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:

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

Absorption (Emission) (Excitation) Fluorescence 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

Intensity

Raman Shift (cm-1)

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. Abstracts from the 19th General Meeting of the International Mineralogical Association, Kobe, Japan, 23-28 July 2006.

Paschotta, R. (2009) Encyclopedia of Laser Physics and Technology. Online .

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 -1 Si-O Be-O

ν (cm Cr-O As-O 1200 V-O Mo-O W-O

1000 Raman Frequency,

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

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

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 the scope of this study (Fritsch and Rossman, 1987). An example is the presence of Mn in beryl; Mn2+ produces a light pink hue (morganite), while Mn3+ 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 (heliodor) is the result of the interaction of Fe3+ with its nearest neighbor oxygen atoms. Intervalence charge transfer between Fe2+ and Ti4+ 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.

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 -1 centered at 1386 and 1240 cm in the Raman spectra of beryl are attributed to CO2.

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

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

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

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

Type-

I H O 2 Type-II

H O 2 Type-I H2O

-1 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 ; 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 (5324095 cm-1) and 685 nm (5324170 cm-1); 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.

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

-1 Raman Shift (cm )

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. Spinolo, G., Fontana, I. & Galli, A. (2007) Optical absorption spectra of Fe2+ and Fe3+ 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.

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 (1.934Å). However, it has been reported that with an increase in pressure and temperature, Cr3+ will occupy the Al1 sites (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

3+ Cr

3+ Cr & V? Cr3+

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 centered at 703 nm (5324573 cm-1) possibly related to V in the Raman spectrum of chrysoberyl; sample R080110, Sri Lanka, unoriented

Cr3+

Intensity

Cr3+ 3+ Cr V?

Cr3+

Raman Shift (cm-1)

Fig. 2.1 C. (left) Further magnification of weak Cr3+ peaks centered at 650 nm (5323370 cm-1), 655 nm (5323540 cm-1), and 664 nm (5323745 cm-1); sample R080110, unoriented

Cr3+

Intensity

Cr3+

3+ -1 Cr V? Raman Shift (cm ) Cr3+

3+ Fig. 2.1 D. (right) Further magnification of weak Cr Intensity 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 (5324573 cm-1); 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.

Corundum

Gem names: ruby, sapphire, padparadascha

Ideal chemistry: Al2O3

Crystal system: trigonal − Point Group: H-M: 3 m S: D 3d − Space Group: R3 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 A E 1g g

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

-1 Raman Shift (cm )

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 3+ positions of Cr 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

Cr3+

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

3+ Cr

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.

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 43m 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: 16 18 v1/v1* = √ m O/ m O 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 -1 values into the equation above you get: v1/1332 cm = 0.96 and so the theoretical peak location in 13 -1 diamond composed of 100% C is v1 = 1278 cm . 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).

Fig. 4.1 (right) Slight shift in the frequency of 5. the primary Raman peak in diamond; 1) natural type IIa diamond, sample R050204, 2) natural 4. type IIa diamond, sample R050206, 3) synthetic yellow diamond, sample X080011; 4) natural type I diamond, sample R050207; 5) natural type IIa diamond, sample R050205; λ = 532 nm, background corrected, Intensity excitation 3. unoriented 2.

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

1. 1336 cm-1

Raman Shift (cm-1)

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

3rd Order Raman 2nd 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

2465 cm-1 2667 cm-1

Intensity

2030 cm-1

-1 Raman Shift (cm )

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

3300 cm-1 3825 cm-1

Intensity

-1 Raman Shift (cm )

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

NV- Intensity 2.

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

-1 Raman Shift (cm )

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

Intensity

4. 3.

1. Raman Shift (cm-1)

41 References

Behr, D., Wagner, J., Wild, C. & Koidl, P. (1993) Homoepitaxial 13C 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 fractionation of calcite from vibrational spectroscopy of 18O-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.

Diopside

Gem names: chrome diopside

Ideal chemistry: CaMgSi2O6

Crystal system: monoclinic

Point Group: H-M: 2/m S: C 2h

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

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

139 323

Ag Ag 390

Ag Intensity 367 Bg See Fig. 5.1 B See Fig. 5.1 C & 357 856 508 530 Ag 710 Ag 917 1048 A A ? g g Bg Bg Bg

-1 Raman Shift (cm )

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 A g

Intensity 465 560

Bg Bg

Raman Shift (cm-1)

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 A B g g

194 Intensity Bg 163 229

Bg Bg

-1 Raman Shift (cm )

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

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?

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

Cr?

-1 Raman Shift (cm )

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 3+ 3+ CaFe (Fe 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.

