-
Evidence for protracted prograde metamorphism followed byrapid
exhumation of the Zermatt-Saas Fee ophiolite
S. SKORA,1 , 2 N. J . MAHLEN,3 C. M. JOHNSON,3 L. P.
BAUMGARTNER,1 T. J . LAPEN,4 B. L . BEARD3
AND E. T. SZILVAGYI31Institute of Earth Sciences, University of
Lausanne, Geopolis, 1015 Lausanne, Switzerland2Institute of
Geochemistry and Petrology, ETH Zurich, Clausiusstrasse, 25NW 8092
Zurich, Switzerland([email protected])3Department of
Geoscience, University of Wisconsin-Madison, 1215 W Dayton St,
Madison, WI 53706, USA4Department of Earth and Atmospheric
Sciences, University of Houston, 4800 Calhoun Road, Houston, TX
77004, USA
ABSTRACT Major and trace-element zoning in garnet, in
combination with Rb–Sr, Sm–Nd and Lu–Hfgeochronology, provide
evidence for a protracted garnet growth history for the
Zermatt-Saas Fee(ZSF) ophiolite, western Alps. Four new Lu–Hf ages
from Pfulwe (c. 52–46 Ma) and one from Cha-mois (c. 52 Ma) are very
similar to a previously published Lu–Hf age from Lago di Cignana.
Overall,the similarity of geochronological and garnet zoning
patterns suggests that these three localities had asimilar prograde
tectonic history, commensurate with their similar structural
position near the top ofthe ZSF. Samples from the lower part of the
ZSF at Saas Fee and St. Jacques, however, producedmuch younger
Lu–Hf ages (c. 41–38 Ma). Neither differences in whole-rock
geochemistry, whichmight produce distinct garnet growth histories,
nor rare-earth-element zoning in garnet, can accountfor the age
differences in the two suites. This suggests a much later prograde
history for the lowerpart of the ZSF, supporting the idea that it
was subducted diachronously. Such a model is consistentwith changes
in subduction vectors based on plate tectonic reconstructions,
where early oblique sub-duction, which produced long prograde
garnet growth, changed to more orthogonal subduction,which
corresponds to shorter prograde garnet growth. Six new Rb–Sr
phengite ages range from c. 42to 39 Ma and, in combination with
previously published Rb–Sr ages, constrain the timing of
thetransition from eclogite to upper greenschist facies P–T
conditions. The proximity of the ZSF in theSaas Fee region to the
underlying continental Monte Rosa unit and the similarity of
peak-metamor-phic ages suggest these two units were linked for part
of their tectonic history. This in turn indicatesthat the Monte
Rosa may have been partly responsible for rapid exhumation of the
ZSF unit.
Key words: geochronology; Lu–Hf; metamorphism; Rb–Sr;
Zermatt-Saas Fee.
INTRODUCTION
High-pressure (HP) and ultrahigh-pressure (UHP)metamorphic
terranes that are associated with conti-nent–continent collision
record subduction of crust todeep levels, followed by rapid
exhumation, asrequired to preserve evidence for (U)HP
conditions(e.g. Coleman & Wang, 1995; Hacker & Liou,
1998;Chopin, 2003). Mechanisms for rapid exhumation ofhigh-density
mafic rocks are debated, and possibilitiesinclude exhumation
through attachment to low-density serpentinites (e.g. Hermann et
al., 2000;Schwartz et al., 2001; Pilchin, 2005) or attachment
tocontinental fragments (e.g. Cloos, 1993; Lapen et al.,2007).
Indeed, many workers argue that buoyancymechanisms are the most
likely means by which rapid(>20 mm yr�1) exhumation of (U)HP
terranes occurs(e.g. Platt, 1987; Wheeler, 1991; Duchêne et
al.,1997b).
The Alpine chain is a classical continent–continentcollisional
orogen, where such theories might betested. Despite intensive
petrological and geochrono-logical studies of the (U)HP terranes of
the westernAlps, there still exists great uncertainty in the
timingof Alpine metamorphism. Part of this uncertainty liesin the
relations between ages and metamorphism –some ages may indeed
reflect peak metamorphism,but mineral growth (of garnet and zircon,
for exam-ple) along the prograde path may produce a range ofages
(e.g. Lapen et al., 2003; Anczkiewicz et al.,2007; Kylander-Clark
et al., 2007; Schmidt et al.,2008; Smit et al., 2010; Zirakparvar
et al., 2011;Kirchenbaur et al., 2012). The aim of this study is
toplace better constraints on the metamorphic historyacross the
Zermatt-Saas Fee (ZSF) unit (Fig. 1) byapplying the Lu–Hf garnet
geochronometer to a rela-tively large number of samples (n = 10)
from differ-ent structural positions in the alpine stack. The
new
© 2015 John Wiley & Sons Ltd 711
J. metamorphic Geol., 2015, 33, 711–734
doi:10.1111/jmg.12148
-
ages are linked to trace-element zonation in garnet,which
together place tight constraints on the timingof prograde and peak
metamorphism. New Rb–Srisochron ages for phengite provide
additional con-straints on the transition between eclogite to
uppergreenschist facies P–T conditions (see also de Meyeret al.,
2014). Last, in the light of the new data set, aswell as previously
published geochronological dataon the underlying continental Monte
Rosa unit, weexamine the timing of metamorphism across the ZSFsheet
to better understand the relations between con-tinental and oceanic
units during subduction-relatedmetamorphism (e.g. Lapen et al.,
2007).
GEOLOGICAL SETTING
Alpine rocks (before collision) can be generally subdi-vided
into: (i) pre-Triassic basement rocks and (ii)Triassic to early
Cenozoic cover sedimentary rocksthat were deposited in a
predominantly shallow watershelf environment (e.g. Tr€umpy, 1980).
During theEarly Jurassic, this carbonate platform broke up
inconjunction with establishment of the Tethyan-
Atlantic junction, which gave rise to the developmentof the
Liguro-Piemont Ocean basin (c. 165–160 Ma,Rubatto et al., 1998;
Schaltegger et al., 2002), andseparation of Europe from
Apulia/Africa. The open-ing of the Atlantic Ocean towards the north
in theLate Jurassic to Early Cretaceous led to shorteningbetween
Europe and Apulia/Africa (e.g. Stampfliet al., 1998).
Southeast-directed subduction andnorthwest-directed thrusting
during the Late Creta-ceous through early Cenozoic eventually
resulted inclosure of the Liguro-Piemont Ocean basin and cre-ation
of the Alpine tectonic nappe system (e.g. Hun-ziker, 1974; Dal Piaz
& Ernst, 1978). The ZSFophiolite represents the (U)HP
metamorphic remnantof that oceanic material. It is comprised of
peri-dotites, serpentinites, eclogitized metagabbros andmetabasalts
that contain local examples of deformedsheeted dyke systems and
well preserved pillow struc-tures, and a cover series of calcareous
and siliceousmetasedimentary rocks (e.g. Bearth, 1967; Dal Piaz
&Ernst, 1978; Barnicoat & Fry, 1986).The paleogeographic
organization used throughout
this paper is from north to south (Fig. 2a): (i) Euro-pean
continental margin (Helvetic: Late Paleozoiccrystalline basement
rocks, covered by various sedi-mentary Mesozoic/Cenozoic nappes),
(ii) Valais Basin(Lower Cretaceous ‘B€undnerschiefer/Schiste
Lustr�e’),(iii) Brianconnais domain (Late Paleozoic
crystallinebasement rocks, covered by various Mesozoic sedi-mentary
rocks), (iv) remnants of the Jurassic Liguro-Piemont oceanic crust
(ophiolitic HP zone, overlain bythe sediment-dominated,
low-pressure Combin zone),(v) northernmost Apulian/African
continental marginor a distal micro-continent (Austroalpine:
pre-Triassiccrystalline rocks), (vi) Apulian/African
continentalmargin (non-metamorphosed Southalpine). Note thatthe
low-P (upper greenschist/lower blueschist facies)Combin zone has
been further subdivided into the (i)Mont Fort nappe
(Triassic/Mesozoic cover sedimen-tary rocks of the Saint Bernard
nappe, which is part ofthe Brianconnais) and (ii) Tsat�e nappe
(ophioliticm�elange zone with overlying Upper Cretaceous
calc-schists, Sartori, 1987). The Tsat�e nappe represents
theaccretionary wedge below which the ZSF was sub-ducted (Sartori,
1987; Marthaler & Stampfli, 1989).Because subduction was
southeast directed (Fig. 2b),
the oldest peak-metamorphic ages are recorded in fel-sic
eclogites from the Sesia-Lanzo zone (Austroalpine)at c. 75–65 Ma
(Inger et al., 1996; Duchêne et al.,1997a; Rubatto et al., 1999,
2011). Subsequent sub-duction of the ZSF unit to eclogite facies
conditionsproduced a suite of metamorphic ages rangingbetween c. 50
and 40 Ma (Duchêne et al., 1997a;Rubatto et al., 1998; Amato et
al., 1999; Dal Piazet al., 2001; Lapen et al., 2003; Rubatto &
Hermann,2003; Gouzu et al., 2006; Herwartz et al., 2008).Focusing
on ages from north of the Aosta fault, theSm–Nd garnet isochron age
of 40.6 � 2.6 Ma mostclosely dates the peak of metamorphism (Amato
et al.,
Zermatt
Allalinhorn
Matterhorn
Breuil
10 km
Sesia zone & Dent Blanche nappeCombin zone (Tsaté & Mt.
Fort nappe)Zermatt-Saas Fee zoneGornergrat zoneMonte Rosa
nappeGrand St. Bernard nappestudy area
St. Jacques
Pfulwe path:P-02,80bP-80c,96
Lago di Cignana:96JA-32, 01NM-45
Chamois:CH-48
Sample location
Saas Fee:SF-25b,2605NM-219
Saas Fee:05NM-214,215
Saas Fee:05NM-212
St. Jacques:SJ-87
Lago di Cignana:08ES-03
Pfulwe pass:P-98,100
SaasFee
Fig. 1. Geological map showing the Liguro-Piemont
oceanicremnants at the western Swiss/Italian border (simplified
afterSteck et al., 1999). The studied samples from Zermatt-Saas
Feeophiolite are from the Pfulwe, Chamois, Lago di Cignana,
St.Jacques and Saas Fee areas, and sample names are listed foreach
locality.
© 2015 John Wiley & Sons Ltd
712 S . SKORA ET AL .
-
1999), because of strong enrichment of Sm at thegarnet rim
(Skora et al., 2009). Rb–Sr metamorphicages of the Tsat�e nappe
range from c. 44 to 37 Ma(Reddy et al., 1999; Cartwright &
Barnicoat, 2002).Peak-metamorphic ages from subducted
continentalslices of the southernmost Brianconnais
continentalmargin such as Monte Rosa, the Gran Paradiso andthe Dora
Maira nappes are c. 42–35 Ma (Tilton et al.,1991; Duchêne et al.,
1997a; Gebauer et al., 1997;Rubatto & Gebauer, 1999; Engi et
al., 2001; Meffan-Main et al., 2004; Lapen et al., 2007).
Subsequentexhumation through upper greenschist facies is datedat c.
40–37 Ma in the Sesia-Lanzo zone (Inger et al.,1996), c. 40–35 Ma
in the ZSF and adjacent Tsat�enappe (M€uller, 1989; Barnicoat et
al., 1995; Reddyet al., 1999; Cartwright & Barnicoat, 2002)
andc. 35–32 Ma in the units from the southernmost Brian-connais
domain (Freeman et al., 1997; Engi et al.,2001; Meffan-Main et al.,
2004).
SAMPLES AND PETROLOGY
Samples were taken to reflect a broad geographic dis-tribution
as well as different structural levels (Fig. 1).The Lago di Cignana
unit is from the structurallyhighest position, just beneath the
Tsat�e nappe. TheChamois locality is only 5 km from Lago di
Cignana,and likely at a similar structural position within theZSF
unit. The Pfulwe samples are also from a struc-turally high
position, although field relations suggestthat it is slightly lower
than Lago di Cignana. The St.Jacques locality lies in the interior
of the ZSF, but itsstructural position is unclear. In contrast, the
Sass-Feesamples clearly reflect the structurally lowest positionin
the ZSF package, directly lying above the MonteRosa nappe. The
petrology of the eclogite samples isdiscussed first, followed by
the metasedimentary sam-ples collected at Lago di Cignana and near
Saas Fee.Thin section images for representative samples areshown in
Figs 3a–f, 4a–f and 5a–d. Sample locationsare given in Table S1
(eclogites only) and Fig. 1.
