Spring 2017 Gems Gemology v2 PDF

SPRING 2017
VOLUME LIII
THE QUARTERLY JOURNAL OF THE GEMOLOGICAL INSTITUTE OF AMERICA
Brazilian Diamonds: Past and Present

Cuprian Liddicoatite
Silicon’s Role in the Color of Corundum
2017 Tucson Report
Editorial Staff
Editor-in-Chief Editors, Lab Notes Contributing Editors
Duncan Pay Thomas M. Moses James E. Shigley
gia.edu/gems-gemology
[email protected] Shane F. McClure Andy Lucas
Donna Beaton Subscriptions
Managing Editor Editors, Micro-World Copies of the current issue may be purchased for
Stuart D. Overlin Nathan Renfro Editor-in-Chief Emeritus $29.95 plus shipping. Subscriptions are $79.99 for one
[email protected] Elise A. Skalwold Alice S. Keller year (4 issues) in the U.S. and $99.99 elsewhere. Cana-
dian subscribers should add GST. Discounts are avail-
John I. Koivula
Editor Customer Service able for group subscriptions, GIA alumni, and current
Jennifer-Lynn Archuleta Editors, Gem News Martha Erickson GIA students. To purchase print subscriptions, visit
[email protected] (760) 603-4502 store.gia.edu or contact Customer Service. For insti-
Emmanuel Fritsch tutional rates, contact Customer Service.
Gagan Choudhary [email protected]
Technical Editors
Tao Z. Hsu Christopher M. Breeding Database Coverage
G&G is abstracted in Thomson Reuters products
[email protected] Editorial Assistants (Current Contents: Physical, Chemical & Earth
Jennifer Stone-Sundberg Brooke Goedert Sciences and Science Citation Index—Expanded, in-
Erin Hogarth cluding the Web of Knowledge) and other databases.
For a complete list of sources abstracting G&G, go to
gia.edu/gems-gemology, and click on “Publication In-
formation.”
Production Staff
Creative Director Photographers Production Specialist Manuscript Submissions
Faizah Bhatti Robert Weldon Juan Zanahuria Gems & Gemology, a peer-reviewed journal, welcomes
the submission of articles on all aspects of the field.
Kevin Schumacher
Image Specialist Multimedia Specialist Please see the Author Guidelines at gia.edu/gems-
Eric Welch Video Production Lynn Nguyen gemology or contact the Managing Editor. Letters
on articles published in G&G are also welcome.
Pedro Padua Please note that Field Reports, Lab Notes, Gem News
Illustrator
Nancy Powers International, Micro-World, and Charts are not peer-
Peter Johnston Betsy Winans reviewed sections but do undergo technical and edito-
rial review.
Copyright and Reprint Permission

Editorial Review Board Abstracting is permitted with credit to the source.
Libraries are permitted to photocopy beyond the lim-
Ahmadjan Abduriyim A.J.A. (Bram) Janse Nathan Renfro
its of U.S. copyright law for private use of patrons.
Tokyo, Japan Perth, Australia Carlsbad, California
Instructors are permitted to reproduce isolated arti-
Timothy Adams E. Alan Jobbins Benjamin Rondeau cles and photographs/images owned by G&G for
San Diego, California Caterham, UK Nantes, France noncommercial classroom use without fee. Use of
photographs/images under copyright by external par-
Edward W. Boehm Mary L. Johnson George R. Rossman ties is prohibited without the express permission of
Chattanooga, Tennessee San Diego, California Pasadena, California the photographer or owner of the image, as listed in
James E. Butler Anthony R. Kampf Andy Shen the credits. For other copying, reprint, or republica-
Washington, DC Los Angeles, California Wuhan, China tion permission, please contact the Managing Editor.
Gems & Gemology is published quarterly by the
Alan T. Collins Robert E. Kane Guanghai Shi Gemological Institute of America, a nonprofit educa-
London, UK Helena, Montana Beijing, China
tional organization for the gem and jewelry industry.
John L. Emmett Stefanos Karampelas James E. Shigley Postmaster: Return undeliverable copies of Gems &
Brush Prairie, Washington Basel, Switzerland Carlsbad, California Gemology to GIA, The Robert Mouawad Campus,
Emmanuel Fritsch Lore Kiefert Elisabeth Strack 5345 Armada Drive, Carlsbad, CA 92008.
Nantes, France Lucerne, Switzerland Hamburg, Germany Our Canadian goods and service registration number
is 126142892RT.
Eloïse Gaillou Ren Lu Fanus Viljoen
Paris, France Wuhan, China Johannesburg, South Africa Any opinions expressed in signed articles are under-
stood to be opinions of the authors and not of the
Gaston Giuliani Thomas M. Moses Wuyi Wang publisher.
Nancy, France New York, New York New York, New York
Jaroslav Hyršl Aaron Palke Christopher M. Welbourn
Prague, Czech Republic Brisbane, Australia Reading, UK
About the Cover

The lead article of this issue covers Brazilian diamond production,which reached its peak in the late 1700s. Our
cover shows a brooch dating back to the mid-19th century. This piece contains a 0.77 ct natural Fancy Light pink
antique pear-shaped diamond set in darkly patinated silver ribbons over gold. The total weight of the diamonds
exceeds 5.20 carats. Photo by Robert Weldon/GIA, courtesy of Lang Antique & Estate Jewelry.
Printing is by L+L Printers, Carlsbad, CA.
GIA World Headquarters The Robert Mouawad Campus 5345 Armada Drive Carlsbad, CA 92008 USA
© 2017 Gemological Institute of America All rights reserved. ISSN 0016-626X
Spring 2017
VOLUME 53, No. 1
EDITORIAL
1 Brazil’s Enigmatic Diamonds, Cuprian Liddicoatite, and Silicon in Corundum
Duncan Pay
FEATURE ARTICLES
2 Brazilian Diamonds: A Historical and Recent Perspective
Darcy Svisero, James E. Shigley, and Robert Weldon
pg. 3 Recounts the history, geology, and notable localities of diamond mining in Brazil, and
outlines plans for future diamond production.
34 Cuprian Liddicoatite Tourmaline
Yusuke Katsurada and Ziyin Sun
Identifies 13 copper-bearing tourmalines, which have traditionally been assumed to be
elbaite species, as liddicoatite.
42 The Role of Silicon in the Color of Gem Corundum
John L. Emmett, Jennifer Stone-Sundberg, Yunbin Guan, and Ziyin Sun
Considers the previously unreported but critical impact of silicon on the coloration of gem-
quality sapphire.
48 Japanese Jadeite: History, Characteristics, and Comparison with Other Sources
Ahmadjan Abduriyim, Kazuko Saruwatari, and Yusuke Katsurada
Examines the sources, color, texture, and chemical composition of jadeite from Japanese
deposits.
68 Causes of Iridescence in Natural Quartz
pg. 49
Xiayang Lin and Peter J. Heaney
Describes the occurrence of iridescence, and the role of differential etching, in quartz
crystals from India’s Jalgaon District.
NOTES AND NEW TECHNIQUES

82 Characteristics of Gem-Quality By-Product Synthetic Zincite
Ji Zhang, Yujie Gao, and Guanghai Shi
Investigates the properties of synthetic zincite produced by industrial kilns in Poland.
pg. 83
REGULAR FEATURES
90 Lab Notes
Three faceted collector stones • Orangy red cordierite • Cape diamond with yellow
pg. 106 phosphorescence • Sapphire filled with bismuth-based glass • H4 defect in CVD synthetic
diamond • HPHT synthetic melee • Intense green HPHT synthetic diamond • Titanium-
diffused synthetic sapphire • Coated Paraíba tourmaline • Rubellite with strong “pink sleeves”
102 G&G Micro-World
Optical dishes in ametrine • Astrophyllite in bastnäsite • Flashes and flames in Ethiopian
opal • Cat’s eye imitation pen pearls • Negative crystals in sapphire • Quarterly crystal:
Mystery inclusion in topaz
109 The Dr. Edward J. Gübelin Most Valuable Article Award
110 2017 Gems & Gemology Challenge
112 Gem News International
Tucson 2017 • New source of Ethiopian emerald • Burmese gem report • Conversation with
pg. 132 Gemfields • Ruby and emerald mining updates • Secondary market for colored stones
• Cultured pearl market • Moroccan amethyst • Emerald dealing in Afghanistan • Montana
sapphire production • Magnesio-axinite from Tanzania • Sustainable vertical integration in
sapphires • Freeform tanzanite and sunstone • Gem paintings • Designs from Paula
Crevoshay • Nigerian tourmaline and sapphire • Turquoise Starship Enterprise • Rutilated
quartz mining community • Jewelry Industry Summit • Color-change pyrope-spessartine
with high grossular component • Plastic imitation of amber • Mixed-type treated red HPHT
synthetic diamond • Filled calcite with strong phosphorescence • Erratum • Letter
Brazil’s Enigmatic Diamonds, Cuprian
Liddicoatite, and Silicon in Corundum
Welcome to the first Gems & Gemology of 2017! This is an unusually diverse issue,
combining Brazilian diamonds, cuprian liddicoatite tourmaline, Japanese jadeite, iris
quartz, synthetic zincite, and a short but important contribution on the fundamental
role of the element silicon in generating blue color in many fine sapphires.
In our lead article, authors Darcy Svisero, James Shigley, and Robert Weldon survey
Brazil’s diamond deposits. This region has produced a significant number of large
rough diamonds. It’s also a source of valuable colored diamonds from predominantly
alluvial deposits, but today’s dia- “The original geologic sources of Brazil’s
mond production is small and
the original geologic sources re- most valuable diamonds remain elusive…”
main mostly elusive. Although economically viable kimberlite pipes have been discov-
ered in recent years, the outlook for increased production appears challenging.
Most researchers assume the bulk of copper-bearing tourmaline on the market today is elbaite, but some is
undoubtedly another tourmaline species: liddicoatite. Yusuke Katsurada and Ziyin Sun report on the chem-
istry of copper-bearing liddicoatite tourmaline submitted to GIA’s Tokyo laboratory. Their study shows ele-
vated gallium and lead content in cuprian liddicoatite, which displays stronger fluorescence under
long-wave ultraviolet light than cuprian elbaite due to high concentration of rare earth elements.
In our third paper, John Emmett, Jennifer Stone-Sundberg, Yunbin Guan, and Ziyin Sun discuss the critical
role of silicon in the color of gem-quality corundum. Although the effect of trace elements such as iron, tita-
nium, and magnesium within the corundum lattice is well understood, the chemical interaction of these
trace elements is not widely appreciated. The authors demonstrate that silicon’s presence is vital to the
process that allows titanium to pair with iron and create blue color in sapphire. To the best of the authors’
knowledge, this topic has not been addressed in the gemological literature.
Our next paper, by Ahmadjan Abduriyim, Kazuko Saruwatari, and Yusuke Katsurada, covers the history
and characteristics of Japanese jadeite and compares this highly valued gem with jadeite from other sources.
Their study examines the chemical composition and chromophores of jadeite varieties from the Itoigawa
and Omi regions of Niigata Prefecture and the Wakasa region of Tottori Prefecture and compares them
with samples from Myanmar, Guatemala, and Russia.
Next, Xiayang Lin and Peter Heaney analyze iris quartz crystals from geodes in the Deccan Trap basalts of west-
central India using scanning and transmission electron microscopy and atomic force microscopy. Their investiga-
tion reveals that certain crystal faces exhibit etching, which acts as a diffraction grating to produce iridescence.
Our final paper, by Ji Zhang, Yujie Gao, and Guanghai Shi, provides a gemological characterization of an in-
triguing industrial by-product: gem-quality synthetic zincite from foundries in Poland.
Be sure to investigate our three regular sections: Lab Notes documents GIA staff gemologists’ most recent
findings, Micro-World reveals fascinating gemstone inclusions, and this issue’s Gem News International re-
ports on the 2017 Tucson gem shows, including an important update on Ethiopian emerald.
We offer congratulations to the winners of our 2016 Dr. Edward J. Gübelin Most Valuable Article Award
(p. 109), and thank all of the readers who voted. Don’t forget to take this year’s G&G Challenge (pp. 110–
111), our annual multiple-choice quiz. Thank you for all your support through 2016!
Duncan Pay | Editor-in-Chief | [email protected]
EDITORIAL GEMS & GEMOLOGY SPRING 2017 1

FEATURE ARTICLES
BRAZILIAN DIAMONDS:
A HISTORICAL AND RECENT PERSPECTIVE
Darcy P. Svisero, James E. Shigley, and Robert Weldon
Brazil, which commanded global production in the 1700s and early 1800s, has remained a continuous
source of diamonds for three centuries. Even though the country represented less than 1% of world pro-
duction in 2015, a number of large famous diamonds, as well as fancy-color diamonds, have originated
there. The sources are primarily alluvial, with diamonds transported by and deposited along a multitude
of rivers. The diamonds are found mainly by independent miners (garimpeiros) in riverbeds, in uncon-
solidated sediments, and in compacted sedimentary conglomerates. After a century of exploration for
the primary sources, some economically viable kimberlite pipes have been discovered in recent years,
with one occurrence now being developed for mining. This article traces the country’s fascinating dia-
mond history before focusing on the geologic setting of the diamond occurrences, as well as the chal-
lenges and future outlook for production. The locations of the secondary deposits, principally in the
states of Minas Gerais and Mato Grosso, are presented.
W
hile Brazil is widely known for the produc- has produced a significant number of large rough di-
tion of colored gemstones, it is often not amonds (Svisero, 1995; Hoover and Karfunkel, 2009;
recognized as a source of gem diamonds also see table 1). The geologic explanation for this
(figure 1). Yet diamonds were discovered in the early abundance of large crystals is not yet clear. Highly
1700s by artisanal miners looking for gold along the valued colored diamonds are also found, including
banks of the Jequitinhonha River near the village of yellows and occasionally greens, pinks, and reds.
Arraial do Tijuco (later named Diamantina) in the
state of Minas Gerais (Calógeras, 1904; Leonardos,
1959; Barbosa, 1991; Cornejo and Bartorelli, 2010). In Brief
For the next 150 years, Minas Gerais was the world’s
major supplier of gem diamonds. Beginning in the • For approximately 150 years between the early 1700s
to the 1870s, Brazil was the world’s most productive
early 1840s, unusual polycrystalline black diamonds
diamond source.
(or carbonados) were also recovered, mainly from the
state of Bahia (Karfunkel et al., 1994). • Although it accounts for less than 1% of global produc-
tion today, Brazil has yielded a number of large dia-
As Africa emerged as the major diamond source monds as well as fancy-color and superdeep diamonds
in the 1870s, Brazilian mining decreased dramati- and carbonado.
cally within a decade, principally because of the low
• Due to the absence of obvious primary sources, the ge-
grade of the alluvial deposits. According to Kimber- ologic origin of these diamonds remains elusive.
ley Process statistics, Brazil produced 31,825 carats
• Production is expected to increase significantly with
in 2015, although this figure is likely a low estimate large-scale kimberlite mining in Bahia State.
considering the lack of accurate reporting there. De-
spite its low output relative to other sources, Brazil
In most countries, diamonds are usually recovered

See end of article for About the Authors and Acknowledgments.
GEMS & GEMOLOGY, Vol. 53, No. 1, pp. 2–33,
from primary kimberlite pipes of volcanic origin.
http://dx.doi.org/10.5741/GEMS.53.1.2 Brazilian diamonds, however, occur in secondary de-
© 2017 Gemological Institute of America posits over large watersheds or sedimentary rock for-
2 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

Figure 1. The Portuguese
Placa dos Três Militares,
or Badge of the Three
Military Orders, was
made in 1789, and is
now part of the crown
jewelry collection of the
Ajuda National Palace
in Lisbon. The badge
consists of Brazilian dia-
monds as well as emer-
alds and rubies, and it is
designed in a way to
conceal the metal
mounting and empha-
size the gems. Photo by
Manuel Silveira Ramos,
courtesy of the Ajuda
National Palace.
mations. The diamonds are recovered by artisanal BACKGROUND

miners called garimpeiros, using simple tools and op- The first European explorers to reach the New World
erating individually or in small groups. The diamonds in the late 1400s were focused on finding spices and
are found as loose crystals obtained either by panning gold rather than diamonds. The Italian explorer and
or dredging in the rivers, or by washing and/or break- navigator Christopher Columbus set sail under the
ing down the sediments that contain them (figure 2). Spanish flag in 1492 in search of a more direct route
Kimberlite bodies have been found in Brazil, but only to the “Indies,” the region of the Indian subcontinent
one in Bahia is now being mined. Despite extensive and eastern Africa that was rich in spices, gemstones,
exploration, none of the known kimberlites seems to pearls, and gold (Fernandez-Armesto, 2009). After a
account for Brazil’s abundance of alluvial diamonds. five-week westward voyage, Columbus discovered an
The long history of diamond production in the coun- island in the Caribbean and called it San Salvador. Por-
try, combined with the absence of obvious primary tugal and Spain, the global powers of the day, under-
sources, has led to various theories on the geologic stood that to avoid future confrontation, they had to
origin of these diamonds (see Chaves et al., 2001). subdivide the recently discovered lands in the New
This article will review the history of diamonds in World. The Treaty of Tordesilhas, drawn up by Pope
Brazil, the locations and geologic settings of the prin- Alexander VI and signed by the two empires in 1494,
cipal diamond deposits, and the mining methods used drew a meridian that would effectively divide South
to recover them. America (which had yet to be discovered by Euro-
BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 3

Figure 2. This selection
of diamonds recovered
from the Abaeté River
in Minas Gerais, weigh-
ing a total of 77.82
carats, is typical of the
material being found
today in Brazil’s allu-
vial deposits. Photo by
Robert Weldon/GIA,
courtesy of Giovani de
Deus Borges.
peans) from north to south (figure 3). Spain would re- nobleman Pedro Álvares Cabral was dispatched in
ceive the lands west of the meridian, including mod- early 1500 with a fleet of ships. His mission was to
ern-day Venezuela, Colombia, Ecuador, and Peru, reach India by traveling south around the tip of
while Portugal gained the eastern side. On his third Africa. His fleet landed instead on the coast of Brazil,
trip to the New World in 1498, Columbus discovered which he claimed for Portugal on April 22 of that
the mouth of the Orinoco River in what is today year. The voyages of Amerigo Vespucci, the Floren-
Venezuela. Here was the key to reaching a vast conti- tine navigator after whom the continent is named,
nental interior concealed in the jungles behind the At- took place around the same time. As he sailed south
lantic coastline. along the Brazilian coast, Vespucci realized that the
To open a new trade route to India that did not re- continent was much larger than previously recog-
quire passing through the Mediterranean, Portuguese nized. These expeditions were the first to bring back
Figure 3. The 1494 Treaty of Tordesilhas effectively divided the South America continent between Portugal and
Spain. The Cantino planisphere map, produced around 1502, shows the north-south Tordesilhas meridian line
passing through South America, whose geography was still largely unknown. Courtesy of Biblioteca Estense.

Figure 4. Ouro Preto,
once the settlement of
Vila Rica in Minas
Gerais, was the site of
Brazil’s first significant
gold discovery, in
1695. This attracted
explorers and fortune-
seekers to Brazil. Over
the next three cen-
turies diamonds and
other gem minerals
were found throughout
the region. Photo by
Robert Weldon/GIA.
shipments to Portugal of pau brasil (Brazilwood), a Jonathan, 2013). Brazil would emerge as a major
tree sought after not only for timber but for a desir- source for dozens of colorful gem species. It would
able red dye (Reid, 2014). Over the next two cen- also become, for a period, the world’s most prolific
turies, more European adventurers came to South source of diamonds (Cornejo and Bartorelli, 2010).
America, most of them seeking riches in the form of
gold, silver, and emeralds. PORTUGUESE EXPLORERS
Between 1534 and 1536, King John III of Portugal Much of the early exploration and colonization of
divided the coastal regions of Brazil into 15 captaincy Brazil was undertaken by the bandeirantes. The ban-
colonies to encourage development; these areas were deiras, funded by the crown and the Catholic
given to Portuguese noblemen to administer and ex- Church, were large quasi-military expeditions com-
plore. Within a few years, most of these captaincies prised of hundreds of free men and slaves who ven-
failed for a number of reasons. With this setback as tured into the interior, capturing and enslaving
well as the presence of French ships along the coast, indigenous groups they encountered (de Galvão
the Portuguese crown decided to turn Brazil into a Bueno, 1922). Along with a Jesuit priest, explorer
royal enterprise in 1549. Several successive gover- Francisco Bruzo Espinoza mounted the first bandeira
nors-general were appointed to administer the in 1554, in search of emerald. It was a short-lived and
colony, which was divided in 1621 into the states of fruitless expedition. Successive expeditions also
Maranhão (in the north) and Brazil (in the south). failed, though in 1572 Sebastião Fernandes Tourinho
The exploration of Brazil’s vast interior was left found green and blue gems (likely tourmaline) along
to Portuguese adventurers, known as bandeirantes the tributaries of the Jequitinhonha and Doce Rivers
(flag bearers), who claimed territories for Portugal in what is now Minas Gerais. The success of Tour-
and the Catholic Church. While Spanish conquests inho’s discoveries attracted more adventurers
in the New World met resistance from powerful em- (Draper, 1950a).
pires—the Aztecs in Mexico and the Incas in Peru— In 1695, at Rio das Velhas, Manuel Borba Gato fi-
the Portuguese were confronted with numerous nally discovered gold, triggering a rush that would
small native tribes and a forbidding interior. lead to the settlement of Vila Rica (later Ouro Preto)
Exploration and conquest on both sides of the in 1698 (figure 4). The town played a leading role in
Tordesilhas meridian would soon rewrite the gemo- Brazilian history for the next two centuries as the
logical texts of the time. For example, Spanish ex- capital of Minas Gerais from 1720 to 1897 (Dodge,
ploitation of emeralds in the territory of New 1922). Settlements at Mariana, São Bento, Serro Frio,
Granada (modern-day Colombia) would vastly and Arraial do Tijuco soon followed the gold discov-
change the world’s understanding and appreciation eries in those areas of Minas Gerais, and dozens more
for the gem, particularly in Europe (Weldon and sprang up in Brazil’s interior. The north-south road

TABLE 1. The largest rough diamond and carbonado discoveries in Brazil.a
Name Weightb (ct) Year Locationc
Diamond
Presidente Vargas 726.6 1938 Santo Antônio do Bonito River, MG
Unnamed 657 1936 Santo Inácio River, MG
Unnamed 630 1938 Santo Inácio River, MG
Santo Antônio 602 1994 Santo Antônio do Bonito River, MG
Goiás 600 (?) 1906 Veríssimo River, GO
Darcy Vargas 460 1939 Santo Antônio do Bonito River, MG
Charneca I 428 1940 Santo Inácio River, MG
Presidente Dutra 407.7 1949 Dourados River, MG
Coromandel VI 400.6 1940 Santo Inácio River, MG
Diário de Minas 375.1 1941 Santo Antônio do Bonito River, MG
Vitória I 375 1945 Abaeté River, MG
Tiros I 354 1940 Abaeté River, MG
Bonito I 346 1948 Santo Antônio do Bonito River, MG
Vitória II 328.3 1943 Abaeté River, MG
Unnamed 328 Unknown Santo Antônio do Bonito River, MG
Patos 324 1937 São Bento River, MG
Unnamed 309 Unknown Santo Antônio do Bonito River, MG
Star of the South 261.4 1853 Bagagem River, MG
Cruzeiro 261 1942 Santo Antônio do Bonito River, MG
Carmo do Paranaíba 245 1937 Bebedouro River, MG
Abaeté 238 1926 Abaeté River, MG
Coromandel III 228 1936 Santo Inácio River, MG
Mato Grosso 227 1963 MT
Regente de Portugal 215 1732 Abaeté River, MG
João Neto de Campos 201 1947 Paranaíba River, GO
Tiros II 198 1935 Abaeté River, MG
Cedro do Abaeté 194 1967 Abaeté River, MG
Tiros III 182 1935 Abaeté River, MG
Coromandel IV 180 1934 Santo Inácio River, MG
Estrela de Minas 179.4 1910 Dourados River, MG
Rio Preto 177 1972 Preto River, MG
Brasília 176.2 1944 Preto River, MG
Juscelino Kubitschek 174.5 1954 Bagagem River, MG
Tiros IV 173 1938 Abaeté River, MG
Minas Gerais 172.5 1937 Santo Antônio do Bonito River, MG
Princesa do Carmo de Paranaíba 165 1986 São Bento River, MG
Abaeté 161.5 1791 Abaeté River, MG
Portuguese 150 1790 Abaeté River, MG
Unnamed 147 1970 Abaeté River, MG
Coromandel V 141 1935 Santo Inácio River, MG
Nova Estrela do Sul 140 1937 Abaeté River, MG
Minas Vermelhas 140 1960 Santo Antônio do Bonito River, MG
Unnamed 139.5 1797 Abaeté River, MG
Charneca III 132 1972 Santo Antônio do Bonito River, MG
Great Brazilian 130 1791 Abaeté River, MG
Dresden Branco 122.5 1857 Bagagem River, MG
Cruzeiro do Sul 118 1929 Bagagem River, MG
Vargem I 110 1940 Santo Inácio River, MG
Jalmeida 109.5 1924 Tesouro, MT
Governador Valadares 108.3 1940 Bagagem River, MG
Independência 107 1941 Tejuco River, MG
Vargem II 105 1942 Santo Inácio River, MG
Abadia dos Dourados 104 1940 Dourados River, MG
Empress Eugénie >100 1760 Chapada Diamantina, BA
Lençóis, BA
Carbonado
Lençóis, BA
Sérgio 3,167 1905
Casco de Burro 2,000 1906
Xique-Xique 931.6 Unknown Andaraí, BA
Abaeté 827.5 1935 Abaeté River, MG
Bahia 350 1851 BA
Santa Ana 319.5 1960 Rosário do Oeste, MT
Pau de Óleo 113 1932 Andaraí, BA
a
Sources: Moreira (1955), Abreu (1973), Svisero (1995), and Hoover and Karfunkel (2009)
b
Carat weights are rounded to the nearest 0.1 ct
c
Abbreviations: BA = Bahia, GO = Goiás, MG = Minas Gerais, MT = Mato Grosso

N
² white stones such as diamonds.” In another account,

Minas Gerais Gabriel Soares de Sousa (1587) noted that fine, eight-
sided crystals—possibly diamond—had been found
Diamantina
during the dry winter months along certain rivers.
English adventurer Anthony Knivet, who was
captured and later escaped during a raid of Por-
tuguese settlements in southern Brazil led by the
English privateer Thomas Cavendish, described see-
Ouro Preto ing what he believed to be diamond crystals while
living among the natives in the late 1590s (Lessa de
Sá, 2015). This would have been another of the earli-
est accounts of Brazilian diamonds, though no details
were provided and the crystals could have been other
gems such as quartz.
Parati Rio de Janeiro
Atlantic In 1695, gold was discovered in the mountainous
Ocean region near Ouro Preto (Southey, 1819), and over the
following decade thousands of miners flocked from
Figure 5. The Estrada Real, or “royal highway,” was the coast to the interior in search of the precious
the principal 18th century route to transport goods metal (Machado and Figueirôa, 2001). Near the vil-
from Minas Gerais to the ports of Parati and Rio de lage of Arraial do Tijuco in the northern part of
Janeiro, a distance of several hundred miles, for Minas Gerais, unusually bright transparent crystals
shipment to Europe. kept showing up in the panned river gravels in the
early 1700s. In some cases, the miners disregarded
them. Dos Santos (1868) recalled that men some-
times used these crystals as small markers in card
system in Minas Gerais, known as the Estrada Real games. A similar account says that in 1721 a gold
or “royal highway,” begun in 1697, was built to pro- miner secured several of these markers, which were
mote trade, economic development, and communi- later recognized as diamonds by someone who had
cation. Minerals, wood, and other natural resources traveled in the Golconda region of India (Oakenfull,
were transported along the road from the interior to 1919; see also Cornejo and Bartorelli, 2010).
the coast for shipment to Lisbon (figure 5). Manufac- The discovery of Brazilian diamonds is supported
tured goods from Portugal were also carried inland. by the account of the Dutch explorer Jacob Roggeven
The road was strictly regulated by the crown to pre- (or Roggewein). As reported in Kerr (1824),
vent smuggling and unauthorized movement of Roggeven’s three ships anchored off the coast near
goods. It was also the only official route for traveling São Paulo for a short time in November 1721 before
inland (Cheney, 2005). resuming their voyage to the Pacific Ocean. Several
As Brazil’s population grew from European settle- crewmembers deserted to go to the diamond mining
ment, larger farms were required to grow crops, and area:
indigenous tribes were used as a source of slave labor.
Slaves were also imported from Portugal’s African A little time before the arrival of Roggewein, the Por-
tuguese had discovered a diamond mine not far from St
colonies to work the farms. The combination of slave Sebastian [a coastal village near São Paulo], of which at
labor and the colony’s immense wealth of minerals that time they were not in full possession, but were
and wood provided Portugal with unique imports for meditating an expedition against the Indians, in order
the European market (Reid, 2014). to become sole masters of so valuable a prize; and with
this view they invited the Dutch to join them, promis-
ing them a share in the riches in the event of success.
EARLY REPORTS ON BRAZILIAN DIAMONDS By these means, nine of our soldiers were tempted to
Early indications of Brazil’s diamond potential were desert. I know not the success of this expedition; but it
sporadic, but there is evidence that crystals were is probable that it succeeded, as great quantities of dia-
found in Bahia within a century of Columbus’s dis- monds have since been imported from Brazil into Eu-
rope. They are said to be found on the tops of
covery of the New World. In one of the earliest de- mountains among a peculiar red earth containing a
scriptions, historian Pero de Magalhães Gândavo great deal of gold; and, being washed down by the great
(1576) mentioned the existence of “certain mines of rains and torrents into the vallies, are there gathered.

Figure 6. The village of
Arraial do Tijuco in
Minas Gerais became
famous for Brazil’s first
alluvial diamond dis-
coveries in the early
1700s. As the settlement
grew into a colonial city,
it became known as
Diamantina. Photo by
Robert Weldon/GIA.
Eventually, reports of diamonds in Minas Gerais needs of both industries. Much like Ouro Preto, Dia-
began to reach Europe. Accounts from the colonial mantina grew into a rich and picturesque city in the
governor came to the attention of both King John V 18th and 19th centuries.
and the Catholic Church in Portugal, and the discov- Oakenfull (1919) reported that between 1732 and
ery was officially announced in 1729 (Ball, 1929). As 1771, “at least 1,666,500 carats of diamonds were ex-
Lisbon’s economic clout in Europe had waned some- ported to Europe.” That figure represents an average
what, this was welcome news. In Minas Gerais, there of about 42,000 carats per year. While miniscule by
was a diamond rush in the many rivers and streams today’s standards, it placed Brazil squarely as the
around Arraial do Tijuco. Portugal moved aggres- world’s top diamond producer at the time, eclipsing
sively to control the area, restricting gold and dia- India’s Golconda region. The glut of Brazilian dia-
mond mining and imposing high taxes. Despite monds pouring into Europe caused for a time a steep
efforts by the crown, clandestine mining and dia- decline in prices. Consequently, there were efforts to
mond smuggling increased. Draper (1950b) noted: protect the value of the Indian diamonds on the mar-
ket by disparaging the quality of Brazilian diamonds
The diamonds played a prominent part in shaping the
destiny of Portugal. Wealth derived from its diamonds (Cassedanne, 1989). Oversupply was put to an end in
not only helped place that country at the zenith of its 1739 when Portugal stepped in to monopolize the
glory but also contributed, at a later stage, to its release Brazilian mines, but by then the Golconda mines had
from the French occupation by paying part of the in- stopped producing. An arrangement with financiers
demnity exacted by France during the Napoleonic wars
(1807-1814).
in Amsterdam ensured a steady supply of Brazilian
rough for the city’s diamond cutters. And despite Por-
The occurrence of alluvial diamonds around Ar- tuguese oversight, smuggled Brazilian diamonds also
raial do Tijuco was reported in the scientific litera- found their way to London, another major center for
ture by de Castro Sarmento (1731). Over time, rough diamonds and jewelry production (Hofmeester,
Arraial do Tijuco became the town of Diamantina 2013a). For more on the early cutting styles for Brazil-
(figure 6), the diamond trading center during Brazil’s ian diamonds, see figure 7 and box A.
tenure as the world’s leading exporter. To properly During this period, additional alluvial sources
dredge, reroute, and mine the rivers around Diaman- were discovered. Diamonds were found in the Mato
tina required hard manual labor. This need coincided Grosso region in 1746, and subsequently in other
with the growth of the sugarcane business in Brazil’s rivers in Minas Gerais. The Jequitinhonha River near
northeast, and slaves were imported from modern- Diamantina and the surrounding area yielded im-
day Angola, Congo, and Mozambique to fill the pressive quantities of alluvial diamonds as well (fig-

Figure 7. This star
motif necklace with
“Old Mine” cut Brazil-
ian diamonds was
made in 1865. Photo by
Luisa Oliveira, cour-
tesy of Ajuda National
Palace.
ure 8), including treasure troves in “potholes” or royal licenses to individuals to work portions of the
traps in the river bedrock. But the primary source of diamond deposits. Guarded mule trains transported
diamonds in Minas Gerais—actual diamondiferous the diamonds along the Estrada Real to Parati and
kimberlites—remains elusive. As Sinkankas (1977) Rio de Janeiro (“Diamond carriers,” 1842). This li-
noted: censing system eventually broke down, as bands of
runaway slaves and garimpeiros clandestinely
The original host rocks of the diamond crystals remain
unknown despite much past speculation, “proof” of worked the areas or stole the recovered diamonds. In
host rock sources by some authorities, and modern in- 1772, the government abandoned this system and
tensive geologic exploration of the regions presently took full control of the workings around Diamantina
yielding diamonds. Brazilian diamonds are found in a (von Spix and von Martius, 1824).
sort of matrix to be sure, this being a conglomerate of The English mineralogist John Mawe is proba-
well-rounded pebbles cemented together with sand and
iron oxides. bly the most celebrated European traveler to
Brazil’s interior, having been the first foreigner to
In 1733, Portuguese authorities began granting reach the diamond mines. In 1809 he received ap-

BOX A: THE EVOLUTION OF THE OLD MINE CUT
Demand for diamonds and experimentation with cutting
styles dovetailed with Brazilian diamond discoveries in
the 1700s. This convergence of marketing and manufac-
turing led to the rapid growth of diamond mining in
Brazil. As diamonds suddenly became available to a
larger audience, cutters in Amsterdam and London em-
barked on a new cutting style that became the precursor
to the modern brilliant cut.
Demand. Jewelry historian Marcia Pointon (1999) wrote
that by the late 1600s, diamond adornment was one of
the “publically recognized criteria for gauging social suc-
cess…” Louis XIV’s thirst for diamonds best illustrated
this trend among Europe’s wealthy classes: “While also
influencing concepts of taste, the demand of the King
himself was hugely significant… in 1669, Louis had pur-
chased an unprecedented 1.5 million livres worth of dia-
monds (about US$21.45 million), including 900,000
livres worth from gem merchant Jean-Baptiste Tavernier;
all of course were Indian diamonds” (McCabe, 2008). In
1691, an inventory of the king’s collection showed it was
composed almost exclusively of diamonds cut in the
brilliant style (Tillander, 1995). By 1715, just prior to his
death, the king displayed a collection of cut diamonds
worth 12 million livres, or more than US$170 million
today (DeJean, 2005).
Global Trade. While diamond rough from India had been Figure A-1. Illustrations of the diamond faceting
sporadic by the early 1700s, the new finds in Brazil stim- arrangements known in France by the mid-1700s.
ulated the global diamond trading network (Inveness, These were compiled by Denis Diderot, chief editor
2014). Within the country, the Portuguese crown author- from 1745 to 1772 of the 28-volume Encyclopédie, ou
ized “exploiters,” who oversaw individual miners, to Dictionnaire Raisonné des Sciences, des Arts et des
work the deposits in Minas Gerais. In turn, these indi- Métiers, one of the principal works of the Enlighten-
viduals paid the Portuguese treasury with a tax per miner ment. This page is taken from his 1771 volume enti-
in their employ (Hofmeester, 2013b). The arrangement tled Recuiel de Plances, sur Les Sciences, Les Arts
worked well. Diamond supplies increased fivefold in Eu- Liberauz, et Les Arts Mechaniques (p. 732).
rope, followed by a 50% drop in prices (Lenzen, 1970;
Yogev, 1978). The Brazilian production triggered a re-
structuring of the global diamond supply chain, as lower
prices only served to compound the demand. More peo-
Soon after the Brazilian discoveries, the most promi-
ple could afford them, thereby stoking a continued ex-
nent cutting styles in Europe were the brilliant and, for
ploration within the country. Furthermore, the
smaller stones, the rose cut (Jeffries, 1750). Originally the
increasing market stirred a growth in the European dia-
“brilliant” cut represented an arrangement of 58 facets,
mond cutting industry.
typically cut in a cushion or somewhat squarish shape
Cutting Styles. Increased supplies of rough diamonds following the outline of the most common type of rough,
also allowed for experimentation in cutting styles by the octahedron. The brilliant was also cut in round, oval,
French cutters (figure A-1). The brilliant cut, introduced and pear shapes, as seen in figure A-2.
around 1660–1670 in France, was also associated with The brilliant style of cutting, favored by Louis XIV in
London by the mid-1700s (Tillander, 1995). the late 1600s, was suddenly available to many. The re-
proval and funding from the Portuguese crown for methods emerges in his 1812 account, Travels in
his visit. His descriptions of the diamond region are the Interior of Brazil:
still some of the most detailed. A vivid portrayal of A shed is erected in the form of a parallelogram…con-
the washing plants and the diamond discovery sisting of upright posts which support a roof thatched

sult was that diamonds cut in this style glittered in the
candlelit ballrooms of the 1700s.
Rise of the Mine Cut. India and Brazil were the classic
sources for high-color diamonds in the 1700s and most
of the 1800s. In comparing the sources, gem expert
Edwin Streeter extolled the virtues of Brazilian stones.
“A fair portion of the South-African stones are diamonds
of the first-water [i.e, best clarity and colorless], rivalling
in beauty and purity the finest Brazilian and Indian
Stones” (Streeter, 1882).
In 1891, one diamond manufacturer wrote: “Many di-
amond dealers sell African diamonds at an enhanced price
Figure A-2. The most common outlines for “brilliant
to the uninitiated public as ‘old mine stones,’ presumably
cut” diamonds, as illustrated in A Treatise on Dia-
referring to diamonds from the ‘old mines’ of Brazil or
monds and Pearls by David Jeffries in 1750. Note that
India… Great variations in color and purity of course
all have 58 facets in a similar layout, and the only
exist, but it is not fair to assume that all the best come
differences are their outlines.
from Brazil or India” (Stonestreet, 1891). Although India
is included in this definition, it does not refer to diamonds
cut in India; rather, it only pertains to Indian diamonds
telle (1903) wrote,“This is now the trade term for blue
fashioned in Europe using the brilliant cutting style.
goods, the name ‘Old Mine’ being confined to the old
Not all of Brazil’s diamonds were “of the first-water,”
cut stones of that quality [referred to in the early 1900s
however. Streeter (1882) did make a distinction between
as blue-white] which occasionally drift back into the
Brazil’s various sources. “The Diamonds of Brazil are
market from private hands” (Cattelle, 1903). More re-
known in commerce as (1) Diamantina diamonds, and
cently, diamond historian Herbert Tillander (1995) indi-
(2) Cincora [sometimes spelled Sincora] diamonds,” he
cated that “Old Mine or Old Miner…were introduced
wrote. “The latter are of less value than the former, be-
at the end of the last century to describe the type of cut
cause they are not of such pure water, nor of so good a
applied to most Brazilian rough.” He later noted that
shape. In Matto-Grosso, the diamonds are small, but of
this term is a “popular but inaccurate term for the cush-
purest water, and in their rough state have a peculiar lus-
ion-shaped brilliant, which dates back to the seven-
tre, which is seen in none other than Brazilian dia-
teenth century.” He pointed out that “Old Mine,” once
monds” (Streeter, 1882). He also noted that diamonds
the epitome of cutting style, came to be viewed with
from Bahia were of lesser quality.
disdain. It was seen as a less desirable cut, referring sim-
By the early 1900s, the “Old Mine” designation had
ply to an old style of shape outline (squarish) and cutting
evolved to refer to aspects of the cut as high-color quality
style, with steeper pavilions and crowns, shorter half
also became associated with diamonds from the mines
facets, and large culets. This cut was replaced with a
in South Africa:
more light-efficient shape and style of cutting known as
Specifically, Brazil diamonds from the old diggings (prob- the round brilliant.
ably the Diamantina area); as generally used, old-cut dia- Today, nostalgia for historic diamonds and that style
monds of good color…. This quality of color was of cut has shed a kinder light on the so-called Old Mine
long-termed ‘Old Mine,’ from Brazil... Since the opening cut. Such gems, when they are found, are often viewed
of the African mines, a new name has been given to ma- as collectors’ items (figure 8). Because Old Mine cuts
terial of this character. A mine called Jagersfontein have enjoyed such a storied history, some contemporary
yielded many bluish-tinted diamonds, and it soon became
jewelry designers continue to use them successfully in
customary in trade circles to call everything of that grade
‘Jagers’ (pronounced Yahgers). (Cattelle, 1903)
their creative expressions. Out of strong demand for this
vintage look, several companies now deliberately cut
Thus, the appearance of African high-quality dia- Old Mine faceted diamonds, but with thicker girdles
monds began to shift the definition of “Old Mine.” Cat- than in their heyday.
with long grass. Down the middle of the area of this thick. On the other side of the area is a flooring of
shed a current of water is conveyed through a canal [sloped] planks… The flooring is divided into about
covered with strong planks, on which the cascalhão [a twenty compartments or troughs…by means of planks
Portuguese term for gravel] is laid two or three feet placed on their edge. The upper ends of these troughs

Figure 8. Crystals such
as this group from the
Jequitinhonha River
were once used as
markers in card games,
as they were not recog-
nized as diamonds.
Photo by Robert Wel-
don/GIA, courtesy of
Gar Mineração.
communicate with the channel, and are so formed that German nobleman Baron Wilhelm Ludwig von
water is admitted into them…The negroes enter the Eschwege fought alongside the Portuguese during the
troughs, each provided with a rake…with which he
Napoleonic Wars, and for his service he was granted
rakes into the trough…[the] cascalhão. The water being
let in upon it, the cascalhão is spread abroad and con- access to Minas Gerais as a consultant for the gov-
tinually raked up to the head of the trough, so as to be ernment. In 1811, he settled in Ouro Preto to lend
kept in constant motion. [After the finer sediment is his qualifications there as a mining engineer. Having
removed], the gravel-like matter is raked up to the end traveled throughout Minas Gerais over the course of
of the trough, and…then the whole is examined with
a decade, he was instrumental in not only document-
great care for diamonds. When a negro finds one, he im-
mediately stands upright and claps his hands, then ex- ing existing techniques used for diamond and gold
tends them, holding the gem between his fore-finger mining, but also offering recommendations for the
and thumb; an overseer receives it from him, and de- use of more modern procedures (von Eschwege,
posits it in a gamella or bowl…In this vessel all the di- 1833).
amonds found in the course of the day are placed, and
Richard Francis Burton, an English explorer, was
at the close of work are taken out and delivered to the
principal officer, who, after they have been weighed, noted for his travel in sub-Saharan Africa in search
registers the particulars in a book kept for that purpose. of the origins of the Nile River as well as other ex-
ploits in the 1850s. In 1865, he was given a diplo-
Mawe’s successful journey prompted other Euro-
matic post in Brazil by the British government. An
pean explorers to follow in his footsteps. His account,
1867 expedition in the Brazilian highlands took
and the illustrations in his book by noted mineralo-
him to Diamantina and the São Francisco River.
gist James Sowerby, inspired many European scien-
His account of reaching the city, and his descrip-
tists from a variety of disciplines to explore Brazil.
tion of the diamond mining region, are well chron-
Swiss botanist Auguste de Saint-Hilaire, who trav-
icled in his 1869 book, Explorations of the
eled there in 1817, described Diamantina as a “valley
Highlands of Brazil.
of diamonds” (de Saint-Hilaire, 1833). He also
When Brazil gained independence in 1822, Por-
painted a dismal image of the conditions for both the
tuguese control of diamond mining ceased in Minas
garimpeiros and slaves around Diamantina. The
Gerais. It was replaced with a system of free exploita-
slaves received little food or rest, and their panning
tion except along the Jequitinhonha River Valley,
and sorting of diamonds was subject to strict punish-
which remained under some government control
ment. The only reprieve was good luck:
until 1845. By the 1880s, production had dropped sig-
From this wretched mode of life a fortunate accident nificantly due to several factors—the low grade of the
sometimes frees a slave. When he happens to find a dia-
mond weighing an octavo, or 17 ½ carats… his value is alluvial deposits, competition from diamonds from
ascertained by the administration, the price is paid to his South Africa, and Brazil’s abolition of slavery in 1888,
owner, he is dressed, and set at liberty. (Mawe, 1812) which meant the loss of cheap labor.

N
A tlantic
² O cean
1. Ancient Archean or Proterozoic “basement”
AM
SL cratons that are covered in many places by ge-
ologically younger sedimentary rocks
E
2. Weathered Proterozoic mountain fold belts that
6 separate the cratons
F C
AM
4 3. Paleozoic sediments that fill topographic basins
D within the cratons
SF
4. Mafic and/or ultramafic igneous intrusions of
3 5 B
Sedimentary units A Mesozoic and Cenozoic age
Neoproterozoic mobile
belts and basement 2
inliers DIAMOND OCCURRENCES
Paleoproterozoic - PP
Diamonds occur in several different geologic settings
LA
Archean nuclei across Brazil. They are usually recovered as loose
Limits of cratons crystals along rivers in unconsolidated alluvial or gla-
At lantic
0 290 580 870 RP 1 Ocean
cial sediments, or they are embedded in conglomer-
km
Sedimentary units - intracontinental basin:

(A) Paraná, (B) San Franciscana, (C) Parnaíba, (D) Parecis, (E) Solimões-Amazonas,
(F) Acre Figure 10. A map showing the distribution of various
Neoproterozoic mobile belts and basement inliers: categories of diamonds found in Brazil. The name of
(1) Dom Feliciano, (2) Ribeira, (3) Paraguai, (4) Tocantins, (5) Brasília, the mining area is shown (with the abbreviation of
(6) Rio Negro-Sergipano
the Brazilian state), preceded by its location number
Paleoproterozoic - Archean nuclei:
(RP) Rio de la Plata, (LA) Luis Alves, (PP) Paranapanema, (SF) São Francisco,
on the map.
(SL) São Luíz, (AM) Amazonas
N A t l ant i c
15
²
Figure 9. A simplified map of the major geological RR

AP Ocean
1
14
regions of Brazil.
AM PA 13 MA 30 CE RN
R N

PI PB
PB
PE
E
GEOLOGIC SETTING AC RO 2 3
32 31 TO
12
29
2 AL
L
6 37
36 28
Brazil contains a variety of rocks and geotectonic 35 MT SE
S
BA S 11
6338
17
structures (see Branner, 1919; de Almeida, 1967; de 4
34 1
16
333 20
0 GO
18 9
Almeida et al., 1981). The present-day land masses in 19
24 MG
2 10
1
233
25
2
8
the Southern Hemisphere were assembled as part of 5 6 7
4 27 ES S
a large Precambrian supercontinent called Gondwana Alluvial diamond
MS 222 26
Lithospheric 2 3
(de Brito-Neves and Cordani, 1991; Cordani et al., diamond
SPP 21
1
RJ J
2003, 2009). Much of the current geological setting of S Carbonado PR 1
Brazil was established in the Pan-African/Brasiliano Superdeep
SC
orogenic cycle (approximately 440–1000 Ma). During diamond
RS A t la n tic
this time, preexisting Archean cratons were brought 0 29
90
9 580 870 O c ea n
together by plate tectonics and assembled along km
m
mountain (or orogenic) fold belts representing the Alluvial diamonds:

zones of continental collision. At about 180 Ma, (1) Tibagi, (2) Franca, (3) Vargem
Vargem Bonita (4) Romaria, (5) Estrela do Sul, (6) Coromandel,
(7) Abaeté, (8) Diamantina, (9) Grão Mogol, (10)
0) Canavieiras, (11) Lençóis, (12) Gilbués,
Gondwana began to break up into separate pieces, (13) Marabá, (14) Vila Nova, (15) Tepequém,
Tepequém, (16)
16) Pimenta Bueno, (17) Juína,
(18) Diamantino, (19) Poxoréu, (20) Barra do Garças
with South America and Africa gradually drifting
Lithospheric diamonds:
apart to create the Atlantic Ocean (Hasui, 2012). (21) Canastra 1, (22) Romaria, (23) T Trrês Ranchhos 4, (24) TTucano,
ucano, (25) Abel Régis,
Since the two continents were connected prior to this (26) Cana Verde, (27) Maravilhas, (28) Salvador or 1, (29) Braúna 3, (30) Moana 11,
(31) Juína 5, (32) Collier 4, (33) Carolina 1, (34) Comet 1, (35) Cosmos 1, (36) Cosmos 2,
breakup, corresponding rock types and certain geo- (37) Cosmos 3, (38) T Tumeleir
umeleiro 3
logical structures occur on either side of the ocean. Carbonado:

Lencóis (BA)
As shown in the simplified map in figure 9, the main Superdeep diamonds:
Juína
geological units consist of:

SW NE
Figure 11. This diagram il-
Santo Inácio River
lustrates the kinds of sec-
ondary diamond deposits
Colluvial deposits Alluvial deposits
found along the Santo Iná-
cio River in Minas Gerais—
alluvial and colluvial
deposits along the river
channel, occasional kim-
berlite pipes (often unrec-
Covering sediments Diamictite (diamond-bearing?)
ognized because they are
Conglomerate (diamond-bearing) Phyllite
deeply weathered and cov-
ered by younger sedi-
Kimberlite intrusion Quartzite ments), and sedimentary
conglomerates exposed on
the adjacent hillsides.
ates or metamorphosed sedimentary rocks. They are the sources in Minas Gerais and Mato Grosso have
also found in small amounts in a few known kim- been the most economically important over the past
berlite bodies. Figure 10 shows the locations of the three centuries.
major mining areas (both historical and current) for The recovery of diamonds has involved simple
the different categories of diamond deposits that will methods used by artisanal miners. Loose diamonds
be discussed. and other heavy minerals are panned directly from a
The general geology of Brazil and its important flowing river, or are washed from sediments through
mineral resources have been reviewed by Berbert et a series of sieves. The diamond-carrying sediments
al. (1981), Delgado et al. (1994), and Machado and are mixtures of clay, sand, silt, gravel, and angular or
Figueirôa (2001). Further information on the dia- rounded pebbles; they are often found in horizontal
mond occurrences themselves can be found in Gor- bands or layers, with their contents washed down
ceix (1882), Pearson (1909), Cassedanne (1989), and from higher terrain.
Cornejo and Bartorelli (2010). The main detrital diamond deposits are called
cascalhos (river gravels). Besides quartz, the gravels
Alluvial Diamonds. Alluvial diamonds are recovered may contain heavier pebbles of iron oxide minerals.
as loose crystals directly from rivers or streams, or They are geologically classified as eluvial, colluvial,
from geologically recent unconsolidated sediments or alluvial gravels depending on their topographic lo-
nearby. These deposits occur throughout Brazil, but cation and distance from the source.
Figure 12. At larger alluvial mining operations, such as the Duas Barras mine on the Jequitinhonha River (left),
sediments are passed through a washing plant (right) where high-pressure water hoses are used to remove larger
pebbles as well as clays and other lighter materials. Diamonds along with heavy minerals are collected as a con-
centrate. Photos by Robert Weldon/GIA.

Figure 13. At the Duas Barras mine, a miner pans the finest-size heavy-mineral concentrate to recover small parti-
cles of gold, the production of which helps offset the cost of diamond recovery. Photos by Robert Weldon/GIA.
Figure 11 illustrates several types of diamond de- the river by suction pumps. The pumped water and
posits in relation to the topography. The diamond- gravel is then passed through a sluice box that re-
bearing gravels occur as deposits along the bottom or moves the lighter material to recover diamonds and
banks of rivers, and along paleo-river channels that other heavy minerals (Barbosa, 1991). Large, expen-
are buried beneath younger rocks. Remnants of the sive dredges are less common because the low grade
buried channels outcrop along riverbanks. The dia- of the deposits makes them too costly to exploit with
monds may also occur in sediments exposed on hill- this kind of equipment.
sides. The garimpeiros use various approaches—past
discoveries, rumors, intuition, and other means—to Jequitinhonha River, Minas Gerais. Following the
identify potential mining areas. They often wait for discovery of diamonds in alluvial gravels around Dia-
the dry season to explore the accessible riverbeds, as mantina in the early 1700s, exploration took place
diamonds can be concentrated in depressions or pot- along the Jequitinhonha River as well as other rivers
holes on the underlying bedrock. In some cases, the and streams in the mountainous Serra do Espinhaço
course of a river is blocked with a dam, or the water (Derby, 1906; Thompson, 1928). Diamonds occur in
is diverted using a large sluice, providing access to metamorphosed sediments of the Mesoproterozoic-
the gravels. Once a potential spot is selected, it may age Espinhaço Supergroup. As a result of the weath-
be necessary to remove overlying layers of silt, clay, ering of these sediments, diamonds are found in the
or sand to reach the diamond-bearing gravels, which rivers and streams and in exposed gravels (Karfunkel
tend to lie on or above the solid bedrock. Larger peb- et al., 1994; Chaves et al., 2001). Mining in this re-
bles are removed, and the gravels are then washed gion has been carried out by garimpeiros and later by
using sieves or pans to pick any diamonds by hand larger companies, though the peak of activity oc-
(figure 12). The finest-size material is then checked curred between about 1740 and 1830.
for diamonds as well as small particles of gold (figure The alluvial deposits of the Jequitinhonha River
13). have been responsible for most of Brazil’s historic di-
Larger hydraulic systems have sometimes been amond production (Chaves and Uhleim, 1991). In
used to wash diamond-bearing sediments exposed in this region, the alluvium consists of a 10- to 45-
an outcrop. In the late 1800s and early 1900s, diving meter-thick layer of gravel capped by a sand-clay
suits were used by some miners to retrieve sediments layer. The diamond content of the gravel is approxi-
from the river depths. These traditional mining ac- mately 0.6 carats per cubic meter of sediment (ct/m3;
tivities have been supplemented by mechanized Dupont, 1991). In 1966, Mineração Tejucana SA, a di-
methods; for example, gravel may be extracted from amond mining company based in Diamantina, began

115
N
350
²
Davinópolis
Lagamar
using dredges to work the alluvium of this river. The Rio Paranaíba
company processed 9 million cubic meters of allu-

Rio Preto
vium per year to recover 70,000 carats of mostly
gem-quality diamonds (along with 150 kg of gold) Três Ranchos
176.2
Rio Sto Antõnio do Bonito
90
over a period of several years (Dupont, 1991). Douradoquara
Rio Santo Inácio 140 Rio Sto
Between December 2007 and April 2008, the 91 261
172.5 Antonio
Coromandel 79.80 328 das Minas
Peçanha mine near Diamantina, owned and operated Grupiara 104 24.70 346
407.68 75 375.1 309 Vermelhas
by Recursos Mineração Ltda. (a subsidiary of Brazil Rio Bagagem 82 10.40 726.6
602
Minerals Inc.) yielded 4,512 carats of diamonds and 11.08 119.5 Star of the 76 56 5.25
428 460
118 South Rio 228
Dourahinho 105 90
60 110
261.38
8.95 kg of gold, with average grades of 0.06 ct and 82.25
80
0.14 g per cubic meter, respectively (Watkins, 2009). 99 Monte Carmelo 138.6
Cities

50 Roads
Mineraçao Rio Novo is a subsidiary of Andrade 108.3
128.48
Rio Dourados
Rivers
Gutierrez SA, a private multinational company head- 67
147 Rio Perdizes Large diamonds
quartered in Belo Horizonte that has been mining 174.5
Romaria Fancy-color
diamonds
along the Jequitinhonha River since 1988. The de-
posit has a diamond content of about 1.1 ct/m3 and Figure 14. The Alto Paranaíba region of western
an average production of 25,000 carats per year Minas Gerais has produced large diamonds along
(Chaves and Uhlein, 1991; Watkins, 2009). Mining river systems near the towns of Coromandel and Ro-
maria. The symbols indicate where individual large
involves removing the overlying sediments and using
diamonds were found; the value shown next to each
a bucket dredger to excavate the mineralized gravel. symbol represents their carat weight. The site where
Diamonds are recovered from the coarser concen- the famous Star of the South diamond was recovered
trate in a vibrational circular sieve, while gold is re- in 1853 along the Bagagem River is indicated in the
trieved from the finer concentrate. lower left corner.
Another important project in the Diamantina re-
gion is run by Mineraçao Duas Barras, a subsidiary
of Brazil Minerals Inc. Data released by the company
showed a yield in 2008 of 32,008 carats, with an av- Poxoréu and Chapada dos Guimarães, both located
erage value of US$145 per carat (Watkins, 2009). Ge- near the state capital of Cuiabá, display a dodecahe-
ological studies of this deposit gave an indicated dral habit, while the rest display octahedral and
resource of approximately 1,639,000 m3 of diamon- mixed habits (Zolinger et al., 2002). All of the dia-
diferous gravel with an average grade of 0.16 ct/m3. monds are recovered from sedimentary rocks of the
A weekly output of nearly 0.5 kg of gold provides a Upper Cretaceous-age Bauru formation (Souza, 1991).
source of revenue to support mining operations. Similarly, diamonds mined elsewhere in southeast-
ern Mato Grosso, around the cities of Tesouro,
Poxoréu, Mato Grosso. Poxoréu is located on the Guiratinga, Alto Garças, Barra do Garças, and Batovi,
northwestern edge of the Paraná Basin. Diamonds are related to conglomerates of the same formation
were discovered in the Coité River in the 1930s, (Weska, 1996).
causing an influx of prospectors. The alluvial de-
posits are distributed along the Coité, São João, Pox- Coromandel, Minas Gerais. The mines around Coro-
oréu, Alcantilados, Pomba, and Jácomo rivers. mandel began operating in the late 18th century (Des
Until the mid-1970s, mining was carried out Genettes, 1859). The town was probably named by
using only picks, sieves, and panning. St. Felix Ltda. Portuguese traders for the Coromandel region of
subsequently conducted studies in the Coité River India, an important diamond source since antiquity
Valley to characterize the diamond reserves—38.6 (Legrand, 1980). Also known as the Alto Paranaíba
million cubic meters of gravels were processed, of province or the Triângulo Mineiro, this area is the
which about 42% contained diamonds. The average second most important source of diamonds in Minas
diamond content was 0.05 ct/m3, with an estimated Gerais, serving as the center of the local diamond
recoverable reserve of 662,000 carats, 27% of it gem- trade (figure 14; Leonardos, 1956; Kaminsky et al.,
quality (Souza, 1991). According to this reference, 2001; Karfunkel et al., 2014).
there were about 2,500 miners and 150 dredges in the The town itself is located on the eroded rocks of
region in the 1980s. the Brasilia fold belt in western Minas Gerais. The
Approximately 70% of the diamond crystals from mines occur in alluvial deposits of Cenozoic age,

US$650,000. The Giacampos Red, a 5.25 ct crystal
found in 2005 in the Santo Inácio River, yielded a
1.57 ct faceted oval. In May 1993, a 602 ct brown di-
amond discovered in the Santo Antonio do Bonito
River (Haralyi et al., 1994) became the second-largest
recorded Brazilian gem diamond (surpassed only by
the colorless Presidente Vargas). In the fall of 2014,
at a site along the Dourahinho River, the mine own-
ers found a flattened 138.6 ct pink crystal, which
they were able to sell almost immediately.
In the Coromandel region, large diamonds have
also been recovered from the Abaeté, São Bento, Bor-
rachudo, and Indaiá rivers, all occurring in the São
Figure 15. The Bagagem River, shown near the town
Francisco River Basin (figure 16). They are also recov-
of Romaria, has yielded numerous large diamonds
over the past two centuries, including the Star of the
ered in the alluvium eroded from Cretaceous con-
South. Photo by Robert Weldon/GIA. glomerates of the Capacete formation (Read et al.,
2004). Examples of large diamonds include the 27.09
ct vivid pink Queen Giacampos and the 13.90 ct
Moussaieff Red, both found in the Abaeté River (King
mainly along rivers that cut through the recent sed- and Shigley, 2003). Other discoveries include the 0.59
iments as well as metamorphic rocks of the crys-
talline basement.
The main diamond-mining areas are situated
along the Paranaíba, Santo Inácio, Dourahinho, Santo Figure 16. This map shows the Abaeté and nearby
Antônio do Bonito, Santo Antônio das Minas Ver- river systems in Minas Gerais that have produced
melhas, and Bagagem rivers, and to a lesser extent large diamonds for many years. Again, the value
the Preto and Dourados rivers. In the Paranaíba shown next to each symbol represents the carat
River, the largest waterway in the region, most min- weight of the rough crystal found at the location. Dis-
coveries of colored diamonds are indicated by colored
ing is done by means of rafts anchored at various
symbols. The rivers shown on the map, some of
points along the riverbed. Heavy mineral concen- which flow into the Tres Marias reservoir, are all lo-
trates from all of these rivers and streams contain tra- cated within the southern portion of the São Fran-
ditional kimberlite indicators: pyrope garnet and cisco River basin.
magnesian ilmenite in similar proportions, followed
by zircon and chromite. Cities N
²
Roads 11.20 31.50
The Coromandel region is famous for its periodic
Rivers
discoveries of large rough diamonds (table 1). The Tréz Marias
Large diamonds
best-known examples are the 726.6 ct Presidente Var- Fancy-color 12
gas and the 261.38 ct Star of the South (Hussak, 1894; diamonds São Gonçalo do Abaeté
Varjão de Minas
Reis, 1959; Smith and Bosshart, 2002; Balfour, 2011; 25.25

Rio Borrachudo
also see Abreu, 1973; Svisero, 1995). The Presidente 4.80

Vargas was found in 1938 along the Santo Antônio
5.55 64
do Bonito River. The Star of the South was discov- 192
58.20

27.09 250 Morada Nova de Minas
ered in 1853 in the Bagagem River, near a village then Rio Areado
300
known by the same name (Dufrenoy, 1855). The area 105
continues to occasionally yield large diamonds, at- 452 Rio Indaiá
testing to the great economic potential of these de- Rio São Bento
Biquinhas
166
posits (figure 15). Rio Abaeté

