Gems and Gemology Summer 2014 PDF

SUMMER 2014
VOLUME L
THE QUARTERLY JOURNAL OF THE GEMOLOGICAL INSTITUTE OF AMERICA
The Future of Botswanas Diamond Industry

Determining Emerald Origin from Multiphase Inclusions
Cause of Color in Rare Blue Amber
Summer 2014
VOLUME 50, No. 2
EDITORIAL
95
Challenges and Opportunities

Duncan Pay
FEATURE ARTICLES
96
Botswanas Scintillating Moment

Robert Weldon and Russell Shor
Examines the history of diamond production in Botswana and the long-term outlook for
beneficiation through value-added industries.
pg. 109
114
Three-Phase Inclusions in Emerald and Their Impact on Origin Determination

Sudarat Saeseaw, Vincent Pardieu, and Supharart Sangsawong
Discusses the role of three-phase inclusions in emerald, in combination with
spectroscopy and trace-element analysis, in determining geographic origin.
NOTES AND NEW TECHNIQUES

pg. 115
134
Color Phenomena of Blue Amber

Yan Liu, Guanghai Shi, and Shen Wang
Investigates the presence of fluorescence, rather than iridescence, as the cause of color in
blue amber from the Dominican Republic and Indonesia.
142
Experimental Studies on the Heat Treatment of Baltic Amber

Yamei Wang, Mingxing Yang, and Yiping Yang
Demonstrates the procedures for heating amber to achieve golden, red, sun spark, and
beeswax effects, and provides identification criteria for these products.
pg. 135
80TH ANNIVERSARY SPECIAL

170
Celebrating G&Gs 80th Year and the Artistry of Harold and Erica Van Pelt
REGULAR FEATURES
141
151
Thank You, Donors

Lab Notes
Chameleon diamond with nickel absorption band Long-term durability of CVD
synthetic film on natural diamond Star opal Shell pearl as pearl imitation Yellow
CVD synthetic diamond Flux-grown synthetic ruby with hydrothermal synthetic seed
crystal Rare faceted wurtzite Tenebrescent zircon
pg. 167
158
Gem News International

Jadeite bangle with the appearance of polymer treatment Prasiolite with inclusion
influenced by Brazil-law twinning Record-size natural moissanite crystals discovered
in Israel Unusual optical effect in blue sapphire Assemblage of synthetic ruby in
calcite matrix Composite coral veneer glued to artificial matrix Dyed bone as a red
coral imitation Bumble Bee jasper from Indonesia Pyrex bracelets Shattuckite
from the DRC Shanghai jewelry show Museum exhibits at GIA New York
Editorial Staff
Editor-in-Chief
Duncan Pay
Associate Editors
Jennifer-Lynn Archuleta
[email protected]
[email protected]
Managing Editor
Justin Hunter
Paula Bonilla
[email protected]
Editorial Assistants
Brooke Goedert
Nathan Renfro
Editor
Stuart D. Overlin
[email protected]
Technical Editors
Tao Z. Hsu
[email protected]
Editors, Lab Notes

Thomas M. Moses
Shane F. McClure
Contributing Editors
James E. Shigley
Andy Lucas
Donna Beaton
Editor-in-Chief Emeritus
Alice S. Keller
Customer Service
Martha Erickson
(760) 603-4502
[email protected]
Jennifer Stone-Sundberg
Database Coverage
Gems & Gemology is abstracted in Thomson
Reuters products (Current Contents: Physical,
Chemical & Earth Sciences and Science Citation
IndexExpanded, including the Web of Knowledge) and other databases. For a complete list of
sources abstracting G&G, go to gia.edu/gemsgemology.
Production Staff
Photographer
Robert Weldon
Image Specialist
Matt Hatch
Production Supervisor
Richard Canedo
Illustrators
Christopher Cruz
Larry Lavitt
Video Production
Pedro Padua
Nancy Powers
Subscriptions
Copies of the current issue may be purchased for
$29.95 plus shipping. Subscriptions are $79.99 for
one year (4 issues) in the U.S. and $99.99 elsewhere. Canadian subscribers should add GST.
Discounts are available for group subscriptions,
GIA alumni, and current GIA students. For institutional rates, contact Customer Service.
To purchase print subscriptions, visit store.gia.edu
or contact Customer Service.
[email protected]
Creative Director
Faizah Bhatti
gia.edu/gems-gemology
Manuscript Submissions
Gems & Gemology welcomes the submission of
articles on all aspects of the field. Please see the
Guidelines for Authors at gia.edu/gandg or contact the Managing Editor. Letters on articles published in Gems & Gemology are also welcome.
Editorial Review Board
Copyright and Reprint Permission

Abstracting is permitted with credit to the source.
Libraries are permitted to photocopy beyond the
limits of U.S. copyright law for private use of patrons. Instructors are permitted to photocopy isolated articles for noncommercial classroom use
without fee. Copying of the photographs by any
means other than traditional photocopying techniques (Xerox, etc.) is prohibited without the express permission of the photographer (where
listed) or author of the article in which the photo
appears (where no photographer is listed). For
other copying, reprint, or republication permission, please contact the Managing Editor.
Ahmadjan Abduriyim
Tokyo, Japan
E. Alan Jobbins
Caterham, UK
Benjamin Rondeau
Nantes, France
Shigeru Akamatsu
Tokyo, Japan
Mary L. Johnson
San Diego, California
George R. Rossman
Pasadena, California
Edward W. Boehm
Chattanooga, Tennessee
Anthony R. Kampf
Los Angeles, California
Kenneth Scarratt
Bangkok, Thailand
James E. Butler
Washington, DC
Robert E. Kane
Helena, Montana
Andy Shen
Wuhan, China
Alan T. Collins
London, UK
Stefanos Karampelas
Lucerne, Switzerland
Guanghai Shi
Beijing, China
John L. Emmett
Brush Prairie, Washington
Lore Kiefert
Lucerne, Switzerland
James E. Shigley
Carlsbad, California
Emmanuel Fritsch
Nantes, France
Ren Lu
Wuhan, China
Elisabeth Strack
Hamburg, Germany
Postmaster: Return undeliverable copies of Gems

& Gemology to GIA, The Robert Mouawad
Campus, 5345 Armada Drive, Carlsbad, CA
92008.
Elose Gaillou
Los Angeles, California
Thomas M. Moses
New York, New York
Christopher P. Smith
New York, New York
Our Canadian goods and service registration number is 126142892RT.
Jaroslav Hyrl
Prague, Czech Republic
Mark Newton
Coventry, UK
Wuyi Wang
New York, New York
Any opinions expressed in signed articles are understood to be opinions of the authors and not of
the publisher.
A.J.A. (Bram) Janse

Perth, Australia
Nathan Renfro
Carlsbad, California
Christopher M. Welbourn
Reading, UK
Gems & Gemology is published quarterly by the

Gemological Institute of America, a nonprofit educational organization for the gem and jewelry
industry.
About the Cover

Botswanas diamond-based economy could continue to thrive even after the last rough has been extracted, thanks
to the governments commitment to developing value-added industries such as polishing and jewelry manufacturing. The cover photo shows diamond cutter Mabedi Motlhasedi examining her work at Laurelton Diamonds, a
manufacturing and retail subsidiary of Tiffany & Co. that opened a plant in Gaborone in 2011. Photo by Robert
Weldon/GIA, courtesy of Tiffany & Co. and Laurelton Diamonds.
Printing is by L+L Printers, Carlsbad, CA.
GIA World Headquarters The Robert Mouawad Campus 5345 Armada Drive Carlsbad, CA 92008 USA
2014 Gemological Institute of America
All rights reserved.
ISSN 0016-626X
Challenges and Opportunities

Welcome to the second issue of our 80th year! In June, GIA held a reception at our Carlsbad
headquarters to celebrate this anniversary and to honor the artistry of Harold and Erica Van Pelt.
The reception marked the opening of a museum exhibit of their G&G cover shots, displayed
alongside many of the fabulous pieces they photographed. The Van Pelts beautiful images
graced the covers of our journal for nearly 30 years and really symbolized the stature and quality
of G&G.
Along with GIA president Susan Jacques, editor-in-chief emeritus Alice Keller, and frequent
contributor John King, I was privileged to offer a few words to our guests. While the others spoke
to G&Gs context, its place at the
Although there are undoubtedly challenges in
heart of GIAs mission, and the
living up to G&Gs 80-year legacy, there are
contributions that enriched the
journal over the years, my topic was the future.
corresponding opportunities to take our journal
Although there are undoubtedly challenges in living up
to a much wider audience.
to G&Gs 80-year legacy, there are corresponding
opportunities to take our journal to a much wider audience. While technology and social media provide much of this
new potential, its also abundantly clear how well people respond to the print journal, and the importance of face-to-face
contact in transforming ideas into reality. In the future, our strategy must incorporate both elements.
Our lead article, by GIAs Robert Weldon and Russell Shor, examines Botswanas efforts to become a major diamond
processing and jewelry manufacturing center, for the benefit of its people. The article draws from Weldons 2013 visit
to Botswana with GIA videographer Kevin Schumacher, whose extensive videos illuminate many of the locations and
people described in the paper. We invite you to view them from your smartphone or tablet by using the quick reference
(QR) code at the end of the article, or from your computer by visiting www.gia.edu/gems-gemology.
In the minds of many practicing gemologists, three-phase inclusions in emerald are inextricably linked to Colombian
origin. Yet emeralds from Afghanistan, China, and Zambia can display strikingly similar inclusions. Sudarat Saeseaw,
Supharart Sangsawong, and Vincent Pardieu, all from GIAs Bangkok lab, present detailed photomicrographs of
samples from these localities combined with spectroscopy and trace-element analysis. The authors demonstrate that a
combination of these techniques has significant potential for determining the geographic origin of emerald.
We also offer a fascinating paper examining the blue color phenomena seen in some Dominican and Indonesian
amber. Yan Liu of Liu Research Laboratories and coauthors Guanghai Shi and Shen Wang provide the gemological
basis for the surface-related blue fluorescence seen in such amber.
Our final feature articlealso on ambercomes from Yamei Wang and Mingxing Yang, both of the China
University of Geosciences in Wuhan, and Yiping Yang, a former postgraduate from the same institution. This paper
documents their heat treatment experiments on a range of amber samples, which successfully reproduced some of
the products found in todays jewelry marketplace. Along with the methods and results of their experiments, the
authors present criteria for identifying heated amber.
One of our continuing priorities is to develop synergy with the GIA website. In this issue, we entreat you to visitin
case you missed themour recent article posts and related videos. Youll hear the voices of researchers, miners, and
gem professionals, and see inside research institutions, mines, museums, and the diverse locations where gems are
bought and sold. Please visit page 172 for QR codes and links to this enhanced content.
We hope you enjoy the summer 2014 edition!
Duncan Pay | Editor-in-Chief | [email protected]
EDITORIAL
GEMS & GEMOLOGY
SUMMER 2014
95
FEATURE ARTICLES
BOTSWANAS SCINTILLATING MOMENT

Robert Weldon and Russell Shor
While South Africas Nelson Mandela inspired the world with his vision of forgiveness and racial harmony, leaders in neighboring Botswana long ago sowed the seeds of cooperation and economic development. They also demonstrated a clear understanding of how to harness the countrys natural
resourcesdiamonds, primarilyfor the good of its people. Botswana has now taken a bold step forward with a long-term plan for value-added industries that will keep the country vibrant long after its
diamond reserves are spent. While Botswanas aspirations of becoming a major diamond center are
great, so too are the challenges that lie ahead.
ust off Airport Road in Gaborone, Botswana,

where the capital citys scrubby outskirts begin, a
new complex erected in the open expanse might
well be mistaken for the airport itself. The main
building, a hulking structure surrounded by high
fences and security cameras, is considerably larger
than Sir Seretse Khama Internationals concourse.
And for good reason: The Debswana complex, representing a fifty-fifty joint venture between De Beers
and Botswana, includes a $35 million state-of-the-art
diamond sorting, valuing, and selling center, called
DTC Botswana, the worlds largest facility of its kind
(Even-Zohar, 2002; De Beers, 2008). It has the capacity to make ready for market almost 45 million carats
of rough per year, or about 40% of the worlds total
annual diamond supply. In value terms, it accounts
for about $6 billion worth of diamond rough annually
(Spektorov et al., 2013; figures 1 and 2).
Behind this momentous change is the worlds
largest diamond producer, De Beers. As of November
2013, all of De Beerss mine productionfrom South
Africa, Namibia, and Botswana, as well as Canada
is consolidated at DTC Botswana, in a pre-selling sorting and valuation process known as aggregation.
While De Beers has fifty-fifty joint partnerships with
governments in other African countries such as

Namibia and South Africa, Botswana is by far its
largest and richest source of diamonds today. Through
agreements with Botswana, the vast majority of the
diamonds are purchased by De Beers for sale to the
companys sightholders. Although De Beers was traditionally Debswanas only client, the renewal in 2006
of the 25-year mining lease for Jwaneng stipulated that
as of 2013, 1015% of production must be sold to the
government-owned Okavango Diamond Company
(ODC). This arrangement gives the government its
own direct sales channel of rough diamond to clients
around the world, bypassing the De Beers channel.
Figure 1. The Debswana complex in Gaborone includes DTC Botswana. Photo courtesy of DTC
Botswana.
See end of article for About the Authors and Acknowledgments.

GEMS & GEMOLOGY, Vol. 50, No. 2, pp. 96113,
http://dx.doi.org/10.5741/GEMS.50.2.96.
96
BOTSWANAS DIAMOND INDUSTRY
GEMS & GEMOLOGY
SUMMER 2014
Figure 2. These diamond

octahedra rough in the 3
ct range are from Jwaneng in southern
Botswana, the worlds
richest diamond mine
by value. Photo by
Robert Weldon/GIA,
courtesy of DTC
Botswana.
Perhaps even more importantly, a new 10-year sales

contract, signed in 2011, included provisions for relocating the DTCthe De Beers sales armfrom London to Gaborone. This agreement essentially locks in
the companys commitment to beneficiation (Diamonds sparkle, 2011).
Aggregation of the diamonds at DTC Botswana
precedes each of the so-called sights, where selected
buyers from around the world convene to purchase
rough allotments. The sights, held in London for
close to a century, moved to Gaborone following reportedly intense mining lease renegotiations with
Botswanas government that took place in 20042005
(Mokone et al., 2013). The move signals a historic
change and a significant upheaval of De Beerss traditional business model, shedding a remnant of the
control over diamond sales exerted by the company
since the colonial days of South Africa. As a result,

more than 60% of its London-based staff relocated to
Botswana in 2013. At a cost of over $120 million, not
to mention the loss of long-established power, it was
surely not an easy decision (Ferreira-Marques, 2013).
But the die has been cast, and both De Beers and
its parent company, Anglo American, have highlighted the positive aspects of the move. De Beers
Group CEO Philippe Mellier (figure 3) says the efforts
underscore the companys commitment to beneficiation. As we make this move, Africa is once again
at the forefront as De Beers and its partners lead the
diamond industry into a sparkling new chapter in its
illustrious history, Mellier noted (De Beers, 2013a).
He added, Africa was home to the diamond industry
when it began in the late 19th century, and De Beers
and its partners were central to that growth.
GEMS & GEMOLOGY
SUMMER 2014
97
CHANGES DOWNSTREAM
Figure 3. Philippe Mellier, CEO of De Beers Group.

Courtesy of De Beers.
For the people of Botswanathe Batswanathe

stakes are high: creating added value above and beyond the mining of resources. They expect the
changes will stimulate the growth of diamond trading
and rough diamond manufacturing, jewelry manufacture, related businesses, and other sectors of their
economy. If the plan succeeds, this evolution will
carry Botswana far beyond 2030, when the countrys
diamond reserves are expected to run out.
The changes have required De Beerss 82 sightholders

and their associates to alter their habits and travel
schedules. Sightholders have begun to make the 10
treks a year to Gaborone instead of London (figure 4).
Reports from the first DTC Botswana sight in November 2013 were largely positive, despite initial
grumbling and some logistical challenges (Bates,
2013). For example, there are currently no direct
flights to Gaborone from the United States, Asia, or
Europe. The dearth of banking infrastructure, Internet service, hotels, restaurants that will accommodate special dietary considerations, taxis, theaters,
and other amenities that sightholders were accustomed to in London is not making the transition
easy. These challenges will have to be worked out
over time.
In a sense, this is exactly what its all about, explains Kago Mmopi, communications and corporate
affairs manager at DTC Botswana (figure 5).
Botswana is a developing country, and infrastructure is also developing, and needs to develop further.
The move is inspiring the Batswana to do that. We
have been speaking to local businessmen and alerting
them of many opportunities that lie ahead. We also
know that we need to fast-track this area.
The government is also pushing hard to create
state-of-the-art infrastructure to facilitate diamond
trading in Botswana. Jacob Thamage (figure 6) is the
coordinator of the Diamond Innovation Hub, an organization set up in part by the government to coordinate banking, security, and transportation for the
Figure 4. Left: Keneilwe Dihutso, a diamond grader at DTC Botswana, displays a 54.29 ct rough octahedron from
Jwaneng prior to a visit by De Beers sightholders. Right: The 54.29 ct crystal is shown next to a 3.00 ct octahedron
from the same mine. Photo by Robert Weldon/GIA.
98
GEMS & GEMOLOGY
SUMMER 2014
Figure 5. Kago Mmopi, DTC Botswanas communications and corporate affairs manager, says the new facility can process 45 million carats of diamond per
year, about twice the capacity of Orapa House, the
sorting facility used by De Beers until recent years.
Photo by Robert Weldon/GIA.
diamond business, as well as to address the concerns

of De Beers sightholders. But it also hopes to seize
the opportunities that will arise for new diamond
manufacturers in Botswana and the attendant businesses that will begin to grow as the diamond sector
expands.
partnerships with many of those companies, Thamage explains. Weve always had aspirations and desires to beneficiate our people and add value to the
country, for the various reasons that brings: skills
building, employment, and the creation of ancillary
services. In 1981, we had only one diamond-manufacturing factory. By 1992, we had two, and four in
1994. Between 2004 and 2006, we licensed 12 more
factories to build.
As of early 2014, 21 factories are licensed to facet
diamonds in Botswana. One of them, the Indian-based
Shrenuj Company, is manufacturing jewelry there as
well (figure 7).
While the aspirations for Botswana and its people
are great, the difficulties are equally so. Logistical
challengessimply the time and distance it takes
buyers to get thereas well as the infrastructure
needed to efficiently run the diamond industry after
the end of mining, are the obvious barriers. More hotels and restaurants accommodating dietary needs of
the diverse global business community will need to
be built. International banking facilities and direct
airline links to Gaborone need to be established,
while unnecessary duplication of bureaucracy in the
import and export sector must be eliminated.
Figure 6. Jacob Thamage, coordinator of the Diamond Innovation Hub, is responsible for attracting
foreign investment in Botswana and helping provide
impetus to beneficiation projects. Photo by Robert
Weldon/GIA.
In Brief
Botswana has emerged as a significant player in the
diamond industry, with Gaborone becoming an increasingly important destination for sightholders and
business interests.
The government of Botswana, historically mindful of
the countrys importance to the world diamond market,
has taken steps to partner with operations of different
sizes to meet both industry demands and the needs of
the Batswana.
Future endeavors include moving beyond production
and into value-added industries such as cutting and
jewelry manufacturing, allowing for long-term economic growth.
Our job at the Diamond Innovation Hub is to attract the kind of companies that will bring innovation to Botswana. The idea is to also be able to go into
GEMS & GEMOLOGY
SUMMER 2014
99
Figure 7. Shrenuj, a diamond and jewelry manufacturing firm based in India, became the first company
to add value in Botswana beyond diamond manufacturing. This star motif in white and pink gold contains diamonds mined and manufactured in
Botswana. Photo courtesy of Shrenuj.
Perhaps the biggest challenge for Botswana lies in

the competitive strength of diamond centers around
the world, which have been cutting diamonds for
decades, if not centuries. Diamonds flow naturally
toward Antwerp, Tel Aviv, Mumbai, and New York.
Even though none of these manufacturing centers
have their own diamond mines (as Botswana does)
they have the necessary infrastructure and skills
base, and they are generally located closer to end consumers. This is not yet the case for Botswana. Today,
almost a million people are employed in the Indian
diamond industry, which commands almost 60% of
the polished market (Mbayi, 2011), compared to the
3,200 cutters in Botswana. Yet the Batswana are confident that this number will grow, as competencies
are developed downstream.
A lot of our youth have come into the diamond
business and excelled, and that has been a pleasant
surprise, explains Thamage. But as we all know, it
is an expensive industry that requires a lot of cash.
The schemes we have in terms of funding our youth
are not enough now for them to be able to buy diamonds for themselves. What we hope to see, as a beginning, is that a lot of those workers will someday
open an office, just like this one, become diamantaires, and start contract polishing.
100
Growth in African manufacturing, however

small, is not going unnoticed in other cutting centers. The Hindu Business Line cited beneficiation
programs in Africa as one of the reasons why the Indian market is losing its sparkle (Ashok, 2013).
Those concerns are premature, given Botswanas
minor role in manufacturing. What the country
needs is a critical mass of talent, manufacturing efficiencies, and an infrastructure that supports a cutting
industry. For consumers of such goods, inevitable
questions arise: Are the diamonds well cut? Is there
an understanding of proper yield? Can African labor
costs compete with those of India? Can the world do
business in Africa? Will consumers around the
world, heeding a call to contribute to beneficiation
in Africa, demand diamonds that are both mined and
cut there? To illuminate these questions requires a
basic understanding of the country, its people, and its
diamond richness (figure 8).
BOTSWANAS DIAMOND BUSINESS

Botswana remains a landlocked, sparsely populated
territory comprising some 600,000 square kilometers,
composed chiefly of arid savannah (figure 9). What
Botswana lacks in agricultural capacity it makes up
for with the worlds richest diamond reserves and productionfour major mines and other deposits identified through explorationas well as a relatively stable
democratic government, and a hardworking, well-edFigure 8. At Laurelton Diamonds Botswana, a cutting
subsidiary of Tiffany & Co., skills transfer brings
much-needed jobs to the local Batswana. Photo by
Robert Weldon/GIA.
GEMS & GEMOLOGY
SUMMER 2014
A F R I C A
BOTSWANA
Z I M B A B W E
NORTH-WEST
Bulawayo
N
Damtshaa mine
Letlhakane mine
Orapa
mine
B O T S W A N A
CENTRAL
Tswapong
mine
GHANZI
KWENENG
Jwaneng
mine
KGALAGADI
SOUTHERN
KGATLENG
Gaborone
SOUTH-EAST
Pretoria
0
300 km
S O U T H
A F R I C A
Johannesburg
Figure 9. Map of Botswanas diamond mines.
ucated population. Its 2013 GDP per capita of approximately $16,400 is among the highest on the continent and closing in on neighboring South Africas,
which is declining (CIA World Factbook, 2013).
At the time of its independence in 1966,
Botswana was one of Africas poorest countries, with
only a few miles of paved road and a largely agrarian
and illiterate population. The average GDP per capita
was $80. A Harvard study written in 2005 noted that
the countrys future in the early 1960s was inestimably bleak. Even Seretse Khama, Botswanas
first president (19661980), conceded that his country was beyond doubt one of the poorest nations in
Africa (Alfaro et al., 2005). Relations between Great
Britain and Botswana remained cordial through the
transition, in part because Botswana offered so little
strategic importance. In 1966, Queen Elizabeth II
knighted Khama, who had studied at Oxford and
married an Englishwoman, Ruth Williams (figure
10).
Just one year later, Botswanas fortunes radically
improved. The countrys diamond reserves, unknown at the time of independence, became evident
when the first diamond-rich kimberlite pipe was discovered by De Beers prospectors at Orapa in 1967 (De
Beers, 2014). Diamond production began in 1970
(Janse, 2007). The sudden windfall could have taken
the poverty-stricken new country in any number of
directions, but a dedication to democracy and ethical
governance prevailed. So did the understanding that
the diamond riches could not be achieved without
technical assistance. In a 1978 speech, Khama declared, We will have to learn how to share aspirations and hopes as one people, united by a common
belief in the unity of the human race. Here rests our
past, our present, and, most importantly of all, our
future.
Khamas foresight appears particularly prescient
today. Following his inauguration, he proposed the
controversial 1967 Mineral Rights Act, aimed at
vesting all mineral resources in the central government rather than in the hands of tribal leaders.
Khamas delicate negotiations, and his appeal for
tribal elders to consider the good of the nation rather
than the good of the tribe, ultimately led to passage
of the act. His skills as a negotiator and leader were
now cemented (Alfaro et al., 2005). This agreement,
as well as a progressive and open attitude toward foreign investment, would lay the foundation for
Botswanas extraordinary growth, and ultimately for
future negotiations with De Beers.
No other mining groups existed in Botswana. In
1968, a jointly owned company called Debswana (85%
owned by De Beers and 15% by the government) was
formed. For Khama and the democratic rulers who
succeeded him, Debswana became the model for
Botswanas negotiations with other diamond and
metal mining companies. By 1975, the discovery of
additional diamond-rich pipessuch as the spectacuFigure 10. In 1948, Seretse Khama caused a considerable
stir, both in England and in Botswana, when he married
an Englishwoman, Ruth Williams. Their son Ian Khama,
far left, has been Botswanas president since 2008. Photo
by Brian Seed, Time & Life Pictures/Getty Images.
GEMS & GEMOLOGY
SUMMER 2014
101
lar Jwaneng deposittruly changed the equation. Two

decades later, Botswana had become the most important diamond producer, supplying one quarter of the
worlds diamonds by value (Janse, 1996). Although the
government could have chosen to nationalize the
mines, as neighboring countries were doing (Alfaro et
al., 2005), Botswanas leadership chose to negotiate
terms that were extremely favorable to the country
while acknowledging that it lacked the capacity to
build the mining operations. Renegotiation with De
Beers in 1969 led to a new agreement, essentially resulting in fifty-fifty joint ownership (Hazelton, 2002).
That, combined with revenues from taxation of its resources, has given Botswana the lions share of the
profits from diamond mining. Conversely, for De
Beers, Botswana became a sort of insurance policy
against growing competitors, particularly in Russia
and Australia (Shor, 1990).
As much as the partnership between government
and private enterprise has been lauded, it has also
faced opposition. Critics contend that the partnership is too lopsided to be efficient and suggest that in
recent economic downturns, as in the mining contraction of 2009 and 2010, many Batswana have been
left without a safety net. Diversification of the economy and less reliance on a single cash crop would
serve as a hedge against downturns, it is believed,
particularly as diamond mining reaches an end in
Botswana. Some suggest that the partnership is in
fact counterproductive, for that very reason. The
relationship serves as a disincentive to economic diversification and political accountability, wrote
Kenneth Good, an Australian professor in global
studies at the University of Botswana who was expelled from the country in 2005 for writing a book
critical of its politics (van Wyk, 2009).
The relationship has not always been smooth, and
some historians note that Botswana has struggled, at
times contentiously, to diversify its economy and to
make De Beers its partner in the manufacturing aspect of the diamond business as well. According to
industry journalist Chaim Even-Zohar (2007), De
Beers was initially against bringing facilities to
Botswana. Even-Zohar notes that the company did
everything in its power to prevent the establishment
of more than two or three token manufacturing
units.
Shortly before the Jwaneng contract renewal,
President Festus Mogae (19982008) expressed his
countrys objective: We are looking for some improvement in the sharing of benefit. The resource is
ours, which is very important, but the investor is also
entitled to a fair return (Idex, 2004). Clearly much

has changed since 2006, when the new agreements
came into play. Today, beneficiation is a word that
ostensibly holds equal weight with De Beers and
Botswana. Mellier often singles out Botswana in
speeches dealing with the companys portfolio, noting the countrys role as a manufacturer and its rising
profile as a rough diamond trading destination (Mellier, 2014).
Seretse Khama and subsequent presidents have all
played significant roles in harnessing the countrys
natural wealth and building institutions of good governance, education (the current literacy rate is almost
85%), and anticorruption. As in several preceding
years, Botswana was rated the least corrupt African
country in 2013 by Transparency International. But
it was not only the rich diamond reserves that affected Botswanas destiny; more recently, copper,
gold, nickel, uranium, and other resources have been
discovered. A focus on bolstering secondary and
higher education, and a commitment to steering the
economy away from total dependence on diamond
revenue, began in earnest in 2008, when Ian Khama
was voted into office. The fruits of an educated populace, good governance, and skillful negotiation for
the extraction of its wealthy natural resourcesparticularly diamondshave kept Botswana on track as
one of the worlds fastest-growing countries, averaging 10% annual growth for almost three decades
(Sanchez, 2006; Farah, 2013; figure 11).
While building the framework for its people, the
government is equally committed to responding to
business needs. Ive had the chance to observe
whats happening in mining in plenty of other countries, says Jim Gowans, outgoing managing director
for Debswana. Government officials in Botswana
may not get enough credit for having done a good job.
They have an idea about what they need to do, and
theyre pretty quick to do it. Another big benefit for
Botswana is to have a democratic system and only
four presidents in their history. Certainly these factors have led to stability.
From the late 1970s into the early 21st century,
well over half of the countrys GDP was from diamonds, but a consistent policy to make the country
less dependent on mining has emerged. In 2013, the
diamond business alonenot counting copper,
nickel and gold, or other industriescontributed to
over a third of the countrys revenues. Over the years,
Botswana has been lucrative for De Beers as well. In
2000 alone, over 80% of the companys revenue
came from its Botswana portfolio53% of that from
102
GEMS & GEMOLOGY
SUMMER 2014
Figure 11. The Three Dikgosi, a frequently visited

monument in Gaborone, depicts the tribal elders who
traveled to Great Britain in 1885 to request the creation of the Bechuanaland Protectorate, the precursor
to the modern state of Botswana. Photo by Robert
Weldon/GIA.
the Jwaneng mine alone (Even-Zohar, 2002). For De

