OH defects in quartz in granitic systems doped with spodumene, tourmaline and/or apatite: experimental investigations at 5–20 kbar

Abstract

The incorporation of OH defects in quartz as a function of Li content in the bulk system and pressures was investigated. Quartz crystals were grown in water-saturated granitic systems, containing various amounts Li, B and P, supplied as accessory phases such as spodumene, tourmaline or apatite in the starting mixtures. High pressure experiments were performed at temperatures between 900 and 1100 °C, and pressures between 5 and 20 kbar with a piston cylinder apparatus, and the synthesized quartz crystals were analyzed by IR spectroscopy, electron microprobe and LA-ICP-MS spectroscopy. All IR absorption spectra revealed absorption features that can be assigned to AlOH (3313, 3379 and 3431 cm⁻¹) and (4H)_Si defects (3585 cm⁻¹), whereas quartz grown in the Li and B systems exhibited two additional bands related, respectively, to LiOH (3483 cm⁻¹) and BOH defects (3596 cm⁻¹). It was further observed that LiOH incorporation increases with higher spodumene content in the starting material and decreases with pressure, until no LiOH defects are observed at pressure higher than 15 kbar. Specifically, the most pronounced reduction of LiOH defects occurs in a rather narrow pressure interval (10–15 kbar) close to the high-quartz/low-quartz transition. However, the link between the transition and the defect incorporation remains unclear. Li total concentrations always exceed the Li-coupled LiOH defects, suggesting the simultaneous presence of dry AlLi defects. Results of this study suggest that LiOH defects are detectable only in quartz crystals grown from middle and upper crustal sections (such as hydrothermal quartz) and not in quartz from deep roots of orogenic granitoids.

Introduction

Quartz is one of the most abundant minerals on the Earth’s crust. None of the less quartz is a phase that occurs in a broad variety of rocks and therefore crystallizes and equilibrates at different petrogenetic conditions. Quartz may contain different amounts of impurities such as H, Li, B, Na, Al, P, K, Ti, Fe, and Ge depending on the formation conditions at different geological environments (e.g., different temperatures, pressures and chemical environments). Some of these trace elements (i.e., Al and Ti) have been successfully applied as geothermometers (Dennen et al. 1970; Wark and Watson 2006; Thomas et al. 2010; Huang and Audétat 2012).

Methodologically, it is possible to determine the quartz grains provenience (1) by means of their luminescence caused by trace metal defects in cathodoluminescence (CL) spectroscopy (Zinkernagel 1978; Matter and Ramseyer 1985; Richter et al. 2003); (2) by chemical analysis of their trace content of impurity metals applied with the electron microprobe or LA-ICP-MS (Müller et al. 2003; Müller and Koch-Müller 2009), or (3) by their inventory of hydrous defects that charge compensate metal impurities involving mono-, tri- or pentavalent cations. In this respect, many studies focused on the determination of species and concentration of hydrous defects in quartz by Fourier transform infrared (FTIR) spectroscopic analyses. Trace elements and hydrous defects are often incorporated in quartz by coupled substitutions of one silicon cation by trivalent (e.g., Al³⁺, B³⁺ and Fe³⁺) and monovalent (e.g., H⁺, Li⁺, K⁺ and Na⁺), or trivalent and pentavalent (e.g., P⁵⁺) cations. Substitutions such as Si⁴⁺ = Al³⁺ + H⁺ cause AlOH defects, which are characterized by three absorption bands at 3313, 3379 and 3431 cm⁻¹ in the infrared range (Kats 1962; Bambauer 1963; Aines and Rossman 1984; Stalder and Konzett 2012). Other coupled substitutions involving the incorporation of protons in the quartz structure such as Si⁴⁺ = B³⁺ + H⁺ and Si⁴⁺ = 4H⁺ generate the BOH (Müller and Koch-Müller 2009) and the (4H)_Si hydrogarnet defect (Paterson 1986; Stalder and Konzett 2012), characterized by the bands at 3596 and 3585 cm⁻¹, respectively. The occurrence of the interstitial hydrous defects such as LiOH causes the absorption band at 3483 cm⁻¹ (Kats 1962; Aines and Rossman 1984; Baron et al. 2015).

