97 \cite{belonoshko96, patel96}.

820: {\it Anderson{\rm 's}} [1998] zero pressure, 300 K value of 1.52 is

821: somewhat higher than ours, but at 400-1800 K his values of

822: $\sim$1.4 are very close to ours.

823: 

824: The volume dependence of $\gamma$, $q$, is 1.03 at 300 K

825: (\callout{Figure~\ref{fig7}}).

826: Our equation of state form, with $\Delta P_{\rm th}$ linear

827: in $T$ and independent of volume, constrains $q$ = 1 if

828: the isochoric heat capacity $C_{V} = 3 n R$, the

829: classical harmonic value.  As $T$ increases,

830: we find that

831: $q$ increases to a value of 1.07 at 3000 K

832: owing to anharmonicity.

833: Values from inverted experimental data sets range from

834: 1.0 \cite{bina95} to 2.0 \cite{shim00}.  

835: {\it Stixrude et al.} [1992] found a high $q$ value

836: of 2.5 to go along with their value for $\gamma$.

837: {\it Akins and Ahrens{\rm 's}}

838: [1999] value was even higher, $q = 4.0 \pm 1.0$ with $C_{V} = 5nR$,

839: but these are preliminary conclusions based on

840: limited data.  {\it Patel et al.} [1996] found a value of 3.0

841: at 0 GPa, decreasing to 1.7 at 100 GPa using a combination

842: of molecular and lattice dynamics.  {\it Anderson{\rm 's}} [1998]

843: Debye calculations gave values close to unity at $T$ $\ge$ 1000 K,

844: with $q$ decreasing slightly with increasing $T$, to 0.82 at 1800 K.

845: 

846: \section{Discussion and Conclusions}

847: Taking sets of experimental P-V-T data, we fit the high

848: temperature data \cite{fiquet00, saxena99} and available experimental

849: MgSiO$_{3}$ data to our thermal Universal EOS form (Table 2)

850: in order to have consistent bases for comparison.

851: The resulting fits have higher statistical uncertainties

852: but compare well with our results.

853: Volumes calculated along the lower mantle geothermal gradient 

854: of {\it Brown and Shankland} [1981], which has a starting $T$ of 1873 K

855: at 670 km (Figure 1), 

856: show that those calculated from

857: our EOS are 3 to 4 \AA$^{3}$ per unit cell larger, $\sim$2.5\%,

858: than those calculated from the inverted experimental data sets

859: (\callout{Figure~\ref{fig8}}).  

860: This corresponds

861: to density differences of $\sim$0.1 g cm$^{-3}$.

862: This is the opposite of the typical error of the

863: local density approximation

864: to density functional theory, on which our method is based.

865: The larger volume is due to the choice of ionic charges

866: and other model assumptions.

867: If we fix $V_{0}$ to the experimentally determined value

868: of 162.49 \AA$^{3}$ \cite{mao91} but otherwise use our

869: EOS parameters, our calculated volumes are very close,

870: within 1\%,

871: to the volumes derived from the inverted experimental data sets.

872: The comparisons also improve as pressure increases.

873: To compare our model with other zero temperature theoretically derived 

874: EOSs, we calculated volumes at 0 K from 0-140 GPa and fit an 

875: isothermal Universal EOS to them (\callout{Table~\ref{tbl3}}).

876: 

877: Compositional models of the lower mantle can be tested

878: against seismological models by examining densities

879: and seismic wave velocities as functions of depth.

880: Candidates include pyrolite \cite{ringwood75}

881: and chondritic or pure perovskite models.

882: Studies support both the former \cite{bina90,stacey96}

883: and latter \cite{butler78,stixrude92}, though uncertainties

884: in the thermodynamic parameters of the constituent minerals

885: make it difficult to resolve this question with certainty.

886: Indeed, several studies have been able to support both

887: models depending on whether high (pure perovskite) 

888: or low (pyrolite) values of $\alpha$ and $\delta_{T}$ 

889: or $\gamma$ and $q$ are used \cite{zhao94,anderson95,karki99}.

890: 

891: Calculating the densities of Mg-pv along a

892: geothermal gradient \cite{brown81} (\callout{Figure~\ref{fig9}}a), 

893: we see that they are $\sim$0.2-0.3 g cm$^{-3}$ ($\sim$4-5\%)

894: too low compared with those from the

895: Preliminary Reference Earth Model (PREM) \cite{dziewonski81}.

896: The calculated bulk sound velocities,

897: $V_{\varphi} = \sqrt{K_{S} / \rho}$, where 

898: $K_{S} = K_{T} ( 1 + \gamma \alpha \, T )$,

899: are, in turn, $\sim$0.4-0.6 km s$^{-1}$ ($\sim$4-7\%) too high compared to

900: the ak135-f seismological model \cite{kennett95,montagner96} (Figure 9b).

901: Using the experimentally determined value of $V_{0}$ =

902: 162.49 \AA$^{3}$ \cite{mao91} in place of our calculated

903: value does increase the densities somewhat, but not enough

904: to match the PREM values.  

905: 

906: As the perovskite found in

907: the lower mantle should be a solid solution of the Mg and 

908: Fe end-members, we added iron by adjusting density by

909: \cite{jeanloz83},

910: \begin{equation}

911:    \rho \, (X_{\rm {Fe}}) = \rho \, (0) (1 + 0.269 \, X_{\rm {Fe}})

912: \end{equation}

913: but we did not change the bulk moduli

914: \cite{mao91}.  Including

915: 10 mol \% Fe does cause densities to agree somewhat with the

916: PREM values (\callout{Figure~\ref{fig10}}a), 

917: but the bulk sound velocities are still much too

918: high (Figure 10b).  Consequently, we find that the lower

919: mantle cannot consist solely of (Mg,Fe)SiO$_{3}$ perovskite.

920: 

921: We also tested pyrolitic compositions consisting of mixtures

922: of perovskite and magnesiow\"ustite (mw).  Densities of 

923: mw were calculated using the thermodynamic data set of 

924: {\it Fei et al.} [1991], as well as $K_{T}$s for the Mg

925: end-member.  $K_{S}$s were then calculated using the

926: relation 

927: \begin{equation}

928:    \alpha (P,T) = \alpha (P_{0},T) \left[

929:    \frac {V (P,T)}{V (P_{0},T)} \right] ^{\delta_{T}} ,

930: \end{equation}

931: with $\delta_{T}$ and $\gamma$ from {\it Inbar and Cohen}

932: [1995].  Iron was accounted for in $K_{S}$ via the

933: relationship \cite{jeanloz83}

934: \begin{equation}

935:    K_{S} \, (X_{\rm {Fe}}) = K_{S} \, (0) + 17 \, X_{\rm {Fe}} .

936: \end{equation}

937: X$_{\rm Mg}$ = 0.93-0.96 for perovskite (pv) and 0.82-0.86

938: for mw \cite{katsura96, martinez97} were used.

939: Densities and bulk sound velocities were calculated

940: using Voigt-Reuss-Hill averaging for

941: two simplified pyrolite models (high [1] and low [2] Fe content)

942: consisting of 67 vol \% pv and 33 vol \% mw and were found to be in 

943: excellent agreement with the seismological models

944: (\callout{Figure~\ref{fig11}}).

945: The two pyrolite models have partitioning coefficients

946: $K^{\rm {pv-mw}}_{\rm {Fe-Mg}}$ of 0.34 and 0.26, where

947: \begin{equation}

948:    K^{\rm {pv-mw}}_{\rm {Fe-Mg}} = 

949:    \frac {X^{\rm {pv}}_{\rm {Fe}}/X^{\rm {pv}}_{\rm Mg}}

950:    {X^{\rm {mw}}_{\rm {Fe}}/X^{\rm {mw}}_{\rm Mg}}.

951: \end{equation}

952: Experimental evidence suggests that for such a bulk

953: composition, $K^{\rm {pv-mw}}_{\rm {Fe-Mg}}$ should increase in the

954: mantle from $\sim$0.20 (660 km) to $\sim$0.35 (1500 km),

955: with X$^{\rm {pv}}_{\rm Mg}$ decreasing and 

956: X$^{\rm {mw}}_{\rm Mg}$ increasing

957: with depth \cite{mao97}.  Compositional

958: models with high Fe content of pv and low

959: Fe content of mw (and vice versa)

960: fall inbetween the results of those shown in Figure 11.

961: Given the range of Fe-Mg partitioning between perovskite

962: and magnesiow\"ustite under the appropriate pressure and

963: temperature conditions, we find pyrolite to be the most likely 

964: compositional model for the lower mantle.

965: 

966: Given that other components (Al, Fe$^{3+}$) and phases

967: (CaSiO$_{3}$ perovskite) should be present in the lower mantle,

968: we do not expect exact agreement of a simplified pyrolite

969: model with any seismological model.

970: It is believed that, under lower mantle conditions,

971: Al$_{2}$O$_{3}$ is mainly incorporated into the Mg-pv

972: structure \cite{irifune94,wood00}.  

973: Generally, the effect of the incorporation of Al into Mg-pv

974: on its physical properties has been considered small

975: [e.g., {\it Weidner and Wang,} 1998].  However, Al, unlike Fe, 

976: causes significant distortion in the Mg-pv lattice \cite{oneill94},

977: which may affect the compressibility.

978: Recent experimental work by {\it Zhang and Weidner} [1999],

979: at pressure of up to 10 GPa and temperatures of up to 1073 K,

980: indicates that compared with MgSiO$_{3}$, Mg-pv with 

981: 5 mol \% Al has a smaller K$_{T}$ value

982: (232-236 GPa) and a $( \partial K_{T} / \partial T )_P$ value

983: more than double that in magnitude.  The values for $\alpha_{0}$,

984: $( \partial \alpha / \partial T )_P$, and $\delta_{T}$ are also larger.

