feature articles geographic origin determination o e · metamor phic emerald group ( i.e ,d av dr...

614 GEOGRAPHIC ORIGIN OF EMERALD GEMS & GEMOLOGY WINTER 2019

When the Spanish conquistadors firstbrought Colombian emeralds (figures 1 and2) onto the international market, they be-

came a global sensation in their day. Emeralds fromCentral Asia and Egypt were known at the time, butthe world had likely never seen emeralds of suchhigh quality and size. Traders soon developed distri-bution channels that brought the Colombian mate-rial all the way from the royal courts in Europe to thepowerful Moguls of India (Giuliani et al., 2000).

The nineteenth and twentieth centuries saw allof this change as new mines sprang up around theworld to challenge the famed Colombian emeralds.Most notable in terms of high-quality productionwere Brazil, Russia, and Zambia, but smaller depositshave been uncovered in Madagascar and Ethiopia andelsewhere. As the market has evolved alongsidethese developments, geographic origin has come tobe an important factor for fine-quality emeralds. Thedemand for emerald origin determination was ini-tially driven by the proliferation of sources. However,this expansion and diversification of emerald sourceshas also complicated origin determination. As thenumber of sources grows, so does the overlap in theircharacteristics. The following sections will detail the

specific criteria used in the laboratory at GIA tomake geographic origin conclusions for emeralds, aswell as potential areas of overlap and how these aredealt with.

SAMPLES AND ANALYTICAL METHODSEmeralds included in this study are dominantly fromGIA’s reference collection, which was assembled overmore than a decade by GIA’s field gemology depart-ment. Stones in GIA’s reference collection were ob-tained from reliable sources and collected as close tothe mining source as possible; see Vertriest et al.(2019), pp. 490–511 of this issue. When necessary, thedata from the reference collection was supplemented

GEOGRAPHIC ORIGIN DETERMINATIONOF EMERALDSudarat Saeseaw, Nathan D. Renfro, Aaron C. Palke, Ziyin Sun, and Shane F. McClure

FEATURE ARTICLES

The gem trade has grown to rely on gemological laboratories to provide origin determination services for emer-alds and other fine colored stones. In the laboratory, this is mostly accomplished by careful observations of in-clusion characteristics, spectroscopic analysis, and trace element profile measurements by laserablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS). Inclusions and spectroscopy can oftenseparate Colombian emeralds from other sources (although there is some overlap between Colombian, Afghan,and Chinese [Davdar] emeralds). For non-Colombian emeralds, trace element analysis by LA-ICP-MS is neededin addition to information from the stone’s inclusions. The relative chemical diversity of emeralds from worldwidedeposits allows confidence in origin determination in most cases. This contribution outlines the methods andcriteria used at GIA for geographic origin determination for emerald.

In Brief• Emerald origin determination can be challenging dueto the number of deposits and their often similar inclu-sion scenes.

• Emeralds can be classified as hydrothermal/metamor-phic or schist-hosted based on UV-Vis-NIR spec-troscopy and inclusions.

• Hydrothermal/metamorphic emeralds, including thosefrom Colombia, have jagged fluid inclusions but caneasily be separated by trace element chemistry.

• For schist-hosted emeralds, a combination of trace ele-ment chemistry and microscopic observations of inclu-sions is required for a conclusive origin determination.

See end of article for About the Authors and Acknowledgments.GEMS & GEMOLOGY, Vol. 55, No. 4, pp. 614–646,http://dx.doi.org/10.5741/GEMS.55.4.614© 2019 Gemological Institute of America

by stones gathered by the authors of this study out-side of the field gemology program or borrowed fromGIA’s museum collection. The trace element datawas obtained from 298 samples total, with 36 fromZambia (25 from Kafubu and 11 from Musakashi), 34from Colombia, 36 from Afghanistan, 22 from Mada-

gascar, 64 from Russia, 49 from Brazil, 32 fromEthiopia, and 25 from China (Davdar). Emeralds fromMadagascar are much less frequently encountered inthe lab. Therefore, these deposits are included in thetrace element section here, but their inclusions arenot described for the sake of brevity and clarity. More

GEOGRAPHIC ORIGIN OF EMERALD GEMS & GEMOLOGY WINTER 2019 615

Figure 1. An 88.4 gemerald crystal fromCoscuez, Colombia.Photo by Robert Wel-don/GIA; courtesy ofthe Roz and GeneMeieran Collection.

information about specific mines within these coun-tries can be found in Vertriest et al. (2019), pp. 490–511 of this issue.

Trace element chemistry was collected at GIAover the course of several years using two differentLA-ICP-MS systems. The ICP-MS was either aThermo Fisher X-Series II or iCAP Qc system, cou-pled to an Elemental Scientific Lasers NWR 213 laserablation system with a frequency-quintupledNd:YAG laser (213 nm wavelength with 4 ns pulsewidth). Ablation was carried out with 55 μm spotsizes with fluence of 8–10 J/cm2 and repetition ratesof either 7, 10, or 20 Hz. 29Si was used as an internalstandard at 313500 ppm using NIST 610 and 612 asexternal standards. Repeat analyses on samples fromsingle locations over time have verified the consis-tency of the analyses with these various setups. Ac-curacy is estimated to be within 10–20% for mostelements analyzed, based on comparisons with elec-tron microprobe data on a small selection of samples.

Inclusions were identified, when possible, usingRaman spectroscopy with a Renishaw inVia Ramanmicroscope system. The Raman spectra of the inclu-sions were excited by a Stellar-REN Modu Ar-ionlaser producing highly polarized light at 514 nm andcollected at a nominal resolution of 3 cm–1 in the2000–200 cm–1 range. In many cases, the confocal ca-pabilities of the Raman system allowed inclusionsbeneath the surface to be analyzed.

UV-Vis spectra were recorded with a Hitachi U-2910 spectrometer or a PerkinElmer Lambda 950 in

the range of 190–1100 nm with a 1 nm spectral reso-lution and a scan speed of 400 nm/min. When possi-ble, polarized spectra of oriented samples werecollected to obtain o- and e-ray absorption spectra.UV-Vis-NIR spectra are presented either as relativeabsorbance with no units or absorption coefficient (a)in units of cm–1 where a = A × 2.303/t, with A = ab-sorbance and t = path length in cm.

RESULTSUV-Vis-NIR Spectroscopy in Emerald Origin Determi-nation. As with rubies and sapphires, emerald’s geo -graphic origin determination starts with a coarseseparation into one of two groups based on their geo-logical conditions of formation: hydrothermal/meta-morphic and schist-hosted/magmatic-related (the TypeII tectonic metamorphic-related and Type I tectonicmagmatic-related groups from Giuliani et al., 2019;Giuliani and Groat, 2019, pp. 464–489 of this issue).This distinction also roughly corresponds to a separa-tion between Colombian emeralds (or those emeraldsthat might have Colombian-like features) and essen-tially everything else. The hydrothermal/metamorphic group contains emeralds from Colom-bia, Afghanistan, and China (Davdar), while the schist-hosted/magmatic group includes emeralds fromZambia (Kafubu), Russia, Ethiopia, and Brazil, amongothers. In many cases, this separation can be madebased on gemological properties. Hydrothermal/meta-morphic emeralds often have jagged multiphase inclu-sions, whereas schist-hosted emeralds generally haveblocky or irregular multiphase inclusions. The differ-ence in appearance is likely related to variations in theway the fluids were originally entrapped as well aspost-entrapment healing and evolution of the inclu-sions while the emeralds were held at high tempera-ture and pressure before being exhumed to the earth’ssurface (e.g., Giuliani et al., 2019).

