Introduction

Subduction zones are major geological structures where life-essential elements, such as H, O, and C, are mobilized by metamorphic fluids1. In forearc regions, which also represent the most hazardous part of convergent margins, tectonic discontinuities promote the eventual return of these fluids to the Earth’s surface and the preservation of habitable conditions in the sub-surface biosphere2,3,4. In turn, fluid migration may promote seismic activity along subduction zones, for example, causing pore pressure fluctuations5,6.

Carbon-bearing fluids are widespread in subduction zones and their speciation is sensitive to the redox state of geological reservoirs7,8. Under oxidized conditions, CO2 and/or other oxidized C-bearing species may be present, while under reduced conditions CH4 and H2 are expected to be dominant9. It is generally accepted that the presence of aqueous fluids weakens rocks, due to the presence of free aqueous fluid at grain boundaries that can facilitate hydrolytic- and reaction weakening during viscous deformation10, or hydrofracturing5,11. Likewise, the infiltration of carbon-bearing aqueous fluids may affect the rheology of the country rock, although its effects are more debated. Experimental studies show that CO2-rich aqueous fluids produce (i) negligible effects on strength when interacting with basaltic rocks at conditions compatible with CO2 storage in basaltic reservoirs12, (ii) strengthening or weakening of quartzite, depending on the oxygen fugacity, in high-pressure and temperature experiments13 and (iii) fracturing in quartzite and dunite, as shown by high-temperature experiments14. Studies of natural anorthosite in lower crustal shear zones evidence that CO2-rich conditions promote metamorphic reactions and precipitate fine-grained metamorphic minerals at grain boundaries, thus limiting the average grain size and favoring grain size-sensitive creep15. In extensional settings, CO2-rich fluid migration may produce metasomatism and strain softening in peridotite, also leading to seismic activity16. In subduction zones, CO2-rich fluids can produce metasomatism of the hydrated mantle and trigger fracturing by dehydration reactions and volume changes17. CO2-rich fluid can accumulate under low-permeability rocks leading to high pore pressure and seismic activity at crustal depth18,19,20,21,22,23. Additionally, the larger wetting angles of CO2 compared to an aqueous fluid (~90° for the former and ~57−67° for the latter, depending on the mineral assemblage24) together with fluid immiscibility can segregate CO2 from the aqueous fluid and generate preferential CO2 migration via hydrofracturing25,26,27,28,29. Finally, CO2-rich fluids expand more than aqueous fluids during the exhumation and decompression of subducted rocks, leading to failure in low-permeability rock types30.

The effect of the infiltration of more reduced CH4-H2-rich aqueous fluids on the stability and rheology of rocks is, in contrast, less studied at crustal and mantle depths, even though reducing conditions can be locally dominant in subduction zones31,32,33,34. Compared to CO2, CH4, and H2 are energy sources playing a fundamental role in metabolic processes in the subsurface biosphere. The mechanisms capable of promoting the migration of these fluids towards shallower reservoirs, including the biosphere, remain largely unknown. Nevertheless, it has recently been documented that the genesis of CH4-rich fluids causes strain localization and may promote seismic activity in subducted rocks35, processes through which the migration of these energetic fluids may be favored.

In this contribution, we show evidence of brecciation of omphacitite, a rock rich in omphacite that represents the characteristic clinopyroxene of eclogite-facies mafic rocks. This rheologically strong rock type is in contact with serpentinites and carbonated serpentinites from which CH4-H2-rich fluids were produced at eclogite-facies conditions31,35,36. Mineral assemblages constrain brecciation to a depth range of 30–80 km in the Alpine subduction zone, thus consistent with the CH4 formation conditions. We suggest that the migration and accumulation of CH4-H2-rich fluids below a permeability seal, possibly promoted by CH4 and H2 immiscibility and segregation in aqueous fluids, led to supralithostatic fluid pressure conditions37 that triggered brittle, and potentially seismic, failure of the host rock, which further facilitated the migration of these fluids in the subduction zone.

Results

Geological and structural setting

The Lanzo Massif is a lithospheric mantle body of ~150 km2 situated in the Western Italian Alps. The massif consists of a core of fresh peridotite rimmed by serpentinites that experienced a complex evolution, from Jurassic exhumation and alteration at the seafloor, to Eocene subduction and Alpine high-pressure metamorphism at ~550–600 °C and 2–2.5 GPa38 (Fig. 1a). The serpentinized rim include bodies of carbonated serpentinites (also called ophicarbonates), which are interpreted to have formed prior to Alpine subduction, during the hydrothermal alteration at the seafloor38,39. Nevertheless, another alteration event occurred at high-pressure conditions in the Alpine subduction zone and resulted in the serpentinization of a fresh peridotite portion and the associated genesis of H2-rich fluids31. This alteration is evidenced by the fluid-mediated reduction of a carbonated serpentinite body, which resulted in the formation of abiotic CH4 and graphite at near peak metamorphic conditions, around 49.6 ± 1.0 Ma, and during exhumation31,36. This process was interpreted to reflect the availability of H2-rich reduced fluids formed through serpentinization and their percolation inside the carbonated serpentinite at high-pressure conditions34. The process of carbonate reduction and CH4 formation was shown to cause strain localization inside the reacted carbonated serpentinite35. The presence of graphite-rich veins inside the surrounding serpentinite was used to suggest that the CH4 produced inside the carbonated serpentinite was capable of migrating outwards and precipitating graphite31.

