j ourna l homepage: www.e lsev ie r .com/ locate /chemgeo
Clinopyroxene and titanomagnetite cation redistributions at Mt. Etnavolcano (Sicily, Italy): Footprints of the final solidification history oflava fountains and lava flows
S. Mollo a,⁎, P.P. Giacomoni b, D. Andronico a, P. Scarlato a
a Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, 00143 Roma, Italyb Departimento di Fisica e Science della Terra, Università di Ferrara, Via Saragat 1, 44122 Ferrara, Italy
Article history:Received 12 February 2015Received in revised form 16 April 2015Accepted 23 April 2015Available online 1 May 2015
Editor: D.B. Dingwell
Keywords:Mt. EtnaLava fountains and lava flowsDegassing and cooling
For a better understanding of the final solidification history of eruptions at Mt. Etna volcano (Sicily, Italy), wehave investigated cation redistributions at the interface between sub-millimetre-sized clinopyroxene andtitanomagnetite crystal rims and coexisting melts. The studied products were scoria clasts from lava fountainsand rock samples from pahoehoe and aa lava flows. Our data indicate that scoria clasts from lava fountainingwere rapidly quenched at the contact with the atmosphere, preserving the original crystal textures and compo-sitions inherited during magma dynamics within the plumbing system. Kinetics and energetics of crystallizationwere instantaneously frozen-in and post-eruptive effects on mineral chemistry were negligible. The near-equilibrium compositions of clinopyroxene and titanomagnetite indicate that lava fountain episodes were sup-plied by high-temperature, H2O-rich magmas ascending with velocities of 0.01–0.31 m/s. In contrast, magmasfeeding lava flow eruptions underwent a more complex solidification history where the final stage of the crystalgrowth was mostly influenced by volatile loss and heat dissipation at syn- and post-eruptive conditions. Due tokinetic effects associated with magma undercooling, clinopyroxenes and titanomagnetites formed by crystal at-tachment and agglomeration mechanisms leading to intricate intergrowth textures. The final compositions ofthese minerals testify to closure temperatures and melt–water concentrations remarkably lower than those es-timated for lava fountains. Kinetically-controlled cation redistributions at the crystal–melt interface suggest thatthe solidification of magma was driven by degassing and cooling processes proceeding from the uppermost partof the volcanic conduit to the surface.
It is widely assumed that changes in the mineral chemistry may po-tentially record the crystallization conditions of magmas during ascent,degassing and cooling (Putirka, 2008). In this framework, Mt. Etna vol-cano (Sicily, Italy) provides a variety of crystallization scenarios wheretextures and compositions of phenocrysts, microphenocrysts, andmicrolites are indicative of processes that initiate atmantle depths, con-tinue in the conduit of the volcano and terminate at subaerial conditions(Ferlito et al., 2008; Viccaro et al., 2010; Ferlito et al., 2011; Kahl et al.,2011; Ferlito and Lanzafame, 2010; Armienti et al., 2013; Corsaroet al., 2013; Lanzafame et al., 2013). The importance of minerals to re-construct magma dynamics is emphasized by the historic and recenteruptions at Mt. Etna characterized by uniform phase assemblages(olivine + clinopyroxene + plagioclase + titanomagnetite) and bulk
); fax: +39 0651860507.
rocks permanently buffered to the compositions of trachybasalts andbasaltic trachyandesites (Armienti et al., 2007; Corsaro et al., 2009). Toprovide magmatic constraints on the early-stage of crystallization, thecompositions of phenocryst cores and bulk rocks are frequently usedas input data for barometers, thermometers, hygrometers and oxygenbarometers (Armienti et al., 2007, 2013; Lanzafame et al., 2013; Ferlitoet al., 2014; Giacomoni et al., 2014; Mollo et al., 2015). However, theseestimates reflect the conditions at which minerals saturate the melt atlow degrees of crystallization (Putirka et al., 2003), and when theerupted products represent near-liquidus compositions (Frey andLange, 2011). Conversely, late-stage solidification conditions are record-ed by the more evolved chemistries of crystal rims and residual glassesthat are frozen-in at the final resting temperature of the system (Coishand Taylor, 1979; Hammond and Taylor, 1982; Ujike, 1982; Venezkyand Rutherford, 1997; Zhou et al., 2000; Hammer, 2008; Baginskiet al., 2009;Mollo et al., 2011, 2013a). This consideration places empha-sis on the fact that volcanic products representative of different eruptivestyles at Mt. Etna may experience a variety of cooling historiescomprised between the liquidus temperature ofmagma and the closure
46 S. Mollo et al. / Chemical Geology 406 (2015) 45–54
temperature of crystal growth (Wall et al., 2014). For example,mineralsin scoria clasts from lava fountains are rapidly quenched duringeruption and their compositions reflect magma dynamics withinthe plumbing system (Andronico and Corsaro, 2011). Similarly,millimetre-to-centimetre-sized phenocrysts in lava flows result fromcrystallization at depth recording pressures and temperatures alongthe different segments of the feeding system (Armienti et al., 2007;Corsaro et al., 2007, 2009; Kahl et al., 2011; Mollo et al., 2015). Incontrast, sub-millimetre-sized phenocrysts show textures and com-positions indicative of crystallization upon degassing and cooling inthe uppermost part of the conduit and at subaerial conditions(Lanzafame et al., 2013). Indeed, during flowage onto the surface,the nucleation and growth of crystals are controlled by time-dependent pathways that influence significantly the rheology andmorphology of both aa and pahoehoe lavas (Kilburn and Guest,1993; Pinkerton and Norton, 1995).
Based on these premises, the final solidification conditions of Etneaneruptions have been investigated in this study through chemical analy-ses conducted at the interface between crystal rims and coexistingglasses. In particular, cation redistributions were measured in sub-millimetre-sized phenocrysts and microphenocrysts of clinopyroxeneand titanomagnetite from lava fountains and lava flows of bothpahoehoe and aa types. These minerals were selected in view of theirclose textural relationships (i.e., nucleation of titanomagnetite onclinopyroxene, inclusion of titanomagnetite in clinopyroxene, andtitanomagnetite and clinopyroxene intergrowth) documented by bothnatural (Corsaro et al., 2007; Mollo et al., 2011, 2013b) and experimen-tal (Del Gaudio et al., 2010;Mollo et al., 2013a, 2015) studies conductedon Etnean magmas. Our data highlight some important textural andcompositional changes among the studied products testifying to thedifferent final solidification paths of magmas.
2. Sampling and methods
Samples object of this study come from 4 historic pahoehoe lavasflowed at the northeast rift of Mt. Etna (Chester et al., 1985), 4 aalavas emplaced during 2002 flank eruption (Andronico et al., 2005;Ferlito et al., 2009) and 4 scoria clasts from the episodic sequence oflava fountains started from January 2011 at the New South-East Crater(Bonaccorso et al., 2013; Behncke et al., 2014). In particular, the historiclavas are related to the 1763 and 1809 eruptions, the 2002 sampleswere collected on lavas emplaced between October 27 and November3 at the northern base of the NE Crater, while the lava fountains arefrom the episodes which occurred on 12/1/11, 19/7/11, 15/11/11,and 5/1/12. These products show an identical phase assemblage of phe-nocrysts,microphenocrysts andmicrolites that have beendiscriminatedon the basis of the longest size dimensions N0.3, 0.3–0.1, and b0.1 mm,respectively (Armienti et al., 1984). The porphyritic index rangesbetween 13 and 40%, and comprises 45–55 vol.% of plagioclase,35–45 vol.% of clinopyroxene, 5–8 vol.% of olivine, and 1–5 vol.% oftitanomagnetite.
Whole-rock major element analyses were carried out at theDepartment of Physics and Earth Sciences of the University of Ferrara(Italy) using X-ray fluorescence (Thermo ARL Advant XP). Intensitieshave been corrected for matrix effects using the method of Lachanceand Trail (1966). Loss on ignition (L.O.I.) has been determined by gravi-metric method assuming Fe2O3 as 15% FeO.
