d/h ratios in basalt glasses from the salas y gomez mantle plume interacting with the east pacific...

D/H ratios in basalt glasses from the Salas y Gomez mantleplume interacting with the East Pacific Rise: Water from oldD-rich recycled crust or primordial water from the lowermantle?

Richard H. Kingsley and Jean-Guy SchillingGraduate School of Oceanography, University of Rhode Island, Narragansett Bay Campus, Narragansett, Rhode Island02882-1197, USA ([email protected]; [email protected])

Jacqueline E. Dixon and Peter SwartRosentheil School of Marine and Atmospheric Sciences, University of Miami, 46000 Rickenbacker Causeway, Miami,Florida 33149, USA ([email protected]; [email protected])

Robert PoredaDepartment of Earth and Environmental Sciences, University of Rochester, 227 Hutchison Hall, Rochester, New York14627, USA ([email protected])

Kyla SimonsRosentheil School of Marine and Atmospheric Sciences, University of Miami, 46000 Rickenbacker Causeway, Miami,Florida 33149, USA

Now at Lamont-Doherty Earth Observatory of Columbia University, 61 Route 9W, Palisades, New York 10964, USA([email protected])

[1] Deuterium/Hydrogen isotope ratios were measured in 22 fresh basalt glasses dredged from young

seamounts along the Easter-Salas y Gomez seamount chain (ESC) and from the spreading centers of the

Easter Microplate (EMP). The dD values decrease regularly from �36% near Salas y Gomez to �63%,

800 km west on the East Pacific Rise (EPR) and the west rift of the EMP. Similar gradients are observed in

H2O, Pb and Sr radiogenic isotope ratios, trace element concentrations, and in the ratios of incompatible to

moderately incompatible trace elements. Fractionation of the D/H ratios by recent low-temperature near

surface processes during the emplacement of these basalts is discounted on the basis of (1) degassing and

hydrothermal alteration modeling, (2) the observed correlation of H2O with refractory incompatible

elements such as Ce and Th, and (3) the young age and freshness of the basalt glasses studied. The dDvariation shows a strong geographical coherence with the long-lived Pb and Sr isotope ratios, suggesting

long-lived upper mantle source heterogeneities in D/H and H2O as well. Binary mixing of the dD and H2O-

rich Salas y Gomez mantle plume with the depleted upper mantle low in dD and H2O is clearly evident in

the covariation of Pb radiogenic-bearing isotope ratios, dD, and H2O along the ESC-EMP. The positive

correlations of dD with Pb isotope ratios indicate that the Salas y Gomez mantle plume with HIMU-affinity

is rich in deuterium and in H2O. A mantle plume model with two possible variations, not mutually

exclusive, namely recycling of old hydrothermally altered and dehydrated crust, with or without

entrainment of primordial water from the lower mantle is proposed and evaluated.

Components: 12,489 words, 12 figures, 4 tables.

Keywords: Easter microplate; hydrogen isotopes; mid-ocean ridge basalts; mantle evolution; crustal recycling; seamounts.

G3G3GeochemistryGeophysics

Geosystems

Published by AGU and the Geochemical Society

AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES

GeochemistryGeophysics

Geosystems

Article

Volume 3, Number 4

13 April 2002

10.1029/2001GC000199

ISSN: 1525-2027

Copyright 2002 by the American Geophysical Union 1 of 26

Index Terms: 1025 Geochemistry: Composition of the mantle; 1040 Geochemistry: Isotopic composition/chemistry; 8121

Tectonophysics: Dynamics, convection currents and mantle plumes; 8124 Tectonophysics: Earth’s interior–composition and

state.

Received 2 July 2001; Revised 17 October 2001; Accepted 19 October 2001; Published 13 April 2002.

Kingsley, R. H., J.-G. Schilling, J. E. Dixon, P. Swart, R. Poreda, and K. Simons, D/H ratios in basalt glasses from the Salas y

Gomez mantle plume interacting with the East Pacific Rise: Water from old D-rich recycled crust or primordial water from the

lower mantle?, Geochem. Geophys. Geosyst., 3(4), 10.1029/2001GC000199, 2002.

1. Introduction

[2] Hydrogen (as water) plays a dominant role in

the Earth’s interior by lowering significantly the

viscosity of the mantle and magmas [Hirth and

Kohlstedt, 1996], and the temperature at which

melting occurs [Hirose and Kawamoto, 1995].

The amount of hydrogen in mantle rocks is related

to the solubility of molecular water or hydroxide

anions in nominally anhydrous mantle minerals

(e.g., olivine, pyroxenes, and garnet and higher

pressure counterparts) and the stability of hydrous

minerals (e.g., phlogopite, richterite, and amphib-

oles) [Green and Ringwood, 1967]. Methane may

also play some minor role, as well as other hydrides

at shallow mantle depths (low pressure), but these

effects are poorly understood and not addressed in

this investigation.

[3] Because of the large relative mass difference of

the two stable hydrogen isotopes 1H and 2H (i.e.,

deuterium, D), large D/H fractionation occurs at

relatively low temperature and pressure during the

hydrothermal alteration, mineralization, and meta-

morphism of rocks [Suzuoki and Epstein, 1976].

The hydrogen isotopic composition of basalts pro-

vides a potentially powerful means of tracing the

origin and source of water present in the plate

tectonic cycle and the different reservoirs in the

mantle.

[4] The few studies of MORB so far available have

revealed a D/H deviation from the Vienna Standard

Mean Ocean Water (VSMOW) value (dD � 0%) in

the range of dD = �40% to �85% [Taylor and

Sheppard, 1986]. These values fall in the same

common range of altered marine volcanic rocks,

marine sediments, and sedimentary rocks, thus

making ambiguous the distinction between so

called ‘‘juvenile’’ (primary) magmatic water from

the upper mantle, seawater recycled by plate tec-

tonics and mantle plumes, or for that matter pri-

mordial water associated with high 3He/4He in the

lower mantle, trapped during the early accretion of

the Earth [Craig and Lupton, 1976]. However, the

combination of D/H ratios with independent evi-

dence from radiogenic Pb, Sr, Nd, and He isotope

ratios on the same basalts can provide new insights

into these questions.

[5] Two schools of thought have developed in

interpreting D/H results:

1. On one hand, it is said that the upper mantle is

uniform with a dD of about �80%. Variations from

this value are primarily the result of secondary

processes occurring during the emplacement of

magmas in the crust, such as seawater alteration

and/or degassing of water at low pressures. The

uniformity of the upper mantle is considered to be

essentially a steady state value resulting from the

plate tectonic cycle, whereby newly formed oce-

anic crust is extensively altered by seawater hydro-

thermal circulation near mid-ocean ridges and

subsequently released by dehydration of the heated

slab at subduction zones [Taylor, 1974; Kyser and

O’Neil, 1984]. According to this view, any original

‘‘juvenile’’ water would have been overwhelmed

by such a process over geological time.

2. In the other school, the mantle is considered

to be heterogeneous and still poorly mixed in water

content and dD values. Radiogenic Pb and Sr

isotope and incompatible element rich MORB

found over or near hotspots, such as along the

Mid-Atlantic ridge near Iceland or Jan Mayen,

which also have higher H2O content and high dDvalues, have been interpreted as reflecting the

2 of 26

GeochemistryGeophysicsGeosystems G3G3 KINGSLEY ET AL.: D/H RATIOS IN BASALT GLASSES 10.1029/2001GC000199

melting of mantle plumes composed of old re-

cycled oceanic crust [e.g., Deloule et al., 1991;

Poreda, 1985; Poreda et al., 1986]. Mantle heter-

ogeneity in H2O and dD is likely to be most

pronounced in the mantle wedge above subduction

zones where seawater-derived fluids percolate

upward after slab dehydration. This is evident in

suprasubduction zone basalts (presumably derived

from the mantle wedge), which indeed contain

high H2O concentrations (>2%) and high dD (up

to �30%) [Poreda, 1985; Hochstaedter et al.,

1990].

[6] In order to further trace the origin of hydrogen

(water) in mantle plumes, as well as degassing/

ingassing of the Earth, we have undertaken a

collaborative study of H2O and CO2 contents, dDand d13C variations in mostly undegassed MORB

influenced by mantle plumes having distinct radio-

genic Pb-Sr-Nd-He isotope signatures. Here we

report on the dD% variation in basaltic glasses

erupted along the Easter-Salas y Gomez seamount

chain — Easter microplate (ESC-EMP) system in

the southeast Pacific (Figure 1).

[7] The ESC-EMP is a 1400 km long volcanic

feature, along which regular gradients in Pb and

Sr radiogenic isotope and trace element contents

reflect progressive dilution of the Salas y Gomez

mantle plume with the depleted upper-mantle, as it

flows along a sublithospheric channel toward the

East Pacific Rise at the latitude of the Easter

Microplate [Schilling et al., 1985; Hanan and

Schilling, 1989; Fontignie and Schilling, 1991;

Kingsley et al., 1997; Kingsley and Schilling,

1998; Pan and Batiza, 1998; Cheng et al.,

1999]. The observed geochemical gradients and

schematic of the working model are shown in

Figure 2.

[8] The isotopic composition of the Salas y Gomez

plume approaches that of the high U/Pb ratio

(HIMU) mantle component (such as the islands

of St. Helena, Tubuaii, and Mangaia [Zindler et al.,

3 of 26

ESC alkalic basaltsESC tholeiitic basaltsEMP east riftEN113 4D picriteEMP west riftEN113 26D picriteEN113 30D plagiophyricEPR N&S of EMP

6000

1000

2500

2800

3000

3200

3400

3600

4200

22

23

24

25

26

27

28

29

30 117 115 113 111 109 107 105 103

Longitude ˚W

Latit

ude

˚S

Figure 1. Dredge station location along the Easter microplate (EMP) and Easter - Salas y Gomez seamount chain(ESC) from which fresh basalt glasses were obtained. See key for meaning of color-coding used throughout this paper.Larger symbols represent samples selected for D/H ratios determined in this study.


1982; Hart, 1988]) but with slightly higher time-

integrated Th/Pb and Rb/Sr ratios and lower

time-integrated U/Pb, Sm/Nd, and Lu/Hf ratios

(R. H. Kingsley et al., manuscript in preparation,

2002). The Salas y Gomez plume is also character-

ized by high 3He/4He ratios reaching up to 17.6 R/Ra

[Poreda et al., 1993a; Poreda et al., 1993b; Poreda

et al., 1997].

[9] The basalt glasses analyzed in this study are

very fresh, young (<3 ma maximum) and erupted

on the seafloor at depth of 2–4 km below sea level,

4 of 26

20

19

18

EM

P

Wes

t Rift EM

P E

ast R

ift

Eas

ter

Isla

nd

Sal

as y

Gom

ez

206 P

b/20

4 Pb

(La/

Sm

) nM

gO (

%)

30

10

117 115 113 111 109 107 105 103

4

3

2

1

10

2

4

6

8

Longitude (°W)Li

thos

pher

eth

ickn

ess

(km

)


H2OLa/SmTh

∆T, X

206Pb204Pb

entraineddepleted

upper-mantle

EMPwes

t r

ift

east

rif

t

Ahu

Eas

ter

Sal

as y

G

omez

lithosphere

hot mantleplume

idealized longitudinal geochemical gradient

a

b δD‰

Figure 2. (a) Longitudinal variation of 206Pb/204Pb, La/Sm (normalized to chondrites), MgO content, and computedlithospheric thickness along the ESC and EMP. Pb isotope data from Hanan and Schilling [1989] and Kingsley andSchilling [1998], La/Sm from Table 3, and MgO data from Pan and Batiza [1998]. Symbols are as defined in Figure 1.(b) Schematic plume flow/ridge interaction model. The buoyant Salas y Gomez mantle plume rises from the deepermantle and is drawn toward the EMP along a sloping lithosphere as a result of the pressure gradient and the spreadingcenter acting as a sink. Binary mixing takes place in the shallow upper mantle by entrainment of depleted uppermantle (blue color) with HIMU-like mantle plume (red color) enriched in incompatible and moderately incompatibletrace elements. The binary mixing results in the longitudinal gradients (idealized profile above cartoon). Modelmodified from Kingsley and Schilling [1998] and Pan and Batiza [1998].


that is mostly at pressures high enough to prevent

significant exsolution of water vapor from basaltic

melts (K. Simons et al., Volatiles in basaltic glasses

from the Easter-Salas y Gomez Seamount Chain

and Easter Microplate: Implications for geochem-

ical cycling of volatile elements, submitted to

Geochemisty, Geophysics, Geosystems, 2001, herin-

after referred to as Simons et al. submitted manu-

script, 2001), and minimizing degassing Rayleigh

fractionation effects on D/H ratios. The glasses

range in composition from tholeiitic west of Easter

Island and on the EMP, to alkalic east of Easter

Island in the intraplate setting [Pan and Batiza,

1998, and Figure 1].

