New insight into polycrystalline diamond genesis from modern nanoanalytical techniques

Earth-Science Reviews 136 (2014) 21–35

Contents lists available at ScienceDirect

Earth-Science Reviews journal homepage: www.elsevier.com/locate/earscirev

New insight into polycrystalline diamond genesis from modern nanoanalytical techniques Dorrit E. Jacob a,⁎, Larissa Dobrzhinetskaya b, Richard Wirth c a b c

Australian Research Council Centre of Excellence for Core to Crust Fluid Systems and Department of Earth and Planetary Sciences, Macquarie University, North Ryde, NSW 2109, Australia Department of Earth Sciences, University of California at Riverside, Riverside, CA 92521-0412, USA Helmholtz Centre Potsdam, GFZ German Research Center for Geosciences, Telegrafenberg, D-14473 Potsdam, Germany

a r t i c l e

i n f o

Article history: Received 23 December 2013 Accepted 11 May 2014 Available online 18 May 2014 Keywords: Diamond Earth's mantle Carbonado Ultra-high pressure metamorphism TEM Subduction

a b s t r a c t Technical developments in analytical methods that reach nanometer spatial resolution have enabled the interrogation of smaller, submicron-sized inclusions in diamond that had previously been elusive. This has inspired and enabled studies of non-classical diamond species from different geological settings, resulting in a strongly faceted and dynamic picture of diamond formation. This article reviews the leap of knowledge achieved by employing state-of-the-art analytical methods with high spatial resolution to polycrystalline diamonds from different settings, i.e. from kimberlite, from crustal ultra-high pressure metamorphic terranes and alluvial carbonados. While crustal metamorphic diamonds are generally formed under oxidizing conditions, polycrystalline diamond from the Earth's mantle and carbonado have inclusion suites reflecting variable, and sometimes extreme, redox conditions. Diamond fluid compositions, however, fall in the same compositional field for worldwide diamond fluids, regardless of their geodynamic environment. On the basis of thermodynamic equilibrium data for C–H–O fluids in the mantle we argue that submicron inclusions in diamonds are products of local remobilization connected to fluid-fluxed partial melting and redox freezing. Thus, evidence from these inclusions complements information from classical work on larger inclusions and allows a unique direct insight into the medium in which diamond formed. © 2014 Published by Elsevier B.V.

Contents 1. 2.

3.

4. 5.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . Modern micro- and nanoanalytical techniques applied to diamond 2.1. X-ray tomography and micro-X-ray computed tomography 2.2. Micro-X-ray fluorescence . . . . . . . . . . . . . . . 2.3. Focused ion beam (FIB) milling . . . . . . . . . . . . . 2.4. Transmission electron microscopy (TEM) . . . . . . . . 2.5. NanoSIMS . . . . . . . . . . . . . . . . . . . . . . Polycrystalline diamonds from kimberlitic sources . . . . . . . 3.1. Mineral intergrowths: macroinclusions . . . . . . . . . 3.2. Oxidation state and clues to formation mechanisms . . . 3.3. Insights from nano-inclusions . . . . . . . . . . . . . Carbonado . . . . . . . . . . . . . . . . . . . . . . . . . Polycrystalline diamonds from ultrahigh-pressure terranes . . . . 5.1. Nanoinclusions in the Erzgebirge polycrystalline diamonds 5.2. Nanoinclusions in the Kokchetav polycrystalline diamonds 5.2.1. Marbles and calcareous-silicate gneisses . . . . 5.2.2. Feldspathic gneisses . . . . . . . . . . . . . . 5.3. Diamond-forming fluid media and oxidation state . . . . 5.4. Source of carbon and nitrogen aggregation . . . . . . . . 5.5. Evidence from fluids for crust–mantle interaction . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

⁎ Corresponding author. Tel.: +61 298508428; fax: +61 298508943. E-mail addresses: [email protected] (D.E. Jacob), [email protected] (L. Dobrzhinetskaya), [email protected] (R. Wirth).

http://dx.doi.org/10.1016/j.earscirev.2014.05.005 0012-8252/© 2014 Published by Elsevier B.V.

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

22 22 22 23 23 24 24 24 24 25 26 27 28 28 29 29 29 29 29 30

22

D.E. Jacob et al. / Earth-Science Reviews 136 (2014) 21–35

6.

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1. Diamond fluids in the Earth's mantle . . . . . . . . . . . . . . . . . . 6.2. Diamond fluids in polycrystalline diamonds . . . . . . . . . . . . . . . 6.3. Oxidation conditions recorded by polycrystalline diamonds: a matter of scale 6.4. Chemical environments for diamond formation . . . . . . . . . . . . . . 7. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction The study of mineral inclusions in diamonds from kimberlites has been critical in shaping our current understanding of the formation of diamond in the Earth's mantle (Navon, 1999; Stachel and Harris, 2008). However, instrumental limitations have, until recently, restricted studies to micrometer-sized inclusions in larger, monocrystalline diamonds. Smaller inclusions and inclusions in small diamonds from kimberlitic and non-kimberlitic settings remained elusive. Progress in instrumental development now enables accurate chemical and structural analyses at unprecedented spatial resolution, which has resulted in a leap in our understanding of small-scale, even atomscale processes in minerals. These new microanalytical methods have allowed diamond inclusion studies to be carried to the next level, as submicron inclusions are now accessible, thus building on a solid foundation from studies of large inclusions. This has inspired a new generation of innovative studies of smaller and rarer diamond species, whose inclusion content had not been accessible before. Here we review recently developed microanalytical techniques that are not yet routinely applied in the Geosciences and illustrate the leap in knowledge in the study of smaller diamonds and their inclusions that these methods have enabled and will continue to support. We compare and contrast evidence from classical studies of micrometer-sized inclusion suites (here referred to as “macroinclusions”) with recent results for submicron inclusions suites (“microinclusions”) focussing on TEM investigations of focused ion beam (FIB) prepared samples and evaluate the contribution of these to our current view of diamond formation. Polycrystalline diamonds are aggregates of diamond crystals with heterogeneous grain sizes and random orientation. An unusual feature is that these diamond aggregates can have a highly porous structure with up to 30% porosity (Heaney et al., 2005), which indicates that they formed from a volatile-rich medium strongly oversaturated in carbon (Sunagawa, 1990). Monocrystalline diamonds can contain a long history of growth and dissolution resulting in complex zonation patterns (Howell et al., 2012). In contrast, polycrystalline diamonds form rapidly (Orlov, 1977; Sunagawa, 2005), presenting snapshots of diamond formation conditions that complement the information from slowly grown monocrystalline diamond. Polycrystalline diamonds are found worldwide in rather contrasting settings. They occur in typical crustal rocks in ultra-high pressure metamorphic (UHPM) terranes as well as in kimberlites, the main source of diamonds from the Earth's mantle, and also in alluvial deposits with unknown sources. Kimberlites are volatile-rich exotic magmatic rocks that contain diamond and mantle xenoliths originating from depths of 150– N450 km. Polycrystalline diamond is found in Group I kimberlites in Africa and Siberia where it can locally amount to ca. 20% of the total diamond production (K. de Corte, pers. comm. 2012). UHPM terranes occur along continental margins delineating a wide spectrum of Phanerozoic collisions of continental lithospheric plates (Ernst, 2006; Ernst and Liou, 2008; Dobrzhinetskaya, 2012; Schertl and Sobolev, 2013). They are composed of metasedimentary and meta-igneous rocks with continental affinities that were subjected to deep subduction up to N250 km followed by exhumation. Microdiamonds have been reported from almost a dozen of UHPM terranes ranging in age from 640 Ma to

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

30 30 31 31 32 32 33 33

2–3 Ma, six of them with well-documented microdiamond occurrences (Dobrzhinetskaya, 2012; Schertl and Sobolev, 2013) and two of these contain polycrystalline diamonds. Lastly, polycrystalline diamond is found in alluvial deposits, most famous perhaps are the carbonado diamonds from Brazil and Central Africa, whose sources are as yet unknown and for which a number of contrasting and mutually exclusive models have been proposed (McCall, 2009). 2. Modern micro- and nanoanalytical techniques applied to diamond 2.1. X-ray tomography and micro-X-ray computed tomography X-ray tomography provides information about the occurrence and spatial distribution of fluid and solid-state inclusions in diamond (Jacob et al., 2011). Amongst the solid inclusions, mineral phases can be identified by applying techniques such as Raman spectroscopy, FTIR spectroscopy, X-ray diffraction analysis (Kopylova et al., 2010) and transmission electron microscopy (TEM), particularly electron diffraction (Wirth et al., 2009). Their chemical composition can be measured by electron microprobe analysis, μ-X-ray fluorescence analysis, TEM energy dispersive analysis and/or electron energy-loss spectroscopy (EELS), and nanometer resolution secondary ion mass spectrometry (Nano-SIMS) analysis. The chemical composition of fluids included in diamond can be determined by infrared spectroscopy (IR, FTIR, Weiss et al., 2010, 2013) and synchrotron-based μ-X-ray fluorescence analysis (Klein-BenDavid et al., 2004; Sitepu et al., 2005). However, if the grain size of the individual phases inside the inclusions is significantly smaller than 100 nm and/or the inclusions are polycrystalline, then most of these techniques render insufficient spatial resolution. Nanometer-sized phases can be reliably determined only by applying TEM techniques (electron diffraction, high-resolution imaging, analytical electron microscopy (AEM)) on site-specific foils prepared by focused ion beam (FIB) milling (Wirth, 2004, 2009; Dobrzhinetskaya et al., 2010). This is why we emphasize FIB/TEM techniques in the following section. The application of various appropriate techniques mentioned above provides answers to important questions, such as: How many inclusions are present in a particular volume? Where are they located in that particular stone? What is the chemical composition of the inclusions and which phases are present? What is the size of the inclusions and that of the constituting phases? How can they be made accessible for analysis, and by which analytical method? Micro-computed X-ray tomography (μCT) is a non-destructive technique that allows identification of inclusions larger than 1 μm per voxel (volume pixel), depending on the original sample size. CT unveils the internal structure of a sample (inclusions, porosity) down to a micrometer scale based on X-ray absorption, which depends on density variations and differences in chemical composition of the sample. A series of two-dimensional X-ray absorption images are acquired and then assembled into a 3D image. X-ray sources are usually synchrotron radiation facilities or high-brilliance rotating anodes. The resolution is basically defined by the original focal spot size and the sample size. The larger the sample the more beam broadening due to scattering effects results, thus reducing spatial resolution. Micro-computed X-ray tomography works well as long as there is a large contrast in X-ray

D.E. Jacob et al. / Earth-Science Reviews 136 (2014) 21–35

attenuation between the host and the inclusions. To date, a number of applications of X-ray computed tomography in Earth Sciences have been published (Mees et al., 2003; Jacob et al., 2011; Cnudde et al., 2012; Ketcham and Koeberl, 2013 and references therein). 2.2. Micro-X-ray fluorescence Micro-X-ray fluorescence analysis is a method that excites X-ray photons in a specific volume defined by the diameter of the primary synchrotron beam and records the X-ray photons exited from atoms present in that particular volume. The penetration depth of the synchrotron beam is limited by the absorption of the incident beam by the target material. The absorption of the excited X-ray photons by the target material defines the detection limit for individual elements. The use of X-ray-fluorescence techniques on micrometer-sized inclusions in minerals requires a focused X-ray beam with very high intensity. Synchrotron radiation provides the intense source and has been used for μ-X-ray fluorescence analysis of inclusions in minerals since the nineteen eighties (Frantz et al., 1988). It is a non-destructive method that leaves the diamond sample unaffected for further investigations. This method allows the collection of spectra with excellent detection limits for individual elements (e.g. 10 ppm for Ti and b1 ppm for Zn; Sitepu et al., 2005). The spectra can also be used for creating elemental distribution maps. Further details of the technique are presented in one of the first studies of that kind on diamond inclusions by Sitepu et al. (2005), and references therein. With μ-X-ray fluorescence analysis the chemical composition of an inclusion in diamond can be determined as well as its location inside the stone. However, to identify the phase and the crystal structure of inclusions another method such as μ-X-ray diffraction analysis can be applied. Identification of micrometer-sized inclusions in diamond using this method is described in detail in recent articles (Kopylova et al., 2010; Smith et al., 2011). The major advantage of this method is that many solid-state inclusions can be identified simultaneously, thus providing an indication of the frequency of individual phases in a single stone. Together with μ-X-ray fluorescence analysis described above it is possible to obtain information on the number of inclusions, their phase compositions and chemical compositions in a comparatively large volume of several hundred μm3. Micro-X-ray diffraction analysis does not require the use of synchrotron radiation. Modern rotating anode X-ray generators can be considered as high-brilliance X-ray sources suitable for this kind of diffraction analysis. 2.3. Focused ion beam (FIB) milling While the X-ray based techniques described above are state of the art methods, they are not appropriate to study nanometer-sized inclusions because of their limited spatial resolution. The leading tool to study the “nano-world” inside inclusions is the focused ion beam (FIB) technique for site-specific TEM sample preparation combined with TEM. Thanks to the advent of this method in the Earth Sciences, we know today that many inclusions with a grain size well below 100 nm exist in diamond. Before FIB sample preparation was available to prepare electron transparent foils from diamond (starting from approximately 2000), there were only two options to produce specimens applicable for TEM studies. One of these is crushing of the sample and placing the crushed material onto a carbon covered TEM grid. This method has many disadvantages, particularly loss of spatial information and inhomogeneous thickness, the latter excluding application of line scans and element mapping. The second method is the argon-ion milling technique that was usually applied in the past to prepare electron transparent foils. However, this method has the tremendous drawback that due to the incident angle of the ions (a few degrees) preferential thinning occurs with the result that the weakest material is always sputtered first. Momentum transfer from Ar ions onto the target atoms is comparatively

