Analytical electron microscopy study of the plessite structure in four IIICD iron meteorites
Explore the source record for details and available documents.
Engineering topics
Publications and source records attributed to Goldstein, J. I..
Explore the source record for details and available documents.
Explore the source record for details and available documents.
Almost all the submicron metal particles in the glass coating of breccia 15286 are rounded and contain two phases, metal and sulfide (FeS). The Ni content of the metal phase as determined by scanning transmission electron microscope X-ray analysis ranged from 9.4 to 15.5 wt %. The sulfide was nearly stoichiometric troilite containing up to 1.3 wt % Ni. The close chemical and microstructural similarities between the coarse (greater than 1-micron) and the submicron sized metal indicate a common origin probably as meteoritic debris. Metal-silicate melt experiments under controlled oxygen partial pressured indicate that the metal particles may have formed from a fine dispersion of immiscible metal-sulfide droplets into an impact-generated silicate melt.
The iron-enriched section of the phase diagram for the ternary alloy Fe-Ni-C of various compositions is determined at 773, 873, 923, and 1003 C. The two-phase tie lines and three-phase tie triangles are measured by electron microprobe analyses. Tie lines in samples without bulk equilibrium are obtained by extrapolated interface compositions under the assumption of local equilibrium at the interface. It is shown that Ni addition somewhat reduces carbon solubility in austenite while decreasing the stability of the carbide phase. In particular, the carbide is always poor in Ni relative to the coexisting metal phase(s).
Measurements of the distribution coefficients of various elements (Ni, Co, P, Pt, Ir, Ge, Cr, and Au) between solid and liquid phases were made for FeNi alloys. A high thermal gradient crystal growing furnace was employed in order to promote plane front solidification, rapid mixing in the liquid and a minimum of diffusion in the solid. Eight measurements of the nickel distribution coefficient gave an average value of 0.91 plus or minus 0.03 and this value was independent of the Ni concentration in the range 5.5 to 8.75 wt. % Ni. Equilibrium distribution coefficients were also calculated by correcting for the diffusion boundary layer present in the liquid. These values are within 20% of the measured distribution coefficient values. Except for Co and Ni, all the measured distribution coefficient values are closer to 1.0 than those calculated by Scott (1972) for the IIIAB irons. Using the experimental values and the calculated equilibrium values for ternary additions to FeNi melts, it appears that fractional crystallization processes will not explain the observed chemical variations in the major iron meteorite groups. Alternative explanations are suggested which also recognize the fact that cooling rates vary widely in several chemical groups.
The metal and sulfide phases in the Luna 24 soil samples were studied with the optical microscope and the electron microprobe. The compositions of the metal particles fall into three groups based on their Ni and Co contents: (1) Samples of meteoritic composition which have undergone metamorphism on the lunar surface. (2) Samples of submeteoritic, low Ni and low Co contents, including most of the metal particles observed. These particles are contained in glass and agglutinate particles and were probably formed by the mixing of meteoritic metal with lunar metal produced by the reduction of silicates during shock-impact. (3) Samples of high-CO content probably formed by mixing of meteoritic material with high-Co metal from the mare basalt or by fractional crystallization from a metal silicate melt. The sulfide minerals were also studied. These are almost pure FeS, and crystallized from a late stage liquid in the mare basalt. Three high-Ni sulfides were also found in the glass phase of agglutinates.
The metallographic structures of eight severely reheated chondrites - Farmington, Ramsdorf, Orvinio, Wickenburg, Lubbock, Rose City, Arapahoe, and Tadjera - were studied using optical, SEM and electron microprobe techniques. The following metallographic criteria were used to estimate the post-shock residual temperature of the chondrites: melted metal-troilite appearance, presence of martensite, phosphorus enrichment of metal and averaging of central metal grain compositions. The presence of phosphides and secondary kamacite are due to slow post-shock cooling rates. Ni rim gradients indicate both extensive remelting of metal grains and relatively fast cooling.
The iron rich portion of the Fe-Ni-Co ternary diagram was determined at four temperatures. The phase boundaries and tie-lines of the (alpha + gamma) phase field were measured by analyzing the alpha and gamma phases with an electron microprobe. Grain boundary allotrimorphs of the alpha phase were observed in the polished and etched sections of samples which were step cooled from the gamma phase into the (alpha + gamma) region. Widmanstaetten-type microstructures composed of gamma-precipitates were observed in samples which were directly heated from room temperature into the (alpha + gamma) region.
A numerical model for diffusion-controlled phase growth has been applied to the exsolution of phosphide lamellae in lunar metal grains. Computer simulations reproduce observed composition profiles, and reveal the influence of cooling rate on dimensional and compositional parameters of phosphide and metal. At lower cooling rates, phosphide lamellae are larger and the concentration of P in the metal host close to the interface is lower. Cooling rates inferred for Apollo 16 samples, based on compositions and dimensions of the phosphide-metal grains, are mostly in the range 1-100 C/day. These rates correspond to burial depths of 5-0.5 m for melt rocks and 3-0.3 m for breccias. This is in good agreement with thicknesses of lunar cooling units determined by other techniques.
Several experiments were performed in order to evaluate the effect of carbon on phosphate reduction in synthetic systems. It was attempted to simulate in the experiments conditions occurring during lunar impact processes, but without shock pressure. Temperature, oxygen fugacity, and bulk chemistry were evaluated separately in order to determine the conditions which are suitable for carbon reduction. It appears on the basis of the results of the reported investigation that carbon can be an effective reducing agent during reheating events such as those encountered by lunar soils and breccias. Phosphate reduction may be viewed as a two-step process in which carbon is mobilized as CO during heating and preferentially dissolved in the metal phase. It then acts as a reducing agent on cooling. Gas phase transport and diffusion of carbon in metal are sufficiently rapid to allow uniform carbon distribution both within and between metal grains. The availability of metal from meteorites and carbon from the solar wind is probably sufficient to make reduction by carbon a significant process on the lunar surface.