985: Smaller bulk moduli and the higher density (0.2\% at 300 K and

986: 0 GPa) would cause seismic velocities to decrease.

987: In addition, the presence of Al tends to equalize the partitioning

988: of Fe into perovskite and magnesiow\"ustite and

989: may allow garnet to coexist with perovskite in the uppermost

990: $\sim$100 km of the lower mantle \cite{wood96,wood00}.

991: These partitioning experiments, however, were done at 

992: high $f_{\rm{O}_{2}}$,

993: so it is uncertain how applicable they are to the mantle.

994: The effect of the presence of Fe$^{3+}$ in Mg-pv on the

995: EOS and elasticity, while known for defect concentrations and

996: electrostatic charge balance \cite{mccammon97}, is unknown.

997: As for CaSiO$_{3}$ perovskite (Ca-pv), 

998: experimental data suggest that its

999: elastic properties are in excellent agreement with

1000: seismological models and thus would be invisible in the

1001: lower mantle \cite{wang96,hama98.2}.

1002: 

1003: Performing MD calculations over the range pressures

1004: and temperatures found in the Earth's mantle, we find

1005: a thermal EOS that is in excellent agreement with experimental

1006: results.  Thermodynamic parameters can be derived that agree well

1007: with experimentally determined values and that can be

1008: confidently interpolated to conditions found in the

1009: lower mantle.  Moreover, no solid-solid phase transitions

1010: or melting was found during the runs under lower mantle

1011: conditions, so orthorhombic MgSiO$_{3}$ perovskite should 

1012: be stable to the core-mantle boundary.

1013: Using these results, we find that pyrolite

1014: with $K^{\rm pv-mw}_{\rm Fe-Mg}$ = 0.26-0.34

1015: is the most likely compositional model for the lower mantle.

1016: 

1017: %%  EQUATIONS

1018: %

1019: %% ------------------------------------------------------ %%

1020: 

1021: %%%% Single-line equations are centered.  For multiline

1022:    % equations, break each line at a sign of operation

1023:    % (+, -, etc.) if possible, with the sign of operation

1024:    % on the new line.  Indent second and subsequent lines

1025:    % to align with the first character following the equal

1026:    % sign on the first line.  If the left side of the

1027:    % equation is longer than one-third of the line,

1028:    % subsequent lines should be indented the same amount

1029:    % as a text paragraph indent (0.38cm, 0.15").  Set the

1030:    % equation number flush right on the last line of the

1031:    % equation with at least 4 mm (5/16 inch, 1 pica) of

1032:    % space between the equation and the equation number;

1033:    % if there is not enough room, set the number flush

1034:    % right on a separate line immediately following the

1035:    % equation.  If you have two equations with the same

1036:    % equation number, set them over three lines of type,

1037:    % with the equation number appearing flush right and

1038:    % in parentheses on the second line.

1039: 

1040: %%%% To create multi-line equations, use the

1041:    % \begin{eqnarray} and \end{eqnarray} environment

1042:    % to break equations into two or more lines, as

1043:    % demonstrated below.  Use an \hspace{} command to

1044:    % insert horizontal space into your LaTeX document.

1045:    % Place an appropriate unit of measure between the

1046:    % curly braces; you may have to experiment to achieve

1047:    % the correct amount of space.  LaTeX accepts several

1048:    % units of measurement (cm, in, etc.); you may use

1049:    % whichever unit you prefer.

1050: 

1051: %\begin{eqnarray}

1052: %   x_{1} & = & (x - x_{0}) \cos \Theta +

1053: %        (y - y_{0}) \sin \Theta \\

1054: %   y_{1} & = & -(x - x_{0}) \sin \Theta +

1055: %        (y - y_{0}) \cos \Theta.

1056: %\end{eqnarray}

1057: 

1058: %%%% You may change equation numbering by resetting

1059:    % the equation counter or by explicitly numbering

1060:    % an equation.

1061:    %

1062:    % To explicitly number an equation, type \eqnum{}

1063:    % (with the desired number between the brackets)

1064:    % after the \begin{equation} or \begin{eqnarray}

1065:    % command.  The \eqnum{} command will affect only

1066:    % the equation it appears with; LaTeX will number

1067:    % any equations appearing later in the manuscript

1068:    % according to the equation counter.

1069:    %

1070:    % To reset the equation counter, place the following

1071:    % command in front of your equation(s).

1072: 

1073: \setcounter{equation}{0}

1074: 

1075:    % Set the equation counter to 0 if the next

1076:    % number needed is 1, or set it to 7 if the

1077:    % next number needed is 8, etc.

1078:    %

1079:    % The \setcounter{equation} command does affect

1080:    % equations appearing later in the manuscript.

1081: 

1082: %%%% If you have a multi-line equation that needs only

1083:    % one equation number, use a \nonumber command in

1084:    % front of the double backslashes (\\) as shown in

1085:    % equations above and below.

1086: 

1087: %%%% To add letters after equation numbers, place your

1088:    % equation or eqnarray within a \begin{mathletters}

1089:    % and \end{mathletters} environment.  This environment

1090:    % can enclose several equations:

1091: 

1092: %\begin{mathletters}

1093: %   \begin{eqnarray}

1094: %      \gamma^\mu & = & \left(\begin{array}{cc} 0 &

1095: %      \sigma^\mu_+ \\

1096: %      \sigma^\mu_- & 0 \end{array} \right),

1097: %      \;\gamma = \left( \begin{array}{cc}

1098: %      \! \! -1 & \! \! 0 \\ \! \! 0 & \! \! 1

1099: %      \end{array} \right), \\ & & \nonumber \\

1100: %      \sigma^\mu_{\pm} & = & ({\bf 1} ,\pm \sigma),

1101: %   \end{eqnarray}

1102: %giving

1103: %   \begin{eqnarray}

1104: %      \not\!a= \left(\begin{array}{cc}0 &

1105: %      (\not\!a)_+\\(\not\!a)_- &

1106: %      0\end{array}\right),

1107: %      \;(\not\!a)_\pm=a_\mu\sigma^\mu_\pm\;,

1108: %   \end{eqnarray}

1109: %\end{mathletters}

1110: 

1111: %% ------------------------------------------------------ %%

1112:    %

1113:    % In the next sections we have a reference to one of

1114:    % the tables (tables appear at the end of the LaTeX

1115:    % document).  There is also some additional math-related

1116:    % markup.  In the second paragraph, note the use of

1117:    % in-text math material between the dollar signs ($$)

1118:    % including a couple of the miscellaneous symbol

1119:    % commands defined in AGU's style file package.

1120:    %

1121:    % Note that keyboard math commands (such as greater

1122:    % than or less than) must appear within a math or

1123:    % equation environment or they will not print properly

1124:    % (i.e., $<$ or $>$).  All other math symbols (such as

1125:    % similar) must be created with appropriate LaTeX symbol

1126:    % commands (i.e., $\sim$ instead of ~) or they will not

1127:    % print properly.  See the aguguide.tex document for a

1128:    % complete list of keyboard and LaTeX math symbols.

1129:    %

1130: %% ------------------------------------------------------ %%

1131: 

1132: %% ------------------------------------------------------ %%

1133: %

1134: %%  CALLOUTS

1135: %

1136: %% ------------------------------------------------------ %%

1137: 

1138: %\callout{Table~\ref{tbl-1}}

1139: %and

1140: %\callout{Table~2}.

1141:    %

1142:    % The first time a table, figure, or plate is

1143:    % mentioned in a manuscript, AGU asks that it

1144:    % be called out (i.e., identified in the right

1145:    % margin of the camera-ready copy).

1146:    %

1147:    % Enclose the first occurrence of a table, figure,

1148:    % or plate within a \callout{} command.  Text within

1149:    % this command will appear in your manuscript text

1150:    % and will also appear in a box in the right margin

1151:    % of your manuscript.  Make sure all callouts appear

1152:    % in numerical sequence (i.e., Table 1, Table 2, etc.)

1153:    %

1154:    % You may use \ref{} commands within callout commands,

1155:    % as shown above.

1156:    %

1157: 

1158: %% ------------------------------------------------------ %%

1159: 

1160: %% ------------------------------------------------------ %%

1161: %

1162: %%  NOTATION TABLES (if any)

1163: %

1164: %% ------------------------------------------------------ %%

1165:    %

1166:    % If your manuscript contains notation tables, use a

1167:    % Level 1 head for the word "Notation" and use a

1168:    % tabular environment without any rules for the

1169:    % notation information, as shown below.

1170:    %

1171:    % If you numbered your sections, place a

1172:    % \setcounter{secnumdepth}{0} in front of

1173:    % the Notation section head, so it will

1174:    % appear flush left and without numbers.

1175:    %

1176:    % Set the symbol columns flush right and the

1177:    % definition column flush left.  You will probably

1178:    % need to adjust the {rp{17.5pc} spacing to create

1179:    % appropriate widths for your notation columns.

1180: 

1181: %% ------------------------------------------------------ %%

1182: %

1183: %%  ACKNOWLEDGMENTS (optional)

1184: %

1185: %% ------------------------------------------------------ %%

1186: 

1187: \acknowledgments

1188: We thank Iris Inbar, Russell Hemley, Stephen Gramsch,

1189: Joe Akins, Mark Woods, Sheila Coleman, and Lars Stixrude for 

1190: helpful discussions and comments.