2. Analytical Methods

2.1. Selection Criteria

[10] Basalt glass samples from 22 dredge stations

were selected primarily on the basis of their

unweathered appearance, depth of eruption, and

geographical coverage. The glassy rind of the

dredged lavas were crushed and cleaned by ultra-

5 of 26

200 300 400 500 600 700 800 900 1000 1100 12000

50

100

150

200

250

300

350

Extraction Temperature (˚C)

200 300 400 500 600 700 800 900 1000 11000

Extraction Temperature (˚C)

10

20

30

40

50a.

b.

EN113 36D-1gtholeiite

GL07 D46-2galkali basalt

δD = -59‰

δD = -46‰

δD = -43‰

δD = -53‰

95% of H2O

98% of H2O

H2O

(µm

)H

2O (

µm)

Figure 3. H2O release spectrum as a function of temperature during step heating under vacuum. (a) Typicaltholeiitic glass and (b) alkali basalt glass. Note the lower temperature at which most of the H2O is outgassed (750�–800�C) for the alkali basalt relative to the tholeiite. Note >95% of the H2O is released between 750�–1000�C.


sonication in deionizedwater. Glass chips (1–5mm)

were then examined using a binocular microscope,

and only those without surficial contamination or

visible (>10 mm) vesicles were selected for analyses.

H2O and CO2 concentrations of these glasses from

the EMP-ESC were measured using infrared spec-

troscopy at the University of Miami. Water contents

vary from 0.07 to 1.50 wt%, and dissolved CO2

contents vary from 50 to 350 ppm. The results of

these analyses have been reported separately

(Simons et al., submitted manuscript, 2001).

2.2. Volatile Extractions

[11] The samples (weighing 0.25–3.0 g depending

on water concentration) were placed into an out-

gassed furnace for heating of the sample in step

increments. Prior to each heating step, the samples

were exposed to oxygen from O2 desorbed from a

molecular sieve in the extraction line for the

purpose of providing an oxidizing atmosphere

during the extraction. Samples were brought to

prescribed temperature and allowed to degas for

30 min at each step. These steps were 600�–800�C

6 of 26

Table 1. Interlaboratory Comparison of SelectedSamplesa

Mohns Ridge Samples dD(SMOW)

(RSMAS)Poreda et al. [1986]

(SIO)

EN026 10D-3g �55.3% �55.8%TR139 31D-2g �62.9% �60.3%TRI39 30D-1g �45.6% �44.1%

aRSMAS measurements relative to a �120% standard gas and

then normalized to the average of three SMOWaliquots run during thistime period (+124.3% relative to the standard gas).

Table 2. D/H Ratios and Water Concentrations in Glasses From the Easter-Salas y Gomez Seamount Chain, theEaster Microplate, and the East Pacific Rise

Sample Identification Latitude,�S

Longitude,�W

ExtractionTemperature

dD%,(SMOW)

H2O wt% Rock Type

Easter-Salas y Gomez ChainGL-07 D30-1g 26.067 103.69 600–1000 �37 1.54 alkalic

700–1000 �42GL-07 D32-4g 26.353 104.304 700–1000 �43 1.09 alkalicGL-07 D42-1g 27.182 106.604 700–1000 �36 1.04 alkalicGL-07 D43-1g 26.792 107.018 600–1000 �44 0.94 tholeiiticGL-07 D47-2g 26.628 108.287 700–1000 �41 0.88 alkalicGL-07 D46-2g 26.206 107.71 600–1000 �44 1.50 tholeiitic

650–1000 �51GL-07 D48-4g 26.425 108.559 600–1000 �40 1.26 alkalic

700–1000 �33GL-07 D50-1ag 26.752 109.453 600–1000 �41 1.01 tholeiiticGL-07 D51-3g 26.738 110.118 700–1000 �52 0.44 tholeiiticGL-07 D56-2g 27.155 109.845 800–1000 �47 0.65 tholeiiticGL-07 D52-5g 26.702 110.84 600–1000 �51 1.00 tholeiiticGL-07 D53-2g 26.562 111.268 600–900 �54 0.77 tholeiitic

Easter Microplate East RiftEN113 41D-1g 26.58 112.65 800–1000 �50 0.49 tholeiiticEN113 4D-1Ag 27.53 112.76 800–1000 �55 0.14 picriticEN113 5D-1Ag 27.21 112.77 800–1000 �49 0.10 tholeiiticEN113 6D-1Ag 26.81 112.64 800–1000 �38 0.76 tholeiitic

700–1000 �37EN113 7D-1g 26.46 112.57 800–1000 �43 0.69 tholeiitic

Easter Microplate West RiftEN113 24D-1Ag 25.9 115.86 600–1000 �57 0.27 tholeiiticEN113 26D-1g 25.28 116.27 700–1000 �63 0.08 picriticEN113 30D-1g 24.11 115.41 700–1000 �40 0.07 tholeiitic

East Pacific RiseEN113 35D-1g 22.53 114.47 700–1000 �57 0.16 tholeiiticEN113 36D-1g 22.68 114.51 700–1000 �59 0.18 tholeiitic


and 800�–1000�C, or in some cases a single step of

700�–1000�C was taken. These increments were

selected because during method development,

glasses were step heated in increments of 100�Cand 200�C from 200�C to 1200�C in order to

determine the release spectrum of H2O and other

volatiles in these basalt glasses. We found that the

bulk of the water in the glasses of tholeiitic compo-

sition is exolved during the 800�–1000�C heating

step (see Figure 3a). This is consistent with several

other studies [Byers et al., 1986; Delaney et al.,

1978; Kyser and O’Neil, 1984], where water lib-

eration was found to peak at 900�C ± 50�. For alkalibasalt glasses, however, we found that the bulk of

H2O is degassed at lower temperature (800�C ± 50�)(Figure 3b). This may be a result of the lower

solubility of H2O in alkali basalts [Dixon et al.,

1997a]. At the end of the 30 min, all gasses liberated

during a particular heating step were transferred

through the extraction line, and CO2, SO2, and H2O

were condensed in a liquid nitrogen (LN2) cold trap

at �195�C. The CO2 (saved for isotopic measure-

ments) and SO2 (discarded) were selectively sepa-

rated from H2O by warming the cold finger to

�130�C (with a mixture of LN2 and pentane) and

to �75�C (with a mixture of LN2 and methanol),

respectively. The cold trap was then heated to

>100�C, and the pressure of the water vapor was

measured using the capacitance manometer to deter-

mine whether there was enough water vapor present

in the particular heating step to measure in the mass

spectrometer. If so, the water vapor was then reacted

with uranium metal to produce pure hydrogen gas.

The H2 was transferred to a sample vessel using a

Toepler Pump (mercury-ram pump) and then

removed from the extraction line and placed on

the mass spectrometer for measuring the D/H ratio.

2.3. Mass Spectrometer Measurements andCorrections

[12] D/H determinations were made on the Finnigan

MAT 251 mass spectrometer at the Rosentheil

School of Marine and Atmospheric Sciences

(RSMAS), University of Miami. Prior to making

the measurements, the instrument was tuned for

optimal sensitivity and characterized for the H3

interference factor. A standard with a dD%(VSMOW)

� 0% was used as the standard gas in the mass

spectrometer except for two samples (GL07 D43-1g

and GL07 D53-2g) for which a dD%(SMOW) =

�120% standard gas was used.

[13] Several microliter-sized aliquots of Standard

Mean Ocean Water were run against the standard

gas over the time period of the glass analyses and

resulted in an average of +8.2 ± 4.0% (n = 9).

Reproducibility of the glass extraction analyses is

estimated at ± 0.9 to ± 5.2% (1s) based on multiple

runs of several samples. We attribute this variability

to sample heterogeneity and/or to variable extrac-

tion parameters that we were unable to control. The

precision error on the mass spectrometric measure-

ments was always less than 1%.

2.4. Step Heating Results

[14] On several occasions, where more than one

temperature step contained enough water to meas-

ure, the lower temperature step contained hydrogen

with a lower dDSMOW value (i.e., lighter) than the

hydrogen liberated at 800�–1000�C (see Figure 3b).

The difference (dD800�C – dD1000�C) averaged

around �20%. This is consistent with what Kyser

and O’Neil [1984] found in their step-heating

experiments of MORB glasses. This effect could

result from (1) preferential extraction of isotopi-

cally distinct species of H (e.g., molecular H2O

over OH�) at a lower temperature, (2) temperature

dependent diffusion of H over D during extraction

(the explanation preferred by Kyser and O’Neil), or

(3) water trapped in vesicles that is liberated during

the 600�–800�C step, which contains isotopically

distinct (lighter) hydrogen. Two studies that meas-

ured the D/H fractionation between vesicles and

glass observed dDSMOW (vesicle-dissolved) = +27%to +40%, indicating heavier water in vesicles

relative to dissolved water in basalt glasses [Kyser

and O’Neil, 1984; Pineau and Javoy, 1986]. This is

contrary to our measured differences in the labo-

ratory during step heating.

[15] In cases where significant amounts of water

was liberated in more than one step, the value taken

to represent the dDSMOW value of magmatic hydro-

gen in the basaltic glasses was calculated using a

weighted average based on the pressures (mass 2) of

7 of 26


the two aliquots measured on an equal volume basis

at the inlet of the mass spectrometer. The choice to

measure only these steps was made in order to

minimize the effect of adsorbed water released at

low temperatures and to minimize the effect of

variable amounts of water in micro vesicles affect-

ing the D/H measurements. Table 1 is an interlabor-

atory comparison of three glasses analyzed with the

above method and compared with the identical glass

samples measured by Poreda et al. [1986] at SIO.

The two laboratory results are within ±3%.

3. Results

[16] Table 2 lists the results of the 22 analyses

carried out during this study. The overall dDSMOW

8 of 26

Figure 4. (a) Longitudinal variation of dD, H2O, Ce, and Th along the ESC-EMP. Th and Ce data are given in Table 3.MgO data is from Pan and Batiza [1998], and H2O is from Simons et al. (submitted manuscript, 2001). (b) Latitudinalvariation along the EMP east and west rifts for 206Pb/204Pb, dD, and La/Sm (normalized to chondrites). Note the peakin plume-influenced basalts at 27�S on the EMP east rift, where the Salas y Gomez plume is entering the spreadingcenter. Dark lines in the profiles connect the samples from the EMP east rift, and the gray dashed line connects thesamples for the west rift.


range observed for the ESC-EMP is �63% to

�33% over a H2O content of 0.07% to 1.54%.

The dD values range from �63% to �57% for

samples located on the EPR north of the EMP and

from the EMP west rift. For the hot spot-affected

east rift of the EMP, the dD values are slightly

heavier on average and range from �55% to

�43%. Even heavier dD values, ranging from

�54% to �33%, characterize the basalt glasses

from the intraplate ESC seamounts which are also

the most radiogenic in Pb and Sr isotope ratios.

For comparison, dD values in previously analyzed

MORB glasses generally range from �90% to

�50% [Craig and Lupton, 1976; Kyser and

O’Neil, 1984; Poreda et al., 1986; Chaussidon

et al., 1991; Pineau and Javoy, 1994; Agrinier et

al., 1995], with the range extending to �44% for

MORB near the Jan Mayen hot spot in the Arctic

[Poreda et al., 1986]. Thus, as along the MAR

affected by the Jan Mayen hot spot, where high

dD values and radiogenic Pb and Sr isotope ratios

occur, basalts on the EPR most affected by the

Salas y Gomez hot spot also have high dD values.

In contrast, lower dD values (�88% to �54%)

are observed on the EPR unaffected by mantle

plumes (e.g., 21�N to 13�S) [Craig and Lupton,

1976; Kyser and O’Neil, 1984; Chaussidon et al.,

1991].