Fig. 1: Geological map, field images and microstructures of the studied omphacitite.
figure 1

a Geological map of the Lanzo Massif, Italian Western Alps. The star highlights the location of the outcrop (GPS coordinates: 45°17'44“N 7°30'36“E), the red square the approximate geological map location (modified from31). Figure created with Adobe Illustrator CS6 (https://www.adobe.com/products/illustrator.html). b Outcrop of omphacitite layers, a few meters thick and tens of meters long, bounded by serpentinites and discontinuous talcschist-rich layers. c Polished slab of omphacitite displaying a fracture network (light green) and graphite-garnet-bearing jadeitite veins (silvery gray). d Enlargement of the contact between omphacitite and talcschist-rich layer. e Thick section scan highlighting the brecciated structure, with omphacitite clasts surrounded by a matrix of jadeite, omphacite, grossular, titanite, and graphite. The yellow arrow shows a graphite-enriched stylolitic structure between two interpenetrated clasts. The location of fluid inclusion shown in Fig. 6a.

Full size image

In this article, we investigate the mechanical behavior of rocks adjacent to the reacted carbonated serpentinite in response to the migration and accumulation of these methane-rich metamorphic fluids. In the study area, the serpentinized Lanzo Massif is separated from the adjacent continental Sesia Zone by an omphacite-rich layer, hereafter omphacitite40,41,42 (Fig. 1). The origin of this rock type, which is mineralogically similar to an eclogite, is ascribed to metasomatic processes affecting the interface between the Lanzo serpentinite and the Sesia gneissic rocks36,37,38. This work focuses on the mechanical behavior of this rock, as a case study similar to the eclogitized mafic crust forming atop of subducting lithospheric mantle sections, whereas its metasomatic origin is beyond the scope of this work and will not be discussed further. The omphacitite layer can be observed with a lateral continuity of a few hundred meters and a thickness up to several meters. Tectonic slicing of this contact produced a few isolated layers of omphacitite within serpentinites (Fig. 1b)30,34. The studied omphacitite lens is up to 5 meters thick and tens of meters long and is bounded by serpentinites. It is located at <20 meters (across the main foliation) from the reduced carbonated serpentinite from which intense eclogite-facies metamorphic CH4 production was previously identified31,35. In this area, the omphacitite shows intense brecciation evidenced by omphacitite angular clasts and fractures sealed by a dark filling material, over a thickness of >2 meters and a length of >5 meters (Fig. 1c, e). Discontinuous talcschist layers occur both at the contact between omphacitite and serpentinite and within the omphacitite (Fig. 1d). Talcschist layers have a maximum thickness of 30 cm and a length of several meters. The main foliation inside the serpentinite, omphacitite, and talcschist is parallel to the main lithological boundaries and dips with low angles to the E-NE.

Microstructural and microchemical constraints of dilation breccia formation

X-Ray Microscopy and petrographic analysis highlight a dilation breccia structure, with internally weakly foliated omphacitite fragments ranging in size from a few microns to several centimetres (Figs. 2 and 3, Supplementary Figs. 1–4, Supplementary information X-ray Microscopy (XRM) statistical data and Supplementary Movies 1-6). These fragments are sealed by a pervasive and interconnected matrix composed of variable proportions of jadeite, omphacite, grossular, titanite, and graphite (Fig. 4c–h and Supplementary Figs. 5–8). The omphacitite breccia can be classified as crackle and mosaic breccia43. Locally, the breccia shows also jadeitite veins oriented subparallel to the foliation visible in the clasts. The veins are made of euhedral jadeite grains growing syntaxially, suggesting a crack-and-seal growth44 (green arrows in Supplementary Fig. 1). Post-brecciation ductile deformation is represented by the local eclogite-facies foliation reworking both the clasts and the matrix. Near the contact with the serpentinite, the omphacitite breccia is locally overprinted by a mylonitic foliation rich in chlorite ± pumpellyite developed at retrograde, greenschist-facies conditions (Supplementary Fig. 4e, f).

Fig. 2: Microstructures, cathodoluminescence, chemical and electron backscatter diffraction maps of the studied omphacitite.
figure 2

a Thin section image showing the contact between omphacitite clasts and matrix. Graphite rims omphacitite clasts and is located inside the matrix and along discrete fractures cutting the clasts (Plane‐polarized light). See location in Supplementary Fig. 1. b X-ray map color coded for the different mineral phases. Ttn: titanite, Ap: apatite, Jd: jadeite, Grt: garnet, Omp: Omphacite. c CL image highlighting oscillatory growth zoning in jadeite grains located in the matrix. d Standardized X‐ray map of the FeO weight % in omphacite. Note the decrease of the FeO content toward the rim of the clast. e, f EBSD maps. In all the EBSD maps white lines represent low‐angle boundaries (2–10° of misorientation), black lines represent high angle boundaries (misorientation > 10°), X is parallel to the stretching lineation (Ls) and Z parallel to the pole of the foliation. e GROD map highlighting irregular grain boundaries and low (<5°) internal lattice distortion. f IPF map, color-coded with respect to the X-kinematic axis, displays no crystallographic preferred orientation.

Full size image
Fig. 3: X-ray Computed Tomography of omphacitite core (size of sample: 5.2 mm height, 3.9 mm diametre).
figure 3

a, b XRM highlights a pervasive network of interconnected fractures. The fractures are sealed by jadeite, garnet, omphacite 2 and graphite. Albite partially replaces jadeite during exhumation. c–h Display of single mineral phases color-coded as in legend. Mineral abbreviations as in the previous figure caption, with additional Ab: albite, Gr: Graphite.

