Magmatic Processes at Euganean Hills (Veneto Volcanic Province, Italy): Clinopyroxene Investigation to Unravel Magmatic Interactions

Abstract

The Euganean Hills (NE Italy) magmatic district represents the final volcanic activity of the Veneto Volcanic Province. Alkaline to subalkaline magmatic suite dominated by intermediate to felsic volcanic rocks characterises the latest volcanic activity of the Euganean Hills. Magmatic (intrusive and volcanic) enclaves are common in Euganean Hills trachytes. We used the ability of clinopyroxene to record variations of P, T, and fO${}_2$ to reconstruct the geological history of the volcanic enclaves and trachytic host. Despite similar major and trace elements composition, clinopyroxene from host is higher in Ca and Na (and Fe${}^{3+}$) and lower in Mg than enclaves and is slightly enriched in trace elements but with the same pattern distribution. Minor differences in geochemistry and crystal structure of clinopyroxene from enclaves and trachytic host suggest similar parental magmas that differs by small degrees of fractional crystallisation. Clinopyroxene geobarometry performed combining X-ray diffraction with mineral geochemistry for volcanic enclaves–trachytic host combined with amphibole geobarometry for intrusive enclaves and crystal mushes points to a crystallisation pressure range between 4.8–2.0 kbars. Our data support the model of a complex system of magma chambers at intermediate to shallow crustal level where mafic magma accumulated, evolved by fractionation processes and mixed.

1. Introduction

The Euganean Hills’ (Figure 1) volcanism represents the final stage of the magmatism developed during the Alpine orogenesis in the Southern Alps and is related to the extensional geodynamic regime active in the area from the middle Eocene to the Lower Oligocene [1,2]. In particular, it represents the final volcanic activity of the Veneto Volcanic Province (VVP). The Euganean Hills volcanic activity started at 42.0 ± 1.5 Ma [3] with the emplacement of submarine lavas (alkaline to subalkaline mafic magma), while a younger volcanic event occurred at 34 ± 2 Ma [4] with the emplacement of intermediate to felsic volcanic rocks [1]. The Euganean Hills volcanic rocks are mostly composed of trachytes and rhyolites with moderate Na-alkaline affinity, and a minor presence of latites and basalts [5]. A peculiarity of the Euganean Hills trachytes is the presence of xenoliths and enclaves of different compositions: magmatic enclaves, mafic and ultramafic cumulates, and metamorphic xenoliths [6]. The variety of enclaves enclosed in the Euganean Hills trachytic rocks testifies to the complex magmatic interaction processes active in the volcanic system: fractional crystallisation, crustal contamination, magma mixing/mingling, etc. [6].

In the present manuscript, we reviewed the literature data available for mafic and ultramafic cumulates, magmatic intrusive enclaves, and metamorphic xenoliths of the Euganean Hills trachytes, adding new data on the magmatic volcanic enclaves to provide further constraints on the architecture of the volcanic plumbing system and mixing processes of the Euganean Hills. In particular, we present new data on clinopyroxene crystals that are one of the mafic phases in magmatic volcanic enclaves and the trachytic host.

Among the rock forming minerals, clinopyroxene is of primary petrogenetic importance in many geodynamical environments. In fact, pyroxene is a liquidus phase in a wide range of volcanic rocks and, due to its complexity of cations substitution, it can record variation in the physico-chemical condition of crystallisation from the variation of intensive parameters such as P, T, and fO2 within the magmatic system in the period preceding eruption to sub-solidus processes (e.g., [7,8,9,10,11,12,13,14]). As a further result, we can use clinopyroxene chemical signature and structural parameters as a geobarometer to reconstruct the pressure of crystallisation (i.e., location of the magma chamber and possible residing in a shallow system before the eruption; [15,16]).

Following this approach, we selected clinopyroxene from magmatic enclaves and host trachytes to investigate their geochemistry and crystal structure to add new data on the petrological processes they testified (mixing/mingling) and to estimate their crystallization pressure. Combining clinopyroxene chemical data with volcanic bulk geochemistry of the magmatic enclaves and with the mafic and ultramafic cumulates, we aim to better constrain the magma storage conditions and the magmatic evolution of the final activity of the Veneto Volcanic Province.

2. Geological and Volcanological Outlines

Among the Permian–Cenozoic magmatic events which occurred during the Alpine orogenesis, the VVP represents the main volcanic activity in the Southern Alps foreland under an extensional tectonic regime [1,2,17,18,19]. A discontinuous magmatic activity characterises the VVP over 30 Ma; it started in the Late Cretaceous [20] with the eruption of Na-alkaline lamprophyre dykes and ended in the Late Oligocene [21], producing four main volcanic districts: the Lessini Mts, the Berici Hills, the Euganean Hills, and the Marostican Hills. The VVP magmas are characterised by ultrabasic to basic compositions, with relatively undifferentiated magmas including (mela)-nephelinites, basanites, and alkaline, transitional, and tholeiitic basalts [2,22,23,24]. The Euganean Hills district has the only occurring intermediate to felsic products [25].

