Primitive Neon Isotopes in Terceira Island (Azores Archipelago)

Earth and Planetary Science Letters 233 (2005) 429–440 www.elsevier.com/locate/epsl

Primitive neon isotopes in Terceira Island (Azores archipelago)

Pedro Madureiraa,b,T, Manuel Moreiraa, Joa˜o Matac, Claude Jean Alle`grea

aLaboratoire de Geóchimie et Cosmochimie, Insitut de Physique du Globe de Paris-Universite´ Paris VII-Denis Diderot, 4 place Jussieu, 75005, Paris, France bCentro de Geofı´sica de E´ vora/Departamento de Geocieˆncias da Universidade de E´ vora, Rua Romaõ Ramalho, 59,7000-671, E´ vora, Portugal cCentro e Departamento de Geologia da Universidade de Lisboa, Faculdade de Cieˆncias, Campo Grande, Edifıćio C6,48 Piso, 1749-016, Lisboa, Portugal Received 27 July 2004; received in revised form 27 January 2005; accepted 15 February 2005 Available online 13 April 2005 Editor: K. Farley

Abstract

We present the first neon data, as well as new helium data, on Terceira Island (Azores archipelago, Portugal). Clear 20Ne and 21Ne excesses compared to air are observed (20Ne/22NeN11.2) and moreover, the samples show a more primitive 21Ne/22Ne ratio than MORB, confirming that the Azores hotspot can be considered as sampling a bprimitiveQ, relatively undegassed, reservoir. Most 4He/3He isotopic ratios range between 80,000 and 63,500 (~ 9 to 11.5 R/Ra), being similar to those previously reported by [1] [M. Moreira, R. Doucelance, B. Dupre´, M. Kurz, C.J. Alle`gre, Helium and lead isotope geochemistry in the Azores archipelago, Earth Planet. Sci. Lett. 169 (1999) 189–205]. A more primitive 4He/3He ratio of ~ 50,000 (R/Ra ~ 15) was obtained in one sample, but we cannot completely exclude the possibility of a cosmogenic 3He contribution for this sample. Our study illustrates that the neon systematics can be more capable than helium to constrain the ultimate origin of hotspots in geodynamic settings dominated by plume–ridge interaction. D 2005 Elsevier B.V. All rights reserved.

Keywords: neon; helium; Azores; plume–ridge interaction

1. Introduction emerged for the MORB and OIB sources (e.g. [2,3] and references therein). 4He production results from From the collection of helium isotopic data in the radioactive decay of U and Th and different time- last 20 yrs, significantly different 4He/3He signatures integrated isotopic ratios can be produced by heterogeneous mantle degassing, through distinct depletion of He as compared with U and Th. MORB samples, T ´ Corresponding author. Centro de Geofisica de Evora/Departa- considered to come from the upper mantle, present a mento de Geocieˆncias da Universidade de E´ vora, Rua Roma˜o 4 3 Ramalho, 59,7000-671, E´ vora, Portugal. Tel.: +351 266745300; narrow range of He/ He ratios with a mean value fax: +351 266745397. around 85,000 to 90,000 (R/Ra=8 to 8.5; where Ra is 3 4 6 E-mail address: [email protected] (P. Madureira). the atmospheric He/ He ratio of 1.38410 ). A

0012-821X/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2005.02.030 430 P. Madureira et al. / Earth and Planetary Science Letters 233 (2005) 429–440 significantly larger variation is found for OIB, but to identify the incorporation of relatively undegassed 4He/3He ratios as low as 20,500 (R/Ra ~ 35) and lower mantle portions in plumes. The Azores was 15,000 (R/Ra ~ 50) measured in samples from Loihi firstly classified as one of the blow 3He hotspotsQ seamount and Iceland hotspot (Baffin island) respec- characterized by 4He/3He ratios higher than the mean tively, are interpreted as reflecting the contribution of MORB ratio [32]. However, Moreira et al. [1] showed a more primitive and less degassed source believed to that Azores Islands have 4He/3He ratios both higher be the lower mantle [4–8]. This interpretation is in and lower than MORB values. Higher values agreement with the geophysical evidence linking the (4He/3HeN140,000; R/Rab5.2) are represented by hotspot activity in these areas to deeply rooted mantle rocks from the Eastern part of S. Miguel, while more plumes [9–11]. The existence of these 4He/3He primitive ratios were recorded in Terceira samples contrasts led to the proposition of a two-layer mantle (around 64,000; R/Ra=11.3). Using Pb isotopic data, model where the upper mantle is degassed, while a these authors considered the S. Miguel end-member relatively undegassed reservoir is sampled by deep ( 206Pb/204Pb=20.00; 207Pb/ 204Pb= 15.75; mantle plumes generating OIB [8,12]. 208Pb/204Pb=40.33) resulting from the presence in More recent studies on noble gases have shown the mantle source of subcontinental lithosphere that the two-layer mantle model is confirmed by delaminated during the opening of the Atlantic. For neon isotopic data (e.g. [13–16]). 20Ne and 22Ne are the Terceira end-member, the radiogenic lead isotopic almost primitive isotopes (the nucleogenic produc- signatures (206Pb/204Pb=20.02; 207Pb/204Pb=15.64) tions of 20Ne and 22Ne can be neglected in mantle coupled with the unradiogenic 4He/3He signatures, rocks) and similar 20Ne/22Ne ratios were found for were assigned to mixing between ancient recycled MORB and OIB samples [13–15,17–21]. Different oceanic crust and high 3He material from the lower 21Ne/22Ne ratios are, however, assigned to MORB mantle. The influence of such low 4He/3He plume and OIB sources. Since 21Ne derives from a particle- signature on the Mid-Atlantic ridge basalts was induced nuclear reactions such as 18O(a,n)21Ne demonstrated by [33], who reported a significant [22], its production is connected with radiogenic He decrease of the MORB 4He/3He ratio in the vicinity formation. The higher 21Ne/22Ne ratio observed in of the Azores archipelago. However, given that the MORB as compared with OIB samples (for a given helium anomaly (4He/3He=64,000; R/Ra=11.3) 20Ne/22Ne ratio), is compatible with a higher time- measured in [1] was not sufficient to clearly constrain integrated (Th+U)/He ratio as previously predicted the origin of the Azores hotspot, we decided to analyze from the 4He/3He signatures for the more degassed both helium and neon isotopes in new olivine upper mantle (e.g [14,16]). Moreover, in oceanic phenocrysts sampled in Terceira lavas. Concerning islands where the mixing between primitive mantle Ne, these are the first data published for Azores and MORB sources or magmas is believed to occur, archipelago. it has been shown that Ne can be more sensitive than He in detecting contributions from relatively undegassed sources [16,20,23–26]. 2. Geology of Terceira and sample locations Notwithstanding that Azores archipelago emerges in the vicinity of the Mid-Atlantic Ridge, it has been The Azores archipelago is located in the North considered as a hotspot resulting from a mantle plume Atlantic near the triple junction where the Eurasian, [27–29] being characterized by a buoyancy flux African and American plates meet (Fig. 1). The similar to that reported, for example, for the Canary boundary between the Eurasian and African plates is Islands (e.g. [30]). Nevertheless, mantle plumes have currently considered as represented by the NW–SE been the locus of an intense debate and controversy lineament of S. Miguel–Terceira–Graciosa or, alter- still exists concerning their ultimate origin: DQ layer, natively, by the direction defined by S. Miguel–S. bottom of mantle transition zone and even the Jorge Islands. Several models have been proposed to asthenosphere have been considered as possible describe the plate kinematics along this boundary: solutions (e.g. [31]). Noble gas isotope geochemistry pure extension [34–37], pure strike-slip [38] and is a key tool to clear up this question given its ability combined extension and strike-slip (e.g. [39–43]). A P. Madureira et al. / Earth and Planetary Science Letters 233 (2005) 429–440 431

