An Investigation Into the Cause of Color in Natural Black Diamonds from Siberia

NOTES & NEW TECHNIQUES

AN INVESTIGATION INTO THE CAUSE OF COLOR IN NATURAL BLACK DIAMONDS FROM SIBERIA

Sergey V. Titkov, Nikolay G. Zudin, Anatoliy I. Gorshkov, Anatoliy V. Sivtsov, and Larisa O. Magazina

with an optical microscope have been described Black and dark gray diamonds from Siberia, previously as graphite (see, e.g., Kammerling et al., Russia, were studied by analytical scanning and 1990). Orlov (1977) suggested that finely dispersed transmission electron microscopy. Their color is graphite microparticles also occur in those black caused by the presence of dark inclusions. Unlike diamonds in which individual inclusions could some previous reports on black diamonds in not be seen with the optical microscope. which the dark inclusions were primarily graphite, Although very rare, facetable black diamonds the Siberian samples with the most intense black are found in primary kimberlite and secondary color contained predominantly magnetite inclu- alluvial deposits worldwide. Such diamonds are sions, while the dark gray diamonds most com- mined in Siberia, eastern Russia, mainly from the monly contained inclusions of hematite and native Mir kimberlite pipe. A number of companies, such iron. Moreover, the black diamonds studied exhib- as Rony Carob Ltd. (Moscow), V. N. Almaz LLC ited anomalously high magnetic susceptibility, which may serve as one criterion for determining (New Jersey), and V. B. International Corp. (New the natural origin of black color. York), cut black diamonds from Siberia for use in jewelry (see, e.g., figures 2 and 3). The aim of this article is to document the iden- atural diamonds occur in almost all colors tification of mineral inclusions in black and near- (Orlov, 1977; Hofer, 1998). Most colors— black diamonds from Siberian deposits using ana- Nincluding brown, pink, red, orange, yellow, lytical scanning and transmission electron green, blue, and violet—are caused by defects in the microscopy methods. In our previous work diamond’s crystal structure, most commonly nitro- (Gorshkov et al., 2000; Titkov et al., 2001; and ref- gen and boron impurities, vacancies, and disloca- erences therein), we used these techniques to tions (Collins, 1982; Fritsch and Rossman, 1988). In investigate microinclusions in polycrystalline dia- contrast, black and gray colors in untreated natural mond aggregates (bort and carbonado) from various diamonds are due to the presence of mineral inclu- deposits. That data expanded our knowledge of sions rather than structural defects. inclusions in natural diamonds, and demonstrated Strange as it may seem, the nature of these their mineralogical diversity. In particular, we inclusions has not been studied in detail, although found that native metals—including common Fe high-quality faceted black diamonds have attracted and rare Cr, Ni, Ag, Au, Cu, Ti, and Fe-Cr, Fe-Ni, considerable interest in the marketplace (see, e.g., Gruosi, 1999; figure 1). Data on mineral inclusions in black diamonds available in the gemological lit- See end of article for About the Authors and Acknowledgments. erature are based mainly on examination with the GEMS & GEMOLOGY, Vol. 39, No. 3, pp. 200–209. optical microscope. Black inclusions clearly seen © 2003 Gemological Institute of America

200 NOTES AND NEW TECHNIQUES GEMS & GEMOLOGY FALL 2003 Figure 1. Black diamonds have become increasingly popular in modern jewelry such as this brooch, which con- tains 992 black diamonds (total weight 54.50 ct) and 170 color- less diamonds (total weight 1.90 ct). The cen- ter diamond weighs 2.51 ct. Courtesy of Chanel Fine Jewelry, from the Collection Privée.

