Mineralogy and mineral chemistry of quartz: A review

Abstract

Quartz (trigonal, low-temperature α-quartz) is the most important polymorph of the silica (SiO₂) group and one of the purest minerals in the Earth crust. The mineralogy and mineral chemistry of quartz are determined mainly by its defect structure. Certain point defects, dislocations and micro-inclusions can be incorporated into quartz during crystallisation under various thermodynamic conditions and by secondary processes such as alteration, irradiation, diagenesis or metamorphism. The resulting real structure is a fingerprint of the specific physicochemical environment of quartz formation and also determines the quality and applications of SiO₂ raw materials. Point defects in quartz can be related to imperfections associated with silicon or oxygen vacancies (intrinsic defects), to different types of displaced atoms, and/or to the incorporation of foreign ions in lattice sites and interstitial positions (extrinsic defects). Due to mismatch in charges and ionic radii only a limited number of ions can substitute for Si⁴⁺ in the crystal lattice or can be incorporated in interstitial positions. Therefore, most impurity elements in quartz are present at concentrations below 1 ppm. The structural incorporation in a regular Si⁴⁺ lattice site has been proven for Al³⁺, Ga³⁺, Fe³⁺, B³⁺, Ge⁴⁺, Ti⁴⁺, P⁵⁺ and H⁺, of which Al³⁺ is by far the most common and typically the most abundant. Unambiguous detection and characterisation of defect structures in quartz are a technical challenge and can only be successfully realised by a combination of advanced analytical methods such as electron paramagnetic resonance (EPR) spectroscopy, cathodoluminescence (CL) microscopy and spectroscopy as well as spatially resolved trace-element analysis such as laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) and secondary-ion mass spectrometry (SIMS). The present paper presents a review of the state-of-the-art knowledge concerning the mineralogy and mineral-chemistry of quartz and illustrates important geological implications of the properties of quartz.

Introduction

The SiO₂ system is very complex. Though it has a simple chemical formula, SiO₂, at least 15 modifications or polymorphs are known, i.e. mineral phases with the same stoichiometric composition but different crystal structures. Quartz is the most important silica polymorph in nature, and occurs as a common constituent of magmatic, metamorphic and sedimentary rocks. In addition, quartz represents an economically important silica raw material. Both single crystals and polycrystalline quartz material are used in industry, for example, as high-purity quartz crystals or sands, refractory materials, or as an ore for silicon metal. For certain highly advanced applications, synthetic quartz crystals or SiO₂ materials are necessary due to increasing quality requirements not being easily met by silica raw materials (Götze and Möckel, 2012; Müller et al., 2012, 2015).

The growing worldwide consumption of a wide range of high-technology applications means there is an increasing demand for high-quality natural quartz raw materials. Such high-purity quartz is defined as quartz containing less than 50 ppm of impurity elements, specifically < 30 ppm Al, < 10 ppm Ti, < 8 ppm Na and K, < 5 ppm Li and Ca, < 3 ppm Fe, < 2 ppm P and < 1 ppm B (Harben, 2002; Müller et al., 2012). The type of impurities controlling the quality of high-purity quartz strongly influences the processing and potential use of the raw materials. Therefore, exploration for high-purity quartz deposits requires a thorough knowledge of the factors influencing the petrological and chemical properties of quartz. Several studies have shown that granitic pegmatites provide potentially the best sources of high-purity quartz (Larsen, 2000; Müller et al., 2007, 2012, 2015).

The mineralogy and mineral chemistry of quartz are determined mainly by its defect structure. Certain point defects, dislocations, planar defects and micro-inclusions of fluids/melts and minerals can be incorporated into quartz during crystallisation under diverse thermodynamic conditions and by secondary processes, such as alteration, irradiation, diagenesis or metamorphism (Götze, 2009). The resulting real structure is not only a fingerprint of the specific physicochemical environment of quartz formation and secondary processes, but also determines the quality of SiO₂ raw materials. Therefore, quartz represents an important industrial raw material, and also serves as an important mineralogical and geochemical indicator for geological and ore-forming processes (e.g. Götze and Möckel, 2012 and references therein).

Quartz is one of the purest minerals, nevertheless minor impurities and defects have potentially significant importance and information, particularly regarding the mineral genesis. However, the detection and characterisation of such defects is generally a technical challenge and can only be successfully accomplished by a combination of advanced analytical methods, such as electron paramagnetic resonance (EPR) spectroscopy, cathodoluminescence (CL) microscopy and spectroscopy, synchrotron X-ray absorption spectroscopy (XAS) together with spatially resolved trace-element analysis such as laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) and secondary-ion mass spectrometry (SIMS). Trace-element analysis provides general information about the types of impurities in quartz and their quantitative abundance. Information about the homogeneity or heterogeneity of quartz grains/crystals and the possible spatial distribution of defects can be obtained from CL measurements. Finally, EPR and XAS measurements yield additional information about the structural incorporation of trace elements and other structural defects not related to impurities, and are increasingly complemented by theoretical data from first-principles calculations. In addition, other spectroscopic techniques such as synchrotron XAS and Mössbauer spectroscopy, which have lower sensitivities than EPR techniques, have contributed useful information about the structural states of iron in quartz. In many cases, microscopic techniques (polarising microscopy, scanning electron microscopy (SEM), transmission electron microscopy (TEM), micro-Raman spectroscopy, or microbeam synchrotron X-ray fluorescence and XAS can complement the investigations.

This paper presents a review of the state-of-the-art knowledge concerning the mineralogy and mineral chemistry of quartz and the importance for better understanding of certain geological processes, together with improvements in technical processing of silica raw materials.

The mineralogy of quartz

Quartz is a mineral with the chemical formula SiO₂, which exists in two modifications in nature. The most common and most important modification is trigonal, low-temperature alpha-quartz, which is stable under surface conditions. At atmospheric pressure, alpha-quartz transforms at ca. 573°C into hexagonal, high-temperature beta-quartz. This transformation is reversible and is accompanied by a change of the density from 2.65 to 2.51 g/cm³. The process also occurs, vice versa, during crystallisation from a melt or high-grade metamorphism during which the hexagonal high-temperature beta-quartz forms and subsequently transforms during cooling into the trigonal low-temperature modification. The structural changes during this transformation typically result in the formation of defects or twinning (Blankenburg et al., 1994).

The crystal structure of alpha-quartz consists of a three-dimensional network of [SiO₄]^4–-tetrahedra, which are all linked via the oxygen atoms and are arranged in helical chains along the c-axis forming structural channels (Fig. 1). The left-handed or right-handed helices in the quartz structure correspond to morphologically right and left crystals, respectively (Donnay and Le Page, 1978; Glazer, 2018), which can be distinguished by the relative position of the trapezohedron face x to the positive main rhombohedron face r. The c-axis in the trigonal symmetry is identical with the optical axis and results in the optical uniaxial character of alpha-quartz. The silicon atom in the quartz structure resides on a two-fold axis, whereas all oxygen atoms are at general positions (Le Page et al., 1980; Kihara, 1990; Baur, 2009). Each SiO₄ tetrahedron has two long and two short Si–O bonds at 1.613(2) Å and 1.603(2) Å, respectively, in quartz measured at 291 K and ambient pressure (Baur, 2009). Le Page et al. (1980) noted that the two Si–O bond distances are 1.611(1) Å and 1.606(1) Å at 94 K, and they do not change in the temperature range from 94 K to 298 K, however the Si–O–Si angle increases from 142.69(4)° to 143.65(5)°. In the EPR literature, the SiO₄ tetrahedron in the quartz structure is typically represented by the notation [SiO₄]⁰ to emphasise its overall neutrality, whereas substitutional groups such as [GeO₄]⁰ and [AlO₄]^–, together with their associated paramagnetic centres [GeO₄]^– and [AlO₄]⁰ are denoted accordingly (Weil, 1984; Mashkovtsev and Pan, 2013; Alessi et al., 2014).

Fig. 1. Projection of morphologically right-handed α-quartz atomic positions onto the (0001) plane perpendicular to the c-axis, showing the EPR coordinate system (xyz), crystallographic axes a₁, a₂ and a₃, and z∥c; the central c-axis large channel is seen to be surrounded by six c-axis small channels; LS denotes left-handed helices.

Several chemical and physical properties of quartz such as trace element and isotopic composition, luminescence behaviour, micro-inclusion inventory or colour are strongly dependent on the specific P–T–x conditions of its formation. These varying properties result in the development of numerous varieties, i.e. colour varieties such as amethyst, smoky quartz and rose quartz, or certain growth phenomena such as fibre and sceptre quartz, or microcrystalline quartz varieties such as chalcedony and quartzine (Rykart, 1995; Blankenburg et al., 1994; Götze, 2009).

Point defects in quartz

Zero-dimensional point defects in quartz can be related to the incorporation of foreign ions in lattice sites and interstitial positions, to different types of displaced atoms, and/or to defects associated with silicon or oxygen vacancies (e.g. Weil, 1984, 1993; Stevens Kalceff, 2009; Nilges et al., 2008, 2009; Pan et al., 2009; Mashkovtsev and Pan, 2012a, 2012b; Mashkovtsev et al., 2013a; Alessi et al., 2014; Mashkovtsev and Pan, 2014). The latter represent pure lattice defects and can be generated without the incorporation of impurity ions into the quartz lattice. Point defects can serve as the starting locations for more extended defects such as dislocations or planar defects (McLaren et al., 1983, 1989; Lin et al., 1994; McConnell et al., 1995).

According to their electronic structure, point defects can be classified into diamagnetic and paramagnetic types and investigations are mainly recognised using EPR spectroscopy in combination with other spectroscopic methods (e.g. luminescence spectroscopy, UV-Vis-IR absorption spectroscopy, synchrotron XAS, Mössbauer spectroscopy) and trace-element analysis. Most of the paramagnetic centres in quartz are metastable, i.e. they are sensitive to irradiation and temperature treatment. However, some stable centres have also been detected, which can be formed by the incorporation of paramagnetic ions such as H^+/0, Ag^2+/0, Li⁰, Cu^2+/0 or Ni⁺, mainly in interstitial positions (Laman and Weil, 1977; Davis and Weil, 1978; Isoya et al., 1983; Weil, 1993). Even these paramagnetic ions in quartz are mostly not intrinsic and are either incorporated by electron diffusion or converted from diamagnetic precursors via irradiation. Indeed, most point defects in quartz are diamagnetic, but can be converted to paramagnetic ones recognisable by EPR analysis, by various physical and chemical treatments (e.g. natural and artificial irradiation; Weil, 1984; Pan and Nilges, 2014).