Full size image
Fig. 4: Backscattered-electron images of omphacitite.
figure 4

a, b Graphite-enriched stylolitic structures within and between omphacitite clasts and c–h dendritic/acicular textures of omphacite, jadeite and cauliflower garnet intergrown in the matrix indicating fast growth and disequilibrium textures.

Full size image

Graphite, as identified by Raman spectroscopy (Supplementary Fig. 5), forms continuous or discontinuous rims around the clasts, with a maximum thickness of 50 µm. Graphite is also located in the matrix, along discrete fractures cutting the clasts and along more diffuse fractured areas located at the edges of the clasts (Figs. 2a, 4a,b). Interpenetrated clasts show graphite-enriched stylolitic structures mainly parallel to the post-brecciation eclogite-facies foliation (yellow arrows in Figs. 1e, 4a,b and Supplementary Figs. 1,2,3). Three generations of omphacite are distinguished based on microstructural observations and electron probe micro-analyser chemical data. Generation 1 and 2 are located inside the clasts, and represent fractured omphacite cores and their overgrowth, respectively (FeO >6.4 wt%, XJd 0.4 XAcm 0.1 XAug 0.5 and FeO 4.8- 6.4 wt%, XJd 0.41 XAcm 0.09 XAug 0.5, respectively; Table 1 and Fig. 2b,d and Supplementary Figs. 9-12). The second generation is more abundant toward the clast edges. Fragments of the first generation dispersed in the breccia filling matrix display a rim of the second generation in optical continuity, suggesting epitaxial overgrowth (arrow in Fig. 2d and S8f,g). The third generation is located either along discrete fractures cutting the clasts, or in the matrix. In the latter case, omphacite-3 forms dendritic/acicular intergrown with jadeite and displays the lowest FeO values (2-6 wt%, XJd 0.42 XAcm 0.12 XAug 0.45; Figs. 2d and 4c-f, Supplementary Fig. 6). Jadeite in the matrix displays rather pure compositions both for idioblastic grains and for dendritic grains intergrown with the third generation of omphacite (XJd 0.91 XAcm 0.08 XAug 0 and XJd 0.95 XAcm 0 XAug 0.05, respectively; Table 1 and Supplementary Figs. 7 and 9c,d). Grossular garnet in the matrix displays complex cauliflower-like habitus intergrown with jadeite and complex oscillatory zoning, with some grains showing increasing grossular and decreasing andradite content toward the rims (XGrs 0.93 XAdr 0.06, and XGrs 0.95 XAdr 0.04, respectively), whereas other grains displaying opposite zoning (Fig. 4g,h and Supplementary Figs. 9,11,13). The presence of jadeite and omphacite crystallizing in the matrix and the absence of albite suggests at least 1 GPa and 400 °C for the fluid-mediated brecciation45. Noteworthy, the methane formation in the massif was constrained between 2 GPa and 550 °C and 1 GPa and 400 °C31,36.

Table 1 Representative spot and average composition analyses (wt%) of the mineral phases
Full size table

Scanning electron microscope (SEM) electron backscatter diffraction (EBSD) analysis was performed close to the edge of an omphacitite clast to determine if a stage of crystal-plastic loading preceded brecciation and to constrain the grain size distribution of the brecciated domains (Fig. 2e, f and Supplementary Fig. 14). Omphacite grain size (measured as the diameter of the equivalent circle) ranges between a few microns to a maximum of 100 microns, with an average value of 19 microns. An exponential decrease is visible in the plot of the frequency versus the grain size of grains (Supplementary Fig. 14e). Larger omphacite grains display irregular grain boundaries and abundant low angle boundaries. Grain reference orientation deviation (GROD) map displays generally low ( < 5°) internal lattice distortion, with most of the grains having values < 2° (Fig. 2e). Inverse pole figures (IPF) map and pole figures highlight that grains do not show a crystallographic preferred orientation (Fig. 2f). The misorientation angle distribution (MAD) of correlated pairs shows major peaks at low angle misorientations (between 2° and 10°) and minor peaks for misorientations values < 30° compared to the theoretical random distribution. A few isolated peaks are present for uncorrelated pairs for misorientations comprised between 60° and 160° (Supplementary Fig. 14f). The peaks located at low angle misorientations are interpreted to reflect fracturing of former single grains with negligible offset and grain rotation46. These data indicate diffuse micro-fracturing of omphacite grains with minimal role of crystal plasticity prior to and during the brecciation event.

Some omphacite grains located close to the edges of the clasts are twinned, whereas no twinned grains were observed in the matrix (Fig. 5 and Supplementary Fig. 11i, j). Transmission Electron Microscope (TEM) analysis shows that these omphacite grains are highly twinned on (100) planes (Fig. 5c–e). The thickness of the twins is highly variable, from a few tens to a few hundreds of nm. The twin walls contain regularly spaced dislocations testifying to the mechanical origin of these twins, which is typically observed to form at critical resolved shear stress higher than 100 MPa47,48,49 (see below for discussion). Some inclusions of carbon-rich amorphous material are located along the twins (Fig. 5e). Their size is around a hundred nm and their shape is elongated in the direction of the twins. Reduced carbon precipitation from the fluid postdates the twinning, as the twinning planes may act as a local preferential weakness where fluid could migrate50. Additionally, the presence of H2-rich fluids may have enhanced the reworking of such discontinuities, as documented in materials science51. Finally, this chronology is also supported by the presence of carbon in the breccia matrix that is, again, successive to twinning.

