Click Here GEOPHYSICAL RESEARCH LETTERS, VOL. 35, L10301, doi:10.1029/2008GL033867, 2008 for Full Article High-pressure polymorphism of Fe2P and its implications for meteorites and Earth’s core Przemyslaw Dera,1 Barbara Lavina,1 Lauren A. Borkowski,2 Vitali B. Prakapenka,1 Stephen R. Sutton,1 Mark L. Rivers,1 Robert T. Downs,3 Nabil Z. Boctor,4 and Charles T. Prewitt3 Received 5 March 2008; revised 7 April 2008; accepted 11 April 2008; published 16 May 2008. [1] Minerals with composition (Fe,Ni)2P, are rare, though important accessory phases in iron and chondritic meteorites. The occurrence of these minerals in meteorites is believed to originate either from the equilibrium condensation of protoplanetary materials in solar nebulae or from the later accretion and condensation processes in the cores of parent bodies. Fe-Ni phosphides are considered a possible candidate for a minor phase present in the Earth’s core, and at least partially responsible for the observed density deficit with respect to pure iron. We report results of high-pressure high-temperature X-ray diffraction experiments with synthetic barringerite (Fe2P) up to 40 GPa and 1400 K. A new phase transition to the Co2Sitype structure has been found at 8.0 GPa, upon heating. The high-pressure phase can be metastably quenched to ambient conditions at room temperature, and then, if heated again, transforms back to barringerite, providing an important constraint on the thermodynamic history of meteorite. Citation: Dera, P., B. Lavina, L. A. Borkowski, V. B. Prakapenka, S. R. Sutton, M. L. Rivers, R. T. Downs, N. Z. Boctor, and C. T. Prewitt (2008), High-pressure polymorphism of Fe2P and its implications for meteorites and Earth’s core, Geophys. Res. Lett., 35, L10301, doi:10.1029/2008GL033867. 1. Introduction [2] Fe-Ni-P minerals, e.g. nickelphosphide, schreibersite, barringerite, as well as analogous P-rich sulfides are often found in meteorites, such as Ollague (Imilac) pallasite [Buseck, 1969], Onello [Britvin et al., 2002], Yamato793274 [Koberl et al., 1991], and CM chondrites [Nazarov et al., 1997]. Equilibrium protoplanetary material condensation modeling [Pataev and Wood, 2000] suggests that these phases may have formed in the solar nebulae, but there is a strong argument that they might also have experienced later alterations, in the parent bodies or after delivery to Earth [Ivanova et al., 2005]. Understanding the thermodynamic properties and pressure-temperature phase relations of these minerals is essential in consideration of 1 Center for Advanced Radiation Sources, University of Chicago, Argonne National Laboratory, Argonne, Illinois, USA. 2 UNLV High Pressure Science and Engineering Center, Argonne National Laboratory, Argonne, Illinois, USA. 3 Department of Geosciences, University of Arizona, Tucson, Arizona, USA. 4 Geophysical Laboratory, Carnegie Institution of Washington, Washington, D. C., USA. the alteration hypothesis, and can provide important insights and constraints into the chemistry and physics of one of the most primitive environments in the Universe. Convincing evidence exists, based on comparison of the silicate Earth model and natural chondritic abundances of elements, that phosphorus is present in Earth’s core [McDonough and Sun, 1995; McDonough, 2003]. [3] In meteoritic samples, the (Fe,Ni)2P grains exhibit a range of Ni:Fe atomic ratios, varying from 0.5 to over 1.5 [Nazarov et al., 1998]. There are two known polymorphs of (Fe,Ni)2P, both found in meteorites. The more common is hexagonal barringerite with the C22 structure type and space group P-62m, the same as synthetic Fe2P. To our knowledge, besides an early compression study on a synthetic powder sample to 10 GPa [Fujiwara et al., 1981], and several low temperature moderate-pressure studies aimed at revealing the effects of pressure on Curie temperature [Abliz et al., 