Ag-Au alloys—occur widely as inclusions in natu- formed using a JEOL JEM-100C transmission elec- ral polycrystalline diamonds. It is noteworthy that tron microscope (TEM) equipped with a goniome- some of these metals, most often Ni and Fe, are ter and a Kevex 5100 energy-dispersive X-ray spec- used as fluxes in diamond synthesis, and synthetic trometer, allowing the detection of elements from crystals often contain them as inclusions Na to U. We also used a JEOL JSM-5300 scanning (Muncke, 1979). electron microscope (SEM) equipped with an Oxford LINK ISIS energy-dispersive X-ray spectrometer, which permitted detection of elements MATERIALS AND METHODS from Be to U (including, most significantly, oxy- We studied four rough black diamonds (samples gen). With these two instruments, we obtained 1–3 and 6), and two rough nearly black or dark gray information on micromorphological features (from diamonds (samples 4 and 5), from which polished SEM backscattered electron images and TEM brilliants could be manufactured. Samples 1–5 images), line and volume structural defects (from were recovered from the kimberlites of western TEM images), structural parameters (from select- Yakutia and weighed 13–25 ct each; sample 6, 1.8 ed-area electron diffraction patterns obtained using ct, came from the Anabar placer deposit in north- the TEM), and chemical compositions (from ener- ern Yakutia. gy-dispersive spectra) of inclusions as small as a All six diamonds were examined with an fraction of a micron. Box A provides more informa- Olympus BX51 optical microscope, and photomi- tion on analytical electron microscopy. crographs were taken using a MIN-8 microscope In preparation for analysis, each sample was with a Pentax-MZ5N camera. crushed after being wrapped in a special thick The six diamonds were then studied with ana- paper to avoid contamination. In large fragments, lytical electron microscopy (Wenk et al., 1995; quantitative energy-dispersive analysis of inclu- Shindo and Oikawa, 2002), and the inclusions sions was performed by SEM on polished surfaces, were identified by their chemical composition and and semi-quantitative analysis was performed on structural parameters. Investigations were per- rough, unpolished surfaces that were quite flat and

NOTES AND NEW TECHNIQUES GEMS & GEMOLOGY FALL 2003 201 BOX A: ANALYTICAL ELECTRON MICROSCOPY

Analytical electron microscopy is the combination of backscattered electrons reflected from a sample sur- transmission or scanning electron microscopy (TEM face. SEM provides information on micromorpholo- or SEM) with energy-dispersive X-ray spectroscopy gy, grain and pore sizes, sample homogeneity, the (EDS). It is one of the most powerful tools available relationship between different phases, and the like. for characterizing the solid-state substances that For both TEM and SEM, the interaction of the occur as micron-sized inclusions in minerals. electron beam with a sample produces characteris- The electron microscope uses a beam of acceler- tic X-ray radiation. Measuring the intensities of ated electrons instead of visible light (as in an optical these X-rays with energy-dispersive detectors (the microscope). Electron microscopy provides for mag- basis of EDS) reveals the chemical composition of nifications of more than 500,000×, and a spatial reso- the sample. lution of 3 angstroms (or 3 × 10−10 m), because the In summary, analytical electron microscopy wavelength of an electron beam is about four orders provides unique information on micromorphologi- of magnitude shorter than that of visible light. cal features (from SEM images and TEM images), In TEM, the electron beam is focused by a mag- line and volume structural defects (from TEM netic lens through a specimen that is thin enough images), structural parameters (from electron to be transparent to the beam. In image mode, it diffraction patterns obtained using the TEM), and provides electron micrographs that reveal various chemical composition (from EDS) of samples as defects (dislocations, misoriented microblocks, small as a fraction of a micron. For more informa- microtwins, etc.) in a specimen. In diffraction tion on analytical electron microscopy, see Wenk mode, the instrument generates selected-area elec- et al. (1995), Shindo and Oikawa (2002), and refer- tron diffraction patterns, calculation of which ences therein. reveals structural characteristics. The general prin- Facilities equipped with SEM+EDS and ciple behind electron diffraction patterns is similar TEM+EDS instruments can be found at some uni- to that of X-ray diffraction patterns. Analysis of versities or research centers, particularly those spe- these patterns is very complicated, because pub- cializing in materials science, chemistry, physics, or lished lattice parameters obtained with this tech- geology. The approximate costs of these instru- nique are quite rare in the mineralogical literature, ments are typically: SEM, $100,000–$300,000; and absent from the gemological literature. The TEM, $500,000–$1,000,000; and EDS detector, analysis and calculation of electron diffraction pat- $70,000–$300,000. terns are discussed in Hirsch et al. (1977). Electron diffraction patterns may be of two types: point and ring. Point diffraction patterns are formed if the sample studied is a perfect single crystal, with each point corresponding to the reflection from a specific set of atomic planes in the lattice. Ring diffraction patterns are produced by polycrystalline aggregates, with each ring corresponding to a certain reflection from many microblocks, or grains, in various orientations (i.e., a ring consists of many points, each of which corresponds to one misoriented block). Discrete-ring diffraction patterns may arise if microblocks in the aggregate have a preferred orientation, or if the aggregate is characterized by a certain texture. Some important limitations of the TEM tech- nique in gemology are that the sample must be very thin (i.e., it must be crushed to a powder or chemi- cally etched), and the area analyzed is only about 1 Figure A-1. The scanning electron microscope used mm in diameter. for this study was equipped with an energy-disper- In a scanning electron microscope (see, e.g., figure sive X-ray spectrometer, for analysis of elements A-1), images are recorded using secondary and ranging from Be to U. Photo by M. K. Sukhanov.

