A Preliminary Gemological Study of Synthetic Diamonds Thin Films

NOTES AND NEW TECHNIQUES

A PRELIMINARY GEMOLOGICAL STUDY OF SYNTHETIC DIAMOND THIN FILMS By Emmanuel Fritsch, Laurie Conner, and John I. Koivula

To evaluate possible problems in separating fern ina- from a carbon-containing, hydrogen-rich gas terials coated with a synthetic diamond thin film (Angus and Hayman, 1988). Variations of this from their iincoated equiv~~lents,we examined one technique have been used successfully to grow sample of a free-standing, polycrystalliiie diamond both synthetic diamond films and a new class of film about one micron (0.001 mm) thick. This film materials called diamond-like carbon (DLC), in could be easily recognized by standard microscopic which the carbon atoms are present in both dia- examination, especially in surface-reflected and polarized light. The potential problems related to de- mond-type and graphite-type bonding arrange- tecting thicker, monocrystolline, or colored synthetic ments. Although the physical and chemical prop- diamor~dthin films and coatings that inight be en- erties of the diamond-like materials can be no- countered in the future 011 faceted stones are also ticeably different from those of diamond, they briefly discussed. seem to be easier to manufacture than synthetic diamond and yet offer many of its advantages. This new growth technique suggests the possibility that a synthetic diamond or DLC coating several microns thick could be placed on various materials One of the most significant developments in to improve their resistance to wear, enhance some gemology in the 1980s has been the commercial other property of the material, or even create new growth and production of gem-quality synthetic types of devices for many useful technological diamonds. All of the samples - both commercially applications. grown and those for research only-described in the gemological literature to date, however, have been synthesized using high pressure and tempera- ABOUT THE AUTHORS ture. Because of this mode of formation, they Dr. Frilsch is research scientist, and Mr. Koivula is chief gem- exhibit characteristic features that allow a skilled ologist, at the Gemological Institute of America, Santa Monica, California. Ms. Conner, who has advanced degrees in both civil gemologist to separate them from natural diamond engineering and business, is director of marketing and sales for by means of conventional testing techniques Crystallume, Menlo Park, California. (Crowningshield, 1971 ; Koivula and Fryer, 1984; Acknowledgments The authors thank Dr Michael Pinneo, Shigley et al., 1986 and 1987). Crystallume's co-founder, vice-president, and chief scientist, for Within the past few years, a completely differ- providing extensive technical data and sharing his personal experience Dr G R Rossman helped lake the total ent type of process has been developed to grow a reflectance infrared absorption spectra of the synthetic thin layer of synthetic diamond on a substrate at diamond film and explain the origin of the coloration Dr James low pressure and moderate temperature. Based on Shigley's constructive comments are much appreciated This work was supported in part by a grant from the Harry Winston the well-known technique of chemical vapor depo- Research Foundalion sition (CVD), this new tecl~nologyinvolves the Gems & Gemology, Vol. 25, No 2, pp 84-90 deposition of tetrahedrally bonded carbon atoms 0 1989 Gemological Institute of America

84 Notes and New Techniques GEMS & GEMOLOGY Summer 1989 Figure 1, The synthetic diamond thin film studied measures approx- irnately 6.8 x 6.8 mm. One of the reasons interference colors are evi- dent here is the fact that the thiclzness (1 y.) of this film is of the same order of magnitude as the wavelengths of visible light. Note the warped aspect of the free-standing part of the film, which measures about 4 mm in diameter. Without the tension imposed by the silicon base, the film relaxes and wrinkles. Photomicrograph by John I. Koivula; magnified 6 x ,

