Missing lawsonite and aragonite found: P–T and fluid composition in meta-marls from the Combin Zone (Western Alps)

We report the first findings of several occurrences of lawsonite and metamorphic aragonite in the meta-sediments from the Combin Zone (Piemonte–Liguria ocean, Western Alps), where the early blueschist-facies episode is poorly documented. New field and metamorphic data (thermodynamic modelling and Raman spectroscopy on carbonaceous material) are used to elucidate the P–T evolution and fluid composition of the Combin Zone and investigate the lawsonite growth and breakdown reactions. Two tectonometamorphic units have been identified within the Combin Zone with distinct geometry, lithological content and P–T conditions. In the higher grade unit, metamorphic aragonite occurs as inclusions in titanite. Lawsonite and garnet were stable at peak P–T conditions (~ 16–17 kbar and 460–480 °C) at very low X(CO2) values. Lawsonite is systematically pseudomorphed, but preserves hourglass zoning or internal fabric associated with the prograde ductile deformation. The lower grade unit (~ 8 ± 1 kbar ~ 370–400 °C) is discontinuously exposed along the western base of the continental Dent Blanche nappe and records P–T conditions similar to those recorded by the Dent Blanche nappe. A metamorphic discontinuity is, therefore, documented between the largest part of the Combin Zone on the one hand, and the Dent Blanche nappe on the other hand. The discovery of lawsonite and metamorphic aragonite allows a better understanding of the large-scale metamorphic structure of the Western Alps.

Lawsonite has been recognized as an essential carrier of H 2 O in subduction zones (Poli and Schmidt 1997;Schmidt and Poli 1998;Tsujimori and Ernst 2014;Whitney et al. 2020), and its potential for geochronology shown in several localities (e.g. Mulcahy et al. 2009;Vitale Brovarone and Herwartz 2013). In addition to specific P-T conditions, lawsonite occurrence is restricted to rocks containing enough CaO and Al 2 O 3 , typically mafic rocks (meta-basalt and meta-gabbro), and Ca-rich sedimentary rocks like metagreywackes, originally containing a significant proportion of volcanic clasts, and meta-marls (e.g. Fornash et al. 2019). Due to its large H 2 O content (typically ca. 12 wt%), the presence of lawsonite in metabasic rocks requires a significant addition of water before the subduction metamorphism, typically by hydrothermal metamorphism following the emplacement of these magmatic rocks (Poli and Schmidt 2002;Staudigel 2014). Unlike metabasic rocks, meta-marls are initially always H 2 O-saturated since their deposition and diagenesis on top of the oceanic crust.
Communicated by Hans Keppler.

Geological setting
The Alps result from the collision of the European and Adriatic plates after the subduction of the Piemonte-Liguria Ocean, a slow-spreading small oceanic domain (Manzotti et al. 2014a;McCarthy et al. 2020;Le Breton et al. 2021) of Jurassic age (e.g. De Wever and Caby 1981;Lombardo et al. 2002;Kaczmarek et al. 2008;Decrausaz et al. 2021).

Lawsonite distribution in the Western Alps
As a general rule, lawsonite (or pseudomorphs after lawsonite) frequently occurs in the ocean-derived units and in the Mesozoic covers of the continent-derived units (e.g. Ellenberger 1960;Michard 1967;Goffé and Chopin 1986;. Some lawsonite may be found in Permian meta-andesites (Saliot 1973;Houfflain and Caby 1987;Schwartz et al. 2000;. Lawsonite is very rare or lacking in units made of reworked (polycyclic) Palaeozoic basement (e.g. Pognante 1989). This primarily reflects the control of the bulk-rock chemical composition (including the availability of H 2 O) on lawsonite development.
Overall, in the Western Alps, the oceanic units may be divided into two main types (e.g. Agard 2021), i.e. those dominated by sediments with few ophiolite complexes (called S-type units), and those largely made of serpentinites (formerly exhumed mantle) associated with meta-gabbros and meta-basalts, with no or restricted occurrences of sediments (referred to as MUM-type units). A metamorphic difference is associated with this lithological distinction ( Fig. 1): the S-type units display low-grade blueschist-facies parageneses, whereas the MUM-type units are characterized by high-grade blueschist-and eclogite-facies parageneses (e.g. Bousquet et al. 2012;Agard 2021). The S-type units are generally considered as recording tectonic processes in an accretionary prism, while the MUM-type units would represent the material of the downgoing oceanic slab (e.g. Marthaler and Stampfli 1989;Vitale Brovarone et al. 2014b;Agard 2021).
Lawsonite (or pseudomorphs after lawsonite) is widespread in metabasic rocks and calcschists of the oceanic units of both the MUM (Fry and Fyfe 1971;Bearth 1973;Dal Piaz 1974;Angiboust et al. 2009) and S type (Bearth 1962;Caron 1974;Pognante and Kienast 1987;Lefeuvre et al. 2020). In the S-type units, lawsonite tends to be largely preserved in the western part of the belt (Fig. 1), whereas it is mostly found as pseudomorphs in the eastern, more internal parts of the belt. This is attributed to increasing peak P-T conditions across the Alpine belt (e.g. Michard 1967;Goffé and Chopin 1986;Agard et al. 2001;Michard et al. 2004;Gabalda et al. 2009;Plunder et al. 2012).
However, the known distribution of lawsonite in the S-type units reveals a paradox. The distribution of lawsonite has been shown on maps with increasing details (Bearth 1962;Kienast and Velde 1970;Bocquet 1971Bocquet , 1974Goffé and Chopin 1986;Goffé et al. 2004;Bousquet et al. , 2008. These maps indicate that lawsonite (or pseudomorphs after lawsonite) is missing in meta-sediments in the area between the east-west trending Aosta Fault to the south and  (Bigi et al. 1983) displaying the occurrences of preserved (red symbols) and pseudomorphed (yellow and white symbols, in low-grade blueschists and high-grade blueschists and eclogites, respectively) lawsonite within the Piemonte-Liguria Ocean (Bearth 1966;Bocquet 1971Bocquet , 1974Goffé and Chopin 1986;. Note the lack of reported lawsonite in the Combin Zone, north of the Aosta Fault. The yellow stars indicate the location of the studied areas. Note also that only two occurrences of metamorphic aragonite are reported in the Western Alps, in the Mesozoic sedimentary cover of the Briançonnais basement in the Vanoise area (Gillet and Goffé 1988) and in the Lanzo Massif (Vitale Brovarone et al. 2017;Giuntoli et al. 2020) the Simplon Fault to the north (Fig. 1). While lawsonite is abundant in the south-western Alps (e.g. Queyras, Ubaye, Montenotte), it has never been reported from the Combin Zone, considered as an equivalent of these regions in the north-western Alps. The proposed lawsonite-glaucophane zone, or the lawsonite 'isograd' running parallel to the belt in the SW Alps, is either absent or drawn as transverse to the belt in the NW Alps (Kienast and Velde 1970;Goffé and Chopin 1986). Similarly, Agard and Lemoine (2005) have drawn three isograds running parallel to the belt in the SW Alps (namely, with increasing grade, lawsonite + , carpholite + , and chloritoid + ). North of the Aosta Fault, these isograds are not identified, and are replaced by a glaucophane + isograd. This difference could suggest that the metamorphic structure of the Alpine belt changes along strike. Indeed, the metamorphic history may have been different in both areas (with peak P-T conditions above the lawsonite stability field in the S-type units of the NW Alps), which would have major implications for the thermal gradient along the Alpine subduction zone.