23

low because of the lower atomic number of Ar ions compared with Ga ions used in FIB milling, and the lower acceleration voltage in argon ion milling devices. For argon ion milling of diamond, this translates into very long milling times, because bond strength in diamond exceeds that of all of its inclusions and consequently will be sputtered first. In contrast, the FIB technique allows preparation of electron transparent foils or membranes from diamond from the location of interest (site-specific) leaving the inclusions untouched. However, because of the extreme hardness of diamond the sputtering yield is significantly smaller than with silicates. To overcome this disadvantage a gas injection system (GIS) is usually employed that injects water vapor close to the target surface to be sputtered thus facilitating the sputtering process. In a FIB device Ga ions are extracted from a liquid gallium source by applying an extraction voltage (6–14 keV) to it. An appropriate suppressor voltage prevents extraction of electrons from the liquid metal ion source (LMIS). Variation of the suppressor voltage adjusts the beam current to a preset value (e.g. 2.2 μA) maintaining the extractor voltage at constant value. A constant ion current is required for reproducible sputtering conditions and automation of the sputtering process. The extracted Ga ions are accelerated to 30 keV and focused onto the target, which they hit with sufficient momentum to sputter atoms from the target surface. Inserting appropriate apertures into the path of the ion beam controls the beam diameter. The sputtering rate depends on the acceleration voltage (momentum of the Ga ions), the target material (bond strength) and the angle of incidence (penetration depth). At high angles such as 90° – the ion beam is normal to the target surface – the sputtering rate is low because the Ga ions penetrate deep into the sample and interact with the sample atoms deep in the sample, but sputtering of target atoms occurs predominantly at the surface. At very low incident angles (1–2°), with the Ga ion beam approximately parallel to the foil surface, sputtering rate is high because the Ga ions do not penetrate deep into the target and sputtering of surface atoms occurs much more frequently. Further details of the FIB technique and the FIB sample preparation of TEM foils with application of a gas injection system can be found elsewhere (Wirth, 2004; Giannuzzi and Stevie, 2005; Desbois et al., 2008; Wirth, 2009, 2010). A typical FIB prepared TEM foil has the dimensions 15 μm × 10 μm × 0.15 μm and takes approximately 4 hours to produce in a fully automated process. Significantly thinner membranes with around 35 nm thickness can be produced in a SEM–FIB combined DualBeam equipment with in-situ lift out technique (Wirth, 2009). However, for the study of inclusions in diamond, it is sometimes more favorable to work with thicker foils to increase the chance of finding nanoinclusions that are still intact. If the diameter of the inclusion is well below 100 nm and the total foil thickness is 200–300 nm, intact inclusions can be observed, while the foil is thin enough to allow electron transparency in diamond. In these cases, the inclusion's chemical composition can be measured with EDX and/or electron energy-loss spectroscopy (EELS) analysis. Subsequently, the electron beam is focused at the inclusion directly, thereby opening it by intensive electron sputtering and releasing the liquid content into the system's vacuum. A second EDX analysis after opening the inclusion can demonstrate which elements were present in a solid crystalline or quench phase, which were dissolved in the fluid or gas phase and, by mass balance, which escaped into the vacuum. FIB-prepared TEM foils have several advantages over other specimen preparation techniques. (1) FIB prepared foils can be prepared at uniform thickness. This allows the acquisition of elemental maps of the inclusions because the recorded X-ray intensities depend only on the concentration of the individual elements and not on sample thickness. Line scan analyses are also possible but are restricted to inclusions that are not sensitive to electron irradiation damage and electron sputtering (sulfides, garnet, olivine). (2) The location of the foil in the sample can be easily reconstructed because the sputtering site is visible on the sample surface under an optical microscope at moderate magnification. (3) No preferential sputtering of the inclusions occurs. (4) No

24

D.E. Jacob et al. / Earth-Science Reviews 136 (2014) 21–35

further carbon coating of the foil is required, which is essential when investigating carbon-based materials. Because of the limited sputtering depth (10–15 μm) inclusions must be located close to the surface to be accessible by FIB otherwise mechanical polishing is required. 2.4. Transmission electron microscopy (TEM) FIB-prepared foils or membranes are studied with TEM methods. TEM is an ideal tool to investigate nanometer-sized inclusions in diamond or any other minerals (Lee, 2010), because it allows the acquisition of chemical, microstructural and structure information from the same object even at nanometer-scale. The chemical information is derived from energy dispersive spectra (EDS) of collected X-ray photons escaping from the sample. Another more qualitative method is the acquisition of Z-contrast sensitive images with a high-angle annular dark field (HAADF) detector in the scanning transmission (STEM) mode. The HAADF images display differences in chemical composition in an inclusion composed of different phases. In some special cases EELS can be used for qualitative chemical analysis. For example, if iron is present in an inclusion there is an overlap of the Fe-L peak in the EDX spectrum with the F-K peak. If we want to prove the presence of fluorine in a particular phase, the fluorine F-L edge in the EEL spectrum can be used because there is no interference with the iron Fe-L edges. The chemical composition of Fe-carbides can be measured with EEL spectroscopy by calculating the Fe/C ratio (Kaminsky and Wirth, 2011). In addition to the chemical composition, grain size, grain shape and defect structure (dislocations, stacking faults, nitrogen platelets) of the inclusion are of interest. This information can be obtained by bright-field and dark-field imaging. Usually, the crystal structure, which unambiguously identifies a solid-state inclusion, is derived from electron diffraction patterns or convergent beam electron diffraction (CBED) patterns. However, for selected area electron diffraction the nanometer-sized crystals are too small and CBED usually cannot be applied because the focused electron beam immediately destroys the crystal structure rendering the inclusion amorphous. The best applicable approach in this case to derive structural information is high-resolution lattice fringe imaging. Due to very short exposure times necessary to acquire images with lattice fringes with a CCD camera (b 1 s) the crystal will not be destroyed before the image is acquired. After acquisition the lattice fringe image is Fourier transformed (FFT), calculating a diffraction pattern. From the diffraction pattern in reciprocal space it is possible to measure the length of several diffraction vectors and angles between them. Transformation of the length of the vectors into real space (d-spacing) allows a comparison of the measured d-spacing with calculated d-spacing of the phase we expect to be present from chemical data. A comparison of the measured d-spacing with the calculated d-spacing and a comparison of the angles between adjacent vectors allow an identification of the phase. Thus, TEM methods enable the identification of nanometer-sized phases inside inclusions in diamond and minerals in general. 2.5. NanoSIMS In addition, FIB-prepared TEM foils can also be used for NanoSIMS studies, which can deliver isotopic information at high spatial resolution. This has been demonstrated with carbonate inclusions in diamond (Wirth et al., 2007), nitride inclusions in coesite (Dobrzhinetskaya et al., 2010) and determinations of the isotopic composition of oxygen and carbon in lower mantle carbonate and silicate inclusions in diamond (Kaminsky et al., 2012). NanoSIMS uses an ion beam (Cs+ or O− depending on the target material to be sputtered) usually b100 nm in diameter that is scanned over a preselected area e.g. 4 × 4 μm. The secondary ions sputtered from the target material are collected and guided into a mass spectrometer for isotopic analysis. Details of this

method are given elsewhere (Stadermann et al., 2005; Floss et al., 2006; Hoppe et al., 2013 and references therein). Many recent applications of these methods focused on polycrystalline diamonds from such contrasting settings as UHP metamorphic terranes, alluvial sources and kimberlites. Metamorphic diamonds from crustal UHP rock are typically in the micron range, thus an ideal target for these methods. Carbonados and polycrystalline diamonds from kimberlites are porous and contain minerals intergrown with the diamond grains in the pores. These minerals have been studied extensively to enlighten the origin of their hosts, but because of the interconnectivity of the pore space, the intergrown minerals are not shielded against alteration and the evidence they bring has to be critically interpreted. Studying submicron minerals and fluids actually included in the diamonds delivers unambiguous evidence and has proved very beneficial, as reviewed in detail below. 3. Polycrystalline diamonds from kimberlitic sources Polycrystalline diamond (PCD) is classified by grain size as framesite (finer; Gurney et al., 1984) or bort (coarser; Orlov, 1977) and typically has a porous structure. Gray to silver in color, these diamond aggregates are often intimately intergrown with silicates, oxides and/or sulfides. Diamond crystals are randomly oriented and range over at least five orders of magnitude in size (Sobolev et al., 2009). PCD is also referred to as “diamondite”, which emphasizes the perspective that these aggregates are rocks, however small, with diamond as the rock-forming mineral while silicates and other phases are minor and accessory components. Polycrystalline diamond is a regular member of the diamond suite in a number of kimberlite pipes worldwide, where it is believed to amount to about 20% of the total production (K. de Corte, pers. comm., 2012). Kimberlite pipes in Africa (Venetia, Premier, Jwaneng, Orapa) yield polycrystalline diamonds, as do some diamond mines in Siberia (Mir, Aikhal; Sobolev, 1977). This variety of diamond is only lately beginning to receive increased attention and sheds light on another facet of diamond formation in the Earth's mantle, namely formation by rapid crystallization from a fluid supersaturated in carbon (Orlov, 1977; Gurney and Boyd, 1982; Jacob et al., 2000, 2011). Much of what we know about them comes from study of phases intergrown with, rather than actually being included in, the diamond crystals. These phases range from several tens to hundreds of micrometers in size and are termed here “macroinclusions” to set them apart from the submicron-sized phases included in the diamond crystals that have only been accessible in recent times. 3.1. Mineral intergrowths: macroinclusions High-pressure minerals intergrown with the diamond aggregates are commonly interpreted as syngenetic to the diamonds, based on the fact that they include small diamond crystals (Kurat and Dobosi, 2000) or display chemical correlations with the diamonds (Jacob et al., 2000, 2004). However, it is noteworthy that unlike inclusions in diamond these intergrown phases are not shielded from later metasomatic overprint. The aggregates contain mostly silicates and oxides of eclogitic, websteritic and peridotitic affinity, broadly similar to the macroinclusion suite found in monocrystalline diamonds, although characteristic differences exist. Chromium-poor (non-peridotitic) garnets in polycrystalline diamonds have a more restricted compositional range than inclusions in monocrystalline diamonds, containing typically less than 6 wt.% CaO (Fig. 1). Thus, most classify as websteritic (Aulbach et al., 2002), and don't reflect the large variation in grossular component seen in eclogitic garnets from mantle xenoliths. Nevertheless, some phases can clearly be classified as eclogitic, for example omphacitic clinopyroxene and rutile in an Orapa sample (Jacob et al., 2011).

D.E. Jacob et al. / Earth-Science Reviews 136 (2014) 21–35

25

Fig. 1. Major element compositions of garnets intergrown with diamond in polycrystalline diamonds from kimberlites compared to garnets from eclogitic xenoliths (Roberts Victor Mine, RSA, large gray field; Jacob, 2004). Gray field labeled A = peridotitic–websteritic garnets; dark gray field B = websteritic–eclogitic compositions. Note the more restricted grossular component compared to typical eclogitic xenoliths. Data from Dobosi and Kurat (2002, 2010), Jacob et al. (2011) and Kurat and Dobosi (2000).