Microprobe and textural analyseswere conducted at the HP-HT Lab-oratory of Experimental Volcanology and Geophysics of the IstitutoNazionale di Geofisica e Vulcanologia (INGV) in Rome, Italy. Chemicalanalyses were carried out on a Jeol-JXA8200 electronic probe microanalysis (EPMA) equipped with five spectrometers. The data werecollected at the crystal–melt interface by analysing the rims of sub-millimetre-sized phenocrysts and microphenocrysts of clinopyroxeneand titanomagnetite, and coexisting glasses a few microns next to thecrystal edges. These compositions are reported in the Excel spreadsheet
submitted online as supplementary material. The analyses wereperformed using an accelerating voltage of 15 kV and a beam currentof 10 nA. For crystals, the beam size was 1 μm with a counting time of20 and 10 s on peaks andbackground, respectively. For glasses, a slightlydefocused electron beam with a size of 3 μm was used with a countingtime of 5 s on background and 15 s on peak. The following standardshave been adopted for the various chemical elements: jadeite (Si andNa), corundum (Al), forsterite (Mg), andradite (Fe), rutile (Ti), ortho-clase (K), barite (Ba), apatite (P), spessartine (Mn) and chromite (Cr).Sodium and potassium were analysed first to prevent alkali migrationeffects. The precision of the microprobe was measured through theanalysis of well-characterized synthetic oxides andminerals. Data qual-ity was ensured by analysing these test materials as unknowns accord-ing to Iezzi et al. (2014). Based on counting statistics, analyticaluncertainties relative to their reported concentrations indicate thatthe accuracy was better than 5%. Images were collected using thebackscattered electron (BSE) mode of a field emission gun-scanningelectron microscopy (FE-SEM) Jeol 6500F equipped with an energy-dispersive spectrometer (EDS) detector. Quantification of EDS datawas done with the same set of standards used for microprobe analyses.Data were processed through the FE-SEM software package in order tocalculate pure element intensities from standards andmatrix correctioncoefficients for samples with unknown compositions. Measured con-centrations ofmajor oxides are reasonably consistentwith the expected(theoretical) values and the discrepancy is generally less than 1 wt.%.BSE imaging combined with X-ray mapping were both adopted for theidentification of the main textural and compositional features of eachsample. X-ray maps were collected by using 15 kV accelerating voltage,8 nA probe current, resolution 1024 × 768 pixel2, and dwell time 10 ms(real time) per pixel. To the single-band images, representing X-raymaps for FeO, MgO and CaO, were assigned different colours (red,green and blue) and combined to form a coloured three-band image.Phases containing a combination of more than one of the elementsbeing mapped are displayed with composite colours.
3. Results
3.1. Textures
Scoria clasts from lava fountains contain phenocrysts and micro-phenocrysts (35–55 vol.%) of clinopyroxene and titanomagnetite em-bedded in abundant glass (30–40 vol.%). The microlite content isrelatively low (15–25 vol.%) suggesting that the late crystallizationwas hampered by rapid quenching of magma (Fig. 1a–d). The habitof large clinopyroxene phenocrysts is prismatic with well-formedplanar edges frequently enclosing sub-rounded titanomagnetitemicrophenocrysts (Fig. 1a) and microlites (Fig. 1b). Titanomagnetitemicrophenocrysts preferentially develop on pre-existing clinopyroxenes(Fig. 1b–c), whereas titanomagnetite phenocrysts grow as isolatedcrystals (Fig. 1d).
Lava flows are characterized by variable phenocryst and micro-phenocryst contents (30–40 vol.%), abundant microlites (N40 vol.%)and scarce residual glass (b25 vol.%). Generally, larger clinopyroxenesand titanomagnetites form through the aggregation and coalescenceof smaller and mutually touching crystals (Fig. 1e–f). This agglomera-tion growth process operates from the micrometre to millimetre scaleleading to the formation of dense mosaics of aggregated phenocrystsand microphenocrysts (i.e., glomerocrysts of clinopyroxene andtitanomagnetite). Sometimes, the growth of a single clinopyroxenephenocryst is interrupted by the contiguous agglomeration oftitanomagnetitemicrophenocrysts andmicrolites (Fig. 1g). Alternative-ly, sub-rounded titanomagnetites show incomplete textures due to therapid crystal growth of clinopyroxene (Fig. 1h). The texturalmaturationof titanomagnetite clearly indicates that the final euhedral habit resultsfrom the coalescence of a great number of microlites growing as vermi-form blebs (Fig. 1i–n).
Fig. 1. Representative textural features of our samples showing single crystal growth in lava fountains (a–d), clinopyroxene and titanomagnetite aggregation and agglomeration inpahoehoe and aa lavas (e–h), and titanomagnetite textural maturation due to coalescence of vermiform blebs (i–n).
47S. Mollo et al. / Chemical Geology 406 (2015) 45–54
3.2. Compositions
Whole-rock analyses plotted on the TAS (total alkali vs. silica) dia-gramshow that our samples can be classified as trachybasalts and basal-tic trachyandesites with SiO2 concentrations constrained in a narrowrange between 48 and 52 wt.% (Fig. 2a). In terms of magma differentia-tion, these products are characterized by a comparable degree of evolu-tion with bulk Mg# values [Mg# = XMg / (XMg + XFe2+) × 100] of47–48, 44–50, and 45–49 for lava fountains, pahoehoe and aa lavas, re-spectively. Conversely, glasses analysed at the interface with the crystalrims show substantial compositional differences. For the case of lavaflows, thematrix glass is highly heterogeneous in terms of major oxides
(e.g., MgO varies from 0.2 to 2.8 wt.%; Fig. 2b) with local chemicalenrichments or depletions responding to post-eruptive kinetic effectsassociated to extensive groundmass crystallization (e.g., Mollo et al.,2013c,d). In contrast, the glass from lava fountains exhibits more prim-itive and homogeneous compositions indicative of instantaneousquenching ofmagma at the contactwith the atmosphere (Fig. 2b).Mov-ing from lava fountains to pahoehoe to aa lavas, clinopyroxene pheno-crysts and microphenocrysts are progressively enriched in Al2O3 anddepleted in CaO+MgO (Fig. 2c), despite the fact that the compositionsremain diopsidic [XCa / (XCa + XMg + XFe2+) N 45; Morimoto et al.,1988]. Similarly, titanomagnetites show decreasing TiO2 contentscounterbalanced by increasing Al2O3 + MgO concentrations (Fig. 2d).
Fig. 2. Chemical data showing whole-rock analyses plotted into the TAS (total alkali vs.silica) diagram (a), TiO2 andMgO concentrations of themelts, (b) compositional variationof clinopyroxene (c) and titanomagnetite (d).
