New Textural Evidence on the Origin of Carbonado Diamond: an Example of 3-D Petrography Using X-Ray Computed Tomography

New textural evidence on the origin of carbonado diamond: An example of 3-D petrography using X-ray computed tomography

Richard A. Ketcham1 and Christian Koeberl2 1High-Resolution X-Ray Computed Tomography Facility, Jackson School of Geosciences, The University of Texas at Austin, Austin, Texas 78712, USA 2Department of Lithospheric Research, University of Vienna, Althanstrasse 14, A-1090 Vienna, Austria, and Natural History Museum, Burgring 7, A-1010 Vienna, Austria

ABSTRACT material that comprises them may not be, –32‰, with major modes at –24‰ and –26‰, and may instead be broken-down remains of rare for diamonds and suggestive of organic car- Three-dimensional textural observations the original included phase(s). While further bon (De et al., 2001; Kagi et al., 2007), although of inclusion and porosity patterns in a 23- verifi cation is needed, a model built around the source for light carbon in the mantle remains carat carbonado diamond using high-reso- this hypothesis may provide the simplest an open question (Deines, 2002). The diamond lution X-ray computed tomography reveal explanation to many of the unusual features material contains H (hydrogen) defects indica- new information bearing on the nature and of carbonado. tive of a hydrogen-rich environment (Garai et al., origin of this enigmatic material. A promi- 2006; Nadolinny et al., 2003), and N (nitrogen) nent patinaed surface is texturally linked to INTRODUCTION defects in various states of aggregation (Fukura a banding and grading of inclusions and pore et al., 2005; Garai et al., 2006). Carbonado hosts space beneath, extending several millimeters Carbonado is an enigmatic polycrystalline extensive porosity ranging from the nanometer up into the specimen. In situ observation dem- diamond variety found in placer deposits in to the millimeter scale. Some carbonado material onstrates that almost all inclusions are poly- Brazil and the Central African Republic (Trueb has been found to be enriched in isotopes char- mineralic and show replacement textures, and Butterman, 1969; Trueb and De Wys, 1969, acteristic of the products of spontaneous fi ssion corroborating previous work indicating that 1971). Carbonados are commonly millimeter to of uranium (Ozima and Tatsumoto, 1997), but the pore network is fully three-dimensionally centimeter sized, and the “Carbonado of Sergio” cathodoluminescence imaging suggests that this (3-D) connected, and that virtually all macro- from Brazil is the largest diamond known (Hag- may be a secondary feature caused by infi ltration inclusions are secondary. Large metal inclu- gerty, 1999) at 3167 carats. Age determinations of the pore network by fl uids enriched in radio- sions are only found immediately adjacent range from 2.6 to 3.8 Ga, with large uncertainties active elements (Kagi and Fukura, 2008; Kagi to the margin of the specimen, and are thus (Ozima and Tatsumoto, 1997; Sano et al., 2002). et al., 2007; Rondeau et al., 2008). also likely to be secondary or even tertiary. The confi ned geographic distribution of carbo- The inclusion suite in carbonado has also However, we also report pseudomorphs after nados, on land masses that were likely adjacent proven enigmatic. It hosts enclosed nano- a phase forming pristinely euhedral rhombic at times in the Archean, suggests that they may inclusions of various metal alloys (Fe, Fe-Ni, dodecahedra, individually and in clusters all have been created in a single event (Heaney Ni-Pt, Si, Ti, Sn, Ag, Cu, and SiC), as well as from 0.3 to 1 mm in diameter; although we et al., 2005), although they are in some respects other minerals (calcite, sylvite, smithsonite) could fi nd no evidence of this phase persist- similar to other polymineralic diamond varieties that may be original (De et al., 1998) and some ing, it nevertheless represents the fi rst “true” such as framesites and yakutites (McCall, 2009). (augite, ilmenite, phlogopite) that may pre- macro-inclusion reported in carbonado, Carbonado is set apart from other diamond date the diamond-forming event (Sautter et al., which almost certainly formed syngenetically varieties by several unusual characteristics. The 2011). It also hosts larger metal particles tens with the diamond material. The pore system diamond material is texturally diverse, being of microme ters in diameter within open pore is essentially trimodal, consisting of single dominantly porphyritic and consisting of small space (De et al., 1998; Fitzgerald et al., 2006). and clustered pseudomorphs, oblate pores (10–250 µm) diamonds cemented together by Its macro-inclusion suite principally features 0.1–0.3 mm in length with a clear preferred even smaller (<1 µm) microdiamonds (De et al., minerals of ostensibly crustal origin, such as orientation, and 20 µm to <1 µm pores that 1998; Petrovsky et al., 2010). The smaller crys- orthoclase, goethite, quartz, kaolinite, hematite, form the connected network. Our observa- tals can have subplanar dislocations that may serpentine, and fl orencite, a hydrous rare-earth tions support recent work suggesting that be interpreted as defect lamellae, whereas the phosphate that commonly forms as an altera- carbonado crystallized from a carbon-super- larger ones are mostly defect free. Other, rarer tion product of monazite. The fact that extensive saturated fl uid and suggest that the second textures found in subregions are homogeneous, leaching can extract virtually all compositional stage may correspond with the creation of fl ow texture (Yokochi et al., 2008) and colum- impurities (Dismukes et al., 1988) and magnetic the pore alignment fabric. We further pos- nar crystals reminiscent of vein infi ll (Rondeau components (Fitzgerald et al., 2006; Kletetschka tulate that, although the present-day macro- et al., 2008). Carbonado carbon isotopes are very et al., 2000) suggests that all macro-inclusions inclusions are certainly secondary, the bulk light, with δ13C values ranging from –21‰ to are probably secondary products precipitated

Geosphere; October 2013; v. 9; no. 5; p. 1336–1347; doi:10.1130/GES00908.1; 9 ﬁ gures; 14 animations. Received 8 February 2013 ♦ Revision received 22 May 2013 ♦ Accepted 25 July 2013 ♦ Published online 14 August 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/5/1336/3344992/1336.pdf by guest on 27 September 2021 3-D petrography of carbonado

