The chromitites of the Herbeira massif (Cabo Ortegal Complex, Spain) revisited

The ultramafic rocks of the Herbeira Massif in the Cabo Ortegal Complex (NW Iberia) host chromitite bodies. The textural and compositional study of the host rocks and the chromitites classified them into: (1) Type-I chromi-tites, forming massive pods of intermediate-Cr chromite (Cr# = 0.60 – 0.66) within dunites; and (2) Type-II chromitites forming semi-massive horizons of high-Cr chromite (Cr# = 0.75 – 0.82) interlayered with dunites and pyroxenites. Minor and trace elements (Ga, Ti, Ni, Zn, Co, Mn, V and Sc) contents in the unaltered chromite cores from both types show patterns very similar to fore-arc chromitites, mimicked by the host dunites and pyroxenites. Calculated parental melt compositions suggest that Type-I chromitites crystallized from a melt akin to fore-arc basalt (FAB), while Type-II chromitites originated from a boninite-like parental melt. Both melts are characteristic of a fore-arc setting affected by extension during rollback subduction and have been related to the development of a Cambrian-Ordovician arc. These chromitites are extremely enriched in platinum-group elements (PGE), with bulk-rock PGE contents between 2,460 and 3,600 ppb. Also, the host dunites and pyroxenites exhibit high PGE contents (167 and 324 ppb, respectively), which are higher than those from the primitive mantle and global ophiolitic mantle peridotites. The PGE enrichment is expressed in positively-sloped chondrite-normalized PGE patterns, characterized by an enrichment in Pd-group PGE (PPGE: Rh, Pt and Pd) over the Ir-group PGE (IPGE: Os, Ir and Ru) and abundant platinum-group minerals (PGM) dominated by Rh-Pt-Pd phases (i.e. Rh-Ir-Pt-bearing arsenides and sulfarsenides, Pt-Ir-Pd-base-metal-bearing alloys, and Pt-Pd-bearing sulfides). The PGM assemblage is associated with base-metal sulfides (mostly pentlandite and chalcopyrite) and occurs at the edges of chromite or embedded within the interstitial (serpentinized) silicate groundmass. Their origin has been linked to direct crystallization from a S-As-rich melt(s), segregated by immiscibility from evolved volatile-rich small volume melts during subduction. At c. 380 Ma, retrograde amphibolite-facies metamorphism occurred during the exhumation of the HP-HT rocks of the Capelada Unit, which affected chromitites and their host rocks but preserved the primary composition of chromite cores of the chromitites. This event contributed to local remobilization of PGE as suggested by the negative slope between Pt and Pd and high Pt/Pd ratios in the studied chromitites, and host dunites and pyroxenites. In addition, it promoted the alteration of primary PGM assemblage and the formation of secondary PGM. Nanoscale observations made by focused ion beam high-resolution transmission electron microscopy (FIB/HRTEM) analysis of a composite grain of Rh-bearing arse-nide with PGE-base-metal bearing alloys suggest the mobilization and accumulation of small nanoparticles of PGE and base-metals that precipitated from metamorphic fluids forming PGE-alloys. Finally, we offer a comparison of the Cabo Ortegal chromitites with other ophiolitic chromitites involved in the Variscan orogeny, from

The ultramafic rocks of the Herbeira Massif in the Cabo Ortegal Complex (NW Iberia) host chromitite bodies.The textural and compositional study of the host rocks and the chromitites classified them into: (1) Type-I chromitites, forming massive pods of intermediate-Cr chromite (Cr# = 0.60-0.66)within dunites; and (2) Type-II chromitites forming semi-massive horizons of high-Cr chromite (Cr# = 0.75-0.82)interlayered with dunites and pyroxenites.Minor and trace elements (Ga, Ti, Ni, Zn, Co, Mn, V and Sc) contents in the unaltered chromite cores from both types show patterns very similar to fore-arc chromitites, mimicked by the host dunites and pyroxenites.Calculated parental melt compositions suggest that Type-I chromitites crystallized from a melt akin to fore-arc basalt (FAB), while Type-II chromitites originated from a boninite-like parental melt.Both melts are characteristic of a fore-arc setting affected by extension during rollback subduction and have been related to the development of a Cambrian-Ordovician arc.These chromitites are extremely enriched in platinum-group elements (PGE), with bulk-rock PGE contents between 2,460 and 3,600 ppb.Also, the host dunites and pyroxenites exhibit high PGE contents (167 and 324 ppb, respectively), which are higher than those from the primitive mantle and global ophiolitic mantle peridotites.The PGE enrichment is expressed in positively-sloped chondritenormalized PGE patterns, characterized by an enrichment in Pd-group PGE (PPGE: Rh, Pt and Pd) over the Irgroup PGE (IPGE: Os, Ir and Ru) and abundant platinum-group minerals (PGM) dominated by Rh-Pt-Pd phases (i.e.Rh-Ir-Pt-bearing arsenides and sulfarsenides, Pt-Ir-Pd-base-metal-bearing alloys, and Pt-Pd-bearing sulfides).The PGM assemblage is associated with base-metal sulfides (mostly pentlandite and chalcopyrite) and occurs at the edges of chromite or embedded within the interstitial (serpentinized) silicate groundmass.Their origin has been linked to direct crystallization from a S-As-rich melt(s), segregated by immiscibility from evolved volatile-rich small volume melts during subduction.At c. 380 Ma, retrograde amphibolite-facies metamorphism occurred during the exhumation of the HP-HT rocks of the Capelada Unit, which affected chromitites and their host rocks but preserved the primary composition of chromite cores of the chromitites.This event contributed to local remobilization of PGE as suggested by the negative slope between Pt and Pd and high Pt/Pd ratios in the studied chromitites, and host dunites and pyroxenites.In addition, it promoted the alteration of primary PGM assemblage and the formation of secondary PGM.Nanoscale observations made by focused ion beam highresolution transmission electron microscopy (FIB/HRTEM) analysis of a composite grain of Rh-bearing arsenide with PGE-base-metal bearing alloys suggest the mobilization and accumulation of small nanoparticles of PGE and base-metals that precipitated from metamorphic fluids forming PGE-alloys.Finally, we offer a comparison of the Cabo Ortegal chromitites with other ophiolitic chromitites involved in the Variscan orogeny, from the Iberian Peninsula to the Polish Sudetes.The studied Cabo Ortegal chromitites are similar to the Variscan chromitites documented in the Bragança (northern Portugal) and Kraubath (Styria, Austria) ophiolitic massifs.
where high-Cr and PGE-rich chromitites formed during successive injections of mantle-derived Cr-and PGE-rich magmas.However, recent research reinterpreted the sequence as a part of the mantle, where the interaction between melts and a refractory harzburgite in a sub-arc mantle environment formed dunites, pyroxenites and chromitites (Tilhac et al., 2016).Therefore, the origin of these chromitites, together with their PGE enrichment, remains unclear.Despite that previous works focused on the geochemistry and petrology of the chromitites and their host rocks (mainly dunites and pyroxenites) many questions remain about the mineralogical setting of PGE on these chromitites.Previous studies on platinum-group minerals (PGM) of the Cabo Ortegal chromitites were carried out more than 20 years ago (Moreno et al., 1999), reporting only a few Os-Ir-Ru-rich PGM and great abundance of Pt-Pd-Rh-rich PGM (i.e. more than 200 grains of PGM within 20 samples), including sulfides [braggite (PdPt 3 S 4 )-cooperite (PtS)], alloys (Pt-Rh-base metals), complex amalgams (Pt-Pd-Au-Hg-Pb), sulfarsenides [hollingworthite (RhAsS)-platarsite (PtAsS); recently disaccredited as variety of S-rich sperrylite (McDonald and Cabri, 2023) − irarsite (IrAsS)], arsenides [sperrylite (PtAs 2 )], bismuthides (Pd-Pt-Bi), tellurides (Pd-Te-Bi), and oxides (Pt-Pd-O).Most of these PGM were found associated with Ni-Fe-Cu sulfides sealed in chromite crystals or more frequently at the margin of the interstitial silicate matrix between those crystals.Moreno et al. (1999) interpreted that PGM located on the edges of chromite (Pt-Pd sulfides) formed by exsolution from primary PGM or PGE-rich sulfides.On the other hand, PGM (PGE-alloys, arsenides, bismuthides, and tellurides) included in the interstitial silicate matrix transformed to serpentine minerals were interpreted as related to postmagmatic processes.
In this paper, we present and discuss new petrographic, mineralogical, and geochemical data from the chromitites of the Herbeira massif, in view of a different interpretation of data suggesting a mantle origin and post-magmatic evolution of these chromitites.We specifically targeted those chromitites hosted in dunites from the upper part of the pseudostratigraphy of the Herbeira massif and we provide the first-ever minor and trace elements data for chromite forming these chromitites, as well as a detailed review about the genesis of the PGM in these chromitites anomalously enriched in PGE.We also present the first nanoscale study of PGM from these chromitites combined with observations of PGM in situ and from hydroseparation concentrates.
The Cabo Ortegal Complex consists of a suite of units that include, from top to bottom (Arenas et al., 2016 and references therein): Upper Units, Ophiolitic Units, Basal Units, and the Somozas Mélange (Fig. 2).The Upper Units are comprised of rocks affected by intermediatepressure (IP) and high-pressure-high-temperature (HP-HT) metamorphism (Arenas et al., 2016).The IP metamorphic rocks, represented by the Cariño Unit, occupy a dominant normal position above the HP-HT Capelada Unit, and are affected by a train of recumbent folds developed during the exhumation of the HP-HT subduction complex (Albert et al., 2013).The Cariño Unit contains a sequence of amphibolite facies turbiditic pelitic to greywackic paragneisses with a maximum depositional age of c. 510 Ma, interlayered with acidic and basic meta-igneous rocks (Albert et al., 2015).The HP-HT metamorphic rocks of the Upper Units occur in the Capelada (above) and Cedeira (below) units.The c. 2000 m thick Capelada Unit comprises (originally in the pre-recumbent folding pile, bottom to top, Arenas et al., 2016, or top to bottom, Puelles et al., 2012 andreferences therein): ultramafic rocks (including the Herberia massif, see below), high-P mafic granulites, eclogites, and eclogitic paragneisses, the latter with inclusions of metagranitic-tonalitic gneisses and eclogites.The Cedeira Unit bears migmatitic gneisses (Chimparra Gneisses), and high-P granulites, high-T amphibolites, and coronitic metagabbros (Candieira Formation) with general slightly lower metamorphic grade (Vogel, 1967;Gil Ibarguchi et al., 1999;Arenas et al., 2016;Beranoaguirre et al., 2022 and references therein).To be noted is that the base the Candieira Formation is a complex tectonic boundary (Carreiro Zone of Tectonic Movement) (Vogel, 1967) that separates the HP-HT Cedeira unit from the Purrido Ophiolite.In this shear zone, mylonitic metaultramafic rocks (garnet-bearing harzburgites and Ticlinohumite-bearing orthopyroxenites) indicate high to ultra-high-P conditions likely developed during subduction (Gil Ibarguchi et al., 1999).On the other hand, the protoliths of the IP and HP-HT units (latest Ediacaran, <550 Ma, to early Ordovician, >470 Ma) are considered to have developed within the framework of a long-lived Cadomian peri-Gonwanan continental-to-transitional magmatic arc setting that evolved through rifting to a passive margin (Arenas et al., 2016;Martínez Catalán et al., 2019;Beranoaguirre et al., 2022 and references therein).
Below the Upper Units, the suprasubduction Purrido and Moeche ophiolites occur, both part of the Upper Ophiolitic Units of Devonian age (Sánchez Martínez et al., 2007, 2011;Arenas et al., 2014).The Purrido Ophiolite is located at the western-most part of the Cabo Ortegal Complex and contains amphibolites and garnet amphibolites.The Moeche Ophiolite consists of greenschists interlayered with scarce phyllites and serpentinites.Underneath the Upper Ophiolitic Units, the Basal Units are represented by the Espasante Unit, which consists of felsic orthogneisses and alternating metabasites and scarce layers of garnet schists.Finally, the Somozas Mélange is the lowest structural unit in the Cabo Ortegal Complex (Arenas et al., 1986(Arenas et al., , 2009;;Marcos et al., 2002;Novo-Fernández et al., 2016).The age of high-pressure metamorphism is ca.400-390 Ma in the HP-HT rocks, ca.390-377 Ma in the mélange, and ca.377-370 Ma in the Basal Units (Novo-Fernández et al., 2022;Beranoaguirre et al., 2022 and references therein, but see Beranoaguirre et al., 2019 andreferences therein for younger events, andFernández-Suárez et al., 2002;Abati et al., 2007 and references therein for older non-Variscan metamorphic events in the allochthonous complexes).

