ABstrACt - Present-day sant’Antioco island (sW- sardinia, italy) hosted one of the southernmost, and youngest, subduction-related, Cenozoic magmatic events of western sardinia. A high-alumina basalt- andesite rock association, the focus of this paper, crops out in the southern portion of the island and represents the mafic end-member of volcanism developed during Miocene. Basaltic andesites and andesites are the dominant rock-types, while basalts and dacites are occasional. Minero-petrographic and geochemical characteristics reveal: i) a clear calc-alkaline signature for these rocks, in spite of high Feo*/Mgo ratios that mimic a tholeiitic affinity; ii) a magma evolution mainly controlled by fractional crystallization and iii) some degree of crustal assimilation. Phase relationships and compositions, mass balance calculations and t, P, X H 2 o estimates consistently indicate that fractional crystallization occurred at low pressure (P=100-400 MPa), under different P H 2 o conditions, which explain the observed wide shifts in major oxides (e.g., Al 2 o 3 and Mgo), for comparable sio 2 contents, and the ambiguous tholeiitic character as well. Phase relationships in the least evolved lithotypes point out that higher H 2 o concentrations (estimated at up to 6 wt%) in the melts reduced the crystallization of plagioclase and favoured that of olivine, whereas lower H 2 o contents (up to 3 wt%) promoted plagioclase fractionation and earlier crystallization of orthopyroxene at the expense of olivine. this X H 2 o imprint was recorded up to the most evolved compositions. Finally, the study rock suite can be derived from a parental (near-primary) magma matching in composition high-Mgo basalts from the Montresta calc-alkaline district in nW sardinia, in the light of the close compositional similarity between the high-Al 2 o 3 basaltic andesites in the two districts. thus, starting from a common parental composition, here we reconstruct liquid lines of descent, where the relative proportions of fractionating plagioclase vs mafic phases, as controlled by X H 2 o in the melts, may account for the compositional differences of sant’Antioco rock-types ab initio. riAssunto - l’isola di sant’Antioco (sardegna sud- occidentale) ospita una delle più recenti manifestazioni del magmatismo orogenico del margine occidentale della sardegna connesso alla subduzione cenozoica. oggetto del presente lavoro sono le vulcaniti affioranti nella porzione meridionale dell’isola, costituite essen- zialmente da andesiti basaltiche alte in allumina e da andesiti (più occasionali basalti e daciti), che rappre- sentano i termini meno differenziati del locale mag- matismo di età miocenica. oltre ad integrare l’ampio PerIodIco di MInerAlogIA established in 1930 Per. Mineral. (2010), 79, 1, 27-55 doi: 10.2451/2010PM0002 http://go.to/permin An International Journal of MINERALOGY, CRYSTALLOGRAPHY, GEOCHEMISTRY, ORE DEPOSITS, PETROLOGY, VOLCANOLOGY and applied topics on Environment, Archeometry and Cultural Heritage Petrogenesis of the High-Alumina Basalt-Andesite suite from Sant’Antioco Island, SW Sardinia, Italy AidA MAriA Conte 1, *, dAnilo MAuro P AllAdino 2 , CristinA Perinelli 3 and eMMAnuelle Argenti 4 1 C.n.r.- i.g.g.- u.o.s. di roma, dipartimento di scienze della terra, sapienza università di roma, P.le Aldo Moro 5, 00185 roma, italy 2 dipartimento di scienze della terra, sapienza università di roma, P.le Aldo Moro 5, 00185 roma, italy 3 dipartimento di scienze della terra, università di Pisa, Via santa Maria 53, 56126 Pisa, italy 4 Parco naturale regionale Bracciano Martignano, Via A. saffi, 4/A, Bracciano, 00062 roma, italy Submitted, November 2009 - Accepted, March 2010 * Corresponding author, E-mail: [email protected]
29
Embed
Petrogenesis of the High-Alumina Basalt-Andesite suite ... · from Sant’Antioco Island, SW Sardinia, Italy AidA MAriA Conte 1,*, d Anilo MAuro PAllAdino 2, ... di fase e i calcoli
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
ABstrACt - Present-day sant’Antioco island (sW-sardinia, italy) hosted one of the southernmost, andyoungest, subduction-related, Cenozoic magmaticevents of western sardinia. A high-alumina basalt-andesite rock association, the focus of this paper, cropsout in the southern portion of the island and representsthe mafic end-member of volcanism developed duringMiocene. Basaltic andesites and andesites are thedominant rock-types, while basalts and dacites areoccasional. Minero-petrographic and geochemicalcharacteristics reveal: i) a clear calc-alkaline signaturefor these rocks, in spite of high Feo*/Mgo ratios thatmimic a tholeiitic affinity; ii) a magma evolutionmainly controlled by fractional crystallization and iii)some degree of crustal assimilation. Phaserelationships and compositions, mass balancecalculations and t, P, XH2o estimates consistentlyindicate that fractional crystallization occurred at lowpressure (P=100-400 MPa), under different PH2o
conditions, which explain the observed wide shifts inmajor oxides (e.g., Al2o3 and Mgo), for comparablesio2 contents, and the ambiguous tholeiitic characteras well. Phase relationships in the least evolvedlithotypes point out that higher H2o concentrations(estimated at up to 6 wt%) in the melts reduced thecrystallization of plagioclase and favoured that of
olivine, whereas lower H2o contents (up to 3 wt%)promoted plagioclase fractionation and earliercrystallization of orthopyroxene at the expense ofolivine. this XH2o imprint was recorded up to the mostevolved compositions. Finally, the study rock suite canbe derived from a parental (near-primary) magmamatching in composition high-Mgo basalts from theMontresta calc-alkaline district in nW sardinia, in thelight of the close compositional similarity between thehigh-Al2o3 basaltic andesites in the two districts.thus, starting from a common parental composition,here we reconstruct liquid lines of descent, where therelative proportions of fractionating plagioclase vs
mafic phases, as controlled by XH2o in the melts, mayaccount for the compositional differences ofsant’Antioco rock-types ab initio.
riAssunto - l’isola di sant’Antioco (sardegna sud-occidentale) ospita una delle più recenti manifestazionidel magmatismo orogenico del margine occidentaledella sardegna connesso alla subduzione cenozoica.oggetto del presente lavoro sono le vulcaniti affiorantinella porzione meridionale dell’isola, costituite essen-zialmente da andesiti basaltiche alte in allumina e daandesiti (più occasionali basalti e daciti), che rappre-sentano i termini meno differenziati del locale mag-matismo di età miocenica. oltre ad integrare l’ampio
and applied topics on Environment, Archeometry and Cultural Heritage
Petrogenesis of the High-Alumina Basalt-Andesite suite
from Sant’Antioco Island, SW Sardinia, Italy
AidA MAriA Conte1,*, dAnilo MAuro PAllAdino2, CristinA Perinelli3 and eMMAnuelle Argenti4
1C.n.r.- i.g.g.- u.o.s. di roma, dipartimento di scienze della terra, sapienza università di roma, P.le Aldo Moro 5,00185 roma, italy
2dipartimento di scienze della terra, sapienza università di roma, P.le Aldo Moro 5, 00185 roma, italy3dipartimento di scienze della terra, università di Pisa, Via santa Maria 53, 56126 Pisa, italy
4Parco naturale regionale Bracciano Martignano, Via A. saffi, 4/A, Bracciano, 00062 roma, italy
28 A.M. Conte, d.M. PAllAdino, C. Perinelli and e. Argenti
panorama degli studi esistenti sul magmatismo orogenicosardo, la caratterizzazione minero-petrografica egeochimica e la parametrizzazione P, t, XH2o dei litotipiinvestigati, consente una ricostruzione dei relativiprocessi petrogenetici, anche alla luce dei recentisviluppi sperimentali sui sistemi calco-alcalini. i risultatiindicano che le vulcaniti in esame rivelano una chiaraaffinità calco-alcalina, nonostante gli alti rapportiFeo*/Mgo tipici di rocce di serie tholeiitica, e che ladifferenziazione magmatica è stata controllata princi-palmente dalla cristallizzazione frazionata, associata adifferenti gradi di assimilazione crostale. le relazionidi fase e i calcoli di bilancio di massa, in accordo conle stime di t e P di cristallizzazione dei fenocristalli edei contenuti di H2o in soluzione nei magmi, sug-geriscono una cristallizzazione a bassa pressione(P=100-400 MPa) con differenti XH2o. Proprio quest’ul-timo parametro spiega sia l’ampia variabilità compo-sizionale di queste rocce (in particolare i differenticontenuti in Al2o3 e Mgo per contenuti in sio2 simili),sia l’apparente carattere tholeiitico. le relazioni difase osservate nei litotipi meno evoluti indicano infattiche alte concentrazioni di H2o nei fusi (stimate intornoal 6% in peso) hanno ridotto l’incidenza del plagioclasio(che tuttavia rappresenta invariabilmente la fase domi -nante, seguita dal clinopirosseno), favorendo la cristal-lizzazione dell’olivina fra le fasi femiche. Viceversa,minori contenuti in H2o (fino a ~3%), oltre a favorireil frazionamento di plagioclasio, hanno comportato laprecoce cristallizzazione dell’ortopirosseno a spesedell’olivina. l’evidenza di diversi contenuti in H2oquale fattore di controllo determinante del processo didifferenziazione dei magmi persiste fino alle compo-sizioni più evolute dell’associazione in esame. infine,le litologie meno differenziate riscontrate a sant’Antiocopossono essere considerate come direttamente derivantida un magma più primitivo di composizione simile aquella dei basalti alti in Mgo affioranti nel distrettovulcanico di Montresta (sardegna nord-occidentale),anche sulla base delle significative analogie compo-sizionali fra le andesiti basaltiche alte in Al2o3 nelledue aree. l’insieme dei dati è compatibile con unmodello in cui, a partire da una comune composizioneparentale, le differenze riscontrate sin dai termini menoevoluti di sant’Antioco possono essere ricondotte apercorsi evolutivi caratterizzati da diversi rapporti trale fasi frazionate (in particolare fra il plagioclasio e lefasi femiche) sotto il controllo cruciale della XH2o.
Key Words: Cenozoic orogenic volcanism, high-
alumina basaltic-andesites, petrology and
geochemistry, Sant’Antioco, Sardinia.
introduCtion
Cenozoic, subduction-related, igneous activityin sardinia has been widely investigated (e.g.,dostal et al., 1982; Beccaluva et al., 1994;Brotzu et al., 1997a, b; Morra et al., 1994, 1997;Mattioli et al., 2000; downes et al., 2001;Franciosi et al., 2003; lustrino et al., 2004,2009). However, information on petrography,mineralogy and chemistry of one of thesouthernmost volcanic districts, sant’Antioco, isstill scarce and incomplete. up to now, the fewgeological and compositional data available onthe sant’Antioco volcanics were reported inluxoro (1987), in the 1:25,000 geo-petrographicmap of sant’Antioco island (Maccioni et al.,1990) and in lecca et al. (1997), who consideredthe andesitic s.l. rocks from sant’Antioco islandas part of the orogenic magmatism in the sulcisarea (sW mainland sardinia).
the sulcis volcanic district is thesouthernmost part of the calc-alkaline magmaticbelt that developed during late eocene - middleMiocene time span (lustrino et al., 2009) alongthe western side of sardinia (Fig. 1), as aconsequence of subduction acting in the WesternMediterranean sea beneath the southerneuropean margin (e.g., doglioni, 1991; doglioniet al., 1999 and references therein).
this work illustrates the mineralogical,geochemical and petrological aspects of thesant’Antioco volcanics in order to integrate thelarge set of studies on Cenozoic orogenicmagmatism in sardinia and to reconstructpetrogenetic processes. in particular, we focus onthe influence of H2o on the magma liquid linesof descent in the light of recent studies on naturaland experimental calc-alkaline systems (sissonand grove, 1993a, b; Pichavant and Macdonald,2007 and references therein; ulmer, 2007 and
29Petrogenesis of the High-Alumina Basalt-Andesite suite from Sant’Antioco Island, SW Sardinia, Italy
Fig. 1 - a) geological sketch map of sardinia, showing the location of sant’Antioco island and other subduction-relatedCenozoic volcanics (after Cherchi and Montadert, 1982, savelli et al., 1979 and Morra et al., 1997, modified). (1) Pliocene-Quaternary sediments; (2) Middle Miocene-Pleistocene anorogenic volcanics; (3) late eocene-Middle Miocene sedimentaryand volcanic rocks: (3a) silicic ignimbrites; (3b) andesites s.l.; (4) continental sedimentary rocks of Cixerri Formation; (5)undifferentiated Paleozoic basement and Mesozoic up to eocenic cover; (6) main Post-Paleozoic regional faults. b) geologicalsketch map of sant’Antioco island. c) enlargement of the area shown in b, where the sampling sites are reported. (* lavaclasts in Pispisia-M.te Arbus polygenic breccias).