2006; Fujiwara et al., 1982], nothing is known about the high-pressure behavior of barringerite. An early highpressure quench study [Senateur et al., 1976] indicated that a second, orthorhombic polymorph of Fe2P, with C23 (Co2Si) type structure may be stable or metastable at ambient conditions. A recent study of samples from the Onello meteorite described an orthorhombic phase, named allabogdanite, with C23 structure type and space group Pnma [Britvin et al., 2002]. The relation between these two polymorphs in the Fe-Ni-P system is not yet completely understood, but a transformation between the same two structure types is known from another binary phosphide system, Co2P. In Co2P the stable, ambient b
phase has the orthorhombic C23-type structure, and upon heating to 1428 K transforms to the hexagonal, C22-type a
phase through a reconstructive mechanism, involving bond breaking and re-formation [Ellner and Mittemeijer, 2001]. The a
Co2P is stable at high temperature, and can be metastably retrieved at ambient conditions by rapid quenching. By analogy to Co2P, the following conclusions have been inferred for the (Fe,Ni)2P system: (1) barringerite is a high-temperature polymorph of (Fe,Ni)2P, and (2) ordering associated with significant Ni content or small Co content is the most likely reason for stabilization of the allabogdanite structure [Britvin et al., 2002; Buseck, 1969]. Here, we report results that change this interpretation. 2. Experiments [4] In order to bring definite answers to the interpretation of the stability of the two polymorphs in pure Fe2P system Copyright 2008 by the American Geophysical Union. 0094-8276/08/2008GL033867$05.00 L10301 1 of 6 L10301 DERA ET AL.: HIGH PRESSURE BEHAVIOR OF Fe2P L10301 Figure 1. SXD diffraction patterns of Fe2P. (a) The C22 type phase at 33 GPa before heating, (b) the C23 type phase at 33 GPa after heating. Grey squares mark locations of peaks of the C22 phase before heating, (c) projection of the two reciprocal spaces of hexagonal phase crystal (red/dark) and orthorhombic crystal (green/light) along the hexagonal 1 0 0 direction, (d) two crystals of the hexagonal phase in a diamond anvil cell at 20 GPa. we performed synchrotron single-crystal X-ray diffraction (SXD) experiments at high pressure and temperature. The starting material was obtained through high-temperature synthesis in evacuated silica tube from stoichiometric mixture of elements, which resulted in a coarse-grained mixture of Fe2P and FeP (part of Fe was lost due to reaction with the SiO2 tube walls). Nevertheless, it was possible to hand-pick several micrograins of approximately 10  10  5 mm size, which were identified using SXD analysis as single crystals of Fe2P. Comparison of unit cell parameters obtained from the single crystal sample (a = 5.8764(3) Å, c = 3.4493(1) Å) with the unit-cell-stoichiometry dependence [Carlsson et al., 1973] confirms that our starting material is close to ideal Fe2.0P1.0 composition. In the first compression experiment two single crystals were loaded into a diamond anvil cell (DAC) with 16:4:1 MeOH:EtOH:H20 pressure medium and ruby pressure standard (Figure 1d). The samples were compressed in small, 3 GPa pressure steps, with SXD data collected at each step. Up to 30 GPa there was no sign of excessive uniaxial strain (auxiliary material Figure S1).1 At 33 GPa significant broadening of single crystal peaks was observed, and one of the crystals was moderately heated 1 Auxiliary materials are available in the HTML. doi:10.1029/ 2008GL033867. with a diode-pumped fiber laser, operating at a wavelength of 1.064 mm [Prakapenka et al., 2008] to an approximate temperature of 1400 K, as estimated from the black body radiation. An SXD pattern collected after the heating revealed that the heated sample retained the single crystal character, but the positions of the diffraction