The origin of the VVP magmatism is still debated and several models have been proposed to explain the petrogenesis of VVP rocks (e.g., [18,19,26,27]). The geochemical signature (trace elements and Sr-Nd isotopes ratios) of most of the rocks suggests an anorogenic intra-plate source [2] associated to an extensional tectonic setting [1,27]. Ref. [26] suggested that a mantle plume rose during the Paleocene, before the subducted European lithospheric slab intercepted it, and then during the middle Eocene, after slab detachment and opening of a plate window. On the other hand, Ref. [19] proposed that the westward rollback of the European slab caused (i) the retreat and steepening of the subducting plate, (ii) the contemporaneous upwelling of portions of the sub-slab mantle from the front of the slab, and (iii) the extensional deformation in the overriding Adria microplate. The geochemistry of some mafic and ultramafic VVP rocks is characterized by an enrichment of LILEs and depletion of HFSEs, suggesting the existence of a multistage metasomatism in the VVP lithospheric mantle [28,29,30]. This major heterogeneity of the VVP lithospheric mantle with respect to typical intraplate tectonic environments is consistent with the proposed geodynamic scenarios (see above). Euganean Hills represents the final volcanic activity of the VVP, characterised by volcanic rocks erupted over a 10 Ma time-span during two main events at 42.0 ± 1.5 Ma [3,19] by emplacement of submarine mafic lavas (breccias, hyaloclastites, and pillow lavas) of alkaline to subalkaline affinity and at 34 ± 2 Ma and 32.0 ± 3.5 Ma [4] by emplacement of intermediate to felsic volcanic rocks (domes, plugs, laccoliths, or dykes) [1]. The magmatic products are mainly alkali trachytes and rhyolites, with minor latites and basalts (Figure 2). Recent ${}^{40}$Ar/${}^{39}$Ar analyses on felsic products from Euganean Hills yielded Oligocenic ages (mean weighted age of 32.21 ± 0.09 Ma; [19]), in agreement with U–Pb zircon ages obtained from silica-rich enclaves (31.0 ± 1.3 and 30.6 ± 1.5 Ma; [6]). The petro-volcanological model proposed for the evolution of the Euganean Hills consists of a mantle-derived magma that produces intermediate to felsic rocks by low-pressure fractional crystallization with limited crustal contamination [25]. A complex system of magma chambers at shallow crustal level where mafic magmas accumulated and evolved towards felsic compositions seems compatible with the extensive cross-cutting fault system of the area and most likely responsible for the extensive evidences of magmas interaction at Euganei Hills [25].

2.1. Evidence of Magmatic Interaction at Euganean Hills

The trachytes of the Euganean Hills are characterised by the presence of magmatic enclaves, mafic and ultramafic cumulates, and metamorphic xenoliths [6,30,33,34,35]. Magmatic enclaves include intrusive volcanic enclaves. The latter group typically shows compositions from basaltic trachyandesite to trachyte and trachytic dacites, overlying the composition of most Euganean volcanic rocks (Figure 2). Intermediate to felsic intrusive enclaves were interpreted as magma chamber solidification fronts, with different degrees of crystallization [6]. Indeed, some intrusive enclaves are characterized by embayments of the host trachyte towards the core of the enclaves and were interpreted as evidence of crystal mush processes (hereafter crystal mushes). Some felsic intrusive show peculiar compositions suggesting a cumulate origin during trachyte–rhyolite evolution [6]. Mafic to ultramafic cumulites, instead, testify for deep-seated cumulates formed at, or close to, the crust–mantle boundary from Alpine subduction-related basaltic magmas. Subsequently, VVP magma dismembers and transports them to shallower levels [30].

Magmatic volcanic enclaves can be easily recognised on the hand specimen by colour; they are from grey to dark grey where the trachytic host is light grey. The contact between the host and the enclaves have an irregular shape, often showing the presence of folds towards the host. The enclaves are generally less porphiritic and have a more fine-grained matrix than the trachytic host. Both host and enclaves have plagioclase, biotite, pyroxene, and amphibole as phenocryst (Figure 3). Some enclaves have rounded vesicles into which are present calcite and zeolite.

Metamorphic xenoliths range from regionally metamorphosed gneissic to granulitic rocks, some of them showing the effect of pyrometamorphism (850–900 °C, 2.5 kbar) after their incorporation in VVP magmas [33,34]. Recently, a migmatitic xenolith gave the first evidence of a Permian (c. 260 Ma) thermal event associated with crustal thinning in the Eastern Southern Alps [35].

2.2. Crystal-Chemistry of Clinopyroxene from Euganean Hills Volcanic Rocks

Pyroxene (diopside/augite) together with amphibole (kaersutitic in composition) is one of the mafic phases present in the Euganean Hills volcanic rocks [36]. The chemical composition of the augitic clinopyroxene does not vary significantly from the different localities of the Euganean Hills, with Fs ranging from 13 to 22, En from 33 to 45, and Wo from 38 to 45. Euganean clinopyroxene are CaO rich (18–22 wt%) with Na${}_2$O in the range 0.51–1.87 wt%, low TiO${}_2$ (up to 2 wt%), and more variable MgO (10–16 wt%) and FeOtot (8–12 wt%) content [32,37,38]. Trace elements do not vary significantly, both for compatible elements such as Sc (97–339 ppm, average 193 ppm), V (15.93–88.5 ppm average 57.54 ppm), and Y (61–160 ppm average 95.6 ppm) and incompatible elements such as Sr (1.51–49 ppm average 18.33 ppm), La (18.56–61.70 ppm average 31.2 ppm), and Ce (68.2–210.9 ppm average 114.31 ppm).

The REE pattern normalised to CI chondrites [39] define a variation range for the Euganean literature clinopyroxene [36] with a slight enrichment of LREE respect to MREE and HREE. A marked negative anomaly in Eu is present [34].

3. Materials and Methods

3.1. Sample Collection

The sampling of the volcanic rocks was mainly performed in queries due to the strong urbanisation or the high vegetation present in the area. Most of the outcrops are strongly altered, making the sampling not trivial. Rock samples showing magmatic enclaves were collected at Montemerlo, Bagnara Alta (M. Madonna) and Lozzo Atesino (M. Lozzo) (see Figure 1). Enclave size varies from a few centimetres to tens of centimetres. The largest enclaves are extremely irregular in shape, where the smallest are typically sub-rounded.