While restricted in their grossular contents, the garnets show a large variation in almandine contents (13–46%), covering nearly the whole variation reported for eclogite xenoliths (Fig. 1). In this respect, their major element composition is similar to low Cr garnet megacrysts found in kimberlites (Gurney et al., 1979). Magnetite, while rarely reported from monocrystalline diamonds (Harris, 1968; Sobolev et al., 1989) coexists with almandine-rich garnet in a sample from Orapa (Jacob et al., 2011) and many other polycrystalline diamond aggregates are magnetic suggesting the presence of magnetite or even native iron (Jacob et al., 2004). Thus, compared to macroinclusions in diamonds, magnetite is overrepresented in polycrystalline diamond aggregates. 3.2. Oxidation state and clues to formation mechanisms Phases intergrown with polycrystalline diamonds show a similarly large range of redox conditions to macroinclusions in monocrystalline diamonds of over ca. 10 log units fO2. Inclusions consisting of native iron and iron carbide, for example, provide evidence for reducing conditions well below the iron–wüstite oxygen buffer (Jacob et al., 2004). The majority of polycrystalline diamonds, however, are compatible with more oxidizing conditions involving carbonates. The origin of the carbon-bearing fluid is debated. Some authors have argued for an entirely mantle-derived carbon fluid, which subsequently undergoes Rayleigh fractionation (Deines, 1980; Maruoka et al., 2004), while others infer a subducted carbon source (Mikhail et al., 2013). The δ13C values (n = 115) overlap with those for eclogitic and metamorphic diamonds and show a broad bimodal distribution from −1.3 to − 29‰ with major frequencies of occurrence at − 5‰ and at − 19‰ (Fig. 2; Table 1). The first peak at the typical mantle value of − 5‰ (Cartigny, 2005) argues for at least partial involvement of mantle-derived carbon. The second peak at −19‰, however, is distinct and typical only for polycrystalline diamond aggregates (Cartigny, 2005). Whether this mean value is indicative of the formation processes or source of carbon or if it is connected to biased sampling of the small database remains to be seen. Nitrogen concentrations and nitrogen isotope ratios in these diamonds cover a large range of 8 to 3635 ppm and δ15N − 6.1 to + 22.6‰ (Gautheron et al., 2005; Mikhail et al., 2013), but neither nitrogen concentrations nor δ15N values are coupled with δ13C values, apparently discounting coupled fractionation processes operating during diamond growth (Deines, 1980; Maruoka et al., 2004; Cartigny, 2005). Nitrogen aggregation states are across the whole spectrum from pure IaA to pure IaB (Mikhail et al., in press).

Fig. 2. Carbon isotopic compositions of polycrystalline diamonds from kimberlites (a), from UHPM terranes ((b): white = Kokchetav, gray = Erzgebirge) and of carbonado diamonds (c) compared to the typical range of monocrystalline diamonds with peridotitic inclusions (light gray band) and fibrous diamonds (dark gray band). Data for polycrystalline diamonds from Jacob et al. (2000), Maruoka et al. (2004), and Mikhail et al. (2013), for UHPM diamonds from Cartigny et al. (2001), Imamura et al. (2013), and Dobrzhinetskaya et al. (2010), for carbonado from Cartigny (2010) and for fibrous diamonds Cartigny (2005) and references therein.

The only study to date on radiogenic isotope systems of minerals included in polycrystalline diamonds reported unradiogenic Nd isotopic ratios of −15.9 to − 21.7 εNd, typical for ancient lithospheric material and argued for recent remobilization of this material to form polycrystalline diamonds (Jacob et al., 2000). The authors based their arguments on coupled chemical characteristics of diamonds and intergrown garnets as well as on trace element zonation in the garnets, which would have been erased by diffusion during expanded storage at mantle temperatures and pressures. Contrasting with this apparently recent formation of polycrystalline diamonds at the Venetia Mine are findings that some samples from an unknown South African locality show hightemperature deformation and annealing structures in Electron Backscattered Diffraction (EBSD) that apparently require long mantle residence times (Rubanova et al., 2012). In summary, polycrystalline diamonds form episodically by smallscale, rapidly occurring fluid-dominated redox processes within the subcratonic lithosphere. Intergrown micron-sized phases are most often hybrids, carrying evidence for a mixed derivation involving fluids and peridotite or eclogite solids and preserving inhomogeneity as a result of their rapid formation. Their trace element patterns often show strong enrichment in elements indicative for carbonatite metasomatism, thus arguing for a role for mainly oxidized fluids. However, very

26

Table 1 Summary of some major characteristics of diamonds discussed in this study compared to fibrous and monocrystalline diamonds. Nitrogen characteristics

Macro-inclusions (5–200 μm)

Micro-inclusions (b1 μm)

Carbon isotopes (δ13C values)

Polycrystalline aggregates

Subcratonic lithosphere (Group I kimberlites)

UHPM terranes

Continental collisions

Intergrowths mostly websteritic, but also eclogitic and peridotitic suites None (due to the small size of the diamonds)

Silicates, oxides, sulfides, metals, fluids. Carbonate present but rare Silicates, carbonates, halides, metals, fluids

−2 to −28‰, bimodal with means at −5‰ and −18‰ (see Fig. 2 for references) −10 to −27‰

Carbonado

Alluvial, inferred to be from the Earth's mantle

Rare, mostly intergrown phases

Silicates, carbonates, halides, metals, fluids

Fibrous

Subcratonic lithosphere

None

Fluids, silicates, carbonates, halides. All K-bearing Unstudied

Monocrystalline Mostly subcratonic lithosphere, some asthenospheric

Peridotitic, eclogitic, websteritic suites

0–4000 ppm, δ15N = −5.7 to +22.5‰ (Mikhail et al., 2013) IaA–IaB (Mikhail et al., in press) 0–11, 150 ppm, δ15N = −1.8 to +14‰ (Cartigny et al., 2001) Ib–IaA (De Corte et al., 1999) −5 to −30‰, most −22 to −30‰ 100–300 ppm (Cartigny, 2010) δ15N = −17 to +8‰ (Heaney et al., 2005), Ib–IaA (Cartigny, 2010) (Cartigny, 2010) −4 to −8‰ (Boyd et al., 1994) 0–3000 ppm, δ15N = −2 to −9‰, IaA, some Ib–IaA (Cartigny, 2010) +1 to −40‰, mean for peridotitic 0–3000 ppm, δ15N = +5 to −10‰ at −5‰ and wide distribution for (Cartigny, 2005), all aggregation states, but mostly IaA–IaB (Shirey et al., 2013) eclogitic (Stachel et al., 2009)

P and T

fO2 range of inclusions

Diamond stability field 6–9 GPa, 900– 1100°C

Mostly FMQ to IW buffers, some beyond Around FMQ/CCO buffers

Unknown

Mostly FMQ to IW

Diamond stability field 4.5–7.5 GPa, 950– 1350°°C (and above)

Around FMQ FMQ to IW but mostly around FMQ

reducing inclusions can also be found in these rocks and underline pronounced heterogeneity in their source regions upon formation as a possible scenario.

3.3. Insights from nano-inclusions

Taking these earlier studies to the nano-scale with the aid of FIB– TEM enabled a more complete and direct characterization of nanoinclusions and, most importantly, of coexisting fluids. Furthermore, micrometer-sized intergrowth phases could be compared to those shielded by diamond to evaluate whether the former were chemically overprinted or even formed later and are thus unrelated to the formation of the diamonds. A sample from the Orapa kimberlite (Jacob et al., 2011) showed that although major nano-phases included in the diamond were largely identical to the micro-sized phases intergrown with the diamond crystals, the latter were considerably altered at a late stage. Magnetite, pyrrhotite, omphacite and rutile were found as both nano- and microinclusions, thus confirming that this paragenesis is syngenetic to diamond. Magnetite, volumetrically the most prominent inclusion phase, displayed a distinctive polycrystalline and porous texture (Fig. 3a), which was interpreted as evidence for formation in the presence of the diamond-forming fluid. Although cavities still containing fluid were not observed in the magnetites, this may be due to the fact that

Fig. 3. TEM HAADF images of inclusions in polycrystalline diamond grains. (a) shows a magnetite included in polycrystalline diamond with porous structure, (b) intact fluid inclusion in a polycrystalline diamond (see Jacob et al., 2011 for an online animated version). Scale bars 200 nm.

D.E. Jacob et al. / Earth-Science Reviews 136 (2014) 21–35

Geological setting

D.E. Jacob et al. / Earth-Science Reviews 136 (2014) 21–35

a typical TEM foil represents only a very small volume of the total crystal or an approximate 2D section. Similarly, several open cavities were encountered in the diamonds, which were interpreted to have lost their fluid content upon FIB sectioning. These cavities contained rutile and/or pyrrhotite and omphacite and often non-stoichiometric phases that were interpreted as quench phases. One occurrence of an intact inclusion enabled identification of a fluid by its characteristic continuous changes in diffraction contrast due to density fluctuations caused by the electron beam (Fig. 3b). This fluid was associated with fine-grained pyrrhotite, a silicate phase rich in Fe, P, Mg, Al, Ca and K and a quench phase rich in non-stoichiometric amounts of Fe, P and Si. Amongst all inclusions only one ca. 10 nm large inclusion was found whose chemical composition suggested it to be a carbonate, although phase identification was not possible. The fluid found in the polycrystalline diamonds from Orapa is thus characterized by silicate rich in Fe, Mg, Al, Ca, K and P, while carbonate is rare and halide phases are absent (Fig. 4). The majority of polycrystalline diamonds bear evidence for having precipitated from an oxidized C–H–O medium, while only a minor proportion contain very reduced inclusions. Experimental studies suggest

Fig. 4. Demonstrates the characteristic microstructure of carbonado inside a diamond grain from the Central African Republic (CAR2#1414, Kaminsky et al., 2013). (a) TEM bright field image of several diamond grains with irregular grain boundaries. Some grains show straight dislocation lines. (b) and (c) TEM dark field images taken at the 111 reflex and displaying three individual diamond grains (copy of a film negative, therefore, 111 crystal faces are dark in the image). Note the irregular grain boundaries and the locally changing inclination of the grain boundary planes.

27

that at typical upper mantle pressures and temperatures the C–H–O medium from which these diamonds formed is subcritical, being either a carbonatitic melt under oxidizing conditions (Dasgupta et al., 2004) or a CH4-rich fluid at low oxygen fugacity (Stalder et al., 2001). A halide component that represents one of the end-members identified in diamond fluids worldwide (Klein-BenDavid et al., 2004) has not yet been encountered. 4. Carbonado Carbonado is a smooth-surfaced fine-crystalline (1–10 μm) and highly porous diamond aggregate, mostly black in color, which lacks most of the typical macroinclusions found in monocrystalline diamonds (Heaney et al., 2005; McCall, 2009). Their origin is still under debate with models ranging between crustal derivation, genesis in the upper Earth's mantle and an extraterrestrial origin (see Heaney et al., 2005; McCall, 2009; Cartigny, 2010; Haggerty, 2014 for detailed summaries). Characteristic features of carbonado are inclusions of crustal minerals, such as graphite, quartz, rutile, florencite (a water-bearing aluminium phosphate), and clay minerals that occur along partially open grain boundaries and in pores. Carbonado (sensu stricto) is found in alluvial deposits in Brazil and in Central Africa, and its primary source is unknown; there is a high chance that some of the inclusions in carbonado are in fact epigenetic, and are thus unable to provide any insights into the origin of this enigmatic diamond (Heaney et al., 2005). Modern submicron methods now provide the means to test this by comparing the nanoinclusion suite in the carbonado with the chemistry of the macrocinclusions. Carbonado is similar to polycrystalline diamond aggregates from kimberlites and the latter are sometimes mistakenly classified as carbonados. Both have a porous structure and random crystallographic orientation of the individual diamond grains within their polycrystalline aggregates (Ishibashi et al., 2012; Rubanova et al., 2012; Kaminsky et al., 2013). Furthermore, some polycrystalline diamond aggregates can show equally small grain sizes and the microporphyritic textures that are typical of carbonado (see Heaney et al., 2005; Haggerty, 2014 for reviews), and carbonado in turn can display coarse-grained microtexture (e.g. Trueb and Butterman, 1969). However, carbonado has much higher porosity (Vicenzi et al., 2003) and displays a very distinct and unusual nanostructure recognized with TEM, displaying column-like diamond grains that form linear and zigzag patterns with low-angle and even partially open grain boundaries. A characteristic feature of individual diamond grains in carbonado is the presence of uneven and rough grain boundary planes as displayed in Fig. 4 (Rondeau et al., 2008), very unlike the microstructure of polycrystalline diamonds from kimberlites. In addition, carbonado occurs in remarkably large sizes, much bigger than those reported for polycrystalline diamond aggregates from kimberlites. Svisero (1995) pointed out that ten of the eleven largest diamonds reported from Brazil are carbonados, the largest weighing 3167 carats. Isotopic dating using Pb isotopes of bulk diamonds (Ozima and Tatsumoto, 1997) and Pb–Pb SHRIMP dating of quartz and rutile inclusions (Sano et al., 2002) of both Brazilian and Central African carbonados yielded Archean ages, which served as a basis to hypothesize that both populations are derived from the same eroded source (McCall, 2009). Carbon isotopes define a distinct range of δ13C = − 23.5 to −33‰ (Table 1; but note two outliers at higher, typical mantle values, Fig. 2) which led a number of authors to argue for a crustal origin involving organic carbon (Smith and Dawson, 1985) or irradiated carburanium (Kaminsky, 1987, 1991; Ozima et al., 1991). Carbonados are rich in hydrogen (Garai et al., 2006), contain single nitrogen impurities (type Ib–IaA; Garai, 2012) and polycyclic aromatic hydrocarbons (Garai et al., 2006), which are otherwise rarely encountered in terrestrial diamond, but do occur in presolar diamond grains from meteorites. These characteristics and the apparent lack of primary terrestrial silicate