202 NOTES AND NEW TECHNIQUES GEMS & GEMOLOGY FALL 2003 Figure 2. These two faceted diamonds are from the Siberian deposits in Russia, which were also the source of the samples examined for this study. The hexagon-shaped stone measures approximately 8.4 mm in diameter. Photo by M. A. Bogomolov.

oriented nearly perpendicular to the electron Figure 3. The dark gray and black diamonds in these beam. These samples were carbon coated to dis- brooches came from Siberia. The length of the brooch perse electrostatic charge in the SEM. For TEM on the right is 6.8 cm. Courtesy of Rony Carob Ltd.; analysis, the crushed diamond powder was placed photo by M. A. Bogomolov. into a test tube with distilled water, and treated ultrasonically to obtain micron-size particles. Drops of this suspension were placed onto a copper grid with an electron-transparent film and dried RESULTS under a lamp. In each suspension sample, about Samples 1–5 were irregular rounded diamonds (see, 100–200 particles were analyzed, and for each par- e.g., figure 4). Examination with an optical micro- ticle, a TEM image, an electron diffraction pattern, scope showed that they consisted of polycrystalline and an energy-dispersive spectrum were obtained. aggregates of intergrown grains from 0.5 to 5.0 mm The relative abundance of various types of inclu- in size, with parallel striations. Some of their sur- sions was estimated based on the frequency of faces showed characteristic post-growth dissolution their presence in the suspension sample. features (e.g., block sculpture, ditrigonal striations, The magnetic susceptibility of the largest frag- etc.; see Orlov, 1977). Sample 6 was a rounded, dis- ments of the six samples was analyzed with a torted crystal with octahedral habit that also dis- Kappabridge KLY-2 magnetometer. Box B provides played dissolution features (block- and drop-like information on magnetic susceptibility. hillocks; see Orlov, 1977). All six diamonds were

Figure 4. These two rough polycrystalline diamonds were among the six studied to deter- mine their cause of color. Shown are sample 2 (left, 25.26 ct) and sample 5 (right, 22.36 ct). Photos by M. A. Bogomolov.