AZORES 42º N

2000m

M.A.R

40º N C Fl Terceira ri G Latitude SJ ft F T 38º N P SMi

SMa east Azores fracture zone

32º W28º W Longitude

Fig. 1. Location of Terceira Island in the Azores archipelago (adapted from [33]). These Islands are the surface expression of a large submarine plateau (M.A.R.—Mid-Atlantic Ridge. Islands from east to west: SMa—Santa Maria; SMi—Sa˜o Miguel; T—Terceira; G—Graciosa; SJ—Sa˜o Jorge; P—Pico; F—Faial; C—Corvo; Fl—Flores). discussion concerning the validity of these models is the volcanism associated with that boundary. As far beyond the aim of our work, but an important stated above, the locus of this extension remains extensional component is obvious taking into account controversial, but some authors argued that it is well

Fig. 2. Geological map of Terceira Island (adapted from [44]) with location of the samples analysed in this study. CP—Cinco Picos caldera; GM—Guilherme Moniz Caldera; PA—Pico Alto volcano; SB—Santa Ba´rbara caldera). 432 P. Madureira et al. / Earth and Planetary Science Letters 233 (2005) 429–440 expressed in Terceira Island, in the so-called Terceira basaltic to mugearitic composition (see Table 1 and rift (see Fig. 1; [34–37]). Appendix A). For one sample (PMT-24), a fraction of Terceira Island is slightly elongated in the E–W clinopyroxene was also analysed. Olivine was prefer- direction with an emerged area of about 382 km2. The entially chosen given its significantly lower U and Th relief increases towards west, where the highest contents as compared with clinopyroxene (e.g. [49]) elevation reaches 1021 m, and the topography is preventing the in-situ production of radiogenic 4He dominated by four main strato-volcanoes (Fig. 2; [44– and nucleogenic 21Ne. Moreover, clinopyroxenes are 46]): Cinco Picos, Guilherme Moniz, Pico Alto and more likely to retain crustal He (radiogenic) incorpo- Santa Ba´rbara. At least the later two are considered to rated into magma en route to the surface [50], given be active and a fissure zone connecting Guilherme the higher diffusivity of He in pyroxenes than in Moniz and Santa Ba´rbara calderas has been regarded olivines at magmatic temperatures [51]. Different lava as the surface expression of Terceira rift. Cinco Picos flows were sampled during the first field campaign in area comprises the oldest volcanic centre, currently 2001. Some olivine poor specimens were re-sampled dominated by a partially eroded caldera with a in 2002 and referenced with the same code (appended diameter of almost 7 km, the largest of the Azores with the suffix A). In order to prevent a hypothetical archipelago. Smaller cinder cones located in this area influence of cosmogenic helium and neon on the materialize minor and more recent eruptions that determined isotopic ratios, most samples were col- partially fill this caldera; their deposits are particularly lected from road-cuts or young sea-cliff faces (see well exposed in the southeast coast of Terceira [45]. Appendix A). Fresh whole rock samples were gently From the geological map (Fig. 2)itcanbe crushed and sieved (sieves with apertures of 1, 1.4, 2 concluded that both pyroclasts and lava flows are and 4 mm) regarding the different dimensions of common,rangingincompositionfromprimitive olivine crystals. Grains were then handpicked with the basalts to comendites and pantellerites [46,47]. aid of a table light and a binocular microscope. All Despite the work of Self [46], where stratigraphic crystals with iddingsitic alteration rims, visible intra- relations between several volcanic units younger than granular fractures and/or attached basaltic matrix were 23,000 yrs are described, absolute age data is still very excluded for analysis. Moreover, homogeneous oli- scarce. Recent basaltic lavas are recognized to have vine separates concerning their petrographic charac- mainly erupted from the fissure zone (the youngest teristics were, whenever possible, always prepared. eruption occurred in 1761) sometimes forming exten- Mineral separates were ultrasonically cleaned in sive flows towards both northern and southern coasts, distilled water and submitted to HF reaction for 15 but little is known about the early evolution of the to 30 min to remove the external surface that could be island. Nevertheless, K/Ar measurements from Feraud in contact with the basaltic matrix. The final step of et al. [48] have shown that Cinco Picos caldera was cleaning was achieved with ethanol and acetone. Most formed less than 300,000 yrs ago. Sample locations phenocryst separates were prepared with a minimum are given in Fig. 2 and a brief petrographic description weight of ~ 0.7 g, in order to obtain a good accuracy of the samples can be found in Appendix A. Most in the analytical results. samples resulted from eruption events described by Volatiles were extracted by crushing in vacuum Self [46] (younger than 23,000 yrs). For samples (150 to 220 strokes) because this method mainly PMT-24 and PMT-71A to 74A the ages are more releases magmatic gases trapped in phenocryst melt uncertain. inclusions (e.g. [52]). Helium and neon were analyzed in Paris during 2003 with the mass spectrometer ARESIBO I [53], equipped with a new fully automated 3. Analytical procedure and results line with low He and Ne blanks. The new analytical procedure will be given elsewhere [54]. Mean 4He and 3.1. Analytical procedure 22Ne blanks were 3.6F0.71010 cm3STP and 1.5F0.310l3 cm3STP respectively (uncertainties Noble gases were mainly measured from green to are 1r), with isotopic ratios similar to the air ratios. yellow olivine phenocrysts, representing 18 samples of 22Ne blank proportions were between 2% and 75% P. Madureira et al. / Earth and Planetary Science Letters 233 (2005) 429–440 433

