CHARACTERIZATION OF VIBRATIONAL AND ELECTRONIC

FEATURES IN THE RAMAN SPECTRA OF GEM MINERALS

by

Renata Jasinevicius

A Prepublication Manuscript Submitted to the Faculty of the

DEPARTMENT OF GEOSCIENCES

In Partial Fulfillment of the Requirements

for the Degree of

MASTER OF SCIENCE

In the Graduate College

THE UNIVERSITY OF ARIZONA
2009
ii
ACKNOWLEDGMENTS

This work could not have been completed without the sponsorship of the Israel
Diamond Institute (IDI).

A special thanks to the following contributors for making this work possible:

James Shigley, Ph.D.-Gemological Institute of America (GIA)

Dr. M. Bonner Denton-University of Arizona, Dept. of Chemistry
Bear Williams, Stone Group Labs (photos)
Tom Tashey-Professional Gem Sciences, Inc.
Charlene Estrada, photography
Sue Robison, RRUFF Project, University of Arizona
To my advisor, Bob Downs, you have challenged me in new ways and taught me how to network.
You have helped me fine tune my writing skills and master the art of thinking scientifically. You
have made me a better student and opened doors for my future. Thank you.

I’d like to extend my gratitude to all the people who provided me with personal support and
encouragement throughout my academic career:

To my future husband, Jason Lafler, I never would have made it without you. I am the luckiest
lady in the world.

Madison Barkley, you are a dear friend, and having you in my life made all of my graduate school
experiences more memorable.

To my mentor and friend, Elizabeth Gordon, your guidance has helped me become the person I
am today. You are an inspirational teacher and I appreciate everything you have done for me.

To my best friend, Andrea Mikenas, you always believed I could do it and never failed to remind
me.

To my former roomie, Lindsay Draeger, your Sunday morning polka phone calls made me smile
and made me feel a little less lonely.

Last, but certainly not least, I’d like to thank my family. To my Mom and Dad who have always
pushed me to achieve excellence, I love you and appreciate everything you have sacrificed to
help me achieve my goals. Rich, you are a great big brother.

To my future parents, Val and Eunice, since the first day we met you have taken me under your
wings and treated me like your daughter. I love you both and feel incredibly blessed to have you
in my life.

iii
CONTENTS
INTRODUCTION 1
Raman Spectroscopy 2
IR Spectroscopy 14
OH in Minerals 14
Luminescence Spectroscopy 15
Color Theory 16

RAMAN ANALYSIS

1. Beryl 19
2. Chrysoberyl 26
3. Corundum 30
4. Diamond 35
5. Diopside 43
6. Garnet 48
7. Olivine 74
8. Quartz 79
9. Spinel 82
10. Spodumene 87
11. Titanite 90
12. Topaz 97
13. Tourmaline 102
14. Zircon 114
15. Zoisite 125

APPENDICES

Appendix A & B: Features in Raman Spectra 132

Appendix C: Unit Conversions 141
Appendix D: Cause of Color Chart 143
Appendix E: Point Symmetry Notations 147

iv
Introduction

Raman spectroscopy, a non-destructive technique used to interpret atomic vibrations, has become a
popular tool for the rapid identification of materials. Raman spectra produce unique vibrational
fingerprints useful in identifying a multitude of materials. With the advent of numerous gem treatments
and a variety of methods available for mineral synthesis, Raman spectroscopy is particularly useful in
identifying and characterizing gemstones. Micro-inclusions in minerals can be analyzed using Raman
spectroscopy providing evidence of mineral genesis or geologic origin. More recently, fluorescence
features attributed to chromophoric ions and trace elements have been observed in Raman spectra,
revealing important information about crystal chemistry. Analysis and interpretation of these features may
help distinguish between natural, treated, and synthesized materials. Advances in optical technologies
are bringing hand-held Raman spectrometers to the forefront of materials research. As new instruments
are developed, with both increases in portability and decreases in production costs, hand-held Raman
units will likely be fundamental to laboratory and field-based Geoscience and gemological research in the
future. Therefore, the development of databases and interpretation of spectra in anticipation of the new
instrumentation is required. Raman spectra and associated interpretation of spectral features for the
important gemstones are presented in this study.

In this study I will present a characterization of the vibrational and electronic features present in the
Raman spectra of gem minerals including:

ƒ The effects of orientation

-
ƒ Spectral features associated with the vibrational modes of OH and H2O
ƒ The causes of color in various gem minerals and the electronic spectral features associated with
particular color-inducing cations
ƒ Spectral features associated with luminescence of REE
ƒ The effects of metamictization

1
Basic Raman Theory

Raman spectroscopy is a type of vibrational spectroscopy. Intense electromagnetic radiation (typically

generated by a laser) interacts with a substance and is scattered into radiation of different wavelengths
associated with nuclear motion and producing a unique spectral fingerprint of atomic vibrations (Smith
and Dent, 2005). An intense light source is necessary in Raman spectroscopy because Raman
scattering is a very weak process; only one out of every 106-108 photons will Raman scatter (Smith and
Dent, 2005). In general, gem minerals at room temperature are in the ground vibrational state (lowest
energy vibrational level). When light interacts with a crystal the incident radiation can be scattered in
several ways that include: 1) Rayleigh scattering, in which the scattered photon retains the energy of the
incident beam (no energy change, elastic scattering) (Fig. I1) and 2) Raman scattering, in which the
scattered photon experiences a change in energy (inelastic). There are two types of Raman scattering:
1) Stokes and 2) anti-Stokes. Stokes scattering occurs when atoms in the crystal (at the ground
vibrational state) absorb the energy from the incident photon and are ultimately promoted to a higher
energy vibrational state; the incident photon loses energy relative to its original state and the wavelength
of the scattered light is shifted towards the red end of the electromagnetic spectrum (Fig. I1). During anti-
Stokes scattering, energy is transferred from the already excited atoms to the incident photon and
subsequently, the atoms associated with this specific vibration are demoted to the ground vibrational
state; the scattered light is higher in energy than the Rayleigh line and therefore, the wavelength of the
scattered light is shifted towards the blue end of the spectrum (Fig. I1) (Smith and Dent, 2005). Anti-
Stokes scattering occurs less frequently than Stokes scattering because it requires that the atoms already
be in a higher energy vibrational state when the laser interacts with it (Smith and Dent, 2005). There are
characteristic temperatures at which certain vibrations are activated in different sets of bonded atoms.
For example, strongly bonded atoms like Si-O are not vibrationally active at room temperature; they are
considered high energy vibrations.

Fig. I1 Energy diagram showing transitions in various types of spectra (modified from Smith and Dent,
2005).
IR Rayleigh Stokes Anti-Stokes Fluorescence
Raman Raman
Electronic Vibrational Relaxation &
States Internal Conversion

Virtual
States
Fluorescence
(Excitation)
Absorption

(Emission)

Vibrational
States

Ground State

Whether or not a vibrational mode is Raman active depends on the polarization of the vibrating bonded
atoms. When the incident beam interacts with the crystal, the atoms begin to oscillate at the same
frequency of the incident radiation. As an atom oscillates, its electrons are pulled in various directions,
depending on their distribution in the electron cloud, resulting in deformation of the cloud. As the
electrons move, so does the atom’s nucleus producing a separation of charges in the atom called a dipole

2
(the atom becomes polarized). Changes in the polarization of bonded atoms produce Raman active
vibrational modes (Ferraro et al., 2003). In minerals, chemistry and crystal structure dictate the types of
vibrations that can occur. The way bonded atoms in a crystal can bend, stretch, or rotate, i.e. their
degrees of freedom of movement, depends on the crystal symmetry (Smith and Dent, 2005). Complex
correlation matrices involving specific site symmetries can predict Raman active modes in minerals
(Ferraro et al., 2003). The details of this process are beyond the scope of this study. However, for a
more detailed discussion of Raman selection rules, character tables, factor group analysis, and crystal
field theory see the following sources:

“Infrared and Raman Selection Rules for Molecular and Lattice Vibrations: The Correlation Method,” W.G.
Fateley and F.R. Dollish, 1972

“Molecular Vibrations: The Theory of Infrared and Raman Vibrational Spectra,” E.B. Wilson, J.C. Decius,
and P.C. Cross, 1955

“Chemical Applications of Group Theory,” F.A. Cotton, 1971

“Symmetry in Bonding and Spectra: An Introduction,” B.E. Douglas and C.A. Hollings

Note: Raman mode analysis tables in this study were generated using the Raman mode prediction tool
found on the Bilbao Crystallographic Server available online at http://www.cryst.ehu.es/.

References

Bilbao Crystallographic Server II: Representations of crystallographic point groups and space groups”. Acta Cryst.
(2006), A62, 115-128.
Ferraro, J.R., Nakamoto, K. & Brown, C.W. (2003) Introductory Raman Spectroscopy, 2nd edn, Academic press, San
Diego, CA.
Smith, E. & Dent, G. (2005) Modern Raman spectroscopy: a practical approach, John Wiley & Sons, Chichester,
West Sussex, England.

3
Experimental Procedures and the RRUFF Database

All Raman spectra and X-ray diffraction data utilized in this study have been taken from the RRUFF
database (found at www.rruff.info). This database, directed by Dr. Robert T. Downs at the University of
Arizona, provides a large set of chemical and spectral data of minerals. All information is available to the
public including the downloadable software, CrystalSleuth, used in this study to view, stack, and process
Raman spectra (Laetsch and Downs, 2006). The development of the RRUFF database is beneficial to
the advancement of Raman spectroscopic research, particularly to applications in geoscience, gemology,
and material science.

Two Raman spectrometers are used to collect spectral data in the RRUFF lab: 1) Thermo Nikolet Almega
microRaman with 532 nm and 780 nm lasers (partially polarized, 1 μm spot size, 4 cm-1 resolution) and 2)
a customized open beam system with a 514 nm tunable Ar laser and a Jobin Yvon Spex HR 460
spectrometer equipped with a liquid nitrogen cooled CCD and a 1200 grooves per mm grating; used
specifically to collect Raman spectra of oriented samples.

Important Notes:

Polarization: The 514 nm Ar laser in the open beam Raman system has vertical linear polarization,
meaning the electric field of the laser oscillates vertically (up and down) with respect to the direction of
propagation of the laser beam (Fig. I2). When the light leaves the laser enclosure it is polarized,
however, interaction of the laser with various optical elements such as lenses, gratings, and mirrors can
change the laser polarization due to geometrical changes along the laser’s path. Therefore, it is
important to note that the optical components of a spectrometer must be very deliberately and precisely
installed in order to maximize laser polarization over the entire optical bench.

Background Correction: Raman spectra are available on the RRUFF database in two forms: 1) raw
and 2) processed. Raw spectra are loaded into the CrystalSleuth program and any cosmic ray present in
the spectrum, non-Raman modes, or features associated with the notch filter are removed (we call this
process is called trimming) (Hill and Rogalla, 1992; Laetsch and Downs, 2006). Processed files are
“background corrected,” meaning the spectra are forced to the baseline by an algorithm in the
CrystalSleuth program (Laetsch and Downs, 2006). This allows for easy comparison of Raman spectra.
Any Raman spectra used in this study that have been background corrected are marked as such. This
process does NOT alter the Raman mode locations. An example is provided in Fig. I3.

Crystal Structures: All crystal structure figures and images have been generated using XtalDraw
software developed by Downs and Hall-Wallace (2003).

Fig. I2 (below) Picture demonstrating vertical linear polarization of a laser; image modified from Paschotta (2009)

Laser beam

4
Fig. I3 Raman spectra of spinel: 1) raw and 2) processed (background corrected); sample R050392, 780 nm laser

1.
Intensity

2.

-1
Raman Shift (cm )

Notation of Spectral Features

The positions of luminescence features in the Raman spectra of gem minerals are wavelength
dependent. For convenience, the peak positions of luminescence centers observed in this study will be
presented in both nm (as described in the literature) and Raman shifts (cm-1) using the following notation
(see also Appendix C for unit conversions):
laser wavelength (nm)
Raman shift cm-1

For example: “Luminescence bands located at 693 nm (5324366 cm-1) and 694 nm (5324396 cm-1) are
attributed to octahedrally coordinated Cr3+ luminescence centers in corundum (Gaft et al., 2005).”

Luminescence Features in this Study

In this study, the luminescence features appear in the extended range (>1500 cm-1) of the Raman spectra
of various minerals. In many cases the luminescence features are so intense that you can no longer see
the Raman peaks associated with the atomic vibrations of the mineral. An example of this is provided
below in the Raman spectrum of topaz (Fig. I4).

Fig. I4 Extended range Raman spectrum of topaz; note how the luminescence features overwhelm the Raman
peaks; sample R060024, Minas Gerais, Brazil, 532 nm laser, unoriented

Luminescence Features

Raman Peaks

5
Minimization of Luminescence

In a gated Raman study, Gaft et al. (2009) reported that by pulsing the laser, luminescence features that
would otherwise overwhelm the Raman spectrum of a mineral sample could be minimized (Gaft et al.,
2009).

References

Gaft, M. & Nagli, L. (2009) Gated Raman spectroscopy: potential for fundamental and applied mineralogy. European
Journal of Mineralogy, Vol. 21, No. 1, pp. 33.

Hill, W. & Rogalla, D. (1992) Spike-correction of weak signals from charge-coupled devices and its application to
Raman spectroscopy. Analytical Chemistry, Vol. 64, No. 21, pp. 2575-2579.

Downs, R.T. & Hall-Wallace, M. (2003) The American Mineralogist Crystal Structure Database. American
Mineralogist 88, 247-250.

Laetsch T. & Downs R. (2006) Software For Identification and Refinement of Cell Parameters From Powder
Diffraction Data of Minerals Using the RRUFF Project and American Mineralogist Crystal Structure Databases.
th
Abstracts from the 19 General Meeting of the International Mineralogical Association, Kobe, Japan, 23-28
July 2006.

Paschotta, R. (2009) Encyclopedia of Laser Physics and Technology. Online <http://www.rp-

photonics.com/encyclopedia.html>.

6
Relationship of Raman Peaks and Bond Length

A study comparing bond lengths to Raman peak positions of asymmetrical stretching modes in anionic
groups was conducted by Estrada et al. (2009) and a preliminary plot is provided and discussed here
(Fig. I5). Overall, the data displays a negative correlation, in which increases in bond lengths are
associated with smaller Raman shifts, and this data can be fit to a power law curve:
ν = 2328.9 R-1.95
Within each anionic group there are also negative correlations, however, the slopes of these trends are
much steeper.

In general, the following statements have been supported by the correlations reported in this study:
Shorter, stronger bonds produce Raman peaks located at higher Raman shifts. Longer, weaker bonds
appear at lower Raman shifts and are much more difficult to correlate because the atoms are not as well-
constrained as in short, strong bonds, therefore oscillating in many different ways (Estrada, 2009). This
information is very useful in approximately predicting Raman mode assignments in minerals containing
the anionic groups presented in this study.
.

Fig. I5 Plot of bond length (Å) vs Raman peak positions of asymmetrical stretching modes of anionic groups; bonds
specified in the key, reproduced with permission from (Estrada, 2009).

N-O
1600 C-O
III B-O
IV B-O
S-O
P-O
1400
Raman Frequency, ν (cm )
-1

Si-O
Be-O
Cr-O
As-O
1200 V-O
Mo-O
W-O

1000

800

600
1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
R(M-O) (Ǻ)

Reference

Estrada, C. (2009) The Correlation of M-O Bond Length to Raman Stretching Frequency in Mineral Anionic Groups.
unpublished.

7
Orientational Dependence of Raman Spectra

The optical indicatrices of minerals dictate optical properties along the crystallographic axes including the
orientational dependence of Raman spectra. Optical indicatrices are imaginary three-dimensional
surfaces in which the lengths of the axes are proportional to the refractive index (Dyar et al., 2008).
Examples of optical indicatries are presented below.

Modified from (Dyar et al., 2008)

Crystal Systems Optical Class

Isometric (cubic): a = b = c
Isotropic
α = β = γ = 90o
Tetragonal: a = b ≠ c
Uniaxial
α = β = γ = 90o
Hexagonal (trigonal): a = b ≠ c
Uniaxial
α = β = 90o γ = 120o
Orthorhombic: a ≠ b ≠ c
Biaxial
α = β = γ = 90o
Monoclinic: a ≠ b ≠ c
Biaxial
α = γ = 90o β ≠ 90o
Triclinic: a ≠ b ≠ c
Biaxial
α ≠ β ≠ γ≠ 90o

Isotropic Optical Indicatrix:

The Raman spectra of minerals with an isotropic indicatrix do not exhibit changes in peak intensity with
changes in the polarization of the incident beam. This is because no matter which may the crystal is
oriented, the cross-section of the indicatrix is a sphere. An example is diamond.

8
Uniaxial Optical Indicatrix:

Tetragonal and hexagonal minerals have uniaxial optical indicatrices. The Raman spectra of minerals
with uniaxial indicatrices exhibit changes in Raman peak intensity with a change in the polarization of the
laser. In the uniaxial optical indicatrix there is only one orientation in which the cross-section of the
indicatrix is circular. An example is beryl. When a beryl crystal is oriented such that the polarization of
the laser is parallel to the c-axis, there is no change in the intensity of the Raman peaks (the cross-
section through the indicatrix is a circle). However, when the laser is polarized parallel to the a-axis, the
peak intensities change (the cross-section of the indicatrix is an ellipse).

Uniaxial Positive Indicatrix Circular Cross-section

Uniaxial Negative Indicatrix

Biaxial Optical Indicatrix:

Orthorhombic, monoclinic, and triclinic minerals have biaxial optical indicatrices. A bixial indicatrix looks
like an ellipsoid, however, unlike the uniaxial indicatrix, there are two angles that produce a circular cross-
section.

Reference

Dyar, M., Gunter, M. & Tasa, D. (2008) Mineralogy and Optical Mineralogy. Mineralogical Society of
America. Chantilly, VA.

9
Factors Influencing the Raman Spectra of Minerals

During collection, analysis, and processing of hundreds of Raman spectra of various minerals, some
basic experimental observations have been made about factors that can affect the quality of Raman data.

Heating

When the incident laser beam interacts with a mineral surface, the surface is heated. Factors such as
size, chemistry, and color of a mineral can alter the way it is capable of absorbing and dispersing the
thermal energy produced by the laser. The effects of heating are illustrated by the following examples.

Burning

When a crystal is incapable of adequately absorbing and dispersing the energy of the incident beam, the
sample can burn. Dark brown to black craters that are the size of the beam spot are indicative of burning.
Sample size, chemistry, and color can influence the probability of burning.

Size: Any very small or very thin crystals are at risk of burning, for example, single crystals
prepared for X-ray diffraction analysis (50-100 microns in size). In addition, crystals that are
poorly embedded in matrix can burn (the mineral may not be able to dissipate the heat of the
laser). When dealing with small, thin, or poorly embedded crystals, the laser intensity is always
reduced to the lowest laser intensity, 10%, to minimize the risk of burning. Samples that are
embedded in matrix are often safe from burning.

Chemistry: Gem minerals typically do not burn because they are hard, stable, strongly bonded,
and usually transparent to translucent. However, there are many minerals that are at risk of
burning due to their chemical compositions. Soft, weakly bonded minerals such as silver
sulphides burn very easily. Acanthite (Ag2S) burns even under the lowest laser intensity setting.

Color: The Raman system used in this study has two possible laser wavelengths, 532 nm
(green) and 780/785 (red). Some crystals may burn under exposure to one of the lasers, but not
the other due to the color of the mineral. An example is chalcanthite (CuSO4·5H2O).
Chalcanthite does not burn under exposure to the higher energy 532 nm (green) laser, but does
burn under the 780/785 nm (red) laser. This preferential burning occurs because chalcanthite is
a blue mineral and when the red 780/785 nm laser hits the sample’s surface the sample absorbs
more energy from the red light than from the green. The sample cannot disperse the thermal
energy produced by the red laser because it is overwhelmed by absorption, and so it burns under
only one laser. Not all blue minerals burn under the 780/785 nm laser. Color is typically
combined with other factors such as size and chemistry in order to result in preferential burning.

Dehydration

Minerals containing water may dehydrate due to the heat produced by the lasers. The removal of water
or hydroxyl from certain minerals can result in a phase change. Discoloration of mineral surface where
the laser beam is focused on the sample is usually indicative of dehydration. In some cases, there is no
visual evidence of dehydration. If the Raman spectrum of a sample unexpectedly changes during data
during data collection, the sample has likely burned or dehydrated.

Phase Changes (not related to water)

Some minerals undergo a phase change when exposed to the heat of the lasers. An example is crocoite
(PbCrO4). After less than 1 second of exposure to the 532 nm laser, the Raman spectrum of crocoite
changes without any visual evidence on the sample’s surface to suggest alteration (there was no
discoloration). Knight et al. (2000) discovered that crocoite experiences a phase change at 1068 K and
transitions from the ambient temperature monazite structure to the barite structure. It is possible that the

10
laser heated the crocoite sample to 1068 K and ultimately caused the sample to change into its high
temperature phase.

Fluorescence

Minerals containing rare-earth elements (REE) frequently produce fluorescence features that can
overwhelm the Raman spectrum of a sample (Fig. I6 A & B). Minerals containing Ca, for example
tremolite (†Ca2Mg5Si8O22(OH)2), frequently incorporate REE into their structures due to the size and
coordination of the Ca-site and therefore, more commonly exhibit fluorescence in the Raman spectra.

Rock-salt structure

Minerals with the halite (or rock-salt) crystal structure produce Raman spectra with intensity equal to zero
due to the presence of an inversion center (Fig. I7 & I8). An inversion center is a symmetry element
describing a point within a crystal through which any straight line extends to points on opposite surfaces
of the crystal at equal distances. When the incident beam causes the atoms in a mineral with the rock-
salt structure to oscillate, the change in polarization of the atoms is negated because the distortion of the
electron cloud in one direction is equal to the distortion of the cloud in another direction (a sort of
destructive interference). The result is a Raman spectrum with intensity equal to zero.

Metamictization

Incorporation of radioactive elements like U and Th into the crystal structure of a mineral can result in
metamictization (a loss of crystallinity). There are several gem minerals that are affected by
metamictization, including titanite and zircon (see also sections on titanite and zircon). Metamictization
can have dramatic effects on the Raman spectrum of a sample including loss of peak intensity, an
increase in peak width, disappearance of peaks, and shifts in peak frequencies.

11
Fig. I6 A. Raman spectra of minerals exhibiting fluorescence.

B.

Fig. I7 A. Raman spectra of minerals with Raman intensities = zero due to symmetry

B.

12
Fig. I8 The halite structure; purple
spheres are Cl atoms, teal spheres are
Na atoms

References

Knight, K.S. (2000) A high temperature structural phase transition in crocoite (PbCrO4) at 1068 K: crystal structure
refinement at 1073 K and thermal expansion tensor determination at 1000 K. Mineralogical Magazine, Vol.
64, No. 2, pp. 291-300.

13
Infrared (IR) Spectroscopy

Infrared spectroscopy, like Raman spectroscopy, is a tool used to measure the unique vibrational
fingerprints of materials. In IR spectroscopy, radiation spanning a range of IR wavelengths is focused
onto a sample (Smith and Dent, 2005). When this energy is absorbed by sets of bonded atoms, the
atoms are promoted to a higher vibrational state. The energy of the incident light must match the energy
of a specific atomic vibration in order for absorption to occur (Smith and Dent, 2005). Vibrations are IR
active if a change in the dipole moment of the atoms (charge separation) is induced during absorption of
the incident radiation (Ferraro et al., 2003). IR spectra are commonly plotted with the x-axis equal to
either wavelength or wave number and the y-axis equal to either absorbance or transmission. It is
possible for vibrations to be both IR and Raman active. An example of equivalent vibrational features in
both the Raman spectra and the IR spectra of a gem mineral is a series of peaks attributed to molecular
water occupying the structural channels of beryl (see also beryl section).

References

Ferraro, J.R., Nakamoto, K. & Brown, C.W. (2003) Introductory Raman Spectroscopy, 2nd edn, Academic press, San
Diego, CA.
Smith, E. & Dent, G. (2005) Modern Raman spectroscopy: a practical approach, John Wiley & Sons, Chichester,
West Sussex, England.

Importance of OH- Incorporation in Minerals

Investigation of the substitution of OH- and H2O molecules into the structure of minerals, principally those
considered nominally anhydrous (those not requiring hydrous species to balance their stoichiometry), has
been of great interest over the past decade. Incorporation of OH- can affect a mineral’s diffusion and
dielectric properties (Rossman, 1996). Of even greater scientific interest is the incorporation of OH into
nominally anhydrous minerals present in the mantle including garnets and olivine (Beran and Libowitsky,
2006). The incorporation of water into the structure of minerals can affect a wide variety of important
mineral properties at depth including: melting conditions, bulk modulus (resistance to compression), and
phase transition processes. Therefore, characterizing the influences of OH- on minerals in the mantle can
help better constrain mantle dynamics (Beran and Libowitsky, 2006). Common crustal minerals such as
quartz, feldspar, and pyroxenes can also contain trace OH- (Johnson, 2006). The primary method utilized
in OH- studies is IR spectroscopy, however, Raman peaks associated with OH- and H2O in various
orientations have also been observed in my study and are discussed in the following sections: beryl,
garnet, titanite, topaz, tourmaline, zircon, and zoisite.

References

Beran, A. & Libowitzky, E. (2006) Water in natural mantle minerals II: olivine, garnet and accessory minerals.
Reviews in Mineralogy and Geochemistry, Vol. 62, No. 1, pp. 169-191.
Johnson, E.A. (2006) Water in nominally anhydrous crustal minerals: speciation, concentration, and geologic
significance. Reviews in Mineralogy and Geochemistry, Vol. 62, No. 1, pp. 117-154.
Rossman, G.R. (1996) Studies of OH in nominally anhydrous minerals. Physics and Chemistry of Minerals, Vol. 23,
No. 4, pp. 299-304.

14
Luminescence Spectroscopy

Additional spectral features are observed in the Raman spectra of mineral samples from the RRUFF
database. These features are not Raman modes, they are luminescence features. The wavelengths of
the lasers (514, 532, 780 and 785 nm) used to collect Raman spectra have excited multiple luminescence
centers in various minerals. A basic introduction to the theory of luminescence spectroscopy (based on
Gaft et al., 2005) is provided here.

When electromagnetic radiation (light) interacts with a luminescent material, an electron is excited to a
higher energy state. When the electron drops back down to the ground state it releases light of a specific
wavelength or range of wavelengths (called radiative decay) (Fig. I1, p.2). Luminescence can be excited
in minerals by radiation of many different wavelengths; luminescence induced by light sources in the UV-
visible range is called photoluminescence (this is what we are observing in the Raman spectra of this
study). Other means of exciting luminescence in minerals include excitation by a beam of electrons,
called cathodoluminescence, and excitation by heat, called thermoluminescence.

The ion responsible for a particular luminescence feature is called a luminescent center or activator.
Correlating luminescence features with their associated centers is not a straightforward task.
Simultaneous emission of multiple centers, luminescence decay, and symmetry constraints can affect
interpretation and correlation of activators to their associated features. The majority of luminescence
centers in minerals are transition elements and REE.

Luminescence spectroscopy is used to measure the energy levels of luminescence centers. Energy
levels are defined by characteristic states (ground state: lowest energy, excited states: higher energy).
Electronic and vibrational transitions are of greatest interest in mineral luminescence. Luminescence
emission occurs when an excited electron jumps from a higher energy level to a lower one, releasing a
photon. Luminescent minerals emit radiation only when the excitation energy is absorbed.

Symmetry and directional properties of orbitals determine the luminescent properties of a substance.
Every atom has a scheme of energy levels which changes when the atoms combine to form the crystal
structures of minerals. In my study, the majority of luminescence activators are transition elements and
REE. The orbitals of interest in luminescence caused by transition elements and REE are the d and f
orbitals. Crystal field and ligand theory discuss the manner in which bonded atoms are influenced by
their nearest neighbors based on electon distributions and symmetry.

For more information on crystal field or ligand theory, consult the following texts:

“Chemical Applications of Group Theory,” F.A. Cotton, 1971

“Symmetry in Bonding and Spectra: An Introduction,” B.E. Douglas and C.A. Hollings

References

Gaft, M., Reisfeld, R. & Panczer, G. (2005) Modern Luminescence Spectroscopy of Minerals and Materials, Springer,
Heidelberg, Germany.

15
Causes of Color in Minerals

The unique and variable combinations of hues, tones, and saturations of gem minerals are the most
important characteristics dictating gemstone value and demand. Here is a brief discussion of the
mechanisms producing color in gems (see also “Cause of Color Chart” in Appendix). For more detailed
information see, “The Physics and Chemistry of Color,” by Kurt Nassau.

There are five mechanisms responsible for color in gem minerals (Fritsch and Rossman, 1987):

1) Metal ions
2) Charge transfer
3) Color centers
4) Band theory
5) Physical optics

1) Metal Ions

The most common cause of color in minerals is the presence of color-inducing metal ions, sometimes
called chromophores. Color is produced by chromophores when electrons undergo transitions between
orbitals confined to a single metal ion (Fritsch and Rossman, 1988). When light interacts with a mineral,
and consequently the ions dispersed throughout, the electrons become excited jumping from the ground
state to an excited state as the ions absorb the incident energy. This excited state is unstable and the
electrons can return back to the stable ground state in two ways: 1) by releasing the energy into the
crystal lattice by atomic vibration (heat), or 2) by emitting the energy as a photon, a process called
luminescence (see also section on “Luminescence Spectroscopy”). Colors produced by metal ions are
the result of absorption of visible light. Common color-inducing ions present in gemstones include: Cu,
Ni, Cr, V, Mn, Fe, Co, and Ti (Fritsch and Rossman, 1987). Copper, for example, produces the green
color of malachite as well as the blue color of azurite. Color-inducing ions typically contain unpaired
electrons in d or f orbitals (Nassau, 2001). This is significant because unpaired electrons can be easily
excited by the energy of visible light, resulting in absorption. When ions are bonded together such that all
of the electrons are paired off, having only completely filled or completely empty shells, the paired
electrons become incredibly difficult to excite, requiring a large amount of energy to do so (UV).
Accordingly, substances with all of their electrons paired off cannot selectively absorb visible light
because visible light is not high enough in energy to excite the electron pairs and therefore, these
substances remain colorless (Nassau, 2001). An example is halite (NaCl).

The valence state or charge of an ion influences the color of a mineral due to the probability of
occurrence of certain electronic transitions, as determined by quantum mechanic rules that are beyond
2+
the scope of this study (Fritsch and Rossman, 1987). An example is the presence of Mn in beryl; Mn
3+
produces a light pink hue (morganite), while Mn produces a bright red color (Fritsch and Rossman,
1987). Other factors affecting the color of gems containing chromophores include nearest neighbor
atoms, coordination of the ion (the number of atoms the ion is bonded to), and the local symmetry (Fritsch
and Rossman, 1987). Common coordinations referred to in this paper include tetrahedral (four
neighboring atoms), octahedral (six neighboring atoms), and distorted cubic (eight neighboring atoms).
Frequently these metal ions substitute (take the place of) atoms of similar size and charge in the crystal
structure.

2) Charge Transfer

While the mechanism producing color via dispersed metal ions involves electronic transitions within a
single ion, the process of charge transfer occurs when electrons jump from one atom to another (Fritsch
and Rossman, 1988). Transfer of electrons can occur between an ion and its nearest neighbors and
even its next nearest neighbors (this process is called intervalence charge transfer). Yellow beryl
3+
(heliodor) is the result of the interaction of Fe with its nearest neighbor oxygen atoms. Intervalence
2+ 4+
charge transfer between Fe and Ti give sapphire its distinctive blue color (Fritsch and Rossman,

16
1988). Intervalence charge transfer between two different metal ions is called heteronuclear intervalence
charge transfer (Nassau, 2001). Homonuclear intervalence charge transfer derives from interactions
between two of the same metal ions, but with different valence state, for example between Fe2+ and Fe3+
(Nassau, 2001).

