Graphite Precipitation Research Articles

Neutron diffraction study of austempered ductile iron

Crystallographic properties of an austempered ductile iron (ADI) were studied by using neutron diffraction. A quantitative phase analysis based on Rietveld refinements revealed three component phases, α-Fe (ferrite), γ-Fe (austenite), and graphite precipitate, with weight fractions of 66.0, 31.5, and 2.5 pct, respectively. The ferrite phases of the samples were found to be tetragonal,14/mmm, with ac/a ratio of about 0.993, which is very close to the body-centered cubic (bcc) structure. The austenite phase had C atoms occupying the octahedral site of the face-centered cubic (fcc) unit cell with about 8 pct occupancy ratio. A strong microstrain broadening was observed for the two Fe phases of the samples. The particle sizes of the acicular ferrite phase were studied by using small angle neutron scattering. The analysis suggested a mean rod diameter of 700 A. The scattering invariant predicts a ferrite volume fraction consistent with the powder diffraction analysis. A textbook case of nodular graphite segregation, with average diameters ranging from 10 to 20 μm, was observed by optical micrography.

Methane‐rich fluids in the oceanic crust

Analyses of fluid inclusions in plutonic rocks recovered from the slow‐spreading Southwest Indian Ridge (SWIR) record CH4 concentrations of 15–40 times those of hydrothermal vent fluids and of basalt‐hosted volcanic gases and provide the first direct sampling of CO2‐CH4‐H2O‐H2‐C‐bearing fluids in the oceanic crust. Compositional, thermal, and spatial analyses of these fluids are used to model the evolution of volatiles during the crystallization and cooling of mid‐ocean ridge magma chambers and to assess the potential importance of carbon‐bearing fluids in geochemical processes in the lower crustal component of hydrothermal systems. Results from these analyses show that the earliest fluids to be exsolved from the melts are dominated by CO2‐rich vapors, which with progressive fractionation evolved to more H2O‐rich compositions. These later fluids were most likely exsolved under immiscible conditions and involved the development of CO2‐H2O‐rich vapors and CO2‐H2O‐NaCl brines that were trapped during mineral growth, as well as during later high‐temperature fracturing events. CO2 + CH4 ± H2O‐rich fluids in olivine and plagioclase minerals that contain up to 30–50 mol % CO2 and 33 mol % CH4 may reflect respeciation of magmatic CO2 during cooling and attendant graphite precipitation. Phase equilibria suggest that these fluids reequilibrated at ∼500°–600°C and at fO2's ∼ −3 log units below, to close to QFM conditions. Alternatively, the inclusions may record respeciation of magmatic fluids attendant with the inward diffusion of H2 into the inclusions and reduction of entrapped CO2 during degassing and cooling of the gabbros throughout the subsolidus regime. Subsequent seawater reaction, at minimum temperatures of ∼400°C, with mafic‐rich layers within the gabbroic rocks or with ultramafic material that underlies the plutonic sequence resulted in the formation of CH4‐H2O fluids that contain up to 40 mol % CH4, molecular H2, and graphite(?) daughter minerals. These data provide strong evidence that the CH4‐H2O‐rich fluids produced during serpentinization reactions were trapped under equilibrium conditions in the presence of graphite at very near to QFM conditions. The ubiquitous occurrence of CH4‐rich fluids in oceanic crustal layer 3 rocks from the SWIR indicates that these fluids may be a significant and previously unrecognized source for these volatile species in some hydrothermal systems venting on the seafloor and that the deep‐seated fluids most likely play an important role in the transfer of carbon from the lithosphere to the hydrosphere.

γ-Fe中のCの固相エレクトロトランスポートに及ぼすグラファイトの析出ならびに析出しているセメンタイトの影響

γ-Fe中のCの固相エレクトロトランスポートに及ぼすグラファイトの析出ならびに析出しているセメンタイトの影響

Graphite precipitation in C--O--H fluid inclusions: closed system compositional and density changes, and thermobarometric implications