[17] The longitudinal profiles along the ESC-EMP

(Figures 2a and 4a) show that both dD and H2O

decrease regularly from highs near Salas y Gomez

where the mantle plume is located, to typical

upper mantle values over the EMP west rift and

the EPR north and south of the EMP. The dD and

H2O profiles essentially mimic the profiles of

radiogenic Pb isotope ratios or ratios of highly

incompatible to moderately incompatible trace

elements such as La/Sm and Th/H2O, as well

as the concentration of incompatible elements

such as Ce and Th (Figures 2a and 4a and Table 3).

This behavior extends to the latitudinal profile

along the EPR-EMP north to south as well, with

dD, H2O, and Pb and Sr isotope ratios peaking at

27�S on the east rift of the EMP, which is the ridge

segment most affected by the Salas y Gomez

mantle plume (Figure 4b). The east rift geochem-

ical anomaly has been interpreted to reflect the

approximate width and compositional structure of

the lateral channel of asthenospheric plume mate-

rial entering the spreading center at this location

[Schilling et al., 1985; Hanan and Schilling, 1989;

Fontignie and Schilling, 1991; Kingsley et al.,

1997; Kingsley and Schilling, 1998]. The associa-

tion of high dD values with the long-term Pb

radiogenic isotope anomalies is a clear indication

that the mantle beneath Salas y Gomez is also rich

in deuterium as well as H2O and that mantle

heterogeneity in D/H and H2O do indeed occur,

as first suggested by Poreda et al. [1986] from their

study of MORBs in the vicinity of the Jan Mayen

hot spot.

[18] The close geographical coherence of dD with

radiogenic Pb isotopes is also evident in their

essentially linear covariation (Figure 5), as well

as with ratios of a highly incompatible element

relative to H2O (e.g., Th/H2O, Figure 6a). On one

hand, the covariation of dD with H2O could

conceivably be considered as resulting from pro-

gressive seawater hydrothermal alteration of mag-

mas while penetrating the crust/lithosphere lid,

whose thickness increases eastward along the

ESC [Liu, 1996; Kruse et al., 1997; Rappaport

et al., 1997]. On the other hand, covariation with

D/H and H2O shown on Figures 5 and 6 are

consistent with the previously proposed binary

mixing model based on Pb and Sr isotope ratios,

invoking a progressive dilution of the Salas y

Gomez mantle plume by entrainment of depleted

upper-mantle material, as the plume flows toward

the East Pacific rise along a shallow sublitho-

spheric channel [Schilling et al., 1985; Hanan and

Schilling, 1989; Fontignie and Schilling, 1991;

Kingsley and Schilling, 1998]. The fact that dDcovariation with 1/H2O is not linear (Figure 7), as

would be anticipated if mixing was the last

process to take place, suggests that binary mixing

occurred not in magma chambers or during erup-

tion but prior to or during partial melting.

[19] The observed Pb and Sr isotopic composition

of the Salas y Gomez mantle plume (206Pb/204Pb �20 � 21.5 and 87Sr/86Sr � 0.7030 � 0.7035

[Fontignie and Schilling, 1991; Kingsley et al.,

1997; Kingsley and Schilling, 1998; Cheng et al.,

9 of 26


Table 3. Trace Element Concentrations and La/Sm Ratio in Basalts From the Easter-Salas y Gomez SeamountChain and From the Easter Microplate Boundariesa [The full Table 3 is available in ASCII tab-delimited format athttp://www.g-cubed.org.]

Sample ID Latitude �S Longitude �W La (ppm) Ce (ppm) Sm (ppm) Th (ppm) (La/Sm)N

Easter-Salas y Gomez Seamount ChainGL-07 D30-1g 26.07 103.69 40.1 87.9 9.44 4.41 2.90GL-07 D31-2g 26.16 104.35 87.7 199.7 18.6 9.98 3.27GL-07 D32-2g 26.35 104.30 21.8 45.7 5.23 2.55 2.92GL-07 D32-4g 26.35 104.30 22.1 48.9 5.23 2.55 2.94GL-07 D34-1g 26.51 104.93 29.8 66.9 7.66 3.45 2.87GL-07 D34-3g 26.51 104.93 31.6 69.4 7.44 3.51 2.81GL-07 D35-7 26.37 105.19 38.4 76.3 7.10 5.35 3.64GL-07 D36-1g 26.38 105.83 41.8 95.2 12.8 4.53 2.24GL-07 D37-2 26.55 106.26 45.1 99.9 10.9 4.57 2.74GL-07 D37-2g 26.55 106.26 45.5 102.2 11.6 4.73 2.76GL-07 D38-5g 26.76 106.15 20.9 47.5 5.64 2.27 2.63GL-07 D39-1g 26.70 105.96 47.5 109.1 12.3 5.10 2.62GL-07 D42-1g 27.18 106.60 31.7 71.2 6.87 3.56 3.08GL-07 D42-3g 27.18 106.60 31.6 67.2 6.71 3.55 3.13GL-07 D43-1g 26.79 107.02 22.0 49.7 6.55 2.21 2.24GL-07 D43-2g 26.79 107.02 20.4 47.5 6.45 2.19 2.27GL-07 D43-4g 26.79 107.02 22.5 49.3 6.36 2.18 2.29GL-07 D45-1g 26.39 107.49 32.8 72.5 7.19 3.65 3.09GL-07 D45-2g 26.39 107.49 31.8 67.8 7.02 3.62 3.15GL-07 D46-2g 26.21 107.71 31.8 71.9 7.14 3.76 3.11GL-07 D46-3g 26.21 107.71 33.7 72.7 7.46 3.95 3.12GL-07 D47-2g 26.63 108.29 23.3 51.2 5.49 2.44 2.78GL-07 D47-4g 26.63 108.29 24.1 51.6 5.50 2.66 3.00GL-07 D48-2g 26.43 108.56 24.6 54.5 7.65 2.57 2.22GL-07 D48-4g 26.43 108.56 25.0 57.8 7.61 2.51 2.22GL-07 D48-5g 26.43 108.56 24.6 55.4 7.67 2.57 2.22GL-07 D49-1g 26.28 108.75 17.4 39.4 5.11 1.86 2.42GL-07 D49-8g 26.28 108.75 17.1 39.9 5.09 1.86 2.39GL-07 D50-1g 26.75 109.45 22.2 50.7 6.69 2.25 2.32GL-07 D51-1g 26.74 110.12 9.55 24.7 4.79 0.775 1.38GL-07 D51-3g 26.74 110.12 9.49 24.3 4.79 0.778 1.37GL-07 D51-5g 26.74 110.12 9.36 24.0 4.81 0.777 1.38GL-07 D52-1g 26.70 110.84 11.0 28.2 5.66 0.924 1.34GL-07 D52-4g 26.70 110.84 11.0 28.3 5.84 0.948 1.31GL-07 D52-5g 26.70 110.84 12.2 31.3 6.14 1.03 1.39GL-07 D53-1g 26.56 111.27 13.8 34.0 6.40 1.17 1.53GL-07 D53-2g 26.56 111.27 14.1 36.2 6.41 1.18 1.54GL-07 D53-5g 26.56 111.27 5.93 17.2 4.54 0.350 0.89GL-07 D55-1g 27.04 110.32 17.8 47.0 8.27 1.57 1.55GL-07 D56-2g 27.16 109.85 13.0 31.1 4.82 1.17 1.89GL-07 D56-3g 27.16 109.85 13.0 31.1 4.85 1.18 1.88GL-07 D57-2g 27.32 108.69 14.1 34.9 5.75 1.23 1.73GL-07 D57-3g 27.32 108.69 14.4 35.1 5.82 1.26 1.66GL-07 D58-2g 27.46 108.31 14.6 34.4 5.15 1.32 2.01GL-07 D59-2g 27.62 108.03 17.5 41.6 6.05 1.66 2.05GL-07 D59-3g 27.62 108.03 16.9 41.2 6.15 1.68 2.02GL-07 D60-1g 27.38 107.99 2.60 8.32 2.23 0.096 0.80GL-07 D60-3g 27.38 107.99 2.75 8.73 2.29 0.099 0.82GL-07 D61-2g 27.27 108.50 5.41 14.70 3.30 0.399 1.06

Easter IslandEA P3BA 26.92 109.35 45.6 107.3 14.2 5.46 2.25EA P4 26.92 109.35 21.6 47.7 6.16 2.60 2.41

aThe ‘‘g’’ in sample identification stands for glass; otherwise, interior rock chips were analyzed. Th, and Sm concentrations analyzed by isotope

dilution ICP-MS. La and Ce concentrations were measured by ICP-MS using internal and external standardization. Precision of the method is basedon 11 replicate analyses of the three U.S. Geological Survey (USGS) standards (BIR-1, BHVO-1, BCR-1) and ranges from 0.5% to 3.1% s.d.depending on the element and concentrations measured. (La/Sm)N = La/Sm ratio normalized to C1 chondrites.


1999]) similar to some basalts from Comores

[Reisberg et al., 1993] is generally interpreted as

representing 1.5–2 Ga old recycled hydrothermally

altered oceanic crust with a minor amount of

recycled sediments [Chase, 1981; Hofmann and

White, 1982; Zindler et al., 1982; Hart, 1984].

Self-consistency would require that the corre-

sponding high dD and H2O contents apparently

observed in the Salas y Gomez plume would have

been caused by relatively low temperature D/H

fractionation 1–2 Ga ago, during the initial for-

mation closely followed by subduction of such

recycled oceanic crust.

[20] It is also possible that the high D/H and H2O

observed represents entrainment by the Salas y

Gomez mantle plume of ‘‘primordial’’ water trap-

ped in the lower mantle since the early accretion of

the Earth, considering that many of the EMP and

ESC basalts have 3He/4He ratios greater than

normal MORB unaffected by mantle plumes [Por-

eda et al., 1993a, 1993b, 1997]. These two possi-

bilities are presented below, and their pros and cons

further evaluated.

[21] The anomalous sample EN113 30D-1g devi-

ates from the general trends of dD versus radio-

genic isotope ratios or dD versus H2O concen-

tration and is significantly deuterium enriched

relative to the other depleted MORBs in the

data set. It is possible that this highly plagiophyric

glass may have suffered some addition of seawater,

as at its low H2O concentrations (0.07%) a very

small amount of hydrothermal alteration will have a

large effect on the dD values. Although the H2O/Ce

ratio of this sample (168) does not indicate added

water (see section 4), as little as 0.017% added

seawater (dD = 0%) to a glass of �56% would be

enough to raise the dD value to �42% and still

keep the H2O/Ce ratio within the range of that

observed for the microplate glasses.

[22] Two unusual picritic glasses represented in

this data set need special attention. These are

EN113 4D-1g, a sample enriched in Pb and Sr

radiogenic isotopes but depleted in the light rare

earths and other large ion lithophile elements, as

well as in H2O; and EN113 26D-1g, a picritic

basalt, which is the most depleted sample in terms

11 of 26

-65

17.5 18.0 18.5 19.0 19.5 20.0 20.5206 Pb /204 Pb

-35

-40

-45

-50

-55

-60

-65

δD‰

(sm

ow)

-30

EN113 30D-1g

EN113 4D-1g

EN113 26D-1g


Figure 5. Covariation of dD with 206Pb/204Pb isotope data from Hanan and Schilling [1989] and Kingsley andSchilling [1998]. Symbols are as defined in Figure 1.


of radiogenic isotope ratios and in incompatible

trace elements, including the light rare earth

elements. These picritic glasses may be the prod-

uct of large degrees of fractional melting [e.g.,

Wendt et al., 1999]. In the case of EN113 26D-1g,

the significant gap evident between this sample

and the rest of the mixing array on the Pb-Pb

isotope diagram (Figure 6b) could be explained by

fractional melting beyond the point where any

enriched mantle component is exhausted (namely

a veined mantle).

4. Effects of Magma EmplacementProcesses on D/H

[23] Considering that the ESC basalts must have

penetrated a crust and lithosphere of increasing

thickness from the EPR to Salas y Gomez [Liu,

12 of 26

(Th/H2O)x105

-35

-40

-45

-50

-55

-60

-65

-30

3530252015105

41

40

39

38

37

3617.5 18.0 18.5 19.0 19.5 20.0 20.5

binary mixing

EN113 30D-1g

EN113 4D-1g

EN113 26D-1g


206Pb / 204Pb

208 P

b/2

04P

bδD

‰(s

mow

)

a.

b.