NOTES AND NEW TECHNIQUES GEMS & GEMOLOGY FALL 2003 203 BOX B: MAGNETIC SUSCEPTIBILITY OF MINERALS

All substances, including minerals, possess magnetic two groups, but is only about 10−6–10−9 for the last. properties to some degree. These properties may be For diamagnetic minerals (e.g., diamond, quartz, characterized by magnetic susceptibility, as well as by feldspar), magnetic susceptibility is negative, being on magnetization, magnetic permeability, magnetic the order of 10−8–10−9. structure, and the like. Magnetic susceptibility (χ) is a Magnetic susceptibility may be measured with a physical value that characterizes the relationship special magnetometer, called a χ-meter (or kappa between the induced magnetization of a substance meter). In this device, a tube containing the material and a magnetic field applied. Volume magnetic sus- under examination is suspended from the arm of a ceptibility is defined by the ratio of the magnetization balance between the poles of an electromagnet. per unit volume of a substance to the strength of the When the magnetic field is activated, it will produce magnetic field applied. For most practical purposes, a certain amount of downward magnetic force, specific magnetic susceptibility is expressed—that is, which is measured by the weight necessary to com- the volume of magnetic susceptibility calculated per pensate for it. From the value of the downward mag- kilogram of the material (in units of m3¼kg). For fer- netic force and the strength of the magnetic field romagnetic (e.g., native Fe and Ni), ferrimagnetic (e.g., applied, the magnetic susceptibility of the mineral magnetite, chromite), and paramagnetic (e.g., garnet, can be calculated. mica, tourmaline) minerals, magnetic susceptibility is For more information on magnetic susceptibility, positive. Its value reaches the order of 106 for the first see O’Reilly (1984).

almost opaque, due to a combination of numerous number of microfractures also occurred. inclusions and microfractures. Both SEM and TEM images revealed that the inclusions were diverse in size, ranging from 0.1 to Inclusions. Examination of the six samples using an 100 µm. Most consisted of magnetite, hematite, and optical microscope revealed large numbers of very native iron, which were identified by their chemical small dark inclusions (figure 5), the apparent cause composition as obtained with energy-dispersive of the black or dark gray color. To the unaided eye, spectroscopy and by their structural parameters as these inclusions generally appeared to be evenly dis- calculated from TEM electron diffraction patterns. tributed throughout the sample. However, examina- 2+ 3+ tion with the optical microscope revealed that they Magnetite. Magnetite (Fe Fe2 O4) was identified in typically formed clusters or linear boundaries five of the six diamonds (samples 1–3, 5, and 6), and between individual diamond grains, where a great was predominant in the two samples that showed the darkest black color (1 and 2). SEM backscattered electron images (see, e.g., figure 6, left) revealed that mag- Figure 5. Examination with an optical microscope netite inclusions occurred as both individual irregular revealed numerous black inclusions in the sample grains (1–4 µm) and clusters of these grains, in the dia- diamonds. Photomicrograph by M. A. Bogomolov, monds themselves and along microfractures. Energy- oblique illumination; the field of view is 2 mm. dispersive spectra of these inclusions usually indicated the presence of only Fe and O (figure 6, right), as would be expected for magnetite. Some grains from sample 6 contained Mn impurities (about 8 wt.%). In TEM images obtained from suspension samples, the magnetite appeared as irregular particles with varying degrees of perfection (i.e., single crystal vs. polycrystalline aggregates; see figure 7, left). The electron diffraction patterns of most of the magnetite particles showed a series of discrete ring reflections (figure 7, right), which is characteristic of a polycrystalline aggregate with preferred orientation of grains, in sizes smaller than 1 µm. For some of the magnetite particles, the diffraction patterns exhibit-

