RESEARCH ARTICLE Experimental constraints on the origin of pahoehoe Bcicirara^ lavas at Mt. Etna Volcano (Sicily, Italy) F. Vetere 1,2 & S. Mollo 3 & P. P. Giacomoni 4 & G. Iezzi 3,5 & M. Coltorti 4 & C. Ferlito 6 & F. Holtz 2 & D. Perugini 1 & P. Scarlato 3 Received: 7 February 2015 /Accepted: 11 April 2015 # Springer-Verlag Berlin Heidelberg 2015 Abstract We present results from phase equilibria experi- ments conducted on the most primitive pahoehoe Bcicirara^ trachybasaltic lava flow ever erupted at Mt. Etna Volcano. This lava is characterized by a pahoehoe morphology in spite of its high content of phenocrysts and microphenocrysts (>40 vol%) with the occurrence of centimetre-sized plagio- clases (locally named cicirara for their chick-pea-like appear- ance). Our experiments have been performed at 400 MPa, 1100–1150 °C and using H 2 O and CO 2 concentrations corre- sponding to the water-undersaturated crystallization conditions of Etnean magmas. Results show that olivine does not crystal- lize from the melt, whereas titanomagnetite is the liquidus phase followed by clinopyroxene or plagioclase as a function of melt–water concentration. This mineralogical feature contrasts with the petrography of pahoehoe cicirara lavas sug- gesting early crystallization of olivine and late formation of titanomagnetite after plagioclase and/or in close association with clinopyroxene. The lack of olivine produces MgO-rich melt compositions that do not correspond to the evolutionary behaviour of cicirara magmas. Moreover, in a restricted thermal path of 50 °C and over the effect of decreasing water concen- trations, we observe abundant plagioclase and clinopyroxene crystallization leading to trace element enrichments unlikely for natural products. At the same time, the equilibrium composi- tions of our mineral phases are rather different from those of natural cicirara phenocrysts and microphenocrysts. The com- parison between our water-undersaturated data and those from previous degassing experiments conducted on a similar Etnean trachybasaltic composition demonstrates that pahoehoe cicirara lavas originate from crystal-poor, volatile-rich magmas under- going abundant degassing and cooling in the uppermost part of the plumbing system and at subaerial conditions where most of the crystallization occurs after the development of pahoehoe surface crusts. Keywords Mt. Etna . Pahoehoe Bcicirara^ lavas . Water exsolution . Degassing Introduction Since ca. 220 ka, the magmatic activity of Mt. Etna Volcano (Sicily, Italy) is characterized by an alkaline affinity with abun- dant eruption of trachybasaltic and basaltic trachyandesitic lava flows (Corsaro et al. 2007, 2009). The plumbing system is fed by continuous supply from mantle depths of primitive, volatile- rich magmas that mix with more evolved, degassed melts re- siding at shallow crustal levels (Armienti et al. 2007, 2013; Ferlito et al. 2008; Corsaro et al. 2013; Giacomoni et al. Editorial responsibility: M. Manga Electronic supplementary material The online version of this article (doi:10.1007/s00445-015-0931-1) contains supplementary material, which is available to authorized users. * S. Mollo [email protected]1 Dipartimento di Fisica e Geologia, Università di Perugia, Piazza Università 1, 06100 Perugia, Italy 2 Institute for Mineralogy, Leibniz University of Hannover, Callinstr. 3, 30167 Hannover, Germany 3 Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, 00143 Roma, Italy 4 Dipartimento di Fisica e Science della Terra, Università di Ferrara, Via Saragat 1, 44122 Ferrara, Italy 5 Dipartimento di Ingegneria & Geologia, Università G. d’Annunzio, Via Dei Vestini 30, 66013 Chieti, Italy 6 Dipartimento di Scienze Biologiche, Geologiche e Ambientali, Università di Catania, Via A. Longo 19, 95125 Catania, Italy Bull Volcanol (2015) 77:44 DOI 10.1007/s00445-015-0931-1
Transcript
RESEARCH ARTICLE
Experimental constraints on the origin of pahoehoe Bcicirara^lavas at Mt. Etna Volcano (Sicily, Italy)
F. Vetere1,2 & S. Mollo3 & P. P. Giacomoni4 & G. Iezzi3,5 & M. Coltorti4 & C. Ferlito6 &
F. Holtz2 & D. Perugini1 & P. Scarlato3
Received: 7 February 2015 /Accepted: 11 April 2015# Springer-Verlag Berlin Heidelberg 2015
Abstract We present results from phase equilibria experi-ments conducted on the most primitive pahoehoe Bcicirara^trachybasaltic lava flow ever erupted at Mt. Etna Volcano.This lava is characterized by a pahoehoe morphology in spiteof its high content of phenocrysts and microphenocrysts(>40 vol%) with the occurrence of centimetre-sized plagio-clases (locally named cicirara for their chick-pea-like appear-ance). Our experiments have been performed at 400 MPa,1100–1150 °C and using H2O and CO2 concentrations corre-sponding to the water-undersaturated crystallization conditionsof Etnean magmas. Results show that olivine does not crystal-lize from the melt, whereas titanomagnetite is the liquidusphase followed by clinopyroxene or plagioclase as a functionof melt–water concentration. This mineralogical feature
contrasts with the petrography of pahoehoe cicirara lavas sug-gesting early crystallization of olivine and late formation oftitanomagnetite after plagioclase and/or in close associationwith clinopyroxene. The lack of olivine produces MgO-richmelt compositions that do not correspond to the evolutionarybehaviour of cicirara magmas.Moreover, in a restricted thermalpath of 50 °C and over the effect of decreasing water concen-trations, we observe abundant plagioclase and clinopyroxenecrystallization leading to trace element enrichments unlikely fornatural products. At the same time, the equilibrium composi-tions of our mineral phases are rather different from those ofnatural cicirara phenocrysts and microphenocrysts. The com-parison between our water-undersaturated data and those fromprevious degassing experiments conducted on a similar Etneantrachybasaltic composition demonstrates that pahoehoe ciciraralavas originate from crystal-poor, volatile-rich magmas under-going abundant degassing and cooling in the uppermost part ofthe plumbing system and at subaerial conditions where most ofthe crystallization occurs after the development of pahoehoesurface crusts.
Since ca. 220 ka, the magmatic activity of Mt. Etna Volcano(Sicily, Italy) is characterized by an alkaline affinity with abun-dant eruption of trachybasaltic and basaltic trachyandesitic lavaflows (Corsaro et al. 2007, 2009). The plumbing system is fedby continuous supply frommantle depths of primitive, volatile-rich magmas that mix with more evolved, degassed melts re-siding at shallow crustal levels (Armienti et al. 2007, 2013;Ferlito et al. 2008; Corsaro et al. 2013; Giacomoni et al.
Editorial responsibility: M. Manga
Electronic supplementary material The online version of this article(doi:10.1007/s00445-015-0931-1) contains supplementary material,which is available to authorized users.