A.M. Conte, d.M. PAllAdino, C. Perinelli and e. Argenti30
references therein). Furthermore, possiblerelationships with compositionally analogous, oreven less differentiated, Cenozoic orogenic rock-types of western sardinia will be considered,aiming at identifying a reliable near-primaryparental magma for the studied rock association.
geologiCAl BACKground
the sant’Antioco island, located immediatelyoffshore of southwestern mainland sardinia(italy; Fig. 1), extends 17.5 km in a n-sdirection, with a maximum W-e width of 10 km.it is part of the calc-alkaline volcanic belt thatstretches along the western coast of sardinia,which developed during late eocene-middleMiocene (38-12 Ma, lustrino et al., 2009 andreferences therein) along the main branches ofthe “sardinian rift system” (Main BranchVolcanic Zone) and in sulcis (south-WesternVolcanic Zone; sensu lecca et al., 1997). thesardinian rift system and related magmatismare interpreted in the frame of the kinematicprocesses acting in the western Mediterranean,which involved the Hercynian paleo-margin ofsouthern europe. these processes occurred inconnection with the back-arc spreading producedby the eastward roll-back of the westward-subducting plate related to convergence betweenAfrica and europe (doglioni, 1991; doglioni et
al., 1999 and references therein). this caused theseparation and counterclockwise (i.e.,southeastward) rotation of the Corsica-sardiniacontinental microplate up to its present position(upper oligocene-middle Miocene). in thiscontext, the Cenozoic calc-alkaline magmatismof sardinia is directly related to the subductionof oceanic crust. the earliest calc-alkalinemagmatic events along the present western sideof sardinia predate the rotation of the Corsica-sardinia microplate; then magmatism migratedsoutheastward following the west-directed,eastward retreating, subduction zone beneathsardinia until middle Miocene. the Cenozoic
orogenic magmas in sardinia were emplacedthrough a variably thinned continental crust (ca.30 to 20 km thick, from north to south; eggeret al., 1988; sartori et al., 2004; sau et al., 2005),which may account for the differences incomposition and eruptive style of magmaticactivity from north to south along the sardinianrift (e.g., lustrino et al., 2002).
the sulcis volcanic district, which includes asouth-western coastal area of mainland sardiniaand the nearby islands of san Pietro andsant’Antioco, represents the southernmostexpression of orogenic magmatism in sardinia(Fig. 1) and one of its youngest occurrences aswell. Volcanic stratigraphy is characterized by: 1)lava and subordinate pyroclastic successions,including andesites, basaltic andesites and rarebasalts, of oligo-Miocene age, which crop out insulcis mainland (i.e., s. giovanni suergiu,narcao, giba) and in the southern part ofsant’Antioco, and 2) silicic welded ignimbritesand subordinate lava flows and domes, ranging incomposition from dacites to comendites, ofMiocene age, exposed in mainland sulcis(Portoscuso area), in the northern and westernparts of sant’Antioco, and throughout the wholesan Pietro island (Assorgia et al., 1990; Morra et
al., 1994; Pioli and rosi, 2005; Pioli et al., 2008).the pre-volcanic substrate on sant’Antioco
includes biohermal-biostromal and ooliticlimestones and silty marls of Cretaceous age,cropping out in the se part of the island(Maccioni et al., 1990). Almost all of the islandis made up of volcanic rocks of Miocene age fromboth silicic, mainly explosive, and mafic tointermediate, mainly effusive, activities.Published K-Ar age data indicate values clusteredat around 16-18 Ma (Araña et al., 1974; savelliet al., 1979; Montigny et al., 1981; Maccioni et
al., 1991). A younger age of about 12 Ma wasrecently obtained by40Ar/39Ar determinations forthe trachytic neck of isola del toro, south ofsant’Antioco (lustrino et al., 2007), which isintermediate between the age of the last products
Petrogenesis of the High-Alumina Basalt-Andesite suite from Sant’Antioco Island, SW Sardinia, Italy 31
of the oligo-Miocene orogenic igneous phase insW sardinia and the early volcanism of the lateMiocene-Quaternary phase in se sardinia (ca.6.64 Ma, lustrino et al., 2007).
Field CHArACteristiCs
oF tHe studied roCK-tyPes
the present study investigates the mafic tointermediate volcanic rocks that crop out in thesouthern part of sant’Antioco island. these arerepresented by high- and low-aspect ratio lavaflows and small lava domes (typically less than1 km across), associated with minor pyroclastic
deposits, locally cut by lava dykes. in order tocharacterize this volcanic activity onsant’Antioco, a set of sixty-seven representativerock samples has been analyzed, as listed intABle 1 (sample localities are shown in Fig. 1).tABle 1 also summarizes relevant fieldcharacteristics of the studied volcanics, alongwith their correlation with the units identified byprevious studies (i.e., luxoro, 1987; Maccioni et
al., 1990). Following the stratigraphic scheme ofluxoro (1987), the studied samples include: i)twenty-two samples from the “lave iniziali”(early lavas, hereon abbreviated as li),consisting of thick (up to 10 m), massive, lava
sAMPles Field AsPeCt luxoro (1987) Maccioni et al. (1990)
flows and lava domes; ii) eleven samples fromlava injections and lava blocks within thePispisia-M.te Arbus polygenic volcanic brecciadeposits (hereon P); iii) twenty-seven samplesfrom thin (usually 1.5 m-thick) lava flows, partof the “lave terminali” (final lavas, hereon lt);iv) seven samples from lava dykes intersectingthe above volcanics (hereon F).
PetrogrAPHy
the studied sant’Antioco volcanics range incomposition from basalts to dacites (seegeochemistry section). Basaltic andesites largelyprevail within lt rock-types, while andesites aredominant within li lithologies. Both basalticandesites and andesites occur in the Pispisia-M.te Arbus lava types (P), as well as in lavadykes (F). Hereafter, for the sake of simplicity,the products from P and F units are groupedwithin either the li or lt main-types, on thebasis of bulk rock composition and of commonmafic phenocryst assemblages (plagioclase isinvariably the main phase and opaques areubiquitous, see below).
lava rocks from both li and lt mostly showporphyritic or glomeroporphyritic (15-48 vol%phenocrysts), isotropic to pilotaxitic or seriate,textures and microcrystalline groundmass.Phenocryst and groundmass assemblages, asdetermined from modal analyses in thin section,are reported in tABle 2 for a set of representativesamples.
lt rocks (also including part of F dykes)consist of basaltic andesites and occasionalbasalts characterized by the ubiquitous, althoughrelatively scarce, presence of olivine (always <5vol%) and by the prevalence of clinopyroxene(Cpx) on orthopyroxene (opx). in a few cases,within the least evolved lavas, which show verylow contents of mafic phenocrysts (about 5%),opx is lacking and olivine (3-4 vol%) prevailson Cpx.
li rocks (also including P lavas and part of Fdykes) comprise basaltic andesites and occasionaldacites characterized by: i) the ubiquitouspresence of pyroxene phenocrysts, with opx oftenprevailing on Cpx; ii) the lack or occasionalpresence of scarce amounts of olivine; iii) theoccurrence of hornblende in two samples from the
A.M. Conte, d.M. PAllAdino, C. Perinelli and e. Argenti32
rock-type lt lt lt lt li li li li
sample: sA7 sA40 sA35 sA3 sA12 sA15 sA31 sA32 HABA HABA HABA HABA And And And dAC
tABle 2Modal analyses of the most representative LT and LI rock-types (at least 1000 points counted for each sample).
Petrogenesis of the High-Alumina Basalt-Andesite suite from Sant’Antioco Island, SW Sardinia, Italy 33
most evolved andesites and in the dacite sample. Phenocryst glomeri (mostly Cpx and opx,
associated with coarse-grained Fe-ti oxides,subordinate plagioclase and occasional olivine)are more common in li than lt and typicallyoccur in two-pyroxene andesites (cfr. “type a”microinclusions in Brotzu et al., 1997a).
Mineral chemistry and geochemical features ofwhole rock samples are reported in the followingsections. Analytical methods are illustrated inAppendix A.
MinerAlogy
Plagioclase
Plagioclase is the most abundant phenocrystphase (up to 80% by volume of total phenocrystsand microphenocrysts) in both lt and li groups.it usually occurs as individual, variably sized,euhedral grains or, mainly in andesites, inglomerocrystic aggregates. Plagioclase crystalsoften show continuous zoning and clear surfaces.dusty or cellular textures may be present, wherecrystal portions may enclose glassy material.Plagioclase phenocrysts also enclose magnetiteand minute plagioclase crystals or, more rarely,fine skeletal Cpx grains.
in li rock-types, plagioclase crystals showcommon intracrystalline compositional ranges,with either normal (core: An75-67; rim: An65-59) orreverse (core: An64-68; rim: An72-74) zoning.Although usually normally zoned, ltplagioclases are characterized by a widercompositional pattern: phenocryst corecompositions even >An90 are observed in severalexamples, while phenocryst rim andmicrophenocryst mostly range An60-75 incomposition (representative analyses arereported in tABle 3). As a whole, thecompositional ranges of lt plagioclase varysignificantly, even within rocks with similarevolution degree. For example, if we considerthe least evolved lithotypes, the prevailing
plagioclase compositions vary from An75-85
(sA7), through An78-90 (sA40), to An85-93 (sA35)(tABle 3; Fig. 2), concomitant with increasingAl2o3 contents in the corresponding whole rocks(cfr. geochemistry section).
Pyroxenes
in lt rocks, Cpx is the dominant maficphenocryst phase (5-15% by volume), while opxusually does not exceed 5 vol%. With increasingdegree of differentiation, the amount of opx vs
Cpx increases to become prevalent in most of theolivine-free li samples. in these latter, Cpx andopx phenocrysts are often associated inpolycrystalline aggregates (glomeri). As a whole,Cpx may occur as: i) euhedral pheno- tomicrophenocryst grains; ii) coarser, roundedgrains, sometimes enclosing corroded relics ofolivine; iii) microgranular coronas that mantleboth opx and Cpx phenocrysts, mainly inevolved li andesites; iv) microliths in thegroundmass. Clinopyroxene is always augitic,with a narrow compositional range (Wo37-45 en39-
45 Fs11-20). no clear compositional variations areobserved as a function of the degree of evolutionfrom basalts to andesites: both Al2o3 (1.6-6.5wt%) and tio2 (0.33-1.10 wt%) contents do notwell correlate with the magnesium number [mg#= Mgo/(Mgo+Feo*) molar], as they showdifferent amounts for comparable mg# values(e.g., lt in Fig. 3 and tABle 4).
orthopyroxene (Wo2-4 en54-72 Fs25-42, tABle 4)occurs as either euhedral to subhedralphenocrysts or microphenocrysts (either isolatedor in polycrystalline glomeri), in some casesmantled by Cpx grains. solid inclusions,commonly of magnetite and, more rarely, ofplagioclase laths, are also found.
Olivine
olivine mostly occurs in lt rocks and usuallydoes not exceed 2-3 vol%. in a few, less evolved,
A.M. Conte, d.M. PAllAdino, C. Perinelli and e. Argenti34
sA7
s
A7
s
A7
s
A7
s
A7
s
A7
s
A35
sA
35
sA
35
sA35
sA
35 s
A35
sA
40
sA40
s
A40
sA
40 s
A40
uni
t
lt
lt
lt
lt
lt
lt
lt
lt
lt
lt
lt
lt
l
t
l
t
l
t
l
t
l
t
1/c
1/r
2/c
2/r
mph
g
1/c
1/r
2/c
3/c
mph
g
1
/c
2
/c
2
/r
m
ph
g
sio
2
45.