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

2+ almandine Fe 3Al2(SiO4)3

grossular Ca3Al2(SiO4)3

2+ spessartine Mn 3Al2(SiO4)3

3+ andradite Ca3Fe 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

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 peak 6 (centered at ~350 cm-1), 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).

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

2+ 2+ Almandine (Fe 3Al2(SiO4)3): The red color-inducing cation in almandine is Fe 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).

2+ 2+ Spessartine (Mn 3Al2(SiO4)3): The orange color of spessartine is produced by the presence of Mn 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).

3+ 3+ Andradite (Ca3Fe 2(SiO4)3): Yellow-green andradite contains Fe 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 3+ Table 6.1 Peaks related to Cr 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, 6.1 & Gaft et al., 2005 (topaz & R040159 Pyrope 683 5324166 Cr3+ Bohemia 6.3 B beryl) Andrut and Wildner, 6.1 & 686 5324225 Cr3+ 2001; Gaft et al., 2005 6.3 B (spinel) 6.1 & 689 5324290 Cr3+? 6.3 B 6.1 & Gaft et al., 2005 693 5324360 Cr3+ 6.3 B (chrysoberyl & corundum) 6.1 & Struve and Huber, 1985; 696 5324431 Cr3+ or V? 6.3 B Gaft et al., 2005 Gaft et al., 2005 R060448 red, India Pyrope 693 5324366 Cr3+ 6.2 (chrysoberyl & corundum) 694 5324393 Cr3+ 6.2 Gaft et al., 2005 colorless, Piedmont, Gaft et al., 2005 (topaz & R070637 Pyrope 683 5324172 Cr3+ 6.3 Italy beryl) Andrut and Wildner, 686 5324236 Cr3+ 6.3 C 2001; Gaft et al., 2005 (spinel) 690 5324316 Cr3+? 6.3 C Gaft et al., 2005 693 5324380 Cr3+ 6.3 C (chrysoberyl & corundum) Struve and Huber, 1985; 695 5324427 Cr3+ or V? 6.3 C Gaft et al., 2005 699 5324492 Cr3+? 6.3 C 701 5324550 Cr3+ 6.3 C Andrut and Wildner, 2001 729 5325095 V? 6.3 C R060099 red, Alaska Almandine 688 5324277 Cr3+? 6.4 Gaft et al., 2005 693 5324372 Cr3+ 6.4 (chrysoberyl & corundum) 694 5324401 Cr3+ 6.4 Gaft et al., 2005 706 5324642 V? 6.4 red-brown, Shigar Gaft et al., 2005 R060279 Spessartine 693 5324370 Cr3+ 6.5 Valley, Pakistan (chrysoberyl & corundum) 710 5324705 V? 6.5 Gaft et al., 2005 R080053 orange, East Africa Spessartine 693 5324370 Cr3+ 6.6 (chrysoberyl & corundum) 694 5324395 Cr3+ 6.6 Gaft et al., 2005 Andrut and Wildner, transparent tan, R060382 Grossular 697 5324450 Cr3+ 6.7 2001; Struve and Huber, Tanzania 1985 701 5324544 Cr3+ 6.7 Andrut and Wildner, 2001 717 5324840 V 6.7 Gaft et al., 2005 721 5324930 V? 6.7 6.7 A. 605 5322262 Mn Gaft et al., 2005 & C. green, Ural Mts., Struve and Huber, 1985; R060477 Uvarovite 696 5324440 Cr3+or V? 6.8 Russia Gaft et al., 2005 701 5324542 Cr3+ 6.8 Andrut and Wildner, 2001

721 5324930 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

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 -1 532 -1 683 nm (4166 cm ), 686 nm ( 4225 cm ), 689 nm V? (5324290 cm-1), and 693 nm (5324360 cm-1), related to Cr3+ centers and a peak located at 696 nm (5324431 cm-1) possibly related to a V center; sample R040159, red pyrope, Meronitz, Bohemia; unoriented

Intensity

Cr3+

Raman Shift (cm-1)

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), 695 nm (5324427 cm-1), 693 nm (5324380 cm-1), 699 (5324492 cm-1), and 701 nm (5324550 cm-1), are likely related to Cr, while a single peak centered at 729 nm (5325095 cm-1) may be related to V sample; note that the luminescence peaks overwhelm the Raman peaks in this spectrum; sample R080060, synthetic pyrope, unoriented

3+ Cr

V 3+ Cr

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

3+ Cr

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

Cr3+

4. Cr3+ 3.

Intensity

3+ Cr 2. V Cr3+ 1.