The EPR techniques, with superior sensitivity unmatched by any other structural methods, have identified a large number of paramagnetic defects (and their diamagnetic precursors) in quartz (Supplementary Tables S1 and S2, see details below). In particular, the common detection and analysis of diagnostic hyperfine or superhyperfine structures of paramagnetic defects, which arise from interactions with non-zero-spin nuclei, not only aid in their unambiguous identification, but also typically provide definitive information about their geometric and electronic configurations. For example, ²⁹Si hyperfine structures, which is the only stable isotope of silicon with a non-zero nuclear spin (I = 1/2) and a natural isotope abundance of 4.7%, are often detected by EPR. However, analysis of ¹⁷O hyperfine structures, which is the only stable isotope of oxygen with a non-zero nuclear spin (I = 5/2) and an exceedingly low abundance of 0.04%, generally requires artificially ¹⁷O-enriched quartz samples (McEachern and Weil, 1994; Mashkovtsev et al., 2013).

Aluminium

Aluminium is the prevailing impurity in quartz. The [AlO₄]⁰ centre (Fig. 2) was first described in smoky quartz by Griffiths et al. (1954). This paramagnetic centre is the most common trace-element defect in quartz, which is caused by substitution of Si⁴⁺ by Al³⁺ with an electron hole localised at one of the four nearest O^2– ions. The precursor state for this paramagnetic centre is the diamagnetic [AlO₄/M⁺]⁰ defect with an adjacent charge compensating cation (M⁺ = H⁺, Li⁺, Na⁺) at an interstitial position in the c-axis channel (Botis et al., 2009). These diamagnetic defects have been confirmed by the low-temperature EPR characterisation of the paramagnetic [AlO₄/M⁺]⁺ centres in quartz irradiated and measured at 77 K (Table S1; Mackey, 1963; Mackey et al., 1970; Nuttall et al., 1981a,b; Dickson and Weil, 1990; Walsby et al., 2003; Botis and Pan, 2009, 2011). The paramagnetic [AlO₄/M⁺]⁺ centres are known to convert to the [AlO₄]⁰ centre upon warming to room temperature (Cohen, 1956; Nuttall and Weil, 1981a; Walsby et al., 2003). The optically active [AlO₄]⁰ centre is responsible for the colour of smoky quartz (O'Brien, 1955; Meyer et al., 1984). Nuttall and Weil (1981c) also reported the triplet-state (S = 1) centre [AlO₄]⁺ in quartz measured at ~35 K after X-ray irradiation at 77 K (Table S1), which contains two electron holes, one at each of two symmetry-related oxygen atoms bonded to an aluminium substituting for silicon and apparently formed from the diamagnetic [AlO₄/M⁺]⁰ precursors as well.

Fig. 2. Representative powder EPR spectrum of a high-purity quartz from Kyshtym (Russia) containing 4.2 ppm Al (Götze et al., 2017) measured at a microwave frequency (ν) of ~9.39 GHz and temperature of 85 K, showing the [AlO₄]⁰ centre. Also shown for comparison is the simulated spectrum of this centre using spin Hamiltonian parameters from Walsby et al. (2003). Insert is a room-temperature spectrum of the same sample containing <0.41 ppm Fe (i.e. the detection limit of LA-ICP-MS), showing a rhombic Fe³⁺ centre at the effective g = ~4.28.

Gemanium

Although the abundance of germanium in nature is low, the isovalent substitution of Ge⁴⁺ for Si⁴⁺ is common. The resulting diamagnetic [GeO₄]⁰ defect in quartz can be transformed by ionising radiation at <100 K into two paramagnetic [GeO₄]^– centres (Table S1), which are distinguished by the orbital of the unpaired spin lying along the different O–Ge–O bisectors of the GeO₄ tetrahedron (Mackey, 1963; Isoya et al., 1978; McEachern et al., 1992; McEachern and Weil, 1994). On warming above 150 K or irradiation at room temperature, the [GeO₄]^– centres can capture M⁺ ions (Li⁺, Na⁺ and Ag⁺) released from associated [AlO₄/M⁺]⁰ centres and form the [GeO₄/M⁺]⁰ centres (Table S1; Mackey, 1963; Weil, 1971; Weil, 1984; Rakov et al., 1985; Dickson et al., 1991; McEachern et al., 1992; McEachern and Weil, 1994; Claridge et al., 2008).

In addition, multiply-compensated germanium electron centres such as [GeHLi₂] and [GeH^–H₁⁺H₂⁺] have been reported. They are interpreted to derive from the diamagnetic [GeO₄/M₁M₂] precursors, containing a divalent Ge²⁺ ion and two monovalent charge compensators (M = Li⁺, H⁺, Na⁺), by capturing a paramagnetic atomic hydrogen during irradiation (Weil, 1971; Laman and Weil, 1978; Weil, 1984).

Iron

Iron can be incorporated into quartz in three different valence states (Fe²⁺, Fe³⁺ and Fe⁴⁺; Cox, 1976, 1977; Cohen, 1985; Cressey et al., 1993; Weil, 1994; Schofield et al., 1995; Dedushenko et al., 2004; Di Benedetto et al., 2009; SivaRamaiah et al., 2011; SivaRamaiah and Pan, 2012), in addition colloidal clusters of metallic iron have also been reported to occur in quartz (Daniels and Morton, 1981; SivaRamaiah and Pan, 2012). EPR studies have established at least six distinct Fe³⁺ centres substituting for Si, as either the charge uncompensated [FeO₄]^– centre or charge-compensated [FeO₄/M⁺]⁰ centres (M⁺ = Li⁺, H⁺, Na⁺, Table S1; Scala and Hutton, 1976; Mombourqette et al., 1986, 1989; Halliburton et al., 1989; Minge et al., 1989, 1990; Weil, 1984, 1994). The larger ionic radius of Fe³⁺ (0.64 Å) compared to Si⁴⁺ (0.42 Å), means that the tetrahedra will be distorted. Therefore, the structural incorporation of Fe³⁺ is only possible in marginal parts of the quartz crystals or in fissures (Götze and Plötze, 1997). Higher amounts of iron can cause lattice defects promoting the formation of highly disordered areas such as in the case of amethyst (Mineeva et al., 1991).

The [FeO₄/M⁺]⁰ centres can be converted to the [FeO₄]^– centre by irradiation (Hutton and Troup, 1966; Stegger and Lehmann, 1989). This process also results in the conversion of Fe³⁺ to Fe⁴⁺ accompanied by the formation of the typical amethyst colour (Fig. 3; Lehmann and Moore, 1966; Stegger and Lehmann, 1989; Dedushenko et al., 2004; Di Benedetto et al., 2009). However, details concerning the Fe⁴⁺ site and its local environment have not been determined unequivocally (Cox, 1976, 1977; Dedushenko et al., 2004; Di Benedetto et al., 2009). Similarily, very little is known about the local structural environment of Fe²⁺ in quartz (Cressey et al., 1993; Schofield et al., 1995; Di Benedetto et al., 2009), because the magnetic properties of Fe²⁺ with oxygen neighbours preclude its EPR observation at conventional X-band frequencies (Weil, 1994).

Fig. 3. Representative single-crystal EPR spectrum of a natural amethyst measured with B∥y and at 294 K. Letters S and I mark lines belonging to the S₁ ([FeO₄/Li]⁰; Han and Choh, 1989; Halliburton et al., 1989) and I centres, respectively. Note that the so-called interstitial (I) centre is actually [FeO₄]^– with substitutional Fe ions at the Si sites (Mombouquette et al., 1986). Also present are the broad signal at g = 4.28 and the E′₁ centre (modified from SivaRamaiah et al., 2011).

Titanium

Titanium is a typical trace impurity in quartz and isovalently substitutes as Ti⁴⁺ for Si⁴⁺ in the SiO₄ tetrahedra. The Ti⁴⁺ in the diamagnetic [TiO₄]⁰ centre can be transformed to the paramagnetic [TiO₄]^– centre during irradiation at low temperatures (Table S1; Bershov, 1970; Isoya and Weil, 1979; Weil, 1984; Bailey et al., 1992; Bailey and Weil, 1992a,b). This [TiO₄]^– centre is unstable above 120 K and, similar to the [GeO₄]^– centres, can capture diffusing M⁺ ions (H⁺, Li⁺, Na⁺) to form charge compensated [TiO₄/M⁺]⁰ centres (Table S1; Wright et al., 1963; Rinneberg and Weil, 1972; Weil, 1984; Matyash et al., 1987; Isoya et al., 1988; Bailey and Weil, 1991, 1992).

There is an ongoing discussion about the role of substitutional Ti as an origin for the rose colour in quartz (Wright et al., 1963). Smith et al. (1978) suggested that the colour originates from a charge transfer between substitutional Ti⁴⁺ and interstitial Fe²⁺. Investigations by Lehmann and Bambauer (1973) and Cohen and Makar (1985) showed that additional Ti³⁺ at interstitial structural channel sites is responsible for the rose colouration. However, further studies have shown that nano-inclusions of dumortierite can also produce the colour in massive rose quartz (Goreva et al., 2001; Ma et al., 2002).

Phosphorus

Phosphorus as a trace element has been shown to substitute for Si⁴⁺ in the quartz structure as EPR has detected several radiation-induced [PO₄]⁰ centres (Table S1). In particular, Mashmeyer and Lehmann (1983) reported the [O₃AlO^–PO₃] centre with a hole trapped on the bridging oxygen atom between substitutional Al and P atoms in quartz, providing unambiguous evidence for the coupled substitution: P⁵⁺ + Al³⁺ = 2 Si⁴⁺. In addition, Maschmeyer and Lehmann (1983) suggested that the [O₃AlO^–PO₃] centre is responsible for radiation-induced rose colouration of quartz.

The aforementioned examples illustrate that there typically exists a close relation between point defects and the different colours of quartz (Lehmann, 1978; Rossman, 1994). Although several allochromatic colours in quartz such as green, blue and red are caused by micro-inclusions of impurities (e.g. dickite, rutile, hematite, goethite, celadonite and dumortierite), idiochromatic colours are mainly generated by chromophores (Rossman, 1994; Yang et al., 2007; Scholz et al., 2012). This group consists of ions of 3d elements (e.g. synthetic blue quartz: Co²⁺, green chrysoprase: Ni²⁺) or paramagnetic centres, which can be activated by irradiation (Lehmann, 1978; Meyer et al., 1984). However, the discussion concerning the origin of colours in quartz is very complex and partly controversial and will not be made here in detail. A thorough review about the origin of colours in quartz is given by Rossman (1994).