Fig. 5: Optical and transmission electron microscopy of twinned omphacite grain.
figure 5

a Omphacitite clast displaying a pervasive network of fractures coated by graphite. Note the location of the omphacite grain shown in (b; thin section photo, plane-polarized light). b Detail of twinned omphacite grain on which TEM was performed (thin section photo, crossed-polarized light). c Selected area of electron diffraction pattern illustrating mirror law of twins along (100), here with the direction [021]*. d TEM bright field image showing numerous planar defects identified as twins on (100). e Carbon-rich inclusions along the twins (arrows).

Full size image

SEM Back-Scattered Electron (BSE) imaging highlights dendritic textures of the garnet, jadeite and omphacite intergrows sealing the breccia fragments (Fig. 4). These growth structures in metamorphic rocks suggest fast precipitation under disequilibrium conditions52,53,54,55. Cathodoluminescence (CL) imaging of euhedral jadeite and garnet in the matrix shows oscillatory zoning (Fig. 2c and Supplementary Figs. 8,10,11,13), as observed in jadeitites by Harlow et al.45, and suggests precipitation from multiple fluid pulses56 or from an evolving fluid not in equilibrium with the surrounding rock.

Raman Spectroscopy highlights CH4, H2, and H2O fluid inclusions located in the euhedral jadeite sealing the breccia (Fig. 6). Based on the collected Raman spectra, the gas bubble in the fluid inclusions consists of 86 mol% of CH4 and 14 mol% H2. No CO2 was detected. Using the obtained \({{{{{{\rm{X}}}}}}}_{{{{{{{\rm{CH}}}}}}}_{4}}\) and H2 proportions, we calculated the \({{{{{{\rm{X}}}}}}}_{{{{{{{\rm{CH}}}}}}}_{4}}\) (CH4/(CH4 + CO2 + H2O + H2) of the trapped fluid. The calculation did not consider fluid-inclusion respeciation during exhumation. The results suggest that the fluid trapped in the fluid inclusions had an \({{{{{{\rm{X}}}}}}}_{{{{{{{\rm{CH}}}}}}}_{4}}\) = 0.65 (XH2O = 0.25; XH2 = 0.1; Supplementary Fig. 15). This corresponds to strongly reducing conditions – fO2 7 units below the fayalite-magnetite-quartz (FMQ) buffer. This fO2 value is similar to previous estimates on the CH4-forming conditions in the adjacent carbonated serpentinite31, and indicate that the fluid present during the brecciation event was a CH4-rich aqueous fluid (see also discussion below). This feature suggests a link between CH4 and H2-rich fluids originated in the adjacent carbonated serpentinite at eclogite-facies conditions and those present during omphacite brecciation.

Fig. 6: Jadeite-hosted two‐phase (liquid‐gas) fluid inclusions from the matrix of omphacitite breccias.
figure 6

a, b Thin section photos, (plane-polarized light). c Representative Raman spectra of CH4–H2–H2O fluid inclusions.

Full size image

Discussion

The eclogite-like, omphacitite metasomatic horizon bounding the Lanzo serpentinites displays evidence of ductile and brittle deformation occurred at near peak, eclogite-facies, to early retrograde, blueschist-facies conditions (2 GPa and 550 °C and 1 GPa and 400 °C) during Alpine subduction. The formation of dilation breccia attests for high pore pressure57. Fluid inclusions and the presence of fluid-deposited graphitic carbon in the breccia matrix indicate that the fluid was aqueous and CH4 and H2-rich. The matrix microstructures, including dendritic and cauliflower-like crystal growths, suggest fast sealing in plausible disequilibrium conditions53,54,55. The relationships between CH4-bearing fluid circulation and brecciation can be derived from the relative chronology of deformation and metamorphic events reconstructed in the natural samples, which here below are discussed starting from the oldest.

Mechanical twinning of clinopyroxene grains located in the omphacitite clasts pre-dated the brecciation event and the ingression of carbon-bearing fluids. Twinning of clinopyroxene requires high differential stresses, with a critical resolved shear stress around 140–150 MPa, as reported in refs. 47,58. Assuming 140 MPa as critical shear stress, Trepmann & Stockhert48 found differential stresses (Δσ) of 0.5 GPa for twinned jadeite grains in metagranite from Montestrutto (Western Italian Alps), with comparable results also obtained for omphacite49. These authors interpreted the high Δσ as peak stresses resulting from synseismic loading. In the studied omphacitite, Mohr-Coulomb diagrams show that brittle failure in a thrust regime would occur at Δσ of ~1.4 and 3.4 GPa for low pore fluid factor approaching hydrostatic values (λ = 0.4) and vertical stresses (σV) of 1 and 2 GPa, respectively, corresponding to the minimum compressive stress (σ3) during deformation and metamorphism of the omphacitite (see section “Microstructural and microchemical constraints of dilation breccia”; Fig. 7). Δσ of more than 1 GPa were most likely not attained, but instead Δσ in the order of some hundreds of MPa triggered mechanical twinning in omphacite. The mechanical twinning is interpreted to represent a prerupture loading event, thus a strain response to instantaneous stress changes48,58.

Fig. 7: Mohr-Coulomb diagram (normal stress σn - shear stress σs) for the studied omphacitite.
figure 7

The diagrams were computed for vertical stresses (σV) of 1 and 2 GPa (a, b, respectively), corresponding to the minimum compressive stress (σ3; see methods for parameters and assumptions). Failure occurs at Δσ of ~ 1.4 a and 3.4 b GPa for hydrostatic pore fluid factor values (λ = 0.4); shear failure occurs at Δσ of 300 MPa for sublithostatic λ values, both hybrid and extension failure occur at Δσ of 114 MPa and 100 MPa for supralithostatic λ values.