2014;Mollo et al. 2015).Most of the crystallization is driven bydecompression, degassing and cooling of magmas during as-cent in the conduit and eruption onto the surface (Viccaro et al.2010; Corsaro et al. 2013; Mollo et al. 2013a; Lanzafame et al.2013; Applegarth et al. 2013). Upon eruption, lavas may occuras pahoehoe and/or aa flows in response to a number of chemo-physical parameters such as magma composition, temperature,crystal content, viscosity of suspension, terrain slope, magmaoutput and strain rates (e.g. Cashman et al. 1999; Hon et al.2003; Kilburn 2004; Iezzi and Ventura 2005). In particular,most of the primitive pahoehoe lava flows at Mt. Etna containcentimetre-sized plagioclases, locally named Bcicirara^ due totheir chick-pea-like appearance (Fig. 1a–c). Althoughpahoehoe lava flows are relatively rare in contrast to very com-mon aa (e.g. Branca et al. 2011), the cicirara lavas outcrop inmany sectors of the volcano, and their peculiar texture has beeninterpreted as the result of a long residence time of magma in ashallow reservoir (e.g. Corsaro and Pompilio 2004). Theselavas show phase assemblages (olivine+clinopyroxene+pla-gioclase+titanomagnetite) and bulk compositions (fromtrachybasalt to basaltic trachyandesite) that are comparable tothose of prehistoric and historical Etnean products (Fig. 1d, e).The most intriguing feature of pahoehoe cicirara lava flows isthe marked abundance of phenocrysts and microphenocrysts(>40 vol%) that is much higher than the value suggested forthe transition from pahoehoe to aa morphology. In fact, fieldobservations (Cashman et al. 1999; Hon et al. 2003), rheolog-ical data (Sehlke et al. 2014), analog measurements (Soule andCashman 2005) and melting/crystallization experiments
(Sato 1995; Hoover et al. 2001) on basaltic melts have docu-mented that the aa morphology is associated with a crystallinity>35 vol%. While the influence of both strain rate and groundslope on the development of pahoehoe surface crusts has beenthe focus of several studies on Etnean lavas (Chester et al. 1985;Hughes et al. 1990; Kilburn and Guest 1993; Guest et al. 2012),little or no attention has been given to the crystallization behav-iour of cicirara magmas before eruption to the surface.Recently, it has been argued that the majority of phenocrystsand microphenocrysts found in pahoehoe cicirara lavas did notequilibrate at depth with volatile-bearing magmas, but rather, asignificant portion of these minerals (mainly their mantle andrim portions; Fig. 1c) crystallize in the shallowest part of theconduit or even during the flow emplacement upon the effect ofwater exsolution and degassing (Lanzafame et al. 2013). Inorder to verify this hypothesis, we have conducted phase equi-librium experiments using a primitive trachybasaltic ciciraralava as starting composition (Fig. 1d, e). Crystallization exper-iments were equilibrated at relative high pressure using H2Oand CO2 concentrations corresponding to the crystallizationconditions of water-undersaturated magmas residing in thedeeper parts of the plumbing system (e.g. Spilliaert et al.2006; Collins et al. 2009). Results have been then comparedwith the natural bulk rock and mineral compositions ofpahoehoe cicirara lavas as well as with data from low-pressure,water-saturated experiments of Metrich and Rutherford (1998)reproducing open-system degassing conditions at Mt. EtnaVolcano. The information gained through this comparison dem-onstrate, in terms of phase assemblage, crystal content, and
Fig. 1 Example of pahoehoecicirara lava flow at Mt. Etna (a)containing high phenocryst andmicrophenocryst contents(>40 vol%) mostly due to theoccurrence of centimetre-sizedplagioclases, locally namedcicirara as a result of their chick-pea-like appearance (b).Plagioclases show clearcore-to-rim textural variations asobserved in the optical photomi-crograph that was acquired undercross-polarized light (c).Pahoehoe cicirara lavas are char-acterized by SiO2 vs alkali (d) andSiO2 vs MgO concentrationscomparable to those of prehistoricand historical Etnean products (e)
major and trace element concentrations, that pahoehoe ciciraralavas originate from deep-seated, crystal-poor magmas whosecrystallization is driven by abundant degassing in the upper-most part of the plumbing system and during emplacement tothe surface.
Experiments and analyses
The selected starting material is the most primitivetrachybasaltic pahoehoe cicirara lava (49.99 wt% SiO2;5.51 wt% Na2O+K2O) ever sampled at Mt. Etna (i.e. theCICG1 sample studied by Lanzafame et al. 2013). This lavaflow is associated with lateral eruptions at 2000 m a.s.l. on theflank of the volcanic edifice. The rock sample used in this studywas collected at 1 m from the bottom of the lava flow. The rocktexture is holocrystalline, and clinopyroxene, olivine,titanomagnetite and plagioclase are common. The total amountof phenocrysts, microphenocrysts and microlites is about 33, 9and 58 vol%, respectively.We stress that, according to previousanalyses conducted by Lanzafame et al. (2013), samples col-lected at different heights of the inner portion of the lava flowunit show limited compositional and textural variations. Thestarting glass was produced by melting twice the powderedrock at 1600 °C and atmospheric pressure for 120 min. Theglass was analysed by scanning electron microscopy, and nocrystalline phases were detected. Twenty electron microprobeanalyses of this glass yielded the following averagetrachybasaltic composition (in wt%): SiO2=49.93 (±0.41),TiO2=1.58 (±0.03), Al2O3=17.68 (±0.25), FeO=9.27(±0.18), MnO=0.18 (±0.02), MgO=5.18 (±0.16), CaO=10.25 (±0.21), Na2O=3.64 (±0.08), K2O=1.75 (±0.05) andP2O5=0.54 (±0.03) where the value in parenthesis is the 2sigma error. Phase equilibrium crystallization experiments wereperformed at 400 MPa in an internally heated pressure vessel(IHPV) using Ar as pressure medium at the Institute forMineralogy, Leibniz University of Hannover (Germany). Thestarting materials, composed of the powdered glass, deionisedwater and silver oxalate (Ag2C2O4), were loaded in Au80Pd20capsules. The proportions of volatile components were 1 and3 wt% H2O, and 0, 0.1 and 0.3 wt% CO2 (Table 1; details oncapsules preparation and relative techniques refer to Vetereet al. 2011, 2014). According to melt inclusion data trackingthe evolutionary behaviour of volatiles at Mt. Etna (Spilliaertet al. 2006; Collins et al. 2009), our experimental conditionsreproduced the H2O+CO2 concentrations commonly observedfor water-undersaturated magmas residing at depth (Corsaroand Pompilio 2004), when volatiles are dissolved into the meltphase. The charges were heated directly to the target tempera-tures of 1150 and 1100 °C for 48 h (Table 1). Experiments werequenched using a rapid-quench device leading to a cooling rateof approximately 150 °C/s. The run products were mounted inepoxy and then a polished thin section was produced from the T
epoxy block. All experiments were performed at the intrinsichydrogen fugacity of the vessel. Under intrinsic conditions inthe IHPV, the oxygen fugacity in capsules containing water-saturated melts (pure H2O fluid) is close to the NNO+3.7 buff-er (i.e. nickel-nickel oxide; Berndt et al. 2002). However, underwater-undersaturated conditions and with increasing CO2 con-tents, the oxygen fugacity of the system decreases with decreas-ing of water activity (e.g. Botcharnikov et al. 2008).Consequently, the oxygen fugacity within the charges was es-timated at the end of experimental runs through the oxygenbarometers of France et al. (2010) and Ishibashi (2013) basedon the plagioclase–clinopyroxene–melt and spinel–melt equi-libria, respectively (Table 1). These models were calibratedusing alkaline datasets and provided estimates comparable tothe redox state of Etnean magmas (i.e. from NNO to NNO+1.5; Armienti et al. 1994a, b; Collins et al. 2009; Giacomoniet al. 2014; Mollo et al. 2011a, 2015).
Images of run products and major element concentrationsof crystals and glasses were obtained at the HP-HT Laboratoryof Experimental Volcanology and Geophysics of the IstitutoNazionale di Geofisica e Vulcanologia in Roma (Italy).Images were collected at the micrometre scale as back-scattered micrographs (magnifications of×500–1000) usinga Jeol FE-SEM 6500 F equipped with an energy dispersionmicroanalysis system. Microprobe analyses were performedwith a Jeol-JXA8200 equipped with five spectrometers(Table 2). For glasses, a slightly defocused electron beamwitha size of 3 μm was used with a counting time of 5 s onbackground and 15 s on peak. For crystals, the beam sizewas 1 μm with a counting time of 20 and 10 s on peaks andbackground, respectively. The following standards have beenadopted for the various chemical elements: jadeite (Si andNa), corundum (Al), forsterite (Mg), andradite (Fe), rutile(Ti), orthoclase (K), barite (Ba), apatite (P), spessartine (Mn)and chromite (Cr). Sodium and potassium were analysed firstto prevent alkali migration effects. The precision of the micro-probe was measured through the analysis of well-characterized synthetic standards. Data quality was ensuredanalysing these standards as unknowns. Based on countingstatistics, analytical uncertainties relative to their reported con-centrations indicate that precision and accuracy were betterthan 5 % for all cations (Iezzi et al. 2008, 2011).