45
48.7
8
47.4
3
47.5
4
47.
73
48.9
9
44.8
9
49.6
4
43.6
1 4
5.89
46
.44
49
.21
44.
42
45.5
0 4
7.65
46
.82
47
.86
Al 2o
3
3
4.88
31
.26
33
.05
32
.92
3
2.38
31
.65
34
.05
31
.19
35
.92
34.
68
33.3
0
31.9
2 3
5.55
31
.74
32.
97
33.7
1
32.5
9Fe
o
0.
84
0.6
3
0.8
5
0.6
4
0.
90
0.7
6
0.5
4
0.9
7
0.6
2
0.6
4
0.8
1
0.9
3
0.5
8
0.6
6
0.6
1
0.6
7
0.8
3C
ao
17.
77
16.1
8
16.6
2
15.8
5
15.
70
14.9
7
19.7
0
15.7
3
19.2
5 1
7.87
16
.91
14
.99
18.
68
17.7
1 1
5.80
16
.83
15
.57
na 2
o
1.5
6
2.8
8
2.1
9
2.7
3
2.
77
3.2
7
0.8
7
3.0
3
0.7
4
1.5
3
2.1
8
3.2
5
1.1
4
1.7
1
2.8
7
2.2
1
2.8
6K
2o
-
0.
09
0.1
0
0.1
0
0.
10
0.1
0
0.0
4
0.1
3
-
-
-
0.1
2
-
-
-
-
0.
18To
tal
100.4
9 9
9.8
2 1
00.2
4 9
9.7
8 99.5
8
99.7
4 1
00.0
9 1
00.6
9 100.1
4100.6
2 9
9.6
4 1
00.4
1100.3
7 9
7.3
2 9
9.9
0 1
00.2
4 9
9.8
9
An
(mol
%)
86
.32
75
.26
80
.28
75
.80
7
5.37
71
.26
92
.39
73
.62
93
.50
86.
57
81.0
8
71.3
5 9
0.05
85
.13
75.
26
80.8
0
74.2
8A
b (m
ol %
)
13.6
8
24.2
4
19.1
4
23.6
3
24.
06
28.1
7
7.3
8
25.6
6
6.5
0 1
3.43
18
.92
27
.99
9
.95
14
.87
24.
74
19.2
0
24.6
9o
r (m
ol %
)
0.
00
0.5
0
0.5
8
0.5
7
0.
57
0.5
7
0.2
2
0.7
2
0.0
0
0.0
0
0.0
0
0.6
6
0.0
0
0.0
0
0.0
0
0.0
0
1.0
2
Kd
2.40
1
.18
1
.59
1
.22
1.19
0
.96
5
.12
1
.17
5
.88
2
.63
1
.75
1
.04
4
.12
2
.60
1
.38
1
.91
1
.37
s
AV
11sA
V11
sA
V11
sAV
11 s
AV
11 s
AV
11 s
AV
15sA
V15
sA
V15
sAV
15sA
V15
sA
V5
sA
V5
sA
V5
sA
V5
s
AV
5 s
AV
5u
nit
lt
l
t
l
t
l
t
l
t
lt
P
P
P
P
P
li
li
li
li
l
i
l
i
1/c
1/r
2/c
2/r
mph
g
1/c
1/r
2/c
m
ph
g
1/
c
1
/r
2
/c
2/
r
m
ph
g
sio
2
46.
54
50.8
7
45.1
8
45.3
2
53.
86
55.2
4
46.1
5
47.3
6
47.4
2 4
8.70
59
.11
49
.32
53.
02
50.3
2 5
2.43
51
.31
59
.31
Al 2o
3
3
4.12
30
.80
34
.73
34
.53
2
8.42
27
.33
34
.23
33
.12
33
.33
32.
38
23.9
2
32.2
2 2
9.62
31
.40
29.
94
30.5
1
24.3
6Fe
o
0.
42
0.7
4
0.5
6
0.8
1
1.
28
1.5
1
0.7
2
0.6
1
0.6
9
0.7
8
1.0
3
0.3
3
0.4
4
0.5
2
0.3
0
0.6
4
0.8
6C
ao
17.
07
13.5
1
18.1
4
17.9
8
10.
88
9.6
6
17.2
5
16.5
3
16.4
1 1
5.12
5
.94
15
.15
12.
10
14.1
3 1
2.47
13
.19
6
.26
na 2
o
1.7
2
3.7
9
1.0
6
1.2
8
5.
09
5.4
5
1.5
4
1.9
7
2.0
4
2.7
2
5.5
3
2.8
4
4.4
9
3.4
0
4.3
1
3.9
7
5.3
1K
2o
-
-
-
-
0
.29
0
.72
-
-
0.1
3
-
3.6
6
-
0.2
2
0.1
3
0.2
0
-
3.
88To
tal
99.8
7 9
9.7
1 9
9.6
7 9
9.9
2 99.8
2
99.9
1 9
9.8
9 9
9.5
9 100.0
2 99.7
0 9
9.1
9 9
9.8
6 99.8
9 9
9.9
0
99.6
5 9
9.6
2 9
9.9
8
An
(mol
%)
84
.58
66
.33
90
.44
88
.59
5
3.24
47
.40
86
.09
82
.26
81
.01
75.
44
29.2
5
74.6
7 5
9.06
69
.14
60.
81
64.7
4
30.5
5A
b (m
ol %
)
15.4
2
33.6
7
9.5
6
11.4
1
45.
07
48.3
9
13.9
1
17.7
4
18.2
2 2
4.56
49
.28
25
.33
39.
66
30.1
0 3
8.03
35
.26
46
.90
or
(mol
%)
0.00
0
.00
0
.00
0
.00
1.69
4
.21
0
.00
0
.00
0
.76
0
.00
21.
46
0.0
0
1.2
8
0.7
6
1.1
6
0.0
0
22.5
5
Kd
2.42
0
.87
4
.16
3
.42
0.52
0
.43
2
.96
2
.22
2
.13
1
.47
0
.28
1
.73
0
.87
1
.35
0
.94
1
.08
0
.38
n
ote:
c: p
heno
crys
t cor
e; r
: phe
nocr
yst r
im (
1, 2
, 3: p
heno
crys
t ide
ntif
icat
ion)
; mph
: mic
roph
enoc
ryst
; g: g
roun
dmas
s. t
he p
lagi
ocla
se-m
elt C
a-n
aex
chan
ge d
istr
ibut
ion
coef
fici
ents
(K
d)
are
also
rep
orte
d (s
ee te
xt f
or c
alcu
latio
n de
tails
).
tAB
le
3R
epre
senta
tive
ele
ctro
n m
icro
pro
be
analy
ses
of
pla
gio
clase
fro
m t
he
Sant’
Anti
oco
stu
dy
rock
-typ
es.
Petrogenesis of the High-Alumina Basalt-Andesite suite from Sant’Antioco Island, SW Sardinia, Italy 35
rocks it may attain up to 5 vol% and representthe main mafic phase. overall, Fo componentranges 50-59 mol%; individual crystals areusually normally zoned (tABle 5). generally,olivine crystals are altered to iddingsite, even ifthey preserve the original habit. in a few cases,fresh olivine relics are observed at the core ofaltered crystals. some olivine crystals arebordered by granular Cpx.
Amphibole
euhedral to subhedral amphibole phenocrysts,showing opacitic rims, are found in scarceamounts (<2 vol%) in two li andesite sampleswith sio2 contents >58 wt% and in the dacitesample. representative analyses, reported intABle 5, allow a classification as Mgo-hornblende, following leake et al. (2004).Compositional (i.e., low Al2o3 content, ~8 wt%on average) and textural relationships testifytheir late stage crystallization under low pressureconditions.
Fe-Ti oxides
Fe-ti oxides (tABle 5) are presentubiquitously both as microphenocrysts (eitherdispersed or bordering pyroxene phenocrysts)and, in higher amounts, as groundmass phases.Magnetite-titanomagnetite is common in basalticandesites, while ilmenite only occurs inandesites.
geoCHeMistry
representative bulk rock compositions fromthe sant’Antioco volcanics are reported in tABle
6. essentially they plot in the basaltic andesiteand andesite fields of the tAs classificationdiagram (le Bas et al., 1986), with quitesubordinate basalts and dacites. According toother classification diagrams, i.e. K2o vs sio2
(Peccerillo and taylor, 1976; Fig. 4a), and Mgovs Al2o3 (Kersting and Arculus, 1994; Fig. 4b),the studied products are mainly high-aluminabasaltic andesites and andesites (with quitesubordinate basalts and dacites: only one datapoint each) of medium-K calc-alkaline affinity.on the other hand, on the mere basis of theFeo*/Mgo vs sio2 diagram (Miyashiro, 1974)all the investigated rocks would show a tholeiiticto mildly tholeiitic affinity (Fig. 4c). However,the plot of Feo*/Mgo vs Feo* (inset in Fig. 4c)indicates that the relatively high values ofFeo*/Mgo in the sant’Antioco rock suite arenot accompanied by an iron enrichment trendtypical of tholeiitic series. Furthermore, apossible tholeiitic affinity for the sant’Antiocovolcanics conflicts with the lack of pigeonitemicrophenocrysts (Kuno, 1959) and of Fe-enrichment in plagioclase (Aramaki and Katsura,1973), as well as with the observed tio2 andAl2o3 contents, which pertain more properly tocalc-alkaline rock-series. on these grounds, therelatively high Feo*/Mgo ratios, which mimica tholeiitic affinity, may be likely linked to therelative proportions of fractionating plagioclase
Fig. 2 - Frequency histograms displaying the variation of Ancontents in plagioclase of the sA7, sA35 and sA40 high-alumina basaltic andesite (HABA) samples from lt unit,which stand for the least evolved sant’Antioco products(frequency expressed as number of analyses). note that eachsample shows a dominant plagioclase composition related tothe Al2o3 content in the whole rock (cfr. tABle 6).
A.M. Conte, d.M. PAllAdino, C. Perinelli and e. Argenti36(a
)
sAV
12
sA7
s
AV
11 s
AV
11
sAV
15 s
AV
15
sA
V15
sAV
15 s
AV
13
sA
V13
s
A10
sAV
5
sA
V5
s
AV
8
sA
V8
uni
t
l
t
l
t
lt
lt
P
P
P
P
l
i
li
l
i
li
l
i
li
l
i
r
c
c
r
c
r
m
ph
g
r
mph
c
c
r
c
r
sio
2
52
.18
5
3.77
5
1.25
5
3.55
52.4
8
51.4
9
52.
80
52.7
2
51.
40
52
.19
5
2.19
5
1.68
51.7
3
52
.16
52.1
7ti
o2
0.2
5
0.
33
0
.35
0.27
0.3
1
0.4
2
0
.30
0
.25
0.29
0.3
8
0
.21
0.2
8
0
.16
0.3
8
0
.30
Al 2o
3
1.
58
1.
05
2
.62
1.23
1.3
9
2.2
4
2
.10
1
.77
0.38
0.4
4
0
.83
1.3
6
1
.37
0.6
2
0
.39
Feo
19.9
9
16.
99
18.
73
18.
39
19
.58
19
.77
1
7.45
19
.99
2
6.44
24.7
1
21.
58
24.
51
24
.86
23.5
8
24
.41
Mno
0.3
5
0.
48
0
.40
0.62
0.5
4
0.3
5
0
.34
0
.35
0.93
1.0
0
0
.69
0.8
8
0
.75
0.7
8
0
.93
Mgo
2
3.53
2
5.64
2
4.61
2
5.20
24.1
2
24.0
3
24.
63
23.7
7
19.
21
19
.81
2
2.94
1
9.96
19.7
2
20
.43
20.1
8C
ao
1
.51
1.82
1.5
4
1.
50
1
.61
1
.75
1.5
9
1.5
3
1.