A number of point defects can be attributed to oxygen and silicon vacancies or oxygen excess. Oxygen-excess centres in quartz include the peroxy linkage (≡Si–O–O–Si≡) and the peroxy radical (≡Si–O–O), an oxygen associated hole centre consisting of an O₂^– ion bound to a single silicon on three oxygen atoms (Friebele et al., 1979). In addition, hydrogen excess from the H₂O crystallisation medium can also result in the formation of OH^– centres in quartz. This defect consists of a proton bound on a regular O^2– ion of the SiO₄ tetrahedron (Weil, 1984, 1993). In addition, the EPR detection of two H-trapped hole centres ([H₄O₄]⁺ and [H₃O₄]⁰) (Table S1) has been used to suggest the presence of the neutral diamagnetic precursor with four hydrogen atoms located at the tetrahedral Si site (i.e. the [H₄O₄]⁰ or hydrogarnet defect; Nuttall and Weil, 1980; Lin et al., 1994; Lees et al., 2003; Jollands et al., 2020). Similarly, the EPR detection of the [HLi₂O₄]⁰) centre in gamma-ray-irradiated quartz suggests the presence of analogous diamagnetic precursors [HLi₃O₄]⁰ or [H₂Li₂O₄]⁰ (Table S1; Lees et al., 2003).

The group of oxygen-vacancy electron centres

The group of oxygen-vacancy electron centres (OVEC) involves an oxygen vacancy (≡Si–Si≡) and can be divided into the E′ and E′′ types, denoting one and two unpaired electrons (S = 1/2 and 1), respectively (Table S2). The E′ (the •Si≡O moiety) centres are the most frequent defects in quartz apart from Al centres and can occur in different stages depending on thermal stability and sensibility to irradiation (Griscom, 1985). The most common E′₁ centre can be formed by various irradiations involving alpha particles, electron beams, neutrons, gamma- and X-rays, and has been used as palaeo-dosimeters and geochronometers due to its high thermal stability (> 550°C) (Moiseev, 1985; Toyoda, 2011, 2016). In addition, the E′₁ centre in nature, induced by diverse radiation sources from U/Th-bearing rocks/fluids to potassic alteration, can be used as an indicator for the formation of uranium deposits when the dominant source of radiation is uranium (Pan et al., 2006; Botis et al., 2006, 2008; Cerin et al., 2017). X-ray or fast-electron irradiation of Ge-doped quartz at room temperature yielded at least three Ge analogues of E′ centres, including the Ge E′₁ centre stable up to 700 K, the Ge E′₂ centre with a hydrogen impurity stable up to 500 K, and another Ge(IV) centre unstable at room temperature (Fig. 4; Feigl and Anderson, 1970; Mashkovtsev et al., 2013).

Fig. 4. Representative single-crystal EPR spectrum of an electron-irradiated, ¹⁷O-enriched quartz (JC324) measured, after annealing at 573 K, with B‖c, room temperature and X-band frequency (~9.3 GHz; modified from Mahskovtsev et al., 2013), showing three ¹⁷O hyperfine sextets and one ⁷³Ge hyperfine line of the Ge E′₁ centre. Insert shows the 3D spin density of the Ge E′₁ centre calculated from the tri-vacancy with an Al impurity model.

Understanding of the E′₁ and Ge E′₁ centres in quartz has evolved with increasingly complete EPR data and sophisticated first-principles calculations (Fig. 4; Li and Pan, 2012; Mashkovtsev et al., 2013). The tri-vacancy model with an Al impurity, which best reproduces all available EPR data, involves the removal of one Si and two O atoms as well as the presence of a neighbouring Al impurity atom (Fig. 4; Li and Pan, 2012; Mashkovtsev et al., 2013). Other well-characterised E′ centres in quartz include the E′₄ and E′_11,16 centres (Table S2; Perlson and Weil, 2008; Mashkovtsev and Pan, 2012a, 2016, 2018; Mashkovtsev et al., 2019). The former contains a proton near the oxygen vacancy, wheras the latter with diagnostic ²⁷Al superhyperfine structures provide experimental proof for the hypothesis that Al impurity plays an important role in the formation of E′ centres (Jani et al., 1983; Mashkovtsev and Pan, 2018; Mashkovtsev et al., 2019). In addition, various E′′ centres arising from interactions between two neighbouring E′ centres have been discovered and characterised with the triplet-state model in recent years (Table S2; Mashkovtsev et al., 2007; Mashkovtsev and Pan, 2011, 2012a, 2012b, 2014).

A family of silicon-vacancy hole centres is represented by different localisations of the unpaired electron on one, two or three oxygen atoms of the tetrahedra with silicon vacancies, which have been denoted as the O^–, O^2–, O₂^3– or O^3– type radicals (Fig. 5; Table S2; Mashkovtsev et al., 1978; Botis et al., 2005; Nilges et al., 2008, 2009; Pan et al., 2008, 2009; Mashkovtsev and Pan, 2013; Alessi et al., 2014). By analogy to the notation for oxygen-vacancy electron centres, Mashkovtsev and Pan (2013) relabelled the silicon-vacancy hole centres as H′₁ to H′₇ (Table S2). The O^– centres with a hole trapped in a single nonbonding 2p orbital (i.e. the non-bridging oxygen hole centres or NBOHC; ≡Si–O) are very common defects in almost all quartz types, which can be detected by CL spectroscopy (Stevens Kalceff, 2009). However, the majority of the O^– centres in quartz are those associated with impurity ions such as Al³⁺ described above (Table S1) and are not linked with any silicon vacancy.

Fig. 5. Representative single-crystal EPR spectrum of an electron-irradiated quartz measured at W-band frequency (~94 GHz), B^c = ~140°, T = 110 K and a microwave power of 0.2 mW, illustrating three silicon-vacancy hole centres (H′₁, H′₄(I) and H′₅) (modified from Nilges et al., 2009). Note that the H′₁ (alias #1) centre has six main lines corresponding to six magnetically inequivalent sites and that ²⁹Si hyperfine satellites are marked on four main lines. Also note that the two remaining main lines at ~3330 mT have irregular line shapes due to an incompletely resolved ²⁷Al superhyperfine structure (Nilges et al., 2009). The H′₄(I) centre is characterised by a well-resolved ²⁷Al superhyperfine structure.

Several studies showed that defects associated with oxygen and silicon vacancies are most frequent in quartz crystallised at relatively low temperature (< 250°C) from amorphous silica precursors due to rapid cooling (Götze et al., 1999, 2015a, 2020). Therefore, these defects are common in microcrystalline quartz (agate, chalcedony) and in some hydrothermal quartz formed with high growth rates. In contrast, pegmatite quartz, in general, has a remarkably low abundance of defects associated with oxygen or silicon vacancies indicating growth of quartz from a parent melt/fluid under more or less constant physicochemical conditions (Götze et al., 2004, 2005).

Moreover, partial dissolution experiments demonstrated that the silicon-vacancy hole centres in detrital quartz from the Athabasca Basin (Canada) are concentrated preferentially along the grain margins. This observation is consistent with the limited penetration distances of alpha particles emitted from the decay of the uranium series present in palaeo-uranium-bearing fluids, providing the best linkage with mineralisation processes in uranium deposits (e.g. tracing the conduits of uranium-bearing fluids and constraining the sources of uranium; Pan et al., 2006; Hu et al., 2008; Cerin et al., 2017).

In addition, abundant data exist on the health effects of quartz. This varies widely in emphasis from particle sizes/morphologies to surface properties and structural states. Long-term exposures to quartz and other silica forms (e.g. cristobalite and amorphous silica) are known to cause serious and potentially life-threatening diseases such as silicosis, fibrosis and lung cancer (Goldsmith, 1994). Of particular interest are the identification of various defects such as the peroxy and superoxide radicals and NBOHC on the surfaces of, and in, bulk quartz and their effects on toxicity and carcinogenic activity (Fubini et al., 1990; Giordano et al., 2007; Di Benedetto et al., 2019, 2021). For example, Fubini et al. (1990) proposed the presence of these radicals on quartz surfaces or chemical functionalities as a possible mechanism for fibrogenecity via their reactions with macrophage oxygen metabolites, triggering the abnormal production of fibroblast stimulating factors and ultimately silicosis. However, a comprehensive synthesis of the voluminous literature on the health effects of quartz is beyond the scope of this review in which the emphasis is on geological contexts.

Cathodoluminescence properties of quartz

Cathodoluminescence (CL) is a powerful method, which enables the visualisation of the defect structure of minerals and reveals visually internal features caused by the varying spatial distribution of point defects that are not discernible by other analytical methods (Götze, 2012a; Götze et al., 2013; Götze and Hanchar, 2018). In general, quartz and other SiO₂ modifications (including amorphous silica) show similar main luminescence emission bands. This is due to the fact that short-range order structural defects are caused mainly by silicon–oxygen and silicon–silicon interactions rather than by interactions between oxygen atoms (Walker, 1985). However, the band positions of the CL emissions can vary depending on the specific structure of the SiO₂ polymorph and experimental conditions (e.g. Walker, 1985; Luff and Townsend, 1990; Remond et al., 1992).

The CL emissions for quartz are mostly weak but variable and the relation of specific luminescence emission bands to different defect centres causes a diversity of CL characteristics and visible CL colours, both in natural and synthetic quartz, depending on the processes of mineral formation or alteration (Table 1).

Table 1. Main emission bands in CL spectra of quartz and suggested activators (modified after Götze, 2012a).

Therefore, knowledge about the origin of different luminescence centres can help to reconstruct geological processes and to reveal different growth generations (Fig. 6). However, the interpretation of the origin of CL emission bands in quartz is difficult due to a lack of stringent quantitative correlations between the intensities of CL bands and the concentrations of specific defects or trace elements.

Fig. 6. Photomicrographs of quartz in transmitted light (crossed polars – a) and CL (b) in a rhyolite from Kemmlitz (Saxony, Germany); different quartz generations are distinguishable by different CL colours: (1) primary quartz phenocryst with blue CL, (2) reddish volcanic quartz of a second generation, (3) secondary microcrystalline quartz of hydrothermal origin with yellow CL. (c) The volcanic quartz exhibits two main emission bands at 450 and 620 nm with varying intensity ratios, whereas the hydrothermal quartz has a main emission at 570 nm and a subordinate shoulder at 620 nm; the circles in (b) mark the positions of spectral CL analyses.

In general, structural defects are much more important than trace elements as CL activators in quartz. Although correlations between CL zoning and several trace elements have been observed (e.g. Watt et al., 1997; Müller et al., 2003; Rusk et al., 2006, 2008; Leeman et al., 2012), the origin of the CL is related mostly to intrinsic lattice defects. The most common CL emission bands in natural quartz are bands with maxima at 450 and 650 nm (Ramseyer et al., 1988; Götze et al., 2001a). The visible luminescence colours of quartz in most igneous and metamorphic rocks and in some authigenic quartz depend on the relative intensities of these two dominant emission bands (Fig. 6).

