Chemical microanalysis

The phlogopites were analysed on MEB-FEG Zeiss Ultra 55 and Cameca SX-five microprobe with EDS Brüker, using beam conditions of 15 kV and 10 nA. Standards were albite (Na), orthoclase (K, Al), diopside (Si, Ca, Mg), hematite (Fe), pyrophanite (Ti, Mn), zircon (Zr), baryte (Ba), apatite (P), scapolite (Cl) and fluorite (F). The results in Table 1 (talc-absent zone) and Table 2 (talc-bearing zone) have been selected among 87 and 367 microanalyses, respectively. The standard deviations associated with the electron microprobe (EMPA) results are indicated in the tables, as well as the detection limits for elements at low concentration such as Mn and Ca, in both cases as weight percent of oxide. They allowed for the calculation of uncertainties on the structural formulae, mentioned in the text below, and especially the interlayer vacancy (Tables 1, 2).

Table 2.

Selected EMPA results of Ba-rich phlogopites from La Creuse, talc zone.

σ = Standard deviation in italics below each constituent; na = not analysed; bd = below detection.

(1) Fe total calculated as FeO; (2) detection limit; (3) calculation after Rieder et al. (Reference Rieder, Cavazzini, D'yakonov, Frank-Kamenetskii, Gottardi, Guggenheim, Koval, Mueller, Neiva, Radoslovich, Robert, Sassi, Takeda, Weiss and Wones1998).

The Fe³⁺/ΣFe ratio was quantified with the Cameca SX-five electron microprobe, using a peak-shift method protocol described in Fialin et al. (Reference Fialin, Wagner, Metrich, Humler, Galoisy and Bezos2001, Reference Fialin, Bezos, Wagner, Magnien and Humler2004, Reference Fialin, Wagner and Pascal2011). The Fe peak positions were determined on Ba-rich micas (sample B5b) and Ba-poor micas (sample LC2), using seven raw single crystals of mica from the Collection de minéraux de Sorbonne Université as a reference, with ten replicate measurements per spot. Fe total and Fe²⁺ in the reference single crystals were analysed by potentiometric titration.

In the Ba-rich, Al-poor phlogopites, the ratio Fe³⁺/ΣFe shows a good inverse correlation with Al₂O₃ (wt.%) (Fig. 3) expressed by the relation:

$${\rm Fe}^{3 +} /{\rm \Sigma Fe} = - 0.263{\rm Al}_2{\rm O}_3 + 3.448 \,\,\lpar {R^2 = 0.9652} \rpar $$

This correlation only corresponds to the phlogopites with <13 wt.% Al₂O₃. No similar correlation was observed in the Al-rich, Ba-poor phlogopites, which are virtually devoid of Fe³⁺.

Fig. 3.

Correlation of the ratio Fe³⁺/ΣFe with the Al content of the barian phlogopites from La Creuse.

X-ray diffraction analysis

X-ray diffraction (XRD) analysis was performed at the Plateforme de diffraction des rayons X, IMPMC-Sorbonne-Université. A single crystal was carefully extracted from a thin section and then mounted on a glass fibre for the XRD experiment. Diffracted intensities were collected at 293 K on an Oxford Diffraction Xcalibur S 4-circle diffractometer equipped with a Sapphire 3 CCD detector and using MoKα radiation (λ = 0.71073 Å, graphite monochromator).

CrysalisPro software (Agilent Technologies Ltd, Yarnton, England), was used for data acquisition and processing: data reduction, cell refinement, space-group determination, scaling, analytical absorption correction (Clark and Reid, Reference Clark and Reid1995). A preliminary experiment rapidly allowed us to determine the monoclinic unit cell and a strategy combining different omega scans was then automatically calculated to perform a complete measurement with an optimized redundancy. The data analysis lead to a monoclinic C2/m cell (R _int = 5.31%) with refined parameters a = 5.3185(5) Å, b = 9.2136(9) Å, c = 10.1349(11) Å and β = 100.131(11)°.

After correction for Lorentz-polarization effects and absorption, the structure was then solved through the Olex2 program (Dolomanov et al., Reference Dolomanov, Bourhis, Gildea, Howard and Puschmann2009) by direct methods using SHELXT (Sheldrick, Reference Sheldrick2015) and gave rise to starting atomic coordinates consistent with previously described analogous structures (Bailey, Reference Bailey and Bailey1984).

The anisotropic structural refinement was then carried out with SHELXL2013 (Sheldrick, Reference Sheldrick2008) by full-matrix least squares minimization and difference-Fourier methods. Neutral atomic scattering factors were used. The crystallographic information files have been deposited with the Principal Editor of Mineralogical Magazine and are available as Supplementary material (see below).

Talc-absent zone

Phlogopites in the talc-absent zone are Al-rich, with variations possibly due to their occurrence in Al-rich, cordierite-bearing metasilitites, or in metarhyolites. Such variations are expressed by octahedral vacancies ^VI□ ranging from 0.1 to 0.4 pfu (for 44 valencies), due to the muscovitic (^VI□^VIAl₂)(Mg₃)_–1 substitution in which ^VIAl is inversely correlated (–2:3) to (Mg + Fe²⁺) (Fig. 4). Another feature of these phlogopites is the presence of interlayer vacancies ranging from 0.10 to 0.46 ± 0.08 (Fig. 1a, Table 1), as already reported from metapelites from a contact metamophic aureole (Guidotti and Dyar, Reference Guidotti and Dyar1991; see below).