Trace element analyses of glasses were conducted at theCNR-IGG–Pavia with a laser ablation inductively coupledplasma mass spectrometry (LA-ICP-MS). The chosen traceelements are representative of rare earth elements (REE) com-prising heavy rare earth elements (HREE) and light rare earthelements (LREE). The laser source used for the analyses con-sists of a Q-switched Nd:YAG laser (Brilliant, Quantel), witha fundamental emission in the near-IR region (1064 nm),which is converted into 266 nm by two harmonic generators.Using mirrors, the laser beam is carried into a petrographicmicroscope, focused above the sample and then projected
onto it. Optimum average instrumental operating conditionsare as follows: RF power 800–900 W, cooling gas 12.08 l/min, sample gas 0.9–1.1 l/min, auxiliary gas 1.00 l/min andcarrier gas 0.9–1.1 l/min. The total scan time is about 700 ms,the settling time is about 340 ms, and hence the acquisitionefficiency is estimated at about 50 %. A typical analysis con-sists of acquiring 1 min of background and 1 min of ablatedsample using 10-μm spot size; thus, approximately 170sweeps are required. The mean integrated time for acquisitionis about 0.9 s for each element. The ablated material wasanalysed with a single-collector double-focusing sector fieldICP–MS (Element, Finnigan Mat, Bremen, Germany). NIST-SRM610 was used as an external standard, whereas 43Ca wereadopted as internal standard. In each analytical run, the USGSreference sample BCR2 was analysed together with the un-knowns for quality control. The precision of individual anal-yses varied depending upon a number of factors, e.g. the ele-ment and isotope analysed as well as the homogeneity of thesample. However, the 1 sigma errors calculated from varia-tions in replicate analyses of crystals and glasses were invari-ably several times larger than the fully integrated 1 sigmaerrors determined from counting statistics alone (Table 2).
Results
Mineralogical features
In Fig. 2, back-scattered images of our experimental productsshow the most important textural changes as a function oftemperature and volatile concentration. Modal phase propor-tions (wt%) were derived by mass balance calculations(Stormer and Nicholls 1978) yielding acceptable residualsum of squares (Σr2<0.25; see Table 1). For a better compar-ison of these estimates with vol% data from natural studiesdealing with lava flows, the volume proportions wererecalculated using average densities of clinopyroxene (3.4 g/cm3), plagioclase (2.6 g/cm3), titanomagnetite (5.1 g/cm3) andmelt (2.6–2.7 g/cm3 calculated for each run through the modelof Lange and Carmichael 1987). As expected, both wt% andvol% estimates show virtually identical values with differ-ences not exceeding 2 % (Table 1); notably, for the case ofnatural cicirara samples, the same conclusion can be reachedby recalculating vol% from crystal size distribution analysisinto wt%.
The phase assemblage is characterized by the ubiquitousoccurrence of titanomagnetite and glass, whereas plagioclaseand clinopyroxene crystallize only at lower temperatures and/or volatile contents. Charges do not contain a free fluid phase,according to H2O and CO2 solubility data (e.g. Collins et al.2009). Both the liquidus temperatures of the melt and thecrystal content decrease with decreasing water activity (cf.Ghiorso and Sack 1995). At 1150 °C and 3 wt% H2O,
titanomagnetite is the liquidus phase. At 1150 °C and 1 wt%H2O, plagioclase and titanomagnetite cosaturate the melt.Plagioclase is the most abundant mineral phase but disappearscompletely when CO2 is added to the charge. At 1100 °C and3 wt% H2O, clinopyroxene and titanomagnetite coprecipitatefrom the melt showing lower crystal contents with increasingCO2. At 1100 °C and 1 wt% H2O, the melt experiences abun-dant crystallization of plagioclase, clinopyroxene andtitanomagnetite, suggesting that the crystal content of ciciraramagmas abruptly increases up to 38 vol% in a restricted ther-mal range of only 50 °C.
Chemical features
Major element concentrations of minerals and glasses are re-ported in Table 2. Due to the low crystal content of experi-ments conducted at 1150 °C, the compositions of glasses donot show substantial changes with respect to the startingtrachybasaltic melt (Fig. 3a). At 1100 °C and 3 wt% H2O,the glass exhibits variable degrees of differentiation as a
function of CO2, and its composition ranges betweentrachybasalt and basaltic trachyandesite. At 1100 °C and1 wt%H2O, the glass is basaltic trachyandesite in compositionshowing silica and alkali enrichments with increasing CO2. Interms of REE concentrations (see Table 2), all glasses arecharacterized by preferential enrichments for HREE relativeto LREE, resembling typical equilibrium Etnean patterns(Scarlato et al. 2014). However, the amount of REE progres-sively increases with increasing crystal content of experimen-tal charges. At 1150 °C and 1 wt% H2O, plagioclase is signif-icantly enriched in CaO with respect to K2O (Fig. 3b), where-as lower CaO and K2O contents are measured in plagioclasecrystals formed at 1100 °C with decreasing fO2 (cf. Molloet al. 2011b). Clinopyroxene crystallizes only at 1100 °Cshowing compositions enriched in MgO and depleted inAl2O3 with decreasing H2O and/or with increasing CO2
(Fig. 3c). The amount of TiO2 in titanomagnetite decreasesas a function of temperature and/or H2O, counterbalanced byincreasing concentrations of FeOtot (Fig. 3d). Conversely,TiO2 is preferentially incorporated into titanomagnetite crystal
Fig. 2 BSE images showingtextural features of our phaseequilibria experiments. Note thevariable proportions ofclinopyroxene (cpx), plagioclase(plg), titanomagnetite (timt) andglass (gl) as a function oftemperature and H2O+CO2
lattice as the concentration of CO2 increases, responding to thelower buffering condition of the system (cf. Mollo et al.2013b; Ishibashi 2013).
Achievement of equilibrium
According to previous experiments conducted by Molloet al. (2013b) on a similar trachybasaltic composition,the time duration of 48 h should be enough to guaran-tee steady state levels and the supply of equilibriumcation proportions to the growing crystals. To verify thiscondition, we have checked for equilibrium composi-tions in plagioclase, clinopyroxene and titanomagnetite,and results from calculations are reported in Table 1.The equilibrium crystallization of plagioclase has beentested through the Ab–An (albite–anorthite) exchangereaction proposed by Putirka (2008). According to theauthor, the equilibrium constant is constrained within twotemperature-dependent intervals of plg–meltKdAb–An=0.10±0.05 at T<1050 °C and plg–meltKdAb–An=0.27±0.11 at T>1050 °C. Our estimates show that plg–meltKdAb–An yieldsvalues between 0.29 and 0.32, suggesting equilibrium crystal-lization for experimental plagioclase crystals formed at T>1050 °C.
For clinopyroxene, we used the temperature-sensitivecpx-meltKdFe–Mg model derived by Putirka (2008) andbased on deviations in observed and calculated Fe–Mgcation partitioning between crystals and coexistingmelts. cpx-meltKdFe–Mg values of clinopyroxene-melt pairsfrom this study are comprised between 0.26 and 0.27,in agreement with the equilibrium range of 0.27±0.03documented in literature (e.g. Putirka et al. 2003). As afurther test for clinopyroxene, we have also adopted themodel of Mollo et al. (2013c) calibrated through thedifference (Δ) between diopside+hedenbergite (DiHd)components predicted for clinopyroxene via regressionanalysis of clinopyroxene–melt pairs in equilibrium con-ditions and those measured in the analysed crystals.ΔDiHd values derived for our experimental composi-tions are remarkably low (0.01–0.06) and comparableto those expected at equilibrium, where ΔDiHd should beclose to zero.