With few exceptions, the synthetic diamond and DLC films grown today are polycrystalline; that is, they are made up of many tiny crystallites. To date, only limited experimental work has been done (e!g.; by Sumitomo Electric Industries) to deposit.single-crystal synthetic diamond films on a substrate of a crystal of synthetic diamond (Hara and Fujimori, 1989). Even so, there has been some concern that this technique could be used in the jewelry industry, for example to coat faceted stones of various diamond simulants so that they would show a high thermal conductivity and might then test as diamond on a thermal inertia probe. These films Figure 2. The thin film studied was grown from might also show some other gemological proper- this gas plasma, using a technique known as ties that would make the simulant behave more plasma-enhanced chemical vapor deposition like a diamond. One jeweler has already reported (PECVD). Photo courtesy of Crystallume. encountering a diamond simulant that was al- legedly coated in this manner (Koivula, 1987). light, as would be done if it were deposited on a To investigate the potential problems that transparent substrate such as a diamond simulant. such a product might pose, we examined a poly- This thin film was tested by classic gemologi- crystalline synthetic diamond thin film, about one cal methods and instrumentation to document micron (1 k) thick, that had been deposited on a ways by which it could be detected. The results are silicon substrate (figure 1). The particular tech- reported below. Although some background on the nique used to grow this thin film, plasma-en- history of synthetic diamond thin films is pro- hanced chemical vapor deposition (PECVD; see vided, this article will not present a detailed figure 2), is one of the most common methods explanation of the various and sometimes complex currently used to grow such materials. The sample growth processes used to produce such films. we studied was manufactured by Crystallume of Readers interested in learning more about this Men10 Park, California. The center part of the technology and its expected industrial applica- silicon square was chemically etched away, so that tions outside of gemology are referred to DeVries part of the thin film would be free standing. This (1987), Spear (1987), Angus and Hayman (1988), allowed us to exanline the film by transmitted and Bachmann and Messier (1989).

Notes and New Techniques GEMS & GEMOLOGY Summer 1989 85 HISTORY OF film production to date has focused on the many DIAMOND THIN FILMS industrial, scientific, and military applications, It is difficult to tell from the published scientific organizations with significant interest in the gem literature who was first to grow synthetic diamond market worldwide are following this new technol- at low pressure. W. Eversole received a patent in ogy closely. Kyocera, a Japanese company well 1962 to produce diamond by a low-pressure pro- known for the many synthetic gem materials cess, but his methodology is very different from grown by its subsidiary Inamori, has submitted the techniques used presently (Eversole, 1962).The several patents on the manufacture of synthetic low-pressure CVD diamond technology in greatest diamond by CVD, including one claiming the use today got its start in the Soviet Union in the production of a 1.2-ct "inclusion-free, flawless and early 1960s. Not until 1968, however, did crystal blemish-free" single crystal (Hiroshi, 1986).When growth scientist Boris Derjaguin and his co- contacted, Kyocera admitted that "work in this workers report their success in depositing syn- area is still in the developmental stage." De Beers thetic diamond at low pressure (Derjaguin et al., is also "closely monitoring developments in the 1968). Although many Western scientists were CVD area" (De Beers, 1988). aware at that time of Derjaguin's work, they Why all the fuss? Until now, diamond's unique questioned the validity of the material's identifica- combination of superior hardness, very high ther- tion as diamond. mal conductivity, and optical transparency was In the late 1970s and early 1980s, however, unavailable in the form of a film or coating over Japanese researchers indicated that they had dupli- large surface areas on a three-dimensional object cated Derjaguin's earlier work, and proved that the with good adherence to the substrate. Although deposition of a synthetic diamond thin film at low adherence problems have not been solved com- pressure was technically feasible (see, e.g., Mat- pletely, this new CVD process does allow synthetic sumoto et al., 1982 a and b; Kamo et al., 1983). diamond to be deposited on a variety of surfaces in These reports made a major impact in the world of thin coatings, thereby opening up a wide variety of materials science. Many organizations worldwide, applications, including the possibility of a syn- such as the Royal Signal & Radar Establishment in thetic diamond coating on diamond sim~ilantsor the United Kingdom (Lettington, 1988) and com- other gem materials. panies like Sumitomo, Kobe Steel, Fujitsu, and Toshiba in Japan, started research and then produc- A BRIEF DESCRIPTION OF tion of synthetic diamond or DLC thin films. In THE CVD PROCESS 1988, for example, approximately US$20 million A schematic illustration of a typical CVD diamond was spent on CVD synthetic diamond research in reactor is given in figure 3. Quite simply, the Japan (M. Yoder, pers. comm., 1989). process involves the use of hydrocarbon (typically Individual researchers in the U.S. began work methane) and hydrogen gases and a source of on growing tiny synthetic diamond crystals by the energy. The energy is most commonly derived CVD method in the 1960s (Eversole, 1962; Angus from microwave, radio frequency (RF), or direct et al., 1968). Not until the early 1980s, however, current (dc) sources, or from a heated tungsten was a major collective effort directed toward the filament. This energy serves to ionize the input growth of synthetic diamond and DLC thin films. gases, creating a carbon-rich gas "plasma" of Prominent organizations such as General Electric, charged particles. Typically, a pressure of 50 to 150 IBM, Pennsylvania State University, the Massa- Torr and a temperature of 700Â to 1000Â° are chusetts Institute of Technology, and the Naval needed for a gas composition of 0.1 to 1 vol.% Research Laboratory, among others, began signifi- methane (CH4)in hydrogen. Under these condi- cant research programs in this area, supported for tions, the hydrocarbon gas breaks down, and car- the most part by grants from the Department of bon atoms are deposited on the substrate material Defense Strategic Defense Initiative ("Star Wars"). as a thin film of synthetic diamond. Atomic One company, Crystallume, was created in 1984 hydrogen is believed to be necessary near the for the sole purpose of developing "engineered" deposition area to maintain the carbon atoms in products coated with synthetic diamond or dia- the proper type of tetrahedral bonding and thereby mond-like thin films. minimize the formation of graphite. The resulting Although most of the interest in diamond thin coated surface can have all the inherent properties