Beers, reliance on Botswanas richesand some
might argue overrelianceis also evident.
Of course, there are still enormous challenges
Botswana must contend with. In the 1990s, the
scourge of HIV/AIDS virus hit hard in the sub-Saharan
country (Weldon, 2001). Today, approximately 24% of
the population is infected with the virus, according to
the World Health Organization. Botswanas relatively
tiny population and workforce can ill afford such a devastating blow. Nevertheless, Botswanas leaders have
met the challenge with education programs on HIV
prevention, partnerships and grants with pharmaceutical companies such as Bristol-Myers Squibb, and the
construction of hospitals and treatment centers.
Grants from the Bill & Melinda Gates Foundation are
aimed at managing the symptoms of the disease or preventing its progression. Over 80% of Batswana infected
with HIV/AIDS are covered by free antiretroviral medication, and 100% have access to it. President Quett
Masire (19801998) was the first Batswana leader to
face the problem at the turn of the century, and in 2009
he declared it a national imperative to wipe out the disease. He established 2016 as the target year when no
new infections would occur. In 2011 alone, the
Botswana government spent some $385.5 million on
AIDS prevention, care, and treatment, with a special
focus on children stricken with AIDS or orphaned by
the disease. The results have been dramatic in terms
of life expectancy through earlier detection and treatment, and reductions of HIV transmission from
mother to child (figure 12). In 2012, President Khama
noted in a state of the nation address that the country

was on track to achieve a less than 1% transmission
from mother to child by 2015. Botswanas example has
been lauded throughout Africa and the rest of the
world. The expenditures come at a considerable
cost, according to Khama, though they are universally
considered utterly necessary for Botswana to move forward economically and socially.
Another challenge was the attempted relocation of
indigenous peoples in the Kalahari Desert to make
room for a potential new mine at Gope. Survival International, a UK-based non-governmental organization, campaigned for several years against developing
the mine and, in late 2010, attempted to block its licensing. According to reports, a lawsuit ended the crisis by allowing the Kalahari Bushmen to return to
their land (Botswana outraged, 2011).
In 2012, De Beers produced less than 30 million
carats worldwideanother year of retrenchment from
a high of 34 million carats in 2006due mainly to flagging global diamond demand, a lingering after-effect of
the 2008 global recession. The following year saw
major changes at the sources, such as the sale of the
Finsch mine and a slope failure at Jwaneng (Greve,
2013). Liquidity problems for many Indian manufacturers, which still cut most of Botswanas diamonds,
also slowed the momentum. For Debswana, a company dedicated to addressing safety concerns, it was
also a difficult year. The slope failure at Jwaneng in
June 2012 killed an employee, and investigations into
its cause shuttered the mine for several months. Debswana has since taken an even more aggressive stance
Figure 12. The government of Botswana ensures that
its citizens are eligible for free antiretroviral medications to treat HIV infection. Additionally, a series of
programs aimed at reducing transmission of HIV
from mother to child have been instituted. Photo by
Lucian Coman, Shutterstock.
GEMS & GEMOLOGY
SUMMER 2014
103
on safety, promulgating a zero harm culture while

improving its production figures (J. Gowans, pers.
comm., 2013).
In tandem with declining production, 2012 sales
totaled $5.5 billion, a billion less than 2011 (De Beers,
2013b). Nevertheless, the goal of transferring De
Beerss sales arm to Botswana was accomplished in
2013. The business lull may have been advantageous
by allowing a more gradual buildup of Gaborone as a
diamond sorting, valuing, and selling center. Debswana is poised for growth, particularly as new dia-
Figure 13. Pauline Paledi, executive director of the

Botswana Diamond Manufacturers Association.
Photo by Robert Weldon/GIA.
mond customers emerge in India, China, and other developing markets. De Beers estimates 5% growth in
China, Hong Kong, and India by 2017, while the U.S.
marketwhich now represents over a third of global
diamond saleswill slip by about a third of its present
portion.
Beyond mining, the creation of long-term, valueadded products, training, and employment for
Botswana is beginning to take effect, though there is
still a ways to go, says Pauline Paledi, the executive
director of the Botswana Diamond Manufacturers Association (figure 13). The organization, established in
2007, consists of some 17 full members who actually
manufacture diamonds in the country today, plus recent applications for an additional four members at
the time of this writing. The association lobbies the
Botswana government as a single voice, explains
Paledi, and works out compliance matters regarding
labor laws and visa and residency issues for foreign
manufacturing companies. The goal is to attract more
foreign investment, and eventually homegrown diamond manufacturing as well. We see growth ahead,
says Paledi. The market is getting stronger for manufacturing, and it is diversified, as we can all see.
The notion that diamonds mined on African soil
could also be manufactured and sold there, bringing
long-term sustainability to Africans, has largely
eluded the continent since South African diamonds
were discovered in 1867. For the Batswana, such
dreams now appear enticingly possible.
DIAMOND MINES OF BOTSWANA

From Gaborone, a car can reach the front gate of Jwaneng, the worlds richest diamond mine, in about
104
GEMS & GEMOLOGY
SUMMER 2014
Figure 14. This

panoramic view of the
Jwaneng mine, Debswanas largest and
most lucrative, shows a
manmade pitvisible
from outer spacethat
covers approximately
520,000 square meters
at the surface and is
more than 400 meters
deep. Photo by Robert
Weldon/GIA.
two and a half hours, a ride straight through a desolate, hilly region of savannah and occasional cattle
ranches. For those lucky enough to receive an invitation to visit the mine and proceed through security
checks and briefings, the overwhelming desire is to
peer over the edge into this open pit, a gigantic manmade hole visible from outer space (figure 14).
Jwanengwhose name means place of small
stones in Setswanawas discovered in 1973. The
pit currently descends nearly 400 meters into three
separate kimberlite pipes. It is expected to reach almost 700 meters in depth by 2017, as new cuts dig
into the kimberlitic ore in the search for diamonds.
The latest is Cut 8, which extends laterally into the
kimberlite pipes at Jwaneng and is expected to produce over 100 million carats and extend the life of
the mine until 2028. A birds-eye view of this massive removal of earth reveals the rugged face of
Botswanas principal mine by value, which has the
capacity to produce some 2.5 million carats of diamond on a monthly basis. During the recession, and
because of the 2012 slope failure, production dropped
to about half of that (Mmopi, pers. comm., 2013).
Once fully operational, Cut 8 will transform Jwaneng
into one of the worlds super-pit mines.
Jwaneng employs over 2,500 people, though they
are rarely visible, partly due to the pits immense size
and partly for security and safety reasons. Cut 8 is
expected to expand the workforce by at least 50%.
At the bottom of the pit, kimberlitic ore is broken up
through dynamite blasting on a regular basis. Miners
are also busy maneuvering the massive Komatsu 930
trucks, each the size of a house and designed to cart
out 250 tons of kimberlitic ore per haul (figure 15). A
staggering 60 million tons of earth are thus removed

in a high production year; on any given workday, traffic from the trucks forms a constant stream into and
out of the mine. Where the earth goes after processing is another logistical and engineering feat.
First, the ore is crushed into much smaller particles at a plant on the surface. The clay and mud are
removed, and the diamonds are separated from the
ore by X-ray at a completely automated recovery
plant (CARP) where, as a theft countermeasure, no
human hands touch the ore. After that, the diamonds
GEMS & GEMOLOGY
Figure 15. A Komatsu 930 truck can remove up to 250

tons of diamond-bearing rock and earth in a single
haul. Photo by Robert Weldon/GIA.
SUMMER 2014
105
produced 20.22 million carats in 2012 and treated

21.87 million tons of earthjust over a carat per ton
(De Beers, 2013b). In 2013, Debswana produced over
22 million carats; these numbers are expected to increase as global economic recovery takes hold.
Botswanas other diamond mines are independently owned rather than joint ventures between De
Beers and the government. One of these is the Lerala
mine, owned by DiamonEx, which has been stalled
due to restructuring and is expected to reopen in 2014.
Others such as the Karowe mine, operated by Lucara
Diamond Corporation, are selling diamonds through
auctions. Karowe made news by yielding 26,196 carats
for auction (Lucara completes first sale of Botswana
diamonds, 2012) and announcing the find of a 4.77 ct
blue diamond as well as several large diamonds
(Golan, 2013). Auctions are also the governments preferred method of selling diamonds, at least for now.
THE OKAVANGO DIAMOND COMPANY:

THE MARKET EXPERIMENT
Figure 16. Mixed rough from Debswanas mines must

be collected and sorted into categories by size, color,
and clarity. Photo by Robert Weldon/GIA.
are presorted, cleaned, and packaged using laser technology (again with no hands touching the rough).
The diamonds are then ready for transport to
Gaborone for sorting and valuing, and ultimately for
sale. The rough is classified by weight, shape, clarity,
and color, regardless of source. According to Kago
Mmopi, Debswanas mines alone supply between
200,000 and 500,000 carats to DTC Botswana on a
weekly basis (figure 16).
De Beers was traditionally the sole purchaser of diamonds mined in Botswana. A renegotiation in September 2011 changed that, with the government
obtaining the right to purchase a progressive 1015%
allotment of the rough to sell through its own channels over the next few years. To that end, the government established the ODC (figure 17) in 2012 and
Figure 17. The Okavango Diamond Company, a

wholly owned government company that derives its
name from Botswanas rich Okavango Delta, purchases over 10% of the rough diamonds produced by
Debswana for sale to clients around the world. This
margin will increase gradually to 15%. Photo by
Robert Weldon/GIA.
A similar procedure takes place at Debswanas

other mines, all of which are open pit. The Orapa
mine is Botswanas oldest, established in 1971, a few
years after diamonds were first discovered in a dry
northeastern part of the country. As with all De Beers
projects, the mine is owned fifty-fifty with the government. Orapa has the capacity to produce almost 20
million carats a year; in 2012, however, production
slowed to 11 million carats due to global economic
conditions. Letlhakane and Damtshaa are both much
smaller-capacity mines, producing less than a million
carats annually. Combined, Debswanas four mines
106
GEMS & GEMOLOGY
SUMMER 2014
Figure 18. Marcus Terhaar, deputy managing director

of the Okavango Diamond Company. Photo by
Robert Weldon/GIA.
appointed Toby Frears as its managing director that

same year.
As diamonds arrive weekly at Debswana, De
Beers and government negotiators settle on a value
for the goods. After the purchases are made, De Beers
aggregates Botswanas goods with its allotments from
Namibia, South Africa, and Canada.
The 10%-plus that the ODC buys on a weekly
basis is gathered and sold to clients around the world
via auction. Sales occur about 10 times a year and include parcels and single stones 10 ct or larger. It is thus
a different system from the fixed allocation contracts
implemented by De Beers and its sightholders. As
would be expected, ODC clients must be properly vetted and comply with strict guidelines to register and
qualify as buyers. Details about each companys role
in the diamond value chain, top personnel, and compliance with all diamond sector regulations are required by ODC.
For all its advantages, the auction system tends to
be risky: volatile during good economic times, and
slow to nonproductive during bearish economies.
Under such conditions, fixed contracts make it easier
to fine-tune and control the market. Marcus Terhaar
(figure 18), ODCs deputy managing director, explains that the auctions help the government derive
current market value for the products.
Auctions are going to be far more volatile than
other selling mechanisms. But thats not necessarily
a bad thing, because over a consolidated period of
time, you will find that they do very well despite the
boom-and-bust cycle.
Jacob Thamage, who also serves on the board of
the ODC, agrees, alluding to current economic cli-
mate. For ODC to launch under these conditions is

actually very good, because, like they say, Rough
seas make the best sailors. It is conditions such as
these that will really teach us what the market is,
and what it could be.
ODC officials seem open to the idea of a dual selling system that incorporates both ideologies. A lot
of companies use dual systemsmeaning that they
have options and contracts. That tends to mitigate
the risk, Thamage explains. By the middle of 2014,
we will have completed a full cycle of selling and understood the possibilities of auctions and contracts.
Terhaar highlights the beneficiation aspects that
are already taking place through the ODC: What we
have in Botswana, insofar as the diamond sector is
concerned, is a very strong diamond mining industry.
We have a very well-established sorting and valuing
operation in Debswana, and we have a good five to
seven years into diamond manufacturing. Rough
trading is clearly the missing link for Botswana, and
that is the primary focus of the ODC. It will bring
jobs. This means that 150 people are coming to our
premises every five weeks or so. That means 150
more hotel accommodations, 150 more seats on local
airlines, and 150 more customers looking for taxis,
making dinner reservations at restaurants. It will foster the growth of ancillary services: diamond reports,
banking, rough valuing, and marketing services, and
so many other types of related services.
For now, ODC is busy establishing its operation,
conducting sales, and promoting itself as a viable separate channel for rough diamond buying. This is a big
feather in Botswanas cap, says Terhaar. Ten years
ago this was all bush country. Botswana has proven
that it can be done. And while there is a lot of learning
to capture, there is certainly the motivation, and the
will, for Botswana to perform as a benchmark diamond centeras efficiently as Antwerp, Israel, or any
other diamond center in the world.
CAN A CUTTING INDUSTRY BE SUSTAINED IN

BOTSWANA?
While Botswana prospered by carefully redirecting its
diamond wealth into nation-building, establishing a
domestic diamond manufacturing industry remained
an elusive goal until recently. Neighboring South
Africa benefited from a well-established diamond
cutting industry that employed several thousand
workers. Even so, much of this industry survived because De Beers subsidized local cutting operations by
providing rough at a 10% discount (actually sparing
GEMS & GEMOLOGY
SUMMER 2014
107
buyers export taxes on rough). Some of these operations existed mainly as a means for their owners to
obtain rough allocations from De Beers. They performed minimal work on stones in the local plants
before exporting them to Israel, Antwerp, or India for
actual manufacturing (Shor, 1990; figure 19).
In the past, De Beers actively discouraged diamond manufacturing in Botswana, claiming the
costs of doing business there were not low enough to
compete with India and China. In addition, De Beers
noted, the country was too far removed from major
consumer markets (Grynberg, 2013).
De Beerss opposition began to crumble in 1990
when the company entered negotiations to renew its
five-year contract with the government to sell rough
mined by Debswana. With De Beers owning 50% of
Debswana, such contracts would seem to be a routine
matter, but in 1989 the government had raised the
issue of starting a domestic manufacturing operation.
At the same time, a number of major rough diamond
dealers were lobbying Botswanas parliament to sell
20% of the companys production outside De Beerss
network, claiming they would be willing to pay 57%
higher prices than De Beers. One of those dealers,
Maurice Tempelsman of Lazare Kaplan International
(LKI), offered to build a polishing factory if the government was willing to supply diamonds to the operation
(Shor, 1990). The government made no such promise,
but LKI built the factory anyway.
De Beers, which already employed 5,000 workers
at its three mines in Botswana, eventually agreed to
develop a cutting factory, mainly to forestall the parliament from considering this 20% market window (Shor, 1990). The operation, named Teemane,
opened in 1993, but a decade later only three others
had been built, the largest being the Lazare Kaplan
International operation (J. Thamage, pers. comm.,
2013).
Botswana assumed a much greater role in De
Beerss operations after 2001 when, as part of the
companys reorganization, it assumed a 15% stake in
the firms ownership. Botswana also captured a much
greater strategic role after De Beers began phasing out
its agreement with Alrosa, the Russian diamond
mining company, to market its productionaccounting for about 25% of De Beerss rough sales
and sold off its aging South African mines and its $5
billion diamond stockpile (Even-Zohar, 2007). By
2005, Botswana was by far the largest and most profitable part of De Beerss operations, accounting for
70% of its earnings (Grynberg, 2013). This fact was
not lost on the countrys leadership, which leveraged
108
Figure 19. Laurelton, a joint venture manufacturing

branch of Tiffany & Co., has set up shop close to the
DTC Botswana complex and the Diamond Technology Park. Today, it works with Indian cutting experts
to bring about a transfer of skills to the Batswana.
The manufacturing plant has ample room for growth,
which it clearly expects. Photo by Robert
Weldon/GIA.
an aggressive series of beneficiation measures in its

negotiations for both the sales contract with Debswana and the renewal of the 25-year lease at Jwaneng (Even-Zohar, 2007).
The agreements, reached in 2006, covered a thousand pages but contained three main requirements
that would transform operations for both sides:
De Beers would directly supply local firms that
qualified for a diamond manufacturing license.
De Beers would relocate all of its sorting and
rough sales operations from London to
Gaborone by 2009.
A certain percentage of De Beers sight goods
would be reserved for local polishing operations.
In short, the government of Botswana, which had
owned a 15% stake in De Beers since 2000, now had
sufficient clout to force the company to supply local
polishing operations, and to require large diamond
manufacturers to develop factories on its soil if they
wanted to obtain supplies of rough diamonds (Grynberg, 2013).
While the 2006 agreement provided for supplies
to 10 factories, as selected by De Beers, Botswana
GEMS & GEMOLOGY
SUMMER 2014
moved aggressively beyond De Beerss plans and issued licenses to 16 companies, which began taking
rough supplies through separate sights domestically
in January 2007. By fall of that year, six had gone into
production (Golan, 2007). On the second point, the
20082009 world financial crisis hit De Beers hard,
requiring the company to borrow $500 million from
shareholders (including $200 million from the Oppenheimer Trust, which owned 40% of the company). The crisis also forced De Beers to postpone its
relocation plans until 2013 (Krawitz, 2008; Anglo
American, 2009).
In 2008, in preparation for De Beerss arrival and
the growth of a domestic cutting industry, the diamond industry, led by South African manufacturer
Safdico with the Botswanan governments cooperation, commissioned the Diamond Technology Park.
The secure 35,000-square-meter facility was located
next to the new De Beers building and just four kilometers from Sir Seretse Khama International Airport.
It was designed to house about 15 medium-size manufacturing operations in two buildings, plus industry
services such as banking, manufacturing equipment,
and diamond gradingGIAs Gaborone laboratory
was one of the first tenantsin a third building. The
Diamond Technology Park website notes that by
2012, these operations had outgrown the complex,
causing a number of manufacturers to locate in the
industrial parks nearby.
Statistics for 2013 indicate that Botswanas drive
to create a local diamond manufacturing industry has
been a success. More than 3,000 workers are employed in polishing operations (compared to fewer
than 500 in 2006) and several thousand more through
ancillary businesses serving the diamond sector. Polished diamond exports neared $800 million and were
forecast to top $1 billion by 2015, compared to $100
million in 2008. The Diamond Technology Park was
fully rented, and the government commissioned
plans for a fourth building of 4,000 square meters to
be completed by the end of 2014 (Grynberg, 2013;
Shor, 2013). By the end of 2013, 24 manufacturers
were operating in Botswana, 21 of them receiving
rough from Debswana through local sights (figure
20).
One reason for this success is that even before De
Beers transferred all of its sorting and sales operations
to Gaborone, the Botswana sights consisted primarily
of higher-value rough that could be profitably polished (Grynberg, 2013). But the rush to open diamond
manufacturing operations in Botswana rekindled the
debate over whether diamond manufacturing there
Diamond manufacturing costs in Botswana range

from just under $40 to $60 per carat, depending on
the efficiency and technological capabilities of a
given operation. These costs include labor, utilities,
maintenance and technology support, and transportation. This cost range is much lower than
Canadas ($80 per carat) but still more than double
that of China ($17 per carat) and four to six times that
of India ($10 per carat), which polishes 92% of world
production by volume (Gregorian, 2013; figure 21).
Visits to nearly all of Botswanas large manufacturing facilities found them fully committed to developing sustainably profitable operations. These operations
began between 2006 and 2010, using expatriates from
India, China, South Africa, and Israel to train local polishers. Each operation had to be started from scratch
because no skilled workers, equipment, or financing
were available locally (R. Moses, pers. comm., 2013).
Over the next four to five years, however, local polishers largely replaced foreign workers at the wheels,
and now comprise 8090% of the workforce (figure
22). For many local Batswana, the employment opportunities have helped them extricate themselves from
poverty. In some cases, the salaries have helped feed
entire families.
The countrys largest diamond factory is run by
Eurostar, an Antwerp firm that has its roots in India.
The company now employs 520 workers; 485 of
these are locals who received a minimum of six
GEMS & GEMOLOGY
Figure 20. A cutter at Diacore Botswana examines the

initial facets made on a fancy yellow diamond. Photo
by Robert Weldon/GIA.
was economically sustainable, particularly in light of

Canadas unsuccessful attempts to develop large-scale
cutting operations (Yellowknife diamond-cutting
plants future in limbo, 2009; Danylchuk, 2013).
MANUFACTURING COSTS
SUMMER 2014
109
DIAMOND MANUFACTURING COSTS, BY COUNTRY

US$ per polished carat
100
100
80
80
70
60
40
Botswana
processing
cost
40
40
30
25
20
20
17
10
Figure 21. This chart,

prepared for the World
Bank, analyzes the percarat costs of diamond
manufacturing (including labor and operating
charges) in diamond
centers worldwide
against Botswanas costs
(Grigorian, 2012).
0
New York
Canada
Belgium
Israel
South
Africa
2,000
3,000
Russia and Armenia

CIS
Thailand
China
India
10,000
110,000
1 million
Estimated employment
600
n/a
1,100
10,000
3,500
months of training and are polishing precision-cut

round and princess diamonds 0.25 ct and up. The
company recently spent $7.5 million to upgrade its
high-tech processing equipment.
Expatriates still dominate the highly technical aspects of the cutting process: rough planning, bruting
and laser sawing, and shaping. Here, too, Batswana
trainees work beside the technicians, learning how
to use advanced equipment (figure 23).
Another large operation, employing 300 workers,
is Shrenuj, which produces half-carat and larger
rounds and has also begun a jewelry manufacturing
Figure 22. Diacore Botswana offers diamond manufacturing jobs to hundreds of Batswana. Photo by
Robert Weldon/GIA.
110
operation. The company has a program to hire hearing-impaired workers, who now comprise one-quarter of its labor forcebut is still working to match
the productivity of its Indian operations (K. Lanny,
pers. comm., 2013). Nevertheless, the operation has
become profitable as performance improves.
Labor costs are less of an issue with Safdico and
Diacore Botswana (formerly Steinmetz), which specialize in large, high-quality diamonds. Both companies have developed rigorous training programs
because each diamond must meet very high, exacting
standards. Safdico and Diacore both added that they
are in Botswana for the long term (K. Teichman and
R. Moses, pers. comms., 2013).
The government has been diligently auditing
companies to ensure that diamond factories provide
full training and employ Batswana to perform all of
the work necessary to produce finished stones (R.
Moses, pers. comm., 2013). Diamond executives give
the government high marks for introducing regulations and policies that support equipment imports,
funds transfer, building licenses, and transparency.
Several diamond firms have taken their beneficiation role beyond offering training and employment
to local workers. Safdico supports programs that train
computer technicians, schoolteachers, and other
community-building professionals in villages outside
the capital. Diacore Botswana provides microloans
to employees for school fees and conducts events for
community improvement projects. They have also
supported sporting events and employ prominent
local athletes (figure 24).
GEMS & GEMOLOGY
SUMMER 2014
Figure 23. The use of highly sophisticated planning,

sawing, and polishing technology is prevalent at the
new manufacturing plants in Gaborone. Photo by
Robert Weldon/GIA.
Beyond the high cost of labor, significant challenges to sustaining a diamond processing industry
remain. Infrastructure is still lacking. Power outages
are common, Internet service remains sluggish, and
importing or repairing equipment is still very costly
and inefficient.
Executives of manufacturing facilities in Botswana,
while acknowledging the economics and challenges,
are divided over whether the countrys diamond industry can become fully self-sufficient. But even the most
skeptical see significant potential for improving
Botswanas competitive position so that it may become a successful niche producer, like New York or
Antwerp. It is still early days, initial stages, said Erik
Figure 24. Diacore Botswana manager Kefir Teichman
(right) poses with gemologist Monica Alfred, a fiveyear veteran of the company and a member of
Botswanas national volleyball team. Photo by Robert
Weldon/GIA.
van Pul, manager of Eurostar in Botswana. We are seeing improvements all the time as workers gain more
expertise and government services and infrastructure
improves.
In 2010, for example, most Botswana diamond operations could not profitably produce polished diamonds below one carat because training and high-tech
processing equipment were not fully in place. By 2012,
the more advanced operations were producing 0.40 ct
stonesat a profit, according to van Pul. At least 10
factories were producing precision triple-excellent
cuts comparable to any diamond polishing operation
in the world. While that represents true progress, one
analyst believes that the ability to polish melee (0.10
0.20 ct) profitably is necessary to create a manufacturing industry that sustains large-scale employment (M.
van den Brande, pers. comm., 2013).
THE FUTURE
The year 2027 is the governments benchmark for developing a diamond-polishing industry that does not
depend on domestic rough (figure 25). Jwaneng is
scheduled to be redeveloped into an underground
mine that same year, which will sharply reduce production. By comparison, Australias Argyle mine fell
from a peak annual output of 42 million carats as an
open pit to 20 million carats as an underground mine
(Argyle Diamond Mine, 2013). Production will likely
continue 30 to 40 years beyond this date, but with
much smaller volume and higher cost.
Both the Botswana government and the diamond
community anticipate that rough auctions by Okavango Diamond Company will stimulate manufacturing and help build the countrys trading base.
Okavango is a government-affiliated company that
has begun selling between 12% and 15% of the countrys diamond production through monthly tender
sales. Existing diamond manufacturers say the tenders
will cut costs and provide greater access to supplies
(R. Moses, pers. comm., 2013). Another goal, according to Terhaar, is attracting smaller diamond companies and rough traders to set up business in the
country. Okavango sales have much lower purchase
requirements than De Beers sights, plus a fairly simple
application process that will enable such firms to participate. Building this trading environment will be
critical, and the government understands that it needs
to work with the diamond industry to lower infrastructure costs and improve services for the countrys
diamond polishing industry to survive beyond supplies from its mines (Grynberg, 2013).
GEMS & GEMOLOGY
SUMMER 2014
111
Indeed, one World Bank study (Grigorian, 2013)

noted that [diamond] producing countries hoping to
establish a viable cutting industry are squeezed by
competition from two directions: one from low-cost
economies such as India and China, another from
high-skills economies such as the United States, Belgium, Israel, and Canada. For any latecomer, the
challenge is plain: Either be cheaper (and work
harder) than the former, or be more knowledgeable
and skills-intensive than the latter. The way
through this competition, the study concluded, was
for these countries to achieve a manufacturing niche
by branding to provide added value to their products.
It remains to be seen whether Botswana-branded
cut diamonds will captivate consumer attention
globally. Buyers are increasingly conscious of the
products they purchase and the supply chain involved. The storyline for a Botswana brand is undoubtedly strong. Consumers, drawn to African
diamonds for well over a century, could find assurance in knowing that the diamonds they purchase
Figure 25. Botswana is capitalizing on its diamond resource riches by adding value through diamond manufacturing. Though not yet large by global standards,
the cutting industry in this sub-Saharan country
shows considerable promise. De Beers now holds its
rough diamond sales in Gaborone and has encouraged large diamantaires to set up manufacturing
plants there. Photo by Robert Weldon/GIA.
today have contributed to skills transfer in Africa,

poverty alleviation, and the dignity of employment
for Botswanas people.
For More on Botswanas Diamond Industry

