Reduced methane-bearing fluids as a source for diamond.

Diamond formation in the Earth has been extensively discussed in recent years on the basis of geochemical analysis of natural materials, high-pressure experimental studies, or theoretical aspects. Here, we demonstrate experimentally for the first time, the spontaneous crystallization of diamond from CH4-rich fluids at pressure, temperature and redox conditions approximating those of the deeper parts of the cratonic lithospheric mantle (5–7 GPa) without using diamond seed crystals or carbides. In these experiments the fluid phase is nearly pure methane, even though the oxygen fugacity was significantly above metal saturation. We propose several previously unidentified mechanisms that may promote diamond formation under such conditions and which may also have implications for the origin of sublithospheric diamonds. These include the hydroxylation of silicate minerals like olivine and pyroxene, H2 incorporation into these phases and the “etching” of graphite by H2 and CH4 and reprecipitation as diamond. This study also serves as a demonstration of our new high-pressure experimental technique for obtaining reduced fluids, which is not only relevant for diamond synthesis, but also for investigating the metasomatic origins of diamond in the upper mantle, which has further implications for the deep carbon cycle.

temperature range of our experiments ( Figure S1). A significant change in CH4 mole fraction only occurs upon cooling below 1000°C and the extent is ƒO2-dependent. Figure S1: Fluid composition calculated from the model of Huizenga [39] at 7 GPa, for ∆logƒO2 between IW and IW+2 for unit activity of carbon (aC=1). Note that decreasing temperature only has a significant effect on fluid composition at ∆logƒO2 > IW+1.0 and T <1100°C. At IW+0.5 (solid line) the fluid has a methane mole fraction of 0.96, similar to experimental results reported in this study and by Matjuschkin et al. [33]. Figure S2: Diamond-bearing experiments in a P-T-field. The graphite-diamond boundary was calculated after Kennedy&Kennedy [2] and Day [4] and melting point of pure gold after by Akella [58].
Analytical techniques. The recovered capsules were mounted in epoxy and ground down to expose the sample material. They were then polished with paraffin, Al2O3 since liquid oils can penetrate deep into graphite pockets and is difficult to remove, even when placed under vacuum. Samples were treated in an ultrasonic bath to remove the grinding material from the surface. No Al2O3 was detected during the subsequent microprobe investigation. The samples were investigated promptly after experiments using a Renishaw micro-Raman spectrometer (RM-1000), which is equipped with a Leica DMLM microscope, a 1800 groove per mm grating and a CCD detector. The spectra were obtained using the 532nm emission line of a Nd:YAG laser or the 633nm emission line of a HeNe laser that were calibrated using the 519 cm -1 band of a Si wafer. Spectra were collected over several accumulations operating at 10-50% laser power with 3-6min acquisition time for each scan. Analysis of fluid inclusions was mostly performed in confocal mode, while other measurements were made in non-confocal mode to investigate the olivine host for the inclusions. Large areas of each sample were investigated and in several cases additional measurements were made after repeated regrinding and polishing, allowing the sample to be sequentially investigated in the third dimension.
The run products, including the Ir-Fe redox sensors were chemically analysed using a JEOL JXA-8900 Superprobe at the University Frankfurt am Main. Analyses were performed at 15kV and 20nA for the silicate phases and at 20kV and 20nA for the Ir-Fe alloys employing a 1-3µm diameter beam. Fayalite, forsterite, wollastonite, albite, KTiPO4 as well as pure Ir, Fe, Au and Ni were used as standards and peak calibration.
The Fe 3+ /∑Fe content of the orthopyroxene starting material was determined by 57 Fe Mössbauer spectroscopy on optically clean, hand-picked separates following the procedure given by Woodland et al. [59].
Unpolarised FTIR spectra were obtained on selected samples using a Bruker Tensor 27 infrared spectrometer at the Australian National University in Canberra. Samples were prepared as doubly-polished 80-200 µm thick sections of the capsule assemblage to avoid the presence of fluid inclusions, During analysis, the sample chamber was purged with dry air to minimise interferences, while the background information was collected prior to each analysis. The crystal thickness was calculated by integrating the absorbance in the silicate overtone between 1625 and 2150 cm -1 divided by the 0.553 coefficient for unpolarised measurements [60]. Note that diamond is not in contact with graphite. Diamonds in figures f-h are exposed to the atmosphere. Therefore, the absence of CH4 peak and/or additional presence of H2O or other higher hydrocarbons can be expected. Unlike fluid inclusion in (c) trapped in olivine, which demonstrates the absence of H2O and a sharp peak of CH4.

Other experimental techniques
A number of previous studies with similar starting materials were apparently unsuccessful in producing diamond, such as we observe in our experiments [25,61,62]. Although we have no exact answer for these differences, there are number of possible reasons that can be considered. An overriding aspect is the formation and maintenance of CH4-rich fluids over at least 2 hours [25]. The following factors can influence the fluids stability and composition. (i) Catalytic reaction of fluid with metal capsules (e.g. Fe, Ni, Pd, Pt), as observed and described by Sokol et al. [25], Matveev et al. [63] Matjuschkin et al. [33] leading to fluid disequilibrium and methane instability [33]. In addition, the formation of carbides with Fe and Ni capsules can lead to a net loss of carbon from the fluid (sample). (ii) Hydrogen loss from the sample related to the choice of pressure medium material surrounding the capsule [33,64,65,66,67]. Such loss causes fluid oxidation and disequilibrium. (iii) Configuration of the experimental assembly is important for achieving as close to equilibrium conditions as possible and maintaining the system during the experiment. For instance, the position of H2-metal buffer is crucial for minimizing the H2-loss (see above). The inner buffer capsule should not be in contact with external outer capsule in order to prevent direct H2-diffusion out of the sample (rather than into the sample) [33,67]. On the other hand, placing the buffer external to the sample, can lead to hydrogen-loss from the assembly rather than imparting the ƒH2 on the sample [43,62]. (iv) Instable buffering due to use of talc, to produce higher ƒH2 [43,61,67]. Dehydration of talk does not guarantee a stable ƒH2 in experiment, which can affect the fluid composition. (v) Finally, but not least, an adequate experimental duration is required in order for solid organic materials (e.g. stearic acid) to produce the fluid phase. Run durations of at least 2 hours appear to be necessary [25,62]. All these aspects together may help to explain the differences between our study and previous work. However, there are maybe additional reasons, which we did not consider here. For example, even the production of the identical assembly may lead to differences due to use of materials from different suppliers (e.g. dense polycrystalline CaF2 vs. pressed CaF2 powder [33,67]. Table S1: Starting mix composition obtained by mixing of 40% F7-olivine and 60% F7orthopyroxene from the Finsch mine, South Africa [36]. Rare Earth Elements (REE) Zr, Hf, Sc, V, Ce,Yb were added to the starting mix in 130 ppm concentrations in form of oxides. b.d.l. = below detection limit.

SiO2
TiO2  Conservative uncertainty of ƒO2 measurements is ±0.05 log units, as determined by the standard deviation calculated from 5 to 20 analyses in each sample. a Fe-FeO buffered experiment [33] b Unspecified higher hydrocarbons, including C2H6 c Estimated for the temperature range 1250 -850°C