Figure 6. (a) Covariation of dDwith Th/H2O ratio. The linearity observed is consistent with binary mixing of a waterand deuterium rich end member (plume) with a water poor and light hydrogen end member (depleted asthenosphere),as previously inferred by the highly linear covariation of 208Pb/204Pb with 206Pb/204Pb [Kingsley and Schilling, 1998]shown in Figure 6b. H2O data is from Simons et al. (submitted manuscript, 2001), Th data unpublished from this lab.Symbols are as defined in Figure 1. Symbols with bars show dD average and range obtained for duplicate analyses.


1996; Kruse et al., 1997; Rappaport et al., 1997], it

follows that both dD and H2O likely also correlate

with lithospheric thickness (Figure 2). A positive

correlation of dD with H2O, such as observed in the

ESC-EMP region (Figure 8), has previously been

attributed by Kyser and O’Neil [1984] either to (1)

progressive degassing of isotopically heavy water-

rich magmas, (2) progressive addition of seawater

to water poor magmas, or (3) increase in hydro-

thermal alteration. Thus it is conceivable that the

intensity of these three processes might increase

with increasing lithosphere thickness along the

ESC. We briefly explore these postulates quantita-

tively and show how such hypotheses can be ruled

out as the primary cause of the D/H and H2O trends

observed along the ESC-EMP.

4.1. Degassing

[24] ‘‘Open’’ and ‘‘closed’’ system degassing mod-

els [Broecker and Oversby, 1971; Taylor et al.,

1983; Taylor and Sheppard, 1986] are superim-

posed on the observed dD versus H2O data in

Figure 8. In order to explain the total variation

observed along the ESC-EMP, the Dvapor-melt

(where Dvapor-melt = dDvapor � dDmelt) would have

to be +5% to +25% (avapor-melt = 1.005 to 1.026),

that is, within the range expected in systems at

magmatic temperatures and crustal pressures deter-

mined from experimental and natural systems [e.g.,

Kyser and O’Neil, 1984; Pineau and Javoy, 1986;

Taylor and Sheppard, 1986]. However, both these

degassing models, as applied to the EMP-ESC, can

be ruled out on the following bases:

1. The maximum degassing predicted (90%, i.e.,

lowest dD and H2O content) would have to have

occurred where the lithosphere is thinnest (i.e., the

EMP) which is contrary to the above postulate of

thickening lithosphere eastward along the ESC.

2. Up to 90% degassing predicted by the open

and closed system models is clearly excessive and

ruled out on several other grounds. There is no

13 of 26

-30

-40

-50

-60

-70

2 4 6 8 10 12 141/ H2O (%)

EN113 30D-1g

EN113 4D-1g

EN113 26D-1g

0.05wt% H O2±

δD‰

(sm

ow)

Figure 7. H isotope ratios versus 1/H2O in ESC-EMP basalts. Pure binary mixing should produce linear arrays onsuch diagrams. Observed curvilinear relationship indicate that the absolute concentration of the H2O has been affectedby recent emplacement processes after mixing. Note that the greatest departure from linearity occurs at the depleted endof the data array where H2O content is lowest and analytical errors or geologic variability is maximized. Error bars showthe effect of as little as 0.05 wt% variability on the inverse of H2O concentration. Symbols are as defined in Figure 1.


correlation between dD and 2–4 km ocean depth of

eruptions (200–400 bars) at which these basalt

glasses were quenched (not shown). Only minor

H2O degassing by vesiculation would be expected

at such ocean depths unless unusually high CO2

concentrations were initially present [Dixon and

Stolper, 1995], but this does not appear to be the

case (R. J. Poreda et al., manuscript in preparation,

2002). Nor is there a correlation between dD and

the volume fraction of vesicles observed in these

basalt glasses (not shown and see Simons [2000]

for vesicularity data). For example, the maximum

H2O degassing expected from the most vesicular

basalt glass studied (GL07 D42-3g with �25%

vesicles) is calculated to be 18%, using Dixon and

Stolper [1995] T, P, H2O-CO2 vapor/melt thermo-

dynamical model and the H2O and CO2 content

observed in this basalt glass.

3. Perhaps, the strongest argument against vari-

able degassing controlling the dD and H2O along the

EMP-ESC is the observed strong covariation of H2O

(volatile species) with Ce (refractory) (r2 = 0.88

14 of 26

10%

30%

50%

% hydrothermal water 70%60%50%40%30%20%10% 80% 90%

+50

‰

+25‰

∆(v-m) = +5‰

∆(v-m) = +30‰

2H O (%)

-80

-70

-60

-50

-40

-30

-20

1.8 2.01.61.41.21.00.80.60.40.20.0

Closed System Degassing

Taylor 1977

openclosed

2% H O Degassed 020406080100

δD‰

(sm

ow) hydrothermal alteration

Ray

leig

hdis

tillati

on

Figure 8. Comparison of covariation of H2O and dD in basalt glasses along the EMP-ESC with modeled trendspredicted for certain magma emplacement processes. Blue lines indicate open and closed system hydrothermalalteration models [Taylor, 1977]. Closed system alteration: dDr,f = (D(r�w)(W/R) + dDr,i)/(1 + W/R), where dDr,f = dDof altered rock for a given W/R.W/R is water/rock mass ratio; D(r�w) is fixed dD fractionation factor, assumed equal to�25%, dDr,i = dD of water poor unaltered basalt, assumed equal to �65%. Open system alteration: (W/R)open =ln[(W/R)closed + 1] [Taylor, 1977]. Solid green lines show open system degassing models (Rayleigh distillation) forvarious values of the fractionation factor, agas-melt = [(D/H)gas/(D/H)melt] [Taylor and Sheppard, 1986]: dDmelt =dDinitial + (1000 + dDinitial)((1 -F)(a � 1)�1), where F is mass fraction of H2O degassed, and 1000lna � dDvesicle gas �dDbasalt melt = D(v-m). A magma with 1.8% H2O and dD value of �35% will exhibit highly curvilinear behavior duringopen system degassing. Dashed green line indicates closed system degassing model: dDmelt = dDinitial � 1000Flna[Taylor, 1986]. This model will produce a straight line. In either type of degassing, >90% degassing of H2O would berequired to explain the observations.


and Figure 9), which suggests a similar effective

crystal/melt partitioning behavior of OH- anion

and the Ce+3 cation during magmatic processes

[Dixon et al., 1988; Michael, 1988; Dixon and

Stolper, 1995; Michael, 1995; Danyushevsky et al.,

2000; Simons, 2000; Simons et al., submitted

manuscript, 2001].

[25] The average of H2O/Ce ratios for the EMP

glasses is 175 ± 23 and for the ESC glasses is 210 ±

20 (Simons et al., submitted manuscript, 2001). It is

similar to those found in undegassed deeply erupted

MORB glasses worldwide (150 to 213), with the

exception of the north Atlantic MORB which aver-

age around 240–280 [Michael, 1995;Danyushevsky

et al., 2000]. H2O-degassed MORB glasses would

have low H2O/Ce ratios, such as basalt erupted on

the shallow portion of the Reykjanes Ridge near

Iceland (e.g., TR101 11D-2g with H2O/Ce = 87)

[Moore and Schilling, 1973, J.-G. Schilling, unpub-

lished data, 1998;Kingsley, 1989]. In contrast, back-

arc basin basalts have high H2O/Ce ratios ranging

from 470 to >3000 [Hochstaedter et al., 1990;

Danyushevsky et al., 1993; Kamenetsky et al.,

2001], resulting fromwater addition from the mantle

wedge by dehydration of the subducted hydrated

oceanic crust. Thus the high degree of correlation

between H2O and Ce in the EMP-ESC glasses

effectively rules out any effect of significant vapor

phase degassing (vesiculation) because the H2O/Ce

ratio would not be expected to be constant had

variable degassing of H2O occurred or had variable

amounts of hydrous minerals been present in the

melting regime. For the same reasons, significant

addition of seawater to magmas causing the linear

correlation between dD values and H2O variation in

the ESC-EMP region is also ruled out (with the

possible exception of EN113 30D-1g as discussed

previously in section 3).

4.2. Postemplacement Seawater Interaction

[26] A thicker crust may lead to longer magma

residence times and/or a greater chance for sea-

15 of 26

1.8

1.6

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0.0

H2O

(w

t.%)

0 10 20 30 40 50 60 70 80 90 100 110 120Ce (ppm)


Figure 9. Covariation of Ce with H2O. The strong covariation of the volatile H2O with the refractory Ce+3 cationindicates similar crystal/melt partitioning during magmatic processes for H2O and Ce. The correlation rules out anyrecent secondary alteration by seawater or degassing of H2O. H2O and Ce data are from Simons et al. (submittedmanuscript, 2001), and Ce unpublished this lab. Symbols are as defined in Figure 1.


water hydrothermal alteration or assimilation of

altered wall rock during magma ascent. Following

Taylor [1977], we have modeled seawater hydro-

thermal alteration for both closed or open system

conditions (Figure 8). Unaltered basalt glasses at

the EPR are assumed to have a dD of �65%, as

observed, and a fraction factor Dseawater-basalt =

�27% [e.g., Satake and Matsuda, 1979]. The

mismatch between the data and the two models is

clear. The dD values observed in the EMP-ESC

basalts are not high enough to be explained by

hydrothermal mixing models as the overlay in

Figure 8 indicates (again, with the possible excep-

tion of EN113 30D-1g).

[27] More importantly, the strong correlation of

dD with 206Pb/204Pb, which can hardly be affected

by seawater interaction, as well as the strong

geographic coherence of these two parameters, is

not consistent with such alteration effects. This, of

course, does not rule out the possibility of hydro-

thermal action in ancient times imparting high dDvalues on altered crust prior to its recycling

through the mantle (see crust recycling model

below).

4.3. Fractional Crystallization

[28] Surprisingly perhaps, dD as well as Pb iso-

tope ratios and trace element ratios anticorrelate

with MgO content (Figure 10). Although one

might be tempted to suggest that the D/H may

be fractionated during fractional crystallization, it

is generally accepted that little fractionation exists

between solid and melt at magmatic temperatures

greater than 900�C (Dmineral-melt = 0% ± 10%,

Dmineral-melt = dDmineral � dDmelt) [Suzuoki and

Epstein, 1976; Graham et al., 1984; Kyser and

O’Neil, 1984; Taylor, 1986]. Fractional crystalli-

zation modeling based on the major element

trends for theses glasses using the MELTS algo-

rithm [Ghiorso and Sack, 1995; Pan and Batiza,

1998] (F = 0 to 65% at MgO = 8.5% to 4%,

respectively) provided melt temperatures ranging

from 1250�C to 1100�C, clearly ruling out any

significant D/H fractionation. More importantly,

the covariation of Pb radiogenic isotope ratios

with dD, H2O, and trace element ratios cannot be

explained by fractional crystallization. A more

detailed examination of the isotope ratios and

trace element ratio trends with MgO reveal the

lack of trends on a more local, level (see fields

drawn around localized data in Figure 10). The

lower MgO content in glasses near Salas y Gomez,

where the crust is also thicker, are likely the result

of longer magma residence times at higher pres-

sures with more evolved crystallization histories

prior to eruption, than more westward on the ESC

and on the EMP and EPR [Pan and Batiza, 1998].

Thus the isotopic covariations with MgO are

merely an artifact of the geographical variation of

magma types and binary mixing of mantle sources

along the ESC-EMP. We conclude that recent,

shallow depths, magma emplacement fractionation

effects cannot be the primary causes of the observed

covariations of dD with H2O and Pb isotopes

observed along the EMP-ESC.

5. Discussion

[29] The covariation and geographical coherence

along the ESC-EMP of the D/H ratio and H2O

content with the long-lived Pb radiogenic isotopes

clearly indicate that the mantle is heterogeneous

with respect to D/H and H2O. This has also been

previously established along the MAR affected by

the Jan Mayen mantle plume [Poreda et al., 1986].

These two occurrences reinforce the idea that the

distribution of D/H and H2O in the upper mantle is

not uniform. Having also ruled out recent emplace-

ment processes as the cause of the high D/H

observed over Salas y Gomez region, we now turn

to our preferred plume model with two possible

variations, not mutually exclusive, namely recy-

cling of hydrothermally altered and dehydrated

oceanic crust, with or without entrainment of

primordial water from the lower mantle. For the

sake of clarity, we discuss these two possibilities

separately.

