Introduction

Tourmaline is the most common and earliest boron mineral formed on Earth, recently gaining an interest from the geoscience community (Grew et al. 2016; Dutrow and Henry 2018; Henry and Dutrow 2018). The importance of tourmaline as a powerful geological tool for probing P–T–X conditions at all crustal levels in the Earth, resides in its very large P–T stability field, common occurrence, and negligible intra-volume element diffusion rates (e.g., Henry and Dutrow 1996; van Hinsberg and Schumacher 2007). Moreover, due to its complex crystal chemistry, which allows a large number of chemical substitutions (e.g., Henry and Dutrow 1996, 2018; Bosi 2018), tourmaline is extremely sensitive to its chemical environment and it is, therefore, valuable as both a petrogenetic and provenance indicator (e.g., Henry and Dutrow 1992, 1996; Dutrow and Henry 2011; van Hinsberg et al. 2011a; Bosi et al. 2018a, 2019a; Ertl et al. 2008, 2012, 2018; Ahmadi et al. 2019; Sipahi 2019).

The tourmaline general chemical formula can be written as XY3Z6T6O18(BO3)3V3W, where X = Na+, K+, Ca2+, □ (= vacancy); Y = Al3+, Fe3+, Cr3+, V3+, Mg2+, Fe2+, Mn2+, Zn2+, Li+, Ti4+, □; Z = Al3+, Fe3+, Cr3+, V3+, Mg2+, Fe2+; T = Si4+, Al3+, B3+; B = B3+; V = (OH)–, O2–; W = (OH)–, F–, O2–. The letters X, Y, Z, T and B represent groups of cations accommodated at the [9]X, [6]Y, [6]Z, [4]T and [3]B crystallographic sites (identified with italicized letters); the letters V and W represent groups of anions accommodated at the [3]O(3) and [3]O(1) crystallographic sites, respectively. The H atoms occupy the H(3) and H(1) sites, which are related to O(3) and O(1), respectively. It is worth mentioning that, unlike amphibole or spinel, the structural and the chemical formula of tourmaline coincide: each crystallographic site in the structural formula matches a (non-italicized) letter in the chemical formula (Bosi et al. 2019b).

Tourmaline-supergroup minerals are currently classified into three groups based on the X-site occupancy: vacant, alkali and calcic (Henry et al. 2011). The X-site occupancy is usually related to both paragenesis and crystallization conditions of the rock in which tourmaline crystallized, and these relations may be used to reconstruct the host-rock thermal history (Henry and Dutrow 1996; van Hinsberg et al. 2011a, b, 2017; Dutrow and Henry 2018; Bosi et al. 2018a, 2019a). A further level of classification into subgroups is based on charge arrangements at the Y and Z sites, where small cations such as Al and Mg and relatively larger cations such as Fe2+ can be accommodated.

Using a large set of structural data, Bosi and Lucchesi (2007) presented a structural-stability field for tourmaline as a function of < Y–O > and < Z–O > mean bond lengths. The endpoints of this field indicate that the tourmaline structure can tolerate only a limited mismatch in dimensions between < Y–O > and < Z-O > (i.e., their difference, Δ(Y–Z) = < Y–O > – < Z–O > , should be smaller than 0.15 Å). The empirical validation of the occurrence of a long-range structural constraint and its effect on the occupancy of Y and Z sites are of particular interest. In fact, cation ordering in tourmaline structure may be exploited for geothermometric purposes (e.g., Henry and Dutrow 1992; Ertl et al. 2008; Filip et al. 2012; Bosi et al. 2016a, b; Ertl et al. 2018), while its variation can affect tourmaline physical properties, such as intrinsic dipole moment (Kim et al. 2018) and bulk moduli (Berryman et al. 2019). The possibility of Fe2+ to partially disorder across the Y and the Z sites, in particular, has been deeply investigated and largely discussed in several papers (see for example Andreozzi et al. 2008; Bosi and Andreozzi 2013, and references therein).

In the present study, a natural schorl with Fe2+ concentrations very close to that of schorlitic end-members was studied with an inclusive, multi-analytical approach with the aim to confirm that the crystal-chemical behavior of Fe2+ in tourmaline is controlled by structural stability requirements.