3.2. Pyroxene Selections from Magmatic Enclaves

We selected four specimens (CEM6, CEM15, CEM18, and CEM33) having a clear boundary between enclaves and host rock, that have unaltered pyroxene crystals in both host (CEM6H, CEM15H, CEM18H, and CEM33H, respectively) and enclave (CEM6E, CEM15E, CEM18E, and CEM33E, respectively) portion. Thick rock sections (about 100 μm), showing the host and enclaves boundary have been made; then, under optical microscope, clinopyroxenes pairs from host and enclaves portions were hand-picked.

We named the pyroxene picked up from the enclave portion enclaves-clinopyroxene (enclaves-cpx), and the pyroxene picked up from the host portion host-clinopyroxene (host-cpx) (see Figure 4). Selected clinopyroxene crystals were used for the crystal chemical study combining X-ray diffraction, electron microprobe, and trace elements chemical analyses on the same single crystal.

3.3. Single-Crystal X-ray Diffraction (SCXRD)

Nine pyroxene single crystals were mounted on a four-circle Phillips PW1100 diffractometer (located at the Department of Physics and Geology, University of Perugia) equipped with graphite-monochromatized MoK$\alpha$ radiation for the crystallographic investigation. Data collection was carried out in the 3–31° $\theta$ range and the equivalent pairs of hkl and h-kl reflections were measured using the $\omega$-scan methods. An accurate measure of the lattice parameters were performed using the least-squares methods applied to the theta values of about 60 reflections. After the correction for Lorentz-polarization and absorption factors, the intensity data were merged to obtain a set of about 500 independent reflections having I > 4$\sigma \left(I\right)$. Anisotropic refinements were carried out in the space group C2/c starting from the diopside coordinate [9] using the least-squares program SHELXL [40]. Ionised scattering factors were refined to obtain occupancies of Mg${}^{2+}$, Fe${}^{2+}$, Al${}^{3+}$, and Ti${}^{4+}$ at the M1 site, Ca${}^{2+}$, Na${}^{+}$, Mg${}^{2+}$, and Fe${}^{2+}$ at the M2 (and M2’) site, mixed ion scattering factors of Si${}^{2.5+}$ at the T site, and O${}^{1.5-}$ at the O1, O2, and O3 sites. The difference Fourier map showed the presence of a residual electron density located about 0.6 Angstrom from the M2 site (site M2’), where Mg${}^{2+}$ and Fe${}^{2+}$ occupancy were refined. Further least-squares refinements were carried out. Less than 1 electron is found as residue in the electron density map in all the refinements. The final agreement factors R1 of the refinement and geometrical parameters are listed in

Table 1. Refinement details and crystallographic data for the studied clinopyroxenes. Lattice parameters, bond distances (in Å), polyhedral volumes (in Å${}^3$), angle (in degree), site distortions (quadratic elongation, $\lambda$ and angle variance, $\sigma 2$, [41]) are reported. Goof, Goodness of fit; R1, agreement factor (in %); Rint agreement factor of equivalent reflections (in %), and N number of unique reflections measured.

Sample	CEM6E	CEM6H	CEM15E	CEM15H	CEM18E	CEM18H	CEM33E	CEM33H	CEM33Hb
Sample	Enclave	Host	Enclave	Host	Enclave	Host	Enclave	Host	Host
a	9.740(6)	9.751(4)	9.742(5)	9.749(5)	9.747(5)	9.751(7)	9.7445(4)	9.751(4)	9.755(5)
b	8.916(5)	8.923(3)	8.916(4)	8.918(5)	8.928(4)	8.921(4)	8.920(3)	8.933(3)	8.929(4)
c	5.257(4)	5.261(2)	5.258(3)	5.262(4)	5.253(3)	5.259(4)	5.256(2)	5.256(3)	5.262(3)
$\beta$	106.19(3)	106.05(3)	106.21(4)	106.02(3)	106.13(4)	106.0(6)	106.23(3)	106.15(3)	106.15(4)
V	438.4	438.9(3)	438.6(4)	439.8(4)	439.1(4)	439.8(6)	438.7(3)	439.7(3)	440.2(4)
Rint	1.8	2.6	2.4	1.7	2.2	3.6	2.2	2.3	5.1
R1	2.6	2.3	2.2	2.1	2.3	2.6	2.5	2.3	5.3
2$\theta$ max	62	62	62	62	62	62	62	62	64
Goof	1.25	1.14	1.22	1.26	1.26	1.24	1.16	1.29	1.085
T-O2	1.589(2)	1.591(2)	1.592(2)	1.592(2)	1.590(2)	1.590(2)	1.593(1)	1.590(2)	1.595(1)
T-O1	1.607(2)	1.607(2)	1.611(2)	1.610(2)	1.604(2)	1.608(2)	1.607(2)	1.605(2)	1.610(1)
T-O3	1.662(2)	1.665(2)	1.662(2)	1.665(1)	1.664(2)	1.664(2)	1.664(1)	1.661(2)	1.666(1)
T-O3	1.679(2)	1.679(2)	1.680(2)	1.682(1)	1.680(2)	1.681(2)	1.679(1)	1.681(2)	1.683(1)
<T-O>	1.635(2)	1.635(2)	1.636(2)	1.637(2)	1.634(2)	1.636(2)	1.636(2)	1.634(2)	1.639(2)
VT	2.226(2)	2.227(2)	2.230(1)	2.233(1)	2.224(1)	2.229(1)	2.229(1)	2.223(2)	2.234(1)
$\sigma 2$	1.0057	1.0058	1.0056	1.0059	1.0058	1.0057	1.0057	1.0057	1.01
$\lambda$	24.060	24.691	23.753	25.125	24.844	24.243	24.117	24.394	24.65
M1-O2	2.045(2)	2.043(1)	2.042(2)	2.039(1)	2.049(2)	2.044(2)	2.045(1)	2.050(2)	2.046(1)
M1-O1	2.061(2)	2.068(2)	2.061(2)	2.066(2)	2.065(2)	2.067(2)	2.062(2)	2.069(2)	2.066(1)
M1-O1	2.132(2)	2.131(1)	2.130(2)	2.130(1)	2.135(2)	2.130(2)	2.134(1)	2.136(2)	2.137(1)
<M1-O>	2.078(2)	2.081(1)	2.078(2)	2.082(2)	2.083(2)	2.080(2)	2.080(2)	2.085(2)	2.084(2)
VM1	11.881(5)	11.916(4)	11.869(2)	11.873(1)	11.964(3)	11.912(3)	11.918(3)	11.997(3)	11.970(3)
$\sigma 2$	1.0055	1.0056	1.0056	1.0057	1.0052	1.0056	1.0053	1.0052	1.0052
$\lambda$	17.350	17.963	17.608	18.264	16.540	17.942	16.706	16.513	16.468
M2-O2	2.311(2)	2.329(2)	2.312(2)	2.332(2)	2.305(2)	2.332(2)	2.301(2)	2.305(2)	2.306(1)
M2-O1	2.346(2)	2.358(2)	2.346(2)	2.358(2)	2.344(2)	2.355(2)	2.341(1)	2.344(2)	2.344(1)
M2-O3	2.578(2)	2.569(2)	2.577(2)	2.566(2)	2.588(2)	2.569(2)	2.587(2)	2.588(2)	2.589(1)
M2-O3	2.740(2)	2.738(2)	2.742(2)	2.735(2)	2.739(2)	2.737(2)	2.741(2)	2.745(2)	2.740(1)
<M2-O>	2.494(2)	2.498(2)	2.494(2)	2.473(2)	2.494(2)	2.498(2)	2.492(2)	2.495(2)	2.496(2)
VM2	25.558(8)	25.737(7)	25.576(5)	25.725(6)	25.557(4)	25.741(6)	25.502(5)	25.599(5)	25.629(5)