72
1
.88
1.7
7
1
.17
1.4
7
1
.86
1.6
4n
a 2o
0.27
-
0.1
9
-
0
.18
-
0.4
6
0.2
8
-
-
0
.03
0.1
8
0
.34
-
-To
tal
9
9.6
6 100.0
8 9
9.6
9 100.7
6 100.2
1 1
00.0
5 9
9.6
7 1
00.6
6 1
00.3
7 100.4
1 100.2
4 1
00.0
2 100.4
0 99.8
1 100.0
2
W
o (m
ol%
)
3
.02
3.56
3.0
5
2.
93
3
.19
3
.46
3.2
0
3.0
3
3.
52
3
.82
3.5
2
2
.40
3.0
1
3
.78
3.3
3e
n (m
ol%
)
65
.55
6
9.79
6
7.92
6
8.39
66.5
6
66.1
3
68.
89
65.5
3
54.
55
56
.00
6
3.52
5
6.95
56.1
1
57
.78
57.0
6Fs
(m
ol%
)
3
1.43
2
6.65
2
9.03
2
8.69
30.2
5
30.4
0
27.
92
31.4
4
41.
94
40
.18
3
2.95
4
0.65
40.8
9
38
.44
39.6
0
m
g#
0
.68
0.73
0.7
0
0.
71
0
.69
0
.68
0.7
2
0.6
8
0.
56
0
.59
0.6
5
0
.59
0.5
9
0
.61
0.6
0
(b
) s
AV
12
sA
7
sA
35
sAV
15
sAV
15 s
AV
15
sAV
10sA
15
sA
15
sA
15
sA
15
sA
10
sA
10
sA
V8
sAV
8u
nit
lt
lt
l
t
P
P
P
li
li
l
i
li
l
i
li
l
i
li
l
i
c
c
c
c
r
m
ph
c
c
r
m
ph
r
c
r
c
r
si
o2
50
.77
5
1.01
4
8.03
4
8.57
49.5
7
50.6
3
49.
66
49.6
6
50.
44
51
.51
5
0.44
4
9.63
50.9
1
51
.51
50.5
9ti
o2
0.5
7
0.
52
1
.09
0.88
0.8
5
0.7
9
0
.67
0
.67
0.79
0.3
3
0
.79
0.7
0
0
.33
0.3
3
0
.58
Al 2o
3
2.
67
1.
86
5
.12
6.32
5.7
9
3.4
3
3
.96
3
.96
3.88
2.5
5
3
.88
3.1
3
1
.64
2.5
5
2
.63
Feo
10.4
4
11.
42
9
.74
7.34
7.8
8
10.6
7
11.
77
11.7
7
10.
45
10
.05
1
0.45
1
1.30
12.9
4
10
.05
11.9
4M
no
0
.16
0.32
0.3
0
0.
15
0
.20
0
.46
0.3
9
0.3
9
0.
23
0
.00
0.2
3
0
.30
0.4
5
0
.00
0.2
3M
go
15.
27
16.
06
14.
08
14.
37
14
.78
15
.11
1
4.11
14
.11
1
3.98
14.9
5
13.
98
14.
27
14
.14
14.9
5
13
.55
Cao
19.0
8
18.
36
20.
76
20.
93
20
.53
18
.45
1
8.46
18
.46
1
9.45
19.6
8
19.
45
20.
36
19
.65
19.6
8
19
.98
na 2
o
0.
39
0.
24
0
.48
0.55
0.4
4
0.2
7
0
.52
0
.52
0.40
0.3
6
0
.40
0.3
4
0
.31
0.3
6
0
.25
Cr 2
o3
0.1
3
Tota
l 99.3
5 99.7
9 9
9.6
0 99.1
1 100.0
4 9
9.9
3 9
9.5
4 9
9.5
4 9
9.6
2 99.4
3 99.6
2 1
00.0
3 100.3
7 99.4
3 99.7
5
W
o (m
ol%
)
39.
26
37.
19
43.
08
44.
75
43
.31
38
.31
3
8.79
38
.79
4
1.17
40.7
2
41.
17
41.
58
39
.98
40.7
2
41
.33
en
(mol
%)
43.7
2
45.
27
40.
66
42.
75
43
.39
43
.65
4
1.26
41
.26
4
1.18
43.0
5
41.
18
40.
55
40
.03
43.0
5
39
.01
Fs (
mol
%)
17.
03
17.
54
16.
27
12.
50
13
.31
18
.04
1
9.95
19
.95
1
7.65
16.2
3
17.
65
17.
87
20
.00
16.2
3
19
.66
mg#
0.7
2
0.
71
0
.72
0.78
0.7
7
0.7
2
0
.68
0
.68
0.70
0.7
3
0
.70
0.6
9
0
.66
0.7
3
0
.67
c: p
heno
crys
t cor
e ; r
: phe
nocr
yst r
im; m
ph: m
icro
phen
ocry
st; g
: gr
ound
mas
s; m
g#: M
go/(
Mgo
+Feo
) m
olar
, with
all
iron
as
Fe2+
.
tAB
le
4R
epre
senta
tive
ele
ctro
n m
icro
pro
be
analy
ses
of
ort
hopyr
oxe
nes
(a)
and c
linopyr
oxe
nes
(b)
from
the
Sant’
Anti
oco
stu
dy
rock
-typ
es.
Petrogenesis of the High-Alumina Basalt-Andesite suite from Sant’Antioco Island, SW Sardinia, Italy 37
and mafic phases during magma evolution,leading to the observed low-Mgo and high-Al2o3 compositions (e.g., Pichavant andMcdonald, 2007; also see below).
Variation diagrams (Fig. 5) exhibit broadcorrelations of major oxides with Mgo, besideswide scatter (e.g., Al2o3 and Cao). As a whole,in spite of distinct field appearance andpetrographic features, li and lt do not defineseparate compositional trends, whereas both rockgroups display correlations consistent withcrystallization and removal of theaforementioned phenocryst phases (i.e.,
plagioclase, pyroxenes, Fe-ti oxides, olivine). infact, Feot, tio2, Cao, and Al2o3 are positivelycorrelated with Mgo, while sio2 and na2o arenegatively correlated with Mgo. the extent ofAl2o3 and Cao scattering, however, implies atleast different mineral phase proportions and/orcompositions during fractional crystallization.
trace elements depict broad correlations withsilica, despite considerable scattering. inparticular, rb, Ba, Zr, y and nb (tABle 6, Fig.6) exhibit an incompatible behaviour, as alsoevidenced by their mutual positive correlations(not shown). Conversely, V and sr behave ascompatible elements and reflect the progressiveremoval of Fe-ti oxides and plagioclase. Cr andni display low concentrations (<100 and <60ppm, respectively; tABle 6), a typical feature oflow-Mgo, high-Al2o3 arc magmas worldwide,which is related to the prolonged removal ofmafic phases during magma differentiation (e.g.,Pichavant and Macdonald, 2007 and referencestherein).
spider diagrams of trace elementconcentrations normalized to n-MorB (sun andMc donough, 1989; Fig. 7a), show depletion ofHFse (High Field strength elements; withpronounced troughs at nb and ti) with respectto lile (large ion lithophile elements) andlree (light rare earth elements), as it istypical of arc magmas. the whole set of samplesshares very similar patterns, with a generalenrichment toward the most differentiated rocksand only slight changes in theincompatible/compatible element ratios. thechondrite-normalized ree patterns are alsosimilar for rocks with different degree ofevolution (Fig. 7b) and show a slight increase of(la/yb)n (from 3.6 to 5.9; subscript n indicatesnormalized abundances), for near constant(sm/yb)n, in the most differentiated rocks. thelatter also show a weak eu trough (eu/eu*=0.83-0.88, where eu* is (sm+gd)/2), which insteadis lacking in less differentiated rock types(eu/eu*=0.93-0.97).
Fig. 3 - Variation diagrams of magnesium number (mg#) vs
Al2o3 (a) and tio2 (b) for clinopyroxenes of representativesant’Antioco samples. note the difference in Al2o3 and tio2
contents of clinopyroxenes in lavas with similar evolutiondegree (i.e., sA7 and sA35 considered for petrogeneticmodeling). HAB: high-alumina basalt; HABA: high-aluminabasaltic andesite; And: andesite.
A.M. Conte, d.M. PAllAdino, C. Perinelli and e. Argenti38
overall, trace element patterns indicate thatmagmatic evolution was mostly controlled byfractional crystallization, consistent withevidence from major element variations. thewide data point scatter, however, implies thatfractional crystallization cannot fully account forthe observed chemical variations in thesant’Antioco volcanics and that an additional
process, such as crustal contamination, waslikely involved in magma evolution (also seebelow). this seems to be confirmed by the87sr/86sr initial (i.e. at 15 Ma) isotopic ratios forsant’Antioco whole rocks, which range from0.70698 to 0.70858±1 (tABle 6). Crustalassimilation is also suggested by the positivecorrelations of the initial sr isotope ratios with
Olivine
sAV12 sAV12 sA7 sA35 sA35 sA40 sA40 unit lt lt lt lt lt lt lt c mph c c r c r sio2 35.73 38.70 35.53 37.56 38.58 36.7 38.13 Feo 37.82 32.27 37.86 34.76 36.1 37.7 39.047 Mno 0.64 0.68 0.75 0.52 0.24 0.64 0.50 Mgo 26.37 25.94 24.99 25.91 24.18 24.94 22.06 nio Cao 0.64 0.71 0.43 0.47 0.70 0.29 0.41 Total 100.56 97.58 99.13 98.75 99.10 99.98 99.74
sA34 sA25 sA32 sAV8 sAV8 sAV11 unit li F li li li lt sio2 47.53 47.71 47.41 - 0.21 - tio2 1.44 1.34 1.45 1.70 45.84 3.12 Al2o3 8.68 8.33 8.18 1.43 0.12 1.34 Feo 14.58 13.61 13.79 89.99 50.32 88.39 Mno 0.07 0.12 0.12 0.35 2.13 0.37 Mgo 13.75 14.41 14.37 - 0.1 0.33 Cao 10.34 10.54 10.97 na2o 1.36 1.61 1.37 K2o 0.21 0.31 0.36 Total 97.96 97.98 98.02 93.47 98.72 93.55 mg# 0.63 0.65 0.65
c: phenocryst core ; r: phenocryst rim; mph: microphenocryst; mg# = Mgo/(Mgo+Feo) molar, with all iron asFe2+.
tABle 5Representative electron microprobe analyses of olivine, amphibole and Fe-Ti oxides
from the Sant’Antioco study rock-types.
Petrogenesis of the High-Alumina Basalt-Andesite suite from Sant’Antioco Island, SW Sardinia, Italy 39
sio2 and with some incompatible trace elementssensitive to crustal contamination (i.e., rb, Ba,Zr), along with the negative correlation with srabundance, sensitive to crustal contamination +plagioclase fractionation (Fig. 8).
estiMAtes oF MAgMA teMPerAtures
And H2o ConCentrAtions
geothermometric determinations based onmagnetite/ti-magnetite and ilmenite pairs(Buddington and lindsley, 1964) are preventedfor the sant’Antioco suite, due to the rare
occurrence of coexisting, unexsolved crystals ofthe two opaque phases in the analyzed samples.therefore, in order to constrain thecrystallization temperatures of magmas, the two-pyroxenes equilibrium geothermometer oflindsley (1983) was applied. this approach mayprovide a general indication of temperature evenin the lack of coexisting low-Ca pyroxene, byplotting individual augite phenocrystcompositions. the projection onto lindsley’sisotherms for P = 500 MPa (preferred to the 1gPa and 1 atm schemes to constrain themaximum crystallization depth at sant’Antioco,
Fig. 4 - Classification diagrams of K2o vs sio2 (after Peccerillo and taylor, 1976) (a), Mgo vs Al2o3 (after Kersting andArculus, 1994) (b), Feo*/Mgo vs sio2 (after Miyashiro, 1974) (c), for the investigated sant’Antioco volcanics. inset in (c)shows the plot of Feo* vs Feo*/Mgo (* = all iron as Feo). the three samples representative of the least evolved compositions(i.e., sA7, sA35 and sA40) are marked. solid lines enclose the compositional field of Cenozoic orogenic volcanic rocks ofsardinia (data from literature cited in the text). HMB: high-magnesia basalt; HAB: high-alumina basalt; a: medium-K basalt;b: medium-K basaltic andesite; g: medium-K andesite; d: medium-K dacite.