peaks were quite different than before (Figures 1a and 1b). In contrast, the diffraction pattern of the unheated crystal remained unchanged. Three-dimensional reconstruction of reciprocal space and peak indexing revealed that the heated crystal transformed to an orthorhombic structure with unit cell a = 5.3814(6) Å, b = 3.447(3) Å, c = 6.422(2) Å (at 33 GPa) and space group Pnma. The new structure was easily identified as isostructural with b-Co2P. Comparison of the orientation matrices of the hexagonal and orthorhombic crystals revealed no obvious crystallographic relation (Figure 1c, auxiliary material Table S2), thus we conclude that the transition occurs through a reconstructive mechanism and re-crystallization. The new phase was further pressurized to 40 GPa with laser annealing at each pressure increase, to determine the equation of state (EOS). No further transformation or crystal orientation changes were found. After reaching 40 GPa a decompression experiment was conducted, without heating. The new, C23-type phase could be metastably quenched to ambient conditions without signs of excessive uniaxial 2 of 6 L10301 DERA ET AL.: HIGH PRESSURE BEHAVIOR OF Fe2P L10301 Figure 2. Mechanism of the phase transition in Fe2P. (a) and (b) The crystal structures of the hexagonal and orthorhombic phases, respectively. The triangles depict the equatorial triangle of the P coordination polyhedra. Yellow triangles are located at fractional atomic coordinate 0.5 along the viewing direction, whereas purple triangles are at coordinate 0. In Figure 2b the dashed triangles show the P atoms that changed location (moved along the viewing direction) during the transition. (c) The coordination polyhedron and the two possible P sites between which the transition occurs. (d) Equation of state of the hexagonal (green) and orthorhombic (red) phases of Fe2P. The vertical axis represents unit cell volume divided by the Z number. 3 of 6 L10301 DERA ET AL.: HIGH PRESSURE BEHAVIOR OF Fe2P Figure 3. EOS parameters for main candidates for lightelement-containing component of the Earth’s core. In EOS calculations there is an intrinsic correlation between the bulk modulus K0 and its pressure derivative K’, which can be visualized by plotting the values of K0 obtained for different fixed values of K’. The bulk modulus of both of Fe2P phases is 18% higher than that of Fe3S [Seagle et al., 2006] and Fe3P [Scott et al., 2007] and is comparable with hcp Fe [Dewaele et al., 2006; Mao et al., 1990] and Fe3C [Li et al., 2002]. The higher value of bulk modulus means that compared to the other candidate phases with lower K0, a much smaller content of Fe2P would need to be present in the core to account for the observed density deficit. strain (auxiliary material Figure S2). After completely releasing pressure, the sample was heated again to about 1400 K, and subsequent SXD analysis revealed that it transformed back to the original hexagonal barringerite single-crystal, but again, without an obvious relation of the orientation matrices (auxiliary material Table S2.). Since in the first experiment, the pressure at which the orthorhombic phase becomes stable was not constrained, we conducted a second experiment with a polycrystalline sample (containing hexagonal Fe2P with a small amount of FeP impurity) and NaCl pressure medium. This time the sample was pressurized in smaller, 1 GPa steps, and heated to ~1400 K at each step. Conversion to the C23 structure was observed at 8.0 GPa. Since the experiments conducted with both single crystal and polycrystalline samples, as well as with different pressure media produced the same transformation sequence, we rule out chemical, non-hydrostaticity or sample-orientation effects as possible causes of the transformation. 