The intensity of these absorption bands depend on the prevailing conditions during crystallization or annealing and may be used to define the provenance of an individual quartz grain. For example, the hydrogarnet defect is a common point defect in synthetic quartz crystallized in a chemically pure system (Chakraborty and Lehmann 1976; Paterson 1986; Rovetta 1989). Concentrations of hydrogarnet in quartz from granitic systems are correlated to pressure (Stalder and Konzett 2012) and therefore may be used to distinguish between low- and high-pressure origins, while the concentration of AlOH defects is both temperature and pressure dependent, and cannot give additional information about the genesis conditions. In contrast, the concentration of BOH defects is not pressure dependent (Baron et al. 2015). Generally, quartz crystals of magmatic origin are mainly characterized by AlOH defects and therefore by higher incorporation of H⁺ with respect to other rocks (Müller and Koch-Müller 2009; Stalder and Neuser 2013). In addition, rocks of pegmatitic origins are enriched in LiOH and BOH defects (Stalder and Neuser 2013). The major purpose of this work is a more detailed study to assess LiOH defects in the quartz structure depending on pressure and chemical system, with the aim to use this feature as petrological indicator. To achieve this goal, several high-pressure experiments were performed on a granite–spodumene system at pressures ranging from 5 to 20 kbar, which represents the pressure conditions from upper to lower crust in collision zones. Furthermore, few additional experiments on granite containing tourmaline and/or apatite were performed, in order to evaluate whether the charge neutrality equation H⁺ + A⁺ + P⁵⁺ = M³⁺ + B³⁺, where A = alkali and M = metal ions (Müller and Koch-Müller 2009; Baron et al. 2015), is achieved.

Experimental and analytical methods

Starting materials

Quartz crystals were grown in the following systems: (1) quartz–granite–spodumene–water, (2) quartz–granite–tourmaline–water, (3) quartz–granite–apatite–water, (4) quartz–granite–spodumene–apatite–water and (5) quartz–granite–tourmaline–apatite–water. All starting materials consisted of synthetic quartz (Alfa Aesar 99.995 % purity) and natural granite from Sardinia/Italy, which was derived from the institute rock collection (68.90 wt% SiO₂, 14.87 wt% Al₂O₃, 6.35 wt% K₂O, 3.05 wt% Na₂O, 2.51 wt% Fe₂O₃, 1.56 wt% CaO, 0.50 MgO, 0.27 wt% TiO₂, 0.08 wt% MnO, 0.04 wt% P₂O₅). Both synthetic quartz and natural granite powders were mixed in proportion of 1:1, partially molten in a platinum crucible at 1100 °C for four hours, in order to promote the homogenization of the granitic components, and rapidly quenched to glass. The resulting glass + quartz mixture was successively ground to powder and finally mixed with different amounts of powdered spodumene, tourmaline and/or apatite, for doping the system with Li, B, and P, respectively. Details concerning the starting material are reported in Table 1.

Table 1 Experimental conditions and starting material

High pressure experiments

Each solid mixture was filled together with 12–25 wt% water in Pt capsules with an outer (inner) diameter of 4.0 (3.6) mm. The capsule was then closed, welded shut, weighted and checked for leakage in an oven at 120 °C. The capsule was finally inserted in a BN sleeve, which produces an oxygen fugacity between CoCoO and WM (Truckenbrodt et al. 1997). All syntheses were performed in an end-loaded piston cylinder press using a NaCl assembly with graphite furnace. For experiments with pressures from 5 to 7.5 kbar a 22-mm assembly in combination with Pyrex sleeve was used, while a half-inch NaCl assembly was used for experiments at pressures from 10 to 20 kbar. Temperatures were monitored by using Ni–CrNi thermocouples. In order to promote crystal growth, the temperature was initially raised to 1100–1000 °C (Table 1), followed by a ramp down 900 °C at a cooling rate of 20 °C/h. Temperatures were hold for 50–120 h (Table 1), and runs were terminated by shutting off the power supply.