Equilibrium C–O–H fluid speciation calculations predict that graphite will precipitate from initially graphite saturated fluid inclusions during cooling and exhumation of metamorphic rocks. In the case that no mass is gained or lost by the inclusions, the original XO ratio [O/(O+H)] of the fluid phase must be maintained. Given this closed system constraint, the down-temperature progress of graphite precipitation can easily be monitored as a function of the varible XO, and produces some effects that are of significance to fluid inclusion studies: 1. Variation of the H2O : CO2 : CH4 relationship in the graphite-saturated COH fluid, namely increase of XH2O and decrease of the carbonic fraction; 2. Decrease of fluid density due to precipitation of graphite, which is denser than the residual fluid; 3. Alteration of the CO2 : CH4 ratio of the fluid, depending on the initial O : H ratio of the fluid: for XO>1/3, fluids increase their CO2 : CH4 ratio with decreasing temperature, and vice-versa. This implies that the CO2 : CH4 ratio measured at room T will not represent the trapping value, which is in any case closer to unity. As a consequence of density reduction, isochores extrapolated from densities observed at room temperature do not pass through the pressure-temperature conditions at which the inclusion was trapped, with pressure underestimates of up to 2 kbar. Actual P-T trapping conditions are located along the equilibrium “bulk isochore” (curve of constant-XO, constant-volume) of the fluid. Alteration of the CO2 : CH4 ratio is a mechanism by which a CO2-rich or CH4-rich carbonic phase can be formed from aqueous fluids that are slightly off the neutral XO=1/3 value. Subsequent segregation of this phase from the aqueous counterpart may account for the formation of pure CO2 and CH4 fluids in the upper crust.

A study of carbon distribution on the contact zone between two solid solutions

The diffusivity of carbon at 1,273 K in solid solutions of Ni, Co, and Fe was determined by measuring the initial graphite precipitation temperature on the interface between regions of low and high carbon content during cooling. The results show that diffusivity is independent of carbon concentration for the range of concentrations investigated. When two samples of the same metal with different carbon concentrations are brought into contact and held isothermally, diffusion of carbon occurs across the contact interface between the samples from the alloy with the higher carbon content to the alloy with the lower carbon content. The resulting carbon concentration profile in the area of the interface is a function of the initial carbon concentrations, C[sub 1] and C[sub 2], the annealing temperature T, and the annealing time t. The present study includes the experiment determination of the carbon concentration profile and the calculation of carbon diffusivity.

Phase diagram methods for graphitic rocks and application to the system C−O−H−FeO−TiO2−SiO2

Carbon-saturated C−O−H (GCOH) fluids have only one compositional degree of freedom. This degree of freedom is specified by the variable X o that expresses the atomic fraction of oxygen relative to oxygen+hydrogen. The only valid constraint on the maximum in the activity of GCOH fluid species is related to the bulk composition of the fluid, as can be expressed by X o. In fluid-saturated graphitic rocks, mineral devolatilization reactions are the dominant factor is determining the redox state of the metamorphic environment. X o is directly proportional to the fo2 of GCOH fluid, and because its value can only be affected by fluid-rock interaction, it is an ideal measure of the redox character and composition of GCOH fluid. Phase diagrams as a function of X o are analogous to the P-T-Xco2 diagrams used for binary H2O−CO2 fluids; this analogy can be made rigorously if the C−O−H fluid composition is projected through carbon into the O−H subcomposition. After projection, the fluid is described as a binary fluid with the components O and H, and the compositional variable X o. Description of GCOH fluids in this manner facilitates construction of phase diagram projections that define the P-T stability of mineral assemblages for all possible fluid compositions as well as fluid-absent conditions. In comparison to phase diagrams with variables based on the properties of fluid species, P-T-X o diagrams more clearly constraint accessible fluid compositions and fluid evolution paths. Calculated P-T-X o projections are presented for the C−O−H−FeO−TiO2−SiO2 system, a limiting model for the stability of Fe−Ti oxides in graphitic metapelites and phase relations in metamorphosed iron-formations. With regard to the latter, the stability of the assemblage qtz+mag+gph has been a source of controversy. Both the calculated C−O−H−FeO−TiO2−SiO2 system petrogenetic grid and natural examples, suggest that this assemblage has a large P-T stability field. Discrepancies between earlier C−O−H−FeO−SiO2 system phase diagram topologies are reconciled by the qtz+mag+gph=sid+fa phase field, a barometric indicator for metamorphosed-iron formations. A more general implication of calculated P-T-X o phase relation is that few inorganic mineral-fluid equilibria appear to be capable of generating hydrogen-rich, fO2, GCOH fluids at crustal metamorphic conditions. The utility of P-T-X o diagrams derives from the use of a true compositional variable to describe fluid composition, this approach can be extended to the treatment of carbon-undersaturated systems, and provides a simple means of understanding metasomatic processes of graphite precipitation.

