Skip to Main Content

Abstract

It is said that during a voyage to Europe in the summer of 1921, the Indian physicist Chandrasekhara Venkata Raman (1888–1970) looked at the wonderful blue opalescence of the Mediterranean Sea and questioned where the sea's blue colour came from and why it should be different from the sky's blue. Raman started a series of experiments to address these questions, and he found the blue colour of the sea was not merely due to simple reflection of the sky in water, as most people imagined, but was additionally affected by molecular scattering of light. This led to the discovery of a new inelastic scattering process that is the optical analogue of the “Compton effect”; it is nowadays known as the “Raman effect”. It describes a change in the wavelength of light that occurs when a light beam interacts with molecular vibrations. The possibility for such interaction between matter and light had already been predicted theoretically by Smekal (1923). The first verification was obtained by Raman and Krishnan (1928) in light scattering experiments on liquids. Only two years later, Sir C.V. Raman (who was knighted in 1929) was the Nobel laureate in physics, honoured for his work on the scattering of light and the discovery of the effect named after him. In his Nobel lecture, given on 11th December 1930, Sir C.V. Raman said “The frequency differences determined from the spectra, the width and character of the lines appearing in them, and the intensity and state of polarization of the scattered radiations enable us to obtain an insight into the ultimate structure of the scattering substance. […] It follows that the new field of spectroscopy has practically unrestricted scope in the study of problems related to the structure of matter” In 1948, he founded the Raman Research Institute in Bangalore, India, with funds from private sources.

Introduction

It is said that during a voyage to Europe in the summer of 1921, the Indian physicist Chandrasekhara Venkata Raman (1888–1970) looked at the wonderful blue opalescence of the Mediterranean Sea and questioned where the sea's blue colour came from and why it should be different from the sky's blue. Raman started a series of experiments to address these questions, and he found the blue colour of the sea was not merely due to simple reflection of the sky in water, as most people imagined, but was additionally affected by molecular scattering of light. This led to the discovery of a new inelastic scattering process that is the optical analogue of the “Compton effect”; it is nowadays known as the “Raman effect”. It describes a change in the wavelength of light that occurs when a light beam interacts with molecular vibrations. The possibility for such interaction between matter and light had already been predicted theoretically by Smekal (1923). The first verification was obtained by Raman and Krishnan (1928) in light scattering experiments on liquids. Only two years later, Sir C.V. Raman (who was knighted in 1929) was the Nobel laureate in physics, honoured for his work on the scattering of light and the discovery of the effect named after him. In his Nobel lecture, given on 11th December 1930, Sir C.V. Raman said “The frequency differences determined from the spectra, the width and character of the lines appearing in them, and the intensity and state of polarization of the scattered radiations enable us to obtain an insight into the ultimate structure of the scattering substance. […] It follows that the new field of spectroscopy has practically unrestricted scope in the study of problems related to the structure of matter” In 1948, he founded the Raman Research Institute in Bangalore, India, with funds from private sources.

Raman spectroscopy has become an important and versatile spectroscopic technique that is nowadays commonly used in many scientific and industrial disciplines. The traditional fields in which Raman spectroscopy has a long and well-established history and tradition are condensed matter physics and chemistry. In the Earth sciences, in contrast, Raman was for long applied with restraint, even though the second publication dealing with the Raman effect (Landsberg & Mandelstam, 1928) had already described the Raman scattering observed in mineralogical samples. With this statement, we would not like to discount many excellent Raman studies on minerals that were done prior to the 1980s. It is, however, a fact that for a long time there existed only a limited number of Raman laboratories dealing with geoscientific problems (in contrast, there exists at least one infrared spectrometer system at almost every mineralogy institute across Europe). The formerly hesitant use of Raman spectroscopy in mineralogy was most probably related to experimental difficulties, which in turn are due to the fact that natural minerals are, in contrast to synthetic chemical substances or semiconductor samples, rich in chemical impurities, microinclusions and structural defects. These sample-related problems do still exist, but they can be managed today owing to the availability of particularly powerful Raman spectrometer systems equipped with sensitive detectors.

This paper is written for mineralogists and geologists who are interested in Raman spectroscopy and would perhaps like to start using this technique in their own research. Correspondingly, this paper will be focussed on the two major questions that are of interest for “Raman beginners”, namely, how Raman works, and to what kinds of samples it may be applied. The theoretical and experimental sections will indeed remain on an introductory level, and techniques such as RRS (resonance Raman spectroscopy), SERS (surface-enhanced Raman scattering) and nano-Raman spectroscopy (also known as near-field Raman) will not be discussed. We consider that it is more important to deal in depth with selected aspects and potential problems related to the recording and interpretation of Raman spectra, and to provide the reader with some basic information on the Raman terminology that is needed when working with the Raman literature. The overview of applications is a selection of examples chosen by the authors, aiming to underline the versatility of the Raman technique. Even though this paper provides the reader with an extensive list of references, these references do not claim to be comprehensive, as they cover only a small fraction of what has been done thus far. As a demonstration of the practical use of particular analytical advantages of the Raman technique, we discuss in more detail five examples of its application to mineralogy.

Theoretical background and practical aspects

The Raman Effect

Electrodynamical model

The electric field vector of visible light (wavelength λ = 400-750 nm) oscillates at a high frequency v0 in the range 4.0-7.5 · 1014 s-1. When such radiation is applied to a molecule or crystal lattice, a coherence between the electrons rotating around the atomic nuclei and the radiation is established, and a charge separation oscillating at the same frequency as the electromagnetic wave (v0) is produced. This vibrational energy is in most cases immediately released by the production of diffuse, elastically scattered light having the same frequency and wavelength as the incident beam of light. This process is called Rayleigh scattering.

Atoms in a mineral, as well as in liquids and gases and other forms of matter, vibrate at frequencies v1 on the order of 1012−1014 s−1 (i.e., v1 << v0). Due to the considerable frequency difference, the excitation of vibrations of the nuclei through simple absorption of incident light is impossible. However, the oscillating electric field of the light can interact with atomic vibrations in an inelastic scattering process, the so-called Raman scattering. The possibility of such interaction is visualised by the following consideration. The nuclei of atoms in the sample are too heavy to follow the high-frequency vibration (v0) of the electric field vector of the incident light. As a result, a time-dependent dipole moment (μ) is induced by the electromagnetic wave. (In contrast, in the absence of radiation on average there is no charge separation and no dipole moment is created, or no variation of the dipole moment is created if the molecule possesses a dipole moment.) This induced dipole moment changes with time according to  

formula
in which forumla cos(v0t) describes the strength of the oscillating electric field of the light, and α is the electric polarisability tensor of the molecule or crystal. The polarisability is also not a constant but varies with time, because the ability of electrons to be displaced with respect to their corresponding nuclei must depend on the actual positions of nuclei. Therefore, α is strongly controlled by the vibrations of atoms (v1) in the sample.

Raman scattering can, largely simplified, be described as the product of the interaction of the time-dependent dipole moment induced by the electromagnetic wave and the electromagnetic wave itself. It is possible that the light-induced deformation of the electron cloud excites the nuclei to vibrate. However, deformation of the electron cloud as induced by vibrations of nuclei may also have an exciting effect on the electric field and, with that, the vibration of the light. Both of these cases have in common that two vibrations with significantly different frequencies (v0 and v1) modulate one another. Their interaction - called the Raman effect - is controlled by the polarisability α. This is the main difference to infrared (IR) absorption, where vibrations of light and vibrations of the sample have the same frequency and the interaction depends on the dipole moment μ (see Fig. 1).

Fig. 1.

Raman and infrared activity depending on the dipole moment (μ) and polarisability (α), elucidated with the example of stretching vibrations of the linear CO2 molecule. The symmetric stretching vibration is Raman-active (α oscillates) but not IR-active (no induced dipole moment). By contrast, the anti-symmetric stretching vibration is IR-active but nearly Raman inactive. Picture redrawn from Schmidt (1994).

Fig. 1.

Raman and infrared activity depending on the dipole moment (μ) and polarisability (α), elucidated with the example of stretching vibrations of the linear CO2 molecule. The symmetric stretching vibration is Raman-active (α oscillates) but not IR-active (no induced dipole moment). By contrast, the anti-symmetric stretching vibration is IR-active but nearly Raman inactive. Picture redrawn from Schmidt (1994).

It can be seen from Figure 1 that in a first, very rough, approximation, symmetric vibrations (which are connected with significant changes of α) are strongly Raman-active whereas antisymmetric vibrations (connected with the induction of a dipole moment μ) are strongly IR-active. This rule, however, holds strictly only for molecules possessing a centre of symmetry. The Raman and/or IR activity or inactivity of vibrational modes, respectively, is described by the “selection rules” (e.g. Rousseau et al.,1981).

Quantum-mechanical model

Another way to describe the Raman effect is based on the quantisation of the matter by taking into account the vibrational states of molecules and crystals. Similar to light energy, vibrational levels cannot have any arbitrary amount of energy; rather there exist smallest possible energy portions (vibrational quanta, in solids, are usually called phonons). Correspondingly, the vibrational states of a molecule or crystal are characterised by a limited number of allowed, discrete energies. Upon irradiation with a beam of light, a molecule or crystal can be transferred to a higher energy (excited) state through the absorption of a light photon and the excitation of a vibrational phonon. A necessary condition for the absorption, however, is that the energy of the incoming photon (E = hv0; where h is Planck's constant) is equivalent to the energy difference between two allowed vibrational states of the molecule or crystal. This is the case for light in the infrared (IR) range, and such genuine absorption is detected using the IR absorption spectroscopy technique (Fig. 2, sketch 1).

Fig. 2.

Elucidation of the light-molecule interaction using a simplified energy level diagram. (1) Light can be absorbed when its photon energy corresponds to the energy difference between two allowed vibrational levels. The absorption of a photon generates a phonon (vibrational quantum) with the same energy. Such interaction of molecules is possible with middle to far infrared light. (2) Visible light, by contrast, cannot be absorbed through phonon excitation in the molecule, because its photon energy is much higher than energy differences between vibrational states of the molecule. The excitation with visible light leads to a virtual electronic state from which the system recovers immediately. The molecule will in most cases have the same vibrational state as before the interaction and, correspondingly, the photon energy of the scattered light (Rayleigh scattering) corresponds to the initial value. (3) Very rarely, the molecule may reach a higher or lower vibrational state than before the interaction. The phonon energy of such scattered light (Raman scattering) is either somewhat decreased or increased with respect to the exciting photon. This photon energy difference (energy shift; Raman shift) corresponds to the energy difference between vibrational levels of the molecule.

Fig. 2.

Elucidation of the light-molecule interaction using a simplified energy level diagram. (1) Light can be absorbed when its photon energy corresponds to the energy difference between two allowed vibrational levels. The absorption of a photon generates a phonon (vibrational quantum) with the same energy. Such interaction of molecules is possible with middle to far infrared light. (2) Visible light, by contrast, cannot be absorbed through phonon excitation in the molecule, because its photon energy is much higher than energy differences between vibrational states of the molecule. The excitation with visible light leads to a virtual electronic state from which the system recovers immediately. The molecule will in most cases have the same vibrational state as before the interaction and, correspondingly, the photon energy of the scattered light (Rayleigh scattering) corresponds to the initial value. (3) Very rarely, the molecule may reach a higher or lower vibrational state than before the interaction. The phonon energy of such scattered light (Raman scattering) is either somewhat decreased or increased with respect to the exciting photon. This photon energy difference (energy shift; Raman shift) corresponds to the energy difference between vibrational levels of the molecule.

The phonon energy of visible (and ultraviolet and near IR) light is, by contrast, much higher than energy differences between the vibrational states of molecules and crystals and, therefore, direct photon absorption under simultaneous excitation of a phonon is impossible. The incident light will excite the system to a virtual high-energy state from which it recovers immediately. As a rule, the diffusely scattered light will then have the same photon energy and, thus, the same frequency as the incident light (Rayleigh scattering; Fig. 2, sketch 2). It may, however, also happen (even though with low probability) that the system recovers to a higher or lower energetic state when compared to the initial value. If the system gains energy through the excitation of a phonon (E = hv1), the scattered photon [E = h(v0 – v1)] has lost the same energy portion (Stokes-type Raman; Fig. 2, sketch 3a). At temperatures above absolute zero, all matter vibrates. Therefore, it is also possible that an already vibrating system is excited, and such a system may instead lose vibrational energy through the interaction with the light. Here, the scattered photon (anti-Stokes Raman; Fig. 2 sketch 3b) has increased in energy [E = h(v0 + v0)]. It is clear that the intensity ratio of Stokes and anti-Stokes type Raman light must depend on the ratio between molecules in the ground and excited state; it therefore depends on temperature.

Interpretation of spectra

The Raman spectrum

The above considerations describe the Raman effect as comprising two possible interactions. When a portion of the energy of the exciting light is used to excite a vibration in the sample, the Raman scattered light (Stokes; v = v0 - v1) will, due to its partial energy loss, experience a red shift in the electromagnetic spectrum. The opposite case is the loss of vibrational energy in the sample in favour of an increase in light energy. The anti-Stokes Raman light (v = v0 + v1) is, therefore, blue-shifted with respect to the excitation frequency. Consequently, the spectrum of scattered light obtained from a sample irradiated with an incident beam of light consists of three principal parts, namely, the intense Rayleigh line and weak Raman bands in the Stokes and anti-Stokes parts of the spectrum (Fig. 3). For a liquid, the ratio between Rayleigh and Raman scattered light is expected to be about 105:1. Note that generally only the Stokes Raman bands are recorded. This is because the shifts of a Stokes Raman band and its corresponding anti-Stokes counterpart are equal in energy [ΔE = ±(hv1)] but Stokes bands are always higher in intensity (Fig. 3) and, therefore, are more easily detected.

A Raman spectrum is a plot of light intensity (usually given in counts, counts per second or arbitrary units) versus photon energy. In vibrational spectroscopy, it is unusual to express the photon energy by the frequency or wavelength of the light (an exception is Brillouin spectroscopy; light scattered through the interaction with low-frequency acoustic vibrations is mostly plotted as intensity versus frequency). Instead, frequencies or wavelengths are generally transformed into wavenumbers (forumla). The wavenumber is defined as  

formula
where c is the speed of light; the unit of the wavenumber is cm-1. Wavenumbers (or more exactly, absolute wavenumbers) are generally also used for the presentation of IR absorption spectra. In Raman spectroscopy, however, the use of absolute wavenumbers would be impractical, because the wavelength and, with that, the absolute wavenumber of each obtained Raman band forumla must always depend on the wavenumber of the incident light forumla. However, only the wavenumber shift forumla, which corresponds to a specific vibration in the sample, is of analytical interest. It has, therefore, become usual to express Raman shifts using their relative wavenumber (i.e., the wavenumber difference between incident and scattered light; forumla). By convention, the Rayleigh line is set at zero Raman shift and anti-Stokes Raman bands have negative and Stokes Raman bands have positive relative wavenumbers (despite the fact that Stokes Raman bands have a decreased absolute wavenumber). For the interdependence of relative and absolute wavenumber, wavelength and frequency see Figure 3.

Fig. 3.

Comparison of the Stokes and anti-Stokes parts of the Raman spectrum. Spectrum of crocoite (Callenberg, Germany) excited with the Ar+ 514.5 nm line (which corresponds to 19435 cm-1). Raman bands in the Stokes and anti-Stokes parts of the spectrum have the same Raman shift values (= relative wavenumbers, with 0 cm-1 Raman shift set at the Rayleigh line). Note that Stokes bands are always higher in intensity than their anti-Stokes counterparts, with the intensity ratio increasing with increasing Raman shift. Note also that some workers use an opposite convention, with Stokes called negative and anti-Stokes called positive on the grounds that the Stokes Raman bands are actually lower in absolute cm-1 than the Rayleigh line. To elucidate the interrelation of absolute and relative wavenumbers, frequencies and wavelengths, the x axis is given with three additional scales at the top.

Fig. 3.

Comparison of the Stokes and anti-Stokes parts of the Raman spectrum. Spectrum of crocoite (Callenberg, Germany) excited with the Ar+ 514.5 nm line (which corresponds to 19435 cm-1). Raman bands in the Stokes and anti-Stokes parts of the spectrum have the same Raman shift values (= relative wavenumbers, with 0 cm-1 Raman shift set at the Rayleigh line). Note that Stokes bands are always higher in intensity than their anti-Stokes counterparts, with the intensity ratio increasing with increasing Raman shift. Note also that some workers use an opposite convention, with Stokes called negative and anti-Stokes called positive on the grounds that the Stokes Raman bands are actually lower in absolute cm-1 than the Rayleigh line. To elucidate the interrelation of absolute and relative wavenumbers, frequencies and wavelengths, the x axis is given with three additional scales at the top.

Vibrations of molecules and crystal lattices

We have elucidated above that Raman shifts correspond to frequencies of vibrations in the sample, i.e., each band in a Raman spectrum represents the interaction of the incident light with a certain vibration of the nuclei. The vibrations of the nuclei, in turn, are controlled by the sizes, valences and masses of the atomic species of which the sample is composed, the bond forces between these atoms, and the symmetry of their arrangement in the crystal structure (bond directions). These factors affect not only the frequencies of atomic vibrations and the observed Raman shifts, respectively, but also the number of observed Raman bands, their relative intensities, their widths (typically expressed as FWHM, “full width at half band maximum intensity”) and their polarisations. Therefore, Raman spectra are highly specific for a certain type of sample and can be used for the identification and structural characterisation of unknown samples.

One of the most challenging tasks in vibrational spectroscopy (Raman as well as IR absorption and Brillouin spectroscopy) is the reliable assignment of observed bands to certain vibrations in the sample. All vibrations, i.e., collective movements of atoms, are complicated combinations of the so-called normal vibrations or normal modes of the respective molecule or crystal. The number of potentially occurring normal vibrations (note: not necessarily all of them are Raman-active) depends on the number of dynamical degrees of freedom of the system. For a single molecule consisting of n atoms, there are 3n - 6 degrees of vibrational freedom (three for each atom, minus three rotational and three translational principal movements of the entire molecule in the three dimensions of space). Linear molecules have 3n - 5 degrees of vibrational freedom, because the rotation around the main molecule axis does not produce any change of rotational energy. To give two examples, for the hydroxyl group (a diatomic molecule) we calculate only (3 × 2) - 5 = 1 vibration (stretching along the O-H bond) and for the triatomic, non-linear water molecule (3 × 3) -6 = 3 vibrations (symmetric and antisymmetric stretching along the O-H bonds and bending of the H-O-H angle; see Fig. 1 in Beran et al., 2004, in this volume). A mineral or other crystal (normally consisting of >> 1020 atoms) can be considered as a single molecule of almost infinite size, which would result in a correspondingly huge number of degrees of vibrational freedom. In contrast, the number of observed vibrations is always limited. This is due to the periodic arrangement of atoms in the crystal lattice (i.e., identical environments give identical energy shifts), which leads to a comparably small number of longitudinal and transversal lattice vibrations. In general, consideration of the vibration of a mineral lattice can be reduced to the corresponding primitive unit cell and its degrees of vibrational freedom.

It has become commonplace to subdivide the vibrations of a crystal lattice into internal vibrations of molecular units and external vibrations. An example is given in Figure 4, which elucidates the general band assignment for four different minerals containing forumla or forumla molecular groups. Internal stretching and bending vibrations of these tetrahedral XO4 groups have Raman shifts in the range 400-1200 cm-1 whereas external vibrations (involving movements of the entire XO4 groups as well as their neighbouring ions) are observed well below 400 cm-1. Differences between the four spectra are due to (1) the tetrahedra having different central ions, (2) different degrees of distortion of the XO4 groups in the four crystal structures, and (3) different bond strengths between tetrahedra and neighbouring atoms. Other typical examples for internal vibrations in minerals are hydroxyl stretching vibrations in OH-bearing species, internal CO3 vibrations in carbonates, or internal SiO4 vibrations in nesosilicates. By contrast, it is not meaningful to discuss internal SiO4 modes in tectosilicates. Here, internal and external bond forces are rather similar and, therefore, the SiO4 tetrahedron is decidedly not a more or less isolated molecule. It is likewise difficult to discuss the occurrence of independent internal vibrations of (Al,Mg,Fe)(O,OH)6 octahedrons in sheet silicates. Due to strong Si-O bond forces, any atomic movement in the octahedral sheet may be accompanied by similarly intense atomic movements in the tetrahedral sheet and, consequently, octahedrons ought not be treated as if they would occur isolated in the structure (McKeown et al., 1999a, 1999b). Such coupling of internal vibrations with the surrounding lattice does always exist but varies widely in strength. The subdivision into internal and external modes is meaningful only if internal bond forces of the molecular unit are much stronger than bond forces to the surrounding atoms in the mineral structure.

Fig. 4.

Assignment of Raman bands, shown with the example of spectra obtained from apatite (Slyudyanka, Siberia), monazite (Moss, Norway), anglesite (Tsumeb, Namibia) and barite (Freiberg, Germany). Despite the four minerals having different lattice types and containing tetrahedral XO4 groups (X = S, P) that are surrounded by different ions, all four spectra show widely similar fingerprint patterns of four groups of internal vibrations of the XO4 groups, and all spectra are dominated by an intense v1(XO4) band (symmetric stretching of XO4 tetrahedra) in the range 960-990 cm-1. Examples for movements of atoms as they would occur in an isolated tetrahedron are shown with small sketches (compare Siebert, 1966; Smith & Carabatos-Nédelec, 2001).

Fig. 4.

Assignment of Raman bands, shown with the example of spectra obtained from apatite (Slyudyanka, Siberia), monazite (Moss, Norway), anglesite (Tsumeb, Namibia) and barite (Freiberg, Germany). Despite the four minerals having different lattice types and containing tetrahedral XO4 groups (X = S, P) that are surrounded by different ions, all four spectra show widely similar fingerprint patterns of four groups of internal vibrations of the XO4 groups, and all spectra are dominated by an intense v1(XO4) band (symmetric stretching of XO4 tetrahedra) in the range 960-990 cm-1. Examples for movements of atoms as they would occur in an isolated tetrahedron are shown with small sketches (compare Siebert, 1966; Smith & Carabatos-Nédelec, 2001).

There exists in the Raman literature several nomenclatures for the description of Raman modes and their corresponding vibrations and, unfortunately, these are not used uniformly in the literature. We cannot give a complete overview here; rather we will briefly discuss three of the most common nomenclatures. Vibrational modes are often described by symbols that refer to irreducible representations (cf. group theory; references listed below); these symbols consist of capital letters with subscript letters and numbers (for example, A1g mode or Eu mode). Here, the capital letter gives information on the degeneration (if two different vibrations have the same frequency, i.e., they are equal in phonon energy, this is referred to one degenerate mode) and symmetry. A and B modes are single vibrations (expressed as either not degenerate or single degenerate), with A modes being symmetric and B modes antisymmetric with respect to the main symmetry axis. E modes are doubly and F modes (the latter only occurring in lattices with high symmetry) are triply degenerate. The subscripts g and u are used to describe modes that are respectively symmetric (g = gerade) and antisymmetric (u = ungerade) with respect to the symmetry centre. It is clear from the above elucidations that g-type vibrations are in general Raman-active whereas u-type vibrations are IR-active (see again Fig. 1). The subscript numbers may refer to symmetries that are different with respect to the main symmetry axis, or they are simply used to number consecutively types of vibrations which otherwise would not be sufficiently distinguished. Another common nomenclature is the v notation for modes. The v notation should not be confused with the Greek letters symbolising types of vibrational movements (Table 1). Vibrations (or rather groups of vibrations) in a molecule or crystal are numbered consecutively v1, v2, v3, … to avoid the need to specify symmetry details. For internal vibrations, it has become usual to put the respective molecule or group in brackets. Although a generally accepted international standard does not exist, there are agreements especially for v notations of molecular or internal vibrations. For instance, it is generally agreed to describe the symmetric stretching vibration of phosphate groups (an A mode) as v1(PO4) mode (cf. Fig. 4).

Table 1.

Basic types of vibrations and their description with Greek symbols.

SymbolDescription

vstretching(vs = symmetric; va or vas = antisymmetric)
δbendings = symmetric; δa or δas = antisymmetric)
ρrocking
π, ωwagging
τtwisting
SymbolDescription

vstretching(vs = symmetric; va or vas = antisymmetric)
δbendings = symmetric; δa or δas = antisymmetric)
ρrocking
π, ωwagging
τtwisting

It is beyond the scope of the present paper to elucidate the band assignment procedure and how frequencies of normal vibrations can be estimated. The least difficult task is the assignment of internal modes (provided there exist more or less isolated molecular units in the mineral structure), whereas the assignment of external modes is much more challenging. Conclusions by analogy may be drawn from the Raman spectra of unknowns through comparison with spectra of known phases with more simple structures; however, such conclusions may remain tainted with some uncertainty. One empirical approach is to systematically synthesise minerals with just one atom type replaced by another of identical charge but different atomic mass and hence ionic radius, such as Si4+ by Ge4+ or Al3+ by Ga3+ (e.g. Tlili and Smith, 1996), or by another isotope of the same atom type, such as H+ by D+. In the latter case one can for example easily identify O-H deformation vibrations (which occur at low wavenumbers and may be easily confused with other Raman bands) as O-D shifts distinctly (Tlili, 1990). The mathematical prediction of theoretical Raman spectra for a given mineral (including information on band frequencies, intensities and polarisation) is extremely complex and involves group theoretical calculations (e.g. Banerjee et al., 2003). This complexity is underlined by the fact that it is not a static, but rather a dynamic system which is dealt with (i.e., one does not only need to know atomic positions but also precise values for inter-atomic bond forces, directions of atomic movements etc.). As a consequence, fully reliable and detailed band assignment may be difficult, especially for minerals having a low-symmetry lattice or/and a large number of atoms per unit cell.

In a given mineral structure with only one kind of cation occupying a specific site (e.g. Al3+ in the octahedra of garnet) a certain Raman band may be attributed to vibration of this atom or of other atoms affected by this atom (regardless of precisely which atoms vibrate, and how in terms of symmetry). This is the general case and is called a unimodal vibration. In a solid solution like garnet there are numerous homovalent cation exchanges, especially replacement of Ca2+ by Mg2+, Mn2+ or Fe2+and also of Al3+ by Cr3+ or Fe3+. Many natural garnets contain all of these elements but some Raman bands remain unimodal (e.g. internal SiO4 modes) and thus only shift in their wavenumber which varies continously from one end-member to another (Pinet & Smith, 1993). However, other bands are duplicated (bimodal behaviour) but at different wavenumbers as a function of the different atomic masses and ionic radii, and hence vibrational energies, of each of two isomorphous cations. Trimodal behaviour was deduced in the Al3+-Cr3+-Fe3+ Ca-garnet ternary system by Pinet & Smith (1993). Likewise in micas, a strong unimodal mode has often been described as T–O–T vibration across the bridging oxygen between two adjacent TO4 tetrahedra. Regardless whether this description is correct or not (there is some discussion on this topic as the bridging oxygens are basal oxygens which compose a forumla [Si2O3]2+ sheet network), this unimodal vibration becomes trimodal as soon as some Si4+ is replaced by Al3+ and it has been possible to distinguish Si-O-Si, Si-O-Al and Al-O-Al vibrations (Tlili et al., 1989); this becomes hexamodal when some Si4+ is replaced by Ge4+ or Al3+ is replaced by Ga3+ (Tlili, 1990; Tlili & Smith, 1996). Thus for a single physically defined Raman vibration mode there are six chemically different Raman bands, and it is of course rather difficult to distinguish which is which without systematically synthesising pure end-members and also many mixed compositions step by step along binary joins in order to be sure to correctly identify a band when several Raman bands are simultaneously moving in wavenumber and also in relative intensity (as well as overlapping).

Directional dependence of Raman scattering

The Raman scattering process is strongly controlled by geometrical factors (i.e., it depends on the polarisation of the atomic vibration and the geometry of the scattering experiment). An example of this is a hydroxyl stretching vibration. Here, Raman scattering is only possible if the electric field vector of an incident beam of light is not perpendicular to the O-H bond direction (the oscillating electric field must have a vector component other than zero parallel to the direction of the stretching movement to be excited). In a trioctahedral mica such as phlogopite, all O-H bonds of hydroxyl groups in the octahedral sheets are oriented along c* (which is the direction perpendicular to the crystallographic a–b plane; cf. Schroeder, 1990). Consequently, if a phlogopite is irradiated with a laser beam propagation direction parallel to c*, no interaction is possible, independent of the polarisation of the laser light (Fig. 5, beam 1). By contrast, the hydroxyl stretching Raman band is obtained with maximum intensity if the incident laser beam is irradiated along a direction within the crystallographic a–b plane and the electric field vector is polarised perpendicular to this plane and thus vibrates parallel to c* (Fig. 5, beam 3). This has been documented, for example, by Loh (1973) and Tlili et al. (1989). Two analogous examples, showing extensive intensity changes of observed Raman bands depending upon the sample orientation with respect to the polarisation plane of the laser light, are presented in Figure 6.

Fig. 5.

An incident beam of light does only interact with the hydroxyl molecule (shown black) and excites it to vibrate along the O–H bond if the vibrating electric field has at least a vector component parallel to the O–H bond direction. In our case (simplified sketch of a part of a sheet silicate structure), beams 1 and 2 cannot interact with the hydroxyl group. The OH Raman band is detected when the sample is excited with beam 3.

Fig. 5.

An incident beam of light does only interact with the hydroxyl molecule (shown black) and excites it to vibrate along the O–H bond if the vibrating electric field has at least a vector component parallel to the O–H bond direction. In our case (simplified sketch of a part of a sheet silicate structure), beams 1 and 2 cannot interact with the hydroxyl group. The OH Raman band is detected when the sample is excited with beam 3.

Fig. 6.