101 152 Paineiras
Colored diamonds finds are not rare in Brazil, and 0.59

13.90
65
they are often not officially registered. Still, there are Carmo do Paranaíba
Tiros
some notable examples. In 1998, a 7.0 ct pink dia- Arapuá
65
46.60
mond found in the Vargem mining area near Coro- 12.30
98 Cedro do Abaeté Abaeté
mandel was purchased by a diamond dealer for

ct Rob Red, from the São Bento River, and two rough try’s largest producer of industrial diamonds. At its
stones of 41.44 and 58.2 carats from the Abaeté River peak, some 30,000 miners produced around 400,000
(see references listed in table 1). carats per month (Costa, 2013). Examination of 2,200
The remarkable Bragança diamond—either 1,680 crystals by one of the authors (DPS) in 1985 revealed
ct (Mawe, 1812) or 1,730 ct (Ferry, in Reis, 1959), de- that 90% were of industrial quality, gray to dark
pending on the source—was found in 1797 in the brown in appearance. Most were malformed or cor-
Abaeté region. For many years, the Bragança be- roded crystals and crystal fragments.
longed to the Portuguese crown jewels (Balfour, Late in the decade, large variations were recognized
2011). Because of its large size and pale appearance, in the abundance of alluvial diamonds in the area. The
it was once considered a topaz by some historians Cinta Larga River basin and its tributaries—the São
and traders (Reis, 1959; Balfour, 2011). More recently, Luiz, Porcão, Samanbaia, Mutum, and Central rivers,
Galopim de Carvalho (2006) discussed the possibility as well as the Sorriso and Duas Barras streams—all
that it was a pale aquamarine. Yet this region’s geol- contained large amounts of alluvial diamonds (up to
ogy is not favorable to the occurrence of either topaz 6–7 ct/m3). In the Juína Mirim River basin, however,
or aquamarine, and it is more likely that the Bra- there were much lower amounts (0.6–0.8 ct/m3). Be-
gança was indeed a diamond. Its current whereabouts sides the high yield of industrial diamonds, the allu-
are unknown. vial deposits were also noteworthy for their large
The origin of these large diamonds continues to diamonds, similar to Coromandel and Abaeté in west-
elude both researchers and mining companies. The ern Minas Gerais. Haralyi (1991) reported finding eight
geological literature reflects two opposing views. crystals between 49.5 and 232 ct, along with six in-
One theory holds that diamonds of all sizes, weath- dustrial diamonds between 52 and 263 ct. According
ered from kimberlite pipes, were recycled into the al- to Costa (2013), in September 1992 a garimpeiro found
luvium from the conglomerates of the Capacete a 413 ct diamond in the village of Poção. Le Noan
formation and distributed throughout the Coroman- (2008) mentioned the discovery of a 452 ct “megadia-
del region (Svisero et al., 1981; Sgarbi and Chaves, mond” in alluvial deposits of the São Luiz River. Al-
2005). Based on field evidence that seems incompat- though sporadic, these finds indicate that the area
ible with a kimberlite “pipe” origin, the other theory around Juína also has potential for large diamonds.
proposes that the diamonds were transported west- Watkins (2009) reported that Juína remains the coun-
ward into the region from the São Francisco craton try’s leading producer of industrial diamonds. In the
by glacial events in the Neoproterozoic and Paleozoic last decade, the estimated output has been on the
periods (Tompkins and Gonzaga, 1987; Gonzaga and order of five million carats.
Tompkins, 1991). This alternate view has been diffi-
cult for some to accept because there is little under- Other Sources. Diamonds were discovered in the
standing of the transport, concentration, and state of Roraima in the early 20th century in the Mau
dispersion mechanisms of diamonds in glacial sys- River region and later near Vila do Tepequém (located
tems in South America (Gonzaga et al., 1994). in the Serra do Tepequém). The deposits occur in sed-
iments that originated from conglomerates of the
Juína, Mato Grosso. The Juína diamond region lies in Neoproterozoic Tepequém formation. Present-day al-
the southern part of the Amazonian craton and in the luvial placers occur where the diamondiferous gravel
northwest of Mato Grosso. Alluvial diamond deposits ranges up to 0.80 m thick, and where indicator min-
are located in the basins of the Juína Mirim, Vinte e erals are present. In some locations, gold is also re-
Um de Abril, and Cinta Larga rivers, all situated covered (Grazziotin and Andrade, 2011).
southwest of the city of Juína. The alluvium comes In the state of Piauí, diamonds mainly occur near
from erosion of the crystalline basement rocks and the city of Gilbués in alluvial deposits along the Ri-
from overlying sediments of the Parecis basin, located achão stream, and in colluvium near the towns of
on the southeastern edge of the Amazonian craton Gioninha, Bom Jardim, Compra Fiado, and Bo-
(Tassinari and Macambira, 1999). The Parecis basin, queirão. At these locations, the diamonds are usually
encompassing an area of 500,000 km2, contains sedi- small but of good quality, sometimes reaching values
ments about 6,000 m deep. The Pimenta Bueno and of US$100 per carat or more (Watkins, 2009).
Paranatinga regions in the neighboring state of
Rondônia also contain diamond-bearing alluvium. Sublithospheric Diamonds. Juína drew the attention
During the 1980s, the Juína region was the coun- of geoscientists with the discovery of so-called su-

Figure 17. A section of the
diamond-bearing sedi-
ments left by past mining
operations at the aban-
doned Romaria mine in
Minas Gerais. The dia-
mondiferous conglomer-
ates lie beneath the lower
portion of the reddish
brown sediments along
the center of the photo;
they are covered by about
10 meters of brown sedi-
ments. These conglomer-
ates extend across a large
area of the countryside
near the town of Romaria.
Photo by Robert Wel-
don/GIA.
perdeep diamonds in the São Luiz River area in the which operated it from 1969 until 1983, when the
early 1990s (Wilding et al., 1991; Stachel et al., 2005; mine was closed for financial reasons.
Harte, 2010; Kaminsky, 2012). They are believed to The Romaria mine, located near the town of the
have formed at 400 to 800 km, considerably deeper same name, was excavated using tractors to remove
than most diamonds that formed in the continental the overlying sediment layer. Trucks transported the
lithosphere at approximately 140 to 200 km (Thom- ore for washing at a plant that processed up to 600
son et al., 2014). These “sublithospheric” diamonds m3 of ore per day, with the final separation of the di-
are characterized by unusual syngenetic mineral in- amonds performed manually (Svisero et al., 1981).
clusions indicating greater formation depth (such as The mine, located in the northeastern edge of the
majorite, periclase-wustite, wollastonite, diopside, Paraná basin, covers an area of approximately 1.2
moissanite, tetragonal garnet, and perovskite). They km2. On the property, the Uberaba formation con-
are mainly low-nitrogen type IIa diamonds. More re- sists of a sequence of conglomerates, sandstones, and
cently, similar mineral inclusions have been found in mudstones; at the base of this formation are dia-
samples from the Collier 4 kimberlite in southwest mond-bearing conglomerates approximately six me-
Juína (Bulanova et al., 2010). ters thick (Suguio et al., 1979; Coelho, 2010; see
figure 17). The concentrates obtained by washing the
Diamonds in Conglomerates. In several locations, conglomerates contain indicator minerals. Chemical
sedimentary conglomerates are the host rocks for di- analysis revealed that the garnet and ilmenite are of
amonds. Diamonds are seldom recovered directly a kimberlitic origin (Svisero, 1979; Svisero and
from conglomerates, since most mines are developed Meyer, 1981; Coelho, 2010). In 1981, during the final
in unconsolidated sediments. The classic example of phase of the EXDIBRA operation, one of the authors
a diamond-bearing conglomerate is the Romaria (DPS) examined a parcel of 450 carats of diamonds
mine (figure 17), located in western Minas Gerais comprising 5,250 stones of various size, shape, and
about 500 km from Belo Horizonte. color. The crystals were predominantly rhombo-
According to Des Genettes (1859), diamonds were dodecahedral in shape; most were colorless, and 70%
first discovered in the Bagagem River in 1722 by the were of gem quality (Svisero et al., 1981).
explorer João Leite da Silva Hortis. Large-scale explo- During this period, drill core samples were taken
ration in the Romaria region began in the second half to determine the extent of the mineralized conglom-
of the 19th century (Hussak, 1894). In the early 20th erates (Feitosa and Svisero, 1984). The results re-
century, this mine was visited by many foreigners vealed the presence of small “basins” formed by
(Porcheron, 1903; Draper, 1911). It was subsequently depressions in both the sandstone and the underlying
controlled by various companies, the last of which mica schist. These paleostructures on the crystalline
was Extratífera de Diamantes do Brazil (EXDIBRA), basement were filled by mud flows, giving rise to di-

Figure 18. The town of Lençois in the Chapada Diamantina region of Bahia has long been an important source of di-
amonds and carbonado. Garimpeiros found diamonds in conglomerates (left) or in the bottoms of water-worn pot-
holes formed in the river bed. The conglomerate specimen on the right, called the “Madonna” because of its shape,
contains a small diamond. Photos by Robert Weldon/GIA; conglomerate sample courtesy of Gar Mineração.
amond-rich layers within the conglomerate. Samples ments in the riverbed. The average value of the dia-
also revealed that the diamond-bearing layers meas- mond mined by these companies is US$150 per carat.
ured up to 9.6 m thick, while the average diamond This region is characterized by small crystals (0.1–
content ranged from 0.33 to 0.69 ct/m3. 0.3 ct), and historically there have not been discov-
eries of large diamonds there (Reis, 1959). Sampling
Diamonds in Glacial Deposits. Most researchers now of heavy-mineral concentrates from the Tibagi,
accept that diamonds in the Tibagi region of Paraná Laranjinha, Cinzas, Itararé, and Verde river valleys
come from sedimentary glacial deposits (Oppenheim, revealed no traditional kimberlite indicator minerals
1936; Maack, 1968; Liccardo et al., 2012). Tibagi is (Chieregati and Svisero, 1990).
one of the country’s oldest centers of diamond min- In 1985, a diversion of the river channel allowed
ing, having been discovered in 1754 by gold prospec- mining of the sediments in the Tibagi riverbed. This
tors (Leonardos, 1959). afforded one of the authors (DPS) the opportunity to
The mines of the Tibagi River basin and the adja- examine a group of 2,210 diamonds (about 90% of
cent areas have been worked since the latter half of them gem quality). The crystals were generally well
the 18th century (Derby, 1878). Diamonds occur in the formed and euhedral, with a small number of cleav-
riverbeds in both old and recent alluvium and collu- age fragments.
vium. The deposits are distributed in small groups in Existing field evidence of glacial striations on
the southeastern border of the Paraná basin (Per- local rocks, and the absence of kimberlite indicator
doncini et al., 2010), and there have been alternating minerals, indicates that diamonds in this region
periods of intense mining activity and neglect. The were dispersed as the result of glacial transport
Tibagi deposits were especially active in the first (Maack, 1968; Chieregati, 1989; Chieregati and Svis-
decades of the 1900s, when Oppenheim (1936) noted ero, 1990; Liccardo et al., 2012). The diamonds’
60 alluvial occurrences in production. small size, euhedral shape, and overall high quality
In the 1980s, Paraná Minerals (MINEROPAR) and all suggest that these deposits were subjected to
the Companhia de Pesquisa de Recursos Minerais transport processes that favored the retention of the
(CPRM) conducted fieldwork to assess the deposits in better crystals. In addition, studies have shown that
the Tibagi region. They found that the gravels had an the region contains diamonds with mineral inclu-
average grade of 0.62 ct/m3 (Watkins, 2009). Currently, sions characteristic of a peridotitic paragenesis,
Jezzini Minerals and Tibagiana Mineração are using pointing to a kimberlitic initial source (Meyer and
ferries to dredge and concentrate the alluvial sedi- Svisero, 1975). Paleogeographic reconstructions (dos

Santos et al., 1996) indicate that the original centers
where glacial flow took place were located in south-
ern Africa.
Diamonds in Metasedimentary Rocks. Diamonds

from metasedimentary deposits occur mainly around
the town of Diamantina in Minas Gerais and in the
larger region known as Chapada Diamantina in Bahia
Morro do Chapéu
(figure 18). Both areas are located along the Serra do
Palm
P meiras
Espinhaço, a mountain range extending approxi- Le
ençois
An
n
ndaeaí
mately 1,100 km from the center of Minas Gerais Muucugê
north into Bahia (figure 19). Minor deposits are lo- Salvador
cated at Grão Mogol and Serra do Cabral. Bahia
Diamantina. This town is located near the Jequitin-

honha River in Minas Gerais. Diamonds occur in
metaconglomerates and metabreccias of the Sopa- Jequitinhonha
ha River
Minas Gerais
Brumadinho formation (Mesoproterozoic-age Espin- Grão
Mogol
haço Supergroup; see Chaves and Uhlein, 1991) and
in alluvial placers. The primary sources of these dia-
Jequitaí
monds still have not been found. Atlan tic
The diamond-bearing metaconglomerates were Serra
Dia
amantina
a
Ocean
do Cabral
deposited in plains and alluvial fans, while the
metabreccias were deposited by debris flows, possi- Espírito
p
S
Santo
bly in flat tidal environments (Chaves and Uhlein,
1991). In the Diamantina district, the most impor- Belo Espinhaço Supergroup
Horizonte (Mesoproterozoic)
tant diamond fields are located near São João da Cha-
São Francisco craton
pada, Sopa, Guinda, Extração, and Datas (Chaves,
1997; Haralyi et al., 1991).
The Sopa metaconglomerates, which are about 12 Figure 19. A simplified map of the São Francisco cra-
to 15 m thick, have diamond concentrations ranging ton, showing the areas covered by sediments of the
from 0.04 to 0.076 ct/m3. Most of the heavy minerals geological formation called the Espinhaço Super-
in these sediments are derived from the crystalline group. The locations of several diamond districts are
basement rocks. Noteworthy is the absence of the also shown, including Diamantina, Grão Mogol, and
traditional kimberlite indicator minerals that are Serra do Cabral in Minas Gerais, and the Chapada
common in the Coromandel region (Chaves, 1997; Diamantina region (Lençóis, Palmeiras, Andaraí, Mu-
cugê and Morro do Chapéu) in Bahia.
Svisero et al., 2005). Diamonds from this district are
generally small: 90% weigh less than 2 ct, with only
1% above 8 ct. The typical size range is from 0.55 to
0.85 ct (Chaves and Uhlein, 1991).
Large “megadiamond” crystals, such as those re- dodecahedral habit followed by other shapes, and the
covered near Coromandel and the Abaeté River, have majority is of gem quality (Chaves, 1997).
apparently been found in the Diamantina district
(Barbosa, 1991). While there are references to stones Chapada Diamantina. Bahia was Brazil’s largest dia-
as large as 140 ct (Renger, 2005), these discoveries oc- mond producer in the second half of the 19th century.
curred during the 18th century, when record keeping The most important areas are in Chapada Diaman-
was inadequate or nonexistent. In recent decades, tina in the eastern portion of the state, including the
crystals above 10 carats have occasionally been found towns of Lençóis, Palmeiras, Andaraí, Mucugê,
(Chaves, 1997; Renger, 2005). Utinga, and Morro do Chapéu. Around 1821, dia-
Most of the diamond crystals from the area are monds were discovered by German explorers Johann
colorless; some exhibit a green or brown surface col- Baptist von Spix and Carl Friedrich Phillipp von Mar-
oration. The predominant crystal morphology is a tius in the Sincorá Mountains of the Mucugê River

recorded. The diamond content at this location was
1 ct per 28 m3 of washed gravel; the carbonado con-
tent was 1 ct per 14 m3 (see Abreu, 1973).
Carbonado. Around 1841, miners in Chapada Dia-

mantina discovered that some dark grains in the
heavy mineral concentrates had the same hardness
as conventional diamonds. They had discovered a
new type of “diamond” with completely different
physical properties, which they called carbonado
(Rivot, 1848; Leonardos, 1937; Herold, 2013; see fig-
ure 20). Derby and Branner (1905) and Branner (1909)
described the washings in Bahia where carbonado
and diamond are found.
Carbonado is a polycrystalline aggregate—opaque,
solid but porous—consisting of a mixture of dia-
mond, graphite, and amorphous carbon (Haggerty,
Figure 20. The polycrystalline form of diamond
known as carbonado was first found in Bahia in
2014). Compared with monocrystalline diamonds,
about 1841. During the late 19th century, this very which have perfect cleavage and a density of 3.51
hard material was of great economic importance for g/cm3, carbonados are free of cleavage and have
industrial use. Shown here is a privately held collec- slightly lower densities (ranging from 3.0 to 3.4
tion of samples in the town of Lençois. Photo by g/cm3). They can display a glass film, or patina, that
Robert Weldon/GIA. is never observed on diamond crystals (Haggerty,
2014). Besides having the same hardness of conven-
tional diamond, carbonado has great toughness, and
region (Abreu, 1973). The explorers had previously its resistance to impact made it an excellent material
visited the deposits of the Diamantina district; based for industrial use by the late 1800s (Herold, 2013).
on geological similarities, they searched for and dis- Carbonados are generally about 1 cm in size but
covered diamonds in Bahia. Mining began in the Mu- occasionally weigh tens or even hundreds of carats.
cugê River and then spread throughout the Decades ago, several “megacarbonados” emerged
headwaters of the Upper Paraguaçu River. Around from the mines of Chapada Diamantina. The greatest
1842, discovery of the Mucugê River placers caused a was the Sérgio—at 3,167 carats, the largest diamond
rush of about 30,000 miners into the region (Leonar- ever found in nature—discovered at Brejo da Lama
dos, 1937; Abreu, 1973). The secondary sources of near Lençóis in 1905. Another example is the 2,000
these diamonds are the metaconglomerates of the ct Casco de Burro, found in Lençóis a year later
Tombador formation (Chapada Diamantina Group, (Leonardos, 1937).
part of the Mesoproterozoic-age Espinhaço Super- Although Chapada Diamantina is the traditional
group). source of carbonado, this material is also mined in
In general, these diamonds have dodecahedral Minas Gerais, Mato Grosso, Goiás, Paraná, and Ro-
habits, or curved forms intermediate between octa- raima (Reis, 1959; Chaves, 1997). Some famous car-
hedral and dodecahedral. Colorless crystals are the bonados are the Pontesinha (267.53 ct) from West
most common, followed by browns and grays. As in Rosário, Mato Grosso (Reis, 1959), and an 827.5 ct
Minas Gerais, there are no kimberlite indicator min- specimen found in the Abaeté River in 1934 (Leonar-
erals. The primary source of diamonds from the Serra dos, 1937). In each of these localities, carbonados
do Espinhaço remains unknown. occur in alluvial deposits associated with normal
Diamond production in Chapada Diamantina crystalline diamond (Chaves, 1997).
peaked between 1850 and 1860, reaching 70,000 The production and trade of carbonados in Cha-
carats per year, followed by alternating phases of ac- pada Diamantina peaked in the second half of the
tivity and decline. In 1929, mechanized mining 19th century. At the time France maintained a con-
equipment was installed on the Paraguaçu River. sulate in Lençóis to facilitate the purchase and export
After two years of operation, a yield of 2,008 carats of this material to Europe (Abreu, 1973; Herold,
of diamond and 4,109 carats of carbonado was 2013). Production went into decline after 1929, when

N
A t l ant i c
²
O c e an
the price fell by nearly 97% due to increased avail-
ability of industrial diamonds from African countries
(Reis, 1959). Prospecting dwindled before officially Amazonas craton São Lui
Luis cratonn
1
10
ending on September 17, 1985, with the establish-
ment of Chapada Diamantina National Park (An- 12 111 9
drade, 1999). 1617
17
15 8
The origin of carbonados remains controversial. 41
14 18
133
Typical monocrystalline diamond is related to three São
ão
4 Frrancis
sc
co
c
distinct geological phenomena: (1) transport from Kimberlites 3 5 craton
6
the earth’s deep interior by explosive volcanism (vol- Diamond-bearing 2 7
kimberlites
canic diamonds), (2) continental collision followed 1
Indicated limits
by subduction and exhumation (metamorphic dia- Paranap
apanema
ap
ap ma
Inferred limits craton
cra
monds), and (3) meteoritic impact (impact dia- Luis Alves
L
monds). Recent evidence indicates that carbonado craton
c
Rio de la Plata
ata
ta
does not fall into any of these categories. Haggerty craton A t la n tic
O c ea n
(2014) suggested an extraterrestrial origin, in which 0 29
90
9 580 870
km
m
the carbonados formed in a white dwarf star, a car-
bon-rich exoplanet, or a supernovae explosion dur- Diamond-mineralized kimberlites:
(1) Canastra 1, (2) Romaria, (3) T Trrês Ranchos 4, (4) TTucano,
ucano, (5) Abel Régis,
ing a cosmic bombardment between 4.1 and 3.8 (6) Cana Verde,
Verde (7) Maravilhas, (8) Salvador,
Salvador (9)) Braúna 3, (10) Moana 11,
(11) Juína 5, (12) Collier 4, (13) Carolina 1, (14)) Comet 1, (15) Cosmos 1,
billion years ago. (16) Cosmos 2, (17) Cosmos 3, (18) T Tumeleir
umeleiro
o 3.
Kimberlite. The systematic exploration of Brazil’s Figure 21. The distribution of kimberlites in the cra-
primary kimberlite sources only began in the late tonic areas of Brazil. According to Companhia de
1960s. Figure 21 presents a country map showing the Pesquisa de Recursos Minerais (2004), there are 1,212
locations of known kimberlite pipes. Karfunkel et al. known kimberlite bodies distributed throughout the
(1994) cited several reasons for this lack of detailed country, most of them concentrated in the Alto
exploration, including the wide areas over which al- Paranaíba region of Minas Gerais.
luvial diamonds are found, the historical control over
these areas by garimpeiros who often prevented ex-
ploration, the intense tropical weathering and ero-
sion that removed the upper sections of kimberlite mostly intruded into rocks of the Brasília fold belt.
pipes, and (until recently) the absence of published The Vargem 1 kimberlite, in the floodplain of the
information from mining companies searching for Vargem Fazenda near Coromandel, was the first body
kimberlite bodies. found in the region (Meyer et al., 1991). It is 700 m
The Redondão kimberlite was the first such body in diameter and consists of serpentinized kimberlitic
found, during geological mapping in the southern rocks that intrude the Vazante phyllite formation
Parnaíba Basin in the 1960s. This pipe is located (figure 22). The pipe contains abundant pyrope gar-
southeast of the city of Santa Filomena in Piauí state. net, magnesian ilmenite, diopside, and chromite, as
It has a circular shape, with a diameter of 1,000 m well as some zircon. The diamond potential of this
and a negative relief of about 70 m in relation to the body is still being investigated.
surrounding rocks. Other kimberlite bodies in the Coromandel region
Several mining companies have undertaken are Três Ranchos 4 (Fazenda Alagoinha), Limeira, and
prospecting for kimberlites in Brazil, especially Indaiá. These consist of hypabyssal kimberlite rock
Pesquisa e Exploração de Minérios S/A (SOPEMI), a facies with isotopic characteristics intermediate be-
De Beers subsidiary that has operated for more than tween Group 1 and Group 2 kimberlites (Bizzi et al.,
three decades. After finishing its exploration activi- 1993; Melluso et al., 2008; Guarino et al., 2013).
ties, this company transferred its database on 1,212 Três Ranchos 4 was the first kimberlite in the
kimberlites to the state-owned CPRM. About 700 of area known to contain diamonds (Gonzaga and
these kimberlites occur in the Alto Paranaíba region Tompkins, 1991), though not in sufficient quantity
in western Minas Gerais and southeastern Goiás. for commercial exploitation. Another well-known
The kimberlites of the Alto Paranaíba province, kimberlite is Abel Regis, a mineralized intrusion in
located southwest of the São Francisco craton, were the Serra da Mata da Corda (Cookenboo, 2005). This

Figure 22. An exposed portion of the Vargem 1 kimberlite seen on the banks of the Santo Inácio River near
Coromandel. The kimberlite pipe, which has decomposed and altered to serpentine, giving it a brownish green
appearance, is only exposed by removing several meters of overlying sediments. Material removed from the
weathered kimberlite is panned to collect the heavy mineral concentrate. When the pan is turned over, the
concentrate is seen as grains of red pyrope garnet and black ilmenite. The concentrate is then checked for dia-
monds. Photos by Robert Weldon/GIA.
pipe is located 20 km northwest of Coromandel Other mineralized kimberlite pipes include Tu-
(again, see figure 14), and it is intruded into the Neo- cano, located west of Carmo do Paranaíba (Cooken-
proterozoic-age Bambuí Group (CPRM, 2013). It has boo, 2005), and the Romaria kimberlite located
an elliptical shape and an area of approximately 1 northeast of the Romaria mine (F.M. Coelho, pers.
km2. Serpentinization of the kimberlite yielded a comm., 2008). Another important diamond-bearing
green saprolite clay containing fragments of pyrope, kimberlite is the Canastra 1, located in the headwa-
magnesian ilmenite, and diopside. Most of the gar- ters of the São Francisco River near São José do Bar-
nets from this kimberlite, and from several others of reiro in the Serra da Canastra region (Cookenboo,
the Upper Paranaíba region, are lherzolitic in nature 2005). It intrudes quartzites of the Canastra Group
(G9). They are followed by eclogitic garnets (G4 and (Barbosa et al., 1970) that are part of the Brasília fold
G5); G10 garnets are either rare or absent (Cooken- belt of Mesoproterozoic age (Chaves et al., 2008).
boo, 2005; Svisero et al., 2005). These G-numbers Canastra 1, dated at 120 million years, comprises two
represent a statistical classification of garnets from subcircular bodies oriented along the regional north-
kimberlites and associated xenoliths, developed by west-southeast trend of the Canastra Group metased-
Dawson and Stephens (1975), that is used in diamond iments. Trial mining conducted in 2007 indicated the
exploration. According to local miners, diamonds up presence of 16 carats of diamond per 100 tons of rock
to 20 carats have been found in ravines in the area of (Chaves et al., 2008). Eighty percent of the recovered
the intrusion. diamonds were gem quality; the crystals were octa-

hedral in shape and colorless to faint yellow. These Diamond-bearing kimberlites also occur in Bahia
characteristics make the diamonds from this area State (Watkins, 2009) in the northeastern São Fran-
some of Brazil’s most valued. Other mineralized bod- cisco craton. The Salvador 1 kimberlite, located near
ies include the Maravilhas kimberlite located near Barra do Mendes, intrudes metaconglomerates of the
the city of the same name, and the Cana Verde pipe Tombador formation. Rb/Sr dating of the intrusion
located north of the city of Luz, both in the southern indicates an age of 1,152 Ma, making it the oldest
portion of the São Francisco craton (O. Barbosa, pers. known intrusion in the country. Also, the pipe’s in-
comm., 1978). trusion into rocks in the Serra do Espinhaço provides
Several diamond-bearing kimberlites occur in evidence of a period of diamond-bearing kimberlite
Juína, where they intrude into the Rio Negro–Juruena magmatism generated in the region.
mobile belt (1.55 to 1.8 Ga). Among others, the Col- The most important kimberlite pipe at present is
lier 4 kimberlite has an average grade of 0.62 ct/m3. Braúna 3 (see box B), which contains diamonds in
Mining companies have recovered 6,000 diamonds commercial concentrations. This pipe is located
from this pipe, including a 17.08 ct stone (Costa, about 260 km from Salvador, in the northeast portion
2013). The pipe contains sublithospheric diamonds, of the São Francisco craton. It occurs as part of a kim-
which have mineral inclusions and other features berlite field consisting of three pipes and 19 dikes, all
that suggest they formed at great depth (Bulanova et diamond bearing and intruding the Nordestina gran-
al., 2010). Another pipe, Juína 5, also contains sub- odiorite (a batholith of Paleoproterozoic age; see Do-
lithospheric diamonds—mainly contact-twinned and natti-Filho et al., 2013). The dikes range from 0.5 to
octahedral crystals—that appear corroded and are cov- 5.0 m thick, and they are controlled by northwest-
ered with a graphite film (Thomson et al., 2014). trending fractures that can be followed for 15 km. All
Rondônia has 62 kimberlites distributed in the Pi- volcanic pipes and dikes of the Braúna field belong
menta Bueno, Colorado do Oeste, and Ariquemes to the hypabyssal facies. The age of the kimberlites,
fields. According to Neto et al. (2014), it ranks fourth obtained by U/Pb dating of perovskites, is 642 +/- 6
in terms of known occurrences after Minas Gerais Ma. The Braúna kimberlites have mineralogical and
(765), Goiás (197), and Mato Grosso (102). Among the geochemical similarities to those in the Guaniamo
mineralized kimberlites in Rondônia, perhaps the region of Venezuela (Donatti-Filho et al., 2013). The
most important is the Carolina 1 pipe, located near Braúna 3 pipe is now being commercially exploited
Pimenta Bueno. It intrudes into rocks of the Rio by Lipari Mineração Ltda.
Negro–Juruena mobile belt (Tassinari and Macam-
bira, 1999); the age of the intrusion is 230 Ma (Hunt PRODUCTION
et al., 2009). Although diamond-bearing, this kimber- Since the discovery of diamonds in South Africa,
lite is characterized by the absence of G10 garnets in Brazil has never regained its standing, and in recent
the heavy mineral concentrates obtained by washing decades it has accounted for less than 1% of global
the ore. Such garnets are rare in other kimberlites in production. According to Minérios & Minerales
Minas Gerais, Bahia, and Mato Grosso (Cookenboo, (2014), the estimated production between 1725 and
2005; Svisero et al., 2005). Analysis of a small batch 1870 was 50,000 to 250,000 carats per year; if the
of diamonds indicated that the predominant color is larger value is assumed, this would yield a total of
gray (48%), followed by colorless (23%) and others 37.5 million carats over the 150-year period. As most
(Hunt et al., 2009). Other mineralized pipes include Brazilian occurrences are alluvial, it has always been
Cosmos 1, 2, and 3; Comet 1; and Tumeleiro 3 (Neto difficult to determine the amount and commercial
et al., 2014). value of what is recovered.
There are two areas with kimberlite concentra- Production increased with the discovery of dia-
tions in Piauí. The first is located in the southern re- monds in other Brazilian states in the 19th and 20th
gion where, besides the famous Redondão kimberlite, centuries, but under limited government oversight.
several intrusions are distributed in the municipali- Production in the early 1970s was approximately
ties of Santa Filomena, Gilbués, Currais, and Baixa 300,000 carats per year, before jumping to 670,000
Grande do Ribeiro. The second is in the eastern re- carats in 1980. This significant increase was due to
gion, covering the municipalities of Picos, Ipiranga do the Juína mines, and Mato Grosso became the coun-
Piauí, Elesbão Veloso, Jardim do Mulato, and Regen- try’s principal source of diamonds. In recent years,
eração. The Moana 11 kimberlite near Picos contains production has declined significantly, to less than
diamonds, but it has been little studied to date. 32,000 carats in 2015 (http://kimberleyprocess

BOX B: THE FIRST KIMBERLITE MINE IN BRAZIL
Figure B-1. A recent view of the open pit at the Lipari diamond mine near Nordestina in Bahia (left). The
Braúna 3 kimberlite pipe lies beneath serpentinized soil and sedimentary rocks. Also shown is the processing
plant on-site where kimberlite ore is broken down to recover the diamonds (right). Photos by Andrew C. John-
son, courtesy of Lipari Mineração Ltda.
Explosive volcanic eruptions are the transport mecha- covered through the work of garimpeiros (K. Johnson,
nism that brings diamonds to the earth’s surface. Most pers. comm., 2014).
of these kimberlite pipes are not diamondiferous or com- This situation is changing. The consequences of
mercially viable to exploit. But some are, and most of these transformations may have implications for Brazil’s
the annual production of diamonds today comes through reemergence as a global producer of diamonds.
mining of these pipes, particularly in southern Africa
(Linde et al., 2014). Nordestina Kimberlite Diamonds. Following a few years
Economic diamondiferous pipes in Brazil have until of exploration and development by Lipari Mineração
recently remained elusive. With the discovery of dia- Ltda., a new mine has been established there to exploit a
mond-bearing kimberlites in South Africa in 1867, Brazil kimberlite field called the Braúna complex in the São
lost its standing as the world’s primary source. Its total Francisco craton. De Beers initially explored the largest
production, an estimated 45 million carats over three of these kimberlites—Braúna 3, with a surface area of
centuries, is alluvial in nature and has mostly been re- 15,000 square meters—in the early 1990s (figure B-1). The
Figure B-2. Left: A selection of rough Brazilian diamonds, 262 carats total, photographed in Antwerp. The crys-
tal at the center weighs 7.46 ct. The diamonds are from the Braúna kimberlite field near Nordestina. Right:
These larger diamond crystals from the Lipari mine include a 3.87 ct resorbed yellow crystal, a 3.49 ct octahe-
dral crystal, and a 2.94 ct macle. Photos by Robert Weldon/GIA, courtesy of Aftergut N. & Zonen and Lipari
Mineração Ltda.

statistics.org). This number is expected to increase
operating license was sold in 2004 to Valdiam Resources
significantly with the mining of the Braúna 3 kim-
Ltd., which runs the operation today as the Lipari mine.
“We have invested $45 million in exploration, de- berlite in Bahia. Open-pit mining is expected until
velopment, and drilling,” says Ken Johnson, president 2021, with the potential for mining to continue un-
and CEO of Lipari. “Our attention is on Braúna 3, where derground.
we’ve developed a resource containing an estimated 2.5 Brazil has participated in the Kimberley Process
million carats of high-quality diamonds. In 2014, our Certification Scheme (KPCS) since 2003. Underre-
initial bulk sampling activities yielded some 2,500 dia- porting of annual production, along with corruption
monds with an average value of over US$322 per carat. and criminal activity in diamond exporting and sales,
Since then, much has changed, with our accrued explo- has been endemic (see Partnership Africa Canada,
ration and development costs at about US$ 70 million 2005, 2006). Experts believe the stringent certification
by early 2017.”
requirements of the KPCS have caused exports to be
Commercial production at Braúna started in July
underreported, particularly for valuable material such
2016, and the projected production for 2017 is 210,000
carats valued at US$42 million. “We are now recovering as fancy-color diamonds (Gomes dos Santos, 2015).
diamonds from the central zone of the kimberlite pipe, Considering the widespread distribution and
and our current average sale value of $200 per carat is mode of occurrence of Brazilian diamonds, and the
in line with our expectations” (figure B-2). lack of historical controls over mining, underreport-
The new mine is expected to produce up to 340,000 ing is not surprising. In countries where diamonds
carats a year during an open-pit mine lifetime of seven are recovered from kimberlite pipes, control is more
years, and Johnson estimates a nearly US$750 million easily maintained over the small area. But Brazil is
revenue stream. “The Braúna 3 project has excellent po- one of the world’s largest countries, and the second-
tential to transition from an open pit to an underground ary deposits cover vast areas. Alluvial mining and
mine. Our current exploration efforts are focused on de-
initial diamond sales are carried out by garimpeiros,
velopment of additional diamond resources below the
individually or in small groups, with no large com-
bottom of the open pit (at a depth of 260 meters)—iden-
tification of these resources will support a decision to panies exerting control. With Brazil’s long coastline
transition to an underground mine.” In addition to the and remote borders with neighboring countries, dia-
current open pit, 22 other kimberlite pipes have been mond smuggling is an ever-present concern.
discovered so far in the vicinity.
CONCLUSIONS
Future of Large-Scale Kimberlite Mining in Brazil. More Brazil’s significance as a diamond producer spans
than 1,300 kimberlites are known in Brazil, though only
three centuries, making it the world’s longest con-
about 70 are expected to yield diamond on a commer-
cial basis. Kimberlite mining is a new scenario, but one
tinuous commercial source of diamonds (figure 23).
that offers advantages over small-scale artisanal mining. Brazil reached its peak production in the mid-1700s
For one, mining occurs in a known location, and the reas adventurers and explorers made their way into the
serves can be far better estimated than in the scattered country’s vast interior. Most of its diamond supply—
alluvial deposits around Brazil. “In 2014, revenue from then and now—comes from the multitude of alluvial
taxation on small-scale activities was very small (about sources spread across the country.
US$78 per carat on average value for export), with even Alluvial diamonds have been mined in most
lower revenue in 2015” Johnson explained. He believes Brazilian states since the beginning of the 18th cen-
it could be, and should be, much higher. Preliminary tury, when the colony was still under Portuguese
2016 data from the Brazilian Ministry of Mines and En- rule. Besides alluvial occurrences, there are other im-
ergy (MME) indicate an average value of US$727 per
portant deposits related to conglomerates (Romaria,
carat, which is a significant increase in value over pre-
vious years. This would rank Brazil fourth in the world
Minas Gerais), glacial rocks (Tibagi, Paraná), and
for the average value of rough diamond production. mid-Paleozoic metasedimentary rocks (Diamantina
Many of the traditional obstacles to diamond min- District, Minas Gerais and Chapada Diamantina,
ing in Brazil will be overcome, added Johnson. “We will Bahia). Among the alluvial deposits, the Coromandel
put kimberlite production on the map and will go on and Abaeté regions have occasionally yielded “mega-
record as the first company to develop a diamond mine diamonds” reaching hundreds of carats in size, as
in a kimberlite primary source of rock. As such it is an well as crystals with pink, yellow, green, red, and
historic event for Brazil, and for us.” blue “fancy colors.” The mines of the Juína region in
Mato Grosso are known for the quantity of their pro-
duction.

Figure 23. Brazilian di-
amonds were shipped
to Europe from the
1720s through the
1860s. The rose-cut dia-
monds, seen in this
necklace and brooch,
are part of the Danish
crown jewels on dis-
play in the Rosenborg
Castle in Copenhagen.
The necklace was origi-
nally worn by Queen
Louise, consort of King
Frederik VIII (1843–
1912). The brooch
could be taken apart
and the components
worn separately. The
jewelry was fashioned
by the firm of C.M
Weishaupt & Söhne of
Hanau, Germany in
1840 using diamonds
most likely of Brazilian
origin. Photo by Iben
Kaufmann.
Although systematic exploration of Brazilian kim- The past three centuries have shown the impor-
berlites began relatively late, there are several known tance of mining to Brazil’s economic development.
mineralized bodies in the states of Minas Gerais, Since the beginning, garimpeiros have worked the re-
Bahia, Mato Grosso, Rondônia, Goiás, and Piauí. mote regions, founding towns that later became
Large mining groups such as De Beers have worked cities such as Diamantina, Coromandel, and Juína.
in Brazil but never at a reported kimberlite source. In the ongoing search for new mineral deposits, the
However, more than 1,300 kimberlite pipes have been garimpeiros pushed west beyond the boundaries of
identified through exploration in the last century. the Treaty of Tordesilhas, helping expand Brazilian

territory. Beginning in the 1700 and 1800s, Brazilian confusion. A lack of expertise in large-scale diamond
diamonds were used in European jewelry—the larger mining, and industry regulation, have also hampered
and better-quality diamonds found their way into growth in the sector (A. Epstein, pers. comm., 2015).
royal jewelry collections at a time when production According to Constance (2011), recent production has
from India had ceased and the African deposits had stalled for several reasons: government efforts to com-
yet to be discovered. ply with the Kimberley Process Certification Scheme,
Another important development was the discov- its crackdown in 2006 on diamond smuggling, and the
ery of carbonado in Chapada Diamantina in the 2008 global financial crisis. When combined with the
1840s. In the early 1990s, sublithospheric diamonds shortage of skilled labor and the limited availability
were discovered in the kimberlites of Juína in Mato of geological resources that can be applied to mineral
Grosso. These unusual diamonds contain inclusions exploration, an expanded diamond mining program
of minerals formed in the transition zone and lower seems unlikely in the near future.
mantle. Their study has increased our awareness of However, Brazil’s diamond production may yet be
the mineralogy of Earth’s interior. rewritten in the early 21st century. Despite consid-
Brazil’s diamond production remains enigmatic. erable geological fieldwork, the sources of most allu-
The country’s enormous expanse, and the fact that di- vial diamonds in Brazil are still uncertain. As
amonds have been found at numerous alluvial large-scale mining will be conducted in at least one
sources, weathered over the millennia from their pri- location, the traditional scene of garimpeiros panning
mary sources and dispersed throughout the continent for alluvial diamonds likely represents a vanishing
in riverbeds and conglomerates, contributes to the way of life.
ABOUT THE AUTHORS Brazilian kimberlites. Al Gilbertson of GIA in Carlsbad for the his-
Dr. Svisero is a professor in the Institute of Geosciences of the Uni- tory of the Old Mine cut presented in box A, and for information
versity of São Paulo, Brazil. Dr. Shigley is a distinguished research on Anthony Knivet. Francisco de Assis Ribeiro and Robledo Dela-
fellow, and Mr. Weldon is manager of photography and visual com- torre Ribeiro of Gar Mineração in Belo Horizonte for the opportu-
munications, at GIA in Carlsbad, California. nity to visit alluvial diamond mining operations in Minas Gerais,
and for selections of rough diamonds made available for photog-
ACKNOWLEDGMENTS raphy. Dr. Joachim Karfunkel of Federal University of Minas Gerais
The authors wish to thank the following individuals: Ken Johnson, in Belo Horizonte, and Dr. Donald B. Hoover of Springfield, Mis-
CEO and president of Lipari Mineração Ltda.; Lauro de Freitas; and souri, for discussions of Brazilian diamond geology. Luis Alberto
José Paulo Donatti Filho, Lipari exploration manager at Nordestina, de Deus Borges of Petrus Mineração, and Giovanni de Deus
for the opportunity to visit the Lipari kimberlite mine. Ari Epstein of Borges of Gem Eireli Company, Patos de Minas, for information
Antwerp World Diamond Centre for insight into Brazil’s diamond in- on diamond occurrences in Minas Gerais and for selections of
dustry. Dr. Stephen Haggerty of Florida International University in rough diamonds made available for photography. Rui Galopim de
Miami for information on carbonado. Prof. Miguel A.S. Basei for Carvalho in Lisbon and Niels Ruddy Hansen in Brøndby, Den-
suggestions on the geological map in figure 9. Dr. Otavio Barbosa, mark, for assistance in obtaining images of European historical
Fernando M. Coelho, and Rafael Rodrigues for information on jewelry.
REFERENCES
Abreu S.F. (1973) Recursos Minerais do Brasil. Editora Edgard Balfour I. (2011) Famous Diamonds. Antique Collectors’ Club
Blucher, São Paulo, two volumes. Ltd., London.
de Almeida F.F.M. (1967) Origem e evolução da Plataforma Ball S.H. (1929) Diamond mining in Brazil. Jewelers’ Circular, Vol.
Brasileira. Bulletin of the National Department of Mineral Pro- 99, No. 11, pp. 37–39, 45.
duction (DNPM)/Division of Geology and Mineralogy (DGM), Barbosa O. (1991) Diamante no Brasil: Histórico, Ocorrência,
No. 241, Rio de Janeiro. Prospecção e Lavra. Companhia de Pesquisa de Recursos Min-
de Almeida F.F.M., Hasui Y., de Brito-Neves B.B. (1981) Brazilian erais (CPRM), Brasília.
structural provinces: An introduction. Earth-Science Reviews, Barbosa O., Braun O.P.G., Dyer R.C., Cunha C.A.B.R. (1970) Ge-
Vol. 17, No. 1/2, pp. 1–29, http://dx.doi.org/10.1016/0012- ologia da região do Triângulo Mineiro. Bulletin DNPM, No.
8252(81)90003-9 136, Rio de Janeiro.
Andrade C.M. (1999) Aspectos mineralógicos, gemológicos e Berbert C.O., Svisero D.P., Sial A.N., Meyer H.O.A. (1981) Upper
econômicos de diamantes e carbonados da Chapada Diaman- mantle material in the Brazilian Shield. Earth-Science Reviews,
tina, Bahia. MSc. thesis, Institute of Geosciences, University Vol. 17, No. 1/2, pp. 109–133, http://dx.doi.org/10.1016/0012-
of São Paulo, 165 pp. 8252(81)90008-8

Bizzi L.A., Smith C.B., Wit M.J., Armstrong R.A., Meyer H.O.A. Coelho F.M. (2010) Aspectos geológicos e mineralógicos da mina
(1993) Mesozoic kimberlites and related alkali rocks in the de diamantes de Romaria, Minas Gerais. MSc thesis, Institute
south-western São Francisco craton, Brazil. In H.O.A. Meyer of Geosciences, University of São Paulo.
and O.H. Leonardos, Eds., Proceedings of the 5th International Companhia de Pesquisa de Recursos Minerais (2004) Cartas Ge-
Kimberlite Conference, Araxá (MG), CPRM Special Publica- ológicas do Brasil ao Milionésimo, Escala 1:1.000.000, Brasília,
tion, Vol. 2, pp. 156–171. Distrito Federal, Brasil.
Boxer C.R. (1962) The Golden Age of Brazil, 1695–1750: Growing Constance D. (2011) Brazilian diamond mining stalls. Rapaport
Pains of a Colonial Society. University of California Press, Magazine, Vol. 34, No. 11, pp. 56–60.
Oakland. Cookenboo H.O. (2005) Exploration for diamond-bearing kimber-
Branner J.C. (1909) The diamond-bearing highlands of Bahia. En- lite in the Brasilia Belt of Minas Gerais. Proceedings of the
gineering and Mining Journal, Vol. 87, No. 20, pp. 981–986 and Fourth Brazilian Diamond Geology Symposium, Bulletin No.
No. 21, pp. 1029–1033. 14, Diamantina, Brazil, pp. 51–53.
——— (1919) Outlines of the geology of Brazil to accompany the Cordani U.G., de Brito-Neves B.B., D’Agrella-Filho M.S. (2003)
geologic map of Brazil. Bulletin of the Geological Society of From Rodinia to Gondwana: A review of available evidence
America, Vol. 30, pp. 189–338. from South America. Gondwana Research, Vol. 6, No. 2, pp.
de Brito-Neves B.B., Cordani U.G. (1991) Tectonic evolution of 275–283, http://dx.doi.org/10.1016/S1342-937X(05)70976-X
South America during the late Proterozoic. Precambrian Re- Cordani U.G., Teixeira W., D’Agrella-Filho M.S., Trindade R.I.
search, Vol. 53, No. 1/2, pp. 23–40, http://dx.doi.org/10.1016/ (2009) The position of the Amazonian craton in superconti-
0301-9268(91)90004-T nents. Gondwana Research, Vol. 15, No. 3/4, pp. 396–407,
Bulanova G.P., Walter M.J., Smith C.B., Kohn S.C., Armstrong, http://dx.doi.org/10.1016/j.gr.2008.12.005
L.S., Blundy J., Gobbo L. (2010) Mineral inclusions in sublithos- Cornejo C., Bartorelli A. (2010) Minerals and Precious Stones of
pheric diamonds from Collier 4 kimberlite pipe, Juina, Brazil: Brazil. Solaris Cultural Publications, São Paulo.
Subducted protoliths, carbonated melts and primary kimberlite Costa V.S. (2013) Mineralogia e petrologia de xenólitos mantélicos
magmatism. Contributions to Mineralogy and Petrology, Vol. da Província Kimberlítica de Juína, MT. PhD thesis, Institute
160, No. 4, pp. 489–510, http://dx.doi.org/10.1007/s00410-010- of Geosciences, University of São Paulo.
0490-6 Dawson J.B., Stephens W.E. (1975) Statistical classification of gar-
Burton R.F. (1869) Explorations of the Highlands of the Brazil, nets from kimberlites and associated xenoliths. Journal of Ge-
with a Full Account of the Gold and Diamond Mines. Tinsley ology, Vol. 83, No. 5, pp. 589–607, http://dx.doi.org/ 10.1086/
Brothers, London, two volumes. 628143
Calógeras J.P. (1904) As Minas do Brasil e sua Legislação, Vol. 1. DeJean J.E. (2005) The Essence of Style: How the French Invented
National Press, Rio de Janeiro. High Fashion, Fine Food, Chic Cafes, Style, Sophistication and
Cassedanne J.P. (1985) Chercheurs de diamant au Brésil. Monde et Glamour, Simon Shuster, New York.
Mineraux, No. 70, pp. 36–39. Delgado I. de M., Pedreira A.J., Thorman C.H. (1994) Geology and
——— (1989) Diamonds in Brazil. Mineralogical Record, Vol. 20, mineral resources of Brazil: A review. International Geology
No. 5, pp. 325–336. Review, Vol. 36, No. 6, pp. 503–544, http://dx.doi.org/10.1080/
de Castro Sarmento J. (1731) A letter from Jacob de Castro Sar- 00206819409465474
mento, M.D. and F.R.S., to Cromwell Mortimer, M.D. and Sec. Derby O.A. (1878) Geologia da região diamantífera do Paraná,
R.S., concerning diamonds found in Brazil. Philosophical Brazil. Archivo Museu Nacional do Rio de Janeiro, Vol. 3, pp.
Transactions of the Royal Society of London, Vol. 37, pp. 199– 89–98.
201. ——— (1898) Brazilian evidence on the genesis of the diamond.
Cattelle W.R. (1903) Precious Stones: A Book of Reference for Jew- Journal of Geology, Vol. 6, No. 2, pp. 121–146.
ellers, J.B. Lippincott & Co., Philadelphia. ——— (1906) The Serra do Espinhaço, Brazil. Journal of Geology,
Chaves M.L.S.C. (1997) Geologia e mineralogia do diamante da Vol. 14, No. 5, pp. 374–401.
Serra do Espinhaço em Minas Gerais. PhD thesis, Institute of Derby O.A., Branner J.C. (1905) The geology of the diamond and
Geosciences, University of São Paulo. carbonado washings of Bahia, Brazil. Economic Geology, Vol.
Chaves M.L.S.C., Uhlein A. (1991) Depósitos diamantíferos da 1, No. 2, pp. 134–142.
região do alto/médio Rio Jequitinhonha, Minas Gerais. In C. Des Genettes H.R. (1859) Diamantes na Bagagem. Revista do
Schobbenhaus, E.T. Queiroz, and C.E.S. Coelho, Eds., Princi- Archivo Publico Mineiro, Vol. 4, Imprensa Official de Minas
pais Depósitos Minerais do Brasil, Vol. 4, DNPM/CPRM, Gerais, Belo Horizonte, Brazil, pp. 287–292.
Brasília, pp. 117–138. Diamond carriers (1842) Penny Magazine, Vol. 11, (March 12), pp.
Chaves M.L.S.C., Karfunkel J., Hoppe A., Hoover D.B. (2001) Di- 100–101.
amonds from the Espinhaço Range (Minas Gerais, Brazil) and Diderot D. (1771) Recuiel de Plances, sur les Sciences, les Arts
their redistribution through the geologic record. Journal of Liberauz, et les Arts Mechaniques. Briasson, Paris.
South American Earth Sciences, Vol. 14, No. 3, pp. 277–289, Dodge F.H. (1922) Ouro Preto – The old capital of Minas Gerais,
http://dx.doi.org/10.1016/S0895-9811(01)00024-4 Brazil. The Pan-American Magazine and New World Review,
Chaves M.L.S.C., Brandão P.R.G., Girodo A.C., Benitez L. (2008) Vol. 34, No. 4, pp. 29–38.
Kimberlito Canastra-1 (São Roque de Minas, MG): Mineralogia Donatti-Filho J.P., Tappe S., Oliveira E.P., Heaman L.M. (2013) Age
e reservas diamantíferas. Revista Escola de Minas, Vol. 61, No. and origin of the Neoproterozoic Brauna kimberlites: Melt gen-
3, pp. 357–364, http://www.scielo.br/pdf/rem/v61n3/a14 eration within the metasomatized base of the São Francisco
Cheney G.A. (2005) Journey on the Estrada Real – Encounters in Craton, Brazil. Chemical Geology, Vol. 353, No. 1, pp. 19–35,
the Mountains of Brazil. Chicago Review Press, Chicago. http://dx.doi.org/10.1016/j.chemgeo.2012.06.004
Chieregati L.A. (1989) Aspectos mineralógicos, genéticos e Draper D. (1911) The diamond-bearing deposits of Bagagem and
econômicos das ocorrências diamantíferas da região nordeste Agua Suja in the State of Minas Gerais, Brazil. Transactions of
do Paraná e sul de São Paulo. MSc thesis, Institute of Geo- the Geological Society of South Africa, Vol. 14, pp. 8–19.
sciences, University of São Paulo. Draper T. (1949) Diamond mining in Brazil. G&G, Vol. 7, No. 8,
Chieregati L.A., Svisero D.P. (1990) Estudo de minerais pesados da pp. 231–242.
região diamantífera do Rio Tibagi (PR) e as possibilidades de ex- ——— (1950a) The gemstones of Brazil. G&G, Vol. 6, No. 12, pp.
istência de fontes primárias de diamante na região. Acta Ge- 369–375.
ológica Leopoldênsia, Vol. 13, No. 10, pp. 171–186. ——— (1950b) The origin and distribution of diamonds in Brazil.

G&G, Vol. 6, No. 10, pp. 298–306. 2010.074.2.189
——— (1951) The diamond mines of Diamantina – Past and pres- Hasui Y., (2012) Compartimentação geológica do Brasil. In Y.
ent. G&G, Vol. 7, No. 2, pp. 49–57. Hasui, C.D. Carneiro, F.F.M. de Almeida, and A. Bartorelli,
Dufrenoy A. (1855) Note sur un cristal de diamant provenant du Eds., Geologia do Brasil, Editora Beca, São Paulo, pp. 112–122.
district Bogagem, au Brésil. Comptes Rendus Hebdomadaires Herold M.W. (2013) The black diamonds of Bahia (carbonados)
des Séances de l'Académie des Sciences, Vol. 40, No. 1, pp. 3– and the building of Euro-America: A half-century supply mo-
5. nopoly (1880-1930s). Commodities of Empire Working Paper,
Dupont H. (1991) Jazida aluvionar de diamante do rio Jequitin- No. 21, 37 pp., Ferguson Center for African and Asian Studies,
honha em Minas Gerais. In C. Schobbenhaus, E.T. Queiroz, and The Open University, Milton Keynes, United Kingdom,
C.E.S. Coelho, Eds., Principais Depósitos Minerais do Brasil, http://commoditiesofempire.org.uk/publications/working-pa-
Vol. 4. DNPM/CPRM, Brasília, pp. 139–148. pers/working-paper-21/
von Eschwege W.L. (1833) Pluto Brasiliensis. G. Reimer, Berlin. Hofmeester K. (2013a) Working for diamonds from the 16th to the
Fausto B. (1999) A Concise History of Brazil. Cambridge Univer- 20th century. From Working on Labor: Essays in Honor of Jan
sity Press, Cambridge. Lucassen. Brill Publishers, Leiden, pp. 19–46.
Feitosa V.M.N., Svisero D.P. (1984) Conglomerados diamantíferos Hofmeester K. (2013b) Shifting trajectories of diamond processing:
da região de Romaria, Minas Gerais. Proceedings of the 33rd From India to Europe and back, from the fifteenth century to
Brazilian Congress of Geology, Rio de Janeiro, Vol. 10, pp. the twentieth. Journal of Global History, Vol. 8, No. 1, pp. 25–
4995–5005. 49.
Fernandez-Armesto F. (2009) Columbus. Oxford University Press, Hoover D.B., Karfunkel J. (2009) Large Brazilian diamonds. Aus-
Oxford. tralian Gemmologist, Vol. 23, No. 10, pp. 440–446.
Galopim de Carvalho R. (2006) The Bragança ‘diamond’ discov- Hunt L., Stachel T., Morton R., Grütter H., Creaser R.A. (2009)
ered? G&G, Vol. 42, No. 3, pp. 132–133. The Carolina kimberlite, Brazil – Insights into an unconven-
de Galvão Bueno A. (1922) The bandeirantes: Their deeds and de- tional diamond deposit. Lithos, Vol. 112, Supp. 2, pp. 843–851,
scendants. Bulletin of the Pan American Union, Vol. 54, pp. http://dx.doi.org/10.1016/j.lithos.2009.04.018
456–480. Hussak E. (1894) Sobre o depósito diamantífero de Água Suja perto
Gomes dos Santos E. (2015) The Kimberley Process Certification de Bagagem, Minas Gerais. In L. Cruls, Ed., Relatório da
System – KPCS and diamond production changes in selected Comissão Exploradora do Planalto Central do Brasil, Annex
African countries and Brazil. Revista Escola de Minas, Vol. 5, Rio de Janeiro, pp. 281–319.
68, No. 3, pp. 279–285, http://dx.doi.org/10.1590/0370- Inveness J (2014). Reinterpreting the role of non-Europeans in the
44672015680024 ‘English private trade’: The globalisation of diamonds via East
Gonzaga G.M., Tompkins L.A. (1991) Geologia do diamante. In C. India Company networks in the late-seventeenth century. Un-
Schobbenhaus, E.T. Queiroz, and C.E.S. Coelho, Eds., Princi- published manuscript, accessed July 17, 2016,
pais Depósitos Minerais do Brasil, Vol. 4. DNPM/CPRM, https://www.academia.edu/8372856/The_globalisation_of_
Brasília, pp. 53–116. diamonds_via_East_India_Company_networks_in_the_late-
Gonzaga G.M., Teixeira N.A., Gaspar J.C. (1994) The origin of di- seventeenth_century_Reinterpreting_the_role_of_non-Euro-
amonds in western Minas Gerais, Brazil. Mineralium Deposita, peans_in_the_English_private_trade
Vol. 29, No. 5, pp. 414–421, http://dx.doi.org/10.1007/ Janse A.J.A. (2007) Global rough diamond production since 1870.
BF01886959 G&G, Vol. 43, No. 2, pp. 98–119, http://dx.doi.org/10.5741/
Gorceix H. (1882) Brazilian diamonds and their origins. Popular GEMS.43.2.98
Science Monthly, Vol. 21, No. 37, pp. 610–620. Jeffries D. (1750) A Treatise on Diamonds and Pearls, C. and J.
Grazziotin H.F., Andrade J.B.F. (2011) Recursos minerais e indi- Ackers, London.
cadores metalogenéticos da folha NA.20-X-A-III, Vila de Tepe- Kaminsky F. (2012) Mineralogy of the lower mantle: A review of
quém. Thirteenth Brazilian Congress of Geochemistry, ‘super-deep’ mineral inclusions in diamond. Earth-Science Re-
Expanded Abstracts, pp. 1725–1728. views, Vol. 110, No. 1/4, pp. 127–147, http://dx.doi.org/
Guarino V., Wu F-Y., Lustrino M., Melluso L., Brotzu P., Gomes 10.1016/j.earscirev.2011.10.005
C. de B., Ruberti E., Tassinari C.C.G., Svisero D.P. (2013) U-Pb Kaminsky F.V., Zakharchenko O.D., Khachatryan G.K., Shiryaev
ages, Sr-Nd-isotope geochemistry, and paragenesis of kimber- A.A. (2001) Diamonds from the Coromandel area, Minas
lites, kamafugites and phlogopite-picrites of the Alto Paranaíba Gerais, Brazil. Revista Brasileira de Geociências, Vol. 31, pp.
igneous province, Brazil. Chemical Geology, Vol. 353, No. 1, 583–596.
pp. 65–82, http://dx.doi.org/10.1016/j.chemgeo.2012.06.016 Kaminsky F.V., Sablukov S.M., Sablukova L.I., Zakharchenko O.D.
Haggerty S.E. (2014) Carbonado: Physical and chemical properties, (2009) The Fazenda Largo off-craton kimberlites of Piauí State,
a critical evaluation of proposed origins, and a revised genetic Brazil. Journal of South American Earth Sciences, Vol. 28, No.
model. Earth-Science Reviews, Vol. 130, pp. 49–72, 3, pp. 288–303.
http://dx.doi.org/10.1016/j.earscirev.2013.12.008 Karfunkel J., Chaves M.L.S.C., Svisero D.P., Meyer H.O.A. (1994)
Haralyi N.L.E. (1991) Os diamantes de Juína, Mato Grosso. In C. Diamonds from Minas Gerais, Brazil: An update on sources,
Schobbenhaus, E.T. Queiroz, and C.E.S. Coelho, Eds., Princi- origin and production. International Geology Review, Vol.
pais Depósitos Minerais do Brasil, Vol. 4. DNPM/CPRM, 36, No. 11, pp. 1019–1032, http://dx.doi.org/ 10.1080/
Brasília, pp. 155–160. 00206819409465502
Haralyi N.L.E., Hasui Y., Morales N. (1991) O diamante pré-cam- Karfunkel J., Hoover D.B., Fernandes A.F., Sgarbi G.N.C., Kam-
briano da Serra do Espinhaço, Minas Gerais. In C. Schobbenhaus, brock K., Oliveira G.D. (2014) Diamonds from the Coromandel
E.T. Queiroz, and C.E.S. Coelho, Eds., Principais Depósitos Min- Area, West Minas Gerais State, Brazil: An update and new data
erais do Brasil, Vol. 4. DNPM/CPRM, Brasília, pp. 209–222. on surface sources and origin. Brazilian Journal of Geology,
Haralyi N.L.E., Hasui Y., Rodrigues A. S. (1994) O segundo maior Vol. 44, No. 2, pp. 325–338, http://dx.doi.org/10.5327/Z2317-
diamante brasileiro: 602 Quilates. Geociências, Vol. 13, No. 1, 4889201400020011
pp. 213–221. Kerr R. (1824) A General History and Collection of Voyages and
Harte B. (2010) Diamond formation in the deep mantle - The Travels, Arranged in Systematic Order, Vol. 11, William Black-
record of mineral inclusions and their distribution in relation wood, Edinburgh, pp. 65–200.
to mantle dehydration zones. Mineralogical Magazine, Vol. 74, King J.M., Shigley J.E. (2003) An important exhibition of seven rare
No. 2, pp. 189–215, https://dx.doi.org/10.1180/minmag. gem diamonds. G&G, Vol. 39, No. 2, pp. 136–143,