The influence of silicon on carbon redistribution in steel weldments

Carbon redistribution was measured in ST1/ST2  Fe-2.5Si-0.8C/Fe-0.32Si-0.49C steel weldments in the temperature range 500–1000 °C. At the temperatures where austenite exists, carbon diffuses from ST1 into ST2; when ferrite is present, the diffusion flow reverses from ST2 into ST1. This effect is attributed to the degree of the silicon influence on the graphite precipitation and carbon activity in ST1 and ST2 steels. The opposite signs of the activity gradients in austenite and ferrite cause the reversal of the carbon diffusion when the annealing temperature is changed from the austenite to the ferrite temperature region. The carbon diffusion coefficients D C and the thermodynamic interaction coefficients ϵ C Si in austenite have been assessed from the experimental data for ST1 and ST2 steels.

Interlayers for diamond-coated cutting tools

Boron-containing chemical vapor deposition (CVD) films are being explored for use as interlayers between WC-Co substrates and CVD diamond coatings to enhance the adherence between these materials. The wear resistance and hardness of diamond coatings on tough WC-Co substrates gives an ideal coated tool material for many machining applications. CVD diamond applied directly onto WC-Co exhibits very poor adherence. Factors contributing to this problem include the solubility and diffusivity of carbon in cobalt, resulting in limited diamond nucleation and precipitation of graphite, and the large thermal expansion mismatch between WC-Co and diamond. Boron-containing films act as diffusion barriers to carbon diffusion into the cobalt and enhance chemical bonding at the coating-substrate interface. Compositional grading of the interlayer maximizes its effectiveness as an aid to adherence. The experimental results leading to these observations are discussed in this paper.

Petrologic phase equilibria and stable isotope fractionations of carbonate-silicate parageneses from granulite-grade rocks of Sri Lanka

The calc-silicate rocks and marbles of central and southern Sri Lanka can be characterized by three groups of parageneses: Cc+Dol+Phl±Fo±Spl±Di+Acc in the northern and eastern part of the Highland Complex, Di+Scp+Spn+Cc+Acc following to the west and Wo+Di+Scp+Spn±Cc+Acc in the most southwestern part. This distribution mirrors the compositional variation of the sedimentary precursors and the regional pressure gradient. During prograde metamorphism decarbonation with internal fluid buffering increased XCO2 to 0.75. Graphite precipitation at near-peak metamorphic conditions at constant oxygen fugacity reduced XCO2 to ∼0.5. The variation of δ18O (+26 to +9‰) and δ13C (+4 to −5‰) of carbonates can largely be explained by primary differences in the isotopic composition and in the silicate/carbonate contents of the protoliths. The particular light isotopic compositions are due to fractionation processes during decarbonation of the rocks. The dolomite-free scapolite-bearing rocks in general are isotopically lighter than dolomite-bearing parageneses. Calcite-graphite temperatures in the range of 750–900°C were calculated, using the experimentally derived thermometer of Scheele (1991). Calcite-silicate temperatures (150–900°C) were calculated with the increment method (Hoffbauer et al., 1994). These temperatures reflect near-peak equilibrium as well as different degrees of retrograde isotope resetting.

Dimensional changes in a sinter hardening alloy steel: R.W. Kiefer et al (Pitney Bowes Inc, Stamford, Connecticut, USA)

Different powder mixtures have been used to analyse the effect of presintering heat treatments on the composition and microstructure of TiMoCN–Ni based cermets. When methane is present in the atmosphere C contents after presintering can be associated to the C activity calculated from the C + 2H2 = CH4 equilibrium. H2 based atmospheres produce high C losses whereas N2–CH4–H2 gas mixtures induce strong carburization of the compacts. When no methane is present in the atmosphere, the C activity could be related to the C content of the compacts by the Boudouard equilibrium. Changes in C and N contents are also observed to depend critically on the phases present in the powder mixtures. Those containing free C, TiC, Mo2C or TiMoC phases show higher C and N changes after presintering than those based on quasi-monophasic TiMoCN powders. These data confirm that presintering treatments can be designed to modify the C/N ratio of TiMoCN–Ni cermets in order to avoid graphite precipitation and to have a closer control on the composition of carbonitride and metallic binder phases after sintering.