Material and experimental methods

The sample used for this study is a black tourmaline from a quartz-tourmaline orbicule found in the leucogranites of the Seagull batholith in the Yukon Territory, Canada (Fig. 1). According to Sinclair and Richardson (1992), the orbicules are considered to have developed during the final crystallization stages in which a hydrous melt/fluid (enriched in B, F, Cl and Fe) separated from a peraluminous granite, and tourmalines from the Seagull batholith are schorl with up to ~ 20 wt.% FeO and normalized Fe2+ cations up to 2.78 atoms per formula unit (apfu). To the best of our knowledge, these Fe2+ concentrations are the highest found in natural schorl so far.

Fig. 1
figure 1

Quartz-tourmaline pegmatitic orbicules included in leucogranites from Seagull batholith (Yukon Territory, Canada). Black tourmaline crystals up to 2-cm large are visible in the right inset

Full size image

In the present study, the Fe oxidation state was obtained by 57Fe Mössbauer Spectroscopy (MS); the chemical composition was obtained by using Electron MicroProbe (EMP) for major and minor elements and Laser-Ablation Inductively-Coupled-Plasma Mass-Spectrometry (LA-ICP-MS) for trace elements (including Li) and B; crystal-structure details were obtained by Structure REFinement (SREF) of X-ray single-crystal diffraction data.

MS

A fragment of the black Seagull tourmaline was ground under ethanol with an agate mortar, and 15 mg of the recovered fine powder was packed to make a disk of 10 mm of diameter. The absorber was then loaded in a Plexiglas® sample holder, and measurements were performed using a conventional spectrometer with a 57Co source of 0.99 GBq (25 mCi) embedded in a Rh matrix available at the Earth Sciences Department, Sapienza University of Rome (Italy). The spectrum was collected in transmission mode, at room temperature and at velocities between –4 to + mm/s, and recorded in a multichannel analyzer with 512 channels. The velocity was calibrated with a 25-µm thick α-iron foil. The obtained absorption spectrum was fitted to Lorentzian line-shapes using the RECOIL 1.04 fitting program (Lagarec and Rancourt 1998), with results reported in Table 1.

Table 1 Room temperature 57Fe Mössbauer parameters for Seagull schorl from Yukon Territory (Canada)
Full size table

EMP

Electron microprobe analysis for the Seagull tourmaline sample was obtained using a wavelength-dispersive spectrometer (WDS mode) with a Cameca SX50 instrument at the “Istituto di Geologia Ambientale e Geoingegneria (Rome, Italy), CNR”, operating at an accelerating potential of 15 kV, with a 15 nA current and a 10 μm beam diameter. Minerals and synthetic compounds were used as standards: wollastonite (Si, Ca), magnetite (Fe), rutile (Ti), corundum (Al), vanadinite (V), fluorphlogopite (F), periclase (Mg), jadeite (Na), orthoclase (K), sphalerite (Zn), rhodonite (Mn), metallic Cr, Ni and Cu. The PAP correction procedure for quantitative electron probe microanalysis was applied (Pouchou and Pichoir 1991). Results are reported in Table 2 and represent the mean values of 10 spot analyses across the crystal used for SREF study. Vanadium, Cr, Ni and Cu were below their respective detection limits (0.03 wt%) in the studied sample.

Table 2 Chemical composition of Seagull schorl from Yukon Territory (Canada)
Full size table

LA-ICP-MS

In-situ trace-element and boron analyses were collected using Laser-Ablation Inductively-Coupled-Plasma mass-spectrometry at Laurentian University (Canada). Samples were ablated using Resonetic Resolution M-50 coupled to a Thermo Electron XSeries II quadrupole ICP-MS. A 193 nm argon fluoride excimer laser was operated at a rate of 8 Hz. Line scans were completed using a 40 μm beam width and a scan speed of 20 μm/s with a measured fluence of ~ 3 j/cm2. Line scans were utilized to test for chemical zonation, which in turn was not present. Dwell times for rare-earth elements were increased to 15 ms from 10 ms (for other trace elements) to improve detection limits. External reference materials include: (1) NIST 610 (Jochum et al. 2011), (2) NIST 612 (Jochum et al., 2011) and (3) BHVO2G (Raczek et al. 2001). Standards were ablated every 10–15 analyses of the unknowns. Drift correction was applied using the baseline reduction scheme in IOLITE (Paton et al. 2011). Reference standards were utilized to assess the accuracy of the analyses. Results are reported in Table 3, together with elemental detection limits, and represent the mean values of 9 analyses across the sample.