http://dx.doi.org/10.5741/GEMS.39.2.136 sional Paper No. 13, Ottawa, http://www.pacweb.org/
Le Noan C. (2008) Technical Report on the Juina Diamond Proj- images/PUBLICATIONS/Conflict_Diamonds_and_KP/13_Fugi
ect, Report No. 43-101, Diagem Inc., Quebec. tives_and_Phantoms_Revised_2015.pdf
Legrand J. (1980) Diamonds: Myth, Magic and Reality. Crown Pearson H. (1909) The diamond fields of Brazil. Journal of the
Publishers, New York. Royal Society of Arts, Vol. 58, No. 2978, pp. 101–129.
Lenzen G. (1970) History of Diamond Production and the Dia- Perdoncini L.C., Soares P.C., Góis J.R. (2010) Geologia e ocorrên-
mond Trade. Barrie and Jenkins, New York. cias diamantíferas da região de Tibagi. Proceedings of the Fifth
Lessa de Sá V.K., ed. (2015) The Admirable Adventures and Brazilian Diamond Geology Symposium, pp. 92–101.
Strange Fortunes of Master Anthony Knivet – An English Pi- Pointon M. (1999) Jewellery in Eighteenth Century England. In M.
rate in Sixteenth-Century Brazil. Cambridge University Press, Berg and H. Clifford, Eds., Consumer and Luxury: Consumer
New York. Culture in Europe, 1650–1850. Manchester University Press,
Leonardos O.H. (1937) Diamante e carbonado no Estado da Bahia. Manchester.
Mineração e Metalurgia, pp. 183–192. Porcheron H. (1903) Rapport sur les Mines de Diamants de Agua
——— (1956) Recursos minerais do Triângulo Mineiro. Engen- Suja dans l’Etat de Minas Gerais, Brésil. Lecog et Mathorel,
haria, Mineração e Metalurgia, Vol. 14, No. 140, pp. 71–77; No. Paris.
141, pp. 133–142; No. 142, pp. 219–226. Read G., Grütter H., Winter S., Luckman N., Gaunt F., Thomsen
——— (1959) Diamante. Revista Engenharia, Mineração e Met- F. (2004) Stratigraphic relations, kimberlite emplacement and
alurgia, No. 175, pp. 5–8. lithospheric thermal evolution, Quiricó Basin, Minas Gerais
Liccardo A., Barbosa T.A., Hornes K.L. (2012) Diamante de Tibagi State, Brazil. Lithos, Vol. 77, No. 1/4, pp. 803–818,
no Paraná – Patrimônio geológico-mineiro e cultural. Anuário http://dx.doi.org/10.1016/j.lithos.2004.04.011
do Instituto de Geociências, Vol. 35, No. 1, pp. 142–151, Reid M. (2014) Brazil: The Troubled Rise of a Global Power. Yale
http://dx.doi.org/10.11137/2012_1_142_151 University Press, New Haven, Connecticut.
Linde O., De Meo R., Epstein A., Fischler S. (2014) The Global Di- Reis E. (1959) Os grandes diamantes brasileiros. Bulletin No. 191,
amond Report 2014 – Diamond: Timeless Gem in a Changing Rio de Janeiro, DNPM/DGM.
World. Bain & Company and Antwerp World Diamond Centre, Renger F.E. (2005) Regimes de extração e produção de diamantes
Dec. 9, http://www.bain.com/publications/articles/global- do Serro Frio no século XVIII. Proceedings of the Fourth Brazil-
diamond-report-2014.aspx ian Diamond Geology Symposium, Bulletin No. 14, pp. 139–
Maack R. (1968) Diamante no tilito de Tibagi, Paraná. Anais da 142.
Academia Brasileira de Ciências, Vol. 40, p. 96. Rivot L.E. (1848) D’un diamant en masse amorphe et compacte
Machado I.F., Figueirôa S.F. de M. (2001) 500 years of mining in provenant du Brésil. Annales des Mines, Vol. 57, pp. 419–422.
Brazil: A brief review. Resources Policy, Vol. 27, No. 1, pp. 9– de Saint-Hilaire A. (1833) Voyage dans le District des Diamans et
24, http://dx.doi.org/10.1016/S0301-4207(01)00004-6 sur le Littoral du Brésil. Librairie Gide, Paris, two volumes (ex-
Mawe J. (1812) Travels in the Interior of Brazil. Longman, Hurst, cerpt from The Museum of Foreign Literature and Science, Vol.
Rees, Orme, Brown, and Green, London. 26, 1835, pp. 6–15).
McCabe I.B. (2008) Orientalism in Early Modern France, Eurasian dos Santos J.F. (1868) Memorias do Districto Diamantino da Cor-
Trade, Exoticism and the Ancient Règime. Bloomsbury Aca- marca do Serro Frio (Provincia de Minas Geraes). Americana,
demic Press, Oxford. Rio de Janeiro.
Melluso L., Lustrino M., Ruberti E., Brotzu P., Gomes C.B., Mor- dos Santos P.R., Rocha-Campos A.C., Canuto J.R. (1996) Patterns
bidelli L., Morra V., Svisero D.P., D’Amelio F. (2008) Major and of late Paleozoic deglaciation in the Paraná Basin, Brazil.
trace-element composition of olivine, perovskite, clinopyrox- Palaeogeography, Palaeoclimatology, Palaeoecology, Vol.
ene, Cr-Fe-Ti oxides, phlogopite and host kamafugites and kim- 125, No. 1/4, pp. 165–184, http://dx.doi.org/10.1016/S0031-
berlites, Alto Paranaíba, Brazil. Canadian Mineralogist, Vol. 0182(96)00029-6
46, No. 1, pp. 19–40, http://dx.doi.org/ 10.3749/canmin.46.1.19 Sgarbi G.N.C., Chaves A.O. (2005) Kimberlitos e diamantes no
Meyer H.O.A., Svisero D.P. (1975) Mineral inclusions in Brazilian oeste mineiro: uma discussão. Proceedings of the Fourth Brazil-
diamonds. Physics and Chemistry of the Earth, Vol. 9, pp. 785– ian Diamond Geology Symposium, Bulletin No. 14, pp. 153–
795, http://dx.doi.org/10.1016/0079-1946(75)90051-8 157.
Meyer H.O.A., Waring M., Posey E.F. (1991) Diamond deposits of Sinkankas J. (1977) Historical notes on South American gem-
the Santo Inácio river and the Vargem intrusions, near Coro- stones. G&G, Vol. 15, No. 11, pp. 334–344.
mandel, Minas Gerais. In Fifth International Kimberlite Con- Smith C.P., Bosshart G. (2002) Star of the South: A historic 128 ct
ference Field Guidebook, pp. 50–57. diamond. G&G, Vol. 38, No. 1, pp. 54–64, http://dx.doi.org/
Minérios & Minerales (2014), Vol. 38, No. 363, pp. 14-18. 10.5741/GEMS.38.1.54
Moreira P.S. (1955) Os grandes diamantes Brasileiros. Gemologia, Southey R. (1819) History of Brazil, Vol. 3. Longman, Hurst, Rees,
Vol. 1, No. 1, pp. 5–12. Orme and Browne, London.
Neto I.C., Castro C.C., Silveira F.V. (2014) Intrusões kimberlíticas de Souza G.S. (1587) Tratado Descritivo do Brasil em 1587. Ty-
de Rondônia: uma síntese com base no conhecimento atual. pographia de João Ignacio da Silva, Rio de Janeiro,
Proceedings of the Sixth Brazilian Symposium on Diamond https://archive.org/stream/tratadodescriti00varngoog#page/n7/
Geology, pp. 97–102. mode/1up
Oakenfull J.C. (1919) Brazil: Past, Present, and Future. John Bayles Souza N. B. (1991) Depósitos diamantíferos de Poxoréo, Mato
and Sons and Danielsson Ltd., London. Grosso. In C. Schobbenhaus, E.T. Queiroz, and C.E.S. Coelho,
Oppenheim V. (1936) Sedimentos diamantíferos do Paraná. Bul- Eds., Principais Depósitos Minerais do Brasil, Vol. 4.
letin No. 9, Rio de Janeiro, DNPM/DGM. DNPM/CPRM, Brasília, pp. 149–154.
Partnership Africa Canada (2005) The Failure of Good Intentions: Stachel T., Brey G.P., Harris J.W. (2005) Inclusions in sublithos-
Fraud, Theft and Murder in the Brazilian Diamond Industry. The pheric diamonds: Glimpses of deep earth. Elements, Vol. 1, No.
Diamonds and Human Security Project, Occasional Paper No. 2, pp. 73–78, http://dx.doi.org/10.2113/gselements.1.2.73
12, Ottawa, http://www.pacweb.org/images/PUBLICATIONS/ Stonestreet G.D. (1891) Diamond Mining in South Africa. Factory
Conflict_Diamonds_and_KP/12_Failure_good_intentions_electr- and Industrial Management, Vol. 1, No. 5, p. 590.
Eng.pdf Streeter E.W. (1882) Precious Stones and Gems. George Bell &
——— (2006) Fugitives and Phantoms: The Diamond Exporters Sons, London, 3rd ed.
of Brazil. The Diamonds and Human Security Project, Occa- Suguio K., Svisero D.P., Felitti-Filho W. (1979) Conglomerados

polimíticos diamantíferos de idade cretácea de Romaria (MG): alogy and Petrology, Vol. 168, No. 6, http://dx.doi.org/
Um exemplo de sedimentação de leques aluviais. Minutes of 10.1007/s00410-014-1081-8
the 2nd Regional Geology Symposium, Vol. 1, pp. 217–229. Tillander H. (1995) Diamond Cuts in Historic Jewellery,
Svisero D.P. (1979) Piropos cromíferos da mina de diamantes de 1381–1910. Art Books International, Slovenia.
Romaria: composição química e origem. Bulletin Mineralógico, Tompkins L.A., Gonzaga G.M. (1989) Diamonds in Brazil and a
Vol. 6, pp. 7–14. proposed model for the origin and distribution of diamonds in
——— (1995) Distribution and origin of diamonds in Brazil: An the Coromandel Region, Minas Gerais, Brazil. Economic Ge-
overview. Journal of Geodynamics, Vol. 20, No. 4, pp. 493–514, ology, Vol. 84, No. 3, pp. 591–602, http://dx.doi.org/10.2113/
http://dx.doi.org/10.1016/0264-3707(95)00017-4 gsecongeo.84.3.591
Svisero D.P., Meyer H.O.A. (1981) Ilmenitas kimberlíticas da mina von Spix J.B., von Martius C.F.P. (1824) Travels in Brazil in the
de diamantes de Romaria, Minas Gerais. Revista Brasileira de Years 1817–1820. Longman, Hurst, Rees, Orme, Browne, and
Geociências, Vol. 11, No. 4, pp. 217–221. Green, London.
Svisero D.P., Felliti-Filho W., Almeida J.S. (1981) Geologia da mina Watkins J.M. (2009) Relatório Técnico 50, Perfil do Diamante
de diamantes de Romaria, município de Romaria, Minas (Gema e Diamante Industrial), Vol. 1. Ministry of Mines and
Gerais. Mineração e Metalurgia, Vol. 425, pp. 4–14. Energy, Brasília.
Svisero D.P., Ulbrich M.N.C., Vlach S.R. (2005) Composição, Weldon R., Jonathan C. (2013) The Museum of London’s extraor-
origem e significado geológico de minerais resistatos de in- dinary Cheapside Hoard. G&G, Vol. 49, No. 3, pp. 126–137,
trusões kimberlíticas da região de Coromandel, Minas Gerais. http://dx.doi.org/10.5741/GEMS.49.3.126
Proceedings of the Fourth Brazilian Congress of Diamond Ge- Weska R.K. (1996) Geologia da região diamantífera de Poxoréo e
ology, Bulletin No. 14, Diamantina, Brazil, pp. 169–172. áreas adjacentes, Mato Grosso. PhD thesis, Institute of Geo-
Tassinari C.C.G., Macambira M.J.B. (1999) Geochronological sciences, University of São Paulo, 219 pp.
provinces of the Amazonian Craton. Episodes, Vol. 22, No 3, Wilding M.C., Harte B., Harris J.W. (1991) Evidence for a deep ori-
pp. 174–182, http://www.episodes.org/index.php/epi/article/ gin for São Luiz diamonds. Fifth International Kimberlite Con-
view/62773/48972. ference Extended Abstracts, Araxa, pp. 456–458.
Thompson L.S. (1928) The upland diamond deposits of the Dia- Yogev G. (1978) Diamonds and Coral: Anglo-Dutch Jews and Eigh-
mantina District, Minas Gerais, Brazil. Economic Geology, Vol. teenth Century Trade. Holmes & Meier Publishers, New York.
23, No. 7, pp. 705–723. Zolinger I.T., Svisero D.P., Weska R.K. (2002) Morfologia cristalina
Thomson A.R., Kohn S.C., Bulanova G.P., Smith C.B., Araujo D., de diamantes provenientes das regiões de Chapada dos
EIMF, Walter M.J. (2014) Origin of sub-lithospheric diamonds Guimarães, Poxoréo, Paranatinga, Diamantino e Alto Paraguai,
from the Juina-5 kimberlite (Brazil): Constraints from carbon Mato Grosso. Revista do Instituto Geológico, Vol. 23, pp. 23–
isotopes and inclusion compositions. Contributions to Miner- 33, http://dx.doi.org/10.5935/0100-929X.20020007.
For More on Brazilian Diamonds

Watch slideshows and videos from this historical diamond
source, where large-scale production could resume. Visit
https://www.gia.edu/gems-gemology/spring-2017-brazilian-
diamonds, or scan the QR code on the right.

FEATURE ARTICLES
CUPRIAN LIDDICOATITE TOURMALINE

yusuke Katsurada and Ziyin Sun
Cuprian (copper-bearing) tourmaline, known as “Paraíba” tourmaline in the trade, has been an important
gem since its discovery in 1989. Until now, almost all of the material reported has been classified as the
elbaite species of the tourmaline supergroup. Chemical analyses by laser ablation–inductively coupled
plasma–mass spectroscopy (lA-ICP-mS), a common technique for origin determination of Paraíba tour-
malines, revealed that 13 copper-bearing samples submitted to GIA’s Tokyo laboratory contained sub-
stantial amounts of Ca in the X-site. Consequently, they are classified as liddicoatite tourmaline. The
origin of these stones is unknown.
m
inerals of the tourmaline supergroup are cy- species in 1977 (Dunn et al., 1977). Gem-quality lid-
closilicates with the general formula dicoatite tourmaline from Madagascar has been
XY3Z6(T6O18)(BO3)3V3W, where X = (Na+, prized for its remarkable color zoning that is typically
Ca , K , and vacancy); Y = (Fe2+, Mg2+, Mn2+, Cu2+, Al3+,
2+ +
characterized by triangular zones and three-rayed
Li+, Fe3+, and Cr3+); Z = (Al3+, Fe3+, Mg2+, and Cr3+); T = stars surrounded by oscillatory zonings when it is cut
(Si4+, Al3+, and B3+); B = (B3+); V = (OH– and O2–), and W perpendicular to the c-axis (e.g., Dirlam et al., 2002;
= (OH–, F–, and O2–) (e.g., Henry et al., 2011). Tourma- Pezzotta and Laurs, 2011).
line-supergroup minerals can be classified into three Cuprian (copper-bearing) tourmalines with vivid
primary groups based on the X-site occupancy: the al- blue, green-blue, green, and violet colors were first
kali group (Na+ and K+ dominant), the calcic group reported in 1989 from Paraíba State in northeastern
(Ca2+ dominant), and the X-site vacant group (vacancy Brazil (Koivula and Kammerling, 1989). They are
dominant). In each primary tourmaline group, specific now commonly called “Paraíba” or “Paraíba-type”
species are further determined based on the occupancy tourmaline in the trade. Fritsch et al. (1990) and Bank
of other sites. Thirty-four species have been recog- et al. (1990) reported that the blue to green colors
nized by the International Mineralogical Association’s were primarily due to the presence of trace (or minor)
Commission on New Minerals, Nomenclature and amounts of copper. Years after the initial discovery
Classification (IMA-CNMNC; e.g., Henry et al., 2011; in the state of Paraíba, similar gem-quality elbaite
Hawthorne and Dirlam, 2011; B. Dutrow, pers. tourmalines colored by copper and manganese were
comm., 2017). Elbaite is a sodium-, lithium-, and alu- found elsewhere in Brazil (Shigley et al., 2001; Fu-
minum-rich species in the alkali group, with the gen- ruya, 2007), in Nigeria (Smith et al., 2001), and in
eral formula (Na)(Li1.5Al1.5)Al6Si6O18(BO3)3(OH)3(OH). Mozambique (Abduriyim and Kitawaki, 2005; Laurs
Liddicoatite is a calcium- and lithium-rich species et al., 2008). The composition of cuprian tourmaline
in the calcic group, with the general formula from all of these deposits has been determined with
(Ca)(Li2Al)Al6Si6O18(BO3)3(OH)3(OH). Most gem-qual- energy-dispersive X-ray fluorescence spectrometry
ity tourmaline has been reported as elbaite. Liddi- (EDXRF), electron microprobe analysis, and laser ab-
coatite was first distinguished as a separate mineral lation–inductively coupled plasma–mass spec-
troscopy (LA-ICP-MS; see Abduriyim et al., 2006;
Okrusch et al., 2016). No mineral species other than
elbaite has been previously reported for “Paraíba”
GemS & GemoloGy, Vol. 53, No. 1, pp. 34–41,
tourmalines except for four pieces of Cu-bearing lid-
http://dx.doi.org/10.5741/GemS.53.1.34 dicoatite first reported by Karampelas and Klemm
© 2017 Gemological Institute of America (2010) showing Ca-rich X-site occupancy and three
34 CUPRIAN LIDDICOATITE TOURMALINE GEMS & GEMOLOGY SPRING 2017

Figure 1. Six of the 13
cuprian liddicoatite
tourmalines from this
study. The faceted
stones weigh 1.59–9.63
ct (average 3.95 ct).
They are transparent
and have greenish blue,
green-blue, and green
colors. Photos by Ma-
sumi Saito.
pieces of possible liddicoatite conjectured from the ditional cuprian tourmaline samples submitted by
qualitative analysis indicating high Ca (Leelawatana- clients were analyzed with LA-ICP-MS for compari-
suk and Jakkawanvibul, 2011). This article reports son. Their geographical origin was identified using
on the same type of cuprian tourmalines that belong GIA’s tourmaline database. Eighteen of these addi-
to the liddicoatite species based on the evidence sup- tional samples were from Brazil, one from Nigeria, and
ported by quantitative analysis. seven from Mozambique.
The data was converted to wt.% oxides and nor-
MATERIALS AND METHODS malized to 100 wt.% and then converted back to
Thirteen tourmaline samples of Paraíba-type colors ppmw to obtain individual element concentrations,
(figure 1) from unknown geographic origins were ex-
amined in GIA’s Tokyo laboratory. These were sub-
mitted by different clients in 2016. Standard
gemological testing was followed by analyses of the In Brief
major, minor, and trace element concentration for • Almost all copper-bearing tourmaline has been classi-
each sample with LA-ICP-MS. Fluorine was not meas- fied as the elbaite species. Cuprian liddicoatite exists,
ured. A Thermo Fisher Scientific’s iCAP Qc ICP-MS however, and may have entered the “Paraíba” tourma-
line market.
was connected to an Electro Scientific Industries
NWR213 laser ablation unit with a frequency-quintu- • only sophisticated quantitative chemical analysis can
effectively separate liddicoatite from elbaite.
pled Nd:YAG laser (213 nm wavelength) running at 4
ns pulse width. NIST SRM 610 and 612 glass stan- • Cuprian liddicoatite shows high Ga and high Pb.
dards were used for external calibration. Ablation was Under long-wave UV, it displays stronger fluorescence
than cuprian elbaite due to high concentrations of rare
achieved using a 40 μm diameter laser spot size, a flu-
earth elements.
ence (energy density) of approximately 10 J/cm2, and
a 7 Hz repetition rate. Three laser spots were acquired
from the girdle of each sample. The composition was
initially internally standardized with 29Si using a cal- based on 27 O2– and 4 OH– anions per formula. LA-
culated amount of Si based on the weight percent of ICP-MS analysis for tourmaline is an incomplete char-
pure elbaite in the chemical formula. Twenty-six ad- acterization, with critical light elements (H and F) and
CUPRIAN LIDDICOATITE TOURMALINE GEMS & GEMOLOGY SPRING 2017 35

TABLE 1. Chemical composition of 13 cuprian tourmaline samples, obtained by lA-ICP-mS.
Sample CT1 CT2 CT3 CT4 CT5 CT6 CT7 CT8 CT9 CT10 CT11 CT12 CT13
no
ppmw
li 11,600 11,700 12,300 13,600 12,300 12,800 12,500 13,000 13,900 11,700 12,100 12,600 15,300
B 33,900 34,400 33,200 35,200 32,500 34,700 35,200 31,500 33,600 34,400 37,600 34,100 40,300
Na 8920 8480 8050 8390 8960 7600 9130 9860 8170 7910 7120 8220 7810
Al 220000 218,000 230,000 220,000 224,000 217,000 218,000 228,000 225,000 217,000 216,000 222,000 197,000
Si 169,000 169,000 158,000 163,000 166,000 169,000 161,000 164,000 161,000 168,000 164,000 166,000 171,000
Ca 21,000 21,300 24,100 24,400 21,800 21,700 22,800 20,400 23,500 22,200 20,800 22,800 20,400
mn 2+
2160 3110 5780 3610 3480 3340 12,300 2130 3480 6530 8000 2860 10,100
Cu 1490 1610 2120 1510 2140 1600 2790 1550 1460 2300 2510 1330 2970
atoms per formula unit, 27 O + 4 OH anions normalization

X-site
X site 0.14 0.15 0.09 0.08 0.11 0.17 0.08 0.11 0.10 0.15 0.22 0.12 0.20
vacancy
Ca 0.50 0.50 0.57 0.58 0.52 0.51 0.54 0.48 0.56 0.52 0.49 0.54 0.48
Na 0.37 0.35 0.33 0.35 0.37 0.31 0.38 0.41 0.34 0.33 0.29 0.34 0.32
X-site 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
total
y-site
Al 1.42 1.36 1.46 1.22 1.49 1.30 1.15 1.59 1.37 1.31 1.08 1.37 0.63
Cu 0.02 0.02 0.03 0.02 0.03 0.02 0.04 0.02 0.02 0.03 0.04 0.02 0.04
li 1.58 1.60 1.70 1.86 1.68 1.74 1.72 1.79 1.90 1.60 1.65 1.73 2.08
2+
mn 0.04 0.05 0.10 0.06 0.06 0.06 0.21 0.04 0.06 0.11 0.14 0.05 0.17
y-site 3.06 3.04 3.29 3.16 3.26 3.12 3.13 3.44 3.36 3.06 2.91 3.17 2.92
total
Z-site
Al 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00
Z-site 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00
total
T-site
Si 5.70 5.71 5.35 5.50 5.62 5.68 5.46 5.55 5.46 5.67 5.53 5.58 5.74
Al 0.30 0.29 0.65 0.50 0.38 0.32 0.54 0.45 0.54 0.33 0.47 0.42 0.26
T-site 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00
total
B-site
B 2.97 3.01 2.92 3.08 2.85 3.04 3.10 2.77 2.95 3.02 3.29 2.99 3.50
B-site 2.97 3.01 2.92 3.08 2.85 3.04 3.10 2.77 2.95 3.02 3.29 2.99 3.50
total
Sum
Cation 18.89 18.90 19.12 19.17 19.01 18.99 19.15 19.10 19.20 18.93 18.98 19.03 19.22
sum
Anion 31.00 31.00 31.00 31.00 31.00 31.00 31.00 31.00 31.00 31.00 31.00 31.00 31.00
sum
Detection limits (ppmw) are: Li (0.95–1.31), B (8.78–10.42), Na (18.91–19.34), Al (3.41–3.48), Ca (339–364), Mn (0.55), and Cu (0.42–0.45). Si is
used as internal standard.

A: Primary tourmaline groups - X-site B: Liddicoatite-feruvite-uvite subsystem C: Dravite-schorl-elbaite subsystem
Ca2+ 1.5Li+ 2Li+
0.00 1.00 0.00 1.00 0.00 1.00
Calcic Elbaite
group Liddicoatite
0.50 0.50 0.50 0.50 0.50 0.50
X-vacant Alkali Feruvite Uvite Schorl Dravite

group group
1.00 0.00 1.00 0.00 1.00 0.00

0.00 0.50 1.00 0.00 0.50 1.00 0.00 0.50 1.00
X-site vacancy Na (+K+)
+
Y-site Fe2+ Y-site Mg2+ Y-site Fe2+ Y-site Mg2+
Cuprian liddicoatite tourmaline Cuprian liddicoatite tourmaline Cuprian elbaite tourmaline from Brazil
Cuprian elbaite tourmaline from Brazil Cuprian elbaite tourmaline from Mozambique
Cuprian elbaite tourmaline from Mozambique Cuprian elbaite tourmaline from Nigeria
Cuprian elbaite tourmaline from Nigeria
Figure 2. A: The 13 cuprian tourmaline samples belong to the calcic group, while the 26 additional cuprian tour-
malines belong to the alkali group based on the dominant occupancy of X-site. B: Li, Y-site Fe2+, and Y-site Mg2+
distinguish the tourmaline species as liddicoatite in a liddicoatite-feruvite-uvite subsystem ternary diagram. C: Li,
Y-site Fe2+, and Y-site Mg2+ of elbaite samples were further plotted in a dravite-schorl-elbaite subsystem ternary di-
agram (modified after Henry et al., 2011).
the oxidation states of transition elements (Fe and Mn) fluid inclusions typical of tourmaline. Standard
undetermined. In this article, it is assumed that Fe and gemological testing resulted in the general range of
Mn are divalent, that the B-site is fully occupied by B, gem tourmalines such as refractive indices of 1.62–
and that the X-site is occupied by Na, Ca, and a va- 1.64 and specific gravity of approximately 3.06, but
cancy equal to 1 atom per formula unit (apfu). Addi- the fluorescence under long-wave ultraviolet light
tional assumptions are that the Z-site is only occupied was stronger than the usual Paraíba tourmalines, as
by Al3+ and equal to 6 apfu, while the T-site is occupied described later. Chemical analysis demonstrated a
by Si4+ and Al3+ and equal to 6 apfu. If Si4+ is greater calcium-dominant composition. The representative
than 6, the T-site is only occupied by Si4+. The excess data in table 1 was selected to show the best stoi-
Al goes into the Y-site. The priority of ions with dif- chiometry. All the data points showed liddicoatite,
ferent valence states entering the Y-site is (R2+ > R3+ > and none were classified as elbaite. In all analyses,
R+ > R4+), such that the Y-site is occupied by Mn, Cu, Mg was below the detection limit and Fe was less
Al, and Li. The common assumption that all iron is than 0.005 apfu. Based on the primary tourmaline
ferrous and that OH + F = 4 apfu can result in the group classification, all 13 calcium-rich cuprian sam-
misidentification of buergerite as “fluor-schorl” as ples are classified as calcic-group tourmaline and the
well as misidentification of the oxy- and fluor- species. 26 additional sodium-rich cuprian samples as alkali-
The chemical composition presented in this article is group tourmaline (table 2, figure 2A). In addition, the
not affected by the minor amounts of Fe in the tour- 13 calcic-group samples plot as liddicoatite tourma-
maline (Clark, 2007). The assumption of 4 OH– does line in a liddicoatite-feruvite-uvite subsystem ter-
not allow for the fluor- or oxy- species to be deter- nary diagram (figure 2B). The 26 tourmalines in the
mined. The major elements, including Ca, Na, Si, Al, alkali group are shown as elbaite in a dravite-schorl-
and Mg, were verified by comparison of LA-ICP-MS elbaite subsystem ternary diagram (figure 2C; Henry
data with electron microprobe data of a secondary et al., 2011).
tourmaline standard with similar Ca:Na ratios Comparing the data in table 1 with that of liddi-
(Dutrow and Henry, pers. comm., 2017). coatitic tourmaline from Madagascar (Dirlam et al.,
2002) shows that these cuprian samples have more
RESULTS AND DISCUSSIONS sodium in the X-site—with a greater elbaitic compo-
Microscopic examination of the 13 samples revealed nent. Other liddicoatite tourmalines from Canada
two-phase inclusions, needle-like growth tubes, and (Teertstra et al., 1999) have similar sodium (0.365–

TABLE 2. X-site occupancy of 26 cuprian elbaite samples in molecular proportions, obtained by lA-ICP-mS.
Na K Ca Vacancy
Brazil 0.705 (0.513–0.861) 0.003 (0.001–0.005) 0.063 (0.000–0.112) 0.229 (0.073–0.381)
mozambique 0.596 (0.519–0.661) 0.002 (0.002–0.003) 0.057 (0.025–0.131) 0.346 (0.205–0.444)
Nigeria 0.724 (0.699–0.748) 0.003 (0.003–0.003) 0.111 (0.103–0.116) 0.162 (0.145–0.185)
0.395 apfu) and similar or lower calcium (0.420–0.498 centration than heavier REE (Tb, Dy, Ho, Er, Tm, Yb,
apfu). The sodium contents of liddicoatite from and Lu). The samples show an exceptionally low Eu
Madagascar and Canada vary with their zonation. content. The high concentration of REEs is inter-
Table 3 shows the samples’ averaged chemical preted as the cause of their comparatively strong flu-
composition for selected minor and trace elements. orescence under long-wave UV (figure 4).
In cuprian tourmalines from different origins, these Although the geographic origin of these cuprian
have some distinguishable trends (Abduriyim et al., liddicoatite is unknown, this unique chemical prop-
2006; Okrusch et al., 2016). For example, Brazilian erty offers directions for further research. Geological
and Nigerian cuprian tourmalines tend to show studies of the elbaitic cuprian tourmaline deposits in
higher concentrations of Cu than those from Brazil have been published (e.g., Shigley et al., 2001;
Mozambique—which typically have higher Ga than Soares et al., 2008; Beurlen et al., 2011), but no de-
the other two sources. Nigerian stones tend to have tailed study has been carried out for Nigerian and
higher Pb than those from Brazil and Mozambique. Mozambican occurrences. Teertstra et al. (1999) re-
Figure 3 shows the Ga-Pb distribution of the 13 ported that some liddicoatite crystals have cores that
cuprian liddicoatite and 26 cuprian elbaite samples correspond to elbaite with rims of liddicoatite. They
analyzed in this study. The cuprian liddicoatite sam- proposed three possibilities for the calcium needed
ples show both high Ga (297–433 ppmw) and high Pb to form liddicoatite: mobilization of Ca from earlier-
(420–827 ppmw). These combinations of trace ele- formed pegmatite minerals, introduction of Ca from
ments plot well outside the ranges of any known ref- host rocks, and conservation of Ca through the crys-
erence samples in GIA’s database, and therefore their tallization of minerals in the pegmatite magma.
geographic origin could not be determined. Major elements can also be used to determine prove-
Another remarkable point was the samples’ high nance. Henry and Guidotti (1985) established distinct
concentration of rare earth elements (REE; table 2)— regions to define potential different source rock types
in other words, lanthanides except Pm. Lighter REE of tourmaline using Al-Fe(tot)-Mg and Ca-Fe(tot)-Mg
(La, Ce, Pr, Nd, Sm, and Gd) showed a higher con- ternary diagrams. Figure 5 indicates that the rock
Pb-Ga CONCENTR ATION Figure 3. Pb vs. Ga con-

centration plot of the 13
1300
Cuprian liddicoatite tourmaline
urmaline cuprian liddicoatites of
1200 Cuprian elbaite tourmaline
aline from Brazil unknown origin and 26
1100 Cuprian elbaite tourmaline
aline from Mozambique
Cuprian elbaite tourma aline from Nigeria cuprian elbaite tourma-
1000
lines from Brazil,
900
Mozambique, and Nige-
Pb (ppmw)
800 ria. The 13 cuprian lid-

700 dicoatites have high Pb
600 and high Ga. Cuprian
elbaite tourmalines
P
500
400 from Brazil have low Pb

300 and low Ga. Cuprian
200 elbaites from Mozam-
100
bique have high Ga and
0
low Pb, while those
0 200 400 600 from Nigeria have low
Ga (ppmw) Ga and high Pb.

TABLE 3. minor and trace element concentration (ppmw) of 13 cuprian liddicoatite samples and
26 cuprian elbaites from Brazil, mozambique, and Nigeria.
Sample CT1 CT2 CT3 CT4 CT5 CT6 CT7 CT8 CT9 CT10 CT11 CT12 CT13 Brazil mozambique Nigeria
(18 samples) (7 samples) (1 sample)
mg bdl 1.90 bdl bdl bdl bdl bdl bdl bdl bdl bdl bdl 0.25 112 bdl 0.73
(0–1170)
mn2+ 2220 3140 5740 3570 3510 3440 12300 2130 3430 6490 7900 2790 9820 11800 4360 18700
(420–26300) (99.5–12100)
Fe2+ bdl 261 72.4 66.0 34.9 20.0 208 bdl 36.0 155 241 42.9 201 164 bdl 209
(0–1060)
Cu 1530 1580 2100 1490 2140 1650 2850 1560 1530 2250 2550 1220 2860 9580 2310 2470
(420–23200) (1140–3430)
Ga 397 373 333 368 382 375 297 433 359 324 319 376 349 139 321 143
(61.2–232) (255–401)
la 4.99 17.0 76.9 42.6 27.9 42.2 108 4.28 42.8 87.3 99.9 27.5 93.2 0.69 0.01 0.04
(bdl–3.72) (bdl–0.03)
Ce 20.3 54.2 230 131 88.1 131 316 20.0 137 277 304 92.0 319 0.97 0.02 0.04
(bdl–5.79) (bdl–0.06)
Pr 4.89 9.62 31.2 18.9 15.5 19.6 40.5 5.06 19.8 32.6 39.9 11.8 39.5 0.07 0.00 0
(bdl–0.42) (bdl–0.00)
Nd 36.5 47.1 111 76 72.2 71.9 126 37.3 72.4 115 123 65.0 120 0.12 bdl bdl
(bdl–0.82)
Sm 40.0 29.5 27.0 26.7 58.3 26.6 26.0 45.7 27.8 25.6 28.8 23.0 28.0 0.02 bdl bdl
(bdl–0.13)
eu 1.26 0.79 0.63 0.64 1.57 0.75 0.64 1.64 0.73 0.77 0.71 0.69 0.71 0.00 bdl bdl
(bdl–0.01)
Gd 12.2 8.25 7.65 6.84 15.1 6.43 7.59 13.3 7.48 7.15 8.22 6.44 6.71 0.01 0.00 bdl
(bdl–0.06) (bdl–0.02)
Tb 0.89 0.51 0.43 0.40 1.18 0.41 0.48 0.86 0.39 0.45 0.51 0.34 0.46 0.00 bdl 0
(bdl–0.01)
Dy 1.48 1.00 0.82 0.66 1.67 0.79 0.79 1.44 0.71 0.77 0.92 0.84 0.75 0.00 bdl bdl
(bdl–0.01)
Ho 0.08 0.06 0.04 0.04 0.07 0.05 0.05 0.10 0.05 0.05 0.06 0.03 0.06 0.00 0.00 bdl
(bdl–0.00) (bdl–0.00)
er 0.09 0.09 0.10 0.06 0.10 0.13 0.09 0.07 0.10 0.09 0.09 0.07 0.08 0.00 0.00 bdl
(bdl–0.00) (bdl–0.01)
Tm bdl 0.00 0.01 0.01 0.00 0.01 0.00 0.01 bdl 0.00 bdl 0.00 0.01 0.00 0.00 bdl
(bdl–0.01) (bdl–0.00)
yb 0.08 0.05 0.02 bdl 0.05 0.04 bdl 0.07 0.03 0.01 bdl 0.05 bdl 0.00 0.00 bdl
(bdl–0.01) (bdl–0.01)
lu 0.01 bdl 0.01 0.00 bdl bdl 0.01 0.01 bdl bdl bdl 0.00 bdl 0.02 bdl bdl
(bdl–0.40)
Ta 7.31 6.52 5.30 7.18 5.06 6.85 4.18 8.05 7.92 6.53 6.70 7.70 5.83 2.23 1.58 6.66
(0.04–8.35) (0.54–3.62)
Pb 703 672 550 773 659 774 420 761 827 642 578 567 684 44.4 43.9 467
(2.74–188) (3.85–123)
∑Ree 123 168 486 304 282 300 626 130 309 547 606 228 609 1.89 0.03 0.1
Abbreviation: bdl = below detection limit. Detection limits (ppmw) are: Mg (0.13–0.56), Mn (0.16–1.01), Fe (6.70–34.02), Cu (0.73–3.00), Ga (0.05–0.19),
La (0.00–1.26), Ce (0.00–0.01), Pr (0.00–0.02), Nd (0.00–0.23), Sm (0.00–0.04), Eu (0.00–0.01), Gd (0.00–0.02), Tb (0.00–0.01), Dy (0.00–0.02), Ho
(0.00), Er (0.00–0.01). Tm (0.00–0.01), Yb (0.01–0.02), Lu (0.00–0.01), Ta (0.00), and Pb (0.05–0.14).
types in which the cuprian liddicoatite was found are dominant, they must be from a Ca-rich host rock.
Li-rich granitoid pegmatites and aplites. Because Karampelas and Klemm (2010) noted that liddicoat-
these copper-bearing liddicoatite tourmalines are Ca- ite rough had been found near the Paraíba-type el-

Figure 4. Randomly selected samples under daylight-equivalent light (left) and under long-wave UV light (right).
Copper-bearing liddicoatite (CT05, CT03, and CT06 at the top) show stronger fluorescence than the cuprian
elbaite (MZ04, MZ06, NG01, BR15, BR17, and BR08 at the middle and bottom) due to high REE concentrations.
Photos by Yusuke Katsurada.
baite mine in Mozambique. Qualitative EDXRF reportedly near Nampula, Mozambique, showed a
analysis of Paraíba-type tourmaline from a new mine Ca peak (Leelawatanasuk and Jakkawanvibul, 2011).
Figure 5. A: This Al-Fe(tot)-Mg diagram (in molecular proportions) reveals that the likely source rock of cuprian lid-
dicoatite tourmaline is Li-rich granitoid pegmatite and aplite. Fe(tot) represents the total Fe. This diagram is di-
vided into regions that define the compositional range of tourmaline from different rock types (modified after Henry
and Guidotti, 1985). B: The Ca-Fe(tot)-Mg diagram (also in molecular proportions) shows that the likely source rock
of cuprian liddicoatite tourmaline is Li-rich granitoid pegmatite and aplite. The rock types defined by the fields in
this diagram (also modified after Henry and Guidotti, 1985) are somewhat different from those in figure 5A.
A: Al-Fe(tot)-Mg diagram for tourmaline from B: Ca-Fe(tot)-Mg diagram for tourmaline from
various source rock types various source rock types
Al Ca
Elbaite Liddicoatite
1 Alkali-free dravite
7 1
2 4
Schorl 5 Dravite
Buergerite
8 Uvite
3 6 3 5
Uvite
Schorl
2 4 6
Buergerite Dravite
Al50Fe(tot)50 Al50Mg50 Fe(tot) Mg
Cuprian liddicoatite tourmaline Cuprian liddicoatite tourmaline
1. Li-rich granitoid pegmatites and aplites 1. Li-rich granitoid pegmatites and aplites
2. Li-poor granitoids and their associated pegmatites and aplites 2. Li-poor granitoids and their associated pegmatites and aplites
3. Fe3+-rich quartz-tourmaline rocks (hydrothermally altered granites) 3. Ca-rich metapelites, metapsammites, and calc-silicate rocks
4. Metapelites and metapsammites coexisting with an Al-saturating phase 4. Ca-poor metapelites, metapsammites, and quartz-tourmaline rocks
5. Metapelites and metapsammites not coexisting with an Al-saturating phase 5. Metacarbonates
6. Fe3+-rich quartz-tourmaline rocks, calc-silicate rocks, and metapelites 6. Metaultramafics
7. Low-Ca-metaultramafics and Cr, V-rich metasediments
8. Metacarbonates and meta-pyroxenites

Analysis of the host rock geology around cuprian by high REE concentrations, the samples’ gemolog-
tourmaline mines in Mozambique may be an impor- ical properties were also similar to cuprian elbaite
tant approach. tourmalines. Only sophisticated quantitative chem-
ical analyses can effectively separate liddicoatite
CONCLUSIONS from elbaite.
Thirteen cuprian tourmaline samples were identi- Cuprian liddicoatite tourmaline is not well
fied as liddicoatite tourmaline. Their Cu and Mn known, but the material may have already entered
concentrations were within the range of other the “Paraíba” tourmaline market. Discovering its
cuprian tourmalines—consistent with their similar origin could provide new insights into the geologic
blue to green colors. Except for their stronger fluo- growth conditions and chemical variations of tour-
rescence under long-wave UV, presumably caused maline crystals.
ABOUT THE AUTHORS ACKNOWLEDGMENTS

Dr. Katsurada is a scientist and a staff gemologist at GIA in We wish to thank Dr. Makoto Miura and Takuya Sunaoshi of GIA’s
Tokyo. Mr. Sun is a staff gemologist at GIA in Carlsbad, laboratory in Tokyo for assisting with LA-ICP-MS testing. We also
California. appreciate the helpful comments and suggestions from Shane
McClure of GIA Carlsbad and Dr. Barbara Dutrow of Louisiana
State University.
REFERENCES
Abduriyim A., Kitawaki H. (2005) Gem News International: Cu- 1–15.
and Mn-bearing tourmaline—More production from Mozam- Henry D.J., Novák M., Hawthorne F.C., Ertl A., Dutrow B.L., Uher
bique. G&G, Vol. 41, No. 4, pp. 360–361. P., Pezzotta F. (2011) Nomenclature of the tourmaline-super-
Abduriyim A., Kitawaki H., Furuya M., Schwarz D. (2006) group minerals. American Mineralogist, Vol. 96, No. 5–6, pp.
“Paraíba”-type copper-bearing tourmaline from Brazil, Nigeria, 895–913, http://dx.doi.org/10.2138/am.2011.3636
and Mozambique: Chemical fingerprinting by LA-ICP-MS. Karampelas S., Klemm L. (2010) Gem News International: “Neon”
G&G, Vol. 42, No. 1, pp. 4–21, http://dx.doi.org/10.5741/ blue-to-green Cu- and Mn-bearing liddicoatite tourmaline.
GEMS.42.1.4 G&G, Vol. 46, No. 4, pp. 323–325.
Bank H., Henn U., Bank F.H., von Platen H., Hofmeister W. (1990) Koivula J.I., Kammerling R.C., Eds. (1989) Gem News: Paraíba
Leuchtendblaue Cu-führende Turmaline aus Paraiba, Brasilien. tourmaline update. G&G, Vol. 25, No. 4, pp. 248–249.
Zeitschrift der Deutschen Gemmologischen Gesellschaft, Vol. Laurs B.M., Zwaan J.C., Breeding C.M., Simmons W.B., Beaton D.,
39, No. 1, pp. 3–11 (in German). Rijsdijk K.F., Befi R., Falster A.U. (2008) Copper-bearing
Beurlen H., de Moura O. J.M., Soares D. R., Da Silva M. R.R., Rhede (Paraíba-type) tourmaline from Mozambique. G&G, Vol. 44,
D. (2011) Geochemical and geological controls on the genesis No. 1, pp. 4–30, http://dx.doi.org/10.5741/GEMS.44.1.4
of gem-quality “Paraíba tourmaline” in granitic pegmatites Leelawatanasuk T., Jakkawanvibul J. (2011) New Paraiba-type
from northeastern Brazil. The Canadian Mineralogist, Vol. 49, tourmaline from Mozambique. http://www.git.or.th/2014/
No. 1, pp. 277–300, http://dx.doi.org/10.3749/canmin.49.1.277 eng/testing_center_en/lab_notes_en/glab_en/2011/GIT_Paraiba-
Clark C.M. (2007) Tourmaline: Structure formula calculations. new_Mine.pdf.
The Canadian Mineralogist, Vol. 45, No. 2, pp. 229–237, Okrusch M., Ertl, A., Schüssler U., Tillmanns E., Brätz H., Bank
http://dx.doi.org/10.2113/gscanmin.45.2.229 H. (2016) Major- and trace-element composition of Paraíba-type
Dirlam D.M., Laurs B.M., Pezzotta F., Simmons W.B. (2002) Liddi- tourmaline from Brazil, Mozambique and Nigeria. The Journal
coatite tourmaline from Anjanabonoina, Madagascar. G&G, Vol. of Gemmology, Vol. 35, No. 2, pp. 120–139.
38, No. 1, pp. 28–53, http://dx.doi.org/10.5741/GEMS.38.1.28 Pezzotta F., Laurs B.M. (2011) Tourmaline: the kaleidoscopic gem-
Dunn P.J., Appleman D.E., Nelen J.E. (1977) Liddicoatite, a new stone. Elements, Vol. 7, No. 5, pp. 333–338, http://dx.doi.org/
calcium end-member of the tourmaline group. American Min- 10.2113/gselements.7.5.333
eralogist, Vol. 62, pp. 1121–1124. Shigley J.E., Cook B.C., Laurs B.M., Oliveira Bernardes M. (2001)
Fritsch E., Shigley J.E., Rossman G.R., Mercer M.E., Muhlmeister An update on “Paraíba” tourmaline from Brazil. G&G, Vol. 37,
S.M., Moon M. (1990) Gem-quality cuprian-elbaite tourmalines No. 3, pp. 260–276, http://dx.doi.org/10.5741/GEMS.37.4.260
from São José da Batalha, Paraíba, Brazil. G&G, Vol. 26, No. 3, Smith C.P., Bosshart G., Schwartz D. (2001) Gem News Interna-
pp. 189–205, http://dx.doi.org/10.5741/GEMS.26.3.189 tional: Nigeria as a new source of copper-manganese-bearing
Furuya M. (2007) Copper-bearing tourmaline from new deposits tourmaline. G&G, Vol. 37, No. 3, pp. 239–240.
in Paraíba State, Brazil. G&G, Vol. 43, No. 3, pp. 236–239, Soares D.R., Beurlen H., Barreto S. d. B., Da Silva M.R.R., Ferreira
http://dx.doi.org/10.5741/GEMS.43.3.236 A.C.M. (2008) Compositional variation of tourmaline-group
Hawthorne F.C., Dirlam D.M. (2011) Tourmaline the indicator minerals in the Borborema pegmatite province, northeastern
mineral: From atomic arrangement to Viking navigation. Ele- Brazil. The Canadian Mineralogist, Vol. 46, No. 5, 1097–1116,
Teertstra D.K., Černý P., Ottolini L. (1999) Stranger in paradise: Lid-

ments, Vol. 7, No. 5, pp. 307–312, http://dx.doi.org/10.2113/ http://dx.doi.org/10.3749/canmin.46.5.1097
gselements.7.5.307
Henry D.J., Guidotti C.V. (1985) Tourmaline as a petrogenetic in- dicoatite from the High Grade Dike pegmatite, southeastern
dicator mineral: an example from the staurolite-grade Manitoba, Canada. European Journal of Mineralogy, Vol. 11,
metapelites of NW Maine. American Mineralogist, Vol. 70, pp. No. 2, pp. 227–235, http://dx.doi.org/10.1127/ejm/11/2/0227

FEATURE ARTICLES
THE ROLE OF SILICON IN THE

COLOR OF GEM CORUNDUM
John L. Emmett, Jennifer Stone-Sundberg, Yunbin Guan, and Ziyin Sun
Natural corundum is colored by a variety of trace-element impurities and by the chemical reactions
among them. Important to nearly all of the colors is the role played by the reactions among silicon, ti-
tanium, and magnesium. Characterization of the interactions between these aliovalent ions has been
hindered by the fact that the laser ablation–inductively coupled quadrupole–mass spectrometry equip-
ment used in most gem labs does not resolve the interferences for silicon. Thus Si has not been measured
at relevant concentrations. Recent development of ion implant standards and the application of sec-
ondary ion mass spectrometry has allowed the accurate measurement of silicon in corundum for the
first time. As an example of its importance, detailed analyses of sapphire from Montana’s Yogo Gulch
deposit are presented. These analyses show that without silicon, Yogo sapphire would not be blue.
C
orundum, like many other transparent gem- tions will be expressed as atomic concentrations in
stones, is inherently colorless if pure. In fact, units of parts per million atomic (ppma), not weight
pure corundum has an excellent transmission concentrations (ppm or ppmw). These units are cho-
window that can achieve good transparency between sen because it is the concentration of trace elements
approximately 0.18 and 5 µm (Khattack and Schmid, that determines how they chemically interact, not
2001). Pure corundum is aluminum oxide, Al2O3. their relative weights.1
Thus, the only cation (positive ion) in pure corun- Al3+ is one of the smallest trivalent cations, with
dum is Al3+ and the only anion (negative ion) is O2–. a radius of only 53 picometers (pm; 1 pm = 10 –12 m)
Neither of these two ions arrayed in the corundum in sixfold coordination, which means six oxygen ions
lattice structure absorbs light in the visible region of surround a single aluminum ion. This cation site in
the spectrum. The color of corundum gems is there- corundum is quite small, limiting the size and thus
fore determined by trace element impurities that are the variety of trace elements that can be easily incor-
in solution in the corundum lattice. Such gems are porated. Whereas corundum has only a single small
referred to as allochromatic (“other colored”). cation site, other gems (beryl, for example) can have
“In solution” is an important distinction to make. multiple cation sites of different sizes.
Sugar dissolved in water is “in solution.” Fine sand For trivalent cations substituting for Al3+ at geo-
stirred into water is not. “In solution” in corundum logic temperatures, size is the primary factor in deter-
means that if the impurities are cations, they will pri- mining the solubility (Blundy and Wood, 2003; Karato,
marily substitute for Al3+ in the corundum lattice. If 2016). Trivalent ions in corundum are termed isova-
they are anions, they will primarily substitute for O2– lent (same valence). Isovalent ions that are larger than
in the corundum structure. aluminum stretch the lattice. The energy required to
This paper discusses how trace elements chemi- stretch the lattice results in a reduced solubility. Iso-
cally interact with each other, and thus concentra- valent ions that are smaller than Al3+ also create sub-
stantial lattice distortion that reduces solubility.
Cations in corundum with a valence different
than the +3 of aluminum are termed aliovalent (dif-
ferent valence). Since a solid must be rigorously elec-
http://dx.doi.org/10.5741/GEMS.53.1.42 trically neutral, the incorporation of an aliovalent
© 2017 Gemological Institute of America trace element requires either the incorporation of an-
42 THE ROLE OF SILICON IN CORUNDUM COLOR GEMS & GEMOLOGY SPRING 2017
other aliovalent ion to achieve an average charge of
+3 for the pair, or the creation of a charged crystal de- CONDUCTION BAND
fect such as an interstitial ion, a vacancy, or a change
Si3+
in valence of an existing ion. These requirements are
Ti3+
energy intensive and result in very low solubilities
ENERGY (eV)
for aliovalent ions as compared to isovalent ions of V2+
the same size (Blundy and Wood, 2003). Fe2+
Cr3+
The overall result of size and valence constraints V3+
Co2+
in corundum is that in nature only a small fraction Fe3+ 3+
Mn2+ Ni2+
of the elements in the periodic table are found in so- Co
Mn3+ Mg2+ Be2+
Ni3+
lution. The majority of these elements are listed in
table 1, along with their radii in the sixfold coordi- VALENCE BAND
nation of corundum.
The foregoing rules are useful for growth of corun- Figure 1. In this energy level diagram of trace ele-
dum at low temperatures, such as in geologic growth. ments in corundum, trivalent donor ions are shown
However, it should be pointed out that for growth on the left and divalent acceptor ions are shown on
from a corundum melt at 2050°C (Czochralski, the right. A charge-compensating donor-acceptor pair
Verneuil, etc.), other factors can modify these general can form when the donor level is significantly above
rules. the acceptor level. There is no significance to the lat-
eral positions of the energy levels, but the vertical po-
sitions are very significant. Data from Kröger (1984b).
In Brief
• Due to advances in analytical techniques, the silicon based on a summary paper (Kröger, 1984a). That
concentration of gem corundum samples has been paper shows that the silicon donor level lies well
accurately measured for the first time. below the titanium donor level, and we used that
• Si can have a major impact on all colors of corundum, position in our analysis of the effect of silicon on the
as illustrated through the analysis of blue sapphire color of corundum.
from Yogo Gulch, Montana. Unknown at the time was that a second summary
paper by the same author (Kröger, 1984b) corrected
the silicon donor level based on measured data in a
concurrent paper by Kröger and one of his students
The trace elements in corundum chemically in- (Lee and Kröger, 1985). These later two publications
teract with each other, and this interaction is, in clearly show that the correct silicon donor level lies
many cases, key to the resulting color we observe. well above the titanium donor level. The second
In Emmett et al. (2003), this interaction and the re- summary paper also corrects the relative level posi-
sulting colors produced were discussed at length. tions of some of the transition metals. Figure 1 of this
Rather than recreate that discussion here, the reader paper presents data from the second summary paper
is referred to that paper. However, as presented in and clearly shows that the silicon donor level is well
that paper, the chemical reaction among trace ele- above the titanium donor level.
ments is based on their relative energy level posi- While the position of the silicon donor level with
tions in the band gap as shown in figure B-1 in regard to the position of the titanium donor level
Emmett et al. (2003). The positions shown were may seem unimportant, it surely is not. Their rela-
tive positions have a significant impact on how trace
1
The more commonly used units for trace element analyses are ppmw
elements react with each other, and in certain cases
(parts per million by weight), usually written as ppm. One ppm it is only the presence of silicon that allows the tita-
means that there is one microgram of impurity in one gram of crystal. nium to pair with iron, creating the blue coloration
In this article we choose to use the unit ppma (parts per million
atomic) to state trace element concentrations. One ppma means that
of the Fe2+-Ti4+ pair in sapphire.
there is one trace element atom for each million atoms. In corundum, To understand how this comes about, we need to
that is 400,000 Al atoms + 600,000 O atoms. examine the corrected band gap chart in figure 1.
(molecular weight of Al2O3) ⁄5 This is a greatly simplified version of the correct
ppma = • ppmw
(atomic weight of the element) chart from Kröger (1984b). The convention of label-
THE ROLE OF SILICON IN CORUNDUM COLOR GEMS & GEMOLOGY SPRING 2017 43
TABLE 1. Radii of selected elements found in
natural corundum.a
1. If there are two acceptors and a single donor,
Cation Valence Ionic radii (pm) the lowest acceptor preferentially pairs with
Berylliumb +2 45 the donor unless both acceptors are energeti-
cally close together in the band gap, in which
Magnesium +2 72
case the donor will pair with both.
Aluminum +3 53.5
2. If there are two donors at different levels and a
Nickel +3 56
single acceptor, the highest donor will prefer-
Iron +3 55 entially pair with the acceptor. If the two donor
Manganese +3 58 energy levels are energetically close together in
Chromium +3 61.5 the band gap, they will both take part in charge
Gallium +3 62
compensation of the acceptor.
Vanadium +3 64
3. If corundum contains Ti, Mg, and Fe, Ti will
pair with Mg before Fe.
Silicon +4 40
4. If corundum contains Si, Mg, and Fe, Si will
Titanium +4 60.5
pair with Mg before Fe.
a
Data from Shannon (1976) and Dyar and Gunter (2008) 5. If corundum contains Si, Ti, Mg, and Fe, Si will
b
Beryllium is often found in natural corundum, but it is associated with
inclusion clouds. Whether Be exists in solution is not yet clear.
pair with both Mg and Fe before Ti will.
6. When the concentration of Mg exceeds the sum
of both Si and Ti, the excess Mg will be charge
compensated by trapped holes (see Emmett et
al., 2003, p. 92) in oxidizing conditions, or by
ing these levels is to show the donor level before do- oxygen vacancies under reducing conditions, or
nating an electron (e.g., Ti as Ti3+), and the acceptor by both.
levels after accepting the electron (e.g., Fe as Fe2+). As
such, the vertical distance between the donor and ac- Why are we still uncertain about the role of silicon
ceptor states is a measure of the binding energy of in corundum? We certainly understand the roles of
the donor-acceptor pair. Ti, Fe, and Mg. It is because we have not been able to
The beryllium acceptor level in figure 1 is shown measure the silicon content of the sapphire samples
in red, as it was not presented in Kröger (1984b) and we have studied. The instruments used for laser ab-
has not been accurately measured. It is estimated lation–inductively coupled plasma–quadrupole mass
from the original beryllium diffusion experiments spectrometry (LA-ICP-QMS) in use in most laborato-
(Emmett et al., 2003). ries show significant interferences for the three sili-
From the defect chemical reactions and energy con isotopes. Thus, when a very high-purity sample
level positions, we can formulate a series of rules of synthetic sapphire is measured on these instru-
about impurity interactions (Smyth, 2000). These ments, it shows nominally several tens to several
rules will not be valid for all concentration levels or hundreds of ppma silicon that is not there (Shen,
all temperatures. They become less valid as concen- 2010). The reason for this is that in the plasma, small
trations of key elements (such as Mg, Si, and Ti) ex- quantities of unusual molecules are formed with a
ceed a few hundred ppma or at temperatures mass close to that of silicon; these molecules are
substantially above room temperature. A correct for- recorded by the instrument as silicon. For example,
28
mulation requires solving the equilibrium chemical Si is the most abundant isotope of silicon, at 92.23%
defect reactions for all of the impurities simultane- abundance. Its mass is 27.97693 atomic mass unit
ously, but measurements of the equilibrium con- (amu). When ablating corundum, aluminum carried
stants for all the reactions are unavailable. However, by the plasma can combine with hydrogen from even
the following rules are useful for a wide variety of sit- a miniscule amount of water vapor to form 27AlH
uations and, more importantly, will serve to illus- with a mass of 27.98936. The mass difference be-
trate the types of chemical reactions that occur tween 28Si and 27AlH is only about 4 parts in 10,000,
among impurities in corundum. These rules correct which is too small for the instrument to resolve, and
those presented in Emmett et al. (2003), written so the instruments record both as silicon. The mass
when it was believed that the silicon donor level was resolution of quadrupole mass spectrometers used in
below that of titanium. normal gemological laboratories at mass 28 is only 2–
3 parts in one hundred in the normal resolution which was done by Scientific Materials Corporation.
mode, and only about 1 part per hundred in the high- The calibration was then transferred from the ion
resolution mode. The required resolution of 4 parts implant standards to the multiply doped crystals
in 10,000 far exceeds the capability of a commercial with Caltech’s SIMS facility (GIA, 2016). A silicon
“desktop” design ICP-QMS in normal gemological ion implant standard was also prepared with the
laboratories. High-resolution QMS is not available hope that advanced instrumentation would eventu-
commercially. This example dramatically shows why ally allow the analysis of silicon.
a zero trace element concentration, matrix-matched The plan for the standards was to have a three-
standard is critical for the interpretation of data from point calibration—a zero, a mid-range value, and a
these and other analytical instruments. Similar inter- high value for each element. Sapphire crystals were
ferences have been measured for the other two silicon grown that accomplished these goals except for the
isotopes (Shen, 2010). high-range iron value. Iron and its oxides all have
A few years ago, GIA initiated a project to develop high vapor pressure at the Czochralski growth tem-
an improved set of matrix-matched standards for the perature (~2300°C) and thus distill out of the melt.
analysis of corundum with the existing LA-ICP-QMS All of the multi-element doped crystals were cal-
instruments. The approach was to make ion implant ibrated at the SIMS facility against the ion implant
standards in sapphire for each element of interest and standards. When the extremely high-purity sapphire
calibrate them with Rutherford backscattering spec- crystal (the zero standard) was measured by SIMS, it
trometry (RBS). For those trace elements lighter than demonstrated that SIMS could resolve the isotopic
or close in mass to aluminum (i.e., Be, Mg, and Si), interferences for silicon. SIMS could therefore meas-
secondary ion mass spectrometry (SIMS) was used ure silicon in corundum samples against an ion im-
against multiple highly characterized standards kept plant standard.
by Evans Analytical. With calibrated ion implant To create a high-range iron standard, a different
standards, the next step was to grow multi-element approach had to be employed. Natural sapphires
doped sapphire crystals for the laboratory standards, from Montana’s Yogo deposit (figure 2) were evalu-
Figure 2. Paula
Crevoshay’s “Yogo
Columbine” pendant
displays the remarkably
consistent blue color of
Yogo sapphires: 243 in
all, with a total weight
of 13.89 carats. The
piece also contains 0.59
ct of yellow sapphires
and 1.37 ct of dia-
monds, set in 18K yel-
low and white gold.
Photo by Orasa Weldon.
TABLE 2. SIMS analysis of a typical Yogo sapphire sample (in ppma).
Trace elements Mg Si Mg – Si = Ti Ti – available Mg = V Cr Fe Ga
available Mg available Ti
Concentration 84.3 24.6 59.7 79.7 20.0 2.7 2.2 1130 10.4
Standard 1.32 0.44 1.4 0.72 1.6 0.05 0.08 20 0.3
deviation (ppma)
ated for uniformity. The Yogo sapphire is almost Surprisingly, it also turned out that the Yogo sap-
unique for the extreme uniformity of its blue color, phires provided an excellent example of the role of
which indicates a high trace element concentration silicon in gem corundum. The trace element concen-
uniformity, at least for the Fe2+-Ti4+ chromophore. A trations and color are extremely uniform, and thus
typical sample is shown in figure 3. It was hoped that the standard deviations (SDs) of the concentration
this uniformity would also extend to Fe3+. The SIMS measurements are quite small. SIMS analysis of the
analyses of nine pieces of Yogo sapphire at nine points individual trace elements against corundum ion im-
each did show a very high iron uniformity, and thus it plant standards is shown in table 2. The data pre-
was chosen to be the high iron concentration standard. sented are the average of three measurement points
The nine Yogo samples were calibrated with SIMS for on this particular sample.
all elements for which ion implant standards had been Since Si is the highest-lying donor and Mg the
prepared. Since we can easily resolve the mass inter- lowest-lying acceptor, Si will pair with Mg first, leav-
ference for silicon in corundum using the high mass ing 59.7 ppma Mg2+ available to interact with other
resolving power of SIMS, accurate silicon measure- trace elements. Then Ti as a donor will pair with all
ments of natural corundum samples have been ob- the remaining available Mg, leaving 20.0 ppma Ti to
tained for the first time. pair with Fe, producing the blue coloration. It is in-
teresting to note that without silicon, Yogo sapphire
would not be blue because Mg>Ti, and thus no un-
Figure 3. A typical polished Yogo wafer sample meas- paired Ti would be available to pair with Fe. The
uring 0.93 mm thick. The c-axis is perpendicular to analysis of this Yogo sapphire is not unusual but is
the wafer surface. Note the distinct color uniformity typical of all of the Yogo samples analyzed.
and the lack of any color zoning. Ten LA-ICP-QMS This clearly demonstrates the critical role of sili-
measurement points can also be seen on this wafer. con in the color of corundum, a factor that, to the
Photo by John L. Emmett. best of our knowledge, has not been correctly ad-
dressed in the literature.
The SDs presented in table 2 only relate to the
point-to-point analysis variations of the sample. For
determining the SDs of the absolute measurement, it
is necessary to correctly combine the point-to-point
SDs with the SDs of the ion implant standards. When
this is done, the SD result for the “available titanium”
becomes 6.7 ppma, which is well below the stated
concentration value of 20 ppma. While this addition
to the SD is rarely applied in the gemological litera-
ture, it is presented here to emphasize the validity of
the absolute result.
The SIMS analysis of Si conducted on nine Yogo
samples resulted in Si concentrations from 22 to 26
ppma. In addition, ten other samples from Myan-
mar, the Rock Creek deposit in Montana (United
States), the Garba Tula deposit in Kenya, the Subera
deposit in Queensland (Australia), the Pailin deposit
in Cambodia, and the Montepuez ruby deposit in
Mozambique were also analyzed for silicon. The con-
centrations measured ranged from 3.5 to 150 ppma. tributing substantially to the resulting color. The
In each case the silicon concentration was in the current analytical capabilities of most of the world’s
same range as the concentrations of the other aliova- gem labs do not allow the quantification of silicon
lent ions, and thus in each case it had a significant at relevant concentration levels. Consideration
impact on color. should be given to the possibility of obtaining
higher-resolution mass detectors for the existing LA-
CONCLUSIONS ICP equipment, or to evaluating other analytical
It is clear from the foregoing study that silicon is an techniques such as laser-induced breakdown spec-
important trace element in gem corundum, con- troscopy (LIBS).
ABOUT THE AUTHORS ACKNOWLEDGMENTS