The recovered capsules were then weighted, pierced, dried and weighted again, for verifying any fluid loss and quantifying the amount of water retained in the run products. If the products were fluid-saturated, it was possible to measure the fluid’s pH, by pH indication paper, after piercing the capsule. The pH of the excess fluid was neutral or very weak alkalic (pH 6–7) for all the runs. Most of the run products consisted exclusively of quartz crystals and quench material. In some runs, also spodumene (run: E30) and apatite crystals (runs: E29 and E30) could be observed (Fig. 1c), but may have also been present in very small proportions in other runs. It was also observed that quartz’s crystal size changes in function of pressures and bulk composition of the starting material. Quartz crystals were then individually handpicked, mounted in thermoplastic resin and oriented parallel to the c axis with a polarizing microscope. Birefringence values of Δn = 0.009 in the orthoscopic view and “flash figures” in the conoscopic view confirmed a successful orientation of the quartz crystal. A λ-plate compensator was employed to determine the direction of the ordinary (n _o) and extraordinary (n _e) refractive indices, an essential prerequisite to perform polarized measurements with FTIR spectroscopy. Oriented crystals were manually ground, polished on both sides and finally removed from the resin by dissolution in acetone. Crystal thicknesses were measured with a mechanical Mitutoyo micrometer with an accuracy of ±2 µm. Characteristics of the prepared crystals including water concentrations are reported in Table 2.

Fig. 1

BSE image of quartz (Qz) crystals in water-saturated melt from run E17 (a) and E20 (b) and euhedral apatite (Ap) crystals from runs E29 (c)

Table 2 IR measurements

FTIR spectroscopy

Mid-infrared spectra were recorded at room temperature in transmission mode by using a Bruker Vertex 70 FTIR spectrometer coupled to a Hyperion 3000 microscope, a nitrogen-cooled MCTD316-025 (mercury cadmium telluride) detector, a globar light source, a KBr beamsplitter and a wire grid polarizer. Background and sample measurements were taken by averaging 32–300 scans, depending on the measured area that varied from 100 × 100 to 40 × 40 µm². Polarized measurements were taken along the ordinary (n _o) and extraordinary (n _e) refractive indexes on the same spot by turning the polarizers’ direction by 90°. All IR bands related to OH defects are strongly polarized in the direction parallel to n_o and exhibit almost no absorbance in the direction parallel to n _e. Fluid/melt inclusions exhibit weak and broad absorption features in the OH range. Because of their isotropic behavior, the contribution of molecular water trapped as fluid micro-inclusions or in quench hydrous melt inclusions was eliminated by subtracting the n_e component from the n_o component of the OH absorption pattern according to the procedure described by Stalder and Konzett (2012). An additional correction described from Baron et al. (2015) was required for spectra containing BOH absorption bands, since the B-related OH defects exhibit a significant component for E||n _e. Water concentrations were calculated by integrating the polarized background-corrected IR spectra (normalized to thickness) across the region of absorption due to OH dipole vibrations (3700–3200 cm⁻¹). The integrated extinction coefficients from mineral-specific (Aines et al. 1984; Thomas et al. 2009) and general wavelength-specific (Libowitzky and Rossman 1997) calibrations were used. All three calibrations yield for most samples results in excellent agreement to each other. Significant deviations only occur for spectra with strong absorptions in the high wavenumber range (Stalder and Konzett 2012; Baron et al. 2015), which are not relevant for the present study.