204 NOTES AND NEW TECHNIQUES GEMS & GEMOLOGY FALL 2003 Figure 6. An SEM backscattered electron image (left) of the broken surface of one of the black diamonds studied revealed that magnetite inclusions (bright white spots) occurred as both individual grains and clusters in the diamond matrix (dark area). The energy-dispersive spectrum of such inclusions (right) typically indicated the presence of only Fe and O. The carbon (C) is from the carbon coating on the sample. ed continuous ring reflections (characteristic of poly- dant than the inclusions of Fe oxides. The native crystalline aggregates with random grain orientation) iron formed individual isometric grains and clusters and point reflections (characteristic of single crys- (figure 11, left); most were essentially free of impuri- tals). Processing of these patterns yielded diffraction ties (figure 11, right), though some showed minor characteristics typical of the magnetite structure. Cr, Mn, Ni, and/or Cu. In TEM images, native Fe was seen as irregular particles (figure 12, left). Their Hematite. In the two dark gray diamonds (samples 4 electron diffraction patterns yielded a series of dis- and 5), hematite (Fe2O3) inclusions were predomi- crete ring reflections (figure 12, right). nant. SEM images revealed that the hematite inclu- In addition to native Fe, sample 1 contained sions usually formed scale-like particles (figure 8, inclusions of native Cu, sample 2 exhibited inclu- left). Energy-dispersive spectra of these particles sions of an Fe0.9Cr0.1 alloy, and sample 3 displayed showed peaks for Fe and O, and rarely a weak peak inclusions of native Ag and native Zn. related to Mg impurities (figure 8, right). TEM images revealed that the particles were actually sin- Other Dark Mineral Phases. Inclusions of an exotic gle crystals (figure 9, left) that showed point reflec- phase consisting of Cu (55–65 wt.%), Sn (7–16 wt.%), tions in the diffraction patterns (figure 9, right); they Ni (4–8 wt.%), and O were found in samples 1 and 2. also formed rounded polycrystalline aggregates (fig- This mineral filled cracks in the black diamonds and, ure 10, left) that gave a discrete-ring type of diffrac- in some cases, formed irregular grains. Analysis of the tion pattern (figure 10, right). elemental distribution within these inclusions— which was obtained using characteristic X-ray radia- Metallic Elements. Inclusions of native iron were tion of Cu, Sn, Ni, and O—indicated that it was quite detected in all six samples, but they were less abun- homogeneous, rather than a mixture of several phases.

Figure 7. This TEM image (left) of a suspension sample from a black diamond shows that most of the magnetite inclusions consist of irregular aggregates. The electron diffraction pattern (right) of this grain shows a series of discrete ring reflections, which is characteristic of a polycrystalline aggregate with preferred orientation of the grains. The diffraction reflections of the corresponding lattice planes have been indexed.

NOTES AND NEW TECHNIQUES GEMS & GEMOLOGY FALL 2003 205 Figure 8. Hematite inclusions, which were most common in the dark gray diamonds, usually occurred as scale- like particles, as shown by the bright white grains in this SEM image (left). The energy-dispersive spectrum of the hematite (right) showed peaks for Fe and O and, rarely, Mg. The Al peak is due to interference, and the C peak corresponds to the carbon coating.

Electron diffraction patterns of this phase showed a filled microfractures and microcavities. Also detect- series of discrete ring reflections, consistent with a ed were common calcite (CaCO3) and dolomite polycrystalline structure. Processing of these electron ([Ca,Mg]CO3), as well as rare olivine ([Mg,Fe]Si2O4), diffraction patterns revealed structural characteristics anhydrite (CaSO4), fluorite (CaF2), halite (NaCl), that, combined with the chemical composition, have cuprite (CuO), quartz (SiO2), feldspars, and some not been reported previously to our knowledge. other minerals. Most of these have been previously In sample 1, small quantities of the sulfide miner- identified as inclusions in diamond (see Harris, als chalcocite (Cu2S) and pyrite (Fe2S) were detected. 1992; Gorshkov et al., 2000; Titkov et al., 2001). Graphite inclusions were extremely rare in our samples. A few scale-like graphite particles were Magnetic Susceptibility. The measured specific found, but only in samples 1 and 3. magnetic susceptibility of the dark gray and black diamonds varied over a wide range, from –0.6 × 10−8 Non-black Inclusions. A minor amount of Fe- m3/kg to +36 × 10−8 m3/kg. (By comparison, pure hydroxide minerals formed inclusions in samples 1, diamond has a specific magnetic susceptibility of 2, and 6. Calculation of their structural parameters −0.62 × 10−8 m3/kg; Novikov, 1993.) The maxi- from electron diffraction patterns identified some mum magnetic susceptibility was measured in as goethite (rhombic modification of FeO-OH) and sample 1, which also had the darkest black color. others as akaganéite (tetragonal modification of However, the relationship between the depth FeO-OH). of black color and magnetic susceptibility was Inclusions that apparently were not related to not straightforward, because our measurements of the black coloring of the diamonds were also found the volume magnetic susceptibility were influ- in the samples. Most of the diamonds contained enced by all of the various mineral inclusions in a inclusions of chrysotile (Mg3[Si2O5]OH), which sample.