Dr. Emmett is the director of Crystal Chemistry in Brush Prairie, Tom Moses and GIA have provided long-term support for numer-
Washington; he is also a consultant to GIA. Dr. Stone-Sundberg ous studies like this one. Tim Thomas initiated GIA’s development
is the managing director of Crystal Solutions in Portland, Oregon, of the new set matrix-matched standards on which this work de-
and a consultant to GIA. Dr. Guan is the laboratory director of the pends. He had the foresight to include an ion implant standard for
Microanalysis Center, Division of Geological & Planetary Sciences, silicon, which made this work possible. Dr. George Rossman has
at the California Institute of Technology in Pasadena, California. facilitated access to Caltech’s excellent SIMS analytical facility. In
Mr. Sun is a staff gemologist at GIA in Carlsbad, California. addition, he continues to provide an extremely helpful debate of
the technical issues, which is greatly appreciated.
REFERENCES
Blundy J., Wood B. (2003) Partitioning of trace elements between Kröger F.A. (1984a) Defect related properties of doped alumina.
crystals and melts. Earth and Planetary Science Letters, Vol. Solid State Ionics, Vol. 12, pp. 189–199, http://dx.doi.org/
210, No. 3-4, pp. 383–397, http://dx.doi.org/10.1016/S0012- 10.1016/0167-2738(84)90148-6
821X(03)00129-8 Kröger F.A. (1984b) Electrical properties of alpha-Al2O3. In W.D.
Dyar M.D., Gunter M.E. (2008) Mineralogy and Optical Mineral- Kingery, Ed., Advances in Ceramics, Volume 10: Structure and
ogy. Mineralogical Society of America, Chantilly, Virginia. Properties of MgO and Al2O3 Ceramics. American Ceramic So-
Emmett J.L., Scarratt K., McClure S.F., Moses T., Douthit T.R., ciety, Columbus, Ohio, pp. 1–15.
Hughes R., Novak S., Shigley J.E., Wang W., Bordelon O., Kane Lee C.H., Kröger F.A. (1985) Electrical conductivity of polycrys-
R.E. (2003) Beryllium diffusion of ruby and sapphire. G&G, Vol. talline Al2O3 doped with silicon. Journal of the American Ce-
39, No. 2, pp. 84–135, http://dx.doi.org/10.5741/GEMS.39.2.84 ramic Society, Vol. 68, No. 2, pp. 92–99, http://dx.doi.org/
Gemological Institute of America (2016). Final report for corun- 10.1111/j.1151-2916.1985.tb15271.x
dum standard sets for GIA LA-ICP-MS instruments. Internal Shannon R.D. (1976) Revised effective ionic radii and systematic
report, June 10. studies of interatomic distances in halides and chalcogenides.
Karato S.-I. (2016) Physical basis of trace element partitioning: A Acta Crystallographica Section A, Vol. 32, No. 5, pp. 751–767,
review. American Mineralogist, Vol. 101, No. 12, pp. 2577– http://dx.doi.org/10.1107/S0567739476001551
2593, http://dx.doi.org/10.2138/am-2016-5665 Shen A.H. (2010) Silicon in sapphires – Its role and detectability
Khattak C.P., Schmid F. (2001) Growth of the world’s largest sap- by LA-ICP-QMS. Goldschmidt Conference Abstracts, p. A945.
phire crystals. Journal of Crystal Growth, Vol. 225, No. 2-4, pp. Smyth D.M. (2000) The Defect Chemistry of Metal Oxides. Ox-
572–579, http://dx.doi.org/10.1016/S0022-0248(01)00955-1 ford University Press, New York.
FEATURE ARTICLES
JAPANESE JADEITE:
HISTORY, CHARACTERISTICS, AND
COMPARISON WITH OTHER SOURCES
Ahmadjan Abduriyim, Kazuko Saruwatari, and Yusuke Katsurada
Even though Japanese jadeite lacks the transparency of the highest-quality Burmese imperial jadeite, its
rarity and natural features make it a highly valued gemstone. In this study, jadeite from the Itoigawa and
Omi regions in Niigata Prefecture and the Wakasa region in Tottori Prefecture, both on Japan’s western
coast, were divided into several color varieties corresponding to chromophores and mineral phases: white
(nearly pure jadeite), green (Fe-rich, Cr-bearing), lavender (Ti-bearing), blue (Ti- and Fe-bearing), and black
(graphite-bearing). White jadeite from Itoigawa-Omi was close to pure jadeite (XJd = 98, or 98% jadeite
composition). Green jadeite from the same location had an XJd range from 98 to 82. The maximum CaO
content in green jadeite was 5 wt.%, and its chromophores were Fe and Cr. Whereas lavender samples
had a jadeite composition of XJd = 98 to 93 and tended to be high in TiO2 and FeOtot and low in MnO con-
tent, blue jadeite showed the highest TiO2 concentration at 0.65 wt.% and had an XJd range of 97 to 93. A
blue jadeite from Wakasa had a range of 97 to 91 and a similarly high TiO2 concentration. In trace-element
analysis, chondrite-normalized and primitive mantle–normalized patterns in lavender, violetish blue, and
blue jadeite from Japan showed higher large-ion lithophile element contents (Sr, Ba) and higher field strength
element contents (Zr, Nb) than those in green jadeite, while white and black jadeite had relatively low REE
contents. The Japanese jadeites were compared to samples from Myanmar, Guatemala, and Russia.
J apan is an important source of jadeite, much of

which comes from the Itoigawa and Omi regions in
Niigata Prefecture. The Kotaki area upstream of the
Hime River in Itoigawa-Omi was the first docu-
This study introduces the historical background
and sources of Japanese jadeite (figure 1). It describes
Figure 1. A large, attractive jadeite boulder from

mented source of gem-quality jadeite and jadeite-bear- Itoigawa, Japan, characterized by mixed white and
ing rocks in Japan (Kawano, 1939; Ohmori, 1939). This green colors. This boulder weighs 40.5 kg and meas-
area is located in the high-pressure, low-temperature ures approximately 39 cm high, 32 cm long, and 26
metamorphic Renge belt within a Late Paleozoic sub- cm wide. Photo by Ahmadjan Abduriyim.
duction zone (Shibata and Nozawa, 1968; Nishimura,
1998). Tsujimori (2002) suggested that blueschist to
eclogite metamorphism was related to the subduction
of oceanic crust. Miyajima et al. (1999, 2001, 2002) and
Morishita (2005) proposed that the ﬂuids that facili-
tated the formation of jadeite in Itoigawa-Omi were
related to subduction zones. U-Pb zircon dating of
jadeite-natrolite rocks in the area indicated that the
age of jadeitization is about 519±17 Ma (Kunugiza et
al., 2002; Tsutsumi et al., 2010).

http://dx.doi.org/10.5741/GEMS.53.1.48
© 2017 Gemological Institute of America
48 JAPANESE JADEITE GEMS & GEMOLOGY SPRING 2017

Kamuikotan
Serpentinite
Itoigawa-Shizuoka
tectonic line
century, the ancient state of Koshi (in modern-day Ni-
Jade mine
Hokkaido igata Prefecture) was ruled by a beautiful empress
Renge-Sangun belts who wore a mysterious curved green jadeite (figure
Sanbagawa belts 3). Koshi produced a variety of beautiful stones and
cultivated a thriving trade with many other parts of
Japan. Typically excavated from the tombs of power-
ful people, magatama jadeite appears to have been a
Itoigawa
sacred ornament as well as a symbol of wealth and
and Omi prestige. Magatama carvings spread to the Korean
Renge Peninsula, where they have been excavated at many
belt
Oya archaeological sites (barnes, 1999).
Sangun Wakasa Yorii
Yorii and
Chichibu Thousands of years of jadeite culture went into
belt
decline during the mid and late Kofun period (3rd to
O
Oosa
Nagasaki 7th century AD) before disappearing in the 6th cen-
Mikkabi tury. Jadeite was rediscovered in Japan in 1938, more
than a thousand years after vanishing, when re-
Engyoji and Sanbagawa belts searcher Eizo Ito uncovered it at the Kotaki River in
Inocho the city of Itoigawa. The following year, Dr. Yoshi-
Figure 2. Japanese jadeite is found in eight locations, Figure 3. In this mosaic painting, made with pieces of
though the only significant source of gem-quality ma- Itoigawa jadeite, the magatama carving is worn by
terial is the Itoigawa-Omi region in Niigata Prefec- an empress of the ancient state of Koshi, in modern-
ture. Modified after Fossa Magna Museum. day Niigata Prefecture. Courtesy of the Jade Ore Mu-
seum (Hisui Gensekikan).
the material’s color varieties, internal texture, and

chemical features using quantitative electron micro-
probe (EPMA) and laser ablation–inductively coupled
plasma–mass spectroscopy (LA-ICP‐MS) analysis.
HISTORICAL BACKGROUND
In addition to Japan, major jadeite localities include
Myanmar, Russia, Central America, and the United
States. Some of the world’s earliest jadeite jade arti-
facts emerged from the Olmec, Maya, and Aztec civ-
ilizations of modern-day Mexico and Guatemala,
which flourished from about 1200 bC until the Span-
ish conquest in the 16th century (Foshag and Leslie,
1955; Umehara, 1971; Taube, 2004). During the
Jomon era, about 5,500 years ago, Japan’s Itoigawa re-
gion became the birthplace of jadeite carving (figure
2), and it is no exaggeration to say that the Japanese
gem culture was derived from this area. In the middle
of the Jomon era, pendant-like jadeite pieces called
taishu were produced and traded throughout many
parts of Japan. Rough jadeite fashioning techniques,
including spherical bead carving, were passed on in
the late Jomon era. In the Yayoi era, curved maga-
tama and tube-shaped kudatama beads became pop-
ular. According to legend dating from the early 8th
JAPANESE JADEITE GEMS & GEMOLOGY SPRING 2017 49

Itoigawa N
²
H
im
iver
Omi
e
between the Pacific plate and the Eurasia plate con-
Ri
ive
Tomi R
ve
iR
r
Uta taining the Japanese islands, under which the cold Pa-
Om
cific plate subducts. This area is thought to have a
Ichiburi SHashidate S
Jadeite Valley high-pressure, low-temperature condition that pro-
Sa
Agero
duces jadeite. Japan has eight jadeite occurrences in all
ka
Negoya
iR
S
(again, see figure 2). Most of the jadeite from the Renge
ive
S
S
r
S S
S
SS
and Sangun belts on the western side (Itoigawa, Oosa,
SS
S
Kotaki S Oya, and Wakasa) is very pure, composed of more than
S
Jadeite Valley
S
S
S
S
S
S
90% jadeite (including similar omphacite). Material
S S S
Sedimentary rock S S
and granitic/
S
S
S SS S S from other parts of Japan very rarely contains more
metamorphic rocks S S SS Hiraiwa
S
than 80% jadeite. Most contains large amounts of al-
Serpentinite
S
SS
bite, kyanite, and analcime and no more than 50%
Limestone S
SS jadeite (Yokoyama and Sameshima, 1982; Takayama,
Jadeite mine S
Kita-Otari
1986; Miyazoe et al., 2009; Fukuyama et al., 2013).
S Albitite mine
S
Fault
Renge and Sangun Belts. The Itoigawa region is as-
Itoigawa-Shizuoka Renge-Onsen
tectonic line
S
signed to the Renge belt, a serpentinite mélange zone
0 5 km
S Mountain
S S with various types of tectonic blocks, high-pressure
Figure 4. Jadeite from Itoigawa is found in serpenti-
and low-temperature metamorphic rocks, metamor-
nite along a fault as blocks near the Kotaki River (up- phosed sedimentary rocks, amphibolites, and rodin-
stream of the Hime River) and the Omi River. The gites (Nakamizu et al., 1989). Gem-quality jadeite has
Kotaki and Hashidate valleys are the main sources of been found only at the Kotaki and Hashidate districts
gem-quality jadeite. White is the most common color, in Itoigawa-Omi, occurring as boulders in the serpen-
followed by green. Lavender, violet-blue, and blue tinite located at the fault border between the Permian-
jadeite are also found in Itoigawa-Omi. Source: Fossa Carboniferous limestone and Cretaceous sandstone
Magna Museum. and shale. Jadeite boulders range from one meter to
several meters in size and are mostly distributed in an
area several hundred meters long. Jadeite rocks in Ko-
taki show concentric zoning, toward the rim, of al-
nori Kawano and his colleagues at Tohoku Univer- bitite (with or without quartz), white jadeite, green
sity published a study of the samples (Kawano, 1939; jadeite, soda-rich calciferous amphibole, and host ser-
Ohmori, 1939). Subsequent research led to additional pentinite. Omi jadeite rock shows a “distinct strati-
discoveries in the Kotaki area, upstream of the Hime form structure,” sometimes with alternating coarse
River, as well as in the Hashidate area of Itoigawa and fine compact layers and often containing lavender
(figure 4). In 1954, some of these areas were desig- jadeite (Chihara, 1991).
nated as preservation sites, but jadeite is still found Sources other than Itoigawa in the Renge and San-
along these rivers or their estuaries. gun belts (Oosa, Oya, and Wakasa) produce limited
Jadeite from Itoigawa, especially along the coast, amounts of jadeite, most of it white with a few green
is beautiful even in its rough state. Colors include areas. Green jadeite with high transparency has not
white, green, violet, blue, and black. but because the been found in these areas. Considering that it con-
sites are protected and mining is not allowed, there tains similar minerals as well as zircons that are
is little supply of this material in the market. In Sep- about 500 million years old (Tsutsumi et al., 2010),
tember 2016, Itoigawa jadeite was chosen as Japan’s the material from Oosa, Oya, and Wakasa presum-
national stone by the Japan Association of Miner- ably formed through the same process as the
alogical Sciences. Itoigawa jadeite. These fine-grained specimens can-
not be differentiated microscopically from those of
JADEITE FROM JAPANESE LOCALITIES Itoigawa.
Jadeite is found in high-pressure, low-temperature The Wakasa region in Tottori Prefecture of western
metamorphic belts (Essene, 1967; Chihara, 1971; Har- Japan is a source of blue jadeite. Jadeite and jadeite py-
low and Sorensen, 2005). It is associated with kyanite, roxene occur in serpentinites and metagabbros related
an indicator mineral of high-pressure, low-tempera- to the Sangun regional metamorphic belt (Kanmera et
ture metamorphism. The Japan Trench is a boundary al., 1980; Chihara, 1991). In this locality, jadeite rock

formed as a vein in part of a serpentinite body ranging jadeite occurs in clusters, and samples with more
from 5 to 30 cm in diameter. The rocks are mostly than 80% jadeite content are rare. The jadeite from
weathered but still hard and compact. Most of the this area is not suitable for jewelry and, like material
jadeite has a violetish blue to blue, lavender, and milky from Hokkaido, cannot be differentiated from the
white color and is associated with albite, quartz, and surrounding metamorphic rocks. Mikkabi in
chlorite. Wakasa jadeite is also known to have a blue Shizuoka Prefecture also produces jadeite, found as a
color, but production is very limited. white vein 2 to 3 cm thick in metagabbro, but it too
is unsuitable for fashioning.
Hokkaido. The northern Japanese island of Hokkaido
is shown in figure 2. The Kamuikotan belt, a high- Kochi and Nagasaki. Rocks containing jadeite have
pressure metamorphic belt, extends north to south occasionally been found within serpentinite in the
in Hokkaido. In the serpentine area of this metamor- city of Kochi. These include gray quartz-bearing
phic belt in the Asahikawa district, jadeite-bearing rocks and green rocks containing pumpellyite and
rocks are very rare but may locally contain more kyanite. Neither contains more than 60% jadeite or
than 80% jadeite. Most of the jadeite around 10 cm possesses transparency, and thus cannot be visually
in size contains less than 50% jadeite content, how- differentiated from common hard metamorphic
ever. Various mineral components deprive the rocks. In Nagasaki, rocks containing jadeite associ-
jadeites of their transparency, making them difficult ated with serpentine have been reported. Parts of the
rock contain more than 80% jadeite, but the content
is often as low as 50%. Only limited amounts of the
jadeite-bearing rocks have been produced.
In Brief
• Japan’s only source of gem-quality jadeite is the SAMPLE DESCRIPTIONS AND
Itoigawa-Omi region in Niigata Prefecture. The area ANALYTICAL METHODS
belongs to a serpentinite mélange zone with high-
To examine the color varieties of Japanese jadeite, we
pressure and low-temperature metamorphic rocks,
amphibolites, and jadeitites. collected representative samples from the field in
Itoigawa and Wakasa and from the Jade Ore Museum
• Some 5,500 years ago, jadeite carvings were traded
throughout Japan. Since the rediscovery of the jadeite (Hisui Genseki Kan) in Tokyo and the Fossa Magna
source in 1938, limited quantities have been available. Museum in Itoigawa (see the top table at
https://www.gia.edu/gems-gemology/spring-2017-
• These Japanese jadeites display a variety of colors, in-
cluding white, green, lavender, violetish blue, blue, japanese-jadeite-tables). The 39 rough and cut Japan-
and black. White jadeite shows a very pure jadeite ese samples consisted of white, green, dark green,
component, while green jadeite has a low percentage lavender, violetish blue, blue, and black jadeite. They
of omphacite ranging from 2% to 18% and is colored were from three different sources: the Kotaki River
by Fe and Cr. Blue samples are enriched with Ti. area (36º55'33N, 137º49'18E; 32 samples) and
• Trace element analysis by LA-ICP-MS confirmed that Hashidate (36º58'35N, 137º45'51E; five samples),
lavender, violetish blue, and blue jadeite showed both in Itoigawa, and the Wakasa region (35º32'17N,
higher large-ion lithophile elements and higher field 134º44'52E; two samples).
strength elements than green, white, and black jadeite.
To compare their optical features, petrographic
structures, and geochemistry with samples from
other parts of the world, we also tested Russian white
to distinguish from surrounding green rocks of law- and green jadeite from the Polar Urals (four samples);
sonite-albite facies retrograded in greenschist facies. dark yellowish green and lavender jadeite from the
Motagua region of Guatemala (six samples); and
Sanbagawa Belt. Two deposits of jadeite-bearing rock burmese white, green, and lavender jadeite from
have been reported in the Yorii and Chichibu dis- Kachin State (38 samples). The samples were pro-
tricts in the Kanto Mountains of Saitama Prefecture. vided by the Jade Ore Museum (Hisui Gensekikan)
One of the locations forms a dome with serpentine, and Miyuki Co., Ltd. Examples are shown in figure 5.
and the maximum jadeite content there is about All 87 samples were observed by visual and mi-
50%. Another occurrence, accompanied by actinolite croscopic means, and their refractive indices were ex-
rocks that supposedly replaced serpentinite, is simi- amined by either normal reading from flat wafers or
lar to the surrounding metamorphic rocks where by the spot method. Specific gravity was determined

A
Figure 5. Representative
samples from four loca-
tions. A: Magatama
carvings (25.51–63.88
ct) from Itoigawa-Omi.
B: Variously colored
rough jadeite boulder,
cobble, and pebbles
from Itoigawa-Omi
(15.3–622.2 g) and a pol-
ished violetish blue
B jadeite from the
Wakasa region (458.2 g).
C: Green and lavender
cabochons (4.75–15.45
ct) and polished slices
from Kachin, Myanmar
(17.84–136.4 g). D:
Rough grayish green
C and lavender jadeite
blocks from the Mo-
tagua region of Guate-
mala (223–1250 g). E: A
green jadeite block (304
g) and two polished
translucent to opaque
jadeites (139.08 and
5.93 ct) from the Polar
Urals of Russia. The
magatama from
Itoigawa-Omi and the
Russian and Guate-
malan samples are
D
courtesy of the Jade
Ore Museum (Hisui
Gensekikan). The
Burmese jadeites are
courtesy of Miyuki Co.,
Ltd. Photos by Masumi
Saito and Ahmadjan
E Abduriyim.
hydrostatically for all the samples, and their absorp- Itoigawa and Wakasa, three from Kachin, two from
tion spectra were observed by handheld prism spec- Motagua, and two from the Polar Urals were cut and
troscope. Four inclusion-bearing samples from polished into thin sections for petrographic structure

analysis. The micro-texture of these petrographic Mg(Kα), and Ni(Kα), respectively.
thin sections was observed using a Nikon Optiphot Trace element and rare earth element (REE) analy-
polarized light microscope. ses were performed with LA-ICP‐MS using a Thermo
To identify the inclusions and reveal the distribu- Scientific iCAP Q quadrupole ICP-MS with an ESI
tion of jadeite and other minerals in jadeite rock, we UP213 Nd-YAG laser. The laser repetition rate was
used a two-dimensional micro-Raman mapping spec- 7 Hz, with an energy density of 10 J/cm2 and a spot
troscope (Horiba Jobin Yvon XploRA LabRAM HR size of 40 μm, using a carrier gas mixture of helium
Evolution) equipped with a 532 nm Nd:YAG laser and argon. It was possible to detect the signal of all
and an optical microscope (Olympus bX41) under isotope ratios and achieve an analytical precision of
real usage conditions. The laser beam was narrowed less than 10% relative standard deviation (RSD).
and focused through a 300 μm aperture and a 100× Three to ten spots were ablated for each sample, and
objective lens, yielding a spatial resolution of about averaged data was calibrated. NIST SRM 610 and 612
1 μm. Raman spectra were acquired with a single were used as external standards. before analysis, the
polychrome spectrometer equipped with a Si-based samples were cleaned with acetone and aqua regia in
charge-coupled device (CCD) detector (1024 × 256 an ultrasonic bath to eliminate surface contamina-
pixels). A composite spectrum in the range of 200– tion.
1800 cm–1 was obtained with LabSpec software, using
a 600 gr/mm grating with a spectral resolution of RESULTS AND DISCUSSION
about ±2.5 to 3.5 cm–1. before performing the meas- Gemological Observations. Jadeite pebbles from
urements, we calibrated the spectrometer by the Si Itoigawa-Omi tend to have rounded corners, result-
520 cm–1 peak. Raman spectral mapping was con- ing from erosion by fluvial processes, and a glittering
ducted in point-by-point macro mode using XY step- whitish surface. because of surface weathering, the
ping motors. It took 15 minutes to obtain a 44 × 30 rough rock does not have a brown skin. These stones
mm spectral macro mapping image with a step size are mainly white, with unevenly distributed pale
of 500 μm in the 200–1800 cm–1 range. green to green areas, and they feel rigid, compact, and
UV-Vis-NIR spectroscopy was performed on par- heavy. Most of the white rocks mixed with some
allel polished wafers of green, lavender, and blue green were in boulder, pebble, and nodule form,
samples from Kotaki, Itoigawa; a violetish blue sam- transparent to semi-translucent to opaque, and finely
ple from the Wakasa region; green and lavender sam- textured, with some coarse texture in eye-visible sin-
ples from Kachin and Motagua; and a green sample gle crystals. The largest rough specimen, found in the
from the Polar Urals. The analyses were performed Hashidate district, weighed 102 tons. The author has
using a Hitachi U-2900 spectrophotometer at 1 nm also observed a 4.6 ton jadeite rock from the Kotaki
resolution. The parallel polished plates were ori- district that is housed in the Fossa Magna Museum
ented with the main polished face perpendicular to (figure 6, left). In this large jadeite boulder, most of
the instrument beam, and the polarizer was not ro- the white and green parts were jadeite jade, while the
tated. The thin sections ranged from 3.63 to 10.08 ct fibrous black portion was composed of amphibole.
and from 0.86 to 3.70 mm thick. Some small green areas were translucent and
To investigate the various colors of the jadeites, gemmy. Some minor faults were filled with white
very precise quantitative chemical composition meas- minerals such as prehnite, pectolite, and zeolite-
urements were obtained by electron microprobe with group minerals that formed within fluids from the
wavelength-dispersive spectrometry (WDS) mode deeper part of the earth.
(JEOL LXA-8900) at the University of Tokyo and In lavender jade, the violet color may be dispersed
Waseda University. Eleven thin sections and polished irregularly over the white matrix. This color is semi-
thin plates from Itoigawa and Wakasa, two specimens translucent to opaque, with a fine to medium texture
from Motagua, and one specimen from the Polar Urals (figure 6, center). The blue jadeite samples, found in a
were analyzed with 15 kV accelerating voltage, using variety of beautiful colors, were rounded and semi-
a beam current of 12 nA and beam diameter of 10 μm. translucent to opaque, with fine to coarse texture (fig-
Data were processed with ZAF correction software. ure 6, right). Aggregates of minute crystals were
The standards were natural albite for Al(Kα) and observed through a loupe but lacked crystal form.
Na(Kα), wollastonite for Ca(Kα) and Si(Kα), orthoclase Specimens from Itoigawa and Tottori had a spot RI of
for K(Kα), chromite for Cr(Kα), Mn-olivine for Mn(Kα), 1.65 to 1.66 and SG values ranging from 3.10 to 3.35.
and TiO2, Fe2O3, MgO, and NiO for Ti(Kα), Fe(Kα), Green jadeite samples were inert under long-wave

Figure 6. Left: A 4.6 ton jadeite rough boulder is displayed at the Fossa Magna Museum in Itoigawa. This eroded
and rounded boulder from the Kotaki area is mostly white, with some green areas of jadeite; the fibrous structure
in the black area is amphibole. Thin fault-like veins are filled with white minerals. Center: A 30 kg rounded boul-
der of predominantly lavender jadeite was found along the Hime River in Itoigawa. The lavender color is dis-
persed irregularly over the white matrix. Right: These rounded jadeite pebbles, found along the coast in
Itoigawa-Omi, are approximately 2 to 15 cm long. Photos by Ahmadjan Abduriyim, courtesy of the Fossa Magna
Museum and the Jade Ore Museum (Hisui Gensekikan).
(365 nm) and short-wave (254 nm) UV radiation. The 14; see figure 7-A1) revealed colorless, semitranspar-
lavender jadeite exhibited a stronger reddish fluores- ent jadeite crystals distributed in the white area,
cence than burmese lavender jadeite, which showed mostly as fine cryptocrystalline grains around 0.05–
a weak reddish fluorescence to long-wave UV. The 0.3 mm in size. Under cross-polarized light, the fine
Japanese blue jadeite was inert to both long- and short- jadeite grains showed both high- and low-order inter-
wave UV. The absorption spectra of all Japanese jadeite ference colors, caused by the different orientation of
samples, measured by a handheld spectroscope, re- each grain. Under plane-polarized light, we occasion-
vealed weaker lines at 690, 650, and 630 nm. In addi- ally observed in matrix large pale green grains over 2
tion, green jadeite from Itoigawa showed a very sharp mm (figure 7-A2) that were well-formed jadeite sin-
line at 437 nm. The lavender jadeite showed weak gle crystals. Their well-developed cleavages inter-
bands at around 530 and 600 nm and a narrow band sected at 87° angles, which is characteristic of
at 437 nm. The blue jadeite showed a very broad band pyroxene. This thin section of green jadeite showed
from the yellow to red portion of the spectrum, as well a prismatic crystalloblastic texture, indicating meta-
as a weak narrow band at 437 nm. morphism under nondirectional pressure. Micro-
The representative Guatemalan jadeite samples se- Raman spectrometry in microfolds and veinlets
lected for this study were grayish and dark green, identified minor amounts of pectolite and prehnite
white, and violetish blue. The green rough was semi- as component minerals.
translucent and opaque, with a fine to medium- The thin section of lavender jadeite from Itoigawa-
grained texture but also somewhat coarsely grained Omi was almost colorless under plane-polarized light
texture in visible crystals. The “Olmec blue” rough (see figure 7-b1). The sample was semi-transparent
from Guatemala was variegated violet to blue mixed to translucent and mainly composed of fine to micro-
with abundant white color. It was translucent to grained crystals around 0.1–0.3 mm in size, showing
opaque, with a fine texture. Its color distribution a prismatic crystalloblastic texture. Ultramylonitic
closely resembled that of Japanese lavender and blue zones with radiating aggregates of fine jadeite grains
jade. randomly cutting through the matrix were observed
Jadeite from the Polar Urals occurs in different in this sample (figure 7-b2). This texture indicates
shades of green. The material usually has a more even that the sample underwent lithostatic and possibly
color distribution than Japanese green jadeite, and it subsequent directional pressure during the metamor-
is highly valued. The samples from this source were phic process. Prehnite and analcime, the main con-
semitransparent to translucent, with a fine to medium stituents of the veinlets that cut through the jadeite
texture. black spots of magnetite could be observed. rock (figure 7-b3), were formed by hydrothermal flu-
ids (Shoji and Kobayashi, 1988). A long prismatic
Petrographic Observation. In plane-polarized light, a vesuvianite crystal with high relief was also found as
white and green jadeite slice from Itoigawa (K-IT-JP- a component mineral.

Jd Figure 7. Petrographic microscope images of jadeite sam-
ples from Itoigawa. The left images are under plane-polar-
ized light, the right images under cross-polarized light. Jd,
Preh, Ves, An, and Ti indicate jadeite, prehnite, vesuvian-
ite, analcime, and titanite, respectively. A1: Green jadeite
A1 K-IT-JP-14 shows a prismatic-granular crystalloblastic tex-
ture, a distribution of fine colorless cryptocrystalline grains
of jadeite, and a predominance of grains around 0.05–0.3
mm in size. A2: Coarse pale green grains larger than 2 mm
can also be observed in the same matrix, which is a well-
Jd
formed large jadeite single crystal. The prominent (110)
cleavage planes intersecting at 87° are characteristic of py-
roxene. B1: Lavender jadeite K-IT-JP-25 shows near-color-
less fine and micro-grained jadeite crystals with a
A2
prismatic crystalloblastic texture. B2: Ultramylonitic zones
with radiating aggregates of fine jadeite grains cut ran-
domly through the center of the matrix, indicating a litho-
static pressure during the metamorphic process. B3: Dark
Jd gray prehnite and analcime, the main constituents of the
veinlets that cut through this lavender jadeite, were formed
by hydrothermal fluids. A high-relief prismatic vesuvianite
crystal was also found as a component mineral. C1: K-IT-
B1 JP-16 is a predominantly blue specimen, translucent with
fine cryptocrystalline grains. C2: Crushed preexisting min-
erals produce a flow structure with granoblastic and my-
lonitic texture. The component minerals analcime and a
very minor amount of euhedral titanite grains are observed
Jd in the matrix. Photomicrographs by Ahmadjan Abduriyim.
B2 In the blue jadeite sample (K-IT-JP-16), the blue

area was larger than the white area, and the colors
Preh gradually blended. It was translucent and granular,
and the fine cryptocrystalline grains from 0.1 to 0.5
Jd
mm (figure 7-C1) showed granoblastic and mylonitic
texture. In this specimen, preexisting minerals were
Ves
crushed and slipped to produce a flow structure (fig-
ure 7-C2). Component minerals included analcime
B3 and titanite, as well as very minor amounts of euhe-
dral titanite grains in the matrix that do not con-
tribute to the blue color in this type of jadeite.
Morishita et al. (2007) proposed that jadeite from
Jd
Itoigawa-Omi formed either by direct precipitation of
minerals from aqueous fluids or by complete metaso-
matic modification of the precursor rocks by fluids.
The burmese green and lavender jadeite samples
C1 showed amphibole, albite, kosmochlor, and nepheline,
while vesuvianite was rare. The Guatemalan green
An
and lavender jadeite in this study showed grossular
Ti garnet, albite, and rutile, while the Russian green
jadeite contained magnetite and analcime.
We used two-dimensional point-by-point Raman
macro mapping to reveal the distribution of jadeite
and other component minerals in the jadeite rocks.
C2

Figure 8 shows a white and green/dark green sample trations (57, 19, 2.3, and 0.4 ppma, respectively; see
from Itoigawa; the red, green, and blue colors corre- the bottom table at https://www.gia.edu/gems-
spond to the integrated intensity of jadeite, amphi- gemology/spring-2017-japanese-jadeite-tables).
bole (richterite), and prehnite, respectively. This The UV-Vis spectra of lavender jadeite from
image reveals that the jadeite and amphibole grains Itoigawa-Omi show features that correspond with Mn,
are mixed within the matrix, while prehnite occurs Ti, and Fe (figure 10). A broad Mn3+-related absorption
in the vein. The rapid point-by-point confocal map- band centered at 530 nm is often observed in burmese
ping technique can be performed on a whole speci- lavender jadeite (Lu, 2012), as well as a characteristic
men or a region of interest to examine finer details, broad band of paired Ti4+-Fe2+ charge-transfer ions cen-
making it possible to classify the distribution of tered at 610 nm and a narrow Fe3+-related 437 nm ab-
jadeite vs. omphacite and/or amphibole. sorption band. The color-causing transition elements
were analyzed in this lavender jadeite. The results
UV-Vis Spectroscopy. Japan. UV-Vis absorption spec- showed that Ti (534 ppma average) and Fe (550 ppma
troscopy was performed on the green, violet, violetish average) were clearly responsible for its blue hue. Mn
blue, and blue jadeite wafers from the Itoigawa-Omi concentrations averaged 18 ppma and produced a
and Wakasa regions. Chemical analysis was carried weak pink or purple hue. The Japanese lavender
out on similarly colored areas of the sample to confirm jadeite showed a violet color, owing to the combina-
each chromophore’s concentration. tion of minor pink and significant blue hues caused
Green portions of Itoigawa jadeite are generally col- by Mn3+ and Ti4+-Fe2+ absorption. Shinno and Oba
ored by chromium and iron, showing a 691 nm ab- (1993) discussed the substitution of Ti3+ at 545 nm in
sorption line (the Cr3+ “chromium line”) and another lavender jadeite from Itoigawa-Omi. However, the Ti3+
absorption line that originates at around 437 nm (the ion is very unstable in nature and is found only in me-
Fe3+ “jadeite line”); see figure 9. The chromophore con- teorites and lunar samples formed in more reducing
centrations in the tested area, a 5 mm circle, were an- conditions (burns, 1981). In terms of ionic radius, iso-
alyzed by LA-ICP-MS, and a concentration was valent Ti3+ is noticeably larger than Al3+ and does not
averaged from three to four laser ablation spots. The replace it in the six-fold coordinated octahedral site
green area contained relatively high Cr and Fe (280 and (figure 11). The chromophore concentrations of Ti and
810 ppma), and the isovalent chromophores Cr3+ and Fe in the violet parts reached 550 and 534 ppma, and
Fe3+ clearly contributed to the green color (Rossman, this combination caused a noticeable blue color.
1974; Harlow and Olds, 1987). The less significant The UV-Vis spectra of blue jadeite from Itoigawa
chromophores Ti, Mn, V, and Co had lower concen- show a very broad band from 500 to 750 nm, a weak
INTENSITY (ARB. UNITS)
-15000 0.70 Figure 8. A two-dimen-

-10000 0.60 sional Raman spec-
-5000
Jadeite troscopy mapping
0.50
Y (μm)
0 0.40 image of a jadeite boul-

5000 0.30
der from Itoigawa. The
Richterite red area corresponds to
10000 0.20
Prehnite
the jadeite distribution,
15000 0.10
the green area to dark
-20000 -10000 0 1000
00 20000 100 300 500 700
0 900 1100 1300 1500 1700
green richterite (amphi-
WAVENUMBER (cm–1)
AVENUMBER bole group), and the
blue area to the min-
eral prehnite, which is
Jd located at the veinlet
Rich Preh and crosses through the
stone. The mapping
Jd area is 44 × 30 mm.
1 cm

UV-VIS SPECTRUM
2.5
Figure 9. The UV-Vis
V Co Mn Cr Fe Ti spectrum of a green
Green jadeite NaAlSi206 (ppmw) 5.8 1.3 52 735 2290 134
(ppma) 2.3 0.4 19 280 810 57
sample from Itoigawa
(K-IT-JP-14) shows the
OEFFICIENT (cm–1)
Fe3+ features corresponding

2
to Cr- and Fe-bearing
green jadeite: the char-
acteristic narrow 691
nm absorption, two
Cr3+-related weak
ABSORBANCE CO
15
1.5
Cr3+
shoulders at 650 and
630 nm, and a sharp,
Cr3+ narrow absorption of
1 Fe3+ at 437 nm. The sat-
urated green area of the
K-IT
T--JP-14, 2.90 mm thick
spectrum corresponds
to the concentration of
0.5
the chromophores Cr
400 450 500 550 600 650 700 750 800 and Fe (280 and 810
WAVELENGTH (nm)
WA ppma, on average).
Fe3+ absorption at 437 nm, and a cutoff above 350 nm color. by comparison, Mn was too low (64 ppma) to
(figure 12). This absorption pattern is similar to the produce a pinkish component.
spectra of blue sapphire and can be attributed to a The violetish blue jadeite from Wakasa in Tot-
charge transfer between Ti4+-Fe2+ pairs (Ferguson and tori Prefecture showed a similar spectral character-
Fielding, 1971). Significant amounts of Ti (1943 istic, with lower Ti, Fe, and Mn concentrations than
ppma) and Fe (4212 ppma) produced a noticeable blue blue jadeite but higher concentrations than violet
UV-VIS SPECTRUM
V Co Mn Fe Ti
2 Lavender jadeite NaAlSi206 (ppmw) 1.9 0.9 51 1510 1280
(ppma) 0.7 0.3 18 534 550
spectrum of a lavender
OEFFICIENT (cm–1)
sample from Itoigawa-

Omi (K-IT-JP-25). Two
broad bands centered
Mn3+ at 530 and 610 nm cor-
respond to Mn and Ti-
Fe3+ Ti4+-Fe
Fe2+
Fe charge transfer, and
ABSORBANCE CO
1.5
there is also a weak
narrow band at 437
nm. The violet color re-
flects the chromophore
combination of low Mn
K-IT
T-JP-25,
-JP-25, 3.70 mm thick
(18 ppma) and much
higher Ti (534 ppma)
1
and Fe (550 ppma).
380 430 480 530 580 630 680 730 780
WAVELENGTH (nm)
WA

narrow Fe3+-related absorption band at 437 nm that is
very common in natural green jadeite. The Cr and Fe
absorption feature generally overlapped with the spec-
trum of Japanese green jadeite, but the absorption in-
tensity was much higher in burmese jadeite due to its
color saturation and transparency (figure 13A). The
burmese lavender jadeite (K-MYA-20) with a predom-
inantly purple color component had a broad absorp-
tion centered at 570 nm, related to Mn concentration
(figure 13b). The Fe and Ti concentration was much
lower than in Japanese lavender jadeite and might not
Figure 11. Illustration of the atomic arrangement of a cause any noticeable blue component.
single-chain pyroxene crystal structure, from the a-
axis to b-axis direction. In the SiO4 tetrahedrons (in- Guatemala. Only the characteristic Fe3+-related nar-
dicated in yellow and brown), four oxygen atoms row absorption band at 437 nm was found in the
surrounding a silicon atom are connected to form a Guatemalan grayish green jadeite spectrum, which
strand, and other six-coordinate octahedron atoms (in lacked the Cr3+ absorption (figure 13C). A very closely
green) are arranged so that they link the strands. matched absorption spectrum was observed in
Larger atoms (orange spheres) fill the spaces. Modified Guatemalan lavender jadeite, which showed multiple
after Miyawaki (2004). broad bands centered at 530 and 610 nm and a weak
narrow band at 437 nm (figure 13D). This absorption
feature, related to Mn3+, Ti4+-Fe2+ pairs, and Fe3+, gen-
jadeite from Itoigawa (see the bottom table at erally overlapped with the bands observed in Itoigawa
https://www.gia.edu/gems-gemology/spring-2017- lavender jadeite. The concentrations of V, Cr, and Co
japanese-jadeite-tables). were too low to create any noticeable color.
Myanmar. The UV-Vis spectrum of the burmese green Russia. A highly saturated vivid green jadeite from
jadeite (K-MYA-16) showed the characteristic narrow the Polar Urals showed an Fe3+ band and strong mul-
Cr3+-related absorption band at 691 nm and the sharp, tiple chromium lines in the 580–700 nm range, a
UV-VIS SPECTRUM
V Co Mn Cr Fe Ti spectrum of a blue
2.5 Blue jadeite NaAlSi206 (ppmw) 41 19 175 nd 11900 4520 sample from Itoigawa
(ppma) 16 6.5 64 nd 4212 1943
(K-IT-JP-16). The ab-
OEFFICIENT (cm–1)
sorption shows a very

wide broad band from
2
500 to 750 nm that
overlaps the Mn-related
broad bands at 530 and
570 nm observed in
ABSORBANCE CO
1.5 Ti4+-Fe2+ Burmese lavender

Fe3+
jadeite. The chro-
mophores Ti and Fe
show a significant con-
1
centration at 1943 and
K-IT
T-JP-16,
-JP-16, 1.74 mm thick
4212 ppma, respec-
tively, and this jadeite’s
blue color could be
0.5
350 400 450 500 550 600 650 700 750 800
mainly due to the Ti4+-
Fe2+ charge transfer.
WAVELENGTH (nm)
WA

5
COEFFICIENT (cm–1) A Green: Kachin 2.9 B Lavender (purple): Kachin
COEFFICIENT (cm–1)
4.5
2.7
ABSORBANCE
ABSORBANCE
4.0 2.5 Fe3+ Mn3+
Fe3+
3+ 2.3
3.5 Cr3+ Cr Cr3+
2.1
3.0
1.9
2.5
1.7
2.0 1.5
350 400 450 500 550 600 650 700 750 800 350 400 450 500 550 600 650 700 750 800
WAVELENGTH (nm) WAVELENGTH (nm)
1.2 1.0
C Grayish green: Motagua D Lavender (violet): Motagua
1.1
0.9
ABSORBANCE
ABSORBANCE
1.0
0.9 0.8
Fe3+
0.8 0.7 Fe3+ Mn3+
Ti4+-Fe2+
0.7
0.6
0.6
0.5
0.5
0.4 0.4
350 400 450 500 550 600 650 700 750 800 350 400 450 500 550 600 650 700 750 800
3.0
E Vivid green: Polar Urals
2.5
ABSORBANCE
Fe3+
2.0
Cr3+
1.5
1.0
0.5
350 400 450 500 550 600 650 700 750 800
WAVELENGTH (nm)
Figure 13. A: The UV-Vis spectrum of a green Burmese jadeite (K-MYA-16) shows the characteristic chromium
lines at 630, 650, and 691 nm and the sharp, narrow Fe3+ absorption band at 437 nm that is commonly seen in
natural green jadeite. The Cr3+- and Fe3+-related feature generally overlaps with the spectrum of Japanese green
jadeite, but the absorption intensity is much higher due to its color saturation and transparency. B: Burmese
lavender jadeite (K-MYA-20) showed a dominant broad absorption band centered at 570 nm, related to Mn3+
concentration. C: The narrow Fe3+ absorption band at 437 nm, often present in Guatemalan green jadeite (M-
GUA-02). The absorption of Cr3+ is not detectable in this 2.32-mm-thick sample. D: The spectrum of
Guatemalan lavender jadeite (M-GUA-03) shows multiple broad bands centered at 530 and 610 nm and a
weak narrow band at 437 nm. The absorption feature related to Mn3+, Ti4+-Fe2+, and Fe3+ generally overlaps with
the bands observed in Itoigawa lavender jadeite. E: A vivid green Polar Ural jadeite shows an Fe3+ band and
strong multiple chromium lines in the 580–700 nm range, a combination that typically produces a highly satu-
rated green color. The concentration of Cr (maximum of 3042 ppma) is much higher than in Japanese green
jadeite.

Ca(Mg,Fe2+)Si2O6
XJd= 0%
diopside, augite,
from any other jade color (0.26, 0.12, and 0.44 wt.%,
hedenbergite respectively). Values for Cr2O3, MnO, K2O, and NiO
CaAl2SiO6
XJd= 20% were below the detection limit of the analysis. TiO2
(0.03 wt.%) was lower than values analyzed from vi-
olet and blue jadeite elsewhere. The white jadeite
was very pure.
LA-ICP-MS analyses of the white jadeite consis-
aegirine-
omphacite
augite
tently identified 19 minor and trace elements (Li,
Mg, K, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Cu, Sc, Ni, Zn,
XJd= 80% Ga, Se, Sr, and Zr). Other trace elements (b, Rb, Y,
jadeite aegirine Nb, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, Hf, Ta,
XJd= 100% W, Th, and U) were above the detection limits. Al-
J
NaAlSi2O6 NaFe3+Si2O6 though Itoigawa “white” jadeite generally had
lower Mg and Ca contents (3841 and 8495 ppmw,
respectively) than green, blue, and black jadeite, al-
Figure 14. This classification of pyroxene based on most all of the detectable minor and trace element
chemical composition shows the relationship be- contents were higher than in white jadeite from
tween jadeite and other pyroxenes in the main iso- Kachin State in Myanmar (see the bottom table at
morphous substitution. Source: National Museum of https://www.gia.edu/gems-gemology/spring-2017-
Nature and Science, Tokyo. japanese-jadeite-tables).
Green Jadeite. Microprobe analyses of four green

combination that typically produces a highly satu- jadeites revealed significant Fe content, from a mini-
rated green color (figure 13E). The Cr concentration mum value of 0.22 wt.% to a maximum of 0.864
(up to 3042 ppma) was much higher than in Japanese wt.%, and a slightly low Cr value of 0.01–0.57 wt.%.
green jadeite. Values for MgO (0.16–2.83 wt.%) and CaO (0.24–4.18
wt.%) were relatively high, but the compositions fit
Chemical Analysis. Quantitative chemical data col- within the jadeite range of XJd = 98.7 to 82.4 (figure
lected by EPMA for the samples from Japan, 15). Samples from Itoigawa-Omi showed slight differ-
Guatemala, and Russia are summarized in table 1. ences in major element composition between crystal
The results are described below for each representative aggregates and discrete single-mineral grains. This
color: white, green, lavender, and blue (including vio- study indicates that the green jadeite is also nearly
letish blue). XJd, X(Ae+Ko), and XQuad (Dio+Aug+Hed) were calcu- pure jadeite, though some discrete single-mineral
lated as mol.% of Al/(Na + Ca), Fe3+/(Na + Ca), and grains showed a composition closer to omphacite
Ca/(Na + Ca), respectively (figure 14). The highest and within the jadeite-dominant matrix.
lowest concentrations of jadeite (XJd) for 11 tested sam- The 13 green specimens from Itoigawa revealed the
ples are listed in table 1. The 87 specimens from the remarkable transport of large-ion lithophile elements
four countries in this study (Japan, Myanmar, in subduction zones, such as Li, b, K, Sr, and ba, as
Guatemala, and Russia) were analyzed by LA-ICP-MS, well as elements that are considered more refractory,
with averaged data calculated based on three to ten such as rare earth elements (La, Ce, Pr, Nd, Sm, Eu,
laser ablation spots for each specimen; minimum and Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu) and Hf, Ta, W, Tl,
maximum values of each element concentration are Pb, Th, and U. The Mg and Ca contents were also rel-
listed in the bottom table at https://www.gia.edu/ atively high, ranging from 2383 to 77,100 ppmw for
gems-gemology/spring-2017-japanese-jadeite-tables, Mg (averaged to 19,957 ppmw) and 4400 to 82,700
with averaged data in parentheses. ppmw for Ca (averaged to 39,206 ppmw). Mg and Ca
contents were much higher in the dark green areas,
White Jadeite. White jadeite from Itoigawa belongs which indicates that the dark green omphacite com-
to the clinopyroxene group and is close to the ideal ponent is more abundant in trace elements (except Li
jadeite composition (again, see table 1). All analyses and Ga) than jadeite. To establish a useful chemical
(more than five spots) were close to the end member fingerprint diagram for separating omphacite jade from
composition, up to XJd = 98 mol.%. The CaO, MgO, jadeite jade, we plotted two different combinations of
and FeOtot contents were lower than those tested major and minor elements. As seen in figure 16, plot-

Quad (Ca+Mg+Fe2+)2
similar ranges. Ti and Fe were dominant in Russian
and Guatemalan green jadeite. The Russian green
jadeite showed the highest Cr content (up to 7940
-1
4 ppmw, averaged to 2872 ppmw) of any samples.
1
JP
-1
12
IT-
JP
Lavender Jadeite. EPMA performed on a violet sam-

P-
3
K-
IT-
-1
J
ple from Itoigawa (K-IT-25) yielded significant TiO2

IT-
K-
JP
K-
IT-
(up to 0.362 wt.%) and FeOtot (up to 0.694 wt.%),

K-
whereas MnO was relatively low (up to 0.019 wt.%).

The color of the Japanese lavender jadeite likewise
0 0.2 0.4 0.6 0.8 1 should correlate to the chromophores Ti4+, Fe2+, and
Jd Ae+Ko
(NaAl) Na(Fe3+,Cr) Mn3+. The contents of MgO (up to 0.864 wt.%) and
CaO (up to 1.879 wt.%) were relatively low. The
jadeite composition ranged from XJd = 98.7 to 93.3,
Figure 15. This ternary diagram for jadeite (Jd)-aegirine
close to pure jadeite.
+ kosmochlor (Ae + Ko)-Ca-Fe-Mg pyroxene (diopside
+ augite + hedenbergite) indicates the chemical con-
LA-ICP-MS detected noticeably high amounts of
centration data of four green jadeites from Itoigawa- Ti and Fe in all of the violet jadeite. Other metal ele-
Omi by EPMA, based on Morimoto et al. (1988). ments such as Li, b, K, Sr, and ba, as well as REEs,
Their compositions fit the jadeite range of XJd = 98.7 were higher than in white or green jadeite from the
to 82.4. Three to 12 spots were tested on the green same geological source in Itoigawa-Omi. Lavender
area of each specimen. jadeite from Guatemala showed a similar violetish
hue, and trace element concentrations revealed sim-
ilarly high levels of Ti and REEs. but the transition
ting Al/Fe vs. Ca/Na clearly separates the jadeite ac- metal ions V, Cr, and Co were below detection limits,
cording to chemical concentration. consistent with its original light violet color
by comparison, the green samples from Itoigawa- (Sorensen et al., 2003). by comparison, 16 burmese
Omi showed higher Li, b, Mg, K, Ca, Ni, and Sr than lavender samples showed appreciable Mn, explaining
green jadeite from Myanmar, while the transition their dominant pinkish/purplish component. Ober-
metal elements Ti, V, Cr, Mn, Fe, and Co showed hänsli et al. (2007) did not observe a Ti phase in the
1
Figure 16. Chemical
fingerprinting of Al/Fe
Omp
vs. Ca/Na indicates
the separation of
jadeite (Jd) from om-
10
phacite (Omp) jade ac-
cording to chemical
Al/Fe
concentration. Al-
though the data plot-
ted here were collected
100
White jadeite Jd from LA-ICP-MS
Green jadeite analysis, this type of
Lavender jadeite diagram can also be
Blue jadeite adapted to EDXRF or
Black jadeite
electron microprobe
1000 data.
10 1 0.1 0.01
Ca/Na

TABLE 1. Electron microprobe analyses of major element compositions of jadeite jade from Japan, Guatemala, and Russia.
Itoigawa Itoigawa Itoigawa Itoigawa Itoigawa Itoigawa
K-IT-JP-14 (5 spots) K-IT-JP-14 (12 spots) K-IT-JP-11 (3 spots) K-IT-JP-12 (3 spots) K-IT-JP-13 (3 spots) K-IT-JP-25 (6 spots)
White Green Green Green Green Lavender (Violet)
Oxides (wt.%) Averaged Max XJd% Min XJd% Max XJd% Min XJd% Max XJd% Min XJd% Max XJd% Min XJd% Max XJd% Min XJd%
SiO2 58.82 60.466 58.95 58.43 57.83 59.25 58.4 59.31 57.11 59.542 59.621
TiO2 0.03 0 0.02 0.012 0 0 0 0 0.012 0.362 0.325
Al2O3 24.3 24.822 24.3 22.38 20.59 21.87 20.83 21.98 21.27 24.518 22.997
Cr2O3 0 0.036 0.09 0.16 0.085 0.013 0.01 0.152 0.577 0.083 0.008
FeO 0.44 0.225 0.45 0.561 0.864 0.744 0.744 0.425 0.481 0.105 0.694
MnO 0 0.013 0.01 0.033 0 0.02 0.022 0.007 0.008 0.019 0
MgO 0.12 0.16 0.54 1.8 2.83 1.83 2.78 2.2 2.41 0.064 0.864
CaO 0.26 0.24 0.92 2.7 4.18 3.02 4.15 3.8 5.1 0.644 1.879
Na2O 15.49 14.113 14.91 13.94 12.77 13.55 13.03 13.22 11.88 14.278 13.31
K2O 0 0 0 0.01 0 0 0 0 0 0.001 0
NiO 0 0 0 0 0.032 0 0.025 0.022 0.025 0.011 0
Total 99.46 100.075 100.2 100.025 99.181 100.296 99.992 101.116 98.873 99.627 99.895
Cations (O=6)
Si 2 2.026 1.991 1.989 1.994 2.009 1.996 1.999 1.975 2.001 2.02
Ti 0.001 0 0 0 0 0 0 0 0 0.008 0.008
Al 0.974 0.98 0.967 0.898 0.837 0.874 0.839 0.872 0.867 0.991 0.919
Cr 0 0.001 0.003 0.004 0.002 0 0 0.004 0.016 0 0
Fe* 0.013 0.006 0.013 0.016 0.024 0.02 0.021 0.012 0.014 0.002 0.019
Mn 0 0 0 0.001 0 0.001 0.001 0 0 0 0
Mg 0.006 0.008 0.027 0.091 0.145 0.093 0.142 0.11 0.124 0.001 0.04
Ca 0.009 0.009 0.033 0.098 0.154 0.11 0.152 0.137 0.189 0.023 0.068
Na 1.021 0.917 0.977 0.92 0.854 0.891 0.864 0.863 0.796 0.932 0.875
K 0 0 0 0 0 0 0 0 0 0.003 0.008
Ni 0 0 0 0 0.001 0 0.001 0.001 0 0 0
Total 4.023 3.947 4.012 4.019 4.013 3.999 4.016 3.997 3.982 3.961 3.965
End members (mol%) after Morimoto et al. (1988)
XJd 98 98.7 95.5 88.5 82.6 88.9 83.3 86.6 82.4 98.7 93.3
X(Ae+Ko) 1.2 0.1 1.5 2 2.3 1.7 2.1 0.4 1.6 0 0
XQuad 0.8 1.2 3 9.5 15.1 10.3 14.6 13 16 1.3 6.7
high-pressure assemblage, which may contain the Chondrite-Normalized REE and Primitive Mantle–
same accessory phases (amphibole, feldspar, and law- Normalized Heavy Trace Element Pattern. To com-
sonite) as jadeite from Itoigawa-Omi. pare trace element compositions in different colors of
Japanese jadeite, we studied their chondrite-normal-
Blue Jadeite. The six blue samples from Itoigawa and ized REE patterns and primitive mantle–normalized
the two violetish blue specimens from Wakasa had trace element patterns. Figures 17 and 18 show the
significantly high TiO2, with maximum values of averaged REE and heavy trace element data of white,
0.649 and 0.745 wt.%, respectively. These corre- black, green, lavender, and blue jadeite from Itoigawa-
sponded with the intense blue area. The CaO con- Omi, along with two violetish blue specimens from
tents were slightly higher in the light blue to blue Wakasa. The REEs in Japanese jadeite tended to be
areas (0.6% to 1.4 wt.%) than in the white areas. Vi- more abundant in lavender, violetish blue, and blue
oletish blue and blue areas revealed the highest con- specimens than in green, white, and black jadeite. In
centration of Ti measured in Japanese jadeite (up to all colors, the light rare earth element (LREE: La, Ce,
4520 ppmw in blue jadeite from Itoigawa, and up to Nd, and Sm) contents tended to be higher than the
3636 ppmw in violetish blue jadeite from Wakasa), as heavy rare earth element (HREE: Eu, Gd, Dy, Y, Er,
well as enriched Fe (up to 11,900 ppmw). Most of the Yb, and Lu) contents. From this chondrite-normalized
REE contents were higher than in lavender jadeite. REE pattern, the lavender, violetish blue, and blue

Itoigawa Wakasa Matuaga Matuaga Polar Urals
K-IT-JP-16 (6 spots) W-TO-JP-02 (6 spots) M-GUA-03 (5 spots) M-GUA-02 (8 spots) PU-RUS-03 (11 spots)
Blue Violetish blue Lavender (Violet) Grayish green Green
Max XJd% Min XJd% Max XJd% Min XJd% Max XJd% Min XJd% Max XJd% Min XJd% Max XJd% Min XJd%
60.728 59.768 58.041 58.899 61.733 58.149 60.898 59.518 58.789 58.518
0.106 0.649 0.745 0.173 0.04 0.07 0.28 0.05 0.073 0.106
24.872 22.995 22.867 24.731 22.927 24.614 23.672 19.48 20.482 19.517
0.001 0.052 0 0.048 0 0.005 0 0 0.385 0.139
0.214 0.445 0.615 0.6 0 0.187 0.836 1.891 0.725 0.761
0.026 0 0.051 0.039 0 0.006 0.039 0.013 0.071 0.019
0.478 1.105 1.532 0.225 0.057 0.015 0.314 2.354 2.639 3.441
0.629 1.436 2.24 0.659 0.108 0.231 0.634 3.718 3.786 4.801
14.113 13.767 13.02 13.43 13.097 14.022 14.595 11.876 12.515 11.347
0.027 0.011 0 0.036 0.149 0.01 0 0 0.01 0.032
0.024 0.006 0 0 0 0.007 0.039 0.012 0.088 0.03
101.218 100.234 99.111 98.84 98.108 97.316 101.307 98.912 99.563 98.711
2.019 2.034 1.997 2.004 2.09 2.005 2.031 2.071 2.038 2.05
0.002 0.006 0.019 0.004 0.001 0.002 0.007 0.001 0.002 0.003
0.974 0.916 0.927 0.992 0.915 1 0.93 0.799 0.837 0.806
0 0.001 0 0.001 0 0.001 0 0 0.011 0.004
0.006 0.013 0.018 0.017 0 0.005 0.023 0.057 0.021 0.022
0.001 0 0.001 0.001 0 0 0.001 0 0.002 0.001
0.024 0.056 0.078 0.011 0.003 0.001 0.016 0.122 0.136 0.18
0.022 0.052 0.083 0.024 0.004 0.009 0.023 0.139 0.141 0.18
0.909 0.902 0.869 0.886 0.86 0.938 0.944 0.801 0.841 0.771
0.001 0 0 0.002 0.006 0 0 0 0 0.001
0.001 0 0 0 0 0 0.001 0 0.002 0.001
3.958 3.981 3.993 3.944 3.879 3.962 3.976 3.99 4.031 4.019
97.4 93.7 91.2 97.3 99.6 99.2 96.1 83.4 84 80.5

0 1 0 0.1 0 0.1 1.4 0.3 1.1 0.4
2.6 6.2 8.8 2.6 0.4 0.7 2.5 16.3 14.9 19.1
jadeite from Japan can be characterized by a high the fluid related to the formation of Itoigawa-Omi
LREE/HREE ratio and a low Eu concentration relative jadeite in the subduction zone was uniquely enriched
to other REE (again, see figure 17). in both LILE1 and HFSE2 brought in by the fluids in
Interestingly, the primitive mantle–normalized the subduction zone, and that these elements were re-
trace element patterns of all colors of Japanese jadeite cycled into serpentinized peridotites.
showed strong positive anomalies of the large-ion by comparison, Russian white and green jadeite re-
lithophile elements (LILE) Sr and ba, as well as the vealed the highest REE and heavy trace element values
high field strength elements (HFSE) Zr and Nb. Green in this study. Green jadeite from Japan and Myanmar
jadeite REE patterns were more depleted, but much showed a very close overlap, whereas the REE and
higher than white and black jadeite, with strong pos- heavy element contents for white samples from Myan-
itive anomalies of Sr, Zr, and Hf. This result is consis- mar were very depleted and had the lowest value.
tent with the conclusion by Morishita et al. (2007) that In lavender and blue samples, Guatemalan jadeite
1
LILE refers to lithophile trace elements (K, Rb, Cs, Sr, Ba, Pb, and Eu), which have an ionic radius to charge ratio that is greater than those of
Ca2+ and Na+, the largest cations common to rock-forming minerals.
2
HFSE refers to high field strength elements (Ti, Zr, Hf, Nb, and Ta), which do not have a large ionic radius. Because of their high charge and the
consequent difficulty in achieving a charge balance, they are typically incompatible.

100
JP-white
JP-green
JP-lavender
JP-blue
10
JP-violetish blue
JP-black
Detection limit (ppmw)
1 Figure 17. Chondrite-

normalized rare earth
element (REE) patterns
are shown for each
0.1 color of Japanese
jadeite.
0.01
0.001
La Ce Nd Sm Eu Gd Dy (Y) Er Yb Lu
was characterized by the highest REE and heavy trace was mainly white, with unevenly distributed
element concentrations. Japanese material showed a pale green to green and lavender to blue color.
lower value but could be separated from burmese 2. Petrographic observations showed that Japan-
lavender jadeite based on the chondrite-normalized ese jadeite was composed of aggregates of long,
and primitive mantle–normalized patterns. semi-euhedral prismatic crystals and granular
single crystals, which combined to produce a
CONCLUSIONS prismatic-granular crystalloblastic texture. Pec-
Japanese jadeite from the Itoigawa-Omi region is tolite, prehnite, and analcime were often pres-
characterized by mixtures of white with green and ent in folding, faults, and veinlets, while the
other colors such as lavender, blue, and black. Al- minor component minerals vesuvianite and ti-
though jadeite mining has been prohibited there tanite were found in the matrix.
since 1954, small pebbles can be found along the 3. Quantitative analysis by electron microprobe
rivers or estuaries. In this study, a large number of showed that the white jadeite was close to pure
samples from Itoigawa-Omi and Wakasa were ana- jadeite (XJd = 98). Green jadeite was in the range
lyzed to characterize the chromophores, optical ab- of XJd =98–82 and XAeg = 2–8, and the chro-
sorption features, and quantitative chemical mophores Fe and Cr were responsible for the
composition of major and trace elements in each green color. Lavender color was produced by a
color variety. This gemological, petrographic, and combination of higher Ti and Fe and lower Mn.
chemical study also included burmese, Guatemalan, In blue jadeite, the Ti4+-Fe2+ charge-transfer
and Russian jadeite samples for comparison. The re- played a significant role in coloration.
sults are summarized below.
4. LA-ICP-MS analysis detected 19 minor and
1. Jadeite boulders and pebbles from Itoigawa- trace elements. The chondrite-normalized rare
Omi tended to have rounded corners from flu- earth element and primitive mantle–normal-
vial erosion and showed a glittering whitish ized heavy trace element patterns of all colored
surface, but the rough rock did not show a jadeite showed higher light REE than heavy
weathered brown skin. The Japanese jadeite REE, along with positive large-ion lithophile el-

10000 JP-white
JP-green
JP-lavender
1000 JP-blue
JP-violetish blue
JP-black
100 Detection limit (ppmw)
Figure 18. Primitive

10 mantle–normalized
heavy trace element
patterns are shown for
1 each color of Japanese
jadeite.
0.1
0.01
0.001
Rb Sr Y Zr Nb Sn Ba La Ce Pr Nd Sm Eu Gd Er Yb Lu Hf Ta W U
ements and high field strength element anom- by high Ti and Fe and low Mn. Investigation of thin
alies. Lavender and blue (including violetish sections revealed that burmese jadeite contains kos-
blue) jadeite had dominant REE compared to mochlor, amphibole, albite, nepheline, and vesuvian-
green jadeite, whereas white and black jadeite ite. Russian jade has the very common mineral
had the lowest REE contents. magnetite in matrix and analcime in veinlets.
Guatemalan jadeite was characterized by grossular
Our studies confirmed that green jadeite from garnet, albite, rutile, and other minerals. While
Itoigawa, Myanmar, and Russia have similar gemo- Guatemalan jadeite contained abundant REEs and
logical properties such as RI and SG and absorption heavy trace elements, burmese jade had much lower
spectra of Cr and Fe, while Guatemalan grayish green values. Japanese jadeite shows characteristic trace el-
jadeite does not contain Cr. Lavender jadeite from ements and matrix inclusion varieties that may be
Japan and Guatemala showed similar color, caused useful in establishing the precise country of origin.
ABOUT THE AUTHORS Co., Tokyo), and Prof. Miyajima (Fossa Magna Museum,
Dr. Abduriyim is president of Tokyo Gem Science, LLC and di- Itoigawa) for information and specimens that made this study
rector of GSTV Gemological Laboratory. He is a former senior possible. We sincerely thank Dr. Mikouchi, associate professor
manager and senior scientist of GIA’s laboratory in Tokyo. Dr. at the University of Tokyo, and Prof. Ogasawara and Dr. Saka-
Saruwatari and Dr. Katsurada are scientists and staff gemolo- maki at Waseda University for providing EPMA testing. We are
gists at GIA’s laboratory in Tokyo. also grateful to GIA colleagues Dr. Shoko Odake and Dr.
Supharart Sangsawong, who assisted with data collection. The
ACKNOWLEDGMENTS authors wish to thank Dr. F. Lin Sutherland, a former curator
The authors thank Nobuyuki Tsurumi (The Jade Ore and research scientist at the Australian Museum in Sydney, for
Museum/Hisui Gensekikan, Tokyo), Minoru Kameyama (Miyuki his critique of the manuscript.