-1 Raman Shift (cm ) 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 (5324372 cm-1), and 694 nm (5324401 cm-1) 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+

Fig. 6.5 B. Magnification of a peak associated with Cr3+ located at 693 nm (5324370 cm-1) and a peak located at 710 nm (5324705 cm-1) possibly related to V in spessartine; sample R060279, Pakistan, unoriented

Cr3+

V? Intensity

-1 Raman Shift (cm )

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

3+ Cr

57 Fig. 6.6 B. Magnification of peaks located at 693 nm (5324370 cm-1) and 694 nm (5324395 cm-1) associated with Cr3+ centers in spessartine; sample R080053, East Africa, unoriented

3+ Cr

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

3+ CrCr3+

Fig. 6.7 B. (right) Magnification of peaks located at 697 nm (5324450 cm-1) and 701 nm (5324544 cm-1) related to Cr3+ centers and peaks located at 717 nm (5324840 cm-1), 721 nm (5324930 cm-1) possibly related to V centers; sample V? R060382, grossular from Tanzania, unoriented

Intensity

3+ Cr

-1 Raman Shift (cm )

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, unoriented

Intensity

-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

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

3+ Cr

Intensity

Raman Shift (cm-1)

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 peaks are likely related to Nd3+. According to luminescence studies of titanite conducted by Gaft et al. (2003), nearly all Nd3+ luminescence emission bands appear in the IR region. Nd3+ has been assigned to 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

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

3+ Nd ? ?

Nd3+ 3+ Nd Nd3+

3+ Nd

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 cm-1), 911 nm (7851763 cm-1), and 950 nm (7852216 cm-1) likely related to luminescence centers of unidentified REE; sample R060382, Lalatema, Tanzania, unoriented

3+ Nd ?

Nd3+ ? ? Intensity

3+ Nd Nd3+

-1 Raman Shift (cm )

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 3+ Nd Nd3+

3+ Nd

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+

3+ Nd 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), 910 nm (7801830 cm-1), and 948 nm (7802278 cm-1) likely related to luminescence centers of unidentified REE in the Raman spectrum of uvarovite; sample R060477, Ural Mts., Russia; unoriented

? Nd3+ ? Intensity Nd3+ ?

Nd3+ Nd3+

Pyrope Raman Shift (cm-1)

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+

? ?

3+ ? Nd 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 located at 894 nm (7801633 cm-1), 903 nm (7801750 cm-1), and 951 nm (7802312 cm-1) 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 (7851777 cm-1) likely related to luminescence centers of unidentified REE in the Raman spectrum of spessartine; sample R060279, red-brown spessartine from Pakistan, unoriented

Nd3+ 3+ Nd

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 nm cm-1 transparent tan, R060382 Grossular 860 7851123 Nd3+ 6.10 A & B Gaft et al., 2003 (titanite) Lalatema Tanzania 875 7851314 Nd? 6.10 A & B 879 7851367 Nd3+ 6.10 A & B Gaft et al., 2003 (titanite) 888 7851490 Nd3+ 6.10 A & B Gaft et al., 2003 (titanite) 895 7851568 ? 6.10 A & B 911 7851763 ? 6.10 A & B 933 7852023 Nd3+ 6.10 A & B Gaft et al., 2003 (titanite) 950 7852216 ? 6.10 A & B 1041 7853135 Nd? 6.10 A & C 1061 7853315 Nd3+ 6.10 A & C Gaft et al., 2003 (titanite) 1070 7853396 Nd3+ 6.10 A & C Gaft et al., 2003 (titanite) 1082 7853502 Nd3+ 6.10 A & C Gaft et al., 2003 (titanite) yellowish tan, unknown R060443 Grossular 790 5326142 ? 6.14 locality 801 5326321 ? 6.14 807 5326408 ? 6.14 818 5326584 Nd? 6.14 Gaft et al., 2005 (zircon) R060477 green, Ural Mts., Russia Uvarovite 861 7801206 Nd3+ 6.11 Gaft et al., 2003 (titanite) 874 7801387 Nd? 6.11 878 7801441 Nd3+ 6.11 Gaft et al., 2003 (titanite) 888 7801559 Nd3+ 6.11 Gaft et al., 2003 (titanite) 894 7801632 ? 6.11

910 7801830 ? 6.11 932 7802091 Nd3+ 6.11 Gaft et al., 2003 (titanite) 932 7802091 Nd3+ 6.11 Gaft et al., 2003 (titanite) 948 7802278 ? 6.11 green, Outokumpu, R061041 Uvarovite 1580 5326411 ? 6.16 Finland 1598 5326484 ? 6.16