Figure 9. Single-crystal particles of hematite (as shown in a TEM image, left) produced point electron diffraction patterns (right). The (001)* index shows the general orientation of the particle.

206 NOTES AND NEW TECHNIQUES GEMS & GEMOLOGY FALL 2003 Figure 10. Polycrystalline aggregates of hematite (shown in a TEM image, left) produced a discrete- ring type of electron diffraction pattern (right).

DISCUSSION observations that the inclusions were completely Our results suggest that the color of the black and isolated in diamond and were not accompanied by dark gray diamonds studied was caused primarily by fractures; furthermore, they showed growth steps inclusions of magnetite, hematite, and native Fe. characteristic of the host diamond (Prinz et al., Inclusions of Cu-based oxides detected in some sam- 1975). Inclusions of native Fe, as well as other ples also may have contributed to the black color of native metals, are rather widespread in polycrys- these diamonds. In samples of the darkest color, mag- talline and translucent diamonds from various netite was the most common inclusion. In the dark deposits worldwide (Gorshkov et al., 2000; Titkov gray diamonds, inclusions of hematite and native Fe et al., 2001; and references therein), so their appear- were predominant. Further detailed studies of black ance in the samples studied was not unexpected. diamonds from other localities are required to deter- Graphite, commonly presumed to be the cause mine whether the presence of magnetite, hematite, of black color in diamonds, was detected in very and native Fe inclusions are merely a characteristic small amounts in only two of our samples, so it feature of black and dark gray diamonds from Siberia, cannot be responsible for their coloring. Sulfide or if they are typical of all such diamonds worldwide. minerals also were very rare, and so they did not Nevertheless, magnetite has been reported as abun- have a significant impact on the color of our sam- dant inclusions in some dark gray diamond aggre- ples. It should be noted that in transparent single- gates from African deposits (Jeynes, 1978). crystal diamonds, dark sulfides are common (Harris, In previous studies, magnetite was observed as 1992), but so far they have not been documented in isolated inclusions in single crystals of transparent large enough quantities to cause dark gray or black diamond (Prinz et al., 1975; Harris, 1992). Its syn- coloration. Sulfides were absent in the opaque sin- genetic character (i.e., formed at the same time) gle-crystal black diamond studied here (sample 6). with diamond was established on the basis of The inclusions of magnetite, hematite, and

Figure 11. Most native-iron inclusions (bright white in the SEM image on the left) were essentially free of impurities, as shown by the energy-dispersive spectrum on the right. The Al peak is due to interference, and the C peak corresponds to the carbon coating.