Discover more about Botswanas strategies and challenges as it
develops a thriving diamond industry. G&Gs exclusive online content
provides access to video interviews and additional resources to
enhance your knowledge of this countrys importance to the diamond
and jewelry markets.
Visit www.gia.edu/gems-gemology, or scan the QR code on the right.
ABOUT THE AUTHORS

Mr. Weldon is manager of photography and visual communications, and Mr. Shor is senior industry analyst, at GIA in Carlsbad,
California.
ACKNOWLEDGMENTS
The authors wish to thank DTC Botswana for kindly allowing GIA
to interview staff and take photos on the diamond floor. Warm
thanks to Alan Bronstein for his encouragement, insights, and introductions. Thanks also to Marc van den Brande for his invaluable assistance, Drs. Roman Grynberg and Bram Janse for their
insights, and the executives of Botswanas diamond-manufacturing facilities who opened the doors to their operations to the authors. Claudia DAndrea, director of GIAs Bangkok campus, and
Anthony Licht of GIAs Johannesburg lab provided firsthand
112
knowledge of the Botswana diamond industry. Many thanks to

Kevin Schumacher, GIAs digital resources specialist, who provided video expertise during the interviews in Botswana.
We dedicate this article to our colleague and friend, Tawfic Farah,
GIAs vice president of international operations until his untimely
death in 2013. Tawfic had a passion for Botswana, and he nurtured a strong desire to help the people of Africa by providing educational opportunities. We need to find a way to be on the right
side of history, he would often advise us. Tawfic wrote a scholarly
paper on Botswana that was not published but served as the
seed for this story.
GEMS & GEMOLOGY
SUMMER 2014
REFERENCES
Alfaro L., Spar D., Allibhoy F. (2005) Botswana: A diamond in the
rough. HBS case no. 9-703-027. Harvard Business School Publishing, Boston, http://altongp.com/currentprojects/Zimbabwe%20
Diamond/Botswana%20Diamonds.pdf.
Anglo American (2009) Annual report 2009. http://ar09.angloamerican.solutions.investis.com/financial_statements/note
s/note38.html.
Argyle Diamond Mine (2013) The future: Argyle as an underground operation. http://www.argylediamonds.com.au/index_
operations.html.
Ashok T. (2013) Indian diamond industry losing sparkle. The
Hindu Business Line, Jan. 29, http://www.thehindubusinessline.com/opinion/indian-diamond-industry-losing-sparkle/
article4357705.ece.
Bates R. (2013) What sightholders really think about De Beers move
to Botswana. JCK, Oct. 15, http://www.jckonline.com/blogs/
cutting-remarks/2013/10/15/what-sightholders-really-thinkabout-de-beers-move-to-botswana.
Botswana looking to raise rent on Jwaneng mining lease. (2004) Idex,
No. 174, pp. 7576, http://www.idexonline.com/portal_FullMazalUbracha.asp?id=23157.
Botswana outraged over boycott of diamonds, tourism (2011)
Bangkok Gems & Jewellery, Vol. 24, No. 6, p. 47.
Danylchuk J. (2013) Yellowknife diamond row factory deal still
not iced. Northern Journal, Feb. 26, http://norj.ca/2013/02/yellowknife-diamond-row-factory-deal-still-not-iced.
De Beers (2008) Largest, most sophisticated sorting facility in the
world begins operations in Botswana. March 18, http://www.debeersgroup.com/ImageVaultFiles/id_686/cf_5/8-03.PDF.
(2013a) De Beers holds inaugural international sight in
Botswana. Nov. 11, http://www.debeersgroup.com/ImageVaultFiles/id_2126/cf_5/De_Beers_holds_inaugural_sight_in_
Gaborone.PDF.
(2013b) Operating and financial review 2012. http://www.
debeersgroup.com/ImageVaultFiles/id_2064/cf_5/2012_OFR.PDF.
(2014) Debswana. http://www.debeersgroup.com/
operations/mining/mining-operations/debswana.
Diamonds sparkle (2011) African Mining, Vol. 16, No. 6, November-December, p. 66.
Even-Zohar C. (2007) From Mine to Mistress: Corporate Strategies
and Government Policies in the International Diamond Industry.
Mining Communications Ltd., London.
Farah T. (2013) The politics of development in Africa: Botswana as
a case study. Working paper, GIA Library, Carlsbad, California.
Ferreira-Marques C. (2013) Goodbye London, hello Gaborone: De
Beers sales head to Africa. Reuters, Oct. 3, http://www.reuters.com/
article/2013/10/03/us-debeers-idUSBRE99204C20131003.
Golan E. (2007) DTC sightholders list: Whos on and what now?
Idex Online, Dec. 20, http://www.idexonline.com/portal_FullNews.asp?TextSearch=&KeyMatch=0&id=29129.
(2013) Lucara recovers 4.77 ct blue diamond. Idex, Vol. 28,
No. 274, p. 125, http://www.idexonline.com/portal_FullNews.asp?id=37641.
Greve N. (2013) De Beers 2012 gem production dips to 27.9m, sales
down 16%. Mining Weekly, Feb. 15, http://www.miningweekly.com/article/de-beers-2012-gem-production-dips-to-
279m-2013-02-15.
Grigorian K. (2012) Diamond production and processing: What
Armenia can learn from an intra-regional exchange on the
diamond trade. World Bank, http://siteresources.worldbank.org/
INTECALEA/Resources/1040211-1251905880790/KB47_
ArmeniaDiamond.pdf.
Grynberg R. (2013) Some like them rough. Botswana Institute for
Development Policy. Discussion Paper 142.
Hazelton R. (2002) Diamonds: Forever or for good? The Diamonds
and Human Security Project, Partnership Africa Canada, Occasional Paper #3, March, http://pacweb.org/Documents/
diamonds_KP/3_diamonds_Forever_Eng_March2002.pdf.
Janse A.J.A. (1996) History of diamond sources in Africa, Part II.
G&G, Vol. 32, No. 1, pp. 230, http://www.gia.edu/gems-gemology/spring-1996-diamond-history-africa-janse.
(2007) Global rough diamond production since 1870. G&G,
Vol. 43, No. 2, p. 98119, http://www.gia.edu/gemsgemology/summer-2007-global-rough-diamond-productionjanse.
Krawitz A. (2008) De Beers postpones setting up aggregation in
Botswana. Diamonds.net, Dec. 9, http://www.diamonds.net/
news/NewsItem.aspx?ArticleID=24441.
Lucara completes first sale of Botswana diamonds (2012) Solitaire,
No. 61, September, p. 28.
Mbayi L. (2011) Linkages in Botswanas diamond cutting and polishing industry. MMCP Discussion Paper No. 6, http://www.prism.
uct.ac.za/Papers/MMCP%20Paper%206_0.pdf.
Mellier P. (2014) We cannot lose consumer trust in diamonds. Professional Jeweller, Jan. 28, http://www.professionaljeweller.
com/article-14143-we-cannot-lose-consumer-trust-in-diamonds/.
Mokone O., Mogapi S. (2013) The treacherous home coming journey
of Botswana diamonds. Sunday Standard Online, Nov. 25,
http://www.sundaystandard.info/article.php?NewsID=18502&
GroupID=4.
Sanchez L. (2006) Botswana: A rags to riches story. Jewellery News
Asia, No. 256, January, pp 141152.
Shor R. (1990) Prosperity & politics. JCK, Vol. 161, No. 6, June, pp.
5672.
(2013a) Botswana: Making beneficiation work. GIA Research & News, June 24, http://www.gia.edu/research-newsbotswana-making-beneficiation-work.
(2013b) Industry Analysis: Demand improves after slow autumn, but challenges remain. GIA Insider, http://www.gia.edu/
gia-news-research-march-2013-demand-improves-slow-autumn.
Spektorov Y., Linde O., Cornelissen B., Khomenko R. (2013) The
global diamond report 2013: Journey through the value chain.
Bain & Company Insights, Aug. 27, http://www.bain.com/publications/articles/global-diamond-report-2013.aspx.
Transparency International (2013) Botswana. http://www.transparency.org/country/#BWA
Weldon R. (2001) Diamonds for good. Professional Jeweler, January,
http://www.professionaljeweler.com/archives/articles/2001/
jan01/0101dn1.html.
Yellowknife diamond-cutting plants future in limbo (2009) CBC News,
June 11, http://www.cbc.ca/news/canada/north/yellowknife-diamondcutting-plant-s-future-in-limbo-1.800202.
GEMS & GEMOLOGY
SUMMER 2014
113
FEATURE ARTICLES
THREE-PHASE INCLUSIONS IN EMERALD AND

THEIR IMPACT ON ORIGIN DETERMINATION
Sudarat Saeseaw, Vincent Pardieu, and Supharart Sangsawong
Until now, the observation of three-phase inclusions in emeralds has been considered a potential indicator of Colombian origin. Nevertheless, emeralds from Afghanistan (Panjshir Valley), China (Davdar),
and Zambia (Kafubu and a new deposit at Musakashi) may contain three-phase inclusions resembling
those often found in specimens from Colombian deposits (Muzo, Chivor, La Pita, Coscuez, and Peas
Blancas). This article presents detailed photomicrographs of samples from these localities, with a focus
on their multiphase inclusions. Also investigated are spectroscopic features and trace-element data. For
the gemologist studying the origin of emeralds hosting three-phase inclusions, a powerful set of tools to
aid in this determination comes from the combination of detailed microscopic examination, UV-VisNIR and FTIR spectroscopy, and LA-ICP-MS trace-element analysis.
he traditional practice of assigning Colombian

origin to emeralds exhibiting three-phase inclusions (Giuliani et al., 1993) is in need of updating, as we now know that such inclusions are often
seen in emeralds from other locations, including
China, Zambia (Schwarz and Pardieu, 2009), and
Afghanistan (Hammarstrom, 1989; Bowersox et al.,
1991). The characterization of these inclusions using
spectroscopy and chemical composition has not been
documented until now.
A new Zambian deposit at Musakashi was first
reported in 2005 (Zwaan et al., 2005), and field trip
reports were subsequently published (Klemm, 2009
and 2010). These articles presented historical and geological information about the mine (see box A), examined the gemological properties of these new
emeralds, and presented chemical analysis using
electron microprobe (Zwaan et al., 2005).
The present study characterizes emeralds containing multiphase inclusions from all of the localities above by using microscopy, spectroscopy, and
trace-element analysis to create a framework for geographic origin determination.

114
EMERALD ORIGIN DETERMINATION
MATERIALS AND METHODS

A total of 84 emeralds were studied, gathered from
the GIA Field Gemology Collection and from gem
merchants of Zambian and Colombian emeralds.
Color ranged from light to deep green. The samples
were studied in three forms: 55 were rough samples
with one or two polished surface windows, 28 were
fabricated as optical wafers oriented either perpendicular or parallel to the crystals c-axis, and one was
faceted. After fabrication, the samples ranged from
0.18 to 4.85 ct. The group consisted of:
Eleven samples reportedly from Musakashi (figure 1), acquired from gem dealer Guy Clutterbuck, who visited the mines several times.
Weight range: 0.574.85 ct.
Ten samples collected from the Kagem mine in
the Kafubu area of Zambia by author VP in the
summer of 2013. Weight range: 0.292.22 ct.
Fifteen samples collected from Afghanistans
Panjshir Valley by author VP in 2011. Weight
range: 0.181.80 ct.
Thirteen samples collected from Davdar,
China, by author VP in 2005 and 2009. Weight
range: 0.161.77 ct.
Thirty-five Colombian samples from Muzo
(12), Coscuez (12), La Pita (3), Peas Blancas (2),
and Chivor (6), acquired from Flavie Isatelle, a
GEMS & GEMOLOGY
SUMMER 2014
Figure 1. These emeralds from Musakashi, Zambia, range from 0.57 to 4.85 ct. The rough samples were
polished with one window on each. The faceted sample measures 6.29 5.35 3.53 mm and weighs 0.97 ct.
Photo by N. Kitdee.
field gemologist/geologist who visited in May

2010, and gem merchant Ron Ringsrud. Weight
range: 0.124.14 ct.
Standard gemological properties were measured
with a Rayner refractometer (yttrium-aluminum-garnet prism) equipped with a near sodiumequivalent
light source to measure refractive index and birefringence, a Chelsea filter, and four-watt long-wave (365
nm) and short-wave (254 nm) UV lamps to observe
fluorescence. Due to fracturing in the samples, coupled with their relatively small sizes, no useful specific
gravity measurement was acquired. The wafers optical path lengths were measured using a Mitutoyo Series 395 spherical micrometer with an accuracy of 2
microns. Microscopic examination was performed
with GIA binocular microscopes at 1070 magnification, using both darkfield and brightfield illumination. Other lighting techniques, including fiber-optic
illumination, were employed to investigate internal
characteristics. Photomicrographs were captured at up
to 180 magnification with a Nikon SMZ 1500 sys-
tem using darkfield, brightfield, and oblique illumination with a fiber-optic light.
Ultraviolet through visible and near-infrared (UVVis-NIR) spectra were collected with a Hitachi U2900 spectrophotometer (for polarized ordinary ray
spectra) at 1 nm resolution and a PerkinElmer
Lambda 950 spectrophotometer at 0.3 nm resolution,
operating with a 60 mm integrating sphere accessory
and a Lambda polarizer accessory. Fourier-transform
infrared (FTIR) spectroscopy was performed using a
Thermo Nicolet 6700 FTIR spectrometer operating
with a 4 beam condenser accessory at 4 cm1 resolution. A Renishaw inVia Raman microscope fitted
with a 514 nm argon ion laser was used for identifying
inclusions.
For laser ablationinductively coupled plasma
mass spectrometry (LA-ICP-MS) chemical analysis,
we used a Thermo Scientific X Series II ICP-MS combined with a Nd:YAG-based laser ablation device operating at a wavelength of 213 nm. For the ICP-MS
operations, the forward power was set at 1300 W and
the typical nebulizer gas flow was approximately
GEMS & GEMOLOGY
SUMMER 2014
115
BOX A: EMERALDS FROM MUSAKASHI, ZAMBIA

Little information is available about the Musakashi deposit (figure A-1) reportedly discovered in 2002, and much
of it is conflicting. It appears to have originally been
worked by local miners from 2002 to 2004. Emeralds
from Musakashi were first reported in 2005 by Antonin
Seifert, Stanislav Vrna, Bjrn Anckar, and Jaroslav Hyrl
(Zwaan et al., 2005). Seifert and his colleagues reported
that these emeralds had a significantly different color
from those found in the Kafubu area. The Musakashi
stones were said to show an intense bluish green reminiscent of emeralds from Muzo, Colombia. The similarity did not end there:
Production and Distribution

According to the mine director, the area produced about
1520 kg of emeralds from its discovery through late
2010 (Klemm, 2010). Gem merchant Guy Clutterbuck,
who supplied some of the samples for this study, reported to one of the authors in 2013 and 2014 that limited emerald mining was still taking place near
Musakashi.
AFRICA
The most interesting characteristic was the presence in

all the stones of three-phase inclusions, consisting of a
bubble and a cube-shaped crystal in a liquid, almost identical to those commonly seen in Colombian emeralds.
Zwaan et al. also identified sphene, iron oxides, feldspar,

and quartz as mineral inclusions.
Production reportedly resumed in 2006, starting with
hand tools and progressing to the use of an excavator for
a few months in 2009 (Klemm, 2009 and 2010). These
reports, published after visits to Musakashi in July 2009
and August 2010, provide details adding to the history
of emerald mining there.
D E M. R E P.
OF THE CONGO
ZAMBIA
LUAPULA
N
Musakashi
Area
NORTH WESTERN
Kafubu
Area
COPPERBELT
ZAMBIA
EASTERN
CENTRAL
LUSAKA
MOZAMBIQUE
Lusaka
WESTERN
SOUTHERN
ZIMBABWE
Location and Access

During the dry season, the Musakashi area is accessible
by a five-hour, 165 km (102 mile) drive from Ndola,
mainly on the poorly maintained Kasempa road (Klemm,
2009 and 2010).
NORTHERN
150 km
Figure A-1. This map shows the location of the Kafubu and the Musakashi emerald deposits.
tribution, though color-zoned areas were also sampled (figure 2).
0.90 L/min. The carrier gas used in the laser ablation

unit was He, set at approximately 0.78 L/min. The
alignment and tuning sequences were set to maximize Be counts and keep the ThO/Th ratio below
2%. Laser ablation conditions consisted of a 40 m
diameter laser spot size, a fluence of 10 1 J/cm2, and
a 7 Hz repetition rate. For quantitative analysis, samples must be calibrated against an external standard
of known composition, which meant measuring the
signals for the elements of interest in the sample and
comparing them to the signals for a standard with
known concentrations of those elements. Generally,
NIST 610 and 612 glasses were used for calibration
standards. All elemental concentrations were calculated by applying 29Si as an internal standard, with Si
concentration calculated from the theoretical value
of beryl (31.35 wt.%). Laser spots were applied in the
same area where UV-Vis spectra were collected,
which was usually clean and had an even color dis-
Gemological Properties. The samples gemological

properties are summarized in table 1. The refractive
indices varied from 1.570 to 1.588 for n and 1.576 to
1.593 for n, with birefringence between 0.006 and
0.010. Emeralds from Musakashi, Davdar, and all the
Colombian deposits had lower RI than the samples
from Kafubu and Panjshir. The Musakashi emeralds
displayed a particularly strong pink reaction in the
Chelsea color filter, while a pink to strong pink display was typical in samples from all other localities
except for Kafubu, in which there was no reaction.
116
GEMS & GEMOLOGY
RESULTS
Microscopic Characteristics. Musakashi, Zambia. In

the 11 samples reportedly from Musakashi, the most
common inclusions were three-phase, usually con-
SUMMER 2014
6
7
8
9
10
1
3
5
2
4
Figure 2. This optical wafer from a Muzo emerald

provides an example of the LA-ICP-MS sample
points selected in this study. Heavily included areas
were avoided in favor of well-crystallized zones in
the sample that best represent the chemical composition of that emerald. Photo by S. Engniwat.
taining at least two colorless transparent crystals and

a rather small gas bubble suspended in a fluid (figures
3 and 4). Such inclusions were found in every sample
we studied. The gas bubble always measured within
the size range of the two crystals in the multiphase
inclusion. In those inclusions containing two crystals, both crystals appeared to be singly refractive
under cross-polarized illumination. The larger crystal
was essentially cubic and bright, while the smaller
one was rounded and faint (figure 3, A and B). In many
samples, the multiphase inclusions also hosted one
or more small dark to black crystals, and occasionally
tiny colorless or whitish crystals (figures 3 and 4).
Some of these inclusions were identified by Raman
spectroscopy (figure 4); the gas phase was identified
as CO2 (peaks at 1284 and 1387 cm1). The singly refractive cubic crystals displayed only the emerald
spectrum and no Raman lines, suggesting they could
be halides (Venkateswaran, 1935). Daughter crystals
were identified as calcite (peaks at 283, 713, and 1085

cm1). An additional peak at 2328 cm1 was assigned
to N2. These results match those for fluid inclusions
found in Colombian emeralds (Giuliani et al., 1994),
but SEM is needed to identify the different complex
mixed salts and daughter minerals. The shape of the
inclusions was usually irregular, with numerous
branches. In rare cases they appeared needle-like (figure 3C) or jagged (figure 3D).
Under darkfield illumination, tube-like growth
features (figure 5) were found in 10 of the 11
Musakashi samples. The tubes were oriented parallel
to the prism faces of the crystal, a common feature
in emeralds. Each sample had an evenly distributed
green color except for one that displayed color banding ranging from strong green to colorless perpendicular to the c-axis and the growth tubes.
Solid inclusions were quite rare. Of the 11 emeralds, only four hosted mineral inclusions that were not
part of multiphase inclusions. The most common
In Brief
Three-phase inclusions in emerald, once considered a
reliable indicator of Colombian origin, have also been
observed in specimens from Zambia, Afghanistan, and
China.
Microscopy, spectroscopy, and trace-elemental analysis
were used to characterize 84 emeralds with multiphase
inclusions.
A promising basis for determining the geographic origin of emeralds comes from a combination of these
methods.
crystal inclusions identified using Raman spectroscopy were colorless and transparent euhedral calcite crystals (figure 6, left). Also found were several
opaque and metallic euhedral inclusions resembling
TABLE 1. Gemological properties of emeralds from various sources.

Property
Musakashi, Zambia
Kafubu, Zambia
All Colombia
Davdar, China
Panjshir, Afghanistan
Refractive indices
n
1.5721.578
1.5821.588
1.5701.573
1.5771.580
1.5721.580
1.5801.582
1.5901.593
1.5761.580
1.5831.588
1.5801.590
Birefringence
0.0060.008
0.0050.008
0.0060.008
0.0050.008
0.0070.010
Chelsea filter
Strong pink
Inert
Pink to strong pink
Pink
Inert to pink
GEMS & GEMOLOGY
SUMMER 2014
117
Figure 3. Three-phase inclusions in Musakashi emeralds. A: The larger multiphase inclusion displays
several crystals and a small
gas bubble. B: Along with
the gas bubble, at least
three colorless crystals and
one or two tiny dark crystals are visible. C: Two colorless transparent crystals
and a small gas bubble in
elongated multiphase inclusions. D: Two colorless
transparent crystals with a
small gas bubble and a tiny
black crystal in irregularly
shaped multiphase inclusions. Photomicrographs by
S. Saeseaw, brightfield illumination. Image widths 9
mm (A) and 7 mm (BD).
iron oxide, possibly hematite or the martite variety of

hematite (figure 6, center), and some whitish translucent crystals identified as cryolite (figure 6, right). The
iron oxide minerals identified usually occur in granitic
pegmatites, but here they appeared in an emerald that
also hosted pyrite and calcite inclusions.
Kafubu, Zambia. Most of the inclusions in the Kafubu samples were multiphase and rectangular in
shape (figure 7), but some were quite irregular in outline (figures 8 and 9). These inclusions typically
hosted some liquid and a gas bubble, but in several
cases it was possible to see a solid third phase. Usu-
RAMAN SPECTRA
Figure 4. This
Musakashi emerald
clearly displays a gas
bubble, a cubic crystal,
two rounded colorless
crystals, a tiny black
crystal, and two tiny
whitish rhombohedra.
Raman spectroscopy
was used to identify
the host emerald
(green), the CO2 gas
bubble (purple), the
square halide crystal
(orange), and the
smaller carbonate crystal (red). Photomicrograph by S. Saeseaw;
brightfield illumination.
1
1085
1387
2
4
283
1726
4) calcite
3) halide
2328
2) gas phase (CO2)
1) fluid phase
emerald
0
500
1000
1500
2000
2500
RAMAN SHIFT (cm-1)
118
GEMS & GEMOLOGY
SUMMER 2014
Figure 5. Growth tubes parallel to the prism faces of

the crystal are visible in this Musakashi emerald.
Photomicrograph by S. Saeseaw, brightfield illumination. Image width 27 mm.
ally the associated crystal had very low relief. When

the inclusions were studied between crossed polarizers, the crystal showed evidence of being doubly refractive (figure 9). This was very different from our
observations of the Musakashi emeralds, which had
at least two singly refractive crystals (halides) trapped
within the multiphase inclusions.
Another interesting observation of the Kafubu
emeralds was that when the inclusion and the gas
bubble were flat, we could estimate the size ratio between them. In these samples, the gas bubble appeared
to contain one-third to one-half the volume of the
multiphase inclusion at room temperature. This was
much larger than the gas bubble observed in emeralds
from Musakashi, which indicates considerably different temperature and pressure conditions of formation
and possibly very different fluid composition. These
observations suggest a different geologic background.
Figure 7. Top: Rectangular multiphase inclusions were

visible in emeralds from Kafubu. Bottom: Detail of
rectangular multiphase inclusions in an emerald from
the Kagem mine reveals relatively large gas bubbles,
with and without solid inclusions. Photomicrographs
by V. Pardieu, brightfield illumination. Image widths
27 mm (top) and 11 mm (bottom).
Solid inclusions with no associated liquid or gas

inclusions were quite common in all the samples
from Kafubu. One of the authors (VP) visited Kagem
Figure 6. Single-phase solid inclusions found in the emeralds from Musakashi vary significantly and consist of: a colorless transparent crystal identified by Raman spectroscopy as calcite (left); a metallic opaque crystal, possibly hematite
or martite (center); and whitish translucent crystals identified using Raman spectroscopy as cryolite (right). Photomicrographs by S. Saeseaw, darkfield illumination. Image widths 13 mm (left), 11 mm (center), and 10 mm (right).
GEMS & GEMOLOGY
SUMMER 2014
119
Figure 8. Irregularly shaped multiphase inclusions in emeralds from Kafubu showed liquid containing a gas bubble and in some cases a solid third phase. Photomicrographs by V. Pardieu, brightfield illumination. Image widths
13 mm (left) and 10 mm (right).
in 2011 and 2012 and found that the emeralds there

were commonly associated with quartz, pyrite, dark
brown to black tourmaline, mica, and chlorite (figure
10). The most common crystal inclusions found had
dark, opaque, and dendritic features, identified by
Raman spectroscopy as pyrolusite, amphibole, and
tourmaline (figure 11), as well as chlorite and mica.
These inclusions offer telltale features that reflect
the mineral associations found at the site.
Panjshir Valley, Afghanistan. Most of the inclusions
in emeralds from the Panjshir Valley are multiphase.
These typically have an elongated needle-like shape
(figures 12 and 13), but they may also display an irregular (figure 14) or even jagged aspect (figure 12).
These inclusions often host several cubic to rounded
Figure 9. Irregular multiphase inclusions in emeralds

from Kafubu. Under cross-polarized illumination, a
small doubly refractive crystal inclusion becomes visible as a bright spot (see red arrow). Photomicrograph by
V. Pardieu; image width 13 mm.
transparent crystals, and sometimes small, dark

opaque crystals. Examined under crossed polarizers,
the cubic crystals appeared to be singly refractive and
the small rounded crystals appeared doubly refractive
(again, see figure 12). These observations were identical to those previously reported (Bowersox et al.,
1991). The gas bubble appeared to be smaller than the
main crystal inclusion (which was usually cubic).
Davdar, China. Most of the inclusions in the samples
from Davdar were multiphase. These were often jagged
or irregular in shape (figures 1517) and occasionally
needle-like (figure 16). Some fluid inclusions contained
two halide cubes (figure 16), as reported previously
Figure 10. An interesting scene observed at the Kagem
mine: an emerald crystal associated with black tourmaline, quartz, mica, and pyrite. It was not surprising
to see this mineral association at the site, as it can
also be seen inside the gem. This offers a fine example
of the direct association between the geological background and the internal world of a gemstone. Photo
by V. Pardieu.
Mica
Emerald
Quartz
Tourmaline
Pyrite
120
GEMS & GEMOLOGY
SUMMER 2014
Figure 11. Various single-phase solid inclusions were observed in emeralds from the Kagem mine. Left: A dendritic black opaque mineral inclusion (probably pyrolusite). Center: A dark brownish transparent mineral inclusion identified by Raman spectroscopy as amphibole. Right: A dark brownish mineral inclusion identified as
tourmaline with Raman. Photomicrographs by V. Pardieu, brightfield illumination. Image widths 20 mm (left)
and 40 mm (center and right).
(Marshall et al., 2012). In these emeralds, the gas bubble was smaller than the whole inclusion and usually
somewhat smaller than the associated cubic crystals.
Colombia. The most common inclusion in the
Colombian emeralds was a jagged multiphase inclusion hosting a gas bubble and one or more cubic crystals (figures 1820). In these stones, the gas bubble was
usually smaller than the whole inclusion and no larger
than the associated cubic crystal. Many of these emeralds also contained a tiny dark opaque crystal (figures
18 and 20, right) and clusters of daughter crystals, usually irregularly shaped (figures 19, bottom, and 20, left).
These daughter crystals are related to various carbonate compounds (Giuliani et al., 1994). Most of the multiphase inclusions had the classic jagged shape
associated with Colombian emeralds. In several cases
the shape was more elongated, like a blade (figure 20,
center), and occasionally irregular (figure 20, right).
UV-Vis-NIR Spectroscopy. The samples collected for
UV-Vis-NIR spectroscopy were fabricated as optical
wafers oriented either perpendicular or parallel to the
crystals c-axiswith the exception of emeralds from
Musakashi, which were either faceted or contained
a single polished window. In this study, UV-Vis-NIR
Figure 12. These Panjshir

emeralds contain a needle-like
multiphase inclusion hosting
crystals and gas bubbles (left),
and jagged, irregularly shaped
multiphase inclusions hosting
crystals and gas bubbles
(right). Photomicrographs by
V. Pardieu, brightfield illumination. Image widths 8 mm
(left) and 13 mm (right).
Figure 13. Elongated needlelike multiphase inclusions,
hosting several crystals and
gas bubbles, in emeralds
from the Kamar Safeed area
near Khenj in the Panjshir
Valley. Photomicrographs by
S. Saeseaw, brightfield illumination (left) and crossed
polarizers (right). Image
widths 9 mm.
GEMS & GEMOLOGY
SUMMER 2014
121
Figure 14. Left: This emerald from the Kamar Safeed area in the Panjshir Valley of Afghanistan contains an irregular blocky multiphase inclusion hosting gas bubbles, a liquid phase, and several crystals. Center: Irregularly
shaped multiphase inclusion hosting a gas bubble and several crystals, from the Koskanda area near Khenj.
Right: Irregularly shaped multiphase inclusion hosting a small gas bubble and several crystals (one cubic and
one more rounded), from Kamar Safeed. Photomicrographs by V. Pardieu, brightfield illumination. Image widths
9 mm (left and center) and 7 mm (right).
spectra characterized the following chromophores for

emerald (Wood and Nassau, 1968):
Cation
Identifying absorption
characteristic
Possible geographic
origin
Cr3+
Bands at approximately 430 and Musakashi, Kafubu,