2011
).The studied area is located in the Herbeira Massif (Fig. 3), which comprises an eastern and a western domain separated by the Herbeira Trans-Fault (Moreno et al., 2001).Nevertheless, Tilhac et al. (2016) suggested that the main contact between both domains is igneous rather than tectonic.
The origin of the ultramafic rocks of the Herbeira Massif is associated with the infiltration and differentiation of Si-undersaturated picritic to boninitic melts in an old subcontinental lithospheric mantle beneath an arc (Tilhac et al., 2016(Tilhac et al., , 2017(Tilhac et al., , 2020)).In this model the interaction between these melts and the host harzburgites generated dunites and chromitites, the former replaced by pyroxenites as the melt differentiated.Subsequently, a late amphibolitization stage occurred due to the cooling of hydrated residual melts, which produced the metasomatism of pyroxenites and peridotites.Other models involve batches of mantlederived magmas forming a stratiform pyroxenite-dunite association at the crust-mantle interface below an arc (Moreno et al., 2001), or partial melting of residual oceanic tectonites or sub-arc mantle caused by the addition of supercritical fluids/melts in a suprasubduction environment with accumulation of magma to form pyroxenites and dunites from various parental liquids (Peucat et al., 1990;Girardeau and Gil Ibarguchi, 1991;Santos Zalduegui et al., 2002).Elemental and isotopic data indicate an enriched component in the pyroxenites, introduced into the mantle source during subduction (Santos Zalduegui et al., 2002;Tilhac et al., 2016Tilhac et al., , 2017Tilhac et al., , 2020)).The age of this magmatic event and of the varied associated protoliths is however controversial.
Internal Sm-Nd and Rb-Sr clinopyroxene-garnet-whole rock isochrons from garnet pyroxenites have yielded ages of 390 and 280 Ma and ca. 229 Ma (Santos Zalduegui et al., 2002;Tilhac et al., 2017, respectively).Tilhac et al. (2017) et al. (2002) bears a clear geological meaning for it is consistent with rutile ages in a garnet-rutile clinopyroxenite and with ages of HP-HT metamorphism in the Upper Units of the Cabo Ortegal Complex (Santos Zalduegui et al., 1996Zalduegui et al., , 2002;;Ordóñez Casado et al., 2001;Novo-Fernández et al., 2022 and references therein).This age would indicate generalized recrystallization at ca. 800 • C and 16-18 kbar at 400-390 Ma of the garnet pyroxenites, and likely of other ultramafic rocks, during the subduction-related metamorphic event that produced HP-HT granulites and eclogitic rocks in the Capelada Unit.Amphibolitization may have taken place latter, at 380 Ma, when the HP-HT rocks of the Capelada Unit exhumed (Arenas et al., 2016 and references therein).
Sm-Nd whole-rock-clinopyroxene isochrons in pyroxenites have yielded Cambrian ages (ca.500 Ma), interpreted as the age of the magmatic-metasomatic protoliths (Santos Zalduegui et al., 2002;Tilhac et al., 2017).It merits mention that Cambrian-early Ordovician ages characterize those of the protoliths of all other formations of the HP-HT units, including the U-Pb zircon crystallization ages of the gabbroic protoliths of the granulites, eclogites, and orthogneisses (Fernández-Suárez et al., 2007;Ordóñez Casado et al., 2001;Albert et al., 2013;Beranoaguirre et al., 2020Beranoaguirre et al., , 2022) ) and the U-Pb detrital zircon 521-510 Ma maximum depositional age of the paragneisses (Albert et al., 2015).All these ages imply the presence of different levels of an entire early Paleozoic crust-mantle arc section constructed onto the extended continental margin of Gondwana or a pre-Gondwanan continental block (Albert et al., 2015;Tilhac et al., 2017).However, all these data and interpretations are opposed to ca. 390 Ma magmatic zircons from spinel harzburgite and garnet pyroxenite that indicate a Devonian (rather than Cambrian) melting event (Ordóñez Casado et al., 2001).Either these zircons indicate re-melting during the Devonian subduction-related event at high pressure (mantle wedge/subarc depths) or could be reinterpreted as metamorphic (formed during subduction?) rather than magmatic.In any case, these zircons indicate an important thermal imprint of the ultramafic rocks during Devonian times.