Two examples demonstrating that Raman-active vibrations may have strong directional dependence. (a) Raman spectra of a gem-quality zircon crystal (non-metamict) from Haddam, Connecticut. For the band assignment cf. Dawson et al. (1971) and Kolesov et al. (2001). (b) Raman spectra of a synthetic α-quartz single crystal (from Nasdala et al., 2004a; redrawn and modified). For the band assignment cf. Scott & Porto (1967) and Etchepare et al. (1974). Scattering geometries are reported using the so-called Porto notation (cf. Damen et al., 1966). Note that apparently different Raman band patterns may simply be due to internal variations of band intensity ratios, which must be considered when applying Raman as a fingerprinting technique.

Fig. 6.

Two examples demonstrating that Raman-active vibrations may have strong directional dependence. (a) Raman spectra of a gem-quality zircon crystal (non-metamict) from Haddam, Connecticut. For the band assignment cf. Dawson et al. (1971) and Kolesov et al. (2001). (b) Raman spectra of a synthetic α-quartz single crystal (from Nasdala et al., 2004a; redrawn and modified). For the band assignment cf. Scott & Porto (1967) and Etchepare et al. (1974). Scattering geometries are reported using the so-called Porto notation (cf. Damen et al., 1966). Note that apparently different Raman band patterns may simply be due to internal variations of band intensity ratios, which must be considered when applying Raman as a fingerprinting technique.

The geometry of the scattering experiment, including the polarisations and directions of propagation of incident and analysed scattered light with respect to the crystallographic orientation of the sample, is usually given by the so-called Porto notation (Damen et al., 1966). For example, in X(YZ)Y, or analogously x(yz)y or a(bc)b, the first and last letter describe the directions of the incident beam and the analysed Raman scattered light, and the two letters inside the brackets describe their respective polarisation. The Porto notation forumla (see Fig. 6a) contains the following information: In the brackets, information on the polarisation plane of the laser light with respect to the crystallographic c axis is given, and it is said that no polariser was used for the scattered light (i.e., all polarisation directions perpendicular to the crystallographic a axis were allowed). According to the Y before the brackets, the light was irradiated along the crystallographic b axis. The forumla after the brackets gives the information that incident and analysed light had opposite directions, i.e., the measurement was done in 180° backscattering geometry. For completeness, the two other principal scattering geometries are 90° scattering (Raman light collected perpendicular to the direction of the incident beam) and 0° forward scattering (Raman light parallel to the incident beam, i.e., analysed straight behind the sample). Note that backscattering (or, more exactly, quasi-backscattering, because genuine backscattering requires to use a parallel beam instead of a convergent beam) is mostly applied for the Raman analysis of minerals, especially via a microscope objective, mainly because the application of the other two geometries presumes suitable preparation of the sample.

The Porto notation gives complete geometric information for a certain scattering experiment. However, a part of this information (namely, both of the light directions) is merely of interest for the experimentalist, whereas only the information on the polarisation is necessary for the band assignment. Therefore, it has become usual in the more recent literature to use expressions that could be understood as “abbreviated Porto notations”. An expression such as “aa geometry” or “aa polarisation” means that both the incident light and the Raman scattered light were polarised parallel to the crystallographic a axis. This expression represents six principally possible scattering geometries, namely b(aa)b, forumla, b(aa)c, c(aa)b, c(aa)c and forumla, and in all of them Raman spectra would be obtained in which bands show the same polarisation behaviour. Analogously, each “abbreviated Porto notation” giving non-identical polarisation directions in brackets represents five principally possible scattering geometries (for instance, b(ab)a, b(ab)c, c(ab)a, c(ab)c and forumla will yield the same polarisational information and can be summarised by “ab polarisation”). There exists a total of 48 principal permutations of the full Porto notation. Since, for instance, scattering experiments under ab and ba geometry will give the same results, not more than six combinations (aa, ab, ac, bb, bc, cc) are necessary to obtain the full information on the polarisation behaviour of Raman bands.

In the following, we will show briefly with the example of fluorapatite how the directional dependence of vibrations and Raman bands affects the obtained spectra. This mineral (Ca5(PO4)3F; Z = 2; hexagonal space group P63/m) has 33 Raman-active in addition to 20 IR-active modes according to  

formula
(e.g.Kravitz et al., 1968; Boyer & Fleury, 1974; Devarajan & Klee, 1981). The polarisation and directional dependence of Raman-active modes is usually described by the Raman tensors (Loudon, 1964). For the Raman-active vibrations of fluorapatite, the following three tensors are applicable:  
formula
(cf.Rousseau et al., 1981). It follows that with zz polarisation, only A modes are obtained. The xz and yz symmetries would be the best candidates for the analysis of E1 modes, because A and E2 modes are then forbidden. Finally, E2 modes are separately analysed under xy geometry. For the simple fingerprinting analysis (such as the study of accessory apatite grains in rock thin sections done under backscattering geometry) it follows, for instance, that (independent of the polarisation) E2 modes can best be observed if the apatite crystal is cut perpendicular to the crystallographic c axis: both the incident and the registered Raman light paths would then be oriented parallel to c and, therefore, no z polarisation is possible.

For the same experimental setup, it is analogously concluded for the 12 Raman-active modes in zircon (ZrSiO4, Z = 4, tetragonal space group I41/amd) that, according to  

formula
(cf. Loudon, 1964; Dawson et al., 1971), all B-type Raman bands are not seen when the laser light is polarised parallel to the crystallographic c axis. This is in particular true for the strong B1g-type vibration [v3(SiO4) mode] at ∼ 1008 cm-1 (cf. Fig. 6a), which is routinely analysed to estimate the degree of radiation damage of zircon (Nasdala et al., 1995; Nasdala et al., 2003a, and references therein). For such purposes, it is therefore advantageous to orient mounts in such a way that the laser light polarisation is perpendicular to the crystallographic c axis of the zircon crystal to be analysed, in order to get maximum intensity of the B1g mode. For the detailed band assignment procedure on the basis of oriented Raman spectra see, for example, Boyer & Fleury (1974) and Iqbal et al. (1977) for apatite-group minerals, and Dawson et al. (1971), Syme et al. (1977) and Kolesov et al. (2001) for zircon.

It is also important to appreciate the consequences of changes in relative band intensities due only to the “optical trajectory orientation effect” (Smith, 1996) due to the angular dependence of reflection efficiencies of mirrors and other optical components in each Raman spectrometer. For example if a crystal is placed vertically under a microscope and analysed with forumla geometry and then analysed again with the crystal rotated by 90° about the vertical axis (Y) and the incident laser polarisation also rotated 90° by a half-wave plate, then the Porto notation will still be the same (as X and Z have been interchanged twice concerning the laser and crystal relationships) but the spectra will be somewhat different in relative intensities as X vibrations will be preferentially diminished by the spectrometer with one crystal orientation but Z vibrations with the other (as X and Z have been interchanged only once concerning crystal and spectrometer relationships).

Another related point is that the entrance slit to many Raman spectrometers is parallel (after various mirror reflections) to the polarisation direction of the incident laser; when this is not the case (e.g. the DILOR® model XY®) and when the Raman scattered light does not pass through a polariser, the normally intense A1g band of garnet is significantly diminished in intensity relative to other bands (Smith, 1996). Note that garnet, although cubic and hence optically isotropic, is nevertheless orientation-dependent with Raman spectroscopy because of its space group (which, for example, lacks a normal tetrad axis).

In regard to the above examples, we would thus like to emphasise that the observation of significant variations in relative intensities of Raman bands may simply be due to the crystal orientation. The absence of a certain Raman band in a spectrum does not necessarily mean the vibration does not exist. For the simple fingerprinting analysis of unknowns it is, therefore, often advantageous to choose truly random sample orientations containing vector components of all possible directions (for instance, the main axis of prismatic crystals should neither be oriented parallel nor perpendicular to the laser polarisation). Alternatively, after one spectral acquisition, a half-wave plate may be used to rotate the incident laser polarisation by 90° and a new acquisition made immediately without changing any other parameter (cf. Smith, 1996); this clearly displays the degree of the orientation dependency (for the specific crystal vs. laser orientations employed) and adding the two spectra gives a very first approximation to a spectrum which is “average” in terms of polarisation. For more details on the theory of the Raman effect and aspects related to the band assignment and interpretation of spectra, the reader is referred to the literature on this subject (e.g. Wilson et al., 1955; Weeny, 1963; Fateley et al., 1972; Wooster, 1973; Long, 1977; Hayes & Loudon, 1978; Marfunin, 1979; Orlov et al., 1985; McMillan, 1985; Griffith, 1987; McMillan & Hofmeister, 1988; Weidlein et al., 1988; Gardiner & Graves, 1989; Kuzmany, 1989; Schmidt, 1994; Marfunin, 1995; Lewis & Edwards, 2001; Loudon, 2001).

Potential analytical artefacts

In addition to varying band intensity ratios as controlled by the geometry of the scattering experiment, a certain mineral sample may yield quite different spectra when being excited with a laser beam due to analytical artefacts. In the following, we will discuss only two of these problem cases as examples. We do this aiming to direct the reader's attention to the necessity of a critical evaluation of obtained spectra.

A common problem of Raman spectroscopy, especially in the analysis of natural minerals, is the emission of photoluminescence (PL) excited by the laser beam. We have discussed above that the Raman scattering process has comparably little probability and, thus, Raman bands are commonly low in intensity. The simultaneous emission of luminescence, by contrast, can easily reach much higher intensity and, depending on the sample, it may surpass the intensity of the Raman scattered light by several orders of magnitude. If this should be the case, the Raman spectrum is likely to be fully obscured by the much stronger luminescence signal. An example is the strong red ruby luminescence (e.g. Nasdala et al., 2004b, in this volume), whose occurrence makes it almost impossible to obtain the Raman spectrum of ruby with red laser excitation. However, it is often possible, especially if the luminescence emission consists of narrow bands, to find a suitable excitation wavelength with which the Raman spectrum lies in a spectral range that is not affected by the luminescence signal (Fig. 7a). For example, natural apatite may show broad-band luminescence in the range 475-660 nm, which can be avoided by choosing an excitation wavelength either below (e.g. Ar+ 457.9 nm) or above (e.g. Kr+ 676.4 nm) this range (see Nasdala, 1992). Analogously, the Raman spectrum of ruby is obtained without difficulty with blue excitation.

Fig. 7.

Analytical artefacts I: Potential effects of laser-induced PL on the Raman spectrum. (a) Photoluminescence bands (marked “PL”) have certain frequencies independent of the excitation whereas Raman bands (marked “R”) have certain shifts with respect to the excitation band. It is, therefore, possible to avoid PL bands in the obtained Raman spectrum by selecting a suitable excitation wavelength. (b) Example Raman spectra taken from the metamict zircon sample #1486 (Sri Lanka; sample courtesy of R.C. Ewing and C.S. Palenik). The upper Raman spectrum, which was obtained with 514.5 nm excitation, is obscured by narrow REE emission bands (marked with asterisks). This analytical artefact may corrupt the interpretation because narrow PL bands are easily mistaken and erroneously interpreted as Raman bands of additional phases. The lower spectrum, obtained with 632.8 nm excitation, shows only the band pattern of ZrO2. This observation reveals decomposition of radiation-damaged ZrSiO4 into crystalline ZrO2 and amorphous SiO2 (the Raman spectrum of the latter phase is not seen due to its low intensity), which can be caused by heat treatment of metamict zircon at ∼ 900-1100 °C (compare Nasdala et al., 2002).

Fig. 7.

Analytical artefacts I: Potential effects of laser-induced PL on the Raman spectrum. (a) Photoluminescence bands (marked “PL”) have certain frequencies independent of the excitation whereas Raman bands (marked “R”) have certain shifts with respect to the excitation band. It is, therefore, possible to avoid PL bands in the obtained Raman spectrum by selecting a suitable excitation wavelength. (b) Example Raman spectra taken from the metamict zircon sample #1486 (Sri Lanka; sample courtesy of R.C. Ewing and C.S. Palenik). The upper Raman spectrum, which was obtained with 514.5 nm excitation, is obscured by narrow REE emission bands (marked with asterisks). This analytical artefact may corrupt the interpretation because narrow PL bands are easily mistaken and erroneously interpreted as Raman bands of additional phases. The lower spectrum, obtained with 632.8 nm excitation, shows only the band pattern of ZrO2. This observation reveals decomposition of radiation-damaged ZrSiO4 into crystalline ZrO2 and amorphous SiO2 (the Raman spectrum of the latter phase is not seen due to its low intensity), which can be caused by heat treatment of metamict zircon at ∼ 900-1100 °C (compare Nasdala et al., 2002).

If the laser-induced luminescence has only moderate intensity, it may still interfere with the Raman spectrum and affect band fitting and interpretation. Luminescence is often characterised by broad-band emissions with FWHMs ≥ 50 nm (compare Nasdala et al., 2004b, in this volume), which converts to several thousand cm-1. Because such luminescence bands are much broader than Raman bands, it is easily possible to correct Raman spectra for the broad-band luminescence background. In the case of narrow luminescence emissions, however, it may be difficult to distinguish between Raman and luminescence bands. For instance, laser-induced photoluminescence emissions of 4felectron elements can have narrow FWHMs on the order of 0.1 nm. This converts to ∼ 3 cm-1 for bands in the visible range, which would be a typical value for FWHMs of Raman bands. Narrow PL emissions are, therefore, often mistaken as Raman bands such that it is not possible to simply assign narrow bands in the obtained spectra to Raman modes and broad bands to luminescence emissions. Reliable distinction is only possible by obtaining multiple Raman spectra with different excitation wavelengths. Raman bands will appear with equal Raman shifts in all of the spectra (except for some semi-metals like gallium and the D and D' bands in disordered graphite). Luminescence emissions, by contrast, are characterised by a certain fixed wavelength and absolute wavenumber, and must show different apparent Raman shifts in spectra obtained with different excitation wavelengths. Thus, the unwanted interference of Raman scattered light and luminescence emission can be avoided by choosing a suitable excitation wavelength. An example of this is shown in Figure 7b.

Another common problem includes effects of local temperature increase due to strong light absorption and, connected with that, alteration or decomposition of the sample. This problem occurs in particular in the Raman microprobe analysis. Recall that focusing a nominally “weak” laser beam of only 1 mW power to a focal-spot area 1 μm in diameter results in a large power density of > 1000 W/mm2. It is true that such power density is in general not problematic when transparent minerals are analysed, except for powders where it may be necessary to place the mineral grains under water to conduct away some of the heat (cf. Smith et al., 1999a). To avoid any misinterpretation of data, however, the experimentalist needs to check carefully for potential effects of local temperature increase in the sample, which may result in notable temperature-induced shifts of Raman bands towards lower wavenumbers due to longer bond distances (an example was reported by Balan et al., 2001). More serious are cases in which significant light absorption results in irreversible damage of the sample. Effects such as local dewatering, degassing or complete decomposition (often recognised from colour changes or the appearance of little “burned” holes at the sample surface) are especially likely to occur in richly coloured or non-transparent minerals and are, therefore, a common problem in the analysis of mineral pigments. An example for permanent sample damage caused by significant laser light absorption is presented in Figure 8.

Fig. 8.

Analytical artefacts II: Local heating of the sample due to strong light absorption may result in local decomposition. A crystal of magnetite, forumla, from Callenberg, Germany, excited with a fully focused but considerably weakened laser beam (632.8 nm), gave the expected magnetite Raman spectrum. When irradiated with higher laser energy, the very same micro-area underwent local oxidation, as indicated by a Raman pattern reminiscent of forumla (compare the reference spectrum of haematite from Elba, Italy).

Fig. 8.

Analytical artefacts II: Local heating of the sample due to strong light absorption may result in local decomposition. A crystal of magnetite, forumla, from Callenberg, Germany, excited with a fully focused but considerably weakened laser beam (632.8 nm), gave the expected magnetite Raman spectrum. When irradiated with higher laser energy, the very same micro-area underwent local oxidation, as indicated by a Raman pattern reminiscent of forumla (compare the reference spectrum of haematite from Elba, Italy).

There are many experimental ways to avoid such unwanted effects and, with that, any misinterpretation of results. Sample alteration and disintegration due to strong light absorption can simply be avoided by decreasing the laser power. For instance, Nasdala et al. (2001b) found dark brownish radiation haloes to be more sensitive to light absorption than their host biotite, and they did Raman microprobe analyses using fully focused visible laser beams with only 0.03 and 0.08 mW power. The significant weakening of the exciting laser beam, however, is often connected with experimental difficulties. As a rule of thumb, if the laser power is decreased to 1% of the initial value, the scattered light needs to be accumulated for a 100× longer time period to receive about the same signal intensity to background noise ratio. However, considering the unavoidable statistical variation of the intensities of the signal and of the background, and also the rule relating the standard deviation to the square root of the intensity of a single measurement (i.e., improving the ratio of an intensity to its standard deviation by ×10 requires a × 100 increase in the time of a single acquisition), for a given fixed total measurement time it is clearly better to maximise the single measurement time and minimise the number of repeated accumulations. It seems, therefore, more worthwhile to avoid strong light absorption by choosing an excitation wavelength that is less absorbed by the sample (however, the use of another wavelength may not always be possible, for example, because of the occurrence of luminescence phenomena). An excellent example is the Raman analysis of amorphous selenium, which is extremely sensitive to light absorption in the visible range. It was found by A.K. Bandyopadhyay and L. Nasdala (unpublished results) that even the irradiation of amorphous selenium with an extremely weak blue laser beam (Ar+ 488 nm, 0.0013 mW, focal spot diameter 10 μm) led to temperature-induced alteration, including local transformation to the crystalline state. In contrast, the same sample did not show any alteration and yielded the Raman spectrum of amorphous selenium (compare Lucovsky et al., 1967) when excited with an NIR laser beam (Ti-sapphire 890 nm, 5 mW, focal spot diameter 2.5 μm), even though the effective laser power density at the sample surface was almost five orders of magnitude higher in the latter case.

There are other, and more advanced, possibilities to avoid temperature-induced effects, such as the application of cooling stages and sample spinning techniques. These techniques are connected with much greater experimental effort and are, therefore, mostly not considered for routine Raman analysis of geological samples. The main point we aim to make in this subchapter is, however, that the successful avoidance of any misinterpretation of artefacts presumes first of all that the artefact is recognised by the analyst, which in turn always requires a critical checking of samples and spectra.

Instrumentation for Raman analysis

Generalities on Raman systems

In this and the following subchapters, we will refer briefly to the main technical aspects related to the Raman spectroscopic analysis, but only as far as concerning the routine analysis of geological samples. More details are provided in the literature listed above, and especially via the internet (see the web pages of the leading spectrometer manufacturers).

The main components of each Raman system are (1) the source of the incident beam of light, (2) optical components used to illuminate the sample and to collect the Raman scattered light, (3) components for the spectral analysis of the light, and (4) a device for the detection of the light. Raman systems are subdivided into two principal types according to the spectral analysis of the Raman light, namely, Fourier-transform (FT) systems using an interferometer, and dispersive systems. Dispersive systems are usually assigned to two major subgroups according to the way the Raman spectrum is separated from the much stronger Rayleigh line. This can be done either only dispersively using conventional gratings (double and triple monochromator systems) or using a holographic Rayleigh rejection filter (also called notch filter); the latter is usually combined with a single grating monochromator, or an “échelle” grating.

Fourier-transform systems are rarely used in geoscience research. These systems are reasonably priced and can be interfaced to an FTIR spectrometer. Another advantage is that with 1064 nm excitation, luminescence problems rarely occur. Disadvantages include the comparably poor volume resolution even if the spectrometer is coupled with a powerful IR microscope (no confocality possible, see below), the low spectral resolution and the fact that only two excitation wavelengths are available. Also, due to its λ-4 dependence, the Raman effect is generally weak in the near-IR (the Raman scattering intensity is about 23 times higher with Ar+ 488 nm excitation compared with Nd:YAG 1064 nm excitation). This needs to be compensated for by relatively high excitation energies in the NIR.

Dispersive double and triple monochromator systems reach the best spectral resolution (better than ∼ 0.05 cm-1), and they provide the option to analyse bands close to the Rayleigh line (bands with Raman shifts as small as only about 1 cm-1) with the use of a photomultiplier (PMT) detector. The main disadvantage of these systems (in addition to their comparably high price) is the strong light intensity loss along the beam path to the detector. This is of special importance for geoscience research, as experimentalists are here often confronted with poorly scattering and/or light sensitive samples that can only be excited with low laser energy. As a result, Raman light that is particularly low in energy needs to be handled quite often in the analysis of mineralogical and geological samples.

For this reason, notch filter systems seem to be the best choice. In these systems, the light is only dispersed once and reflected by a lower number of mirrors, which results in a much better optical throughput (or light efficiency). Notch systems are also reasonably priced, and their only significant disadvantage (bands with Raman shifts smaller than 50-100 cm-1 are cut off by the notch filter and cannot be analysed) seems of minor relevance for most geoscience research tasks.

Samples are nowadays generally excited using laser light sources. Therefore, it seems somehow unnecessary to describe Raman spectroscopy by the term “laser-Raman spectroscopy” (emphasising the use of a laser source), which can still be found in many papers. Lasers provide a variety of nearly monochromatic excitation wavelengths (the full range is available using a tunable laser) in the UV, visible or near-IR. Unwanted laser emissions, such as the broad-band spontaneous emission of tunable lasers or discharge emissions (plasma lines) of gas ion lasers, are usually rejected by the use of interference filters. Nevertheless, the beam irradiated to the sample is never fully monochromatic, rather the excitation has a certain bandwidth depending on type (and age and adjustment) of the laser.

For quantitative conclusions (especially on the shapes and FWHMs of narrow bands), one needs to consider both the intrumental broadening of Raman bands (described by the so-called “apparatus function” of the Raman system) and the spectral resolution with which the analysis was done. Any detected Raman signal is broadened due to experimental limitations including the quality (sharpness) of the excitation source, the spectral analysis (groove density on the grating used to disperse the light, focal length, widths of internal slits), and the physical resolution of the detector. Most of the modern systems used in geoscience research are equipped with charge-coupled device (CCD) detectors in favour of other older detector types (e.g. diode arrays, PMTs). Here, the physical resolution of the detector is controlled by the pixel density (in cm-1). As a rule of thumb, the actual spectral resolution of a Raman system is generally about three times the physical resolution. Since obtained bands are always broader than the actual Raman signal, real FWHMs need to be calculated by correcting measured FWHMs for the apparatus function and other factors mentioned above. Band correction procedures have been described in detail, for example, by Irmer (1985) and Verma et al. (1995).

Confocality and the Raman microprobe

One of the most important aspects in the application of Raman spectroscopy to the study of geological samples is the opportunity to perform analyses with high lateral resolution on a micro-scale (i.e., “microprobe” analyses). The first Raman microprobes were constructed in the mid-1970s in France (e.g. the “MOLE” by Delhaye & Dhamelincourt, 1975; see also Dhamelincourt & Bisson, 1977; Dhamelincourt et al., 1979; Dhamelincourt, 1987) and in the USA (Rosasco et al., 1975a, 1975b).

The effective lateral resolution achieved by modern Raman microprobe systems is about 1-1.5 μm (Markwort et al., 1995; Nasdala et al., 1996), i.e., roughly twice the excitation wavelength. Note, however, that coupling of a Raman spectrometer (or interferometer) with a powerful optical microscope is a prerequisite, but clearly not sufficient, to create a genuine Raman microprobe. The highest lateral and volume resolution is only possible with a confocal arrangement of the optical pathway (Fig. 9). Due to the usually high numerical aperture of highly magnifying microscope objectives, the beam that is irradiated and focused on to the sample is significantly convergent and diverges again behind its focal plane. As a consequence, the excited sample volume has the shape of an “hour glass” or, if the focal plane is adjusted to the sample surface, the shape of a truncated cone. Even if the top of such a truncated cone (corresponding to the focal plane) has a diameter of only 1 μm, its base must be much broader, and it is rather the diameter of the latter that determines the true lateral resolution of the measurement. The spatial resolution is improved by effectively cutting off the truncated cone at its broad base, i.e., only light scattered from areas in, or slightly above and below, the focal plane is allowed to get to the detector, whereas the rest is rejected by a narrow confocal diaphragm (or confocal hole; Fig. 9). Thus, both depth resolution and lateral resolution are improved simultaneously. Modern confocal systems can operate at a depth resolution of about ± 2 μm. When focussing at the sample surface, this results in an effective volume resolution of ≤ 5 μm3 for transparent samples (e.g. Markwort et al., 1995).

The detection limit for many minerals and related phases is much smaller than the volume resolution in the confocal mode. A sample volume of 5 μm3 corresponds to ∼ 10-11 g, which is analysed with ease even if these 5 μm3 should contain several phases. Even much smaller amounts of sample can be analysed in some cases (e.g. analysis of nanotubes). To give an example from the area of routine geoscience analysis, the thin carbon coat on the surface of an electron microprobe mount is easily detected if only one square micrometer is excited by the laser beam (this corresponds to an amount of sample on the order of 10-14 g). This ability has made Raman spectroscopy a powerful tool for the analysis of small objects (powder particles, inclusions, aerosols, tiny samples in diamond anvil cells etc.) and heterogeneities in minerals (e.g. separate analysis of zones, reaction rims, coats/thin films).

Fig. 9.

Principle of confocal Raman measurements. High depth resolution is achieved by cutting off the light that is scattered from points outside the focal plane with a narrow diaphragm.

Fig. 9.

Principle of confocal Raman measurements. High depth resolution is achieved by cutting off the light that is scattered from points outside the focal plane with a narrow diaphragm.

The rejection of a major portion of the Raman scattered light by the confocal diaphragm results in significant intensity loss at the detector. In the case of heterogeneous samples, this disadvantage is more than compensated for by the improved spatial resolution. Raman bands of the phase to be analysed are now less affected by bands of neighbouring phases or zones (Fig. 10), and in addition luminescence and other unwanted light originating from the surroundings of the analysed volume is significantly suppressed. If homogeneous samples are analysed, however, the intensity loss as caused by a narrow confocal hole is not accompanied by any significant gain in information. Therefore, Raman systems are usually operated in the fully confocal mode only if necessary, whereas finding a compromise between spatial resolution and signal yield seems more efficient for routine analyses.

Fig. 10.

As a simple example for confocality effects, we present two Raman spectra taken from a carbon dioxide inclusion (∼ 12 μm diameter) inside a topaz grain in a polished rock slice of a greisen from Cínovec, Czech Republic. In the regular, i.e., the non-confocal mode (upper spectrum), a complex Raman spectrum is obtained, which consists of vibrational bands of the included CO2 gas, the host mineral and also the araldite epoxy used to attach the polished section to a glass slide (about 20 μm behind the CO2 inclusion). With a narrow confocal diaphragm placed in the optical beam path (lower spectrum), internal CO2 vibrations are clearly obtained whereas topaz bands are largely suppressed and araldite bands are completely suppressed.

Fig. 10.

As a simple example for confocality effects, we present two Raman spectra taken from a carbon dioxide inclusion (∼ 12 μm diameter) inside a topaz grain in a polished rock slice of a greisen from Cínovec, Czech Republic. In the regular, i.e., the non-confocal mode (upper spectrum), a complex Raman spectrum is obtained, which consists of vibrational bands of the included CO2 gas, the host mineral and also the araldite epoxy used to attach the polished section to a glass slide (about 20 μm behind the CO2 inclusion). With a narrow confocal diaphragm placed in the optical beam path (lower spectrum), internal CO2 vibrations are clearly obtained whereas topaz bands are largely suppressed and araldite bands are completely suppressed.

Mobile Raman microscopy for truly in situ analysis

The introduction of a microscope into a Raman spectrometer to create the “Raman microprobe” (RMP) made it possible to analyse individual micrometre-sized crystals inside natural rocks or synthetic mineral assemblages (Dhamelincourt & Bisson, 1977). Subsequently the manufacturers turned to creating another exciting development, mobility, even if the commercial objectives were essentially industrial. It was individual scientists who seized on the opportunity to explore new applications in the geological and archæological sciences. The mobility was achieved by the miniaturisation of many parts of Raman spectrometers which, along with smaller air-cooled laser sources and photon detectors, not to mention portable computers and mobile telephones, made it possible to envisage analysing minerals really in situ by carrying the Raman system to the material rather than transporting a sample to an analytical laboratory.

If this did not sound especially interesting to many geologists who already had a standard “thin section” available to be carried to any laboratory without problem, mobility was especially interesting to archæologists and art historians, along with restorers and curators, as it became possible to analyse their precious objects (or rather parts thereof) not only non-destructively (as with several other modern chemical techniques) but also without displacing the object (which always carries risks of damage or loss). Archæometricians and their colleagues have long been faced with a terrible dilemma: to obtain useful scientific information by deliberately damaging part of a priceless artefact, or not to make any damage whatsoever and hence getting no scientific information at all. For this reason the greater part of all archæological objects have never had their mineral (or molecular) constitution verified, which in many cases left the scientist, historian or general public frustrated by the lack of key information since many artefacts in exhibitions are merely labelled “rock” or some supposed mineral species name. A similar situation applied to gemmologists faced with verifying the mineral species of gemstones mounted in an ornate structure like a crown or an altar, which could not possibly be unmounted for traditional gemmological analysis. With the remarkable combination of non-destructivity and of mobility afforded by Raman analysis, a new era has emerged: “The new age of ‘don’t move it, don't even touch it' archæometry has now arrived to allow remote non-destructive characterisation in all the domains of ARCHÆORAMAN and in situ almost anywhere” (Smith, 2002a).

In order to counteract the terminological confusion caused by the variable use of the term “in situ”, such as describing analysis of individual grains within a multiphase sample which was not prepared in any way, but was nevertheless extracted, perhaps destructively, from its primary source (e.g. a geological locality or an archæological site), or its secondary source (e.g. a museum display or archived collection), and then transported into an analytical laboratory and hence not at all analysed really in situ at its remote site, various workers have employed terms like “remote” analysis, analysis “at distance” or “mobile” analysis. “Mobile Raman Microscopy” (MRM) was carefully defined by Smith (1999) as including analysis by either “portable” systems, which can be carried by one man (e.g. the Kaiser Holoprobe®), or “transportable” systems, which need four men (e.g. the Jobin Yvon LabRAM®). Several kinds of configurations are now commercially available; in general terms the smaller apparatus have a lower spectral resolution or a shorter spectral range, or less features like having only one laser source or no accompanying microscope nor video accessories.