Auger Electron Spectroscopy Study of Grain Boundary Segregation in Alloy K-500: Part II. Behavior in Slow Strain Rate Tensile Test

An auger electron spectroscopy (AES) study was performed to characterize grain boundary segregation behavior and grain boundary precipitation occurring during slow strain rate tensile tests (SSRTTs) of alloy K-500. It was discovered that flake-like precipitates formed on grain boundaries of some test alloys which had a high bulk Cu level. Auger electron (AE) spectra taken from precipitates and AE mapping of grain boundary facets containing precipitates indicated that these precipitates were most likely titanium carbides. The grain boundary segregation that existed in as-processed alloys was also changed in the course of SSRTTs. A significant reduction in C segregation and a noticeable increase in Cu segregation were observed in test alloys after SSRTTs. Graphite precipitation was not detected in any of the test alloys before or after SSRTTs. The carbide precipitation and change in segregation pattern are qualitatively understandable on the basis of available thermodynamic data, but further study is necessary to fully explain the observed phenomena.

A Carbon Isotope Study of Graphites from the Kerala Khondalite Belt, Southern India: Evidence for $CO_{2}$ Infiltration in Granulites

Field relations of arrested charnockite patches, felsic pegmatite dikes, and coarse cordierite + graphite veins in the granulite facies supracmstal belt of southern Kerala suggest paleo-fluid channels. We have investigated the stable isotopic composition of graphites associated with these structures, as well as those in association with charnockites adjacent to calcsilicate horizons. The remarkably high values displayed by graphites from the structurally controlled zones ( - 8.6 to - ), together with the narrow spread in individual localities (), correlate with precipitation from isotopically homogeneous -rich fluids that either infiltrated along structural pathways (faults/ shears) or were transferred through magmatic conduits. Graphites from charnockite-calcsilicate associations show extreme enrichment ( up to -4.8%). Changing fluid regimes are detected from up to variation in among graphites sampled adjacent to calcsilicate horizons and within single crystals of coarse graphite associated with cordierite-rich veins. Steep isotopic gradients across fluid channels, and the preservation of biogenic carbon signatures in disseminated graphite flakes within adjacent metapelites (), suggest that influx was strongly channelized. Our study provides evidence for graphite precipitation in reduced crustal lithologies infiltrated by -rich fluids and illustrates that graphite can be used as a potential monitor to detect infiltration in granulites.

Microscale isotopic zonation in graphite crystals: Evidence for channelled CO influx in granulites

Coarse graphite in association with cordierite and quartz within a vein in a metapelite locality at Chittikara (CHT) in southern India illustrates precipitation of graphite from CO 2-rich fluids. Carbon isotope traverses within individual graphite crystals using a microsampling technique reveal strong isotopic zonations in graphite crystal from the vein, with δ 13C values oscillating between −13‰ in the core and −8.5‰ in the rim. CO 2 trapped within fluid inclusions in the cordierite shows δ 13C= −5‰, suggesting isotopic equilbrium with the core composition of the graphite. A graphite crystal from the metapelite immediately adjacent to the vein is also zoned, with a smooth rimward increase in 13C. Disseminated graphite flakes in the host metapelites ca. 1 m from the vein are, however, unzoned, and possess strongly negative δ 13C values (−34.4‰). The rimward increase in 13C, the absence of other carbon-bearing phases and the sluggish exchange rates for carbon in graphites preclude diffusive re-equilibration, and suggest that the different zoning patterns in the coarse graphite flakes within and adjacent to the vein assemblage at CHT have resulted from precipitation in different parts of a highly channelled fluid regime, where fluids are non-wetting and move episodically by a hydraulic fracture mechanism. Our study provides the first unambiguous insight into graphite precipitation mechanisms, and has potential implications for the nature of fluid interconnectivity and CO 2 transport and equilibration length scales within the crust.

Effects of pressure, time, and various additives on the crystallization of graphite and (Fe 1− xNi x) 3C carbide in the FeNiC system

Crystallization of graphite and (Fe 1− x Ni x ) 3C carbide from the FeNiC system has been studied at pressures up to about 5.5GPa. Pressure and time as crystallization factors were investigated. Numerous elements have been added to the base metal to study their individual effects on graphite precipitation. The microstructures of several specimens from different experiments are illustrated. It was found that pressure stabilizes (Fe 1− x Ni x ) 3C carbide and time promotes the formation of large graphite crystals. Flake-type graphite is obtained with Si, Se, Cd, Zn, Cu, I, and Al additions. Specific shapes are promoted by Te, Sb, Pb, Ba, and BaSi. Spheroidal graphite crystals are observed when Mg and MgSi are introduced.