600 nm (), and at approximately Davdar, Panjshir,
420 and 630 nm ()
Colombia
Lines at 476, 680, and 683 nm
Fe2+
A band at 620 nm (), and a

broad band at approximately
810 nm (, )
Strong: Kafubu
Moderate: Davdar,
Panjshir
Fe3+
Narrow bands at 372 () and

465 nm (, ), and a band at
approximately 400 nm (, )
Kafubu, Panjshir
V3+
Bands at approximately 400 nm

() and 654 nm ()
Strong: Davdar,
Colombia
Zambia. In the 11 unoriented samples from Musakashi, the UV-Vis spectra displayed minimum absorptions at about 348 and 510 nm for the ordinary
ray. For octahedral Al3+ in emeralds, substituted Cr3+
exhibited bands at 430 and 600 nm for the ordinary
ray and lines at 476, 680, and 683 nm. These emeralds showed no significant Fe2+-related absorption
features in the NIR region, and no Fe3+ was observed
(figure 21A).
The nine fabricated samples from Kafubu all displayed the same absorptions: UV-Vis minima at 367
and 514 nm for the ordinary ray, and at 390 and 500
nm for the extraordinary ray. Moreover, a narrow Fe3+
band was observed at 372 nm in the ordinary ray but
not the extraordinary ray, and a strong Fe2+ band at
around at 810 nm was recorded (figure 21B).
Figure 15. Jagged multiphase inclusions hosting

gas bubbles and several
crystals in emeralds
from Davdar, China.
Photomicrographs by S.
Saeseaw, brightfield illumination. Image
widths 7 mm.
Figure 16. Irregular and
needle-shaped multiphase
inclusions found in emeralds from Davdar, hosting
single gas bubbles, multiple cubic crystals, rounded
transparent crystals, and
tiny black opaque solids.
Photomicrographs by S.
Saeseaw, brightfield illumination. Image widths 8
mm (left) and 7 mm (right).
122
GEMS & GEMOLOGY
SUMMER 2014
Figure 17. Irregular multiphase inclusions with a gas

bubble and a cubic crystal in an emerald from Davdar.
Photomicrograph by S. Saeseaw, brightfield illumination. Image width 8 mm.
Afghanistan. Nine fabricated samples were studied.

Their UV-Vis spectra showed the Fe3+ peak at 372 nm
for the ordinary ray and moderate Fe2+ absorption at
810 nm, similar to Kafubu emeralds. Yet some of the
Panjshir samples appeared to contain no Fe3+, due to
a low iron concentration, and they also exhibited
peak heights at 810 nm that were lower than in the
Kafubu samples (figure 21C).
China. Ten samples were measured, but only three
were fabricated perpendicular or parallel to the c-axis.
A representative UV-Vis spectrum is displayed in figure 21D. Davdar and Musakashi emeralds presented
similar spectra, but with higher Fe2+ absorption in the
NIR region of the former. But no Fe3+ features were
observed, and strong V3+ absorptions were detected.
Colombia. The eight fabricated samples from Chivor,
Coscuez, and Muzo exhibited UV-Vis spectra similar
to those of Musakashi emeralds. Neither Fe2+ nor Fe3+
Figure 19. Examples of multiphase inclusions in emeralds

from Colombia. Top: This multiphase inclusion hosting a
gas bubble and a large colorless cubic crystal with barely
visible edges, which appears larger than the bubble, is
seen in an emerald from the La Pita area. Bottom: Another example of jagged multiphase inclusions hosting a
gas bubble, a large colorless cubic crystal (larger than the
bubble), and possibly a smaller cluster of daughter crystals. This sample was from the Coscuez area. Photomicrographs by S. Saeseaw, brightfield illumination. Image
widths 9 mm (top) and 10 mm (bottom).
were observed. Absorptions corresponding to V3+ were

present at about 400 and 654 nm (figure 21E).
FTIR Spectroscopy. FTIR spectra were recorded in
the 4007800 cm1 region, with the major area of
Figure 18. Two classic

jagged multiphase inclusions in emeralds
from Muzo, Colombia,
hosting a gas bubble
with two cubic crystals
and a tiny dark opaque
crystal. Photomicrographs by S. Saeseaw,
brightfield illumination.
Image widths 9 mm
(left) and 7 mm (right).
GEMS & GEMOLOGY
SUMMER 2014
123
Figure 20. Left: In Colombian emeralds such as this one from Muzo, liquid was often associated with a gas bubble
as well as a large colorless cubic crystal. In this sample, some smaller colorless and dark crystals are also visible.
Center: Elongated blade-like multiphase inclusions hosting a gas bubble and a large colorless cubic crystal
(slightly larger than the bubble) in an emerald from Coscuez. Right: In some cases, the multiphase inclusions were
not jagged or elongated blades but irregularly shaped. Here the multiphase inclusions from a Coscuez emerald
host a gas bubble, a large colorless cubic crystal (about the same size as the bubble), and some smaller crystals.
Photomicrographs by S. Saeseaw, brightfield illumination. Image widths 13 mm (left) and 8 mm (center and right).
interest between 4500 and 7500 cm1specifically,

the area related to water molecule vibrations. The
resulting spectra indicated that the emeralds could
be divided according to type I and II water molecule
assignments (Wood and Nassau, 1968; Schwarz and
Henn, 1992). Type I water molecules are oriented in
the channels, not linking to other alkali metal ions,
and their H-H direction is parallel to the c-axis. Type
II molecules are located adjacent to alkali metal ions
in the channel, and their H-H direction is perpendicular to the c-axis.
All of the emeralds contained type I water molecules except the samples from Kafubu (figure 22).
The extraordinary ray plots of type I emeralds each
exhibited a 7140 cm1 peak that was more intense
than the peaks found at 7095 and 7072 cm1. The
plots from type II emeralds displayed their highestintensity peak at 7095 cm1. These results matched
previous reports (Wood and Nassau, 1968; Schwarz
and Henn, 1992) that assigned the 7140 cm1 peak to
type I water molecules and the 7095 cm1 peak to
type II water molecules.
Trace-Element Analysis. All 84 emeralds from the
five different localities were analyzed using LA-ICPMS (table 2).
content varied from 680 to 1,490 ppmw. Trace

amounts of Ni were also present.
The samples from Kafubu were rich in the trace
elements Li, Na, K, Rb, Cs, Mg, and Fe, as well as Ti,
Sc, Mn, Ni, and Zn. Total alkali ion concentration
was as high as 34,747 ppmw; Mg and Fe contents averaged 15,004 and 8,621 ppmw, respectively. The
chromophore elements Cr and V varied from 733 to
4,330 ppmw for Cr and from 71 to 180 ppmw for V;
the Cr/V ratio was between 8 and 40. Small amounts
of Sc were detected, from 12 to 75 ppmw. Significant
traces of Mn, Ni, and Zn were also present.
Afghanistan. Afghan emeralds from Panjshir contained on average 10,780 ppmw of alkali ions. The
chromophore elements Cr and V ranged from 118 to
4,730 ppmw and from 255 to 3,680 ppmw, respectively. The Cr/V ratio was 0.3 to 3.3. Iron contents
ranged from 1,010 to 9,820 ppmw. Scandium contents were detected up to 2,290 ppmw, the highest
concentration for this element in our study.
China. The emeralds from Davdar contained relatively lower amounts of combined alkali ions, 8,835
ppmw on average. Their green color was caused by
the presence of both Cr and V, which varied from 146
to 5,630 ppmw for Cr and from 657 to 6,960 ppmw
for V. The ratio for Cr/V was 0.1 to 1.0. Iron concentrations ranged up to 4,350 ppmw.
Zambia. The Musakashi emeralds shared similar

trace-element chemistry with the Colombian samples. They contained low concentrations of alkali
metals (Li, Na, K, Rb, and Cs); Mg, and Fe, as well as
minor traces of Ni. Total alkali ion concentrations
averaged 4250 ppmw. In terms of chromophore elements, the concentrations of Cr were greater than V,
and the Cr/V ratio ranged between 1.7 and 5.3. Fe
Colombia. All the Colombian emeralds contained

low concentrations of alkali metals (Li, Na, K, Rb, and
Cs), as well as Mg and Fe. Total alkali ion concentrations averaged 4,725 ppmw. The chromophore elements Cr and V ranged from 172 to 10,700 ppmw for
124
GEMS & GEMOLOGY
SUMMER 2014
UV-VIS-NIR ABSORPTION SPECTRA

B
ABSORPTION COEFFICIENT (cm1)
Musakashi: unoriented sample

Cr 3+
10
Cr 3+
8
4
Cr 3+
2
0
250
300
350
400
450
500
550
600
650
700
750
800
850
12
12
Kafubu: E | | c
10
Fe 2+
Cr 3+
Cr 3+
4
Cr 3+
Fe 3+
2
0
300
250
350
400
WAVELENGTH (nm)
450
500
550
600
650
700
750
800
850
WAVELENGTH (nm)
10
4
Cr 3+
2
Cr 3+
Fe 3+
Cr
3+
Fe 2+
0
300
250
350
400
450
500
550
600
650
700
750
800
850
12
Panjshir: E | | c
Davdar: E | | c
10
8
Cr 3+
6
Cr 3+
4
V 3+
Cr 3+
0
250
300
350
400
WAVELENGTH (nm)
450
500
550
600
650
700
750
800
850
WAVELENGTH (nm)
E
12
Coscuez: E | | c
10
6
Cr 3+
4
Cr 3+
V 3+
Cr 3+
Figure 21. These UV-Vis-NIR absorption spectra

are representative of the five different emerald localities in this study. Solid line = ordinary ray ()
and dashed line = extraordinary ray (). Note: LAICP-MS analysis gave values corresponding to the
different localities:
(A) 5,526 ppm Cr, 1,312 ppm V, and 1,400 ppm Fe
(B) 3347 ppm Cr, 116 ppm V, and 9,443 ppm Fe
(C) 1,236 ppm Cr, 687 ppm V, and 6,250 ppm Fe
(D) 650 ppm Cr, 2,494 ppm V, and 2,188 ppm Fe
(E) 822 ppm Cr, 1882 ppm V, and 723 ppm Fe
0
250
300
350
400
450
500
550
600
650
700
750
800
850
WAVELENGTH (nm)
Cr and from 218 to 10,100 ppmw for V; the Cr/V ratio

was between 0.04 and 3.5. La Pita contained higher Cr
and V than the other Colombian mines. Overall, the
Colombian emeralds contained the lowest Fe concentrations in this study, from 117 to 2,030 ppmw.
DISCUSSION
GEMS & GEMOLOGY
Geology. Several summaries describing the geology of

major emerald deposits are available in the gemological and geological literature (Giard, 1998; Behmenburg, 2002; Groat et al., 2008), but very few have dealt
SUMMER 2014
125
TABLE 2. Chemical composition (average ppmw) of 84 emerald samples by LA-ICP-MS.a

Element
Musakashi,
Zambia
Kafubu,
Zambia
Panjshir,
Afghanistan
Davdar,
China
Coscuez,
Colombia
Muzo,
Colombia
11 samples,
55 spots
10 samples,
165 spots
15 samples,
205 spots
13 samples,
103 spots
12 samples,
100 spots
12 samples,
89 spots
Li
5597
(73)
360872
(576)
78254
(113)
72332
(109)
52139
(83)
2892
(45)
Be
40,70055,6000
(47,260)
40,60059,000
(49,630)
44,50057,900
(51,578)
45,80071,400
(54,515)
45,20057,100
(50,448)
43,40056,800
(51,053)
Na
1,1505,900
(4,109)
14,20020,100
(17,204)
2,47015,600
(9,822)
4,20012,500
(8,356)
2,5306,500
(4,885)
1,8909,860
(4,527)
Mg
9285,740
(3,652)
12,60017,400
(15,004)
1,79015,700
(9,197)
3,42011,700
(7,366)
2,1005,840
(4,176)
1,7206,590
(3,869)
75,6000108,000
(90,895)
64,10092,400
(77,865)
71,900110,000
(90,390)
81,200153,000
(101,220)
83,400112,000
(94,655)
81,300109,000
(97,888)
25102
(59)
305890
(506)
521,590
(670)
102609
(340)
646
(14)
749
(26)
Sc
40240
(119)
1275
(31)
492,290
(580)
54916
(372)
2258
(41)
34706
(224)
5882,100
(1,165)
71180
(109)
2553,680
(1,444)
6576,960
(2,867)
6566,210
(2,204)
6736,920
(2,062)
Cr
1,3406,170
(4,211)
7334,330
(2,287)
1184,730
(1,832)
1465,630
(1,333)
1726,330
(1,156)
2084,890
(1,620)
Fe
6801,490
(1,224)
5,90011,600
(8,621)
1,0109,820
(3,890)
1,2304,350
(2,440)
2851,130
(650)
1882,030
(588)
Ga
1329
(20)
919
(15)
1238
(24)
1043
(23)
1354
(28)
1744
(30)
15
(3)
17105
(65)
4110
(50)
329
(15)
0.73
(2)
0.85
(3)
133
Cs
310
(6)
5272,210
(1,391)
1197
(49)
641
(15)
719
(12)
419
(11)
47
Ti
bdl30
(6)
625
(12)
bdl67
(14)
bdl63
(7)
bdl27
(6)
bdl14
(4)
Ni
bdl12
(6)
738
(20)
bdl14
(2)
bdl
bdl3
(2)
bdl3
(1)
Zn
bdl2
(1)
1244
(29)
bdl5
(1)
bdl8
(2)
bdl2
(1)
bdl5
(1)
23
24
27
Al
39
45
51
53
57
69
85
Rb
60
66
Data reported in minimum and maximum values, with average concentration in parentheses; ppmw = parts per million by weight; bdl = below detection limit.
with the newer deposits such as Davdar or Musakashi.

The mention by Klemm (2010) of sandstone as a possible host rock for Musakashi emeralds is quite interesting, as calcareous sandstone was also identified as
the main host rock of the emeralds from Davdar (Pardieu and Soubiraa, 2006; Michelou and Pardieu, 2009;
Schwarz and Pardieu, 2009; Marshall et al., 2012). It is
Multiphase Inclusions. The discovery of three-phase

inclusions in samples reportedly from Musakashi
was rather surprising. While these inclusions are
similar to those in emeralds from Colombia, Afghan-
126
GEMS & GEMOLOGY
very intriguing to observe similar multiphase inclusions in the stones from these two deposits.
SUMMER 2014
Peas Blancas,
Colombia
La Pita,
Colombia
Chivor,
Colombia
Detection
limit
2 samples,
10 spots
3 samples,
15 spots
6 samples,
35 spots
3065
(48)
2480
(44)
40113
(67)
0.2
49,90054,700
(52,480)
52,90070,700
(58,666)
46,90056,000
(50,365)
2.3
2,3904,100
(3,226)
1,9807,000
(4,677)
1,3406,090
(3,448)
8.8
2,2803,740
(2,888)
1,8406,830
(4,475)
1,1005,280
(2,859)
0.4
96,600109,000
(102,750)
92,300129,000
(105,493)
84,600110,000
(97,188)
3.6
811
(10)
bdl40
(24)
bdl30
(14)
3.0
3682
(57)
94344
(248)
6173
(87)
0.8
3173,960
(2,000)
6,00010,100
(8,004)
2182,020
(846)
0.3
1,8102,540
(2,183)
2,94010,700
(5,645)
6712,880
(1,334)
2.5
164656
(400)
200383
(280)
117862
(366)
18.4
1230
(20)
3236
(34)
740
(19)
0.1
0.61
bdl4
(3)
0.53
(2)
0.1
310
512
616
(11)
0.1
47
(4)
bdl5
(4)
bdl7
(4)
2.1
bdl
bdl
bdl
0.5
bdl
bdl
bdl
0.4
phases. Usually at least two crystals could be seen:

one cubic, one more rounded, and often several
smaller ones ranging from transparent colorless to
opaque black. This was not typically observed in our
Colombian samples. The cubic crystals, the gas
phase, and the daughter crystals were identified by
Raman spectroscopy as a halide, CO2, and carbonate,
respectively. These results are identical to those reported for fluid inclusions in Colombian emerald
(Giuliani et al., 1994). The shape of the cavity housing the multiphase inclusions in our Musakashi samples was usually quite irregular, not unlike those
seen in those from Colombia, Afghanistan, and
China. In some cases the cavitys shape was similar
to the jagged form commonly found in Colombian
emeralds or the elongated needle shape typical of
Panjshir emeralds. But for the most part, the cavities
in the Musakashi samples were similar to those associated with Davdar.
Unlike the Kafubu emeralds, the multiphase inclusions in our Musakashi samples also contained
more crystals with single refraction than double refraction. The gas bubbles in Musakashi multiphase
inclusions were smaller than those found in Kafubu
material. These results reflect the possibility that different temperature and pressure conditions, geologic
backgrounds, and fluid compositions existed when
the emeralds from these two neighboring locations
formed.
istan, and China (Xin Jiang), specimens from the

neighboring Kafubu area of Zambia contain twophase inclusions or three-phase inclusions with crystals that are only visible under cross-polarized light
(Zwaan et al., 2005). The stones reportedly from
Musakashi had multiphase inclusions hosting not
only liquid and gas bubbles but also several solid
Spectroscopy. UV-Vis-NIR. The UV-Vis-NIR spectra

were characterized by absorptions that resulted from
the presence of the chromophores chromium, vanadium, and iron, illustrated in figure 21. The Musakashi
emeralds shared similar absorption characteristics with
those from Colombia and Davdar, but with lower V3+
absorption. Furthermore, the Musakashi samples
showed no significant Fe2+-related absorption features
at about 810 nm, and no Fe3+ absorption at about 372
nm, which distinguished them from Davdar and most
of the Kafubu material. Kafubu and Panjshir samples
showed a strong narrow absorption band at 372 nm corresponding to Fe3+, as well as a broad band around 810
nm corresponding to Fe2+. The Davdar emeralds displayed only the Fe2+-related broad band. The 810 nm
peak is related to Fe content, though their relationship
has not been clearly established yet. With higher iron
concentration, the peak intensity at 810 nm increased,
along with the peak at 372 nm. In this study, we measured the 810 nm peak height of the ordinary ray, which
ranged from 7 to 11 cm1 (Kafubu) and from 1 to 5 cm1
(Panjshir).
GEMS & GEMOLOGY
SUMMER 2014
127
FTIR SPECTRA
28
30
Type I
30
26
24
22
20
18
5268 (,)
16
14
12
10
8
6
5430 ()
7095 (,)
7072 ()
6815 ()
7265 ()
7140 (,)
5205
4
2
0
28
Type II
5273 (,)
26
24
22
20
18
16
14
12
10
8
7095 (,)
7076 ()
6
4
2
5340 ()
7140 ()
7265 ()
6840 (,)
5200 (,)
5594 ()
0
7500
7000
6500
6000
5500
5000
4500
7500
7000
-1
6500
6000
5500
5000
4500
WAVENUMBER (cm-1)
WAVENUMBER (cm )
Figure 22. Representative FTIR spectra (solid line = ordinary ray , dashed line = extraordinary ray ) illustrating
the type I and II water observed in emeralds from each origin. Type I samples showed peaks at 7140 (, ), 7095 (,
), 7072 (), 7042 (), 6843 (), 6815 (), 5430 (), 5268 (, ), and 5205 (, ) cm1. Type II samples (which were
only from Kafubu) showed peaks at 7265 (), 7140 (), 7095 (, ), 7076 (), 6840 (, ), 5594 (), 5340 (), 5273 (, ),
and 5200 (, ) cm1.
FTIR. The FTIR spectra recorded for the type I emerald samples all showed a 7140 cm1 peak, higher than
those at 7095 and 7072 cm1 for the extraordinary ray,
related to low alkali metals contents in the channels
of the beryl structure. (Again, Kafubu was the only
locality that yielded type II samples.) Total alkaline
ion concentration, determined using LA-ICP-MS,
showed no relationship with the peak at 7140 cm1.
This was because some of the alkali metals may exist
in sites other than those that influence the water
spectrum. For example, Li may occur in Be or Si sites,
depending on ion sizes.
area. This plot may also help differentiate between

Davdar and Panjshir, as the areas overlap only
slightly. We compare this to Schwarz and Pardieu
(2009), whose diagram of Ga versus Fe showed nearly
complete overlapping between Davdar and Panjshir,
rendering it impossible to distinguish between the
two locations. As the population fields in the log-log
plot of Li vs. Cs overlap for Musakashi and all
Colombian emeralds, a plot of Fe vs. K and Fe vs. Ga
proved useful in separating between these two origins (figures 24 and 25). In addition, the recorded Ni
content was just above the detection limit in emeralds from Musakashi but below the detection limit
in most Colombian samples. Average Fe concentrations were highest in samples from Kafubu. Moreover, the Panjshir emeralds contained higher Sc,
which was very useful in distinguishing them.
Trace-Element Variations. The samples from

Musakashi were characterized by LA-ICP-MS as having low concentrations of alkali metals and Mg, and
high Al content. Most of the Kafubu emeralds
seemed to have an abundance of the alkali metals Li,
Na, K, Rb, and Cs. Various log-log population fields
generated from the trace-element data collected
(with multiphase inclusions) showed great potential
in distinguishing geographic origin (figures 2325).
To explore this further, we added published data for
Brazilian emeralds from Rio Grande do Norte
(Zwaan et al., 2012), which confirmed the usefulness
of this separation criterion. A log-log plot of Li versus
Cs content can easily separate the two Zambian deposits (figure 23). The same type of plot can generally
distinguish Panjshir and Davdar from Musakashi and
Colombian emeralds, even with a small overlapping
For many years, the observation of three-phase inclusions in emeralds was considered a reliable indicator
of Colombian origin. But with the arrival in the market of emeralds from Panjshir (Afghanistan), Davdar
(China), and Kafubu and Musakashi (Zambia) that
may also display three-phase inclusions, origin determination became more complicated. There are some
notable differences, however.
Emeralds from the Musakashi deposit are of particular interest, as their internal features are quite different from those found in the well-known deposits
128
GEMS & GEMOLOGY
CONCLUSIONS
SUMMER 2014
CHEMICAL COMPOSITION
10,000
Musakashi, Zambia
Peas Blancas, Colombia
Kafubu, Zambia
La Pita, Colombia
Coscuez, Colombia
Chivor, Colombia
Rio Grande do Norte, Brazil

(Zwaan et al., 2012)
Muzo, Colombia
Davdar, China
Cs (ppmw)
1,000
Figure 23. Log-log plot of

lithium (Li) versus cesium (Cs) concentrations
in emeralds from the five
different localities.
100
10
1
100
10
1,000
Li (ppmw)
in the neighboring Kafubu area. The inclusion scene

within the Musakashi emeralds is indeed dominated
by multiphase inclusions that are commonly associated with Colombian origin. Nevertheless, the outline and make-up of these multiphase inclusions may
be more irregular than those typical of Colombian de-
posits. Furthermore, the tendency of Musakashi

emeralds to have at least two crystals associated with
a gas bubble may raise questions about a possible
Zambian origin.
Emeralds from Panjshir, Afghanistan, are often
characterized by elongated needle-shaped multiphase
10,000
Musakashi, Zambia
Kafubu, Zambia
La Pita, Colombia
Coscuez, Colombia
Chivor, Colombia