Eclogites
All eclogites contain abundant omphacitic clinopy-roxene and
garnet porphyroblasts. Other minerals,such as white mica (mostly
paragonite, some phen-gite), glaucophane, epidote–clinozoisite
solid solu-tions (collectively called epidote hereafter
ineclogites), lawsonite (pseudomorphs), carbonate,quartz and rutile
also occur in the ZSF eclogites.These minerals may or may not be
part of the peak-metamorphic assemblage, co-existing with
garnet.Minor retrogression is present in all samples. It
ischaracterized by transformation of omphacite intoNa–Ca hornblende
� albite; garnet rims into Na–Cahornblende � albite � chlorite �
epidote; transfor-mation of glaucophane into Na–Ca hornblende
(Na–Ca hornblende is either barrosite, taramite or kato-phorite);
as well as rutile that is retrograded intotitanite and occasionally
ilmenite at the rims.
Pfulwe
The samples P-80b, P-80c, P-02, P-96 come from asmall outcrop
just below Pfulwe pass (~0.5 kmENE), whereas samples P-98 and P-100
come fromthe actual Pfulwe pass, which is located ~7.5 km eastof
Zermatt. The geology and petrology of this area isoutlined in, for
example, Bearth (1967), Ernst & DalPiaz (1978), Oberh€ansli
(1982), Barnicoat & Fry(1986) and Barnicoat (1988). Both
outcrops are dis-cussed together as the Pfulwe area hereafter.
Ompha-cite often occurs as small grains that have
undulatoryextinction in these samples, preserving a
radial,flower-like growth structure (Fig. 3a,b). This appear-ance
suggests that these rocks record little or nodeformation. Nearly
idiomorphic garnet is coarse(Fig. 3c,d) reaching up to 1 cm.
Omphacite, rutile,glaucophane and quartz inclusions occur in
garnet,whereas epidote and ilmenite inclusions are moreabundant in
the cores. Matrix paragonite and glauco-phane likely belong to the
prograde assemblage for
Paleogeographic setting in Jurassic times
Today
Valais basin Brianconnais domain Liguro-Piemont ocean
Apulia/Africa
Europe
2
1: Zone Houillere/Pontis nappe2: Grand Saint Bernard nappe3:
Mont Fort nappe 4. Monte Rosa nappe5: Zermatt-Saas Fee zone6: Tsaté
nappe 7: Dent Blanche nappe
Matterhorn
NW SE
6 5
1
43 7
sub-continental mantle
2 43 7
sediments
5 6
?
1
5
astheno-sphere
(a)
(b)Fig. 2. (a) Schematic diagram showing thepaleogeographic
setting of the Zermatt-SaasFee and related units in a profile from
NWto SW during Jurassic oceanic rifting(modified after Labhart,
1992). (b)Geological profile approximating thecurrent tectonic
situation of the WesternAlps (modified after Labhart, 1992).
© 2015 John Wiley & Sons Ltd
GEOCHRONOLOGY OF THE ZERMATT -SAAS OPHIOL I TE 713
-
omp
grt
grt
96JA-32
omp
grt
grt
lwspseudom.
P-96
grt
rt omp
P-96
grt
rt omp
P-80b
Radialomp
grt
gln
omp
Na-Cahbl
rt
P-80b
Radialomp
grt
gln
omp
Na-Cahbl
rt
96JA-32
1 mm 1 mm
lwspseudom.
1 mm 1 mm
1 mm 1 mm
(a) (b)
(c) (d)
(e) (f)
Fig. 3. Photomicrographs illustrating the metamorphic textures
of dated samples (left columns are in PPL, and right columns
aresame views in XP). All ZSF eclogites contain omphacitic
clinopyroxene and garnet porphyroblasts that vary considerably in
size.Other minerals, such as white mica (phengite or paragonite),
glaucophane, epidote, lawsonite, carbonate, quartz and rutile
alsooccur. (a, b) Eclogites from the Pfulwe area are relatively
undeformed. In places, omphacite even preserve a radial,
flower-likegrowth structure. (c, d) Garnet can grow exceptionally
large at Pfulwe, occasionally reaching up to 1 cm in diameter. (e,
f)Lawsonite pseudomorphs are commonly observed at Lago di Cignana,
as both inclusions in garnet and in the matrix. grt, garnet;omp,
omphacite; Na–Ca hbl, Na–Ca hornblende; gln, glaucophane; rt,
rutile; lws pseudom., lawsonite pseudomorphs.
© 2015 John Wiley & Sons Ltd
714 S . SKORA ET AL .
-
reasons discussed in Skora et al. (2008). Epidote,Fe-rich
dolomite, rutile, apatite and quartz often occurin various but
small amounts. The peak-metamorphic
assemblage is: omphacite + garnet + paragonite +glaucophane +
epidote + rutile (+ carbonate +quartz).
SF-25b
grt
grt
gln
rt->ttn
SF-25b
grt
grt
gln
lwspseudom.
rt->ttn
CH-48
omp
grt
grt
pgrt
CH-48
omp
grt
grt
pgrt
SJ-87
grt
omp
rt
gln
ttn incl.
SJ-87
grt
omp
rt
gln
ttn incl.
1 mm
1 mm 1 mm
1 mm
lwspseudom.
1 mm 1 mm
(a) (b)
(c) (d)
(e) (f)
Fig. 4. Photomicrographs illustrating the metamorphic textures
of dated samples (left columns are in PPL, and right columns
aresame views in XP). The degree of deformation in these samples is
strongly increased when compared with textures of samples
fromPfulwe and Lago di Cignana (Fig. 3a–d), although the mineral
paragenesis is comparable. (a, b) The sample from Chamoisexhibits a
strong metamorphic foliation defined by oriented omphacite and
rutile. (c, d) Garnet at Saas Fee is sometimes ruptured,suggesting
that deformation was very strong. (e, f) The sample from St.
Jacques differs from all other samples in that garnet ismuch
smaller on average. Abbreviations as in Fig. 3; pg, paragonite;
ttn, titanite.
© 2015 John Wiley & Sons Ltd
GEOCHRONOLOGY OF THE ZERMATT -SAAS OPHIOL I TE 715
-
Lago di Cignana
The petrology of this sample is described in detail inAmato et
al. (1999), Lapen et al. (2003) and Skoraet al. (2009). Briefly,
the matrix is composed of rela-tively coarse-grained omphacite,
glaucophane, epi-dote, paragonite and phengite, which define a
weakmetamorphic foliation (data not shown). Garnet por-phyroblasts
are ~0.4–4 mm in diameter. Lawsonitepseudomorphs are relatively
common in this sample(Fig. 3e,f). The peak-metamorphic assemblage
of96JA-32 is garnet + omphacite + glaucophane +epidote + rutile +
carbonate (+ phengite).
Chamois
One sample (CH-48) taken from the Nuarsax/Cha-mois area in the
Valtournenche is layered and well
foliated. The layering is marked by the presence orabsence of
lawsonite pseudomorphs (data not shown),and foliation is defined by
oriented omphacite(Fig. 4a,b). In lawsonite-free layers, the
foliationwraps around ~1–3 mm sized garnet. Garnet
inclusionpatterns differ from Pfulwe samples in that titanite is
avery frequent inclusion (besides omphacite), especiallyin garnet
cores. Rutile supplants titanite as the Ti-richmineral inclusion in
garnet rims. The few paragonitegrains that occur in the matrix
often crosscut the folia-tion (Fig. 4a,b), hence, they are likely
to have grownpost-deformational. Given that paragonite is stable
atblueschist facies conditions, it is likely that it wouldhave been
deformed alongside omphacite if paragonitebelonged to the prograde
assemblage. Rutile grainsare also present in the matrix. The
peak-metamorphicassemblage of the omphacite-rich layers is
omphacite+ garnet + rutile. The other layers contain additional
08ES-03
1 mm 1 mm
0.5 mm 0.5 mm
08ES-03
ph
qtz
ph
qtz
pmt pmt
cbcbphph
05NM-21905NM-219
qtzqtz
bt bt
(a) (b)
(c) (d)
Fig. 5. Photomicrographs illustrating the metamorphic textures
of dated samples (left columns are in PPL, and right columns
aresame views in XP). The appearance of metasedimentary rocks
varies from quartz-rich schists to calcsilicates. All samples
containrelatively coarse-grained phengite that was used for Rb–Sr
dating. (a, b) Sample 08ES-03 from Lago di Cignana is
unusuallymanganiferous, as suggested by the presence of piemontite
(=manganiferous epidote). (c, d) All samples from the Saas Fee
areaare relatively rich in carbonate in addition to phengite and
quartz. ph, phengite; cb, carbonate; qtz, quartz; pmt, piemontite;
bt,biotite.
© 2015 John Wiley & Sons Ltd
716 S . SKORA ET AL .
-
coarse-grained paragonite + epidote � albite assem-blages, in
which minerals are not oriented. Theseassemblages are similar to
the lawsonite pseudo-morphs described by Skora et al. (2009) for
Lago diCignana (Fig. 3e,f), except that the lawsonite habit isoften
not clearly preserved. The matrix lawsonite isnevertheless
interpreted to have formed during earlyprograde metamorphism
followed by transformationon the retrograde path, given the
undeformed andunoriented appearance of mica.
Saas Fee
Two samples were collected close to the BritanniaHut (Saas Fee).
The matrix of the eclogites consistsof fine-grained omphacite,
epidote and rare glauco-phane, all of which define a very strong
foliation.Garnet in these samples is up to 3 mm in diameter,and
extremely inclusion rich. Grains are partly bro-ken, suggesting
that some deformation occurred aftergarnet growth ceased. Titanite
is more abundant thanrutile in the inclusion population, especially
in garnetcores. In addition, titanite (�ilmenite) surrounds
orreplaces rutile in the matrix. Lawsonite pseudo-morphs, marked by
epidote + paragonite � albite,occur in both samples as garnet
inclusions (Fig. 4c,d).Matrix lawsonite pseudomorphs appear to
berestricted to sample SF-26. Saas Fee samples arecharacterized by
the peak-metamorphic assemblageomphacite + garnet + rutile +
glaucophane + epidote.
St. Jacques
One sample was taken northeast of St. Jacques in theVal d0Ayas
(SJ-87). Matrix omphacite is aligned anddefines a foliation (Fig.
4e,f). Garnet in this samplediffers from all other samples in that
it is much smal-ler (≤1 mm, Fig. 4e,f) but very abundant. In terms
ofinclusion patterns, it has abundant titanite in thecores, and
rutile in the rims. Other minerals that arealigned in the matrix
include epidote, omphacite,glaucophane and paragonite. Glaucophane
com-monly has rutile inclusions, in contrast to garnetcores,
suggesting that it started to grow close to peakconditions. This
sample also contains lawsonite pseu-domorph textures (paragonite +
epidote � albite),very similar to the Saas Fee samples. The
peak-meta-morphic assemblage is omphacite + garnet + glauco-phane +
epidote + rutile + paragonite.
Metasedimentary rocks
Metasedimentary rocks in the ZSF were sampled forRb–Sr
geochronology at Lago di Cignana, and nearSaas Fee, representing
the lower and upper parts ofthe ZSF respectively. The mineralogy
and composi-tions of metasedimentary rocks in the ZSF varygreatly,
as primarily reflected in variable proportionsof quartz, garnet,
carbonate and mica.
Lago di Cignana
Samples are garnet-bearing, quartz-rich schists(quartz: 40–65%,
phengite: 15–25%, carbonate:
-
Petrological studies conducted on eclogites fromthe Western Alps
all indicate that Tpeak and Ppeakwere reached at roughly the same
point in time (e.g.Reinecke, 1998). This is in contrast to studies
oneclogites from the Central Alps (e.g. Brouwer et al.,2005) that
appear to have reached their Tpeak duringthe collisional stage.