REFERENCES
barnes G.L. (1999) The Rise of Civilization in East Asia: The Ar- Miyazoe T., Nishiyama T., Uyeta K., Miyazaki K., Mori Y. (2009)
chaeology of China, Korea and Japan. Thames and Hudson, Coexistence of pyroxenes jadeite, omphacite, and diopside-
New York. hedenbergite rock from a serpentinite mélange in the
burns R.G. (1981) Intervalence transitions in mixed-valence min- Kurosegawa zone of central Kyushu, Japan. American Miner-
erals of iron and titanium. Annual Review of Earth and Plan- alogist, Vol. 94, No. 1, pp. 34–40.
etary Sciences, Vol. 9, No. 1, pp. 345–383, http://dx.doi.org/ Morimoto N., Fabries J., Ferguson A.K., Ginzburg I.V., Ross M.,
10.1146/annurev.ea.09.050181.002021 Seifert F.A, Zussman J., Aoki K., Gottardi G. (1988) Nomencla-
Chihara K. (1971) Mineralogy and paragenesis of jadeites from the ture of pyroxenes. American Mineralogist, Vol. 73, pp. 1123–
Omi-Kotaki area, Central Japan. Mineralogical Society of 1133.
Japan, Special Paper, No. 1, pp. 147–156. Morishita T. (2005) Occurrence and chemical composition of bar-
Chihara K. (1991) Jade in Japan. In R. Keverne, Ed., Jade. Van Nos- ian feldspar in a jadeitite from the Itoigawa-Ohmi district in
trand Reinhold, New York, pp. 216–217. the Renge high-P/T-type metamorphic belt, Japan. Mineralog-
Essene E.J. (1967) An occurrence of cymrite in the Franciscan for- ical Magazine, Vol. 69, No. 1, pp. 39–51, http://dx.doi.org/
mation, California. American Mineralogist, Vol. 52, pp. 1885– 10.1180/0026461056910237
1890. Morishita T., Arai S., Ishida Y. (2007) Trace element compositions
Ferguson J., Fielding P.E. (1971) The origins of the colours of yel- of jadeite (+omphacite) in jadeitites from the Itoigawa-Ohmi
low, green, and blue sapphires. Chemical Physics Letters, Vol. district, Japan: Implications for fluid processes in subduction
10, No. 3, pp. 262–265, http://dx.doi.org/10.1016/0009- zones. Island Arc, Vol. 16, No. 1, pp. 40–56, http://dx.doi.org/
2614(71)80282-8 10.1111/j.1440-1738.2007.00557.x
Foshag W.F., Leslie R. (1955) Jadeite from Manzanal, Guatemala. Nakamizu M., Okada M., Yamazaki T., Komatsu M. (1989) Meta-
American Antiquity, Vol. 21, No. 1, pp. 81–83, http://dx.doi.org/ morphic rocks in the Omi-Renge serpentinite mélange, Hida
10.2307/276111 Marginal Tectonic belt, Central Japan. Memoirs of the Geolog-
Fukuyama M., Ogasawara M., Horie K., Lee D.C. (2013) Genesis ical Society of Japan, Vol. 33, pp. 21–35 (in Japanese with Eng-
of jadeite-quartz rocks in the Yorii area of the Kanto Moun- lish abstract).
tains, Japan. Journal of Asian Earth Sciences, Vol. 63, pp. 206– Nishimura Y. (1998) Geotectonic subdivision and areal extent of
217, http://dx.doi.org/10.1016/j.jseaes.2012.10.031 the Sangun belt, inner zone of southwest Japan. Journal of
Harlow G.E., Olds E.P. (1987) Observations on terrestrial ureyite Metamorphic Geology, Vol. 16, No. 1, pp. 129–140,
and ureyitic pyroxene. American Mineralogist, Vol. 72, pp. http://dx.doi.org/10.1111/j.1525-1314.1998.00059.x
126–136. Oberhänsli R., bousquet R., Moinzadeh H., Moazzen M., Arvin
Harlow G.E., Sorensen S.S. (2005) Jade (nephrite and jadeitite) and M. (2007) The field of stability of blue jadeite: A new occur-
serpentinite: Metasomatic connections. International Geology rence of jadeitite at Sorkhan, Iran, as a case study. Canadian
Review, Vol. 47, No. 2, pp. 113–146, http://dx.doi.org/ Mineralogist, Vol. 45, No. 6, pp. 1501–1509, https://doi.org/
10.2747/0020-6814.47.2.113 10.3749/canmin.45.6.1501
Kanmera K., Hashimoto M., Matsuda T. (1980) Geology of Japan. Ohmori K. (1939) Optical properties of jade (jadeite) newly oc-
Vol. 15 of Iwanami Koza Chikyukagaku series. Iwanami curred in Japan. Journal of the Japanese Association of Miner-
Shoten, Tokyo, 387 pp. (in Japanese). alogists, Petrologists and Economic Geologists, No. 22, pp.
Kawano Y. (1939) A new occurrence of jade (jadeite) in Japan and 201–212 (in Japanese).
its chemical properties. Journal of the Japanese Association of Rossman G.R. (1974) Lavender jade. The optical spectrum of Fe3+
Mineralogists, Petrologies and Economic Geologists, No. 22, and Fe2+￫Fe3+ intervalence charge transfer in jadeite from Myan-
pp. 195–201 (in Japanese). mar. American Mineralogist, No. 59, pp. 868–870.
Kunugiza K., Nakamura E., Miyajima H., Goto A., Kobayashi K. Shibata K., Nozawa T. (1968) K-Ar age of Omi schist, Hida Moun-
(2002) Formation of jadeite-natrolite rocks in the Itoigawa- tains, Japan. Bulletin of the Geological Survey of Japan, No. 19,
Ohmi area of the Hida marginal belt inferred from U-Pb-zircon pp. 243–246.
dating. Abstract of annual meeting of the Japanese Association Shinno I., Oba T. (1992) Absorption and photo-luminescence
of Mineralogists, Petrologists and Economic Geologists in spectra of Ti3+ and Fe3+ in jadeites. Mineralogical Journal, Vol.
2002, GP20 (in Japanese) 16, No. 7, pp. 378–386, http://dx.doi.org/10.2465/minerj.
Lu R. (2012) Color origin of lavender jadeite: An alternative ap- 16.378
proach. G&G, Vol. 48, No. 4, pp. 273–283, http://dx.doi.org/ Shoji T., Kobayashi S. (1988) Fluid inclusions found in jadeite and
10.5741/GEMS.48.4.273 stronalsite, and a comment on the jadeite-analcime boundary.
Miyajima H., Matsubara S., Miyawaki R., Ito K. (1999) Itoigawaite, Journal of Mineralogy, Petrology and Economic Geology, Vol.
a new mineral, the Sr analogue of lawsonite, in jadeite from the 83, No. 1, pp. 1–8, http://dx.doi.org/10.2465/ganko.83.1
Itoigawa-Ohmi district, central Japan. Mineralogical Magazine, Sorensen S.S., Harlow G.E., Rumble D. III. (2003) SIMS oxygen iso-
Vol. 63, No. 6, pp. 909–916, http://dx.doi.org/10.1180/ tope analyses of jadeitite: trace element correlations, fluid com-
002646199548899 positions, and temperature estimates. Geological Society of
Miyajima H., Matsubara S., Miyawaki R., Yokoyama K., Hirokawa America Abstract with Program, Vol. 35, No. 6, pp. 90–97.
K. (2001) Rengeite, Sr4ZrTi4Si4O22, a new mineral, the Sr-Zr ana- Takayama M. (1986) Mode of occurrence and significance of
logue of perrierite from Itoigawa-Ohmi district, Niigata Prefec- jadeite in the Kamuikotan metamorphic rocks, Hokkaido,
ture, central Japan. Mineralogical Magazine, Vol. 65, No. 1, pp. Japan. Journal of Metamorphic Geology, Vol. 4, No. 4, pp. 445–
111–120, http://dx.doi.org/10.1180/002646101550163 454, http://dx.doi.org/10.1111/j.1525-1314.1986.tb00363.x
Miyajima H., Miyawaki R., Ito K. (2002) Matsubaraite, Taube K.A. (2004) Olmec art at Dumbarton Oaks. Dumbarton
Sr4Ti5(Si2O7)2O8, a new mineral, the Sr-Ti analogue of perrierite Oaks Research Library and Collection, Washington, D.C.
in jadeitite from the Itoigawa-Ohmi district, Niigata Prefecture, Tsujimori T. (2002) Prograde and retrograde P-T paths of the Late
Japan. European Journal of Mineralogy, Vol. 14, No. 6, pp. Paleozoic glaucophane eclogite from the Renge metamorphic
1119–1128, http://dx.doi.org/10.1127/0935-1221/2002/0014- belt, Hida Mountains, southwestern Japan. International Ge-
1119 ology Review, Vol. 44, No. 9, pp. 797–818, http://dx.doi.org/
Miyawaki R. (2004) Main minerals of jadeite jade. Gemmology, 10.2747/0020-6814.44.9.797
Vol. 35, No. 420, pp. 14–15 (in Japanese). Tsutsumi Y., Yokoyama K., Miyawaki R., Matsubara S., Terada K.,

Hidaka H. (2010) Ages of zircons in jadeitite and jadeite-bearing Hirobunkan, Tokyo (in Japanese), 354 pp.
rocks of Japanese islands. Bulletin of the National Museum of Yokoyama K., Sameshima T. (1982) Miscibility gap between
Nature and Science, Vol. 36, pp. 19–30. jadeite and omphacite. Mineralogical Journal, Vol. 11, pp. 53–
Umehara S. (1971) Japan Ancient Jade Memorandum. Yoshikawa 61, http://dx.doi.org/10.2465/minerj.11.53
ADDITIONAL READING
Abduriyim A., Kitawaki H. (2006) Determination of the origin of Chemical Geology, Vol. 120, Nos. 3–4, pp. 223–253,
blue sapphire using laser ablation inductively coupled plasma http://dx.doi.org/10.1016/0009-2541(94)00140-4
mass spectrometry (LA-ICP-MS). The Journal of Gemmology, Miyajima H. (2014) Story of the Much-Valued Jadeite. Fossa
Vol. 30, No. 1/2, pp. 23–42. Magna Museum and Education Committee of Itoigawa,
Abduriyim A., Kitawaki H., Furuya M., Schwarz D. (2006) Itoigawa, 96 pp. (in Japanese).
“Paraiba”-type copper-bearing tourmaline from brazil, Nigeria, Oba T., Nakagawa Y., Kanayama K., Watanabe T. (1992) Notes on
and Mozambique: Chemical fingerprinting by LA-ICP-MS. rock-forming minerals in the Joetsu district, Niigata Prefecture,
G&G, Vol. 42, No. 1, pp. 4–21, http://dx.doi.org/10.5741/ Japan: (5) Lavender jadeite from the Kotaki river. Bulletin of
GEMS.42.1.4 Joetsu University of Education, No. 11, pp. 367–375.
blodgett T., Shen A.H. (2011) Application of discrimination analy- Ouyang Q. (2001) Characteristic of violet jadeite jade and its col-
sis in gemology: Country-of-origin separation in colored stones oration mechanism. Journal of Gems and Gemmology, Vol. 3,
and distinguishing HPHT-treated diamonds. G&G, Vol. 47, No. No. 1, pp. 1–6 (in Chinese).
2, p. 145. Rankin A.H., Greenwood J., Hargreaves D. (2003) Chemical fin-
breeding C.M., Shen A.H. (2010) LA-ICP-MS analysis as a tool for gerprinting of some East African gem rubies by laser ablation-
separating natural and synthetic malachite. News from Re- ICP-MS. The Journal of Gemmology, Vol. 28, No. 8, pp.
search, Oct. 11, http://www.gia.edu/gia-news-research-nr101410 473–482.
Fisher R.A. (1936) The use of multiple measurements in taxo- Saeseaw S., Pardieu V., Sangsawong S. (2014) Three-phase inclu-
nomic problems. Annals of Eugenics, Vol. 7, No. 2, pp. 179– sions in emeralds and their impact on origin determination.
188, http://dx.doi.org/10.1111/j.1469-1809.1936.tb02137.x. G&G, Vol. 50, No. 2, pp. 114–132, http://dx.doi.org/10.5741/
Harlow G.E., Shi G.H. (2011) An LA-ICP-MS study of lavender GEMS.50.2.114
jadeite from Myanmar, Guatemala, and Japan. G&G, Vol. 47, Saminpanya S., Manning D.A.C., Droop G.T.R., Henderson
No. 2, pp. 116–117. C.M.b. (2003) Trace elements in Thai gem corundums. The
Iwao S. (1953) Albitite and associated jadeite rock from Kotaki dis- Journal of Gemmology, Vol. 28, No. 7, pp. 399–415.
trict, Japan: A study in ceramic raw material. Bulletin of the Shen A.H., blodgett T., Shigley J.E. (2013) Country-of-origin deter-
Geological Survey of Japan. No. 153, pp. 1–26. mination of modern gem peridots from LA-ICP-MS trace-ele-
Kobayashi S., Miyake H., Shoji T. (1986) A jadeite rock from Oosa- ments chemistry and linear discriminant analysis (LDA).
cho, Okayama Prefecture, southwestern Japan. Mineralogical Geological Society of America Abstracts with Programs, Vol.
Journal, Vol. 13, No. 6, pp. 314–327, http://dx.doi.org/ 45, No. 7, p. 525.
10.2465/minerj.13.314 Sutherland F.L., Schwarz D., Jobbins E.A., Coenraads R.R., Webb
Luo Z.M, Yang M.X., Shen A.H. (2015) Origin determination of G. (1998) Distinctive gem corundum suites from discrete basalt
dolomite-related white nephrite through Ib-LDA. G&G, Vol. 51, fields: A comparative study of barrington, Australia, and West
No. 3, pp. 300–311, http://dx.doi.org/10.5741/GEMS.51.3.300 Pailin, Cambodia, gemfields. The Journal of Gemmology, Vol.
McDonough W.F., Sun S.S. (1995) The composition of the Earth. 26, No. 2, pp. 65–85.
For online access to all issues of GEMS & GEMOLOGY from 1934 to the present, visit:

FEATURE ARTICLES
CAUSES OF IRIDESCENCE IN
NATURAL QUARTZ
Xiayang Lin and Peter J. Heaney
Iris quartz specimens from geodes in the Deccan Trap basalts of west-central India have been described
since 1860. These quartz crystals exhibit dominant terminal faces, and iridescence is visible on only
the minor z {011} faces and not the major r {101} faces. For this study, we analyzed iris quartz crystals
from India’s Jalgaon District using scanning and transmission electron microscopy (SEM and TEM) and
atomic force microscopy (AFM). SEM and AFM imaging revealed that the iridescent z faces exhibit pe-
riodic ridges with wavelengths from 400 to 700 nm, but no surface topography was apparent on the
non-iridescent r faces. TEM examination of a section removed from a z face by focused ion beam (FIB)
milling suggests that the ridge and valley structures were produced by preferential etching of periodic
defects on the z faces, creating a diffraction grating. We interpret these defects as paired Brazil twin
boundaries.
I
ridescent solids exhibit a color change in re- BACKGROUND: PHYSICS OF IRIDESCENCE
sponse to variations in the lighting angle and Iridescence may be caused by the interference of light
viewing direction (figure 1). As exemplified by either by thin films or by repetitive substructures.
opal and labradorite, this phenomenon can trans-
form an otherwise lackluster mineral into a gem, Thin-Film Iridescence. Thin-film interference effects
and many techniques have been developed to gener- are seen when a liquid or solid is coated by a film that
ate a play of light in non-iridescent stones. Applica- is nanometers to micrometers in thickness. When in-
tions for iridescent materials have attracted
attention in diverse research fields, including pho-
tonics and computer graphics, while studies of the Figure 1. Iridescence is clearly visible on the minor z
physics of iridescence have informed the design of face of this iris quartz from Jalgaon, India. Photomi-
cosmetics, paints, and anti-counterfeiting devices crograph by Peter J. Heaney; field of view 4.5 mm.
(Parker and Townley, 2007; Vigneron et al., 2007).
Moreover, the functional significance of iridescence
in organic structures is an active area of exploration
in evolutionary and developmental biology (re-
viewed in Meadows et al., 2009). Here we describe
the two physical mechanisms by which iridescence
occurs in natural materials, and we report the results
of our investigations into the cause of iridescence in
naturally formed quartz crystals from India’s Jalgaon
District.

http://dx.doi.org/10.5741/GEMS.53.1.68
© 2017 Gemological Institute of America
68 IRIDESCENCE IN NATURAL QUARTZ GEMS & GEMOLOGY SPRING 2017

Incident Light Destructive Interference
Incident Light
Light Wave 1 Constructive Interference Light Wave 1
Light Wave 2
Light Wave 2
Air Air
Film Film
Substrate Substrate
Figure 2. Light rays that reflect from the upper boundary and the lower boundary of the thin film interfere con-
structively or destructively and form a new wave. Left: In this illustration of constructive interference, light waves
1 and 2 are in phase. Right: With destructive interference, the two waves are out of phase.
cident light strikes the thin film’s surface, a portion monochromatic, the interference pattern generated
of the light is transmitted through the interface and by the reflected light waves will appear as bright and
the remainder is reflected. After the transmitted light dark bands as a function of the viewing angle.
impinges the lower boundary of the film, some or all When the incident light is polychromatic, only
of that light wave may be reflected. The two waves one wavelength will be reinforced by constructive in-
reflected from the thin film’s upper and lower bound- terference for a given reflection angle. As a result,
thin-film interference is selective for specific wave-
lengths depending on the angle. Longer wavelengths
interfere at steeper angles according to Bragg’s law,
In Brief and thus a full rainbow spectrum is evident when the
• Iridescent quartz from India has been known since the film is viewed from a perspective that samples a
mid-19th century. range of reflection angles. In addition, the colors as-
• In this study, microscopic investigation of the iridescent sociated with the iridescence depend on the thick-
z {011} faces of two quartz samples revealed periodic ness of the thin film and on the relative refractive
ridges, which created diffraction gratings. indices (RIs) of the film and substrate. Refractive in-
• The ridge-and-valley structures were produced by the dices play a role because reflected light waves expe-
preferential etching of parallel defects. These defects rience a phase shift of 180° when the RI of the thin
are interpreted as paired Brazil twin boundaries.
film is lower than that of the substrate. Examples of
minerals showing thin-film iridescence include bor-
nite (Buckley and Woods, 1983; Vaughan et al., 1987)
aries interfere with each other to produce a new light and fire obsidian (Ma et al., 2001, 2007). Likewise,
wave (figure 2). If, after its extended travel path and many fractures in minerals, filled or unfilled, can dis-
possible phase shift, the ray that reflects off the thin play iridescence because of thin-film effects.
film/substrate boundary (ray 2 in figure 2, left) is ex-
actly in phase with the wave that reflects only from Iridescence from Diffraction Gratings. The disper-
the thin film surface (ray 1 in figure 2, left), the two sion of light can also be induced by interference ef-
reflected waves will constructively interfere and fects that arise from natural diffraction gratings
combine into a light wave with high intensity; oth- within or on the surfaces of minerals. The interac-
erwise, the emergent waves interfere destructively tion of light with the diffraction grating can be purely
and are canceled (figure 2, right). If the source is reflective from surface grooves (figure 3), or it can be
IRIDESCENCE IN NATURAL QUARTZ GEMS & GEMOLOGY SPRING 2017 69

Ƨi ƴ =ƴƧm
light angle changes, the diffracted wavelength also
Diffracted Beam
changes, generating iridescence. The equations that
Incident Beam Ƨi govern the conditions for transmissive diffraction
Ƨm through repetitive substructures, as with iris agate,
differ slightly from those that describe reflective iri-
descence, but they similarly define the conditions for
constructive interference of emergent light waves
(Kinoshita et al., 2008).
Mineralogical iridescence produced by intrinsic
ƴ
ƴ diffraction gratings is of particular interest because
the substructures that create the gratings often yield
insights into crystal growth processes. These in turn
d may allow geologists to quantify aspects of a min-
eral’s crystallization history, suggesting new path-
Figure 3. Schematic representation of a reflective dif- ways toward the self-assembly of hierarchically
fraction grating. Iridescence occurs when diffracted ordered structures. Four well-known examples of iri-
beams interfere constructively, as in figure 2, left. The descence produced by repetitive transparent sub-
dashed lines indicate the incident normals to the structures are precious opal (Sanders, 1964; Darragh
grooves and base of the diffraction grating. et al., 1966; Gaillou et al., 2008; Rondeau et al., 2010),
labradorite (Nissen, 1971; McConnell, 1974; Cliff et
al., 1976; Miura and Tomisako, 1978), moonstone
transmissive through transparent modular substru- (Akizuki and Sugawara, 1970), and iris agate (Taijing
tures with periodic variations in refractive indices and Sunagawa, 1994; Heaney and Davis, 1995).
(Kinoshita et al., 2008). When light reflects from a
blazed surface grating, as shown schematically in fig- IRIDESCENT QUARTZ
ure 3, waves that reflect from adjacent grooves will Naturally iridescent quartz has been recognized for
constructively interfere only when the emergent more than a century and is known by many names:
waves are in phase. When the reflected waves are out iris quartz, rainbow quartz, schiller quartz, anan-
of phase, destructive interference ensues. The crite- dalite, and adularescent quartz. Petrov and Tanaka
rion for constructive interference is met when the (2011) published online the only detailed review of
difference in the distances traveled by light waves the history of iridescent quartz. Although there is
scattered from adjacent grooves is equal to an integral some ambiguity regarding the source locality for this
number of wavelengths. material, most of the iridescent quartz on the market
In the construction shown in figure 3, the math- comes from India’s Deccan Traps in an area west of
ematical equation that governs diffraction from Burhanpur in Madhya Pradesh State, near the border
blazed gratings is dependent on the steepness of the with Maharashtra State. Indeed, the first known ref-
blaze angle between the inclined surface and the grat- erence to iridescent quartz is the 1860 donation of an
ing plane (labeled φ in figure 3), the distance between iris quartz cluster to the British Museum of Natural
grooves (d in figure 3), and the angles (θi and θm) that History by an engineer who worked on the construc-
the incident and reflected waves make with respect tion of the Central India Railway. Iridescent quartz
to the normals to the grating plane (Laufer, 1996). has also been reported from Washington State (USA),
The condition for constructive interference is math- Uruguay, Brazil, Germany, Turkey, and Morocco
ematically expressed as (Gübelin and Koivula, 2005; Petrov and Tanaka,
2011; Renfro and Koivula, 2011).
The iris quartz in the British Museum (number
sin (2q – ei) – sin ei = mh/d
55795) is described by vom Rath (1873) and Scharff
(1875), and their descriptions precisely match the
where m is an integer that represents the so-called characteristics of the specimens examined in the
diffraction order. For a given angle θi, only a single present study (figure 4). These samples occur as clus-
wavelength of light satisfies this equation for first- ters of euhedral quartz crystals that evidently filled
order diffraction (where m = 1). Thus, as the incident gaseous vugs (open spaces) within the Deccan Trap

{101} faces. The spectral colors vary from red to blue
depending on the angle of incident light and the
Euhedral
quartz
viewing direction. Etch pits are commonly located
on both the z and r faces, and iridescence is particu-
larly strong within the etch pits on the z faces. Some
crystals are broken, however, and rainbow colors are
not visible in broken portions of otherwise iridescent
z faces. The prismatic m {100} faces are poorly devel-
oped in these crystals. All of the single quartz crys-
tals observed on this sample measure about 0.5–1 cm
in length.
Chalcedony
Until about 2009, one Japanese company con-
trolled the market for iridescent quartz, importing
several tons of material. Since that time, production
1 cm Basalt from India has exceeded the importer’s capacity and
the market has opened internationally (Petrov and
Tanaka, 2011). Much of this quartz is of relatively
Figure 4. Iridescent quartz from Jalgaon, India, was low quality, but a small brilliant piece exhibiting
used in this study. Photo by Xiayang Lin. strong colors can easily cost up to $2,000. In light of
the growing market for iridescent quartz, specimens
have been produced artificially through the deposi-
basalts. At the base of these clusters is a rind of chal- tion of thin films of various metals (as with “aurora
cedony, 0.2–0.4 cm thick, that abruptly transitions quartz”). As a result, gem buyers have become skep-
to quartz crystals, as is characteristic of many agate tical of the authenticity of iridescent quartz from
geodes. The quartz crystals exhibit strongly ex- India. Nevertheless, Petrov and Tanaka (2011) re-
pressed terminal faces, and iridescence is visible only ported that the Japanese gemologist Koichi Iida
on the smaller z {011} faces and in zones closely sur- (Japanese Saitama Jewelry Institute) examined Indian
rounding them (figure 5), but not on the dominant r iris quartz and confirmed that the iridescence is nat-
Figure 5. The quartz crystals exhibit strongly expressed terminal faces, and iridescence is only visible on smaller z
{011} faces and not on r {101} faces. The sample measures approximately 5.0 × 4.0 × 3.5 cm. Photo by Xiayang Lin.
c-axis
z
r

c-axis Right-handed quartz
Left-handed quartz
Taijing and Sunagawa, 1990). Raman puzzled over the
absence of iridescence in Brazil-twinned amethyst and
recognized that the mere change in handedness at the
twin boundary was insufficient to generate a schiller
interference effect, as is seen in labradorite, since the
RIs of Brazil twins are identical. Consequently, he
speculated that the twin boundaries in the Indian iris
quartz were associated with “extremely thin layers of
impurity material.”
Raman was unable to confirm his deductions re-
garding the role that Brazil twins play in generating
iris effects in quartz. Even though Brazil twin bound-
z aries occur parallel to the {101} and {011} rhombohe-
dral planes of quartz, studies of amethyst have
r determined that a specific Brazil twin boundary is
never oriented parallel to a rhombohedral face with
the same indices—in other words, the (101) face of
quartz is not underlain by Brazil twin boundaries par-
Figure 6. A schematic representation of C.V. Raman’s
model for lamellar Brazil twinning in iridescent
allel to (101) (Balakirev et al., 1975). This observation
quartz. contradicts Raman’s model of lamellar twin bound-
aries parallel to the iridescent crystal faces (figure 6).
In the absence of a follow-up to Raman’s work, many
members of the gem community have presumed that
ural. In light of the historical record for this material, iridescence in these quartz crystals arises from a thin
the natural origin of iridescent quartz from the Dec- film of opal coating the surface (based on our discus-
can Traps of India seems irrefutable. sions with collectors).
C.V. Raman, the 1930 Nobel Laureate in physics, Using more sophisticated analytical techniques
published the only scientific analysis of iridescent than Raman had at his disposal, we sought to deter-
quartz we are aware of. Raman (1950) obtained spec- mine the cause of iridescence in quartz crystals from
imens from a jeweler in Bombay and examined them the Deccan Traps. For our investigation, we em-
using reflections of polarized light and visible light ployed a combination of scanning electron mi-
spectroscopy. He argued that the iridescence is not croscopy (SEM), atomic force microscopy (AFM),
the result of a superficial film coating the quartz sur- focused ion beam (FIB) milling, and transmission
face, as he observed the effect outside an iridescent electron microscopy (TEM).
face due to reflection within the crystal. Rather, he
attributed the iridescence to “a layer of material MATERIALS AND METHODS
which lies below the surface of the crystal and forms Specimens. We purchased two specimens of iris
an integral part of its structure. The observations in- quartz clusters from Georges Claeys (Geonic Min-
dicate the thickness of the layer to be of the order of eralen Collectie, Mariakerke, Belgium) at the 2014
a quarter of a millimeter.” Tucson Gem and Mineral Show. Figures 4 and 5
Based on his spectroscopic study, Raman con- show the same specimen from different angles. Mr.
cluded that the iridescence is caused by “a stratified Claeys reported that he obtained the specimens from
medium consisting of a great many parallel layers of a dealer who collected them from the Jalgaon District
extreme thinness,” and he estimated that each layer of India. The physical properties of the two speci-
was 0.34 μm thick (figure 6). He further hypothesized mens precisely match the descriptions of other iris
that the striations are regularly spaced polysynthetic quartz clusters from this locality, and we feel confi-
Brazil twins oriented parallel to the minor rhombohe- dent of its provenance.
dral z faces. Polysynthetic Brazil twins consist of al-
ternating layers of right- and left-handed quartz, and Scanning Electron Microscopy and Energy Dispersive
they are particularly common in amethyst (Schlössin Spectroscopy. All analytical testing was performed at
and Lang, 1965; McLaren et al., 1967; McLaren and Pennsylvania State University’s Materials Character-
Pitkethly, 1982; Baran et al., 1987; Schmetzer, 1987; ization Laboratory. For the first stage of our study, we

A Amorphous carbon strap B C
4 µm 5 µm 5 µm
Figure 7. SEM images of the iridescent surface during the focused ion beam (FIB) milling process. The foil was ex-
tracted from a z {011} face and oriented perpendicular to the ridges that compose the diffraction grating. An amor-
phous carbon strap was deposited to protect the surface texture (A), and trenches were excavated in front of and
behind the strap to generate the foil (B and C).
used an FEI Quanta 200 environmental scanning the average distance between two adjacent ridges was
electron microscope to map the surface topography calculated through a fast Fourier transform (FFT) al-
of both iridescent and non-iridescent pyramidal faces gorithm using MATLAB (MathWorks, Inc).
of iris quartz. Selected quartz crystals were pried
from the cluster, cleaned in methanol, and attached Focused Ion Beam Milling and Transmission Elec-
to a standard scanning electron microscope alu- tron Microscopy. To ascertain any relationship be-
minum mounting stub with double-sided carbon tween iridescent behavior and Brazil twinning, we
tape. Because quartz is not conductive and was not prepared an electron-transparent mineral section, or
coated by carbon, SEM images were taken at low vac- foil, for TEM using focused ion beam lift-out (Heaney
uum with an accelerating voltage of 20 kV. An Ox- et al., 2001). Focused ion beam milling was per-
ford Instruments INCA x-act (Model 51-ADD0001) formed with an FEI Helios NanoLab 660 FIB. A single
EDS detector on the microscope was used for surface quartz crystal was removed from a cluster and coated
chemical analysis. EDS data were analyzed using Ox- with conductive carbon paint to avoid charging.
ford Instruments’ AZtec nanoanalysis software (ver- After the sample was mounted within the FIB, an
sion 2.4). We used three different accelerating amorphous carbon strap was deposited over the area
voltages (20, 10, and 5 kv) to acquire spectra for the of extraction to protect the foil and preserve the sur-
same sites. face structure during the milling process (figure 7). A
Ga+ ion beam was used to excavate the material on
Atomic Force Microscopy. We next employed AFM both sides of the foil. The initial Ga beam voltage of
to construct high-resolution three-dimensional topo- 30 kV was reduced to 5 kV and then 2 kV for the final
graphic maps of the pyramidal faces. By rastering a thinning. Beam currents operated at 0.23 nA for the
probe across the sample’s surface, AFM can reveal sur- amorphous carbon deposition, 21 nA and 9.3 nA for
face structure at the nanoscale. We removed two intermediate milling stages, and finally 2.5 nA for
quartz single crystals from the cluster and oriented milling prior to foil lift-out.
one iridescent face and one non-iridescent face parallel Next, the foil was soldered to a glass probe tip and
to the flat stage in the AFM. A PeakForce Tapping deposited onto a V-shaped TEM half-grid post. Milling
model with ScanAsyst was used for these surface the sample on the grid began with a 30 kV ion beam
measurements. The peak force set point ranged from with the stage tilted to 53.5°. As an angle of 52° is nor-
2.5 to 7.5 nN for AFM imaging, and the scan rate was mal to the ion beam, this yielded an over-tilt of 1.5°.
1 to 0.5 Hz. The AFM probe used in these analyses The over-tilt was increased to 3.5° for 5 kV milling,
was a Bruker ScanAsyst-Air probe, which has a silicon and finally to 5.0° for 2 kV milling. The thickness of
tip on a nitride lever. The front angle of the tip was the final foil was less than 100 nm to allow electron
15°, and the back angle was 25°. The data were col- transparency, and the area of the quartz foil was ap-
lected as line scans with 512 points per line, with 256 proximately 7 μm × 4.5 μm. The entire milling and
lines collected in total. NanoScope Analysis software extraction process was monitored by secondary elec-
(version 1.50) was used to process the AFM data, and tron imaging. The milled sample was loaded in a

Unetched Surface
Etch Pit
z {011}
r {101}
500 µm
50 µm
Figure 9. An SEM image of an iridescent z face that

shows the boundary of an etch pit and an unetched
surface. The striations are continuous across the
boundary.
r {101}
descent r faces were apparent in SEM images of the
crystals (figure 8). In particular, a high-magnification
backscattered electron (BSE) image of an apparent
etch pit across both faces revealed that the iridescent
z {011}
z face features parallel ridge-and-valley structures
(figure 8, bottom), whereas the non-iridescent face
was relatively flat and smooth. The SEM images sug-
gested that the average distances between adjacent
ridges fell below a micron, but AFM analysis (see
below) allowed for more rigorous quantification. Al-
though the striations were more pronounced in the
20 µm apparent etch pits, an examination of the unetched
surfaces of the z faces also revealed a substructure
Figure 8. SEM images of the iridescent z {011} face
that consisted of alternating lamellae (figure 9).
and the non-iridescent r {101} faces. Top: A lower-
magnification view reveals the prevalence of etch pits
Compositional analyses on the iridescent faces
on both faces. The dashed line separates the z and r obtained by EDS offered little evidence of a thin-film
faces. Bottom: A higher-magnification image within coating that might be responsible for the iridescence
an etch pit clearly shows the parallel ridge-and-valley (figure 10). Trace amounts of Na, Al, and Ca were de-
topography of the z face and the absence of ridges on tected, but these elements were present in equal
the r face. amounts on the iridescent and non-iridescent faces.
Al commonly substitutes for Si in natural quartz
(Heaney, 1994).
AFM confirmed the existence of distinctly peri-
Philips double-tilt holder with a Be stage and exam- odic ridge-and-valley structures on the iridescent z
ined using a Philips 420 TEM at 120 kV. faces and the absence of such surface modulations on
the non-iridescent r faces (figure 11). The ridges were
RESULTS oriented parallel to the edge between the m and z
Surface Morphology and Composition. Topographic faces. We calculated an average distance between ad-
differences between the iridescent z and the non-iri- jacent ridges by processing the AFM images through

EDS SPECTRUM
8
Si Map sum spectrum 20 kv
COUNTS PER SECOND (eV)
7 Map sum spectrum 10 kv

Map sum spectrum 5 kv Figure 10. An EDS spec-
6
trum of an iridescent
5 face. Data were col-
4
lected at 20 kv (yellow),
O 10 kv (red), and 5 kv
3
(blue). Higher-voltage
2 spectra reflect chemical
composition at greater
1 Ca
C
K depths.
F Na Mg Al K Ca Ca
0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
ENERGY (keV)
an FFT algorithm using MATLAB. As the ridge shown in figure 13. The first few high-amplitude
shapes were not identical from top to bottom, we peaks were caused by signal leakage in the FFT and
sliced the images into 256 cross sections from the top
down (figures 12A and 12B), performing FFT on each
section (figure 12C), and stacked the 256 FFT results Figure 12. Fast Fourier transform (FFT) analysis of the
to extract the most dominant frequency of ridge os- spatial variation for an iridescent face as determined
cillation. The results of our FFT calculations are by AFM. A: An AFM image of an iridescent face. This
image provided 256 cross sections from top to bottom
for FFT analysis. B: The variation in topography for
Figure 11. Atomic force microscopy (AFM) images the cross section indicated by the blue line in A. C:
show the difference in topography between an irides- An FFT spectrum of B to calculate the most dominant
cent and a non-iridescent face. The height varies from frequency of the ridge oscillation.
500 to –500 nm on an iridescent face, and the valley-
and-ridge structure can be clearly observed. Non-iri- A
descent faces have a much narrower height range.
The bright spots are dirt on the surface, and the black
holes are pits.
507.2 nm
B
400
HEIGHT (nm)
300
–496.9 nm 200
99.99
0
-99.99
Iridescent face 2.0 µm -200
-300
1 2 3 4 5 6 7 8 9 10
17.2 nm
SPECTRAL FREQUENCY (1/μm)
AMPLITUDE (nm)
C
SPECTRAL RMS
70
60
50
40
30
20
–31.4 nm 10
5 10 15 20 25 30
Non-iridescent face 2.0 µm SPECTRAL FREQUENCY (1/μm)

FFT SPECTRUM
1,000,000 Figure 13. The final FFT
spectrum for the irides-
900,000
cent face in figure 12A.
STACKED RMS AMPLITUDE (nm)
800,000 The low-frequency

peaks are caused by
700,000
“signal leakage” and
600,000 should be ignored. The
500,000
most dominant fre-
Dominant frequency quencies that represent
400,000 the typical distances
300,000
between two adjacent
ridges on the z faces are
200,000
in the 1.59–2.29 μm–1
100,000 range. Therefore, the
wavelength of the oscil-
0
0.00 0.99 1.99 2.98 3.98 4.97 5.96 6.96 7.95 8.95 9.94
lation is on the order of
437 to 629 nm.
FREQUENCY (1/µm)
should be ignored. The most dominant frequencies order of 437 to 629 nm. These distances fall within
that reflect the wavelengths of the surface modula- the range of visible-light wavelengths, explaining the
tions ranged from 1.59 to 2.29 μm–1. Therefore, the effectiveness of the substructure as a diffraction grat-
real-space wavelength of the oscillation was on the ing for visible light. These values are impressively
close to Raman’s 1950 estimate of 0.34 μm (340 nm)
for the periodicity of the iris quartz striations.
Figure 14. This brightfield TEM image of the FIB sec-
tion was extracted from an iridescent face perpendi- TEM Characterization. Using FIB milling, we ex-
cular to the surface striations, which appear as the tracted a thin foil from an iridescent z face; the foil
jagged surface near the top of this cross-sectional foil
was oriented perpendicular to the grooved surface
beneath the carbon cap. The lighter striations indi-
striations (figure 7). Brightfield TEM imaging of the
cated by arrows are artifacts of FIB milling.
foil revealed the surface valley-and-ridge structure
documented by SEM and AFM very clearly beneath
the protective carbon cap (figure 14). Two varieties of
Ca defect structures associated with the surface ridges
rbo
zs
nc
ap are evident in these images. First, corridors of miss-
urf
ace ing material can be observed oriented normal to the
ridged surface of the foil (arrowed in figure 14). As
striking as these features are, they are not meaning-
ful, because they were produced artificially during
the sample preparation process. It is commonly ob-
served that when solids with uneven surfaces are
sputter milled by the focused ion beam technique,
the Ga ions are channeled by the valley structures
Bo downward into the specimen during milling (Volkert
tto
mo
f fo
and Minor, 2007). As a result, the section is selec-
il
tively scoured along distinct channels, leading to the
uneven excavation of material in parallel rows. This
undesirable artifact creates an appearance of parallel
slats within the foil.
The second set of visible defect features is struc-
1 µm
turally more significant. A higher-magnification

c-axis
the inclined planar defects observed in brightfield
38.2° TEM images. Electron diffraction patterns from the
foil revealed only quartz, with no evidence of an opa-
line phase.
Consequently, it seems clear that the iridescence
observed on the z faces of quartz from the Deccan
Traps of India is caused by a combination of the sur-
face grooving with an effective blaze angle of 38.2°
and an underlying lamellar structure, which together
create a diffraction grating for visible light. The ab-
sence of either opal or compositional impurities de-
finitively rules out iridescence by a thin-film effect,
sensu stricto. In addition, our observation contradicts
Raman’s 1950 model for iridescence insofar as the
500 nm
planar lamellae we observed were oriented parallel
Figure 15. At higher magnification, planar defects ori- to m {010} rather than z {011}. On the other hand, we
ented at 38.2° from the surface of the z face are appar- did find some support for Raman’s interpretation that
ent in this TEM image (arrows). The valleys on the z diffraction arises from a surface layer of the quartz
face coincide with the intersection of these defects crystals. After noting that broken z surfaces were not
with the surface. We hypothesize that the defects iridescent, we polished iridescent faces with a dia-
were preferentially etched and are responsible for the mond polishing compound (8 μm grit) to see whether
surface grooving that gives rise to the diffraction- the iridescence could be eliminated by abrasion. In
based iridescence.
fact, the iridescence at first intensified and then dis-
appeared. Although our study could not quantita-
tively constrain the depth of the diffracting layer, we
image of the foil (figure 15) shows that the slopes of estimate the thickness of the stratified medium as
the surface ridges are inclined ~38° from the top sur- less than 500 μm but greater than the 5 μm depth of
face of the foil, which corresponds to the z (011) plane our FIB TEM foil, consistent with Raman’s estimate
of the original crystal. Parallel line defects extend of 250 μm.
from the inclined ridges at the top of the foil through
to the bottom of the section (arrowed in figure 15). Origin of the Periodic Substructure. We pose two
We interpret these defects as traces of defect planes models for the existence of the repetitive structures
oriented parallel to the m (010) face of the quartz that create the diffraction grating. One possibility is
crystal. The calculated angle between m (010) and z that the lamellae observed in figure 15 represent sub-
(011) for quartz is 38.2°, consistent with the inclina- divided growth faces of quartz. Quartz is known to
tion observed in figure 15. The defects generally run grow with “artichoke” or “cathedral” morphologies
parallel to each other, but in places near the surface (figure 16) that involve the emergence of split, lay-
they create vein-like networks. Micropores measur- ered crystals from the rhombohedral faces of a parent
ing tens of nanometers in diameter can also be seen crystal. The cause of the phenomenon is not known,
as material gaps along the defects. but it may result from the poisoning of growth faces
by impurities, leading to the division of the rhombo-
DISCUSSION hedral faces into multiple members, whose c-axes
Cause of Iridescence. Our SEM and AFM images re- point roughly in the same direction.
vealed that the z faces of iris quartz are marked by a Although we initially favored this genetic mech-
topographic valley-and-ridge structure with remark- anism, we have subsequently grown skeptical of it.
ably regular intervals. The ridges were inclined from The artichoke morphology has been described for
the z (011) surface at an angle of 38°, and the distance macroscopic crystals with layered outgrowths that
from the top to the bottom of the grooving ranged are several millimeters in thickness. We know of no
from approximately 100 to 600 nm. Fourier analysis analogue for the nanoscale periodicity displayed by
of the AFM images placed the periodicity within the the iris quartz crystals from the Jalgaon District.
range of visible light, between 437 and 629 nm. That Moreover, it is not clear to us how an undulating sur-
magnitude is consistent with the spacing between face topography might emerge on the z faces and not

the r faces of quartz during crystal growth. If the
quartz druses grew in response to pulses of silica-rich
fluids into the basaltic cavities of the Deccan Traps,
then periodic growth behavior might be attributed to
pulsing of the fluids, but episodic infusions of fluids
would seem more likely to produce growth layers
stacked normal to the z and r faces, which was not
observed.
The second and more plausible explanation attrib-
utes the observed microstructures to polysynthetic
Brazil twinning, as Raman (1950) originally specu-
lated, but with a few departures from his original
model. Whereas Raman proposed that Brazil twins
were stacked parallel to the z faces and contained im-
purities at the twin boundaries, we argue that the
Brazil twin boundaries occur as pairs, that they are
il
oriented parallel to m {010}, and that diffraction of fo
of
m
light occurs at the twin boundaries as the result of tto
Bo
etching by aqueous fluids along the boundaries.
We base this interpretation on the observation 100 nm
that when one traces the planar defects along the FIB
Figure 17. In this TEM image, paired twin boundaries
TEM foil from the surface of the crystal toward the
(see arrows) are visible in the lower portion of the
interior, the slightly curved and wandering defect crystalline quartz foil.
traces grade into strictly flat planar defects. Those de-
fect planes occur as paired boundaries separated by
~30 nm (figure 17). Bragg thickness contours crossed

Figure 16. Schematic cross section of an “artichoke” these boundaries without interruption, and diffrac-
quartz crystal. tion patterns that include these regions appeared as
single-crystal patterns. Consequently, we found no
evidence that these intergrowths represent exsolu-
tion lamellae or intergrowths of another phase, and
we conclude that they are paired twin boundaries.
The two most common twins in quartz are the
Dauphiné and the Brazil twins, and it is notoriously
difficult to distinguish between these by diffraction
methods (Heaney, 1994).
McLaren and Phakey (1965, 1966) demonstrated
that interference fringes can be observed in bright-
field and darkfield TEM images across Brazil twin
boundaries. We examined a TEM foil from a Brazilian
amethyst (Smithsonian National Museum of Natu-
ral History, USNM #R1453) to confirm the appear-
ance of these fringes under the imaging conditions
we applied to the Indian iris quartz, and the fringes
were immediately apparent. We saw no such fringes
during our examination of the parallel defect planes
in the iris quartz from Jalgaon. That might weigh
against an interpretation of these features as Brazil
twins, but we are unaware of Dauphiné twins occur-
ring as polysynthetic, nanoscale intergrowths except
when crystals are heated to the transition tempera-

and Giovanoli, 1988). The paired Brazil twin bound-
aries in iris quartz is proposed by us to represent
c-axis Paired Brazil twin strained regions that were particularly susceptible to
boundaries dissolution, and we suggest that post-growth
episodes of etching by invasive fluids preferentially
altered the paired Brazil twin boundaries where they
intersected the surface z faces, sculpting the valley-
and-ridge topography we observed. Moreover, the
channeling of these fluids along the twin boundaries
within the crystal resulted in dissolution and recrys-
tallization of the twin boundaries, producing both
fluid-filled nanopores and the vein-like defect sur-
faces observed near the top of our TEM foil.
Consequently, we interpret the diffraction effect
z r as a result of surface etching and subsurface alter-
ation by post-depositional fluids. The iridescence
thus represents a combined effect of reflective inter-
Figure 18. Our model for iridescence in quartz. Paired ference from the grooved surface and transmissive
Brazil twin lamellae oriented parallel to m {010} interference at the weathered twin boundaries just
intersect the z {011} faces and were preferentially below the surface (figure 17). When the etched sur-
etched by aqueous fluids to create a surface grating. face layer is removed, as on the naturally fragmented
and the artificially polished z faces, then the irides-
cence disappears. Thus, the depth of the iridescence
effect traces the reach of aqueous alteration, not the
ture between low and high quartz (573°C; van Ten- interior extent of the twinning. In turn, this observa-
deloo et al., 1976). Moreover, interference fringes are tion explains why iris quartz crystals occur globally
not expected at Brazil twin boundaries when the yet remain elusive to collectors. Many quartz crys-
TEM section of quartz is oriented exactly perpendi- tals—particularly amethysts—are extensively
cular to the twin boundary—the electron beam from twinned according to the Brazil law and contain pe-
one twin member does not interfere with that of the riodic defects separated by distances close to the
other, as happens when the twin walls are oblique to wavelengths of visible light. But it is only when these
the surface of the TEM foil. crystals have experienced a subsequent stage of dis-
McLaren (1967) reported an absence of telltale solution, with differential etching along the defect
Brazil twin fringes in his TEM analysis of quartz planes (figure 18), that they will transform to iris
crystals exposed to a shear deformation parallel to quartz.
(001). Deformation generated paired Brazil twin In “To the Rainbow,” the Scottish poet Joanna
boundaries approximately 30 nm apart, and these Baillie (1762–1851) lamented the fate of natural won-
boundary pairs were themselves separated by hun- ders when their magic is reduced to empirical hy-
dreds of nanometers. Brightfield images from TEM potheses:
foils cut normal to the twins exhibited no fringe in-
terference. The defect structures described by When science from creation’s face
Enchantment’s veil withdraws,
McLaren (1967) are virtually identical to those we What lovely visions yield their place
characterized for the Indian iris quartz. Although we To cold material laws!
do not suggest that the quartz druses from the Dec-
can Traps were exposed to deformational shear, we As C.V. Raman understood, the search for an ex-
speculate that another process (possibly incorpora- planation of rainbows in gem materials is not simply
tion of Al or rapid growth rates) may have generated an exercise to recast a beautiful phenomenon as a
periodic paired Brazil twin boundaries oriented par- mathematical equation. Gem rainbows can appear
allel to m {010}. when crystal structures invisible to the unaided eye
It is well known that twin boundaries in minerals are arrayed in repetitive patterns that will diffract
are vulnerable to preferential weathering (e.g., Cor- light. Thus, modern scientific analysis withdraws en-
nell et al., 1974; Wilson and McHardy, 1980; Cornell chantment’s veil only to yield new enchantments.

ABOUT THE AUTHORS tional Science Foundation (NSF-EAR11-47728). SEM, FIB, and
Ms. Lin received her M.S. in geosciences at Pennsylvania State AFM work were performed at Pennsylvania State University’s Ma-
University in 2015 and is working with the Rapaport Diamond terials Characterization Laboratory with the assistance of Dr.
Corporation. Dr. Heaney is a professor in the Department of Geo- Trevor Clark, Julie Anderson, Joshua Maier, and Timothy Tighe.
sciences at Pennsylvania State University. Special thanks are due to Drs. James Kubicki, Maureen Feine-
man, Jeffrey Post, Long-Qing Chen, Yi Wang, and Mr. Phil Kong
ACKNOWLEDGMENTS for their useful suggestions.
This research was made possible through funding from the Na-
REFERENCES
Akizuki M., Sugawara H. (1970) The lamellar structure in moon- colors. Reports on Progress in Physics, Vol. 71, No. 7, pp.
stone and anorthoclase from Korea. Contributions to Mineral- 076401, http://dx.doi.org/10.1088/0034-4885/71/7/076401
ogy and Petrology, Vol. 29, No. 1, pp. 28–32, http://dx.doi.org/ Laufer G. (1996) Introduction to Optics and Lasers in Engineering.
10.1007/BF00387996 Cambridge University Press, Cambridge, England.
Balakirev V.G., Tsinober L.I., Tsyganov E.M. (1975) Electron-mi- Ma C., Gresh J., Rossman G.R., Ulmer G.C., Vicenzi E.P. (2001)
croscope study of Brazilian twins in synthetic amethysts. Micro-analytical study of the optical properties of rainbow and
Kristallografiya, Vol. 19, No. 4, pp. 517–520. sheen obsidians. The Canadian Mineralogist, Vol. 39, No. 1,
Baran Z., Godwod K., Warminski T. (1987) X-ray study of Brazil pp. 57–71, http://dx.doi.org/10.2113/gscanmin.39.1.57
twins in natural amethyst. Physica Status Solidi (a), Vol. 404, Ma C., Rossman G.R., Miller J.A. (2007) The origin of color in
No. 4, pp. 9–24, http://dx.doi.org/10.1002/pssa.2211010102 “fire” obsidian. The Canadian Mineralogist, Vol. 45, No. 3, pp.
Buckley A.N., Woods R. (1983) X-ray photoelectron spectroscopic 551–557, http://dx.doi.org/10.2113/gscanmin.45.3.551
investigation of the tarnishing of bornite. Australian Journal McConnell J.D.C. (1974) Electron-optical study of the fine struc-
of Chemistry, Vol. 36, No. 9, pp. 1793–1804, http://dx.doi.org/ ture of a schiller labradorite. In W.S. MacKenzie and J. Zuss-
10.1071/CH9831793 man, Eds., Feldspars. Manchester University Press.
Cliff G., Champness P.E., Nissen H.U., Lorimer G.W. (1976) Ana- McLaren A.C., Phakey P.P. (1965) A transmission electron micro-
lytical electron microscopy of exsolution lamellae in plagio- scope study of amethyst and citrine. Australian Journal of
clase feldspars. In H.-R. Wenk, Ed., Electron Microscopy in Physics, Vol. 18, No. 2, pp. 135–142, http://dx.doi.org/
Mineralogy. Springer, Berlin, Heidelberg, pp. 258–265. 10.1071/PH650135
Cornell R.M., Giovanoli R. (1988) Acid dissolution of akaganéite McLaren A.C., Phakey P.P. (1966) Electron microscope study of
and lepidocrocite: The effect on crystal morphology. Clays and Brazil twin boundaries in amethyst quartz. Physica Status So-
Clay Minerals, Vol. 36, No.5, pp. 385–390. lidi (b), Vol. 13, No. 2, pp. 413–422, http://dx.doi.org/10.1002/
Cornell R.M., Posner A.M., Quirk J.P. (1974) Crystal morphology pssb.19660130213
and the dissolution of goethite. Journal of Inorganic and Nuclear McLaren A.C., Pitkethly D.R. (1982) The twinning microstructure
Chemistry, Vol. 36, No. 9, pp. 1937–1946, http://dx.doi.org/ and growth of amethyst quartz. Physics and Chemistry of Min-
10.1016/0022-1902(74)80705-0 erals, Vol. 8, No. 3, pp. 128–135, http://dx.doi.org/10.1007/
Darragh P.J., Gaskin A.J., Terrell B.C., Sanders J.V. (1966) Origin of BF00311283
precious opal. Nature, Vol. 209, pp. 13–16, http://dx.doi.org/ McLaren A.C., Retchford J.A., Griggs D.T., Christie J.M. (1967)
10.1038/209013a0 Transmission electron microscope study of Brazil twins and
Gaillou E., Fritsch E., Aguilar-Reyes B., Rondeau B., Post J., Bar- dislacations[sic] experimentally produced in natural quartz.
reau A., Ostrooumov M. (2008) Common gem opal: An inves- Physica Status Solidi (b), Vol. 19, No. 2, pp. 631–644,
tigation of micro- to nano-structure. American Mineralogist, http://dx.doi.org/10.1002/pssb.19670190216
Vol. 93, No. 11–12, pp. 1865–1873, http://dx.doi.org/10.2138/ Meadows M.G., Butler M.W., Morehouse N.I., Taylor L.A.,
am.2008.2518 Toomey M.B., McGraw K.J., Rutowski R.L. (2009) Iridescence:
Gübelin E.J., Koivula J.I. (2005) Photoatlas of Inclusions in Gem- views from many angles. Journal of The Royal Society Inter-
stones. Opinio Verlag, Basel, Switzerland, Vol. 2. face, Vol. 6, Suppl. 2, pp. S107–S113, http://dx.doi.org/10.1098/
Heaney P.J. (1994) Structure and chemistry of the low-pressure sil- rsif.2009.0013.focus
ica polymorphs. Reviews in Mineralogy and Geochemistry, Miura Y., Tomisako T. (1978) Ion microprobe mass analysis of ex-
Vol. 29, No. 1, pp. 1–40. solution lamellae in labradorite feldspar. American Mineralo-
Heaney P.J., Davis A.M. (1995) Observation and origin of self-or- gist, Vol. 63, No. 5–6, pp. 584–590.
ganized textures in agates. Science, Vol. 269, No. 5230, pp. Nissen H.-U. (1971) End member compositions of the labradorite
1562–1565, http://dx.doi.org/10.1126/science.269.5230.1562 exsolution. Naturwissenschaften, Vol. 58, No. 9, pp. 454–454,
Heaney P.J., Vicenzi E.P., Giannuzzi L.A., Livi K.J.T. (2001) Focused http://dx.doi.org/10.1007/BF00624619
ion beam milling: A method of site-specific sample extraction Parker A.R., Townley H.E. (2007) Biomimetics of photonic nanos-
for microanalysis of Earth and planetary materials. American tructures. Nature Nanotechnology, Vol. 2, No. 6, 347–353,
Mineralogist, Vol. 86, No. 9, pp. 1094–1099, http://dx.doi.org/ http://dx.doi.org/10.1038/nnano.2007.152
10.2138/am-2001-8-917 Petrov A., Tanaka Y. (2011) Iris quartz. Mindat.org, www.mindat.
Kinoshita S., Yoshioka S., Miyazaki J. (2008) Physics of structural org/article.php/1335/Iris+Quartz

Raman C.V. (1950) Crystals of quartz with iridescent faces. Pro- Mineralogical Journal, Vol. 17, No. 2, pp. 53–76,
van Tendeloo G., van Landuyt J., Amelinckx S. (1976) The α→ β

ceedings of the Indian Academy of Sciences, Section A, Vol. http://dx.doi.org/10.2465/minerj.17.53
31, No. 5, pp. 275–279.
Renfro N., Koivula J.I. (2011) Gem News International: Spectral in- phase transition in quartz and AlPO4 as studied by electron mi-
terference in quartz from India. G&G, Vol. 47, No. 1, pp. 58–59. croscopy and diffraction. Physica Status Solidi (a), Vol. 33, No.
Rondeau B., Fritsch E., Mazzero F., Gauthier J.-P., Cenki-Tok B., 2, pp. 723–735, http://dx.doi.org/10.1002/pssa.2210330233
Bekele E., Gaillou E. (2010) Play-of-color opal from Wegel Tena, Vaughan D.J., Tossell J.A., Stanley C.J. (1987) The surface proper-
Wollo Province, Ethiopia. G&G, Vol. 46, No. 2, pp. 90–105, ties of bornite. Mineralogical Magazine, Vol. 51, No. 360, pp.
http://dx.doi.org/10.5741/GEMS.46.2.90 285–293, http://dx.doi.org/10.1180/minmag.1987.051.360.11
Sanders J.V. (1964) Colour of precious opal. Nature, No. 204, pp. Vigneron J.P., Pasteels J.M., Windsor D.M., Vértesy Z., Rassart M.,
1151–1153, http://dx.doi.org/10.1038/2041151a0 Seldrum T., Dumont J., Deparis O., Lousse V., Biró L.P., Ertz
Scharff F. (1875) Neues Jahrbuch für Mineralogie, Geologie, D., Welch V. (2007) Switchable reflector in the Panamanian tor-
Paläontologie, Stuttgart, Germany, E. Schweizerbart’sche Ver- toise beetle Charidotella egregia (Chrysomelidae: Cassidinae).
lagshandlung (in German). Physical Review Series E, Vol. 76, No. 3, p. 031907,
Schlössin H.H., Lang A.R. (1965) A study of repeated twinning, http://dx.doi.org/10.1103/PhysRevE.76.031907
lattice imperfections and impurity distribution in amethyst. Volkert C.A., Minor A.M. (2007) Focused ion beam microscopy
Philosophical Magazine, Vol. 12, No. 116, pp. 283–296, and micromachining. MRS Bulletin, Vol. 32, No. 05, pp. 389–
http://dx.doi.org/10.1080/14786436508218871 399, http://dx.doi.org/10.1557/mrs2007.62
Schmetzer K. (1987) Microscopic observation of twinning mi- vom Rath G.(1873) Mineralogische Mittheillungen von Gerhard
crostructure in natural amethyst. Neues Jahrbuch fur Miner- vom Rath. Annalen der Physik und Chemie, Vol. 6, No. 3, pp.
alogie-Monatshefte, Vol. 1, pp. 8–15. 337–386 cited in A. Petrov and Y. Tanaka (2011) Iris quartz.
Taijing L., Sunagawa I. (1990) Structure of Brazil twin boundaries Mindat.org, www.mindat.org/article.php/1335/Iris+Quartz
in amethyst showing Brewster fringes. Physics and Chemistry Wilson M.J., McHardy W.J. (1980) Experimental etching of a mi-
of Minerals, Vol. 17, No. 3, pp. 207–211, http://dx.doi.org/ crocline perthite and implications regarding natural weather-
10.1007/BF00201451 ing. Journal of Microscopy, Vol. 120, No. 3, pp. 291–302,
Taijing L., Sunagawa I. (1994) Texture formation of agate in geode. http://dx.doi.org/10.1111/j.1365-2818.1980.tb04149.x
For online access to all issues of GEMS & GEMOLOGY from 1934 to the present, visit:

NOTES & NEW TECHNIQUES
CHARACTERISTICS OF GEM-QUALITY
BY-PRODUCT SYNTHETIC ZINCITE
Ji Zhang, Yujie Gao, and Guanghai Shi
physical properties in applications such as photo-

Gem-quality synthetic zincite from Poland, an in- detectors (Soci et al., 2007), LED lights (Dong et al.,
dustrial by-product of zinc oxide processing, has 2007), laser diodes (Liang et al., 2010), transparent
been available in the market since the early 1980s. transistors, and substrates (Gahtar et al., 2013). The
In this investigation, conventional gemological wide-ranging industrial applications of ZnO have
methods, spectroscopic techniques, and advanced driven crystal growth efforts. Zincite crystals have
analytical testing were carried out on 10 rough been synthesized in labs through vapor deposition
and faceted samples. The chromophore man- (Pasko et al., 1976), hydrothermal methods (Karipidis
ganese (Mn), which is significant in natural zincite et al., 2008), and pressurized melt growth methods
and can reach 1 wt.% MnO, appears to be negli- (Nause and Nemeth, 2005). Yet there have been no
gible in by-product crystals. These crystals also ap- reports of any effort to pursue synthetic gem crystal
pear to have low to insignificant levels of other growth. Crystals grown by vapor deposition often ap-
potential chromophores such as iron and copper. pear one-dimensional, needle-like, or columnar due
Inclusions matching previously reported oval and to rapid growth along the c-axis (Pasko et al., 1976).
“tadpole-like” inclusions in Polish synthetic zincite Hydrothermal crystals display a two-dimensional
were observed and tested by Raman and IR spec- plate-like appearance because of the seed orientation
troscopy. These inclusions were identified as car- (Kortounova et al., 2001; Karipidis et al., 2008), and
bon-bearing material, possibly belonging to soot pressurized melt growth normally produces round
and/or gases captured during the growth of the boules (Nause and Nemeth, 2005).
synthetic zincite host. During the 1980s, a new form of synthetic zincite
emerged from an unnamed foundry in Poland’s
Z incite (ZnO) was discovered in 1810 (Dunn, 1995;

Welch, 2008) and mined from marble-hosted Zn-
Fe-Mn skarn-type deposits (Hague et al., 1956; Web-
In Brief
ster, 1983) at the Franklin and Sterling Hill mines of • The colors of by-product synthetic zincite from the
foundry in Lower Silesia, Poland, may be related to the
Sussex County, New Jersey. These mines are still con- material’s structure.
sidered the only locality for gem-quality natural
• Raman spectroscopy may be useful in distinguishing
zincite (Dunn, 1995; Peck et al., 2009; Wilson, 2013).
the Polish specimens from natural zincite.
The natural zincites were always aggregates, both
• Oval and “tadpole-like” inclusions in the by-product
massive and small (Nowak et al., 2007). Only a few
synthetic zincite were identified by Raman as carbon-
large gem-quality crystals displaying strong luster and bearing material, possibly from soot captured during
a beautiful bright red color have been faceted (Welch, the growth process.
2008). Some collectors seek the stone for its rarity and
attractive appearance, although a relatively low hard-
ness (4 on the Mohs scale) limits its use in jewelry
Lower Silesia Province (Crowningshield, 1985; Kam-
(Henderson, 1945; Nowak et al., 2007; Wilson, 2013).
merling and Johnson, 1995). This material grew as an
Zincite is also an important semiconductor ma-
industrial by-product in the chimney vents of indus-
terial, doped by impurity elements to enhance its
trial kilns producing zinc oxide powder, through a
process similar to chemical vapor deposition (Nowak
et al., 2007). We will refer to it as “by-product
zincite” hereafter. These crystals possess better
http://dx.doi.org/10.5741/GEMS.53.1.82 transparency and brighter color than natural zincite.
© 2017 Gemological Institute of America By-product zincites were sold in the jewelry market
82 NOTES & NEW TECHNIQUES GEMS & GEMOLOGY SPRING 2017

in Lower Silesia have previously been studied (e.g.,
Crowningshield, 1985; Kammerling and Johnson,
1995; Nowak et al., 2007). Insights into their color
and the nature of their unusual oval or “tadpole-like”
inclusions (Nowak et al., 2007) have remained elu-
sive. In this investigation, rough by-product zincite
crystal samples from the foundry in Lower Silesia
were chosen for standard gemological observations,
elemental analysis by X-ray fluorescence (XRF) and
inductively coupled plasma–mass spectrometry (ICP-
MS), and spectroscopic analysis in order to find the
characteristics that distinguish them from natural
material, identify their inclusion features, and provide
an explanation for the cause of color.
Figure 1. Most of the by-product synthetic zincites MATERIALS AND METHODS

in this study had a reddish orange to yellow color; Figure 1 shows four of the 10 rough by-product zincite
one sample (ZnO-G) was yellowish green. Top row, crystal samples (0.34–13.85 g) that were analyzed. Two
left to right: samples ZnO-LC (13.85 g) and ZnO-SC were faceted by author YG (labeled Zincite-L and
(6.34 g). Bottom row, left to right: samples ZnO-G Zincite-S). Sample zin-10 is a group of five un-
(0.06 g) and zin-7 (1.12 g). Composite photo by Ji processed rough fragments from a yellow crystal; the
Zhang. remainder of this crystal is the faceted sample Zincite-
S. All rough samples were purchased at the Warsaw
Mineral Show in September 2010; the seller claimed
from the early 1980s until the 2000s, but now they they were produced at an undisclosed foundry in
can only be acquired from existing stocks (Kammer- Lower Silesia, Poland.
ling and Johnson, 1995; Nowak et al., 2007). Standard gemological testing, visible spec-
The properties and formation processes of by-prod- troscopy, and Fourier-transform infrared (FTIR) spec-
uct synthetic zincite crystals from the Olawa foundry troscopy were performed at the School of Gemology,
A B
t Figure 2. The main crystal

faces of the by-product
p t zincite: hexagonal prism m,
hexagonal pyramids t and p,
m single face c, and contact
m twins along the (0001) direc-
tion. These features follow
the wurtzite structure of
zincite (Zhao et al., 2006). A:
sample ZnO-LC, field of
C c D view 12.5 mm. B: sample
ZnO-LC in another direc-
t
t tion, field of view 13.0 mm.
C: sample ZnO-SC, field of
m view 3.5 mm. D: sample
m t ZnO-SC, field of view 4.5
p mm. Photomicrographs by Ji
Zhang.
m
NOTES & NEW TECHNIQUES GEMS & GEMOLOGY SPRING 2017 83

China University of Geosciences in Beijing (CUGB).
Visible spectra of samples Zincite-L, Zincite-S, and TABLE 1. Chemical composition of potential chro-
mophore elements of by-product synthetic zincite
ZnO-G were recorded using a LabTech UV-Vis spec-
tested by ICP-MS (in ppmw).
trophotometer in the 400–700 nm range with a sam-
pling interval of 1 nm at a rate of 600 nm/min. ZnO-LC Zincite-L Zincite-S zin-10 ZnO-G
Mid-infrared (4000–400 cm–1) spectra of rough sample (orange- (reddish (yellow) (yellow) (yellowish
zin-10 were obtained with a Bruker Tensor 27 FTIR red) orange) green)
spectrophotometer, operating in absorption and re- Ti 8.3 10.8 5.8 0.0 0.8
flectance modes with a resolution of 4 cm–1 and 50 V 0.2 1.4 5.2 1.4 0.7
scans per sample at room temperature using the KBr Cr 1.6 3.8 7.7 11.0 2.3
pellet method (100 mg KBr and 1.0 mg sample).
Mn 1.8 1.6 4.3 7.5 1.5
XRF results for nine samples (ZnO-LC, Zincite-
L, Zincite-S, zin-10, ZnO-G, zin-fs, zin-fc, zin-7, and Fe 63.3 51.0 104.9 86.2 38.3
zin-8) were tested at the Shanghai laboratory of the Co 0.7 0.7 1.9 0.6 0.7
National Gemstone Testing Center (NGTC), using a Ni 9.2 6.1 19.7 6.1 10.8
Shimadzu EDX-7000 from 11Na to 92U with a 3 mm Cu 16.1 6.8 29.1 6.9 9.3
collimator and a two-minute testing time.
ICP-MS of five samples (ZnO-LC, Zincite-L,
Cd bdl bdl bdl bdl bdl
Sample detection limits (ppmw) were Ti (0.336), V (0.058), Cr
Zincite-S, three specimens of the zin-10 group, and (0.245), Mn (0.014), Fe (3.383), Co (0.007), Ni (0.023), Cu (0.020),
ZnO-G) was conducted at Key Laboratory of Subma- and Cd (0.021).bdl = below detection limit
rine Geosciences of the Second Institute of Oceanog-
raphy, State Oceanic Administration in Hangzhou,
using an Elan DRC-e ICP-MS with the U.S. Geolog-
ical Survey (USGS) standard BHVO-2 measured to RESULTS AND DISCUSSION
monitor accuracy. Small quantities of the five sam- Crystallography and Gemological Properties. The by-
ples were ground into a powder and sieved to obtain product crystals exhibited a hexagonal prismatic and
a 200 mesh size. Aqueous solutions of the powders pyramid appearance, consisting of hexagonal prism
were created by dilution of solutions of the powders m {1010}, hexagonal pyramids t {1124} and p {1011},
created using first aqua regia in a Teflon bomb with and single face c {0001} (figure 2), with contact twin-
heating followed by HClO4 and HNO3 in a Teflon ning along (0001) following wurtzite structure (Zhao
bomb with heating. The unit of ICP-MS results was et al., 2006). The samples’ colors ranged from reddish
ppmw; detection limits for all elements were lower orange to yellow to greenish yellow, and all of them
than 4 ppmw. were transparent, with a subadamantine to vitreous
Raman spectroscopy was performed at the Chi- luster (figures 1 and 3). Their RIs were over the limit
nese Academy of Geological Sciences (CAGS) in Bei- of the refractometer (RI = 1.78). With a handheld
jing. Six samples (Zincite-L, Zincite-S, ZnO-LC,
ZnO-SC, zin-fs, and zin-fc) were analyzed using a
Renishaw 2000 laser Raman microspectroscopy sys- Figure 3. The 36.54 ct reddish orange Zincite-L (left)
tem with an argon-ion laser (514.5 nm excitation), 20 and the 2.87 ct yellow Zincite-S (right) were faceted
mW power, minimum spot size of 1 μm, and spectral from the by-product zincite rough. Photo by Yujie Gao.
resolution of 1–2 cm–1. The oval inclusion in Zincite-
S and another unidentified inclusion belonging to zin-
fc were also analyzed in situ by Raman spectroscopy.
We compared our results with the Raman spectra
and chemical data of two previously tested natural
red zincite samples (R050419 and R050492) from the
RRUFF database (http://rruff.info). Sample R050419
is a deep red crystal from the Sterling Hill mine, and
sample R050492 is a massive deep red crystal from
the Franklin mine. No natural zincite was examined
in this study due to the difficulty of obtaining such
material.