Blue emission band at ~450 nm (2.75 eV)

The ~450 nm (2.75 eV) blue emission band (Fig. 6c) is related to oxygen deficiency centres (ODC) and is similar in amorphous and crystalline SiO₂ (Skuja, 1998). Recombination of the so-called self-trapped exciton (STE) involves an irradiation-induced electron hole pair (oxygen Frenkel pair) consisting of an oxygen vacancy and a peroxy linkage (Stevens Kalceff and Phillips, 1995). The blue emission at ca. 450 nm is probably the most common CL emission in natural quartz and can be detected in almost all quartz types.

In addition to the evidenced activation of the 450 nm band by a structural defect, several studies reported a strong correlation between the intensity of luminescence and the concentration of Ti in quartz, with the brightest CL corresponding to the highest Ti concentrations (e.g. Müller et al., 2002, 2003; Van den Kerkhof et al., 2004; Rusk et al., 2008; Leeman et al., 2012; Drivenes et al., 2016). Up to now there has been no serious spectroscopic evidence that Ti is responsible for the activation of the blue emission in quartz. However, recent systematic studies of Ti-rich natural quartz from igneous rocks, as well as synthetic quartz from Ti-diffusion experiments could prove that Ti might also be responsible for the activation of the 450 nm luminescence emission band (unpublished data – compare with Fig. 22).

Red emission band at 620–650 nm

The red emission band at 620–650 nm (1.95–1.9 eV; Fig. 6c) is attributed to the recombination of electrons in the non-bridging oxygen band-gap state with holes in the valence-band edge (Siegel and Marrone, 1981; Stevens-Kalceff, 2009). A number of different precursors of this NBOHC has been proposed such as strained silicon–oxygen bonds, hydrogen or sodium impurities, or peroxy linkages (Stevens-Kalceff and Phillips, 1995). The formation of the NBOHC from these different defects causes variations of the band position. It is assumed that the 620 nm (1.95 eV) component is characteristic for hydroxyl precursors (:Si–OH), which are common in hydrothermal and authigenic quartz as well as silicified wood (Stevens-Kalceff et al., 2000).

The CL emission at 1.9 eV (650 nm) commonly increases during electron irradiation due to the formation of NBOHC from different precursors. A high state of lattice damage can be attained in radiation-damaged quartz. The creation of NBOHC defects by bond breaking due to the α-particles was observed in both natural radiation-damaged quartz samples and in radiation experiments (e.g. Komuro et al., 2002; Krickl et al., 2008). In natural quartz, such lattice damage causes halos around U- and Th-bearing accessory minerals or migration tracks of U-bearing fluids (Fig. 7), which cannot be detected by conventional microscopic techniques (e.g. Owen, 1988; Ramseyer et al., 1988; Meunier et al., 1990; Götze et al., 2001a ; Botis et al., 2005; Cerin et al., 2017).

Fig. 7. Photomicrographs in transmitted light (crossed polars, a) and CL (b) of drusy quartz from the Arrow U-deposit, Athabasca basin (Saskatchewan, Canada); the CL image reveals a continuous yellow orange radiation rim around the crystal as well as radiation halos around radioactive inclusions (arrow). (c) The CL spectra display strong development of the 650 nm emission band (NBOHC) due to radiation induced lattice damage; the circles in (b) mark the positions of spectral CL analyses.

Yellow emission band at 570 nm (2.15 eV)

Another defect-related CL emission in quartz is a broad yellow emission band centred at 570 nm (2.15 eV) that was first observed by Rink et al. (1993) in natural quartz of hydrothermal origin and Götze et al. (1999) in agates of acidic volcanic rocks and hydrothermal vein quartz (Fig. 6c). Götze et al. (2015b) showed that the appearance of the 570 nm emission band can be attributed to high oxygen deficiency and local structural disorder in quartz. The proposed luminescence centre model implies self-trapped exciton (STE) emissions from strongly disordered regions in quartz. Additional geochemical data proved that quartz with yellow CL occurs exclusively in a low-temperature hydrothermal environment (mostly <250°C) and is related to fast crystallisation in an environment with oxygen deficiency.

Blue emission band at 385 nm (3.15 eV)

Trace-element activated CL in quartz is rare, but can be observed for specific geological environments, in particular in hydrothermal and pegmatite quartz. A blue emission at the edge of the UV region with a maximum at ca. 385 nm (3.15 eV; Fig. 8) correlates well with the Al content and the concentration of paramagnetic [AlO₄/M⁺] centres (Alonso et al., 1983; Luff and Townsend, 1990; Perny et al., 1992). The short-lived blue CL is the typical feature of natural and synthetic hydrothermal quartz (Ramseyer et al., 1988; Götze et al., 2001a, 2009). Because of the sensitivity to electron irradiation, this emission has been attributed to the recombination of a hole trapped adjacent to a substitutional, charge-compensated aluminium-alkali ion centre (Stevens Kalceff and Phillips, 1995). The rapid attenuation of the 385 nm emission under an electron beam results from the dissociation and electromigration of the charge compensating cations out of the interaction volume under the influence of the irradiation induced electrical field (Perny et al., 1992). However, Gorton et al. (1996) found a reduced 390 nm band even in ultra-pure synthetic quartz and suggested that the 3.15 eV emission is probably not completely related to Al.

Fig. 8. Photomicrographs in transmitted light (crossed polars, a) and CL (b) of an Al-doped, synthetic hydrothermal quartz; CL reveals growth zones not visible in transmitted light. (c) The CL image after 60 s of electron irradiation reveals the transient character of the CL, which turns from initial blue (b) to reddish-brown (c). The related CL spectra (d) show a strong decrease of the blue emission band, whereas the red band increases due to the conversion of precursor centres (e.g. silanol groups :Si–OH) into the NBOHC.

Blue green emission at ~500 nm (2.45 eV),

Another trace-element related luminescence emission band in quartz is the short-lived blue-green CL (Fig. 9) at ca. 500 nm (2.45 eV), which can be related to the alkali-compensated trace-element centres in the quartz structure (Ramseyer and Mullis, 1990; Perny et al., 1992, Götze et al., 2005). Ramseyer and Mullis (1990) and Perny et al. (1992) performed CL measurements, microprobe analyses, temper and electro-diffusion experiments and concluded that the CL can be related to the uptake of positively charged interstitial cations (H⁺, Na⁺, Li⁺) associated with the substitution of Al for Si. Several studies showed that the transient blue-green CL is a characteristic feature of pegmatite quartz and can also occur in hydrothermal and even in metamorphic quartz (Ramseyer and Mullis, 1990, Perny et al., 1992; Götze et al., 2001a, 2005; Sittner and Götze, 2018). Recent studies showed the predominance of Li⁺ as a charge-balancing cation for the 500 nm CL emission in pegmatite quartz (Sittner, 2019).

Fig. 9. Photomicrographs in transmitted light (crossed polars, a) and CL (b) of a pegmatite quartz from Heftetjern, Tørdal region (Norway); the sample exhibits the characteristic transient blue-green CL; fluid trails are visible due to dark CL. (c) The CL spectra show a drastic drop of the CL intensity of the 500 nm emission band during electron irradiation; the circle in (b) marks the position of spectral CL measurements.

The intensity of the blue-green CL falls off rapidly within 30 to 60 seconds during electron irradiation. This transient behaviour can be related to ionisation-enhanced diffusion of luminescence centres as was proved by electro-diffusion experiments (Ramseyer and Mullis, 1990). The short-lived CL can be restored by heating the quartz to 500°C for one day (Perny et al., 1992) indicating an opposite mechanism to that of thermoluminescence. In contrast, gamma-irradiation causes a decrease in the intensity of the short-lived blue-green CL and an increase in smoky colouration (Ramseyer and Mullis, 1990). This can be explained by the conversion of [AlO₄/M⁺]⁰ luminescence centres into [AlO₄]⁰ colour centres.

Green luminescence at ~500–600 nm

A remarkable feature of some microcrystalline silica varieties as well as macrocrystalline quartz in agates is greenish luminescence emitted in CL and under short-wave UV (<300 nm) excitation, which is uncommon in quartz of magmatic and metamorphic rocks (Fig. 10). There have also been a few reports about UV-excited green luminescence in opal and chert that have elevated uranium concentrations (Gaillou et al., 2008). The green luminescence is due to the electron transition from an excited to a ground state of the uranyl ion (UO₂²⁺) and is shown by a typical emission line at ~500 nm (2.45 eV) accompanied by several equidistant lines (Götze et al., 2015a; Fig. 10c). These are due to the harmonic vibrations of oxygen atoms along the uranyl axis (Gorobets and Rogojine, 2002). Recent studies have revealed that U is mainly incorporated in the form of a uranyl-silicate complex in opal and microcrystalline quartz (Pan et al., 2021). The mechanism of the uranyl luminescence is very effective and can be detected in SiO₂ phases with U contents as low as 1 ppm.

Fig. 10. Photomicrographs in transmitted light (crossed polars, a) and CL (b) of a chalcedony sample from Kardzali, Bulgaria; microcrystalline quartz exhibits a bright green CL. (c) The CL spectra reveal that uranyl (UO₂²⁺) is responsible for the emission peak at ~500 nm accompanied by several equidistant lines; there is an additional CL emission band at 650 nm due to the non-bridging oxygen hole centre (NBOHC). The circle in (b) marks the position of spectral CL analyses.

Red luminescence at 705 nm (1.65 eV)

A CL emission band in the red on the edge to the infrared around 705 nm (1.65 eV) was first reported from synthetic quartz crystals (Pott and McNicol, 1971). It is assumed that this ~705 nm emission can be related to the substitutional incorporation of Fe³⁺ into the quartz lattice similar to other silicates such as feldspar minerals (Pott and McNicol, 1971; Gorobets and Rogojine, 2002). Some red luminescing quartz was also found in metasomatically overprinted granites from Khaldzan Buregte (Mongolia), where the ferric iron was probably mobilised during fenitisation processes (Kempe et al., 1999). Because of the relatively large ionic radius of Fe³⁺ its incorporation into the quartz lattice is limited and therefore, the 705 nm emission is rare. Its origin has to be further verified by measurements of the luminescence lifetime and/or the luminescence excitation spectra.

Mineral chemistry of quartz

Quartz is one of the purest minerals in the Earth's crust in terms of chemical impurities expressed as trace-element contents. Due to limitations in charge and ionic radii only a limited number of ions can substitute for Si⁴⁺ in the crystal lattice or can be incorporated in interstitial positions. Therefore, most elements are present in quartz in concentrations below 1 ppm (Müller et al., 2003, 2012; Götze, 2009). The structural incorporation in a regular Si⁴⁺ lattice position has been proven for Al³⁺, Ga³⁺, Fe³⁺, B³⁺, Ge⁴⁺, Ti⁴⁺ and P⁵⁺ (e.g. Weil, 1984), in which Al is probably the most common ion (up to a few 1000 ppm) because of its abundance in the Earth's crust and the similar ionic radii of Si⁴⁺ and Al³⁺. Mono- and divalent ions, primarily H⁺, Li⁺, Na⁺, K⁺, Be⁺, Rb⁺ and Fe²⁺ may enter interstitial lattice positions in small amounts and charge-compensate trivalent substitutional ions, mainly Al³⁺ (Bambauer, 1961, Kats, 1962, Perny et al., 1992, Stalder et al., 2017, Potrafke et al., 2019). For most other elements (e.g. Ca, Mg, Ba, Sr, REE, Mn, U and Th) capture by nano- (< 100 nm) or micro-inclusions (0.1–1 μm) of fluids or minerals is the most important mechanism (e.g. Gerler and Schnier, 1989; Müller et al., 2003; Götze, 2009).