28

D.E. Jacob et al. / Earth-Science Reviews 136 (2014) 21–35

inclusions led to speculations of an extraterrestrial origin of these enigmatic diamonds (Haggerty, 1999, 2014). FIB–TEM analyses provided the means to search for smaller silicate or oxide inclusions inside individual diamond crystals in order to exclude any potentially epigenetic material (De Souza Martins, 2006; Sautter et al., 2011; Kaminsky et al., 2013). The inclusions found with this method were generally relatively small (in the hundred nanometer range). Phases comprised almandine–pyrope garnet, augite, apatite (including fluorapatite), phlogopite, SiO2, Ca–Mg–Sr–and Ca–Bacarbonates, halides (sylvite and bismocolite BiOCl), native Ni and metal alloys (Fe–Ni, Cr–Fe–Mn, and Pb–As–Mo), oxides (FeO, Fe–Sn– O, TiO2, SnO2, FeTiO3 and PbO2), and Fe–sulfides, as well as fluid inclusions. These results are similar to an earlier TEM study of syngenetic inclusions in carbonado by De et al. (1998) who reported native metals and alloys in addition to SiC, sylvite, calcite, smithsonite (ZnCO3) and Mg–Al silicates. The fluid inclusions are predominantly silicate–carbonate compositions; halide-rich fluids were also observed, but more rarely. Molecular H2O, detected by FTIR, was found in a carbonado from Central Africa (Ishibashi et al., 2012). The nanoinclusion suites of carbonados thus show much more similarity to findings in mantle-derived diamonds (Klein-BenDavid et al., 2007; Cartigny, 2010; Jacob et al., 2011) and metamorphic diamonds from ultrahigh pressure terranes (e.g. Dobrzhinetskaya, 2012; Dobrzhinetskaya et al., 2013) than the larger inclusions. In contrast, low-pressure crustal phases identified in earlier studies, such as florencite, were not encountered, pointing to an epigenetic origin of these minerals. Therefore, the solid phase inclusions reported from carbonados so far (while maybe not yet being representative) do not support the extraterrestrial origin suggested by Haggerty (1999, 2014) and Garai et al. (2006). 5. Polycrystalline diamonds from ultrahigh-pressure terranes Polycrystalline diamonds (Fig. 5) were identified recently within felsic quartz–feldspathic gneisses of the Erzgebirge massif, Germany (Dobrzhinetskaya et al., 2013) where they occur in addition to abundant monocrystalline microdiamonds. Massonne (2003, 2006) showed that these rocks were subjected to UHP metamorphism during continent– continent collision at T = 1100 °C and P = 7–8 GPa, which corresponds to depths of ~200–210 km. While these represent one of the most recent finds, polycrystalline diamonds have been known for more than a decade from rocks of the Kokchetav massif as so-called “S- and R-type” diamonds (Ishida et al., 2003; Ogasawara and Aoki, 2005) or “overgrown” diamonds (Sitnikova and Shatsky, 2009). These diamond-bearing metasedimentary rocks of the Kokchetav UHP massif are thought to have formed under similar conditions as those in the Erzgebirge, namely in a deep subduction zone at T = 980–1200 °C and P = 6–9 GPa, corresponding to depths of ca. 160–250 km (Dobrzhinetskaya et al., 2005; Ogasawara and Aoki, 2005). 5.1. Nanoinclusions in the Erzgebirge polycrystalline diamonds The diamonds included in felsic gneisses contain abundant nanoinclusions of 10–200 nm in size. These can be differentiated into two suites, based on their occurrence within the diamonds: along the grain boundaries of microcrystalline diamond aggregates (suite 1), or on the dislocations that are widely developed inside diamond crystals (suite 2). Suite 1 includes crystalline and amorphous nanoinclusions of SiO2, rare CaCO3, BaCO3 and KAlSi3O8 as well as abundant amorphous quench phases of an approximate composition of NaSO4 and KCl. All phases contain traces of Pb, K, Ca, Cr, Fe, Al, P, Cl, S, Ti, and Zr in different combinations. Structures of crystalline phases of SiO2, CaCO3 and KAlSiO8 could not be determined due to their very small size (~10 nm). Suite 2 is composed of rare crystalline nanoinclusions of ZrSiO4, NaSO4, BaSO4 and countless b 1–10 nm amorphous inclusions, mostly SiO2 and

Fig. 5. Typical diamonds from ultrahigh pressure metamorphic terranes: (A) diamond inclusion in zircon from the Erzgebirge massif, Germany (modified from Dobrzhinetskaya et al., 2007), (B) diamond from the garnet–biotite gneisses of Kokchetav massif, Kazakhstan (collection Dobrzhinetskaya).

KAlSi3O8 that are associated with perforated cavities. The amorphous matter is assumed to be the residue of evaporated remnants of former fluid inclusions that were perforated by either the FIB sputtering during foil preparation or the TEM electron-beam sputtering upon focussing the beam to form a very small probe of only several nanometers in diameter. The process of fluid bubble penetration by electron sputtering has been observed several times during numerous TEM study sessions and detailed descriptions of this effect can be found elsewhere (e.g. Dobrzhinetskaya et al., 2005). EDX analysis combined with TEM observations showed that the suite of amorphous inclusions contains traces of Cl, S, P, Al, K, Fe, As, Hg, Mo, Co and Ti mixed in different proportions and combinations (Dobrzhinetskaya et al., 2013). The histogram in Fig. 6a comprises the combined compositional data for all suites of fluid inclusions in the Erzgebirge gneisses (for details on the data reduction see Dobrzhinetskaya et al., 2013). Data analysis suggests that the integral composition of the Erzgebirge diamond-forming fluids correspond to a high-density C–O–H fluid, a hydrous-silicic fluid rich in Al and K and a hydrous-saline fluid rich in Cl, K and Na. The overall composition of these fluids is similar to that of fluids in fibrous diamonds plotted in Fig. 9. Traces of carbonate, sulfate, phosphate and metallic cations such as Fe, Mg, Ti, Cr, Co, Mo, Zn, Zr, Pb, As and Hg reflect the local geochemical diversity of the metasedimentary rocks which host these diamonds.

D.E. Jacob et al. / Earth-Science Reviews 136 (2014) 21–35

29

The nanoinclusions are situated in cavities along diamond–diamond interfaces or even create films of amorphous matter between individual grains of the polycrystalline aggregates, but they are also observed inside individual diamond crystals. Most of these crystalline nanoinclusions are oxides (SiO2, TiO2, FexOy, Cr2O3, ThxOy), or silicates, e.g. KAlSi3O8, and rare inclusions of ZrSiO4, BaSO4, and CaCO3–aragonite occur as well (Fig. 6b). The inclusions contain variable amounts of K, Na, P, S, Pb, Zn, Nb, Al, Co, Mo, P, and Cl (e.g. de Corte et al., 1999; Dobrzhinetskaya et al., 2005). Amorphous nanoinclusions associated with cavities have similarly diverse compositions and exhibit large variations of trace elements. They are considered to be non-stoichiometric “quench” products of fluid inclusions that were opened during preparation or analysis. Diamonds hosted by felsic rocks have abundant Si- and Al-bearing inclusions and contain rare carbonate inclusions, whereas diamonds from marble and calcareous-silicate gneisses are mostly rich in carbonate inclusions and only rarely contain Si-, or Al-bearing minerals. This compositional diversity of the crystalline and amorphous nanoinclusions shown by the Kokchetav polycrystalline diamonds in all rock types clearly indicates that in general the fluid from which diamonds crystallized has a local origin in UHP terranes. 5.3. Diamond-forming fluid media and oxidation state

Fig. 6. Histograms showing average (n) elemental composition of fluids in diamonds measured as raw counts per second TEM-EDX data: (a) garnet–quartz–feldspathic gneisses of the Erzgebirge massif, Germany (n = 40 inclusions), (b) garnet–biotite gneisses of the Kokchetav massif, Kazakhstan (n = 50 inclusions), (c) marbles of the Kokchetav massif, Kazakhstan (n = 45 inclusions).

5.2. Nanoinclusions in the Kokchetav polycrystalline diamonds In the UHP Kokchetav massif, polycrystalline diamonds are found as inclusions in diopside and phlogopite from metacarbonate rocks (marbles and calcareous rocks) and in feldspathic gneisses. 5.2.1. Marbles and calcareous-silicate gneisses Diamonds from these carbonaceous rocks contain abundant crystalline and amorphous nanoinclusions, and their chemical composition clearly indicates that the source for the diamonds is locally derived. They have distinct morphologies and are known as “S- and R”-type diamonds (Ishida et al., 2003). The nanoinclusions are aragonite, magnesite and rarely also TiO2. No SiO2 inclusions nor any Al–Si-bearing minerals nor their quench products have been observed. The crystalline nanophases and the amorphous matter lining the walls of small cavities (remnants of decrepitated fluid inclusions) contain a diverse spectrum of lithophile and siderophile trace elements such as Co, Cu, Zn, Mn, Pb, K, Ba, Fe, Sr, Th, Zn, Cl, S and P (Fig. 6c). 5.2.2. Feldspathic gneisses Monocrystalline and polycrystalline diamonds from feldspathic gneisses of the Kokchetav massif are similar to those from felsic gneisses of the Erzgebirge massif in that they also contain two suites of nanoinclusions of which one is crystalline while the other is amorphous.

The concept of UHPM diamond crystallization from a supercritical C–O–H fluid in the diamond stability field is agreed upon by a majority of researchers (see review papers: Dobrzhinetskaya, 2012; Schertl and Sobolev, 2013). This has received considerable support from detailed analytical studies of nanoinclusions in natural UHPM diamonds and from experiments, which successfully synthesized diamond from different carbon species in the presence of H2O at pressures and temperatures corresponding to the diamond stability field (e.g. Dobrzhinetskaya et al., 2009). The diamond-hosting sedimentary rocks are rich in H2O and diamonds themselves contain nanoinclusions that are enriched with almost all elements available in the host rocks. These observations show that the origin of the diamond-forming fluid is local. The presence of carbonate inclusions in polycrystalline diamonds and in single diamond crystals from both the Erzgebirge and the Kokchetav UHPM terranes showed that the oxygen fugacity must have been close to the CCO and the FMQ oxygen buffers (logfO2 = − 6 in Fig. 7; e.g. Dobrzhinetskaya, 2012; Dobrzhinetskaya et al., 2013). 5.4. Source of carbon and nitrogen aggregation Polycrystalline and single crystal diamonds from UHPM terranes originate from the same fluid source (e.g., Dobrzhinetskaya, 2012; Dobrzhinetskaya et al., 2013). The diamonds from the Kokchetav massif, Kazakhstan have δ13C = −10.57‰ and N contents of up to 11,385 ppm (Cartigny et al., 2001), whereas diamonds from dolomitic marble are characterized by δ13C = − 10.19‰ and N = 2,762 ppm (Table 1, Fig. 8). These δ13C values are interpreted to be a mixture between crustal and mantle carbon reservoirs (Cartigny et al., 2001; Ogasawara and Aoki, 2005), which is supported by He and Ne isotope studies in these diamonds (Sumino et al., 2011). The Erzgebirge microdiamonds (Fig. 8) are characterized by significantly lighter carbon in comparison with the Kokchetav terrane, ranging in δ13C between − 17 to − 27‰ and N = 100–4,647 ppm (Dobrzhinetskaya et al., 2010). This carbon isotopic variation suggests that the Erzgebirge diamonds, in contrast to the Kokchetav suite, crystallized from an entirely crustal carbon reservoir, although whether that carbon was organic or inorganic may be debated. The wide variation of nitrogen abundances was explained by an inhomogeneous distribution of N-bearing fluid inclusions, or by the presence of Nbearing mica such as phengite, which is observed in association with microdiamonds or as inclusions (e.g. Stöckhert et al., 2001;

30

D.E. Jacob et al. / Earth-Science Reviews 136 (2014) 21–35

residence times in the mantle (e.g. Jones et al., 1992; Cartigny et al., 2001). 5.5. Evidence from fluids for crust–mantle interaction

Fig. 7. Composition and oxidation state of C–O–H fluids in equilibrium with diamond for P = 5 GPa, T = 1400 K (modified after Taylor, 1990; Odling, 1989). The heavy curve labeled “diamond saturation line” delineates the stability field of diamond and coexisting fluid at this P and T. The oxidized fluid-only region lies above, and the reduced fluid-only region is to the lower left of the saturation curve. They are separated by a diamond + fluid field that extends to very water-rich fluid compositions labeled “water maximum”. Positions of the magnetite–wüstite (MW) and the iron–wüstite (IW) oxygen buffers are indicated by dashed lines. Reducing fluids coexisting with diamond at fO2 = IW and below consist of mixtures of CH4 N H2O N H2 and C2H6 progressing to mixtures of CH4 N C2H6 N H2 at fO2 well below IW.