Chromitite bodies
The chromitite bodies of the Herbeira massif are mostly hosted in dunites of the western domain of the massif (Fig. 3).According to  Moreno et al. (2001), these dunites can be grouped into two units; lower and upper dunite, separated by a garnet-bearing pyroxenite dominated unit.The chromitite bodies in the lower dunite unit are scarce and occur as a few irregular and discontinuous layers up to 5 cm thick.They are PGE depleted with values up to 108 ppb (Moreno et al., 2001).On the other hand, the upper dunite unit hosts the highest amount of chromitites, which are highly enriched in PGE with contents up to 13,000 ppb (Moreno et al., 1999).The chromite mineralization occurs as disseminated grains, irregular pods, thin layers, and bands up to 0.50 m thick (Moreno et al., 2001 and references therein).
The studied chromitites are located in the upper dunite unit of the western domain of the Herbeira massif (Fig. 3; black star).They correspond to massive bodies (>80 vol% chromite) of a few meters long with irregular thickness (up to few tens of cm) hosted in dunites (Fig. 4a) and semi-massive horizons (50-80 vol% chromite) of up to a few tens of centimeters length, and < 1 to 3 cm of thickness, mainly hosted in dunites intercalated with minor pyroxenite layers (Fig. 4b, c).In this study, the massive chromitite pods and semi-massive horizons will be referred to as Type-I and Type-II chromitites, respectively.

Studied samples and analytical techniques
The samples collected for this study represent different types of PGErich chromitites and their associated lithologies from the upper dunite unit of the western domain of the Herbeira Massif.Several polished thin sections (from Type-I chromitites, from the contact between Type-II chromitites and dunites, and pyroxenites, and from the contact between dunite and pyroxenite) were studied by means of optical and scanning electron microscopy (SEM).The SEM images and qualitative identification of platinum-group minerals (PGM) and base-metal sulfides (BMS) by means of energy-dispersive spectroscopy (EDS) were obtained with a field emission scanning electron microscope (FE-SEM) JEOL JSM-7100 at the Centres Científics i Tecnològics de la Universitat de Barcelona (CCiTUB), Spain, and a ZEISS Sigma 300 VP at the Laboratories for Quantitative Target Mineralogy (QanTmin) of the Luleå University of Technology (LTU), Sweden.Operating conditions were 20-25 kV accelerating voltage.
The mineral chemistry analyses of chromite, PGM, and BMS were obtained with a JEOL JXA-8230 electron microprobe (EMP) equipped with five wavelength-dispersive spectrometers (WDS) and one EDS spectrometer at the CCiTUB.The analytical conditions are described in detail in Appendix A. Minor and trace elements in chromite from both types of chromitites, dunites, and pyroxenites, were analyzed by laser ablation inductively coupled mass spectrometry (LA-ICP-MS) at the Laboratorio de Estudios Isotópicos (LEI) from the Centro de Geociencias, UNAM (Mexico), using a Resolution M− 50 Excimer laser coupled to a ThermoICap Qc ICP-MS.The analytical procedure is described in detail in Appendix A. Results of electron microprobe analyses and LA-ICP-MS are presented in Appendix B and Appendix C, respectively.
Whole-rock PGEs analyses were performed on two (Type-I and Type-II) chromitites, one dunite, and one pyroxenite samples at Genalysis Ltd (Perth, Western Australia) by nickel sulfide fire assay collection, following the method described by Chan and Finch (2001).The analytical procedure is described in detail in Appendix A and results are presented in Appendix D.
A thin-foil sample was prepared and extracted from a PGM grain that was included in a polished monolayer from a Type-II chromitite sample by focused ion beam-scanning electron microscope (FIB-SEM) in the Laboratorio de Microscopías Avanzadas (LMA) at the Instituto de Nanociencia de Aragón (INA), University of Zaragoza, Spain.Subsequently, the thin-foil sample was studied by scanning transmission electron microscopy (STEM) and high-resolution electron microscopy (HRTEM) at the same institution.The analytical details for FIB-SEM and STEM-HRTEM analyses are described in Appendix A.
Type-II chromitites have semi-massive textures (50-80 vol% chromite) and are formed by subhedral to anhedral chromite grains up to 450 μm that show pervasive pull-apart fractures (Fig. 5e).Some chromite grains are surrounded by thin altered brighter Fe-rich rims (Fig. 5fg), and this alteration also occurs in fractures.In comparison with Type-I chromitites, the degree of alteration in Type-II chromitites is higher.Chromite grains have abundant subrounded inclusions of silicates (olivine, clinopyroxene, and amphibole) with sizes between 20 and 150 μm (Fig. 5h).Some of these silicate inclusions are affected by fractures and are serpentinized.Fe-Ni-Cu sulfides (mainly pentlandite and chalcopyrite) and PGM are also present (Fig. 5h), mostly located at the edge of chromite grains, in contact with the serpentinized interstitial matrix.The fractures are filled with serpentine group minerals and chlorite.
Dunites hosting the Type-II chromitites display mesh textures and preserve relics of primary olivine, orthopyroxene, minor clinopyroxene, and accessory chromite (Fig. 6a).Some disseminated magnetite can also be observed in the serpentinized matrix.Accessory chromite occurs as euhedral to subhedral grains, some of them larger than 100 μm, showing thin alteration rims (Fig. 6b).Some Fe-Ni-Cu sulfide inclusions (pentlandite and chalcopyrite) occur on the edge of accessory chromite and in the silicate matrix (Fig. 6c-d).
The pyroxenite layers intercalated with Type-II chromitites are coarse-grained and strongly serpentinized.Clinopyroxene is anhedral, up to 1 mm, and might locally show rims of secondary amphibole.Orthopyroxene occurs with sizes up to 1 mm and exhibits undulose extinction (Fig. 6e).Minor olivine grains in the pyroxenite have sizes of ~ 200 μm and show signs of deformation.Amphibole also occurs in the matrix as subhedral grains ~ 150 μm in size.Accessory disseminated chromite forms euhedral to anhedral grains with sizes up to 400 μm.Some chromite grains show fine alteration rims as observed in the accessory chromite in dunites (Fig. 6f).BMS are present as fractured grains enclosed in clinopyroxene (Fig. 6g) and chromite, or frequently on the edges of the latter (Fig. 6h).
Altered chromite rims and fractures in both types of chromitites have clear chemical differences (Appendix B: Table B.3). Type-I chromitites show higher Cr and Fe 2+ but lower Al and Mg than their unaltered chromite cores.The Fe 3+ # remains similar to that of unaltered chromite, defining an alteration trend characterized by an increase of Cr (Fig. 8a).On the other hand, alteration rims in chromite from Type-II chromitites have higher Fe 3+ # and Fe 2+ but lower Cr, Al, and Mg than their unaltered cores, defining an alteration trend towards compositions richer in Fe 3+ (Fig. 8b).

