Mantle Xenoliths from Huanul Volcano (Central-West Argentina): A Poorly Depleted Mantle Source under Southern Payenia

Huanul is a shield volcano with several lava flows hosting mantle xenoliths erupted during the Pleistocene (0.84 ± 0.05 Ma). It is located in the southern part of the Payenia Volcanic Province, which is among the largest Neogene-Quaternary volcanic provinces of South America. The volcanism here has been ascribed as the northernmost expression of the back-arc volcanism of the Andean Southern Volcanic Zone. We present the first petrographic and mineral chemistry study of mantle xenoliths collected from Huanul lavas with the aim of reconstructing directly the mantle source of the Payenia Volcanic Province. Xenoliths are commonly small (<5 cm in radius) but scarcely crossed by basaltic veins. All xenoliths have a fertile lherzolitic modal composition and are equilibrated in the spinel-facies. Most of them exhibit an almost primitive-mantle geochemical affinity, characterized by slightly depleted clinopyroxene REE patterns reproducible by partial melting degrees between 0 and 4% of a PM source. Geothermobarometric P-T estimates of clinopyroxene-orthopyroxene couples form a linear trend between 10 and 24 kbar with constant increase of T from 814 to 1170 ◦C along a 50–60 mW/m2 geotherm. Evidences of interaction with the host basalts occur as spongy textures in clinopyroxene and reacted spinel, which tend to became more restitic in composition and show chromatographic or complete overprinting of the trace element compositions. The presence of plagioclase and calculated P-T values constrain this melt/rock reaction process between 6 and 14 kbar, during magma ascent, and fit the mantle adiabat model. Calculated melts in equilibrium with the primary clinopyroxenes do not fit the composition of the host basalt and, together with the geothermobarometric estimations, point to an asthenospheric mantle source for the magmatism in southern Payenia. The PM geochemical affinity of the xenoliths of Huanul is an extremely rare finding in the South America lithospheric mantle, which is commonly extensively refertilized by subduction-derived melts.


Introduction
The territories of South America located south of 33 • S are famous for the occurrence of mantle xenoliths within basaltic lavas and pyroclastic structures. Many of these volcanic fields and xenoliths have been recognized and investigated in the last two decades increasing our knowledge on the general characteristics of the subcontinental lithospheric mantle (SCLM), as well as the detailed structure of the mantle column at several localities (e.g., [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15]). A complicating factor in the study of these mantle xenoliths is the development of the Pacific subduction, currently represented by the Nazca and Antarctic oceanic plates, below the South American continent. This process has profoundly influenced at many locations the chemistry of the subcontinental mantle up to ca. 700 km of distance from the Chilean trench, due to the introduction of slab-derived components along the back-arc region. Therefore, over time, volcanism and subduction have contributed in several episodes to the depletion, metasomatism and refertilization of the SCLM, and the results of these processes are commonly archived by mantle xenoliths transported to the surface by the Tertiary-Quaternary magmas. According to various studies, multiple events of metasomatism, depletion and refertilization are widely recorded by xenoliths from the entire Patagonia, and only very rarely are xenoliths here representative of a primary primitive SCLM (e.g., [7,9,[11][12][13][14][16][17][18][19] and references therein).
In this paper, we present mineral chemistry data of the first occurrence of xenoliths located in the Southern Payenia Volcanic Province (Mendoza province, Argentina), where it was recently proposed, through a study of the lavas, that the mantle source has been affected by refertilization processes [20]. Xenoliths were collected from lava flows of the Huanul shield volcano. Mineral phases were analysed for major (olivine, orthopyroxene, clinopyroxene, and spinel) and trace elements (clinopyroxene). We aim to reconstruct the chemical and geothermobarometric features of the Huanul mantle column prior to the xenoliths' ascent and prior to the interaction with the alkaline melts transporting media.