3) Color Centers

A defect that causes light absorption is called a color center. Defects are commonly caused by natural or
artificial radiation and include vacancies (missing atoms), extra atoms, changes in valence state of metal
ions, or extracted electrons placed into an existing defect (Fritsch and Rossman, 1988). When a color
center has one less electron than it would have had prior to irradiation, the center is called a hole center
(Nassau, 2001). Vacancies in diamond, removal of carbon atoms from the structure, produced by natural
irradiation can produce green diamonds (Fritsch and Rossman, 1988). Substitution of Al3+ for Si4+ in
quartz combined with natural irradiation produces a hole center and creates the color of smoky quartz
(Fritsch and Rossman, 1988; Nassau, 2001). In general, gemstones colored by centers tend to fade
easily with exposure to heat or light because the electrons that have been displaced during the formation
of the color center are weakly held by surrounding cations. Therefore, it does not take much energy to
free the electron from its trap. The freed electron returns to its original position, and the mineral to its
original color (Fritsch and Rossman, 1988).

4) Band Theory

A less common cause of color in gem minerals is the result of delocalized electrons interacting with visible
light (Fritsch and Rossman, 1988). The study of this interaction is called band theory and it is most
commonly used to describe the properties of metals and semiconductors. In some minerals billions of
atoms contribute to the possible energy levels of the substance producing an energy band. The low-
energy valence band is populated only by electrons and the high energy conduction band is typically
empty. The energy separating the two bands is called a band gap. Instead of transitions between energy
levels of atoms, transitions between bands produce colors in these minerals (Fritsch and Rossman,
1988). Transitions between bands occur when the electrons in the valence band absorb light that
provides a sufficient amount of energy for the electrons to jump over the band gap into the conduction
band. There are three possible transitions that can occur: 1) the incident visible light does not provide the
electrons with enough energy to jump the band gap (band gap is greater than energy of visible light), all
visible light is transmitted (none is absorbed) and the gemstone appears colorless (corundum, topaz,
quartz, quartz, diamond, beryl); 2) the energy of the band gap is in the visible range (violet, blue, green
light are absorbed) and the gemstone can appear a range of colors from deep yellow to deep red
(depending on specific energy of the band gap) i.e. the red color of cinnabar; 3) the energy of the band
gap is less than the lowest energy of the visible light resulting in total absorption (the sum of all colors is
black) and in a metallic luster due to remission of the light by the electrons; i.e. gold, copper, pyrite, and
platinum (Fritsch and Rossman, 1988). Trace amounts of impurities of atoms in these minerals can
produce energy levels that are between the valence and conduction band resulting in different colors.
Examples of this are incorporation of boron and nitrogen (Type Ib) in diamond producing blue and yellow
colors, respectively (Fritsch and Rossman, 1988).

5) Physical Optics

There are other physical mechanisms besides absorption of light that can produce various colors in
gems. Scattering, dispersion, intereference and diffraction of light can produce various optical effects in
gem materials including the play-of-colors in opal and labradorite (diffraction) (Fritsch and Rossman,
1988). These optical phenomena are not significant to the scope of this study as it relates to Raman
spectra.

17
References

Fritsch, E. & Rossman, G. (1988) An update on color in gems. Part 3: colors caused by bandgaps and physical
phenomena. Gems & GemoA logy, Vol. 24, No. 2, pp. 81-102.
Fritsch, E. & Rossman, G.R. (1988) An update on color in gems. Part 2: Colors involving multiple atoms and color
centers. Gems & Gemology, Vol. 24, No. 1, pp. 3–15.
Fritsch, E. & Rossman, G.R. (1987) An Update on Color in Gems. Part 1. Introduction and Colors caused by
Dispersed Metal Ions. Gems and Gemology, Vol. 23, No. 3, pp. 126–139.
Nassau, K. (ed) (2001) The Physics and Chemistry of Color: The Fifteen Causes of Color, 2nd edn, John Wiley and
Sons, New York.

18
Beryl
Gem names: emerald, aquamarine, heliodor,
morganite, goshenite, green beryl

Ideal chemistry: Be3Al2Si6O18

Crystal system: hexagonal

Point Group: H-M: 6/mmm S: D6h

emerald from Habachtal, Salzburg, Austria
Space Group: P6/mcc

Table of Atomic Coordinates (Hazen and Finger, 1986):

atom x y z
Be .5 0 .25
Al 1/3 2/3 .25
Si .3876 .1159 0
O1 .3103 .2369 0
O2 .4985 .1456 .1453

Raman mode analysis:

Raman Active Modes

Atom Wyckoff Position Point Symmetry A1g E2g E1g

O2 24m 1 3 6 6
Si, O1 12l m 2 4 2
Be 6f 222 - 1 2
Al 4c 32 - 1 1

(1 × 24m) + (2 × 12l) + (1 × 6f) + (1 × 4c)

Raman mode analysis predicts the existence of 36 active Raman modes in beryl:

7A1g + 16E2g + 13E1g = 36

19
Orientational Dependence of Spectra

When beryl is oriented with the a-axis parallel to the incident laser beam (the cross-section of the optical
indicatrix is an ellipse), the intensities of the peaks change dramatically with rotation due to a change in
the degree of freedom of atomic vibration (Fig. 1.1A). When the sample is oriented with the c-axis
parallel to the laser (the cross-section of the optical indicatrix is a circle), there is no noticeable change in
peak intensities due to an equal distribution of atomic vibration in all directions (FIG. 1.1B). In spite of the
orientational dependence of its Raman peaks, beryl can be accurately identified by the spectrum of a
randomly oriented sample. Note that in collecting the Raman spectra of beryl it is possible that the
Raman peak at 1100 cm-1 may vary drastically in intensity due to orientation.

FIG. 1.1 Raman spectra of oriented beryl crystals; sample R050368 (red beryl), processed

A. Raman spectra showing peak intensities as a function of orientation; 514.5 nm laser parallel to a* (through prism
face); at 0° laser is polarized ll to c-axis; at 90° laser is polarized perpendicular to c-axis

B. Raman spectra showing peak intensities as a function of orientation. Notice that the intensities do not change as
the direction of polarization changes; 514.5 nm laser parallel to c-axis; at 0° laser is polarized ll to b-axis; at 90° laser
is polarized perpendicular to b-axis

FIG. 1.2 A. B.

The crystal structure of beryl. Yellow spheres: channel-filling atoms; purple tetrahedra: BeO4 groups; green
tetrahedra: SiO4 groups; blue octahedra: AlO6 groups (A) Looking down: c-axis; (B) Looking down a-axis.

20
Spectral Features Related to H2O

As depicted in Fig. 1.2A, stacked six-membered rings of Si-tetrahedra linked by Be-tetrahedra and Al-
octahedra form channels parallel to the c-axis in beryl. These channels are commonly occupied by
cations (most commonly, but not limited to Na, K, Li, Cs), water, CO2 or some combination of the above
(Lodzinksi et al., 2005).

Channel-filling water molecules in beryl are designated either as type-I or type-II** (Wood & Nassau,
1968). These designations are orientationally dependent; type-I water molecules are oriented with the 2-
fold axis of the water molecule perpendicular to the 6-fold axis of beryl; type-II molecules are oriented with
the 2-fold axis of the water molecule parallel to the 6-fold axis of beryl (Aurisicchio et al. 1994).
Aurisicchio et al. (1994) suggest that in alkali-rich beryl, OH groups interact with alkali cations in the
channels, resulting in a dominance of type-II water, over type-I. Lodzinski et al. (2005) report that peaks
centered at 1386 and 1240 cm-1 in the Raman spectra of beryl are attributed to CO2.

OH- Peaks (based on IR data reported by Lodzinski et al., 2005):

Type-I H2O: 1598 cm-1, 3609-3606 cm-1 (stretching vibration), 3692/3696 cm-1, 3880 cm-1 (Figs. 1.3-1.4)

Type-II H2O: 1628/1634 cm-1, 3594/3597 cm-1 (stretching vibration), 3651/3657 cm-1 (Figs. 1.3-1.4)

In many experimental techniques, such as electron microprobe, OH- can be difficult to detect. Therefore,
the OH- content of a mineral is often calculated, not measured. OH- can be detected by X-ray or neutron
diffraction. However, a simpler, quicker, non-destructive method for detecting OH- in a sample is Raman
spectroscopy.

**R. I. Mashkovtsev and A. S. Lebedev (1993) proposed a third water type associated with channel-filling alkali
cations (Type-III). Type-III water is oriented along the same axis as Type-I, however, the cations are situated at a
greater distance from the water molecules producing IR active bands observed at 3705 and 1604 cm-1. Type-III
water was not observed in this study.

FIG. 1.3 Raman spectra of light-blue beryl demonstrating a type-I H2O peak centered at 3600 cm-1 and a type-II
-1
H2O peak centered at 3660 cm ; sample R050065, 532 nm laser, unoriented

Type-
I H2O

Type-II
H2O
Type-I
H2O

FIG. 1.4 (right) Magnification of a type-I H2O peak centered at 3600 cm-1
and a type-II H2O peak centered at 3660 cm-1 ; sample R050065, 532 nm
laser, unoriented, processed
Type-II
H2O

21
Spectral Features Related to Chromophores and Other Ions

Minute amounts of chromophoric (color-inducing) trace elements produce the many colors in beryl.
Colorless beryl, goshenite, has low concentrations of color-inducing atoms as compared to colored
beryls. Golden beryl, heliodor, is the result of Fe3+ substituting for either Be2+ or Al3+. A variety of color
centers responsible for the hues of light blue beryl, also known as aquamarine, are attributed to ferrous
and ferric iron, while red and pink beryl contain trace amounts of color-inducing manganese (Gaft et al.,
2005).

Distiguishing between an emerald and a green beryl is a controversial issue in gemology. The presence
of Cr3+ substituting for Al3+ in a deep green-colored beryl was the original defining characteristic of an
emerald. However, V3+ substituting for Al3+ can also produce deep green-colored beryl (Gaft et al., 2005).
Whether or not a gemstone is marketed as an emerald or simply a green beryl can greatly affect the
gemstone’s value. Luminescence studies reported by Gaft et al. (2005) attribute two distinct
luminescence centers located at 680 and 685 nm in beryl to octahedrally coordinated Cr3+ (Fig. 1.5).

FIG. 1.5 Peaks in the Raman spectra of beryl associated with Cr3+ luminescence centers located at 680 nm
532 -1 532 -1
( 4095 cm ) and 685 nm ( 4170 cm ); note that the luminescence peaks overwhelm the Raman peaks in this
spectrum; sample R060943, green beryl from Xinjiang, China, unoriented

A. Cr3+

Note: Microprobe data shows that this beryl sample contains a significant amount of vanadium, thus, some might
argue that it is not an emerald, but rather a green beryl. Although vanadium is present in the crystal, there are no
visible luminescence features associated with vanadium in this spectrum.

22
Spectra of Minerals with Inclusions

Gemstones frequently contain mineral inclusions. These inclusions provide scientists and gemologists
alike with valuable information about mineral genesis. Raman spectroscopy is a useful tool for identifying
mineral inclusions in gemstones. In an emerald from Habachtal, Salzburg, Austria, electron microprobe
analysis reveals the presence of actinolite micro-inclusions (Fig. 1.8). The resulting spectrum displays
Raman peaks belonging to both beryl and actinolite; note the doublet around 670 cm-1 and the presence
of actinolite peaks in the 150-200 cm-1 range that do not normally exist in a beryl spectrum (Fig. 1.7 &
1.9). Recognizing spectra containing Raman peaks of multiple minerals may help to distinguish between
gemstones from various localities.

FIG. 1.7 Spectrum of emerald from Habachtal, Austria with a split peak centered at 670 cm-1 demonstrating that
Raman peaks from both beryl and actinolite are present in this spectrum; sample R060944, 532 nm laser, unoriented,
processed

FIG. 1.8 (right) Polished surface of microprobe mount of an emerald crystal

from Austria; the light colored fibers are micro-inclusions of actinolite in the
emerald

FIG. 1.9 (below) Comparison of the Raman spectra of 1) Austrian emerald,

2) aquamarine R040002, 3) actinolite R040063; demonstrates that the Raman
spectrum of the Austrian emerald (1) contains peaks belonging to both beryl
and actinolite; unoriented; 532 nm laser, processed
Intensity

3.

2.

1.

Raman Shift (cm-1)

23
Fluorescence of Natural Beryl

Raman spectra of natural emerald collected with a 785 nm laser exhibit fluorescence that overwhelms the
Raman signal (Fig. 1.10A). In natural beryl of any other color, fluorescence generated by the 785 nm
incident beam is minimal (Fig. 1.10B), making this optical phenomenon diagnostic of green-colored
beryls. The presence of Cr3+ is likely the cause of this fluorescence. The Raman spectra of green-colored
beryls collected using a laser of a higher energy, such as 532 nm, minimizes this effect (Fig. 1.10A).

FIG. 1.10

A. Fluorescence in the Raman spectrum of a green-colored beryl with 785 nm laser (black spectrum) and 532 nm
laser (green spectrum); sample R060942, unoriented
Intensity

785 nm

532 nm

Raman Shift (cm-1)

B. The presence of minimal fluorescence in the Raman spectra of a colorless beryl with 785 nm laser and 532 nm
laser; sample R040002, unoriented
Intensity

532 nm
785 nm

Raman Shift (cm-1)

24
References

Aurisicchio, C., Grubessi, O. & Zecchini, P. (1994) Infrared spectroscopy and crystal chemistry of the beryl group.
Canadian Mineralogist, Vol. 32, No. 1, pp. 55.
Gaft, M., Reisfeld, R. & Panczer, G. (2005) Modern Luminescence Spectroscopy of Minerals and Materials, Springer,
Heidelberg, Germany.
Hazen, R.M., Au, A.Y. & Finger, L.W. (1986) High-pressure crystal chemistry of beryl (Be3Al2Si6O18) and euclase
(BeAlSiO4OH). American Mineralogist, Vol. 71, No. 7-8, pp. 977-984.
Łodziński, M., Sitarz, M., Stec, K., Kozanecki, M., Fojud, Z. & Jurga, S. (2005) ICP, IR, Raman, NMR investigations of
beryls from pegmatites of the Sudety Mts. Journal of Molecular Structure, Vol. 744, pp. 1005-1015.
Mashkovtsev, R. & Lebedev, A. (1993) Infrared spectroscopy of water in beryl. Journal of Structural Chemistry, Vol.
33, No. 6, pp. 930-933.
Wood, D. & Nassau, K. (1968) The characterization of beryl and emerald by visible and infrared absorption
spectroscopy. American Mineralogist, Vol. 53, No. 5/6, pp. 777-800.

Additional Information

Adams, D.M. & Gardner, I.R. (1974) Single-crystal vibrational spectra of beryl and dioptase. Journal of the Chemical
Society, Dalton Transactions, Vol. 1974, No. 14, pp. 1502-1505.
Charoy, B., de Donato, P., Barres, O. & Pinto-Coelho, C. (1996) Channel occupancy in an alkali-poor beryl from
Serra Branca (Goias, Brazil): Spectroscopic characterization. American Mineralogist, Vol. 81, No. 3-4, pp. 395-
403.
Fritsch, E. & Rossman, G. (1988) An update on color in gems. Part 3: Colors caused by band gaps and physical
phenomena. Gems & Gemology, Vol. 24, No. 2, pp. 81-102.
Fritsch, E. & Rossman, G.R. (1988) An update on color in gems. Part 2: Colors involving multiple atoms and color
centers. Gems & Gemology, Vol. 24, No. 1, pp. 3–15.
Fritsch, E. & Rossman, G.R. (1987) An Update on Color in Gems. Part 1. Introduction and Colors caused by
Dispersed Metal Ions. Gems & Gemology, Vol. 23, No. 3, pp. 126–139.
Goldman, S., Rossman, G.R. & Parkin, K.M. (1978) Channel constituents in beryl. Physics and Chemistry of
Minerals, Vol. 3, No. 3, pp. 225-235.
Hagemann, H., Lucken, A., Bill, H., Gysler-Sanz, J. & Stalder, H.A. (1990) Polarized Raman spectra of beryl and
bazzite. Physics and Chemistry of Minerals, Vol. 17, No. 5, pp. 395-401.
Hofmeister, A., Hoering, T. & Virgo, D. (1987) Vibrational spectroscopy of beryllium aluminosilicates: Heat capacity
calculations from band assignments. Physics and Chemistry of Minerals, Vol. 14, No. 3, pp. 205-224.
Johnson, E.A. (2006) Water in nominally anhydrous crustal minerals: speciation, concentration, and geologic
significance. Reviews in Mineralogy and Geochemistry, Vol. 62, No. 1, pp. 117-154.
Johnson, M.L., Elen, S. & Muhlmeister, S. (1999) On the identification of various emerald filling substances. Gems &
Gemology, Vol. 35, No. 2, pp. 82–107.
Kim, C.C., Bell, M.I. McKeown, D.A. (1995) Vibrational analysis of beryl (Be3Al2Si6O18) and its constituent ring
(Si6O18). Physica B: Physics of Condensed Matter, Vol. 205, No. 2, pp. 193-208.
Kloprogge, J.T. & Frost, R.L. (2000) Raman microscopic study at 300 and 77 K of some pegmatite minerals from the
Iveland–Evje area, Aust-Agder, Southern Norway. Spectrochimica Acta Part A: Molecular and Biomolecular
Spectroscopy, Vol. 56, No. 3, pp. 501-513.
Moroz, I., Roth, M., Boudeulle, M. & Panczer, G. (2000) Raman microspectroscopy and fluorescence of emeralds
from various deposits. Journal of Raman Spectroscopy, Vol. 31, No. 6, pp. 485-490.
Rossman, G.R. (2006) Analytical methods for measuring water in nominally anhydrous minerals. Reviews in
Mineralogy and Geochemistry, Vol. 62, No. 1, pp. 1-28.
Rossman, G.R. (1996) Studies of OH in nominally anhydrous minerals. Physics and Chemistry of Minerals, Vol. 23,
No. 4, pp. 299-304.
2+ 3+
Spinolo, G., Fontana, I. & Galli, A. (2007) Optical absorption spectra of Fe and Fe in beryl crystals. Physica Status
Solidi B, Basic Research, Vol. 244, No. 12, pp. 4363.
Taran, M.N. & Rossman, G.R. (2001) Optical spectroscopic study of tuhualite and a re-examination of the beryl,
cordierite, and osumilite spectra. American Mineralogist, Vol. 86, No. 9, pp. 973-980.

25
Chrysoberyl
Gem names: alexandrite, cat’s eye

Ideal chemistry: BeAl2O4

Crystal system: orthorhombic

Point Group: H-M: mmm S: D2h

Space Group: Pnma

Synthetic alexandrite gemstone

Table of Atomic Coordinates (Weber et al., 2007):

atom x y z
Al1 0 0 0
Al2 0.27282 0.25 -0.00503
Be 0.09289 0.25 0.43360
O1 0.09022 0.25 0.78822
O2 0.43316 0.25 0.24167
O3 0.16324 0.01554 0.25728

Raman mode analysis:

Raman Active Modes

Atom Wyckoff Position Point Symmetry Ag B1g B2g B3g

O3 8d 1 3 3 3 3
Al2, Be, O1, O2 4c m 2 1 2 1
Al1 4a 1 - - - -

(1 x 8d) + (4 x 4c) + (1 x 4a)

Raman mode analysis predicts the existence of 36 active Raman modes in chrysoberyl:

11Ag + 7B1g + 11B2g + 7B3g = 36

26
Spectral Features Related to Chromophores and Other Ions

Perhaps the most famous gem variety of chrysoberyl is alexandrite. Chrysoberyl gemstones that exhibit a
color change from green or bluish-green in daylight to purplish-red under incandescent light are called
alexandrite. This optical phenomenon also occurs in other gemstones such as sapphire, garnet, spinel,
zoisite, and fluorite, though the colors may vary (Liu et al., 1994). In alexandrite, color-change is
attributed to octahedrally coordinated Cr3+ substituting for Al3+ (Fritsch and Rossman, 1988). The
substitution of Cr3+ for Al3+ is more complicated, however, because there are two different symmetries
associated with the Al3+ sites, half of the Al3+ atoms are located at an inversion center (the Al1 site) and
half are located on a mirror plane (the Al2 site) (Gubelin, 1976; Hassan and El-Rakhawy, 1974). Cr3+ is
slightly larger than Al3+ and therefore, prefers to occupy the more spacious Al2 site (<R(Al2-O)>1.934Å).
However, it has been reported that with an increase in pressure and temperature, Cr3+ will occupy the Al1
sites (<R(Al1-O)>1.890Å) (Gubelin, 1976; Hassan and El-Rakhawy, 1974). The ratio of Cr3+ occupying
the Al1 site to Cr3+ occupying the Al2 site determines whether or not the alexandrite will exhibit color-
change. The greater the ratio of Cr3+ in Al1 to Cr3+ in Al2, the more pronouned the color-change becomes
(Gubelin, 1976; Hassan and El-Rakhawy, 1974). The alexandrite effect in gemstones is not solely
dependent on the Cr3+ content, but rather on the positions and intensities of the absorption and
transmission regions resulting from the presence of Cr3+ in various structural sites (Gubelin and
Schmetzer, 1982).

Gaft et al. (2005) report the presence of multiple luminescence centers attributed to Cr3+ substituting for
Al3+ in chrysoberyl located at 650, 655, 664, 679, 680, 693, 694, 700, 707, and 716 nm. The positions of
Cr3+ peaks in the Raman spectra of chrysoberyl in this study are centered at: 650 nm, 655 nm, 664 nm
679 nm, 680 nm, 693 nm, and 700 nm (Fig. 2.1 A-D). Gaft et al. (2005) hypothesize that a luminescence
center located at 703 nm may be attributed to V2+ (Fig. 2.1 A., B., & D.).

Fig. 2.1 A. Peaks in the Raman spectra of chrysoberyl related to Cr3+ luminescence centers centered at 650, 655,
664, 679, 680, 693, and 700 nm; note that the luminescence peaks overwhelm the Raman peaks in this spectrum;
sample R080110, Sri Lanka, λexcitation = 532 nm, unoriented

Cr3+

3+
Cr & V?
3+
Cr

27
Fig. 2.1 B. Magnification of Cr3+ peaks centered at 650 nm (5323370 cm-1), 655 nm (5323540 cm-1), 664 nm (5323745
cm-1), 679 nm (5324062 cm-1), 680 nm (5324100 cm-1), 693 nm (5324367 cm-1), and 700 nm (5324510 cm-1) and a peak
532 -1
centered at 703 nm ( 4573 cm ) possibly related to V in the Raman spectrum of chrysoberyl; sample R080110, Sri
Lanka, unoriented

Cr3+
Intensity

3+
Cr3+
Cr V?
3+
Cr

Raman Shift (cm-1)

Fig. 2.1 C. (left) Further magnification of weak Cr3+

peaks centered at 650 nm (5323370 cm-1), 655 nm
532 -1 532 -1
( 3540 cm ), and 664 nm ( 3745 cm ); sample
R080110, unoriented
Intensity

Cr3+

Cr3+

Raman Shift (cm-1) Cr3+ V?

Intensity

Cr3+

Fig. 2.1 D. (right) Further magnification of weak Cr3+

peaks centered at 693 nm (5324367 cm-1), and 700 nm (5324510
cm-1) and possibly a peak attributed to V2+ centered at 703 nm
532 -1
( 4573 cm ); sample R080110, unoriented

Raman Shift (cm-1)

28
References

Fritsch, E. & Rossman, G. (1988) An update on color in gems. Part 3: Colors caused by band gaps and physical
phenomena. Gems & Gemology, Vol. 24, No. 2, pp. 81-102.
Gaft, M., Reisfeld, R. & Panczer, G. (2005) Modern Luminescence Spectroscopy of Minerals and Materials, Springer,
Heidelberg, Germany.
Gubelin, E. (1976) Alexandrite from Lake Manyara, Tanzania. Gems & Gemology, pp. 203-209.
Gubelin, E. & Schmetzer, K. (1982) Gemstones with alexandrite effect. Gems & Gemology, Vol. 18, pp. 197-203.
Hassan, F. & El-Rakhawy, A. (1974) Chromium III centers in synthetic alexandrite. American Mineralogist, Vol. 59,
pp. 159-165.
Liu, Y., Shigley, J., Fritsch, E. & Hemphill, S. (1994) The "alexandrite effect” in gemstones. Color Research &
Application, Vol. 19, No. 3, pp. 186-191.
Weber, S.U., Grodzicki, M., Lottermoser, W., Redhammer, G.J., Tippelt, G., Ponahlo, J. & Amthauer, G. 57 Fe
Mössbauer spectroscopy, X-ray single-crystal diffractometry, and electronic structure calculations on natural
alexandrite. Physics and Chemistry of Minerals, pp. 1-9.

29
Corundum
Gem names: ruby, sapphire,
padparadascha

Ideal chemistry: Al2O3

Crystal system: trigonal

−
Point Group: H-M: 3 m S: D3d
−
Space Group: R 3 c
Synthetic sapphire

Table of Atomic Coordinates (Lewis et al., 1982):

atom x y z
Al 0 0 0.35216
O 0.30624 0 0.25

Raman mode analysis:

Raman Active Modes

Atom Wyckoff Position Point Symmetry A1g Eg

O 18e 2 1 3
Al 12c 3 1 2

Raman mode analysis predicts the existence of 7 active Raman modes in corundum:

2A1g + 5E g = 7

30
Introduction to Raman Spectrum of Corundum

The crystal structure of corundum consists of dense, closest packed layers of oxygen and octahedrally
coordinated aluminum (Fig. 3.2). In general, we have found that minerals composed solely of octahedral
polyhedra are poor Raman scatterers. A perfect octahedron has an inversion center at the cation site,
therefore, no significant change in polarization of the atom can occur. An exception to this observation is
corundum. In the structure of corundum, edge and face sharing of the Al-octahedra cause distortion,
resulting in appreciable polarization of the atoms and intense Raman peaks. The Raman spectrum of
corundum is provided below.

Fig. 3.1 The Raman spectrum of corundum with mode assignments as reported by Porto and Krishnan (1967);
sample X080003, synthetic yellow corundum, 780 nm laser, processed, unoriented

418
A1g
Intensity

378
Eg
432
Eg 645 751
578 A1g Eg
Eg
451 Eg

Raman Shift (cm-1)

Fig. 3.2 The crystal structure of corundum; blue octahedra: AlO6 groups

A. View down c-axis B. View down a-axis

31
Orientational Dependence of Spectra

Xu et al. (1995) and Porto and Krishnan (1967) report that when the incident beam is polarized parallel to
the c-axis, the following three Raman peaks in corundum disappear: 576 (Eg), 643(A1g), and 749 (Eg)
cm-1. When corundum is oriented with the a-axis parallel to the incident laser beam (the cross-section of
the optical indicatrix is an ellipse), the intensities of the peaks change dramatically with rotation due to a
change in the degree of freedom of atomic vibration (Fig. 3.3 A). When the sample is oriented with the c-
axis parallel to the laser (the cross-section of the optical indicatrix is a circle), there is no noticeable
change in peak intensities due to an equal distribution of atomic vibration in all directions (FIG. 3.3 B). In
spite of the orientational dependence of its Raman peaks, corundum can be accurately identified by the
spectrum of a randomly oriented sample.

Fig. 3.3. Raman spectra of oriented corundum crystals; sample R040096, purple corundum, Sri Lanka, processed

A. Raman spectra showing peak intensities as a function of orientation; 514.5 nm laser parallel to a*; at 0° laser is
polarized ll to c-axis; at 90° laser is polarized perpendicular to c-axis, sample R040096, processed

B. Raman spectra showing peak intensities as a function of orientation. Notice that the intensities do not change as
the direction of polarization changes; 514.5 nm laser parallel to c-axis; at 0° laser is polarized ll to a-axis; at 90° laser
is polarized perpendicular to a-axis

32
Spectral Features Related to OH-

An OH- band centered at 3310 cm-1 was recently reported by Beran and Rossman (2006) in natural
corundum from worldwide localities during an IR study. This band is most often found in blue-colored
corundum (sapphire) suggesting that the OH- is involved in a chemical reaction with Fe in the structure.
Prior to this study, OH- bands in corundum were rarely observed in natural crystals, appearing instead in
synthetic material (Beran and Rossman, 2006). There are no observable peaks associated with OH-
modes in the Raman spectra of corundum in this study.

Spectral Features Related to Chromophores and Other Ions

Gem quality corundum comes in a wide variety of colors, the mechanisms of which vary (Appendix D). In
blue sapphires a charge transfer between Fe2+-O-Ti4+ in combination with charge transfer between ferric
and ferrous iron create the diagnostic sapphire color. In rubies, octahedrally coordinated Cr3+ is the
dominant color-inducing ion (Smith et al., 1997; Fritsch and Rossman, 1988). Trace amounts V3+ and
Fe3+ may offer minor contributions to the red color in some rubies (Fritsch and Rossman, 1988). The
positions of Cr3+ luminescence centers located at 693 nm and 694 nm, denoted R2 and R1 respectively,
have been well established in the literature (Fig. 3.4 A. & B.) (Collins et al., 1960). The positions of the
peaks in this doublet shift systematically with pressure (Collins et al., 1960; Mao et al., 1978). Therefore,
the positions of Cr3+ peaks have been calibrated to determine pressures in diamond anvil cells (Mao et
al., 1978).

Fig. 3.4 A. Peaks in the Raman spectrum of corundum associated with Cr3+ luminescence centers located at 693
532 -1 532 -1
nm ( 4366 cm ) (R2) and 694 nm ( 4396 cm ) (R1); note that the luminescence peaks overwhelm the Raman
peaks in this spectrum; sample R060020, colorless sapphire from Montana, unoriented

R1

Cr3+
R2

33
Fig. 3.4 B. Magnification of peaks in Raman spectrum of corundum associated with Cr3+ luminescence centers;
sample R060020, colorless sapphire from Montana, λexcitation = 532 nm, unoriented

R1

Cr3+

R2
Intensity

Raman Shift (cm-1)

References

Beran, A. & Rossman, G.R. (2006) OH in naturally occurring corundum. European Journal of Mineralogy, Vol. 18, No.
4, pp. 441.
Collins, R.J., Nelson, D.F., Schawlow, A.L., Bond, W., Garrett, C.G.B. & Kaiser, W. (1960) Coherence, narrowing,
directionality, and relaxation oscillations in the light emission from ruby. Physical Review Letters, Vol. 5, No. 7,
pp. 303-305.
Gaft, M., Reisfeld, R. & Panczer, G. (2005) Modern Luminescence Spectroscopy of Minerals and Materials, Springer,
Heidelberg, Germany.
Fritsch, E. & Rossman, G. (1988) An update on color in gems. Part 3: Colors caused by band gaps and physical
phenomena. Gems & Gemology, Vol. 24, No. 2, pp. 81-102.
Lewis, J., Schwarzenbach, D. & Flack, H.D. (1982) Electric field gradients and charge density in corundum, Al2O3 .
Crystal Physics, Diffraction, Theoretical and General Crystallography, Vol. 38, No. 5, pp. 7394.
Mao, H.K., Bell, P.M., Shaner, J.W. & Steinberg, D.J. (1978) Specific volume measurements of Cu, Mo, Pd, and Ag
and calibration of the ruby R fluorescence pressure gauge from 0.06 to 1 Mbar. Journal of Applied Physics, Vol.
49, pp. 3276.
Smith, C.P., Gübelin, E.J., Bassett, A.M. & Manandhar, M.N. (1997) Rubies and fancy-color sapphires from Nepal.
Gems & Gemology, Vol. 33, No. 1, pp. 24–41.
Xu, J., Huang, E., Lin, J. & Xu, L.Y. (1995) Raman study at high pressure and the thermodynamic properties of
corundum; application of Kieffer's model. American Mineralogist, Vol. 80, No. 11-12, pp. 1157-1165.

Additional Information

Fritsch, E. & Rossman, G.R. (1988) An update on color in gems. Part 2: Colors involving multiple atoms and color
centers. Gems & Gemology, Vol. 24, No. 1, pp. 3–15.
Fritsch, E. & Rossman, G.R. (1987) An Update on Color in Gems. Part 1. Introduction and Colors caused by
Dispersed Metal Ions. Gems and Gemology, Vol. 23, No. 3, pp. 126–139.
Mao, H.K. & Bell, P.M. (1976) High-Pressure Physics: The 1-Megabar Mark on the Ruby R1 Static Pressure Scale.
Science, Vol. 191, No. 4229, pp. 851-852.
Porto, S.P.S. & Krishnan, R.S. (2004) Raman Effect of Corundum. The Journal of chemical physics, Vol. 47, pp.
1009.
Szabo, A. (1970) Laser-induced fluorescence-line narrowing in ruby. Physical Review Letters, Vol. 25, No. 14, pp.
924-926.