Platinum group elements (PGEs) geochemistry
Bulk-rock PGE contents of the Cabo Ortegal chromitites show that Type-I chromitites have 2,460 ppb and Type-II chromitites have 3,600 ppb (Appendix D).Chondrite-normalized (Naldrett and Duke, 1980) PGE patterns for both types of chromitites (Fig. 11) exhibit an enrichment in PPGEs (Rh, Pt, and Pd) relative to IPGEs (Os, Ir, and Ru), defining a positive slope from Ru to Pt and a negative slope between Pt and Pd.
Regarding the host ultramafic rocks, dunites contain 167 ppb total PGEs, whereas pyroxenites contain 324 ppb total PGEs.In both cases, the PGE content is lower than that from Type-I and Type-II chromitites (Appendix D), but much higher than that of the primitive mantle (i.e.23.5 ppb; after McDonough and Sun, 1995) and global ophiolitic mantle peridotites (see Table 4 in Bhat et al., 2021).The chondrite-normalized PGE contents of the ultramafic rocks also show an enrichment in PPGEs relative to IPGEs.These rocks display a negative Ru anomaly with respect to Ir, and Rh (Fig. 11).
The Pt/Pd ratio (Appendix D) in Type-I chromitites is higher (Pt/Pd = 5.1) than Type-II chromitites (Pt/Pd = 2.1) and the host ultramafic rocks (dunite; Pt/Pd = 0.9, and pyroxenite; Pt/Pd = 2.0).The distribution of PGEs observed both for the chromitites and for the host ultramafic rocks agrees with the PGE mineralogy observed in these rocks, which consists mostly of PGM enriched in Pt and Pd (see below).

Platinum-group minerals (PGM) and base-metal sulfides (BMS)
The PGM assemblage observed in situ on thin sections consists of Rh-Ir-Pt-bearing arsenides and sulfarsenides, Pt-Ir-Pd-base-metal-bearing alloys, and Pt-Pd-bearing sulfides (Figs. 12 and 13).PGM from Type-I and Type-II chromitites are mostly located on the edges of chromite grains and in the serpentinized silicate matrix.Due to their small size (<5 µm) and complex textures (replacement/intergrowth), these PGM could only be analyzed qualitatively (see Appendix E).

PGM and associated BMS included on the edge of chromite grains
The PGM assemblage on the edge of chromite is either enclosed within chromite grains or in direct contact with the silicate matrix and comprises Rh-Ir-Pt bearing sulfarsenides and arsenides, Pt-Ir-Pd-basemetal bearing alloys, and Pt-Pd sulfides (Fig. 12; Appendix F).
The PGM observed within chromite are a few grains of Pt-Fe-Cu-Ni alloys with Cu-Fe and Ni-Fe sulfides (probably chalcopyrite and pentlandite, respectively), and Pt-Pd sulfides, all close to the edge of chromite (Fig. 12a, b).The Pt-Fe-Cu-Ni alloys occur as irregular shaped grains enclosed on the edges of the Ni-Fe sulfide associated with Cu-Fe sulfide, amphibole, and chlorite (Fig. 12a).In contrast, Pt-Pd sulfides consist of irregular aggregates (<10 µm) of particles of micrometric size (Fig. 12b).Some of these microparticles occur with needle-like and blocky morphologies.
Rh-Ir-Pt sulfarsenides occur as small polygonal grains (<7 µm) that contain platy-shaped microparticles of Ir-Fe-Ni alloys (Fig. 12c), whereas Pt-Pd sulfides included within Pt-Pd-Fe alloys are found associated with partially oxidized Ni-Fe sulfides (Fig. 12d).Also, Pt-Pd sulfides occur as single grains on the edge of chromite (Appendix F: Regarding BMS, they are completely included in chromite and usually occur as single or composite grains of Ni, Ni-Fe and Cu-Fe sulfides without PGM.Generally, they are found as euhedral (Fig. 5d

PGM and associated BMS in fractures and in the serpentinized silicate matrix
The PGM assemblage located in chromite fractures and in the interstitial silicate matrix correspond to Pt-Pd sulfides, Rh-Ni bearing arsenides, and Pt-Ir-Fe-Ni bearing alloys (Fig. 13).These PGM are often embedded in secondary silicates (e.g., chlorite and/or serpentine).Pt-Pd-S grains occur in the silicate matrix; they are fractured and show irregular morphologies, with less than 5 µm in size (Fig. 13a ).Ir-Fe-Ni alloys occurs as small micrometric to nanometric particles embedded in a Rh-Ni-As matrix (Fig. 13d), whereas Pt-Fe-Ni-Cu (− Ir-Pd) alloys appear as aggregates of micrometer to nanometer-sized particles on the edge of the composite grain of Rh-Ni-As with Ir-Fe-Ni (Fig. 13e).With respect to BMS, they occur in fractures or in the interstitial silicate matrix as irregular grains of Fe-Cu-Ni sulfides (probably pentlandite and chalcopyrite; Fig. 13f).

PGM and associated BMS in mineral concentrates obtained after the hydroseparation process
22 PGM grains were found in the hydroseparated concentrates (Table 1).The composition of PGM grains was obtained quantitatively (Appendix B) and qualitatively (in small grains or mineral phases < 5 µm and/or with complex textures; Appendix E).The PGM in Type-I chromitites (n = 5) are restricted to fine size fractions (<25 and 25-45 µm).

Micro-and nanostructure of PGM
For a better understanding of these complex grains, we selected the composite grain of RhNiAs with intergrowths of Fe-Cu-Os-Ir-Pt-Ni from Fig. 15m (from Type-II chromitite) to obtain a FIB thin-foil cut across the whole grain to obtain a representative section.The TEM-EDS elemental mapping shows that the matrix is mainly composed of Rh, Ni, As, and Pd, whereas the intergrowths are composed of Fe, Cu, Os, Ir, Pt, and Ni (Fig. 19).A high-resolution TEM image (HRTEM) of one of these intergrowths reveals that is composed of different nanometric-sized polycrystalline aggregates (Fig. 20a).TEM-EDS mapping of the same intergrowth shows that these polycrystalline aggregates of nanoparticles have compositions of Pt-Ni-Fe and Fe-Cu-Ru-Os-Ir (Fig. 20b).(Naldrett and Duke, 1980) patterns of the Cabo Ortegal chromitites, dunites, and pyroxenites, and comparison with chromitites hosted in the ophiolitic mantle (Proenza et al., 2007 and references therein).Data for Ural-Alaskan-type complexes of the Urals are from Garuti et al., (2003Garuti et al., ( ,2005) ) and those for layered UG-2 Bushveld chromitites are from Naldrett et al. (2012).

Effects of metamorphism and alteration on mobility of major and minor elements
In the studied chromitites, the alteration is restricted to thin rims and along fractures (Fig. 5a, f-g).The alteration in chromite from Type-I chromitites shows that rims and fractures are enriched in Cr and Fe 2+ and depleted in Al and Mg when compared to chromite cores (Fig. 8a).On the other hand, the alteration rims in chromite from Type-II chromitites are enriched in Fe 3+ (Fig. 8b) and depleted in Cr, Al and Mg in comparison with unaltered chromite cores.
The major element composition of unaltered chromite cores from Type-I and Type-II chromitites overlap with the field defined for ophiolitic chromitites (Fig. 7b, d).The minor and trace element contents in unaltered chromite cores of the studied chromitites show that Ti, Ni, Co, and Sc are within the range of primary igneous high-Cr ophiolitic chromitites, with exception of Mn, V, and Zn contents, which are higher than those typical for ophiolitic chromitites (Fig. 9).Such enrichments in Mn, V, and Zn in Type-I and Type-II chromitites could suggest subsolidus reequilibration between olivine and chromite promoting an increase in Zn, Co, Mn, Ni, including V and Fe 3+ as seen in chromite from the podiform chromitites at Kempirsai (Kazakhstan; Hu et al., 2022).On the other hand, the Ga contents in chromite from Type-II chromitites plot within the field of high-Cr ophiolitic chromitites (Ga < 30 ppm), whereas the chromite from Type-I chromitites have higher values, similar to those from high-Al ophiolitic chromitites (Ga > 40 ppm; Farré-de-Pablo et al., 2020) (Fig. 9).This suggests that chromite from Type-I and Type-II chromitites have Ga contents similar to high-Al and high-Cr ophiolitic chromitites, respectively.
Therefore, major and less mobile minor and trace elements contents in both types of chromitites are within the range of primary igneous chromite, suggesting that the primary chromite composition is preserved.
On the other hand, accessory chromite from dunites and pyroxenites shows similar Cr# and TiO 2 to the Type-II chromitites (Fig. 7b-c) but are more enriched in Fe 2 O 3 (Fig. 8b), V, Zn, and Mn (Fig. 9).This enrichment could be linked to the same alteration processes related to subsolidus reequilibration (i.e.promoting an increase in V, Zn, and Mn) and circulation of oxidizing fluids (i.e. increase in Fe 3+ ) (Fig. 8b).
The MORB-normalized trace elements patterns of unaltered chromite cores from Type-I and Type-II chromitites from Cabo Ortegal Complex are similar to those high-Cr chromitites from the oceanic suprasubduction zone (SSZ) mantle, specifically to fore-arc chromitites (Fig. 10a-b).Moreover, the MORB-normalized trace element patterns of accessory chromite from dunites and pyroxenites are similar to that from Type-II chromitites (Fig. 10c-d).All these evidences suggest that the magmatic fingerprint of a fore-arc environment has been preserved in the studied rocks.