Electron microprobe

Textures from run products were inspected under a JEOL JXA-8100 electron microprobe (Fig. 1), and major and minor element chemistry of quartz crystals and quenched melt were determined using 15 kV acceleration voltage. Concerning the beam current different analytical conditions were used depending on the measured material, 50 nA beam current on quartz for acquisition of the Al-Kα line using a raster size of 12 × 9 µm, and 5 nA beam on quenched melt using a 6 × 8 µm raster size. As standards, the following minerals were used: Si: quartz, Al: corundum, K: orthoclase, Na: jadeite, P: apatite, Fe: almandine, Mg: MgO, Ca: diopside, Ti: rutile, Mn: tephroite.

LA-ICP-MS

Trace element concentrations in quartz grains were analyzed by LA-ICP-MS (laser ablation unit: ESI NWR 193; ICP-MS: Agilent 7500ce) at the Central Lab for Water, Minerals and Rocks, NAWI Graz, Karl-Franzens-University of Graz and Graz University of Technology). The material was ablated by using a 193 nm laser pulsed at 10 Hz, 50 µm spot size with energy of ~7.5 J cm⁻² and 11 Hz for measurements of 35 µm spot size, with energy of ~8.5 J cm⁻². Helium was used as carrier gas at ~0.6 l/min flow, and data were acquired in time-resolved mode. For each analysis a 30 s gas blank was obtained for background correction, followed by 60 s of active laser and a wash out time of 45 s. NIST612 glass was used for standardization and drift correction, while NIST614 and SL-1 (G-Probe proficiency program, www.geoanalyst.org) were analyzed as unknowns to monitor the accuracy of the measurements. Both standards could be reproduced within ± 10 % of the recommended values. Data reduction for quartz crystal analyses was performed using the software Glitter with silicon as internal standard, while element concentrations in the quench material was determined successively using the software SILLS (Guillong et al. 2008), which was gently granted from Dr. Markus Wälle from ETH, Zürich. For NIST612, the concentration values reported by Jochum et al. (2011) were used.

Results

IR band assignment and determination of water content as OH defects

All the IR absorption spectra revealed at least two specific absorption features, in particular a dominating triplet at 3313, 3379 and 3431 cm⁻¹, and a much weaker band at 3585 cm⁻¹, which can be assigned to AlOH defects (Kats 1961, 1962; Bambauer 1963) and hydrogarnet (4H)_Si defects, respectively. In addition, quartz crystals from the spodumene-bearing system show another band at 3483 cm⁻¹ at pressures ≤10 kbar. This band corresponds to the LiOH defects and its absorbance is negatively correlated to pressure (Fig. 2). Quartz crystals grown in the tourmaline-bearing systems show an additional band at 3596 cm⁻¹ that can be assigned to BOH defects (Müller and Koch-Müller 2009). The existence of a further band at 3469 cm⁻¹ is not as clear. If this band is considered real, it may represent another Li-related defect (Kats 1961). Crystal’s grown in systems containing apatite did not show any additional bands; however, it is observed that the presence of apatite component in the bulk system show a weak negative effect on the intensities of OH bands related to trivalent cations, e.g., the B-related band at 3596 cm⁻¹ and the Al-related band at 3379 cm⁻¹ (Fig. 3).

Fig. 2

IR spectra from quartz grown at different pressures (all with 2 % spodumene in the system) and different spodumene content in the system (all at 10 kbar)

Fig. 3

IR spectra from quartz synthesized at 10 kbar in different bulk compositions containing different amounts of apatite, tourmaline and spodumene

Influence of chemical and physical parameters on OH incorporation

At 10 kbar, the LiOH concentration is correlated to the spodumene content in the starting material (Fig. 4). As this trend is followed up to high spodumene contents (i.e., 20 wt% in the bulk system), no evidence for saturation with respect to LiOH in quartz is apparent. At 5 kbar, very high LiOH concentrations in quartz are already observed at much lower spodumene contents (i.e., 2 wt%) in the system, suggesting that the saturation is reached at lower concentration levels and spodumene is less dissolved in the melt at lower pressure. Another experiment performed at 5 kbar with higher concentrations of spodumene in the bulk system would be critical to understand whether LiOH saturation already occurred. Such an experiment was performed (E33, Table 1). However, no crystals (neither spodumene nor quartz) could be observed in the run product neither under petrographic microscope, nor by XRD. Either the system was super-liquidus or nucleation was too sluggish to form crystals.