Aluminium

The structural incorporation of aluminium is evidenced by numerous studies using EPR spectroscopy (Alonso et al., 1983). The substitution of Si⁴⁺ by Al³⁺ is mostly accompanied by charge compensating cations in interstitial positions. Combined studies by EPR, CL, spatially resolved trace-element analyses and SIMS mapping have revealed that Li⁺ seems to be the dominant charge compensating ion in igneous and pegmatite quartz (Götze et al., 2005; Beurlen et al., 2011; Sittner, 2019), whereas H⁺ plays an additional significant role in hydrothermal quartz and authigenic quartz in sedimentary/diagenetic environments (Bambauer, 1961; Miyoshi et al., 2005; Müller and Koch-Müller, 2009; Jourdan et al., 2009; Götte et al., 2011; Lehmann et al., 2011; Götte, 2016; Stalder et al., 2017; Fig. 11). The preferred incorporation of H⁺ in interstitial structural positions in natural and synthetic hydrothermal quartz can be explained by its input from the mineralising aqueous fluids (Bambauer et al., 1962; Stenina, 2004). Further possible mechanisms of structural incorporation of Al³⁺ are the coupled substitution with a pentavalent ion or the charge compensation by electron defects on vacancies.

Fig. 11. Element correlation for Al vs. Li (ppma = atoms per 10⁶ atoms Si) in quartz of different origin: (a) pegmatite quartz from Norway; (b) pegmatite quartz from different deposits in Norway and Namibia; (c) hydrothermal quartz from the Ural region, Russia. The pegmatite quartz shows almost 1:1 ratios of Al to Li, whereas the hydrothermal quartz is Li deficient (data from Götze et al., 2004, 2017; Sittner, 2019).

The concentration of Al in quartz in magmatic systems seems to be mainly controlled by the aluminium saturation index of the melt rather than by the temperature (Jacomon and Larsen, 2009; Breiter et al., 2013, 2020). Accordingly, the Al concentration in quartz increases during melt evolution from metaluminous to peraluminous compositions. In hydrothermal systems, factors such as temperature, growth rate and Al concentration in the aqueous solution influence the Al content in quartz. It was shown that changes in pH have strong effects on Al solubility and the aqueous Al concentration, and thus cause variations of the Al concentration in hydrothermal quartz (e.g. Perny et al., 1992; Rusk et al., 2008; Müller et al., 2010; Luo et al., 2020).

The incorporation of Al into SiO₄ tetrahedra is limited by the ionic radius of the Al³⁺ ion (0.54 Å), which can occur both in fourfold and sixfold coordination with oxygen. Therefore, high amounts of Al substitution in quartz result in a distortion of the tetrahedra and destroy the perfection of the crystal lattice. Nevertheless, previous EPR studies of [AlO₄]⁰ centres in chalcedony and the macrocrystalline quartz in agates have revealed abnormal high concentrations of structural Al in these SiO₂ phases (Götze et al., 2001b). There are strong indications that the formation of some hydrothermal quartz and/or agate proceeds through several structural states of SiO₂ with amorphous silica as the first solid phase (Dong et al., 1995, Götze et al., 2020). This might result in areas or zones with a high degree of disorder and large numbers of lattice defects (oxygen and silicon vacancies), which enable the incorporation of high concentrations of ‘defect’ Al. Such a scenario could also explain the extreme high concentrations of Al in neighbouring individual growth zones or generations of hydrothermal quartz, which reflect abrupt changes in the physicochemical conditions (open hydrothermal system).

Titanium

In contrast to Al, titanium can be both incorporated as substitutional ion for Si⁴⁺ and bound on mineral micro-inclusions such as rutile or anatase. The geochemical compatibility of Ti and its relatively large ionic radius (Ti⁴⁺ 0.61 Å) results in a preferred uptake in early formed quartz at high temperatures and pressures. Therefore, the highest Ti contents in quartz can be found in quartz phenocrysts in volcanic rocks and early crystallised quartz grains in igneous rocks. In granitic melts, decrease in the Ti concentrations in quartz can be observed during increasing fractionation and decreasing temperature (Larsen et al., 2004; Breiter and Müller, 2009). The lowest Ti concentrations in quartz were measured in quartz crystallised from aqueous solutions at relatively low temperatures (Blankenburg et al., 1994; Müller et al., 2003). Using the temperature dependence of Ti uptake in quartz, Wark and Watson (2006) created a Ti-in-quartz geothermometer based on the Ti concentration in quartz. Subsequently, Huang and Audétat (2012) discovered, in addition, that Ti incorporation in quartz is slightly controlled by the crystallisation pressure, and modified the geothermometer to a Ti-in-quartz geothermobarometer. The geothermobarometer is, however, very sensitive to the Ti saturation of the melt in which the quartz crystallises and thus has to be applied with caution.

The limitations in structural compatibility of Ti frequently cause exsolution phenomena at lower temperatures, such as the formation of TiO₂ micro-inclusions (rutile, anatase) in quartz crystals. Ti⁴⁺ is commonly sixfold coordinated by oxygen because of its ionic radius. The substitution of fourfold coordinated Ti⁴⁺ for Si⁴⁺ in the tetrahedral position is only possible at elevated temperatures. Therefore, a straight correlation between bulk Ti contents in quartz and structural Ti measured by EPR spectroscopy is often missing (Fig. 12). However, it should also be considered that Ti can be incorporated as a diamagnetic [TiO₄]⁰ centre that cannot be detected by EPR spectroscopy.

Fig. 12. Plot of Ti content in quartz measured by ICP-MS vs. concentration of structural Ti ([TiO₄/Li⁺]⁰ centres) determined by EPR spectroscopy (concentrations were determined as peak to base intensity under constant analytical settings according to Moiseev, 1985). Pegmatite quartz samples from the Rubicon Mine, Namibia (red circles) exhibit a linear correlation with zero intercept indicating a complete presence of Ti as structural substituent for Si; quartz samples from Norway (blue squares) are mostly far away from the correlation and show higher Ti contents; in these samples micro-inclusions of rutile were detected (data from Götze et al., 2004).

Germanium

Germanium shows a similar geochemical behaviour compared to Si, in addition Ge⁴⁺ has a similar ionic radius (0.53 Å). Therefore, Ge⁴⁺ can substitute isovalently for Si⁴⁺ in the SiO₄ tetrahedra. However, Ge belongs to the group of elements occurring in low concentrations in the Earth's crust (Clark value 1.4 ppm) mostly resulting in low concentrations in quartz. Germanium does not partition into any particular mineral phase and is therefore enriched in residual melts or fluids. In particular the presence of halogens (e.g. fluorides) in the crystallisation medium favours the development of durable and volatile compounds (Walenczak, 1969). Thermodynamic considerations suggest that the transport and accumulation of Ge as the GeF₄ complex during chemical transport reactions could probably explain elevated Ge concentrations in quartz in specific geological environments (Götze et al., 2012).

Germanium contents in quartz from magmatic and metamorphic rocks are mostly in the range between 0.5 and 1.5 ppm (Walenczak, 1969; Blankenburg et al., 1994). In most hydrothermal quartz, the Ge content is similar. Only some hydrothermal quartz veins in paragenesis with metalliferous deposits show elevated Ge contents up to several ppm (Blankenburg et al., 1994; Müller et al., 2018; Breiter et al., 2020). Concentrations of >1.5 ppm up to 23 ppm Ge were measured in quartz crystallised from fractionated granitic and pegmatitic magmas (Götze et al., 2004; Beurlen et al., 2011; Breiter et al., 2014). Extraordinary Ge concentrations of partially > 90 ppm were found in micro- and macrocrystalline quartz of agates, which exceeded the average concentration of the Earth's crust and also the Ge content in the surrounding host rocks (Götze et al., 2016, 2020).

Phosphorus

Phosphorus can substitute as P⁵⁺ with an ionic radius of 0.38 Å for Si⁴⁺ in the quartz structure. A coupled mechanism 2 Si⁴⁺ ↔ P⁵⁺ + M³⁺ (M³⁺ = Al³⁺, B³⁺) is assumed for the structural incorporation with Al³⁺ as the most frequent charge compensator, confirmed by the paramagnetic [O₃AlOPO₃]⁺ centre (Maschmeyer and Lehmann, 1983). However, P is generally not considered in analytical measurements. Therefore, a specified procedure for the precise determination of P in quartz was developed by Müller et al. (2008).

The absolute concentration of P in quartz is mostly low and rarely exceeds several ppm (e.g. Larsen et al., 2004; Müller et al., 2003, 2012; Beurlen, 2011; Breiter et al., 2013, 2020; Sittner, 2019). The concentration of P is typically buffered by the presence of fractionating mineral phases that particularly incorporate it at the expense of the residual melt and quartz. Therefore its distribution in quartz seems to be erratic (Beurlen et al., 2011). Nevertheless, it seems to be slightly enriched during the differentiation of granitic and pegmatitic systems in late crystallisation phases (Larsen et al., 2004).

Iron

Iron can be incorporated structurally by substituting for Si⁴⁺ in the tetrahedra or in interstitial sites. However, the ionic radii of Fe²⁺ (0.78 Å) and Fe³⁺ (0.65 Å) are much larger than that of Si⁴⁺. Therefore, the incorporation is limited and mostly occurs in marginal parts or damaged areas of quartz crystals (Mineeva et al., 1991; Götze and Plötze, 1997). Therefore, elevated contents of Fe in quartz can mostly be related to micro-inclusions of Fe-bearing minerals (e.g. hematite, goethite) or nano-clusters of amorphous Fe-compounds.

Boron

Boron mostly occurs in low concentrations (< 1 ppm) in quartz (Blankenburg et al., 1994). Elevated contents of > 10 ppm B have been measured in quartz of pegmatites (e.g. Götze et al., 2004; Beurlen et al., 2011; Müller et al., 2012), and remarkably high concentrations of up to 46 ppm were detected in quartz of agates from different locations (Götze et al., 2020). Similar to Ge, a preferred transport of boron as a volatile BF₃ complex during chemical transport reactions could be responsible for the accumulation of B in quartz, if halogens were available in the participating fluids (Götze et al., 2012).