Dobrzhinetskaya et al., 2010). Kinetic effects upon diamond crystallization may also have played a role. Nitrogen aggregation characteristics show that all known diamonds from UHPM terranes belong to the type 1b–1aA, suggesting a short residence time (ca. 5 Ma) at high temperature conditions (~ 900– 1100 °C) (e.g. Cartigny et al., 2001; Dobrzhinetskaya et al., 2006). This distinguishes them from many kimberlitic monocrystalline diamonds, which belong to type 1aAB and are interpreted to have had longer

Fig. 8. Comparative diagram of nitrogen content versus δ13C in diamonds from ultrahigh pressure metamorphic terranes, polycrystalline diamonds from kimberlite (black spots, Jacob et al. unpublished SIMS data, open circles, Mikhail et al., 2013) and worldwide monocrystalline diamonds from kimberlitic and lamproitic sources (Field 3). Fields 1–3 represent Kokchetav massif, Kazakhstan and are adopted from Cartigny et al. (2001, 2003). Field 1 (black squares): diamonds from grt pyroxenites; Field 2 (gray squares): alluvial microdiamonds, and diamonds from marbles (open squares). Field 4 is from Dobrzhinetskaya et al. (2010) and represents Erzgebirge microdiamonds included in garnets from quartz–feldspathic gneisses.

Though most observations have pointed to a crustal origin there are several lines of evidence that the diamond-bearing fluid in at least some UHPM terranes is a product of crust–mantle interaction. One of them is the presence of eskolaite (Cr2O3) inclusions that contain traces of Al, Si, P, Ni, and Fe in Kokchetav diamonds from felsic gneisses (Dobrzhinetskaya et al., 2003). Dobrzhinetskaya and Wirth (2009) postulated that Ni- and Fe-contents as well as the Cr-content of eskolaite were probably derived from a mantle wedge reservoir. Noble gas studies (Sumino et al., 2011) supported this proposition. These authors found that the Kokchetav diamonds contain primordial He and Ne, suggesting that a deep mantle plume affected the subducted continental slab. Though no evidence of crust–mantle interaction has yet been found in the Erzgebirge diamonds (potentially mostly because of the lack of He–Ne isotopes studies), the mechanism of diamond formation from both UHPM terranes is characterized by active “crustal” fluid migration through the subduction zones. In Kokchetav, the process was complicated by the interaction of this local, crust-derived fluid with mantle fluids. 6. Discussion 6.1. Diamond fluids in the Earth's mantle While mantle and crustal environments for the formation of polycrystalline diamonds appear strikingly different, all of these diamond crystals contain fluid inclusions, thus emphasizing the prominent role of fluids in diamond formation in general (Navon et al., 1988). Much of what we know about diamond-forming fluids comes from intensive studies of fibrous and cloudy diamonds from the Earth's mantle (Schrauder and Navon, 1994; Izraeli et al., 2001; Klein-BenDavid et al., 2004; Zegdenizov et al., 2004; Rege et al., 2005; Tomlinson et al., 2006, 2009; Logvinova et al., 2008; Weiss et al., 2009). Fibrous diamonds form in the same environments in the Earth's mantle as “normal” monocrystalline diamonds, but represent an extreme end of this process, providing a different perspective. Although they are only a minor component of the diamond population in Group I kimberlites (Boyd et al., 1994), fibrous stones are ideal study objects for nanomethods, because they can contain thousands of fluid inclusions. For the first time, direct observation of diamond fluids was possible and these studies significantly extended our knowledge, which had previously been based solely upon indirect evidence from carbon and nitrogen and their isotopes within the diamond itself. Many studies worldwide revealed that the fluids trapped in these stones range compositionally between three end-members, namely a carbonatitic end-member rich in Ca, Mg, Fe, K and CO2, a hydroussilicic end-member rich in Si, Al, K and H2O, and a hydrous-saline fluid rich in Cl, K, Na and H2O (Fig. 9; Klein-BenDavid et al., 2004). Most recently, high Mg/Ca carbonatitic fluid inclusions were reported (Klein-BenDavid et al., 2008; Kopylova et al., 2010; Weiss et al., 2011), which perhaps represent an additional end-member for these fluids, situated deeper in the lithosphere. A characteristic feature of these fluids is a pronounced miscibility gap between the silicate-dominated and the halide-rich end-members at pressures and temperatures relevant to diamond formation in the Earth's mantle, described from case studies (Izraeli et al., 2001; KleinBenDavid et al., 2004, 2006, 2007) and substantiated experimentally (Safonov et al., 2007). Most fibrous diamond fluid compositions fall along mixing lines between either the carbonatitic-hydrous saline or the carbonatitic-hydrous silicic end-members for which contradictory explanations exist. Silicates are the liquidus phases in carbonatite melts (Safonov et al., 2007; Litasov and Ohtani, 2009); therefore diamond crystallization upon successive cooling should produce a

D.E. Jacob et al. / Earth-Science Reviews 136 (2014) 21–35

31

6.3. Oxidation conditions recorded by polycrystalline diamonds: a matter of scale

Fig. 9. Fluid end-member compositions derived from studies on fibrous diamonds (bold dashed line: Izraeli et al., 2001; Klein-BenDavid et al., 2004) and experimental results of Safonov et al. (2007, solid line — 1600 °C, dashed line — 1500 °C, stippled line — 1400 °C) showing the compositional range of fluid inclusions in diamonds. Note the large miscibility gap between the saline and the siliceous end-member in this system. Overlain are the general range of inclusion compositions for UHP metamorphic diamond inclusions from metacarbonate rocks (Kokchetav Massif — light gray field) and feldspathic gneisses (Kokchetav Massif and Erzgebirge — dark gray fields) and the approximate composition for the polycrystalline diamond aggregate from Orapa (silicate inclusions as dark gray field connected with a tie line to the single carbonatic inclusion (star)).

progression from silicate-rich to carbonate-rich inclusions. However, this contradicts other models (Schrauder and Navon, 1994) that argued for fractional crystallization of a carbonatitic liquid, which would result in an opposite trend in inclusion chemistry. 6.2. Diamond fluids in polycrystalline diamonds The fluid compositions from polycrystalline diamonds from Orapa plot towards the hydrous-silicic end-member of these compositions and lack the halide component (Fig. 4). Potassium is present in many of the nano-inclusions in this sample, but also Ca, Mg and particularly Fe, which are more typical for the carbonate-dominated fluids in fibrous and cloudy diamonds. Although it is too early to draw general conclusions for the polycrystalline diamonds from kimberlites, these observations show that this particular fluid bears general similarities with those described from the fibrous and cloudy diamonds in representing a composition close to the silicate-dominated end-member, although different in divalent cation composition. In contrast, fluid inclusions in the carbonado suite studied here (not shown) cover the broad spectrum of fluid compositions depicted in Fig. 9, including the apparent lack of compositions plotting into the miscibility gap. The hypothesis that worldwide diamond fluids comprise silicates, water, carbonate and halogens also holds for polycrystalline diamonds from crustal rocks. Diamonds from metacarbonates of the Kokchetav UHP terrane contain hydrous-saline and carbonatic inclusions, while silicate-bearing fluid inclusions are rare. Felsic gneisses from the Kokchetav and Erzgebirge massifs, on the other hand, have inclusions in diamond that are of hydrous-silicic or hydrous-saline compositions. The exact composition for each individual nano-inclusion in diamond from the felsic gneisses was difficult to assess, and the exact compositional range distribution between the individual end-members remains unclear as of yet. The large miscibility gap at 5 GPa in Fig. 9 persists towards lower temperatures and pressures (Suk, 2001); however, the relevant experimental studies did not include water. Safonov et al. (2007) and Litasov and Ohtani (2009) suggested that the presence of water would significantly increase the miscibility of Si- and Cl-bearing liquids.

Macro-inclusions indicate that diamond has a large stability field stretching over more than 10 log units in oxygen fugacity (fO2), from conditions of about six units below the iron–wüstite oxygen buffer, where moissanite (SiC) is stable (Moore and Gurney, 1989; Ulmer et al., 1998) to distinctly oxidized conditions where carbonate inclusions are present, at an oxygen fugacity of about 4 log units above the iron–wüstite equilibrium (Fig. 7). Depending on the exact bulk composition, even higher fO2 is suggested by rare pure CO2 inclusions in diamond (Schrauder and Navon, 1994; Chinn, 1995). An important insight from the application of FIB–TEM methods to diamond was that in some cases the nano- and micro-inclusion suite within a single diamond can trace fO2 conditions covering the entire 10 log units fO2 of the diamond stability field (Klein-BenDavid et al., 2007). Most microinclusions (fluid and solid alike) record oxidized conditions; however, there is also a pronounced record of heterogeneous of fO2 conditions from microinclusions. Similarly, macro-inclusion suites also suggest that most of their host diamonds formed by reduction of an oxidized medium (Stachel and Harris, 2008), but inclusion evidence for very reducing conditions can be found, too. However, this skewed distribution towards oxidized carbon sources is not reflected by the inclusions in polycrystalline diamonds from the Earth's mantle. Here, inclusions that reflect very reducing conditions appear to be just as common as those formed under oxidizing conditions. Native metals and alloys (e.g. Fe-Ni alloy, Sn) are found in carbonados but so are carbonates (Kaminsky et al., 2013). Carbonates and magnetite or cohenite and native iron occur in polycrystalline diamonds from kimberlites (Jacob et al., 2004, 2011). The large range of oxygen fugacity recorded by polycrystalline diamonds suggests a prominent role of redox gradients and transient, small-scale equilibria in their formation. Compositions of mantle volatiles in equilibrium with diamond vary at the relevant pressures and temperatures depending on oxygen fugacity. While they consist of pure CO2 at oxidizing conditions, they comprise mixtures of CH4, H2O and H2 close to the iron–wüstite buffer and below (Fig. 7; Saxena, 1989; Taylor, 1990; Sokol et al., 2009; Zhang and Duan, 2009). Under oxidizing conditions the CO2-rich volatiles form very mobile carbonatitic melts that provide efficient transport media for carbon in the mantle (Green and Wallace, 1988). In a reducing environment, the main carriers of carbon are methanerich fluids (Jakobsson and Holloway, 1986; Sokol et al., 2009; Foley, 2011), because reducing silicic melts dissolve b100 to 2000 ppm carbon (Taylor and Green, 1987; Ardia et al., 2013). Reduction of carbonatitic melts or oxidation of methane-bearing fluids results in the formation of carbon-saturated aqueous fluids (Fig. 7) from which diamond can precipitate across a wide pressure and temperature range. For reducing conditions more than 3 orders of magnitude below the water maximum in Fig. 7, fO2 changes upon diamond precipitation occur from a fluid with near-constant C/H ratio, which facilitates the precipitation of suites of inclusions with variable oxidation states. Here, rapid diamond crystallization may occur with essentially unchanging fluid composition if the fluid throughput is high. The rapid change in fluid composition 2 orders of magnitude below the water maximum would be associated with rapid diamond crystallization and entrapment of inclusions slightly above iron–wüstite. The situation is different for oxidizing conditions above the water maximum, where a fluid in equilibrium with diamond has quite constant fO2, but evolves from CO2- to H2O-rich compositions. Thus, inclusion suites in diamonds formed on the oxidized side of this diagram will show a more uniform oxidation state. Particularly these latter conditions apply to UHPM polycrystalline diamonds. Neither native metals nor alloys and carbides have been found in these diamonds; instead, they are dominated by inclusions that indicate high oxygen fugacity (e.g. carbonates), close to the FMQ