Fig. 4.

Compositional variations in the octahedral site (Al, Mg + Fe²⁺) of the phlogopites from La Creuse correspond to variable muscovitic substitution (^VI□^VIAl₂)(Mg₃)_–1 in the talc-free rocks. Ba-rich phlogopites from the talc–baryte rocks have a much smaller muscovite content.

The gap in phlogopite compositions associated with the talc ‘front’

At the talc front, phlogopite is present on both sides but with strong differences both in composition and textural relations. Phlogopites from the talc zone differ without transition from their counterparts from the downstream talc-free zone, not only by their high Ba contents but also by higher Mg, lower Fe and Al, and partial replacement of K by Na (Tables 1, 2, Figs 4, 5). Each zone corresponds in various plots (Figs 4 to 7) to a dense, almost continuous, cluster of points, both clusters being separated by a gap.

Fig. 5.

Part of a triangular plot of K, Na and Ba proportions in the phlogopites from La Creuse. In the Ba-rich phlogopites K is partly replaced not only by Ba but also by Na.

The increased amount of Mg in phlogopite occurs mainly as a replacement of Fe²⁺ but also to a large extent of octahedral Al. Two classical substitutions involving replacement of ^VIAl by Mg are known as Tschermak (MgSi)(^VIAl^IVAl)_–1 substitution, with a positive 1:1 correlation between Si and (Mg + Fe²⁺) as a diagnostic feature, and muscovitic substitution (Mg₃)(^VI□^VIAl₂)_–1. The small increase in Si (<0.2 apfu, Fig. 1b, hardly higher than the standard deviation 0.18) that accompanies the large increase in (Mg + Fe²⁺) (>0.6, much larger than the standard deviation 0.14) indicates that the loss of Al is only marginally due the Tschermak substitution, and the change from one zone to the other mainly corresponds to the effect of the muscovitic substitution (Fig. 4).

Ba-phlogopites in the talc zone

Within the talc zone, phlogopite has a limited compositional range characterized by strikingly constant Mg + Fe²⁺ (Fig. 6a) whereas the Ba contents range from 0.22 ± 0.04 to 0.82 ± 0.08 apfu. Three comparatively homogeneous samples or groups of samples, A, B and C respectively, can be defined according to the Ba contents of the phlogopites. The increase in Ba from group C to group A is accompanied by an increase in octahedral vacancies and a correlated decrease in ^VIAl (Figs 7b,c), without change in either Mg + Fe²⁺ (Fig. 6a) or interlayer vacancy (Fig. 6c). The change in Ti with increasing Ba, hardly larger than the 2σ of the linear regression (Fig. 6b), may be ignored in the first approximation. With regard to Si (Fig. 7a), apart from a correlation with Ba within group A (a single sample) that can be accounted for by the Wendlandt substitution (Ba^IVAl)(KSi)_–1 (see text below), no change is observed from one group to the others. Therefore, the whole range of compositions occurring in the talc zone corresponds approximately to a substitution involving only Ba, K, Al and octahedral vacancies, i.e. (Ba^VI□_⅓)(K^VIAl_⅓)_–1. The corresponding vector indicated on Figs 7a,b,c suggests that the set of analyses is adequately described by this substitution.

Fig. 6.

Elements virtually independent of the Ba content in the Ba-phlogopites from La Creuse: Mg + Fe²⁺(a); Ti (b); and interlayer vacancies (c). The linear regressions with standard deviations are indicated on the diagrams.

Fig. 7.

Variations correlated to Ba in the Ba-phlogopites from La Creuse. (a) Si remains in the same range in different sample groups, but in individual samples there is a –1:1 correlation consistent with the Wendlandt substitution. In contrast, octahedral Al (b) and vacancies (c) show trends consistent with the substitution (Ba^VI□_⅓)(K^VIAl_⅓)_–1 .

Before getting further precision, note that both ends, along this substitution, of the whole compositional range of the talc-zone phlogopites correspond to essential crystal chemical constraints, i.e. positive or null values for octahedral Al (Fig. 7b) and octahedral vacancies (Fig. 7c), respectively. These observations indicate that phlogopite in the talc zone requires sufficient Ba to be stable, the lower Ba contents corresponding to an entirely trioctahedral character of the phlogopite. At the other end of the trend, further incorporation of Ba is blocked by exhaustion of octahedral Al. This is probably the reason why, in the Ba-rich sample referred to as group A, a few compositions with higher Ba contents depart from the average through the Wendlandt substitution (Ba^IVAl)(KSi)_–1 (Fig. 7), however, with a small extent due to the lack of Al in the host rock. This lack of Al is expressed by the good negative correlation observed between Fe³⁺ and Al, with significant amounts of Fe³⁺ restricted to the few analyses lowest in Al and highest in Ba (Fig. 3, Table 2).