Geological Setting
The Huanul volcano is located in the southern part of the Payenia Volcanic Province (hereafter Payenia), at 37 •   The Cenozoic magmatism that formed this province affected the central-western region of Argentina and produced volcanic fields in Mendoza, La Pampa, and Neuquén provinces. This igneous activity was almost continuous from the Miocene to the Holocene and mainly formed a wide basaltic plateau that covers the extra-Andean back-arc [21][22][23][24][25][26][27][28]. A gap in the volcanic activity related to regional geodynamics has been identified between the late Miocene and the early Pliocene. Here the stress regime has changed from compressive during Miocene [29,30] to extensional in the Pliocene-Holocene [23,28,30]. This change and the migration of the volcanic activity toward the west were likely controlled by the steepening of the subducted Nazca plate that led to the ascent of the asthenosphere [29,[31][32][33][34].
Payenia is located between 460 to 540 km east of the Chile Trench and comprises the largest Neogene-Quaternary volcanic province of South America. The volcanism is manifested as numerous basaltic cinder cones, few andesitic-dacitic composite volcanoes, and lava flows, several of which are extraordinarily long, up to 180 km (e.g., [25,[35][36][37]). Its southern part is occupied by a small volcanic field named Río Colorado Volcanic Field [25,28,[37][38][39] constituted by pahoehoe lava flows in a northwest-southeast direction. Here, the most important eruptive center is Huanul, a shield volcano ( Figure 2) built from basaltic lava flows and pyroclastic aggregates. Its crater is 4 km in diameter and rises ca. 20 m from the surrounding [21]. The basaltic flows overlay the sedimentary rocks of the Neuquén Basin (Jurassic to Paleogene) and the Quaternary sediments of the Colorado river. Eight lava units are currently recognized in the field, and the oldest one, named El Corcovo lava flow, covers a distance of ca. 70 km for an area of ca. 415 km 2 [21]. El Corcovo consists of alkaline basalts with a K-Ar age of 0.84 ± 0.05 Ma [40]. Detailed information on the Huanul volcanism and eruption types are reported by [28,35,41,42]. Here, we summarize the characteristics of the Huanul lava flow from which our xenoliths were collected. They are porphyric basalts with holocrystalline, intergranular and felted groundmass, and subordinate subophitic and intersertal microtextures. The latter is more frequent in the superficial layer of the lava flow. Phenocrysts are mainly olivine and plagioclase (15 to 30%) and minor clinopyroxene (<2%). Olivine is on average 1.2 mm in size (up to 2.5 mm). It commonly exhibits subhedral to euhedral shapes, with short columnar habits with well-developed pseudohexagonal and hexagonal sections that include equidimensional opaque minerals of the spinel group. It is partially alterated to low temperature iddingsite (IBT) at the edges and fractures, and sometimes the central sectors develop high temperature iddingsite (IAT). The plagioclase phenocrysts are subhedral, mainly aligned along the flux direction, and with maximum sizes up to 6 mm, mainly around 1 to 2.5 mm. Clinopyroxene phenocrysts and microphenocrysts are characterized by subhedral morphologies and columnar to prismatic habits, up to 1.5 mm and on average 0.6 mm in size.
The geochemical data of the host lava were reported by [41] with compositions in the range of variability of ocean island basalts (OIB) and transitional basalt of Patagonia [43].