Table 1 Helium and neon isotopes in Terceira samples 4 4 3 20 22 21 22 22 21 22 Sample Diameter Weight He R/Ra F He/ He F Ne/ Ne s Ne/ Ne s Ne Ne/ Necorr (mm) (g) (ncm3STP/g) (pcm3STP/g) PMT8 1–1.4 0.88 31.0 9.8 0.2 73729 1505 9.92 0.04 0.0292 0.0004 3.8 PMT9 1–1.4 0.80 46.0 9.8 0.1 73729 752 9.87 0.05 0.0299 0.0003 2.9 PMT24-1 1.4–2 0.89 7.2 11.4 0.7 63381 3892 9.89 0.09 0.0276 0.0007 0.6 PMT24-2 2–4 0.67 37.0 10.5 0.1 68814 655 9.97 0.06 0.0301 0.0004 3.0 PMT24-3 1–3 0.45 13.0 10.2 0.7 70768 4713 9.89 0.05 0.0293 0.0006 1.5 PMT24-4 1–1.4 0.79 10.0 9.0 0.7 80283 6244 9.84 0.04 0.0300 0.0007 1.3 PMT24 Pyr 1.4–2 0.83 13.0 9.1 0.5 79400 4363 9.85 0.04 0.0297 0.0003 5.7 PMT79 1–1.4 0.99 35.0 9.8 0.1 73729 752 PMT19 1–1.4 1.00 32.0 9.7 0.1 74489 768 10.11 0.08 0.0307 0.0005 0.9 0.051 PMT4 1–1.4 0.89 26.0 10.0 0.1 72254 723 10.68 0.18 0.0321 0.0017 0.2 0.043 PMT61 1–1.4 1.81 39.0 9.6 0.1 75265 784 10.45 0.05 0.0316 0.0005 0.7 0.045 PMT89 1–1.4 0.98 63.0 9.3 0.1 77693 835 9.93 0.05 0.0298 0.0004 2.7 PMT88 1–1.4 1.00 33.0 9.4 0.1 76866 818 10.19 0.07 0.0327 0.0004 0.7 0.067 PMT98 1–1.4 1.08 38.0 9.9 0.1 72984 737 11.28 0.23 0.0373 0.0020 0.2 0.052 PMT27 0.5–1 1.60 40.0 9.3 0.3 77693 2506 9.84 0.02 0.0291 0.0002 7.7 PMTlOl-1 1–2 0.27 24.0 8.6 0.7 84017 6839 9.86 0.15 0.0294 0.0012 0.9 PMTlOl-2 1–2 0.36 30.0 9.6 0.6 75265 4704 10.35 0.4 0.0353 0.0049 0.2 0.075 PMTlOl-3 1–2 0.67 37.0 9.7 0.2 74489 1536 9.78 0.04 0.0291 0.0003 10.7 PMT101-4 1–2 0.40 14.0 9.1 0.5 79400 4363 9.74 0.04 0.0292 0.0002 16.4 (Brown oliv) PMT57 1.4–2 0.71 50.0 9.8 0.1 73654 751 9.87 0.03 0.0290 0.0003 3.9 PMT56 1.4–2 1.03 15.0 9.6 0.3 75265 2352 9.85 0.06 0.0287 0.0006 0.9

Sta Ba´rbara sequencea PMT7lA 1–1.4 0.90 5.0 10.6 0.9 68164 5788 9.93 0.07 0.0303 0.0006 1.0 PMT72 1–3 0.72 4.8 14.8 1.1 48820 3629 9.92 0.06 0.0300 0.0007 0.9 PMT73 1–3 0.73 13.0 10.8 0.4 66902 2478 9.87 0.05 0.0289 0.0003 2.7 PMT74A 1–3 0.98 10.0 8.7 0.4 83051 3818 9.73 0.04 0.0288 0.0005 1.3 4He and 22Ne concentrations are in ncm3 STP/g and pcm3 STP/g respectively. Errors are given at the 1r uncertainty level. The extrapolated 21Ne/22Ne to a solar 20Ne/22Ne ratio is given in the last column (i.e. corrected for air contamination). Only samples with 20Ne/22Ne ratios higher than 10 are considered for the extrapolation. Several noble gas analyses were performed on olivine separates coming from samples PMT- 24 and PMT-101 regarding their different petrographic characteristics. List of samples are given from E to W along the island. a Samples from Santa Ba´rbara sequence are given from base to top.

(sample PMT-101). During the period of analyses, the cps of 40Ar2+ that corresponds generally to only few + 40 + 20 signal of CO2 and Ar were ~ 1000 cps and 150 cps percent of the neon signal (usually NeN1000 cps in 2+ respectively. This implied ~ 8 cps of CO2 and ~ 15 samples).

Table 2 Helium isotopes in Terceira samples analyzed by M.D. Kurz and M. Moreira at the Woods Hole Oceanographic Institution (WHOI) Sample Diameter Weight 4He R/Ra F 4He/3He F (mm) (g) (ncm3STP/g) PMT24 1.4–2 0.32 16.6 10.0 0.1 71988 617 PMT71A 1.4–2 0.33 14.4 10.5 0.1 68618 613 PMT72A 1–3 0.22 4.7 9.9 0.3 73139 1866 PMT73 1–3 0.15 5.6 10.1 0.2 71264 1736 PMT74A (Brown oliv) 1–3 0.14 19.2 10.0 0.1 72596 1028 ACO95-17 1–2 0.27 9.5 9.9 0.1 72676 716 ACO95-18 1–2 0.23 17.6 9.4 0.1 76622 528 AC095-17 and AC095-18 are new data from collection of [1]. 434 P. Madureira et al. / Earth and Planetary Science Letters 233 (2005) 429–440