Eclogites from the Central Alpsare heavily overprinted. Their
overprint assemblage isconsistent with exhumation through
amphibolitefacies, in contrast to eclogites from the Western
Alpsthat only show a greenschist facies overprint. The lat-ter is
indicative of cooling during exhumation.Because of the shape of
garnet isopleths in P–T spacein eclogites (e.g. Hoschek, 2001),
garnet cannot havegrown during exhumation unless it was subjected
toheating upon exhumation. The term peak metamor-phism is thus used
to describe the point of deepestburial, which equates to Tpeak and
Ppeak in the Wes-tern Alps, and it is assumed that garnet growth
musthave ceased upon reaching this point.
ANALYTICAL METHODS
Whole-rock (WR) powders were prepared from slabsof rock, which
were trimmed to remove weatheredportions, crushed in a steel jaw
crusher and reduced tosand-sized particles using an alumina-lined
disc mill(University of Wisconsin). The samples were thensplit,
where a portion was saved for mineral separates,a portion was
powdered using an alumina-lined shat-terbox and a portion was
powdered using a tungsten-carbide shatterbox. WR data given in
Table 1 areXRF data (University of Lausanne), obtained on thesame
WR powder that was used for geochronology.
Mineral chemistry
Garnet
Central cuts of garnet in thick sections of eclogiteswere
prepared for major and trace-element analyses,using X-ray
tomography at the University of Lau-sanne. Major-element X-ray maps
of all measuredgrains were obtained prior to acquiring
wavelength-dispersive quantitative analyses using a Cameca SX-
50 (5 spectrometers) electron microprobe (Lausanne).The same
thick sections and garnet profiles werethen used to obtain
trace-element profiles using laserablation inductively coupled
plasma mass spectrome-try (LA-ICP-MS). Trace-element data were
acquiredon a Perkin-Elmer ELAN 6100 DRC ICP-MS atLausanne. Samarium
data for one Pfulwe garnetwere obtained by secondary ion mass
spectrometry(SIMS) at the Max Planck Institute of Chemistry(Mainz).
A detailed description on garnet sectioning,EMPA, LA-ICP-MS and
SIMS is published inSkora et al. (2006). Garnet chemical data are
plottedin Fig. 6a–c.
Phengite
Chemical compositions of phengite in the metasedi-mentary rocks
were measured using the CamecaSX51 electron microprobe at 15 keV
and 20 nA atthe University of Wisconsin-Madison. Data are givenin
Table S2.
Rb–Sr, Sm–Nd and Lu–Hf isotopes
Mineral separates and WR samples of eclogites wereprocessed for
Sm–Nd and Lu–Hf geochronology fol-lowing the methods in Lapen et
al. (2004) at theUniversity of Wisconsin-Madison. Analytical
detailsare only briefly summarized here, whereas a fulldescription
is given in Appendix S1. Eclogite samplesand minerals were
dissolved in Parr bombs, followedby several dry-down steps.
Complete dissolution wasconfirmed after centrifuging each sample.
The Amatoet al. (1999) leaching procedure was used on
garnetfractions spiked only with the mixed 149Sm–150Ndtracer.
Garnet-rich fractions spiked with both mixed176Lu–178Hf and
149Sm–150Nd tracers were not sub-jected to sequential dissolution
procedures as out-lined in Amato et al. (1999), because Mahlen et
al.(2008) found that this method could potentially frac-tionate Lu
and Hf. There was further concern thatleaching or partial
dissolution may preferentially dis-solve certain garnet
populations, and our geochrono-logic modelling is intended to model
total garnetevolution.
wt% P-80b P-80c P-96 P-02 P-98 P-100 96JA-32a CH-48 SF-25b SF-26
SJ-87
SiO2 48.4 50.0 50.4 48.7 49.7 51.4 50.0 50.3 46.2 47.7 48.5
TiO2 1.8 1.6 2.5 1.7 2.2 2.4 1.8 1.9 1.9 2.1 2.2
Al2O3 14.6 15.4 15.3 15.5 15.9 16.0 14.8 15.5 15.7 15.7 14.6
Fe2O3 9.4 8.3 10.0 9.9 11.1 10.8 9.5 8.6 10.4 9.8 11.4
MnO 0.2 0.1 0.2 0.1 0.1 0.1 0.2 0.2 0.2 0.1 0.2
MgO 5.9 4.9 4.2 5.4 4.3 4.4 6.4 4.9 4.4 4.9 5.7
CaO 9.2 10.5 7.8 8.8 8.1 7.0 10.5 11.7 15.6 13.5 11.3
Na2O 6.2 6.9 6.7 6.4 5.8 5.8 3.9 5.4 3.8 4.2 3.9
K2O 0.1 0.0 0.1 0.0 0.2 0.3 0.2 0.0 0.0 0.1 0.1
P2O5 0.2 0.3 0.4 0.2 0.3 0.4 0.2 0.2 0.2 0.2 0.2
LOI 3.0 1.0 1.7 2.3 1.2 0.5 1.7 0.3 0.7 0.8 0.8
Sum 99.0 99.0 99.3 99.0 98.9 99.1 99.2 99.0 99.1 99.1 98.9
Major elements determined by XRF; LOI, loss of ignition.aSample
from Amato et al. (1999) and Lapen et al. (2003).
Table 1. Whole-rock data for eclogites (wt%).
© 2015 John Wiley & Sons Ltd
718 S . SKORA ET AL .
-
Lutetium and Hf isotopes were measured on aMicromass Isoprobe.
The 176Lu decay constant ofS€oderlund et al. (2004) was used,
at1.865 9 10�11 yr�1, which is essentially identical tothat of
Scherer et al. (2001). Samarium and Nd iso-topes were measured on a
Micromass Sector 54 ther-
mal ionization mass spectrometer (TIMS). The decayconstant for
147Sm used was 6.54 9 10�12 yr�1. Iso-tope data for both Lu–Hf and
Sm–Nd are given inTables 2 and 3, and are plotted in Fig.
7a–s.Mineral separates and WR samples of metasedi-
mentary rocks were analysed for Rb–Sr isotopes at
0.4
0.5
0.5
1.0
1.5
20
40
60
80
100
120
140
160
0.1
0.2
0.3
LuSm (ICP)Sm (SIMS)
Lu [p
pm]
Rim-to-rim zoning [mm]
MnFeMgCa
Cat
ions
[pfu
] C
atio
ns [p
fu]
Cat
ions
[pfu
]
0.5
1.0
1.5
2.0
Pfulwe (sample P-80b)
Chamois (sample CH-48)
10
20
30
40
50Lu
[ppm
]
1.0
1.5
2.0
Sm
[ppm
]S
m [p
pm]
Saas Fee (sample SF-25b)
0.5
1.0
1.5
2.0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.5 1.0 1.5 2.0 2.5 3.0 3.5
0.5 1.0 1.5 2.0 2.5 0.5 1.0 1.5 2.0 2.5
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.2 0.4 0.6 0.8 1.0 1.2 1.4
1.6
2
4
6
8
10
12
14
16
18
Lu [p
pm]
0.5
0.6
0.7
0.8
0.9
0.1
0.2
0.3
0.4 Sm
[ppm
]average quantificationlimit of Sm
average quant.limit of Sm
(a)
(b)
(c)
Fig. 6. Dual plots of major elements (EMPA) and Lu and Sm
(LA-ICP-MS, SIMS) zoning pattern for representative garnetsamples,
one from each investigated location. All data were collected on
garnet central sections, as determined by X-raytomography. Most Sm
concentrations of garnet cores are below the limit of
quantification for LA-ICP-MS. The data are stillplotted, but this
only serves to illustrate that the garnet core is depleted in Sm
when compared with the rim compositions.
© 2015 John Wiley & Sons Ltd
GEOCHRONOLOGY OF THE ZERMATT -SAAS OPHIOL I TE 719
-
the University of Wisconsin. A 87Rb–84Sr mixed iso-tope tracer
was added to the samples; two mixedspikes were used, one for high
Rb/Sr ratio phasessuch as mica (spike molar Rb/Sr = 191) and one
sui-ted for materials with lower Rb/Sr ratios (spikemolar Rb/Sr =
3.8). Rubidium and Sr were separatedusing cation-exchange
chromatography and analysedusing a VG Instruments Sector 54 TIMS.
The 87Rbdecay constant of Rotenburg et al. (2012) was used,1.3971 9
10�11 yr�1, which is close to other recentdeterminations by Kossert
(2003) and Nebel et al.(2011). For samples in the age range of 40
Ma, thedecay constant of Rotenburg et al. (2012) producesages c.
0.6 Ma older than those calculated using theold decay constant of
Steiger & J€ager (1977). Rb–Srdata for the metasedimentary
rocks are given inTable 4 and are plotted in Fig. 8a–f.
Errors used in isochron calculations reflect mea-sured
uncertainties of individual analyses (internal 2standard error, SE)
for 143Nd/144Nd, 176Hf/177Hf and87Sr/86Sr. The uncertainties in all
parent/daughter
Table 2. Sm–Nd isotope data for Zermatt-Saas Fee eclogites.
Mat. Sm (ppm) Nd (ppm) 147Sm/144Nd 143Nd/144Nd 2SE eNd
Pfulwe sample: P-80b
grt l 0.432 0.181 0.2482 0.513076 �15 8.5grt la – – – 0.513077
�17 8.6WR 17.0 4.51 0.1618 0.513069 �7 8.4gln 8.93 2.25 0.1536
0.513061 �8 8.3Pfulwe sample: P-80c
grt l 0.433 0.282 0.3887 0.513131 �64 9.6WR 13.1 3.50 0.1627
0.513067 �8 8.4Pfulwe sample: P-02
grt l 0.804 0.281 0.2120 0.513099 �92 9.0grt la – – – 0.513081
�38 8.6grt la – – – 0.513088 �15 8.8grt c. 3.87 1.07 0.1684
0.513052 �12 8.1WR 15.3 4.24 0.1687 0.513078 �8 8.6omp 9.38 2.53
0.1637 0.513056 �9 8.2Pfulwe sample: P-98
grt 1.50 0.54 0.2190 0.513053 �14 8.1grta – – – 0.513058 �31
8.2WR 21.38 5.76 0.1638 0.513059 �8 8.2Pfulwe sample: P-100
grt 0.84 0.54 0.3912 0.513119 �24 9.4WR 19.95 6.03 0.1608
0.513049 �7 8.0omp 26.74 6.94 0.1577 0.513055 �8 8.1Chamois sample:
CH-48
grt 2.95 1.20 0.2469 0.513084 �8 8.7WR 16.0 4.36 0.1656 0.513053
�5 8.1Saas Fee sample: SF-25b
grt 1.44 3.42 0.2558 0.513120 �7 9.4WR 5.30 16.36 0.1970
0.513090 �7 8.8Saas Fee sample: SF-26
grt l 0.702 1.14 0.3722 0.513100 �23 9.0grt la – – – 0.513103
�34 9.1grt 1.96 4.25 0.2799 0.513090 �13 8.8WR 6.30 20.3 0.1889
0.513086 �7 8.7omp 3.06 9.76 0.1906 0.513073 �9 8.5gln 4.03 13.5
0.1821 0.513074 �7 8.5St. Jacques sample: SJ-87
grt 7.69 2.92 0.2305 0.513094 �6 8.9WR 19.4 6.13 0.1927 0.513078
�7 8.6
Mat., material; grt, garnet; omp, omphacite; gln, glaucophane;
WR, whole rock; l, lea-
ched; c, core.
2SE are from in-run statistics.
Where replicate analyses occur, isochrons are calculated from
averages.aAnalyses that are repeat measurements of the same
dissolution.
Table 3. Lu–Hf isotope data for Zermatt-Saas Fee eclogites.