But advanced analytical techniques are often morereliable than relying solely on inclusions. Schist-hosted emeralds generally have much higher Fe con-tent than hydrothermal/metamorphic emeralds. Thischemical distinction is most easily seen using UV-Vis-NIR spectroscopy. Representative UV-Vis-NIR o-rayspectra of hydrothermal/metamorphic emeralds areshown in figure 3 (top). The most prominent featuresin this group are the Cr3+ broad absorption bands atabout 430 and 600 nm and the sharp Cr3+ band ataround 683 nm. A typical Colombian UV-Vis-NIRspectrum is characterized by the absence of any ab-sorption bands in the near-IR region at wavelengths


Figure 2. Colombian emerald from Chivor, 4.50 ct.Photo by Robert Weldon/GIA.

greater than 700 nm (bottom spectrum in figure 3,top). Other members of the hydrothermal/metamorphic emerald group (i.e., Davdar, China, andAfghanistan), which generally have higher concentra-tions of Fe, usually show a noticeable to significantabsorption band at 810 nm caused by Fe2+, possiblyenhanced by nearby Fe3+. This feature has traditionallybeen used to separate Colombian emeralds from thosefrom all other deposits (e.g., Saeseaw et al., 2014).However, as shown in figure 3 (top), there is someoverlap between some Colombian emeralds and other

members of the hydrothermal/metamorphic group.On average, Colombian emeralds have lower Fe con-tent than stones from any other deposit. However, thefew stones at the higher range of Fe content forColombian emeralds have a noticeable Fe2+ absorptionband at 810 nm. Similarly, while Afghan emeraldstend to have higher Fe content than most Colombianstones, at their lower Fe concentration range somestones may show only a slight Fe-related absorptionat 810 nm, which could be mistaken for a classicColombian emerald spectrum. Great care must be


Figure 3. Top: Represen-tative UV-Vis-NIRspectra of metamor-phic/hydrothermalemeralds. The low-Feand high-Fe Colombianemeralds have 200 and1900 ppm Fe, respec-tively, while the low-Feand high-Fe Afghanemeralds have 1100and 2400 ppm Fe, re-spectively. Bottom:Representative UV-Vis-NIR spectra of schist-hosted emeralds. Allspectra are o-ray spec-tra and have beenrenormalized in orderto compare the relativeintensities of absorp-tion features and,hence, absolute intensi-ties cannot be com-pared between spectra.The spectra are also off-set by various amountsin order to facilitatecomparison.

AB

SOR

BAN

CE

UV-VIS-NIR SPECTRA

WAVELENGTH (nm)300 400 500 600 700 800 900 1000

Cr3+, 430 nm

Cr3+, 600 nm Cr3+, 639, 683 nmFe2+/Fe3+, 810 nm

Afghanistan(high-Fe)

Afghanistan(low-Fe)

China

Colombia (low-Fe)

Colombia(high-Fe)

AB

SOR

BAN

CE

WAVELENGTH (nm)300 400 500 600 700 800 900 1000

Cr3+, 430 nm

Fe3+, 372 nm

Cr3+, 600 nm

Cr3+, 639, 683 nm

Fe2+/Fe3+, 810 nm

Zambia

Madagascar

Russia

EthiopiaBrazil

taken in interpreting UV-Vis-NIR spectra to identifypossible Colombian emeralds, and multiple lines ofevidence, including trace element chemistry, must beused to accurately ascertain a Colombian origin.

Due to their much higher Fe content, schist-hostedemeralds are easy to separate from Colombian andother hydrothermal/metamorphic stones based ontheir o-ray UV-Vis-NIR absorption spectra (figure 3,bottom). The most obvious distinction is the signifi-cantly increased intensity of the Fe-related 810 nm ab-sorption band in schist-hosted emeralds. Additionally,these higher-iron emeralds also tend to have a moreprominent Fe3+ absorption band at 372 nm than the hy-drothermal/metamorphic emeralds. As with the hy-drothermal/metamorphic group, making geographicorigin determinations for schist-hosted emeralds re-quires the use of multiple lines of evidence includingUV-Vis-NIR analysis, trace element chemistry, and mi-croscopic observation of a stone’s inclusions.

Microscopic Internal Inclusions. Emeralds, even top-quality material, tend to be fairly heavily included.Fluid inclusions, partially healed fissures, needles,and/or solid mineral crystals are frequently encoun-tered. Inclusions are an extremely important tool forgeographic origin determination as well as separatingnatural and synthetic emeralds. However, there areseveral areas of overlap. Emeralds from some de-posits may have inclusion scenes that resemble thoseseen in other important deposits. Some stones maylack any diagnostic inclusions, making it impossibleto provide an origin opinion if there is no definitivetrace element data. Gemologists often classify emer-ald into two broad groups based on fluid inclusions:jagged or blocky. This also corresponds roughly to thehydrothermal/metamorphic and schist-hosted clas-sification presented above. Most fluid inclusions inemerald are multiphase, but it is often difficult to ob-serve all the distinct phases in an inclusion using agemological microscope. Jagged fluid inclusions usu-ally appear to be composed of a liquid, a gas, and oneor more solid phases. Sometimes there is even morethan one liquid phase present as well. Regardless,when there are clearly observed solid, liquid, and gasphases present, these are typically referred to asthree-phase inclusions. Blocky fluid inclusions oftenappear to be composed of only two phases, a liquidand a gas. However, there is often a solid phase pres-ent in these inclusions, which is hard to observe dueto similarity in refractive index between the solidand fluid. Sometimes the use of cross-polarized light

reveals hidden solid crystals in these fluid inclusions.Additional information about the internal featuresof emeralds from major world deposits can be foundin the following references: Giuliani et al. (1993,2019), Vapnik et al. (2006), Groat et al. (2008), Mar-shall et al. (2012), Zwaan et al. (2012), Saeseaw et al.(2014), and Vertriest and Wongrawang (2018).

Geographic Origin Determination of Hydrother-mal/Metamorphic Emeralds. Colombian emeraldsare the most economically important from the hy-drothermal/metamorphic type of deposit, and jaggedfluid inclusions are their hallmark features. Suchfluid inclusions are typically elongate, with jaggedsawtooth edges. Unfortunately, emeralds from otherhydrothermal/metamorphic deposits such asAfghanistan and China may have similar jagged fluidinclusions. Among the mineral inclusions found inthese emeralds are calcite, pyrite with various mor-phologies, albite, and carbonate minerals. This sec-tion will document the internal features of emeraldsfrom Colombia, Afghanistan, and China. Emeraldsfrom a minor deposit at Musakashi in Zambia willalso be mentioned briefly at the end of this section.Further information on the inclusion scenes in emer-alds from these deposits can be found in Bowersox etal. (1991), Bosshart (1991), and Saeseaw et al. (2014).

The Internal World of Colombian Emeralds.Colom-bian emeralds (figure 4) have long been held in highesteem for their rich green color and for the history


Figure 4. Matched pair of Colombian emeralds withno oil, approximately 10 carats total. Photo by RobertWeldon/GIA; courtesy of Amba Gem Corporation.

and legends associated with these stones. The mostreadily identifiable feature of Colombian emeraldsare classic three-phase jagged fluid inclusions. Thesecontain a gas bubble and a cubic crystal (typicallyhalite and/or sylvite) and generally range in size from100 μm to 1 mm (figures 5 and 6). However, as men-tioned above, jagged fluid inclusions can be seen inemeralds from other deposits such as Afghanistanand China (see also Saeseaw et al., 2014). Large jaggedfluid inclusions (greater than about 500 μm) areunique to Colombia and can be taken as diagnosticevidence of origin. However, when only smallerjagged fluid inclusions are observed, care must be

taken to ascertain a stone’s origin and additional sup-porting evidence should be sought out. Gota deaceite is a growth feature sometimes seen in Colom-bian emeralds that can be used as a strong indicationof origin (figure 7). Similar growth features are some-times observed in emeralds from other sources suchas Zambia, but it is by far most common in emeraldsfrom Colombia. This growth feature’s name is Span-ish for “drop of oil,” in reference to the roiled and tur-bid appearance it lends. Gota de aceite grows parallelto the basal faces, often in planes that do not extendthroughout the stone. Common solid inclusions inColombian emeralds are carbonates, pyrite (figure 8),


Figure 5. Small jagged inclusions in Colombian emer-alds. Photomicrograph by Sudarat Saeseaw; field ofview 0.9 mm.