1191: We acknowledge the support of the National Science

1192: Foundation through grant EAR-9870328.

1193: The computations were performed on the Cray SV1 at the Geophysical

1194: Laboratory, with the support of NSF grant EAR-9975753 and the

1195: W. M. Keck Foundation.

1196: 

1197:    % Type your acknowledgment text (if any)

1198:    % immediately after the \acknowledgments

1199:    % command.  Acknowledgments may be only

1200:    % one paragraph in length.

1201:    %

1202:    % Be sure to spell "\acknowledgments"

1203:    % exactly as it appears above; if you spell

1204:    % it differently then LaTeX will not recognize

1205:    % the command and it will not work.  There is

1206:    % no "end acknowledgments" command.

1207:    %

1208:    % If your manuscript contains only one

1209:    % acknowledgment, use an \acknowledgment

1210:    % command to produce a singular

1211:    % "Acknowledgment" subhead.

1212:    %

1213:    % If your manuscript contains Editor's

1214:    % acknowledgments, set them in a new

1215:    % paragraph one line below the regular

1216:    % acknowledgments, as shown.

1217: 

1218: %% ------------------------------------------------------ %%

1219: 

1220: % \newpage

1221: 

1222:    % If desired, you may place a \newpage command after

1223:    % the acknowledgments section; this will force the

1224:    % references section to begin on a new page.

1225: 

1226: 

1227: %% ------------------------------------------------------ %%

1228: %

1229: %%  REFERENCES

1230: %

1231: %% ------------------------------------------------------ %%

1232: 

1233:    % Use "\cite" commands to call out citations,

1234:    % and then use LaTeX's "thebibliography" environment

1235:    % for the reference list.  Please note that the

1236:    % \begin{thebibliography}{} command is followed by a

1237:    % null argument {} (see the example below).

1238:    %

1239:    % Each reference has a \bibitem command to define

1240:    % the citation format and the symbolic tag, as well

1241:    % as a \reference command which sets up formatting

1242:    % parameters for the reference list itself.  Authors

1243:    % choose what markup appears in the text, but it must

1244:    % match the tags used with the corresponding \bibitem

1245:    % command, for example, \cite{can95}

1246:    %

1247:    % Try to use \cite for all references in your document.

1248:    % Use \markcite commands for the format such as

1249:    % {\it Kurtz and King} [1994].

1250:    % HOWEVER, THIS FORMAT SHOULD BE USED AS LESS AS POSSIBLE.

1251:    %

1252: 

1253: \begin{thebibliography}{}

1254: 

1255: \bibitem[{\it Akins and Ahrens,} 1999]{akins99}

1256: \reference % (abstract)

1257: Akins, J.~A., and T.~J.\/ Ahrens, Shock wave equation of

1258: state measurements on MgSiO$_{3}$ perovskite at lower

1259: mantle pressures indicate $\gamma_{0}$ = 1.5 and q = 

1260: $\partial {\:\ln \gamma} / \partial {\:\ln V} >$  3,

1261: \eos {\it 80} (46), Fall Meet. Suppl., F757, 1999. 

1262: 

1263: \bibitem[{\it Anderson,} 1980, 1984]{anderson80}

1264: \reference % (paper)

1265: Anderson, O.~L., An experimental high-temperature thermal

1266: equation of state bypassing the Gr\"uneisen parameter,

1267: \pepi {\it 1,} 173-183, 1980.

1268: 

1269: \bibitem[{\it Anderson,} 1984]{anderson84}

1270: \reference % (paper)

1271: Anderson, O.~L., A universal thermal equation-of-state,

1272: {\it J. Geodyn., 1,} 185-214, 1984.

1273: 

1274: \bibitem[{\it Anderson,} 1998]{anderson98}

1275: \reference % (paper)

1276: Anderson, O.~L., Thermoelastic properties of MgSiO$_{3}$

1277: perovskite using the Debye approach, {\it Am. Mineral.,

1278: 83,} 23-35, 1998.

1279: 

1280: \bibitem[{\it Anderson and Suzuki,} 1983]{anderson83}

1281: \reference % (paper)

1282: Anderson, O.~L., and I.\/ Suzuki, Anharmonicity of three

1283: minerals at high temperature: Forsterite, fayalite, and

1284: periclase, \jgr {\it 88,} 3549-3356, 1983.

1285: 

1286: \bibitem[{\it Anderson et al.,} 1995]{anderson95}

1287: \reference % (paper)

1288: Anderson, O.~L., K.\/ Masuda, and D.\/ Guo, Pure silicate

1289: perovskite and the PREM lower mantle model: a thermodynamic

1290: analysis, \pepi {\it 89,} 35-49, 1995.

1291: 

1292: \bibitem[{\it Belonoshko,} 1994]{belonoshko94}

1293: \reference % (paper)

1294: Belonoshko, A.~B., Molecular dynamics of MgSiO$_{3}$ perovskite

1295: at high pressures: Equation of state, structure, and melting 

1296: transition, \gca {\it 58,} 4039-4047, 1994.

1297: 

1298: \bibitem[{\it Belonoshko and Dubrovinsky,} 1996]{belonoshko96}

1299: \reference % (paper)

1300: Belonoshko, A.~B., and L.~S.\/ Dubrovinsky, Equations of state

1301: of MgSiO$_{3}$-perovskite and MgO (periclase) from computer

1302: simulations, \pepi {\it 98,} 47-54, 1996.

1303: 

1304: \bibitem[{\it Bina,} 1995]{bina95}

1305: \reference % (paper)

1306: Bina, C.~R., Confidence limits for silicate perovskite equations

1307: of state, {\it Phys. Chem. Miner., 22,} 375-382, 1995.

1308: 

1309: \bibitem[{\it Bina,} 1998]{bina98}

1310: \reference % (paper)

1311: Bina, C.~R., Lower mantle mineralogy and the geophysical

1312: perspective, {\it Rev. Mineral., 37,} 205-239, 1998.

1313: 

1314: \bibitem[{\it Bina and Silver,} 1990]{bina90}

1315: \reference % (paper)

1316: Bina, C.~R., and P.~G.\/ Silver, Constraints on lower mantle

1317: composition and temperature from density and bulk sound

1318: velocity profiles, \grl {\it 17,} 1153-1156, 1990. 

1319: 

1320: \bibitem[{\it Birch,} 1952]{birch52}

1321: \reference % (paper)

1322: Birch, F., Elasticity and constitution of the Earth's

1323: interior, \jgr {\it 57,} 227-286, 1952.

1324: 

1325: \bibitem[{\it Brown and Shankland,} 1981]{brown81}

1326: \reference % (paper)

1327: Brown, J.~M., and T.~J.\/ Shankland, Thermodynamic parameters in the

1328: Earth as determined from seismic profiles, \gjras {\it 66,}

1329: 579-596, 1981.

1330: 

1331: \bibitem[{\it Butler and Anderson,} 1978]{butler78}

1332: \reference % (paper)

1333: Butler, R., and D.~L.\/ Anderson, Equations of state fits to

1334: the lower mantle and outer core, \pepi {\it 17,} 147-162, 1978.

1335: 

1336: \bibitem[{\it Chao et al.,} 1961]{chao61}

1337: \reference % (paper)

1338: Chao, E.~C.~T., H.~T.\/ Evans, B.~J.\/ Skinner, and

1339: C.\/ Milton, Neighborite, NaMgF$_{3}$, a new mineral from the

1340: Green River Formation, South Ouray, Utah, {\it Am. Mineral.,

1341: 46,} 379-393, 1961.

1342: 

1343: \bibitem[{\it Cohen,} 1987]{cohen87}

1344: \reference % (paper)

1345: Cohen, R.~E., Elasticity and equation of state of MgSiO$_{3}$

1346: perovskite, \grl {\it 14,} 1053-1056, 1987.

1347: 

1348: \bibitem[{\it Cohen et al.,} 1989]{cohen89}

1349: \reference % (article in book)

1350: Cohen, R.~E., L.~L.\/ Boyer, M.~J.\/ Mehl, W.~E.\/ Pickett, and

1351: H.\/ Krakauer, Electronic structure and total energy calculations

1352: for oxide perovskites and superconductors, in {\it Perovskite:

1353: A Structure of Great Interest to Geophysics and Materials Science,

1354: Geophys. Monogr. Ser.}, 45,

1355: edited by A.\/ Navrotsky and D.~J.\/ Weidner, pp. 55-66,

1356: AGU, Washington, D. C., 1989.

1357: 

1358: \bibitem[{\it Cohen et al.,} 2000]{cohen00}

1359: \reference % (paper)

1360: Cohen, R.~E., O.\/ G\"ulseren, and R.~J.\/ Hemley, Accuracy

1361: of equation-of-state formulations, {\it Am. Mineral., 85,}

1362: 338-344, 2000.

1363: 

1364: \bibitem[{\it Dubrovinsky et al.} 1998]{dubrovinsky98}

1365: \reference % (abstract)

1366: Dubrovinsky, L., S.~K.\/ Saxena, R.\/ Ahuja, and B.\/

1367: Johansson, Theoretical study of the stability of 

1368: MgSiO$_{3}$-perovskite in the deep mantle, \grl {\it 25,}

1369: 4253-4256, 1998.

1370: 

1371: \bibitem[{\it Duffy and Ahrens,} 1993]{duffy93}

1372: \reference % (abstract)

1373: Duffy, T.~S., and T.~J.\/ Ahrens, Thermal properties of

1374: perovskite above 100 GPa from shock and static compression

1375: data, \eos {\it 74} (43), Fall Meet. Suppl., 571, 1993.