Mat. Lu (ppm) Hf (ppm) 176Lu/177Hf 176Hf/177Hf 2SE eHf
Pfulwe sample: P-80b
grt 3.00 3.12 0.1363 0.283184 �5 14.1grta – – – 0.283189 �8
14.3grta – – – 0.283187 �7 14.2WR 0.492 3.90 0.0180 0.283081 �7
10.5WRa – – – 0.283079 �11 10.4omp 0.131 3.44 0.0054 0.283062 �9
9.8gln 0.052 3.41 0.0021 0.283062 �8 9.8Pfulwe sample: P-80c
grt 3.93 3.17 0.1764 0.283235 �5 15.9grta – – – 0.283233 �12
15.8grta – – – 0.283231 �8 15.8WR 0.352 3.77 0.0130 0.283076 �7
10.3WR 0.381 3.90 0.0139 0.283079 �6 10.4omp 0.100 3.85 0.0037
0.283062 �8 9.8Pfulwe sample: P-96
grt 4.66 4.88 0.1356 0.283184 �6 14.1grta – – – 0.283198 �6
14.6grta – – – 0.283195 �7 14.5grta – – – 0.283197 �6 14.6grta – –
– 0.283191 �6 14.4WR 0.710 5.83 0.0174 0.283093 �5 10.9omp 0.152
5.49 0.0040 0.283076 �8 10.3Pfulwe sample: P-02
grt 2.85 3.41 0.1190 0.283175 �6 13.8grta – – – 0.283173 �8
13.7grt 2.80 3.53 0.1132 0.283176 �6 13.8grta – – – 0.283176 �6
13.8grta – – – 0.283176 �6 13.8grt c 4.24 3.01 0.2006 0.283249 �10
16.4WR 0.442 3.74 0.0167 0.283088 �9 10.7WRa – – – 0.283083 �13
10.5omp 0.051 1.29 0.0055 0.283078 �7 10.4ompa – – – 0.283083 �6
10.5Pfulwe sample: P-98
grt 2.47 4.20 0.0837 0.283164 � 7 13.4grta – – – 0.283160 � 5
13.3grta – – – 0.283163 � 8 13.4WR 0.660 5.21 0.0179 0.283096 � 7
11.0Pfulwe sample: P-100
grt 1.95 4.44 0.0624 0.283122 �8 11.9grta – – – 0.283126 �7
12.1grta – – – 0.283122 �8 11.9WR 0.661 5.62 0.0166 0.283088 �5
10.7omp 0.162 5.76 0.0038 0.283080 �6 10.4Chamois sample: CH-48
grt 3.13 3.05 0.1458 0.283221 �7 15.4grta – – – 0.283222 �6
15.6grt 3.05 2.95 0.1466 0.283224 �6 15.5grta – – – 0.283220 �8
15.4WR 0.524 4.31 0.0173 0.283097 �7 11.0WRa – – – 0.283100 �5
11.1WRa – – – 0.283096 �6 11.0omp 0.0621 3.37 0.0024 0.283077 �5
10.3ompa – – – 0.283080 �6 10.4Saas Fee sample: SF-25b
grt 3.87 3.89 0.1412 0.283195 �8 14.5grta – – – 0.283196 �6
14.5WR 0.661 4.00 0.0236 0.283111 �5 11.5WRa – – – 0.283112 �7
11.6WRa – – – 0.283113 �5 11.6Saas Fee sample: SF-26
grt 6.47 3.71 0.2482 0.283299 �11 18.2grta – – – 0.283289 �7
17.8grta – – – 0.283286 �8 17.7WR 0.720 4.69 0.0218 0.283120 �10
11.6WRa – – – 0.283119 �5 11.8omp – 1.62 – 0.283112 �10 11.6gln –
3.70 – 0.283117 �10 11.7St. Jacques sample: SJ-87
grt 4.53 2.92 0.2208 0.283284 �6 17.7grta – – – 0.283286 �7
17.7WR 0.771 4.91 0.0224 0.283139 �5 12.5omp 0.049 3.54 0.0020
0.283126 �7 12.1
Mat., material; grt, garnet; omp, omphacite; gln, glaucophane;
WR, whole rock; c, core.
2SE are from in-run statistics.
Where replicate analyses occur, isochrons are calculated from
averages.aAnalyses that are repeat measurements of the same
dissolution.
© 2015 John Wiley & Sons Ltd
720 S . SKORA ET AL .
-
ratios were set equal to 0.5%, but this has little effecton the
propagated errors for the isochrones due tothe relatively young age
of the samples.
RESULTS
Results are presented first for garnet and phengitechemistries,
which were used for Sm–Nd and Lu–Hfand Rb–Sr geochronology
respectively. Sm–Nd andLu–Hf ages for eclogites and Rb–Sr ages
formetasedimentary rocks are summarized in Tables 5and 6
respectively.
Garnet chemistry
Major-element profiles determined by electron micro-probe
exhibit concentric prograde growth zoning thatis characterized by
spessartine and grossular contents
that are highest in the cores, and pyrope and alman-dine
contents that are highest at the rims. The exactzoning pattern, as
well as the magnitude, variesslightly by location.
Rare-earth-element profiles for allgarnet grains vary
systematically from Lu to Sm, verysimilar to that described in
Skora et al. (2006). OnlyLu (enriched in early grown cores) and Sm
(enrichedin late-grown rims) data are shown; these elements arethe
most important for understanding the geochrono-logical results.
Selected examples of characteristic zon-ing patterns of each
locality are given in Fig. 6a–c.Note that no central garnet section
of the St. Jacquessamples was obtained because they are too
small.
Phengite chemistry
Phengite from Lago di Cignana exhibits Si that variesfrom 3.26
to 3.48 atoms per formula unit (apfu; aver-
97 ± 320 MaMSWD = 8.2
0.51322
0.51298
0.51314
0.51306grt
leachedgrt
WRomp
–10 ± 51 Ma
0.51322
0.51298
0.51314
0.51306WR grt
P-96 Pfulwe
143 N
d/14
4 Nd
20 ± 21 MaMSWD = 1.7
43 ± 43 Ma
0.51322
0.51298
0.51314
0.51306
0.51298
0.51314
0.51306
147Sm/144Nd
0.1 0.3 0.5 0.7
leachedgrt
WR
leachedgrt
WR
glc
0.51322
Sm-Nd
46.7 ± 2.1 MaMSWD = 2.2
grtcore
grt
WRomp
0.28335
0.28305
0.28325
0.28315
54.0 ± 6.4 Ma
0.28335
0.28305
0.28325
0.28315 grt
WR
grt
46.1 ± 2.2 MaMSWD = 1.3
WR
0.28335
0.28305
0.28325
0.28315
omp17
6 Hf/1
77H
f
grt
49.6 ± 2.5 MaMSWD = 0.68
51.7 ± 2.1 MaMSWD = 1.04
WRomp
glc
WRomp
grt
0.28335
0.28305
0.28325
0.28315
0.28335
0.28305
0.28325
0.28315
176Lu/177Hf
0.04 0.12 0.20 0.28
Lu-Hf
20 ± 14 MaMSWD = 2.4
grt
omp
WR
glc
leachedgrt
78 ± 25 Ma
grtWR
58 ± 17 Ma
grt
WR
44 ± 16 MaMSWD = 1.7
grt
WRomp
147Sm/144Nd
P-02 Pfulwe
P-98 Pfulwe
P-80b Pfulwe
P-80c Pfulwe
SF-26 Saas Fee
SF-25bSaas Fee
CH-48 Chamois
SJ-87 St. Jacques
P-100 Pfulwe
65 ± 37 Ma
0.1 0.3 0.5 0.7
grtWR
40.7 ± 1.8 Ma
grt
WR
38.1 ± 2.7 Ma
grt
WR
52.6 ± 1.7 MaMSDW = 1.5
grt
WRomp
40.3 ± 5.9 MaMSWD = 0.18
grtWR
omp
176Lu/177Hf
0.04 0.12 0.20 0.28
39.2 ± 1.6 MaMSWD = 0.20
grt
WRomp
P-02 Pfulwe
P-98 Pfulwe
P-80b Pfulwe
P-80c Pfulwe
SF-26 Saas Fee
SF-25b Saas Fee
CH-48 Chamois
SJ-87 St. Jacques
P-100 Pfulwe
no Sm-Nd data availablefor sample P-96 (Pfulwe)
(l)
(c)
(d)
(a)
(b)
(h)
(g)
(f)
(i)
(e)
(m)
(n)
(j)
(k)
(r)
(q)
(p)
(s)
(o)
Fig. 7. (a–i) Sm–Nd and (j–s) Lu–Hf isochrons for Alpine
eclogites. Data are from Tables 2 and 3. Note that all isochrons
are atthe same scale for Sm-Nd and Lu-Hf respectively. grt, garnet;
WR, whole rock; omp, omphacite; glc, glaucophane.
© 2015 John Wiley & Sons Ltd
GEOCHRONOLOGY OF THE ZERMATT -SAAS OPHIOL I TE 721
-
age = 3.36 � 0.05; calculations are based on 11 oxy-gen),
indicating moderately high celadonite contents(Table S2). Saas Fee
samples display generallylower average phengite contents and larger
variation(Si range = 3.05–3.34 apfu; average = 3.24 �
0.09).Paragonite contents in all samples are low, varyingbetween
~0.04 and 0.08 apfu. Phengite in the piemon-tite-bearing sample
08ES-03 has measurable Mn con-tents (0.3 � 0.1 wt%). Calcium
contents are alwayslow, often below detection. All phengite in
themetasedimentary rocks at Lago di Cignana and SaasFee is
interpreted to reflect eclogite to upper green-schist facies P–T
conditions.
Sm–Nd and Lu–Hf geochronology – eclogites
The Sm–Nd and Lu–Hf results obtained on the exactsame samples
and dissolutions are compared inFig. 7a–s, and it is immediately
apparent that virtu-ally none of the Sm–Nd ages (Fig. 7a–i)
produceduseful isochron ages due to very low measured147Sm/144Nd
ratios for garnet, which in turn pro-duced ages of very low
precision. Errors in Sm–Ndages range from 14 to 320 Ma. In contrast
to thegenerally unsuccessful Sm–Nd geochronologicalresults from the
ZSF, most samples produced Lu–Hfisochrons that are geologically
meaningful (Fig. 7j–s).
We note that although some samples produced ageswith relatively
high uncertainties, sequential dissolu-tion methods were not used
for Lu–Hf garnetanalyses for reasons detailed in the methods
section.Lu–Hf isochron precision is proportional to the Lu/Hf
ratios that were obtained on garnet, as expected.At Pfulwe, four
samples produced Lu–Hf ages thathad uncertainties of
-
ages, a second age group of ‘40 Ma’ is defined. Abroadly similar
age range (c. 45–42 Ma) wasobtained by Herwartz et al. (2008) for
the upperValle di Gressonay.
Rb–Sr geochronology – metasedimentary rocks
Rb–Sr isochrons for the metasedimentary rocks arelargely
controlled by phengite, the highest Rb/Sr min-eral in the samples
(Table 4). The intercepts are con-trolled by carbonate and an
epidote-group mineral.WR analyses provide an assessment of
isochronintegrity and may identify open-system behaviour.Isochron
ages that include the WR for Lago di Cig-nana samples have large
errors. Phengite–clinozoisite(01NM-45) and phengite–piemontite
(sample 08ES-03) ages at Lago di Cignana, however, have
relativelysmall errors of 41.6 � 0.2 Ma (Fig. 8a) and39.8 � 0.2 Ma
(Fig. 8b) respectively. Phengite–(clino)zoisite ages for the Saas
Fee samples are39.1 � 0.3 Ma (05NM-212; Fig. 8c), 38.7 � 0.2
Ma(05NM-214; Fig. 8d), 39.4 � 0.2 Ma (05NM-215;Fig. 8e) and 40.6 �
0.2 Ma (05NM-219; Fig. 8f).The weighted average of these samples
is39.7 � 0.2 Ma. All these ages are indistinguishablewith isochron
ages constructed using all components,including whole rocks (Table
6). Although very smallage uncertainties are reported here (0.3 Ma
or less,
based on analytical 87Rb/86Sr and 87Sr/86Sr errors),the
uncertainty in the geological age interpretation iscertainly much
larger; this will be important to con-sider in the discussion
section. Collectively, all of theRb–Sr phengite ages are close to
the 40 Ma agegroup, as defined in the previous section for
eclogites.The initial 87Sr/86Sr ratios are all high,
>0.710(Table 4), consistent with significant sourcing of
themetasedimentary rocks within the ZSF from conti-nental basement,
as reflected by the local graniticnappes (Mahlen et al., 2005).