The studied area
This study focuses on the north-western Alps, where the oceanic units derived from the Piemonte-Liguria Ocean are classically divided into two main groups on the basis of lithological, metamorphic and tectonic criteria (Kienast 1973;Dal Piaz 1974;Ernst and Dal Piaz 1978;Burri et al. 1999;Dal Piaz et al. 2015).
The Zermatt Zone is made of dismembered ophiolites dominated by mantle-derived serpentinized peridotites, meta-gabbros and meta-basalts, with subordinate metasediments. Because these lithologies display spectacular eclogite-facies assemblages, they attracted a large number of studies (e.g. Bearth 1967; Reynard and Ballèvre 1988;Angiboust et al. 2009;Bucher and Grapes 2009;Groppo et al. 2009;Dragovic et al. 2020;Kempf et al. 2020). The Zermatt Zone is a typical example of a MUM-type unit.
The Combin Zone is essentially made of meta-sediments, with minor, strongly dismembered, ophiolitic bodies. Several subunits may be recognized in the Combin Zone, like the Tsaté nappe (Sartori 1987), the Aiguilles Rouges d'Arolla and the Mont de l'Etoile units (Decrausaz et al. 2021). However, the metamorphic history of the Combin Zone is poorly known, and rather deceptive, because it is dominated by a greenschist-facies metamorphism. An earlier blueschist-facies episode is accepted in most syntheses (e.g. Bousquet et al. 2012;Dal Piaz et al. 2015;Agard 2021;Tajčmanová et al. 2021). However, key HP mineral assemblages are rarely reported, with the exception of manganese-rich minerals in meta-cherts (e.g. Baldelli et al. 1983). Rare sodic amphiboles of varying composition (magnesioriebeckite, ferro-glaucophane and glaucophane) have been found in meta-basalts (Sperlich 1988;Martin and Cortiana 2001;Angiboust et al. 2014), although derivation of P-T conditions using them is strongly dependent on the Fe oxidation state . Overall, this suggests that blueschist-facies conditions have been attained in the Combin Zone. However, as emphasized by Epstein et al. (2021), neither lawsonite nor pseudomorphs after lawsonite have been reported in the Combin Zone, in stark contrast with the S-type units in the SW Alps.
Two occurrences of metamorphic aragonite have been found in the European Alps, namely in the Mesozoic cover of the External Vanoise (Gillet and Goffé 1988) and in ophicarbonates from the Lanzo Massif (Vitale Brovarone et al. 2017;Giuntoli et al. 2020). In particular, metamorphic aragonite is unknown in all other oceanic units from the Western Alps, whether they belong to the Combin-, S-type or Zermatt-, MUM-type units.

Mineral chemistry
Mineral analyses were performed with a Cameca SX100 electron microprobe (Microsonde Ouest, Brest, France) operating in the wavelength-dispersive mode. Operating conditions for spot analyses were set to 15 kV, 20 nA and 10 s counting time on the peak (spot size = 1 μm). The ϕ(ρZ) matrix correction was applied based on Pouchou and Pichoir (1985). Standards were natural albite (Na, Si), orthoclase (K), corundum (Al), wollastonite (Ca), forsterite (Mg), MnTiO 3 (Mn, Ti), andradite (Fe), Cr 2 O 3 (Cr), NiO (Ni). A JEOL JXA-8530 electron probe microanalyzer at the Institute of Earth Science of the University of Lausanne (Switzerland) was used to acquire X-ray compositional maps in wavelength-dispersive mode. Operating conditions for compositional mapping were set to 50 ms on 0.7 μm pixel, using an acceleration voltage of 15 kV and a current of 20 nA. Mineral abbreviations and symbols are those used by THERMOCALC

Bulk-rock chemistry
Major-element compositions of selected samples (Table S6 in the Supplementary Information) were determined in whole-rock samples by ICP-AES (CRPG, Nancy). Bulkrock glasses were prepared by mixing appropriate proportions (1:5) of fine-grained rock powder with di-lithium tetraborate. Details about the method used for the analyses are available in Carignan et al. (2001). The composition of two calcschist samples (OL30 and OL32) displaying pseudomorphs after lawsonite has been also measured by the area scan method using a scanning electron microscope (JSM-7100 F, University of Rennes 1). The chemical composition of areas of 2 × 3 mm in size with pseudomorphs after lawsonite was analysed and averaged.

Raman spectroscopy of inclusions in titanite
Mineral inclusions in titanite were identified using Raman micro-spectroscopy at the Laboratoire Magmas et Volcans (Clermont-Ferrand, France). The analyses were carried out in back-scattered geometry using a Renishaw InVia confocal micro-spectrometer equipped with a 532.1 ± 0.3 nm diode laser, a Peltier-cooled CCD detector, a Rayleigh rejection edge filter, and a Leica DM 2500M optical microscope with a motorized XYZ stage. Daily calibration of the spectrometer was performed using a silicon standard (520.5 cm −1 peak). The analyses were carried out using a grating of 2400 grooves/mm, a 20-μm slit aperture (high confocality setting), a 100× microscope objective, and ~ 1 mW laser power on the sample. These analytical conditions result in a spectral resolution of ~ 0.4 cm −1 . Each analysis lasted 30 s (i.e., 3 accumulations of 10 s). The WiRE™ 4.4 software recorded the spectra in the ~ 90-1350 cm −1 wavenumber range. Spectra of all host crystals (titanite) were acquired, and, if necessary, subtracted from the spectra of very small inclusions (few μm) that were often contaminated by the signal coming from the host.

Raman spectroscopy on carbonaceous material (RSCM) method
Regional metamorphic processes produce a gradual transformation of the organic matter present in rocks to CM (carbonaceous material). The peak T reached by samples can be estimated using this graphitization process; regional metamorphic T thus controls the degree of graphitization (Beyssac et al. 2002(Beyssac et al. , 2003. Because graphitization is considered as an irreversible process, the T estimates are not affected by metamorphic reactions associated with retrogression. Carbonaceous material was analysed in 34 samples. Raman spectra of CM in 20 samples were obtained at the Institute of Earth Sciences (University of Lausanne) with a Horiba HR Raman-FTIR spectrometer from HORIBA Scientific (532.15 nm wavelength), an integrated Raman microprobe consisting of an Olympus BX41 confocal microscope coupled to an 800-mm focal-length spectrograph (Tables S7 and S8 in the Supplementary information). Raman spectra of CM in 14 other samples were obtained at the Laboratoire Magmas et Volcans (Clermont-Ferrand, France) using a Renishaw InVia confocal micro-spectrometer (Table S8 in the Supplementary Information), equipped with a 532nm diode laser. Because RSCM can be affected by several analytical mismatches, we followed closely the analytical procedures described by Beyssac et al. (2002Beyssac et al. ( , 2003. Measurements were carried out on polished thin sections using a grating of 2400 grooves/mm, a 20-μm slit aperture (high confocality setting), a 100× microscope objective, and 1-5 mW laser power on the sample surface. The laser was generally focused at 5-15 μm depth to analyze graphite that was located below different minerals and not exposed to the surface. Each analysis consisted of six or ten accumulations of 20 s each in the ~ 900-2000 cm −1 wavenumber range, containing the two principal vibrations of graphite. A spectrum of the mineral phase on top of graphite (mainly calcite, quartz, albite, muscovite, and chlorite) was systematically acquired in the ~ 90-1350 cm −1 wavenumber range with an acquisition time of 15 s. On each sample, we measured at least 15-20 different grains of graphite. The spectra were processed using the WiRE™ 4.4 software: first, they were truncated, then a linear baseline correction was applied in the 1200-1800 cm −1 spectral range, and finally curve fitting was performed. Three bands (corresponding to G, D1, and D2 components, respectively at ~ 1580, ~ 1350, and ~ 1620 cm −1 ) were generally necessary to fit the graphite spectrum; few spectra required a fourth component centered at 1500-1520 cm −1 . The degree of organization of CM is quantified by the relative area of the G, D1, and D2 components of CM in Raman spectra. This is equal to D1⁄(G + D1 + D2) and it is defined as the peak area ratio (R2 ratio).
To evaluate the inter-laboratory reproducibility and accuracy, two independent measurements were carried out on a graphite-bearing sample (OL30, Table S8 in the Supplementary Information) at both University of Lausanne and Laboratoire Magmas et Volcans and gave consistent results (calculated T of 482 ± 20 °C and 477 ± 12 °C, respectively). Another sample (VT08142, a Mesozoic meta-sediments from the Dent Blanche Tectonic System, Fig. 7) was analyzed at the University of Bern (Manzotti et al. 2014b) and at the Laboratoire Magmas et Volcans: the obtained T are consistent (434 ± 7 °C and 444 ± 8 °C, respectively).