Analytical Methods
Nineteen xenoliths were collected for petrographic and geochemical investigations. Thirteen samples were selected for major element analyses of the mineral phases. Among them, ten samples were selected for trace elements analyses on clinopyroxene.
Major elements of the main mineral phases (olivine, orthopyroxene, clinopyroxene, and spinel) were determined at the Dipartimento di Scienze della Terra of the Università di Milano (Italy) by a JEOL JXA-8200 electron microprobe with 15 kV accelerating voltage, 15 nA beam current, 1-3 µm beam diameter, 30 s counting time on the peaks and 10 s on the background. Natural minerals (olivine for Mg; omphacite for Na; ilmenite for Ti; rodonite for Mn; K-feldspar for K; anorthite for Al and Ca; wollastonite for Si; fayalite for Fe, and nicolite for Ni) and synthetic chromite for Cr were used as standards. The results were corrected for matrix effects using the conventional ZAF method provided by the JEOL suite of programs. Results are considered to be accurate within 2-6%. Details on the method are reported in [44]. Data are presented in the Supplementary Materials Table S1.
Trace elements spot analyses on the selected clinopyroxene samples were determined by Laser Ablation-Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS), using a Thermo Fisher Scientific mass spectrometer coupled to a laser ablation New Wave UP 213 at the Laboratory of the Centro Interdipartimentale Grandi Strumenti (CIGS) of the Università di Modena e Reggio Emilia. Detailed method is reported in [45]. Data reduction was performed with the Thermo Fisher Scientific PlasmaLab ® software using NIST610, NIST612 and NIST614 as external standards. The isotope 44 Ca was used as internal standard for clinopyroxene. Laser spot size was calibrated at 50 µm and laser beam fluency at 20 microJoule for cm 2 . Data are reported in the Supplementary Materials Table S2.

Petrography
The Huanul xenoliths are small, commonly smaller than 5 cm, and rounded. The modal composition of the nineteen samples indicates that all samples are lherzolite equilibrated in the spinel-facies ( Figure 3a; Table 1).  Textures after [46] vary from porphyroclastic (10 samples) to mosaic-porphyroclastic (6 samples) to granuloblastic (3 samples) ( Figure 4A-D). No correlation is recognized between modal composition and texture ( Figure 3). Basaltic veins are present in most of the xenoliths in percentages varying from 1.02 to 9.45 vol.% (Table 1) with only two samples being vein-free. Veins are thin, only few µm thick, and develop at the borders with minerals ( Figure 4A). We commonly recognize their shapes up to the contact with the host basalt. Similar to the host basalt, basaltic veins are holocrystalline and formed by a groundmass with association of olivine ± plagioclase ± clinopyroxene in variable abundances. Olivine at the contact with the vein does not show any change while orthopyroxene, clinopyroxene and spinel tend to form reaction structures. Two populations of olivine are commonly observed. The larger olivines (of ca. 4-5 mm) are anhedral and show kink-bands. Smaller, recrystallized olivine grains (0.5-2 mm) do not have kink-bands and develop triple junctions. Orthopyroxene is anhedral with maximum length, ranging from 2 to 8 mm. Commonly, it develops exsolution lamellae of clinopyroxene, but, when in contact with the hosting lava, it shows resorption rims with disappearance of exsolution lamellae and formation of lobate shapes ( Figure 4D), thus evidencing dissolution features ( Figure 4F). Clinopyroxene has maximum width ranging between 1.5 and 3.5 mm, is anhedral and exhibits exsolution lamellae of orthopyroxene. When in contact with basaltic veins or the host basalt, clinopyroxene tends to assume a spongy texture, with formation of new, un-exsolved, pecilitic clinopyroxene with inclusions of olivine and plagioclase ( Figure 4G,I). This recrystallization is either complete or partial, with cores commonly characterized by exsolution lamellae and spongy rims. Not all samples show this feature, which depends on the presence of basaltic veins in contact with clinopyroxene. Two samples are characterized only by spongy clinopyroxene: mosaic sample HU14, where clinopyroxenes are almost completely spongified, and porphyroclastic sample HU15, where clinopyroxene cores are surrounded by spongy domains.
Spinel is interstitial, up to 2 mm, anhedral, and frequently holly leaf shaped. It is brown in color and can form crystal aggregates ( Figure 4H). Similar to clinopyroxene, spinels show reaction textures with the formation of spinel-2 coronas in a matrix of olivine and plagioclase when in contact with basaltic veins ( Figure 4I).