Muzo, Colombia
Davdar, China
K (ppmw)
1,000
Figure 24. Log-log plot

of iron (Fe) versus
potassium (K) concentrations in emeralds
from the five different
localities.
100
10
1
10
100
1,000
Fe (ppmw)
GEMS & GEMOLOGY
SUMMER 2014
129
TABLE 3. Summary of physical properties, UV-Vis-NIR absorption data, and trace-element chemistry of
emeralds from five different localities.
Locality
Musakashi,
Zambia
Refractive
index
1.5721.582
Color filter
reaction
Strong pink
Microscopic
characteristics
Multiphase inclusions
tend to be more irregular
than those in Colombian
gems; multiphase inclusions with at least two
crystals associated with
a gas bubble may
indicate Zambian rather
than Colombian origin
UV-Vis spectroscopy
(cations in addition to Cr3+)
No significant Fe2+-related
absorption features in the
NIR region, and no Fe3+
observed
Trace-element analysis
(84 samples)
Low alkali metal concentrations (1,530
6,060 ppmw)
Cr > V; Cr/V ratio 1.75.3
Fe concentration: 6801,490 ppmw
Li vs. Cs log-log plot can help separate
Musakashi from Kafubu and slightly
separate from Colombia
Fe vs. K log-log plot can help separate
Musakashi from Colombia
Kafubu,
Zambia
Panjshir
Valley,
Afghanistan
Davdar,
China
1.5821.593
(RI values
typically
higher than
other
localities)
Inert
1.5721.590
Inert to pink
1.5771.588
Pink
are usually rectangular
in shape, but may be
irregular in outline; solid
crystals of pyrolusite,
chlorite, mica, amphibole, and tourmaline
Strong broad Fe2+ band at

approximately 810 nm
High alkali metal concentrations

(15,83423,294 ppmw)
O-ray: narrow Fe3+ band at

Cr >> V; Cr/V ratio 840
tend to be more
elongated or needleshaped than other
localities; each may
host several cubic to
rounded transparent
crystals, and sometimes
small, dark opaque
crystals
Medium broad Fe2+ band at Moderate to high alkali metal

concentrations (3,94617,505 ppmw)
Fe concentration: 5,90011,600 ppmw

Kafubu from others
O-ray: Fe3+ peak at 372 nm

in some samples
Cr and V concentrations may be

similar:
Cr/V ratio 0.33.3
Panjshir from Kafubu, Musakashi, and
Colombia (and Davdar, to a lesser
extent)
Similar to Musakashi:
are usually jagged or
Small broad Fe2+ band at
irregular in outline, and approximately 810 nm
some are needle-like
Similar to Colombia:
Strong V3+ bands at approximately 400 and 654 nm
Low to moderate alkali metal

concentrations (5,19012,620 ppmw)
Slightly more V than Cr; Cr/V ratio
0.11.0
Fe vs. K log-log plot can help separate
Davdar from Musakashi and Colombia
Davdar from Panjshir, to a lesser extent
Colombia
(Muzo,
Coscuez, La
Pita, Peas
Blancas, and
Chivor)
130
1.5701.580
(typically
lower than
other
localities)
Pink to
strong pink
Classic jagged
multiphase inclusions
host a gas bubble and
one or more cubic
crystals; gas bubbles are
usually smaller than the
whole inclusion, and
also smaller or about
the same size as the
associated cubic crystal
No Fe2+ or Fe3+ observed
Low alkali metal concentrations

(1,5158,115 ppmw)
Weak to medium V3+ bands
at approximately 400 and
Cr and V concentrations may be
654 nm
similar:
Cr/V ratio 0.04 and 3.5
Fe concentration: 1172,030 ppmw
Fe vs. Ga and Fe vs. K log-log plot can
help separate Colombia from others
GEMS & GEMOLOGY
SUMMER 2014
100
Musakashi, Zambia
Kafubu, Zambia
La Pita, Colombia
Coscuez, Colombia
Chivor, Colombia

Muzo, Colombia
Davdar, China
(Ga) in ppmw
Figure 25. Log-log plot of

iron (Fe) versus gallium
(Ga) concentrations in
emeralds from the five
different localities.
10
100
100
1,000
[Fe] in ppmw
inclusions rather than the traditional jagged type often

seen in Colombian material. Furthermore, Afghan
emeralds often host up to 10 cubic to rounded transparent crystals, and sometimes small dark, opaque
crystals. When examined between crossed polarizers,
the cubic crystals appeared to be singly refractive and
the small rounded crystals appeared doubly refractive.
Such features are not typical of Colombian emeralds.
The multiphase inclusions in emeralds from
Davdar, China, can be very similar to those in emeralds from Colombian deposits. Many of the multiphase inclusions found in Davdar specimens are
jagged and have only one cubic crystal. But the shape
of these multiphase inclusions is usually more irregular (sometimes needle-like) than the jagged outline
typically observed in Colombian emeralds.
ABOUT THE AUTHORS

Ms. Saeseaw is a senior manager of colored stones, Mr. Pardieu
is a senior manager of field gemology, and Dr. Sangsawong is a
research scientist, all at GIA in Bangkok.
ACKNOWLEDGMENTS
The authors thank Mr. Guy Clutterbuck for providing Musakashi
emeralds. We also thank GIAs Jonathan Muyal for lapidary assis-
Studying the inclusion scene in these emeralds

can be quite confusing, but spectroscopy and traceelement chemistry can provide additional valuable
tools to distinguish between them. The presence or
absence of Fe2+ and Fe3+ and the varying intensity of
V3+ absorption may suggest different countries of origin. More generally, plotting the log-log results from
the LA-ICP-MS data for Cs, Li, K, Ga, and Fe resulted in population fields that also appear very
promising as a tool in determining the origin of
emeralds.
In many cases, the combination of the inclusion
scene, UV-Vis-NIR absorption data, and trace-element chemistry (summarized in table 3) can help the
gemologist determine the origin of emeralds with
three-phase or multiphase inclusions.
tance. John Emmett, Kenneth Scarratt, Emily Dubinsky, and

Nathan Renfro provided many helpful discussions. Assistance
from Nattida Ng-Pooresatien and Charuwan Khowpong with data
collection is also greatly appreciated.
GEMS & GEMOLOGY
SUMMER 2014
131
REFERENCES
Behmenburg C. (2002) Beryl and emerald: History and myth. In
Extralapis: Emeralds of the World, extraLapis English No. 2,
Lapis International, East Hampton, CT.
Bowersox G., Snee L.W., Foord E.E., Seal II R.R. (1991) Emeralds
of the Panjshir Valley, Afghanistan. G&G, Vol. 27, No. 1, pp.
2639.
Giard D., Ed. (1998) Lemeraude. Association Franaise de Gemmologie, Paris.
Giuliani G., Cheilletz A., Dubessy J., Rodriguez C.T. (1993) Chemical composition of fluid inclusions in Colombian emerald deposits. 8th Quadrennial IAGOD Symposium, pp. 159168.
Groat L.A., Giuliani G., Marshall D.D., Turner D. (2008) Emerald
deposits and occurrences: A review. Ore Geology Reviews, Vol.
34, No. 1-2, pp. 87112, http://dx.doi.org/10.1016/j.oregeorev.
2007.09.003.
Hammarstrom J.M. (1989) Mineral chemistry of emeralds and some
associated minerals from Pakistan and Afghanistan: An electron
microprobe study. In A.H. Kazmi and L.W. Snee, Eds., Emeralds
of Pakistan: Geology, Gemology & Genesis. pp. 125150.
Klemm L. (2009) Fieldtrip to emerald mines in Zambia, summer
2009. From http://www.gubelingemlab.ch/PDF/Fieldtrips/2009_
Field_report_LK_Zambia_forInternetx.pdf.
(2010) Fieldtrip to emerald mines in Zambia, summer 2010.
From http://www.gubelingemlab.ch/PDF/Fieldtrips/2010_Field_
report_LK_Zambia_Internet.docx.pdf.
Marshall D., Pardieu V., Loughrey L., Jones P., Xue G. (2012) Conditions for emerald formation at Davdar, China: Fluid inclusion,
trace element and stable isotope studies. Mineralogical Magazine, Vol. 76, No. 1, pp. 213226, http://dx.doi.org/10.1180/
minmag.2012.076.1.213.
Michelou J.-C., Pardieu V. (2009) A glimpse of the new fine emerald deposit at Xinjiangs Davdar Mine. InColor, No. 10, Spring,
pp. 2628, 30.
Pardieu V., Soubiraa G. (2006) From Kashmir to Pamir, Summer
2006: Gemmological expedition report to ruby, emerald and
spinel mining areas in Central Asia. Part 4: China (Xin Jiang):
Emeralds from the silk roads. http://www.fieldgemology.org/
gemology%20china%20emerald%20davdar.php.
Schwarz D., Henn U. (1992) Emeralds from Madagascar. Journal
of Gemmology, Vol. 23, No. 3, pp. 140149.
Schwarz D., Pardieu V. (2009) Emeralds from the Silk Road countries A comparison with emeralds from Colombia. InColor,
No. 12, Fall/Winter, pp. 3843.
Schwarz D., Mendes J.C., Klemm L., Lopes P.H.S. (2011) Emeralds
from South America Brazil and Colombia. InColor, No. 16,
Spring, pp. 3646.
Venkateswaran C.S. (1935) The Raman spectra of some metallic
halides. Proceedings of the Indian Academy of Sciences Section A, Vol. 1, No. 11, pp. 850858.
Wood D.L., Nassau K. (1968) The characterization of beryl and
emerald by visible and infrared absorption spectroscopy. American Mineralogist, Vol. 53, May-June, pp. 777799.
Zwaan J.C., Seifert A.V., Vrna S., Laurs B.M., Anckar B., Simmons
W.B., Falster A.U., Lustenhouwer W.J., Muhlmeister S., Koivula
J.I., Garcia-Guillerminet H. (2005) Emeralds from the Kafubu
area, Zambia. G&G, Vol. 41, No. 2, pp. 116148.
Zwaan J.C., Jacob D.E., Hger T., Cavalcanti Neto M.T.O., Kanis
J. (2012) Emeralds from the Fazenda Bonfim region, Rio
Grande do Norte, Brazil. G&G, Vol. 48, No. 1, pp. 217,
http://dx.doi.org/10.5741/GEMS.48.1.2
For online access to all issues of GEMS & GEMOLOGY from 1934 to the present, visit:
gia.edu/gems-gemology
132
GEMS & GEMOLOGY
SUMMER 2014
NOTES & NEW TECHNIQUES
COLOR PHENOMENA OF BLUE AMBER

Yan Liu, Guanghai Shi, and Shen Wang
The greenish blue color observed in some amber

from the Dominican Republic and Indonesia is
actually fluorescence stimulated by ultraviolet
(UV) light. A previous study described the color
as iridescence over a yellow background, but it
is in fact a surface fluorescence. Since this amber
can be considered a long-wavelength pass filter
with a half-pass wavelength at about 530 nm in
its spectral transmittance, it does not transmit either UV or short-wave visible light. As this material completely absorbs light in the UV range and
strongly absorbs light in the short-wave visible
range, the stimulated blue color is confined exclusively to the surface. The amber from this study
also showed the Usambara effect, the phenomenon in which color varies with the path length of
light through a sample.
Although resinites of different ages exist, available biostratigraphic and paleogeographic data suggest that the main amber deposits in the Dominican
Republic (including those famous for yielding biological inclusions) were formed in a single sedimentary
basin during the late Early Miocene through early
Middle Miocene (15 to 20 million years ago). There
is little evidence of extensive reworking or redeposition of the ambers. Before the studies of Iturralde-Vinent and MacPhee (1996), amber from the northern
area was thought to have formed during the Early
Eocene to Early Miocene epochs (Baroni-Urbani and
Saunders, 1982; Lambert et al., 1985; Poinar and Cannatella, 1987; Grimaldi, 1996), while published estimates for the eastern area ranged from the Cretaceous
to Holocene epochs (Burleigh and Whalley, 1983;
Poinar and Cannatella, 1987; Grimaldi, 1996).

Bellani et al. (2005) studied blue amber from the

Dominican Republic and identified the aromatic hydrocarbon component perylene as the source of the
UV-stimulated fluorescence emission in the visible
wavelength range from 430 to 530 nm, resulting in
the observed greenish blue color. They also described
the blue color as iridescence under natural daylight
or daylight-equivalent lighting.
Iridescence is an optical phenomenon in which a
surface appears to shift color as the viewing angle or
angle of illumination changes (Nassau, 1983). Although most iridescence is caused by interference,
Liu et al. (1999b) found that in pearls and shells it results from diffraction. Each layer of shell and pearl is
optically heterogeneous. Light is strongly diffused as
it passes through the layers, causing the milky white
color observed. The diffused light cannot produce interference to cause the iridescence color. For this to
happen, each layer would have to be optically uniform and have a thickness on the order of visiblelight wavelengths.
UV fluorescence is also known to cause significant visible color change in some diamonds. Strong
UV fluorescence under daylight caused such an effect
in the 56.07 ct Tavernier diamond (Liu et al., 1998).
134
GEMS & GEMOLOGY
mber is fossilized tree resin used for jewelry, decoration, medicine, and perfume. Specimens with
inclusions of insects and plants are of great scientific
significance and highly esteemed by collectors.
Amber is usually yellow to brown, and some specimens display red to brownish red or reddish brown
colors. Blue amber is rare, found mainly in the Dominican Republic with some production from Indonesia and Mexico. This variety comes from the
resin of the extinct tree species Hymenaea protera
(Iturralde-Vinent and MacPhee, 1996; Poinar and
Poinar, 1999). According to Iturralde-Vinent and
MacPhee (1996), most Dominican amber occurs in
two zones: north of Santiago de los Caballeros (the
northern area) and northeast of Santo Domingo
(the eastern area).
SUMMER 2014
AT L A N T I C
OCE A N
Northern amber site
Santiago
HAITI
Eastern amber site - 1
DOMINIC AN
Eastern amber site - 2
RE P UB L IC
Santo Domingo
La Romana
CA R I BBE A N SE A
0
80 km
Figure 1. Map of the Dominican amber mines visited

by one of the authors in 2013.
The UV fluorescence changed the entire diamonds

bodycolor from light brown under incandescent light
to light pink under daylight.
The phenomenon by which bodycolor varies with
a gemstones thickness is called the Usambara effect.
Liu et al. (1999a) determined that this color variation
is caused by the hue angle change corresponding
with the path length of light through a gem material.
This article presents the color phenomena of blue
amber from the Dominican Republic and Indonesia.
Fluorescence and photoluminescence were measured
at room temperature to show the characteristic features. As a result, we found that the ambers greenish
blue color is only superficial. By calculating the colorimetric data at different depths within a sample,
Figure 3. These blue amber beads from the Dominican

Republic, 6.012 mm in diameter, are shown under
daylight-equivalent flashlight. Their greenish blue
color is evenly distributed. Photo by Guanghai Shi.
we determined that the hue angle changed with

depth, an indication of the Usambara effect in blue
amber.

In 2013, one of the authors visited amber deposits in
the northern and eastern parts of the Dominican Republic (figure 1). Figure 2 shows a typical mining operation, while figure 3 shows blue amber beads
acquired from the local miners.
In the northern area, the upper 300 meters of the
La Toca Formation of 1,200-meter-thick clastic rocks
contains the amber-bearing unit, composed of sandstone with occasional conglomerate that accumulated in a deltaic to deep-water environment. Amber
Figure 2. Left: Amber is mined by hand in the Dominican Republic. Right: After digging an open pit inside a hill,
the miners typically dig a well more than 10 meters deep to look for an amber layer. Blue amber can occasionally
be found with other varieties of amber. Photos by Guanghai Shi.
GEMS & GEMOLOGY
SUMMER 2014
135
fragments in this deposit, which measure up to 40

cm in length, show few surface signs of transportation. Lignite occurs in the form of thin lamellae
within the sandstones. These clastic rocks grade into
flyschoid, with deeper-water deposits containing detrital amber.*
In the eastern area, the amber-bearing sediments
in the approximately 100 m thick Yanigua Formation
are composed of organically rich laminated sand,
sandy clay, and some intercalated lignite layers up to
1.5 m thick. Amber pieces occur in the lignite and
the sandy clay. In addition to indicative sedimentary
features, the characteristics of the invertebrate and
vertebrate fossils from these bedscrocodiles, sirenians, and turtlesimply that deposition occurred
in a near-shore context, probably in coastal lagoons
fronting low, densely forested hills (Brouwer and
Brouwer, 1982).
Two blue amber samples were studied. Sample
1, a cabochon from the Dominican Republic, appeared to have a greenish blue surface color and a
yellow bodycolor (figure 4). Sample 2, a rough Indonesian amber with one surface polished, showed
blue color on the polished surface and a red bodycolor (figure 5). Both samples showed very strong
greenish blue fluorescence under long-wave UV radiation at 365 nm, and weak fluorescence under
short-wave UV at 254 nm. Both also showed greenish
blue phosphorescence that lasted several seconds.
An MDIS-f8 multifunction dual integrating
sphere spectrometer was used to measure the samples spectral transmittance and UV fluorescence so
Figure 4. Sample 1 was a 1.25 ct blue amber cabochon from the Dominican Republic, measuring 10.98
7.57 4.68 mm. Photo by Yan Liu.
Figure 5. Sample 2, a rough amber from Indonesia,

weighed 35.15 ct and measured 26.21 25.96 20.42
mm. Photo by Yan Liu.
that we could calculate hue and saturation. The optical arrangement for measuring the greenish blue
fluorescence is shown in figure 6. The polished surface of sample 2 was held at a 45 angle to the sam-
In Brief
Blue amberamber that exhibits a greenish blue color
under certain conditionsowes this color component
to fluorescence.
The greenish blue fluorescence color is essentially confined to the surface, as both the excitation and emission wavelengths are completely absorbed.
Blue amber also displays the Usambara effect, whereby
its bodycolor changes from yellow to red with increasing thickness.
ple stage. A 365 nm light-emitting diode (LED) was

used to illuminate the polished surface at a 45 angle
to excite blue fluorescence. A portion of the blue fluorescence passed through the aperture into the integrating sphere, where it was collected by a
collimator and sent to the spectrometer for measurement. The spectrometer was calibrated by a radio-
*Clastic refers to a sedimentary rock composed of broken pieces of

older rocks. Liginite, a brownish black coal, is considered the lowest
grade of coal due to its relatively low heat content. Flyschoid formations consist of cyclic sedimentary deposits and formed under deep
marine circumstances, in a low-energy depositional environment.
136
GEMS & GEMOLOGY
SUMMER 2014
10
2
10
11
5
3
2
3
4
6
8
5
Figure 6. The optical arrangement for measuring the

ambers blue fluorescence featured a multifunction
spectrometer with a 365 nm LED. Components include: (1) light measurement integrating sphere, (2)
sample stage, (3) aperture, (4) collimator, (5) light trap,
(6) baffle, (7) light source, (8) spectrometer, (9) computer, (10) blue amber sample, and (11) 365 nm LED.
metric calibration lamp to measure the relative spectral power distribution of the blue fluorescence
rather than the digital counts. Therefore, the measured blue fluorescence spectrum was true in spectral
power distribution.
Figure 7 shows the optical arrangement for measuring spectral transmittance using the multifunction
spectrometer with a xenon-filled tungsten lamp at
the top of the integrating sphere. We placed an unpolished area chipped from sample 2 that was about
1 mm thick on the measurement stage, where the diameter of the aperture was 3 mm. The light from the
lamp at the top of the measurement integrating
sphere passed through both the slice and the aperture, into the measurement integrating sphere. The
collimator collected the transmitted light from the
amber slice and sent it to the spectrometer for measurement. To measure the percentage of spectral
transmittance, the spectrometer is calibrated to
100% without the sample on the stage, and to 0% by
turning off the tungsten lamp.
12
7
13
Figure 7. The optical arrangement for measuring transmittance uses a multifunction spectrometer, with a
xenon-filled tungsten lamp at the top of the integrating
sphere. Components include: (1) light measurement integrating sphere, (2) sample integrating sphere, (3) sample stage, (4) aperture, (5) disk, (6) collimator, (7) light
trap, (8) baffle, (9) light source for reflectance measurement, (10) light source for transmittance measurement,
(11) blue amber sample, (12) spectrometer, and (13)
computer.
The blue light with a wavelength shorter than 450

nm is completely absorbed, and the transmittance of
visible light from 450 to 760 nm increases gradually.
Thus, the 1 mm slice appears yellow, and the bodycolor of the amber appears orangy yellow to brown
in the thicker area.
Figure 8. This slice, taken from sample 2, is 1 mm

thick at the center. This piece was used for measuring
the spectral transmittance of blue amber. Photo by
Yan Liu.
SURFACE COLOR
Figure 8 shows the unpolished slice taken from sample 2. Figure 9 shows the slices spectral transmittance, which is essentially a typical long-pass
transmittance spectrum with a half-pass wavelength
at about 530 nm and a cutoff wavelength at about
450 nm. The slope is not steep at the short-wavelength end of the long-pass transmittance spectrum.
GEMS & GEMOLOGY
SUMMER 2014
137
TRANSMITTANCE SPECTRUM
SPECTRAL TRANSMITTANCE (%)
100
90
80
70
60
50
40
30
20
10
0
400
500
600
700
WAVELENGTH (nm)
Figure 9. Spectral transmittance of the slice from sample 2. This is a long-wavelength pass spectrum with a
cutoff wavelength at about 450 nm, and a half-pass
wavelength at about 530 nm. Only light with wavelength longer than 450 nm can partially pass through
the amber slice, and the transmittance increases
gradually from 450 to 760 nm.
Figure 10 shows the UV fluorescence spectrum of

sample 2, measured with a 365 nm UV light source.
The UV light was incident to the surface of the sample, and the UV fluorescence was collected from the
surface by the measurement integrating sphere component of the dual integrating sphere spectrometer.
The UV fluorescence spectrum showed three peaks
at around 450, 485, and 500 nm. This UV fluorescence spectrum was similar to the one recorded by
Bellani et al. (2005). It is this short-wavelength fluorescence spectrum that causes the greenish blue surface color of blue amber.
When UV light struck the blue amber, it excited

the greenish blue fluorescence. Because of the total
absorption of UV light, however, it cannot penetrate
into the blue amber. It can only cause blue fluorescence at the surface.
With increasing depth below the surface, the cutoff wavelength shifts to a long wavelength, and the
greenish blue fluorescence is eventually totally absorbed. As this depth is still very shallow, the fluorescent light penetration is confined near the surface,
limiting the ability to change bodycolor.
Figure 11 shows a lateral view of the greenish
blue fluorescence of sample 2. To emphasize the
greenish blue surface UV fluorescence, the photo
was taken with a 365 nm UV light source. The UV
light illuminated the blue amber from the right side,
while incandescent light illuminated the top and
bottom to show the bodycolor. The observed fluorescence agreed with the theoretical fluorescence
based on the fluorescence measurements in the
short-wavelength visible range (see figure 10).
UV light cannot penetrate the ambers surface,
but fluorescence can penetrate to a shallow depth;
Figure 11. Lateral view of the greenish blue fluorescence of sample 2. Long-wave UV illumination from
the right produced the greenish blue fluorescence at
the surface; incandescent illumination from above
and below was used to show its bodycolor. Due to the
short-wavelength cutoff, UV light can only excite the
greenish blue fluorescence at the surface. The greenish blue fluorescence cannot penetrate deep into the
blue amber. Photo by Yan Liu.
Figure 10. The UV fluorescence spectrum of the bulk

piece of sample 2, generated by 365 nm long-wave
UV light.
UV FLUORESCENCE SPECTRUM
RELATIVE INTENSITY
365 nm UV light
400
500
600
700
WAVELENGTH (nm)
138
GEMS & GEMOLOGY
SUMMER 2014
therefore, the very strong greenish blue fluorescence

cannot cause the alexandrite effect of bodycolor unless the sample is very thin. This is different from
the alexandrite effect previously mentioned in the
Tavernier diamond, where the greenish blue fluorescence penetrated the stone and caused a change in
its bodycolor. The greenish blue fluorescence color
of our amber samples was superficial.
UV fluorescence is generally very weak compared to the reflected or transmitted light of a materials bodycolor under daylight or daylightequivalent light, even if the UV fluorescence is very
strong. This statement holds true for the greenish
blue fluorescence of blue amber. Under daylight, the
surface greenish blue fluorescence is difficult to observe, and only the yellow to brownish orangy red
bodycolor is apparent. The surface fluorescence is
usually observed when the blue amber is held
against a black background. The black background
limits light reflection, enhancing the fluorescence.
THE USAMBARA EFFECT

One surface of sample 2 was polished to form a
wedge shape to show color change with different
thicknesses. With increasing thickness, sample 2s
cutoff wavelength moves to longer wavelengths, and
its transmission color changes from yellow to red
(figure 12). The hue is yellow where the sample is
thin and changes to brownish orangy red with increasing thickness.
Table 1 shows the calculated relationship between the thickness and hue angle of sample 2 in the
CIELAB color space under the CIE standard illuminant A at a color temperature of 2856 K, which represents the incandescent light used to observe the
transmission color (Wyszecki and Stiles, 2000).
Where the amber is 1 mm thick, its hue angle is 88.1
and its chroma is 33.4. The 88.1 hue angle represents a yellow hue, the 33.4 chroma a medium saturation. As the thickness increases, the hue angle
becomes smaller and the chroma becomes higher.
When the thickness reaches 7 mm, the hue angle is
lowered to 76.7 and the chroma achieves its maximum saturation at 75.0. This hue angle and chroma
represent an orange color with high saturation. When
the thickness reaches 25 mm, the hue angle is reduced to 62.8, observed as an orangy red hue, and
the chroma becomes 55.1, representing a mediumhigh saturation.
Figure 13 shows the hue angle change of the blue
amber in the CIELAB color space under the CIE
TABLE 1. Calculated relationship between thickness

and hue angle and chroma of a blue amber sample.
Thickness (mm)
Hue angle
Chroma
88.1
33.4
85.6
51.7
83.3
62.0
81.4
68.0
79.6
71.6
76.7
75.0
10
75.3
73.1
15
70.2
68.7
20
65.4
62.9
25
62.8
55.1
standard illuminant A at 2856 K. The locus of these

hue angles is a closed curve, with a hue angle variation of more than 20 in CIELAB color space. This
locus is typical of the Usambara effect, which describes color variation with material thickness. In
fact, many red and reddish brown ambers we have
observed, with or without the blue fluorescence,
clearly displayed the Usambara effect. As sample
thickness increased, their bodycolor went from yel-
Figure 12. The bodycolor of sample 2 changes from

yellow to orangy red where the stone is thicker. Photo
by Yan Liu.
GEMS & GEMOLOGY
SUMMER 2014
139
b*
CONCLUSION
a*
Figure 13. In CIELAB color space, the hue angle locus

of sample 2 shows a variation of more than 20, attributable to the Usambara effect.
low to brownish orange or orangy brown, brownish

red, or even red.
ABOUT THE AUTHORS

Mr. Liu is president of Liu Research Laboratories in South El
Monte, California. Dr. Shi is a professor at the School of Gemology, China University of Geosciences in Beijing. Mr. Wang is a
student at Swarthmore College in Swarthmore, Pennsylvania.
The greenish blue color of blue amber from the Dominican Republic and Indonesia is caused by UVstimulated fluorescence. The greenish blue fluorescence color is confined to the surface. There are two
reasons for this: (1) UV light can only stimulate the
greenish blue fluorescence on the surface of amber,
and (2) amber strongly absorbs the greenish blue fluorescence light and does not allow it to penetrate much
below the surface. Where a sample is thicker, the cutoff wavelength of the long-wavelength pass spectrum
moves to a longer wavelength, and the greenish blue
fluorescence can only penetrate a very thin layer of
the amber. Thus, the blue fluorescence only appears
on the surface.
Blue ambers bodycolor changes with increasing
thickness: yellow in areas less than 3 mm thick,
orangy yellow to yellowish orange at approximately
36 mm, orange at 68 mm, and reddish orange and
orangy red where the stone is more than 8 mm
thick.
ACKNOWLEDGMENTS
This research was partially supported by Grant No. 41373055
from the National Science Foundation of China.
REFERENCES
Baroni-Urbani C., Saunders J.B. (1982) The fauna of the Dominican
Republic amber: the present status of knowledge. In W. Snow et
al., Eds., Transactions of the 9th Caribbean Geology Conference,
Vol. 1, pp. 213223.
Bellani V., Giulotto E., Linati L., Sacchi D. (2005) Origin of the blue
fluorescence in Dominican amber. Journal of Applied Physics,
Vol. 97, No. 1, p. 016101, http://dx.doi.org/10.1063/1.1829395.
Brouwer S.B., Brouwer P.A. (1982) Geologia de la region ambarifera
oriental de la Republica Dominicana. In W. Snow et al., Eds.,
Transactions of the 9th Caribbean Geology Conference, Vol. 1,
pp. 305322.
Burleigh R., Whalley P.J. (1983) On the relative geological ages of
amber and copal. Journal of Natural History, Vol. 17, pp. 919
921.
Grimaldi D.A. (1996) Amber: Window to the Past. Harry Abrams,
New York.
Iturralde-Vinent M.A., MacPhee R.D.E. (1996) Age and paleogeographical origin of Dominican amber. Science, Vol. 273, pp.
18501852, http://dx.doi.org/10.1126/science.273.5283.1850.
Lambert J.B., Frye J.S, Poinar G.O. (1985) Amber from the Dominican Republic: Analysis by nuclear magnetic resonance spectroscopy. Archaeometry, Vol. 27, pp. 4351.
Liu Y., Shigley J., Moses T., Reinitz I. (1998) The alexandrite effect
of the Tavernier diamond caused by the fluorescence under daylight. Color Research & Application, Vol. 23, No. 5, pp. 323327,
http://dx.doi.org/10.1002/(SICI)1520-6378(199810)23:5<323::
AID-COL8>3.0.CO;2-Y.
Liu Y., Shigley J., Halvorsen A. (1999a) Color hue change of a gem
tourmaline from the Umba Valley, Tanzania. Journal of Gemmology, Vol. 26, No. 6, pp. 386396.
Liu Y., Shigley J., Hurwit K. (1999b) Iridescent color of a shell of the
mollusk Pinctada margaritifera caused by diffraction. Optics
Express, Vol. 4, No. 5, pp. 177182.
Nassau K. (1983) The Physics and Chemistry of Color. John Wiley
and Sons, New York.
Poinar G.O., Cannatella D.C. (1987) An Upper Eocene frog from
the Dominican Republic and its implications for Caribbean
biogeography. Science, Vol. 237, No. 4819, pp. 12151217,
http://dx.doi.org/10.1126/science.237.4819.1215.
Poinar G., Poinar R. (1999) The Amber Forest: A Reconstruction of
a Vanished World. Princeton University Press, Princeton, New
Jersey.
Wyszecki G., Stiles W. (2000) Color Science, 2nd ed. John Wiley and
Sons, New York.
140
GEMS & GEMOLOGY
SUMMER 2014
EXPERIMENTAL STUDIES ON THE HEAT TREATMENT OF