by fl uids moving through a connected pore net- 13 mm × 10 mm, with a deltoid shape (Fig. 1). work. None of the inclusion phases found thus One broad side and a smaller adjacent edge have far is typical of kimberlitic diamond, such as a silvery, vitreous patina, a common though not 5 mm pyrope, pyrrhotite, and chromian clinopyroxene ubiquitous feature in carbonados that has been (Heaney et al., 2005), although olivine has been interpreted by some as a fusion crust (e.g., reported (Trueb and De Wys, 1971). Ishibashi Shelkov et al., 1997), although overall the side et al. (2012) report a 0.3 µm void with euhe- is rough. The other side is a dull dark gray and dral walls that they interpreted as a former fl uid rougher. Pores are visible to the naked eye on inclusion, though it was not preserved as such all surfaces. due to sample preparation. We imaged the specimen using the Xradia The origin of these diverse features in carbo- MicroXCT scanner at the University of Texas nado has eluded consensus for over 40 yr. High-Resolution X-Ray CT Facility (http:// Hypotheses have included meteorite impact www.ctlab.geo.utexas.edu). The MicroXCT (Smith and Dawson, 1985); formation during hot has a unique design for a micro–computed Archean subduction of early organic material; tomography scanner. Whereas most instru- volcanic hydrothermal systems with metal-cata- ments utilize large detectors and geometric lyzed diamond growth (McCall, 2009); extrater- magnifi cation from a small X-ray focal spot to restrial origin (Garai et al., 2006); and irradiation achieve high resolution, the MicroXCT derives of carbonaceous materials (Ozima and Tatsu- its resolution from a set of specialized detec- moto, 1997). Recent work has favored formation tors consisting of a 35 mm camera lens or vari- at mantle conditions, likely in the presence of a ous microscope objectives coated with scintil- C-O-H fl uid supersaturated in carbon (Ishibashi lating material. This arrangement produces Figure 1. Carbonado specimen as mounted et al., 2012; Petrovsky et al., 2010), a C-H–rich very sharp imagery and facilitates “zooming for micro–computed tomography (CT) fl uid in lower crust transiently metamorphosed at in” to image small subvolumes within larger imaging, along with samples of calcite, apa- mantle conditions (Sautter et al., 2011), or asso- specimens by simply switching detectors, in tite, and almandine garnet for comparison ciated with komatiite magma intrusion into the much the same manner one switches objec- of X-ray attenuation. The patinaed side of continental lithosphere (Cartigny, 2010). tives with a petrographic microscope (Figs. the specimen is shown. Because carbonado is among the tough- 2A, 2B, and 2C). With each such increase in est substances in nature, it is diffi cult to study magnifi cation, however, acquisition time rises using traditional petrographic techniques. With considerably, due to lower signal and the need Following scanning, a single cut was made in the exception of early X-ray imaging (Trueb to acquire more densely sampled data (i.e., the sample using a laser at the Carnegie Insti- and Butterman, 1969; Trueb and De Wys, more views) to reduce the effect of parts of the tute. After polishing the surface on a diamond 1969, 1971), most studies have concentrated sample not in the region of interest but still in wheel, scanning electron microscopy (SEM) on microchemical analyses of small pieces the X-ray path. images and chemical analyses were obtained of material that were powdered, fractured, The specimen was scanned at a range of reso- using the JEOL JSM-6490LV at the University Focused Ion Beam– (FIB) extracted, laser-cut, lutions with different detectors, using 29.2 µm of Texas. or polished with laborious effort, providing a voxels to image the whole sample and 5.4 µm Visualization was a particularly important spatially limited and largely two-dimensional and 1.0 µm voxels to image selected subvol- aspect of our methodology. It is not an exaggera- perspective. Most macro-inclusions have been umes within it. X-ray energies were varied from tion to say that many of the insights in this study studied by disaggregating carbonado speci- 40 to 140 kV, and some scans were repeated at can only be obtained by viewing the CT imagery mens. It is particularly challenging to examine different energies to help with discrimination of in motion, either as slice animations stepping the macro-inclusions and diamond in concert on phases. Crystals of almandine, apatite, and cal- through the data or as rotating volume render- anything but a fractured or laser-cut surface, due cite were also included in the fi eld of view when ings allowing structures and fabrics to be viewed to the extreme range of material hardness. As a imaging the entire specimen to aid in interpre- from many orientations. We thus provide a num- result, relating these features to each other has tation of computed tomography (CT) numbers ber of these animations as supplemental material. been problematic. Here, we report on a textural (image gray levels). These animations are also available at http://www analysis using high-resolution X-ray computed Data were reconstructed as up to 900 ~1000 × .ctlab.geo.utexas.edu/pubs/ketcham_koeberl tomographic (HRXCT) imaging of a relatively 1000 16 bit grayscale images per scan, and soft- /ketcham_koeberl.htm. large specimen. Three-dimensional nondestruc- ware corrections were applied to compensate tive imaging of pores and inclusions in situ for ring and beam hardening artifacts. The volu- Interpretation of CT Numbers for enables us to look at various textural details in metric data were visualized with Avizo (VSG, Dual-Energy CT Scans their spatial context, providing a holistic view, Inc., versions 6.1 and 7.0) and VG Studio Max and essentially allowing “three-dimensional (Volume Graphics, Inc., version 2.0). Measure- CT numbers refl ect the linear X-ray attenua- (3-D) petrography” to be performed. ments of metal inclusions were carried out with tion coeffi cient of the materials being scanned, Blob3D software, accounting for partial-volume which are a function of density, atomic number SPECIMEN AND METHODS effects to compensate for their small size (Ket- (Z), and X-ray energy (Ketcham and Carlson, cham, 2005a, 2006), and 3-D determination of 2001). Due to the complexities of scanning, such Our sample is from the Central African inclusion preferred orientation was conducted as the polychromatic nature of the X-ray source Republic, and it was purchased from a dealer in using star volume analysis with Quant3D soft- and beam hardening and other artifacts, as well Belgium. It is 23.45 carats (4.690 g), ~18 mm × ware (Ketcham, 2005b). as the variability of natural materials, which can

Geosphere, October 2013 1337

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/5/1336/3344992/1336.pdf by guest on 27 September 2021 Ketcham and Koeberl

A B

2 mm 1 mm

C D E

0.2 mm 0.2 mm 0.2 mm

Figure 2. (A) 29.2-µm-resolution computed tomography (CT) image encompassing entire specimen cross section, with patinaed surface on right. Circle indicates position of part B. (B) 5.3-µm-resolution image showing polymineralic nature of most inclusions. Circle indicates position of C and D. (C) 1.0-µm-resolution, 140 kV scan showing pseudomorphs; bright material in center is probably ﬂ orencite. (D) Same ﬁ eld of view as C, imaged with X-rays set at 40 kV, increasing contrast of inclusion to lower right (further imagery in Animation 14). (E) Vol- ume rendering of the high-attenuation material in pseudomorph in C, indicating replacement of a rhombic dodecahedron. Slice data are given in Animations 1 and 2.

feature zoning, impurities, and microporosity, other machine-specifi c factors such as detec- at progressively higher energies, up to 63.3 keV absolute mapping of CT number to phases is tor effi ciency as a function of energy. It is also for Lu. By scanning below and substantially often not straightforward. As a result, CT data affected by beam hardening within the sample above 40 kV (Animations 1 and 2), the effective in general do not have information suffi cient and measures taken to correct for it. attenuation coeffi cient of REE-rich phases will for independent mineral identifi cation, neces- The relationship between X-ray energy and change dramatically with respect to other min- sitating prior knowledge of which phases are linear attenuation coeffi cient for the comparison erals. In this case, Figure 3 suggests that when likely to be present. The effective attenuation minerals used and some of the phases discussed the entire X-ray spectrum used is below 40 keV, coeffi cient for a mineral is essentially the inte- in this study is shown in Figure 3. The step then fl orencite will have an effective attenua- gral of all attenuation coeffi cients over the range function for fl orencite is caused by the X-ray tion coeffi cient similar to almandine, whereas of X-ray energies used, weighted by the relative absorption K edge of Ce at 40.4 keV; heavier when a large portion of the spectrum is above intensity of the X-ray fl ux at each energy, and rare earth elements (REEs) have their K edges 40 keV, then fl orencite will be substantially

1338 Geosphere, October 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/5/1336/3344992/1336.pdf by guest on 27 September 2021 3-D petrography of carbonado

10000

diamond florencite 1000 spinel kaolinite Fe almandine Animation 1. Individual slice animation for 100 calcite a 29.2-µm-resolution data set of carbonado acquired with X-rays set at 40 kV, at two apatite

) contrast levels. Patinaed side of specimen is –1 on right side of images. Images are 17.9 mm

μ(cm wide (1224 × 612 pixels; 10.4 MB). If you are 10 viewing the PDF of this paper or reading it ofﬂ ine, please visit http://dx.doi.org/10.1130 /GES00908.S1 or the full-text article on www.gsapubs.org to view Animation 1.