P-T pseudosections
P-T pseudosections were calculated in the chemical system MnO-Na 2 O-CaO-K 2 O-FeO-MgO-A l 2 O calculated with the Theriak-Domino software (de Capitani and Brown 1987;de Capitani and Petrakakis 2010), using the internally consistent thermodynamic data set 5.5 Powell 1998, updated November 2003).

Isobaric T-X(CO 2 ) phase diagrams
Isobaric T-X(CO 2 ) projections in the CaMASCH and CaFASCH systems were calculated using THERMOCALC 3.37  and the internally consistent thermodynamic dataset 5.5 updated November 2003). To investigate the mineral equilibria and fluid composition during the prograde evolution and at peak P conditions, pressure was fixed at 12, 14 and 16 kbar. The last value corresponds to the peak P estimated in this study for the micaschist OL38 and interpreted as the peak P reached by the Combin Zone. To understand the reactions that control the stability domains of the critical minerals and retrieve the first-order message, only mineral end-members (rather than solid solutions) were considered in the calculations: aragonite, calcite, kyanite, lawsonite, pyrophyllite, quartz, zoisite, either dolomite or ankerite and either clinochlore or daphnite. The fluid was treated as a binary H 2 O-CO 2 solution.

Field observations
This study focuses on two main areas in north-western Italy where the Combin Zone is exposed (Fig. 1). The first one is located between the underlying Grand Saint Bernard nappe and the overlying Dent Blanche Tectonic System in the Ollomont valley , where detailed field work was carried out in the Conca de By, in the area limited by the Ollomont village to the south and the Mont Avril (i.e. the Italian-Swiss border) to the north (Fig. 2). The second area is located to the south-east of the Dent Blanche nappe in the St Barthelemy valley (Fig. 1).
In both areas, the Combin Zone mainly consists of metasediments (~ 70%), greenschist-facies meta-basalts (~ 20%), meta-gabbros (~ 5%), and minor serpentinized ultramafic rocks (~ 5%) . Three different types of meta-sediments have been distinguished: (i) Rare (< 2%) whitish to silvery quartz-rich garnetbearing micaschists (Fig. S1) associated with quartzites are exposed in metre-sized outcrops close to the Bivouac Rosazza ( Fig. 2), along the Arête de la Bonne Mort and at the Tête Blanche de By. The micaschists contain minute (ca. 0.5 mm in diameter) garnet and pseudomorphs after sodic amphibole. They display a pervasive foliation marked by mmthick quartz-rich layers alternating with muscoviteand chlorite-rich layers. We discovered pseudomorphs after lawsonite in the brownish and greyish calcschists exposed in the By valley, Fig. 2 Sketch map of the studied area NW of the Dent Blanche nappe. From west to east (i.e. from the bottom to the top), the Grand Saint Bernard nappe, the Combin Zone and the Dent Blanche nappe are exposed. The Mesozoic meta-sediments and a few ophiolitic bodies of the Combin Zone are largely outcropping in the Ollomont valley. Red stars indicate the location of the studied samples. Metamorphic aragonite and pseudomorphs after lawsonite are indicated by orange rectangle and yellow lozenges, respectively. T estimated in this study by RSCM are also reported. In most of the Combin Zone the RSCM temperatures are in the range ~ 450-500 °C (By Unit). However, a thin, discontinuous zone at the contact with the Dent Blanche nappe records lower T in the range ~ 370-400 °C (Cornet Unit) between Balme de Bal and Col du Filon (Fig. 2) as well as in the greyish calcschists cropping out in the southern St Barthelemy valley (Fig. 1). Pseudomorphs made of greyish to blackish clusters of submillimetre-sized crystals are conspicuous, because they are standing out on weathered surfaces. In some cases, the pseudomorphs have a prismatic or lozenge shape (Fig. 3c, d), reminiscent of lawsonite. Their modal amount reaches ~ 30%. In strongly deformed calcschists, pseudomorphs after lawsonite are flattened parallel to the main foliation (Figs. 3b and S1).

Petrography and mineral chemistry
A detailed study of three meta-sedimentary samples from the Ollomont valley has been performed. The garnet micaschist (OL38) is a good candidate for investigating the P-T evolution of the meta-sediments, whereas samples OL30 and OL32 are used to study the nature and the distribution of the pseudomorphs in the meta-sediments. In addition, the petrography and mineral chemistry of lawsonite-and garnetbearing calcschists from the St Barthelemy valley are also described.
Subhedral garnet porphyroblasts (up to 500 μm in size; Fig. 4a) display a growth zoning with a significant core-torim decrease in spessartine (from 54 to 17 mol%), compensated by an increase of almandine (from 27 to 59 mol%). Grossular content is rather constant in the core (~ 17 mol%) with a slight increase (~ 22 mol%) in the rim (Fig. 4c). Crystals locally display a snowball structure with an internal rotated foliation (S 1 ) defined by tiny inclusions of rutile and minor quartz. Tourmaline and apatite also occur as inclusions in garnet, with the latter preferentially located in the garnet core.
Garnet is mostly enveloped by the dominant foliation S 2 (stage 2) (Fig. 4a), which is marked by the alternation of discontinuous mm-thick layers of Si-rich Fig. 3 a Alternating layers of brownish pure marbles (meta-limestone, blue square), greyish schists (meta-shale, pink square) and brownish impure marbles (meta-marl, green square). The latter displays pseudomorphs after lawsonite. b Distribution of pseudomorphs after lawsonite, abundant in the phyllitic marbles and absent in the greyish schists, except in a narrow zone at the contact between the two lithologies. This field observation emphasizes how the bulk-rock chemistry controls lawsonite development. c A weathered surface where dissolution of calcite has put in relief pseudomorphs after lawsonite essentially made of graphite, muscovite, paragonite, and minor calcite and therefore less soluble than the matrix. Note that some of the pseudomorphs display the characteristic shape of lawsonite (short rhombohedral prisms). d Brownish pseudomorphs after lawsonite muscovite (Si 3.31-3.41 a.p.f.u., Fig. S2) and minor chlorite (X Mg = 0.44-0.48, X Mn = 0.01-0.02), and several millimetres thick quartz lithons. Apatite, tourmaline, and ilmenite replacing rutile, are found in the muscovite layers and oriented parallel to the main foliation S 2 .
In summary, it is interpreted that the main mineral assemblage comprises garnet, chlorite, muscovite, rutile, quartz and probably glaucophane. Garnet texture and growth zoning reflect syntectonic crystallisation along a prograde P-T path. Albite, epidote, muscovite, chlorite and the minerals pseudomorphing glaucophane, interpreted as retrograde phases, crystallised subsequently.