Figure 6. This large jagged multiphase inclusion ischaracteristic of Colombian emerald. Photomicro-graph by Charuwan Khowpong; field of view 1 mm.

Figure 7. Unique gota de aceite growth features inColombian emerald. Photomicrograph by PatchareeWongrawang; field of view 0.9 mm.

Figure 8. Group of pyrite crystals and rhombohedralcarbonate crystal in Colombian emerald. Photomicro-graph by Patcharee Wongrawang; field of view 1.6 mm.

quartz, feldspar, and small black particles from thesurrounding black shales (Giuliani et al., 2019; Giu-liani and Groat, 2019, pp. 464–489 of this issue).However, most of these minerals are also found inother emerald deposits. Similar to Kashmir sap-phires, Colombian emeralds also have one very raremineral inclusion that has never been seen in emer-alds from other deposits: parisite (Gübelin andKoivula, 2008). When observed, it can be considereda diagnostic indicator of Colombian origin; unfortu-nately, this inclusion is not frequently encountered.

The Internal World of Afghan Emeralds. WhileAfghan emeralds have never held the market shareof their Colombian brethren, many fine stones havebeen produced from the mines in the Panjshir Valley(figure 9). As for their internal characteristics, multi-phase fluid inclusions are the most common inclu-sion in Afghan emeralds. They often have anelongate, needle-like shape and host several daughterminerals (figures 10 and 11), which can distinguishthem from Colombian and Chinese emeralds.Daughter minerals in fluid inclusions in emeralds are


Figure 9. Emeralds fromAfghanistan: a 23.43 ctemerald-cut stone fromZarajet and an emeraldcrystal measuring ap-proximately 31.7 mmlong. Photo by RobertWeldon/GIA; courtesyof Himalayan Gems &Jewelry.

often assumed to be halite when they have a cubichabit; however, when there are several daughter min-erals, as in the inclusions in Afghan emeralds, theirexact identity is often hard to determine, even withthe use of confocal Raman spectroscopy. The typicalinclusion scene in these emeralds from the PanjshirValley consists of small jagged fluid inclusions (fig-ures 12 and 13), similar to those seen in Colombianemeralds, and scarce crystalline inclusions. Solid in-clusions, when present, include pyrite, limonite,beryl, carbonate minerals, and feldspar.

The Internal World of Chinese Emeralds.The depositat Davdar, China, was discovered late in the twenti-eth century. The geology of Davdar is not well under-stood, but it is reported that this deposit shares somesimilarities with other metamorphic deposits (Giu-liani et al., 2019). Fluid inclusions in these emeraldshave a jagged shape and are composed of a roundedgas bubble and a cubic crystal (figures 14–17). Theyare often small and can resemble those seen in emer-alds from Colombia and Afghanistan. While there islikely very little production from the Chinese mines


Figure 10. Elongate needle-like multiphase inclusionscontaining several daughter crystals and a gas bubbleare typically seen in Afghan emeralds. Photomicro-graph by Charuwan Khowpong; field of view 0.7 mm.

Figure 11. Needle-like multiphase inclusions lie paral-lel to each other in Afghan emerald. Photomicrographby Charuwan Khowpong; field of view 0.9 mm.

Figure 12. Small jagged inclusions in Afghan emerald.Photomicrograph by Nattida Ng-Pooresatien; field ofview 0.7 mm.

Figure 13. Irregular jagged inclusions in Afghan emer-ald. Photomicrograph by Nattida Ng-Pooresatien;field of view 0.9 mm.

reaching the market, their potential for overlap withColombian emeralds makes it important to be awareof their identifying characteristics to avoid potentialproblems in origin determination.

Inclusion Scenes Gone Wrong. Many Colombianemeralds are easily identified by large jagged inclu-sions (figure 6) or gota de aceite growth features (fig-ure 7). Multiphase inclusions containing numerousdaughter crystals (figure 10) are considered conclu-sive evidence of an Afghan origin. However, it can becomplicated in smaller or less-included stones wherethere may not be much information available to the

microscopist. For example, the inclusion scenes oftwo Afghan emeralds in figures 18 and 19 showedjagged edges with a rounded gas bubble and cubiccrystal(s). These inclusions could easily be misinter-preted as Colombian. Similarly, emeralds from aminor deposit in Zambia at Musakashi can also havejagged three-phase inclusions similar to those seenin Colombian emeralds, making it necessary tosearch for further evidence (figures 20 and 21). Aswith the Chinese emerald deposit at Davdar, fewstones have emerged from Musakashi. Nonetheless,it is important to be aware of this potential overlapin properties. UV-Vis-NIR analysis can be helpful to


Figure 14. Jagged fluid inclusions with a transparentcubic crystal and a gas bubble in emerald fromChina. Photomicrograph by Sudarat Saeseaw; field ofview 0.7 mm.

Figure 15. Jagged fluid inclusions in Chinese emerald.Photomicrograph by Charuwan Khowpong; field ofview 1.05 mm.

Figure 16. Elongate jagged inclusions in emerald fromChina. Photomicrograph by Nattida Ng-Pooresatien;field of view 0.7 mm.

Figure 17. Irregular multiphase inclusions in Chineseemerald. Photomicrograph by Nattida Ng-Pooresatien;field of view 0.7 mm.

separate between Colombian and Afghan, Chinese,or Musakashi emerald. Many Afghan, Chinese, andMusakashi emeralds have obvious Fe-related absorp-tion bands distinguishing them from Colombianstones. However, there will be some overlap in thehigh-Fe Colombian emeralds and the low-Fe Afghanand Chinese emeralds, both of which may showminor Fe-related absorption bands. For these border-line cases, chemical analysis is needed for a conclu-sive result.

Geographic Origin Determination for Schist-HostedEmeralds. Schist-hosted emerald deposits are the re-

sult of magmatic processes, including pegmatiticevents. Emeralds formed in these deposits throughinteraction of pegmatites or other magmatic bodieswith mafic, ultramafic, and/or metamorphic countryrocks. This type of emerald includes importantsources such as Zambia (Kafubu), Brazil, Russia, andEthiopia. These emeralds tend to have a darker greencolor than Colombian, Afghan, and Chinese emer-alds owing to their generally higher iron content.However, some lower-iron schist-hosted emeralds,especially those from Russia, may have lighter-tonedcolors. Schist-hosted emeralds often have blocky fluidinclusions that can be either two-phase, three-phase,


Figure 18. Jagged multiphase inclusions in an Afghanemerald appear similar to features seen in someColombian emeralds. Photomicrograph by Nattida Ng-Pooresatien; field of view 0.8 mm.

Figure 19. Irregularly shaped multiphase inclusionhosting a small gas bubble and several crystals in anAfghan emerald. Photomicrograph by GIA; field ofview 0.3 mm.

Figure 20. Jagged three-phase fluid inclusions in anemerald from Musakashi, Zambia. Photomicrographby Charuwan Khowpong; field of view 0.7 mm.