A.M. Conte, d.M. PAllAdino, C. Perinelli and e. Argenti40
eA
rly
lA
VA
s (l
i)
Pis
Pisi
A -
M.te
Ar
Bu
s u
nit
(P)
sam
ple
sAV
7
sA1
1
sA1
2
sAV1
3
sA15
s
A10
sA31
s
AV18
sA17
sAV5
sAV8
s
A34
s
A33
sA3
2
sA27
* s
AV15
* s
AV14
*sAV
19**
sAV1
7**s
AV16
**sA
V20*
a roc
k-ty
pe H
ABA
HAB
A
And
A
nd
An
d
And
And
And
And
And
And
A
nd
An
d
d
AC
HAB
A H
ABA
HA
BA
HABA
H
ABA
An
d
And
sio
2
54
.60
54.
89
55.1
9 5
5.28
55.
90
56.8
2
57.0
4 5
7.73
58
.02
58.
99
59.5
1 6
0.33
61.
52
65.
04
5
2.68
53
.18
53
.91
54.
51
55.5
5 5
6.00
57.
22ti
o2
0.
90
1.0
3
0.9
6
0.9
0
0.9
0
0.9
3
0.8
3
0.6
9
1.0
0
0.7
0
0.7
5
0.7
1
0.6
7
0.6
0
1.07
1
.05
0
.99
0
.96
0
.90
0
.91
0.
70A
l 2o3
1
9.43
18.
94
19.8
2 1
9.09
18.
51
18.0
9
18.4
8 1
9.62
17
.29
17.
86
17.3
6 1
7.17
17.
00
15.
33
1
8.82
19
.17
19
.47
18.
70
18.9
6 1
8.48
19.
71Fe
2o3
5.0
0
3.6
1
4.0
8
3.3
6
3.9
2
2.9
0
5.0
7
3.9
5
4.6
0
3.1
7
4.0
4
4.7
9
4.6
3
4.2
7
3.43
4
.42
4
.56
3
.69
3
.82
3
.63
3.
39Fe
o
3.4
5
4.7
8
3.1
9
4.5
5
4.1
7
5.3
6
2.9
0
2.6
5
3.7
1
3.9
1
3.6
4
2.2
4
2.4
7
2.1
0
5.53
4
.59
4
.06
4
.99
4
.27
4
.53
3.
36M
no
0.14
0
.16
0
.12
0
.17
0
.16
0
.17
0
.14
0
.12
0
.19
0
.15
0
.15
0
.13
0
.12
0
.13
0.
16
0.1
9
0.1
7
0.1
7
0.1
6
0.1
6
0.13
Mgo
2.
61
3.2
3
2.1
9
2.1
2
3.1
8
2.8
8
2.6
8
1.7
1
2.4
5
2.1
6
2.3
0
2.1
7
1.7
9
1.7
4
4.41
3
.76
3
.19
3
.07
2
.58
2
.59
1.
78C
ao
8.7
4
8.6
3
9.0
2
8.6
7
8.5
7
7.6
6
7.4
2
8.4
9
7.3
1
7.0
4
6.8
4
6.6
3
6.2
9
5.5
6
9.37
9
.24
9
.44
8
.63
8
.69
8
.44
8.
57n
a 2o
2.7
5
2.6
5
2.6
2
2.6
2
2.6
0
2.7
4
2.8
2
2.9
6
3.0
1
2.9
3
2.7
7
2.5
7
2.6
6
2.5
5
2.30
2
.49
2
.53
2
.30
2
.48
2
.18
2.
78K
2o
1.3
0
1.1
4
1.2
4
1.4
5
1.1
3
1.3
7
1.4
9
0.8
9
1.0
8
1.7
3
1.5
7
1.2
7
1.7
0
1.5
8
1.00
0
.72
0
.97
1
.11
1
.07
1
.42
0.
95P 2
o5
0.
19
0.2
0
0.1
8
0.2
1
0.2
3
0.2
5
0.2
0
0.2
5
0.2
8
0.2
2
0.2
1
0.2
2
0.1
9
0.1
9
0.16
0
.16
0
.17
0
.20
0
.17
0
.17
0.
26l
.o.i
.
0.
86
0.6
9
1.3
6
1.5
6
0.7
1
0.7
8
0.9
0
0.9
1
1.0
3
1.1
0
0.8
3
1.7
5
0.9
4
0.8
9
1.01
0
.99
0
.51
1
.63
1
.31
1
.43
1.
11
99.9
7 9
9.95
99
.97
99.
98 9
9.98
99
.95
99
.97
99.
97
99.9
7 9
9.96
99
.97
99.
98 9
9.98
9
9.98
99.
94
99.9
6
99.9
7 9
9.96
99
.96
99.
94 9
9.96
F
eOto
tal 7
.95
8.0
3 6
.86
7.5
7 7.7
0 7
.96
7
.47
6
.21
7
.85
6
.76
7
.28
6
.55
6.6
4 5.9
5 8
.62
8
.57
8.1
6 8.3
1 7
.71
7.8
0 6
.41
mg#
36.
9
41.
7
36.
3
33.
3
42.
4
39.
2
39.
0
32.
9
35.
7
36.
3
36.
0
37.
1
32.
5
34.
3
47.7
4
3.9
4
1.1
3
9.7
3
7.4
3
7.2
33
.1
V
2
12
222
180
1
91
186
1
91
1
88
236
209
153
220
1
05
13
6
8
3
16
2
180
148
1
33
14
2
154
2
02C
r
1
0
1
2
9
6
12
8
11
10
8
15
6
5
5
6
22
18
14
12
12
10
12C
o
1
9
2
0
1
3
1
6
1
2
16
18
18
16
14
14
13
12
20
10
24
20
13
15
19
11n
i
12
10
7
9
9
8
7
5
5
19
5
6
15
15
13
10
8
7
7
9
5r
b
3
8
3
5
4
3
4
2
4
2
41
48
45
43
60
58
55
59
56
27
29
29
33
47
45
36sr
317
2
99
34
3
357
3
13
316
292
3
29
3
76
28
7
28
9
279
268
250
3
16
319
301
3
74
39
3
379
3
64y
21
28
27
27
26
27
23
2
4
29
2
5
2
7
2
6
3
0
3
3
2
2
2
3
2
5
2
8
2
6
2
3
28
Zr
104
1
20
14
1
134
1
39
133
122
1
40
1
30
13
2
12
1
145
129
123
98
11
4
11
2
125
133
1
32
207
nb
5
7
8
7
8
7
7
9
8
8
6
8
7
7
6
5
5
7
6
6
9B
a
43
5
326
448
4
05
412
4
21
3
57
530
379
685
448
4
71
44
6
50
3
266
4
26
33
0
357
452
4
00
407
th
3.9
4
.5
4
.7
4.
5
4.
3
4.
1
4
.1
5
.1
4
.9
5
.8
4
.7
5.
4
6.
5
5.7
2
.2
3
.1
2
.6
3.
5
3.
9
4.
8
5.2
la
18
.9
22.
0
24.
1
26.
2
2
2.1
15
.0
Ce
36
.9
44.
4
46.
7
44.
9
4
3.3
30
.4
Pr
4
.47
4
.96
5
.25
5
.53
4.7
6
3
.45
nd
21
.3
23.
7
24.
2
24.
7
2
2.1
17
.7
sm
4.6
6
5.0
9
5.2
4
5.1
8
4
.54
3.8
3
e
u
1.4
5
1.5
1
1.5
95
1.48
1.2
8
1
.32
gd
4
.89
4
.96
4
.99
5
.37
4.3
4
4
.12
tb
0
.80
0
.80
0
.79
0
.80
0.6
6
0
.65
dy
4
.76
4
.79
4
.86
4
.93
3.9
8
4
.29
Ho
0
.98
0
.97
1
.00
1
.03
0.8
0
.87
er
2
.91
2
.95
3
.03
3
.11
2.3
3
2
.69
tm
0
.43
0
.43
0
.45
0
.47
0.3
4
0
.39
yb
2
.65
2
.60
2
.98
3
.06
2.2
2
2
.39
lu
0
.43
0
.43
0
.43
0
.48
0.3
4
0
.38
Hf
3
.18
3
.51
3
.67
3
.45
3.2
1
2
.26
ta
0.
44
0.4
6
0.5
5
0.6
5
0
.56
0.4
9
u
0.5
9
0.7
4
0.8
1
1.0
5
1
.24
0.6
1
(87
sr/86
sr) i0
.707
64
0.7
0807
0.7
0858
0.7
0763
tAB
le
6R
epre
senta
tive
analy
ses
of
majo
r and t
race
ele
men
ts f
rom
Sant’
Anti
oco
stu
dy
rock
-typ
es.
87Sr/
86Sr
isoto
pe
rati
os
of
sele
cted
sam
ple
s are
als
o r
eport
ed.
Petrogenesis of the High-Alumina Basalt-Andesite suite from Sant’Antioco Island, SW Sardinia, Italy 41
Fi
nA
l l
AV
As
(lt
)
dy
Ke
s (F
)sa
mpl
e
sAV
12
sA4
0
sA7
sA35
s
AV11
s
A36
sA3
sA1
sA3
8
sA
9
s
A29
s
A21
sA
V8b
s
A25
a roc
k-ty
pe H
AB
HAB
A H
ABA
HAB
A H
ABA
HAB
A
HABA
HA
BA
HAB
A H
ABA
H
ABA
An
d
And
And
sio
2
50
.86
51.
12 5
1.13
51.
68
52.6
5 5
3.16
53
.23
53.
70
54.
14
54.3
8
5
2.58
55.
26 5
8.93
6
0.60
tio
2
1
.18
1
.07
1
.04
1
.03
1
.03
1
.00
1
.00
1
.00
0
.90
1
.03
1
.12
1
.04
0
.68
0.85
Al 2o
3
20
.23
19.
64 1
8.78
20.
68
20.3
2 1
9.00
18
.36
17.
98
20.
80
19.6
6
1
8.19
18.
26 1
7.99
1
6.79
Fe2o
3
4
.43
5
.09
4
.50
4
.62
4
.42
3
.88
3
.80
5
.77
2
.95
3
.45
5
.37
4
.23
3
.48
1.99
Feo
5
.31
4
.48
5
.07
3
.94
4
.07
5
.18
4
.98
3
.30
4
.51
4
.40
4
.07
4
.49
3
.64
5.13
Mno
0.
16
0.1
8
0.1
8
0.1
6
0.1
6
0.1
8
0.1
8
0.1
8
0.1
4
0.1
6
0.1
8
0.1
5
0.1
4
0.
17M
go
2.78
3
.56
4
.61
3
.09
2
.68
3
.92
3
.85
4
.00
2
.17
3
.03
4
.19
2
.80
2
.00
1.57
Cao
10.
65
9.9
0
9.8
7 1
0.46
10
.11
9
.15
9
.16
8
.79
9
.68
8
.89
9
.45
8
.97
7
.13
5.78
na 2
o
2
.37
2
.49
2
.07
2
.36
2
.46
2
.45
2
.59
2
.46
2
.65
2
.53
2
.44
2
.33
2
.80
3.55
K2o
0
.76
0
.96
0
.69
0
.57
1
.03
0
.97
1
.19
1
.17
0
.93
1
.00
0
.69
0
.98
1
.57
1.94
P 2o
5
0
.17
0
.18
0
.16
0
.21
0
.21
0
.20
0
.22
0
.23
0
.23
0
.19
0
.22
0
.20
0
.19
0.28
l.o
.i.
1.04
1
.31
1
.85
1
.16
0
.81
0
.86
1
.39
1
.05
0
.85
1
.24
1
.46
1
.25
1
.41
1.30
99
.94
99.
97 9
9.95
99.
96
99.9
5 9
9.95
99
.95
99.
63
99.
95
99.9
6
9
9.96
99.