3. Results and Discussion [5] The two polymorphs have a very similar compressibility, as shown in Figure 2d. The bulk moduli of the C22 and C23 phases, determined using the 3-rd order Birch Murnaghan EOS are 174(7) GPa and 177(3) GPa, respec- L10301 tively (at K’ = 4.0, auxiliary material Table S1 and Figure 3). The ambient pressure density of the orthorhombic phase is marginally lower than that of the hexagonal phase, which explains the ambient stability of the latter. During compression at pressure of about 5 GPa the two V(p)/Z curves cross, and the orthorhombic phase assumes a higher density, which is consistent with the observed pressure at which the transition occurs. [6] The fact that at ambient temperature both phases can be compressed and decompressed by as much as 20 GPa past the transition pressure without inducing the transition constitutes a remarkable example of metastability. Our observation of the C23 phase in pure Fe end-member demonstrates that the presence of Ni is not necessary for the orthorhombic structure to be metastable at ambient pressure. While it is possible that Ni may have a stabilizing effect on the C23 structure, it seems unlikely, as (1) C23 phase has never been observed in pure Ni2P, and (2) based on the comparison of ionic radii of Fe and Ni one would expect the pressure of C22->C23 transition to be higher in the Ni-end member. In order to understand the origin of metastability of the two phases of Fe2P it is necessary to take a close look at the relation between their crystal structures. A common supercell can be conveniently defined for the purpose of such comparison, as shown in Figures 2a and 2b. Both structures share basically the same, pseudohexagonal arrangement of Fe atoms, but differ in the organization of the P substructure. P atoms are located in voids coordinated by 9 Fe atoms. If viewed along the hexagonal c direction, for every two neighboring, face-sharing voids only every other one is occupied. A conversion between the two structure types can be accomplished though diffusion of half of the P atoms along the c-hexagonal direction from the sites occupied in the C22 structure, to empty voids. The P atoms have to move by ~1.7 Å to the neighboring void (Figure 2c), with the transformation preserving the topology of the Fe lattice, which undergoes a relatively minor distortion to accommodate the new coordination scheme. While such transformation is topologically very simple, and the resulting structures most likely differ very little in total energy, breaking of the three Fe-P bonds, needed for the P atoms to move to the new voids requires substantial kinetic energy to overcome the activation barrier. [7] There is a significant similarity of the metastability and polymorphism in Fe2P to the high-pressure polymorphism of silica phases. The octahedral high-pressure polymorph of SiO2, stishovite, was shown to undergo first a structural distortion to the CaCl2 structure [Kingma et al., 1995] (at 50 GPa), and then a transformation to the a-PbO2 structure [Murakami et al., 2003] (at 121 GPa and high temperature), which involves diffusion of every other Si cation along the octahedral edge-sharing chains into previously unoccupied octahedral voids [Dera et al., 2002]. Interestingly, similar to Fe2P, the high-pressure a-PbO2type polymorph of silica (seifertite) can be metastably quenched to ambient conditions. SiO2 seifertite is also found in meteoritic samples, and its occurrence serves as a useful proxy of the pressure-temperature history of meteorites. It should be emphasized that the remarkable metastability of both hexagonal as well as orthorhombic Fe2P provides an indicator of thermodynamic history of meteor- 4 of 6 L10301 DERA ET AL.: HIGH PRESSURE BEHAVIOR OF Fe2P ites, which is even more convenient than SiO2, since the Fe2P phases do not amorphize, but instead undergo a singlecrystal-to-single-crystal transition. Occurrence of allabogdanite clearly indicates one of