86 Notes and New Techniques GEMS & GEMOLOGY Summer 1989 ing in air because it gives rise to interference colors (lilze an oil sliclz on water-again! see figure 1). This results from factors such as the contrast in index of refraction between diamond and the underlying silicon (or air! for the free-standing part! and a film thiclzness close to the wavelengths of visible light. Also! the smoothness of the silicon- diamond film interface might have an influence. Haziness. In darlzfield illuminatioi~~the thin film appears to be hazy rather than transparent (figure 41. Because the film is polycrystalline and the individual synthetic diamond crystallites are uni- formly approximately 1 p. (again! comparable in size to wavelengths in the visible range)! the material scatters visible light. This scattering results in a hazy appearance. Color. When viewed against a white baclzground in diffused light! this film has a brownish appearance that is barely visible to the nalzed eye (figure5). We could not detect any absorption features with a hand-held spectroscope. However! the optical absorption spectrum obtained by transmission mode with a Pye Unicam 8800 spectrophotometer shows a gradual rise in absorption toward the ultravioletl . Figure 3. This schematic diogrl~mill~~strotcs the which is what would be expected from a brownish PECVD p

Fig~ire4. In dorlzficld illiln~inotion,th~s syrl- of diamond. DLC films are produced with varia- tl~etzcdiamond film is not tronsporent b~zt tions of CVD technology that use a higher methane somewl~a~hazy. This is due to light scattering concentration in the gas! a much lower tempera- at the grain 60~1nclariesof this polycrystalline ture (down to room temperature)! or ion beam material. Photomicrograph by Io11n I. I

GEMOLOGICAL EXAMINATION OF THE DIAMOND THIN FILM The film we examined was deposited on a square ' silicon substrate that is 6.8 mm on each edge and 0.55 mm thick. The central free-standing portion is about 4 mm in diameter (again! see figure I). This film (Crystallume sample EASI 23K)! approximately 1 p thick! was deterlnined by Raman spectroscopy to be crystalline diamond with one part per thousand of a "gra~hitic~~component. It wo~~ld,therefore! be readily considered a syilthetic diamond film rather than a diamond-lilze carbon film [see Williams and Glass! 19891.