Seven samples were also analyzed for helium at the the same sample or lava flow. Although most of these Woods Hole Oceanographic Institution (WHOI) with ratios can be considered similar at the 1r uncertainty the MS2 instrument. The analytical procedure can be level, a detailed petrographic and electron-micro- found in Kurz et al. [55]. 4He blank was ~ 910 11 probe study [56] shows that, within a single rock cm3STP and represented between 2% and 10% of the sample, olivine crystals with different major element signal. compositions can be characterized by different He isotopic signatures. 3.2. Helium Concentrations of helium in the samples analyzed at WHOI are between 4.7 and 19.2 ncm3STP/g (Table Helium results from all samples analyzed in Paris 2), consistent with the measurements made in Paris. are shown in Table 1 and a plot of He concentration The 4He/3He ratio varies between 68,620F610 and versus 4He/3He is given in Fig. 3.Total4He 76,620F530 (R/Ra=10.5F0.1 and 9.4F0.1 respec- concentrations vary from 4.8 to 63 ncm3STP/g. The tively). Data are also reported in Fig. 3. Similar isotopic ratios range between typical lower MORB isotopic ratios were obtained for samples PMT-24, values (e.g. sample PMT-74A, 4He/3He=83,050F PMT-71A and PMT-73 in both laboratories. However, 3820; R/Ra=8.7F0.4) and a more primitive 4He/3He a significant difference in 4He/3He ratio can be ratio of 48,820F3630 for sample PMT-72 (R/ observed for sample PMT-74A (83,050 vs. 72,600) Ra=14.8F1.1). There is not a simple correlation and, particularly, for the most primitive sample (PMT- between isotopic ratios and He concentration, 72) analyzed in Paris (48,820 vs. 73,140). For sample because the later depends on the volume of melt PMT-74A, the difference is of the same order as that inclusions trapped in phenocrysts as well on the obtained for some olivine groups coming from one crushing efficiency. However, there is a tendency for single rock sample and can also be ascribed to some low 4He/3He ratios to appear in low He concentration distinctive petrographic features (namely the color, samples. Also relevant are the different ratios see Table 2). Sample PMT-72 is a helium poor sample obtained for distinct olivine separates coming from (4.8 ncm3STP/g) and therefore the difference in

Terceira This study Paris 120000 Moreira et al., 1999

This Study WHOI 100000 7

MORB

80000 9 R/Ra He 3 60000 12 He/ 4

40000 18

20000 36 Baffin island 10 20 30 40 50 60 4He (nCCSTP/g)

Fig. 3. 4He concentration versus 4He/3He in Terceira samples. Data from this study reach a minimum 4He/3He ratio of 48,820F3630 (see text). 1r uncertainties are shown. P. Madureira et al. / Earth and Planetary Science Letters 233 (2005) 429–440 435