Table 3 Boron and trace-elements contents of Seagull schorl from Yukon Territory (Canada)
Full size table

SREF

A representative crystal fragment of the Seagull tourmaline was selected for X-ray diffraction measurements on a Bruker KAPPA APEX-II single-crystal diffractometer (Sapienza University of Rome, Earth Sciences Department), equipped with a CCD area detector (6.2 × 6.2 cm active detection area, 512 × 512 pixels) and a graphite-crystal monochromator, using MoKα radiation from a fine-focus sealed X-ray tube. The sample-to-detector distance was 4 cm. A total of 3681 exposures (step = 0.2°, time/step = 20 s) covering a full reciprocal sphere with a redundancy of about 12 was collected. Final unit-cell parameters were refined using the Bruker AXS SAINT program on 9981 reflections with I > 10 σ(I) in the range 5° < 2θ < 80°. The associated intensities were processed and corrected for Lorentz and background effects plus polarization, using the APEX2 software program of Bruker AXS. The data were corrected for absorption using a multi-scan method (SADABS). The absorption correction led to a significant improvement in Rint. No violation of R3m symmetry was detected.

Structure refinement was done using the SHELXL-2014 program (Sheldrick 2015). Starting coordinates were taken from Bosi et al. (2015). Variable parameters were: scale factor, extinction coefficient, atom coordinates, site-scattering values (for X, Y and Z) and atomic-displacement factors. Regarding the atomic model refinement, the X site was modeled using the Na scattering factor. The occupancies of the Y and Z sites were obtained considering the presence of Al versus Fe. The T, B and anion sites were modeled, respectively, with Si, B and O scattering factors and with a fixed occupancy of 1, because refinement with unconstrained occupancies showed no significant deviations from this value. A final refinement was then performed by modelling the site occupancy of the O(1) site with O and F fixed to the value obtained from the empirical formula (see below). Similar chemical constraints were applied to refine the H(1) and H(3) sites. There were no correlations greater than 0.7 between the parameters at the end of the refinement. Table 4 lists crystal data, data-collection information and refinement details; Table 5 displays the fractional atom coordinates, site occupancy factors and equivalent isotropic-displacement parameters; Table 6 shows selected bond lengths. A CIF file is included in supplemental material.

Table 4 Single-crystal X-ray diffraction data details for Seagull schorl from Yukon Territory (Canada)
Full size table
Table 5 Fractional atom coordinates (x,y,z), site occupancy factors (s.o.f.) and equivalent-isotropic (Ueq) and isotropic (Uiso) displacement parameters (in Å2) for Seagull schorl from Yukon Territory (Canada)
Full size table
Table 6 Selected bond distances (Å) for Seagull schorl from Yukon Territory (Canada)
Full size table

Results and discussion

Chemical composition and iron speciation

The black tourmaline sample from Seagull batholith (Yukon Territory, Canada) is chemically homogeneous and was established to be a schorl (see below) with concentrations of Fe2+ among the highest currently found in nature (FeOtot ~ 18 wt.%, Table 2). The Mössbauer absorption spectrum of the Seagull tourmaline shows two dominant absorptions between –1 and +3 mm/s (Fig. 2). The spectrum was deconvoluted using Lorentzian doublets in agreement with models already used in the existing literature (Andreozzi et al. 2008; Bosi et al. 2019b). Accordingly, a model with five doublets was initially adopted and their hyperfine parameters were tentatively refined. Four doublets gave center shift (δ) values around 1.10(2) mm/s and were assigned to [6]Fe2+. They are distinguished by their quadrupole splitting (ΔEQ) values: the first three doublets, conventionally labelled Y1, Y2 and Y3, have ΔEQ = 2.43(9), 2.11(6), 1.69(11) were interpreted as Fe2+ at the Y sites with different nearest and next-nearest neighbor coordination environments. The fourth Fe2+ doublet, which has ΔEQ = 1.20(5) mm/s was interpreted as [6]Fe2+ in a different chemical environment (or crystallographic-site symmetry) from the Y sites, that is Fe2+ at the Z site. This interpretation is in agreement with the model described in Andreozzi et al. (2008) and Bosi et al. (2015). A potential fifth doublet, centered around 0.3 mm/s and representing [6]Fe3+, was quantified to be ≤ 2% of the Fetot, but its area was smaller than the experimental uncertainty (± 3%) and its hyperfine parameters could not be refined. The contribution of [6]Fe3+ was therefore considered to be negligible and all the Fe was eventually considered as [6]Fe2+ (Table 1). Notably, the same conclusions were also obtained by Sinclair and Richardson (1992). Concerning site distribution of Fe2+, the absorption doublet assigned to ZFe2+ was quantified at 25(5)% of the Fetot, which corresponds to 0.65(13) apfu and converges to ZFe2+ obtained by SREF data (see below).