NOTES AND NEW TECHNIQUES GEMS & GEMOLOGY FALL 2003 207 native Fe apparently contributed to the unusual about 150 km, and then were carried to the surface magnetic properties of the diamonds studied. The by kimberlite or lamproite magma (Orlov, 1977; wide range of values we measured was probably due Kirkley et al., 1991; Harris, 1992). According to this to the variable content of magnetic inclusions, and theory, the occurrence of the inclusions revealed in to the presence of numerous inclusions of other our samples may be explained only as a result of minerals. In previous work (Jeynes, 1978), diamond post-growth formation within fractures in the host polycrystalline aggregates (e.g., bort) that could be diamonds. However, most of the black inclusions separated from other diamonds by an electromagnet were entirely surrounded by the diamond matrix, were referred to as stewartite. indicating that they were incorporated into the host The unusual magnetic properties of these natural- diamonds during growth. Previous work has also color black diamonds may be useful in separating established that magnetite and native-iron inclu- them from their artificially colored counterparts. sions in diamond can be syngenetic (Prinz et al., Black may be produced by irradiation with neutrons 1975; Harris, 1992; Gorshkov et al., 2000; Titkov et or gamma rays in a nuclear reactor (Collins, 1982), by al., 2001). Furthermore, these black inclusions were ion implantation in a linear accelerator (Moses et al., abundant in our samples, while inclusions of sili- 2000), and by heating in a vacuum to cause internal cate minerals that are typical of mantle rocks (e.g., graphitization (Notari, 2002). None of these processes garnet, olivine, pyroxene) were nearly absent— induces magnetism, and the relatively inclusion-free despite the fact that silicate inclusions should be natural diamonds that are irradiated to produce a present in polycrystalline diamonds if they actually black color are not magnetic. Therefore, magnetic crystallized from silicate melts. properties could provide an important criterion for The abundance of magnetite, hematite, and identifying natural black coloration. It is important to native-iron inclusions in the diamonds that we stud- note that anomalously high magnetic susceptibility ied might be better explained by another mechanism is also characteristic of synthetic diamonds grown of diamond formation. This hypothesis assumes that from Fe melts and containing metallic flux inclu- diamond crystallized as a result of the injection of sions (Novikov, 1993). One magnetic synthetic deep-derived hydrocarbon-bearing fluids into more- “black” diamond was described by Reinitz (1999), oxidized rocks of the lithosphere and their subse- but its color was actually very dark blue, caused by quent interaction (Taylor and Green, 1989; Navon, abundant boron impurities rather than inclusions. 2000). Most of the inclusions in our samples, in par- It is interesting to speculate that the weak mag- ticular native Fe and magnetite, might be products netism of these natural black diamonds may be of redox (oxidation-reduction) interactions between responsible for the positive health benefits ascribed hydrocarbon fluids and lithospheric rocks (eclogite to black diamonds in ancient times. and peridotite). In the process of these interactions, The inclusions in natural black and dark gray the redox potential could change dynamically, caus- diamonds are also interesting when we consider the ing the formation of inclusions representing very dif- genesis of these diamonds in the earth. The widely ferent redox conditions, for native metals on the one accepted hypothesis is that most gem-quality dia- hand and for oxide and carbonate minerals on the monds crystallized from silicate mantle melts (of other. Understanding the occurrence of these inclu- eclogite and peridotite compositions) at a depth of sions requires further detailed study.

Figure 12. In TEM images (e.g., on the left), native Fe often appeared as irregular particles. Their electron diffraction patterns (e.g., right) gave a series of discrete ring reflections.

208 NOTES AND NEW TECHNIQUES GEMS & GEMOLOGY FALL 2003 SUMMARY AND CONCLUSION Black diamonds, cut and rough, continue to gain pop- ularity in contemporary jewelry (figure 13). Yet there has been little detailed research to date on the cause of color in these unusual gems. This investigation by analytical electron microscopy has shown that the coloring of the black and dark gray diamonds studied was caused primarily by the presence of inclusions of magnetite, hematite, and native Fe. The diamonds with the darkest black color contained predominantly magnetite inclusions, while the dark gray samples most commonly contained inclusions of hematite and native Fe. These mineral inclusions are responsible for the anomalously high magnetic susceptibility of our samples, which may prove to be of value as a criterion for the natural origin of black color.

Figure 13. Even rough black diamonds have found a niche in contemporary jewelry. This ring, set with a 5.8 ct black diamond, is designed by and courtesy of Pam Keyser.

ABOUT THE AUTHORS ACKNOWLEDGMENTS Dr. Titkov ([email protected]) is senior research scientist, Prof. The authors thank Dr. J. E. Shigley, of GIA Research, Carlsbad, Gorshkov is principal research scientist, Mr. Sivtsov is research for substantially improving the manuscript. We are also grateful to scientist, and Mrs. Magazina is research scientist at the E. S. Grigorian of the All-Russia Institute of Mineral Resources for Institute of Geology of Ore Deposits, Petrography, Mineralogy, magnetic susceptibility measurements, and Dr. M. A. Bogomolov and Geochemistry, of the Russian Academy of Sciences in for taking the photos. This work was supported by grant No. 01- Moscow. Dr. Zudin is director of Rony Carob Ltd. in Moscow. 05-64457 from the Russian Foundation of Basic Researches.