Full size image

The infiltration of carbon-bearing reduced fluids and the brecciation rapidly post-dated the twinning, as carbon precipitated along the twin boundaries of the strongly deformed omphacite crystals and in the breccia matrix (Fig. 5). The successive dilation brecciation of omphacitite could require much lower Δσ values compared to the case of fracturing under hydrostatic λ values, provided that supralithostatic λ values were reached. For σV of 1 and 2 GPa, shear failure occurs at Δσ of 300 MPa for sublithostatic λ values of 0.9–0.97, respectively, hybrid failure occurs at Δσ of 114 MPa for supralithostatic λ values of 1.02–1.01 and extension failure occurs at Δσ of 100 MPa for same λ values (Fig. 7). High pore pressure is also supported by the fact that omphacite twinning is rare and not pervasive, differently to what observed in dry lower crustal conditions, where clinopyroxene deformation twinning is pervasive along pseudotachylite-bearing faults formed at high differential stress and low pore pressure59. Accordingly, the formation of the omphacitite breccia requires pore pressure increase to supralithostatic values. Fluid inclusions in the breccia matrix, as well as the precipitation of graphitic carbon in it, indicate that the overpressured fluid was CH4-H2-rich (Fig. 6). Talcschists and antigorite serpentinites adjacent to the omphacitite are characterized by low permeability and may have represented a permeability barrier for the CH4-rich fluids generated in the underlying carbonated serpentinite. A recent experimental study showed that if serpentinites and talcschists are deformed, permeability increases in the former but not in the latter49,60. Consequently, talcschists could act as capping seals until CH4-H2-rich aqueous fluids produced supralithostatic pore pressure values, leading to brecciation of the surrounding rocks and migration of CH4-H2-rich fluids in the newly formed porosity6,61,62. Subsequent ductile deformation still occurring at high-pressure conditions and, during the retrograde path, at greenschist facies conditions might be related to subsequent aseismic creep (Supplementary Fig. 4e, f).

Brecciation has often been considered a geological record of seismic failure63,64, although some caution should be used in the absence of other distinctive textures, such as pseudotachylites65 or fluidized ultracataclastic veins66. In particular, cataclastic products that display strong shape fabrics may also be related to aseismic shearing flow64. Noteworthy, in the study samples, there is no evidence of internal organization of the clasts or the matrix producing an anisotropy, such as a foliation, but only of local and successive ductile overprint. Pseudotachylites were not identified due either to their absence or to the fact that they are easily altered and overprinted in geological settings rich in fluids, such as subduction zones67. Nonetheless, several authors related the cauliflower structure of garnet to fast growth and seismic activity54,55. Moreover, the brecciation of omphacite-rich, strong rock types at subduction zone high-pressure conditions has been in some cases associated with deep seismic ruptures and transiently high pore pressure56,68,69. Additionally, dilation breccias similar to the ones described in this study are also found at shallower (crustal) depths in carbonate rocks and have been related to seismic failure20,57. Furthermore, it is conceivable that hydrofracturing also involved transient frictional shear deformation, analogous to experiments in serpentine where reaction-induced extension and shear fractures developed simultaneously70 and to natural conditions where hydrofracturing and frictional slip along phyllosilicate-rich foliations occurred simultaneously71. As in the dilation breccias example described above, the studied omphacitite shows evidence of one major brecciation event, possibly because the sealed breccia became harder than the surrounding omphacitite. An alternative interpretation may imply that only a main stage of fluid migration through this rock volume occurred. Additionally, fluids kept open the fracture network until complete sealing, as no significant collapse structures are visible except for the graphite-enriched stylolitic structures between interpenetrated clasts. The fracture network was rapidly sealed at disequilibrium conditions, as evidenced by cauliflower garnet and dendritic/acicular intergrowths of jadeite and omphacite (Fig. 4). Therefore, based on the above features, we suggest that brecciation of omphacitite was plausibly associated with seismic activity at deep forearc conditions.