two possible scenarios: (1) the phase was formed at elevated pressure (above 8.0 GPa) and high temperature and after formation, it has never experienced heating above 1400 K at low pressure, or (2) Fe2P was initially formed as barringerite at moderate pressure and high temperature, and then transformed to allabogdanite most likely during a shock event on the parent body. In the second scenario any high-temperature ambient-pressure alteration after the impact can be clearly ruled out. Presence of hexagonal barringerite, on the other hand, indicates that the peak pressure and temperature of impact shock did not exceed 8.0 GPa and 1400 K. [8] The outer core of the Earth is believed to contain less than 1 wt. % of phosphorus [Stewart and Schmidt, 2007; Stewart et al., 2007], which is certainly less than the content required to stabilize the Fe2P structure. However, the estimate of the combined content of the light elements in the core (5 wt. %) [Wood et al., 2006], the ability of the C23 structure to accommodate multi-component solid solutions involving all top candidates for the light element core component [Chandra et al., 1980; Jernberg et al., 1984; Stewart and Schmidt, 2007], and the recent evidence of heterogeneity of the outer core density [Soldati et al., 2003] all make the possibility of Fe2P presence in the core more plausible. With the bulk modulus higher and the ambient density lower than that of Fe3P and Fe3S (Figure 3) much less Fe2P would need to be present in the core to account for the observed density deficit. [9] The C23 structure type is more common among binary and ternary systems than the C22 type, with examples including many silicides, germanides, and phosphides (often containing cations relevant for the deep Earth, such as Fe, Ni, Ti, Mn or Co), but relatively few sulfides. Since more compounds can accommodate the allabogdanite structure than the barringerite structure, it can be expected that the structural compatibility of Fe2P for alloying with other siderophile elements at high pressure and temperature should increase with respect to the ambient compatibility. However, with the limited pressure range covered in our experiments, it remains to be verified whether the C23 structure remains stable to the outer core pressure and temperature conditions. [10] In addition to sulfur and phosphorus, silicon is considered one of the key candidates for the light-element core constituent [Lin et al., 2003, 2002]. It has been demonstrated that quite a large amount (up to 35% atomic ratio) of phosphorus in the diiron phosphide structure can be substituted by silicon [Jernberg et al., 1984]. Interestingly, introduction of Si into the Fe2P structure opens a possibility of purely temperature-induced C22!C23 phase transition [Jernberg et al., 1984]. While Earth science researchers start to address the issues of interplay between different light elements in ternary systems between Fe, S and Si [Sherman, 1997] and Fe, S and P [Stewart and Schmidt, 2007], we are still very far from understanding the relevant multi-component phase diagrams. In particular, by analogy to the Co2P-Co2Si system, a solid solution between (Ni,Fe)2(P,S) and (Fe,Ni)2Si is possible, providing ways for L10301 formation of new compounds that could be stable at Earth’s core conditions. 