Mass transfer and element redistribution during chloritization of metamorphic biotite in a metapelite: insights from compositional mapping

Mass balance calculations can be carried out by comparing whole-rock composition between altered and least-altered samples (e.g., Grant, 1986, 2005; Osterberg et al. 1987; Ague 1994, 2011) or by writing non-stoichiometric equations for fluid-driven reactions (e.g., Gresens 1967; Centrella et al. 2015, 2018). Despite the differences, all approaches to mass balance calculations assume one single fluid phase with a constant fluid chemical composition. Gresens (1967) argued that the comparison of whole-rock composition lacks insight into mineral forming processes and element redistribution between parent and product phases accompanied by volume and/or mass changes. For instance, if veins of epidote or quartz cut through altered portions of a rock, whole-rock chemistry might not produce reliable results and careful textural and mineralogical studies would be more suitable for studying specific reactions. Gresens (1967) proposed a set of equations that use chemical composition, mineral densities, and solid volume change. To establish gains and losses, knowledge of volume change or geochemical behavior of elements is needed. When N minerals are compared in a single alteration equation, N-1 additional facts or assumptions are required, adding uncertainties regarding fluid-driven reactions.

Grant (1986, 2005) provided a new method for the solution of Gresens’ (1967) equations to determine element concentration changes, called the isocon analysis. On an isocon diagram element concentration in the altered sample is plotted against that in the precursor. A straight line can be plotted through immobile elements to define an isocon. The slope of an isocon reveals overall mass change during alteration, while the distance of each point to the isocon illustrates concentration change for each element. This approach has been used to resolve mass change during hydrothermal alteration (Osterberg et al. 1987), replacement reactions (Plümper, Putnis 2009; Moore et al. 2019), formation of shear zones (Condie, Sinha, 1996) or migmatites (Olsen, Grant 1991). The application of the isocon method relies on an appropriate selection of immobile elements. Baumgartner and Olsen (1995) proposed a statistical approach to that problem that considers uncertainties from rock heterogeneity and analytical errors. Using Gresens’ (1967) and Grants’ (1986) calculations, Ague (1994, 2003, 2011) proposed a new graphical method for investigations of mass balance. On Ague’s (1994, 2003, 2011) concentration ratio diagram, immobile elements define a reference frame and a straight line, indicating a total mass change. Ti or Zr are often considered immobile during metasomatism and become mobile only under high P-T conditions or in halogens-rich fluids.

Chloritization of biotite was the focus of multiple studies and many attempts have been made to understand element redistribution and solid volume change during this reaction. Previous studies proposed multiple mechanisms of biotite replacement across geological settings (Ferry 1979; Parry, Downey 1982; Veblen, Ferry 1983; Baños, Amouric 1984; Eggleton, Banfield 1985; Parneix et al. 1985; Kogure, Banfield 2000; Yuguchi et al. 2015, Xiao, Chen 2020). Veblen and Ferry (1983) suggested that a single layer of chlorite is produced from two biotite layers as it inherits an octahedral sheet. This change is accompanied by the loss of tetrahedral sheet and potassium interlayer sheet in biotite. The reaction results in vacancy and disarray of sheets in the crystal structure of biotite with subsequent volume loss (fulfilled by new phases). Another mechanism was presented by Baños and Amouric (1984), where one layer of biotite is altered into one layer of chlorite. During this process, the potassium interlayer in biotite is immediately replaced by a brucite-like sheet.

Parry and Downey (1982) concluded that during the isovolumetric replacement of biotite, Al is conserved and K, Mg, and Fe are added or subtracted. Considering the same reaction, Ferry (1979) showed that to conserve Al, 13% of volume loss is needed and the addition of Mg and Fe is required. Further studies by Veblen and Ferry (1983) argued for higher volume loss during chloritization. In contrast, Parneix et al. (1985) considered the conservation of Ti during the reaction and argued against volume loss during biotite breakdown. Yuguchi et al. (2015) applied singular value decomposition (SVD) analysis to chloritization for eight mineral assemblages. These authors primarily distinguish two mechanisms for chlorite formation: (1) the first one involving small volume decrease and large inflow of Al, Fe, Mn, and Mg with hydrothermal fluid, and (2) the second one with large volume decrease and large outflow of ions into hydrothermal fluids. These authors link variance in the composition of hydrothermal fluids to gradual changes accompanying progressing reaction.