Parental melt compositions
In the Cr# versus Mg# diagram, the composition of chromite from Type-I chromitites falls between the fields defined for accessory Crspinel in mid-ocean ridge basalts (MORB) and boninites (Fig. 7c), whereas the composition of chromite from Type-II chromitites falls near the field defined for Cr-spinel in boninites.In the TiO 2 versus Al 2 O 3 diagram (Fig. 21a), the composition of chromite from Type-I chromitites overlaps the fields for chromite in abyssal and suprasubduction zone (SSZ) peridotites, whereas the composition of chromite from Type-II chromitites overlaps the fields for chromite in SSZ peridotites and island arc basalt (IAB).It should be noted that the composition of accessory chromite from dunites and pyroxenites hosting Type-II chromitites also plots within the fields defined for chromite in SSZ peridotites and IAB (Fig. 21a).The composition of this accessory chromite in terms of their TiO 2 and Al 2 O 3 contents, and its proximity to the Type-II chromitites bodies suggest that they formed in the same environment.
Because the studied chromite preserves a primary chemical composition, it is possible to estimate their parental melts and, hence, the origin of the chromitites.The composition of the melts in equilibrium with chromite has been estimated using the Al 2 O 3 and the TiO 2 contents of chromite from both types of chromitites and the set of equations from  2014), which were partially derived from data by Kamenetsky et al. (2001).It is worth noting that different sets of equations to estimate the parental melts in equilibrium with chromite have been derived from the approach of Kamenetsky et al. (2001) (e.g., Rollinson, 2008;Pagé and Barnes, 2009;Zhou et al., 2014;Zhu and Zhu, 2020), showing similar results.The Zaccarini's equations use the Al 2 O 3 and TiO 2 contents of chromite to calculate the Al 2 O 3 and TiO 2 contents of the parental melts using the following equations for high-Cr chromite: The FeO/MgO ratios of the parental melts were estimated using the Maurel (1984, cited by Augé, 1987) empirical expression: with FeO and MgO in wt.%, and Al# = Al/(Cr + Al + Fe 3+ ) and Fe 3+ # = Fe 3+ /(Cr + Al + Fe 3+ ), both in atomic proportions.
Summarizing, the estimated parental melt compositions suggests that Type-I chromitites formed from FAB-like melts, whereas Type-II chromitites crystallized from boninitic melts.These melts typically occur in fore-arc region of island arcs affected by extension during subduction.We propose that these same boninitic melts reacted with peridotites to form the pyroxenites of the Herbeira massif (Tilhac et al., 2016).In this context, we suggest that the origin of the Cabo Ortegal chromitites reflects the development of a Cambrian arc (ca.500 Ma), as proposed by Tilhac et al. (2017) for the Herbeira pyroxenites.Indeed, some authors have suggested that the terranes in north-western Iberia may represent fragments of an ensialic island arc of at least Cambrian-Ordovician age (Andonaegui et al., 2002;Santos Zalduegui et al., 2002).

Origin of the PGE enrichment in the Cabo Ortegal chromitites and associated rocks
The PPGE enrichment in the studied chromitites resembles that of the Type II chromitites (especially Type IIA) described by González-Jiménez et al., (2014a,2014b).According to these authors, this PPGE enrichment may be related to PGE collection in sulfide melt(s) eventually segregated through immiscibility from evolving, volatile-rich, small-volume melts.These melts are also capable of concentrating S, along with PPGE (González-Jiménez et al., 2011), favoring the segregation of immiscible S-As rich melt(s), thereby leading to the collection of base-metals and further incompatible PGE.
These small melt fractions, enriched in volatile components, generated after extensive melt-rock reaction and fractionation, could be responsible of crystallization of the chromitites in Cabo Ortegal, as a result of percolation and differentiation of these melts from FAB-like to boninitic affinities in a sub-arc mantle setting.This is also consistent with the fact that PGM and PGM-bearing sulfides are systematically located at the edge of chromite grains and included in the interstitial silicate matrix (Figs. 12 and 13), evidencing that segregation of sulfide melt(s) probably occurred at late stages of, and after, crystallization of chromite.Furthermore, sulfide segregation could be favored by decreasing S and volatile solubility in the magma as a result of increasing polymerization degree by differentiation processes (e.g., González-Jiménez et al., 2011 and references therein), resulting in Siricher melts of boninitic affinity parental of the Type-II chromitites (Fig. 21), producing a higher concentration of PGE in the latter compared to Type-I chromitites.These same boninite-like melts reacted with peridotites to form the pyroxenites of the Herbeira massif (Type 2 and 4 pyroxenites; Tilhac et al., 2016).The fact that the shape of chondrite-normalized PGE patterns of the pyroxenite and dunite hosting Type-II chromitites (Fig. 11) are very similar to Type-II chromitites, showing the same PPGE enrichment reinforces this interpretation.Additionally, both Type-I and Type-II chromitites, show similar PGEdistribution patterns.This may be explained by the presence of sulfide mineralization (PGE and base-metals) in these rocks.According to O'Driscoll and González-Jiménez (2016), sulfide-mineralized rocks exhibit enrichment in PPGE over IPGE, represented by steep positive chondrite-normalized PGE patterns, irrespective of the relative PGE abundances.This reflects the immiscible segregation of sulfide melt(s) after extensive fractionation of the silicate melt.Thus, the segregation of the sulfide melt(s) would scavenge the incompatible PGE (Rh, Pt, Pd), due to the chalcophile affinity of these elements (Barnes et al., 1985;Li et al., 1996), and, to a lesser extent, with their higher mobility in volatile-rich systems (Fleet and Wu, 1995 and references therein).