Calcschists with pseudomorphs after lawsonite
Meta-sediments from the Combin Zone display a large range of compositions, from metapelites (micaschists without calcite) to meta-limestones (almost pure marbles with very minor mica), with all compositional intermediates (phyllitic marbles and calcschists). This difference reflects the sedimentary layering, with alternating layers a few cm to a few dm thick (Fig. 3a). At first sight, the distribution of the pseudomorphs is highly heterogeneous, because they are absent in the marbles and the micaschists, and may be abundant in the interbedded calcschists ( Fig. 3a, b). This clearly reflects a bulk-rock control on the modal abundance of the pseudomorphed phase (see below the discussion on the bulk-rock chemistry). Besides this control by the original sedimentary character, two types of pseudomorph-bearing calcschists have been distinguished and are described below.

Calcschist with graphite-rich pseudomorphs after lawsonite (OL30)
Sample OL30 is a brownish calcschist (Fig. 5) collected from a 15-cm-thick layer interbedded with pure marble layers (~ 10-cm thick). In the field, the alternation of calcschist and pure marble layers is parallel or slightly oblique to the main regional foliation (Fig. S1). A mm-thick layering (S 0 ) is deformed by irregularly spaced crenulations (S 1 ). The layering is defined by discontinuous ~ 1 to 2-mm-thick whitish and brownish domains (Fig. 5a). The whitish domains comprise inclusion-free calcite (FeO = 0.07-0.09 wt%, MgO = 0.24-0.26 wt%), minor quartz and white mica. In the brownish domains, calcite is Fe-richer (FeO = 0.44-1.38 wt%, MgO = 0.18-0.75 wt%) and typically cloudy, because riddled with inclusions of iron oxides and minor rutile. The domains with or without inclusions are independent of the actual grain boundaries, suggesting that prior to the final recrystallisation the brownish domains possibly contained ankerite-dolomite in addition to calcite. The calcite matrix shows dispersed crystals of Si-rich muscovite (Si = 3.17-3.38 a.p.f.u.; X Mg = 0.61-0.80, Fig. S2) and paragonite (X Na = 0.86-0.97). Graphite occurs as minute inclusions in matrix white mica, locally concentrated in the core of the crystals.
Polished surfaces of the calcschists reveal spectacular sector zoning in the pseudomorphs (Fig. 5). The hourglass textures also correspond to a chemical sector distribution, as displaying an hourglass distribution of graphite inclusions. c Photomicrograph and X-ray map for K and Ti of a pseudomorph after lawsonite showing the characteristic sector zoning with preferential incorporation of graphite (Fig. 5). Sector zoning in itself is not diagnostic of lawsonite, and may be observed in many other metamorphic minerals (e.g. chloritoid, staurolite, andalusite, garnet, diopside/omphacite, …). However, given (i) the consistency of their orthorhombic shapes (which excludes garnet), (ii) the nature of the breakdown products, and their similarity to those found in partial pseudomorphs observed in similar rocks from the Western Alps (e.g. Caron 1974;Lefeuvre et al. 2020), and (iii) their occurrence in Jurassic or Cretaceous (hence monocyclic) meta-sediments, which exclude some potential precursors like andalusite or diopside/omphacite, the aggregates are confidently interpreted as pseudomorphs after lawsonite.
Pseudomorphs after lawsonite are euhedral and mostly parallel to the axial planes of the crenulation (S 1 ). Pseudomorphs after lawsonite are bounded by paragonite layers on the two opposite sides parallel to the S 1 , and are surrounded by large crystals of sparry calcite on the two sides perpendicular to S 1 . Mg-chlorite (X Mg = 0.61-0.62; X Mn < 0.01) is frequently found along the four edges of the lozenges.

Calcschist with graphite-poor pseudomorphs after lawsonite (OL32)
Sample OL32 is a greyish calcschist with a discontinuous, spaced foliation (S 2 ) marked by the shape preferred orientation of Si-rich muscovite (Si 3.42-3.48 a.p.f.u.; X Mg = 0.56-0.65, Fig. S2), chlorite (X Mg = 0.49-0.51; X Mn < 0.01), and graphite (Fig. 6a). Graphite is mainly concentrated in the S 2 planes. Microlithons between the S 2 planes display a relict foliation (S 1 ) oriented at high angle to the main foliation S 2 , and defined by graphite, muscovite, titanite, with dispersed grains of quartz in a matrix of calcite. Cloudy, lozenge-shaped pseudomorphs (up to 0.5 mm) are dispersed in the calcite matrix (Fig. 6b), and are generally parallel to S 1 . The pseudomorphs are made of calcite containing a large amount of minute oxides. They may derive from a calcic (tremolite) or sodic (glaucophane) amphibole.
Aggregates (up to 1 cm in size) with an orthorhombic shape and containing unoriented crystals of calcite, albite, quartz, inclusion-free Si-rich muscovite (Si = 3.28-3.39 a.p.f.u.; X Mg = 0.51-0.74), and titanite are wrapped by the S 2 foliation (Fig. 6a), with calcite-rich pressure shadows. Because of their characteristic shapes, they are interpreted as pseudomorphs after lawsonite. In the pseudomorphs, albite and calcite are cloudy due to the presence of tiny graphite inclusions.
Titanite (up to 0.5 mm in size; Al 2 O 3 = 2.93-3.39 wt%) in the pseudomorphs and in the matrix displays abundant elongated or rounded inclusions of carbonates (aragonite, calcite, dolomite-ankerite), quartz, Si-rich muscovite, pyrite, and apatite. The shape of the inclusions may record the finegrained foliation S 1 (Fig. 6c, d). Aragonite occurs as rounded or slightly elongate inclusions (up to 10 μm in size) in the core as well as in the rim of titanite. No cracks are present around the aragonite inclusions.