0.1 10 60 110 160

Energy (keV)

Figure 3. X-ray linear attenuation coeffi cient µ vs. X-ray energy for Animation 2. Individual slice animation for minerals discussed in this study, over the X-ray spectrum used. 29.2-µm-resolution data sets of carbonado acquired with X-rays set at 40 kV (left) and 140 kV (right). Patinaed side of specimen is more attenuating . Because previous carbonado OBSERVATIONS on right side of images. Images are 17.9 mm work indicates that only REE-rich phases will wide (1224 × 612 pixels; 6.0 MB). If you are have heavy elements in suffi cient quantity to Mineralogy viewing the PDF of this paper or reading it affect their X-ray attenuation, and other REE- offl ine, please visit http://dx.doi.org/10.1130 rich minerals aside from fl orencite have also Figure 4 and Animations 3–6 show a series of /GES00908.S2 or the full-text article on been reported (e.g., rhabdophane, xenotime; De volume renderings of the inclusion phases in the www.gsapubs.org to view Animation 2. et al., 1998), in this study we generically refer 29.2-µm-resolution data, in which the diamond to minerals identifi ed by reduced relative X-ray is rendered transparent except for the sample attenuation below 40 keV as REE rich. boundary, and other phases are rendered trans- rendered opaque is expanded to encompass the At energies below 40 keV, the most attenuat- parent or partially or fully opaque based on CT values for the apatite and calcite samples (Figs. ing macrophase we expect to fi nd based on prior number. The highest-attenuation phases in the 4C and 4D; Animations 5 and 6), more inclusion investigations of carbonado is native Fe. The data gathered at 140 kV (Fig. 4A; Animation 3) phases become visible. diamond material itself has very low attenu- are expected to be REE-rich minerals such as Native metal inclusions were identified ation at low kV because of its extremely low fl orencite, and native Fe, the latter of which by their high CT numbers in the 40 kV, mean atomic number relative to other phases. is far less abundant. REE-rich phases occur 29.2-µm-resolution data (Fig. 5A, upper left). In As X-ray energy increases, and the dominant throughout the sample. When data are acquired total, 16 were identifi ed and measured in the com- attenuation mechanism transitions from photo- at 40 kV (Fig. 4B; Animation 4), the REE-rich plete data set, with sphere-equivalent diameters electric absorption to Compton scattering, the phases are indistinguishable from the almandine ranging from 80 µm to 148 µm, and a mean value attenuation coeffi cient becomes a linear func- crystal included in the scan fi eld, as predicted of 107 µm. Every metal inclusion was directly tion of energy, and other phases with higher from the relations shown in Figure 3. The simi- adjacent to the boundary of the specimen, with mean atomic numbers but lower density (i.e., larity of the inclusions rendered opaque in Fig- the largest distance from sample exterior to inclu- spinel, kaolinite) come to have similar or lower ures 4A and 4B indicates that there is little mate- sion being 300 µm (see Animation 7). All of these attenuation coeffi cients compared to diamond. rial similar to almandine (i.e., a relatively dense inclusions appear to be linked to the outside of the For example, a euhedral inclusion faintly visible Fe-bearing phase) in the carbonado, and this specimen via open porosity, though the local arti- in the lower right corner in Figure 2D, which corroborates the interpretation that the phases facts cast by the inclusions in the HRXCT data, was imaged at 40 kV, is almost indistinguish- shown in Figure 4A are REE-bearing due to as seen for example in Figure 5A, prevent this able from the diamond matrix when imaged at their substantial change in attenuation rela- statement from being made unequivocally. No 140 kV in Figure 2C. tive to almandine. As the range of CT numbers smaller metallic inclusions were observed within

Geosphere, October 2013 1339

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/5/1336/3344992/1336.pdf by guest on 27 September 2021 Ketcham and Koeberl

Figure 4. Volume renderings A 140 kV, μeff > almandine B 40 kV, μeff ≥ almandine from 29.2 µm resolution data sets showing various inclusion populations based on their effective attenuation relative to the index minerals scanned with the carbonado. Diamond is rendered transparent with purple boundary; inclusion colors are based on index minerals (green—REE- rich minerals more attenuating than almandine at X-ray energies above ~40 keV; red—almandine; yellow—apatite; white— calcite). Rotation Animations 3–6 allow viewing from a range of angles. (A) 140 kV, computed tomography (CT) number > almandine. See also Animation 3 (http://dx.doi.org/10.1130 /GES00908.S3); 903 × 1013 C 40 kV, μeff ≥ apatite D 40 kV, μeff ≥ calcite pixels; 7.8 MB. (B) 40 kV, CT number ≥ almandine. See also Animation 4 (http://dx.doi.org /10.1130/GES00908.S4); 903 × 1013 pixels ; 7.2 MB. (C) 40 kV, CT number ≥ apatite. See also Animation 5 (http://dx.doi.org /10.1130/GES00908.S5); 903 × 1013 pixels; 6.7 MB. (D) 40 kV, CT number ≥ calcite. See also Animation 6 (http://dx.doi.org /10.1130/GES00908.S6); 903 × 1013 pixels; 6.0 MB. If you are viewing the PDF of this paper or reading it ofﬂ ine, please visit the individual doi links or the full- text article on www.gsapubs.org to view Animations 3–6.