The Ba substitution

The element-element plots of Figs 4 to 7 are clear enough to illustrate the general tendencies outlined above, but they do not provide full evidence for complex substitutions involving at least five independent end-members in micas. In order to obtain undisputable information on the substitution(s) involved in the Ba-rich phlogopite observed in the talc zone, we used a simple method of recalculating the phlogopite compositions on the basis of a set of independent exchange vectors applied to the additive component phlogopite K₂Mg₆Al₂Si₆O₂₀(OH)₄. In the present case, we first reduce the number of parameters by defining

$$\eqalign {&{\rm K}{^\ast} = {\rm K} + {\rm Na},\; \cr & {\rm Mg}{^\ast} = {\rm Mg} + {\rm Fe}^{2 +} + \hbox{ Mn and }\cr & {\rm Al}{^\ast} = {\rm Al} + {\rm Fe}^{3 +}.} $$

Our choice of exchange vectors is

For each analysis, the set of values (x, y, z, t, u) has been determined from the composition (Si, Al*, Mg*, ^VI□, Ti, K*, ^XII□) by a linear regression. As the interlayer vacancy (z) remains quasi constant in the talc zone, compositional variations are entirely described by x, y and t (Fig. 8), with a possible small contribution of u. Linear regressions of x(t), y(t) and u(t) (Figs 8a,b, 6b) yield the main exchange vector controlling phlogopite compositions from the talc zone: –0.98 x + 0.29 y + 0.03 u + t, i.e. (Ba^VI□_0.29Mg*_0.08Ti_0.03^IVAl_0.08)(K*^VIAl*_0.40Si_0.08)_–1. This exchange vector confirms, within uncertainties taken as 2σ of the linear regressions, the simplified vector anticipated above (Ba^VI□_0.33)(K^VIAl_0.33)_–1. The extent of replacement of K by Ba is constrained in the range ~0.24 to 0.80 apfu by the sum octahedral Al + octahedral vacancies, itself constrained by the constant (buffered) contents in other octahedral components, Mg + Fe (~5.62 ± 0.14) and Ti (~0.14 ± 0.014).

Fig. 8.

Linear regressions observed between the Wendlandt substitution t and: (a) the Tschermak substitution x; (b) the muscovitic substitution y, in the Ba-phlogopites from La Creuse. The standard deviations are indicated in brackets.

It must be emphasized that this substitution is observed within the talc zone, but cannot be extended to join the Al-rich phlogopite of the talc-absent zone. The Ba-richest compositions correspond approximately, apart for the small Ti content, to Ba_0.76□_0.56K*_0.68Mg*₆Al_2.20Si_5.80O₂₀(OH)₄ i.e. kinoshitalite_0.38phlogopite_0.34talc_0.28, the interlayer vacancy being accounted for by the ‘talc’ substitution. The whole compositional range can be described chemically as solid solutions between talc and Ba-rich micas ranging from phlogopite_0.79kinoshitalite_0.21 to phlogopite_0.47kinoshitalite_0.53.

If the phlogopite–kinoshitalite solid solution is well documented (e.g. Bailey, Reference Bailey and Bailey1984; Brigatti and Poppi, Reference Brigatti and Poppi1993), the existence of a solid solution between talc and phlogopite is debated. There is evidence that in some cases the total number of atoms (K, Na and Ba) in the interlayer of trioctahedral micas is less than 2, down to 1.4 and even 1.3, as shown for instance by biotites from pelitic rocks belonging to a contact-metamorphic aureole in Maine, USA (Guidotti and Dyar, Reference Guidotti and Dyar1991; Brigatti et al., Reference Brigatti, Guidotti, Malferrari and Sassi2008). In that occurrence the number of vacant interlayer sites in the biotite seems to depend on the metamorphic grade, ranging on average from 0.34 ± 0.04 in the lower grades to 0.22 ± 0.06 in the sillimanite and anatectic zones. Si-rich phlogopites with up to 0.30 pfu vacancy in the interlayer region have also been described from peridotites. They correspond to the incorporation of significant amounts of talc component, tentatively interpreted as stabilized at high H₂O pressure by the translation of talc (+H₂O) to a 10 Å phase (Comodi et al., Reference Comodi, Nazzareni, Fumagallo and Capitani2011). The only described case of a large range of solid solution with talc occurs in the interlayer-deficient Na-mica wonesite (Spear et al., Reference Spear, Hazen and Rumble III1981).

If the Ba-rich phlogopite of La Creuse is a single phase, it is an interesting case. Other a priori possibilities are an interstratified compound with talc layers (such compounds however have never been described) or a non-homogeneous crystal with talc-rich nano-domains.

Refinement 1

As the Al/Si content at the tetrahedral site does not significantly influence the refinement, it was fixed to the usually accepted 1:3 ratio for phlogopite. The interlayer site was fixed with 100% K and the two octahedral sites were considered fully occupied by a 1:1 Mg/Fe ratio. Then, only the Mg/Fe ratios were refined and gave rise to a 92:8 Mg/Fe ratio for each site (R1 factor = 5.02%). These values were consistent with those obtained from microprobe analysis.

The refinement of K occupancy led to 109%, which partially confirmed the presence of a heavier atom within the interlayer space. We then introduced Ba and the K/Ba fraction was refined: a 96:3 K/Ba ratio was obtained (R ₁ = 4.86%). As the expected proportions of Ba from EMPA are generally higher, this result is in favour of a vacancy in this site. Thus, we systematically refined the occupancies of Mg/Fe and K/Ba exploring different vacancies. The solutions giving the best R factor (4.77%), goodness-of-fit (1.06) and residuals (highest peak = 1) values were obtained with 15%<vacancy < 40% at the interlayer site. Such values for R, goodness-of-fit and residuals are fairly acceptable for this type of structure. Details for the data collection and structure determination parameters for an arbitrary retained 35% vacancy are given in Table 3, supplemented with Tables 4 and 5.

