helium and lead isotope geochemistry of the azores archipelago

17
ELSEVIER Earth and Planetary Science Letters 169 (1999) 189–205 Helium and lead isotope geochemistry of the Azores Archipelago Manuel Moreira a,b,L , Re ´gis Doucelance a , Mark D. Kurz b , Bernard Dupre ´ a,c , Claude Jean Alle `gre a a Laboratoire de Ge ´ochimie et Cosmochimie, URA CNRS 1758, Institut de Physique du Globe de Paris, Universite ´ Denis Diderot Paris 7, 4 Place Jussieu, 75252, Paris Cedex 05, France b Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, 360 Woods Hole Road, MS25, Woods Hole, MA 02543, USA c Laboratoire de Geochimie, OMP, UMR-CNRS 5563, 38 rue des 36 ponts, 31400, Toulouse, France Received 21 April 1998; revised version received 10 March 1999; accepted 11 March 1999 Abstract New helium and lead isotopic data for basalts from the Azores archipelago (North Atlantic) show that the Azores have 4 He= 3 He ratios both higher and lower than MORB values. Good covariations of helium and lead isotopes are observed at the scale of the archipelago, and suggest the coexistence of two mantle components in the Azores which are identified by data from Sa ˜o Miguel and Terceira. The eastern part of Sa ˜o Miguel island displays radiogenic helium ( 4 He= 3 He > 140,000, R= R a < 5:1) and lead (20.00, 15.75 and 40.33 for 206 Pb= 204 Pb, 207 Pb= 204 Pb and 208 Pb= 204 Pb). The 207 Pb= 204 Pb and 208 Pb= 204 Pb ratios for Sa ˜o Miguel are unusually radiogenic for oceanic basalts. Terceira basalts contain relatively unradiogenic=primitive 4 He= 3 He ratios, with a minimum value of 64,000 ( R= R a D 11:3), and relatively high lead isotopic ratios ( 206 Pb= 204 Pb D 20.02, 207 Pb= 204 Pb D 15.64 and 208 Pb= 204 Pb D 39.35). We propose that the Terceira source has a composition produced by a mixing between recycled oceanic crust (high 206 Pb= 204 Pb) and entrained lower mantle (high 3 He) material. The Sa ˜o Miguel island isotopic signature may be due to sampling of local (km-size) heterogeneity located at relatively shallow depth. The preferred origin of this heterogeneity is the Jurassic delamination of subcontinental lithosphere, which occurred during rifting and opening of the North Atlantic. The primitive helium ratios were also observed on the Mid Atlantic ridge at 38.5ºN, reflecting plume–ridge interaction, whereas radiogenic ratios (>100,000) were observed at latitude higher than 40ºN and may reflect the influence of the Sa ˜o Miguel component at the ridge. 1999 Elsevier Science B.V. All rights reserved. Keywords: helium; lead; isotope ratios; Azores; magmas; mixing 1. Introduction Helium isotopic ratios measured on Oceanic Is- land Basalts (OIB) display very large variations L Corresponding author. Fax: C1 508 457 2193; E-mail: [email protected] compared to Mid Oceanic Ridge Basalts (MORB). The relatively homogeneous MORB source reser- voir has a 4 He= 3 He mean value of 88,000 š 5000 ( R= R a D 8) [1] whereas the 4 He= 3 He isotopic ratios observed in OIB range from 25,000 ( R= R a D 30) for Loihi seamount samples [2] to values higher than 150,000 ( R= R a D 4:8) for the Gough–Tristan da 0012-821X/99/$ – see front matter 1999 Elsevier Science B.V. All rights reserved. PII:S0012-821X(99)00071-0

Upload: manuel-moreira

Post on 01-Nov-2016

212 views

Category:

Documents


0 download

TRANSCRIPT

ELSEVIER Earth and Planetary Science Letters 169 (1999) 189–205

Helium and lead isotope geochemistry of the Azores Archipelago

Manuel Moreira a,b,Ł, Regis Doucelance a, Mark D. Kurz b, Bernard Dupre a,c,Claude Jean Allegre a

a Laboratoire de Geochimie et Cosmochimie, URA CNRS 1758, Institut de Physique du Globe de Paris,Universite Denis Diderot Paris 7, 4 Place Jussieu, 75252, Paris Cedex 05, France

b Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, 360 Woods Hole Road, MS25,Woods Hole, MA 02543, USA

c Laboratoire de Geochimie, OMP, UMR-CNRS 5563, 38 rue des 36 ponts,31400, Toulouse, France

Received 21 April 1998; revised version received 10 March 1999; accepted 11 March 1999

Abstract

New helium and lead isotopic data for basalts from the Azores archipelago (North Atlantic) show that the Azoreshave 4He=3He ratios both higher and lower than MORB values. Good covariations of helium and lead isotopes areobserved at the scale of the archipelago, and suggest the coexistence of two mantle components in the Azores whichare identified by data from Sao Miguel and Terceira. The eastern part of Sao Miguel island displays radiogenic helium(4He=3He > 140,000, R=Ra < 5:1) and lead (20.00, 15.75 and 40.33 for 206Pb=204Pb, 207Pb=204Pb and 208Pb=204Pb). The207Pb=204Pb and 208Pb=204Pb ratios for Sao Miguel are unusually radiogenic for oceanic basalts. Terceira basalts containrelatively unradiogenic=primitive 4He=3He ratios, with a minimum value of 64,000 (R=Ra D 11:3), and relatively high leadisotopic ratios (206Pb=204Pb D 20.02, 207Pb=204Pb D 15.64 and 208Pb=204Pb D 39.35). We propose that the Terceira sourcehas a composition produced by a mixing between recycled oceanic crust (high 206Pb=204Pb) and entrained lower mantle(high 3He) material. The Sao Miguel island isotopic signature may be due to sampling of local (km-size) heterogeneitylocated at relatively shallow depth. The preferred origin of this heterogeneity is the Jurassic delamination of subcontinentallithosphere, which occurred during rifting and opening of the North Atlantic. The primitive helium ratios were alsoobserved on the Mid Atlantic ridge at 38.5ºN, reflecting plume–ridge interaction, whereas radiogenic ratios (>100,000)were observed at latitude higher than 40ºN and may reflect the influence of the Sao Miguel component at the ridge. 1999 Elsevier Science B.V. All rights reserved.

Keywords: helium; lead; isotope ratios; Azores; magmas; mixing

1. Introduction

Helium isotopic ratios measured on Oceanic Is-land Basalts (OIB) display very large variations

Ł Corresponding author. Fax: C1 508 457 2193; E-mail:[email protected]

compared to Mid Oceanic Ridge Basalts (MORB).The relatively homogeneous MORB source reser-voir has a 4He=3He mean value of 88,000 š 5000(R=Ra D 8) [1] whereas the 4He=3He isotopic ratiosobserved in OIB range from 25,000 (R=Ra D 30)for Loihi seamount samples [2] to values higher than150,000 (R=Ra D 4:8) for the Gough–Tristan da

0012-821X/99/$ – see front matter 1999 Elsevier Science B.V. All rights reserved.PII: S 0 0 1 2 - 8 2 1 X ( 9 9 ) 0 0 0 7 1 - 0

190 M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205

Cunha–St Helena island group [3–5]. A two-layeredmantle best explains these results, where the uppermantle represents the MORB source, which is highlydegassed and well mixed by convection [1], whereasthe primordial signal observed in Loihi seamountbasalts is attributed to the presence of a less de-gassed reservoir, which is probably located in thelower mantle [2,6].

The significance of the radiogenic helium signa-tures observed in the Gough–Tristan da Cunha–StHelena island group remains controversial. Thesesignatures have been attributed to the injection,through subduction, of oceanic crust and sedimentsinto the upper mantle, which will present high (U CTh)=3He ratios and be stored for some time at the670 km boundary layer, prior to eruption [3,5,7]. Analternate view attributing the high 4He=3He ratiosto shallow, magma chamber contamination processes(unrelated to the mantle), originally discussed byCondomines et al. [8] and Zindler and Hart [9,10],was more recently defended by Hilton et al. [11].Based on results from Heard island (Indian Ocean),where both primitive and radiogenic helium isotopicratios have been measured, Hilton and co-work-ers propose that most the hotspots are in fact ‘low4He=3He’ hotspots, and the observed radiogenic he-lium signature may reflect shallow depth contamina-tion, either by interaction of the plume with oceaniccrust or radiogenic 4He production in a degassedmagma chamber.