Figure 21. Jagged three-phase fluid inclusions in anemerald from Musakashi, Zambia. Photomicrographby Charuwan Khowpong; field of view 0.7 mm.

or multiphase. The most common solid crystal inclu-sion is mica, which occurs in a variety of forms andcolors. Mica can be formed before the emerald (i.e.,protogenetic) or at the same time as the host crystal(syngenetic). Other mineral inclusions can also befound: dendritic black inclusions and quartz, carbon-ate minerals, talc, pyrite, emerald, chlorite, and spinel.The following sections will describe typical inclusionscenes for each origin. Note that emeralds from Mada-gascar would be included in the schist-hosted group.However, these emeralds are much less frequently en-countered in the lab and so they are not included inthe discussion of inclusion scenes. Nonetheless, origindetermination of Madagascar emeralds will be consid-ered later in the trace element chemistry section. Fur-ther information on the inclusion scenes inschist-hosted emeralds can be found in Cassedanneand Sauer (1984), Hänni et al. (1987), Zwaan et al.(2005), Saeseaw et al. (2014), Vertriest and Won-grawang (2018), and Palke et al. (2019a).

The Internal World of Zambian Emeralds from Ka-fubu. Zambian emeralds (figure 22) are found in theKafubu area in the southern part of the historicallyimportant copper mining area known as the Copper-belt. Emerald mineralization occurs at the phlogo-pite-biotite contact zone between pegmatites and atalc-magnetite schist. Zambian emerald fluid inclu-sions are typically blocky (figure 23) or irregularlyshaped (figure 24) multiphase inclusions. Mica is acommon inclusion in Zambian emeralds from Ka-fubu, occurring with a brownish color and roundedshape (figure 25) or in pseudo-hexagonal greenplatelets (figure 26). Black platelets, or dendritic in-

clusions composed of oxide minerals such as mag-netite, hematite, or ilmenite (figure 27), can be seenin Zambian and other schist-hosted emeralds. Elon-gate amphibole crystals are also observed occasion-ally in Zambian emeralds (figure 28). Other mineralssuch as apatite, pyrite, talc, barite, albite, and calcitehave also been reported (Saeseaw et al., 2014).

The Internal World of Brazilian Emeralds. Emeraldshave been discovered in several Brazilian localitiesincluding Carnaíba and Socotó (Bahia), Santa Terez-inha (Goiás), and Itabira (Minas Gerais). This articlewill focus on production from Itabira, or the Belmont


Figure 22. Zambian emeralds at 8.14 ct (left) and 1.76ct (right). Photo by Robert Weldon/GIA; courtesy ofMark Kaufman, Kaufman Enterprises.

Figure 23. Blocky fluid inclusions are common inZambian emeralds, seen perpendicular to the opticaxis of the host. Photomicrograph by Nattida Ng-Pooresatien; field of view 1.1 mm.

Figure 24. Irregularly shaped multiphase inclusionsin emeralds from Kafubu, Zambia, as seen parallel tothe optic axis. Photomicrograph by Nattida Ng-Pooresatien; field of view 2 mm.

mine, as it is the main producer of Brazilian emeraldtoday. Characteristic inclusions for emeralds fromBelmont include blocky fluid inclusions and paralleltiny tubes described as “rain-like” (figure 29) that, ifdense enough, can occasionally produce chatoyancy.The fluid inclusions typically show a blocky shape(figure 30) and may have multiple liquid/gas/solidphases apparent within (figure 31). Some irregular


Figure 25. Brownish micas are common in Zambianemerald but can also be found in emerald from otherdeposits. Photomicrograph by Nattida Ng-Pooresatien;field of view 1.3 mm.

Figure 26. Pseudo-hexagonal fluid inclusions andgreenish mica are common in Zambian emerald. Pho-tomicrograph by Nattida Ng-Pooresatien; field ofview 0.9 mm.

Figure 27. Opaque thin platelets showing dendriticinclusions are seen in Zambian and other schist-hosted emeralds. Photomicrograph by Nattida Ng-Pooresatien; field of view 1.0 mm.

Figure 28. Greenish rod-like inclusions identified asamphibole in Zambian emerald. Photomicrograph byNattida Ng-Pooresatien; field of view 1.6 mm.

Figure 29. Cloud of fine elongate tubes with variousfillings form a rain-like inclusion in a Brazilian emer-ald. Photomicrograph by Charuwan Khowpong; field

of view 1.30 mm.

fluid inclusions in Brazilian emeralds may have con-centric equatorial fractures (figure 32). Mica can befound as both syngenetic pseudo-hexagonal (figure33) and protogenetic rounded brown crystals (figure34). Other solid mineral inclusions such as quartz,magnetite, chromite spinel, calcite, or pyrite can alsobe observed, similar to other deposits of schist-hostedemerald.


Figure 30. This Brazilian emerald contains blockyand rectangular fluid inclusions and a tube filledwith aggregate calcite. Photomicrograph byCharuwan Khowpong; field of view 0.80 mm.

Figure 31. Irregular three-phase inclusion (liquid/liq-uid/gas) with immiscible liquids and a gas bubble ina Brazilian emerald. Photomicrograph by CharuwanKhowpong; field of view 1.00 mm.

Figure 32. Elongate three-phase inclusion with con-centric equatorial fractures in a Brazilian emerald.Photomicrograph by Patcharee Wongrawang; field ofview 1.30 mm.

Figure 33. Brownish mica flakes (syngenetic) showinga nearly ideally formed pseudo-hexagonal shape in aBrazilian emerald. Photomicrograph by CharuwanKhowpong; field of view 0.70 mm.

Figure 34. Crystals of rounded dark brownish bi-otite-phlogopite mica flakes in a Brazilian emerald.Photo micrograph by P. Wongrawang; field of view2.70 mm.

The Internal World of Ethiopian Emeralds. In late2016, emerald was discovered in Ethiopia near the vil-lage of Shakiso (figures 35 and 36). Their fluid inclu-sions often reveal blocky, elongate, or irregularmultiphase inclusions (figures 37–39) and nail-likegrowth blockage inclusions (figure 40). The blocky in-clusions are very similar to those seen in emeraldsfrom Zambia (Kafubu), Brazil, and Russia. Some fluidinclusions in Ethiopian stones contain two separateliquid phases, one gas bubble, and often solid crystals


Figure 36. Ethiopian emerald weighing 4.12 ct. Photoby Robert Weldon/GIA; courtesy of Mayer & Watt.

Figure 37. Blocky and irregular fluid inclusions in anEthiopian emerald. Photomicrograph by JonathanMuyal; field of view 1.44 mm.

Figure 38. Fluid inclusion with two separate liquidsand a gas bubble in an Ethiopian emerald. Photomi-crograph by Jonathan Muyal; field of view 0.72 mm.

Figure 39. Irregular fluid inclusion with large gas bub-ble in an Ethiopian emerald. Photomicrograph byUngkhana Atikarnsakul; field of view 2.0 mm.

Figure 35. Left to right: Ethiopian emeralds weighing2.00, 9.72, and 4.13 ct. Photo by Robert Weldon/GIA;courtesy of Mayer & Watt.

(liquid/liquid/gas) (figures 37–39). When these stonesare examined in a microscope using an intense incan-descent light, one of the fluids can evaporate andmerge with the gas, which has been identified byRaman spectroscopy as a CO2 gas. Some inclusionsmay also have granular fringes (figure 39). Minute par-ticles and iridescent thin films (figure 41) are seen oc-casionally and can be confused with inclusion scenesmore typical of Russian emerald. Internal growth fea-tures are typically straight with angular color zoningfollowing crystal prism faces. Some stones exhibitroiled growth (figure 42) that is distinct from the gotade aceite growth seen in Colombian emeralds. Othermineral inclusions such as brown mica platelets,clinochlore, magnetite spinel, calcite, quartz, and talccan be observed in Ethiopian emerald.