DISCUSSION
Below, the discussion commences with some impor-tant issues
related to Rb–Sr, Sm–Nd and Lu–Hf iso-chron geochronology, such as
isotopic closure(Dodson, 1973) and equilibrium assemblages, and
thevery high errors obtained for the Sm–Nd techniqueare addressed
in the context of REE abundances ingarnet. This is followed by a
short explanation onhow Lu–Hf and Rb–Sr ages need to be
interpretedbased on REE zoning pattern and other
petrologicalconsiderations. Individual Lu–Hf and Rb–Sr ages arethen
placed in the context of previously publisheddata, as well as their
structural position in the alpinestack. Lastly, the implications of
the new data for thesubduction history of the ZSF unit are
discussed.
Assessment of Rb–Sr, Sm–Nd and Lu–Hf geochronology
Geochronological investigations aimed at determiningthe P–T path
of (U)HP terranes require rocks thatdid not exceed the closure
temperatures of appropri-ate geochronometers for significant
periods of time.In addition, meaningful geochronology
requireslithologies that do not contain inherited components(e.g.
Scherer et al., 2000), as well as samples thatexhibit minimal
retrograde overprints. Retrogradegarnet resorption is especially
problematic for Lu–Hfgeochronology (Kelly et al., 2011). The
eclogites ofthe ZSF unit, however, satisfy all of the above
crite-
Table 5. Sm–Nd and Lu–Hf ages (Ma) for Alpine eclogites.
Sample Location Sm–Nd Age Initial 143Nd/144Nd 2SE eNd Lu–Hf Age
Initial176Hf/177Hf 2SE eHf
P-80b Pfulwe 20 � 21 0.513045 �24 7.9 49.6 � 2.5 0.283061 �5
10.8P-80c Pfulwe 43 � 43 0.513021 �47 7.5 51.7 � 2.1 0.283063 �4
11.0P-96 Pfulwe – – – – 46.1 � 2.2 0.283076 �5 11.3P-02 Pfulwe 97 �
320 0.51296 �36 6.2 46.7 � 2.1 0.283075 �4 11.2P-98 Pfulwe �10 � 51
0.513069 �58 8.4 54.0 � 6.4 0.283078 �9 11.6P-100 Pfulwe 44 � 16
0.513006 �18 7.2 40.3 � 5.9 0.283076 �5 11.296JA-32 Lago di Cignana
40.6 � 2.6a 0.513046 �11 8.0 48.8 � 2.1b 0.283101 �5 12.3CH-48
Chamois 58 � 17 0.512990 �22 6.9 52.6 � 1.7 0.283079 �3 11.5SF-25b
Saas Fee 78 � 25 0.512989 �38 8.1 38.1 � 2.7 0.283095 �4 12.2SF-26
Saas Fee 20 � 14 0.513054 �19 6.9 40.7 � 1.8 0.283103 �6 11.8SJ-87
St. Jacques 65 � 37 0.512996 �52 6.9 39.2 � 1.6 0.283123 �4
12.9
Sm–Nd and Lu–Hf isochrons are calculated with relative
uncertainties of 0.5% for 147Sm/144Nd and 176Lu/177Hf, and the
analytical uncertainties (internal 2SE) for 143Nd/144Nd
and176Hf/177Hf as given in Tables 2 and 3.aAmato et al.
(1999).bLapen et al. (2003).
Table 6. Rb–Sr ages (Ma) for Alpine metasedimentary rocks.
Sample Location Rb–Sr agea Rb–Sr ageb Initial 87Sr/86Src 2SE
01NM-45 Lago di Cignana 41 � 19 41.6 � 0.2 0.711079 �1508ES-03
Lago di Cignana 39.6 � 8.4 39.8 � 0.2 0.710033 �1005NM-212 Saas Fee
40.3 � 3.8 39.1 � 0.3 0.711512 �1305NM-214 Saas Fee 38.5 � 1.2 38.7
� 0.2 0.710166 �905NM-215 Saas Fee 39.2 � 0.9 39.4 � 0.2 0.710002
�1305NM-219 Saas Fee 40.8 � 0.7 40.6 � 0.2 0.709470 �13
Rb–Sr isochrons are calculated with relative uncertainties of
0.5% for 87Rb/86Sr, and theanalytical uncertainties (internal 2SE)
for 87Sr/86Sr as given in Table 4.aAge calculated using the WR and
all mineral fractions.bAge calculated using phengite–(clino)zoisite
only.cGiven are initials and 2SE that are calculated from
phengite–(clino)zoisite isochrons;they overlap within error with
calculated intercepts using multiple minerals.
© 2015 John Wiley & Sons Ltd
GEOCHRONOLOGY OF THE ZERMATT -SAAS OPHIOL I TE 723
-
ria for robust geochronology. For example, all sam-ples show
only minor indications of retrogression. Inaddition, the peak
temperatures of the ZSF unit didnot exceed ~600 °C, which is below
the minimumblocking temperatures of both Sm–Nd and Lu–Hfgarnet
geochronometers (e.g. Th€oni & Miller, 1996;Duchêne et al.,
1997a; Ganguly et al., 1998; VanOrman et al., 2002; Tirone et al.,
2005; Skora et al.,2008), and this is confirmed by the preservation
ofprograde REE zoning in garnet (Fig. 6a–c). Hence,Lu–Hf and Sm–Nd
ages should represent timeswithin the prograde-metamorphic
cycle.
The precursors of the eclogites were young, oceanicmaterial that
should contain minimal or no inheritedcomponents nor relict
minerals that would complicatethe isotope systematics. The high eNd
(+6 to +8) andeHf (+10 to +13) values (Table 5) indicate that
theZSF eclogites were derived from depleted mantle andtherefore
should contain no inherited components.This is important because
eclogitic garnet is inclusionrich, which could potentially supply
non-radiogenicHf through, for example, sub-microscopic
zirconinclusions, or inherited Nd from epidote inclusions.Although
zircon was not observed in the samples,LA-ICP-MS data suggest its
presence at sub-micronscale. Such zircon most likely grew during
progrademetamorphism when formerly Zr-rich host mineralsbroke down.
This is corroborated by a publishedaverage U–Pb zircon age from the
Lago di Cignanaunit, which is relatively young (c. 44 Ma, Rubattoet
al., 1998), and only slightly older than the Sm–Ndage determined
from the same area (c. 41 Ma; Amatoet al., 1999). Baxter &
Scherer (2013) show thatalthough inclusions of the same age as the
dated min-eral may decrease the isochron precision, they wouldnot
alter the accuracy of the age. Hence, we concludethat the small
amount of zircon dissolved alongsidegarnet cannot significantly
alter the Lu–Hf agesexcept to decrease precision.
It is commonly assumed that failure of Sm–Ndgarnet geochronology
in terms of spread in147Sm/144Nd ratios reflects the impact of
LREE-richinclusions such as metamorphic epidote, which is
arefractory sink for all LREE, and this has prompteddevelopment of
leaching techniques (e.g. Amatoet al., 1999; Baxter et al., 2002;
Anczkiewicz & Thirl-wall, 2003; Pollington & Baxter, 2011).
The methodof Amato et al. (1999) was adopted here for
selectedsamples (Pfulwe, Saas Fee), given its previous successin
removing the majority of LREE-rich inclusionsfrom the Lago di
Cignana area. It is immediatelyapparent from Fig. 7a–i, however,
that this leachingprocedure was unsuccessful. The measured range
of147Sm/144Nd ratios of leached garnet fractions of~0.21–0.39, with
absolute Nd concentrations as highas 0.18–1.14 ppm (Table 2), is
indicative of contami-nation by LREE-rich inclusions (e.g. Th€oni,
2002;Baxter & Scherer, 2013). A key question is: why werethe
Sm–Nd geochronology attempts at localities
other than at Lago di Cignana so unsuccessful? Oneexplanation is
that the leaching methods are rathersensitive to garnet sizes, the
sizes of inclusions, differ-ent inclusion populations, and the
duration of acidleaching and leaching temperatures, raising the
possi-bility that every sample potentially behaves differ-ently
(e.g. Pollington & Baxter, 2011). Success orfailure in leaching
methods applied to Sm–Nd garnetgeochronology may also be dependent
on absoluteSm–Nd concentrations in garnet. Garnet enriched inSm and
Nd is more resistant to the influence ofLREE-rich inclusions before
147Sm/144Nd ratios aredecreased to levels too low to provide useful
iso-chrons. In contrast, garnet that contains very low Smand Nd
will be much more difficult to use for Sm–Nd geochronology, and
leaching methods may be lesseffective. This is illustrated for the
different examplesin Fig. S1.Measured REE profiles across garnet
indicate that
Sm concentrations in Pfulwe (Fig. 6a), Chamois(Fig. 6b) and Saas
Fee (Fig. 6c) are enriched towardsthe rim, similar to that found in
Lago di Cignana(Skora et al., 2009). Absolute concentrations of
Sm,however, vary significantly among samples fromspecific
localities, as well as between localities, reflect-ing differences
in garnet growth histories. Rim-Smconcentrations are highest at
Lago di Cignana(2.5 ppm, Skora et al., 2009) and Chamois (2
ppm,Fig. 6b), lower in Saas Fee samples (0.8 ppm,Fig. 6c) and
lowest in Pfulwe samples (
-
amphibole), accessory phases (e.g. apatite, titanite)and whole
rocks. Complexities in Rb–Sr isochronsmay reflect differential
diffusional resetting of miner-als, or disturbance of the isotopic
system due tointeractions with late, alkali-rich fluids or
re-crystal-lization (e.g. Giletti, 1991; Jenkin et al., 1995;
K€uhnet al., 2000; de Jong, 2003; di Vincenzo et al., 2006;Glodny
et al., 2008). In this study, all Saas Fee sam-ples have
well-equilibrated phengite–WR–(clino)-zoisite–carbonate isochrons,
and moderate to smalluncertainties. In contrast,
phengite–WR–clinozoisiteand phengite–WR–piemontite ages for Lago di
Cig-nana samples have larger errors (8–19 Ma). This sug-gests that
individual minerals were affected byadditional processes such as
recrystallization, fluid–mineral interaction or partial
re-equilibration at theLago di Cignana (U)HP locality. Similar
observa-tions were also reported by Mahlen et al. (2005) forother
ZSF metasedimentary rocks, where there wasclear evidence for Rb–Sr
mobility during Alpinemetamorphism. Previous studies have found
that pro-grade Sr zoning can be preserved in epidote in UHPterranes
that have recorded significantly higher peak-metamorphic
temperatures (>700 °C; Nagasaki &Enami, 1998). This suggests
that Sr in epidote dif-fuses very slowly at eclogite facies
conditions, andthat epidote-group minerals will likely remain
closedat all times in the ZSF unit. We therefore use
(clino)-zoisite–phengite and piemontite–phengite pairs forthe most
accurate Rb–Sr isochron ages in all samplesin this study.
The isotopic closure temperature for Rb–Sr inphengite is not
well constrained. J€ager et al. (1967)empirically determined a
closure temperature of~500 � 50 °C for phengite-bearing alpine
metamor-phic rocks. This is close to the
peak-metamorphictemperatures of the ZSF unit of ~550–660 °C
(e.g.Oberh€ansli, 1982; Barnicoat & Fry, 1986; Reinecke,1998;
Bucher et al., 2005; Angiboust et al., 2009;Groppo et al., 2009),
suggesting that phengite Rb–Srisochrons either date
peak-metamorphic conditions(eclogite facies) or the onset of
cooling (upper green-schist facies). We note that the availability
of fluidsfor recrystallization, as well as the presence orabsence
of an Rb and Sr exchange partner in thematrix, will also play a
role in isotopic resetting ofphengite at peak temperatures (e.g.
Giletti, 1991; Jen-kin et al., 1995; K€uhn et al., 2000; Glodny et
al.,2008). The Rb–Sr ages in this study are thereforeinterpreted to
date a period that encompassed eclog-ite to upper greenschist
facies P–T conditions,depending on the individual rock sample and
its P–T-fluid-deformation history.
