Earth and Planetary Science Letters 395 (2014) 205–216

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

www.elsevier.com/locate/epsl

Rb–Sr ages from phengite inclusions in garnets from high pressurerocks of the Swiss Western Alps

Caroline M.C. de Meyer a,∗, Lukas P. Baumgartner a, Brian L. Beard b, Clark M. Johnson b

a Institute of Earth Sciences, University of Lausanne, Géopolis, CH-1015 Lausanne, Switzerlandb Department of Geoscience, University of Wisconsin, 1215 W. Dayton, Madison, WI 53706, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 14 October 2013Received in revised form 21 March 2014Accepted 22 March 2014Available online 13 April 2014Editor: T.M. Harrison

Keywords:Rb–SrgeochronologyWestern Alpsgarnethigh pressure metamorphismZermatt-Saas Fee ophiolite

The Zermatt-Saas Fee Zone (ZSZ) was subducted to eclogite-facies conditions, reaching peak pressuresand temperatures of 20–28 kbar and 500–630 ◦C. The rocks were partially overprinted under greenschist-facies conditions during exhumation. Previous Rb–Sr isochron ages obtained on matrix phengites inmetasediments of the ZSZ have been interpreted to date early exhumation of the ZSZ. Here wepresent new Rb–Sr geochronology on phengite inclusions in garnets to date prograde growth ofgarnets. We show that garnet acted as a shield for the included phengites, limiting Rb and Sr isotopeexchange with the bulk rock, upon complete enclosure of the mica, during garnet growth, even ifpeak metamorphism exceeded the Rb–Sr blocking temperature. Similarly, garnet isolated the micasfrom the matrix during subsequent recrystallization due to fluid infiltration or deformation duringexhumation. Phengite inclusion ages for two metapelitic samples from the same locality (Triftji) are44.86 ± 0.49 Ma and 43.6 ± 1.8 Ma, and are about 4 m.y. older than the corresponding matrix micaages of 40.01 ± 0.51 Ma and 39.5 ± 1.1 Ma, respectively. The results confirm previous Sm–Nd and Lu–Hfgeochronology on the ZSZ that indicated protracted garnet growth during prograde metamorphism, andconfirm that at least parts of the ZSZ underwent peak metamorphic HP conditions less than 43 m.y. ago,followed by rapid exhumation to upper greenschist-facies conditions around 40 Ma ago.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Understanding the time-temperature pathways of progrademetamorphism has been a long-standing goal of studies in oro-genic belts. Garnet is common in the metamorphic assemblagein metamorphic mafic and metasedimentary rocks in Barrovianand subduction zone metamorphism, and hence has been tar-geted by quite a few geochronological studies (e.g. Baxter andScherer, 2013). Garnet often grows along a prograde path (e.g.Spear, 1993), and major-element zoning patterns may retain in-formation on the pressures and temperatures attained duringthe prograde history. The advances made in Lu–Hf and Sm–Ndgeochronology of garnet (e.g. Lapen et al., 2003; Skora et al., 2006;Harvey and Baxter, 2009; Baxter and Scherer, 2013) have madedirect dating of some garnets possible, thus permitting recon-struction of the timing of prograde metamorphic events relatingspecific geochronological data to periods of garnet growth. Garnetgrowth rates have been inferred from modelling of Lu–Hf and Sm–Nd and other REEs (Lapen et al., 2003; Skora et al., 2006, 2009)

* Corresponding author. Tel.: +41 216924300.E-mail address: demeyerc.caroline@gmail.com (C.M.C. de Meyer).

http://dx.doi.org/10.1016/j.epsl.2014.03.0500012-821X/© 2014 Elsevier B.V. All rights reserved.

or by direct dating of rim and core (Pollington and Baxter, 2011;Dragovic et al., 2012). Although these isotopic systems have con-tributed greatly to our understanding of the timescales of garnetgrowth during prograde metamorphism, age errors are generallyfairly high for these isotopic systems in contrast to other isotopicsystems where the parent/daughter ratios may be quite high, suchas the Rb–Sr isotopic system. In the case of garnet, however, Rb–Srgeochronology of garnets has rarely been utilized, because Sr con-centrations in garnet are low and its Rb/Sr ratio is relatively small,making direct dating of garnet extremely difficult (e.g., Chris-tensen et al., 1994, 1989; Vance and O’Nions, 1990; Kohn, 2009;Sousa et al., 2013).

Potassium-rich, high Rb/Sr phases such as mica, alkali-feldsparor even amphiboles have been successfully used to constrain thetime of mineral growth. Isotopic resetting due to diffusive ex-change or recrystallization during deformation and fluid flow canoverprint the age of prograde crystallization of white micas. There-fore Rb–Sr mica ages often represent exhumation ages. Here wepropose a new approach: the use of the Rb–Sr system to determinegarnet growth timing by dating phengite inclusions in the garnet.Since diffusion in garnet is generally considered to be slow, thiswill likely prevent re-equilibration with matrix minerals, if micas

206 C.M.C. de Meyer et al. / Earth and Planetary Science Letters 395 (2014) 205–216

Fig. 1. Tectonic map of the Western Alps. The Zermatt-Saas Fee Zone (ZSZ) (seeFig. 2) is a lithologically defined unit within the northern end of the Piemonte-Ligurian oceanic basin. It underwent eclogite facies metamorphism. After Barnicoatand Fry (1986).

are entirely enclosed within garnet. This will result in an appar-ent increase in temperature for the Rb–Sr “closure”, approachingthe closure temperature of garnet for Rb–Sr dating. The conceptof a texturally controlled shielding of minerals as a means for pre-venting isotopic resetting has been used amongst others by Montelet al. (2000), who showed that monazite does not exchange if itis shielded by garnet even well above its “Dodson closure tem-perature”. Gouzu et al. (2006) applied Ar/Ar in situ geochronologyon phengite included in garnet of the Lago di Cignana locality inthe Zermatt-Saas Fee Zone (ZSZ; see Fig. 2). They found that theinclusions are about 6 m.y. older than the matrix phengites, inagreement with the shielding concept discussed above.

The study area is the Zermatt-Saas-Fee Zone (ZSZ) in theWestern Alps (Figs. 1 and 2). The ZSZ is composed of deeplysubducted rocks that were subjected to eclogite-facies metamor-phism. Whereas the mafic eclogitic rocks have been the focus ofmuch dating activity (e.g. Rubatto et al., 1998; Lapen et al., 2003;Gouzu et al., 2006; Skora et al. 2006, 2009), only a few attemptshave been made to date the high-pressure peak metamorphic as-semblages of the associated metasediments (e.g. Rubatto et al.,1998). Here, we analyzed metasediments containing garnets withphengite inclusions. The metasediments are part of the ZSZ, whichare included in – and contain – eclogites. These rocks provide anideal test case for the use of Rb–Sr geochronology on phengite in-clusions for determining the timing of garnet growth.

2. Sample selection and analytical techniques

Garnet-bearing metasediments of the ZSZ were collected andcharacterized by micro-computed X-ray tomography (μ-X-CT) and

electron microprobe analysis (EMPA). Samples were selected onthe basis of the following criteria: 1) most garnets contain severalwhite mica inclusions per grain 2) mica inclusions are phengites,rather than paragonites; and 3) garnet porphyroblasts are not frac-tured or strongly poikilitic. For each sample the crystal size distri-bution and the general shape and amount of garnet fracturing wasdetermined on a representative volume of the sample from 3D-images reconstruction from μ-X-CT. Qualitative and quantitativechemical analyses of mica inclusions were obtained by EMPA onthin sections to evaluate the textural context and chemistry of thewhite mica. Only samples containing phengite inclusions isolatedfrom the matrix were selected on the basis of both 3D tomogra-phy images and 2D thin sections.

Mica inclusions. Entire garnet crystals were separated from samplesusing a SelfragTm electro disintegrator at the University of Bern.SelfragTm fractures the rock samples mostly along grain boundariesusing high-energy electrical discharges. Garnet grains, along withthe other minerals grains, are mainly obtained as whole crystals.About ∼2 kg of each sample were disintegrated. The fragmentedmaterial was wet sieved into several fractions (>4 mm, 2–4 mm,1–2 mm and 0.5–1 mm) and dried in an oven at 60 ◦C. Garnetswere handpicked from each of those fractions. The garnet fractionswere separated into poikilitic and non-poikilitic fractions whenboth textures occurred in the sample. For sample 09CDM105 allof the garnets, and for sample 09CDM110 only the non-poikiliticgarnets were selected for further processing. Garnets bigger than 2mm were mechanically abraded by packing them into small plas-tic tubes filled with water and the tubes placed into a shakerapparatus (see Supplementary Materials for images of garnets be-fore and after abrasion). This procedure removes the outermostrim of the garnet, together with any attached micas. The garnetswere checked under a binocular after 15–30 min for attached micagrains and the process was repeated until all external mica wasremoved. Up to two hours of shaking was required. Only those gar-nets without any visible micas or other minerals at their surface orin cracks were selected for further processing.