Fig. 2
figure 2

Room temperature 57Fe Mössbauer spectrum for Seagull schorl from Yukon Territory (Canada). Experimental spectrum (exp) is represented by open circles, calculated spectrum (calc) by thick red curve (reduced χ2 = 0.68). Lorentzian absorption doublets assigned to [6]Fe2+ are represented by thin green lines

Full size image

Site populations

The Li content was considered negligible for crystal-chemical purposes because it is below the detection limit of LA-ICP-MS analysis. The B content was assumed to be stoichiometric (B = 3.00 apfu corresponding to B2O3 = 10.08 wt.%) as corresponding to the LA-ICP-MS results, within the experimental uncertainty (Table 3). A stoichiometric B content is further supported by the SREF results (Tables 5, 6): the values of the B- and T-site occupancy factors and the < B–O > and < T–O > distances are effectively consistent with the B and T sites fully occupied by B and Si, respectively (e.g., Bosi and Lucchesi 2007). Based on MS results, the oxidation state of all Fe atoms was considered exclusively +2. The (OH) contents and the atomic fractions were then calculated by charge balance under the assumption of (T + Y + Z) = 15.00 apfu and 31 anions. The excellent agreement between the number of electrons per formula unit (epfu) derived from chemistry and from SREF (258.3 epfu versus 258.6 epfu, respectively) supports the above stoichiometric assumptions.

With regard to the site populations at X, B, T, O(3) (≡ V) and O(1) (≡ W), the standard site preference suggested for tourmaline (e.g., Henry et al. 2011) were actually satisfied, while the Y and Z site populations were optimized according to the procedure of Bosi et al. (2017) and by fixing the minor elements Ti4+, Mn2+ and Zn at Y. The robustness of this optimized cation distribution was successively confirmed by the procedure of Wright et al. (2000) in which the default setting was assumed, but the chemical variability was constrained by electroneutrality. The resulting empirical formula for the Seagull tourmaline is:

X(Na0.74□0.24K0.01Ca0.01)Σ1.00Y(Fe2+2.05Al0.92Ti0.02Mn0.01Zn0.01)Σ3.00Z(Al5.41Fe2+0.53Mg0.06)Σ6.00(Si6O18)(BO3)3 V(OH)3 W[(OH)0.46F0.41O0.13]Σ1.00.

The refined site-scattering values and those calculated from the optimized site-populations are in excellent agreement (Table 7). It is worth noting the occurrence at Z site of significant amounts of Fe2+ (0.53 apfu), which is a cation heavier than Al and it is thus required to justify the refined Z-site scattering value (Z-m.a.n, where m.a.n. is the mean atomic number) of 14.14(3), considerably larger than the expected value for a Z site fully occupied by Al (Z-m.a.n = 13). Such a cation distribution is also corroborated by the refined < Z–O > value (1.924 Å), which is larger than the typical < Z–O > values observed for a Z site fully occupied by Al (1.902–1.913 Å; Bosi and Andreozzi 2013) and reflects the presence of cations larger than Al at the Z site (as, for example, Fe2+). Finally, as previously highlighted, the amounts of Fe2+ at Z retrieved by SREF (0.53 apfu) are consistent with the absorption area of the MS doublet assigned to ZFe2+ (Table 1), which corresponds to ~ 0.65(13) apfu.

Table 7 Site-scattering values and optimized site-populations for Seagull schorl from Yukon Territory (Canada)
Full size table

Influence of the tourmaline structural stability on the Al-Fe2+ order–disorder

The empirical formula of the Seagull tourmaline is consistent with a tourmaline belonging to the alkali group (Henry et al. 2011): Na-dominant at the X position of the tourmaline general chemical formula and (OH)-dominant at W (OH > F). The Y and Z positions are dominated by Fe2+ and Al, respectively. In accord with the IMA-CNMNC rules (Bosi et al. 2019d), the end-member formula is X(Na)Y(Fe2+)3Z(Al)6(Si6O18)(BO3)3(OH)3(OH), corresponding to schorl (Henry et al. 2011). By our most comprehensive understandings, the Seagull schorl is the Fe2+-richest schorl investigated by SREF thus far, even richer than the fluor-schorl from Steinberg, Germany, for which FeOtot = 16.5 wt.% and Fe2+ ≤ 96% of Fetot (Ertl et al. 2016).