34
Diamond
Gem names: canary, cape

Ideal chemistry: C

Crystal system: isometric

Point Group: H-M: m3m S: Oh

Space Group: Fd3m

Table of Atomic Coordinates (Wyckoff, 1963):

atom x y z
C 0 0 0

Raman mode analysis:

Raman Active Modes

Atom Wyckoff Position Point Symmetry T2g

C 8a 4 3m 1

Raman mode analysis predicts the existence of 1 active Raman mode in diamond:

1T2g = 1

35
Raman Features of Diamond

Diamonds are likely the most valuable of all the gemstones. The covalently bonded carbon atoms in the
face-centered cubic crystal structure of diamond (Fig. 4.2) make it the hardest natural substance on
Earth. For this reason, diamonds are commonly used as an industrial abrasive. The hardness of
diamond, combined with its rarity and optical properties, also make it a popular gemstone.

Spectra of diamond have one diagnostic active Raman mode, the carbon-carbon stretching mode, that
produces a peak centered at 1332 cm-1 (Knight and White, 1989). Slight shifts in the frequency of the
primary diamond peak occur depending on several factors including origin of the diamond and perfection
of the diamond lattice (Fig. 4.1) (Huong, 1992). Frequency and width of this peak are also temperature
dependent (Fig. 4.3) (Liu et al., 2000). In addition, isotopic effects can result in a frequency shift of the
primary diamond peak. According to Chrenko (1988), natural diamonds typically contain 1 at.% of 13C.
To investigate isotopic effects on the primary peak in the Raman spectra of diamond, several diamonds
containing up to 91at.% 13C were grown by the temperature gradient method (Chrenko, 1988). In all
cases, there was a single Raman peak in each spectrum, however, the location of this peak shifted from
1332.5 cm-1 at 1% 13C to 1288.7 cm-1 at 91% 13C. The absence of a second peak suggests that in this
experiment the carbon isotopes were distributed homogenously throughout the diamond structure
(Chrenko, 1988). In experimentally synthesized 12C:13C diamond films grown on natural 12C diamond
substrates, two 1st order Raman peaks were observed (Behr et al., 1993). In the synthesized pure 13C
film, in addition to the peak at 1333 cm-1 associated with the 12C substrate, a peak attributed to 13C
appeared at 1283 cm-1 (Behr et al., 1993). In the 50%12C/50%13C film, the 13C peak shifted to 1313 cm-1
(Behr et al., 1993). 12C and 13C spectral differences have not been observed in this study.

Additional weaker peaks at 1817, 1864, 2025, 2177, 2254, 2333, 2458, 2519, and 2667 cm-1 represent
second-order Raman scattering (Fig. 4.4 A. & B.) (Solin and Ramdas, 1970). Analysis of these peaks is
complex however, increasing the 13C content of diamond results in a frequency shift of these peaks
suggesting that these peaks are related to carbon and not to trace impurities (Chrenko, 1988). Very weak
peaks centered at 3300 and 3825 cm-1 represent third-order Raman scattering (Fig. 4.4 A. & 4.5)
(Bormett et al., 1995). In this study, no visible differences were observed in the Raman spectra of natural
and synthetic diamonds (Fig. 4.6).

The shift in frequency of a Raman peak due to isotopic effects can be predicted using a simple equation.
Gillet et al. (1996) used the following calculation to predict the positions of Raman peaks in calcite
containing various amounts of 16O and 18O:
v1/v1* = √ m16O/ m18O
v1= theoretical peak postion; v1*= actual peak position; m = mass of the isotopes

This equation can also be used to predict the peak position of the primary Raman peak in diamonds
composed of 13C. For example, the mass of 12C is 12 and the mass of 13C is 13. When you plug these
values into the equation above you get: v1/1332 cm-1 = 0.96 and so the theoretical peak location in
diamond composed of 100% 13C is v1 = 1278 cm-1. This calculated position is very close to the peak
position of diamond composed of 91% 13C centered at 1288.7 cm-1 reported by Chrenko (1988).
5.

Fig. 4.1 (right) Slight shift in the frequency of 5.

the primary Raman peak in diamond; 1) natural
4.
type IIa diamond, sample R050204, 2) natural
type IIa diamond, sample R050206, 3)
Intensity

synthetic yellow diamond, sample X080011;

4) natural type I diamond, sample R050207;
5) natural type IIa diamond, sample R050205;
λexcitation = 532 nm, background corrected, 3.
unoriented
2.

1.

Raman Shift (cm-1)

36
Fig. 4.2 The face-centered cubic crystal structure of diamond; blue spheres: carbon atoms

Fig. 4.3 (left) Shift in frequency and change in width of primary Raman
Intensity

-1
peak in diamond when cooled to liquid nitrogen temperature; yellow synthetic
2. 1334 cm
diamond, sample X080011, λexcitation = 532 nm, unoriented; 1) room
temperature, 2) liquid nitrogen temperature

1. 1336 cm-1

-1
Raman Shift (cm )

Fig. 4.4 A. Second order Raman scattering and third order Raman scattering (not visible in this figure); natural type
IIa diamond; sample R050205; λexcitation = 532 nm, unoriented

rd
3 Order Raman
nd
2 Order Raman Scattering
Scattering

37
Fig. 4.4 B. Magnification of second order Raman scattering; natural type IIa diamond; sample R050205; λexcitation =
532 nm, unoriented

Intensity

-1 -1
2465 cm 2667 cm

-1
2030 cm

Raman Shift (cm-1)

Fig. 4.5 Magnification of third order Raman scattering; natural type IIa diamond; sample R050206; λexcitation = 532
nm, unoriented
Intensity

-1 -1
3300 cm 3825 cm

Raman Shift (cm-1)

Fig. 4.6 Spectra of natural and synthetic diamond showing negligible spectral differences; 1) synthetic diamond,
sample X080011; 2) natural diamond, sample R050207; λexcitation = 532 nm, processed, unoriented
Intensity

2.
1.

Raman Shift (cm-1)

38
Causes of Color in Diamond

Diamonds are divided into two types depending on the dominant chromophores present: Type I or Type II
(Deljanin and Simic, 2007).

Type I diamonds-

Type I diamonds are subdivided based on the way nitrogen is dispersed throughout the crystal structure.
Type Ia diamonds contain aggregates of nitrogen: IaA contain a pair of nitrogen atoms (A-center); IaB
contain four nitrogen atoms surrounding a vacancy (B-center); IaAB contain a mixture of both B-centers
and A-centers. 97% of all diamonds are Type Ia. They can appear colorless, near-colorless, yellow, or
brown. Type Ib diamonds account for less than 1% of all diamonds. Diamonds of this type contain
isolated atoms of nitrogen that replace carbon atoms (C-centers). Type Ib diamonds are yellow, orange,
or brown (Deljanin and Simic, 2007).

Type II diamonds-

Type II diamonds are generally absent of nitrogen (containing less than 1ppm). Type II diamonds make
up approximately 2% of all diamonds. Diamonds containing no nitrogen and no boron are called Type IIa.
They can be colorless, near-colorless, brown, and pink. Type IIb diamonds contain boron. When boron
replaces carbon in the diamond lattice the nonequivalent charge creates a hole in the diamond lattice
allowing for Type IIb diamonds to conduct a positive electrical charge. Type IIb diamonds are blue, gray,
light brown, or near-colorless (Deljanin and Simic, 2007).

Non-nitrogen Related Colors:

Pink/red: Plastic deformation of the diamond lattice can result in displacement of the carbon atoms along
glide planes. This is the cause of both pink and red colors in diamond (Fritsch et al., 2007b; Shigley,
1993).

Purple: Like pink diamonds, purple diamonds are also believed to be the result of plastic deformation.
The details of color-inducing defects in purple diamonds are still being explored (Titkov et al., 2008).

Brown: Brown diamonds typically contain nitrogen impurities and graining associated with deformation of
the diamond lattice (Massi, 2005).

Black: Black to gray diamonds contain micro-inclusions of dark minerals such as graphite, magnetite,
hematite, and native iron (Titkov et al., 2003).

Green: Green colored diamonds are produced by irradiation. Surface “staining” of natural diamonds is
commonly caused by the interaction of radioactive alpha and beta particles with the diamond surface,
producing a green-colored “skin” on the diamond (Kane et al., 1990). The historic Dresden green
diamond has green color throughout the body of the stone and not just at the surface therefore the color
of this famous gemstone is likely the result of deeper penetrating ionizing radiation. Exposure to this type
of radiation forced some of the carbon atoms out of their sites producing vacancies in the diamond
structure. These vacancies produce what is known as the GR1 color center. This center absorbs light
from the red portion of the spectrum, resulting in a diamond that is green (Kane et al., 1990).

Chameleon Diamonds: As the name suggests, chameleon diamonds change color, from grey-green to
yellow, with a change in temperature (thermochromic behavior) or with change in light exposure
(photochromic). These diamonds contain high concentrations of hydrogen along with some nitrogen and
nickel. Chameleon diamonds are type IaA/B and the suggested model for the color-change involves an
electron trap caused by the interaction of hydrogen atoms with A-aggregate nitrogen (Fritsch et al.,
2007a).

39
Spectral Artifacts Related to Nitrogen in Synthetic Diamonds

The 638 nm center, or NV- center, is the result of an isolated nitrogen atom trapping a vacancy in the
diamond structure (Zaitsev, 2001). An isolated nitrogen atom substituting for a carbon atom is bound to
the nearest vacancy filled with a single electron resulting in the nitrogen atom relaxing away from the
vacancy at a distance equal to approximately 8% of the normal C-C bond (Zaitsev, 2001; Loubser and
van Wyk, 1978; Mainwood, 1994). This center occurs naturally in any nitrogen-containing diamonds that
have been irradiated with high-energy ions, however, it is especially noticeable in type Ib diamonds
(Zaitsev, 2001; Nishida et al., 1989). In this study, it has only been observed in synthetic diamonds (Fig.
4.7). At liquid nitrogen temperature the 638 nm peak becomes more resolved (Fig. 4.8). The weaker
peaks accompanying the 638 nm peak are also related to the NV- center (Zaitsev, 2001).

A note on HPHT and GE POL treatments:

Fisher and Spits (2000) report that diamonds exposed to HPHT (high-pressure and high-temperature)
treatment commonly display the 638 nm peak and its neutrally charged counterpart at 575 nm. They add
that the presence of isolated nitrogen in type IIa diamonds is indicative of HPHT treatment, a technique
used to change the color of undesirable diamonds. Isolated nitrogen atoms present in GE POL diamonds
(natural diamonds enhanced by General Electric using undisclosed methods of HPHT and annealing
techniques) were likely generated from the disaggregation of A- and B-centers at temperatures above
1960oC and 2240oC respectively (Fisher and Spits, 2000). The presence of the luminescence features
described above provides valuable information about synthetic diamonds and further investigation of
these features may lead to the development of a method to distinguish between synthesized, treated, and
natural diamonds.

Fig. 4.7 A comparison of the spectral features related to the 638 nm (5323100 cm-1) NV- center in the Raman
spectra of a natural diamond and two synthetic diamonds; 1) natural type IIa diamond, sample R050205; 2) synthetic
yellow diamond, sample X080011; 3) DeBeers Element6 type IIa synthetic diamond, sample X07001; λexcitation = 532
nm, room temperature, unoriented

3.
Intensity

NV-
2.

1.

Raman Shift (cm-1)

40
Fig. 4.8 The NV- center (638 nm) at 1) room temperature and at 2) liquid nitrogen temperature; synthetic yellow
diamond, sample X080011; λexcitation = 532 nm, unoriented

Intensity

NV-

Raman Shift (cm-1)

Diamond Simulants

Common diamond simulants such as colorless zircon, cubic zirconia, yttrium aluminum garnet (YAG),
gadolinium gallium garnet (GGG), and strontium titanate are easily identified by Raman spectroscopy
(Fig. 4.9).

Fig. 4.9 Comparison of the Raman spectra of natural diamond with the spectra of common diamond simulants. 1)
Diamond, sample R050204; 2) Zircon, sample R050488; 3) YAG, sample X090003; 4) GGG, sample X090005; 5)
GGG, sample X090007; 6) Cubic zirconia, sample R040142; 7) Strontium titanate (synthetic tausonite); sample
X090004; λexcitation = 532 nm, processed, unoriented

7.

6.
Intensity

5.

4.
3.
2.

1.
-1
Raman Shift (cm )

41
References
13
Behr, D., Wagner, J., Wild, C. & Koidl, P. (1993) Homoepitaxial C diamond films studied by micro-Raman and
photoluminescence spectroscopy. Applied Physics Letters, Vol. 63, No. 22, pp. 3005-3007.
Bormett, R.W., Asher, S.A., Witowski, R.E., Partlow, W.D., Lizewski, R. & Pettit, F. (1995) Ultraviolet Raman
spectroscopy characterizes chemical vapor deposition diamond film growth and oxidation. Journal of Applied
Physics, Vol. 77, pp. 5916.
Chrenko, R.M. (1988) 13C-doped diamond: Raman spectra. Journal of Applied Physics, Vol. 63, No. 12, pp. 5873-
5875.
Deljanin, B. & Simic, D. (2007) Laboratory-Grown Diamonds: Information Guide to HPHT-grown and CVD-grown
Diamonds, 2nd ed, Gemology Headquarters International, India.
Fisher, D. & Spits, R.A. (2000) Spectroscopic evidence of GE POL HPHT-treated natural type IIA diamonds. Gems &
Gemology, Vol. 36, No. 1, pp. 42-49.
Fritsch, E., Massi, L., Rossman, G.R., Hainschwang, T., Jobic, S. & Dessapt, R. (2007a) Thermochromic and
photochromic behaviour of “chameleon” diamonds. Diamond & Related Materials, Vol. 16, No. 2, pp. 401-408.
Fritsch, E., Rondeau, B., Hainschwang, T. & Quellier, M.H. (2007b) A contribution to the understanding of pink color
in diamond: The unique, historical Grand Condé. Diamond & Related Materials, Vol. 16, No. 8, pp. 1471-1474.
Gillet, P., McMillan, P., Schott, J., Badro, J. & Grzechnik, A. (1996) Thermodynamic properties and isotopic
18
fractionation of calcite from vibrational spectroscopy of O-substituted calcite. Geochimica et Cosmochimica
Acta, Vol. 60, No. 18, pp. 3471-3485.
Huong, P.V. (1992) Diamond and diamond simulants as studied by micro-Raman spectroscopy. Materials science &
engineering.B, Solid-state materials for advanced technology, Vol. 11, No. 1-4, pp. 235-242.
Kane, R.E., McClure, S.F. & Menzhausen, J. (1990) The legendary Dresden green diamond. Gems & Gemology, Vol.
26, No. 4, pp. 248–266.
Knight, D.S. & White, W.B. (1989) Characterization of diamond films by Raman spectroscopy. Journal of Materials
Research, Vol. 4, No. 2, pp. 385-393.
Liu, M.S., Bursill, L.A., Prawer, S. & Beserman, R. (2000) Temperature dependence of the first-order Raman phonon
line of diamond. Physical Review B, Vol. 61, No. 5, pp. 3391-3395.
Loubser, J.H.N. & van Wyk, J.A. (1978) Electron spin resonance in the study of diamond. Rep.Prog.Phys, Vol. 41,
No. 8, pp. 1201.
Mainwood, A. (1994) Nitrogen and nitrogen-vacancy complexes and their formation in diamond. Physical Review B,
Vol. 49, No. 12, pp. 7934-7940.
Massi, L., Fritsch, E., Collins, A.T., Hainschwang, T. & Notari, F. (2005) The “amber centres” and their relation to the
brown colour in diamond. Diamond & Related Materials, Vol. 14, No. 10, pp. 1623-1629.
Nishida, Y., Mita, Y., Okuda, S., Mihara, T., Kato, R., Ashida, M., Sato, S. & Yazu, S. (1990) Color centers in
synthetic Ib diamonds and their application to opto-electronics in Science and Technology of New Diamond,
eds. S. Saito, O. Fukunaga & M. Yoshikawa, KTK Scientific, , pp. 363-367.
Shigley, J.E. & Fritsch, E. (1993) A notable red-brown diamond. Journal of Gemmology, Vol. 23, No. 5, pp. 259–266.
Solin, S.A. & Ramdas, A.K. (1970) Raman Spectrum of Diamond. Physical Review B, Vol. 1, No. 4, pp. 1687-1698.
Titkov, S.V., Shigley, J.E., Breeding, C.M., Mineeva, R.M., Zudin, N.G. & Sergeev, A.M. (2008) Natural-Color Purple
Diamonds from Siberia. Gems & Gemology, Vol. 44, No. 1, pp. 56.
Titkov, S.V., Zudin, N.G., Gorshkov, A.I., Sivtsov, A.V. & Magazina, L.O. (2003) An investigation into the cause of
color in natural black diamonds from Siberia. Gems and Gemology, Vol. 39, No. 3, pp. 200–209.
Wyckoff, R.W.G. (1963) Crystal Structures. Vol. 1, Interscience Publishers.
Zaitsev, A.M. (2001) Optical properties of diamond, Springer New York.

Additional Information

Collins, A.T. (1980) Vacancy enhanced aggregation of nitrogen in diamond. J.Phys.C: Solid St.Phys, Vol. 13, pp.
2641-2650.
Collins, A. (1982) Colour centres in diamond. Journal of Gemmology, Vol. 18, pp. 37-75.
King, J.M., Shigley, J.E., Gelb, T.H., Guhin, S.S., Hall, M. & Wang, W. (2005) Characterization and Grading of
Natural-Color Yellow Diamonds. Gems & Gemology, Vol. 41, No. 2, pp. 88.
McNamara, K.M., Gleason, K.K., Vestyck, D.J. & Butler, J.E. (1992) Evaluation of Diamond Films by Nuclear
Magnetic Resonance and Raman Spectroscopy. Diamond & Related Materials, Vol. 1.
Moss, T.M., King, J.M., Wang, W. & Shigley, J.E. (2002) A highly unusual, 7.34 ct, Fancy vivid purple diamond.
Journal of Gemmology, Vol. 28, No. 1, pp. 7-12.

42
Diopside
Gem names: chrome diopside

Ideal chemistry: CaMgSi2O6

Crystal system: monoclinic

Point Group: H-M: 2/m S: C2h

Space Group: C2/c Chrome diopside, Photo courtesy of Stone Group Labs

Table of Atomic Coordinates (Redhammer, 1998):

atom x y z
MgM1 0 0.9078 0.25
CaM2 0 0.3014 0.25
SiT 0.2860 0.0916 0.2302
O1 0.1184 0.0871 0.1432
O2 0.3623 0.2498 0.3175
O3 0.3487 0.0201 0.9983

Raman mode analysis:

Raman Active Modes

Atom Wyckoff Position Point Symmetry Ag Bg

Si, O1, O2, O3 8f 1 3 3

Mg, Ca 4e 2 1 2

(4 x 8f) + (2 x 4e)

Raman mode analysis predicts the existence of 30 active Raman modes in diopside:

14Ag + 16B g = 30

43
Introduction to the Raman Spectrum of Diopside

Diopside is an inosilicate consisting of single chains of SiO4 groups (Fig. 5.2). Diopside is a pyroxene and
is commonly found in ultramafic igneous and metamorphic rocks. Phase equilibrium studies indicate that
diopside is stable at temperature and pressure conditions similar to those of the upper mantle (Swamy et
al., 1997). The Raman spectrum of diopside is presented below (Fig. 5.1 A-C). The most intense peak in
the spectrum, centered at 1014 cm-1, is associated with Si-O stretching vibrations with non-bridging
oxygen atoms (Richet et al., 1997). Swamy et al. (1997) report that the frequencies of Raman peaks in
diopside decrease with an increase in temperature. Richet et al. (1997) report that increases in
temperature also affect the peak width. The peaks centered at ~600 cm-1, associated with Si-O-Si
bending modes, widen with increases in temperature (Richet et al., 1997).

Fig. 5.1 A. The Raman spectrum of diopside with mode assignments reported by Swamy et al. (1997); sample
R060171, 532 nm laser, processed, unoriented

1014
Ag
667
Ag
Intensity

139 323
Ag Ag 390
Ag
367
Bg See Fig. 5.1 B
&
See Fig. 5.1 C
357 856
Ag 508 530 710 917 1048
Ag
Ag Ag Bg Bg Bg
?

Raman Shift (cm-1)

Fig. 5.1 B. Magnification of Raman peaks in 460-580 cm-1 range; mode assignments reported by Swamy et al.
(1997) sample R060171, 532 nm laser, processed, unoriented

515
Bg
530
508 Ag
Ag
Intensity

465 560
Bg Bg

Raman Shift (cm-1)

44
Fig. 5.1 C. Magnification of Raman peaks in 160-300 cm-1 range; mode assignments reported by Swamy et al.
(1997) sample R060171, 532 nm laser, processed, unoriented

181
Ag
254 301
Ag Bg
Intensity

194
Bg
163 229
Bg Bg

Raman Shift (cm-1)

Fig. 5.2 The crystal structure of diopside; green octahedra: MgO6 groups, blue tetrahedra: SiO4 groups, orange
ellipsoids: Ca

45
Spectral Features Related to Chromophores and Other Ions

Gem quality diopside is most commonly green or yellowish green though the mechanisms responsible for
these colors can vary. Octahedrally coordinated Cr3+, V3+, a combination of Cr and V, or a charge
transfer between Fe2+-Fe3+ can produce green color in diopside (Fritsch and Rossman, 1988; Andrut et
al., 2003). A rare purple color in diopside is generated by charge transfer between Fe2+ and Ti4+ (Herd et
al., 2000). Comparison of the peak positions of spectral features in diopside, centered at 671 nm, 679
nm, 684 nm, 700 nm and 722 nm (Fig. 5.3 A. & B.), with the positions of luminescence centers in the
spectra of other minerals studied by Gaft et al. (2005) suggests that these peaks are likely related to Cr
and V. This has yet to be confirmed by experimental luminescence studies.

Fig. 5.3 A. Peaks centered at 671 nm (5323900 cm-1), 679 nm (5324082 cm-1), 684 nm (5324196 cm-1), 700 nm
(5324526 cm-1), and 722 nm (5324958 cm-1) in the Raman spectrum of diopside likely related to Cr and V luminescence
centers; note that the luminescence peaks overwhelm the Raman peaks in this spectrum; sample R060085, green
diopside, Tyrol, Austria, unoriented

Cr?

V?

Cr?

Fig. 5.3 B. Magnification of features likely related to Cr and V luminescence centers in the Raman spectrum of
diopside; sample R060085, green diopside, Tyrol, Austria, λexcitation = 532 nm, unoriented

Cr?
Intensity

V?

Cr?

Raman Shift (cm-1)

46
References

Andrut, M., Brandstätter, F. & Beran, A. (2003) Trace hydrogen zoning in diopside. Mineralogy and Petrology, Vol.
78, No. 3, pp. 231-241.
Fritsch, E. & Rossman, G. (1988) An update on color in gems. Part 3: Colors caused by band gaps and physical
phenomena. Gems & Gemology, Vol. 24, No. 2, pp. 81-102.
Fritz, E.A., Laurs, B.M., Downs, R.T. & Costin, G. (2007) Yellowish green diopside and tremolite from Merelani,
Tanzania. GEMS AND GEMOLOGY, Vol. 43, No. 2, pp. 146.
Gaft, M., Reisfeld, R. & Panczer, G. (2005) Modern Luminescence Spectroscopy of Minerals and Materials, Springer,
Heidelberg, Germany.
Herd, C.D.K., Peterson, R.C. & Rossman, G.R. (2000) Violet-colored diopside from southern Baffin Island, Nunavut,
Canada. Canadian Mineralogist, Vol. 38, No. 5, pp. 1193.
Redhammer, G.J. (1998) Mossbauer spectroscopy and Rietveld refinement on synthetic ferri-Tschermak's molecule
CaFe3+(Fe3+Si)O6 substituted diopside. European journal of mineralogy, Vol. 10, No. 3, pp. 439-452.
Richet, P., Mysen, B.O. & Ingrin, J. (1998) High-temperature X-ray diffraction and Raman spectroscopy of diopside
and pseudowollastonite. Physics and Chemistry of Minerals, Vol. 25, No. 6, pp. 401-414.
Swamy, V., Dubrovinsky, L.S. & Matsui, M. (1997) High-temperature Raman spectroscopy and quasi-harmonic lattice
dynamic simulation of diopside. Physics and Chemistry of Minerals, Vol. 24, No. 6, pp. 440-446.

47
Garnet
General Formula: X3Z2(SiO4)3

Crystal system: isometric

Gem names: pyrope, almandine, grossularite,

spessartite, andradite (demantoid), uvarovite,
rhodolite, tsavorite, hessonite
−
Point Group: H-M: m 3 m S: Oh
Spessartine, Fujian Province,China
Space Group: Ia3d

Table of Garnet Chemistries

Garnet Variety Ideal Chemical Formula

pyrope Mg3Al2(SiO4)3

almandine Fe2+3Al2(SiO4)3

grossular Ca3Al2(SiO4)3

spessartine Mn2+3Al2(SiO4)3

andradite Ca3Fe3+2(SiO4)3

uvarovite Ca3Cr2(SiO4)3

Note: The members of the garnet group are isostructural (they have the same crystal structure), therefore, atomic
coordinates and Raman mode analysis only for pyrope (the most common gem garnet) have been provided because
its structure is representative of the other garnet species.

Table of Atomic Coordinates (Merli et al., 2000):

atom x y z
Mg 0.125 0 0.25
Al 0 0 0
Si 0.375 0 0.25
O 0.033 0.0503 0.6533

48
Raman mode analysis:

Raman Active Modes

Atom Wyckoff Position A1g Eg T2g

O 96h 3 6 9
Si 24d - 1 3
Mg 24c - 1 2
Al 16a - - -

Raman mode analysis predicts the existence of 25 active Raman modes in all garnet species:

3A1g + 8E g + 14T2g = 25

49
Garnet Chemistry

The garnet group consists of multiple isostructural species that are chemically distinguishable from one
another based on a variety of possible atomic substitutions in the X (divalent dodecahedral) and Y
(trivalent octahedral) sites of the crystal structure (see “Table of Garnet Chemistries” on previous page)
(Novak and Gibbs, 1971). Pyrope, almandine, and spessartine are three end-member species of a
garnet solid-solution series and are denoted, chemically separate from the other garnet species, as
‘pyralspite’ (Winchell, 1933). The calcium-rich garnet species uvarovite, grossular, and andradite are
classified as ‘ugrandite’ (Winchell, 1931). Rarely are crystals of pure end-member chemistry found in
nature, therefore, garnets are frequently described as having multiple chemical components (Novak and
Gibbs, 1971).

A recent study conducted by Henderson (2009) at the University of Arizona has demonstrated that the
Raman spectra of garnets can be used to calculate their crystal chemistry. Raman spectroscopy of
garnets can be utilized as a quick, non-destructive alternative technique to microprobe analysis for
constraining the chemical composition of a sample. Using a correlation matrix to compare shifts in the
frequency of six peaks associated with stretching modes in the Raman spectra of garnet to the change in
chemical composition based on microprobe data, the chemistry of multiple garnet varieties can be
predicted to within 5% (overall error of bulk composition) of the microprobe values (Fig. 6). Henderson
(2009) also noted that reported a correlation between chemistry and peak intensity: in Raman spectra of
calcic garnets (those containing >50% Ca in the X-site, such as the ugrandites) the most intense peak is
-1
peak 6 (centered at ~350 cm ), while in garnets containing less than 50% Ca (pyralspites), the most
intense peak is peak 2 (centered at ~850 cm-1) (Fig. 6). Note: the octahedral (Y) site sits on a center of
inversion and therefore, there are no Raman peaks associated with the vibrations of atoms occupying this
site (Hofmeister and Chopelas, 1991).

Fig. 6 Raman spectra of end-member garnets; the six labeled peaks are the ones used in the chemical composition
calculation (reproduced with permission from Henderson, 2009)

50
Cause of Color

Garnets come in a wide variety of colors and a discussion of the various causes of color is provided
below (see also Table of Color Causes, Appendix).

Pyrope (Mg3Al2(SiO4)3): Brown-red pyropes contain Fe2+ while pure red pyropes typically contain a
combination of Fe2+ and octahedrally coordinated Cr3+ (Fritsch and Rossman, 1988; Manning, 1967).

Almandine (Fe2+3Al2(SiO4)3): The red color-inducing cation in almandine is Fe2+ and a garnet containing
components of both pyrope and almandine, well-known to gemologists as ‘rhodolite’, also owes its
distinctive reddish purple color to Fe2+ (Fritsch and Rossman, 1988; Manning, 1967).

Spessartine (Mn2+3Al2(SiO4)3): The orange color of spessartine is produced by the presence of Mn2+ in
distorted cubic coordination (Fritsch and Rossman, 1988; Gubelin, 1982; Manning, 1967).

Uvarovite (Ca3Cr2(SiO4)3): The intense green color of uvarovite is due to the presence of the common
chromophore, octahedrally coordinated Cr3+ (Fritsch and Rossman, 1988; Manning, 1969).

Grossular (Ca3Al2(SiO4)3): The green variety of grossular, commonly called tsavorite, contains
octahedrally coordinated V3+, while orange grossular (also known as hessonite) typically contains Mn2+ or
Fe2+ (Fritsch and Rossman, 1988; Manning, 1970).

Andradite (Ca3Fe3+2(SiO4)3): Yellow-green andradite contains Fe3+ in the octahedral site and the green
variety called demantoid contains octahedrally coordinated Cr3+ (Fritsch and Rossman, 1988).

Color-change garnets: Several garnet species exhibit a color-change, or alexandrite effect (Carstens,
1973; Gubelin, 1982; Schmetzer and Bernhardt, 1999). According to Schmetzer and Bernhardt (1999),
there are two divisible groups of color-change garnets, 1) chromium-rich (> 3 wt%) pyropes with a green
to blue-green color change, and 2) pyralspites containing octahedrally coordinated V and/or Cr with a red
to purple-red color change. Gubelin (1982) describes chromium-rich pyropes that are blue-green in
daylight and wine red under incandescent light due to the presence of Cr and/or V. Less Cr and V are
necessary to cause color-change in garnets with a major spessartine component (Gubelin, 1982).

51
Spectral Features Related to Chromophores

There are multiple luminescence features in the Raman spectra of garnet associated with both Cr and V.
Gaft et al. (2005) attribute luminescence features located at 690, 695, 698, 703, and 717 nm to the
presence of vanadium in grossular and a band at 694 nm to Cr3+ in rhodolite (pyrope-almandine mixture).
Polarized UV-VIS absorption spectra of uvarovites conducted by Andrut and Wildner (2001) attribute
features at 686, 701, and 695 nm to octahedrally coordinated chromium. Absorption studies of
chromium-doped gallium garnet laser crystals report the presence of peaks associated with Cr3+ centers
located at 692 and 696 nm (Struve and Huber, 1985).

Based on the location of Cr3+ and V luminescence features in garnet described by the aforementioned
authors and the location of features in other chromium- and vanadium-bearing mineral species such as
topaz, beryl, spinel, corundum, chrysoberyl, and zoisite (see Appendices A & B), the following table has
been provided to address the luminescence features present in each garnet specie and assign them to
the corresponding chromophores (Table 6.1).

In this study, there are more observable luminescence features present in the Raman spectra of colorless
pyrope, than there are in deep red-colored pyrope (Fig. 6.3 C.). Typically, colorless minerals are
colorless due to a lack of chromophores or color centers. The fact that there are actually more
luminescence features in the Raman spectra of colorless pyrope samples is, therefore, unexpected.
Colorless pyrope samples exhibit eight well-resolved peaks attributed to luminescence centers. Peaks
centered at 683 nm, 686 nm, 690 nm, 695 nm, 693 nm, 699 nm, and 701 nm are likely related to Cr, while
a single peak centered at 729 nm may be related to V (Gaft et al., 2005; Andrut and Wildner, 2001).
There are only five peaks associated with Cr luminescence centers in red pyrope located at 683 nm, 686
nm, 689 nm, 692 nm, and 695 nm (Gaft et al., 2005; Andrut and Wildner, 2001). These peaks appear
broader and less resolved than the corresponding peaks present in colorless pyrope.