Phengite inclusions were extracted by crushing the cleaned gar-nets in a steel hand crusher followed by handpicking under abinocular microscope. A total of 0.2 to 2.1 mg of white mica wasobtained for each sample. The size of the included micas variesfrom less than 100 μm (sample 09CDM105) to a maximum sizeof ∼300 μm (sample 09CDM110). The chemical sample selectionprocedure (e.g. EMPA) ascertained that these micas were mainlyphengite, with less than 10% paragonite.

Matrix minerals. The fragmented, wet sieved fraction smaller than600 μm was further dry-sieved and separated into four size frac-tions of 600–500 μm, 500–400 μm, 400–300 μm, and 300–180 μm.Matrix minerals (zoisite and tourmaline) were hand-picked fromthe whole rock fractions under a binocular microscope. White micawas magnetically separated from the whole rock fractions using aFrantz separator. A second pass in the Frantz magnetic separatorand hand picking under a binocular microscope was necessary toincrease purity of the phengite separates.

Rb and Sr analysis. All Rb–Sr isotope analyses were performed atthe University of Wisconsin-Madison using a Micromass Sector 54thermal ionization mass spectrometer (TIMS). Samples were spikedusing a mixed 87Rb–84Sr spike that was calibrated against multiplenormal solutions to determine Rb and Sr concentrations by iso-tope dilution mass spectrometry (IDMS) and dissolved in a 10:1HF:HNO3 solution. Rubidium and strontium were subsequently pu-rified using EiChrome Sr-specific resin and HNO3 to obtain a pureSr cut, followed by purification of Rb using BioRad AG MP-50macroporus cation exchange resin with 2 M HCl (e.g. Beard et al.,2013). Strontium was loaded onto Re filaments with a TaF activator

C.M.C. de Meyer et al. / Earth and Planetary Science Letters 395 (2014) 205–216 207

Fig. 2. Overview of published geochronological ages for the prograde metamorphic path of the ZSZ (and Monte Rosa nappe: Gressonay Valley), North of the Aosta Valley. Notethe considerable spread obtained with different methods, but also with identical isotopic systems. rt = rutile; phg = phengite; zrn = zircon. Sources of the ages are Bowtellet al. (1994), Rubatto et al. (1998), Amato et al. (1999), Mayer et al. (1999), Lapen et al. (2003, 2007), Mahlen et al. (2005, 2006), Gouzu et al. (2006), Liati and Froitzheim(2006) and Herwartz et al. (2008). Tectonic map modified after Steck et al. (2000).

following the methods of Charlier et al. (2006) and Beard et al.(2013). Finally, Sr isotope analyses were obtained using a multi-dynamic analysis with internal exponential normalization to an86Sr/88Sr value of 0.1194. In general, the reported isotopic ratiosare the average of 120 determinations measured at an 88Sr ion in-tensity of 3 × 10−11 A; uncertainties reported for each analysis are2σ , based on in-run statistics. Repeat analysis of the NIST SRM-987Sr isotope standard run under similar conditions as the samplesyielded an average 87Sr/86Sr value of 0.710267 ± 0.000014 (2σ ,n = 17). Rubidium was loaded onto Ta filaments with H3PO4, and87Rb/85Rb values were determined by static multi collector analy-sis. The long-term reproducibility of multiple analyses of 85Rb/87Rbfor the NIST SRM-984 Rb standard is 0.5%. Because the chemicalseparation procedure produces very clean Rb cuts that are free ofother ions, the uncertainties in Rb isotopic measurement of theNIST SRM-984 standard can be used to estimate the uncertaintiesin the 87Rb/86Sr ratio of the sample.

Total procedural blanks were determined by IDMS, and blanksare reported as grouped with the specific types of samples thatwere analyzed. Blanks measured during chemical separation ofwhole rocks were 434 pg for Rb and 227 pg for Sr, during measure-ment of matrix phengite blanks were 17.5 pg for Rb and 51.5 pgfor Sr. For analysis of tourmaline and zoisite blanks were 17.3 pgfor Rb and 82.0 pg for Sr. For mica inclusions in garnet, whichwere further purified using twice the EiChrome Sr-specific resin,procedural blanks were 13.9 pg for Rb and 62.5 pg for Sr forsample 09CDM110 and 19.3 pg for Rb and 20.5 pg for Sr for

sample 09CDM105. These blank levels result in Sr sample: blankratios of at least 103 for whole rock and matrix mineral analy-ses, and the Rb and Sr contents and 87Sr/86Sr ratios are reportedas measured because blank correction is entirely negligible. Thevery small quantities of Rb and Sr measured for the phengite in-clusions in garnet, however, make these analyses more susceptibleto blanks; Sr sample to blank ratios for blanks determined duringprocessing of these samples were 174 × 101 and 19 for samples09CDM105 and 09CDM110, respectively (Table 2). The 87Sr/86Sr ra-tio of the blank was determined to be 0.7091 ± 0.0030 (Beard etal., 2013). We considered a 50% uncertainty on the average valueof the Rb and Sr blanks of the mineral separates.

To calculate 87Rb/86Sr ratios, Sr isotope abundances were deter-mined using an 84Sr/86Sr ratio equal to that of our average mea-sured NIST SRM-987 standard (0.056500 ± 0.000020), an 86Sr/88Srratio of 0.1194, and the spike-subtracted 87Sr/86Sr ratios. For Rbisotope abundances, we used the values determined in the NISTSRM-984 Rb standard as reported by Steiger and Jaeger (1977).Because mixed Rb–Sr spikes were used, we estimate that the87Rb/86Sr ratio for samples is precise to ±0.5% on the basis ofrepeated measurements of NIST SRM-984. The tourmaline analysisof sample 09CDM110 is estimated to have a larger uncertainty of0.7% because the sample was under-spiked for Rb. Isochron ageswere calculated using Isoplot version 3.72 (Ludwig, 2008), using thedecay constant (λ

87Rb) = 1.3971 × 10−11 yr−1 of Rotenburg et al.(2012), the above estimated uncertainties for Rb/Sr ratios and thein-run statistics for Sr isotope analyses. We note that this value

208 C.M.C. de Meyer et al. / Earth and Planetary Science Letters 395 (2014) 205–216

for the decay constant lies within error of other recent determina-tions by Kossert (2003) and Nebel et al. (2011), and use of thesealternative decay constants would not change the calculated ageswithin the analytical uncertainties of the current study.

3. Geological setting

The composite Zermatt-Saas Fee Zone (ZSZ) and the overlap-ping Tsaté nappe crop out in the Western Alps over an area of ca.1000 km2, along the border between Switzerland and Italy (Stecket al., 2000). These nappes are the remnants of the Piemonte-Ligurian oceanic basin (Fig. 1), which was the western-most partof the Tethyan Ocean. The latter separated the African continentfrom the European promontory (Stampfli et al., 2001). The ophi-olite sequence of the ZSZ consists of serpentinites, metagabbros,metabasalts, and a metasedimentary cover, containing quartzitesand calcschist (Bearth, 1962; Ernst and Dal Piaz, 1978; Barnicoatand Fry, 1986). The ZSZ tectonically underlies the Tsaté nappe(part of the Combin zone of Bearth, 1967). It contains oceaniccrust and sediments (Sartori, 1987; Steck, 1987; Marthaler andStampfli, 1989) and has been interpreted to represent remnantsof the accretionary prism (e.g. Marthaler and Stampfli, 1989;Reddy et al., 2003). The ZSZ is separated from the Tsaté nappeby rocks of probable Triassic and Jurassic age of unclear continen-tal origin, called the Cimes Blanches unit (Vannay and Allemann,1990), or Theodul-decollement nappe (Bucher et al., 2004). The ZSZwas subducted to eclogite-facies conditions with peak pressuresand temperatures up to 20–24 kbar and 500–600 ◦C, followed by agreenschist-facies overprint during exhumation (Oberhansli, 1982;Meyer, 1983; Ganguin, 1988; Angiboust et al., 2009) (see Fig. 5).The preservation of coesite inclusions (Reinecke, 1991) and evendiamond (Frezzotti et al., 2011) in garnets in at least one local-ity indicates that at least part of the ZSZ underwent ultra-high-pressure metamorphic conditions with pressures reaching 26–28kbar at temperatures of 590–630 ◦C. Some authors argue that theentire ZSZ underwent pressures higher than 25 kbar (Bucher etal., 2005). The overlying Tsaté nappe underwent medium- to high-pressure greenschist-facies metamorphism (Ernst and Dal Piaz,1978; Vannay and Allemann, 1990; Cartwright and Barnicoat,2002; Reddy et al., 2003).