32

D.E. Jacob et al. / Earth-Science Reviews 136 (2014) 21–35

oxygen buffer. As predicted by the C–H–O speciation diagram in Fig. 7, fO2 variability is minimal in inclusion suites from these environments. The recent discovery by Bali et al. (2013) that water and hydrogen are likely to be immiscible under the conditions of the mantle geotherm may help to provide an explanation for the extreme small-scale heterogeneities in fO2 conditions represented by the inclusions in polycrystalline diamonds. During precipitation of diamond along the saturation curve (Fig. 7), fluids become enriched in H2O, so that hydrogen is likely to be released: the differential mobility of water and hydrogen (Bali et al., 2013) will lead to the production of very reducing environments on the scale of nanometers to centimeters, promoting the crystallization of nanophases such as metals, carbides and nitrides. 6.4. Chemical environments for diamond formation Decades of studies of carbon and nitrogen in diamond have underlined the complications associated with deciphering the origins of its ingredients (for recent reviews see Stachel and Harris, 2008; Shirey et al., 2013 and references therein). As carbon is subjected to efficient deep recycling on Earth (Javoy et al., 1982), there is no question that sources of carbon in diamond are multiple, comprising both mantle and subducted crustal components. Source characteristics for carbon and nitrogen, however, are often encrypted by mixing and fractionation processes upon subduction and at mantle temperatures (Deines, 1980). It is likely that the silicic, carbonatitic and saline compositions of fluids included in diamonds are subjected to similarly complex processes, which complicate interpretations. Silicate and carbonatitic liquids have long been recognized as major metasomatic agents in the mantle. In comparison, the roles of halides and halogens in the Earth's mantle are much more enigmatic and as of yet data on the deep halogen cycle is virtually non-existent (Izraeli et al., 2001; Kamenetsky et al., 2004; Litasov and Ohtani, 2009). Volatiles, even when present in only small amounts, have large effects on the melting points of peridotite and ultramafic rocks in the mantle (Green, 1973; Eggler, 1976; Wyllie, 1978; Taylor and Green, 1988; Dasgupta and Hirschmann, 2006; Foley, 2011). Precipitation of diamond from a C–H–O fluid via redox reaction rapidly increases the water activity of the fluid (Fig. 7), which may promote melting of the surrounding mantle material. This process couples the precipitation of diamond with melting caused by the increase in water activity of the fluid. It occurs in both reduced and oxidized conditions, but will be particularly effective in oxidized conditions where the depression of melting temperatures in the presence of both H2O and CO2 is stronger (Foley et al., 2009). This process results in the generation of silicate partial melts simultaneous with diamond precipitation that mix with the diamond fluid and, thus, are likely to be included into the growing diamonds. It is therefore hypothesized here that syngenetic solid phase inclusions in diamonds generally represent mixtures of components carried in the fluid and material produced by partial melting of surrounding rocks at the site of diamond precipitation on a very local scale. The composition of a given inclusion suite would then depend on both the chemical load of the carbon-bearing fluid which may vary depending on its oxidation state and on the degree of partial melting of the local rock. Higher degrees of melting lead to inclusion suites dominated by silicate material typical for the local environment, thus with compositions representing the Earth's mantle (eclogitic–peridotitic). Lower degrees of melting of local rock generate inclusions that are more representative for the chemical composition of the original C–H–O fluid and its solutes. However, in either case, the redox conditions recorded by the syngenetic inclusion suite are a result of the diamond precipitation process and bear little information on the original oxygen fugacity of the diamond's environment. For the UHPM polycrystalline diamonds, the marble host rocks present potential near-infinite reservoirs of carbon. Indeed, the major and trace element chemistry of the solid nano-inclusions precipitated

from the carbon-bearing fluids, in both the marbles and the felsic gneisses corresponds closely to those of their respective hosts. Major fluid species identified by FTIR spectroscopy are carbonate and water in inclusions from marbles and felsic gneisses alike (de Corte et al., 1999; Kikuchi et al., 2005; Dobrzhinetskaya et al., 2006; Dobrzhinetskaya and Wirth, 2009; Sitnikova and Shatsky, 2009). In order to explain the differences in diamond grade between directly neighboring lithologies, however, cm-scale heterogeneities in fluid species, e.g. in XCO2, must have existed during peak metamorphism (Ogasawara et al., 2000; Ogasawara and Aoki, 2005). Nevertheless, despite the majority of evidence pointing to local sources for the carbon-bearing fluids in these UHP terranes, noble gas isotopic studies (Sumino et al., 2011) suggested the involvement of an exotic, mantle-derived component at least for the Kokchetav massif, which is the only UHPM locality for which a noble gas study was carried out to date. A major challenge for the future will be to carefully decipher these sources of mixed information.

7. Conclusions State-of-the-art analytical techniques with high spatial resolution have deepened our understanding of natural diamond and its inclusions, while at the same time they highlight the complexity of the processes at work. Taking microanalysis to the submicron scale in diamond research up to now benefited in particular from TEM analysis using FIB prepared electron-transparent samples. Its application to carbonado and to polycrystalline diamond from kimberlite allowed syngenetic inclusions to be differentiated from epigenetic material, which proved particularly critical to test models for the origin of carbonado, finally resolving the Earth's mantle as its origin. In situ sampling of diamond fluids revealed a large heterogeneity in redox-conditions and fluid chemical compositions at small scale, which is not reflected in the macro-inclusion suite. However, despite originating from such chemically different environments as the mantle and the crust, fluids from polycrystalline diamonds have compositions that conform with the fluid endmembers established by studies on fibrous diamonds, thus suggesting a universally important role of a limited number of basic ingredients, namely carbonates, silicates, halides and water. Strong redox gradients reflected by the micro-inclusions indicate diamond precipitation via small-scale, ephemeral redox processes driven by the contrasting oxidation states of fluids and their depositional environment. The susceptibility of the melting point of Earth's mantle rocks to the presence of even small amounts of volatiles promotes melting simultaneous to diamond precipitation, further enhanced by the changing of fluid composition towards higher water activity. This creates a chemically heterogeneous environment, in which diamonds and their inclusions are precipitated via redox-freezing processes. Recent experimental work implies that the generation of ephemeral small-scale redox gradients may be even further enhanced by the immiscibility of water and hydrogen in the diamond fluids at these pressures and temperatures (Bali et al., 2013). The submicron inclusion suites are therefore products of redox reactions and mixing of components from the C–H–O fluid, which can be foreign, and those derived locally from the surrounding mantle rocks. Simple mass balance considerations predict that in instances where partial melting of the host rock dominates, the fluid component is occluded in the resulting inclusion chemistry. In these cases the particular inclusion only bears information on the depositional environment rather than on the carbon-bearing medium. In contrast to polycrystalline diamonds from the mantle, those from UHPM terranes carry evidence for homogeneously oxidizing formation conditions and carry a strong local signature in their submicron inclusion suite. Still, rare gas isotopes suggest the seemingly occluded involvement of a mantle fluid in their formation, suggesting that a

D.E. Jacob et al. / Earth-Science Reviews 136 (2014) 21–35

foreign fluid component may merely be masked by dilution in the major and trace element compositions. Development in modern microanalytical techniques allows in situ analysis of diamonds and their inclusions spanning roughly five orders of magnitude in spatial resolution. Information derived from inclusion suites at either end of the size range complements each other. Macroinclusions, for once, give critical insight into the chemical environment of diamond formation as well as the pressure and temperature range of diamond growth. On the other hand, micro-inclusions at the lower end of the size distribution deliver direct information on the diamond growth medium represented by included minerals, quench phases and fluid, which cannot be derived from macroinclusions alone. Acknowledgments DJ gratefully acknowledges financial support from DFG grant JA781/ 9-1 (Heisenberg Professur), the Earth System Science Research Centre at Johannes-Gutenberg University, Mainz and Macquarie University, Sydney. De Beers is thanked for the accessibility of suitable sample material. LD acknowledges support from US National Science Foundation (grants EAR-0408505, EAR-1118796) and the Lab Fee Research Award (No. 09-LR-05-116946-DOBL). Anja Schreiber (GFZ Potsdam) is thanked for skilful FIB sample preparation. We are grateful to Oded Navon and Andy Moore who reviewed an earlier version of this ms. This is contribution 460 from the ARC Centre of Excellence for Core to Crust Fluid Systems (http://www.ccfs.mq.edu.au). References Ardia, P., Hirschmann, M.M., Withers, A.C., Stanley, B.D., 2013. Solubility of CH4 in a synthetic basaltic melt, with applications to atmosphere-magma ocean-core partitioning of volatiles and to the evolution of the Martian atmosphere. Geochim. Cosmochim. Acta 114, 52–71. Aulbach, S., Stachel, T., Viljoen, K.S., Brey, G.P., Harris, J.W., 2002. Eclogitic and websteritic diamond sources beneath the Limpopo Belt — is slab-melting the link? Contrib. Mineral. Petrol. 114, 331–348. Bali, E., Audetat, A., Keppler, H., 2013. Water and hydrogen are immiscible in Earth's mantle. Nature 495, 220–223. Boyd, S.R., Pineau, F., Javoy, M., 1994. Modelling the growth of natural diamonds. Chem. Geol. 116, 29–42. Cartigny, P., 2005. Stable isotopes and the origin of diamond. Elements 1, 79–84. Cartigny, P., 2010. Mantle-derived carbonados? Geochemical insights from diamonds from the Dachine komatiite (French Guiana). Earth Planet. Sci. Lett. 296, 329–339. Cartigny, P., De Corte, K., Shatsky, V.S., Ader, M., De Paepe, P., Sobolev, N.V., Javoy, M., 2001. The origin and formation of metamorphic microdiamonds from the Kokchetav massif, Kazakhstan: a nitrogen and carbon isotopic study. Chem. Geol. 176, 265–281. Cartigny, P., Harris, J.W., Taylor, A., Davies, R., Javoy, P., 2003. On the possibility of a kinetic fractionation of nitrogen stable isotopes during natural diamond growth. Geochim. Cosmochim. Acta 67, 1571–1576. Chinn, I.L., 1995. A Study of Unusual Diamonds from the George Creek K1 Kimberlite Dyke, Colorado. (Ph.D. Thesis) University of Cape Town, p. 94. Cnudde, V., Dewanckele, J., De Boever, W., Brabant, L., De Kock, T., 2012. 3D characterization of grain size distributions in sandstone by means of X-ray computed tomography. In: Sylvester, P. (Ed.), Quantitative Mineralogy and Microanalysis of Sediments and Sedimentary Rocks. short course series, vol. 42, pp. 99–113. Dasgupta, R., Hirschmann, M.M., 2006. Melting in the Earth's deep upper mantle caused by carbon dioxide. Nature 440, 659–662. Dasgupta, R., Hirschmann, M.M., Withers, A.C., 2004. Deep global cycling of carbon constrained by the solidus of anhydrous, carbonated eclogite under upper mantle conditions. Earth Planet. Sci. Lett. 227, 73–85. De Corte, K., Cartigny, P., Shatsky, V.S., De Paepe, P., Sobolev, N.V., Javoy, M., 1999. Characteristics of microdiamonds from UHPM rocks of the Kokchetav massif (Kazakhstan). In: Gurney, J.J., Gurney, J.L., Pascoe, M.D., Richardson, S.H. (Eds.), Proceedings of the VIIth International Kimberlitic Conference. Red Roof Design cc, Cape Town, South Africa, pp. 174–182. De Souza Martins, M., 2006. Geologia dos Diamantes e Carbonados Aluvionares da Bacia do Rio Macaubas (MG). Tese de DoutoramentoUniversidade Federal De Minas Gerais, Instituto de Geociencias, Belo Horizonte, Minas Gerais, Brazil, (Ph.D. thesis). De, S., Heaney, P.J., Hargraves, R.B., Vicenzi, E.P., Taylor, P.T., 1998. Microstructural observations of polycrystalline diamond: a contribution to the carbonado conundrum. Earth Planet. Sci. Lett. 164, 421–433. Deines, P., 1980. The carbon isotopic composition of diamonds — relationship to diamond shape, color, occurrence and vapor composition. Geochim. Cosmochim. Acta 44, 943–961. Desbois, G., Urai, J.L., Burkhardt, C., Drury, M.R., Hayles, M., Humbel, B.M., 2008. Cryogenic vitrification and 3D serial sectioning using high cryo FIB SEM technology for brinefilled grain boundaries in Halite: first results. Geofluids 8, 60–72.