Lead isotopes also display different histogramsfor MORB and OIB. Ocean island samples usuallypresent more radiogenic signatures and, again, ex-hibit more variability, than lavas from spreading cen-ters [10,12,13]. Differences in lead isotopic composi-tions between oceanic islands are still debated. Theycould reflect plume source heterogeneities [14,15],material entrainment during plume ascent [16,17] orinteraction between plumes and lithospheric mantle[18]. Due to the different decay constants, lead iso-topes are less sensitive to in situ production, andcannot be related to magma chamber outgassing.

This study focusses on the Azores archipelago,which is considered to belong to the ‘high 4He=3He’hotspot category [19–21], and presents new he-lium and lead isotopic results for six islands fromthe Azores archipelago: Faial, Graciosa, Pico, SaoMiguel, Santa Maria and Terceira. Good covaria-

tions between helium and lead isotopes are observedat the scale of the archipelago, and suggest thecoexistence of two components in the Azores; thetwo end members are best identified by data fromSao Miguel and Terceira islands. The Eastern partof Sao Miguel island displays radiogenic helium(4He=3He > 140,000) and lead (20.00, 15.75 and40.33 for 206Pb=204Pb, 207Pb=204Pb and 208Pb=204Pb)signatures, whereas Terceira basalts present primi-tive 4He=3He ratios, with a minimum value around64,000 (R=Ra D 11:3), similar 206Pb=204Pb ratio(20.02), but significantly lower 207Pb=204Pb (15.64)and 208Pb=204Pb (39.35). All other island results areinterpreted in terms of binary mixing between thelocal MORB mantle source and these two plumecomponents.

2. Geographical situation and sample locations

The Azores archipelago is composed of nine is-lands (Corvo, Faial, Flores, Graciosa, Pico, SantaMaria, Sao Jorge, Sao Miguel and Terceira), whichrepresent the emerged part of a large oceanic plat-form (Fig. 1). Located at the triple junction be-tween the Eurasian, African and American plates,the plateau is crossed over by the Mid-Atlanticridge; seven islands (Faial, Graciosa, Pico, SantaMaria, Sao Jorge, Sao Miguel and Terceira) standat the East of the ridge, two (Corvo and Flores) arelocated to the west, on the American plate.

The oldest rocks are observed on Santa Maria(8 Myr) and Sao Miguel (4 Myr for the Easternpart) islands [22–25]. Other volcanics, including thecentral part of Sao Miguel, were erupted duringthe Pleistocene and Holocene. Historic lavas havebeen found in most areas of the archipelago (in-cluding Sao Miguel island), and the most recentvolcanism occurred on Faial in 1957=58 (Capelinhossite). The lava compositions range from basalts totrachy-andesites and some peridotitic nodules havebeen observed on Pico, Faial and Terceira islands[26,27].

Samples from the islands of Santa Maria, SaoMiguel, Terceira, Pico, Graciosa and Faial, wereanalyzed. Sample locations are given on Fig. 2 andin Appendix A.

M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205 191

F

Fl

C

P

G

T

SMi

SMa

Triple junction

Terceira riftM

.A.R

east Azores fracture zone

2000m

42 N

40 N

38 N

AZORES

32 W 28 W

?

o

o

o

o

o

Fig. 1. Location map of the Azores archipelago (modified after Turner et al. [34]). Triple junction has been proposed to be located at38.5ºN [41,57], rather than in the straight continuity of the Terceira rift. Abbreviations for the island names: Santa Maria (SMa), SaoMiguel (SMi), Terceira (T), Graciosa (G), Pico (P), Faial (F), Flores (Fl) and Corvo (C).

3. Analytical procedure

Helium analyses were performed on millimetersized olivine and pyroxene phenocrysts separatedfrom fresh lavas. Samples from the Eastern sideof Sao Miguel island were ankaramites with largeolivine and pyroxene crystals, up to 2 cm in size;most of the other lavas were alkali-basalts. Measure-ments were preferentially done by crushing, in orderto avoid possible in-situ radioactive decay or produc-tion of 3He cosmogenic and to obtain the inheritedhelium trapped in inclusions [3,28]. The powderof the crushed olivines and pyroxenes from sampleACO95-68 (ankaramite from Sao Miguel) were alsomelted in vacuo to release the gases located in thematrix, which allows an evaluation of the contribu-tion from radiogenic and cosmogenic helium.

Most of the helium measurements were per-formed in Paris using a glass mass spectrometerand extraction line. Due to diffusion, prior mea-surements with two purifications using hot titaniumgetters and trapping on a cold head in our KO-

VAR glass ARESIBO II mass spectrometer yieldedsignificant 4He blanks for very low helium con-centration samples such as phenocrysts (olivines,pyroxenes) from oceanic basalts [29,30]. Thus, wedeveloped a shorter crushing procedure which nowyields a 4He blank of 5 ð 10�10 š0.5 cm3 STP,a significant improvement to previous blank valuesof 1:5 š 0:5 ð 10�8 cm3 STP. A charcoal trap wasinstalled close to the crusher which retains all ac-tive gases, except He and Ne, during 30 minutes ofcryo-pumping at liquid nitrogen temperature. 3He isanalysed on an electron multiplier and 4He, depend-ing of the absolute amount of gas, on a Faraday cup,or on the electron multiplier. Using only 30 minutesof purification in a very small volume, the helium in-troduced into the extraction line by diffusion is verysmall, leading to low blanks. This procedure makespossible the measurement on this mass spectrometerof very low level samples using 1 or 2 g of olivine orpyroxene.

This new procedure was tested by repeated analy-sis of a oceanite sample from Reunion island (Indian

192 M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205

T2

T7

T9

Terceira Island

T28

Pico

ACO95-33

S14

S9

S6

ACO95-30

ACO95-20

Pico Island

ACO95-14

ACO95-16

ACO95-9ACO95-10

ACO95-11

ACO95-12

Faial Island

São Miguel Island

Santa Maria Island

ACO95-3ACO95-56

ACO95-66 ACO95-62

ACO95-47ACO95-49

ACO95-53

ACO95-s

ACO95-52

FA7

FA6

FA12

FA3

FA2

FA1

ACO95-55

ACO95-68

SM7 SM319

SM139

FA10

SM169SM150

FA44

ACO95-26 -40

1718-1720

1563

1761

nordeste

Capelinhos

5km

5km

5km

5km

5km

Fig. 2. Faial, Pico, Santa Maria, Sao Miguel and Terceira islandsampling maps, modified after Forjaz et al. [67].

ocean), whose olivines are known to have a veryhomogeneous 4He=3He isotopic ratio of 56,000 š1000 (R=Ra D 12:9 š 0:2) [31,32]. Using differentamounts of olivines (0.5, 0.8 and 1.3 g, respectively),we obtained reproducible 4He=3He results of 55,296š 3254, 55,918 š 4152 and 55,591 š 2740.

Some helium analyses were also performed inthe Woods Hole Oceanographic Institution where theblank is typically 3 to 5 ð 10�11 cm3 STP [33];these data are italicized in Table 1. There is a goodagreement between the data obtained in the two lab-oratories. Sample ACO95-3, which was collected atthe 1563 lava flow from Queimado peak eruption,have given 4He=3He ratios of 149,600 and 133,830similar to the ratio of 133,800 of one sample fromthe same lava flow analyzed by M.D. Kurz (sam-ple #SM88-45, unpublished results) indicating goodinter-laboratory agreement.

Lead chemical separation was performed on ¾0.5g of powdered samples using the procedure of Man-hes et al. [34], in a clean room under controlled at-mosphere. This permits one to maintain the total Pbblank below 0.35 ng. Lead isotopic ratios were mea-sured on a Thomson THN 206 mass spectrometerand calibrated against SRM981 NBS [35] standardmeasurements, with a statistical mass discriminationfactor of 1 š 0:3‰ per amu. The total precision,tested by duplicate analysis, is better than 0.05%(2¦ ) per mass unit difference in the isotopic ratio.