In the case of titanomagnetite, the achievement ofequilibrium can be tested using the Ti–Fe cation redis-tribution found by Mollo et al. (2013b). When both Tiand Fe are supplied at the equilibrium proportion,titanomagnetites growing from Etnean trachybasaltic magmasshow relative high TiO2/FeOtot ratios of 0.25–0.30. Over the
Fig. 3 Compositional variationsof residual melt (a), plagioclase(b), clinopyroxene (c) andtitanomagnetite (d) from ourphase equilibria experiments as afunction of temperature andH2O+CO2 concentrations
effect of disequilibrium growth conditions, Ti is lessincorporated into the crystal lattice and counterbalancedby increasing Fe concentrations. Due to this disequilibriumpartitioning, the TiO2/FeOtot ratio in titanomagnetite de-creases down to 0.05. Our experimental titanomagnetitecompositions exhibit TiO2/FeOtot ratios between 0.27and 0.30 corresponding to values derived for equilibriumgrowth conditions.
Discussion
Textural maturation
Figure 4 shows that our experimental charges reproduce sev-eral textural features observed in natural pahoehoe ciciraralavas. Generally, larger and near-regular plagioclases formby attachment of smaller and isolated crystals (Fig. 4a)
Fig. 4 Comparison between ourexperimental textures and naturaltextures of pahoehoe cicirara lavaflows showing crystal attachmentof plagioclase (a) from the run at1150 °C, 1 wt% H2O and 0 wt%CO2, sieve-textured plagioclasedue to entrapment of melt inclu-sions (b) from the run at 1100 °C,1 wt% H2O and 0 wt% CO2, andcrystal growth by agglomerationof clinopyroxene (c) andtitanomagnetite (d) from the runat 1100 °C, 3 wt% H2O and0 wt% CO2
Bull Volcanol (2015) 77:44 Page 11 of 18 44
paralleling the crystallization mechanism documented by re-cent studies dealing with plagioclase nucleation and growth(Mollo et al. 2011b; Iezzi et al. 2008, 2011, 2014). Due to theequilibrium condition imposed by our experiments, the attach-ment of crystals proceeds along the longest crystallographicaxis in order to minimize the interfacial energy and structuralmismatches (Kostov and Kostov 1999; Deer et al. 2001). Thistextural maturation observed in laboratory agrees with that ofnatural pahoehoe cicirara samples (cf. Lanzafame et al. 2013),indicating that the growth of plagioclase phenocrysts (longestsize dimensions >0.3 mm) is driven by the attachment ofearly-formed microphenocrysts and microlites. Moreover,our experimental textures highlight that the melt next to theboundary of small plagioclase crystals is frequently entrappedduring attachment, testifying to incomplete crystal growth(Fig. 4b). Melt inclusions may or may not crystallize tinyclinopyroxenes and titanomagnetites, and in the process, thefinal habit of plagioclase becomes sieve-textured as for thecase of large phenocrysts erupted at Mt. Etna Volcano(Giacomoni et al. 2014). According to the maximum growthrate (xlsGmax) measured for (i) natural samples from basalticlava flows at Mauna Loa and Kilauea volcanoes (Cashmanand Marsh 1988; Crisp et al. 1994) and (ii) experimen-tal samples obtained for both alkaline and calc-alkalinecompositions (Lasaga 1997; Burkhard 2002; Baker et al.2008; Pupier et al. 2008; Mollo et al. 2011b, 2012), wecan assume that plgGmax (10−8 cm/s)>cpxGmax (10−9 cm/s)≅ timtGmax (10−9 cm/s). The final euhedral habits oflarger clinopyroxenes and titanomagnetites result from the ag-glomeration of smaller touching crystals. This mutual crystalgrowth process can be explained by the almost identical growthrates calculated for clinopyroxene and titanomagnetite.Moreover, in pahoehoe cicirara lava samples, crystal inter-growth proceeds from micrometre to millimetre scale formingglomerocrysts that appear as dense mosaics of aggregatedclinopyroxenes and titanomagnetites (Lanzafame et al.2013). Notably, crystal growth by agglomeration has beendocumented by previous ex-situ (Pupier et al. 2008) and insitu (Schiavi et al. 2009) experiments and may also producelarge, unzoned and chemically homogeneous phenocrysts(Iezzi et al. 2011, 2014).
Magma crystallization and differentiation
In Fig. 5a, the bulk rock compositions of natural pahoehoecicirara lavas (see the Excel spreadsheet provided assupplementary material) have been compared with melts fromwater-undersaturated experiments presented in this study andfrom water-saturated experiments of Metrich and Rutherford(1998). This latter experimental dataset was obtained using aprimitive Etnean trachybasalt (47.92 wt% SiO2; 4.15 wt%Na2O+K2O) equilibrated at temperatures, pressures, melt-water contents and buffering conditions of 1009–1135 °C,
27–120 MPa (PH2O=Ptotal), 1.3–3.1 wt% H2O and G-CH-QFM-NNO (i.e. graphite-methane, quartz–fayalite–magne-tite, and nickel–nickel oxide, respectively). Despite the factthat the starting compositions used for water-undersaturatedand water-saturated experiments are more (Mg#50) and less(Mg#63) differentiated, respectively, the experimentally de-rived melts lie on the same evolutionary trend. In other words,the most differentiated Mg#45–55 trachybasalts can be derivedby the early crystallization of olivine and clinopyroxene fromthe Mg#58–68 magmas (Armienti et al. 2007; Mollo et al.2015). Consequently, the comparison of water-saturated andwater-undersaturated data allows to constrain the whole crys-tallization history of the Etnean plumbing system. In terms ofphase assemblage, the most important difference betweenwater-undersaturated and water-saturated experiments is that,at PH2O=Ptotal, olivine is the liquidus phase followed byclinopyroxene, plagioclase and titanomagnetite (note that thislatter mineral forms only over the effect of abundant H2Oloss). Due to the lack of olivine crystallization in our experi-mental charges, MgO decreases weakly (4.7–5.4 wt%) duringmelt differentiation, whereas SiO2 increases remarkably(49.9–54.6 wt%) as shown in Fig. 5a. Thus, residual meltsshow MgO-rich compositions that do not correspond withthose of pahoehoe cicirara lavas. Through hygrometric esti-mates conducted on large cicirara plagioclases, Lanzafameet al. (2013) found that these phenocrysts formed in equilibri-um with ~1.5 wt% H2O. Thermodynamic simulations con-ducted by the same authors at 400 MPa, 1.5 wt% H2O, andQFM or NNO+1.5 buffer confirm that olivine does not crys-tallize from the cicirara magma. Conversely, water-saturatedexperiments of Metrich and Rutherford (1998) faithfully cap-ture most of the evolutionary trend of cicirara magmas ac-counting for the early formation of olivine (1–5 vol%). In lightof this, we have performed mass balance calculations bysubtracting to our water-undersaturated melts 1–7 vol% of atypical olivine phenocryst (75 mol.% Fo, i.e. forsterite) frompahoehoe cicirara lavas (see the Excel spreadsheet provided assupplementary material). Results show that the stability ofolivine is crucial to successfully reproduce the differentiationof pahoehoe cicirara magmas. For the case of cal-alkalinebasalts, it was found that, as PH2O increases, the piercingpoints in the forsterite–diopside–anorthite system shifts to-wards the plagioclase pole, expanding both the olivine andclinopyroxene primary phase abundances (Sisson and Grove1993). In trachybasaltic systems, the stability field of olivineincreases with increasing PH2O especially at the expense ofclinopyroxene (Metrich and Rutherford 1998). Such a featureis also consistent with experiments performed on Kilauea ba-salts where olivine is the liquidus phase at 100 MPa and1100 °C and is joined by clinopyroxene 30 °C lower(Spulber and Rutherford 1983). The lack of olivine in ourwater-undersaturated charges is clearly in contrast with thepetrography of pahoehoe cicirara lavas where olivine