the specimen in any of the higher-resolution sub- 1 and 2) with up to three components distin- patinaed surface, and in all other faces. How- volume scans; the nano-inclusions observed in guished by the abundance of small inclusions: ever, smaller inclusions with maximum dimen- transmission electron microscopy (TEM) studies a 300–500-µm-wide inclusion-depleted outer sions in the tens to hundreds of micrometers (De et al., 1998; Sautter et al., 2011) are below rim; an ~500-µm-wide inclusion-rich middle have an apparent gradation from sparse near the the resolution of the HRXCT data. band that features overall slightly higher X-ray patinaed surface to increasingly abundant with SEM analysis on the cut section showed attenuation than surrounding material; and an increasing distance from it, as seen both in the kaolinite and possibly other clay minerals, with inner inclusion-depleted zone with variable example slice image (Fig. 5A) and most clearly minor iron, to be by far the most common inclu- thickness. The outermost zone may correspond in a 3-D volume rendering (Fig. 5B; Animation sion mineral in this specimen. Disseminated to the 200-µm-thick outer rind investigated with 6). There is also an apparent dearth of small ﬂ orencite was often distinguishable by signiﬁ - SEM by Shelkov et al. (1997), who found it to inclusions along the non-patinaed surface, but cant Ce peaks. Other probable phases included be free of microporosity, which they interpreted closer inspection reveals both open porosity on ilmenite and quartz. to be a result of annealing just below what they the same size scale as the small inclusions, and interpreted to be the fusion crust. many inclusions of lower attenuation that do not Macrotextures Large, millimeter-scale inclusions occur show up as clearly in the volume rendering, so throughout the specimen, with no readily dis- this is most likely a surface leaching effect. The volume directly beneath the patinaed cernible pattern, even crossing the three-layer When the region near the patinaed surface is side of the specimen has an evident rim tex- rim texture. These inclusions are responsible observed more closely with 5.4-µm-resolution ture (Fig. 5A, right side; also see Animations for most of the megaporosity observed on the data (Fig. 6), it is also clear that there are low-

1340 Geosphere, October 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/5/1336/3344992/1336.pdf by guest on 27 September 2021 3-D petrography of carbonado

A B

Figure 5. Example micro–computed 5 mm tomography imagery of carbonado specimen based on 40 kV, 29.2 µm data. Both images have patinaed side on right, normal to image plane. (A) Slice image showing the three-layered texture (divided arrow marking zone boundaries), with small inclusions becoming more abundant from right to left. Top arrow indicates native metal; all such grains were close to the edge of the specimen. (B) Volume rendering showing graded small inclusion pattern. Diamond is rendered transparent, and inclusion colors are based on scheme in Figure 4. X-ray attenuation of inclusions increases from purple to white to yellow to orange. 2 mm

(Animation 8). The band has a higher density There are two distinct populations of inclu- of small inclusions near the resolution limit of sions evident in the 5.4-µm-resolution HRXCT those data, but it retains an overall brightening data (Fig. 6C; Animations 9 and 10). One is that cannot be ascribed to individual inclusions. smaller and irregular, but often roughly ellip- Based on the progression observed across data soidal and with a preferred alignment. The other resolutions, however, it can be inferred that the is larger and features many perfectly euhedral brightening is probably caused by yet smaller boundaries in contact with the diamond matrix. inclusions. In other words, there are within the diamond Cursory viewing suggests that the inclusions material perfect negative crystal forms much have a preferred orientation that is oblique to larger than the diamond crystallites, which host the overall apparent grading direction. This is polymineralic assemblages with disequilibrium supported by 3-D fabric analysis that combines textures. These latter minerals are thus clearly all phases signifi cantly more attenuating than pseudomorphs replacing an earlier phase. The diamond in the 40 kV, 29.2-µm-resolution data most distinct pseudomorphed habit is rhombic (Fig. 6B). As suggested by the fabric analysis dodecahedral {011}, which tends to occur in and corroborated by data visualization, the pre- clumps of 0.2–1 mm crystal forms (Figs. 2E dominant shape of elongated pores corresponds and 7; Animations 11–13). to blades (ellipsoid axes a > b > c). Similar Aside from the near-edge occurrences of Animation 7. Three-dimensional (3-D) vol- fabrics were also observed in early X-ray imag- inclusions interpreted to be native metal, all ume rendering animation for 40 kV, 29.2 µm ing (Trueb and Butterman, 1969; Trueb and inclusions containing high-Z phases that we data set, with only computed tomography De Wys, 1969, 1971). observed in detail were polymineralic; the only (CT) numbers ≥ metal shown (small orange inclusions we observed that appeared to be particles around specimen edges) (903 × Microtextures monomineralic were relatively large and con- 1013 pixels; 6.7 MB). If you are viewing the tained low-Z material. The 1-µm-resolution PDF of this paper or reading it offl ine, please The higher-resolution 3-D images (Fig. 6; images, which include one low-Z inclusion visit http://dx.doi.org/10.1130/GES00908.S7 Animations 8–10) of the inclusions show that (Fig. 2D, lower right; Animation 14), reveal or the full-text article on www.gsapubs.org nearly all of them (>99%) are composites with extremely sharp grain boundaries and a rhombic to view Animation 7. multiple minerals. They are in many cases a mix dodecahedral habit. We positioned our laser cut between low-attenuation and high-attenuation to expose this latter inclusion and found it to con- attenuation inclusions and open pores, but the phases, the latter in most cases likely to be sist of kaolinite, and thus another pseudomorph. overall grading pattern of lower-inclusion-num- fl orencite. Grain boundaries of phases within The polymineralic nature of many inclusions ber density toward the patinaed surface remains inclusions are diffuse and intergrown or mottled, complicated 3-D visualization by volume ren- clear. The inclusion-rich band near the patinaed indicating alteration or replacement and sug- dering, and thus required segmentation by hand surface seen in Figure 5A can also be detected gesting disequilibrium and material exchange to fully discern pseudomorph faces. Selected in some regions of the 5.4-µm-resolution data after the diamond had formed. examples are shown in Figure 7 and Anima-

Geosphere, October 2013 1341

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/5/1336/3344992/1336.pdf by guest on 27 September 2021 Ketcham and Koeberl

AB1 max SVD/SVD

Figure 6. Computed tomogra- 5 mm phy data of patinaed border region and inclusion preferred orientation. (A) Index image from 29.2 µm data, show- C ing position of C. (B) Three- dimensional (3-D) rose diagram of inclusions from center of specimen, showing preferred orientation; north is up in A, east is left, etc. Colors and distance from the origin indicate relative strength of fabric according to the star volume distribution metric (Ketcham, 2005b). (C) 5.4-µm-resolution data showing close-up of region near patinaed surface, showing preferred orientation of pores and inclusions, and increasing inclusion density from right to left. Full data set is provided in Animation 10.

patinaed surface

internal 1 mm open pores

1342 Geosphere, October 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/5/1336/3344992/1336.pdf by guest on 27 September 2021 3-D petrography of carbonado

Animation 8. Individual slice animation for Animation 9. Individual slice animation Animation 10. Individual slice animation longer-acquisition 40 kV, 5.4 µm data set for 40 kV, 5.4 µm data set encompassing for 40 kV, 5.4 µm data set encompassing of specimen interior. Denser angular sam- different regions of the specimen interior. different regions of the specimen interior. pling reduced interference artifacts, mak- Images are 5.3 mm wide (1006 × 1024 pixels ; Images are 5.3 mm wide (1006 × 1024 pixels; ing porosity more apparent. Patinaed side 15.1 MB). If you are viewing the PDF of this 14.9 MB). If you are viewing the PDF of this appears along right edge. Images are 5.3 mm paper or reading it offl ine, please visit http:// paper or reading it offl ine, please visit http:// wide (992 × 1013 pixels; 57 MB). If you are dx.doi.org/10.1130/GES00908.S9 or the full- dx.doi.org/10.1130/GES00908.S10 or the viewing the PDF of this paper or reading it text article on www.gsapubs.org to view full-text article on www.gsapubs.org to view offl ine, please visit http://dx.doi.org/10.1130 Animation 9. Animation 10. /GES00908.S8 or the full-text article on www .gsapubs.org to view Animation 8.