.

3.4. Electron Microprobe Data

Electron probe microanalysis was carried out by a Cameca SX-50 electron microprobe by WDS at the IGGCNR (Padova, Italy). The operating condition was 15 kV and 15 nA, and an average of about five spots were analysed on each crystal. Site populations were obtained combining the microprobe and single-crystal refinement data, as described by [42], the cation recalculation is reported in

Table 2. Compositions (in weight percent) of major oxides for the studied clinopyroxenes. Cation partitioning based on six oxygen expressed in atoms per formula units.

Sample	CEM6E	CEM6H	CEM15E	CEM15H	CEM18E	CEM18H	CEM33E	CEM33H
Sample	Enclave	Host	Enclave	Host	Enclave	Host	Enclave	Host
SiO${}_2$	53.21(62)	52.43(37)	51.99(47)	51.92(60)	52.02(48)	53.36(54)	52.33(22)	51.64(31)
TiO${}_2$	0.26(7)	0.30(4)	0.38(7)	0.33(2)	0.29(8)	0.25(11)	0.26(5)	0.45(9)
Al${}_2$O${}_3$	0.58(32)	1.86(10)	1.90(4)	2.01(5)	1.71(17)	0.71(52)	0.95(6)	1.67(29)
Cr${}_2$O${}_3$	-	0.01(1)	0.02(1)	0.02(2)	-	0.01(2)	-	0.03(3)
FeO	9.43(0.25)	9.89(013)	10.03(27)	9.87(20)	11.20(41)	8.92(35)	11.30(86)	9.58(31)
MnO	1.29(7)	0.47(5)	0.89(4)	0.74(3)	1.07(9)	0.52(6)	1.21(7)	0.59(11)
MgO	14.37(28)	14.04(18)	12.54(23)	1 3.67(7)	11.87(19)	14.92(14)	12.92(43)	14.85(20)
CaO	20.81(28)	20.27(21)	21.37(20)	20.44(25)	21.23(20)	21.35(65)	20.46(20)	20.47(16)
Na${}_2$O	0.58(5)	0.81(5)	1.09(5)	0.84(2)	1.17(8)	0.48(7)	0.71(5)	0.66(7)
Oxide total	100.53(40)	100.09(63)	100.20(84)	99.84(88)	100.57(70)	100.52(56)	100.12(41)	99.92(41)
Fe${}_2$O${}_3$	1.77(59)	2.37(56)	3.55(49)	3.01(65)	4.03(1.09)	1.56(35)	2.46(34)	3.37(35)
FeO	7.84(0.56)	7.76(46)	6.84(32)	7.17(54)	7.58(85)	7.51(5)	9.08(95)	6.55(51)
Total	100.71(37)	100.32(65)	100.56(86)	100.14(86)	100.97(71)	100.68(57)	100.37(41)	100.25(43)
Si	1.974(17)	1.947(8)	1.937(8)	1.935(9)	1.940(17)	1.971(16)	1.962(6)	1.920(13)
Ti	0.007(2)	0.009(1)	0.011(2)	0.009(1)	0.008(2)	0.007(3)	0.007(2)	0.013(2)
Al	0.025(14)	0.081(4)	0.083(1)	0.088(2)	0.075(8)	0.031(23)	0.042(2)	0.073(12)
Cr	-	-	-	0.001((1)	-	-	-	0.001(1)
Fe${}^{3+}$	0.050(17)	0.066(15)	0.100(14)	0.084(18)	0.113(30)	0.043(10)	0.070(9)	0.094(9)
Fe${}^{2+}$	0.243(16)	0.241(15)	0.213(11)	0.223(16)	0.237(26)	0.232(3)	0.285(30)	0.204(16)
Mn	0.041(2)	0.015(2)	0.028(2)	0.023(1)	0.034(2)	0.016(2)	0.038(3)	0.019(4)
Mg	0.795(14)	0.777(7)	0.697(8)	0.759(4)	0.660(12)	0.822(5)	0.722(24)	0.823(10)
Ca	0.827(13)	0.806(5)	0.853(1)	0.816(6)	0.848(6)	0.845(24)	0.822(8)	0.815(6)
Na	0.041(4)	0.058(3)	0.078(3)	0.061(1)	0.085(6)	0.035(5)	0.051(4)	0.047(5)
Mg No	77	76	77	77	74	78	72	80
Wollastonite	44	44	48	45	49	44	45	44
Enstatite	43	43	40	42	38	43	39	45
Ferrosilite	13	13	12	12	14	12	16	11

. The mean number of electrons calculated for the site populations obtained on the basis of the chemical data differs with those from the site-occupancy refinement of less than one electron, suggesting a very good agreement between SCXRD and EMPA population.