48 S. Mollo et al. / Chemical Geology 406 (2015) 45–54
4. Discussion
4.1. Clinopyroxene and titanomagnetite intergrowth
Both the petrography of natural products (Métrich et al., 2004;Corsaro et al., 2013) and the phase assemblage of decompression exper-iments (Metrich and Rutherford, 1998) suggest that, at Mt. Etna volca-no, titanomagnetite forms at the end of the crystallization sequencemostly driven by volatile exsolution. This has been also confirmed byMELTS simulations conducted on primitive (Mg# = 50) and moreevolved (Mg# = 63) Etnean trachybasaltic compositions equilibratedat 100–400 MPa, 1150–1050 °C, and with 1–2 wt.% H2O (Mollo et al.,2015). The textural relationship between titanomagnetite andclinopyroxene reflects the early nucleation of clinopyroxene and thesubsequent incorporation of titanomagnetite (Fig. 1). The entrapmentof titanomagnetite can be explained as the result of the rapid growthof clinopyroxene (10−7 cm/s; Baker, 2008; Mollo et al., 2013a) that isup to three orders of magnitude faster than that of titanomagnetite(10−10 cm/s; Cashman and Marsh, 1988). While clinopyroxenesand titanomagnetites from lava fountains grow as isolated crystals,the textural maturation of sub-millimetre-sized phenocrysts andmicrophenocrysts from pahoehoe and aa lava flows is rather different.X-ray maps conducted on lava samples show that chemically homoge-neous titanomagnetite phenocrysts grow by agglomeration of smallercrystals with virtually identical compositions (Fig. 3a). The chemistryof agglomerated titanomagnetites does not differ from that of singlecrystals hosted in clinopyroxene phenocrysts (Fig. 3b), thus excludingthe crystallization of titanomagnetite on the liquidus. Clinopyroxeneedges are frequently embedded in titanomagnetite phenocrysts provid-ing that clinopyroxene aggregation and titanomagnetite agglomerationare both coeval crystal growth processes (Fig. 3b–d). Crystal inter-growth is generally attained when clinopyroxene microphenocrystsattaching on themselves to form a single phenocryst are entrapped dur-ing the process of titanomagnetite agglomeration (Fig. 3b–d). Again, no
chemical differences are observed between small clinopyroxenes in-corporated into titanomagnetite and large clinopyroxene phenocrysts.This points out that titanomagnetite and clinopyroxene form througha cooperative mechanism of nucleation and growth, where the agglom-eration rate of titanomagnetite blebs competes with the attachmentrate of clinopyroxene crystals, and vice versa. Crystal growth mecha-nisms by attachment and agglomeration have been documented byex-situ (Pupier et al., 2008) and in-situ (Schiavi et al., 2009) laboratoryobservations conducted on basaltic liquids. Additionally, it has been ex-perimentally demonstrated that crystal agglomeration can lead to theformation of large chemically homogeneous crystals (Iezzi et al., 2011,2014) in response to the emerging theory of aggregation by self-orientation of sub-micrometric early-formed primary crystals (Teng,2013 and references therein). One interesting result from X-ray map-ping is that clinopyroxene sub-millimetre-sized phenocrysts resultfrom the aggregation of microphenocrysts with identical compositions(Fig. 3e–f) and/or occasionally by the growth of Mg-poor crystals onpre-existing Mg-rich ones (Fig. 3g–i). The attachment proceeds alongthe longest crystal axis and the face of attachment is correlated withthe shape of the crystals so as to minimize the interfacial energy andstructural mismatches (Kostov and Kostov, 1999; Deer et al., 2001).Mg-poor crystals overgrowth around Mg-rich clinopyroxenes andapparently seem to reproduce the zoned textures documented bymagma mixing processes at Mt. Etna (Armienti et al., 2007; Corsaroet al., 2013; Mollo et al., 2015). However, Mg-rich and Mg-poor crystaldomains are separated by sharp interfaces where Mg-rich crystals areattached only to one side (Fig. 3g–i). This asymmetric crystal zoningclearly excludes recharge andmagmamixing, otherwise clinopyroxenephenocrysts should exhibit axially symmetric core-to-rim chemical var-iations. A similar feature has been attributed to successive nucleationevents in cooled magmas that become progressively more differentiat-ed (Mollo et al., 2012; Iezzi et al., 2014); in this view, after the early for-mation of primitive minerals, more evolved crystals heterogeneouslynucleate onto the pre-existing substrates with ongoing differentiationof the feeding melt. Both textural and compositional analyses of somepahoehoe lavas from pre-historic and historic eruptions at Mt. Etna in-dicate that Mg-rich clinopyroxenes can originate at depth from high-temperature, H2O-rich magmas, whereas most of the heterogeneousnucleation and aggregation of Mg-poor microphenocrysts may occurat syn-eruptive conditions over the effect of undercooling (Lanzafameet al., 2013). In the case of Hawaiian lava flows, massive crystallization(up to ~60 vol.%) ofmicrophenocrysts andmicrolites has been observedduringflowage onto the surface (Crisp and Baloga, 1990; Cashmanet al.,1999; Soule et al., 2004), together with the formation of clinopyroxenephenocrysts with maximum length of 0.5 mm (Crisp et al., 1994). Thisconsideration agrees with results from fractional crystallization model-lingwhere the segregation of 10–20 vol.% of phenocrysts at depth is themaximum admissible value to explain the persistent eruption of Etneanmagmas buffered to trachybasaltic compositions (Corsaro et al., 2013;Mollo et al., 2015; Vetere et al., 2005. Abundant crystallization atdepthwould also imply the eruption of high yield strength lavas, in con-trast with the flow surface morphology of pahoehoe types (Hon et al.,2003). If the attachment process involves microphenocrysts withcomparable compositions, clinopyroxenes show indistinct crystalcontacts and appear in perfect continuity as one single phenocryst(Fig. 3e–f). The entrapment at the crystal boundary of wedged inte-rstitial melts (moderately to highly crystallized upon lava cooling) re-veals the incomplete aggregation of clinopyroxenes by their faces(Fig. 3e–f). A similar attachment featurewas observed for the formationof large clinopyroxene phenocrysts due to volatile exsolution anddegassing of a fluid-saturated magma rising from depth (Frey andLange, 2011). Moreover, crystals in pahoehoe lavas are generallysurrounded by an almost continuous film of glass, whereas the ground-mass of aa lavas shows micrometre-sized glass pockets isolated bydense microlite crystallization and intricate intergrowth patterns. Thisdifferent glass contentfinds an interesting analogue in the experimental
Fig. 3.X-raymapsof lava samples showing crystal growthby aggregation and coalescence of clinopyroxene and titanomagnetite crystalswith virtually identical compositions (a–f), aswellas the crystallization of Mg-poor clinopyroxenes onto early-formed Mg-rich crystals (g–i).
49S. Mollo et al. / Chemical Geology 406 (2015) 45–54
results of Sato (1995) which showed that undercooling due to H2O ex-solution leads to the formation of aa lavas with a crystal density higherthan that of pahoehoe types. At Mt. Etna, the amount of water in equi-librium with clinopyroxene decreases from about 4 to 1.5 wt.% as pres-sure decreases from 400 to 0.1 MPa (Collins et al., 2013; Mollo et al.,2015). Thus, volatile-richmagmas at pressures b 100MPamay undergostrong degassing while travelling from the conduit to the surface(Métrich et al., 2004; Spilliaert et al., 2006; Collins et al., 2013). This isalso corroborated by P–T arrays indicative of an upward accelerationof magma due to abundant loss of water in the shallower parts of theplumbing system (Armienti et al., 2013 and references therein).
4.2. The cooling history of lava fountains and lava flows
From lava fountains to pahoehoe to aa lavas, TiO2 in titanomagnetiteprogressively decreases counterbalanced by increasing MgO and Al2O3
contents (Fig. 2d). This compositional variation was observed for rapid-ly cooled crystals from core-to-rim in MORB pillow lavas at the Juan deFuca Ridge (Zhou et al., 2000) and from the outermost to the innermostportions of trachybasaltic dikes at Mt. Etna volcano (Mollo et al., 2011).Ti–Al–Mg cation substitutions can be addressed to the incorporation ofelements incompatible with the crystal lattice upon the effect of
dynamic crystallization conditions (Zhou et al., 2000). When the meltsupplies Ti and Fe to the growing crystals at equilibrium proportions,the TiO2/FeOtot ratio of Etnean titanomagnetites is comprised between0.18 and 0.25 (cf. Mollo et al., 2013c). However, over the effect of dis-equilibrium growth conditions, Ti is less favourably incorporated intothe crystal lattice and the TiO2/FeOtot ratio progressively decreases(Fig. 4a). On this basis, Mollo et al. (2013c) have calibrated ageospeedometer to estimate the cooling rate (CR) as follows:
Using Eq. (1), we found that crystals from lava fountains yieldrelatively low cooling rates ranging between 0.001 and 0.01 °C/s(Fig. 4a). In contrast, lava flows are characterized by low to high coolingrates up to 0.06 °C/s and 0.11 °C/s for pahoehoe and aa lavas, respective-ly. As reported in literature, thin-to-thick lava crusts can provideminimum-to-maximum insulation leading to highly variable coolingrates in the order of 10−1–10−5 °C/s (Crisp and Baloga, 1990; Harrisand Rowland, 2009). Perhaps the fast cooling rates measured for aalavas are related to high degrees of undercooling due to abundant vola-tile degassing; indeed, this causes strong crystal nucleation favouring aamorphologies (Sato, 1995; Lanzafame et al., 2013). In general, our
Fig. 4. Cooling rate (CR) predicted by Eq. (1) vs. TiO2/FeOtot ratios of titanomagnetite (a).Comparison of ascent velocities (AV) calculated for lava fountains (b), aa (c) andpahoehoe(d) lavas using both Eqs. (2) and (3).
