The Geochemistry of Gems
and Its Relevance to Gemology:
Different Traces, Different Prices
George R. Rossman1
1811-5209/09/0005-0159$2.50 DOI: 10.2113/gselements.5.3.159
I
n colored gems, minor and trace chemical components commonly
determine the difference between a common mineral specimen and
a gemstone. Also, these components are often responsible for the
color, and may provide a “fingerprint” for determining the provenance
of the gemstone. The minor elements that are incorporated will depend
on local geologic conditions such as temperature, redox conditions, and,
particularly, chemistry.
Keywords : gemstone, provenance, color, geochemistry
COLOR IN GEMSTONES
Metal ions from the first row of transition elements in the
periodic table, especially Ti, V, Cr, Mn, Fe, and Cu, are the
most important causes of color in oxide and silicate
gemstones. V3+, Cr3+, Mn3+, and Cu 2+ can produce strong
coloration when present at concentrations of tenths of a
weight percent. Color comes from electronic transitions
involving only the electrons in the d-orbitals (referred to
as ligand-field transitions or crystal-field transitions). When
present by themselves, Fe2+, Fe3+, and Mn2+ typically require
higher concentrations to cause significant color. Intervalence
charge transfer (IVCT) interactions, which involve an
exchange of an electron between two cations with different
valences (for example, between Fe2+ and Fe3+ or between
Fe 2+ and Ti4+) are a major source of color in gems and
require only a small amount of the interacting couple to
produce intense color. In some systems, charge transfer
from oxygen to the metal ion also contributes to the color.
Green color can also occur in
andradite garnet, Ca 3Fe 2 (SiO 4 ) 3.
Andradite is pale yellow-green when
it has exactly the end member
composition, but commonly, minor
amounts of Ti4+ coupled with Fe2+
turn andradite to brown or black.
A beautiful green variety of andradite occurs when minor amounts
of Cr3+ enter the garnet (Mattice
1998). These stones, known as the
variety demantoid, are highly
valued (Fig. 1b).
The stoichiometric components of garnets also depend on
the geologic setting. In lithium pegmatites, minerals that
crystallize late in the formation of the gem pockets in the
pegmatites can be nearly devoid of iron. In this setting,
nearly pure end member spessartine garnet, Mn3Al2 (SiO4) 3,
can occur. This garnet has a beautiful orange color due to
Mn2+ in a cation site of eight-coordination (Fig. 2). If the
garnet grows while some iron is still present in the pegmatitic fluids, the color becomes a much less valuable brown-orange
due to solid solution with the almandine end member,
Fe3Al2 (SiO4) 3.
A
Garnets
Good examples of the compositional dependence of color
are provided by the garnet group. When grossular garnet,
Ca3Al2 (SiO4) 3, is composed of just the end member components, it is colorless. Ca 2+, Al3+, Si4+, and O2- ions do not
absorb light in the range of the visible spectrum. However,
low concentrations of minor elements can dramatically
modify the color. Small amounts of V3+ with some Cr3+
turn grossular into the green tsavorite variety (Fig. 1a).
Spectacular examples of these garnets occur in marble
seams in graphitic gneisses of the Mozambique belt in
northeastern Tanzania and southeastern Kenya. There,
metamorphic fluids were able to mobilize traces of vanadium and chromium from the host rock and incorporate
them in the grossular garnets. The unusual beauty of these
garnets was recognized after their discovery in 1967 and
they were given the trade name tsavorite, in honor of the
nearby Tsavo National Park in Kenya (Bancroft 1984).
B
(A) The tsavorite variety of grossular garnet owes its color
to the substitution in the Al site of about a percent level
of
accompanied by a lesser amount of Cr3+. This stone
weighs 7.4 carats. (B) The color of the demantoid variety of andradite
is due to a minor amount of Cr3+ but is somewhat modified by the
presence of the stoichiometric constituent Fe3+. This highly valued
variety of andradite has classically come from the southern Ural
Mountains of Russia. Photos : Wimon Manorotkul, Palagems.com
Figure 1
V3+ usually
1 Division of Geological and Planetary Sciences
California Institute of Technology
Pasadena, CA 91125-2500, USA
E-mail: [email protected]
E lements , V ol . 5,
pp.
159–162
159
J une 2009
PROVENANCE OF GEMS
Minor and Trace Elements
Over time, gems from certain localities have been recognized as having greater beauty, and thus greater value. Even
as new sources of gems are located, gems from the classic
localities may still be perceived to have a higher value than
more recently discovered stones of similar color and quality.