Magmatic PGM and BMS assemblages
The studied PGM assemblage consists mostly of Rh-Ir-Pt-bearing arsenides and sulfarsenides, Pt-Ir-Pd-base-metal-bearing alloys, and Pt-Pd sulfides.
According to our observations, PGE-bearing sulfarsenides (irarsitehollingworthite series) and arsenides (sperrylite) are located at the edges of chromite (Fig. 12c) and some of them associated with BMS (pentlandite and chalcopyrite).In contrast, Moreno et al. (1999) observed that PGE-bearing sulfarsenides and arsenides of the Cabo Ortegal chromitites were located in the serpentinized silicate matrix, and interpreted them as secondary in origin (e.g., formed during serpentinization).But it could also be speculated that the occurrence of PGE-bearing sulfarsenides and arsenides in the studied chromitites (Fig. 12c and 15i-l) suggests direct crystallization of these PGM from the basaltic silicate melt before, or at the same time, to the segregation of the immiscible sulfide melt(s) (González-Jiménez et al., 2020 and references therein), as has been experimentally demonstrated (Helmy et al., 2013;Maier et al., 2015).
The experiments by Helmy et al. (2013) showed that Pt-As nanoassociations form at high temperatures (950-1,180 • C) in both the sulfide melt and the crystallizing monosulfide solid solution (MSS).On the other hand, González-Jiménez et al., (2014a,2014b) suggested that crystallization of chromite in an open-system (dynamic) environment promote changes of T-fO 2 -fS 2 -a As over short time spans.In this context, sudden changes of fS 2 and/or a As as result of the influx of fluids/melt could destabilize PGM-alloys that crystallized earlier, promoting the formation of PGE-sulfides or PGE-sulfarsenides.Liang et al. (2019) argued that euhedral shape and similar composition of PGE-bearing sulfarsenides and arsenides in BMS suggest a direct crystallization from the sulfide melt before crystallization of MSS and intermediate solid solution (ISS).
Additionally, the Pt-Pd sulfides (probably braggite-cooperitevysotskyte series) observed in Type-I and Type-II chromitites (Fig. 12b, 13a; Appendix F: Figs.F.1, F.6) could suggest that they crystallized at high temperatures directly from basaltic magmas in equilibrium with sulfide Ni-Fe-Cu melts or could represent exsolved products from BMS. Cooperite is stable at temperatures above 1,100 • C, while braggite and vysotskyte are stable below 1,100 • C and 1,000 • C, respectively (Verryn and Merkle, 2000) suggesting that some Pt-Pd sulfides in the studied chromitites could crystallized at magmatic conditions.However, discrete Pt-Pd sulfides as inclusions in Pt-Pd-Fe alloys in contact with Fe-oxides (probably as a result of desulfurization of Ni-Fe sulfides) at the edge of chromite (Fig. 12d) could suggest that the Pt-Pd sulfides were exsolved from BMS after crystallization (O'Driscoll and González-Jiménez, 2016).In this context, the Pt-Pd-Fe alloys could represent alteration products from desulfurization of Pt-Pd sulfides originally exsolved from BMS (Fig. 12d).Similar observations were made by Moreno et al. (1999), who interpreted that Pt-Pd sulfides associated with pentlandite and chalcopyrite, may have formed by exsolution from solid solution in BMS.Nevertheless, Moreno and coauthors, suggested that the origin of Pt-Pd sulfides not associated with BMS is uncertain and may represent alteration products of primary PGM.In view of all these observations, we suggest the possibility that some Pt-Pd sulfides could have crystallized under magmatic condition, as well as that some of them may represent exsolved products from BMS and/or alteration of primary PGM.
The observation of a grain with a composition close to isoferroplatinum (Fig. 15e and 16) may suggest crystallization from a sulfide melt.Isoferroplatinum and Pt-Pd sulfides (cooperite) can crystallize directly from a sulfide melt fraction (e.g., Genkin and Evstigneeva, 1986) at higher fS 2 and temperatures around 1,100-1,000 • C as demonstrated experimentally by Makovicky and Karup-Møller (2000) and Majzlan et al. (2002) for associations of Pt-rich alloys and Pt-Pd-Rh-Ir sulfides.
An ISS is the product of the recrystallization of a Cu-rich liquid at around 850 • C as it cools, which can recrystallize into Cu-Fe sulfides (chalcopyrite, bornite, and cubanite) while segregating a semi-metallicrich melt (enriched in Te, Bi, As, and/or Sb) capable of forming PGM of Pt and Pd (Holwell and McDonald, 2010).Crystallization of this residual semi-metallic-rich melt forms discrete PGM around the margins of sulfide grains and could explain the occurrence of small Pt-Pd-bearing arsenides on the edges of Ni-Fe and Cu-Fe sulfides (pentlandite and chalcopyrite) (Fig. 12f Bi) and tellurides (Pd-Te-Bi) described by Moreno et al. (1999).Similar observations on the pyroxenites of the Herbeira massif by Tilhac et al. (2020), show that BMS (pyrrhotite, pentlandite, and chalcopyrite) with minor PGM (Pt-Pd-rich tellurides, bismuthides, and arsenides) occur in the interstitial matrix.These authors interpreted them to be exsolved products after crystallization of primary sulfides.

Secondary PGM and BMS assemblages
After crystallization in the magmatic stage PGM and PGM-bearing sulfides located at the edges of chromite and in the silicate matrix (Figs. 12 and 13) are more vulnerable to post-magmatic processes.As previously reported, the ultramafic rocks of the Herbeira massif have undergone retrograde amphibolite facies metamorphism (~8 kbar, 500 • C; Tilhac et al., 2016) related to the exhumation of the HP-HT rocks of Capelada Unit, which may have triggered a local PGE redistribution, alteration of primary PGM, and formation of secondary PGM in the studied chromitites.
The observation of some Pt-Fe-Cu-(Ni) alloys with porous textures (Figs. 12e,13b,14d,f) suggests that they were formed after the alteration of PGE-rich BMS or preexisting PGM.Strongly reducing conditions (e.g., serpentinization) promote the alteration (desulfurization) of magmatic (PGE-rich) BMS or PGM resulting in the formation of PGE-and/or base-metal-alloys (Cabri et al., 2022 and references therein).Porous grains with compositions between tetraferroplatinum and tulameenite (Fig. 15d, f and 16) suggest that they are formed by lowtemperature alteration of primary PGM (O'Driscoll and González-Jiménez, 2016 and references therein).Moreover, the occurrence of discrete Pt-Fe alloys on the edge of Ni-Fe sulfides (Fig. 12a) could suggest that they are products of alteration of pentlandite (e.g., Prichard et al., 2008) or formed by the replacement of Pt-Pd sulfides during desulfurization processes with sulfide-undersaturated metamorphic fluids (Li and Ripley, 2006).Their location at the edge of chromite crystals or interstitial silicate matrix favored interaction with metamorphic fluids, whereas porous textures (Fig. 15a-d, f) suggest the accumulation of small nanoparticles of Pt, Fe, Cu, and Ni precipitated from metamorphic fluids (Farré-de-Pablo et al., 2022).
Our HRTEM observations regarding the intergrowths of Pt-Ni-Fe and Fe-Cu-Ru-Os-Ir with Rh-Ni-As (interpreted as zaccariniite) (Fig. 15m, 19, and 20), reveal that these intergrowths are formed by aggregates of polycrystalline particles of nanometric size (i.e.nanoparticles) (Fig. 20a).In this context, minerals can exist in various nanostructures such as nanofilms, nanorods, or nanoparticles, depending on whether one, two, or three dimensions are reduced to the nanoscale, respectively.They are considered mineral nanoparticles when they coexist in larger (micrometer-scale) sizes and are termed nanominerals when they exclusively exist within the nanometric size range (González-Jiménez and Reich, 2017).Jiménez-Franco et al. ( 2020) noted that hydrous fluids caused the desulfurization of laurite, releasing minute particles of Ru-Os-Ir alloys < 50 nm in diameter.Jiménez-Franco and coauthors based their observations on an intergrowth of nano-sized (<5 nm) laurite with Ru-Os-Ir-(S-As) and Ru-Os-Ir alloys (5-10 nm) within a Fe-Mn oxide/hydroxide matrix.They suggested that these nanoparticles are the products of the desulfurization process of magmatic laurite, which decomposed into smaller laurite nanoparticles.These smaller particles then underwent a progressive transformation into Ru-Os-Ir-(S-As) particles and, eventually, into Ru-Os-Ir nanoalloys.The reorganization (coarsening) of the desulfurizing laurite within a matrix transforming to Fe-Mn oxide/hydroxide resulted in the formation of nanoscale intergrowths of Ru-Os-Ir-(S-As) particles and Ru-Os-Ir alloys with Fe-Mn oxide/hydroxide (Jiménez-Franco et al., 2020).According to this, it could be speculated that the observed intergrowths of Pt-Ni-Fe and Fe-Cu-Ru-Os-Ir alloys with Rh-Ni-As in the Cabo Ortegal chromitites were formed by desulfurization of primary PGM by metamorphic fluids that released PGE-nanoparticles and base-metal alloys, thus allowing the reorganization and accumulation of these nanoparticles into complex intergrowths.
Zaccariniite is usually interpreted as a secondary mineral formed during post-magmatic processes (e.g., serpentinization; Vymazalova et al., 2012).Observations that Pt-Ir-Rh-As-S occur with zaccariniite and complex intergrowths of PGE-base metal alloys (Fig. 15n  Summarizing, we propose that the Cabo Ortegal chromitites crystallized from volatile-rich small-volume melts that concentrate S, Pt, and Pd and favored the segregation of immiscible sulfide melt(s) after chromite crystallization, producing further concentration of PGE (especially Pt and Pd).The PGM assemblage in Type-I and Type-II chromitites suggests that the origin of the PGM is a combination of direct crystallization from silicate/sulfide melt, followed by the fractionation of sulfide melt, and subsequent sub-solidus reequilibration.Post-magmatic processes related to the percolation of hydrated metamorphic fluids responsible for retrograde metamorphism at amphibolite facies (e.g., exhumation of the HP-HT rocks of the Capelada Unit; Tilhac et al., 2016 and references therein) would promote local remobilization of PGE, which gave place to the alteration of primary PGM and formation of secondary PGM.This is recorded by the negative slope between Pt and Pd and high Pt/Pd ratios in chromitites, dunites, and pyroxenites of the Herbeira massif (Fig. 11; Appendix D).Pd has a greater tendency to be mobilized by metamorphic fluids rather than Pt as experimentally demonstrated by Wood et al. (1992) and Evstigneeva and Tarkian (1996).This suggests that high Pt/Pd ratios and negative slope between Pt and Pd in chromitites and associated host rocks would result from remobilization of Pd by metamorphic fluids leaving residual Pt.It should also be noted that similar PGM assemblages and their textural location were described by Bridges et al. (1993) in the PGE-rich chromitites (up to 11,200 ppb) from the Bragança Massif in northern Portugal.It is thus possible that the same process that led to PGE enrichment and the formation of a PGM assemblage in the Cabo Ortegal chromitites took place in the relatively close Bragança chromitites.Moreno et al. (2001) interpreted the ultramafic rocks of the Herbeira massif as a stratiform pyroxenite-dunite association at the crust-mantle interface below an arc, originated from batches of Cr-and PGE-rich mantle-derived magmas to explain the formation of Pt-Pd-rich chromitites.Other authors proposed that partial melting of residual tectonites or sub-arc mantle caused by the addition of supercritical fluids/melts in a suprasubduction environment with accumulation of magma formed pyroxenites and dunites from different parental liquids (Peucat et al., 1990;Girardeau and Gil Ibarguchi, 1991;Santos Zalduegui et al., 2002).More recently, Tilhac et al., (2016Tilhac et al., ( ,2017Tilhac et al., ( ,2020) ) suggested that the origin of the ultramafic rocks is related to the percolation and differentiation of Si-undersaturated picritic to boninitic melts in an old subcontinental lithospheric mantle beneath an arc.