General Appearance. The film can be easily recognized either on the silicon substrate or free stand-

Notes and New Tcchniq~ics GEMS & GEMOLOGY Slimnier 1989 87 Nomarski Differential Interference Contrast Microscopy. This ex;~mii~ationtechniq~~e (Mc- Crone et al.) 19781 l~iglllightsvery small details on a surface! down to 1 nm or so) by means of different interference colors (thus the orange color of the illustrations in figure 6). Not only is the granular nature of the thin film readily visible wit11 this technique (figure 6) left)! but one can actiially measure the approximate diameter of the individual grains! 1.25 p. on average. The surface of the table of a natural polished diamondl examined wit11 t11e same techniq~ieat the same magnification, has a very different appearance (figure 6) right]. The few lines visible are s~irfacegraining) features inherent to the ~inderly- ii~gstructure of this particular diamond. Fig~l 'he17 viewed against a 'te bacl<- Luminescence. No l~iminescenceto either short- gro~~ndin cllffused light, this synthetic dia- wave or long-wave ~iltravioletradiation was ob- mond thin film appears light browz~.Photo- served on this particular film. micrograph by \oh11 I. Koivufa; magnified 3 x. Thermal Conductivity. Using a GIA-GEM Di~otes- ter) we tested several spots on tlle synthetic dia- resulting from the peculiar electronic structure of mond thin film for thermal conductivity. The the minor ltgraplliticll component (Koidl! 1989). readings on the diamond tester fell in the range 60 Also! a thin-film interference phenomenon could to 80 on the scale! which is at tlle lower end of the give rise to a brown interference color evenly green zone labeled "diamo~~d."However! the sili- distributed on the film (as is commonly observed con substrate gave the exact same readings. Tllere- on camera lenses with optical coatings). fore) we believe that the film is too thin to test as diamond with this instrument (on which the val~ie Reaction to Polarized Light. When examined be- should really be between 90 and 1lo), and that the tween crossed polarizer~~tlle thin film is not measured readings actually correspond to the extinct, Altho~igl~diamond is singly refractivel an thermal cond~ictivityof the substratel since dia- aggregate reaction is shown by polycrystalline thin mond has a higher thermal conductivity than films because of light scattering at the grain silicon. This film (at this thiclzness] could not be boundaries and also possibly strain. ~isedto ttdisguisella diamo~ldsimulant.

Figure 6. With Nomarski differential interference contrast, the graininess of the surface of the synthetic diamond thin film is readily apparent (left). The only observable features on the surface of the table of

88 Notes a11d Ne\v Techlliq~~es GEh4S t3 GEMOLOGY SLIIIII~C~1989 Fig~lre7. Thin-filn~tec~iiiology has advo~lcedto tlie point of producing this 500-p (0.5-mm) thick transparent sytithet~cciiai?iolld film. Bec(1~1se this filni hos been polished 011 one side onlx there is some residual haziness ond color caused by inip~l- rities from tlie substrote on the un- polished s~irfoce,Photo courtesy of Thotnas A~ithonj~Gelierol Elec- tric co.