The Internal World of Russian Emeralds. Fineemeralds have been produced from the Ural Moun-tains since the mid-nineteenth century, placingRussia as one of the classic sources. Russian emer-alds can harbor unique inclusion scenes. Iridescentthin films that lie parallel to the basal pinacoid areespecially indicative of origin (figure 43). Fluid in-clusions take on different forms including elongate(figure 44), irregularly edged multiphase (figures 45and 46), to blocky (figure 47), although blocky fluidinclusions are uncommon in Russian emeralds.Some fluid inclusions also have patchy, granularfringes (figure 48). Long needles or growth tube in-clusions can also be found (figure 49). Brown mica(figure 50) and rod-shaped or needle-like amphibolecrystals have been observed. Russian emerald typi-


Figure 41. Scattered thin films in an Ethiopian emer-ald. Photomicrograph by Ungkhana Atikarnsakul;field of view 2.0 mm.

Figure 40. Elongate fluid inclusion originating arounda mineral inclusion, resembling “nail head spicules”in an Ethiopian emerald. Photomicrograph byUngkhana Atikarnsakul; field of view 0.8 mm.

Figure 42. Roiled internal growth in an Ethiopianemerald. Photomicrograph by Ungkhana Atikarn-sakul; field of view 2.7 mm.

Figure 43. Plane of numerous interference thin filmsaligned parallel to the basal pinacoid in a Russianemerald. Photomicrograph by Charuwan Khowpong;field of view 1.40 mm.


Figure 47. Blocky and irregular fluid inclusions in aRussian emerald. Photomicrograph by Aaron Palke;field of view 1.26 mm.

Figure 44. Elongate fluid inclusion with a small solidinclusion at the end in a Russian emerald. Photomicro-graph by Suwasan Wongchacree; field of view 1.20 mm.

Figure 45. Irregularly shaped two- and three-phase in-clusions in a Russian emerald. Photomicrograph bySuwasan Wongchacree; field of view 1.05 mm.

Figure 48. Irregular fluid inclusions with granularfringes in a Russian emerald. Photomicrograph byAaron Palke; field of view 1.26 mm.

Figure 49. Long growth tubes in a Russian emerald.Fiber-optic illumination. Photomicrograph bySuwasan Wongchacree; field of view 1.05 mm.

Figure 46. Irregularly shaped fluid inclusions in aRussian emerald. Photomicrograph by CharuwanKhowpong; field of view 1.75 mm.

cally has few inclusions, and origin determinationcan be challenging.

Inclusion Scenes Gone Wrong. Owing to their simi-lar geological genesis, emeralds from Zambia (Ka-fubu), Brazil, Russia, and Ethiopia may be difficult todistinguish. Making an accurate origin determina-tion requires an experienced gemologist with a ro-bust reference database to compare against unknownstones. Reaching an origin conclusion is easier whenthe stone has abundant inclusions as opposed to aclean stone. However, every emerald must be ob-served carefully, as stones from geographically dis-

parate deposits can have similar inclusion scenes. Forexample, how would you determine origin on theseblocky inclusions shown in figures 51 and 52? Oneis from Zambia and the other from Ethiopia. Irregularfluid inclusions that might have once given an im-pression of Zambian origin (again, see figure 24) arenow also occasionally found in Ethiopian emerald(figure 53). Elongate or thin rod-like fluid inclusionswere first seen in Brazilian stones, but similar inclu-sions have been found in emeralds from Ethiopia andRussia (figure 54). Similarly, rain-like inclusions wereonce considered diagnostic for Brazilian emerald butcan now be observed in Ethiopian and Russian emer-


Figure 50. Lath-shaped brown inclusions identified asphlogopite mica in a Russian emerald. Photomicro-graph by Suwasan Wongchacree; field of view 1.40 mm.

Figure 51. Rectangular blocky fluid inclusions typi-cally seen in Zambian emeralds. Photomicrograph byNattida Ng-Pooresatien; field of view 1.1 mm.

Figure 52. Blocky fluid inclusions in Ethiopian emer-ald might lead to misinterpretation as Zambian ori-gin. Photomicrograph by Ungkhana Atikarnsakul;field of view 1.1 mm.

Figure 53. Irregular fluid inclusions in this Ethiopianemerald were initially interpreted as an indicator ofZambian origin. Photomicrograph by UngkhanaAtikarnsakul; field of view 1.3 mm.

alds as well (figure 55). Clusters of brownish mica arecommon inclusions in all of these schist-hostedemeralds (figure 56). In this class there are no mineralinclusions that can be used to indicate origin. Inmany cases it is impossible to give an origin opinionbased on microscopy. For emeralds from schist-hostrock, UV-Vis-NIR analysis is not useful in separationas they all display similar spectra. In this case, chem-ical analysis by LA-ICP-MS is required to make anaccurate origin determination.

Trace Element Chemistry. Trace element analysisis a much more powerful tool for origin determina-

tion of emeralds than it is for gem corundum. Thislikely has its roots in the respective crystallinestructures. Beryl has several unique crystal sites ofvarying sizes and geometry into which trace ele-ments can substitute. This allows easier incorpora-tion and greater variety of substitutional traceelements than for corundum, which has only onecrystallographic site that can accept foreign elements.Therefore, emeralds appear to be more sensitive toslight changes in their geological environment, whichcan impart unique trace element signatures forstones from different geographic localities. Giventhe overlapping inclusion scenes for emeralds fromthe deposits discussed here, trace element chem-istry analysis is crucial for making accurate origincalls. Several examples of trace element plots usedin the GIA laboratory for emerald origin determina-tion are shown in figure 57, and table 1 gives thegeneral ranges and averages of the trace elementsused in origin determination at GIA. Commonlyused trace elements include Li, K, V, Cr, Fe, Rb, andCs. Colombian emeralds tend to be the purestchemically, possessing generally lower concentra-tions of the alkalis Li, K, Rb, and Cs as well as Fe.The other hydrothermal/metamorphic emeraldsfrom Afghanistan and China are more enriched inthese trace elements, allowing them to be clearlyseparated from Colombian stones. The most charac-teristic feature of the schist-hosted emeralds is theirenrichment in Fe relative to the hydrothermal/meta-morphic group. However, most schist-hosted emer-alds also have much higher concentrations of thealkali metals, and members of this group can be gen-erally differentiated from each other by their dis-


Figure 55. Parallel fine needles in Russian emerald re-semble inclusions seen in Brazilian stones. Photomicro-graph by Suwasan Wongchacree; field of view 1.05 mm.

Figure 54. Elongate fluid inclusions in emerald fromEthiopia. These can be also seen in Brazilian andRussian emerald. Photomicrograph by UngkhanaAtikarnsakul; field of view 1.1 mm.

Figure 56. Cluster of brownish phlogopite mica crys-tals in Ethiopian emerald. These can also be seen inother schist-hosted emeralds. Photomicrograph byUngkhana Atikarnsakul; field of view 4.0 mm.

tinctive trace element profiles. For instance, Russ-ian emeralds tend to have lower Fe and higher Li;Zambian emeralds from Kafubu tend to have highCs, Fe, and Li; Madagascar emeralds tend to havehigh K and Fe; Ethiopian emeralds tend to be mod-erately enriched in most trace elements; and Brazil-ian stones tend to occupy the lower range of manyof the trace elements.

Despite these general trace element profiles, thereis still significant overlap when considering only oneor two trace elements at a time. However, the selec-tive plotting method employed by GIA for blue sap-phires and rubies is also routinely used to providegreater confidence in emerald origin determinationbased on trace element chemistry; see Palke et al.(2019b), pp. 536–579 of this issue, for a discussion ofthis method. Essentially, selective plotting makes for

an easier comparison between an unknown stoneand the vast accumulation of reference data availableby filtering out data with dissimilar trace elementprofiles and comparing only against reference stonesthat match the full trace element signature of the un-known stone. Elements used in the selective plottingmethod include Li, K, Fe, Rb, and Cs. The use of se-lective plotting often allows trace element data to beinterpreted more easily and provides greater confi-dence in origin determination based on these data(figure 58). In most cases trace element chemistryprovides solid evidence about a stone’s origin, butsometimes there is still too much overlap in the traceelement data (figure 59). When trace element dataand inclusion information are ambiguous or contra-dictory, an “inconclusive” origin determination is al-ways warranted.