The geographical origin of gems, in a general sense, is
becoming an important commercial factor. More value is
ascribed to particular deposits of gems compared to others
with similar geology.
Minor and trace elements are often different or incorporated
differently in gems of the same species but from different
localities. Thus, they may provide a readily available tool for
determining the locality of origin of gems. The following
examples illustrate this concept.
One question that must be addressed is what can be done
with a faceted stone to determine its locality of origin. The
need to avoid visually destructive analytical methods
restricts the use of many standard geochemical methods
and presents demanding analytical challenges. A variety
of tools are now available, including minimally or
non­destructive chemical analysis for major and trace
elements, luminescence, and isotopic analysis. Other
avenues of investigation, such as inclusions and growth
features, are discussed in Fritsch and Rondeau (2009 this
issue) and Devouard and Notari (2009 this issue).
Tourmaline
Most gem tourmalines owe their color to Fe2+ (most blue
tourmalines), Fe2+ plus Fe2+ –Ti4+ IVCT (green), Mn3+ (pink),
Mn2+ –Ti4+ IVCT (yellow), or a combination of these factors
(Fig. 3). At a few localities, such as in Kenya and Tanzania, Cr3+
and v3+ are the minor components responsible for the color.
In 1988, a new find of gem-quality elbaite with unusually
saturated shades of green and blue was made in the
Brazilian state of Paraíba. The unusual blue color comes
from the copper content, which can range up to 1.7 wt%
CuO (Rossman et al. 1991). The stones became an instant
success in the commercial market (Fig. 4, inset). Later, tourmalines were found in Nigeria and Mozambique that also
contained copper and had blue colors approaching those
of the tourmaline gems from Paraíba. The question was
raised about the possibility of distinguishing the provenance
of copper-containing tourmalines once they had been
faceted and entered the market.
Quantitative laser ablation–inductively coupled plasma–
mass spectrometry (LA–ICP–MS) analysis can be used to
differentiate tourmalines from the various localities by
comparing concentrations or proportions of selected minor
and trace elements such as Cu, Mn, Ga, Pb, Be, Mg, and Bi.
For example, the Brazilian stones generally have more Mg,
Zn, Bi, and Sb, while the Nigerian stones generally contain
elevated levels of Ga and Pb (Abduriyim et al. 2006). By
comparing the relative proportions of Bi, Pb, and Ga in such
tourmalines, one can, in most cases, distinguish between the
three main geographical provenances (Fig. 4). However, there
is still a small overlap between the compositions of tourmalines from Mozambique and Brazil (Krzemnicki 2007).
Corundum
For many years, rubies from the Mogok region of Burma
were considered the finest in the world and commanded
a high price (Hughes 1997). Beginning in the early 1990s,
rubies from a different source in Burma appeared in markets
in Bangkok. The Möng-Hsu rubies generally are unsaleable
as mined. They usually must be heated, often to high
temperatures, to remove a naturally occurring dark blue
color that arises from a combination of Fe2+ –Ti4+ and Fe2+ –
Fe3+ IVCT in the core of the stones. Heating oxidizes the
Fe2+ to Fe3+, which disrupts the IVCT couple. Furthermore,
the heating of rubies from Möng-Hsu introduces flux into
cracks in the stones (Peretti et al. 1995; Emmett 1999).
Although beautiful, the rubies from Möng-Hsu are generally valued less than rubies from Mogok because they have
been treated to enhance their appearance (Drucker 1999).
A crystal of orange spessartine and a faceted gem from
the Little Three mine near Ramona, San Diego County,
California. The orange color is due to Mn2+ in the eight-coordinated
cation site of the garnet. Photo: Wimon Manorotkul, www.palagems.com
Figure 2
A suite of tourmalines illustrating the tremendous variety
of colors displayed by this mineral group. All the gems
in this figure probably belong to the tourmaline species elbaite. End
member elbaite’s ideal composition is Na(Li1.5Al1.5)Al6(BO3)3Si6O18 (OH)4,
a species that would be devoid of color if it were exactly the ideal
composition. The gems in this photo are colored by traces of iron,
manganese, and titanium. Photo: Wimon Manorotkul, www.palagems.com
Figure 3
E lements
Thus it would be useful to be able to distinguish rubies
from different localities. Apart from specific microscopic
features, it has been shown that rubies from the Mogok
and Möng-Hsu localities can be differentiated on the basis
of their Ti/V ratio (Muhlmeister et al. 1998; Mittermayr et
al. 2008) as determined using X-ray fluorescence (XRF) or
LA–ICP–MS. In other examples, key elements, such as V,
Ti, Ga, and Fe, have been used to separate rubies from
Vietnam/Burma versus Thailand or Tanzania. For a more
general distinction of ruby localities, the ratios of Fe, Ga,
and Cr have proven useful (Rankin et al. 2003; Peretti
2008; Schwarz et al. 2008). Likewise, the source of blue
sapphires can be determined using the trace element ratios
160
J une 2009
of Zn, Sn, Ba, Ta, and Pb as determined by LA–ICP–MS
analysis of element concentrations down to levels
approaching ppb (Guillong and Günther 2001; Rankin et
al. 2003; Abduriyim and Kitawaki 2006a).