52
Table 6.1 Peaks related to Cr3+ and V luminescence centers in garnets, Raman spectra: λexcitation = 532 nm,
unoriented

Spectral
Sample Description/Origin Garnet Ion Figure Reference
Features
nm cm-1
red, Meronitz, 532 6.1 & Gaft et al., 2005 (topaz &
R040159 Pyrope 683 4166 Cr3+
Bohemia 6.3 B beryl)
Andrut and Wildner,
532 6.1 &
686 4225 Cr3+ 2001; Gaft et al., 2005
6.3 B
(spinel)
532 6.1 &
689 4290 Cr3+?
6.3 B
532 6.1 & Gaft et al., 2005
693 4360 Cr3+
6.3 B (chrysoberyl & corundum)
532 6.1 & Struve and Huber, 1985;
696 4431 Cr3+ or V?
6.3 B Gaft et al., 2005
532 Gaft et al., 2005
R060448 red, India Pyrope 693 4366 Cr3+ 6.2
(chrysoberyl & corundum)
532 3+
694 4393 Cr 6.2 Gaft et al., 2005
colorless, Piedmont, 532 3+ Gaft et al., 2005 (topaz &
R070637 Pyrope 683 4172 Cr 6.3
Italy beryl)
Andrut and Wildner,
532
686 4236 Cr3+ 6.3 C 2001; Gaft et al., 2005
(spinel)
532
690 4316 Cr3+? 6.3 C
532 3+ Gaft et al., 2005
693 4380 Cr 6.3 C
(chrysoberyl & corundum)
532 Struve and Huber, 1985;
695 4427 Cr3+ or V? 6.3 C
Gaft et al., 2005
532
699 4492 Cr3+? 6.3 C
532
701 4550 Cr3+ 6.3 C Andrut and Wildner, 2001
532
729 5095 V? 6.3 C
532
R060099 red, Alaska Almandine 688 4277 Cr3+? 6.4
532 3+ Gaft et al., 2005
693 4372 Cr 6.4
(chrysoberyl & corundum)
532
694 4401 Cr3+ 6.4 Gaft et al., 2005
532
706 4642 V? 6.4
red-brown, Shigar 532 3+ Gaft et al., 2005
R060279 Spessartine 693 4370 Cr 6.5
Valley, Pakistan (chrysoberyl & corundum)
532
710 4705 V? 6.5
532 3+ Gaft et al., 2005
R080053 orange, East Africa Spessartine 693 4370 Cr 6.6
(chrysoberyl & corundum)
532
694 4395 Cr3+ 6.6 Gaft et al., 2005
Andrut and Wildner,
transparent tan, 532
R060382 Grossular 697 4450 Cr3+ 6.7 2001; Struve and Huber,
Tanzania
1985
532
701 4544 Cr3+ 6.7 Andrut and Wildner, 2001
532
717 4840 V 6.7 Gaft et al., 2005
532
721 4930 V? 6.7
532 6.7 A.
605 2262 Mn Gaft et al., 2005
& C.
green, Ural Mts., 532 Struve and Huber, 1985;
R060477 Uvarovite 696 4440 Cr3+or V? 6.8
Russia Gaft et al., 2005
532
701 4542 Cr3+ 6.8 Andrut and Wildner, 2001
532
721 4930 V? 6.8

53
Pyrope

Fig. 6.1 A. Peaks likely related to Cr3+ and possibly V luminescence centers; note that the luminescence peaks
overwhelm the Raman peaks in this spectrum; sample R040159, red pyrope, Meronitz, Bohemia; λexcitation = 532 nm,
unoriented

V?

Cr3+

Fig. 6.2 A. Peaks likely related to Cr3+ centers; note that the luminescence peaks overwhelm the Raman peaks in
this spectrum; sample R060448, red pyrope, India; λexcitation = 532 nm, unoriented

Cr3+

Fig. 6.1 B. (right) Magnification of peaks located at

683 nm (4166 cm-1), 686 nm (5324225 cm-1), 689 nm V?
(5324290 cm-1), and 693 nm (5324360 cm-1), related to
3+ 532
Cr centers and a peak located at 696 nm ( 4431
-1
cm ) possibly related to a V center; sample
Intensity

R040159, red pyrope, Meronitz, Bohemia;

unoriented

Cr3+

Raman Shift (cm-1)

54
Fig. 6.2 B. (left) Magnification of peaks located at 693 (5324366 cm-1)
and 694 nm (5324396 cm-1) associated with Cr3+ luminescence
centers; sample R060448, red pyrope, India, unoriented
Intensity

Cr3+

-1
Raman Shift (cm )

Fig. 6.3 A. Peaks in Raman spectrum of colorless pyrope located at 683 nm (5324172 cm-1), 690 nm (5324316 cm-1),
532 -1 532 -1 532 -1 532 -1
695 nm ( 4427 cm ), 693 nm ( 4380 cm ), 699 ( 4492 cm ), and 701 nm ( 4550 cm ), are likely related to Cr,
532 -1
while a single peak centered at 729 nm ( 5095 cm ) may be related to V sample; note that the luminescence peaks
overwhelm the Raman peaks in this spectrum; sample R080060, synthetic pyrope, unoriented

Cr3+

V
Cr3+

Fig. 6.3 B. Peaks in Raman spectrum of red pyrope located at at 683 nm (5324171 cm-1), 686 nm (5324225 cm-1),
689 nm (5324289 cm-1), 692 (5324358 cm-1), and 695 nm (5324429 cm-1) related to Cr3+ centers; sample R040159,
Sunset Crater, AZ, unoriented

V?

Cr3+

55
Fig. 6.3 C. Comparison of magnified luminescence features associated with Cr and V centers in red pyrope (3-4)
and colorless pyrope (1-2); 1) sample R070637, colorless pyrope, Piedmont, Italy, 2) sample R080060, synthetic
colorless pyrope, 3) sample R040159, red pyrope, Bohemia, 4) sample R050446, red pyrope, Sunset Crater, AZ; 532
nm laser, unoriented

V?

Cr3+
Intensity

4.
Cr3+
3.

Cr3+ 2.
V
3+
Cr 1.

Raman Shift (cm-1)

Almandine

Fig. 6.4 A. Peaks in Raman spectrum of almandine associated with Cr3+ and V luminescence centers; sample
R060099, red almandine, Alaska, λexcitation = 532 nm, unoriented

Fig. 6.4 B. (right) Magnification of peaks associated

with Cr3+ located at 688 nm (5324277 cm-1), 693 nm
532 -1 532 -1
( 4372 cm ), and 694 nm ( 4401 cm ) and a peak
located at 706 nm (5324642 cm-1) related to a V center;
sample R060099, Alaska, unoriented
Cr3+
Intensity

V?

Raman Shift (cm-1)

56
Spessartine

Fig. 6.5 A. Peaks in the Raman spectrum of spessartine associated with Cr3+ and possibly V luminescence centers;
sample R0600279, spessartine, Pakistan, λexcitation = 532 nm, unoriented

Cr3+

V?

Fig. 6.5 B. Magnification of a peak associated with Cr3+ located at 693 nm (5324370 cm-1) and a peak located at 710
532
nm ( 4705 cm-1) possibly related to V in spessartine; sample R060279, Pakistan, unoriented

Cr3+
Intensity

V?

Raman Shift (cm-1)

Fig. 6.6 A. Peaks in the Raman spectrum of spessartine associated with Cr3+ luminescence centers; sample
R080053, East Africa, λexcitation = 532 nm, unoriented

Cr3+

57
Fig. 6.6 B. Magnification of peaks located at 693
nm (5324370 cm-1) and 694 nm (5324395 cm-1)
3+
associated with Cr centers in spessartine; sample
R080053, East Africa, unoriented

Cr3+

Intensity
Raman Shift (cm-1)

Grossular

Fig. 6.7 A. Peaks associated with Cr3+, Mn, and possibly V centers in the Raman spectrum of grossular; note that
the luminescence peaks overwhelm the Raman peaks in this spectrum; sample R060382, Tanzania, λexcitation = 532
nm, unoriented

V?

Cr3+3+
Cr

Mn

Fig. 6.7 B. (right) Magnification of peaks located at 697 nm

(5324450 cm-1) and 701 nm (5324544 cm-1) related to Cr3+
532 -1
centers and peaks located at 717 nm ( 4840 cm ), 721 nm
(5324930 cm-1) possibly related to V centers; sample V?
R060382, grossular from Tanzania, unoriented
Intensity

Cr3+

Raman Shift (cm-1)

58
Fig. 6.7 C. (left) Magnification of a broad feature
Mn located at 605 nm (5322262 cm-1) likely related to Mn;
sample R060382, grossular from Tanzania,
Intensity

unoriented

-1
Raman Shift (cm )

Uvarovite

Fig. 6.9 A. Peaks in the Raman spectrum of uvarovite associated with Cr3+ and possibly V centers as well as
centers associated with unidentified REE; note that the luminescence peaks overwhelm the Raman peaks in this
spectrum; sample R060477, Ural Mts., Russia, λexcitation = 532 nm, unoriented

V?

Cr3+

Fig. 6.9 B. Magnification of peaks located at 696 nm (5324440 cm-1) and 701 nm (5324542 cm-1) attributed to Cr3+
centers and a peak possibly related to V located at 721 nm (5324930 cm-1) in the Raman spectrum of uvarovite;
sample R060477, Ural Mts., Russia, unoriented

?
V

Cr3+
Intensity

-1
Raman Shift (cm )

59
Spectral Features Associated with Rare-earth Elements

Incorporation of REE’s into the structure of garnet is of great importance to geoscientists because of its
applications to geochemical and thermodynamic modeling. Trace-element partitioning and diffusion
between garnets and silicate melts are vital to understanding the dynamics of the mantle (Quatieri et al.,
2002). Very few studies have been conducted to determine the site preferences and coordination of trace
and rare-earth elements in garnets (Quatieri et al., 2002). Although garnets incorporate relatively low
concentrations of rare-earth elements into their structures compared to other silicate minerals such as
titanite and zircon, isotopic studies (Nd-Sm, U-Pb, Rb-Sr, Lu-Hf) of metamorphic garnets still provide
valuable data necessary to constrain temperature and pressure conditions as well as dates of
metamorphic and igneous events (Prince et al., 2000).

There are several luminescence features present in the Raman spectra of garnet that are likely related to
trace amounts of REE. Although many of these features have not been described in the literature,
comparison of spectral features in the various garnet species with luminescence features of other REE-
bearing minerals (like titanite and zircon) studied by Gaft et al. (2005), suggests that several of these
3+
peaks are likely related to Nd . According to luminescence studies of titanite conducted by Gaft et al.
(2003), nearly all Nd luminescence emission bands appear in the IR region. Nd3+ has been assigned to
3+

features located at 860, 870, 878, 880, 888, 906/907, 930, 940/942, 1047, 1060, 1070/1071, 1080,
1089/1090, 1100, 1115, and 1131 nm in titanite (Gaft et al., 2003). Additional luminescence studies done
on zircon also attribute bands at 817 and 885 nm to Nd3+ (Gaft et al., 2005). Denisov et al. (1986) provide
images of spectral features associated with Nd3+ in their study of gallium garnet crystals synthesized for
use in lasers. Although no table of assigned centers was provided, the band locations appear to be
consistent with the observations made by Gaft et al. (2003, 2005) for titanite and zircon (Denisov et al.,
1986).

Based on the location of Nd3+ spectral features provided by the aforementioned authors, the following
table and corresponding figures address the various Nd3+ peaks present in the Raman spectra of each
garnet variety as well as describe additional peaks likely related to unidentified REE luminescence
centers (Table 6.3). Laser ablation ICP-MS data for sample R060382, grossular from Tanzania, is
provided in Table 6.2 (Breeding, 2007).

Table 6.2 Laser ablation ICP-MS data for grossular, sample R060382, Lalatema, Tanzania; numerical values
represent concentrations in parts per million (ppm) of each element taken at three different spots on the sample
(Breeding, 2007).

7Li 24Mg 29Si 31P 43Ca 44Ca 45Sc 48Ti 51V

3.7 2966 18.5 68.2 28.4 27.4 16.0 2612.0 212.6
3.4 2974 18.2 62.5 27.9 26.7 15.4 2650.0 230.6
3.7 2872 17.7 67.6 27.5 26.6 16.4 2672.0 231.7

52Cr 55Mn 56Fe 57Fe 69Ga 72Ge 88Sr 89Y 90Zr

13.2 2802.0 1110.0 1306.0 24.5 2.6 1.2 95.6 24.7
24.5 2721.0 992.7 1159.0 23.2 2.5 1.2 93.7 25.4
35.5 2620.0 949.2 1121.0 22.5 2.4 1.2 99.7 26.3

140Ce 141Pr 146Nd 147Sm 153Eu 157Gd 159Tb 163Dy 165Ho

0.9 0.4 5.4 5.5 2.9 8.0 1.9 14.6 3.4
0.9 0.4 5.4 5.1 2.8 7.9 1.9 14.8 3.4
0.9 0.4 5.6 5.4 2.8 7.7 1.9 15.4 3.4

166Er 169Tm 172Yb 175Lu 178Hf

10.5 1.6 13.5 1.7 1.1
10.4 1.6 13.4 1.8 1.0
11.2 1.6 14.2 1.9 1.3

60
Grossular

Fig. 6.10 A. Peaks associated with Nd3+ luminescence centers in the Raman spectrum of grossular; note that the
luminescence peaks overwhelm the Raman peaks in this spectrum; sample R060382, Lalatema, Tanzania, λexcitation =
785 nm, unoriented

Nd3+ ?
?

Nd3+ 3+
Nd

Nd3+

Nd3+

Fig. 6.10 B. Magnification of peaks centered at 860 nm (7851123 cm-1), 875 nm (7851314 cm-1), 879 nm (7851367
cm-1), 888 nm (7851490 cm-1), 933 nm (7852023 cm-1) attributed to Nd3+ centers and peaks located at 895 nm (7851568
-1 785 -1 785 -1
cm ), 911 nm ( 1763 cm ), and 950 nm ( 2216 cm ) likely related to luminescence centers of unidentified REE;
sample R060382, Lalatema, Tanzania, unoriented

Nd3+ ?
Intensity

Nd3+ ? ?

Nd3+
Nd3+

Raman Shift (cm-1)

Fig. 6.10 C. Magnification of peaks centered at 1041 nm (7853135 cm-1), 1061 nm (7853315 cm-1), 1070 nm (7853396
cm-1), and 1082 nm (7853502 cm-1) attributed to Nd3+ centers; sample R060382, Lalatema, Tanzania, unoriented
Intensity

Nd3+

Nd3+

Nd3+

Raman Shift (cm-1)

61
Uvarovite

Fig. 6.11 A. Peaks associated with Nd3+ and other unidentified REE luminescence centers in the Raman spectrum
of uvarovite; note that the luminescence peaks overwhelm the Raman peaks in this spectrum; sample R060477, Ural
Mts., Russia, λexcitation = 780 nm, unoriented

Nd3+
Nd3+
Nd3+

Nd3+

Fig. 6.11 B. Magnification of peaks centered at 861 nm (7801206 cm-1), 874 nm (7801387 cm-1), 878 nm (7801441
cm-1), 888 nm (7801559 cm-1), and 932 nm (7802091 cm-1) related to Nd3+ and peaks located at 894 nm (7801632 cm-1),
780 -1 780 -1
910 nm ( 1830 cm ), and 948 nm ( 2278 cm ) likely related to luminescence centers of unidentified REE in the
Raman spectrum of uvarovite; sample R060477, Ural Mts., Russia; unoriented

?
Intensity

3+
Nd
?
3+
Nd ?

3+
Nd
Nd3+

-1
Pyrope Raman Shift (cm )

Fig. 6.12 A. Peaks associated with Nd3+ and other unidentified REE luminescence centers in the Raman spectrum
of pyrope; note that the luminescence peaks overwhelm the Raman peaks in this spectrum; sample R080060 λexcitation
= 780 nm, unoriented

Nd3+

? Nd3+

Nd3+

62
Fig. 6.12 B. Magnification of peaks centered at 863 nm (7801239 cm-1), 871 nm (7801348 cm-1), 877 nm (7801426
cm-1), 879 nm (7801446 cm-1), 885 nm (7801526 cm-1), and 942 nm (7802207 cm-1) related to Nd3+ centers and peaks
780 -1 780 -1 780 -1
located at 894 nm ( 1633 cm ), 903 nm ( 1750 cm ), and 951 nm ( 2312 cm ) related to luminescence centers
of unidentified REE in the Raman spectrum of pyrope; sample R080060, unoriented

Nd3+

?
Intensity

?
?

Nd3+

Raman Shift (cm-1)

Spessartine

Fig. 6.13 A. Peaks associated with Nd3+ and other unidentified REE luminescence centers in the Raman spectrum
of spessartine; sample R060279, red-brown spessartine from Pakistan, λexcitation = 785 nm, unoriented

Nd3+

Fig. 6.13 B. Magnification of peaks centered at 877 nm (7851335 cm-1), 886 nm (7851457 cm-1), 930 nm (7851991
cm-1), and 942 (7852126 cm-1) related to Nd3+ centers and peaks located at 904 nm (7851675 cm-1) and 912 nm
785 -1
( 1777 cm ) likely related to luminescence centers of unidentified REE in the Raman spectrum of spessartine;
sample R060279, red-brown spessartine from Pakistan, unoriented

Nd3+
Nd3+
Intensity

Raman Shift (cm-1)

63
Feature Related to Unidentified REE Luminescence Centers:

Fig. 6.14 A. Peaks likely related to unidentified REE luminescence centers in the Raman spectrum of grossular;
sample R060443, unknown locality, λexcitation = 532 nm, unoriented

Fig. 6.14 B. Magnification of peaks centered at 790 nm (5326142 cm-1), 801 nm (5326321 cm-1), and 807 nm
(5326408 cm-1) likely related to luminescence centers of unidentified REE in the Raman spectrum of grossular; sample
R060443, unknown locality, unoriented
Intensity

?
?

Raman Shift (cm-1)

Fig. 6.15 A. Peak likely related to an unidentified REE luminescence center in the Raman spectrum of spessartine;
R050063, China, λexcitation = 785 nm, unoriented

64
Fig. 6.15 B. Magnification of peak centered at
968 nm (7852414 cm-1) likely related to an
unidentified REE luminescence center; sample
R050063, spessartine from China, unoriented

Intensity
Raman Shift (cm-1)

Fig. 6.16 A. Peaks likely related to unidentified REE luminescence centers in the Raman spectrum of uvarovite;
note that the luminescence peaks overwhelm the Raman peaks in this spectrum; sample R06104, Finland, λexcitation =
532 nm, unoriented

Fig. 6.16 B. Magnification of peaks centered at 1580 nm (5326411 cm-1), 1598 nm (5326484 cm-1), and 1621 nm
(5326571 cm-1) likely related to unidentified REE luminescence centers; sample R061041, uvarovite from Finland,
unoriented
Intensity

Raman Shift (cm-1)

65
Table 6.3 Table of spectral features in garnet related to the presence of Nd3+ and possibly other unidentified REE

Sample Garnet
Description/Origin Spectral Feature Ion Figure Reference
No. Specie
-1
nm cm
transparent tan, 785
R060382 Grossular 860 1123 Nd3+ 6.10 A & B Gaft et al., 2003 (titanite)
Lalatema Tanzania
785
875 1314 Nd? 6.10 A & B
785
879 1367 Nd3+ 6.10 A & B Gaft et al., 2003 (titanite)
785
888 1490 Nd3+ 6.10 A & B Gaft et al., 2003 (titanite)
785
895 1568 ? 6.10 A & B
785
911 1763 ? 6.10 A & B
785
933 2023 Nd3+ 6.10 A & B Gaft et al., 2003 (titanite)
785
950 2216 ? 6.10 A & B
785
1041 3135 Nd? 6.10 A & C
785
1061 3315 Nd3+ 6.10 A & C Gaft et al., 2003 (titanite)
785
1070 3396 Nd3+ 6.10 A & C Gaft et al., 2003 (titanite)
785
1082 3502 Nd3+ 6.10 A & C Gaft et al., 2003 (titanite)
yellowish tan, unknown 532
R060443 Grossular 790 6142 ? 6.14
locality
532
801 6321 ? 6.14
532
807 6408 ? 6.14
532
818 6584 Nd? 6.14 Gaft et al., 2005 (zircon)
780
R060477 green, Ural Mts., Russia Uvarovite 861 1206 Nd3+ 6.11 Gaft et al., 2003 (titanite)
780
874 1387 Nd? 6.11
780 3+
878 1441 Nd 6.11 Gaft et al., 2003 (titanite)
780 3+
888 1559 Nd 6.11 Gaft et al., 2003 (titanite)
780
894 1632 ? 6.11
780
910 1830 ? 6.11
780 3+
932 2091 Nd 6.11 Gaft et al., 2003 (titanite)
780 3+
932 2091 Nd 6.11 Gaft et al., 2003 (titanite)
780
948 2278 ? 6.11
green, Outokumpu, 532
R061041 Uvarovite 1580 6411 ? 6.16
Finland
532
1598 6484 ? 6.16
532
1621 6571 ? 6.16
colorless, synthetic, P =
780
R080060 23.5 kb, T = 1000 deg C Pyrope 863 1239 Nd3+ 6.12 Gaft et al., 2003 (titanite)
for 23 hours
780
871 1348 Nd3+ 6.12 Gaft et al., 2003 (titanite)
780 3+
877 1426 Nd 6.12 Gaft et al., 2003 (titanite)
780
879 1446 Nd3+ 6.12 Gaft et al., 2003 (titanite)
780 3+
885 1526 Nd 6.12 Gaft et al., 2005 (zircon)
780
894 1644 ? 6.12
780
903 1750 ? 6.12
780 3+
942 2207 Nd 6.12 Gaft et al., 2003 (titanite)
780
951 2312 ? 6.12

66
Sample Garnet
Description/Origin Spectral Feature Ion Figure Reference
No. Specie
nm cm-1
785
R060279 red-brown, Pakistan Spessartine 877 1335 Nd3+ 6.13 Gaft et al., 2003 (titanite)
785
886 1457 Nd3+ 6.13 Gaft et al., 2005 (zircon)
785
904 1675 ? 6.13
785
912 1777 ? 6.13
785 3+
930 1991 Nd 6.13 Gaft et al., 2003 (titanite)
785 3+
942 2126 Nd 6.13 Gaft et al., 2003 (titanite)
785
R050063 red-brown, China Spessartine 968 2414 ? 6.15

67
Spectral Features Related to OH-

The study of structurally incorporated water in garnets is of particular interest to geologists studying the
properties of the mantle. These nominally anhydrous minerals represent storage sites for hydrogen
(Andrut et al., 2002). The presence of hydrogen can strongly affect the physical properties of minerals,
and therefore, investigation of hydrogen substitution in minerals, such as garnet, may provide vital
seismic and thermodynamic insight into the mechanics of the mantle (Andrut et al., 2002). Garnets can
contain anywhere from <1-200 ppm of hydroxyl, with the average being under 60 ppm, depending on the
garnet member (Bell and Rossman, 1992). Pyralspites typically contain less than .5 wt% OH- while
ugrandites can hold up to 20 wt% OH- (Lager et al., 1989). In garnets with relatively high amounts of OH-
(> 5 wt%), such as members of the grossular-hydrogrossular series, IR modes associated with hydroxyl
are strong, broad, overlapping peaks centered at approximately 3660 and 3600 cm-1 (Beran and
Libowitsky, 2003; Rossman and Aines, 1991). Substitution of an (OH)4 for a SiO4 group, known as the
hydrogrossular substitution, produces these modes (Beran and Libowitsky, 2003). Lager et al. (1989)
describes a slightly different set of IR peak locations due to hydrogrossular substitution than the
aforementioned authors at 3598 and 3677 cm-1 (Lager et al., 1989). The hydrogrossular substitution has
been observed not only in members of the grossular-hydrogrossular series, but also in Ti-rich andradites
(Beran and Libowitsky, 2003). The details of hydrogen substitution in garnets with low OH- concentrations
remain ambiguous (Beran and Libowitsky, 2003; Johnson, 2006).

In general, garnets with low concentrations of OH- display multiple, fine-structured IR bands which may be
affected by cation substitutions in proximity of the OH- (Johnson, 2006). IR studies of synthesized pyrope
conducted by Withers et al. (1998) reveal an IR mode centered at 3630 cm-1, likely representing
hydrogrossular substitution. In the same study, bands located at 3622 (shoulder at 3612), 3561, 3602,
3656, and 3665 cm-1 were attributed to hydroxyl incorporation in synthesized grossulars (Withers et al.,
1998). Rossman et al. (1989) conducted IR experiments on pyrope from Dora Maira Massif in the
Western Alps and reported peaks located at 3661, 3651, 3641 and 3602 cm-1 in the spectra, modes
similar to those found in low-OH- grossular. Multiple spectroscopic studies of uvarovite reveal fourteen
hydroxyl modes in the IR spectra (Andrut et al., 2002). Modes located at 3559/3540, 3572/3565,
3595/3588, and 3618 cm-1 are likely associated with hydrogrossular substitution, while modes located at
3652/3602 and 3640 cm-1 may represent substitution of OH- into cation vacancies resulting in SiO3(OH)
tetrahedral groups (Andrut et al., 2002).

Vibrational modes associated with hydroxyl were observed in the Raman spectra of grossular (Fig. 6.17),
spessartine (Fig. 6.18), and andradite (Fig. 6.19 & 6.20). A doublet in the spectrum of spessartine with
peaks at 3583 and 3632 cm-1 (Fig. 6.18) and a single mode located at 3578 cm-1 in andradite (Fig. 6.19)
may represent hydrogrossular substitution (Lager et al., 1989; and Andrut et al., 2002). The more
complex features in grossular (3534, 3570, 3610, 3645, and 3686 cm-1, Fig. 6.17) and andradite (3538,
3564, 3596, 3611, and 3634 cm-1, Fig. 6.20) are more difficult to interpret. Due to the complexity of the
features it is likely that the sample contains low concentrations of OH- (Johnson, 2006).

Arrendondo and Rossman (2002) conducted IR and Raman studies on two suites of garnets containing
OH to determine whether or not water content could be calculated using Raman spectra. They concluded
that Raman spectra are not well suited for the quantitative determination of water in garnet (Arrendondo
and Rossman, 2002). Studies conducted by Thomas et al. (2008) contradict this conclusion, claiming
that OH content can be quantitatively analyzed in garnets by confocal Raman spectroscopy and
comparison with glass standards of known chemical compositions, also known as the “Comparator
Technique”. Di Muro et al. (2006) confirm the accuracy and reliability of Raman spectroscopy to quantify
water content in glass chips.

68
Fig. 6.17 A. Peaks in the Raman spectrum of grossular associated with OH-; sample R050312, Eden Mills,
Vermont, 532 nm laser, unoriented

OH-

Fig. 6.17 B. (below) Magnification of peaks centered at 3534,

-1
3570, 3610, 3645, and 3686 cm attributed to OH in grossular;
sample R050312, Eden Mills, Vermont, 532 nm laser,
unoriented Intensity

OH-
Intensity

OH-
Raman Shift (cm-1)

Fig. 6.18 B. (above) Magnification of doublet

centered at 3583 and 3632 cm-1 associated with OH
Raman Shift (cm-1) in sppessartine, sample R050063, Fujian Province,
China, 532 nm laser, unoriented

Fig. 6.18 A. Peaks in the Raman spectrum of spessartine associated with OH-; sample R050063, Fujian Province,
China, 532 nm laser, unoriented

OH-

69
Fig. 16.19 A. Peaks in the Raman spectrum of andradite associated with OH-; sample R060350, San Benito Cty.,
CA, 532 nm laser, unoriented

OH-

Fig. 6.19 B. (right) Magnification of peak centered at 3578 cm-1

attributed to OH- in andradite, sample R060350, San Benito Cty.,
CA, 532 nm laser, unoriented

Intensity

OH-

Raman Shift (cm-1)

Fig. 16.20 A. Peaks in the Raman spectrum of andradite associated with OH-; sample R040001, Stanley Butte,
Graham Cty., AZ, 532 nm laser, unoriented

OH-

70
Fig. 6.20 B. Magnification of peaks centered at 3538, 3564, 3596, 3611, and 3634 cm-1 attributed to OH- in
andradite, sample R040001, Stanley Butte, Graham Cty., AZ, 532 nm laser, unoriented

Intensity

OH-

Raman Shift (cm-1)

71
References

Andrut, M., Wildner, M. & Beran, A. (2002) The crystal chemistry of birefringent natural uvarovites. Part IV. OH defect
incorporation mechanisms in non-cubic garnets derived from polarized IR spectroscopy. European Journal of
Mineralogy, Vol. 14, No. 6, pp. 1019.
Andrut, M. & Wildner, M. (2001) The crystal chemistry of birefringent natural uvarovites: Part I. Optical investigations
and UV-VIS-IR absorption spectroscopy. American Mineralogist, Vol. 86, No. 10, pp. 1219-1230.
Arredondo, E.H. & Rossman, G.R. (2002) Feasibility of determining the quantitative OH content of garnets with
Raman spectroscopy. American Mineralogist, Vol. 87, No. 2-3, pp. 307-311.
Bell, D.R. & Rossman, G.R. (1992) Water in Earth's mantle: The role of nominally anhydrous minerals. Science, Vol.
255, No. 5050, pp. 1391-1397.
Beran, A. & Libowitzky, E. (2003) IR spectroscopic characterization of OH defects in mineral phases. Phase
transitions, Vol. 76, No. 1-2, pp. 1-15.
Breeding, M. (2007) Personal communication.
Carstens, H. (1973) The red-green change in chromium-bearing garnets. Contributions to Mineralogy and Petrology,
Vol. 41, No. 3, pp. 273-276.
Denisov, A.L., Ostroumov, V.G., Saidov, Z.S., Smirnov, V.A. & Shcherbakov, I.A. (1986) Spectral and luminescence
3+ 3+
properties of Cr and Nd ions in gallium garnet crystals. Journal of the Optical Society of America B, Vol. 3,
No. 1, pp. 95-101.
Di Muro, A., Villemant, B., Montagnac, G., Scaillet, B. & Reynard, B. (2006) Quantification of water content and
speciation in natural silicic glasses (phonolite, dacite, rhyolite) by confocal microRaman spectrometry.
Geochimica et Cosmochimica Acta, Vol. 70, No. 11, pp. 2868-2884.
Fritsch, E. & Rossman, G. (1988) An update on color in gems. Part 3: Colors caused by band gaps and physical
phenomena. Gems & Gemology, Vol. 24, No. 2, pp. 81-102.
Gaft, M., Nagli, L., Reisfeld, R. & Panczer, G. (2003) Laser-induced time-resolved luminescence of natural titanite
CaTiOSiO4 . Optical Materials, Vol. 24, No. 1-2, pp. 231-241.
Gaft, M., Reisfeld, R. & Panczer, G. (2005) Modern Luminescence Spectroscopy of Minerals and Materials, Springer,
Heidelberg, Germany.
Gubelin, E. & Schmetzer, K. (1982) Gemstones with alexandrite effect. Gems & Gemology, Vol. 18, pp. 197-203.
Henderson, R.R. (2009) Determining chemical composition of the silicate garnets using Raman spectroscopy.
Prepublication Manuscript, University of Arizona.
Hofmeister A.M. & Chopelas A., (1991) Vibrational spectroscopy of end member silicate garnets, Physics and
Chemistry of Minerals 17, 503-526.
Johnson, E.A. (2006) Water in nominally anhydrous crustal minerals: speciation, concentration, and geologic
significance. Reviews in Mineralogy and Geochemistry, Vol. 62, No. 1, pp. 117-154.
Lager, G.A., Armbruster, T., Rotella, F. & Rossman, G.R. (1989) OH substitution in garnets: X-ray and neutron
diffraction, infrared, and geometric-modeling studies. American Mineralogist, Vol. 74, pp. 840-851.
Manning, P.G. (1969) Optical absorption studies of grossularite, andradite (var. colophonite) and uvarovite. Canadian
Mineralogist, Vol. 9, No. 5, pp. 723.
Manning, P.G. (1967) The optical absorption spectra of the garnets almandine-pyrope, pyrope, and spessartine and
some structural interpretations of mineralogical significance. Canadian Mineralogist, Vol. 9, No. 2, pp. 237.
Manning, P.G. & Harris, D.C. (1970) Optical-absorption and electron-microprobe studies of some high-Ti andradites.
Canadian Mineralogist, Vol. 10, No. 2, pp. 260.
Merli, M., Callegari, A., Cannillo, E., Caucia, F., Leona, M., Oberti, R. & Ungaretti, L. (1995) Crystal-chemical
complexity in natural garnets; structural constraints on chemical variability. European Journal of Mineralogy,
Vol. 7, No. 6, pp. 1239.
Novak, G.A. & Gibbs, G.V. (1971) The crystal chemistry of the silicate garnets. American Mineralogist, Vol. 56, pp.
791-825.
Prince, C.I., Kosler, J., Vance, D. & Günther, D. (2000) Comparison of laser ablation ICP-MS and isotope dilution
REE analyses—implications for Sm–Nd garnet geochronology. Chemical Geology, Vol. 168, No. 3-4, pp. 255-
274.
Quartieri, S., Boscherini, F., Chaboy, J., Dalconi, M.C., Oberti, R. & Zanetti, A. (2002) Characterization of trace Nd
and Ce site preference and coordination in natural melanites: a combined X-ray diffraction and high-energy
XAFS study. Physics and Chemistry of Minerals, Vol. 29, No. 7, pp. 495-502.
Rossman, G.R. & Aines, R.D. (1991) The hydrous components in garnets; grossular-hydrogrossular. American
Mineralogist, Vol. 76, No. 7-8, pp. 1153-1164.
Rossman, G.R., Beran, A. & Langer, K. (1989) The hydrous component of pyrope from the Dora Maira Massif,
Western Alps. European Journal of Mineralogy, Vol. 1, No. 1, pp. 151.
Schmetzer, K. & Bernhardt, H.J. (1999) Garnets from Madagascar with a color change of blue-green to purple. Gems
& Gemology, Vol. 35, No. 4, pp. 196–201.
3+
Struve, B. & Huber, G. (1985) The effect of the crystal field strength on the optical spectra of Cr in gallium garnet
laser crystals. Applied Physics B: Lasers and Optics, Vol. 36, No. 4, pp. 195-201.