crystallizes in the main phase assemblage and in the ground-mass. Olivine is also a FeO-bearing mineral, and consequent-ly, its absence in our water-undersaturated experiments can beaddressed to the early formation of titanomagnetite as al iquidus phase. However, in Etnean lava f lows,titanomagnetite generally occurs at the end of the crystalliza-tion sequence where crystallization is driven by magma de-compression and volatile exsolution (Métrich et al. 2004;Corsaro et al. 2013; Mollo et al. 2013b). Considering thatour experimental conditions closely match with those derivedfor the early crystallization of natural cicirara magmas(Lanzafame et al. 2013), the lack of olivine in the experimen-tal charges can be explained taking into account two possiblescenarios: (1) the crystallization of olivine occurs at low pres-sure when the exsolution of water enlarges the olivine stabilityfield (Sisson and Grove 1993), and (2) olivine phenocrystscrystallize from primitive trachybasaltic liquids that, at shal-low crustal levels, mix with more differentiated magmas.Although the input of deep-seated primitive magmas intoshallow reservoirs has been documented at Mt. Etna by anumber of studies (Corsaro et al. 2009, 2013; Armienti et al.2007, 2013; Kahl et al. 2013; Mollo et al. 2015), this second
hypothesis is excluded by the compositions of olivinephenocrysts and microphenocrysts found in cicirara lavaflows. In fact, using equilibrium and thermobarometricmodels, Mollo et al. (2015) have demonstrated that primitiveMg#58–68 trachybasaltic liquids equilibrate with Fo81–88 oliv-ines at 650–1050 MPa and 1150–1250 °C, whereas Fo70–80olivines saturate the more differentiated Mg47–57 magmas at0.1–450 MPa and 1100–1150 °C. The most primitive ciciraratrachybasalt object of this study shows Mg#51 compositionthat is in equilibrium with the Fo73–77 olivine phenocrystsand microphenocrysts hosted in the lava flow (seeLanzafame et al. 2013 for the natural compositions). Thisimplies that the crystallization of olivine cannot be addressedto the input of mafic magmas into the shallower portions of theplumbing system. Conversely, it is important to note that thelow-pressure, water-saturated experiments of Metrich andRutherford (1998) successfully crystallize Fo77 olivines inequilibrium with Mg#51 trachybasaltic melts (see also the dis-cussion below). On the other hand, the nucleation and growthof phenocrysts certainly start at high-temperature and high-pressure conditions when volatiles are dissolved in magma(Armienti et al. 2013; Mollo et al. 2015). This is also
Fig. 5 SiO2 vs. MgO diagram inwhich melts from our water-undersaturated experiments iscompared with cicirara magmasas well as with melts fromwater-saturated experiments ofMetrich and Rutherford (1998)(a). Mass balance calculationsshow that our melts can reproducethe evolutionary behaviour ofcicirara magmas only bysubtracting olivine. La vs. Dy di-agram in which melts from ourwater-undersaturated experimentsis compared with cicirara magmas(b). Results from fractional crys-tallization modelling indicate thatnatural magmas experience max-imum crystallization of 22 vol%before eruption
supported by the good correspondence between our experi-mental textures and those observed for natural cicirara products,giving reason for melt entrapments in sieve-textured plagio-clases, and clinopyroxene and titanomagnetite intergrowths(Fig. 4). However, thermodynamic and thermobarometric cal-culations have suggested that only small amounts (≤15 vol%)of plagioclase and clinopyroxene phenocrysts equilibrate atdepth with the host cicirara magma (Lanzafame et al. 2013).The crystal content increases (≤25 vol%) before eruption whenmagma rapidly rises from the deeper parts of the plumbingsystem to the vent. Thus, magma decompression, water exso-lution and heat dissipation induce abundant crystallization dur-ing magma ascent in the conduit and lava flowage onto thesurface (Lanzafame et al. 2013).
A similar conclusion can be reached through the compari-son between REE concentrations of experimental melts fromthis study and those of pahoehoe ciciraramagmas (Fig. 5b). Athigh degrees of crystallization (≥35 vol%), our experimentsindicate that the trace element signature of the original magmaabruptly increases towards levels never reached by the naturalproducts. Conversely, at relative low degrees of crystallization(≤6 vol%), our experimental melts reproduce about 35 % ofpahoehoe cicirara compositions. During magma differentia-tion, REE are highly incompatible in olivine (ol-meltKdREE<0.0018), moderately incompatible in clinopyroxene (cpx-meltKdREE<0.38) and plagioclase (plg-meltKdREE<0.74), andweakly compatible in titanomagnetite (timt-meltKdREE<1.3).Thus, increasing REE concentrations generally reflect higherdegrees of crystallization and/or magma evolution. On thisbasis, we have used the Excel spreadsheet of Ersoy andHelvaci (2010) to constrain the evolutionary behaviour of Lavs. Dy in pahoehoe cicirara lavas (see the Excel spreadsheetprovided as supplementary material). A fractional crystalliza-tion process has been modelled using the internal set of parti-tion coefficients for basaltic compositions and consideringolivine, clinopyroxene, plagioclase and titanomagnetite in2:5:7:1 relative proportions. Results indicate that natural
REE concentrations are reproduced by the removal of a solidphase content ≤22 vol% that, importantly, corresponds to themaximum crystallization (~25 vol%) of cicirara magmas esti-mated before eruption by Lanzafame et al. (2013).
In Fig. 6, the experimental temperature and the crystal con-tent of our charges have been compared with those obtainedby Metrich and Rutherford (1998). Data from this study showthat, at 1150 °C, the degree of crystallization is low(≤5.4 vol%) irrespective of the amount of volatiles added tothe charges. However, at 1100 °C, the crystal content abruptlyincreases up to 38 vol% as the amount of H2O decreases from3 to 1 wt%. Thermal conditions <1100 °C more likely agreewith direct measurements of inner lava flow temperatures(1050–1080 °C) conducted at the eruptive vents (e.g.Tanguy and Crocchiati 1984). Additionally, both melt inclu-sion analyses (Spilliaert et al. 2006) and hygrometric estimates(Armienti et al. 2013; Mollo et al. 2015) on Etnean productssuggest that H2O exsolution starts at 400 MPa and the amountof water dissolved in magmas decreases from 4.5 to 1 wt%with decreasing pressure. Thus, following the crystallizationpath depicted by our experiments, it should be expected thatthe phenocryst and microphenocryst content of deep-seatedcicirara magmas drastically increases up to ~40 vol% uponand immediately after eruption to the surface (e.g. Crispet al. 1994). On one hand, if crystals remain in equilibriumwith magmas and are transported to the surface, the eruptedlavas typically would not develop pahoehoe surface crusts dueto their very high crystal contents (see Fig. 6). On the otherhand, if crystals are separated from the magmatic suspensionbefore eruption (i.e. Rayleigh fractionation), pahoehoecicirara lavas would achieve unlikely REE enrichments thatare never observed in nature (see Fig. 5b). We note that, atPH2O=80 MPa and within the eruptive temperature range ofEtnean lavas (1050–1080 °C), water-saturated charges ofMetrich and Rutherford (1998) are characterized by low de-grees of crystallization (12–20 vol%), in agreement with therheological conditions to form pahoehoe morphologies. The
Fig. 6 Temperature vs. crystalcontent diagram in which datafrom our water-undersaturatedexperiments is compared withthose reported by Metrich andRutherford (1998) at water-saturated conditions. We also re-port direct measurements of innerlava flow temperatures at Mt.Etna (1050–1080 °C)and the phenocryst content(0–26 vol%) of lavas showingpahoehoe morphologies
effect of water on magma is to depress the crystallizationtemperature of plagioclase and clinopyroxene reducing theirproportions in magmas (Spulber and Rutherford 1983; Sissonand Grove 1993). According to the above discussion, thecrystal size distribution analysis and volatile solubility model-ling of pahoehoe cicirara magmas suggest (1) weak crystalli-zation at depth uponwater-undersaturated conditions, (2) low-to-moderate crystallization driven by water exsolution in theconduit, (3) moderate-to-high crystallization at the vent due tothe loss of water still retained in the melt and (4) abundantmicrolite crystallization imposed by heat dissipation duringflowage (Lanzafame et al. 2013).