It is well-established that aqueous fluids can cause hydrofracturing and promote seismic activity in subduction zones5. Likewise, it has been shown that carbon-bearing aqueous fluids, such as CO2-rich aqueous fluids, deriving from orogenic degassing can accumulate under geological traps at crustal depths leading to fluid overpressure and the so-called carbo-fracturing18,19,20,21,22,29. The term overpressure is generic37 and literature data point to either sublithostatic or supralithostatic conditions. In the present case study, Mohr-Coulomb diagrams (Fig. 7) point to supralithostatic fluid pressure. To assess the potential of CH4-bearing aqueous fluids for supralithostatic hydrofracturing at subduction zone conditions, we performed thermodynamic modeling in the P-T range of CH4 formation in the Lanzo Massif. Thermodynamic modeling was achieved by computing a pseudosection along the retrograde path of the Lanzo Massif (y-axis; passing through the 2 GPa, 600 °C and 1 GPa, 400 °C conditions) for a COH fluid72 at carbon saturation conditions. The latter condition is suggested by the precipitation of graphite in both the source area of the CH4-H2-rich fluid –the reduced carbonated serpentinite–, and in the omphacitite breccia matrix (Fig. 1e). Although the petrographic analysis strongly suggests disequilibrium conditions during breccia sealing, the modeling was performed considering fluid-graphite equilibrium in order to expand the potential implication of the modeling to other graphite-bearing rock systems, such as metapelitic rocks. In Fig. 8a, the x-axis refers to the redox state of the graphite-saturated COH fluid [XO = nO/(nO + nH + nC) in the COH fluid72]; For each point along the P-T gradient, the molar volume change between a graphite-saturated COH fluid and pure H2O was computed. For XO = 0.333, a fluid in equilibrium with graphite contains its maximum amount of water and its minimum amount of carbon, so that the difference relative to pure H2O is minimized. Moving across the x-axis towards XO values higher or lower than 0.333, the carbon content in the fluid increases, and the H2O fraction decreases (Fig. 8b). It can be observed that this increase of carbon in the graphite-saturated COH fluid (XO ≠ 0.333) results in a substantial increase in the molar volume of the fluid relative to the water maximum (XO = 0.333). This means that, to reach hydrofracturing conditions, a smaller amount of carbon-bearing fluid is required relative to pure H2O. Because CH4 has a molar volume higher than H2O, a CH4-rich fluid as the one observed in the study rocks (\({{{{{{\rm{X}}}}}}}_{{{{{{{\rm{CH}}}}}}}_{4}}\)  =  0.65 = 0.65) would reach fluid overpressure conditions suitable for hydrofracturing with a much smaller amount of fluid relative to pure H2O, or even relative to a graphite-saturated fluid at water maximum30. For example, Fig. 8a shows that the volume change of a CH4 rich fluid relative to a H2O fluid is about 70% for XO values comparable to those estimated from the natural fluid inclusions.

Fig. 8: Thermodynamic modeling results.
figure 8

a Colormap and contours (white lines) of the fluid volume (in % increase) relative to a pure H2O fluid. Fluid immiscibility at lower P-T conditions is displayed on top of the volume colormap for both CH4-rich fluids (below XO = 0.33) and CO2-rich fluids (above XO = 0.33). The star refers to the XO value derived from the fluid inclusions (Methods for details about carbon-saturation/undersaturation conditions). b Evolution of the fluid composition at 2 GPa and 570 °C in terms of molar fractions of CO2 (red), CH4 (green), and H2O (blue).

Full size image

Figure 8 also shows that the hydrofracturing potential of a graphite-saturated COH fluid is higher for CH4-rich, reduced fluids relative to CO2-rich, oxidized fluids. Additionally, CH4 and CO2 expand more compared to H2O when the confining pressure is reduced30,73. Therefore, CH4 could expand and produce hydrofracturing as soon as it reaches a more porous medium or when the confining pressure drops. Figure 8 also suggests that, if the CH4-rich fluids remained trapped below the talcschist seal during decompression, fluid volume expansion would promote hydrofracturing for even smaller amounts of trapped fluid.

H2O, H2, CO2, and CH4 are immiscible over a wide range of P–T conditions in subduction zones25,26,27,28,74. Evidence of immiscibility has been described in the carbonated serpentinites representing the source rock28,75. Fluid phase separation can occur in immiscible fluids, and H2O fluid can move much easier compared to CO2-rich fluid due to the higher wetting angles of the latter that preclude a thin film at grain boundaries24,76,77. Unfortunately, to the best of our knowledge, no data on the wetting angle of CH4 are available so far, preventing the understanding of its behavior. Finally, seismic activity can enhance the production of effervescing fluid, thus further favoring the migration of carbon-rich fluids18.

Summarizing, we propose the following stages of deformation of omphacitite layers inside serpentinites of the Lanzo Massif, occurring during Alpine subduction around 50 Ma36 at eclogite-facies conditions (Fig. 9):

  1. 1.

    At low pore pressure conditions (hydrostatic), brittle failure in omphacitites would require differential stresses of ~1.4 GPa and 3.4 GPa for vertical stresses of 1 and 2 GPa, respectively. These conditions were not attained, but high differential stresses, in the order of some hundreds of MPa, produced mechanical twinning in omphacite. These high differential stresses likely reflect prerupture loading.

  2. 2.

    The migration and accumulation of CH4–H2-rich fluids under low permeability talcschists produced pore pressure increase to supralithostatic values, leading to brittle failure and brecciation of the omphacitites. Differential stresses around 100–114 MPa are required for extension failure and hybrid failure, respectively. The high molar volume of CH4 leads to fluid overpressure more easily compared to aqueous fluids. Additionally, immiscibility could also enhance fluid phase separation and accumulation.

  3. 3.

    The fracture network was rapidly sealed at disequilibrium conditions by jadeite, omphacite, garnet and graphite. No evidence of further eclogite-facies brittle failure is recorded, suggesting (i) a hardening of the rock volume, (ii) a single major stage of CH4-H2 migration through the rock volume at peak conditions or (iii) a migration of the rupture in an adjacent volume of undeformed omphacitite. Probably the brecciation was associated with seismic failure.

  4. 4.

    Subsequent local ductile deformation of the breccias occurred still at eclogite-facies conditions and during the retrograde path at greenschist facies conditions, recording aseismic creep.

Fig. 9: Interpretation of the studied structures related to pore pressure fluctuations.
figure 9

1. High differential stresses produced mechanical twinning in omphacite. This stage marks progressive prerupture loading. 2. Brecciation of omphacitites occurred at supralithostatic λ values. 3. Fractures were sealed by jadeite, omphacite, garnet, titanite and graphite growth. Fluid inclusions in omphacite recorded the hydrocarbon-bearing fluids composition. Figure created with Adobe Illustrator CS6 (https://www.adobe.com/products/illustrator.html).