1376: 

1377: \bibitem[{\it Dziewonski and Anderson,} 1981]{dziewonski81}

1378: \reference % (paper)

1379: Dziewonski, A. M., and D. L.\/ Anderson, Preliminary reference

1380: Earth model, \pepi {\it 25,} 297-356, 1981.

1381: 

1382: \bibitem[{\it Fei et al.,} 1991]{fei91}

1383: \reference % (paper)

1384: Fei, Y., H.~K.\/ Mao, and B.~O.\/ Mysen, Experimental 

1385: determination of element partitioning and calculation of

1386: phase relations in the MgO-FeO-SiO$_{2}$ system at

1387: high pressure and high temperature, \jgr {\it 96,}

1388: 2157-2169, 1991.

1389: 

1390: \bibitem[{\it Fiquet et al.,} 1998]{fiquet98}

1391: \reference % (paper)

1392: Fiquet, G., D.\/ Andrault, A.\/ Dewaele, T.\/ Charpin,

1393: M.\/ Kunz, and D.\/ Ha\"usermann, P-V-T equation of state

1394: of MgSiO$_{3}$ perovskite, \pepi {\it 105,} 21-31, 1998.

1395: 

1396: \bibitem[{\it Fiquet et al.,} 1998, 2000]{fiquet00}

1397: \reference % (paper)

1398: Fiquet, G., A.\/ Dewaele, D.\/ Andrault, M.\/ Kunz, and T.\/ Le Bihan,

1399: Thermoelastic properties and crystal structure of MgSiO$_{3}$ 

1400: perovskite at lower mantle pressure and temperature conditions,

1401: \grl {\it 27,} 21-24, 2000.

1402: 

1403: \bibitem[{\it Funamori et al.,} 1996]{funamori96}

1404: \reference % (paper)

1405: Funamori, N., T.\/ Yagi, W.\/ Utsumi, T.\/ Kondo, and

1406: T.\/ Uchida, Thermoelastic properties of MgSiO$_{3}$

1407: perovskite determined by in situ X ray observations up

1408: to 30 GPa and 2000 K, \jgr {\it 101,} 8257-8269, 1996.

1409: 

1410: \bibitem[{\it Gear,} 1971]{gear71}

1411: \reference % (book)

1412: Gear, C.~W., {\it Numerical Initial Value Problems

1413: in Ordinary Differential Equations,} 253 pp., 

1414: Prentice-Hall, Old Tappan, N. J., 1971.

1415: 

1416: \bibitem[{\it Gordon and Kim,} 1972]{gordon72}

1417: \reference % (paper)

1418: Gordon, R.~G., and Y.~S.\/ Kim, Theory for the forces between

1419: closed-shell atoms and molecules, {\it J. Chem. Phys., 56,}

1420: 3122-3133, 1972.

1421: 

1422: \bibitem[{\it Hama and Suito,} 1998a]{hama98}

1423: \reference % (paper)

1424: Hama, J., and K.\/ Suito, Equation of state of MgSiO$_{3}$

1425: perovskite and its thermoelastic properties under lower 

1426: mantle conditions, \jgr {\it 103,} 7443-7462, 1998a.

1427: 

1428: \bibitem[{\it Hama and Suito,} 1998b]{hama98.2}

1429: \reference % (paper)

1430: Hama, J., and K.\/ Suito, High-temperature equation of 

1431: state of CaSiO$_{3}$ perovskite and its implications for

1432: the lower mantle conditions, \pepi {\it 105,} 33-46, 1998b.

1433: 

1434: \bibitem[{\it Hedin and Lundqvist,} 1971]{hedin71}

1435: \reference % (paper)

1436: Hedin, L., and B.~I.\/ Lundqvist, Explicit local-exchange

1437: correlation potentials, {\it J. Phys. C. Solid State Phys., 4,} 

1438: 2064-2083, 1971.

1439: 

1440: \bibitem[{\it Heinz and Jeanloz,} 1987]{heinz87}

1441: \reference % (paper)

1442: Heinz, D.~L., and R.\/ Jeanloz, Measurement of the melting

1443: curve of Mg$_{0.9}$Fe$_{0.1}$SiO$_{3}$ at lower mantle

1444: conditions and its geophysical implications, {\jgr}

1445: {\it 92,} 11,437-11,444, 1987.

1446: 

1447: \bibitem[{\it Hemley and Cohen,} 1996]{hemley96}

1448: \reference % (paper)

1449: Hemley, R.~J., and R.~E.\/ Cohen, Structure and bonding in

1450: the deep mantle and core, {\it Philos. Trans. R. Soc. London, Ser. A,

1451: 354,} 1461-1479, 1996.

1452: 

1453: \bibitem[{\it Hemley et al.,} 1987]{hemley87}

1454: \reference % (paper)

1455: Hemley, R.~J., M.~D.\/ Jackson, and G.~G.\/ Gordon, Theoretical

1456: study of the structure, lattice dynamics, and equations of

1457: state of perovskite-type MgSiO$_{3}$ and CaSiO$_{3}$,

1458: {\it Phys. Chem. Miner., 14,} 2-12, 1987.

1459: 

1460: \bibitem[{\it Inbar and Cohen,} 1995]{inbar95}

1461: \reference % (paper)

1462: Inbar, I., and R.~E.\/ Cohen, High pressure effects on thermal

1463: properties of MgO, \grl {\it 22,} 1533-1536, 1995.

1464: 

1465: \bibitem[{\it Irifune,} 1994]{irifune94}

1466: \reference % (paper)

1467: Irifune, T., Absence of an aluminous phase in the upper part

1468: of the Earth's lower mantle, {\it Nature, 370,} 131-133, 1994.

1469: 

1470: \bibitem[{\it Jackson and Rigden,} 1996]{jackson96}

1471: \reference % (paper)

1472: Jackson, I., and S.~M.\/ Rigden, Analysis of P-V-T data: Constraints

1473: on the thermoelastic properties of high-pressure minerals,

1474: \pepi {\it 96,} 85-112, 1996.

1475: 

1476: \bibitem[{\it Jeanloz,} 1988]{jeanloz88}

1477: \reference % (paper)

1478: Jeanloz, R., Universal equation of state, \prb {\it 38,} 805-807, 1988.

1479: 

1480: \bibitem[{\it Jeanloz and Thompson,} 1983]{jeanloz83}

1481: \reference % (paper)

1482: Jeanloz, R., and A.~B.\/ Thompson, Phase transitions and

1483: mantle discontinuities, {\it Rev. Geophys., 21,} 51-74, 1983.

1484: 

1485: \bibitem[{\it Karki and Stixrude,} 1999]{karki99}

1486: \reference % (paper)

1487: Karki, B.~B., and L.\/ Stixrude, Seismic velocities of major

1488: silicate and oxide phases of the lower mantle, \jgr

1489: {\it 104,} 13,025-13,033, 1999.

1490: 

1491: \bibitem[{\it Karki et al.,} 1997]{karki97}

1492: \reference % (paper)

1493: Karki, B.~B., L.\/ Stixrude, S.~J.\/ Clark, M.~C.\/ Warren,

1494: G.~J.\/ Ackland, and J.\/ Crain, Elastic properties of

1495: orthorhombic MgSiO$_{3}$ perovskite at lower mantle pressures,

1496: {\it Am. Mineral., 82,} 635-638, 1997.

1497: 

1498: \bibitem[{\it Kato et al.,} 1995]{kato95}

1499: \reference % (paper)

1500: Kato, T., E.\/ Ohtani, H.\/ Morishima, D.\/ Yamazaki,

1501: A.\/ Suzuki, M.\/ Suto, T.\/ Kubo, T.\/ Kikegawa, and

1502: O.\/ Shimomura, In situ X ray observation of high-pressure

1503: phase transitions of MgSiO$_{3}$ and thermal expansion

1504: of MgSiO$_{3}$ perovskite at 25 GPa by double-stage

1505: multianvil system, \jgr {\it 100,} 20,475-20,481, 1995.

1506: 

1507: \bibitem[{\it Katsura and Ito,} 1996]{katsura96}

1508: \reference % (paper)

1509: Katsura, T., and E.\/ Ito, Determination of Fe-Mg partitioning

1510: between perovskite and magnesiow\"ustite, \grl {\it 23,}

1511: 2005-2008, 1996.

1512: 

1513: \bibitem[{\it Kennett et al.,} 1995]{kennett95}

1514: \reference % (paper)

1515: Kennett, B.~L.~N., E.~R.\/ Engdahl, and R.\/ Buland,

1516: Constraints on seismic velocities in the Earth from 

1517: traveltimes, \gji {\it 122,} 108-124, 1995.

1518: 

1519: \bibitem[{\it Knittle and Jeanloz} 1987]{knittle87}

1520: \reference % (paper)

1521: Knittle, E., and R.\/ Jeanloz, Synthesis and equation of state of

1522: (Mg,Fe)SiO${3}$ perovskite to over 100 gigapascals,

1523: {\it Science, 235,} 668-670, 1987.

1524: 

1525: \bibitem[{\it Knittle and Jeanloz} 1989]{knittle89}

1526: \reference % (paper)

1527: Knittle, E., and R.\/ Jeanloz, Melting curve of 

1528: (Mg,Fe)SiO${3}$ perovskite to 96 GPa: Evidence 

1529: for a structural transition in lower mantle melts,

1530: \grl {\it 16,} 421-424, 1989.