Mineral chemistry, water exsolution and degassing
In Fig. 7, the composi t ions of phenocrysts andmicrophenocrysts found in pahoehoe cicirara lavas have beencompared with those presented in this study and reported byMetrich and Rutherford (1998). As described above, olivinedoes not crystallize in our water-undersaturated experiments.However, it is interesting to observe that natural olivine crystalsshow Fo contents comprised between 73 and 77 (Fig. 7a), cor-responding to crystallization temperatures of 1045–1100 °C forthe water-saturated experiments of Metrich and Rutherford(1998). The same conclusion can be reached by comparingthe Di and Hd contents of natural clinopyroxenes with thoseobtained by the same authors (Fig. 7b). In contrast,clinopyroxenes from our water-undersaturated experiments ex-hibit low Di and Hd concentrations due to the early crystalliza-tion of titanomagnetite and plagioclase causing FeO and CaOdepletions in the melt. Intriguingly, the An content of plagio-clase phenocrysts and microphenocrysts widely ranges between55 and 77 (Fig. 7c) and is reproduced in full only when the state
of the system changes from water-undersaturated to water-saturated regimes. At Mt. Etna Volcano, the Ab–An exchangein plagioclase is generally ascribed to the continuous replenish-ment of volatile-rich magmas and/or H2O–CO2 fluxing fromdepth (Armienti et al. 2004; Spilliaert et al. 2006; Kamenetskyet al. 2007; Collins et al. 2009; Ferlito et al. 2014). A high wateractivity at very shallow crustal levels favours An enrichments inplagioclase (Moore and Carmichael 1998; Brugger andHammer 2010), and consequently, crystallization experimentsof Metrich and Rutherford (1998) successfully reproduce thecompositions of An-rich phenocrysts and microphenocrystsfound in natural products. The role played by magmaundercooling at the late stage of crystallization is mostly record-ed by the relatively low ulvospinel (Usp) contents oftitanomagnetites from pahoehoe cicirara lavas with respect tothose measured in our phase equilibria experiments (Fig. 7d).Using the geospeedometer of Mollo et al. (2013b) based onintra-crystal cation exchanges in rapidly growingtitanomagnetites, we found that the cooling path of pahoehoecicirara lavas is comprised between 2 and 5 °C/min, resemblingvalues measured for both lava flows and dikes at Mt. EtnaVolcano (Mollo et al. 2013b, c) and other volcanic settings(e.g. Tenerife, Kenya and Hawaii; Crisp and Baloga 1990;Wilding et al. 1996; Cashman et al. 1999; Gottsmann et al.2004; Harris and Rowland 2009).
Arrays of P-T estimates for Etnean eruptions identifydifferent ascent velocities of magmas throughout the en-tire length of the vertically developed feeding system(Fig. 8). Deep-level (>6 km) magma ascent velocitiesare relatively low with values between 0.42 and 4.2×10−3 m/s (Armienti et al. 2013), whereas shallow-level(1–4 km) magma ascent velocities are very fast and com-prised between 0.04 and 0.4 m/s (Aloisi et al. 2006).
Fig. 7 The compositions ofolivine (a), clinopyroxene (b),plagioclase (d) andtitanomagnetite (e) from pahoehoecicirara lavas are comparedwith those from ourwater-undersaturated experimentsand water-saturated experimentsof Metrich and Rutherford (1998).Cooling rate conditions drivingtitanomagnetite crystallizationhave been estimated through thegeospeedometer of Mollo et al.(2013b). Fo forsterite,Di diopside,Hd hedenbergite, Or orthoclase,An anorthite, Usp ulvospinel
According to Armienti et al. (2013), when volatiles arereleased to a vapour phase retained in the magmatic sus-pension, the buoyancy of the whole magma increasesleading to an acceleration in the uppermost parts of theplumbing system. Coherently, H2O and CO2 melt inclu-sion data (cf. Métrich et al. 2004; Spilliaert et al. 2006;Collins et al. 2009) indicate that volatile-rich magmasresiding at shallow depths (~4 km) undergo strong vola-tile exsolution while travelling from the conduit onto thesurface (Fig. 8). With respect to the equilibrium growthrates of deep-seated crystals, degassing-induced crystalli-zation in the central conduit can increase the growth rateof plagioclase by 3–4 orders of magnitude over ascenttimes of 0.2–7 h (Applegarth et al. 2013). This changein the crystal growth regime provides further explanationfor the sieve-textures of plagioclase phenocrysts found inpahoehoe cicirara lavas, especially for the entrapment ofabundant melt inclusions into the crystal rims (Giacomoniet al. 2014). Additionally, magma degassing can produce20–30 vol% of crystallization (Applegarth et al. 2013), inclose correspondence with the amount of crystals mea-sured upon water-saturated conditions (Metrich andRutherford 1998). We can therefore conclude that if thecrystallization of pahoehoe cicirara magmas starts atdepth, the degree of crystallization should remain low(<15 vol%) so that the original magma composition doesnot change significantly in terms of major and trace ele-ment concentrations and thus lavas can be erupted aspahoehoe type. While this gas-rich magma rises to shal-low crustal levels, the water activity increases favouringthe early formation of olivine. During eruption to the sur-face, most of the crystal l izat ion of plagioclase,clinopyroxene and titanomagnetite is driven by degassing
in the uppermost part of the plumbing system and coolingat subaerial conditions.
Conclusions
Results from our equilibrium experiments show that, for thecase study of cicirara pahoehoe lavas, the early crystallizationof titanomagnetite as liquidus phase prevents the formation ofolivine in the experimental charges. As a consequence, theresulting melts are significantly enriched in MgO showingcompositions that do not correspond to the evolutionary be-haviour of natural cicirara magmas. Moreover, with respect tonear-liquidus crystal contents (2–5 vol%), a moderate temper-ature and water decrease cause the abundant crystallization ofplagioclase and clinopyroxene (33–37 vol%) leading to REElevels unlikely for natural cicirara products. The comparisonbetween our water-undersaturated experimental data andthose conducted under water-saturated conditions demon-strates that cicirara lavas originate from crystal-poor magmasthat, during emplacement at the surface, develop pahoehoesurface crusts upon the effect of degassing and cooling thatcontrol most of the crystallization of minerals.
Acknowledgments The authors are grateful to Tim McClinton andHiroaki Sato (as Reviewers) and Michael Manga (as Associate Editor)for their useful and constructive suggestions. We kindly thank A. Cavallofor assistance during electron microprobe analysis. The research activitiesof the HP-HT laboratory of the INGV were supported by the EuropeanObserving System Infrastructure Project (EPOS). F. Vetere would like toacknowledge the Marie Curie Fellowship 297880 SolVoM and D.Perugini the European Research Council for the Consolidator GrantERC-2013-CoG Proposal No. 612776 - CHRONOS.