5.1. Recycling of Altered Oceanic Crust

[30] In the recycling scenario based on the HIMU

affinity of the Pb and Sr isotope composition in

the Salas y Gomez mantle plume, the high dDvalue also observed would have been the result of

relatively low temperature D/H fractionation

16 of 26


1.5–2 Ga ago, during the initial formation and

seawater hydrothermal alteration of such oceanic

crust, prior to subsequent subduction and recy-

cling through the mantle. Justification of this

model must rely on Hutton’s principle of unifor-

mitarianism, by considering the effect on dD and

H2O caused by modern seawater-derived hydra-

tion of oceanic crust formed at present-day ridge

axes and its subsequent dehydration at modern

subduction zones. The average extent of hydration

of oceanic crust by seawater (dD = 0%) has been

estimated to be 1–2 wt% H2O [Ito et al., 1983;

Peacock, 1990], whereas the average H2O content

of fresh upper mantle-derived N-MORB glasses is

17 of 26

11MgO (%)

10987654

-35

-40

-45

-50

-55

-60

-65

-30

EN113 30D-1g

EN113 4D-1g

EN113 26D-1g

20.0

19.5

19.0

18.5

18.0

17.5

3.0

2.0

1.0

(La/

Sm

) n

EN113 4D-1g

EN113 26D-1gEN113 30D-1g

ESC east of 109 W

ESC west of 109 W

EMP east rift

EMP west riftand EPR


δD‰

(sm

ow)

206

Pb

/204

Pb

Figure 10. Covariation of dD, 206Pb/204Pb, and La/Sm (normalized to chondrites) with MgO. MgO is taken as anindicator of extent of fractional crystallization and/or partial melting. Note that dD varies in a similar manner as the206Pb/204Pb, which is not affected by fractional crystallization or partial melting and reflects mantle source variation.Fields define distinct regions with similar 206Pb/204Pb ratios. Pb isotope and rare earth data source as in Figure 2.Symbols are as defined in Figure 1 and lines with bars as in Figure 6.


0.1–0.2% (i.e., 10 times less), with a dD = �70%to �90% [e.g., Kyser and O’Neil, 1984]. Thus the

dD of water in contact or in equilibrium with

hydrated minerals near a mid-ocean ridge is esti-

mated to be close to modern seawater (i.e., 0%).

At equilibrium, the dD values of hydrous minerals

would then have to be displaced by an amount

D(mineral-water) = dDmineral � dDwater � 1000lna(min-

(mineral-water) (where the fractionation factor a(min-

(mineral-water)) = (D/H)mineral/ (D/H)waterspecific to each phase). A compilation of experi-

mentally determined hydrogen isotope fractionation

factors by hydration currently available, between

hydrous minerals and water are shown in Figure 11.

The range of D(mineral-water) is �80% to +20%, for

metamorphic grades involving greenschist, blues-

chist, and/or amphibolite facies conditions over a

temperature range of 100� to 1000�C and pressure

range of 1 bar to 2 Kb, thus within the temperature

and pressure range of crustal hydration from near

mid-ocean ridges to onset of subduction. The com-

pilation indicates that the D(mineral-water) variations

are also a complex function of temperature, pres-

sure, and composition. Where polybaric measure-

ments are available, D(mineral-water) generally

increases with increasing pressure [Mineev and

Grinenko, 1996; Driesner, 1997; Xu and Zheng,

1999], whereas D(mineral-water) variations with tem-

perature is either positive or negative depending on

the hydrated mineral type. In general, however, at

the temperatures expected during hydrothermal

alteration near the ridge crest (i.e., 300�–600�C),the common minerals of serpentine, hornblende,

chlorite, and epidote would incorporate hydrogen

with dD values ranging between �70% and 0%.

These equilibrium and mass balance constraints

might leave circulating hydrothermal water slightly

enriched in deuterium, as is indeed observed at

hydrothermal vents (dDSMOW = +2.5%) [Craig

et al., 1980; Welhan and Craig, 1983; Kawahata

et al., 1987]. In confirmation, observed values in

hydrothermally altered oceanic crust obtained from

dredging or drilling generally range from �20% to

�60% (see Figure 11 and references in caption).

[31] Can the high D/H ratio imprinted during

hydrothermal alteration at the ridge crest survive

the recycling process? The affirmative requires that

no significant D/H fractionation takes place during

the slab recycling processes in the mantle, includ-

ing dehydration during subduction and phase trans-

formations throughout the transition zone on the

way down or on the way up as a buoyant mantle

plume. If the recycling lithosphere remains rela-

tively closed but for water escape, the bulk D/H

ratio of the recycling slab need not change drasti-

cally. It may be that dehydration of the crust alters

the D/H ratio of residual water but the evidence

(albeit limited) suggests this fractionation is not

significant. The fractionation factor a in mineral-

water systems decreases with increasing pressure

[Mineev and Grinenko, 1996; Driesner, 1997; Xu

and Zheng, 1999], thus apparently minimizing D/H

fractionation at the depths where dehydration takes

place. The dD in H2O contained in back arc basin

basalts (BABB) range from �46% to �30%[Poreda, 1985; Hochstaedter et al., 1990]. These

basalts are derived from the mantle wedge by

partial melting triggered from fluid released from

dehydration of the subducting slab. The BABB dDvalues are consistent with and indeed suggest a

seawater origin. Since the D/H values of BABB

(fluid) are not significantly different from those of

the downgoing hydrothermally altered rocks

(�20% to �60%), then a mass balance argument

requires that the residual D/H values of water,

possibly left behind in hydrous and nominally

anhydrous minerals, must be of a similar range,

that is, similar to that found in basalts derived from

the Salas y Gomez mantle plume (�36%) and

significantly higher than normal upper mantle

range of �90% to �60%.

[32] This hypothesis is highly unconstrained for

lack of hydrous or nominally anhydrous mineral-

water D/H fractionation experiments at pressure

greater than 4 Kb. Bell and Ihinger [2000] have

shown on theoretical grounds and observed stretch-

ing frequencies of OH� bonds in mantle minerals,

that very low dD values are predicted in nominally

anhydrous minerals relative to hydrous minerals

such as phlogopite. The predicted (Dbspinel-phlogopite)

is of the order of �35%. Conversely, little D/H

fractionation may result between hydrous minerals

with similar OH� bond characteristics such as for

example between serpentine and phase A, a dense

18 of 26


hydrous magnesian silicate (DHMS). Regardless of

the distribution of water and D/H among phases,

water is likely to be homogenized during decom-

pression melting of the mantle plume. Small-scale

heterogeneities between mineral phases or separate

lithologic units in the mantle plume thus become

irrelevant when plume magmas are sampled.

[33] Finally, it is legitimate to ask if the recycling

model based on D/H is consistent with the Salas y

19 of 26

-80

0 2 4 6 8 10 12

-60

-40

-20

0

20

1000

700

500

400

300

200

100

25-100

hornblendechloritebruciteepidoteserpentinezoisite

1-2Kb

1bar

-80

-60

-40

-20

0

-100

Salas y Gomez plume

Tem

pera

ture

rang

e of

hydr

othe

rmal

alte

ratio

n

range of dredgedand drilled

altered oceanic crust

unaltered

oceanic crust

Temperature range in subductedslab from 3 to 22 GPa

subductionhot slab

cold slab

hydro-thermalalteration

T˚C

P (

GP

a)

0

10

20

5000 1000 1500

106/T2

1000

lnα

(min

eral

-wat

er)

Temperature (˚C)

4Kb

1bar

0.5Kb

2Kb

1bar

mea

sure

d δD

‰(s

mow

)

basaltic

Figure 11. Experimental hydrous mineral-water D/H fractionation factors (1000lna(mineral-water)) versus temperature(T and 1/T 2), for hydrous minerals common to basaltic systems affected by hydrothermal alteration at ridge ordehydration during subduction. The fractionation factors are a complex function of temperature, composition, andpressure. Vertical gray regions show expected temperature ranges for region of alteration in the lower crust nearspreading centers [Kawahata et al., 1987; Stakes, 1991; Zheng et al., 1998; Putlitz et al., 2000] and for the expectedtemperature range experienced by subducting slabs (see inset) [Staudigel and King, 1992]. Arrows with pressurelabels show effect of pressure on 1000lna(mineral-water) for reactions shown [Driesner, 1997]. Data compiled fromWenner and Taylor [1973], Suzuoki and Epstein [1976], Sakai and Tsutsumi [1978], Graham et al. [1980] Grahamet al. [1984], Satake and Matsuo [1984] Mineev and Grinenko [1996], Vennemann and O’Neil [1996], Driesner[1997], Chacko et al. [1999], and Xu and Zheng [1999]. Right-hand axis scale shows corresponding range of dDobserved. Observed dD values in metabasalts and metagabbros can be directly compared to 1000lna(mineral-water) �D(mineral-water), by assuming to first order that the dD value of the hydrothermal water is equal to SMOW = 0%. Therange of altered oceanic crust observed from dredged or drilled rocks agree well with fractionation expected underequilibrium conditions. Data for the measured range of dD in dredged and drilled altered crust are from Sheppard andEpstein [1970],Wenner and Taylor [1973], Satake and Matsuda [1979], Stakes and O’Neil [1982], and Stakes [1991].Note that the dD values measured for the Salas y Gomez mantle plume can be reasonably explained by fractionationduring hydrothermal alteration by seawater.


Gomez mantle plume 87Sr/86Sr ratio of 0.7035

[Fontignie and Schilling, 1991; Kingsley et al.,

1997]. Indeed, a three-stage discontinuous evolu-

tion model can readily generate this value, using

reasonable assumptions for the parameters

involved and listed in Table 4. Namely, these

are (1) mean parent/daughter ratios, (2) age of

fractionation of the continental crust, the depleted

upper mantle, and the recycled oceanic litho-

spheric slab (consisting of an upper basaltic and

lower gabbroic crust and residual harzburgite),

and (3) the 87Sr/86Sr seawater evolution curve of

Veizer and Compston [1976]. The model is illus-

trated in Figure 12 and detailed in the caption as

well as in Table 4. It is similar to the recycled

crust model of Hart and Staudigel [1989] for the

source of OIB. From 4.56 to 2.5 Ga the mantle

evolves with bulk earth Sr isotope systematics

[Allegre et al., 1983a]. The mean age of depletion

of the upper mantle by continental crust extraction

is assumed to be 2.5 Ga. At 1.5 Ga the oceanic

crust/lithosphere precursor of the Salas y Gomez

plume is formed at a ridge by partial melting. The

upper oceanic crust (UOC) is immediately hydro-

thermally altered by seawater and dehydrated

during its subduction, over a timescale of less

than 100 Myr. The 87Rb/86Sr fractionations for the

lower crust (LOC) and the harzburgite restite

formation assumes Hart et al.’s [1999] Rb, Sr,

and isotopic values recommended in their Table 7

and a batch melting model as defined in the

caption of our Figure 12. The 87Rb/86Sr increase

for the hydrothermal alteration of the UOC fol-

lows that of modern conditions modeled by Hart

and Staudigel [1989]. Fractionation effects during

subduction are omitted. The 87Sr/86Sr similar to

20 of 26

Table 4. Three Stage Sr Evolution Model With Recycling of Oceanic Crust

Model Aa Model Ba

Begin first stage primitiveb

mantle formation4.56Ga 4.56Ga

Begin second stagec

extraction of continental crust2.5Ga 3.5Ga

Begin third stage, oceanic crustformation/alteration

1.5Ga 2.0Ga

Seawater 87Sr/86Sr 0.705 0.704

AlteredUOC

AlteredLOC

Harzburgite(Hbz)

AlteredUOC

AlteredLOC

Harzburgite(Hbz)

Third stage 87Rb/86Srd 0.223 0.006 0.0013 0.223 0.006 0.0013Alteration mixe

%Sr(sw):%Sr(basalt)28:72 2:98 0:100 28:72 2:98 0:100

Beginning 87Sr/86Sr for third stage 0.7030 0.7024 0.7023 0.7021 0.7014 0.7013Final 87Sr/86Sr: recycled in plume 0.7078 0.7025 0.7023 0.7085 0.7016 0.7014Mixing proportions, (%)f 6.4 18.6 75 6.4 18.6 75