The existence of nano-inclusions of tourmaline has been discussed as a possible reason for high B contents in pegmatite quartz (Müller et al., 2012). However, quartz commonly lacks such impurities and a structural incorporation of boron must be assumed by analogy with other silicates (Grew and Hinthorne, 1983). Because of the small ionic radius of the B³⁺ ion (0.27 Å), the formation of planar BO₃-groups in minerals is commonly favoured (Jolland et al., 2020), however the incorporation as BO₄ tetrahedra is also possible. Therefore, a coupled substitution of B³⁺ with pentavalent ions (P⁵⁺) in two neighbouring MO₄ tetrahedra or in combination with charge-balancing cations and electron defects, respectively, has to be considered.

Gallium

Gallium occurs at very low concentrations in natural quartz (< 0.1 ppm) and is, thus, rarely detected. Walenczak (1969) was probably the first who reported Ga concentrations of igneous, metamorphic and hydrothermal quartz. In most of the quartz investigated, Ga showed concentrations between 0.02 and 0.4 ppm and amethyst had generally high Ga up to 25 ppm. Recent publications, applying in situ micro-beam techniques with very low detection limits for trace-element determination, have not confirmed such high Ga contents in natural quartz, and only chalcedony and macrocrystalline quartz of agates had Ga contents of 1–2 ppm (Götze et al., 2015a). Müller et al. (2021) has shown, however that Al and Ga concentrations in quartz correlate (Fig. 13).

Fig. 13. Aluminium vs. Ga in pegmatite quartz of different origin (data from Müller et al., 2021; NYZ = Nb/Y/F-pegmatite, LCT = Li/Cs/Ta-pegmatite according to Černỳ and Ercit, 2005).

Alkali and alkali earth elements

Alkali and alkali earth elements can occur in elevated concentrations (up to >100 ppm) in quartz. In principle, cations such as Li⁺, Na⁺ or K⁺ can be incorporated in interstitial positions (Botis et al., 2009), where Li⁺ is the most common charge-compensating ion. The lithophile character of Li probably favours its partition into silicate melts and quartz rather than in aqueous fluids. Therefore, it prefers the incorporation into structural channels for the charge compensation of Al rather than the concentration in fluid inclusions (Götze et al., 2004). Hence, the uptake of Li in the quartz lattice seems to be mostly limited by the amount of substitutional Al (Müller and Koch-Müller, 2003). The availability of Li at time of quartz crystallization is thereby also controlled by the crystallization order of Li minerals and quartz in granites and granitic pegmatites.

Although Na and K were also proven to be potential charge compensators (e.g. Heaney and Davis, 1995; Botis et al., 2009; Luo et al., 2020), these ions are more commonly hosted by micro-inclusions of minerals and fluids or accumulated along micro-fissures (e.g. Günther et al., 1998, Müller et al., 2003; Götze et al., 2004, 2017; Sittner, 2019; Ladenburger et al., 2020). For instance, K is commonly associated with Al in quartz of granitic igneous rocks and pegmatites as nano-inclusions of K-feldspar or sheet silicates (e.g. Zolensky et al., 1988; Jacamon, 2006; Jacamon and Larsen, 2009; Seifert et al., 2011). Müller et al. (2012), for instance, suggested a formation of micro-inclusions with muscovite composition due to accumulation of Al and K during deformation processes. However, the majority of K, Na, Rb, Cs, Ca, Mg and Sr seems to be bound preferentially to fluid inclusions.

In general, the trace amounts of most alkali and alkali earth elements in quartz show parallel trends because of their similar geochemical behaviour (Figs 14, 15). The data indicate that the majority of elements is bound preferentially to micro-inclusions. Chemical investigations of inclusion fluids in quartz revealed that those inclusions play a major role in hosting alkali and alkali earth elements (e.g. Rossman et al., 1987; Gerler, 1990; Gerler and Schnier, 1998; Günther et al., 1998; Götze et al., 2004; Ladenburger et al., 2020). For instance, Bottrell et al. (1988) reported evidence for the presence of Na, K, Mg and Ca in fluid inclusions of different quartz samples.

Fig. 14. Contents of alkali elements in quartz: (a) K vs. Na; (b) Rb vs. Na (red symbols) and Rb vs. K (blue symbols); (c) Rb vs. Cs. All log plots show positive trends and similar absolute concentrations of K and Na with nearly 1:1 ratio (circles = hydrothermal quartz, rhombs = pegmatite quartz; data from Götze et al. 2004, 2017).

Fig. 15. Contents of alkali and alkali earth elements in quartz: (a) Na vs. Mg; (b) Rb vs. Sr; (c) Ba vs. Mg (red symbols) and Ba vs. Sr (blue symbols). All log plots show similar trends with varying absolute concentrations (Mg > Sr); note that pegmatite quartz mostly contains < 0.1 ppm Sr and hydrothermal quartz > 0.1 ppm Sr (circles = hydrothermal quartz, rhombs = pegmatite quartz; data from Götze et al. 2004, 2017).

Rare earth elements

Rare earth elements (REE) generally have very low concentrations in quartz. Therefore, their absolute abundance is not typically determined. However, REE distribution patterns are commonly used for the reconstruction of mineral forming processes and thus, the distribution of REE in quartz can provide useful information about the geological environment of quartz formation (Monecke et al., 2000, 2002; compare with Fig. 20).

Because of their crystal-chemical properties, REE cannot be incorporated into the quartz crystal structure. It is assumed that REE in quartz are hosted preferentially in fluid inclusions. Rossman et al. (1987), Ghazi et al. (1993) and McCandless et al. (1997) reported a correlation of REE together with Rb and Sr with fluid-inclusion water indicating that these elements are mostly contained in the inclusion fluids. These results were confirmed by Götze et al. (2004) for pegmatite quartz, who found a correlation of REE contents in quartz with the absolute amount of inclusion fluid (Fig. 16). On the basis of these considerations, REE may provide important fingerprints concerning the composition of mineral forming fluids.

Fig. 16 Bulk REE content in pegmatite quartz plotted vs. absolute amount of inclusion fluid in the quartz samples; the correlation trend indicates the preferred accumulation of REE in fluid inclusions; the scatter can be explained by variations of their absolute concentrations in the fluid (data from Götze et al., 2004).

Antimony

Antimony is another trace element in quartz that can probably be related to a supply by mineralising fluids. Although the concentration of Sb in most quartz of igneous, volcanic and metamorphic rocks is far below 1 ppm (Blankenburg et al., 1994), Rusk et al. (2011) found detectable Sb contents in quartz of epithermal ore deposits. Antimony was measured at concentrations from <5 to ~120 ppm and concentrations fluctuated dramatically among quartz generations and growth zones. Similarly, Monnier et al. (2021) found highest Sb contents (up to 200 ppm) in quartz from stibnite-bearing veins, considerably higher than the values measured in quartz unrelated to Sb mineralisation (mainly ≤ 1 ppm). The lack of a consistent relationship between Sb and other trace elements in quartz lead to the conclusion that Sb incorporation in quartz is probably controlled by the content of Sb in the mineralising fluid (Rusk et al., 2011; Pacák et al., 2019; Li et al., 2020; Monnier et al., 2021). Therefore, the Sb content in quartz might be used as a potential indicator for tracing Sb deposits.

Niobium, tantalum, uranium and thorium

For a number of elements the nature of incorporation into quartz is very difficult to determine. In particular these are elements that are present in extremely low concentrations and sometimes below the detection limit of some trace-element analytical methods (e.g. U, Th, Nb and Ta). It can be assumed that the ions are either too large to substitute for the small Si⁴⁺ ion in the quartz lattice or that they are not soluble in the mineralising fluids that accumulate in fluid inclusions. This can be observed for the immobile elements niobium and tantalum in quartz of Nb–Ta mineralisation, where the extreme low concentrations of these trace elements in associated quartz do not reflect the mineralisation and, therefore, cannot be used as petrogenetic indicators (Götze et al., 2004).

Uranium and thorium are two elements that are typically present in very low concentrations in quartz. Quartz from magmatic, metamorphic rocks and pegmatites has U and Th contents at sub-ppm levels (e.g. Gerler, 1990; Blankenburg et al., 1994; Götze et al., 2004; Götze, 2009). These characteristic low concentrations in quartz can be explained by the fact that the crystal-chemical properties of U and Th (in particular ion sizes) prevent a substitutional incorporation into the quartz structure. The correlation of the U and Th concentrations (Fig. 17a) indicates that traces of both elements are situated on interstitial places in the lattice or bound on submicroscopic mineral inclusions (e.g. on grain boundaries).

Fig. 17 (a) Th vs. U concentrations in quartz samples from different parent rocks and localities (data from Götze et al., 2004, 2017; circles = hydrothermal quartz, rhombs = pegmatite quartz); (b) model of the uranyl–silicate complex in microcrystalline SiO₂ with neighbouring SiO₄ tetrahedra and uranyl polyhedra (modified after Pan et al., 2021).

Unusually high U concentrations of up to >70 ppm were detected in some microcrystalline quartz and even in macrocrystalline quartz of agates from different worldwide occurrences (Götze et al., 2009, 2015a). Results of CL spectroscopy indicate that U is incorporated into quartz as uranyl ions (UO₂²⁺), which are bound to the silica surface (Götze et al., 2015a). The presence of uranyl ions can be confirmed easily by its characteristic green luminescence (compare Fig. 10). Recent data confirmed the binding mechanism of uranyl silicate complexes in microcrystalline SiO₂ and suggested a new mechanism of uranium deposition (Fig. 17b; Pan et al., 2021). This model involves uranyl co-precipitation during silicification without the putative reduction process and provided new aspects concerning the formation of uranium deposits.

OH

In addition to water contained in fluid or melt inclusions, quartz may incorporate different amounts of OH point defects in its structure (e.g. Kats, 1962; Bambauer et al., 1962; Chakraborty and Lehmann, 1976; Stalder, 2021). The specific OH defects can be distinguished and quantified by Infrared (IR) spectroscopy based on their characteristic absorption bands. The most common type is the Al–OH defect followed by the Li–OH defect, whereas B–OH and hydrogarnet defects (4H) are less common. Systematic investigations of OH contents in different quartz types revealed that the incorporation of OH defects depends on the specific conditions of formation, but shows no correlation with the amount of molecular water incorporation (Baron et al., 2015; Jaeger et al., 2019; Stalder, 2021).