Fig. 8 Schematic representation of magma dynamics driving thecrystallization of cicirara magmas at Mt. Etna Volcano. Thecrystallization proceeds from a first region at depth where high-temperature, volatile-bearing magmas rise with low ascent velocities toa second shallow region where low-temperature, degassing magmas are
rapidly erupted onto the surface. Ascent velocities have been estimated byAloisi et al. (2006) and Armienti et al. (2013), whereas H2O+CO2 con-tents come from melt inclusion analyses of Métrich et al. (2004)),Spilliaert et al. (2006) and Collins et al. (2009)
Aloisi M, Bonaccorso A, Gambino S (2006) Imaging composite dikepropagation (Etna, 2002 case). J Geophys Res Solid Earth 111:B06404. doi:10.1029/2005JB003908
Applegarth LJ, Tuffen H, James MR, Pinkerton H, Cashman KV (2013)Direct observations of degassing-induced crystallization in basalts.Geology 41:243–246
Armienti P, Pareschi MT, Pompilio M, Innocenti F (1994a) Effects ofmagma storage and ascent on the kinetics of crystal growth: the caseof the 1991-1993 Mt. Etna eruption. Contrib Mineral Petrol 115:402–414
Armienti P, Clocchiatti R, D'Orazio M, Innocenti F, Petrini R, PompilioM, Tonarini S, Villari L (1994b) The long-standing 1991–1993Mount Etna eruption: petrography and geochemistry of lavas. ActaVulcanol 4:15–28
Armienti P, Tonarini S, D’Orazio M, Innocenti F (2004) Genesis andevolution of Mount Etna alkaline lavas: petrological and Sr-Nd-Bisotope constraints. Period Mineral 73:29–52
Armienti P, Tonarini S, Innocenti F, D’Orazio M (2007) Mount Etnapyroxene as tracer of petrogenetic processes and dynamics of thefeeding system. In: Beccaluva L, Bianchini G, Wilson M (Eds.)Cenozoic Volcanism in the Mediterranean, Geological Society ofAmerica Special Papers 418: 265–276
Armienti P, Perinelli C, Putirka KD (2013) A new model to estimatedeep-level magma ascent rates, with applications to Mt. Etna(Sicily, Italy). J Petrol 54:795–813
Baker DR (2008) The fidelity of melt inclusions as records of melt com-position. Contrib Mineral Petrol 156(3):377–395
Berndt J, Liebske C, Holtz F, Freise M, Nowak M, Ziegenbein D,Hurkuck W, Koepke J (2002) A combined rapid-quench and H2-membrane setup for internally heated pressure vessels: descriptionand application for water solubility in basaltic melts. Am Mineral87:1717–1726
Botcharnikov RE, Almeev RR, Koepke J, Holtz F (2008) Phase relationsand liquid lines of descent in hydrous ferrobasalt— implications forthe Skaergaard intrusion and Columbia river flood basalts. J Petrol49:1687–1727
Branca S, Coltelli M, Groppelli G, Lentini F (2011) Geological map ofEtna volcano, 1:50000 scale. Ital J Geosci 130:265–291
Brugger CR, Hammer JE (2010) Crystallization kinetics in continuousdecompression experiments: implications for interpreting naturalmagma ascent processes. J Petrol 51:1941–1965
Burkhard DJM (2002) Kinetics of crystallization: example of micro-crystallization in basalt lava. Contrib Mineral Petrol 142:724–737
Cashman KV, Marsh BD (1988) Crystal size distribution (CSD) in rocksand the kinetics and dynamics of crystallization. Part II: Makaopuhilava lake. Contrib Mineral Petrol 99:292–305
Cashman KV, Thornber C, Kauahikaua JP (1999) Cooling and crystalli-zation of lava in open channels, and the transition of Pahoehoe Lavato Aa. Bull Volcanol 61:306–323
Chester DK, Duncan AN, Guest JE, Kilburn CRJ (1985) Mount Etna.The anatomy of a volcano. CambridgeUniversity Press, Cambridge,404 pp
Collins SJ, Pyle DM, Maclennan J (2009) Melt inclusions track pre-eruption storage and dehydratation of magmas at Etna. Geology 6:571–574
Corsaro RA, PompilioM (2004) Dynamics of magmas atMount Etna. In:Bonaccorso A, Calvari S, Coltelli M, Del Negro C, Falsaperla S(eds) Mt. Etna volcano laboratory, 143: AGU GeophysicalMonograph Series, Washington, 91–110
Corsaro RA, Miraglia L, Pompilio M (2007) Petrological evidence of acomplex plumbing system feeding the July-August 2001 eruption ofMt. Etna, Sicily, Italy. Bull Volcanol 69:401–421
Corsaro RA, Metrich N, Allard P, Andronico D, Miraglia L,Fourmentraux C (2009) The 1974 flank eruption of Mount Etna:an archetype for deep dike-fed eruptions at basaltic volcanoes anda milestones in Etna’s recent history. J Geophys Res 114, B07204
Corsaro RA, Di Renzo V, Distefano S, Miraglia L, Civetta L (2013)Relationship between petrologic processes in the plumbing systemof Mt. Etna and the dynamics of the eastern flank from 1995 to2005. J Volcanol Geotherm Res 251:75–89
Crisp J, Baloga S (1990) A model for lava flows with two thermal com-ponents. J Geophys Res 95:1255–1270. doi:10.1029/89JB01696
Crisp J, Cashman CK, Bonini JA, Hougen SB, Pieri DC (1994)Crystallization history of the 1984 Mauna Loa lava flow. JGeophys Res 99:7177–7198
Ferlito C, Coltorti M, Lanzafame G, Giacomoni PP (2014) The volatileflushing triggers eruptions at open conduit volcanoes: evidence fromMount Etna volcano (Italy). Lithos 184–187:447–455
Ferlito C, Viccaro M, Cristofolini R (2008) Volatile-induced magma dif-ferentiation in the plumbing system of Mt. Etna volcano (Italy):evidence from glass in tephra of the 2001 eruption. Bull Volcanol70:455–473
France L, Ildefonse B, Koepke J, Bech F (2010) A new method to esti-mate the oxidation state of basaltic series from microprobe analyses.J Volcanol Geotherm Res 189:340–346
Ghiorso MS, Sack RO (1995) Chemical mass-transfer in magmatic pro-cesses IV. A revised and internally consistent thermodynamic modelfor the interpolation and extrapolation of liquidus-solid equilibria inmagmatic systems at elevated temperatures and pressures. ContribMineral Petrol 119:197–212
Giacomoni PP, Ferlito C, Coltorti M, Bonadiman C, Lanzafame G (2014)Plagioclase as archive of magma ascent dynamics on BOpenConduit^ volcanoes: the 2001-2006 eruptive period at MountEtna. Earth Sci Rev 138:371–393
Gottsmann J, Harris AJL, Dingwell DB (2004) Thermal histrory ofHawaian pahoehoe lava crusts at the glass transition: implicationsfor flow rheology and emplacement. Earth Planet Sci Lett 228:343–353
Guest JE, Duncan AM, Stofan ER, Anderson SW (2012) Effect of slopeon the development of pahohehoe flow fields: evidence fromMountEtna. J Volcanol Geotherm Res 219–220:52–62
Harris AJL, Rowland SK (2009) Effusion rate controls on lava flowlength and the role of heat loss: a review. In: Thordarson T, Self S,Larsen S, et al (eds) Studies in volcanology: the legacy of GeorgeWalker. Special Publications of IAVCEI, 2, Geological Society,London, pp 33–51
Hon K, Gansecki C, Kauahikaua J (2003) The transition from aa topahoheoe crust on flows emplaced during the Pu’u O’o-Kupaianaha eruption. US Geol Surv Prof Pap 1676:89–104
Hoover SR, Cashman KV, Manga M (2001) The yield strength ofsubliquidus basalts: experimental results. J Volcanol Geotherm Res107:1–18
Hughes JW, Guest JE, Duncan AM (1990) Changing styles of effusiveeruption on Mount Etna since AD 1600. In: Ryan MP (ed) Magmatransport and storage. Wiley, London, pp 385–405
Iezzi G, Ventura G (2005) The kinematics of lava flows inferred from thestructural analysis of enclaves: a review. In: Manga M, Ventura G(eds) Kinematics and dynamics of lava flows. Geological Society ofAmerica, Boulder, pp 15–28