Full size image

Concluding, this study suggests that at a depth range between 30 and 80 km in subduction zones the genesis and migration of fluids carrying deep energy sources can lead to supralithostatic pore-fluid pressure and trigger brittle failure in omphacite-rich, mechanically strong rock types similar to eclogite. Modeling results suggest that these CH4-H2-rich fluids can lead to brittle failure much more easily compared to water-dominated fluids, with a CH4-H2-rich fluid having a volume 70% higher than pure H2O at these conditions (Fig. 8). Immiscibility and phase separation may have favored preferential accumulation of carbon-hydrogen fluids via hydrofracturing. We propose that talcschist and serpentinites acted as low-permeability barriers and seal horizons, thus allowing pore pressure to increase until (supra)lithostatic conditions. We envisage that genesis and migration of CH4-H2-rich aqueous fluids may trigger seismic activity in subduction zones at forearc depths. These processes may play an important role in promoting the migration of deep energy sources from deep source areas towards shallower reservoirs, including the subsurface biosphere where microbial life can take advantage of them through metabolic processes2,3,4.

Methods

Optical cathodoluminescence (CL)

CL was performed at the University of Bologna (Italy), using a NewTec Scientific Cathodyne motorized optical cathodoluminescence. Images were acquired at 15 kV, ~300 μA/mm2 under a pressure of ~0.01 Bar.

Scanning electron microscopy (SEM) and electron backscatter diffraction (EBSD)

SEM data were acquired with a Zeiss Ultra 55 field emission gun at the Sorbonne Université, Paris (France). High vacuum was used, with an accelerating voltage of 15 kV and a working distance of 7.5 mm. Backscattered electron (BSE) data were acquired with an Angle Selective Backscattered Detector or an Energy Selective Backscattered Detector.

EBSD data were acquired with a Jeol JSM 6610LV SEM at the Electron Microscopy Centre of the University of Plymouth (United Kingdom). EBSD patterns were collected using a 20 kV accelerating voltage, 18-23 mm working distance, 1 µm step size and a 70° sample tilt. AZtec software (Oxford Instruments) automatically indexed diffraction patterns. Raw maps were processed with HKL Channel 5 (Oxford Instruments), using the noise reduction procedure of78. Grains smaller than 3 times the step size were removed from the dataset. The mean angular deviation value was 0.66, the raw indexing rate of 85%, as only omphacite was indexed. Crystallographic orientation data were plotted on pole figures (stereographic projection; upper and lower hemispheres), with X parallel to the stretching lineation and Z parallel to the pole of the foliation. The grain orientation spread maps (GOS maps) were used to display the intensity of internal strain of individual grains. GOS is defined as the average misorientation angle between each pixel in a grain and that grain’s average orientation. Grain reference orientation deviation (GROD) map displays the misorientation angle measured at every pixel in a grain with respect to the mean orientation of the grain. Grain size in maps and diagrams was defined as the diameter of the equivalent circle.

Transmission electron microscopy (TEM)

As described by79, focused ion beam (FIB) foils were prepared for Transmitted Electron Microscopy (TEM) using a FEI strata DualBeam 235 FIB (at IEMN, Lille). The surface was protected with platinum strip, ∼25 µm long and ∼2 µm thick. Material on each side of the region of interest was removed by a gallium (Ga) ion beam (30 kV, 7 nA). Then, FIB sections were lifted out and attached onto a copper TEM grid by depositing platinum at the contact(s) between the foil and the grid. The section was thinned to ∼150 nm using low beam current (1 nA to 300 pA and 100 pA) grazing on both side of the section. Secondary electrons images were taken to control the thinning process. The amorphous material redeposited by the plasma was removed by scanning the foil with a 5 kV ion beam at an angle of 4–7° with the foil surface. FIB sections were studied by TEM using a Thermo Fisher Scientific Tecnai G2-20 (LaB6 filament) operating at 200 kV at the electron microscopy platform of the University of Lille. The study of the microstructure was performed using bright and dark field imaging in conventional TEM mode. Structural information was obtained by selected area electron diffraction (SAED). Semi-quantitative compositions were measured using energy dispersive X-ray spectroscopy (EDXS).

Electron probe micro-analyser (EPMA)

As described by35, EPMA data were acquired at the Camparis Platform, Sorbonne Université, using a Cameca SX Five equipped with five spectrometers. Spot analyses and X-ray maps were acquired with wavelength dispersive spectrometers (WDS). Firstly, spot analyses were acquired for each mineral phase, successively X-ray maps were collected on overlapping areas. Spot analyses were acquired with 15 KeV accelerating voltage, 10 nA specimen current and ~1 µm beam diameter. The following standards were used to measure ten oxide compositions: orthoclase (Al2O3, K2O), garnet (SiO2, MgO, FeO), albite (Na2O), diopside (CaO), manganese titanate (TiO2, MnO), chromium oxide (Cr2O3). X-ray maps were acquired with 15 KeV accelerating voltage, 10–100 nA specimen current, dwell times of 100 ms and step size of 1 µm. Ten elements (Si, Ti, Al, Fe, Mn, Mg, Na, Ca, K and Cr) were collected in two series for intensity X-ray maps. Successively, intensity X-ray maps were processed using XMapTools 3.4.180, using spot analyses as internal standard to obtain concentration maps of oxide weight percentage.