Clinopyroxene Trace Elements
The trace elements composition of Huanul clinopyroxenes is very homogeneous within sample. Thus, in the diagrams, we use average values ( Table 2) as representative of each sample. The complete dataset is reported in Supplementary Materials Table S2. Clinopyroxene REE patterns are mainly depleted in LREE and flat-like in M-and H-REE ( Figure 6).  [48]. The pattern of a primitive mantle-clinopyroxene (solid grey line) of Sun and McDonough [49] and patterns (dotted gray lines) after 1, 3, 5, and 10% of depletion by partial nonmodal melting are also shown. Pattern of MORB [50] and OIB [49] are reported for comparison.
These patterns are similar to those of clinopyroxenes produced by low degree of partial melting of a PM source (up to 3%), calculated from the composition of a PM [50] using the partition coefficients of Ionov et al. [51] for a spinel-facies mantle peridotite. The LILE and HSFE abundances of these clinopyroxenes are low with negative anomalies of Pb, Ti, Ba and Nb ( Figure 6).
The porphyroclastic cores with spongy rims of HU15 clinopyroxenes are characterized by almost flat M-to H-REE patterns with enrichments in La and Ce. These patterns from Lu to Eu follow closely those of a clinopyroxene derived from a PM source after 5% of partial melting ( Figure 6). The extended trace elements pattern is somewhat similar to the other porphyroclastic clinopyroxenes but with positive anomalies for U, Th and Sr.
The HU14 spongified clinopyroxene shows a convex-upward REE pattern ( Figure 6). Its extended trace element pattern is more enriched compared to porphyroclastic clinopyroxenes ( Figure 6) and shows a strong positive Ba anomaly and a negative Pb and moderately negative Zr-Hf anomalies.
An extensive discussion about the precision and accuracy of the geothermometers used in Table 1 is reported in [56,57]. According to the latter, the best P estimates for spinel-facies peridotites can be obtained with the geobarometer of [53]. The P values calculated for Huanul using equations of [53] were then used to estimate T, according to each thermometer. We discuss the T data of primary pyroxenes using the two-pyroxenes geothermometer of Brey and Köhler [54] because it is the most suitable for peridotite xenoliths in spinel-facies ( [56]) and is widely used in the literature for Patagonia mantle xenoliths (e.g., [4,10,11]). We are using instead the geothermobarometer equations of Wang et al. [58] to discuss spongy clinopyroxene because these equations are an implementation of those of Putirka [59] and are suitable for samples equilibrated in the plagioclase facies. The estimated T obtained by different geothermometers are in good agreement, showing parallel trends and almost constant difference in values (Table 3).
Primary pyroxenes show almost linear trends from 25 to 10 kbar (Figure 7; Table 3).   Porphyroclastic and mosaic textured samples form a unique trend, while cores of spongy clinopyroxenes from granuloblastic sample HU14 form a parallel trend at higher T. Temperature estimates for mosaic and porphyroclastic samples range between 814 and 1117 • C, while granuloblastic sample HU14 varies between 1089 and 1222 • C (Figure 7; Table 3). It is interesting to note that xenoliths from the Agua Poca volcano show a geotherm trend higher than Huanul samples and comparable to the trend of the granuloblastic sample HU14 (Figure 7). Comparing Huanul xenoliths T-P estimations with the geotherm calculated with the equations of Hasterok and Chapman [60] based on surface heat estimations, mosaic and porphyroclastic samples show a steeper trend between 50 and 60 mW/m 2 , while granuloblastic and Agua Poca samples are comprised between 60 and 70 mW/m 2 (Figure 7). Spongy clinopyroxenes show a higher geotherm at lower P compared to primary pyroxenes and Agua Poca (Figure 7). Although the equations of Wang et al. [58] are fitted for P between 0-12 kbar, the good agreement of values up to 14 kbar (better than values obtained from Putirka, [59] Table 3) suggests that these values are not meaningless. Spongy clinopyrox-ene estimates range between 6.3 and 14 kbar and between 1238 and 1319 • C, forming a trend that fits the geotherm adiabat curve of an ascending asthenospheric mantle (Figure 7).