3.5. Laser Ablation Inductively Coupled Plasma Mass Spectrometry

Trace element compositions were estimated by Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry (LA-ICP-MS) at the Department of Physics and Geology, University of Perugia. The instrumentation consisted of a Teledyne Photon Machine G2 laser ablation device coupled to a Thermo Fischer Scientific iCAP Q quadrupole mass spectrometer. A circular spot size with a diameter of 65 $\mathsf{\mu }$m, a repetition rate of 8 Hz, and a laser fluence of 3.5 Jcm${}^{-1}$ have been utilized. Ablation times were 50 s per spot, preceded by a 30 s background measurement and followed by 30 s of washout. Data reduction was performed by the Iolite 3 software [43]. The NIST SRM 610 [44] and the USGS BCR2G [45] reference materials were used as calibrator and quality control, respectively. Si was used as an internal standard. Under the reported analytical conditions, the precision is better than 10 percent for concentrations above 0.1 ppm, and better than 5 percent above 2 ppm; accuracy is always better than 10 percent [46,47]. An average over three analyses are reported in

Table 3. Compositions of trace elements (in ppm) for the studied clinopyroxenes. $1\sigma$ in parenthesis.

Sample	CEM18E	CEM15E	CEM33E	CEM6E	CEM15H	CEM33H
Sc	32.31(1.41)	92.20 (2.35)	122.77(2.00)	95.25(5.00)	180.708(3.90)	251.20(2.80)
Ti	1265(60)	1963(48)	2682(40)	1158(49)	2455(84)	1279(24)
V	65.00(4.23)	78.20(2.43)	86.47(1.77)	36.75(2.35)	78.40(2.30)	33.05(0.87)
Cr	4.50(8.30)	1.60(5.25)	5.10(4.53)	3.40(7.35)	0.40(4.40)	1.10(2.90)
Mn	3940(207)	5855(172)	4380(115)	8455(290)	7480(160)	10040(175)
Co	38.40(2.77)	23.25(1.11)	21.45(84)	25.00(1.15)	15.10(88)	9.51(53)
Cu	0.50(1.28)	0.77(1.01)	0.12(73)	0.95(1.55)	0.60(82)	bdl
Ga	5.66(78)	11.48(1.04)	9.58(76)	6.42(67)	13.02(94)	7.65(60)
Rb	0.01(24)	0.18(26)	0.29(24)	0.26(48)	0.12(19)	bdl
Sr	13.20(98)	41.10(1.40)	41.35(97)	5.19(90)	12.61(61)	3.31(20)
Y	82.80(5.00)	87.13(2.43)	93.67(1.57)	167.05(7.40)	77.60(2.30)	135.40(2.05)
Zr	40.53(2.23)	120.60(3.15)	155.43(2.73)	50.10(4.65)	164.10(6.00)	84.40(1.70)
Nb	0.520(136)	0.707(175)	0.424(115)	0.790(225)	1.650(230)	0.319(91)
Sn	1.993(770)	2.635(387)	1.720(270)	2.765(685)	5.940(590)	1.920(270)
Cs	0.0220(563)	0.0340(320)	0.0050(327)	0.0500(705)	bdl	0.0010(275)
Ba	0.120(197)	0.070(0.133)	0.190(187)	0.445(370)	0.050(160)	0.075(150)
La	23.753(1.470)	20.355(690)	17.920(477)	55.800(4.050)	34.510(0.700)	34.060(0.690)
Ce	91.70(4.83)	74.10(1.75)	69.00(1.23)	204.50(15.50)	122.70(2.80)	134.50(1.90)
Pr	16.030(967)	13.030(492)	12.673(350)	3.70(2.80)	19.48(49)	24.045(485)
Nd	80.20(5.10)	69.775(2.900)	72.167(1.900)	169.50(12.50)	98.00(3.00)	129.35(2.75)
Sm	23.767(3.167)	22.475(1.475)	23.680(1.093)	48.550(3.700)	26.80(1.80)	39.25(1.35)
Eu	3.333(343)	5.415(420)	5.840(263)	3.350(425)	4.830(310)	4.345(235)
Gd	22.100(2.333)	22.45(1.70)	24.17(1.10)	46.80(4.55)	22.00(1.30)	37.25(1.25)
Tb	3.140(317)	3.312(222)	3.683(167)	6.930(545)	3.350(220)	5.580(175)
Dy	20.13(1.80)	20.72(1.14)	22.57(76)	37.70(2.70)	20.10(1.00)	33.06(88)
Ho	3.430(337)	3.607(217)	3.803(160)	6.320(565)	3.340(190)	5.830(190)
Er	8.660(927)	9.245(565)	10.027(440)	16.090(1.210)	8.050(530)	14.450(5359
Tm	1.163(167)	1.260(130)	1.329(94)	2.045(255)	1.090(120)	1.810(107)
Yb	7.097(967)	7.420(687)	7.893(517)	12.450(1.450)	6.890(680)	11.425(615)
Lu	0.850(153)	0.964(114)	1.051(82)	1.620(190)	0.980(120)	1.635(100)
Hf	2.270(357)	5.747(480)	5.550(363)	3.335(610)	8.150(580)	3.145(230)
Ta	0.0347(267)	0.0963(333)	0.0413(163)	0.0700(470)	0.2720(580)	0.0310(135)
Pb	0.337(180)	0.315(135)	0.263(130)	0.580(325)	0.540(140)	0.500(145)
Th	0.503(77)	0.174(50)	0.095(25)	0.366(123)	0.153(38)	0.054(17)
U	0.071(40)	0.023(16)	0.013(8)	0.083(56)	bdl	0.0053(56)

.