1621 5326571 ? 6.16 colorless, synthetic, P = R080060 23.5 kb, T = 1000 deg C Pyrope 863 7801239 Nd3+ 6.12 Gaft et al., 2003 (titanite) for 23 hours 871 7801348 Nd3+ 6.12 Gaft et al., 2003 (titanite) 877 7801426 Nd3+ 6.12 Gaft et al., 2003 (titanite) 879 7801446 Nd3+ 6.12 Gaft et al., 2003 (titanite) 885 7801526 Nd3+ 6.12 Gaft et al., 2005 (zircon)

894 7801644 ? 6.12 903 7801750 ? 6.12 942 7802207 Nd3+ 6.12 Gaft et al., 2003 (titanite) 951 7802312 ? 6.12

Sample Garnet Description/Origin Spectral Feature Ion Figure Reference No. Specie nm cm-1 R060279 red-brown, Pakistan Spessartine 877 7851335 Nd3+ 6.13 Gaft et al., 2003 (titanite) 886 7851457 Nd3+ 6.13 Gaft et al., 2005 (zircon) 904 7851675 ? 6.13 912 7851777 ? 6.13 930 7851991 Nd3+ 6.13 Gaft et al., 2003 (titanite) 942 7852126 Nd3+ 6.13 Gaft et al., 2003 (titanite) R050063 red-brown, China Spessartine 968 7852414 ? 6.15

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 -1 - 3652/3602 and 3640 cm 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.

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, 3570, 3610, 3645, and 3686 cm-1 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-

-1 Raman Shift (cm )

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 properties of Cr3+ and Nd3+ 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. Struve, B. & Huber, G. (1985) The effect of the crystal field strength on the optical spectra of Cr3+ 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.

Olivine

Gem names: peridot

2+ Ideal chemistry: (Mg,Fe )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 -1 peaks, centered at 824 and 856 cm , 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 Ag

Intensity 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

3+ Cr

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.

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 A E Position Symmetry 1

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 o high temperature β phase (P3121) at 573 C (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 related to an E mode centered at 1066 cm-1 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 A 697 808 1083 1160 1231 1 E 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.

Spinel

Ideal chemistry: MgAl O 2 4

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 3 m - - - Mg 8a 43m - - 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+

3+ Cr

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

3+ Cr Cr3+

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

Intensi 4. 3. 2.

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

Intensity 4.

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 2.

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

Intensity 2.

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 -1 spectrum. The 727 cm 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. Martignago, F., Negro, A.D. & Carbonin, S. (2003) How Cr3+ and Fe3+ 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.

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

3+ Al hole center or Mn2+?

Intensity

-1 Raman Shift (cm )

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. Walker, G., El Jaer, A., Sherlock, R., Glynn, T.J., Czaja, M. & Mazurak, Z. (1997) Luminescence spectroscopy of Cr3+ 2+ and Mn in spodumene (LiAlSi2O6). Journal of Luminescence, Vol. 72, pp. 278-280.

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.

Titanite

Gem names: sphene (old name)

Ideal chemistry: CaTiSiO 5

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

.193 4. Intensity

Intensity .079 3. 2. .078 1. .059

Raman Shift (cm-1) 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

? 3+ Sm ? 3+ Eu ?

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 cm-1), and 620 nm (5322682 cm-1); sample R050114, unoriented

Sm3+ Sm3+

Intensity 3+ Sm

Raman Shift (cm-1)

Fig. 11.4 C. Magnification of peaks attributed to Eu3+ luminescence centers in titanite located at 689nm (5324263 cm-1) and 703 nm (5324545 cm-1 and 5324651 cm-1) as well as peaks likely related to unidentified REE centers located at 639 nm (5323162 cm-1), 647 nm (5323360 cm-1), 658 nm (5323619 cm-1), 664 nm (5323737 cm-1), 718 nm (5324870 cm-1), and 729 nm (5325086 cm-1); sample R050114, Pakistan, unoriented

Eu3+

Intensity ?

? ?

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 nm (7851038 cm-1), 905 nm (7851696 cm-1), 923 nm (7851904 cm-1), 934 nm (7852039 cm-1), and 972 nm (7852455 cm-1) in the Raman spectrum of titanite from Kingman, AZ; sample R050039, unoriented

Nd3+ ?

Intensity

? ? ?

Raman Shift (cm-1)

Fig. 11.6 Magnification of peaks located at 794 nm (5326217 cm-1), 799 nm (5326290 cm-1), and 806 nm (5326393 cm-1) 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.