Geodynamic setting and tectonic implications
Our calculated parental melt compositions suggests that Type-I chromitites crystallized from FAB-like melts, while Type-II chromitites crystallized from boninitic-like parental melts.These melts are characteristic of fore-arc regions affected by extension during subduction.Minor and trace elements contents in the unaltered chromite cores from both types of chromitites show patterns very similar to fore-arc chromitites (Fig. 10).In this context, we propose that the origin of these chromitites are the result of the percolation and differentiation of smallvolume melts of FAB-like to boninitic melts in a sub-arc mantle environment.According to our model, the same boninitic melts reacted with peridotites to form the pyroxenites of the Herbeira massif.Thus, we suggest that the origin of the Cabo Ortegal chromitites and their PGE enrichment occurred during the developing of an arc of (at least) Cambrian-Ordovician age, as suggested by Tilhac et al. (2017) for the origin the Cabo Ortegal pyroxenites.Some authors have suggested that the terranes in north-western Iberia probably represent fragments of an ensialic island arc of Cambrian-Ordovician age (Andonaegui et al., 2002;Santos Zalduegui et al., 2002).Even though the age of this arc is controversial, a cluster of ages around 500 Ma has been reported in ultramafic rocks of Cabo Ortegal (Van Calsteren et al., 1979;Santos Zalduegui et al., 2002) and Sm-Nd whole-rock-clinopyroxene isochrons in pyroxenites have yielded Cambrian ages (ca.500 Ma), interpreted as the age of the magmatic-metasomatic protoliths (Santos Zalduegui et al., 2002;Tilhac et al., 2017).

Comparison with other ophiolitic chromitites involved in the European Variscan orogeny
Table 2 reports the occurrence of chromitite orebodies within ophiolitic complexes involved in the European Variscan orogeny.All these chromitites are related to suprasubduction zones, generated in fore-arc or back-arc environments.Most of these chromitite bodies show massive textures and variable sizes (from a few meters to 100′s meters long) and are hosted in dunites and harzburgites.Their Cr# is highly variable from 0.22 (Calzadilla de los Barros, Spain; Merinero et al., 2013) to 0.93 (Avren, Bulgaria;Colás et al., 2014;Colás, 2015).The following massifs host high-Cr chromitites: Bragança (Portugal; Cr# = 0.62-0.85;Bridges et al., 1995), Cabo Ortegal-Herbeira (Spain; Cr# = 0.60-0.82;this study), Kraubath and Hochgrössen (Austria; Cr# = 0.80-0.86 and 0.74-0.84,respectively; Malitch et al., 2003), and Avren Fig. 21.A) chemical composition of chromite from chromitites, dunites, and pyroxenites of cabo ortegal complex in terms of tio 2 vs. Al 2 O 3 compositions.For comparison, compositional fields of chromite from different geotectonic settings are shown.Data of compositional fields from Kamenetsky et al. (2001): LIP = large igneous province; OIB = ocean island basalt; IAB = island arc basalt; MORB = mid-ocean ridge basalt; SSZ = suprasubduction zone.b) TiO 2 and Al 2 O 3 (wt.%) content of the melt calculated to be in equilibrium with chromite from the Cabo Ortegal chromitites compared to the fields for chromites from boninites and MORB sources (after Pagé and Barnes, 2009).Literature data of the Cabo Ortegal chromitites taken from García-Izquierdo et al. (2011) for comparison.c) and d) Al 2 O 3 and TiO 2 contents of the melt in equilibrium with the Cabo Ortegal chromitites.Regression lines were obtained from Zaccarini et al. (2011) and are based on data from chromite-melt inclusions in MORB and arc lavas described by Kamenetsky et al. (2001) and Rollinson (2008).Literature data of the Cabo Ortegal chromitites taken from García-Izquierdo et al. (2011) for comparison.The deep and shallow fields correspond to the chromite and melt compositions of the Wadi Rajmi mantle chromitites in Oman (Rollinson, 2008) and are used for comparison.
The ages for the protoliths of the ophiolitic massifs involved in the European Variscan Orogen ranges between Proterozoic to Devonian (Table 2).However, the age of protoliths for the ophiolites of the Central and Eastern Rhodopes (Dobromirtsi, Chernichevo, Golyamo Kamenyane, Avren, and Yakovitsa) remains undetermined.Indeed, the protoliths of these metamorphosed mafic-ultramafic rocks have been considered to be Precambrian, Paleozoic or even Mesozoic (see Colás, 2015;González-Jiménez et al., 2015a and references therein).Interestingly, the ophiolite bodies of the Iberian massif, Herbeira (Cabo Ortegal) and Bragança have similar protolith ages of ca.~ 500 Ma (Santos Zalduegui et al., 2002;Tilhac et al., 2017) and 495-488 Ma (Mateus et al., 2016), respectively.The mafic-ultramafic rocks of Bragança are equivalent to those from the Herbeira massif of Cabo Ortegal (Santos Zalduegui et al., 2002 and references therein) and have also been linked to the development of a Cambrian-Ordovician arc (Andonaegui et al., 2002;Santos Zalduegui et al., 2002).
In terms of their geodynamic setting, host rocks, textures, PGE contents, PGM assemblage, and Cr#, the studied Cabo Ortegal chromitites are similar to other Variscan chromitites, particularly those documented in the Bragança massif in northern Portugal and those in the Kraubath massif in Austria.

