Lu G&G Winter_Layout 1 12/19/12 10:44 AM Page 273 FEATURE ARTICLES COLOR ORIGIN OF LAVENDER JADEITE: AN ALTERNATIVE APPROACH Ren Lu The market value of jadeite has risen dramatically in recent decades, often rivaling that of fine ruby and sapphire. Understanding the color origin of jadeite and reliably determining treatments have become increasingly important in the trade. This study uses single-crystalline analogs in conjunction with poly- crystalline jadeite to examine the color origin of lavender jadeite through quantitative spectroscopy and modern trace-element analytical techniques. Several previously proposed chromophores are assessed for their possible contribution to jadeite coloration. Quantitative analysis confirms that green and laven- der colorations are caused by chromium and manganese, respectively. The relative significance of these two chromophores is compared to determine their impact on observable coloration. The findings on color origin are applied to the identification of treated material on the current market. adeite is a highly regarded gemstone, particularly and Ti4+-Fe2+—have been proposed based on UV-visi- in Asian markets. Some of the finest pieces com- ble spectroscopic data, chemical analyses, and com- Jmand premium values, often surpassing those for parisons to similarly colored minerals (Rossman, top-quality ruby, sapphire, and emerald, as evidenced 1974; Shinno and Oba, 1993; Chen et al., 1999; by recent auctions (Leblanc, 2012). At Christie’s Ouyang, 2001; Harlow and Shi, 2011). Hong Kong sale on May 29, 2012, a lavender jadeite Quantitative analysis relies on the precise deter- bangle fetched US$453,003. mination of chromophore concentration and the op- The value of a gem material largely depends on whether it is of natural, treated, or synthetic origin Figure 1. Most intense lavender-color jadeite has been (figure 1). Gemological testing and detection of color treated to achieve that saturation of color but this enhancement rely on a clear understanding of color cabochon is natural color. Photo by Tino origin. The detection of chromophore(s) appropriate Hammid/GIA. for the observed color is required for a natural color determination. Trivalent chromium (Cr3+) and iron (Fe3+) have long been known as the source of “emerald” and “grassy” green colors in jadeite, respectively (Harlow and Olds, 1987; Rossman, 1977). Yet the origin of lavender color has been a subject of debate among various studies over the past 30 years. Various chromophores—in- cluding single transition metal ions Mn3+, Mn2+, Ti3+, Fe3+, and V3+, and paired charge-transfer ions Fe2+-Fe3+, See end of article for About the Author and Acknowledgments. GEMS & GEMOLOGY, Vol. 48, No. 4, pp. 273–283, http://dx.doi.org/10.5741/GEMS.48.4.273. © 2012 Gemological Institute of America COLOR ORIGIN OF LAVENDER JADEITE GEMS & GEMOLOGY WINTER 2012 273 Lu G&G Winter_Layout 1 12/19/12 10:44 AM Page 274 tical path length that light travels through a region tive results are consequently instructive to the analy- of particular absorption characteristics. Such direct sis of color origin and to determining enhancement and quantitative correlation between proposed chro- of lavender jadeite mophores and observed lavender jadeite color has In terms of technical approach, two key compo- been lacking, however. Three main challenges intrin- nents of this study are quantitative absorption spec- sic to jadeite have hindered our understanding of the troscopy and trace-element analysis at the gem’s chromophores: parts-per-million level (Box A). This is achieved through laser ablation–inductively coupled plasma– 1. The polycrystalline and sometimes near mass spectrometry (LA-ICP-MS), a technique that crypto-crystalline nature of the finest jadeite has become practical only in recent years. These poses fundamental difficulties. In polycrys- mass spectrometers provide point-by-point chemical talline materials, light does not follow a direct analysis with micrometer-size spatial resolution and path. The path length is not simply the thick- concentrations better than parts-per-million, which ness of the material, but rather an indirect and can be fully correlated to quantitative absorption complicated path through all the irregularities spectroscopy in color analysis. of crystal grains. 2. Chromophore characterization has tradition- ally relied on electron microprobe analysis, MATERIALS AND METHODS which is best suited for major elements but in- Nine natural jadeite slabs ranging from ~16 to 88 ct sufficient for detecting trace elements. Yet with well-known provenance (Nant Maw mine 109, chromophores are often trace elements at parts- Myanmar; Kotaki-Gawa Itoigawa, Japan; and near per-million (ppm) levels, rather than main ele- Saltan and La Ensenada, Guatemala) were provided ments at percent (parts-per-hundred) levels. For by Dr. George Harlow of the American Museum of instance, a trace amount of chromium at only Natural History in New York. These materials were a few hundred ppm can produce appreciable mostly whitish, with zones of pinkish lavender colors in ruby (McClure 1962; Eigenmann