BALTIC AMBER
Yamei Wang, Mingxing Yang, and Yiping Yang
Amber samples were subjected to heat treatment

experiments aimed at reproducing heated products found in the jewelry market today. Using a
furnace and applying specific conditions, the authors produced four types of heated amber:
golden, red, flower, and beeswax. This paper
describes the procedures and parameters that influence the characteristics of heated amber.
mbers naturally soft color is appreciated by connoisseurs the world over. In China, the gem has
been regarded as sacred since ancient times. It is one
of Chinas seven Buddhist treasures, along with gold,
silver, pearl, coral, Tridacna clams, and liuli (glass).
Huge quantities of amber are consumed each year,
especially in the Chinese market, but high-quality
natural material is rare and expensive. As a result,
some dealers seek to improve its appearance and
quality through heat treatment, which expands the
variety available but poses identification challenges.
Numerous studies have dealt with the identification of heat-treated amber (e.g., Brody et al., 2001;
Zhang, 2006; Feist et al., 2007; Abduriyim et al.,
2009). Yet the methods of heat treatment are closely
guarded and quite complex. Details surrounding enhancement mechanisms, the evolution of flow striations, and the reworking of inclusions, as well as
changes in gemological properties under different
treatment conditions, are not well understood.
In recent years, a few reports have focused on reproducing the heat treatment of amber (e.g., Zhang
2003; Abduriyim et al., 2009; Ross, 2010). In one
study, samples were placed in plant oil or fine-grained

sands and heated, improving their transparency and generating sun spark inclusions. Such methods are not
easily duplicated, however. This investigation adopts
standard equipment and procedures for commercialscale heat treatment and explores the mechanisms
behind a series of changes resulting from the treatment.
To simulate the heat-treated varieties available in
the Chinese market, the authors chose raw Baltic
amber specimens from Kaliningrad, Russia, and performed a series of heating experiments at an anonymous processing facility. We succeeded in producing
a full range of heated amber varieties, including
golden, red, flower, and beeswax (figure 1). These
Figure 1. These photos show four types of heated

Baltic amber found in the market. Clockwise from
top left: a 21.34 g golden amber, a 42.80 g red amber,
a 23.21 g beeswax amber, and a 11.95 g golden
amber with flower inclusions. Photos by Yamei Wang,
courtesy of Shengji Co.

142
GEMS & GEMOLOGY
SUMMER 2014
7
1
2
1
JB
JA
JC
1
5
Figure 2. Raw amber

specimens from Kaliningrad, Russia, were collected for heat treatment
experiments. The whole
specimens ranged from
27.23 g (sample JB) to
74.70 g (sample JE).
Photo by Yamei Wang.
JF
JD
JE
experiments yielded useful information about procedures, conditions, and influencing factors for heat
treatment. The gemological parameters and inclusion characteristics may provide valuable references
for the identification of heat-treated amber.

Six good-quality Russian amber specimens were obtained from Beata Amber (Vilnius, Lithuania) for heat
treatment. The raw samples weighed 380 g total and
were labeled JAJF. They ranged from orange-yellow
to honey yellow in color and from transparent to
opaque, with naturally weathered surfaces (table 1).
For this study, each specimen was cut into several
subsamples, designated as JA-1 through JA-5, etc. (figure 2). This enabled us to track and observe changes
in structure and properties at different experimental
stages.
For the heat treatment experiments, we used a
Kapex AB/3/KAP furnace (figure 3). The automatic
control and protection system can be connected to
the pressure furnace through a stop valve, depending
on the desired enhancement effect. Slowly charging
inert gas (N2) into the pressure furnace and heating
it removes cloudy bubbles inside the amber, improving its transparency and hardness. By regulating the
gas medium used for enhancement, the resulting
color can also be adjusted.
eters and overall effectiveness. It should be noted

that these parameters are given in averages or approximate values, as detailed data were not divulged by
the processing facility.
Purification Process. Purification removes bubbles
from amber in an inert gas medium, improving transparency. This process requires careful regulation of
temperature and pressure inside the autoclave to
avoid melting the amber. Thick specimens with poor
transparency may need to undergo several rounds of
purification, or one lengthy process at higher pressure and temperature. Purification is the primary
method for producing golden and pearly beeswax
amber, and it is commonly used as a pretreatment
before oxidation or sun sparking.
Figure 3. A Kapex AB/3/KAP amber furnace was used in

the heat treatment experiments. Photo by Yamei Wang.
RESULTS
The heat treatment of amber generally consists of
four stages: preparation, dosing, enhancement, and
blowing-out. Details of the stages are described in
table 2. Table 3 shows the samples appearance before
and after treatment, as well as the treatment param-
GEMS & GEMOLOGY
SUMMER 2014
143
TABLE 1. Characteristics of the amber samples used in heat treatment experiments.

Sample
code
Description
JA
Golden specimen, rectangular and sub-angular in

shape. Transparent, with a yellowish orange and finegrained surface. The interior is clean, occasionally
showing small silky and cloudy inclusions, a few tiny
gas bubbles, and an apparent disc-shaped crack.
JB
Beeswax-colored specimen, nearly trigonal to subtriangular in shape, showing a color variation of golden
yellow to milky white from the edge to the center.
Transparent to opaque, exhibiting a brownish yellow
and saccharoidal surface. The center shows smooth flow
structure, gas bubbles, and a disc-shaped structure.
JC
Beeswax-colored specimen. Slightly transparent to

semitransparent, showing a brownish dendritic
surface. The inside displays flow striations and clearly
defined boundaries, with bubbles visible in cloudy
inclusions.
JD
Beeswax-colored specimen, nearly elliptical in shape.

Slightly transparent to semitransparent, showing a
brownish yellow dendritic surface. The inside displays
smooth flow striations and bubbles.
JE
Beeswax-colored specimen, nearly rectangular.

Opaque, with a brownish yellow surface. The inside
shows mottled or smooth distribution of light yellow
and white veins with clearly defined boundaries.
JF
Beeswax-colored specimen, nearly trigonal and subangular. Milky white, opaque, with a brownish yellow
dendritic surface and flow structures inside.
For the first test, we chose samples JA-3, JA-4, JA5, and JC-7; the JA samples are transparent, naturally
golden amber, and JC-7 has an opaque yellow
beeswax color. We started with a gas pressure of 4.5
MPa and a room temperature of 27C, which was increased over a three-hour period to 200C. That temperature was maintained for about two hours, then
cooled down to 35C over a period of 14 hours, at
which point the samples were removed. After purification, all the samples showed better transparency.
Sample JC-7 exhibited the most dramatic improvement, changing from near opaque to transparent at
the edges, with slight transparency at the center for
a pearly beeswax effect. Its color did not change.
Samples JA-3, JA-4, and JA-5 changed from golden
yellow to brownish yellow. After purification, the JA
samples were reserved for oxidation and sparking.
Theoretically, less-transparent amber can be purified in a single round by setting the proper conditions, but since the process is performed in a sealed
tank, the effects cannot be observed in real time. To

prevent overheating and damaging the ambers physical-chemical properties, multiple purifications may
144
GEMS & GEMOLOGY
In Brief
Portions of six raw amber specimens were subjected to
heat treatment experiments to understand the conditions necessary to create specific effects in amber.
Purification, oxidation, and sparking processes yielded
four distinct types of heat-treated amber: golden, red,
flower, and beeswax.
The wide variety and quality of amber on the market
depends on the materials inherent properties and the
treatment parameters.
be used for transparency enhancement. For the second test, semitransparent to opaque sample JB-2 and
SUMMER 2014
TABLE 2. General procedures for the heat treatment of amber.

Stage
Description
1. Preparation
Check the autoclave operation system.

Determine operating parameters such
as temperature, pressure, and atmosphere for heat treatment, based on
thickness, size, and transparency of
semi-finished and finished amber
products.
2. Dosing
Arrange the samples on an iron tray

and place inside the autoclave. Close
and tighten the cover and insert the
sensor.
3. Enhancement Set the pressure, fill the autoclave with

an inert gas (or oxygen) until the initial
pressure is reached, and shut off the
gas valve. Turn on the autoclave power
supply and set the temperature, heating
rate, and duration. After heating, the
system will automatically shut off.
4. Blowing-out
When the autoclave naturally cools

down to about 40C, shut off the
power supply. Remove the temperature sensor, release the gas inside
the stove, open the cover, and retrieve
the specimens.
opaque beeswax samples JB-4, JC-1, JC-4, JD-3, JD-4,

JD-5, JD-6, JE-2, JF-1, and JF-2 were selected for two
purification cycles, using the procedure described for
the first test. The initial pressure was set at 5 and 5.5
MPa, respectively; the temperature was set at 200
210C, with heating times of five and six hours, respectively.
After two purification cycles, the cloudy inclusions disappeared, turning the samples transparent.
Samples JB-2 and JD-3 became transparent brownish
yellow golden amber, while the others turned brownish red. Polishing off the superficial red oxidized layers of samples JC-4, JD-4, JE-2, and JF-2 revealed a
golden yellow color. After polishing, samples JD-4,
JD-5, and JD-6 were set aside for sparking and oxidation. The red oxidation is thought to be related to
remnants of air left inside the autoclave from incomplete purging or from the presence of oxygen impurity in the nitrogen gas, either of which would
oxidize the ambers surface and turn it red during pu-
rification. Charging air into the autoclave to a pressure of 0.2 MPa and then releasing it would improve
air exhaust and help prevent such oxidation. When
the preset temperature is reached inside the autoclave, slowly releasing some air and exhaust vapor
would help accurately regulate pressure for the purification test.
GEMS & GEMOLOGY
Oxidation Process. Oxidation at an elevated temperature is the main process used to create red amber. It
produces a red to deep reddish brown oxidized layer
on the surface. Purified samples JA-4, JA-5, JB-4, JC2, JD-6, and JF-3 were selected for oxidation in a
sealed autoclave. The process was somewhat different
from the purification test, in that a small amount of
oxygen was charged into the inert gas media, according to a strict volume ratio of the two gases. Too little
oxygen would produce an insufficient oxidation effect
on the amber surface, while too much oxygen would
be dangerous to the test equipment and the material
SUMMER 2014
145
due to carbonization. Pressure was set at 4.5 MPa and

the temperature was raised to 210C over three hours,
with inert gas charged into the autoclave together
with a small amount of oxygen.
After testing, all the samples had transformed
into red to blackish red amber. A longer heating time
generally means a deeper color of red amber. If one
oxidation did not produce the expected effect, another could be performed. Only gas pressure needs to
be increased (under the same temperature) by 0.51.0
MPa above the previous oxidation test; otherwise,
the amber samples tend to show sparking inside.
Sun Spark Inclusions. To create small sun spark
inclusions, the bubbles present in amber need to be
burst. After heating the amber as in a purification
step, the gas from the autoclave is released, causing a
rapid decrease of interior pressure and breaking down
the equilibrium between the inner and outer pressure
of a bubble. This results in the expansion and sparking
of bubbles and the formation of disc-shaped internal
cracks, since the pressure inside the bubbles exceeds
the pressure outside them. These disc-shaped cracks
are termed sun spark inclusions. Amber with these
inclusions is known as flower amber. Multiple rounds
of sparking are needed to produce golden flower and
red flower amber varieties.
Conventional sparking consists of deep-frying and
baking in sand, and the sparking effect is visually regulated. Because these processes are time-consuming
and quantities of amber containing the appropriate
amount of gas/liquid inclusions are limited, most
flower amber is manufactured by autoclave.
In our testing, sample JE-5 achieved the best results,

which came from two rounds of sparking.
Red Flower Amber. Red flower amber is produced by
two different methods: (1) sparking red amber and (2)
oxidizing golden flower amber. The first process involves manufacturing red amber, stopping the heating,
and directly releasing the air. This results in an instantaneous pressure decrease that produces sparking inside the red amber; when the superficial red oxidized
layer is polished off, red flower amber is revealed. In
the second method, golden flower amber undergoes oxidation treatment and its oxidized skin is polished off.
Although the first method is more efficient, both yield
identical results.
To produce red flower amber, we chose samples
JC-3, JD-4, JE-3, and JE-4. The parameters for the
sparking were an initial pressure of 3.0 MPa and a
maximum temperature of 200C after three hours of
heating. Sample JC-3 was successful, while the other
three failed. Sample JA-5 underwent oxidation, followed by sparking, after which the dark red layer was
partially polished off to expose the golden yellow
bodycolor. In this manner, golden and red bicolored
flower amber was generated.
Beeswax Amber. The method for manufacturing
beeswax amber is relatively simple but consumes
considerable time and energy. The process requires
prolonged heating and slow oxidation under constant
pressure and low temperature. First, samples are
arranged in an iron tray that is spread with a layer of
fine-grained sand and placed in the furnace. The temperature is set at approximately 5060C, for a heating period of 60100 days. During the baking process,
the temperature must be controlled within a certain
range. Samples JD-2 and JE-6 were chosen to produce
beeswax samples. The specimens were heated at
about 60C for 60 and 90 days, respectively. After
heating, the samples had a dull yellow color like ancient beeswax.
Golden Flower Amber. Sparking for golden flower

amber starts with the same procedures used in purification but differs in the latter stages. After heating
is complete, the power supply is immediately
switched off, releasing gas inside the autoclave to
cool down the pressure chamber. Samples JA-3, JD5, and JE-5, which contain relatively abundant gas
bubbles, were selected for this test. To prevent excessive sparking, which would result in cracking, all
samples underwent purification treatment beforehand. For the first sparking test, the initial pressure
was set at 2.0 MPa. The maximum temperature was
set at 200C, which was reached within two hours
and maintained for one hour. Samples that failed to
produce a flower effect after the first sparking test underwent a second and in some cases a third sparking
test, at successively higher pressure and temperature
conditions, until a golden flower effect was obtained.
Heat treatment of amber is aimed at altering or improving color, enhancing transparency, or producing
inclusions with special effects. Different results were
targeted, including purification, oxidation, sun
spark, and beeswax effects. Based on these tests,
several conclusions can be drawn.
Beeswax-colored amber can be purified as transparent golden material using an initial pressure of 4.55.0
MPa and a maximum temperature of 200210C
146
GEMS & GEMOLOGY
DISCUSSION
SUMMER 2014
TABLE 3. Enhancement parameters and characteristics of amber samples.

Sample
Before test
After test
Enhancement method(s) and

parameters
Assessment
JA-3
(9.44 g)
Golden yellow
Bicolored golden and red

flower
Purification, sparking
Temperature: 200C
Time: 2.5 hours
Pressure: 3.0 MPa
Ambient medium: inert gas
The sun spark was small

and only slightly visible.
JA-4
(8.94 g)
Golden yellow
Red
Purification, oxidation
Temperature: 210C
Time: 3.0 hours
Pressure: 4.5 MPa
Ambient medium: inert gas +
oxygen
The expected effect was

achieved.
JA-5
(6.90 g)
Golden yellow
Bicolored red and golden

flower
Purification, oxidation, sparking

Temperature: 210C
Time: 3.0 hours
Pressure: 4.5 MPa
oxygen

achieved.
JB-2
(6.82 g)
White and golden yellow

pearly beeswax
Brownish yellow golden
Purification
Temperature: 200C
Time: 5.0 hours
Pressure: 5.0 MPa

largely achieved; only the
right lower corner part was
not completely transparent.
JB-4
(5.42 g)
White and golden yellow

beeswax
Blackish red
Temperature: 210C
Time: 3.0 hours
Pressure: 4.5 MPa
oxygen

achieved.
JC-1
(6.58 g)
White and honey yellow

beeswax
Dark red
Purification
Temperature: 210C
Time: 6.0 hours
Pressure: 5.5 MPa
Ideal color was produced,

but transparency was not
ideal.
JC-2
(7.82 g)

beeswax
Red
Temperature: 210C
Time: 3.0 hours
Pressure: 4.5 MPa
oxygen

achieved.
JC-3
(8.71 g)

beeswax
Golden with red flower
Purification, sparking, oxidation

Temperature: 200C
Time: 3.0 hours
Pressure: 3.0 MPa
oxygen

achieved, with the middle
part showing attractive sun
spark inclusions.
GEMS & GEMOLOGY
SUMMER 2014
147
TABLE 3 (continued). Enhancement parameters and characteristics of amber samples.

Sample
Before test
After test

parameters
Assessment
JC-4
(8.78 g)

beeswax
Bright yellow golden
Purification
Temperature: 210C
Time: 6.0 hours
Pressure: 5.5 MPa

achieved, due to oxidation
of the polished surface after
three rounds of purification.
JC-7
(25.64 g)

beeswax
Pearly beeswax
Purification
Temperature: 200C
Time: 5.0 hours
Pressure: 4.5 MPa

achieved.
JD-2
(5.66 g)
Honey yellow beeswax
Beeswax
Beeswax baking
Temperature: 60C
Time: 60 days
Pressure: constant
Ambient medium: air

achieved.
JD-3
(8.36 g)
Brownish yellow golden
Purification
Temperature: 210C
Time: 5.5 hours
Pressure: 5.5 MPa

achieved after three rounds
of purification.
JD-4
(6.18 g)
Bright yellow golden
Temperature: 200C
Time: 3.0 hours
Pressure: 3.0 MPa
After three rounds of purification, the expected effect

was achieved. The sparking
test failed, as no sun spark
inclusions were produced.
JD-5
(5.26 g)
Golden flower
Temperature: 200C
Time: 2.5 hours
Pressure: 3.0 MPa

achieved.
JD-6
(6.82 g)
Red
Temperature: 210C
Time: 3.0 hours
Pressure: 4.5 MPa
oxygen

achieved.
JE-2
(11.50 g)
White and yellow beeswax
Bright golden
Purification
Temperature: 210C
Time: 6.0 hours
Pressure: 5.5 MPa
After three rounds of purification and polishing of the

oxidized surface, the
expected effect was largely
achieved, though cracks
formed inside the amber.
148
GEMS & GEMOLOGY
SUMMER 2014
Sample
Before test
After test

parameters
Assessment
JE-3
(11.25 g)
Golden and red

Temperature: 200C
Time: 3.0 hours
Pressure: 4.0 MPa
oxygen
The test failed, as no sun

spark inclusions were
produced.
JE-4
(13.87 g)
Red flower

Temperature: 200C
Time: 3.0 hours
Pressure: 4.0 MPa
oxygen
The test failed, as no sun

spark inclusions were
produced.
JE-5
(12.53 g)
Golden flower
Temperature: about 200C
Time: 2.5 hours
Pressure: Approx. 3.0 MPa
Ideal sun spark inclusions

were created after three
rounds of purification.
JE-6
(12.00 g)
Dull yellow beeswax
Beeswax baking
Temperature: 60C
Time: 90 days
Pressure: constant
Ambient medium: indoor
open environment

achieved.
JF-1
(11.78 g)
Milky white beeswax
Dark red
Temperature: 210C
Time: 6.0 hours
Pressure: 5.5 MPa
After three rounds of purification, red color resulted

from slow oxidation, but the
transparency was not ideal.
JF-2
(9.76 g)
Milky white beeswax
Golden
Purification
Temperature: 200C
Time: 5.0 hours
Pressure: 5.0 MPa
After three cycles of purification, the transparency was

not noticeably improved.
JF-3
(7.45 g)
Milky white beeswax
Red
Temperature: 210C
Time: 3.0 hours
Pressure: 4.5 MPa
oxygen

largely achieved.
GEMS & GEMOLOGY
SUMMER 2014
149
(reached in three to six hours and maintained for two

hours), followed by 14 hours of cooling. Achieving the
desired effect, such as pearly beeswax, may require
multiple rounds of purification. Purification is also the
pretreatment for other enhancement processes.
Well-controlled oxidation generates red amber by
producing a red to dark red surface layer. The parameters are an initial pressure of 4.5 MPa (involving a
specific volume ratio of inert gas and oxygen) and a
maximum temperature of 210C, to be reached in approximately three hours. The key parameter for oxidation is the amount of oxygen added: Too little
oxygen produces a very light color, while too much
increases the likelihood of carbonizing the test materials and damaging the equipment.
Samples containing certain levels of gas/liquid inclusions that undergo purification followed by a rapid
pressure release acquire internal sun spark inclusions, converting them to flower amber. The parameters are: initial pressure of 2.0 MPa and a maximum
temperature of 200C, reached over a period of two
hours. Sparking generally requires several rounds of
treatment before the effect can be considered satisfactory, with higher pressure and temperature with
each additional treatment. Red flower amber can be
manufactured by two methods: the sparking of red
amber, or the oxidation and polishing of golden
flower amber.
Beeswax-colored amber is created by prolonged oxidation (color darkening) on the surface at constant
pressure and low temperature, which involves a tem-
ABOUT THE AUTHORS

Ms. Wang is vice professor at the Gemological Institute, China
University of Geosciences in Wuhan, and the director of the universitys gem testing center in Guangzhou. Dr. Yang
([email protected]) is dean of the Gemological Institute. Ms. Yang,
formerly a postgraduate at the Gemological Institute, works in the
amber trade in Guangzhou.
perature of about 5060C and a baking time of two

to three months.
Amber heat treatment is affected by several factors, including:
1. Material properties: sample color, transparency,
and bulk size, which determine the initial heat
treatment parameters.
2. Temperature and pressure conditions: the time
allowed for heating, maintaining temperature,
and cooling, as well as the pressure and the rate
of pressure decreases. These are the key parameters for achieving satisfactory amber products.
3. Atmospheric inert gas and oxygen, and the
ratio between the two, which can determine
the color of heat-treated amber.
CONCLUSION
This paper reports on methods of amber heat treatment and discusses several influencing factors. The
study suggests that the variety and quality of heated
ambers are influenced by the inherent properties
(color, transparency, and bulk size) of the natural material, as well as the heat treatment parameters, including temperature, pressure, and atmosphere.
Using a furnace and various parameters, four types of
heated amber were produced: golden, red, flower,
and beeswax. The study documented the various
procedures for amber heat treatment, yielding valuable data on how different effects are achieved.
ACKNOWLEDGMENTS
This work was financially supported by the Gemological Institute,
China University of Geosciences in Wuhan. We are grateful to the
peer reviewers and technical editors for constructive feedback.
The authors would like to thank Beata Amber for providing the
samples used in this study.
REFERENCES
Abduriyim A, Kimura H, Yokoyama Y., Nakazono H., Wakatsuki
M., Shimizu T., Tansho M., Ohki S. (2009) Characterization
of green amber with infrared and nuclear magnetic resonance spectroscopy. G&G, Vol. 45, No. 3, pp. 159171,
Brody R.H.E., Howell G.M., Pollard A.M. (2001) A study of amber
and copal samples using FT-Raman spectroscopy. Spectrochimica Acta Part A, Vol. 57, No. 6, pp. 13251338, http://dx.doi.org/
10.1016/S1386-1425(01)00387-0.
Feist M., Lamprecht I., Mller F. (2007) Thermal investigations of
150
amber and copal. Thermochimica Acta, Vol. 458, No. 1-2, pp.
162170, http://dx.doi.org/10.1016/j.tca.2007.01.029.
Ross A. (2010) Amber: The Natural Time Capsule. Firefly Books
Ltd., Ontario, Canada.
Zhang P.L. (2006) Systematic Gemology, 2nd ed. Geology Press,
Beijing (in Chinese).
Zhang S.H. (2003) Can heat treatment change refractive index?
Journal of Gems and Gemmology, Vol. 5, No. 2, p. 1415 (in
Chinese with English abstract).
GEMS & GEMOLOGY
SUMMER 2014
Editors
Thomas M. Moses | Shane F. McClure
DIAMOND
Chameleon, with Nickel
Absorption Band
Editors note: All items were written by staff

members of GIA laboratories.
Figure 1. The chameleon diamond

is shown before and after heating
(left and right).
Figure 3. DiamondView imaging

showed uneven blue and green
fluorescent zones, proving the
stones natural origin.
of a collection of chameleon diamonds, Spring 2005 G&G, pp. 20

34), its role in the color-change effect
is unknown. Nickel has also been re-
ported as a cause of green coloration

in certain diamonds (W. Wang et al.,
Natural type Ia diamond with greenyellow color due to Ni-related de-
Figure 2. The absorption spectrum of the chameleon diamond shows a

nickel-related band at 685 nm.
VISIBLE-NIR SPECTRUM
.04
ABSORBANCE (a.u.)
The Carlsbad laboratory recently examined a 0.31 ct Light greenish yellow

marquise-cut diamond. The stone displayed the strong yellow fluorescence,
persistent yellow phosphorescence,
and green color component of a
chameleon diamond. A chameleon diamond has a green or greenish color
component under normal conditions.
When heated or left in the dark for
long periods of time, the green component temporarily disappears, giving
way to an orange component. Figure 1
shows the diamond before and immediately after heating, in which the removal of the green component leads
to an orange-dominant hue, demonstrating that the stone was in fact a
chameleon diamond. Unfortunately,
the stones low saturation makes the
effect less noticeable.
Examination of the visible-NIR
spectrum revealed a noticeable
nickel-related absorption band at 685
nm (figure 2). DiamondView images
were taken to prove the stone was
not synthetic, as nickel is a common
catalyst in HPHT synthetics (figure
3). While nickel has long been known
to occur as a trace element in
chameleon diamonds (T. Hainschwang et al., A gemological study
.03
Ni-related
absorption band
.02
.01

500
LAB NOTES
550
600
650
700
750
800
WAVELENGTH (nm)
GEMS & GEMOLOGY
SUMMER 2014
151
fects, Fall 2007 G&G, pp. 240243),

but this may be the first time it has
been identified as the major cause of
a green component in a chameleon.
The rarity of chameleon diamond,
combined with the rarity of Ni-related natural green diamonds, makes
this a truly unique specimen.
Troy Ardon
Long-Term Durability of CVD