The composition of the studied schorl can be approximated as NaY(Fe2+2Al)Z(Al5Fe2+)(Si6O18)(BO3)3(OH)3(OH,F). With respect to the ideal ordered schorl (Fe2+ ordered at Y and Al at Z, see above), the observed partial disorder of Fe2+ over the Y and the Z sites (68% and 9% atoms/site, respectively) can be explained by the intracrystalline order–disorder reaction YAl + ZFe2+ → YFe2+ + ZAl, which is needed to satisfy the long-range structural stability of tourmaline (Bosi 2018). In fact, in the ideal ordered schorl the expected difference between the < YFe2+–O > and < ZAl–O > would be Δ(Y–Z) = 0.232 Å [< YFe2+–O > = 2.136(1) Å and < ZAl–O > = 1.904(3) Å, calculated from the ionic radii of Bosi (2018)], that is too large and inconsistent with the structural stability limits of the tourmaline supergroup minerals: Δ(Y–Z) less than 0.15 Å [as proposed by Bosi and Lucchesi (2007) and successively confirmed by Bosi (2018) using 322 SREF data sets]. Consequently, the structure of the ordered schorl would be unstable (Fig. 3). In the studied schorl, the misfit between < Y–O > and < Z–O > is reduced to 0.127 Å by the reaction YAl + ZFe2+ → YFe2+ + ZAl, which mitigates the potential misfit by shortening < Y–O > up to 2.051 Å (introducing YAl) and increasing < Z–O > up to 1.924 Å (introducing ZFe2+). Therefore, the data are consistent with the partial disordering of significant amounts of Fe2+ into the Z sites as a mechanism to establish long-range structural stability (Fig. 3).

Fig. 3
figure 3

Modified from Bosi (2018)

Relation between < Z–O > and < Y–O > showing the structural-stability limits for tourmaline minerals. Solid diagonal lines: left = ratio 1:1 between < Z–O > and < Y–O > ; right = ratio 1:1 shifted by 0.15 Å. The full red circle represents the partially disordered Seagull schorl from Yukon Territory (Canada); the red cross is the ideal ordered schorl; black dots represents 322 data sets with SREF.

Full size image

The site redistribution of Fe–Mg–Al over the Y and Z sites was experimentally investigated at high temperature for schorl, dravite, Fe-bearing oxy-dravite, lucchesiite and Fe-rich fluor-elbaite by Filip et al. (2012), Bosi et al. (2016a,b, 2018b, 2019c). These studies, conducted at both oxidizing and reducing conditions, showed significant Fe–Mg–Al intersite exchanges when Fe-bearing tourmalines were heated up to 700–800 °C, with Fe always involved in the disordering process (up to 0.37 apfu of ZFe for schorl and lucchesiite). Although the possibility of a geothermometric exploitation of tourmaline intersite exchanges is encouraging, the present results (with ZFe2+ = 0.53 apfu) suggest adopting some precautions, at least for very high Fe contents, due to the important role of the long-range structural constraints on Fe2+ site distribution.

Conclusions

The comprehensive multi-analytical approach of this study of a schorl sample from Seagull batholith (Yukon Territory, Canada) constrains the crystal-chemical behavior of Fe2+ in tourmaline. The studied sample is the Fe2+-richest schorl with SREF data known in the literature.

From a crystal-chemical viewpoint, a partial disorder of Fe2+ over the Y and Z sites was observed, and this is required for the long-range structural constraints to be fulfilled. Similar arguments were applied to explain the observed partial disorder of Mg over Y and Z for dravite and fluor-dravite, as well as the failure in synthesizing the tsilaisite end-member composition (Bosi and Lucchesi 2007; Clark et al. 2011; Bosi et al. 2012; Bosi 2018). For tourmaline to be stable, the 3-D framework of ZO6 octahedra must be able to accommodate the structural islands made of (XO9 + YO6 + BO3 + TO4) polyhedra. With Z site fully occupied by Al, very high amounts (>> 2 apfu) of Mg2+, Fe2+ and Mn2+ ordered at the Y site would make tourmaline unstable, because the resulting difference between < Y–O > and < Z–O > would be too large [i.e., Δ(Y–Z) > 0.15 Å]. A partial disorder of Al and R2+-cations over Y and Z is, therefore, necessary to reduce the misfit between YO6 and ZO6 and to guarantee tourmaline structure stability.