Surface precipitation of graphite layers on carbon‐doped nickel and their stabilization effect against chemisorption and initial oxidation

AbstractA new method for coating the inner surface of ultrahigh‐vacuum (UHV) vessels with an inert material is proposed, namely coating with highly oriented graphite layers which precipitate on the surface of a carbon‐doped nickel substrate. By using various surface analytical methods, characterization of the precipitated graphite has been performed. AES measurements at elevated temperatures clarified the surface precipitation process on polycrystalline Ni–C. Low‐energy electron diffraction studies on Ni(111) showed that the surface‐precipitated graphite grew epitaxially with the substrate and its structure was found to agree with that of highly oriented pyrolytic graphite. Reflection high‐energy electron diffraction studies revealed that the basal planes (0001) of precipitated graphite preferentially cover the surface even in the case of polycrystalline Ni–C. AES measurements in an O2 atmosphere showed that the graphite‐covered surface was hardly oxidized by prolonged oxygen exposure. Thermal desorption spectroscopy measurements suggested that the desorption of CO after O2 dosing was much smaller from the graphite‐precipitated surface than from a stainless steel surface. From the standpoint of applications to problems in engineering, graphite precipitation on metals is found to be a promising method for making an inert and low‐outgassing surface for UHV use.

Microstructural investigation of a rapidly solidified FeSiC alloy

An FeCSi alloy treated with a nodularizer, Fe-45at.%Si-3at.%Mg was melt quenched by single roller melt spinning. Two new metastable phases were found in the as-quenched condition. The major metastable phase is a simple cubic structure with a lattice constant of 0.63 nm. By using the convergent beam electron diffraction technique, the detailed structure was analyzed to be point group 432 and space group P4 132 or P4 332. The second metastable phase is a body centred cubic (b.c.c.) structure with a lattice constant of 0.89 nm. Decomposition of these two metastable phases was studied by in situ heating in a transmission electron microscope. At around 300°C, graphite particles precipitate out in the simple cubic structured metastable phase. However, no phase transition was noted in the b.c.c. structured metastable phase after heating at around 300 °C for 2 h.

Effect of surface precipitation of titanium carbide on adherence of alumina film on steels

Alumina (Al 2O 3) was deposited on two types of iron-based alloy, type 14C and type 47C, with an r.f. magnetron sputtering method. Type 14C contains 0.14% carbon and 0.7% titanium. Type 47C contains 0.47% carbon and 0.7% titanium. Samples were annealed at 900 or 1100 K in a high vacuum. Al 2O 3 films on type 47C blistered or exfoliated. However, Al 2O 3 films on tyoe 14C exhibited neither of these behaviours at these two temperatures. Therefore the adherence of Al 2O 3 films on type 14C was better than that of Al 2O 3 on type 47C. Depth profile analyses with Auger electron spectroscopy and X-ray photoelectron spectroscopy showed that titanium carbide (TiC) precipitated at the interface between type 14C and Al 2O 3, and that graphite precipitated or sulphur segregated at the interface between type 47C and Al 2O 3. Precipitated TiC could relax thermal stress, had an anchoring effect and reacted with Al 2O 3. Therefore Al 2O 3 films on surfaces where TiC precipitated had good adherence. However, graphite precipitation at the interface between type 47C alloy and Al 2O 3 film was accompanied by volume expansion and the distribution of precipitated graphite was not uniform. The sulphur segregation at the interface made the Al 2O 3 film sensitive to thermal stress. Therefore Al 2O 3 films on the surfaces where graphite precipitated or sulphur segregated did not adhere well.