72
Thomas, S.M., Thomas, R., Davidson, P., Reichart, P., Koch-Muller, M. & Dollinger, G. (2008) Application of Raman
spectroscopy to quantify trace water concentrations in glasses and garnets. American Mineralogist, Vol. 93, No.
10, pp. 1550.
Winchell, A.N. (1933) Optical Mineralogy II, Wiley and Sons, New York.
Withers, A.C., Wood, B.J. & Carroll, M.R. (1998) The OH content of pyrope at high pressure. Chemical Geology, Vol.
147, No. 1-2, pp. 161-171.

Additional Information

Aines, R.D. & Rossman, G.R. (1984) The hydrous component in garnets; pyralspites. American Mineralogist, Vol. 69,
No. 11-12, pp. 1116-1126.
Beran, A., Langer, K. & Andrut, M. (1993) Single crystal infrared spectra in the range of OH fundamentals of
paragenetic garnet, omphacite and kyanite in an eklogitic mantle xenolith. Mineralogy and Petrology, Vol. 48,
No. 2, pp. 257-268.
Beran, A. & Libowitzky, E. (2006) Water in natural mantle minerals II: olivine, garnet and accessory minerals.
Reviews in Mineralogy and Geochemistry, Vol. 62, No. 1, pp. 169-191.
Gillet, P., Fiquet, G., Malezieux, J.M. & Geiger, C.A. (1992) High-pressure and high-temperature Raman
spectroscopy of end-member garnets; pyrope, grossular and andradite. European Journal of Mineralogy, Vol. 4,
No. 4, pp. 651.
Kolesov, B.A. & Geiger, C.A. (1998) Raman spectra of silicate garnets. Physics and Chemistry of Minerals, Vol. 25,
No. 2, pp. 142-151.
Manning, P.G. (1972) Optical absorption spectra of Fe3+ in octahedral and tetrahedral sites in natural garnets.
Canadian Mineralogist, Vol. 11, No. 4, pp. 826.
3+ 3+
Mazurak, Z. & Czaja, M. (1995) Optical properties of tsavorite Ca3Al2 (SiO4)3: Cr , V from Kenya. Journal of
Luminescence, Vol. 65, No. 6, pp. 335-340.
O'Donnell, K.P., Marshall, A., Yamaga, M., Henderson, B. & Cockayne, B. (1989) Vibronic structure in the
photoluminescence spectrum of Cr3+ ions in garnets. Journal of Luminescence, Vol. 42, No. 6, pp. 365-373.
Rossman, G.R. (1996) Studies of OH in nominally anhydrous minerals. Physics and Chemistry of Minerals, Vol. 23,
No. 4, pp. 299-304.
van Westrenen, W., Blundy, J. & Wood, B. (1999) Crystal-chemical controls on trace element partitioning between
garnet and anhydrous silicate melt. American Mineralogist, Vol. 84, No. 5-6, pp. 838-847.

73
Olivine
Gem names: peridot

Ideal chemistry: (Mg,Fe2+)2SiO4

Crystal system: orthorhombic

Point Group: H-M: mmm S: D2h

Space Group: Pbnm

Burmese peridot; Photo courtesy of Stone Group Labs

Table of Atomic Coordinates - forsterite (Kirfel et al., 2005):

atom x y z
Mg1 0 0 0
Mg2 0.50846 0.77742 0.25
Si 0.07353 0.59403 0.25
O1 0.73408 0.59155 0.25
O2 0.22160 0.44704 0.25
O3 0.22253 0.66316 0.46697

Raman mode analysis:

Raman Active Modes

Atom Wyckoff Position Point Symmetry Ag B1g B2g B3g

O3 8d 1 3 3 3 3
Mg2, Si, O1, O2 4c m 2 1 2 1
Mg1 4a 1 - - - -

(1 x 8d) + (4 x 4c) + (1 x 4a)

Raman mode analysis predicts the existence of 36 active Raman modes in forsterite:

11Ag + 7B1g + 11B2g +7B3g = 36

74
Introduction to Raman Spectrum of Forsterite

Forsterite is an isosilicate and consists of hexagonal closest packed SiO4 groups and octahedrally
coordinated Fe2+ and Mg (Fig. 7.2). It is the most important mineral in the upper 400 km of the Earth
because it reflects the Fe component in the mantle. Forsterite crystals containing 8-10% of Fe are a
desirable yellow-green color, known to gemologists as peridot. Pure, end-member forsterite is colorless.
Innumerable studies have been conducted involving the structural and chemical properties of olivine due
to its geologic importance. The Raman spectrum of forsterite is provided below. The two most intense
peaks, centered at 824 and 856 cm-1, are associated with SiO4 stretching modes (Chopelas, 1991).

Fig. 7.1 A. The Raman spectrum of forsterite with mode assignments reported by Chopelas (1991); sample
R040052, synthetic forsterite; 514.5 nm laser, processed

824
Ag 856
Ag
Intensity

920
See Fig. 7.1 B B3g
965
592 Ag
226 374 545 B3g
Ag B3g Ag 608
Ag

Raman Shift (cm-1)

Fig. 7.1 B. Magnification of Raman peaks in 200-440 cm-1 range; mode assignments reported by Chopelas (1991);
sample R040052, synthetic forsterite; 514.5 nm laser, processed

374
329 339 B3g
Ag Ag
226
Ag 304
Intensity

Ag

315 422
B3g 410 Ag
B3g

Raman Shift (cm-1)

75
7.2 The crystal structure of forsterite; green octahedra: MgO6 groups, blue tetrahedra: SiO4 groups

A. Viewed down the c-axis B. Viewed down the b-axis

C. Viewed down the a-axis

76
Spectral Features Related to Chromophores

The yellowish-green gem variety of forsterite is well-known as peridot. Peridot contains octahedrally
coordinated Fe2+, with occasional trace amounts of Cr3+ producing its characteristic color (Fritsch and
Rossman, 1987; Fritsch and Rossman, 1988). Forsterite and chrysoberyl are isostructural and therefore,
it is likely that the weak peaks centered at 693 nm (5324373 cm-1) and 694 nm (5324400 cm-1) in the Raman
spectra of forsterite (Fig. 7.3 A. & B.) are associated with octahedrally coordinated Cr3+ luminescence
centers also present in chrysoberyl (Gaft et al., 2005).

Fig. 7.3 A. Peaks centered at 693 and 694 nm that are likely associated with Cr3+ luminescence centers; sample
R060539, brown forsterite gem from Sri Lanka, λexcitation = 532 nm, unoriented

Cr3+

Fig. 7.3 B. Magnification of peaks associated with Cr3+

luminescence centers; sample R060539, brown forsterite gem from
Sri Lanka, λexcitation = 532 nm, unoriented
Cr3+
Intensity

Raman Shift (cm-1)

77
References

Chopelas, A. (1991) Single crystal Raman spectra of forsterite, fayalite, and monticellite. American Mineralogist, Vol.
76, No. 7-8, pp. 1101-1109.
Fritsch, E. & Rossman, G. (1988) An update on color in gems. Part 3: Colors caused by band gaps and physical
phenomena. Gems & Gemology, Vol. 24, No. 2, pp. 81-102.
Fritsch, E. & Rossman, G.R. (1987) An Update on Color in Gems. Part 1. Introduction and Colors caused by
Dispersed Metal Ions. Gems and Gemology, Vol. 23, No. 3, pp. 126–139.
Gaft, M., Reisfeld, R. & Panczer, G. (2005) Modern Luminescence Spectroscopy of Minerals and Materials, Springer,
Heidelberg, Germany.
Kirfel, A., Lippmann, T., Blaha, P., Schwarz, K., Cox, D.F., Rosso, K.M. & Gibbs, G.V. (2005) Electron density
distribution and bond critical point properties for forsterite, Mg2SiO4, determined with synchrotron single crystal
X-ray diffraction data. Physics and Chemistry of Minerals, Vol. 32, No. 4, pp. 301-313.
Kolesov, B.A. & Geiger, C.A. (2004) A Raman spectroscopic study of Fe–Mg olivines. Physics and Chemistry of
Minerals, Vol. 31, No. 3, pp. 142-154.

Additional Information

Kolesov, B.A. & Tanskaya, J.V. (1996) Raman spectra and cation distribution in the lattice of olivines. Materials
Research Bulletin, Vol. 31, No. 8, pp. 1035-1044.

78
Quartz
Gem names: amethyst, citrine, ametrine,
praseolite, smoky, rose, milky, rutilated,
tourmalinated

Ideal chemistry: SiO2

Crystal system: trigonal

Point Group: H-M: 32 S: D3

Synthetic citrine gem
Space Group: P3221

Table of Atomic Coordinates (Ikuta et al., 2007)

atom x y z
Si 0.4696 0 0
O 0.4132 0.2679 0.1191

Raman mode analysis:

Raman Active
Modes
Wyckoff Point
Atom A1 E
Position Symmetry
O 6c 1 3 6
Si 3a 2 1 3

Raman mode analysis predicts the existence of 13 active Raman modes in quartz at room conditions:

4A1 + 9E = 13

79
Introduction to the Raman Spectrum of Quartz

Quartz is one of the most common minerals on the Earth’s surface, second only to feldspar. It is a
framework silicate consisting entirely of SiO4 groups (Fig. 8.2 A. & B.). Quartz is one of the few minerals
in this study that does not have an inversion center, therefore, all of the predicted vibrational modes can
be represented by peaks in the Raman spectra (Fig. 8.1). The most intense peak, centered at 463 cm-1,
is associated with Si-O-Si bending (Sato and McMillan, 1987). The second most intense peak, centered
at 205 cm-1, is associated with a soft mode (Jayaraman et al., 1967). This vibrational mode is much more
flexible than the mode associated with the 463 cm-1 peak, and therefore, it exhibits both temperature and
pressure dependence. Quartz undergoes a transition from the low temperature α phase (P3221) to the
high temperature β phase (P3121) at 573oC (Shapiro et al., 1967). With an increase in temperature, the
205 cm-1 peak exhibits a decrease in frequency as quartz approaches the α-β transition (Shapiro et al.,
1967). The peak is nonexistent in the Raman spectrum of β-quartz (Shapiro et al., 1967). Jayaraman et
al. (1967) report that the 205 cm-1 peak exhibits a large initial increase in frequency and a decrease in
peak width with pressure.

Fig. 8.1 The Raman spectrum of quartz with mode assignments as reported by Sato and McMillan (1987), a peak
-1
related to an E mode centered at 1066 cm is not visible in this spectrum; sample R040031, 514 nm laser,
processed, unoriented

463
A1
Intensity

205
A1 401
&
128 263 354 393
E E A1 E 697 808 1083 1160 1231
E E A1 E E

Raman Shift (cm-1)

Fig. 8.2 The crystal structure of quartz; blue tetrahedra: SiO4 groups

A. Viewed down the a-axis B. Viewed down the c-axis

80
Causes of Color in Quartz

The mechanisms responsible for the various colors of quartz are still controversial. Hole centers (missing
electrons in the crystal structure) created by the presence of various ions produce a variety of colors in
quartz. The distinctive purple color of amethyst is the result of a hole center created by incorporation of
Fe2+ and Fe3+ into the crystal structure combined with irradiation (Paradise, 1982). Exposure to heat
destabilizes this color center, and in the presence of Fe3+, amethyst changes color becoming yellow
citrine. If Fe2+ is present in amethyst, heat will alter the quartz to green-colored prasiolite (Paradise,
1982; Fritsch and Rossman, 1988). A hole center created by substitutional AlO4 groups produces the
diagnostic black color of smoky quartz, though the details of this process are still debated (Maschmeyer
et al., 1980; Fritsch and Rossman, 1988). The color of rose quartz is attributed to pink fibrous micro-
inclusions, likely a mineral species closely related to dumortierite (Goreva et al., 2001). The pink color of
these inclusions is the result of charge transfer between Fe2+ and Ti4+ in the M1 site of the inclusions (Ma
et al., 2002). There are no observable peaks in the Raman spectra of quartz associated with the
aforementioned color centers in this study.

References

Fritsch, E. & Rossman, G. (1988) An update on color in gems. Part 3: Colors caused by band gaps and physical
phenomena. Gems & Gemology, Vol. 24, No. 2, pp. 81-102.
Goreva, J.S., Ma, C. & Rossman, G.R. (2001) Fibrous nanoinclusions in massive rose quartz: The origin of rose
coloration. American Mineralogist, Vol. 86, No. 4, pp. 466-472.
Ikuta, D., Kawame, N., Banno, S., Hirajima, T., Ito, K., Rakovan, J.F., Downs, R.T. & Tamada, O. (2007) First in situ
X-ray identification of coesite and retrograde quartz on a glass thin section of an ultrahigh-pressure
metamorphic rock and their crystal structure details. American Mineralogist, Vol. 92, No. 1, pp. 57-63.
Jayaraman, A., Wood, D.L. & Maines, R.G. (1987) High-pressure Raman study of the vibrational modes in AlPO4 and
SiO2(α-quartz). Physical Review B, Vol. 35, No. 16, pp. 8316-8321.
Ma, C., Goreva, J.S. & Rossman, G.R. (2002) Fibrous nanoinclusions in massive rose quartz: HRTEM and AEM
investigations. American Mineralogist, Vol. 87, No. 2-3, pp. 269-276.
Maschmeyer, D., Niemann, K., Hake, H., Lehmann, G. & Räuber, A. (1980) Two modified smoky quartz centers in
natural citrine. Physics and Chemistry of Minerals, Vol. 6, No. 2, pp. 145-156.
Paradise, T.R. (1982) The natural formation and occurrence of green quartz. Gems & Gemology, Vol. 18, No. 1, pp.
39.
Sato, R.K. & McMillan, P.F. (1987) An infrared and Raman study of the isotopic species of alpha-quartz. Journal of
Physical Chemistry, Vol. 91, No. 13, pp. 3494-3498.
Shapiro, S.M., O'Shea, D.C. & Cummins, H.Z. (1967) Raman scattering study of the alpha-beta phase transition in
quartz. Physical Review Letters, Vol. 19, No. 7, pp. 361-364.

81
Spinel
Ideal chemistry: MgAl2O4

Crystal system: isometric

Point Group: H-M: m3m S: Oh

Space Group: Fd3m

Synthetic spinel

Table of Atomic Coordinates (Martignago et al., 2003):

atom x y z
Mg 0.125 0.125 0.125
Al 0.5 0.5 0.5
O 0.26338 0.26338 0.26338

Raman mode analysis:

Raman Active Modes

Atom Wyckoff Position Point Symmetry A1g Eg T2g

O 32e 3m 1 1 2
Al 16d 3m - - -
Mg 8a 4 3m - - 1

Raman mode analysis predicts the existence of 5 active Raman modes in spinel:

1Ag + 1Eg + 3T2g = 5

82
Spectral Features Related to Chromophores and Other Ions

Natural Spinel

The metal ions that produce the various colors of natural spinel include Cr3+, Fe3+, Fe2+, and Co2+ (Fritsch
and Rossman, 1988) (See Appendix D). Octahedrally coordinated Cr3+ produces pink and red hues in
spinel (Fritsch and Rossman, 1988). The addition of tetrahedrally coordinated Fe2+ to octahedrally
coordinated Cr3+ in the spinel structure creates a purple hue (Fritsch and Rossman, 1988). In addition to
these ions, synthetically grown spinels can contain trace amounts of Cu, V, and Mn that may contribute to
color (Cain, 1988). Multiple prominent peaks centered at 676 nm, 686 nm, 698 nm, 708 nm, and 718 nm
(Fig. 9.1 A. & B.) in the Raman spectra are related to Cr3+ luminescence centers in spinel (Gaft et al.,
2005). These features are present in the Raman spectra of spinel samples of several colors, but not in
the spectra of black spinel (Fig. 9.2). The peaks in the Raman spectra of black spinel are also poorly
resolved when compared to the Raman peaks in the spectra of the other colored samples (Fig. 9.3).

Fig. 9.1 A. Peaks centered at 676 nm (5324000 cm-1), 686 nm (5324211 cm-1), 698 nm (5324466 cm-1), 708 nm
(5324657 cm-1), and 718 nm (5324859 cm-1) in the Raman spectrum of natural spinel are related to the Cr3+
luminescence center; note that the luminescence peaks overwhelm the Raman peaks in this spectrum; sample
R050392, pink Burmese spinel, unoriented

Cr3+

Cr3+

Fig. 9.1 B. Magnification of peaks attributed to Cr3+ luminescence centers; sample R050392, natural pink Burmese
spinel, λexcitation = 532 nm, unoriented

Cr3+
3+
Cr
Intensity

Raman Shift (cm-1)

83
Fig. 9.2 Peaks associated with Cr3+ luminescence centers in natural spinel of a variety of colors; note the absence
of Cr3+ peaks in black spinel; 1) sample R060799, blue spinel from Pakistan; 2) sample R050392, pink Burmese
spinel; 3) sample R050411, very pale pink Burmese spinel; 4) sample R050259, purple Burmese spinel; 5) sample
R070013, black Thai spinel; λexcitation = 532 nm, unoriented

5.
Intensity

4.

3.

2.

1.

Raman Shift (cm-1)

Fig. 9.3 Comparison of the Raman spectra of spinel of various colors (both synthetic and natural) with the spectra
of natural black spinel; 1) natural pink spinel, Mogok, Burma, sample R050411; 2) natural blue spinel, Pakistan,
sample R060799; 3) natural black spinel, Cuba, sample R060798; 4) natural black spinel, Thailand, sample R070013;
5) synthetic blue spinel, sample X080014; 6) synthetic red spinel, sample X080013; λexcitation = 532 nm, processed,
unoriented

6.

5.
Intensity

4.

3.

2.
1.

Raman Shift (cm-1)

84
Synthetic Spinel

The peaks attributed to Cr3+ luminescence centers in natural spinel visibly differ from the peaks present in
the spectra of synthetic spinel samples (Fig. 9.4 A. & B.). The peaks in synthetic spinel appear broader
and poorly resolved when compared to the spectra of natural spinel. The highest intensity Cr3+ peak in
the spectra of synthetic spinel is also shifted to a higher frequency, by 83 cm-1, than the peak with the
highest intensity in natural spinel. Based on luminescence studies conducted by Tijero and Ibarra (1993),
when chromium-containing spinel are heated and annealed (700-950oC), the spectral features associated
with the luminescence centers can change dramatically. The peaks associated with Cr3+ centers in spinel
that have been heated and annealed are broader and less resolved. The changes in the spectra are
related to the disordering of the Mg and Al in the crystal structure. As the proportion of four and six
coordinated Mg and Al changes, there is a decrease in the distance between Cr3+ and O2-. As the spinel
are heated, the population of ordered centers associated with Cr3+ decreases, and therefore, the peaks
associated associated with the Cr3+ disappear (Tijero and Ibarra, 1993).

Fig. 9.4 A. Peaks related to Cr3+ centers in spinel; 1) synthetic red spinel gem, sample X080013; 2) synthetic blue
spinel gem, sample X080014; 3) natural pink Burmese spinel, sample R050392; notice the differences in shape and
width of the highest intensity peak; note that the luminescence peaks overwhelm the Raman peaks in this spectrum;
λexcitation = 532 nm, unoriented
Intensity

3.

2.

1.

Raman Shift (cm-1)

Fig. 9.4 B. Magnification of the peaks related to Cr3+ centers in spinel; 1) synthetic red spinel gem, sample
X080013; 2) synthetic blue spinel gem, sample X080014; 3) natural pink Burmese spinel, sample R050392; notice
the differences in shape and width of the highest intensity peak; λexcitation = 532 nm, unoriented

3.
Intensity

2.

1.

Raman Shift (cm-1)

85
Structural Effects on Raman Spectra

The ordered structure of spinel, called the normal structure, has the following cation distribution: X[Y2]O4,
where X are divalent tetrahedrally coordinated cations (Mg), and Y are octahedrally coordinated trivalent
cations (Al) (Uchida et al., 2005). Most natural spinels, and all synthetic spinels, are disordered, meaning
Al occupies the tetrahedral site and both Al and Mg occupy the octahedral site (Uchida et al., 2005). This
disordering produces structural defects, electron traps, and vacancies, in the crystal structure (Cain et al.,
1988). These defects can complicate the interpretation of the spectra of spinel. In nature, spinel can cool
slowly enough to create completely ordered crystals. An example are spinel of Burma. There are five
predicted Raman-active modes in spinel, however, only four peaks have been observed in the Raman
spectra of natural spinel. Based on analysis of synthetic spinel, the missing fifth peak should appear at
492 cm-1. Unlike natural spinel, synthetic spinel frequently produce six to seven Raman peaks including
an additional peak centered at 727 cm-1 (Cynn, 1992). In natural, ordered spinel, at ambient conditions,
this peak is not visible, however, once a sample is heated and then quenched, the peak appears in the
spectrum. The 727 cm-1 peak is attributed to the stretching vibration of AlO4 groups created by the
rearrangement of some Al ions from octahedral to tetrahedral sites (Cynn, 1992). This additional peak
has not been observed in this study. At this time, no definitive study relating the ordering of Mg and Al to
the Raman spectra of spinel has been conducted.

References

Cain, L.S., Pogatshnik, G.J. & Chen, Y. (1988) Optical transitions in neutron-irradiated MgAl2O4 spinel crystals.
Physical Review B, Condensed matter, Vol. 37, No. 5, pp. 2645-2652.
Chopelas, A. & Hofmeister, A.M. (1991) Vibrational spectroscopy of aluminate spinels at 1 atm and of MgAl2O4 to
over 200 kbar. Physics and Chemistry of Minerals, Vol. 18, No. 5, pp. 279-293.
Cynn, H., Sharma, S.K., Cooney, T.F. & Nicol, M. (1992) High-temperature Raman investigation of order-disorder
behavior in the MgAl2O4 spinel. Physical Review B, Condensed matter, Vol. 45, No. 1, pp. 500-502.
Fritsch, E. & Rossman, G. (1988) An update on color in gems. Part 3: Colors caused by band gaps and physical
phenomena. Gems & Gemology, Vol. 24, No. 2, pp. 81-102.
Gaft, M., Reisfeld, R. & Panczer, G. (2005) Modern Luminescence Spectroscopy of Minerals and Materials, Springer,
Heidelberg, Germany.
3+ 3+
Martignago, F., Negro, A.D. & Carbonin, S. (2003) How Cr and Fe affect Mg–Al order–disorder transformation at
high temperature in natural spinels. Physics and Chemistry of Minerals, Vol. 30, No. 7, pp. 401-408.
Tijero, J.M.G. & Ibarra, A. (1993) Use of luminescence of Mn2+ and Cr3+ in probing the disordering process in
MgAl2O4 spinels. Journal of Physical Chemistry, Solids, Vol. 54, No. 2, pp. 203-207.
Uchida, H., Lavina, B., Downs, R.T. & Chesley, J. (2005) Single-crystal X-ray diffraction of spinels from the San
Carlos Volcanic Field, Arizona: Spinel as a geothermometer. American Mineralogist, Vol. 90, No. 11-12, pp.
1900-1908.

86
Spodumene
Gem names: kunzite, hiddenite

Ideal chemistry: LiAlSi2O6

Crystal system: monoclinic

Point Group: H-M: 2/m S: C2h

Space Group: C2/c

Kunzite, Photo courtesy of Stone Group Labs

Table of Atomic Coordinates (Cameron et al., 1973):

atom x y z
Si 0.2941 0.0935 0.2560
Al1 0 0.9066 0.25
Li2 0 0.2752 0.25
O1 0.1099 0.0823 0.1402
O2 0.3646 0.2673 0.3009
O3 0.3565 0.9871 0.0578
Raman mode analysis:

Raman Active Modes

Atom Wyckoff Position Point Symmetry Ag Bg

Si, O1, O2, O3 8f 1 3 3

Al, Li 4e 2 1 2

(4 x 8f) + (2 x 4e)

Raman mode analysis predicts the existence of 30 active Raman modes in spodumene:

14Ag + 16B g = 30

87
Spectral Features Related to Chromophores and Other Ions

Spodumene is isostructural with diopside. Spodumene contains two nonequivalent octahedral sites
occupied by Al and Li (Souza, 2004). The Li site is distorted. Al and Li are frequently replaced by color-
inducing ions such as Mn, Cr, and Fe (Souza, 2004). The pink color of kunzite is associated with trace
amounts of manganese, while the diagnostic green color of hiddenite is associated with octahedrally
coordinated Cr and V (Fritsch and Rossman, 1988).

Luminescence studies of spodumene conducted by Walker et al. (1997), Souza et al. (2003), and Gaft et
al. (2005) report the presence of a broad luminescence feature located between 600-630 nm. The
interpretations of this feature vary. Walker et al. (1997) and Gaft et al. (2005) theorize that this peak is
associated with a Mn2+ center. However, Souza et al. (2003) report that this feature is not related to the
presence of Mn. Instead, Souza et al. (2003) theorize that the broad feature is likely related to a hole
captured by Al3+ atoms in the crystal structure. During experimentation with doped, artificial spodumene
crystals, Souza et al. (2003) observed that the presence of Mn intensifies this hole-center luminescence
band, while Fe quenches it. A broad peak located between centered at 620 nm is present in the Raman
spectra of every spodumene sample in this study and is likely related to the theorized centers described
above (Fig. 10.1).

Fig. 10.1 A broad peak located at 620 nm (5322670 cm-1) in the Raman spectrum of spodumene may be related to a
Al3+-hole center or Mn2+ center; note that the luminescence peaks overwhelm the Raman peaks in this spectrum;
sample R040050, variety kunzite, unoriented

Al3+
hole center
2+
or Mn ?
Intensity

Raman Shift (cm-1)

References

Cameron, M., Sueno, S., Prewitt, C.T. & Papike, J.J. (1973) High-temperature crystal chemistry of acmite, diopside,
hedenbergite, jadeite, spodumene, and ureyite. American Mineralogist, Vol. 58, pp. 594-618.
Fritsch, E. & Rossman, G. (1988) An update on color in gems. Part 3: Colors caused by band gaps and physical
phenomena. Gems & Gemology, Vol. 24, No. 2, pp. 81-102.
Gaft, M., Reisfeld, R. & Panczer, G. (2005) Modern Luminescence Spectroscopy of Minerals and Materials, Springer,
Heidelberg, Germany.
Souza, S.O., Ferraz, G.M. & Watanabe, S. (2004) Effects of Mn and Fe impurities on the TL and EPR properties of
artificial spodumene polycrystals under irradiation. Nuclear Inst.and Methods in Physics Research, B, Vol. 218,
pp. 259-263.
Souza, S.O., Watanabe, S., Lima, A.F. & Lalic, M.V. (2007) Thermoluminescent mechanism in lilac spodumene.
ACTA PHYSICA POLONICA SERIES A, Vol. 112, No. 5, pp. 1001.
3+
Walker, G., El Jaer, A., Sherlock, R., Glynn, T.J., Czaja, M. & Mazurak, Z. (1997) Luminescence spectroscopy of Cr
2+
and Mn in spodumene (LiAlSi2O6). Journal of Luminescence, Vol. 72, pp. 278-280.

88
Additional Information

Chandrasekhar, B.K. & White, W.B. (1992) Polarized luminescence spectra of kunzite. Physics and Chemistry of
Minerals, Vol. 18, No. 7, pp. 433-440.
Pommier, C.J.S., Denton, M.B. & Downs, R.T. (2003) Raman spectroscopic study of spodumene (LiAlSi2O6) through
the pressure-induced phase change from C2/c to P2/c. Journal of Raman Spectroscopy, Vol. 34, pp. 769-775.

89
Titanite
Gem names: sphene (old name)

Ideal chemistry: CaTiSiO5

Crystal system: monoclinic

Point Group: H-M: 2/m S: C2h

Space Group: P21/a or C2/c

Table of Atomic Coordinates (Ghose et al., 1991) (P21/a)

atom x y z
Ca 0.2421 0.4185 0.2511
Ti 0.5137 0.2540 0.7496
Si 0.7483 0.4327 0.2491
O1 0.7494 0.3219 0.7491
O2 0.9097 0.3161 0.4332
O3 0.0881 0.1849 0.0644
O4 0.3829 0.4609 0.6456
O5 0.6191 0.0399 0.8532

Raman mode analysis:

Raman Active Modes

Atom Wyckoff Position Point Symmetry Ag Bg

Ca, Ti, Si, O1, O2,

4e 1 3 3
O3, O4, O5

(8 x 4e)

Raman mode analysis predicts the existence of 48 active Raman modes in titanite:

24Ag + 24B g = 48

90
Metamictization Due to Radiation Damage

The crystal chemical properties of titanite not only make this stone valuable to collectors, but also to
geologists. The reaction of titanite crystals to radioactive elements can provide valuable information
about the geologic conditions under which they form (see also “Spectral Features Related to REE,” p.
69). Incorporation of U and Th into the crystal structure of titanite results in damage of the crystal, called
metamictization, due to radioactive decay (Vance and Netson, 1985). Zircon also experiences
metamictization, however, X-ray diffraction studies conducted by Vance and Netson (1985) reveal that
titanite is two to three times more sensitive to damage by α-decay. The reactivity of titanite to certain
radioactive elements may play a role in the disposal of nuclear fuel waste in the future. Vance and
Netson propose that titanite-containing glass ceramics could act as the immobilizing host of nuclear
waste. An understanding of the metamictization of titanite by radioactive compounds is not only of
mineralogical interest, but also of importance to this proposed waste disposal process (Vance and
Netson, 1985).