The eclogites of the ZSZ have been the subject of detailed inves-tigations to determine the age of prograde and retrograde mineralsin order to constrain the metamorphic pressure–temperature–timepath. Fig. 2 gives an overview of the ages available in the litera-ture, obtained by various methods, for the prograde path of thenorthern part of the ZSZ. Previous geochronological studies of theZSZ used eclogite garnets. Ages obtained are between ∼40 Ma and∼50 Ma for Lu–Hf and Sm–Nd (see Fig. 2 and references therein).Comparison of Lu–Hf and Sm–Nd ages obtained from the samesample of the ZSZ (e.g., Lago di Cignana area) yields differencesof up to 10 Ma, depending upon the geochronological system used(Skora et al., 2009). Laser ablation-ICPMS and SIMS measurementsof REE zoning in garnet have led Skora et al. (2006) to proposethat a bulk garnet isochron age will be a function of the Lu andSm core-to-rim zonation in garnet. Skora et al. (2006) found thatactual measured profiles deviate significantly from those predictedby simple Rayleigh fractionation models (Lapen et al., 2003). Theyargued for diffusion-limited uptake of REE by the growing garnet.Lutetium–Hf garnet isochron ages are interpreted to date an av-erage age for the prograde path, roughly weighted towards theearliest 2/3 of the overall garnet growth time, whereas Sm–Ndgarnet isochron ages reflect the termination of the garnet growth(peak conditions), reflecting Sm enrichment in the rim. Notably,the Sm–Nd age is 4 m.y. (±2.6 m.y.) younger than the U–Pb zirconage from the same locality (Lago di Cignana area), where the zir-

con age had been previously interpreted by Rubatto et al. (1998)to date peak metamorphic conditions.

In addition to the range in ages that different geochronome-ters yield in the ZSZ, different locations within the ZSZ yieldeddifferent ages for a given geochronometer. Mahlen et al. (2006)report differences in Lu–Hf ages of up to 10 Ma when ZSZ sam-ples near Saas-Fee are compared to those obtained in Val Tour-nanche and Val St. Jacques. Collectively, these data indicate thatthe ZSZ had a protracted and diachroneous prograde metamorphicevolution. In contrast, ages that likely reflect exhumation or thepost-peak greenschist-facies overprint are rather uniform. Rb–Srgeochronology on white micas from the matrix of metasedimentsof the ZSZ resulted in ages clustering about ∼39 Ma (Hunzikeret al., 1992; Barnicoat et al., 1995; Amato et al., 1999; Reddyet al., 1999; Cartwright and Barnicoat, 2002; Lapen et al., 2003;Reddy et al., 2003; Szilvagyi, 2010). These ages are generally in-terpreted to record early exhumation, when the rocks cooled tobelow 500 ± 50 ◦C, or shearing under greenschist facies condi-tions. When corrected using the 87Rb decay constant of Rotenburget al. (2012), the ages obtained in these earlier studies increase byroughly 0.6 Ma.

4. Sample description

The samples of this study are metasediments of the ZSZ that areclosely associated with eclogites. These metapelitic rocks and calc-schists were deposited on the oceanic crust (Bearth and Schwan-der, 1981; Seydoux, 2013) or accreted, and underwent the samemetamorphic history and P-T conditions as the associated maficand ultramafic rocks. It is important to note that the eclogite-hosted metasediments studied here should not be confused withthe calcschists from the Tsaté unit, given that in the literature bothoccurrences of metasediments are often termed “Bündnerschiefer”.

Two samples were selected for analysis. Both samples,09CDM105 (Swiss coordinate system 624.358/089.527) and09CDM110 (624.590/089.757), were selected from an outcrop atTriftji, at the base of the Breithorn. The samples share a commonpressure–temperature–time (P–T–t) history, but show significantdifferences in mineral assemblage and texture, reflecting differentwhole-rock chemistry.

4.1. Petrography

Thin section images show typical textures for each sample(Fig. 3a). BSE-images of garnets with phengite inclusions are rep-resented in Figs. 3b and 4. The petrographic characteristics of thesamples are summarized below.

Sample 09CDM105 is a garnet-bearing, graphitic schist. It con-tains phengite, cloudy quartz, and poikilitic, millimetre-sized gar-net porphyroblasts. Minor phases are clinozoisite, calcite, parag-onite, sphene, apatite, graphite, and oxides. Many garnets arepoikilitic, rich in inclusions of rutile, quartz and phengite. Someinclusions of graphite and chloritoid, and rare inclusions of parago-nite, tourmaline and zircon are present. The garnets contain aggre-gates of clinozoisite, paragonite, phengite, and chlorite ± calcite,which are likely pseudomorphs after lawsonite (see Fig. 4d). Thematrix contains lozenge-shaped pseudomorphs after lawsonite;consisting of mainly phengite, quartz, and minor clinozoisite andcalcite. Some lawsonite pseudomorphs contain only paragonite andclinozoisite. The garnet is almandine-rich and part of the progradeassemblage Phg + Qtz + Grt + Rt (± Ap ± Gr). The presence ofregrowth after fracturing and compositional zoning of the garnetcould suggest that the outermost rim of the garnet grew duringretrograde metamorphism. Finally, some chlorite is present as aretrograde replacement of the garnet.

C.M.C. de Meyer et al. / Earth and Planetary Science Letters 395 (2014) 205–216 209

Fig. 3. (a) Representative sketch of a thin section of each sample studied; (b) back scattered electron image (BSE) of a garnet with phengite inclusions. It is likely that thebiotite in the lower figure (sample 09CDM110) formed retrograde in a crack of the garnet, partially at the expense of phengite. However the separation method (see –Section 2, Sample selection and analytical techniques) assured that mostly phengites not connected to the surface through cracks were selected for age dating. Ap = apatite;Bt = biotite; Cld = chloritoid; Phg = phengite; Qtz = quartz.

Sample 09CDM110 is a garnet-mica-bearing quartzite. Millime-tre- to cm-sized porphyroblasts of almandine-rich garnet are en-closed in a matrix of small-grained (∼100 μm diameter) cloudyquartz and up to one-mm-long phengite crystals. The matrix con-tains tourmaline, often included in the phengite. The phengitesdefine the main schistosity. Two texturally different types of gar-net can be distinguished: a poikilitic garnet with mainly quartzinclusions, and an inclusion poor, euhedral garnet. Note that theincluded phengites used for Rb–Sr dating were recovered fromthe inclusion-poor garnets. The inclusions in garnet are primar-ily quartz and ilmenite, with minor tourmaline, apatite and ru-tile, some phengite and rare biotite and chlorite. Tabular ilmeniteand barroisite are found from the core to the rim of the garnets,whereas rutile and chloritoid are included only towards the out-ermost rims. On the basis of this, garnet seems to have formedduring prograde metamorphism and is also part of the peak as-semblage: Qtz + Phg + Grt + Rt + Cld (± Ap ± Tur). The outer-most surfaces of the garnets and rims of phengites in the matrixare retrogressed to biotite and chlorite. Some retrograde feldspar(albite) and calcite (partially altered) are scattered throughout thematrix.

4.2. Chemistry of white micas

Fig. 4 shows the silica content of the phengites analyzed in thisstudy relative to the sum of the divalent octahedral cations. Inclu-sions of mica in garnet were classified on the basis of their radialposition within the garnet, and occurring inclusion assemblage,based on observations in thin sections. Representative phengitecompositions, obtained by EMPA, are given in Table 1.

There is a general trend visible. Inclusions have higher phen-gite contents towards the rim of the metapelitic garnets than inthe centre. It is well known that phengite content of white mica

increases with pressure (Massonne and Schreyer, 1987) for theassemblage white mica–k-feldspar–quartz–phlogopite. Below thewhite mica of each sample is discussed in detail.

Matrix white micas of sample 09CDM105 are mainly phengite.Paragonite only occurs in the pseudomorphs after lawsonite. Theinclusions were grouped with respect to the included assemblages(see Fig. 4a). An evolution towards higher phengite content can beobserved relative to their position towards the rim of the garnet.The lowest phengite content is observed for phengites occurringin pseudomorphs after lawsonite containing mainly clinozoisite orparagonite (group C, Fig. 4d). Phengites of group B (Fig. 4c) arezoned. The core of those phengites is less phengitic. They likelyrepresent partially equilibrated phengites during increasing meta-morphic conditions before entrapment in garnet. They are includedin the quartz inclusion-rich part of the garnet. Finally, small phen-gites included at, or outside of, the rim of the quartz-rich part ofgarnet have the highest phengite content (group A, Fig. 4b). Thephengite inclusions selected for Rb–Sr dating were recovered fromgroups A and B, it was possible to differentiate them under thebinocular microscope.