4. Results

4.1. Clinopyroxene Crystal-Chemistry

Clinopyroxene from both host and enclaves has a diopside-augite composition with a ferrosilitic component varying from 11 to 16 (Table 2; Figure 5A). Crystals have low to absent Cr${}_2$O${}_3$ and low TiO${}_2$ (0.25–0.45 wt%). FeOtot content is quite similar in host and enclaves clinopyroxene (8.92–11.3 wt%) but a higher amount of Fe${}_2$O${}_3$ is calculated for host-clinopyroxene respect to enclaves-clinopyroxene (Table 2). Despite the similar major element composition some minor difference is represented by the Mg, Ca, and Na amount together with the Fe${}^{3+}$ (Table 2). In particular, host-cpx has, in general, higher content of Ca and Na (and Fe${}^{3+}$) and lower content of Mg (Figure 5B).

Some variability can be observed in the trace elements distribution with host-cpx enriched with respect to enclaves-cpx. Compatible elements such as V, Cr, and Co are higher in enclaves-cpx than in host-cpx, but Sc behave opposite (82.42 ppm average in enclaves-cpx and 175.72 ppm average in host-cpx). Incompatible elements such as Y, La, Ce, Sm, and Nd are higher in host-cpx than enclaves-cpx, but Sr is higher in enclaves-cpx (32 ppm average) than in host-cpx (7.0 ppm average) as reported in Table 3. Despite the slight enrichment in trace elements of host-cpx respect to enclaves-cpx the pattern distribution normalised to CI chondrites [39] is very similar with a slight enrichment of LREE respect to MREE (Figure 6). Depletion in Eu observed in some pyroxenes may be related to feldspar fractionation and/or different degree of magma evolution.

X-ray mappings do not reveal evident patterns of zoning in crystals from either host and enclaves; microchemical analyses, including raw compositional EDS and WDS lines, indicate a slight oscillatory variation in the chemical composition of clinopyroxenes from the host, whereas those from enclaves are homogeneous.

The major and trace elements distribution of clinopyroxene and bulk rocks suggest that the difference between host and enclave parental magmas could be related to different degrees of evolution (by FC) of a similar parental magma.

4.2. Clinopyroxene Structural Data

Pyroxene crystallographic structure (bond lengths and polyhedral volumes) has the ability, through chemical substitutions, to record the variations in P-T-X conditions of the melt from which they crystallize.

In the Euganean trachyte pyroxenes, the cations substitution scheme is the same in both enclaves- and host-cpx and can be described as follows:

• (Al${}^{3+}$, Si${}^{4+}$)T = (Ca${}^{2+}$, Na${}^{+}$)M2;
• (Fe${}^{3+}$, Ti${}^{4+}$, Mg${}^{2+}$)M1 = (Al${}^{3+}$, Si${}^{4+}$)T;
• (Ca${}^{2+}$)M2 = (Mg, Fe${}^{2+}$)M2.

The difference is mainly related to the extent of the substitution occurring in slightly different parental melt for the enclaves with respect to the host rocks. As described in the above paragraph, the chemical content of the enclaves- and host-cpx are quite similar, with small variation in the Ca+Na content resulting in a higher (Mg + Fe${}^{2+}$) M2 content in the host-cpx.

The tetrahedron geometry is quite similar in both enclaves- and host-cpx, having mean <T-O> bonds length in the range 1.637(2)–1.634(3) Å (Table 1), without significant variation in the host or enclaves samples pair. Despite the slight variation range in volume values for host-cpx (2.222(2)–2.233(2) Å${}^3$) and enclaves-cpx (2.224(2)–2.229(2)Å${}^3$), the influence of the nearby M2 site configuration produces a more pronounced difference on site distortion with a higher tetrahedral angle variance [41] for host- than enclaves-cpx (Table 1). Between the two groups, the larger difference occurs in M1 and M2 sites: the short and intermediate M1-O2 and M1-O1A1 bonds have similar length values in both enclaves- and host-cpx, whereas the longest M1-O1A2 bond length is shorter in enclaves- than in host-cpx; the shorter M2-O1 and M2-O2 bonds length increases from enclaves- to host-cpx, whereas the longest M2-O3C1 and M2-O3C2 bond values partially overlap without a clear relationship from enclaves- to host-cpx. As a consequence, the octahedral M1 volume values for host- and enclaves-cpx overlap in the range 11.884(5)–11.997(5) Å${}^3$ while the polyhedral M2 volumes is smaller in enclaves-cpx than in host-cpx (Table 1).

The lattice parameters also reflect the pyroxene chemical variation; in particular, the crystallographic $\left(\beta \right)$ angle decreases as Fe${}^{2+}$ enters the M1 octahedral site. The pyroxene structure thus has the capability to record the changes in parental magma chemistry by the major (and trace) elements exchange occurring in the tetrahedral and polyhedral sites, thus this mineral can be used to infer the parental magma chemical characteristics such as degree of evolution or affinity [48]. In particular, the variation of $\left(\beta \right)$ angle respect to VM2/VM1 polyhedral volumes ratio reflect the affinity of the parental magma [48], and can thus describe possible difference in the parental enclaves magmas respect to the host magmas. The host-cpx plots in the field defined by the alkaline clinopyroxene from the literature in agreement with the bulk rock composition, whereas the enclaves-cpx has values plotting towards the transitional rocks field of [48], suggesting a slightly more transitional affinity for the enclaves parental melt respect to trachytic host melt (Figure 7).