1531: 

1532: \bibitem[{\it Kudoh et al.,} 1987]{kudoh87}

1533: \reference % (paper)

1534: Kudoh, Y., E.\/ Ito, and H.\/ Takeda, Effect of pressure on

1535: the crystal structure of perovskite-type MgSiO$_{3}$,

1536: {\it Phys. Chem. Miner., 14,} 350-354, 1987.

1537: 

1538: \bibitem[{\it Lewis and L\'epine,} 1989]{lewis89}

1539: \reference % (paper)

1540: Lewis, L.~J., and Y.\/ L\'epine, Structural transitions in

1541: fluorine-based perovskites: A molecular dynamics study of

1542: KMnF$_{3}$, {\it Phys. Rev. B Condens. Matter, 40,} 3319-3322, 1989.

1543: 

1544: \bibitem[{\it Mao et al.,} 1991]{mao91}

1545: \reference % (paper)

1546: Mao, H.~K., R.~J.\/ Hemley, Y.\/ Fei, J.~F.\/ Shu, L.~C.\/ Chen,

1547: A.~P.\/ Jephcoat, and Y.\/ Wu, Effect of pressure, temperature,

1548: and composition on lattice parameters and density of 

1549: (Fe,Mg)SiO$_{3}$-perovskites to 30 GPa, \jgr {\it 96,}

1550: 8069-8079, 1991.

1551: 

1552: \bibitem[{\it Mao et al.,} 1997]{mao97}

1553: \reference % (paper)

1554: Mao, H.~K., G.\/ Shen, and R.~J.\/ Hemley, Multivariable

1555: dependence of Fe-Mg partitioning in the lower mantle,

1556: {\it Science, 278,} 2098-2100, 1997.

1557: 

1558: \bibitem[{\it Martinez et al.,} 1997]{martinez97}

1559: \reference % (paper)

1560: Martinez, I., Y.\/ Wang, F.\/ Guyot, R.~C.\/ Liebermann, and 

1561: J.~C.\/ Doukhan, Microstructures and iron partitioning in 

1562: (Mg,Fe)SiO$_{3}$ perovskite - (Mg,Fe)O magnesiow\"ustite assemblages: 

1563: An analytical transmission electron microscopy study, \jgr

1564: {\it 102,} 5265-5280, 1997.

1565: 

1566: \bibitem[{\it Martyna et al.,} 1994]{martyna94}

1567: \reference % (paper)

1568: Martyna, G.~J., D.~J.\/ Tobias, and M.~L.\/ Klein,

1569: Constant pressure molecular dynamics algorithms,

1570: {\it J. Chem. Phys., 101,} 4177-4189, 1994.

1571: 

1572: \bibitem[{\it Masuda and Anderson,} 1995]{masuda95}

1573: \reference % (paper)

1574: Masuda, K., and O.~L.\/ Anderson, The isentropic density profile

1575: of silicate perovskite computed by thermal pressure, \grl

1576: {\it 22,} 2211-2214, 1995.

1577: 

1578: \bibitem[{\it Matsui,} 1988]{matsui88}

1579: \reference % (paper)

1580: Matsui, M., Molecular dynamics study of MgSiO$_{3}$ perovskite,

1581: {\it Phys. Chem. Miner., 16,} 234-238, 1988.

1582: 

1583: \bibitem[{\it Matsui and Price,} 1991]{matsui91}

1584: \reference % (paper)

1585: Matsui, M., and G.~D.\/ Price, Simulation of the pre-melting

1586: behaviour of MgSiO$_{3}$ perovskite at high pressures and

1587: temperatures, {\it Nature, 351,} 735-737, 1991.

1588: 

1589: \bibitem[{\it McCammon,} 1997, 1998]{mccammon97}

1590: \reference % (paper)

1591: McCammon, C., Perovskite as a possible sink for ferric iron

1592: in the lower mantle, {\it Nature, 387,} 694-696, 1997.

1593: 

1594: \bibitem[{\it McCammon,} 1998]{mccammon98}

1595: \reference % (paper)

1596: McCammon, C., The crystal chemistry of ferric iron in

1597: Fe$_{0.05}$Mg$_{0.95}$SiO$_{3}$ perovskite as determined by 

1598: M\"ossbauer spectroscopy in the temperature range 80-293 K,

1599: {\it Phys. Chem. Miner., 25,} 292-300, 1998.

1600: 

1601: \bibitem[{\it Meade et al.,} 1995]{meade95}

1602: \reference % (paper)

1603: Meade, C., H.~K.\/ Mao, and J.\/ Hu, High-temperature phase

1604: transition and dissociation of (Mg,Fe)SiO$_{3}$ perovskite

1605: at lower mantle pressures, {\it Science, 268,} 1743-1745, 1995.

1606: 

1607: \bibitem[{\it Montagner and Kennett,} 1996]{montagner96}

1608: \reference % (paper)

1609: Montagner, J.-P., and B.~L.~N.\/ Kennett, How to reconcile

1610: body-wave and normal-mode reference earth models, \gji

1611: {\it 125,} 229-248, 1996.

1612: 

1613: \bibitem[{\it Nos\'e and Klein,} 1989]{nose89}

1614: \reference % (paper)

1615: Nos\'e, S., and M.~L.\/ Klein, Structure and dynamics of

1616: the fluorperovskite, RbCaF$_{3}$, {\it J. Chem. Phys., 90,}

1617: 5005-5010, 1989.

1618: 

1619: \bibitem[{\it O'Neill and Jeanloz,} 1994]{oneill94}

1620: \reference % (paper)

1621: O'Neill, B., and R.\/ Jeanloz, MgSiO$_{3}$-FeSiO$_{3}$-Al$_{2}$O$_{3}$

1622: in the Earth's lower mantle: Perovskite and garnet at 1200 km depth,

1623: \jgr {\it 99,} 19,901-19,915, 1994.

1624: 

1625: \bibitem[{\it Patel et al.,} 1996]{patel96}

1626: \reference % (paper)

1627: Patel, A., G.~D.\/ Price, M.\/ Matsui, J.~P.\/ Brodholt, and

1628: R.~J.\/ Howarth, A computer simulation approach to the high 

1629: pressure thermoelasticity of MgSiO$_{3}$ perovskite, \pepi

1630: {\it 98,} 55-63, 1996.

1631: 

1632: \bibitem[{\it Poirier,} 1989]{poirier89}

1633: \reference % (paper)

1634: Poirier, J.~P., Lindemann law and the melting temperature

1635: of perovskites, \pepi {\it 54,} 364-369, 1989.

1636: 

1637: \bibitem[{\it Ringwood,} 1975]{ringwood75}

1638: \reference % (book)

1639: Ringwood, A.~E., {\it Composition and Petrology of the

1640: Earth's Mantle,} 618 pp., McGraw-Hill, New York, 1975.

1641: 

1642: \bibitem[{\it Ross and Hazen,} 1990]{ross90}

1643: \reference % (paper)

1644: Ross, N.~L., and R.~M.\/ Hazen, High-pressure crystal chemistry

1645: of MgSiO$_{3}$ perovskite, {\it Phys. Chem. Miner., 17,} 228-237, 1990. 

1646: 

1647: \bibitem[{\it Saxena et al.,} 1996]{saxena96}

1648: \reference % (paper)

1649: Saxena, S.~K., L.~S.\/ Dubrovinsky, P.\/ Lazor, Y.\/ Cerenius,

1650: P.\/ H\"aggkvist, M.\/ Hanfland, and J.\/ Hu, Stability of

1651: perovskite (MgSiO$_{3}$) in the Earth's mantle, {\it Science,

1652: 274,} 1357-1359, 1996.

1653: 

1654: \bibitem[{\it Saxena et al.,} 1998]{saxena98}

1655: \reference % (paper)

1656: Saxena, S.~K., L.~S.\/ Dubrovinsky, P.\/ Lazor,, and J.\/ Hu,

1657: In situ X-ray study of perovskite (MgSiO$_{3}$): Phase

1658: transition and dissociation at mantle conditions,

1659: {\it Eur. J. Mineral., 10,} 1275-1281, 1998.

1660: 

1661: \bibitem[{\it Saxena et al.,} 1999]{saxena99}

1662: \reference % (paper)

1663: Saxena, S.~K., L.~S.\/ Dubrovinsky, F.\/ Tutti, and T.\/ Le Bihan,

1664: Equation of state of MgSiO$_{3}$ with the perovskite structure

1665: based on experimental measurement, {\it Am. Mineral., 84,}

1666: 226-232, 1999.

1667: 

1668: \bibitem[{\it Serghiou et al.,} 1998]{serghiou98}

1669: \reference % (paper)

1670: Serghiou, G., A.\/ Zerr, and R.\/ Boehler, 

1671: (Mg,Fe)SiO$_{3}$-perovskite stability under lower mantle

1672: conditions, {\it Science, 280,} 2093-2095, 1998.

1673: 

1674: \bibitem[{\it Shim and Duffy,} 2000]{shim00}

1675: \reference % (paper)

1676: Shim, S.-H., and T.~S.\/ Duffy, Constraints on the P-V-T equation

1677: of state of MgSiO$_{3}$ perovskite, {\it Am. Mineral., 85,}

1678: 354-363, 2000.

1679: 

1680: \bibitem[{\it Stacey,} 1996]{stacey96}

1681: \reference % (paper)

1682: Stacey, F.~D., Thermoelasticity of (Mg,Fe)SiO$_{3}$ perovskite

1683: and a comparison with the lower mantle, \pepi {\it 98,}

1684: 65-77, 1996.