Table 3.

Crystal data and structure refinement of Mica4_18092014*.

*For an arbitrary retained 35% vacancy in interlayer site.

Table 4.

Fractional atomic coordinates and equivalent isotropic displacement parameters (Ų) of Mica4_18092014*.

* for an arbitrary retained 35% vacancy in interlayer site

** U _eq is defined as ⅓ of the trace of the orthogonalized U ^ij tensor.

Table 5.

Anisotropic displacement parameters (Ų)* of Mica4_18092014**.

* The anisotropic displacement factor exponent takes the form: –2π²[h ²a*²U ¹¹+2hka*b*U ¹²+…].

** for an arbitrary retained 35% vacancy in interlayer site.

Refinements 2 (based on chemical analysis)

We then tried out a series of compositions determined from electron microprobe analysis (Table 6) for the crystal-structure refinement. We mainly observed that the models from the groups B and C (samples T14 to Ta) significantly stand out according to crystallographic agreement factors. The best fits were obtained with MC and LC4 compositions. These compositions were consistent with those obtained from our first refinement. The refinements assuming composition LC4 for the studied crystal, with empirical formula Ba_0.36□_0.52K_0.90Na_0.22Mg_5.6Fe_0.48Ti_0.10Al_2.42Si_5.80O₂₀(OH)₄, are given in Tables 7, 8 and 9. Bond lengths are given in Table 10 and chemical crystal parameters, in Table 11, are based on the model of Mercier et al. (Reference Mercier, Rancourt, Redhammer, Lalonde, Robert, Berman and Kodama2006).

Table 6.

Mica compositions* tested in structural refinements.

* from EMPA. n = number of analyses; nm = not measured.

Table 7.

Crystal data and structure refinement* for LC4 composition.

* of Mica4_18092014.

Table 8.

Fractional atomic coordinates and equivalent isotropic displacement parameters (Ų)* for LC4 composition.

*U _eq is defined as ⅓ of the trace of the orthogonalized U ^ij tensor.

Table 9.

Anisotropic displacement parameters (Ų)* for LC4 composition.

*The anisotropic displacement factor exponent takes the form: –2π²[h ²a*²U ¹¹+2hka*b*U ¹²+…]

Table 10.

Bond lengths (Å) for LC4 composition.

Symmetry codes: (i) –½+x,½+y,+z; (ii) ½+x,½+y,+z; (iii) 3/2–x,½–y,1–z; (iv) ½+x,½–y,+z; (v) ½–x,–½+y,–z; (vi) 1–x,+y,–z; (vii) ½+x,–½+y,+z; (viii) ½–x,½–y,–z; (ix) 1–x,–y,–z; (x) +x,–y,+z; (xi) 3/2–x,½–y,–z; (xii) –½+x,–½+y,+z; (xiii) 1–x,–y,1–z; (xiv) –1 + x,+y,+z; (xv) 1–x,+y,1–z; (xvi) –½+x,½–y,+z; (xvii) 1 + x,+y,+z; (xviii) 3/2–x,–½+y,1–z; (ix) 2–x,–y,1–z; (xx) +x,1–y,+z

Table 11.

Chemical crystal parameters for LC4 composition.

α: Tetrahedral rotation angle; τ: tetrahedral flattening angle; TAV: tetrahedral angle variance (Robinson et al., Reference Robinson, Gibbs and Ribbe1971); TQE: tetrahedral quadratic elongation (Robinson et al., Reference Robinson, Gibbs and Ribbe1971); Ψ: octahedral flattening angle (Donnay et al., Reference Donnay, Donnay and Takeda1964; Toraya, Reference Toraya1981); OQE: octahedral quadratic elongation (Robinson et al., Reference Robinson, Gibbs and Ribbe1971); OAV: octahedral angle variance (Robinson et al., Reference Robinson, Gibbs and Ribbe1971).

Existence or not of a solid solution between phlogopite and talc

As mentioned above, a large range of solid solution with talc has been shown to occur in the interlayer-deficient Na-mica wonesite (Schreyer et al., Reference Schreyer, Abraham and Kulke1980; Spear et al., Reference Spear, Hazen and Rumble III1981), observed to contain exsolutions of talc lamellae (Veblen, Reference Veblen1983a,Reference Veblenb; Kogure et al., Reference Kogure, Miyawaki and Banno2005). Wonesite is structurally closer to talc than K-phlogopite as regards the layer offset (Kogure et al., Reference Kogure, Miyawaki and Banno2005) which is caused by repulsion between facing basal oxygen atoms across the interlayer region. The amount of layer offset is limited by the interlayer cations located in the cavity space of the oxygen hexagonal (ditrigonal) rings. In accordance with the small size of Na⁺ ions, the interlayer-deficient Na mica wonesite contains a large layer offset (1.25 Å), thus more resembling talc (1.55 Å, Ďurovič Reference Ďurovič and Wilson1992) than K-rich micas. In accordance, talc exsolutions observed in wonesite are believed to indicate an appreciable range of solid solution between talc and Na-rich phlogopite (Veblen, Reference Veblen1983a). In the case of large cations (K⁺ and Ba²⁺), even if the hexagonal rings are not reduced in size by tetrahedral rotation, no layer offset occurs (Bailey, Reference Bailey and Bailey1984).