Lawsonite and garnet calcschists from the St Barthelemy valley
Samples VB13, VB17, VB19 are calcschists and VB21 is a phyllitic marble, all collected in the St Barthelemy valley (Fig. S3).
Sample VB13 is a typical calcschist distinguished by the occurrence of cm-sized blackish clusters moulded by the main chlorite-and white mica-bearing foliation S 2 . The clusters are made of calcite and albite aggregates overgrowing an earlier foliation (S 1 ) defined by elongated trails of graphite and titanite, and at high angle with respect to S 2 . The clusters are interpreted as pseudomorphs after lawsonite.
Samples VB17 and VB19 contain equal amounts of quartz, carbonate, chlorite and white mica, and are distinguished from the other calcschists by the presence of albite and garnet porphyroblasts. The latter deserves some attention, because garnet is exceedingly rare in the Combin calcschists. Garnet crystals display honeycomb textures, with narrow lines of garnet separating large, rounded grains of calcite and/or quartz (Fig. S3). Calcite and quartz have different orientations from one inclusion to the other, and are typically undeformed, although they may display some undulose extinction or minor recrystallisation along their margins when in contact with the matrix foliation. The matrix foliation wraps around the garnet porphyroblasts. These textures are very similar to those described by Hawkins et al. (2007) and Okamoto et al. (2017) in micaschists and calcschists of similar metamorphic grade as the ones studied here. Overall, the modal proportion of garnet in samples VB17 and VB19 is very small (three to five 'grains' per thin section, each grain being about 1 mm in diameter, meaning a modal proportion less than 1 vol. %). Garnet is spessartine-and grossular-rich and poor in pyrope (spss = 26-38 mol%, grs = 18-28 mol%, alm = 34-42 mol%, prp = 2-3 mol%; Fig.  S3), similar to blueschist-facies garnet-bearing marbles from the Queyras (e.g. Ballèvre and Lagabrielle 1994).
Sample VB21 comprises calcite with minor amounts of quartz and white mica, the latter defining the main foliation. Millimetre-sized dark clots are made of aggregates of calcite and minor zoisite, the latter being only found in these aggregates. The clots are wrapped by the main foliation S 2 and generally contain a large amount of graphitic material, locally clearly displaying a relict straight or sigmoidal schistosity (S 1 ) at high angle to S 2 . This suggests that the clots derive from syn-S 1 porphyroblasts, replaced by calcite and zoisite during or after D 2 . They are interpreted as former crystals of lawsonite.  Fig. 6 a Photomicrographs (plane-polarised light) of a calcschist displaying pseudomorphs after lawsonite (sample OL32). Crenulation cleavage S 2 , overprinting a faint pervasive foliation S 1 , is moulding the pseudomorphs, with development of calcite-rich strain shadows. b An example of the small-sized, lozenge-shaped pseudomorph dispersed within the recrystallized calcite matrix, and interpreted as representing former (calcic or sodic) amphibole. c and d An example of a titanite crystal displaying numerous inclusions of carbonates (aragonite, calcite, ankerite-dolomite), quartz, Si-rich muscovite, pyrite, and apatite. The red arrows indicate three metamorphic aragonite crystals. The white rectangle is shown with a larger magnification in (d). e Raman spectra of metamorphic aragonite included in titanite and the host titanite. f Raman spectra of calcite analysed in the matrix and as inclusion in titanite Only one sample, namely the calcschist OL51, sampled close to the contact with the Grand Saint Bernard Unit in the Chesal area, displays a slightly higher T of ~ 515 °C (R2 = 0.21-0.37). According to the acquired data, the temperature distribution is rather homogeneous within each zone (i.e. 373-403 °C and 462-494 °C, respectively). The samples used for detailed petrological investigations belong to the higher T subset. The garnet micaschist (OL38) contains Fig. 7 Simplified tectonic map of the studied area in the north-western Alps. The distribution of some key minerals (garnet and lawsonite), first described in this study, is indicated. In addition, RSCM data from this study are shown within rectangles (mean values of up to 20 measurements per studied sample are given in °C). These are consistent with RSCM values taken from the literature (Negro et al. 2013;Angiboust et al. 2014;Manzotti et al. 2014b;Decrausaz et al. 2021), here shown in rectangles without contours. Note also that garnet and lawsonite occurrences have only been found in the higher T (450-500 °C) domain. P Pillonet; EL Etirol-Levaz very rare graphite crystals yielding a T of 457 ± 21 °C (Figs. 2, 7). The lawsonite-bearing calcschists OL30 and OL32 yield a RSCM T of 482 ± 20 °C and 467 ± 28 °C, respectively.

Micaschists (metapelites)
Because the Piemonte-Liguria meta-sediments equilibrated at low-grade blueschist (T < 450 °C) or high-grade blueschist and eclogite (T > 450 °C) conditions, two schematic compatibility diagrams have been drawn (Fig. 8a, b). In the low-grade blueschist-facies, carpholite and/or chloritoid may be stable. Garnet is not stable, except when stabilized by Mn, like in the meta-cherts. In the AK(FM) projection, the micaschist OL38 (red star in Fig. 8a) shows lower Al 2 O 3 compared with the average pelite compositions (Shaw 1956;Mahar et al. 1997;Tinkham et al. 2001) and Piemonte-Liguria meta-pelites taken from the literature (see caption of Fig. 8), and plots close to the muscovite-chlorite tie-line. The highly aluminous Piemonte-Liguria metapelites display carpholite and chloritoid. The relatively high MnO content (0.28%) of sample OL38 results in stabilizing a spessartinerich garnet at relatively low T (see below). With increasing T (i.e. in the high-grade blueschist and eclogite-facies), almandine-pyrope garnet becomes stable in pelitic compositions (Fig. 8b).

Calcschists (meta-marls)
In the Combin Zone, calcschists are layered meta-sedimentary rocks with a large range of compositions resulting from mixing of a carbonate component (i.e. limestone) with a pelitic component (i.e. shale of various Al 2 O 3 contents). The sedimentary layering is still recognizable in all outcrops (Fig. 3a). For instance, Fig. 3a displays a characteristic field example, where brownish pure marbles (meta-limestone, blue square) are interbedded with greyish micaschists (metashale, pink square) and brownish impure marbles (marly meta-limestone or meta-marl, green square).
Of major interest for our purpose is the distribution of pseudomorphs after lawsonite with respect to the layering and schistosity (Fig. 3b). The pseudomorphs are lacking in the pure marble layer (blue square) as well as in the micaschist (pink square). In contrast, they are abundant in the brownish impure marble (green square) (Fig. 3a, b) and at the contact between the greyish micaschist and the brownish pure marble, where they are concentrated in a 1-2-cmthick layer. The modal proportions of lawsonite may vary in a single bed, like in the brownish impure marble on top of the outcrop (Fig. 3a), the modal variations of lawsonite defining internal layers inside the impure marble bed.
The above field observations reflect directly the relation between the composition of the rocks and that of lawsonite, which necessitates both Al and Ca for its growth. This may be rationalized using an ACF projection (Fig. 8c, d). The range of chemical compositions in the studied meta-sediments is represented based on (i) their Ca content (from a pure limestone to a pure shale), and (ii) mixing lines between a pure limestone and four shales with increasing Al 2 O 3 content (red to yellow lines in Fig. 8d). Consider three hypothetical bulk compositions representative of the three different meta-sedimentary layers shown by our reference outcrop (Fig. 3a). Neither a pure limestone (blue square) nor a pure shale (pink square) can contain lawsonite. In contrast, all intermediate bulk compositions (resulting from the mixing of a carbonate component with aluminous shales), like the brownish impure marble (green square), are favourable for the growth of lawsonite under appropriate P-T conditions.
In such rocks, the lawsonite modal proportions are depicted in Fig. 8d as thin dashed lines. Along each mixing line between a pure limestone and the various shales, the largest amount of lawsonite is predicted in rocks plotting along the lawsonite-chlorite tie-line, the lawsonite modal amount increasing from 0% to more than 50% (Fig. 8d). This rationale may be used to understand the field observations.
Lawsonite is absent in pure marbles and pure shales because of the lack of an appropriate component (Al and Ca, respectively). Varying amounts of lawsonite reveal subtle differences of the bulk chemical composition between or inside marly beds, preserving and emphasizing the sedimentary bedding in a system that can be considered, to first order, as isochemical. The large amount of lawsonite at the contact between the greyish micaschist and the brownish pure marble may be interpreted in two ways. It can correspond to a thin marly layer deposited at the transition between a carbonate-and a clay-dominated sedimentation. It may also result from pre-metamorphic processes, an originally sharp transition between shale and limestone being obscured by syn-depositional bioturbation and/or early diagenesis. Alternatively, the lawsonite-rich layer may be the result of metasomatism due to localized fluid flow, with exchange of Ca and Al.
Reporting measured bulk compositions in the ACF diagram gives also some relevant information. First, the three  Shaw (1956), Mahar et al. (1997) and Tinkham et al. (2001). Bulk compositions from low-grade blueschist are taken from the literature as follows (i) Queyras (Terra Nera-Est Fraiteve, Assietta, Albergian): Agard 1999, Bebout et al. 2013, Henry et al. 1996(ii) Shaw (1956), Mahar et al. (1997), Tinkham et al. (2001) are plotted for comparison average pelite compositions may display a minor amount of lawsonite. Second, the studied samples OL30 and OL32 may be classified as meta-marly limestones, with the modal amount of lawsonite in the studied calcschist of ~ 10% (Fig. 8d). Because the pseudomorphs after lawsonite, considered as isovolumic, are necessarily non-isochemical, some elemental exchange must have taken place, and this small-scale metasomatism, favoured by the dehydration of lawsonite during its breakdown, may have changed the bulk composition observed today. For comparison, the available bulk compositions of metamorphic (Bebout et al. 2013) and metasomatic (Vitale Brovarone et al. 2014a) lawsonite-bearing meta-sediments from the Piemonte-Liguria ocean have been plotted. In these rocks, the modal amount of lawsonite varies from ~ 5 to ~ 60%.
To conclude, the distribution of the pseudomorphs with respect to the layering at outcrop scale nicely matches the first-order links between bulk chemistry and mineral assemblages, as depicted in the ACF projection. In addition, this suggests a nearly isochemical metamorphism of a layered sedimentary sequence.