Primary Poorly Depleted Spinel-Facies Mantle Column beneath Huanul Volcano
The mineralogical association suggests that Huanul xenoliths have been equilibrated in the spinel-facies stability field. Accordingly, the calculated P of equilibration ranges between 10 and 25 kbar. No evidence of a deeper provenance such as garnet breakdown or garnet relics was found.
The mineral modal composition (i.e., spinel-facies lherzolites; Figure 3) coupled with major and trace elements composition of the mineral phases ( Figures 5 and 6) suggest that melting and metasomatic processes did not affect or only slightly affected the spinel-facies mantle column in this region. A similar hypothesis was suggested for xenoliths from Agua Poca volcano (also located in Payenia) [10], which also show similar major elements minerals composition with the exception of the spongy clinopyroxenes ( Figure 5). Noticeably, a group of xenoliths from Agua Poca have spinels with high Cr# and low Mg#, similar to our reacted spinels. However, no reaction textures were reported for Agua Poca xenoliths.
Clinopyroxene REE patterns are consistent with residues after 1 to 3% of nonmodal partial melting of a PM composition ( Figure 6). Similar degrees of partial melting have been obtained using different equations based on spinel Cr# (i.e., [61][62][63]), with F (%) values mainly between 0 and 4% ( Figure 8; Table 3) and by OSMA diagram ( [64]). As expected, a direct relationship between F% and olivine content is recognized.
Temperatures estimated for the spinel-facies Huanul xenoliths are lower compared to other Patagonian localities, in agreement with a colder mantle column that underwent low degrees of partial melting (Figure 7). Noticeably, Huanul xenoliths mainly plot along a geotherm colder than Agua Poca. However, granuloblastic sample HU14 follows a trend similar to Agua Poca (Figure 7), thus suggesting that the granuloblastic texture could be related to localized heating of the mantle column. This is also in agreement with degrees of melting up to 13% estimated for Agua Poca [10].
Calculated melts in equilibrium with the LREE-depleted xenoliths have a MORB-like geochemical affinity (Figure 9), different from the host basalt HU06.
This suggests that the source of the Huanul magmatism was deeper than 25 kbar, possibly in the asthenosphere. The steeper T-P trend of Huanul and Agua Poca xenoliths compared to common geotherms of the subcontinental lithospheric mantle indicates possible melt upwelling pathways from the asthenosphere (Figure 7). This is also in agreement with the spongy clinopyroxene estimations which fit the modeled adiabat curve.  This evidence also suggests that the slight depletion of the mantle column was an inherited feature and was not induced by percolation of the host basalt melt. According to Chilson-Parks et al. [20], part of the Payenia basalts derive from an asthenospheric source with geochemical affinities to the Discovery-Shona MORB source. Based on major and trace element abundances and isotope ratios of the Payenia basalts, they suggest that the Payenia SCLM was twice metasomatized 50 to 150 Ma. One melt was generated from the Discovery-Shona MORB source, while the second one derived from the mantle-wedge during subduction of the Pacific plate. In this framework, the primary composition of our xenoliths could represent the source of this magmatism with MORB-affinity. Regional geophysical [65] and geodynamic [66] studies of the Payenia region have revealed a low seismic velocity structure in the region down to 200 km, indicative of upwelling of deep, hot asthenospheric material. This Neogene structure has been related either to the steepening of the subducted Nazca plate [65] or to the subduction of the hypothetical 'Payenia' mantle plume [66]. On this basis, the primary texture and mineralogical and geochemical features of our xenoliths might have recorded the original characteristics of the subcontinental lithospheric mantle prior to the Neogene asthenospheric upwelling, while the spongy and reaction textures might reflect the latest interaction with the upwelling mantle.