fl orencite and kaolinite. Figure 8B shows a pair specimen, though some are present in the region tions 11 and 12. Among the 13 inclusions thus of elliptical, irregular inclusions that are also ~1 mm inward from the patinaed surface. Some processed , some were undeterminable, but the polymineralic. of these megapores have a similar size and shape only habit identifi ed was rhombic dodecahedral. to nearby inclusions (Fig. 6C), suggesting disso- SEM imagery of the laser-cut surface (Fig. 8) Porosity lution of previous grains or infi lling of previous corroborates the inferences from the CT data pores, or both. Similarly, open pores in a smaller concerning the two distinct morphologies of Open megapores hundreds of micrometers size range (tens of micrometers) often have inclusions. Figure 8A shows a euhedral pseudo- and larger were relatively sparse, and these the same preferred orientation as nearby inclu- morph, in which the infi lling minerals include were principally observed near the rim of the sions (Fig. 6C). With increasing HRXCT scan

A B

Figure 7. Volume renderings of pseudomorphs after dodecahedral phase based on 5.4 µm resolution data. Opaque white is REE-rich phase (probable fl orencite); semitransparent green indicates minerals with lower attenuation coeffi cients. Scale bars are 50 µm. (A) See also Animation 11 (http://dx.doi.org/10.1130/GES00908.S11); 512 × 512 pixels; 2.7 MB. (B) See also Animation 12 (http://dx.doi.org/10.1130/GES00908.S12); 1024 × 620 pixels; 4.2 MB. If you are viewing the PDF of this paper or reading it offl ine, please visit the individual doi links or the full-text article on www.gsapubs.org to view Animations 11–12.

Geosphere, October 2013 1343

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/5/1336/3344992/1336.pdf by guest on 27 September 2021 Ketcham and Koeberl

morphing minerals are present, from almost pure kaolinite in some to fl orencite mixed with other phases in others, at least some chemical components from the original phase must have migrated through the pore network. The only large metallic inclusions were observed extremely close to the outside of the specimen, and probably connected to the exterior via relatively wide open porosity. The resolution of our data is not suffi cient to image the nanoscale metallic inclusions documented by De et al. (1998), which were not observed to be in contact with any pore network. We thus infer that there are two populations of metallic particles in carbonado: Megaparticles are probably Animation 13. Individual slice animation secondary pore fi llings, whereas fully enclosed for 40 kV, 1.0 µm data set encompassing nano-inclusions are primary. The observation smaller subvolume of specimen interior. Animation 14. Individual slice animation for that the metal inclusions are confi ned to the very The animation shows various pseudomorphs 40 kV, 1.0 µm data set encompassing smaller edge of this specimen suggests that they may containing abundant florencite (bright subvolume of specimen interior. The anima- even be a tertiary feature acquired after the other material). Images are 1 mm wide (341 × 318 tion shows a pseudomorph containing exclu- interior inclusions had already formed. pixels ; 4.3 MB). If you are viewing the PDF sively kaolinite. Images are 1 mm wide (454 × of this paper or reading it offl ine, please visit 495 pixels; 17.8 MB). If you are viewing the Implications from Textural Observations http://dx.doi.org/10.1130/GES00908.S13 or PDF of this paper or reading it offl ine, please the full-text article on www.gsapubs.org to visit http://dx.doi.org/10.1130/GES00908.S14 A striking feature of this sample is the spa- view Animation 13. or the full-text article on www.gsapubs.org to tial coherence of the patinaed surface paralleled view Animation 14. with banding just underneath and apparent grading of pores extending throughout the specimen. This coherence implies that a single process or resolution, smaller pores become visible, and event was responsible for all of these textural are observed in every 1-µm-resolution scan we the pore network is almost fully 3-D connected, features. The patinaed surface has been inter- obtained. However, their appearance is isolated, as suggested previously based on leaching exper- preted in other carbonados as possibly repre- and the HRXCT data were not able to resolve iments (Dismukes et al., 1988). senting a fusion crust, which may be formed by directly continuous channels forming a con- However, the rhombic dodecahedral pseudo- passage through the atmosphere at high veloc- nected network. The SEM imagery documented morphs almost certainly represent a phase that ity, either as an incoming meteorite (Shelkov micropores of various sizes throughout the dia- did form as an original inclusion in carbonado et al., 1997; Smith and Dawson, 1985), or as mond matrix along the cut surface. Pores from but has since been lost. Insofar as these pseudo- ejecta from an impact (Kletetschka et al., 2000). tens of micrometers down to <1 µm width can be morphs are many times larger than the diamond However, we consider it unlikely that either of seen, and there were no broad areas that lacked crystallites comprising their boundaries, we these mechanisms would result in the observed pores at some level. Only the megapores were consider it unlikely that they could be some form bulk rearrangement of the internal volume of the observed to have other minerals within them. of negative diamond crystal; we also consider it diamond material, extending several millimeters unlikely that they are themselves resorbed dia- inwards, i.e., much thicker than typical fusion DISCUSSION monds. The pristine faces and corners of the crusts. For this to occur, there would have to be a dodecahedra, combined with their presence transfer of carbon from the rim several millime- Inclusions throughout the sample volume, strongly suggest ters into the interior through the confi ned pore that they formed coevally with the diamond, network, and recrystallization there as diamond, The polymineralic nature and disequilibrium and moreover did not suffer any resorption dur- without forming graphite, a similar physical texture common to all REE-containing inclu- ing diamond growth. Of the 30 or so minerals argument to the one that rules out an impact sions indicate that they are exclusively sec- previously identifi ed in carbonado, none has a origin (DeCarli, 1997). ondary. Florencite is often a hydrothermal altera- rhombic dodecahedral crystal form. A variety A more likely alternative is that the small- tion product of monazite, which also tends to of minerals may form rhombic dodecahedra inclusion patterns refl ect an original pore net- host high amounts of thorium and uranium. The (including, rarely, diamond or spinel), and this work, which grades from almost nonporous near occurrence of fi ssion products (Fukura et al., habit is commonly associated with garnet, a the patinaed surface to progressively more abun- 2005) and radiogenic lead in carbonado diamond common inclusion phase in mantle diamonds dant and larger pores with increasing distance matrix (i.e., Ozima and Tatsumoto, 1997) can be (e.g., Spetsius and Taylor, 2008). However, it from this surface. The connectivity of the net- attributed to recoil implantation from these inclu- also seems evident that this phase must have work implies that the pores were formed with a sions and pore fl uids carrying components that subsequently become unstable, insofar as it has common, intercommunicating mechanism, such have been mobilized. Overall, the observations in been entirely removed from this large speci- as a trapped grain-boundary fl uid phase consis- this study support the conclusion that all mega- men, and all other carbonados studied to date. tent with recent conclusions of carbonado form- inclusions in carbonado are secondary, and that Moreover, insofar as wide varieties of pseudo- ing in a fl uid medium (Ishibashi et al., 2012).