50 S. Mollo et al. / Chemical Geology 406 (2015) 45–54
estimates are comparable to those determined by means of relaxationgeospeedometry applied to lava flows from a number of volcanic com-plexes, e.g., Tenerife, Kenya and Hawaii (Wilding et al., 1995; Cashmanet al., 1999; Gottsmann et al., 2004). This good correspondence suggeststhat, at the final stage of crystallization, titanomagnetite crystals did nothave enough time to re-equilibrate with the melt, faithfully recordingthe cooling history of the eruption. However, the last portion of theEtnean feeding system is an open-conduit persistently filled withmagma experiencing continuous loss of gas supplied by a steady streamof H2O-rich fluids from depth (Ferlito et al., 2011). Therefore, the crys-tallization process is potentially controlled by the efficiency ofdegassing that, in turn, depends on the velocity of magma decompres-sion (Ferlito and Lanzafame, 2010). During magma ascent, the degreeof undercooling slightly decreases as water is exsolved in the melt andthen rapidly increases accompanied by abundant water exsolution.Since magma may undergo both decompression and cooling duringascent, the fingerprint of one process or another can make difficult toestimate the ascent velocity using geospeedometers. To verify if cationredistributions in titanomagnetite crystals can be related to waterloss during decompression, we have recalculated cooling rates fromEq. (1) as decompression rates and then as ascent velocities throughtwodifferent approaches.Médard andGrove (2008) provided a polyno-mial function describing the effect of H2O on the liquidus temperaturedecrease of basaltic melts (ΔT):
ΔT �Cð Þ ¼ 40:4 �H2O–2:9 �H2O2 þ 0:77 �H2O
3 wt:%ð Þ: ð2Þ
By means of Eq. (2), Applegarth et al. (2013) found that thedegassing-induced undercooling of Etnean magmas can be estimated asa function of water loss. Considering that the pressure of starting waterdegassing is ~400 MPa and magmas undergo a water loss of ~4 wt.%while travelling upward under pure adiabatic conditions (Armientiet al., 2013; Mollo et al., 2015), Eq. (2) indicates that the melt is cooled
of ~164 °C upon the effect of decompression. On the other hand, kineticexperiments of Shea andHammer (2013)weredesigned so that the effectofmagmaundercoolingwas equalwhether in the direction of cooling anddecompression. The authors found that the solidification behaviour ofmagma was very similar texturally and compositionally. The only excep-tionwas represented by the appearance of amphibole at highwater pres-sures that, however, it is rare at Mt. Etna and does not crystallize atshallow levels due to water exsolution (Corsaro et al., 2007, 2013). Theregression analysis of the data of Shea and Hammer (2013) indicatesthat, for the same degree of undercooling, the decompression rate (DR)is well correlated (R2 = 0.96) to the cooling rate by the expression:
DR MPa=sð Þ ¼ 1:1954 � CR °C=sð Þ þ 9 � 10−5: ð3Þ
Using both Eqs. (2) and (3), we have estimated the ascent velocitiesofmagmas feeding lava fountains (Fig. 4b), pahoehoe lavas (Fig. 4c) andaa lavas (Fig. 4d), obtaining values that are almost comparable. The as-cent velocities of lava fountains (0.01–0.31 m/s) closely match withthose measured for eruptions at Mt. Etna (0.04–0.4 m/s; Aloisi et al.,2006), Hawaii (0.01–0.04 m/s; Rutherford, 2008; Gonnermann andManga, 2013), Unzen (0.01–0.07 m/s; Toramaru et al., 2008) andMount St. Helens (0.01–0.15m/s; Rutherford and Hill, 1993). Converse-ly, the ascent velocities of pahoehoe and aa lavas increase up to 2 m/s,finding similar estimates only for mantle-derived magmas erupted atthe Fogo volcano (0.7–6.7m/s; Hildner et al., 2011) and Pali-Aike volca-nic field (1–9 m/s; Demouchy et al., 2006). However, these fast ascentvelocities cannot be addressed to the upward migration of Etneanmagmas fromdepth. First, our estimateswere derived using the compo-sitions of titanomagnetite rims that formed at the end of the solidifica-tion history when the crystallinity of magma was higher than 50 vol.%.Second, Armienti et al. (2013) demonstrated that the P–T paths ofdeep-seated (N6 km) magmas at Mt. Etna are directly correlated tovery slow ascent velocities (0.4–4°10−3 m/s). According to the authors,if volatiles are released to a vapour phase retained in themagmatic sus-pension, the buoyancy of the whole magma increases significantly pro-viding explanation for the fast ascent velocities in the uppermost parts(1–4 km) of the plumbing system. In fact, fluid mechanic processesgoverning magma ascent indicate that the volume expansion ofmagma by the growth of gas bubbles is balanced by an increasing accel-eration towards the surface (Gonnermann and Manga, 2013). On thisbasis, we infer that kinetics and energetics of the crystal growth werealmost instantaneously frozen-in at the time of lava fountaining.Centimetre-sized scoria clasts were quenched in a few ten of secondsand this time interval was too short for the onset of post-eruptive crys-tallization (cf. Hort and Gardner, 2000; Giordano and Dingwell, 2003).Consequently, titanomagnetite compositions were mostly controlledbymagmadynamicswithin the plumbing systemof the volcano. In con-trast, magmas feeding lava flows underwent more complex crystalliza-tion paths that started at crustal depths, proceeded along the volcanicconduit, and terminated at subaerial conditions. As a direct result, thefinal cation redistributions measured at the crystal–melt interfacewere driven by both volatile exsolution during magma ascent andheat loss during lava flowage. Thus, the effect of cooling at the finalstage of the crystal growth succeeded to the crystallization kineticsdue to early decompression. It is also not excluded that post-eruptivedegassing could have influenced the final compositions of minerals inthe pahoehoe flow units (Mollo et al., 2013b). This is also confirmedby H2O–CO2–S–Cl solubility data suggesting that ascending magmasmay retain variable water contents that are liberated at subaerial condi-tions favouring crystallization (Lanzafame et al., 2013). Clues on the dif-ferent crystallization conditions of titanomagnetites from lava fountainsand lava flows are provided by the redox state of the system. We haveestimated the oxygen fugacity of the magmatic suspension throughthe spinel-based oxygen barometer of Ishibashi (2013) derived fromthe formalism of Ariskin and Nikolaev (1996) for the Fe2+–Fe3+
partitioning between spinel and mafic melt, and the model of Sack
51S. Mollo et al. / Chemical Geology 406 (2015) 45–54
et al. (1980) relating the Fe3+/Fe2+ ratio and oxygen fugacity in silicatemelt. Results indicate that lava fountains equilibrated at buffering con-ditions (i.e., NNO + 1) close to those estimated for the crystallizationof Etnean magmas at depth (Armienti et al., 1994, 2004, 2013). Con-versely, pahoehoe and aa lavas are more oxidized (i.e., NNO + 2–NNO+ 3) showing values found for degassed lava flows during prefer-ential hydrogen loss at subaerial conditions (Mathez, 1984; Sisson andGrove, 1993) or exposure to atmospheric oxygen (Furukawa et al.,2010; Mollo et al., 2013b).