Fig. 4

OH content in quartz as LiOH defects against the spodumene content in the starting material (as wt ppm water)

As far as the pressure effect on the OH incorporation in quartz is concerned, crystals grown from the starting material containing 2 % spodumene show not a clear trend, but a jump for the AlOH defects from ca. 60 wt ppm water at 5 kbar to a constant value of ca. 40 wt ppm water starting from 12.5 kbar (Fig. 5). This result slightly modifies previous experimental results on quartz in granitic systems, where a negative relation between AlOH concentrations and pressure was observed (Stalder and Konzett 2012; Baron et al. 2015).

Fig. 5

Water concentrations as total amount of OH defects (open circles) and as Al-specific OH defects (black circles) from quartz grown in the 2 % spodumene system. The concentrations are expressed as wt ppm water using the calibration of Libowitzky and Rossman (1997). Error bars correspond to an absolute error of 20 %

Similar to AlOH, the LiOH content in quartz shows an evident negative correlation with pressures, and—at 2 wt% spodumene in the starting material—basically drops to zero at 12.5 kbar (Fig. 6). If we assume that similar trends occur at higher spodumene contents, LiOH defects could be present up to 15 kbar, but would disappear even in extremely Li-rich systems at even higher pressures.

Fig. 6

Content of LiOH defects against pressure from quartz grown in the granitic system containing 2, 10 and 20 % spodumene

Electron microprobe and LA-ICP-MS results

Quartz, melt and accessory phases such as apatite were observed in BSE images (Fig. 1). Comparison of quantitative analyses from electron microprobe and LA-ICP-MS was used as tool to discriminate between measurements of inclusion-free quartz and those containing melt inclusions (see Fig. 1a). The following two criteria were interpreted as sign for melt contamination: (1) an unreasonable high concentration of K (>100 ppm), as extremely incompatible element, and/or (2) a deviation in Al content measured by LA-ICP-MS by more than 3σ from the value determined by electron microprobe. Data points that showed such characteristics were considered as contaminated by tiny melt inclusions and were omitted from the data set. Data points from inclusion-free quartz and melt compositions from electron microprobe analyses are reported in Table 3. At constant temperature, Al in quartz shows a negative trend against pressure and decreases from 1000 ppm at 5 kbar to values around 250 ppm at 20 kbar (Fig. 7), in good agreement with the results of Stalder and Konzett (2012). The analytical totals from the measurements taken on melts are below 100 wt%, suggesting dissolution of water in the melt. Melts that formed between 7.5 and 12.5 kbar (i.e., below the second critical endpoint) typically show analytical totals of ~90 wt%, indicating ~10 wt% water content, whereas amorphous quench formed at 20 kbar show analytical totals of ~70 wt%, which corresponds to ~30 wt% water (Table 3).

Table 3 Trace elements content in quartz and melt from electron microprobe analyses

Fig. 7

Al contents in quartz against pressure from EMPA (open circles) and LA-ICP-MS (black circles) from runs containing 2 % spodumene. Areas in gray represent the results from Stalder and Konzett (2012)

LA-ICP-MS analytical results are reported in Table 4. Quartz crystals from some runs could not be analyzed, either because the grain size was too small to perform analyses on inclusion-free spots either due to the explosion of the sample caused by the overheating of fluid inclusions beneath the measured surface during laser irradiation. LA-ICP-MS data points from the quench material exhibit a very high standard deviation, which in some cases was even higher than the mean value. However, the ratio Al/Li within one sample was nearly constant, and therefore, results from the quench material are reported as Al/Li ratio.