4He/3He between the two laboratories could be attributed to an analytical problem. Our glass machine SW ARESIBO I has a low resolution power (~ 200) that 13.0 does not allow a peak separation between HD+ and 3He+. Nevertheless, the HD+ peak is always lower + Iceland line MORB line than 2–3 cps (H2 is lower than 8000 cps in the tube) Ne 12.0 and the HD contribution is corrected in all the 22 fractionation line samples, blanks and standards by measuring H 2 Ne/ 3 4 together with He and He. We introduced in the 20 mass spectrometer ~ 7 cps of 3He for sample PMT-72 11.0 and the 4He blank corresponded to ~ 9% of the 4He signal. However, sample PMT-71A, that gives the exactly same ratio in both laboratories (68,160 vs. 10.0 68,620) has roughly the same introduced 3He (9.5 AIR cps) with also ~ 9% of 4He blank. Moreover, the two 0.03 0.04 0.05 0.06 0.07 samples were analyzed in the same week and the 21Ne/22Ne machine was very stable during the period of Fig. 4. Neon three-isotope diagram. Ne isotopic data for Terceira analyses. Therefore we exclude an analytical problem Island define a trend line lying above the typical MORB trend [57]. 4 3 for the large difference in He/ He measured in 1r uncertainties are shown. samples PMT-72 and PMT-72A. Despite the possibility for the lowest 4He/3He contribution of a lower mantle reservoir to the isotopic ratio to represent the Azores plume signature, Terceira magma source. The 20Ne/22Ne ratios are it is important to emphasize that the low He lower than the mantle 20Ne/22Ne ratio (N 12.5) and concentration and the uncertainty relative to the age can be currently interpreted as reflecting atmospheric of PMT-72 lava flow make the 4He/3He ratio contamination (e.g. [14,58]). particularly prone to possible 3He cosmogenic con- A regression line (R2 ~0.9) was calculated for tributions. Indeed, experiments led by Yokochi et al. Terceira samples and projected in Fig. 4. The obtained [57] show that a significant proportion (up to 25%) of trend line passes through the point with the highest crystalline matrix-sited He can be released during 20Ne/22Ne ratio (sample PMT-98), which simultane- crushing trough lattice distortion and shock- and ously represents the best analytical accuracy as a result dislocation-induced diffusion. Notwithstanding that of minor atmospheric contamination. It is possible to an extremely low 3He concentration (lower than extend this line up to the primitive mantle 20Ne/22Ne 2.210 14 cm3STP/g, similar to blank values) was ratio in order to obtain a 21Ne/22Ne ratio corrected for 21 22 measured after melting the olivine powder, we cannot atmospheric contamination ( Ne/ Necorr) [15].A completely discard the possibility of a 3He cosmo- corrected ratio of 0.052 was obtained using a solar genic contribution for the low 4He/3He isotopic ratio 20Ne/22Ne ratio of 13.8 [59]. This value is significantly obtained in Paris for this sample. lower than the MORB ratio (0.075) [60]. It should be noticed that the trend line was calculated after 3.3. Neon excluding the olivine separate PMT-101-2 as a result of its large uncertainty at the 1r level. Neon results are shown in Table 1. 22Ne concentrations vary between 0.2 and 16.4 pcm3STP/g and the three Ne isotope diagram is given in Fig. 4. Six of the 4. Discussion olivine separates, representing three different lava flows from the fissure zone (Table 1 and Fig. 2), are 4.1. Origin of the Azores hotspot placed above the atmospheric values at the 1r uncertainty level. These samples plot above the Anomalous topography, gravity, crustal thickness, MORB line as determined by [58], confirming the S-wave velocities and geochemical signatures have 436 P. Madureira et al. / Earth and Planetary Science Letters 233 (2005) 429–440 been used as evidence for the existence of a plume in (e.g. [14]). In the following, we will discuss two the Azores region ([11,27,61–63] and references possibilities for the variations of the He and Ne therein). However, the relatively shallow character of isotopic ratios. One is a closed system evolution of a the S-wave negative anomalous zone (b 300 km), the heterogeneous mantle plume, whereas a second lack of a well defined hotspot track, the recent 3D re- possibility is mixing between two different sources. interpretation of admittance in the area [64] and some geochemical arguments (e.g. [65]) have been used 4.2.1. Heterogeneous 3He/22Ne plume source against a deep plume hypothesis applied to Azores. The assessment of mantle signatures can be made, Moreover, Anderson [66] proposed that hotspots could if an initial isotope composition of the solid Earth is be explained by asthenospheric upwelling as a passive assumed, through the determination of the time- response to breakup which, given the tectonic com- integrated (degassing independent) ratio (designated 3 22 plexity of the Azores region, could be considered as an as He/ Necalc by [24]) inferred from isotope attractive working hypothesis. compositions of He and Ne [67]: Noble gas signatures, enabling the discrimination 3He=22Ne between involvement of degassed and undegassed calc (primitive) mantle reservoirs in the magma genesis 21 22 21 22 4 21 ¼ Ne= Necorr Ne= Neinit He= Ne4prod must be considered of utmost importance to unravel the 4 3 4 3 ultimate origin of the Azores magmatism. The pre- 1= He= Heobs He= Heinit viously existent noble gas determinations in the Azores 4 21 basalts have shown the existence of some unradiogenic where He/ Ne*prod represents the mantle production He signatures as compared with MORB. For Terceira ratio, considered constant and estimated to be equal to 4 3 7 4 3 Island, He/ He ratios as low as 64,000 (R/Ra=11.3) 2.210 [68], and He/ Heobs stands for the meas- were interpreted as reflecting the existence of a ured He isotope ratio. The primordial mantle 4 3 21 22 contribution from the deep mantle [1]. A deep plume He/ Heinit and Ne/ Neinit ratios can be assigned component is strengthened by the first Ne data now to the solar nebula or to the meteoritic composition obtained for the Azores archipelago through the study Ne–B [21,69]. It is beyond the scope of the paper to 21 22 of Terceira Island ( Ne/ Necorr =0.052), which are discuss the exact primitive ratios of the mantle and, to clearly indicative of the involvement of the lower be consistent, we will consider solar-like lower mantle mantle in the Azores magmas. In this context, the ratios. Nevertheless, it should be noticed that consid- shallow character of the S-wave anomaly in the area ering the primitive ratios as Ne–B, the conclusions of shall be considered as reflecting the presence of a dying this paper do not change. 3 22 and nowadays untailed plume, the head of which The mean He/ Necalc ratios estimated for Terceira underplates the Azores Plateau [63]. samples with 20Ne/22Ne greater than 10 at 1r In the next section we will use He–Ne systematics uncertainty level, range between 10.6F1.4 and to better constrain the Terceira mantle source. 3.4F2.6 with a mean value around 5 (Fig. 5). These values are intermediate to those estimated for MORB 4.2. Mixing vs. heterogeneous plume source and Iceland (7 and 4 respectively) [20,60,67] and could be interpreted as reflecting the presence of long- Given that 3He and 22Ne isotopes are considered to lived heterogeneities in the mantle plume source in be primordial one might expect that the mantle terms of the 3He/22Ne and (Th+U)/3He ratios (Fig. 5). 3He/22Ne ratio of a mantle reservoir is constant over However, considering the proximity of Terceira time (they are both highly incompatible elements), Island to Mid-Atlantic Ridge, we will now evaluate the provided that the system remained closed and no He– hypothesis of mixing to generate the low 21Ne/22Ne Ne fractionation occurred. On the other hand, mantle ratios in association with MORB-like helium ratios. sources characterized by a given 3He/22Ne will be affected by coupled changes in the 4He/3He and 4.2.2. Binary mixing 21Ne/22Ne relative to their initial values, since both The binary mixing model can be tested in a plot of 4 21 4 3 21 22 He and Ne are produced from decay of U and Th He/ He vs. Ne/ Necorr, where the end-members are P. Madureira et al. / Earth and Planetary Science Letters 233 (2005) 429–440 437

of magnitude degassing estimated by Moreira and 0.08 MORB Alle`gre [33] for MORB in the vicinity of the Azores Plateau. The 3He/22Ne ratios determined for Terceira 0.07 11 magmas are thus envisaged as resulting from mixing between a degassed MORB magma and an unde- 7

Ne 0.06 gassed plume-like magma (or distillated, see below). 22 5 It remains to find an explanation for the apparently

Ne/ 0.05 less degassed nature of the Terceira plume magma 21 3 when compared with local MORB. This can be argued 0.04 in terms of the model proposed by Dixon et al. [72] and Clague et al. [73] for Kilauea and Puna ridge 0.03 (Hawaii). The model states that the degassed and, as a Solar consequence, denser shallow magmas can sink 0 25000 50000 75000 through a long (~4 km) magma column, thus enabling 4He/3He magma mixing at deeper levels. More recently, this model was also envisaged by Natland [74] for Juan 21 22 4 3 Fig. 5. Ne/ Necorr ratios plotted against He/ He. Straight lines Fernandez Islands (SE Pacific). Such a model can represent the temporal evolution of reservoirs with different explain the less fractionated nature of plume related 3He/22Ne ratios (indicated on the figure) [20]. magmas, providing a mechanism for magma mixing to occur at a sufficient depth to prevent significant considered to be MORB upper mantle and a primitive fractionation induced by volatile loss. In our study, we plume component with solar-like He and Ne isotopic cannot decide if the plume-like magma is not ratios (Fig. 6).Thedegreeofrelative3He/22Ne degassed or only degassed following a distillation elemental fractionation is described by the parameter process as it is the case for Shona and Discovery ridge 3 22 3 22 r =[( He/ Ne)MORB/( He/ Ne)plume] [20,24]. Hyperbolic mixing lines with r-values close to 10 encompass the majority of data implying that, during MORB mixing, the MORB 3He/22Ne ratio was ~ 10 times 0.08 higher than in the plume component, here considered as represented by the solar value. r-values similar to 0.07 those here determined for Terceira were reported by r=1