Phengites in the matrix of sample 09CDM110 are folded. Phen-gites included in garnets are commonly found in the rims of gar-nets (e.g. Fig. 4f) and plot towards higher phengite contents (seeFig. 4e), as expected for prograde growth of garnet and phen-gite. Matrix phengites scatter more in composition than inclusions.Recrystallized, retrograde white micas occurring together with bi-otite and chlorite, are less silica-rich compared to the other ma-trix phengites. The strong pressure dependence of the phengiticcomponent in white mica is illustrated in Fig. 5 for this sample.The phengitic composition of the white mica inclusions in gar-net indicates a pressure of 20 to 22 ± 2 kbar for a temperatureof 500 ± 50 ◦C. Phase petrology based on the garnet compositionand the phases included in garnet suggest increasing metamorphic

210 C.M.C. de Meyer et al. / Earth and Planetary Science Letters 395 (2014) 205–216

Fig. 4. Plot of phengite contents for mica included in garnets and matrix mica. (a), (b), (c) and (d) sample 09CDM105; (e) and (f) sample 09CDM110. Note that phengitecontent increases in mica inclusions towards the rim of the garnet, indicating increasing pressures as garnet grows. Cld = chloritoid; Czoi = clinozoisite; Pg = paragonite;Phg = phengite; Qtz = quartz; Rt = rutile.

conditions during garnet growth with pressure and temperatureestimates of 470 ± 50 ◦C and 14 ± 2 kbar for the garnet core, and500 ± 50 ◦C and 22 ± 2 kbar (see Fig. 5b) for the garnet rim.

5. Results of Rb–Sr geochronology

Two Rb–Sr isochrons are calculated for each sample, onefor phengite inclusions in garnet plus whole rock, defined asMicaINCL–WR, and the other isochrons for matrix phengites, othermatrix minerals plus whole rock, defined as MicaMATRIX-WR(Fig. 6). Because preparation of phengite mineral separates fromgarnet inclusions was extremely challenging – it requires largequantities of carefully selected garnets – only one phengite in-clusion analysis per sample was obtained. Tourmaline was omittedfrom the age calculations because they are strongly zoned in chem-ical compositions, in any event, the blocking temperature for Rb–Srin tourmaline is not well known. Addition of the tourmaline anal-ysis to the age calculation of sample 09CDM110 results in an ageof 39.26 ± 0.66 Ma with a high MSWD of 169. There was no at-tempt made to determine Rb–Sr ages using garnet, because thevery low Sr content of garnet would make any garnet mineral sep-arate highly susceptible to the effects of inclusions, and sufficientlypure garnet could not be obtained.

The blank-subtracted ratios were used for the age calculationsof the MicaINCL–WR. The effect of Rb and Sr blank subtraction isgreatest on phengite inclusions from sample 09CDM110, where useof the measured data produces an age of 43.07 ± 0.32 Ma, and useof blank-subtracted Rb and Sr contents and 87Sr/86Sr ratios pro-duce an age of 43.6 ± 1.8 Ma. For sample 09CDM105 use of themeasured data produces an age of 44.13 ± 0.48 Ma, and use ofblank-subtracted Rb and Sr contents and 87Sr/86Sr ratios producean age of 44.86 ± 0.49 Ma.

The MicaINCL–WR ages of samples 09CDM105 (44.86±0.49 Ma)and 09CDM110 (43.6 ± 1.8 Ma) are respectively 4.85 ± 1.00 and4.1±2.9 m.y. older than the corresponding MicaMATRIX–WR ages of40.01 ± 0.51 Ma and 39.5 ± 1.1 Ma, respectively. The MicaINCL–WRages reflect 2-point isochrons, and thus age errors only reflect an-alytical uncertainties. The relatively high Rb/Sr ratios of the phen-gite inclusions provide great leverage to the isochrons, permittinghigh precision in the Rb–Sr age, within the limits of the 2-pointisochrons approach.

6. Discussion

The goal of this study was to determine white mica crystal-lization ages that yield prograde ages, despite having maybe been

C.M.C. de Meyer et al. / Earth and Planetary Science Letters 395 (2014) 205–216 211

Table 1Representative EMPA analyses of matrix and included phengites.

Sample 09CDM105 Sample 09CDM110

Matrix Matrix Group A Group B Group C Matrix Matrix Inclusion Inclusion

SiO2 49.6 51.8 51.6 48.4 49.6 50.8 51.9 52.7 52.3Al2O3 28.7 26.2 26.0 32.3 30.0 26.4 23.2 25.3 24.8TiO2 0.32 0.16 0.08 0.18 0.26 0.32 0.22 0.19 0.18FeO 2.16 3.09 3.59 1.94 2.32 4.03 4.73 3.76 4.32MnO 0.00 0.03 0.00 0.00 0.01 0.00 0.00 0.00 0.00MgO 3.04 3.90 3.76 1.73 2.29 3.60 4.31 3.68 3.88CaO 0.04 0.01 0.03 0.00 0.04 0.01 0.00 0.04 0.04Na2O 0.30 0.00 0.01 0.71 0.54 0.22 0.00 0.22 0.24K2O 9.50 10.2 9.95 9.87 9.83 10.2 10.8 9.23 9.49Total 93.7 95.4 95.0 95.1 94.9 95.6 95.2 95.1 95.3

Si 3.31 3.41 3.41 3.20 3.29 3.35 3.47 3.46 3.45Al IV 0.69 0.59 0.59 0.80 0.71 0.65 0.53 0.54 0.55Al VI 1.56 1.44 1.43 1.71 1.62 1.40 1.30 1.42 1.37Ti 0.02 0.01 0.00 0.01 0.01 0.02 0.01 0.01 0.01Fe3+ 0.12 0.15 0.16 0.08 0.08 0.22 0.23 0.10 0.17Fe2+ 0.00 0.02 0.04 0.03 0.05 0.00 0.04 0.10 0.06Mg 0.30 0.38 0.37 0.17 0.23 0.35 0.43 0.36 0.38Mn 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00Ca 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00Na 0.04 0.00 0.00 0.09 0.07 0.03 0.00 0.03 0.03K 0.81 0.85 0.84 0.83 0.83 0.86 0.92 0.77 0.80Total 6.85 6.85 6.84 6.92 6.89 6.88 6.93 6.79 6.82

Mg + Fe2+ 0.30 0.40 0.41 0.20 0.28 0.35 0.47 0.46 0.44Mg/Mg + Fe2+ 1.00 0.95 0.90 0.85 0.82 1.00 0.91 0.78 0.86Na/Na + K 0.05 0.00 0.00 0.10 0.08 0.03 0.00 0.04 0.04

Sample 09CDM105: Groups A, B and C refer to the groups defined in Fig. 4.

above the Rb–Sr closure temperature. The ages retained their fi-delity to the matrix or bulk sample. Below we first discuss thecontrast in white mica Rb–Sr ages for inclusions relative to thoseof the matrix in the context of the P-T-t history of the rocks, fol-lowed by a discussion of the implications for the tectonic evolutionof the ZSZ.

6.1. Shielding of phengite by garnet

The closure temperature for the Rb–Sr system in white micashas been estimated at 500 ± 50 ◦C, on the basis of resetting ofthe Rb–Sr isotope systematics in muscovites from the Central Alps(Jaeger et al., 1967; Jaeger, 1973). Cooling rates and the size ofmicas will affect the closure temperature as defined by Dodson(1973). However, the Dodson closure temperature for the Rb–Srsystem can be highly underestimated, in analogy to experimentsfor the Ar/Ar system in muscovites (Harrison et al., 2009), re-sulting in closure at significantly higher temperatures, see alsoVilla (1998). For example, slow movement of elements along grainboundaries (e.g. Skora et al., 2006) will limit diffusion towardsthe surroundings of the mineral independent of the temperature,or the matrix may not absorb the elements released, depend-ing on element distribution coefficients. These effects can resultin higher effective closure temperatures, which depend on themodal composition of the rock and the presence or absence offluids (Eiler et al., 1992; Jenkin et al., 1995; Kuhn et al., 2000;Glodny et al., 2003; Meffan-Main et al., 2004; Di Vicenzo et al.,2006; Glodny et al., 2008a, 2008b). Dynamic recrystallization canled to re-opening of the system, even at temperatures lower thanthe closure temperature (Freeman et al., 1997).

Garnet is Sr poor, containing typically less than 1 ppm Sr(Jenkin et al., 1995), and garnet has very low Rb/Sr ratios, henceproviding little leverage on Rb–Sr isochrons. The Sr diffusion ingarnet is slow (e.g. Burton and O’Nions, 1991, 1992; Burton etal., 1995; Baxter and De Paolo, 2000; Sousa et al., 2013), andBurton et al. (1995) concluded that the closure temperature forRb–Sr in garnet should be at least equal to, or higher than, the

closure temperature for Sm–Nd in garnet on the basis of the simi-larity in ages obtained with both isotopic systems. In the ZSZ, theSm–Nd closure temperature is above the peak metamorphic condi-tions (Skora et al., 2006), indicating that garnet should have beenclosed during peak metamorphic conditions of the ZSZ for Rb–Sr.It therefore seems likely that once white mica became includedinto garnet during prograde garnet growth in the ZSZ, the whitemica inclusions would have evolved as a closed system even ifpeak metamorphic conditions of the ZSZ exceeded the Rb–Sr clo-sure temperature for white mica, or if recrystallization occurreddue to fluid infiltration or deformation.