Table 4 LA-ICP-MS measurements

Discussion

Li saturation

Based on the observation of Li-specific (3483 cm⁻¹) and Al-specific (3379 cm⁻¹) OH bands, different OH environments exist. The linear correlation of Li-specific OH and spodumene content in the starting material observed at 10 kbar (Fig. 4) suggests that the system is not saturated with respect to Li even at highest Li content investigated at 10 kbar. If spodumene is added to undoped granite up to 10 wt% spodumene in the starting material, the resulting quartz increases its Li-specific OH without changing the Al-specific OH significantly (Fig. 8). Further input of spodumene leads to a further increase in Li-specific OH accompanied with a decrease of Al-specific OH. This leads to the conclusion that Li is incorporated in different local environments, one of which increasing the total OH (presumably as LiOH) and the other counteracting OH incorporation by combining with Al (Si⁴⁺ = Al³⁺ + Li⁺) and reducing the coupling capacity of Al to protons (Si⁴⁺=Al³⁺ + H⁺). The kink in the LiOH–AlOH-plot may further suggest that saturation with respect to Li was reached (see dotted line in Fig. 4).

Fig. 8

Relation of LiOH defect and AlOH defect concentrations in quartz at 10 kbar. The kink in the trend suggests a change in the saturation regime (see text for discussion)

Charge balance involving H and other impurities

According to previous studies charge balance of impurities in quartz is achieved by H⁺ + A⁺ + P⁵⁺ = M³⁺ + B³⁺, with A = alkali ions and M = metal ions (Müller and Koch-Müller 2009; Baron et al. 2015). In order to account for the contrasting role of Li, two different species for Li have to be defined: (I) Li_I decreasing the OH content by forming a dry AlLi defect, and (II) Li_II increasing the OH content by incorporation of LiOH and formation of non-bridging oxygens (Kats 1961). Li_II was calculated from the absorption band in the IR spectrum, and Li_I was calculated by subtracting Li_II from total Li determined by LA-ICP-MS. Data points show a linear correlation (Fig. 9), but deviate systematically by some 15 % toward the trivalent cations. The reason for this mismatch is unknown. A major contribution of incorrect Li assignment is excluded due to the rather low Li concentrations. A better explanation is the accuracy of the IR calibration for the quantification of water contents.

Fig. 9

Coupled substitutions of H⁺, Li⁺, K⁺, Al³⁺, P⁵⁺ and B³⁺. Li_I corresponds to AlLi defects and Li_II to LiOH defects

Pressure effect

IR spectra of OH defects in quartz show systematic variations with pressure and the Li-specific band at 3483 cm⁻¹ is reduced with increasing pressure at a given starting composition. If we consider endmember compositions for coupled substitutions such as HAlO₂ (for Si⁴⁺ = Al³⁺ + H⁺) and LiAlO₂ (for Si⁴⁺ = Al³⁺ + Li⁺), increasing pressure shifts the reaction

$${\text{HAlO}}_{2} + {\text{ LiOH}} = {\text{LiAlO}}_{2} + {\text{ H}}_{2} {\text{O}}$$

(1a)

or

$$2{\text{ SiO}}_{2} + {\text{HAlO}}_{2} + {\text{LiOH}} = {\text{LiAlSi}}_{2} {\text{O}}_{6} + {\text{H}}_{2} {\text{O}}$$

(1b)

to the right and in this way reduces the total amount of OH defects. Based on the present results most of the shift in the reaction seems to occur at a rather narrow pressure interval, namely between 10 and 15 kbar. In this pressure interval falls the transition from high-quartz to low-quartz (Mirwald and Massonne 1980). If the high-/low-quartz transition were relevant for the preferred incorporation of LiAl defects, we would expect lower LiAl defects at high pressure, because LiAlSi₂O₆ can crystallize in high-quartz structure (London 1984), which is stable below 13 kbar. However, according to reaction (1a) and (1b), the LiAl component is better incorporated above 13 kbar, which makes a connection of this hypothesis and the high-/low-quartz transition questionable. On the other hand, as far as total concentrations of Li and Al are concerned, the negative correlation to pressure (Fig. 7) seems to be in accord to the preferred LiAl coupling in the high-quartz field (i.e., at low pressure). The appearance of a dominant LiOH absorption band at 3483 cm⁻¹ seems to be related to quartz that crystallized at rather low pressures. Samples from nature with these characteristics exist, such as large hydrothermal quartz crystals that grew in the roof region of late stage magmatic systems and show a high LiOH/AlOH band ratio, e.g., the rock crystal from Brazil reported in Stalder and Neuser (2013).