Quartz from granitic systems may show large variations in OH content with average values of ca. 20 wt. ppm (Stalder, 2021). It has been shown that Variscan samples with 20–35 wt. ppm OH contents have much higher values than Proterozoic samples from Scandinavia with ~3 wt. ppm (Müller and Koch-Müller, 2009; Stalder et al., 2017; Potrafke et al., 2020). Granitic pegmatites contain average OH contents of ca. 20 wt. ppm (Müller and Koch-Müller, 2009). OH contents of up to 13 wt. ppm (average of 6 wt. ppm) were determined for quartz from felsic volcanic rocks. Individual phenocrysts can show strong zoning with a decrease in OH from core to rim (Tollan et al., 2019; Jollands et al., 2020). Hydrothermal quartz crystals are also heterogeneous with respect to the OH content (0–225 wt. ppm) with the highest concentrations in the crystal centre (Kats, 1962; Chakraborty and Lehmann, 1976). The lowest OH contents (mostly <5 wt. ppm) were measured in quartz from metamorphic rocks (Müller and Koch-Müller, 2009; Stalder et al., 2017).

A thorough review about OH defects in quartz has been given recently by Stalder (2021).

Geological implications of quartz properties

Depending on the specific conditions of formation, quartz can develop characteristic chemical, physical and morphological properties (indicator properties), which range from the atomic scale (point defects) up to macroscopic appearance and crystal morphology. Knowledge about such characteristic features is indispensable for the recognition of any relationships between quartz genesis, its specific properties and possible industrial applications.

Trace elements in quartz as petrogenetic indicators

The trace-element distribution in quartz from different rock types can provide useful information about the formation history of these rocks, as they are highly sensitive to crystallisation and fractionation processes. For instance, trace elements in granites and granitic pegmatites reflect the evolution of the melt and can therefore be used as petrogenetic indicators. Elements such as K, Fe and Ti are geochemically compatible and therefore have the highest concentrations in early formed quartz. In contrast, P, Ge, Rb and Li are relatively incompatible resulting in elevated concentrations in quartz that formed in more evolved granitic melts and/or residual fluids (Schrön et al., 1982, 1988; Larsen et al., 2004; Beurlen et al., 2011; Monnier et al., 2018; Müller et al., 2017; Breiter et al., 2020).

Larsen et al. (2004) suggested the Ge/Ti ratio is robust during sub-solidus processes in igneous systems and concluded that it represents a strong index for the evolution of melt differentiation in granitic igneous rocks, and has potential as an igneous geothermometer (Jacomon and Larsen, 2009). The Ge/Ti ratio is a reliable melt fractionation index as Ti in quartz is a function of magma temperature and Ti melt saturation and, thus, decreases with melt differentiation, whereas Ge increases with the melt fractionation degree. A similar behaviour was proposed for the Ge/Fe ratio in quartz (Schrön, 1982, 1988). However, as noted above most of the detected elevated contents of Fe in quartz are related to micro-inclusions of Fe-bearing minerals. Thus, the ratio should not be used for genetic interpretations.

Schrön et al. (1988) noticed that elevated Ge concentrations are found in pegmatitic quartz compared to quartz from granites and rhyolites, and established the Ti ‒ Al/10 – 10Ge ternary discrimination diagram, which is still used widely. Most recently Breiter et al. (2020) proposed that this diagram does not only enable the discrimination between pegmatitic and granitic quartz, but also permits, to some extent, discrimination between common and rare-metal, and between S- and A-type granites (Fig. 18).

Fig. 18. Triangle for discrimination of S-type granites, A-type granites and pegmatites (modified after Breiter et al., 2020); the arrows indicate evolution trends during magma fractionation.

A steady increase in the Ge/Al values in quartz during the transition from magmatic to hydrothermal stages related to rare metal mineralised systems was reported by Müller et al. (2018) and Breiter et al. (2020). In the case of the Zinnwald/Cinovec Sn–W–Li greisen-type deposit the value of Ge/Al = 0.008 is a discriminator between magmatic and hydrothermal quartz (Müller et al., 2018). In the case of the Panasqueira W deposit, the value of Ge/Al = 0.01 was found to discriminate most of the magmatic quartz from hydrothermal quartz (Breiter et al., 2020).

Breiter and Müller (2009) also detected decreasing Ti content in granitic quartz with increasing fractionation. In contrast, Al contents in quartz increased from the early to the late population independent of the peraluminosity of the melt, resulting in a negative correlation between the Al and Ti contents. Accordingly, the Ti vs. Al diagram seems to be the most valuable indicator for the evolution of the melt from which the quartz crystallised (Müller et al., 2002, 2003, 2017; Breiter and Müller, 2009; Jacamon and Larsen, 2009; Breiter et al., 2012; Peterková and Dolejš, 2019). These general trends were also found in quartz from plutons of different geochemical character (Breiter et al., 2013; Hong et al., 2019). The Ge/Ti vs. Al/Ti plot shows a positive correlation reflecting the fractionation trend of granitic melts (Breiter et al., 2020).

The trace-element composition of quartz from hydrothermal environments also provides information about primary crystallisation conditions, the mineralisation regime in hydrothermal ore deposits as well as about processes of secondary hydrothermal overprint in different types of host rocks (e.g. Monecke et al., 2002; Rusk et al., 2008, 2012; Jourdan et al., 2009; Luo et al., 2020). For instance, Jourdan et al. (2009) reported Al and Li concentrations of more than several hundreds of ppm for distinct growth zones within hydrothermal quartz crystals formed at temperatures of about 300°C or less. These quartz crystals also displayed patterns of cyclic growth under CL. In contrast, quartz crystals formed at temperatures closer to 400°C and without visible cyclic growth have low concentrations of Al, Li and other trace elements.

Such variations were also reported in different quartz generations of ore mineralisation (Rusk et al., 2006, 2008, 2011; Luo et al., 2020). The variations of trace elements such as Ti, Al, Li, K or Na might be related to variations in the pressure and temperature regime, but can also result from fluctuations of the pH and composition of the mineralising solutions or crystallisation rate rather than changes in temperature. Rusk (2012) investigated quartz from almost 30 hydrothermal ore deposits including porphyry-type (Cu–Mo–Au) deposits, orogenic Au deposits, and epithermal deposits and concluded that each of these ore types can be distinguished from one another on the basis of Al and Ti concentrations alone. The temperature dependence of Ti incorporation was used to establish an Al versus Ti discrimination diagram of quartz based on the observation that the Ti content, and to the same extent the Al content, vary systematically among these ore deposit types (Fig. 19).

Fig. 19. Log plot of Ti vs. Al contents in quartz for the discrimination of hydrothermal quartz originating from epithermal, orogenic Au and porphyry type deposits (modified after Rusk, 2012).

Most porphyry-type quartz contains 1–200 ppm Ti and 50–500 ppm Al showing a linear correlation with an Al/Ti ratio between 1 and 10. In contrast, quartz from epithermal ore deposits always contains < 3 ppm Ti with a wide scatter of Al between 20 and 4000 ppm (Al/Ti ratios 100–10,000), whereas quartz from orogenic Au deposits mostly has 100–1000 ppm Al and 1–10 ppm Ti (Al/Ti ratio 10–100; Rusk, 2012). These systematic variations in trace-element concentration reflect the different physical and chemical conditions during quartz formation.

Concentrations of REE in quartz can also be used as valuable indicators for geological and/or geochemical processes, as they can reflect the evolution of mineralising fluid and melt compositions. Thus, REE distribution patterns contain important information concerning the primary conditions of quartz formation and also information about secondary effects (Fig. 20). The general shape of the REE distribution and the relationships between light, middle and heavy REE indicate the source of the silica-bearing fluids and the influence of the complexation of the REE during transport (Lipin and McKay, 1986). Moreover, negative and positive anomalies of Ce and Eu can reflect adopted properties from the mineralising fluids and/or variations in redox conditions resulting in a change of the valence state of the ions (Fig. 20).

Fig. 20. Chondrite-normalised REE distribution patterns of quartz from different geological environments: (a) hydrothermal quartz from the TAG mound (Mid-Atlantic Ridge), high-purity quartz (HPQ) from Vjazovka (Ural region, Russia) and from the gold deposit Muruntau/Myutenbai (Uzbekistan); (b) pegmatite quartz from Hittero (Norway) and macrocrystalline (agate_qtz) as well as chalcedony (agate_chal) of an agate from Hohenstein-Ernstthal (Saxony, Germany) both showing positive Ce-anomalies, negative Eu-anomalies and tetrad effects; (c) metamorphic quartz from the Middle Erzgebirge (Germany) and overprinted counterpart from the area of tin mineralisation; (data from Monecke et al., 2000, 2002; Götze et al., 2004, 2016, 2017).

The hydrothermal quartz samples in Fig. 20a represent different hydrothermal systems and accordingly display characteristic REE distribution patterns. The chondrite-normalised REE distribution pattern of the hydrothermal quartz from the Trans-Atlantic Geotraverse (TAG) mound in the Mid-Atlantic Ridge (Fig. 20a) is bell-shaped with no significant Ce or Eu anomaly. This REE distribution can be related to the origin and evolution of the fluids within the hydrothermal system that is related to an actively forming sulfide deposit (Monecke et al., 1999). In contrast, the high-purity quartz from Vjazovka (Ural region, Russia) and the quartz from the hydrothermal gold deposit show typical crustal signatures with a smooth chondrite-normalised REE pattern with decreasing REE concentrations from La to Lu. The negative Ce- and positive Eu-anomalies, result from the evolution of the hydrothermal fluids in the deposits and their interactions with the surrounding wall rocks.

Unusual REE fractionation and the development of tetrad effects (Fig. 20b) can be related to fluid immiscibility and preferential partitioning of the REE between the vapour and coexisting liquid (Monecke et al., 2007). Tetrad effects were observed both in quartz of pegmatites (Götze et al., 2004) and quartz of agates in acidic volcanic rocks (Götze et al., 2016) and probably reflect the participation of volatiles during SiO₂ accumulation and crystallisation. The marked increase of heavy REE indicates the complexation of REE by fluorine or carbonate complexes (Wood, 1990).

The hydrothermal overprint of crystalline host rocks in the vicinity of ore deposits can also result in a change of the REE distribution pattern. An example from quartz in unaltered metamorphic host rocks located distal to the alteration halo within a tin deposit is shown in Fig. 20c. The hydrothermally overprinted quartz shows REE signatures that deviate significantly from the unaltered equivalent. These differences are interpreted to result from the interaction of the pre-existing quartz with the hydrothermal fluids (Monecke et al., 2004).

Several studies have shown that the trace-element composition of quartz can be used to fingerprint the source of detrital quartz grains in clastic sediments (Götze and Zimmerle, 2000). Quartz can retain chemical signatures of the source rocks because of its common occurrence and relative mechanical and chemical stability during sedimentary processes. These geochemical signatures can be used to reveal the origin and provenance of individual quartz grains within sediments. Differences in the distribution of selected trace elements and/or trace-element ratios may be related to inherent characteristics of the source and can be used as discriminating tools (e.g. Dennen, 1967; Hallbauer, 1992; Götze and Lewis, 1994; Müller and Knies, 2013; Ackerson et al., 2015; Stalder et al., 2019).