Synthetic Film on Natural Diamond
In 2005, a thin film (< 10 m) of synthetic diamond grown by chemical
vapor deposition was deposited on the
pavilion of a 0.27 ct natural diamond.
The CVD layer was deliberately
doped with boron to create a bluish
color. No subsequent polishing was
performed. Rather than blue, the resulting color was a less desirable gray.
This natural diamond with CVD
overgrowth was then reset in a ring
and worn daily by the author for the
next eight years. Recently, this hybrid
diamond was removed from its setting for gemological and spectroscopic
characterization and to assess the current state of the CVD film.
Well-controlled durability studies
have examined non-diamond coatings
on diamond (e.g., A.H. Shen et al.,
Serenity coated colored diamonds:
Detection and durability, Spring 2007
G&G, pp. 1634) and nanocrystalline
diamond coatings on non-diamond
materials (J.E. Shigley et al., Characterization of colorless coated cubic zirconia [Diamantine], Spring 2012
G&G, pp. 1830). To our knowledge,
however, none have been performed
on CVD diamond films grown on natural diamond. The assessment of this
one sample, given the exposure a ring
is usually subjected to, can indicate the
long-term stability of such coatings on
diamond. The expectation was that
such films would remain stable, as
there is no lattice mismatch in the
crystal structure between the CVD
synthetic film and the natural diamond. Lattice mismatch and the resulting low cohesion at the diamond/
non-diamond interface typically result
in moderate to very poor durability.
152
LAB NOTES
It was difficult to find evidence of

the CVD layer using spectroscopic
techniques. FTIR absorption (which
generally provides data from the bulk
of the diamond) revealed no CVDspecific features, only that the underlying diamond was type Ia. Using PL
spectroscopy, which collects data
from a limited area, it was difficult to
find spectroscopic evidence of the
CVD film, even when the PL data
were collected in confocal mode. No
CVD-related 3123 cm1 peak was detected in IR absorption, and the silicon peak at 737 nm was not observed
in PL. The best evidence of the CVD
layers continued existence on the diamond was from its unusual appearance in visual observation (figure 4)
and microscopic analysis (figure 5).
The pavilion also showed electrical
conductivity, which provides another
indication of the boron doping.
By examining the pavilion of the
diamond at high magnification, we
could assess the condition of the CVD
layer after eight years of daily wear.
The film was still largely intact, covering the entire pavilion and containing only a few sporadic pinholes.
There was no noticeable degradation
at the facet junctions. The obvious
color zoning in figure 4 that appears
to divide the diamond into quadrants
Figure 4. This 0.27 ct diamond
with a boron-doped CVD synthetic diamond film on the pavilion received a color grade of
Fancy Dark brownish yellowish
gray. The distinct color zoning
seen in this photo makes it apparent that the diamond has been
treated in some way and is not a
natural stone.
GEMS & GEMOLOGY
Figure 5. This differential interference contrast image shows the irregular texture of the CVD overgrowth
film on the pavilion facets of the
0.27 ct natural diamond. Each facet
is a different color, and the boundaries between colors are facet junctions. Field of view 246 microns.
is likely due to differences in deposition and boron incorporation rates on
the various facets. Recently, the diamond was annealed twice, at 600C
and 700C for one hour each in a reducing atmosphere, in an effort to
make it appear bluer. This treatment
changed the color from Fancy Dark
brownish yellowish gray to Fancy
Dark gray.
As treatments become more sophisticated and move from the laboratory into the trade, the need for
accurate detection only intensifies.
Nevertheless, such semipermanent
films are identifiable and reiterate the
need to use both spectroscopic and
traditional methods for gemological
characterization.
Sally Eaton-Magaa
Star OPAL
Opal is best known for displaying
play-of-color. It may also display asterism, though star opal has only been
reported from Idaho, and a perfect sixrayed star is exceptionally rare (J.V.
Sanders, The structure of star opals,
Acta Crystallographica, Vol. A32,
1976, pp. 334338). The Bangkok laboratory recently had the opportunity
to examine a transparent star opal (figure 6). The 2.39 ct light brownish yellow cabochon displayed a distinct
six-rayed star.
Standard gemological testing gave
a spot RI reading of 1.43 and a hydro-
SUMMER 2014
Figure 6. This 2.39 ct transparent

opal with a light brownish yellow bodycolor possessed a sixrayed star.
Figure 7. This photomicrograph

of the opal shows the intersection of large parallel planes with
play-of-color, producing asterism. Magnified 15.
static SG of 2.10. When exposed to ultraviolet radiation, the stone fluoresced strong bluish white under
long-wave and weak bluish white
under short-wave UV. It phosphoresced green after exposure to longwave UV. Advanced gemological
testing by energy-dispersive X-ray fluorescence (EDXRF) confirmed a silicarich material with some additional
trace elements, including aluminum
and iron. All of these properties were
consistent with opal.
Microscopic examination revealed
large parallel planes with play-of-color
intersecting to form a hexagonal pattern (figure 7). This is responsible for
producing the six-rayed star. This unusual stone serves as a reminder that,
unlike asterism in other gem materials, the star in opal is caused by diffraction of light from faults or
imperfections in the packing arrangement of silica spheres.
Wasura Soonthorntantikul
178). The shell beads are often coated

with artificial materials to simulate a
wide variety of natural and cultured
pearls in the market. Recent submissions of such imitations to GIA
prompted researchers in the lab to obtain several samples from a commercial website. Labeled as shell pearls
(figure 8), they resembled Tahitian
black, white, and yellow/golden South
Sea cultured pearls.
Although the samples were similar in appearance and heft to cultured
pearls, routine gemological tests revealed unnatural surface characteristics. The black shell pearl necklace
also exhibited multicolored but rather
oily orient. Magnification showed numerous minute particles with a glittering effect (figure 9, top), as well as
a lack of the obvious overlapping
nacre platelets commonly found in
nacreous pearls. These features suggested that an artificial coating had
been applied to the bead. Inspection
Shell PEARL as a Pearl Imitation

Imitation pearls made of shell, or shell
pearls, have a long history in the jewelry market and have been reported in
previous Lab Notes columns (Fall 1984,
p. 170; Winter 1986, p. 239; Summer
2001, pp. 135136; Summer 2004, p.
Figure 9. Top: The surface of the

imitation pearls contained a glittery coating rather than overlapping nacre platelet structures
(magnified 112.5). Bottom: White
bead-like shell material was exposed at a damaged area near the
drill hole (magnified 10).
Figure 8. These shell pearls range from 10 to 12 mm and display black,

white, and golden colors.
LAB NOTES
GEMS & GEMOLOGY
SUMMER 2014
153
near the drill holes of some samples

revealed chipping and peeling of this
very thin coating, exposing white
bead-like shell material underneath
(figure 9, bottom).
X-radiography showed only the
inner bead, with occasional parallel
banding and crackssimilar to the
banding and trematode tunnels
common in some saltwater shell
beadswhile the outer coating remained transparent (figure 10).
EDXRF analysis detected a high concentration of bismuth, which usually
does not occur in pearls but had been
reported previously in coated natural
and cultured pearls (Fall 2005 Gem
News International, pp. 272273;
Winter 2011 Lab Notes, pp. 313314).
Finally, Raman spectroscopic analysis
of the exposed white bead identified it
as aragonite. These results, coupled
with the beads inert reaction under
X-ray fluorescence, confirmed it was
made of saltwater shell.
We have noted the use of the
terms shell pearl or just pearl for such
imitations in product descriptions on
numerous commercial websites. Such
inappropriate nomenclature could be
Figure 10. X-radiography revealed

banding and cracks in the inner
shell beads.
misleading to inexperienced buyers.

According to the CIBJO Blue Book on
pearls, Imitations or simulants of
natural pearls and cultured pearls . . .
shall be immediately preceded by the
word imitation or simulated, with
equal emphasis and prominence . . . as
those of the name itself. Imitation
pearl would be the proper name of
this material, and consumers need to
be aware of its true identity.
Jessie (Yixin) Zhou and
Chunhui Zhou
Yellow CVD SYNTHETIC

DIAMOND
Gem-quality CVD-grown synthetic diamonds are usually type IIa and occasionally type IaA. In addition, a faceted
type Ib yellow CVD synthetic of 3 mm
in diameter was reported more than a
decade ago (J.E. Butler, Chemical
vapor deposited diamond: Maturity
and diversity, Interface, Spring 2003,
pp. 2226). To the best of our knowledge, no other type Ib examples have
been documented in the gem industry
since then. The New York laboratory
Figure 11. This 0.40 ct square-shaped

type Ib CVD synthetic diamond was
color graded as Fancy yellow.
recently examined a gem-quality type

Ib CVD-grown synthetic diamond.
This Fancy yellow square weighed
0.40 ct and measured 4.20 4.12 2.55
mm (figure 11).
As is typical for CVD synthetics,
the strong silicon-vacancy ([Si-V]) center was detected as a doublet at 736.5
and 736.9 nm in the 633 nm PL spectrum. The mid-IR spectrum showed a
single substitutional nitrogen atom
(N0s ) at 1130 and 1344 cm1 (figure 12).
The nitrogen content, which contributed to the yellow color, was calculated from absorption bands in the
Figure 12. The yellow CVD synthetics mid-IR spectrum revealed single
substitutional nitrogen atoms (Nso) at 1130 and 1344 cm1. The diamond
Raman band was detected at 1331 cm1. A hydrogen-related defect was
detected at 3107 cm1, and C-H defects could be observed at 2948, 2920,
2905, 2873, and 2849 cm1. An unknown doublet was also observed at
3030 and 3032 cm1.
MID-IR ABSORPTION SPECTRUM
Matrix
3107
3030
3032
2920
ABSORBANCE (a.u.)
2948
Vein
2905
2873
4500
4000
3500
1344
1331
1130
2849
3000
2500
2000
1500
1000
WAVENUMBER (cm1)
154
LAB NOTES
GEMS & GEMOLOGY
SUMMER 2014
Figure 13. DiamondView imaging revealed curved green lines, reflecting changes in growth conditions (left). Pronounced growth events were also observed in green and bluish green regions near the culet (middle). Microscopic
examination revealed a graphitized stress halo (right, magnified 112.5), along with small feathers on the girdle
(not shown).
10001400 cm1 mid-IR spectral region

as 4.5 ppm. Nitrogen can be unintentionally introduced during the growth
process; it can also be added deliberately to control the growth rate of single-crystal CVD synthetic diamond.
DiamondView images showed curved
green lines, indicating changes in conditions during growth (figure 13, left).
The DiamondView also revealed green
and greenish blue regions of pronounced growth events near the culet,
as in the middle image of figure 13. A
graphitized stress halo (figure 13, right)
and small feathers on the girdle were
observed under the microscope. Typical CVD straina localized tatamilike structure with low-strain
interference colorswas visible under
cross-polarized light.
Post-growth treatment was detected: a hydrogen-related defect at
3107 cm1 and C-H defects at 2948,
2920, 2905, 2873, and 2849 cm1
(again, see figure 12). We did not observe the neutral charge state of the
nitrogen vacancyhydrogen complex
([N-V-H]0) at 3123 cm1. These results
suggested that the sample was HPHTtreated. An unknown doublet at
30303032 cm1 was also found.
Strong N-V (nitrogen-vacancy) defects
at 575 (NV0) and 637 (NV) nm were
detected in PL spectrum using a 514
nm laser. The negatively charged nitrogen vacancyhydrogen complex
([N-V-H]) is usually contained in asgrown nitrogen-doped single-crystal
CVD synthetic diamond, but it is unstable. After HPHT treatment, the hydrogen atoms were dissociated,
LAB NOTES
causing C-H and N-V defects to form.

Additional defects, such as H3 and
H2, were also detected. Both centers
could be introduced in as-grown CVD
synthetic diamonds by HPHT annealing. The H3 centers were responsible for the observed green
dislocations in the DiamondView
image. As-grown CVD synthetic diamonds usually possess an unattractive brown color component. This
can be removed by HPHT treatment
to achieve an attractive yellow color,
as observed in this sample.
Type Ib CVD synthetics are also of
interest for technological applications, especially for quantum computing (see E. Gibney, Flawed to
perfection: Ultra-pure synthetic diamonds offer advances in fields from
quantum computing to cancer diagnostics, Nature, Vol. 505, 2014, pp.
472474). The N-V electron spin
could be used as a basic unit of quantum computing, known as a quantum
bit or qubit. Qubits are not limited
to the binary 0 or 1 states of bits
in traditional computers, but can simultaneously exist in both binary
states and any state in betweena
difficult concept to grasp but one that
will lead to vastly enhanced computational capacity. Type Ib yellow CVD
diamonds of micrometer size were
grown and subsequently electron
beamirradiated and annealed by P.
Neumann to create NV centers, as
noted in his 2012 PhD thesis, Towards a room temperature solid state
quantum processor: The nitrogen-vacancy center in diamond. After irra-
GEMS & GEMOLOGY
diation, the color changed from the

as-grown yellow to green, then to purplish pink after annealing.
Kyaw Soe Moe, Wuyi Wang, and
Ulrika DHaenens-Johansson
Flux-Grown SYNTHETIC RUBY

with Hydrothermal Synthetic Seed
Crystal
The Carlsbad laboratory recently received a 1.18 ct transparent red octagonal step-cut stone for ruby report
service. Standard gemological testing
established the following properties:
RI1.762 to 1.770; birefringence
0.008; optic signuniaxial negative;
pleochroismorangy red to purplish
red; specific gravity4.01; fluorescence reactionstrong red to longwave, weak red to short-wave UV
radiation. Examination with a deskmodel spectroscope revealed a typical
ruby spectrum. All of these properties
were consistent with natural or synthetic ruby.
Under magnification, the most distinctive internal characteristic in the
crown was the presence of strong irregular growth features: zigzag- or mosaiclike striated patterns (figure 14), typical
of a hydrothermal synthetic. Other
areas of the ruby lacking these irregular growth features contained hexagonal metallic platelets and high-relief,
whitish flux inclusions (figure 15), typical of a flux-grown synthetic. Flux and
hydrothermal inclusions have not
been previously documented in the
same specimen.
SUMMER 2014
155
Figure 14. The zigzag-like growth

structure observed in this 1.18 ct
synthetic ruby is characteristic of
hydrothermal growth. Field of
view 1.42 mm.
Laser ablationinductively coupled
plasmamass spectrometry (LA-ICPMS) analysis revealed traces of Ca, Ti,
Cr, Fe, Mo, Rh, Sn, W, and Pt. The low
amount of Fe and Ti, the absence of V
and Ga, and the presence of Pt were
consistent with flux-grown corundum.
Both natural and flame-fusion
synthetic ruby have been used as seed
crystals in the flux growth of ruby (J.I.
Koivula, Induced fingerprints, Winter 1983 G&G, pp. 220227; Summer
1991 Lab Notes, p. 112). The seed
crystals are generally removed during
the cutting process but may, on rare
occasions, be detected in finished
specimens. Upon close microscopic
examination, we noted that several of
Figure 15. When the synthetic

ruby was examined under diffused fiber-optic lighting and
darkfield illumination, hexagonal
platinum platelets and trapped
flux residue became apparent.
Field of view 1.42 mm.
156
LAB NOTES
the flux-filled healed fractures (wispy

veils) extended into the areas showing hydrothermal graining. These
observations led us to conclude that
the hydrothermal material was a seed
crystal and that the flux healing was
a secondary process to the hydrothermal growth. There was an irregular
separation between the materials
under brightfield illumination.
This unusual combination of a hydrothermal ruby seed with flux ruby
overgrowth is the first of its kind examined by GIA.
Ziyin Sun and Dino DeGhionno
Rare Faceted WURTZITE

Recently the Carlsbad laboratory examined a 3.97 ct transparent brownish
red pear-shaped modified brilliant for
identification service (figure 16). Standard gemological testing revealed a refractive index that was over the limit
of the RI liquid and a specific gravity
(obtained hydrostatically) of 3.96.
There was no fluorescence observed
with exposure to long- and short-wave
UV light. The stone also displayed an
adamantine luster and a uniaxial optic
figure when examined with polarized
light. Microscopic examination with a
fiber-optic light source showed fine
needles and a large reflective decrepitation halo surrounding a negative
crystal. Examination also revealed
that the decrepitation halo was perpendicular to the optic-axis direction,
suggesting the stone had basal cleavage (figure 17).
Figure 16. This 3.97 ct faceted

wurtzite has an adamantine luster and brownish red bodycolor.
GEMS & GEMOLOGY
Figure 17. A decrepitation halo

surrounds a negative crystal in
the faceted wurtzite. Field of
view 2.15 mm.
Raman spectroscopy and X-ray

powder diffraction conclusively identified the stone as wurtzite. EDXRF
analysis detected zinc and sulfur, which
further supported this identification.
Wurtzite, a polymorph of sphalerite, commonly occurs in hydrothermal vein deposits associated
with barite and sphalerite. Facet-grade
wurtzite has been reported to occur in
Merelani, Tanzania (Winter 2013
GNI, p. 261). This is the first wurtzite
examined by the GIA laboratory.
Amy Cooper and Nathan Renfro
Tenebrescent ZIRCON
The Bangkok laboratory recently received for identification a 13.16 ct reddish orange stone, submitted as
tenebrescent zircon (figure 18, left).
Standard gemological testing indicated
typical properties for the stable form of
zircon: RI was over the detection limit
of the standard refractometer, SG was
approximately 4.50, and the sample
was inert to both long- and short-wave
UV. After exposure to short-wave UV
for approximately 30 minutes, however, its tone became darker. Chemical
analysis using EDXRF showed significant amounts of zirconium and silicon,
as well as a trace amount of hafnium.
Tenebrescence, the phenomenon of
reversible photochromism, is common
in gemstones such as hackmanite (a variety of sodalite) but unusual in zircon.
In September 2011, GIA examined two
zircons from central Australia that
SUMMER 2014
Figure 18. This 13.16 ct zircon is shown before (left) and just after 30 minutes of short-wave UV exposure (center).
The SWUV lighting darkened the stone, while darkness restored the orange color. The image on the right shows
the color of the stone after it was left in the dark for three weeks.
minished, and the stone appeared
brown (figure 18, center). There was no
further change after 30 minutes.
The UV-visible spectrum after
SWUV exposure showed a significant
increase in absorption in the 590750
nm range, reducing transmission of
the orange to red color and making
the stone darker (figure 19).
As noted above, the reaction is re-
turned orange in darkness and faded to

near colorless when exposed to light
(S.F. McClure, Tenebrescent zircon,
Winter 2011 G&G, pp. 314315). This
zircon displayed a different phenomenon: Its reddish orange color did not
fade under fluorescent lighting or fiberoptic light. Exposure to short-wave UV
(SWUV) caused it to darken. The reddish orange component gradually di-
Figure 19. After 30 minutes of SWUV exposure, the zircons absorption

spectrum revealed more absorption in the 590750 nm range, which relates to an orange to red color.
UV-VISIBLE SPECTRA
3.0
2.5
Before SWUV exposure
versible. The reddish orange color returned when the stone was stored in
the dark at room temperature (see figure 18, right). The color change was
obvious in the first 10 to 20 hours.
After that, no further color change was
observed.
To determine whether the process
was repeatable, the stone was again
subjected to SWUV. Once again the
reddish orange color darkened in tone.
In summary, the reddish orange
zircon was darkened by short-wave
UV exposure, and the color returned
when the stone was stored in the
dark. The darkening reaction was
much faster than the recovery. Further studies incorporating other conditions that might affect the rate of
color modification in tenebrescent
zircon are being conducted.
Ratima Suthiyuth
After SWUV exposure

2.0
1.5
653
690
750
1.0
0.5
0
400
500
600
700
800
900
1000
1100
1200
PHOTO CREDITS:
Robison McMurtry1, 4; Troy Ardon3;
Lhapsin Nillapat6; Charuwan Khowpong7; Jian Xin (Jae) Liao8; Chunhui
Zhou9, top; Jessie (Yixin) Zhou9, bottom; Nathan Renfro5, 14, 15, and 17;
Sood Oil (Judy) Chia11; Kyaw Soe
Moe13; Don Mengason16; Sasithorn
Engniwat18.
WAVELENGTH (nm)
LAB NOTES
GEMS & GEMOLOGY
SUMMER 2014
157
Contributing Editors
Emmanuel Fritsch, CNRS, Team 6502, Institut des Matriaux Jean Rouxel (IMN), University of Nantes,
France ([email protected])
Kenneth Scarratt, GIA, Bangkok ([email protected])
COLORED STONES AND ORGANIC MATERIALS

Jadeite bangle with the appearance of polymer treatment.
Jadeite jade always comes under close scrutiny in the market
because it is so frequently treated. The treatment options
for jadeite jade are numerous: bleaching, dyeing, polymer
impregnation, or a combination of these. Polymer impregnation is a particular favorite, a process that is not always
easy to observe by magnification. A 244.29 ct translucent
light to deep green and brownish yellow bangle measuring
approximately 70 14 mm was recently submitted to the
Lai Tai-An Gem Lab in Taipei. The client was concerned
about a more transparent light-colored area (figure 1) that
appeared to be possible evidence of impregnation.
Standard gemological testing indicated a spot RI reading of 1.54 at the more transparent light-colored area and
1.66 elsewhere on the bangle, an SG of approximately 3.32,
and a characteristic jadeite spectrum exhibiting a band at
Figure 1. This 244.29 ct translucent light to deep green

and brownish yellow bangle measures approximately
70 14 mm. It contained a transparent, light-colored
area that suggested the possibility of a polymer impregnation process. Photo courtesy of Lai Tai-An Gem Lab.
Figure 2. Microscopic examination of the suspect area

revealed natural inclusions and a crystalline form.
Neither would be found in an impregnated specimen.
Photo courtesy of Lai Tai-An Gem Lab.
about 437 nm. FTIR and Raman spectroscopy were also applied, producing spectral features that confirmed the bangles identity as jadeite jade. Microscopic examination
revealed natural inclusions within the transparent area
that appeared to be polycrystalline (figure 2). Neither
would be seen in an impregnated specimen, ruling out this
treatment. Further analysis with Raman microscopy was
conducted to identify its true nature.
Sharp absorption peaks at around 207, 347, 400, 465,
and 990 cm1 were a match for nepheline (figure 3), a min-
Editors note: Interested contributors should send information and illustrations to Justin Hunter at [email protected] or GIA, The Robert
Mouawad Campus, 5345 Armada Drive, Carlsbad, CA 92008.
GEMS & GEMOLOGY, VOL. 50, NO. 2, pp. 158169,
158
GEM NEWS INTERNATIONAL
GEMS & GEMOLOGY
SUMMER 2014
INTENSITY (a.u.)
RAMAN SPECTRUM
2095
1885
1675
1465
1255
1045
835
625
415
205
RAMAN SHIFT (cm-1)
Figure 3. Sharp absorption peaks around 207, 347,

400, 465, and 990 cm1 were consistent with
nepheline, a mineral sometimes associated with albite in jadeite.
eral sometimes associated with albite in jadeite. The

nepheline crystal remained intact during the fashioning
process, and its presence led to confusion over whether the
bangle had been treated.
Larry Tai-An Lai ([email protected])
Lai Tai-An Gem Laboratory, Taipei
Prasiolite with inclusion influenced by Brazil-law twinning. The Indian Gemological Institute Gem Testing Laboratory recently examined a 6.43 ct transparent grayish
yellowish green oval mixed-cut specimen. Standard gemological examination gave an RI of 1.5451.553 with a uniaxial positive optic sign and a birefringence of 0.008, and a
hydrostatic SG of 2.65. These readings readily identified
the stone as prasiolite, a green variety of quartz.
The prasiolite displayed prominent Brazil-law twinning, with very sharp twinning planes when the sample
was observed parallel to the optic axis between crossed polarizing filters, though only two sections of twinning were
clearly visible in the stone (figure 4). The presence of such
sharp and prominent Brazil-law twinning confirmed the
samples natural origin. When synthetics do exhibit such
patterns, it has a flame-like or irregular shape (J.I. Koivula
and E. Fritsch, The growth of Brazil-twinned synthetic
quartz and the potential for synthetic amethyst twinned
on the Brazil law, Fall 1989 G&G, pp. 159164).
Magnification showed no inclusions, but illumination
from the side using a fiber-optic light source revealed
minute white particles that scattered the light (figure 5).
These pinpoint inclusions followed the pattern of the
Brazil-law twinning, which was apparent without the aid
of polarized filters (E.J. Gbelin and J.I. Koivula, Photoatlas
Figure 4. Between crossed polarizers, the prasiolite

showed two sections of Brazil-law twinning in the
optic-axis direction. Photo by Meenakshi Chauhan,
magnified 15.
of Inclusions in Gemstones Volume 2, Opinio Verlag,

Basel, Switzerland, 2005, p. 573).
Minor rhombohedron faces that did not feature Brazillaw twinning were filled with these minute light-scattering
particles. These inclusions were visible on the faces with
some tilting of the fiber-optic light source but did not show
any pattern (figure 6). The minute particles were not visible
in both zones simultaneously.
In diffused transmitted light, the minor rhombohedron
faces displayed a slightly darker shade of the grayish yellowish green color seen in prasiolite. These faces also
showed a wavy internal growth structure between crossed
polarizers. The major rhombohedron faces were even
lighter in color. No color bands were observed in the stone.
Although the owner claimed the stones color was natural, the presence of Brazil-law twinning raised suspicion,
since this twinning has only been associated with the
Figure 5. Fine dotted inclusions demonstrated a

Brazil-law twinning pattern along the optic axis, exactly where this twinning was visible under polarized
filters. Photo by Meenakshi Chauhan, magnified 10.
GEMS & GEMOLOGY
SUMMER 2014
159
Figure 7. The largest known natural moissanite crystal, found along the Kishon River in northeastern Israel, measures 4.1 mm in its longest dimension.
Photo by Vered Toledo.
Figure 6. Fine dotted inclusions were visible in this
minor rhombohedral face where Brazil-law twinning
was absent. Photo by Meenakshi Chauhan, magnified 10.
amethyst variety of quartz. There is still no way to separate

naturally heated from heat-treated prasiolite.
Meenakshi Chauhan
Indian Gemological Institute
Gem Testing Laboratory, GJEPC, New Delhi
Record-size natural moissanite crystals discovered in Israel.
Synthetic silicon carbide (SiC) is well known in the gem
and jewelry market as synthetic moissanite. Natural SiC
is very rare and has been found as tiny crystals (usually less
than 1.5 mm) in only a few deposits worldwide (Y. Bauer
et al., Moissanite from middle mountains of Czechoslovakia, International Geology Review, Vol. 7, No. 7, 1965,
pp. 11941206; A.A. Shiryaevet et al., Moissanite (SiC)
from kimberlites: Polytypes, trace elements, inclusions
and speculations on origin, Lithos, Vol. 122, No. 3-4,
2011, pp. 152164, and references therein).
Over the past 14 years, however, more than 2,500 crystals of natural moissanite have been discovered by Shefa
Yamim, an Israeli exploration and mining company. The
crystals come from primary and alluvial deposits along the
Kishon River, near Haifa in northern Israel. Shefa Yamim
began prospecting in this area following the 1988 prophetic
statement of the Lubavitcher Rebbe that precious stones
and gems would be discovered in the valley next to Haifa
(see video at http://youtu.be/_uzY5mV8rdY). In 2000, the
company unearthed moissanite crystals ranging from 0.1
to 1 mm in size; two years later, 2.2 mm crystals were
found. In 2009, a 3.5 mm moissanite discovery set a world
record, only to be surpassed in August 2012 by a crystal
measuring 4.1 mmthe largest specimen to date (figure 7
and http://goo.gl/JZFiwe).
There is evidence of even larger crystals, as many broken crystals were discovered. Some crystals have also been
found in situ in the volcanic rock of the Rakefet magmatic
160
complex, one of the magmatic bodies of Mount Carmel,

just southeast of Haifa, which is drained by a small tributary to the Kishon River.
We studied 30 Israeli crystals ranging from less than 1
mm to 3.5 mm. All were identified as moissanite (SiC 6H)
by Raman spectroscopy. Their morphology was hexagonal,
Figure 8. These natural moissanite crystals are hexagonal to pyramidal, showing rounded areas that may
result from either growth or dissolution. The crystals
range up to 2.5 mm (not to scale). Composite photo
by Aurlien Delaunay.
GEMS & GEMOLOGY
SUMMER 2014
Figure 10. This 1.36 ct light bluish violet sapphire displayed an interesting color shift when exposed to the
UV radiation of the DiamondView. Photo by Igor
Iemelianov; magnified 15.
Figure 9. The colors seen in moissanite from the