We performed a series of reconstructions of precession images (Fig. 9) showing different hkl planes, in which can be noticed the absence of any systematic Bragg spots incompatible with the unit-cell parameters determined for the crystal LC4. These images reveal a good crystalline quality, although it remains impossible to detect any intercalated nano-domain using the XRD technique.

Fig. 9.

Series of (hkl) precession images of the Ba-phlogopite from La Creuse (Mica4_18092014).

The hypothesized presence of talc layers in amounts as high as 28 mol.% would result in a departure from monoclinic symmetry towards hexagonal symmetry and in appreciable changes in crystallographic parameters, in particular β and c. β in LC4 is 100.131°, larger than in both phlogopite (100.08°) and kinoshitalite (99.99°), contrary to the expected effect of talc (98.68° for triclinic talc). The c parameter of a solid-solution phlogopite–kinoshitalite 72:28 is calculated as 10.190 Å. Assuming that the measured c parameter of a hypothetical mixture of this Ba-mica and talc (c = 9.46 Å) can be inferred from their proportions, in order to obtain the value of c measured in LC4 (10.1349 Å) the proportion of talc should be as low as 6.5%, far below 28%.

The cell parameters of LC4 can be compared to phlogopites from literature (review in Brigatti and Guggenheim, Reference Brigatti, Guggenheim, Mottana, Sassi, Thompson and Guggenheim2002) selected according to: (1) the octahedral sheet, as similar as possible to that of La Creuse, Fe- and Ti-poor, in order to minimize differences in dimension of the octahedral sheet; and (2) the interlayer filling, either mainly K, or including Ba in phlogopite–kinoshitalite solid solution, or including vacancies. The tetrahedral rotation 7.1° is small, consistent with that in Si-rich potassic phlogopites, in agreement with the composition of the tetrahedral sheet close to Si:Al = 3 (Fig. 10a,b). In contrast, in kinoshitalites characterized by an increased substitution of Al to Si balancing the substitution of Ba to K, the tetrahedral rotation is increased to counterbalance the larger mean size of the tetrahedra (Bailey, Reference Bailey and Bailey1984; Brigatti and Poppi, Reference Brigatti and Poppi1993; Brigatti and Guggenheim, Reference Brigatti, Guggenheim, Mottana, Sassi, Thompson and Guggenheim2002). The interlayer separation in LC4 is smaller than in phlogopite, and consistent with that in kinoshitalites with similar Ba content (Fig. 10c), in agreement with the stronger attraction on tetrahedra exerted by Ba²⁺ than by K⁺ (Brigatti and Poppi, Reference Brigatti and Poppi1993). This apparent lack of influence of the assumed vacancies on the interlayer separation may seem surprising because with a partly vacant interlayer space, repulsion between facing basal oxygen atoms is expected to result in an enlarged interlayer separation. However, the cavity space for the interlayer cations is larger in LC4 than in a phlogopite–kinoshitalite solid solution having the same Ba content because the tetrahedral rotation is smaller in LC4, in accordance with the general inverse correlation between interlayer separation and tetrahedral rotation (Brigatti and Guggenheim, Reference Brigatti, Guggenheim, Mottana, Sassi, Thompson and Guggenheim2002).

Fig. 10.

Comparison of structural parameters between the Ba-phlogopite from La Creuse (LC4) and selected trioctahedral micas from literature listed below: relationships between the tetrahedral rotation angle α and: (a) the Ba content in kinoshitalites and phlogopites with or without interlayer vacancies; (b) the composition of tetrahedral sheet in a large range of compositions. (c) Dependence of the interlayer separation on the Ba content.

1 – (K_0.70Na_0.22Ba_1.16)(${\rm Fe}_{{\rm 0}{\rm. 10}}^{{\rm 3 +}} $ Al_0.44Mg_4.14${\rm Mn}_{{\rm 1}{\rm. 04}}^{{\rm 2 +}} $)(Si_4.10Al_3.88Ti_0.02)O_20.66F_0.10(OH)_3.24        Kato et al. (Reference Kato, Miùra, Yoshii and Maeda1979)

2 – (K_1.16Na_0.18Ba_0.70)(${\rm Fe}_{{\rm 0}{\rm. 08}}^{{\rm 3 +}} $ Al_0.70Mg_4.20${\rm Mn}_{{\rm 1}{\rm. 04}}^{{\rm 2 +}} {\rm Mn}_{{\rm 0}{\rm. 44}}^{{\rm 3 +}} $)(Si_4.66Al_3.30Ti_0.02)O_21.50F_0.14(OH)_2.36        Guggenheim and Kato (Reference Guggenheim and Kato1984)

3 – (K_0.12Na_0.02Ba_1.98)(Al_0.08Mg_5.28Mn_0.62)(Si_4.06Al_3.94)O₂₀F_0.74Cl_0.04(OH)_3.22         Gnos and Armbruster (Reference Gnos and Armbruster2000)