REFERENCES monds. The Loupe, Vol. 9, No. 4, pp. 16–19. Collins A.T. (1982) Color centers in diamond. Journal of Gemmology, Navon O. (2000) Diamond formation in the earth’s mantle. In J.J. Vol. 18, No. 1, pp. 37–75. Gurney et al., Eds., Proceedings of the 7th International Fritsch E., Rossman G.R. (1988) An update on color in gems. Part 2: Kimberlite Conference, Vol. 2, Cape Town, South Africa, pp. Colors caused by band gaps and physical phenomena. Gems & 584–604. Gemology, Vol. 24, No. 2, pp. 81–102. Notari F. (2002) Black diamond treatment by “internal” graphitiza- Gorshkov A.I., Bao Y.N., Titkov S.V., Ryabchikov I.D., Magazina tion. Revue de Gemmologie, Nos. 145/146, pp. 42–60. L.O., Sivtsov A.V. (2000) Composition of mineral inclusions and Novikov N.V., Ed. (1993) Physical Properties of Diamond. Ohsha formation of polycrystalline diamond aggregates (bort) from the Publishers Ltd., Tokyo, 240 pp. Shengli kimberlite pipe, China. Geochemistry International, Vol. O’Reilly W. (1984) Rock and Mineral Magnetism. Blackie, Glasgow. 38, No. 7, pp. 698–705. Orlov Yu. L. (1977) The Mineralogy of the Diamond. John Wiley & Gruosi F. (1999) The Black Diamond. De Grisogono, Geneva, 72 pp. Sons, New York, 235 pp. Harris J.W. (1992) Diamond geology. In J. E. Field, Ed., The Properties Prinz M., Manson D.V., Hlava P.F. (1975) Inclusions in diamonds: of Natural and Synthetic Diamonds, Academic Press, London, pp. Garnet lherzolite and eclogite assemblages. Physics and 345–393. Chemistry of the Earth, Vol. 9, pp. 797–816. Hirsch P.B., Howie A., Nicholson R.B. (1977) Electron Microscopy of Reinitz I. (1999) Gem trade lab notes: Synthetic diamond, black. Thin Crystals. Robert E. Krieger, Malabar, FL, 563 pp. Gems & Gemology, Vol. 35, No. 2, p. 138. Hofer S.C. (1998) Collecting and Classifying Coloured Diamonds. Shindo D., Oikawa T. (2002) Analytical Electron Microscopy for Ashland Press, New York. Material Science. Springer-Verlag, Berlin, 152 pp. Jeynes C. (1978) Natural polycrystalline diamond. Industrial Taylor W.R., Green D.H. (1989) The role of reduced C-O-H fluids in Diamond Review, No. 1, pp. 14–23. mantle partial melting. In J. Ross, Ed., Kimberlites and Related Kammerling R.C., Kane R.E., Koivula J.I., McClure S.F. (1990) An Rocks—Proceedings of the 4th International Kimberlite investigation of a suite of black diamond jewelry. Gems & Conference, Vol. 1, Blackwell Scientific Publishers, Melbourne, Gemology, Vol. 26, No. 4, pp. 282–287. pp. 592–602. Kirkley M.B., Gurney J.J., Levinson A.A. (1991) Age, origin, and Titkov S.V., Gorshkov A.I., Vinokurov S.F., Bershov L.V., Solodov emplacement of diamonds: Scientific advances in the last decade. D.I., Sivtsov A.V. (2001) Geochemistry and genesis of carbonado Gems & Gemology, Vol. 27, No. 1, pp. 2–25. from Yakutian diamond deposits. Geochemistry International, Muncke G. (1979) Physics of diamond growth. In J.E. Field, Ed., The Vol. 39, No. 3, pp. 228–236. Properties of Diamonds, Academic Press, London, pp. 473–499. Wenk H.-R., McLaren A.C., Pennock G.M. (1995) Electron Moses T.M., Reinitz I.M., Koivula J.I., Buerki P.R., McClure S.F., microscopy of minerals. In A.S. Marfunin, Ed., Advanced Shigley J.E. (2000) Update on the new treated black and green dia- Mineralogy, Vol. 1, Springer-Verlag, Berlin, pp. 264–298.

NOTES AND NEW TECHNIQUES GEMS & GEMOLOGY FALL 2003 209