1344 Geosphere, October 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/5/1336/3344992/1336.pdf by guest on 27 September 2021 3-D petrography of carbonado

surface, which was down-facing. Also of note is A that euhedral pores and elongate irregular pores both intersect the patinaed surface, and that the patinaed surface, while relatively smooth, is still rough in the manner of an impression on rock rather than rounded. The elongation of the irregular inclusions and pores also requires explanation, as it implies that either these pores formed with this shape and orientation or were deformed by strain in the diamond matrix. We can speculatively link this texture with the two-stage carbonado growth model suggested by Petrovsky et al. (2010), in which the diamond material begins as a loose mush of 10–100 µm crystals, which is then sintered with smaller diamonds and cryptocrystalline material by a large increase in nucleation rate caused by a sudden change in some environmental factor, such as temperature B decrease, pressure increase, or crystallization of a new phase. If this mush contained bubbles of liquid, which were then sheared during a relatively sudden event with an associated pressure increase that drove pressure signifi cantly away from the diamond-graphite stability boundary, the deformation of the bubbles could be frozen in place by the rapid crystallization of microdiamonds. The contemporaneity of the shear strain and sintering provides an explanation for deformation lamellae (De et al., 1998) and high residual stress (Kagi et al., 2007) being confi ned to the smaller, later diamond population, which would be deforming even as it was crystallizing, transiently being a weaker and more mobile material, while the larger crystals remained undamaged. Another important con- straint is provided by the pristine nature of the Figure 8. Scanning electron microscope images of (A) euhedral and pseudomorphs, which show no signs of strain, (B) irregular-oriented inclusions in carbonado. Scale bars are both suggesting that the diamond medium surround- 50 µm. ing the dodecahedral phase was weaker than the dodecahedral phase itself. The two-stage hypothesis also explains the different carbon iso- Overall, the porosity in carbonado has a dis- below the patinaed surface as identifi ed by Shel- topes of the small and large diamond populations tinctive trimodal character. The fi rst two modes kov et al. (1997), which could have formed a (De et al., 2001; Petrovsky et al., 2010). Opening consist of large euhedral pores up to millime- local trap for mineral-laden fl uids, forcing them of fractures in the crystalline mush during defor- ter scale caused by dissolution of originally to precipitate their material in-place. mation and contemporaneous crystallization included minerals, and preferentially elongated If the pore network indeed bears some may also provide a mechanism for the formation pores from 0.1 up to 0.3 mm in length with genetic relation to the patinaed surface, the of columnar diamond (Rondeau et al., 2008). irregular boundaries, which appear to be isolated question arises of whether growth was toward islands of original porosity. The third is micro- or away from this surface. We favor the idea that Implications of the Present-Day to nanoscale pores that constitute the connected growth was away from it. One possible explana- Inclusion Suite network. Precipitation of secondary minerals tion is that the initial layers of carbonado could appears to be almost exclusively within the large more easily precipitate porosity free, and that as Even if the present-day mega-inclusions in pore classes, probably owing to larger spaces carbon was consumed from the fl uid, residual carbonado are secondary, the unusual abundance being more favorable due to surface energy con- fl uid remained trapped inside the pores, inhibit- of U, Th, REEs, and other incompatible elements siderations. The principal exception is within ing infi ll. An alternative possibility is that the in carbonado still calls for explanation. We sug- the bright band adjacent to the patinaed surface, large, irregular pores represent bubbles of fl uid gest that for the sake of parsimony, this explana- in which smaller pore spaces may be fi lled. This trapped in a diamond crystal mush during the tion should optimally be linked to the formation may be due to the possible presence of a locally early formation stages of carbonado, and that event. For this evidently rare diamond creation pore-free, impermeable boundary immediately these bubbles buoyantly rose from the patinaed mechanism to be followed by a similarly rare

Geosphere, October 2013 1345

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/5/1336/3344992/1336.pdf by guest on 27 September 2021 Ketcham and Koeberl

incompatible-element enrichment event in a garnet with 350 ppm U and Th/U ratio of 4 will tainly secondary, it does not necessarily follow different environment and location, millions or accumulate a dose of 4.2 × 1018 alpha decays per that the materials that comprise them are. The billions of years later, in samples now distrib- gram, which is enough to cause signifi cant swell- radiogenic elements in the fl uid may have also uted between two continents, is akin to lightning ing in zircon (Weber, 1990). However, whatever helped in diamond crystallization (e.g., Ozima striking twice. On the other hand, carbonatitic phase housed U and Th would have to grow in the and Tatsumoto, 1997), but it is not clear if such high-density fl uids trapped in fi brous diamonds highly reducing conditions implied by the metal catalysis is necessary. are frequently rich in incompatible elements nano-inclusions; Ishibashi et al. (2012) placed the Figure 9 shows a schematic summary of car- (Rege et al., 2010; Weiss et al., 2011). likely oxygen fugacity of the diamond-forming bonado formation as postulated here. In addition An obvious place to look is to the now-absent fl uid at ~3 log units below the FMQ buffer, while to incorporating all of the textural observations dodecahedral phase. If that phase was suffi - Sautter et al. (2011) placed it even lower, at ~15 in this study, we believe that it is also consis- ciently enriched in U and Th, over time it could log units below the IW buffer based on the pres- tent with the multitude of observations by other become unstable due to metamictization from ence of metallic Ti. authors discussed in the foregoing text. radiation damage, causing the crystal structure to We thus postulate that the carbon-supersatu- Left unspecifi ed in this narrative is whether collapse and new minerals to form in its place. rated fl uid from which the diamond grew was the carbonado formation process took place in As one possibility, garnet is capable of hosting also enriched in incompatible elements, perhaps the crust or the mantle. Most evidence points high concentrations of U, Th, and REEs in hydro- in the manner of a pegmatite from the fl uid to a mantle origin, as the known processes that thermal conditions; Smith et al. (2004) document residuum of a crystallizing carbonatitic magma catalyze diamond formation at crustal pres- local U concentrations of up to 358 ppm and body, albeit in a possibly very unusual setting, sures either include features not observed in REE concentrations up to 4724 ppm in zoning and that these elements partitioned into the carbo nado or do not explain features that are bands within skarn garnets from the Beinn an dodecahedral phase, and perhaps others that we observed (Petrovsky et al., 2010); an alternative Dubhaich granitic aureole in Skye, Scotland, are not able to identify from the pseudomorphs. possibility is a transient event producing man- which they attributed to periods of closed-system The enrichment may have also included other tle conditions in the deep crust (Sautter et al., crystallization. Garnet retains all fi ssion damage elements that are also represented in the 2011). However, the positing of a deep liquid at temperatures below 200 °C (Haack and Potts, present-day carbonado inclusion suite, such as so highly enriched in incompatible elements as 1972), and this ability to accumulate radiation K, P, and Al, with the fi rst now helping comprise to allow precipitation of millimeter-scale grains damage without annealing facilitates metamicti- the kaolinite, the second the fl orencite, and so of a U-, Th-, and light (L) REE–rich phase is zation. After 1.7 b.y., the age of some conglom- on. In other words, even though the present-day unconventional. A possible compensating cir- erates bearing carbonados (Sano et al., 2002), a inclusion minerals in carbonado are almost cer- cumstance is that carbonado formation occurred