Electrical Conductivity. Two steel electrodes were slice-has been grown on a silicoi~substrate (fig- applied to different points of the film still on the ure 7). While this material currently is not com- substrate to test for electrical conductivity. A very mercially available' it does indicate the speed with wealz conductivity was detected in certain areas which this techiiology is progressi~~g. (readings' (round 1 or 2 compared to 130 when the The most identifiable feature of the syi~thetic two electr'cides are put in contact with each other). diamond thin film we studied is its granular This might be due to minute amounts of a more textureI the result of its polycrystalline nature. "graphitic'" phase at the grain boundariesl as graph- What ifl one might l~ypothesi~e~someone devel- ite is a good electrical conductor, oped a monocrystalline syi~theticdiamond film thiclz enough to give the thermal cond~ictivityof DISCUSSION diamond that co~~ldbe deposited 011 a transparent From this restricted set of observationsl one can substrate? There should be no haziness and no still predict that as long as a synthetic diamond or aggregate reaction in polarized light. The case DLC film grown by the CVD technique is thin! the most feared by the gem trade would be a mono- thermal tester remains the fastest instrument to crystalline film on cubic zirconia. It is highly separate diamond from a synthetic diamond- improbablel howeverl that such a synthetic dia- coated simulant. According to Michael Pinneo mond film wo~~ldadhere to this particular subs- (pers. coinm.l 1988)/the film wo~~ldhave to be at trate! since the crystal structures of the two least 5 p thiclz to test as diamond with the diamond materials are very different. Although similar thermal probe. This thiclzness requirement could adhesion problems also exist with polycrystalline be even greater; the fact that the film is poly- thin filnls! an intermediate layer of DLC as thin as crystalline makes its thermal behavior very un- 10 nm could be used to improve adhesionl as it has predictable. Currently' this behavior is even more in a variety of industrial applications requiring a uncertain for DLC. nonsilicon substrate. Interference colors also readily reveal the pres- Another possible application of a synthetic ence of a very thin film such as that examined for diamond thin film would be to modify the appar- this st~~dy!although they disappear on a film 3-4 p ent color of a faceted dianlond. S~~initomoElectric thiclz (M. Pinneo, pers. comm.' 1988). General Industries is already able to deposit a blue syn- Electric's Thomas Anthony announced at the thetic diamond coating as thiclz as 20 p on a April 1989 Materials Research Society Short natural near-colorless diamond octahedron (N. Course on Diamond and Diamond Thin Films that Fujimoril pers. comm.! 1989). The resulting blue a transparentl near-colorless polycrystalline syn- product is electrically conductive, One can imag- thetic dian~ondthin film 500 p thiclz - one-third ine that this type of coating might be easy to detect the thiclzness of a Sumitomo synthetic diamond by in~mersionof the diamond in methylene iodide.

Notes and New Techniques GEMS & GEMOLOGY S~~~nnier1989 89 The color in a natural blue diamond is distributed add a point of weight to push a 0.99-ct stone over in a slightly nonhomogeneous patchy pattern! the 1.00-ct barrier. whereas the sharp edges of the coating would be readily visible) as they are on a diffusion-treated CONCLUSION sapphire with immersion in methylene iodide, The polycrystalline synthetic diamond or DLC Currently it is unclear whether this technology films cominercially available today as represented could provide a new way to balance a small yellow by the sample we examined! appear to be easily component in a faceted near-colorless diamond by detectable and sho~~ldnot be a source of concern at depositing small amounts of blue synthetic dia- this time to the jewelry industry. mond in certain areas of the cut stone) as has been However) even scientists intimately familiar done in the past with metallic fl~lorides. with the subject are in strong disagreement about Many more applications of synthetic diamond possible future developments. There isl therefore! a and DLC coatings in the gem trade can be imag- need to closely monitor the progress of this devel- ined. Perhaps the most obvious is to coat gem- oping technology. It is interesting to note that we stones that have a low scratch hardness to improve have begun to see advertisements for diamond- their resistance to wear. A patent already exists in coating equipment and also for companies offering Germany for this very application on! for example! to coat various substrates with synthetic diamond apophyllite and lzyanite (Streclzert et ala11988). or DLC thin films (both of which are often-and Alsol DLC thin films could potentially be used to erroneously-referred to simply as ''diamond "seal11 water in natural opals to prevent crazing. filmst1).At this timel however! we are not aware of Should diamond coating evolve into a very inex- the commercial availability of any synthetic dia- pensive technologx it might even be economical to mond- or DLC-coated gem material.