4. Results

Results are given in Tables 1 and 2. Heliumconcentrations obtained by crushing range from

Footnote to Table 1: R=Ra is the 3He=4He ratio normalized tothe AIR value of 1:384ð 10�6. ¦ is the uncertainty. Uncertaintyin the helium concentrations is approximately 5%. The last col-umn gives the percentage of 4He blank contribution (which was4He ³ 5ð10�10 cm3 STP in Paris and 3 to 5ð10�11 at WHOI).Data in italics were analyzed in the Woods Hole OceanographicInstitution isotope facilities. Measurements were conducted onolivine (Ol.) and clinopyroxene (CPX). Re-crush means thatsamples were crushed a second time, sequentially, in an attemptto insure complete extraction, and also to resolve any contri-butions from radiogenic helium. The experiments conducted bymelting of powder remaining after crushing are abbreviated mp.

M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205 193

Table 1Helium concentration (in 10�8 cm3 STP=g) and isotopic ratios (corrected for blank) for Azores samples

Sample Mineral Weight 4He R=Ra4He=3He Blank

(g) (ð10�8) (%)

FaialACO95-sand Ol. 1.04 4.9 7.5 š 0.5 95960 š 6630 1

Ol. 1.33 6.8 8.0 š 0.2 90890 š 1930 14ACO95-47 Ol. 1.79 0.53 8.3 š 1.2 87160 š 12090 4ACO95-49 Ol. 1.58 43 8.7 š 0.1 86320 š 1210 3ACO95-52 Ol. 1.08 4.5 7.2 š 0.4 99940 š 5940 1ACO95-53 Ol. 0.38 1.2 8.4 š 0.1 85970 š 700 1ACO95-55 Ol. 1.01 0.54 5.6 š 1.7 129720 š 40520 11FA10 Ol. 1.13 2.0 7.8 š 0.5 92280 š 5780 3FA12 Ol. 0.19 1.2 8.0 š 0.1 89601 š 2378 2

PicoACO95-20 Ol. 0.35 1.7 10.0 š 0.1 72545 š 415 2ACO95-20 CPX 1.84 3.6 10.0 š 0.2 72110 š 1730 1ACO95-26 Ol. 1.32 1.4 10.0 š 0.3 72550 š 2480 2ACO95-30 Ol. 0.96 36 8.8 š 0.1 81820 š 870 0.2ACO95-30 b Ol. 1.01 13 10.3 š 0.1 70150 š 890 1ACO95-30 b CPX 1.08 7.6 9.7 š .0.1 74320 š 1070 1ACO95-33 Ol. 1.17 3.2 8.9 š 0.5 80910 š 4620 1ACO95-40 Ol. 0.97 1.0 7.5 š 0.9 95960 š 10830 3

Santa MariaS6 Ol. 0.17 0.19 7.7 š 0.3 94376 š 3242 17

Sao MiguelACO95-3 Ol. 1.19 1.4 4.8 š 0.6 149600 š 17650 2

Ol. 1.46 1.8 5.4 š 0.5 133830 š 13490 3ACO95-56 Ol. 1.02 0.50 4.5 š 1.2 162070 š 43520 29ACO95-62 Ol. 1.97 0.23 6.1 š 1.9 118450 š 36900 12ACO95-62 CPX 1.07 0.60 3.0 š 1.1 240850 š 88310 9ACO95-66 Ol. 1.42 0.33 5.9 š 2.0 121640 š 41160 16ACO95-68 Ol. 0.33 0.20 4.1 š 0.1 174317 š 5047 7

Re-crush 0.06 3.8 š 0.4 189644 š 20408 25mp 0.31 0.37 4.4 š 0.1 163180 š 4310 4

ACO95-68 CPX 1.61 3.3 2.6 š 0.1 276840 š 11670 1CPX 2.82 3.4 4.0 š 0.3 180550 š 11910 1

ACO95-68 (melt) CPX 3.02 3.3 4.4 š 0.7 163470 š 24600ACO95-68 CPX 0.32 2.4 4.06 š 0.03 178200 š 1360 1

Re-crush 0.02 4.3 š 0.8 170000 š 30520 82mp 0.18 0.48 1.6 š 0.1 461980 š 18610 6

TerceiraT2 Ol. 1.08 0.41 10.4 š 1.6 69530 š 10570 11

Ol. 1.88 0.59 11.3 š 0.8 63780 š 4620 8T7 Ol. 1.02 1.5 11.0 š 0.4 65510 š 2140 3T9 Ol. 0.34 1.4 9.3 š 0.1 77550 š 430 1T28 Ol. 1.74 0.23 7.0 š 1.8 103070 š 26910 9T28 Ol. 0.30 0.13 8.5 š 0.2 85165 š 2350 13ACO95-9 Ol. 0.10 1.6 9.5 š 0.1 75858 š 709 3ACO95-10 Ol. 0.20 0.93 9.4 š 0.1 76965 š 680 3ACO95-11 Ol. 1.60 2.2 9.4 š 0.5 77200 š 3710 2ACO95-12 š Ol. 0.27 0.14 9.7 š 0.3 74604 š 2326 13ACO95-14 Ol. 0.99 1.2 8.0 š 0.8 90420 š 8520 5ACO95-16 Ol. 0.14 0.50 10.3 š 0.2 70437 š 1415 7

GraciosaG1 Ol. 0.19 0.26 8.1 š 0.2 89600 š 2380 10

194 M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205

Table 2Lead isotopic compositions determined for some Azores samples

Sample 206Pb=204Pb 207Pb=204Pb 208Pb=204Pb

FaialFA1 19.332 15.555 38.796

19.318 15.541 38.746FA2 19.607 15.574 38.943FA3 19.137 15.601 38.814

19.130 15.605 38.816FA6 19.471 15.619 39.123FA7 19.607 15.642 39.314FA10 19.707 15.635 39.296FA12 a 19.20 15.60 38.89FA44 a 19.60 15.59 39.10

GraciosaG1 19.80 15.60 39.18

PicoP1 a 19.96 15.61 39.37

Santa MariaS6 a 19.14 15.56 39.07S9 a 19.01 15.56 39.01S14 a 19.28 15.58 39.30

Sao MiguelWestSM7 a 19.47 15.59 39.32SM319 a 19.48 15.62 39.37NordesteSM139 a 20.00 15.78 40.33SM150 a 19.96 15.75 40.19SM169 a 19.96 15.74 40.31

a Previously published by Dupre (PhD thesis).

1:7 ð 10�9 cm3 STP=g for olivines from sampleS6 (Santa Maria) to 4:3 ð 10�7 cm3 STP=g forolivines analysed in a xenolith from Faial island(ACO95-49). Helium isotopic ratio vary between63,780 (R=Ra D 11:3) and 276,840 (R=Ra D 2:6) inT2 (Terceira) and ACO95-68 (Sao Miguel) samplesrespectively.

Helium concentrations obtained for melting aftercrushing of sample ACO95-68 were 3:7 ð 10�9 cm3

STP=g for the olivines, with a 4He=3He ratio of163,180 š 4310 (R=Ra D 4:4š0:1) which is similarto the crushed ratio (4:1 š 0:1), and 4:8 ð 10�9

cm3 STP=g for the pyroxenes with a helium ratioof 461,980 š 18,610 (R=Ra D 1:6 š 0:1). Thisratio is much more radiogenic than the crushedratio (4:06š0:03). The agreement between 4He=3Heratios obtained by crushing in this sample, despite

the presence of radiogenic 4He in the clinopyroxenestrongly suggest, that the crushing data reflects themagmatic helium. This point is illustrated in Fig. 3c,where good agreement in 4He=3He ratios is obtainedfor several different extractions and over a widevariation in 4He concentrations.

Three distinctive island groups have been dis-tinguished on the basis of helium isotopic results(Fig. 3):

(1) Sao Miguel island, for which 4 samples havebeen analysed, has the most radiogenic 4He=3He ra-tios. One sample (ACO95-3) was from a historicallava flow (Queimado peak eruption, 1563) whereasthe 3 others are Pliocene ankaramites [23]. Theyall display very radiogenic 4He=3He ratios, from121,640 š 41,160 to 276,840 š 11,670, with aweighted mean value of 140,660 for the histori-cal sample ACO95-3 (crushed samples). These dataagree with previous measurements from Sao Miguelby Kurz (fig. 8.10 of [36]).