Iezzi G, Mollo S, Ventura G, Cavallo A, Romano C (2008) Experimentalsolidification of anhydrous latitic and trachytic melts at different
cooling rates: the role of nucleation kinetics. Chem Geol 253:91–101
Iezzi G, Mollo S, Torresi G, Ventura G, Cavallo A, Scarlato P (2011)Experimental solidification of an andesitic melt by cooling. ChemGeol 283:261–273
Iezzi G, Mollo S, Shaini E, Cavallo A, Scarlato P (2014) The coolingkinetics of plagioclase revealed by electron microprobe mapping.Am Mineral 99:898–907
Ishibashi H (2013) Spinel-melt oxygen barometry: a method and appli-cation to Cenozoic alkali basaltic magmas from the Higashi-Matsuura district, NW Kyushu, Japan. Geosci Rep Shizuoka Univ40:21–32 (in Japanese with English abstract)
Kahl M, Chakraborty S, Costa F, Pompilio M (2013) Dynamic plumbingsystem beneath volcanoes revealed by kinetic modeling, and theconnection to monitoring data: an example from Mt. Etna. EarthPlanet Sci Lett 308:11–22
Kamenetsky VS, Pompilio M, Métrich N, Sobolev AV, Kuzmin DV,Thomas R (2007) Arrival of extremely volatile-rich high-Mgmagmas changes explosivity of Mount Etna. Geology 35:255–258
Kilburn CRJ (2004) Fracturing as a quantitative indicator of lava flowdynamics. J Volcanol Geotherm Res 132:209–224
Kilburn CRJ, Guest JE (1993) Aa lavas ofMount Etna, Sicily. In: KilburnCRJ (ed) Active lavas. UCL Press, London, pp 73–106
Kostov I, Kostov RI (1999) Crystal habits of minerals. BulgarianAcademic Monographs, Sophia
Lange RA, Carmichael ISE (1987) Densities of Na2O-K2O-CaO-MgO-FeO-Fe2O3-Al2O3-TiO2-SiO2 liquids: new measurements and de-rived partial molar properties. Geochim Cosmochim Acta 51:2931–2946
Lanzafame G, Mollo S, Iezzi G, Ferlito C, Ventura G (2013) Unravelingthe solidification path of a pahoehoe Bcicirara^ lava from MountEtna volcano. Bull Volcanol 75:1–16
Lasaga AC (1997) Kinetic theory in the earth sciences. PrincetonUniversity Press, Princeton
Metrich N, Rutherford MJ (1998) Low pressure crystallization paths ofH2O-saturated basaltic-hawaitic melts from Mt Etna: implicationsfor open-system degassing of basaltic volcanoes. GeochimCosmochim Acta 62:1195–1205
Métrich N, Allard P, Spilliaert N, Andronico D, Burton M (2004) 2001flank eruption of the alkali- and volatile-rich primitive basalt respon-sible for Mount Etna’s evolution in the last three decades. EarthPlanet Sci Lett 228:1–17
Mollo S, Lanzafame G,Masotta M, Iezzi G, Ferlito C, Scarlato P (2011a)Cooling history of a dike as revealed by mineral chemistry: a casestudy from Mt. Etna volcano. Chem Geol 288:39–52
Mollo S, Putirka K, Iezzi G, Del Gaudio P, Scarlato P (2011b)Plagioclase-melt (dis)equilibrium due to cooling dynamics: implica-tions for thermometry, barometry and hygrometry. Lithos 125:221–235
Mollo S, Misiti V, Scarlato P, Soligo M (2012) The role of cooling rate inthe origin of high temperature phases at the chilled margin of mag-matic intrusions. Chem Geol. doi:10.1016/j.chemgeo.2012.05.029
Mollo S, Scarlato P, Lanzafame G, Ferlito C (2013a) Deciphering lavaflow post-eruption differentiation processes by means of geochem-ical and isotopic variations: a case study from Mt. Etna volcano.Lithos 162–163:115–127
Mollo S, Putirka K, Iezzi G, Scarlato P (2013b) The control of coolingrate on titanomagnetite composition: implications for ageospeedometry model applicable to alkaline rocks from Mt. Etnavolcano. Contrib Mineral Petrol 165:457–475
Mollo S, Putirka K, Misiti V, SoligoM, Scarlato P (2013c) A new test forequilibrium based on clinopyroxene-melt pairs: clues on the solidi-fication temperatures of Etnean alkaline melts at post-eruptive con-ditions. Chem Geol 352:92–100
Mollo S, Giacomoni PP, Coltorti M, Ferlito C, Iezzi G, Scarlato P (2015)Reconstruction of magmatic variables governing recent Etnean erup-tions: constraints from mineral chemistry and P-T-fO2-H2O model-ling. Lithos 212–215:311–320. doi:10.1016/j.lithos.2014.11.020
Moore GM, Carmichael ISE (1998) The hydrous phase equilibria(to 3 kbar) of an andesite and basaltic andesite from WesternMexico: constraints on water content and conditions of phenocrystgrowth. Contrib Mineral Petrol 130:304–319
Pupier E, Duchene S, Toplis MJ (2008) Experimental quantification ofplagioclase crystal size distribution during cooling of a basaltic liq-uid. Contrib Mineral Petrol 155:555–570
Putirka K (2008) Thermometers and barometers for volcanic systems. In:Putirka KD, Tepley F (Eds.) Minerals, inclusions and volcanic pro-cesses. Rev Mineral Geochem 69: 61–120
Putirka KD, Mikaelian H, Ryerson F, Shaw H (2003) New clinopyroxene–liquid thermobarometers for maric, evolved, and volatile-bearing lavacompositions, with applications to lavas from Tibet and the SnakeRiver Plain, Idaho. Am Mineral 88:1542–1554
Sato H (1995) Textural difference between pahoehoe and aa lavas of Izu-Oshima volcano, Japan: an experimental study on population den-sity of plagioclase. J Volcanol Geotherm Res 66:101–113
Scarlato P, Mollo S, Blundy JD, Iezzi G, Tiepolo M (2014) The role ofnatural solidification paths on REE partitioning betweenclinopyroxene and melt. Bull Volcanol 76:1–4
Schiavi F, Walte N, Konschak A, Keppler H (2009) First in situ observa-tion of crystallization processes in a basaltic-andesitic melt with themoissanite cell. Geology 37:963–966
SehlkeA,WhittingtonA, Robert B, Harris A, Gurioli L,Médard E (2014)Pahoehoe to `a`a transition of Hawaiian lavas: an experimentalstudy. Bull Volcanol 76:876
Sisson TW, Grove TL (1993) Experimental investigations of the role ofH2O in calcalkaline differentiation and subduction zonemagmatism.Contrib Mineral Petrol 113:143–166
Soule SA, CashmanKV (2005) Shear rate dependence of the pahoehoe to`a`a transition: analog experiments. Geology 33:361–364
Spilliaert N, Allard P, Metrich N, Sobolev AV (2006) Melt inclusionrecord of the conditions of ascent, degassing, and extrusion ofvolatile-rich alkali basalt feeding the powerful 2002 flank eruptionof Mount Etna (Italy). J Geophys Res 111, B04203
Spulber SD, Rutherford MJ (1983) The origin of rhyolite andplagiogranite in oceanic crust: an experimental study. J Petrol 24:1–25
Stormer JC, Nicholls J (1978) XLFRAC: a program for the interactivetesting of magmatic differentiation models. Comput Geosci 4:143–159
Tanguy JC, Crocchiati R (1984) The Etnean lavas, 1977-1983: petrologyand mineralogy. Bull Volcanol 47:879–894
Vetere F, Botcharnikov R, Behrens H, Holtz F, De Rosa R (2011)Solubility of H2O and CO2 in shoshonitic melts at 1250 °C andpressure from 50 to 400 MPa. J Volcanol Geotherm Res 202:251–261
Vetere F, Behrens H, Botcharnikov R, Holtz F, Fanara S (2014) The roleof alkalis in the solubility of H2O and CO2 in silicate melts.Implication for phonotephritic magmas. Contrib Mineral Petrol167:1014
Viccaro M, Giacomoni PP, Ferlito C, Cristofolini R (2010) Dynamics ofmagma supply at Mt. Etna volcano (Southern Italy) as revealed bytextural and compositional features of plagioclase phenocrysts.Lithos 116:77–91
Wilding M, Webb S, Dingwell DB (1996) Tektite cooling rates: calori-metric relaxation geospeedometry applied to a natural glass.Geochim Cosmochim Acta 60:1099–1103