Most manufacturers now provide a variety of configurations of which several can be considered as mobile. The most relevant options are briefly mentioned as follows. A vertical microscope can be fixed to the spectrometer (ideal for traditional point analysis, or for motorised x-y-z 3-dimensional mapping). A horizontal microscope can be fixed to the spectrometer (good for analysing very tall or very heavy objects like statues or other sculptured rocks (Fig. 11a) (e.g. Smith & Bouchard, 2000; Smith, 2000, in print, a). Optical fibres can carry a laser beam from a laser source to a remote head, and also carry the Raman diffused light (and video signal from a TV camera) from the remote head to a spectrometer (Fig. 11b; ideal for analysing anywhere within the perimeter defined by the length of the fibres, currently over 100 m long, and in any orientation) (e.g. Smith, 2001a; Rondeau & Smith, 2002), and furthermore inside complex open structures like a crown. With the spectrometer placed on a cart, extra mobility is achieved being limited only by the length of an ordinary electric power cable. A vertical microscope may be not fixed to the spectrometer and hence be easily portable alone; connected by optical fibres to a portable spectrometer this allows microscopic point analysis or Raman mapping on site. The remote head can have a choice of objectives or longer focal length lenses; fitted to a special tripod, the remote head can be positioned, rotated and shifted in any direction; also it can have a built-in CCD camera allowing observation of the zone under analysis. One or more lasers may be built in, but extra external lasers can also be coupled to the Raman system. With an appropriate configuration, MRM analytical operations can now be achieved in situations unimaginable a decade ago, such as verifying mineral pigments really in situ in paintings on walls or ceilings of ecclesiastical buildings, tombs or caves (e.g. Prehistoric rock art; cf. Smith et al., 1999a, whose early work in this field was on extracted microsamples).

Fig. 11.

Two examples of the non-destructive analysis of art specimens using mobile Raman systems. (a) A Teotihuacan sculptured mask in greenish white-to-grey rock positioned aside the horizontal microscope of a Raman system. All analyses of whitish or greyish parts confirmed the presence of calcite. Museum of Mankind, Paris, 1999. (b) Using a 100 m long optical fibre, a probe head on a tripod was connected with a portable MRM standing in a different room. The remote head was orientated sub-horizontally to send the laser beam (visible here as a tiny green spot) on to the blue-pigmented surface of this 3 m high wooden statue from Oceania on permanent display inside the Louvre Museum, Paris in 2000. This constitutes real in situ “at distance” non-destructive analysis. Photographs taken by D.C. Smith.

Fig. 11.

Two examples of the non-destructive analysis of art specimens using mobile Raman systems. (a) A Teotihuacan sculptured mask in greenish white-to-grey rock positioned aside the horizontal microscope of a Raman system. All analyses of whitish or greyish parts confirmed the presence of calcite. Museum of Mankind, Paris, 1999. (b) Using a 100 m long optical fibre, a probe head on a tripod was connected with a portable MRM standing in a different room. The remote head was orientated sub-horizontally to send the laser beam (visible here as a tiny green spot) on to the blue-pigmented surface of this 3 m high wooden statue from Oceania on permanent display inside the Louvre Museum, Paris in 2000. This constitutes real in situ “at distance” non-destructive analysis. Photographs taken by D.C. Smith.

Remote MRM analysis employing optical fibres (but without a video camera in order to have no electricity at the end of the fibres) with the spectrometer placed in a boat and the tripod set up under water by a diver or a robot has been proposed as a novel way of conducting subaquatic archæometry (Smith, 2003) at depths not exceeding a few hundred metres (which covers a great number of buried cities and sunken ships); this is a very different approach from that of sending an entire MRM system to depth inside some kind of watertight vessel (indeed a special submarine) which is rather similar to sending another kind of special MRM apparatus into Space. Such configurations are already being designed by other research groups (Wang et al., 1996; Wang & Haskin, 2000; Brewer et al., 2002; Wang et al., 2003), and a remarkable telescope system has recently been developed for analysing at distances up to 60 m away from the remote head (Sharma et al., 2002, 2003).

Images based on Raman scattered light

Raman imaging

There are two principally different ways to generate an image from Raman scattered light. The first of these two ways is generally known as Raman imaging, direct imaging, or global imaging technique. Here, the CCD detector of the spectrometer system is used as a camera that photographs a rectangular area of the sample. Imaged areas are mostly several tens of μm across; their sizes depend on the spectrometer optics (especially the magnification of the objective). It is clear that, since a certain area is to be imaged in the Raman mode, this whole sample area needs to be excited, which is normally done by illuminating the sample with a largely defocused laser beam. To avoid the contribution of all of the scattered light (consisting of Rayleigh and polychromatic Raman light) to the obtained image, a notch filter and a band-pass filter are placed in the optical path. The latter allows only the light in a small wavenumber range (“spectral window”) to pass through whereas the rest is discarded. Finally, only a small portion of the Raman light scattered from the excited sample area reaches the CCD and contributes to the image. Consequently, Raman imaging is widely similar to regular photography, with the main difference being that only a particular colour is used: each pixel of the CCD detects the integral light intensity in the pre-set, narrow spectral window that was scattered from a certain micro-area of the sample.

Using angle-tunable band-pass filters, the spectral window can be adjusted to correspond to different Raman bands of the sample. It is, therefore, possible to obtain multiple images of a certain sample area for the Raman signal of different phases occurring in that area. Often some image correction is applied, for instance, to correct for non-uniform laser intensity distribution over the imaged sample area and other instrumental factors. The final images will yield intensity distribution patterns of Raman light in a certain wavenumber range (and, with that, potentially a certain mineral phase). Examples are shown in Figures 12c and d.

The greatest advantage of the direct imaging technique is that imaging is fast. To obtain an image, it takes only about as long as it takes to record a single Raman spectrum, which is mostly on the order of seconds. The disadvantage of the direct imaging technique is the limitation of spectral information. It is clear that the spectral window of an image is fixed once the image has been taken, and the only non-uniform spectral property left is then intensity assigned to lateral coordinates, whereas other spectral information (band FWHMs, band asymmetries, background slopes etc.) is lost. Consequently, it may become difficult to assign the obtained signal to a certain phase if two minerals have a Raman line within the pre-set spectral window, and it may also be difficult to distinguish between Raman and luminescence light. Raman imaging is, therefore, mostly applied to obtain quickly qualitative or semi-quantitative information on low-luminescent mineral samples. This technique is used much more intensely in scientific and industrial disciplines outside the geosciences, such as materials science (e.g. homogeneity check of semiconductors), pharmacy (for a discussion see Bugay, 2001) and in the forensic analysis (e.g. quick detection of drugs covered by sugar). For more details on the imaging technique see, for example, Lehnert (2000).

Fig. 12.

Three examples of Raman-based image generation. (a) Photomicrograph of a coesite inclusion that was partially transformed to α-quartz in garnet from a high-pressure gneiss (Saidenbach, Saxonian Erzgebirge, Germany). The light is not fully cross-polarised, to make visible the internal sub-radial fracture pattern of the host garnet (caused by the volume expansion upon coesite → α-quartz transformation). Sample courtesy of H.-J. Massonne. (b) Raman spectra of coesite and α-quartz. The two SiO2 polymorphs are unambiguously distinguished from their typical fingerprint patterns of Raman bands. Intensity distribution images for the 521 cm-1 coesite band (red) in (c) and the 464 cm-1 quartz band (blue) in (d) revealed that most of the inclusion still consists of coesite, with the transformation to α-quartz having started from the outer rim and internal fractures of the inclusion. (e) Raman spectrum of moganite-bearing agate from St. Egidien, Germany (solid graph; sample courtesy of J. Götze). The most striking difference between the obtained spectrum and that of moganite-free quartz (dotted) is the additional appearance of an intense band at 502 cm-1 (compare Kingma & Hemley, 1994; Götze et al., 1998). (f) Raman map generated from the band integral ratio of the 502 and 464 cm-1 bands, indicating rhythmic growth zoning. Bright areas (high moganite content) are assigned to quickly grown, microcrystalline zones. Dark areas consist mainly of coarse-grained α-quartz. (g) Raman map of a gem-quality feldspar with “moonstone” cat's eye effect from Tanzania, revealing a perthitic internal texture (for sample description see Milisenda et al., in print). It is possible to distinguish between orthoclase and albite because the Raman shift of the main feldspar band varies slightly with the chemical composition (513 cm-1, orthoclase; 509 cm-1, albite) (compare Mernagh, 1991). Even though the Raman bands of the two feldspar minerals overlap widely one another, albite lamellae in an orthoclase matrix are clearly resolved using the ratio of signal intensities at 513 cm-1 and 509 cm-1 Raman shift.

Fig. 12.

Three examples of Raman-based image generation. (a) Photomicrograph of a coesite inclusion that was partially transformed to α-quartz in garnet from a high-pressure gneiss (Saidenbach, Saxonian Erzgebirge, Germany). The light is not fully cross-polarised, to make visible the internal sub-radial fracture pattern of the host garnet (caused by the volume expansion upon coesite → α-quartz transformation). Sample courtesy of H.-J. Massonne. (b) Raman spectra of coesite and α-quartz. The two SiO2 polymorphs are unambiguously distinguished from their typical fingerprint patterns of Raman bands. Intensity distribution images for the 521 cm-1 coesite band (red) in (c) and the 464 cm-1 quartz band (blue) in (d) revealed that most of the inclusion still consists of coesite, with the transformation to α-quartz having started from the outer rim and internal fractures of the inclusion. (e) Raman spectrum of moganite-bearing agate from St. Egidien, Germany (solid graph; sample courtesy of J. Götze). The most striking difference between the obtained spectrum and that of moganite-free quartz (dotted) is the additional appearance of an intense band at 502 cm-1 (compare Kingma & Hemley, 1994; Götze et al., 1998). (f) Raman map generated from the band integral ratio of the 502 and 464 cm-1 bands, indicating rhythmic growth zoning. Bright areas (high moganite content) are assigned to quickly grown, microcrystalline zones. Dark areas consist mainly of coarse-grained α-quartz. (g) Raman map of a gem-quality feldspar with “moonstone” cat's eye effect from Tanzania, revealing a perthitic internal texture (for sample description see Milisenda et al., in print). It is possible to distinguish between orthoclase and albite because the Raman shift of the main feldspar band varies slightly with the chemical composition (513 cm-1, orthoclase; 509 cm-1, albite) (compare Mernagh, 1991). Even though the Raman bands of the two feldspar minerals overlap widely one another, albite lamellae in an orthoclase matrix are clearly resolved using the ratio of signal intensities at 513 cm-1 and 509 cm-1 Raman shift.

Raman mapping

The second basic way to generate an image from Raman scattered light is the Raman mapping technique. A map (i.e., a colour-coded image) is rather a mathematical product, generated on the basis of a large number (typically 1000–50000) of single spectra. Here, the Raman system is operated in the confocal mode, and a full Raman spectrum is obtained for each pixel of the image to be generated. This is mostly done using software-controlled x-y stages, i.e., the sample is moved step-by-step relative to the fixed microscope objective. The step width can be chosen by the experimentalist. It is mostly adjusted depending on the size of the mapped area, and in order to get a meaningful compromise between sufficient lateral resolution and size of the resulting data file. The whole data set is then processed (e.g. background correction and band fitting/deconvolution for all spectra), which results in a complex data array. The full spectral information for each Raman band is available for every x-y coordinate of the mapped area. Finally, multiple colour-coded maps can be generated for any parameter. An example is presented in Figure 12f. Since single spectra are obtained in the confocal mode, maps with an excellent depth resolution may be generated (compare Fig. 19 below), and it is even possible to generate tomographic images of two-dimensional planes inside minerals (Nasdala et al., 2003b) (see Fig. 17 below and the cover figure of this volume). The main disadvantage of Raman maps is that obtaining a large number of single spectra in succession may be extremely time-consuming. For instance, if a single spectrum is recorded in only 5 seconds, a map consisting of 150 × 150 pixels/spectra will require a total of more than 31 hours of laboratory time. To decrease the experimental time, some companies work on the development of line scanning techniques (a line of spectra is simultaneously obtained and an area is then mapped line-by-line). Raman mapping is, therefore, unsuitable for quick homogeneity checks. However, the wealth of detailed spectral information that is available has made Raman mapping an extremely valuable technique for detailed studies of internally heterogeneous minerals.

Applications of Raman spectroscopy

Generalities on applications in mineralogy and geology

Raman spectroscopy has been used successfully in nearly all geoscience disciplines and virtually all kinds of samples have been studied using this technique. The application of Raman spectroscopy seems especially influenced by its experimental advantages (see below). The simple identification of tiny particles of minerals, or inclusions in minerals, and related substances is widely applied. This is mostly done in cases where the more common techniques (e.g. electron microprobe or X-ray diffraction analysis) cannot be used, for example, because of the impossibility to separate or prepare the sample to be studied. The identification of rock-forming or accessory minerals is based on the numerous papers describing the Raman spectra of these mineral species (these papers can be found in the relevant literature databases; example references are not cited here). One obvious advantage of Raman spectroscopy is that polymorphs with the same chemical composition can be easily distinguished (e.g. Etchepare et al., 1974, 1978; Sharma & Simons, 1981; Mernagh, 1991; Rodgers, 1993). Chemical information on the composition of minerals and solid solution members can be obtained (e.g. Mernagh, 1991; Wopenka et al., 1999; Kreisel et al., 2000; Wang et al., 2001), and it is also possible to study the isotopic composition of minerals and included phases (e.g. Sato & McMillan, 1987; Champagnon et al., 1997; Irmer & Graupner, 2002).

An important field for the use of Raman analysis is based on its sensitivity to short-range order. Melts, quenched melts, glasses and other amorphous or amorphised materials have been extensively studied (e.g. Sharma, 1972; Brawer & White, 1974, 1976, 1978; Furukawa & White, 1978, 1981; Furukawa et al., 1978, 1981; Videau et al., 1981; Piriou et al., 1981; McMillan et al., 1982; McMillan & Piriou, 1983; McMillan, 1984, Minser et al., 1984; Matson & Sharma, 1985; Mukherjee & Sharma, 1985; Mysen & Virgo, 1985; Sprenger et al., 1993; Mysen, 1995; Henderson & Fleet, 1995; Frantz & Mysen, 1995; Alberto et al., 1995; Sprenger, 1996), for instance to obtain information on the coordination of ions in such “X-ray amorphous” samples. Ions in solutions have been studied successfully as well (e.g. Sharma & Reed, 1976; Piriou & Svoronos, 1985; Bondarenko & Gorbaty, 1999), which makes Raman spectroscopy a valuable tool in geochemical investigations. More recently, Raman analyses have also been applied to the study of minerals that were fully or partially amorphised due to the impact of radioactivity, as for instance radiation-damaged zircon (Nasdala, 1995), monazite (Seydoux-Guillaume et al., 2002) and biotite (Nasdala et al., 2001b), as well as spent nuclear fuel including its alteration and corrosion products (Amme et al., 2002). Here, Raman spectra provide information on the present substances and their degree of short-range order and crystallinity, respectively. Similar structural information is obtained in the study of order-disorder phenomena in minerals (Keramidas et al., 1975; McMillan et al., 1984; Bischoff et al., 1985; Tlili et al., 1989).

There are numerous applications of Raman spectroscopy in biomineralogy, as for instance the study of fossils, corals, nacre, human bones and dental enamel, and coatings applied to medical implants (e.g. Vénec-Péyré & Jaeschke-Boyer, 1979; Daudon et al., 1981; Nelson & Williamson, 1982; Urmos et al., 1991; Silvé et al., 1992; Pasteris et al., 1999; Dietrich et al., 2001; Freeman et al., 2001; Miyazaki et al., 2002; Perrin & Smith, 2002; Silva et al., 2003; Taddei et al., 2003; Martini et al., 2003; Balz et al., submitted). An example is presented in Figure 13. Raman spectroscopy has also been used to investigate environmental processes such as weathering and corrosion (e.g. Refait et al., 2003).

In crystallography and materials science, Raman spectroscopy is routinely used to identify and check the quality and homogeneity of synthetic growth products (e.g. Boyer et al., 1985; Sweegers et al., 2001). Because of the opportunity to perform analyses non-destructively, Raman has become an extremely valuable tool in the study of gemstones, which includes their identification and the identification of inclusions, and the detection of potential treatments done to enhance colour and clarity (e.g. Délé-Dubois et al., 1980, 1986a, 1986b; Maestrati, 1989; Schmetzer et al., 1996, 1997; Nassau et al., 1997; Krzemnicki, 1999; Ostroumov et al., 1999; Chalain et al., 1999, 2000); overviews are given by Coupry & Brissaud (1996), Kiefert et al. (2001) and Smith (in print, a). Finally, the generally increased acceptance and use of Raman analysis is also documented by the fact that this technique is more and more applied in the description of new mineral species or their redefinition (for instance Grice et al., 1986; Nasdala et al., 1993, 1998; Holtstam, 1997; Bühn et al., 1999; Brugger et al., 1999; Birch et al., 2001; Witzke et al., 2001; Wallwork et al., 2002; Kolitsch, 2003, Krause et al., 2003).

Fig. 13.

Raman micro-spectroscopy applied to biominerals. (a) Scanning electron microscope image of assemblages of synthetic Ca-carbonate minerals, grown at the inner skin of a chicken egg shell. Sample courtesy by M. Balz. (b) Crystals with sizes of down to < 1 μm are easily analysed, as Raman spectra allow the unambiguous differentiation among CaCO3 polymorphs. Note in particular the clearly different low-frequency lattice vibrations of aragonite and calcite. Aragonite and calcite can also be clearly distinguished by the weak 702 & 707 & 717 cm-1 triplet of the former and the 713 cm-1 singlet of the latter, and again by further Raman bands in the 1400–1600 cm-1 spectral range.

Fig. 13.

Raman micro-spectroscopy applied to biominerals. (a) Scanning electron microscope image of assemblages of synthetic Ca-carbonate minerals, grown at the inner skin of a chicken egg shell. Sample courtesy by M. Balz. (b) Crystals with sizes of down to < 1 μm are easily analysed, as Raman spectra allow the unambiguous differentiation among CaCO3 polymorphs. Note in particular the clearly different low-frequency lattice vibrations of aragonite and calcite. Aragonite and calcite can also be clearly distinguished by the weak 702 & 707 & 717 cm-1 triplet of the former and the 713 cm-1 singlet of the latter, and again by further Raman bands in the 1400–1600 cm-1 spectral range.

Applications to the study of inclusions in minerals

Because of its high volume resolution and the ability to measure efficiently major fluid inclusion components such as CO2 and CH4 in situ, Raman micro-spectroscopy has become an important, in some cases the only, tool to determine composition and density of fluid inclusions. A list of Raman-active species in fluid inclusions and some applications are given by Burke (2001). The primary applications in the geosciences are: (1) qualitative identification of gaseous, liquid, supercritical, and solid components of fluid inclusions as a “fingerprint” method; (2) semi-quantitative determination of ratios between two or more gaseous, liquid, or supercritical species inside inclusions (such as CO2-CH4); and (3) estimation of the origin and formation conditions of solid phases inside inclusions.

Qualitative identification of gaseous, liquid, supercritical and solid components of fluid inclusions are some of the major applications that have successfully been applied to the determination of the most frequently occurring polyatomic fluid components using Raman spectroscopy (e.g. 12CO2, CH4 and N2) (Rosasco et al., 1975b; Guilhaumou et al., 1978; Touray et al., 1985; Pasteris et al., 1986; Burke & Lustenhouwer, 1987; van den Kerkhof 1988). In addition, minor and rare components of fluid inclusions have also been investigated (e.g. 13CO2, H2S, SO2, CO, COS, H2, O2 and NH3) (Bény et al., 1982; Touray et al., 1985; Pasteris et al., 1986; Frezzotti et al., 1992; Giuliani et al., 2003). Detection of the important liquid and gaseous H2O components in fluid inclusions is difficult but not impossible (e.g. Wopenka et al., 1990). Major ions in solution, such as those of Na, K, Mg, Fe and Li, are detectable at low temperatures as solid hydrates (Dubessy et al., 1982, 1992; Winter & Roberts, 1993). Detection of hydrocarbons in inclusions is possible with some difficulties (Stephenson, 1974; Guilhaumou, 1982; Guilhaumou et al., 1988; Pironon, 1993; Orange et al., 1996). Raman spectra of clathrates (polymerised open-structured hydrates with cavities incorporating molecules such as CO2, CH4, N2, O2 and H2S that are formed at low temperatures in multicomponent inclusions) have been presented by Sum et al. (1997). Champagnon et al. (1997) distinguished isotopes of N2 and O2 in clathrates in ice cores. Numerous daughter minerals, or accidentally trapped solids, have been identified (e.g. Dhamelincourt et al., 1979; Beny et al., 1982; Andersen et al., 1984, 1989; Cervelle & Moëlo, 1990; Phillipot & Selverstone, 1991; Mernagh & Trudu, 1993; Mernagh & Hoatson, 1995; Burke, 1998; Barrie et al., 1999; Vapnik & Moroz, 2002; Koděra et al., 2003).

Semi-quantitative detection of various gas mixtures within fluid inclusions were carried out in the chemical systems CO2–H2S–H2O–S (Beny et al., 1982), CH4–CO2 and CO2–N2 (Guilhaumou et al., 1982; Darimont et al., 1988; van den Kerkhof, 1988; Frezzotti et al., 1992), in systems involving additional H2O (e.g. Seitz et al., 1987; Leng et al., 1998) and in complex systems including H2O, NaCl and further components (e.g. Pasteris et al., 1986; Dubessy et al., 1989; Thomas et al., 1990; Nwe & Morteani, 1993; Dubessy et al., 1999; Siemann & Ellendorf, 2001; Giuliani et al., 2003). Complementary to microthermometry, the dissolved salt content of aqueous inclusions at room temperature has been determined using Raman analysis (Mernagh & Wilde, 1989). Furthermore, ions in aqueous inclusions such as forumla and HS can be quantified (e.g. Rosasco & Roedder, 1979; Dubessy et al., 1983, 1992; Murata et al., 1997; Benison et al., 1998; Boiron et al., 1999).

Graphite or carbonaceous material in fluid inclusions have been investigated by Reutel (1992), Wopenka & Pasteris (1993), Frezzotti et al. (1994), Andersen & Burke (1996), Cesare & Maineri (1999) and Kaindl et al. (1999). Furthermore, there are numerous studies that have used Raman spectroscopy for the identification and characterisation of solid inclusions in minerals (e.g. Smith, 1984; Liu et al., 1990; Yang et al., 1998; Izraeli et al., 1999; Nasdala & Massonne, 2000; Sobolev et al., 2000; Ye et al., 2001; Kunz et al., 2002; Gillet et al., 2002; Massonne & Nasdala, 2003; Chopin, 2003).

Applications in high-pressure and high-temperature studies

The Raman spectroscopic study of vibrational properties of minerals, glasses, melts and fluids at high pressure and high temperature has important applications in material and geosciences (Ferraro, 1984; Gillet, 1996; Gillet et al., 1998). The development of the diamond anvil cell (DAC) has opened up new possibilities for in situ Raman spectroscopic experiments in minerals at high pressures and at high as well as low temperatures (Jayaraman, 1983, 1986; Ferraro, 1984; Chervin et al., 1992). Raman spectra can be obtained from micrometre-scale specimens compressed by pressures of up to > 135 GPa in the DAC (e.g. over the pressure range of the entire Earth's mantle) (Gillet et al., 1998). The sample chamber of a DAC consists of a hole (usually < 300 μm in diameter and < 150 μm in height) in a gasket, which is pressed between two diamond anvils. This very small volume contains the sample inside a pressure-transmitting medium, and in most cases also an in situ pressure sensor such as a small ruby chip. Ruby can be used for pressure determination by monitoring the frequency of the laser-induced photoluminescence (Piermarini et al., 1975; Mao et al., 1986). Two different techniques are used for heating up samples in the DAC: a small resistance heater for temperatures up to 1200 °C and laser heating using a CO2 laser for temperatures up to 1700 °C in the hot spot of the DAC (Boehler & Chopelas, 1992; Gillet, 1996). Recently the Bassett-type DAC became a powerful tool for hydrothermal studies at simultaneous high pressures up to 2.5 GPa and temperatures from –190 °C to 1200 °C (Bassett et al., 1993; Bassett et al. 1996; Shen et al., 1992; Bassett, 2003). Synthetic moissanite (hexagonal SiC) has been recently proposed to be suitable as a substitute for diamond in anvils in spectroscopic high pressure cells (Xu & Mao, 2000).

Raman spectroscopy mostly in conjuction with the DAC technique has been used to investigate the structure and the high-pressure high-temperature behaviour, including phase transitions, of Earth-interior related minerals (Williams et al., 1992; Reynard et al., 1997; Schmidt & Ziemann, 2000; Shim & Duffy, 2001; Chopelas & Serghiou, 2002; Kleppe et al., 2002, 2003), glasses (Farber & Williams, 1996), and melts and liquids (Frantz et al., 1994; Richet et al., 1996; Williams & Knittle, 2003; Ziemann et al., submitted). Moreover, the study of vibrational properties of minerals as a function of pressure and temperature (based on Raman and IR spectroscopic data) allows the derivation of fundamental properties such as thermal conductivity (Hofmeister, 1999, 2001) and the thermodynamic functions heat capacity, vibrational entropy, internal energy and Helmholtz free energy (Kieffer, 1979; Gillet, 1996; Gillet et al., 1998; Hofmeister & Mao, 2002). A further application is the calibration of spectroscopic pressure sensors, which can be used for in situ experiments at high temperatures beyond the range of other pressure measurement techniques (Schiferl et al., 1997; Schmidt & Ziemann, 2000).

Applications in archaeometry: The Raman Microscope (RM)

The symbol RM (where the M variably refers to Microscope, Microscopy, Microprobe or Microspectroscopy) is becoming popular in archæometry because it emphasises an enormous advantage over most spectroscopic techniques: it is possible to observe a sample under high magnification, to choose the precise microcrystal to be analysed and then to analyse it immediately. Research applying RM to gemstones or pigments in art history, archæology, conservation or restoration, whether to verify a supposed mineral species or to recognise fakes, began in the 1980s (e.g. for gemstones: Délé-Dubois et al., 1980; for pigments: Delhaye et al., 1985). The first significant catalogue of the Raman spectra of gemstones appeared in Pinet et al. (1992) and of pigments in Bell et al. (1997). Only in the late 1990s did RM become introduced as a valuable new non-destructive analytical technique applicable to a wider range of geomaterials (e.g. polished ceremonial eclogite and jade axes: Smith & Gendron, 1997), or biomaterials (e.g. skin, resin, linen: Edwards et al., 1996a, 1996b, 1996c), or inorganic and organic pigments from Prehistoric rock art (e.g. Smith et al., 1999a; Edwards et al., 2000).

The pseudo-acronym “ARCÆORAMAN” was coined by Smith & Edwards (1998) to cover this new sub-discipline; Table 7 in Smith & Carabatos-Nedelec (2001) summarises the early bibliography in this new field. A series of international congresses has promoted interest in this sub-discipline for RM, mostly without the help of any complementary analytical technique (GEORAMAN-1996 in Nantes, -1999 in Valladolid, -2002 in Prague, -2004 in Honolulu; ICORS-1998 in Cape Town, -2004 in Gold Coast, and other meetings in London in 2001 and Ghent in 2003); the present situation is literally an explosion of publications. It is however clear that the majority of past (and present) work has been (and still is) on pigments of one kind or another, and that the majority of researchers in ARCHÆORAMAN are chemists or physicists, with an insufficient number of geologists and biologists involved as there is an evident need to identify natural mineral, animal or vegetable species and to try to determine their geological or biological provenance.

Table 2.

Ten domains of the application of Raman analysis to archaeological samples (after Smith, 2002a).

DomainExample materials

gemsgemstones (rough, cut or mounted), cameos, corals, intaglios, jewellery etc.
ceramicschina, earthenware, faience, glass, porcelain, pottery, slags, tiles etc.
rocksaxe heads, building columns, ceremonial stones, inlaid rock, millstones, mosaics, necklaces, sculptures, vitrified forts etc.
corroded metalscorroded bracelets, coins, cutlery, necklaces, statues, swords, tools etc.
resins s.l.non-cellular organic material composed of only a few different molecules or of amorphous hydrocarbons without a growth texture: amber, bitumen, coal, glue, gum, oil, putty, wax etc.
tissues s.l.cellular organic molecules or biominerals with a growth texture: bone, cotton, feather, fur, hair, horn, ivory, leather, linen, nail, papyrus, parchment, silk, skin, teeth, wool, wood etc.
pigments/inks/dyes on or in an inorganic substratebrick, ceramic, plaster, stone, stucco etc.
pigments/inks/dyes on or in an organic substratebone, canvas, paper, skin, textile, wood etc.
coloured vitreous materialspigments on or in enamel, glass or glaze etc.
climatic deterioration of any of these materialscorrosive agents involved, original, intermediate and final products
DomainExample materials

gemsgemstones (rough, cut or mounted), cameos, corals, intaglios, jewellery etc.
ceramicschina, earthenware, faience, glass, porcelain, pottery, slags, tiles etc.
rocksaxe heads, building columns, ceremonial stones, inlaid rock, millstones, mosaics, necklaces, sculptures, vitrified forts etc.
corroded metalscorroded bracelets, coins, cutlery, necklaces, statues, swords, tools etc.
resins s.l.non-cellular organic material composed of only a few different molecules or of amorphous hydrocarbons without a growth texture: amber, bitumen, coal, glue, gum, oil, putty, wax etc.
tissues s.l.cellular organic molecules or biominerals with a growth texture: bone, cotton, feather, fur, hair, horn, ivory, leather, linen, nail, papyrus, parchment, silk, skin, teeth, wool, wood etc.
pigments/inks/dyes on or in an inorganic substratebrick, ceramic, plaster, stone, stucco etc.
pigments/inks/dyes on or in an organic substratebone, canvas, paper, skin, textile, wood etc.
coloured vitreous materialspigments on or in enamel, glass or glaze etc.
climatic deterioration of any of these materialscorrosive agents involved, original, intermediate and final products

The ten topics classified by Smith (2002a) may be summarised as shown in Table 2. The first four topics and the ninth are those most concerned with minerals and of these the first has by far the greatest number of publications. The advantages for gemstones are considerable as RM can be employed for several different purposes: to verify the nature of the gemstone itself, to examine for treatments (e.g. heating, resin impregnation, pigmentation), to explore solid or fluid microinclusions, or to detect synthetic and imitation stones (e.g. Délé-Dubois et al., 1980; Lasnier 1989; Maestrati, 1989; Pinet et al., 1992; Schmetzer et al., 1996; Smith & Robin, 1997; Hänni et al., 1998; Chalain et al., 1999; Kiefert et al., 2001; Smith, in print, a). Ivory is arguably a gemstone or a tissue; distinguishing real from fake ivory is easy with RM (Brody et al., 1998). Examples of work on rocks were published in Smith & Bouchard (2000) and Smith (in print, b). Ceramics was the last topic to be examined, both from the point of view of the minerals constituting pottery (e.g. Fry et al., 1998) or the pigments in glazes (e.g. Liem et al., 2000; Colomban & Treppoz, 2001). In the meantime new projects were launched to evaluate the potential of RM to corroded metals (e.g. Bouchard & Smith, 1999, 2000, 2001; McCann et al., 1999; Smith & Bouchard, 2002; Frost, 2003) and to stained glass (e.g. Smith et al., 1999b; Bouchard & Smith, 2002). Burgio & Clark (2001) presented an updated catalogue of the Raman spectra of pigments of which many are minerals; new catalogues of the Raman spectra of minerals involved in corroded metals, stained glass or Prehistoric pigments were recently published by Bouchard & Smith (in print).