TRACE-ELEMENT CONCENTRATION VISIBLE SPECTRA
120
ZnO-LC
Zincite-S
(orange-red)
100 ZnO-G
Zincite-L
CONCENTRATION (ppmw)
(reddish orange)
Zincite-S
80
(yellow)
ABSORBANCE
Zin-10
60 (yellow)
ZnO-G
(yellowish green)
40
20
0
Ti V Cr Mn Fe Co Ni Cu Cd
400 450 500 550 600 650 700

ELEMENT
WAVELENGTH (nm)
Figure 4. The nine potential chromophore elements of Figure 6. The visible spectra showed that yellow by-
five by-product zincite samples were tested by ICP- product samples Zincite-S and ZnO-G also absorbed
MS. In all five samples, Fe content was higher than light in the blue to violet regions.
the other eight elements; the concentrations of each
of the non-iron elements are lower than 30 ppm.
Chemical Composition. XRF testing suggested that

spectroscope, absorption bands were only seen in the the samples were composed of almost pure zinc and
blue to violet regions. The samples’ specific gravity, oxygen. To investigate the potential chromophores
measured hydrostatically, ranged from 5.66 to 5.70. of by-product zincite, the powders of five samples
These properties were consistent with previous re- were tested by ICP-MS. The results showed that Mn
ports (Crowningshield, 1985; Kammerling and John- levels were negligible, ranging from 1.53 to 7.55 ppm
son, 1995; Nowak et al., 2007). with a mean of 3.38 ppm, corresponding to previous
reports (Kammerling and Johnson, 1995; Nowak et
al., 2007). These levels were all dramatically lower
than Mn contents in natural material (approximately
Figure 5. The visible spectrum of reddish orange by-
1 wt.% MnO; see Webster, 1983; http://ruff.info).
product sample Zincite-L showed that it absorbed
light in the green to violet region, from 550 to 400 nm. Iron contents in by-product samples were relatively
high among nine potential chromophores, ranging
from 38.31 to 104.92 ppmw with a mean of 68.74
VISIBLE SPECTRUM ppm; the contents of other elements were lower than
20
30 ppmw (table 1; figure 4).
18
16 Visible Spectra and Cause of Color. The visible spec-

REFLECTANCE (%)
14 trum of reddish orange Zincite-L displayed a wide ab-

12 sorption band from 550 to 400 nm, covering the
10 green to violet regions (figure 5). Yellow Zincite-S re-
8
vealed a broad absorption band between 480 and 420
6
nm in the blue to violet region, while yellowish
4
green ZnO-G absorbed only violet light from 450 to
400 nm (figure 6).
2
There is no clear relationship between the mate-
0
400 450 500 550 600 650 700 rial’s color and its chemical composition, and the
WAVELENGTH (nm) contents of potential chromophore elements were
low, although the yellow samples had relatively

R AMAN SPECTRA
Raman Spectra. Raman spectra were obtained from
521
477 437
sample Zincite-L with its c-axis oriented parallel and
perpendicular to the laser beam, together with five
R040592,
562 other samples (Zincite-S, ZnO-LC, ZnO-SC, zin-fs,
532 nm
1091 1068
and zin-fc) whose c-axes were randomly oriented with
331
respect to the laser. In addition, the spectra from two
natural samples, taken from the RRUFF database,
522 were analyzed. The unoriented Raman spectra were
R050419, similar, so one representative spectrum of Zincite-S
532 nm
568 was chosen (figure 7). The Raman peaks of zincite
478
1088 1064 436 were divided into three groups:
331
(1) Shared peaks in natural and by-product zincite
INTENSITY (COUNTS)
spectra at approximately 331 and 437 cm–1

1149
410
(2) Unique peaks in natural zincite spectra at
1104 438 377 about 477, 522, 568, 1065, and 1090 cm–1
c
580
331 (3) Unique peaks in by-product synthetic spectra
647 546
993 at approximately 379, 410, 540, 580, 660, and
1150 cm–1
436 The latter two groups of peaks could serve to dis-
tinguish by-product zincite from natural zincite
|| c
using Raman spectroscopy.
1147 575 332 Oval Inclusion. Nowak et al. (2007) reported unusual
oval and “tadpole-like inclusions” in Polish by-prod-
437 uct synthetic zincite; in this study, two similar inclu-
Unoriented
sions were observed and tested. The first oval
1149 411 inclusion, belonging to sample Zincite-S (figure 8),
575 378
332 showed Raman peaks at 2938, 2923, 2882, 2854, 1613,
986 660 539 206 1546, 1507, 1442, and 1098 cm–1, which indicated
1200 1000 800 600 400 200 some carbon-bearing material stretching modes (figure
RAMAN SHIFT (cm–1) 9). The second inclusion, located at the cleavage plane
of sample zin-fc (the figure was not obtained success-
fully due to the rough cleavage surface), displayed
Figure 7. These Raman spectra of zincite, from top to Raman peaks at 2929, 2882, 2850, 1613, 1580, 1430,
bottom, belong to natural samples R050492 and
1392, 1346, 1296, and 1204 cm–1 (figure 10).
R050419 from the RRUFF database (both unoriented)
and three by-product samples from this study:
Combined with the C-H stretching mode indi-
Zincite-L (⊥c-axis), Zincite-L ( || c-axis), and Zincite-S cated by the IR spectra of rough sample zin-10 (figure
(unoriented). Spectra are vertically shifted for clarity. 11) and excluding the shellac used in the cutting
process (whose Raman peaks are not a match), it is
suggested that there are some carbon-bearing mate-
rials that make up the oval inclusions in the by-
higher Fe contents than orange to red or yellowish product synthetic zincite. Although the origin of the
green samples (again, see table 1 and figure 4). The oval inclusion was not clear, possible sources in-
cause of various colors observed may be related to clude (1) a group of carbon-bearing gases such as CH4
structure rather than chemical composition. These and CO2 in the crystal formation atmosphere; and (2)
structural factors include increasing amounts of band the complex fuel ashes (carbon-bearing soot) pro-
gap absorption from defect states (Kammerling and duced by the smelting process (Nowak et al., 2007).
Johnson, 1995), lattice defects with electron donors On one hand, Raman spectra of individual gases
such as neutral zinc atoms (Nowak et al., 2007), and (Dickinson and Rasetti, 1929) have one or two peaks
either oxygen deficit or zinc excess within zincite near the peaks of the oval inclusion, such as the
(Mikami et al., 2005; Karipidis et al., 2008). 1285 cm–1 peak of CO2 and the 2914 and 3022 cm–1

unusual oval inclusion Figure 8. A: Small inclusions
A B near surface
were observed in the pavilion
of Zincite-S; field of view 2.5
mm. B: When the internal fea-
tures of Zincite-S were exam-
ined with magnification, the
small inclusions
inside inclusions shown in figure 8A
came into focus, but the oval
inclusion near the surface was
100 µm blurred. C: The round dot in
the middle is the same oval
C D inclusion as in figure B, shown
in reflected light. D: This oval
inclusion in Zincite-S was also
observed and photographed in
situ when Raman spec-
troscopy was performed. The
colored dots represent analysis
spots. Photomicrographs by
100 µm 20 µm Yujie Gao.
peaks of CH4. On the other hand, some peaks in the These are at 1580 cm–1 (same as the G band of soot),
oval inclusion’s spectra are the same as or similar to 1346 cm–1 (near the 1350 cm–1 D1 band), 1507 cm–1
the peaks in Raman spectra of soot (e.g., Ramya et (near the 1500 cm–1 D3 band), 2800–2900 cm–1 (near
al., 2013; Kim et al., 2005; Sadezky et al., 2005). the 2700, 2900, and 3100 cm–1 Lorentz-shaped
R AMAN SPECTRA
2882
2853
Figure 9. Raman spec-
1507
1546 1442
tra of the unusual oval
177
2923 1098 300
inclusion in sample
2854
438
Zincite-S showed peaks
at 2938, 2923, 2882,
2854, 1613, 1546, 1507,
INTENSITY
2936 1442, and 1098 cm–1.

1442 These peaks indicated
1613
some carbon-bearing
438
material stretching
2938 modes. The colors of
the traces correspond to
the analysis spots
1508 1442
shown in figure 8D.
1098 Spectra are vertically
1046 shifted for clarity.
432
3500 3000 2500 2000 1500 1000 500`
RAMAN SHIFT (cm–1)

R AMAN SPECTRA
1580 1430
1392
1613 1442
1710
1346
1296
1204
2882
2850 1550
2934 1280 1254 Figure 10. Raman spec-
1520
2962 tra of the inclusion lo-
2000 1750 1500 1250 1000 cated on the cleavage
2727
plane of sample zin-fc
2929
INTENSITY
showed peaks at 2929,

2882, 2850, 1613, 1580,
546 144
1430, 1392, 1346, 1296,
and 1204 cm–1. Spectra
are vertically shifted
1597 for clarity.
1650
1450
2940 2884
1613 462
3500 3000 2500 2000 1500 1000 500
RAMAN SHIFT (cm–1)
bands), 1613 cm–1 (near 1620 cm–1), and 1204 cm–1 those of individual organic gases. Thus, the Raman
(near 1200 cm–1). results might be contributed mainly by soot and sup-
The Raman peaks of the oval inclusion more ported by some individual gases.
closely matched peaks of carbon-bearing soot than The oval inclusions reported by Nowak et al.
(2007) and in this paper were quite different from the
fluid inclusions, liquid film negative crystals, growth
tubes, and cloud inclusions in natural zincite
Figure 11. In the IR spectrum of rough sample zin-10, (Kawano, 2008). Although lab-grown synthetic zincite
the peak at 438 cm–1 belongs to the zincite crystal, has not been reported for gemstone use, the oval in-
while a series of bands at 2850 and 2920 cm–1 indi-
clusion composed of carbon-bearing materials (possi-
cate the C-H stretching mode of some possible car-
bly soot and/or gases) might not occur in lab-grown
bon-bearing materials.
zincite because of the accurately controlled original
materials and reaction processes (Pasko et al., 1976;
IR SPECTRUM
Nause and Nemeth, 2005; Karipidis et al., 2008).
1.4
438
ABSORBANCE (ARB. UNITS)
1.2 CONCLUSIONS
1.0 The by-product synthetic zincite has much less man-
532
ganese than natural zincite, and the causes of color of
0.8
by-product synthetic and natural zincite crystals ap-
0.6 pear to be different. Besides Mn, the contents of other
2923 2853 1634
0.4 1039 potential chromophores are also at levels low enough
3425
to indicate that the various colors of the by-product
0.2
material may be affected by structure-related factors.
0.0 Oval inclusions consistent with a previous report on
3500 3000 2500 2000 1500 1000 500
by-product synthetic zincite were observed, and testing
WAVENUMBER (cm–1) suggested they were soot and/or gases captured during
the growth of host by-product synthetic zincite.

ACKNOWLEDGMENTS
ABOUT THE AUTHORS
The authors would like to thank Prof. Pu Wang and Drs. Yingxin
Mr. Zhang, a postgraduate student at the State Key Laboratory of
Liu, Mingyue He, Xiaoyan Yu,Ye Yuan, and Xuemei He at CUGB,
Geological Processes and Mineral Resources at China University
along with Prof. Wenyi Xue at CAGS, for their help in analysis and
of Geosciences in Beijing at the time of this study, is currently a
discussions. Dr. Jie Li and Dr. Jihao Zhu at Key Laboratory of
teaching assistant at the School of Gemology, Shanghai Jian Qiao
Submarine Geosciences of the Second Institute of Oceanogra-
University. Mr. Gao, a postgraduate student at the School of
phy, State Oceanic Administration in Hangzhou, China, are
Gemology, China University of Geosciences in Beijing at the time
thanked for their help with ICP-MS testing. Ms. Ling Du at the
of this study, is an assistant appraiser at National Gems & Art-
Shanghai laboratory of the NGTC is thanked for her assistance
works Appraisal Company Limited of Shenzhen. Dr. Shi
with XRF testing. The thoughtful and constructive comments by
([email protected]) is a professor of gemology and mineralogy
reviewers and editors are gratefully acknowledged. This research
at the State Key Laboratory of Geological Processes and Mineral
was partially supported by Grant No. 41373055 from the National
Resources at China University of Geosciences in Beijing.
Science Foundation of China.
REFERENCES
Crowningshield G.R. (1985) Gem Trade Lab Notes: Synthetic zinc oxide by chemical vapor transport. Journal of Crystal
zincite. G&G, Vol. 21, No. 4, pp. 237–238. Growth, Vol. 276, No. 3–4, pp. 389–392, http://dx.doi.org/
Dickinson R.G., Dillon R.T., Rasetti F. (1929) Raman spectra of 10.1016/j.jcrysgro.2004.11.431
polyatomic gases. Physical Review, Vol. 34, No. 4, pp. 582–589, Nause J., Nemeth B. (2005) Pressurized melt growth of ZnO
http://dx.doi.org/10.1103/PhysRev.34.582 boules. Semiconductor Science and Technology, Vol. 20, No.
Dong C.K., Han W.S., Kong B.H., Cho H.K., Hong C.H. (2007) Fab- 4, pp. S45–S48, http://dx.doi.org/10.1088/0268-1242/20/4/005
rication of the hybrid ZnO LED structure grown on p-type GaN Nowak J.W., Braithwaite R.S.W., Nowak J., Ostojski K., Krystek
by metal organic chemical vapor deposition. Physica B: Con- M., Buchowiecki W. (2007) Formation of large synthetic zincite
densed Matter, Vols. 401–402, No. 4, pp. 386–390, (ZnO) crystals during production of zinc white. Journal of
http://dx.doi.org/10.1016/j.physb.2007.08.194 Gemmology, Vol. 30, No. 5/6, pp. 257–267.
Dunn P.J. (1995) Franklin and Sterling Hill, New Jersey: The Pasko P.G., Kidyarov B.I., Avdienko K.I. (1976) Influence of the
World’s Most Magnificent Mineral Deposits. The Franklin-Og- vapor phase composition and reagent flow on the growth mech-
densburg Mineralogical Society. anism and morphology of ZnO crystals. Journal of Crystal
Gahtar A., Benramache S., Benhaoua B., Chabane F. (2013) Prepara- Growth, Vol. 33, No. 2, pp. 298–302, http://dx.doi.org/
tion of transparent conducting ZnO:Al films on glass substrates 10.1016/0022-0248(76)90056-7
by ultrasonic spray technique. Journal of Semiconductors, Vol. Peck W.H., Volkert R.A., Mansur A.T., Doverspike B.A. (2009) Sta-
34, No. 7, pp. 26–30. ble isotope and petrologic evidence for the origin of regional
Hague J.M., Baum J.L., Herrmann L.A., Pickering R.J. (1956) Ge- marble-hosted magnetite deposits and the zinc deposits at
ology and structure of the Franklin-Sterling area, New Jersey. Franklin and Sterling Hill, New Jersey highlands, United
Bulletin of the Geological Society of America, Vol. 67, No. 4, States. Economic Geology, Vol. 104, No. 7, pp. 1037–1054,
pp. 435–474, http://dx.doi.org/10.1130/0016-7606(1956)67% http://dx.doi.org/10.2113/econgeo.104.7.1037
5B435:GASOTF%5D2.0.CO;2 Ramya K., John J., Manoj B. (2013) Raman spectroscopy investiga-
Henderson E.P. (1945) Zincite. G&G, Vol. 5, No. 3, pp. 251–256. tion of camphor soot: Spectral analysis and structural informa-
Kammerling R.C., Johnson M.L. (1995) An examination of tion. International Journal of Electrochemical Science, Vol. 8,
“serendipitous” synthetic zincite. Journal of Gemmology, Vol. pp. 9421–9428.
24, No. 8, pp. 563–568. Sadezky A., Muckenhuber H., Grothe H., Niessner R., Pöschl U.
Karipidis T.K., Maltsev V.V., Volkova E.A., Chukichev M.V., (2005) Raman microspectroscopy of soot and related carbona-
Leonyuk N.I. (2008) Thermal stability of zincite single crys- ceous materials: spectral analysis and structural information.
tals. Crystallography Reports, Vol. 53, No. 2, pp. 326–330, Carbon, Vol. 43, No. 8, pp. 1731–1742, http://dx.doi.org/
http://dx.doi.org/10.1134/S1063774508020247 10.1016/j.carbon.2005.02.018
Kawano J. (2008) Research Lab. Report: Natural zincite GAAJ- Soci C., Zhang A., Xiang B., Dayeh S.A., Aplin D.P.R., Park J., Bao
ZENHOKYO Laboratory 2008.7.11, http://grjapan.ddo.jp/ X.Y., Lo Y.H., Wang D. (2007) ZnO nanowire UV photodetectors
gaaj_report/2008/2008_07en.html with high internal gain. Nano Letters, Vol. 7, No. 4, pp. 1003–
Kim U.J., Furtado C.A., Liu X., Chen G., Eklund P.C. (2005) Raman 1009, http://dx.doi.org/10.1021/nl070111x
and IR spectroscopy of chemically processed single-walled car- Webster R. (1983) Gems: Their Sources, Descriptions and Identi-
bon nanotubes. Journal of the American Chemical Society, fication, 4th ed. Revised by B.W. Anderson, Butterworth-Heine-
Vol. 127, No. 44, pp. 15437–15445, http://dx.doi.org/ mann Ltd., London.
10.1021/ja052951o Welch F.S. (2008) Identification of 1850s brown zinc paint made
Kortounova E.V., Lyutin V.I., Dubovskaya V.D., Chvanski P.P. with franklinite and zincite at the U.S. Capitol. APT Bulletin,
(2001) The growth of zinc oxide crystals with impurities. High Vol. 39, No. 1, pp. 17–30, http://dx.doi.org/10.2307/25433934
Pressure Research, Vol. 20, No. 1–6, pp. 175–183, Wilson W.E. (2013) The best zincite specimens from Franklin,
http://dx.doi.org/10.1080/08957950108206165 New Jersey. The Mineralogical Record, Vol. 44, No. 5, pp. 509–
Liang H.K., Yu S.F., Yang H.Y. (2010) Directional and controllable 534.
edge-emitting ZnO ultraviolet random laser diodes. Applied Zhao J., Jin Z.G., Li T., Liu X.X. (2006) Preparation and character-
Physics Letters, Vol. 96, No. 10, pp. 101116–101116-3, ization of ZnO nanorods from NaOH solutions with assisted
http://dx.doi.org/10.1063/1.3356221 electrical field. Applied Surface Science, Vol. 252, No. 23, pp.
Mikami M., Eto T., Wang J., Masa Y., Isshiki M. (2005) Growth of 8287–8294, http://dx.doi.org/10.1016/j.apsusc.2005.10.050

Editors
Thomas M. Moses | Shane F. McClure
Three Faceted
COLLECTOR SPECIMENS
Three stones recently submitted to
GIA (figure 1) were found to be rare
mineral species that had not previ-
ously been examined at the Carlsbad
laboratory. Standard gemological test-
ing and Raman analysis identified
these specimens as wagnerite, thaum-
asite, and strontianite.
A 0.60 ct transparent golden yel-
low pear brilliant cut with vitreous
luster was identified as the phos-
phate monoclinic crystal wagnerite,
(Mg,Fe2+)2(PO4)F. Though uncommon,
wagnerite occurs in various geologic
settings, such as granitic pegmatites, Figure 1. Left to right: 0.60 ct golden yellow wagnerite, 1.72 ct colorless
gneisses, eclogites, and hydrothermal thaumasite, and 1.41 ct pale yellow strontianite.
environments (P. Korbel and M. Novak,
The Complete Mineral Encyclopedia,
Gramercy, Lisse, the Netherlands, optic nature. It was inert to long-wave step cut with vitreous luster. It proved
2003). It had a specific gravity (SG) of and short-wave UV. These properties, to be strontianite, SrCO3, an or-
3.15, a refractive index (RI) of 1.568– along with Raman testing, led to its thorhombic carbonate mineral mem-
1.580, and a biaxial positive optic na- identification as thaumasite, a silicate ber of the aragonite group with a
ture. Wagnerite is a relatively soft and mineral with the chemical formula Mohs hardness of 3.5. The 1.41 ct
brittle mineral with a Mohs hardness of Ca3(SO4)[Si(OH)6](CO3)·12H2O belong- specimen was highly birefringent,
5.0–5.5. The stone was inert under both ing to the hexagonal crystal system. with an RI of 1.511–1.661, an SG of
long-wave and short-wave UV radia- Magnification revealed strong dou- 3.76, and a biaxial negative optic na-
tion. Multiphase fluid inclusions, fin- bling, spectacular radiating fibrous in- ture. It fluoresced red under long-
gerprints, graining, and color zoning clusions, reflective films, tiny crystal wave UV but displayed a weak
features were observed with magnifica- inclusions, and strong hexagonal an- whitish reaction under short-wave
tion. gular graining. Thaumasite has a vit- UV. The stone exhibited strong dou-
The largest stone, a 1.72 ct trans- reous luster and a brittle nature due to bling and fluid inclusions under mag-
parent colorless octagonal step cut, its low hardness of 3.5 on the Mohs nification. Raman spectroscopy
had a relatively low SG of 1.90, an RI scale. Thaumasite occurs in geother- confirmed its identity. Strontianite
of 1.470–1.515, and a biaxial negative mal waters or seawaters reacting with occurs together with calcite and zeo-
basalt and tuffs, or in areas of contact lites in hydrothermal, low-tempera-
metamorphism (J.W. Anthony et al., ture veins and in cavities of volcanic
Editors’ note: All items were written by staff Handbook of Mineralogy: Volume II: rocks (Korbel and Novak, 2003).
members of GIA laboratories. Silica, Silicates: Part 2, Mineral Data Wagnerite, thaumasite, and stron-
Publishing, Tucson, Arizona, 1995, p. tianite are rarely seen as faceted gems
GEMS & GEMOLOGY, Vol. 53, No. 1, pp. 90–101.
790). due to their softness and brittleness.
© 2017 Gemological Institute of America The last stone we observed was a These collector specimens were the
transparent pale yellow octagonal first of their kind to be submitted to
90 LAB NOTES GEMS & GEMOLOGY SPRING 2017

R AMAN SPECTRUM
GIA’s Carlsbad laboratory for testing. 100000
Although challenging to identify 90000
without advanced testing techniques,
INTENSITY (COUNTS)
80000
they bring excitement to gemologists 70000

searching for unusual faceted gems. 60000
Jonathan Muyal and Nicole Ahline 50000
40000
30000
Unusual Dark Orangy Red 20000
CORDIERITE 10000
The New York lab recently examined 0

the transparent dark orangy red 150 350 550 750 950 1150 1350
cordierite shown in figure 2. RAMAN SHIFT (cm–1)

Cordierite, a magnesium aluminum
cyclosilicate with the formula Figure 3. Raman spectroscopy identified the orangy red stone as a
Mg2Al4Si5O18, is better known by its cordierite.
gemological name of iolite, which
comes from the Greek ios (violet). The
term is reserved for the violet through
using 514 nm laser excitation conclu- (figure 4). The larger surface-reaching
slightly violetish blue gem-quality va-
sively identified the stone as fractures in the stone appeared to
riety of cordierite. Iolite’s typical colors
cordierite (figure 3) with a strong have a “heat wave” appearance (fig-
have been compared to blue sapphires
match to the cordierite spectra in the ure 4B), that could be characteristic of
and tanzanites. These colors are
RRUFF database (http://rruff.info). the clarity-enhancing material.
thought to come from the charge
Microscopic examination with Cordierite, which is structurally sim-
transfer between Fe2+ and Fe3+ ions.
fiber-optic illumination revealed fin- ilar to beryl, will easily develop frac-
The stone’s RI ranged from 1.515
gerprints, rows of flaky whitish inclu- tures and is commonly clarity
to 1.547, with a birefringence of 0.037.
sions, and large fractures throughout enhanced with oil-based fluids. Both
This refractive index is much lower
the stone that had been filled with an cordierite and beryl are structurally
than what is listed in the literature for
unknown clarity-enhancing material similar in that they are cyclosilicates,
iolite, and we believe the discrepancy
is due to the stone’s lower Fe content.
SG was measured as 2.56. The stone
showed strong trichroism ranging Figure 4. Various inclusions within the dark orangy red cordierite. Numer-
from orangy red to dark brown to pur- ous partially enhanced fractures are visible (B), along with low visibility
ple. There was no fluorescence ob- on the surface (D). Field of view 4.77 mm (A), 14.48 mm (B), 3.57 mm (C),
served with exposure to long- and and 3.57 mm (D).
short-wave UV. Raman spectroscopy
Figure 2. This 8.13 ct cordierite

had an unusual dark orangy red
color.
A B
C D
LAB NOTES GEMS & GEMOLOGY SPRING 2017 91

VISIBLE ABSORPTION SPECTRA
30 Cape DIAMOND with Yellow
Orangy red Phosphorescence
Violet
ABS. COEFFICIENT (cm–1)
25
Phosphorescence in a cape diamond is
very rare, especially after exposure to
20
long-wave UV light. The New York
laboratory recently examined a cape
15
diamond exhibiting yellow phospho-
rescence after exposure to a long-wave
10
UV source. The 0.75 ct Faint yellow-
green round brilliant is shown in fig-
5
ure 6.
0
Cape absorption peaks were ob-
400 450 500 550 600 650 700 750 800 850 served at 415 and 477 nm, along with
WAVELENGTH (nm) a GR1 peak (figure 7, left). The IR spec-
trum confirmed that this was a type Ia
diamond, with nitrogen aggregates de-
Figure 5. In these visible absorption spectra, notice the large window be-
tected in the one-phonon region. Hy-
tween 600 and 800 nm for the orangy red cordierite sample and the ab-
drogen-related peaks were also
sorption band centered at 580 nm for the GIA reference spectrum of violet
detected at 1405, 2785, 3107, 3237,
iolite.
4169, and 4496 cm–1. Brown radiation
stains were found on the girdle. The di-
amond contained contacted partially
containing linked six-member rings iolite. The violet color of iolite is pro-
graphitized omphacite crystals with
of tetrahedra that result in channels duced by a strong absorption band
cuboctahedral morphology (figure 7,
within the crystal structure. These centered at approximately 580 nm,
right). Strong blue fluorescence and
channels are capable of trapping resulting from intervalence charge
medium yellow phosphorescence were
water along with cations such as transfer to either valence state of Fe
observed under a desktop long-wave
from the alkali and alkaline earth cations that might have occupied the
UV light source. Under short-wave
metals. The high water content divalent and trivalent sites in the
UV, medium yellow fluorescence was
within the stone was represented by crystal structure (substituting for Mg
noted but phosphorescence was ab-
the strong overtone bands seen in the and Al, respectively; see M. O’-
sent. DiamondView imaging (<225
Fourier-transform infrared (FTIR) Donoghue, Gemstones, Chapman
nm excitation) also showed greenish
spectrum, with very similar features and Hall Ltd., London, 1988, p. 77).
blue fluorescence but no phospho-
to that of hydrothermal natural beryl Although the red cordierite con-
rescence. Fiber-optic illumination
spectra. Laser ablation–inductively tained a considerable amount of iron,
revealed the blue “transmission”
coupled plasma–mass spectrometry this charge transfer did not appear to
luminescence that occurs when a
(LA-ICP-MS) detected trace levels of be taking place. The potential causes
alkali and alkaline metals. of the color are still unknown but
A similar red cordierite, exam- can be linked to manganese (up to
ined in December 2015, was said to 486 ppmw in the orangy red sample Figure 6. This yellow-green 0.75
be from the Iakora district of Fia- and 248 ppmw in the violet sample) ct cape diamond displayed yel-
narantsoa Province in southeast or iron, both of which were detected low phosphorescence to long-
Madagascar (Spring 2016 GNI, pp. by LA-ICP-MS. This orangy red wave UV light.
97–98). The visible absorption spec- cordierite appeared to have roughly
trum we observed was similar to the half as much iron as a saturated vio-
one described by Fritsch et al. (figure let iolite (approximately 25,000 and
5). The spectrum had a decrease in 40,000 ppmw, respectively). No other
absorption in the 400–500 nm region transition metals varied significantly
and low absorption between 600 and between the orangy red and violet
800 nm, resulting in a transmittance samples. Regardless of the origin of
window in the orange to red region color for the orangy red cordierite,
that produced the color of the stone. naming this variety “iolite” could be
These regions of low absorption are misleading according to the defini-
characteristic of the trichroic color tion of the term.
components noted in the stone. We This is the first time GIA has en-
compared this visible spectrum to a countered a cordierite of this color.
GIA reference spectrum of ordinary Augusto Castillo and Akhil Sehgal

VISIBLE-NIR ABSORPTION SPECTRUM
ABSORBANCE
415
477
GR1
400 450 500 550 600 650 700 750 800
WAVELENGTH (nm)
Figure 7. The diamond displayed “cape” absorption peaks at 415 and 477 nm along with a GR1 peak (left). It con-
tained contact crystal inclusions of green omphacite (right; field of view 1.41 mm).
strong light travels through a dia- served throughout the stone regard- Greenish yellow phosphorescence
mond. The UV-Vis spectrum of such less of its bicolor nature. Its central in a chameleon diamond has been
diamonds shows an absorption peak colorless region exhibited weak cape systematically measured using a spec-
at 415 nm and a luminescence peak at lines and strong blue transmission, trometer (see S. Eaton-Magaña et al.,
the lower energy end of the peak. similar to the green diamond in this “Fluorescence spectra of colored dia-
Such is the case with this stone. note. Unlike our sample, the bicolor monds using a rapid, mobile spec-
Nearly 30 years ago, a bicolor dia- diamond also showed very weak yel- trometer,” Winter 2007 G&G, pp.
mond with a near-colorless central low phosphorescence to short-wave 332–351). The peak maximum
portion and light yellow tips was re- UV. A chameleon diamond may dis- recorded for the chameleon diamond
ported in G&G (Winter 1989 Lab play strong yellow phosphorescence was 557 nm. We used an Ocean Optics
Notes, p. 237). It showed very weak to a long-wave UV light source (Sum- USB2000 charge-coupled device (CCD)
chalky yellow phosphorescence for mer 1992 Lab Notes, p. 124; Spring spectrometer similar to the one de-
approximately 10 seconds after the 2000 Lab Notes, pp. 60–61). Our sam- scribed in Eaton-Magaña et al. (2007),
long-wave UV lamp was turned off. ple was not a chameleon diamond, but with a different UV source. In
This yellow phosphorescence was ob- however. place of a deuterium source, we used
Figure 8. Phosphorescence spectra (left) show a peak maximum at 570 nm, which is responsible for the cape dia-
mond’s yellow phosphorescence. Shown on the right is the phosphorescence decay curve at 570 nm.
PHOSPHORESCENCE SPECTRA PHOSPHORESCENCE DECAY CURVE OF 570 NM PEAK

1600
1200
1400
INTENSITY (COUNTS)
INTENSITY (COUNTS)
1000
1200 570
1000 800
800
600
600
400
400
200
200 2
6
E(s)
10
14
TIM
0 0
400 500 600 700 800 900 1000 0 5 10 15 20 25
WAVELENGTH (nm) DECAY TIME (s)

GIA’s new LED desktop long-wave muth-based glass (“Ruby treatments
UV (365 nm excitation) light source. – revisited, with Mozambique’s per-
A broad peak at approximately 570 spective,” Lab Information Circular,
nm, which is responsible for the yel- Gem Testing Laboratory, Vol. 63, De-
low phosphorescence, was observed cember 2011). Due to the large differ-
in the emission spectra (figure 8, left). ence in refractive indices between the
The phosphorescence decay curve at bismuth-based glass filled area and
570 nm, representing the rate of de- the surrounding corundum, this treat-
creasing intensity with time, is plot- ment was easier to visually detect
ted in figure 8 (right). Half-life is than lead-glass filling (figure 11).
defined as the time required for the Fracture filling with lead glass and
initial peak intensity to decrease to Figure 10. With the aid of fiber- non-lead glass has been widely used
one-half its original value (again, see optic illumination, magnification in low-grade ruby (S.F. McClure et al.,
Eaton-Magaña et al., 2007). The half- revealed these trapped gas bub- “Identification and durability of lead
life value measured for our sample bles of various shapes and sizes glass–filled rubies,” Spring 2006
was 5.0 seconds. in the light pink sapphire. Field of G&G, pp. 22–34). Pink sapphires such
Yellow phosphorescence is a very view 2.00 mm. as this 2.77 ct stone are less com-
rare feature in cape diamonds. The de- monly encountered in the laboratory
fect responsible for this optical feature for examination and identification.
remains unknown. fractures. The stone’s RI of 1.76 and
Ratima Suthiyuth
UV fluorescence reaction (medium red
Kyaw Soe Moe, Jon Neal, and
to long-wave and weak red to short-
Paul Johnson
wave) were consistent with corun-
dum. Its SG of 3.87 was lower than First Observation of H4 Defect in
that of corundum and possibly a result CVD SYNTHETIC DIAMOND
Pink SAPPHIRE Filled with of the filler. Due to the sapphire’s very Diamond consists of a relatively sim-
Bismuth-Based Glass light color, its visible-range absorption ple lattice of carbon atoms wherein
Recently, GIA’s Bangkok laboratory spectrum was difficult to observe. only a few impurities can substitute
received a 2.77 ct light pink sapphire Qualitative chemical analysis into the structure. Despite the appar-
(figure 9) for an identification report. using energy-dispersive X-ray fluores- ent simplicity, the diversity of defects
The sample was a semi-transparent cence (EDXRF) spectroscopy showed that can exist in the diamond lattice
cabochon with natural inclusion fea- the presence of bismuth in addition to is astounding. Nitrogen is the most
tures: silk needles and particles, the expected Al, Cr, Fe, and Ga. No abundant impurity in diamond; one of
twinned planes, and thin films. Also lead was detected in the sample, even the defects commonly seen in natural
observed throughout the stone were with the aid of the thick palladium fil- diamonds, the H4 defect, consists of
numerous gas bubbles of various ter. The chemical composition of the four nitrogen atoms and two vacan-
shapes and sizes that were trapped filler indicated that the stone had un- cies. It is strongly associated with B-
along surface-reaching fractures and dergone fracture filling using bis- aggregated nitrogen impurities, which
cavities (figure 10). These indicate the consist of four nitrogen atoms sur-
presence of glass filling. However, no rounding a single vacancy. H4 is cre-
flash effect was evident in the filled ated through a fairly straightforward
Figure 11. Under reflected illumi- formation mechanism where a single
nation, the pink sapphire’s large vacancy is trapped adjacent to a B-ag-
Figure 9. This 2.77 ct light pink sap- cavity filling was readily visible gregate of nitrogen. H4 is often seen in
phire cabochon was treated with a with the unaided eye. Field of irradiated and annealed diamonds
bismuth-based glass filling. view 2.70 mm. with suitable amounts of B-form ni-
trogen, and is a mature aggregate of ni-
trogen that is very difficult to achieve
in synthetic diamond growth or even
post-growth treatment.
We were very surprised to observe
an H4 peak at 496 nm in a pink syn-
thetic diamond grown by chemical
vapor deposition (CVD) that was re-
cently submitted to GIA’s Carlsbad
laboratory for a synthetic colored dia-
mond grading report. This 0.26 ct

was grown using gas that was doped
with nitrogen, allowing its incorpora-
tion into the diamond lattice as single
substitutional nitrogen. The pink
color was produced by NV– centers
created during post-growth irradiation
and annealing treatment. Nitrogen ag-
gregation from single substitutional
nitrogen to the more complex B-form
nitrogen requires very high tempera-
tures and pressures: 2500°C and 9.5
Figure 14. These four near-color-
GPa for several hours (I. Dobrinets et
less round brilliant melee were
al., HPHT-Treated Diamonds: Dia-
identified as HPHT synthetic dia-
Figure 12. This 0.26 ct pink syn- monds Forever, Springer Science &
monds lacking the typical Si- and
thetic diamond contains the first Business Media, 2013, p. 41) and gen-
Ni-related defects.
H4 defect GIA has observed in a erally requires a high starting concen-
CVD synthetic diamond. tration of nitrogen. This CVD
synthetic diamond had low total ni-
have created significant concern in the
trogen content and was grown under
industry, and various screening meth-
vacuum and likely irradiated and an-
round brilliant (figure 12) was type Ib ods have been developed to take into
nealed at atmospheric pressure, mak-
(single substitutional nitrogen impu- account their most reliable distin-
ing the occurrence of H4 even more
rities) with a very low total nitrogen guishing features. In general, the spec-
puzzling. It is possible that the H4 was
content and no detectable A or B form imens described in previous Lab Notes
created during the irradiation and an-
nitrogen. FTIR spectroscopy also (e.g. Summer 2015, pp. 183–184) were
nealing process, but among the many
showed a peak at 3107 cm–1, indicat- very small round faceted goods ranging
samples GIA has examined, this was
ing the presence of a defect consisting between 0.005 and 0.01 ct.
the first observation of this defect in a
of three nitrogen atoms surrounding a Recently GIA’s Bangkok laboratory
CVD synthetic diamond.
vacancy along with a hydrogen atom. received four loose near-colorless
In addition to the observation of Troy Ardon and round brilliants for quality assurance
the H4 defect (figure 13, left), the syn- Christopher M. Breeding service (figure 14). They weighed be-
thetic nature of the sample was re- tween 0.07 and 0.09 ct, slightly more
vealed by the 737 nm SiV– defect in than previously submitted melee. All
the photoluminescence (PL) spectrum HPHT Synthetic Diamond Melee were identified as type IIb HPHT-
(figure 13, right). The presence of (sub- Without Si or Ni Defects grown material using FTIR spectrom-
tle) growth layers in the Diamond- Near-colorless synthetic diamond etry. Their most interesting feature
View fluorescence images provided melee grown using the high-pressure, was the absence of silicon- or nickel-
further confirmation. This diamond high-temperature (HPHT) method related emission peaks when analyzed
Figure 13. Photoluminescence spectra of the pink CVD synthetic diamond showed the occurrence of an H4 defect
(left) as well as a SiV– defect (right).
PL SPECTRUM (488 NM LASER EXCITATION) PL SPECTRUM (633 NM LASER EXCITATION)
260000
INTENSITY (COUNTS)
INTENSITY (COUNTS)
800
240000
600
220000
400
200000
200 H4 defect SiV– doublet
180000
0
160000
495 496 497 498 499 735 740

PL SPECTRA
200000
160000
Diamond Diamond Raman
180000
140000 Raman
INTENSITY (COUNTS)
INTENSITY (COUNTS)
160000
120000 140000
120000
100000
100000
80000
80000
60000 60000
584.4 40000
40000 NV0
20000
20000
0
0 -20000
525 575 625 675 725 775 825 850 870 890 910 930 950
Figure 15. The PL spectra features observed at liquid nitrogen temperature lacked the 736.6/736.9 nm emission
doublet related to the SiV– defect (left, 514 laser excitation) and the Ni-related defect doublet at 883/884 nm (right,
830 nm laser excitation).
at liquid nitrogen temperature by PL microscope revealed metallic and rod- HPHT Synthetic Diamond with
spectroscopy (figure 15). The spectra like inclusions (figure 17) that also as- Intense Green Color
typically observed in HPHT synthetic sisted in identifying these melee as HPHT-grown synthetic diamonds are
diamonds show an emission doublet synthetic diamonds. predominantly colorless, yellow, and
related to the negatively charged sili- The study of these four melee-size blue, depending on the control of im-
con split-vacancy defect SiVˉ at HPHT synthetic diamonds revealed purities during growth. Other colors
736.6/736.9 nm and an associated Ni- atypical PL spectra. However, other are very rare. In this study, we report
related defect at 883.0/884.7 nm (Sum- means of identification (inclusions, on an intensely colored green HPHT
mer 2015 Lab Notes, pp. 183–185). growth patterns as seen via the Dia- synthetic diamond identified in the
The growth patterns observed in mondView, and their phosphorescence New York laboratory.
the DiamondView were an identifying reactions) still played an important This square-cut synthetic dia-
feature that did remain the same. The role in their identification. It is inter- mond, which weighed 0.42 ct and
characteristic angular growth patterns esting to note that HPHT synthetic measured 4.42 × 4.34 mm with a
and blue-green fluorescence, together melee diamonds being submitted to depth of 2.59 mm, was graded as
with an associated green phosphores- GIA for identification are getting Fancy Deep green (figure 18). Initial
cence, still provide important evi- larger, indicating the continuous im- observation revealed strong color zon-
dence that aid in the separation of provement of synthetic technology. ing with narrow, colorless growth sec-
natural and synthetic diamonds (figure Piradee Siritheerakul and tors on the pavilion that were visible
16). Examination with a gemological Wuyi Wang without magnification. An abundance
of small metallic inclusions observed
under magnification (figure 19) caused
the sample to exhibit magnetism.
Figure 16. These DiamondView Magnetism is a common but not diag-
fluorescence and phosphores- Figure 17. Darkfield illumination nostic feature in HPHT synthetic dia-
cence images of a 0.081 ct sample reveals rod-like inclusions in a monds. Natural diamonds may also
show growth patterns character- 0.087 ct HPHT-grown melee dia- exhibit magnetism, though only in ex-
istic of HPHT synthetic diamond. mond. tremely rare cases (G.R. Rossman and
J.L. Kirschvink, “Magnetic properties
0.1 mm of gem-quality synthetic diamonds,”
Fall 1984 G&G, pp. 163–166).
Upon further examination, the di-
amond was determined to be type IIb
(containing the impurity boron), as
seen from its mid-FTIR absorption
spectrum (figure 20). The concentra-
tion of uncompensated boron in this
diamond was 0.013 ± 0.003 ppma.

FTIR SPECTRUM
ABSORBANCE
Boron
1332
4500 4000 3500 3000 2500 2000 1500

WAVENUMBER (cm–1)
Figure 20. The green synthetic diamond’s mid-FTIR spectrum, which de-
termined that the material was type IIb, showed an uncompensated
Figure 18. This 0.42 ct green boron concentration at ~0.013 ppma.
HPHT synthetic diamond dis-
played strong growth sectors
that were observed without Gemology, No. 15, 2015). Diamond- strong nickel-related absorption, with
magnification. View imaging revealed the typical a band observed at ~685 nm (figure 22).
hourglass growth structure of HPHT This center, due to a Ni+ interstitial
synthetic diamonds (figure 21), along vacancy, creates a “transparent” win-
This amount, though detectable, did with extremely strong phosphores- dow at approximately 555–585 nm, re-
not contribute to the green color. cence for which the presence of boron sulting in an observed green color. This
Many synthetic diamonds grown by might be responsible. Noted in the diamond was most likely HPHT
New Diamond Technology (NDT) in FTIR spectrum was an absorption fea- treated post-growth to activate the Ni+
Russia reportedly also contain this ture related to nitrogen impurity, a rel- and cause this intense absorption. It is
level of boron impurity, as it is incor- atively strong peak at 1332 cm–1 caused very rare to see HPHT synthetic dia-
porated during the growth process (B. by N+. It is unusual to see such features mond with an intense green color at-
Deljanin et al., “NDT breaking the 10 in today’s synthetic diamonds. A UV- tributed to a nickel-related impurity.
carat barrier,” Contributions to Vis-NIR absorption spectrum showed In addition to boron and nickel,
this synthetic diamond also contained
the SiV– impurity in its crystal lattice.
Figure 19. The green synthetic diamond displayed strong magnetism, Recent studies have shown that while
caused by abundant metallic inclusions. Field of view ~1.4 mm. this impurity is common in CVD-
grown diamonds, it may also be ob-
served in HPHT-grown diamonds
(U.F.S. D’Haenens-Johansson et al.,
“Near-colorless HPHT synthetic dia-
monds from AOTC Group,” Spring
2014 G&G, pp. 30–45). The distribu-
tion of Si and Ni impurities were
mapped using Raman mapping tech-
niques, and the results proved consis-
tent with previous studies of near-
colorless HPHT-grown diamonds also
observed in the trade (P. Johnson et al.,
“Discovery and distribution of the [Si-

{111}
{111}
{111}
{111}
1 mm 1 mm
633 (SiV–) 785 (Ni+)
Figure 21. An oriented DiamondView image (left) and 633 and 785 nm PL maps (right) show the SiV– and Ni+ de-
fects confined to the {111} growth sectors.
V]- defect in HPHT-grown gem-qual- contribute to the color of these syn- synthetic diamond. This is the first
ity diamonds,” 2015 GSA Annual thetic diamonds, and it is either in- instance in which nickel has been re-
Meeting, paper No. 300-12), with both tentionally or unintentionally ported to contribute to the deep green
the Ni+ and SiV– confined to the {111} incorporated during crystal growth color of a synthetic diamond. This
growth sectors. Typically observed in (again, see figure 21). finding could indicate a new develop-
HPHT-grown diamonds, these {111} These spectroscopic and gemolog- ment in growth technology and a new
sectors are the most developed and ical features, with the exception of material making its way into the dia-
dominate the diamond crystal’s mor- the green color resulting from nickel, mond trade. This example was sub-
phology. The SiV– center does not were consistent with typical HPHT mitted as an undisclosed synthetic
diamond and was subsequently issued
an identification report.
Figure 22. The HPHT synthetic diamond’s UV-Vis-NIR spectrum dis- Developments in synthetic dia-
played a very intense Ni-related absorption. mond growth technologies (both CVD
and HPHT) continue to emphasize
the need for careful observation and
UV-VIS -NIR SPECTRUM
testing techniques to identify the new
types of materials encountered in the
gem trade today.
Paul Johnson and Elina Myagkaya
ABSORBANCE
Ni Multi-Elemental Diffused and Melt-

Grown SYNTHETIC SAPPHIRE
GIA’s Carlsbad laboratory examined a
4.21 ct faceted blue oval. Standard
gemological testing determined that
the stone was uniaxial negative and
had an RI of nε = 1.759 and nω = 1.769,
with a corresponding birefringence of
0.010. Hydrostatic SG was measured
400 500 600 700 800 900 as 4.00. The stone was inert to long-
WAVELENGTH (nm) wave UV and showed medium chalky
blue color to short-wave UV. These

ties. In melt-grown corundum, the
high growth temperature allows Ti4+
to pair with Al vacancies more easily
than other charge compensating ions;
this creates the chalky fluorescence.
In contrast, Pairing of Ti4+ with other
ions (usually Fe2+ or Mg2+) and the
presence of Fe3+ ions prevent fluores-
cence. That could explain why the
chalky fluorescence was not present
where the blue diffusion color was
concentrated (R. Hughes and J. Em-
mett, “Heat seeker: UV fluorescence
as a gemological tool,” 2005,
www.r uby-sapphire.com/heat_
seeker_uv_fluorescence.htm).
LA-ICP-MS analysis was per-
formed on the girdle of the stone.
Three laser ablation spots were
drilled six times down into the stone.
Figure 23. Top left: Color concentration was seen along the pavilion facet The concentration vs. depth profile is
junctions of the synthetic sapphire, and blue halos were shown on the shown in figure 24. From the surface
table facet. Top right: A spherical gas bubble was observed. Bottom left: of the stone to around 180 µm depth,
Numerous clouds were visible throughout the stone. Some formed finger- the Be concentration decreased from
print-like patterns. Bottom right: Under the DiamondView's short-wave 76.8 to 13.9 ppma, Mg from 7.8 to 0.8
UV radiation, blue fluorescence was absent where there were concentra- ppma, Ti from 580 to 391 ppma, Fe
tions of the diffusion color in the stone. Fields of view: 9.61 mm (top left), from 390 to 83.4 ppma, and Ga from
1.25 mm (top right), 2.9 mm (bottom left), and 8.58 mm (bottom right). 6.12 to 1.36 ppma. The results con-
firmed diffusion treatment, not only
by Ti and Fe but also by Be, Mg, and
Ga. The presence of Mo provided
properties were consistent with iodide under crossed polarizers and some indication of synthetic origin.
heated sapphire, either natural or syn- viewing down the optic axis, no Plato In a GIA identification report, the
thetic. Under magnification and dif- lines were observed (see W. Plato, stone was described as titanium and
fused transmitted lighting, a blue “Oriented lines in synthetic corun- beryllium diffused. It is the first time
color concentration was observed dum,” Fall 1952 G&G, pp. 223–224). GIA has examined a multi-elemental
along the facet junctions on both the The absence of Plato lines does not diffused and melt-grown synthetic
pavilion and crown facets, along with rule out flame-fusion growth. The sapphire.
a few large blue halos on the table heat from diffusion treatment can re-
Ziyin Sun, Jonathan Muyal, and
facet (figure 23, top left), which indi- duce the visibility of Plato lines (S.
Nicole Ahline
cated that the stone had been tita- Elen and E. Fritsch, “The separation
nium diffused (R.E. Kane et al., “The of natural from synthetic colorless
identification of blue diffusion-treated sapphire,” Spring 1999 G&G, pp. 30–
sapphire,” Summer 1990 G&G, pp. 41). More advanced testing is needed TOURMALINE
115–133; Summer 2015 GNI, pp. 203– to reveal all the physical and chemical Paraíba Tourmaline with Unusual
205). Under darkfield lighting, indi- properties of this unique stone. Coating
vidual large spherical gas bubbles Under the short-wave radiation (a Paraíba tourmaline, named after the
were present (figure 23, top right). Nu- 20-nm-wide band centered at 225 nm) Brazilian state where it was discov-
merous clouds observed throughout in a DiamondView device, the stone ered in the late 1980s, is a precious
much of the stone could have easily displayed a strong chalky blue fluores- copper-bearing tourmaline variety
been mistaken for natural clouds. cence. Hughes and Emmett (2005) (e.g., J.E. Shigley et al., “An update on
Some of them grouped to form natu- concluded that the chalky blue fluo- ‘Paraíba’ tourmaline from Brazil,”
ral-looking fingerprint-like inclusions rescence was due to isolated Ti4+ ions, Winter 2001 G&G, pp. 260–276).
(figure 23, bottom left). The stone’s in- or Ti-Al vacancy pairs. The strength Treatments such as heating and clar-
ternal characteristics indicated a of the chalky fluorescence depends on ity enhancement have been described
melt-grown synthetic origin. How- the growth temperature and Ti4+ con- (e.g., S.F. McClure et al., “Gemstone
ever, when immersing in methylene centration relative to other impuri- enhancement and its detection in the

CONCENTRATION VS. DEPTH PROFILE
spectroscopy were performed to con-
1000
firm the presence of copper within the
CONCENTRATION (ppma)
15 material, as well as the material’s
100
Mg cause of color.
Ga
Viewed face-up with the unaided
10
eye, the stone appeared normal. How-
CONCENTRATION (ppma)
Ti
Fe
Be ever, the coated surface looked differ-
10
1 ent under reflected lighting and
0 30 60 90 120 150 180
DEPTH (μm) darkfield illumination (figure 26, left).
A brassy, uneven submetallic luster
was observed under reflected light,
while whitish features in the sub-
metallic luster area (figure 26, right)
5
appeared under darkfield illumina-
tion. Coatings typically show signs of
wear due to their low durability. This
is often seen at the facet junctions.
This coating did not show signs of
0 wear and could not be removed using
0 30 60 90 120 150 180
a standard ink pen eraser.
DEPTH (μm) LA-ICP-MS chemical analyses de-
tected elements such as B, Li, and Na,
indicating elbaite tourmaline. How-
Figure 24. The synthetic sapphire’s concentration vs. depth profile re- ever, strong early (near-surface) peaks
vealed that concentrations of Be, Mg, Ti, Fe, and Ga decreased signifi- of elements Pt, W, and Mo, which are
cantly when the laser ablation spots went deeper into the stone. not normally contained in natural
tourmaline, were detected. We deter-
mined that this chemistry difference
probably resulted from the submetal-
2000s,” Fall 2010 G&G, pp. 218–240), consistent with tourmaline. The lic surface coating. As with Pt- and Pt-
but coating has never been reported. stone showed medium to strong related coatings, it is possible that the
A Paraíba tourmaline ring with an dichroism displaying greenish blue coating was used to improve abrasion
unusual coating was recently submit- and light blue colors. Natural inclu- resistance (e.g., K. Schmetzer, “Sur-
ted to GIA’s Tokyo laboratory for sions such as trichites and finger- face treatment of gemstones, espe-
identification and origin service (fig- prints were observed. We identified cially topaz – an update of recent
ure 25). Standard gemological proper- the stone as tourmaline using Raman patent literature,” Journal of Gem-
ties revealed an RI of 1.640–1.620, spectroscopy. EDXRF and UV-Vis mology, Vol. 31, No. 1, 2008, pp. 7–13)
Figure 25. The Paraíba tourma- Figure 26. Left: An uneven submetallic coating with an irregular pattern is
line in this ring had an unusual revealed with reflected light. The white dashed box represents the area
submetallic surface coating. shown in the right photo. Field of view 3.33 mm. Right: Under brightfield
illumination, a part of the left table facet shows whitish surface struc-
tures surrounding the metallic luster domains. Field of view 1.31 mm.

or to assist with scanning electron mi-
croscopy (SEM) observation (e.g., I.
Stokroos et al., “A comparative study
of thin coatings of Au/Pd, Pt and Cr
produced by magnetron supporting
for FE-SEM,” Journal of Microscopy,
Vol. 189, No. 1, 1998, pp. 79–89).
The reason for applying this coat-
ing is unknown. This is the first time
GIA’s laboratory has observed coated
Paraíba tourmaline.
Kazuko Saruwatari and Figure 27. Left: Surface-reaching growth tubes and fractures associated
Philip G. York with strong pink sleeves, whose color became weaker as they penetrated
the stone’s interior; field of view 2.83 mm. Right: Looking down the length
of the growth tubes, bleeding of pink color was observed; field of view
RUBELLITE with Incredibly Strong 1.13 mm.
“Pink Sleeves”
GIA’s Hong Kong laboratory recently
examined a purplish red cabochon-cut It is different from dye concentra- knowledge, the incredibly strong pink
rubellite tourmaline. Standard gemo- tions, which would only be located sleeves seen in this rubellite have not
logical tests yielded a spot RI of 1.64 inside the surface-reaching growth been previously reported in either
and SG of 3.06, which are consistent tubes and fractures. Paraíba or rubellite tourmaline.
with tourmaline. Magnification re- Radiation-induced pink sleeves Xiaodan Jia and Mei Mei Sit
vealed strong doubling, fluid inclu- surrounding surface-reaching growth
sions, and parallel growth tubes, all tubes and fractures have been re-
typical of tourmaline. What made this ported in Paraíba tourmalines from
specimen unusual were its localized Mozambique. The formation is be-
zones of strong pink coloration sur- lieved to be due to a radioactive solu- PHOTO CREDITS:
rounding surface-reaching growth tion entering the growth tubes and Robison McMurtry—1, 12; Jian Xin (Jae)
tubes and fractures (figure 27, left). fractures by capillary action, and the Liao—2, 18; Augusto Castillo—4; Sood Oil
Bleeding of pink color from the presence of pink sleeves proves that (Judy) Chia—6; Kyaw Soe Moe—7 (right);
growth tubes into the body of the the stone has not undergone heat Nuttapol Kitdee—9, 14; Charuwan Khow-
stone was best observed looking down treatment. (J.I. Koivula et al., “Solu- pong—10, 11, 16, 17; Ungkhana Atikarn-
the length of the growth tubes (figure tion-generated pink color surrounding sakul—17; Paul Johnson—19, 21; Jonathan
27, right); the pink color gradually growth tubes and cracks in blue to Muyal—23; Nicole Ahline—23; Masumi
Saito—25; Kazuko Saruwatari—26; Xiaodan
faded away into the surrounding tour- blue-green copper-bearing tourma-
Jia—27.
maline host. This type of zoning has lines from Mozambique,” Spring 2009
come to be known as “pink sleeves.” G&G, pp. 44–47). To the best of our

Editor
Nathan Renfro
Contributing Editors
Elise A. Skalwold and John I. Koivula
Ametrine Optical Dishes: Windows into the

Effects of Crystal Structure
During exploration of a gem’s interior, it is important to un-
derstand how crystal structure affects the incorporation and
orientation of inclusions, how an inclusion can affect its
host, how interpretation of observations can depend on
point of view, and the nature of light as it passes through a
crystal—host or inclusion. As proposed earlier in this col-
umn (Spring 2016, p. 81), polished inclusion study blocks
are ideal for showcasing, studying, and reviewing basic con-
cepts relating to the micro-world. Here we take a look at
light and crystal structure.
Of all the varieties of quartz, ametrine is perhaps the
most fascinating to observe. While its stunning color zon-
ing makes it a popular designer gemstone, it also encom-
passes several optical properties unique to different quartz Figure 1. Left: An 85 × 25 mm ametrine crystal from
varieties that can all be explored in one specimen. Almost Bolivia’s Anahí mine exhibits the deeply etched prism
unrecognizable as quartz, ametrine crystals are deeply faces typical of this quartz variety. Right: The crystal’s
etched (figure 1, left) and have abundant healed fractures. termination, viewed in the direction of its c-axis, dis-
The only clues to the trigonal symmetry typical of quartz plays a rare “Y”-shaped manifestation of the crystal’s
are occasional rhombohedral faces and, very rarely, an enig- trigonal symmetry. Slices made perpendicular to this
matic flat termination manifesting a three-armed figure direction will reveal this symmetry with alternating
(figure 1, right). Only when a basal section is made by cut- amethyst and citrine sectors. Since the optic axis par-
ting a slice perpendicular to the c-axis is the improbable allels the c-axis located at the intersection of the “Y”
yet beautiful combination of citrine and amethyst revealed. arms, all paths of light found across this termination
face will be along the optic axis. Photos by Elise A.
Skalwold.
About the banner: In this scene, pyrite inclusions have dissolved, leaving
behind hematite and goethite, both of which react dramatically to polarized
light. Photomicrograph by John Koivula; field of view 10 mm. Beauty and function are married in an ametrine study
Editors’ note: Interested contributors should contact Nathan Renfro at block oriented so that the view through its two largest faces
[email protected] and Jennifer-Lynn Archuleta at [email protected] is directly along the optic axis of a crystal, as seen in figure
for submission information. 1. The innovation of adding a set of concave dishes for sam-
GEMS & GEMOLOGY, VOL. 53, NO. 1, pp. 102–108. pling each color zone, as well as the borders between them,
© 2017 Gemological Institute of America reveals truly remarkable optical phenomena (figures 2 and
3). These optical dishes allow the viewer to see light passing
102 MICRO-WORLD GEMS & GEMOLOGY SPRING 2017

Figure 2. An ametrine study block with concave Figure 3. Placing the study block between crossed
dishes cut and polished into its surface, including one polarizing filters in transmitted light results in a va-
on the opposite side. Viewed in transmitted light riety of bull’s-eye optic figures that form in the di-
along the optic axis direction (i.e., all paths parallel to vergent light produced by the concave dishes. The
the c-axis), as in figure 1 (right), the trigonal symmetry citrine sectors under the minor rhombohedral faces
of the crystal defines the citrine and amethyst sectoral (z) are free of twinning and show a classic bull’s-eye
zoning and allows observation of the effects of the optic figure with varying colors due to slightly dif-
optic axis on light. This specimen was fashioned by ferent thicknesses. The amethyst sectors display an
Nathan Renfro. Photo by Kevin Schumacher. Airy’s spiral (middle row, right dish) produced by
the alternating layers of left- and right-handed
quartz of the Brazil-law twinning that formed under
the crystal’s major rhombohedral faces (r); broadly
through the specimen in many directions, as does the strain- angled colors in these latter areas are also indica-
free glass sphere familiar to gemologists for use with a po- tive of this twinning. The dish on the upper right is
lariscope (figure 4). When viewed between crossed polarizing a distorted figure on the border of the sectors. Photo
filters, these dishes dramatically display the different uni- by Kevin Schumacher.
Figure 4. Left: A section through part of the study axial optic figures unique to quartz, which vary depending
block. The dish serves as a concave (negative) lens on thickness and placement in the sectors: the classic
that causes the light rays to diverge. Right: A gemolo- “bull’s-eye” optic figure seen in the untwinned citrine sec-
gist’s strain-free glass sphere is used to observe an in- tors, an Airy’s spiral in the amethyst sectors, and a distorted
terference figure. Although the rays first converge at figure produced in a border region. All of these observed phe-
the base of the sphere, they ultimately diverge just nomena are produced by optical activity (i.e., the rotatory
like the rays in the figure on the left. Illustration by dispersion of light as it travels along the optic axis direction
William A. Bassett. within the quartz crystal). By analyzing these figures, we can
determine handedness and the effects of thickness, while
observations of the amethyst’s Brazil-law twinning may help
distinguish natural versus synthetic origin.
a b For an in-depth analysis of the phenomenon of optical
activity and the origin of quartz’s optic figure, see E.A.
Skalwold and W.A. Bassett, Quartz: A Bull’s Eye on Op-
tical Activity, Mineralogical Society of America, 2016,
www.minsocam.org/msa/OpenAccess_publications/
Skalwold/Quartz_Bullseye_on_Optical_Activity.pdf.
Elise A. Skalwold and William A. Bassett
Ithaca, New York
MICRO-WORLD GEMS & GEMOLOGY SPRING 2017 103

Figure 5. This 6.92 mm bastnäsite-(Ce) from Pakistan Figure 6. Magnification affords a more detailed view of
contains numerous straight and curving acicular in- the bastnäsite-(Ce) inclusions. Photomicrograph by
clusions of astrophyllite. Photo by Kevin Schumacher. Nathan Renfro; field of view 2.85 mm.
Astrophyllite in Bastnäsite-(Ce) ure 6), both straight and curved, that were identified by
Pakistan’s Zagi Mountains are known as a source for both Raman analysis. Some of the included fibers and stalks ap-
astrophyllite and bastnäsite-(Ce). The two minerals often pear to have been partially altered or coated with iron com-
grow together, with bastnäsite-(Ce) overgrowing stalks and pounds.
fibers of astrophyllite; the latter is then sometimes incor- John I. Koivula
porated as inclusions in the bastnäsite-(Ce). We refer to GIA, Carlsbad
these types of inclusions as “protogenetic.”
The author recently had the opportunity to examine
the brownish orange bastnäsite-(Ce) seen in figure 5. The Flashes and Flames in Ethiopian Opal
1.40 ct oval modified brilliant, provided by Luciana Barbosa The relatively new source for precious opal located in
(Gemological Center, Asheville, North Carolina), hosted Ethiopia’s Wollo Province produces material with vibrant
an abundance of eye-visible inclusions of astrophyllite (fig- play-of-color and occasionally with interesting inclusions.
Figure 7. This 2.68 ct opal displayed play-of-color when viewed face-up (left) and flame-like chatoyancy on its
slightly domed base (right). A more pronounced dome might have had the effect of sharpening the sheen into a
cat’s eye. Photos by Kevin Schumacher.