Concluding remarks
We interpret that the studied chromitites of the Cabo Ortegal Complex formed in a sub-arc mantle environment probably in a fore-arc region.Besides, we interpret that Type-I chromitites crystallized from FAB-like melts in an early stage and Type-II chromitites from boninitic melts in the ensuing stage.The studied chromitites are enriched in PGE, showing PPGE enrichment with respect to IPGE.Observations that PGM associated with BMS are systematically located at the edges of chromite and in the silicate matrix of the chromitites suggest that the origin of the PGE enrichment is related to S-As rich melt(s) that collected base-metals and incompatible PGE.These melts were segregated by immiscibility from volatile-rich small volume melts.The PGM assemblage observed in the studied chromitites reflects the direct crystallization from silicate/ sulfide melt(s), followed by fractionation of sulfide melt(s) and subsequent subsolidus reequilibration.The origin of these chromitites together with their PGE enrichment is linked to the developing of an arc of at least Cambrian-Ordovician age.
Post-magmatic processes related to the percolation of hydrated fluids during retrograde amphibolite metamorphism (~8 kbar, 500 • C) during the exhumation at c. 380 Ma of the HP-HT rocks of the Capelada Unit affected dunites, pyroxenites, and chromitites of the Herbeira massif.This led to local remobilization of PGE, promoting alteration of primary PGM and formation of secondary PGM, but preserving the magmatic composition of the chromite cores in the chromitites.
Finally, the studied Cabo Ortegal chromitites in terms of their geodynamic setting, host rocks, textures, PGE contents, PGM assemblage, and Cr# are similar to the Variscan chromitites documented in the Bragança (northern Portugal) and Kraubath (Styria, Austria) ophiolitic massifs.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 3 .
Fig. 3. Geological map of the Herbeira massif showing the main geological units and the location of the studied chromitites.The black star indicates the sampling location.Modified from Sánchez Martínez et al. (2007), Arenas et al. (2016), and Tilhac et al. (2020).
; Appendix F: Fig. F.6). Besides, Pt arsenides and Pt-Fe-Ni sulfides are found in the silicate matrix as single irregular shaped grains (Appendix F: Fig. F.7). On the other hand, Pt-Ni-Fe and Pt-Fe alloys appear as single aggregates that can reach up to 7 µm.They are mostly found in open fractures of chromite grains (Fig. 13b).Composite grains of Rh-Ni-As are irregular in shape and show complex intergrowths with Ir-Fe-Ni and Pt-Fe-Ni-Cu (− Ir-Pd) alloys (Fig. 13c-e; Appendix F: Fig. F.8

Fig. 12 .
Fig. 12. Backscattered electron images of representative PGM and base-metal sulfides included on the edge of chromite from thin sections of the Cabo Ortegal chromitites.a) Pt-Fe-Cu-Ni alloys with base-metal sulfides, amphibole and chlorite in chromite.b) Pt-Pd sulfide grain included in chromite.c) Polygonal composite grain of Rh-Ir-Pt bearing sulfarsenide with Ir-Fe-Ni alloy.d) Composite grain of Pt-Pd and Ni-Fe sulfides with Pt-Pd-Fe alloys and Fe-oxides.e) Porous Pt-bearing alloy.f) Pt-Pd arsenide with Ni-Fe sulfide and Fe-oxides.g) Subrounded composite grain of Ni, Ni-Fe and Cu-Fe sulfides.h) Partially desulfurized Ni-Fe sulfide.Abbreviations: Amph: amphibole, Chr: chromite, Chl: chlorite.

Fig. 13 .
Fig. 13.Backscattered electron images of representative PGM and base-metal sulfides in fractures and in the altered silicate groundmass from thin sections of the Cabo Ortegal chromitites.a) Pt-Pd sulfide embedded in chlorite.b) Pt-Ni-Fe alloy.c) Composite grain of Rh-Ni-As with Ir-Fe-Ni and Pt-Fe-Ni-Cu alloys.e) and f) detail pictures of the grain in c) showing the textural relation between Rh-Ni-As, Ir-Fe-Ni and Pt-Fe-Ni-Cu.f) Fe-Ni-Cu sulfide in a fracture of chromite.Abbreviations: Chr: chromite.

Fig. 14 .
Fig. 14.Backscattered electron images of representative PGM and base-metal sulfides from hydroseparation concentrates of the studied chromitites.a) Subidiomorphic grain of RhAsS (hollingworthite).b) Same grain in a) but polished.c) Pt-Fe-Ni alloy.d) Porous Pt-Fe-Ni alloy grain.e) Pt-As grain.f) Composite grain of Rh-Ni-As with Ir-Fe-Ni, Ir-Fe-Pt-Ni-S, Pt-Fe-Ni-S-As, Pt-S-Fe-Ni-Cu and Ni-Fe-Co sulfide.h) Same grain in f) showing the distribution of the Ir-Fe-Pt-Ni-S and Ir-Fe-Ni phases.The red square shows a detailed view in i).i) Rh-Ni-As with intergrowths of Ir-Fe-Ni alloys.Grains of Ni-As showing triple junctions.

(
Fig. 15.Backscattered electron images of representative PGM and base-metal sulfides from hydroseparation concentrates of the studied chromitites.a), b) and c) Porous Pt-Fe-Ni alloys.d) Same grain in c) but polished.e) and f) Polished grains of Pt-Fe alloys with porous textures.g) Cu-Pt-Fe (Pd) alloy with Fe-Cu-Ni sulfide.h) Polished grain of Cu-Pt-Fe-Pd alloy with Ni-Fe sulfide.i) Subidiomorphic grain of Pt-Ir-Ru-Rh-As-S.j) Same grain of Pt-Ir-Ru-Rh-As-S from i) but polished.k) Idiomorphic grain of Pt-Ir-Ru-Rh-As-S.l) Idiomorphic grain of PtAs 2 (sperrylite).m) Composite grain of Rh-Ni-As with intergrowths of Pt-Ni-Fe and Fe-Cu-Ru-Os-Ir.n) Composite grain of Rh-Ni-As with Ir-Fe-Ni, Pt-Fe-Ni-Cu-Ir and Pt-Ir-Rh-As-S.

Fig. 16 .
Fig. 16.Compositions of Pt dominated alloys from chromitites of Cabo Ortegal, in the Pt-Fe-Ni + Cu ternary diagram.Colored squares show the composition of the Pt-bearing alloys in the inset.Blue, red, and violet squares correspond to analyses 1-3, 4-6 and 22-29 respectively (Table B.5).

Fig. 17 .
Fig. 17.Element distribution map of the free PGM of Fig. 14f from the Type-I chromitites of Cabo Ortegal.

Fig. 18 .
Fig. 18.Element distribution map of the free PGM of Fig. 15n from the Type-II chromitites of Cabo Ortegal.
and 18) could suggest that zaccariniite and PGE-base metal alloys represent the products of desulfurization of Pt-Ir-Rh-As-S PGM.According to Farré-de-Pablo et al. (2022), the loss of S of primary PGM is compensated by the incorporation of Fe, Ni, Cu (Mn) from metamorphic fluids.In this context, desulfurization of PGM of Pt-Ir-Rh-As-S and addition of PGE and base-metals (i.e.Ni and Fe) by metamorphic fluids could have triggered the formation of zaccariniite and PGE alloys in the studied chromitites.

Fig. 19 .
Fig. 19.Scanning transmission electron microscopy (STEM) image of the whole thin-foil obtained from the PGM of Fig. 15m and corresponding TEM-EDS elemental mapping.

Fig. 20 .
Fig. 20.A) high-magnification hrtem image of one intergrowth (see text for details), showing polycrystalline aggregates of nanoparticles.b) stem image of the same intergrowth and corresponding tem-eds elemental mapping.
also reported Sm-Nd and Rb-Sr clinopyroxene-amphibole "isochrons" from pyroxenites in the ranges 332-298 Ma and 352-322 Ma, respectively.All ages except the ca.390 Ma Sm-Nd garnet-clinopyroxene-whole rock age indicate isotopic disequilibrium, while the Devonian 390 Ma age of Santos Zalduegui

Table 1
Summary of PGM grains found in the Cabo Ortegal chromitites after hydroseparation.