Figure 57. Typical traceelement plots used fororigin determination ofemeralds.

Fe (

ppm

w)

TRACE ELEMENT DISCRIMINATION

Rb (ppmw)0.2 2 20 200

30

300

3000

Rb

(ppm

w)

Cs (ppmw)1 10 100 1000 10000

0.2

2

20

200

Li (

ppm

w)

Fe (ppmw)30 300 3000

5

50

500

Li (

ppm

w)

K (ppmw)1 10 100 1000 10000

5

50

500

Colombia

RussiaBrazil

Zambia

Ethiopia

China Afghanistan Madagascar


Schist-Hosted Emeralds

TABLE 1. Generalized trace element profiles of major world emerald deposits in ppmw.

24.2-139

62.9

61.0

bdl-359

17.7

14.3

117-2030

553

414

bdl-4.92

2.03

1.96

3.17-19.0

10.9

10.7

Range

Average

Median

Colombia

*bdl = below the detection limit of the LA-ICP-MS analysis

Li K Fe Rb Cs

78.2-268

115

99.7

51.7–1590

639

625

849–9820

3780

3120

3.87–110

46.9

44.1

11.3–97.0

46.2

43.6

Range

Average

Median

Afghanistan

Li K Fe Rb Cs

68.6–332

114

110

91.8–1320

314

303

1170–6430

2711

2470

3.47–31.5

15.1

14.6

5.96–41.2

16.3

13.8

Range

Average

Median

China (Davdar)

Li K Fe Rb Cs

31.3–359

80.8

59.4

33.2–1150

231

205

2460–9120

5120

5035

7.42–91.8

31.4

30.0

16.8–1130

148

79.7

Range

Average

Median

Li K Fe Rb Cs

Hydrothermal/Metamorphic Emeralds

Brazil

183–446

310

306

98.6–889

307

306

1970–5320

3841

3870

14.7–166

56.9

58.0

151–544

345

351

Range

Average

Median

Li K Fe Rb Cs

Ethiopia

50.3–320

144

127

107–2660

1200

1220

6720–15800

9400

8990

12.7–395

157

174

110–1670

555

541

Range

Average

Median

Madagascar

Li K Fe Rb Cs

300–1640

751

723

bdl–8830

319

108

864–8800

2613

2330

9.10–361

39.9

28.5

107–2180

714

521

Range

Average

Median

Russia

Li K Fe Rb Cs

360–1140

604

600

305–890

500

513

4620–11600

8139

8410

17.4–116

65.9

72.6

375–2430

1344

1480

Range

Average

Median

Zambia (Kafubu)

Li K Fe Rb Cs

0.016–0.71 0.27–3.15 0.71–5.57 0.003–0.040 0.005–0.99Range

Detection Limits (ppmw)

Li K Fe Rb Cs


Figure 58. Example of aRussian emerald’s ori-gin determined usingthe selective plottingmethod (the unknownstone shown as a redcircle). The trace ele-ment profile of an un-known emerald can beeasier to interpretthrough the selectiveplotting method usedby GIA. Selective plot-ting uses coarse,medium, and fine win-dows to filter out dis-similar reference data,making the plots easierto interpret.

K (

ppm

w)


Cs (ppmw)1 10 100 1000 10000

1

100

10

1000

10000

K (

ppm

w)

Cs (ppmw)1 10 100 1000 10000

1

100

10

1000

10000

Li (

ppm

w)

Rb (ppmw)0.2 2 20 200

5

50

500

Li (

ppm

w)

Rb (ppmw)0.2 2 20 200

5

50

500

K (

ppm

w)

Cs (ppmw)1 10 100 1000 10000

1

100

10

1000

10000

K (

ppm

w)

Cs (ppmw)1 10 100 1000 10000

1

100

10

1000

10000

Li (

ppm

w)

Rb (ppmw)0.2 2 20 200

5

50

500

Li (

ppm

)

Rb (ppmw)0.2 2 20 200

5

50

500

All data All data

Coarse Coarse

Medium Medium

Fine Fine

Colombia

Russia

Brazil

Zambia

Ethiopia

China

Afghanistan

Madagascar

Unknown


Figure 59. An exampleof an emerald with an“inconclusive” origin(analyses shown as redcircles). Even with theselective plottingmethod, there may stillbe too much overlap inthe data and an “incon-clusive” origin wouldbe warranted.

Fe (

ppm

w)

Cs (ppmw)1 10 100 1000 10000

30

300

3000

Fe (

ppm

w)

Cs (ppmw)1 10 100 1000 10000

30

300

3000

Fe (

ppm

w)


Rb (ppmw)0.2 2 20 200

30

300

3000

Fe (

ppm

w)

Rb (ppmw)0.2 2 20 200

30

300

3000

Fe (

ppm

w)

Cs (ppmw)1 10 100 1000 10000

30

300

3000

Fe (

ppm

w)

Cs (ppmw)1 10 100 1000 10000

30

300

3000

Fe (

ppm

w)

Rb (ppmw)0.2 2 20 200

30

300

3000

Fe (

ppm

w)

Rb (ppmw)0.2 2 20 200

30

300

3000

All data All data

Coarse Coarse

Medium Medium

Fine Fine

Colombia

Russia

Brazil

Zambia

Ethiopia

China

Afghanistan

Madagascar

Unknown


Figure 60. This ring contains a 12.09 ct sugar loaf cabochon emerald. Photo by Robert Wel-don/GIA; courtesy of Pioneer Gems.

CONCLUSIONSWith the proliferation of sources across the worldin the twenty-first century, emerald origin determi-nation has become an increasingly important serv-ice offered by gemological laboratories. At GIA,origin determinations are supported by years ofmeticulous efforts by the field gemology depart-ment in assembling a robust reference collectionand collecting data on these samples (Vertriest et al.,2019, pp. 490–511 of this issue). This continual ac-cumulation of information has forced the laboratoryto adapt its methods and criteria for origin determi-nation to keep up with the evolution of the worldemerald market (figures 60 and 61). Given the find-ing of jagged three-phase fluid inclusions in low-Feemeralds from Afghanistan and China, any stonethat gives an initial impression of a Colombian ori-gin must be carefully scrutinized to avoid inaccu-rate origin conclusion. The rise of Ethiopianemerald in the last few years has also forced the labto come to terms with these new stones and their

inclusion scenes, which can overlap with Brazilian,Russian, and Zambian emeralds. The lab has alsohad to understand the increased likelihood of seeingRussian emeralds from the Mariinsky (Malysheva)mine near Ekaterinburg in the Russian Urals. Forany non-Colombian emeralds, trace element chem-istry is always needed to support an origin determi-nation, given the potential for overlap in theirinternal features. And yet, despite the years of ex-perience collecting and analyzing data and docu-menting changes to the emerald mining scene, thereare still some stones for which the lab is obligatedto issue “inconclusive” origin opinions. Most of thetime this is due to a trace element profile that doesnot match any of the reference data or because ofambiguous or contradictory inclusion scenes. Ef-forts by the laboratory and the field gemology de-partment will be crucial to closing gaps in ourknowledge and keeping on top of further develop-ments in the world of emeralds—whatever the fu-ture may hold.


Figure 61. The emer-alds in this bracelet,48.91 carats total, arefrom the Chivor minein Colombia. Photo byRobert Weldon/GIA;courtesy of PioneerGems.