(2) Faial, Graciosa and Santa Maria islands all ex-hibit a MORB-like helium signature, with 4He=3Heratios ranging from 85,970 š 1210 to 94,380 š3240 (Faial ACO95-55 sample displays an higherbut uncertain value of 129,720 š 40,520).

(3) Pico and Terceira islands have a more dis-persed helium signature and range from typicalMORB ratios to relatively primitive values, suchas 63,780 š 4620 for Terceira T2 basalt and 70,150š 890 for Pico ACO95-30 sample.

Lead isotopic results (Table 2) for the Azoresislands are all relatively radiogenic in compositioncompared to local MORB. However, the respectivepositions of the five islands in Pb–Pb space suggestsa similar grouping of Azores islands (Fig. 4), withgood helium and lead covariations at the scale of thearchipelago:

(1) Eastern Sao Miguel samples have relativelyhigh 206Pb=204Pb ratios, around 20, and very ra-diogenic 207Pb=204Pb and 208Pb=204Pb values, upto 15.78 and 40.33 respectively, which are clearlydifferent from those measured on other islands.

(2) Faial and Santa Maria basalts represent theleast radiogenic compositions of the archipelago,with a mean value of 19.50 for 206Pb=204Pb, 15.60for 207Pb=204Pb and 39.00 for 208Pb=204Pb.

(3) Pico and Terceira samples exhibit lead isotoperatios ranging from Faial–Santa Maria values (19.50,

M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205 195

10-9 10-8 10-7

10-9 10-8 10-7

4He (ccSTP/g)

Terceira

Sao Miguel

Pico

Kurz (unpublish)

rc

rc

rc

mp

c

c

c

mp

mp

100000

150000

100000

150000

200000

300000

400000

4 He/

3 He

4 He/

3 He

4 He/

3 He

ACO95-68CPX

Olivine

MORB

MORB

radiogenic

crush

re-crush

melt powder

inherited helium

Graciosa

Santa Maria

Faial

R/R

aR

/Ra

R/R

a

7.2

4.8

7.2

4.8

14.5

14.5

3.6

2.4

1.8

Fig. 3. 4He=3He ratios vs. 4He concentrations (in cm3 STP=g) for Azores samples. Grey lines correspond to the MORB 4He=3Hemean value of ¾90,000 [1,68]. Top figure represents the Pico, Terceira and Sao Miguel islands. Some Sao Miguel data from Kurz(unpublished) were added. Middle figure shows Faial, Santa Maria and Graciosa basalts. One can see the isotopic ratios are exactlythe same as the mean MORB ratio. Bottom figure represents two experiments on the ankaramite ACO95-68 from the Nordeste area onSao Miguel. Step crushing and powder melting extractions were performed on olivines and clinopyroxene to constrain the post eruptionradioactive production of 4He. The results show that the 4He=3He ratios are always the same during crushing (olivine and CPX), butdifferent for CPX melting due to the higher U content of CPX. This implies the 4He=3He ratios obtained by crushing are representativeof the magmatic values, and have not been altered by post-eruptive radiogenic helium (see text).

15.60 and 39.00) to radiogenic compositions simi-lar to Sao Miguel basalts for 206Pb=204Pb (20.02),lower for 207Pb=204Pb (15.64) and 208Pb=204Pb(39.35) [37,38]. Moreover, similarly to helium,

Pico–Terceira and Faial–Santa Maria associatedfields are not totally distinctive.

These three groups were also proposed by Turneret al. [38] based on trace element-isotope systematics

196 M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205

TT

SM

T

SM

SMSM

15.5

15.6

15.7

15.8

17.5 18.5 19.5 20.5 21.5

17.5 18.5 19.5 20.5 21.5

207 P

b/20

4 Pb

208 P

b/20

4 Pb

206Pb/204Pb

38.0

38.5

39.5

40.5

37.5

39.0

40.0

M

M

Tu

Tu

BE

MORB36-40˚N

MORB36-40 N

MORB36-40˚N

MORB36-40 N

FaialPicoSanta MariaSao MiguelPrevious studies

Terceira

Graciosa

FaialPicoSanta MariaSao MiguelPrevious studies

Terceira

M

M

Tu

Tu

T

BE

NHRL

NHRL

NHRL

15.5

15.6

15.7

15.8

17.5 18.5 19.5 20.5 21.5

17.5 18.5 19.5 20.5 21.5

206Pb/204Pb

38.0

38.5

39.5

40.5

37.5

39.0

40.0

BE

ENA

ENA

BE

a

b d

c

o

o

Fig. 4. Lead isotopic diagrams, showing end-member positions of Sao Miguel and Terceira [37] islands. (a) 206Pb=204Pb vs. 207Pb=204Pb.(b) 206Pb=204Pb vs. 208Pb=204Pb. Also reported are previous Sao Miguel [38,48] results, the Northern Hemisphere Reference Line(NHRL, as defined in [69]), Bulk Earth average value (BE) [58], and representative fields for local MORB [70–73], Tubuaii and Mangaiaislands [74–76]. SM and T points are postulated to be representative of the Azores plume compositions. (c) 206Pb=204Pb vs. 207Pb=204Pb.(d) 206Pb=204Pb vs. 208Pb=204Pb. ENA fields (East North America tholeiitic samples [59]) can be interpreted as reflecting binary mixingbetween a end-member with low lead isotopic ratios and Sao Miguel end-member SM.

where for example, the Th=Nb is higher for SaoMiguel island [38].

Fig. 5 shows the 4He=3He ratio against the 206Pb=204Pb ratios for the different Azores archipelagoislands. The three groups can be easily distinguishedon this figure.

5. Discussion

5.1. The Eastern Sao Miguel radiogenic heliumsignature

Since Sao Miguel island is one of the oldest is-lands in the archipelago (up to 4 Myr [23,25]), high

M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205 197

18

19

20

21

206 P

b/20

4 Pb

50,000 100,000 150,000 200,000

14.5 7.2 4.8 3.6

4He/3He

R/Ra

Gough

Loihi

Heard 1

Heard 2

St Helena

MORB 35˚-40˚NReunion

Recycled oceanic crust

Faial - St Maria-Graciosa

Terceira

Faial

Santa Maria

Terceira

Graciosa

Sao Miguel (Kurz)

Sao Miguel

n-MORBs

Fig. 5. 4He=3He ratios versus 206Pb=204Pb ratios for Azores Archipelago basalts and some other islands. Some Sao Miguel data fromKurz (unpublished) were added. Other data are from the literature.

4He=3He ratios measured in some samples couldbe explained by post-eruptive decay of Th and U,particularly for samples with low helium and highU and Th contents. Sample crushing procedures areperformed to liberate the inherited gases contained influid and melt inclusions, but 4He production couldoccur in the matrix with subsequent diffusion into theinclusions, or partial release by crushing. However,both the low concentration Pliocene ankaramites (2.3to 6:0ð 10�9 cm3 STP=g for ACO95-62, 66 and 68)and historical lava ACO95-3 (434 yr), whose olivinescontain around 1:6 ð 10�8 cm3 STP=g 4He, display

radiogenic helium isotope ratios, which strongly sug-gests that the unique radiogenic signature of SaoMiguel cannot totally result from such a decayprocess. Only old samples could have been mod-ified, starting from a minimum ACO95-3 olivineslike radiogenic value (140,000, we will consider,in the following, as the Nordeste Sao Miguel he-lium signature) to reach ratios as high as 276,840(ACO95-68 CPX) as illustrated by the melting ex-periment (Fig. 3).

The melt of the ACO95-68-CPX powder hasgiven a 4He=3He ratio of 461,980 š 18,610, much

198 M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205

more radiogenic than the crushed value of 178,200.This can easily be explained by post-eruptive 4Heproduction since this sample has an age between 2and 4 Ma [23]. Using a U concentration of 0.012ppm and a Th=U ratio of 2.9 for CPX [39] we canobtain a U=He age of 1.2 Ma for sample ACO95-68.This age is in good agreement with the K=Ar age[23] considering the uncertainty on the U contentand the Th=U ratio. However, for olivines, crushingand melting have given the same helium isotopicratios (Fig. 3), indicating that 4He production from Uand Th decay is negligible for olivines. This showsthat the radiogenic helium isotopic ratios observed inthe samples from the Nordeste area do not representpost-eruptive decay of Th and U.