The effects of metamictization in titanite are apparent in various laboratory techniques and include:
broadening of X-ray diffraction (XRD) peaks, decrease in XRD peak intensity, loss of anisotropy or
orientational dependence of spectral features, and broadening and loss of resolution in IR bands (Zhang
-
et al., 2002). In addition, metamict titanite generally contains more OH than its crystalline counterpart
-
(see below, “Spectral Features Related to OH “) (Hammer et al., 1996; Zhang and Salje, 2003). Although
an in-depth Raman spectroscopic study of metamict titanite has yet to be conducted, Raman scattering of
partially amorphous titanite from the RRUFF database produce poorly resolved, broader Raman peaks
and concomitant widened diffraction peaks, as is the case in metamict zircons (Fig. 11.1) (Nasdala et al.,
1995).

Note: A structural phase transition occurs in titanite with increase in temperature. Raman spectroscopic
studies conducted by Salje et al. show that at temperatures above 860K, the structure of titanite changes
from P21/a symmetry to A2/a, resulting in peak differences in the Raman spectra (Salje et al., 1992).

Fig. 11.1 (Right) Portion of XRD pattern with associated band width values and (left) corresponding portion of
Raman spectra of titanites showing increase in metamictization from bottom to top; 1) sample R040033, 2) sample
R050124, 3) sample R050114, 4) R050039 (this sample has undergone the most damage); 514 nm laser, Laser
parallel to -b* (0 -1 0). Fiducial mark perpendicular to laser is parallel to -c [0 0 -1], background corrected

Intensity

.193
Intensity

4.
.079
3.
2.
.078
1.
.059
-1
Raman Shift (cm )
2 Theta

91
Spectral Features Related to OH-

When exposed to radiation, titanite becomes metamict (slightly amorphous). Radioactive decay produces
particles that smash through the crystal structure, breaking bonds and destroying crystallinity.
Metamictization alters the titanite structure allowing OH- to replace the oxygen atoms (O1) shared by the
bent chains of TiO6 groups in the crystal structure (Fig. 11.2) (Hammer et al., 1996). The O1 site is
underbonded making the substitution of O1 for OH- and F common (Frost et al., 2001). Degree of
metamictization may determine how much OH- a titanite sample contains; damaged titanites contain more
OH- than those without exposure to radiation (Hammer et al., 1996). A peak centered at 3485 cm-1 in the
Raman spectra of titanite corresponds to an IR band associated with OH- reported by Isetti and Penco
(1968) at the same location (Fig. 11.3 A. & B.).

Fig. 11.2 (Right) Crystal structure of titanite; arrows

-
pointing to pink O1 atoms where OH substitutes; green
octahedra: TiO6 groups, blue tetrahedra: SiO4 groups, red
spheres: oxygen atoms, teal polyhedra: 7-coordinated Ca
atoms

Fig. 11.3 A. Weak peak located at 3485 cm-1 attributed to OH-;

titanite from Brazil, sample R040033, 532 nm laser, unoriented
Intensity

OH
-

Raman Shift (cm-1)

Fig. 11.3 B. (right) Magnification of weak OH- peak located

at 3485 cm-1; titanite from Brazil, sample R040033, 532 nm
laser, unoriented
Intensity

OH-

Raman Shift (cm-1)

92
Cause of Color in Titanite

Gem-quality titanite (sphene) is typically green, yellow-green, yellow, brown and less frequently, pink, in
color. Iron is usually responsible for yellow, green, and brown colors in titanite. However, octahedrally
coordinated Cr3+ can also produce a green color in titanite (Fritsch and Rossman, 1988). The presence
of pink carbonate inclusions gives some crystals a pink color. In addition, Mn-rich titanite is pink due to
octahedrally coordinated Mn2+ (Fritsch and Rossman, 1988). No spectral features attributed to the
aforementioned chromophores have been observed in this study.

Spectral Features Related to Rare-earth Elements

The large, 7-fold Ca site in the crystal structure of titanite (Fig. 11.2) can incorporate many different large
atoms such as U, Th, Pb and Mn, as well as a variety of rare Earth elements (REE) including Sm, Eu, and
Nd (Frost et al., 2001; Gaft et al., 2005). In pegmatite-derived titanites, REE’s can account for over 4 wt
% of a crystal, although a typical titanite crystal usually contains less than .02 wt% of REE (Frost et al.,
2001). In addition, titanite can incorporate anywhere from 10 to over 100 ppm’s of U into its structure.
The geochemistry of titanite, in combination with its high closure temperature (max. 700oC), make it an
excellent geochronometer providing important U-Pb isotopic data used to date geologic events (Frost et
al., 2001).

There are multiple features in the Raman spectra of titanite that are associated with a variety of REE
luminescence centers, particularly Sm3+, Eu3+, and Nd3+. Gaft et al. (2005) attribute peaks centered at
574, 578, 589, 600, 613, 617, and 620 nm to Sm3+ centers (Fig. 11.4 A. & B.). Peaks centered at 689
and 703 nm are attributed to Eu3+ luminescence features with corresponding features located at 4263,
4545, and 4651 cm-1 in the Raman spectra (Fig. 11.4 A. & 11.4 C.) (Gaft et al., 2005). Luminescence
features located at 867, 883, and 894 nm are attributed to Nd3+ centers (Fig. 11.5 A. & B.) (Gaft et al.,
2005; Gaft et al., 2003).

There are multiple peaks present in the Raman spectra of titanite that have not been presented in the
literature, but are likely related to the luminescence centers of unidentified REE. The locations of the
peaks are as follows: 639 nm, 647 nm, 658 nm, 664 nm, 718 nm, and 729 nm (Fig. 11.4 A. & C.); 854
nm, 905 nm, 923 nm, 934 nm, and 972 nm (Fig. 11.5 A. & B.); 794 nm, 799 nm, and 806 nm (Fig. 11.4 A.
& 11.6).

Fig. 11.4 A. Peaks in the Raman spectra of titanite associated with Sm3+ and Eu3+ luminescence centers as well as
peaks likely related to unidentified REE centers; sample R050114, Pakistan, λexcitation = 532 nm, unoriented

?
Sm3+

?
Eu3+ ?

93
Fig. 11.4 B. Magnification of peaks attributed to Sm3+ luminescence centers in titanite located at 574 nm (5321394
cm-1), 578 nm (5321526 cm-1), 589 nm (5321720 cm-1), 600 nm (5322151 cm-1), 613 nm (5322491 cm-1), 617 nm (5322597
-1 532 -1
cm ), and 620 nm ( 2682 cm ); sample R050114, unoriented
Intensity

Sm3+
Sm3+
Sm3+

-1
Raman Shift (cm )

Fig. 11.4 C. Magnification of peaks attributed to Eu3+ luminescence centers in titanite located at 689nm (5324263
-1 532 -1 532 -1
cm ) and 703 nm ( 4545 cm and 4651 cm ) as well as peaks likely related to unidentified REE centers located
at 639 nm ( 3162 cm ), 647 nm ( 3360 cm ), 658 nm (5323619 cm-1), 664 nm (5323737 cm-1), 718 nm (5324870
532 -1 532 -1
-1 532 -1
cm ), and 729 nm ( 5086 cm ); sample R050114, Pakistan, unoriented
Intensity

Eu3+

?
?

Raman Shift (cm-1)

Fig. 11.5 A. Peaks related to Nd3+ as well as peaks likely related to unidentified REE centers in titanite from
Kingman, AZ; sample R050039, λexcitation = 785 nm, unoriented

Nd3+

?
?

94
Fig. 11.5 B. Magnification of peaks related to Nd3+ luminescence centers located at 867 nm (7851269 cm-1), 883 nm
(7851374 cm-1), and 894 nm (7851588 cm-1) and multiple peaks likely related to unidentified REE centers located at 854
785 -1 785 -1 785 -1 785 -1 785 -1
nm ( 1038 cm ), 905 nm ( 1696 cm ), 923 nm ( 1904 cm ), 934 nm ( 2039 cm ), and 972 nm ( 2455 cm )
in the Raman spectrum of titanite from Kingman, AZ; sample R050039, unoriented

Nd3+
?
Intensity

?
? ?

-1
Raman Shift (cm )

Fig. 11.6 Magnification of peaks located at 794 nm (5326217 cm-1), 799 nm

532 -1 532 -1
( 6290 cm ), and 806 nm ( 6393 cm ) likely related to unidentified REE
centers in titanite from Pakistan; sample R050114, unoriented

Intensity

? ?

Raman Shift (cm-1)

95
References

Fritsch, E. & Rossman, G. (1988) An update on color in gems. Part 3: Colors caused by band gaps and physical
phenomena. Gems & Gemology, Vol. 24, No. 2, pp. 81-102
Frost, B.R., Chamberlain, K.R. & Schumacher, J.C. (2001) Sphene (titanite): phase relations and role as a
geochronometer. Chemical Geology, Vol. 172, No. 1-2, pp. 131-148.
Gaft, M., Nagli, L., Reisfeld, R. & Panczer, G. (2003) Laser-induced time-resolved luminescence of natural titanite
CaTiOSiO4 . Optical Materials, Vol. 24, No. 1-2, pp. 231-241.
Gaft, M., Reisfeld, R. & Panczer, G. (2005) Modern Luminescence Spectroscopy of Minerals and Materials, Springer,
Heidelberg, Germany.
Ghose, S., Ito, Y. & Hatch, D.M. (1991) Paraelectric-antiferroelectric phase transition in titanite, CaTiSiO5 . Physics
and Chemistry of Minerals, Vol. 17, No. 7, pp. 591-603.
Hammer, V.M.F., Beran, A., Endisch, D. & Rauch, F. (1996) OH concentration in natural titanites determined by FTIR
spectroscopy and nuclear reaction analysis. European Journal of Mineralogy, Vol. 8, No. 2, pp. 281.
Isetti, G. & Penco, A.M. (1968) La posizione dell’Idrogeno ossidrilico nella titanite. Mineralogica et Petrographica
Acta, Vol. 14, pp. 115–122.
Nasdala, L., Irmer, G. & Wolf, D. (1995) The degree of metamictization in zircon; a Raman spectroscopic study.
European Journal of Mineralogy, Vol. 7, No. 3, pp. 471.
Salje, E., Schmidt, C. & Bismayer, U. (1993) Structural phase transition in titanite, CaTiSiO5: A Raman spectroscopic
study. Physics and Chemistry of Minerals, Vol. 19, No. 7, pp. 502-506.
Vance, E.R. & Metson, J.B. (1985) Radiation damage in natural titanites. Physics and Chemistry of Minerals, Vol. 12,
No. 5, pp. 255-260.
Zhang, M. & Salje, E.K.H. (2003) Spectroscopic characterization of metamictization and recrystallization in zircon and
titanite. Phase Transitions: A Multinational Journal, 76, Vol. 1, No. 2, pp. 117-136.
Zhang, M., Salje, E.K.H., Bismayer, U., Groat, L.A. & Malcherek, T. (2002) Metamictization and recrystallization of
titanite: An infrared spectroscopic study. American Mineralogist, Vol. 87, No. 7, pp. 882-890.

Additional Information

Beran, A. & Libowitzky, E. (2006) Water in natural mantle minerals II: olivine, garnet and accessory minerals.
Reviews in Mineralogy and Geochemistry, Vol. 62, No. 1, pp. 169-191.
Beran, A. & Libowitzky, E. (2003) IR spectroscopic characterization of OH defects in mineral phases. Phase
transitions, Vol. 76, No. 1-2, pp. 1-15.
Gaft, M., Nagli, L., Reisfeld, R. & Panczer, G. (2003) Laser-induced time-resolved luminescence of natural titanite
CaTiOSiO4. Optical Materials, Vol. 24, No. 1-2, pp. 231-241.
Johnson, E.A. (2006) Water in nominally anhydrous crustal minerals: speciation, concentration, and geologic
significance. Reviews in Mineralogy and Geochemistry, Vol. 62, No. 1, pp. 117-154.
Rossman, G.R. (2006) Analytical methods for measuring water in nominally anhydrous minerals. Reviews in
Mineralogy and Geochemistry, Vol. 62, No. 1, pp. 1-28.
Rossman, G.R. (1996) Studies of OH in nominally anhydrous minerals. Physics and Chemistry of Minerals, Vol. 23,
No. 4, pp. 299-304.

96
Topaz
Gem names: sherry topaz, Imperial topaz

Ideal chemistry: Al2SiO4F2

Crystal system: orthorhombic

Point Group: H-M: mmm S: D2h

Space Group: Pbnm topaz from Antonio Peireira Topaz mine,

Minas Gerais, Brazil

Table of Atomic Coordinates (Gatta et al., 2006):

atom x y z
Al 0.9053 0.13131 0.0816
Si 0.39976 0.94088 0.25
O1 0.7064 0.0311 0.25
O2 0.4516 0.7564 0.25
O3 0.2114 0.98979 0.09284
F 0.5982 0.25263 0.05988

Raman mode analysis:

(3 x 8d) + (3 x 4c)

Raman Active Modes

Atom Wyckoff Position Point Symmetry Ag A1g B2g B3g

Al, O3, F 8d 1 3 3 3 3
Si, O1, O2 4c m 2 1 2 1

Raman mode analysis predicts the existence of 54 active Raman modes in topaz:

15Ag + 12B1g + 15B2g +12B3g = 54

97
Spectral Features Related to OH-

In the crystal structure of topaz, two F atoms bond with an AlO4 group to form distorted octahedra.
Hydroxl (OH-) commonly substitutes for fluorine in varying concentrations, depending on the origin of the
topaz sample (Beny and Piriou, 1987). Multiple peaks associated with OH- are can be observed in
Raman spectra of topaz. Pinheiro et al. (2002) report the presence of a single peak centered at 3647
cm-1 in the Raman spectra of topaz containing less than 10 wt% OH- and associate this peak with the
substitution of F for OH- (Fig. 12.5). In samples containing greater than 10 wt% OH-, this peak becomes
asymetrical. The asymmetry is related to the presence of a second OH- Raman peak centered at 3639
cm-1 (Pinheiro et al., 2002). Pinheiro et al. (2002) suggest that the appearance of the second OH- peak at
3639 cm-1 is related to a change in the local symmetry from D162h to C92v producing two nonequivalent F
sites that can be occupied by OH-. A Raman peak centered at 1165 cm-1 has been assigned to the in-
plane stretching mode of hydroxl (Fig. 12.1) (Beny and Piriou, 1987). Pinheiro et al. (2002) report the
presence of two IR bands associated with hydroxyl in topaz; 1) a band centered at 1150 cm-1, related to
Al-OH- bending and 2) a band centered at 3600 cm-1 associated with an OH- stretching mode. OH-
content is not related to the various colors of topaz (see next section on chromophores in topaz).

FIG. 12.1 Raman peaks associated with hydroxyl in topaz; a peak centered at 3650 cm-1 is related to the
substitution of F for OH- and a peak centered at 1165 cm-1 is assigned to a stretching mode of OH- ; sample
R050176, colorless topaz from Nigeria, λexcitation = 532 nm, unoriented

OH-

-
OH

98
Spectral Features Related to Chromophores and Other Ions

Unlike beryl, the color varieties of topaz are dominantly the result of radiation-induced color centers,
rather than the result of electronic transitions produced by chromophoric ions. The pink color in topaz is
undoubtebly produced by substitution of octahedrally coordinated Al3+ for Cr3+ however, the details
surrounding irradiation-induced centers in other color varieties of topaz remain ambiguous (Taran et al.,
2003; Fritsch and Rossman, 1988). Gaft et al. (2005) attribute yellow color in topaz to Cr3+ in combination
with a radiation induced O- hole. Cr3+ combined with a F-center (vacancy with a trapped electron)
produce orange-red topaz (Gaft et al., 2005). Red-brown Cr-deficient topaz is colored by O- and F center
combinations (Gaft et al., 2005). Blue topaz is produced by R-centers: two F-centers with two trapped
electrons (Gaft et al., 2005). In a study of topaz from various Brazilian localities, Taran et al. (2003)
discovered that Cr3+ is also responsible for pink to violet colors, while Cr4+ produces a red-orange Imperial
topaz color, and a combination of Cr3+ and Cr4+ results in pink-orange colors.

Luminescence studies reported by Gaft et al. attribute peaks located at 680, 696, 712 and 730 nm to
multiple Cr3+ luminescence centers (Fig. 12.2) (Gaft et al., 2005).

FIG. 12.2 Two distinct peaks centered at 680 nm (5324075 cm-1) and 683 nm (5324160 cm-1) related to Cr3+ centers;
note that the luminescence peaks overwhelm the Raman peaks in this spectrum; sample R040121, pink topaz from
Mexico, unoriented

Cr3+
Cr3+

The Raman spectra of two brown topaz crystals from different mines in Minas Gerais, Brazil have a
unique set of peaks distinguishing them from the other color varieties of topaz in this study (Fig. 12.3 A. &
B.). Peaks centered at 678 nm, 683 nm, 712 nm, and 733 nm are likely related to the Cr3+ luminescence
centers observed by Gaft et al. (2005).

FIG. 12.3 A. Peaks centered at 678 nm (5324055cm-1), 683 nm (5324175 cm-1), 712 nm (5324762cm-1), and 733 nm
532 -1 3+
( 5155 cm ) in the Raman spectrum of topaz attributed to Cr luminescence centers; note that the luminescence
peaks overwhelm the Raman peaks in this spectrum; sample R060024, brown topaz from Minas Gerais, Brazil,
unoriented

Cr3+

99
FIG. 12.3 B. Magnification of peaks in Raman spectra attributed to Cr3+ luminescence centers; brown topaz from
Minas Gerais, Brazil, sample R060026 (blue spectrum), sample R060024 (black spectrum), λexcitation = 532 nm,
unoriented

Cr3+

Intensity

Cr3+

Raman Shift (cm-1)

References

Beny, J.M. & Piriou, B. (1987) Vibrational spectra of single-crystal topaz. Physics and Chemistry of Minerals, Vol. 15,
No. 2, pp. 148-159.
Fritsch, E. & Rossman, G. (1988) An update on color in gems. Part 3: Colors caused by band gaps and physical
phenomena. Gems & Gemology, Vol. 24, No. 2, pp. 81-102.
Gaft, M., Reisfeld, R. & Panczer, G. (2005) Modern Luminescence Spectroscopy of Minerals and Materials, Springer,
Heidelberg, Germany.
Gatta, G.D., Nestola, F., Bromiley, G.D. & Loose, A. (2006) New insight into crystal chemistry of topaz: A multi-
methodological study. American Mineralogist, Vol. 91, No. 11-12, pp. 1839-1846.
Pinheiro, M.V.B., Fantini, C., Krambrock, K., Persiano, A.I.C., Dantas, M.S.S. & Pimenta, M.A. (2002) OH/F
substitution in topaz studied by Raman spectroscopy. Physical Review B, Vol. 65, No. 10, pp. 104301.
Ribbe, P.H. & Gibbs, G.V. (1971) The crystal structure of topaz and its relation to physical properties. American
Mineralogist, Vol. 56, pp. 24-30.
Taran, M.N., Tarashchan, A.N., Rager, H., Schott, S., Schürmann, K. & Iwanuch, W. (2003) Optical spectroscopy
study of variously colored gem-quality topazes from Ouro Preto, Minas Gerais, Brazil. Physics and Chemistry of
Minerals, Vol. 30, No. 9, pp. 546-555.

Additional Information

Dickinson, A.C. & Moore, W.J. (1967) Paramagnetic resonance of metal ions and defect centers in topaz. Journal of
Physical Chemistry, Vol. 71, No. 2, pp. 231-240.
Fritsch, E. & Rossman, G.R. (1988) An update on color in gems. Part 2: Colors involving multiple atoms and color
centers. Gems & Gemology, Vol. 24, No. 1, pp. 3–15.
Fritsch, E. & Rossman, G.R. (1987) An Update on Color in Gems. Part 1. Introduction and Colors caused by
Dispersed Metal Ions. Gems & Gemology, Vol. 23, No. 3, pp. 126–139.
3+
Gaft, M., Nagli, L., Reisfeld, R., Panczer, G. & Brestel, M. (2003) Time-resolved luminescence of Cr in topaz
Al2SiO4 (OH, F) 2. Journal of Luminescence, Vol. 102, pp. 349-356.
Kloprogge, J.T. & Frost, R.L. (2000) Raman microscopic study at 300 and 77 K of some pegmatite minerals from the
Iveland–Evje area, Aust-Agder, Southern Norway. Spectrochimica Acta Part A: Molecular and Biomolecular
Spectroscopy, Vol. 56, No. 3, pp. 501-513.
Komatsu, K., Kagi, H., Okada, T., Kuribayashi, T., Parise, J.B. & Kudoh, Y. (2005) Pressure dependence of the OH-
stretching mode in F-rich natural topaz and topaz-OH. American Mineralogist, Vol. 90, No. 1, pp. 266-270.
Krambrock, K., Ribeiro, L.G.M., Pinheiro, M.V.B., Leal, A.S., Menezes, M. B.C. & Spaeth, J.M. (2007) Color centers
in topaz: comparison between neutron and gamma irradiation. Physics and Chemistry of Minerals, Vol. 34, No.
7, pp. 437-444.
Londos, C., Vassilkou-Dova, A., Georgiou, G. & Fytros, L. (1992) Infrared studies of natural topaz. Physica status
solidi A, Applied research, Vol. 133, No. 2, pp. 473-479.

100
Rossman, G.R. (2006) Analytical methods for measuring water in nominally anhydrous minerals. Reviews in
Mineralogy and Geochemistry, Vol. 62, No. 1, pp. 1-28.
Rossman, G.R. (1996) Studies of OH in nominally anhydrous minerals. Physics and Chemistry of Minerals, Vol. 23,
No. 4, pp. 299-304.
Schott, S., Rager, H., Schurmann, K. & Taran, M. (2003) Spectroscopic study of natural gem quality “Imperial"
topazes from Ouro Preto, Brazil. European Journal of Mineralogy, Vol. 15, No. 4, pp. 701.
Shinoda, K. & Aikawa, N. (1997) IR active orientation of OH bending mode in topaz. Physics and Chemistry of
Minerals, Vol. 24, No. 8, pp. 551-554.
Souza, D.N., Fernandes de Lima, J., Valerio, M.E.G., Fantini, C., Pimenta, M.A., Moreira, R.L. & Caldas, L.V.E.
(2002) Influence of thermal treatment on the Raman, infrared and TL responses of natural topaz. Nuclear Inst.
and Methods in Physics Research, B, Vol. 191, No. 1-4, pp. 230-235.
3+
Tarashchan, A.N., Taran, M.N., Rager, H. & Iwanuch, W. (2006) Luminescence spectroscopic study of Cr in
Brazilian topazes from Ouro Preto. Physics and Chemistry of Minerals, Vol. 32, No. 10, pp. 679-690.

101
Tourmaline
General Formula:
X Y3 Z6 Si6O18 (BO3)3 (OH,O, F)4

Crystal system: Trigonal

Gem names: rubellite, indicolite, Paraiba,

watermelon

Point Group: H-M: 3m S: C3v

Zoned Elbaite (“watermelon” tourmaline)
Space Group: R3m

List of Ideal Chemistries of Major Gem Tourmalines:

Elbaite: Na(Al1.5Li1.5)Al6(BO3)3Si6O18(OH)4

Liddicoatite: Ca(Li2Al)Al6(BO3)3Si6O18(OH)3F

Dravite: NaMg3Al6(BO3)3Si6O18(OH)4

Uvite: CaMg3(Al5Mg)(BO3)3Si6O18(OH)3F

Note: Raman mode analysis and atomic coordinates for elbaite are provided below

Table of Atomic Coordinates (Bosi et al., 2005):

atom x y z
NaX 0 0 0.23267
AlY 0.12318 0.06159 0.62744
AlZ 0.29758 0.26086 0.61151
B 0.10969 0.21938 0.45457
SiT 0.19192 0.18996 0
O1 0 0 0.78262
O2 0.06091 0.12182 0.48293
O3 0.26857 0.13429 0.50990
O4 0.09335 0.18670 0.07149
O5 0.18670 0.09335 0.09304
O6 0.19695 0.18684 0.77565
O7 0.28554 0.28578 0.08043
O8 0.20989 0.27071 0.44137
H3 0.2664 0.1332 0.3818

102
Raman mode analysis:

Raman Active
Modes
Wyckoff
Atoms A1 E
Position
AlZ, SiT, O6,
18c 3 6
O7, O8
AlY, B, O2, O3,
9b 2 3
O4, O5, H3
NaX, O1 3a 1 1

(5 × 18c) + (7 × 9b) + (2 × 3a)

Raman mode analysis predicts the existence of active Raman modes in :

31A1 + 53E = 84

103
Spectral Features Related to OH-

Multiple studies have been conducted to interpret IR bands associated with hydroxyl in various tourmaline
species. OH- can substitute into two positions in the crystal structure of tourmaline (Gonzales-Carreño,
1988; Castañeda et al., 2000). The first position, denoted OH1, is located at the center of the hexagonal
rings with the oxygen coordinated by three octahedral Y-site cations (Fig. 13.1) (Gonzales-Carreño, 1988;
Castañeda et al., 2000). In the second position, denoted OH3, the hydroxyl occupies a position along the
edge of the hexagonal columns where the oxygen is coordinated by one Y and two Z cations (Fig. 13.1)
(Gonzales-Carreño, 1988; Castañeda et al., 2000). The type of cations occupying the Y- and Z-sites in
the crystal structure affect the frequency and bandwidth of the OH- bands (Gonzales-Carreño, 1988). In
general, the highest frequency bands (3600+ cm-1) are associated with OH1 while those appearing at
3600-3400 cm-1 are assigned to OH3 vibrations (Castañeda et al., 2000). Factor group analysis
conducted by Castaneda et al. (2000) predicts three IR active modes for OH- in tourmaline, one for OH1
and two for OH3; only two of these modes are Raman active. Two tables comparing the locations of IR
bands attributed to OH- observed in the literature with Raman peaks positions observed in this study are
provided below (Table 13.1 & 13.2).

Table 13.1 Locations of IR bands attributed OH- and their associated coordinated cations in tourmaline

Tourmaline
OH1 OH3 H2O Reference
Variety
Location Coordinated Location Coordinated
(cm-1) Cations (cm-1) Cations
Dravite 3738 MgYMgYMgY 3568 R2+YAlZAlZ
Schorl 3738 MgYMgYMgY
Schorl 3633 FeYFeYFeY 3553 R2+YAlZAlZ
3650 LiYAlYAlY 3583 LiYAlZAlZ
Elbaite
3463 AlYAlZAlZ
3650/3680 LiYAlYAlY 3586/3604 LiYAlZAlZ
Ca-Elbaite Gonzales-
3474/3507 AlYAlZAlZ
Y Y Y Carreño,
3670 Li Mn Al 3594 LiYAlZAlZ 1988
3568 R2+YAlZAlZ
Mn-Elbaite
3492 AlYAlZAlZ

3692 LiYAlYFeY 3594 LiYAlZAlZ

Fe-Elbaite 3558 R2+YAlZAlZ
3492 AlYAlZAlZ

3641, 3457-60, 3170,

Elbaite (Al,Li) YAlYAlY (Al,Li)YAlZAlZ
3646-47 3580-82 3340
3165, Castañeda et
3478-79,
Y Y Y Y Z Z 3175, al., 2000
Fe-Elbaite 3585, 3592 Li Al Fe 3484, 3557, (Fe,Li,Al) Al Al
3360,
3560
3380

3468-70,
3645,
(Al,Li) YAlYAlY 3478, AlYAlZAlZ
Tourmaline 3648-51 Zhang et al.,
3490-92
(various) Y Y Y Y Z Z 2008
3628-31 Fe Al Al 3559, 3564 Fe Al Al
3585-3597 LiYAlZAlZ

104
Table 13.2 Raman peaks in tourmaline associated with hydroxyl and the possible cation coordination assignments
based on IR data; 532 nm laser, unoriented

OH
Tourmaline Possible
Sample # peaks Figure Reference
Variety Assignment
cm-1
Gonzales-Carreño,
3496 OH3 AlYAlZAlZ
1988
R050487, R060652, 13.2 B. Zhang et al., 2008;
Elbaite OH3 FeYAlZAlZ,
R050260 3566 & C. Gonzales-Carreño,
OH3(R2+YAlZAlZ)
1988
3598 OH3 LiYAlZAlZ Zhang et al., 2008
Gonzales-Carreño,
3476 OH3 AlYAlZAlZ
R060560, R050119, 13.2 A. 1988
Elbaite 3592 OH1 LiYAlYFeY Castañeda et al., 2000
R060003, R060566 & C.
3656 ?
3511 ? 13.3 A.
Liddicoatite R060635
3609 ? & C.
3489 OH3 AlYAlZAlZ Zhang et al., 2008
Y Z Z 13.3 B.
Liddicoatite R060969 3595 OH3 Li Al Al Zhang et al., 2008
& C.
3657 ?
3571 ?
13.4 A.
Dravite R050077 Gonzales-Carreño,
3740 OH1MgYMgYMgY & B.
1988
OH1 LiYAlYFeY, Castañeda et al., 2000;
3584 13.5 A.
Uvite R050172 OH3 LiYAlZAlZ Zhang et al., 2008
& B.
3743 ?

Fig. 13.1 The crystal structure of elbaite

viewed down the c-axis showing sites of
hydroxyl substitution; red spheres: oxygen;
blue spheres: Al; green spheres:
substitutional site for OH3; pink sphere
substitutional site for OH1; teal cation below
OH1 substutional site is Na OH3

OH1

105
Fig. 13.2 A. Peaks associated with OH in Raman spectra of elbaite; 1) sample R050119, 2) R060560, 3)
R060003, 4) R060566; 532 nm laser, unoriented

-
OH

4.

3.

2.

1.

Fig. 13.2 B. Peaks associated with OH in Raman spectra of elbaite; 5) sample R060562, 6) R050260, 7)
R050487; 532 nm laser, unoriented

OH-

7.

6.

5.

106
Fig. 13.2 C. Magnification of peaks associated with OH in elbaite, note the variation in frequency and peak shape;
1) sample R050119, 2) R060560, 3) R060003, 4) R060566, 5) R060562, 6) R050260, and 7) R050487; 532 nm laser,
unoriented

7.

Intensity

6.

5.

4.

3.

2.

1.

Raman Shift (cm-1)

Fig. 13.3 A. Peaks associated with OH in the Raman spectrum of liddicoatite; 1) sample R060635, brown gem of
unknown locality; 532 nm laser, unoriented

OH-

107
Fig. 13.3 B. Peaks associated with OH in the Raman spectrum of liddicoatite; sample R060969, red crystal from
Madagascar; 532 nm laser, unoriented

OH-

Fig. 13.3 C. Magnification of peaks associated with

OH in liddicoatite, note the variation in frequency and
peak shape; 1) sample R060635; centered at 3511
-1
and 3609 cm (no literature match), brown gem of
unknown locality; 2) sample R060969, centered at
3489 and 3595 cm-1, red crystal from Madagascar,
Intensity

532 nm laser, unoriented

2.

1.

Raman Shift (cm-1)

Fig. 13.4 A. Peaks associated with OH in dravite; sample R050077, dark brown crystal from Western Australia, 532
nm laser, unoriented

OH-

108
Fig. 13.4 B. (left) Magnification of peaks centered at 3571 and 3740 cm-1
associated with OH in dravite; sample R050077, dark brown crystal from
Western Australia, 532 nm laser, unoriented
Intensity

OH-

Raman Shift (cm-1)

Fig. 13.5 A. Peaks associated with OH in uvite; sample R050172, 532 nm laser, unoriented

OH-

Fig. 13.5 B. (right) Magnification of peaks centered 3584 and

3743 cm-1 associated with OH in uvite; sample R050172, 532
nm laser, unoriented
Intensity

OH-

Raman Shift (cm-1)

109
Spectral Features Related to Chromophores

Tourmaline comes in a wide variety of colors due to its chemical complexity (see “Table of Color Causes”,
Appendix). Several members of the tourmaline group are often multi-colored and zoned due to
compositional changes during crystal growth (Dunn et al., 1977). Liddicoatite (the structural analog of
elbaite) from Madagascar is famous for its striking color patterns (Dunn et al., 1977). Elbaite, uvite, and
dravite commonly incorporate iron into their crystal structures, Fe2+ into the Z-site and Fe3+ into the Y- and
Z-sites, and this greatly affects the color (Mattson and Rossman, 1987). There are many colors of
tourmaline for which the dominant color-inducing mechanism is charge transfer involving Fe including
blue, green, and brown elbaite and liddicoatite and yellow to brown dravite (Fritsch and Rossman, 1988).
Green dravite and uvite are produced by octahedrally coordinated Cr3+ and V3+ (Fritsch and Rossman,
1988). The presence of Mn is likely responsible for yellow-green, pink, and red tourmalines (Fritsch and
Rossman, 1988). Recent studies of the newly discovered cuprian elbaite from Paraiba, Brazil attribute its
distinctive, nearly neon blue and green hues to a high concentration (greater than 0.1 wt%) of Cu2+
(Rossman et al., 1991). Although extensive studies have been conducted to understand the processes
that produce color in tourmaline, further investigation is necessary in order to fully comprehend the
details.