100 μm

A Accumulation B Deformation and sintering C Redistribution of included material

Figure 9. Schematic illustration of model for formation and evolution of carbonado diamond based on textural observations in this study. (A) In a fl uid-fi lled cavity associated with a crystallizing magma, crystallizing solids including 10–100 µm diamond crystallites (orange circles), the dodecahedral phase (pink hexagons), and assorted nano-inclusions settle to the fl oor, trapping fl uid between mineral grains and occasionally as bubbles in the process (white ovals). While the assemblage remained a mush, the bubbles would tend to slowly rise due to positive buoyancy. (B) A shearing event then deforms the bubbles, but not the dodecahedra, while catalyzing crystallization of micrometer-scale diamond from the trapped fl uid (light orange shading). (C) Over the subsequent billions of years, once the carbonado has been emplaced in a near-surface environment, accumulation of radiation damage from U and Th breaks down the dodecahedral phase, and material migrates through the pore network to form new minerals that are stable at crustal conditions (mottled ovals and hexagons).

1346 Geosphere, October 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/5/1336/3344992/1336.pdf by guest on 27 September 2021 3-D petrography of carbonado

relatively early in the evolution of Earth, possi- line diamond: A contribution to the carbonado conun- McCall, G.J.H., 2009, The carbonado diamond conundrum: drum: Earth and Planetary Science Letters, v. 164, Earth-Science Reviews, v. 93, p. 85–91, doi:10.1016 bly at 3.8 Ga or even earlier, at a time when the p. 421–433, doi:10.1016/S0012-821X(98)00229-5. /j.earscirev.2009.01.002. segregation of incompatible elements into the De, S., Heaney, P.J., Vicenzi, E.P., and Wang, J., 2001, Nadolinny, V.A., Shatsky, V.S., Sobolev, N.V., Twitchen, crust was far from complete (Rollinson, 2006). Chemical heterogeneity in carbonado, an enigmatic D.J., Yuryeva, O.P., Vasilevsky, I.A., and Lebedev, V.N., polycrystalline diamond: Earth and Planetary Sci- 2003, Observation and interpretation of paramagnetic ence Letters, v. 185, p. 315–330, doi:10.1016/S0012 defects in Brazilian and Central African carbonados: CONCLUSIONS -821X(00)00369-1. The American Mineralogist, v. 88, p. 11–17. DeCarli, P.S., 1997, More on the possibility of impact origin Ozima, M., and Tatsumoto, M., 1997, Radiation-induced High-resolution X-ray CT enables “3-D of carbonado: American Institute of Physics Confer- diamond crystallization: Origin of carbonados and its petrography” to reveal a number of previously ence Proceedings, v. 429, p. 681–683. implications on meteorite nano-diamonds: Geochimica Deines, P., 2002, The carbon isotope geochemistry of mantle et Cosmochimica Acta, v. 61, no. 2, p. 369–376, doi: unrecognized features in a large specimen of xenoliths: Earth-Science Reviews, v. 58, p. 247–278, 10.1016/S0016-7037(96)00346-8. carbonado diamond. These include a grading doi:10.1016/S0012-8252(02)00064-8. Petrovsky, V.A., Shiryaev, A.A., Lyutoev, V.P., Sukharev, Dismukes, J.P., Gaines, P.R., Witzk, H., Leta, D.P., Kear, A.E., and Martins, M., 2010, Morphology and defects of oriented inclusions that is spatially associated B.H., Behal, S.K., and Rice, S.B., 1988, Demineraliza- of diamond grains in carbonado: Clues to carbo- with the patinaed surface, and the occurrence tion and microstructure of carbonado: Materials Sci- nado genesis: European Journal of Mineralogy, v. 22, of dodecahedral pseudomorphs refl ecting the ence and Engineering, ser. A, v. 105–106, p. 555–563, p. 35–47, doi:10.1127/0935-1221/2010/0022-1978. doi:10.1016/0025-5416(88)90743-4. Rege, S., Griffi n, W.L., Pearson, N.J., Araujo, D., Zedgeni zov, fi rst syngenetic inclusion phase above the nano- Fitzgerald, C.B., Venkatesan, M., Douvalis, A.P., and Coey, D., and O’Reilly, S.Y., 2010, Trace-element patterns of meter scale recognized in carbonado, albeit in J.M.D., 2006, Magnetic properties of carbonado dia- fi brous and monocrystalline diamonds: Insights into its absence. From these observations, we build monds: Journal of Magnetism and Magnetic Materials, mantle fl uids: Lithos, v. 118, p. 313–337, doi:10.1016 v. 300, p. 368–372, doi:10.1016/j.jmmm.2005.05.019. /j.lithos.2010.05.007. upon recent work proposing that carbonado Fukura, S., Nakagawa, T., and Kagi, H., 2005, High spatial Rollinson, H., 2006, Crustal generation in the Archean, in grew from a C-supersaturated fl uid, and we pro- resolution photoluminescence and Raman spectroscopic Brown, M., and Rushmer, T., eds., Evolution and Dif- measurements of a natural polycrystalline diamond, car- ferentiation of the Continental Crust: Cambridge, UK, pose that the irregular pores may correspond to bonado: Diamond and Related Materials, v. 14, p. 1950– Cambridge University Press, p. 173–230. trapped bubbles in a 10–100 µm loose or lightly 1954, doi:10.1016/j.diamond.2005.08.046. Rondeau, B., Sautter, V., and Barjon, J., 2008, New colum- bound diamond slurry, and that the observed Garai, J., Haggerty, S.H., Rekhi, S., and Chance, M., 2006, nar texture of carbonado: Cathodoluminescence study: Infrared absorption investigations confi rm the extra- Diamond and Related Materials, v. 17, p. 1897–1901, fabric in these pores may refl ect a strain event terrestrial origin of carbonado diamonds: The Astro- doi:10.1016/j.diamond.2008.04.006. that sparked the growth and deformation of the physical Journal, v. 653, no. 2, p. L153–L156, doi: Sano, Y., Yokochi, R., Kentaro, T., Chaves, M.L., and Ozima, smaller, <1 µm diamond population that holds 10.1086/510451. M., 2002, Ion microprobe Pb-Pb dating of carbonado, Haack, U.K., and Potts, M.J., 1972, Fission track annealing polycrystalline diamond: Precambrian Research, v. 113, carbonado together. We further postulate that in garnet: Contributions to Mineralogy and Petrology, p. 155–168, doi:10.1016/S0301-9268(01)00208-X. this fl uid may have also been highly enriched v. 34, p. 343–345, doi:10.1007/BF00373764. Sautter, V., Lorand, J.P., Cordier, P., Rondeau, B., Leroux, Haggerty, S.H., 