Zindler and Hart [9,10] and Hilton et al. [11] haveproposed that high 4He=3He ratios measured in someOIB could reflect shallow depth contamination athigh levels in the crust. The presence of ankaramitesamples with low pressure pyroxenes (high total Aland low AlVI) on Sao Miguel island is consistentwith the existence of shallow magma chambers be-neath the Azores archipelago [23]. Plume sourcecontamination is modeled by a two-component mix-ing and the expression of the mass ratio betweenplume (Mi) and contaminant (oceanic crust) material(Mc) is then given by:

Mc

Mi

8>>><>>>:� 4He

3He

�SM

�� 4He

3He

�i� 4He

3He

�c

�� 4He

3He

�SM

9>>>=>>>;ðð

3HeŁ

ið3He

Łc

where .4He=3He/SM, .4He=3He/i and .4He=3He/crefers to the Sao Miguel island, Sao Miguel plumeand contaminant representative helium values, [3He]i

and [3He]c to the plume and contaminant concentra-tions (cm3 STP=g).

We used results obtained by Staudacher and Al-legre [40] on 22 My old oceanic crust (which corre-sponds to the approximate age of the nearby Azorescrust [41]) as a contaminant material ([3He]c D5ð10�14 cm3 STP=g and .4He=3He/c D 2ð107), thevalue of 140,000 for Sao Miguel island composition,and Terceira-like 4He=3He ratio (60,000) for SaoMiguel ‘plume’. The 3He concentration in the plumematerial is thought to be intermediate between lower(2:5ð10�9 cm3 STP=g) [42] and upper mantle (1:1ð

10�10 cm3 STP=g) [43,44] values because the heliumratio we took is very different from the lower mantleratio of 20,000. Let us take 5ð 10�10 cm3 STP=g forthe 3He content for the Terceira plume. The Mc=Mi

result (4000%) suggests that important magmachamber outgassing, at least 99.9%, is required tobe consistent with a plume contamination scenario.

In the previous calculations, we do not considerthe possibility of radiogenic decay in a degassedmagma chamber. Using 3He concentration estimatedabove (5 ð 10�10 cm3 STP=g) in the following 4Heradioactive production equation (t in Ma, U in ppmand He in cm3 STP=g):� 4He

3He

½t

D� 4He

3He

½0

C

2:8ð 10�8

²4:35C Th

U

¦[U]

[3He]ð t

we can determine the time interval t necessary toincrease the initial helium ratio (4He=3He)0 from60,000 (Terceira less primitive value) to the present-day value .4He=3He/t of 140,000. Assuming a Th=Uratio of 2.75 [38] and a U concentration of 0.040ppm (we postulate Terceira plume uranium to beintermediate between lower mantle material [6] andrecycled altered oceanic crust [15] based on lead iso-tope systematics, cf. Section 5.2), we need to invokevery large outgassing (at least 99.999%) to obtainmagma chamber interval time consistent with classi-cal estimates of magma residence time (between 10and 10,000 yr [45,46]).

Similar calculations can be done with lead iso-topes. However, using a Terceira-like isotopic com-position (207Pb=204Pb D 15.64) for the Sao Miguelplume and the most radiogenic value (207Pb=204PbD 15.64) measured in old oceanic crust (14–37 My)near the Azores archipelago [47] as a contaminantdoes not allow to obtain 207Pb=204Pb ratios that areas high as those determined on Sao Miguel sam-ples (207Pb=204Pb D 15.78). Contamination by recentAtlantic sediments is also unlikely, given the high208Pb=204Pb ratios [48].

There is no evidence for such an extensive out-gassing in the Sao Miguel olivine samples, whichshow similar He content to olivines from other is-lands, in particular from Terceira. Kurz et al. [20]have observed a variation of the 4He=3He ratio with

M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205 199

the longitude on Sao Miguel island (with a decreasewestward), similar to the 87Sr=86Sr and 206Pb=204Pbvariations [20,48,49]. Because these variations andcorrelations are difficult to explain by shallow depthcontamination, we propose a deeper origin for thehigh 4He=3He ratios measured in Sao Miguel basalts.A source effect also appears to be consistent with thehelium–lead covariations observed at the scale of thearchipelago. Kurz et al. [3], Vance et al. [4], Grahamet al. [5], Hanan and Graham [16] and Hanyu andKaneoka [7] have postulated that the radiogenichelium measured in Gough, Tristan, St Helena andsome Pacific islands can be explained by recyclinginto the mantle of oceanic crust or sediments, andreinjection of this material in plume sources. Sub-ducted plates should have high (U C Th)=3He ratiosbecause helium is outgassed to the atmosphere dur-ing ridge magmatism and during subduction. Afterstorage for some time in the mantle, this will givehigh 4He=3He ratios by U and Th α-decay. Lead andstrontium isotope signatures are consistent with sucha scenario [38,49], especially the high 207Pb=204Pb ra-tios which imply old U=Pb fractionation. Recyclingof subducted altered oceanic crust and sedimentsallows to increase the U=Pb ratio [15,50], and thengive very radiogenic 207Pb=204Pb ratios with time.

5.2. The Terceira helium signature

Kurz et al. [19] first proposed the presenceof a “high 4He=3He” plume beneath the Azoresarchipelago, based on helium isotopes measurementsin MORB glasses near the Azores platform. How-ever, it turns out that the Azores hotspot is alsocharacterized by ‘low 4He=3He’, since primitive val-ues (4He=3He ³ 65,000) have been measured in thisstudy in some Terceira and Pico samples. Produc-tion of cosmogenic helium [28] in Terceira or Picosamples remains very unlikely, as all the analysedbasalts are relatively young (some are historical,see Table 1 and Appendix A) and were sampledat low altitude, generally in freshly exposed sur-faces. Moreover, we have analyzed these samplesby crushing which primarily releases the inheritedhelium trapped in inclusions; only a small fractionof the cosmogenic helium is released by crushingand should be negligible for these samples. There-fore, the primitive helium has to originate within

the source of the plume. It may come from thelower mantle, source of primitive helium for most ofoceanic islands (Hawaii, Iceland, Galapagos or Reu-nion islands). The [3He]LM=[3He]UM concentrationratio between the lower and upper mantle shouldbe roughly 10 [6,51] so even a small amount ofentrained lower mantle could transfer a primitivehelium isotopic signal. A helium signature similarto Terceira was found on nearby Atlantic MORBglasses (Fig. 6, [52]) thereby confirming the plume–ridge interaction already observed for REE, Sr, Ndand Th=U systematics [53–56]. The somewhat dif-ferent helium observations by Kurz et al. [19], couldsimply be due to the different sampling scale ofthis study. For example, most of the MORB samplesfrom 39ºN are very degassed, and may have beendredged off the ridge axis since the ridge is not welldefined in this area. In fact, the helium signaturerecorded by MORB glasses displays very localizedprimitive values (around 38.5ºN; Moreira and Alle-gre [52]) which seems to correspond to the locationof the triple junction [41,57] (Fig. 6).

Most primitive helium ratios measured on Ter-ceira samples could be explained by mixing be-tween primitive mantle having high 3He (4He=3HeD 20,000) and two other discrete mantle sources:MORB mantle (4He=3He D 90,000) and a low 3Heplume best represented by data from Sao Miguel(4He=3He D 140,000). Helium alone cannot resolvebetween the two hypotheses. However, lead isotopicratios (Fig. 4a,b) show that the Terceira signaturecannot result from binary mixing between the SaoMiguel plume and lower mantle material. The pre-sent-day composition of the lower mantle certainlycorresponds to a binary mixing between Bulk Earthmaterial and about 17% of depleted mantle [58],and then is few different from Bulk Earth composi-tion. The Terceira signature probably reflects mixingof ‘high ¼’ material, lower mantle material andentrained depleted asthenospheric material, whereasthe observed trend for Terceira island may representshallow interaction between this specific materialand the local MORB mantle source.