1685: 

1686: \bibitem[{\it Stixrude and Bukowinski,} 1992]{stixrude92.2}

1687: \reference % (paper)

1688: Stixrude, L., and M.~S.~T.\/ Bukowinski, Stability of

1689: (Mg,Fe)SiO$_{3}$ perovskite and the structure of the

1690: lowermost mantle, \grl {\it 19,} 1057-1060, 1992.

1691: 

1692: \bibitem[{\it Stixrude and Cohen,} 1993]{stixrude93}

1693: \reference % (paper)

1694: Stixrude, L., and R.~E.\/ Cohen, Stability of orthorhombic

1695: MgSiO$_{3}$ perovskite in the Earth's lower mantle,

1696: {\it Nature, 364,} 613-616, 1993.

1697: 

1698: \bibitem[{\it Stixrude et al.,} 1992]{stixrude92}

1699: \reference % (paper)

1700: Stixrude, L., R.~J.\/ Hemley, Y.\/ Fei, and H.~K.\/ Mao,

1701: Thermoelasticity of silicate perovskite and magnesiow\"ustite

1702: and stratification of the Earth's mantle, {\it Science, 257,}

1703: 1099-1101, 1992.

1704: 

1705: \bibitem[{\it Utsumi et al.,} 1995]{utsumi95}

1706: \reference % (paper)

1707: Utsumi, W., N.\/ Funamori, T.\/ Yagi, E.\/ Ito, T.\/ Kikegawa,

1708: and O.\/ Shimomura, Thermal expansivity of MgSiO$_{3}$ perovskite

1709: under high pressures up to 20 GPa, \grl {\it 22,} 1005-1008, 1995.

1710: 

1711: \bibitem[{\it Vinet et al.,} 1987]{vinet87}

1712: \reference % (paper)

1713: Vinet, P., J.\/ Ferrante, J.~H.\/ Rose, and J.~R.\/ Smith,

1714: Compressibility of solids, \jgr {\it 92,} 9319-9325, 1987.

1715: 

1716: \bibitem[{\it Wang et al.,} 1994]{wang94}

1717: \reference % (paper)

1718: Wang, Y., D.~J.\/ Weidner, R.~C.\/ Liebermann, and Y.\/ Zhao,

1719: P-V-T equation of state of (Mg,Fe)SiO$_{3}$ perovskite:

1720: Constraints on composition of the lower mantle, \pepi {\it 83,}

1721: 13-40, 1994.

1722: 

1723: \bibitem[{\it Wang et al.,} 1996]{wang96}

1724: \reference % (paper)

1725: Wang, Y., D.~J.\/ Weidner, and F.\/ Guyot, Thermal equation

1726: of state of CaSiO$_{3}$ perovskite, \jgr {\it 101,} 661-672, 1996.

1727: 

1728: \bibitem[{\it Weidner and Wang,} 1998]{weidner98}

1729: \reference % (paper)

1730: Weidner, D.~J., and Y. Wang, Chemical- and Clapeyron-induced buoyancy

1731: at the 660 km discontinuity, \jgr {\it 103,} 7431-7441, 1998.

1732: 

1733: \bibitem[{\it Wentzcovitch et al.,} 1995]{wentzcovitch95}

1734: \reference % (paper)

1735: Wentzcovitch, R.~M., N.~L.\/ Ross, and G.~D.\/ Price, Ab

1736: initio study of MgSiO$_{3}$ and CaSiO$_{3}$ perovskites

1737: at lower-mantle pressures, \pepi {\it 90,} 101-112, 1995.

1738: 

1739: \bibitem[{\it Wolf and Bukowinski,} 1987]{wolf87}

1740: \reference % (article in book)

1741: Wolf, G.~H., and M.~S.~T.\/ Bukowinski, Theoretical study of 

1742: the structural properties and equations of state of MgSiO$_{3}$

1743: and CaSiO$_{3}$ perovskites: Implications for lower mantle

1744: composition, in {\it High-Pressure Research in Mineral Physics,

1745: Geophys. Monogr. Ser., 39},

1746: edited by M.~H.\/ Manghnani and Y.\/ Syono, pp.\/ 313-331,

1747: AGU, Washington, D. C., 1987.

1748: 

1749: \bibitem[{\it Wolf and Bukowinski,} 1988]{wolf88}

1750: \reference % (paper)

1751: Wolf, G.~H., and M.~S.~T.\/ Bukowinski, Variational stabilization

1752: of the ionic charge densities in the electron-gas theory of

1753: crystals: Applications to MgO and CaO, {\it Phys. Chem. Minerals,

1754: 15,} 209-220, 1988.

1755: 

1756: \bibitem[{\it Wood,} 2000]{wood00}

1757: \reference % (paper)

1758: Wood, B.~J., Phase transformations and partitioning relations in

1759: peridotite under lower mantle conditions, \epsl {\it 174,}

1760: 341-354, 2000.

1761: 

1762: \bibitem[{\it Wood and Rubie,} 1996]{wood96}

1763: \reference % (paper)

1764: Wood, B.~J., and D.~C.\/ Rubie, The effect of alumina on phase

1765: transformations at the 660 kilometer discontinuity from Fe-Mg

1766: partitioning effects, {\it Science, 273,} 1522-1524, 1996.

1767: 

1768: \bibitem[{\it Yeganeh-Haeri,} 1994]{yeganeh94}

1769: \reference % (paper)

1770: Yeganeh-Haeri, A., Synthesis and re-investigation of the

1771: elastic properties of single-crystal magnesium silicate

1772: perovskite, \pepi {\it 87,} 111-121, 1994.

1773: 

1774: \bibitem[{\it Zhang and Weidner,} 1999]{zhang99}

1775: \reference % (paper)

1776: Zhang, J., and D.~J.\/ Weidner, Thermal equation of state of

1777: aluminum-enriched silicate perovskite, {\it Science, 284,}

1778: 782-784, 1999.

1779: 

1780: \bibitem[{\it Zhao and Anderson,} 1994]{zhao94}

1781: \reference % (paper)

1782: Zhao, Y., and D.~L.\/ Anderson, Mineral physics constraints 

1783: on the chemical composition of the Earth's lower mantle,

1784: \pepi {\it 85,} 273-292, 1994.

1785: 

1786: \end{thebibliography}{}

1787: \end{article}

1788:    % You must type an \end{article}

1789:    % command after the references.

1790: 

1791: \newpage

1792: 

1793: %%  FIGURE CAPTIONS

1794: \begin{figure}

1795: \centerline{\epsfig{file=01.eps,width=5in}}

1796: \caption{Experimental pressure-temperature coverage of

1797: (Mg,Fe)SiO$_{3}$ perovskite.  Open symbols are for static

1798: experiments on X$_{\rm Mg}$ = 1.0, solid symbols

1799: are for X$_{\rm Mg}$ = 0.9, and

1800: open symbols with crosses are shock wave data.

1801: Hexagons, {\it Kudoh et al.,} [1987];

1802: right-facing triangles, {\it Ross and Hazen,} [1990]; 

1803: left-facing triangles, {\it Wang et al.,} [1994];

1804: upward pointing triangles, {\it Fiquet et al.,} [1998];

1805: downward pointing triangles, {\it Fiquet et al.,} [2000];

1806: diamonds, {\it Saxena et al.,} [1999];

1807: open circles, {\it Utsumi et al.,} [1995];

1808: solid circles, {\it Knittle and Jeanloz,} [1987];

1809: squares, {\it Funamori et al.,} [1996];

1810: stars, {\it Mao et al.,} [1991];

1811: pentagons, {\it Kato et al.,} [1995];

1812: circle with cross, {\it Duffy and Ahrens,} [1993].

1813: Solid curve is the lower mantle geothermal gradient of

1814: {\it Brown and Shankland} [1981]. 

1815: Static experimental work has direct

1816: coverage of approximately the upper one third.  The shock wave 

1817: data point that falls near the geotherm is perovskite plus

1818: magnesiow\"ustite.}

1819: \label{fig1}

1820: \end{figure}

1821: \clearpage

1822: 

1823: \begin{figure}

1824: \centerline{\epsfig{file=02.eps,width=5in}}

1825: \caption{Energetics of octahedral rotations:

1826: calculated total energies relative to the cubic structure

1827: ($Pm$3$m$) using VIB (curves) and LAPW [{\it Stixrude and

1828: Cohen,} 1993; {\it Hemley and Cohen,} 1996]

1829: (symbols) as a function of (left) $R$ and (right) $M$ point

1830: rotation angles as represented by the fractional change

1831: in the oxygen coordinate ($\delta_{R}$ and $\delta_{M}$).

1832: The orthorhombic structure ({\it Pbnm}) 

1833: occurs at $\delta_{R}$ = 0.0912 and $\delta_{M}$ = 0.0766

1834: at $P$ = 0 GPa.}

1835: \label{fig2}

1836: \end{figure}

1837: \clearpage

1838: 

1839: \begin{figure}

1840: \centerline{\epsfig{file=03a.eps,width=5in}}

1841: \centerline{\epsfig{file=03b.eps,width=5in}}

1842: \clearpage

1843: \centerline{\epsfig{file=03c.eps,width=5in}}

1844: \caption{Structural parameters and volumes of MD results:

1845: (a) $b/a$ axis ratio and (b) $c/a$ axis ratio.

1846: At low pressures and high temperatures the structure 

1847: begins to deviate from orthorhombic {\it Pbnm}, but no

1848: phase transitions, with the possibility of the 

1849: structure heading toward cubic at 12 GPa and 3000 K.