REFERENCES Angus J.C., Hayman C.C. (1988) Low-press~~re,metastable Koivula ].I., Fryer C.W (1984)Identifyinggem-quality synthetic growth of diamond and "diamond-lilze" phases. Science, diamonds: An update, Cezns ed Cemologx Vol. 20, No. 3, Vol. 241, pp. 913-921. pp. 146-158. Angus J.C., Will H.A., Stanlzo WS. (1968) Growth of diamond Lettington A.H. (19881 Optical applications and improved seed crystals by vapor deposition. lournal of Applied deposition processes for diamo~ldlikecarbon coatings. Physics, Vol, 39, No. 6, pp. 2915-2922. Paper presented at the Society of PhotoOpticaI lnstrumen- Bachnlann PK., Messier R. (19891 Emerging technology of tation Engineers Conference on Diamond Optics, San diamond thin films, Chemicol ond Engineering AJews, Vol. Diego, CA, August 16 and 17. 67, No. 20, pp. 24-39. McCrone WC., McCrone L.B., Delly J.G. (1978)Polarized Light Crowningshielcl R, (1971)General Electric's c~~ttablesynthetic Mjcroscopy Ann Arbor Science P~lblishers,Ann Arbor, diamonds. Gems d Gemologx Vol. 13, pp. 302-314. Ml, pp. 54-57, De Beers Consolidated Mines, Ltd. (1988)Annual Report 1987. Matsumoto S., Sat0 Y., Kamo M., Setalza N. (1982a) Vapor Kimberley, South Africa. deposition of diamond particles from methane. lopanese Derjaguin B.V, Fedoseev D,V, Lulzyanovicl~VM., Spitzin B.V, /ourno1 of Applied Physics, Vol. 21, No. 4, pp. Ll83-Ll85. Ryabov VA., Lavrentyev A.V (1968)Filamentary diamond Matsumoto S., Sat0 Y., Tsutsumi M., Setalza N. (1982bJGrowth crystals. lourno1 of Crystal Growth, Vol. 2, pp. 380-384. of diamond particles from methane-hydrogen gas. lournal DeVries R.C. (1987) Synthesis of diamond under metastable of MaLeri~lsScier~ce, VoI. 17, pp. 3 106-3 112. conditions. Annual Review in Materials Science, Vol. 17, Shigley J.E., Fritsch E., Stoclzton C,M., Koivula J.I., Fryer C.W, pp. 161-187. Kane R.E. (1986)The gen~ologicalproperties of the Sum- Eversole W (19621 Synthesis of diamond. U.S. Patent Office, itomo gem-quality synthetic yellow diamonds. Gems el Nos, 3030187 and 3030188. Gemologx Vol. 22, No. 4, pp. 192-208. Hara A., Fujimori N. (19891 Tcchnological application of CVD Shigley J.E., Fritsch E., Stoclzton C.M., Koiv~~laJ.I., Fryer C.W, diamond. In Spring Meeting of the Electrochemical Svci- Kane R.E., Hargett D.R., Welch C.W (1987).The gemologi- etj( Exter~dedAbstracts> Electrochemical Society, Vol. cal properties of the De Beers gem-quality synthetic 89-1, p. 108. diamonds. Gems 01 Gemologx Vol. 23, No. 4, pp, 187-206, Hiroshi A. (19861 Growth of ornamental diamonds. (Japanese Spear K.E. (19871 Growth of crystalline diamond from low- patent) Japan Kolzai Tolzlzyo Koho JP 611158898 A2, July pressure gases. Earth ond 1UineraI Sciences, Vol. 56, No. 4, 18. pp. 53-59. Kamo M., Sat0 Y., Matsumoto S., Setaka N. (1983)Diamond Streclzert G.G., Franch H,-G., Collin G. (1988)Scl~muclzstei~~e synthesis from gas phase in microwave plasma. [ournal of und Verfahren zu ihrer Herstellung. Bu~~desrepublilzDeu- Crystal Growth, Vol. 62, pp. 642444. tschland, Deutsches Patentamt Offenlegungsschrift, Nu- Koidl P (1989) Critical assessment of diamondlilze carbon. ln mnler DE 3708171Al. Spring Meeting of the Electrochemical Societx Extended Williams B.E., Glass J.T (19891 Characterization of diamond Abstracts, Electrochemical Society, Vol. 89-1, p. 134. thin films: Diamond phase identificatio~~,surface mor- Koivula J.I. (19871 Gem news: Cubic zirconia coated by syn- phology, and defect structures. 10~1rnolof Materials Re- thetic diamond? Gems 01 Gemologx Vol. 23, No. 1, p, 52. sel~rch,Vol. 4, No. 2, pp. 373-384.

90 Notes and New Techniques GEMS & GEMOLOGY Sum111er 1989