Figure CS 1-1. This0.57 ct step-cutemerald is at thecenter of this origindetermination casestudy. Photo byDiego Sanchez.

Figure CS 1-2. UV-Vis-NIR absorption spectroscopyindicates a schist-hosted deposit such as Zambia,Ethiopia, or Brazil.

AB

S. C

OEF

F. (

cm–1

)

UV-VIS-NIR SPECTRUM

WAVELENGTH (nm)

300 400 500 700 800 900600

0

2

4

6

8

10

12

14

16

18

20

Figure CS 1-3. Blocky fluid inclusions also indicate aschist-hosted deposit such as Zambia, Ethiopia, orBrazil. Photomicrograph by Aaron Palke; field of view1.72 mm.

In the first emerald case study, we will analyzethe 0.57 ct step-cut emerald shown in figure CS1-1. The first step in determining this stone’s ge-ographic origin is to collect a UV-Vis-NIR ab-sorption spectrum. The absorption spectrum(figure CS 1-2) shows a pronounced 850 nmabsorption band in the near-infrared region.This identifies the stone as a schist-hosted emer-ald, essentially narrowing the options for its ge-ographic origin to Zambia, Brazil, Russia,Ethiopia, and Madagascar. Microscopic obser-vation shows a field of blocky fluid inclusions(figure CS 1-3), typical for schist-hosted emer-alds. At this point, trace element chemistry isrequired to ensure an accurate origin determi-nation. Representative trace element plots usingdata from LA-ICP-MS measurements are shownin figure CS 1-4. Use of the selective plottingmethod (see Palke et al., 2019b, pp. 536–579of this issue) clearly shows that this stone is con-sistent with a Zambian origin. Taken together,the analytical data collected allows a conclu-sive origin determination of Zambia.

CASE STUDY 1: ZAMBIAN EMERALD


Figure CS 1-4. Trace element analysis supports a Zambian origin. The selective plottingmethod was used with coarse, medium, and fine filters.

Colombia

RussiaBrazil

Zambia

Ethiopia


Unknown

Rb

(ppm

w)


Cs (ppmw)

1 10 100 1000 10000

2

20

200

0.02

Rb

(ppm

w)

Cs (ppmw)

1 10 100 100001000

2

20

200

0.2

Rb

(ppm

w)

Cs (ppmw)

1 10 100 100001000

2

20

200

0.2

Rb

(ppm

w)

Cs (ppmw)

1 10 100 100001000

2

20

200

0.2

All data

Medium

Coarse

Fine


Figure CS 2-1. Thiscase study involvesthe origin determi-nation process for a2.04 ct step-cutemerald. Photo byDiego Sanchez.

Figure CS 2-2. UV-Vis-NIR absorption spectroscopysuggests an origin in a metamorphic-related depositsuch as Colombia, Afghanistan, or China.

AB

S. C

OEF

F. (

cm–1

)

UV-VIS-NIR SPECTRUM

WAVELENGTH (nm)

300 400 500 700 800 900600

0

5

10

15

20

25

Figure CS 2-3. Jagged fluid inclusions strongly indi-cate a Colombian origin. Photomicrograph by NathanRenfro; field of view 1.31 mm.

The 2.04 ct step-cut emerald shown in figure CS2-1 is the focus of this case study. The origin de-termination procedure starts with the stone’sUV-Vis-NIR spectrum (figure CS 2-2). There isonly a very slight rise in the absorption spectrumbeyond 700 nm and into the near-infrared re-gion. This indicates a very likely Colombian ori-gin, though an Afghan or Chinese origin cannotbe completely ruled out, as some emeralds fromthose locales show similar UV-Vis-NIR spectra.The most noteworthy inclusions are jagged fluidinclusions, which also give a distinct impressionof Colombian pedigree (figure CS 2-3). Onceagain, this information must be considered care-fully as some emeralds from China, Afghanistan,and Zambia (Musakashi) can show somewhatsimilar inclusions. Therefore, trace elementanalysis is needed. Representative trace elementplots (figure CS 2-4) clearly indicate a Colom-bian origin. Note that the selective plottingmethod was not employed here and is generallynot needed for Colombian emeralds.

CASE STUDY 2: COLOMBIAN EMERALD


Figure CS 2-4. Trace element analysis corroborates a Colombian origin for this emerald,shown as the red circle.

Colombia

Russia

Brazil

Zambia

Ethiopia

China

Afghanistan

Madagascar

Unknown

Fe (

ppm

w)


Rb (ppmw)

0.2 2 20 200

30

300

3000


Figure CS 3-1. This1.05 ct step-cutemerald is thefocus of this origindetermination casestudy. Photo byDiego Sanchez.

Figure CS 3-2. UV-Vis-NIR absorption spectroscopyindicates a schist-hosted deposit such as Zambia,Russia, Ethiopia, Brazil, or Madagascar.

AB

S. C

OEF

F. (

cm–1

)

UV-VIS-NIR SPECTRUM

WAVELENGTH (nm)

300 400 500 700 800 900600

0

5

10

15

20

25

Figure CS 3-3. Irregular fluid inclusions also suggest aschist-hosted emerald deposit such as Zambia, Rus-sia, Ethiopia, Brazil, or Madagascar. Photomicrographby Aaron Palke; field of view 2.37 mm.

Figure CS 3-1 shows a 1.05 ct step-cut emeraldunder consideration for a geographic origin de-termination. GIA’s standard procedure for emer-ald origin determination is to start by collectinga UV-Vis-NIR absorption spectrum (figure CS 3-2). The rise in absorption beyond 700 nm andinto the near-infrared clearly indicates that thisstone belongs to the schist-hosted emeraldgroup. Therefore, the origins we will considerare Zambia, Brazil, Russia, Ethiopia, and Mada-gascar. Microscopic observations of the inclu-sion characteristics reveal irregularly shapedfluid inclusions, which are typical of manyschist-hosted emeralds but do not clearly indi-cate any specific origin (figure CS 3-3). Finaliz-ing the geographic origin determination on thisstone requires trace element chemistry measure-ments. Figure CS 3-4 shows representative traceelement plots in comparison with referencestones of known provenance from GIA’s coloredstone reference collection. Unfortunately, thisstone does not seem to uniquely match any spe-cific geographic location. Without the use of se-lective plotting, the Fe-Li plot seems to suggesta Madagascar origin. However, other plots showthat this stone lies outside the field of the Mada-gascar reference data. Therefore, with the refer-ence collection data we have so far and with theapplication of selective plotting, Madagascardoes not appear to be a viable option. Withmore reference data, hopefully stones such asthis one will be able to match to data. Untilthen, given the ambiguity in the data, the onlypossible option today for this emerald is an “in-conclusive” origin call.

CASE STUDY 3: INCONCLUSIVE EMERALD


Figure CS 3-4. Trace element analysis of this unknown emerald (shown as the red circle)failed to conclusively match it with reference data of stones with known provenance. Theultimate origin call would be “inconclusive.” The selective plotting method was used withcoarse, medium, and fine filters.

Colombia

RussiaBrazil

Zambia

Ethiopia


Unknown

Li (

ppm

w)


Fe (ppmw)

30 300 3000

50

500

5

Li (

ppm

w)

Fe (ppmw)

30 300 3000

50

500

5

Li (

ppm

w)

Fe (ppmw)

30 300 3000

50

500

5

Li (

ppm

w)

Fe (ppmw)

30 300 3000

50

500

5

All data

Medium

Coarse

Fine


Figure CS 4-1. Theorigin determina-tion process for a1.04 ct emerald isthe subject of thiscase study. Photoby Diego Sanchez.

Figure CS 4-2. UV-Vis-NIR absorption spectroscopyprovides evidence of a schist-hosted emerald that didnot originate in Colombia.