Garnet geochronology constraints on prograde-metamorphic
history
Lapen et al. (2003) first proposed that the very highLu/Hf
distribution coefficient ratio for garnet should
result in Lu enrichment in early grown cores. Thiswould bias
Lu–Hf garnet ages towards the onset ofgarnet growth, in contrast to
Sm–Nd ages that wouldrepresent a volumetric mean age in the absence
of anysignificant growth zoning. Assuming a Rayleigh modeland batch
nucleation, the estimated duration of pro-grade garnet growth was
c. 12 Ma based on the con-trast in Lu–Hf age (48.8 � 2.1 Ma) and
Sm–Nd age(40.6 � 2.6 Ma) measured on the same sample. Thisled to
the recognition that measured Sm–Nd and Lu–Hf bulk garnet ages
reflect the integrated effects oftheir growth histories (see also
Kohn, 2009).It was subsequently shown that REE profiles of
garnet from the ZSF unit, and some other areasworldwide, do not
follow those predicted by Rayleighfractionation, but rather reflect
diffusion-limitedtransport of REEs, which may produce REE
deple-tion halos around growing garnet (e.g. Skora et al.,2006;
Moore et al., 2013). The study of Skora et al.(2009) showed that
although Lu is enriched in earlygrown cores due to
diffusion-limited transport, Lu–Hf ages do not date the onset of
garnet growth sim-ply because core concentrations are of minor
abun-dance when weighted over a spherical geometry. TheLu–Hf ages
may still be slightly skewed towards theonset of growth, but only
when garnet grows overprolonged periods of time. Thus,
interpretation ofLu–Hf ages are a complex matter as they can
datenear-peak metamorphism, if garnet grew over a shortinterval (10
Ma). If thisis not the case, then Lu–Hf and Sm–Nd ages may be
similardespite enrichment of Lu in the early grown cores.
© 2015 John Wiley & Sons Ltd
GEOCHRONOLOGY OF THE ZERMATT -SAAS OPHIOL I TE 725
-
reflect an earlier, prograde-metamorphic age if garnetgrowth was
protracted.
Diffusion-limited transport also explains Smenrichments in the
late-grown rims (Fig. 6a–c),reflecting relaxation of depletion
halos in the matrixand faster diffusion rates at peak- or near-peak
con-ditions. With the recognition that the Sm/Nd ratiosare highest
in the rims of garnet in the Lago di Cig-nana sample dated by Sm–Nd
(Amato et al., 1999),garnet growth modelling in Skora et al. (2009)
pro-vides strong support for an age of c. 40 Ma for peakUHP
metamorphism at Lago di Cignana. Note thatthis estimate is strongly
dependent upon the confi-dence of the Sm–Nd age, which has a
reported preci-sion of 2.6 Ma, and this uncertainty assumes thatany
surviving LREE-rich inclusions that remainedafter leaching did not
influence the age accuracy (e.g.prograde epidote is of roughly
similar age as garnet).The age uncertainty noted by Amato et al.
(1999)permits an integrated Sm–Nd age as old as 43.2 Ma,which would
place the end of garnet growth andpeak metamorphism at c. 42 Ma. In
contrast, if thetrue Sm–Nd age at Lago di Cignana lay towards
theyounger end of the uncertainty, peak metamorphismcould be
younger than 40–38 Ma.
Garnet crystal-size distributions (CSD) will alsoexert an
important control on bulk garnet ages(Skora et al., 2009).
Differently sized garnet grainsare typically interpreted to have
grown during a pro-grade-metamorphic event, nucleating at
differenttimes (continuous nucleation and growth). Large gar-net
reflects early nucleation and the entire garnetgrowth history,
whereas the smallest ones record latenucleation and growth towards
the end of the pro-grade path (e.g. Jones & Galwey, 1966;
Kretz, 1966;Cashman & Ferry, 1988). These studies also showthat
intermediate to small (late-grown) garnet tend todominate over
large (old) garnet in the population.
Comparison of Lu–Hf and Sm–Nd isochron agesto their respective
model ages, taking into accountREE zoning in garnet and CSDs,
suggests that garnetin the Lago di Cignana area grew over a longer
per-iod (30 � 10 Ma, Skora et al., 2009). Although suchlong
prolonged garnet growth times may sound sur-prising, other studies
have subsequently found com-parable results (e.g. Cheng et al.,
2011). Note that nonew age modelling was performed for this
study,given the similarities of measured REE zoning pat-tern in all
ZSF garnet. The essence of the quantita-tive modelling of Skora et
al. (2009) is thusreproduced in Fig. 9, and our interpretation of
themetamorphic history of the ZSF is based on theserelations.
Constraints on peak metamorphism and initial uplift
The relations between model ages and garnet growthduration shown
in Fig. 9 highlight the importance ofknowing the age of peak
metamorphism and hence
cessation of garnet growth when interpreting Sm–Ndand Lu–Hf
garnet isochron ages. Rb–Sr phengiteages provide important
constraints on the transitionfrom eclogite to upper greenschist
facies P–T condi-tions. The average of four samples from Saas Fee
is39.5 � 0.1 Ma, whereas the average Rb–Sr age forthe two samples
from Lago di Cignana is40.7 Ma � 0.1 Ma. Another published Rb–Sr
ageusing phengite–WR from Lago di Cignana is slightlyyounger at
38.5 � 0.1 Ma, using the decay constantof Rotenburg et al. (2012;
original published age ofAmato et al., 1999; 37.9 � 0.1 Ma). The
Rb–Sr agesfor other ZSF samples span a large range of c. 48–37 Ma
(e.g. Barnicoat et al., 1995; Amato et al.,1999; Reddy et al.,
1999; Cartwright & Barnicoat,2002). This large age range is in
part related to thefact that varying mineral assemblages were used
forgeochronology, and different studies selected theirrocks for
different purposes, including samples thatare heavily overprinted.
If, however, comparisons arerestricted to Rb–Sr ages of
epidote–phengite pairs,the results converge to the same range
observed inthis study. Recalculating the epidote–phengite
Rb–Srisochron ages of Barnicoat et al. (1995) for theT€aschalp
area, using Rotenburg et al. (2012), pro-duces a range of 43.0–40.1
Ma with a weighted aver-age of 41.4 � 0.5 (published ages using old
decayconstant: 42.3–39.5 Ma). This area is spatially closeto Pfulwe
(~1–2 km NNE of Pfulwe). Anotheryounger Rb–Sr age for T€aschalp is
reported inAmato et al. (1999). They proposed an age of38.6 � 0.1
Ma (original ages using old decay con-stant: 38.0 � 0.1 Ma) based
on a phengite–WR pair.Last, Reddy et al. (1999) published an
epidote–phen-gite age of 41.2 � 0.6 Ma (original age using olddecay
constant: 40.5 � 0.6 Ma) for the Val d’Ayas.The sample location is
~5 km north of St. Jacques.Hence, all previously published Rb–Sr
ages are prin-cipally consistent with the results obtained in
thisstudy.Although Rb–Sr ages can be difficult to relate with
the specific P–T conditions, a pattern emerges for thedata
presented here. The age distributions at Lago diCignana (c. 41–38.5
Ma) and T€aschalp (c. 41–38.6 Ma) are consistent with the
interpretation thatRb–Sr may effectively record a time period
thatencompasses peak eclogite to post-peak upper green-schist
facies P–T conditions. The oldest Rb–Sr ages(c. 43–40 Ma) fully
overlap with what is interpreted aspeak metamorphism in these
structurally highest units(e.g. with the Sm–Nd age of Lago di
Cignana).Younger ages in the order of c. 38.5 Ma are just out-side
the error bracket for peak ages, and are thereforeinterpreted to
reflect a fluid event during exhumationthrough upper greenschist
facies. Average Rb–Sr agesin Saas Fee samples closely overlap with
average Lu–Hf ages at c. 39.4 Ma. This would suggest that SaasFee
samples are slightly younger in terms of peakmetamorphism (c. 39
Ma) than samples from Pfulwe,
© 2015 John Wiley & Sons Ltd
726 S . SKORA ET AL .
-
T€aschalp and Lago di Cignana where Rb–Sr (andSm–Nd) ages are
closer to c. 41 Ma.
Collectively, our results suggest that peak meta-morphism
occurred at c. 42–40 Ma, and uplift andgreenschist facies
metamorphism occurred between c.39 and 38 Ma in the structurally
highest units (e.g.Pfulwe, T€aschalp, Lago di Cignana). Peak
metamor-phism might have been reached c. 1–2 Ma later inthe
structurally lowest unit represented by Saas Fee
rocks (c. 40–39 Ma). The greenschist facies overprintis not
constrained in these samples. Our resultspotentially indicate very
rapid initial uplift fromeclogite to greenschist facies conditions
within analmost irresolvable time interval, perhaps in the orderof
only c. 1 Ma, although possibly as much as 4 Ma.
Evidence for diachronous prograde metamorphism in theZSF
unit
The broad correlation between the two age groupsdefined by the
Lu–Hf garnet geochronology at c.50 Ma (c. 52–46 Ma: Lago di
Cignana, Pfulwe andChamois) and 40 Ma (c. 41–38 Ma: Saas Fee andSt.
Jacques) suggest distinctly different progradegarnet-growth
histories. The age differences could berelated to diachronous
subduction of the ZSF zoneor they could be related to petrological
differences.For example, garnet in different samples could
havestarted nucleating at very different times yet follow-ing
broadly similar P–T–t paths. The latter hypothe-sis might be tested
through garnet inclusionassemblages and bulk chemistry effects on
garnet-inphase relations. For example, the presence of
titaniteinstead of rutile is often associated with lower
P–Tconditions (blueschist instead of eclogite facies;e.g. John et
al., 2011). Hence if garnet started togrow significantly earlier in
one rock compared toanother, one could expect titanite inclusions
to bemore abundant in garnet cores. Titanite-rich garnetcores are
common in Lago di Cignana, Chamois andSaas Fee samples. Garnet
cores that have very lowtitanite abundances are exclusively
characteristic ofPfulwe samples. Inclusion populations therefore
donot correlate with the distinct Lu–Hf age groups,suggesting that
there are no systematic differences ingarnet growth timing relative
to P–T paths. In termsof bulk-rock chemistry, it is well known that
Mncontents, as well as Mg/(Fe + Mg) ratios, exert astrong influence
on the garnet-in phase relationsduring prograde metamorphism (e.g.
Spear, 1993). Itis shown in Fig. 10a,b that the distinct Lu–Hf
agegroups do not correlate with either of these parame-ters.
Although there are some differences in bulk-rock compositions in
terms of CaO and Na2Oconcentrations, likely reflecting various
degrees ofspilitization of the protolith basalts (Table 1), there
isno clear correlation with Lu–Hf ages (Fig. 10c). Wetherefore
conclude that there is no evidence frominclusion populations or
bulk chemical compositionsfor distinctly different prograde garnet
growth histo-ries relative to P–T conditions.Previous modelling of
Lu–Hf garnet ages has
demonstrated that the core-rim Lu distribution pat-terns can
have a strong influence on the Lu–Hf agemeasured for bulk garnet
(Lapen et al., 2003; Kohn,2009; Skora et al., 2009). All Lu zoning
profiles dis-play narrow central peaks that are consistent
withdiffusion-limited Lu uptake (Fig. 6a–c). The minor
MnO
[wt%
]
0.11
0.13
0.15
0.17
0.19
MgO
(MgO
+FeO
)
0.25
0.30
0.35
0.40
Age [Ma]35 5040 45
CaO
/Na 2
O
1.0
2.0
3.0
4.0
(a)
(b)
(c)
Fig. 10. Whole rock v. age relations, showing that there is
nosimple correlation between (a) Mn, (b) Mg/(Mg + Fe), whichare
parameters that are known to influence when the garnet-inreaction
is crossed during prograde metamorphism. (c) Thereis also no
obvious correlation with Ca/Na, which characterizesthe different
degrees of ocean–floor spilitization. These resultsshow that the
different garnet growth intervals do not reflectsystematic changes
in bulk chemical compositions.
© 2015 John Wiley & Sons Ltd
GEOCHRONOLOGY OF THE ZERMATT -SAAS OPHIOL I TE 727
-
differences between samples, such as the extent of Lushoulders
towards the rim, cannot account for majorLu–Hf age differences.