Selected examples of Raman applications in the Geosciences

Semi-quantitative micro-Raman spectroscopy of a gas inclusion

Fluid inclusions in minerals can be used as indicators of pressure-temperature and fluid composition, provided that chemical composition and density of individual inclusions are well known (e.g. Rankin, 2004). The typically small size (1–10 μm diameter) and low weight (10-9–10-12 g) renders impossible in many cases reliable microscopic observation, this being a prerequisite for the application of standard microthermometric methods. Modern confocal micro-Raman systems reach an effective volume resolution better than 5 μm3 (Markwort et al., 1995; Nasdala et al., 1996) and allow semi-quantitative analysis of inclusions with sizes down to 2 μm diameter (Fricke et al., 1990). Notch filter systems coupled with comparatively low power (40–100 mW) 514.5 or 532 nm excitation laser sources and a CCD detector guarantee high spectral efficiency and resolution (Burke, 2001). Low laser power furthermore reduces the probability of reactions in CO2-CH4 inclusions that could be induced by local laser heating effects (e.g. Huizenga & Touret, 1999). Semi-quantitative estimations of two or more component fluids within inclusions requires knowledge of the so-called Raman scattering “cross sections”, a measure for the Raman activity of a certain component in mixtures (Schrötter & Klöckner, 1979). Empirical calibration of the Raman spectrometer by gas mixtures of known composition and density is nowadays used for the quantification of fluid mixtures (e.g. Wopenka & Pasteris, 1986, 1987; Dubessy et al. 1989; Chou et al., 1990; Seitz et al., 1993, 1996). The reproducibility of an analysis is usually better than 5% (van den Kerkhof & Kisch, 1993).

The following example illustrates a semi-quantitative estimation of the gas composition of a natural CO2-CH4-N2 inclusion (supercritical at room temperature) in quartz (Fig. 14a). The studied sample was taken from an Archean amphibolite-facies lode gold mineralisation in Western Australia (Neumayr et al., 1993). The Raman spectrum of the inclusion was excited with the 514.5 nm emission of an Ar+ laser. It can be seen in Figure 14b that, in the spectral range between 600 and 3400 cm-1, bands of three gas phases have been obtained, in addition to a number of low-intensity bands of the surrounding host quartz. Band fitting, assuming symmetric Gaussian-Lorentzian peak shapes, yielded the parameters given in Table 3. Note that the observed Raman shifts of CO2, N2 and CH4 bands are slightly lowered when compared to the 1 bar values (1285, 1388, 2331 and 2917 cm-1; Burke, 2001), which is taken as evidence for elevated internal gas pressure in the inclusion. The determined Raman band integrals are the basic input parameters to calculate relative molar fractions of the three fluid species. Due to numerous influencing factors it is more accurate to calculate relative molecule ratios in contrast to absolute numbers. The following formula is based on Placzek's polarisability theory (Placzek, 1934; Schrötter & Klöckner, 1979; Dubessy et al., 1989):  

formula

Fig. 14.

Raman spectroscopy applied to gas analysis. (a) Photomicrograph of a gas inclusion in a quartz crystal from a gold deposit in the Pilbara Block, Western Australia (for sample description see Neumayr et al., 1993). (b) Raman spectrum taken from the central inclusion in (a). Bands of CO2, N2 and CH4 were obtained. Raman bands of the host quartz are marked with asterisks.

Fig. 14.

Raman spectroscopy applied to gas analysis. (a) Photomicrograph of a gas inclusion in a quartz crystal from a gold deposit in the Pilbara Block, Western Australia (for sample description see Neumayr et al., 1993). (b) Raman spectrum taken from the central inclusion in (a). Bands of CO2, N2 and CH4 were obtained. Raman bands of the host quartz are marked with asterisks.

Table 3.

Band parameters for four Raman bands obtained from a fluid inclusion in quartz, and calculated molar gas portions.

SpeciesObserved Raman shift [cm–1]Integrated peak area Ai [cts cm–1]Raman cross section σi (514 nm)Instrumental efficiency ζiGas portion [mol%]

CO2(V)12823.51.0120
CO2(V+)13857.11.51
N223271.5117
CH42912116.87.5173
SpeciesObserved Raman shift [cm–1]Integrated peak area Ai [cts cm–1]Raman cross section σi (514 nm)Instrumental efficiency ζiGas portion [mol%]

CO2(V)12823.51.0120
CO2(V+)13857.11.51
N223271.5117
CH42912116.87.5173

The parameters Xa, Aa, σa and ζa are molar fraction, integrated band area, Raman cross-section and instrumental efficiency of species a in a fluid mixture, respectively.

For our sample inclusion, the integrated band areas Ai, Raman cross-sections σi and instrumental efficiency parameters ζi are given in Table 3. The ζi values, an empirical calibrated instrumental parameter using synthetic gas mixtures of known composition and density (van den Kerkhof, 1988), equals 1 for all species and the specific instrument in our example. Insertion of the appropriate parameter for CO2 into Equation 6 leads to  

formula

According to Dubessy et al. (1989), the CO2 component should be calculated using the sum of the two peak areas and the sum of their σ (2.5 in case of the 514.5 nm laser). In the same way the N2 component is determined as 7 mol% and the CH4 component as 73 mol%.

Semi-quantitative chemical analysis by Raman spectroscopy

Semi-quantitative chemical analysis using Raman spectroscopy is usually assumed to involve some kind of calibration of the absolute intensity or surface area of a specific Raman band and its proportionality to the concentration of a particular chemical group, on condition that the variation of spectral intensity with crystal orientation is adequately dealt with (cf. FTIR using finely ground powders in KBr pellets). A completely different approach is discussed here; this uses the wavenumber shifts and not at all the intensities, since the wavenumber shift of a Raman band due to a chemical substitution is independent of crystal orientation, and hence of polarisation effects (except where modes overlap and may lead to confusion), and is a viable indicator of chemical composition. It should be noted that the chemical elements detected may have no direct involvement in the Raman vibration being used. For example wavenumber shifting of the supposed Si–O–Si vibration in inosilicates like pyroxene and internal SiO4 vibrations in nesosilicates like garnet is very dependent upon the cation charge and ionic radius of the cations in nearby octahedral sites which do indeed have an indirect effect upon the bond lengths and angles of the Si–O–Si and SiO4 vibrations respectively; thus one can identify other cations simply by observing the behaviour of Si–O bonds.

Early work on binary systems was made on the natural system jadeite–diopside of the pyroxene group, (Na1–xCax)(Al1–xMgx)Si2O6, by Smith & Boyer (1985) to examine the heterovalent substitution of Na+Al3+ (jadeite: Jd) by Ca2+Mg2+ (diopside: Di), and on the synthetic system spinel–magnesiochromite of the spinel group, Mg(CrxAl2–x)O4, by Malézieux et al. (1983) to examine the homovalent substitution of Al3+ by Cr3+. In both systems the variation of Raman wavenumber with chemical composition was seen to be at least close to linear, if not truly linear. A mathematical and statistical method for determining the composition of an unknown sample in an n-dimensional chemical space using the intersection of (n – 1) different Raman bands, and assuming linear variations throughout, was presented by Smith & Pinet (1989). It was used by Smith et al. (1988) for the ternary garnet system pyrope–almandine–grossular, and by Smith et al. (1989) for the ternary pyroxene system diopside–hedenbergite–aegirine. Of course in a ternary system where only two Raman bands are required (because the constant sum of all considered end-members provides one constraint), the analysis may be done graphically (Fig. 15); for a greater number of end-members in the solid solution, a mathematical treatment by simultaneous equations is necessary.

Fig. 15.

A ternary system of end-members (A, B and C) where the wavenumber isopleths W1 for Raman peak P1 are subhorizontal and the wavenumber isopleths W2 for peak P2 are subvertical. The intersection of the measured wavenumbers of an unknown sample u for these two peaks (W1u and W2u) gives the semi-quantitative analysis in terms of proportions of A, B and C.

Fig. 15.

A ternary system of end-members (A, B and C) where the wavenumber isopleths W1 for Raman peak P1 are subhorizontal and the wavenumber isopleths W2 for peak P2 are subvertical. The intersection of the measured wavenumbers of an unknown sample u for these two peaks (W1u and W2u) gives the semi-quantitative analysis in terms of proportions of A, B and C.

The methodological approach and several related problems were discussed by Smith (2001b, 2002b) where some example Raman analyses were presented using ternary, quaternary, pentary or hexary garnet systems (i.e. including grossular, andradite and uvarovite). For example, a composition of Gro0.53And0.43Uva0.03 was determined using Raman bands I and II (Pinet & Smith, 1993, 1994) in the ternary system Gro–And–Uva whereas Gro0.51And0.46Uva0.00 was determined for the same sample from electron microprobe (EMP) analysis (i.e., an absolute discrepancy of ±3 mol%). A Raman-based analysis (using bands I, II, IV and XII) of a garnet in the pentary system Pyr–Alm–Spe–Gro–And gave Pyr0.01Alm0.20Spe0.53Gro0.26And0.01, whereas the EMP gave Pyr0.04Alm0.27Spe0.43Gro0.20And0.06 (i.e., an absolute discrepancy of –7 to +10 mol%).

Of the eight problems cited in Smith (2001b), the three most important ones are mentioned here. Firstly, the method is based on the wavenumber values of the pure end-member compositions, but these are not always known; hence they have to be estimated statistically in multivariate chemical space or observed from experimental syntheses (which do not always produce well-crystallised on-composition grains). Secondly, the necessity to be sure that band Z of the unknown corresponds to band Z of the reference data set. For many mineral groups this is not at all obvious and requires a time-consuming calibration by means of obtaining both Raman spectra and electron microprobe analyses on precisely the same region in a long series of natural standards in order to be able to follow band Z “step by step” from one end-member to the other (Pinet & Smith, 1993, 1994), since extrapolating from one to the other is very risky. Plotting wavenumber shifts against chemical exchanges in order to observe the behaviour of each Raman band in garnets reveals valuable information that can be used to select which bands show the most consistent spectral appearances and the most linear trends, and should provide the sharpest intersections because of varying in a different way (cf Fig. 15). Thirdly, the most difficult problem encountered is the low intersection angle of some iso-wavenumber isopleths in multidimensional chemical space (when no alternative Raman bands are available); this can lead to large uncertainties and often carries the intersection point out of the system's boundaries and thus yield negative values (obviously impossible chemically but perfectly normal mathematically). Work on this and other problems is continuing gradually. At the moment it is not possible to give error bars with this method, not just because of the complexity of error propagation through simultaneous equations, but because it is difficult to put a value on the estimated error of each Raman or EMP value.

A simple formula for obtaining the jadeite proportion (cJd) of jade in the ternary diopside–hedenbergite–jadeite was given by Smith (in print, b) as:  

formula
with vUnk being the Raman shift of the unknown in cm-1. Applying this simple formula to a newly discovered Guatemalan jade (Gendron et al., 2002), a jadeite proportion of 95 mol% was calculated; subsequent EMP analysis confirmed 97 mol% Jd. Here the accuracy is rather good, around ±2 mol% Jd, despite the fact that this is a short-cut method of less accuracy for estimating the jadeite proportion in jadeite-jade in the Di–Hd–Jd ternary with only one Raman band, because the difference between the Si–O–Si symmetrical stretch wavenumbers for diopside (666 cm-1) and hedenbergite (660 cm-1) are similar (Pinet et al., 1992) and their small difference is ignored; using a second band for a proper ternary solution will give more accurate results, but ±2 mol% is already the best that can be expected. This short-cut is recommended only for rapid rough work on jade. In the Di–Hd–Ae ternary, proper full use of two Raman bands yielded the following results: this semi-quantitative Raman method gave Di0.56Hd0.02Ae0.42 whereas EMP analysis gave Di0.56Hd0.06Ae0.38 (Smith et al., 1989). Some recent work on clinopyroxenes and orthopyroxenes was presented by Mernagh & Hoatson (1997), but this only involved binary systems.

Smith & Périn (2003) used the same method to show non-destructively that many Middle-Age Barbarian garnets in cloisonné gold jewellery from Vicq, France, were solid solutions between pyrope and almandine (often called “rhodolite” by gemmologists although this term is not recognised by the IMA). However when dealing with similar incrusted stones from Brut in North Ossetia, Russian Federation, Smith et al. (2003) noted that the ternary system Pyr–Alm–Spe, and also the quaternary system with grossular added, was totally useless as they yielded extreme negative mol% values. It was necessary to include andradite, and hence a pentary system, and this revealed approximately 80 mol% andradite even if minor negative values occurred for some of the other members and the accuracy is at least as bad as ±20 mol%. Thus whatever the true composition, it is quite certain that the andradite makes up the bulk, and this was a significant (non-destructive!) discovery for archæology as andradite was previously unknown in cloisonné gold jewellery. If the reader has some reservations about the validity of this method, it is sufficient to observe that andradite has the lowest wavenumber of all six common natural garnets for several key Raman bands (e.g. bands I & II at 994 & 873 cm-1, respectively, for andradite; similarly 1014 & 874 cm-1 for grossular, 1027 & 904 cm-1 for spessartine, 1034 & 914 cm-1 for almandine, 1062 & 925 cm-1 for pyrope; values from Pinet & Smith, 1993, 1994). The wavenumbers for certain crystals from Brut were so close to the pure end-member andradite that no other algebraic combination of end-members could possibly give rise to such low wavenumbers. Figure 16 shows the Raman spectra of a crystal from Brut with bands I & II at respectively 995 & 876 cm-1 and of an andradite-rich standard (99.99 mol% And), to display their similarity, and of two pyrope–almandine solid solutions also from Brut to display their considerably higher wavenumbers.

Fig. 16.

Raman spectra of cloisonné-gold garnets in a fibula from Brut, North Ossetia, Russian Federation. From the top downwards a pyrope-rich Pyr–Alm solution, an almandine-rich Pyr–Alm solution, an andradite (from Smith et al., 2003; Smith et al., in prep.) and finally an andradite reference (And0.999Uva0.001) from Pinet & Smith (1993). Note the down wavenumber shifts of band I (on the right) and of band II (the most intense band in this spectral zone) as one approaches andradite. This confirms that the crystal from Brut, deduced to be very rich in andradite from the semi-quantitative method, really is very rich in andradite from comparison with a standard.

Fig. 16.

Raman spectra of cloisonné-gold garnets in a fibula from Brut, North Ossetia, Russian Federation. From the top downwards a pyrope-rich Pyr–Alm solution, an almandine-rich Pyr–Alm solution, an andradite (from Smith et al., 2003; Smith et al., in prep.) and finally an andradite reference (And0.999Uva0.001) from Pinet & Smith (1993). Note the down wavenumber shifts of band I (on the right) and of band II (the most intense band in this spectral zone) as one approaches andradite. This confirms that the crystal from Brut, deduced to be very rich in andradite from the semi-quantitative method, really is very rich in andradite from comparison with a standard.

The usefulness of the semi-quantitative analytical method of Smith & Pinet (1989) depends on the problem at hand; it was designed for the non-destructive analysis of precious samples whether from the cultural heritage or of rare scientific samples such as meteorites, or inclusions of olivine, pyroxene or garnet inside diamonds without the need to extract them. It is very useful for sub-binary jadeite jade and has been shown to be useful for pentary garnets. “Step-by-step” data sets have been built up for other mineral groups such as olivines, spinels and tourmalines (Smith, in prep.). The method is easy and moderately accurate for binary and ternary solid solutions, but becomes less easy and less accurate with each new extra chemical dimension.

Characterisation of the real structure of natural carbon

Internal heterogeneities of diamond crystals

The Raman spectrum of diamond (β-C; cubic space group forumla) is dominated by one intense first-order Raman band with a Raman shift of ∼ 1332 cm-1 at room temperature and ambient pressure, with second-order Raman bands occurring mainly in the range 2100–2700 cm-1 (Solin & Ramdas, 1970). The 1332 cm-1 band is assigned to the main zonecentre optical phonon of diamond, and it is often referred to as LO=TO mode (longitudinal optical and transversal optical lattice vibrations are degenerate, i.e., they have the same frequency, because of the high symmetry of the diamond lattice). Even though consisting virtually of only one band, the Raman spectrum is highly typical of diamond and allows the unambiguous identification of this mineral. For instance, petrologists use Raman microprobe analyses to verify the identity of microdiamonds in high-pressure rocks and, with that, to conclude about depths of rock formation and metamorphic conditions (e.g. Sobolev & Shatsky, 1990; Izraeli et al., 1999; Nasdala & Massonne, 2000; Massonne & Nasdala, 2003; for an overview see Chopin, 2003, and references therein). Many recent applications of the Raman technique are related to materials science research, for instance the analysis of thin films, substrates and layers of (cubic) diamond, as well as (amorphous) diamond-like carbon (DLC), that are applied to the surface of various materials (e.g. Shroder et al., 1990; Knight & White, 1996).

As an example for the potential of Raman spectroscopy for the investigation of natural diamond, we discuss recent Raman studies on diamond crystals from the Panda kimberlite in the Ekati diamond mine, Northwest Territories, Canada, that contain large, single-crystal graphite inclusions. These graphite inclusions exhibit a pseudo-hexagonal, plate-like habit (cf. Fig. 18a below) and are oriented with their (001) face parallel to a (111) face of their diamond host. Glinnemann et al. (2003) found that the graphite inclusions have remnant pressures up to 2.6 GPa (estimated from unit cell parameters). Raman spectra obtained from the neighbouring host diamond have confirmed these pressure estimates (Nasdala et al., 2003b); this was possible because it is well known how much the Raman shift of the LO=TO mode increases with increasing pressure (Grimsditch et al., 1978; Hanfland et al., 1985; Boppart et al., 1985).

Raman maps produced by Nasdala et al. (2003b) showed that inclusions are surrounded by haloes of enhanced intracrystalline pressure/strain (cf. “inhomogeneously-distributed isobarsd”; Smith, 1984), several hundred μm across (cover picture). It is obvious that the external pressure relaxation during the uplift of the diamond crystals must have resulted in heterogeneous expansion of the diamond-inclusion couples. Caused by particularly extensive volume expansion of graphite crystals along their c axes (Zhao & Spain, 1989), complete pressure relaxation in the neighbouring diamond was hindered and, therefore, distinct haloes of enhanced remnant pressures (compressive strain) in the diamond are observed in areas next to graphite (001) faces (see blue-black areas in the cover picture). The volume expansion of graphite along [001] has also led to the opening of disc-shaped cracks in the surrounding diamond parallel to the graphite (001) plane (cf. Fig. 18a below). Diamond areas close to the ends of such cracks are affected by strong dilative strain (see red-yellow areas in the cover picture).

It has been discussed controversially whether the well-shaped graphite single crystals are primary and diamond is secondary in nature, or graphite and diamond are syngenetic and grew more or less simultaneously. Nasdala et al. (submitted) found that Raman maps based on the FWHM of the LO=TO mode show often patterns that reveal the growth zoning of diamond. Based on such patterns indicating that the internal diamond growth texture starts from, and virtually surrounds, the graphite inclusions (Fig. 17), it is now possible to conclude that graphite must be the primary phase and has been overgrown by diamond.

Fig. 17.

Two-dimensional (2D) tomographic Raman mapping used to reveal internal growth textures inside diamond crystals from the Panda kimberlite, Ekati diamond mines, Canada. Samples by courtesy of J.W. Harris and J. Glinnemann. For a more detailed sample description see Glinnemann et al. (2003). (a) Diamond PAG02, view along [110] (modified from Nasdala et al., 2003b). (b) Diamond PAG03, view along [111] (modified from Nasdala et al., submitted). Both diamond crystals contain one large graphite inclusion. Slight variations of the FWHM of the diamond LO=TO mode reveal patterns that are interpreted as growth zoning of the diamond crystals. Areas with particularly high FWHMs (caused by additional local strain) indicate the locations of the graphite crystals and cracks in the host diamond. It can be seen that in both cases the growth zoning originates at the graphite inclusion. This observation characterises the graphite inclusion as the primary mineral and the host diamond as overgrowth.

Fig. 17.

Two-dimensional (2D) tomographic Raman mapping used to reveal internal growth textures inside diamond crystals from the Panda kimberlite, Ekati diamond mines, Canada. Samples by courtesy of J.W. Harris and J. Glinnemann. For a more detailed sample description see Glinnemann et al. (2003). (a) Diamond PAG02, view along [110] (modified from Nasdala et al., 2003b). (b) Diamond PAG03, view along [111] (modified from Nasdala et al., submitted). Both diamond crystals contain one large graphite inclusion. Slight variations of the FWHM of the diamond LO=TO mode reveal patterns that are interpreted as growth zoning of the diamond crystals. Areas with particularly high FWHMs (caused by additional local strain) indicate the locations of the graphite crystals and cracks in the host diamond. It can be seen that in both cases the growth zoning originates at the graphite inclusion. This observation characterises the graphite inclusion as the primary mineral and the host diamond as overgrowth.

Estimation of the order/disorder of graphitic carbon

Single-crystal graphite (α-C; hexagonal space group P63/mmc) has nine vibrational modes, of which only two (E2g type modes) at 42 and 1581 cm-1 are Raman active (Tuinstra & Koenig, 1970; Nemanich et al., 1977; Dresselhaus & Dresselhaus, 1981). Since only few mineralogical laboratories routinely obtain spectra in the low frequency range below 100 cm-1, the first-order Raman spectrum of well-crystallised, natural graphite is generally known to be dominated by a single, intense band at ∼ 1580 cm-1 (Fig. 18d). This relatively sharp band (FWHM < 20 cm-1) is assigned to stretching vibrations of carbon-carbon bonds within the sp2 layers, which are degenerate due to the high symmetry of the hexagonal graphite {001} planes, and it is usually referred to as the graphite G band (e.g. Knight & White, 1989; Kawashima & Katagiri, 1995; Ferrari & Robertson, 2000). Its first overtone was described as weak but sharp band at 3248 cm-1 by Nemanich & Solin (1977), in addition to two more intense second-order bands at ∼ 2440 and ∼ 2730 cm-1 (for detailed band assignment see, for instance, Kawashima & Katagiri, 1995).

The Raman spectra of microcrystalline graphite and disordered carbon show additional, intense bands at ∼ 1350 (the D band; a normally Raman inactive A1g mode that is activated due to the finite crystal size) and ∼ 1620 cm-1 (the D' band; related to the disorder of graphite). Since these two bands (Fig. 18b, d) are normally not observed from single crystal graphite, they are referred to as “disordered modes”. In disordered carbon, the D' mode overlaps with the G mode and, as a result, only one broad band at ∼ 1600 cm-1 is obtained. The width and asymmetry of this overlap peak is, therefore, often used as a measure of the order/disorder of carbon. Note that the intensity ratio of the G and D modes provides only a rough estimate of the crystallinity of carbonaceous samples as long as the orientation of the sp2 carbon layers with respect to the electric field vector of the laser beam is unknown. This is because the D mode is also observed (even though with lower intensity) in the edge plane of macroscopic graphite crystals, which was explained by Katagiri et al. (1988) as due to the breakdown of translational and local lattice symmetries. Note also that the Raman shifts of the D and D' bands, as well as further first-order disordered modes of lower intensity, their overtones and combinations with other modes, vary in dependence with the excitation frequency (Vidano et al., 1981). For instance, the D band shifts from 1365 cm-1 with 457.9 nm excitation down to 1284 cm-1 with 1064 nm excitation (Kawashima & Katagiri, 1995; Wang et al., 1998). The appearance and spectral features (band positions, relative intensities, widths) of the D and D' bands (and their overtones and combination bands) are routinely used to characterise and estimate the crystallinity of naturally formed graphite (for example, Wopenka & Pasteris, 1993; Yui et al., 1996; Beyssac et al., 2002) and other carbonaceous samples such as bitumen (Jehlicka et al., 1997) and aerosols (Sze et al., 2001). In the following, we will briefly discuss three examples.

Fig. 18.

Two examples for the Raman-based structural characterisation of natural carbonaceous samples. (a) Optical photomicrograph of a single-crystal graphite inclusion in diamond PAG07 (view along diamond [111]) from the Ekati diamond mines, Canada (Glinnemann et al., 2003; Nasdala et al., 2003b). The inclusion is surrounded by two disc-like shaped cracks in the host diamond. (b) Two corresponding Raman spectra (632.8 nm excitation). The large graphite crystal (measurement A) is, as expected, well ordered as only a narrow graphite G band is obtained in addition to the intense diamond LO=TO band. The surrounding inner crack (measurement B) shows the presence of disordered graphitic carbon. This is indicated by the graphite D' band, seen as shoulder of the broadened graphite G band. The tails of the strong diamond LO=TO band appear broadened due to graphite D band (at ∼ 1340 cm-1) as well as an additional band at the low-frequency side. (c) Scanning electron microscope image showing the (001) plane of a graphite crystal from a uranium mineralisation in Saskatchewan, Canada (for sample description see Wang et al., 1989). Hollow points were formed by either chemical alteration or radiation effects. (d) Two corresponding Raman spectra (514.5 nm excitation). The main body of the crystal is well ordered. The measurement placed inside a hollow point reveals strong structural disorder. The broad band at ∼ 1600 cm-1 is interpreted to consist of an overlap of the broadened graphite G and D' bands.

Fig. 18.

Two examples for the Raman-based structural characterisation of natural carbonaceous samples. (a) Optical photomicrograph of a single-crystal graphite inclusion in diamond PAG07 (view along diamond [111]) from the Ekati diamond mines, Canada (Glinnemann et al., 2003; Nasdala et al., 2003b). The inclusion is surrounded by two disc-like shaped cracks in the host diamond. (b) Two corresponding Raman spectra (632.8 nm excitation). The large graphite crystal (measurement A) is, as expected, well ordered as only a narrow graphite G band is obtained in addition to the intense diamond LO=TO band. The surrounding inner crack (measurement B) shows the presence of disordered graphitic carbon. This is indicated by the graphite D' band, seen as shoulder of the broadened graphite G band. The tails of the strong diamond LO=TO band appear broadened due to graphite D band (at ∼ 1340 cm-1) as well as an additional band at the low-frequency side. (c) Scanning electron microscope image showing the (001) plane of a graphite crystal from a uranium mineralisation in Saskatchewan, Canada (for sample description see Wang et al., 1989). Hollow points were formed by either chemical alteration or radiation effects. (d) Two corresponding Raman spectra (514.5 nm excitation). The main body of the crystal is well ordered. The measurement placed inside a hollow point reveals strong structural disorder. The broad band at ∼ 1600 cm-1 is interpreted to consist of an overlap of the broadened graphite G and D' bands.

We have already described above that extensive volume expansion along the c axis of single-crystal graphite inclusions in diamond crystals from the Panda kimberlite, Canada, has typically resulted in the opening of disc-shaped cracks in the neighbouring host diamond (Fig. 18a). Inspection under a high-power optical binocular microscope indicated that the cracks may be partially filled by carbonaceous matter. This was confirmed by Nasdala et al. (2003b) who showed that the pseudo-hexagonal graphite crystal is well-ordered whereas cracks are partially filled with poorly ordered sp2 (i.e., graphitic) carbon (Fig. 18b). Note that spectrum B in Figure 18b gives also some indication for the additional presence of another carbon species. The D band of disordered graphitic carbon should have its maximum intensity at ∼ 1340 cm-1 with He-Ne excitation and, therefore, most of its intensity should be expected at the high-frequency tail of the strong diamond LO=TO band (signal from the diamond host). The additional intensity at the low-frequency tail points to the presence of a sp3 carbon species, possibly disordered nanometre-sized diamond (e.g. Yoshikawa et al., 1995).

In a comprehensive study, Wang et al. (1989) investigated graphitic samples from a metamorphised uranium deposit in Saskatchewan, Canada. These authors observed moderate graphite degeneration only at the surface and interpreted this as a result of a secondary alteration process undergone by the deposit. In addition, some graphite samples that were collected from places close to uranium concentrations exhibit numerous hollow points with sizes of several μm and smaller (Fig. 18c). These hollow points were found to show severe structural damage, indicated by an almost amorphous structure of the carbon (Fig. 18d). Wang et al. (1989) discussed that the higher degree of graphite alteration observed inside the hollow points might be due to either irradiation or chemical damage.