4.3. The thermal path of eruptions
To test whether or not, at the time of the eruption, Etneanclinopyroxenes were in equilibrium with the melts, we have used themodel derived by Mollo et al. (2013a) and based on the difference (Δ)between diopside + hedenbergite (DiHd) contents measured in theanalysed crystals with those predicted for clinopyroxene via regressionanalysis of clinopyroxene–melt pairs in equilibrium condition. Assum-ing near-equilibrium crystallization within 10% of the one-to-one line(cf. Jeffery et al., 2013), most of the crystal compositions from lavafountains equilibrated with the host magma (Fig. 5a). Conversely,clinopyroxene–melt pairs from lava flows indicate both equilibriumand disequilibrium crystallization conditions, with the majority of datashowingnet departure from the one-to-one line. According to a numberof studies, under kinetically-controlled crystallization conditions, Al3+
substitutes for Si4+ in clinopyroxene tetrahedral site (see Fig. 2c) caus-ing charge imbalances that are compensated by increasing concentra-tions of highly charged cations, such as Ti (Smith and Lindsley, 1971;Grove and Bence, 1977; Grove and Raudsepp, 1978; Coish and Taylor,1979; Gamble and Taylor, 1980; Blundy et al., 1998; Hill et al., 2000;Wood and Blundy, 2001; Lofgren et al., 2006; Mollo et al., 2013d;Scarlato et al., 2014). In our compositions, the partitioning of Ti between
Fig. 5. Test for equilibrium between clinopyroxene and melt based on thediopside + hedenbergite (DiHd) model of Mollo et al. (2013b) (a). Thermal path (TP)predicted by Eq. (4) vs. the partition coefficient of Ti (cpx–meltKTi)measured at the interfacebetween clinopyroxene and coexistingmelt (b). Clinopyroxene compositional variation interms of MgO and FeO concentrations (c). H2O estimated by the hygrometer of Armientiet al. (2013) vs. FeO/MgO ratio of clinopyroxene.
clinopyroxene and melt compositions (i.e., cpx–meltKTi) increases by afactor of four from lava fountains to pahoehoe to aa lavas (Fig. 5b),showing values out of the equilibrium range experimentally-derivedfor Etnean compositions (cf. Mollo et al., 2013d). Additionally, takinginto account the close textural relationship between clinopyroxeneand titanomagnetite, it is interesting to note that as long as Ti is prefer-entially incorporated in clinopyroxene (Fig. 5b), its concentration de-creases in titanomagnetite with the effect of cooling (Fig. 4a). Thesecation redistributions are generally controlled by kinetic or time-dependent pathways in which crystal compositions attempt to equili-bratewith an ever-changing temperature bounded between theminer-al saturation temperature and the closure temperature of crystalgrowth. Under such conditions, the thermal path (TP) of the systemcan be estimated trough the following clinopyroxene-based expression(Mollo et al., 2013a):
TP °Cð Þ ¼ −20:87þ 650:39 � ΔDiHd: ð4Þ
As anticipated above, scoria clasts erupted during lava fountainingwere rapidly quenched at the contact with the atmosphere, thuspreventing post-eruption crystallization processes. Coherently, Eq. (4)indicates that the closure temperature of clinopyroxene crystal growthis constrained in a narrow thermal path with maximum value of25 °C, suggesting that most of the crystallization occurred close to theclinopyroxene saturation temperature (Fig. 5b). In contrast, the finalcompositions of clinopyroxenes from lava flows responded to greatchanges in temperature caused by degassing in the conduit and coolingat subaerial conditions. As a consequence, Eq. (4) predicts more extend-ed thermal paths showing maximum values of 70 and 140 °C forpahoehoe and aa lavas, respectively (Fig. 5b). The broad thermal rangeobtained for aa samples can be addressed to radiative cooling of theflow interior due to continuous crustal disruption (cf. Crisp andBaloga, 1994), whereas the rapid formation of a thick crust in pahoehoelavas may have insulated the flow interior by conductive heat transfer(cf. Cashman et al., 1999).
Although H2O exsolves with decreasing pressure, ascending magmasatMt. Etna can retain part of their original volatile contentwith the resultof abundant degassing at subaerial conditions (Lanzafame et al., 2013;Mollo et al., 2013b). Through phase equilibrium experiments, Gardneret al. (1995) found a positive correlation between the Fe/Mg ratio ofclinopyroxene and the water dissolved in the melt. The mechanism bywhich water might reduce the activity of MgO relative to FeO in themelt can be expressed by the reaction:
Fe OHð Þ2 þMgO→Mg OHð Þ2 þ FeO: ð5Þ
Thermodynamic calculations conducted by Frey and Lange (2011)suggest that hydroxyl groups preferentially complex with Mg2+ ratherthan Fe2+ and that dissolved hydroxyl groupsmay reduce the activity ofMgO relative to FeO in themelt. Looking at our compositions, we noticethat clinopyroxenes from lava fountains are enriched in FeO, whereasthose from pahoehoe and aa lavas show FeO depletions (Fig. 5c). As aconsequence, it would be expected that magmas feeding lava flowswere more degassed than those supplying lava fountains. The melt–water content was derived through the clinopyroxene-based hygrome-ter of Armienti et al. (2013) whose calibration dataset consists only ofEtnean compositions. According to Eq. (5), our estimates indicate thatlava fountains equilibrated with moderate water contents (1.7–2.8wt.%), whereas lavaflowsweremore degassed (0–1.9wt.%) perhapsdue to continuous loss of gases duringflow and cooling (Fig. 5d). To bet-ter constrain the thermal paths of lava fountain and lava flow magmas,we have used the equilibrium clinopyroxene–melt pairs as input datafor the thermometer of Putirka et al. (1996) that shows the advantageto be independent to both pressure and melt–water concentration.The histogram of lava fountains indicates that clinopyroxene crystalsequilibrated at 1050–1125 °C (Fig. 6), with a unimodal distribution
Fig. 6. Crystallization temperatures of lava fountains, aa and pahoehoe lavas estimatedthrough the P- and H2O-independent thermometer of Putirka et al. (1996).
52 S. Mollo et al. / Chemical Geology 406 (2015) 45–54
peaked at 1075 °C and close to the saturation temperature of Etneantrachybasalts (Mollo et al., 2015). Conversely, the histograms of lavaflows indicate crystallization conditions of 950–1075 °C (unimodal dis-tribution peaked at 1000 °C) and 875–1000 °C (bimodal distributionwith the maximum peak at 900 °C) for pahoehoe and aa lavas, respec-tively. The closure temperature of the crystal growth estimated for aalavas is generally lower than that of pahoehoe types. This differentthermal condition agrees with the temperature decrease measured,for example, at Kilauea volcano during the transition from pahoehoeto aa (Cashman et al., 1999) as well as with the crystallization and vis-cosity increase necessary to produce high yield strength lavas (Crispand Baloga, 1994; Crisp et al., 1994)
5. Conclusions
The following key conclusions can be drawn from this study:
(1) lava fountains were rapidly quenched at the time of eruptionpreventing kinetic effects due to post-eruptive cooling;
(2) the compositions of clinopyroxene and titanomagnetite rims re-corded slowcooling rates during fast ascent of high-temperature,H2O-rich magmas within the plumbing system of the volcano;
(3) in contrast, pahoehoe and aa lavas underwent more complexcrystallization histories that were influenced by degassing inthe conduit and cooling during flowage onto the surface;
(4) the final crystallization of pahoehoe lavas was driven by coolingrates generally lower than those measured for aa lavas, whereasthe opposite applies to the closure temperature of crystal growthkinetics;
(5) the residual melts of both pahoehoe and aa lavas are highlydegassed and oxidized suggesting abundant volatile loss at sub-aerial conditions.
Acknowledgments
The authors are grateful to B. S. Ellis, an anonymous reviewer andD. B. Dingwell as editor-in-chief for their useful and constructive sug-gestions. A. Cavallo is acknowledged for assistance during electron
microprobe analysis. The research activities of the HP–HT laboratoryof the INGV were supported by the European Observing System Infra-structure Project (EPOS) Grant agreement no. 262229.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.chemgeo.2015.04.017.