4 – (K_0.81Na_0.08Ca_0.02Ba_1.09)(Al_0.33${\rm Fe}_{{\rm 0}{\rm. 54}}^{{\rm 2 +}} $ Mg_5.07Ti_0.05)(Si_4.34Al_3.66)O_19.88F_1.43(OH)_2.70         Brigatti and Poppi (Reference Brigatti and Poppi1993) #27

5 – (K_0.95Na_0.08Ca_0.01Ba_0.96)(Al_0.40${\rm Fe}_{{\rm 0}{\rm. 59}}^{{\rm 2 +}} $ Mg_4.88Mn_0.04Ti_0.09)(Si_4.40Al_3.60)O_19.94F_1.34(OH)_2.72        Brigatti and Poppi (Reference Brigatti and Poppi1993) #26

6 – (K_1.77Ba_0.23)(Al_0.49${\rm Fe}_{{\rm 0}{\rm. 47}}^{{\rm 3 +}} {\rm Fe}_{{\rm 1}{\rm. 53}}^{{\rm 2 +}} $ Mg_3.16Mn_0.01Ti_0.34)(Si_5.17Al_2.83)O_21.05F_0.52(OH)_2.43        Brigatti and Poppi (Reference Brigatti and Poppi1993) #25

7 – (K_1.76Na_0.16Ba_0.07)(Al_0.23${\rm Fe}_{{\rm 0}{\rm. 93}}^{{\rm 3 +}} {\rm Fe}_{{\rm 0}{\rm. 84}}^{{\rm 2 +}} $ Mg_3.70Mn_0.02Ti_0.28)(Si_5.30Al_2.70)O_21.12F_0.03(OH)_2.85        Brigatti and Poppi (Reference Brigatti and Poppi1993) #23

8 – (K_1.75Na_0.14Ca_0.05Ba_0.06)(Al_1.86${\rm Fe}_{{\rm 0}{\rm. 82}}^{{\rm 3 +}} {\rm Fe}_{{\rm 0}{\rm. 78}}^{{\rm 2 +}} $ Mg_2.21Mn_0.05Ti_0.28)(Si_5.36Al_2.64)O_22.71F_0.29(OH)_1.00        Brigatti and Poppi (Reference Brigatti and Poppi1993) #20

9 – (K_1.96Na_0.02)($\; {\rm Fe}_{{\rm 0}{\rm. 16}}^{{\rm 3 +}} {\rm Fe}_{{\rm 0}{\rm. 26}}^{{\rm 2 +}} $ Mg_5.46Ti_0.12)(Si_5.64Al_2.08${\rm Fe}_{{\rm 0}{\rm. 28}}^{{\rm 3 +}} $)O_20.02F_0.22(OH)_3.76        Brigatti and Guggenheim Reference Brigatti, Guggenheim, Mottana, Sassi, Thompson and Guggenheim2002) #73

10 – (K_1.86Na_0.10Ba_0.04)($\; {\rm Fe}_{{\rm 0}{\rm. 32}}^{{\rm 3 +}} {\rm Fe}_{{\rm 0}{\rm. 18}}^{{\rm 2 +}} $ Mg_3.30Ti_0.16)(Si_5.68Al_2.08${\rm Fe}_{{\rm 0}{\rm. 24}}^{{\rm 3 +}} $)O_20.34F_0.02(OH)_3.64        Brigatti et al. (Reference Brigatti, Medici, Saccani and Vaccaro1996) #Tae23-1a

11 – (K_1.92Na_0.06Ba_0.02)($\; {\rm Fe}_{{\rm 0}{\rm. 42}}^{{\rm 3 +}} {\rm Fe}_{{\rm 0}{\rm. 14}}^{{\rm 2 +}} $ Mg_5.28Mn_0.02Ti_0.12)(Si_5.70Al_2.20${\rm Fe}_{{\rm 0}{\rm. 10}}^{{\rm 3 +}} $)O_20.26F_0.12(OH)_3.62        Brigatti et al. (Reference Brigatti, Medici, Saccani and Vaccaro1996) #Tas27-Bb

12 – (K_1.80Na_0.04)(Al_0.14${\rm Fe}_{{\rm 0}{\rm. 32}}^{{\rm 2 +}} $ Mg_5.40Ti_0.06)(Si_5.82Al_2.18)O_19.80F_2.26(OH)_1.94        Brigatti and Guggenheim (Reference Brigatti, Guggenheim, Mottana, Sassi, Thompson and Guggenheim2002) #70

13 – (K_1.86Na_0.08)(Al_0.48${\rm Fe}_{{\rm 0}{\rm. 18}}^{{\rm 3 +}} {\rm Fe}_{{\rm 0}{\rm. 24}}^{{\rm 2 +}} $ Mg_4.96Mn_0.02Ti_0.04)(Si_5.48Al_2.52)O_19.98F_0.12(OH)_3.90        Alietti et al. (Reference Alietti, Brigatti and Poppi1995) #1a

14 – (K_1.86Na_0.08)(Al_0.48${\rm Fe}_{{\rm 0}{\rm. 14}}^{{\rm 3 +}} {\rm Fe}_{{\rm 0}{\rm. 22}}^{{\rm 2 +}} $ Mg_5.10Mn_0.02Ti_0.04)(Si_5.30Al_2.70)O_19.92F_0.18(OH)_3.90        Alietti et al. (Reference Alietti, Brigatti and Poppi1995) #1b