Moreover, all samples appearto contain a garnet population that is
consistent witha single continuous nucleation and growth event
dur-ing prograde metamorphism, as opposed to, forexample, bimodal
garnet size distributions that arecommonly found in the Central
Alps (e.g. Herwartzet al., 2011). The only difference is that
Pfulwe sam-ples have, on average, significantly fewer but
largergarnet crystals compared to all other samples, butthis
distinction does not correlate with the differentLu–Hf age groups.
We therefore rule out distinct dif-ferences in Lu profiles or
garnet populations as anexplanation for the distinct Lu–Hf age
groups.
Below we summarize the age groups and their link-ages to
structural positions in the nappe stack. All theage data discussed
below are summarized in Fig. 11.
Pfulwe, Lago di Cignana and Chamois samples
Lu–Hf ages from four samples from a single outcropjust below the
Pfulwe pass are relatively well
constrained (uncertainties: 2.1–2.5 Ma), and rangefrom 51.7 to
46.1 Ma. This is very similar to the Lu–Hf age obtained by Lapen et
al. (2003) for Lago diCignana (48.8 � 2.1 Ma). The Lu–Hf age of
Cha-mois sample CH-48 is also well constrained to be52.6 � 1.7 Ma,
again falling into the 50 Ma age cate-gory. Structurally, all these
locations reflect the topof the exhumed ZSF sequence, beneath the
Tsat�enappe, which is thought to represent the accretionaryprism
below which the ZSF unit was subducted.Peak-metamorphic constraints
are provided by Sm–Nd and Rb–Sr ages for Lago di Cignana (also
closeto Chamois) and T€aschalp (close to Pfulwe). They allroughly
fall within an age bracket of c. 42–40 Ma.The similarities in ages,
zoning pattern and CSD sug-gest that these units have shared the
same tectonichistory. It is important to stress that as in the
Lu–Hfage at Lago di Cignana, the disparate Lu–Hf
andpeak-metamorphic ages require a prolonged garnetgrowth interval
of 30 � 10 Ma (Fig. 9). This hasbeen further discussed in Skora et
al. (2009).
Saas Fee samples
Two samples have yielded considerably younger Lu–Hf ages (c.
41–38 Ma), overlapping with Rb–Sr agesobtained from four
metasedimentary samples fromthe same area (average: 39.5 � 0.1 Ma).
The samplescome from an area that lies at the structurally
lowestposition in the obducted ZSF unit, directly above theMonte
Rosa nappe, which reflects the distal part ofthe European
continent. The age modelling (Fig. 9)suggests that this is best
explained by a very shortgarnet growth period (
-
place them in the 40 Ma age group. Trace-elementzoning data and
a CSD are not available, but petro-logical descriptions of the
samples are close to thatobserved in other locations. If a
peak-metamorphicage of 40 Ma is assumed for this area, similar to
allother areas, then it is clear that garnet must havegrown for a
relatively short period of time. This isstrikingly similar to that
found for the Saas Fee sam-ples. Because the upper Valle di
Gressonay area alsojust overlies the Monte Rosa nappe, this places
thesesamples into a similarly low structural position as theSaas
Fee samples.
Implications for the ZSF and related units
The discussion here, which builds on the modellingof Skora et
al. (2009), that Lu–Hf ages date differentpoints in time, either
near-peak metamorphism or atime that reflects integrated prograde
growth, pro-vides an alternative interpretive context for
Lu–Hfgeochronology studies of other units in the region.Herwartz et
al. (2008) studied the ‘Balma unit’ in theupper Valle di Gressonay,
and they could not recon-cile their c. 42 Ma Lu–Hf age with a
peak-metamor-phic age of c. 40 Ma of the ZSF unit, which ledthem to
suggest that the ‘Balma unit’ originated inthe Valais trough. Our
new results document youngerLu–Hf ages (41–38 Ma: Saas Fee) in
samples fromthe lower ZSF, a position that is structurally
similarto rocks studied by Herwartz et al. (2008) in thatboth just
overly the Monte Rosa unit. These relationssuggest that Lu–Hf age
information must be inter-preted within the context of the
structural and meta-morphic histories of the units as determined
bymultiple geochronometers and P–T constraints. Wetherefore suggest
that it is not appropriate to inferthat the upper Valle di
Gressonay must have origi-nated in the Valais based on the Lu–Hf
ages alone.In fact, the originally proposed paleoposition of theZSF
unit (Bearth, 1967) is very well reconciled withour new
interpretation that the lower ZSF units hadprograde-metamorphic
histories that were distinctfrom those of the upper ZSF.
The results presented here suggest that the dis-parate ages for
peak metamorphism reflect the pro-tracted nature of prograde
metamorphism, as well asthe diachronous subduction that occurred
across theorogen. The relatively young age of the Saas Fee(and
upper Valle di Gressonay) samples in terms ofinitiation of garnet
growth suggests that the lowersection of the ZSF unit may have been
subducted lastwhen compared with the Pfulwe, Lago di Cignanaand
Chamois areas, all of which require a prolongedprograde garnet
growth history. This suggests twodifferent origins: the Pfulwe,
Lago di Cignana andChamois areas must have subducted early,
placingthem close to the African continental margin. In con-trast,
the Saas Fee and the ‘Balma unit’ must haveoriginated much closer
to the European continental
margin, near the Monte Rosa unit, and was sub-ducted
significantly later (Fig. 12b). This is differentfrom current
estimates for the paleopositions in theWestern Alps, where the
entire ZSF unit is placed inthe northern realm (Fig. 12a, e.g.
Labhart, 1992). Wenote that our interpretation requires a change in
thepaleogeographic origin of the Tsat�e nappe, whichmust be placed
towards the northern realm of theLiguro-Piemont Ocean, close to the
European conti-nental margin (Fig. 12b). This is required to
explainthe relatively young Rb–Sr ages (range: c. 44–37 Ma;without
distinction in the different minerals used forisochrons, Reddy et
al., 1999) in the light of thegreenschist to lower blueschist
facies rocks of theTsat�e nappe that are unlikely to have exceeded
theirblocking temperature of ~500 � 50 °C.Reconstruction of the
relative movements of the
Apulian/African plate and the European plate sug-gests an
initially strong sinistral movement, followedby a rotational
movement and subsequently a moreperpendicular subduction (Fig.
13a,b, see also e.g.Dewey et al., 1989; Rosenbaum et al., 2002;
Rosen-baum & Lister, 2005). Average subduction rates of0.3–1.0
cm yr�1 were estimated to explain the pro-grade history of the Lago
di Cignana area (Skoraet al., 2009), which lies within the
estimated range of0.4–1.5 cm yr�1 that is based on independent
paleo-magnetic studies (e.g. Dewey et al., 1989; Schmidtet al.,
1997). The initially oblique subduction shouldresult in slow
subduction rates, reflecting a small per-pendicular subduction
vector, and slow garnetgrowth. This is exactly what is proposed for
the Lagodi Cignana, Pfulwe and Chamois areas, which weresubducted
early based on the geochronological resultsobtained here. The
protracted garnet growth periods,in turn, explain the disparate
Lu–Hf, Sm–Nd andRb–Sr isochron ages. In contrast, oceanic crust
thatwas originally located close to the European margin
Traditional position of the remnants of the Piemont-Ligurian
ocean
NW SE
SesiaBriannconnais(e.g. Monte Rosa)
Zermatt-Saas Feehigh P remnants
Tsaté nappe low P remnants
Reinterpretation of the remnants of the Piemont-Ligurian
ocean
NW SE
SesiaBriannconnais(e.g. Monte Rosa)
Zermatt-Saas Feehigh P remnants
Tsaté nappe low P remnants
(b)
(a)
Fig. 12. (a) Schematic illustration of the
traditionalpaleoposition of the Zermatt-Saas Fee zone (white
circles) andTsat�e nappe (white crosses). (b) Schematic
illustration of there-interpretation that is based on geochronology
in this study.
© 2015 John Wiley & Sons Ltd
GEOCHRONOLOGY OF THE ZERMATT -SAAS OPHIOL I TE 729
-
would have subducted last. Plate motion vectorstowards the end
of subduction were almost perpen-dicular, which would have led to
fast convergencerates and hence shorter garnet growth intervals
forthe Saas Fee and the upper Valle di Gressonay areas.Finally, our
model may be accommodated in the agerange and thermal structure
inferred for the Liguro-Piemont basin, which initially formed c.
165–160 Ma(Rubatto et al., 1998; Schaltegger et al., 2002) andhad
started to close at c. 110–90 Ma (e.g. Deweyet al., 1989). We
speculate that the very young age ofthe oceanic crust that was
subducted, combined withits small size, reflects alpine subduction
that was dri-ven by compression from the different plate
motions,rather than slab pull.
The ZSF peak age estimates are reasonably similarto a U–Pb
rutile age, obtained from an eclogite faciestension crack in the
Monte Rosa unit(42.6 � 0.6 Ma, Lapen et al., 2007). This
suggeststhat the Monte Rosa and the ZSF zone were partlycoupled
during prograde metamorphism and uplift(Lapen et al., 2007). In
addition to the overlap inpeak-metamorphic ages of the Monte Rosa
unit withthose determined for the ZSF unit, there is an addi-tional
overlap of the greenschist to amphibolite faciesoverprint in both
units, which has been estimated tohave occurred between c. 40 and
32 Ma. It is possiblethat both units closely shared paleogeographic
posi-tions during basin closure, and subsequent exhuma-tion.
Coupling of the granitic nappes of GranParadiso and Monte Rosa to
the Western Alps ophi-olites during eclogite facies metamorphism
could pro-vide an important buoyancy force, in addition to
theserpentinite units, for rapid exhumation of the high-density
eclogites.
CONCLUSIONS
New Lu–Hf geochronology on garnet, in concertwith Rb–Sr and
Sm–Nd geochronology, indicates
that peak-metamorphic conditions from the WesternAlps define
distinct age groups of 50 and 40 Ma. It isnot possible to correlate
the disparate ages withpetrological differences such as REE zoning
thatmight suggest different paragenesis. Instead, it isshown that
the age record is a consequence of differ-ent origins within the
Liguro-Piemont realm, com-bined with diachronous subduction and
significantlydifferent prograde garnet growth intervals. The
ZSFunit is suggested to reflect slivers of oceanic crust
thatstarted subducting at different moments in time. Theindividual
slivers nevertheless appear to have had abroadly common
peak-metamorphic age of c. 42–39 Ma. Areas that have the oldest
Lu–Hf ages(50 Ma age group: Pfulwe, Lago di Cignana andChamois)
represent the structurally highest subunitsin the exhumed ZSF unit.
They are adjacent to theoverlaying Cime Blanche and Tsat�e units,
which inturn are overlain by remnants of the African conti-nental
margin. Because their inferred prograde garnetgrowth interval is
>20 Ma, close to that permitted byplate tectonic
reconstructions, these areas must haveoriginated at the
southernmost realm of the Liguro-Piemont Ocean, close to the
African continental mar-gin. The early onset of garnet growth
inferred for thisgroup indicates that these units were subducted
first.The upper Valle di Gressonay and the Saas Fee unitthat have
younger Lu–Hf ages (40 Ma age group)are structurally in similar
positions in that they bothdirectly overly the Monte Rosa unit
(Briannconais).Hence their paleoposition was likely the
northern-most realm of the Liguro-Piemont Ocean, where itwas
subducted as the last piece of the PiemontOcean. The distinctly
different age groups andinferred prograde garnet growth intervals
are wellexplained by rotation of plate motion of the Liguro-Piemont
Ocean with time. An initially oblique sub-duction resulted in slow
burial rates and long garnetgrowth durations. Subduction then
gradually turnedtowards a near-perpendicular orientation, resulting
in
Plate motion trajectories
42 Ma
81 MaSmith, 1971
80 Ma
63 Ma
Dewey et al., 1973
95 Ma
70 Ma
Livermore & Smith, 1985 Rosenbaum et al., 2002
60 Ma
110 Ma
Dewey et al., 1989
92 Ma
83 Ma
46 Ma
Europe
Africa/ Apulia
Alp-
Teth
ys
Vardar
ValaisBrianconnais
Paleogeographic reconstruction beforeinitiation of
subduction
(a) (b)
Fig. 13. (a) Paleogeographic reconstruction of the Alpine area
in the Early Cretaceous, before the initiation of subduction
(afterStampfli et al., 2002). (b) Reconstruction of the relative
movements of the Apulian/African against the European plate
suggests aninitially strong sinistral movement, which should result
in initially slow subduction rates (Smith, 1971; Dewey et al.,
1973;Livermore & Smith, 1985; Dewey et al., 1989; Rosenbaum et
al., 2002). This is followed by a rotational movement
andsubsequently a more perpendicular subduction. This study does
not differentiate between the different proposed models for
platemotion, but finds that the geochronological results are fully
consistent with rapid convergence rates near the end of
subductionbased on ages from garnet and phengite in structurally
lowest units (e.g. Saas Fee).