Raman spectroscopy

Raman spectra were acquired with a Renishaw InVIA Reflex microspectrometer at the Laboratoire de Géologie of the Ecole Normale Supérieure, Paris, France. As described by35, 514 nm laser was used, in polarized mode, delivering 20 mW on the sample, a long-working-distance ×50 Leica objective lens with 0.5 numerical aperture, and 1800 grooves/mm gratings. Analyses were conducted on polished thin sections. A silicon standard was used for calibration. An acquisition time of 5–10 s was used, with 2–4 accumulations per spot. Based on the results of fluid inclusion analysis, the areas of the gaseous molecules were used to estimate the proportion of the main gaseous molecules following the method of81. That the gas bubble contains 14 mol% of H2 and 86 mol% of CH4 (no CO2 detected). The method does not allow us to estimate the proportion of H2O.

Mohr-Coulomb failure diagrams

The Mohr-Coulomb failure diagrams (normal stress σn - shear stress σs) were plotted using the software MohrPlotter v 3.082 available at https://www.rickallmendinger.net/mohrplotter. Pore fluid factor λ was defined as the ratio of fluid pressure to vertical stress σv83. A coefficient of internal friction (μ) of 0.58, a cohesive strength (C) of 50 MPa, and a tensile strength (T) of 25 MPa were used for omphacite (values from ref. 84). Diagrams were computed assuming an Andersonian stress field and a thrusting regime with the vertical stress σV corresponding to the minimum compressive stress σ3.

X-ray microscopy (XRM)

X-Ray Microscopy (XRM) was performed using a ZEISS Xradia Versa 610, available at the Research Center on Nanotechnologies Applied to Engineering (CNIS)—Sapienza University of Rome (Italy). This instrument overcomes the limits of traditional X-ray Computed Tomography (CT) featuring a two-stage magnification architecture (Geometric Magnification + Optics) and a high flux X-ray source to produce sub-micron scale resolution images with enhanced contrast. XRM relies on the same X-ray CT approach to collect a set of 2D projections (radiographs) of the specimen at different viewing angles, acquired by rotating the sample and exposing it to the X-ray beam. The resulting projections are then reconstructed using a reconstruction algorithm to obtain a 3D dataset. Here, the Feldkamp–Davis–Kress (FDK) algorithm was used. Absorption contrast tomography was performed to obtain a three-dimensional reconstruction (dataset) of the specimens using a voltage of 50/60 kV, a power of 4.50/6.50 W, a 0.4× objective for scanning the sample in its entirety, and 4× objective for investigating specific regions of interest (ROIs) with higher resolution to provide an isotropic voxel size of 7 µm and 1.3 µm for ROIs, respectively. The sample was 5.2 mm height and 3.9 mm in diameter; ROI 1 1.1 mm and 0.8 mm and ROI2 1.2 mm and 1.1 mm. The exposure time was set between 1 s and 13 s to reach the desired intensity detection and 1601 projections were acquired for low resolution scans whilst 3001/4501 projections were set for higher resolution investigations. The identification of ROIs was conducted using the Scout-and Zoom procedure. Then, image processing, analysis, and segmentation were performed using Dragonfly Pro software. High resolution datasets were filtered using a non-local means filter (Kernel Size 9; Smoothing 0.5) to reduce noise from images while preserving sharp edges.

Thermodynamic modeling

Thermodynamic modeling of fluid inclusion component speciation [e.g., \({{{{{{\rm{X}}}}}}}_{{{{{{{\rm{CH}}}}}}}_{4}}\) (CH4/(CH4 + CO2 + H2O + H2)] and redox state (XO and fO2) was performed using the thermodynamic database of Huizenga85 and Connolly72 for the COH system. The measured CH4-H2 proportions in the gas phase derived from the fluid inclusions suggest that the fluid trapped in the fluid inclusions is carbon-undersaturated. The XO value calculated for this fluid is 0.07 which, at 550 °C and 2 GPa corresponds to an fO2 value about 7 log units below the FMQ buffer. These estimates were obtained under assumption that the fluid inclusions preserved their original XO value86. Post-entrapment respeciation of the fluid inclusions should have moved the fluid towards carbon saturation (Cesare 1995)86, suggesting that our estimates are conservative with respect to the reduced nature of the fluid. Although the fluid inclusion composition suggests carbon-undersaturated conditions, petrographic evidence (graphite precipitation in the breccia matrix) indicates that the fluid was at least transiently carbon saturated. These fluctuations may be explained by transient pressure variations during the breccia formation, or disequilibrium conditions, as also suggested by the breccia sealing materials.

The thermodynamic modeling for Fig. 8 was done using Perple_X87 version 7.07.71 and the thermodynamic database of Holland and Powell88, modified in 2002. The system of components was modified to have C, O2, and H2 using ctransf. The x axis was set between 1 mole of H2 at x = 0 and 1 mole of O2 at x = 1. The graphite saturation was achieved by having one mole of C for the entire pseudosection. The fluid was modeled using the solid solution COH-Fluid.

The molar volume of COH fluid relative to pure H2O at any P and T along the retrograde path was computed as follows:

$$V{relative}\,{to}\,{H}_{2}O=\frac{{V}_{{COH}-{fluid}}^{P,T}}{{V}_{{H}_{2}O}^{P,T}}$$
(1)

Where:

\({V}_{{COH}-{fluid}}^{P,T}\) is the molar volume of COH fluid in equilibrium with graphite (in m3⋅mol−1) at P and T, \({V}_{{H}_{2}O}^{P,T}\) is the molar volume of pure H2O at the same P and T conditions in m3⋅mol−1.