White mica inclusions are scattered throughout the garnets inthe studied samples, and hence different white mica grains be-came completely enclosed in garnet at different times. The effectof continuously including new mica grains during garnet growth isillustrated in Fig. 7. We assume that phengite begins to be includedinto garnet at a temperature higher than its closure temperature,but below the closure temperature for Sr in garnet. Hence, Sr iso-topes in partially enclosed micas will continue to exchange withthe matrix. Assuming Sr isotopic exchange continues to occur be-tween matrix and phengite inclusions prior to complete isolationinside garnet, the 87Sr/86Sr ratio will increase according to the87Rb/86Sr ratio of the bulk rock. As the first white mica grainsare totally enclosed in garnet (Fig. 7), the 87Sr/86Sr ratios of thesegrains will increase faster, reflecting the high Rb/Sr ratio of thewhite mica (inclusion 1 line segment in Fig. 7). As additional whitemicas are enclosed in garnet (e.g. inclusion 2 in Fig. 7), thesewill have initial 87Sr/86Sr ratios that reflect radiogenic Sr growththat is controlled by the bulk rock 87Rb/86Sr. Although newly in-cluded phengites may have the same 87Rb/86Sr ratio of the earliestincluded white micas, at the moment of inclusion the newly in-cluded phengite will have a lower 87Sr/86Sr ratio than the earliestincluded phengites. This lower 87Sr/86Sr ratio reflects isotopic equi-librium with the bulk rock before being completely captured bythe garnet. For simplicity, all of the included phengites are mod-elled with identical Rb/Sr ratios so the pooled composition wouldevolve on a growth curve that is parallel to those of the individ-

212 C.M.C. de Meyer et al. / Earth and Planetary Science Letters 395 (2014) 205–216

Tabl

e2

Rb–S

rre

sult

s.

Sam

ple

nam

eM

iner

alse

para

teSa

mpl

em

ass

(mg)

Rb (ppm

)Sr (p

pm)

87Rb

/86Sr

2-s

(%)

87Sr

/86Sr

2-SE

(%)

Sr (ng)

Blan

kRb

(pg)

Blan

kSr

(pg)

Sam

ple/

Blan

kRb

Sam

ple/

Blan

kSr

09CD

M10

5W

hole

rock

57.3

209

78.0

17.

756

0.5

0.71

401

0.00

1544

7×

1043

422

727

6×

102

197

×10

2

Phg

180–

300

7.7

391

51.7

321

.90.

50.

7220

40.

0018

398

17.5

51.5

172

×10

377

3×

10Ph

g30

0–40

013

.530

529

.79

29.6

80.

50.

7262

40.

0013

402

17.5

51.5

235

×10

378

1×

10Ph

g40

0–50

09 .

341

532

.59

36.9

20.

50.

7303

10.

0014

303

17.5

51.5

221

×10

358

9×

10Zo

8 .5

6.03

15.5

21.

124

0.5

0.71

032

0.00

1613

217

.382

.029

6×

102

161

×10

Phg

inG

rt2 .

194

.716

.94

16.1

90.

50.

7192

10.

0013

35.6

19.3

20.5

103

×10

217

4×

10Bl

ank-

subt

ract

edPh

gin

Grt

16.0

80.

50.

7192

30.

0016

09CD

M11

0W

hole

rock

51.4

126

21.6

116

.89

0.5

0.71

852

0.00

1411

1×

1043

422

714

9×

102

489

×10

Phg

300–

400

8 .3

415

9.17

131.

90.

50.

7815

50.

0021

76.1

17.5

51.5

197

×10

314

8×

10Ph

g40

0–50

011

.132

66.

492

146.

50.

50.

7904

20.

0016

72.1

17.5

51.5

207

×10

314

0×

10Ph

g50

0–60

08 .

438

17.

211

154.

20.

50.

7941

0.00

2560

.617

.551

.518

3×

103

118

×10

Tur

111 .

842

5.5

0.01

224

0.7

0.70

990.

0013

468

×10

17.3

82.0

114

×10

571

×10

2

Phg

inG

rt0.

211

65.

877

57.3

0.5

0.74

284

0.00

581.

1813

.962

.516

7×

1018

.8Bl

ank-

subt

ract

edPh

gin

Grt

59.5

1.7

0.74

447

0.11

76

Fig. 5. (a) Phengitic component of muscovite in a PT-diagram for sample 09CDM110.The whole rock composition (in weight%: 74.94 SiO2; 9.76 Al2O3; 0.40 TiO2; 3.26Fe2O3; 2.84 FeO; 0.06 MnO; 1.65 MgO; 1.78 CaO; 0.00 NaO; 2.71 K2O; 0.06 P2O5)modelled was first corrected for apatite, and all Fe3+ considered as Al3+ . Thepseudosection was calculated with Perple_X (Connolly, 1990), using the databasehpver04 (Holland and Powell, 1998) and the following solution models: TiBio(HP)(Powell and Holland, 1999); Ctd(HP) (White et al., 2000); GlTrTsPg (White et al.,2003); Omph(HP) (Holland and Powell, 1996); feldspar (Fuhrman and Lindsley,1988); Chl(HP), Ep(HP), Gt(HP), Pheng (HP) (Holland and Powell, 1998); and IlHm(A)and MtUl(A) (Andersen and Lindsley, 1988). Fluid saturation of H2O was assumed.(b) PT-estimates during garnet growth for sample 09CDM110 (shaded in grey). Thearrow indicates the PT-path for the ZSZ as proposed by Angiboust et al. (2009).

ual enclosed phengites. The final composition of the bulk includedphengite would extrapolate to an initial 87Sr/86Sr ratio, and anaverage age, that is intermediate between the first and the lastinclusion, determined by mass balance of the inclusions.

In contrast to the Rb–Sr isotope evolution of the white mica in-cluded in garnet, isotopic evolution of the white mica in the matrixwill be dependent upon the time that post-peak metamorphismattained the Rb–Sr closure temperature upon cooling (matrix linein Fig. 7). Rubidium-Sr ages for matrix white mica may record atime that is significantly lower than the time that cooling passedthrough the Rb–Sr closure temperature if further mineral reactionsor recrystallization occurred, which would effectively maintain anopen system for Rb–Sr (e.g. Freeman et al., 1997; see also above).The initial 87Sr/86Sr ratios of the matrix will control the startingpoints for Rb–Sr isotope evolution of the matrix phengites as longas the samples remained below the Rb–Sr closure temperature.Relatively low-Rb/Sr, high-Sr phases such as tourmaline, apatite,± carbonates, ± epidote group minerals are the likely mineralsto have controlled the initial Sr isotope composition of the matrixphengites in the samples of this study.

The changing mineral assemblage during garnet growth, andhence enclosure of phengite, did not substantially influence the

C.M.C. de Meyer et al. / Earth and Planetary Science Letters 395 (2014) 205–216 213

Fig. 6. Rb–Sr geochronology results. Figures (a) and (b) are calculated isochrons ofresp. sample 09CDM105 and 09CDM110, along with the data. The isochron obtainedby combining white micas included in garnets with the whole rock is presentedin the upper left of each graph, whereas the right lower half shows the isochroncalculated using the phengite fractions of the matrix and the high-Sr/low-Rb ma-trix mineral(s). Minerals within parentheses were analyzed but not included forthe age calculation. (c) Overview of the isochron ages obtained. The weighted micainclusions – whole rock isochrons are 2-point fits and thus reflect the analytical un-certainties only. Calculated and plotted with the Isoplot program of Ludwig (2008),using Model 3 fits and the λ

87 Rb-value of Rotenburg et al. (2012).

Rb and Sr balance of the whole rock. The main inclusions of thegarnets are quartz ± rutile ± ilmenite. Those minerals as wellas garnet do not contain considerable amounts of Sr and Rb. Theother phases included in the garnet (see – Section 4.1, Petrography)are only in minor quantities included relative to their occurrencein the matrix. So although some – calcite, tourmaline and apatite –do contain considerable amounts of Sr, inclusion of those mineralsin garnet during the prograde metamorphism will not affect the Rband Sr balance of the whole rock. Sr-rich minerals in the matrix ofsample 09CDM110 are tourmaline, carbonates, apatite and albite.In sample 09CDM105 the Sr-rich minerals are epidote-group min-erals, apatite and calcite. The retention of Sr in those minerals ispoorly known and it is likely that the effective whole rock assem-blage with whom the included phengites exchanged for Sr was notthe same as for the matrix phengites, in analogy to the discussionin Sousa et al. (2013). However, because of the high Rb/Sr ratio ofthe analyzed phengites the effect on both ages will be small.

From the relations illustrated in Fig. 7, we can conclude thatthe age obtained from Rb–Sr isochron geochronology on mica in-clusions in garnet, tied to a bulk rock analysis, will be a volumet-rically averaged age that is dependent on the growth mechanismand timing of garnet growth. Different end member growth mech-anisms have been proposed for garnet growth (e.g. Kretz, 1974;Carlson, 1989): diffusion or surface kinetics controlled or heat flowcontrolled. Each mechanism leads to a different garnet volume-time relationship. Hence, the garnet growth mechanisms will havean effect on the Rb–Sr ages obtained from the white mica inclu-sions, as does garnet growth affect Sm–Nd and Lu–Hf geochronol-ogy (see detailed discussion by Skora et al., 2009).