Another common feature in sedimentary and diagenetic environments is secondary quartz neomorphism as quartz overgrowth cements in sandstones or isolated euhedral quartz crystals in various sedimentary host rock. Authigenic quartz cements play a major role in controlling the hydraulic properties and quality of reservoir sandstones (McBride, 1989). The combination of trace-element studies with other analytical methods (e.g. CL, fluid inclusion and isotope studies) can provide important information about timing of quartz cementation and the burial history of sandstones or concerning temperature and salinity of fluids during precipitation (e.g. Burley et al., 1989; Kelly et al., 2007; Lehmann et al., 2011).

Moreover, some investigations have used complex geochemical and mineralogical studies for the reconstruction of silica sources and diagenetic conditions during the formation of euhedral authigenic quartz crystals in soil, limestones and carbonates, sulfate and salt deposits as well as bituminous coal and lignite (e.g. Füchtbauer, 1961; Grimm, 1962; Fabricius, 1987; Richter, 1971; Friedman and Shukla, 1980; Ruppert et al., 1985; Soong and Blattner, 1986; Götze, 2012b). The resulting mineralogical and geochemical characteristics show that the authigenic quartz from the various sedimentary environments differs clearly from quartz of crystalline rocks and a distinction between quartz from different sedimentary environments is possible.

Interpretation of CL and trace-element data

In general, quartz commonly shows growth zoning in CL, reflecting either variations of the composition of the crystallisation medium (melt or fluid) or changing physicochemical conditions (temperature, Eh, pH). Therefore, combined CL and trace-element studies have significant potential for the reconstruction of crystallisation sequences in magmatic and hydrothermal systems (e.g. Kempe et al., 1999; Landtwing and Pettke, 2005; Müller et al., 2005, 2010; Jourdan et al., 2009; Götte et al., 2011; Audétat, 2013; Mao et al., 2017; Wertich et al., 2018; Wang et al., 2019; Luo et al., 2020; Lan et al., 2021).

Although a correlation of the CL response and the concentration of trace elements has been noted by several authors (e.g. Watt et al., 1997; Müller et al., 2003; Rusk et al., 2006, 2008, 2011; Lehmann et al., 2011; Leeman et al., 2012), CL zoning (in particular in panchromatic CL images) in quartz cannot be related solely to variations in the trace-element concentration. For instance, Watt et al. (1997) analysed volcanic quartz phenocrysts and xenocrysts by ion microprobe (SIMS) and found higher Al contents in areas of bright CL emission, though they suggested that other factors (e.g. the effect of growth rate on defect density) may also play a major role in controlling CL intensity. Several important aspects limit the possible correlation between the abundance of specific trace elements and visible CL colours or CL zoning.

In summary, it must be emphasised that most of the CL emissions in quartz are related to intrinsic lattice defects and not to trace elements. Therefore, a correlation of CL response to trace elements is often very equivocal. Among the detectable trace elements in quartz only a few of them act as activators for CL emission (compare Table 1). Only Al, Ti and charge compensating alkali ions in hydrothermal and pegmatite quartz sometimes show relations to visible CL behaviour. Most of the inherited trace elements in quartz have no influence on the CL behaviour. To attempt to correlate the concentration of selected trace elements with the visible CL is not reasonable. Spectral CL measurements are necessary and it is only possible to compare trace-element variations with specific individual CL emission bands. In most cases, the bulk CL emission consists of different emission bands and thus, represents a mixture of point defects responsible for the CL signal. Therefore, correlative relations between the CL and single trace elements are very difficult or impossible. In addition, the complexity of the luminescence process and analytical conditions hamper a clear quantitative evaluation of the CL spectra (Habermann, 2002; Götze, 2012a).

The example in Fig. 21 illustrates these complex relationships and demonstrates possible misinterpretations. The hydrothermal quartz crystal exhibits marked zoning in panchromatic CL (Fig. 21a), visible by variations in the intensity of the bulk (cumulative) CL signal (bright and dark zones). Local measurements of the Al content show the dark zones have the highest concentrations (300 ppm) and the bright zones have low (<50 ppm) and not detectable Al concentrations. Attempting to correlate the Al concentration with the panchromatic CL intensity would result in the conclusion that Al correlates negatively with the CL intensity (brightness). However, the corresponding true colour CL image (Fig. 21b) demonstrates that the dark zones exhibit blue CL, whereas bright areas are dominated by yellow CL. Related CL spectra of these zones (Fig. 21c) reveal that the dark areas (transient blue CL) are caused by activation due to Al defects ([AlO₄/M⁺] centres), whereas the bright areas (yellow CL) have no relation to trace elements and are exclusively due to structural defects (e.g. STE). In conclusion, the visible zoning in brightness in panchromatic CL (Fig. 21a) is caused mainly by lattice defects and is essentially not related to trace-element variations. This interpretation was only possible by the application of CL spectroscopy.

Fig. 21 Photomicrographs of intensely zoned hydrothermal quartz from Chemnitz (Saxony, Germany) in panchromatic CL (a) and true colour optical microscopy–cathodluminescence image (b); the dark luminescent zones in panchromatic CL (blue CL zones in b) have elevated Al concentrations up to 300 ppm, whereas the bright zones in panchromatic CL (yellow in b) have mostly low (< 50 ppm) or no detectable Al contents; Ti concentrations are generally < 10 ppm. (c) The spectra reveal that the blue CL (1) is caused by the incorporation of structural Al ([AlO₄/M⁺] centre), the yellow CL (2) is related mainly to structural defects (e.g. STE) and not to trace elements; the circles in (a) mark the positions of spectral CL analyses.

This example shows the essential problems with CL data, in particular for the interpretation of panchromatic CL images. Even the combination of CL spectroscopy with local trace-element analysis never provides a quantitative relationship between trace elements and the intensity of specific emission bands. For instance, Götte et al. (2011) demonstrated the correspondence of Li abundances with [AlO₄|Li⁺]^– defects and the transient CL emission at 395 nm in hydrothermal quartz, however they found inconsistencies with a simple causal relationship between Al-Li centres and the 395 nm emission band. They assumed that the conversion of [AlO₄|Li⁺] centres to [AlO₄]⁰ centres by natural irradiation and/or the existence of intrinsic defects such as oxygen deficiency centres can affect the luminescence properties. Similar results were reported by Sittner (2019) for the CL of pegmatite quartz. The 500 nm emission band could clearly be ascribed to structurally incorporated Li, but a clear quantitative correlation between Li concentration and intensity of the 500 nm emission band was missing. The incorporation of other monovalent cations and the charge-compensation of Al³⁺ by electron defects were discussed as possible explanations.

Another problem is the ongoing discussion about the role of Ti as an activator of CL in quartz. Observations about correlations between the Ti content in quartz and the CL intensity resulted in the conclusion that the blue CL emission band at 450 nm might be activated by Ti (e.g. Müller et al., 2002, 2003; Van den Kerkhof et al. 2004; Rusk et al., 2008; Drivenes et al., 2016). First attempts have been made to quantify the Ti concentration in quartz by using the intensity of the 450 nm emission band (Leeman et al., 2012). These assumptions are in sharp contrast to previous investigations, which have provided evidence by spectroscopic methods that the blue 450 nm emission can be related to oxygen deficiency centres (ODC) (e.g. Stevens-Kalceff and Phillips, 1995; Skuja, 1998; Stevens-Kalceff, 2009).

Recent studies on quartz of different origin and containing different concentrations of Ti revealed that structural defects not related to trace elements (self-trapped exciton – STE) and also incorporation of Ti into the quartz structure can both activate a blue CL emission at the same position (450 nm, 2.75 eV; unpublished data). The STE-activated CL emission is unstable under the electron beam and in particular detectable in quartz with low Ti contents (e.g. high-purity hydrothermal quartz), whereas high Ti concentrations (e.g. in quartz phenocrysts of rhyolites) result in a stable CL emission (Fig. 22). In conclusion, even in the case of Ti activation of the 450 nm luminescence emission in quartz, the superposition with the STE emission at the same position would prevent any serious quantitative evaluation of the CL signal. A reliable distinction of the two different CL emissions at 450 nm is only possible by time-resolved CL spectroscopy combined with EPR spectroscopy and trace-element analysis of the Ti concentration. This example emphasises that the correlation of CL and trace-element data has to be done with caution.

Fig. 22. CL emission spectra of (a) a quartz phenocryst in the Leisnig porphyry (Saxony, Germany) and (b) high-purity quartz from Kuznechikhinsk (Ural, Russia). Both quartz samples exhibit a blue CL (see insets) with a dominant emission band at 450 nm. The blue CL emission band of the Ti-rich quartz phenocryst (up to 80 ppm Ti and paramagnetic [TiO₄/M⁺]^– centres) is stable during electron irradiation and can probably be related to the activation by substitutional Ti. In contrast, the intensity of the STE-activated 450 nm emission band of the HPQ quartz (4.52 ppm Ti, no detectable Ti-centres) shows a strong decrease due to electron bombardment.

Conclusions

Quartz is one of the most common minerals in the Earth crust and occurs ubiquitously in magmatic, metamorphic and sedimentary rocks. This common occurrence makes it an important pathfinder mineral for the reconstruction of geological processes and palaeoclimatic conditions as well as the exploration of mineral deposits and other natural resources.

Although quartz has a simple composition and represents a very pure mineral with low abundances of defects and impurities, it develops significant indicator properties depending on its formation history and the influence of secondary processes. Analytical techniques with low detection limits and/or high spatial resolution can provide a wealth of data concerning the real structure and chemistry of natural and synthetic quartz. The analytical combination of CL microscopy and spectroscopy, EPR spectroscopy, as well as spatially resolved trace-element analysis has been demonstrated to characterise the type and abundance of structural defects and trace elements in quartz, and permit reconstruction of processes responsible for their formation and secondary effects. Each of these analytical techniques can provide specific information about the point defects and trace-element uptake in quartz.

It has been shown that quartz from different geological settings might have different trace-element composition related to the different supply and mechanisms of incorporation of certain elements. Therefore, trace elements in quartz are important petrogenetic indicators in geosciences. In addition, intrinsic types of point defects detected by EPR can provide valuable genetic information. Cathodoluminescence can help to visualise the real structure of quartz and to relate internal textures to possible formation processes.

However, interpretation of analytical data might be limited by the ability of quartz to regenerate during secondary alteration processes occurring under metamorphic or hydrothermal conditions. Therefore, primary genetic information concerning the real structure or trace-element composition might be obliterated during quartz alteration or regeneration (Kempe et al., 2012). In these cases care has to be taken when interpreting the genetic information encoded. Distinction of features related to primary growth or secondary alteration of quartz is not simple and requires application of complementary analytical techniques.