Several spectral features related to chromophores have been observed in the Raman spectra of
tourmaline. Gaft et al. (2005) attribute luminescence emission centers located at 697, 707, 691/692 nm
to the presence of Cr3+ in tourmaline. Peaks likely related to the presence of Cr3+ centers were observed
in the Raman spectra of uvite centered at 680 nm and 683 nm (Fig. 13.6 A. & B.), and dravite centered at
4185 cm-1 (684 nm) (Fig. 13.7 A. & B.). Although the positions of these peaks do not match those
provided by Gaft et al. (2005) for Cr3+ luminescence centers in tourmaline, they do correspond to the
locations of trivalent chromium centers in several other minerals including beryl, chrysoberyl, and topaz.

Fig. 13.6 A. Doublet associated with Cr3+ luminescence centers in uvite; note that the luminescence peaks
overwhelm the Raman peaks in this spectrum; sample R050301, Bahia, Brazil, λexcitation = 532 nm, unoriented

Cr3+

110
Fig. 13.7 A. Peak centered at 684 nm (5324185 cm-1) associated with a Cr3+ luminescence center in dravite; note
that the luminescence peaks overwhelm the Raman peaks in this spectrum; sample R040088, brown crystal from
Western Australia; unoriented

Cr3+

Fig. 13.6 B. (right) Magnification of doublet centered at 680

3+
(5324110 cm-1) and 683 nm (5324160 cm-1) associated with Cr in
uvite; sample R050301 from Bahia Brazil, unoriented

Cr3+
Intensity

Raman Shift (cm-1)

Cr3+
Intensity

Fig. 13.7 B. (left) Magnification of peak centered at 684 nm

532 -1 3+
( 4185 cm ) associated with a Cr luminescence center in
dravite; sample R040088, brown crystal from Western Australia;
unoriented

Raman Shift (cm-1)

111
Spectral Features Related to Rare-earth Elements

The geochemistry of REE distribution in tourmalines is a valuable tool for modeling the development of
various hydrothermal ore deposits, particularly gold deposits (Jiang et al., 2004). Sm-Nd isotopic data
can help constrain pressure, temperature, and depth conditions as well as preserve geochemical
information regarding ore-forming fluids (Jiang et al., 2004). Although tourmaline typically contains only
very small concentrations of REE (ppm’s), its resistance to weathering and ability to precipitate from
hydrothermal systems over a broad range of conditions make it a viable marker mineral for economic
geologists (Jiang et al., 2004).

Gaft et al. (2003) report the presence of multiple luminescence centers attributed to Nd3+ in titanite
located at 860 nm, 869 nm, 878 nm, 896 nm, and 1053 nm. The same centers are observed in the
Raman spectra of elbaite and liddicoatite in this study (Fig. 13.8 & 13.9).

Fig. 13.8 Peaks located at 860 nm (7801196 cm-1), 869 nm (7801314 cm-1), 878 nm (7801433 cm-1), 896 nm (7801668
-1 780 -1 3+
cm ), and 1053 nm ( 3323 cm ) associated with Nd centers in elbaite; note that the luminescence peaks
overwhelm the Raman peaks in this spectrum; sample R060631, unknown locality, unoriented

Nd3+
Nd3+

Fig. 13.9 Peaks located at 860 nm (7801196 cm-1), 869 nm (7801314 cm-1), 878 nm (7801433 cm-1), 896 nm (7801668
3+
cm-1), and 1053 nm (7803323 cm-1) associated with Nd centers in liddicoatite; note that the luminescence peaks
overwhelm the Raman peaks in this spectrum; sample R060635, unknown locality, unoriented

Nd3+
Nd3+

112
References

Bosi, F., Agrosi, G., Lucchesi, S., Melchiorre, G. & Scandale, E. (2005) Mn-tourmaline from island of Elba (Italy):
Crystal chemistry. American Mineralogist, Vol. 90, No. 10, pp. 1661-1668.
Castañeda, C., Oliveira, E.F., Gomes, N. & Soares, A.C.P. (2000) Infrared study of OH sites in tourmaline from the
elbaite-schorl series. American Mineralogist, Vol. 85, No. 10, pp. 1503-1507.
Dunn, P.J., Appleman, D.E. & Nelen, J.E. (1977) Liddicoatite, a new calcium end-member of the tourmaline group.
American Mineralogist, Vol. 62, No. 11-12, pp. 1121-1124.
Fritsch, E. & Rossman, G. (1988) An update on color in gems. Part 3: colors caused by bandgaps and physical
phenomena. Gems & GemoA logy, Vol. 24, No. 2, pp. 81-102.
Gaft, M., Nagli, L., Reisfeld, R. & Panczer, G. (2003) Laser-induced time-resolved luminescence of natural titanite
CaTiOSiO4 . Optical Materials, Vol. 24, No. 1-2, pp. 231-241.
Gaft, M., Reisfeld, R. & Panczer, G. (2005) Modern Luminescence Spectroscopy of Minerals and Materials, Springer.
Gonzalez-Carreño, T., Fernandez, M. & Sanz, J. (1988) Infrared and electron microprobe analysis of tourmalines.
Physics and Chemistry of Minerals, Vol. 15, No. 5, pp. 452-460.
Jiang, S.Y., Yu, J.M. & Lu, J.J. (2004) Trace and rare-earth element geochemistry in tourmaline and cassiterite from
the Yunlong tin deposit, Yunnan, China: implication for migmatitic–hydrothermal fluid evolution and ore genesis.
Chemical Geology, Vol. 209, No. 3-4, pp. 193-213.
2+ 3+
Mattson, S.M. & Rossman, G.R. (1987) Fe -Fe interactions in tourmaline. Physics and Chemistry of Minerals, Vol.
14, No. 2, pp. 163-171.
Rossman, G.R., Fritsch, E. & Shigley, J.E. (1991) Origin of color in cuprian elbaite from São José de Batalha,
Paraíba, Brazil. American Mineralogist, Vol. 76, No. 9-10, pp. 1479-1484.
Zhang, A., Wang, R., Li, Y., Hu, H., Lu, X., Ji, J. & Zhang, H. (2008) Tourmalines from the Koktokay No. 3 pegmatite,
Altai, NW China: spectroscopic characterization and relationships with the pegmatite evolution. European
Journal of Mineralogy, Vol. 20, No. 1, pp. 143.

Additional Information

De Oliveira, E.F., Castañeda, C., Eeckhout, S.G., Gilmar, M.M., Kwitko, R.R., De Grave, E. & Botelho, N.F. (2002)
Infrared and Mossbauer study of Brazilian tourmalines from different geological environments. American
Mineralogist, Vol. 87, No. 8-9, pp. 1154-1163.
Gasharova, B., Mihailova, B. & Konstantinov, L. (1997) Raman spectra of various types of tourmaline. European
Journal of Mineralogy, Vol. 9, No. 5, pp. 935.
Manning, P.G. (1973) Effect of second-nearest-neighbour interaction on Mn3+ absorption in pink and black
tourmalines. Canadian Mineralogist, Vol. 11, No. 5, pp. 971.
Peng, M., Mao, H.K., Chen, L.G. & Chan, E.E.T. (1988) The Polarized Raman Spectra of Tourmaline. Annual Report
of the Director of the Geophysical Laboratory, Carnegie Inst.Washington, Vol. 1989.

113
Zircon
Ideal chemistry: ZrSiO4

Crystal system: tetragonal

Point Group: H-M: 4/mmm S: D4h

Space Group: I41/amd

Photo courtesy of Stone Group Labs

Table of Atomic Coordinates (Finch et al., 2001):

atom x y z
Si 0 0.75 5/8
Zr 0 0.75 1/8
O 0 0.0657 0.1961

Raman mode analysis:

Raman Active Modes

Atom Wyckoff Position Point Symmetry A1g B1g B2g Eg

O 16h m 2 2 1 3
Si 4b 4 2m - 1 - 1
Zr 4a 4 2m - 1 - 1

Raman mode analysis predicts the existence of 12 active Raman modes in zircon:

2A1g + 4B1g + 1B2g + 5E g = 12

114
Introduction to Raman Spectrum of Zircon

Zircon is an isosilicate consisting of isolated SiO4 groups and dodecahedrally coordinated Zr atoms
creating open channels parallel to the c-axis in the crystal structure (Fig. 14.2). The Raman spectrum of
zircon is presented here (Fig. 14.1). The most intense peak in the spectrum, centered at 357 cm-1, is
associated with restricted rotation of SiO4 groups (Kolesov et al., 2001). The next most intense peak,
centered at 1008 cm-1, is associated with Si-O stretching modes, while the peak centered at 438 cm-1, is
associated with O-Si-O bending (Kolesov et al., 2001).

Fig. 14.1 The Raman spectrum of zircon; mode assignments reported by Kolesov et al. (2001); sample R050203,
532 nm laser, processed, unoriented

357
Eg
439 1008
A1g B1g
Intensity

215
B1g
225
Eg 974
A1g
202 393
Eg B1g

Raman Shift (cm-1)

Fig. 14.2 The crystal structure of zircon; blue tetrahedra: SiO4 groups, pink polyhedra: ZrO8 groups

A. View down the a-axis B. View down the c-axis

115
Orientational Dependence of Spectra

When zircon is oriented with the a-axis parallel to the incident laser beam (the cross-section of the
uniaxial optical indicatrix is an ellipse), the intensities of the peaks change dramatically with rotation due
to a change in the degree of freedom of atomic vibration (Fig. 14.3 A). Notice that when the laser is
polarized parallel to the c-axis, the peaks centered at 968 and 1001 cm-1, associated with Si-O stretching
modes nearly disappear. When the sample is oriented with the c-axis parallel to the laser (the cross-
section of the uniaxial optical indicatrix is a circle), there is no noticeable change in peak intensities due to
an equal distribution of atomic vibration in all directions (FIG. 14.3 B). In spite of the orientational
dependence of its Raman peaks, zircon can be accurately identified by the spectrum of a randomly
oriented sample.

Fig. 14.3 Raman spectra of oriented corundum crystals; sample R050034, reddish brown zircon, Oaxaca, Mexico,
processed

A. Raman spectra showing peak intensities as a function of orientation; 514.5 nm laser parallel to a*; at 0° laser is
polarized ll to c-axis; at 90° laser is polarized perpendicular to c-axis; sample R050034, processed

B. Raman spectra showing peak intensities as a function of orientation. Notice that the intensities do not change as
the direction of polarization changes; 514.5 nm laser parallel to c-axis; at 0° laser is polarized ll to b-axis; at 90° laser
is polarized perpendicular to b-axis

116
Cause of Color in Zircon

The color centers of zircon are complex and the details surrounding the color-inducing mechanisms are
still debated. Zircon comes in a variety of colors and most of these colors are generated by natural
irradiation. Zircon, like titanite, frequently incorporates radioactive atoms such as U and Th into its crystal
structure. Red zircon have radiation-induced color centers in which Nb4+ substitutes for Zr4+ (Fielding,
1970; Fritsch and Rossman, 1988). Blue zircon is attributed to the presence of U4+ (Mackey et al., 1975;
Fritsch and Rossman, 1988). No spectral features attributed to these color centers have been observed
in this study.

Metamictization Due to Radiation Damage

Zircon typically incorporates radioactive atoms like U and Th into its crystal structure. As these elements
radioactively decay over time, the crystal structure is compromised resulting in metamictization
(breakdown of the crystal structure) (Johnson, 2006). Metamictization can make a crystal slightly or
completely amorphous depending on the amount of radiation exposure; this can affect the X-ray
diffraction patterns, unit cell parameters, and IR and Raman spectra of the crystal (Zhang et al., 2000;
Nasdala et al., 1995). Raman spectroscopic studies of radiation-damaged zircons conducted by Zhang et
al. (2000) and Nasdala et al. (1995) show widening of Raman peaks as well as frequency shifts due to an
increase in the irregularity of bond lengths and angles in the crystal structure. In particular, the peaks
associated with the stretching modes of Si-O, located at 900-1000 cm-1, broadened and although the
zircon structure was damaged, Zhang et al. (2000) conclude that in comparison with nearly amorphous
silica, the SiO4 groups in metamict zircon are less polymerized. The Raman spectra of metamict zircon
display peaks associated with both a damaged crystalline zircon, and an amorphous zircon (observed
between 850-1100 cm-1) (Zhang et al., 2000). Comparisons of diffraction peaks and Raman spectra of
zircon samples that may have undergone very slight radiation damage are shown in Fig. 14.4. The unit
cell parameters of the zircon samples included in this study are provided in Table 14.1.

Fig. 14.4 (Right) Portion of the XRD pattern of zircon with associated band width values and (left) portion of the
Raman spectra of zircon showing possible slight radiation damage; 1) sample R050034, 2) sample R050286, 3)
sample R050203, 4) R050488 (this sample has likely been exposed to the most radiation); 514 nm laser, processed

.126
Intensity

Intensity

4.
.120

3. .099

2.
.095
1.

-1
Raman Shift (cm ) 2 Theta

117
Table 14.1 Unit cell parameters of zircon samples from RRUFF database

Zircon color, Origin, &

Sample number
Unit cell parameters Structural State Reference

a (Å) c (Å) V (Å3)

red-brown; Oaxaca, likely well
6.6077(1) 5.9957(2) 261.78(1) RRUFF database
Mexico (R050034) crystallized
brownish red; Renfrew
likely well
Cty., Ontario, Canada 6.6132(3) 6.0038(3) 262.57(2) RRUFF database
crystallized
(R050286)
dark red; Eastern Thailand likely well
6.6049(1) 5.9801(2) 260.88(1) RRUFF database
(R050203) crystallized
brown; Sigulani Village,
likely well
Tambani Area, Nyassaland 6.6058(5) 5.9818(8) 261.03(4) RRUFF database
crystallized
(R050488)

Spectral Features Related to OH-

The details behind the incorporation of OH- and H2O into various structural sites of zircon remain
controversial. As with titanite, an increase in metamictization results in an increase in OH- concentration.
Well-crystallized zircon exhibit sharp, anisotropic IR peaks associated with OH-, whereas the IR spectra of
damaged crystals usually display an additional peak associated with the presence of H2O molecules
(Beran and Libowitsky, 2003). IR and Raman studies performed by Nasdala et al. (2001) confirm the
presence of at least three peaks centered at 3180, 3385, and 3420 cm-1 associated with OH- defects in
crystalline zircon (Fig. 14.5 A & B) (Dawson et al., 1971).

Fig. 14.5 A. Raman peak centered at 3424 cm-1 attributed to OH-; sample R050286, natural red zircon, Ontario,
Canada, λexcitation = 532 nm, unoriented

OH-

118
Fig. 14.5 B. Magnification of Raman peak centered at 3424
cm-1 attributed to OH-; sample R050286, natural red zircon,
Ontario, Canada, λexcitation = 532 nm, unoriented

OH-

Intensity
Raman Shift (cm-1)

119
Spectral Features Related to Rare-earth Elements

Like titanite, zircon crystals commonly contain a wide variety of trace elements including, but not limited
to, REE, Y, Hf, P, and radioactive atoms, such as U and Th (Hanchar et al., 2001). U-Pb dating
constrained by zircon crystals has provided the timing of nearly all geologic events. The high closure
temperature of zircon along with its ability to isotopically record geologic processes, make zircon
invaluable to geoscientists (Hanchar et al., 2001).

There are numerous luminescence features associated with REE present in the Raman spectra of zircon,
specifically peaks associated with Eu3+, Sm3+, and Nd3+ centers. Luminescence studies conducted by
Gaft et al. (2005) report the presence of Eu3+ centers located at 591, 596, 604, and 614 nm (Fig. 14.6 A.
& B.). Additional luminescence features located at 702 and 707 nm in the Raman spectra may also be
related to Eu3+ (Fig. 14.6 A. & C.) (Gaft et al., 2005). Peaks related to Sm3+ luminescence centers in
zircon are centered at 660 and 666 nm (Fig. 14.6 A. & D.) (Gaft et al., 2005). Gaft et al. (2005) report the
presence of peaks located at 870 nm, 879 nm, and 891 nm that are likely related to luminescence centers
produced by Nd3+ (Fig. 14.6 A. & B.) (Gaft et al., 2005).

There are multiple peaks in the Raman spectra of zircon that have not been presented in the literature,
but are likely related to the presence of unidentified REE luminescence centers. The locations of the
peaks are as follows: 719 nm, 722 nm, 728 nm (Fig. 14.6 A. & 14.5 A.); 797 nm, 803 nm, and 810 nm
(Fig. 14.6 A & 14.8 B.); 909 nm, 917 nm, 928 nm, and 976 nm (Fig. 14.7 A & 14.9).

Fig. 14.6 A. Peaks in the Raman spectra of zircon associated with: Eu3+ and Sm3+ luminescence centers, a peak
associated with OH-, and multiple peaks of unknown origins; sample R050203, natural red zircon, Thailand, λexcitation =
532 nm, unoriented

Eu3+
?
?
3+
Sm
OH-
Eu3+

Fig. 14.6 B. Magnification of peaks in the Raman spectrum of zircon attributed to Eu3+ centers located at 591 nm
532 -1 532 -1 532 -1 532 -1
( 1910 cm ), 596 nm ( 2012 cm ), 604 nm ( 2260 cm ), and 614 nm ( 2560 cm ); sample R050203, natural
red zircon, Thailand, unoriented

Eu3+
Intensity

Eu3+

-1
Raman Shift (cm )
120
Fig. 14.6 C. Magnification of peaks in the Raman spectrum of zircon related to Eu3+ centers located at 702 nm
(5324572 cm-1) and 707 nm (5324728 cm-1); sample R050203, natural red zircon, Thailand, unoriented

Eu3+
Intensity

-1
Raman Shift (cm )

Fig. 14.6 D. Magnification of a peak attributed to OH- centered at 3424 cm-1 and peaks related to Sm3+
luminescence centers located at 660 nm (5323654 cm-1) and 666 nm (5323772 cm-1) in the Raman spectrum of zircon;
sample R050203, natural red zircon, Thailand, unoriented

Sm3+
Intensity

OH-

-1
Raman Shift (cm )

Fig. 14.7 A. Peaks attributed to Nd3+ luminescence centers and unidentified REE; sample R050286, natural red
zircon, Ontario, Canada, λexcitation = 780 nm, unoriented

Nd3+

121
Fig. 14.7 B. Magnification of peaks attributed to Nd3+ luminescence centers located at 870 nm (7801327 cm-1), 879
nm (7801450 cm-1), and 891 nm (7801600 cm-1); sample R050286, natural red zircon, Ontario, Canada, unoriented

Nd3+
Intensity

Nd3+

Raman Shift (cm-1)

Fig. 14.8 A. (right) Magnification of peaks centered at 719 nm

(5324882 cm-1), 722 nm (5324950 cm-1), and 728 nm (5325073 cm-1)
likely associated with the luminescence centers of unidentified
REE; sample R050203, natural red zircon, Thailand, unoriented
Intensity

Raman Shift (cm-1)

Intensity

Fig. 14.8 B. (left) Magnification of peaks centered at

532 -1 532 -1
797 nm ( 6254 cm ), 803 nm ( 6348 cm ), and 810
532 -1
nm ( 6450 cm ) likely associated with the
luminescence centers of unidentified REE; sample
R050203, natural red zircon, Thailand, unoriented

Raman Shift (cm-1)

122
Fig. 14.9 Magnification of peaks centered at 909 nm (7801819 cm-1), 917 nm (7801922 cm-1), 928 nm (7802047 cm-1),
and 976 nm (7802580 cm-1) likely associated with the luminescence centers of unidentified REE; sample R050286,
natural red zircon, Ontario, Canada, unoriented
Intensity

Raman Shift (cm-1)

References

Beran, A. & Libowitzky, E. (2003) IR spectroscopic characterization of OH defects in mineral phases. Phase
transitions, Vol. 76, No. 1-2, pp. 1-15.
Dawson, P., Hargreave, M.M. & Wilkinson, G.R. (1971) The vibrational spectrum of zircon (ZrSiO4). Journal of
Physics C: Solid State Physics, Vol. 4, pp. 240-256.
Finch, R.J., Hanchar, J.M., Hoskin, P.W.O. & Burns, P.C. (2001) Rare-earth elements in synthetic zircon: Part 2. A
single-crystal X-ray study of xenotime substitution. American Mineralogist, Vol. 86, No. 5-6, pp. 681-689.
Gaft, M., Panczer, G., Reisfeld, R. & Uspensky, E. (2001) Laser-induced time-resolved luminescence as a tool for
rare-earth element identification in minerals. Physics and Chemistry of Minerals, Vol. 28, No. 5, pp. 347-363.
Gaft, M., Reisfeld, R. & Panczer, G. (2005) Modern Luminescence Spectroscopy of Minerals and Materials, Springer,
Heidelberg, Germany.
Hanchar, J.M., Finch, R.J., Hoskin, P.W.O., Watson, E.B., Cherniak, D.J. & Mariano, A.N. (2001) Rare earth
elements in synthetic zircon: Part 1. Synthesis, and rare earth element and phosphorus doping. American
Mineralogist, Vol. 86, No. 5-6, pp. 667-680.
Johnson, E.A. (2006) Water in nominally anhydrous crustal minerals: speciation, concentration, and geologic
significance. Reviews in Mineralogy and Geochemistry, Vol. 62, No. 1, pp. 117-154.
Kolesov, B.A., Geiger, C.A. & Armbruster, T. (2001) The dynamic properties of zircon studied by single-crystal X-ray
diffraction and Raman spectroscopy. European Journal of Mineralogy, Vol. 13, No. 5, pp. 939.
4+
Mackey, D.J., Runciman, W.A. & Vance, E.R. (1975) Crystal-field calculations for energy levels of U in ZrSiO4.
Physical Review B, Vol. 11, No. 1, pp. 211-218.
Nasdala, L., Beran, A., Libowitzky, E. & Wolf, D. (2001) The incorporation of hydroxyl groups and molecular water in
natural zircon (ZrSiO4). American Journal of Science, Vol. 301, No. 10, pp. 831.
Nasdala, L., Irmer, G. & Wolf, D. (1995) The degree of metamictization in zircon; a Raman spectroscopic study.
European Journal of Mineralogy, Vol. 7, No. 3, pp. 471.
Zhang, M., Salje, E.K.H., Farnan, I., Graeme-Barber, A., Daniel, P., Ewing, R.C., Clark, A.M. & Leroux, H. (2000)
Metamictization of zircon: Raman spectroscopic study. Journal of Physics-Condensed Matter, Vol. 12, No. 8,
pp. 1915-1926.

123
Additional Information

Ashbaugh, C.E. (1989) Gemstone irradiation and radioactivity. Gems and Gemology, Vol. 25, No. 4, pp. 196-213.
Beran, A. & Libowitzky, E. (2006) Water in natural mantle minerals II: olivine, garnet and accessory minerals.
Reviews in Mineralogy and Geochemistry, Vol. 62, No. 1, pp. 169-191.
Götze, J., Kempe, U., Habermann, D., Nasdala, L., Neuser, R.D. & Richter, D.K. (1999) High resolution
cathodoluminescence combined with SHRIMP ion probe measurements of detrital zircons. Mineralogical
Magazine, Vol. 63, No. 2, pp. 179-187.
Nasdala, L., Zhang, M., Kempe, U., Panczer, G., Gaft, M., Andrut, M. & Plotze, M. (2003) Spectroscopic methods
applied to zircon. Reviews in Mineralogy and Geochemistry, Vol. 53, No. 1, pp. 427-467.
Rossman, G.R. (2006) Analytical methods for measuring water in nominally anhydrous minerals. Reviews in
Mineralogy and Geochemistry, Vol. 62, No. 1, pp. 1-28.
Rossman, G.R. (1996) Studies of OH in nominally anhydrous minerals. Physics and Chemistry of Minerals, Vol. 23,
No. 4, pp. 299-304.
Tomašić, N., Bermanec, V., Gajović, A. & Linarić, M.R. (2008) Metamict Minerals: an Insight into a Relic Crystal
Structure Using XRD, Raman Spectroscopy, SAED and HRTEM. Croatica Chemica Acta, Vol. 81, No. 2.
Zhang, M. & Salje, E.K.H. (2001) Infrared spectroscopic analysis of zircon: radiation damage and the metamict state.
Journal of Physics Condensed Matter, Vol. 13, No. 13, pp. 3057-3072.
Zhang, M., Salje, E.K.H. & Ewing, R.C. (2002) Infrared spectra of Si-O overtones, hydrous species, and U ions in
metamict zircon: radiation damage and recrystallization. Journal of Physics Condensed Matter, Vol. 14, No. 12,
pp. 3333-3352.
Zhang, M., Salje, E.K.H., Ewing, R.C., Farnan, I., Ríos, S., Schluter, J. & Leggo, P. (2000) Alpha-decay damage and
recrystallization in zircon: evidence for an intermediate state from infrared spectroscopy. Journal of Physics
Condensed Matter, Vol. 12, No. 24, pp. 5189-5200.

124
Zoisite
Gem names: tanzanite

Ideal chemistry:
Ca2Al3(Si2O7)(SiO4)(OH)

Crystal system: orthorhombic

Point Group: H-M: mmm S: D2h

Tanzanite gem
Space Group: Pnma

Table of Atomic Coordinates (Comodi and Zanazzi, 1997):

atom x y z
Ca1 0.3668 0.25 0.4373
Ca2 0.4518 0.25 0.1150
Si1 0.0813 0.25 0.1150
Si2 0.4105 0.75 0.2824
Si3 0.1600 0.25 0.4357
Al1 0.2497 0.9970 0.1897
Al2 0.1055 0.75 0.3004
O1 0.1307 -0.0006 0.1453
O2 0.1011 0.0137 0.4309
O3 0.3587 0.9897 0.2450
O4 0.2193 0.75 0.3004
O5 0.2276 0.25 0.3119
O6 0.2718 0.75 0.0600
O7 0.9916 0.25 0.1639
O8 0.9960 0.75 0.2952
O9 0.4211 0.75 0.4431
O10 0.2682 0.25 0.0754
H 0.263 0.25 0.976

125
Raman mode analysis:

Raman Active Modes

Wyckoff Point
Atom Ag B1g B2g B3g
Position Symmetry
Al1, O1, O2, O3 8d 1 3 3 3 3
Ca1, Ca2, Si1, Si2, Si3,
Al2, O4,
4c m 2 1 2 1
O5, O6, O7, O8, O9,
O10, H

(4 x 8d) + (14 x 4c)

Raman mode analysis predicts the existence of 132 active Raman modes in zoisite:

40Ag + 26B1g + 40B2g + 26B3g = 132

126
Spectral Features Related to OH-

The crystal structure of zoisite, a member of the epidote group, is composed of chains of edge-
sharing AlO6 groups bound by SiO4 groups. Hydroxl frequently creates a weak bridging bond
oriented parallel to the c-axis between the AlO6 octahedra (Fig. 15.1). Changes in the energy of
this OH stretching mode can reveal valuable information about the crystal structure (Winkler et
al., 1989). This bridging OH produces a peak in both IR and Raman spectra centered at 3160-
3170 cm-1 (Fig. 15.2) (Winkler et al., 1989; Langer and Lattard, 1980; Winkler et al., 2008). With
an increase in pressure, this peak shifts to lower frequencies at a rate of approximately 34 cm-1
per GPa due to shortening of the hydrogen bond (Winkler et al., 2008). The frequency shift per
GPa is much larger in zoisite than in the structurally similar mineral, clinozoisite, because the
bridging hydrogen bond in zoisite is considerably straighter and shorter than the hydrogen bond
in clinozoisite. The OH bond in clinozoisite is bent at an angle of greater than 30o, thus, reducing
its compressibility with pressure (Winkler et al., 2008). Winkler et al. also observed two IR bands
at 2330 and 2900 cm-1 and suggest that these bands are associated with CO2 and C-H stretching
of organic material, respectively (Winkler et al., 1989). Neither of these peaks were observed in
the Raman spectra of zoisite in this study.

Fig. 15.1 Crystal structure of zoisite viewed down b-axis;

note location of hydrogen (red spheres) bonds bridging AlO6
groups (green octahedra); orange tetrehedra: SiO4 groups,
purple spheres: Ca cations

Fig. 15.2 Peak centered at 3166 cm-1 in the Raman

spectra of zoisite associated with bridging OH- oriented
parallel to the c-axis; sample R050038, λexcitation = 532 nm,
unoriented
Intensity

OH-

Raman Shift (cm-1)

127
Spectral Features Related to Chromophores and Other Ions

Although it has been established that the ion responsible for the distinct blue-violet color of
tanzanite is vandium, the valence and the specific nature of the color center remain controversial
(Franz and Liebscher, 2004). Fritsch and Rossman (1988) attribute the color to a combination of
octahedrally coordinated V4+ and V3+. Green color in zoisite is produced by the presence of Cr3+
and pink color is due to Mn3+ (Fritsch and Rossman, 1988). Luminescence data collected by Gaft
et al. (2005) and Koziarska et al. (1994) both attribute peaks located at 692 and 710 nm to
vanadium centers, however, the valence is still in dispute (Fig. 15.3 A. & B.). Koziarska et al.
(1994) attribute the features to V3+, while Gaft et al. (2005) attribute them to V2+.

Zoisite can incorporate a variety of trace and rare-earth elements into its structure including, but
not limited to, Cr, Sr, Eu, Tb, Dy, Nd, Y and Hf (Frei et al., 2004). Luminescence features in the
Raman spectra of a tanzanite from Umba Valley, Tanzania, located at 872 nm, 880 nm, 890 nm,
and 897 nm are likely related to the presence of Nd3+ (Fig. 15.4 A. & B.) (Gaft et al., 2005). Laser
ablation ICP-MS data for this sample is provided below (Table 15.1, Breeding, 2007). Peaks
located centered at 586, 587, 589, 591, 611, and 617 nm are likely produced by REE, however,
no luminescence or infrared data have been published to confirm this (Fig. 15.5). Gaft et al.
(2005) report the presence of peaks centered at 575, 544, 440 nm related to Dy, Tb, and Eu
luminescence centers respectively, however, no equivalent peaks in the Raman spectra were
observed in this study.