A different behaviour is observed for CEM33 sample that has the highest $\left(\beta \right)$ values among the host-cpx overlapping the enclaves-cpx $\left(\beta \right)$ values, with a chemical composition and the Fe${}^{2+}$/Mg distribution in the M2 and M1 sites more similar to the enclaves-cpx than to host-cpx (Figure 7).

5. Discussion

Euganean Hills trachytes are characterised by the presence of mafic cumulites, metamorphic xenoliths and magmatic enclaves. In particular, two types of magmatic enclaves have been found in the Euganean Hills trachytes:

• Intrusive (and crystal mushes) enclaves;
• Volcanic enclaves.

Intrusive enclaves represent the magma chamber solidification front at different degrees of crystallisation [6] while volcanic enclaves testify mixing processes occurring at magma chamber depth up to surface. Chemical composition and mineral paragenesis are very similar for trachytic rocks and intrusive/volcanic enclaves and crystal mushes [6]. Although amphibole and clinopyroxene are not always present in the Euganean magmatic rocks, they can be very useful to reconstruct the origin and evolution of the parental magma. Clinopyroxene can be used to characterise the parental magma, especially when complex processes such as magma mixing occur, and both clinopyroxene and amphibole can be useful to reconstruct the crystallisation pressure of magma.

Clinopyroxene from host and volcanic enclaves analysed in this study have major and trace elements content in the range defined by the literature dataset for clinopyroxene from Euganean trachytes [32,36,37,38]. A slight enrichment of LREE in respect to MREE and HREE can be observed in the spider diagram of Figure 6 and characterise literature trachytic pyroxene and our samples. The presence of a strong negative anomaly in Eu in some Euganean pyroxenes (host-enclaves of this study and literature data), seems related to feldspar fractionation in the parental magma [36].

The geochemical dataset by [36] is the most comprehensive and representative for the Euganean volcanics, and reports the chemical variability range for the bulk volcanic rocks and minerals (i.e., clinopyroxene) of the Euganean Hills. Host- and enclaves-cpx trace elements patterns plot inside the range defined by the dataset of [36] with host-cpx slightly enriched in trace elements with respect to enclaves-cpx (Figure 6), suggesting a crystallisation from a similar parental magma that possibly differs in the degrees of fractional crystallisation.

Pyroxene (and amphibole) from mafic and ultramafic cumulates [30] although, with a similar pattern in the spider diagram (Figure 6), has one order of magnitude lower content of trace elements than pyroxene from volcanic rocks and enclaves. Such a difference reflects the origin of the parental magma, in fact mafic and ultramafic cumulates have a parental magma derived by subduction-related metasomatized lithospheric mantle [30] while alkaline volcanic rocks and enclaves are related to intra-plate magmatic activity [2].

The Euganean Hills volcanic enclaves-host trachytes pair differs in chemical composition and crystal structure as a response to the different parental magma characteristics: the host-cpx (with a higher Al, Fe${}^{3+}$, Ti, and Ca on average) recorded a more alkaline affinity of the parental magma, whereas the enclaves-cpx records a more transitional affinity of the alkaline parental magma, in good agreement with the bulk rock chemistry. Nevertheless, the CEM33 host-enclaves cpx pair data clearly suggest that even if we can observe mixing structure in the bulk rock the clinopyroxene in both host and enclaves portions are very similar and have similar characteristics (Figure 7). Not only can the strong sensitivity of clinopyroxene chemistry be thus very helpful to identify possible migration of clinopyroxene during the mixing processes from one end member to the other [13,49,50,51,52,53,54,55], but the crystal structure can also be a powerful tool to detect the crystal mobility and parental magma origin.

Euganean Hills clinopyroxenes have been studied by [37] with a similar approach as that reported in this study. Trachytes from Mt Rosso, Zovon quarry, Roccapendice, Mt Gemola, Mt Lozzo, and Montemerlo quarry were studied by the authors by picking up clinopyroxene crystals from rock granular powder lacking the petrographic textural data on the possible occurrence of mixing textures (Carbonin personal com.). As a consequence of the blind sampling, we can observe a distribution of the clinopyroxene crystallographic parameters values in two groups that almost overlap with our enclaves-host cpx crystal chemical values (Figure 7, Table 1 and Table 2 (and Tables 1 and 2 in [37])). As clinopyroxene crystals of [37] were selected from a rock granular powder, we could consider both that enclaves and host rocks portions were mixed in the granular material and a random selection of enclaves- and host-rocks were picked up for the study; or, if the authors selected optically homogeneous rocks, that they found crystals of the enclaves that were migrated into the host. Aside from which is the applicable hypothesis, the crystal structure geometry proved to be a powerful tool to support and complete the clinopyroxene major and trace elements chemistry analysis to unravel complexity in the magmatic system.

5.1. Geobarometric Considerations

5.1.1. Geobarometry Based on the Structural Geometry of the Analyzed Clinopyroxenes

Clinopyroxene single-crystal geobarometer is based on the property that pyroxene crystallographic and crystal–chemical signature is function of P-T-X variables, particularly sensitive to crystallization pressure (e.g., [7,11,56,57,58]). In fact, irrespective to magma bulk composition, a linear relationship between unit cell and octahedral M1 volumes with pressure of crystallization has been observed: both volumes increase decreasing pressure [42,48,56,59].