Isotopic Methods
The provenance of gems has always been important to
some degree. However, now that provenance has increased
in importance for commercial reasons, the tools to determine origin have been refined. Chemical composition, inclusions, growth features, luminescence, and trace elements
may all have a role in the determination of provenance.
While stable isotopes have proven highly useful in
geochemistry for studying the geological history of rocks
and minerals, they have, to date, found little practical
application in the determination of the provenance of
gemstones. In principle, isotopes should provide information about the origin of gems, but the cost, time required,
and destructive nature of these tests have, until now,
prevented isotopic methods from gaining wide application
in gemology. A few examples demonstrate the utility of
isotopic methods when applied to gem minerals.
Emerald
Emerald is a green variety of beryl, Be 3Al 2 Si6O18, that
contains Cr3+ and, occasionally, some V3+ as the chromophore.
It forms from hydrothermal fluids. The isotopic composition of these fluids varies with locality (Giuliani et al. 1998;
Zwaan et al. 2004). In an elegant study, Giuliani et al.
(2000) used the isotopic composition of oxygen in emerald
to trace international trade routes since antiquity (Fig. 5).
wide use in commercial gem laboratories, but holds much
promise for the future. As is the case with many analytical
methods, the overlapping ranges of oxygen isotope ratios,
especially for the classical or commercially important
deposits such as Mogok, Kashmir, Sri Lanka, and Madagascar,
mean that no single analytical method will provide the
answers to all problems of provenance.
SYNTHETIC CRYSTALS
Many of the same analytical methods used to differentiate
the geographic or geologic source of a gem can also be
applied to distinguish synthetic from natural stones. Such
distinctions will become increasingly important as the
quality of synthetic materials rises to nearly match that of
their natural counterparts.
Synthetic Amethyst
Hydrogen is an important trace element in many natural
minerals. It is a common charge-balancing cation (in the
form of an OH group). Its mode of incorporation can vary
depending on the geologic conditions of formation of the
host crystal. The intensity and shape of absorption bands
in the OH region of the electromagnetic spectrum provide
a test for synthetic amethyst. A band at 3595 cm-1 is present
in the infrared spectrum of all natural amethysts but only
rarely in synthetic ones. If present in synthetic amethyst,
its full width at half maximum (FWHM) is about 7 cm-1,
whereas it is about 3 cm-1 in all natural samples. This
absorption band difference provides a method to separate
natural from synthetic amethysts (Karampelas et al. 2005).
Synthetic Ruby
Corundum
Because both ruby and sapphire occupy an important place
in the gem market, the origin of corundum gems is a matter
of interest. In addition to the use of chemical element
ratios, as discussed above, to distinguish among localities,
certain classes of corundum show large isotopic differences
between different localities (Yui et al. 2003; Giuliani et al.
2005). Oxygen isotopes in carbonate-hosted corundum
show wide variations, whereas oxygen isotopes in mantlederived corundum vary much less. Because of the time
required for isotope analysis, its expense, and the destructive nature of the technique, the approach has not gained
Minor and trace components commonly found in nature
can be lacking in some synthetic gems. Such differences
can be detected by some of the same testing methods previ-
The oxygen isotope composition of emeralds varies
among important gem-producing regions of historical
importance. The colored bands indicate the range of composition at
each locality, and the white rectangles indicate the composition of
the ancient emeralds studied by Giuliani et al. (2000). 1: Gallo-Roman
earring; 2: Holy Crown of France; 3: Haüy’s emeralds; 4: Spanish
galleon wreck; 5: “old mine” emeralds. The isotopic variations allowed
these authors to trace the flow of emeralds in world commerce from
antiquity to the late 18th century. Graph modified from Giuliani et al. (2000)
Figure 5
A plot of the relative proportions of Bi, Pb, and Ga, all
present as trace elements in Cu-containing tourmalines
(inset), can in most cases distinguish the provenance of such
tourmalines: Nigeria, Mozambique, or Brazil. These data were
obtained using LA–ICP–MS. Graph modified from Mickael Krzemnicki
(2007); Photo (inset): Wimon Manorotkul, www.palagems.com
Figure 4
E lements
161
J une 2009
ously discussed. For example, for several years, synthetic
ruby was made from purified aluminum oxide from which
most of the naturally occurring gallium had been removed
in the industrial purification process. Thus, natural rubies
were readily distinguished by the presence of trace concentrations of gallium. However, soon after the gallium test
became widely known, gallium began to appear in some
synthetic stones.