The third example is related to an application that has been controversially discussed in the past two years, namely, the Raman analysis of potential carbonaceous remnants of very old microfossils (e.g. Schopf et al., 2002; Brasier et al., 2002). Schopf et al. (2002) proposed that poor ordering of carbonaceous unknowns might be indicative of kerogen (i.e., maturated bio-organic matter). In Figure 19, we present Raman spectra and maps obtained from graphite flakes occurring in an early Archaean, metasomatic rock (metacarbonate) from the Isua Supracrustal Belt, southern West Greenland (Lepland et al., 2002; van Zuilen et al., 2002, 2003). Analyses done on graphite flakes exposed to the polished surface yielded Raman spectra that are typical of highly disordered carbon (Fig. 19a). However, it is well known that graphite (Wang et al., 1989) and other minerals (e.g. Libowitzky, 1994) may experience superficial structural damage and disorder due to the mechanical thinning and polishing process, which draws into question the high disorder being deduced as a natural feature of graphite having become exposed at the surface of polished sections. These doubts were confirmed by Raman maps (Figs. 19b, c) obtained from a graphite flake that is partly exposed at the surface of the section and partly buried under the thin cover of chlorite. It can be seen that the graphite flake is generally well ordered in areas where it is covered by chlorite while it is highly disordered only in three areas in which the graphite is exposed at the surface and was affected by the mechanical polishing process. Raman spectroscopy cannot be thus applied as an unambiguous biodiagnostic tool because the reliable distinction between kerogen and highly disordered graphite of inorganic origin is not possible from the spectra alone (e.g. Pasteris & Wopenka, 2002).

Fig. 19.

Raman maps of a graphite flake in a metacarbonate (sample no. AL8-1; courtesy of A. Lepland) from the 3.8 Ga Isua Supracrustal Belt, Greenland, showing that graphite disorder can be induced by the polishing process. (a) The two Raman spectra show that graphite that has been exposed to the surface of the thin section is disordered (spectrum A) whereas the main graphite flake, analysed through a thin chlorite cover, appears well-ordered (spectrum B). (b) A Raman map of the surface, generated from the integral intensity of the band at ∼ 1600 cm-1, shows three micro-areas in which graphite is exposed at the surface. This map corresponds widely to an optical microphotograph taken in the reflected light mode (not shown). (c) A Raman map, recorded with the focus of the fully focused beam adjusted ∼ 2 μm below the surface and generated from the broadening of the ∼ 1600 cm-1 band, shows that only the three exposed areas have experienced structural disorder (large FWHMs) as a result of the mechanical polishing process. Surrounding micro-areas that are still covered by chlorite are well ordered (smaller FWHMs).

Fig. 19.

Raman maps of a graphite flake in a metacarbonate (sample no. AL8-1; courtesy of A. Lepland) from the 3.8 Ga Isua Supracrustal Belt, Greenland, showing that graphite disorder can be induced by the polishing process. (a) The two Raman spectra show that graphite that has been exposed to the surface of the thin section is disordered (spectrum A) whereas the main graphite flake, analysed through a thin chlorite cover, appears well-ordered (spectrum B). (b) A Raman map of the surface, generated from the integral intensity of the band at ∼ 1600 cm-1, shows three micro-areas in which graphite is exposed at the surface. This map corresponds widely to an optical microphotograph taken in the reflected light mode (not shown). (c) A Raman map, recorded with the focus of the fully focused beam adjusted ∼ 2 μm below the surface and generated from the broadening of the ∼ 1600 cm-1 band, shows that only the three exposed areas have experienced structural disorder (large FWHMs) as a result of the mechanical polishing process. Surrounding micro-areas that are still covered by chlorite are well ordered (smaller FWHMs).

Gemstone identification by Raman spectroscopy analysis through glass

Three Florentine tables in stone marquetry, a large one of white marble inlaid with precious stones representing mainly flowers and birds or insects, and two smaller ones of black marble inlaid mainly with flowers and fruit, are conserved inside the high-security “Trésor” of the Muséum National d'Histoire Naturelle in Paris. These beautiful early XVIIth century tables are each covered today by a heavy sheet, 1.6 cm thick, of protective glass. Although the mineral species of many of the gemstones had previously been recognised [e.g. purple amethyst (quartz), blue lapis-lazuli (lazurite)], their identity had never been definitively proved, partly because of the overlying plate glass and partly because of the obvious refusal to allow the extraction of any crystal or part thereof for precise gemmological analysis. For several of the gemstones recognition was not obvious, especially the whitish and greenish ones composing different kinds of flowers.

A mobile Raman system equipped with optical fibres and a green 532 nm laser and a separate unit with a red 785 nm laser were carried into the Trésor. The protective glass was not removed from the tables, partly to avoid possibly damaging the table and partly because of its considerable weight. Successively for each table the small box containing the laser source, photon detector and spectrometer unit was placed on a trolley on the floor near the table and a photograph's tripod was placed on the glass. The remote head was suspended vertically from the tripod and the laser, arriving by an optical fibre, was focussed through the remote lens on to a specific crystal of interest 7.5 cm away (Fig. 20). The Raman scattered light returned by the same optical path. It was very easy to just slide the tripod over the glass and analyse any crystal at will, often with no need to readjust the focus. The results were obtained rapidly by means of their Raman spectra appearing on the computer screen, which was also placed on top of the glass. With this configuration, all parts of the mobile Raman unit and of the table were within reach of the analyst.

Fig. 20.

Raman analysis though a glass cover I. (a) Part of a mobile Raman system sitting on top of a very large precious Florentine white marble table in stone marquetry protected by 1.6 cm thick plate glass (reflections visible). Note the incrusted gemstones forming various ornamental designs, especially birds and flowers. Note the tripod suspending the remote head with its long focal length lens (black tube with white rim at bottom reflected on to the computer box at the side and on to the glass below). Treasure vault of the National Museum of Natural History, Paris, in 2000. (b) A close-up view of remote identification through glass of a blue tulip-like flower inlaid in the same marble Florentine table. In this case, the obtained PL spectrum confirmed the expected mineral nature: lazurite (the key colorant of lapis lazuli). The white mark is at the end of the remote lens and the shadows come from the tripod legs. (c) Verification of the mineral species of a yellowish-white petal of an inlaid flower in a magnificent Florentine table of black marble; note that the computer keyboard and mouse as well as the tripod were able to be placed over this precious work because it is protected by a glass plate (invisible). The collection lens is suspended 7.5 cm above the crystal. All photographs were taken by D.C. Smith.

Fig. 20.

Raman analysis though a glass cover I. (a) Part of a mobile Raman system sitting on top of a very large precious Florentine white marble table in stone marquetry protected by 1.6 cm thick plate glass (reflections visible). Note the incrusted gemstones forming various ornamental designs, especially birds and flowers. Note the tripod suspending the remote head with its long focal length lens (black tube with white rim at bottom reflected on to the computer box at the side and on to the glass below). Treasure vault of the National Museum of Natural History, Paris, in 2000. (b) A close-up view of remote identification through glass of a blue tulip-like flower inlaid in the same marble Florentine table. In this case, the obtained PL spectrum confirmed the expected mineral nature: lazurite (the key colorant of lapis lazuli). The white mark is at the end of the remote lens and the shadows come from the tripod legs. (c) Verification of the mineral species of a yellowish-white petal of an inlaid flower in a magnificent Florentine table of black marble; note that the computer keyboard and mouse as well as the tripod were able to be placed over this precious work because it is protected by a glass plate (invisible). The collection lens is suspended 7.5 cm above the crystal. All photographs were taken by D.C. Smith.

The protective plate glass could be assumed to constitute an obstacle, as it indeed is for applying almost all, if not all, other microanalytical techniques. Five reasons explain why the glass does not create any insurmountable problem, and in most cases no problem at all. Firstly, it is well known that the Raman effect concerns the transmission of a beam of light and a transparent medium like glass or water allows the beam to cross it, and to return if 180° geometry is used (as in this case where the laser beam successively travelled through an optical fibre, the remote lens, the air above the table, the glass on the table, and finally the crystal being analysed; the Raman diffused light that was collected returned by the same glass, same air and same lens and then another optical fibre). Secondly, the glass yields its own Raman spectrum, but, although being superposed on the Raman spectrum of the crystal being analysed, this is no problem if the glass spectrum is weaker and/or its Raman bands are wider relative to those of the crystal. Thirdly, the Raman bands of the glass may be in spectral zones where there are no Raman bands from the crystal, and vice versa, hence no problematic overlap. Fourthly, if there is a problematic overlap, it is easy to acquire a Raman spectrum of the pure glass and then simply subtract it from the combined spectrum. Lastly, but not least, the laser beam is focussed by the remote lens and the photon collection system collects mostly light emanating from the level of the focal point on or in the crystal, such that most of the Raman signal from the overlying glass is not recorded (cf. confocal analysis of a microinclusion where it is possible to exclude the Raman signal from the host mineral).

In the larger table the matrix is a sheet of white marble, calcite being confirmed by its Raman spectrum (Fig. 20a). The green thorax of an insect initially posed a problem since its identification was far from obvious (pyroxene?); in fact its Raman spectrum quickly revealed ordinary α-quartz (Smith & Rondeau, 2001). Several parts of flowers or birds or ornamental borders are composed of a deep blue material, for example in the petals of blue tulip-like flowers (Fig. 20b; Smith, 2001a); all parts analysed yielded an intense identical luminescence with the red laser which corresponded to the luminescence from the lazurite in a type specimen of lapis-lazuli from Afghanistan. Thus here the identification of lazurite in the table was made by luminescence with a Raman spectrometer and not by Raman spectra which were completely eclipsed by the luminescence.

In both of the smaller tables the matrix is of black marble and there are more fruit representations than on the large white table, as well as a greater proportion of incrustations to matrix (Fig. 20c). Purple grapes gave, as expected, the Raman spectrum of α-quartz (i.e., amethyst from its characteristic colour). The petals of some flowers were yellowish white and could have been of calcite, quartz or even of ivory or yet some other phase like feldspar; in fact they are of dolomite as proved rapidly by its key Raman bands at 178, 301, 724 and 1098 cm-1 (Fig. 21; Smith, in print, a). A pomegranate has many red pips which were presumed all to be of garnet; in fact some were of garnet but others were of ruby; this was especially interesting as it indicated the use of two different mineral species for the same motif and also the use of a very precious stone.

This project, the first of its kind reported, was in fact incredibly easy as the apparatus was perfect for the job to be done and it is clear that various other projects necessitating analysing through glass should be also feasible. The real problem was statistical, as there were too many crystals for all of them to be analysed in the time available; for example all of the hundreds of similar-looking blue crystals, but with varying hues, may well be of lazurite, but to what extent can this statement be justified on the basis of analysing only a handful of crystals? However this problem is no worse than that for all other mixtures containing innumerable grains (cf. mineral pigments and of course natural rocks).

Fig. 21.

Raman analysis though a glass cover II. The inset photograph (corresponding to the centre of Fig. 20c) is a close-up of a yellowish-white petal being analysed under the thick protective glass by the red 632.8 nm laser beam (spot is seen as bright circle). The Raman spectrum obtained from one of the yellowish-white petals of the flower (corrected for backgound luminescence; solid graph) is shown in comparison with the spectrum of dolomite from Freiberg, Germany (dotted). The obtained spectrum reveals the four principal bands of dolomite.

Fig. 21.

Raman analysis though a glass cover II. The inset photograph (corresponding to the centre of Fig. 20c) is a close-up of a yellowish-white petal being analysed under the thick protective glass by the red 632.8 nm laser beam (spot is seen as bright circle). The Raman spectrum obtained from one of the yellowish-white petals of the flower (corrected for backgound luminescence; solid graph) is shown in comparison with the spectrum of dolomite from Freiberg, Germany (dotted). The obtained spectrum reveals the four principal bands of dolomite.

In situ Raman spectroscopy of quartz: A new pressure sensor

Quartz is one of the early and most extensively studied minerals in Raman spectroscopy (e.g. Raman & Nedungadi, 1940; Gillet et al., 1990). Due to the importance of quartz in crustal processes as well as for technical applications, a number of experimental studies of the temperature and pressure variations in the Raman spectrum of quartz have been carried out (e.g. Raman & Nedungadi, 1940; Pine & Tannenwald, 1969; Briggs & Ramdas, 1977; Dean et al., 1982; Jayaraman et al., 1987; Gillet et al., 1990; Liu & Mernagh, 1992; Castex & Madon, 1995). Furthermore, SiO2 exhibits a number of high-pressure and/or high-temperature polymorphs. These phases (Smith, 1984) and the phase boundaries among them (Gillet et al., 1998) can be well identified by Raman spectroscopy. Thermodynamic properties of quartz have been inferred from vibrational modelling (Kieffer, 1979; Gillet et al., 1990; Castex & Madon, 1995). Such models are based on Grüneisen parameters, which are calculated from experimentally derived parameters like pressure- and temperature-induced shifts of the frequencies of vibrational modes. To improve such data, pressure sensors are needed that work reliably at high temperatures, too.

Coupling a powerful Raman spectrometer with a hydrothermal diamond anvil cell (HDAC; Bassett et al., 1996) permits precise in situ Raman spectroscopy of α-quartz at temperatures from 23 to 800 °C and simultaneously at hydrostatic pressures ranging between 0.1 MPa and 2.1 GPa (Schmidt & Ziemann, 2000). Due to the specific measurement procedure reported in detail in that study, the errors of data are comparably small in the investigated temperature and pressure ranges. The accuracy and reproducibility of the temperature measurement is < ±1.5 °C. Determined pressures have typical errors of < ±20 MPa even at high temperatures up to 800 °C. To improve the accuracy of the Raman line position determination, the wavenumbers are calibrated using the plasma lines of the Ar+ laser by removing the interference filter (plasma lines at 116.0 cm-1, 266.3 cm-1, and 520.3 cm-1; see lower spectra in Fig. 22). The resulting wavenumber accuracy is about ±0.2 cm-1.

Fig. 22.

Raman spectra of α-quartz, presenting examples for pressure-induced (top pair of spectra) and temperature-induced (below) shifts of Raman-active modes. Note that in particular the 206 cm-1 band (A1 mode) shows dramatic changes in frequency and width. In the lower pair of spectra, the 128 cm-1 α-quartz band is partially obscured by a plasma line emitted by the Ar+ laser (apparent Raman shift 116 cm-1; marked with asterisks).

Fig. 22.

Raman spectra of α-quartz, presenting examples for pressure-induced (top pair of spectra) and temperature-induced (below) shifts of Raman-active modes. Note that in particular the 206 cm-1 band (A1 mode) shows dramatic changes in frequency and width. In the lower pair of spectra, the 128 cm-1 α-quartz band is partially obscured by a plasma line emitted by the Ar+ laser (apparent Raman shift 116 cm-1; marked with asterisks).

The lines of the three most intense Raman modes of α-quartz (464, 206 and 128 cm-1 at ambient conditions) shift towards higher wavenumbers with increasing P (pressure) and towards lower wavenumbers with increasing T (temperature; Fig. 22). The pressure-induced frequency shifts of these modes show very different slopes and behaviours of linearity at room temperature (Fig. 23a). No significant deviations were found among the data obtained from the two samples used in that study, which was done on a synthetic and a natural quartz crystal.

Fig. 23.

Pressure-induced shifts of Raman bands of α-quartz. (a) Frequency shifts of the three most intense Raman bands in the α-quartz spectrum as a function of pressure (observed at room temperature). Frequency shifts are plotted relative to the band positions at 0.1 MPa. Note that pressure-induced frequency changes of some quartz bands exceed the pressure-induced frequency shift of the ruby luminescence (e.g. Piermarini et al., 1975; Mao et al., 1986). (b) Pressure-induced shift of the 464 cm-1 A1 mode as a function of temperature. Provided the temperature is known, this plot can be used for pressure determination in HP-HT experiments. Plots from Schmidt & Ziemann (2000), modified.

Fig. 23.

Pressure-induced shifts of Raman bands of α-quartz. (a) Frequency shifts of the three most intense Raman bands in the α-quartz spectrum as a function of pressure (observed at room temperature). Frequency shifts are plotted relative to the band positions at 0.1 MPa. Note that pressure-induced frequency changes of some quartz bands exceed the pressure-induced frequency shift of the ruby luminescence (e.g. Piermarini et al., 1975; Mao et al., 1986). (b) Pressure-induced shift of the 464 cm-1 A1 mode as a function of temperature. Provided the temperature is known, this plot can be used for pressure determination in HP-HT experiments. Plots from Schmidt & Ziemann (2000), modified.

The most intense A1 Raman mode of quartz (at 464 cm-1) shows simultaneously a relatively large and quasi-linear pressure dependence and a still moderate shift with temperature in the studied P-T range (Fig. 23b). The pressure-induced frequency change of this band is slightly higher than that of the ruby fluorescence scale (Fig. 23a). Therefore, the Raman shift of this band is proposed as pressure sensor in HDAC studies of SiO2-saturated systems at temperatures between –269 °C and 560 °C. Equations for the pressure calculation were presented by Schmidt & Ziemann (2000).

The A1 Raman mode of quartz at 206 cm-1 is strongly anharmonic (Gillet et al., 1990; Castex & Madon, 1995). With increasing pressure, the frequency of this mode increases non-linearly with a negative curvature at room temperature (Fig. 23a). The pressure-induced frequency shift relative to the line position at 0.1 MPa forumla exceeds that of the ruby fluorescence by about 3 times. Therefore, this mode can be used alternatively to the ruby sensor for room-temperature experiments at pressures up to about 5 GPa with the equation of Schmidt & Ziemann (2000):  

formula

A first application of pressure-induced shifts of α-quartz Raman bands as a pressure sensor was presented by Ziemann & van den Kerkhof (2002). These authors have analysed a quartz specimen from an amphibolite-grade metamorphosed migmatite from the Colorado Front Range, USA (for the sample description, see Olsen, 1987). The quartz crystals were found to contain several superdense CO2 inclusions (van den Kerkhof & Olsen, 1990). In their Raman study, Ziemann & van den Kerkhof (2002) used the host quartz itself as a “pressure gauge” to measure the pressure field around the CO2 inclusions. They found that in Raman linescans across one of the “superdense” CO2 inclusions, the A1 α-quartz mode at 206 cm-1 shifted towards higher wavenumbers, indicating significantly increased internal pressure of the host quartz in micro-areas close to the inclusion. The measured band upshift corresponded to internal pressures on the order of 40 MPa (Ziemann & van den Kerkhof, 2002). The observed field of enhanced pressure in the host quartz is assumed to stabilize the superdense inclusions.

Summary: Advantages and disadvantages

Instead of attempting to present conclusions about the Raman technique and its applications, we complete this paper with a brief summary of the analytical disadvantages and advantages that are most relevant for researchers who apply Raman analysis to the study of minerals and other geological objects. It is remarkable that, in spite of the great research progress and technical developments during the last decade, including the advent of several new analytical methods, the principal disadvantages and advantages of Raman spectroscopy are essentially the same as those listed almost 20 years ago by Herman et al. 1985; for more recent discussions see for instance Hope et al., 2001; Nasdala et al., 2001a; Bugay, 2001).

One of the main disadvantages is still the limited availability of reliable reference data for minerals. Most geoscientific applications use Raman spectroscopy as a fingerprinting tool for the identification of minerals and related phases. In such cases it is not absolutely necessary to know which vibrations the observed bands are assigned to, rather it would be sufficient to have Raman data of the relevant standard species available for comparison. Even though quite a number of institutions and Raman laboratories have been working hard over the last few years to build up their own Raman databases for minerals, a fully reliable and comprehensive library does not yet exist. Several catalogues of Raman spectra have been published already, and many spectra collections can be found on the internet. Critical evaluation of these data, however, implies that a significant number of spectra (according to a cautious estimation of about 5%) are incorrect. Searching for the spectrum of a certain mineral is also made difficult because of different terminologies. For instance, to find the spectra of the minerals azurite, cuprorivaite, lazurite, goethite and orpiment in a mineral pigment library, one may need to know respectively names such as Mountain blue, Egyptian blue, Ultramarine, Yellow ochre and King's yellow. Finally, none of the presently existing databases are complete (i.e., none contains the spectra of all common rock-forming minerals). This problem, however, will hopefully be remedied in the not too distant future.

If a Raman analysis yields a spectrum that is unknown to the researcher, there are basically three practical ways for the identification of the sample, namely (1) finding a sufficiently matching spectrum by a library search (problems addressed above), (2) partially identifying an unknown, for example as a carbonate or as a sulphate on the basis of a strong Raman band in the appropriate range for the most intense CO3 or SO4 vibration, using existing spectra of similarly composed/structured minerals, or (3) producing one's own standards (i.e., obtaining a number of spectra from available or synthesised minerals whose chemical composition might be similar to that of the unknown). The hypothetical fourth way, namely, the calculation of the theoretical Raman spectrum of a suspected mineral species, is mostly impossible or at least highly impractical. There exist a number of computer codes with which it is possible to calculate quickly a theoretical X-ray diffraction pattern from the unit cell parameters and atomic positions of a mineral. By contrast, the calculation of a fully reliable theoretical Raman spectrum is costly in terms of time for simple mineral structures and still almost impossible for complex mineral structures, especially in view of the non-ideal chemical and structural composition of naturally formed minerals. This difficulty is also affected by another general problem of Raman spectroscopy, which is the often complicated band assignment.

Furthermore, it must be mentioned that not all mineral species can be analysed using the Raman technique. Some minerals have no first-order Raman spectrum (due to their lattice symmetry, an example is halite). Others are in general difficult to analyse, like manganese oxides, because of their thermal sensitivity and/or strong light absorption, or because of their poor transparency (especially minerals with a metallic lustre). Finally, other effects such as the occurrence of intense luminescence (of various origins, e.g. trace elements such as Cr3+; organic detritus as occurs in biominerals, or in dead microbes or algae in ancient pigments; patina on climate-exposed archaeological artefacts) and the so-called “Raman background” (an extremely broad Raman signal related to defects; see Pilz & Kriegsmann, 1987; Splett et al., 2000) may hinder the successful Raman analysis of unknowns.

In spite of these problems, Raman spectroscopy, favoured by a number of analytical advantages, has been successfully used in all geoscientific disciplines, and the number of Raman-related publications per year increases rapidly. First of all, analysis is mostly possible without any sample preparation. There is no need for powdering, thinning, sawing, scraping, breaking, drilling, coating/sputtering or dissolving of the sample, and often the object to be analysed does not even need to be removed from its matrix. Sample polishing is mostly also unnecessary, but it is usually done if the same samples are to be analysed with other techniques anyway, or to improve the beam quality through the suppression of diffuse reflection/scattering of light at a rough sample surface. Another important point is that Raman analysis can be done non-destructively. This is most advantageous for complex studies applying a variety of techniques to the very same micro-areas in minerals. Since a Raman analysis does not cause any permanent damage to the sample (of course only under the prerequisite that any local strong light absorption, photochemical reaction etc. was avoided by the experimentalist), it will not affect the results of other techniques that are performed later. In contrast, it is well known that an electron microprobe analysis is not non-destructive on a scale of single analysis pits. Raman measurements should, therefore, generally be done before electron microprobe point analyses. A non-destructive technique (which does not even require any sample preparation) is also very well suitable for the analysis of valuable and rare objects such as gemstones or unique historic treasures. Furthermore, Raman analysis can be done without any direct contact with the sample. This advantage is used, for example, in the in situ analysis of inclusions inside their host mineral (e.g. included liquids and gases, fossilised pressures in and around HP inclusions), the analysis inside many kinds of cells and containers (e.g. diamond anvil cells, heating experiments, industrial processes) and truly at-distance analysis (e.g. environmental analysis).

The Raman technique has an excellent volume resolution down to the micrometer scale. This allows for the analysis of tiny amounts of sample and truly microprobe analysis. Due to the excellent volume resolution, information on the homogeneity/heterogeneity of samples can be obtained, and images can be generated, both with a lateral resolution of ∼ 1 μm. Another advantage of the Raman technique is its sensitivity to short-range order in materials, which allows for the investigation of glasses, fully metamict minerals, melts and other “X-ray amorphous” samples (for example, structural information on the initial stages of recrystallisation processes, and bonding and coordination of ions in molten rock can be obtained). Last but not least, powerful Raman systems are comparably affordable in terms of acquisition and maintenance (for instance, an electrical socket and a regular office table are sufficient nowadays to run a Raman microprobe system, and beam calibration and adjustment procedures have become really undemanding).

It is clearly not our goal to claim that Raman yields more information than other analytical techniques. This may be true in some cases, but in other cases it is other techniques that yield more useful data (e.g. isotopes or trace elements). No technique is perfect for everything, and the applied technique(s) will always depend on the analytical problem to be solved. The difference from most other analytical techniques and one of the greatest advantages of Raman spectroscopy is that the information (mineral identification or mineral characterisation) can often be obtained in a really straightforward and non-destructive way. Hence it does not need much insight to foresee that in future, with an increasing number of spectrometer systems at universities and other research institutions and more extensive reference data available, Raman will soon become a routine technique in mineralogy, as common as IR spectroscopy, which in several research fields is gradually being replaced by Raman. Whereas most analytical techniques provide only the chemical composition or the physical structure of an analysed volume, with Raman specroscopy both are obtained simultaneously, even if indirectly (cf. IR absorption and X-ray diffraction). Also there are not many techniques that can analyse solids, liquids or gases, crystalline, molecular, vitreous or amorphous inorganic or organic materials, and especially mixtures of these such as composite fluid inclusions, biominerals or pigments. If there were one keyword which best summarises the generally advantageous nature of the Raman spectroscopic technique, it would probably be “versatility”.

References

Alberto
,
H.V.
Ayres De Campos
,
N.
Mysen
,
B.O.
(
1995
):
The structural role of titanium in silicate glasses: a Raman study of the system CaO-SiO2-TiO2
.
Phys. Chem. Glasses
 ,
36
:
112
114
.
Amme
,
M.
Renker
,
B.
Schmid
,
B.
Feth
,
M.P.
Bertagnolli
,
H.
Döbelin
,
W.
(
2002
):
Raman microspectrometric identification of corrosion products formed on UO2 nuclear fuel during leaching experiments
.
J. Nucl. Mater.
 ,
306
:
202
212
.
Andersen
,
T.
O'Reilly
,
S.Y.
Griffin
,
W.L.
(
1984
):
The trapped fluid phase in upper mantle xenoliths from Victoria, Australia; implications for mantle metasomatism
.
Contrib. Mineral. Petrol.
 ,
88
:
72
85
.
Andersen
,
T.
Burke
,
E.A.J.
(
1996
):
Methane inclusions in shocked quartz from the Gardnos impact breccia, south Norway
.
Eur. J. Mineral.
 ,
8
:
927
936
.
Andersen
,
T.
Burke
,
E.A.J.
Austrheim
,
H.
(
1989
):
Nitrogen-bearing, aqueous fluid inclusions in some eclogites from the Western Gneiss Region of the Norwegian Caledonides
.
Contrib. Mineral. Petrol.
 ,
103
:
153
165
.
Balan
,
E.
Neuville
,
D.R.
Trocellier
,
P.
Fritsch
,
E.
Muller
,
J.-P.
Calas
,
G.
(
2001
):
Metamictization and chemical durability of detrital zircon
.
Am. Mineral.
 ,
86
:
1025
1033
.
Balz
,
M.
Therese
,
H.A.
Li
,
J.
Gutmann
,
S.
Kappl
,
M.
Nasdala
,
L.
Hofmeister
,
W.
Butt
,
H.-J.
Tremel
,
W.
:
Crystallization of raterite nanowires by the cooperative interaction of tailor-made nucleation surfaces and polyelectrolytes
.
Adv. Mater.
 