Multiple generations of phengites are observed petrographicallyand geochemically (see Fig. 4) in the matrix. Differences in equi-libration between the fractions may exist upon cooling, due totheir size or to reaction during deformation, resulting in the rel-atively high scatter of the analyzed phengites, and hence in thehigh MSWD.

6.2. Interpretation of the obtained ages

The weighted MicaMATRIX–WR matrix age of 39.92 ± 0.45 Mafor the two samples (see Fig. 6c) is consistent with earlier stud-ies on the ZSZ, which, when using the new 87Rb decay constantof Rotenburg et al. (2012) produces an average age of 39.6 Ma, asnoted above, relative to an age of 39.0 Ma when using the Rb-decay constant of Steiger and Jaeger (1977). Combined, our results,and the recalculated MicaMATRIX–WR ages of previous studies ofthe ZSZ (Hunziker et al., 1992; Amato et al., 1999; Barnicoat et al.,1995; Cartwright and Barnicoat, 2002; Lapen et al., 2003; Szilvagyi,2010), indicate cooling below the closure temperature of 500 ±50 ◦C by ca. 40 Ma. That similar MicaMATRIX–WR ages are foundthroughout the ZSZ, at various structural levels, is significant. Itdemonstrates that early exhumation of the entire ZSZ occurredconcomitantly. The similarity in MicaMATRIX–WR ages to the Sm–Nd age of 40.6 ± 2.6 Ma that Amato et al. (1999) and Skora et al.(2009) interpret to record peak metamorphism, however poses aninteresting puzzle. These ages imply very rapid exhumation fromUHP conditions, although the relatively large error of the Sm–Nd

214 C.M.C. de Meyer et al. / Earth and Planetary Science Letters 395 (2014) 205–216

Fig. 7. Schematic sketch of the evolution of Sr-isotopes in some model phengiteinclusions in garnet, assuming closure with respect to the Rb–Sr system upon com-plete inclusion in garnet. A closed, low Rb/Sr system with garnet, phengite andthe whole rock, which functions as an infinite Sr reservoir, is considered. See textfor discussion. A garnet enclosing phengite during growth is sketched. Phengites inbrighter gray scale values correspond to phengites earlier closed in time for Rb–Srisotopic resetting.

age must be considered. It is likely that the MicaMATRIX-WR agesdo reflect recrystallization during fluid infiltration. Significant fluidflow occurred during breakdown of eclogite- and blueschist-faciesassemblages to barroisite-bearing simplectites in mafic rocks. Inplaces this led to a complete greenschist facies recrystallizationunder relatively high-P greenschist facies conditions (Bearth, 1973;Bucher et al., 2005; Cartwright and Barnicoat, 2003; Ernst and DalPiaz, 1978). It is noteworthy that Cartwright and Barnicoat (2002)and Reddy et al. (1999, 2003) report Rb–Sr MicaMATRIX–WR defor-mation ages in the ZSZ that become younger towards shear zonesthat were active during exhumation.

The MicaINCL–WR ages obtained from the Triftji locality are∼4 Ma older than all MicaMATRIX–WR ages from the ZSZ. Theseages are skewed towards the end of the garnet growth, but areolder than peak metamorphism, as the garnet which included thephengites grew during prograde metamorphism. The 1 Ma differ-ence between the two samples from Triftji is easily explained bythe differences in inclusion patterns. There are more inclusions to-wards the rim of garnets in sample 09CDM110 than in sample09CDM105, and garnet was stable at lower P and T conditionsfor sample 09CDM105, and hence could have started growth atan earlier time. This indicates that the mica inclusions should beexpected to record different time points along the prograde garnet-growth path.

The Rb–Sr geochronology on the micas included in garnets ofthe ZSZ can be compared with the Ar/Ar ages of 43.2 ± 1.1 Maand 44.4 ± 1.5 Ma on phengite inclusions from the Lago di Cig-nana (Gouzu et al., 2006). Those authors suggest that the Ar/Arages are too old, interpreting this to reflect excess argon. Instead,we suggest that the relatively old Ar/Ar ages may not reflect excessAr in the sense of inheritance from old radiogenic crustal terranes,but instead reflects the effective closure to Ar loss upon inclusionin garnet. Hence, this “excess argon” actually dates the true age oftotal inclusion of the mica along the prograde garnet growth path.This is entirely consistent with the Rb–Sr ages measured here formica inclusions from the Triftji locality. Hence, if all garnet grewduring prograde metamorphism, the peak was likely to be attainedafter 44 Ma at the Lago di Cignana locality, as proposed by Skoraet al. (2009), as inferred from Sm–Nd and Lu–Hf geochronology.

7. Conclusions

Geochronology of minerals encased in garnet is an effective ap-proach to constrain the age of mineral inclusion and hence a timepoint on prograde garnet growth and metamorphic P-T-t paths. Inthe absence of in-situ measurement techniques that are sufficientlyaccurate to decode garnet growth history in detail, we present herea bulk crystal method to estimate timing for garnet growth. Theresults of Rb–Sr geochronology on phengite inclusions in intactgarnet from the Zermatt-Saas Fee Zone (ZSZ) demonstrates thatinclusions are efficiently isolated from the matrix once includedentirely in the garnet. The age differences of up to 4 m.y. betweenmatrix and included phengites is best explained by protracted gar-net growth during prograde metamorphism, where peak metamor-phic HP conditions were attained before 40 Ma, but after 43 Ma.Our findings are consistent with previous Ar/Ar, Sm–Nd, and Lu–Hf studies on the ZSZ, where Sm–Nd and Lu–Hf geochronologysuggests 10s of m.y. for prograde garnet growth, and previouslyinterpreted “excess Ar” may in fact reflect mica formation agesduring prograde metamorphism. The uniformity of matrix phen-gite ages of ∼40 Ma across the ZSZ, from this and previous studies,indicates rapid exhumation to upper greenschist-facies conditionsafter peak HP metamorphism.

Acknowledgements

The authors would like to acknowledge the insightful commentsof the reviewers Bob Cliff, Erik Scherer, and an anonymous re-viewer. Finally we thank the editor for the handling of this paper.Their comments significantly improved the quality of the paper.LPB and CDM would like to thank the Swiss National Science Foun-dation (Grant No. 200020-125261) for its support. CDM would liketo thank Benita Putlitz for her useful advice during sample prepa-ration.

Appendix A. Supplementary material

Supplementary material related to this article can be found on-line at http://dx.doi.org/10.1016/j.epsl.2014.03.050.

References

Amato, J.M., Johnson, C.M., Baumgartner, L.P., Beard, B.L., 1999. Rapid exhumationof the Zermatt-Saas ophiolite deduced from high-precision Sm–Nd and Rb–Srgeochronology. Earth Planet. Sci. Lett. 171, 425–438.

Andersen, D.J., Lindsley, D.H., 1988. Internally consistent solution models for Fe–Mg–Mn–Ti Oxides – Fe–Ti oxides. Am. Mineral. 73, 714–726.

Angiboust, S., Agard, P., Jolivet, L., Beyssac, O., 2009. The Zermatt-Saas ophiolite: thelargest (60-km wide) and deepest (c. 70–80 km) continuous slice of oceaniclithosphere detached from a subduction zone?. Terra Nova 21, 171–180.

Barnicoat, A.C., Fry, N., 1986. High-pressure metamorphism of the Zermatt-Saasophiolite zone, Switzerland. J. Geol. Soc. 143, 607–618.

C.M.C. de Meyer et al. / Earth and Planetary Science Letters 395 (2014) 205–216 215

Barnicoat, A.C., Rex, D.C., Guise, P.G., Cliff, R.A., 1995. The timing of and nature ofgreenschist facies deformation and metamorphism in the Upper Pennine Alps.Tectonics 14, 279–293.

Baxter, E.F., DePaolo, D.J., 2000. Field measurement of slow metamorphic reactionrates at temperatures of 500 degrees to 600 degrees C. Science 288, 1411–1414.

Baxter, E.F., Scherer, E.E., 2013. Garnet geochronology: timekeeper of tectonometa-morphic processes. Elements 9 (6), 433–438.

Beard, B.L., Ludois, J.M., Lapen, T.J., Johnson, C.M., 2013. Pre-4.0 billion year weather-ing on Mars constrained by Rb–Sr geochronology on meteorite ALH84001. EarthPlanet. Sci. Lett. 361, 173–182.

Bearth, P., 1962. Versuch einer Gliederung alpinmetamorpher Serien der Westalpen.Schweiz. Mineral. Petrogr. Mitt. 42, 127–137.

Bearth, P., 1967. Die Ophiolite der Zone von Zermatt-Saas Fee. Beitrage geolischeKarte Schweiz, vol. 132.

Bearth, P., 1973. Gesteins-und mineralparagenesen aus den Ophioliten von Zermatt.Schweiz. Mineral. Petrogr. Mitt. 44, 15–26.