References
Aloisi, M., Bonaccorso, A., Gambino, S., 2006. Imaging composite dike propagation (Etna,2002 case). J. Geophys. Res. 111, B06404.
Andronico, D., Corsaro, R.A., 2011. Lava fountains during the episodic eruption of South-East Crater (Mt. Etna), 2000: insights into magma-gas dynamics within the shallowvolcano plumbing system. Bull. Volcanol. 73 (9), 1165–1178.
Andronico, D., Branca, S., Calvari, S., Burton, M.R., Caltabiano, T., Corsaro, R.A., Del Carlo, P.,Garfì, G., Lodato, L., Miraglia, L., Murè, F., Neri, M., Pecora, E., Pompilio, M., Salerno, G.,Spampinato, L., 2005. A multi-disciplinary study of the 2002–03 Etna eruption:insights for a complex plumbing system. Bull. Volcanol. 67, 314–330.
Applegarth, L.J., Tuffen, H., James, M.R., Pinkerton, H., Cashman, K.V., 2013. Direct observa-tions of degassing-induced crystallization in basalts. Geology 41, 243–246.
Ariskin, A.A., Nikolaev, G.S., 1996. An empirical model for the calculation of spinel–meltequilibria in mafic igneous systems at atmospheric pressure: 1. Chromian spinels.Contrib. Mineral. Petrol. 123, 282–292.
Armienti, P., Barberi, F., Innocenti, F., Pompilio, M., Romano, R., Villari, L., 1984.Compo- sitional variation in the 1983 and other recent Etnean lavas: insightson the shallow feeding system. Bull. Volcanol. 47, 995–1007.
Armienti, P., Pareschi, M.T., Pompilio, M., Innocenti, F., 1994. Effects of magma storage andascent on the kinetics of crystal growth: the case of the 1991–1993Mt. Etna eruption.Contrib. Mineral. Petrol. 115, 402–414.
Armienti, P., Tonarini, S., D'Orazio, M., Innocenti, F., 2004. Genesis and evolution of MountEtna alkaline lavas: petrological and Sr–Nd–B isotope constraints. Period. Mineral. 73,29–52.
Armienti, P., Tonarini, S., Innocenti, F., D'Orazio, M., 2007. Mount Etna pyroxene as tracerof petrogenetic processes and dynamics of the feeding system. In: Beccaluva, L., et al.(Eds.), Cenozoic Volcanism in the Mediterranean Area. Geological Society of America144, pp. 265–276.
Armienti, P., Perinelli, C., Putirka, K.D., 2013. A new model to estimate deep-level magmaascent rates, with applications to Mt. Etna (Sicily, Italy). J. Petrol. 54, 795–813.
Baker, D.R., 2008. The fidelity of melt inclusions as records of melt composition. Contrib.Mineral. Petrol. 156, 377–395.
Behncke, B., Branca, S., Corsaro, R.A., De Beni, E., Miraglia, L., Proietti, P., 2014. The2011–2012 summit activity of Mount Etna: birth, growth and products of the newSE crater. J. Volcanol. Geotherm. Res. 270, 10–21.
Blundy, J., Robinson, J., Wood, B., 1998. Heavy REE are compatible in clinopyroxene on thespinel lherzolite solidus. Earth Planet. Sci. Lett. 160, 493–504.
Bonaccorso, A., Calvari, S., Currenti, G., Del Negro, C., Ganci, G., Linde, A., Napoli, R., Sacks,S., Sicali, A., 2013. From source to surface: dynamics of Etna lava fountains investigat-ed by continuous strain, magnetic, ground and satellite thermal data. Bull. Volcanol.75, 690.
Cashman, K.V., Marsh, B.D., 1988. Crystal size distribution (CSD) in rocks and the kineticsand dynamics of crystallization II: Makaopuhi lava lake. Contrib. Mineral. Petrol. 99,292–305.
Cashman, K.V., Thornber, C., Kauahikaua, J.P., 1999. Cooling and crystallization of lava inopen channels, and the transition of pahoehoe to aa. Bull. Volcanol. 61, 306–323.
Chester, D.K., Duncan, A.M., Guest, J.E., Kilburn, C.R.J., 1985. Mount Etna: The Anatomy of aVolcano. Chapman and Hall, London.
Coish, R.A., Taylor, L.A., 1979. The effects of cooling rate on texture and pyroxene chemis-try in DSDP Leg 34 basalt: a microprobe study. Earth Planet. Sci. Lett. 42, 389–398.
Collins, S.J., Pyle, D.M., Maclennan, J., 2013. Melt inclusions track pre-eruption storage anddehydration of magmas at Etna. Geology 37, 571–574.
Corsaro, R.A., Miraglia, L., Pompilio, M., 2007. Petrological evidence of a complex plumb-ing system feeding the July–August 2001 eruption of Mt. Etna, Sicily, Italy. Bull.Volcanol. 69, 401–421.
Corsaro, R.A., Metrich, N., Allard, P., Andronico, D., Miraglia, L., Fourmentraux, C., 2009. The1974 flank eruption of Mount Etna: an archetype for deep dike-fed eruptions at ba-saltic volcanoes and a milestones in Etna's recent history. J. Geophys. Res. B07204http://dx.doi.org/10.1029/2008JB006013.
Corsaro, R.A., Di Renzo, V., Distefano, S., Miraglia, L., Civetta, L., 2013. Relationship be-tween petrologic processes in the plumbing system of Mt. Etna and the dynamicsof the eastern flank from 1995 to 2005. J. Volcanol. Geotherm. Res. 251, 75–89.
Crisp, J.A., Baloga, S.M., 1994. Influence of crystallization and entrainment of cooler materialon the emplacement of basaltic aa lava flows. J. Geophys. Res. 99 (B6), 11819–11831.
Crisp, J., Baloga, S., 1990. A model for lava flows with two thermal components.J. Geophys. Res. 95, 1255–1270. http://dx.doi.org/10.1029/89JB01696.
Crisp, J., Cashman, K.V., Bonin, J.A., Hougen, S.B., Pieri, D.C., 1994. Crystallization history ofthe Mauna Loa lava flow. J. Geophys. Res. 99, 7177–7198.
53S. Mollo et al. / Chemical Geology 406 (2015) 45–54
Del Gaudio, P., Mollo, S., Ventura, G., Iezzi, G., Taddeucci, J., Cavallo, A., 2010. Cooling rate-induced differentiation in anhydrous and hydrous basalts at 500 MPa: implicationsfor the storage and transport of magmas in dikes. Chem. Geol. 270, 164–178.
Demouchy, S., Jacobsen, S.D., Gaillard, F., Stern, C.R., 2006. Rapid magma ascent recordedby water diffusion profiles in mantle olivine. Geology 34, 429–432.
Ferlito, C., Lanzafame, G., 2010. The role of supercritical fluids in the potassium enrich-ment of magmas at Mount Etna volcano (Italy). Lithos 119, 642–650.
Ferlito, C., Viccaro, M., Cristofolini, R., 2008. Volatile-induced magma differentiation in theplumbing system of Mt. Etna volcano (Italy): evidence from glass in tephra of the2001 eruption. Bull. Volcanol. 70, 455–473.
Ferlito, C., Coltorti, M., Cristofolini, R., Giacomoni, P.P., 2009. The contemporaneous emis-sion of low-K and high-K trachybasalts and the role of the NE Rift during the 2002eruptive event, Mt. Etna, Italy. Bull. Volcanol. 71, 575–587.
Ferlito, C., Viccaro, M., Nicotra, E., Cristofolini, R., 2011. Regimes of magma recharge on theeruptive behaviour during the period 2001–2005 at Mt. Etna volcano. Bull. Volcanol.74 (2), 533–543.
Ferlito, C., Massimo, C., Lanzafame, G., Giacomoni, P.P., 2014. The volatile flushing triggerseruptions at open conduit volcanoes: evidence fromMount Etna volcano (Italy). Lith-os 184–187, 447–455.