(submitted)
.
Banerjee
,
A.
Jensen
,
J.O.
Jensen
,
J.L.
(
2003
):
A theoretical study of As4S4: Bonding, vibrational analysis and infrared and raman spectra
.
J. Mol. Struct.
 ,
626
:
63
75
.
Barrie
,
I.J.
Touret
,
J.L.R.
(
1999
):
Fluid inclusion studies of gold-bearing quartz veins from the Yirisen Deposit, Sula Mountains greenstone belt, Masumbiri, Sierra
.
Ore Geol. Rev.
 ,
14
:
203
225
.
Bassett
,
W.A.
(
2003
):
High pressure-temperature aqueous systems in the hydrothermal diamond anvil cell (HDAC)
.
Eur. J. Mineral.
 ,
15
:
730
780
.
Bassett
,
W.A.
Shen
,
A.H.
Bucknum
,
M.
Chou
,
I.-M.
(
1993
):
A new diamond anvil cell for hydrothermal studies to 2.5 GPa and from –190 to 1200 °C
.
Rev. Sci. Instrum.
 ,
64
:
2340
2345
.
Bassett
,
W.A.
Wu
,
T.-C.
Chou
,
I.-M.
Haselton
,
H.T.
Jr.
Frantz
,
J.
Mysen
,
B.
Huang
,
W.-L.
Sharma
,
S.K.
Schiferl
,
D.
(
1996
):
The hydrothermal diamond anvil cell (HDAC) and its applications
. In
Dyar
,
M.D.
McCammon
,
C.
Schaefer
,
M.W.
(eds.):
Mineral spectroscopy: A tribute to Roger G. Burns /Geochem. Soc. Spec. Publ.
 ,
5
/.
Houston (Tex.)
:
Geochem. Soc.
,
261
272
.
Bell
,
I.M.
Clark
,
R.J.H.
Gibbs
,
P.J.
(
1997
):
Raman spectroscopic library of natural and synthetic pigments (pre ∼ 1850 AD)
.
Spectrochim. Acta A
 ,
53
:
2159
2179
.
Benison
,
K.C.
Goldstein
,
R.H.
Wopenka
,
B.
Burruss
,
R.C.
Pasteris
,
J.D.
(
1998
):
Extremely acid Permian lakes and ground waters in North America
.
Nature
 ,
392
:
911
914
.
Bény
,
C.
Guilhaumou
,
N.
Touray
,
J.-C.
(
1982
):
Native sulphur-bearing fluid inclusions in the CO2-H2S-H2O-S system - microthermometry and Raman microprobe (MOLE) analysis - thermochemical interpretations
.
Chem. Geol.
 ,
37
:
113
127
.
Beran
,
A.
Voll
,
D.
Schneider
,
H.
(
2004
):
IR spectroscopy as a tool for the characterisation of ceremic precursor phases
. In
Beran
,
A.
Libowitzky
,
E.
(eds.):
Spectroscopic methods in mineralogy /EMU Notes Mineral.
 ,
6
/.
Budapest
:
Eötvös Univ. Press
,
189
226
.
Beyssac
,
O.
Rouzaud
,
J.-N.
Goffé
,
B.
Brunet
,
F.
Chopin
,
C.
(
2002
):
Graphitization in a high-pressure, low- temperature metamorphic gradient: a Raman microspectroscopy and HRTEM study
.
Contrib. Mineral. Petrol.
 ,
143
:
19
31
.
Birch
,
B.
Kolitsch
,
U.
Witzke
,
T.
Nasdala
,
L.
Bottrill
,
R.
(
2001
):
Petterdite, the Cr-dominant analogue of dundasite, a new mineral from Dundas, Tasmania, Australia and Callenberg, Saxony, Germany
.
Can. Mineral.
 ,
38
:
1467
1476
.
Bischoff
,
W.D.
Sharma
,
S.K.
Mackenzie
,
F.T.
(
1985
):
Carbonate ion disorder in synthetic and biogenic magnesian calcites: A Raman spectral study
.
Am. Mineral.
 ,
70
:
581
589
.
Boehler
,
R.
Chopelas
,
A.
(
1992
):
Phase transitions in a 500 kbar – 3000 K gas apparatus
. In
Syono
,
Y.
Manghnani
,
M.H.
(eds.):
High-pressure research: Application to Earth and planetary sciences /Geophys. Monogr.
 ,
67
/.
Washington (D.C.)
:
Am. Geophys. Soc., Terra Scientific
,
55
60
.
Boiron
,
M.-C.
Moissette
,
A.
Cathelineau
,
M.
Banks
,
D.
Monin
,
C.
Dubessy
,
J.
(
1999
):
Detailed determination of paleofluid chemistry: an integrated study of sulphate-volatile rich brines and aqueo-carbonic fluids in quartz veins from Ouro Fino (Brazil)
.
Chem. Geol.
 ,
154
:
179
192
.
Bondarenko
,
G.V.
Gorbaty
,
Yu. E.
(
1999
):
In situ Raman spectroscopic study of sulfur-saturated water at 1000 bar between 200 and 500°C
.
Geochim. Cosmochim. Acta
 ,
61
:
1413
1420
.
Boppart
,
H.
van Straaten
,
J.
Silvera
,
I.F.
(
1985
):
Raman spectra of diamond at high pressures
.
Phys. Rev., B
 ,
32
:
1423
1425
.
Bouchard
,
M.
Smith
,
D.C.
(
1999
):
Metalloraman: corroded antique and modern coins studied by Raman Microscopy
. In
Georaman'99 (Valladolid), Abstr.
 /
Univ. Valladolid Press, Spec. Publ.
/
31
32
.
Bouchard
,
M.
Smith
,
D.C.
(
2000
):
Metalloraman
. “
Archéométrie”, Doss. Archéol.
 ,
253
:
157
170
.
Bouchard
,
M.
Smith
,
D.C.
(
2001
):
Evaluating Raman Microscopy for the non-destructive archaeometry of corroded coins: a powerful technique for conservation studies
.
Asian Chem. Lett.
 ,
5
,
3
:
157
170
.
Bouchard
,
M.
Smith
,
D.C.
(
2002
):
Archaeological and experimental stained glass: a non-destructive Raman microscopic (RM) study
. In
“Art 2002”, 7th Int. Conf. on non-destructive testing and microanalysis for the diagnostics and conservation of the cultural and environmental heritage (Antwerp), Abstr.
 ,
Univ. of Antwerp
,
152
.
Bouchard
,
M.
Smith
,
D.C.
(
2003
):
Catalogue of 45 reference Raman spectra of minerals concerning research in art history or archaeology, especially on corroded metals & coloured glass
.
Spectrochim. Acta A
 ,
59
:
2247
2266
.
Bouchard
,
M.
Smith
,
D.C.
:
Database of 74 Raman spectra of standard minerals of relevance to metal corrosion, stained glass or prehistoric rock art
. In
Edwards
,
H.G.M.
Chalmers
,
J.M
: (eds.):
Raman spectroscopy in archaeology and art history /RSC Anal. Spectrosc. Monogr./.
 
Cambridge
:
R. Soc. Chem.
,
in print
.
Boyer
,
L.L.
Fleury
,
P.A.
(
1974
):
Determination of interatomic interactions in Ca10(PO4)6F2 (fluorapatite) from structural and lattice-dynamical data
.
Phys. Rev., B
 ,
9
:
2693
2700
.
Boyer
,
H.
Smith
,
D.C.
Guyot
,
F.
(
1985
):
RMP (Raman microprobe) characterisation of micron-sized run products in high P-T experimental mineralogy
. In
Conf. Exhib. on Analytical Chemistry and Applied Spectroscopy (Pittsburgh), Abstr.
 ,
1244
.
Brasier
,
M.D.
Green
,
O.R.
Jephcoat
,
A.P.
Kleppe
,
A.K.
Van Kranendonk
,
M.J.
Lindsay
,
J.F.
Steele
,
A.
Grassineau
,
N.V.
(
2002
):
Questioning the evidence for Earth's oldest fossils
.
Nature
 ,
416
:
76
81
.
Brawer
,
S.A.
White
,
W.B.
(
1974
):
Raman spectroscopic investigation of the structure of silicate glasses. I. The binary alkali silicates
.
J. Chem. Phys.
 ,
63
:
2421
2432
.
Brawer
,
S.A.
White
,
W.B.
(
1976
):
Raman spectroscopic investigation of the structure of silicate glasses (II). Soda- Alkaline earth - Alumina ternary and quaternary glasses
.
J. Non-Cryst. Solids
 ,
23
:
261
278
.
Brawer
,
S.A.
White
,
W.B.
(
1978
):
Structure and crystallization behaviour of LiO-Fe2O3-SiO2 glasses
.
J. Mater. Sci.
 ,
13
:
1907
1920
.
Brewer
,
P.G.
Pasteris
,
J.D.
Malby
,
G.
Peltzer
,
E.
White
,
S.
Freeman
,
S.
Wopenka
,
B.
Brown
,
M.
Cline
,
D.
(
2002
):
Laser Raman spectroscopy used to study the ocean at 3600–m depth
.
EOS Trans. Am. Geophys. Union
 ,
83
:
469
470
.
Briggs
,
R.J.
Ramdas
,
A.K.
(
1977
):
Piezospectroscopy of the Raman spectrum of α-quartz
.
Phys. Rev., B
 ,
16
:
3815
3826
.
Brody
,
R.H.
Edwards
,
H.G.M.
Farwell
,
D.W.
O'Connor
,
S.
Pollard
,
A.M.
(
1998
):
Real or fake: FT-Raman study of ivories
. In
Heyns
,
A.M.
(ed.):
ICORS Capetown'
 
98
.
Chichester
:
Wiley
,
22
523
.
Brugger
,
J.
Bonin
,
M.
Schenk
,
K.J.
Meisser
,
N.
Berlepsch
,
P.
Ragu
,
A.
(
1999
):
Description and crystal structure of nabiasite, BaMn9[(V,As)O4]6(OH)2, a new mineral from the central Pyrenees (France)
.
Eur. J. Mineral.
 ,
11
:
879
890
.
Bugay
,
D.E.
(
2001
):
Characterization of the solid state: spectroscopic techniques
.
Adv. Drug Delivery Rev.
 ,
48
:
43
65
.
Bühn
,
B.
Rankin
,
A.H.
Haller
,
M.
Radtke
,
M.
Knöchel
,
A.
(
1999
):
Burbankite, a (Sr,REE,Na,Ca)-carbonate in fluid inclusions from carbonatite-derived fluids: identification and characterization using laser Raman spectroscopy, SEM-EDX and synchrotron micro-XRF analysis
.
Am. Mineral.
 ,
84
:
1117
1125
.
Burgio
,
L.
Clark
,
R.J.H.
(
2001
):
Library of FT-Raman spectra of pigments, minerals, pigment media and varnishes, and supplement to existing library of Raman spectra of pigments with visible excitation
.
Spectrochim. Acta A
 ,
57
:
1491
1521
.
Burke
,
E.A.J.
(
1998
):
New data on zircon from Matongo (Burundi)
.
Aardk. Meded. KU Leuven
 ,
9
:
17
20
.
Burke
,
E.A.J.
(
2001
):
Raman microspectrometry of fluid inclusions
.
Lithos
 ,
139
:
139
158
.
Burke
,
E.A.J.
Lustenhouwer
,
W.J.
(
1987
):
The application of a multichannel laser Raman microprobe (Microdil-28®) to the analysis of fluid inclusions
.
Chem. Geol.
 ,
61
:
11
17
.
Castex
,
J.
Madon
,
M.
(
1995
):
Test of the vibrational modelling for the λ-type transitions: application to the α-β quartz transition
.
Phys. Chem. Miner.
 ,
22
:
1
10
.
Cervelle
,
B.
Moëlo
,
Y.
(
1990
):
Advanced microspectroscopy
. In
Jambor
,
J.L.
Vaughan
,
D.
(eds.):
Advanced microscopic studies of ore minerals
./
MAC Short Course Handb.
 ,
17
/,
Ottawa
:
Mineral. Assoc. Can
,
379
408
.
Cesare
,
B.
Maineri
,
C.
(
1999
):
Fluid-present anatexis of metapelites at El Joyazo (SE Spain): constraints from Raman spectrosopy of graphite
.
Contrib. Mineral. Petrol.
 ,
135
:
41
52
.
Chalain
,
J.-P.
Fritsch
,
E.
Hänni
,
H.A.
(
1999
):
Detection of GE POL diamonds: a first step
.
Rev. Gemmol., Assoc. Fr. Gemmol.
 ,
138–139
:
30
33
.
Chalain
,
J.-P.
Fritsch
,
E.
Hänni
,
H.A.
(
2000
):
Identification of GE POL diamonds: a second step
.
J. Gemmol.
 ,
27
:
73
78
.
Champagnon
,
B.
Panczer
,
G.
Chazallon
,
B.
Arnaud
,
L.
Duval
,
P.
Lipenkov
,
V.
(
1997
):
Nitrogen and oxygen guest molecules in clathrates hydrates: different sites revealed by Raman Spectroscopy
.
J. Raman Spectrosc.
 ,
28
:
711
716
.
Chervin
,
J.C.
Canny
,
B.
Gauthier
,
M.
Pruzan
,
P.
(
1992
):
Micro-Raman at variable low-temperature and very high pressure
.
Rev. Sci. Instrum.
 ,
64
:
203
206
.
Chopelas
,
A.
Serghiou
,
G.
(
2002
):
Spectroscopic evidence for pressure-induced phase transitions in diopside
.
Phys. Chem. Miner.
 ,
29
:
403
408
.
Chopin
,
C.
(
2003
):
Ultrahigh-pressure metamorphism: tracing continental crust into the mantle
.
Earth Planet. Sci. Lett.
 ,
212
:
1
14
.
Chou
,
I.M.
Pasteris
,
J.D.
Seitz
,
J.C.
(
1990
):
High-density volatiles in the system C-O-H-N for the calibration of a laser Raman microprobe
.
Geochim. Cosmochim. Acta
 ,
54
:
535
543
.
Colomban
,
P.
Treppoz
,
F.
(
2001
):
Identification and differentiation of ancient and modern European porcelains by Raman macro- and micro-spectroscopy
.
J. Raman Spectrosc.
 ,
32
:
93
102
.
Coupry
,
C.
Brissaud
,
D.
(
1996
):
Applications in Art, Jewellery and Forensic Science
. In
Turell
,
G.
Corset
,
J.
(eds.):
Raman microscopy, developments and applications
 .
London
:
Academic Press
,
421
453
.
Damen
,
T.C.
Porto
,
S.P.S.
Tell
,
B.
(
1966
):
Raman effect in zinc oxide
.
Phys. Rev.
 ,
142
:
570
574
.
Darimont
,
A.
Burke
,
E.
Touret
,
J.
(
1988
):
Nitrogen-rich metamorphic fluids in Devonian metasediments from Bastogne, Belgium
.
Bull. Minéral.
 ,
111
:
321
330
.
Daudon
,
M.
Protat
,
M.F.
Reveillaud
,
R.J.
Jaeschke-Boyer
,
H.
(
1981
):
Anwendung des Lasers für die Harnsteinanalyse
.
Fortschr. Urol. Nephrol.
 ,
17
:
311
321
.
Dawson
,
P.
Hargreave
,
M.M.
Wilkinson
,
G.F.
(
1971
):
The vibrational spectrum of zircon (ZrSiO4)
.
J. Phys. C. Solid State Phys.
 ,
4
:
240
256
.
Dean
,
K.J.
Sherman
,
W.F.
Wilkinson
,
G.R.
(
1982
):
Temperature and pressure dependence of the Raman active modes of vibration of α-quartz
.
Spectrochim. Acta
 ,
38A
:
1105
1108
.
Délé-Dubois
,
M.L.
Dhamelincourt
,
P.
Schubnel
,
H.J.
(
1980
):
Etude par spectrométrie Raman d'inclusions dans les diamants, saphirs et émeraudes
.
Rev. Gemmol., Assoc. Fr. Gemmol.
 ,
63–64
:
11
14
.
Délé-Dubois
,
M.L.
Poirot
,
J.P.
Schubnel
,
H.J.
(
1986
a
):
Identification de micro-inclusions dans des rubis et émeraudes de synthése par spectroscopie Raman
.
Rev. Gemmol., Assoc. Fr. Gemmol.
 ,
88
:
15
17
.
Délé-Dubois
,
M.L.
Dhamelincourt
,
P.
Poirot
,
J.P.
Schubnel
,
H.J.
(
1986
b
):
Differentiation between natural gems and synthetic minerals by laser Raman micro-spectrometry
.
J. Mol. Struct.
 ,
143
:
135
138
.
Delhaye
,
M.
Dhamelincourt
,
P.
(
1975
):
Raman microprobe and microscope with laser excitation
.
J. Raman Spectrosc.
 ,
3
:
33
.
Delhaye
,
M.
Guineau
,
B.
Vezin
,
J.
Coupry
,
C.
(
1985
):
La microsonde Raman au secours des oeuvres d'art
.
Measures
 ,
11
:
119
124
.
Devarajan
,
V.
Klee
,
W.E.
(
1981
):
A potential model for fluorapatite
.
Phys. Chem. Miner.
 ,
7
:
35
42
.
Dhamelincourt
,
P.
(
1987
):
Laser Raman and fluorescence microprobing techniques
.
Anal. Chim. Acta
 ,
195
:
33
43
.
Dhamelincourt
,
P.
Bisson
,
P.
(
1977
):
Principe et réalisation d'un microscope optique utilisant l'effet Raman
.
Microsc. Acta
 ,
79
:
267
276
.
Dhamelincourt
,
P.
Beny
,
J.
Dubessy
,
J.
Poty
,
B.
(
1979
):
Analyse d'incluisons fluides à la microsonde MOLE à effet Raman
.
Bull. Minéral.
 ,
102
:
600
610
.
Dietrich
,
D.
Witke
,
K.
Rößler
,
R.
Marx
,
G.
(
2001
):
Raman spectroscopy on Psaronius sp.: a contribution to the understanding of the permineralization process
.
Appl. Surface Sci.
 ,
179
:
231
234
.
Dresselhaus
,
M.S.
Dresselhaus
,
G.
(
1981
):
Intercalation compounds of graphite
.
Adv. Phys.
 ,
30
:
139
326
.
Dubessy
,
J.
Audeoud
,
D.
Wilkins
,
R.
Kosztolanyi
,
C.
(
1982
):
The use of the Raman microprobe mole in the determination of the electrolytes dissolved in the aqueous phase of fluid inclusions
.
Chem. Geol.
 ,
37
:
137
150
.
Dubessy
,
J.
Geisler
,
D.
Kosztolanyi
,
C.
Vernet
,
M.
(
1983
):
The determination of sulphate in fluid inclusions using the M.O.L.E. Raman microprobe. Application to a Keuper halite and geochemical consequences
.
Geochim. Cosmochim. Acta
 ,
47
:
1
10
.
Dubessy
,
J.
Poty
,
B.
Ramboz
,
C.
(
1989
):
Advances in C-O-H-N-S fluid geochemistry based on micro- Raman spectrometric analysis of fluid inclusions
.
Eur. J. Mineral.
 ,
1
:
517
534
.
Dubessy
,
J.
Boiron
,
M.C.
Moissette
,
A.
Monnon
,
C.
Sretenskaya
,
N.
(
1992
):
Determinations of water, hydrates and pH in fluid inclusions by micro-Raman spectrometry
.
Eur. J. Mineral.
 ,
4
:
885
894
.
Dubessy
,
J.
Moissette
,
A.
Bakker
,
R.J.
Frantz
,
J.D.
Zhang
,
Y.G.
(
1999
):
High-temperature Raman spectroscopic study of H2O-CO2-CH4 mixtures in synthetic fluid inclusions; first insights on molecular interactions and analytical implications
.
Eur. J. Mineral.
 ,
11
:
23
32
.
Edwards
,
H.G.M.
Barry
,
B.W.
Williams
,
A.C.
Rull
,
F.
(
1996
a
):
The iceman: an FT-Raman spectroscopic study of 5200-year-old human skin
.
Terra Abstr., Abstr. Suppl. 2, to Terra Nova
 ,
8
:
1
.
Edwards
,
H.G.M.
Falk
,
M.J.
Farwell
,
D.W.
Janaway
,
R.C.
(
1996
b
):
FT-Raman spectroscopic studies of linen samples from archaeological burial sites
.
Terra Abstr., Abstr. Suppl. 2, to Terra Nova
 ,
8
:
2
.
Edwards
,
H.G.M.
Daffner
,
L.A.
Falk
,
M.J.
Farwell
,
D.W.
Heron
,
C.
Quye
,
A.
(
1996
c
):
FT-Raman spectroscopic analyses of ancient resins of archaeological significance
.
Terra Abstr., Abstr. Suppl. 2, to Terra Nova
 ,
8
:
1
.
Edwards
,
H.G.M.
Newton
,
E.M.
Russ
,
J.
(
2000
):
Raman spectroscopic analysis of pigments and substrata in Prehistoric rock art
.
J. Mol. Struct.
 ,
550–551
:
245
256
.
Etchepare
,
J.
Merian
,
M.
Smetankine
,
L.
(
1974
):
Vibrational normal modes of SiO2. I α and β quartz
.
J. Chem. Phys.
 ,
60
:
1873
1876
.
Etchepare
,
J.
Merian
,
M.
Kaplan
,
P.
(
1978
):
Vibrational normal modes of SiO2. II. Cristobalite and tridymite
.
J. Chem. Phys.
 ,
68
:
1531
1537
.
Farber
,
D.L.
Williams
,
Q.
(
1996
):
An in situ Raman spectroscopic study of Na2Si2O5 at high pressures and temperatures: Structures of compressed liquids and glasses
.
Am. Mineral.
 ,
81
:
273
283
.
Fateley
,
W.G.
Dollish
,
F.R.
McDevitt
,
N.T.
Benley
,
F.F.
(
1972
):
Infrared and Raman selection rules for molecular and lattice vibrations: the correlation method
 .
New York (N.Y.)
:
Wiley
.
Ferrari
,
A.C.
Robertson
,
J.
(
2000
):
Interpretation of Raman spectra of disordered and amorphous carbon
.
Phys. Rev., B
 ,
61
:
14095
14107
.
Ferraro
,
J.R.
(
1984
):
Vibrational spectroscopy at high external pressures – The diamond anvil cell
 .
New York (N.Y.)
:
Academic Press
,
264
p.
Frantz
,
J.D.
Mysen
,
B.O.
(
1995
):
Raman spectra and structure of BaO-SiO2, SrO-SiO2 and CaO-SiO2 melts to 1600° C
.
Chem. Geol.
 ,
121
:
155
176
.
Frantz
,
J.D.
Dubessy
,
J.
Mysen
,
B.O.
(
1994
):
Ion-pairing in aqueous MgSO4 solutions along an isochore to 500 °C and 11 kbar using Raman spectroscopy in conjunction with the diamond-anvil cell
.
Chem. Geol.
 ,
116
:
181
188
.
Freeman
,
J.J.
Wopenka
,
B.
Silva
,
M.J.
Pasteris
,
J.D.
(
2001
):
Raman spectroscopic detection of changes in bioapatite in mouse femora as a function of age and in-vitro fluoride treatment
.
Calcif. Tissue Int.
 ,
68
:
156
162
.
Frezzotti
,
M.L.
Burke
,
E.A.J.
De Vivo
,
B.
Stefanini
,
B.
Villa
,
I.M.
(
1992
):
Mantle fluids in pyroxenite nodules from Salt Lake Crater (Oahu, Hawaii)
.
Eur. J. Mineral.
 ,
4
:
1137
1153
.
Frezzotti
,
M.L.
Di Vincenzo
,
G.
Ghezzo
,
C.
Burke
,
E.A.J.
(
1994
):
Evidence of magmatic CO2-rich fluids in peraluminous graphite-bearing leucogranite from Deep Freeze Range (northern Victoria Land, Antarctica)
.
Contrib. Mineral. Petrol.
 ,
117
:
111
123
.
Fricke
,
A.
Medenbach
,
O.
Schreyer
,
W.
(
1990
):
Fluid inclusions, planar elements and pseudotachylites in the basement rocks of the Vredefort structure, South Africa
.
Tectonophysics
 ,
171
:
169
183
.
Frost
,
R.L.
(
2003
):
Some interesting lead compounds of importance in archaeology: a Raman spectroscopic study
. In
Int. Congr. Application of Raman spectroscopy in art and archaeology (Ghent), Abstr.
 ,
57
.
Fry
,
R.L.
Frost
,
R.L.
Turner
,
J.
Hallam
,
D.
(
1998
):
Analysis of ceramics from the wreck of the “Pandora” using Raman and FTIR spectroscopy
. In
Heyns
,
A.M.
(ed.):
ICORS Capetown'
 
98
.
Chichester
:
Wiley
,
546
547
.
Furukawa
,
T.
White
,
W.B.
(
1978
):
Structure and crystallisation of glasses in the Li2Si2O5-TiO2 system determined by Raman spectroscopy
.
Phys. Chem. Glasses
 ,
20
:
69
80
.
Furukawa
,
T.
White
,
W.B.
(
1981
):
Raman spectroscopy of heat-treated B2O3-SiO2 glasses
.
J. Am. Ceram. Soc.
 ,
64
:
443
447
.
Furukawa
,
T.
Brawer
,
S.
White
,
W.B.
(
1978
):
The structure of lead silicate glasses determined by vibrational spectroscopy
.
J. Mater. Sci.
 ,
13
:
268
282
.
Furukawa
,
T.
Fox
,
K.E.
White
,
W.B.
(
1981
):
Raman spectroscopic investigation of the structure of silicate glasses. III. Raman intensities and structural units in sodium silicate glasses
.
J. Chem. Phys.
 ,
75
:
3226
3237
.
Gardiner
,
D.J.
Graves
,
P.R.
(eds.) (
1989
):
Practical Raman spectroscopy
 .
Berlin
:
Springer-Verlag
.
Gendron
,
F.
Smith
,
D.C.
Gendron-Badou
,
A.
(
2002
):
Discovery of jadeite-jade in Guatemala confirmed by non-destructive Raman microscopy
.
J. Archaeol. Sci.
 ,
29
:
837
851
.
Gillet
,
P.
(
1996
):
Raman spectroscopy at high pressure and high temperature. Phase transitions and thermodynamic properties of minerals
.
Phys. Chem. Miner.
 ,
23
:
263
275
.
Gillet
,
P.
Le Cléach
,
A.
Madon
,
M.
(
1990
):
High-temperature Raman spectroscopy of SiO2 and GeO2 polymorphs: anharmonicity and thermodynamic properties at high temperatures
.
J. Geophys. Res.
 ,
95
:
21635
21655
.
Gillet
,
P.
Hemley
,
R.J.
McMillan
,
P.F.
(
1998
):
Vibrational properties at high pressures and temperatures
. In
Hemley
,
R.J.
(ed.):
Ultrahigh-pressure mineralogy – Physics and chemistry of the Earth's deep interior /Rev. Mineral.
 ,
37
/.
Washington (D.C.)
:
Mineral. Soc. Am.
,
525
590
.
Gillet
,
P.
Sautter
,
V.
Harris
,
J.
Reynard
,
B.
Harte
,
B.
Kunz
,
M.
(
2002
):
Raman spectroscopic study of garnet inclusions in diamonds from the mantle transition zone
.
Am. Mineral.
 ,
87
:
312
317
.
Giuliani
,
G.
Dubessy
,
J.
Banks
,
D.A.
Hoang Quang
,
Vinh
Lhomme
,
T.
Pironon
,
J.
Garnier
,
V.
Phan Trong
,
Trinh
Pham Van
,
Long
Ohnenstetter
,
D.
Schwarz
,
D.
(
2003
):
CO2-H2S-COS-S8-AlO(OH)- bearing fluid inclusions in ruby from marble-hosted deposits in Luc Yen area, North Vietnam
.
Chem. Geol.
 ,
194
:
167
185
.
Glinnemann
,
J.
Kusaka
,
K.
Harris
,
J.W.
(
2003
):
Oriented graphite single-crystal inclusions in diamond
.
Z. Kristallogr.
 ,
218
:
733
739
.
Götze
,
J.
Nasdala
,
L.
Kleeberg
,
R.
Wenzel
,
M.
(
1998
):
Occurrence and distribution of “moganite” in agate/chalcedony: a micro-Raman, Rietfeld, and cathodoluminescence study
.
Contrib. Mineral. Petrol.
 ,
133
:
96
105
.
Grice
,
J.D.
Van Velthuizen
,
J.
Dunn
,
P.J.
Newbury
,
D.E.
Etz
,
E.S.
Nielsen
,
C.H.
(
1986
):
Moydite (yttrium, rare earth elements) [B(OH)4](CO3). A new mineral species from the Evans-Lou pegmatite, Quebec
.
Can. Mineral.
 ,
24
:
665
673
.
Griffith
,
W.P.
(
1987
):
Advances in the Raman and infrared spectroscopy of minerals
. In
Clark
,
R.J.H.
Hester
,
R.E.
(eds.):
Spectroscopy of inorganic-based materials
 .
New York (N.Y.)
:
Wiley
,
119
187
.
Grimsditch
,
M.H.
Anastassakis
,
E.
Cardona
,
M.
(
1978
):
Effect of uniaxial stress on the zone-center optical phonon of diamond
.
Phys. Rev., B
 ,
18
:
901
904
.
Guilhaumou
,
N.
(
1982
):
Accurate analysis of fluid inclusions by the laser molecular microprobe (MOLE) and by microthermometry
.
Trav. Lab. Geol. Ec. Norm. Super.
 ,
14
.
Guilhaumou
,
N.
Dhamelincourt
,
P.
Touray
,
J.-C.
Barbillat
,
J.
(
1978
):
Raman microprobe analysis of gas inclusions in the N2-CO2 system
.
C.R. Acad. Sci., D: Sci. Nat.
 ,
287
:
1317
1319
.
Guilhaumou
,
N.
Jouaffre
,
D.
Velde
,
B.
Bény
,
C.
(
1988
):
Raman microprobe analysis on gaseous inclusions from the diagenetically altered Terres Noires (SE France)
.
Bull. Minéral.
 ,
111
:
577
585
.
Hanfland
,
M.
Syassen
,
K.
Fahy
,
S.
Louie
,
S.G.
Cohen
,
M.L.
(
1985
):
Pressure dependence of the first-order Raman mode in diamond
.
Phys. Rev., B
 ,
31
:
6896
6899
.
Hänni
,
H.
Schubiger
,
B.
Kiefert
,
L.
Häberli
,
S.
(
1998
):
Raman investigations on two historical objects from Basel cathedral: the Reliquary Cross and Dorothy Moinstrance
.
Gems Gemol.
 ,
34
:
102
113
.
Hayes
,
W.
Loudon
,
R.
(
1978
):
Scattering of light by crystals
 .
New York (N.Y.)
:
Wiley
.
Henderson
,
G.S.
Fleet
,
M.E.
(
1995
):
The structure of Ti silicate glasses by micro-Raman spectroscopy
.
Can. Mineral.
 ,
33
:
399
408
.
Herman
,
R.G.
Bogdan
,
C.E.
Sommer
,
A.J.
(
1985
):
Laser Raman microprobe study of the identification and thermal transformations of some carbonate and alumosilicate minerals
. In
Snyder
,
R.L.
Condrate
,
R.A.
Johnson
,
P.F.
(eds.):
Advances in materials characterization II/Mater. Sci. Res.
 ,
19
/.
New York (N.Y.)
:
Plenum Press
,
113
130
.
Hofmeister
,
A.M.
(
1999
):
Mantle values of thermal conductivity and the geotherm from phonon lifetime
.
Science
 ,
283
:
1699
1706
.
Hofmeister
,
A.M.
(
2001
):
Thermal conductivity of spinels and olivines from vibrational spectroscopy: Ambient conditions
.
Am. Mineral.
 ,
86
:
1188
1208
.
Hofmeister
,
A.M.
Mao
,
H.-K.
(
2002
):
Redefinition of the mode Grüneisen parameter for polyatomic substances and thermodynamic implications
.
Proc. Natl. Acad. Sci. USA
 ,
99
:
559
564
.
Holtstam
,
D.
(
1997
):
Hiaerneite, a new, Zr-Sb oxide mineral isostructural with calzirtite, from Langban, Sweden
.
Eur. J. Mineral.
 ,
9
:
843
848
.
Hope
,
G.A.
Woods
,
R.
Munce
,
G.C.
(
2001
):
Raman microprobe mineral identification
.
Miner. Eng.
 ,
14
:
1565
1577
.
Huizenga
,
J.M.
Touret
,
J.L.R.
(
1999
):
Fluid inclusions in shear zones - The case of the Umwindsi Shear Zone in the Harare-Shamva-Bindura Greenstone-Belt, NE Zimbabwe
:
Eur. J. Mineral.
 ,
11
:
1079
1090
.
Iqbal
,
Z.
Tomaselli
,
V.P.
Fahrenfeld
,
O.
Möller
,
K.D.
Ruszala
,
F.A.
Kostiner
,
E.
(
1977
):
Polarized Raman scattering and low frequency infrared study of hydroxyapatite
.
J. Phys. Chem. Solids.
 ,
38
:
923
927
.
Irmer
,
G.
(
1985
):
Zum Einfluß der Apparatefunktion auf die Bestimmung von Streuquerschnitten und Lebensdauern aus optischen Phononenspektren
.
Exp. Tech. Phys.
 ,
33
:
501
506
.
Irmer
,
G.
Graupner
,
T.
(
2002
):
Isotopes of C and O in CO2: a laser Raman study using gas standards and natural fluid inclusions
. In
Georaman 2002 (Prague), Abstr., Acta Univ. Carolinae, Geol.
 