A recently examined 2.68 ct cabochon contained a scatter-
ing of octahedral metal sulfide inclusions (relatively com-
mon in Ethiopian opal) along with an unusual combination
of optical phenomena: play-of-color and chatoyancy. Face-
up, the stone showed the typical flashes of colors one
would expect in precious opal (figure 7, left). However, the
back of this cabochon contained an interesting structure
that produced a sheen-like chatoyancy over its slightly
rounded surface (figure 7, right). This chatoyant zone was
strikingly similar to the flame structure seen in some non-
nacreous pearls, which is produced by an intricate struc-
ture of aragonite crystal laths (figure 8). In the case of this
opal, the flame-like structure’s cause remains an enigma,
though it could be due to the structural influence of a pre-
existing fibrous mineral replaced by opal. Unfortunately,
Raman analysis was unable to detect any presence of an-
other mineral phase in the chatoyant area, so the exact
cause of chatoyancy is still unknown.
Chatoyancy is not unheard of in opal, though it is quite
rare in precious material. Cat’s-eye common opal from Tan-
zania owes its optical effect to arrays of parallel needles,
while the asterism and chatoyancy unique to the precious Figure 9. With fiber-optic illumination, a sharp,
opal of Spencer, Idaho, are due to the arrangement of the bright pink chatoyant band extends across the sur-
same silica micro-spherules responsible for its play-of-color. face of a 1.89 ct dark brown button-shaped imita-
At least one cat’s-eye precious opal was observed to even tion pearl perpendicular to its internal fibrous
“wink” at the viewer when viewed with opposing lighting structure. Photo by Robert Weldon/GIA.
(Summer 2003 Lab Notes, p. 148). This Ethiopian opal cabo-
chon joins the rarefied group of chatoyant precious opal
with its unusual phenomenal display of flashes and flames.
Nathan D. Renfro Imitation Cat’s-Eye Pen “Pearls”
GIA, Carlsbad
Chatoyancy is an optical phenomenon commonly seen in
Elise A. Skalwold gemstones that have an abundance of parallel solid or hol-
low long acicular inclusions in which an eye-like band is
produced when viewed in strong light. Fine cat’s-eye gem-
stones of various mineral species are prized; however, nat-
Figure 8. The fibrous nature of the opal’s unusual cha- ural chatoyancy in pen pearls would be an extraordinary
toyant zone is clearly revealed, along with the numer- occurrence. During a recent study on reportedly non-
ous minute octahedral black metal sulfide crystals nacreous pearls from the Dr. Eduard J. Gübelin Collection,
that are a common inclusion in opal from Wollo. the author found three dark brown button-shaped samples
Oblique fiber-optic illumination. Photomicrograph by stated to be from a Pinna species mollusk. All three sam-
Nathan Renfro; field of view 2.73 mm. ples displayed an attractive cat’s-eye effect when illumi-
nated under a fiber-optic light or any single white light
(figure 9). In each specimen, the chatoyant band manifested
as a single bright pink line extending across the surface in
a direction perpendicular to the internal fibrous structure.
Evidence of human manipulation includes fine linear
features present in random directions across cellular struc-
tures; these are proof of work carried out during the fashion-
ing of the samples into cabochons. Microscopic examination
revealed the fibrous structure as numerous well-formed,
minute parallel acicular crystals. Infrared spectroscopy sub-
sequently confirmed the crystals to be calcite. The calcite
was arranged in a vertical columnar orientation rather than
in a radial and concentric structure characteristic of non-
nacreous Pinna (pen) pearls, which led to further doubts con-
cerning the “pearl” identity of the samples.

Figure 10. Fluorescence images reveal banded features, typical of shell, at right angles to the vertical columnar
structure of the calcite crystals (left). A mosaic pattern (right) is apparent looking down the hexagonal cross-
section of the calcite columns. Photomicrographs by Artitaya Homkrajae; fields of view 2.88 mm and 1.44 mm,
respectively.
Microradiography of all three samples revealed tight in- reveal a hexagonal cross-section, creating a striking mosaic
ternal structures. The obvious radial structures that would or cellular pattern (figure 10, right). The cells varied in
be expected in pen pearls were not observed (N. Sturman form, and their approximate diameters ranged from 15.50
et al., “Observations on pearls reportedly from the Pin- to 38.00 microns.
nidae family (pen pearls),” Fall 2014 G&G, pp. 202–215). These samples were appealing for the magnificent kalei-
Based on its external and internal features, this material doscopic colors that were seen using fiber-optic illumination
was almost certainly the same type of imitation pearl pre- (figure 11). The semitranslucent to translucent nature of the
viously reported in G&G (Winter 2011 GNI, pp. 330–332). crystals allowed light to pass through and produce vibrant,
Fluorescence images taken of one specimen using a po- colorful scenes owing to the interference of the light rays that
larizing microscope and an ultraviolet excitation filter with were reflected and refracted when interacting with the crys-
a wavelength range of 330–380 nm (figure 10, left) reveal a tals. As one can see, exploration of the micro-world is not
prominent banded structure at right angles to the calcite only important to forensics—it can also be quite beautiful.
columns. When viewed at right angles to their lengths Artitaya Homkrajae
under the same conditions, the long, thin calcite crystals GIA, Carlsbad
Figure 11. Under fiber-

optic light, magnificent
reflective kaleidoscopic
colors of pink, yellow,
and green are revealed
from the well-formed
acicular calcite crystals
after mineral oil is ap-
plied to the sample’s
surface. Photomicro-
graph by Artitaya
Homkrajae; field of
view 3.57 mm.

Figure 12. A beautiful iridescent rosette pattern Figure 13. These two large primary fluid inclusions
made up of flattened negative crystals appears containing a black solid phase are associated with
around the primary negative crystals when illumi- fingerprints that have a variable orientation. The
nated by oblique fiber-optic lighting. Photomicro- photo was taken along the c-axis in darkfield illumi-
graph by Victoria Raynaud; field of view 1.07 mm. nation. Photomicrograph by Victoria Raynaud; field
of view 1.07 mm.
Negative Crystals in Sapphires descence under oblique fiber-optic illumination. Other

Negative crystals, or cavities within a crystal, and fluid in- negative crystals associated with healed fissures are ob-
clusions occur in almost every mineral. They are generally served in random orientation. The smaller, more delicate
too small to notice without high magnification, but these fluid inclusions in these structures have a smoother out-
modest features hold great beauty and scientific value. line, while the fingerprints themselves are often associated
These microscopic features are valuable to understanding with the primary negative crystals.
the growth of continents, the rise of mountains, and the The multiphase inclusion in figure 14 contains a gas, a
formation of gems. The following photos show beautiful liquid, and at least one solid. When the specimen was gen-
examples of such inclusions in sapphires from the Baw Mar tly heated by the microscope’s well light, the gas and the
mining area in Mogok, Myanmar. Some of these fluid in- liquid homogenized. Raman spectroscopy identified the
clusions are arranged in a “rosette” pattern consisting of a gas phase as pure CO2 and the opaque solid as marcasite.
large flattened negative crystal surrounded by smaller ones. Using the same method, we also identified diaspore fibers,
This structure is always seen perpendicular to the c-axis seen as faint lines in the inclusions, and a miniscule
(figure 12). The larger primary inclusions like those seen siderite crystal. Although fluid inclusions are common in
in figure 13 are characterized by their high relief and may most blue sapphires, the examples from Baw Mar reveal
display triangular growth markings as well as beautiful iri- these fascinating inclusions in an appealing way.
Figure 14. Left: A large fluid inclusion containing a CO2 gas bubble and a marcasite crystal. The primary inclusion
is surrounded by smaller inclusions that make up a healed fissure. Right: When the sample was gently heated, the
gas bubble disappeared as the gas and liquid homogenized. The sample is illuminated using diffused backlighting.
Photomicrographs by Victoria Raynaud; field of view 1.20 mm.

Analysis of the opaque black plates was a more diffi-
cult matter. GIA’s Carlsbad lab staff employed Raman to
identify the plates, but the inclusion was too deep to get
any signal other than the spectral lines of the topaz host.
Energy-dispersive X-ray fluorescence (EDXRF) was at-
tempted to pick up any hints of the plates’ chemical com-
position, but useful information continued to elude such
efforts. It became clear that destructive analysis would be
needed to clearly identify the black plates. Since this was
an inclusion combination, we decided that this topaz
should be preserved unaltered for future exploration.
Surely one day technological advances will allow the iden-
tification of this mystery inclusion, but for now it is
enough to enjoy the aesthetics of such a beautiful included
crystal and the fascinating exploration of its inner world.
John I. Koivula
Figure 16. While the light green crystal was identified

as fluorite by Raman analysis, the opaque platy in-
clusions were too deep to be identified. Photomicro-
graph by Nathan Renfro; field of view 5.87 mm.
Figure 15. The interior of this 5.10 ct Sri Lankan topaz

is dominated by a dramatic inclusion scene. Photo by
Kevin Schumacher.
These photos were taken as part of a research project

to document the blue sapphires from the Baw Mar mine.
The complete report of this study can be found at
https://www.gia.edu/gia-news-research/blue-sapphires-
baw-mar-mine-mogok-myanmar.
Victoria Raynaud and Wim Vertriest
GIA, Bangkok
Quarterly Crystal: Mystery in Topaz

Occasionally we run into an inclusion identification prob-
lem we just can’t solve. Such was the case with the 5.10 ct
topaz from Elahera, Sri Lanka, shown in figure 15. This
transparent colorless crystal, acquired from Kusum S. Nao-
tunne (Colombo, Sri Lanka), hosts a prominent translucent
light green crystal and a flowery cluster of hexagonal
opaque silvery black plates (figure 16). The bodycolor and
characteristic form of the green inclusion suggest fluorite;
this was confirmed by laser Raman micro-spectrometry.

First Place
JAIPUR, INDIA: THE GLOBAL GEM AND JEWELRY POWER OF THE
PINK CITY
WINTER 2016
Andrew Lucas, Nirupa Bhatt, Manoj Singhania, Kashish Sachdeva,
Tao Hsu, and Pedro Padua
Andrew Lucas is a field gemologist in GIA’s content strategy department in
Carlsbad, California. Mr. Lucas researches and documents the entire mine-to-
market gem and jewelry industry for GIA education; he also presents seminars Andrew Lucas Nirupa Bhatt Manoj Singhania
on colored stones and diamonds. Nirupa Bhatt is managing director for GIA
India and Middle East. Prior to her tenure with GIA, Ms. Bhatt worked for Rio Tinto for 20 years. Manoj Singhania is director of edu-
cation at GIA India and Middle East. Kashish Sachdeva is a gemology instructor at GIA India. Mr. Sachdeva began his career as a mer-
chandiser and subsequently gained experience in gem identification, diamond grading, and gemstone treatments. Tao Hsu is the tech-
nical editor of Gems & Gemology. Dr. Hsu received her PhD in geology from the University of Southern California. Pedro Padua is the
senior video producer at GIA. He received his master’s degree in communications from Loyola Marymount University.
Second Place
AN INTRODUCTION TO PHOTOLUMINESCENCE SPECTROSCOPY
FOR DIAMOND AND ITS APPLICATIONS IN GEMOLOGY
SPRING 2016
Sally Eaton-Magaña and Christopher M. Breeding
Sally Eaton-Magaña is a research scientist at GIA in Carlsbad, where she inves-
tigates the physics of natural, treated, and synthetic diamonds. Dr. Eaton-
Magaña received her PhD in chemical engineering from Case Western
Reserve University. Christopher M. Breeding is a senior research scientist and Sally Eaton-Magaña Christopher M.
manager at GIA in Carlsbad, where he investigates origin of color in diamond Breeding
and other gems. Dr. Breeding, who is a co-editor of G&G’s Gem News International section,
holds a PhD in geology from Yale University.
Third Place
INCLUSIONS IN NATURAL, SYNTHETIC, AND TREATED EMERALD
WINTER 2016
Nathan D. Renfro, John I. Koivula, Jonathan Muyal, Shane F. McClure,
Kevin Schumacher, and James E. Shigley
Nathan Renfro is the analytical manager of the gem identification department
and analytical microscopist in the inclusion research department at GIA in
Carlsbad. John I. Koivula is the analytical microscopist at GIA in Carlsbad. He
is the co-author of the three-volume Photoatlas of Inclusions in Gemstones and Nathan Renfro John I. Koivula Jonathan Muyal
the author of MicroWorld of Diamonds. Jonathan Muyal is a staff gemologist
GIA in Carlsbad, where he also serves as the librarian for the research reference stone collection. Mr. Muyal specializes in pho-
tomicrography of inclusions in gem materials. Shane McClure is director of identification services at GIA in Carlsbad. He is well
known for his many articles and lectures on gem identification. He is also co-editor of G&G’s Lab Notes section. Kevin Schumacher
serves on the G&G production staff as a photographer, videographer, and image specialist. He has 37 years of professional photog-
raphy experience, the past 11 with GIA. James E. Shigley is distinguished research fellow at GIA in Carlsbad. He is the author of
numerous articles on diamonds and colored gemstones. Dr. Shigley holds a PhD from Stanford University.
Thank you to all the readers who voted. In addition to our winning authors, we congratulate
Sandra Wen of Wuhan, China, whose name was randomly drawn from the entries to win a
one-year subscription to G&G.
MOST VALUABLE ARTICLE AWARD GEMS & GEMOLOGY SPRING 2017 109
The following 25 questions are from the four 2016 issues of G&G. Refer to
the articles in those issues to find the single best answer for each question.
Mark your choice on the response card provided in this issue or visit
gia.edu/gems-gemology to take the Challenge online. Entries must be
received no later than Friday, August 11, 2017. All entries will be acknowl-
edged with an e-mail, so please remember to include your name and e-mail
address (and write clearly).
Score 75% or better, and you will receive a certificate of completion (PDF
file). Earn a perfect score, and your name also will be listed in the Fall 2017
issue of GEMS & GEMOLOGY.
formed.
1. Which statement about pink CVD synthetics by B. can be overgrown with a thin
synthetic diamonds is usually true? A. diamond type. diamond coating of a much
A. They demonstrate no fluores- B. viewing strain patterns in younger age.
cence to long-wave UV light. crossed polarizers. C. are found beneath Lac de Gras.
B. They show weak red fluores- C. IR absorption spectroscopy. D. all of the above
cence to long-wave UV light. D. the absence of a SiV– doublet.
C. They show blue fluorescence 8. The Sutara Placer is significant
to long-wave UV light. 5. What risk does an opal cutter take because
D. They start out as light brown when trying to maximize play-of A. it is a new source of facet-
CVD material. color from a piece of rough? grade sapphire.
A. Viewing the color bars from B. it produces trapiche and
2. Which statement about ruby the side may give a false sense trapiche-type sapphire.
mining in northern Mozambique is of the play-of-color present at C. it is one of only two known
true? the top of the bar. gem-bearing sapphire placer
A. Montepuez Ruby Mining plans B. Color bars may vary in thick- deposits in Russia.
to incorporate more automa- ness or even disappear within D. its sapphire’s color range is ri-
tion in the sorting process. a stone. valed only by Yogo sapphire
B. Activity is slowing down. C. Sand may be embedded in the from Montana.
C. The highest-quality rubies are silica gel, resulting in blem-
coming from Napula. 9. Which statement about
ishes and disrupting the poten-
D. Garimpeiros will play a more Colombian emeralds is true?
tial for a smooth top surface.
significant role as the existing D. All of the above. A. The oxidation of organic matter
deposits are further developed. released the elements Be, Cr,
6. Flux residues were found in and V, which were incorporated
3. In general, rough diamonds from synthetic emeralds donated by into the growing emeralds.
the Diavik mine demonstrate Richard Nacken to the Deutsches B. The emeralds formed at 330°C
A. cubic morphology and near Museum. This find indicates what from hot evaporitic brines.
colorless to yellow colors. is true about these samples? C. The emeralds’ fluid inclusions
B. dodecahedral morphology and A. None of them were grown generally contain doubly re-
brown color. hydrothermally. fractive crystals.
C. octahedral morphology and B. All of them contain both D. A and B
D–Z color. molybdenum and vanadium.
D. octahedral morphology ac- 10. Which statement about the zero-
C. Only gold crucibles were used.
companied by complete or phonon line (ZPL) in photolumi-
D. Only synthetic emerald seeds
partial coatings. nescence spectroscopy is false?
were used.
A. With increasing temperatures,
4. CVD synthetic diamond can 7. Diavik diamonds the ZPL emission decreases.
usually be separated from HPHT A. are among the oldest diamonds B. The ZPL photon emission re-
110 G&G CHALLENGE GEMS & GEMOLOGY SPRING 2017

sults when only phonons are material with an SG of about 3.0 joined with animal hide glue.
involved. shows a B-O-Si bending vibration D. The preferred carving material
C. At room temperature, it is hard at a Raman shift of 612 nm? was nephrite.
to distinguish the ZPL from A. Danburite
21. How does phosphorescence differ
other emission features. B. Grandidierite
from fluorescence?
D. The wavelength of a given ZPL C. Variscite
is exact. A. Phosphorescence is a form of
D. Tourmaline
thermoluminescence and fluo-
11.The Fabergé hardstone figures of the 16. Val Malenco serpentine is rescence is a form of photolu-
Cossack guards Pustynnikov and A. composed of lizardite, antig- minescence.
Kudinov can be distinguished by all orite, and chrysotile. B. Phosphorescence occurs after
of the following features except B. composed of lizardite and sim- the excitation light source is
A. the color of their kiver bags ilar to that from China. turned off and fluorescence oc-
and belts. C. composed of antigorite and curs while it is turned on.
B. the gems used for their eyes. similar to that from Korea. C. Fluorescence lasts longer than
C. the inscription on the bottom D. distinguished from Chinese and phosphorescence.
of their boots. Korean material by RI and SG. D. Fluorescence is delayed lumi-
D. Kudinov’s Gold Badge of Office. nescence; phosphorescence is
17. Which of the following is a differ- not.
12. In 1943, Richard Nacken filed a ence between natural and
patent application for a method non-bead-cultured (NBC) pearls? 22. What gem-quality minerals
that would influence synthetic A. NBC pearls do not contain contain BO3 and SiO4 groups,
crystal growth for technical appli- beads. have an SG of approximately 3.0,
cations. What did the patent cover? B. NBC pearls only form in a graft and are found in Madagascar?
A. A hydrothermal process for location. A. Grandidierite and tourmaline
synthetic emerald growth. C. Only hatchery-raised mollusks B. Danburite and tourmaline
B. A hydrothermal process for pri- can produce NBC pearls. C. Danburite and grandidierite
marily synthetic quartz growth. D. There is no difference. D. None of the above
C. A hydrothermal process for pri-
18. Which statement regarding the 23. Ballerina Pearl Co. treats pearls
marily synthetic sapphire
blue color in blue zircon is true? from P. margaritifera to achieve a
growth.
A. It is the natural color of zircon. pistachio color. Which statement
D. An autoclave for growing
B. It is stable to long-wave UV about their color is true?
quartz and emerald from flux.
exposure. A. It is altered by acetone.
13. Which statement about variscite is C. It can be altered and restored. B. It is enhanced by isopropanol.
false? D. It is assumed to result from the C. It results primarily from dyeing.
A. It was used in Neolithic jewelry. treatment of yellow material. D. None of the above are true.
B. The main source today is cen-
19. Some companies preforming emer- 24. What percentage of the gold
tral Tajikistan.
ald rough in Jaipur maximize the refined by Hoover and Strong
C. Its phosphate component comes
weight yield of finished stones by comes from recycled material?
from guano decomposition.
D. Relative concentrations of iron A. using hand-controlled jam pegs. A. 80%
and chromium are the main B. using high-tech equipment and B. 20%
color contributors. computerized cutting. C. 100%
C. outsourcing material to ensure D. 50%
14. In peridot, cleavage separations they get the best cutters.
that result from the presence of a D. using traditional bow-powered 25. Finding chrysoberyl and sapphire
small solid inclusion paired with laps. together associated with host rocks
a fluid-filled negative crystal are of volcanic origin such as basalts
called ____________ inclusions. 20. Which statement regarding the is related to
materials used by Fabergé to A. pegmatitic activity, which is
A. fingerprint
construct the hardstone Cossack common.
B. tension halo
figures is false? B. high-grade metamorphism,
C. lily pad
D. three-phase A. All the materials are natural. which is somewhat rare.
B. All gemstones are from the C. tertiary volcanic host rocks,
15. Which exceptionally rare B- and Russian Empire. which is very rare.
Si-containing orthorhombic gem C. Individual gem pieces were D. none of the above.
G&G CHALLENGE GEMS & GEMOLOGY SPRING 2017 111

Contributing Editors
Emmanuel Fritsch, University of Nantes, CNRS, Team 6502, Institut des
Matériaux Jean Rouxel (IMN), Nantes, France ([email protected])
Gagan Choudhary, Gem Testing Laboratory, Jaipur, India ([email protected])
Christopher M. Breeding, GIA, Carlsbad ([email protected])
Dealers also talked about the October 2016 removal of

TUCSON 2017 U.S. sanctions on Myanmar, legalizing the import of
Burmese jadeite and rubies. According to Edward Boehm
of RareSource, the lifting of the ban was widely welcomed,
The 2017 Tucson gem and mineral shows drew eager buy- but Myanmar government reforms of the gem mining sec-
ers from all over the globe. Although many dealers reported tor have some way to go. Conditions on the ground are
lower traffic this year, most were content with the volume much improved, however. Boehm told us new production
of business and noted that buyers were “serious.” might take some time to appear, which would impact
As in previous years, we identified some strong trends: prices and selection of available goods in the short term.
Large multinational companies have significantly im-
• Demand for special one-of-a-kind pieces in both pacted colored gemstone mining and distribution over the
pearls and colored gemstones last decade. Gemfields’ Kagem and Montepuez mines sup-
• Continuing investment by multinational companies ply approximately one-third of global emerald and 70% of
in colored gemstone mining and distribution global ruby production, respectively. At this year’s show
• A growing focus on ethically sourced gemstones and we were able to interview Gemfields CEO Ian Harebottle
beneficiation for his perspective on the company’s operations and its
• Strong demand for high-end gems and a softening of 2013 merger with luxury brand Fabergé.
demand for commercial goods Representing another publicly traded company intend-
ing to mine ruby in Mozambique, Christiaan Jordaan of
• Continuing importance of the secondary market in
Mustang Resources LLC told us about Mustang’s initial
the U.S. for exceptional pieces
2016 bulk sampling and exploration activities. The com-
• Innovative partnerships emerging between individ- pany hopes to become an important supplier of commer-
ual colored gemstone mines, designer cutters, and cial and gem-quality ruby.
television merchandisers Marcello Ribeiro of Belmont Group updated us on de-
This year’s big story was the emergence of Ethiopia as velopments at the company’s Belmont and Canaan emer-
a potentially major source of gem-quality emerald. These ald mines at Itabira, Brazil. He emphasized the importance
new gems resemble other schist-hosted emeralds, espe- of sound mine planning through fieldwork and trace-ele-
cially those from Brazil and Zambia. Although most of the ment analysis of potential emerald host rocks.
material is less saturated and included, dealers are excited David Bindra of B&B Fine Gems confirmed the impor-
that the new source also produces fine gem-grade crystals tance of the secondary market for exceptional gems, which
of exceptional size, color, and clarity. are in very high demand. He noted less consumption of
commercial to mid-grade material.
A partnership between Morocco’s Geostone Group and
Editors’ note: Interested contributors should send information and illustra- U.S. gem carver Glenn Lehrer illustrates another interest-
tions to Stuart Overlin at [email protected] or GIA, The Robert Mouawad ing development. Production from Geostone’s Moroccan
Campus, 5345 Armada Drive, Carlsbad, CA 92008. amethyst mine is cut to Lehrer’s high standards in an In-
dian factory, featuring a branded designer cut with high
GEMS & GEMOLOGY, VOL. 53, NO. 1, pp. 112–142. weight retention. The resulting faceted stones—and fin-
© 2017 Gemological Institute of America ished amethyst jewelry—are sold directly to consumers
through television merchandising. The approach helps
112 GEM NEWS INTERNATIONAL GEMS & GEMOLOGY SPRING 2017

A
E
D
F
G H
A: 59.28 ct quartz with iron staining. B: 11.77 ct black

opal cabochon from Lightning Ridge, Australia. C:
100 ct Mogok ruby crystal with a perfect termination
at one end. D: Pyrite at the heart of a 46.98 ct color-
less quartz. E: 73.22 ct “diamondback-cut” tourma-
line from Afghanistan. F: 4.31 ct pair of matched
Burmese pink spinels. G: 49.36 ct purple trapiche
sapphire from India. H: 3.61 ct trapiche sapphire from
Kashmir. I: 53.77 ct papagoite. Photos by Robert Wel-
don/GIA. Courtesy of Rare Earth Mining Co. (A, D,
and I); David Humphrey (B); RareSource (C); Stephen
M. Avery (E); and Mayer & Watt (F, G, and H).
I
build a niche for Moroccan amethyst and brings the bene- of science fiction and traditional inlay work is one of the
fits of scale to designer cutting, which is usually associated show’s unexpected finds.
with unique pieces. Finally, no survey of the Tucson shows would be com-
Eternity Emerald’s Arthur Groom shared his rough plete without Paula Crevoshay’s one-of-a-kind designs.
emerald buying expertise with us, especially his years of ex- This year, she showcased a bracelet featuring five large
perience negotiating with miners in Afghanistan’s Panjshir freeform fire opals, a swallowtail butterfly pin with yellow
Valley. He noted that the quality of Afghan emerald is not and black diamonds, an elephant pin with mother-of-pearl
realized in the West because large, fine crystals are often tusks, and a stunning cuff bracelet featuring a spectacular
damaged by improper blasting and extraction techniques. boulder opal centerpiece.
Potentate Mining’s Warren Boyd showed us sapphire Duncan Pay, Andy Lucas, Tao Hsu,
production from Montana’s Rock Creek alluvial deposits, Eric Welch, and Albert Salvato
including a remarkable 12.61 ct blue sapphire cut from a GIA, Carlsbad
6.37 g rough crystal. Although most of the production has
greenish or brownish pastel colors, it reacts well to heat A new discovery of emeralds from Ethiopia. In recent years
treatment. Boyd told us bright natural colors are in high Ethiopia has gained considerable attention in the gem trade
demand. He showed several suites of fancy sapphires cut for large amounts of high-quality opal from an area near
by his client, Americut. Montana’s alluvial sapphire de- Wegel Tena (B. Rondeau et al., 2010, “Play-of-color opal
posits will be the subject of an upcoming G&G field report. from Wegel Tena, Wollo Province, Ethiopia,” Summer
The trend toward ethical, sustainable business prac- 2010 G&G, pp. 90–105). Apart from opal, emeralds have
tices and transparent supply chains is exemplified by Shea- been sporadically mined, near Dubuluk, for more than a
han Stephen of Sheahan Stephen Sapphires, Inc. His decade. This deposit is located about 80 km from the
company documents and guarantees the integrity of the Kenyan border. Gemfields has been exploring this deposit
gems it sells from the mine, through treatment and cutting since July 2015 (Fall 2012 GNI, pp. 219–220).
and directly to the customer.
A new deposit of high-quality emeralds (see figure 1)
Nigeria is an under-documented source of fine rubellite
has been found in the rural villages of Kenticha and Dermi,
and indicolite tourmaline and predominantly blue basalt-
in the Seba Boru district (figure 2). In November 2016, au-
hosted sapphire. For this reason, we especially welcomed
thor MN and business partner Daniel Kifle visited the local
talking to Zoe Michelou, who represented a Nigerian min-
trading town of Shakiso, where Ethiopian gem merchants
ing company. She updated us on production of these gem-
gather to legally buy and sell emeralds that are mined sev-
stones from that country’s Oyo, Kwara, and Taraba states.
eral kilometers away. Shakiso is located about 160 km
Gem paintings, which GIA has documented in Viet-
north of the Dubuluk emerald deposit. The mining area is
nam and Mogok, made their first appearance at Tucson.
divided into a few “associations.” Each consists of a man-
Wanlaya Suwannapirom’s Than Thong Arts booth featured
ager and several members who control the actual mining
an array of art-inspired handmade portraits and miniatures.
and distribution of the emerald rough. After the rough has
The technique converts otherwise unusable natural gem
been sorted, it makes its way first through Shakiso before
fragments into valuable art objects and wearable art such
as pendants and pins.
Gem artist Alexander Kreis showed us a spectacular
27.20 ct freeform tanzanite complemented by a custom ring Figure 1. This suite of untreated emeralds is from a
mount made by his mother Sonja, a master goldsmith. new find in Ethiopia’s Seba Boru district. The largest
They related the importance of the story behind a jewelry faceted stone weighs 10.64 ct. The largest rough crys-
piece for their clients—how details of the stone’s cut and tal weighs 63.12 g. Photo by Robison McMurtry, cour-
the ring’s design represent the toil and effort of the Tanzan- tesy of Michael Nemeth Inc.
ian miners and imbue the piece with added significance.
Fran Mastoloni provided a cultured pearl market up-
date, explaining how careful selection and matching of the
different cultured pearl types into a range of varied neck-
laces help him cater to the market’s desire for distinctive
“fashion-forward” yet flexible jewelry.
Dealer Bill Vance of Vance Gems exhibited rare gem-
quality magnesio-axinites from just one location in
Merelani, Tanzania. This material displayed remarkable
fluorescence, and we hope to report further on its chem-
istry in the near future.
We were delighted to find a Southwest-inspired Star-
ship Enterprise, handmade by David Freedland of David R.
Freedland Jr. Designs in sterling silver. This quirky blend

Red Sea
Atlantic
Yemen Ocean
Sudan
Gulf of Aden
A B
Addis Ababa
Ethiopia
Shakiso
Somalia
C D
Indian Ocean
Kenya
Figure 4. Microscopic observation of the Ethiopian
Figure 2. The new deposit of Ethiopian emerald is lo- emeralds revealed (A) biotite crystals; (B) blocky
cated near the town of Shakiso, about 12 hours from multiphase inclusions; (C) multiphase inclusions con-
Addis Ababa. taining a liquid, gas, and multiple daughter crystals
seen parallel to the c-axis; and (D) a multiphase fluid
inclusion showing liquid and gaseous CO2 phases
being sold to dealers in the capital city of Addis Ababa, and an immiscible aqueous liquid (parallel to this in-
about a 12-hour drive from the mining area. clusion are several growth tubes). Photomicrographs
According to Tewoldebran Abay, the mineral market- by Nathan Renfro (A and B) and Victoria Raynaud (C
ing director of the Ministry of Mines, Petroleum and Nat- and D). Field of view 2.04 mm (A), 1.79 mm (B), 1.07
ural Gas, more than 100 kilograms of emerald rough have mm (C) and 1.03 mm (D).
been produced to date. Mining still is done the traditional
way using hand tools, without heavy machinery.
Samples from the new deposit, acquired from multiple
in dark biotite crystals, but had an extremely pleasant
independent sources, were examined at GIA’s Carlsbad and
green color when examined with transmitted light. How-
Bangkok laboratories. Even though most of the material is
ever, these Ethiopian crystals often do not yield large clean
commercial grade, lighter in saturation, and moderately to
stones because their interior is riddled with dense, dark bi-
heavily included, fine gem-grade crystals of exceptional
otite mica crystals. Some show a double termination, but
size, color, and clarity (see figures 1 and 3) are obtainable
most are broken and heavily included on one end. Usually
and can produce stones that do not require clarity enhance-
only one end of the crystal is clean enough to yield faceted
ment. Many of the rough crystals were completely covered
gems. The matrix minerals attached to some of the emer-
alds were identified as dark brown to black biotite flakes,
quartz, and kaolinite.
Figure 3. Author Michael Nemeth sorts through These emeralds are very similar in appearance to other
parcels of gem-quality rough emeralds from the open schist-hosted emeralds—in particular, those from Brazil
market in Shakiso. Rough crystals can weigh more and Zambia. Among the faceted and rough samples exam-
than 20 g, with reports of some weighing almost 100 ined, blocky multiphase inclusions and irregular biotite
g. Photo by Michael Nemeth. crystals were the most common microscopic features ob-
served (figure 4). Otherwise, the gemological properties
were very consistent with emeralds, including an average
specific gravity of 2.73 and a refractive index of 1.581–
1.589. These emeralds were generally inert to long- and
short-wave UV exposure due to their moderately high iron
content, which is typical of schist-hosted emeralds. UV-
Vis-NIR spectroscopy results (figure 5) were consistent
with emeralds colored by chromium and vanadium. The
Fourier transfer infrared (FTIR) spectrum was consistent
with beryl, as expected, but did not reveal any other diag-
nostic features.
Quantitative trace element chemical analysis was per-
formed with a Thermo Scientific iCap Q inductively cou-
pled plasma–quadrupole mass spectrometer combined
GEM NEWS INTERNATIONAL GEMS & GEMOLOGY SPRING 2017 115

UV-VIS -NIR SPECTRA
ABS. COEFFICIENT (cm–1) 20 need written permission to enter the Shakiso area for buy-
18 o ray ing. The law is vigorously enforced, and penalties are se-
16 e ray
14
vere.
12 This exciting discovery in Ethiopia will provide a new
10 source of large, high-quality emeralds for the gem and jew-
8 elry trade. So far, this deposit appear to be quite promising,
6
as significant production was seen in the recent gem shows
4
2 in Tucson, Bangkok, and Hong Kong. Only time will tell
0 how significant this deposit will be.
300 400 500 600 700 800 900 1000 1100
Nathan Renfro and Ziyin Sun
WAVELENGTH (nm) GIA, Carlsbad
Michael Nemeth
Figure 5. The characteristic UV-Vis-NIR spectra of the MLN Gems, San Diego, California
new emerald production from Ethiopia. Wim Vertriest, Victoria Raynaud, and
Vararut Weeramonkhonlert
GIA, Bangkok
with a New Wave Research UP-213 laser ablation unit. Fine gems and update from Myanmar. At the AGTA Gem-
The analyses were compared to data from other known Fair, Edward Boehm (RareSource, Chattanooga, Tennessee)
sources using GIA reference samples, including Zambian showed us a succession of top-quality gems. Boehm is a ge-
and Brazilian schist-hosted emeralds. Based on the results, ologist, gemologist, and accomplished gem dealer who
it was possible to separate the new find of Ethiopian emer- works with rare and higher-end gems. He noted that the
alds from other sources by comparing trace alkali metals price of spinel from Myanmar has increased dramatically,
and some transition metals (figure 6). with recent prices at the source more than double those of
Due to heightened tensions and fear of price instability, previous years. He explained that spinel has gained favor in
most of the mine area was temporarily closed by a joint ef- Myanmar, which is reflected by the higher prices at this
fort of the mining associations and the local government year’s show.
from early November through December 2016. It has been Boehm noted that a growing appreciation for spinel from
reopened, but now all dealers, including Ethiopian dealers, all sources is driving higher market prices. As an example,
Figure 6. LA-ICP-QMS quantitative trace-element composition analysis of alkali and some transition metals
(measured in ppmw) proved useful in separating this new deposit of Ethiopian emeralds from Brazilian and Zam-
bian schist-hosted deposits.
Cs-Li-Rb Cs-Sc-Rb
Cs Cs
0 Zambia 0 Zambia
100 Brazil 100 Brazil
10 Ethiopia 10 Ethiopia
90 90
20 20
80 80
30 30
70 70
40 40
60 60
50 50
50 50
60 60
40 40
70 70
30 30
80 80
20 20
90 90
10 10
100 100
0 0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
LI Rb×8 Sc×5 Rb×5

Figure 7. Suites combining blue-to-violet pastel col- Figure 8. This cushion-cut 20.41 ct Burmese peridot
ored spinels with red to rose-colored gems of similar shows a softer appearance due to a multitude of tiny
tone. The blue gems range from 3.63 to 8.30 ct and inclusions. Photo by Albert Salvato; courtesy of
the pink to red examples from 3.16 to 5.70 ct. Photo RareSource.
by Albert Salvato; courtesy of RareSource.
he showed us a beautiful oval 21.56 ct Sri Lankan pink in the pavilion in the face-up position, which is considered
spinel cut to make the most of the gem’s high clarity and less desirable. Buyers want to ensure the stone is oriented
moderate dispersion. The size, clarity, and brightness are in such a way that doubling isn’t visible through the table.
reminiscent of Tajik spinel. He compared this bright lilac Next, he showed us a rich green, 4.81 ct cushion-cut
pink Sri Lankan spinel with a strong orangy red square cush- demantoid garnet from Russia, which contained a golden-
ion-cut 9.82 ct Burmese gem, which he described as having colored horsetail inclusion. Such a gem would wholesale
a “flame” color. The per-carat price of the flame-colored for $15,000 per ct, he said. According to Boehm, gems of
Burmese stone is almost double that of the Sri Lankan such pure green color rarely come out of the ground—they
spinel—$12,000 per carat wholesale versus $7,000 per ct. A are typically heated to this color. This treatment has been
strong red “flame” color makes the price jump significantly. going on for 10–15 years, he explained. Despite this treat-
If the stone were above 10 ct, Boehm added, the price would ment, the demantoid is still an exceptional stone.
further jump to around $15,000 per carat. Boehm explained that judicious heating converts a yel-
Pastel blue to violet spinels are also selling well for lowish green gem to a vibrant green but removes some of
Boehm. Designers like to mix these delicate blues with the stone’s characteristic fire, making it an almost “electric”
pinks or “rose” colors of similar tone and saturation to pro- green color. He prefers a balance of color and fire, acknowl-
duce very attractive suites and sets. The pinks help to high- edging that the gem’s inherent fire—flashes of red and blue
light the violet or purple in the pale blues, he added. The coming off a green bodycolor—is part of its unique appeal.
bluish violet examples he showed us hailed from Sri Lanka In today’s market, any demantoid over 2 ct is extremely rare,
and wholesaled for $1,200–$1,500 per ct (figure 7). Com- so this gem’s size of almost 5 ct makes it very desirable.
plementary rose-colored gems came from a variety of The demantoid was from new production rather than
sources—including Vietnam, Sri Lanka, and Myanmar— the secondary market, Boehm explained. He added that
and ranged from $1,000 to $1,800 per ct. there is still newly mined supply and treaters are perfecting
Boehm explained that the most sought-after color is the heating techniques to make gems much more vivid, al-
“electric” haüyne blue, which comes from Luc Yen, Viet- though not all respond like this one.
nam, and is only sporadically available in small sizes. He Finally, he showed us a striking 31.79 ct Sri Lankan
added that the light-toned, gray-to-violet blue colors are pear-shaped sapphire with padparadscha color—a delicate
more available than in the past—partly because there’s pink flushed with a bloom of orange (figure 9). He ex-
more demand so more people are bringing it to market. plained that the gem had been recut three times to perfect
Like Burmese spinel, Burmese peridot is currently in its shape and proportions. The objective of recutting the
vogue. Boehm showed us a fine 20.41 ct cushion-cut gem original old-style cut was to close the window on the pavil-
(figure 8). The presence of a multitude of tiny inclusions ion, but any further work would affect the brilliance and
lend it a softness and reduce extinction. By comparison, might even lose the cherished pinkish orange color. He
fine Pakistani peridot—which is also available in large added that when cutting padparadscha sapphire, it is very
sizes—has a more “crystalline” appearance and deeper hard to keep that balance of pink and orange, because
color, but shows more extinction. He explained that the sometimes the orange might be in just one portion of the
price of Burmese peridot has recently gone up sharply. A stone, which could easily be removed accidentally. A gem
top gem like the one in figure 8, which would formerly of this size is enormous for a padparadscha, he noted, and
have been in the $250–$350 per carat range, is now whole- the wholesale asking price would be in the neighborhood
saling for $450–500 carat and even up to $600 per carat. He of $30,000–$35,000 per ct.
cautioned buyers to check whether a stone shows doubling Boehm noted that on October 7, 2016, President Barack

Mogok mining licenses have been suspended as the
Burmese parliament and ministries work together to
rewrite mining laws in Mogok and the jade mining areas
in Pagan. These laws have been in place for more than 100
years and have changed little.
When Aung San Suu Kyi’s National League of Democ-
racy party won a majority of seats in parliament during the
November 2015 election, it was able to effect change with
a special focus on the environment. The new administra-
tion’s objective is to make reclamation integral to obtaining
a mining or prospecting license, so that when mining fin-
ishes, the licensee will be obligated to return the land to its
Figure 9. This 31.79 ct Sri Lankan padparadscha sap- original condition. The government wants to ensure these
phire shows the cherished combination of a pink laws are enforced and that environmental and ethical stan-
bodycolor suffused with orange. Photo by Albert Sal- dards conform to Western expectations for the gem mining
vato; courtesy of RareSource. sector. The aim is for the new mining laws to meet the in-
ternationally accepted standards outlined by the Organiza-
tion for Economic Cooperation and Development (OECD),
Obama signed an executive order lifting U.S. sanctions on a group of 34 democratic countries that develop economic
Myanmar, allowing Burmese jadeite and rubies, and any and social policy to support free market economies. These
jewelry containing them, to be imported into the United principles are laid out in the policy paper OECD Due
States. The sanctions had been imposed in 1997 (see Diligence Guidance for Responsible Supply Chains of
https://www.treasury.gov/resource-center/sanctions/Pro- Minerals from Conflict-Affected and High-Risk Areas
grams/pages/burma.aspx). (http://dx.doi.org/10.1787/9789264252479-en). For more in-
Boehm felt fortunate to be in Mogok (figure 10) when formation on corporate social responsibilities and ethical
the ban was lifted and was able to celebrate with many of colored stone supply chains, please see J. Archuleta, “The
the local miners and brokers. He hoped the change would color of responsibility: Ethical issues and solutions in col-
significantly improve the supply of Mogok gems but antic- ored gemstones,” Summer 2016 G&G, pp. 144–160).
ipated that any increase in gems reaching the U.S. from Boehm mentioned the recent AGTA October 2016 del-
this source would likely take some time. Most if not all egation (figure 11) to Myanmar that visited the cities of Yan-
gon, Nay Pyi Taw, Mandalay, and Mogok. Along with other
U.S. trade and industry bodies, AGTA is seeking to reestab-
Figure 10. In October 2016, an AGTA delegation lish responsible gemstone trade between the two countries.
visited gem markets and mines around the world- This is only possible because of the political and economic
famous locale of Mogok. Their mission was to report progress Myanmar has made in the past six years, but it also
back on conditions within the local gem and min- depends on Myanmar’s gemstone sector meeting interna-
ing industry after the new government took office. tional standards such as OECD’s guidelines.
Photo by James E. Shigley. The delegation, which included GIA researcher Dr.
James Shigley and representatives from Jewelers of Amer-
ica, met with a cross section of stakeholders, including the
U.S. ambassador, ministers and representatives of the
Myanmar government, international and local organiza-
tions, trade associations, gem dealers and traders, and mine
owners. They visited a number of gem markets and mines,
including the Baw Maw mine (figure 12) and the Taung
ruby and sapphire mine.
According to Boehm, the delegation was an impartial
group sent to look at the situation “on the ground” in Myan-
mar, particularly in Mogok. He explained that it was an
“eye-opening” experience and although conditions were
much better than the group anticipated, there was room for
improvement. The delegation made a series of observations
that were set out in their white paper, Step by Step – Myan-
mar Gem Sector Emerges from Isolation and U.S. Sanc-
tions, which is available from AGTA’s website at
http://agta.org/info/docs/burmawhitepaper2016.pdf

age the gem sector properly and that the Burmese people
benefit from the gems. He cautioned that a lot of mining li-
censes remain suspended. Therefore it’s very unlikely that
U.S. businesses will see much production until these are re-
instated, possibly in late 2017. He forecasts that the U.S.
trade at large will have to wait at least another year before
production appears in earnest. At that point, he expects
many more U.S. buyers will travel to Mogok and competi-
tion for scarce goods will drive prices up, but at least the
goods can be legally imported into the U.S. once more.
He reported that because of the government’s suspen-
sion of mining licenses, Mogok’s mines are currently
Figure 11. The October 2016 AGTA delegation at closed again after a brief reopening. Boehm suspects several
Mogok’s Taung ruby and sapphire mine. Edward reasons for caution on the part of the authorities. Firstly,
Boehm and GIA’s Dr. Jim Shigley are at left. Photo it would be very tempting for miners to produce as much
courtesy of James E. Shigley. material as possible now that dealers from the U.S. can
legally buy gems in Mogok, so the government wants to
Boehm indicated that Mogok was still beholden on the buy time to regulate mining ethically to make sure the
U.S. gem industry’s ongoing efforts to ensure that Myan- Burmese people are protected (figure 13). Secondly, they
mar’s government follows through with its promise to man- want to prevent large foreign investors from taking over
the mining sector. Finally, there are some security issues
with traveling the new road leading into Mogok at night,
Figure 12. Workers sorting heavy mineral concen- which the authorities want to correct before mining li-
trate from the Baw Mar mine near Mogok. Photo by censes are reinstated. Boehm considers these good signs
James E. Shigley. that demonstrate the Myanmar government’s commit-
ment to security and corporate social responsibility for the
good of its citizens.
Duncan Pay
Conversation with Gemfields. UK-based Gemfields is

widely considered the largest producer and distributor of
responsibly sourced colored gemstones. This year in Tuc-
Figure 13. During the AGTA trip, delegates watched

villagers searching for peridot in the waste tailings of
the Pyaung Gaung mine near Mogok. Photo by James
E. Shigley.

Figure 14. The massive
open-pit Kagem emer-
ald mine in Zambia is
perhaps the largest
colored gemstone min-
ing operation in the
world. Photo by
Andrew Lucas, cour-
tesy of Gemfields.
son, we had an exclusive interview with CEO Ian Harebot- According to Gemfields, Fabergé’s sales orders jumped 95%
tle to get his insights on the company and the colored gem- during the second quarter of fiscal year 2017.
stone industry overall. When it comes to corporate social responsibility, Gem-
Unlike the diamond industry, almost every segment of fields sets a good example across its mining locations
the colored stone trade has traditionally lacked scale and through successful joint ventures and community pro-
systematic practices. Gemfields has dramatically changed grams. The Kagem emerald mine is one of the only mining
this dynamic in some of the most influential colored stone companies that has paid federal taxes in the past 15 years
categories. The company is actively working on three min- in Zambia. The Montepuez ruby mine was named Mozam-
ing operations: the Kagem emerald mine (figure 14) and the bique’s most transparent tax payer and the most important
Kariba amethyst mine in Zambia, and the Montepuez ruby job generator in its province. Harebottle described the dif-
mine in Mozambique. Among the three, Kagem and Mon- ferent challenges the company has faced and will be facing
tepuez are world-class operations, supplying about one- in the future. Since the sites are usually located in areas
third of the global emerald production and 70% of ruby
production. Before, both stone types experienced fluctuat-
ing supply and chronic shortages. Gemfields is very opti- Figure 15. The merger of Gemfields and Fabergé
mistic about the expected supply from Kagem and should continue to inspire one-of-a-kind colored gem-
Montepuez, based on scientific field surveys and detailed stone pieces. Photo courtesy of Fabergé.
geophysical and geochemical studies.
To stabilize supply for the global market, Gemfields re-
quires the mines to reserve one year’s production in case
yield declines significantly. Based on the authors’ research,
many Gemfields clients appreciate the standards set by the
company on emerald and ruby mining and distribution.
The consistent supply and well-tuned rough grading sys-
tems grant buyers more time to focus on faceting, jewelry
making, and sales. This practice also has a positive influ-
ence on other mining companies that aim to meet the
same standards.
In 2013, Gemfields purchased the iconic luxury brand
Fabergé. Harebottle explained that the deal does not reflect
an interest in vertical integration but in creativity and
uniqueness in the market, an area where Fabergé has a rich
heritage (figure 15). Therefore, he hopes that the brand can
be a platform of magnificent jewelry designs, igniting the
passion for colored gemstones in the younger generation.

that lack robust mining regulations or laws and some
countries have a very long mining history, the company
often applies different strategies to localize itself. To cham-
pion responsibility and sustainability on location, a foreign
company such as Gemfields needs to make the local people
feel confident about their future, and this is fueled by job
opportunities.
Since its founding less than 10 years ago, Gemfields has
always sought to rekindle the passion for colored gem-
stones. Harebottle likes to compare the company’s role in
colored stones with that of De Beers in the diamond sector,
especially when the latter heavily promoted diamonds. He
also feels that colored gemstones should have a stronger
position in the luxury goods market. With this mission
under way, and active exploration for new production sites
in Colombia and Ethiopia, more accomplishments are ex-
pected from this major colored gem supplier. Figure 16. The washing process used by Mustang Re-
sources is different from most jigs used for colored
Andrew Lucas and Tao Hsu
stone mining. Photo courtesy of Mustang Resources.
New ruby miner in Mozambique. Mustang Resources LLC,

a publicly listed Australian company, is developing a ruby
mining operation in northern Mozambique. With the suc- well-known cutters. Mustang will also test the ruby mar-
cess of the colored stone mining giant Gemfields, Mozam- ket in Asia and other parts of the world. Jordaan reassured
bique has secured its position in today’s global ruby the authors that Mustang will focus on improving quality
market. Mustang’s concession is located about 8 km north- of life in the mining community and promoting their ruby
west of Gemfields’ Montepuez ruby mine. Both properties as an ethically mined product. Since the beginning of this
are located on the ruby mineralization belt in the Mon- ruby mining project, the company has hired and trained 70
tepuez Complex, where ideal physical and chemical con- local employees to work together with foreign experts.
ditions favored the formation of gem-quality ruby. In Andrew Lucas and Tao Hsu
between the two concessions there are some designated
ruby-bearing areas where individual small-scale miners can New emerald production from Belmont. The Belmont
legally work. mine in Brazil has been a key emerald supplier in the global
While the geology of the Mustang operation is generally market for the past 40 years. It is one of the very few com-
similar to that of the Gemfields concession, the bulk sam- panies that can guarantee a complete mine-to-market chain
pling strategy and washing process are slightly different
(figure 16). Currently, the miners focus on exploring and
sampling the secondary deposits. They open pits of 20 × 5
Figure 17. Taking a cue from Gemfields, Mustang
meters or 20 × 7 meters and immediately refill the first pit
Resources grades rough rubies into specific quality
before opening the adjacent ones; this way, no large pit is
and size parcels. Photo by Andrew Lucas and Tao
left open on the surface of the mining area. Mustang has
Hsu, courtesy of Mustang Resources.
been using scrubbers instead of log washers to break the
clay bodies in the ore. A new customized scrubber system
will soon be in place to optimize this process. Unlike the
jig system applied by many colored gemstone mining op-
erations around the world, the team uses a giant rotary pan
to concentrate the heavy minerals, including ruby, through
the centrifugal forces generated during rotation.
Since the team from Mustang started to develop the
project in early 2016, the exploration, bulk sampling and
ruby recovery have made significant progress. Christiaan
Jordaan, the managing director of Mustang’s Montepuez
ruby project, informed the authors that the washing plant
had been successfully relocated and tuned to operate effi-
ciently and that grading into parcels had begun (figure 17).
The team has test marketed these rubies in the United
States, where some large, significant stones will be cut by

of custody for the emeralds it produces. As open-pit mining sustainability, the company plans to add two more optical
became more expensive over the past decade, underground sorters and further expand the cutting facility. Belmont
mining gradually generated a larger share of production at will also keep promoting its mine-to-market product and
Belmont. Today, all mining activity at the original open pit better communicate with younger buyers to make them
is underground. Open-pit mining will continue in the two feel proud of their emerald purchases.
pits to the north of the original site. General manager Andrew Lucas and Tao Hsu
Marcelo Ribeiro revealed some exciting news about the
Canaan mine, about 2 km away from the original pit, and
expansion plans for the sorting and cutting facilities. Importance of the secondary market for fine gems. At the
AGTA show, Dave Bindra (B&B Fine Gems, Los Angeles)
Belmont acquired interest in the Canaan mine through
showed us some exceptional items he recently sourced
a partnership with its original owners. Exploration and un-
from the secondary market. These are previously sold
derground mining infrastructure are still ongoing at this
goods recirculated back into the market by the former
site. Although the operation just started about a year ago,
owners or their heirs.
Ribeiro and his team have learned a lot about this deposit
through field studies and chemical analysis. The interest- Bindra also updated us on his perception of U.S. traders’
ing finding is that there is plenty of beryllium in this area, reaction the withdrawal of sanctions against Myanmar
which makes the survey of chromium concentration crit- (Burma) and what that means for U.S. dealers like him. Ac-
ical to locating the mining targets. This is the exact oppo- cording to Bindra, most U.S. dealers greeted lifting of the
site of the situation at the original Belmont site, where trade embargo with enthusiasm as they can legally import
beryllium is the critical factor. and trade Burmese rubies again. Prior to this they were lim-
According to Ribeiro, two geological events were re- ited to trading items that had come into the country “pre-
sponsible for emerald crystallization in the area. The orig- ban.” Bindra doesn’t expect prices to come down; rather,
inal Belmont mine is mainly influenced by the first event, he projects they will strengthen as American demand and
Canaan by the second. Based on research, the second event consumption of Burmese goods grows. He noted that dur-
was much stronger at the Canaan mine, destroying most ing the ban, some consumers and certain major brands
of the crystals formed during the first event. The second- stopped consuming rubies altogether, basically because
generation emeralds have a lower chromium concentra- they weren’t open to consuming rubies from sources like
tion, and therefore lighter green colors, but they are often Mozambique, which has become a prolific source of com-
larger and cleaner than those from the Belmont mine (fig- mercial-to-fine grade gems. Now that the ban is lifted,
ure 18). The new production is sometimes slightly yellow- Bindra hopes many of these companies will start buying
ish or bluish. Ribeiro is very excited about the Canaan and selling Burmese rubies again. The main issue now is
production, which allows the company to provide a greater the very limited supply coming out, but he says there’s def-
selection to its clientele. initely a strong market for fine-quality stones.
Belmont currently cuts 50% of the Canaan production Bindra highlighted some exceptional pieces at his booth
by value. All medium- to high-quality stones are cut on- to illustrate the importance of the secondary market for his
site to add more value. To increase the capacity of the op- inventory. The company also procures gem materials from
eration and make mining cheaper for long-term currently active mining areas, but the secondary market is
important for larger, more important stones, which are
scarce as a percentage of new production. As an example, he
Figure 18. Second-generation stones from Brazil’s shows us a fine, oval-cut, 55.52 ct copper-bearing tourmaline
Canaan mine are often cleaner and lighter in color from Mozambique (figure 19). He notes this particular stone
than those from the first geological event. Photo by came out of the original 2007 production from Nampula
Eric Welch/GIA. Province, northeastern Mozambique, near the village of in
Mavuco (see B.M. Laurs et al., “Copper-bearing (Paraíba-
type) tourmaline from Mozambique,” Spring 2008 G&G,
pp. 4–30).
B&B actually sold this stone about seven years ago, and
was lucky to reacquire it when the opportunity arose in
2016. Bindra considers this fortunate as there is no produc-
tion from this deposit today and it would be impossible for
them to find an item like this, other than the secondary
market.
Another example he cites is a beautiful 5.06 ct trillion-
cut benitoite (figure 20). This gem—a barium titanium sil-
icate—is typically colorless to blue, and is noteworthy for
its high refractive indices, moderate birefringence, and
strong dispersion. It comes from just one mine, located in

value, which is why B&B is often able to reacquire items
over a period of time. He always keeps this in mind when-
ever selling rare, esoteric gemstones.
Next, Bindra showed us one of the most exceptional
stones he’s had in years: a 10.22 ct cushion-cut Brazilian
alexandrite. He adds that B&B has trouble fulfilling orders
for 6 mm (approximately 1.00 ct) round alexandrite with
current Brazilian production, so finding a 10 ct size is really
an anomaly. According to Bindra, it came out of an old col-
lection in Japan, where it sat in a safe for 20 years. He said
Figure 19. This exceptional oval-cut, 55.52 ct copper- he “couldn’t write the check fast enough” to get it into
bearing tourmaline from Mozambique was previously B&B’s inventory. In this case, there was no need for repol-
sold by B&B. They recently reacquired it through the ishing. The gem is unusual not just for its size and the qual-
secondary market. Photo by Eric Welch; courtesy of ity of its color change, but also because of its brilliance. He
B&B Fine Gems. notes that Brazilian alexandrite of this rich blue-green to
reddish purple color tend to be “over dark” and lack life.
This stone was faceted with a typical Portuguese pavilion,
San Benito County, California (see B.M. Laurs et al., “Ben- which confers considerable brilliance. Due to these factors,
itoite from the New Idria District, San Benito County, Cal- Bindra adds, the wholesale value of a truly fine gemstone
ifornia,” Fall 1997 G&G, pp. 166–187). The benitoite mine like this is well over $55,000 per carat.
has been closed for years. Bindra describes this 5 ct as an Finally Bindra shows us a 20-ct-plus star sapphire with
“astronomical size for this deposit”—the typical size range a sharp six-rayed star, strong blue bodycolor, exceptional
of commercially available finished gems from this mine translucency, and a well-proportioned, high-domed cabo-
was between 0.30 and 1.70 ct. He noted that this gem was chon shape. It’s another piece that’s “come out of recircu-
a prized possession of an original mine owner. lation,” says Bindra. He suspects the current marketplace
Bindra also showed us a couple more recirculated scarcity of fine blue star sapphires is due to a newly dis-
pieces: a fine 15.06 ct double-sided black opal cabochon covered heating process, leading to a lot of fine-color ex-
(figure 21) with broad flashes of play-of-color from Aus- amples being recut into faceted gems. He cites an example
tralia’s Lightning Ridge and a 19 ct antique-cut Brazilian of a 23 ct star sapphire, which might be heated and refash-
Imperial topaz. This last stone was old, unheated material ioned into a top-gem royal blue color faceted stone in the
from Ouro Preto. The stone came out of an old collection 10–15 ct range. For some people, that’s more desirable, so
and was a little over 20 ct. They were able to recut it with stars are becoming rarer. Not every stone can be heated and
minimal loss of weight to significantly improve luster and improved, as they might break due to the unforeseen way
brilliance. Bindra told us these gems underscore the impor- inclusions react to the heat.
tance of the secondary market for his inventory. While it Bindra sees strong demand from the top one percent of
doesn’t supply the day-to-day production B&B needs, it consumers. He notes that exceptional gemstones are in
does produce truly exceptional items from time to time. very high demand, as more than ever people find value in
He noted that fine colored gemstones always retain their
Figure 21. This 15.06 ct double-sided black opal cabo-

Figure 20. This fine-quality 5.06 ct trillion-cut gem is chon from Lightning Ridge was another recirculated
exceptionally large for benitoite. Photo by Robert gem offered by B&B. Photo by Eric Welch; courtesy of
Weldon/GIA; courtesy of B&B Fine Gems. B&B Fine Gems.

hard assets—fine art, classic automobiles, fine wine, high-
end watches, important diamonds—and colored gemstones
are “in that same conversation.” Bindra professes some
skepticism about the middle of the market, where com-
mercial to mid-grade material is not moving as well. At the
top end of the market, high-end colored gemstones are dif-
ficult to find because of competition from overseas—from
new markets, from new capital—especially in Southeast
Asia, where there is a cultural affinity for consuming fine
colored gemstones.
Duncan Pay
Figure 22. The range of necklaces at Mastoloni’s