33

Dobosi, G., Kurat, G., 2002. Trace element abundances in garnets and clinopyroxenes from diamondites — a signature of carbonatitic fluids. Mineral. Petrol. 76, 21–38. Dobosi, G., Kurat, G., 2010. On the origin of silicate-bearing diamondites. Mineral. Petrol. 99, 29–42. Dobrzhinetskaya, L.F., 2012. Microdiamonds — frontiers of ultrahigh-pressure metamorphism: a review. Gondwana Res. 21, 207–223. Dobrzhinetskaya, L.F., Wirth, R., 2009. Ultradeep rocks and diamonds in the light of advanced scientific technologies. In: Cloetingh, S., Negendank, J. (Eds.), New Frontiers in Integrated Solid Earth Sciences. Springer, pp. 373–395. Dobrzhinetskaya, L.F., Green II, H.W., Bozhilov, K.N., Mitchell, T.E., Dickerson, R.M., 2003. Crystallization environment of Kazakhstan microdiamond: evidence from nanometric inclusions and mineral associations. J. Metamorph. Geol. 21, 425–437. Dobrzhinetskaya, L.F., Wirth, R., Green II, H.W., 2005. Direct observation and analysis of trapped COH fluid growth medium in metamorphic diamond. Terra Nova 17, 472–477. Dobrzhinetskaya, L.F., Liu, Z., Cartigny, P., Zhang, J., Tchkhetia, N.N., Green II, H.W., Hemley, R.J., 2006. Synchrotron infrared and Raman spectroscopy of microdiamonds from Erzgebirge, Germany. Earth Planet. Sci. Lett. 248, 340–349. Dobrzhinetskaya, L.F., Wirth, R., Green, H.W., 2007. A look inside of diamond-forming media in deep subduction zones. Proc. Natl. Acad. Sci. U. S. A. 104, 9128–9132. Dobrzhinetskaya, L.F., Wirth, R., Yang, J., Hutcheon, I.D., Weber, P.K., Green II, H.W., 2009. High-pressure highly reduced nitrides and oxides from chromitite of a Tibetan ophiolite. Proc. Natl. Acad. Sci. U. S. A. 106 (46), 19233–19238. Dobrzhinetskaya, L.F., Green, H.W., Takahata, N., Sano, Y., Shirai, K., 2010. Crustal signature of delta C-13 and nitrogen content in microdiamonds from Erzgebirge, Germany: ion microprobe studies. J. Earth Sci. 21, 623–634. Dobrzhinetskaya, L.F., Wirth, R., Green II, H.W., Schreiber, A., Obannon, E., 2013. First find of polycrystalline diamond in ultrahigh-pressure metamorphic terrane of Erzgebirge, Germany. J. Metamorph. Geol. 31 (1), 5–18. Eggler, D.H., 1976. Does CO2 cause partial melting in the low-velocity layer of the mantle? Geology 4, 69–72. Ernst, W.G., 2006. Preservation/exhumation of ultrahigh-pressure subduction complexes. Lithos 92, 321–335. Ernst, W.G., Liou, J.G., 2008. High- and ultrahigh-pressure metamorphism — past results, future prospects. Am. Mineral. 93, 1771–1786. Floss, C., Stadermann, F.J., Bradley, J.P., Zu, R.D., Bajt, S., Graham, G., Lea, A.S., 2006. Identification of isotopically primitive interplanetary dust particles: a NanoSIMS isotopic imaging study. Geochim. Cosmochim. Acta 70, 2371–2399. Foley, S.F., 2011. A reappraisal of redox melting in the Earth's mantle as a function of tectonic setting and time. J. Petrol. 52, 1363–1391. Foley, S.F., Yaxley, G.M., Rosenthal, A., Buhre, S., Kiseeva, E.S., Rapp, R.P., Jacob, D.E., 2009. The composition of near-solidus melts of peridotite in the presence of CO2 and H2O between 40 and 60 kbar. Lithos 112S1 (Special issue 9IKC), 247–283. Frantz, J.D., Mao, H.K., Zhang, Y., Wu, Y., Thompson, A.C., Underwood, J.H., Giauque, R.D., Jones, K.W., Rivers, M.-L., 1988. Analysis of fluid inclusions by X-ray fluorescence using synchrotron radiation. Chem. Geol. 69, 235–244. Garai, J., 2012. Ionized nitrogen mono-hydride bands are identified in the presolar and carbonado diamond spectra. Meteorit. Planet. Sci. 47, 1–7. Garai, J., Haggerty, S.E., Rekhi, S., Chance, M., 2006. Infrared absorption investigations confirm the extraterrestrial origin of carbonado diamonds. Astrophys. J. 653, L153–L156. Gautheron, C., Cartigny, P., Moreira, M., Harris, J.W., Allegre, C.J., 2005. Evidence for a mantle component shown by rare gases, C and N isotopes in polycrystalline diamonds from Orapa (Botswana). Earth Planet. Sci. Lett. 240, 559–572. Giannuzzi, L.A., Stevie, F.A., 2005. Introduction to Focused Ion Beams: Instrumentation, Theory, Techniques and Practice. Springer. Green, D.H., 1973. Experimental melting studies on a model upper mantle composition at high pressures under water-saturated and water-undersaturated conditions. Earth Planet. Sci. Lett. 19, 37–53. Green, D.H., Wallace, M.E., 1988. Mantle metasomatism by ephemeral carbonatite melts. Nature 336, 459–462. Gurney, J.J., Boyd, F.R., 1982. Mineral Intergrowths with Polycrystalline Diamonds from the Orapa Mine, Botswana. Carnegie Institution Year Book, pp. 267–273. Gurney, J.J., Jacob, W.R.O., Dawson, J.B., 1979. Megacrysts from the Monastery kimberlite pipe, South Africa. In: Boyd, F.R., Meyer, H.O.A. (Eds.), Proceedings 2nd International Kimberlite Conference, 2. AGU, pp. 227–243. Gurney, J.J., Harris, J.W., Rickard, R.S., 1984. Silicate and oxide inclusions in diamonds from the Orapa Mine, Botswana. In: Kornprobst, J. (Ed.), Kimberlites II: The Mantle and Crust–Mantle Relationships. Elsevier, Amsterdam. Haggerty, S.E., 1999. A diamond trilogy: superplumes, supercontinents and supernova. Science 289, 852–860. Haggerty, S.E., 2014. Carbonado: physical and chemical properties, a critical evaluation of proposed origins, and a revised genetic model. Earth-Sci. Rev. 130, 49–72. Harris, J.W., 1968. The recognition of diamond inclusions. Industrial Diamond ReviewDeBeers Industrial Diamond Division. Heaney, P.J., Vicenzi, E.P., De, S., 2005. Strange diamonds: the mysterious origins of carbonado and framesite. Elements 1 (2), 85–89. Hoppe, P., Cohen, S., Meibom, A., 2013. NanoSIMS: technical aspects and applications in cosmochemistry and biological geochemistry. Geostand. Geoanal. Res. 37, 111–154. Howell, D., O'Neill, C.J., Grant, K.J., Griffin, W.L., Pearson, N.J., O'Reilly, S.Y., 2012. μ-FTIR mapping: distribution of impurities in different types of diamond growth. Diam. Relat. Mater. 29, 29–36. Imamura, K., Ogasawara, Y., Yurimoto, H., Kusakabe, M., 2013. Carbon isotope heterogeneity in metamorphic diamond from the Kokchetav UHP dolomite marble, northern Kazakhstan. Int. Geol. Rev. 55, 453–467. Ishibashi, H., Kagi, H., Sakuai, H., Ohfuji, H., Sumino, H., 2012. Hydrous fluid as the growth media of natural polycrystalline diamond, carbonado: implication from IR spectra and microtextural observations. Am. Mineral. 97, 1366–1372.

34

D.E. Jacob et al. / Earth-Science Reviews 136 (2014) 21–35

Ishida, H., Ogasawara, Y., Ohsumi, K., Saito, A., 2003. Two stage growth of microdiamond in UHP dolomite marble from Kokchetav Massif, Kazakhstan. J. Metamorph. Geol. 21, 515–522. Izraeli, E.S., Harris, J.W., Navon, O., 2001. Brine inclusions in diamond: a new upper mantle fluid. Earth Planet. Sci. Lett. 187, 323–332. Jacob, D.E., 2004. Nature and origin of eclogites from the Earth's mantle. Lithos 77, 295–316. Jacob, D.E., Viljoen, K.S., Grassineau, N., Jagoutz, E., 2000. Remobilization in the cratonic lithosphere recorded by polycrystalline diamond (framesite). Science 289, 1182–1185. Jacob, D.E., Kronz, A., Viljoen, K.S., 2004. Cohenite, native iron and troilite inclusions in garnets from polycrystalline diamond aggregates. Contrib. Mineral. Petrol. 146, 566–576. Jacob, D.E., Wirth, R., Enzmann, F., Kronz, A., Schreiber, A., 2011. Nano-inclusion suite and high resolution micro-computed-tomography of polycrystalline diamond (framesite) from Orapa, Botswana. Earth Planet. Sci. Lett. 308 (3–4), 307–316. Jakobsson, S., Holloway, J., 1986. Crystal-liquid experiments in the presence of a C–O–H fluid buffered by graphite + iron + wüstite: experimental method and near liquidus relations in basanite. J. Volcanol. Geotherm. Res. 29, 265–291. Javoy, M., Pineau, F., Allegre, C.J., 1982. Carbon geodynamic cycle. Nature 300, 171–173. Jones, R., Briddon, P.R., Oberg, S., 1992. First-principles theory of nitrogen aggregates in diamond. Philos. Mag. Lett. 66, 67–74. Kamenetsky, M.B., Sobolev, A.V., Kamenetsky, V.S., Maas, R., Danyushevsky, L.V., Thomas, R., Sobolev, N.V., Pokhilenko, N.P., 2004. Kimberlite melts rich in alkali chlorides and carbonates: a potent metasomatic agent in the mantle. Geology 32, 845–848. Kaminsky, F.V., 1987. Origin of polycrystalline carbonado diamond aggregates. Dokl. Earth Sci. 294, 122–123. Kaminsky, F.V., 1991. Carbonado and yakutite: properties and possible genesis. In: Meyer, H.O.A., Leonardo, O.H. (Eds.), Proc. of the 5th International Kimberlite Conference, 2, pp. 136–143. Kaminsky, F.V., Wirth, R., 2011. Iron carbide inclusions in lower-mantle diamond from Juina, Brazil. Can. Mineral. 49 (2), 555–572. Kaminsky, F., Matzel, J., Jacobsen, B., Hutcheon, I.D., Wirth, R., 2012. First data on isotopic compositions of oxygen and carbon in lower-mantle carbonates and silicates. Abstract, Goldschmidt Conference, Montreal, Canada, June 2012. Kaminsky, F.V., Wirth, R., Morales, L., 2013. Internal texture and syngenetic inclusions in carbonado. Can. Mineral. 51, 39–55. Ketcham, R.A., Koeberl, C., 2013. New textural evidence on the origin of carbonado diamond: an example of 3D petrography using X-ray computed tomography. Geosphere. http://dx.doi.org/10.1130/GES00908.1. Kikuchi, M., Kagi, H., Ogasawara, Y., Dobrzhinetskaya, L., 2005. FTIR study of microdiamonds from garnet–biotite gneisses of the Kokchetav massif, Kazakhstan. EOS, AGU Meeting, p. F-1237. Klein-BenDavid, O., Wirth, R., Izraeli, E.S., Hauri, E., Navon, O., 2004. Brine and carbonatitic melts in a diamond from Diavik — implications for mantle fluid evolution. Geochim. Cosmochim. Acta 68, A276. Klein-BenDavid, O., Wirth, R., Navon, O., 2006. TEM imaging and analysis of microinclusions in diamonds: a close look at diamond-growing fluids. Am. Mineral. 91 (2–3), 353–365. Klein-BenDavid, O., Wirth, R., Navon, O., 2007. Micrometer-scale cavities in fibrous and cloudy diamonds — a glance into diamond dissolution events. Earth Planet. Sci. Lett. 264, 89–103. Klein-BenDavid, O., Logvinova, A.M., Schrauder, M., Spetius, Z.V., Weiss, Y., Hauri, E.H., Kaminsky, F.V., Sobolev, N.V., Navon, O., 2008. High-Mg carbonatitic microinclusions in some Yakutian diamonds—a new type of diamond-forming fluid. Lithos 112, 648–859. Kopylova, M., Navon, O., Dubrovinsky, L., Khachatryan, G., 2010. Carbonatitic mineralogy of natural diamond-forming fluids. EPSL 291, 126–137. Kurat, G., Dobosi, G., 2000. Garnet and diopside-bearing diamondites (framesites). Mineral. Petrol. 69, 143–159. Lee, M.R., 2010. Transmission electron microscopy (TEM) of Earth and planetary materials: a review. Mineral. Mag. 74, 1–27. Litasov, K.D., Ohtani, E., 2009. Phase relations in the peridotite–carbonate–chloride system at 7.0–16.5 GPa and the role of chlorides in the origin of kimberlite and diamond. Chem. Geol. 262, 29–41. Logvinova, A.M., Wirth, R., Fedorova, E.N., Sobolev, N.V., 2008. Nanometre-sized mineral and fluid inclusions in cloudy Siberian diamonds: new insights on diamond formation. Eur. J. Mineral. 20, 317–331. Maruoka, T., Kurat, G., Dobosi, G., Koeberl, C., 2004. Isotopic composition of carbon in diamonds of diamondites: record of mass fractionation in the upper mantle. Geochim. Cosmochim. Acta 68, 1635–1644. Massonne, H.J., 2003. A comparison of the evolution of diamondiferous quartz-rich rocks from the Saxonian Erzgebirge and the Kokchetav Massif: are so-called diamondiferous gneisses magmatic rocks? Earth Planet. Sci. Lett. 216, 347–364. Massonne, H.J., 2006. Early metamorphic evolution and exhumation of felsic highpressure granulites from the northwestern Bohemian Massif. Mineral. Petrol. 177, 177–202. McCall, G.J.H., 2009. The carbonado diamond conundrum. Earth Sci. Rev. 93 (3–4), 85–91. Mees, F., Swennen, R., van Geet, M., Jacobs, P., 2003. Applications of X-ray computed tomography in the geosciences. Geol. Soc. Lond. Spec. Publ. 215, 1–6. Mikhail, S., Dobosi, G., Verchovsky, A.B., Kurat, G., Jones, A.P., 2013. Peridotitic and websteritic diamondites provide new information regarding mantle melting and metasomatism induced through the subduction of crustal volatiles. Geochim. Cosmochim. Acta 107, 1–11. Mikhail, S., Howell, D., McCubbin, F.M., 2014. Evidence for multiple diamondite-forming events in the mantle. Am. Mineral. http://dx.doi.org/10.2138/am.2014.4778 (in press).