Mount Carmel area of northern Israel range from
dark blue to light green. Note the broken or rounded
morphology. Composite photo by Aurlien Delaunay.
bipyramidal to platy, with the pinacoid generally present

(figures 8 and 9). The crystals were often broken, particularly the larger ones with rounded, shiny surfaces (again,
see figure 7). As shown in figure 9, the crystals were transparent and ranged from deep blue (the most common color)
to light green. Except for size, these characteristics are similar to those of other natural moissanites documented in
the literature. The crystals are often associated with small
bits of creamy white or red complex matrix.
The crystals we examined were inert under long- and
short-wave UV radiation, with the exception of five small
green to light green samples. In these five specimens,
medium to strong orange luminescence was seen under either long- or short-wave UV. These were found to be magnetic when placed next to a powerful permanent alnico
magnet. All of the samples contained small metallic inclusions, the largest approximately 30 microns in size, which
could be the source of magnetism. We will continue to follow moissanite exploration in this area, and a more in-depth
research paper is forthcoming.
Unusual optical effect in blue sapphire. A 1.36 ct light

bluish violet sapphire (figure 10) was presented to the laboratory of the State Gemological Centre of Ukraine for examination. The stone exhibited an RI of 1.7621.770, with
a birefringence of 0.008, and a hydrostatic SG of 3.97. The
samples displayed typical pleochroism and a very weak
pinkish orange fluorescence to long-wave UV. Qualitative
analysis using EDXRF spectroscopy showed major amounts
of Fe, Cr, and Mg, and minor amounts of Ti and Ga. Microscopic examination revealed dot-like dissolved rutile, proof
of heat treatment. The sapphire showed a small color shift
from bluish violet in the daylight to pale blue in incandescent light. A band with pink luminescence (apparently
due to Cr 3+ impurity) was identified using the DiamondView (figure 11). When removed from the DiamondView,
Figure 11. The DiamondView recorded this band with

pink luminescence in the light bluish violet sapphire.
Image by Igor Iemelianov.
Emmanuel Fritsch
Vered Toledo
Shefa Yamim
Akko, Israel
Antoinette Matlins
Antoinette Matlins, LLC
Woodstock, Vermont
GEMS & GEMOLOGY
SUMMER 2014
161
Figure 12. When removed from the

DiamondView, the sapphire had completely
turned brown (left).
After about 12 hours, the
stone reverted to its original light bluish violet
color (right). Photos by
Igor Iemelianov, magnified 15.
Assemblage of synthetic ruby in calcite matrix. Recently,

the Dubai Central Laboratory received a white and cream-
colored rough sample for identification (figure 13). The 124.4

107.7 65.8 mm, 5917.35 ct specimen had two openings,
the larger one about 25 mm long and 10 mm wide. A hexagonal red stone embedded in the matrix could be seen
through these openings (figure 14). Only the prism side was
visible, and the rough concealed both terminations.
The red stones refractive index and specific gravity
were impossible to determine, and only a very small portion was visible for obtaining other information. A spectroscope was the only option, and the stone clearly showed
a ruby spectrum. It also displayed strong and medium red
reactions to long- and short-wave UV radiation, respectively. Microscopic examination with fiber-optic light revealed a group of round gas bubbles and tiny fractures, but
curved striae were not seen. These properties indicated a
flame-fusion synthetic ruby cleverly embedded in matrix
to imitate natural ruby.
Raman spectroscopy identified the matrix as calcite.
Close examination indicated that the cream-colored side
of the stone had been drilled to make a hole for inserting
the synthetic ruby. After the drilled area was sealed with a
mixture of glue and calcite powder, two openings were cut
Figure 13. This synthetic ruby was fixed in a white

and cream-colored matrix and represented as natural
ruby. Photo by Sutas Singbamroong, Dubai Central
Laboratory.
Figure 14. The synthetic ruby was embedded in a calcite matrix, seen here under incandescent light. Photo
by Nazar Ahmed, Dubai Central Laboratory.
162
GEMS & GEMOLOGY
the stone had completely changed color to brown (figure

12, left). After about 15 minutes in natural daylight, the
light bluish violet hue began to return (figure 12, right).
About 12 hours later, its original blue-violet color was restored. Such color instability is extremely unusual in blue
sapphires.
It should be noted that the UV lamp in the DiamondView has a wavelength less than 225 nm (according to the
users manual). The energy of UV radiation is considered
too weak for the color treatment of corundum (see K. Nassau, Gemstone Enhancement, Butterworth-Heinemann,
1984, p. 221). In this sapphire, the DiamondView apparently caused the color change. Such a reversible change of
color with exposure to light is called photochromism.
Iurii Gaievskyi ([email protected]),
Igor Iemelianov, and Elena Belichenko
State Gemological Centre of Ukraine, Kiev
SYNTHETICS AND SIMULANTS
SUMMER 2014
Figure 15. Under long-wave UV radiation, the drilled

area showed strong greenish blue fluorescence. Photo
by Nazar Ahmed, Dubai Central Laboratory.
to reveal the synthetic ruby inside. Under long-wave UV

radiation, the drilled area showed strong greenish blue fluorescence, while the rest of the stone was inert or exhibited
a light greenish blue reaction (figure 15).
These types of artificial assemblages should alert
traders to exercise caution when purchasing rough stones,
especially in newer, less established markets.
Nazar Ahmed ([email protected]) and
Sutas Singbamroong
Gemstone Unit, Dubai Central Laboratory
United Arab Emirates
Composite coral veneer glued to artificial matrix. Red
coral, one of the most precious organic gems, has become
increasingly popular in the Chinese jewelry market. With
natural coral resources being depleted, ornaments of large
size and high quality are becoming even more valuable. As
a result, the market for treated red coral and imitation
products is rapidly growing.
Composite red coral normally consists of coral fragments glued together with cement compounds. In 2010,
Figure 16. These 14 ornaments, represented as red

coral products, consisted of a coral veneer glued to a
matrix. The ten bangles (top) weigh 37.3457.43 g,
and the four snuff bottles (bottom) range from 121.95
to 145.07 g. Photo by Haibo Li.
the National Gemstone Testing Center Gem Laboratory

in Beijing tested one bangle that was represented as red
coral (Summer 2010 GNI, pp. 158159). Because the client
did not allow the lab to take a powder sample, further testing was restricted. There was some concern about whether
the bangle had an artificial matrix, but this could not be
determined without destructive tests.
Recently, the NGTC received 10 bangles and four snuff
bottles (figure 16) presented as red coral. The bangles resembled the one tested in 2010. Each of the 14 pieces was
covered with a very thin surface layer of red coral. The
composition of the material below this coral veneer was
analyzed.
Magnification revealed the boundaries between the
coral fragments and the cement compounds, which tended
to be recessed (figures 17 and 18). This clearly demonstrated that the veneer was composite red coral. The coral
Figure 17. Left: A closeup of a bangle with

coral veneer. Right: The
bangles inner surface is
covered with rectangular coral slices, but
gaps filled with cement
compounds are still visible. Photos by Haibo
Li; magnified 6.5.
GEMS & GEMOLOGY
SUMMER 2014
163
Coral
Coral
Cement
Cement
Coral
Cement and Bubbles
Cement
Coral
Artificial matrix material

Coral
Figure 18. A: Reflected light showed the difference in luster between the cement compounds and the coral fragments; magnified 10. B: Under transmitted light, air bubbles can be found in the cement compounds; magnified
20. C: The boundary between the matrix and the composite coral veneer is very clear and sharp; magnified 10.
D: The fluorescence reaction of the same sample under short-wave UV. Photomicrographs by Haibo Li.
fragments had a vitreous luster and showed wavy growth

patterns typical of natural coral. No growth patterns were
seen in the cement compounds except for air bubbles.
Whitish material seen inside the neck of the bottles was
identified as an artificial matrix. The composite red coral veneer, which was only about 12 mm thick, had been glued
to the surface of the whitish matrix. Under a short-wave UV
lamp, strong blue-white fluorescence from the cement compounds and the matrix was visible (figure 18).
Micro-infrared spectrometry (with 4.0 cm1 resolution
and 64 scans in the 6754000 cm1 range) identified the cement compounds as epoxy resin (figure 19A). Two different
kinds of artificial matrix material were detected: ethyl
cyanoacrylate (figure 19B) and alkyd synthetic resin mixed
with carbonates. The main component of the carbonates
was established as calcite by subtracting the spectra (figure
19C). The calcite could have come from the manufacturing
process of the composite red coral.
The SG of the synthetic resin was about 12 g/cm3,
much lower than that of red coral (2.65 g/cm3). Carbonates
could have been added to the matrix to make it heavier and
approximate the heft of coral. The snuff bottles all had similar size but different weights. The three that contained
alkyd resin mixed with calcite, which has a relatively high
SG, weighed 140145 g. The matrix in the fourth snuff bottle was ethyl cyanoacrylate, which has a lower SG than
calcite. As a result, it only weighed 121 g.
In todays market, oversized red coral products are often
constructed from an artificial matrix. After the matrix is
shaped, a veneer of red coral fragments is glued to its surface. This makes it easier to manufacture larger ornaments
while preserving natural coral reefs. Combining magnification and UV fluorescence observation with infrared spectrometry, we can identify composite coral assemblages
quickly and effectively.
164
GEMS & GEMOLOGY
Haibo Li ([email protected]), Zhoujing Yue, Jie Liang,

Taijin Lu, Jun Zhang, and Jun Zhou
National Gemstone Testing Center Gem Laboratory,
Beijing
Dyed bone as a red coral imitation. Numerous materials
have been widely used to imitate red coral: glass, plastic,
SUMMER 2014
ceramics, composites, and dyed shells. Recently, the Gem

Testing Laboratory in Jaipur examined an 8.86 ct orangered cabochon measuring 16.56 12.19 7.34 mm (figure
Figure 19. A: The IR spectrum of the cement compound in one of the bangles identified it as epoxy
resin. B: The IR spectrum of the artificial matrix material in one of the snuff bottles indicated ethyl
cyanoacrylate. C: The artificial matrix material in the
other three snuff bottles was identified, by subtracting IR spectra, as alkyd resin mixed with carbonates.
IR SPECTRA
Figure 20. This 8.86 ct orange-red cabochon, submitted as coral, was identified as dyed bone. Photo by
Gagan Choudhary.
0.60
ABSORBANCE
0.50
0.40
0.30
0.20
0.10
0.00
4000
3000
2000
1000
2.60
2.40
2.20
2.00
ABSORBANCE
1.80
1.60
1.40
1.20
1.00
Figure 21. The dyed bone illustrated in figure 20 displayed subparallel veins oriented along the length of
the cabochon; also note the orange-red color concentrations against the samples whitish bodycolor. Photomicrograph by Gagan Choudhary; magnified 48.
0.80
0.60
0.40
0.20
0.00
4000
ABSORBANCE
20) that was presented as red coral but proved to be dyed

bone, an unusual imitation.
Initial observation showed a strong resemblance to red
coral, but careful microscopic observation proved otherwise. The sample lacked the typical tree ring or striated
growth structure seen in corals, though it did display a network of fine veins throughout. Viewed from above, these
veins appeared subparallel and were oriented along the
length of the cabochon (figure 21). The veins also displayed
orange-red color concentrations against the cabochons
whitish bodycolor (again, see figure 21), suggesting that the
1.00
3000
2000
1000
Artificial matrix material
0.00
1.00
Alkyd resin
0.00
1.00
Subtraction: Carbonate material
0.00
4000
3000
2000
1000
-1
WAVENUMBERS (cm )
GEMS & GEMOLOGY
SUMMER 2014
165
a strong peak at approximately 960 cm1, with weaker peaks

at approximately 430, 586, and 1072 cm1. Also present
were a few indistinct features in the 12001700 cm1 region.
The peaks at approximately 960, 586, and 430 cm1 were attributed to PO43 bending and stretching, and the one at 1072
cm1 to CO43 stretching. The indistinct features between
1200 and 1700 cm1 were due to collagenthe essential
component of bone tissue (C. Kontoyannis et al., Analysis
of bone composition with Raman spectroscopy,Proceedings of the 13th Panhellenic Pharmaceutical Congress,
2012, http://nemertes.lis.upatras.gr/jspui/bitstream/10889/
5187/1/PMO031.pdf).
Bones have been used as jewelry since ancient times,
and some specimens are dyed or stained to make them appear older (Summer 2006 Lab Notes, p. 160). Bone is used
to imitate ivory, but this is the first time we have seen a
bone dyed red and presented as coral, an interesting and
unusual imitation. Although the market penetration of
this material is unknown, we cannot rule out the possibility of it being mixed in packets of red coral, and this separation would pose a challenge for jewelers and gemologists.
Gagan Choudhary ([email protected])
Gem Testing Laboratory, Jaipur, India
MORE FROM TUCSON 2014
material was both dyed and organic. Detailed observations

of the sample from all sides revealed that the subparallel
veins were confined to circular concentric planes oriented
along the length of the cabochon (figure 22, top); the individual veins were further surrounded by concentric
whitish rings (figure 22, bottom). These structural features
ruled out the possibility of coral, but the identity had yet
to be established.
The samples spot RI was approximately 1.56, without
any noticeable birefringence blink, while its hydrostatic SG
was 2.01. Under a UV lamp, it displayed pinkish orange
patchy fluorescence (which was stronger under long-wave
UV). This characteristic is typically associated with dyes.
Qualitative EDXRF analysis revealed the presence of phosphorus and calcium. Chemical, gemological, and observational features indicated bone, which was further confirmed
by Raman spectroscopy. Raman spectra (figure 23) obtained
using a 785 nm laser in the 2002000 cm1 region displayed
166
Figure 23. The bones Raman spectrum showed major

PO43 peaks at about 960, 586, and 430 cm1; a CO43
peak at 1072 cm1; and collagen-related peaks in the
12001700 cm1 region.
RAMAN SPECTRUM
4000
3500
3000
COUNTS
Figure 22. Subparallel veins were confined to circular

concentric planes oriented along the length of the
cabochon (top), while the individual veins were surrounded by concentric whitish rings (bottom). Photomicrographs by Gagan Choudhary; magnified 40
(top) and 56 (bottom).
Bumble Bee jasper from Indonesia. At the Tucson Gem

and Mineral Show, All in Vein (Quartzsite, Arizona) exhibited slabs and pairs of Bumble Bee jasper, which was sold
elsewhere at the show as Eclipse jasper. The term jasper
is a misnomer, as this vibrantly colored orange, yellow, and
2500
960
2000
1500
1000
430
586
1072
500
0
200
400
600
800
1000
1200
1400
1600
1800
RAMAN SHIFT (cm-1)
GEMS & GEMOLOGY
SUMMER 2014
Figure 24. This Bumble Bee jasper, measuring 32

54 mm, is a mixture of volcanic lava and sediment.
Photo by Robert Weldon/GIA; courtesy of Robert E.
Kane.
Figure 26. This Pyrex bracelet displays a striking

orbicular pattern that appears three-dimensional.
Photo by Eric Welch; courtesy of Kevin OGrady.
black material (figure 24) actually formed from a mixture

of Indonesian volcano lava and sediment. A carbonate-rich
rock first discovered on the island of Java during the 1990s,
the material is soft, with a Mohs hardness of 5 or below.
The porous rock is easily cut and polished, and most specimens are filled with Opticon (H. Serras-Herman, Bumble
Bee jasper: A colorful volcanic lapidary material, Rock
& Gem, Vol. 43, No. 8, pp. 2629).
ety of dramatic effects. He crafts each piece freehand using

a torch, adding elements such as cobalt to achieve colors and
embedding flower mosaics and other patterns that appear
three-dimensional. OGradys signature bracelets contain
gold, silver, or a combination of both.
Stuart Overlin
Pyrex bracelets. At the Pueblo Gem & Mineral Show, Kevin
OGrady (Scottsdale, Arizona) exhibited a line of glass
bracelets and other jewelry (figures 25 and 26). OGrady has
worked with borosilicate glass, better known by the brand
name Pyrex, since the late 1980s. He touted the durability
and versatility of this medium, which lends itself to a vari-
Stuart Overlin
Shattuckite from the DRC. At the Tucson Gem and Mineral Show, Brett Kosnar of Mineral Classics (Black Hawk,
Colorado) exhibited rare specimens of shattuckite from the
Democratic Republic of Congo. A new find of this intense
medium to dark blue copper silicate (figure 27) occurred in
the Kambove District in October 2013. The mineral has a
Mohs hardness of 3.5 and is used for cabochons and carvings. Shattuckite has been reported from various African
and European sources, as well as copper mines in the
Figure 25. Pyrex bracelets with attractive colors and patterns were
on display at the Tucson show.
Photo by Eric Welch; courtesy of
Kevin OGrady.
GEMS & GEMOLOGY
SUMMER 2014
167
Figure 27. These intense blue shattuckite specimens

are from a recent find in the Democratic Republic of
Congo. Photo by Brett Kosnar; courtesy of Brett Kosnar, Kosnar Gem Co.
American West. The type locality is the Shattuck-Denn

mine in Bisbee, Arizona (R. Bowell and R. Cook, Connoisseurs Choice: Shattuckite: Kunene District, Kaokoveld,
Namibia, Rocks & Minerals, Vol. 84, November/December 2009, pp. 544550).
Kosnars Congolese shattuckite is being stabilized prior
to jewelry manufacturing. He plans to debut cabochons of
this very limited material at Tucson in 2015.
Stuart Overlin
Figure 28. Top-quality stones, including purple and

green jade, were offered at Jewelry Shanghai 2014.
Photo by Jennifer Stone-Sundberg.
Figure 29. This clear quartz carving seen at the

Shanghai show demonstrates the exceptional craftsmanship sought by Chinese collectors. Photo by
Jennifer Stone-Sundberg.
CONFERENCE REPORT
Jewelry Shanghai 2014. Organized by the Gems & Jewelry
Association of China, Jewelry Shanghai 2014 took place
May 812 at the Shanghai World Expo Exhibition & Convention Center. The two exposition halls held over 1,000
vendors representing more than 22 countries in 4,200
square meters of exhibitor space. A wide variety was offered, including exceptional examples of tourmaline, jade
(figure 28), and ruby, as well as phenomenal stones such as
cats-eye chrysoberyl and star sapphire. Additionally, superb craftsmanship was on display in the form of fine jade
and quartz carvings (figure 29), gold work, and cloisonn
enameling. This year saw the addition of top-level fossil,
mineral, and gem collector specimens. Other popular items
for sale were amber, turquoise, and moonstone.
The show featured lectures on topics such as Chinese
jade carving, diamond, and Burmese rubies. As Shanghai is
168
GEMS & GEMOLOGY
SUMMER 2014
Figure 30. Fantastic Cross, an

18K gold brooch featuring turquoise, variscite, chrysocolla, cultured pearl, and enamel, is part of
the museum display at GIAs
New York laboratory and campus.
Photo by Orasa Weldon/GIA,
courtesy of the Anthony and
Elizabeth Duquette Foundation
for the Living Arts.
Chinas diamond import center, the Shanghai Diamond Exchange (which handles every diamond imported into the
country) and the Hoge Raad voor Diamant (HRD) were present. Several other gem- and jewelry-related agencies, including the Shanghai Gold Exchange and the International
Gemological Institute (IGI), also attended. Mainland China
posted jewelry retail sales of 295.9 billion yuan (approximately US$47 billion) in 2013 (http://www.stats.gov.cn/
tjsj/zxfb/201401/t20140120_502082.html), and this show
certainly indicated the quality and diversity of items the
Chinese gem and jewelry market demands today.
Jennifer Stone-Sundberg
Crystal Solutions, LLC
Portland, Oregon
Ren Lu
China University of Geoscience
Wuhan
MISCELLANEOUS
GIA Museum exhibits in New York. GIAs recently opened
laboratory and campus at the International Gem Tower in
New Yorks Diamond District features a permanent museum exhibit of gems, jewelry, and mineral sculptures.
Fourteen display cases highlight nearly 100 items from the
GIA collection, ranging from large aquamarine crystals to
a diminutive objet dart golden cactus and a necklace featuring akoya, Tahitian, and South Sea cultured pearls (figure 30). The exhibit showcases the wonder of gems to
visitors and students.
ERRATUM
In the Gem News International section of the Spring 2014
issue, the photographer of the large oolitic opal block
should have been listed as Don Mengason.
GEMS & GEMOLOGY
SUMMER 2014
169
1: The exhibits location at the GIA library entrance. 2: Colombian emerald necklace, Summer
1981 issue. 3: Kathryn Kimmel (left) with Harold and Erica Van Pelt. 4: 80th anniversary decor.
5: Rubellite necklace and ring, Winter 1988 cover. 6: Bill Larson, Susan Jacques, and Elose
Gaillou. 7: Tom Overton, Brendan Laurs, Alice Keller, and John King.
170
GEMS & GEMOLOGY
SUMMER 2014
Celebrating G&Gs 80th Year

and the Artistry of Harold and
Erica Van Pelt
On June 11, G&G contributors, supporters, and staff toasted the opening of
the new GIA Museum exhibit, The Beauty of Science: Gems & Gemology
Celebrates 80 Years, Featuring the Artistry of Harold and Erica Van Pelt.
The exhibit spotlights gems and jewelry featured on selected G&G covers
since 1981. The cover photos are the work of the Van Pelts, the husbandand-wife team who contributed nearly three decades of photography to
the journals pages. In addition to gathering treasures from the
Smithsonian and other sources, the collection offers intriguing examples
of cutting-edge research, locality reports, and industry coverage over the
years. The exhibit runs through December 14.
1
11
12
13
10
14
15
8: Kathryn Kimmel, Janine Castro, Laila Sharif, Susan Jacques, Tess Lecklitner, and Amanda Bilberry. 9: Pat Syvrud, Jeff Post, and Terri Ottaway.
10: John King. 11: Qing dynasty jade vase, Spring 1986 cover. 12: The Van Pelts and Stan and Leslie Weinberg. 13: Kate Donovan and Gabriel
Mattice. 14: Jeff Post and Zee Allred. 15: Marcelo Souza, Diane Caldwell, Dona Dirlam, and Charlie Carmona.
Photos by Kevin Schumacher and Matt Hatch/GIA.
GEMS & GEMOLOGY
SUMMER 2014
171
RESEARCH AND NEWS

Visit G&G online to explore free multimedia content. Scan the QR codes with your smartphone or tablet, or
enter the links below.
Gemological Field Expeditions: GIAs Andy Lucas and Vincent Pardieu

discuss the challenges and rewards of gemology fieldwork. Contains
slideshows from local mines and markets. www.gia.edu/gia-newsresearch-behind-scenes-gemological-field-expedition.
Mogok Expedition Three-Part Series: Our resident field gemologists

reminisce about the sights, the people, and the gem markets in Myanmars
famed Valley of Rubies. Contains slideshows showing local sites, miners,
and gem markets. www.gia.edu/gia-news-research-expedition-to-thevalley-of-rubies-part-1.
Chinese Gem and Jewelry Industry: The authors explore the development, impact, and future of Chinas domestic and international gemstone
and jewelry market. Includes artisan video interviews, on-site
slideshows, and a Chinese language PDF. www.gia.edu/gemsgemology/spring-2014-lucas-chinese-gem-industry.
Understanding the Geology of Diamonds: Delves into the relationship

between natural diamond formation and the geological formation of
Earth. Includes a diamond geology glossary, slideshows, and author
interviews. www.gia.edu/gems-gemology/WN13-diamond-geologyshirey.
The Museum of Londons Cheapside Hoard: An inside look at this

extraordinary jewelry and gemstone collection dating from the 17th
century. Contains slideshows of the exhibit and video interviews with
museum curators. www.gia.edu/gems-gemology/FA13-cheapsidehoard-weldon.
Occurrences of Oregon Sunstone: GIA visits three major sources of

Oregon sunstone to collect samples for research and provide background
into the geological relevance of this material. Includes video interviews and
slideshows from the sites. www.gia.edu/gems-gemology/FA13-oregonsunstone-pay.
A Look at Tucson 2014: Industry insight, creative innovations, and

unusual finds from this years gem and mineral shows. Includes video
interviews. www.gia.edu/gems-gemology/spring-2014-gemnewstucson-2014-overview.

Gems and Gemology Summer 2014 PDF

Uploaded by

Copyright:

Available Formats

Gems and Gemology Summer 2014 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

Gems and Gemology Summer 2014 PDF

Uploaded by

Copyright:

Available Formats

SUMMER 2014

THE QUARTERLY JOURNAL OF THE GEMOLOGICAL INSTITUTE OF AMERICA

The Future of Botswanas Diamond Industry

Challenges and Opportunities

Botswanas Scintillating Moment

Three-Phase Inclusions in Emerald and Their Impact on Origin Determination

NOTES AND NEW TECHNIQUES

Color Phenomena of Blue Amber

Experimental Studies on the Heat Treatment of Baltic Amber

80TH ANNIVERSARY SPECIAL

Thank You, Donors

Gem News International

Editors, Lab Notes

Editorial Review Board

Copyright and Reprint Permission

Postmaster: Return undeliverable copies of Gems

Our Canadian goods and service registration number is 126142892RT.

A.J.A. (Bram) Janse

Gems & Gemology is published quarterly by the

About the Cover

Challenges and Opportunities

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

GEMS & GEMOLOGY

BOTSWANAS SCINTILLATING MOMENT

ust off Airport Road in Gaborone, Botswana,

governments in other African countries such as

See end of article for About the Authors and Acknowledgments.

BOTSWANAS DIAMOND INDUSTRY

GEMS & GEMOLOGY

Figure 2. These diamond

Perhaps even more importantly, a new 10-year sales

since the colonial days of South Africa. As a result,

BOTSWANAS DIAMOND INDUSTRY

GEMS & GEMOLOGY

Figure 3. Philippe Mellier, CEO of De Beers Group.

For the people of Botswanathe Batswanathe

The changes have required De Beerss 82 sightholders

BOTSWANAS DIAMOND INDUSTRY

GEMS & GEMOLOGY

diamond business, as well as to address the concerns

BOTSWANAS DIAMOND INDUSTRY

GEMS & GEMOLOGY

Perhaps the biggest challenge for Botswana lies in

BOTSWANAS DIAMOND INDUSTRY

Growth in African manufacturing, however

BOTSWANAS DIAMOND BUSINESS

GEMS & GEMOLOGY

Figure 9. Map of Botswanas diamond mines.

BOTSWANAS DIAMOND INDUSTRY

GEMS & GEMOLOGY

lar Jwaneng deposittruly changed the equation. Two

entitled to a fair return (Idex, 2004). Clearly much

GEMS & GEMOLOGY

BOTSWANAS DIAMOND INDUSTRY

Figure 11. The Three Dikgosi, a frequently visited

the Jwaneng mine alone (Even-Zohar, 2002). For De

BOTSWANAS DIAMOND INDUSTRY

noted in a state of the nation address that the country

GEMS & GEMOLOGY