Model A Model B

Present day 87Sr/86Srg 0.7036 0.7030of evolved, mixed lithologies

(SyG = Salas y Gomezplume source)

(SyG = 0.7035) (SyG = 0.7035)

aModel A and Model B are for different ages for beginning of the second and third stages (see Figure 12).

bFor the first stage, all models evolve from BABI (0.69899) at 4.56Ga and 87Rb/86Sr = BSE = 0.0892 [Allegre et al., 1983a].

cFor the second stage (post continental crust extraction) mantle evolves with 87Rb/86Sr = DMM = 0.0456 [Allegre et al., 1983b].

dThird stage 87Rb/86Sr ratios from Hart et al. [1999]. Harzburgite 87Rb/86Sr based on estimated ratio in restite using batch partial melting of 14%

and concentration in the melt = sum of Rb and Sr in the basalt + gabbro (see Hart et al. [1999] values).eSr mixing proportions at time of hydrothermal alteration of the crust from Hart and Staudigel [1989] and Hart et al. [1999].

fMixing proportions are (1.8 parts), lower crust (5.2 parts) (i.e., 7 km of crust in proportions of 0.24 and 0.73 [e.g., Rait, 1963]), and depleted

harzburgite lithosphere (21 parts (km) [e.g., Oxburgh and Parmentier, 1977]).gSr isotope ratio resulting from mixing of evolved upper crust. CSr (UOC)M(UOC) I(UOC) + CSr (LOC)M(LOC) I(LOC) + CSr (Hbz)M(Hbz) I(Hbz) = CSr (Mix)

M(Mix)I(Mix), where CSr is the concentration of Sr in the UOC, LOC, and Hbz or 117 ppm, 158 ppm, and 1 ppm, respectively (e.g., Hart et al. [1999]and see text for estimation of Hbz); I(UOC), I(LOC), and I(Hbz) are the modeled present day isotope ratios.


that of the Salas y Gomez plume are readily

achieved by mixing the three recycled lithologic

units at any time after subduction, and prior

surfacing, using mixing proportions analogous to

1.8 km basalt, 5.2 km gabbro, and 21 km harz-

burgite [Hill, 1957; Raitt, 1963; Oxburgh and

Parmentier, 1977]. The model is not unique.

Sensitivity tests indicate that these proportions

are not critical nor is the degree of partial meting

and timing of the Rb/Sr fractionations chosen, as

for example shown in Figure 12b (see the caption

of Figure 12 and Table 4).

[34] The recycling model requires a 87Rb/86Rb for

the Salas y Gomez plume of 0.05. This would

require that the observed 87Rb/86Sr of the ESC

basalts be produced by 1% partial melting (assum-

ing an accumulated fractional melting model

[Shaw, 1970]) (Kingsley et al., manuscript in prep-

aration, 2002). This degree of melting is very low

compared to 5–8% usually considered to produce

alkali basalts. The problem is not uncommon, as

most OIB have a time integrated 87Rb/86Sr lower

than actually observed, even once corrected for

partial melting. Two stage metasomatic melt

enrichments have been proposed to alleviate this

discrepancy [e.g., Schilling et al., 1980].

5.2. Entrainment of Primordial H2O Fromthe Lower Mantle

[35] The high 3He/4He ratios observed in some

mantle plume-derived basaltic glasses relative to

the N-MORB average of R/Ra 8.5 ± 1.4 [Kurz,

1991; McDougall and Honda, 1998], such as those

found in Hawaii, Iceland, and Reunion, have com-

monly been attributed to entrainment of a high3He/4He, more primitive, and undegassed compo-

nent of the lower mantle by mantle plumes rising

from the core/mantle boundary [Farley and Craig,

1992; Farley et al., 1992;Hauri et al., 1994; Allegre

et al., 1995; Hanan and Graham, 1996]. Basalt

glasses from the EMP-ESC also contain high

(3He/4He)R/Ra reaching 17 on the Ahu lava field

[Poreda et al., 1993b, 1997]. Thus by analogy with3He, it is possible that the lower mantle may also

contain some primordial deuterium-rich H2O (or

other hydrogen compounds) which would have

been stored from the early accretion stage of the

Earth. C1 chondrites have dD values ranging from

+500 to +2000% [Yang and Epstein, 1983].

Although not very well constrained, entrainment

and dilution of such material by a rising mantle

plume from the core-mantle thermal boundary layer

could conceivably account for the high D/H ratio of

the Salas y Gomez mantle plume, provided that the

entrained water is subsequently well mixed with the

recycled component during the plume ascent, or its

bending once reaching the Nazca plate lithosphere.

In fact, it has also been suggested that a significant

chondritic water reservoir may exist in the lower

mantle bound in stable DHMS minerals [Abe and

Matsui, 1986; Ahrens, 1989]. Thus it is not possible

to rule out the prospect that the high D/H and H2O

content observed in the Salas y Gomez mantle

plume may have a similar origin from the partly

undegassed lower mantle. However, if as suggested

by Javoy [1997], the lower mantle is composed of

enstatite chondritic material with a dD value

�490% [Yang and Epstein, 1983], then the model

of entrainment of primitive water from the lower

mantle would not apply.

[36] In summary, we emphasize that the two pos-

sible variations just discussed for the high D/H, and3He/4He of the Salas y Gomez mantle plume model

need not be mutually exclusive. It could involve

recycling of old hydrothermally altered oceanic

crust (and associated gabbroic and harzburgitic

sections) with entrainment of primordial water

from the lower mantle.

[37] Fluid dynamic models of mantle entrainment

by rising mantle plumes [Griffiths and Campbell,

1991a, 1991b; Hauri et al., 1994] have been pro-

posed to explain spatial geochemical zonation over

hot spots, such as for the Galapagos [White and

Hofmann, 1978; White, 1979; Geist et al., 1999],

Marquesas [Duncan et al., 1986], Afar [Schilling

et al., 1992], Hawaii [Hauri, 1996; Dixon and

Clague, 2001]. In the case of the Salas y Gomez

mantle plume, if such a model were to be invoked to

account for the along EMP-ESC gradients observed

in radiogenic Pb isotopes, H2O, and D/H ratios, and

more irregular 3He/4He ratios, some very ad hoc

condition of melting and mixing at shallow depth

along the sublithospheric channel would have to be

advanced. This is particularly true considering that

21 of 26


22 of 26

0.708

0.710

0.706

0.704

0.702

0.700

0.698

87S

r/86

Sr

0 1 2Time(Ga)

3 4

0.708

0.710

0.706

0.704

0.702

0.700

0.698

87S

r/86

Sr

0 1 2Time(Ga)

3 4

LOCHarzburgite

UOC

SeawaterSyG

Stage 1Stage 2Stage 3

a

mix

ing

87Rb/86Sr = 0.006

87Rb/ 86

Sr = 0.22

DM evolution

BSE evolution

BSE evolution

OC

ext

ract

ion

CC

ext

ract

ion

and

alte

ratio

n

LOCHarzburgite

UOC

Seawater

SyG

Stage 1Stage 2Stage 3

b

mix

ing

87Rb/86Sr = 0.006

87Rb/ 86

Sr = 0.22

87Rb/86Sr= 0.0013 DM evolution

OC

ext

ract

ion

CC

ext

ract

ion

and

alte

ratio

n

PM

form

atio

nP

M fo

rmat

ion

87Rb/86Sr= 0.0013

Figure 12. Evolution diagram for a three-stage Sr isotope evolution model. Figure 12a depicts model A of Table 4.First stage: 4.56 Ga to 2.5 Ga initial 87Sr/86Sr of 0.69899 and 87Rb/86Sr ratio of BSE = 0.0892 [Allegre et al., 1983b].Second stage: the continental crust is extracted from the upper mantle and the 87Rb/86Sr ratio becomes that of thedepleted upper mantle = 0.0456 [Allegre et al., 1983b]. Third stage: at 1.5 Ga formation of the oceanic lithosphereoccurs, along with hydrothermal alteration by seawater circulation. The 87Rb/86Sr ratio of UOC is 0.223 (as measuredin DSDP site 417/418 altered basalts by Hart and Staudigel [1989]). The 87Sr/86Sr of the altered UOC is assumed tobe a 28%:72% alteration mix of Sr derived from seawater and fresh basalt [Hart and Staudigel, 1989] using a87Sr/86Sr = 0.705 for seawater at that time [Veizer and Compston, 1976] and the upper mantle 87Sr/86Sr at that time.The 87Rb/86Sr ratio of the LOC = 0.006 is assumed to be as measured by [Hart et al., 1999] in DSDP site 735 alteredolivine gabbros. The 87Sr/86Sr ratio of the LOC is based on a mix of 2% seawater Sr: 98% upper mantle Sr ascalculated from Hart et al. [1999, Table 7], instead of 28%:72% for the UOC. The 87Sr/86Sr of the seawater and uppermantle is the same as for the UOC case. The Hbz 87Rb/86Sr ratio is estimated by 14% batch partial melting of uppermantle, using the Rb and Sr contents from the combined fresh basalt and gabbro given by Hart et al. [1999, Table 7].Using 10% or 20% melting does not affect the results. After alteration, the three sections of the lithosphere aresubducted, stored, mixed, and then erupted as a result of a rising mantle plume at present. The results of mixing thethree lithologies in proportions based on present-day crustal thickness analogues (see text and Table 4 caption) areconsistent with the 87Sr/86Sr in the Salas y Gomez (SyG) mantle plume source (0.7035) (see arrow in diagram).The model results are insensitive to the relative amount of harzburgite due to its low Sr concentration (�1ppm).Figure 12b shows the results of model B in Table 4 and is a test of the sensitivity of the model to different mean ageassumed for continental crustal extraction (3.5 Ga) and oceanic crust formation (2.0 Ga). The resulting 87Sr/86Sr in therecycled mantle plume is 0.7030 relative to the Salas y Gomez value of 0.7035.


the 3He/4He variation along the EMP/ESC is not

fully coupled with that of Pb, Sr and D/H isotopes

ratios and H2O [Poreda et al., 1993b, 1997]. Such

modeling is beyond the scope of this paper.

[38] As noted, no simple correlation is found

between 3He/4He and D/H, Pb, Nd, and Sr isotope

ratios in the region [Poreda et al., 1993a, 1993b,

1997] nor for that matter between ‘‘high’’ and

‘‘low’’ 3He/4He hotspots and EM-1, EM-2 and

HIMU Pb, Nd, Sr, Hf, and Os isotope compositional

types worldwide. In fact, the high D/H found on the

Mohns ridge in the vicinity of the Jan Mayen hot

spot are accompanied by low 3He/4He ratios relative

to N-MORB [Poreda et al., 1986].

6. Conclusions

1. Recent, shallow depth magma emplacement

fractionation effects cannot be the primary causes

of the observed dD, H2O, and Pb isotope

variations along the EMP-ESC and their strong

covariations.

2. The close geographical coherence and covar-

iation of dD and H2O with Pb isotope ratios as well

as with ratios of highly incompatible elements

relative to H2O (e.g., Th/H2O) are fully consistent

with binary mixing of water and deuterium-rich

Salas y Gomez plume with the surrounding water

and deuterium-depleted upper mantle.

3. No definite conclusions can be reached as to

the ultimate source and origin of the high D/H and

H2O content of the Salas y Gomez mantle plume.

The two variations in the plume model, invoking

recycling of hydrothermally altered and dehydrated

oceanic crust generated 1–2 Ga ago, with or

without entrainment of primordial water from the

lower mantle, remain equally viable working

hypotheses at this point.

4. Whether the high dD ratios observed for the

Salas y Gomez mantle plume are reflecting

‘‘recycled ancient seawater’’ or ‘‘entrainment of

deuterium-rich primordial water’’ from the lower

mantle, or a mixture of both, the implication arising

from this study is that there exists one or more

mantle minerals capable of harboring a significant

amount of deuterium-rich water at mantle pressures

where sinking slabs are stored as a source of mantle

plumes, that is, at the depth of the transition zone if

it is proved to be a thermal boundary layer, or more

likely at the core/mantle boundary.

Acknowledgments

[39] This work was supported in part by National Science

Foundation grants to J.-G.S. (OCE 9626924) and J.E.D. (OCE

9530373). Comments by editors Bill White and Yoshiyuki

Tatsumi and reviewers Tetsu Kogiso and Catherine Chauvel

were very helpful and greatly appreciated.