Improved ductility of Ni3Si by microalloying with boron or carbon

The effects of boron and carbon additions on the tendency for intergranular fracture in trinickel silicide intermetallics are reported. Melt spinning of Ni77Si23 alloyed with 0.1 at. pet boron results in full bend ductility and complete transgranular fracture compared with brittle intergranular fracture for the unmodified compound. Alloying with 0.1 at. pet carbon also produces full bend ductility but a mixed mode failure (≈30 pct transgranular). For both carbon and boron additions, reducing the Ni concentration of the base compound results in a greater percentage of intergranular fracture. The boron solubility limit depends on the Ni concentration of the base compound. For Ni77Si23, the solubility limit is between 0.1 and 0.2 at. pet boron. For compounds with silicon concentrations of 23.5 and 24.0 at. pct, the solubility limit is less than 0.1 at. pct boron. Boron additions above the solubility limit result in Ni3B precipitates which degrade the bend ductility and increase the percentage of integranular fracture. Alloying with carbon above the solubility limit (<0.1 at. pct) produces graphite precipitation. For Ni77Si23, increasing the carbon concentration from 0.1 to 1.0 at. pct resulted in no change in the ductility. Auger examination of the grain boundary composition showed strong segregation of both boron and carbon. Enrichment in silicon concentration was also observed.

Mass transfer during wall-rock alteration: An example from a quartz-graphite vein, Black Hills, South Dakota

Mass transfer and fluid-rock interaction have been evaluated along two sample traverses in low-sillimanite grade quartz-mica schist adjacent to a synmetamorphic quartz-graphite vein in the southern Black Hills, South Dakota. In an ~ 17 cm halo between apparently unaltered schist and the vein contact is an outer zone of cryptic alteration and three inner zones of visible alteration. The cryptic zone consists of the original prograde metamorphic mineral assemblage (quartz + biotite ± muscovite + plagioclase + microcline) plus anomalously high amounts of tourmaline. The outermost visible zone contains abundant graphite. The second visible zone is defined by intensive bleaching of the schist. The innermost visible zone, immediately adjacent to the vein, is tourmaline + quartz + plagioclase + limonite + graphite. The vein is composed almost entirely of quartz, but also contains trace amounts of graphite. Mass balance calculations indicate that Al was essentially inert. The predominant chemical changes during wall-rock alteration were addition of B and C from the vein-forming fluid along with loss of K from the wall rocks, corresponding to precipitation of tourmaline and graphite, and the progressive destruction of microcline, biotite, and muscovite toward the vein. In addition, the elements V, Cr, Cu, Zn, Pb, As, Sb, W, and Au were introduced into the country rock, whereas Si, Rb, Ba, and Cs were removed. On the basis of a constant Al reference frame, calculations indicate a net volume loss of 21–34% within one centimeter of the vein with little or no volume loss further from the vein. Fluid-rock interaction modeling suggests that between one and four equivalent masses of fluid interacted chemically with the most altered mineral assemblages. In addition, greater than one equivalent mass of reactive fluid penetrated to distances of at least 5 cm from the vein contact.

Role of late magmatic fluids in Merensky-type platinum deposits: A discussion

A number of features of platinum group element (PGE)-rich layers in layered intrusions provide evidence for the involvement of late magmatic fluids. The presence of pegmatitic textures, intergrowth of sulfides with intercumulus volatile-rich phases, high Cl content in volatile-bearing phases, and association with graphite and fluid inclusions have been taken by some authors as evidence for postcumulus transport and deposition of PGE by a fluid phase of late magmatic derivation. We argue that these apparently hydrothermal features are the result of postcumulus processes superimposed on layers whose extensive, uniform PGE concentrations were formed by magmatic cumulus processes. The association with pegmatites and volatiles arises from the relatively low solidus temperatures of orthocumulate layers, which cause these layers to act as traps for fluids migrating through the crystal pile. These fluids are mostly derived by vapor exsolution from fractionated intercumulus melt and are consequently enriched in CI. Redissolution of these fluids into vapor-undersaturated trapped liquid in orthocumulate layers causes recrystallization and formation of gabbro pegmatites. Immiscible sulfide-oxide liquid solidifies at near-solidus temperatures in the interstitial pore space between early-crystallizing silicate minerals, and it is forced into areas occupied by late-crystallizing CI-bearing silicates. This accounts for intergrowths of Cu and PGE-rich sulfides with amphiboles and micas, a ubiquitous feature of sulfide-bearing gabbros. Further exsolution of a vapor phase during late stages of crystallization causes deuteric alteration, formation of fluid inclusions, and possible local redistribution of sulfides and PGE. Cooling of this vapor phase may lead to precipitation of graphite at or below the orthocumulate solidus. Graphite deposition may be catalyzed by sulfides and PGE.