Frey, H.M., Lange, R., 2011. Phenocryst complexity in andesites and dacites from theTequila volcanic field, Mexico: resolving the effects of degassing vs. magma mixing.Contrib. Mineral. Petrol. 162, 415–445.
Furukawa, H., Ko, N., Go, Y.B., Aratani, N., Choi, S.B., Choi, E., Özgür Yazaydin, A., Snurr,R.Q., O'Keeffe, M., Kim, J., Yaghi, O.M., 2010. Ultrahigh porosity in metal–organicframeworks. Science 424–428.
Gamble, R.P., Taylor, L.A., 1980. Crystal/liquid partitioning in augite: effects of cooling rate.Earth Planet. Sci. Lett. 47, 21–33.
Gardner, J.E., Rutherford, M., Carey, S., Sigurdsson, H., 1995. Experimental constraints onpre-eruptive water contents and changing magma storage prior to explosive erup-tions of Mount St. Helens volcano. Bull. Volcanol. 57, 1–17.
Giacomoni, P.P., Ferlito, C., Coltorti, M., Bonadiman, C., Lanzafame, G., 2014. Plagioclase asarchive of magma ascent dynamics on “open conduit” volcanoes: the 2001–2006eruptive period at Mount Etna. Earth Sci. Rev. 138, 371–393.
Giordano, D., Dingwell, D.B., 2003. Non-Arrhenian multicomponent melt viscosity: amodel. Earth Planet. Sci. Lett. 208, 337–349.
Gonnermann, H.M., Manga, M., 2013. Magma ascent in the volcanic conduit. In: Fagents,S.A., Gregg, T.K.P., Lopez, R.C. (Eds.), Modeling Volcanic Processes: The Physics andMathematics of Volcanism. Cambridge Univ Press, pp. 55–84.
Gottsmann, J., Harris, A.J.L., Dingwell, D.B., 2004. Thermal history of Hawaiian pahoehoelava crusts at the glass transition: implications for flow rheology and emplacement.Earth Planet. Sci. Lett. 228, 343–353.
Grove, T.L., Bence, A.E., 1977. Experimental studyof pyroxene–liquid interaction inquartz-normative basalt 15597. Proceedings of Lunar and Planetary Science Conference, 8th,pp. 1549–1579.
Grove, T.L., Raudsepp, M., 1978. Effects of kinetics on the crystallization of quartz-normative basalt 15597: an experimental study. Proceedings of Lunar and PlanetaryScience Conference, 9th, pp. 585–599.
Hammer, J.E., 2008. Experimental studies of the kinetics and energetics of magma crystal-lization. In: Putirka, K.D., Tepley, F.J. (Eds.), Minerals, Inclusions and Volcanic Process-es. Rev. Mineral. Geochem. 69, pp. 9–59.
Hammond, P.A., Taylor, L.A., 1982. The ilmenite–titanomagnetite assemblage: kinetics ofre-equilibration. Earth Planet. Sci. Lett. 61, 143–150.
Harris, A., Rowland, S., 2009. Effusion rate controls on lava flow length and the role of heatloss: a review. In: Thordarson, T., Self, S., Larsen, G., Rowland, S.K., Hoskuldsson, A.(Eds.), Studies in Volcanology: The Legacy of George Walker. Special Publications ofIAVCEI, pp. 33–51.
Hildner, H., Klügle, A., Hauff, F., 2011.Magma storage and ascent during the 1995 eruptionof Fogo, Cape Verde Archipelago. Contrib. Mineral. Petrol. 162, 751–772.
Hill, E., Wood, B.J., Blundy, J.D., 2000. The effect of Ca-Tschermaks component on traceelement partitioning between clinopyroxene and silicate melt. Lithos 53,203–215.
Hon, K.A., Gansecki, C., Kauahikaua, J., 2003. The transition from 'a'a ̄ to pa h̄oehoe crust onflows emplaced during the Pu'u ‘O’o- Kupaianaha eruption. In: Heliker, C.A., et al.(Eds.), The Pu'u ‘O’o-Kupaianaha Eruption of Kilauea Volcano, Hawai'i: The First20 Years. U.S. Geological Survey Professional Paper 1676, pp. 89–104.
Hort, M., Gardner, J., 2000. Constraints on cooling and degassing of pumice during Plinianvolcanic eruptions based on model calculations. J. Geophys. Res. 105, 25981–26001.
Iezzi, G., Mollo, S., Torresi, G., Ventura, G., Cavallo, A., Scarlato, P., 2011. Experimental so-lidification of an andesitic melt by cooling. Chem. Geol. http://dx.doi.org/10.1016/j.chemgeo.2011.01.024.
Iezzi, G., Mollo, S., Shaini, E., Cavallo, A., Scarlato, P., 2014. The cooling kinetics of plagio-clase revealed by electron microprobe mapping. Am. Mineral. http://dx.doi.org/10.2138/am.2014.4626.
Ishibashi, H., 2013. Spinel–melt oxygen barometry: a method and application to Cenozoicalkali basaltic magmas from the Higashi–Matsuura district, NW Kyushu, Japan.Geosci. Repts. 40, 21–32.
Jeffery, A.J., Gertisser, R., Troll, V.R., Jolis, E.M., Dahren, B., Harris, C., Tindle, A.G., Preece, K.,O'Driscoll, B., Humaida, H., Chadwick, J.P., 2013. Magma plumbing system of the2007–2008 dome-forming eruption of Kelut volcano, East Java, Indonesia. Contrib.Mineral. Petrol. 166, 275–308.
Kahl, M., Chakraborty, S., Costa, F., Pompilio, M., 2011. Dynamic plumbing system beneathvolcanoes revealed by kinetic modelling and the connection to monitoring data: anexample from Mt. Etna. Earth Planet. Sci. Lett. 308, 11–22.
Kilburn, C.R.J., Guest, J.E., 1993. AA lavas of Mount Etna, Sicily. In: Kilburne, C.R.J., Luongo,G. (Eds.), Active Lavas:Monitoring andModelling. University College of London Press,pp. 73–106.
Kostov, I., Kostov. R.I., 1999. Crystal Habits of Minerals. Bulgarian Academic Monographs,Ed. Pensoft Publishers, Academic Publishing House, Sofia, 415 pp.
Lachance, G.R., Trail, R.J., 1966. Practical solution to the matrix problem in X-ray analysis.Can. Spectrosc. 11, 43–48.
Lanzafame, G., Mollo, S., Iezzi, G., Ferlito, C., Ventura, G., 2013. Unrevealing the solidifica-tion path of a pahoehoe “cicirara” lava from Mount Etna volcano. Bull. Volcanol. 75,703.
Lofgren, G.E., Huss, G.R., Wasserburg, G.J., 2006. An experimental study of trace elementpartitioning between Ti–Al–clinopyroxene and melt: equilibrium and kinetic effectsincluding sector zoning. Am. Mineral. 91, 1596–1606.
Mathez, E.A., 1984. Influence of degassing on oxidation states of basaltic magmas. Nature310, 371–375.
Médard, E., Grove, T., 2008. The effect of H2O on the olivine liquidus of basaltic melts: ex-periments and thermodynamic models. Contrib. Mineral. Petrol. 155, 417–432.
Metrich, N., Rutherford, M.J., 1998. Low pressure crystallization paths of H2O-saturatedbasaltic–hawaitic melts from Mt Etna: implications for open-system degassing ofbasaltic volcanoes. Geochim. Cosmochim. Acta 62, 1195–1205.
Métrich, N., Allard, P., Spilliaert, N., Andronico, D., Burton, M., 2004. 2001 flank eruption ofthe alkali- and volatile-rich primitive basalt responsible for Mount Etna's evolution inthe last three decades. Earth Planet. Sci. Lett. 228, 1–17.
Mollo, S., Lanzafame, G., Masotta, M., Iezzi, G., Ferlito, C., Scarlato, P., 2011. Cooling historyof a dike as revealed bymineral chemistry: a case study fromMt. Etna volcano. Chem.Geol. 288, 39–52.