References

Abe, Y., and T. Matsui, Early evolution of the Earth: Accre-

tion, atmosphere formation, and thermal history, Proc. Lunar

Planet. Sci. Conf. 16th, J. Geophys. Res., 91 suppl., E291–

E302, 1986.

Agrinier, P., R. Hekinian, D. Bideau, and M. Javoy, O and H

stable isotope compositions of oceanic crust and upper man-

tle rocks exposed in the Hess Deep near the Galapagos triple

junction, Earth Planet. Sci. Lett., 136, 183–196, 1995.

Ahrens, T. J., Water storage in the mantle, Nature, 342, 122–

123, 1989.

Allegre, C. J., S. R. Hart, and J.-F. Minster, Chemical structure

and evolution of the mantle and continents determined by

inversion of Nd and Sr isotopic data, I, Theoretical methods,

Earth Planet. Sci. Lett., 66, 177–190, 1983a.

Allegre, C. J., S. R. Hart, and J.-F. Minster, Chemical structure

and evolution of the mantle and continents determined by

inversion of Nd and Sr isotopic data, II, Numerical experi-

ments and discussion, Earth Planet. Sci. Lett., 66, 191–213,

1983b.

Allegre, C. J., M. Moreira, and T. Staudacher, 4He/3He disper-

sion and mantle convection, Geophys. Res. Lett., 22, 2325–

2328, 1995.

Bell, D. R., and P. D. Ihinger, The isotopic composition of

hydrogen in nominally anhydrous mantle minerals, Geo-

chim. Cosmochim. Acta, 64, 2109–2118, 2000.

Broecker, W. S., and V. M. Oversby, Chemical Equilibrium in

the Earth, 318 pp., McGraw-Hill, New York, 1971.

Byers, C. D., M. O. Garcia, and D. W. Muenow, Volatiles in

basaltic glasses from the East Pacific Rise at 21�N: Implica-

tions for MORB sources and submarine lava flow morphol-

ogy, Earth Planet. Sci. Lett., 79, 9–20, 1986.

Chacko, T., L. R. Riciputi, D. R. Cole, and J. Horita, A new

technique for determining equilibrium hydrogen isotope

fractionation factors using the ion microprobe: Application

to the epidote-water system, Geochim. Cosmochim. Acta, 63,

1–10, 1999.

Chase, C. G., Ocean island Pb: Two-stage histories and mantle

evolution, Earth Planet. Sci. Lett., 52, 277–284, 1981.

Chaussidon, M., S. M. F. Sheppard, and A. Michard, Hydro-

gen, sulphur and neodymium isotope variations in the mantle

beneath the EPR at 12�500N, Geochem. Soc. Spec. Publ., 3,325–337, 1991.

23 of 26


Cheng, Q. C., J. D. Macdougall, and P. Zhu, Isotopic con-

straints on the Easter seamount chain source, Contrib.

Mineral. Petrol., 135, 225–233, 1999.

Craig, H., and J. E. Lupton, Primordial neon, helium and hy-

drogen in oceanic basalts, Earth Planet. Sci. Lett., 31, 369–

385, 1976.

Craig, H., J. A. Welhan, K. Kim, R. Poreda, J. E. Lupton, and

Anonymous, Geochemical studies of the 21 degrees N EPR

hydrothermal fluids, Eos Trans. AGU, 61, 992, 1980.

Danyushevsky, L. V., T. J. Falloon, T. J. Sobolev, A. J. Craw-

ford, M. Carroll, and R. C. Price, The H2O content of basalt

glasses from southwest Pacific backarc basins, Earth Planet.

Sci. Lett., 117, 347–362, 1993.

Danyushevsky, L. V., W. M. Eggins, T. J. Falloon, and D. M.

Christie, H2O abundance in depleted to moderately enriched

mid-ocean ridge magmas, Part I, Incompatible behaviour,

implications for mantle storage, and origin of regional varia-

tions, J. Petrol., 41, 1329–1364, 2000.

Delaney, J. R., D. W. Muenow, and D. G. Graham, Abun-

dance and distribution of water, carbon, and sulfur in the

glassy rims of submarine pillow basalts, Geochim. Cosmo-

chim. Acta, 42, 581–594, 1978.

Deloule, E., F. Albarede, and S. M. F. Sheppard, Hydrogen

isotope heterogeneities in the mantle from ion probe analysis

of amphiboles from ultramafic rocks, Earth Planet. Sci. Lett.,

105, 543–553, 1991.

Dixon, J. E., and D. A. Clague, Volatiles in basaltic glasses

from Loihi Seamount, Hawaii: Evidence for a relatively dry

plume component, J. Petrol., 42, 627–654, 2001.

Dixon, J. E., and E. M. Stolper, An experimental study of

water and carbon dioxide solubilities in mid-ocean ridge

basaltic liquids, Part II, Applications to degassing, J. Petrol.,

36, 1633–1646, 1995.

Dixon, J. E., E. M. Stolper, and J. R. Delaney, Infrared spectro-

scopic measurements of CO2 and H2O in the Juan de Fuca

Ridge basaltic glasses, Earth Planet. Sci. Lett., 90, 87–104,

1988.

Dixon, J. E., D. A. Clague, P. Wallace, and R. Poreda, Volatiles

in alkalic basalts from the North Arch volcanic field, Hawaii:

Extensive degassing of deep submarine-erupted alkalic series

lavas, J. Petrol., 38, 911–939, 1997.

Driesner, T., The effect of pressure on deuterium-hydrogen

fractionation in high-temperature water, Science, 277,

791–794, 1997.

Duncan, R. A., M. T. McCulloch, H. G. Barsczus, and D. R.

Nelson, Plume versus lithospheric sources for melts at Ua

Pou, Marquesas Islands, Nature, 322, 534–538, 1986.

Farley, K. A., and H. Craig, Mantle plumes and mantle

sources, Science, 258, 821–822, 1992.

Farley, K. A., J. H. Natland, and H. Craig, Binary mixing of

enriched and undegassed (primitive?) mantle components

(He, Sr, Nd, Pb) in Samoan lavas, Earth Planet. Sci. Lett.,

111, 183–199, 1992.

Fontignie, D., and J.-G. Schilling, 87Sr/86Sr and REE varia-

tions along the Easter Microplate boundaries (South Pacific):

Application of multivariate statistical analyses to ridge seg-

mentation, Chem. Geol., 89, 209–241, 1991.

Geist, D., W. White, T. Naumann, and R. Reynolds, Illegiti-

mate magmas of the Galapagos; insights into mantle mixing

and magma transport, Geology, 27, 1103–1106, 1999.

Ghiorso, M. S., and R. O. Sack, Chemical mass transfer in

magmatic processes, IV, A revised and internally consistent

thermodynamic model for the interpolation and extrapolation

of liquid-solid equilibria in magmatic systems at elevated

temperatures and pressures, Contrib. Mineral. Petrol., 119,

197–212, 1995.

Graham, C. M., S. M. F. Sheppard, and T. H. E. Heaton,

Experimental hydrogen isotope studies, I, Systematics of

hydrogen isotope fractionation in the systems epidote -

H2O, zoisite - H2O, and AlO(OH) - H2O, Geochim. Cosmo-

chim. Acta, 44, 353–364, 1980.

Graham, C. M., R. S. Harmon, and S. M. F. Sheppard, Experi-

mental hydrogen isotope studies: Hydrogen isotope ex-

change between ampibole and water, Am. Mineral., 69,

128–138, 1984.

Green, D. H., and A. E. Ringwood, The stability fields of

aluminous pyroxene peridotite and garnet peridotite and their

relevance in upper mantle structure, Earth Planet. Sci. Lett.,

3, 151–160, 1967.

Griffiths, R. W., and I. H. Campbell, Interaction of mantle

plume heads with the Earth’s surface and onset of small-scale

convection, J. Geophys. Res., 96, 18,295–18,310, 1991a.

Griffiths, R. W., and I. H. Campbell, On the dynamics of long-

lived plume conduits in the convecting mantle, Earth Planet.

Sci. Lett., 103, 214–227, 1991b.

Hanan, B. B., and D. W. Graham, Lead and helium isotope

evidence from oceanic basalts for a common deep source of

mantle plumes, Science, 272, 991–995, 1996.

Hanan, B. B., and J.-G. Schilling, Easter microplate evolution:

Pb isotope evidence, J. Geophys. Res., 94, 7432–7448,

1989.

Hart, S. R., A large-scale isotope anomaly in the Southern

Hemisphere mantle, Nature, 309, 753–757, 1984.

Hart, S. R., Heterogeneous mantle domains: Signatures, gen-

esis and mixing chronologies, Earth Planet. Sci. Lett., 90,

273–296, 1988.

Hart, S. R., J. Blusztajn, H. Dick, P. S. Meyer, and K. Mue-

helenbachs, The fingerprint of seawater circulation in a 500-

meter section of ocean crust gabbros, Geochim. Cosmochim.

Acta, 63, 4059–4080, 1999.

Hart, S. R., and H. Staudigel, Isotopic characterization and

identification of recycled components, in Crust/mantle Re-

cycling at Convergence Zones, edited by S. R. Hart and

L. Gulen, pp. 15–28, Kluwer Acad., Norwell, Mass., 1989.

Hauri, E. H., Major-element variability in the Hawaiian mantle

plume, Nature, 382, 415–419, 1996.

Hauri, E. H., J. A. Whitehead, and S. R. Hart, Fluid dynamic

and geochemical aspects of entrainment in mantle plumes,

J. Geophys. Res., 99, 24,275–24,300, 1994.

Hill, M. N., Recent geophysical exploration of the ocean floor,

Phys. Chem. Earth, 2, 129–163, 1957.

Hirose, K., and T. Kawamoto, Hydrous partial melting of Iher-

zolite at 1 GPa: The effect of H2O on the genesis of basaltic

magmas, Earth Planet. Sci. Lett., 133, 463–473, 1995.

Hirth, G., and D. L. Kohlstedt, Water in the oceanic upper

mantle: Implications for rheology, melt extraction and the

24 of 26


evolution of the lithosphere, Earth Planet. Sci., Lett., 144,

93–108, 1996.

Hochstaedter, A. G., J. B. Gill, M. Kusakabe, S. Newman,

M. Pringle, B. Taylor, and P. Fryer, Volcanism in the Sumisu

Rift, I, Major element, volatile, and stable isotope geochem-

istry, Earth Planet. Sci. Lett., 100, 179–194, 1990.

Hofmann, A. W., and W. M. White, Mantle plumes from an-

cient oceanic crust, Earth Planet. Sci. Lett., 56, 421–436,

1982.

Ito, E., D. M. Harris, and A. T. Anderson Jr., Alteration of

oceanic crust and geologic cycling of chlorine and water,

Geochim. Cosmochim. Acta, 47, 1613–1624, 1983.

Javoy, M., The major volatile elements of the Earth: Their

origin, behavior and fate, Geophys. Res. Lett., 24, 177–

180, 1997.

Kamenetsky, V. S., R. A. Binns, J. B. Gemmell, A. J. Craw-

ford, T. P. Mernagh, R. Maas, and D. Steele, Parental basaltic

melts and fluids in eastern Manus backarc basin; implica-

tions for hydrothermal mineralisation, Earth Planet. Sci.

Lett., 184, 685–702, 2001.

Kawahata, H., M. Kusakabe, and Y. Kikuchi, Strontium, oxy-

gen, and hydrogen isotope geochemistry of hydrothermally

altered and weathered rocks in DSDP Hole 504B, Costa Rica

Rift, Earth Planet. Sci. Lett., 85, 343–355, 1987.

Kingsley, R. H., Carbon dioxide and water in Mid-Atlantic

Ridge basalt glasses, MS thesis, Univ. of Rhode Island, Nar-

ragansett, 1989.

Kingsley, R. H., and J.-G. Schilling, Plume-ridge interaction in

the Easter-Salas y Gomez seamount chain - Easter micro-

plate system: Pb isotope evidence, J. Geophys. Res., 103,

24,159–24,177, 1998.

Kingsley, R. H., J.-G. Schilling, and D. Fontignie, The Salas y

Gomez Plume Connection to the EPR: Pb, Sr, and Nd Iso-

tope Evidence, Eos Trans. AGU, 78(46), Fall Meet. Suppl.,

F688, 1997.

Kruse, S. E., Z. J. Liu, D. F. Naar, and R. A. Duncan, Effective

elastic thickness of the lithosphere along the Easter seamount

chain, J. Geophys. Res., 102, 27,305–27,317, 1997.

Kurz, M. D., Noble gas isotopes in oceanic basalts: controver-

sial constraints on mantle models, in Applications of Radio-

genic Isotope Systems to Problems in Geology, edited by

L. Heaman and J. N. Ludden, pp. 259–286, Mineral. Assoc.

of Can. Short Course Handbook, Toronto, Can., 1991.

Kyser, T. K., and J. R. O’Neil, Hydrogen isotope systematics

of submarine basalts, Geochim. Cosmochim. Acta, 48,

2123–2133, 1984.

Liu, Z. J., The origin and evolution of the Easter Seamount

Chain, Ph. D. thesis, Univ. of South Florida, St. Petersburg,

1996.

McDougall, I., and M. Honda, Primordial solar noble-gas com-

ponent in the Earth: Consequences for the origin and evolu-

tion of the Earth and its atmosphere, in The Earth’s Mantle

Composition, Structure, and Evolution, edited by I. Jackson,

pp. 159–190, Cambridge Univ. Press, New York, 1998.

Michael, P. J., The concentration, behavior and storage of H2O

in the suboceanic upper mantle: Implications for mantle me-

tasomatism, Geochim. Cosmochim. Acta, 52, 555–566,

1988.

Michael, P. J., Regionally distinctive sources of depleted

MORB: Evidence from trace elements and H2O, Earth Pla-

net. Sci. Lett., 131, 301–320, 1995.

Mineev, S. D., and V. A. Grinenko, The pressure influence on

hydrogen isotope fractionation in the serpentine-water sys-

tem, J. Conference Abstr., 1, 404, 1996.

Moore, J. G., and J.-G. Schilling, Vesicles, water and sulphur

in Reykjanes Ridge basalts, Contrib. Mineral. Petrol., 41,

105–118, 1973.

Oxburgh, E. R., and E. M. Parmentier, Compositional and

density stratification in oceanic lithosphere: Causes and con-

sequences, J. Geol. Soc. London, 133, 343–355, 1977.

Pan, Y., and R. Batiza, Major element chemistry of volcanic

glasses from the Easter Seamount Chain: Constraints on

melting conditions in the plume channel, J. Geophys. Res.,

103, 5287–5304, 1998.

Peacock, S. M., Fluid processes in subduction zones, Science,

248, 329–337, 1990.

Pineau, F., and M. Javoy, The volatile record of a ‘‘popping’’

rock from the Mid-Atlantic Ridge at 15�N: Concentrationsand isotopic compositions, Terra Cognita, 6, 191, 1986.

Pineau, F., and M. Javoy, Strong degassing at ridge crests; the

behaviour of dissolved carbon and water in basalt glasses at

14 degrees N, Mid-Atlantic Ridge, Earth Planet. Sci. Lett.,

123, 179–198, 1994.

Poreda, R., Helium-3 and deuterium in back-arc basalts: Lau

Basin and Mariana Trough, Earth Planet. Sci. Lett., 73,

244–254, 1985.

Poreda, R., J.-G. Schilling, and H. Craig, Helium and hydro-

gen isotopes in ocean-ridge basalts north and south of Ice-

land, Earth Planet. Sci. Lett., 78, 1–17, 1986.

Poreda, R. J., J.-G. Schilling, R. Batiza, and D. Naar, Geo-

chemistry of volcanism along the Easter Seamount Chain,

Eos Trans. AGU, 74(43), Fall Meet. Suppl., F672, 1993a.

Poreda, R. J., J.-G. Schilling, and H. Craig, Helium isotope

ratios in Easter Microplate basalts, Earth Planet. Sci. Lett.,

119, 319–329, 1993b.

Poreda, R. J., R. H. Kingsley, J.-G. Schilling, J. E. Dixon, and

P. Swart, Plume-Ridge interactions along the Easter Sea-

mount Chain: Mixing and degassing of volatiles, Eos Trans.

AGU, 78(46), Fall Meet. Suppl, F688, 1997.

Putlitz, B., A. Matthews, and J. W. Valley, Oxygen and hydro-

gen isotope study of high-pressure metagabbros and metaba-

salts (Cyclades, Greece): Implications for the subduction of

oceanic crust, Contrib. Mineral. Petrol., 138, 114–126,

2000.

Raitt, R. W., The crustal rocks, in The sea, edited by M. N.

Hill, pp. 85–102, Wiley-Interscience, New York, 1963.

Rappaport, Y., D. F. Naar, C. C. Barton, Z. J. Liu, and R. N.

Hey, Morphology and distribution of seamounts surrounding

Easter Island, J. Geophys. Res., 102, 24,713–24,728, 1997.

Reisberg, L., A. Zindler, F. Marcantonio, W. White, D. Wy-

man, and B. Weaver, Os isotope systematics in ocean island

basalts, Earth Planet. Sci. Lett., 120, 149–167, 1993.

Sakai, H., and M. Tsutsumi, D/H fractionation factors between

serpentine and water at 100 to 500�C and 2000 bar water

pressure and the D/H ratios of natural serpentines, Earth

Planet. Sci. Lett., 40, 231–242, 1978.

25 of 26


Satake, H., and J.-I. Matsuda, Strontium and hydrogen isotope

geochemistry of fresh and metabasalt dredged from the Mid-

Atlantic Ridge, Contrib. Mineral. Petrol., 70, 153–157,

1979.

Satake, H., and S. Matsuo, Hydrogen isotopic fractionation

factor between brucite and water in the temperature range

from 100 to 510�C, Contrib. Mineral. Petrol., 80, 19–24,

1984.

Schilling, J.-G., M. B. Bergeron, and R. Evans, Halogens in

the mantle beneath the North Atlantic, Philos. Trans. R. Soc.

London, Ser. A, 297, 147–178, 1980.

Schilling, J.-G., H. Sigurdsson, A. N. Davis, and R. N. Hey,

Easter Microplate evolution, Nature, 317, 325–331, 1985.

Schilling, J.-G., R. H. Kingsley, B. B. Hanan, and B. L.

McCully, Nd-Sr-Pb isotopic variations along the Gulf of

Aden: Evidence for Afar mantle plume-continental litho-

sphere interaction, J. Geophys. Res., 97, 10,927–10,966,

1992.

Shaw, D. M., Trace element fractionation during anatexis,

Geochim. Cosmochim. Acta, 34, 237–243, 1970.

Sheppard, S. M. F., and S. Epstein, D/H and 18O/16O ratios of

minerals of possible mantle or lower crustal origin, Earth

Planet. Sci. Lett., 9, 232–239, 1970.

Simons, K., Volatiles in basaltic glasses from the Easter-Salas

y Gomez Seamount Chain and Easter Microplate: Implica-

tions for geochemical cycling of volatile elements, M.S.

Thesis, Univ. of Miami, Miami, 2000.

Stakes, D. S., Oxygen and hydrogen isotope compositions of

oceanic plutonic rocks: high temperature deformation and

matamorphism of oceanic layer 3, in Stable Isotope Geo-

chemistry: A Tribute to Samuel Epstein, Geochem. Soc. Spec.

Publ.3, edited by H. P. Taylor, J. R. O’Neil, and I. R. Kaplan,

pp. 77–90, Geochem. Soc., San Antonio, Tex., 1991.

Stakes, D. S., and J. R. O’Neil, Mineralogy and stable isotope

geochemistry of hydrothermally altered oceanic rocks, Earth

Planet. Sci. Lett., 57, 285–304, 1982.

Staudigel, H., and S. D. King, Ultrafast subduction: The key to

slab recycling efficiency and mantle differentiation?, Earth

Planet. Sci. Lett., 109, 517–530, 1992.

Suzuoki, T., and S. Epstein, Hydrogen isotope fractionation

between OH-bearing minerals and water, Geochim. Cosmo-

chim. Acta, 40, 1229–1240, 1976.

Taylor, B. E., Magmatic volatiles: Isotopic variation of C, H,

and S, in Stable Isotopes in High Temperature Geological

Processes, Reviews in Mineralogy, edited by J. W. Valley, H.

P. Taylor, and J. R. O’Neil, pp. 185–226, Mineral. Soc. of

Am., Chelsea, Mich., 1986.

Taylor, B. E., J. C. Eichelberger, and H. R. Westrich, Hydrogen

isotopic evidence of rhyolitic magma degassing during shal-

low intrusion and eruption, Nature, 306, 541–545, 1983.

Taylor, H. P. J., Oxygen and hydrogen isotope evidence for

large-scale circulation and interaction between ground waters

and igneous intrusions, with particular reference to the San

Juan volcanic field, Colorado, in Geochemical Transport

and Kinetics, edited by A. W. Hofmann, B. J. Giletti, H. S.

Yode, Jr., and R. A. Yund, pp. 299–324, Carnegie Inst. of

Washington, Washington, D.C., 1974.

Taylor, H. P. J., Water/rock interactions and the origin of H2O

in granitic batholiths, J. Geol. Soc. London, 133, 509–558,

1977.

Taylor, H. P., and S. M. F. Sheppard, Igneous Rocks, I, Pro-

cesses of isotopic fractionation and isotope systematics, in

Stable Isotopes in High Temperature Geological Processes,

Reviews in Mineralogy, edited by J. W. Valley, H. P. Taylor,

and J. R. O’Neil, pp. 227–271, Mineral. Soc. of Am., Chel-

sea, Mich., 1986.

Veizer, J., and W. Compston, 87Sr/86Sr in Precambrian carbo-

nates as an index of crustal evolution, Geochim. Cosmochim.

Acta, 40, 905–914, 1976.

Vennemann, T. W., and J. R. O’Neil, Hydrogen isotope ex-

change reactions between hydrous minerals and molecular

hydrogen, I, A new approach for the determination of hydro-

gen isotope fractionation at moderate temperatures, Geo-

chim. Cosmochim. Acta, 60, 2437–2451, 1996.

Welhan, J. A., and H. Craig, Methane, hydrogen and helium in

hydrothermal fluids at 21 degrees N on the East Pacific Rise,

in Hydrothermal Processes at Seafloor Spreading Centers,

NATO Conference Series; IV, Marine Sciences, edited by

P. A. Rona, K. Bostroem, L. Laubier, and K. L. Smith Jr.,

pp. 391–409, Plenum Press, New York, 1983.

Wendt, J. I., M. Regelous, Y. Niu, R. Hekinian, and K. D.

Collerson, Geochemistry of lavas from the Garrett transform

fault: Insights into mantle heterogeneity beneath the eastern

Pacific, Earth Planet. Sci. Lett., 173, 271–284, 1999.

Wenner, D. B., and H. P. Taylor, Oxygen and hydrogen isotope

studies of the serpentinization of ultramafic rocks in oceanic

environments and continental ophiolite complexes, Am. J.

Sci., 273, 207–239, 1973.

White, W. M., Pb isotope geochemistry of the Galapagos Is-

lands, Year Book Carnegie Inst. Wash., 78, 331–335, 1979.

White, W. M., and A. W. Hofmann, Geochemistry of the Ga-

lapagos Islands: Implications for mantle dynamics and evo-

lution, Year Book Carnegie Inst. Wash., 77, 596–606, 1978.

Xu, B.-L., and Y.-F. Zheng, Experimental studies of oxygen

and hydrogen isotope fractionations between precipitated

brucite and water at low temperatures, Geochim. Cosmo-

chim. Acta, 63, 2009–2018, 1999.

Yang, J., and S. Epstein, Interstellar organic matter in meteor-

ites, Geochim. Cosmochim. Acta, 47, 2199–2216, 1983.

Zheng, Y.-F., B. Fu, L. Yiliang, Y. Xiao, and S. Li, Oxygen and

hydrogen isotope geochemistry of ultrahigh-pressure eclo-

gites from the Dabie Mountains and the Sulu terrane, Earth

Planet. Sci. Lett., 155, 113–129, 1998.

Zindler, A., E. Jagoutz, and S. Goldstein, Nd, Sr, and Pb iso-

topic systematics in a three-component mantle: a new per-

spective, Nature, 298, 519–523, 1982.

26 of 26


d/h ratios in basalt glasses from the salas y gomez mantle plume interacting with the east pacific...

Documents