1850: (c) Volume.  Symbols are MD results and curves are Universal EOS

1851: fits every 300 K.}

1852: \label{fig3}

1853: \end{figure}

1854: \clearpage

1855: 

1856: \begin{figure}

1857: \centerline{\epsfig{file=04.eps,width=5in}}

1858: \caption{Volume coefficient of thermal expansion.  Solid curves are

1859: from this work.  Other symbols, from experimental work as follows:

1860: dashed curve (0 GPa), {\it Wang et al.} [1994]; 

1861: circle (0 GPa), {\it Fiquet et al.} [1998]; 

1862: square (0 GPa) and dashed-dotted line (25 GPa),{\it Funamori et al.}

1863: [1996]}

1864: \label{fig4}

1865: \end{figure}

1866: \clearpage

1867: 

1868: \begin{figure}

1869: \centerline{\epsfig{file=05.eps,width=5in}}

1870: \caption{Anderson-Gr\"uneisen parameter as a function of 

1871: temperature.  The circle is the 

1872: experimentally determined value

1873: from {\it Wang et al.} [1994]. The square is the value

1874: determined by {\it Masuda and Anderson} [1995] from the

1875: experimental data of {\it Utsumi et al.} [1995].

1876: Open circles are $\delta_{T}$s determined from Debye 

1877: theory by {\it Anderson} [1998].}

1878: \label{fig5}

1879: \end{figure}

1880: \clearpage

1881: 

1882: \begin{figure}

1883: \centerline{\epsfig{file=06.eps,width=5in}}

1884: \caption{Gr\"uneisen parameter as a function of

1885: temperature.  The solid circle is the 

1886: experimentally determined value from {\it Wang et al.} [1994],

1887: and the square is the value determined by {\it Masuda and Anderson}

1888: [1995] from the experimental data of {\it Utsumi et al.} [1995].

1889: Other symbols are from inversion of multiple experimentally

1890: determined data sets: triangle, {\it Bina} [1995];

1891: inverted triangle, {\it Jackson and Rigden} [1996]); and 

1892: diamond, {\it Shim and Duffy} 2000].  Open circles are

1893: $\gamma$s determined from Debye theory by {\it Anderson} [1998].}

1894: \label{fig6}

1895: \end{figure}

1896: \clearpage

1897: 

1898: \begin{figure}

1899: \centerline{\epsfig{file=07.eps,width=5in}}

1900: \caption{The ln $\gamma$ versus ln V.  The slope of each curve

1901: gives $q$: solid curve and squares, 300 K; dashed curve and circles,

1902: 1200 K; dotted curve and triangles, 2100 K; dash-dotted curve and

1903: inverted triangles, 3000 K.  The lower temperatures are all close to

1904: 1, with a slight rise to 1.07 at 3000 K.}                  

1905: \label{fig7}

1906: \end{figure}

1907: \clearpage

1908: 

1909: \begin{figure}

1910: \centerline{\epsfig{file=08.eps,width=5in}}

1911: \caption{Volumes of MgSiO$_{3}$ perovskite along the

1912: geothermal gradient of {\it Brown and Shankland} [1981].

1913: Dash-dotted curve is volume from our Universal EOS fit.  Dashed

1914: curve is from our Universal fit with $V_{0}$ fixed to 162.49 \AA$^{3}$

1915: [{\it Mao et al.,} 1991].  Dotted curve is for experimental

1916: data sets with $V_{0}$ fixed to 162.49 \AA$^{3}$ (see Table 2).  

1917: Both high temperature (F98+F00+S99) and all MgSiO$_{3}$ data

1918: sets fall on the same line.

1919: Shaded areas indicate the uncertainties in the fits of

1920: the experimental data sets.}

1921: \label{fig8}

1922: \end{figure}

1923: \clearpage

1924: 

1925: \begin{figure}

1926: \centerline{\epsfig{file=09a.eps,width=10cm}}

1927: \centerline{\epsfig{file=09b.eps,width=10cm}}

1928: \caption{(a) Densities and (b) bulk sound velocities

1929: of MgSiO$_{3}$ perovskite calculated

1930: along the lower mantle geotherm of {\it Brown and Shankland} [1981]

1931: using the thermal Universal EOS.

1932: Solid curve is for uncorrected volumes and 

1933: dashed curve is for $V_{0}$ set to the experimental value of

1934: {\it Mao et al.,} [1991].

1935: Squares are densities from PREM [{\it Dziewonski and Anderson,} 1981].

1936: Circles are velocities from the ak135-f seismological model

1937: [{\it Kennett et al.,} 1995; {\it Montagner and Kennett,} 1996].}

1938: \label{fig9}

1939: \end{figure}

1940: \clearpage

1941: 

1942: \begin{figure}

1943: \centerline{\epsfig{file=10a.eps,width=10cm}}

1944: \centerline{\epsfig{file=10b.eps,width=10cm}}

1945: \caption{(a) Densities and (b) bulk sound velocities

1946: of (Mg,Fe)SiO$_{3}$ perovskite calculated

1947: along the geotherm of {\it Brown and Shankland} [1981]

1948: using the experimental value of $V_{0}$

1949: {\it Mao et al.,} [1991] for X${\rm ^{pv}_{Fe}}$ = 0.00 (solid curve),

1950: 0.05 (dashed curve), and 0.10 (dotted curve).

1951: Squares are densities from PREM [{\it Dziewonski and Anderson,} 1981].

1952: Circles are velocities from the ak135-f seismological model

1953: [{\it Kennett et al.,} 1995; {\it Montagner and Kennett,} 1996].}

1954: \label{fig10}

1955: \end{figure}

1956: \clearpage

1957: 

1958: \begin{figure}

1959: \centerline{\epsfig{file=11a.eps,width=10cm}}

1960: \centerline{\epsfig{file=11b.eps,width=10cm}}

1961: \caption{(a) Densities and (b) bulk sound velocities

1962: of two simplified pyrolite models (67 vol \% pv, 33 vol \% mw)

1963: along the geotherm of {\it Brown and Shankland} [1981].

1964: Solid curve is X$^{\rm pv}_{\rm Mg}$ = 0.93, X$^{\rm mw}_{\rm Mg}$ = 0.82.

1965: Dashed curve is X$^{\rm pv}_{\rm Mg}$ = 0.96, X$^{\rm mw}_{\rm Mg}$ = 0.86.

1966: Squares are densities from PREM [{\it Dziewonski and Anderson,} 1981].

1967: Circles are velocities from the ak135-f seismological model

1968: [{\it Kennett et al.,} 1995; {\it Montagner and Kennett,} 1996].}

1969: \label{fig11}

1970: \end{figure}

1971: 

1972: \clearpage

1973: 

1974: %%  TABLES

1975: 