P-T pseudosection for micaschist OL38
Two pseudosections were calculated for the carbonatefree garnet micaschist in the system MnNCKFMASHTO in the P-T range 6-20 kbar and 350-500 °C (Fig. 9). Comparison of the measured garnet core composition (alm = 27 mol%, spss = 54-55 mol%, grs = 17 mol%) and the corresponding isopleths suggests that the garnet core crystallised at ~ 410 °C ~ 15 kbar, in the g-chl-mu-am-qru-law field, compatible with the observed mineral assemblage. Lawsonite has not been identified in the sample, but calculated lawsonite modal abundance is less than 1 vol% in this field (0.2-0.4 vol%). The modelled presence Fig. 9 H 2 O-saturated P-T pseudosections calculated for the garnet micaschist OL38 using the unfractionated bulk-rock composition (see Tab. S6) (a) and a composition from which the garnet core has been subtracted (b). Some fields are not labelled for the sake of clarity; their assemblages can be deduced from the assemblages in adjacent fields. The yellow field in (b) corresponds to the P-T space where the observed paragenesis characteristic of stage 2 is stable. The grey arrow is the inferred prograde P-T path of a sodic amphibole is consistent with the pseudomorphs observed (Fig. 4b). In this field, the modelled Si content in muscovite reaches 3.35-3.38 a.p.f.u., consistent with the values measured in the studied sample. The measured rimward decrease of spessartine and increase of almandine, and the position of the garnet isopleths point towards a prograde growth of garnet. This is also compatible with the observed constant or only slightly rimward increasing amount of grossular.
To estimate the peak conditions reached by the sample (corresponding to the end of garnet growth), the strong fractionation of manganese in garnet has been taken into account and a second pseudosection has been calculated after having subtracted the garnet cores from the bulk-rock composition (Fig. 9b). The corresponding garnet-muscovite-chlorite-amphibole-quartz-rutile stability field spans ~ 14-17.5 kbar and ~ 445-500 °C. The presence of garnet provides a lower bound to minimum P, whereas the chloritoid-in line sets the upper pressure limit. Peak conditions inferred from the garnet rim composition (alm = 54-59 mol%, spss = 17-22 mol%, grs = 21-22 mol%) are ~ 16-17 kbar and ~ 460-480 °C. The Si content of muscovite modelled in this field is of the order of 3.32-3.33 a.p.f.u., the lower range of the measured values. The inferred peak temperature is in agreement with the maximum T of 457 ± 21 °C (Table S8) calculated for this sample using the RSCM method (see above).

T-X(CO 2 ) projections
To investigate the stability of lawsonite in CO 2 -bearing systems, isobaric T-X(CO 2 ) projections have been calculated in the CaMASCH and CaFASCH model systems at 12, 14 and 16 kbar (Figs. 10 and S4). This P range covers the prograde to peak P conditions reached by the Combin Zone as constrained by the P-T modelling on the micaschist OL38 (Fig. 9). Additionally, the following reactions have also been plotted: Whereas the topology of the invariant points and univariant curves is the same in the CaMASCH and CaFASCH systems, the equilibria are shifted towards higher T with increasing pressure (Figs. 10a, b and S4). The stability field of law + arag + q (blue field) expands with increasing P towards higher T, due to the displacement of the reaction The reaction law = arag + ky + q restrains this field (and the stability of lawsonite in general) to X(CO 2 ) < 0.022. For the calcschists, the maximum temperatures estimated by RSCM are in the range ~ 460-500 °C. This suggests that the presence of lawsonite, aragonite and quartz requires (1) albite = jadeite + quartz (2) calcite∕aragonite + rutile + quartz = titanite + CO 2 (3) 3 lawsonite + aragonite = 2 zoisite + 5H 2 O + CO 2 Fig. 10 Isobaric T-X(CO 2 ) projections/grids in the system CMASH and CFASH, calculated at a pressure of 16 kbar (a, b). The reaction albite = jadeite + quartz is plotted as a reference. The decarbonation equilibrium ru + arag + q = sph + CO 2 is also shown by a dashed red line. The grey dashed lines show the effect of pressure on the reaction zo = law + arag. The blue and yellow fields indicate the T-X(CO 2 ) space where lawsonite + aragonite + quartz and zoisite + aragonite + quartz, respectively, can be stable fluids with very low X(CO 2 ) and a minimum P of the order of 12 kbar.
In addition, the coexistence of titanite with quartz and aragonite in the pseudomorphed lawsonite-bearing calcschist OL32, further constrains the X(CO 2 ) to extremely low values (< 0.001 at 12 kbar, and < 0.006 at 16 kbar, Fig. 10). Considering the measured composition of titanite and consequently its reduced activity (a sph = 0.87) has only a negligible effect on the position of equilibrium (2) (X(CO 2 ) increases by 0.001, T decreases by less than 5 °C).

P-T history
On the basis of new field and petrological data two tectonic units may be identified within the Combin Zone. They differ in size, lithological content, and Alpine P-T conditions. New petrological and field data also indicate that there is a major tectonic discontinuity between the two units constituting the Combin Zone west of the Dent Blanche nappe (Fig. 7). Each unit displays different, but homogeneous (within the error of the thermobarometric estimates) metamorphic peak P-T conditions, with a T gap between the two units of ~ 50-80 °C.

Cornet Unit
The upper unit, hereafter called the Cornet Unit, is structurally located just below the contact with the Dent Blanche nappe, and consists of 50-100-m-thick blackish phyllites and minor brownish calcschists. The last few metres below the contact with the overlying Dent Blanche nappe are made of a diverse array of ocean-derived lithologies, including serpentinite, and interpreted as a potential melange along the subduction interface (Angiboust et al. 2014;. The Cornet Unit is discontinuously exposed along the western base of the Dent Blanche nappe. It crops out in the Col Cornet and Fenêtre de Durand areas (Figs. 2,  7). RSCM estimates (Fig. 7) show that the Cornet Unit experienced relatively low-T conditions (370-400 °C) during the Alpine metamorphism (this study, Angiboust et al. 2014). The peak P conditions reached by this unit during the Alpine history are difficult to assess, due to the lack of diagnostic or P-T-sensitive mineral assemblages. North of the studied area, RSCM T estimates may be used to define the same subunit within the Combin Zone, e.g. in the Aiguilles Rouges d'Arolla (Decrausaz et al. 2021) and Bréonaz (Negro et al. 2013;Angiboust et al. 2014).