5.3. Terceira–Sao Miguel: same hotspot?

Since we excluded a shallow contamination originfor explaining the He data in Sao Miguel island, the

200 M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205

60,000

80,000

100 000

120,000

36 38 40 42 44 46

LATITUDE ( N)

4 He/

3 He

N MORB

o

Fig. 6. 4He=3He ratios measured in MORB as a function of the latitude along mid Atlantic ridge between 35 and 40ºN, modified afterKurz et al. [19] (crosses) and Moreira and Allegre [52] (diamonds). Note that the minimum of the 4He=3He ratio is close to 38.5ºN,which is the supposed triple junction between African, American and European plates [41,57].

Azores archipelago helium signature, with coexis-tence of both radiogenic and primitive ratios, seemsto be paradoxical. Identification of recycled terrige-nous sediments in the source of Sao Miguel basaltsdoes not give information about the depth of thisspecific material and there are two models for thegeneration of the Azores islands.

The first one proposes the Azores plume deepsource may be isotopically heterogeneous, contain-ing a mixture of recycled subducted terrigenoussediments, altered oceanic crust and lower mantlematerial. A low degree of melting has permittedto exhibit the sedimentary signature (radiogenic) ofthe source (the fertile part of the mixture, i.e. sed-iments melt preferentially) in Sao Miguel basalts(Nordeste), whereas other islands result from higherextent of melting and reflect mixing of recycledoceanic crust, lower mantle and upper mantle mate-rials. The origin of this heterogeneous source mayresult of the entrainment by a plume coming fromthe lower mantle of material stored at the 670 kmdiscontinuity or higher or reflects the heterogeneityof the source itself.

The second model proposes the Azores plumedeep source is isotopically homogeneous, with a Ter-ceira like signature (mixing of lower mantle mate-rial and ‘himu’ material). Sao Miguel island basaltssample a very localized upper mantle heterogene-

ity, melted by the rising of the Terceira plume orblob. This heterogeneity could correspond to sub-ducted oceanic crust and sediments stored underthe north america continent during the pre-Grenvillesubduction [59], recycled into the upper mantle bydelamination occurring during the opening of theNorth Atlantic ocean (Fig. 4c,d and Fig. 7). Again,intermediate helium and lead compositions reflect in-teraction with the local MORB mantle source (Faial,Graciosa, Pico : : : ). Such a model can be validatedby the fact that East North American basalts (ENA),that correspond to the nearest subcontinental mate-rial from the Azores, show linear trend in Pb–Pbdiagrams (Fig. 4b) which could reflect mixing be-tween a source with low 206Pb=204Pb 207Pb=204Pband 208Pb=204Pb ratios, which could be subconti-nental mantle [60,61], and a component with SaoMiguel-like signature which could be a mixing ofsediments and recycled oceanic crust. We proposethis component was stored below the North Amer-ica before injection in the North Atlantic mantleby delamination during the opening of the NorthAtlantic. This model is similar to the proposal ofWidom et al. [48], involving shallow interaction ofthe plume source with delaminated subcontinentallithosphere enriched by metasomatism. Radiogenichelium ratios measured in some continental xeno-liths or continental basalts that sample the subcon-

M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205 201

ENA High AtlasMAR

MARAmerica

America

Africa

Africa

SMiT

>1Ga

1Ga

200 Ma

200 -10 Ma

10 Ma

Pre-Grenville subduction

Grenville continent - continent collision

ENA / High Atlas

delamination

Rifting

Sea floor spreading

a

b

c

d

e

"Terceira" plume or blob

Fig. 7. Schematic evolution of the North Atlantic to explain the Azores isotopic data (model 2). (a) Pre-Grenville subduction [59] bringssome oceanic crust and sediments into the subcontinental lithospheric mantle. (b) End of the subduction with the Grenville collision. (c)and (d) Opening of the Atlantic ocean is accompanied by some delamination, which provides regional enrichment of the Azores mantle.(e) Sao Miguel volcanism samples some localized km-size heterogeneity corresponding to old recycled material.

202 M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205

tinental lithosphere can account for such a model[62,63].

In both model 1 and model 2, we postulate a re-gional enrichment of the asthenosphere beneath theAzores. The observed radiogenic helium signature isnot restricted to Sao Miguel island. Kurz et al. [19]have measured 4He=3He ratios higher than 100,000in North Atlantic MORB glasses (for latitudes com-prised between 40ºN and 50ºN; Fig. 6), which areconsistent with contamination of the local MORBmantle source by delaminated subcontinental litho-sphere (occurring during the opening of the Atlanticocean). Such a hypothesis has been already proposedto explain specific Pb and Sr radiogenic signaturesof the Oceanographer transform zone basalts [64] at35ºN, and could also be proposed to explain simi-lar isotopic anomalies identified on the Mid AtlanticRidge at 14ºN [43,65], 43ºN and 46ºN [56]. Recentdiscovery of very old zircons in the Mid AtlanticRidge are also consistent with such a model [66].

6. Conclusions

The Azores archipelago corresponds to the sec-ond hotspot where three distinctive helium isotopicsignatures have been observed: MORB values foundat Santa Maria, Graciosa and Faial islands, primitivevalues for some Terceira and Pico basalts, and radio-genic values in Eastern Sao Miguel. The only otheroceanic island where 4He=3He ratios both higher andlower than MORB are found is Heard Island [11].The helium data are correlated with lead isotopiccompositions observed on the same islands and areinterpreted in terms of source variations. The exis-tence of such a multiple isotopic signature can beexplained by two scenarios. The first model pro-poses that the heterogeneous Azores plume sourcehas evolved from Sao Miguel to Terceira composi-tion, by various degrees of melting of heterogeneousmantle. The second hypothesis is that the Azoresplume is homogeneous with a Terceira like com-position. The origin of the Sao Miguel signaturewould then correspond to sampling of very localkm-size heterogeneities, resulting from the delami-nation of enriched subcontinental lithosphere, whichoccurred during the Jurassic opening of the NorthAtlantic.

Acknowledgements

M.M. would like to thank Pascale Louvat verymuch for his help during the mission in Azores andpatience during rock sampling, and the team of theEarth Sciences department from Universidade dosAcores, in particular Pr. Victor Forjaz and J.C. Nunesfor discussions and help during the Azores trip. B.Bourdon, J. Kunz and Th. Staudacher are thankedfor discussion and improvement of the manuscript.D. Graham, I. Kaneoka and R. Poreda helped to im-prove the quality of the manuscript by their reviewcomments. Josh Curtis is thanked for his help duringanalyses in the Woods Hole Oceanographic Insti-tution. This is IPGP contribution 1574 and WHOIcontribution 9893. [CL]

Appendix A. Sample locations

Sample locations are also given on Fig. 2. Some location arein [37].

FaialACO95-sand Olivine rich sand from the Capelinho place. Re-

sults from projection of submarine eruption duringthe 1957 eruption.

ACO95-47 Collected on the chimney in the rim that separatesthe historical eruption of the Capelinho (1957=58)from the Complexe do Capelo (<800 yr).

ACO95-49 Xenolith taken on the beach of the Baia da Ribeiradas Cabras. Unknown age.

ACO95-52 Sand with olivines and augites taken in the beachof Varadouro.

ACO95-53 Collected on the road leading to the caldera nearthe CabeodosTrinta Pic.

ACO95-55 Porto de Feiteira.

PicoACO95-20 Olivine and pyroxene rich basalts from the

1718-20 flow (Fase de Sta Luzia), collected nearthe airport.

ACO95-26 Basalt collected on the road between Sao Roquedo Pico and Lajes. Bottom of the Pic de Felipe(complexe da Madalena).

ACO95-30 Xenolith collected in the Misterio de Silveira(phase de Sta Luzia, 1718-20).

ACO95-33 Collected near Caelano (Ponta da Faca) (complexeda Madalena).

ACO95-40 Same location as 26.

Sao MiguelACO95-3 1563 Queimado Pic lava flow.ACO95-56 Collected on the road between Ribeira Grande and

Ponta Delgada near the Dr Ferreira Pic.

M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205 203

NordesteACO95-62 Ankaramite collected near the Guihermo ou dos

Moinhos river. South of Lomba da Fazenda.ACO95-66 Ankaramite collected near the Mulher river. South

of Santana.ACO95-68 Ankaramite collected near Faial da Terra.