Synthetic Emerald
For some time, a minor absorption band at 2293 cm-1 in
the infrared spectrum of natural emeralds was found to be
absent in the spectrum of synthetic emeralds and could
therefore be used as a test to distinguish between natural
and synthetic stones. However, such tests are not always
long lasting. In the case of emeralds, Russian hydrothermal
synthetic emeralds now contain the 2293 cm-1 band (DurocDanner 2006). Bands caused by water trapped in the c-axis
channels of beryl are present in natural emeralds and aquamarines (the blue variety of beryl) but are absent in fluxgrown synthetic emeralds. Fortunately, other methods for
distinguishing natural from synthetic emeralds are available,
and these are based on trace element analysis using
methods such as XRF and particle-induced X-ray emission
(PIXE) (Yu et al. 2000).
REFERENCES
Abduriyim A, Kitawaki H (2006a)
Determination of the origin of blue
sapphire using Laser Ablation Inductively
Coupled Plasma Mass Spectrometry
(LA-ICP-MS). The Journal of Gemmology
30: 23-36
Abduriyim A, Kitawaki H (2006b)
Applications of laser ablation–inductively
coupled plasma–mass spectrometry
(LA-ICP-MS) to gemology. Gems &
Gemology 42: 98-118
Abduriyim A, Kitawaki H, Furuya M,
Schwarz D (2006): “Paraíba”-type copperbearing tourmaline from Brazil, Nigeria,
and Mozambique: Chemical fingerprinting
by LA-ICP-MS. Gems & Gemology 42: 4-21
Bancroft P (1984) Tsavorite. In: Gem and
Crystal Treasures. Western Enterprises/
Mineralogical Record, Fallbrook, CA, pp
298-302. Available online at: http://
palagems.com/tsavorite_bancroft.htm
Devouard B, Notari F (2009) The identification of faceted gemstones: From the
naked eye to laboratory techniques.
Elements 5: 163-168
Drucker RB (1999) Ruby: Why the source
affects the price. www.jckonline.com/
article/CA635628.html
Duroc-Danner JM (2006) The identification
value of the 2293 cm-1 infrared absorption
band in natural and hydrothermal
synthetic emeralds. The Journal of
Gemmology 30: 75-82
Emmett JL (1999) Fluxes and the heat
treatment of ruby and sapphire. Gems
& Gemology 35: 90-92
Ertl A, Rossman GR, Hughes JM, Ma C,
Brandstätter F (2008) V3+ -bearing,
Mg-rich, strongly disordered olenite
from a graphite deposit near Amstall,
Lower Austria: A structural, chemical
and spectroscopic investigation. Neues
Jahrbuch für Mineralogie 184: 243-253
Fritsch E, Rondeau B (2009) Gemology:
The developing science of gems.
Elements 5: 147-152
E lements
TREATED NATURAL GEMS
Many of the tests to determine the geological or geographic
origin of a stone can also be used to find out if a stone has been
subjected to laboratory processes to change its color or other
properties. As an example, consider the corundum gems,
which are commonly heated to clarify and modify their color.
A recent development is the diffusion of beryllium, at the
level of 10 ppm or less, into the stones to change their color
to extents that range from subtle to dramatic. This treatment
was initially difficult to detect, but now a variety of analytical methods have been developed. Laser-induced breakdown spectroscopy (LIBS), LA–ICP–MS, and secondary ion
mass spectrometry (SIMS) now make it possible to detect
these low levels of beryllium in treated stones (Krzemnicki
et al. 2004; Abduriyim and Kitawaki 2006a, b).
CONCLUSIONS
The examples cited illustrate just a few of the methods, both
common and sophisticated, that are employed to determine
the origin of gem materials. In many cases, rigorous tests
prove to be too expensive compared to the value of the item
tested, or currently are too destructive for routine use. In
several instances, the geochemical reasons for some of the
observed differences are not fully understood. In other
cases, suitable tests are still lacking.