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 -1 16 9 3639 cm is related to a change in the local symmetry from D 2h to C 2v 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-

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+

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 (5325155 cm-1) in the Raman spectrum of topaz attributed to Cr3+ 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

3+ Cr

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. Gaft, M., Nagli, L., Reisfeld, R., Panczer, G. & Brestel, M. (2003) Time-resolved luminescence of Cr3+ 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.Â 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. Tarashchan, A.N., Taran, M.N., Rager, H. & Iwanuch, W. (2006) Luminescence spectroscopic study of Cr3+ 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

H-M: 3m S: C Point Group: 3v 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 A E Position 1

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 H O Reference Variety 2 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 3474/3507 AlYAlZAlZ Gonzales- Y Y Y Y Z Z Carreño, 3670 Li Mn Al 3594 Li Al Al 1988 3568 R2+YAlZAlZ Mn-Elbaite 3492 AlYAlZAlZ

3692 LiYAlYFeY 3594 LiYAlZAlZ Fe-Elbaite 3558 R2+YAlZAlZ 3492 AlYAlZAlZ

3641, Y Y Y 3457-60, Y Z Z 3170, Elbaite (Al,Li) Al Al (Al,Li) Al Al 3646-47 3580-82 3340 3165, Castañeda et 3478-79, 3175, al., 2000 Fe-Elbaite 3585, 3592 LiYAlYFeY 3484, 3557, (Fe,Li,Al)YAlZAlZ 3360, 3560 3380

3468-70, 3645, (Al,Li) YAlYAlY 3478, AlYAlZAlZ 3648-51 Tourmaline 3490-92 Zhang et al., (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 Y Z Z Gonzales-Carreño, 3496 OH3 Al Al Al 1988 R050487, R060652, Y Z Z 13.2 B. Zhang et al., 2008; Elbaite OH3 Fe Al Al , 3566 2+Y Z Z Gonzales-Carreño, R050260 OH3(R Al Al ) & C. 1988 3598 OH3 LiYAlZAlZ Zhang et al., 2008

Y Z Z Gonzales-Carreño, 3476 OH3 Al Al Al R060560, R050119, 13.2 A. 1988 Elbaite Y Y Y R060003, R060566 3592 OH1 Li Al Fe & C. Castañeda et al., 2000 3656 ? 3511 ? 13.3 A. Liddicoatite R060635 3609 ? & C. 3489 OH3 AlYAlZAlZ Zhang et al., 2008 13.3 B. Liddicoatite R060969 3595 OH3 LiYAlZAlZ Zhang et al., 2008 & C. 3657 ? 3571 ? 13.4 A. Dravite R050077 Y Y Y Gonzales-Carreño, 3740 OH1Mg Mg Mg & B. 1988 OH1 LiYAlYFeY, Castañeda et al., 2000; 3584 Y Z Z 13.5 A. Uvite R050172 OH3 Li Al Al 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-

3. 2.

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-

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

Intensity 5.

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 and 3609 cm-1 (no literature match), brown gem of unknown locality; 2) sample R060969, centered at 3489 and 3595 cm-1, red crystal from Madagascar, 532 nm laser, unoriented

Intensity 2.

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

-1 Raman Shift (cm )

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 (5324110 cm-1) and 683 nm (5324160 cm-1) associated with Cr3+ in uvite; sample R050301 from Bahia Brazil, unoriented

Cr3+

Intensity

Raman Shift (cm-1) 3+ Cr

Intensity Fig. 13.7 B. (left) Magnification of peak centered at 684 nm (5324185 cm-1) associated with a Cr3+ 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 cm-1), and 1053 nm (7803323 cm-1) associated with Nd3+ 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 cm-1), and 1053 nm (7803323 cm-1) associated with Nd3+ 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. Mattson, S.M. & Rossman, G.R. (1987) Fe2+-Fe3+ 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: I4 /amd 1

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 42m - 1 - 1 Zr 4a 42m - 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 225 974 Eg A1g 202 393 Eg B1g

-1 Raman Shift (cm )

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

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

4. Intensity Intensity .120 3. .099

2. .095 1.

Raman Shift (cm-1) 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 (5321910 cm-1), 596 nm (5322012 cm-1), 604 nm (5322260 cm-1), and 614 nm (5322560 cm-1); sample R050203, natural red zircon, Thailand, unoriented

Eu3+

Intensity

Raman Shift (cm-1) 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+

3+ Nd Intensity

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)

Fig. 14.8 B. (left) Magnification of peaks centered at 532 -1 532 -1

Intensity 797 nm ( 6254 cm ), 803 nm ( 6348 cm ), and 810 nm (5326450 cm-1) 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 A B B B Position Symmetry g 1g 2g 3g 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 -1 at 2330 and 2900 cm 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-

-1 Raman Shift (cm )

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

Intensity

Raman Shift (cm-1)