Phosphorus is often present in lunar rocks and soils in bulk concentrations in excess of 0.5 wt % P205. The minerals apatite and whitlockite account for most of the phosphorus. However, it may also be present in metal grains. The relations between metal particles and the phosphates commonly found adjacent to these particles are investigated, taking into account studies involving three Apollo 17 rocks. It appears that phosphorus in lunar metal particles is related to the phosphate minerals by a redox reaction. The independent phosphates probably formed over much of the cooling period of the rock. While the rock was at high temperature, however, phosphorus was free to diffuse and dissolve in the metal droplets providing conditions were sufficiently reducing. If the rock was partially molten, this process would take less than 1 hr. If cooling occurred slowly under less reducing conditions, some of the phosphorus in the metal particles would be oxidized to form whitlockite.
Results are reported for two sets of experiments conducted to investigate the distribution and origin of phosphorus in lunar metal particles. The first set measured the equilibrium oxygen fugacity as a function of temperature for synthesized Fe-Ni and Fe-Ni-P alloys; the second set imposed various oxygen pressures so that the P distribution among the coexisting phases could be observed. The conditions of temperature, oxygen fugacity, and time necessary to produce P contents similar to those found in lunar metal particles are determined. The results show that: (1) the P distribution in lunar-type metal is controlled by oxygen fugacity, temperature, and bulk composition; (2) the P distribution is limited by the reaction rate at the metal surface and by the amount of phosphate in contact with the metal; (3) the nucleation and growth rate of phosphate controls the rate of P loss during oxidation; and (4) an oxygen fugacity of 10 to the -20th power atm at 950 C is required to saturate iron with P. It is concluded that a reducing species such as carbon may establish a local equilibrium and prevent oxygen communication with the surrounding rock.
Carbides and the surrounding metal of eight particles from Apollo 14, 16, and 17 soils were analyzed for C, Fe, Ni, and Co with the electron microprobe. In addition, carbide-containing metal particles in two Apollo 17 rocks were studied. Carbide-metal assemblages were found in anorthositic hornfels, melt rock, and agglutinate, but not in any primary igneous rocks. Carbon measurements show that cohenite is the carbide phase present in all the samples. Cohenite exsolves from the carbon-rich metal phase during cooling of the lunar metal particle in the host rocks or soil fragments. Measurements of two-phase interface compositions indicate equilibration for carbides and metal during cooling below 600 C and possibly below 500 C.
The moderately magnetic fractions of soils from a bright ray of South Ray Crater and from the continuous ejecta of North Ray Crater are described. The clasts in these fractions are classified as agglutinate, ANT melt rocks, glassy melt rocks, and ANT hornfels/breccia. Determinations are made of the Ni-Co and P contents of single-phase metal grains in the clasts. It is found that most of these grains have compositions in the meteoritic range and that the compositions tend to cluster at about 6% Ni. A significant amount of Fe-rich metal is also detected in some clasts along with a clear P-enrichment in metal from one soil sample. Meteoritic contamination of lunar metal having compositions in or close to the meteoritic range is considered. A model involving nine separate factors is proposed as an explanation of the observed metal compositions.
Explore the source record for details and available documents.
The grape cluster metal particle 63344,1 found in lunar coarse fines is examined using the scanning electron microscope (SEM), electron microprobe, and an optical microscope. This metal particle is approximately 0.5 cm in its largest dimension and consists of hundreds of metallic globules welded together to form a structure somewhat like a bunch of grapes. Electron microprobe analysis for Fe, Ni, Co, P, and S in the metal was carried out using wavelength dispersive detectors. No primary solidification structure is observed in the globules, and the particle is slow cooled from the solidification temperature (nearly 1300 C) taking days to probably months to reach 600 C. Two mechanisms for the formation of globules are proposed. One mechanism involves the primary impact of an iron meteorite which produces a metallic liquid and vapor phase. The second mechanism involves the formation of a liquid pool of metal after impact of an iron meteorite projectile followed by a secondary impact in the liquid metal pool.
The microstructure and chemical composition of metallic spherules from the lunar soil were studied by optical, electron-probe, and scanning-electron-microscope techniques. Metallic spherules produced experimentally were compared to the lunar spherules with respect to structure, chemistry, and cooling rates. The structure of the lunar spherules indicates an origin by solidification of molten globules of metal. The spherules are probably produced from both lunar and meteoric sources, the major source being the metal phase present in lunar rocks and soil.
Calculated compositions for metal crystallized from anorthositic melt rocks or feldspathic basalts such as 14310, 68415, and 77017 are (1) meteoritic in the initial stages or if metal is the only phase crystallized, and (2) higher in Co than meteoritic metal as a result of fractional crystallization of metals and silicates. Metal in coarse anorthositic rocks 15415, 15455, 60015, 64435, and 78238 is not of meteoritic composition and has Co contents from 0 to 9% with a Ni/Co ratio of about 1. The formation of anorthosites and the composition of the earliest metal are discussed with attention to Ni concentrations and the Ni/Co ratio. It is shown that anorthositic hornfels and breccias contain metal of a composition not found in their source material, coarse anorthosite; it is suggested that the hornfels includes metal grains probably derived with little change in composition from ancient meteorites.