By Unit
The lower unit (hereafter referred to as the By Unit) is located on top of the Grand Saint Bernard nappe and underlies the Cornet Unit, or, if the latter is lacking, directly the Dent Blanche nappe. The By Unit contains by far most of the oceanic sediments and associated ophiolites within the Combin Zone (Figs. 2, 7), and it may prove upon further research to be a southern equivalent of the Tsaté nappe. The By Unit is dominated by carbonate rocks (i.e. calcschists interbedded with impure to pure marbles) with subordinate metabasalts and serpentinite slices, and minor meta-gabbros of N-MORB affinity . The metamorphic history of the By Unit is known through three types of data.
-Our RSCM estimates on the studied calcschists indicate a peak T in the range 450-500 °C for the Alpine metamorphism. These values are in good agreement with maximum T estimated by RSCM in the Combin Zone to the west of the Dent Blanche nappe (Negro et al. 2013;Angiboust et al. 2014). -The prograde P-T evolution of the studied garnet-bearing micaschist (sample OL38) during the Alpine cycle is marked by the growth of the garnet core at ⁓15 kbar, 410 °C (Fig. 9a) and the subsequent prograde evolution that peaked at 16-17 kbar and 460-480 °C (Fig. 9b). This stage represents the peak P-T conditions reached by the By Unit during the Alpine cycle. -Metamorphic aragonite inclusions in titanite and spectacular pseudomorphs after lawsonite in calcschists; these deserve a separate discussion.
The meta-sediments from the Saint Barthelemy valley (east of the Dent Blanche nappe) display characteristics similar to those of the By Unit, including the pseudomorphs after lawsonite.

A first report of metamorphic aragonite inclusions in titanite
Metamorphic aragonite is seldom reported in blueschistfacies rocks, and this is generally ascribed to the fast kinetics of the polymorphic transition (Brown et al. 1962). Following experimental studies (Snow and Yund 1987;Hacker and Kirby 1993;Hacker et al. 2005) and theoretical considerations on the exact mechanism of the polymorphic transformation (Carlson and Rosenfeld 1981;Madon and Gillet 1984;Gillet et al. 1987;Sotin and Madon 1988), it is assumed that aragonite preservation is favoured if the transition is crossed along the exhumation path at a T lower than about 200-250 °C, in the absence of fluid and deformation. Such conditions have been rarely met in the Western Alps, and indeed the only occurrences so far reported is located in the External Vanoise (Fig. 1), the coldest HP unit along this traverse (Gillet and Goffé 1988), and in the Lanzo Massif, where metamorphic aragonite is included in magnetite and garnet (Vitale Brovarone et al. 2017;Giuntoli et al. 2020).
No pseudomorphs after metamorphic aragonite (like those of Syros- Brady et al. 2004) have been found in the field, despite extensive research during mapping. This is attributable to the full aragonite recrystallisation (e.g. Seaton et al. 2009) during the pervasive greenschist-facies ductile deformation. Consequently, chances are weak to find relict aragonite in the matrix. On the other hand, aragonite could be preserved in a rheologically strong host mineral van der Molen and van Roermund 1986;Tajčmanová et al. 2021). Such a host happened to be titanite in our rocks. However, only very small crystals (5-10 µm), entirely included, distant from fractures, have yielded Raman spectra of aragonite, most other inclusions having been converted to calcite.
Interestingly, in the eclogite-facies rocks from the Piemonte-Liguria Ocean, metamorphic aragonite has been documented in Alpine Corsica (Chopin et al. 2008). There, rare marble samples display graphite-bearing parageneses with aragonite-grossular intergrowths. Such a growth mechanism is unrelated to the one observed in our samples, where aragonite is making globular inclusions inside its host, and there is no observation in support of a previous fibrous habit. Preservation of metamorphic aragonite in the Corsican samples has also been related to the mechanical behaviour of the host garnet. Although metamorphic aragonite inclusions in titanite are, to the best of our knowledge, reported here for the first time, aragonite inclusions have been already found in a large number of host species (garnet, e.g. Kouketsu and Enami 2010;omphacite, e.g. Proyer et al. 2013;diamond, e.g. Dobrzhinetskaya et al. 2006;zircon, e.g. Liu et al. 2001;allanite, e.g. Liu et al. 2015). We are quite confident that further examination of a larger amount of marbles or calcschists from the Western Alps may reveal other metamorphic aragonite occurrences in titanite crystals.

A first report of lawsonite in the Combin Zone
Many studies have been conducted in the Combin Zone in the last decades. Despite the report of the local occurrence garnet in manganese-rich meta-cherts (Burri et al. 1999) and of sodic amphibole in metabasites (i.e. magnesio-riebeckite: Angiboust et al. 2014), the recent detailed petrological and geochemical studies (Angiboust et al. 2014;Jaeckel et al. 2018;Epstein et al. 2020Epstein et al. , 2021 have not reported lawsonite, and its lack along the section studied in this paper was even specifically emphasized. We are pleased now to have found this missing lawsonite. Indeed, this study presents the first description of pseudomorphs after lawsonite in the calcschists of the Combin Zone (By Unit) and demonstrates that lawsonite was stable during its prograde to peak P-T evolution (Fig. 10). Our thermodynamic modelling shows that at the peak P-T conditions derived from the RSCM data and the garnet micaschist, lawsonite-aragonite-quartz was the stable assemblage in the meta-marls, at the condition that X(CO 2 ) did not exceed 0.005 mol per cent in the titanitebearing meta-marls. Further studies on the primary fluid inclusions (see e.g. Herviou et al. 2021) may provide additional constraints on the fluid composition in equilibrium with lawsonite.

Lawsonite breakdown reactions
In a sense, lawsonite-breakdown reactions should be easy to investigate, because the products are under our eyes. However, great care should be taken in identifying breakdown products, because some of the phases actually occurring in the pseudomorphs may be former inclusions in the lawsonite porphyroblasts, known to contain a large amount of inclusions (e.g. Caron 1974;Lefeuvre et al. 2020). At least three phases can be safely considered as being co-stable with lawsonite rather than a breakdown product. The first two ones are titanite (e.g. sample OL32) and rutile (e.g. sample OL30), because of the immobility of titanium. The third one is graphite, especially when it defines curved lines interpreted as inclusion trails formed during synkinematic lawsonite growth (VB13, Fig. S3). In a single pseudomorph, one can note that graphite trails persist in albite, but are lacking in calcite, suggesting that carbon was consumed during the breakdown reaction, enhancing the growth of calcite.
Another point of concern is the fact that the material inside the pseudomorph is never isochemical with respect to lawsonite, and may vary considerably in adjacent lithologies (e.g. Felix and Fransolet 1972;Ballèvre et al. 2003;Hamelin et al. 2018). The presence of muscovite + paragonite (OL30) or muscovite + albite (OL32) within the pseudomorphs requires the addition of K and Na to the original lawsonite. In both samples, white micas in the pseudomorphs have a much lower celadonite content than those in the matrix, suggesting growth at lower P-T conditions.
The best example of a lawsonite-breakdown reaction was found in sample from the St Barthelemy valley (VB13, Fig.  S3), where the pseudomorphs consist of zoisite and calcite and where zoisite is only found in the pseudomorphs, thus suggesting reaction (3) to have taken place.