Fig. 15.4 A. Peaks in the Raman spectrum of zoisite likely related to the presence of REE luminescence centers; peaks centered at 872 nm (7851270cm-1), 880 nm (7851378 cm-1), 890 nm (7851504 cm-1), and 897 nm (7851602 cm-1) are related to Nd3+; 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 3+ Nd

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 -1 nm ion/molecule mineral reference shift (cm ) elbaite & 7801196 860 Nd3+ Gaft et al., 2003 liddicoatite 7851269 872 Nd3+ titanite Gaft et al., 2005

7851270 872 Nd3+ zoisite Gaft et al., 2005 elbaite & 7801314 869 Nd3+ Gaft et al., 2003 liddicoatite 7801327 870 Nd3+ zircon Gaft et al., 2005

7851374 880 Nd3+ titanite Gaft et al., 2005

7851378 880 Nd3+ zoisite Gaft et al., 2005 elbaite & 7801433 878 Nd3+ Gaft et al., 2003 liddicoatite 7801450 879 Nd3+ zircon Gaft et al., 2005

7851504 890 Nd3+ zoisite Gaft et al., 2005

7851588 897 Nd3+ titanite Gaft et al., 2005

7801600 891 Nd3+ zircon Gaft et al., 2005

7851602 897 Nd3+ zoisite Gaft et al., 2005 elbaite & 7801668 896 Nd3+ Gaft et al., 2003 liddicoatite OH- 1165 N/A topaz Beny & Piriou, 1987 (stretching) 1280-1289 N/A 13C diamond McNamara et al., 1992

5321394 574 Sm3+ titanite Gaft et al., 2005

5321526 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

5321720 585 Sm3+ titanite Gaft et al., 2005 2nd order 1817 N/A diamond Solin & Ramdas, 1970 Raman 2nd order 1864 N/A diamond Solin & Ramdas, 1970 Raman

132 laser (nm) Raman -1 nm ion/molecule mineral reference shift (cm ) 5321910 591 Eu3+ zircon Gaft et al., 2005

5322013 596 Eu3+ zircon Gaft et al., 2005 2nd order 2025 N/A diamond Solin & Ramdas, 1970 Raman 5322151 600 Sm3+ titanite Gaft et al., 2005 2nd order 2177 N/A diamond Solin & Ramdas, 1970 Raman 2nd order 2254 N/A diamond Solin & Ramdas, 1970 Raman 5322260 604 Eu3+ zircon Gaft et al., 2005 2nd order 2333 N/A diamond Solin & Ramdas, 1970 Raman 2nd order 2458 N/A diamond Solin & Ramdas, 1970 Raman 5322491 614 Eu3+ titanite Gaft et al., 2005 2nd order 2519 N/A diamond Solin & Ramdas, 1970 Raman 5322560 614 Sm3+ zircon Gaft et al., 2005

5322597 617 Sm3+ titanite Gaft et al., 2005 2nd order 2667 N/A diamond Solin & Ramdas, 1970 Raman Al3+-hole or Gaft et al., 2005; 5322670 620 spodumene Mn2+? Souza et al., 2003 5322682 620 Sm3+ titanite Gaft et al., 2005

5323100 638 NV- center diamond Zaitsev, 2001

3160-3170 N/A bridging OH- zoisite Winkler et al., 1989 3rd order 3300 N/A diamond Bormett et al., 1995 Raman 5323370 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

5323540 655 Cr3+ chrysoberyl Gaft et al., 2005

133 laser (nm) Raman -1 nm ion/molecule mineral reference shift (cm ) Zhang et al., 2008; Gonzales- 3566 N/A OH3 elbaite Carreño, 1988 3592 N/A OH1 elbaite Castañeda et al., 2000 Type-II H O 3594/3597 N/A 2 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 H O 3606-3609 N/A 2 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

5323654 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

5323745 664 Cr3+ chrysoberyl Gaft et al., 2005

5323772 666 Sm3+ zircon Gaft et al., 2005 3rd order 3825 N/A diamond Bormett et al., 1995 Raman

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

5323900 671 Cr3+ diopside Gaft et al., 2005

5324000 676 Cr3+ spinel Gaft et al., 2005

5324055 678 Cr3+ topaz Gaft et al., 2005

5324062 679 Cr3+ chrysoberyl Gaft et al., 2005

5324075 679 Cr3+ topaz Gaft et al., 2005

5324082 679 Cr3+ diopside Gaft et al., 2005

5324095 680 Cr3+ beryl Gaft et al., 2005

5324100 680 Cr3+ chrysoberyl Gaft et al., 2005

5324110 680 Cr3+ uvite Gaft et al., 2005

5324160 683 Cr3+ uvite Gaft et al., 2005

5324160 683 Cr3+ topaz Gaft et al., 2005

134 laser (nm) Raman -1 nm ion/molecule mineral reference shift (cm ) 5324170 683 Cr3+ beryl Gaft et al., 2005