1999, A diamond trilogy: Superplumes, H., Ferraris, C., and Pont, S., 2011, Petrogenesis of in a number of incompatible elements that were supercontinents, and supernovae: Science, v. 285, mineral micro-inclusions in an uncommon carbonado: partitioned into primary inclusion phases, which p. 851–860, doi:10.1126/science.285.5429.851. European Journal of Mineralogy, v. 23, no. 5, p. 721– subsequently broke down due to their high U Heaney, P.J., Vicenzi, E.P., and De, S., 2005, Strange diamonds: 729, doi:10.1127/0935-1221/2011/0023-2154. The mysterious origins of carbonado and framesite: Ele- Shelkov, D., Verchovsky, A.B., Pillinger, C.T., Hutchison, and Th contents. While some rough edges and ments, v. 1, p. 85–89, doi:10.2113/gselements.1.2.85. R., and Milledge, H.J., 1997, Carbonado: More clues loose ends remain, this combined model encom- Ishibashi, H., Kagi, H., Sakuai, H., Ohfuji, H., and Sumino, to a common impact origin from samples from Brazil passes a great number of the enigmatic observa- H., 2012, Hydrous fl uid as the growth media of natural and the Central African Republic: Lunar and Planetary polycrystalline diamond, carbonado: Implication from Science, v. 26, p. 1281–1282. tions that have made the origin of carbonado a IR spectra and microtextural observations: The Ameri- Smith, J.V., and Dawson, J.B., 1985, Carbonado diamond mystery for several decades, and it does so with can Mineralogist, v. 97, p. 1366–1372, doi:10.2138/am aggregates from early impacts of crustal rocks: Geology, .2012.4097. v. 13, p. 342–343, doi:10.1130/0091-7613(1985)13 a single formation mechanism and setting. In Kagi, H., and Fukura, S., 2008, Infrared and Raman spec- <342:CDAFEI>2.0.CO;2. addition, the proposed incompatible-element– troscopic observations of Central African carbonado Smith, M.P., Henderson, P., Jeffries, T.E.R., Long, J., and rich fl uid may provide a compact explanation and implications for its origin: European Journal of Williams, C.T., 2004, The rare earth elements and ura- Mineralogy, v. 20, p. 387–393, doi:10.1127/0935-1221 nium in garnets from the Beinn an Dubhaich Aureole, for the observed differences between carbo- /2008/0020-1817. Skye, Scotland, UK: Constraints on processes in a nado and other polycrystalline diamonds such Kagi, H., Sato, S., Akagi, T., and Kanda, H., 2007, Genera- dynamic hydrothermal system: Journal of Petrol- as yakutites and framesites, suggesting that, in tion history of carbonado inferred from photolumines- ogy, v. 45, no. 3, p. 457–484, doi:10.1093/petrology cence spectra, cathodoluminescence imaging, and car- /egg087. other respects, the mechanism for their origin bon-isotopic composition: The American Mineralogist, Spetsius, Z., and Taylor, L.A., 2008, Diamonds of Siberia: may have been similar. v. 92, p. 217–224, doi:10.2138/am.2007.1957. Photographic Evidence of their Origin: Lenoir City, Ketcham, R.A., 2005a, Computational methods for quantita- Tennessee, Tranquility Base Press, 278 p. ACKNOWLEDGMENTS tive analysis of three-dimensional features in geologi- Trueb, L.F., and Butterman, W.C., 1969, Carbonado: A micro- cal specimens: Geosphere, v. 1, no. 1, p. 32–41, doi: structural study: The American Mineralogist, v. 54, This research was supported in part by National 10.1130/GES00001.1. p. 412–425. Science Foundation grant EAR-0948842, and by the Ketcham, R.A., 2005b, Three-dimensional textural measure- Trueb, L.F., and De Wys, E.C., 1969, Carbonado: Natural John A. and Katherine G. Jackson School of Geo- ments using high-resolution X-ray computed tomog- polycrystalline diamond: Science, v. 165, p. 799–802, sciences and the Owen-Coates Fund of the Geology raphy: Journal of Structural Geology, v. 27, p. 1217– doi:10.1126/science.165.3895.799. 1228, doi:10.1016/j.jsg.2005.02.006. Trueb, L.F., and De Wys, E.C., 1971, Carbon from Ubangi— Foundation at the University of Texas at Austin. We Ketcham, R.A., 2006, Accurate three-dimensional measure- A microstructural study: The American Mineralogist, thank Joseph Lai for performing the laser cut, Travis ments of features in geological materials from X-ray v. 56, p. 1252–1268. Clow for helping with inclusion segmentation, and computed tomography data, in Desrues, J., Viggiani, Weber, W.J., 1990, Radiation-induced defects and amor- J. Barnes, J. Lin, and D. Smith for helpful discussions. G., and Besuelle, P., eds., Advances in X-ray Tomogra- phization in zircon: Journal of Materials Research, v. 5, We appreciate the efforts of the editors and two anony- phy for Geomaterials: London, ISTE, p. 143–148. no. 11, p. 2687–2697, doi:10.1557/JMR.1990.2687. mous reviewers in helping to improve the manuscript. Ketcham, R.A., and Carlson, W.D., 2001, Acquisition, Weiss, Y., Griffi n, W.L., Bell, D.R., and Navon, O., 2011, optimization and interpretation of X-ray computed High-Mg carbonatitic melts in diamonds, kimberlites REFERENCES CITED tomographic imagery: Applications to the geosciences: and the sub-continental lithosphere: Earth and Plan- Computers & Geosciences, v. 27, p. 381–400, doi: etary Science Letters, v. 309, p. 337–347, doi:10.1016 Cartigny, P., 2010, Mantle-related carbonados? Geochemi- 10.1016/S0098-3004(00)00116-3. /j.epsl.2011.07.012. cal insights into diamonds from the Dachine komatiite Kletetschka, G., Taylor, P.T., Wasilewski, P.J., and Hill, Yokochi, R., Ohnenstetter, G., and Sano, Y., 2008, Intragrain (French Guiana): Earth and Planetary Science Letters, H.G.M., 2000, Magnetic properties of aggregate variation in δ13C and nitrogen concentration associated v. 296, p. 329–339, doi:10.1016/j.epsl.2010.05.015. polycrystalline diamond: Implications for carbonado with textural heterogeneities of carbonado: Canadian De, S., Heaney, P.J., Hargraves, E.P., Vicenzi, E.P., and Taylor, history: Earth and Planetary Science Letters, v. 181, Mineralogist, v. 46, p. 1283–1296, doi:10.3749/canmin P.T., 1998, Microstructural observations of polycrystal- p. 279–290, doi:10.1016/S0012-821X(00)00213-2. .46.5.1283.

Geosphere, October 2013 1347

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/5/1336/3344992/1336.pdf by guest on 27 September 2021