AB

S. C

OEF

F. (

cm–1

)

UV-VIS-NIR SPECTRUM

WAVELENGTH (nm)

300 400 500 700 800 900600

0

5

10

15

20

25

Figure CS 4-3. Blocky fluid inclusions point to aschist-related emerald deposit, indicating an origin ofZambia, Ethiopia, or Brazil. Photomicrograph byNathan Renfro; field of view 1.89 mm.

In this case study we examine the 1.04 ct step-cut emerald shown in figure CS 4-1. The UV-Vis-NIR absorption spectrum shows asignificant rise beyond 700 nm, leading to alarge peak in the near-infrared region (figure CS4-2). This allows clear assignment of this stoneto the schist-related emerald group. Hence, themain geographic origins to consider are Zam-bia, Brazil, Russia, Ethiopia, and Madagascar.Microscopic examination shows abundant rec-tangular, blocky fluid inclusions (figure CS 4-3),an inclusion scene consistent with most schist-hosted emeralds. However, the observation ofsuch inclusions alone cannot lead to an accu-rate origin call. To finalize the origin determi-nation, we need accurate trace elementanalysis. Figure CS 4-4 shows representativetrace element plots from LA-ICP-MS measure-ments. Use of the selective plotting methoddemonstrates a clear correlation with Zambianemeralds from GIA’s colored stone referencecollection. Along with the spectroscopic andinclusion data, this allows a confident conclu-sion of Zambian origin.

CASE STUDY 4: ZAMBIAN EMERALD


Figure CS 4-4. Trace element analysis provides conclusive evidence of Zambian origin. Theselective plotting method was used with the coarse, medium, and fine filters.

Colombia

RussiaBrazil

Zambia

Ethiopia


Unknown

Fe (

ppm

w)


Cs (ppmw)

1 10 100 1000 10000

300

3000

30

Fe (

ppm

w)

Cs (ppmw)

1 10 100 100001000

300

3000

30

Fe (

ppm

w)

Cs (ppmw)

1 10 100 100001000

300

3000

30

Fe (

ppm

w)

Cs (ppmw)

1 10 100 100001000

300

3000

30

All data

Medium

Coarse

Fine


ABOUT THE AUTHORSMs. Saeseaw is senior manager of identification at GIA inBangkok. Dr. Palke is a senior research scientist, Mr. Renfro ismanager of colored stone identification, Mr. Sun is a research as-sociate, and Mr. McClure is global director of colored stone serv-ices, at GIA’s laboratory in Carlsbad, California.

ACKNOWLEDGMENTS The authors would like to thank GIA colleagues in Bangkok andCarlsbad for their valuable assistance. We also thank James Dayand Raquel Alonso-Perez and an anonymous reviewer for theirthorough and helpful peer reviews which greatly strengthened thismanuscript.

Bosshart G. (1991) Emeralds from Colombia (Part 2). Journal ofGemmology, Vol. 22, No. 7, pp. 409–438.

Bowersox G., Snee L.W., Foord E.E., Seaf II R.R. (1991) Emeraldsof the Panjshir Valley, Afghanistan. G&G, Vol. 27, No. 1, pp.26–39, http://dx.doi.org/10.5741/GEMS.27.1.26

Cassedanne J.P., Sauer D.A. (1984) The Santa Terezinha de Goiásemerald deposit. G&G, Vol. 20, No. 1, pp. 4–13,http://dx.doi.org/10.5741/GEMS.20.1.4

Giuliani G., Groat L.A. (2019) Geology of corundum and emeraldgem deposits: A review. G&G, Vol. 55, No. 4, pp. 464–489,http://dx.doi.org/10.5741/GEMS.55.4.464

Giuliani G., Cheilletz A., Dubessy J., Rodriguez C.T. (1993) Chem-ical composition of fluid inclusions in Colombian emerald de-posits. Proceedings of the Eighth Quadrennial IAGODSymposium, pp. 159–168.

Giuliani G., Chaussidon M., Schubnel H.-J., Piat D.H., Rollion-Bard C., France-Lanord C., Giard D., de Narvaez D., RondeauB. (2000) Oxygen isotopes and emerald trade routes since an-tiquity. Science, Vol. 287, No. 5453, pp. 631–633,http://dx.doi.org/10.1126/science.287.5453.631

Giuliani G., Groat L.A., Marshall D., Fallick A.E., Branquet Y.(2019) Emerald deposits: A review and enhanced classification.Minerals, Vol. 9, No. 2, p. 105, http://dx.doi.org/10.3390/min9020105

Groat L.A., Giuliani G., Marshall D.D., Turner D.B. (2008) Emer-ald deposits and occurrences: A review. Ore Geology Reviews,Vol. 34, No. 1-2, pp. 87–112, http://dx.doi.org/10.1016/j.orege-orev.2007.09.003

Gübelin E.J., Koivula J.I. (2008) Photoatlas of Inclusions in Gem-stones, Volume 3. Opinio Publishers, Basel, Switzerland, 672pp.

Hänni H.A., Schwarz D., Fischer M. (1987) The emeralds of theBelmont mine, Minas Gerais, Brazil. Journal of Gemmology,Vol. 20, No. 7, pp. 446–456.

Marshall D., Pardieu V., Loughrey L., Jones P., Xue G. (2012) Con-

ditions for emerald formation at Davdar, China: Fluid inclu-sion, trace element and stable isotope studies. MineralogicalMagazine, Vol. 76, No. 1, pp. 213–226, http://dx.doi.org/10.1180/minmag.2012.076.1.213

Palke A.C, Lawley F.J., Vertriest W., Wongrawang P., Katsurada Y.(2019a) The Russian emerald saga: The Mariinsky Priisk mine.InColor, No. 44, pp. 36–46.

Palke A.C., Saeseaw S., Renfro N.D., Sun Z., McClure S.F. (2019b)Geographic origin determination of blue sapphires. G&G, Vol.55, No. 4, pp. 536–579, http://dx.doi.org/10.5741/GEMS.55.4.536

Saeseaw S., Pardieu V., Sangsawong S. (2014) Three-phase inclu-sions in emerald and their impact on origin determination.G&G, Vol. 50, No. 2, pp. 114–132, http://dx.doi.org/10.5741/GEMS.50.2.114

Vapnik Y., Moroz I., Roth M., Eliezri I. (2006) Formation of emer-alds at pegmatite-ultramafic contacts based on fluid inclusionsin Kianjavato emerald, Mananjary deposits, Madagascar. Min-eralogical Magazine, Vol. 70, No. 2, pp. 141–158,http://dx.doi.org/10.1180/0026461067020320

Vertriest W., Wongrawang P. (2018) A gemological description ofEthiopian emeralds. InColor, No. 40, pp. 72–73.

Vertriest W., Palke A., Renfro N.D. (2019) Field gemology: Buildinga research collection and understanding the development ofgem deposits. G&G, Vol. 55, No. 4, pp. 490–511,http://dx.doi.org/10.5741/GEMS.55.4.490

Zwaan J.C., Seifert A.V., Vrana S., Laurs B.M., Anckar B., SimmonsW.B., Falster A.U., Lustenhouwer W.J., Muhlmeister S., KoivulaJ.I., Garcia-Guillerminet H. (2005) Emeralds from the Kafubuarea, Zambia. G&G, Vol. 41, No. 2, pp. 116–148,http://dx.doi.org/10.5741/GEMS.41.2.116

Zwaan J.C., Jacob D.E., Hager T., Cavalcanti M.T.O., Kanis J.(2012) Emeralds from the Fazenda Bonfim region, Rio Grandedo Norte, Brazil. G&G, Vol. 48, No. 1, pp. 2–17,http://dx.doi.org/10.5741/GEMS.48.1.2

REFERENCES

feature articles geographic origin determination o e · metamor phic emerald group ( i.e ,d av dr...

Documents