The studied sample was collected in Ny Friesland, Svalbard Archipelago, Norway, from an outcrop of garnet-bearing metapelite (KK15-4; coordinates: 79.20248ºN, 16.50917ºE) belonging to the Atomfjella Complex. The pre-Devonian basement of Svalbard is divided into three provinces with different stratigraphy and tectono-metamorphic history: Southwestern, Eastern, and Northwestern Province (e.g., Gee, Teben’kov 2004; Fig. 1). Svalbard’s basement provinces are cut by N-S trending strike-slip faults that facilitated Silurian to Devonian amalgamation of Svalbard (Harland et al. 1992; Gee, Teben’kov 2004; Majka, Kośmińska 2017; Faehnrich et al. 2020). The studied sample comes from the eastern shore of the Wijdefjorden and was collected along the strike of a major fault zone – Billefjorden Fault Zone (BFZ, Fig. 1). The BFZ has a long history of reactivation, with ductile structures overprinted by brittle deformation (e.g., Harland et al. 1974; McCann, Dallmann 1996; Dallmann, Piepjohn 2020). The joints that cut garnet-bearing metapelite acted as fluid pathways causing the alteration of biotite with the visible presence of chlorite. The alteration zone is narrow and restricted to 5 cm away from the joint. The sample KK15-4 was collected next to the joint surface, preserves the alteration front and metamorphic protolith (Fig. 2), providing an opportunity to study fluid-driven reactions in metapelites during brittle deformation.

Figure 1.

Figure 2.

In the unaltered (parent) rock (zone C), the metamorphic assemblage consists of garnet, biotite, plagioclase, and quartz, with accessory tourmaline, and zircon. Garnet forms euhedral to subhedral grains with inclusions of quartz and biotite, more abundant in the core (Fig. 3A). In plagioclase, early stages of sericitization are recognized. In the unaltered zone (zone C), biotite do not show signs of alteration, while 4 cm from the joint (zone B) only one occurrence of chlorite was noted (Fig. 3B). In the most altered zone (zone A), biotite is breaking down to chlorite, K-feldspar, titanite, and rutile. Reaction either occurs completely, without remnants of biotite, or chlorite forms intercalations within the parent mineral (Figs. 3C,D and 4).

Figure 3.

Figure 4.

The whole-rock composition is reported in Table 1. There are notable differences in the concentration of major elements between the three sections of the sample. For instance, SiO₂ is 60.70 wt% in the unaltered sample (C) and 61.14 wt% in the altered one (A), similarly MgO changes from 2.65 wt% to 3.01 wt% and Loss on Ignition (LOI) from 1.2 wt% to 2.2 wt%. Trace elements such as Ba show variation from 653 ppm to 589 ppm and Sr from 117 ppm to 132 ppm in the unaltered to the altered zone. In contrast, Zr does not show much variation and changes between 245 and 258 ppm (the analytical uncertainty is 5 ppm). For this reason, ratios of elements to relatively immobile Zr are plotted in Fig. 5. The changes in Fe/Zr, Mg/Zr, and K/Zr are most significant between zone A and C and show little variation between zone B and the unaltered protolith (C).

Figure 5.

In biotite, X_Fe (=Fe_tot/Fe_tot+Mg) varies from 0.54 to 0.63, and the Ti content from 0.28 to 0.49 a.p.f.u. (Fig. 6A, Table S1). In chlorite, X_Fe changes from 0.27 to 0.61 within and between the zones (Fig. 6B, Table S2). Two types of chlorite can be distinguished, chlorite from zone C predominately falls within pennine/diabanite and ripidolite fields on Hey’s (1954) classification diagram, whereas chlorite from zone A plots mostly within brunsvigite and pennine/diabanite fields (Fig. 6B). In secondary K-feldspar, potassium content varies from 0.82 to 0.97 a.p.f.u. with Ca and Na changing from 0.01 to 0.12 a.p.f.u. (Table S3). Plagioclase is dominated by albite with subordinate anorthite component (X_an is up to 0.29, Table S4). Titanite displays elevated Al content (up to 0.12 a.p.f.u.) and Fe (0.03 to 0.04 a.p.f.u.) (Table S5). Comparably, minor amounts of Si, Al, Mg, and Fe in rutile (Table S5) could reflect substitution or errors in the EMP analyses caused by small crystal size against the spot size.

Figure 6.

The first set of mass balance calculations was conducted on the whole-rock composition (comparing zone A to C) under the assumption that Zr is immobile and without consideration of solid volume variation or density change during the hydration reaction. The concentration ratio diagram is presented in Fig. 7 and follows the approach of Ague (1994, 2003, 2011). This diagram shows that a small “mass loss” occurred during metasomatism (1.2 wt%). However, the calculated “mass loss” is within the analytical uncertainty of Zr concertation that drive these calculations (245±5 ppm versus 248±5 ppm in the altered and unaltered zone respectively) and needs to be taken into consideration with caution. Al₂O₃, K₂O, CaO, MnO, Ba, Ni, Y were removed from the system, whereas SiO₂, Fe2O₃, MgO, Na₂O, TiO₂, P₂O₅, Sr, Nb, Sc were introduced with the fluid. However, SiO₂ and Cr₂O₃ behave almost as immobile components. The highest losses are in CaO (-17.85%) and K₂O (-15.42%) and the highest gains are in LOI= H₂O (+85.58%) and MgO (+14.98%).

Figure 7.