Mollo, S., Iezzi, G., Ventura, G., Cavallo, A., Scarlato, P., 2012. Heterogeneous nucleationmechanisms and formation of metastable phase assemblages induced by differentcrystalline seeds in a rapidly cooled andesitic melt. J. Non-Cryst. Solids 358,1624–1628. http://dx.doi.org/10.1016/j.jnoncrysol.2012.04.010.
Mollo, S., Putirka, K., Misiti, V., Soligo, M., Scarlato, P., 2013a. A new test for equilibriumbased on clinopyroxene–melt pairs: clues on the solidification temperatures ofEtnean alkaline melts at post-eruptive conditions. Chem. Geol. 352, 92–100.
Mollo, S., Scarlato, P., Lanzafame, G., Ferlito, C., 2013b. Deciphering lava flow post-eruption differentiation processes by means of geochemical and isotopic variations:a case study from Mt. Etna volcano. Lithos 162–163, 115–127.
Mollo, S., Putirka, K., Iezzi, G., Scarlato, P., 2013c. The control of cooling rate ontitanomagnetite composition: implications for a geospeedometry model applicableto alkaline rocks from Mt. Etna volcano. Contrib. Mineral. Petrol. 165, 457–475.http://dx.doi.org/10.1007/s00410-012-0817-6.
Mollo, S., Blundy, J., Scarlato, P., Iezzi, G., Langone, A., 2013d. The partitioning of trace el-ements between clinopyroxene and trachybasaltic melt during rapid cooling andcrystal growth. Contrib. Mineral. Petrol. 166, 1633–1654.
Mollo, S., Giacomoni, P.P., Coltorti, M., Ferlito, C., Iezzi, G., Scarlato, P., 2015. Reconstructionof magmatic variables governing recent Etnean eruptions: constraints from mineralchemistry and P–T–fO2–H2O conditions. Lithos 212–215, 311–320.
Morimoto, N., Fabries, J., Ferguson, A.K., Ginzburg, I.V., Ross, M., Seifert, F.A., Zussman, J.,Aoki, K., Gottardi, G., 1988. Nomenclature of pyroxenes. Am. Mineral. 73, 1123–1133.
Pinkerton, H., Norton, G., 1995. Rheological properties of basaltic lavas at sub-liquidustemperatures: laboratory and field measurements on lavas from Mount Etna.J. Volcanol. Geoth. Res. 68, 307–323.
Pupier, E., Duchene, S., Toplis, M.J., 2008. Experimental quantification of plagioclase crystalsize distribution during cooling of basaltic liquid. Contrib. Mineral. Petrol. 155, 555–570.
Putirka, K., 2008. Thermometers and barometers for volcanic systems. In: Putirka, K.D.,Tepley, F. (Eds.), Minerals, Inclusions and Volcanic Processes. Rev. Mineral. Geochem.vol. 69. Mineralogical Society of America, pp. 61–120.
Putirka, K., Johnson, M., Kinzler, R., Walker, D., 1996. Thermobarometry of mafic igneousrocks based on clinopyroxene–liquid equilibria, 0–30 kbar. Contrib. Mineral. Petrol.123, 92–108.
Putirka, K., Ryerson, F.J., Mikaelian, H., 2003. New igneous thermobarometers for maficand evolved lava compositions, based on clinopyroxene+ liquid equilibria. Am. Min-eral. 88, 1542–1554.
Rutherford, M.J., 2008. Magma ascent rates. Rev. Mineral. Geochem. 69, 241–271.Rutherford, M.J., Hill, P.M., 1993. Magma ascent rates from amphibole breakdown:
experiments and the 1980–1986 Mount St. Helens eruptions. J. Geophys. Res. 98,19667–19685.
Sack, R.O., Carmichael, I.S.E., Rivers, M.L., Ghiorso, M.S., 1980. The ferric–ferrous equilibriain natural silicate liquids at 1 bar. Contrib. Mineral. Petrol. 75, 369–376.
Sato, H., 1995. Textural difference between pahoehoe and aa lavas of Izu-Oshima volcano,Japan: an experimental study on population density of plagioclase. J. Volcanol.Geotherm. Res. 66, 101–113.
Scarlato, P., Mollo, S., Blundy, J., Iezzi, G., Tiepolo, M., 2014. The role of natural solidifica-tion paths on REE partitioning between clinopyroxene and melt. Bull. Volcanol. 76:810. http://dx.doi.org/10.1007/s00445-014-0810-1.
Schiavi, F., Walte, N., Keppler, H., 2009. First in situ observation of crystallization processesin a basaltic–andesitic melt with the moissanite cell. Geology 37, 963–966.
Shea, T., Hammer, J.E., 2013. Kinetics of decompression and cooling-induced crystallization ofmafic-intermediate hydrous magmas. J. Volcanol. Geotherm. Res. 260, 127–145.
Sisson, T.W., Grove, T.L., 1993. Experimental investigations of the role of H2O incalcalkaline differentiation and subduction zonemagmatism. Contrib. Mineral. Petrol.113, 143–166.
Smith, D., Lindsley, D.H., 1971. Stable andmetastable augite crystallization trends in a sin-gle basalt flow. Am. Mineral. 56, 225–233.
Spilliaert, N., Allard, P., Metrich, N., Sobolev, A.V., 2006. Melt inclusion record of the con-ditions of ascent, degassing, and extrusion of volatile-rich alkali basalt feeding thepowerful 2002 flank eruption of Mount Etna (Italy). J. Geophys. Res. 111, B04203.
54 S. Mollo et al. / Chemical Geology 406 (2015) 45–54
Teng, H.R., 2013. How ions and molecules organize to form crystals. Elements 9 (3),189–194.
Toramaru, A., Noguchi, S., Oyoshihara, S., Tsune, A., 2008. MND (microlite number densi-ty) water exsolution rate meter. J. Volcanol. Geotherm. Res. 175, 156–167.
Ujike, O., 1982. Microprobe mineralogy of plagioclase, clinopyroxene and amphibole asrecords of cooling rate in the Shirotori–Hiketa dike swarm, northeastern Shikoku,Japan. Lithos 15, 281–293.
Venezky, D.Y., Rutherford, M.J., 1997. Preeruption conditions and timing of dacite–andes-ite magma mixing in the 2.2 ka eruption at Mount Rainier. J. Geophys. Res. 102,20069–20086.
Vetere, F., Mollo, S., Giacomoni, P.P., Iezzi, G., Coltorti, M., Ferlito, C., Holtz, F., Perugini, D.,Scarlato, P., 2005. Experimental constraints on the origin of pahoehoe “cicirara” lavasat Mt. Etna volcano (Sicily, Italy). Bull. Volcanol. 77:44. http://dx.doi.org/10.1007/s00445-015-0931-1.
Viccaro, M., Giacomoni, P.P., Ferlito, C., Cristofolini, R., 2010. Dynamics ofmagma supply atMt. Etna volcano (Southern Italy) as revealed by textural and compositional featuresof plagioclase phenocrysts. Lithos 116, 77–91.
Wall, K.T., Rowe, M.C., Ellis, B.S., Schmidt, M.E., Eccles, J.D., 2014. Determining volcaniceruption styles on Earth and Mars from crystallinity measurements. Nat. Commun.5, 5090. http://dx.doi.org/10.1038/ncomms6090.
Wilding, M., Webb, S.L., Dingwell, D.B., 1995. Evaluation of a relaxation geospeedometerfor volcanic glasses. Chem. Geol. 125, 137–148.
Wood, B.J., Blundy, J.D., 2001. The effect of cation charge on crystal-melt partitioning oftrace elements. Earth Planet. Sci. Lett. 188, 59–72.
Zhou, W., Van der Voo, R., Peacor, D.R., Zhang, Y., 2000. Variable Ti-content and grain sizeof titanomagnetite as a function of cooling rate in very young MORB. EPSL 179, 9–20.