TerceiraACO95-9 1761 lava flow at the bottom of the Pico do Fogo.ACO95-10 Collected near the road between Biscoitos and

Angra Do Heroismo (near the Pico do Fogo).ACO95-11 Collected in the Guiherme Moniz Caldeira (north).ACO95-12 Collected near Faieis in a river (Rib. das Pedras).ACO95-14 Collected near the road between Sao Sebastiao and

Feteira (near Boavista).ACO95-16 Collected near the road near Rossio de Sant Ana.

References

[1] C.J. Allegre, M. Moreira, T. Staudacher, 4He=3He disper-sion and mantle convection, Geophys. Res. Lett. 22 (17)(1995) 2325–2328.

[2] M.D. Kurz, W.J. Jenkins, S.R. Hart, D. Clague, Heliumisotopic variations in volcanic rocks from Loihi Seamountand the island of Hawaii, Earth Planet. Sci. Lett. 66 (1983)388–406.

[3] M.D. Kurz, W.J. Jenkins, S.R. Hart, Helium isotopic sys-tematics of oceanic islands and mantle heterogeneity, Na-ture 297 (1982) 43–47.

[4] D. Vance, J.O.H. Stone, R.K. O’Nions, He, Sr and Ndisotopes in xenoliths from Hawaii and other oceanic islands,Earth Planet. Sci. Lett. 96 (1989) 147–160.

[5] D.W. Graham, S.E. Humphris, W.J. Jenkins, M.D. Kurz,Helium isotope geochemistry of some volcanic rocks fromSaint Helena, Earth Planet. Sci. Lett. 110 (1993) 121–131.

[6] C.J. Allegre, T. Staudacher, P. Sarda, Rare gas systematics:formation of the atmosphere, evolution and structure of theearth’s mantle, Earth Planet. Sci. Lett. 81 (1986) 127–150.

[7] T. Hanyu, I. Kaneoka, The uniform and low 3He=4He ratiosof HIMU basalts as evidence for their origin as recycledmaterials, Nature 390 (1997) 273–276.

[8] M. Condomines, K. Gronvold, P.J. Hooker, K. Muehlen-bachs, R.K. O’Nions, N. Oskarsson, E.R. Oxburgh, Helium,oxygen, strontium and neodymium isotopic relationships inIcelandic volcanic, Earth Planet. Sci. Lett. 66 (1983) 125–136.

[9] A. Zindler, S. Hart, Helium: problematic primordial signals,Earth Planet. Sci. Lett. 79 (1986) 1–8.

[10] A. Zindler, S.R. Hart, Chemical geodynamics, Annu. Rev.Earth Planet. Sci. 14 (1986) 493–571.

[11] D.R. Hilton, J. Barling, G.E. Wheller, Effect of shallow-level contamination on the helium isotope systematics ofocean-island lavas, Nature 373 (1995) 330–333.

[12] C.J. Allegre, Isotope geodynamics, Earth Planet. Sci. Lett.86 (1987) 175–203.

[13] C.J. Allegre, E. Lewin, Isotopic systems and stirring timesof the Earth’s mantle, Earth Planet. Sci. Lett. 136 (1995)629–646.

[14] B.L. Weaver, The origin of ocean island basalt end-membercompositions: trace element and isotopic constraints, EarthPlanet. Sci. Lett. 104 (1991) 381–397.

[15] C. Chauvel, A.W. Hofmann, P. Vidal, HIMU-EM: theFrench Polynesian connection, Earth Planet. Sci. Lett. 110(1992) 99–119.

[16] B.B. Hanan, D.W. Graham, Lead and Helium evidencefrom oceanic basalts for a common deep source of mantleplumes, Science 272 (1996) 991–995.

[17] S.R. Hart, E.H. Hauri, L.A. Oschmann, J.A. Whitehead,Mantle plumes and entrainment: isotopic evidence, Science256 (1992) 517–520.

[18] C. Class, S.L. Goldstein, Plume–lithosphere interactions inthe ocean basins: constraints from the source mineralogy,Earth Planet. Sci. Lett. 150 (1997) 245–260.

[19] M.D. Kurz, W.J. Jenkins, J.-G. Schilling, S.R. Hart, Heliumisotopic variation in the mantle beneath the central NorthAtlantic Ocean, Earth Planet. Sci. Lett. 58 (1982) 1–14.

[20] M.D. Kurz, A. Gulessarian, R.B. Moore, Isotopic variationswithin oceanic islands: He, Sr and Pb isotopes in basaltsfrom Sao Miguel, Azores, Eos 71 (1990) 657.

[21] R.H. Kingsley, J.G. Schilling, Carbon in Mid-AtlanticRidge basalt glasses from 28º to 63ºN: evidence for acarbon-eriched Azores mantle plume, Earth Planet. Sci.Lett. 142 (1996) 175–189.

[22] A.A. Abdel-Monem, L.A. Fernandez, G.M. Boone, K–Arages from the eastern Azores group (Santa Maria, SaoMiguel and the Formigas Islands), Lithos 8 (1975) 247–254.

[23] L.A. Fernandez, Geology and petrology of the Nordestevolcanic complex, Sao Miguel, Azores: summary, Geol.Soc. Am. Bull. 91 (1980) 675–680.

[24] G. Feraud, I. Kaneoka, C.J. Allegre, K=Ar ages and stresspattern in the Azores: geodynamic implications, EarthPlanet. Sci. Lett. 46 (1980) 275–286.

[25] R. Moore, Volcanic geology and eruption frequency, SaoMiguel, Azores, Bull. Volcanol. 52 (1990) 602–614.

[26] R.C. Mitchell-Thome, Geology of the Middle Atlantic Is-lands, Gebruder Borntraeger, Berlin, Stuttgart, 1976, 382pp.

[27] Z. Franca, M.H. Almeida, N. Wallenstein, Ocorrencia dexenolotos mantelicos numa lava da ilha do Pico (Acores),GAIA- Revista de Geociencias, Univ. Lisboa, 1995.

[28] M.D. Kurz, Cosmogenic helium in a terrestrial igneousrock, Nature 320 (1986) 435–439.

[29] T. Staudacher, C.J. Allegre, Terrestrial xenology, EarthPlanet. Sci. Lett. 60 (1982) 389–406.

[30] M. Moreira, T. Staudacher, P. Sarda, J.-G. Schilling, C.J.Allegre, A primitive plume neon component in MORB:The Shona ridge-anomaly, South Atlantic (51–52ºS), EarthPlanet. Sci. Lett. 133 (1995) 367–377.

[31] T. Staudacher, P. Sarda, C.J. Allegre, Noble gas systematicsof Reunion Island, Indian Ocean, Chem. Geol. 89 (1990)1–17.

204 M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205

[32] D. Graham, J.G. Lupton, F. Albarede, M. Condomines, Ex-treme temporale homogeneity of Helium isotopes at Pitonde la Fournaise, Reunion Island, Nature 347 (1990) 545–548.

[33] M.D. Kurz, T.C. Kenn, J.C. Lassiter, D.J. DePaolo, Heliumisotopic evolution of Mauna Kea volcano: first results fromthe 1 km drill core, J. Geophys. Res. 101 (1996) 11781–11791.

[34] G. Manhes, J.-F. Minster, C.J. Allegre, Comparative U–Th–Pb and Rb–Sr study of the St Severin amphoterites: conse-quence for early solar system chronology, Earth Planet. Sci.Lett. 39 (1978) 14–24.

[35] E.J. Cantanzaro, T.J. Murphy, W.R. Shields, E.L. Garner,Absolute isotopic abundance ratios of common, equal-atom,and radiogenic lead isotopic standards, J. Res. Natl. Bur.Stand. 72A (1968) 261–267.

[36] M.D. Kurz, Noble gas isotopes in oceanic basalts: contro-versial constraints on mantle models, Miner. Assoc. Can.Short Course 19 (1991) 259–286.

[37] B. Dupre, B. Lambret, C.J. Allegre, Isotopic variationswithin a single oceanic island: the Terceira case, Nature299 (1982) 620–622.

[38] S. Turner, C. Hawkesworth, N. Rogers, P. King, U–Thisotope disequilibria and ocean island basalt generation inthe Azores, Chem. Geol. 139 (1997) 145–164.

[39] E.H. Hauri, T.P. Wagner, T.L. Grove, S.F. Foley, S.R. vander Laan, Experimental and natural partitioning of Th, U,Pb and other trace elements between garnet, clinopyroxeneand basaltic melts, Chem. Geol. 117 (1234) (1994) 149–166.

[40] T. Staudacher, C.J. Allegre, Recycling of oceanic crust andsediments: the noble gas subduction barrier, Earth Planet.Sci. Lett. 89 (1988) 173–183.

[41] R. Searle, Tectonic pattern of the Azores spreading centreand triple junction, Earth Planet. Sci. Lett. 51 (1980) 415–434.

[42] C.J. Allegre, T. Staudacher, P. Sarda, Rare gas systematics,formation of the atmosphere, evolution and structure of theEarth’s mantle, Earth Planet. Sci. Lett. 81 (1986) 127–150.

[43] T. Staudacher, P. Sarda, S.H. Richardson, C.J. Allegre, I.Sagna, L.V. Dmitriev, Noble gases in basalt glasses from aMid-Atlantic ridge topographic high at 14ºN: geodynamicconsequences, Earth Planet. Sci. Lett. 96 (1989) 119–133.

[44] M. Moreira, J. Kunz, C.J. Allegre, Rare gas systematicson popping rock: estimates of isotopic and elemental com-positions in the upper mantle, Science 279 (1998) 1178–1181.

[45] M. Condomines, C. Hemond, C.J. Allegre, U–Th–Raradioactive disequilibria and magmatic processes, EarthPlanet. Sci. Lett. 90 (1988) 243–262.

[46] D.M. Pyle, The volume and residence time of magmabeneath active volcanoes determined by decay-series dise-quilibria methods, Earth Planet. Sci. Lett. 112 (1992) 61–73.

[47] B. Dupre, C. Gopel, H. Bougault, Lead isotopic variationsin old ocean crust near the Azores, Init. Reports DSDP 82(1985) 497–500.

[48] E. Widom, R.W. Carlson, J.B. Gill, H.U. Schmincke, Th–Sr–Nd–Pb isotope and trace element evidence for the originof the Sao Miguel, Azores, enriched mantle source, Chem.Geol. 140 (1997) 49–68.

[49] C.J. Hawkesworth, M.J. Norry, J.C. Roddick, R. Vollmer,143Nd=144Nd and 87Sr=86Sr ratios from the Azores andtheir significance in LIL-element enriched mantle, Nature280 (1979) 28–31.

[50] A.W. Hofmann, W.M. White, Mantle plumes from ancientoceanic crust, Earth Planet. Sci. Lett. 57 (1982) 21–436.

[51] C.J. Allegre, T. Staudacher, P. Sarda, M.D. Kurz, Con-straints on evolution of Earth’s mantle from rare gas sys-tematics, Nature 303 (1983) 762–766.

[52] M. Moreira, C.J. Allegre, Rare gases in Mid Atlantic Ridgebasalts near the Azores: injection of atmospheric rare gasesin the mantle?, in prep.

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

[54] W.M. White, J.G. Schilling, S.R. Hart, Evidence for theAzores mantle plume from strontium isotopes geochemistryof the Central North Atlantic, Nature 263 (1976) 659–663.

[55] B. Bourdon, C. Langmuir, A. Zindler, Ridge–hotspot in-teraction along the Mid-Atlantic Ridge between 37º300 and40º300N: the U–Th disequilibrium evidence, Earth Planet.Sci. Lett. 142 (1996) 175–189.

[56] D. Yu, D. Fontignie, J.-G. Schilling, Mantle plume–ridgeinteraction in the Central North Atlantic: A Nd isotopestudy of Mid-Atlantic Ridge basalts from 30º to 50ºN, 146(1997) 256–272.

[57] J. Freire Luis, J.M. Miranda, A. Galdeano, P. Patriat, J.C.Rossignol, L.A. Mendes Victor, The Azores triple junctionevolution since 10 Ma from an aeromagnetic survey of theMid Atlantic Ridge, Earth Planet. Sci. Lett. 125 (1994)439–459.

[58] C.J. Allegre, E. Lewin, Chemical structure and history ofthe Earth: evidence from global non-linear inversion ofisotopic data in a three box model, Earth Planet. Sci. Lett.96 (1989) 61–88.

[59] W. Pegram, Development of continental lithospheric mantleas reflected in the chemistry of the Mesozoic AppalachianTholeites, USA, Earth Planet. Sci. Lett. 97 (1990) 316–331.

[60] C.J. Hawkesworth, P.D. Kempton, N.W. Rogers, R.M. El-lam, P.W. Van Calsteren, Continental mantle lithosphere,and shallow level enrichment processes in the Earth’s man-tle, Earth Planet. Sci. Lett. 96 (1990) 256–268.

[61] J. Mahoney, A.P. LeRoex, Z. Peng, R.L. Fisher, J.H. Nat-land, Southwestern Limits of Indian Ocean Ridge Mantleand the Origin of Low 206Pb=204Pb Mid-Ocean RidgeBasalt: Isotope Systematics of the Central Southwest IndianRidge (17º–50ºE), J. Geophys. Res. 97 (13) (1992) 19771–19790.

[62] T. Dunai, H. Baur, Helium, neon and argon systematicsof the European subcontinental mantle: implications forits geochemical evolution, Geochim. Cosmochim. Acta 59(1992) 2767–2784.

[63] M.R. Reid, D.W. Graham, Resolving lithospheric and sub-lithospheric contributions to helium isotope variations in

M. Moreira et al. / Earth and Planetary Science Letters 169 (1999) 189–205 205

basalts from the Southwestern US, Earth Planet. Sci. Lett.144 (12) (1996) 213–222.

[64] S.B. Shirey, J.F. Bender, C.H. Langmuir, Three-componentisotopic heterogeneity near the Oceanographer transform,Mid-Atlantic Ridge, Nature 325 (1987) 217–223.

[65] L. Dosso, H. Bougault, J.-L. Joron, Geochemical morphol-ogy of the North Mid-Atlantic Ridge, 10º–24ºN: Traceelement-isotope complementarity, Earth Planet. Sci. Lett.120 (1993) 443–462.

[66] J. Pilot, C.-D. Werner, F. Haubrich, N. Baumann, Paleo-zoic and Proterozoic zircons from the Mid Atlantic Ridge,Nature 393 (1998) 676–679.

[67] V. Forjaz, A. Serralheiro, J.C. Nunes, Carta Vulcanologicados Acores, Grupo Central, Universidade Dos Acores, 1sted., Ponta Delgada, 1990.

[68] T. Staudacher, M. Moreira, C.J. Allegre, Heterogeneity andscatter of helium isotopic ratios in MORB, Mineral. Mag.58A (1994) 874.

[69] S.R. Hart, A large-scale isotope anomaly in the SouthernHemisphere mantle, Nature 309 (1984) 753–757.

[70] B. Dupre, C.J. Allegre, Pb–Sr–Nd isotopic correlation and

the chemistry of the north atlantic mantle, Nature 286(1980) 17–22.

[71] B. Dupre, B. Lambret, D. Rousseau, C.J. Allegre, Limita-tions on the scale of mantle heterogeneities under oceanicridges, Nature 294 (5841) (1981) 552–554.

[72] B. Hamelin, B. Dupre, C.J. Allegre, Lead–strontium iso-topic variations along the East Pacific rise and the Mid-Atlantic Ridge: a comparative study, Earth Planet. Sci. Lett.67 (1984) 340–350.

[73] E. Ito, W. White, C. Gopel, The O, Sr, Nd and Pb isotopegeochemistry of MORB, Chem. Geol. 62 (1987) 157–176.

[74] P. Vidal, C. Chauvel, R. Brousse, Large mantle heterogene-ity beneath French Polynesia, Nature 307 (1984) 536–538.

[75] E.H. Hauri, S.R. Hart, Re–Os isotope systematics of HIMUand EMII oceanic island basalts from the south PacificOcean, Earth Planet. Sci. Lett. 114 (1993) 353–371.

[76] J.D. Woodhead, Extreme HIMU in an oceanic setting: thegeochemistry of Mangaia Island (Polynesia), and tempo-ral evolution of the Cook-Austral hotspot, J. Volcanol.Geotherm. Res. 72 (1996) 1–19.