Bearth, P., Schwander, H., 1981. The post-Triassic sediments of the ophiolite zoneZermatt-Saas Fee and the associated manganese mineralizations. Eclogae Geol.Helv. 74, 189–205.

Bowtell, S.A., Cliff, R.A., Barnicoat, A.C., 1994. Sm–Nd isotopic evidence on the age ofeclogitization in the Zermatt-Saas ophiolite. J. Metamorph. Geol. 12, 187–196.

Bucher, K., Dal Piaz, G.V., Oberhänsli, R., Gouffon, Y., Martinotti, G., Polino, R., 2004.Blatt 1347 Matterhorn – Geologischer Atlas der Schweiz, 1:25.000, Erläuterun-gen 107.

Bucher, K., Yvonne, F., De Capitani, C., Grapes, R., 2005. Blueschists, eclogites, anddecompression assemblages of the Zermatt-Saas ophiolite: high-pressure meta-morphism of subducted Tethys lithosphere. Am. Mineral. 90, 821–835.

Burton, K.W., Onions, R.K., 1991. High-resolution garnet chronometry and the ratesof metamorphic processes. Earth Planet. Sci. Lett. 107, 649–671.

Burton, K.W., Onions, R.K., 1992. The timing of mineral growth across a regionalmetamorphic sequence. Nature 357, 235–238.

Burton, K.W., Kohn, M.J., Cohen, A.S., Onions, R.K., 1995. The relative diffusion of Pb,Nd, Sr and O in Garnet. Earth Planet. Sci. Lett. 133, 199–211.

Carlson, W.D., 1989. The significance of intergranular diffusion to the mechanismsand kinetics of porphyroblast crystallization. Contrib. Mineral. Petrol. 103, 1–24.

Cartwright, I., Barnicoat, A.C., 2002. Petrology, geochronology, and tectonics of shearzones in the Zermatt-Saas and Combin zones of the Western Alps. J. Meta-morph. Geol. 20, 263–281.

Cartwright, I., Barnicoat, A.C., 2003. Geochemical and stable isotope resetting inshear zones from Taschalp: constraints on fluid flow during exhumation in theWestern Alps. J. Metamorph. Geol. 21, 143–161.

Charlier, B.L.A., Ginibre, C., Morgan, D., Nowell, G.M., Pearson, D.G., Davidson, J.P.,Ottley, C.J., 2006. Methods for the microsampling and high-precision analysisof strontium and rubidium isotopes at single crystal scale for petrological andgeochronological applications. Chem. Geol. 232, 114–133.

Christensen, J.N., Rosenfeld, J.L., Depaolo, D.J., 1989. Rates of tectonometamor-phic processes from rubidium and strontium isotopes in Garnet. Science 244,1465–1469.

Christensen, J.N., Selverstone, J., Rosenfeld, J.L., Depaolo, D.J., 1994. Correlation byRb–Sr geochronology of garnet growth histories from different structural levelswithin the Tauern Window, Eastern Alps. Contrib. Mineral. Petrol. 118, 1–12.

Connolly, J.A.D., 1990. Multivariable phase-diagrams – an algorithm based on gener-alized thermodynamics. Am. J. Sci. 290, 666–718.

Di Vincenzo, G., Tonarini, S., Lombardo, B., Castelli, D., Ottolini, L., 2006. Compari-son of Ar-40–Ar-39 and Rb–Sr data on phengites from the UHP Brossasco-Isascaunit (Dora Maira Massif, Italy): implications for dating white mica. J. Petrol. 47,1439–1465.

Dodson, M.H., 1973. Closure temperature in cooling geochronological and petrolog-ical systems. Contrib. Mineral. Petrol. 40, 259–274.

Dragovic, B., Samanta, L.M., Baxter, E.F., Selverstone, J., 2012. Using garnet to con-strain the duration and rate of water-releasing metamorphic reactions duringsubduction: an example from Sifnos, Greece. Chem. Geol. 314, 9–22.

Eiler, J.M., Baumgartner, L.P., Valley, J.W., 1992. Intercrystalline stable isotope diffu-sion – a fast grain-boundary model. Contrib. Mineral. Petrol. 112, 543–557.

Ernst, W.G., Dalpiaz, G.V., 1978. Mineral parageneses of eclogitic rocks and relatedmafic schists of Piemonte ophiolite nappe, Breuil-St-Jacques Area, Italian West-ern Alps. Am. Mineral. 63, 621–640.

Freeman, S.R., Inger, S., Butler, R.W.H., Cliff, R.A., 1997. Dating deformation using Rb–Sr in white mica: greenschist facies deformation ages from the Entrelor shearzone, Italian Alps. Tectonics 16, 57–76.

Frezzotti, M.L., Selverstone, J., Sharp, Z.D., Compagnoni, R., 2011. Carbonate dissolu-tion during subduction revealed by diamond bearing rocks from the Alps. Nat.Geosci. 4, 703–706.

Fuhrman, M.L., Lindsley, D.H., 1988. Ternary-feldspar modeling and thermometry.Am. Mineral. 73, 201–215.

Ganguin, J., 1988. Contribution à la caractérisation du métamorphisme polyphase dela zone de Zermatt-Saas Fee (Alpes Valaisannes). ETHZ, Zurich, p. 311.

Glodny, J., Austrheim, H., Molina, J.F., Rusin, A.I., Seward, D., 2003. Rb/Sr record offluid-rock interaction in eclogites: the Marun-Keu complex, Polar Urals, Russia.Geochim. Cosmochim. Acta 67, 4353–4371.

Glodny, J., Kuhn, A., Austrheim, H., 2008a. Geochronology of fluid-induced eclogiteand amphibolite facies metamorphic reactions in a subduction-collision system,Bergen Arcs, Norway. Contrib. Mineral. Petrol. 156, 27–48.

Glodny, J., Kuhn, A., Austrheim, H., 2008b. Diffusion versus recrystallization pro-cesses in Rb–Sr geochronology: isotopic relics in eclogite facies rocks, WesternGneiss region, Norway. Geochim. Cosmochim. Acta 72, 506–525.

Gouzu, C., Itaya, T., Hyodo, H., Matsuda, T., 2006. Excess Ar-40-free phengite inultrahigh-pressure metamorphic rocks from the Lago di Cignana area, WesternAlps. Lithos 92, 418–430.

Harrison, T.M., Celerier, J., Aikman, A.B., Hermann, J., Heizler, M.T., 2009. Diffusion ofAr-40 in muscovite. Geochim. Cosmochim. Acta 73, 1039–1051.

Harvey, J., Baxter, E.F., 2009. An improved method for TIMS high precisionneodymium isotope analysis of very small aliquots (1–10 ng). Chem. Geol. 258,251–257.

Herwartz, D., Munker, C., Scherer, E.E., Nagel, T.J., Pleuger, J., Froitzheim, N., 2008.Lu–Hf garnet geochronology of eclogites from the Balma Unit (Pennine Alps):implications for Alpine paleotectonic reconstructions. Swiss J. Geosci. 101,S173–S189.

Holland, T., Powell, R., 1996. Thermodynamics of order-disorder in minerals. 2. Sym-metric formalism applied to solid solutions. Am. Mineral. 81, 1425–1437.

Holland, T.J.B., Powell, R., 1998. An internally consistent thermodynamic data set forphases of petrological interest. J. Metamorph. Geol. 16, 309–343.

Hunziker, J., Desmons, J., Hurford, A.J., 1992. Thirty-two years of geochronologicalwork in the Central and Western Alps: a review on seven maps. Mém. Géol.Lausanne 13, 1–59.

Jaeger, E., 1973. Die alpine Orogenese im Lichte der radiometrischen Altersbestim-mung. Eclogae Geol. Helv. 66, 11–21.

Jaeger, E., Niggli, E., Wenk, E., 1967. Rb–Sr Alterbestimmungen an Glimmern derZentralalpen. Beitrage zur Geologischen Karte der Schweiz, vol. 134.

Jenkin, G.R.T., Rogers, G., Fallick, A.E., Farrow, C.M., 1995. Rb–Sr closure temperaturesin bi-mineralic rocks – a mode effect and test for different diffusion-models.Chem. Geol. 122, 227–240.

Kohn, M.J., 2009. Models of garnet differential geochronology. Geochim. Cosmochim.Acta 73, 170–182.

Kossert, K., 2003. Half-life measurements of 87Rb by liquid scintillation counting.Appl. Radiat. Isot. 59, 377–382.

Kretz, R., 1974. Some models for the rate of crystallization of garnet in metamorphicrocks. Lithos 7, 123–131.

Kuhn, A., Glodny, J., Iden, K., Austrheim, H., 2000. Retention of Precambrian Rb/Srphlogopite ages through Caledonian eclogite facies metamorphism, Bergen ArcComplex, W-Norway. Lithos 51, 305–330.