1976: \begin{planotable}{rrcccc}

1977: \tablewidth{41pc}

1978: \tablecaption{MD P-V-T Data Used in EOS Fits}

1979: \tablenum{1}

1980: \tablehead{

1981:    \multicolumn{1}{c}{P, GPa}&

1982:    \multicolumn{1}{c}{T, K}&

1983:    \multicolumn{1}{c}{a, \AA}&

1984:    \multicolumn{1}{c}{b, \AA}&

1985:    \multicolumn{1}{c}{c, \AA}&

1986:    \multicolumn{1}{c}{V, \AA$^{3}$}

1987: }

1988: \label{tbl1}

1989: \startdata

1990:   0 &  300 & 4.8166 & 4.9283 & 6.9649 & 165.33 \nl

1991:   0 &  600 & 4.8299 & 4.9397 & 6.9787 & 166.50 \nl

1992:   0 &  900 & 4.8439 & 4.9519 & 6.9944 & 167.77 \nl

1993:   0 & 1200 & 4.8586 & 4.9648 & 7.0107 & 169.11 \nl

1994:   0 & 1500 & 4.8732 & 4.9769 & 7.0289 & 170.48 \nl

1995:   0 & 1800 & 4.8886 & 4.9901 & 7.0455 & 171.87 \nl

1996:   0 & 2100 & 4.9109 & 5.0145 & 7.0578 & 173.80 \nl

1997:   0 & 2400 & 4.9311 & 5.0354 & 7.0688 & 175.52 \nl

1998:   0 & 2700 & 4.9525 & 5.0538 & 7.0831 & 177.28 \nl

1999:   0 & 3000 & 4.9672 & 5.0633 & 7.1076 & 178.76 \nl

2000:  12 &  900 & 4.7790 & 4.8856 & 6.8895 & 160.86 \nl

2001:  12 & 1200 & 4.7917 & 4.8951 & 6.9024 & 161.90 \nl

2002:  12 & 1500 & 4.8036 & 4.9051 & 6.9165 & 162.97 \nl

2003:  12 & 1800 & 4.8169 & 4.9157 & 6.9288 & 164.06 \nl

2004:  12 & 2100 & 4.8295 & 4.9256 & 6.9441 & 165.18 \nl

2005:  12 & 2400 & 4.8429 & 4.9367 & 6.9586 & 166.36 \nl

2006:  12 & 2700 & 4.8628 & 4.9539 & 6.9690 & 167.88 \nl

2007:  12 & 3000 & 4.9035 & 4.9531 & 6.9695 & 169.27 \nl

2008:  25 &  300 & 4.6971 & 4.8093 & 6.7728 & 152.99 \nl

2009:  25 &  600 & 4.7073 & 4.8169 & 6.7830 & 153.80 \nl

2010:  25 &  900 & 4.7181 & 4.8250 & 6.7948 & 154.68 \nl

2011:  25 & 1200 & 4.7285 & 4.8330 & 6.8055 & 155.53 \nl

2012:  25 & 1500 & 4.7390 & 4.8409 & 6.8166 & 156.38 \nl

2013:  25 & 1800 & 4.7497 & 4.8487 & 6.8291 & 157.27 \nl

2014:  25 & 2100 & 4.7609 & 4.8572 & 6.8408 & 158.19 \nl

2015:  25 & 2400 & 4.7728 & 4.8658 & 6.8514 & 159.11 \nl

2016:  25 & 2700 & 4.7850 & 4.8736 & 6.8645 & 160.08 \nl

2017:  25 & 3000 & 4.8015 & 4.8859 & 6.8773 & 161.34 \nl

2018:  50 &  900 & 4.6194 & 4.7307 & 6.6445 & 145.20 \nl

2019:  50 & 1200 & 4.6280 & 4.7366 & 6.6531 & 145.84 \nl

2020:  50 & 1500 & 4.6366 & 4.7414 & 6.6632 & 146.48 \nl

2021:  50 & 1800 & 4.6450 & 4.7474 & 6.6726 & 147.14 \nl

2022:  50 & 2100 & 4.6545 & 4.7524 & 6.6821 & 147.81 \nl

2023:  50 & 2400 & 4.6627 & 4.7589 & 6.6920 & 148.49 \nl

2024:  50 & 2700 & 4.6728 & 4.7644 & 6.7014 & 149.20 \nl

2025:  50 & 3000 & 4.6814 & 4.7700 & 6.7129 & 149.90 \nl

2026: 140 &  300 & 4.3572 & 4.5023 & 6.2698 & 123.00 \nl

2027: 140 &  600 & 4.3627 & 4.5046 & 6.2768 & 123.35 \nl

2028: 140 &  900 & 4.3687 & 4.5076 & 6.2839 & 123.74 \nl

2029: 140 & 1200 & 4.3747 & 4.5097 & 6.2903 & 124.10 \nl

2030: 140 & 1500 & 4.3806 & 4.5121 & 6.2967 & 124.46 \nl

2031: 140 & 1800 & 4.3867 & 4.5146 & 6.3027 & 124.82 \nl

2032: 140 & 2100 & 4.3923 & 4.5164 & 6.3104 & 125.18 \nl

2033: 140 & 2400 & 4.3986 & 4.5186 & 6.3170 & 125.55 \nl

2034: 140 & 2700 & 4.4048 & 4.5209 & 6.3235 & 125.92 \nl

2035: 140 & 3000 & 4.4109 & 4.5230 & 6.3309 & 126.31

2036: \end{planotable}

2037: 

2038: \clearpage

2039: 

2040: \begin{planotable}{lllllll}

2041: \tablewidth{41pc}

2042: \tablecaption{Equation of State Parameters$^{\rm a}$}

2043: \tablenum{2}

2044: \tablehead{

2045:    \multicolumn{1}{c}{}&

2046:    \multicolumn{1}{c}{V$_{0}$, \AA$^{3}$}&

2047:    \multicolumn{1}{c}{K$_{T_{0}}$, GPa}&

2048:    \multicolumn{1}{c}{K$^{\prime}_{T_{0}}$}&

2049:    \multicolumn{1}{c}{a, GPa}&

2050:    \multicolumn{1}{c}{b, GPa K$^{-1}$}&

2051:    \multicolumn{1}{c}{$\chi^{2}$, GPa}

2052: }

2053: \tablecomments{$^{\rm a}$MD are best fit EOS parameters to

2054: molecular dynamics P-V-T data.  Experimental results

2055: are the best fit Universal 

2056: EOS parameters to experimental data sets, except data taken

2057: from {\it Mao et al.} [1991] (M91). 

2058: Remaining experimental results are 

2059: F98, {\it Fiquet et al.} [1998] (27 data points);

2060: F00, {\it Fiquet et al.} [2000] (38 data points); 

2061: S99, {\it Saxena et al.} [1999], (37 data points);

2062: All, all X$_{\rm Mg}$ = 1.0 data (363 data points; see Figure 1).

2063: Values in parentheses are standard deviations.}

2064: \tablecomments{$^{\rm b}$Fixed equal to 4.}

2065: \tablecomments{$^{\rm c}$Fixed to $V_{0}$ from {\it Mao et al.} [1991].}

2066: \tablecomments{$^{\rm d}$Fixed to $K_{T_0}$ from {\it Yeganeh-Haeri} [1994].}

2067: \label{tbl2}

2068: \startdata

2069: This work \nl

2070: $\:\:\:\:$Birch-Murnaghan EOS & 165.40(6)& 274(1) & 3.73(2) & -2.00(2) & 0.00667(5) & 0.08561 \nl

2071: $\:\:\:\:$Universal EOS & 165.40(5) & 273(1) & 3.86(2) & -1.99(1) & 0.00664(5) & 0.06712 

2072: \vspace{0.7em}\nl

2073: Experimental results \nl

2074: $\:\:\:\:$M91 & 162.49(7) & 261(4) & 4$^{\rm b}$ \nl

2075: $\:\:\:\:$F98 & 162.65(15) & 262(6) & 3.41(22) & -1.79(21) & 0.00597(70) & 1.47438 \nl

2076: & 162.49$^{\rm c}$ & 267(3) & 3.25(10) & -1.79(21) & 0.00597(70) & 1.41566 \nl

2077: & 162.60(9) & 264$^{\rm d}$ & 3.33(2) & -1.79(21) & 0.00596(69) & 1.41333 \nl

2078: & 162.49$^{\rm c}$ & $264^{\rm d}$ & 3.41(3) & -1.80(35) & 0.00601(73) & 1.36143 \nl

2079: $\:\:\:\:$F00 & 162.74(56) & 239(8) & 4.41(18) & -1.85(6) & 0.00618(20) & 0.61052 \nl

2080: & 162.49$^{\rm c}$ & 245(4)& 4.29(11) & -1.85(6) & 0.00618(20) & 0.59369 \nl

2081: & 161.68(18) & 264$^{\rm d}$ & 3.88(2) & -1.85(6) & 0.00617(20) & 0.60570 \nl

2082: & 162.49$^{\rm c}$ & $264^{\rm d}$ & 3.52(4) & -1.73(8) & 0.00577(28) & 1.23307 \nl

2083: $\:\:\:\:$S99 & 163.25(63) & 242(6) & 4.45(10) & -1.72(1) & 0.00574(2) & 0.26719 \nl

2084: & 162.49$^{\rm c}$ & 256(5) & 4.20(10) & -1.72(1) & 0.00572(2) & 0.26227 \nl

2085: & 161.10(28) & 264$^{\rm d}$ & 4.06(4) & -1.71(1) & 0.00572(2) & 0.26645 \nl

2086: & 162.49$^{\rm c}$ & $264^{\rm d}$ & 3.89(2) & -1.75(2) & 0.00584(7) & 0.31819 \nl

2087: $\:\:\:\:$F98 + F00 + S99 & 162.93(33) & 235(6) & 4.72(13) & -1.85(12) & 0.00618(38) &

2088: 1.56307 \nl

2089: & 162.49$^{\rm c}$ & 244(4) & 4.52(13) & -1.86(12) & 0.00621(40) & 1.55478 \nl

2090: & 161.69(22) & 264$^{\rm d}$ & 4.08(1) & -1.87(12) & 0.00622(38) & 1.61901 \nl

2091: & 162.49$^{\rm c}$ & $264^{\rm d}$ & 3.75(2) & -1.73(15) & 0.00577(7) & 2.22349 \nl

2092: $\:\:\:\:$All & 162.31(10) & 249(6) & 4.40(20) & -1.92(10) & 0.00639(32) & 0.60337 \nl

2093: & 162.49$^{\rm c}$ & 243(8) & 4.55(28) & -1.90(10) & 0.00635(34) & 0.61628 \nl

2094: & 162.01(22) & 264$^{\rm d}$ & 3.92(8) & -1.88(10) & 0.00628(38) & 0.67860 \nl

2095: & 162.49$^{\rm c}$ & $264^{\rm d}$ & 3.75(2) & -1.67(20) & 0.00557(65) & 1.04352

2096: \end{planotable}

2097: 

2098: \vspace{3pc}

2099: 

2100: \begin{planotable}{llcl}

2101: \tablewidth{21pc}

2102: \tablecaption{Zero Temperature Equation of State Parameters}

2103: \tablenum{3}

2104: \tablehead{

2105:    \multicolumn{1}{c}{}&

2106:    \multicolumn{1}{c}{V, \AA$^{3}$}&

2107:    \multicolumn{1}{c}{K, GPa}&

2108:    \multicolumn{1}{c}{K$^{\prime}$}

2109: }

2110: \tablecomments{$^{\rm a}$Universal fit.  Other 

2111: theoretical fits are third-order Birch-Murnaghan fits.}

2112: \tablecomments{$^{\rm b}${\it Cohen} [1987].}

2113: \tablecomments{$^{\rm c}${\it Stixrude and Cohen} [1993].}

2114: \tablecomments{$^{\rm d}${\it Wentzcovitch et al.} [1995].}

2115: \label{tbl3}

2116: \startdata

2117: This work$^{\rm a}$ & 164.22 & 280 & 3.84 \nl

2118: PIB$^{\rm b}$ & 164.78 & 252 & 4.05 \nl

2119: LAPW$^{\rm c}$ & 160.74 & 266 & 4.2 \nl

2120: PWPP$^{\rm d}$ & 157.50 & 259 & 3.9

2121: \end{planotable}

2122: 

2123: \clearpage

2124: 

2125: \end{document}

2126: 

2127:   % All LaTeX input files must conclude with an

2128:   % \end{document} command.  LaTeX cannot process

2129:   % a file that does not contain this command.

2130: