Earth and Planetary Science Letters 281 (2009) 70–80

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

j ourna l homepage: www.e lsev ie r.com/ locate /eps l

Helium isotopes in hydrothermal volcanic fluids of the Azores archipelago

P. Jean-Baptiste a,⁎, P. Allard b, R. Coutinho c, T. Ferreira c, E. Fourré a, G. Queiroz c, J.L. Gaspar c

a LSCE, CEA-CNRS-UVSQ, CEA/Saclay, 91191 Gif-sur-Yvette cedex, Franceb Laboratoire Pierre Sue, CNRS-CEA, CEA/Saclay, 91191 Gif-sur-Yvette cedex, Francec Centro de Vulcanologia e Avaliação de Riscos Geologicos, Universidade dos Açores, Ponta Delgada, Portugal

⁎ Corresponding author. Tel.: +33 169087714; fax: +E-mail address: Philippe.Jean-Baptiste@cea.fr (P. Jea

0012-821X/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.epsl.2009.02.009

a b s t r a c t

a r t i c l e i n f o

Article history:

We present the first helium Received 9 July 2008Received in revised form 4 February 2009Accepted 6 February 2009Available online 12 March 2009

Keywords:helium isotopesAzoreshydrothermal fluids

isotope data for thermal waters and gas emissions on the islands of Terceira,Graciosa, Faial, Pico and Flores, as well as new data for Sao Miguel. The results allow us to track current mantledegassing associated with the Azores hot spot, to delineate its spatial distribution and to discuss its possibleorigin. As a general rule, we find that free gases tend to display somewhat higher 3He/4He ratio thangroundwaters.We argue that this difference is likely due to radiogenic helium inputs to aquifers duringwater–rock interactions and, therefore, that gas phases are the fluid carriers with the most representative of mantlesource signature. The measured 3He/4He ratios (normalized to the air ratio, Ra) range from lower-than-MORBvalues (5.23–6.07 Ra) on central SaoMiguel, toMORB values on Faial (8.53 Ra) and Flores (8.04 Ra) – located oneither side of the Mid-Atlantic Ridge – and to plume-type values on Graciosa (11.2 Ra) and Terceira (13.5 Ra)where free gases also display ten times higher-than-MORB CO2/3He ratios (1.8–2.6×1010). Such a wide Heisotopic range and its spatial distribution corroborate with available data for volcanic rocks, indicating thatplume's head presently underlies the central part of the archipelago. The plume-type 3He/4He ratios onTerceira and Graciosa agreewith geochemical and seismic evidence of a deep-rootedmantle plume feeding theAzores hot spot. Our finding that high 3He/4He ratios correspond to low 3He concentrations and high (arc-type) CO2/3He values exclude a simple plume supply of 3He-rich primitivemantle. Instead, the simultaneity ofboth elevated CO2/3He and 3He/4He ratios is best explained by a 3He-rich contribution from the lower mantlediluted in a CO2-rich feeding plume that contains a recycled altered oceanic plate component. The alternativepossibility of an enhanced time-integrated 3He/(U+Th) ratio in the Azores plume due to a greatercompatibility of helium relative to U and Th during melting events is difficult to reconcile with the enrichedpattern of volcanic rocks from the central islands. In any case, the Azores plume should derive from a mantlereservoir that could escape convective homogenization for a very long period of time, in agreement with sub-chondritic osmium isotopic ratios in volcanic rocks from the central islands of the archipelago.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

The Azores archipelago, composed of nine active volcanic islands(Fig. 1), is the emerged part of a huge oceanic basaltic province (seeGente et al., 2003 for a recent review), located at the Triple Junction ofthe North American, the African and the Eurasian plates between 35°and 40°N. The Eurasian/African branch of the Triple Junction is acomplex system of tectonic faults and ridges (Searle, 1980; Vogt andJung, 2004). It forms a linear structure trending ESE-WNW on top ofwhich are found most of the Azores Islands, except Flores and Corvowhich are located to the west of the Mid-Atlantic Ridge (Fig. 1). TheAzores basaltic province has been a region of major ocean floor uplift,intense seismicity and extensive volcanism since about 5Ma ago (Boothet al., 1978). More than twenty volcanic eruptions have occurred sincethe 15th century (Booth et al.,1978), and themost recent one happenedin 1998–2001 offshore of Terceira Island (Gaspar et al., 2003). Active

33 169087716.n-Baptiste).

ll rights reserved.

volcanism is also evidenced by numerous persistent gas emissions andhydrothermal manifestations on the islands as well as on the seafloor(Jean-Baptiste et al., 2008).

The Azores intra-plate oceanic basaltic (OIB) magmatism whichdeveloped close to the Mid-Atlantic Ridge, is a typical example ofhotspot–ridge interaction (Schilling, 1975; Montagner and Ritsema,2001). Although initially considered as a low 3He/4He hotspot(Tolstikhin et al., 1991; Kurz et al., 1982a), its basalts were subsequentlyfound to display 3He/4He ratios both above and below the MORBdomain (Moreira et al., 1999; Madureira et al., 2005). Ocean islandbasalts worldwide show awide range of 3He/4He values between b5 Raand 50 Ra — where Ra is the atmospheric 3He/4He ratio (Kurz et al.,1982b; Farley and Neroda, 1998; Graham, 2002; Stuart et al., 2003),which sharply contrasts with the relative homogeneity (8±1 Ra) ofMid-Ocean Ridge Basalts (MORB) derived from the upper mantle.Classically, high 3He/4He ratios have been attributed to plume uprise ofprimitive material supposedly residing in the lower mantle (Craig andLupton, 1976; Lupton, 1983; Allègre et al., 1983; Allègre and Turcotte,1985), whereas lower-than-MORB ratios are thought to reflect the

Fig. 1. Bathymetry of the Azores Plateau and location map of the nine islands of the Azores archipelago (Lourenço et al., 1998).

71P. Jean-Baptiste et al. / Earth and Planetary Science Letters 281 (2009) 70–80

recycling of crustal material (sediments, oceanic crust or continentallithosphere) with high time-integrated (U+Th)/3He ratios (Kurz et al.,1982b; McKenzie and O'Nions, 1983; Hanyu and Kaneoka, 1997) or, insome cases, shallow crustal contamination during magma ascent andponding (Zindler and Hart, 1986; Hilton et al., 1995). However, even thenature of the high 3He/4He OIB source itself remains controversial (e.g.Anderson,1998). Recent advances in seismological studies (e.g. Van derHilst et al., 1997; Grand, 2002) have shown that oceanic plates can besubducted down to the base of the lower mantle and, therefore, haveraised new questions on the reality of the canonic two-layered mantle,opening the way to alternative scenarios. In particular, recent measure-ments of the partition coefficients of noble gases and U–Th duringmelting and crystallisation (Parman et al., 2005; Heber et al., 2007) haveled to the alternative hypothesis that high 3He/4He ratios inOIB samplescould track recycled upper mantle residues left behind with higher-than-MORB time-integrated 3He/(U+Th) ratios rather than primitivemantle reservoirs (Parman et al., 2005).

Here,wepresent new 3He/4Hedata for bothhydrothermal gases andwaters in the Azores archipelago. Up to now, only a few helium isotopedata were available for fluids in Sao Miguel island (Tolstikhin et al.,1991). In addition to providing additional results for SaoMiguel, herewereport the first He isotope data for hydrothermal fluids on five otherislands: Terceira, Graciosa, Faial, Pico and Flores. The results allow us tocharacterize current mantle degassing across the Azores domain and,togetherwithavailableHe isotopedata for the local volcanic rocks (Kurz,1991; Moreira et al., 1999; Madureira et al., 2005), to delineate thehorizontal extent of the mantle plume feeding the Azores hot spot.

2. Sample location and methods

Our study of the main hydrothermal features of the volcanic archi-pelago concerned twentyfive sites on the islands of SaoMiguel, Terceira,Graciosa, Pico, Faial and Flores. In Sao Miguel, Terceira and Graciosaislands, the main hydrothermal manifestations and degassing fieldsoccur on central volcanoes. The other islands are characterized byminoror more punctual degassing sites (Ferreira et al., 2005). Fluids werecollected from natural manifestations (fumaroles, solfataras, thermal

springs and pools, gas bubbles in water), geothermal production wells,and cold water wells (Fig. 2).

Water samples were taken in copper tubes equipped with metalclamps at both ends. The tubes were flushedwith water prior to closureusing a peristaltic pump connected with Tygon tubing, except forsamples taken at geothermal well heads whichwere flushed directly bythe water flow. Fumarolic and solfataric gases, as well as gas bubbles inwater – including one submarine sample taken by scuba diving off thecoast of Ribeira Quentes (SaoMiguel) –were collected in pre-evacuatedglass bulbs with stopcocks using an inverted funnel. All gases areessentially composed of carbon dioxide (up to 99.7 vol.%), plus minoramounts of N2, CH4, and ofH2S in fumarolic gases (Pasquier-Cardin et al.,1999; Ferreira and Oskarsson, 1999; Ferreira et al., 2005).

The isotopic analyses were made at LSCE — Saclay (Jean-Baptisteet al., 1992). Helium and neon isotopes in the gas samples were directlymeasured on a MAP-215 mass spectrometer connected to a high-vacuum inlet system. Helium dissolved in the water samples was firstextracted under vacuum into sealed glass tubes and then analysed on aVG-3000mass spectrometer (neonwas notmeasured on these samplesdue to Hall probe problems on the mass spectrometer at that time).Typical He blank for both spectrometers is b5–6×10−10 cm3 STP. Theaccuracy is 1% for 4He concentrations inwater samples and 2.5% for 4Heand 20Ne in gas samples (due to the additional uncertainty in aliquot'scalibration). The analytical uncertainty on measured 3He/4He ratios,calibrated against an atmospheric air standard, is better than 0.8%.

Helium concentrations and isotopic ratios in gas samples werecorrected for possible air contamination using the measured 20Ne/4Heratios and considering a pure atmospheric origin for neon, as classicallydone, since magmatic 20Ne/4He ratios are very much lower than theatmospheric ratio (20Ne/4He)a — Craig and Lupton (1976); Madureiraet al. (2005). The air-corrected 3He/4He ratio, Rc, is thus obtained fromthe following equation:

Rc = Ra = R= Rað Þ− 20Ne= 4He� �

=20Ne=4He

� �a

h i

� 1− 20Ne= 4He� �

=20Ne=4He

� �a

h i:

Fig. 2. Fluid sample locations on the islands of Flores, Graciosa, Faial, Terceira, Pico and Sao Miguel islands.

72 P. Jean-Baptiste et al. / Earth and Planetary Science Letters 281 (2009) 70–80

As no neon data are available for the water samples (except forsample 20 from Terceira which was sampled later on and analysed onthe MAP-215), the 3He/4He ratio (Rm) of the dissolved magmatichelium component is computed as follows:

The measured helium isotope concentrations can be divided inthree components,

4Heh i

= 4Hesolh i

+ 4Heah i

exc+ 4Hem

h i

3Heh i

= 3Hesolh i

+ Ra4Heah i

exc+ Rm

4Hemh i

where [4Hesol] and [3Hesol] correspond to the solubility equilibrium atthe temperature and salinity conditions of the waters (Weiss, 1971;Potter and Clynne, 1978), [4Hea]exc is a possible helium excess ofatmospheric origin, and [4Hem] is the magmatic component. The airexcess component is usually noted as Δ/100×[4Hesol], hence Rm willbe given by the following formula:

Rm = 3Heh i

− 4Hesolh i

− RaΔ = 100 × 4Hesolh i� �

� 4Heh i

− 4Hesolh i

− Δ = 100 × 4Hesolh i� �

:ð1Þ

If the water samples are collected with great care in order to avoidany air bubble, as we did, the air excess Δ is typically of the order ofa few % only (e.g. Hohmann et al., 1998). In that case Eq. (1) is wellapproximated by:

RVm = 3Heh i

− 3Hesolh i� �

=4Heh i

− 4Hesolh i� �

ð2Þ

where Rm′ is the slope of the correlation between 3He and 4He con-centrations when plotting [3He]–[3Hesol] versus [4He]–[4Hesol]. Com-paring Eqs. (1) and (2) leads to the following equation linking the truevalue Rm and the approximate value Rm′ :

Rm = Ra = RVm = Ra − Δ= 100= 4Heh i

=4Hesolh i

− 1� �h i

� 1− Δ= 100= 4Heh i

=4Hesolh i

− 1� �h i

:ð3Þ

Eq. (3) shows that if the term Δ/100/([4He]/[4Hesol]−1) is muchless than 1, Rm′ and Rm tend to be identical. Practically, the value of theslope Rm′ is strongly influenced by the samples with the highest heliumcontent (corresponding in the present data set to ([4He]/[4Hesol]−1)values between 2.8 (on Sao Miguel) and 34.2 (on Flores). For an aircomponent excess in the upper range (Δ=5%) and the less favourable([4He]/[4Hesol]−1) value of 2.8, we thus calculate that the maximumdifference between Rm and Rm′ is on the order of 0.1 Ra. The reliability ofthis correction procedure is illustrated by sample 20 (for which the neonconcentration is available) where there is a good agreement betweenthe air-corrected 3He/4He ratio computed by the present procedure(10.31 Ra) and that based on the neon concentration (10.36 Ra).

3. Results

He isotope results for water and free gas samples are respectivelygiven in Tables 1 and 2. Fig. 3 shows scatterplots of 3He versus 4Heconcentrations for each island, as well as data points for the localvolcanic rocks when available. In the following, we first examine theresults island by island before discussing their overall regional pattern.

Table 1Helium isotope results for water samples.

Island Samplenumber

Sample location Type of manifestation T Conductivity pH R/Ra(uncorrected)

[4He] [4He]solub(°C) (μS/cm) (10−8 cm3 STP/g) (10−8 cm3 STP/g)

Sao Miguel 1 Lagoa das Furnas/North Boiling pool 98.7 4680 2.0 3.36 3.84 0.312 Lagoa das Furnas/East Bubbling waters 17.6 146.4 6.0 2.65 7.20 4.505 Furnas/Terra Nostra Bubbling waters 18.3 323 6.4 3.82 16.94 4.497 Ribeira Quente Submarine bubbling waters 15 – – 1.89 5.50 3.889 Caldeira Vehla Bubbling pool 86.5 6900 7.8 2.73 2.32 2.1510 Fogo/Ribeira Grande Geothermal plant CL3 86.1 23,200 7.9 1.01 2.32 2.16

Flores 11 Lajedo/Ponta Negra Warm spring 29.0 32,400 6.5 7.70 142.5 4.05Faial 12a Podo do Capelo Well 39.9 11,190 6.1 7.60 35.7 4.14

12b Podo do Capelo Well 39.9 11,190 6.1 7.57 30.0 4.1413 Varadouro Well 29.2 8980 7.1 2.83 6.21 4.28

Pico 14a Silveira Well 18.5 2560 6.6 4.74 8.20 4.4614b Silveira Well 18.5 2560 6.6 4.45 8.80 4.46

Terceira 17 Ladeira da pena/Fontinhas Well 15.6 167 8.2 0.98 5.37 4.5318 Pico Celeiro/Praia de Vitoria Well 18.1 703 7.9 1.26 6.60 4.4919 Juncal/Praia de Vitoria Well 22.9 1701 7.2 6.02 20.11 4.4220 Terra Cha Well 25 726 7.5 7.38 14.00 4.41

Graciosa 21a Baia dos Homiciados Bubbling hot spring 58.0 81,000 5.9 5.57 5.87 3.4421b Baia dos Homiciados Bubbling hot spring 58.0 81,000 5.9 5.36 4.14 3.4422a Furna do Enxofre Subterranean lake 14.7 729 6.7 7.68 52.59 4.5422b Furna do Enxofre Subterranean lake 14.7 729 6.7 7.86 61.96 4.5423a Furo Pontal Well 25.2 1152 8.4 4.48 8.85 4.4023b Furo Pontal Well 25.2 1152 8.4 4.15 7.99 4.4024a Furo Corelas Well 41.9 710 5.1 6.59 20.17 4.1824b Furo Corelas Well 41.9 710 5.1 6.92 28.66 4.1825 Termas do Carapacho Hot spring 41.4 16,300 6.3 7.08 18.77 4.09

Helium solubility was calculated at water in situ temperature and conductivity conditions using the solubility data fromWeiss (1971) in the range 0–40 °C and from Potter and Clynne(1978) above 40 °C.

73P. Jean-Baptiste et al. / Earth and Planetary Science Letters 281 (2009) 70–80

3.1. Sao Miguel

Sao Miguel, located 400 km east of the Mid-Atlantic-Ridge (MAR)is the largest island of the Azores archipelago. It has three activestrato-volcanoes (Furnas, Agua de Pau, and Sete Cidades), withsummit calderas partly filled by a lake, which have a long record ofboth effusive and explosive eruptions (Booth et al., 1978). The lasteruption in Furnas caldera, in 1650 AD, caused about 200 casualties(Booth et al., 1978). Whereas Sete Cidades volcano has been at restfor quite a long (unknown) time, both Furnas and Agua de Pauvolcanoes display intense hydrothermal manifestations along theirmain volcano-tectonic structures, tapping active hydrothermalsystems at depth. All thermal springs and emitted gases are CO2-dominated (Ferreira and Oskarsson, 1999; Cruz et al., 1999; Ferreiraet al., 2005). CO2 is also released through widespread diffuse soil

Table 2Helium isotope results for gas samples (subscript “c” indicates atmospheric correction base

Island Samplenumber

Sample location Type of manifestation

Sao Miguel 1 Lagoa das Furnas/North Boiling pool2 Lagoa das Furnas/East Bubbling waters3 Furnas/Caldeira Grande Boiling pool4 Furnas/Serra do Trigo Bubbling waters5 Furnas/Terra Nostra Bubbling waters6 Ribeira Quente Bubbling waters7 Ribeira Quente Bubbling seawater8 Ribeira Quente Bubbling pool9 Caldeira Vehla Bubbling pool10 Fogo/Ribeira Grande Geothermal plant CL2

Flores 11 Lajedo/Ponta Negra Warm spring/bubbleTerceira 15 Furnas do Enxofre Solfatara/fumaroles

16 Furnas do Enxofre Solfatara/fumarolesGraciosa 21 Baia dos Homiciados Hot spring/bubbles

22a Furna do Enxofre Bubbling mud pool22b Furna do Enxofre Bubbling mud pool

As CO2 constitutes almost 100% of the gas phase at all sites (see text), 3He concentrations d

emanations across Furnas caldera's floor (Pasquier-Cardin et al.,1999) and Fogo's northern slopes (Ferreira et al., 2005). Carbon istypically mantle-derived, with a mean δ13C value of −4‰ (Ferreiraand Oskarsson, 1999). We collected water and/or gas samples withtemperature ranging from 17.6 °C to 99.4 °C at ten sites in thesoutheastern and central parts of the island (Fig. 2). Gas sampleswere obtained from fumaroles, mud pools, and bubbling springs orpools. In addition, we could collect submarine bubbling gas at tenmeters depth off the coast of Ribeira Quente (sample no. 7).

All gas samples display a narrow 3He/4He range, between 5.23 and6.07 Ra, (Fig. 3). These values are higher than those (4.06–4.9 Ra)initially reported by Tolstikhin et al. (1991) for three sites (FurnasCaldera, Caldeira Vehla and Ribeira Grande), but the latter were un-corrected for air contamination. In the water samples, dissolved heliumin excess of the solubility equilibriumhas a 3He/4He ratio lower than the

d on 20Ne — see text).

[4He] 4He/20Ne [4He]c R/Ra Rc/Ra(ppm) (ppm)

12.2 6.79 13.0 5.09 5.29±0.0916.8 14.88 17.6 5.14 5.23±0.0811.2 3.75 12.5 5.23 5.62±0.1017.0 7.29 18.9 5.85 6.07±0.1054.2 18.67 64.6 5.64 5.72±0.0916.7 17.65 17.4 5.38 5.46±0.0912.4 12.78 12.8 5.33 5.44±0.0911.7 6.00 12.6 5.31 5.55±0.0915.9 19.53 16.5 5.25 5.32±0.0914.4 16.67 14.9 5.65 5.74±0.09

s 661 44.53 6306 7.99 8.04±0.123.0 0.71 2.2 7.92 13.53±0.702.2 2.41 2.0 11.24 12.79±0.258.2 2.55 8.9 7.90 8.88±0.164.0 1.64 3.8 8.85 10.73±0.253.6 9.25 3.5 10.88 11.23±0.18

irectly translate into CO2/3He ratios.

Fig. 3. Scatter diagrams of 3He versus 4He content in free gas phases and [3He]–[3Hesol] versus [4He]–[4Hesol] in water samples. Scales for water and gas results are in black and redrespectively. Data of Moreira et al. (1999) for volcanic rocks have been plotted for comparison (“+” symbols). For convenience, we use the same scale for water samples (in cm3 STP/gof water) and rock samples (in cm3 STP/g of rock). Therefore, rock concentrations were multiplied by 4 for Sao Miguel and Terceira and by 100 for Graciosa.

74 P. Jean-Baptiste et al. / Earth and Planetary Science Letters 281 (2009) 70–80

gas phase, with a mean value of 4.5±0.8 Ra, but share comparable Heisotopic ratios with the local volcanic rocks for which post-eruptiveradiogenic heliumcontaminationwas invoked (Moreira et al.,1999).We

have shown in Section 2 that the residual influence of the aircontribution upon the air-corrected 3He/4He value is ≤0.1 Ra and so,cannot explain the isotopic difference with the gas phase. Therefore,

75P. Jean-Baptiste et al. / Earth and Planetary Science Letters 281 (2009) 70–80

since the 3He/4He ratio of the gas phase more faithfully represents thelocal magmatic end-member, as we argue in Section 4.1, this isotopicdiscrepancy suggests a significant (~20%) radiogenic 4He input to thegroundwaters from the host volcanic rocks during water–rockinteractions.

3.2. Terceira

The most active volcanic site on Terceira, Furnas de Enxofre, is asolfataric field located in the central part of the island. It consists ofnumerous small fumaroles distributed along fissures and cracks withsulphur deposits, cutting a highly altered soil. Two gas samples collectedfrom the most active fumaroles reveal plume-type 3He/4He ratios of13.53±0.7 Ra and 12.79±0.25 Ra (Table 2 and Fig. 3b). These ratios areeven higher than the values (up to 11.4 Ra) reported for Terceira volcanicrocks by Moreira et al. (1999) and Madureira et al. (2005), with theexception of a value of 14.8 Ra in one He-poor sample for which theauthors did not exclude a cosmogenic 3He contribution. Therefore, ourresults strengthen the idea that plume-type mantle He degassing isactively occurring through Terceira Island. This is confirmed by thehigher-than-MORB 3He/4He ratios in ourwater samples (up to 10.3Ra insample 20), in spite of their partial contamination by radiogenic heliumduring water–rock exchange, as previously observed on Sao Miguel.

3.3. Graciosa

Themain hydrothermalmanifestations occur in the Caldeira volcano(southeasternpart of the island), andparticularly inside thehuge (150mwide, 80 m high) Furna do Enxofre lava cave located in the caldeira,where a bubbling mud pool releases steam and gases. This leads to theaccumulation of deadly (N15%) CO2 concentrations at the bottom of thecave, filled by a cold water subterranean lake. Two gas samples collectedfrom this bubbling mud pool give comparably high 3He/4He ratios of10.7 Ra and 11.2 Ra, which again suggest a plume-type contribution(Table 2 and Fig. 3). In contrast with these relatively high ratios, heliumdissolved in the subterranean lake located in the same cave (Table 1:samples 22a and 22b) displays a surprisingly more radiogenic 3He/4Hevalue (8.3–8.4 Ra). This ratio closely matches that of all other waterscollected on Graciosa, which define a quite homogeneous air-corrected3He/4He signature (mean value=8.31 Ra) comparable to the value of(8.1±0.2) Ra measured on one olivine sample by Moreira et al. (1999).This strongly suggests that the subterranean lakewaters are outcroppinggroundwaters directly connected to the underlying aquifer and littleinfluenced by the nearby gas bubbling. An intermediate value of 8.9 Rais observed in the gas phase of the Baia dos Homiciados seashorehydrothermal system (sample 21).

3.4. Pico

The island is dominated by the 2300 m high and steep cone of thePico volcano, which last erupted in 1718–1720. At present, surfacehydrothermal manifestations are scarce on the island. A single waterwell, sampled on the southern coast in the Silveira area, gives aMORB-type 3He/4He ratio of 8.5±0.5 Ra (Fig. 3). This value plots at the lowerside of the He isotopic rangemeasured byMoreira et al. (1999) for Picolavas (9.4±0.9 Ra).

3.5. Faial

The main hydrothermal activity on Faial is located on the outerwestern flank of the central caldera, in the vicinity of the Capelhinosvolcanowhich started erupting under seawater in 1957 then subaeriallythroughout 1957–1958. No gas sample could be collected in this area,fuming activity onCapelhinos being too residual. Instead,water samples(12 and 13) could be collected using hydrological sampling bottles fromtwo wells located at Capelo (180 m deep) and Varadouro (6 m deep).

Both waters have similar He isotopic ratio, 8.53 Ra, in good agreementwith the data of Moreira et al. (1999) for Faial volcanic rocks (Fig. 3).

3.6. Flores

The island of Flores is one of the two islands of the Azores grouplocated on thewestern side of theMAR. The sole geothermal feature thatwe were able to find, thanks to the help of local people, is located in avery remote place (Punta Negra) on the seashore close to Lajedo. Itconsists in a tiny thermal pool, a few tens of liters in size only, hidden atthe base of a highly fractured rock. The pool is fed from beneath by asmall inflowofwarmwater (29 °C).Gasbubbles also escape froma smallfissure at the bottom of the pool. Lying one and a half meter only abovesea level, the pool is regularly submerged by waves especially at hightide. In spite of the difficult access and of the very small size of thisgeothermal feature, the results are quite interesting: bothwater and gassamples show a spectacularly high helium content (6300 ppm for gasbubbles), which suggests gas separation and concentration from adeeper thermal aquifer, and display a similar 3He/4He ratio (8.04 Ra —Fig. 3). Such a value points to a pure MORB origin for helium. Nocomparison is possible with volcanic rocks on Flores, as no He isotopicdata are yet available for this island.

4. Discussion

4.1. Representativity of the analysed samples with respect to the heliumisotope signature of the mantle source

Noble gas isotope composition of olivine phenocrysts in volcanicrocks iswidely used in geochemistry and considered as representative ofmantle sources, even though different processes (e.g., partial melting,crystal-melt-bubble partitioning, outgassing, crustal contamination,etc.) can potentially affect the noble gas composition of silicate liquidsand solid products (Dymond and Hogan, 1978; Honda and Patterson,1999 and references therein). Noble gases can also be tracked from themagma-derived gas phase, which carries the vast majority of the initialgas content of the magma. As a matter of fact, there is a general goodagreement worldwide between 3He/4He ratios measured in volcanicrocks samples (phenocrysts) and in emitted volcanic gases, as illustratedby the results formid-ocean ridge submarinevolcanism(Graham, 2002;Schlosser and Winckler, 2002) and for subaerial volcanoes such asMount Etna (Allard et al.,1997), Pantelleria (Parello et al., 2000), and theRoman Comagmatic Province in Italy (see Fig. 7 in Martelli et al., 2004).

In the Azores, our helium isotopic data set for hydrothermal fluidsoffers a consistent picture with that for volcanic rocks (Moreira et al.,1999;Madureira et al., 2005), but also shows locally some discrepanciesbetween gas, water and rock results. In particular, the most intensethermal gas emissions from central vents on Sao Miguel, Terceira andGraciosa volcanic islands tend to display higher 3He/4He ratio thanthermal and cold groundwaters and, in some cases, than volcanic rocksthemselves. Therefore, the origin of these isotopic discrepancies needsfirst to be understood if one wants to correctly characterize the Heisotopic signature of the underlyingmagmatic source. Twomain classesof processes may potentially be involved:

(i) Kinetic fractionation of helium isotopes during either magmadegassing processes or/and gas transport through aquifers,which, due to the mass difference between the two isotopes,can enrich the gas phase in 3He relative to 4He. Such a frac-tionation may principally occur during diffusion processes(Graham law) and can generate significant changes in the 3He/4He ratio of the emitted and residual phases if Raleigh-type gasdistillation affects a reservoir of limited size.

(ii) Contribution of radiogenic 4He produced by U and Th decay inboth the ~23Myold oceanic crust (Searle,1980) and the volcanicrocks of the Azores Plateau, which would preferentially affect

76 P. Jean-Baptiste et al. / Earth and Planetary Science Letters 281 (2009) 70–80

hydrothermal aquifers and possibly pondingmagma bodies (andsome of their erupted products).

Belowwe briefly outline a series of observations which support theidea of a limited bearing (if any) of kinetic fractionation of He isotopesonto the measured 3He/4He ratios of Azores gas phases and, instead, asignificant contribution of radiogenic 4He to the local groundwaters.

4.1.1. Kinetic fractionation during magma degassingKinetic fractionation of He isotopes may occur during Rayleigh gas

distillation from ascendingmagma batches (i.e. during helium diffusionfromthemelt into escapinggas bubbles), andwas invoked to explain thetransient 3He/4He variations (on the order of ±0.5 Ra) recorded inMount Etna gas emissions at the onset of several magmatic eventsbetween 2001 and 2005 (Rizzo et al., 2006). However, eruptive activityin the Azores archipelago ismuchmore occasional than on Etna, and thehydrothermal manifestations we studied correspond to steady statedegassing taking place during the long repose intervals of Azoresvolcanoes. Moreover, as already discussed by Moreira et al. (1999) andMadureira et al. (2005), helium isotopic data for the Azores volcanicrocks shownoor little dependencyof 3He/4Heratios to theHecontentofolivine crystals. For instance, most of the rock samples from Faial andTerceira respectively display similar R/Ra ratio, within analyticaluncertainty, despite olivine's 3He contents that vary by as much astwo orders of magnitude. These different observations, and the globalconsistency between the hydrothermal fluids and rocks data sets, argueagainst any significant He isotopic fractionation between magmas,phenocrysts and emitted gases in the Azores data set.

4.1.2. Kinetic fractionation during gas transportIn theory, kinetic fractionation of helium isotopes between gas and

water can occur in aquifers during bubble formation and/or gasbubbling. However, field data published for steady-state hydrothermalsystems with short water residence time (i.e., with no suspicion ofradiogenic helium contamination) generally show an excellent agree-ment between helium isotopes ratios in the gas and the water phases(e.g. Allard et al., 1997; Parello et al., 2000), and hence show no sign ofany such Rayleigh kinetic fractionation. In the Azores too, such afractionation process is quite unlikely, especially for the most activethermal gas emissions in Sao Miguel, Graciosa and Terceira: theseemanations, characterized by intense advective gas flow, are directlyconnected to the active conduits of the volcanic systems; their nearlypure CO2-composition is consistent with a continuous gas supply fromdeep magmatic sources through CO2–He oversaturated hydrothermalsystems, allowing almost direct gas transit towards the surface.We thusconsider their 3He/4He ratios as the least exposed to possible isotopicfractionation. Here we also emphasize that, at Furnas volcano, wefind similar 3He/4He ratios for both intense fumarolic gas emissions(sample 1) and gentle gas bubbling across Furnas lake (sample 2;Table 2) or ponding waters (samples 4 and 5), although differingmarkedly in their temperature, He content and He/CO2 ratio. Similarly,at Fogo volcano, we find identical 3He/4He ratios for both surfacebubbling gas (Caldera Velha, sample 9) and overpressured gas from the1.4 kmdeep CL2 geothermal drill of Ribeira Grande (sample 10). Similar3He/4He ratios in the gas phase and in corresponding water are alsofound on Flores. Such similarities thus suggest minor kinetic fractiona-tion of He isotopes during gas transport across the studied aquifers. Themost spectacular He isotopic contrast in our data set is between thesubterranean cold water lake located at the bottom of the Furna doEnxofre lava cave (Graciosa) and nearby mud-pool bubbling gases (seeSection 3.3), for which we provide an explanation below.

4.1.3. Crustal contamination of groundwatersAs shownabove, heliumdissolved inhot springs andgroundwaters of

the Azores archipelago tends to display lower 3He/4He ratios than themost intense gas emissions. Although little is known concerning the

water residence time in the various sampled aquifers and concerning theradiogenic 4He production rate in the host formations, themost straight-forward explanation for this discrepancy is a contamination by radio-genic helium, as observed in many instances worldwide (Ballentine andBurnard, 2002; Kipfer et al., 2002). Lower 3He/4He ratios in ground-waters and/or remote gas emissions, due to increasing dilution of themagmatic component by radiogenic 4He derived from host rocks, arealso a common observation made at many volcanoes (Sano et al., 1984;Welhan et al., 1988; Marty et al., 1989; Sakamoto et al., 1992). On SaoMiguel, Moreira et al. (1999) evidenced a post-eruptive radiogenicingrowth in bulk lavas, which would allow the transfer of 4He togroundwaters duringwater–rock exchanges. This contaminationprocessof groundwaters should be favoured by the long repose intervals ofAzores volcanoes and by the strong alteration of host rocks at depth byacid gases of magmatic derivation, such as documented at Furnasvolcano (Ferreira and Oskarsson, 1999; Cruz et al., 1999). Accepting thisframework, groundwaters at both Furnas and Fogo volcanoes couldcontain ~20% radiogenic 4He from host rocks. We interpret in the sameway the lower 3He/4He ratio (8.3–8.4 Ra) for Graciosa groundwaterswith respect to the free gases (up to 11.2 Ra). This interpretation extendsto the subterranean lake located in the Furna do Enxofre lava cave ofGraciosa's caldera (Section 3.3), whose 3He/4He ratio markedly differsfrom that of the nearby bubbling mud-pool but is identical to that of thelocal aquifer, and which may thus consist of outcropping groundwatersdirectly connected to the underlying aquifer and little influenced by thenearby gas bubbling.

In conclusion, while helium dissolved in thermal and cold ground-waters of the Azores archipelago is likely affected to various extents byradiogenic contamination from the host rocks, we are quite confidentthat steady-state gas emissions from the most active vents in eachisland are reliable carriers of the helium isotope composition of theunderlying magmatic end-member.

4.2. Regional He isotopic pattern

Our results on Azores hydrothermal fluids, together with He iso-topic data for the volcanic rocks (Kurz, 1991; Moreira et al., 1999;Madureira et al., 2005), reveal not only a wide range but also a quitecoherent spatial distribution of 3He/4He ratios across the Azoresarchipelago.

Sao Miguel is clearly distinct from the other islands, with 3He/4Heratios consistently lower than the MORB reference in both fluids androcks. The hydrothermal gas phases from both Furnas and Agua de Pauvolcanoes indicate a homogeneous magmatic end-member, with mean3He/4He ratio of 5.54±0.25 Ra (Fig. 3), which points to a magmaticsource with higher time-integrated (U+Th)/3He ratio relative tonormal MORB material. Whether this ratio is representative of wholemagmatism in Sao Miguel island however remains uncertain. In fact,Kurz (1991) reported more extreme 3He/4He values of 3.5 Ra and 8.4Ra respectively, for rock samples from the northeastern (NordesteComplex) and northwestern (Sete Cidades) volcanic complexes. Weemphasize that Kurz's data correlate with a westward decrease ofradiogenic Sr and Pb in volcanic rocks across Sao Miguel island (Kurz,1991), which thus strongly suggests a trend in the underlying mantlesource. In terms of He isotopes, themagmatic source feeding Furnas andAgua de Pau central volcanoes would then be intermediate betweentypicalMORB to thewest (SeteCidades) andhighly radiogenic under theeasternmost part of the island. As amatter of fact, lavas from SaoMiguelhave long been recognized to display highly radiogenic Sr and Pbisotopic ratios among OIBs (Hawkesworth et al., 1979; White et al.,1979), implying a crustal component in their source. It has beenproposed that this crustal component could represent either a mantlecontaminated by subducted terrigenous sediment (Hawkesworth et al.,1979; Turner et al.,1997) or delaminated subcontinental lithosphere leftbehind at the openingof theNorth Atlantic (Widomet al.,1997;Moreiraet al., 1999). However, more recent data for Sr–Pb–Hf–Nd isotopes and

Fig. 5. Longitudinal distribution of 3He/4He ratios in Azores volcanic fluids and rocks.Rock data (in blue) are from Kurz (1991), Moreira et al. (1999) and Madureira et al.(2005). Our gas data are in red. Data for the Mid-Atlantic Ridge (37°N–40°N) rocks arefrom Moreira and Allègre (2002) and those for submarine hydrothermal fluid ventingare from Jean-Baptiste et al. (1998) and Charlou et al. (2000).

77P. Jean-Baptiste et al. / Earth and Planetary Science Letters 281 (2009) 70–80

trace elements (Elliot et al., 2007; Beier et al., 2007) rather suggest anenriched crustal source containing recycled oceanic crust, plus somesubducted seamount (for the easternmost volcanism in SaoMiguel). Heisotope study of volcanic fluids in eastern SaoMiguel would certainly beof interest in the future to discuss these various options.

The other islands of the Azores archipelago exhibit a helium isotopicsignature varying fromtypicalMORBvalues (Flores, Faial) toplume-typevalues at Terceira, Graciosa and Pico. This He isotopic pattern broadlyparallels the variations of other isotopic markers in correspondingvolcanic rocks (Fig. 4). In addition to the radiogenic componentmentioned above, the observed isotopic variations require a mixingwith at least two other mantle components: (i) normal MORB mantlesimilar to that feeding the nearby Mid-Atlantic Ridge, and (ii) a plume-type component with moderately high 3He/4He ratio and intermediate87Sr/86Sr and 206Pb/204Pb values compared to the most ‘primitive’ end-member feeding Iceland or Loihi seamount in Hawaii (e.g. Graham,2002).

The occurrence of plume-type 3He/4He ratios in fluids and rocks ofthe central volcanic islands demonstrates that the Azores plumecomponent is concentrated under the central part of the archipelago(Fig. 5). Here we find the highest value (13.5 Ra) in gas emissionsof Terceira, where volcanic rocks also display the least radiogenic21Ne/22Ne ratios compared toMORB (Madureira et al., 2005). However,the actual centre of the “high” 3He/4He plume might well be locatedbeneath the islandof Sao Jorge, just south of Terceira,whereMoreira andAllègre (2002) found one rock sample with a ratio as high as 15.9 Ra.

Fig. 4. He–Sr and He–Pb isotopic relations for Azores volcanic products (adapted fromGraham et al., 1998). The helium isotope data set includes our present results for Azoresfluids and those for volcanic rocks fromKurz (1991), Moreira et al. (1999), Madureira et al.(2005). Sr–Pb isotope data for Azores lavas are from White et al. (1976), Hawkesworthet al. (1979), Dupré et al. (1982), Davies et al. (1989), Widom et al. (1997), Turner et al.(1997), Moreira et al. (1999), França et al. (2006), and Beier et al. (2007).

4.3. Constraints on the origin of the Azores plume component

The origin of the Azores plume component is complex and stilldebated. Plume-related OIBs from the central Azores islands displaygeochemical evidence of mixed contributions from the lower mantle(Moreira et al., 1999; Madureira et al., 2005) and a recycled radiogeniccomponentwhich, depending on the considered indicators, may consistof altered oceanic crust and sediments (Turner et al.,1997;Moreira et al.,1999) or delaminated oceanic mantle lithosphere of Archean age(Schaefer et al., 2002). Plume supply from a primitive lower mantlesource is indicated by helium isotopic data for volcanic rocks and fluidsfrom the central islands (Moreira et al.,1999;Madureira et al., 2005; thiswork) and by the lower-than-MORB 21Ne/22Ne ratios in volcanic rocksfrom Terceira (Madureira et al., 2005). A deep derivation of the Azoresplume is also supported by the very low (sub-chondritic) Os isotopicratios in volcanic rocks from Terceira and especially Faial–Pico islands(Schaefer et al., 2002) and by the seismic tomography inversion ofMontelli et al. (2004), which shows a well-resolved plume originatingfrom near the bottom of the mantle. However, an interesting featurewhich arises fromour data set is that the gas emissionswith the highest3He/4He ratios correspond to the lowest 3He concentrations; there is infact a broad anti-correlation between both parameters in Azores gasphases when considering all islands (Fig. 6a). Moreover, because CO2

constitutes almost 100% of the gas phase at all sites, such a trend directlytranslates into corresponding variations of the CO2/3He ratios, whichincrease fromanaverage of (9.3±2.1)×109 on SaoMiguel to ~17.5×109

onGraciosa and 22±4×109 onTerceira (Fig. 6b). These latter values areone order of magnitude higher than the average MORB CO2/3He molarratio (~2.2×109; Marty and Tolstikhin, 1998). Therefore, the trends inFig. 6a–b could reflect either a dilution of 3He by a CO2-rich gas from acarbon enrichedmantle plume or alternatively a lower-than-MORB 3Hecontent of the mantle source.

A CO2-rich Azores plume was previously inferred by Kingsley andSchilling (1995) who measured high CO2 contents and higher-than-MORBCO2/3He ratios (upto4.8×109) in submarinebasaltic glasses fromthe Mid-Atlantic ridge at the latitude of the Azores and estimated aminimal C/3He molar ratio of ~1010 for the plume source. The co-variations of these ratios with both Sr and Nd isotopic ratios in the samebasalts, and their independence fromthedegreeof basalt degassing and/or CO2 oversaturation (1.8 to 6.4), were taken as an evidence for ridgeinteractionwith amantle plume highly enriched in carbon dioxide (by a

Fig. 6. Relationship between 3He/4He, 3He concentrations (a) and CO2/3He (b) in ourAzores gas samples (plus, squares and diamonds are for Sao Miguel, Graciosa andTerceira, respectively; MORB and ARCS domains are from Marty and Tolstikhin, 1998;Kilauea and Hengill data are from Craig and Lupton, 1976 and Marty et al., 1991respectively).

78 P. Jean-Baptiste et al. / Earth and Planetary Science Letters 281 (2009) 70–80

factor 3.3 to 8.8 relative to depleted asthenosphere) and other volatilessuch as H2O and Cl (Kingsley and Schilling, 1995). The high CO2/3Heratios (1.8–2.6×1010) combined with high 3He/4He ratios we measuredin gas emissions from Terceira and Graciosa further argue in support of aCO2-rich mantle plume feeding the Azores. Carbon recycling in themantle through oceanic plate subduction is the main process capable ofexplaining higher-than-MORB C/3He ratios in the products of arc-volcanismand, likely, of plume-related hot spots (e.g.Marty and Jambon,1987; Trull et al., 1993; Marty and Tolstikhin, 1998). Therefore, weinterpret the coexistence of high (arc-type) CO2/3He ratio andmoderately high 3He/4He ratio in the Azores plumematerial as evidenceof a mixing between a carbon-rich, He-poor component of recycledorigin and a 3He-rich component from the lower mantle. In agreementwith other geochemical data, the recycled component (with high initialCO2/3He ratio) could be an ancient subducted oceanic plate, previouslyenriched in CO2 and other volatiles during hydrothermal alteration. Notethat the CO2-rich plume would also contribute to mantle degassingthrough Sao Miguel, where CO2/3He ratios are lower than on Terceiraand Graciosa but still high compared to the MORB reference.

However, one cannot entirelyexclude that the low3Heconcentrationin volcanic gases of Terceira and Graciosa (Fig. 6a–b) could alternativelyresult from a lower-than-MORB 3He content in the Azores plumematerial. As a matter of fact, it is a general observation (the so-called“helium paradox” — Anderson,1998) that OIBs presumably fed by deepmantle plumes tend to have lower helium concentrations than MORBs,in contrast to the expectation that a plumeof primitivematerial from thelower mantle should be less degassed and hence richer in 3He than the

depleted upper mantle. If real, a relative 3He depletion in the Azoresplume material might result from two possible processes:

i) Previous degassing of the plume upon ascent. Such a process wasinvoked by Hilton et al. (1997) for explaining the helium–carbonisotope relationship at Kilauea volcano and, more recently, forexplaining the differences in noble gas elemental ratios betweenMORB and OIB glasses (Gonnermann and Mukhopadhyay, 2007;Hopp and Trieloff, 2008). Note, however, that because carbondioxide is less soluble than helium in the melt, plume pre-degassing would be expected to lower the CO2/3He ratio of theplume, whereas we find particularly high ratios for the Azores;

ii) Fractionation of He and U–Th during partial melting events.Recent experimental results unexpectedly revealed that heliumcould be more compatible (i.e. less efficiently removed) than U andTh during partial mantle melting (Parman et al., 2005), therebyallowing enhanced 3He/4He ratio (less affected by radiogenic Heingrowth from U–Th decay) in ancient mantle reservoirs that,otherwise, have been depleted by successive melting events. Theseresults thus offer a potential key to resolve the apparent heliumparadox between plume-related OIBs and MORBs. This possibilityalso extends to other noble gases, since new results fromHeber et al.(2007) show that the partition coefficients of noble gases in olivineare one to two orders of magnitude greater than for U and Th.Therefore, as far ashelium is concerned, partialmantlemelting couldlead to preferential extraction of U and Th, leaving behind a residualreservoir with a higher-than-MORB time-integrated 3He/(U+Th)ratio. The only condition to get high 3He/4He ratios is that these U–Th-depleted mantle domains must have formed before these ratiosin the evolving mantle have dropped too low, hence relatively earlyin the Earth history. Therefore this scenario, sometimes presented asthe “whole mantle convection” alternative to the classic “layeredmantle” model, equally requires that mantle domains with distinctgeochemical signatures could escape convective homogenization forvery long periods of time, something which is difficult to reconcilewith the whole mantle convection assumption.

With regard to the Azores hot spot, in addition to explaining thenegative trend between 3He concentrations and the 3He/4He ratios(Fig. 6a), this “residual reservoir” scenario would support Schaeferet al.'s (2002) conclusions based on osmium isotopes, that the Azoresplume may derive from the deep subduction of a depleted Archeanoceanic mantle lithosphere. We emphasize, however, that the Azoreslava present enriched trace element patterns (Turner et al., 1997),suggesting that the source of the Azores hotspot is more enriched inincompatible elements such as U and Th thanMORB. It seems thereforedifficult to produce a lower-than-MORB 3He content in the Azoressource whatever the relative incompatibilities of He and U–Th.

Hence, we rather favour the conclusion that the Azores plumematerial is highly enriched in CO2 and other volatiles (rather thanbeing depleted in 3He), and contains ancient altered oceanic materialwith high initial CO2/3He ratio that was recycled deep in a 3He-richlower mantle. Future systematic investigations of carbon dioxide andnoble gases in both volcanic rocks and fluids of the Azores archipelagoshould provide further insight into this matter.

5. Conclusions

Helium isotope ratios measured in both thermal waters and gasesof the Azores archipelago lead to the following main observations andconclusions:

— Hydrothermal gas emissions from the most active and central ventsof the volcanic systems tend to display higher 3He/4He ratios thanlocal groundwaters, and likely carry themost representative isotopicsignature of the underlying mantle source. Thermal groundwatersrather share similar isotopic ratio with the host volcanic rocks

79P. Jean-Baptiste et al. / Earth and Planetary Science Letters 281 (2009) 70–80

(Moreira et al., 1999; Madureira et al., 2005), implying slight tosignificant addition of radiogenic helium to the aquifers duringwater–rock interactions.

— Hydrothermal fluids in central Sao Miguel island clearly distin-guish from those in the other islands, with 3He/4He ratios of 5.23–6.07 Ra much lower than the MORB reference. Such ratios, togetherwith Sr–Nd–Pb–Hf–Os isotopic compositions and trace elementpatterns in corresponding volcanic rocks, imply the contribution ofa radiogenic crustal component in the local degassing mantlesource. The actual nature of this crustal component however re-mains a matter of debate (e.g. Elliot et al., 2007).

— Fluids on Faial and Flores islands, located closer to and on either sideof the nearby Mid-Atlantic Ridge, actually display MORB-type 3He/4He ratios, indicating essentially pure upper mantle degassing.

— Instead, plume-type 3He/4He ratios characterizehydrothermalfluidson Terceira, Graciosa and, to a lesser degree, Pico islands. Similarlyhigh ratios in corresponding volcanic rocks indicate no or little Heisotopic fractionationduringmagmadegassing. Adeepmantle originof helium in the analysed fluids is consistent with tomographicevidence of a deep-rooted plume beneath the Azores hot spot. Thespatial distribution of 3He/4He ratios demonstrates that plume'shead is located beneath the central part of the archipelago.

— The highest 3He/4He ratios in Terceira and Graciosa islands cor-respond to the lowest 3He concentrations and highest CO2/3Heratios. This observation is interpreted as indicative of a high CO2

content (and consequent 3Hedilution) in theAzores plumematerial,in agreement with previous measurement of high C/3He ratios inbasaltic glasses from the nearby Mid-Atlantic Ridge (Kingsley andSchilling,1995). This CO2 enrichmentof theAzores plume, associatedwith plume-type 3He/4He values, is consistent with other geochem-ical evidence for the recycling of ancient oceanic material in thelower mantle beneath the Azores.

References

Allard, P., Jean-Baptiste, P., D'Alessandro, W., Parello, F., Parisi, B., Flehoc, C., 1997.Mantle-derived helium and carbon in groundwaters and gases of Mount Etna, Italy.Earth Planet. Sci. Lett. 148, 501–516.

Allègre, C.J., Turcotte, D.L., 1985. Geodynamic mixing in the mesosphere boundary layerand the origin of oceanic islands. Geophys. Res. Lett. 12, 207–210.

Allègre, C.J., Staudacher, T., Sarda, P., Kurz, M., 1983. Constraints on evolution of theEarth's mantle from rare gas systematics. Nature 303, 762–766.

Anderson, D.L., 1998. A model to explain the various paradoxes associated with mantlenoble gas geochemistry. Proc. Natl. Acad. Sci. U.S.A. 95, 9087–9092.

Ballentine, C.J., Burnard, P.G., 2002. Production, release and transport of noble gases inthe continental crust. In: Porcelli, D., Ballentine, C.J., Wieler, R. (Eds.), Noble Gases inGeochemistry and Cosmochemistry. Rev. Mineral. Geochem., vol. 47, pp. 481–538.