4. Methods [11] X-ray diffraction data were collected at the bending magnet station 13BMD of GSECARS facility, at the Advanced Photon Source, ANL. The monochromatic incident beam wavelength was 0.3344Å. The incident beam was focused using a pair of Kirkpatrik Baez mirrors to a focal spot of approximately 0.008 by 0.015 mm. A MAR345 image plate detector, placed approximately 350 mm from the sample was used to record the diffraction images. The sample was rotated during the exposure within a range of w = ±20o Typical exposure time was 30s/degree. The data was processed using GSE-ADA software [Dera, 2007]. The pressure inside the DAC was measured using the ruby fluorescence method [Mao et al., 1986]. [12] Acknowledgments. This project was supported by a grant from the MRI Program, Division of Materials Research, National Science Foundation (NSF-DMR-0521179). X-ray data were collected at GSECARS (Sector 13), Advanced Photon Source (APS), Argonne National Laboratory. GSECARS is supported by the National Science Foundation, the U.S. Department of Energy, the W.M Keck Foundation, the U.S. Department of Agriculture and the State of Illinois. Use of the APS was supported by DOE-BES, under contract DE-AC02-06CH11357. References Abliz, M., Y. Uwatoko, T. Ohki, and H. Fuji (2006), New pressure-induced phase transitions in Fe2P, J. Phys. Soc. Jpn., 75, 123,706, 2006. Britvin, S. N., N. S. Rudashevsky, S. V. Krivovichev, P. C. Burns, and Y. S. Polekhovsky (2002), Allabogdanite, (Fe,Ni)2P, a new mineral from the Onello meteorite: The occurrence and crystal structure, Am. Mineral., 87, 1245 – 1249. Buseck, P. (1969), Phosphide from meteorites: Barringerite, a new ironnickel mineral, Science, 165, 169 – 171. Carlsson, B., M. Gölin, and S. Rundquist (1973), Determination of the homogeneity range and refinement of the crystal structure of Fe2P, J. Solid State Chem., 8, 57 – 67. Chandra, R., S. Bjarman, T. Ericsson, L. Häggström, C. Wilkinson, and R. Wäppling (1980), A Mössbauer and X-ray study of Fe2P1 – xBx compounds (x<0.15), J. Solid State Chem., 34, 389 – 396. Dera, P. (2007), GSE-ADA data analysis program for monochromatic single crystal diffraction with area detector, GeoSoilEnviroCARS, Argonne, Ill. Dera, P., C. T. Prewitt, N. Z. Boctor, and R. J. Hemley (2002), Characterization of a high-pressure phase of silica from the Martian meteorite Shergotty, Am. Mineral., 87, 1018 – 1023. Dewaele, A., P. Loubeyre, F. Occeli, M. Mezouar, P. I. Dogokupets, and M. Torrent (2006), Quasihydrostatic equation of state iron above 2 Mbar, Phys. Rev. Lett., 97, 215504. Ellner, M., and E. J. Mittemijer (2001), The reconstructive phase transformation b-Co2P!a-Co2P and the structure of the high-temperature phosphide b-Co2P, Z. Anorg. Allg. Chem., 627, 2257 – 2260. Fujiwara, H., M. Nomura, H. Kadomatsu, N. Nakagiri, T. Nishizaka, Y. Yamamoto, H. Fuji, and T. Okamoto (1981), Anisotropic lattice compression in Fe2P, J. Phys. Soc. Jpn., 11, 3533 – 3534. Fujiwara, H., H. Kadomatsu, and H. Fujjiwara (1982), Effect of uniaxial stress on the Curie temperature in Fe2P, J. Phys. Soc. Jpn., 51, 1401 – 1405. Ivanova, M. A., M. A. Nazarov, F. Brandstätter, L. V. Moroz, T. Ntaflos, and G. Kurat (2005), Mineralogical differences between metamorphosed and non-metamorphosed CM chondrites, Lunar Planet. Sci. Conf., XXXVI, 1054. Jernberg, P., A. A. Yousif, L. Häggström, and Y. Andersson (1984), A Mössbauer study of Fe2P1 – xSix (x < 0.35), J. Solid State Chem., 53, 313 – 322. Kingma, K. J., R. E. Cohen, R. J. Hemley, and H. K. Mao (1995), Transformation of stishovite to a denser phase at lower mantle pressures, Nature, 374, 243 – 245. Koberl, C., G. Kurat, and C. Brandstätter (1991), Lunar meteorite Yamato793274: Mixture of mare and highland components, and barringerite from the Moon, Proc. NIPR Symp. Antarct. Meteorites, 4, 33 – 55. 5 of 6 L10301 DERA ET AL.: HIGH PRESSURE BEHAVIOR OF Fe2P Li, J., K. H. Mao, Y. Fei, E. Gregoryanz, M. Eremets, and C. S. Zha (2002), Compression of Fe3C to 30 GPa at room temperature, Phys. Chem. Miner., 29, 166 – 169. Lin, J. F., D. L. Heinz, A. J. Campbell, J. M. Devine, and G. Shen (2002), Iron-silicon alloy in the Earth’s core?, Science, 295, 313 – 315. Lin, J., J. Campbell, L. Heinz, and G. Shen (2003), Static compression of iron-silicon alloys: Implications for silicon in the Earth’s core, J. Geophys. Res., 108(B1), 2045, doi:10.1029/2002JB001978. Mao, H. K., J. Xu, and P. M. Bell (1986), Calibration of the ruby pressure gauge to 800 kbar under quasi-hydrostatic conditions, J. Geophys. Res., 91, 4673 – 4676. Mao, H. K., Y. Wu, L. C. Chen, J. F. Shu, and A. P. Jephcoat (1990), Static compression of iron to 300 GPa and Fe0.8Ni0.2 alloy to 260 GPa: Implications for the composition of the core, J. Geophys. Res., 95, 21,737 – 21,742. McDonough, W. F. (2003), Compositional model for the Earth’s core, in The Mantle and Core, Treatise Geochem., vol. 2, edited by R. W. Carlson, pp. 547 – 568, Elsevier, Oxford. U.K. McDonough, W. F., and S. S. Sun (1995), The composition of the Earth, Chem. Geol., 120, 223 – 253. Murakami, M., K. Hirose, S. Ono, and Y. Ohishi (2003), Stability of CaCl2type and a-PbO2-type SiO2 at high pressure and temperature determined by in-situ X-ray measurements, Geophys. Res. Lett., 30(5), 1207, doi:10.1029/2002GL016722. Nazarov, M. A., F. Brandstätter, and G. Kurat (1997), Comparative chemistry of P-rich opaque phases in CM chondrites, Lunar Planet. Sci. Conf., XXVIII, 1466. Nazarov, M. A., F. Brandstätter, G. Kurat, and T. Ntaflos (1998), Chemistry of P-rich sulfides in Murchinson, cold Bokkelveld and Nagoya CM chondrites, Lunar Planet. Sci. Conf., XXIX, 1466. Pataev, M. I., and J. A. Wood (2000), The condensation origin of zoned metal grains in Bencubbin/CH-like chondrites: Thermodynamic model, Lunar Planet. Sci. Conf., XXXI, 1608. Prakapenka, V. B., A. Kubo, A. Kuznetsov, A. Laskin, O. Shkurikhin, P. Dera, M. L. Rivers, and S. R. Sutton (2008), Advanced flat top laser heating system for high pressure research at GSECARS: Application to the melting behavior of germanium, High Pressure Res., in press. L10301 Scott, H. P., S. Huggins, M. R. Frank, S. J. Maglio, C. D. Martin, Y. Meng, J. Santillán, and Q. Williams (2007), Equation of state and highpressure stability of Fe3P-schreibersite: Implications for phosphorus storage in planetary cores, Geophys. Res. Lett., 34, L06302, doi:10.1029/ 2006GL029160. Seagle, C. T., A. J. Campbell, D. L. Heinz, G. Shen, and V. B. Prakapenka (2006), Thermal equation of state of Fe3S and implications for sulfur in Earth’s core, J. Geophys. Res., 111, B06209, doi:10.1029/2005JB004091. Senateur, J. P., A. Rouault, R. Fruchart, J. J. Capponi, and M. Perroux (1976), Etude par spectrometrie Mossbauer des transformations cristallographiques sous hautes pressions de MnFeAs et Fe2P, Mater. Res. Bull., 11, 631 – 635. Sherman, D. M. (1997), The composition of the Earth’s core: Constraints on S and Si vs. temperature, Earth Planet. Sci. Lett., 153, 149 – 155. Soldati, G., L. Boschi, and A. Piersanti (2003), Outer core density heterogeneity and the discrepancy between PKP and PcP travel time observations, Geophys. Res. Lett., 30(4), 1190, doi:10.1029/2002GL016647. Stewart, A. J., and M. W. Schmidt (2007), Sulfur and phosphorus in the Earth’s core: The Fe-P-S system at 23 GPa, Geophys. Res. Lett., 34, L13201, doi:10.1029/2007GL030138. Stewart, A. J., M. W. Schmidt, W. von Westrenen, and C. Liebske (2007), Mars: A new core-crystallization regime, Science, 316, 1323 – 1325. Wood, B. J., M. J. Walter, and J. Wade (2006), Accretion of the Earth and segregation of its core, Nature, 441, 825 – 833.





















N. Z. Boctor, Geophysical Laboratory, Carnegie Institution of Washington, 5251 Broad Branch Road, N. W., Washington, DC 20015, USA. L. A. Borkowski, UNLV High Pressure Science and Engineering Center, Argonne National Laboratory, Building 434A, 9700 South Cass Avenue, Argonne, IL 60439, USA. P. Dera, B. Lavina, V. B. Prakapenka, M. L. Rivers, and S. R. Sutton, Center for Advanced Radiation Sources, University of Chicago, Argonne National Laboratory, Building 434A, 9700 South Cass Ave, Argonne, IL 60439, USA. (dera@cars.uchicago.edu) R. T. Downs and C. T. Prewitt, 209 Gould-Simpson Building, University of Arizona, Tucson, AZ 85721-0077, USA. 6 of 6