Fig. 15.3 A. Peaks centered at 692 nm (5324340cm-1) and 710 nm (5324695 cm-1) associated with
vanadium luminescence centers in zoisite; note that the luminescence peaks overwhelm the Raman peaks
in this spectrum; sample R060567, variety tanzanite, Umba Valley, Tanzania, unoriented

128
Fig. 15.3 B. Magnification of peaks centered associated with vanadium luminescence centers in zoisite;
sample R060567, variety tanzanite, Umba Valley, Tanzania, λexcitation = 532 nm, unoriented

V
Intensity

Raman Shift (cm-1)

Fig. 15.4 A. Peaks in the Raman spectrum of zoisite likely related to the presence of REE luminescence
785 -1 785 -1 785 -1
centers; peaks centered at 872 nm ( 1270cm ), 880 nm ( 1378 cm ), 890 nm ( 1504 cm ), and 897
785 -1 3+
nm ( 1602 cm ) are related to Nd ; sample R060567, tanzanite from Umba Valley, Tanzania, unoriented

Nd3+

Fig. 15.4 B. Magnification of peaks in the Raman spectrum of zoisite likely related to Nd3+ luminescence
centers; sample R060567, tanzanite from Umba Valley, Tanzania, λexcitation = 785 nm, unoriented
Intensity

Nd3+

Nd3+

Raman Shift (cm-1)

129
Table 15.1 Laser ablation ICP-MS data from zoisite, sample R060567, Umba Valley, Tanzania; numerical
values represent concentrations in parts per million (ppm) of each element taken at three different spots on
the sample; (Breeding, 2007)

24Mg 27Al 29Si 31P 43Ca 45Sc 48Ti 51V 52Cr

240.3 19.5 18.0 77.0 18.1 10.5 424.5 2209.0 137.1
220.2 20.7 18.4 87.5 18.0 12.7 430.5 2675.0 166.3
225.7 20.1 18.2 58.2 17.9 11.4 423.7 2499.0 158.0

55Mn 69Ga 72Ge 88Sr 89Y 90Zr 137Ba 139La 140Ce

17.5 163.8 16.5 1355.0 61.7 0.9 2.8 8.7 18.3
18.9 185.0 14.7 1340.0 94.5 2.0 2.9 15.5 33.3
18.4 177.9 16.5 1375.0 79.0 1.5 3.2 11.7 25.1

141Pr 146Nd 147Sm 153Eu 157Gd 159Tb 163Dy 165Ho 166Er

2.5 13.0 4.7 1.6 5.5 1.2 8.4 1.8 4.7
4.6 21.8 8.0 2.7 9.2 1.8 13.4 2.8 7.4
3.5 16.8 6.3 2.3 7.3 1.5 10.8 2.3 6.2

169Tm 172Yb 175Lu 208Pb 232Th 238U

0.5 3.6 0.4 0.6 1.6 6.8
0.8 5.3 0.7 0.2 2.4 13.1
0.7 4.7 0.6 0.2 2.0 10.9

Fig. 15.5 A. Peaks of unknown origins centered at 586 nm (5321748 cm-1), 587 nm (5321788 cm-1), 589 nm
(5321830 cm-1), 591 nm (5321887 cm-1), 611 nm (5322440 cm-1), and 617 nm (5322593 cm-1); peaks may be
related to REE luminescence centers; note that the luminescence peaks overwhelm the Raman peaks in
this spectrum; sample X060001, variety tanzanite, unoriented

130
Fig. 15.5 B. Magnification of peaks of unknown origins; sample X060001, variety tanzanite, λexcitation = 532
nm, unoriented

Intensity

Raman Shift (cm-1)

References

Breeding, M. (2007) Personal communication.

Comodi, P. & Zanazzi, P.F. (1997) The pressure behavior of clinozoisite and zoisite: An X-ray diffraction
study. American Mineralogist, Vol. 82, No. 1-2, pp. 61-68.
Franz, G. & Liebscher, A. (2004) Physical and chemical properties of the epidote minerals-an introduction.
Reviews in Mineralogy and Geochemistry, Vol. 56, No. 1, pp. 1-81.
Frei, D., Liebscher, A., Franz, G. & Dulski, P. (2004) Trace element geochemistry of epidote minerals.
Reviews in Mineralogy and Geochemistry, Vol. 56, No. 1, pp. 553-605.
Fritsch, E. & Rossman, G. (1988) An update on color in gems. Part 3: Colors caused by band gaps and
physical phenomena. Gems & Gemology, Vol. 24, No. 2, pp. 81-102.
Gaft, M., Reisfeld, R. & Panczer, G. (2005) Modern Luminescence Spectroscopy of Minerals and Materials,
Springer, Heidelberg, Germany.
Koziarska, B., Godlewski, M., Suchocki, A., Czaja, M. & Mazurak, Z. (1994) Optical properties of zoisite.
Physical Review-Section B-Condensed Matter, Vol. 50, No. 17, pp. 12297-12300.
Langer, K. & Lattard, D. (1980) Identification of a low-energy OH-valence vibration in zoisite. American
Mineralogist, Vol. 65, No. 7-8, pp. 779-783.
Winkler, B., Gale, J.D., Refson, K., Wilson, D.J. & Milman, V. (2008) The influence of pressure on the
structure and dynamics of hydrogen bonds in zoisite and clinozoisite. Physics and Chemistry of
Minerals, Vol. 35, No. 1, pp. 25-35.
Winkler, B., Langer, K. & Johannsen, P.G. (1989) The influence of pressure on the OH valence vibration of
zoisite. Physics and Chemistry of Minerals, Vol. 16, No. 7, pp. 668-671.

131
Appendix A
Features in Raman Spectra of Gem Minerals*
Sorted by increasing Raman shift

* garnets are excluded see charts in garnet section for mode locations and assignments

laser (nm)
Raman
shift (cm-1)
nm ion/molecule mineral reference
780 elbaite &
1196 860 Nd3+ Gaft et al., 2003
liddicoatite
785
1269 872 Nd3+ titanite Gaft et al., 2005
785
1270 872 Nd3+ zoisite Gaft et al., 2005

780 elbaite &

1314 869 Nd3+ Gaft et al., 2003
liddicoatite
780
1327 870 Nd3+ zircon Gaft et al., 2005
785
1374 880 Nd3+ titanite Gaft et al., 2005
785
1378 880 Nd3+ zoisite Gaft et al., 2005

780 elbaite &

1433 878 Nd3+ Gaft et al., 2003
liddicoatite
780
1450 879 Nd3+ zircon Gaft et al., 2005
785
1504 890 Nd3+ zoisite Gaft et al., 2005
785
1588 897 Nd3+ titanite Gaft et al., 2005
780
1600 891 Nd3+ zircon Gaft et al., 2005
785
1602 897 Nd3+ zoisite Gaft et al., 2005

780 elbaite &

1668 896 Nd3+ Gaft et al., 2003
liddicoatite
-
OH
1165 N/A topaz Beny & Piriou, 1987
(stretching)
13
1280-1289 N/A C diamond McNamara et al., 1992
532
1394 574 Sm3+ titanite Gaft et al., 2005
532
1526 579 Sm3+ titanite Gaft et al., 2005

1598 N/A Type-I H2O beryl Lodzinski et al., 2005

1628/1634 N/A Type-II H2O beryl Lodzinski et al., 2005

532
1720 585 Sm3+ titanite Gaft et al., 2005
nd
2 order
1817 N/A diamond Solin & Ramdas, 1970
Raman
2nd order
1864 N/A diamond Solin & Ramdas, 1970
Raman

132
laser (nm)
Raman
shift (cm-1)
nm ion/molecule mineral reference
532
1910 591 Eu3+ zircon Gaft et al., 2005
532
2013 596 Eu3+ zircon Gaft et al., 2005
nd
2 order
2025 N/A diamond Solin & Ramdas, 1970
Raman
532
2151 600 Sm3+ titanite Gaft et al., 2005
nd
2 order
2177 N/A diamond Solin & Ramdas, 1970
Raman
2nd order
2254 N/A diamond Solin & Ramdas, 1970
Raman
532
2260 604 Eu3+ zircon Gaft et al., 2005
nd
2 order
2333 N/A diamond Solin & Ramdas, 1970
Raman
2nd order
2458 N/A diamond Solin & Ramdas, 1970
Raman
532
2491 614 Eu3+ titanite Gaft et al., 2005
2nd order
2519 N/A diamond Solin & Ramdas, 1970
Raman
532
2560 614 Sm3+ zircon Gaft et al., 2005
532
2597 617 Sm3+ titanite Gaft et al., 2005
nd
2 order
2667 N/A diamond Solin & Ramdas, 1970
Raman
532 Al3+-hole or Gaft et al., 2005;
2670 620 spodumene
Mn2+? Souza et al., 2003
532
2682 620 Sm3+ titanite Gaft et al., 2005
532
3100 638 NV- center diamond Zaitsev, 2001

3160-3170 N/A bridging OH- zoisite Winkler et al., 1989

rd
3 order
3300 N/A diamond Bormett et al., 1995
Raman
532
3370 650 Cr3+ chrysoberyl Gaft et al., 2005

3424 N/A OH- zircon Dawson et al., 1971

3476 N/A OH3 elbaite Gonzales-Carreño, 1988

Castañeda et al., 2000; Zhang
3584 N/A OH1 or OH3? elbaite et al., 2008
OH-
3485 N/A titanite Hammer et al., 1996
(stretching)
3489 N/A OH3 liddicoatite Zhang et al., 2008

3496 N/A OH3 elbaite Gonzles-Carreño, 1988

532
3540 655 Cr3+ chrysoberyl Gaft et al., 2005

133
laser (nm)
Raman
shift (cm-1)
nm ion/molecule mineral reference
Zhang et al., 2008; Gonzales-
3566 N/A OH3 elbaite Carreño, 1988
3592 N/A OH1 elbaite Castañeda et al., 2000
Type-II H2O
3594/3597 N/A beryl Lodzinski et al., 2005
(stretching)
3595 N/A OH3 liddicoatite Zhang et al., 2008

3598 N/A OH3 elbaite Zhang et al., 2008

Type-I H2O
3606-3609 N/A beryl Lodzinski et al., 2005
(stretching)
OH- substituting
3650 N/A topaz Pinheiro et al., 2002
for F
3651/3657 N/A Type-II H2O beryl Lodzinski et al., 2005
532
3654 660 Sm3+ zircon Gaft et al., 2005

3673 N/A OH- andalusite Rossman, 1996

3692/3696 N/A Type-I H2O beryl Lodzinski et al., 2005

3740 N/A OH1 dravite Gonzales-Carreño, 1988

532
3745 664 Cr3+ chrysoberyl Gaft et al., 2005
532
3772 666 Sm3+ zircon Gaft et al., 2005
rd
3 order
3825 N/A diamond Bormett et al., 1995
Raman
3880 N/A Type-I H2O beryl Lodzinski et al., 2005
532
3900 671 Cr3+ diopside Gaft et al., 2005
532
4000 676 Cr3+ spinel Gaft et al., 2005
532
4055 678 Cr3+ topaz Gaft et al., 2005
532
4062 679 Cr3+ chrysoberyl Gaft et al., 2005
532
4075 679 Cr3+ topaz Gaft et al., 2005
532
4082 679 Cr3+ diopside Gaft et al., 2005
532
4095 680 Cr3+ beryl Gaft et al., 2005
532
4100 680 Cr3+ chrysoberyl Gaft et al., 2005
532
4110 680 Cr3+ uvite Gaft et al., 2005
532
4160 683 Cr3+ uvite Gaft et al., 2005
532
4160 683 Cr3+ topaz Gaft et al., 2005

134
laser (nm)
Raman
shift (cm-1)
nm ion/molecule mineral reference
532
4170 683 Cr3+ beryl Gaft et al., 2005
532
4175 683 Cr3+ topaz Gaft et al., 2005
532
4185 684 Cr3+ dravite Gaft et al., 2005
532
4196 684 Cr3+ diopside Gaft et al., 2005
532
4211 686 Cr3+ spinel Gaft et al., 2005
532
4263 689 Eu3+ titanite Gaft et al., 2005
532
4340 692 V zoisite Gaft et al., 2005
532
4366 693 Cr3+ corundum Gaft et al., 2005
532
4367 693 Cr3+ chrysoberyl Gaft et al., 2005
532
4373 693 Cr3+ forsterite Gaft et al., 2005
532
4396 694 Cr3+ corundum Gaft et al., 2005
532
4400 694 Cr3+ forsterite Gaft et al., 2005
532
4466 698 Cr3+ spinel Gaft et al., 2005
532
4510 700 Cr3+ chrysoberyl Gaft et al., 2005
532
4526 700 Cr3+ diopside Gaft et al., 2005
532
4545 701 Eu3+ titanite Gaft et al., 2005
532
4572 702 Eu3+ zircon Gaft et al., 2005
532
4573 703 V2+ chrysoberyl Gaft et al., 2005
532
4651 706 Eu3+ titanite Gaft et al., 2005
532
4657 708 Cr3+ spinel Gaft et al., 2005
532
4695 709 V zoisite Gaft et al., 2005
532
4728 710 Eu3+ zircon Gaft et al., 2005
532
4762 712 Cr3+ topaz Gaft et al., 2005
532
4859 718 Cr3+ spinel Gaft et al., 2005
532
4958 722 V3+ diopside Gaft et al., 2005
532
5155 733 Cr3+ topaz Gaft et al., 2005

135
Appendix B
Features in Raman Spectra of Gem Minerals*
Sorted alphabetically by mineral name

*garnets are excluded see charts in garnet section for mode locations and assignments

laser (nm)
Raman
mineral shift (cm-1)
nm ion/molecule reference

beryl 1598 N/A Type-I H2O Lodzinski et al., 2005

beryl 1628/1634 N/A Type-II H2O Lodzinski et al., 2005

Type-II H2O
beryl 3594/3597 N/A Lodzinski et al., 2005
(stretching)
Type-I H2O
beryl 3606-3609 N/A Lodzinski et al., 2005
(stretching)

beryl 3651/3657 N/A Type-II H2O Lodzinski et al., 2005

beryl 3692/3696 N/A Type-I H2O Lodzinski et al., 2005

beryl 3880 N/A Type-I H2O Lodzinski et al., 2005

532
beryl 4095 680 Cr3+ Gaft et al., 2005

532
beryl 4170 683 Cr3+ Gaft et al., 2005

532
chrysoberyl 3370 650 Cr3+ Gaft et al., 2005

532
chrysoberyl 3540 655 Cr3+ Gaft et al., 2005

532
chrysoberyl 3745 664 Cr3+ Gaft et al., 2005

532
chrysoberyl 4062 679 Cr3+ Gaft et al., 2005

532
chrysoberyl 4100 680 Cr3+ Gaft et al., 2005

532
chrysoberyl 4367 693 Cr3+ Gaft et al., 2005

532
chrysoberyl 4510 700 Cr3+ Gaft et al., 2005

532
chrysoberyl 4573 703 V2+ Gaft et al., 2005

532
corundum 4366 693 Cr3+ Gaft et al., 2005

532
corundum 4396 694 Cr3+ Gaft et al., 2005

136
laser (nm)
Raman
mineral shift (cm-1)
nm ion/molecule reference
13 McNamara et al.,
diamond 1280-1289 N/A C 1992

2nd order Solin & Ramdas,

diamond 1817 N/A
Raman 1970

2nd order Solin & Ramdas,

diamond 1864 N/A
Raman 1970

2nd order Solin & Ramdas,

diamond 2025 N/A
Raman 1970

2nd order Solin & Ramdas,

diamond 2177 N/A
Raman 1970

2nd order Solin & Ramdas,

diamond 2254 N/A 1970
Raman

2nd order Solin & Ramdas,

diamond 2333 N/A
Raman 1970

2nd order Solin & Ramdas,

diamond 2458 N/A
Raman 1970

2nd order Solin & Ramdas,

diamond 2519 N/A
Raman 1970

2nd order Solin & Ramdas,

diamond 2667 N/A
Raman 1970
532
diamond 3100 638 NV- center Zaitsev, 2001
rd
3 order
diamond 3300 N/A Bormett et al., 1995
Raman
3rd order
diamond 3825 N/A Bormett et al., 1995
Raman
532
diopside 3900 671 Cr3+ Gaft et al., 2005
532 3+
diopside 4082 679 Cr Gaft et al., 2005
532 3+
diopside 4196 684 Cr Gaft et al., 2005
532 3+
diopside 4526 700 Cr Gaft et al., 2005
532 3+
diopside 4958 722 V Gaft et al., 2005
532
forsterite 4373 693 Cr3+ Gaft et al., 2005

532
forsterite 4400 694 Cr3+ Gaft et al., 2005

532
spinel 4000 676 Cr3+ Gaft et al., 2005

532
spinel 4211 686 Cr3+ Gaft et al., 2005

137
laser (nm)
Raman
mineral shift (cm-1)
nm ion/molecule reference
532
spinel 4466 698 Cr3+ Gaft et al., 2005

532
spinel 4657 708 Cr3+ Gaft et al., 2005

532
spinel 4859 718 Cr3+ Gaft et al., 2005

532 Al3+-hole or Gaft et al., 2005;

spodumene 2670 620
Mn2+? Souza et al., 2003

785
titanite 1269 872 Nd3+ Gaft et al., 2005

785
titanite 1374 880 Nd3+ Gaft et al., 2005

785
titanite 1588 897 Nd3+ Gaft et al., 2005

532
titanite 1394 574 Sm3+ Gaft et al., 2005

532
titanite 1526 578 Sm3+ Gaft et al., 2005

532
titanite 1720 589 Sm3+ Gaft et al., 2005

532
titanite 2151 600 Sm3+ Gaft et al., 2005

532
titanite 2491 613 Sm3+ Gaft et al., 2005

532
titanite 2597 617 Sm3+ Gaft et al., 2005

532
titanite 2682 620 Sm3+ Gaft et al., 2005

titanite 3485 N/A OH- (stretching) Hammer et al., 1996

532
titanite 4263 689 Eu3+ Gaft et al., 2005

532
titanite 4545 703 Eu3+ Gaft et al., 2005

532
titanite 4651 706 Eu3+ Gaft et al., 2005

OH-
topaz 1165 N/A Beny & Piriou, 1987
(stretching)
532
topaz 5155 733 Cr3+ Gaft et al., 2005

OH- substituting
topaz 3650 N/A Pinheiro et al., 2002
for F

138
laser (nm)
Raman
mineral shift (cm-1)
nm ion/molecule reference

532
topaz 4055 678 Cr3+ Gaft et al., 2005

532
topaz 4075 679 Cr3+ Gaft et al., 2005

532
topaz 4160 683 Cr3+ Gaft et al., 2005

532
topaz 4175 683 Cr3+ Gaft et al., 2005

532
topaz 4762 712 Cr3+ Gaft et al., 2005

780
tourmaline 1196 860 Nd3+ Gaft et al., 2003

780
tourmaline 1314 869 Nd3+ Gaft et al., 2003

780
tourmaline 1433 878 Nd3+ Gaft et al., 2003

780
tourmaline 1668 896 Nd3+ Gaft et al., 2003

Gonzales-Carreño,
tourmaline 3476 N/A OH3 1988

tourmaline 3489 N/A OH3 Zhang et al., 2008

Gonzales-Carreño,
tourmaline 3496 N/A OH3 1988
Zhang et al., 2008;
tourmaline 3566 N/A OH3 Gonzales-Carreño,
1988
Castañeda et al.,
tourmaline 3592 N/A OH1 2000

tourmaline 3595 N/A OH3 Zhang et al., 2008

tourmaline 3598 N/A OH3 Zhang et al., 2008

532
tourmaline 4110 680 Cr3+ Gaft et al., 2005

532
tourmaline 4160 683 Cr3+ Gaft et al., 2005

532
tourmaline 4185 684 Cr3+ Gaft et al., 2005

780
zircon 1327 870 Nd3+ Gaft et al., 2005

780
zircon 1450 879 Nd3+ Gaft et al., 2005

780
zircon 1600 891 Nd3+ Gaft et al., 2005

532
zircon 1910 591 Eu3+ Gaft et al., 2005

139
laser (nm)
Raman
mineral shift (cm-1)
nm ion/molecule reference
532
zircon 2013 596 Eu3+ Gaft et al., 2005

532
zircon 2260 604 Eu3+ Gaft et al., 2005

532
zircon 2560 614 Eu3+ Gaft et al., 2005

zircon 3424 N/A OH- Dawson et al., 1971

532
zircon 3654 660 Sm3+ Gaft et al., 2005

532
zircon 3772 666 Sm3+ Gaft et al., 2005

532
zircon 4572 702 Eu3+ Gaft et al., 2005

532
zircon 4728 710 Eu3+ Gaft et al., 2005

785
zoisite 1270 872 Nd3+ Gaft et al., 2005

785
zoisite 1378 880 Nd3+ Gaft et al., 2005

785
zoisite 1504 890 Nd3+ Gaft et al., 2005

785
zoisite 1602 897 Nd3+ Gaft et al., 2005

zoisite 3160-3170 N/A bridging OH- Winkler et al., 1989

532
zoisite 4340 692 V Gaft et al., 2005

532
zoisite 4695 709 V Gaft et al., 2005

140
Appendix C

Unit Conversions

Raman spectroscopy utilizes Raman shift (cm-1) to describe the location of Raman modes. The
abbreviation cm-1 can be misleading because Raman shift represents a change in wave numbers, a shift
from the frequency of the laser. The abbreviation Δcm-1 would be more appropriate, however it is
convention to describe Raman shift with cm-1.

When comparing Raman and IR modes to luminescence features, a conversion from cm-1 to nm is
required. The simple conversion is as follows:

Note: these conversions are laser wavelength dependent.

From nm to cm-1:

1) Convert the wavelength of the laser used for Raman spectroscopy from nm to wave numbers.
Do this by converting nm into cm and then to wave numbers.

1 nm = 1.0 x 10-7 cm

ex. 532 nm laser = 532 x 1.0 x 10-7 = .0000532 cm

From cm to cm-1: 1/.0000532 cm = 18796.99248 cm-1

The 532 nm laser is equal to 18796.99248 cm-1

2) Convert the luminescence feature (nm) to wave numbers by same process described in step 1).

ex. Luminescence feature at 693 nm = 14430.01443 cm-1

3) Since Raman shift represents the shift in energy from the initial energy of the laser beam, the final
step requires that you subtract the converted luminescence feature in wave numbers from the
converted laser wavelength in wave numbers. This will give you the location in the Raman
spectrum where you would expect to see the luminescence feature you have observed.

ex. (laser) 18796.99248 cm-1 - (feature) 14430.01443 cm-1 = 4366.978051 cm-1 (Raman shift)

To switch between Raman spectra from different laser wavelengths, simply replace the laser wavelength
with the wavelength being utilized; ie. 514 nm, 780 nm, 785 nm, etc.

From cm-1 to nm (reverse the process):

1) Convert your value in Raman shift to wave numbers by subtracting the Raman shift value from
the wavelength of the laser.

ex. Raman shift value: 1378 cm-1 Laser: 532 nm = 18796.99248 cm-1
-1 -1 -1
18796.99248 cm – 1378 cm = 17418.99 cm (wave numbers)

2) Convert wave numbers into cm and then nm.

-1 -5
1/17418.99 cm = 5.74 x 10 cm

5.74 x 10-5 cm x 107 = 574 nm

141
Wave lengths of radiation are frequently described in angstroms (Å). The simple conversion from
angstroms to nm is provided here:

Angstroms to nm:

1 Å = .1 nm To convert from Å to nm simply take the wavelength in angstroms and multiply by 0.1

Example: 5448Å = 544.8 nm

Wave lengths of radiation can also be described in electron volts (eV). The simple conversion from eV to
nm is provided here:

eV to nm:

1239.8424121/ x eV = nm

Example: 1239.8424121/2.16 eV = 574 nm

142
Appendix D

Cause of Color Chart

(modified from Fritsch and Rossman, 1988)

Mineral Color Cause Reference

octahedrally Fritsch and Rossman,
beryl emerald green
cooridnated Cr3+ 1988; Gaft et al., 2005
octahedrally Fritsch and Rossman,
green beryl
cooridnated V3+ 1988; Gaft et al., 2005
O2- Fe3+ charge
green with yellow Fritsch and Rossman,
transfer, Fe2+ filling
or blue undertones 1988; Gaft et al., 2005
channels
O2- Fe3+ charge Fritsch and Rossman,
heliodor (golden)
transfer 1988; Gaft et al., 2005
color centers due to
dark blue (Maxixe) Fritsch and Rossman, 1988
irradiation
light blue Fritsch and Rossman,
channel-filling Fe2+
(aquamarine) 1988; Gaft et al., 2005
Fe2+O Fe3+
Fritsch and Rossman,
dark aquamarine intervalence charge
1988; Gaft et al., 2005
transfer
octahedrally Fritsch and Rossman,
pink (morganite)
coordinated Mn2+ 1988; Gaft et al., 2005
octahedrally
red Fritsch and Rossman, 1988
coordinated Mn3+
Fritsch and Rossman,
octahedrally 1988; Gubelin, 1976;
chrysoberyl alexandrite
coordinated Cr3+ Hassan and El-Rakhawy,
1974
octahedrally
yellow Fritsch and Rossman, 1988
coordinated Fe3+
octahedrally
coordinated Cr3+ with
corundum ruby red some V3+ and Fe3+ in Fritsch and Rossman, 1988
octahedral
coordination
Fe2+ O Ti4+charge
blue sapphire Fritsch and Rossman, 1988
transfer
Fe2+ O Ti4+ charge
transfer; with
purple Fritsch and Rossman, 1988
octahedrally
coordinated Cr3+
still debated;
presence of Cr
orange-pink
possible with other Fritsch and Rossman, 1988
(padparadscha)
color centers, valence
is debated
octahedrally
orange to brown coordinated Cr3+with Fritsch and Rossman, 1988
some Fe3+
octahedrally
pink Fritsch and Rossman, 1988
coordinated Cr3+

143
Mineral Color Cause Reference
multiple possibilities
relating to Fe
corundum yellow Fritsch and Rossman, 1988
(possible charge
transfer or Fe-pairs)
several possibilities
green involving either Fe Fritsch and Rossman, 1988
and/or Cr3+
yellow, orange, Type I diamonds;
diamond brown, colorless, contain nitrogen Deljanin and Simic, 2007
near colorless aggregates
brown, pink,
Type IIa diamonds;
colorless, near Deljanin and Simic, 2007
nitrogen absent
colorless
blue, gray, light Type IIb diamonds;
brown, near nitrogen absent, Deljanin and Simic, 2007
colorless contain boron)
Fritsch et al., 2007; Shigley,
pink/red deformation of lattice
1993
purple deformation of lattice Titkov et al., 2008
nitrogen impurities
brown and lattice Massi, 2005
deformation
dark mineral
black Titkov et al., 2003
microinclusions
chameleon (color
electron trap due to H
change green to Fritsch et al., 2007
and N interactions
yellow)
octahedrally
green (chrome Fritsch and Rossman,
diopside coordinated Cr3+, V3+,
diopside) 1988; Andrut et al., 2003).
or a combination
charge transfer Fe2+ Fritsch and Rossman,
yellowish green
Fe3+ 1988; Andrut et al., 2003).
charge transfer Fe2+
purple (rare) Herd et al., 2000
and Ti4+
garnet
Fe2+ in distorted cubic Fritsch and Rossman,
almandine red (rhodolite)
coordination 1988; Manning, 1967
octahedrally
andradite green (demantoid) Fritsch and Rossman, 1988
coordinated Cr3+
octahedrally
yellow green Fritsch and Rossman, 1988
coordinated Fe3+
octahedrally Fritsch and Rossman,
grossular green (tsavorite)
coordinated V3+ 1988; Manning, 1970
Fritsch and Rossman,
orange (hessonite) Mn2+ or Fe2+
1988; Manning, 1970
octahedrally
Fritsch and Rossman,
pyrope red coordinated Cr3+and
1988; Manning, 1967
some Fe2+
Fritsch and Rossman,
brownish-red Fe2+
1988; Manning, 1967

144
Mineral Color Cause Reference
Gubelin, 1982; Manning,
spesartine orange Mn2+
1967
octahedrally Fritsch and Rossman,
uvarovite green
coordinated Cr3+ 1988; Manning, 1970
green to blue, red octahedrally Gubelin, 1982; Schmetzer
color-
to purple, green to coordinated Cr3+and and Bernhardt, 1999;
change
red V3+ Carstens, 1973
octahedrally
green, yellow-
olivine coordinated Fe2+with Fritsch and Rossman, 1988
green (peridot)
minor Cr3+
hole center created
Paradise, 1982; Fritsch and
quartz purple (amethyst) by presence of Fe
Rossman, 1988
and radiation
O2- Fe3+ charge
yellow (citrine) Fritsch and Rossman, 1988
transfer
Paradise, 1982; Fritsch and
green (praseolite) Fe2+
Rossman, 1988).
Maschmeyer et al., 1980;
black (smoky) hole center
Fritsch and Rossman, 1988
pink (rosy) pink microinclusions Goreva et al., 2001
octahedrally
spinel pink/red Fritsch and Rossman, 1988
coordinated Cr3+
octahedrally
coordinated Cr3+with
purple Fritsch and Rossman, 1988
tetrahedrally
coordinated Fe2+
Co and Fe in
cobalt blue tetrahedral Fritsch and Rossman, 1988
coordination
tetrahedrally
blue green coordinated Fe2+and Fritsch and Rossman, 1988
Fe3+
pink to purple
spodumene Mn3+ Fritsch and Rossman, 1988
(kunzite)
octahedrally
green (hiddenite) coordinated Cr3+ and Fritsch and Rossman, 1988
V3+, with some Mn
possible Fe charge
pale green Fritsch and Rossman, 1988
transfer and some Mn
titanite green Fe Fritsch and Rossman, 1988
chrome sphene octahedrally
Fritsch and Rossman, 1988
(green) coordinated Cr3+
pink Mn Fritsch and Rossman, 1988

yellow Fe Fritsch and Rossman, 1988

brown Fe Fritsch and Rossman, 1988

octahedrally Taran et al., 2003; Fritsch
pink
coordinated Cr3+ and Rossman, 1988)

145
Mineral Color Cause Reference
3+
Cr and a radiation
topaz yellow Gaft et al., 2005
induced O- hole
Cr3+combined with a
orange-red Gaft et al., 2005
F-center
O- and F center
red-brown Gaft et al., 2005
combinations
blue R-centers Gaft et al., 2005

purple Cr3+ Taran et al., 2003

red-orange
Cr4+ Taran et al., 2003
(imperial)
pink-orange Cr3+ and Cr4+ Taran et al., 2003

tourmaline
octahedrally
dravite green coordinated V3+ and Fritsch and Rossman, 1988
Cr3+
Fe and Ti charge
yellow to brown Fritsch and Rossman, 1988
transfer
red Fe Fritsch and Rossman, 1988
elbaite and Fe and charge
blue (indicolite) Fritsch and Rossman, 1988
liddicoatite transfer
Fe and Ti charge
green Fritsch and Rossman, 1988
transfer
Fe and Ti charge
brown Fritsch and Rossman, 1988
transfer
Mn and Ti charge
yellow-green Fritsch and Rossman, 1988
transfer
pink to red Mn and possible
Fritsch and Rossman, 1988
(rubellite) irradiation center
neon blue/green
Cu2+ Rossman et al., 1991
(Paraiba)
octahedrally
uvite green coordinated Cr3+ and Fritsch and Rossman, 1988
V3+
4+
Nb and irradiation Fielding, 1970; Fritsch and
zircon red
center Rossman, 1988
Mackey et al., 1975; Fritsch
blue presence of U4+
and Rossman, 1988
blue-violet
zoisite V4+ and V3+ Fritsch and Rossman, 1988
(tanzanite)
green Cr3+ Fritsch and Rossman, 1988

pink Mn3+ Fritsch and Rossman, 1988

146
Appendix E

Point Group Symmetry Notations

There are two notations used to describe point symmetry in crystals and molecules: 1) Hermann-Manguin
notation (preferred by crystallographers) and 2) Schönflies notation (preferred by spectroscopists and
chemists). A table correlating crystallographic point groups with the corresponding Schönflies notation is
provided below (Boisen and Gibbs, 1990).
HM S HM S HM S
1 C1 1 Ci m Cs
2 C2 2/m C2h 4 S4
3 C3 3 C3i 6 C3h
4 C4 4/m C4h mm2 C2v
6 C6 6/m C6h 3m C3v
222 D2 mmm D2h 4 2m D2d
32 D3 3m D3d 4mm C4v
422 D4 4/mmm D4h 6 2m D3h
622 D6 6/mmm D6h 6mm C6v
23 T m3 Th 4 3m Td
432 O m3m 0h
235 I m3 5 Ih

A brief description of Schönflies notation is provided below (Gaft et al., 2005; Ferraro et al., 2003).

Schönflies notation:

C: (cyclic) one rotational axis; C1: identity, Cn: has an n-fold axis

D: (dihedral) orthogonal axes (perpendicular); Dn: n-fold axis of symmetry perpendicular to n 2-fold axes

Cubic: T (tetrahedral) 4 axes, O (octahedral) 8 axes, I (icosahedra, describes molecules, not crystals) 20
axes

i: inversion center

E/I: identity

Planar Symmetry:

-one subscript letter follows the rotation symmetry; if letter is h (horizontal): mirror plane parallel to the
rotation axis; v (vertical): mirror plane perpendicular to rotation axis; d (diagonal): diagonal mirror

ex. C3h: has one 3-fold rotational axis with a mirror parallel to the 3-fold axis

-Sn: S is mirror symmetry and invariance (remains unchanged) by a n-fold rotation followed by reflection
in plane perpendicular to axis

147