5324175 683 Cr3+ topaz Gaft et al., 2005

5324185 684 Cr3+ dravite Gaft et al., 2005

5324196 684 Cr3+ diopside Gaft et al., 2005

5324211 686 Cr3+ spinel Gaft et al., 2005

5324263 689 Eu3+ titanite Gaft et al., 2005

5324340 692 V zoisite Gaft et al., 2005

5324366 693 Cr3+ corundum Gaft et al., 2005

5324367 693 Cr3+ chrysoberyl Gaft et al., 2005

5324373 693 Cr3+ forsterite Gaft et al., 2005

5324396 694 Cr3+ corundum Gaft et al., 2005

5324400 694 Cr3+ forsterite Gaft et al., 2005

5324466 698 Cr3+ spinel Gaft et al., 2005

5324510 700 Cr3+ chrysoberyl Gaft et al., 2005

5324526 700 Cr3+ diopside Gaft et al., 2005

5324545 701 Eu3+ titanite Gaft et al., 2005

5324572 702 Eu3+ zircon Gaft et al., 2005

5324573 703 V2+ chrysoberyl Gaft et al., 2005

5324651 706 Eu3+ titanite Gaft et al., 2005

5324657 708 Cr3+ spinel Gaft et al., 2005

5324695 709 V zoisite Gaft et al., 2005

5324728 710 Eu3+ zircon Gaft et al., 2005

5324762 712 Cr3+ topaz Gaft et al., 2005

5324859 718 Cr3+ spinel Gaft et al., 2005

5324958 722 V3+ diopside Gaft et al., 2005

5325155 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 -1 nm ion/molecule reference shift (cm )

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 H O beryl 3594/3597 N/A 2 Lodzinski et al., 2005 (stretching) Type-I H O beryl 3606-3609 N/A 2 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

beryl 5324095 680 Cr3+ Gaft et al., 2005

beryl 5324170 683 Cr3+ Gaft et al., 2005

chrysoberyl 5323370 650 Cr3+ Gaft et al., 2005

chrysoberyl 5323540 655 Cr3+ Gaft et al., 2005

chrysoberyl 5323745 664 Cr3+ Gaft et al., 2005

chrysoberyl 5324062 679 Cr3+ Gaft et al., 2005

chrysoberyl 5324100 680 Cr3+ Gaft et al., 2005

chrysoberyl 5324367 693 Cr3+ Gaft et al., 2005

chrysoberyl 5324510 700 Cr3+ Gaft et al., 2005

chrysoberyl 5324573 703 V2+ Gaft et al., 2005

corundum 5324366 693 Cr3+ Gaft et al., 2005

corundum 5324396 694 Cr3+ Gaft et al., 2005

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

13 McNamara et al., diamond 1280-1289 N/A C 1992

2nd order Solin & Ramdas, diamond 1817 N/A Raman 1970

nd 2 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 Raman 1970

2nd order Solin & Ramdas, diamond 2333 N/A Raman 1970

2nd order Solin & Ramdas, diamond 2458 N/A Raman 1970

nd 2 order Solin & Ramdas, diamond 2519 N/A Raman 1970

2nd order Solin & Ramdas, diamond 2667 N/A Raman 1970 diamond 5323100 638 NV- center Zaitsev, 2001 3rd order diamond 3300 N/A Bormett et al., 1995 Raman 3rd order diamond 3825 N/A Bormett et al., 1995 Raman diopside 5323900 671 Cr3+ Gaft et al., 2005 diopside 5324082 679 Cr3+ Gaft et al., 2005 diopside 5324196 684 Cr3+ Gaft et al., 2005 diopside 5324526 700 Cr3+ Gaft et al., 2005 diopside 5324958 722 V3+ Gaft et al., 2005 forsterite 5324373 693 Cr3+ Gaft et al., 2005 forsterite 5324400 694 Cr3+ Gaft et al., 2005

spinel 5324000 676 Cr3+ Gaft et al., 2005

spinel 5324211 686 Cr3+ Gaft et al., 2005

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

spinel 5324466 698 Cr3+ Gaft et al., 2005

spinel 5324657 708 Cr3+ Gaft et al., 2005

spinel 5324859 718 Cr3+ Gaft et al., 2005

3+ 532 Al -hole or Gaft et al., 2005; spodumene 2670 620 Mn2+? Souza et al., 2003