Lapen, T.J., Johnson, C.M., Baumgartner, L.P., Mahlen, N.J., Beard, B.L., Amato, J.M.,2003. Burial rates during prograde metamorphism of an ultra-high-pressure ter-rane: an example from Lago di Cignana, western Alps, Italy. Earth Planet. Sci.Lett. 215, 57–72.

Lapen, T.J., Johnson, C.M., Baumgartner, L.P., Dal Piaz, G.V., Skora, S., Beard, B.L.,2007. Coupling of oceanic and continental crust during Eocene eclogite-faciesmetamorphism: evidence from the Monte Rosa nappe, western Alps. Contrib.Mineral. Petrol. 153, 139–157.

Liati, A., Froitzheim, N., 2006. Assessing the Valais ocean, Western Alps:U-PbSHRIMP zircon geochronology of eclogite in the Balma unit, on top of theMonte Rosa nappe. Eur. J. Mineral. 18, 299–308.

Ludwig, K., 2008. Isoplot version 3.7; BGC’s Visual Basic Add-in for Microsoft’s Ex-cel®.

Mahlen, N.J., Skora, S., Johnson, C.M., Baumgartner, L.P., Lapen, T.J., Beard, B.L., Pilet,S., 2005. Lu–Hf geochronology of eclogites from Pfulwe, Zermatt-Saas ophiolite,western Alps, Switzerland. Geochim. Cosmochim. Acta 69, A305.

Mahlen, N.J., Johnson, C.M., Baumgartner, L.P., Lapen, T.J., Skora, S., Beard, B.L., 2006.The protracted subduction history and HP/UHP metamorphism of the Zermatt-Saas ophiolite, western Alps, as constrained by Lu–Hf geochronology. In: AGUFall Meeting, Supplement, p. V41E-05.

Marthaler, M., Stampfli, G.M., 1989. Lustrous schists with ophiolites of the tsate-nappe – an early accretionary prism from the apulian active margin, Schweiz.Mineral. Petrogr. Mitt. 69 (2), 211–216.

Massonne, H.J., Schreyer, W., 1987. Phengite geobarometry based on the limitingassemblage with K-Feldspar, Phlogopite, and Quartz. Contrib. Mineral. Petrol. 96,212–224.

Mayer, A., Abouchami, W., Dal Piaz, G.V., 1999. Eocene Sm–Nd age for the eclogiticmetamorphism of the Zermatt-Saas ophiolite in the Ayas Valley, Western Alps.In: European Union of Geosciences Meeting, p. A809.

Meffan-Main, S., Cliff, R.A., Barnicoat, A.C., Lombardo, B., Compagnoni, R., 2004.A tertiary age for Alpine high-pressure metamorphism in the Gran Paradisomassif, Western Alps: a Rb–Sr microsampling study. J. Metamorph. Geol. 22,267–281.

Meyer, J., 1983. Mineralogie und Petrologie des Allalinsgabbros. Universität Basel,Basel.

Montel, J.M., Kornprobst, J., Vielzeuf, D., 2000. Preservation of old U–Th–Pb ages inshielded monazite: example from the Beni Bousera Hercynian kinzigites (Mo-rocco). J. Metamorph. Geol. 18, 335–342.

Nebel, O., Scherer, E.E., Mezger, K., 2011. Evaluation of the (87) Rb decay constantby age comparison against the U–Pb system. Earth Planet. Sci. Lett. 301, 1–8.

216 C.M.C. de Meyer et al. / Earth and Planetary Science Letters 395 (2014) 205–216

Oberhansli, R., 1982. The P-T history of some pillow lavas from Zermatt. Ofioliti 2,431–436.

Pollington, A.D., Baxter, E.F., 2011. High precision microsampling and preparationof zoned garnet porphyroblasts for Sm–Nd geochronology. Chem. Geol. 281,270–282.

Powell, R., Holland, T., 1999. Relating formulations of the thermodynamics of min-eral solid solutions: activity modeling of pyroxenes, amphiboles, and micas. Am.Mineral. 84, 1–14.

Reddy, S.M., Wheeler, J., Cliff, R.A., 1999. The geometry and timing of orogenic ex-tension: an example from the Western Italian Alps. J. Metamorph. Geol. 17,573–589.

Reddy, S.M., Wheeler, J., Butler, R.W.H., Cliff, R.A., Freeman, S., Inger, S., Pickles, C.,Kelley, S.P., 2003. Kinematic reworking and exhumation within the convergentAlpine Orogen. Tectonophysics 365, 77–102.

Reinecke, T., 1991. Very-high-pressure metamorphism and uplift of coesite-bearingmetasediments from the Zermatt-Saas Zone, Western Alps. Eur. J. Mineral. 3,7–17.

Rotenburg, E., Davis, D.W., Amelin, Y., Ghosh, S., Bergquist, B.A., 2012. Determina-tion of the decay-constant of 87Rb by laboratory accumulation of 87Sr. Geochim.Cosmochim. Acta 85, 41–57.

Rubatto, D., Gebauer, D., Fanning, M., 1998. Jurassic formation and Eocene sub-duction of the Zermatt-Saas-Fee ophiolites: implications for the geodynamicevolution of the Central and Western Alps. Contrib. Mineral. Petrol. 132,269–287.

Sartori, M., 1987. Structure of the Combin Zone between the Diablons and Zermatt(Valais). Eclogae Geol. Helv. 80, 789–814.

Seydoux, L., 2013. Interactions fluide-roche dans les ultramafiques de la zone deZermatt-Saas Fee. Master thesis. University of Lausanne, Lausannne, p. 188.

Skora, S., Baumgartner, L.P., Mahlen, N.J., Johnson, C.M., Pilet, S., Hellebrand, E., 2006.Diffusion-limited REE uptake by eclogite garnets and its consequences for Lu–Hfand Sm–Nd geochronology. Contrib. Mineral. Petrol. 152, 703–720.

Skora, S., Lapen, T.J., Baumgartner, L.P., Johnson, C.M., Hellebrand, E., Mahlen, N.J.,2009. The duration of prograde garnet crystallization in the UHP eclogites atLago di Cignana, Italy. Earth Planet. Sci. Lett. 287, 402–411.

Sousa, J., Kohn, M.J., Schmitz, M.D., Northrup, C.J., Spear, F.S., 2013. Strontiumisotope zoning in garnet: implications for metamorphic matrix equilibra-tion, geochronology and phase equilibrium modelling. J. Metamorph. Geol. 31,437–452.

Spear, F., 1993. Metamorphic Phase Equilibria and Pressure–Temperature–TimePaths. Mineralogical Society of America, Washington, DC. 799 pp.

Stampfli, G.M., Mosar, J., Favre, P., Pillevuit, A., Vannay, J.C., 2001. Permo-Mesozoicevolution of the western Tethys realm: the Neo-Tethys East Mediterranean Basinconnection. In: Cavazza, W., Robertson, A.H.F.R., Ziegler, P.A. (Eds.), PeritethyanRift/Wrench Basins and Passive Margins (IGCP 369), Peri-Tethys Memoir 6. Bull.Museum Nat. Hist. Nat., Paris, pp. 51–108.

Steck, A., 1987. Le massif du Simplon – Réflexions sur la cinématique des nappes degneiss. Schweiz. Mineral. Petrogr. Mitt. 67, 27–45.

Steck, A., Bigioggero, B., Dal Piaz, G.V., Escher, A., Martinotti, G., Masson, H., 2000.Carte tectonique des Alpes de Suisse occidentale et des régions avoisinantes1:100.000. Carte géologique speciale No. 123.. Serv. hydrol. géol. natl., Berne.

Steiger, R.H., Jaeger, E., 1977. Subcommission on geochronology – convention on useof decay constants in geochronology and cosmochronology. Earth Planet. Sci.Lett. 36, 359–362.

Szilvagyi, E.T., 2010. Rb–Sr geochronology of HP/UHP metasediments and furtherconstraints on peak metamorphism and exhumation rates, Zermatt-Saas ophio-lite. Western Alps. Master thesis. University of Wisconsin, Madison.

Vance, D., Onions, R.K., 1990. Isotopic chronometry of zoned garnets – growth-kinetics and metamorphic histories. Earth Planet. Sci. Lett. 97, 227–240.

Vannay, J.C., Allemann, R., 1990. The piedmontese zone in the Haut-Valtournanche(Aosta Valley, Italy). Eclogae Geol. Helv. 83, 21–39.

Villa, I.M., 1998. Isotopic closure. Terra Nova 10, 42–47.White, R.W., Powell, R., Holland, T.J.B., Worley, B.A., 2000. The effect of TiO2 and

Fe2O3 on metapelitic assemblages at greenschist and amphibolite facies con-ditions: mineral equilibria calculations in the system K2O–FeO–MgO–Al2O3–SiO2–H2O–TiO2–Fe2O3. J. Metamorph. Geol. 18, 497–511.

White, R.W., Powell, R., Phillips, G.N., 2003. A mineral equilibria study of the hy-drothermal alteration in mafic greenschist facies rocks at Kalgoorlie, WesternAustralia. J. Metamorph. Geol. 21, 455–468.