Beier, C., Stracke, A., Haase, K.M., 2007. The peculiar geochemical signatures of SaoMiguel (Azores) lavas: metasomatised or recycled mantle sources? Earth Planet.Sci. Lett. 259, 186–199.

Booth, B., Croasdale, R., Walker, G.P.L., 1978. A quantitative study of five thousand yearsof volcanism on Sao Miguel, Azores. Phil. Trans. R. Soc. Lond. A288, 271–319.

Charlou, J.L., Donval, J.P., Fouquet, Y., Douville, E., Ondreas, H., Jean-Baptiste, P.,Dapoigny, A., Stievenard, M., 2000. Compared geochemical signatures andevolution of Menez Gwen (37°50N) and Lucky Strike (37°17N) hydrothermalfluids, south of the Azores Triple junction on the Mid-Atlantic Ridge. Chem. Geol.171, 49–75.

Craig, H., Lupton, J.E., 1976. Primordial neon, helium, and hydrogen in oceanic basalts.Earth Planet. Sci. Lett. 31, 369–385.

Cruz, J.V., Coutinho, R.M., Carvalho, R., Oskarsson, N., Gislason, S., 1999. Chemistry ofwaters from Furnas volcano, Sao Miguel, Azores: fluxes of carbon dioxide andleached material. J. Volcanol. Geotherm. Res. 92, 151–167.

Davies, G.R., Norry, M.J., Gerlach, D.C., Cliff, R.A., 1989. A combined chemical and Pb–Sr–Nd isotopes study of the Azores and Cape Verde hot-spots: the geodynamicimplications. In: Saunders, A.D., Norry, M.J. (Eds.), Magmatism in the Ocean Basins.Geol. Soc. London Spec. Publ., vol. 42, pp. 231–255.

Dupré, B., Lambert, B., Allègre, C.J., 1982. Isotopic variations within a single oceanicisland: the Terceira case. Nature 299, 620–622.

Dymond, J., Hogan, L., 1978. Factors controlling the noble gas abundance patterns ofdeep-sea basalts. Earth Planet. Sci. Lett. 38, 117–128.

Elliot, T., Blichert-Toft, J., Heumann, A., Koetsier, G., Forjaz, V., 2007. The origin of enrichedmantle beneath Sao Miguel, Azores. Geochim. Cosmochim. Acta 71, 219–240.

Farley, K.A., Neroda, E., 1998. Noble gases in the Earth's mantle. Ann. Rev. Earth Planet.Sci. 26, 189–218.

Ferreira, T., Oskarsson, N., 1999. Chemistry and isotopic composition of fumaroledischarges of Furnas caldera. J. Volcanol. Geotherm. Res. 92, 169–179.

Ferreira, T., Gaspar, J.L., Viveiros, F., Marcos, M., Faria, C., Sousa, F., 2005. Monitoring offumarole discharge and CO2 soil degassing in the Azores: contribution to volcanicsurveillance and public health risk assessment. Ann. Geophys. 48, 787–796.

França, Z.T.M., Tassinari, C.C.G., Cruz, J.V., Aparicio, A.Y., Arana, V., Rodrigues, B.N., 2006.Petrology, geochemistry and Sr–Nd–Pb isotopes of the volcanic rocks from PicoIsland — Azores (Portugal). J. Volcanol. Geotherm. Res. 156, 71–89.

Gaspar, J.L., Queiroz, G., Pacheco, J.M., Ferreira, T., Wallenstein, N., Almeida, M.H.,Coutinho, R., 2003. Basaltic lava balloons produced during the 1998–2001 Serretasubmarine ridge eruption (Azores), in: explosive subaqueous volcanism. AGUGeophys. Monogr. 140, 205–212.

Gente, ¨P., Dyment, J., Maia, M., Goslin, J., 2003. Interaction between the Mid-AtlanticRidge and the Azores hotspot during the last 85 Myr: emplacement and rifting ofthe hotspot derived plateaus. Geochem. Geophys. Geosyst. 4, 8514. doi:10.1029/2003GC000527.

Gonnermann, H.M., Mukhopadhyay, S., 2007. Non-equilibrium degassing and primor-dial source for helium in ocean–island volcanism. Nature 449, 1037–1040.

Graham, D.W., 2002. Noble gas isotope geochemistry of Mid-Ocean Ridge and OceanIsland Basalts: characterization of mantle source reservoirs. In: Porcelli, D.,Ballentine, C.J., Wieler, R. (Eds.), Noble Gases in Geochemistry and Cosmochemistry.Rev. Mineral. Geochem., vol. 47, pp. 247–317.

Graham, D.W., Larsen, L.M., Hanan, B.B., Storey, M., Pedersen, A.K., Lupton, J.E., 1998.Helium isotope composition of early Iceland mantle plume inferred from theTertiary picrites of West Greenland. Earth Planet. Sci. Lett. 160, 241–255.

Grand, S.P., 2002. Mantle shear-wave tomography and the fate of subducted slabs. Phil.Trans. R. Soc. Lond. A360, 2475–2491.

Hanyu, T., Kaneoka, I., 1997. The uniform and low 3He/4He ratios of HIMU basalts asevidence for their origin as recycled materials. Nature 390, 273–276.

Hawkesworth, C.J., Norry, M.J., Roddick, J.C., Vollmer, R., 1979. 143Nd/144Nd and 87Sr/86Nd ratios from the Azores and their significance in LIL-element enriched mantle.Nature 280, 28–31.

Heber, V.S., Brooker, R.A., Kelley, S.P., Wood, B.J., 2007. Crystal-melt partitionning ofnoble gases (helium, neon, argon, krypton, and xenon) for olivine and clinopyr-oxene. Geochim. Cosmochim. Acta 71, 1041–1061.

Hilton, D.R., Barling, J., Whelier, G.E., 1995. Effect of shallow-level contamination on thehelium isotope systematics of ocean–island lavas. Nature 373, 330–333.

Hilton, D.R., McMurtry, G.M., Kreulen, R., 1997. Evidence for extensive degassing of theHawaiian mantle plume from helium–carbon relationships at Kilauea volcano.Geophys. Res. Lett. 24, 3065–3068.

Hohmann,R.,Hofer,M., Kipfer, R., Peeters, F., Imboden,D.M., Baur, H., Shimareev,M.N.,1998.Distribution of helium and tritium in Lake Baikal. J. Geophys. Res. 103, 12823–12838.

Honda, M., Patterson, D.B., 1999. Systematic elemental fractionation of mantle-derivedhelium, neon, and argon in mid-oceanic ridge glasses. Geochim. Cosmochim. Acta63, 2863–2874.

Hopp, J., Trieloff, M., 2008. Helium deficit in high-3He/4He parent magmas:predegassing fractionation, not a “helium paradox”. Geochem. Geophys. Geosyst.9, Q03009. doi:10.1029/2007GC001833.

Jean-Baptiste, P., Mantisi, F., Dapoigny, A., Stievenard, M., 1992. Design and performanceof a mass spectrometric facility for measuring helium isotopes in natural watersand for low-level tritium determination by the 3He ingrowth method. Appl. Radiat.Isot. 43, 881–891.

Jean-Baptiste, P., Bougault, H., Vangriesheim, A., Charlou, J.L., Radford-Knoery, J.,Fouquet, Y., Needham, D., German, C., 1998. Mantle 3He in hydrothermal vents andplume of the Lucky Strike site (MAR 37°17N) and associated geothermal heat flux.Earth Planet. Sci. Lett. 157, 69–77.

Jean-Baptiste, P., Fourré, E., Charlou, J.L., Donval, J.P., Dapoigny, A., 2008. Deepwatermantle3He plumes over the Northern Mid-Atlantic Ridge (36°N–40°N) and the AzoresPlatform. Geochem. Geophys. Geosyst. 9, Q03010. doi:10.1029/2007GC001765.

Kingsley, R.H., Schilling, J.G., 1995. Carbon in Mid-Atlantic Ridge basalt glasses from28°N to 63°N: evidence for a carbon-enriched Azores mantle plume. Earth Planet.Sci. Lett. 129, 31–53.

Kipfer, R., Aeschbach-Hertig, W., Peeters, F., Stute, M., 2002. Noble gases in lakes andground waters. In: Porcelli, D., Ballentine, C.J., Wieler, R. (Eds.), Noble Gases inGeochemistry and Cosmochemistry. Rev. Mineral. Geochem., vol. 47, pp. 615–700.