titanite 7851269 872 Nd3+ Gaft et al., 2005

titanite 7851374 880 Nd3+ Gaft et al., 2005

titanite 7851588 897 Nd3+ Gaft et al., 2005

titanite 5321394 574 Sm3+ Gaft et al., 2005

titanite 5321526 578 Sm3+ Gaft et al., 2005

titanite 5321720 589 Sm3+ Gaft et al., 2005

titanite 5322151 600 Sm3+ Gaft et al., 2005

titanite 5322491 613 Sm3+ Gaft et al., 2005

titanite 5322597 617 Sm3+ Gaft et al., 2005

titanite 5322682 620 Sm3+ Gaft et al., 2005

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

titanite 5324263 689 Eu3+ Gaft et al., 2005

titanite 5324545 703 Eu3+ Gaft et al., 2005

titanite 5324651 706 Eu3+ Gaft et al., 2005

OH- topaz 1165 N/A Beny & Piriou, 1987 (stretching)

topaz 5325155 733 Cr3+ Gaft et al., 2005

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

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

topaz 5324055 678 Cr3+ Gaft et al., 2005

topaz 5324075 679 Cr3+ Gaft et al., 2005

topaz 5324160 683 Cr3+ Gaft et al., 2005

topaz 5324175 683 Cr3+ Gaft et al., 2005

topaz 5324762 712 Cr3+ Gaft et al., 2005 tourmaline 7801196 860 Nd3+ Gaft et al., 2003 tourmaline 7801314 869 Nd3+ Gaft et al., 2003 tourmaline 7801433 878 Nd3+ Gaft et al., 2003 tourmaline 7801668 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 tourmaline 5324110 680 Cr3+ Gaft et al., 2005 tourmaline 5324160 683 Cr3+ Gaft et al., 2005 tourmaline 5324185 684 Cr3+ Gaft et al., 2005

zircon 7801327 870 Nd3+ Gaft et al., 2005

zircon 7801450 879 Nd3+ Gaft et al., 2005

zircon 7801600 891 Nd3+ Gaft et al., 2005

zircon 5321910 591 Eu3+ Gaft et al., 2005

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

zircon 5322013 596 Eu3+ Gaft et al., 2005

zircon 5322260 604 Eu3+ Gaft et al., 2005

zircon 5322560 614 Eu3+ Gaft et al., 2005

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

zircon 5323654 660 Sm3+ Gaft et al., 2005

zircon 5323772 666 Sm3+ Gaft et al., 2005

zircon 5324572 702 Eu3+ Gaft et al., 2005

zircon 5324728 710 Eu3+ Gaft et al., 2005

zoisite 7851270 872 Nd3+ Gaft et al., 2005

zoisite 7851378 880 Nd3+ Gaft et al., 2005

zoisite 7851504 890 Nd3+ Gaft et al., 2005

zoisite 7851602 897 Nd3+ Gaft et al., 2005

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

zoisite 5324340 692 V Gaft et al., 2005

zoisite 5324695 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

18796.99248 cm-1 – 1378 cm-1 = 17418.99 cm-1 (wave numbers)

2) Convert wave numbers into cm and then nm.

1/17418.99 cm-1 = 5.74 x 10-5 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 2+ Fritsch and Rossman, channel-filling Fe (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 3+ coordinated Cr 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 2+ 2+ Fritsch and Rossman, orange (hessonite) Mn or Fe 1988; Manning, 1970 octahedrally Fritsch and Rossman, pyrope red coordinated Cr3+and 1988; Manning, 1967 some Fe2+ 2+ Fritsch and Rossman, brownish-red Fe 1988; Manning, 1967

144 Mineral Color Cause Reference

2+ Gubelin, 1982; Manning, spesartine orange Mn 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 2+ Paradise, 1982; Fritsch and green (praseolite) Fe 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 3+ spodumene Mn 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 Cr3+ 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 4+ Cr 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 2+ Cu Rossman et al., 1991 (Paraiba) octahedrally uvite green coordinated Cr3+ and Fritsch and Rossman, 1988 V3+ Nb4+ and irradiation Fielding, 1970; Fritsch and zircon red center Rossman, 1988 4+ Mackey et al., 1975; Fritsch blue presence of U and Rossman, 1988 blue-violet + 3+ zoisite V4 and V 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 C C C 3 3 3i 6 3h 4 C4 4/m C4h mm2C2v 6 C6 6/m C6h 3m C3v

222 D2 mmm D2h 4 2m D2d

32 D D 4mm C 3 3 m 3d 4v 422 D4 4/mmm D4h 6 2m D3h

622 D6 6/mmm D6h 6mm C6v

23 T T T m 3 h 4 3m d 432 O m 3 m 0h

235 I m 3 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