15 – (K_1.86Na_0.04Ba_0.02)(Al_0.36${\rm Fe}_{{\rm 0}{\rm. 30}}^{{\rm 3 +}} {\rm Fe}_{{\rm 0}{\rm. 06}}^{{\rm 2 +}} $ Mg_5.26Ti_0.04)(Si_5.20Al_2.80)O_19.86F_0.22(OH)_3.92        Alietti et al. (Reference Alietti, Brigatti and Poppi1995) #2a

16 – (K_1.90Na_0.04Ba_0.02)(Al_0.94${\rm Fe}_{{\rm 0}{\rm. 30}}^{{\rm 3 +}} {\rm Fe}_{{\rm 0}{\rm. 14}}^{{\rm 2 +}} $ Mg_4.46Mn_0.08Ti_0.02)(Si_5.00Al_3.00)O_20.04F_0.08(OH)_3.88        Alietti et al. (Reference Alietti, Brigatti and Poppi1995) #3a

17 – (K_1.54Na_0.32Ba_0.10)Mg_6.00 (Si_5.90Al_2.10)O₂₀F_2.60(OH)_1.40        Hazen and Burnham (Reference Hazen and Burnham1973)

18 – (K_1.94Na_0.02Ba_0.04)($\; {\rm Fe}_{{\rm 0}{\rm. 06}}^{{\rm 2 +}} $ Mg_4.92${\rm Ti}_{{\rm 0}{\rm. 18}}^{{\rm 3 +}} $ Li_0.46Na_0.22) (Si_6.62Al_0.08${\rm Fe}_{{\rm 1}{\rm. 30}}^{{\rm 3 +}} $)O₂₀F_4.0        Hazen et al. (Reference Hazen, Finger and Velde1981)

19 – (K_1.72Na_0.04)(Al_0.96${\rm Fe}_{{\rm 1}{\rm. 20}}^{{\rm 2 +}} $ Mg_3.38Mn_0.06Zn_0.02Ti_0.22)(Si_5.50Al_2.40${\rm Fe}_{{\rm 0}{\rm. 10}}^{{\rm 3 +}} $)O_20.28F_0.24(OH)_3.48        Brigatti et al. (Reference Brigatti, Guidotti, Malferrari and Sassi2008) #O-K-9

20 – (K_1.50Na_0.12Ba_0.12)(Al_0.2Cr_0.06${\rm Fe}_{{\rm 0}{\rm. 04}}^{{\rm 3 +}} {\rm Fe}_{{\rm 0}{\rm. 32}}^{{\rm 2 +}} $ Mg_5.42Ti_0.12)(Si_5.62Al_2.38)O_20.24F_0.06Cl_0.04(OH)_3.66        Comodi et al. (Reference Comodi, Nazzareni, Fumagallo and Capitani2011) #MO10-1

21 – (K_1.48Na_0.18Ba_0.12)(Al_0.24Cr_0.06${\rm Fe}_{{\rm 0}{\rm. 04}}^{{\rm 3 +}} {\rm Fe}_{{\rm 0}{\rm. 32}}^{{\rm 2 +}} $ Mg_5.34Ti_0.10)(Si_5.70Al_2.30)O_20.24F_0.06Cl_0.04(OH)_3.66        Comodi et al. (Reference Comodi, Nazzareni, Fumagallo and Capitani2011) #MO10-2

22 – (K_1.62Na_0.06Ba_0.02)(Al_0.28Cr_0.06${\rm Fe}_{{\rm 0}{\rm. 02}}^{{\rm 3 +}} {\rm Fe}_{{\rm 0}{\rm. 26}}^{{\rm 2 +}} $ Mg_5.54Ti_0.02)(Si_5.66Al_2.34)O_20.14F_0.04Cl_0.04(OH)_3.86        Comodi et al. (Reference Comodi, Nazzareni, Fumagallo and Capitani2011) #300b-1

23 – (K_1.62Na_0.06Ba_0.02)(Al_0.32Cr_0.06${\rm Fe}_{{\rm 0}{\rm. 04}}^{{\rm 3 +}} {\rm Fe}_{{\rm 0}{\rm. 30}}^{{\rm 2 +}} $ Mg_5.30Ti_0.06)(Si_5.68Al_2.32)O_20.14F_0.04Cl_0.04(OH)_3.86        Comodi et al. (Reference Comodi, Nazzareni, Fumagallo and Capitani2011) #300b-2

24 – (K_0.17Na_0.95)(Al_0.78Cr_0.01Fe_0.75Mg_4.21Ti_0.07)(Si_6.20Al_1.80)O₂₀(OH,F)₄        Kogure et al. (Reference Kogure, Miyawaki and Banno2005)

Reasons for doubting the likelihood of the talc substitution (charge imbalance on the basal oxygen atoms, large expected interlayer cavity following the presence of vacancies compared to that in kinoshitalite) probably set limits on this solid solution. What is observed is that the mole fraction of talc is neither smaller nor larger than 0.28 (±0.03). This value thus may reflect the saturation in talc component. Most importantly, if the Ba-mica were a mixture of two phases, talc and Ba-phlogopite, with talc nano-domains or layers, the relative amounts of both components would be unconstrained, as the bulk chemical composition of the rocks is highly variable, especially as regards the Ba content. In contrast, there is no exception to the constant character of the mole fraction of the talc component, in spite of the large number of available analyses (201). Therefore, the observed composition must correspond to a single phase, or possibly to a former single phase secondarily unmixed as in the case of wonesite with talc exsolutions. Owing to the probably moderate temperature of the talc-forming metasomatic process, such a single phase may well have been metastable at the time of its formation.