Giuliani G, France-Lanord C, Coget P,
Schwarz D, Cheilletz A, Branquet Y,
Giard D, Martin-Izard A, Alexandrov P,
Piat DH (1998) Oxygen isotope systematics
of emerald: relevance for its origin and
geological significance. Mineralium
Deposita 33: 513-519
Giuliani G, Chaussidon M, Schubnel H-J,
Piat DH, Rollion-Bard C, France-Lanord C,
Giard D, de Narvaez D, Rondeau B (2000)
Oxygen isotopes and emerald trade routes
since antiquity. Science 287: 631-633
Giuliani G, Fallick AE, Garnier V, FranceLanord C, Ohnenstetter D, Schwarz D
(2005) Oxygen isotope composition as
a tracer for the origins of rubies and
sapphires. Geology 33: 249-252
Guillong M, Günther D (2001) Quasi ‘nondestructive‘ laser ablation-inductively
coupled plasma-mass spectrometry
fingerprinting of sapphires. Spectrochimica
Acta B - Atomic Spectroscopy 56: 1219-1231
Hughes RW (1997) Ruby & Sapphire.
Chapter 12, World Sources. RWH
Publishing, Fallbrook, CA
Karampelas S, Fritsch E, Zorba T,
Paraskevopoulos KM, Sklavounos S (2005)
Distinguishing natural from synthetic
amethyst: the presence and shape of the
3595 cm-1 peak. Mineralogy and Petrology
85: 45-52
Krzemnicki MS (2007) “Paraiba”
tourmalines from Brazil and Africa.
Origin determination based on LA-ICP-MS
analysis of trace elements. SSEF Facette
14: 9. Available online at www.ssef.ch/
en/news/facette_pdf/Facette14_e.pdf
Krzemnicki MS, Hänni HA, Walters RA
(2004) A new method for detecting Be
diffusion–treated sapphires: Laser-induced
breakdown spectroscopy (LIBS). Gems
& Gemology 40: 314-322
Mattice G (1998) Demantoid From the Ural
Mountains of Russia. The Gem Spectrum.
http://palagems.com/gem_spectrum4.1.htm
Mittermayr F, Konzett J, Hauzenberger C,
Kaindl R, Schmiderer A (2008) Trace
element distribution, solid- and fluid
162
inclusions in untreated Mong Hsu
rubies. Geophysical Research Abstracts
10, EGU2008-A-10706, 2008 SRef-ID:
1607-7962/gra/EGU2008-A-10706
Muhlmeister S, Fritsch E, Shigley JE,
Devouard B, Laurs BM (1998) Separating
natural and synthetic rubies on the basis
of trace-element chemistry. Gems &
Gemology 34: 80-101
Peretti A (2008) New important gem
discovery in Tanzania: the Tanzanian
“Winza”-(Dodoma) rubies. Contributions
to Gemmology 7. www.gemresearch.ch/
news/Tanzania/Tanzania.htm
Peretti A, Schmetzer K, Bernhardt, H-J,
Mouawad F (1995) Rubies from Mong
Hsu. Gems & Gemology 31: 2-25
Rankin AH, Greenwood J, Hargreaves D
(2003) Chemical fingerprinting of some
East African gem rubies by Laser Ablation
ICP-MS. The Journal of Gemmology
28: 473-482
Rossman GR, Fritsch E, Shigley JE (1991)
Origin of color in cuprian elbaite from
São José de Batalha, Paraíba, Brazil.
American Mineralogist 76: 1479-1484
Schwarz D, Pardieu V, Saul JM, Schmetzer
K, Laurs BM, Giuliani G, Klemm L,
Malsy A-K, Erel E, Hauzenberger C,
Du Toit G, Fallick AE, Ohnenstetter D
(2008) Rubies and sapphires from
Winza, central Tanzania. Gems &
Gemology 44: 322-347
Yu KN, Tang SM, Tay TS (2000) PIXE
studies of emeralds. X-Ray Spectrometry
29: 267-278
Yui T-F, Zaw K, Limtrakun P (2003)
Oxygen isotope composition of the
Denchai sapphire, Thailand: a clue to
its enigmatic origin. Lithos 67: 153-161
Zwaan JC, Cheilletz A, Taylor BE (2004)
Tracing the emerald origin by oxygen
isotope data: the case of Sandawana,
Zimbabwe. Comptes Rendus Geoscience
336: 41-48
J une 2009