96 9
9.96
9
9.95
F
eOto
tal
9.2
9 9.0
5 9.1
2 8.1
0 8
.05
8
.67
8
.40
8.4
9 7.1
7 7
.50
8
.91
8.2
9 6.7
8 6.9
2m
g#
3
4.8
4
1.2
4
7.4
4
0.5
3
7.2
4
4.6
4
5.0
4
5.6
3
5.0
4
1.9
4
5.6
3
7.6
3
4.5
28.8
V
279
2
92
280
2
52
22
3
250
265
2
49
1
85
20
6
257
1
75
11
7
6
4C
r
30
17
100
19
1
6
1
6
1
6
3
1
26
1
5
4
8
1
6
6
6C
o
21
27
28
17
25
18
23
27
15
19
22
18
15
14
ni
16
2
8
4
1
3
5
2
0
3
4
1
4
1
5
15
7
1
9
7
12
5
rb
23
2
7
2
1
2
2
2
6
2
5
2
8
3
3
25
3
4
2
3
3
5
5
7
6
7sr
338
3
52
336
3
98
36
2
315
355
3
40
3
68
30
2
337
3
78
28
0
31
4y
23
24
24
25
23
26
25
25
28
26
24
22
24
35
Zr
89
8
2
8
4
8
0
9
9
8
3
1
02
101
101
126
1
06
108
121
185
nb
3
4
4
6
4
4
6
7
6
8
6
7
7
10B
a
402
3
22
380
3
45
39
7
291
343
3
01
3
10
35
4
378
5
20
53
9
48
2t
h
1.
9
2.
8
2.
4
3.
2
2.
9
2.
6
2
.7
2.
8
3
.1
4
.1
2
.3
4.
4
5.
6
6
.6l
a
13.
6
13.
4
13.
9
13.
4
13.
2
15
.9
22.9
29.9
Ce
2
7.2
2
7.8
2
8.6
2
8.2
2
7.4
33.1
41
.4
63
.4Pr
3.2
4
3.2
9
3.4
9
3.4
6
3.3
5
4.
13
5.1
2
7.
12n
d
16.
8
16.
6
17.
0
17.
3
16.
7
20
.2
22.4
32.3
sm
3.
67
3.8
7
3.8
8
3.9
6
3.8
3
4.
74
4.9
9
6.
69e
u
1.
24
1.2
4
1.2
4
1.2
9
1.2
4
1.
42
1.6
8
1.
78g
d
3.
99
4.0
8
4.1
9
4.2
5
4.2
0
4.
58
4.8
6
6.
42t
b
0.
65
0.6
9
0.6
9
0.7
1
0.7
0
0.
79
0.7
4
0.
98d
y
4.
08
4.1
1
4.1
5
4.4
0
4.3
4
4.
74
4.7
1
5.
94H
o
0.
84
0.8
3
0.8
3
0.8
9
0.8
9
0.
95
1.0
3
1.
19e
r
2.
39
2.4
7
2.5
5
2.6
8
2.7
0
2.
86
2.9
9
3.
47t
m
0.3
5
0.3
7
0.3
5
0.4
2
0.3
9
0.
42
0.4
4
0.
52y
b
2.
25
2.3
5
2.2
8
2.5
0
2.5
2
2.
74
3.0
1
3.
36l
u
0.
36
0.3
5
0.3
5
0.4
0
0.3
9
0.
42
0.4
5
0.
52H
f
1.
73
2.2
3
2.2
7
2.2
8
2.4
1
2.
72
4.3
1
4.
63ta
0.2
2
0.4
3
0.2
9
0.4
4
0.6
4
0.
59
0.8
7
0.
91u
0.4
1
0.5
2
0.4
9
0.3
7
0.5
4
0.
58
0.9
8
1.
13(87
sr/86
sr) i
0.70
716
0.
7069
8
0.7
0844
tAB
le
6C
onti
nued
...
a H
AB
: hig
h-al
umin
a ba
salt;
HA
BA
: hig
h-al
umin
a ba
salti
c an
desi
te; A
nd
: and
esite
;d
AC
: dac
ite.
* la
va in
ject
ion
in p
olyg
enic
bre
ccia
;**
lava
cla
sts
in p
olyg
enic
bre
ccia
;m
g# =
Mgo
/(M
go+F
eo*)
mol
ar, w
ith a
llir
on a
s Fe
o.
A.M. Conte, d.M. PAllAdino, C. Perinelli and e. Argenti42
Fig. 5 - Variation diagrams of selected major oxides vs Mgo for the investigated sant’Antioco volcanic suite. the threesamples representative of the least evolved compositions (i.e., sA7, sA35 and sA40) are marked.
Fig. 6 - Variation diagrams of selected trace elements vs sio2 for the investigated sant’Antioco volcanic suite. the threesamples representative of the least evolved compositions (i.e., sA7, sA35 and sA40) are marked.
according to the observed phase relationshipsand Melts modeling; ghiorso et al., 1994; seebelow in this section) indicates that mosttemperatures fall in the range of 1000-1150°C.
one of the least evolved basaltic-andesites, sA7(lt), shows the highest equilibrationtemperature at ~1200°C (Fig. 9).
in addition, crystallization temperatures were
estimated by applying the plagioclase-liquidgeothermometer (Putirka, 2008) to theplagioclase cores-whole rock pairs. For thesecalculations, basaltic andesite and andesitesamples characterized by the lowest contents ofphenocrysts (either isolated or in glomeri) wereselected as the best representative of liquidcompositions. the calculated temperatureinterval of 1100-1165°C broadly matches thatobtained through lindsley’s (1983) method.
the amount of dissolved H2o in the systemcan be also estimated based on the plagioclase-melt equilibrium equation (Putirka, 2008).Calculations yield H2o concentrations in the 2.3-3.9 wt% range, where the highest value refers tothe highest-Al2o3 lava sample (sA35). A similarrange of H2o contents (2.0-4.1 wt%) is derivedthrough the Pichavant and Macdonald (2007)equation (an empirical geohygrometer based onmelt compositions for arc basalt melts), using tvalues obtained from pyroxene geothermometry.notably, among the least evolved lavas (mg# ≥0.4), sA35 records both the highest H2o content(~4 wt%) and the lowest crystallizationtemperature (tABle 7).
the obtained H2o contents are consistent withthe observed plagioclase compositionalvariations, the An/Ab ratio being sensitive to
XH2o in the melt. From experimental evidence(sisson and grove, 1993a, b), the Kd
Ca−na (= XCaPl
* Xnaliq / Xna
Pl * XCaliq) for plagioclase-melt
equilibria of high-alumina basalts and basalticandesites is strongly dependent on H2o contentsand relatively independent on pressure, as itincreases from 1.1 (anhydrous conditions) up to5.5 (H2o-saturated magmas).
on these grounds, the plagioclase-liquidCa/na (molar) diagram provides constraints onthe pre-eruptive XH2o conditions of the studiedproducts (Fig. 10). Concerning high-aluminabasaltic andesites, Kd
Ca−na values formicrophenocrysts and phenocryst rims vs bulkgroundmass compositions range from ca. 1.3 to2.3, most of the data points plotting between theexperimental Kd
Ca−na equilibrium lines foranhydrous (Kd
Ca−na=1.1) and 2 wt% H2o-hydrated melts (Kd
Ca−na=1.7; Fig. 10).Phenocryst cores vs whole rock equilibria yieldhigher Kd
Ca−na values (from ca. 2.6 to 4.5),corresponding to the interval between 2 wt%H2o (Kd
Ca−na=1.7) and the H2o-saturatedconditions (6 wt% H2o; Kd
Ca−na=5.52).Phenocryst cores vs whole rock data for selectedandesites (Fig. 10) reveal lower H2oconcentrations than in basaltic andesites
Petrogenesis of the High-Alumina Basalt-Andesite suite from Sant’Antioco Island, SW Sardinia, Italy 43
Fig. 7 - normal MorB-normalized diagram of trace element concentrations (a) and chondrite-normalized ree patterns (b)for representative sant’Antioco volcanics (HABA: high-alumina basaltic andesite; And: andesite).
A.M. Conte, d.M. PAllAdino, C. Perinelli and e. Argenti44
Fig. 8 - Variation diagrams of 87sr/86sr vs sio2 and selected trace elements for representative sant’Antioco mafic volcanics.data for volcanic rocks from Montresta (HMB: high-Mgo basalt, HAB: high-Al2o3 basalt, BA: basaltic andesite and And:andesite) and sarroch (basaltic andesite and andesite) districts (after Morra et al., 1997 and Conte, 1997) are also reported forcomparison.
Fig. 9 - representative analyses of pyroxenes from the sant’Antioco study suite, projected onto the di-en-Fs-Hd quadrilater(following lindsley, 1983). isotherms are shown for P = 500 MPa, chosen, in preference to 1 gPa or 1 atm, as the mostappropriate pressure conditions for magma crystallization in our case study (see text).
Petrogenesis of the High-Alumina Basalt-Andesite suite from Sant’Antioco Island, SW Sardinia, Italy 45
(comparable to the values obtained from crystalrims in these latter).
Furthermore, the relationships between thecalculated H2o contents in the melts andplagioclase compositions for some of the leastevolved lavas (i.e., sA7, sA40 and sA35) werechecked following silicate melt modeling(ghiorso et al., 1994) for different amounts ofdissolved water in the system (1-6 wt%), withinthe load pressure interval of 100-400 MPa.Model results broadly confirm the aboveestimated water contents, the latter being themost critical parameter controlling Kd values(rather than mere load pressure). For example, arun at P=100 MPa, t=1110°C and 2 wt% H2ofits the calculated Kd
Ca−na for phenocryst coresvs whole rock for sample sA7, while a run atP=200 MPa, t=1040°C and 5 wt% H2o yields aKd
Ca−na matching that obtained for sample sA35.Finally, from pilot experiments performed on the
tABle 7Calculated water contents (H2O wt%) for selected
Sant’Antioco samples, following the methods of Pichavant
and McDonald (2007) and Putirka (2008).
Fig. 10 - Comparison of [Ca]/[na] (molar) in plagioclase and coexisting liquid for selected samples from sant’Antiocovolcanics (HABA: high-alumina basaltic andesite; And: andesite). the three samples representative of the least evolvedcompositions (i.e., sA7, sA35 and sA40) are marked. each bar represents the range of values for a single rock-sample.Crystal-liquid equilibria are referred to either phenocryst cores vs whole rock compositions or phenocryst rims andmicrophenocrysts vs bulk groundmass compositions. straight lines represent experimental equilibrium Kd
Ca−na values foranhydrous (Kd
Ca−na=1.1) to H2o-saturated (KdCa−na=5.5) melts after sisson and grove (1993a).
rock Pichavant and Putirka samples Macdonald (2008) (2007)
A.M. Conte, d.M. PAllAdino, C. Perinelli and e. Argenti46
sant’Antioco least evolved rocks, it has beeninferred that sA35 experienced crystallizationunder significantly higher PH2o with respect tosA40 and sA7 (Conte et al., 2000).
PetrogenetiC inFerenCes
the paucity of primitive magma compositionsin subduction-related volcanism worldwideprecludes the assessment of the origin of primarymagmas, and makes it difficult to distinguish theeffects of subcrustal vs crustal processes.similarly, as concerns the rather differentiatedlavas from sant’Antioco, only low-pressureprocesses involved in their petrogenesis can beaddressed here. indications on source-relatedprocesses can be only tentatively drawn bycomparison with less-evolved Cenozoic calc-alkaline rocks of sardinia.