Mg-metasomatism vs protolith chemistry

In the talc-absent zone, Al-rich phlogopite formed through reaction between the Mg-rich infiltrating fluid and K-feldspar or muscovite. In the part of the metasomatic sequence observed in La Creuse, muscovite (and pseudomorphs after cordierite) persist as residual islands. The presence of Ba is restricted to the occurrence of hyalophane replacing K-feldspar close to the baryte zone (apart from a few late-stage crosscutting baryte geodic veinlets). In the talc + baryte zone, the Ba-rich phlogopite was produced by a similar reaction between feldspar, now present as celsian, and the talc-producing magnesian fluid. The Ba contents of both phlogopite and feldspars are related by a reciprocal equilibrium between celsian, K-feldspar, phlogopite and kinoshitalite:

$$\eqalign{& {\rm K}_2{\rm Mg}_6{\rm Al}_2{\rm Si}_6{\rm O}_{20}\lpar {{\rm OH}} \rpar _4{\rm +} 2{\rm BaAl}_2{\rm Si}_2{\rm O_8} = \cr & \quad {\rm Ba}_2{\rm Mg}_6{\rm Al}_4{\rm Si}_4{\rm O}_{20}\lpar {{\rm OH}} \rpar _4 + 2{\rm KAlSi}_3{\rm O}_8} $$

which involves the Wendlandt substitution (BaAl)(KSi)_–1 as an exchange vector in both solid solutions. However, this substitution is not that observed in the phlogopite. The remarkable invariance of both Si and Mg + Fe²⁺ contents in phlogopites from the talc zone, emphasized above as the major control on the compositional range of the Ba-rich phlogopite, is suggested to result from buffering by the metasomatic conditions that produced large amounts of talc and preserved the persistence of quartz, especially inside some baryte lenses. The observed composition of the Ba-rich phlogopite is therefore controlled by the Mg-metasomatic process that produced talc in a Ba-rich and Si-rich, Al-poor rock, and this mineral is essentially a Ba-mineral, therefore not the same mineral (from a petrological viewpoint) as the Al-rich phlogopite.

Interlayer vacancies

A distinctive feature of the metasomatic process is the leaching of K, evidenced by: (1) the replacement of K by Na between the Ba-poor, talc-absent zone and the talc zone (Fig. 5); (2) the replacement of K by Ba in both the feldspar and the phlogopite in the Ba-rich zone in which no K-feldspar is left; and finally (3) the absence of any K-bearing mineral in the upstream, phlogopite-absent, talc-chlorite zone. This tendency is consistent with the high level of interlayer vacancy in the phlogopites from La Creuse (0.56 pfu in phlogopite in equilibrium with talc), compared to more common Ba-rich phlogopites. Whereas the micas from granites, pegmatites and volcanic or subvolcanic alkaline rocks, all rocks in which the mica is in equilibrium with K-feldspar or leucite, have usually two atoms in the interlayer site (Foster, Reference Foster1960), a total number of atoms (K, Na and Ba) less than 2, down to 1.4, can be found in micas from rocks devoid of K-feldspar such as the metapelites belonging to the contact-metamorphic aureole mentioned above (Guidotti and Dyar, Reference Guidotti and Dyar1991; Brigatti et al., Reference Brigatti, Guidotti, Malferrari and Sassi2008) or the talc-absent metarhyolites and metapelites at La Creuse. A possible interpretation could be that the chemical potential of K₂O would be too low to saturate the biotite interlayer site. Such an interpretation is consistent with the peculiarities of phlogopite in the talc zone of La Creuse, undoubtedly stabilized by the availability of Ba in conditions of K leaching out.

To conclude, this Ba-phlogopite is original by a composition chemically corresponding to the vacancy-producing substitution (Ba_0.5□_0.5)(K)_–1 which maintains the maximum Si content. This substitution can be decomposed as ½ Wendlandt substitution (BaAl)(KSi)_–1 + ½ ‘talc’ substitution (□Si)(KAl)_–1. The interlayer vacancy is therefore essential to the charge balance in a phlogopite rich in both Ba and Si. This result is supported by the XRD analysis and is consistent with a single phase close to phlogopite for most cell parameters and with an interlayer vacancy. It is also supported by the petrology of the Ba-rich rocks at La Creuse in which Si-rich and Mg-rich conditions are evidenced by the important development of talc. As counter examples, phlogopites from rocks containing baryte and devoid of talc are unknown at La Creuse but are commonly observed in the nearby sulfide deposit of Chessy (Fig. 2a), where the mineralized area is also located in a Mg metasomatic aureole. Only one sample of ore containing Ba-rich phlogopite has been reported (Poupon, Reference Poupon1989), and this Ba-phlogopite belongs to the ordinary type with filled interlayer (Fig. 1a).