46
:
35
36
.
Izraeli
,
E.S.
Harris
,
J.W.
Navon
,
O.
(
1999
):
Raman barometry of diamond formation
.
Earth Planet. Sci. Lett.
 ,
173
:
351
360
.
Jayaraman
,
A.J.
(
1983
):
Diamond anvil cell and high-pressure physical investigations
.
Rev. Mod. Phys.
 ,
55
:
65
85
.
Jayaraman
,
A.J.
(
1986
):
Ultrahigh pressures
.
Rev. Sci. Instrum.
 ,
57
:
1013
1031
.
Jayaraman
,
A.J.
Wood
,
D.L.
Maines
,
R.G.
(
1987
):
High pressure Raman study of the vibrational modes in AlPO4 and SiO2 (α-quartz)
.
Phys. Rev., B
 ,
35
:
8316
8321
.
Jehlicka
,
J.
Beny
,
C.
Rouzaud
,
J.-N.
(
1997
):
Raman microspectrometry of accumulated non-graphitized solid bitumens
.
J. Raman Spectrosc.
 ,
28
:
717
724
.
Kaindl
,
R.
Hoinkes
,
G.
Knoll
,
P.
Abart
,
R.
(
1999
):
Fluid inclusions related to Variscan and Alpine metamorphism in the Austroalpine Ötztal Basement, Eastern Alps
.
Mineral. Petrol.
 ,
65
:
29
49
.
Katagiri
,
G.
Ishida
,
H.
Ishitani
,
A.
(
1988
):
Raman spectra of graphite edge planes
.
Carbon
 ,
26
:
565
571
.
Kawashima
,
Y.
Katagiri
,
G.
(
1995
):
Fundamentals, overtones, and combinations in the Raman spectrum of graphite
.
Phys. Rev., B
 ,
52
:
10053
10059
.
Keramidas
,
V.G.
Deangelis
,
B.A.
White
,
W.B.
(
1975
):
Vibrational spectra of spinels with cation ordering on the octahedral sites
.
J. Solid State Chem.
 ,
15
:
233
245
.
Kiefert
,
L.
Hänni
,
H.A.
Ostertag
,
T.
(
2001
):
Raman spectroscopic applications to gemmology
. In
Lewis
,
I.
Edwards
,
H.G.M.
(eds.):
A handbook on Raman spectroscopy
 .
New York (N.Y.)
:
Marcel Dekker
,
469
489
.
Kieffer
,
S.W.
(
1979
):
Thermodynamics and lattice vibrations of minerals
.
Rev. Geophys. Space Phys.
 ,
17
:
1
59
.
Kingma
,
K.J.
Hemley
,
R.J.
(
1994
):
Raman spectroscopic study of microcrystalline silica
.
Am. Mineral.
 ,
79
:
269
273
.
Kleppe
,
A.K.
Jephcoat
,
A.P.
Smyth
,
J.R.
(
2002
):
Raman spectroscopic study of hydrous γ-Mg2SiO4 to 56.5 GPa
.
Phys. Chem. Miner.
 ,
29
:
473
476
.
Kleppe
,
A.K.
Jephcoat
,
A.P.
Welch
,
M.D.
(
2003
):
The effect of pressure upon hydrogen bonding in chlorite: A Raman spectroscopic study on chlinochlore to 26,5 GPa
.
Am. Mineral.
 ,
88
:
567
573
.
Knight
,
D.S.
White
,
W.B.
(
1996
):
Characterization of diamond films by Raman spectroscopy
.
J. Mater. Res.
 ,
4
:
385
393
.
Koděra
,
P.
Murphy
,
P.J.
Rankin
,
A.H.
(
2003
):
Retrograde mineral reactions in saline fluid inclusions: The transformation ferropyrosmalite ↔ clinopyroxene
.
Am. Mineral.
 ,
88
:
151
158
.
Kolesov
,
B.A.
Geiger
,
C.A.
Armbruster
,
T.
(
2001
):
The dynamic properties of zircon studied by single- crystal X-ray diffraction and Raman spectroscopy
.
Eur. J. Mineral.
 ,
13
:
939
948
.
Kolitsch
,
U.
(
2003
):
Mg-rich wolfeite, (FeII,Mg)2(PO4)(OH): structure refinement and Raman spectroscopic data
.
Acta Cryst.
 ,
E59
:
i125
i128
.
Krause
,
W.
Bernhardt
,
H.-J.
Effenberger
,
H.
Kolitsch
,
U.
Lengauer
,
C.
(
2003
):
Redefinition of arhbarite, Cu2Mg(AsO4)(OH)3
.
Mineral. Mag.
 ,
67
:
1099
1107
.
Kravitz
,
L.C.
Kingsley
,
J.D.
Elkin
,
E.L.
(
1968
):
Raman and infrared studies of coupled
forumla vibrations.
J. Chem. Phys.
 ,
49
:
4600
4610
.
Kreisel
,
J.
Glazer
,
A.M.
Jones
,
G.
Thomas
,
P.A.
Abello
,
L.
Lucazeau
,
G.
(
2000
):
An X-ray diffraction and Raman spectroscopy investigation of A-site substituted perovskite compounds: the (Na1-xKx)0.5Bi0.5TiO3 (0≤x≤1) solid solution
.
J. Phys., Condens. Matter
 ,
12
:
3267
3280
.
Krzemnicki
,
M.S.
(
1999
):
Diopside needles as inclusions in demantoid garnet from Russia: a Raman microspectrometric study
.
Gems Gemol.
 ,
35
:
192
195
.
Kunz
,
M.
Gillet
,
P.
Fiquet
,
G.
Sautter
,
V.
Graafsma
,
H.
Conrad
,
P.
Harris
,
J.
(
2002
):
Combined in situ X-ray diffraction and Raman spectroscopy on majoritic garnet inclusions in diamonds
.
Earth Planet. Sci. Lett.
 ,
198
:
485
493
.
Kuzmany
,
H.
(
1989
):
Festkörperspektroskopie: Eine Einführung
 .
Berlin
:
Springer-Verlag
.
Landsberg
,
G.
Mandelstam
,
L.
(
1928
):
Eine neue Erscheinung bei der Lichtzerstreuung in Krystallen
.
Naturwissenschaften
 ,
16
:
557
558
.
Lasnier
,
B.
(
1989
):
Rubis, spinelle ou grenat en inclusion dans un diamant
? In
Georaman-89
  (
Paris
), Abstr.,
/Assoc. Natl. Rech. Tech. Spec. Publ.
17
/.
Lehnert
,
R.
(
2000
):
Beyond imagination – image formation based on Raman spectroscopy
.
GIT Imag. Microsc.
 , (
2
):
6
9
.
Leng
,
J.
Sharma
,
A.
Bodnar
,
R.J.
Pottorf
,
R.J.
Vityk
,
M.O.
(
1998
):
Quantitative analysis of synthetic fluid inclusions in the H2O-CH4 system using Raman spectroscopy
. In
Vanko
,
D.A.
Cline
,
J.S.
(eds.):
PACROFI - Pan-American Conf. on Research on Fluid Inclusions, Progr. Abstr.
 ,
7
:
59
.
Lepland
,
A.
Arrhenius
,
G.
Cornell
,
D.
(
2002
):
Apatite in early Archean Isua supracrustal rocks, southern West Greenland: its origin, association with graphite and potential as a biomarker
.
Precambrian Res.
 ,
118
:
221
241
.
Lewis
,
I.R.
Edwards
,
H.G.M.
(eds.) (
2001
):
Handbook of Raman spectroscopy. From the research laboratory to the process line
./
Pract. Spectrosc. Ser.
 ,
28
/.
New York (N.Y.)
:
Marcel Dekker
,
1074
p.
Libowitzky
,
E.
(
1994
):
Anisotropic pyrite: A polishing effect
.
Phys. Chem. Miner.
 ,
21
:
97
103
.
Liem
,
N.Q.
Sagon
,
G.
Quang
,
V.X.
Tan
,
H.V.
Colomban
,
P.
(
2000
):
Raman study of the microstructure, composition and processing of ancient Vietnamese porcelains and celadons
.
J. Raman Spectrosc.
 ,
31
:
933
942
.
Liu
,
L.G.
Mernagh
,
T.P.
(
1992
):
High-pressure Raman study of the α-quartz forms of SiO2 and GeO2 at room temperature
.
High Temp. High Press.
 ,
24
:
13
21
.
Liu
,
L.G.
Mernagh
,
T.P.
Jaques
,
A.L.
(
1990
):
A mineralogical Raman spectroscopy study on eclogitic garnet inclusions in diamonds from Argyle
.
Contrib. Mineral. Petrol.
 ,
105
:
156
161
.
Loh
,
E.
(
1973
):
Optical vibrations in sheet silicates
.
J. Phys. C. Solid State Phys.
 ,
6
:
1091
1104
.
Long
,
D.A.
(
1977
):
Raman spectroscopy
 .
New York (N.Y.)
:
McGraw-Hill
,
276
p.
Loudon
,
R.
(
1964
):
Raman scattering from crystals
.
Adv. Phys.
 ,
13
:
423
482
.
Loudon
,
R.
(
2001
):
The Raman effect in crystals
.
Adv. Phys.
 ,
50
:
813
864
.
Lucovsky
,
G.
Mooradian
,
A.
Taylor
,
W.
Wright
,
G.B.
Keezer
,
R.C.
(
1967
):
Identification of the fundamental vibrational modes of trigonal, α-monoclinic and amorphous selenium
.
Solid State Commun.
 ,
5
:
113
117
.
Maestrati
,
R.
(
1989
):
Contribution à l'édification du catalogue Raman des gemmes
.
Ph.D. thesis
 ,
Univ. Nantes
,
France
.
Malézieux
,
J.M.
Barbillat
,
J.
Cervelle
,
B.
Coutures
,
J.-P.
Couzi
,
M.
Piriou
,
B.
(
1983
):
Etude de spinelles de synthese de la serie MgCrxAl2-xO4 et de chromites naturelles par microsonde Raman-laser. Synthetic spinels of the MgCrxAl2-xO4 series and natural chromites studied by Raman-laser microprobe
.
Tschermaks Mineral. Petrogr. Mitt.
 ,
32
:
171
185
.
Mao
,
H.K.
Xu
,
J.
Bell
,
P.M.
(
1986
):
Calibration of the ruby pressure gauge to 800 kbar under quasi- hydrostatic conditions
.
J. Geophys. Res.
 ,
91
:
4673
4676
.
Marfunin
,
A.S.
(
1979
):
Spectroscopy, luminescence and radiation centres in minerals
 .
Berlin
:
Springer-Verlag
,
352
p.
Marfunin
,
A.S.
(ed.) (
1995
):
Methods and instrumentations /Adv. Mineral.
 ,
2
/.
Berlin
:
Springer-Verlag
,
441
p.
Markwort
,
L.
Kip
,
B.
Da Silva
,
E.
Rousell
,
B.
(
1995
):
Raman imaging of heterogeneous polymers: a comparison of global versus point illumination
.
Appl. Spectrosc.
 ,
49
:
1411
1430
.
Martini
,
D.
Fini
,
M.
Franchi
,
M.
De Pasquale
,
V.
Baccelli
,
B.
Gamberoni
,
M.
Tinti
,
A.
Taddei
,
P.
Giavaresi
,
G.
Ottani
,
V.
Raspanti
,
M.
Guizzardi
,
S.
Ruggeri
,
A.
(
2003
):
Detachment of titanium and fluorohydroxyapatite particles in unloaded endosseous implants
.
Biomaterials
 ,
24
:
1309
1316
.
Massonne
,
H.-J.
Nasdala
,
L.
(
2003
):
Characterization of an early metamorphic stage through inclusions in zircon in diamondiferous quartzofeldspatic rocks from the Erzgebirge, Germany
.
Am. Mineral.
 ,
88
:
883
889
.
Matson
,
D.W.
Sharma
,
S.K.
(
1985
):
Structures of sodium alumino- and gallosilicate glasses and their germanium analogs
.
Geochim. Cosmochim. Acta
 ,
49
:
1913
1924
.
McCann
,
L.I.
Trentelman
,
K.
Possley
,
T.
Golding
,
B.
(
1999
):
Corrosion of ancient Chinese bronze money trees studied by Raman microscopy
.
J. Raman Spectrosc.
 ,
30
:
121
132
.
McKeown
,
D.A.
Bell
,
M.I.
Etz
,
E.S.
(
1999
a
):
Raman spectra and vibrational analysis of the trioctahedral mica phlogopite
.
Am. Mineral.
 ,
84
:
970
976
.
McKeown
,
D.A.
Bell
,
M.I.
Etz
,
E.S.
(
1999
b
):
Vibrational analysis of the dioctahedral mica: 2M1 muscovite
.
Am. Mineral.
 ,
84
:
1041
1048
.
McMillan
,
P.F.
(
1984
):
Structure studies of silicate glasses and melts – applications and limitations of Raman spectroscopy
.
Am. Mineral.
 ,
69
:
622
644
.
McMillan
,
P.F.
(
1985
):
Infrared and Raman spectroscopy
. In
Kieffer
,
S.W.
Navrotsky
,
A.
(eds.):
Microscopic to macroscopic: atomic environments to mineral thermodynamics. /Rev. Mineral
 .,
14
/.
Washington (D.C.)
:
Mineral. Soc. Am.
,
9
63
.
McMillan
,
P.F.
Hofmeister
,
A.M.
(
1988
):
Infrared and Raman spectroscopy
. In
Hawthorne
,
F.C.
(ed.):
Spectroscopic methods in mineralogy and geology /Rev. Mineral
 .,
18
/.
Washington (D.C.)
:
Mineral. Soc. Am.
,
99
159
.
McMillan
,
P.F.
Piriou
,
B.
(
1983
):
Raman spectroscopy of calcium aluminate glasses and crystals
.
J. Non-Cryst. Solids
 ,
55
:
221
242
.
McMillan
,
P.F.
Piriou
,
B.
Navrotsky
,
A.
(
1982
):
A Raman spectroscopic syudy of glasses along the joins silica-calcium aluminate, silica-sodium aluminate, and silica-potassium aluminate
.
Geochim. Cosmochim. Acta
 ,
46
:
2021
2037
.
McMillan
,
P.F.
Putnis
,
A.
Carpenter
,
M.A.
(
1984
):
A Raman spectroscopic study of Al-Si ordering in synthetic magnesium cordierite
.
Phys. Chem. Miner.
 ,
10
:
256
260
.
Mernagh
,
T.P.
(
1991
):
Use of the laser Raman microprobe for discrimination amongst feldspar minerals
.
J. Raman Spectrosc.
 ,
22
:
453
457
.
Mernagh
,
T.P.
Hoatson
,
D.M.
(
1995
):
A laser-Raman microprobe study of platinum-group minerals from the Munni Munni layered intrusion, West Pilbara block, Western Australia
.
Can. Mineral.
 ,
33
:
409
417
.
Mernagh
,
T.P.
Hoatson
,
D.M.
(
1997
):
Raman spectroscopic study of pyroxene structures from the Munni Munni layered intrusion, Western Australia
.
J. Raman Spectrosc.
 ,
28
:
647
658
.
Mernagh
,
T.P.
Trudu
,
A.G.
(
1993
):
A laser Raman microprobe study of some geologically important sulphide minerals
.
Chem. Geol.
 ,
103
:
113
127
.
Mernagh
,
T.P.
Wilde
,
A.R.
(
1989
):
The use of the laser Raman microprobe for the determination of salinity in fluid inclusions
.
Geochim. Cosmochim. Acta
 ,
53
:
765
771
.
Milisenda
,
C.
Häger
,
T.
Henn
,
U.
:
Chatoyance und Asterismus in Feldspat aus Tansania
.
Z. Dtsch. Gemmol. Ges.
 ,
in print
.
Minser
,
D.G.
Walden
,
B.
White
,
W.B.
(
1984
):
Structure of alkali-zinc silicate glasses by Raman Spectroscopy
.
Am. Ceram. Soc. Bull.
 ,
67
:
47
49
.
Miyazaki
,
M.
Onose
,
H.
Moore
,
B.K.
(
2002
):
Analysis of the dentin–resin interface by use of laser Raman spectroscopy
.
Dental Mater.
 ,
18
:
576
580
.
Mukherjee
,
S.P.
Sharma
,
S.K.
(
1985
):
A comparative Raman study of the structures of conventional and gel-derived glasses in the SiO2-GeO2 system
.
J. Non-Cryst. Solids
 ,
71
:
317
325
.
Murata
,
K.
Kawakami
,
K.
Matsunaga
,
Y.
Yamashita
,
S.
(
1997
):
Determination of sulfate in brackish waters by laser Raman spectroscopy
.
Anal. Chim. Acta
 ,
344
:
153
157
.
Mysen
,
B.O.
(
1995
):
Structural behavior of Al3+ in silicate melts: In situ, high-temperature measurements as a function of bulk chemical composition
.
Geochim. Cosmochim. Acta
 ,
59
:
455
474
.
Mysen
,
B.O.
Virgo
,
D.
(
1985
):
Interaction between fluorine and silica in quenched melts on the joins SiO2- AlF3 and SiO2-NaF determined by Raman spectroscopy
.
Phys. Chem. Miner.
 ,
12
:
77
85
.
Nasdala
,
L.
(
1992
):
A Raman study on the so-called “pseudoapatite” from Halsbrücke near Freiberg/Saxony
.
Neues Jahrb. Mineral. Abh.
 ,
164
:
211
227
.
Nasdala
,
L.
Massonne
,
H.-J.
(
2000
):
Microdiamonds from the Saxonian Erzgebirge, Germany: in situ micro- Raman characterisation
.
Eur. J. Mineral.
 ,
12
:
495
498
.
Nasdala
,
L.
Pekov
,
I. V.
Witzke
,
T.
(
1993
):
Raman investigation of naturally occurring C14H10
.
Chem. Erde
 ,
53
:
59
69
.
Nasdala
,
L.
Irmer
,
G.
Wolf
,
D.
(
1995
):
The degree of metamictization in zircons: a Raman spectroscopic study
.
Eur. J. Mineral.
 ,
7
:
471
478
.
Nasdala
,
L.
Pidgeon
,
R.T.
Wolf
,
D.
(
1996
):
Heterogeneous metamictization of zircon on a microscale
.
Geochim. Cosmochim. Acta
 ,
60
:
1091
1097
.
Nasdala
,
L.
Witzke
,
T.
Ullrich
,
B.
Brett
,
R.
(
1998
):
Gordaite, [Zn4Na(OH)6(SO4)Cl·6H2O]: Second occurrence in the Juan de Fuca Ridge, and new data
.
Am. Mineral.
 ,
83
:
1111
1116
.
Nasdala
,
L.
Banerjee
,
A.
Häger
,
T.
Hofmeister
,
W.
(
2001
a
):
Laser-Raman micro-spectroscopy in mineralogical research
.
Microsc. Anal., Eur. ed.
 ,
70
:
7
9
.
Nasdala
,
L.
Wenzel
,
M.
Andrut
,
M.
Wirth
,
R.
Blaum
,
P.
(
2001
b
):
The nature of radiohaloes in biotite: experimental studies and modeling
.
Am. Mineral.
 ,
86
:
498
512
.
Nasdala
,
L.
Lengauer
,
C.L.
Hanchar
,
J.M.
Kronz
,
A.
Wirth
,
R.
Blanc
,
P.
Kennedy
,
A.K.
Seydoux-Guillaume
,
A.-M.
(
2002
):
Annealing radiation damage and the recovery of cathodoluminescence
.
Chem. Geol.
 ,
191
:
119
138
.
Nasdala
,
L.
Zhang
,
M.
Kempe
,
U.
Panczer
,
G.
Gaft
,
M.
Andrut
,
M.
Plötze
,
M.
(
2003
a
):
Spectroscopic methods applied to zircon
. In
Hanchar
,
J.M.
Hoskin
,
P.W.O.
(eds.):
Zircon /Rev. Mineral. Geochem.
 ,
53
/.
Washington (D.C.)
:
Mineral. Soc. Am.
,
427
467
.
Nasdala
,
L.
Brenker
,
F.E.
Glinnemann
,
J.
Hofmeister
,
W.
Gasparik
,
T.
Harris
,
J.W.
Stachel
,
T.
Reese
,
I.
(
2003
b
):
Spectroscopic 2D-tomography: Residual pressure and strain around mineral inclusions in diamonds
.
Eur. J. Mineral.
 ,
15
:
931
935
.
Nasdala
,
L.
Wopenka
,
B.
Lengauer
,
C.L.
(
2004
a
):
Transformation of SiO2 to the amorphous state by shearing at high pressure – Comment
.
Am. Mineral.
 ,
89
:
912
913
.
Nasdala
,
L.
Götze
,
J.
Hanchar
,
J.M.
Gaft
,
M.
Krbetschek
,
M.R.
(
2004
b
):
Luminescence techniques in Earth Sciences
. In
Beran
,
A.
Libowitzky
,
E.
(eds.):
Spectroscopic methods in mineralogy /EMU Notes Mineral.
 ,
6
/.
Budapest
:
Eötvös Univ. Press
,
43
91
.
Nasdala
,
L.
Hofmeister
,
W.
Harris
,
J.W.
Glinnemann
,
J.
:
Growth zoning and strain patterns inside diamond crystals as revealed by Raman tomography
.
Am. Mineral.
 ,
submitted
.
Nassau
,
K.
McClure
,
S.F.
Elen
,
S.
Shigley
,
J.E.
(
1997
):
Synthetic moissanite: a new diamond substitute
.
Gems Gemol.
 ,
33
:
260
175
.
Nelson
,
D.G.A.
Williamson
,
B.E.
(
1982
):
Low-temperature laser Raman spectroscopy of synthetic carbonated apatites and dental enamel
.
Aust. J. Chem.
 ,
35
:
715
727
.
Nemanich
,
R.J.
Solin
,
S.A.
(
1977
):
Observation of an anomalously sharp feature in the 2nd order Raman spectrum of graphite
.
Solid State Commun.
 ,
23
:
417
420
.
Nemanich
,
R.J.
Lucovsky
,
G
Solin
,
S.A.
(
1977
):
Infrared active optical vibrations of graphite
.
Solid State Commun.
 ,
23
:
117
120
.
Neumayr
,
P.
Groves
,
D.I.
Ridley
,
J.R.
Koning
,
C.D.
(
1993
):
Syn-amphibolite facies Archean lode gold mineralisation in the Mt. York District, Pilbara Block, Western Australia
.
Mineral. Deposita
 ,
28
:
457
468
.
Nwe
,
Y.Y.
Morteani
,
G.
(
1993
):
Fluid evolution in the H2O-CH4-CO2-NaCl system during emerald mineralization at Gravelotte, Murchison greenstone belt, Northeast Transvaal, South Africa
.
Geochim. Cosmochim. Acta
 ,
57
:
89
103
.
Olsen
,
S.N.
(
1987
):
The composition and role of the fluid in migmatites: a fluid inclusion study of the Front Range rocks
.
Contrib. Mineral. Petrol.
 ,
96
:
104
120
.
Orange
,
D.
Knittle
,
E.
Farber
,
D.
Williams
,
Q.
(
1996
):
Raman spectroscopy of crude oils and hydrocarbon inclusions: a feasability study
. In
Dyar
,
M.D.
McCammon
,
C.
Schaefer
,
M.W.
(eds.):
Mineral spectroscopy: A tribute to Roger G. Burns /Geochem. Soc. Spec. Publ.
 ,
5
/.
Houston (Tex.)
:
Geochem. Soc.
,
65
81
.
Orlov
,
R.Yu.
Uspenskaya
,
M.E.
Guseva
,
E.V.
(
1985
):
Application of Raman spectroscopy in mineralogy
 .
Moscow
:
Izd. Moscow Univ.
,
112
p (
in Russian
).
Ostroumov
,
M.
Fritsch
,
E.
Lasnier
,
B.
Lefrant
,
S.
(
1999
):
Spectres Raman des opales: aspect diagnostique et aide à la classification [Raman spectra of opals: diagnostic features and an aid to classification]
.
Eur. J. Mineral.
 ,
11
:
899
908
.
Pasteris
,
J.D.
Wopenka
,
B.
(
2002
):
Images of the Earth's earliest fossils
?
Nature
 ,
420
:
476
477
.
Pasteris
,
J.D.
Kuehn
,
C.A.
Bodnar
,
R.J.
(
1986
):
Applications of the laser Raman microprobe Ramanor U-1000 to hydrothermal ore deposits: Carlin as an example
.
Econ. Geol.
 ,
81
:
915
930
.
Pasteris
,
J.D.
Wopenka
,
B.
Freeman
,
J.J.
Young
,
V.L.
Brandon
,
H.J.
(
1999
):
Medical mineralogy as a new challenge to the geologist: Silicates in human mammary tissue
?
Am. Mineral.
 ,
84
:
997
1008
.
Perrin
,
C.
Smith
,
D.C.
(
2002
):
A preliminary Raman spectroscopic study of the earliest steps of diagenesis of mineral and organic material in living scleractinian corals
. In
Georaman 2002
  (
Prague
), Abstr.,
/Acta Univ. Carolinae, Geol.
,
46
/,
66
68
.
Philippot
,
P.
Selverstone
,
J.
(
1991
):
Trace-element-rich brines in eclogitic veins: implications for fluid composition and transport during subduction
.
Contrib. Mineral. Petrol.
 ,
106
:
417
430
.
Piermarini
,
G.J.
Block
,
S.
Barnett
,
J.D.
Forman
,
R.A.
(
1975
):
Calibration of the pressure dependence of the R1 ruby fluorescence line to 195 kbar
.
J. Appl. Phys.
 ,
46
:
2774
2780
.
Pilz
,
W.
Kriegsmann
,
H.
(
1987
):
The nature of the so-called “Raman background”
.
Z. Phys. Chem.
 ,
268
:
215
216
.
Pine
,
A.S.
Tannenwald
,
P.E.
(
1969
):
Temperature dependence of Raman linewidth and shift in α-quartz
.
Phys. Rev.
 ,
178
:
1424
1430
.
Pinet
,
M.
Smith
,
D.C.
(
1993
):
La microspectrométrie Raman des grenats X3Y2Z3O12: I. La série calcique naturelle ouvarovite-grossulaire-andradite
.
Schweiz. Mineral. Petrogr. Mitt.
 ,
73
:
21
40
.
Pinet
,
M.
Smith
,
D.C.
(
1994
):
La microspectrométrie Raman des grenats X3Y2Z3O12: II. La série alumineuse naturelle pyrope-almandin-spessartite
.
Schweiz. Mineral. Petrogr. Mitt.
 ,
74
:
161
179
.
Pinet
,
M.
Smith
,
D.C.
Lasnier
,
B.
(
1992
):
Utilité de la microsonde Raman pour l'identification non-destructive des gemmes, avec une sélection représentative de leurs spectres Raman
. In
Schubnel
,
H.J.
Smith
,
D.C.
(eds.):
La microsonde Raman en gemmologie /Rev. Gemmol., spécial hors série/
 .
Paris
:
Assoc. Fr. Gemmol.
,
61
.
Piriou
,
B.
Svoronos
,
D.R.
(
1985
):
Spectroscopic study of the Eu3+ ion in nitric acid aqueous solutions
.
J. Less-Common Met.
 ,
112
:
275
285
.
Piriou
,
B.
Videau
,
J.J.
Portier
,
J.
(
1981
):
Raman spectroscopic studies of aluminium fluoroberyllate glasses
.
J. Non-Cryst. Solids
 ,
46
:
105
110
.
Pironon
,
J.
(
1993
):
Estimation de la longueur de cha^ne des hydrocarbures des inclusion fluidespar spectrométrie Raman
.
C.R. Acad. Sci. Paris, Ser. II
 ,
316
:
1075
1082
.
Placzek
,
G.
(
1934
):
Rayleigh-Streuung und Raman-Effekt
. In
Marx
,
E.
(ed.):
Handbuch der Radiologie
 , Vol.
6
.
Leipzig
:
Akad. Verlagsges
,
205
374
.
Raman
,
C.V.
Krishnan
,
K.S.
(
1928
):
A new type of secondary radiation
.
Nature
 ,
121
:
501
502
.
Raman
,
C.V.
Nedungadi
,
T.M.K.
(
1940
):
The α-β transformation of quartz
.
Nature
 ,
145
:
147
.
Rankin
,
A.H.
(
2004
):
Fluid inclusions – Tools for geological investigations
. In
Selley
,
R.C
Cocks
,
R.
Plimer
,
I.
(eds.):
Encyclopedia of geology
 ,
1–4
.
Oxford
:
Elsevier
,
in press
.
Refait
,
P.
Memet
,
J.-B.
Bon
,
C.
Sabot
,
R.
Génin
,
J.-M.
(
2003
):
Formation of the Fe(II)–Fe(III) hydroxysulphate green rust during marine corrosion of steel
.
Corros. Sci.
 ,
45
:
833
845
.
Reutel
,
C.
(
1992
):
Krustenfluide in Gesteinen und Lagerstätten am Westrand der Böhmischen Masse
.
Göttinger Arb. Geol. Paläontol.
 ,
53
.
Reynard
,
B.
Remy
,
C.
Takir
,
F.
(
1997
):
High-pressure Raman spectroscopic study of Mn2GeO4, Ca2GeO4, Ca2SiO4, and CaMgGeO4 olivines
.
Phys. Chem. Miner.
 ,
24
:
77
84
.
Rodgers
,
K.A.
(
1993
):
Routine identification of aluminium hydroxide polymorphs with the laser Raman microprobe
.
Clay Miner.
 ,
28
:
85
99
.
Rondeau
,
B.
Smith
,
D.C.
(
2002
):
Non-destructive in situ identification of gemstones on two “Navaratna” jewels by MRM (Mobile Raman Microscopy)
. In “
Art 2002”, 7th Int. Conf. on non-destructive testing and microanalysis for the diagnostics and conservation of the cultural and environmental heritage (Antwerp)
 , abstr.,
Univ. of Antwerp
,
223
.
Rosasco
,
G.J.
Roedder
,
E.
(
1979
):
Application of a new Raman microprobe spectrometer to nondestructive analysis of sulfate and other ions in individual phases in fluid inclusions in minerals
.
Geochim. Cosmochim. Acta
 ,
43
:
1907
1915
.
Rosasco
,
G.J.
Etz
,
E.S.
Cassatt
,
W.A.
(
1975
a
):
The analysis of discrete fine particles by Raman spectroscopy
.
Appl. Spectrosc.
 ,
29
:
396
404
.
Rosasco
,
G.J.
Roedder
,
E.
Simmons
,
J.H.
(
1975
b
):
Laser activated Raman spectroscopy for nondestructive partial analysis of individual phases in fluid inclusions in minerals
.
Science
 ,
190
:
557
560
.
Rousseau
,
D.L.
Baumann
,
R.P.
Porto
,
S.P.S.
(
1981
):
Normal mode determination in crystals
.
J. Raman Spectrosc.
 ,
10
:
253
290
.
Sato
,
R.K.
McMillan
,
P.F.
(
1987
):
An infrared and Raman study of the isotopic species of α-quartz
.
J. Phys. Chem.
 ,
91
:
394
398
.
Schiferl
,
D.
Nicol
,
M.
Zaug
,
J.M.
Sharma
,
S.K.
Cooney
,
T.F.
Wang
,
S.Y.
Anthony
,
T.R.
Fleischer
,
J.F.
(
1997
):
The diamond 13C/12C Isotope Raman pressure sensor system for high-temperature/pressure diamond-anvil cells with reactive samples
.
J. Appl. Phys.
 ,
82
:
3256
3265
.
Schmetzer
,
K.
Hänni
,
H.
Bernhardt
,
H.-J.
Schwarz
,
D.
(
1996
):
Trapiche rubies
.
Gems Gemol.
 ,
32
:
242
250
.
Schmetzer
,
K.
Kiefert
,
L.
Bernhardt
,
H.-J.
Beili
,
Z.
(
1997
):
Characterisation of Chinese hydrothermal synthetic emerald
.
Gems Gemol.
 ,
33
:
277
291
.
Schmidt
,
W.
(
1994
):
Optische Spektroskopie: Eine Einführung für Naturwissenschafter und Techniker
 .
Weinheim
:
VCH
.
Schmidt
,
C.
Ziemann
,
M.A.
(
2000
):
In-situ Raman spectroscopy of quartz: A pressure sensor for the hydrothermal diamond-anvil cell experiments at elevated temperatures
.
Am. Mineral.
 ,
85
:
1725
1734
.
Schopf
,
J.W.
Kudryavtsev
,
A.B.
Agresti
,
D.G.
Wdowiak
,
T.J.
Czaja
,
A.D.
(
2002
):
Laser-Raman imagery of Earth's earliest fossils
.
Nature
 ,
416
:
73
76
.
Schroeder
,
P.A.
(
1990
):
Far infrared, X-ray powder diffraction, and chemical investigation of potassium micas
.
Am. Mineral.
 ,
75
:
983
991
.
Schrötter
,
H.W.
Klöckner
,
H.W.
(
1979
):
Raman scattering cross-sections in gases and liquids
. In
Weber
,
A.
(ed.):
Raman spectroscopy of gases and liquids
 .
Berlin
:
Springer-Verlag
,
123
166
.
Scott
,
J.F.
Porto
,
S.P.S.
(
1967
):
Longitudinal and transversal optical lattice vibrations in quartz
.
Phys. Rev.
 ,
161
:
903
910
.
Seitz
,
J.C.
Pasteris
,
J.D.
Wopenka
,
B.
(
1987
):
Characterization of CO2-CH4-H2O fluid inclusions by microthermometry and laser Raman microprobe spectroscopy: inferences for clathrate and fluid equilibria
.
Geochim. Cosmochim. Acta
 ,
51
:
1651
1664
.
Seitz
,
J.C.
Pasteris
,
J.D.
Chou
,
I.M.
(
1993
):
Raman spectroscopic characterization of gas mixtures. 1. Quantitative composition and pressure determination of CH4, N2 and their mixtures
.
Am. J. Sci.
 ,
293
:
297
321
.
Seitz
,
J.C.
Pasteris
,
J.D.
Chou
,
I.M.
(
1996
):
Raman spectroscopic characterization of gas mixtures. 2. Quantitative composition and pressure determination of the CO2-CH4 system
.
Am. J. Sci.
 ,
296
:
577
600
.
Seydoux-Guillaume
,
A.-M.
Wirth
,
R.
Nasdala
,
L.
Gottschalk
,
M.
Montel
,
J.M.
Heinrich
,
W.
(
2002
):
An XRD, TEM and Raman study of experimentally annealed natural monazite
.
Phys. Chem. Miner.
 ,
29
:
240
253
.
Sharma
,
S.K.
(
1972
):
Raman study of ferric chloride solutions and hydrated melts
.
J.Chem. Phys.
 ,
60
:
1368
1375
.
Sharma
,
S.K.
Reed
,
M.D.
(
1976
):
Raman study of zincate ions in concentrated alkaline solutions
.
J. Inorg. Nucl. Chem.
 ,
38
:
1971
1972
.
Sharma
,
S.K.
Simons
,
B.
(
1981
):
Raman study of crystalline polymorphs and glasses of spodumen composition quenched from various pressures
.
Am. Mineral.
 ,
66
:
118
126
.
Sharma
,
S.K.
Angel
,
S.M.
Ghosh
,
M.
Hubble
,
H.W.
Lucey
,
P.G.
(
2002
):
Remote pulsed laser Raman spectroscopy system for mineral analysis on planetary surfaces to 66 meters
.
Appl. Spectrosc.
 ,
56
:
699
705
.
Sharma
,
S.K.
Lucey
,
P.G.
Ghosh
,
M.
Hubble
,
H.W.
Horton
,
K.
(
2003
):
Stand-off Raman spectroscopic detection of minerals on planetary surfaces
.
Spectrochim. Acta A
 ,
59
:
2391
2407
.
Shen
,
A.H.
Bassett
,
W.A.
Chou
,
I-M
. (
1992
):
Hydrothermal studies in a diamond-anvil cell: Pressure determination using the equation of state of H2O
. In
Syono
,
Y.
Manghnani
,
M.H.
(eds.):
High-pressure research: Application to Earth and planetary sciences /Geophys. Monogr.
 ,
67
/.
Washington (D.C.)
:
Am. Geophys. Soc., Terra Scientific
,
61
68
.
Shim
,
S.-H.
Duffy
,
T.S.
(
2001
):
Raman spectroscopy of Fe2O3 to 62 GPa
.
Am. Mineral.
 ,
87
:
318
326
.
Shroder
,
R. E.
Nemanich
,
R.J.
Glass
,
J.T.
(
1990
):
Analysis of the composite structures in diamond thin films by Raman spectroscopy
.
Phys. Rev., B
 ,
41
:
3738
3745
.
Siebert
,
H.
(
1966
):
Anwendungen der Schwingungsspektroskopie in der anorganische Chemie /Anorg. allg Chem
.
Einzeldarst
 ,
7
/.
Berlin
:
Springer-Verlag
,
210
p.
Siemann
,
M.G.
Ellendorff
,
B.
(
2001
):
The composition of gases in fluid inclusions of Late Permian (Zechstein) marine evaporites in northern Germany
.
Chem. Geol.
 ,
173
:
31
44
.
Silva
,
C.C.
Pinheiro
,
A.G.
Miranda
,
M.A.R.
Góes
,
J.C.
Sombra
,
A.S.B.
(
2003
):
Structural properties of hydroxyapatite obtained by mechanosynthesis
.
Solid State Sci.
 ,
5
:
553
558
.
Silvé
,
C
Lopez
,
E.
Vidal
,
B.
Smith
,
D.C.
Camprasse
,
S.
Camprasse
,
G.
Couly
,
G.
(
1992
):
Nacre initiates biomineralization by human osteoblasts maintained in vitro
.
Calcif. Tissue Int.
 ,
51
:
363
369
.
Smekal
,
A.
(
1923
):
Zur Quantentheorie der Dispersion
.
Naturwissenschaften
 ,
11
:
873
875
.
Smith
,
D.C.
(
1984
):
Coesite in clinopyroxene in the Caledonides
.
Nature
 ,
310
:
641
644
.
Smith
,
D.C.
(
1996
):
The importance of using a half-wave plate for the Raman Spectroscopy of minerals which cannot be rotated under the laser beam, especially large objects of gemmological or archaeological interest
. In
Georaman-96 (Nantes), Abstr., Terra Nova
 ,
8
(Abstr. Suppl.
2
):
23
24
.
Smith
,
D.C.
(
1999
):
Letting loose a laser: MRM (Mobile Raman Microscopy) for archeometry and ethnomineralogy in the next millennium
.
Mineral. Soc. (London) Bull., Invited Feature Article
 , No.
125
:
3
8
.
Smith
,
D.C.
(
2000
):
Pigments rouges et bleus sur cinq oeuvres d'Amérique: Analyse non-destructive par MRM (Microscopie Raman Mobile)
.
Techne
 ,
11
:
69
83
.
Smith
,
D.C.
(
2001
a
):
Recent MRM (Mobile Raman Microscope) analytical operations in situ in four French museums
. In
Colloquium: “Raman spectroscopy in archaeology and art history”
 .
London
:
The British Museum
.
Smith
,
D.C.
(
2001
b
):
Semi-quantitative chemical analysis of jade & garnet jewels by non-destructive Raman microscopy
. In
Colloquium: “Raman spectroscopy in archaeology and art history”
 .
London
:
The British Museum
.
Smith
,
D.C.
(
2002
a
):
Archaeoraman and mobile Raman spectroscopy (MRM): from pigments in aerial wall-paintings to gemstones in submarine archaeometry
. In
Georaman 2002
  (
Prague
), Abstr.
/Acta Univ. Carolinae, Geol.
,
46
/,
84
86
.
Smith
,
D.C.
(
2002
b
):
Semi-quantitative chemical analysis of garnet & jade in mounted jewels by non-destructive Raman microscopy: recent progress with the analytical method
. In
Georaman 2002
  (
Prague
), Abstr. /
Acta Univ. Carolinae, Geol.
,
46
/,
87
89
.
Smith
,
D.C.
(
2003
):
In situ mobile subaquatic archaeometry evaluated by non-destructive Raman microscopy of gemstones lying under impure waters
.
Spectrochim. Acta A
 ,
59
:
2353
2369
.
Smith
,
D.C.
:
Jewellery and precious stones
. In
Edwards
,
H.G.M.
Chalmers
,
J.M
: (eds.):
Raman spectroscopy in archaeology and art history /RSC Anal. Spectrosc. Monogr./
 .
Cambridge
:
R. Soc. Chem.
,
in print (a)
.
Smith
,
D.C.
:
Mesoamericanjade
. In
Edwards
,
H.G.M.
Chalmers
,
J.M
: (eds.):
Raman spectroscopy in archaeology and art history/RSC Anal. Spectrosc. Monogr./.
 