The Vcell vs VM1 relationship is related to the cation substitutions affecting the structure, (Al${}^{3+}$, Mg${}^{2+}$)M1 = (Ca${}^{2+}$, Na${}^{+}$)M2, (Al${}^{3+}$, Mg${}^{2+}$)M1 = (Al${}^{3+}$, Si${}^{4+}$)T, and (Ca${}^{2+}$)M2 = (Mg, Fe${}^{2+}$)M2, which at a given P is related to the crystallization conditions (aCaO, aSiO${}_2$). Thus, the Vcell vs VM1 relationship can record even small variations of the crystallisation pressure, as observed in core-rim pairs [60]. The cell volumes are smaller in enclaves-cpx (438.4(5)–439.4(4) Å${}^3$) than in host-cpx (439.7(3)–440.2(4) Å${}^3$); similarly, polyhedral M2 volumes are smaller in enclaves-cpx (25.469(6)–25.573(7) Å${}^3$) than in host-cpx (25.596(8)–25.746(7) Å${}^3$), octahedral M1 volumes instead cover a similar values range (11.871(4)–11.963(4) Å${}^3$ in enclaves-cpx and 11.872(4)–11.997(5) Å${}^3$ in host-cpx; Table 1). The observed geometrical variation results by the entering of Al and Fe${}^{3+}$ for Mg or Fe${}^{2+}$ in M1, whereas the entering of Na for Ca increases the volume of M2.

The enclaves- and host-cpx having similar M1 volumes but distinct unit cell volumes plot in two clusters in Figure 8A, suggesting a similar crystallisation pressure range. The difference in the cell volume reflects the crystallisation in a parental magma with slightly different composition. In fact, as discussed by several authors [48,59,61], the activity of SiO${}_2$, Al${}_2$O${}_3$, and Na${}_2$O components of a magma strongly influence the chemistry (and structure) of clinopyroxene in a way that decreases the silica activity of the parental magma; the VM1 decreases more than cell volume, producing a shifting in the VM1 vs. Vcell plot. We can observe a shifting of in the VM1 vs. Vcell plot from enclaves-cpx (with smaller cell volumes) to host-cpx (with larger cell volume). Data by [37] overlap our data in Figure 8A, suggesting a similar crystallisation pressure. Moreover, a similar distribution pattern is observed, suggesting that [37] could have sampled enclaves-host pairs in the Euganean Hills area.

The cell and octahedral M1 volumes relationship, as discussed in [48], can thus be used to constrain the crystallization pressure of pyroxenes and the magma storage conditions.

The Vcell-VM1 space is defined for the high-pressure field by clinopyroxene from spinel lherzolites (spinel-lherzolite stability field P = 20–10 kbars [48], also confirmed this by clinopyroxene from pyroxenite mantle nodules enclosed in alkaline lava from Oahu (Hawaii) whose crystallization pressure was estimated to be 17 kbars [60] and for the volcanic field by clinopyroxene phenocrysts of magmatic rocks (see [15] and reference therein)).

Euganean Hills host-enclaves clinopyroxenes plot in the Vcell-VM1 space defined for the volcanic pyroxenes (Figure 8A) in particular distribute around a value of 3 kbars defined by literature data including pyroxene from alkaline lavas [10,15,60], thus an estimation of a crystallisation pressure range between 2 and 4 kbars is suggested.

5.1.2. Geobarometry Based on the Clinopyroxene Geochemistry

Single crystal temperature and pressure estimators calibrated by [58] have been used to provide P and T constraints on the basis of clinopyroxene geochemistry (

Table 4. Estimated temperature and pressure for the studied clinopyroxenes on the basis of geochemistry.

Sample	T (°C)	P (kbar)
CEM6H	904	2.9
CEM6E	1024	3.4
CEM15H	892	2.8
CEM15E	932	2.7
CEM18H	865	2.4
CEM18E	964	3.0
CEM33H	878	2.4
CEM33E	1013	3.3

). We obtained an estimation of temperature between 865 °C and 1024 °C and pressure between 2.4 and 3.4 kbars (Figure 8B). As a general remark we can observe that even if enclaves-cpx has slightly higher P-T values than host-cpx, the differences are below the estimated error (T: 60 °C; P: 2 kbars).

5.1.3. Geobarometry Based on the Amphibole Geochemistry

Temperature and pressure of amphibole-bearing magmatic intrusive enclaves was estimated [6] using the geothermometer calibration of [62] and the Al-in-amphibole geobarometer (calibration of [63]). Using the estimated temperature of 800–850 °C for the crystallisation of the amphibole-bearing enclaves, the crystallisation pressures for the amphibole was estimated in the range 4.5–4.8 kbars and 2.9–3.3 kbars for intrusive enclaves and in the range 4.2–4.9 kbars and 2.6–3.3 kbars for those enclaves inferred to reflect crystal mushes processes.

6. Conclusions

Euganean Hills trachytes enclose two types of magmatic enclaves: (a) intrusive (and crystal mushes) enclaves representing the crystallisation front of the melt in the magma chamber and (b) volcanic enclaves testifying the petrological processes occurred in the magma chamber up to the surface. Clinopyroxene from volcanic enclaves and amphibole from intrusive enclaves and crystal mushes have been used to reconstruct the architecture of the Euganean Hills.

• Clinopyroxene major and trace elements composition of host and volcanic enclaves pairs suggest similar parental magmas differing in small degrees of fractional crystallisation;
• Single-crystal clinopyroxene geobarometer based on crystal structure and geochemistry and amphibole geobarometer agree in the estimation of a pressure of crystallisation between 4.8–2.0 kbars;
• A complex system of magma chambers at intermediate to shallow crustal level can be inferred by using clinopyroxene and amphibole phenocrysts as toolkits. Here, mafic magmas accumulated, evolved by fractionation processes, and mixing occurred before eruption (Figure 9).