By considering the new set of minero-petrographic and geochemical data, in the lightof the huge amount of literature studies on high-alumina arc magmatism, it appears that the widespectrum of low-Mgo (and low Mg#, ni, Cr),high-alumina rock-types from sant’Antiocolikely reflects extensive fractional crystallizationof a parental basaltic liquid, where changes in thefractionating mineral assemblages occurred inresponse to variable PH2o conditions (sisson andgrove, 1993a, b; Pichavant and Macdonald,2007 and references therein; ulmer, 2007 andreferences therein), coupled with crustalinteraction at some extent (as suggested by traceelement and sr isotope data; geochemistrysection). on the other hand, the lack of typicaldisequilibrium textures (i.e., rim resorption,mantling by other phases, reverse or oscillatoryzoning) in pyroxene crystals does not supportadditional low-pressure processes, such asmixing between compositionally differentmagmas. in this regard, plagioclase dustytextures and zoning patterns are not consideredas conclusive evidence of magma mixing andare, instead, thought to be essentially controlled
by H2o conditions (also cfr. lonis et al., 1997).even if, in some cases, the observed plagioclasedisequilibrium textures and the occurrence ofpolycrystalline aggregates in li andesites maypoint out some crystal recycling in the magmareservoir, we exclude that this process may havesignificantly affected the chemical evolution ofthe sant’Antioco magmas. in this regard, theobserved wide scattering in Al2o3 and Cao vs
Mgo contents (Fig. 5) would imply unlikelyhigh amounts (some tens vol%) ofplagioclase+pyroxenes-bearing glomeri (also cfr.Brotzu et al., 1997a), which do not actuallyoccur in the study volcanics. Moreover, traceelement patterns (e.g., the absence of sr, eu andCr positive anomalies) provide no evidence ofsignificant cumulitic processes.
in order to constrain in more detail thepetrogenetic processes involved in the studiedrock suite, we focus now on the least evolvedcompositions (i.e., sA7, sA40 and sA35, high-alumina basaltic andesites), which arecharacterized by different Mgo, Al2o3, K2o andFeo contents, in spite of similar sio2 contents(around 52 wt%). in Fig. 5, the Al2o3-richestsample, sA35, also displays slightly higher Caoand lower Feo contents with respect to theMgo-richest sample, sA7, while sample sA40shows intermediate contents. trace elementcontents are quite similar, except for higher sr insA35 and slightly higher ni and Cr in sA7(tABle 6).
in addition, the three samples are characterizedby distinctive phenocryst assemblages, althoughall plagioclase-dominated (tABle 2): the Al2o3-richest, phenocryst-poorest, sample sA35contains the lowest plagioclase amount, andolivine plus rare Cpx as mafic phases; sA40,with intermediate Al2o3 content, shows higherabundances of plagioclase and mafic phases,with olivine again prevailing on Cpx; in sA7, theAl2o3-poorest and Mgo-richest sample, Cpxprevails on olivine, and minor amounts of opxare also observed.
Petrogenesis of the High-Alumina Basalt-Andesite suite from Sant’Antioco Island, SW Sardinia, Italy 47
the three samples also differ in mineralchemistry, being characterized by distinctplagioclase compositions (i.e. An75-85, An78-90 andAn85-93 for sA7, sA40 and sA35, respectively;Fig. 2) and different Al2o3 contents in Cpx (from~2wt% in sA7 up to ~6wt% in sA35), whichparallel whole-rock Al2o3.
these differences likely indicate thatcorresponding magmas underwent different pre-eruptive conditions of storage andcrystallization, consistent with the aboveestimated values of temperature, pressure ofcrystallization and dissolved water content forthe three cases. in particular, plagioclase-hostliquid equilibria (Fig. 10) provide quantitativeinformation on H2o contents in the threemagmas at different evolutionary stages (theliquid compositions being roughly representedby either the whole-rock or groundmass bulkcompositions with ongoing magmacrystallization).
Water content data broadly fit the above-reported mineralogical features of samples. infact, it has been proven that high dissolved H2oconcentrations in high-alumina melts, not onlyfavour the early formation of spinel and thecrystallization of An-richer plagioclase, but alsoreduce the total proportion of plagioclase in thecrystallizing assemblage, expand the stabilityfield of olivine at the expense of other maficphases such as Cpx and opx, and decrease theen-Fs contents of Cpx in favour of Cats(Kushiro, 1969; sisson and grove, 1993a;Pichavant and Macdonald, 2007).
these considerations lead us to conclude thatchanges in phenocryst proportions andcompositions (and consequent changes in bulkrock chemistry) in the least evolved basalticandesites were determined by different H2oconcentrations in the melts and by the physicalconditions controlling water exsolution from thesystem during magma ascent. in summary: theAl2o3-richest, Mgo-poorest sample (sA35)displays, consistent with the highest XH2o (up to
ca. 6 wt%, Fig.10), the lowest modal amount ofplagioclase, accompanied by the Ca-richestcomposition, a paucity of Cpx, with the Ca-ts-richest augite composition, and the presence ofolivine. Conversely, the Mgo-richest, Al2o3-poorest composition of sA7, may reflect acrystallization path under H2o-poorer conditions(up to ca. 3 wt%), which favoured thecrystallization of higher amounts of Ca-poorerplagioclase and Al2o3-poorer Cpx, as well as theoccurrence of opx at the expense of olivine.sA40, which shows intermediate Al2o3 andMgo contents, consistently accounts forintermediate H2o concentration in the magma(up to ca. 4 wt%). in this picture, the Ab-richerplagioclase rims and microphenocrysts in thethree samples may testify late H2o exsolution.
Moreover, modes of water retention/exsolutionin the system influence the amphibole stabilityfield. notably, amphibole only occurs in two lisamples, which records prolonged H2o retentionin the corresponding magmas; late-stageprocesses of decreasing pressure and H2oexsolution, causing amphibole break-down, aretestified by the presence of opacitic rims.
the effect of XH2o on phase equilibria thusmay represent the key factor in determining theobserved geochemical differences among theleast evolved sant’Antioco rock-types, as aresult of different evolution paths from a possiblecommon parental melt.
Identification of the parental magma
As it is typical of arc-related magmatismworldwide (e.g., nye and reid, 1986; Brophy,1989; eggins, 1993; sisson and grove, 1993b;rohrbach et al., 2005; Pichavant et al., 2002;Pichavant and Macdonald, 2007), in Cenozoicorogenic magmatism of sardinia, high-magnesiabasalts (HMBs, Mgo >8 wt%), possiblyrepresentative of mantle-derived magmas, arerelatively uncommon and limited to a few areas(Mattioli et al., 2000 and references therein). in
A.M. Conte, d.M. PAllAdino, C. Perinelli and e. Argenti48
particular, HMBs from the Montresta district,nW sardinia (18.5 Ma, Montigny et al., 1981;Morra et al., 1997; Fig. 1) could match incomposition the possible precursor forsant’Antioco lt basaltic andesites. in fact, thehigh-Al2o3 basaltic andesites derived fromMontresta HMBs through fractionalcrystallization±assimilation processes (Morra et
al., 1997) show major and trace elementcompositions that closely resemble those fromsant’Antioco (Fig. 11, 12), except forsignificantly lower Ba contents (e.g., 140-250 vs
300-400 ppm, for Montresta and sant’AntiocoHABA, respectively; Fig. 8).
if a typical Montresta HMB (e.g., sampleKB13, Mgo = 9.08 wt%, Morra et al., 1997) ischosen as representative of the sant’Antiocoparental magma composition, major elementmass balance calculations (following stormerand nicholls, 1978) broadly indicate thatsamples sA7, sA40 and sA35 may beconsistently derived by fractional crystallizationunder different PH2o conditions (tABle 8).notably, the results yield similar fractions ofresidual liquid for the three cases (~45-50 wt%),but different proportions of the main crystalphases (especially plagioclase vs mafic phases),i.e.: Plg/(ol+Cpx)=0.52, 0.36 and 0.24 for sA7,sA40 and sA35, respectively.
the similar fractions of residual liquid mayaccount for the similar contents in the majority ofincompatible trace elements in the three cases,whereas the observed differences in the srcontents can be explained by the different amountsof plagioclase fractionation (Fig. 6; 7a, b).
it is also worth noting that relatively high XH2o
conditions ab initio, which determine thefractionation of high proportions of mafic phases(with high Mgo/Feo* ratios), with respect toplagioclase, may even result in the observedambiguous tholeiitic character of thesant’Antioco mafic volcanics (Fig. 4). in ourinterpretation, this inference could also apply tothe “tholeiitic” lavas in the Montresta district
(Morra et al., 1997).
Magma evolution
From the whole compositional spectrum of thestudied sant’Antioco volcanics it appears thatthe differences existing in the least evolvedcompositions also reflect in the differentiationproducts (Fig. 5, 11). Major-element mass-balance calculations (tABle 9) show that thelatter can be derived through prolongedfractional crystallization of Plg + Cpx + Fe-tioxides ± ol ± opx mineral assemblages. in moredetail, if sA35 and sA7 lava-types, whichdisplay the most significant differences in termsof oxide concentrations and H2o contents, areconsidered as starting melts, the results ofcalculations evidence that olivine continues toplay a role in the early stages of the H2o-richdifferentiation trend starting from sA35 (Fig.11), while opx may replace olivine in thefractionating crystal assemblage along the H2o-poor differentiation trend starting from sA7 (Fig.11, tABle 9). these results confirm, also inagreement with experimental evidence(Pichavant and Macdonald, 2007 and referencestherein), that in mafic to intermediate calc-alkaline magmas opx is stable only under lowH2o concentrations, while in more hydrousconditions opx crystallization is delayed andtakes place extensively only in sio2-richermagmas at lower temperatures. indeed, in thesant’Antioco volcanics, like in other typical arc-rock suites (e.g., Macdonald et al., 2000), opxis always present in the most differentiated rocks(sio2 >55 wt%; tABle 9). in more hydrousmelts, opx crystallization is joined by amphiboleduring the late differentiation stages.
However, simple fractional crystallizationcannot fully explain the whole compositionalspectrum of the study rock suite. in this regard,the above reported 87sr/86sr data (0.70698 to0.70858±1; tABle 2; Fig. 8) are in the 0.70398-0.71130 range pertaining to the Cenozoic
Petrogenesis of the High-Alumina Basalt-Andesite suite from Sant’Antioco Island, SW Sardinia, Italy 49
Fro
m H
MB
-KB
13 (
Montr
esta
dis
tric
t) t
o H
AB
A’s
(Sant’
Anti
oco
dis
tric
t)
fro
m K
B13
to s
A35
f
rom
KB
13 to
sA
40
fro
m K
B13
to s
A7
K
B13
*
s
A35
res
iduals
sA
40
r
esid
uals
sA7
res
iduals
sio
2
48.8
8
52.5
4
0.
02
52.0
6
0.
00
52.3
4
0.0
4ti
o2
0
.75
1
.04
0.00
1
.09
0.01
1
.07
0
.01
Al 2o
3
1
5.55
21
.02
0.00
20
.00
0.00
19
.22
0
.02
Feo
*
1
0.16
8
.27
0.02
9
.27
0.04
9
.38
0
.05
Mno
0
.12
0
.16
0.00
0
.18
0.02
0
.18
0
.01
Mgo
9
.08
3
.14
0.00
3
.63
0.02
4
.72
0
.00
Cao
12
.90
10
.63
0.01
10
.08
0.01
10
.11
0
.01
na 2
o
1.81
2
.40
0.14
2
.53
0.05
2
.12
0
.26
K2o
0
.52
0
.58
0.05
0
.98
0.00
0
.70
0
.04
P 2o
5
0.2
3
0.2
1
0.
02
0.1
8
0.
02
0.1
6
0.0
3 10
0.00
100
.00
100
.00
10
0.00
∑
r2
0.25
0.1
6
0.4
8
R
emove
d c
ryst
als
(w
t%)
ol
11.8
9
10.
46
8
.17
C
px
28.
41
28.
18
27.
51
Plg
9.5
7
13.
74
18.
66
Fe-t
i oxi
des
0
.10
0.
75
wt%
of
resi
dual
liqu
id
50.
03
47.
62
44.
91
C
px/o
l
2.
39
2.
70
3
.37
C
px+o
l
40.3
0
38.
63
35.
68
Plg/
(Cpx
+ol)
0.
24
0.
36
0
.52
∑r2
:su
m o
f sq
uare
d re
sidu
als;
oth
er a
bbre
viat
ions
as
in t
AB
le
2; *
com
posi
tions
of
who
le r
ock
and
min
eral
pha
ses
from
Mor
ra e
t
al.