Kurz, M.D., Jenkins, W.J., Schilling, J.G., Hart, S.R., 1982a. Helium isotopic variations inthe mantle beneath the central North Atlantic Ocean. Earth Planet. Sci. Lett. 58,1–14.

Kurz, M.D., Jenkins, W.J., Hart, S.R., 1982b. Helium isotopic systematics of ocean islandsand mantle heterogeneity. Nature 297, 43–47.

Kurz, M.D., 1991. Noble gas isotopes in oceanic basalts: controversial constraints onmantle models. In: Heaman, L., Ludden, J.N. (Eds.), Short Course Handbook onApplications of Radiogenic Isotope Systems to Problems in Geology. MineralAssociation of Canada, Toronto, Chapter 8, pp. 259–286.

Lourenço, N., Miranda, J.M., Luis, J.F., Ribeiro, A., Mendes-Victor, L.A., Madeira, J.,Needham, H.D., 1998. Morpho-tectonic analysis of the Azores volcanic plateau froma new bathymetric compilation of the area. Mar. Geophys. Res. 20, 141–156.

Lupton, J.E., 1983. Terrestrial inert gases: isotope tracer studies and clues to primordialcomponents in the mantle. Ann. Rev. Earth Planet. Sci. 11, 371–414.

Madureira, P., Moreira, M., Mata, J., Allègre, C.J., 2005. Primitive neon isotopes in TerceiraIsland (Azores archipelago). Earth Planet. Sci. Lett. 233, 429–440.

Martelli, M., Nuccio, P.M., Stuart, F.M., Burgess, R., Ellam, R.M., Italiano, F., 2004. Helium–strontium isotope constraints onmantle evolution beneath the Roman ComagmaticProvince, Italy. Earth Planet. Sci. Lett. 224, 295–308.

Marty, B., Jambon, A., 1987. C/3He ratio in volatile fluxes from the solid earth:implications for carbon geodynamics. Earth Planet. Sci. Lett. 83, 16–26.

Marty, B., Tolstikhin, I.N., 1998. CO2 fluxes from mid-ocean ridges, arcs and plumes.Chem. Geol. 145, 233–248.

80 P. Jean-Baptiste et al. / Earth and Planetary Science Letters 281 (2009) 70–80

Marty, B., Jambon, A., Sano, Y., 1989. Helium isotopes and CO2 in volcanic gases of Japan.Chem. Geol. 76, 25–40.

Marty, B., Gunnlaugsson, E., Jambon, A., Oskarsson, N., Ozima, M., Pineau, F., Torssander,P., 1991. Gas geochemistry of hydrothermal fluids, the Hengill area, southwest riftzone of Iceland. Chem. Geol. 91, 207–225.

McKenzie, D., O Nions, R.K., 1983. Mantle reservoirs and ocean island basalts. Nature301, 229–231.

Montagner, J.P., Ritsema, J., 2001. Interactions between ridges and plumes. Science 294,1472–1473.

Montelli, R., Nolet, G., Dahlen, F.A., Masters, G., Engdahl, E.R., Hung, S.H., 2004. Finite-frequency tomography reveals a variety of plumes in the mantle. Science 303,338–343.

Moreira, M., Allègre, C.J., 2002. Rare gas systematics on Mid Atlantic ridge (37–40°N).Earth Planet. Sci. Lett. 198, 401–416.

Moreira, M., Doucelance, R., Kurz, M.D., Dupré, B., Allègre, C.J., 1999. Helium and leadisotope geochemistry of the Azores archipelago. Earth Planet. Sci. Lett.169,189–205.

Parman, S.W., Kurz, M.D., Hart, S.R., Grove, T.L., 2005. Helium solubility in olivine andimplications for high 3He/4He in ocean island basalts. Nature 437, 1140–1143.

Parello, F., Allard, P., D Alessandro, W., Federico, C., Jean-Baptiste, P., Catani, O., 2000.Fluid geochemistry of Pantelleria volcano, Sicily Channel Rift: a mantle volatile end-member for volcanism in southern Europe. Earth Planet. Sci. Lett. 180, 325–339.

Pasquier-Cardin, A., Allard, P., Ferreira, T., Hatté, C., Coutinho, R., Fontugne, M., Jaudon,M., 1999. Magma-derived CO2 emissions recorded in 14C and 13C content of plantsgrowing in Furnas caldera, Azores. J. Volcanol. Geotherm. Res. 92, 195–207.

Potter, R.W., Clynne, M.A., 1978. The solubility of noble gases He, Ne, Ar, and Xe in waterup to the critical point. J. Solution Chem. 7, 837–844.

Rizzo, A., Caracausi, A., Favara, R., Martelli, M., Paonita, A., Paternoster, M., Nuccio, P.M.,Rosciglione, A., 2006. New insights intomagma dynamics during last two eruptionsof Mount Etna as inferred by geochemical monitoring from 2002 to 2005. Geochem.Geophys. Geosyst. 7, Q06008. doi:10.1029/2005GC001175.

Sakamoto, M., Sano, Y., Wakita, H., 1992. 3He/4He ratio distribution in and around theHakone volcano. Geochem. J. 26, 189–195.

Sano, Y., Nakamura, Y., Wakita, H., Urabe, A., Tominaga, T., 1984. Helium-3 emissionrelated to volcanic activity. Science 224, 150–151.

Schaefer, B.F., Turner, S., Parkinson, I., Rogers, N., Hawkesworth, C., 2002. Evidence forrecycled Archean mantle lithosphere in the Azores plume. Nature 420, 304–307.

Schilling, J.G., 1975. Azores mantle blob: rare earth evidence. Earth Planet. Sci. Lett. 25,103–115.

Schlosser, P., Winckler, G., 2002. Noble gases in oceanwaters and sediments. In: Porcelli,D., Ballentine, C.J., Wieler, R. (Eds.), Noble Gases in Geochemistry and Cosmochem-istry. Rev. Mineral. Geochem., vol. 47, pp. 701–730.

Searle, R., 1980. Tectonic pattern of the Azores spreading centre and triple junction.Earth Planet. Sci. Lett. 51, 415–434.

Stuart, F.M., Lass-Evans, S., Fitton, J.G., Ellam, R.M., 2003. High 3He/4He ratios in picriticbasalts from Baffin Island and role of a mixed reservoir in mantle plumes. Nature424, 57–59.

Tolstikhin, I.N., Kamensky, I.L., Forjaz, V., Polyak, B.G., Kononov, V.I., Kudrjavtsev, D.I.,1991. Helium isotopes in the gases of Sao Miguel island (on the Azores Plateauorigin). Izvestiya Akad. Nauk SSSR. Ser. Geol. 9, 137–147.

Trull, T.W., Nadeau, S., Pineau, F., Polve, M., Javoy, M., 1993. C–He systematics in hotspotxenoliths: implications for mantle carbon contents an carbon recycling. EarthPlanet. Sci. Lett. 118, 43–64.

Turner, S., Hawkesworth, C., Rogers, N., King, P., 1997. U–Th isotope disequilibria andocean basalt generation in the Azores. Chem. Geol. 139, 145–164.

Van der Hilst, R.D., Widiyantoro, S., Engdahl, E.R., 1997. Evidence for deep mantlecirculation from global tomography. Nature 386, 578–584.

Vogt, P.R., Jung, W.Y., 2004. The Terceira Rift as hyper-slow, hotspot-dominated obliquespreading axis: a comparison with other slow-spreading plate boundaries. EarthPlanet. Sci. Lett. 218, 77–90.

Weiss, R.F., 1971. Solubility of helium and neon inwater and seawater. J. Chem. Eng. Data16, 235–241.

Welhan, J.A., Poreda, R.J., Rison, W., Craig, H., 1988. Helium isotopes in geothermal andvolcanic gases of the Western United States, II: Long Valley caldera. J. Volcanol.Geotherm. Res. 34, 201–209.

Widom, E., Carlson, R.W., Gill, J.B., Schmincke, H.U., 1997. Th–Sr–Nd–Pb isotope andtrace element evidence for the origin of the Sao Miguel, Azores, enriched mantlesource. Chem. Geol. 140, 49–68.

White, W.M., Schilling, J.G., Hart, S.R., 1976. Evidence for the Azores mantle plume fromstrontium isotope geochemistry of the Central North Atlantic. Nature 263, 659–663.

White, W.M., Tapia, M.D.M., Schilling, J.G., 1979. The petrology and geochemistry of theAzores islands. Contrib. Mineral. Petrol. 69, 201–213.

Zindler, A., Hart, S.R., 1986. Helium: problematic primordial signals. Earth Planet. Sci.Lett. 79, 1–8.