The second set of mass balance calculations was done for the Bt-Chl replacement reaction, neglecting other product phases (K-feldspar, titanite, or rutile). We assume that Cr and Al are immobile during this reaction and use the average composition of biotite and chlorite from EMP spot analyses (Table S6). Calculations use the Isocon method (Grant 1986, 2005) and are presented in Fig. 8. The analyses show which element oxides must be introduced to the system to create chlorite and which are released either with the fluid or to create secondary phases. The H₂O (+11%), MgO (+3%), FeO (+3%) need to be added and K₂O (-6%), TiO₂ (-6%), SiO₂ (-2%), Na₂O (-5%) are released.

Figure 8.

X-Ray maps were obtained for three areas to investigate replacement textures, chemical composition, and proportions of mineral phases in the altered sample KK15-4A, therefore all three areas are from the same distance from the joint. Based on point EMP analyses, mineral proportions and their molar volume, local mass balance was calculated. These calculations assume an isovolumetric reaction for pseudomorphic replacement of Bt by Chl + Kfs ± Rt ± Ttn. The first set of maps was obtained for Si, Al, Fe, K, Mg from a 1.2 mm by 1.2 mm area (Fig. 9). Both Fe and Mg distributions clearly point out areas of epitaxial chlorite replacement within biotite crystals (Figs. 9E,F). Distribution of K suggests two types of K-feldspar; one represented by small grains (~30-60 μm) produced in proximity to biotite, and the second one represented by larger (up to 100 μm) grains without clear association with other phases (Fig. 9D). Within this area, chlorite occupies more than 10%, while rutile and K-feldspar ca. 6.5% (Fig. 10). Assuming that these are the only product phases from the biotite breakdown, that the reaction is isovolumetric, and using molar volume from Tables S1-5 and phase proportions in Fig. 10, the following reaction can be written down: 1.08Bt=>0.62Chl+0.27Kfs+0.12Rt$$1.08{\rm{Bt}} = > 0.62{\rm{Chl}} + 0.27{\rm{Kfs}} + 0.12{\rm{Rt}}$$

Figure 9.

Figure 10.

Using the coefficients in this equation and the structural formula derived from the EMP analysis, we calculated the gains and losses of the elements. Si, Ti, Al, Na, and K were lost during this reaction and Fe, Mn, Mg, and Ca were gained (Table 2).

The second set of chemical maps covers a smaller area (ca. 170 by 170 μm) and X-Ray maps were obtained for Si, Al, K, Fe, Mg, Na, Ca, Ti, and Mn (Fig. 11). This area only hosts one biotite grain, which is surrounded by plagioclase, and both titanite and rutile occur as Ti bearing phases. Chlorite and K-feldspar seem to replace biotite within its initial contour (as a pseudomorphic reaction). Mineral proportions (Fig. 11K) and mass balance (calculated as described above), resulted in the following replacement reaction: 1.17Bt=>0.74Chl+0.15Ttn+0.08Kfs+0.02Rt$$1.17{\rm{Bt}} = > 0.74{\rm{Chl}} + 0.15{\rm{Ttn}} + 0.08{\rm{Kfs}} + 0.02{\rm{Rt}}$$

Figure 11.

As in the case of the previous calculation, Si, Ti, Al, Na, and K were gained while Fe, Mn, Mg, and Ca were lost (Table 2).

The third area (ca. 240 by 120 μm) includes biotite almost completely replaced by chlorite and only a small portion of the biotite is preserved (Fig. 12). The studied crystal is surrounded by plagioclase and quartz. There are two varieties of chlorite: one more Mg- and the second more Fe-rich (Fig. 11), but we use the average composition of both for the replacement reaction. Similar to the second area, both rutile and titanite are present as Ti-bearing phases (Fig. 11). The calculated reaction is as follows: 1.05Bt=>0.53Chl+0.03Ttn+0.40Kfs+0.05Rt$$1.05{\rm{Bt}} = > 0.53{\rm{Chl}} + 0.03{\rm{Ttn}} + 0.40{\rm{Kfs}} + 0.05{\rm{Rt}}$$

Figure 12.

In this area, only Mg and Ca were gained during the replacement reaction, and the remaining components (Si, Ti, Al, Na, K, Mn, and Fe) were lost (Table 2).

A summary of the mass balance calculations is presented in Fig. 13. Relative ratios of the product phases vary between the studied areas (the sum is always 1 mol). In all cases, the dominant product of biotite breakdown is chlorite, ranging from 0.74 mol to 0.53 mol. There is a notable variation in the amount of K-feldspar produced, from 0.4 mol to 0.08 mol. Proportions of Ti-bearing phases also vary. In the first area, titanite was not detected as a product of the biotite breakdown. Instead, all of Ti was incorporated in rutile. The three areas generally show a similar pattern in terms of element redistribution during this reaction with Si, Ti, Al, Na, and K lost to the fluid (Fig. 13B, positive values) and Fe, Mn, Mg, and Ca gained. Only in the third area, Fe was lost during the reactions.

Figure 13.