Timing of lawsonite growth and breakdown with respect to ductile deformation
Porphyroblasts in general (e.g. Zwart 1962;Passchier and Trouw 2005) and lawsonite in particular (e.g. Caron 1974) are useful markers for deciphering the timing of metamorphism with respect to deformation.
In sample OL30 (Fig. 5a, b), the unusually large pseudomorphs (with hourglass sector zoning) overgrow a faint layering (S 0 ) which has been only weakly crenulated (S 1 ). Most pseudomorphs are elongated at high angle with respect to S 0 , and approximately parallel to S 1 . This suggests that the main episode of growth has taken place before S 1 .
In sample OL32, pseudomorphs after lawsonite locally, although rarely, preserve their original lozenge shape. In most cases, however, pseudomorphs after lawsonite are moderately deformed (Fig. 6a), which results in the formation of dissolution seams along some faces, and pressure shadows along the others. In the field, the orientation of these dissolution seams is parallel to the axial planes of F 2 folds, and consequently probably related to stage 2 deformation. Lawsonite growth may have occurred in this sample either during or after the development of the S 1 pervasive foliation.
In samples VB13 (Fig. S3) and VB21, lawsonite porphyroblasts overgrow a pervasive foliation nicely recorded by the alignment of tiny graphite crystals. Because this "internal foliation" (S 1 ) is either straight or slightly curved, and is discontinuous with respect to the "external foliation" (S 2 ), growth of lawsonite porphyroblasts has taken place during development of D 1 .
Lawsonite breakdown has taken place during exhumation of the Combin Zone. Increasing strain associated with S 2 may have led to a complete destruction of the pseudomorphs after lawsonite, that would become unrecognizable. This could explain why lawsonite was not recognized before this study. On the other hand, it may be used to map the domains, at regional scale, where intense greenschist-facies deformation has erased the pseudomorphed lawsonite, and therefore better define the strain gradients within the Combin Zone.

Implications for the metamorphic structure of the Western Alps
These new data have important implications both for the metamorphic history of the oceanic units in the Western Alps and for the building of the nappe stack along the subduction interface.

Metamorphic structure of the Western Alps
After several decades of research, the metamorphic structure of the Western Alps is nicely depicted on the metamorphic maps of the Alps (e.g. Bousquet et al. 2012;. Concerning the oceanic units derived from the Piemonte-Liguria Ocean, Agard (2021) provides a synthesis where two types of units are distinguished (see introduction). The MUM-type units (Zermatt, Viso, Voltri), out of the scope of this paper, record an eclogite-facies metamorphism (Fig. 11). The S-type units, to which the Combin Zone belongs, have been subdivided into low-T blueschists (with preserved carpholite and lawsonite) and high-T blueschists (where most carpholite is replaced by chloritoid and lawsonite by zoisite).
Considering now the section in the north-western Alps (Fig. 1), many authors have proposed that the Combin Zone may represent an equivalent of the Queyras Unit, based on (i) the structural position of the Combin Zone, more internal than the Briançonnais Grand Saint Bernard nappes, and (ii) the lack of eclogite-facies parageneses, and the occurrence of rare relicts of blueschist-facies minerals (especially sodic amphiboles of magnesio-riebeckite to ferro-glaucophane compositions). However, neither lawsonite nor metamorphic aragonite were described in this area. Our findings, therefore, allow us to explain the paradox. The oldest metamorphic parageneses found in the largest part of the Combin Zone (By Unit in Fig. 11b) have crystallised in the lawsonite-aragonite stability field, at about 460-480 °C, 16-17 kbar. The uppermost part of the Combin Zone, occurring as smaller slices just below the overlying Dent Blanche nappe (Cornet Unit in Fig. 11b, but also Aiguilles Rouges d'Arolla Unit further north according to Decrausaz et al. 2021) is characterized by lower P-T conditions than the By Unit. One major difference with respect to the south-western section of the Alps is that the most external domain, with fresh lawsonite and carpholite, is lacking in the north-western Alps, explaining why previous authors did not recognize the lawsonite-glaucophane zone or the lawsonite isograd in this transect.

P-T discontinuities in the nappe stack
In the north-western Alps, from the bottom to the top, the nappe stack involves the various units of the Grand Saint Bernard nappe (including their sedimentary cover), the oceanic rocks that constitute the bulk of the Combin Zone (Tsaté nappe, Aiguilles Rouges d'Arolla, By Unit…) and the overlying Dent Blanche nappe. The P-T history of the Grand Saint Bernard subunits has not been the topic of recent research. Pending further studies (in course), the RSCM data for the Mesozoic cover of the higher subunit of the Grand Saint Bernard nappe (i.e. the Mont Fort nappe: see Pantet et al. 2020 for a discussion) have given values of the order of 450-500 °C (Fig. 7), similar to those in the overlying units. Further work is required to integrate in this scheme the "Série Rousse", which is considered in our scheme (Fig. 7) as part of the detached cover from the Middle Penninic Units (see also Pantet et al. 2020).
In the oceanic units of the Western Alps, the thermobaric structure may be described using three end-member models, namely (i) a continuous T gradient across the oceanic units, without P gaps, (ii) a continuous T gradient with P gaps, and (iii) distinct subunits with gaps in both P and T (e.g. Ballèvre et al. 1990;Agard et al. 2001;Plunder et al. 2012). Concerning specifically the case of the oceanic units of the Combin zone, Angiboust et al. (2014) have suggested the occurrence of three subunits displaying a continuous T gradient, i.e. from the bottom to the top a lower subunit with a max. T in the range 490-460 °C, a median subunit with a max. T in the range 450-420 °C, and an upper subunit with a max. T in the range 410-360 °C. However, it is unclear in the model of Angiboust et al. (2014) whether the subunits  Groppo et al. 2009) are separated by gaps either in P, T, or both and their (tectonic?) boundaries have not been drawn on a map. Given the large and consistent dataset now available on maximum T derived from RSCM data (Negro et al. 2013;Angiboust et al. 2014;Decrausaz et al. 2021; this paper), we recognize in the studied area the existence of two main units (Cornet and By units) with a T gap of the order of at least 50 °C. In agreement with this T difference, and independently of the estimation method, we found garnet and lawsonite only in the higher T, structurally lower, unit (i.e. By unit). The lower T, structurally higher, unit (i.e. Cornet Unit) may be extended further north, out of the studied area, and includes slices located below the basal contact of the Dent Blanche nappe stack, the largest slice being the Aiguilles Rouges d'Arolla (Decrausaz et al. 2021). Given the T gap between them, we consider that the By and Cornet units are separated by a tectonic boundary (Fig. 7).
In the Dent Blanche nappe, P-T conditions of 8 ± 1 kbar ~ 400 °C have been recently estimated by thermodynamic modelling in rocks cropping out in the Ollomont valley close to the contact with the Cornet Unit . These values are consistent with the relatively low T conditions (387 ± 14 °C; Angiboust et al. 2014) previously estimated for the Mont Dolin Series (i.e. the Mesozoic cover) of the Dent Blanche nappe exposed farther north, close to the Arolla village (Fig. 7). On the basis of these data, it may be suggested that the Dent Blanche nappe and the Cornet Unit may have reached similar peak P-T conditions (8 ± 1 kbar and ~ 400 °C) conditions during the Alpine cycle (Fig. 11b). If so, the Cornet and By Units record a P difference of ~ 8 kbar (~ 20-25 km) during the Alpine metamorphism (Fig. 11b) and their interface represents a major tectonic contact.
The timing of the peak P-T conditions remains a matter of debate and is still poorly constrained (see Manzotti et al. 2014a for a detailed discussion). In the Dent Blanche nappe, K-Ar ages scatter from 55 to 30 Ma (Mont Dolin: Ayrton et al. 1982), whereas Rb-Sr ages display a range from 58 to 43 Ma (Angiboust et al. 2014), interpreted as recording a protracted evolution of the deformation under high-pressure metamorphic conditions. In the Combin Zone, Rb-Sr geochronology on white mica constrains the greenschist-facies deformation at 45-36 Ma (Reddy et al. 1999, Reddy et al. 2003Angiboust et al. 2014).
To sum up, there is a significant metamorphic discontinuity between the bulk of the ocean-derived units from the Combin Zone and the overlying Dent Blanche nappe. This contact, whose detailed description is out of scope of this paper, has been interpreted as a ductile normal shear zone post-dating the early, subduction-related, fabrics .
Funding Open access funding provided by Stockholm University.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.