Partitioning of Al, Li and H between quartz and melt

Since concentrations of Li in the melt could not be measured directly, Li concentrations in the quenched melt were calculated by mass balance. In order to do this, in a first step melt fractions were calculated from K₂O and Al₂O₃ content in the melt (the only significant repository of these elements). Furthermore, it was assumed that only insignificant amounts of Li are dissolved in the excess fluid after the experiment, because the pH of the fluid phase was close to neutral. Consequently, also Li was considered to be hosted mainly in the melt, and its concentrations were calculated based on the melt fraction calculated before. Hydrogen concentrations in the melts were calculated by subtracting the electron microprobe’s analytical totals from 100 %. Concentrations were used to determine partition coefficients for Al, Li and H between quartz and melt, which follow the order D ^qz/melt_Li  > D ^qz/melt_Al  > D ^qz/melt_H (Table 5). At 10 kbar, three different correlations are observed between the Log of D ^qz/melt_Li , D ^qz/melt_Al , D ^qz/melt_H and the spodumene content in the starting material (Fig. 10), that is (1) an almost horizontal trend for Li, (2) a clear positive trend for Al and (3) a slightly negative trend for H. In particular, the trends for D ^qz/melt_Li and D ^qz/melt_Al converge at 20 wt% spodumene and a kink for D ^qz/melt_Al between 10 and 15 wt% of spodumene content in the starting material is observed. The convergence and the kink might confirm the hypothesis that the system is saturated in spodumene for these conditions. Pressure does also affect the partitioning of Li, Al and H between quartz and melt. Experiments performed on starting material containing 2 wt% spodumene at pressures between 5 and 20 kbar, show in fact a negative correlation between the Log of D ^qz/melt_Li , D ^qz/melt_Al , D ^qz/melt_H and pressure (Fig. 11). All the trends present the same inclination, suggesting that Li, Al and H incorporations are related.

Table 5 Chemistry of quartz and melt, partition coefficients D ^qz/melt_Li , D ^qz/melt_Al and D ^qz/melt_Li , K and Al content in starting material and degree of melting

Fig. 10

Relations between the Log D(quartz/melt) of Li, Al, H and spodumene concentration in the starting material for experiments performed at 10 kbar. D ^quartz/melt_Al shows a positive trend that kinks between 10 and 20 wt% spodumene and converges with D ^quartz/melt_Li at 20 wt% spodumene. These particular features might confirm the hypothesis of a change in the saturation regime

Fig. 11

Relations between the Log D(quartz/melt) of Li, Al, H and pressure for experiments performed with starting materials containing 2 wt% spodumene. All the trends show the same negative inclination

The effect of apatite on defects incorporation in quartz

In this study it is observed that the presence of apatite in the bulk system has only a weak effect on the intensities of OH bands. LA-ICP-MS measurements revealed constant concentrations of 10–17 ppm P in quartz for all run products, suggesting that even the undoped system was saturated by apatite, even if it was only observed in BSE images at higher concentration levels (Fig. 1c). A weak coupling of P⁵⁺ to B³⁺ and Al³⁺ is possible, as apatite slightly reduces the AlOH and BOH bands in when added to the tourmaline-doped system (Fig. 3). High concentrations of P in the starting material may further influence phase and melting relation and therefore indirectly affect the trace elements incorporation (including OH defects) in quartz.

Compared to quartz from natural samples, phosphorous in quartz is well within the reported range for P₂O₅ enriched granites and monazite-bearing pegmatites that contain 7–10 ppm P (Breiter and Müller 2009) and 9–23 ppm P (Larsen et al. 2004), respectively.