Cultured pearl market update. At the 2017 AGTA show,
AGTA booth included innovative blends of different
Fran Mastoloni (Mastoloni Pearls, New York City) pro-
cultured pearl types. Photo by Eric Welch; courtesy of
vided a cultured pearl market update. Mastoloni is a 30-
Mastoloni Pearls.
year veteran of the Tucson show and well positioned to
place the ups and downs of demand in context. There is a
real desire for cultured pearls, he says, but people have
moved away from the traditional white necklace. Partly tail” necklace to illustrate the point, this time with
this is because cultured pearl sources have become much baroque pearls ranging from 16–17 mm plus down to 9
more diverse in recent years, with French Polynesia, Aus- mm. The organic shapes bring texture and movement to
tralia, the Philippines, Indonesia, and China contributing the piece and “make it fun.”
a diverse array of cultured pearl types, shapes, sizes, and The next necklace he showed us blended 9 mm baroque
colors (figure 22). white South Sea “keshi” pearls with 14 mm dark Tahitian
Mastoloni’s clients are seeking—and selecting—some- round cultured pearls, creating two transitions: from light
thing different. “They all want a unique necklace,” he says. to dark and from organic shapes to regular, round ones (fig-
And since the second quarter of 2016, he’s seen demand ure 24). A customer might own all three of these necklaces,
pick up considerably. Most clients don’t want to spend a he said, and use them for different occasions.
fortune, but they do want the piece they choose to be Also popular are variations on the wave necklace,
unique and individual. His job is to create cultured pearl which blends a uniform strand of round white Japanese
necklaces that meet their needs. akoya cultured pearls with “waves” of graduated silvery
To demonstrate, Mastoloni showed us his “Fantasy” black Tahitian cultured pearls that swell and pinch out
necklace: a standard-length strand combining South Sea and against the white of the akoyas. The mix of the different
Tahitian cultured pearls with lighter gold ones from Indone- cultured pearl types working together is what makes these
sia. This blends a repeating pattern of large (up to 16 mm)
eye-catching, semi-baroque pearls with smaller examples—
down to 8 mm—to create a large, mixed look. The use of Figure 23. The “Cocktail” necklace features a combina-
light and dark pearls also confers the ability to use the piece tion of differently sized round cultured pearls: darker
on a variety of clothing colors for day or night use, he added, Tahitian, golden Philippine, and white Australian South
so it’s very versatile. He added that products must also ap- Sea. Photo by Eric Welch; courtesy of Mastoloni Pearls.
peal to female self-purchasers, so the retail price point—less
than $2,000 per strand—is an important consideration. This
is a very important—if not the most important—market
segment for him today.
Mastoloni showed us another successful cultured pearl
line for the company, which he dubbed the “Cocktail”
necklace (figure 23). It combines Philippine gold pearls,
white Australian South Sea, and lighter Tahitian pearls to
create a long necklace that can be draped twice. Again,
every necklace is individual, so that a customer can “key
into” an individual pearl—maybe a large South Sea or
Tahitian—and identify that necklace as unique. Examples
like these, he explained, transform the concept of what a
cultured pearl necklace might be.
Customers are similarly drawn to baroque pearls for
their uniqueness. Mastoloni used a variation of the “Cock-

designs so successful. They’re the kind of “fashion-for-
ward” pieces that stylish but price-conscious consumers
pick out, he says.
Of course, sourcing pearls for all these necklaces is al-
ways a challenge for Mastoloni. One of his key considera-
tions is maintaining standards. For the “Cocktail” necklaces
he has to source pearls of the same tone, color saturation,
and luster. You just can’t throw pearls together and expect
them to look good, he explains. It requires an understanding
of the product along with consistency in both quality and
look—as well as an enormous amount of pearls.
Mastoloni asked us to look at professional women of
power on television, who frequently wear significant pearl Figure 25. A pair of 15 × 18.5 mm light pistachio
necklaces. Their choice might be a necklace combining dif- Tahitian cultured pearls, A matching-color drop shape
ferent pearl colors, or a uniform one with large round white measuring 16 × 22.7 mm makes up a trio. Photo by
South Sea or Tahitian cultured pearls. Fine pearls can be Eric Welch; courtesy of Mastoloni Pearls.
worn in the boardroom or at social gatherings—they show
that the wearer “has arrived.” No other jewelry is as ver-
satile.
Mastoloni noted that the supply of Tahitian cultured Pairs, such as the 15.0 × 18.5 mm Tahitian cultured
pearls has changed. Although there looks to be a good pearls seen in figure 25, are also perfect for unique pieces.
variety of subtle colors, he’s not seeing the same wide Their light pistachio color is one of the colors Mastoloni
range of colors as 10–15 years ago. This appears to be be- used to love but can no longer easily obtain. Serendipi-
cause the culturing process is improving and there is tously, he found a matching-color larger drop shape meas-
more control over color consistency. This is better from uring 16.0 × 22.7 mm to make up a trio (again, see figure
the point of view of investors, because consistent col- 25). According to Mastoloni, the right designer could make
ored pearls are easier to match and sell. As a result, those an incredible piece out of these.
investing in pearl cultivation find it easier to recoup His next piece was a 16 × 23 mm Tahitian drop pearl
their expenses. with a unusual shape and excellent luster. It was a near-
Next, he showed us single pearls and collection pieces, perfect teardrop with a very attractive slightly asymmetri-
starting with a 23.3 × 28.6 mm Australian baroque South cal tip, shaped almost like a drop of water. It begs attention,
Sea pearl that was solid and unblemished on all its sur- he explains, a pearl like this could be the center of atten-
faces. Single pearls like this one are unique items that a tion in a one-of-a-kind necklace. Mastoloni confirmed that
person can make their signature piece—something they such teardrop-shaped pearls are now difficult to find in
wear every day. Tahitian production. As pearl culturing techniques im-
prove, drop pearls are nowhere near as desirable—or as
profitable to produce—as spherical ones. Pearl farmers are
Figure 24. This necklace features 14 mm round more concerned with return on investment, so they are re-
Tahitian cultured pearls in the front and 9 mm ally striving to produce more round pearls. As pearl cultur-
South Sea “keshi.” Photo by Eric Welch; courtesy of ing methods improve, consequently, it becomes harder and
Mastoloni Pearls. harder to find nice teardrops—producers make a lot of less
attractive, misshapen examples which bring down the
value of the harvest.
Finally, Mastoloni showed us a top-quality pair of 15.3
mm round golden pearls. He noted that golden pearls—
which are generally produced from the gold-lipped oyster
in the Philippines and Indonesia—are still a good business,
with much stronger demand in Asia than in the U.S., al-
though the market probably peaked about a year ago. He
notes that there might be a little oversupply in Asian mar-
kets, but he expects it to be temporary. As far as quality is
concerned, Mastoloni’s advice is to look for a pearl that has
a good “gold” color—he always likens it to a “ring color”—
“you don’t want yellow and you don’t want green, but it’s
got to have a richer gold to it,” he says. Indonesia is always
going to have a steady supply of pearls, because cultivation
takes place over such a large area, he says but cautions that

there’s so much champagne and creamy greenish yellow
production that he cannot use. He notes that their supply
is getting better, but it’s not quite there yet. In his opinion,
cultivators in the Philippines are excellent at producing
gold pearls, as their waters and their oysters are better
suited for these products. Indonesia will continue to pro-
duce all colors, and Australian pearl farmers will concen-
trate on white, while those in Tahiti focus on the different
shades and nuances of black pearls.
Duncan Pay
Moroccan amethyst: Mining and marketing. Characteristic

“hourglass” color zoning (figure 26) and striking red Figure 26. These amethyst crystals from Morocco
hematite “finger” inclusions help distinguish Moroccan show the strong zoning characteristic of stones from
amethyst. Although amethyst is readily available on the this location. Photo by Eric Welch/GIA, courtesy of
market, the material from Boudi in southern Morocco Geostone Group.
stands out. While Moroccan amethyst has been featured in
G&G previously (F. Troilo et al., “Amethyst from Boudi,
Morocco,” Spring 2015, pp. 32–40), we gathered some fresh
insights on the mine and the marketing strategy for this world-renowned gem carver Glenn Lehrer, who became
material from Alaeddine Rafei, who directs grading and their marketing advisor and took Moroccan amethyst in a
sales for Geostone Group in Morocco. completely different direction. Lehrer saw the unique fea-
About 40 years ago, a group of nomads were stuck in a tures in these stones and applied his cutting styles and
storm in the Anti-Atlas mountain range of southern Mo- award-wining patented cuts to show off their beauty. The
rocco. The next morning, they found purple crystals on the special cuts also better use the rough. The yield of Moroc-
washed ground. The nomads kept bringing the crystals to can amethysts at a normal cutting factory is only about
tourist shops in the villages nearby. Seven years ago, a ge- 8%. When Lehrer’s special cuts are applied, the yield was
ologist and a cofounder of Geostone Group finally had the increase to an average of 20% for top-quality stones. He
stones analyzed by GIA and confirmed that they are also introduced these amethysts to Gems TV. Instead of
amethyst. Mining started in 2011–2012 and has continued selling the rough to gem cutters, the company now offers
on a small scale since then. From 2011 to 2015, the miners faceted stones and finished amethyst jewelry directly to
mainly used hand tools to remove the earth and extract the consumers through multiple TV programs in different
crystals. As the wall has gotten pushed further back, exca- countries.
vators have been used since 2015 to move earth. The min- As a result, Moroccan amethyst is not just another
ers are still hand picking crystals out of the removed earth. amethyst. It is now a brand and trademark of Geostone
Rough amethyst crystals are transported to the com- Group. The mine-to-market story is sold with the gems.
pany’s warehouse in Casablanca for washing and grading. The story even landed Moroccan amethyst in the Rio
Washing is done with water and acid. Rafei explained the Grande catalog. In return, the company is actively involved
four grades of Moroccan amethyst. The topmost grade con- in community building in the mining area. It only hires
sists of double-terminated single crystals, which are ex-
tremely rare. Rafei explained that only 1% of the
production is gem quality (figure 27), of which only 2% are
the double-terminated crystals. Next come the AAA grade, Figure 27. Morocco produces deep purple amethyst
which refers to the deep purple stones, and the slightly less with red hematite “finger” inclusions that can cause
purple AA grade. The AA grade is the most common on red flashes of color. Photo by Eric Welch/GIA, cour-
the market. The fourth grade is a category called Rose du tesy of Geostone Group.
Maroc. Amethysts falling to this grade have an even lighter
purple color than AA stones.
Marketing these amethysts is an even greater chal-
lenge. As people can imagine, the rough amethyst market
is very competitive, since there are many sources in differ-
ent countries. In 2013, when Geostone’s founders brought
their Moroccan amethysts to Tucson to test the market,
they went home deflated by the feedback. Large colored
stone cutting companies see little value in these
amethysts. Luckily, the founders got connected with

local Berbers to work in the mine, offering them a good
salary and health care. During the past several years, Geo-
stone Group has kept production quite stable and limited
mining crew sizes to around 20 people. Wells, roads, elec-
tricity, and designated vehicles for schoolchildren and vil-
lagers are all provided using profits from the amethyst
mine. In 2016, Geostone Group received an award from the
Moroccan government to highlight its contributions to so-
cial responsibility and gem mining in Morocco.
Emerald dealing in Afghanistan. Dealing rough gemstones

is never easy. This is true even for gemstones dealers with Figure 28. This 260 g rough emerald parcel was used
decades of experience such as Arthur Groom of Eternity to demonstrate the methodology in grading and nego-
Emerald (Ridgewood, New Jersey). Groom has traveled all tiating the purchase of rough gemstones in the field.
over the world to purchase emerald rough, and his current Photo by Eric Welch/GIA.
focus is Afghanistan. In Tucson, he demonstrated to us the
art of rough purchasing and the critical decisions that must
be made during this process.
Groom started by opening a parcel of about 260 g of $1500 per carat. Old Soviet ammunition from the war
rough emerald (figure 28), which he valued at a minimum three decades ago is still used to blast the emerald-bearing
of $600,000. He recounted the decisions he had to make in rocks due to the shortage of mining dynamite in
the field in Afghanistan when buying this parcel from min- Afghanistan. This is a problem for emerald mining, since
ers. Just as with individual colored gemstones, the first many larger stones are broken during the blasting. The bro-
thing buyers need to look at is the color of the whole par- ken crystals also limit the weight retention. For instance,
cel. Tip number one: Always spread out the stones after a perfectly preserved rough crystal can reach a yield of
you get the first impression of the parcel. This is impor- 60%, compared to only 10–15% for a broken one.
tant, because colors always look better when stones are The “Four Cs” of colored gemstones are generally the
clustered together; spreading out the stones helps the buyer same as for diamonds; however, a fifth “C” is crucial for
see the appearance of individual stones and better sort emeralds. This is clarity enhancement, a common practice
them into several groups. Next, do a quick estimation of in the emerald trade. Imperfections are removed by cutting
the percentage of each group relative to the whole parcel. them out or treating fissures to lower their visibility. In the
For example, a parcel may contain about 20% lighter-col- case of emerald, the latter is more common if higher
ored emeralds. weight retention and better proportions are desired. The
Since evaluation is the most critical step in rough pur- buyer must find the emerald’s fractures and make a deci-
chasing, there are several aspects buyers must pay extra at- sion on the possibility of enhancing them and leaving them
tention to when estimating the value of a certain piece or in the stone. The essence of rough evaluation is to com-
group. The color, the model (shape) of the crystal, and the prehensively consider all the factors and find the balance
clarity features are the three key observations that deter- between them. Mastering the process requires years of
mine your decision. When looking at color, it is especially
important to consider the crystal’s core. This is essential for
Afghan emeralds, because many of them have a lighter core Figure 29. Upon close examination of this emerald
(figure 29). Stones like these will lose color after faceting, crystal, sorted out of a large parcel during a buying
which demands extra planning and often results in a higher trip to Afghanistan, Arthur Groom found the crystal
weight loss. For emeralds, darker rough is more likely to core to be much lighter in color. Photo by Andrew
maintain its color after cutting. The lighter rough tends to Lucas.
lose color, especially when cut into melee.
The rough dealer must visualize how the finished stone
can be placed inside the rough based on the model, as it
dictates the final yield of the rough. Groom demonstrated
the types of emerald that achieve high weight retention
and others that do not. There is a large quantity of pencil-
shaped rough coming from the mountains of Afghanistan,
material once regarded as waste by some miners and buy-
ers. Groom said that with today’s technology, these stones
can be cut into melee that command on average $500–

Magnificent sapphires from Montana’s Rock Creek mine.
Potentate Mining marketing director Warren Boyd pre-
sented two magnificent sapphires (figure 31) and a suite of
fancy sapphires extracted from the company’s Rock Creek
operation in Montana in 2016. He also updated the authors
on this past mining season.
The larger of the two stones is the 12.61 ct blue sap-
phire, one of the largest of its kind from Montana. It was
faceted from a 6.37 g rough that had been heated. The
rough showed strong blue color before heating but had a
core with heavy silk inclusions. Heating dissolved the silk
and enhanced the blue color. The 7.9 ct orange oval is the
Figure 30. Buying rough emerald in the field under largest of this color ever recovered from Montana, to Boyd’s
less than ideal conditions, like a miner’s office in knowledge. This stone was faceted from a 5.2 g rough of
Kabul, requires discipline and a systematic approach. pale greenish color with a dark brownish core. While the
Photo by Andrew Lucas. rough was of very low transparency, the results of heating
were better than expected. The heated rough showed a
royal blue rim surrounding a bright orange core. Both
stones were faceted by Jeff Hapeman (Earth’s Treasury,
practice and familiarity with both the rough and the fin- Westtown, Pennsylvania).
ished stone markets. Most of the Rock Creek sapphire is small, and larger
After evaluation comes the art of negotiation. If the stones with spectacular colors are very rare. Because the bulk
buyer is enthusiastic about purchasing the parcel, the ne- of the production comes out of the ground with greenish pas-
gotiation takes off from there. Based on Groom’s experi- tel colors, natural stones with bright colors are also very valu-
ence, buyers need to stay cool and offer reasonable prices able. Boyd and his clients from Americut showed us a suite
to get what they want. The key takeaway here is that as a of natural fancy sapphires (figure 32) recovered in 2016. The
buyer, you need to know the value of the parcel and show company had produced sapphire rough from its Rock Creek
the miners that you know what you are doing. Ridicu- deposit for two consecutive mining seasons and gradually
lously low offers should be avoided, since that would im- built up a large inventory. Potentate is not planning to in-
mediately cast doubts on your capability and reputation. crease production but to keep it stable in the near future. The
Once both sides realize that a fair deal is going to happen, company installed a new water clarifier in its washing plant
the “sweet spot” for the price can be reached quickly. at Eureka Gulch, which recycles the water used in ore wash-
Dealing with gemstones is ultimately about dealing ing. The company will keep its focus on building inventory
with people. Afghan emerald miners (figure 30) are a very and bridging the gap between supply and demand.
strong-willed group. They always come together to sell Andrew Lucas and Tao Hsu
their stones. These miners do not owe banks money, so
there is no pressure for them to sell the goods. In this situ- Magnesio-axinite from Merelani, Tanzania. Colored gem-
ation, winning their trust is the key to motivating them to stone dealer Bill Vance (Vance Gems, Newark, Delaware)
bring in more stones that fit the buyers’ needs. Groom re- showed us some transparent examples of magnesio-axinite,
minded us that respecting the locals and bringing value and a rare magnesium-dominant member of the axinite group
benefit to them will guarantee good business and long-last- of minerals. Axinite has a general formula of
ing friendship. H(CaFe2+MnMg)3(Al2BSi4O16). Gem-quality magnesio-axi-
Andrew Lucas and Tao Hsu nite—also known as axinite-(Mg)—is transparent to
Figure 31. This 12.61 ct

blue sapphire and 7.90
ct orange sapphire re-
cently mined by Poten-
tate show that large
material still comes out
of Rock Creek, Mon-
tana. Photo by Eric
Welch/GIA, courtesy of
Potentate Mining.

vibrant “merthiolate flame color” red under long-wave UV
(figure 33, right). We intend to report on the chemistry of
this interesting material at a future date.
Vance said he’s attempting to name the gem “vanceite.”
Besides the fluorescence, pink to blue-violet stones display
a color shift: If you view it under a fluorescent light, you’re
going to see a blue stone, he says, under incandescent, it’s
more of a pinkish color. Yellows don’t show a color shift.
Because of the material’s color shift and strong fluores-
cence, Vance believes it appeals to a wide audience. Ac-
cording to Vance, these examples come from a very limited
occurrence in the Karo pit “D” block, Merelani Hills, Tan-
zania, where the chemistry and formation conditions were
just right, and new material is very hard to get. It comes
from the same rocks as green garnet, tanzanite, and chrome
tourmaline. He remarked that while cutting of magnesio-
axinite is straightforward, the biggest problem is getting
Figure 32. As displayed by these rough and cut sap- clean pieces to cut as almost all have inclusions. In terms
phires, Potentate’s sapphire production includes a of size, he said, anything over four carats is a “monster.”
wide variety of colors. Photo by Eric Welch/GIA, Vance also showed us a heart-shaped 10.46 ct Tanzan-
courtesy of Potentate Mining. ian “Merelani mint” garnet. He noted that it displayed a
very strong response to both long-wave and short-wave ul-
traviolet light (figure 34). Although such vanadium-bearing
grossular garnet can come from Kenya or Tanzania, most
translucent, with pale blue to pale violet, light brown to of the Kenyan material displays no long-wave UV reaction.
light pink, or yellow to golden yellow or orange color (fig- According to Vance, a strong long-wave UV reaction from
ure 33, left). More intensely colored bluish violet to purple a green grossular garnet is strong evidence of a Tanzanian
magnesio-axinite tends to have a higher RI (1.659–1.681) stone from D block. Vance explained that supply of this
than lighter-colored pink, purple, or yellow material material is now dwindling. In the beginning, prices ranged
(1.652–1.668). Magnesio-axinite has a birefringence of from $400–$500 per carat up to $800 per carat. Today’s
0.010–0.016, strong dispersion, and a Mohs hardness of 5.5. prices for equivalent material are in the range of $2,500 per
Stones with a blue color component display pale blue to carat. For a particularly fine example over six carats, prices
pale violet and pale gray pleochroic colors. Fluorescence is might be as high as $10,000 per carat. Demand is really
one of the material’s most striking properties—the gem strong, especially in China.
shows a dull red under short-wave ultraviolet light and a Duncan Pay
Figure 33. Left: A suite of eight magnesio-axinite faceted gems and one rough stone. Clockwise from top right:1.21
g orange-yellow rough, 2.72 ct. yellow-orange pear shape, 1.14 ct bicolor marquise, 1.51 ct pair of pale pinkish blue
princess-cut stones, 1.43 ct pink-blue oval, 1.83 ct light pink to light blue triangle, and at center in the pendant, a
4.22 ct. oval pink-blue gem. Right: The same magnesio-axinite gems under long-wave ultraviolet fluorescent light-
ing. Photos by Eric Welch; courtesy of Vance Gems.

Figure 34. Left: A 10.46 ct
heart-shaped green grossular
garnet of the color known as
“Merelani mint” in the
trade. Right: The same
grossular garnet displays a
strong orange reaction to
long-wave ultraviolet light.
Photo by Eric Welch; cour-
tesy of Vance Gems.
Ethical and sustainable vertical integration in the sapphire with his employees and suppliers in Sri Lanka and perfecting
market. When the recession of 2008 struck, Sheahan the processes throughout the supply chain will allow for
Stephen (Sheahan Stephen Sapphires, Inc., San Francisco, slow but sustainable business growth. His vision is to take
California) had been dealing in Sri Lankan gemstones for this model to other gem mining countries, especially in
two years. He saw his monthly cash flow drop to a third of Africa, where he is currently working in Madagascar.
what it had been when he first started traveling to Sri Tao Hsu and Andrew Lucas
Lanka to buy stones for resale in the United States. During
the recession he decided to borrow money and become ver- Exceptional freeform tanzanite and Oregon sunstone rings.
tically integrated. Stephen created a transparent supply At the Gem and Jewelry Exchange (GJX), we caught up
chain and an ethical supply model by becoming directly with gem artist Alexander Kreis (Sonja Kreis Unique Jew-
involved in mining. His new business also included heat elry, Niederworresbach, Germany). As we’ve noted in pre-
treatment, cutting, and sales. vious Tucson reports, theirs is a family business: father
Sustainable practices tie directly into Stephen’s beliefs Stefan buying the rough, mother Sonja designing the jew-
in how businesses should strive to operate. He sees lack of elry, and Alexander cutting the gems. Alexander and Sonja
sustainability as a major cause of many of society’s prob- showed us a new ring (figure 36) set with a spectacular
lems. He works with his customers in what he calls “cus- 27.20 ct vivid blue freeform tanzanite (figure 37). He ex-
tomized” vertical integration. He fills specific orders for plained that he had recently traveled to East Africa seeking
his customers directly from mines and brings the sapphires a singular piece of gem rough to inspire an exceptional jew-
to market in a controlled manner, meaning they are always elry piece. He found his inspiration in Tanzania with a 65
in his hands or an employee’s hands until they reach the ct rough tanzanite of remarkable color and clarity.
final customer. Stephen guarantees the entire supply chain Even with such a top-quality piece, Alexander cautioned
and strives to provide full documentation, including the that success is never guaranteed. Unlike baseball—where
mine the sapphires (figure 35) came from and the name of you get three chances—with cutting, it’s “one strike and
the treater and the cutter.
He focuses on education and training in Sri Lanka to ob-
tain the quality and traceability his customers expect. Train-
ing and complete trust in his Sri Lankan staff allow him to Figure 35. Untreated yellow sapphires such as this
buy stones unseen that will fill his customers’ specifica- 12.18 ct gem command a premium in the market-
tions. Stephen also believes in providing education and other place, especially when a transparent chain of custody
assistance for his employees’ families in Sri Lanka. and sustainable mining practices can be documented
Stephen finds that customers today want more informa- for the customer. Photo by Eric Welch/GIA, courtesy
tion about their jewelry. They want to know where a piece of Sheahan Stephen Sapphires.
came from, who was involved in manufacturing it and
bringing it to market, and the corporate social responsibility
practices of the company involved. With many of his cus-
tomers today willing to pay 30% more for documented eth-
ical sapphires and sapphire jewelry, he focuses on growing
his sales with sustainability-minded clients. At the same
time, he strives to be close to other wholesalers in his pric-
ing to stay competitive and grow his business.
A major part of his business investment is the time he
spends growing relationships with miners and cutters in Sri
Lanka that can take years. That is an essential element of
developing a sustainable business model for artisanal min-
ing. Stephen hopes that building long-term relationships

you’re out!” He added that the gem can never be its best if
the cutter misjudges and makes it too flat. For this one-of-
a-kind rough, classic cutting styles such as oval or cushion
shapes might be counterproductive if too much weight is
lost getting proportions right or if the deep blue color is lost.
For Alexander, cutting is an “intuitive adventure” in
which it might take months to place the first facet. The
finished stone is a nonsymmetrical shield shape with a
mixture of large and small facets that catch the light and
create a strong sense of movement. According to him, cut-
ting a freeform allows more reflections so that the gem will
exhibit maximum sparkle and display the best possible
color from the rough. Figure 37. The gem was cut from a 13 g rough tanzan-
The facet dimensions govern the size of the reflections ite of remarkable color and clarity. Photo courtesy of
from the gem’s surface, and a series of grooves on the base Sonja Kreis Unique Jewelry.
of the gem running at right angles to the facets on the top
create sharp angles, heightening the sense of movement
and drama (figure 38). For Alexander, this represents not
just the movement of the earth, but also the volume of ma- best type of jewelry mounting to accompany it. It’s not just
terial the workers had to shift to recover the gems. the size and shape that influences the gem, she noted, but
He reminded us that mining in Merelani is a challenge, the story behind it, too. This aspect is increasingly impor-
taking place at depths of up to 3,900 ft (1,189 meters). Min- tant for clients who like to buy into the romance and unique
ers can only work for short periods due to temperatures attributes of a jewelry piece. In this case, the rose gold on
reaching 122°F (50°C), and they must take regular breaks the mounting represents the heat the miners endure to re-
in designated cooling areas. cover the gems at depth in the mine, while the white gold
As a goldsmith and jewelry designer, Sonja follows a cradling the tanzanite stands for the designated cooling areas
500-year family tradition. She explained that it’s not just where the workers have a chance to recover. The diamonds
attributes like size, shape, and color that drive the design along the edges of the center gem symbolize the many
of a jewelry piece to complement the stone. She normally stones that have been moved out of the way to reach the
starts thinking about the jewelry design when Alexander tanzanite. The branching metalwork in the galleries at each
completes the stone, noting that it can change substan- end of the mounting represents the air that must be pressed
tially from the rough crystal to the finished piece. into the mine depths to sustain the miners. The little trail
When Sonja saw the finished cut tanzanite, still wet of melee diamonds around the curve of the ring’s shank adds
from the wheel, she held it and examined it from every angle a final touch of elegance to a significant piece with a large
in front of a mirror to get a feeling for the gem and for the center stone.
Figure 36. This ring is set with a spectacular 27.20 ct Figure 38. Angled grooves on the base of the 27.20 ct
vivid blue freeform tanzanite. Photo courtesy of Sonja freeform tanzanite create a dramatic sense of move-
Kreis Unique Jewelry. ment. Photo courtesy of Sonja Kreis Unique Jewelry.

Sonja also shared with us a striking ring centered on a
28.21 ct freeform Oregon sunstone from the Dust Devil
mine (figure 39). The organic outline of the mounting com-
plements the smooth contours of the gem, which rests in
an 18K rose gold “cup.” The cup is highly polished to re-
flect light back through the gem and amplify its rich orange
color. The gem itself is shot through with an array of glit-
tering copper particles that catch the light and add to the
piece’s allure.
Duncan Pay
Gem paintings from Thailand. The 2017 GJX show hosted

a new vendor that manufactures and sells gem paintings,
merchandise not typically seen in international gem and Figure 39. A 28.21 ct freeform Oregon sunstone from
jewelry shows. Only a handful of companies around the the Dust Devil mine is the centerpiece of this ring.
globe are active in this market segment. Than Thong Arts Photo courtesy of Sonja Kreis Unique Jewelry.
operates its own factory in northern Thailand. Owner
Wanlaya Suwannapirom is from a family that has been in
the jewelry industry for over 70 years. She started her gem
painting business more than a decade ago and gradually be- gradually go down to smaller sizes to fill the holes in be-
came the industry leader, with both international and do- tween the fragments. This step continues until no space
mestic clientele (figure 40). left between any fragments and the motif is completely
These paintings are handmade by highly skilled arti- colored by gems. Another challenge is to accurately and
sans in Thailand, using natural uncut gem fragments naturally reflect the human skin tone (again, see figure 41).
sourced from all over the world. The fragments come from To achieve this, painters grind quartz or calcite into very
either mining operations or cutting factories. The paint- fine powder and dye them to slightly different tones for a
ings’ bright colors are generated by colored gem fragments, very gradational transition of skin tone.
while the background and skin tones are created by dyed To cater to a wide range of customers, paintings are pro-
quartz or calcite powder. Artists also use extremely thin duced in many sizes and framed in various ways (figure 42).
24K gold leaf to color human hair and other details (figure Besides wall and desk decor, the paintings are framed in
41). Unlike many other gem paintings, the gemstones are jewelry boxes and wearable jewelry. According to Suwan-
not glued to paper or canvas but attached to the glass and napirom, smaller paintings require more attention to de-
then framed. tail. In addition, crystal glass that contains a higher
The process includes several steps: sketching the paint- concentration of lead is used on the wearables to give them
ing on paper, a reverse transfer to glass, gluing the gem frag- a glossier look than the pyrolytic glass used for decor.
ments and the dyed quartz or calcite powder, drying the Therefore, the price range of the jewelry is similar to some
glue, and framing the artwork. The middle steps are the of the larger desk decor pieces.
most critical for the quality of the final product. The frag- Customized gem paintings based on photos can also be
ments are sorted into different sizes. Painters need to fol- specially created. At the booth, we saw some photo-based
low the order to apply the coarser grains first and then portraits made by Suwannapirom and her team. Some
Figure 40. While her

family had been in-
volved in different sec-
tors of the gem industry
for more than 70 years,
Wanlaya Suwan-
napirom found a pas-
sion for gem paintings
such as the one in this
pendant. Photo by
Albert Salvato/GIA,
courtesy of Than
Thong Arts.

Figure 41. By dyeing ground quartz or calcite in
slightly different tones, a realistic gradation of skin
tones can be achieved. Here, 24K gold leaf is used to
create a beautiful hair color for the mother. Photo by
Eric Welch/GIA.
Figure 42. Than Thong Arts’ gem paintings are offered
in a variety of sizes, frames, and styles. Photo by Tao
Hsu/GIA.
prominent figures have hired her to produce their portrait
in gems, including the kings of Saudi Arabia and Thailand.
The finished products look very accurate and natural. The
gems in the paintings offered at the show came with a cer-
them show some green and blue play-of-color, which com-
tificate of authenticity issued by a Thai gemological lab.
plements the tsavorite and blue zircon accents. In return,
Andrew Lucas and Tao Hsu the strong blue and green color of highly refractive tsa-
vorite and zircon accentuates the opals’ play-of-color. The
New designs from Paula Crevoshay. Paula Crevoshay came intense orange color of the Brazilian fire opal forms a dra-
to Tucson some new designs that once again showcase her matic contrast with the Mexican stones and adds another
unique talent. layer to draw viewers’ attention. The mounting of the
The “Queen Thai” bracelet (figure 43) contains five bracelet follows the profile of the freeform opals, giving the
large freeform fire opals. Accompanying the opal are micro- jewelry its own personality.
pavé brilliant tsavorite accents and blue zircons connecting Crevoshay expressed her love of nature in two new
the bracelet sections. Four of the fire opals are from Mex- pieces from the collection. The “Swallowtail” butterfly
ico, and one is from Brazil. Crevoshay pointed out that the (figure 44, left) combines yellow diamond, black diamond,
four Mexican fire opals have better transparency, while the blue Yogo sapphire, yellow sapphire, and red spinel. She in-
Brazilian one has a more saturated orange color. All of formed us that this swallowtail is modeled specifically
Figure 43. In the “Queen Thai” bracelet, the contrast in orange color between the more intense Brazilian fire opal
(left) and the Mexican fire opals leads the eye directly to the Brazilian stone. All of the opals are complemented by
the green tsavorite. Photo by Eric Welch/GIA, courtesy of Paula Crevoshay.

after those found in Montana, a state whose natural beauty
and sapphires she admires. She noted that she always tries
to find an excuse to involve Montana sapphire in her art-
works. The various stones are mounted over a large surface
area, which allows them to optically interact with each
other. The colors of the stones accurately reproduce the
colors on the wings of these butterflies. When people move
the piece in their hands, it vividly reflects the whimsical
look of the real butterfly. The elephant head in figure 44
(right) evokes Crevoshay’s memories of visiting an ele-
phant orphanage in Sri Lanka. Captured in this pendant is
the extinct East African forest elephant, which is smaller
and less aggressive than most other species. A special tex- Figure 45. The blue play-of-color from the boulder
ture was generated on the gold to depict the elephant’s opal and the blue color from the sapphires dominate
skin. A sapphire and spinel headdress shows the royalty of one’s initial impression of the cuff bracelet. Photo by
this creature. The eyes are tsavorite, while the tasks are Eric Welch/GIA, courtesy of Paula Crevoshay.
carved mother-of-pearl.
Last but not least is the cuff bracelet featuring a large
Australian boulder opal accented with sapphire (figure 45).
To Crevoshay, “all opals are created by Monet,” since each 350 pits. There are usually 10 to 15 miners per pit. The
is a striking blend of colors. The play-of-color pattern of this miners descend 10 to 20 meters by ladder to the bottom of
opal reminds viewer of a spectacular cityscape at night. a pit, and then the pegmatites are blasted and tunneled be-
There is a subtle red linear play-of-color across the middle fore extraction. Some tunnels are linked for better air cir-
of this piece. This long marquise-shaped boulder opal has culation. Pits can be very close together, although some
blue as its deepest play-of-color, and Crevoshay selected the larger open-pit operations have excavators in the pits and
blue sapphire to go with it. According to color theory, blue carts to remove the ore. Rubellite is concentrated in pock-
is a color that recedes. The blue sapphire and the blue play- ets within the pegmatites. There are plans to increase
of-color seem to dominate the cuff bracelet but then recede mechanization, and new pockets are being discovered.
to the background to allow the other colors such as the green Miners are paid a salary and provided with food and
and the red to take center stage. tools, as well as a bonus based on the value of the produc-
Andrew Lucas and Tao Hsu tion. Sometimes miners can recover one ton of rubellite
from a large pocket. Large single crystals can easily reach 18
Tourmaline and sapphire from Nigeria. Zoe Michelou, kg. The rubellite is kept in a safe and then transported to the
marketing manager for a Nigerian mining company, gave city of Ibadan for grading every two to three weeks, depend-
an update in Tucson on the production of gemstones, par- ing on production. Currently there are three to four quality
ticularly tourmaline. She primarily handles sales and mar- categories. Most of the material is now shipped to Bangkok
keting of rubellite from Oyo State, where there are around for final grading and sale. A new facility is planned to per-
Figure 44. Left: The yel-

low diamond, black dia-
mond, blue Yogo
sapphire, yellow sap-
phire, and red spinel in
this swallowtail brooch
create a stunning color
combination. Right: This
brooch was inspired by a
visit to an elephant or-
phanage in Sri Lanka,
which adds an extra
touch of humanity to the
piece. Photos by Eric
Welch/GIA, courtesy of
Paula Crevoshay.

form the final grading in Nigeria, with training of locals to
create parcels for international clientele. Many of the rubel-
lite crystals sent to Bangkok weigh 1 kg or more.
There is also an alluvial deposit where rubellite pebbles
can be recovered without moving too much earth. In
Kwara State there is important production of indicolite and
green tourmaline, including some pink and bicolor. Most
of the mining is still near the surface, with a few deeper
pits. The first production of indicolite tourmaline sent to
Bangkok had crystals of 2–34 grams ranging from commer-
cial to very high quality.
Currently, the Nigerian is focused on selling rough
rubellite (figure 46) through the Bangkok office and trade Figure 46. High-quality rubellite rough is highly
shows to the international market, especially in East Asia. sought after in the Asian market. Photo by Eric
The clients are mainly rough dealers and manufacturers Welch/GIA, courtesy of Zoe Michelou.
from China, Hong Kong, and Thailand. This year, the com-
pany will start a training program for grading at the mining
site in Nigeria. In the future, rough will be sorted and
graded prior to shipment to Bangkok. stands 4.5 cm high. It rests on a sterling silver stand inlaid
The interests Michelou represents just bought 180 with Arizona turquoise (figure 48).
square kilometers of land in the Mambilla Plateau area of The iconic starship’s saucer section is also inlaid with
Taraba State for sapphire mining. Nigeria has produced Arizona turquoise and a concentric band of purple spiny
large quantities of fine-quality basalt-related sapphires. Ex- oyster centered on a lab-grown white opal representing the
traction will start in the spring, with sales expected by the bridge. Other oval lab-grown opal cabochons grace the ends
end of 2017 or beginning of 2018. They will only sell rough of the engine nacelles, and a turquoise cabochon represents
sapphires in Bangkok. Both primary and secondary deposits the ship’s shuttle bay. The starship portion has a loop and
will be mined, with an initial focus on the secondary de- can be detached and worn as a pendant, if desired. Freed-
posits. The washing plant will be able to handle 50 tons of land also has another handmade starship available: the
ore per hour. Good-quality material weighing 3 to 5 grams “Reliant” from the 1982 motion picture Star Trek II: The
is already being recovered. Security will be tight, and deal- Wrath of Khan along with an array of cyberpunk—or
ers from the Cameroon side of the border currently buying steampunk—ray guns that are redolent of classic U.S. sci-
from unlicensed miners will be restricted from the com- ence fiction films of the 1950s.
mercial mining area. Transparent supply from the mine to Freedland told us there’s no other reason behind the
Bangkok is an important part of the marketing strategy. considerable time and effort required to hand fabricate
The company is working out its grading system for rough these unique pieces except that he’s always liked the idea
sapphires.
The company primarily sells rough but also sells some
cut stones to manufacturing jewelers in Europe and the
United States. Michelou sees an increase in manufacturing Figure 47. A sampling of Nigerian indicolites after
jewelers buying rough, including tourmaline and sapphire, fashioning. Photo by Eric Welch/GIA, courtesy of Zoe
and having it cut instead of just purchasing finished stones. Michelou.
By doing so they can guarantee mine-to-market custodian-
ship and meet strict supply chain compliance regulations.
“Boldly go” with a Southwest-inspired Starship Enterprise.

At the Pueblo (Riverpark Inn) show, David Freedland
(David R. Freedland Jr. Designs, Tucson) showed us some
unusual pieces that stood out at his booth. Normally, you
expect Southwest-style jewelry to embrace nature or Na-
tive American influences for pieces featuring detailed in-
lays of Arizona turquoise and other materials. Although
Freedland had these in abundance, what really caught our
eye was this handmade, sterling-silver rendition of the
Starship Enterprise from the cult science fiction television
series Star Trek. The piece measures 6.5 cm in length and

Figure 48. This handmade piece fuses Southwestern Figure 49. The remote Brazilian mining community
style and its characteristic inlay techniques with a of Remedios is the site of the Bahia golden rutilated
science fiction icon: the Starship Enterprise from Star quartz mines. The town is developing a sustainable
Trek. Photo by Kevin Schumacher; courtesy David R. collective of artisanal miners. Photo by Brian Cook.
Freedland Jr. Designs.
of space and the unknown—“what’s out there,” as he says. for ecotourism, a concept that can be combined with
And he thought they would be a lot of fun to make. We gemological study and buying expeditions.
hope these designs live long and prosper.
Duncan Pay
Figure 50. A local miner holds up an example of the
Golden rutilated quartz artisanal mining community. At golden rutilated quartz found in the Remedios area.
the January 2017 Jewelry Industry Summit held in Tucson, Photo by Robert Weldon/GIA.
Brian and Kendra Cook (Nature’s Geometry) discussed
their efforts to promote sustainable mining in Brazil’s
Bahia State. In cooperation with 2,500 miners from the re-
gion, the Cooks are developing a collective in Remedios,
Novo Horizonte (figure 49). Plans to brand the region’s
unique golden rutilated quartz will be supported by a ware-
house and cutting facilities, as well as a visitor center. To
provide food security for the population, the Cooks also in-
tend to bring organic community farming to the area.
Brian Cook first visited the remote site, located in Cha-
pada Diamantina bordering the Atlantic Rain Forest and the
Serrado and Caatinga ecological communities, in 1983 as a
geology student. The trip from Salvador, Bahia’s capital, took
2½ days, and he was shown an example of golden rutilated
quartz (figure 50), which was relatively rare on the gem mar-
ket at the time. He later became an exporter of the quartz
and helped raise its profile. The Cooks have since become
landowners in Remedios, and their property includes a suc-
cessful golden rutilated quartz mine. Over the years they
have visited with their children from their home in Salvador
(now a ten-hour journey thanks to improved infrastructure)
and become trusted members of the community.
The Cooks have already helped locals formalize their
land and mining rights, and now they are turning their at-
tention to other initiatives. Their agenda includes mine
safety, certificates of origin to ensure transparency and con-
sumer satisfaction, and teaching cutting and polishing
gemstones. They especially seek to empower local women,
who already sell rough gems at the local markets, through
lapidary and beadmaking training. The community’s prox-
imity to the Atlantic Rain Forest makes it an ideal location

Brian and Kendra Cook are seeking investors and cor- ested in sharing their stories on this topic should con-
porate sponsorship for their community. Learn more about tact [email protected].
their work at bahiainitiative.com. • Members of the sales associate and consumer edu-
Jennifer-Lynn Archuleta cation team, led by Kevin Reilly (Platinum Guild In-
GIA, Carlsbad ternational, New York), have developed an outline
for training purposes and a proposal for content
Jewelry Industry Summit. The second Jewelry Industry providers and potential sponsors.
Summit was held in Tucson, Arizona, January 29–30. The • The sustainability in jewelry committee, headed by
meeting was a continuation of the first summit, held in Mike Pace (New York), has distributed a survey
New York in March 2016, though new initiatives emerged about perceptions of sustainability to 800 fine jew-
from the Tucson gathering. elry buyers who identify as “millennials.”
Opening remarks by Doug Hucker and Jeff Bilgore • The team working to fight silicosis and industrial
(AGTA) were followed by three speakers from outside the diseases among gemstone cutters, led by Eric Braun-
gem and jewelry industry. Lisa Manley (Sustainability wart (Columbia Gem House, Portland, Oregon), is
Strategy and Cone Communications, Boston) discussed creating pilot programs with three NGOs as well as
exploring a company’s strengths, opportunities, and as- the AGTA.
pirations, calling sustainability “a team sport.” Thea
Polancic (Conscious Capitalism, Chicago) spoke on Among the new initiatives proposed at the summit
imagining the industry’s future, citing specific compa- were an effort to find a safe, affordable alternative for mer-
nies that have used their leverage in their industries to cury in gold mining; a development index to measure the
make a difference in the world. Bob Mitchell (Electronic positive and negative effects of the gem and jewelry indus-
Industry Citizenship Coalition, Washington DC) pre- try on a nation’s economy; and a plan to build an agro-geo-
sented on pinpointing possibilities for successful and sus- ecotourism movement in Thailand.
tainable business practices. The location and dates of the next Jewelry Industry
The majority of the summit was focused on committee Summit will be announced later in 2017.
updates from the first meeting and creating new initiatives
Jennifer-Lynn Archuleta
to be worked on. Carried over from 2016:
• Brian Cook (Nature’s Geometry, Tucson) is develop-
ing a model community for golden rutile miners in
Bahia State, Brazil. A construction site has been cho- REGULAR FEATURES
sen, and negotiations with the municipal govern-
ment for lapidary equipment have begun. See p. 136
COLORED STONES AND ORGANIC MATERIALS
for more information.
• The education committee, headed by Christina
Pyrope-spessartine color-change garnet with a high grossu-
Miller (College Corner, Ohio) is creating a glossary
lar component. During the 2016 Tucson Gem and Mineral
that can help members of the industry and con-
Show, Jeff Aylor provided several transparent rough gar-
sumers understand the language of sustainability,
nets, reportedly from Sri Lanka, to GIA for scientific ex-
ethical sourcing, and responsible practices.
amination. We noticed that the rough garnets showed
• The chain of custody and due diligence committee obvious color change from yellow-green under daylight
drafted guidelines for suppliers seeking responsible equivalent lighting to orange under incandescent illumi-
sources. nation. Three rough garnets were polished into wafers and
• The harmonization team is charged with establish- analyzed to fully understand the chemistry and color be-
ing best practices among the existing standards havior of the material.
across industries. Under the lead of Mark Hanna Standard gemological testing of three garnet wafers re-
(Richline, Providence, Rhode Island) and Richard vealed an RI from 1.76 to 1.77; hydrostatic SG ranged from
Nehls (LA Rocks, Los Angeles), the team has collab- 3.93 to 4.00. Fluorescence was inert to long-wave and
orated with other organizations to learn from their short-wave UV light. The stones did not show any
challenges and successes. pleochroism when utilizing the dichroscope. Using a hand-
• #ResponsibleJewelryStories, the marketing arm of held spectroscope, absorption lines in the blue and violet
the summit, is led by Cleo Zancope Gnatek (Jane section, and very weak absorption bands at 520 and 573
Taylor Jewelry, New York), Dana Bronfman (Dana nm, were observed. Microscopic examination showed in-
Bronfman Fine Jewelry, New York), and Cecilia Gard- tersecting long and short needles throughout. Graphite and
ner (New York). This team has created a website to quartz inclusions were also observed. All of these proper-
share stories of responsible sourcing, ethical prac- ties are consistent with garnet.
tices, and sustainability. Industry members inter- The chemical composition for each wafer was obtained

A D65 A D65 A D65
Figure 51. Top: Photos of three garnet wafers under incandescent light and fluorescent light (a daylight-equiva-
lent imitation). Bottom: Calculated color panels of three garnet wafers under CIE A illumination and CIE D65
illumination. Photos by Ziyin Sun.
with a ThermoFisher iCAP Q ICP-MS coupled with a New in the CIE 1976 color circle found online. The color change
Wave Research UP-213 laser ablation unit. We concluded of this material is not very significant, but is definitely ob-
that this type of material contains 39.79%–52.66% spes- servable.
sartine, 21.71%–32.03% pyrope, 3.17%–7.28% almandine, Ziyin Sun and Jonathan Muyal
16.87%–22.46% grossular, 0.38%–0.72% goldmanite, and GIA, Carlsbad
0.11%–0.19% uvarovite (see detailed table at
Aaron C. Palke
https://www.gia.edu/gems-gemology/spring-2017-gem-
University of Queensland and Queensland Museum
news-pyrope-high-grossular-component). Based on the gar-
Brisbane, Australia
net classification from Stockton and Manson (“A proposed
new classification for gem-quality garnets,” Winter 1985 SYNTHETICS AND SIMULANTS
G&G, pp. 205–218), this material should be classified as
pyrope-spessartine. The V2O3, MgO, and MnO oxide wt.% Plastic amber imitation in a rosary. The Dubai Central
are similar to the group 5 color change garnets previously Laboratory (DCL) receives almost all types of ambers, nat-
reported (K. Schmetzer et al., “Color-change garnets from ural and treated (e.g. heated, dyed, reconstructed) as well
Madagascar: Variation of chemical, spectroscopic and col- as amber imitations for identification. Amber rosaries
orimetric properties, The Journal of Gemmology, Vol. 31, (prayer beads) are increasingly popular in Middle East coun-
No. 5–8, 2009, pp. 258–259). However, these garnets con- tries, especially the United Arab Emirates and Arabian
tains much higher grossular component (16.78%–22.46%) Gulf countries.
than those reported by Schmetzer et al. (3.02%–10.23% Recently DCL received a rosary for amber identifica-
grossular). To our knowledge, the chemical composition of tion. The yellow and yellowish brown color, structure, and
this type of color-change garnet has never been reported size of the 102 beads were identical (figure 52, left). At first
before. glance, even with a well-trained eye, this rosary could eas-
The color of the three wafers was quantitatively calcu- ily be misidentified as exclusively natural amber. Testing
lated by using the visible spectra (Z. Sun et al., “Vanadium revealed that the rosary was strung with 71 heat-treated
and chromium bearing pink pyrope garnet: characteriza- amber beads and 31 Bakelite plastic beads.
tion and quantitative colorimetric analysis,” Winter 2015 Spot RIs of the amber beads were 1.53; observation be-
G&G, pp. 348–369). The color of each wafer can be pre- tween crossed polarization of most beads revealed anom-
sented by using CIE L*a*b* and RGB color coordinates, as alous double refraction with strained colors. Stress spangles
seen online in table 2. The calculated color panels of three and brown thread-like substances are the most noticeable
wafers under daylight-equivalent lighting (CIE D65 illumi- inclusions in these beads; these features are commonly seen
nation) and incandescent lighting (CIE A illumination) in heat-treated amber. The spot RIs of the plastic beads was
were shown in the bottom row of figure 51. The three around 1.64. Aggregate and anomalous double refraction re-
wafers were also photographed under both fluorescent light action was observed under the polariscope and gas bubbles
and incandescent light to compare with the panels in the and swirl marks were the main inclusions.
top row of figure 51. The most interesting aspect of the specimen was its
One way to judge the quality of a color-change stone is reaction under long-wave UV. The amber showed moder-
to plot the color pair in the CIE 1976 color circle. Good ate to strong chalky greenish blue and yellow fluores-
color-change pairings show a large hue angle difference, cence, but the plastic appeared brown (figure 52,
small chroma difference, and large chroma values (see table right).This reaction was very useful for differentiating be-
2). The color coordinates of the three wafers were plotted tween the two materials.

Figure 52. Left: The rosary
shows yellow to yellowish
brown color in daylight-
equivalent lighting (left).
Right: The rosary under
long-wave UV. The amber
shows a moderate to
strong chalky greenish
blue and yellow reaction,
but the plastic appears
brown. Photo by Nazar
Ahmed Ambalathveettil.
To confirm the identity of each material, we analyzed been on the market for over 10 years. To the best of our
their infrared spectra with attenuated total reflectance knowledge, nearly 95% of the studied red HPHT synthetic
(ATR) mode (figure 53). Spectrum A displayed main peaks diamonds treated by the multi-treatment process are clas-
at 1716.4 and 1161.1 cm–1, consistent with amber. The sified as type Ib or type IIa. Recently, we examined a treated
main peaks of Spectrum B were at 1650.8, 1600.7, and red HPHT-grown synthetic diamond which was confirmed
1218.8 cm–1. After comparison with spectrum A and the to be mixed type (IaA + IaB + Ib).
spectra of other plastics, spectrum B confirmed the spot The 0.27 ct round brilliant sample, graded as Fancy
RI’s identification of this sample as Bakelite plastic. Dark red (figure 54), had a metallic inclusion under the
This rosary is an excellent example of the importance table. The DiamondView image showed strong red fluores-
of testing and disclosure, because even an experienced per- cence and cuboctahedral growth structure; the latter is a
son could not differentiate between the amber and Bakelite typical feature of HPHT-grown diamonds (figure 55). Ni-re-
beads with the unaided eye. lated defects at 883.1/884.8 nm and 793.6 nm were revealed
Nazar Ahmed Ambalathveettil ([email protected]), by a photoluminescence (PL) spectrum at liquid-nitrogen
Mohamed Karam, and Sutas Singbamroong temperature with two different laser excitations (532 and
Gemstone Unit, Dubai Central Laboratory 785 nm). The H2 defect corresponding to a sharp peak at
Dubai, United Arab Emirates 986.2 nm and the NV center associated with a strong peak
at 637 nm were also found. In addition, the Fourier-trans-
Mixed-type treated red HPHT synthetic diamond. A mul- form infrared (FTIR) absorption spectrum (figure 56) con-
tiple-treatment process involving irradiation and HPHT firmed the material was an uncommon type of IaA + IaB +
annealing can produce pink to red color in natural and syn- Ib, with absorptions at 1282 cm–1 (A center), at 1175 cm–1
thetic diamonds; specimens treated by this process have (B center), and at 1131 and 1344 cm–1 (C center), respectively
(again, see figure 56). This mixed type is similar to that ob-
served in treated natural diamonds. A moderate 1450 cm–1
Figure 53. IR spectra of the two specimens, with ATR
method applied. Spectrum A represents natural
amber, with typical peaks at 1716.4 and 1161.1 cm–1. Figure 54. This 0.27 ct Fancy Dark red specimen is a
Spectrum B’s main peaks at 1650.8, 1600.7, and multi-step treated HPHT synthetic diamond. Photo
1218.8 cm–1 indicate that it is Bakelite plastic. by Meili Wang.
IR SPECTRA
102
TRANSMITTANCE (%)
98
94
90
86
82
100 1716.4 1161.1
97
94
91
88
85 1218.8
84 1650.8
81
79
76 1600.7
3800 3500 3200 2900 2600 2300 2000 1700 1400 1100 800
WAVENUMBER (cm–1)

Shigley, “The ‘type’ classification system of diamonds and
its importance in gemology,” Summer 2009 G&G, pp, 96–
111), such diamonds with mixed type Ia + Ib are available
in the jewelry market, as exemplified by this study. There-
fore, caution is still needed to distinguish such treated syn-
thetic diamonds from treated stones of natural origin, as
they may show similar characteristics.
Meili Wang and Guanghai Shi
China University of Geosciences, Beijing
Joe C.C. Yuan
Taidiam Technology (Zhengzhou) Ltd., China
Figure 55. The DiamondView fluorescence image

shows cuboctahedral growth structure, a typical fea-
TREATMENTS
ture in HPHT-grown diamonds. Photo by Wen Han.
Filled calcite with pronounced phosphorescence. Asterism,
chatoyancy, color change, and play-of-color are phenomena
absorption (H1a), which correlates with irradiation and sub- that are caused by the interaction of light with the host’s
sequent annealing (Clark et al., “Absorption spectra of nat- structure or from its inclusions. Fluorescence, on the other
ural and irradiated diamonds,” Proceedings of the Royal hand, is the emission of light by a substance that has ab-
Society of London A, Vol. 234, 1956, pp. 363–381), was also sorbed light or other electromagnetic radiation. Fluores-
present. Thus, this diamond was identified as a multi-step cent materials immediately cease to glow when the
treated HPHT synthetic diamond, with irradiation and radiation source is turned off; phosphorescent specimens
HPHT annealing indicated as treatment processes. continue to emit light for variable periods after the radia-
Although most of the recent treated red HPHT syn- tion ceases. Previous studies (Spring 2005 Lab Notes, pp.
thetic diamonds are reported to be type Ib or type IIa (e.g, 46–47; Summer 2012 Lab Notes, pp. 139–140) have re-
Fall 2016 Lab Notes, pp. 308–309; C.M. Breeding and J.E. ported on the misnomer “night glowing pearls,” which are
in fact not pearls at all but round rocks of mineral species
that exhibit phosphorescence. Most of this material in the
Chinese market has been treated in a variety of ways and
Figure 56. FTIR spectroscopy reveals that this dia- in fact does not phosphoresce naturally.
mond is type Ia + Ib, with the existence of aggregated
A partially carved white ovoid with a greasy luster
nitrogen atoms (A center at 1282 cm–1 and B center at
and golden paint within the low-relief carved features
1175 cm–1) and isolated nitrogen (C center at 1131
(figure 57, left) was submitted for identification to the Lai
and 1344 cm–1) impurities. The H1a absorption at
Tai-An Gem Lab. The submitting client claimed it was a
1450 cm–1, correlating with irradiation and subse-
“night glowing pearl.” The object weighed 796.71 ct and
quent annealing, is also observed.
measured 48 mm in diameter. A spot RI reading of 1.59
and an SG of approximately 2.66 indicated that the object
FTIR SPECTRUM was most likely calcite. This was later confirmed by
2.0 FTIR and Raman (figure 58) analysis. Initial exposure to
1282 short-wave (254 nm) and long-wave (365 nm) UV light
1.8
(cabinet and lamp respectively) failed to elicit any initial
1.6
1131 visible reaction; however, both short-wave and long-
ABSORBANCE
1175 wave UV did initiate some fluorescence after about 15

1.4
minutes. When observed in a dark environment, moder-
1.2
1331 1098 ate green phosphorescence lasting more than 10 minutes
1.0 1344 was noted, with strong whitish phosphorescence concen-
0.8 trated in one area (see figure 57, right).
1450
0.6
Microscopic inspection of the area showing the white
phosphorescence revealed that the reaction originated from
0.4
within the object, and did not emanate from the entire
1600 1500 1400 1300 1200 1100 1000 900
outer surface. The microscope revealed a few tiny bubbles
WAVENUMBER (cm–1) comprised of solid epoxy resin (figure 59, left). This translu-
cent epoxy acted as a plug to an opening into the (assumed

Figure 57. The calcite
object (left, in daylight-
equivalent light) show-
ing a moderate green
overall phosphorescence
after short-wave UV ex-
posure (right). The ob-
ject emits a very strong
whitish phosphores-
cence from one area, as
seen in the center of the
right image. Photos by
Lai Tai-An Gem Lab.
hollowed out) interior, where a highly phosphorescent ma- plug the void was itself a piece of jadeite. Since destructive
terial was placed to create the calcite’s glowing effect. It testing was not permitted on either item, analysis of their
was therefore understandable why this small capped area interiors was not possible. This technique of placing a
showed the strongest reaction (figure 59, right). phosphorescent material inside a hollowed-out stone ob-
The trade name “night glowing pearl” used in Asian ject appears to be applicable to any material, so those in
markets mostly refers to natural fluorite, but some fluorite the trade should be aware when examining similar objects.
is treated with a coating containing rare earth elements Larry Tai-An Lai ([email protected])
(REE) such as europium (Eu), dysprosium (Dy), or Lai Tai-An Gem Laboratory, Taipei
neodymium (Nd) to add or enhance phosphorescence. The
sample submitted was calcite, consisting mainly of cal-
cium, as confirmed by energy dispersive X-ray fluorescence ERRATUM
(EDXRF). However, no Eu, Dy, Nd, or any other REE were
detected, proving the material was not luminiferous. It In the J.C. Zwaan et al. Winter 2015 article on alluvial sap-
therefore seems more logical that the phosphorescence phires from Montana, in the lower right corner of the Mg-
originated from material below the ovoid’s surface. Fe-Ti ternary plot (p. 385), the label “Ti × 100” should be
It is interesting to note that the same client also sub- “Ti × 10.”
mitted a jadeite jade sphere that was fashioned and treated
in a similar manner. The cap on the jadeite piece used to
Figure 59. Left: Tiny bubbles were observed in the

cavity under magnification. Field of view 6.2 mm.
Figure 58. Raman analysis identified the material as Right: The filled cavity showed stronger phosphores-
calcite (peaks at 279 and 1086 cm–1, blue trace). Cavi- cence because the epoxy resin is more transparent
ties were filled with epoxy resin (peaks at 638, 1112, than its calcite host. Field of view 11.6 mm. Photos
and 1608 cm–1, pink trace). by Lai Tai-An Gem Lab.
R AMAN SPECTRA
60000
Calcite
INTENSITY (COUNTS)
50000 Epoxy
40000
30000
1112 1608
1086
20000
638
10000 279
0
100 300 500 700 900 1100 1300 1500 1700 1900 2100
WAVENUMBER (cm–1)

Nacken Synthetic Emeralds: Additional Information A new development, however, serves to unveil further
In a recently published study on synthetic emeralds grown details. A document authored by Nacken and dated May
by Prof. R. Nacken (K. Schmetzer et al., Winter 2016 G&G, 30, 1937, was recently found in the historical archive of the
pp. 368–392), three different types of flux-grown crystals city of Idar-Oberstein (Stadtarchiv Idar-Oberstein, Depart-
and faceted samples were described. The vast majority of ment 2b, File I B, No. 94). The four-page summary, written
the synthetic emeralds, comprising several hundred sam- to the head of the local administration for the District of
ples, fell within two groups designated types 1 and 2. Most Birkenfeld, reports goals and tasks of the Institute for Gem-
of these had been obtained either from a parcel donated by stone Research. In the section covering scientific research,
Nacken to the Deutsches Museum in Munich or from ma- the following is written (translated from German): “Exam-
terial that had remained in Nacken’s family and was now ination of properties related to crystal growth and synthesis
in the possession of his grandsons. Upon examination, all of gemstones. These experiments already resulted in the
crystals and faceted samples of types 1 and 2 were found to production of synthetic emeralds.”
contain irregularly shaped seeds of colorless beryl. In con- From this short statement, it becomes apparent that
trast, only a few synthetic emeralds of another type, desig- growth of synthetic gem materials was one of the specific
nated type 3, were discovered. These lacked colorless seeds research topics undertaken or at least planned at the Insti-
and had come from the collection of the late F.H Pough, tute. Equally logical then is the corollary that Wild, being
who in turn had acquired the crystals from G.O. Wild. Ex- the local representative in Idar-Oberstein, would have had
isting documentation shows that Wild possessed a parcel access to the synthetic materials grown in relation to the
of Nacken synthetic emeralds during the period after 1945 Institute’s research projects. Although no additional infor-
and donated various samples to friends and colleagues for mation is provided in Nacken’s report, the timing would
research and related purposes. support a conclusion that at least some experiments related
In 1936, the German government decided to affiliate to the growth of synthetic emeralds were performed in the
the private Institute for Gemstone Research, located in 1930s, with Nacken attempting to improve the technique
Idar-Oberstein and directed by G.O. Wild, with the Univer- by which he had successfully synthesized emeralds in the
sity of Frankfurt. Nacken was named director of the joint 1920s. Presumably, he would have tried to avoid larger col-
entity, with Wild remaining the local leader for activities orless zones within the synthetic crystals, as such areas are
in Idar-Oberstein. A close collaboration between Wild and easily observable, especially in faceted samples. Thus, tak-
Nacken during 1936 and 1937 has long been apparent from ing into account all known circumstances, current evi-
known writings, but no direct information about the syn- dence would seem to suggest that Nacken’s type 3
thetic emeralds in Wild’s possession has previously been flux-grown synthetic emeralds without colorless seeds
available. Wild’s son, K.E. Wild, had seen the parcel with were a result of experiments performed in the 1930s, sub-
Nacken synthetic emeralds in the 1950s and 1960s, while sequent to the growth of type 1 and type 2 synthetic emer-
working in his father’s office, but no further background alds with colorless seeds in the 1920s.
about those samples was given to him at the time (K.E. Karl Schmetzer
Wild, pers. comms., 2016, 2017). Petershausen, Germany
For More Coverage of Tucson 2017

Watch exclusive videos from the gem shows, featuring
interviews and insider insight, visit https://www.gia.edu/
gems-gemology/spring-2017-gemnews-tucson-overview
or scan the QR code on the right.

Spring 2017 Gems Gemology v2 PDF

Uploaded by

Copyright:

Available Formats

Spring 2017 Gems Gemology v2 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Spring 2017 Gems Gemology v2 PDF

Uploaded by

Copyright:

Available Formats

SPRING 2017

THE QUARTERLY JOURNAL OF THE GEMOLOGICAL INSTITUTE OF AMERICA

Brazilian Diamonds: Past and Present

Copyright and Reprint Permission

About the Cover

NOTES AND NEW TECHNIQUES

Duncan Pay | Editor-in-Chief | [email protected]

EDITORIAL GEMS & GEMOLOGY SPRING 2017 1

In most countries, diamonds are usually recovered

2 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

mations. The diamonds are recovered by artisanal BACKGROUND

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 3

4 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 5

6 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

² white stones such as diamonds.” In another account,

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 7

8 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 9

10 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 11

12 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

Sedimentary units - intracontinental basin:

Figure 9. A simplified map of the major geological RR

mountain (or orogenic) fold belts representing the Alluvial diamonds:

logical structures occur on either side of the ocean. Carbonado:

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 13

14 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 15

company processed 9 million cubic meters of allu-

16 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

also see Abreu, 1973; Svisero, 1995). The Presidente 4.80

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 17

18 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 19

20 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

Diamonds in Metasedimentary Rocks. Diamonds

cated at Grão Mogol and Serra do Cabral. Bahia

Diamantina. This town is located near the Jequitin-

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 21

Carbonado. Around 1841, miners in Chapada Dia-

22 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 23

24 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 25

26 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 27

28 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 29

30 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 31

32 BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017

For More on Brazilian Diamonds

BRAZILIAN DIAMONDS GEMS & GEMOLOGY SPRING 2017 33

CUPRIAN LIDDICOATITE TOURMALINE

34 CUPRIAN LIDDICOATITE TOURMALINE GEMS & GEMOLOGY SPRING 2017

CUPRIAN LIDDICOATITE TOURMALINE GEMS & GEMOLOGY SPRING 2017 35

atoms per formula unit, 27 O + 4 OH anions normalization

36 CUPRIAN LIDDICOATITE TOURMALINE GEMS & GEMOLOGY SPRING 2017