Cambridge
:
R. Soc. Chem.
,
in print (b)
.
Smith
,
D.C.
Bouchard
,
M.
(
2000
):
Petroraman et Gemmoraman
.
Archéométrie, Doss. Archéol.
 ,
253
:
54
59
.
Smith
,
D.C.
Bouchard
,
M.
(
2002
):
Bronze disease: non-destructive distinction of three polymorphs of Cu2Cl(OH)3 by Raman Microscopy (RM)
. In
“Art 2002”, 7th Int. Conf. on non-destructive testing and microanalysis for the diagnostics and conservation of the cultural and environmental heritage (Antwerp), abstr.
 ,
Univ. of Antwerp
,
227
.
Smith
,
D.C.
Boyer
,
H.
(
1985
):
Raman microprobe fingerprinting of ordered and disordered pyroxenes in the system diopside-omphacite-jadeite
.
Second Int. Eclogite Conf., Abstr., Terra Cognita
 ,
5
:
429
.
Smith
,
D.C.
Carabatos-Nédelec
,
C
. (
2001
):
Raman spectroscopy applied to crystals: phenomena and principles, concepts and conventions
. In
Lewis
,
I.R.
Edwards
,
H.G.M.
(eds.):
Handbook of Raman spectroscopy. From the research laboratory to the process line. /Pract. Spectrosc. Ser
 .,
28
/.
New York (N.Y.)
:
Marcel Dekker
,
349
422
.
Smith
,
D.C.
Edwards
,
H.G.M.
(
1998
):
A wavenumber-searchable tabular indexed catalogue for “Archaeoraman®”: Raman spectra of geomaterials and biomaterials of interest in archaeology (sensu lato)
. In
Heyns
,
A.M.
(ed.):
ICORS Capetown'
 
98
.
Chichester
:
Wiley
,
510
511
.
Smith
,
D.C.
Gendron
,
F.
(
1997
):
Archaeometric application of the Raman microprobe to the non-destructive identification of two pre-Columbian ceremonial polished “greenstone” axe heads from Mesoamerica
.
J. Raman Spectrosc.
 ,
28
:
731
738
.
Smith
,
D.C.
Périn
,
P.
(
2003
):
Non-destructive in situ MRM (Mobile Raman Microscope) semi-quantitative chemical analysis of garnets in Merovingian cloisonné style jewellery from Vicq, France
. In
Int. Congr. Application of Raman spectroscopy in art and archaeology (Ghent), Abstr
 ,
62
.
Smith
,
D.C.
Pinet
,
M.
(
1989
):
A method for single variable plotting of multi-dimensional chemical data for comparing Raman wavenumbers with chemical variations in solid-solutions
. In
Georaman-89
  (
Paris
), Abstr., /
Assoc. Natl. Rech. Tech. Spec. Publ.
17
/,
24
.
Smith
,
D.C.
Rondeau
,
B
(
2001
):
Non-destructive mineralogical analysis of precious stones in situ under thick glass by MRM (Mobile Raman Microscopy)
. In
Congress Archéométrie 2001 (La Rochelle), Abstr.
 ,
101
.
Smith
,
D.C.
Robin
,
S.
(
1997
):
Early-Roman Empire intaglios from “rescue excavations” in Paris: an application of the Raman Microprobe to the non-destructive characterisation of archaeological objects
.
J. Raman Spectrosc.
 ,
28
:
189
193
.
Smith
,
D.C.
Pinet
,
M.
Boyer
,
H.
(
1988
):
Raman spectra of ternary pyralspite garnets and their preliminary calibration for the chemical analysis of synthetic or natural micron-size crystals
.
Terra Cognita
 ,
8
:
77
78
.
Smith
,
D.C.
Pinet
,
M.
Yang
,
J.
(
1989
):
A Raman microspectrometric calibration for the chemical analysis of pyroxenes in the system diopside-hedenbergite-aegirine
. In
Georaman-89
  (
Paris
), Abstr.,
Assoc. Natl. Rech. Techn. Special Publ.
,
25
.
Smith
,
D.C.
Bouchard
,
M.
Lorblanchet
,
M.
(
1999
a
):
An initial Raman microscopic investigation of Prehistoric Rock Art in caves of the Quercy District S.W. France
.
J. Raman Spectrosc.
 ,
30
:
347
354
.
Smith
,
D.C.
Carabatos-Nedelec
,
C.
Bouchard
,
M.
(
1999
b
):
Vitroraman: establishing a database on the Raman spectra of pigments on and in stained glass
. In
Georaman'99
  (
Valladolid
), Abstr.
/Univ. Valladolid Press Spec. Publ.
/,
36
37
.
Smith
,
D.C.
Périn
,
P.
Kazanski
,
M.
Gabuev
,
T.
(
2003
):
Andradite-rich pentary garnet compositions deduced from non-destructive in situ MRM (Mobile Raman Microscope) semi-quantitative chemical analysis of Vth century cloisonné garnets from an Alan warrior burial mound at Brut, Northern Ossetia, Russian Federation
. In
Int. Congr. Application of Raman spectroscopy in art and archaeology (Ghent), Abstr.
 ,
39
.
Sobolev
,
N.V.
Shatsky
,
V.S.
(
1990
):
Diamond inclusions in garnet from metamorphic rocks: a new environment for diamond formation
.
Nature
 ,
343
:
742
746
.
Sobolev
,
N.V.
Fursenko
,
B.A.
Goryainov
,
S.V.
Shu
,
J.F.
Hemley
,
R.J.
Mao
,
H.K.
Boyd
,
F.R.
(
2000
):
Fossilized high pressure from the Earth's deep interior: The coesite-in-diamond barometer
.
Proc. Natl. Acad. Sci.
 
USA
,
97
:
11875
11879
.
Solin
,
S.A.
Ramdas
,
A.K.
(
1970
):
Raman spectrum of diamond
.
Phys. Rev., B
 ,
1
:
1687
1698
.
Sprenger
,
D.
(
1996
):
Spektroskopische Untersuchungen und Berechnungen zur Struktur anorganischer Gläser
.
Ph.D. thesis
 ,
Johannes Gutenberg-Universität
,
Mainz, Germany
.
Sprenger
,
D.
Bach
,
H.
Meisel
,
W.
Gütlich
,
P.
(
1993
):
Discrete bond model (DBM) of sodium silicate glasses derived from XPS, Raman and NMR measurements
.
J. Non-Cryst. Solids
 ,
159
:
187
203
.
Splett
,
A.
Splett
,
C.
Pilz
,
W.
(
2000
):
Dynamics of the Raman background decay
.
J. Raman Spectrosc.
 ,
28
:
481
485
.
Stephenson
,
D.A.
(
1974
):
Raman cross-sections of selected hydrocarbons and freones
.
J. Quant. Spectrosc. Radiat. Transfer
 ,
14
:
1291
1301
.
Sum
,
A.K.
Burruss
,
R.C.
Sloan
,
F.D.
(
1997
):
Measurement of clathrate hydrates via Raman spectroscopy
.
J. Phys. Chem. B
 ,
101
:
7371
7377
.
Sweegers
,
C.
de Coninck
,
H.C.
Meekes
,
H.
van Enckevort
,
W.J.P.
Hiralal
,
I.D.K.
Rijkeboer
,
A.
(
2001
):
Morphology, evolution and other characteristics of gibbsite crystals grown from pure and impure aqueous sodium aluminate solutions
.
J. Crystal Growth
 ,
233
:
567
582
.
Syme
,
R.W.G.
Lockwood
,
D.J.
Kerr
,
J.
(
1977
):
Raman spectrum of synthetic zircon (ZrSiO4) and thorite (ThSiO4)
.
J. Phys. C. Solid State Phys.
 ,
10
:
1335
1348
.
Sze
,
S.K.
Siddique
,
N.
Sloan
,
J.J.
Escribano
,
R.
(
2001
):
Raman spectroscopic characterization of carbonaceous aerosols
.
Atmos. Environ.
 ,
35
:
561
568
.
Taddei
,
P.
Tinti
,
A.
Reggiani
,
M.
Monti
,
P.
Fagnano
,
C.
(
2003
):
In vivo bioactivity of titanium and fluorinated apatite coatings for orthopaedic implants: a vibrational study
.
J. Mol. Struct.
 ,
651–653
:
427
431
.
Thomas
,
A.V.
Pasteris
,
J.D.
Bray
,
C.J.
Spooner
,
E.T.C.
(
1990
):
H2O-CH4-NaCl-CO2 inclusions from the footwall contact of the Tanco granitic pegmatite: Estimates of internal pressure and composition from microthermometry, laser Raman spectroscopy, and gas chromatography
.
Geochim. Cosmochim. Acta
 ,
54
:
559
573
.
Tlili
,
A.
(
1990
):
Spectroscopies Raman et infrarouge, et cristallochimie des micas naturels et synthétiques dans le système: Li-Na-K-Ca-Mg-Fe-Mn-Al-Ga-Si-Ge-O-H-D
.
Ph.D. thesis
 ,
Muséum National d'Histoire Naturelle
,
Paris
,
299
p.
Tlili
,
A.
Smith
,
D.C.
(
1996
):
Empirical assignment of Raman spectra of synthetic Na-Mg-Al-Si-micas by comparison with their Ca-, Ga-, or Ge-equivalents: II. High wavenumber regions 550–1250 & 3550–3750 cm-1
. In
Georaman-96 (Nantes), Abstr., Terra Nova
 ,
8
(Abstr. Suppl.
2
):
12
.
Tlili
,
A.
Smith
,
D.C.
Beny
,
J.-M.
Boyer
,
H.
(
1989
):
A Raman microprobe study of natural micas
.
Mineral. Mag.
 ,
53
:
165
179
.
Touray
,
J.-C.
Bény
,
C.
Dubessy
,
J.
Guilhaumou
,
N.
(
1985
):
Microcharacterization of fluid inclusions in minerals by Raman microprobe
.
Scanning Electron Microsc.
 ,
1
:
103
118
.
Tuinstra
,
F.
Koenig
,
J.L.
(
1970
):
Raman spectrum of graphite
.
J. Chem. Phys.
 ,
53
:
1126
1130
.
Urmos
,
J.
Sharma
,
S.K.
Mackenzie
,
F.T.
(
1991
):
Characterization of some biogenic carbonates with Raman spectroscopy
.
Am. Mineral.
 ,
76
:
641
646
.
van denKerkhof
,
A.M.
(
1988
):
The system CO2-CH4-N2 in fluid inclusions: theoretical modelling and geological applications
.
Ph.D. thesis
 ,
Free Univ., Amsterdam
,
The Netherlands
.
van den Kerkhof
,
A.M.
Olsen
,
S.N.
(
1990
):
A natural example of superdense CO2 inclusions: Microthermometry and Raman analysis
.
Geochim. Cosmochim. Acta
 ,
54
:
895
901
.
van den Kerkhof
,
A.M.
Kisch
,
H.J.
(
1993
):
CH4-rich inclusions from quartz veins in the Valley-and- Ridge province and the anthracite fields of the Pennsylvania Appalachians – Reply
.
Am. Mineral.
 ,
78
:
220
224
.
van Zuilen
,
M.A.
Lepland
,
A.
Arrhenius
,
G.
(
2002
):
Reassessing the evidence for the earliest traces of life
.
Nature
 ,
418
:
627
630
.
van Zuilen
,
M.A.
Lepland
,
A.
Teranes
,
J.
Finarelli
,
J.
Wahlen
,
M.
Arrhenius
,
G.
(
2003
):
Graphite and carbonates in the 3.8 Ga old Isua Supracrustal Belt, southern West Greenland
.
Precambrian Res.
 ,
126
:
331
348
.
Vapnik
,
Y.
Moroz
,
I.
(
2002
):
Compositions and formation conditions of fluid inclusions in emerald from the Maria deposit (Mozambique)
.
Mineral. Mag.
 ,
66
:
201
213
.
Vénec-Péyré
,
M.T.
Jaeschke-Boyer
,
H.
(
1979
):
Intérêt de la microsonde moléculaire à laser Mole en systématique: étude du Foraminifère Rosalila posidonicola (Colom)
,
C.R. Acad. Sci. Paris
 ,
288
:
819
821
.
Verma
,
P.
Abbi
,
S.C.
Jain
,
K.P.
(
1995
):
Raman-scattering probe of anharmonic effects in GaAs
.
Phys. Rev., B
 ,
51
:
16660
16667
.
Vidano
,
R.P.
Fischbach
,
D.B.
Willis
,
L.J.
Loehr
,
T.M.
(
1981
):
Observation of Raman band shifting with excitation wavelength for carbons and graphites
.
Solid State Commun.
 ,
39
:
341
344
.
Videau
,
J.J.
Portier
,
J.
Piriou
,
B.
(
1981
):
Raman spectroscopic studies of fluorophosphate glasses
.
J. Non-Cryst. Solids
 ,
48
:
385
392
.
Wallwork
,
K.
Kolitsch
,
U.
Pring
,
A.
Nasdala
,
L.
(
2002
):
Decrespignyite-(Y), a new copper yttrium rare earth carbonate chloride hydrate from Paratoo, South Australia
.
Mineral. Mag.
 ,
66
:
181
188
.
Wang
,
A.
Haskin
,
L.A.
(
2000
):
Development of a flight Raman spectrometer for the “Athena” rover scientific instrument payload for Mars Surveyor missions
.
Inst. Phys. Conf. Ser.
 ,
165
:
103
104
.
Wang
,
A.
Dhamelincourt
,
P.
Dubessy
,
J.
Guerard
,
D.
Landais
,
P.
Lelaurian
,
M.
(
1989
):
Characterization of graphite alteration in an uranium deposit by micro-Raman spectroscopy, X-ray diffraction, transmission electron microscopy and scanning electron microscopy
.
Carbon
 ,
27
:
209
218
.
Wang
,
A.
Jolliff
,
B.L.
Haskin
,
L.A.
(
1996
):
Raman system for robotic exploration on planets
.
Terra Abstr., Abstr. Suppl. 2, to Terra Nova
 ,
8
:
25
.
Wang
,
A.
Jolliff
,
B.L.
Haskin
,
L.A.
Kuebler
,
K.E.
Viscupic
,
K.M.
(
2001
):
Characterization and comparison of structural and compositional features of planetary quadrilateral pyroxenes by Raman spectroscopy
.
Am. Mineral.
 ,
86
:
790
806
.
Wang
,
A.
Haskin
,
L.A.
Lane
,
A.L.
Wdowiak
,
T.J.
Squyres
,
S.W.
Wilson
,
R.L.
Hovland
,
L.E.
Manatt
,
K.S.
Raouf
,
N.
Smith
,
C.D.
(
2003
):
Development of the Mars microbeam Raman spectrometer (MMRS)
.
J. Geophys. Res.
 ,
108(E1)
:
5005
.
Wang
,
Z.
Huang
,
X.
Xue
,
R.
Chen
,
L.
(
1998
):
Dispersion effects of Raman lines in carbon
.
J. Appl. Phys.
 ,
84
:
227
231
.
Weeny
,
R.M.
(
1963
):
Symmetry group and introduction to group theory and its applications
 .
Oxford
:
Pergamon Press
.
Weidlein
,
J.
Müller
,
U.
Dehnicke
,
K.
(
1988
):
Schwingungsspektroskopie. Eine Einführung
 .
Stuttgart
:
Georg Thieme Verlag
,
228
p.
Williams
,
Q.
Knittle
,
E.
(
2003
):
Structural complexity in carbonatite liquid at high pressures
.
Geophys. Res. Lett.
 ,
30
:
22
1
4
.
Williams
,
Q.
Collerson
,
B.
Knittle
,
E.
(
1992
):
Vibrational spectra of magnesite (MgCO3) and calcite-III at high pressure
.
Am. Mineral.
 ,
77
:
1158
1165
.
Wilson
,
E.B.
Decius
,
J.C.
Cross
,
P.C.
(
1955
):
Molecular vibrations
 .
New York (N.Y.)
:
McGraw-Hill
.
Winter
,
C.J.
Roberts
,
S.
(
1993
):
Laser Raman studies of low temperature hydrates produced in the highly saline fluids
.
Arch. Mineral.
 ,
49
:
246
.
Witzke
,
T.
Zhen
,
S.
Seff
,
K.
Doering
,
T.
Nasdala
,
L.
Kolitsch
,
U.
(
2001
):
Ronneburgite, K2MnV4O12, a new mineral from Ronneburg, Thuringia, Germany: Description and crystal structure
.
Am. Mineral.
 ,
86
:
1081
1086
.
Wooster
,
W.A.
(
1973
):
Tensors and group theory for the physical properties of crystals
 .
Oxford
:
Clarendon Press
.
Wopenka
,
B.
Pasteris
,
J.D.
(
1986
):
Limitations to quantitative analysis of fluid inclusions in geological samples by laser Raman microprobe spectroscopy
.
Appl. Spectrosc.
 ,
40
:
144
151
.
Wopenka
,
B.
Pasteris
,
J.D.
(
1987
):
Raman intensities and detection limits of geologically relevant gas mixtures for a laser Raman microprobe
.
Anal. Chem.
 ,
59
:
2165
2170
.
Wopenka
,
B.
Pasteris
,
J.J.
(
1993
):
Structural characterization of kerogens to granulite-facies graphite: applicability of Raman microprobe spectroscopy
.
Am. Mineral.
 ,
78
:
533
557
.
Wopenka
,
B.
Pasteris
,
J.D.
Freeman
,
J.J.
(
1990
):
Analysis of individual fluid inclusions by Fourier transform infrared and Raman micro-spectroscopy
.
Geochim. Cosmochim. Acta
 ,
54
:
519
533
.
Wopenka
,
B.
Freeman
,
J.J.
Grew
,
E.
(
1999
):
Raman spectroscopic identification of B-free and B-rich kornerupine (prismatine)
.
Am. Mineral.
 ,
84
:
550
554
.
Xu
,
J.
Mao
,
H.K.
(
2000
):
Moissanite: A window for high-pressure experiments
.
Science
 ,
290
:
783
785
.
<