Moore, R.O., Gurney, J.J., 1989. Mineral inclusions in diamond from the Monastery kimberlite, South Africa. Spec. Publ. Geol. Soc. Aust. 14, 1029–1041. Navon, O., 1999. Diamond formation in the Earth's mantle. Proc 7th Int Kimberlite Conf, 2, pp. 584–604. Navon, O., Hutcheon, I.D., Rossman, G.R., Wasserburg, G.J., 1988. Mantle-derived fluids in diamond micro-inclusions. Nature 355, 784–789. Odling, N.W.A., 1989. The Role of Sulphur in Upper Mantle Processes. (Ph.D. Thesis) University of Tasmania. Ogasawara, Y., Aoki, K., 2005. The role of fluid for diamond-free UHP dolomite marble from Kokchetav Massif. Int. Geol. Rev. 47, 1178–1193. Ogasawara, Y., Ohta, M., Fukusava, K., Katayama, I., Maruyama, S., 2000. Diamond-bearing and diamond-free metacarbonate rocks from Kumdy-Kol from the Kokchetav massif, northern Kazakhstan. Island Arc 9, 400–416. Orlov, J.L., 1977. The Mineralogy of Diamond. Wiley, New York. Ozima, M., Tatsumoto, M., 1997. Radiation-induced diamond crystallization: origin of carbonados and its implications on meteorite nano-diamonds. Geochim. Cosmochim. Acta 61, 369–376. Ozima, M., Zashu, S., Tomura, K., Matsuhisa, Y., 1991. Constraints from noble-gas contents on the origin of carbonado diamonds. Nature 351, 472–474. Rege, S., Jackson, S., Griffin, W.L., Davies, R.M., Pearson, N.J., O'Reilly, S.Y., 2005. Quantitative trace-element analysis of diamond by laser ablation inductively coupled plasma mass spectrometry. J. Anal. At. Spectrom. 20, 601–611. Rondeau, B., Sautter, V., Barjon, J., 2008. New columnar texture of carbonado: cathodoluminescence study. Diamond Relat. Mater. 17, 1897–1901. Rubanova, E.V., Piazolo, S., Griffin, W.L., O'Reilly, S.Y., 2012. Deformation microstructures reveal a complex mantle history for polycrystalline diamond. Geochem. Geophys. Geosyst. 13, Q10010. http://dx.doi.org/10.1029/2012GC004250. Safonov, O.G., Perchuk, L.L., Litvin, Y.A., 2007. Melting relations in the chloride–carbonate– silicate systems at high-pressure and the model for formation of alkalic diamondforming liquids in the upper mantle. Earth Planet. Sci. Lett. 253, 112–128. Sano, Y., Yokochi, R., Terada, K., Chaves, M.L., Ozima, M., 2002. Ion microprobe Pb–Pb dating of carbonado, polycrystalline diamond. Precambrian Res. 113, 155–168. Sautter, V., Lorand, J.P., Cordier, P., Rondeau, B., Leroux, H., Ferraris, C., Pont, S., 2011. Petrogenesis of mineral micro-inclusions in an uncommon carbonado. Eur. J. Mineral. 23, 721–729. Saxena, S.K., 1989. Oxidation state of the mantle. Geochim. Cosmochim. Acta 53, 89–95. Schertl, P.P., Sobolev, N.V., 2013. The Kokchetav Massif, Kazakhstan: "Type locality" of diamond-bearing UHP metamorphic rocks. Journal of Asian Earth Sciences 63, 5–38. Schrauder, M., Navon, O., 1994. Hydrous and carbonatitic mantle fluids in fibrous diamonds from Jwaneng, Botswana. Geochim. Cosmochim. Acta 58, 761–771. Shirey, S.B., Cartigny, P., Frost, D.J., Keshav, S., Nestola, F., Nimis, P., Pearson, D.G., Sobolev, N.V., Walter, M.J., 2013. Diamonds and the geology of mantle carbon. Rev. Mineral. Geochem. 75, 355–421. Sitepu, H., Kopylova, M.G., Quirt, D.H., Cutler, J.N., Kotzer, T., 2005. Synchrotron micro-Xray fluorescence analysis of natural diamonds: first steps in identification of mineral inclusions in situ. Am. Mineral. 90, 1740–1747. Sitnikova, E.S., Shatsky, V.S., 2009. New data of FTIR spectroscopy on the composition of diamond crystallization medium in metamorphic rocks of the Kokchetav massif. Russ. Geol. Geophys. 50, 842–849. Smith, J.V., Dawson, J.B., 1985. Carbonado: diamond aggregates from early impact of crustal rocks? Geology 13, 342–343. Smith, E.M., Kopylova, M.G., Dubrovinsky, L., Navon, O., Ryder, J., Tomlinson, E.L., 2011. Transmission X-ray diffraction as a new tool for diamond fluid inclusion studies. Mineral. Mag. 75, 2657–2675. Sobolev, N.V., 1977. Deep-seated Inclusions in Kimberlites and the Problem of the Composition of the Upper Mantle. AGU, New York. Sobolev, N.V., Yefimova, E.S., Pospelova, L.N., 1989. Native iron in Yakutian diamonds and its paragenesis. Sov. Geol. Geophys. 22, 18–21. Sobolev, N.V., Logvinova, A.M., Efimova, E.S., 2009. Syngenetic phlogopite inclusions in kimberlite-hosted diamonds: implications for role of volatiles in diamond formation. Russ. Geol. Geophys. 50, 1234–1248. Sokol, A.G., Palyanova, G.A., Palyanov, Y.N., Tomilenko, A.A., Melenevsky, V.N., 2009. Fluid regime and diamond formation in the reduced mantle: experimental constraints. Geochim. Cosmochim. Acta 73, 5820–5834. Stachel, T., Harris, J.W., 2008. The origin of cratonic diamonds — constraints from mineral inclusions. Ore Geol. Rev. 34, 5–32. Stachel, T., Harris, J.W., Muehlenbachs, K., 2009. Sources of carbon ins inclusion bearing diamonds. Lithos 112 (2), 625–637. Stadermann, F.J., Croat, T.K., Bernatowicz, T., Amari, S., Messenger, S., Walker, R.M., Zinner, E., 2005. Supernova graphite in the NanoSIMS: carbon, oxygen and titanium isotopic compositions of a spherule and its TiC sub-components. Geochim. Cosmochim. Acta 69, 177–188. Stalder, R., Ulmer, P., Thompson, A.B., Günther, D., 2001. High pressure fluids in the system MgO–SiO2–H2O under upper mantle conditions. Contrib. Mineral. Petrol. 140, 607–618. Stöckhert, B., Duyster, J., Trepmann, C., Massonne, H.J., 2001. Microdiamond daughter crystals precipitated from supercritical COH + silicate fluids included in garnet, Erzgebirge, Germany. Geology 29, 391–394. Suk, N.I., 2001. Experimental study of liquid immiscibility in silicate–carbonate systems. Petrology 9, 477–488. Sumino, H., Dobrzhinetskaya, L.F., Burgess, R., Kagi, H., 2011. Deep-mantle-derived noble gases in metamorphic diamonds from the Kokchetav massif, Kazakhstan. Earth Planet. Sci. Lett. 307, 439–449. Sunagawa, I., 1990. Growth and morphology of diamond crystals under stable and metastable conditions. J. Cryst. Growth 99, 1156–1161.

D.E. Jacob et al. / Earth-Science Reviews 136 (2014) 21–35 Sunagawa, I., 2005. Crystals — Growth, Morphology and Perfection. Cambridge University Press, Cambridge. Svisero, D.P., 1995. Distribution and origin of diamonds in Brazil. J. Geodyn. 20, 493–514. Taylor, W.R., 1990. A reappraisal of the nature of fluids included by diamond — a window to deep-seated mantle fluids and redox conditions. In: Herbert, H.K., Ho, S.E. (Eds.), Stable Isotopes and Fluid Processes in Mineralization. University of Western Australia, Perth. Taylor, W.R., Green, D.H., 1987. The petrogenetic role of methane: effect on liquidus phase relations and the solubility mechanism of reduced C–H volatiles. In: Mysen, B.O. (Ed.), Magmatic Processes: Physicochemical Principles. Geochem. Soc, Washington, pp. 121–138. Taylor, W.R., Green, D.H., 1988. Measurement of reduced peridotite — C–O–H solidus and implications for redox melting of the mantle. Nature 332, 349–352. Tomlinson, E.L., Jones, A.P., Harris, J.W., 2006. Co-existing fluid and inclusions in mantle diamond. Earth Planet. Sci. Lett. 250, 581–595. Tomlinson, E.L., Müller, W., EIMF, 2009. A snapshot of mantle metasomatism: trace element analysis of coexisting fluid (LA-ICP-MS) and silicate (SIMS) inclusions in fibrous diamonds. Earth Planet. Sci. Lett. 279, 362–372. Trueb, L.F., Butterman, W.C., 1969. Carbonado: a microstructural study. Am. Mineral. 54, 412–425. Ulmer, G.C., Grandstaff, D.E., Woermann, E., Goebbels, M., Schonitz, M., Woodland, A.B., 1998. The redox stability of moissanite (SiC) compared with metal–metal oxide buffers at 1773 K and at pressures up to 90 kbar. N. Jahrb. Mineral. 172, 279–307. Vicenzi, E.P., Heaney, P.J., Ketcham, R.A., Rivers, M., 2003. High-Resolution X-ray Microtomography of Polycrystalline Diamond. Goldschmidt Conference Abstracts, p. A513. Weiss, Y., Kessel, R., Griffin, W.L., Kiflawi, I., Klein-BenDavid, O., Bell, D.R., Harris, J.W., Navon, O., 2009. A new model for the evolution of diamond-forming fluids; evidence from microinclusion-bearing diamonds from Kankan, Guinea. Lithos 112, 660–674. Weiss, Y., Kiflawi, I., Navon, O., 2010. IR spectroscopy: quantitative determination of the mineralogy and bulk composition of fluid microinclusions in diamonds. Chem. Geol. 275 (1–2), 26–34.

35

Weiss, Y., Griffin, W.L., Bell, D.R., Navon, O., 2011. High-Mg carbonatitic melts in diamonds, kimberlites and the sub-continental lithosphere. Earth Planet. Sci. Lett. 309, 337–347. Weiss, Y., Kiflawi, I., Navon, O., 2013. The IR absorption spectrum of water in microinclusion-bearing diamonds. In: Pearson, D.G., Grütter, H., Harris, J.W., Kjarsgaard, B.A., O'Brien, H., Chalapathi Rao, N.V., Sparks, S. (Eds.), Proceedings of 10th International Kimberlite Conference. Spec. Issue Journal of the Geological Society of India, vol. 1, pp. 271–280. Wirth, R., 2004. A novel technology for advanced application of micro- and nanoanalysis in geosciences and applied mineralogy. Eur. J. Mineral. 16 (6), 863–876. Wirth, R., 2009. Focused Ion Beam (FIB) combined with SEM and TEM: advanced analytical tools for studies of chemical composition, microstructure and crystal structure in geomaterials on a nanometre scale. Chem. Geol. 261 (3–4), 217–229. Wirth, R., 2010. Focused Ion Beam (FIB): site-specific sample preparation, nano-analysis, nano-characterization and nano-machining. In: Brenker, F.E., Jordan, G. (Eds.), Nanoscopic Approaches in Earth and Planetary Sciences. Wirth, R., Pinti, D.L., Sano, Y., Takahata, N., Kaminsky, F., Wirth, R., Pinti, D.L., Sano, Y., Takahata, N., Kaminsky, F., 2007. FIB/TEM and NanoSIMS investigations of microand nanoinclusions in diamond: new insights into the Earth's mantle and diamond formation. Abstract AGU Fall Meeting, San Francisco. EOS, Transactions, American Geophysical Union, Supplement, 88, (52), pp. 10–14 (December, 2007, Abstract MR23D-10). Wirth, R., Kaminsky, F., Matsyuk, S., Schreiber, A., 2009. Unusual micro- and nanoinclusions in diamonds from the Juina Area, Brazil. Earth Planet. Sci. Lett. 286 (1–2), 292–303. Wyllie, P.J., 1978. Mantle fluid compositions buffered in peridotite — CO2–H2O by carbonates, amphibole, and phlogopite. J. Geol. 86, 687–713. Zegdenizov, D.A., Kagi, H., Shatsky, V.S., Sobolev, N.V., 2004. Carbonatitic melts in cuboid diamonds from Udachnaya kimberlite pipe (Yakutia). Evidence from vibrational spectroscopy. Mineral. Mag. 68, 61–73. Zhang, C., Duan, Z., 2009. A model for C–O–H fluid in the Earth's mantle. Geochim. Cosmochim. Acta 73, 2089–2102.