Xenolith-Melt Interaction during Magma Ascent
The majority of Huanul xenoliths do not show chemical evidences of refertilization or melt/rock interaction processes. However, two main reactions are recognized in the mineral textures of HUA14 and HUA15 samples mainly affecting clinopyroxene and spinel phases. The first one is testified by the replacement of exsoluted clinopyroxene by spongy clinopyroxene (Figure 4G,I). The second reaction produced recrystallization of spinel and formation of reacted spinel structures ( Figure 4I). In both cases, neoblasts of olivine and plagioclase in a fine matrix (in spinel reaction) or as inclusions (in spongy-clinopyroxene) are part of the new mineralogical association. The occurrence of plagioclase as stable phase in the reactions supports a shallow depth for the melt-xenolith interaction, thus suggesting that the migrating melt could be related to the host basalt and that the interaction occurred during the ascent of the xenolith. This hypothesis is also supported by the fact that these structures occur in crystals in contact with veins, which in turn are connected with the host basalt. Further evidence is the occurrence of olivine neoblasts within the recrystallization zone with Mg# values lower than those accepted for mantle olivine (Mg# = 77.5 − 65.1), thus possibly derived from direct crystallization of the migrating melt. The direct crystallization of olivine neoblasts together with the resorption of orthopyroxene, as testified by lobate shapes of orthopyroxenes at the contact with the host basalt, suggests that the migrating melt was Si-poor, thus consistent with the host basalt HU06 [41]. In addition, the trace element composition of partial and complete spongy clinopyroxenes (samples HU14 and HU15) is more enriched in Ba, Nb and Ta compared to the other xenoliths ( Figure 6). In particular, partially spongy clinopyroxene cores from sample HU15 show enrichments in LREE comparable with chromatographic effects related to melt-rock reaction processes as observed in other Patagonian mantle xenoliths (e.g., [11,14]) and in the Fernando de Noronha xenoliths (Figure 10b,c of Rivalenti et al. [67]). Conversely, the convex-upward REE patterns of HU14 spongy clinopyroxenes are commonly interpreted as evidence of a prolonged interaction with an alkaline melt (e.g., [11]). Thus, these spongy clinopyroxenes can provide constrains on the composition of the percolating melt. Calculated melts in equilibrium with these clinopyroxenes using the coefficients of Ionov et al. [51] are similar to the host basalt HU06 (Figure 9), thus suggesting infiltration of the host basalt.

Conclusions
This is the first petrographic and geochemical study of mantle xenoliths entrained in the lava flows of the Huanul shield volcano. We found evidences of a poorly-depleted, almost primitive subcontinental lithospheric mantle beneath the southern Payenia Volcanic Province. This is a rare occurrence among the many xenoliths collected from Argentina and Patagonia, where mantle xenoliths record multiple events of depletion and refertilization related to the Pacific subduction below South America. The Huanul xenoliths' geochemical and mineralogical features are similar to those of xenoliths from the Agua Poca volcano, a close locality that, however, has recorded higher degrees of partial melting.
Xenoliths in Huanul volcano are lherzolitic in composition and equilibrated in the spinel-facies between 10 and 25 kbar (~33-79 km). They mainly fall on a 50-60 mW/m 2 geotherm, with T increasing from 814 to 1170 • C, steeper than geotherm models for subcontinental lithospheric mantle. Interaction of the xenoliths with the host basalt during ascent is recorded by the formation of spongy textures in clinopyroxene and reacted spinel in contact with basalt veins or the host basalt. In these textures, the primary minerals are recrystallized toward more refractory compositions accompanied by segregation of plagioclase and olivine, which document a relatively shallow depth between 6-14 kbar for the melt-rock interaction event. The extremely high T estimations for spongy clinopyroxene follow the adiabat curve of subcontinental lithospheric mantle and are in good agreement with chemical evidences of interaction with the alkaline host lava.
Taken together our data indicate that this xenolith population of Huanul could represent the Discovery-Shona asthenospheric MORB source invoked by Chilson-Parks et al. [20] for the magmatism of the southern Payenia region. In addition, Huanul xenoliths do not show any evidence of a SCLM pertaining to a mantle wedge domain metasomatized by the subducting slab, thus making the Payenia Volcanic Province a unique and important location to investigate the primitive geochemical character of the SCLM of South America.

Data Availability Statement:
The data presented in this study are available in this article and its Supplementary Materials Tables S1 and S2.

Conflicts of Interest:
The authors declare no conflict of interest.