© 2015 John Wiley & Sons Ltd
730 S . SKORA ET AL .
-
much faster burial, as well as rapid garnet growthrates in the
subducted slab.
ACKNOWLEDGEMENTS
Financial support provided by the Swiss NationalScience
Foundation grant SNF2100–066996(L. Baumgartner) and by the
Soci�et�e Acad�emiqueVaudoise (S. Skora) is gratefully
acknowledged. Sup-port was also provided by the U.S. National
ScienceFoundation grant EAR-0309853 to C. Johnson andB. Beard. S.
Skora acknowledges a SNF Ambizonegrant (PZ00P2_142575). The
constructive reviews ofE. Baxter and two anonymous reviewers, as
well asthe editorial handling of D. Robinson, were veryhelpful and
are greatly acknowledged.
REFERENCES
Amato, J.M., Johnson, C.M., Baumgartner, L.P. & Beard,B.L.,
1999. Rapid exhumation of the Zermatt-Saas ophiolitededuced from
high-precision Sm-Nd and Rb-Sr geochronol-ogy. Earth and Planetary
Science Letters, 171, 425–438.
Anczkiewicz, M.F. & Thirlwall, M.F., 2003. Improving
precisionof the Sm–Nd garnet dating by H2SO4 leaching – a
simplesolution to phosphate inclusions problem. In:
Geochronology:Linking the Isotopic Record with Petrology and
Textures (edsVance, D., M€uller, W. & Villa, I.M.), pp. 83–91.
GeologicalSociety London Special Publications, London.
Anczkiewicz, R., Szczepanski, J., Mazur, S. et al., 2007.
Lu-Hfgeochronology and trace element distribution in
garnet:implications for uplift and exhumation of ultra-high
pressuregranulites in the Sudetes, SW Poland. Lithos, 95,
363–380.
Angiboust, S., Agard, P., Jolivet, L. & Beyssac, O., 2009.
TheZermatt-Saas ophiolite: the largest (60 km-wide) and deepest(c.
70-80 km) continuous slice of oceanic lithospheredetached from a
subduction zone? Terra Nova, 21, 171–180.
Barnicoat, A.C., 1988. Zoned high-pressure assemblages in
Pil-low lavas of the Zermatt-Saas ophiolite zone,
Switzerland.Lithos, 21, 227–236.
Barnicoat, A.C. & Fry, N., 1986. High-pressure
metamorphismof the Zermatt-Saas ophiolite zone, Switzerland.
Journal ofthe Geologic Society of London, 143, 607–618.
Barnicoat, A.C., Rex, D.C., Guise, P.G. & Cliff, R.A.,
1995.The timing of and nature of greenschist facies deformationand
metamorphism in the upper Pennine Alps. Tectonics, 14,279–293.
Baxter, E.F. & Scherer, E.E., 2013. Garnet
geochronology:timekeeper of tectonometamorphic processes. Elements,
9,433–438.
Baxter, E.F., Ague, J.J. & DePaolo, D.J., 2002. Prograde
tem-perature–time evolution in the Barrovian type-locality
con-strained by Sm/Nd garnet ages from Glen Clova, Scotland.Journal
of the Geological Society, 159, 71–82.
Bearth, P., 1967. Die Ophiolithe der Zone von Zermatt-SaasFee.
Beitr€age zur Geologischen Karte der Schweiz, 132, 130.
Brouwer, F.M., Burri, T., Engi, M. & Berger, A., 2005.
Eclog-ite relics in the Central Alps: PT – evolution, Lu-Hf agesand
implications for formation of tectonic m�elange zones.Schweizer
Mineralogische und Petrographische Mitteilungen,85, 147–174.
Bucher, K. & Grapes, R., 2009. The eclogite-facies Allalin
gab-bro of the Zermatt-Saas Ophiolite, Western Alps: a recordof
subduction zone hydration. Journal of Petrology, 50,1405–1442.
Bucher, K., Fazis, Y., de Capitani, C. & Grapes, R.,
2005.Blueschists, eclogites, and decompression assemblages of
the
Zermatt-Saas ophiolite: high-pressure metamorphism of sub-ducted
Tethys lithosphere. American Mineralogist, 90, 821–835.
Cartwright, I. & Barnicoat, A.C., 2002. Petrology,
geochronol-ogy, and tectonics of shear zones in the Zermatt-Saas
andCombin zones of the Western Alps. Journal of MetamorphicGeology,
20, 263–281.
Cashman, K.V. & Ferry, J.M., 1988. Crystal size
distribution(CSD) in rocks and the kinetics and dynamics of
crystalliza-tion. 3. Metamorphic crystallization. Contributions to
Miner-alogy and Petrology, 99, 401–415.
Cheng, H., Vervoort, J.D., Li, X., Zhang, C., Li, Q. &
Zheng,S., 2011. The growth interval of garnet in the UHP
eclogitesfrom the Dabie orogen, China. American Mineralogist,
96,1300–1307.
Chinner, G.A. & Dixon, J.E., 1973. Some high-pressure
para-geneses of Allalin Gabbro, Valais, Switzerland. Journal
ofPetrology, 14, 185–202.
Chopin, C., 2003. Ultrahigh-pressure metamorphism:
tracingcontinental crust into the mantle. Earth and
PlanetaryScience Letters, 212, 1–14.
Cloos, M., 1993. Lithospheric buoyancy and collisional
oroge-nesis; subduction of oceanic plateaus, continental
margins,island arcs, spreading ridges, and seamounts.
GeologicalSociety of America Bulletin, 105, 715–737.
Coleman, R.G. & Wang, X., 1995. Ultrahigh Pressure
Meta-morphism. Cambridge University Press, Cambridge, 528 pp.
Dal Piaz, G.V. & Ernst, W.G., 1978. Areal geology and
petrol-ogy of eclogites and associated metabasites of the
Piemonteophiolite nappe, Breuil-St. Jacques area, Italian
WesternAlps. Tectonophysics, 51, 99–126.
Dal Piaz, G.V., Cortiana, G., Del Moro, A., Martin, S.,
Pen-nacchioni, G. & Tartarotti, P., 2001. Tertiary age and
pale-ostructural inferences of the eclogitic imprint in
theAustroalpine outliers and Zermatt-Saas ophiolite, westernAlps.
International Journal of Earth Sciences, 90, 668–684.
Dewey, J.F., Pitman, W.C., Ryan, W.B.F. & Bonnin, J.,
1973.Plate tectonics and the evolution of the Alpine systems.
Geo-logical Society of America Bulletin, 84, 137–180.
Dewey, J.F., Helman, M.L., Turco, E., Hutton, D.H.W. &Knott,
S.D., 1989. Kinematics in the western Mediterranean.In: Kinematics
in the Western Mediterranean (eds Coward,M.P., Dietrich, D. &
Park, R.G.), pp. 265–283. GeologicalSociety London Special
Publications, London.
Dodson, M.H., 1973. Closure temperature in
coolinggeochronological and petrological systems. Contributions
toMineralogy and Petrology, 40, 259–274.
Dragovic, B., Samanta, L.M., Baxter, E.F. & Selverstone,
J.,2012. Using garnet to constrain the duration and rate
ofwater-releasing metamorphic reactions during subduction:
anexample from Sifnos, Greece. Chemical Geology, 314, 9–22.
Duchêne, S., Blichert-Toft, J., Luais, B., T�elouk, P.,
Lardeaux,J.M. & Albar�ede, F., 1997a. The Lu-Hf dating of
garnetsand the ages of the Alpine high-pressure metamorphism.
Na-ture, 387, 586–589.
Duchêne, S., Lardeaux, J.M. & Albar�ede, F., 1997b.
Exhuma-tion of eclogites: insights from depth-time path
analysis.Tectonophysics, 280, 125–140.
Engi, M., Scherrer, N.C. & Burri, T., 2001. Metamorphic
evo-lution of pelitic rocks of the Monte Rosa Nappe:
constraintsfrom petrology and single grain monazite age data.
Schweiz-erische Mineralogische und Petrographische Mitteilungen,
81,305–328.
Ernst, W.G. & Dal Piaz, G.V., 1978. Mineral paragenesis
ofeclogitic rocks and related mafic schists of the
Piemonteophiolite nappe, Breuil-St. Jacques area, Italian
WesternAlps. American Mineralogist, 63, 621–640.
Freeman, S.R., Inger, S., Butler, R.W.H. & Cliff, R.A.,
1997.Dating deformation using Rb-Sr in white mica:
greenschistfacies deformation ages from the Entrelor shear zone,
ItalianAlps. Tectonics, 16, 57–76.
© 2015 John Wiley & Sons Ltd
GEOCHRONOLOGY OF THE ZERMATT -SAAS OPHIOL I TE 731
-
Ganguly, J., Tirone, M. & Hervig, R.L., 1998. Diffusion
kineticsof samarium and neodymium in garnet, and a method
fordetermining cooling rates of rocks. Science, 281, 805–807.
Gebauer, D., Schertl, H.P., Brix, M. & Schreyer, W., 1997.35
Ma old ultrahigh-pressure metamorphism and evidencefor very rapid
exhumation in the Dora Maira Massif, Wes-tern Alps. Lithos, 41,
5–24.
Giletti, B.J., 1991. Rb and Sr diffusion in alkali feldspar,
withimplications for cooling histories of rocks. Geochimica
etCosmochimica Acta, 55, 1331–1343.
Glodny, J., K€uhn, A. & Austrheim, H., 2008. Diffusion
versusrecrystallization processes in Rb–Sr geochronology:
isotopicrelics in eclogite facies rocks, Western Gneiss Region,
Nor-way. Geochimica et Cosmochimica Acta, 72, 506–525.
Gouzu, C., Itaya, T., Hyodo, H. & Matsuda, T., 2006.
ExcessAr-40-free phengite in ultrahigh-pressure metamorphic
rocksfrom the Lago di Cignana area, Western Alps. Lithos,
92,418–430.
Groppo, C., Beltrando, M. & Compagnoni, R., 2009. The
P-Tpath of the ultra-high pressure Lago Di Cignana and adjoin-ing
high-pressure meta-ophiolitic units: insights into the evo-lution
of the subducting Tethyan slab. Journal ofMetamorphic Geology, 27,
207–231.
Hacker, B.R. & Liou, J.G., 1998. When Continents
Collide:Geodynamics and Geochemistry of Ultrahigh-Pressure
Rocks.Kluwer Academic, Dordrecht, 323 pp.
Hermann, J., M€untener, O. & Scambelluri, M., 2000.
Theimportance of serpentinite mylonites for subduction
andexhumation of oceanic crust. Tectonophysics, 327, 225–238.
Herwartz, D., M€unker, C., Scherer, E.E., Nagel, T.J.,
Pleuger,J. & Froitzheim, N., 2008. Lu-Hf garnet geochronology
ofeclogites from the Balma Unit (Pennine Alps): implicationsfor
Alpine paleotectonic reconstructions. Swiss Journal ofGeosciences,
101(Suppl. 1), S173–S189.
Herwartz, D., Nagel, T.J., M€unker, C., Scherer, E.E. &
Froitz-heim, N., 2011. Tracing two orogenic cycles in one
eclogitesample by Lu-Hf garnet chronometry. Nature Geoscience,
4,178–183.
Hoschek, G., 2001. Thermobarometry of metasediments