Ne 0.06

[15,20,26,70] for Iceland, Shona, Discovery and East 22 Pacific Rise (178S) and interpreted in terms of magma Ne/ 0.05 mixing. These authors explained the significantly 21 Loihi 3 22 higher He/ Ne MORB ratio as the result of He–Ne 0.04 fractionation promoted by magma degassing in a r=10 relatively near surface environment. Indeed, since He 0.03 Solar and Ne have different melt solubilities (higher for He) the elemental He/Ne ratio of a basaltic melt increases with progressive volatile loss [71]. In the case of 0 25000 50000 75000 Loihi, located in a clear within-plate setting, near 4He/3He linear mixing (r =1) was interpreted as taking place at Fig. 6. Binary mixing model between the MORB (for the North greater mantle depths, inhibiting the magma degassing Atlantic) and a solar type mantle end-members. To prevent large processes to occur. Considering that, like Iceland and corrections and uncertainties, only samples with 20Ne/22Ne ratios Shona-Discovery, the Azores hotspot is also located greater than 10 are plotted as well samples with uncertainties lower than 30%. Terceira data can be fitted by a hyperbolic mixing line close to an oceanic ridge, a similar model can be 3 22 3 22 3 22 with an r value (=( He/ Ne)MORB/( He/ Ne)solar) close to 10. considered to explain the high He/ Ne MORB ratio Small dots are from [15,20,26,69] and represent different plume– during mixing. This idea is endorsed by a three order ridge interactions. Loihi point is plotted from [19]. 438 P. Madureira et al. / Earth and Planetary Science Letters 233 (2005) 429–440 anomalies (that are also plume-ridge interactions) 15255/99), and by a CRUP/CPU joint research [75]. Argon measurements are required to evaluate program. The authors are indebted to M. Kurz for the most feasible hypothesis. We did not analyze the He analyses performed at WHOI. PM and JM are argon in our samples in order to decrease the 40Ar indebted to G. Queiroz, J. Gaspar and T. Ferreira by memory during neon analyses. their guidance and support during the field work, Another explanation for the high r-value is that the while N. Lourenc¸o, M. Miranda, P. Silva and P. plume source component has a 3He/22Ne ratio lower Moita are thanked for fruitful discussions on the than 2 (i.e. solar-like), 5 to 10 times lower than the Azores geology and for their encouragement. Centro ratio in the MORB source reservoir (~7–10) [60]. de Geologia da Univ. Lisboa (POCA) and Departa- Therefore, in this hypothesis, there is no need to mix mento de Geocieˆncias da Univ. E´ vora are acknowl- degassed MORB magma and the mixing probably edged for facilities provided. Comments by S. occurs at depth. The acceptance of such a model, Niedermann and P. Bumard were highly constructive which attributes a very low 3He/22Ne ratio to the and helped to improve the manuscript. This is IPGP plume source, has fundamental consequences for the contribution no. 2053. origin, distribution and evolution of volatile elements in the Earth as well as for the dynamics of the terrestrial mantle. Appendix A. Supplementary data

Supplementary data associated with this article can 5. Conclusions be found, in the online version, at doi:10.1016/ j.epsl.2005.02.030. We presented the first neon isotopic ratios on the Azores archipelago together with new helium data for Terceira Island. Ne and He isotopic data argue for a References primitive mantle contribution to Terceira magmas, [1] M. Moreira, R. Doucelance, B. Dupre´, M. Kurz, C.J. Alle`gre, thus endorsing the models invoking the role of a deep Helium and lead isotope geochemistry in the Azores archipe- mantle plume in the generation of the Azores Islands. lago, Earth Planet. Sci. Lett. 169 (1999) 189–205. We propose that the Terceira He and Ne signatures [2] C.J. Alle`gre, M. Moreira, T. Staudacher, 4He/3He dispersion result of mixing between two types of magmas: and mantle convection, Geophys. Res. Lett. 22 (17) (1995) MORB and solar-like plume. As indicated by the 2325–2328. [3] K.A. Farley, E. Neroda, Noble gases in the Earth’s mantle, positioning of Terceira samples close to the MORB Annu. Rev. Earth Planet. Sci. 26 (1998) 189–218. end-member (Fig. 6) this component can be consid- [4] F.M. Stuart, S. Lass-Evans, J.G. Fitton, R.M. Ellam, High ered as prevalent in the mixing process, particularly 3He/4He ratios in picritic basalts from Baffin Island and the concerning helium. The large influence of the MORB role of a mixed reservoir in mantle plumes, Nature 424 (2003) end-member can be envisaged as a consequence of the 57–59. [5] M.D. Kurz, W.J. Jenkins, S.R. Hart, Helium isotopic system- proximity to Mid-Atlantic ridge, explaining the 4 3 atics of oceanic islands and mantle heterogeneity, Nature 297 majority of He/ He ratios found for Terceira samples (1982) 43–47. only slightly lower than the typical MORB value. Our [6] M.D. Kurz, W.J. Jenkins, S.R. Hart, D. Clague, Helium isotopic study highlights the importance of neon data in variations in volcanic rocks from Loihi Seamount and the island getting the plume signature of a mixture dominated of Hawaii, Earth Planet. Sci. Lett. 66 (1983) 388–406. [7] D.R. Hilton, K. Grfnvold, C. Macpherson, P. Castillo, by MORB. Extreme 3He/4He ratios in northwest Iceland: constraining the common component in mantle plumes, Earth Planet. Sci. Lett. 173 (1999) 53–60. Acknowledgements [8] C.J. Alle`gre, T. Staudacher, P. Sarda, Rare gas systematics, formation of the atmosphere, evolution and structure of the Earth’s mantle, Earth Planet. Sci. Lett. 81 (1986) 127–150. This work was partially financed by FCT, through [9] Y. Ji, H.C. Nataf, Detection of mantle plumes in the lower the project STAMINA (Study of the Tectono-mag- mantle by diffraction tomography: Hawaii, Earth Planet. Sci. matic Interplay in the Azores) (PDCTM/PP/MAR/ Lett. 159 (1998) 99–115. P. Madureira et al. / Earth and Planetary Science Letters 233 (2005) 429–440 439

[10] H. Bijwaard, W. Spakman, Tomographic evidence for a [29] W.M. White, M.D.M. Tapia, J.-G. Schilling, The petrology narrow whole mantle plume below Iceland, Earth Planet. and geochemistry of the Azores Islands, Contrib. Mineral. Sci. Lett. 166 (1999) 121–126. Petrol. 69 (1979) 201–213. [11] J.-P. Montagner, J. Ritsema, Interactions between ridges and [30] N. Sleep, Hotspots and mantle planes: some phenomenology, plumes, Science 294 (2001) 1472–1473. J. Geophys. Res. 95 (1990) 6175–6736. [12] C.J. Alle`gre, Isotope geodynamics, Earth Planet. Sci. Lett. 86 [31] V. Courtillot, A. Davaille, J. Besse, J. Stock, Three distinct (1987) 175–203. types of hotspots in the Earth’s mantle, Earth Planet. Sci. Lett. [13] M. Honda, I. McDougall, D.B. Patterson, A. Doulgeris, D. 205 (2003) 295–308. Clague, Possible solar noble-gas component in Hawaiian [32] M.D. Kurz, W.J. Jenkins, J.-G. Schilling, S.R. Hart, Helium basalts, Nature 349 (1991) 149–151. isotopic variation in the mantle beneath the central North [14] M. Honda, I. McDougall, D.B. Patterson, A. Doulgeris, D. Atlantic Ocean, Earth Planet. Sci. Lett. 58 (1982) 1–14. Clague, Noble gases in submarine pillow basalt glasses from [33] M. Moreira, C.J. Alle`gre, Rare gas systematics on Mid Loihi and Kilauea, Hawaii: a solar component in the Earth, Atlantic Ridge (378–408), Earth Planet. Sci. Lett. 198 (2002) Geochim. Cosmochim. Acta 57 (1993) 859–874. 401–416. [15] M. Moreira, T. Staudacher, P. Sarda, J.-G. Schilling, C.J. [34] D.C. Krause, N.D. Watkins, North Atlantic crustal genesis in Alle`gre, A primitive plume neon component in MORB: the the vicinity of the Azores, Geophys. J. R. Astron. Soc. 19 Shona ridge-anomaly, South Atlantic (51–528S), Earth Planet. (1970) 261–283. Sci. Lett. 133 (1995) 367–377. [35] A. Udıás, Seismic stresses in the region Azores–Spain– [16] M. Moreira, C.J. Allegre, Helium–Neon systematics and the Western Mediterranean, Rock Mech. 9 (1980) 75–84. structure of the mantle, Chem. Geol. 147 (1998) 53–59. [36] A. Udıás, A.F. Espinoza, J. Mezcua, E. Bufom, R. Vegas, S.P. [17] H. Hiyagon, M. Ozima, B. Marty, S. Zashu, H. Sakai, Noble Nishenko, J.M. Martinez-Solares, A. Lopez-Arroyo, Seismic- gases in submarine glasses from mid-oceanic ridges and Loihi ity and tectonics of the North African–Eurasian plate boundary seamount: constraints on the early history of the Earth, (Azores–Iberia–Tunisia), Open-File Rep. (U. S. Geol. Surv.) Geochim. Cosmochim. Acta 56 (1992) 1301–13016. (1986) 86–626. [18] R.J. Poreda, K.A. Farley, Rare gases in Samoan xenoliths, [37] E. Bufom, A. Udıás, M.A. Colomba´s, Seismicity, source Earth Planet. Sci. Lett. 113 (1992) 129–144. mechanisms and tectonics of the Azores–Gibraltar plate [19] P.J. Valbracht, T. Staudacher, A. Malahoff, C.J. Alle`gre, Noble boundary, Tectonophysics 152 (1988) 89–118. gas systematics of deep rift zone glasses from Loihi seamount, [38] V.H. Forjaz, Azores Study Tour. Field Trip Guide: Seminar on Hawaii, Earth Planet. Sci. Lett. 150 (1997) 399–411. the Prediction of Earthquakes, Econ. Comm. for Europe-UN, [20] M. Moreira, K. Breddam, J. Curtice, M. Kurz, Solar neon in Lisbon, 1988, 26 pp. the Icelandic mantle: evidence for an undegassed lower [39] D. McKenzie, Active tectonics of the Mediterranean region, mantle, Earth Planet. Sci. Lett. 185 (2001) 15–23. Geophys. J. R. Astron. Soc. 30 (1972) 109–185. [21] M. Trieloff, J. Kunz, D.A. Clague, D. Harrison, C.J. Alle`gre, [40] R. Searle, Tectonic pattern of the Azores spreading centre and The nature of pristine noble gases in mantle plumes, Science triple junction, Earth Planet. Sci. Lett. 51 (1980) 415–434. 288 (2000) 1036–1038. [41] J. Madeira, A. Ribeiro, Geodynamic models for the Azores [22] G.W. Wetherill, Variations of isotopic abundances of neon and triple junction: a contribution from tectonics, Tectonophysics argon extracted from radioactive minerals, Phys. Rev. 96 (3) 184 (1990) 405–415. (1954) 679–683. [42] N. Lourenço, J.M. Miranda, J.F. Luis, A. Ribeiro, L.A. Mendes [23] S. Niedermann, W. Bach, J. Erzinger, Noble gas evidence for a Victor, J. Madeira, H.D. Needham, Morpho-tectonic analysis of lower mantle component in MORB from the southern East the Azores Volcanic Plateau from a new bathymetric compi- Pacific Rise: decoupling of helium and neon isotope system- lation of the area, Mar. Geophys. Res. 20 (1998) 141–156. atics, Geochim. Cosmochim. Acta 61 (1997) 2697–2715. [43] A. Ribeiro, Soft Plate and Impact Tectonics, Springer, Berlin, [24] E. Dixon, Interpretation of helium and neon isotopic hetero- 2002, 324 pp. geneity in Icelandic basalts, Earth Planet. Sci. Lett. 206 (2003) [44] G. Zbyszewski, A.C. Medeiros, O.d.V. Ferreira, C.T.d. Assun- 83–99. çaõ, Notıćia explicativa da folha Ilha Terceira na escala de 1/ [25] J. Hopp, M. Trieloff, R. Altherr, Neon isotopes in mantle rocks 50.000, Serviços Geolo´gicos de Portugal, Lisboa, (1971). from the Red Sea region reveal large-scale plume–lithosphere [45] M.S. Rosenbaum, Basaltic volcanism in south-east Terceira, interaction, Earth Planet. Sci. Lett. 219 (2004) 61–76. Azores, Geol. Mag. 111 (5) (1974) 409–420. [26] M.D. Kurz, M. Moreira, J. Curtice, D.E. Lott III, J.J. [46] S. Self, The recent volcanology of Terceira, Azores, Geol. Soc. Mahoney, J.M. Sinton, Correlated helium, neon, and melt Lond. 132 (1976) 645–666. production on the super-fast spreading East Pacific near 178S, [47] J.E. Mungall, R.F. Martin, Petrogenesis of basalt-comendite Earth Planet. Sci. Lett. 232 (2005) 125–142. and basalt-pantellerite suites, Terceira, Azores, and some [27] J.G. Schilling, Azores mantle blob: rare earth evidence, Earth implications for the origin of the ocean–Island rhyolites, Planet. Sci. Lett. 25 (1975) 103–115. Contrib. Mineral. Petrol. 119 (1995) 43–55. [28] W.M. White, J.G. Schilling, S.R. Hart, Evidence for the [48] G. Feraud, I. Kaneoka, C.J. Alle`gre, K/Ar ages and stress Azores mantle plume from strontium isotopes geochemistry of pattern in the Azores: geodynamic implications, Earth Planet. the Central North Atlantic, Nature 263 (1976) 659–663. Sci. Lett. 46 (1980) 275–286. 440 P. Madureira et al. / Earth and Planetary Science Letters 233 (2005) 429–440

[49] B.J. Wood, J.D. Blundy, J.A.C. Robinson, The role of [62] Y.-S. Zhang, T. Tanimoto, High-resolution global upper mantle clinopyroxene in generating U-series disequilibrium during structure and plate tectonics, J. Geophys. Res. 98 (1993) mantle melting, Geochim. Cosmochim. Acta 63 (1999) 9793–9823. 1613–1620. [63] G. Silveira, E. Stutzman, A. Davaille, J.P. Montagner, L. [50] D.R. Hilton, J. Barling, G.E. Wheller, Effect of shallow-level Mendes-Victor, Azores hotspot signature in the upper mantle, contamination on the helium isotope systematics of ocean– J. Volcanol. Geotherm. Res., (in press). island lavas, Nature 373 (1995) 330–333. [64] J.F. Luis, M.C. Neves, The isostatic compensation of Azores [51] T.W. Trull, M.D. Kurz, Experimental measurements of 3He Plateau: a 3D admittance and coherence analysis, J. Volcanol. and 4He mobility in olivine and clinopyroxene at magmatic Geotherm. Res., (in press). temperatures, Geochim. Cosmochim. Acta 57 (1993) [65] E. Bonatti, Not so bHot SpotsQ in the oceanic mantle, Science 1313–1324. 250 (1990) 107–110. [52] P. Scarsi, Fractional extraction of helium by crushing of [66] D.L. Anderson, The statistics and distribution of helium in the olivine and clinopyroxene phenocrysts: effects on the 3He/4He mantle, Int. Geol. Rev. 42 (2000) 289–311. measured ratio, Geochim. Cosmochim. Acta 64 (21) (2000) [67] M. Honda, I. Mc Dougall, Primordial helium and neon in the 3751–3762. Earth—a speculation on early degassing, Geophys. Res. Lett. [53] T. Staudacher, C.J. Alle`gre, Terrestrial xenology, Earth Planet. 25 (1998) 1951–1954. Sci. Lett. 60 (1982) 389–406. [68] I. Yatsevich, M. Honda, Production of nucleogenic neon in the [54] M. Moreira, P. Madureira, S. Doucet, A. Lecomte, C.J. Earth from natural radioactive decay, J. Geophys. Res. 102 Alle`gre, He and Neon isotopes systematics in OIB, Earth (1997) 10291–10298. Planet. Sci. Lett. (in preparation). [69] M. Trieloff, J. Kunz, C.J. Alle`gre, Noble gas systematics of the [55] M.D. Kurz, T.C. Kenna, J.C. Lassiter, D.J. DePaolo, Helium Reunion mantle plume source and the origin of primordial isotopic evolution of Mauna Kea volcano: first results from the noble gases in Earth’s mantle, Earth Planet. Sci. Lett. 200 1 Km drill core, J. Geophys. Res. 101 (1996) 11781–11791. (2002) 297–313. [56] P. Madureira, M. Moreira, J. Mata, G. Queiroz, Multiple [70] P. Sarda, M. Moreira, T. Staudacher, J.-G. Schilling, C.J. olivine helium isotopic signatures from a single rock sample: a Alle`gre, Rare gas systematics on the southernmost Mid- case study from Terceira Island (Azores archipelago) (in Atlantic Ridge: constraints on the lower mantle and the Dupal preparation). source, J. Geophys. Res. 105 (2000) 5973–5996. [57] R. Yokochi, B. Marty, R. Pik, P. Burnard, High 3He/4He ratios [71] A. Jambon, H. Weber, O. Braun, Solubility of He, Ne, Ar, Kr in peridotite xenoliths from SW Japan-Revisited: evidence for and Xe in a basalt melting in the range 1250–1600 8C. cosmogenic 3He released by vacuum crushing, Geochem. Geochemical implications, Geochim. Cosmochim. Acta 50 Geophys. Geosyst. 6 (1) (2005) Q01004. (1986) 401–408. [58] P. Sarda, T. Staudacher, C.J. Alle`gre, Neon isotopes in [72] J.E. Dixon, D. Clague, E. Stolper, Degassing history of water, submarine basalts, Earth Planet. Sci. Lett. 91 (1988) 73–88. sulfur, and carbon in submarine lavas from Hilauea volcano, [59] J.-P. Benkert, H. Baur, P. Signer, R. Wieler, He, Ne and Ar Hawaii, J. Geol. 99 (1991) 371–394. from the solar wind and solar energetic particles in lunar [73] D. Clague, J. Moore, J. Dixon, W. Friesen, Petrology of ilmenites and pyroxenes, J. Geophys. Res. 98 (1993) submarine lavas from Kilauea’s Puna Ridge, Hawaai, J. Petrol. 13147–13162. 36 (1995) 299–349. [60] M. Moreira, J. Kunz, C.J. Alle`gre, Rare gas systematics on [74] J.H. Natland, Capture of helium and other volatiles during the popping rock: estimates of isotopic and elemental composi- growth of olivine phenocrysts in picritic basalts from the Juan tions in the upper mantle, Science 279 (1998) 1178–1181. Femandez Islands, J. Petrol. 44 (2003) 421–456. [61] Y.-S. Zhang, T. Tanimoto, Ridges, hotspots and their [75] M. Moreira, P. Sarda, Noble gas constraints on degassing interaction as observed in seismic velocity maps, Nature 355 processes, Earth Planet. Sci. Lett. 176 (2000) 375–386. (1992) 45–49.