(199
7). F
or e
ach
step
of
mas
s ba
lanc
e, th
e co
mpo
sitio
ns o
f re
mov
ed c
ryst
als
are
take
n fr
om p
heno
crys
t pha
ses
(usu
ally
rim
com
posi
tions
) fr
om th
e co
rres
pond
ing
pare
nt r
ocks
(se
e t A
Bl
e3,
4, 5
).
tAB
le
8M
odel
ed f
ract
ionati
on s
teps
(foll
ow
ing S
torm
er a
nd N
icholl
s, 1
978)
for
the
least
evo
lved
basa
ltic
andes
ites
fro
m S
ant’
Anti
oco
, st
art
ing f
rom
a t
ypic
al
hig
h-m
agnes
ia
basa
lt o
f M
ontr
esta
, in
ferr
ed a
s poss
ible
pare
nta
l m
agm
a.
A.M. Conte, d.M. PAllAdino, C. Perinelli and e. Argenti50
Fro
m S
A7 t
o S
A15b t
o S
A32b
F
rom
SA
35 t
o S
A16 t
o S
AV
18
sA7
sA15
b
r
esid
uals
sA32
b
res
idua
ls
s
A35
sA
16
r
esid
uals
sA
V18
r
esid
uals
sio
2
5
2.34
56.
68
0
.00
6
0.88
0.01
si
o2
52.
54
5
4.50
0.01
5
8.49
0
.04
tio
2
1.0
7
0.95
0.1
2
0.
73
0.
10
tio
2
1.
04
1.0
5
0.
02
0
.70
0
.17
Al 2o
3
19
.22
18.
43
0
.00
1
7.57
0.00
A
l 2o3
21.0
2
2
0.20
0.00
1
9.88
0
.02
Feo
*
9.38
8.
11
0
.00
6.81
0.01
Fe
o*
8
.27
7
.96
0.00
6.3
2
0.0
4M
no
0.18
0.
15
0
.00
0.13
0.00
M
no
0.
16
0.1
4
0.
00
0
.12
0
.00
Mgo
4.
72
3.38
0.0
0
2.
40
0.
00
Mgo
3.14
2
.43
0.00
1.7
4
0.0
0C
ao
1
0.11
8.
32
0
.00
6.90
0.00
C
ao
10.
63
9.5
3
0.
00
8
.60
0
.02
na 2
o
2.12
2.
59
0
.01
2.64
0.03
n
a 2o
2
.40
2
.81
0.00
3.0
0
0.0
0K
2o
0.7
0
1.16
0.0
0
1.
67
0.
01
K2o
0.5
8
1.1
0
0.
10
0
.90
0
.17
P 2o
5
0.1
6
0.22
0.0
0
0.
25
0.
00
P 2o
5
0.
21
0.2
6
0.
00
0
.25
0
.01
100
.00
1
00.0
0
100
.00
s
um
100.
00
10
0.00
1
00.0
0
∑
r2
0
.13
0.16
∑
r2
0.1
5
0
.48
R
emove
d c
ryst
als
(w
t%)
R
emove
d c
ryst
als
(w
t%)
ol
o
l
2.
97
Cpx
6
.84
3.3
8
Cpx
2
.58
3.71
opx
6.
28
4
.21
o
px
2.
93
Pl
g
2
3.7
16.
23
Plg
1
2.46
1
7.94
Fe-t
i oxi
des
2.8
6
2
.19
o
x
0.3
5
2.
88w
t% o
f re
sidu
al li
quid
39
.68
73.9
9
wt%
of
resi
dual
liqu
id
81.6
4
72.
54
∑
r2:
sum
of
squa
red
resi
dual
s; o
ther
abb
revi
atio
ns a
s in
tA
Bl
e2.
Add
ition
al e
xpla
natio
ns a
s in
tA
Bl
e8.
tAB
le
9M
odel
ed f
ract
ionati
on s
teps
(foll
ow
ing S
torm
er a
nd N
icholl
s, 1
978)
for
the
basa
ltic
andes
ites
and a
ndes
ites
of
the
Sant’
Anti
oco
suit
e, s
tart
ing f
rom
SA
7 a
nd S
A35
com
posi
tions,
as
dis
cuss
ed i
n t
he
text
.
Petrogenesis of the High-Alumina Basalt-Andesite suite from Sant’Antioco Island, SW Sardinia, Italy 51
orogenic rocks of sardinia (dupuy et al., 1974,1979; Coulon, 1977; Brotzu et al., 1997a, b;Conte, 1997; Morra et al., 1997; downes et al.,2001), for which different degrees of crustalcontamination, coupled with variable enrichmentof mantle source by subducted sediments, hasbeen proposed. At sant’Antioco, the observedrange of sr isotope data, the positive correlationof 87sr/86sr vs elements with crustal affinity (e.g.,si, rb, Ba; Fig. 8), as well as the wide scatteringof trace elements (Fig. 6), strongly suggestcontamination by crustal wall rocks. inparticular, the sant’Antioco 87sr/86sr values aresignificantly higher with respect to thechemically analogous products of Montresta,whereas they overlap the higher values from the
nearby sarroch volcanic district (Fig. 8). in thelatter, crustal assimilation has been revealed andquantified in about 4-14 wt% (Conte, 1997),where the assimilated material considered, i.e. anaverage composition of the granitoid batolith ofsardinia (from del Moro et al., 1975; secchi et
al., 1991), would account for the observed traceelement enrichment (e.g., Ba, rb and Zr) anddepletion (e.g., sr) trends. Besides theoversimplified assumption concerning thecrustal material involved, considering the similar87sr/86sr ratios and trace elements behaviourobserved for sant’Antioco and sarrochvolcanics, we infer that crustal contaminationaffected magma evolution at sant’Antioco atanalogous, or even slightly higher, extent.
ConCluding reMArKs
the present work integrates the petrologicalknowledge on the Cenozoic orogenicmagmatism of western sardinia, with focus onmafic to intermediate rock-types. new data on
Fig. 11 - Plot of Cao (a) and Al2o3 (b) vs Mgo contents forthe investigated sant’Antioco volcanics, compared withMontresta (data after Morra et al., 1997; rock abbreviationsas in Fig. 8). Arrows represent possible differentiation paths,as considered in mass balance calculations (tABle 8, 9).samples marked are discussed in the text.
Fig. 12 - representative ree patterns for high-Al2o3 basalticandesites (HABA) and andesites (And) from sant’Antiocoand Montresta (shaded field; data after Morra et al., 1997).the pattern for the Montresta high-Mgo basaltic sample(HMB, KB13), considered as parental magma in petrogeneticmodeling, is also reported (data after Morra et al., 1997).
A.M. Conte, d.M. PAllAdino, C. Perinelli and e. Argenti52
sant’Antioco relatively mafic volcanics, mostlyrepresented by basaltic andesites and andesites,reveal a remarkably wide compositionalspectrum in terms of Al2o3, Mgo and Caocontents, as related to the characterizing mineralassemblages. the observed variability in wholerock and mineral chemistry likely correlates todifferent liquid lines of descent, controlled by thePH2o histories of magma systems, starting from acommon parental magma composition. the latterhas been identified as a high-Mgo basalt similarin composition to the broadly coeval Montrestaprimitive lavas. different from the Montrestadistrict, the sant’Antioco area is characterizedby the lack of primitive terms, possiblyindicating different local tectonic regimefavouring magma ponding and differentiation atupper crustal levels.
the overall geochemical features of thesant’Antioco study rock-types evidence a clearcalc-alkaline signature, despite relatively highvalues of Feo*/Mgo that could mimic amisleading tholeiitic affinity. starting from thecomposition of a Montresta HMB, considered asa suitable primitive (possibly near-primary)magma, two end-member differentiation trendsare recognized, which lead respectively to theMgo-richest/Al2o3-poorest and Mgo-poorest/Al2o3-richest basaltic andesites, representing theleast evolved rock-types at sant’Antioco. thesetrends can be respectively explained by higherproportions of plagioclase vs olivine+Cpx in thefractionating assemblage under H2o-poorerconditions and vice versa in H2o-richerconditions. Basically, the XH2o imprint isrecorded up to the most evolved compositions,being also the main controlling factor ofincoming opx crystallization.
Finally, sr isotope data and the trends of traceelements of “crustal affinity” with increasingdegree of magma evolution, indicate thatdifferent degrees of crustal contamination alsoplayed a significant role in the petrogenesis ofsant’Antioco volcanics, consistent with magma
slow ascent and ponding within a tick crust inmultiple reservoirs and/or feeder conduits thatbehaved independently.
ACKnoWledgeMents
sincere thanks to F. secchi and an anonymousreferee for their meaningful and constructive review.
APPendiX A: AnAlytiCAl MetHods
Whole-rock samples were analyzed for majorand selected trace elements (V, Cr, Co, ni, rb,sr, y, Zr, nb, Ba, th) by a Philips PW1480-XrFspectrometer at igg-Cnr, rome, using fuseddishes (major element) and pressed powderpellets (trace elements). Philips X40 softwarewas used for data reduction. Analytical precisionis generally better than 3% and 5% , for majorand trace elements, respectively.
loi was determined after the samples hadbeen kept at 1050°C for 2 h; resulting losseswere corrected for Feo oxidation and Feo wasmeasured by titration with KMno after aciddigestion.
ree, ta, Hf and u were determined by iCP-Ms at the Activation laboratories, Ancaster,Canada.
sr isotope ratios were measured, after sampledissolution and ion exchange separationprocedures, on a Vg354 double collector massspectrometer at the university Federico ii,naples. repeated analyses of nBs-987international reference standard gave an averagevalue of 0.71027±0.00001 (2s).
Mineral analyses were carried out using aCameca sX50 electron microprobe with anaccelerating voltage of 15 keV and 15 nA beamcurrent at C.n.r.-igAg, rome, using the ZAFcorrection procedure.
reFerenCes
ArAMAKi s. and KAtsurA t. (1973) - Petrology and
Petrogenesis of the High-Alumina Basalt-Andesite suite from Sant’Antioco Island, SW Sardinia, Italy 53
liquidus temperature of the magma of the 1970
eruption of Akita-Komagatake Vocano, northeastern
Japan. J. Assoc. Min. Petrol. econ. geol., 68, 101-124.
ArAñA V., BArBeri F. and sAntACroCe r. (1974) -Some data on the comendite type area of S. Pietro
and S. Antioco islands, Sardinia. Bull. Volcanol., 38,1-12.
AssorgiA A., FAddA A., giMeno torrente d., MorrA
V., ottelli l. and seCCHi F. (1990) - Le successioni
dostAl J., Coulon C. and duPuy C. (1982) -Cainozoic andesite rocks of Sardinia (Italy). in:thorpe r.s. (ed.), Andesites: orogenic Andesitesand related rocks. Wiley, Chichester, pp. 353-370.
doWnes H., tHirlWAll M.F. and trAyHorn s.C.(2001) - Miocene subduction-related magmatism in
southern Sardinia: Sr-Nd and oxygen isotopic
evidence for mantle source enrichment. J. Volcanol.geotherm. res., 106, 1-21.
duPuy C., MC nutt r.H. and Coulon C. (1974) -Déterminasion de 87Sr/86Sr dans les andésites
Pioli l. and rosi M. (2005) - Rheomorphic structures
in a high-grade ignimbrite: The Nuraxi tuff, Sulcis
volcanic district (SW Sardinia, Italy). J. Volcanol.geotherm. res., 142, 11-28.
Pioli l., lAnZA r., ort M. and rosi M. (2008) -Magnetic fabric, welding texture and strain fabric in
the Nuraxi Tuff, Sardinia, Italy. Bull. Volcanol., 70,1123-1137.
PutirKA K.d. (2008) - Thermometers and Barometers
for volcanic Systems. in: Putirka K.d. and tepley F.J.iii (eds.), Minerals, inclusions and volcanicprocesses, reviews in Mineralogy & geochemistry,69, the Mineralogical society of America,Washington d.C., pp. 61-120.
roHrBACH A., sCHutH s., BAllHAus C., MünKer C.,MAtVeeV s. and QoPoto C. (2005) - Petrological
implications for mantle composition and processes.in: saunders A.d. and norry M.J. (eds.),Magmatism in the ocean Basins. geol. soc. lond.spec. Publ., 42, 313-346.
ulMer P. (2007) - Differentiation of mantle-derived
calc-alkaline magmas at mid to lower crustal levels:
experimental and petrologic constraints. Per.Mineral., 76, 309-325.