et al., (Burmese) and bluish lavender (Japanese and 1972; and the author’s recent analysis of hun- Guatemalan) colors. Sixteen faceted pieces of known dreds of ruby samples) or green jadeite (analysis impregnated and/or color-enhanced lavender and presented below). Similarly, a few tens of ppm purplish jadeite materials were provided by Chinese of beryllium will readily alter the color of sap- dealers. To test the alternative approach to establish- phire (Emmett et al., 2003). Thus the true chro- ing color origin, three centimeter-size gem-quality mophore(s) responsible for the observed color natural crystals of spodumene (hiddenite and kunzite may not be correctly identified due to limited varieties) from Afghanistan were obtained from GIA sensitivity of analytical techniques. collections. 3. Multiple transition metal ions or pairs are known to produce broad absorption bands in UV-Visible Spectroscopy. Jadeite and spodumene the same general region (near 550–650 nm) re- samples were prepared as wafers with parallel pol- sponsible for a lavender color. ished surfaces and various thicknesses. For single crystals of spodumene, three sets of parallel polished This study takes a completely different approach surfaces with maximum pleochroic colors were pre- to addressing color origin in lavender jadeite by quan- pared using a custom-built optical orientation device. titatively analyzing high-quality single-crystals of UV-visible spectra were collected with a Perkin- closely matched materials. Elmer PE950 spectrometer equipped with mercury Spodumene and jadeite share closely matched and tungsten light sources, and photomultiplier crystallographic structures and optical and spectro- tube/PbS detectors that were built into an integrating scopic properties. Similar to jadeite, spodumene is sphere. A custom-made sample holder specially de- available in both green (hiddenite) and pink/lavender veloped for quantitative analysis was used to ensure (kunzite) color. Unlike jadeite, which is polycrys- the precise positioning of the sampling area in a 3 talline and rarely exhibits large crystals, high-quality mm diameter window. The same sampled area was single crystals of spodumene are widely available, further analyzed by LA-ICP-MS, particularly for which facilitates quantitative spectroscopic and trace-element composition to correlate spectral fea- trace-element (chromophore) analysis. The quantita- tures with potential chromophores. Polarized spectra 274 COLOR ORIGIN OF LAVENDER JADEITE GEMS & GEMOLOGY WINTER 2012 Lu G&G Winter_Layout 1 12/19/12 10:44 AM Page 275 BOX A: QUANTITATIVE CHROMOPHORE ANALYSIS FROM SPECTROSCOPY AND TRACE-ELEMENT CHEMISTRY The combination of UV-visible absorption spectroscopy VISIBLE-RANGE SPECTRUM and chemical analysis allows us to determine the chro- 2.0 mophore(s) that cause the observed color. Figure A-1 il- Ruby wafer, O-ray, 1.42mm thick Cr: 2487 ppmw / 958 ppma lustrates how this is accomplished. Absorption is proportional to the concentration of absorbers through 1.5 which light passes (known as the Beer-Lambert law). A Original sample wafer few relatively simple mathematical steps will lead to 1.0 the following: ABSORBANCE 0.5 Equation 1 Calculated color CIE L*a*b* = 57, 55, -31 0 Equation 2 300 400 500 600 700 800 WAVELENGTH (nm) where: A is absorbance s is the absorption cross section N is the concentration of absorbers 100 ppma 200 ppma 500 ppma 1000 ppma L*a*b* = 77, 28, -19 L*a*b* = 63, 47, -28 L*a*b* = 46, 67, -32 L*a*b* = 36, 71, -21 d is the thickness the light path length and symbols with the subscript “0” are sets of Figure A-1. This visible-range spectrum depicts the deter- known values of these parameters. mination of chromophore (Cr) in a ruby sample. The Absorption cross section is a constant for a particu- color circles (below the spectrum) demonstrate col- lar chromophore. Consequently, absorption for any oration for rubies with various Cr concentrations for a 5 chromophore concentration and sample thickness can mm path length. Rubies with twice the Cr concentration be predicted from the relationships above. For instance, and half the path length appear the same as 200 ppma more saturated color (and correlating absorbance A) can and 2.5 mm path length. be achieved by either increasing chromophore concen- tration (N) or thickness of sample (d). ment of thickness, respectively. Color coordinates (CIE In this sample for ruby (personal data), a known set of L*a*b*) can be calculated from absorbance/transmittance, values A , N , and d are established from the UV-visible and the color of the sample is quantitatively reproduced 0 0 0 absorption spectrum, LA-ICP-MS analysis, and measure- using software such as Adobe Photoshop. were collected in the 200–1400 nm range with a 0.65 was used in the trace-element analysis. NIST (Na- nm spectral resolution at a scan speed of 96 nm/min. tional Institute of Science and Technology) glass Quantitative UV-visible spectroscopic measure- standards SRM 610 and 612 were used for internal ment for colors relies on correctly identifying the calibration (http://www.nist.gov/srm/).